Methods in Enzymology Chromatin and Chromatin Remodeling Enzymes

Preface
A central challenge of the post-genomic era is to understand how the 30,000 to

40,000 unique genes in the human genome are selectively expressed or silenced
to coordinate cellular growth and differentiation. The packaging of eukaryotic
genomes in a complex of DNA, histones, and nonhistone proteins called
chromatin provides a surprisingly sophisticated system that plays a critical role
in controlling the flow of genetic information. This packaging system has
evolved to index our genomes such that certain genes become readily acces-
sible to the transcription machinery, while other genes are reversibly silenced.
Moreover, chromatin-based mechanisms of gene regulation, often involving
domains of covalent modifications of DNA and histones, can be inherited from
one generation to the next. The heritability of chromatin states in the absence
of DNA mutation has contributed greatly to the current excitement in the field
of epigenetics.
The past 5 years have witnessed an explosion of new research on chroma-
tin biology and biochemistry. Chromatin structure and function are now widely
recognized as being critical to regulating gene expression, maintaining genomic
stability, and ensuring faithful chromosome transmission. Moreover, links be-
tween chromatin metabolism and disease are beginning to emerge. The identi-
fication of altered DNA methylation and histone acetylase activity in human
cancers, the use of histone deacetylase inhibitors in the treatment of leukemia,
and the tumor suppressor activities of ATP-dependent chromatin remodeling
enzymes are examples that likely represent just the tip of the iceberg.
As such, the field is attracting new investigators who enter with little
firsthand experience with the standard assays used to dissect chromatin struc-
ture and function. In addition, even seasoned veterans are overwhelmed by the
rapid introduction of new chromatin technologies. Accordingly, we sought to
bring together a useful ‘‘go-to’’ set of chromatin-based methods that would
update and complement two previous publications in this series, Volume 170
(Nucleosomes) and Volume 304 (Chromatin). While many of the classic proto-
cols in those volumes remain as timely now as when they were written, it is our
hope the present series will fill in the gaps for the next several years.
This 3-volume set of Methods in Enzymology provides nearly one hundred
procedures covering the full range of tools—bioinformatics, structural biology,
biophysics, biochemistry, genetics, and cell biology—employed in chromatin
research. Volume 375 includes a histone database, methods for preparation of
histones, histone variants, modified histones and defined chromatin segments,
xv
xvi preface
protocols for nucleosome reconstitution and analysis, and cytological methods

for imaging chromatin functions in vivo. Volume 376 includes electron micro-
scopy and biophysical protocols for visualizing chromatin and detecting chro-
matin interactions, enzymological assays for histone modifying enzymes, and
immunochemical protocols for the in situ detection of histone modifications
and chromatin proteins. Volume 377 includes genetic assays of histones and
chromatin regulators, methods for the preparation and analysis of histone
modifying and ATP-dependent chromatin remodeling enzymes, and assays
for transcription and DNA repair on chromatin templates. We are exceedingly
grateful to the very large number of colleagues representing the field’s leading
laboratories, who have taken the time and effort to make their technical
expertise available in this series.
Finally, we wish to take the opportunity to remember Vincent Allfrey,
Andrei Mirzabekov, Harold Weintraub, Abraham Worcel, and especially Alan
Wolffe, co-editor of Volume 304 (Chromatin). All of these individuals had key
roles in shaping the chromatin field into what it is today.
C. David Allis
Carl Wu
Editors’ Note: Additional methods can be found in Methods in Enzymology,

Vol. 371 (RNA Polymerases and Associated Factors, Part D) Section III
Chromatin, Sankar L. Adhya and Susan Garges, Editors.
METHODS IN ENZYMOLOGY
EDITORS-IN-CHIEF
John N. Abelson Melvin I. Simon

DIVISION OF BIOLOGY
CALIFORNIA INSTITUTE OF TECHNOLOGY
PASADENA, CALIFORNIA
FOUNDING EDITORS
Sidney P. Colowick and Nathan O. Kaplan

Contributors to Volume 376
Article numbers are in parentheses and following the names of contributors.
Affiliations listed are current.
Rhoda M. Alani (12), Department of Juliana Callaghan (10), Department of

Oncology, Johns Hopkins University Biochemistry, University of Cambridge,
School of Medicine, Baltimore, Maryland Cambridge CB2 1GA, United Kingdom
21218
Marek Cebrat (12), Department of
Francisco Asturias (4), Department of Cell Pharmacology and Molecular Sciences,
Biology, The Scripps Research Institute, Johns Hopkins University School of
La Jolla, California 92037 Medicine, Baltimore, Maryland 21218
Andrew J. Bannister (18), Wellcome Julie Chaumeil (27), Mammalian Develop-
Trust/Cancer Research, United Kingdom mental Epigenetics Group, UMR
Institute and Department of Pathology, 218-Nuclear Dynamics and Genome Plas-
University of Cambridge, Cambridge ticity, Curie Institute-Research Section,
CB2 1QR, United Kingdom 75248 Paris, Cedex 05-France
P. B. Becker (1), Adolf Butenandt Institut, Dina Chaya (24), Cell and Developmental
Lehrstuhl für Molekularbiologie, Schil- Biology Program, Fox Chase Cancer
lerstr. 44, 80336 Munich, Germany Center, Philadelphia, Pennsylvania 19111
Martin L. Bennink (6), Biophysical Tech- Peter Cheung (15), Department of Med-
niques Group and MESAþ Research ical Biophysics, University of Toronto,
Institute, Department of Science Technol- Ontario Cancer Institute, Toronto,
ogy, University of Twente, 7500 AE Ontario M5G 2M9, Canada
Enschede, The Netherlands
J. Chin (1), Department of Biochemistry,
Bradley E. Bernstein (23), Department of Northwestern University, Molecular Bio-
Chemistry and Chemical Biology, Harvard logy and Cell Biology, Evanston, Illinois
University, Cambridge, Massachusetts 60208-3500
02138
David N. Ciccone (22), Department of
Margie T. Borra (11), Department of Bio- Molecular Biology, Massachusetts Gen-
chemistry and Molecular Biology, Oregon eral Hospital, Boston, Massachusetts
Health and Science University, Portland, 02114
Oregon 97239
Philip A. Cole (12), Department of
Brent Brower-Toland (5), Biology De- Pharmacology and Molecular Sciences,
partment, Washington University in St. Johns Hopkins University School of
Louis, St. Louis, Missouri 63130 Medicine, Baltimore, Maryland 21218
Michael Bustin (14), Protein Section, Carlos Cordon-Cardo (13), Division of
National Cancer Institute, National Insti- Molecular Pathology, Memorial Sloan
tutes of Health, Bethesda, Maryland Kettering Cancer Center, New York,
20892 New York 10021
ix
x contributors to volume 376
Carolyn A. Craig (25), Biology Depart- Jeffrey C. Hansen (2), Department of

ment, Washington University in St. Louis, Biochemistry and Molecular Biology,
St. Louis, Missouri 63130 Colorado State University, Fort Collins,
Colorado 80523
John M. Denu (11), Department of
Biochemistry and Molecular Biology, Edith Heard (27), Mammalian De-
Oregon Health and Science University, velopmental Epigenetics Group, UMR
Portland, Oregon 97239 218-Nuclear Dynamics and Genome Plas-
ticity, Curie Institute-Research Section,
Meghann K. Devlin (12), Department of 75248 Paris, Cedex 05, France
Oncology, Johns Hopkins University
School of Medicine, Baltimore, Maryland Rachel A. Horowitz-Scherer (3),
21218 Department of Biology, University of
Massachusetts, Amherst, Massachusetts
Marija Drobnjak (13), Division of 01003
Molecular Pathology, Memorial Sloan
Kettering Cancer Center, New York, Emily L. Humphrey (23), Department of
New York 10021 Chemistry and Chemical Biology, Harvard
University, Cambridge, Massachusetts
Brian Dynlacht (20), Department of
02138
Pathology, New York University School
of Medicine, New York, New York 10016 Steven A. Jacobs (9), Department of Bio-
Sarah C. R. Elgin (25), Biology De- chemistry and Molecular Genetics, Uni-
partment, Washington University in St. versity of Virginia, Charlottesville,
Louis, St. Louis, Missouri 63130 Virginia 22908
Chukwudi Ezeokonkwo (4), Department Thomas Jenuwein (16), Research Institute

of Cell Biology, The Scripps Research of Molecular Pathology (IMP), The
Institute, La Jolla, California 92037 ViennaBiocenter,Vienna,A-1030,Austria
Peggy Farnham (21), McArdle Laboratory Monika Kauer (16), Research Institute of
for Cancer Research, University of Wis- Molecular Pathology (IMP), The Vienna
consin, Madison, Wisconsin 53706 Biocenter, Vienna, A-1030, Austria
Wolfgang Fischle (9), Department of W. Kevin Kelly (13), Genitourinary On-

Biochemistry and Molecular Genetics, cology Service and Department of Medi-
University of Virginia, Charlottesville, cine, Memorial Sloan Kettering Cancer
Virginia 22908 Center, New York, New York 10021
Fred K. Friedman (14), Laboratory of Me- Sepideh Khorasanizadeh (9), Depart-

tabolism, National Cancer Institute, ment of Biochemistry and Molecular Gen-
National Institutes of Health, Bethesda, etics, University of Virginia,
Maryland 20892 Charlottesville, Virginia 22908
Philippe T. Georgel (2), Department of Roger D. Kornberg (4), Department of

Biological Sciences, Marshall University, Structural Biology, Stanford University
Huntington, West Virginia 25755 School of Medicine, Stanford, California
94305
Michael Grunstein (19), Department of
Biological Chemistry, School of Medicine Tony Kouzarides (18), Wellcome Trust/
and Molecular Biology Institute, Univer- Cancer Research, United Kingdom Insti-
sity of California, Los Angeles, Los tute, University of Cambridge, Cambridge
Angeles, California 90095 CB2 1QR, United Kingdom
contributors to volume 376 xi
Siavash K. Kurdistani (19), Department Shiraz Mujtaba (8), Department of Physi-

of Biological Chemistry, University of ology and Biophysics, Structural Biology
California, Los Angeles School of Medi- Program, Mt. Sinai School of Medicine,
cine and Molecular Biology Institute, Los New York University, New York, New
Angeles, California 90095 York 10029
G. Längst (1), Adolf Butenandt Institut, Alexey G. Murzin (10), MRC Centre for
Lehrstuhl für Molekularbiologie, Schil- Protein Engineering, Cambridge, CB2
lerstr. 44, 80336 Munich, Germany 2QH United Kingdom
Ernest Laue (10), Department of Bio- Natalia V. Murzina (10), Department of
chemistry, University of Cambridge, Biochemistry, University of Cambridge,
Cambridge CB2 1GA, United Kingdom Cambridge CB2 1GA, United Kingdom
Sanford H. Leuba (6), Department of Peter R. Nielsen (10), Department of
Cell Biology and Physiology, University Biochemistry, University of Cambridge,
of Pittsburgh School of Medicine, Hill- Cambridge CB2 1GA, United Kingdom
man Cancer Center, UPCI Research
Pavilion, Pittsburgh, Pennsylvania Kenichi Nishioka (17), Department of De-
15213-1863 velopmental Genetics, National Institute
of Genetics, Shizuoka, Japan, 411-8540
Yuhong Li (25), University of Iowa, De-
partment of Biochemistry, Iowa City, Matthew J. Oberley (21), McArdle
Iowa 52242 Laboratory for Cancer Research, Univer-
sity of Wisconsin, Madison, Wisconsin
John Lis (26), Cornell University, Ithaca, 53706
New York 14853
Marjorie A. Oettinger (22), Department
Chih Long Liu (23), Department of Chem- of Molecular Biology, Massachusetts
istry and Chemical Biology, Harvard Uni- General Hospital, Boston, Massachusetts
versity, Cambridge, Massachusetts 02138 02114
Yahli Lorch (4), Department of Structural Ikuhiro Okamoto (27), Mammalian Devel-
Biology, Stanford University School of opmental Epigenetics Group, UMR 218 –
Medicine, Stanford, California 94305 Nuclear Dynamics and Genome Plasti-
Paul A. Marks (13), Cell Biology Pro- city, Curie Institute-Research Section,
gram, Memorial Sloan-Kettering Cancer 75248 Paris, Cedex 05, France
Center, New York, New York 10021
Susanne Opravil (16), Research Institute of
Ronen Marmorstein (7), Structural Biol- Molecular Pathology (IMP), The Vienna
ogy Program, The Wistar Institute, Biocenter, Vienna, A-1030, Austria
Philadelphia, Pennsylvania 19104-4268
Barbara Panning (28), Department of
Karl Mechtler (16), Research Institute of Biochemistry and Biophysics, University
Molecular Pathology (IMP), The Vienna of California, San Francisco, San Fran-
Biocenter, Vienna, A-1030, Austria cisco, California 94143-0448
Katrina B. Morshead (22), Massachusetts Laura Perez-Burgos (16), Research Insti-
General Hospital, Department of tute of Molecular Pathology (IMP), The
MolecularBiology,Boston, Massachusetts Vienna Biocenter, Vienna, A-1030,
02114 Austria
xii contributors to volume 376
Antoine H. F. M. Peters (16), Research Paul R. Thompson (12), Department of

Institute of Molecular Pathology (IMP), Pharmacology and Molecular Sciences,
The Vienna Biocenter, Vienna, A-1030 Johns Hopkins University School of
Austria Medicine, Baltimore, Maryland 21218
Danny Reinberg (17), Department of Biol- Julissa Tsao (21), Microarray Centre,
ogy, Howard Hughes Medical Institute, University Health Network, Toronto,
University of Medicine and Dentistry Ontario M5G 2C4, Canada
of New Jersey, Piscataway, NJ
08854-5635 Lori L. Wallrath (25), Department of
Biochemistry, University of Iowa, Iowa
Bing Ren (20), San Diego Branch and De- City, Iowa 52242
partment of Cellular and Molecular
Medicine, Ludwig Institute for Cancer Michelle D. Wang (5), Department of
Research, University of California, San Physics, Laboratory of Atomic and Solid
Diego School of Medicine, La Jolla, State Physics, Cornell University, Ithaca,
California 92093-0653 New York 14853
Victoria M. Richon (13), Discovery Biol- Ling Wang (12), Department of Pharma-
ogy, Aton Pharma, Inc., Tarrytown, New cology and Molecular Sciences, Johns
York 10591 Hopkins University School of Medicine,
Baltimore, Maryland 21218
Richard C. Robinson (14), Laboratory of
Metabolism, National Institutes of Janis K. Werner (26), Cornell University,
Health, National Cancer Institute, Ithaca, New York 14853
Bethesda, Maryland 20892 Jon Widom (1), Northwestern University,
Daniel Robyr (19), Department of Biol- Department of Biochemistry, Molecular
ogical Chemistry, University of California, Biology and Cell Biology, Evanston,
Los Angeles, School of Medicine and Illinois 60208-3500
Molecular Biology Institute, Los Angeles, Christopher L. Woodcock (3), Department
California 90095 of Biology, University of Massachusetts,
Kavitha Sarma (17), Department of Biol- Amherst, Massachusetts 01003
ogy, Howard Hughes Medical Institute,
Patrick Yau (21), Microarray Centre, Uni-
University of Medicine and Dentistry
versity Health Network, Toronto, Ontario
of New Jersey, Piscataway, NJ
M5G 2C4, Canada
08854-5635
Ken Zaret (24), Cell and Developmental
Stuart Schreiber (23), Department of
Biology Program, W. W. Smith Chair in
Chemistry and Chemical Biology, Harvard
Cancer Research, Fox Chase Cancer
University, Cambridge, Massachusetts
Center, Philadelphia, Pennsylvania 19111
02138
Yujun Zheng (12), Department of
BrianE.Schwartz(26),CornellUniversity,
Pharmacology and Molecular Sciences,
Ithaca, New York 14853
Johns Hopkins University School of
J. Paul Secrist (13), Discovery Biology, Medicine, Baltimore, Maryland 21218
Aton Pharma, Inc., Tarrytown, New York
10591 Ming-Ming Zhou (8), Structural Biology
Program, Department of Physiology and
Gena E. Stephens (25), Biology Depart- Biophysics, Mt. Sinai School of Medicine,
ment, Washington University in St. Louis, New York University, New York, New
St. Louis, Missouri 63130 York 10029-6574
contributors to volume 376 xiii
Xianbo Zhou (13), Discovery Biology, Aton Jordanka Zlatanova (6), Department of
Pharma, Inc., Tarrytown, New York Chemical and Biological Sciences and En-
10591 gineering, Polytechnic University, Brook-
lyn, New York 11201
[1] fluorescence anisotropy assays 3
[1] Fluorescence Anisotropy Assays for Analysis of

ISWI-DNA and ISWI-Nucleosome Interactions
By J. Chin, G. Längst, P. B. Becker, and J. Widom
Fluorescence anisotropy is a rapid, sensitive, and quantitative technique

that is well suited to the analysis of protein-protein and protein-DNA inter-
actions in solution. Fluorescence anisotropy is a measure of the depolarization
of emitted fluorescence intensity obtained after excitation by a polarized
light source, and depends directly on the relative rate of fluorescence emis-
sion versus the rate of tumbling in solution. The concept is simple: if a
fluorescent molecule (or, more typically, a molecule to which a fluorescent
probe has been attached) tumbles slowly in solution relative to the lifetime
of fluorescence emission, then the light emitted in response to polarized
excitation will remain highly polarized. However, if the molecules tumble
rapidly in comparison to the emission lifetime, then, prior to emitting, they
will have tumbled sufficiently so as to have ‘‘forgotten’’ their orientation at
the moment of excitation, thus depolarizing (randomizing the polarization
of) the emitted light.
Fluorescence anisotropy is applicable for analysis of macromolecular
interactions because there is a good match between typical fluorescence
lifetimes and typical macromolecular tumbling times. For approximately
spherical molecules, the tumbling time scales as the molecular volume, that
is, as the molecular weight. Thus, binding of an unlabeled macromolecule
can make a significant change to the tumbling time of the molecule to
which the fluorescent probe is attached, and hence to the measured anisot-
ropy. For the studies described in the following, we utilize DNA molecules
labeled at one end with the fluorescent dye fluorescein (these DNA mol-
ecules may be ‘‘naked DNA’’ or they may be incorporated into nucleo-
somes), and we use fluorescence anisotropy to monitor the binding of the
Drosophila ISWI chromatin remodeling protein1–3 to the labeled DNA
or nucleosomes.
Fluorescence anisotropy is especially useful because of its high inherent
sensitivity. Dyes such as fluorescein allow quantitative analysis of emission
polarization from sub-nanomolar concentrations. Since dissociation con-
stants are typically nanomolar or greater, this allows experiments to be
1
T. Tsukiyama, C. Daniel, J. Tamkun, and C. Wu, Cell 83, 1021 (1995).
2
P. D. Varga-Weisz et al., Nature 388, 598 (1997).
3
G. Längst and P. B. Becker, J. Cell Sci. 114, 2561 (2001).
Copyright 2004, Elsevier Inc.

All rights reserved.
METHODS IN ENZYMOLOGY, VOL. 376 0076-6879/04 $35.00
4 chromatin proteins [1]
set up with the probe concentration Kd; consequently the free concentra-
tion of the added macromolecule (ISWI, in our case), which is generally
either difficult to measure or is completely unknown, will be approximately
equal to the total concentration, which can be definitively measured, thus
greatly simplifying the analysis of the binding measurements. Another im-
portant benefit of the sensitivity of the anisotropy measurement is that it
preserves precious reagents. Measurements can be made in small volumes,
and samples can be recovered and reused if desired.
Finally, as discussed later, the experiment can be carried out using inex-
pensive conventional fluorometers such as are found at most biochemical
or chemical research laboratories, or, alternatively, using an inexpensive
instrument specialized for the fluorescence anisotropy experiment.
Investigators planning to carry out such studies should study two par-
ticularly useful references, one on fluorescence theory and methodology
in general4 and one focused on fluorescence approaches to analysis of pro-
tein-DNA interactions in particular.5 These references nicely define and
explain the set of four fluorescence intensity measurements that go into a
single measurement of fluorescence anisotropy; we will not duplicate this
important topic here, but rather refer readers to these other sources.
Fluorescein-Labeled DNA
We use DNA sequences labeled with fluorescein attached at the 50 -end
through a C6 linker. Relatively short sequences are purchased as a pair
of complementary oligonucleotides, one containing 50 -fluorescein. These
are annealed, and the resulting duplex purified away from any remaining
single strand by reverse-phase HPLC on a Zorbax-10 column using a
gradient of 10–20% acetonitrille in 0.1 M triethanolamine-acetate, pH
7.0, 0.1 mM EDTA, developed over 10 min at 1 ml min1. When longer
sequences (e.g., nucleosome-length DNAs) are required, direct synthesis
is not practical. Instead we use preparative scale PCR, with one of
the two primers again containing 50 -fluorescein. The resulting PCR
product is purified by gel electrophoresis in 1% agarose gels with standard
TAE buffer, and extracted from the gel using Ultra-DA (Millipore) gel
extraction kits.
DNA concentrations are quantified by UV absorbance.
4
J. R. Lakowicz, ‘‘Principles of Fluorescence Spectroscopy,’’ 2nd Ed. Kluwer Academic/
Plenum Press, New York, 1999.
5
J. J. Hill and C. A. Royer, Meth. Enzymol. 278, 390 (1997).
Preparation of Nucleosomes
Nucleosomes are formed by salt gradient dialysis using purified histone
octamer and DNA, and the resulting nucleosomes are purified by sucrose
gradient ultracentrifugation, as described.6–8 We typically label a small
amount of the fluorescein-labeled DNA with [ 32] ATP (at the 50 end that
does not have a fluorescein) using T4 polynucleotide kinase to facilitate
following the sample throughout preparation and purification. Reconstitu-
tion reactions typically contain 300 ng of (32P, fluorescein) double-labeled
DNA, 15 g of fluorescein-only labeled DNA, 3 g of histone octamer, in a
300 l volume of 2.5 M NaCl, 0.5 TE (TE is 10 mM Tris, pH 8.0, 1 mM
Na3 EDTA) with 0.5 mM phenylmethylsulfonyl fluoride (PMSF), and
0.1 mM benzamidine (BZA) added as protease inhibitors. The reconstitu-
tion reactions are loaded onto 12 ml 5–30% sucrose gradients in 0.5 TE

and centrifuged at 4 in an SW41 rotor (Beckman) at 41,000 rpm for
22–24 h. (We aim for substoichiometric reconstitution of histone octamer
onto DNA, as this eliminates the possibility of overloading nucleosomes
with excess histones9 while providing useful diagnostics for the reconstitu-
tion and markers for the subsequent sucrose gradient purification.) Gradi-
ents are fractionated from the bottom in 0.5-ml fractions; fractions
containing nucleosomes are identified by scintillation counting, pooled
and exchanged into 0.5 TE buffer on Centricon-30 concentrators, and
analyzed by native polyacrylamide gel electrophoresis. Nucleosome con-
centrations are measured by UV absorbance at 260 nm. Reconstituted
nucleosomes are stored at concentrations of 50 nM or greater, on ice in
0.5 TE, and are used within 2 weeks.
Instrumentation and Technical Considerations

We use a conventional photon counting steady-state fluorometer (ISS
PC-1, L-format) with rotatable polarizers in the excitation and emission
paths. Alternatively, the Panvera Corporation markets a sensitive and rela-
tively inexpensive instrument dedicated specifically to fluorescence anisot-
ropy measurements. We generally increase the sensitivity of the PC-1 by
removing the emission monochrometer and use instead a set of optical filters
chosen to pass a desired broad band of fluorescence emission wavelengths,
as described later.
6
K. J. Polach and J. Widom, J. Mol. Biol. 254, 130 (1995).
7
P. T. Lowary and J. Widom, J. Mol. Biol. 276, 19 (1998).
8
J. D. Anderson, A. Thåström, and J. Widom, Mol. Cell. Biol. 22, 7147 (2002).
9
G. Voordouw and H. Eisenberg, Nature 273, 446 (1978).
Sample Cleanliness
As with any sensitive experimental method in analytical biochemistry,
care must be taken in certain matters to avoid potential pitfalls.
It goes without saying that both buffers and samples must be free from
significant fluorescent contaminants. Fluorescence from the buffer alone
and from unlabeled samples should be checked and shown to be negligible
in comparison to the fluorescence obtained at the desired concentration of
labeled sample. In our experience this has never proven to be a problem
using dilute buffers supplemented with approximately physiological con-
centrations of salts and Mg2þ and small amounts of glycerol; nevertheless,
it should be checked, especially in case of problems with the water or with
contaminated glass- or plastic-ware.
Scattered Light
Even when samples are free of contaminants, one must take care to
eliminate certain additional potential artifacts due to scattered light. Scat-
tered light is particularly problematic in anisotropy measurements because
it is generally perfectly polarized, and hence will systematically distort
measurements of fluorescence anisotropy from the sample.
Two chief types of scattered light need to be considered in anisotropy
measurements: elastic (Rayleigh) scattering and inelastic (chiefly Raman)
scattering. Elastic scattering is a process in which excitation light is scat-
tered in all directions, unshifted in wavelength, by interaction of the excita-
tion light with molecules in the sample. Even pure solvents scatter light
elastically. Such scattering is weak, yet may nevertheless be significant
in comparison to the faint fluorescence from a very dilute fluorophore.
Macromolecules in solution greatly increase the intensity of scattered light,
in proportion to their concentration and molecular weight. Solutions con-
taining a high molecular weight species such as nucleosomes can result in
a scattering intensity that greatly exceeds the intensity of fluorescence
emission from a dye attached to that same macromolecule.
If excitation and emission monochrometers were ‘‘perfect,’’ then elastic
scattering would present no problem: one could simply set the excitation and
emission monochrometers to different wavelengths (e.g., the excitation
and emission maxima, respectively), and there would be no leakage of scat-
tered excitation light through the emission monochrometer. In fact, how-
ever, the finite resolution of the monochrometers, together with optical
imperfections that allow light of colors outside the assumed bandpass to
pass through, albeit at reduced intensity, are such that there can be signifi-
cant excitation intensity at the color chosen for emission measurement,
even though these colors may differ by 30 nm or more. In fact, this leakage
can in practice be so great that, when combined with a relatively strong

elastic light scattering from a macromolecular sample, the intensity of
scattered excitation light reaching the emission detector may be significant
in comparison to the intensity of fluorescence.
Raman (inelastic) scattering occurs when excitation light is scattered
by solvent molecules (water, in biochemical applications) with concomitant
vibrational excitation of the solvent molecules. The Raman scattered light
is thus shifted in color toward the red relative to the excitation color by
an amount corresponding to the vibrational energy change. This is a fixed
amount in energy terms (3600 cm1 for water), but corresponds to a vary-
ing wavelength change because of the reciprocal relationship between
energy and wavelength (E ¼ hc/h ¼ Planck’s constant, c ¼ velocity of
light, ¼ wavelength). For excitation at 280 nm, the Raman peak occurs
at 311 nm, whereas for excitation at 490 nm (our typical choice for fluores-
cein), the Raman peak occurs at approximately 595 nm. The width of the
Raman peak will be identical to that of the excitation light (measured on
an energy axis, not on a wavelength axis). The Raman intensity is generally
low relative to elastic scattering, but may nevertheless become significant
when sample concentrations are low.
Excitation Path Filter

Both kinds of scattered light are readily eliminated with appropriate
optical filters, with or without the use of an emission monochrometer. We
use a bandpass interference filter in the excitation path, placed between the
excitation monochrometer and the sample, to eliminate any remaining
light at colors other than the desired excitation wavelength that happens
to pass through the excitation monochrometer. This filter is chosen such
that its wavelength of maximum transmission matches the excitation mono-
chrometer wavelength setting, and the bandwidth of the filter is chosen to
be comparable to that of the excitation monochrometer, so as to minimize
unnecessary loss of excitation intensity.
Emission Path Filters

The combination of excitation monochrometer and bandpass filter in
the excitation path together ensures that the excitation light is adequately
clean. There remains, however, the possibilities that either light elastically
scattered (at the excitation color) by the sample or Raman scattered exci-
tation light may make it through the emission path and be counted by the
emission detector.
We use a cut-on filter in the emission path to reject elastically scattered ex-
citation light. Such filters absorb or reflect short wavelengths, while passing
longer wavelengths with high transmittance, with a steep rise in transmit-

tance occurring over a relatively narrow wavelength range. In general,
colored glass filters work well for this purpose, and are available in a closely
spaced series of cut-on wavelength ranges. One picks a filter that has essen-
tially zero transmittance over the full bandpass of the excitation light, but
that has high transmittance over much of the width of the fluorescence
emission spectrum. In certain cases, if other constraints dictate that there
will be only small shifts in color between excitation and emission wave-
lengths, it can be beneficial to use specialized multilayer dielectric filters,
which can achieve much steeper cut-on characteristics.
Finally, one must remember to eliminate also the Raman scattered
light, taking into account its full spectral width. Depending on the excita-
tion wavelength and the fluorophore, the Raman scatter may be blue-
shifted or red-shifted relative to the fluorescence emission. Even if the
Raman scattering is superimposed on the fluorescence emission spectrum,
the fluorescence emission spectrum will generally be much broader than
the Raman band, so that it will be possible to choose filters that pass
either the blue-side or the red-side of the fluorescence emission while
rejecting the Raman.
The particular situation dictates the choice of filters to be used.
When the Raman scatter is to the blue of the fluorescence emission, cut-
on filter can be chosen to reject both elastic and Raman scattered light.
When the Raman band is to the red of the fluorescence, one may need to
supplement the cut-on filter with a cut-off filter. If an emission mono-
chrometer is used, this itself may serve effectively as a cut-off filter (be-
cause of the lower intensity of Raman scattering compared to elastic
scattering). Alternatively, or in addition, cut-off (short-pass) filters may
be used. The selection of colored glass cut-off filters is much less extensive
than for cut-on. In general, one will need to use specialized multilayer
dielectric filters instead.
Once a filter combination has been chosen (whether or not an emission
monochrometer will be used for the actual experiment), it is wise to
verify that the filter combination works as planned by recording fluores-
cence emission spectra of buffer alone, of unlabeled sample, and of
labeled sample (at the concentration that will be used for anisotropy
measurement), scanning the emission monochrometer from below the
excitation wavelength to above the fluorescence emission range and
Raman band. Both buffer and unlabeled sample spectra should show
negligible intensity, in comparison to the intensity from the labeled
sample, at all wavelengths that will be monitored during the anisotropy
measurement.
Filter Set for Use with Fluorescein

We find the following combination of filters to be highly effective for
use with fluorescein, whether an emission monochrometer is present or
not. We choose 490 nm as the center wavelength for excitation.
Excitation interference Coherent/Ealing #35-3482 interference filter,

filter center wavelength ¼ 490.0 nm, bandpass ¼7.3
nm (full width at half-maximum transmission)
Emission cut-on Coherent/Ealing #26-4333 OG-515
(long-pass) filter colored glass
Figure 1 shows excitation and emission spectra for fluorescein in panel

A, compared to the transmission characteristics of the three filters in panel
B. The excitation filter is well matched to the excitation maximum for
Fig. 1. Fluorescence spectra of fluorescein compared to transmission spectra of the optical

filters used for measurement of fluorescein anisotropy. (A) Fluorescence excitation and
emission spectra. Left-hand curve: fluorescein excitation spectrum, obtained monitoring
emission at ¼ 520 nM; 0.5 nM fluorescein-labeled oligonucleotide in a buffer containing
20 mM HEPES-KOH, pH 7.6, 80 mM KCl, 2 mM MgCl2, 1 mM DTT, 5% glycerol. Right-
hand curve: the corresponding emission spectrum, with excitation at ¼ 490 nm. (B) Long-
dashed curve: transmission characteristics of the 488.8 nm interference filter used in the
excitation path in conjunction with the excitation monochrometer. Solid and short-dashed
curves, transmission characteristics of the cut-on and bandpass filters, respectively, which are
used in the emission path, typically with no emission monochrometer. The excitation filter
matches the excitation maximum for fluorescein. The cut-on and leading (short wavelength)
edge of the bandpass filters both strongly reject any elastically scattered excitation light; the
falling (long wavelength) edge of the bandpass filter strongly rejects any Raman scattered
light, which is centered at 595 nm.
fluorescein. The colored glass cut-on filter rejects most of the excitation
bandpass, while passing the majority of the emission spectrum. The special-
ized cut-off filter nicely rejects any Raman scatter, which is centered
at 595 nm, while further strongly reducing any elastically scattered excita-
tion light. (The cut-off [bandpass] filter also strongly rejects elastically
scattered light on its own; however, the combination of the two filters gives
far better blocking at the excitation bandpass than either one on its own.
This improved scatter rejection is important in strongly scattering/weakly
fluorescing samples.)
The performance of this filter set can be appreciated from the emission
spectra in Fig. 2. Panel A shows the ability of the filter set to suppress both
elastically scattered light (>10,000-fold reduction) and Raman scattering to
undetectably low levels, while panel B shows that this huge reduction in
background scattering intensity comes at a modest cost (2- to 3-fold) in
the collectible intensity of fluorescence emission.
Fig. 2. Performance of the filter set for fluorescein anisotropy measurement. (A) Rejection
of elastically scattered light and Raman scatter. Emission spectra recorded from a scattering
solution (10 g/ml BSA in buffer), with excitation at 490 nm using the excitation mono-
chrometer plus the 488.8 nm interference filter. Solid line: no filters in the emission path. Long-
dashed line: 515 nm cut-on filter. Short-dashed line (essentially invisible): both 515-nm cut-on
plus bandpass filters. When no filters are used in the emission path, note the strong peak of
scattering intensity detected when emission ¼ excitation. Inset: same spectra plotted on a
500-fold more sensitive scale. The emission filter combination reduces the scattering signal by
>10,000-fold, rendering it immeasurably low. Similarly, there is negligible remaining intensity
at the Raman scattering wavelength (595 nm). (B) Fluorescence emission from fluorescein
with no filter in the emission path (solid black line), 515-nm cut-on filter only (grey line) or
both cut-on and bandpass filters (dashed line). The >10,000-fold reduction in scattering
intensity comes at a cost of only 2-fold in emission intensity at the emission maximum, and
only 3-fold in total emission intensity integrated over all wavelengths (such as could be
measured by a detector in the emission path if no emission monochrometer were present).
pH Sensitivity
Fluorescein has a pKa of 6.5, and its fluorescence intensity is depen-
dent on the charge state. Consequently, it is important to control the pH
of the solution, preferably at a pH 7.5, so that unanticipated small
changes in pH will affect the fluorescence intensity only negligibly.
Intensity Changes
Two additional phenomena concerning the sample fluorescence inten-
sity bear mention. First, fluorescein is sensitive to photobleaching; thus,
prolonged exposure of sample to excitation light, or even room light, can
significantly reduce the fluorescence intensity. Since anisotropy is an inher-
ent property, independent of concentration, modest amounts of sample
bleaching occurring during binding titrations need not be problematic, pro-
vided that there is negligible bleaching during the time required for any
given anisotropy measurement. This can be checked by monitoring the
total intensity during the four individual measurements that make up an
anisotropy measurement.
In certain cases, however, the binding of a protein to a fluorescein-
labeled DNA can affect the fluorescence intensity directly. For example,
the bound protein might happen to quench (or enhance) the fluorescence
quantum yield, or perhaps shift the emission color, so that the intensity
measured over a given wavelength range may increase or decrease. Such
effects invalidate the usual interpretation of the anisotropy changes. If
the intensity changes, the likelihood is that the fluorescence lifetime too
is changing; and if the lifetime changes, then the anisotropy is affected even
if molecular tumbling time were to remain constant. Consequently, it will
no longer be possible to assign a linear relationship between a measured
anisotropy change and a probability of binding site occupancy.
Actually, such cases can be a blessing in disguise: one can often monitor
the binding process simply by measuring the intensity change directly. In
any event, it is necessary to pay attention to the total intensity when carry-
ing out anisotropy measurements. Systematic changes in intensity that cor-
relate with binding invalidate the standard interpretation of the anisotropy
experiment.
Cuvettes
Samples used in fluorescence experiments may be precious, and one
may wish to minimize the volume of sample used. We routinely use
samples of 75–100 l in quartz ultramicro-cuvettes (Hellma, black walled,
#105.251-QS) having a sample chamber of 3 3 mm with a 5-mm tall
aperture, that is, a 45-l illuminated volume. These cuvettes are easy to fill
and clean, and fit in the 1.25-cm square cuvette holders that are standard in
most fluorometers. These cuvettes are available with the sample chamber
placed at various heights above the base of the cuvette. It is important to
pay attention to the height of the optical axis of the fluorometer, and to
make certain that the height of the sample chamber of the cuvette matches
that of the fluorometer.
Experimental Design for Binding Titrations

We use selected high-affinity nucleosome positioning sequences7,10,11 to
simultaneously provide a high degree of homogeneity in nucleosome
positioning while also enhancing the stability of the nucleosomes against dis-
sociation by mass action despite the dilute nucleosome concentrations used.
ISWI is an ATP-dependent nucleosome remodeling factor that induces
nucleosome sliding on nicked DNA.12 It is expressed in Escherichia coli
and purified by gel-filtration to near homogeneity.13 Non-specific inter-
actions of ISWI with DNA alone and with DNA at its entry into the nu-
cleosome have been described qualitatively.3 Fluorescence anisotropy
measurements permit a quantitative description of these interactions.
Binding buffer for ISWI-DNA or ISWI-nucleosome binding reactions
contains: 20 mM HEPES-KOH, pH 7.6, 80 mM KCl, 2 mM MgCl2, 5%
glycerol, 1 mM DTT, supplemented when desired with ATP, ADP, or
other nucleotides or analogs. We typically make up a distinct sample for
each ISWI to be investigated. This allows the DNA or nucleosomes to
remain constant as the ISWI is varied over a titration. Binding reactions
are 100 l final volume, typically with 1 nM fluorescein-labeled DNA or
5 nM nucleosomes, and the desired ISWI. ISWI protein is diluted into
binding buffer as appropriate, such that accurately measurable volumes
are added into the 100-l final binding reaction volumes. Samples are incu-
bated at room temperature for 30 min prior to measurement of fluores-
cence anisotropy to ensure that binding reactions are well equilibrated
(control studies show that binding appears to equilibrate essentially in-
stantaneously, as judged by the absence of further changes in anisotropy,
hence the 30-min equilibration time is more than sufficient). Samples
are placed in quartz ultramicro-cuvettes as described earlier. We use an
excitation wavelength of 490 nm, no emission monochrometer, and the
10
A. Thåström et al., J. Mol. Biol. 288, 213 (1999).
11
J. Widom, Q. Rev. Biophys. 34, 269 (2001).
12
G. Längst and P. B. Becker, Mol. Cell 8, 1085 (2001).
13
D. F. Corona et al., Mol. Cell 3, 239 (1999).
Fig. 3. Raw fluorescence anisotropy data from ISWI binding to naked DNA. Titration of a
5 nM 50 -fluorescein–labeled 35-bp long DNA with increasing concentrations of ISWI protein.
(See text for buffer conditions.) The curve superimposed on the data represents a fit to a
cooperative binding model.
combination of the cut-on and bandpass cut-off filter in the emission path
described previously.
We always include a sample prepared with no ISWI (ISWI ¼ 0) to es-
tablish the experimental ‘‘baseline’’ for each titration, and we extend the
titrations to sufficiently high ISWI to allow accurate determination of the
anisotropy corresponding to complete binding (complete occupancy of
binding sites by bound ISWI). Note that, whatever method is used to moni-
tor binding processes, it is important to carry titrations through the full
range of the binding process. A good practice is to use ‘‘direct’’ plots14 of
the measured signal (anisotropy, in our case) versus the titrant concentra-
tion (ISWI, in our case) plotted on a log scale to allow representation of the
wide range of titrant necessary to explore the full range from fraction
bound 0 (no binding) to fraction bound 1 (100% binding).
We use Kalaidagraph software to fit raw binding data to desired binding
models.
Results of Binding Experiments

Typical raw data resulting from such an experiment are shown in Fig. 3,
for a 50 -fluorescein–labeled 35-bp long DNA, used at 5 nM. Aficionados
of binding studies will recognize immediately from the raw data that the
14
I. M. Klotz, ‘‘Ligand-Receptor Energetics: A Guide for the Perplexed.’’ Wiley, New York,
1997.
binding curve as plotted in this manner is too ‘‘steep’’ to correspond to

simple 1:1 binding of ISWI to DNA, implying positive cooperativity in
the binding of ISWI to DNA. The titration midpoint for these given condi-
tions (EC50) is 30 nM. The concentration of fluorescein-labeled tracer is
small in comparison and hence may safely be neglected (or alternatively
quantitatively accounted for,14 yielding small corrections that convert
[ISWI]total to [ISWI]free). If binding were simple (1:1 ISWI:DNA complex,
noncooperative binding curve), then this measured EC50 would also be
the thermodynamic Kd.
In any experimental analysis of binding processes, it is helpful to
reduce the number of adjustable parameters in the curve fitting. In our an-
isotropy studies, we directly measure both the lower and upper ‘‘baselines’’
for the titrations, and hold these quantities fixed at their measured values
during the curve fitting procedure. We set the anisotropy for the 0-nM
ISWI baseline (the lower baseline for the curve fitting) equal to the average
of several measurements on the same DNA sample lacking any ISWI, and
we set the upper baseline equal to an average of the results for (replicate
measurements on) the last couple or few titration points, having taken care
to extend the titrations to the point at which any further increases in ISWI
do not result in any further significant changes in measured anisotropy. For
a given fixed assumed stoichiometry of the binding process, this leaves only
one free parameter—the apparent affinity or Kd—to be determined by
curve fitting. Alternatively, one may allow both the apparent affinity and
the cooperativity (or ‘‘molecularity’’) to be simultaneously fit.
Figure 4 shows results for a negative control, in which the DNA (18-bp
duplex, 50 -end labeled with fluorescein, 1 nM concentration) proves to be
too short to allow high affinity binding of the ISWI. Plainly, even the raw
data resulting from the anisotropy experiment can distinguish samples in
which binding does occur from samples in which it does not. The EC50
(Kd, if 1:1 ISWI:DNA complex) for ISWI binding to this DNA is 100 nM.
Figure 5 shows the results an experiment monitoring binding to 177-bp–
containing nucleosomes (again, 50 -end labeled with fluorescein, 5 nM
concentration). The raw data are rescaled along the ordinate to represent
the fraction of DNA with bound ISWI, simply by linearly rescaling the
measured anisotropies from 0 (experimentally measured lower baseline)
to 1 (measured upper baseline). The titration midpoint for this particular
dataset (EC50) is 15 nM, slightly lower (i.e., higher affinity) than for
the 35-mer DNA of Fig. 3. An average over many datasets (data not
shown) suggests that this small apparent difference in affinity between
DNA and nucleosome is not statistically significant. Evidently, ISWI binds
to nucleosomes with an affinity that is close to its affinity for long naked
DNA.
Fig. 4. Raw anisotropy data when no binding occurs. A 1 nM solution of a 50 -fluorescein–

labeled 18-bp long (double-stranded) DNA is titrated with increasing concentrations of ISWI
protein. ISWI has negligible affinity for such short DNAs.
Fig. 5. ISWI binding to a nucleosomal DNA. A 5-nM solution of a 50 -fluorescein–labeled

177-bp DNA, assembled into nucleosomes, is titrated with increasing concentration of ISWI.
Raw fluorescence anisotropy data are scaled to fraction bound (fraction of DNAs having
ISWI protein bound): anisotropy data obtained in the absence of any ISWI protein establish
the anisotropy for fractional occupancy ¼ 0; the averaged anisotropy from the highest several
titration points (where the signal appears to have plateaued) defines the fractional occupancy
¼ 1. The curve represents a least-squares fit to a cooperative binding model. The DNA
template used includes a 147-bp selected nucleosome positioning sequence together with
an additional 30 bp of DNA extending beyond one end; a single fluorescein is attached at the
50 -end of this 30-bp extension. The DNA is assembled into nucleosomes and purified as
described (see text).
Finally, in Fig. 6 we study the binding of ISWI to DNA (50 -fluorescein,

1 nM concentration) in the presence of 1 mM ADP. The titration midpoint
for this dataset (EC50) is 28 nM, very close to the measured 30 nM EC50 for
binding to naked DNA in the absence of nucleotide (Fig. 3). Evidently, the
Fig. 6. ISWI titration in the presence of added nucleotide. Titration of a 1 nM 50 -

fluorescein–labeled 150-bp long DNA with increasing concentrations of ISWI protein, in the
presence of 1 mM ADP. The raw anisotropy data are scaled to fraction bound. Nucleotides or
other cofactors or binding partners can be included in the reactions without difficulty.
affinity of ISWI for naked DNA is not influenced by the presence of high
concentrations of ADP. This figure highlights the utility of anisotropy
measurements to monitor binding in solutions containing other additives
such as nucleotides. In contrast, it is difficult or risky to carry out such stud-
ies using a gel electrophoretic mobility shift approach, for example, since
prohibitively costly amounts of nucleotide analogs might be required, and
moreover these compounds may actually electrophorese. In that case, their
concentrations around the complexes during a gel separation would be
undefined, rendering the experiments uninterpretable.
Conclusions
Fluorescence anisotropy is well known to be useful for analysis of pro-
tein-DNA interactions, and it seems likely to be particularly useful for
analysis of nucleosome remodeling factors because it is rapid, quantitative,
highly sensitive (conserving precious reagents), suitable for use in the pres-
ence of cofactors such as ATP, readily measured even during rapid kinetic
experiments, and very broadly applicable. It will allow analysis of the inter-
actions of remodeling factors or their individual proteins or domains with
any other species that can be specifically labeled.
[2] biophysical analysis of genomic loci 17
[2] Biophysical Analysis of Specific Genomic Loci

Assembled as Chromatin In Vivo
By Philippe T. Georgel and Jeffrey C. Hansen
Background
Over the last decade much progress has been made toward understand-
ing the effects of chromatin on nuclear functions. However, virtually all
previous studies of chromatin fiber organization in vivo have been re-
stricted to gathering information about the locations of nucleosomes, his-
tone post-translational modifications, regulatory DNA binding proteins,
and chromatin remodeling machines relative to specific functional DNA
elements, for example, promoters, origins of replications, repair sites.
Despite a vastly improved understanding of the composition and configur-
ation of functionally important genomic loci, very little is known about the
higher-order organization of these chromosomal regions in vivo. Biophys-
ical characterization of specific in vivo–assembled chromatin structures has
not been possible due to technical limitations. Consequently, our knowl-
edge of the functional effects of chromatin folding and higher-order struc-
ture has been obtained almost exclusively through use of in vitro model
systems that mimic the solution behavior of natural chromatin.1–4
Recently, we have adapted the technique of agarose multigel electro-
phoresis (AME)5–8 for analysis of the higher-order nucleoprotein structure
of specific genomic loci that have been isolated as native chromatin in
unfractionated low-salt nuclear extracts.9,10 This approach yields analytical
measurements of average macromolecular radii and surface charge density,
which in turn allows one to evaluate the condensation behavior and
conformational flexibility of the chromatin fragment being studied. Here
1
J. C. Hansen and C. L. Turgeon, Methods Mol. Biol. 119, 127 (1999).
2
J. C. Hansen, Annu. Rev. Biophys. Biomol. Struct. 31, 361 (2002).
3
P. J. Horn, K. A. Crowley, L. M. Carruthers, J. C. Hansen, and C. L. Peterson, Nat. Struct.
Biol. 9, 167 (2002).
4
B. Dorigo, T. Schalch, K. Bystricky, and T. J. Richmond, J. Mol. Biol. 327, 85 (2003).
5
T. M. Fletcher, P. Serwer, and J. C. Hansen, Biochemistry 33, 10859 (1994).
6
T. M. Fletcher, U. Krishnan, P. Serwer, and J. C. Hansen, Biochemistry 33, 2226 (1994).
7
L. M. Carruthers, C. Tse, K. P. Walker, III, and J. C. Hansen, Methods Enzymol. 304,
19 (1999).
8
L. M. Carruthers and J. C. Hansen, J. Biol. Chem. 275, 37285 (2000).
9
P. T. Georgel and J. C. Hansen, Biopolymers 68, 557 (2003).
10
P. T. Georgel, T. M. Fletcher, G. L. Hager, and J. C. Hansen, Genes Dev. 17, 1617 (2003).

we describe how AME can be used as a biophysical method for character-

izing specific in vivo–assembled chromatin fragments. The general con-
cepts presented in this chapter are based on our recent studies of
genomic murine mammary tumor virus (MMTV) promoters.10
To best describe the increasing number of specific types of ‘‘higher
order chromatin structures’’ observed in vitro and in vivo, Woodcock and
Dimitrov11 have introduced a nomenclature that is analogous to that used
for proteins. Primary chromatin structure refers to a linear arrangement of
nucleosomes (i.e., beads-on-a-string), secondary chromatin structure de-
scribes condensed fiber conformations that results from intrinsic and/or
protein-mediated nucleosome-nucleosome interactions (i.e., linker his-
tone-stabilized 30 nm fiber). Tertiary structures chromatin refers to chro-
matin suprastructures formed through interaction of secondary structures
(i.e., long-range fiber-fiber interactions). Throughout this chapter we use
the nomenclature suggested by Woodcock and Dimitrov.11
Agarose Multigel Electrophoresis

Agarose gel electrophoresis generally is thought of as a preparative
technique, but it also can be used to obtain quantitative information about
macromolecular size and charge. In the presence of an applied electrical
field, the mobility, , of a charged macromolecule in solution is directly
proportional to its surface change density.5,6,12 In the presence of agarose,
the ‘‘gel-free’’ mobility, 00 , is reduced by the interaction of the macromol-
ecule with the network of gel pores, Pe (pore size). Interactions with the gel
matrix are referred to as ‘‘sieving,’’ and are dependent on the effective
radius, Re, and conformational flexibility of the macromolecule.5,6,12,13
AME is an easily accessible method that is performed with the commer-
cially available electrophoresis apparatus shown in Fig. 1A. An agarose
multigel consists of a multiple individual agarose running gels embedded
in a 1.5% agarose frame (Fig. 1B). The running gels typically range
from 0.2% to 3.0% agarose. The multigel apparatus minimizes gel to gel
variations in temperature, field strength, and buffer pH, which allows
determination of the , 00 , and Re of macromolecules with analytical pre-
cision.5,12 The relationship between , 00 , Re, and Pe during an agarose gel
electrophoresis experiment is described in Eq. (1):
0
=0 ¼ ð1 Re =Pe Þ2 : (1)
11
C. L. Woodcock and S. Dimitrov, Curr. Opin. Genet. Dev. 11, 130 (2001).
12
G. A. Griess, E. T. Moreno, R. A. Easom, and P. Serwer, Biopolymers 28, 1475 (1989).
13
J. C. Hansen, J. I. Kreider, B. Demeler, and T. M. Fletcher, Methods 12, 62 (1997).
Fig. 1. (A) Multigel apparatus. (B) 9-lane agarose multigel. Percentages of agarose in
running gels are in increment of 0.1%. Note: the figure shows only one half of an 18-lane gel
(mirror image).
In an AME experiment, the sample of interest is first spiked with the spher-
ical bacteriophage T3 (Re ¼ 30.1 nm). The T3 00 is obtained by extrapolat-
ing the linear region of a plot of log T3 versus agarose percentage to 0%
agarose, that is, the y-axis.5,6,13 Using Eq. (1), the Pe of each running gel is
calculated from the and 00 and known Re (30.1 nm) of the T3 internal
standard. For the unknown band(s) of interest, the and 00 are determined
experimentally and the Re in each running gel is calculated using Eq. (1)
and the T3-derived values of Pe for that gel.
Extending AME to In Vivo–Assembled Chromatin

Previous AME studies have been performed with purified bacterio-
phages5–8,13 and defined chromatin model systems assembled in vitro from
pure components.14 Consequently, to apply AME for the analysis of geno-
mic chromatin fragments isolated in low-salt nuclear extracts, it was first
necessary to determine whether any components in the extracts altered
the intrinsic , 00 , and Re of the genomic chromatin bands. The results of
these control experiments have been published9 and are summarized
briefly here. Naked DNA or model 12-mer nucleosomal arrays were added
to nuclear extracts to mimic the release of genomic chromatin fragments
into the same environment. Under these conditions super-shifted smears
were observed for both the DNA and nucleosomal arrays on standard 1%
agarose gels. To abolish deleterious non-specific DNA binding, herring
sperm DNA (HS DNA) was added to the samples. For incubations per-
formed with 10 g standard unfractionated nuclear extracts15 in a final
volume of 10–20 l, addition of up to 100-fold excess HS DNA led to single
discrete bands that migrated the same as pure DNA or nucleosomal arrays
alone (note that the exact amount of required excess HS-DNA is extract de-
pendent). We have reported data for yeast, human carcinoma, Drosophila
embryo, mouse adenocarcinoma cell extracts. Other types of extracts will
require empirical optimization through titration of competitor HS DNA as
described in Georgel and Hansen.9
AME experiments were subsequently performed with DNA and nu-
cleosomal array samples prepared under the same conditions. In all cases,
the measured Re and 00 values were the same within experimental error.9
The quantitative data demonstrate that the intrinsic Re and 00 of nucleoso-
mal arrays can be accurately determined in the presence of unfractionated
low-salt nuclear extracts. To adapt AME for analysis of genomic chromatin
fragments, it was also necessary to show that Southern blotting could be
used to accurately measure the of specific bands in each of the running
gels. This was accomplished by mixing 208-12 DNA or nucleosomal arrays
in a low-salt nuclear extract as described earlier, and then measuring the
band migrations by both fluorescent dyes and Southern hybridization.
Again, the measured Re and 00 values were the same within experimental
error.9
14
J. C. Hansen, J. Ausio, V. H. Stanik, and K. E. van Holde, Biochemistry 28, 9129 (1989).
15
J. D. Dignam, R. M. Lebovitz, and R. G. Roeder, Nucleic Acids Res. 11, 1475 (1983).
Methods
The only aspect of the AME approach that is not general is obtaining
the specific genomic chromatin fragment of interest in a low-salt nuclear
extract. Once this has occurred, the remainder of the method has been
standardized. Presented in the following are the general protocols needed
to:
1. Prepare a multigel.
2. Perform an AME experiment using low-salt nuclear extracts as the
source of the chromatin.
3. Transfer the multigel to membrane and measure the in each
running gel through use of Southern blotting.
4. Calculate the , 00 , and Re of the chromatin-band(s) of interest, and
the Pe of each running gel.
In addition, we describe western assays for determining the compo-
sition of the genomic chromatin fragment, and biophysical assays for
formation of salt-dependent secondary chromatin structure and for the
extent of conformational flexibility. To emphasize the potential of the
AME approach, we conclude the article by briefly describing key aspects
of our recent use of AME to characterize the active and inactive states of
genomic MMTV promoters.
Equipment
Multigel units such as those shown in Fig. 1A are commercially avail-
able from Aquebogue (NY). It is important to keep all the various com-
ponents of the multigel together as a single unit. A high-temperature
water bath is needed to prepare the multigel. A Cole Parmer oscillating
pump run at 40–50% maximum output voltage (i.e., 50–60 V for this pump)
regulated through the use of a rheostat is needed to circulate the running
buffer and maintain temperature control. For the reasons described later,
we strongly recommend that the agarose be molecular biology grade,
Low Electro Endo-Osmosis purchased from Research Organics (cat #
1170A-3). The casting and running buffer is either E buffer (40 mM Tris-
HCl [pH 7.8], 0.25 mM Na2 EDTA) or TAE (40 mM Tris acetate [pH
8.3], 1 mM Na2EDTA). When formation of secondary chromatin struc-
tures is being assayed, E buffer is used and contains MgCl2 at concentra-
tions equal to 0.1–2.0 mM free Mg2þ. Do not use Tris-borate buffers in
combination with low electro endo-osmosis (EEO) agarose, as this will
generate anomalous undesirable electro-osmosis effects.
Multigel Preparation

Set the high-temperature water bath at 65–70 , and fill with enough
water to have the entire tube of agarose (see later) submerged to about
1 cm from the lid. The gel frame is prepared with 1.5% agarose in E or
TAE buffer. Weigh 1.8 g of agarose into a 250-ml flask. Tare the scale.
Add 120 ml of running buffer (measure by weight not by volume; i.e., as-
suming a buffer density of 1.0, add 120 g of buffer) and tare again. Cover
the top of the flask with a piece of plastic wrap, punch a small hole, and
melt the agarose in a microwave oven. After melting, weigh the flask again.
Compensate for evaporation by adding milliQ water. Cool the agarose to a

temperature around 55 (i.e., comfortable to touch). Assemble the slot
former and the desired comb as described in detail,16 and as shown in
Fig. 1A. Secure the end plates in the grooves with the détentes (small plas-
tic parallelepipedic blocks). If needed, seal the plates with a small quantity
of 1.5% framing agarose. Level the unit. To make the frame, pour the
molten agarose into the gel bed, allow the agarose to set for 30 or 60 min
(9- or 18-lane slot formers, respectively).
Begin preparing the agarose running gels. Start by labeling capped 15-ml
Kimble borosilicate (threaded end, 20 125 for 9-lane gels and 16 125 for
18-lane gels) tubes with the chosen agarose concentrations (see Table I
for details). Add running buffer to the tubes (see Table I for volumes)

and place the tubes in the water bath at 65–70 .
After complete polymerization of the frame, carefully remove the comb
and then very gently pull out the slot former. Abrupt removal will result in
breaking the lanes. You may want to practice removing the slot former
using standard agarose prior to using the more expensive low EEO agar-
ose. Using a pair of small forceps, remove the small strips of agarose that
often form at the edges of the wells. Burst any bubbles that may have
formed in the thin layer of agarose beneath each slot. Put the comb back
in its original location.
While the frame is polymerizing, prepare the agarose stock(s) for the gel
dilutions. To prepare a 0.2–1% agarose multigel, start with 60 ml of molten
1% agarose in a 125-ml flask (prepared in the same way as described earlier
for the 1.5% agarose stock). Immediately dilute the stock according to
Table I to prepare each running gel. After adding the appropriate volume
of 1% agarose to the pre-warmed running buffer, cap the tube, vortex
briefly, and immediately pour the agarose into the appropriate multigel slot
using disposable plastic pipettes. Prepare one agarose concentration at a
time. Use a new pipette every time you change agarose concentration.
16
J. C. Hansen, T. M. Fletcher, and J. I. Kreider, Methods Mol. Biol. 119, 113 (1999).
TABLE I
Dilutions
Gel (%) Buffer (ml) Agarose stock (ml)
1% stock
18-lane frame
0.2 4 1
0.3 3.5 1.5
0.4 3 2
0.5 2.5 2.5
0.6 2 3
0.7 1.5 3.5
0.8 1 4
0.9 0.5 4.5
1 0 5
9-lane frame
0.2 8 2
0.3 7 3
0.4 6 4
0.5 5 5
0.6 4 6
0.7 3 7
0.8 2 8
0.9 1 6
1 0 10
3% stock
18-lane frame
0.9 3.5 1.5
1 3.4 1.7
1.3 2.9 2.2
1.6 2.4 2.7
1.9 1.9 3.2
2.2 1.4 3.7
2.5 0.9 4.2
2.8 0.4 4.7
3 0 5
9-lane frame
0.9 7 3
1 6.7 3.3
1.3 5.7 4.3
1.6 4.7 5.3
1.9 3.7 6.3
2.2 2.7 7.3
2.5 1.7 8.3
2.8 0.7 9.3
3 0 10
For a 9-lane frame, use 7 ml of diluted agarose per lane and 2.5 ml for
an 18-lane. Note that to maintain stability of the frame, the 0.2% agarose
should not be poured in the slot at the edge of the multigel frame. For a
0.2–1.0% multigel we suggest the following pouring order: 0.8, 0.9, 1.0,
0.2, 0.3, 0.4, 0.5, 0.6, 0.7. To prepare a 1.0–3.0% agarose 18-lane multigel,
start with 60 ml of molten 3.0% agarose in a 125-ml flask. For each running
gel, immediately dilute the stock according to Table I as described previ-
ously. For a 0.9–3.0% gel we suggest the following pouring order: 0.9, 1.0,
1.3, 1.6, 1.9, 2.2, 2.5, 2.8, 3.0.
Let the running gels set for 1 h. Pull out the comb, end plates, and dé-
tentes. Fill the multigel apparatus with running buffer. The multigels can be
prepared in advance and used the following day. Make sure that the gel box
is covered with plastic wrap to avoid unnecessary evaporation of buffer.
Gels have been used successfully after 3 days. To prevent the gel from
sliding off the gel bed during subsequent electrophoresis, place several
10-l plastic tips in the plate grooves at opposite corners.
Gel Electrophoresis
Prior to electrophoresis, low-salt nuclear extracts (100 g total
nucleic acid) are transferred to fresh tubes containing the appropriate
amount of empirically determined herring sperm DNA, 0.5 g of bacterio-
phage T3, and glycerol (10% w/v final). Samples are loaded in the
appropriate well and electrophoresed at 1.33 V/cm for 6 h (timed manually
and precisely). Alternatively, one can perform the electrophoresis for
1 V/cm for 8 h.13 The average buffer temperature in our experiments

has been 24 3 (room temperature). Using the pump system described
earlier, the running buffer was continuously recirculated throughout the
entire experiment at a slow and steady pace, which prevents the formation
of pH or ion gradients and large temperature fluctuations.
To measure the of the bacteriophage T3 in each running gel,
after electrophoresis a thin plastic sheet is placed under the multigel
to permit transfer of the fragile gel to a tray containing either ethidium
bromide or SYBR green. The tray is placed on a rotating shaker set at
60 rpm (maximum) for 30 min. A 60-min destaining period in running
buffer (or distilled water) is required if ethidium bromide staining is
used. The multigel is then photographed under ultraviolet light using
standard positive/negative film or using a digital camera, with a UV-
compatible ruler placed next to the edge of the multigel for measurement
purposes.
Southern Hybridization
To detect specific genomic chromatin bands by Southern blotting (see
later), the gel was subsequently treated with standard denaturing and neu-
tralizing solutions, and transferred to Hybond N or NX membranes using
20 SSC in combination with the Schleicher and Schuell Turboblotter.
The use of the Turboblotter (transferring from top to bottom) avoids
having to flip the multigel, and in doing so minimizes the risk of separation
of the running gels from the agarose frame. After transfer, a probe covering
the desired genomic region is radioactively labeled by random priming or
nick-translation. The unincorporated radioactivity is removed using gel fil-
tration (BioSpin P30, BioRad) and the recovered probe is subsequentially
phenol-chloroform extracted. The Hybond membrane is pre-hybridized at

42 for 1.5 h in hybridizing solution (6 SSC, 5 Denharts, 0.5% SDS,
50% formamide, and 0.1 mg/ml denatured salmon sperm DNA). The
probe is boiled for 10 min, added to the hybridization solution, and then in-

cubated at 42 overnight. The membranes were washed twice in 2 SSC,
0.1% SDS at 42 and 30 , respectively, and once with 0.2 SSC, 1% SDS

at 27 . The membranes were then exposed to film or to a PhosporImager
screen for 2–5 days.
Data Analysis
For each individual band detected by Southern blotting or fluorescent
staining, the migration in centimeters is measured from the center of the
well to the center of the band using NIH Image17 or equivalent software.
The digitized image of the ruler placed adjacent to the multigel (see
earlier) serves as reference to set the scale (number of pixels per centi-
meters). Migrations (in centimeters) are subsequently converted to (in
cm2/V s). The gel-free mobility, 00 is obtained by extrapolating the linear
region of a plot of log versus agarose percentage to the y-axis. For
an 2.5-kb chromatin fragment, the linear region falls in the range of
0.2–0.9% agarose. The correlation coefficients of the linear regressions
generally are 0.99. Note that the linear region must be determined
empirically for different sized chromatin fragments.
Using Eq. (1), the Pe of each running gel is determined from the meas-
ured and 00 and known Re (30.1 nm) of the bacteriophage T3 internal
standard. The measured Pe was compared to the range that was deter-
mined based on calculated Pe values from 10 random multigels (see
17
R. R. O’Neill, L. G. Mitchell, C. R. Merril, and W. S. Rasband, Appl. Theor. Electrophor. 1,
163 (1989).
TABLE II
T3
0 and E
4
T3
0 (10 cm2/V s) E (105 cm2/V s)
E buffer 0.755 1.02

E buffer þ 2 mM MgCl2 0.65 0.97
Table III). If the Pe values fall outside the expected ranges, the data may be
suspect. The Re of the genomic chromatin band(s) in each running gel is de-
termined from their measured and 00 and the calculated Pe for that gel.
To convert the measured 00 to the actual 0, one must correct for the
effects of electroosmosis on the measured (E). This requires empirical
determination of the E for each specific type of commercial agarose,
which is a time-consuming and expensive process. Toward this end, we
have determined the E for the Research Organics low EEO agarose men-
tioned earlier, and found it to be constant over several lots and many years.
By using this specific agarose in each multigel frame and running gel, ex-
perimental measurements of E can be avoided, and the 00 is converted
to the true 0 according to the following equation
0 0 T3
0 ¼ ð0 ÞðT30 þ E Þ=0 E (2)
where the T3
0 also is a measured constant in the same agarose (see Table II
for T3
0 and E values determined in E and E þ 2 mM Mg2þ buffers).
Assay for Secondary Chromatin Structure Formation

The formation of locally folded secondary chromatin structure is
assayed by determining whether the Re of the genomic chromatin fragment
decreases by 35% in E buffer containing 2 mM free Mg2þ (relative to
the Re in E buffer alone). The decrease in Re between 0 and 2 mM Mg2þ
also should be linear.6,10 Importantly, when assaying for secondary chroma-
tin structure formation, the AME experiments must be performed under
conditions where the Re is 10 times larger than the Re, that is, in
0.2–1.0% low ‘‘concentration’’ multigels. All aspects of the AME experi-
ment are performed as described above, the only difference being the
free Mg2þ concentration in the casting and running buffer. For these pur-
poses E buffer is used because of its low EDTA concentration, and the free
Mg2þ concentration is assumed to be 0.25 mM less than the added MgCl2
concentration.
TABLE III
Empirically Determined Pe Range (nm)
Gel (%) Pe range (nm)
0.2 400–600
0.3 300–460
0.4 200–350
0.5 150–300
0.6 150–250
0.7 150–200
0.8 130–200
0.9 120–190
1 100–150
1.3 90–120
1.6 75–100
1.9 65–80
2.2 60–70
2.5 50–55
2.8 45–50
3 37–43
Flexibility Assay
A relative assay for the conformational flexibility of a chromatin frag-
ment is provided by the slope of a graph of Re versus Pe in the gel range
where Re approaches Pe, that is, in high percentage agarose gels.5,6,12 A
slope near zero is indicative of a particle whose conformational features
are not deformed during electrophoresis. Such a particle is relatively inflex-
ible, for example, unfolded nucleosomal arrays in low salt.5 In contrast, the
Re of a more flexible particle (e.g., DNA, irregularly spaced nucleosomal
arrays in low salt) decreases as the Re approaches Pe. For a 2.5-kb chroma-
tin fragment, this occurs between 1.2% and 3.0% agarose. Thus, relative
conformational flexibility is assayed in high-concentration multigels.
Composition Analysis of the Genomic Fragments

It is very powerful to couple the biophysical measurements made using
AME with western analyses of the composition of the same specific geno-
mic fragment (see later). To accomplish this, the same low-salt nuclear ex-
tracts used in the AME experiments are loaded on duplicate agarose gels
buffered with TAE. The gels are run at 5 V/cm for exactly 150 min. The
first gel is probed with labeled DNA from the genomic fragments of inter-
est to define their location in the gel. The second gel is the source of the
genomic chromatin fragments used for the subsequent western analyses.
Specifically, to confirm the identity of the bands and determine the pre-
cise distance of migration, in one gel, the DNA was transferred to Hybond
NX membrane and probed with the fragment-specific probe. The second gel
is stained with Commassie (GelCode Blue Stain, Pierce) according to the
manufacturer’s specifications. Slices with 1.5 mm-thickness containing the
bands of interest are excised, soaked in 1% SDS, 50 mM Tris-HCl (pH
6.8) and 1% -mercaptoethanol for 15 min at room temperature. The slices
are then transferred to a Pyrex tube containing the same buffer pre-heated

to 95 for 25–35 s (or until the agarose slice becomes translucent), and the
tube immediately chilled on ice. The cooled agarose slice was rapidly placed
in the well of 1.5-mm thick 4–20% SDS gel and electrophoresed at 100 V
for 90 min. The proteins were subsequently transferred to Immobilon-P
PVDF (Amersham) membranes and immunoblotted according to standard
Maniatis western blotting conditions using appropriate antibodies.
Data Interpretation
The final aspect of the AME approach is to interpret the combined
biophysical and compositional data. For illustrative purposes, this section
summarizes selected important observations from our work with genomic
MMTV promoters under different states of transcriptional regulation.10
For example, the inactive form of the genomic MMTV promoter in
mouse 3134 cells was found to contain histone H1, and showed the same
behavior in the AME assays as model H5-containing nucleosomal arrays
assembled in vitro from pure components. Exposure of cells to dexametha-
sone leads to transcriptional activation in vivo.18–20 The transcriptionally
active genomic MMTV promoters isolated from dexamethasone-treated
cells contained RNA Pol II, TBP, Octl, Brgl, and acetylated H3 tail
domains, and its Re increased to 43 nm (compared to 26 nm in the tran-
scriptionally inactive state). Thus, the biophysical measurements support
the conclusion that H1 had been replaced with several large transcription
factors and multiprotein assemblages after hormone treatment. Interest-
ingly, both the inactive and active forms of the genomic MMTV promoter
were capable of forming salt-dependent secondary chromatin structure.
Together, these data support a model in which transcriptional acti-
vation is associated with reorganization of secondary chromatin structures,
18
H. Richard-Foy, F. D. Sistare, A. T. Riegel, S. S. Simons, Jr., and G. L. Hager, Mol.
Endocrinol. 1, 659 (1987).
19
H. Richard-Foy and G. L. Hager, EMBO J. 6, 2321 (1987).
20
T. K. Archer, C. J. Fryer, H. L. Lee, E. Zaniewski, T. Liang, and J. S. Mymryk, J. Steroid
Biochem. Mol. Biol. 53, 421 (1995).
[3] visualization and 3D structure determination 29
rather than decondensation of the promoter chromatin into an unfolded

beads-on-a-string structure.
As with any newly introduced technical approach, the ability to accur-
ately interpret AME data will increase in direct proportion to the number
of systems that are characterized by this method in the future.
[3] Visualization and 3D Structure Determination

of Defined Sequence Chromatin and Chromatin
Remodeling Complexes
By Rachel A. Horowitz-Scherer and Christopher L. Woodcock
Recent progress in preparing and purifying defined sequence chromatin

and chromatin remodeling complexes provides an opportunity to deter-
mine the three-dimensional (3D) changes that accompany major remodel-
ing events, such as histone modification, the binding of transcriptional
repressors and activators, and the action of ATP-dependent remodeling
complexes. The compelling evidence that many of the functional changes
in chromatin are intimately connected to changes in ‘‘higher-order struc-
ture’’ which may be epigenetically inherited, provides a powerful incentive
to understand their structural basis.
The ultimate goal is to understand the different states of chromatin at an
atomic level of resolution, but the range of chromatin structures accessible to
X-ray crystallography is very limited. This limitation can, in principle, be over-
come using lower-resolution 3D volumes generated by an electron microscope
(EM) into which atomic level structures can be ‘‘docked.’’ The range of
transmission EM-based technologies available for visualizing chromatin
and chromatin remodeling complexes is listed in Table I. For each method,
we provide experimental guidelines based on recent work in our laboratory.
General Considerations
With any imaging technique, the information retrievable is limited
by the quality of the sample, and for chromatin, the most important
considerations are as follows.
Purity
The proportion of the sample constituting the structure of interest
should be monitored. A higher level of contamination can be tolerated if
the material under study is structurally distinct. For example, a circular

rather than decondensation of the promoter chromatin into an unfolded

beads-on-a-string structure.
As with any newly introduced technical approach, the ability to accur-
ately interpret AME data will increase in direct proportion to the number
of systems that are characterized by this method in the future.
[3] Visualization and 3D Structure Determination

of Defined Sequence Chromatin and Chromatin
Remodeling Complexes
By Rachel A. Horowitz-Scherer and Christopher L. Woodcock
Recent progress in preparing and purifying defined sequence chromatin

and chromatin remodeling complexes provides an opportunity to deter-
mine the three-dimensional (3D) changes that accompany major remodel-
ing events, such as histone modification, the binding of transcriptional
repressors and activators, and the action of ATP-dependent remodeling
complexes. The compelling evidence that many of the functional changes
in chromatin are intimately connected to changes in ‘‘higher-order struc-
ture’’ which may be epigenetically inherited, provides a powerful incentive
to understand their structural basis.
The ultimate goal is to understand the different states of chromatin at an
atomic level of resolution, but the range of chromatin structures accessible to
X-ray crystallography is very limited. This limitation can, in principle, be over-
come using lower-resolution 3D volumes generated by an electron microscope
(EM) into which atomic level structures can be ‘‘docked.’’ The range of
transmission EM-based technologies available for visualizing chromatin
and chromatin remodeling complexes is listed in Table I. For each method,
we provide experimental guidelines based on recent work in our laboratory.
General Considerations
With any imaging technique, the information retrievable is limited
by the quality of the sample, and for chromatin, the most important
considerations are as follows.
Purity
The proportion of the sample constituting the structure of interest
should be monitored. A higher level of contamination can be tolerated if
the material under study is structurally distinct. For example, a circular

TABLE I
Method Comments Resources needed
Positive stain Restricted to 2D Transmission EM

Mostly for verifying structural integrity Carbon support films
of chromatin in decondensed state
Shadow casting Quasi-3D data from shadow lengths Transmission EM
surface relief Vacuum evaporator
Negative Nucleosomes, large proteins, and Transmission EM with tilt
staining protein complexes resolved capability
Partial 3D information from stereo Image digitization; software
pairs; full 3D from tomography for tomography, SPR
and from single-particle
reconstruction (SPR) of samples
with uniform shape
Cryo-imaging Only method that can be used Transmission EM, ideally with
with unfixed, unstained material. field emission gun (FEG);
Tomography and SPR more cryo-holder; cryo transfer
technically challenging device; image digitization,
3D reconstruction software
‘‘minichromosome’’ is readily distinguished from contaminating linear

chromatin.
Concentration
A chromatin sample containing 50 g/ml of the material of interest will
usually suffice. Concentrations much lower than this will always yield
poorer images, and cryo-imaging will be all but impossible. Small volumes
may be concentrated and the buffer components exchanged if needed using
Minicon or Centricon (Amicon #42409, 4208) systems.
Buffer Components
Materials that alter the surface tension or otherwise interfere with ad-
hesion or staining, such as detergents, sucrose, or glycerol, must be re-
moved by dialysis, and the sample separated from ‘‘carrier’’ proteins such
as BSA. If the sample is to be fixed (see later), Tris-type buffers must be
avoided as they react with aldehyde fixatives. We prefer HEPES or PIPES,
at the lowest concentrations that provide adequate buffering. For chroma-
tin, the concentrations of monovalent and divalent salts have a strong influ-
ence on compaction, and will be dictated by the experiment. Protein
complexes that require high salt (over 200 mM monovalent ions) for sta-
bility are best fixed at the optimum salt, after which dilution or dialysis can
be used to achieve the appropriate ionic strength.
Fixation
With the exception of cryo-imaging, all techniques require the adhesion
of the sample to a support, a process that exposes chromatin to denaturing
surface forces. Fixation in 0.1% EM-grade glutaraldehyde for 4 h followed
by its removal by dialysis is optimal for most chromatin samples. This treat-
ment preserves the sedimentation constant of chromatin after a post-
fixation change in salt concentration. Longer fixation may result in an
irreversible increase in compaction. For small volumes, minidialysers
(Pierce #69570) are very convenient for this procedure.
Adhesion to Support Films

Most of the techniques discussed here require the sample to be deposited
on a thin carbon film. Proper adhesion requires that the film be made
hydrophilic and, for chromatin samples, a monovalent salt concentration
of 50 mM is recommended.1
In our laboratory, carbon films are created by electron beam evapor-
ation in an oil-free, turbopumped vacuum of 10–6 Torr (Baltec BAE80
system), onto freshly cleaved mica sheets, floated onto a double distilled
H2O (ddH2O) surface, and then deposited onto the rhodium side of 400-
mesh copper-rhodium specimen support grids. Film stability is improved
if the carbon is deposited in 3–4 bursts, separated by a few minutes. The
carbon surface is rendered hydrophilic by a 1-min exposure to a glow-
discharge in air at a pressure of 50 mTorr in a Haskins PDC-3XG plasma
cleaner (setting 3). Grids should be used within 1 h of glow-discharge. A
properly treated carbon film will retain a thin layer of liquid when a drop
of buffer is added and then wicked away with filter paper.
Positive Staining
This method is primarily useful for verifying the quality of decondensed
chromatin samples, allowing the investigator to check nucleosome number
and distribution, as well as the overall configuration of the array (circular
or linear). A simple but reproducible procedure is as follows:
1. On a freshly exposed sheet of Parafilm, place a drop of sample
(5–10 l), 3 drops (100 l) of water, 3 drops of 2% aqueous uranyl
acetate (UA), and a final drop of water.
2. Place a glow-discharged carbon film on the sample.
1
C. L. Woodcock, L. Y. Frado, G. R. Green, and L. Einck, J. Microsc. 121, 211 (1981).
3. After 5 min, remove the grid, blot from one side with a flag of filter
paper (verifying visually that the surface remains fully wetted) and
place on the first drop of water for 10 s. Remove, blot, and place on
the next drop, and continue at 10 s intervals through the 3 drops of
stain to the fourth drop of water, which serves to remove unbound
stain. After the grid has air-dried, it is ready for EM observation.
This method provides a faint positive staining, primarily of DNA, and
the microscope should be set up to maximize contrast. Alternatively, the
EM may be set up for tilted beam dark field, which produces excellent
reversed contrast.
A positive staining technique that provides significantly more contrast
and stains protein and DNA but, in our hands, is less reproducible, is to
use ethanolic phosphotungstic acid (PTA). This method is described in
detail elsewhere.2
Shadow Casting after Drying from Glycerol (Fig. 1A)

This method is adapted from the technique described by Tyler and
Branton for filamentous proteins.3a It has the advantage of providing
quasi-3D information about chromatin conformation and tolerates salts
and other buffer components that would normally have to be removed
for EM preparation. The fixed chromatin sample is mixed with glycerol,
and sprayed with an atomizer onto freshly cleaved mica. The mica is placed
in a vacuum evaporator, whereupon the glycerol gradually evaporates,
leaving chromatin particles behind as the meniscus withdraws. The 3D
shape of the sample is protected from air-drying collapse, and salts and
other contaminants are swept into a small area at the center of the original
drop. Without breaking the vacuum, the mica is shadowed with Pt, and
then coated with a layer of carbon. Finally, the mica is removed from the
chamber, and the carbon/platinum replica floated off and transferred to
EM grids for examination. Detailed steps are as follows:
1. The fixed sample (50 g/ml) is made 70% in glycerol (mix
thoroughly).
2. About 20 l is applied to the intake tube of a standard drugstore
atomizer, from which the rubber bulb has been removed. The output
from a ‘‘Dust-off’’ or similar can of compressed gas is connected to
the air input orifice of the atomizer.
2
C. L. Woodcock and R. A. Horowitz, Methods Cell Biol. 53, 187 (1998).
3a
J. M. Tyler and D. Branton, J. Ultrastruct. Res. 71, 95 (1980).
Fig. 1. Examples of the options for EM imaging of chromatin described in the text. In
this case, the methylated DNA-binding protein MECP2 was bound to defined 12-mer
nucleosome arrays (in collaboration with P. Georgel, P. Wade, and J. Hansen) and fixed with
glutaraldehyde. (A) Shadowing after vacuum drying from glycerol preserves the 3D shape of
the complex and permits measurement of height (between white lines) and diameter (between
black lines). Bar ¼ 30 nm. (B) Negative staining with uranyl acetate gives an overall sense of
size and shape, but requires tomographic reconstruction to reveal details of internal
substructure (C, D). A central slice from a reconstruction (C) reveals the outline of individual
nucleosomes (arrow). Slice-by-slice examination of the reconstructed volume allows the
location and orientation of individual nucleosomes in the array to be identified, and
construction of a solid model (D) of the array, in which nucleosomes are represented by 5
10 nm disks. (E) ECM imaging of unstained particles yields low-contrast images of the
solution conformation, without flattening or drying. Individual nucleosomes can be identified
(arrow). Stereo-pair reconstruction3b of one particle (F) yields the 3D coordinates of
individual nucleosomes (arrow), and a solid model (G) can be built from those coordinates. A
stereo pair does not give orientation information, so nucleosomes are represented as 10 nm
diameter spheres. Bar ¼ 10 nm.
3b
A. Beorchia, M. Ploton, M. Menager, M. Thiry, and N. Bonnet, J. Microsc. 163, 221 (1991).
3. A 2 2 cm ‘‘target’’ of freshly cleaved mica is placed about 25 cm

from the atomizer, and a 1-s burst of gas used to propel the sample
as a fine spray. Trial runs should be used to establish the optimum
volume of sample and distance of the mica to produce a range of
individual drop sizes that are closely spaced but do not coalesce.
4. The mica is quickly transferred to a vacuum evaporator, and a high
vacuum maintained for several hours. Platinum is then evaporated

at an angle of 10 , monitoring the thickness with a film thickness
monitor. Trial runs are needed to calibrate the system for an
optimum shadow thickness.
5. To provide accurate size measurements as well as a complete surface
profile of the specimen, the mica is mounted on a rotating platform,
which is kept stationary for the first one-third of the evaporation,
and rotated for the remaining two-thirds.

6. The tilt angle is set to 90 , and a self-supporting layer of carbon is
evaporated onto the mica.
7. The mica is removed from the evaporator, the carbon scored into
squares, and the replicas floated off onto water. 400 mesh grids are
then used to pick up the squares from below.
8. In the transmission EM, the outline of the original droplets are
marked by faint circular ‘‘watermarks,’’ and the drop centers by
heavy accumulations of salt and other buffer components.
Shadowed sample will be found between these delimiters.
9. Excellent images are obtained from very lightly shadowed samples
recorded in dark field image mode (Fig. 1A).
Negative Staining (Fig. 1B)

Selection of the appropriate negative stain requires a certain amount of
trial and error; different samples inevitably demand modifications to
the basic procedure. The standard stain for chromatin and many proteins
is 1–2% aqueous unbuffered UA, while for the yeast SWI/SNF complex
(a case discussed in detail later) and several other chromatin-associated
regulatory proteins, 1.5% PTA buffered to pH 7.0 and with 0.015% glucose
has provided better preservation of fine detail. Other possible stains
are ammonium molybdate, vanadate, and alternate uranyl salts. It is advis-
able to evaluate the sample in several stains, as the appearance4 and subse-
quent reconstructions5 may differ depending on the amount of support the
4
C. L. Woodcock, H. Woodcock, and R. A. Horowitz, J. Cell Sci. 99, 99 (1991).
5
C. L. Woodcock and W. Baumeister, Eur. J. Cell Biol. 51, 45 (1990).
stain provides, and the features it accentuates. A detailed discussion of

the appearance of chromatin in various negative stains may be found in
Woodcock et al. (1991).4 We use the following conditions for staining:
1. On a freshly exposed sheet of Parafilm, place a drop of sample
(5–10 l), 3 drops (50 l) of water, and 3 drops of stain. Place a
glow-discharged carbon film on the sample. (If diluting the sample
to 50 mM NaCl, place the appropriate volume of NaCl solution on
the Parafilm first, then place the grid on the drop and finally ‘‘inject’’
the protein sample into the drop under the grid with a fine-tipped
Eppendorf pipet.)
2. After 5 min, remove the grid, blot from one side with a flag of filter
paper (verifying visually that the surface remains fully wetted), and
place on the first drop of water for 10 s. Remove, blot, and place on
the next drop, and continue at 10-s intervals through the third drop
of stain. The grid surface must not dry out during this sequence.
3. Pick the grid up off the last drop with self-closing forceps, and wick
the stain off the grid. Allow to dry in a dust-free area, then store in a
secure grid box in a desiccator.
For unfixed protein-only samples, an alternate method is recommend.
This method takes some practice to avoid wetting the back of the grid; if
this occurs the grid must be discarded:
1. Fill a Pasteur pipet with negative stain (in this case, usually 1–2%
UA).
2. Apply 3–5 l of sample to a glow-discharged carbon-coated grid
clamped in self-closing forceps. Allow to adsorb for 10–60 s.

3. Hold the grid at a 45 angle, and allow about 10 drops of the stain
to flow dropwise over the face of the grid into a waste container. The
flow should be rapid enough to provide the needed rinsing action to
remove excess salts.
4. Blot lightly as in the previous list and allow to air-dry.
There are several parameters to balance in the negative staining
process: the stain must be thick enough to support the specimen without
artifactual flattening during drying but thin enough for good contrast in
the images. Altering the speed at which the grid dries provides a
certain amount of control; grids can be left to dry in a chamber with up
to 65% RH. Negative staining requires a threshold protein concentration
to achieve an even spreading of the stain, but the specimen must be dis-
persed enough to allow clear discrimination of individual particles and
provide a useable separation between particles for the image processing
to follow.
Electron Cryomicroscopy (Fig. 1E–G)

Electron cryomicroscopy (ECM) permits the observation of specimens
in their solution conformation, without stains, shadowing, or substrate adhe-
sion and the accompanying flattening and drying artifacts. In most instances,
chemical fixation is unnecessary. A drop of the sample is applied to an EM
grid covered with a fenestrated substrate and allowed to spread over the
holes, then blotted to form a thin film of 20–100 nm. The EM grid is then rap-
idly plunged into a cryogen to convert the thin layer of specimen in solution
to vitreous ice. The frozen-hydrated specimen is transferred to the electron
microscope and observed at temperatures that maintain the vitreous layer.
Detailed descriptions of the application of this technique to chromatin
are available.6,7 Some technical advances since those publications appeared
are noteworthy:
. One of the most time-consuming and difficult steps in ECM is the
production of reproducible fenestrated substrates or ‘‘holey films.’’
It is now possible to purchase prepared grids with regularly sized
(selection from 1–3 m) and spaced holes (www.quantifoil.com).
. Manipulation of the variables of blotting, plunging, and humidity
can now be computer controlled through a plunging device, the
Vitrobot, originally developed at the University of Maastricht (sold
by FEI Company, www.vitrobot.com).
. The possibility of performing ECM tomography has been facilitated
by the development of computer interfaces to the electron
microscope for control of tilting and data acquisition (www.tvips.
com, www.feico.com, www.ami.scripps.edu).

. A high-tilt ( 80 ) ECM specimen holder is now available
(www.gatan.com).
Electron Microscope Data Collection
CCD Cameras Versus Film

The theoretical pros and cons of digital versus film recording of data
have been extensively discussed8; here we will consider some of the prac-
tical considerations. For beam-sensitive specimens (negatively stained or
frozen hydrated) the microscope must be able to record low-dose images
via deflection/shuttering of the beam above the sample. For most samples
6
J. Bednar and C. L. Woodcock, Meth. Enz. 304, 191 (1999).
7
C. L. Woodcock and R. A. Horowitz, Methods 12, 84 (1997).
8
J. Hesse, H. Hebert, and P. J. Koeck, Microsc. Res. Tech. 49, 292 (2000).
in negative stain, this is in the range of 10–15 e/A2 per exposure. Many
samples can tolerate up to 25 e/A2. Some proteins and DNA-protein com-
plexes, dependent on a poorly understood combination of the preparative
method and the negative stain, are extremely sensitive to damage. The mi-
gration of stain molecules, producing granulation, is usually obvious to the
eye, but damage to the biological material is subtler. If the microscope is
equipped with a CCD camera for digital data recording, it is useful to
collect a series of extremely low-dose images of a single field, until the cu-
mulative dose is 25 e/A2. The images can then be analyzed by calculating
difference images (arithmetic pixel-by-pixel subtraction of one exposure
from the next) to determine the dose at which damage becomes detectable.
Other factors that must be considered are the target resolution in the
reconstruction, the size and distribution of the protein on the support film,
and the area of the CCD chip. Film collection permits hundreds or thou-
sands of particles to be selected from a single exposure, whereas
the number of particles in a single CCD image is far fewer. However, film
collection necessitates scanning the negatives into digitally accessible data
at the appropriate resolution. The pixel size should be 3–4 smaller than
the desired resolution. It is possible with either scanners or CCD imaging
to greatly oversample; this should be avoided as artifacts can be introduced
when data management requires interpolation to larger pixel sizes. In
negative stain, the final resolution in a reconstruction of an asymmetric
particle is unlikely be better than 25A, necessitating a pixel size in the
raw data of 6–8A. For the yeast SWI/SNF complex, this was achieved using
a 2048 2048 CCD camera, each CCD image containing 50–100 particles.
Without convenient access to a high-resolution film scanner, the time
invested in collecting larger numbers of CCD images may be offset by the
time savings of eliminating darkroom work and the digitizing of negatives.
Microscope Parameters
With the target resolution determining the final pixel size, and there-
fore the magnification at which to record the images, other microscope pa-
rameters need be determined. While the best resolution will be achieved at
intermediate voltages in field emission gun (FEG)-equipped microscopes,
perfectly acceptable data can be acquired with care with LaB6 filament–
equipped microscopes at 80–120 kV. The highest resolution will be
obtained at the highest available voltage. The illumination conditions
(small condenser aperture, coordination of spot size with spread of beam)
should be optimized to provide the most coherent beam. The defocus
for the exposures should be selected so that the first zero of the contrast
transfer function (CTF) falls beyond the target resolution.
Single-Particle Reconstruction from Negatively Stained Images

of Asymmetric Particles
Single-particle reconstruction (SPR) is the method of choice for calculat-
ing 3D reconstructions of individual proteins, or protein complexes that
adopt a consistent 3D shape and have a size range from several hundred
kilodaltons to several megadaltons. With ‘‘ideal’’ images, the best possible
resolution in negatively stained samples will be 25 A and will require a
minimum of 2000 imaged particles. The lack of perfect images usually means
that 5000–10,000 images are required. Other reconstruction strategies are
available for proteins that assemble into helices or 2D crystals.9 Large cellu-
lar components and macromolecular assemblies that have highly variable
individual conformations are best analyzed by electron tomography.
Specific SPR methods exist for icosahedral viruses, for complexes that
exhibit rotational or other symmetries, and for particles that present a
subset of preferred orientations when prepared for electron microscopy.10
Here, we concentrate on the methods employed in this laboratory for yeast
SWI/SNF, a large (1.14 MDa) complex of 12 proteins, which did not
assume a preferred orientation in negative stain, and which does not
possess any symmetry.
Specimen Preparation
Many protein complexes, including ySWI/SNF, are unstable in the con-
ditions required for successful EM and must be chemically cross-linked
(see earlier). For this sample, we used glutaraldehyde fixation following
removal by dialysis of the glycerol and high salt required for long-term
stability, and PTA proved to be the most suitable negative stain.
Overview of the SPR Process
Data Preparation
The first step in calculating a single-particle reconstruction is to select the
individual particles. First, each source image (negative or CCD image) must
be evaluated for quality: images with drift or astigmatism are rejected and
groups of images with like defocus/CTF grouped together. Then the individ-
ual particles are excised from the parent image and centered in individually
‘‘boxed’’ fields. In order for subsequent programs to reorient the selected
particles, the box size should be 25–50% larger than the particle itself.
Most of the image processing packages that include SPR: SPIDER
(www.wadsworth.org/spider), EMAN (ncmi.bcm.tmc.edu), IMAGIC
(www.ImageScience.de), and SUPRIM (ami.Scripps.edu), offer the option
of automating the particle selection process. This can be done by cross-cor-
relation with projections of a previously calculated model if one exists. A
semi-automated method uses a subset of user-identified particles, which
are rotationally averaged and then used as a template for cross-correlation
with the particles in the raw image. This was used successfully for ySWI/
SNF; with sufficiently relaxed criteria, the program will ‘‘overpick’’ and
then the operator needs only to delete improper selections, a considerably
shorter task than hand-selecting each particle.
Alignment and Generation of Class Averages

Each particle must be centered in its box with x, y translations and on-
plane rotations necessary to have the entire collection of particles in align-
ment with respect to each other. One way to align the particles is to
compare them to projections of an existing model. This model can come
from other data sources, or be constructed from geometric primitives based
on the user’s preliminary assessment of the particle shape. Caution must be
maintained at this step, as the reference model can introduce a strong bias.
For ySWI/SNF, an asymmetric particle for which no a priori structural
information was available,11 a reference-free alignment was performed in
EMAN.12 This method employs algorithms that group particles by similar-
ity in their appearance, then averages the grouped images, with the goal of
having one group for each unique view of the particle. The user then exam-
ines the group averages and the members of the group to determine if a
unique view is adequately represented. From these groups, the user
chooses a subset that represents the best possible spread of views of the
particle for the calculation of an initial model. The Fourier common lines
method is employed to determine the relative orientations of the selected
groups in order to build the initial model.13 This initial model is then
9
J. Ruprecht and J. Nield, Prog. Biophys. Mol. Biol. 75, 121 (2001).
10
P. A. Thuman-Commike, FEBS Lett. 505, 199 (2001).
11
C. L. Smith, R. A. Horowitz-Scherer, J. F. Flanagan, C. L. Woodcock, and C. L. Peterson,
Nat. Struct. Biol. 10, 141 (2003).
12
S. J. Ludtke, P. R. Baldwin, and W. Chiu, J. Struct. Biol. 128, 82 (1999).
13
M. van Heel, Ultramicroscopy 21, 111 (1987).
reprojected, and the members of the initial classes are then compared to
this set of projections and reassigned to new classes based on their fit to
the projections. The new classes are averaged, and an iterative process of
model construction, reprojection, and reclassification is begun.
Final Reconstruction Parameters

Within the first few iterations, several parameters should be assessed.
One is to determine whether or not the particles are in a preferred orienta-
tion. The set of calculated Euler angles for the class averages should be
examined to see if any sets of views are excluded (Fig. 2A). Side-by-side
comparison of the class averages and their corresponding projections from
the progressing iterations provides a visual evaluation of the improving
correlation between them (Fig. 2B).
Another important assessment is whether the classes consist of very
small numbers of particles. This can be an indicator of significant hetero-
geneity in the sample. It is possible that a distinct subset of classes will
be evident, indicating an alternate conformation rather than random disso-
ciation of a complex. In this case, it may be possible to segregate the data
and compute independent reconstructions of the alternate state. The itera-
tive refinements are allowed to run until changes between successive
Fig. 2. Class averages: angular distribution and comparison to projections. (A) The dots
in the triangle represent a well-spread distribution of orientations the particles have been
classed into. The apex of the triangle is the view down the z-axis (top) of the particle, and the
base represents views along the x-axis (edge). The brighter the dot, the more particles have
been assigned to that orientation. If the dots appear clustered with gaps in the distribution, the
particle has probably assumed a preferred orientation on the EM support film. (B) A good
indicator of convergence is obtained by noting a positive match when comparing the class
averages (top row) with projections of the refined model at the same orientation (bottom
row). If an arithmetic subtraction is performed between the two, the resulting difference
image should consist of undifferentiated noise.
iterations are minimized, and are monitored by examining the Fourier shell
correlation (FSC) between rounds (Fig. 3A and B).
The resolution of the reconstruction can be estimated by dividing the
data into halves, for instance, by separating then averaging all the even
and odd numbered images in all the classes then independently recon-
structing from the even/odd averages. A FSC between the two reconstruc-
tions is examined (Fig. 3C), and the resolution at the correlation value of
0.5 provides a reasonable estimate of the resolution of the reconstruction.14
If the number of images comprising each class is very small, indicating
either sample or staining heterogeneity, the half-averages will be very
noisy, and the resolution may be underestimated.
Accuracy of the Reconstruction

The final question becomes how to determine if the reconstruction is
correct. For a wholly asymmetric object, the reconstruction algorithms
might be producing only one of many possible solutions and there is no
quantitative way to determine if the one generated is the correct one.
The most direct approach is to calculate reconstructions from more than
one unique data set and comparing the results qualitatively (see section
on Visualization,), and quantitatively, by calculating the FSC between the
wholly independent reconstructions.11
Visualization
The 3D model can be studied by slicing through the volume and exam-
ining contour maps of the densities, and by creating isocontour surface
maps to view the outer shape of the molecule (Fig. 4). Selecting the thresh-
old at which to contour a surface is done by most software packages
from the molecular weight of the sample (assuming a protein density of
1.35 g/cc). The ‘‘true’’ surface may be up to 30% of this value. In negative
stain, it is quite difficult to establish the correct cutoff. For instance, in UA
stained objects, there is always the possibility of partial positive staining of
the complex, especially when nucleic acids are present. These will then
have an opposite density from stain-excluding regions and may be misinter-
preted as ‘‘holes.’’ Surfaces with deep pockets that accumulate stain may
also be difficult to delineate accurately. There are a variety of software
packages designed to perform visualization tasks. Chimera (www.cgl.ucs-
f.edu/chimera) and Vis5D (www.ssec.wisc.edu/billh/vis5d.html) are free-
ware packages we have used; AVS (www.avs.com) and Amira
14
I. I. Serysheva, M. Schatz, M. van Heel, W. Chiu, and S. L. Hamilton, Biophys. J. 77,
1936 (1999).
Fig. 3. Fourier shell correlation (FSC): convergence and estimation of resolution. In

this convergence plot, the thin lines represent the FSC between subsequent iterations of a
refinement. As refinement proceeds, these become more consistent. The thin solid lines (a, b)
are plots between the (a) first and second and the (b) sixth and seventh rounds. Convergence
is observed to begin after the fifth round. The bold solid line (c) is the FSC between two
reconstructions each calculated from one half of the total data. A reliable estimate of
resolution is the radius at 0.5, in this plot equal to 30 A.
(www.tgs.com) are commercial packages widely used by structural biolo-

gists and currently in use in our laboratory.
SPR from ECM Images

ECM data offer the potential for considerably higher resolution than
negative stain. While it is advisable to collect ECM data on an FEG-
equipped microscope, the choice of film or CCD acquisition is subject to
the same concerns as for conventional EM (see the previous section).
The basic flow of image processing for SPR is the same, but consideration
of the CTF becomes more important. Images must be collected at several
defocus values to fill in missing information at the nodes of the CTF15 and
the image data must be corrected for phase reversals.12 Space limitations
15
J. F. Conway and A. C. Steven, J. Struct. Biol. 128, 106 (1999).
Fig. 4. Visualization. The 3D reconstruction can be examined as 2D-slices with contour

lines (A) to convey information about protein density (solid lines). The dotted lines are
contours of negative density, resulting from accumulation of stain at pockets in the structure.
The structure can also be viewed as a solid model, with the surface drawn from a selected
isoline. In (B) the surface for the 3D model has been constructed at the contour ¼ 1.0, which
was calculated to enclose a volume equal to the mass of the particle. Density and volume
information from frozen-hydrated samples is much more reliable than from negative stain.
preclude a full discussion of this aspect of the image processing in the

context of this article; for an excellent review, see Tao and Zhang (2000).16
Electron Tomography (Fig. 1C and D)

While many chromatin remodeling enzymes and other regulators of
transcription may be amenable to SPR, their substrate—chromatin—is
not. Although comprised of similar ‘‘subunits,’’ the nucleosomes, chroma-
tin itself is highly variable in conformation. While an enzyme such as SWI/
SNF may have a unique binding site on the nucleosome, in EM prepar-
ations, no two arrays of SWI/SNF-bound chromatin will appear the same,
and the image averaging that is key to SPR is not possible. To determine
the 3D structure of, for example, a chromatin-bound transcriptional
regulatory complex requires the use of EM tomography.
For tomography, the EM data are collected as a series of images as the

specimen is tilted between 70 and þ70 in a holder designed for that pur-
pose. The most common approach is to use a single tilt axis, as discussed
later. An alternative is to collect data for two tilt axes. This requires a
16
Y. Tao and W. Zhang, Curr. Opin. Struct. Biol. 10, 616 (2000).
tilt/rotate specimen holder and adds to the complexity of data acquisition

and processing, but provides increased structural information.17 The tilt
images are aligned with respect to one another, and then a 3D volume is
calculated via back-projection of the tilt data.18
EM tomography can be performed on either negatively stained or ECM
preparations. The advantage of ECM is the retention of the solution con-
formation of the sample. The significant investment of resources required
to create the infrastructure to perform tomographic data collection (com-
puter controlled stage, high-tilt specimen holder, CCD camera) is further
increased for frozen-hydrated samples.
Specimen Preparation for Tomography

Whether negative stain or frozen-hydrated samples are used, the sup-
port film should be mounted on thin-bar, hexagonal 600-mesh grids. These
optimize mechanical support and open area to accommodate tilting. Prior
to use, colloidal gold beads are applied to the back surface of the carbon.
While tilt series can be aligned by cross-correlation of the imaged particles,
in tilt series of negative stain and frozen-hydrated samples, the dose/image
is so low that the specimen may be barely discernible. The high-density
gold beads serve as fiducial markers for the alignment process. For ECM
tomography, it is also possible to mix the gold beads with the sample.
Grid preparation:
1. Immediately before use, apply a drop of 0.02% polylysine, pH 8.0 to
the back side of the grid for 30 s.
2. Wick off with filter paper.
3. Wash with 5 drops of ddH20, leaving the last drop on the grid for
30 s before wicking and air drying.
4. 3 l of colloidal gold solution (bead size 5–10 nm) is applied to the
same side, wicked, and rinsed with 5 drops ddH2O and air-dried.
5. The sample is applied to the front surface of the grid and negatively
stained as described above.
The concentration of gold solution will need to be experimentally de-
termined so that least 5 beads are present in the recorded image field.
For tomography, the sample should be somewhat more widely spaced than
for conventional imaging, to preclude particles outside the area of interest
entering the field during tilting.
17
P. Penczek, M. Marko, K. Buttle, and J. Frank, Ultramicroscopy 60, 393 (1995).
18
M. Radermacher, in ‘‘Electron Tomography’’ (J. Frank, ed.), p. 91. Plenum, New York,
1992.
Data Collection

Most specimen holders can only be tilted through 60 . High-tilt

holders for tomography, which range from 70 to 80 are available for
room temperature or ECM work (Gatan, www.gatan.com, Fischione,
www.fischione.com). There are also holders designed for double-tilt data
collection, in which after the initial tilt series is collected, the sample is

rotated 90 in the holder and a second series collected.
For acquiring tilt series of negatively stained or frozen-hydrated
specimens, the ability to record extremely low-dose images is required.
For this, a high-sensitivity CCD is invaluable, as images with electron doses
of <1 e/A2 can be recorded. The dose required for a single image with an
optimal damage should be distributed signal/noise ratio (S/N) without spe-
cimen among all the images of a tilt series, resulting in an equivalent S/N in
the back-projected reconstruction.19 With this approach, it is possible to
obtain high-quality reconstructions from images in which the sample is
barely discernible.20
The number of tilt images needed is determined by the desired reso-
lution in the reconstruction, but may be constrained by the beam sensitivity
of the sample. A guideline for the number of tilts required is given by
n ¼ t/r, where t is the thickness of the sample, r is the desired resolution,
and n is the number of equally spaced tilts required.21 A series of low-dose
images taken sequentially over the same field can supply information on
the maximum dose the sample can withstand. The low S/N of images
recorded at 1 e/A2 makes it highly desirable to use automated data
collection, where computer control of the microscope enables successive
fields to be accurately positioned via cross-correlation, and auto-focusing
is available.22
We employ a strategy for data collection which permits the collection of

at least one full 5 tilt series, then a 2.5 interleaved series, and finally, if de-

sired, a 1.25 series. An alternative to using equidistantly spaced angles in
the tilt series is to collect at a more finely spaced increment in the high
tilts.21 This method may be advantageous for very thin, flat samples to
increase resolution in the z-direction of the reconstruction.23
Recording images:
1. In Search mode, select an area to record, preferably in the center
1/3 of the grid.
19
B. F. McEwen, K. H. Downing, and R. M. Glaeser, Ultramicroscopy 60, 357 (1995).
20
R. A. Horowitz, A. J. Koster, J. Walz, and C. L. Woodcock, J. Struct. Biol. 120, 353 (1997).
21
R. A. Crowther, D. J. DeRosier, and A. Klug, Proc. Royal Soc. London A 317, 319 (1970).
22
A. J. Koster, H. Chen, J. W. Sedat, and D. A. Agard, Ultramicroscopy 46, 207 (1992).
23
W. O. Saxton, W. Baumeister, and M. Hahn, Ultramicroscopy 13, 57 (1984).
2. If the object being studied is asymmetric, the longest axis should

be aligned to within 15 of the tilt axis as resolution is degraded
perpendicular to the tilt axis.
3. Adjust eucentric height to the highest possible precision.
4. Observed at maximum tilts to be sure neither grid bars nor other
objects in the sample obscure the area of interest.
5. Ascertain that the focus area has high-contrast features or gold
beads in an easily recognizable pattern to facilitate repositioning
after tilting.

6. Record the first 0 image. Confirm that there are five or more gold
beads in the field to be reconstructed.
7. Go to maximum positive tilt.

8. Record images at 5 intervals through 0 .
9. Go to maximum negative tilt.


11. Take second 0 image.

12. Go to maximum positive tilt, decrease by 2.5 .


14. Go to maximum negative tilt, decrease by 2.5 .


16. Take third 0 image.

17. Repeat from maximum tilts offset by 1.25 .
The purpose of starting at high tilt each half-pass is to acquire the max-
imum amount of high-tilt information before the specimen has accumu-
lated much dose. The series of 0 images is used to assess beam damage
with cumulative dose.
Alignment and Reconstruction

Alignment of the tilt series can be accomplished by cross-correlation,
least squares fitting of fiducial markers, or a combination both methods.
Software packages are available to accomplish either or both of these tasks
(see http://3dem.ucsd.edu/#sw). Alignment proceeds until only sub-pixel
corrections are necessary. The human eye is an exquisitely sensitive de-
tector for the quality of an alignment; viewing the aligned tilt images as a
movie allows one to identify any deficiencies in alignment. This strategy
also provides initial information about the shape of the structure of inter-
est, and whether excessive flattening or other preparative artifacts are
present.
Reconstruction by weighted back-projection is the most common algo-
rithm employed; other methods (e.g., maximum entropy, algebraic recon-
struction ‘‘ART’’ and simultaneous iterative reconstruction ‘‘SIRT’’) are
also in use. Software environments such as SPIDER and SUPRIM have ex-
tensive support for tomographic reconstruction, as do dedicated packages
such as IMOD (http://bio3d.colorado.edu/imod). Several national re-
sources (see later section) have significant tomography development pro-
jects underway, and can provide expert consultation, software, and access
to state-of-the-art equipment.
Determination of Resolution in the Reconstruction

Several factors affect the resolution of tomographic reconstructions.

The ‘‘missing wedge’’ of data from lack of images recorded between 70

and 90 creates smearing or exclusion of features in the z-direction. Also,
the number of tilts and the angular increment determine the ‘‘worst case’’
resolution (r) as r ¼ t/n, the ‘‘worst case’’ being when the volume is
densely packed with information.21,24 Although reconstructed volumes
may be low-pass filtered to exclude specimen details smaller than r, our ex-
perience with chromatin is that this level of filtering may obscure valuable
information. For instance, linker DNA is frequently seen in tomographic
reconstructions of chromatin where the theoretical resolution limit would
preclude its visualization.25,20
For a qualitative evaluation of the tomographic reconstruction, one

compares the 0 projection of the final reconstruction with the first 0 raw
image. It is useful to do the same for several tilts also. A careful examin-
ation of a 1-pixel thick central slice of the reconstructed volume will
provide an indication of what detail is likely to be available in the volume.
Visualization and Analysis

The most accessible way of evaluating the data is to slice the volume
into 2D xy plane slices (Fig. 1C), to create a z-axis series. It is often highly
informative to view this z-series as a movie, to study transitions between
structures in the interior of the volume. The volume can be sliced to 1-pixel
thickness, to generate highly detailed views of substructure not visible in
projections of the whole reconstructed volume.
The most difficult aspect of tomography is understanding the 3D struc-
ture in the volumes. For chromatin, the use of isosurface contouring has
proven unsatisfactory. The most useful information is accessed by viewing
the entire volume as its original greyscale data, rendered with a transpar-
ency gradient to allow seeing into the volume. This task remains difficult
24
C. L. Woodcock, in ‘‘Electron Tomography’’ (J. Frank, ed.), p. 313. Plenum, New York,
1992.
25
R. A. Horowitz, D. A. Agard, J. W. Sedat, and C. L. Woodcock, J. Cell Biol. 125, 1 (1994).
computationally with the typically large data sets involved. An ideal com-
putational visualization environment would permit access to the greyscale
data while having a transparent volume to interactively move on arbitrary
as well as orthogonal axes, and incorporate 3D measurement tools. While
freeware volume visualization packages are available, currently each
falls short in some aspect. Our laboratory currently uses the commercial
package, Amira (www.tgs.com).
National Resources available in the United States.
Boulder Laboratory for 3-Dimensional Fine Structure bio3d.Colorado.edu

Tomography
National Center for Microscopy and Imaging Research www.ncmir.ucsd.edu
Tomograph y
National Center for Macromolecular Imaging ncmi.bcm.tmc.edu/ncmi
SPR
Resource for Visualization of Biological Complexity www.wadsworth.org/rvbc
[4] Electron Microscopic Analysis of the RSC

Chromatin Remodeling Complex
By Francisco J. Asturias, Chukwudi Ezeokonkwo,
Roger D. Kornberg, and Yahli Lorch
Introduction: Use of Electron Microscopy for Analysis of Large

Macromolecular Complexes
Macromolecular electron microscopy (EM) is an ideal technique for
structural characterization of large (>300,000 Da) macromolecular com-
plexes, which may be too difficult to purify to homogeneity in the large
quantities required for more conventional techniques, such as X-ray crys-
tallography. Single-molecule–thick 2 D crystals suitable for EM crystallo-
graphic analysis can be prepared using the lipid layer crystallization
technique,1,2 which requires two to three orders of magnitude less material
(typically 10–100 g of protein) than would be required for preparation of
3D crystals. Even smaller amounts of material (1–5 g) are required for the
preparation of samples from which single-particle images can be obtained
1
F. J. Asturias and R. D. Kornberg,J. Biol. Chem. 274, 6813 (1999).
2
E. E. Uzgiris and R. D. Kornberg, Nature 301, 125 (1983).

computationally with the typically large data sets involved. An ideal com-
putational visualization environment would permit access to the greyscale
data while having a transparent volume to interactively move on arbitrary
as well as orthogonal axes, and incorporate 3D measurement tools. While
freeware volume visualization packages are available, currently each
falls short in some aspect. Our laboratory currently uses the commercial
package, Amira (www.tgs.com).
National Resources available in the United States.
Boulder Laboratory for 3-Dimensional Fine Structure bio3d.Colorado.edu

Tomography
National Center for Microscopy and Imaging Research www.ncmir.ucsd.edu
Tomography
National Center for Macromolecular Imaging ncmi.bcm.tmc.edu/ncmi
SPR
Resource for Visualization of Biological Complexity www.wadsworth.org/rvbc
[4] Electron Microscopic Analysis of the RSC

Chromatin Remodeling Complex
By Francisco J. Asturias, Chukwudi Ezeokonkwo,
Roger D. Kornberg, and Yahli Lorch
Introduction: Use of Electron Microscopy for Analysis of Large

Macromolecular Complexes
Macromolecular electron microscopy (EM) is an ideal technique for
structural characterization of large (>300,000 Da) macromolecular com-
plexes, which may be too difficult to purify to homogeneity in the large
quantities required for more conventional techniques, such as X-ray crys-
tallography. Single-molecule–thick 2D crystals suitable for EM crystallo-
graphic analysis can be prepared using the lipid layer crystallization
technique,1,2 which requires two to three orders of magnitude less material
(typically 10–100 g of protein) than would be required for preparation of
3D crystals. Even smaller amounts of material (1–5 g) are required for the
preparation of samples from which single-particle images can be obtained
1
F. J. Asturias and R. D. Kornberg, J. Biol. Chem. 274, 6813 (1999).
2
E. E. Uzgiris and R. D. Kornberg, Nature 301, 125 (1983).

[4] electron microscopic analysis 49
and later analyzed to generate structures at a resolution of 10–30 Å.3 We

have used single-particle image analysis techniques to characterize the
structure of RSC, an abundant, essential chromatin remodeling complex
from the yeast, Saccharomyces cerevisiae.4 A 3D reconstruction of RSC
obtained from images of negatively stained particles suggests a possible
mode of interaction between the complex and a nucleosome core particle.
Sample Preparation for Electron Microscopy of Single Particles

Preserved in Stain
Preservation in a heavy metal stain offers a quick and simple method to
preserve biological macromolecules for imaging in the electron micro-
scope. A widely used stain is uranyl acetate, which works well with many
molecular complexes. The electron-dense stain is excluded from the
volume occupied by the specimen, and what is actually recorded is there-
fore an imprint of the macromolecule in the stain. The advantages of using
a heavy metal salt stain are the ease and speed of specimen preparation and
imaging. However, the specimen is visualized only indirectly and at low
resolution, and in most cases is partially deformed as a consequence of de-
hydration and interaction with the substrate to which the macromolecules
are adsorbed. Despite its shortcomings, preservation in stain is the best al-
ternative for initial characterization of a macromolecule, and can provide
useful structural information. In addition, a structure calculated from
specimens preserved in stain can in many instances be used as a starting
point for more precise structural characterization based on images of
unstained specimens preserved in amorphous ice.
Specimens are typically prepared using carbon-coated grids. To
prepare the grids, a thin (10 nm) layer of amorphous carbon is evapor-
ated onto a freshly cleaved mica surface. The carbon layer is released
by slowly pushing the carbon-coated mica, carbon-coated side facing

upward, at an angle (45 ) through an air-water interface. The grids to
be carbon coated are previously arranged on a support below the water sur-
face. After release from the mica, the carbon film is carefully manipulated
to position it on top of the grids, and is deposited on them by slowly
lowering the water level until the carbon film comes into contact with
the grids.
3
J. Frank, ‘‘Three-Dimensional Electron Microscopy of Macromolecular Assemblies.’’
Academic Press, San Diego (1996).
4
F. J. Asturias, W. H. Chung, R. D. Kornberg, and Y. Lorch, Proc. Natl. Acad. Sci. USA 99,
13477 (2002).
Specimens are prepared using a protein solution in a low-salt

(<100 mM) buffer, as the presence of high-salt concentration would inter-
fere with staining and create problems for adequate imaging of the speci-
men. A protein solution with a concentration in the range of 10–100 g
protein/ml usually results in an adequate density of particles in the samples.
To facilitate adsorption of the molecules to the carbon support film the
grids are glow-discharged at a residual pressure of 2 10 1 mbar for
30–60 s. The nature of the gas used during glow discharge (air, water vapor,
and amyl amine are typical choices) can be varied to affect the properties
of the carbon film after glow discharge. Two to three microliters of the pro-
tein solution are placed on the surface of a freshly glow-discharged grid
(the effect of glow discharge on the carbon surface only lasts for a few min-
utes), which is held using microforceps. The protein solution is left on the
grid for 10–60 s to give the particles a chance to adsorb onto the support
film, and then blotted away by touching the side of the grid to a piece of
filter paper. The grid is then washed with two to three drops of 1–2% ura-
nyl acetate. The final drop of stain is left on the grid for 30–60 s to ensure
proper staining of the sample. The sample preparation process is com-
pleted by depositing a second carbon layer on top of the stained molecules.
This is accomplished by floating a small piece (5 5 mm) of carbon on
the surface of a pool (5 ml) of stain using the same procedure employed
to release the carbon film used to coat the grids. The stained sample is then
immersed in the stain pool and withdrawn through the interface to pick up
the second carbon layer.5,6 Excess stain is removed by bringing a piece of
filter paper into contact with the backside of the grid.
Preliminary Specimen Evaluation and Data Collection

While specimens preserved in negative stain are relatively resistant to
radiation damage, it is nonetheless important to minimize the electron dose
received by the specimen prior to imaging. Initial evaluation of the speci-
mens entails examining the grids at low (2000) magnification. If the con-
centration of the protein solution used to prepare the specimens was
adequate, properly stained specimens will have a ‘‘grainy’’ appearance,
and examination at higher magnification (35,000) will reveal nicely
stained particles. Properly stained particles should be clearly distinguish-
able from the background and show internal detail. If the amount of pro-
tein adsorbed on the carbon support surface was too low, very little stain
5
G. W. Tischendorf, H. Zeichhardt, and G. Stoffler, Mol. Gen. Genet. 134, 187 (1974).
6
G. Stoffler and M. Stoffler-Meilicke, in ‘‘Modern Methods in Protein Chemistry’’
(H. Tesche, ed.), p. 409. De Gruyter, Berlin (1983).
will be retained and the specimen will resemble a clean carbon film.
Finally, an excessive amount of material adsorbed to the carbon support
film will result in samples where staining is evident, but no individual par-
ticles can be distinguished when an area is examined at higher (e.g.,
35,000) magnification. After determining that the staining and protein
concentration in the samples are adequate, the next step is to collect
images to evaluate the preservation of the particles and the distribution
of particle orientations in the samples. Since individual particles must be
selected for image analysis, it is important to optimize their distribution
by carefully adjusting the protein concentration used during sample prep-
aration. Ideally, particles should be evenly distributed on the support film,
and the separation between individual particles should correspond to 2–3
particle diameters. This will result in a large number of particles being in-
cluded in recorded images, while ensuring that it will be possible to select
individual particles without interference from neighboring ones.
Most of the data for single particle analysis are still collected on film,
but the use of electronic detectors is increasing, and the trend will undoubt-
edly continue. The arguments that follow apply to either type of data, as
the distinction between them largely disappears after film data has been
digitized. Zero-tilt images of the specimens are necessary to assess particle
preservation and orientation. Images for data analysis are typically
recorded at magnifications in the range 35,000–65,000. In the case of film
data, the images should be scanned using a pixel size equivalent to about
one fourth of the desired resolution. The number of pixels available for
alignment of individual particle images is often the more critical parameter
for determining how images should be digitized.
Analysis and Classification of 2D images

The RSC images used in this study were collected at a magnification of
60,000, and scanned using a pixel size corresponding to 2.54 Å in the
specimen. Once the images were digitized, individual particles were cut
out using a window size such that the side of the window corresponded
roughly to twice the largest dimension of an RSC particle. Reference-free
alignment of the particle images followed by multivariate statistical analy-
sis and hierarchical ascendant classification3 provides information about
the distribution of particle orientations, and about potential variability in
particle conformation. RSC particles were found to occur mostly on a
single preferred orientation, but to display a significant amount of variabil-
ity in conformation. At low resolution and in projection, the RSC particle
appears as formed by four globular domains arranged around a central
cavity. The results of multivariate statistical analysis and hierarchical
Fig. 1. Structure of RSC in projection. Images of RSC particles (5880) preserved in uranyl
acetate were computationally aligned. All particles seemed to be in a similar orientation but
differed in conformation and were sorted into homogeneous classes using multivariate
statistical analysis and hierarchical ascendant classification.3 A central area of lower density is
apparent in several of the class averages shown, which include 45% of all particles. Most of
the variation in RSC conformation is related to a domain forming the bottom part of the
structure, which is either missing (35% of particles), or collapsed against the top of the structure
(22% of particles).
ascendant classification of RSC particle images are shown in Fig. 1. While

three of the four domains remain fairly constant in their relative
arrangement and orientation, the fourth one (forming the lower part of
the RSC structure as shown in Fig. 1) varies in position, and in some in-
stances appears to be either absent, or collapsed against the top part of the
RSC structure. Given the large degree of variability shown by the RSC
particles, 3D reconstructions were calculated for homogeneous subsets
of particles defined by the hierarchical ascendant classification analysis.
Calculation of Initial 3-D Reconstructions

From a set of particles that correspond to a single conformation and a
single orientation, an initial 3-D reconstruction can be easily calculated
using the random conical tilt method,7,8 which is implemented by combin-
ing information from untilted and tilted images of the same specimen area.
Briefly, particles in a given orientation are selected from the untilted
images, along with their corresponding tilted pairs. In-plane rotational
alignment parameters for the untilted particle images, the tilt angle for
the corresponding tilted particle images, and information about the relative
orientation between the tilted and untilted images in a tilt pair are used to
calculate a 3D structure from the tilted particle images. If additional homo-
geneous (same conformation and orientation) particle sets can be identi-
fied, the procedure is repeated. The resulting reconstructions are then
combined to make up for the incomplete information in each individual re-
construction resulting from physical limitations in the maximum value of
the tilt angle used for data collection (missing cone problem).
We were able to calculate several reconstructions of the RSC complex
from homogeneous particle subsets identified by statistical analysis. Some
of those reconstructions are shown in Fig. 2. Comparison of the structures
confirms that the differences among them arise mostly from rearrangement
of a domain in the lower part of the structure that moves as a solid module.
The structures could not be combined because they correspond to different
conformations of the RSC complex. An additional limitation of the recon-
structions relates to particle deformation induced by preservation in stain.
This phenomenon is known to be significant, as surface tension forces and
dehydration cause particles preserved in stain to be flattened along the
direction perpendicular to the carbon support surface. The problem can
be corrected when information from different views can be combined
to calculate an improved reconstruction (different reconstructions are
deformed in different directions and the effect of such deformation is alle-
viated in the final combined volume), but the presence of different con-
formations and a single predominant orientation prevented us from using
that approach to improve the quality of the RSC reconstruction.
Interpretation of the RSC Reconstruction

As explained earlier, examination of the different 3D reconstructions of
the RSC complex shown in Fig. 2 led us to conclude that the variability ob-
served was most likely related to a change in the position of the domain
7
M. Radermacher, J. Electron Microsc. Tech. 9, 359 (1988).
8
M. Radermacher, T. Wagenknecht, A. Verschoor, and J. Frank, J. Microsc. 146, 113 (1987).
Fig. 2. Three-dimensional reconstruction of RSC. (A) RSC consists of four modules that
define a central cavity. Two views of the structure (front and back) are shown. (B) The most
significant variation in RSC conformation was due to the collapse (top) or absence (bottom)
of a module that forms the lower part of the RSC structure. The scale bar corresponds to
approximately 100 Å.
that defines the bottom portion of the central cavity. This domain seems to
partially collapse under the sample preservation conditions used in the
study. However, a majority of the RSC particles (45%) show a clearly de-
fined cavity, and that appears to be the most representative conformation
of the complex. Variability in the position of the lower domain may actu-
ally have functional significance, as it could control access to the central
cavity, which would be more accessible when the bottom domain moves
away from the rest of the structure. The shape and size of the central cavity
in the RSC complex matches closely the shape and size of a nucleosome
core particle. Possible binding of a nucleosome in the cavity was tested
by calculating a reconstruction of the RSC/nucleosome complex.
Studying the RSC/Nucleosome Interaction Under EM Sample

Preparation Conditions
Determining the structure of a complex formed by RSC and the nucleo-
some core particle (histone core plus 150 pb of double-stranded DNA)
would likely help to determine the mechanism by which RSC alters the
interaction between the histone core and nucleosomal DNA and renders
the nucleic acid more accessible to components of the transcription
apparatus. RSC interacts with the nucleosome core particle with very high
affinity (Kd 109), and therefore formation of the complex for examination
by electron microscopy should be a straightforward matter. A rough
method for quantitating formation of the RSC/nucleosome complex under
the conditions required for sample preparation was implemented as
follows.
First, samples of RSC alone were prepared, using a protein concentra-
tion that resulted in an adequate particle density. Under the same condi-
tions (same grids, same buffer conditions), samples of nucleosomes alone
were prepared, using a nucleosome concentration that resulted in a density
of nucleosome particles similar to that of the RSC particles. Once both
types of samples had been prepared, the number of particles in an arbitrary
area (the area corresponding to a frame collected with an electronic CCD
camera detector at 66,000 magnification) was determined for both the
RSC and nucleosome samples. Having determined such numbers and
checked that the desired ratio of RSC/nucleosome particles had been
obtained, samples containing both RSC and nucleosomes were prepared
in such a way that the final concentration of both RSC and nucleosomes
in the solution used to prepare the mixed samples (as well as the buffer
conditions, etc.) were identical to those previously used to prepare samples
of the individual components. The RSC/nucleosome samples were then
examined in the microscope, and the number of RSC and nucleosomes in
the previously specified unit area were determined. A decrease in the
number of nucleosomes with respect to that observed in the nucleosome
(no RSC) samples could be considered as indirect evidence for formation
of the RSC/nucleosome complex. The experiment was carried out by using
a solution of RSC containing 160 g protein/ml in a buffer containing
50 mM potassium acetate, 3 mM magnesium chloride, and 15 mM
HEPES, pH 7.5. The nucleosome samples were prepared using a solution
TABLE I
The Number of Particles in an Arbitrary Area is Listed for RSC (First Column),
Nucleosomes (Second Column), RSC in the RSC/Nucleosome Samples (Third
Column), and Nucleosomes in the RSC/Nucleosome Samples. The Last Row
Shows the Average of the Values for Each Column a
RSC/nucleosomes RSC/nucleosomes
RSC Nucleosomes (RSC count) (nucleosome count)
72 70 35 18
65 55 43 21
67 73 41 17
67 57 28 19
72 71 35 27
69 65 36 20
a
Analysis of these numbers indicates that there is an overall decrease in the number of
particles per unit area after mixing of RSC and nucleosomes, probably as a result of
mutual charge neutralization that decreases the affinity of the particles for the
amorphous carbon film support. It also reveals a 40–50% decrease in the expected
number of nucleosomes in the RSC/nucleosome particles, suggesting that about half of
the RSC particles have a bound nucleosome.
with a concentration of 50 g/ml in the same buffer. Formation of the RSC/

nucleosome complex was carried out by mixing RSC and nucleosome solu-
tions resulting in the same final concentration of both components as used
for preparation of single-component samples, and incubating the resulting

mix for 30 min at 30 . The results, summarized in Table I, indicate that
the number of free nucleosomes diminished by about 40–50% when the nu-
cleosomes were incubated with RSC. The conclusion from this analysis is
that the RSC/nucleosome complex seems to be at least partially stable
under the acidic conditions (pH 4) of the uranyl acetate solution used
to preserve the particles.
To further evaluate formation of the RSC/nucleosome complex, images
of particles selected from the RSC/nucleosome samples were aligned
and subject to statistical analysis as described before. While a relatively
large (12) number of distinguishable conformations were detected in
RSC samples (see Fig. 1), variability in the structure of RSC seems to be
significantly reduced in the presence of nucleosomes. As shown in Fig. 3,
only four distinct conformations of RSC were detected when nucleosomes
were present. The decrease in the number of free nucleosomes in the pres-
ence of RSC, and the clear stabilization of the RSC conformation in the
presence of nucleosomes, point to the stability of the RSC/nucleosome
complex under the conditions of our EM analysis. A reconstruction calcu-
lated from the largest set of RSC/nucleosome images is shown in Fig. 4.
Fig. 3. Structure of the RSC/nucleosome complex in projection. The RSC/nucleosome

complex was formed by incubating RSC and nucleosomes in a buffer containing 50 mM
potassium acetate, 3 mM magnesium chloride, and 15 mM HEPES, pH 7.5. A total of 9260
images of particles preserved in uranyl acetate were computationally aligned and sorted into
homogeneous classes using multivariate statistical analysis and hierarchical ascendant
classification.3 As observed for RSC alone, most particles appeared to be in the same
orientation. However, interaction with the nucleosome seems to stabilize the structure of the
RSC complex, as evidenced by the much smaller number of homogeneous groups generated
by statistical analysis.
Additional density is present in the central cavity, in agreement with our

suggestion that the nucleosome might bind in that location. However, the
amount of additional density observed is less than that expected on the
basis of the molecular weight of the nucleosome core particle. Such dis-
crepancy is most likely due to inclusion of unoccupied RSC particles in
the RSC/nucleosome image data set, as indicated by the data summarized
in Table I. Statistical analysis of the untilted RSC/nucleosome images
(where the central cavity is clearly visible and does not overlap significantly
Fig. 4. Three-dimensional reconstruction of the RSC/nucleosome complex. The overall

structure of the RSC complex remains constant after incubation with nucleosomes, but
additional density is apparent in the central cavity. However, the additional density only
corresponds to about half of that expected based on the molecular weight of the nucleosome
core particle. Partial detection of the nucleosome is most likely due to the presence of
unoccupied RSC particles in the data set, and to problems with staining caused by charged
surfaces in the nucleosome and the RSC central cavity.
with the surrounding density that defines it) corresponding to the tilted par-
ticle images used to calculate the RSC/nucleosome reconstruction was
carried out to attempt to separate empty and occupied RSC particles.
The results from such analysis (not shown) were inconclusive, perhaps be-
cause the area corresponding to the central cavity includes only a relatively
small number of pixels. Incomplete formation of the RSC/nucleosome
complex detected by the particle count study (Table I), along with staining
artifacts produced by the highly charged nature of the histone core and
DNA surfaces, must contribute to the detection of an artificially low
density in the central RSC cavity.
The structure of the RSC/nucleosome complex strongly suggests that
interaction of a nucleosome with RSC occurs by binding of the nucleosome
to the central cavity in the RSC complex. The size and shape of the central
cavity match closely the size and shape of a nucleosome core particle. This
is illustrated in Fig. 5, which shows our proposed model for RSC/nucleo-
some interaction. The RSC complex is shown in its prevalent confor-
mation, and a low-resolution (25 Å) structure of the nucleosome core
particle (obtained by low-pass filtering of the atomic-resolution X-ray
structure of the complex9) has been placed in the central RSC cavity.
9
K. Luger, A. W. Mader, R. K. Richmond, D. F. Sargent, and T. J. Richmond, Nature 389,
251 (1997).
Fig. 5. Possible mode of RSC-nucleosome interaction. An X-ray structure of the

nucleosome9 was filtered to 25 Å and manually fitted in the central cavity of the RSC
structure [using the program O (T. A. Jones, J. Y. Zou, S. W. Cowan, and M. Kjeldgaard, Acta
Crystallogr. A 47 (Pt 2), 110 (1991).)]. The close fit between the nucleosome and the RSC
cavity is apparent. The scale bar corresponds to approximately 100 Å.
The figure illustrates the close fit of the nucleosome in the cavity. The
model shown in Fig. 5 is supported by results from nuclease digestion stud-
ies of nucleosomal DNA in an RSC nucleosome complex, which indicate
that digestion of nuclesomal DNA occurs preferentially at positions that
would be more exposed in the model.4
Examination of Unstained RSC and RSC/Nucleosome Particles

While preservation of single particles in a heavy metal stain provides a
simple and convenient way of obtaining structural information about a
macromolecular complex, the results suffer from limitations in resolution
and in the fidelity of the resulting volumes. These problems can be avoided
by examining unstained particles preserved in amorphous ice (cryo-

electron microscopy), which makes possible direct imaging of molecular
features, and has been shown to preserve molecular structure to near
atomic resolution.10–12 Preparation of frozen-hydrated samples has been
extensively described13 and will not be repeated here. Instead, some of
the issues encountered in applying the technique to the study of the RSC
complex will be presented.
The first factor to consider is the type of grids to be used for sample
preparation. Unstained particles can be adsorbed to a thin amorphous
carbon support film, as in the case of stained samples. In fact, some of
the highest resolution structures obtained by single particle analysis
methods have actually been calculated from images of particles adsorbed
to a support film.14 The presence of a support film also makes it possible
to determine more accurately the defocus value of the images, a parameter
that is critical for calculation of a meaningful reconstruction.15 Particles can
also be preserved in a thin, unsupported film of amorphous ice, which is
usually formed across holes in a relatively thick perforated carbon support
film.16 Particles preserved in an unsupported amorphous ice film are not
disturbed by interaction with a substrate, and there is no contribution to
the background signal from a continuous support film. However, under
those conditions particles are often found to preferentially interact with
the air/water interface, and this can interfere with their preservation. Also,
sample preparation usually requires the use of a protein solution with a
much higher (>10-fold higher) concentration, as the particles often tend
to adsorb to the edges of the perforated carbon support.
Frozen-hydrated RSC samples were prepared using grids with a con-
tinuous carbon support film. The experiments were complicated by appar-
ent interaction of the particles with the air/water interface, even in the
presence of the support film. Nonetheless, a number of RSC images suit-
able for analysis were obtained. There are two main differences between
images of RSC particles preserved in amorphous ice and those of RSC par-
ticles preserved in stain. First, the frozen-hydrates particles no longer show
a preferred orientation on the substrate film. Second, the conformation of
10
A. Miyazawa, Y. Fujiyoshi, M. Stowell, and N. Unwin, J. Mol. Biol. 288, 765 (1999).
11
G. Ren, A. Cheng, V. Reddy, P. Melnyk, and A. K. Mitra, J. Mol. Biol. 301, 369 (2000).
12
B. Bottcher, S. A. Wynne, and R. A. Crowther, Nature 386, 88 (1997).
13
J. Dubochet, M. Adrian, J. J. Chang, J. C. Homo, J. Lepault, A. W. McDowall, and
P. Schultz, Q. Rev. Biophys. 21, 129 (1988).
14
J. Frank, P. Penczek, R. K. Agrawal, R. A. Grassucci, and A. B. Heagle, Methods Enzymol.
317, 276 (2000).
15
P. Penczek, M. Radermacher, and J. Frank, Ultramicroscopy 40, 33 (1992).
16
P. Penczek, R. A. Grassucci, and J. Frank, Ultramicroscopy 53, 251 (1994).
the particles appears essentially homogeneous (i.e., there is no variation in

the position of the lower module of the RSC structure). Images of particles
preserved in amorphous ice were classified into homogeneous groups by
aligning them to a set of reference projections of the RSC 3D reconstruc-
tion calculated from images of particles preserved in stain. To eliminate
any possible reference-related bias, reference-free alignment was used to
examine the homogeneity and alignment of each one of the particle groups
defined by the classification. A gallery of the resulting reference-bias–free
averages is shown in Fig. 6. The final step in obtaining a reconstruction
Fig. 6. Structure of unstained RSC in projection. Images of unstained, frozen-hydrated

RSC particles were separated into groups based on their similarity to projections of the RSC
reconstruction calculated from particles preserved in negative stain. The members of each
homogeneous group were then realigned without use of a reference to avoid possible bias.
Projections of the reference structure calculated from stained particles are shown in the four top
rows. The corresponding group averages from unstained particles are shown in the bottom four
rows. Improving the homogeneity of the groups, and determining the relative orientation of the
resulting group averages, will make possible calculation of an improved RSC reconstruction.
of the RSC complex from the unstained particle images entails combining
the information present in this set of projections. This requires precise
knowledge of their relative orientations. Evidently, approximate informa-
tion is provided by comparison with the available stain reconstruction.
However, preliminary analysis indicates that the deformation induced by
preservation in stain is significant, and this prevents direct use of the stain
reconstruction as a meaningful reference for calculation of an improved re-
construction from the unstained particle images. We are currently working
to obtain a suitable reference volume, which will allow us to make full use
of the information contained in images of unstained RSC particles.
[5] Use of Optical Trapping Techniques to Study

Single-Nucleosome Dynamics
By Brent Brower-Toland and Michelle D. Wang
Introduction
Over the past decade, optical trapping techniques have become a stan-
dard part of the repertoire of tools available for the study of biological
molecules.1–5 More recently, optical trapping techniques have been applied
to the study of chromatin structure and even details of the structure of
individual nucleosomes in a chromatin array.6–8
The general experimental design of optical trapping experiments with
chromatin involves immobilization of one end of a linear DNA molecule
on a surface, while the other end of the molecule is attached to a polystyrene
microsphere (bead). The microsphere can then be used as a microscopic
‘‘handle’’ which can be captured and manipulated by the optical trap
(Fig. 1). The optical trap can be used to exert and measure piconewton-scale
1
K. Svoboda, C. F. Schmidt, B. J. Schnapp, and S. M. Block, Nature 21, 365 (1993).
2
H. Yin, M. D. Wang, K. Svoboda, R. Landick, J. Gelles, and S. M. Block, Science 270,
1653 (1995).
3
M. D. Wang, M. J. Schnitzer, H. Yin, R. Landick, J. Gelles, and S. M. Block, Science 282,
902 (1998).
4
J. Liphardt, B. Onoa, S. B. Smith, I. Tinoco, and C. Bustamante, Science 292, 733 (2001).
5
S. J. Koch, A. Shundrovsky, B. C. Jantzen, and M. D. Wang, Biophys. J. 83, 1098 (2002).
6
Y. Cui and C. Bustamante, Proc. Natl. Acad. Sci. USA 97, 127 (2000).
7
M. L. Bennink, S. H. Leuba, G. H. Leno, J. Zlatanova, B. G. de Grooth, and J. Greve, Nat.
Struct. Biol. 8, 606 (2001).
8
B. Brower-Toland, R. C. Yeh, C. Smith, C. L. Peterson, J. T. Lis, and M. D. Wang, Proc.
Natl. Acad. Sci. USA 99, 1960 (2002).

of the RSC complex from the unstained particle images entails combining
the information present in this set of projections. This requires precise
knowledge of their relative orientations. Evidently, approximate informa-
tion is provided by comparison with the available stain reconstruction.
However, preliminary analysis indicates that the deformation induced by
preservation in stain is significant, and this prevents direct use of the stain
reconstruction as a meaningful reference for calculation of an improved re-
construction from the unstained particle images. We are currently working
to obtain a suitable reference volume, which will allow us to make full use
of the information contained in images of unstained RSC particles.
[5] Use of Optical Trapping Techniques to Study

Single-Nucleosome Dynamics
By Brent Brower-Toland and Michelle D. Wang
Introduction
Over the past decade, optical trapping techniques have become a stan-
dard part of the repertoire of tools available for the study of biological
molecules.1–5 More recently, optical trapping techniques have been applied
to the study of chromatin structure and even details of the structure of
individual nucleosomes in a chromatin array.6–8
The general experimental design of optical trapping experiments with
chromatin involves immobilization of one end of a linear DNA molecule
on a surface, while the other end of the molecule is attached to a polystyrene
microsphere (bead). The microsphere can then be used as a microscopic
‘‘handle’’ which can be captured and manipulated by the optical trap
(Fig. 1). The optical trap can be used to exert and measure piconewton-scale
1
K. Svoboda, C. F. Schmidt, B. J. Schnapp, and S. M. Block, Nature 21, 365 (1993).
2
H. Yin, M. D. Wang, K. Svoboda, R. Landick, J. Gelles, and S. M. Block, Science 270,
1653 (1995).
3
M. D. Wang, M. J. Schnitzer, H. Yin, R. Landick, J. Gelles, and S. M. Block, Science 282,
902 (1998).
4
J. Liphardt, B. Onoa, S. B. Smith, I. Tinoco, and C. Bustamante, Science 292, 733 (2001).
5
S. J. Koch, A. Shundrovsky, B. C. Jantzen, and M. D. Wang, Biophys. J. 83, 1098 (2002).
6
Y. Cui and C. Bustamante, Proc. Natl. Acad. Sci. USA 97, 127 (2000).
7
Struct. Biol. 8, 606 (2001).
8
B. Brower-Toland, R. C. Yeh, C. Smith, C. L. Peterson, J. T. Lis, and M. D. Wang, Proc.
Natl. Acad. Sci. USA 99, 1960 (2002).

[5] optical trapping techniques 63
Fig. 1. Instrument and experimental configuration (not to scale). (A) Overall design of
the instrument. The drawing is modified from Wang et al.9 (B) Experimental configuration.
Under feedback control, a nucleosomal array is stretched between the surface of a microscope
coverslip and an optically trapped microsphere.
forces and nanometer-scale displacements on a nucleosomal DNA

molecule.
Optical trapping techniques are single-molecule techniques that allow
mechanical manipulation of a nucleosomal DNA molecule under physio-
logical solution conditions. Because the sample is immobilized, solution
conditions and sample components can actually be varied during the course
of experimentation. The experiments can be nondestructive, permitting
repeated sampling of the same nucleosomal arrays. An added advantage
of the use of an optical trap is the freedom to consider individual nucleoso-
mal structure in the context of a nucleosomal array, rather than on isolated
mononucleosomes. These features distinguish optical trapping techniques
from others in the repertoire of tools available for the study of chromatin.
In this chapter, we describe the optical trapping system and experimental
sample preparation techniques necessary to carry out dynamic structural an-
alysis of individual nucleosomes in nucleosomal arrays. We anticipate that op-
tical trapping experiments will prove valuable in answering questions about
chromatin structure that are difficult to access with traditional techniques.
Chromatin Sample Preparation
Preparation of Histones
Owing to the sensitivity of single molecule studies of chromatin, the
use of highly purified biochemical components is critical to the success of
these experiments. Large quantities of highly purified histone proteins
can be prepared from various chromatin samples using standard hydroxya-
patite (HAP) chromatography. Starting with washed nuclei from the tissue
source of choice, their chromatin content is fragmented by mild MNase
digest, and bound to HAP (BioGel HTP, BioRad Laboratories, Hercules,
CA) by virtue of their nucleic acid component as previously detailed.10,11
Linker histone and nonhistone proteins are removed by washing the
HAP bed at moderate ionic strengths. Finally, core histone proteins are
eluted from the HAP:DNA complex at high ionic strength.
Alternatively, highly purified recombinant histone proteins can be
obtained by expression in bacteria, permitting choice and manipulation
of primary sequence.12 This flexibility permits the design of experiments
9
M. D. Wang, H. Yin, R. Landick, J. Gelles, and S. M. Block, Biophys. J. 72, 1335 (1997).
10
A. P. Wolffe and K. Ura, Methods 12, 10 (1997).
11
G. Schnitzler, in ‘‘Current Protocols in Molecular Biology’’ (F. A. Ausubel et al., eds.), 3,
p. 21.5.7. Wiley, New York, 2001.
12
K. Luger, T. J. Rechsteiner, and T. J. Richmond, Methods Enzymol. 304, 3 (1999).
involving minor histone variant proteins, critical structural mutants, and

the introduction of cross-linking moieties.13
DNA Labeling and Attachment Methods

Selection of a DNA template for chromatin assembly and subse-
quent single-molecule analysis involves two major criteria: template length
and sequence. With the experimental configuration previously described by
Brower-Toland et al.,8 a DNA template of around 4000 bp is optimal.
A shorter DNA template will lead to greater uncertainty in the deter-
mination of the DNA extension and the tension in the DNA.9 This
is especially a concern for a saturated nucleosomal DNA due to its
shortened DNA tether. A longer DNA template will introduce more
Brownian motions of the trapped microsphere. DNA sequence choice is
governed entirely by the goal of the experiment. We have previously
utilized a DNA molecule composed of 17 repetitions of the sea urchin
5S nucleosomal positioning element in order to produce nucleosomal
arrays containing, as far as possible, identical nucleosomal units.8
However, nucleosomes assembled by salt dialysis on a nonrepetitive
sequence with no bias for nucleosome positioning produce single-molecule
data of sufficient quality for analysis.14 As expected, nucleosomes on
such arrays exhibit a broader range of binding energies than those
assembled on naturally occurring positioning elements such as the 5S
sequence.
Labeling and purification of these DNA molecules have been accom-
plished by standard enzymatic manipulation and chemical fractionation
methods. Nonrepetitive sequences can be asymmetrically end-labeled using
the polymerase chain reaction (PCR) with differentially end-labeled
primers. Repetitive sequences not amenable to PCR amplification have
been labeled by Klenow fill-in reaction with the appropriate labeled NTPs.
Experiments performed in our lab have utilized biotin and digoxigenin-
labeled nucleic acids successfully. These two labels are especially convenient
because of the wide commerical availability of avidin-coated microspheres,
and the existence of sufficiently high-affinity anti-digoxigenin anti-
bodies (Roche Applied Science, Indianapolis, IN). Removal of residual
unincorporated label is critical to the success of these experiments.
13
A. Flaus, K. Luger, S. Tan, and T. J. Richmond, Proc. Natl. Acad. Sci. USA 93, 1370 (1996).
14
B. Brower-Toland and M. D. Wang, unpublished data.
Chromatin Assembly
Assembly of chromatin arrays for analysis from highly purified DNA
and histone components can be achieved by chemical (gradient salt
dialysis)15 or by enzymatic means.16 The salt dialysis method of assembly
has the advantage of preserving sample purity and minimizing the amount
of post-assembly purification required before experimentation.
Enzymatic assembly of chromatin has the pitfall of introducing a large
number of protein impurities if an assembly extract is utilized. The recent
development by the Kadonaga lab of a completely recombinant assembly
system minimizes this complication.17 Chromatin assembly by this method
is more rapid than by salt dialysis. Moreover, it has the advantage of pro-
ducing extremely regular arrays of nucleosomes in a sequence-independent
fashion, without introducing artifactual structures such as dinucleosomes.
With either technique, optimization of assembly conditions by post-
assembly electrophoretic analysis is necessary prior to single-molecule
experimentation, both to avoid artifactual structures and to ascertain the
quality of array formation.
Preparation of Experimental Samples

For single-molecule studies, samples are prepared by sequential infu-
sion of solutions into sample chambers for microscopic observation. Cham-
ber volume is approximately 10 l. Fluid flow through the chamber is by
capillary action produced by placing an absorbent wick at one end of the
chamber while delivering solutions by micropipette at the other end. All
procedures are carried out at room temperature, and incubations are per-
formed in a humid chamber to avoid evaporation. Samples prepared in this
way are exquisitely sensitive to changes in physical and chemical condi-
tions, so that consistency in all aspects of sample preparation, especially
temperature, is critical. Once prepared, chromatin samples have a useful
experimental lifetime of about 2 h at room temperature.
Buffer solutions:
1. PBS: 140 mM NaCl, 2.7 mM KCl, 10 mM Na2HPO4, 1.8 mM
NaH2PO4, pH 7.3.
2. Blocking buffer (BB): PBS þ 0.2% purified non-fat milk protein.
3. Chromatin dilution buffer (CDB): 10 mM Tris-HCl (pH 8.0), 1 mM
Na2EDTA, 150 mM NaCl.
15
K.-M. Lee and G. Narlikar, in ‘‘Current Protocols in Molecular Biology’’ (F. A. Ausubel
et al., eds.), 3, p. 21.6.3. Wiley, New York, 2001.
16
M. Bulger and J. T. Kadonaga, Methods Mol. Genet. 5, 241 (1994).
17
M. E. Levenstein and J. T. Kadonaga, J. Biol. Chem. 277, 8749 (2002).
4. Experimental sample buffer (ESB): 10 mM Tris-HCl (pH 8.0),

1 mM Na2EDTA, 100 mM NaCl, 1.5 mM MgCl2, 0.02% (v/v)
Tween-20, and 0.01% (w/v) milk protein.
The blocking agent that has worked best in our hands is a purified milk
protein powder (Biorad Laboratories, Hercules, CA). For attachment of
digoxigenin-labeled DNA samples to sample chamber surfaces we have
used polyclonal sheep anti-digoxigenin (Roche Applied Science, Indian-
apolis, IN) with relatively uniform results. Chromatin assembled by the salt
dialysis method referred to earlier are stored at concentrations 100 ng/l

at 4 , and diluted to 0.2–1 ng/l in CDB immediately before application to
sample chamber for experimentation.
1. Rinse sample chamber with 5 chamber volumes of PBS.
2. Immediately infuse 1 volume of anti-digoxigenin solution (25 ng/l
in PBS). Incubate 10 min.
3. Rinse with 5 volumes of BB. Incubate with residual blocker, 5 min.
4. Rinse with 5 volumes of PBS.
5. Immediately infuse 1 volume of diluted chromatin sample in CDB.
Incubate 7.5 min.
6. Rinse with 5 volumes of CDB.
7. Infuse 1 volume of avidin-coated beads (10 pM in CDB þ 0.01%
(w/v) purified non-fat milk protein). Incubate 5 min.
8. Rinse with 5 volumes of ESB. Seal chamber with nail polish if no
additional solutions will be infused.
Instrumentation
Mechanical measurements of a single nucleosomal array can be
obtained by using a single-beam optical trapping microscope.5,8 Here, we
provide a brief overview of the instrument. The reader should refer to
Wang et al.9 and Koch et al.5 for more detailed descriptions of the design,
construction, and calibration of the optical trapping setup.
Optical Trapping System

A schematic of our optical setup is shown in Fig. 1A. The trapping laser
has a wavelength of 1064 nm (Spectra-Physics Lasers, Mountain View,
CA). The laser beam passes through a single-mode optical fiber (Oz
Optics, Carp, ON) and an acousto-optic deflector (NEOS Technologies,
Melbourne, FL), and is focused onto the sample plane by a 100, 1.4-
NA, oil immersion objective on an Eclipse TE200 DIC microscope (Nikon
USA, Melville, NY). The focus of the laser serves as the center of the trap
for a micron-sized microsphere. The laser light is collected by a 1.4-NA oil
immersion condenser and projected onto a quadrant photodiode (Hama-

matsu, Bridgewater, NJ). The photocurrents from each quadrant of the
photodiode are amplified and converted to voltage signals using a position
detection amplifier (On-Trak Photonics, Lake Forest, CA). The position of
the optical trap relative to the sample can be adjusted with a servo-con-
trolled 1-D piezoelectric stage (Physik Instrumente GmbH & Co., Wald-
bronn, Germany). Analog voltage signals from the position detector and
stage position sensor are anti-alias filtered at 5 kHz (Krohn-Hite, Avon,
MA) and digitized at 7–13 kHz for each channel using a multiplexed
analog to digital conversion PCI board (National Instruments Corporation,
Austin, TX).
Calibration of the Optical Trapping System

The instrument calibration methods may follow those of Wang et al.9 In
brief, the first step of the calibration determines the position of the trap
center relative to the beam waist and the height of the trap center relative
to the coverslip. The second step of the calibration determines the position
detector sensitivity and trap stiffness. The third step of the calibration lo-
cates the anchor position of the DNA tether on the coverslip, and is per-
formed prior to each measurement by stretching the DNA at low load
(<5 pN). These calibrations are subsequently used to convert data into
force and extension.
Experimental Control and Data Acquisition

To disrupt nucleosomes as shown in Fig. 1B, the coverslip is moved
relative to the trapped microsphere with a piezoelectric stage to stretch
the DNA. Once a surface-tethered microsphere is optically trapped, the
coverslip is then moved with a piezoelectric stage to stretch the nucleoso-
mal DNA with either a velocity clamp or a force clamp.5,8 Both of these
clamps may be implemented with digital feedback algorithms, with an
average rate for a complete feedback cycle of 7–13 kHz. In the velocity
clamp mode, the coverslip is moved at a constant velocity relative to
the microsphere, whose position is kept constant by modulating the light
intensity (trap stiffness) of the trapping laser. A disruption event, during
which DNA is released from a histone octamer, is observed as a sudden re-
duction in the tension of the DNA. In the force-clamp mode, the position
of the coverslip is modulated so that the trapping force on the microsphere
is held constant by keeping its position fixed in a trap of constant stiffness.
In this mode, a disruption event is observed as a step in the coverslip
position.
Determination of the DNA Elasticity

Determination of the force-extension relation of a naked DNA is essen-
tial for the conversion of force and extension to number of base pairs of
naked DNA (see section on Data Analysis). Marko and Siggia proposed
the Worm-Like-Chain (WLC) model, which accounts for the entropic elas-
ticity and well describes the force-extension relation in the low-force region
(<5 pN).18 Wang et al. extended this model to also include the enthalpic
elasticity in the high-force region (>5 pN),9 and referred to this modified
form as the Marko-Siggia (MMS) model. This MMS model incorporates
both enthalpic and entropic contributions to stiffness and fits the experi-
mental results extremely well for forces up to 50 pN. In the MMS model,
the elastic stiffness of DNA is parameterized by its contour length under
zero tension, L0, its persistence length, LP, and its elastic modulus, K0.
The force (F) and extension () are simply related:
" #
kB T 1 1 F
F¼ þ :
Lp 4ð1 =L0 þ F=K0 Þ2 4 L0 K0
The elastic parameters of dsDNA can be obtained following Wang

et al.9 Under our experimental conditions, L0 per base is 0.338 nm, LP is
43.1 nm, and K0 is 1205 pN. Therefore, if both the force and extension
are known, this relation can be used to obtain L0, which can be readily
converted to the number of base pairs.
Data Analysis
Nucleosomes can be disrupted in various ways. The two ways presented
here are velocity-clamp stretching and force-clamp stretching. These
two methods are roughly equivalent, but with some subtle differences. Vel-
ocity clamp allows disruption of all nucleosomes at a specified stretching
velocity regardless of the strength of protein-DNA interactions within
nucleosome. However, nucleosomes in an array are disrupted under
slightly different force conditions, which depend on the number of nucleo-
somes remaining in the array at a specific disruption. Force clamp allows
disruption of all nucleosomes under identical force conditions (i.e., the
same force). However, more experimentation is required to determine a
workable range of force: If the force is too small or too large, the time to
disrupt the nucleosomes will be too long or too short to be experimentally
accessible.
18
J. F. Marko and E. D. Siggia, Macromolecules 28, 8759 (1995).
Fig. 2. Stretching a nucleosomal array with a velocity clamp at 28 nm/s. Nucleosomal

arrays were prepared with avian core histones and a 3684-bp DNA fragment containing 17
direct tandem repeats of the sea urchin 5S positioning element. The biotin and digoxigenin
linkers at the two ends of the DNA effectively contribute 50 bp of DNA. (A) Force-
extension curve of a fully saturated nucleosomal array. At higher force (>15 pN), a sawtooth
pattern containing 17 disruption peaks was observed. Force-extension characteristics of a
Velocity Clamp
An example of data taken with a velocity clamp is shown in Fig. 2. At
low force (<15 pN), the force-extension curve starts to deviate from that of
the corresponding naked DNA; at higher force (>15 pN), a distinctive saw-
tooth pattern starts to appear; at even higher force (>40 pN), the force-
extension resembles that of a naked DNA (dotted curve). Previously,
Brower-Toland et al.8 demonstrated that the high-force sawtooth pattern
is indicative of a nucleosomal array, with each peak corresponding to a
single nucleosome. Under the conditions used in our experiments, the
spacing between adjacent peaks is 27 nm. The observed sawtooth pattern
suggests separate disruption of strong DNA-histone interactions in
individual nucleosomes.
To determine the amount of DNA released from a nucleosome, the
MMS model can be applied. This conversion attributes extension only to
naked DNA, that is, linker DNA and DNA peeled from nucleosome core
particles (NCP). This method of conversion from force-extension curve to
number of base pairs of naked DNA is similar to that previously used for
single-molecule studies of transcription.3 The MMS model is only an
approximation for a nucleosomal array. To achieve better precision for
the conversion, a more refined model will be necessary. Conversion of
the data in Fig. 2A is shown in Fig. 2B, where the amount of naked DNA
is plotted as a function of time during stretching. At the beginning of
stretching (0–2 s), the average amount of naked DNA is constant, indicat-
ing no DNA release from NCPs. This should correspond to the amount
of linker DNA for a relaxed nucleosomal array. As force rises in the
low-force region, DNA release is gradual, indicating a simultaneous release
of DNA from all nucleosomes, with 76 bp of DNA release per nucleo-
some. At high force, the sawtooth peaks in Fig. 2A, are converted to steps.
DNA release is sudden, indicating a separate release of DNA from each
nucleosome of 80 bp.
Force Clamp
An example of data taken with a force clamp at 20.2 pN is shown in
Fig. 3. Unlike the velocity clamp measurements, all the nucleosomes
experienced the same force before disruption. Here, sudden disruptions
of nucleosomes resulted in stepwise increases in the DNA extension, with
full-length naked DNA (dotted line) are shown for comparison. (B) Amount of naked DNA
as a function of time derived from data shown in Fig. 2A. The top dotted line is a comparison
with a full-length naked DNA. At higher force, the curves show 17 steps, which correspond to
the 17 disruption peaks in Fig. 2A.
Fig. 3. Stretching a nucleosomal array with a force clamp. The graphs are plots of DNA
extension (left axis) and the amount of naked DNA (right axis) versus time under constant
force of 20.2 pN.
each step corresponding to one nucleosome disruption. The extension (left

axis) is readily converted to number of base pairs of naked DNA
(right axis). These steps ( 80 bp) provide a measure of DNA release per
nucleosome at high force.
Conclusion
Renewed interest in chromatin as a mediator of the structure, mainten-
ance, and regulation of eukaryotic genomes has inspired the development
of a variety of novel chromatin techniques. Optical trapping technology
provides a useful addition to this repertoire of techniques. We anticipate
that single-molecule optical trapping experiments on chromatin structure
will complement more traditional technologies, and aid in the elucidation
of the structural role in chromatin of histone and nonhistone proteins
and their post-translational modifications. Likewise, optical trapping
methods will be adaptable to the study of enzymatic activities such as
RNA polymerases and ATP-dependent chromatin remodelers operating
on chromatin structure.
Acknowledgments
We thank David A. Wacker and Dr. Robert M. Fulbright for their helpful discussions and
technical assistance. Supported by grants from the NIH and the Keck Foundation’s
Distinguished Young Scholar Award to M.D.W.
[6] single-molecule analysis of chromatin 73
[6] Single-Molecule Analysis of Chromatin

By Sanford H. Leuba, Martin L. Bennink, and
Jordanka Zlatanova
The importance of chromatin structure and dynamics in the functioning

of eukaryotic DNA is currently widely accepted. Still, understanding the
structure and how it changes in response to various signals regulating
DNA transactions (transcription, replication, recombination, and repair)
remains a challenging task. Recently, single-molecule techniques have
been added to the arsenal of more traditional biochemical and biophysical
approaches,1 allowing the investigation of one molecule (or molecular
complex) at a time, frequently in real time. This unique capability is very
valuable, since it allows the elucidation of the behavior of individual mol-
ecules and an assessment of their variability. Thus, these techniques repre-
sent a significant improvement over population-average methods, in which
individual differences are masked in the average parameters measured.
Moreover, single-molecule techniques avoid the technically challenging
task of synchronizing a population of molecules at a specific stage of a pro-
cess. Even when such synchronization can be achieved, it is quickly lost
soon after the synchronization block is relieved because of the stochastic
nature of the individual reaction steps. Another major advantage of
the single-molecule approaches is the high spatial and temporal resolution
of the measurements, allowing observation of changes in the nanometer
distance range and millisecond timescale.
Here, we will describe in technical detail the three single-molecule tech-
niques that have been utilized so far to study single chromatin fibers: the
atomic force microscope (AFM), optical tweezers (OT), and magnetic
tweezers (MT). The AFM produces digital topographical images of samples
deposited on flat surfaces by raster-scanning the surface with a sharp tip
mounted on the back of a flexible cantilever2 (Fig. 1A). Atoms on the tip
interact with atoms on the surface, causing the cantilever to deflect upwards
or downwards, depending on whether the tip-sample interaction is repul-
sive or attractive. The deflections of the cantilever are registered by a laser
beam reflected off the back of the cantilever to produce a topographical
image of the sample. The AFM can also be used as a force spectroscopy
1
S. H. Leuba and J. Zlatanova, eds., ‘‘Biology at the Single-Molecule Level,’’ Pergamon,
Amsterdam, 2001.
2
G. Binnig, C. F. Quate, and C. Rohrer, Phys. Rev. Lett. 56, 930 (1986).

Fig. 1. Schematic of operation principles of three single-molecule instruments successfully

used for chromatin research. (A) Atomic force microscope (AFM). (B) Optical tweezers
(OT). (C) Magnetic tweezers (MT).
tool: in this application, the AFM tip is allowed to move only in the
z-direction, up and down. If a molecule that is immobilized on the surface
is also attached to the tip (either through specific linkages or through
nonspecific adhesion), then the upward movement of the cantilever will
stretch the molecule to produce the so-called force-extension curves.3
Using OT is a technique for manipulating single molecules based on the
interaction of light with matter.4 Light can exert forces on small beads of
certain optical properties in such a way that the bead is kept suspended
at a point close to the waist of a laser beam that is focused with an objec-
tive. If such a ‘‘trapped’’ bead is pulled out of this equilibrium position
by some external force, a net restoring force resulting from the bead’s
interaction with light will effectively pull it back to this restoring position
(Fig. 1B; see ‘‘Physics of Optical Trapping’’ and Fig. 2A–C). Thus, the
trap acts as a Hookean spring, with the force on the trapped bead being
proportional to its displacement from the equilibrium position in the trap
(Fig. 2D). In the OT set-up, one can suspend a macromolecule between
an optically trapped bead and another surface (a coverslip or another
bead), which can be moved at will to apply tension to the molecular tether,
and force-extension curves can be recorded.
Finally, in the MT, the macromolecule is attached between a surface
and a magnetic bead (Fig. 1C). Manipulation of an external magnetic field
can be used to apply stretching force to the tethered molecule, and/or to
induce precisely known levels of supercoiling: stretching is achieved by
changing the distance between the external magnet(s) and the cuvette,
while supercoiling is introduced by rotating the external magnetic field
3
J. Zlatanova, S. M. Lindsay, and S. H. Leuba, Prog. Biophys. Mol. Biol. 74, 37 (2000).
4
A. Ashkin, J. M. Dziedzic, J. E. Bjorkholm, and S. Chu, Optics Lett. 11, 288 (1986).
Fig. 2. Principle of optical trapping in a single-beam gradient trap. Schematic showing the
propagation of two high-angle rays as they refract at the surface of a high-index spherical
particle, when positioned just below the focal point (A), just above the focal point (B), or just
off-axis within the focal plane (C). The net force resulting from these two rays is always
pointing toward the focal point, establishing a stable trapping point at the focus. Reflection at
the surface of the bead is neglected (if reflection is taken into account, the center of the trap
will be just below the focal point). (D) The force exerted on the trapped bead increases
linearly with the displacement of the bead from the trap center. Conceptually, this is identical
to a Hookean spring with zero rest length. (E) When the trapped bead is attached to a single
DNA molecule, the force exerted by the trap on the bead is equal to the tension within the
molecule. The deflection of the laser beam is used to accurately measure the displacement of
the bead and, therefore, the tension in the molecule. (F) Optical tweezers optical set-up. The
beam of a 2-W continuous-wave infrared laser (1064 nm) is expanded using lens L1
( f ¼ 50 mm) and L2 ( f ¼ 350 mm) (NOTE 9). The diameter of the beam slightly overfills the
back aperture of the water-immersion objective (Obj), which is a necessary condition to
establish a stable optical trap within the flow cell. The intensity of the laser can be tuned by
turning the quarter-wavelength plate (WP) in front of the first lens. A polarizing beam splitter
cube (PBS) just before the objective separates the two orthogonal polarization directions,
enabling fine control of the intensity of the beam at the objective (light that is reflected by the
polarizing beam splitter is stopped using a beam stop, BS). The transmitted (i.e., refracted)
laser light is collected with a 0.9-NA condenser lens (Con) and projected onto a quadrant
detector. Note that the quadrant detector is not positioned in a conjugate image plane
(NOTE 1). This detector quantifies the deflection, which is calibrated and transformed into a
force signal. A neutral density filter (F1) is used in front of the quadrant detector to reduce
the intensity to measurable levels. Within this experimental set-up, the objective is also used
for viewing the beads in the flow cell. The light of a fiber light source is collimated using a 40-
mm lens (L3) and projected onto the back of the condenser, which results in a homogeneous
illumination of the sample. The objective in combination with an additional lens L4
( f ¼ 170 mm) is used to project the microscopic image onto a CCD array. An additional filter
(F2) is used in front of the camera to fully block the infrared laser light. Dichroic mirrors (D1
and D2) are used to separate the laser-beam path from that of the illumination beam.
either clockwise or counterclockwise.5–7 The magnitude of the applied

force can be determined by the equipartition theorem: F ¼ lkB T/<x2>,
where l is the contour length of the tether, kB, Boltzmann’s constant, T, ab-
solute temperature, and <x2>, the Brownian motion of the bead in the
direction perpendicular to the tether axis.
AFM imaging
Chromatin fibers are challenging to image by AFM because they are
relatively soft (compared to naked DNA, for example), have features that
are high above the surface, and contain histones that are notoriously sticky.
Softness results in indentation of the sample by the tip, while the sticky his-
tones may easily attach to it: both interfere with proper imaging. Distortion
to the image also occurs when the vertical distance traveled by the tip is too
large—which happens with samples containing topographically high fea-
tures. We have recently published detailed protocols of how to image chro-
matin fibers with the AFM.8 Here, we provide some further thoughts based
on our prior experience with the Digital Instruments (DI, www.di.com)
AFM (e.g., Zlatanova et al. (1994),9 Leuba et al. (1994),10 Yang et al.,11
Zlatanova et al. (1998)12 Leuba, Bustamante, Zlatanova, and van Holde
(1998),13 Leuba, Bustamante, van Holde, and Zlatanova (1998)14), and
our more recent work with the Molecular Imaging (MI, www.molec.com)
instrument (e.g., Karymov et al.,15 Tomschik et al.,16 An et al.17).
5
T. R. Strick, J. F. Allemand, D. Bensimon, A. Bensimon, and V. Croquette, Science 271,
1835 (1996).
6
T. R. Strick, J.-F. Allemand, V. Croquette, and D. Bensimon, J. Stat. Phys. 93, 647 (1998).
7
T. Strick, J. Allemand, V. Croquette, and D. Bensimon, Prog. Biophys. Mol. Biol. 74,
115 (2000).
8
S. H. Leuba and C. Bustamante, in ‘‘Methods in Molecular Biology: Chromatin Protocols’’
(P. B. Becker, ed.), p. 143. Humana Press, Totowa, NJ, 1999.
9
J. Zlatanova, S. H. Leuba, G. Yang, C. Bustamante, and K. van Holde, Proc. Natl. Acad.
Sci. USA 91, 5277 (1994).
10
S. H. Leuba, G. Yang, C. Robert, B. Samori, K. van Holde, J. Zlatanova, and
C. Bustamante, Proc. Natl. Acad. Sci. USA 91, 11621 (1994).
11
G. Yang, S. H. Leuba, C. Bustamante, J. Zlatanova, and K. van Holde, Nat. Struct. Biol. 1,
761 (1994).
12
J. Zlatanova, S. H. Leuba, and K. van Holde, Biophys. J. 74, 2554 (1998).
13
S. H. Leuba, C. Bustamante, J. Zlatanova, and K. van Holde, Biophys. J. 74, 2823 (1998).
14
S. H. Leuba, C. Bustamante, K. van Holde, and J. Zlatanova, Biophys. J. 74, 2830 (1998).
15
M. A. Karymov, M. Tomschik, S. H. Leuba, P. Caiafa, and J. Zlatanova, FASEB J. 15,
2631 (2001).
16
M. Tomschik, M. A. Karymov, J. Zlatanova, and S. H. Leuba, Struct. Fold. Des. 9, 1201 (2001).
17
W. An, V. B. Palhan, M. A. Karymov, S. H. Leuba, and R. G. Roeder, Mol. Cell 9,
811 (2002).
There are two different modes of imaging with the AFM, contact mode
and dynamic mode. In the contact mode, the AFM tip is in constant contact
with the sample/surface during the raster-scanning. In the dynamic (so-
called tapping) mode, the cantilever is oscillated at a given frequency
above the surface and the effect of the sample is to dampen the amplitude
of this oscillation. In tapping mode, the tip contacts the sample/surface only
intermittently, thus minimizing the friction/drag forces exerted by the tip in
contact mode. The different instruments achieve a controlled oscillation of
the cantilever in different ways (see later).
Various commercial AFM designs have advantages and disadvantages
in the ease and quality of imaging over an area of 0.5 m 0.5 m. The
most useful image size for chromatin fibers is 1 m 1 m, 512 pixels
512 pixels, as it gives sufficient nucleosome resolution to perform measure-
ments of center-to-center internucleosomal distances, angles between lines
connecting three contiguous nucleosomes, z-heights, and numbers of
nucleosomes per 10-nm fiber length.
Digital Instruments has two designs of the instrument: the Bioscope and
the Multimode. In the Bioscope, the piezo and the cantilever holder are
both above the imaging surface. This scope can be mounted on inverted
fluorescence microscopes: one can easily find large samples (e.g., cells or
metaphase chromosomes) on the surface and then position the scanning
tip directly over the region of interest. The Bioscope comes with a piezo
scanner with a range of 100 m 100 m. Since biological processes occur
in structures in the nanometer range, scan sizes of a couple of micrometers
are needed; such needs are better met by the smaller scanners (typically
16 m 16 m, or 12 m 12 m) of the Multimode instrument. Here,
the piezo scanner is below the imaged surface, creating a risk for buffer
leakage over (and destruction of) the piezo. The DI instrument uses
separate piezos to drive the cantilever oscillation in tapping mode; this
adversely affects the instrumental noise.
The MI AFM uses magnetically driven cantilevers to image biological
samples in air and in buffers. An external solenoid drives an alternating
current (AC) signal that vibrates a cantilever coated with a magnetically
susceptible material (the cantilever responds to the changes in the AC
magnetic field). This method of oscillation is much gentler and reduces
the noise in the instrument (for more technical details, see refs. Zlatanova
et al. [2000],3 Lindsay [2001]18). The MI liquid cell is of a simpler design
than that of the DI instrument, and the piezo is placed above the imaged
18
S. M. Lindsay, in ‘‘Scanning Probe Microscopy and Spectroscopy: Theory, Techniques, and
Applications’’ (D. Bonnell, ed.), p. 289. Wiley, New York, 2001.
surface. Despite these obvious advantages of the MI instrument, it takes

longer to learn how to successfully use it.
A more recent AFM company, Asylum Research (AR, www.asylumre-
search.com), has come up with an x-, y-, z-dimension AFM to complement
their single z-dimension puller. The puller is used in force spectroscopy
experiments3 to apply stretching forces to molecules suspended between
the tip and the surface. The piezo crystal in the AR 3D imaging AFM
is also placed above the imaging surface. We have no experience with this
instrument.
The DI software has gone through several rounds of maturity; however,
the company limits access to the original source code. The MI and AR soft-
ware are more accessible (e.g., the AR AFM is run on IgorPro, Wave-
metrics, Portland, OR, USA), but are less mature. The ease of use of the
software, as well as the ability to archive data, and then reaccess the data
in its original format is an important consideration for scientists planning
to use an AFM.
Materials
. Triethanolamine, TEA (Baker, Phillipsburg, NJ, USA)
. Electron microscope (EM)-grade glutaraldehyde (EM Sciences,
Fort Washington, PA, USA, Cat. No. 16120)
. Mica (Asheville-Schoonmaker Mica Company, Newport News,
VA, USA)
. Paper cutter board (local office supply) to cut mica
. Scotch magic tape (3M)
. Milli-Q (Millipore) water system (Fisher Scientific, Pittsburgh, PA,
USA); all buffers are prepared from 18.2 MV purified water
. Slick microfuge tubes (PGC Scientific, Gaithersburg, MD, USA)
. Tank of argon
. Butane professional torch (VWR, West Chester, PA, USA, Cat.
No. 62379-525)
Chromatin Preparation and Dialysis; Glutaraldehyde Fixation

Chromatin fibers are isolated from cell nuclei using micrococcal nucle-
ase-based procedures (e.g., refs. 10, 15). Extensive dialysis versus 10 mM
TEA-HCl, pH 7.0, 0.1 mM EDTA is crucial to obtaining good AFM
images of extended fibers. For chromatin fibers reconstituted on naked
DNA using crude extracts or purified histones and chromatin-assembly
proteins, it is essential to purify the chromatin fibers through a sucrose or
glycerol gradient before dialysis (e.g., ref. 17).
We have used glutaraldehyde to ‘‘fix’’ chromatin fibers before imaging,
that is, to preserve the specific morphology characteristic of specific
environmental conditions. Fixation also prevents gross structural distor-

tions that may be caused by surface interactions. A final concentration of
only 0.1% glutaraldehyde seems to work fine, in accordance with other ob-
servations (Thoma et al. (1979)19; personal discussion with Jan Bednar). To
lessen the loss of material to the tube walls, we use slick microfuge tubes.
One microliter of glutaraldehyde from a freshly opened ampoule is added
to 99 l of a 0.1-mg/ml solution of chromatin fibers (A260 ¼ 2.0) and the
tube is left overnight on ice; rotation of the tube tends to precipitate the
chromatin fibers onto the tube surface.
Preparation of Mica and Glass Surfaces, Sample Deposition, and

Imaging Conditions
Square sheets of mica (cut with an office supply paper cutter) are
mounted (typically with double-stick tape or superglue—the thick-gel
superglue is preferable to the runny variety) onto a surface. To make a
fresh, atomically flat surface (of roughness not exceeding a few angstroms),
mica is peeled using Scotch magic tape. The mica surface remaining on the
tape is inspected by eye to see whether it is smooth and shiny; a shiny
surface is an indication that the mica surface left behind on the block is
optically flat, and likely to be atomically flat when imaged by the AFM.
An alternative surface is cover glass. The glass is first rinsed with a
stream of 95% ethanol from a squirt bottle, the ethanol is allowed to evap-
orate, and the cover glass is flamed with a torch to burn away impurities,
improving its surface flatness. It is also possible to burn impurities by
baking the coverslips for 5 min in a 600 oven. After flaming or baking,
the cover glass is rinsed briefly with Milli-Q water and allowed to dry.
When the water-rinsing step is omitted, small particles (100 nm sized,
perhaps ash) are observed in the AFM images.
Typically, 2–10 l of sample are deposited on the surface. After 1–5
min, the sample is rinsed either with the same low-ionic strength buffer
or Milli-Q water. Excess liquid is blotted with blotting paper and blown
off for a few seconds with a light flow of argon or nitrogen gas.
Imaging conditions depend on the AFM make, the cantilever, the
surface used and on whether the imaging is performed in air or under
liquid. Typically, we have employed imaging rate of two lines per second;
capturing a 512 pixels 512 pixels image takes 8 min. For more details
see reference 8.
19
F. Thoma, T. Koller, and A. Klug, J. Cell Biol. 83, 403 (1979).
Optical Tweezers
Physics of Optical Trapping

Optical trapping and manipulation of dielectric particles was first dem-
onstrated in 1986.4 Radiation, when changed in direction or intensity as a
result of interaction with matter (i.e., reflection and refraction at an inter-
face), exerts force on the matter. Photons have translational momentum
p ¼ mv ¼ h/c, where m is mass, v, velocity, h, the energy of the photon,
and c is the speed of light. When the momentum of a photon is changed
as a result of reflection and/or refraction, the matter the photon interacts
with also undergoes a change in momentum (the momentum conservation
law). For a continuous flow of photons, this results in a physical force
exerted on the bead. The magnitude of force that radiation (with normally
used light intensities) can exert on objects is rather low. However, for small
particles, whose dimensions are on the order of the wavelength of the radi-
ation, these forces are significant. Figure 2A–C illustrates the radiation
forces exerted on a bead by the rays.
The optical trap is working as a Hookean spring. The force exerted on
the bead is proportional to its displacement in the trap (Fig. 2D). The pro-
portionality factor relating the displacement with the force is the trap stiff-
ness or trap constant k (typically 0.1 pN/nm). When a DNA molecule is
suspended between two beads (Fig. 2E), the force applied to the molecule
through the handle is equal in magnitude and acts in opposite direction
to the force experienced by the optically trapped bead. The force can be
determined by measuring the displacement of the bead from the center
of the trap using video analysis. However, the force can be obtained more
accurately and with higher time resolution by measuring the deflection of
the beam as a result of the displacement of the trapped bead (Fig. 2E).
One can consider the trapped bead as a micron-sized lens, the movement
of which over nanometer distances can be detected very accurately by
measuring the deflection of the laser beam. This piconewton (pN) reso-
lution in force and nanometer resolution in displacement makes the optical
trap an excellent tool for measuring force-extension curves of individual
macromolecules.
Materials
. Quadrant detector (SPOT-9DMI, UDT Sensors, Hawthorne, CA,
USA, www.udt.com)
. Newport 3D translation stage with both manual and piezo-electric
control (ULTRAlign Precision Fiber Optics Linear Stage, 561D,
Newport, Irvine, CA, USA, www.newport.com)
. 2-W infrared laser (1064 nm) (CrystaLaser, Reno, NV, USA,

www.crystalaser.com)
. Millennia 10-W CW laser (Spectra Physics, Mountain View, CA,
USA, www.spectra-physics.com)
. High NA water-immersion objective (100, 1.2 NA, Leica, Wetzlar,
Germany, www.leica-microsystems.com)
. Horizontal vibration-isolated table (vibration-isolated Workstation
07 OWS 112, Melles Griot, Irvine, CA, USA, www.mellesgriot.com)
. Glass capillaries (Borosilicate glass capillaries, GC120TF-15,
1.2 mm OD, 0.94 mm ID, thin wall, with filament, Clark
Electromedical Instruments, Reading, UK)
. Square glass capillary (ID ¼ 50 m, OD ¼ 150 m, VitroCom,
Mountain Lakes, NJ, USA, www.vitrocom.com)
. Professional pipette puller (Flaming/Brown Micropipette Puller,
model P-87, Sutter Instrument Company, Novato, CA, USA,
www.sutter.com)
. Thin fluid line (Intramedic polyethylene tubing, non-sterile, PE 10,
ID ¼ 0.28 mm, OD ¼ 0.61 mm, Becton Dickinson, Franklin Lakes,
NJ, USA)
. 3 mm outer diameter flexible tubing (various suppliers)
. Six-way low-pressure selection valve (Upchurch Scientific, Oak
Harbor, WA, USA, www.upchurch.com)
. Uniform Microspheres, CML, 2.60 m, Bangs Laboratories,
Fishers, IN, USA, www.bangslabs.com; these beads have carboxyl-
ate end-groups at the surface that are activated using EDAC
(1-ethyl-2-(3-dimethylaminopropyl)carbodiimide, Sigma E7750,
Sigma-Aldrich Chemie BV, Zwijndrecht, The Netherlands)
. Chemicals: -DNA (New England Biolabs, Beverly, MA, USA);
streptavidin (Roche, Almere, The Netherlands); dNTPs (100 mM
dGTP, dCTP, Sigma); bio-dATP (0.4 mM in 50 mM Tris-HCl, pH
7.5, Sigma); bio-dUTP (0.3 mM in 50 mM Tris-HCl, pH 7.5, Sigma);
Klenow fragment in 50% glycerol, 100 mM KH2PO4, pH 7.0,
10 mM -mercaptoethanol (Sigma)
Optics
The instrument is based on an optical microscope equipped with a CCD
camera (Fig. 2F; see also Bennink [2000]20). For the optical trap, a 2-W in-
frared laser is introduced. The 3-mm diameter of the emerging laser beam
20
M. L. Bennink, Ph.D. Thesis, Department of Applied Physics, University of Twente,
Enschede, 2000.
with a Gaussian-shaped intensity profile is first expanded using two lenses

in order to get a beam with a diameter of just over 2 cm. This beam is pro-
jected on the back aperture of a high NA water-immersion objective. The
high numerical aperture, of the objective, in combination with the slight
overfilling of the back aperture, ensures that there is enough intensity in

high-angle peripheral rays (having an angle of 70 with the optical axis)
that mainly contribute to the creation of the gradient force within the op-
tical trap.21 In many optical traps, oil-immersion objectives are being used
because the additional refraction of light when it enters the flow cell causes
relatively more intensity in peripheral rays, which will result in a stiffer
trap. However, there is a major drawback in using oil objectives. The focal
point is very close to the glass surface, which limits the depth at which
one can trap. Moreover, since the shape of the focus depends strongly on
the distance of the focal point to the glass, the trap stiffness will depend
on this distance, making calibration very complicated. We use a water-
immersion objective which makes it possible to trap beads up to 100 m
from the cover glass surface. Moreover, since the refractive index of the
immersion fluid is equal to that within the flow cell, the trap stiffness is
totally independent of the distance to the cover slide.
Using the same objective that is infinity-corrected for both imaging and
trapping has an advantage over using two objectives on either side of the
flow cell, one to create the optical trap, and the second one to make a
microscopic image. The plane in which the probe beads are captured is at
a fixed position with respect to the objective lens. Furthermore, since the
objective is infinity-corrected and a second lens is needed in front of the
CCD array to create an image, the object plane is also fixed with respect
to the objective. Now you can position the second lens, L2 ( f ¼ 170 mm),
such that the trapped bead is in focus, ensuring that the bead is always
trapped at the same position and in focus on the CCD camera.
Measurement of forces acting on the trapped bead is achieved by
detecting the deflection of the transmitted light.22,23 Small displacements
of the bead from the fixed position of the trap result in deflections of the laser
beam from the optical axis. This deflected light is collected by a condenser
lens and projected onto a quadrant detector. In a dual-beam set-up,24 all
transmitted and refracted light is captured, allowing direct derivation of
force from the measured change in momentum. In our instrument, how-
ever, the condenser lens is only collecting a small cone of the refracted
21
A. Ashkin, Biophys. J. 61, 569 (1992).
22
L. P. Ghislain and W. W. Webb, Opt. Lett. 18, 1678 (1993).
23
L. P. Ghislain, N. A. Switz, and W. W. Webb, Rev. Sci. Instrum. 65, 2762 (1994).
24
S. B. Smith, Y. Cui, and C. Bustamante, Methods Enzymol. 361, 134 (2003).
and transmitted light (NA ¼ 0.9), and this signal therefore needs experi-
mental calibration of the force.
The position of the quadrant detector has been widely debated within
the optical tweezers community. In our set-up, the detector is positioned
just behind the back focal plane (NOTE 1). The exact position of the
detector behind the back focal plane is not critical, but the beam has to
be within the linear regime of the detector. To determine the linear regime,
you simply put the entire detector onto an x–y translation stage and deter-
mine the signal as a function of the relative position of the beam on the
detector.
The entire set-up is constructed on top of a horizontal vibration-
isolated table to reduce external vibrations (within the building). The beam
pointing stability of the laser is crucial to the measurements (NOTE 2).
Furthermore, the noise due to airflow and dust can be reduced by shielding
the beam and the optical components with a Plexiglas (perspex) box. The
box encloses the beam emerging from the laser all the way to the detector,
leaving an opening at the flow cell to allow easy access. The side branches
for the white light illumination and the CCD array are not critical and
therefore not shielded.
Flow Cell Construction

For the chromatin assembly and stretching experiments, we use a flow
cell that allows exchange of buffers during the course of a single experiment.
The cell is constructed manually using a standard microscope glass (26 mm
76 mm). To create an entry and an exit for the flow cell, two small holes
are drilled within the glass using a diamond drill in a benchtop drill press
(both available in hardware stores) at positions indicated in Fig. 3A.
The next step is to use parafilm for spacing the two glass surfaces. Fold a
piece of parafilm of at least 10 cm 10 cm into a rectangular shape of 5 cm
10 cm and press it with your fingers such that the two layers stick
together. Use a scalpel to cut a rectangle with the dimensions of the micro-
scope slide (you can put a slide on top of the parafilm and use it as a tem-
plate). Now cut a long rectangle of 4 mm 50 mm, at the position
indicated in Fig. 3B. Take out this part and put the double layer of parafilm
on top of a microscope glass with holes. The positions of the entrance and
exit holes are exactly at the ends of the rectangle that is removed. Place the
VitroCom square capillary on top of the parafilm; the length of the capil-
lary should slightly exceed the width of the parafilm on the top. Use a
scalpel to cut it to size (wear safety glasses). Finally, take a thin microscope
glass (26 mm 76 mm) and put this on top of the assembly to seal the
flow cell.
Fig. 3. The flow cell. The custom-made flow cell consists of a sandwich of a standard
microscope glass with two holes (A), two layers of parafilm with a rectangular channel cut out
(B), and a thin cover glass (170 m thick) with the same dimension as the standard
microscope glass. This entire sandwich is positioned onto a metal (i.e., iron) holder (C), and
two plastic brackets are used to clamp the unit onto the holder (D). Inlet and outlet tubing are
glued inside the holes drilled in the plastic brackets. The plastic tubing is pushed over the hole
in the microscope glass and is actually clamping the sandwich on the holder (the plastic
bracket is not touching the glass directly). This also ensures a leak-free connection to the flow
channel. Finally, a second plastic bar (D) is attached on the top of the iron holder, which will
serve as a support on which to glue the inserted micropipette. (E) Exploded view of the flow
cell illustrating how the different parts should be assembled together. M3 nuts and bolts are
used to clamp the brackets holding the flow cell in place and ensure a leak-free connection.
The position of the small capillary between the parafilm and the thin cover slide is also
indicated. The micropipette goes through the capillary into the flow cell, with its tip in the
center of the flow cell. When inserted, the pipette is glued to the horizontal bar, which serves
as a support.
The parafilm is not sticky enough to hold the two glasses together

so you need to heat the sandwich to 60 (NOTE 3) using a clean hot plate.
Put the sandwich flat on the hotplate with the thick glass down. Put a
separate microscope slide on top that allows you to push the entire sand-
wich gently together (if you do not use a second slide, you might break
the thin cover glass). Note the transparency of the parafilm between

the glass slides. It is grey and nontransparent at RT; at 60 it starts
to become transparent and sticky. Push with your finger at the point
where the capillary is integrated until the glass is perfectly flat, that is, no
distortion in the reflection. Take the sandwich off the plate and let it cool;
the sandwich is sealed now.
To reuse the thick microscope glass with the holes, leave the sandwich
in ethanol for at least a few hours. If you want to use it again right away,
place it in ethanol for 5 min, and put the point of a scalpel between the
two glasses. Keeping the sandwich submerged in ethanol, gently push the
scalpel edge between the two pieces of glass. Give the ethanol time to flow
between the glass and the parafilm, and make sure that the thin glass slide
is not bent too much. The sandwich will fall apart in a few minutes. Use a
cotton-tipped applicator to clean the glass while submerged in ethanol.
For the holder you need:
. Iron plate (2 mm thick), cut as in Fig. 3C
. 4 bolts (M3 10) and nuts (M3)
. 2 bolts (M3 6) and nuts (M3)
. 2 Delrin plastic bars (3 mm thick) with three holes as indicated in
Fig. 3D
. 1 Delrin plastic bar (1.5 mm thick) with two holes as indicated in
Fig. 3D
. 3-mm outer diameter flexible tubing
. Thin fluid line (10–20 cm long)
. Glue (5 min-epoxy)
Place this glass sandwich in the iron holder. The U-shaped holder is cut
out using a laser, but standard metal drilling and cutting tools can be used
as well. Position the sandwich as indicated in Fig. 3E with the thin cover
glass facing the iron. The holes created for the entrance and exit tubing
should be exactly on top of the 7-mm holes within the iron holder. In be-
tween the glass and the iron, you can put one or two layers of parafilm
(not shown in Fig. 3E) to prevent the glass from breaking. Cut out small
holes at the position of the entrance and exit tubing holes in this spacer.
This is important for aligning the tubing later.
Cut the 3-mm exit tubing with a very sharp knife such that the exit
plane is exactly perpendicular to the tube axis; this will ensure a sealed con-
nection to the glass later. The tubing is glued through the center hole in the
bar. For the entrance tubing we use thin tubing for better control of the
flow. This thin tubing is glued within the 3-mm tubing, which is making
the connection to the glass (NOTE 4). This tubing is glued in the center
hole of the second bar.
Use bolts (M3 10) and nuts (M3) to attach the bars over the sandwich
(Fig. 3E). The sandwich is pushed against the holder by the tubing only,
which is enough to keep it in place and to ensure a sealed connection.
When looking from the backside through the 7-mm hole in the holder,
you can see if the tubing is correctly aligned and sealed. Connect the
second bar (Fig. 3D, top one) on the backside of the holder using two bolts
(M3 6). This will be the holder for the glass micropipette.
Pulling and Placing Micropipettes in the Flow Cell; Mounting the Flow
Cell into the Setup
For micropipettes, we use glass capillaries that we pull using a profes-
sional pipette puller that has four different parameters, which we optimized
to produce long, tapered micropipettes with a tip diameter of 1–2 m (the
diameter of the micropipette is still less than 50 m at 10 mm away from
the tip). Our optimized parameters are: heat, 323; pull, 100; velocity, 10;
time, 240, but they may vary from one puller to the next (even of the same
model), so use these numbers as a starting point and optimize them further.
Start by breaking the glass tube in the middle (wrap the tube in a tissue)
and clamp one piece in the instrument. After pulling, inspect the tip with
a microscope, and if it looks good, gently break off the long tube, again
holding it in a tissue. Then glue the micropipette to a 20- to 30-cm long
3-mm tubing.
View the holder with the glass sandwich, the tubing and the top bar with a
microscope. Use a metal support plate and small magnets to keep the flow
cell in place. The tubing with the micropipette is mounted on a 3D transla-
tion stage, able to move objects with 10-m accuracy over a range of a few
centimeters. You need to move the tip of the micropipette into the square
capillary that is connecting the outside world with the flow cell. Exert ex-
treme care, since if you touch the tip head-on, it will break. Once you have
the tip within the square capillary, you slowly move the pipette in until the
tip is in the middle of the flow cell. Glue the tip to the plastic bar on top using
two-component glue; the flow cell is ready to be mounted in the set-up.
The flow cell is positioned on a Newport 3D translation stage with both
manual and piezo-electric control. An iron plate with a 4-cm diameter
opening, through which the objective accesses the glass surface, is attached
vertically to the stage and the flow cell is attached to it using small cylin-
drically shaped magnets (5 mm in height, 12 mm in diameter, available in
different stores). This allows you to coarsely align the flow cell such that the
tip of the micropipette is in front of the objective (NOTE 5).
Flow System
A semi-automated flow system controls the exchange of buffers within
the flow cell and the flow rate (Fig. 4). The buffers are kept in containers
ranging from 1 ml (for the cell extract) to 50 ml (for the normal flow
buffer). All containers are screw-capped and have two holes drilled into
Fig. 4. The flow system used to introduce streptavidin-coated beads, biotinylated DNA,
and the cell-free extract at a controlled flow rate. Different containers ranging from 1 to 50 ml
are connected to a closed air-line, whose pressure is controlled using three computer-
controlled pressure valves. These are connected to a high-pressure line (2.0 atm), outside
pressure (1.0 atm) and a barrel, in which low pressure is maintained (0.5 atm). Switching
these valves shortly will change the pressure in the line, and therefore the flow rate in the flow
cell. Thin tubing is used to connect each container to a six-way selection valve, which specifies
which container is to be used. The beads are introduced in the system using a syringe.
the cap. One hole is 3 mm in diameter, and a piece of 3-mm tubing (air-
line) is inserted. A second hole, only 1 mm in diameter, provides access
for a thin fluid line. The connections through the caps are sealed with
5-min epoxy. All airlines are connected to a closed air system. Increasing
the pressure within the airlines will pressurize all the containers and
will maintain buffer flow. The thin tubing is connected to a six-way low-
pressure selection valve (NOTE 6), which enables the operator to manually
select the desired container.
For containers we use transparent disposable vessels:
. 2-ml tube, screw cap, Nalgene cryovials, PP sterile, Nalgene, a
subsidiary of Sybron, Rotherwas, Hereford, UK
. 12-ml tube, blue screw cap, Greiner bio-one, PS tube, sterile,
diameter, 16.8 mm, height, 100 mm, Greiner BV, Alphen a/d Rijn,
The Netherlands
. 50-ml tube, blue screw cap, Greiner bio-one CELLSTAR, PP tubes,

sterile, skirt, diameter 30 mm, height 115 mm, Greiner BV
. 1- and 5-ml syringe, Terumo Europe NV, Leuven, Belgium
. Needles, Micro-Lance 3 (30G1/2), 0.3 mm 13 mm, Becton
Dickinson, Drogheda, Ireland
Changing the pressure within the airlines controls the flow rate and the
direction of the flow. Three solenoid valves are connected to the airline;
the first valve is connected to a high-pressure container (2.0 atm), the
second valve is connected to a low-pressure container (0.5 atm), and the
third one is left open. Opening the valves for a short time will result in a
stepwise increase or decrease of the pressure within the system, thus con-
trolling the flow rate. Opening the third valve will restore the pressure in
the system to atmospheric level, stopping the flow.
The containers containing the beads, the DNA, the cell extract, and the
flow buffer used to flow the DNA molecules in are connected to this system
before an experiment starts. A wash buffer containing 1% Tween is used to
wash the flow cell after each experiment. The beads are introduced using a
syringe coupled through a needle to a fluid line connected to the selector
valve (Fig. 4). A stirring magnet (5 mm long) is put into the syringe to keep
the beads in suspension by moving an external magnet along the syringe.
Do this just before you enter new beads into the flow cell.
Preparation of Streptavidin-Coated Beads, End-Biotinylated DNA,

and Cell Extract
Streptavidin-coated beads are used as attachment surfaces for the ends
of the DNA molecule. On commercial available beads, the amount of
streptavidin coating the surface is too low; for this reason, we use the
following protocol to obtain streptavidin-coated beads.
1. Prepare: (i) 100 ml 50 mM phosphate buffer, pH 7.4, containing
0.9% NaCl (PBS) and (ii) 100 ml 50 mM MES, pH 6.0
2. Dissolve 1 mg streptavidin in 0.5 ml MES (filtered through 0.2 m
pore-size filters) and add this to 2 ml 2% CML beads
3. Shake this mixture for 15 min at RT
4. Add 4 mg EDAC and vortex
5. Adjust the pH to 6.5 using a 0.1-M NaOH (usually you need
50 l)
6. Incubate for 2 h or overnight at RT (on shaker)
7. Add glycine (to a final concentration of 100 mM) to block unreacted
sites
8. Centrifuge for 10 min at 10,000 rpm (Eppendorf centrifuge) and

resuspend the beads in 50 mM PBS solution. Do this three times
9. Resuspend in 2 ml PBS, 1% BSA, 0.013% NaN3. This suspension is kept
as stock. For an experiment, add 5 l of the stock to 1 ml of the main flow
buffer (10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 150 mM NaCl, 0.05%
BSA, and 0.01% NaN3). This final bead suspension has a concentration
of 105 beads per milliliter and can be loaded into the syringe.
We use -DNA, a dsDNA fragment, 48,502 bp with 12-nucleotide
ss overhangs on either end that are complementary to each other. The
ends of the DNA fragment are biotinylated by filling in with Klenow
polymerase.
Buffers and Solutions
. 50 mM Na2HPO4
. 1 M MgCl2 (autoclave both solutions and mix 1 ml Na2HPO4 with
10 l MgCl2 to get the reaction buffer: 50 mM Na2HPO4, 10 mM
MgCl2, pH 7.5)
. Stop solution (0.1 EDTA, pH 8.0)
. TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.5)
Protocol

1. Incubate 100 l (50 g) of -DNA in TE buffer for 5 min at 65 .
2. Add:
. 1 ml reaction buffer
. 0.5 l dNTPs (100 mM dGTP, dCTP)
. 1.25 l bio-dATP
. 1.67 l bio-dUTP
. 1.67 l Klenow fragment (10 units)

3. Incubate for 3 h at 37 .
4. Add 0.2 ml 0.1 M EDTA, pH 8.0.
5. Purify with Centricon concentrators (use a fixed-angle centrifuge,
centrifuge for 20min at 1000g [2200 rpm]; use a maximum of 2 ml
per concentrator).
6. Wash three times with 0.8 M TE buffer.
7. Put the concentrator upside down and centrifuge once at 2200 rpm
for 2 min.
8. Add 2 ml TE buffer.
The concentration of the labeled DNA in the stock solution is 25 g/ml.
The ratio A260/A280 should be 1.8–1.9. For each experiment, 10 l of
the DNA stock solution is added to 1 ml of TE buffer. Final DNA

concentration is 250 ng/ml, 4.5 109 molecules per milliliter.
The cell extract is isolated from Xenopus laevis eggs.25 Twelve
microliters of high-speed supernatant is diluted in 1 ml of assembly
buffer (50 mM HEPES-KOH, pH 7.6, 50 mM KCl, 1 mM EDTA, 2 mM
-mercaptoethanol).
Attaching a Single DNA Molecule Between Beads (Fig. 5)

The main flow buffer (10 mM Tris-HCl, pH 7.5, 1 mM EDTA, 150 mM
NaCl, 0.05% BSA, and 0.01% NaN3, used during the attachment of the
single DNA molecule to the two beads and during the stretching of
the chromatin fiber) is connected to the flow system last. Pressurize the
system to fill all fluid lines and remove any air in the system. Use inputs
of the selection valve that are next to each other. Turning the selector to
an input that is not connected is risky, since that might introduce air
bubbles. Flow the main buffer for a few minutes to ensure that all beads,
DNA, and proteins from the cell extract (from pre-flowing all the buffers)
are removed from the flow cell.
Introduce a very small amount of the bead suspension into the flow cell
by moving the piston in only 1–3 mm. (It is useful to make a small holder,
e.g., out of Delrin plastic, in which to secure the syringe.) Switch back to
the main buffer and put some pressure on the system (vflow ¼ 1–2 mm/s).
At this moment you must have the trapping laser beam on; the quadrant
detector will detect a relatively stable signal. As the beads enter the area
within the field of view of the CCD camera, the signal on the quadrant de-
tector displays sharp peaks, resulting from beads that move through the
laser beam. Stop the flow when these peaks are visible.
Using the piezo-controlled flow cell in combination with the flow
control, you can guide beads toward the position of the optical trap
(Fig. 5A). Try to trap beads that are above the focal plane, that is, closer
to the microscope objective. Beads below the focal plane will be pushed
further down when positioned in the laser beam and thus impossible to
capture. Position your objective as close to the glass as possible, increasing
the chance of capturing a bead. When the bead is captured, move the flow
cell such that you can transfer the bead onto the tip of the micropipette
(Fig. 5B). Position the pipette 0.5 m from the trapped bead and apply
under-pressure using the syringe connected to the micropipette (in some
cases capillary forces will make the bead jump onto the pipette). Repeat
the procedure for capturing a second bead in the optical trap (Fig. 5C);
25
G. H. Leno, Methods Cell Biol. 53, 497 (1998).
Fig. 5. Schematic presentation of how a single DNA molecule is attached between two
polystyrene beads. Arrows indicate the direction of the flow inside the flow cell. The four lines
forming a cross mark indicate the position of the optical trap. (A) The 2.6-m polystyrene
beads are introduced into the flow cell, and one of them is captured in the optical trap.
(B) The micropipette is moved toward the trapped bead to transfer the bead onto its
tip. Suction is applied to hold the bead in place. (C) A second bead is captured in the optical
trap. (D) Buffer containing DNA molecules replaces the buffer containing the beads. (E) A
single DNA molecule attaches to the trapped bead and stretches out in the flow direction.
(F) The bead on the micropipette is moved at a distance of about 16 m (i.e., length of the
DNA) downstream from the trapped bead to connect to the free end of the DNA molecule.
The presence of the invisible DNA molecule suspended between the two beads is attested to
by moving the pipette away from the trap position and observing the trapped bead being
pulled out of its position.
once a second bead is captured, increase the flow rate to flush the other
beads out of the flow cell. To speed up this procedure, you can position
the micropipette in front of the trapped bead. This will reduce the drag
force on the trapped bead, thus allowing the use of higher flow rates.
When all beads have disappeared, stop the flow completely by turning
the selector valve between two inputs. The signal from the quadrant de-
tector shows a noisy pattern, representing the Brownian motion of the
optically trapped bead. This signal should be centered around zero for both
x and y directions (for this alignment, it is convenient to have the quadrant
detector positioned onto a coarse translation table). Set the flow rate to
1 mm/s (at the position of the trap); a constant force of 30 pN will now
be detected, which is a direct result of the drag force on the trapped bead.
Switch to the container with DNA molecules, let the DNA solution flow
for 5-10 s, and then reverse to the main buffer. After 1 to 1.5 min, the
DNA is around the position of the trapped bead (Fig. 5D). When a DNA
molecule attaches to the trapped bead, there is a jump in the force because
the molecule causes a significant increase in drag force (10 pN). If no
jump is detected, repeat the procedure until you observe an attachment.
During attachment, a constant flow is maintained to prevent attachment of
the open DNA end to the trapped bead, resulting in a looped molecule.
Only when both ends are attached (see next step) to separate beads can
you stop the flow.
The last step is to attach the second bead (on the micropipette) to the
free end of the DNA molecule. Move the flow cell back and forth such that
the pipette is at a distance of 16 m downstream from the trapped bead
(Fig. 5E). If the molecule becomes tethered, moving the micropipette away
from the trap position will result in a force increase. Be careful not to
destroy the system at this point, since it is rather easy to pull the bead
out of the trap. You can move the micropipette to the other side and pull
the bead out of the trap in an upstream direction (Fig. 5F).
To ensure that just one individual DNA molecule is attached between
the two beads, the force-extension curve is recorded: the force is measured
while the pipette is moved away from the trapped bead. Video analysis
using a frame grabber and custom-written software applying a centroid
method (NOTE 7) is used to determine the distance between the two
beads, that is, the end-to-end distance of a DNA molecule (NOTE 8).
The typical force-extension curve for a single molecule of dsDNA26,27
exhibits a characteristic horizontal force plateau at 65 pN. Having mul-
tiple DNA molecules attached in a parallel fashion would shift the plateau
to higher forces: 130 pN for two molecules, 195 pN for three, etc. Other
features of the force curve, such as the length of the molecule or the force
development at short distances, are used to determine whether the DNA
molecule is free of bound protein molecules. If this curve deviates from
the expected curve, the molecule is discarded.
Assembly of Chromatin; Stretching and Relaxing of the Chromatin Fiber

To assemble chromatin fibers directly in the flow cell, we introduce the
Xenopus laevis egg extract, a crude cellular fraction containing cell debris
that is easily attracted onto the surface of the trapped bead (attempts to
26
P. Cluzel, A. Lebrun, C. Heller, R. Lavery, J. L. Viovy, D. Chatenay, and F. Caron, Science
271, 792 (1996).
27
S. B. Smith, Y. Cui, and C. Bustamante, Science 271, 795 (1996).
filter these out resulted in loss of activity of the cell extract). Therefore,
the trap must be turned off (i.e., laser is blocked), and the beads are
kept separate using a constant buffer flow. The tension within the DNA
can be estimated from the flow rate using Stoke’s law, but a correction
has to be made for the shielding effect of the micropipette and the bead
attached to it. An independent method to determine the tension within
the molecule is to determine the Brownian motion of the freely suspended
bead. Video analysis software is used to extract the positions of the
two beads at a 25-Hz time resolution (i.e., video rate). These data are used
to calculate the mean square displacement of the bead perpendicular to
the direction in which the DNA molecule is pulled. This value is directly
related to the force on the DNA molecule, with F ¼ lkBT/<x2>, where
l is the length of the biopolymer, kBT the thermal energy, and <x2>
the mean square displacement.
Thus, the assembly is performed under constant flow conditions and ob-
served as shortening of the apparent end-to-end distance as a result of the
formation of nucleosomal particles28 (240 nucleosomal particles along
the length of the -DNA molecule). Carefully controlling the flow
rate allows the observation of chromatin fiber assembly as a function of
the tension within the DNA.28
To stretch the reconstituted chromatin fiber, the cell extract is replaced
by the main flow buffer at a low flow rate so that the force does not exceed
5 pN. At this point, the trap is turned on, its intensity being slowly in-
creased from zero to ensure that the free bead is gently caught in the trap,
again avoiding large forces on the suspended chromatin fiber. Next, the mi-
cropipette is moved away at a constant velocity (typically 1 m/s) while the
force generated in the fiber is continuously monitored as its length in-
creases. When the distance between the two beads reaches 20–22 m
(the contour length of the ds -DNA molecule is 16 m, so the DNA at
this point is overstretched), the direction of the micropipette movement is
reversed to relax the molecule at the same speed.
The deflection signal representing the force exerted on the chromatin
fiber is measured using a LabView data acquisition card at a frequency of
1.2 kHz; such a frequency allows the observation of very small steps of tens
of nanometers. In the DNA stretching experiments, the end-to-end dis-
tance measurements are performed using real-time video analysis of the
precise positions of the two beads, thus limiting the sampling rate to a max-
imum of 25 Hz. For the chromatin stretching experiments, we needed sam-
pling at higher frequencies. Since the micropipette bead is moved away at a
28
M. L. Bennink, L. H. Pope, S. H. Leuba, B. G. de Grooth, and J. Greve, Single Mol. 2,
91 (2001).
Fig. 6. Analysis of the opening events. (A) Enlargement of a portion of a force-extension

curve. Portions of the curve, in which the force is continuously rising, then falling abruptly, can be
clearly distinguished, indicative of opening events. The step sizes involved in the opening events
can be analyzed using a pairwise distribution function. Determining the slope of the apparent
linear portions (i.e., stretch modulus) and assuming that this parameter does not change
significantly from one chromatin structural intermediate to the next, the contour length of the
fiber at each data point can be determined. Then, the contour length data can be analyzed by a
pairwise distribution function. This function determines all distances between any two points,
that is, for each point N, the distances to points N1, N2, etc., as well as to Nþ1, Nþ2, etc., are
determined; then a frequency distribution graph is plotted. The large arrow indicates a typical
point N, and the smaller arrows indicate points N1, N2, N3, etc. The horizontal lines are
the distances between point N and N1, N2, N3, etc. (B) The analysis on the data in this
force-distance curve results in a multi-peak histogram, in which peaks are found at 65 nm, and
multiples thereof, indicating that there is a discrete opening length during stretching; the
presence of the peaks at 130 and 195 nm reflect two and three simultaneous opening events,
respectively. The brackets above the graph in (A) indicate the major occurrences of distances
that account for the major peaks in (B), and connecting arrows help show this relationship.
constant speed, an interpolation algorithm can be successfully used to de-

termine the position of this bead at 1.2 kHz. The precise position of the
second (trapped) bead is calculated from the acquired deflection signal
using the trap stiffness. As depicted in Fig. 2D, the force acting on the
trapped bead and its displacement from the center of the trap are coupled
with the trap stiffness. The position of that bead, and thus the length of the
chromatin fiber, can therefore be obtained at a 1.2-kHz frequency.
The force-extension curve of a chromatin fiber looks quite different from
that of the naked DNA molecule, measured before the addition of the
cell extract (Fig. 6A; see also Fig. 3 of Bennink, Leuba et al. [2001]29).
29
Struct. Biol. 8, 606 (2001).
The tension within the chromatin fiber starts to develop already at 2 to

3 m extension, apparently revealing a wormlike chain behavior as
has been observed for the naked DNA molecule. When the tension
reaches 20 pN, a sudden drop in force is observed on further extension,
indicating a domain opening event within the chromatin structure. Multiple
force peaks are observed on further extension, indicating multiple opening
events. These structural rearrangements continue until the length of the
chromatin fiber reaches the contour length of the naked DNA molecule.
From that point onward the structure behaves like a DNA molecule without
any bound histones, that is, the force-extension curve exhibits the charac-
teristic 65-pN plateau. During the relaxation part of the cycle, the behavior
is that of naked DNA, showing no signs of bound proteins.
Analysis of the Opening Events

To get more detailed information on the exact structural rearrange-
ments within each opening event, the length increase involved is deter-
mined. For this, each ascending part of the force signal must be fitted to
a modified wormlike chain model.30,31 This model includes both entropic
and enthalpic (i.e., intrinsic) elasticity of the polymer:
" #
kB T 1 1 x F
F¼ þ ;
Lp 4ð1 ðx=L0 Þ þ ðF=SÞÞ2 4 L0 S
where Lp is the persistence length, L0 the contour length, x the end-to-end

distance of the polymer, kBT the thermal energy (4.1 pN nm), and S is the
stretch modulus, defined as the force needed to extend the polymer to
twice its contour length.
However, an accurate fit to this model is impossible because data points
are available only between 15 and 40 pN. Within this force range, the en-
tropic contribution to the elasticity of the polymer is negligible, which re-
duces the expression to F ¼ (S/L0)xS. This expression is used to fit the
linear portion of the curve before each drop. Using stretch moduli of sev-
eral discernible fiber intermediates along the stretch curve and assuming
that this parameter does not change significantly on a single opening event,
a contour length for each fiber intermediate can be determined.
Plotting the contour lengths as a function of time clearly reveals the
stepwise increase in length, on each opening event.29 A powerful method
to determine whether the observed steps are quantized is the pairwise
30
J. F. Marko and E. D. Siggia, Macromolecules 28, 8759 (1995).
31
M. D. Wang, H. Yin, R. Landick, J. Gelles, and S. M. Block, Biophys. J. 72, 1335 (1997).
distribution function (Fig. 6). This function bins all possible differences
between each pair of calculated contour lengths in the data set and plots
the result as a histogram. The presence of peaks in this histogram is a clear
signature of a quantized step present in the data. Figure 6B reveals a peak
at 65 nm, and multiples thereof (i.e., 130 nm, 195 nm).29 There is also a
strong peak around zero, which is a direct result of the pairwise distribution
function: Many data points that originally are on one rising part of
the curve are reduced to about the same value for the contour length,
producing differences close to zero.
Magnetic Tweezers
Materials
. Square glass cuvettes, 1 mm 1 mm 50 mm (VitroCom,
Mountain Lakes, NJ, USA, www.vitrocom.com)
. 2.8-m streptavidin-coated magnetic beads (Dynabeads, Dynal,
Oslo, Norway, www.dynal.no)
. Steel ball bearing (Allied Industrial Technologies, Cat. No.
100KSFF, H401)
. Planetary-geared 12 V DC motor, 30 rpm (Cramer, Cat. No.
800HN-DC-3277, available from Digi-Key, Thief River Falls, MN,
USA, Cat. No. CRA203-ND, www.digikey.com)
. Elenco Precision regulated power supply, model XP-620 (1.5–15 V)
. Rubber O-rings 6.7 cm 7.3 cm 0.3 cm (2–5/8 in 2–7/8 in
1/8 in) (Allied Industrial Technologies, Cat. No. 01-231 or 01-146)
. NdFeB/NIB magnets (Indigo Instruments, Waterloo, Canada,
www.indigo.com, Cat. No. 33512)
. 0.8-mm inner diameter silicon tubing (BioRad, Hercules, CA, USA,
Cat. No. 731–8210)
. 3-Com HomeConnect video camera (obsolete; try www.ebay.com
alternative digital cameras from Orange Micro, Anaheim, CA,
USA, www.orangemicro.com, and Logitech, Fremont, CA, USA,
www.logitech.com)
. Six-way low-pressure bulkhead selection valve (Upchurch Scien-
tific, Oak Harbor, WA, USA, Cat. No. V-241)
. Low-pressure injection valve (Upchurch, Cat. No. V451)
. Seiwa Optical America objective micrometer (Cat. No. RET-
PCS81X, available from Fisher Scientific, Pittsburgh, PA, USA,
Cat. No. #S11144)
. Videopoint software (Pasco Scientific, Roseville, CA, USA,

www.pasco.com) for manual particle tracking of a bead in
QuickTime movies in NIH Image
. Chemicals: 3-aminopropyltriethoxysilane (APTES) (Sigma, St.
Louis, MO, USA); toluene, research grade; NP40 (Igepal) (Sigma)
blue dextran (Sigma); mPEG (polyethylene glycol)-NH2 (5000 Da)
(Shearwater, now Nektar, Huntsville, AL, USA); PEG (3350 Da)
(Sigma)
Instrumental Setup
In the horizontal magnetic tweezers,32 a white light source (either a
fiber optic or a 150-W bulb) is focused onto a square glass cuvette
(Fig. 7A) through a condenser lens, placed roughly halfway between the
light source output and the cuvette (focusing can be facilitated by using a
white postcard to view the beam of light at the cuvette). The beam passes
through a hollow steel bearing on which a rare earth magnet is placed.
The cuvette is held by a machined piece of Delrin plastic bridge (Fig. 7B
and C) bolted onto the three-axis flexure stage. We have fashioned a
paperclip (not shown) to hold the cuvette tightly in place in the Delrin
bridge. The focused light is collected by an objective juxtaposed to the
cuvette. Most experiments were performed with 90 oil (1.2 NA) or 40
air (0.6 NA) objectives. If the objective has a fixed focal plane, typically a
distance of 160 mm, then the video camera is placed 160 mm behind
the objective. Inexpensive ($100) digital cameras can be used. These
cameras typically have either USB or Firewire direct connections to a PC
and thus do not need an extra analog/digital computer board. Public-
domain software such as NIH Image or ImageJ can be used to collect
movies in AVI format.
The magnification of the image on the video screen is determined by
the magnification of the objective and the distance from the camera to
the objective. There are a couple of ways to calibrate the number of nano-
meters per pixel in the image. One can use the dimensions of the 2.8-m
magnetic bead for rough calibration. More accurate calibration can be
done with an objective micrometer; the micrometer is a grid of fine lines
on a glass slide typically separated by distances of 10 m. This gridded slide
is placed at the position of the cuvette, and images are captured with the
video camera located at positions used during an actual experiment. It is
possible to slide the camera toward the objective to increase the field of
32
S. H. Leuba, M. A. Karymov, M. Tomschik, R. Ramjit, P. Smith, and J. Zlatanova, Proc.
Natl. Acad. Sci. USA 100, 495 (2003).
Fig. 7. Schematic of the horizontal version of the magnetic tweezers set-up and some
technical parts of the magnetic tweezers instrument. This schematic in (A) indicates some
important distances (in millimeters). The hollow bearing with the mounted magnet is movable
up to 50 mm away from the square glass cuvette in order to adjust the force on the magnetic
bead. (B) Side-view drawing of the Delrin bridge to hold the square glass cuvette. (C) Cross-
section at one end of the Delrin bridge. The three 1.4-mm furrows at the top of the piece are
the three possible locations of the cuvette. (D) Side-view drawing of the DC motor-driven
assembly (fabricated out of aluminum plate and Teflon pulleys) for rotating the external
magnet(s) on the embedded bearing. Upper right side of drawing is space on aluminum plate
to mount DC motor that directly rotates the pulley on the right. The two pulleys are
connected by a rubber O-ring. (E) Bird’seye view of the same assembly. In this view it is
possible to see that the pulley on the left is directly connected to the hollow ball bearing that
is mounted on the upper left side of the aluminum plate in (E). Dimensions are in millimeters.
view. This allows observation of a larger number of beads, finding one that
is tethered, centering it on the screen, and then enlarging the image by
manually moving the camera away from the objective. This is a flexibility
advantage of our horizontal MT set-up over the more traditional set-up
with the magnet and the cuvette positioned above the objective of an
inverted fluorescence microscope. In that set-up, the distance between
the eyepiece (or CCD camera) and the objective is fixed (160 mm); thus,
a video image can only be zoomed in and out by changing the objective;
in doing so one often loses the bead of interest.
DC Motor/Bearing Set-Up to Rotate Magnet (Fig. 7D and E)

and Gravity Flow System
The planetary-geared motor runs a pulley connected to another pulley
via a rubber O-ring. Switching the electrodes in the power supply runs the
motor in the opposite direction, thus allowing to supercoil topologically
constrained DNA molecules either positively or negatively.
Gravity-fed flow systems consistently provide the same speed and avoid
undesirable vibrations of peristaltic pumps. Half a dozen buffer bottles,
placed 70 cm above the instrument, are connected to a six-way low-pres-
sure valve. The outflow from the valve goes to a low-pressure injection
valve, which has an injection loop of 200 l, that can be used to inject
the proteins of interest. The length of time it takes for the injected buffer
to reach the cuvette can be checked by coloring the buffer with blue
dextran, and looking for the appearance of blue color on the screen.
Incorporation of Biotinylated Nucleotides into DNA

The procedure is described in the optical tweezers section. To prepare
DNA fragments that can be suspended between the surface and a bead in a
topologically constrained manner, we use PCR-based procedures (see
Strick [1999]33).
Surface Treatment of Square Glass Cuvette: Activation of the Glass

Surface with APTES (Steps 1–3), Coating the Walls of the Cuvette
with Streptavidin (Steps 4–7), and Attachment of Biotinylated
-DNA Carrying a Magnetic Bead to the Cuvette Walls (Steps 8–10)
Surface treatment of the glass cuvette is the major critical step in per-
forming the experiment. The use of glutaraldehyde to link streptavidin to
the APTES-coated surface is beneficial in stabilizing the attachment of
the tethered complexes. The high stability of the connection to the surface
allows us to routinely use 2 M NaCl to remove bound proteins and utilize
the same tethered DNA molecule for another experiment.
Buffers
. Phosphate-buffered saline (PBS) (150 mM NaCl, 3 mM KCl, 2 mM
KH2PO4, 10 mM NaH2PO4, pH 7.2)
. Washing buffer (used to wash out the unattached beads in the
cuvette after it is assembled into the instrument): PBS with 1%
NP40, 0.2% PEG (3350 Da), 1 mM EDTA, 0.2% biotin
33
T. Strick, Ph.D. Thesis, Department de Physique, Ecole Normale Superieure, Paris, 1999.
. Blue buffer: PBS with blue dextran ( just enough to see the blue),
1% NP40, 0.2% PEG, 2 M NaCl (this buffer is added to the DNA/
beads and used to visualize their entry into the cuvette)
. Histone-dissociation buffer (used to dissociate histones from the
DNA when chromatin fibers are to be disassembled; it can also used
to rinse the DNA/beads in the initial stage of the experiment): PBS
with 0.2% biotin, 1% NP40, 0.2% PEG, 1 mM EDTA, 2 M NaCl
1. Place six cuvettes in 5 ml of 1% APTES in toluene in a 15-ml

polypropylene tube in a rotator at RT for 5 min.
2. Transfer the cuvettes to another tube with 10 ml toluene and rotate
for 10 min.

3. Transfer the cuvettes to glass Petri dish and bake at 110 for 30–
45 min. At this point cuvettes can be stored in the Petri dish for
several days.
4. Pipette 10% glutaraldehyde from a freshly opened ampoule into a
cuvette; after 5 min non-shaking incubation, rinse the glutaralde-
hyde out of the cuvette with a water bottle for 1 min.
5. Mix 4 l of streptavidin solution (5 mg/ml) with 96 l PBS, pipette
the mixture into a cuvette, and let it sit for 10–20 min (we have
reused the diluted streptavidin solution 5–10 times). (If the DNA
is end-labeled with digoxigenin, a much longer deposition time,
20–60 min, is needed for the antidigoxygenin antibody to adsorb
stably.)
6. Rinse the cuvette once with a water bottle.
7. Place the cuvette in 10 ml PBS containing 50 mM glycine and 0.1%
mPEG-NH2 (both reagents block the unreacted primary amino
groups on the walls of the cuvette).
8. Mix in 0.6-ml microfuge tube 2 l of 5 M NaCl and 1 l of 2.8-m
streptavidin-coated Dynabeads suspension. Add 0.5 l of biotiny-
lated -DNA (50 g/ml in 10 mM Tris-HCl, 1 mM EDTA, pH 7.5).
Immediately add blue buffer to 50 l total volume. Wait for 5–10
min (2 M NaCl neutralizes the charges on the DNA phosphate
backbone and allows streptavidin to bind to the biotinylated DNA
more easily). The average number of DNA molecules per bead is
1/2.
9. Put the DNA/beads into the cuvette, connect tubing, and let sit for
3–5 min (NOTE 10).
10. The beads settle on the bottom of the cuvette, so the capillary needs

to be rotated 90 around its longitudinal axis to visualize the bottom
through the objective. Blue buffer is used to wash out free beads/
DNA, and washing buffer brings the NaCl concentration down to

150 mM before Brownian motion analysis is performed.
Locating a Tethered Magnetic Bead

With the flow turned off and the DC motor rotating the ball bearing to
which the external permanent magnet is attached, it is possible to observe
single magnetic beads rotating in a circle (irregular-circular or oval paths
suggest more than one DNA molecule tethering the surface to the bead).
Once a tethered bead is identified, the camera is moved further from the
objective so that the area covered by the rotating bead occupies roughly
the area of the video screen.
Calibrating the Force; Measuring the Change in Tether Length Due

to Chromatin Assembly
With the flow turned off and the external magnet at a fixed position,
Brownian motion analysis can be performed to determine the force acting
on the single DNA molecule. We will explain the particle-tracking analysis
with Videopoint, although much more sophisticated programs are avail-
able. Using the movie function of NIH image, 1000 frames of video are col-
lected at 1 frame per second. The movie is opened within Videopoint, and
the center of the tethered bead for all 1000 frames of the movie is mouse-
clicked. (Manual clicking on 1000 video frames is tedious; it is possible to
use a particle-tracking macro-function in ImageJ after thresholding the
images to remove the background.) The table of x and y coordinates of the
bead is imported into KaleidaGraph or Excel (or other spreadsheet pro-
gram). The variance of the x-data points is automatically computed, and
this variance is the <x2> term of the equipartition formula (see above).
The first movie is collected with the magnet at the closest possible po-
sition to the cuvette. Next, the external magnet stage is moved 0.5 mm
away from the cuvette, and another 1000-frame movie is collected; this
procedure is repeated until the distance to the cuvette is 0.5 cm. At each
distance, we determine <x2>. The other parameter needed to determine
the force at any given distance of the external magnet to the cuvette is the
tether length. Assuming that the DNA molecule is maximally extended
(i.e., approaches its contour length of 16.4 m) when the magnet is at its
closest possible position to the cuvette, a maximum force of 22 pN is cal-
culated using the equipartition theorem. (Using 16 m as the initial esti-
mate for the length of the tether at this position of the magnet is a
reasonable assumption, since at forces >3 pN the extension l is within
10% of the contour length of -DNA.)6 Once a first approximation force
is calculated by using this value of l and the measured <x2>, the actual
extension at this force is estimated by interpolation from the force-exten-

sion curve (Fig. 8 in ref. 6); a second approximation force is calculated by
using this interpolated value of l. After one round of iteration, the accuracy
of our measurements is 10% for forces >3 pN and 20% for forces in the
range of 0.5–3 pN.
Once we have the calibration of forces at each distance of the magnet to
the cuvette, all we need to do in the actual chromatin assembly experi-
ments is determine the shortening of the DNA tether on formation of
nucleosomes. To understand how this is done with our set-up, let us first
explain how changes in tether length are determined in the laboratory of
Croquette and Bensimon (e.g., ref. 5). In that instrument, two magnets
are used instead of one. These magnets, spaced at 2 mm, create a homoge-
neous magnetic field for the tiny magnetic bead, and keep the bead always
at the same x-, y-position on the screen (apart from the low-range Brown-
ian motions around this position). Whatever happens to the tethered DNA
molecule to change its length, the only observable change (apart from the
change in the Brownian motion) is a change in the z-dimension, z.
Fig. 8. Schematic to explain how the measurements of travel of the bead across the screen
are done, and how the tether length is calculated. (A) Cross-section of the cuvette with DNA-
tethered magnetic bead and the placement of the external magnet above the incident light;

such a placement tilts the tether so that the angle ’ is about 60 C. (B) Schematic of the
positions of the magnetic bead in the video image on chromatin assembly when the only
applied external force is the magnetic force; the only movement of the bead during assembly
is in the x-dimension of the video screen. We have defined this movement of the bead across
the screen as the ‘‘travel across the screen.’’ (C) Schematic of how the angle ’ is determined
and how travel is converted to actual changes in tether length. The angle ’ is determined by
trigonometry using xinitial and the initial length of the DNA (16 m). The actual tether
length during the assembly/disassembly experiment is then calculated by dividing the travel
(depicted in [B]) by the cosine of ’.
Following the movements of the bead in z, that is, in direction perpendicu-

lar to the focal plane is technically nontrivial and requires precalibration.
Such calibration is based on observing the size of interference rings around
the bead that become larger as the bead moves away from the focal plane.
We determine z using a different, somewhat simpler (but less accur-
ate) approach that involves no prior calibration. We place the only external
magnet not right above the sample, but to the side, so that the DNA tether

is at an angle (typically 60 ) to the imaged surface (Fig. 8A). Since the
movement of the bead on the screen is only in the x-direction (subject to
only the magnetic force, once the flow is stopped) (Fig. 8B), the travel of
the bead on the screen and knowledge of the angle are sufficient to deter-
mine z, using trigonometry (Fig. 8C). The external magnet pulls on the
DNA tether at an angle, so it is important to keep the angle the same
during the course of the experiment; this can be easily achieved by moving
the external magnet in both the vertical and horizontal dimensions by two
linear micrometer stages.
Assembling a Single Chromatin Fiber and Analysis of the Assembly

Core histones (final total concentration 0.1 mg/ml) and nucleosome as-
sembly protein 1 (NAP-1; 0.2 mg/ml) in 150 mM NaCl, 10 mM Tris-HCl,

pH 8.0, 1 mM EDTA are incubated at 37 for 30 min, diluted tenfold in
the same buffer containing 1% NP-40, 0.2% PEG, and injected into the
200-l loop of the injection valve. Using gravity flow, the mixture is allowed
to flow into the cuvette. The position of the tethered bead is followed until
the bead starts moving against the flow (typically 2–3 min), indicating that
the assembly process has begun: the wrapping of the DNA around the his-
tone octamer leads to shortening of the DNA tether. At this point the flow
is turned off. The tethered bead is now subjected to external force from the
permanent magnet only, and we can follow the travel of the bead across the
video screen as a function of the applied external force (see above).
We can now study the rate of chromatin assembly as a function of applied
force. The instrument allows to rheostatically change the force in the course
of a single experiment by moving the external magnet further and further
away from the cuvette in a stepwise fashion. Reducing the applied force
resulted in an immediate increase in the rate of chromatin assembly.32
Notes
NOTE 1: In many set-ups the quadrant detector is positioned in the
back focal plane of the condenser lens, or a position that is optically
conjugate to this. When properly set up, the detector will be completely
insensitive to lateral displacements of the optical trap and will only sense
displacements of the bead from the trap center. In our set-up, the position
of the trap is fixed, avoiding the need for an optical scheme that will render
the deflection insensitive to trap movements. The quadrant detector is
behind the focal point such that the ring-shaped illumination pattern is
0.5 cm in diameter. This position, however, is not critical.
NOTE 2: We first used a Millennia 10-W CW laser, which has good
beam pointing stability, but is expensive and large. The laser head is
44 cm 15 cm 15 cm and needs to be positioned on top of the optical
table. The controller unit is 60 cm 40 cm 26 cm. Both are connected
with a thick cable, which is not flexible and hard to install. Now, we use a
2-W infrared (1064 nm) laser (CrystaLaser) that is only 16 cm 7 cm
3 cm, with a controller box of only 15 cm 15 cm 5 cm. For a single
trap, a 2-W laser is sufficient to exert forces of 100 pN on a 2.6-m
polystyrene bead.

NOTE 3: To get 60 , start with a hot plate that has not been used for a
while. Put your finger on the plate and turn it on. Put the glass sandwich on
the plate as well. Wait until it feels hot (you have to take your finger off)
and turn off the hot plate. This is about the right temperature you need.
NOTE 4: Cut a piece of 3-mm tubing 1 to 1.5 cm in length. Pull the
thin tubing (10–20 cm in length) through this piece and fill it up with
two-component glue (the thin tubing should be coming through; otherwise
you risk closing the thin tubing with the glue). When solidified, use a sharp
knife to cut the tubing perpendicular to its axis, creating an end that is able
to make a good seal with the glass.
NOTE 5: The high NA objective used for optical trapping and for cre-
ating the microscopic image has a field of view of only 40 m 40 m. This
makes it tedious to find the tip of your micropipette. First align the micro-
pipette (by eye) as well as you can in front of the objective. Then switch to
the microscopic image and use the manual control on the stage to locate
the edge of the flow channel. Scan along the edge until you see the capillary
through which the pipette is entering the flow channel, and follow the
pipette to the tip.
NOTE 6: For connecting tubing to the selector valve, there are small
cone-shaped pieces, provided with the valve. These cones, however, do
not fit the thin tubing that we use. For connecting the thin line, we use
Micro-Lance needles. Using a very sharp scalpel, cut the needle into two
pieces at the point where the thin metal tube ends in the plastic part. This
will provide a smooth end-face in which the metal tubing ends in the center.
Take the plastic gray screws that are on the selector valve and increase the
size of the hole such that the needle is able to go through. The end-face is
still slightly larger in diameter and the needle will stick out of the screw.
Now attach the thin line to the metal needle and clamp it in the selector by
carefully screwing in the gray unit.
NOTE 7: For accurate position detection of the two beads, custom-
written video analysis routines that use a centroid method have been de-
veloped. Beads imaged by the CCD array appear as light, circular objects
surrounded by a black edge. The bead position is determined as an average
of the positions of all the pixels that are within this black circular edge.
For this, an algorithm starts at a pixel inside this edge and moves slowly
outward in all directions, until a preset threshold level is reached (i.e., that
is lower than the value of the edge). This results in 600–800 pixels that are
next to each other, forming the inner part of the imaged bead. The x- and y-
position of the bead is determined as the average x- and y-position of these
pixels (accuracy of determination of 4 nm).
NOTE 8: The end-to-end distance is obtained by subtracting the radii of
the two beads from the center-to-center distance as determined with the
centroid method (see NOTE 7).
NOTE 9: In a beam-expanding system you normally use two lenses that
are separated at a distance equal to the sum of their focal distances. Within
this optical layout, the distance between the two lenses (L1 and L2 in
Fig. 2F) is only 390 mm, instead of the expected 400. The beam coming
off lens L2 is thus slightly converging. Using such a beam will create a clear
image of the trapped bead, since the bead is trapped stably just below the
focal point. In practice, correct aligning can be done by slightly moving lens
L2 back and forth until the bead is in focus on the CCD array.
NOTE 10: Connecting the tubings in a way that avoids bubble forma-
tion requires practice. Passing bubbles appear to irreparably stick the
tethered beads to the surface, making them useless. It is recommended to
agitate the DNA-connected beads during the deposition process by making
a couple of short flow pulses (turning on and off the flow valve).
Acknowledgments
We thank Drs. Mikhail Karymov, Miroslav Tomschik, Paul Smith, and Guoliang Yang,
and Kirsten van Leijenhorst-Groener, Ravi Ramjit, and Waldemar Koscielny for help with
experiments, Dr. Haocheng Zheng for assistance with figures, and Dr. Jan Bednar for
discussion. This research has been supported by an NCI K22 grant and startup funds from the
University of Pittsburgh School of Medicine (SHL), startup funds from Polytechnic University
(JZ), and the Dutch Foundation of Fundamental Research (FOM) and MESAþ Research
Institute, University of Twente (MLB).
[7] Biochemical and Structural Characterization of

Recombinant Histone Acetyltransferase Proteins
By Ronen Marmorstein
Chromatin is modified in distinct ways to modulate gene expression.1,2

These modifications are mediated by ATP-dependent nucleosome remod-
eling proteins that move the nucleosome cores along the DNA,3,4 and his-
tone-modifying enzymes that post-translationally modify the N-terminal
histone tail regions.5 Post-translational histone modifications have been
proposed to effect gene expression through both steric/electrostatic effects
and through a signal-transduction cascade that has been called the ‘‘histone
code hypothesis.’’6,7 These post-translational histone modifications include
methylation, acetylation, phosphorylation, and ubiquitinilation (that occurs
at the histone C-terminal tail).
The enzymes that regulate histone acetylation, called HATs, were one
of the first to be characterized biochemically.8 To date, over 20 different
HAT proteins have been characterized and they fall into distinct families
that generally have high sequence similarity within the catalytic HAT
domain and related substrate specificity within families, but have divergent
sequence and substrate specificity between families (Table I).9 In addition,
HAT proteins usually contain other conserved domains with autonomous
transcription-related functions that also seem to be family specific, consis-
tent with the ability of HAT proteins to coordinate other transcription-
related activities.1 In vivo, HAT enzymes are the catalytic subunits of
multisubunit protein complexes, and distinct HAT proteins acetylate spe-
cific lysine residues within specific N-terminal histone tails.10 There have
been several reports showing that subunits of HAT complexes, among
other transcription-related activities, play a modulatory role on the level
of enzymatic activity and the specificity for substrate.11–13 Nonetheless,
1
R. Marmorstein, Nat. Rev. Mol. Cell. Biol. 2, 422 (2001).
2
Y. Zhang and D. Reinberg, Genes Develop. 15, 2343 (2001).
3
C. L. Peterson, EMBO Reports 3, 319 (2002).
4
K. E. Neely and J. L. Workman, Biochim. Biophys. Acta 1603, 19 (2002).
5
P. A. Grant, Genome Biol. 2, REVIEWS0003 (2001).
6
B. D. Strahl and C. D. Allis, Nature 403, 41 (2000).
7
S. L. Schreiber and B. E. Bernstein, Cell 111, 771 (2002).
8
J. E. Brownell, J. Zhou, T. Ranalli, R. Kobayashi, D. G. Edmondson, S. Y. Roth, and C. D.
Allis, Cell 84, 843 (1996).
9
R. Marmorstein, Cell. Mol. Life Sci. 58, 693 (2001).
10
D. E. Sterner and S. L. Berger, Microbiol. Mol. Biol. Rev. 64, 435 (2000).

[7] biochemical characterization of HATs 107
TABLE I
HAT Families and Their Transcription-Related Functions
HAT Organism Function Histonea
GCN5/PCAF family H3 (K14)

Gcn5 Yeast to Coactivator
human (adaptor)
PCAF Human Coactivator þ Non-histone
proteins
MYST family H4
Sas2 Yeast Silencing
Sas3 Yeast Silencing (H3)
Esa1 Yeast Cell cycle
progression
MOF Fruit fly Dosage
compensation
Tip60 Human HIV Tat interaction
MOZ Human Leukeomogenesis
HBO1 Human Origin recognition
interaction
TAFII250 family Yeast to TBP-associated H3
human factor
CBP/p300 family Worm to Global All þ non-histone
human coactivator proteins
SRC family Mice and Steroid receptor H3/H4
human coactivators
SRC-1 coactivators
ACTR/AIB1/pCIP/TRAM-1/RAC3
SRC-3
TIF-2
GRIP1
ATF-2 Yeast to Sequence specific H4, H2B
human DNA-binding activator
HAT1 family Yeast to Replication-dependent H4
human chromatin assembly
(cytoplasmic)
a
Only preferred histone substrates are indicated.
11
R. Balasubramanian, M. G. Pray-Grant, W. Selleck, P. A. Grant, and S. Tan, J. Biol. Chem.
277, 7989 (2002).
12
L. A. Boyer, M. R. Langer, K. A. Crowley, S. Tan, J. M. Denu, and C. L. Peterson, Mol.
Cell 10, 935 (2002).
13
P. A. Grant, A. Eberharter, S. John, R. G. Cook, B. M. Turner, and J. L. Workman, J. Biol.
Chem. 274, 5895 (1999).
considerable information on HAT function has been derived from

biochemical and structural characterization of the isolated HAT domains.
This chapter will focus on the biochemical and structural characteriza-
tion of recombinant HAT proteins. In particular, I will focus on the meth-
odologies used to prepare HAT proteins for such studies and highlight
what we have learned from these studies. Finally, I will discuss what we
have still to learn about the activity of HAT proteins.
Identification of HAT Domains for Overexpression in Bacteria

Bacterial expression systems have been effectively used to overexpress
the catalytic domain of several HAT proteins. In my laboratory, the ex-
pression of intact HAT protein or proteins containing additional domains
tethered to the catalytic domain, for example, the bromodomain, has re-
sulted in the expression of protein that is largely insoluble or, if soluble,
proteolytically cleaved between the HAT domain and the other associated
domain. HAT proteins from the Gcn5/PCAF family were the first nuclear
HAT proteins to be overexpressed for biochemical and structural stud-
ies.14–16 One often uses sequence homology within a family of related pro-
teins to define structural domains, and this strategy was effectively used for
overexpressing HAT domains from the Gcn5/PCAF family for biochem-
ical analysis. The high degree of sequence homology among the Gcn5/
PCAF family of HAT proteins within a region that colocalized to a region
that was mutationally sensitive for HAT activity and transcription17,18 was
used to define the structural catalytic domain (Fig. 1A). The HAT domain
for overexpression included this region of homology and 2–3 additional
flanking residues. The proteins and amino-acid boundaries chosen for over-
expression were yeast Gcn5 (yGcn5, residues 99–261), human PCAF
(hPCAF, residues 493–658) and Tetrahymena GCN5 (tGCN5, residues
48–210). In the case of Tetrahymena, each of the UAG codons, which code
for Gln in Tetrahymena but code for STOP in bacteria, had to be changed
to CAG, thus coding for Gln when tGcn5 was expressed in bacteria.16
14
A. Clements, J. R. Rojas, R. C. Trievel, L. Wang, S. L. Berger, and R. Marmorstein, EMBO
J. 18, 3521 (1999).
15
R. C. Trievel, J. R. Rojas, D. E. Sterner, R. Venkataramani, L. Wang, J. Zhou, C. D. Allis,
S. L. Berger, and R. Marmorstein, Proc. Natl. Acad. Sci. USA 96, 8931 (1999).
16
J. R. Rojas, R. C. Trievel, J. Zhou, Y. Mo, X. Li, S. L. Berger, D. Allis, and R. Marmorstein,
Nature 401, 93 (1999).
17
M. H. Kuo, J. X. Zhou, P. Jambeck, M. E. A. Churchill, and C. D. Allis, Genes Develop. 12,
627 (1998).
18
L. Wang, L. Liu, and S. L. Berger, Genes Dev. 12, 640 (1998).
Each of the HAT domains was cloned into the PRSET plasmid vector

for overexpression in Escherichia coli BL21 (DE3) cells. Induction at 37
with IPTG resulted in robust protein induction; however, upon cell disrup-
tion by sonication, the majority of protein was found in the insoluble pro-
tein fraction. In contrast, if the protein was induced with IPTG and induced

overnight at 15 , a significant fraction of the protein (between 25% and
50%) was found in the soluble protein fraction. Although the reason for
this is not clear, it has been proposed that the lower temperature allows
the protein to fold more slowly, resulting in more ‘‘native like’’ folding
and less improper folding that typically leads to protein aggregation.
My laboratory was also able to overexpress the HAT domain from
members of the MYST HAT family that were suitable for biochemical
study19; however, this has not been as straightforward as with the HAT do-
mains from the Gcn5/PCAF family. As with the Gcn5/PCAF HAT
domain, the regions chosen for overexpression relied heavily on sequence
homology among the MYST HAT members (Fig. 1B). Various deletion
constructs of the yeast Esa1 members of the MYST HAT family were pre-
pared and bacterial extracts containing these deletion mutants were
screened for HAT activity. This analysis localized the HAT active domain
to residues 147–445 (the C-terminus of the protein). Induction of a yEsa1
protein construct at 15 , harboring residues 147–445, showed good expres-
sion of soluble protein; however, the protein showed visible degradation
during purification. Inspection of the sequence conservation within the
MYST family revealed that residues 147–162 and 434–445 were poorly con-
served, and based on this we prepared a yEsa1 construct harboring residues
160–435 that was amenable to the overexpression of intact soluble protein
that was suitable for biochemical and structural studies. Subcloning analo-
gous regions of the human homologue of yEsa1, we have more recently
been able to overexpress hMOZ and hTIP60 in bacteria (unpublished).
However, in these cases the proteins were prepared as N-terminal
6 His-fusion proteins and also contained a C-terminal KKK sequence
to improve protein solubility.
Purification of HAT Proteins

Purification of the Gcn5/PCAF HAT domains took advantage of the fact
that these proteins use the acetyl-Coenzyme-A (Ac-CoA) cofactor and
employed a CoA agarose affinity column in the purification.14–16 Chromatog-
raphy on CoA-agarose was followed by cation exchange chromatography
19
Y. Yan, N. A. Barlev, R. H. Haley, S. L. Berger, and R. Marmorstein, Mol. Cell 6,
1195 (2000).
Fig. 1. Sequence homology among HAT proteins. (A) Sequence homology among the
Gcn5/PCAF family of HAT proteins. Residues with black and grey shading are identical and
conserved, respectively. Secondary structural elements within the enzymatic core (black) and
(SP-sepharose) and gel filtration chromatography (Superdex-75). CoA-

agarose appeared to be critical for enriching for ‘‘natively folded’’ protein
that was capable of binding CoA and monomeric protein that eluted at an
appropriate position on gel filtration chromatography. We also observed
that the proteins were more stable when purified and stored in a low-pH
buffer, and we found Na citrate, pH 6.0, to be a suitable buffer. We presume
that the low pH has two advantages. First, it is far from the pKa of these
proteins (which is above 8) and therefore increases the surface charge of
the protein for increased solubility. Second, enzymatic studies reveal that
the protein has low activity at this pH, thus disfavoring autoacetylation that
may destabilize the protein. Following purification of the Gcn5/PCAF HAT
domains, they were each concentrated to between 20 and 40 mg/ml in a
buffer containing 20 mM Na citrate, 150 mM NaCl, and 1–10 mM reducing

agent and stored at 70 until use.
Purification of the yEsa1 HAT domain was carried out essentially as de-
scribed for the Gcn5/PCAF HAT domains.19 However, in the case of
yEsa1, CoA-agarose affinity resulted in significant protein loss, and so only
cation exchange and gel filtration chromatography were used. In addition,
protein concentration beyond 8 mg/ml resulted in protein precipitation,
so protein was generally not concentrated beyond this point.
Crystallization of HAT Proteins

The Gcn5, PCAF, and Esa1 HAT proteins were crystallized using hang-
ing drop vapor diffusion and standard crystallization screens.14–16 We have
found that crystals of liganded forms of the HAT proteins are more easily
obtained than the free proteins, whereby the CoA or Ac-CA bound forms
are the most straightforward to obtain. Indeed, the nascent tGcn5 and
yGcn5 crystals that we were able to obtain were not ‘‘true’’ nascent pro-
teins. The nascent yGcn5 contained the C-terminus from a symmetry-
related yGcn5 molecule in the histone-binding cleft15 and nascent tGCN5
contained a HEPES buffer molecule bound to the CoA-binding region.16
This, coupled with the fact that the PCAF HAT domain could only be crys-
tallized in the presence of CoA or Acetyl-CoA in our laboratory, suggests
that the HAT domains of these proteins become more rigid in the presence
of a ligand, making them more amenable to crystallization. This is consist-
ent with another published report showing that the cofactor-liganded forms
N- and C-terminal regions (grey) are color-coded. The ‘‘A’’ and ‘‘H’’ symbols represent
residues that mediate Ac-CoA and histone contacts, respectively, and the catalytic glutamate
general base is indicated with a ‘‘’’ symbol. (B) Sequence homology among the MYST
family of HAT proteins. The color-coding and symbol designation are as described in (A). The
catalytic glutamate and cysteine residues are indicated with a ‘‘’’ symbol.
of Gcn5/PCAF have greater stability in solution (as measured by protease

resistance) than their unliganded counterparts.20
Interestingly, crystallization of a ternary complex with CoA cofactor
product and a histone H3 peptide substrate could only be obtained with
the tGcn5 member of the recombinant Gcn5/PCAF HAT domains. This
does not appear to be related to differential affinities of histone peptides
for members of the Gcn5/PCAF HAT domains, but rather to accessibility
of the histone peptide binding site within the crystal lattice. Indeed, we
have recently been able to co-crystallize tGcn5 with histone H4 and p53
peptides (unpublished), two substrates that have Km/Kcat values 1000- to
10,000-fold lower than histone H3, respectively.21
Overall Structure of HAT Proteins

My laboratory has determined the structure of HAT proteins from the
Gcn5/PCAF14–16,22 and MYST subfamily19,23 in various liganded forms,
and Dutnall and co-workers have determined the structure of the cytoplas-
mic yeast HAT1 enzymes bound to CoA.24 A comparison of these structures
reveals structural homology within a central core domain and structural
divergence in regions amino-and carboxyl-terminal to this core domain
(Fig. 2). The structure of ternary complexes between Tetrahymena Gcn5
(tGCN5) with CoA and either an 11 (16)- or 19-residue (unpublished)
histone H3 peptide cantered around the preferred lysine-14 target reveals
the mode of cosubstrate binding. The protein structure contains an L-shaped
cleft, with CoA bound in the short segment and the histone peptide bond
in the long segment of the L-shaped cleft. A correlation of the structure
with related biochemical and mutagenesis studies reveals that the central
domain plays a particularly important role in Ac-CoA binding and catalysis,
while the regions N- and C-terminal to the catalytic core domain play a
particularly important role in histone substrate specific binding within Gcn5.
Enzymology of HAT Proteins

To study the enzymology of HAT domains, our laboratory has used a
sensitive, fluorescence-based assay using the sulfhydryl-sensitive dye, CPM
(7-diethylamino-3-(40 -maleimidylphenyl)-4-methylcoumarin), that reacts
20
J. E. Herrera, M. Bergel, X. J. Yang, Y. Nakatani, and M. Bustin, J. Biol. Chem. 272,
27253 (1997).
21
R. C. Trievel, F.-Y. Li, and R. Marmorstein, Anal. Biochem. 287, 319 (2000).
22
Y. Lin, C. M. Fletcher, J. Zhou, C. D. Allis, and G. Wagner, Nature 400, 86 (1999).
23
Y. Yan, S. Harper, D. W. Speicher, and R. Marmorstein, Nat. Struct. Biol. 9, 862 (2002).
24
R. N. Dutnall, S. T. Tafrov, R. Sternglanz, and V. Ramakrishnan, Cell 94, 427 (1998).
Fig. 2. Overall structure of HAT proteins. (A) Schematic structure of the Gcn5/PCAF
HAT domain. tGcn5 on complex with CoA (red) and a 19-residue histone H3 peptide (green).
The structurally conserved catalytic core domain is colored in blue, and the structurally
variable N- and C-terminal domains are colored in aqua. (B) Schematic structure of the yEsa1
member of the MYST HAT domain in complex with CoA (red). The color coding is as
indicated in (A). (C) Schematic structure of the yHAT1 HAT domain in complex with Acetyl-
CoA (red). The color coding is as indicated in (A).
with free CoA, that is generated upon substrate acetylation.21 Enzymatic

reactions are typically carried out in the presence of protein, Ac-CoA,
and substrate and quenched by the addition of isoproponal. CPM is then
added to the quenched reaction, which reacts with the free sulfhydryl of
the CoA reaction product. The fluorescence of the samples is then read
with emission and excitation filters of 385 and 465 nm, respectively, which
is appropriate for the CoA-modified CPM chromaphor. To obtain steady-
state parameters, reactions are carried out in the linear range of protein
concentration and reaction time with saturating concentrations of one of
the substrates and varying the concentration of the other substrate. Reac-
tions are also typically carried out in 96-well white plates and analyzed with
a fluorescence microplate reader. This assay is sensitive, reproducible, and

convenient; however, a shortcoming of the assay is that the CPM chroma-
phore reacts avidly with all accessible free sulfhydryl groups and will there-
fore react with free cysteine residues within the protein or substrate. If
either the protein or substrate contains surface-exposed cysteine residues,
they will react with the CPM molecule and thereby increase the back-
ground of the signal. The assay cannot be carried out in real time because
the reaction of CPM with free sulfhydryls on the protein may also modify
the activity of the protein. Nonetheless, the fluorescence-based assay has
been used successfully to obtain steady-state kinetic parameters for
each of the recombinant HAT proteins that we have worked with in the
laboratory, including yGcn5, tGcn5, hPCAF, and yEsa1.19,21,23
Other assays that have been employed to enzymatically characterize
HAT proteins include radioactive assays and a spectrophotometric
enzyme-linked assay. For the radioactive assays, [3H]acetyl-CoA25 or
[14C]acetyl-CoA26 is used as a substrate to generate the labeled histone
products. The amount of labeled histone is then quantitated by either
blotting the labeled histone onto filter paper for scintillation counting
(for [3H]) or separating the labeled histone on SDS Tris-Tricine polyacryl-
amide gels and quantifying by phosphorimage analysis (for [14C]). The
continuous spectrophotometric enzyme-linked HAT assay that has been
developed that uses a coupled enzyme system with either -ketoglutarate
dehydrogenase or pyruvate dehydrogenase.27
The enzymology that has been carried out on the Gcn5/PCAF family
and the yEsa1 member of the MYST family reveals that they use different
catalytic mechanisms. The Gcn5/PCAF HAT proteins use a sequential
ordered mechanism, whereby the acetyl moiety of Ac-CoA is transferred
directly from the cofactor to the N nitrogen of the substrate lysine residue
(Fig. 3A).15,16,25 In contrast, the yEsa1 member of the MYST family of
HAT proteins employs a ping-pong catalytic mechanism, whereby the
acetyl moiety from Ac-CoA is first transferred to a cysteine residue within
the active site of the protein, and in a second step this acetyl moiety is
transferred to the N nitrogen of the substrate lysine residue (Fig. 3B).23
By sequence homology, it appears that all members of the MYST
HAT family may use the same catalytic mechanism as yEsa1. Despite
the different catalytic mechanisms between the two HAT families, both
contain a conserved glutamate residue in the active site that appears to
25
K. G. Tanner, R. C. Trievel, M.-H. Kuo, R. Howard, S. L. Berger, C. D. Allis,
R. Marmorstein, and J. M. Denu, J. Biol. Chem. 274, 18157 (1999).
26
O. D. Lau, T. K. Kundu, R. E. Soccio, S. Ait-Si-Ali, E. M. Khalil, A. Vassilev, A. P. Wolffe,
Y. Nakatani, R. G. Roeder, and P. A. Cole, Mol. Cell 5, 539 (2000).
27
Y. Kim, K. G. Tanner, and J. M. Denu, Anal. Biochem. 280, 308 (2000).
Fig. 3. Enzymatic mechanism of HAT proteins. (A) Proposed catalytic mechanism for the
Gcn5/PCAF HATs. Residue numbering for tGcn5 is indicted. (B) Proposed catalytic
mechanism for the MYST HATs. Residue numbering for yEsa1 is indicated.
deprotonate the N nitrogen of the target lysine substrate prior to acetyl

transfer to substrate.
It is very surprising that the Gcn5/PCAF and MYST family of HAT
enzymes, two functionally and structurally related enzyme families, use dif-
ferent catalytic mechanisms. Interestingly, a recent report from the Cole
group shows that the p300 HAT, a member of yet another HAT family,
appears to use yet a different catalytic mechanism.28 It is attractive to
28
P. R. Thompson, H. Kurooka, Y. Nakatani, and P. A. Cole, J. Biol. Chem. 276, 33721 (2001).
Fig. 4. Substrate binding by HAT proteins. (A) tGcn5/CoA/histone H3 complex. CoA is

colored in purple and the ordered regions of the 19-residue histone H3 peptide (residues 7–21)
are colored in green. An overlay of the histone H3 peptide from the ternary complex with the
11-residue histone H3 peptide (residues 9–19) is shown in grey. (B) Schematic tGcn5-histone
H3 interactions in the ternary tGcn5/CoA/histone H3 complex shown in (A). Regions of the
hypothesize that the difference in catalytic mechanism between these

different HAT families may be related to their different in vivo properties.
Histone-Specific Binding by HAT Proteins

The crystal structure of the ternary complex between tGCN5, CoA, and
an 11-residue histone H3 peptide (H3p11) reveals that the histone sub-
strate sits in a cleft formed by the central protein core and the pantetheine
arm of CoA and is flanked by N- and C-terminal HAT domain segments
that mediate most of the tGCN5-histone H3 contacts.16 The participation
of the pantetheine arm of CoA in histone H3 contacts is consistent with
the important structural role played by CoA in histone recognition.
Coupled with the increased stability of the cofactor-liganded Gcn5/PCAF
enzymes, we propose that Ac-CoA binds the enzyme prior to histone
substrate binding. One surprising feature of the tGCN5/CoA/H3p11 struc-
ture is that most of the peptide-protein interactions are mediated through
the backbone of the peptide, in its C-terminal region, whereas the
N-terminal residues 9–13 are flexible with disordered side chains in the
crystal structure. This result, coupled with the biochemical studies demon-
strating that GCN5/PCAF family members exhibit a high degree of speci-
ficity for lysine 14 of histone H3,17,18,29,30 raised the possibility that there
are other structural determinants for substrate binding specificity by the
GCN5/PCAF HAT domain. In support of this, the PCAF HAT domain
has a greater than 100-fold enhanced substrate specificity (kcat/Km) for a
19–amino acid H3-derived substrate peptide (residues 5–23; H3p19)
when compared to H3p11,21 with only a slightly elevated activity toward
a 27-residue histone H3 peptide (1.9-fold relative to H3p19). Taken
together, these studies suggest that the 19-residue histone H3 peptide is
sufficient to mediate most, if not all, of the histone H3 interactions with
the Gcn5/PCAF HAT domain.
A more recent structure of tGcn5 in ternary complex with CoA and a
19-residue histone H3 peptide shows that 15 of the 19 histone H3
29
M. H. Kuo, J. E. Brownell, R. E. Sobel, T. A. Ranalli, R. G. Cook, D. G. Edmondson, S. Y.
Roth, and C. D. Allis, Nature 383, 269 (1996).
30
R. L. Schiltz, C. A. Mizzen, A. Vassilev, R. G. Cook, C. D. Allis, and Y. Nakatani, J. Biol.
Chem. 274, 1189 (1999).
histone H3 peptide that are ordered in the ternary complex with the 19-residue peptide, but
not ordered in the ternary complex with the 11-residue peptide, are colored in grey. (C)
Superposition of putative substrate binding sites of HAT proteins. The superposition is
generated by superimposing the core domains from the Gcn5/PCAF member, PCAF (pink),
and the MYST member, yEsa1 (blue), with yHAT1 (green). Only the core domain and CoA
of yEsa1 (blue) is shown for clarity.
residues are ordered with only two disordered residues at each end of the
peptide (Fig. 4A and B) (unpublished). Surprisingly, relative to the com-
plex containing H3p11, the H3p19 complex reveals significantly more ex-
tensive protein-peptide interactions. The greatest structural alterations
between the two peptides occur in residues N-terminal to lysine 14, as
H3p19 makes more interactions in this region. Most strikingly, a core of
12 residues of histone H3 centered on lysine 14 is well ordered in the
tGCN5/CoA/histone H3p19 complex. Each of these residues makes
sequence-specific contacts to protein regions N- and C-terminal to the cata-
lytic core of tGCN5. This comparison reveals that residues outside the core
12-residue sequence of the histone H3 peptide anchor and reposition the
core of histone H3 for more optimal enzyme-histone contacts.
The protein regions N- and C-terminal to the catalytic domains of Esa1
and HAT1 show structural divergence to the corresponding regions of
Gcn5/PCAF (Fig. 2). Despite this, a supposition of the respective core
domains of the two protein families reveals that the N- and C-terminal
regions of Gcn5 that specifically contact histone substrate have structurally
overlapping counterparts in the Esa1 structure (Fig. 4C).19 This suggests
that the Gcn5/PCAF and MYST HATs, and possibly other HAT families,
may have a similar structural scaffold for substrate-specific recognition
and that the sequence divergence within this scaffold may contribute to
substrate-specific binding.
Conclusions and Future Prospects

The methodologies that we have outlined for preparing recombinant
HAT domains of the Gcn5/PCAF and MYST family of HAT proteins
should be generally applicable for the preparation of HAT domains from
other HAT families. The functional and structural studies carried out on
the Gcn5/PCAF and MYST HAT families have provided important in-
sights into the overall architecture of HAT proteins and the mode of cata-
lysis and substrate binding. The architectural features of the Gcn5/PCAF
and MYST family lead to the hypothesis that all HAT proteins will have
a structurally conserved catalytic core domain with divergent catalytic
mechanisms and structurally divergent but related N- and C-terminal
domains that are correlated with substrate binding. The structure of
HAT domains from other HAT families needs to be determined in order
to test and extend this hypothesis. In particular, the mode of substrate
binding specificity is an area that requires considerably more attention. In
addition, it is clear that other protein subunits of the in vivo HAT
complexes modify HAT functional at both the level of catalysis and
substrate binding. The structural and mechanistic basis for this is an
[8] use of nuclear magnetic resonance spectroscopy 119
important issue that will require structural studies of relevant HAT com-
plexes. These structures may be at the heart of understanding how HAT
activity is coordinated with other histone-modifying activities to faithfully
modulate transcriptional regulation.
Acknowledgments
I would like to acknowledge all past and present members of my laboratory who have
contributed to the studies discussed in this chapter. In particular, I would like to thank
A. Clements, M. Holbert, A. Poux, J. Rojas, T. Sikorski, R. Trievel, and Y. Yan.
[8] Use of Nuclear Magnetic Resonance Spectroscopy

to Study Structure-Function of Bromodomains
By Shiraz Mujtaba and Ming-Ming Zhou
Characterization of the evolutionarily conserved protein modular

domains in signaling proteins that recognize post-translationally modified
amino acids or unique sequence motifs in a protein has revolutionized
our understanding of regulation of protein-protein interactions or enzyme
activities in signal transduction that govern cell growth, proliferation,
differentiation, and apoptosis.1 Chromatin remodeling represents another
important frontier in cell biology.2,3 While recent studies have identified
numerous conserved protein modules in many proteins and enzymes
linked to chromatin remodeling,3–5 their detailed molecular mechanisms
remain elusive. Nuclear magnetic resonance (NMR) spectroscopy is a
powerful tool not only for determination of high-resolution 3D structures
of protein domains but also for investigation of their biochemical
functions. The resulting structural and functional inferences can help us
gain important insights into the molecular mechanisms underlying chroma-
tin-mediated epigenetic control processes, including transcriptional acti-
vation and repression, as well as gene silencing. Here we describe the
1
T. Pawson and P. Nash, Genes Dev. 14, 1027 (2000).
2
R. D. Kornberg and Y. Lorch, Cell 98, 285 (1999).
3
S. Bjorklund, G. Almouzni, I. Davidson, K. P. Nightingdale, and K. Weiss, Cell 96,
759 (1999).
4
F. Jeanmougin, J. M. Wurtz, B. L. Douarin, P. Chambon, and R. Losson, Trends Biochem.
Sci. 22, 151 (1997).
5
R. Aasland, T. J. Gibson, and A. F. Stewart, Trends Biochem. Sci. 20, 56 (1995).

important issue that will require structural studies of relevant HAT com-
plexes. These structures may be at the heart of understanding how HAT
activity is coordinated with other histone-modifying activities to faithfully
modulate transcriptional regulation.
Acknowledgments
I would like to acknowledge all past and present members of my laboratory who have
contributed to the studies discussed in this chapter. In particular, I would like to thank
A. Clements, M. Holbert, A. Poux, J. Rojas, T. Sikorski, R. Trievel, and Y. Yan.
[8] Use of Nuclear Magnetic Resonance Spectroscopy

to Study Structure-Function of Bromodomains
By Shiraz Mujtaba and Ming-Ming Zhou
Characterization of the evolutionarily conserved protein modular

domains in signaling proteins that recognize post-translationally modified
amino acids or unique sequence motifs in a protein has revolutionized
our understanding of regulation of protein-protein interactions or enzyme
activities in signal transduction that govern cell growth, proliferation,
differentiation, and apoptosis.1 Chromatin remodeling represents another
important frontier in cell biology.2,3 While recent studies have identified
numerous conserved protein modules in many proteins and enzymes
linked to chromatin remodeling,3–5 their detailed molecular mechanisms
remain elusive. Nuclear magnetic resonance (NMR) spectroscopy is a
powerful tool not only for determination of high-resolution 3D structures
of protein domains but also for investigation of their biochemical
functions. The resulting structural and functional inferences can help us
gain important insights into the molecular mechanisms underlying chroma-
tin-mediated epigenetic control processes, including transcriptional acti-
vation and repression, as well as gene silencing. Here we describe the
1
T. Pawson and P. Nash, Genes Dev. 14, 1027 (2000).
2
R. D. Kornberg and Y. Lorch, Cell 98, 285 (1999).
3
S. Bjorklund, G. Almouzni, I. Davidson, K. P. Nightingdale, and K. Weiss, Cell 96,
759 (1999).
4
F. Jeanmougin, J. M. Wurtz, B. L. Douarin, P. Chambon, and R. Losson, Trends Biochem.
Sci. 22, 151 (1997).
5
R. Aasland, T. J. Gibson, and A. F. Stewart, Trends Biochem. Sci. 20, 56 (1995).

procedures recently used to delineate structure-function relationships of

the bromodomains.
The bromodomain of 110 amino acids, first reported in the Drosophila
protein brahma (hence the name),6,7 represents an extensive family of evo-
lutionarily conserved protein modules found in many chromatin-associated
proteins and in nearly all known nuclear histone acetyltransferases
(HATs).4 It has been long implicated from yeast genetic and biochemical
studies that bromodomains play an important role in chromatin remodel-
ing8–10 on the basis of their importance in the assembly and activity of mul-
ti-protein complexes of chromatin remodeling,11,12 as well as by the fact
that the bromodomain module is indispensable for the function of GCN5
in yeast.13,14 For example, it has been shown that deletion of a bromodo-
main in human HBRM, a protein in the SWI/SNF remodeling complex,
causes both decreased stability and loss of nuclear localization.15,16 Bromo-
domains of Bdf1, a Saccharomyces cerevisiae protein, are required for spor-
ulation and normal mitotic growth.17 Finally, bromodomain deletion in
Sth1, Rsc1, and Rsc2, three members of the nucleosome remodeling
complex, can cause a conditional lethal phenotype (in Sth1)18 or a strong
phenotypic inhibition on cell growth (in Rsc1 and Rsc2).19 Notably, the
phenotypic effect observed in Rsc1 and Rsc2 results from deletion of
only the second but not the first bromodomain, suggesting that these two
bromodomains serve distinct functions through interactions with different
6
J. W. Tamkun, R. Deuring, M. P. Scott, M. Kissinger, A. M. Pattatucci, T. C. Kaufman, and
J. A. Kennison, Cell 68, 561 (1992).
7
S. R. Haynes, C. Dollard, F. Winston, S. Beck, J. Trowsdale, and I. B. Dawid, Nucleic Acids
Res. 20, 2603 (1992).
8
J. E. Brownell, J. Zhou, T. Ranalli, R. Kobayashi, D. G. Edmondson, S. Y. Roth, and
C. D. Allis, Cell 84, 843 (1996).
9
P. Filetici, C. Aranda, A. Gonzalez, and P. Ballario, Biochem. Biophys. Res. Commun. 242,
84 (1998).
10
G. A. Marcus, N. Silverman, S. L. Berger, J. Horiuchi, and L. Guarente, EMBO J. 13,
4807 (1994).
11
C. E. Brown, L. Howe, K. Sousa, S. C. Alley, M. J. Carozza, S. Tan, and J. L. Workman,
Science 292, 2333 (2001).
12
D. E. Sterner, P. A. Grant, S. M. Roberts, L. J. Duggan, R. Belotserkovskaya, L. A. Pacella,
F. Winston, J. L. Workman, and S. L. Berger, Mol. Cell. Biol. 19, 86 (1999).
13
T. Georgakopoulos, N. Gounalaki, and G. Thireos, Mol. Gen. Genet. 246, 723 (1995).
14
P. Syntichaki, I. Topalidou, and G. Thireos, Nature 404, 414 (2000).
15
C. Muchardt, B. Bourachot, J. C. Reyes, and M. Yaniv, EMBO J. 17, 223 (1998).
16
C. Muchardt and M. Yaniv, Semin. Cell. Dev. Biol. 10, 189 (1999).
17
P. Chua and G. S. Roeder, Mol. Cell. Biol. 15, 3685 (1995).
18
J. Du, I. Nasir, B. K. Benton, M. P. Kladde, and B. C. Laurent, Genetics 150, 987 (1998).
19
B. R. Cairns, A. Schlichter, H. Erdjument-Bromage, P. Tempst, R. D. Kornberg, and
F. Winston, Mol. Cell 4, 715 (1999).
biological ligands.19 The recent NMR-based structure-function analysis of

the prototypical bromodomain from the transcriptional coactivator PCAF
(p300/CBP-associated factor) demonstrates that bromodomains function
as acetyl lysine–binding domains,20 which offers insights into the molecular
basis of biological functions of bromodomains in a wide variety of cellular
events, including chromatin remodeling and transcriptional activation.21–24
In this chapter, we describe use of the NMR-based methods to study
structure and function of bromodomains, which is generalizable for other
conserved protein modular domains in chromatin remodeling.
Preparation of Protein Samples
Protein Expression and Stable Isotope Labeling

The cDNA construct that encodes the bromodomain of PCAF (residues
719–832) used in the NMR structural analysis is designed on the basis of
sequence analysis, which shows that this region of PCAF contains the con-
served bromodomain present in many other proteins.4,20 The expression
construct is ligated into the pET14b vector (Novagen) between Nde1 and
BamH1 sires. The recombinant protein expressed in Escherichia coli BL21
(DE3) cells contains a hexahistidine tag at its amino terminus followed by a
thrombin cleavage site. After confirming the clone by DNA sequencing,
protein expression studies are conducted in E. coli BL21 (DE3) cells. Ini-
tial optimization of experimental conditions for expression and solubility of
the protein is conducted with small- and large-scale culture in Luria-
Bertani (LB) media. Subsequently, protein samples for the NMR structural
study is expressed in an M9 minimal medium. Ingredients of the minimal
medium consist of NaCl (0.5 g/L), NH4Cl (1 g/L), KH2PO4 (3 g/L), and
Na2HPO4-H2O (6 g/L), which after sterilization by autoclave are added to
glucose (4 g/L), vitamin B1 (0.0005%), MgSO4 (1 M), CaCl2 (100 M)
and ampicillin (100 mg/L). For a typical protein expression experiment,
bacterial BL21(DE3) cells transformed with pET14b-bromodomain are

grown overnight at 30 in 100 ml of the M9 minimal medium containing
ampicillin (100 mg/L). This starter culture is used to inoculate 1 L of fresh

M9 media, which is then incubated at 30 for a few hours until OD600
reaches about 0.5. Protein expression is induced by adding 200–400 M
20
C. Dhalluin, J. E. Carlson, L. Zeng, C. He, A. K. Aggarwal, and M.-M. Zhou, Nature 399,
491 (1999).
21
M. H. Dyson, S. Rose, and L. C. Mahadevan, Front Biosci. 6, 853 (2001).
22
F. Winston and C. D. Allis, Nat. Struct. Biol. 6, 601 (1999).
23
24
L. Zeng and M.-M. Zhou, FEBS Lett. 513, 124 (2001).
isopropyl- -d-thiogalactopyranoside (IPTG) to the cell culture, which is

further incubated for another 10–12 h at 18 . Cells are harvested by centri-

fugation at 3000g for 30 min at 4 , and cell pellets are collected for protein
purification (see later) or quick freezing in liquid nitrogen for storage at
80 . This expression procedure is also used to make various stable
isotope (15N, 13C/15N or 2H/13C/15N)-labeled proteins for the NMR struc-
tural study. Particularly, the labeled proteins are prepared from bacterial
cells grown in a minimal medium containing 15NH4Cl with or without
13
C6-glucose in H2O or 2H2O. 15NH4Cl and 13C6-glucose provide sole
nitrogen and carbon sources for recombinant protein expressed in E. coli.
Protein Purification
The harvested bacterial cells are resuspended in a lysis buffer [50 mM
Tris-HCl of pH 8.0, containing 10% glycerol, 1% NP-40, 300 mM NaCl,
1 mM PMSF, and EDTA free protease inhibitors (one tablet per liter of
cell culture) (Roche)] and subjected to sonication. Cellular debris is re-
moved by centrifugation at 100,000g for 20 min, and the supernatant
obtained is used for subsequent protein purification.
The bromodomain protein is first purified by affinity chromatography
on a nickel-IDA column (Invitrogen) using an FPLC system (Amersham
Pharmacia Biosciences). The cell lysate is applied to the nickel resin
column of 5 ml (for 3–4 L of cell culture) that is pre-equilibrated in a bind-
ing buffer [50 mM Tris-HCl of pH 8.0, containing 250 mM NaCl, 5 mM
-mercaptoethanol ( -ME), 1 mM PMSF, and protease inhibitors], and
subsequently washed with 10–20 column volumes of the binding buffer,
followed by 10 column volumes of a washing buffer [50 mM Tris-HCl of
pH 8.0, containing 250 mM NaCl, 5 mM -ME, 1 mM PMSF, and protease
inhibitors, plus 20 mM imidazole]. The hexahistidine-tagged bromodomain
protein is eluted from the column in 1.0-ml fractions with an elution buffer
with an 20–500-mM imidazole gradient in 50 mM Tris-HCl of pH 8.0, con-
taining 250 mM NaCl, 5 mM -ME, 1 mM PMSF, and protease inhibitors.
Fractions containing the pure protein are pooled and dialyzed to a throm-
bin cleavage buffer [50 mM Tris-HCl of pH 8.0, containing 250 mM NaCl,
and 5 mM -ME]. The hexahistidine tag in the recombinant protein is re-
moved with thrombin treatment (1 unit thrombin/mg protein) overnight

at 4 . The thrombin cleavage reaction is stopped by addition of 1 mM
PMSF. The protein sample is concentrated and applied to a size exclusion
chromatography column for further purification using an FPLC system in a
phosphate buffer [100 mM phosphate of pH 6.5, containing 0.5 mM EDTA
and 1 mM DTT]. Peak fractions are collected, concentrated to 0.5 mM,
and dialyzed to the final NMR buffer in H2O/2H2O (9:1) consisting of
100 mM phosphate, pH 6.5, containing 0.5 mM EDTA, 5 mM perdeuter-

ated DTT. Typically, 5–10 mg of pure bromodomain protein is obtained
from 1 L of cell culture.
Protein Structure Determination by NMR

Three-dimensional structure of a protein can be determined by using
heteronuclear multi-dimensional NMR methods.25 The heteronuclear
NMR methods separate the proton signals of a protein in the NMR spectra
according to chemical shifts of their attached heteronuclei (such as 15N and
13
C), thus minimizing signal overlapping problems in the protein spectra.
Also, NMR resonance assignments of the protein can be obtained in a
sequence-specific manner, which assures the accuracy of data analysis
for high-resolution structure determination. In addition, because of favor-
able 1H, 13C, and 15N relaxation rates caused by partial deuteration of
the protein, factional deuteration in combination with 13C and 15N-labeling
is often employed for protein structure determination by NMR.26,27 All
the NMR data are processed and analyzed using NMR software programs
of NMRPipe28 and NMRView.29 For sequential backbone and side chain
assignments and structure determination of the protein, a set of NMR
experiments is described briefly later.
Backbone Assignments. Sequence-specific backbone assignment is
achieved using a suite of deuterium-decoupled 3D NMR experiments that
include HNCA, HN(CO)CA, HN(CA)CB, HN(COCA)CB, HNCO, and
HN(CA)CO experiments.30 Using the triple-labeled (75% 2H, 13C, and
15
N) protein sample, we perform constant-time experiments to gain higher
digital resolution and use a water flip-back scheme to minimize amide
signal attenuation from water exchange.
Side-Chain Assignments. Sequential side-chain assignments are
accomplished from a series of 3D NMR experiments with alternative
approaches to confirm the assignments. These experiments include 3D
15
N-edited TOCSY-HSQC, HCCH-TOCSY, (H)C(CO)NH-TOCSY, and
H(C)(CO)NH-TOCSY.25
25
G. M. Clore and A. M. Gronenborn, Meth. Enzymol. 239, 249 (1994).
26
M. Sattler and S. W. Fesik, Structure 4, 1245 (1996).
27
M.-M. Zhou, K. S. Ravichandran, E. T. Olejniczak, A. P. Petros, R. P. Meadows, M. Sattler,
J. E. Harlan, W. Wade, S. J. Burakoff, and S. W. Fesik, Nature 378, 584 (1995).
28
F. Delaglio, S. Grzesiek, G. W. Vuister, G. Zhu, J. Pfeifer, and A. Bax, J. Biomol. NMR 6,
277 (1995).
29
B. A. Johnson and R. A. Blevins, J. Biomol. NMR 4, 603 (1994).
30
T. Yamazaki, W. Lee, C. H. Arrowsmith, D. R. Mahandiram, and L. E. Kay, J. Am. Chem.
Soc. 116, 11655 (1994).
NOE Analysis/Distance Restraints. Distance restraints are obtained

from analysis of 15N- and 13C-edited 3D NOESY data, which are collected
with different mixing times to minimize spin diffusion problems. The
nuclear Overhauser effect (NOE)-derived restraints are categorized as
strong (1.8–3 Å), medium (1.8–4 Å), or weak (1.8–5 Å) based on the ob-
served NOE intensities. We also employ the recently developed ARIA
program31 that is integrated with X-PLOR for the iterative automated
NOE analysis. ARIA-assigned NOE peaks are manually checked and con-
firmed to ensure the success of ARIA/X-PLOR-assisted NOE analysis and
structure calculations.
Slow Exchange Amides. Amide protons involved in hydrogen bonds are
identified from an analysis of the amide exchange rates measured from a
series of 2D 1H/15N-heteronuclear single quantum coherence (HSQC)
spectra recorded after adding 2H2O to the protein sample.
Stereospecific Methyl Groups. Stereospecific assignments of methyl
groups of Val and Leu residues are obtained from an analysis of carbon
signal multiplet splitting using 10% 13C-labeled protein sample, which
can be readily prepared using 10% 13C-glucose containing M9 minimal
medium.32
Dihedral Angle Restraints. angle constraints are generated from the
3
JHNH coupling constants measured in a 3D HNHA-J experiment.33
Stereospecific assignments of -methylene protons, which give information
on 1, angles can be obtained from HNHB34 and 15N-edited TOCSY with
a short mixing time.35
Structure Calculations and Refinements. Structures of the protein are
calculated using a distance geometry/simulated annealing protocol with
the X-PLOR program.36–38 The structure calculations employ inter-proton
distance restraints obtained from 15N- and 13C-resolved NOESY spectra of
the protein or protein/peptide complex. The initial structures are typically
calculated with only manually assigned NOE-derived distance restrains,
which are used to assist further NOE assignments and identify hydrogen
bond partners for the slow exchange amide protons. The converged struc-
tures are refined with the experimental restraints of dihedral angles and
31
M. Nilges and S. O’Donoghue, Prog. NMR Spectrosc. 32, 107 (1998).
32
D. Neri, T. Szyperski, G. Otting, H. Senn, and K. Wüthrich, Biochemistry 28, 7510 (1989).
33
G. Vuister and A. Bax, J. Am. Chem. Soc. 115, 7772 (1993).
34
J. C. Matson, O. W. Sörensen, P. Söresen, and F. M. Poulsen, J. Biomol. NMR 3, 239 (1993).
35
G. M. Clore, A. Bax, and A. M. Gronenborn, J. Biomol. NMR 1, 13 (1991).
36
A. T. Brunger, ‘‘X-PLOR Version 3.1: A System for X-Ray Crystallography and NMR.’’
Yale University Press, New Haven, CT, 1993.
37
J. Kuszewski, M. Nilges, and A. T. Brunger, J. Biolmol. NMR 2, 33 (1992).
38
M. Nilges, G. M. Clore, and A. M. Gronenborn, FEBS Lett. 229, 317 (1988).
hydrogen bonds. Final structure calculations employ the manual and the
ARIA-assisted NOE distance restraints, together with hydrogen bond dis-
tance restraints and dihedral angle restraints. The distance restraint force
constant used in the calculations is typically 50 kcal/mol/Å2, and no NOE
is violated by more than 0.3 Å. The torsion restraint force constant is
200 kcal/mol/rad2, and no dihedral angle restraint is violated by more than

5 . Only the covalent geometry terms, NOE, torsion, and repulsive van der
Waals terms are used in the structure refinement. A large, and negative
Lennard-Jones potential energy should be observed for the final structures,
indicating good non-bonded geometry of the structure. Procheck39 analysis
is also performed to show the majority of the protein residues that are in
preferred and allowed regions of the Ramachandran map. Finally, struc-
tures of proteins determined using NOE-derived distance restraints and
dihedral angle restraints can be further refined with use of residual dipolar
couplings, which can be measured in bicelle-based liquid crystalline or
cross-linked polyacrylamide gel medium and implemented in the final
refinement stage of the structure calculations.40–43
The Bromodomain Structure

The 3D structure of a prototypical bromodomain from the transcrip-
tional coactivator PCAF determined by NMR shows that the bromodo-
main adopts an atypical left-handed four-helix bundle (helices Z, A,
20
B, and C) (Fig. 1A). A long intervening loop between helices Z and
A (termed the ZA loop) is packed against the loop connecting helices B
and C (named the BC loop) to form a surface accessible hydrophobic
pocket, which is located at one end of the four-helix bundle, opposite the
amino and carboxy termini of the protein. Mutagenesis studies suggest that
tertiary contacts among the hydrophobic and aromatic residues between
the two inter-helical loops contribute directly to the structural stability of
the protein.20 This unique structural fold is highly conserved in the bromo-
domain family, as supported by several more recently determined struc-
tures of bromodomains from human GCN544 and S. cerevisiae GCN5p45
39
R. A. Laskowski, J. A. Rullmannn, M. W. MacArthur, R. Kaptein, and J. M. Thornton,
J. Biomol. NMR 8, 477 (1996).
40
J. H. Prestegard, Nat. Struct. Biol. 5 Suppl., 517 (1998).
41
J. J. Chou, S. Li, C. B. Klee, and A. Bax, Nat. Struct. Biol. 8, 990 (2001).
42
S. Cavagnero, H. J. Dyson, and P. E. Wright, J. Biomol. NMR 13, 387 (1999).
43
J. J. Chou, S. Gaemers, B. Howder, J. M. Louis, and A. Bax, J. Biomol. NMR 21, 377 (2001).
44
B. P. Hudson, M. A. Martinez-Yamout, H. J. Dyson, and P. E. Wright, J. Mol. Biol. 304,
355 (2000).
45
D. J. Owen, P. Ornaghi, J. C. Yang, N. Lowe, P. R. Evans, P. Ballario, D. Neuhaus,
P. Eiletici, and A. A. Travers, EMBO J. 19, 6141 (2000).
Fig. 1. Ligand binding of PCAF bromodomain. (A) Three-dimensional NMR structure

of the PCAF bromodomain. (B) Superimposed region of the 2D 15N-HSQC spectra of the
bromodomain ( 0.5 mM) in its free form (dark) and complexed to a histone H4 peptide
containing acetylated lysine 8 (SGRGKGG-AcK-GLGK, where AcK is acetyl-lysine) (molar
ratio 1:6) (light). The movement of protein resonances upon ligand binding is indicated by
arrows connecting from the free to the ligand-bound forms. (C) Ribbon and dotted-surface
diagram of the bromodomain depicting the location of the lysine-acetylated H4 peptide-
binding site. Bromodomain residues that exhibit major chemical shift changes of the backbone
amide 1H and 15N resonances upon binding to the AcK histone H4 peptide as observed in the
2D 15N-HSQC spectra are indicated in lighter color. Note that most of these perturbed
residues are located in the ZA and BC loops. [From C. Dhalluin, J. E. Carlson, L. Zeng, C. He,
A. K. Aggarwal, and M.-M. Zhou, Nature 399, 491 (1999).]
as well as the double bromodomain module of human TAFII250.46 The

structural similarity among these bromodomains is very high for the four
helices with pairwise root-mean-square deviations of 0.7–1.8 Å for the
backbone C atoms. The majority of structural deviations are localized in
the loop regions, particularly in the ZA and BC loops. This observation
is in an agreement with relatively high sequence variations in these loops.4
The modular structure supports the notion that bromodomains act as a
functional unit for protein interactions, and multiple bromodomain
modules can be placed sequentially in a protein to serve similar or distinct
functions.4,20
Ligand Binding Study by NMR

The unique advantage of protein structural analysis by NMR is that in
addition to determination of a 3D structure of a protein, resonance assign-
ments obtained in the structural study provide a map of the entire protein
at atomic details-level, which could be used in biochemical analysis of pro-
tein-ligand interactions. Because the NMR resonances of protein residues
are highly sensitive to local chemical and conformational changes, binding
of a ligand to a host protein could be detected by resonance perturbations
of protein residues directly or indirectly involved in interactions with the
ligand.47,48 For highly specific interactions between a host protein and a
ligand, protein resonances change as a function of ligand concentration
until complete saturation of the protein by the ligand. NMR titration of
protein and ligand binding, therefore, can be used to determine binding af-
finity (dissociation constant, KD) of the complex. The ligand-binding study
is most conveniently performed by using 2D 15N-HSQC spectra that record
backbone amide proton and nitrogen resonances of a uniformly 15N-
labeled protein. Because the ligand is typically not 15N-labeled thus invis-
ible in 15N-HSQC spectra, protein resonance changes in the spectra are
relatively easy to monitor. With the resonance assignment of a protein,
using this NMR method, one can identify the location of the ligand-binding
sites on the protein. The NMR titration study of the PCAF bromodomain
shows that the protein can bind to acetyl-lysine–containing peptides
derived from major acetylation sites on histones H3 and H4 in a highly
specific manner, and the interaction is dependent on acetylation of
lysine (Fig. 1B and C).20
46
R. H. Jacobson, A. G. Ladurner, D. S. King, and R. Tjian, Science 288, 1422 (2000).
47
P. J. Hajduk, R. P. Measdows, and S. W. Fesik, Q. Rev. Biophys. 32, 211 (1999).
48
J. M. Moore, Curr. Opin. Biotech. 10, 54 (1999).
This unique ability of NMR spectroscopy to detect relatively weak but

highly specific interactions between a host protein and a ligand is also dem-
onstrated by our recent study of the bromodomain of transcriptional coac-
tivator CBP (CREB-binding protein) and the tumor suppressor protein
p53. It is known that transcriptional activity of p53 in cell cycle arrest, se-
nescence or apoptosis is tightly controlled by acetylation of its C-terminal
lysines, that is, K320 by PCAF,49 and K373 and K382 by CBP,50 which
results in its association with transcriptional coactivators including
CBP.51,52 Our study shows that p53/CBP association involves CBP bromo-
domain binding to p53 at the acetylated lysine 382 but not K373 or K320,
which can be demonstrated by NMR titration with lysine-acetylated p53
peptides containing different acetylation sites (Mujtaba and Zhou, unpub-
lished results). Notably, amino acid sequences of the acetylated K373 and
K382 peptides used in the NMR-binding study are identical except that the
acetylated lysine is in a different position. Remarkably, only the latter p53
peptide causes ligand concentration–dependent resonance perturbations of
protein residues in 2D 15N-HSQC spectra of the bromodomain, underscor-
ing the highly selective nature of the CBP bromodomain/p53 AcK382 rec-
ognition. None of the p53 peptides showed any detectable binding to the
structurally homologous bromodomain from PCAF.20,53
To understand the detailed structural basis of molecular recognition of
acetyl-lysine–containing peptide, one can determine the 3D structure of a
bromodomain in complex with a synthetic, lysine-acetylated peptide de-
rived from the known binding site in its biological binding partner protein.
Intermolecular NOE distance restraints required for structure determin-
ation of the complex can be obtained from a 13C-edited (F1) and 15N, and
13
C-filtered (F3) 3D NOESY spectrum collected for a sample containing
isotope-labeled protein and unlabeled ligand.54 In addition, a 2D 13C/15N-
filtered 1H-1H NOESY spectrum can provide intra-molecular NOEs for
the peptide molecule bound to the protein. Structure calculations of the
complex using both inter- and intra-molecular NOE distance restraints to-
gether with other experimentally determined hydrogen bond distance
restraints and dihedral angle restrains are similar to those described earlier
49
L. Liu, D. M. Scolnick, R. C. Trievel, H. B. Zhang, R. Marmorstein, T. D. Halazonetis, and
S. L. Berger, Mol. Cell Biol. 19, 1202 (1999).
50
W. Gu and R. G. Roeder, Cell 90, 595 (1997).
51
N. A. Barlev, L. Liu, N. H. Chehab, K. Mansfield, K. G. Harris, T. D. Halazonetis, and S. L.
Berger, Mol. Cell 8, 1243 (2001).
52
C. Prives and J. L. Manley, Cell 107, 815 (2001).
53
S. Mujtaba, Y. He, L. Zeng, A. Farooq, J. E. Carlson, M. Ott, E. Verdin, and M.-M. Zhou,
Mol. Cell 9, 575 (2002).
54
M. Sattler, J. Schleucher, and C. Griesinger, Prog. NMR Spectrosc. 34, 93 (1999).
for the structure determination of the protein alone using XPLOR

program.
Protein–Peptide Binding Assays

Recent structural studies of different bromodomains show that bromo-
domains share a conserved left-handed four-helix bundle fold, and acetyl-
lysine–containing peptides are bound between the ZA and BC loops.
Residues important for acetyl-lysine recognition in different bromodo-
mains are conserved; however, ligand selectivity differ due to a few but im-
portant differences in bromodomain sequences. These include variations in
the ZA loops, which have relatively low sequence conservation and amino
acid deletion or insertion in different bromodomains; and differences in
bromodomain residues that directly interact with residues surrounding
acetyl-lysine in a target protein. Structure-based mutational analysis can
be employed to determine the structural basis of molecular recognition of
a bromodomain/ligand complex. For example, in an effort to determine the
molecular determinants of the selective recognition between the PCAF
bromodomain and HIV-1 Tat at the acetylated lysine 50, we performed
structure-based mutational analysis of the complex in an in vitro–binding
assay using the recombinant and purified GST-fusion bromodomain and
an N-terminal biotinylated, p53 peptide containing lysine-acetylated K382
that is immobilized onto Streptavidin agarose beads.53
The cDNA encoding GST-fusion bromodomain of PCAF is cloned into
the pGEX4T-3 vector (Amersham Biosciences). The GST bromodomain is
expressed in E. coli BL21(DE3) codon-plus cells using a procedure similar
to that described earlier for NMR protein sample preparation, except that
the GST protein is prepared in LB medium. The harvested bacterial cells
culture is re-suspended in lysis buffer [20 mM Tris of pH 8.0, 150 mM
NaCl, 1.0% NP-40, 1 mM PMSF, 5 mM -ME, 10% glycerol, 5 mM
EDTA, 1 mM PMSF, and DNAse]. After cell lysis by sonication and re-
moval of cell debris by centrifugation, the supernatant is purified by affinity
chromatography using a glutathione Sepharose resins according to manu-
facturer’s (Amersham Bioscience) instruction. The binding assay is per-
formed by incubating an equal amount of the PCAF bromodomain with

biotinylated HIV Tat peptide for 2 h at 22 in the binding buffer (50 mM
Tris buffer of pH 7.5, containing 50 mM NaCl, 0.1% BSA, and 1 mM
DTT). The protein-peptide complex is pulled down by 10 l of Streptavi-
din agarose (Novagen) by further incubation of 30 min. To minimize
non-specific binding, the beads are washed extensively (thrice) with a high
salt- and detergent-containing washing buffer (50 mM Tris buffer of pH
7.5, containing 300 mM NaCl, 0.1% NP-40, and 1 mM DTT). Proteins
eluted from the agarose beads are separated by SDS-PAGE and protein/
peptide interaction is visualized by western blotting using anti-GST anti-
body (Sigma) and horseradish-peroxidase–conjugated goat anti-rabbit IgG.
Such a GST pull-down assay is also used to assess other bromodomains’
binding or effect of site-directed mutation of PCAF bromodomain residues
on protein binding to the HIV-1 Tat peptide. Furthermore, this binding
assay can be used for mutational analysis of Tat peptide residues in a
peptide competition assay, in which a non-biotinylated peptide carrying a
mutation at a specific amino acid residue competes with the biotinylated
wild-type Tat AcK50 peptide for binding to the GST-fusion PCAF bromo-
domain. The molar ratio of the mutant and wild-type peptides in the mix-
ture is kept at 1:2. Because of the high sensitivity of western blotting
detection, this binding study can be performed at a protein concentration
(10 M) much lower than that required for NMR study (200 M), thus
ensuring specificity of protein-peptide interactions. Using this GST pull-
down–based assay, mutational analyses of the protein and peptide residues
validate the molecular interactions observed in the NMR structure of
the complex, and identify key residues of the PCAF bromodomain that
are important for recognition of the acetyl-lysine and its flanking residues
in the HIV-1 Tat peptide.
Perspective
The NMR-based methods described here have enabled establishment
of the biochemical functions of bromodomains as acetyl-lysine–binding
domains.20,44,53 This new mechanism of regulating protein-protein inter-
actions via lysine acetylation has broad implications in a wide variety
of cellular processes, including chromatin remodeling and transcriptional
activation.21–23,53,55 Such a powerful NMR structure-based approach is
readily applicable to investigate the biological functions of the other evolu-
tionarily conserved protein modular domains that play an important role in
regulation of chromatin remodeling.
Acknowledgment
The work was supported by a grant from the National Institutes of Health to M.-M. Z.
(CA87658).
55
A. Dorr, V. Kiermer, A. Pedal, H.-R. Rackwitz, P. Henklein, U. Schubert, M.-M. Zhou,
E. Verdin, and M. Ott, EMBO J. 21, 2715 (2002).
[9] interaction of the chromodomain with histone peptides 131
[9] Assays for the Determination of Structure

and Dynamics of the Interaction of the
Chromodomain with Histone Peptides
By Steven A. Jacobs, Wolfgang Fischle, and
Sepideh Khorasanizadeh
The histone code hypothesis predicts that multiple histone modifica-

tions, acting in a combinatorial or sequential fashion on one or multiple his-
tone tails, specify unique downstream functions.1 In order for this process
to come to fruition, there must exist enzymatic protein modules which
write the histone code by modifying specific residues of the histone tails,
and protein recognition modules which read the code by binding to specific
modifications. Indeed this appears to be the case, as numerous protein
modules which enzymatically modify histone tails have been identified.
Histone acetyltransferase domains (HAT) acetylate lysines.2,3 Histone
deacetylase domains (HDACs) remove acetyl marks from lysines,4 SET
domains methylate specific lysines,5 histone arginine methyltransferases
methylate arginine residues, and kinases phosphorylate serine and threo-
nine residues. Recognition protein modules discovered to date include
the chromodomain which interacts with methylated lysines6–8 and the bro-
modomain which interacts with acetylated lysines; for review, see Marmor-
stein and Berger (2001).9 To study the specificities and affinities of the
interactions of chromodomains with histone tails we have successfully ap-
plied biophysical methods, including nuclear magnetic resonance (NMR)
spectroscopy, X-ray crystallography, fluorescence spectroscopy, and
isothermal titration calorimetry (ITC), all of which are described here.
NMR spectroscopy is best used to quickly screen peptides for interaction
with a specific protein module and determine the extent of a minimally struc-
tured complex. NMR or X-ray crystallography can subsequently be used to
1
2
S. Y. Roth, J. M. Denu, and C. D. Allis, Annu. Rev. Biochem. 70, 81 (2001).
3
R. Marmorstein and S. Y. Roth, Curr. Opin. Genet Dev. 11, 155 (2001).
4
R. Marmorstein, Structure (Camb) 9, 1127 (2001).
5
M. Lachner and T. Jenuwein, Curr. Opin. Cell Biol. 14, 286 (2002).
6
M. Lachner, D. O’Carroll, S. Rea, K. Mechtler, and T. Jenuwein, Nature 410, 116 (2001).
7
A. J. Bannister, P. Zegerman, J. F. Partridge, E. A. Miska, J. O. Thomas, R. C. Allshire, and
T. Kouzarides, Nature 410, 120 (2001).
8
S. A. Jacobs, S. D. Taverna, Y. Zhang, S. D. Briggs, J. Li, J. C. Eissenberg, C. D. Allis, and
S. Khorasanizadeh, EMBO J. 20, 5232 (2001).
9
R. Marmorstein and S. L. Berger, Gene 272, 1 (2001).

solve the structure of a complex formed between a protein module and

a target peptide in order to gain the details of interaction at atomic
resolution. Binding affinities and thermodynamics of interaction are best
determined by ITC and/or fluorescence spectroscopy.
NMR Spectroscopy
In addition to the determination of high-resolution structure, this
technique can be used to rapidly screen for specific protein-protein,
protein-peptide, or protein-small molecule interactions, map the protein
interaction surface, and calculate the binding affinity. All of these applica-
tions have been used successfully to study the interactions of chromodo-
main and bromodomain chromatin binding modules with histone tails.
For example, NMR has been used to determine the specificity of binding of
the HP1 chromodomain to methylated H3 tail peptides and map the surface
of interaction.8 It has also been used to solve the high-resolution structure of
the HP1 chromodomain bound to a methylated histone H3 tail.10
NMR techniques do have limitations. First of all, the protein module
must be able to be expressed in milligram quantities in a few liters of min-
imal media in order to incorporate 15N and/or 13C labeling at an efficient
cost (every liter of minimal media requires 1 g of 15N ammonium chloride
at $30/g and 4 g of 13C glucose at $200/g). In addition, the protein must be
soluble and stable to >200 M in a buffer suitable for NMR measurements
at room temperature (samples of 50–100 M may be studies when using a
state-of-the-art cryoprobe in the NMR spectrometer). Phosphate buffer in
the pH range 6–7.5 is ideal to ensure that backbone amide hydrogens are
fully occupied for maximum sensitivity of detection.
To screen for protein-peptide interactions by NMR spectroscopy, the
main tool is the two-dimensional [1H-15N]-HSQC experiment. The spec-
trum resulting from this experiment correlates the 1H and 15N chemical
shifts, thus providing a backbone fingerprint of the protein. The value of
the chemical shift of each residue is highly dependent on the local chemical
environment of that residue. Thus, addition of a protein or peptide which
specifically binds to the target protein produces a new spectrum with a de-
fined number of large chemical shift perturbations. For example, the ability
of protein modules to bind modified histone tail peptides can be monitored
by titrating an unlabeled peptide into a 15N-labeled protein solution, and
collecting the [1H-15N]-HSQC spectrum of the protein. Figure 1A shows
the [1H-15N]-HSQC spectra of the free chromodomain (black cross peaks)
10
P. R. Nielsen, D. Nietlispach, H. R. Mott, J. Callaghan, A. Bannister, T. Kouzarides, A. G.
Murzin, N. V. Murzina, and E. D. Laue, Nature 416, 103 (2002).
Fig. 1. NMR analysis of the HP1 chromodomain interaction with the dimethyllysine 9
histone H3 peptide. (A) The [1H-15N]-HSQC spectrum of the free chromodomain (black cross
peaks) and after the addition of the dimethyllysine 9 histone H3 peptide (grey cross peaks).
Each dashed line corresponds to the perturbation of a residue. (B) Chemical shift
perturbation (ave) upon H3 peptide binding for the residues of the chromodomain.
Fig. 2. Chromodomain backbone dynamics by [1H]-15N NOE measurements. Black

triangles represent the free chromodomain while open circles represent the chromodomain-
H3 peptide complex. The residues with [1H]-15N NOE value >0.55 are expected to be rigid
and structured (above the dashed line).
and after addition of an H3 tail peptide (grey cross peaks) containing a di-
methyllysine at residue 9. The large chemical shift perturbations are indica-
tive of formation of a specific protein-peptide complex. In contrast, an
unmethylated peptide or a peptide methylated at lysine 4 produces a few
small chemical shift changes, indicating that these peptides do not specific-
ally interact with the HP1 chromodomain.8 NMR was also used to map the
surface of the chromodomain which interacts with the H3 tail.8 Figure 1B
shows that a few key residues exhibit the greatest change in chemical shift,
implicating them for direct peptide binding.
X-ray crystallography or NMR spectroscopy of a protein-peptide com-
plex is greatly simplified if only minimally structured domains are studied.
Backbone dynamics determined by NMR can establish which segments of
the protein are ordered into a defined structure. Figure 2 shows a plot of
HP1 chromodomain [1H]-15N NOE values for the free protein (filled tri-
angles) and after complexation with a histone H3 peptide dimethylated
at lysine 9 (open circles). Only residues 25–74 of the chromodomain are
rigid in both the free and bound form as judged by the [1H]-15N NOE value
>0.55. This information was then used to prepare a new compact construct,
which dramatically improved growth of suitable crystals for high-resolution
structure determination by X-ray crystallography.11
11
S. A. Jacobs and S. Khorasanizadeh, Science 295, 2080 (2002).
Protein Preparation
The coding regions of Drosophila HP1 residues 1–84 and 17–76 were
sub-cloned into NdeI-BamHI sites of pET16b vector (Novagen) and ex-
pressed in BL21(DE3) Escherichia coli (Stratagene) with an N-terminal
His-tag. Isotopic labeling of the protein (15N/13C) was achieved by grow-
ing these cells in minimal media containing 7 g Na2HPO4, 3 g KH2PO4,
0.5 g NaCl, 0.1 g yeast extract, 4 g 13C-glucose, 1 g 15N-NH4Cl, 50 mg
thiamine-HCl, 0.5 g MgSO4, and 20 mg CaCl2 per liter of water. Cultures
were grown at 37 to an optical density (at 600 nm) of 0.6, and protein ex-
pression was induced by adding isopropyl--d-thiogalactopyranoside
(IPTG) to a concentration of 1 mM. Temperature was reduced and shaking

continued for 4 h at 30 until cells were harvested by centrifugation, and

the pellet was stored in 20 overnight. The frozen pellet (10 g) was resus-
pended in 50 ml of 50 mM potassium phosphate, pH 7.5, 10 mM imidazole,
500 mM NaCl, 5% glycerol, 1 mM benzamidine HCl, 1 mM phenylmethyl-
sulfonylfluoride (PMSF), and the suspension was stirred in the cold room
for 30 min. This suspension was then sonicated on ice with 30 pulses each
30 s using a Misonex sonicator XL. The crude lysate was then cleared by
centrifugation at 35,000 g for 30 min and the supernatant loaded onto a
column containing 3 ml of Niþ2-agarose affinity resin (Qiagen). The column
was pre-equilibrated with 30 ml of the same buffer. After all supernatant
was loaded, the column was washed overnight with 1 L of 50 mM
potassium phosphate, pH 7.5, 10 mM imidazole, 500 mM NaCl, and 5%
glycerol. Bound material was eluted from the column in 1-ml fractions with
50 mM potassium phosphate, pH 7.5, 250 mM imidazole, 100 mM NaCl,
and 5% glycerol, and fractions were analyzed for protein content and
purity by SDS-PAGE and absorption spectroscopy at 280 nm. Protein frac-
tions were pooled and concentrated using a Millipore ultrafree filtration
device and further purified by gel filtration chromatography using a Super-
dex 75 column (Pharmacia) equilibrated with 50 mM potassium phosphate,
pH 6.0, 25 mM NaCl, 1 mM sodium azide, and 0.1 mM DTT. Finally,
column fractions containing HP1 were pooled and the sample concentrated
to 1 mM for NMR studies. The HP1 chromodomain was prepared using
this protocol for all of the assays described later.
Peptide Preparation
Histone tail peptides were prepared synthetically and purified by
reverse-phase HPLC. A nonnative tyrosine residue is placed at the C-
terminus of each histone peptide to allow determination of concentration
by UV absorbance at 280 nm using the extinction coefficient e280 ¼ 1280/
M/cm. All peptides were further purified in order to remove trace organic
solvent by gel filtration chromatography. About 5 mg of the peptide was

dissolved in 100 l of distilled water and passed over a spin column con-
taining 3 ml of Sephadex G10 resin (Pharmacia) equilibrated with distilled
water. The pure peptides were then lyophilized for storage. This protocol
for peptide preparation was used for all assays described later.
Protocol for Determination of the Specificity of the HP1

Chromodomain for K9 Methylated H3 Tail
All data were collected on a Bruker AVANCE (600 MHz) spectrometer
with samples containing 0.2–1 mM HP1 1–84 in a buffer solution of 50 mM
sodium phosphate, pH 6.0, 25 mM NaCl, 1 mM sodium azide, and

0.1 mM DTT at 25 . Spectra were processed with NMRPipe12 and analyzed
with NMRView. First, the [1H-15N]-HSQC spectrum of the free chromo-
13
domain was collected. Next, a purified and dried peptide sample was resus-
pended in the 15N-labeled chromodomain NMR sample in order to achieve
molar ratio of 1:4 protein to peptide. The [1H-15N]-HSQC spectrum of this
complex sample was collected, and compared with that of the free chromo-
domain in order to identify the sites with chemical shift changes associated
with residues on the surface of peptide binding (see Fig. 1A).
Protocol for Mapping the H3 Tail Binding Surface of the Chromodomain

The first step in mapping the interaction surface is to assign the reso-
nance of the protein in the free form as well as in complex with the peptide.
To accomplish this, resonance assignments are carried out (using 15N/13C-
labeled protein) by collecting and analyzing the three-dimensional NMR
spectra HNCA, HNCO, HN(CO)CA, 15N-edited NOESY, and 15N-edited
TOCSY for the free and complex samples.14,15 Eq. (1) was used to tabulate
the weighted average of the HN, N, C, and C0 chemical shift perturbations
upon binding of the methylated H3 tail to HP1:16
( 2 2 )12
2 N c 2 c0
ave ¼ 0:25 ðHN Þ þ þ þ : (1)
5 2 2
12
F. Delaglio, S. Grzesiek, G. W. Vuister, G. Zhu, J. Pfeifer, and A. Bax, J. Biomol. NMR 6,
277 (1995).
13
B. A. Johnson and R. A. Blevins, J. Biomol. NMR 4, 603 (1994).
14
S. Grzesiek and A. Bax, J. Magn. Reson. 96, 432 (1992).
15
D. Marion, P. C. Driscoll, L. E. Kay, P. T. Wingfield, A. Bax, A. M. Gronenborn, and G. M.
Clore, Biochemistry 28, 6150 (1989).
16
S. Grzesiek, S. J. Stahl, P. T. Wingfield, and A. Bax, Biochemistry 35, 10256 (1996).
The residues which experience the greatest change (ave > 0.15) were
then mapped onto the chromodomain surface.
Protocol for Determination of Backbone Dynamics to Establish

a Minimal Interacting Domain of HP1 with Methylated H3
Tail by NMR
The [1H]-15N steady-state heteronuclear NOE values for the backbone
nuclei of the free and complexed chromodomain samples were measured
by collecting a modified two-dimensional [1H-15N]-HSQC NMR spectrum
with the water flip-back method.17 Spectra were collected in an interleaved
manner, one with a recovery delay of 6 s (reference spectrum) and the
other with a recovery delay of 3 s of proton saturation (NOE spectrum).
The [1H]-15N NOE value for each residue was calculated as the ratio of
the cross-peak intensity NOE/reference, and error was estimated from
the baseline noise in the two spectra.
Structure Determination of Protein-Histone Tail Complexes

Several NMR and X-ray crystal structures of protein modules interact-
ing with modified histone tails have been solved.10,11,18–21 These structures
provide detailed views of the chemistry of acetyllysine and methyllysine
recognition, as well as an explanation for the specificity of module/histone
tail interactions. Figure 3 shows the structure of the HP1 chromodomain
bound to a methylated H3 tail peptide at atomic resolution. The
methyl ammonium group of the methyllysine is recognized by a three-
member aromatic cage of the chromodomain. The binding is stabilized by
cation interaction between the electrons of the aromatic rings, and the
positive charge of the methyllysine.11 Increasing methyl groups on the
lysine side chain reduces the number of hydrogens available for hydrogen
bonding. In the binding pocket, the dimethyllysine hydrogen bonds to one
water molecule, whereas trimethyllysine is not capable of hydrogen
bonding.
17
S. Grzesiek and A. Bax, J. Am. Chem. Soc. 115, 12593 (1993).
18
C. Dhalluin, J. E. Carlson, L. Zeng, C. He, A. K. Aggarwal, and M. M. Zhou, Nature 399,
491 (1999).
19
B. P. Hudson, M. A. Martinez-Yamout, H. J. Dyson, and P. E. Wright, J. Mol. Biol. 304,
355 (2000).
20
R. H. Jacobson, A. G. Ladurner, D. S. King, and R. Tjian, Science 288, 1422 (2000).
21
D. J. Owen, P. Ornaghi, J. C. Yang, N. Lowe, P. R. Evans, P. Ballario, D. Neuhaus,
P. Filetici, and A. A. Travers, EMBO J. 19, 6141 (2000).
Fig. 3. The methyllysine binding aromatic cage (in black) of the HP1 chromodomain
surrounding the H3 peptide methyllysine (in grey): (A) dimethyllysine and (B) trimethylly-
sine. Increasing methylation leads to a decrease in the potential to form hydrogen bonds with
a nearby ordered water molecule (in grey).
Protocol for Determination of the Structure of the HP1 Chromodomain

Complexed with a Methylated H3 Tail at Atomic Resolution
The purified HP1 fragment 17–76 was exchanged into a buffer of
10 mM potassium phosphate, pH 7.0, by gel filtration chromatography
using Superdex 75 resin (Pharmacia) as a final purification step before
concentrating to 1 mM. The purified H3 tail peptide residues 1–15 with
dimethyllysine or trimethyllysine at residue 9 was resuspended in the pro-
tein solution to reach a molar ratio of 1:4 protein to peptide. As the HP1
chromodomain had not been previously crystallized, a broad initial screen
of crystallization conditions was necessary. Several commercially available
sparse matrix and grid screens (Hampton Research, Emerald Biostruc-
tures) were employed in initial screening using the hanging drop vapor dif-
fusion method. In each case, 2 l of the protein solution was combined with
2 l of the reservoir solution and equilibrated against a 1-ml reservoir solu-
tion. Hexagonal crystals were obtained in various drops from an ammo-
nium sulfate grid screen (Hampton Research), with optimal crystals
growing when 3.2 M ammonium sulfate and 0.1 M MES buffer at pH 6.1
was used as the reservoir solution. Single crystals in space group C2221
(a ¼ 33 Å, b ¼ 76 Å, c ¼ 75 Å) were obtained. The crystals were cryo-
protected in reservoir solution supplemented with 35% MPD, and flash-
frozen in liquid nitrogen for data collection. Diffraction data were collected
on a Rigaku RUH2R X-ray generator equipped with a Bruker SMART
6000 CCD detector. All data were processed and scaled with HKL2000.22
The structure was solved by molecular replacement using CNS23 and
MOLREP24 using the crystal structure of one monomer of the dimeric
Swi6 chromo shadow domain25 with all residues turned to alanines as the
search model. Chain tracing and structure refinement was carried out using
ARP/wARP,26 CNS, and O.27
Fluorescence Spectroscopy
Macromolecular associations can be monitored by either the intrinsic
tryptophan fluorescence of a protein or the fluorescence anisotropy of an
extrinsic probe attached to one of the interacting species. The use of intrin-
sic fluorescence is somewhat limited since not all macromolecular associ-
ations produce a change in intrinsic fluorescence signal. However,
intrinsic fluorescence does have an advantage in that there are no external
probes introduced to the system. Changes in fluorescence intensity upon
ligand titration can be fit to the following equation in order to attain
dissociation constants:28
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Fs ð½XT þ ½YT þ Kd Þ ð½XT þ ½YT þ KD Þ2 ð4½XT ½YT Þ
Fi ¼ ;
2½XT
(2)
where Fi is the fluorescence change produced by fitting of increasing quan-
tities of peptide, Fs the fluorescence change at saturation, [XT] the total
22
Z. Otwinowski and W. Minor, Methods Enzymol. 276, 307 (1997).
23
A. T. Brunger, P. D. Adams, G. M. Clore, W. L. DeLano, P. Gros, R. W. Grosse-Kunstleve,
J. S. Jiang, J. Kuszewski, M. Nilges, N. S. Pannu, R. J. Read, L. M. Rice, T. Simonson, and
G. L. Warren, Acta Crystallogr. D. Biol. Crystallogr. 54, 905 (1998).
24
A. Vagin and A. Teplyakov, J. Appl. Crystallogr. 80, 1022 (1997).
25
N. P. Cowieson, J. F. Partridge, R. C. Allshire, and P. J. McLaughlin, Curr. Biol. 10,
517 (2000).
26
A. Perrakis, M. Harkiolaki, K. S. Wilson, and V. S. Lamzin, Acta Crystallogr. D. Biol.
Crystallogr. 57, 1445 (2001).
27
T. A. Jones and M. Kjeldgaard, ‘‘O Version 5.9.’’ 1994.
28
K. Croce, S. J. Freedman, B. C. Furie, and B. Furie, Biochemistry 37, 16472 (1998).
Fig. 4. Fluorescence anisotropy binding data. Binding curves for the interaction of the
HP1 chromodomain (open circles) and SuVar(3-9) chromodomain (open squares) with
fluoresceinated-H3 peptide residues 1–15 trimethylated at lysine 9. The HP1 binding assay
with the unmodified H3 peptide is plotted with filled triangles. All binding measurements

were performed at 15 in 50 mM potassium phosphate, pH 8.0, 25 mM NaCl, and 2 mM DTT.
concentration of protein, [YT] the total concentration of titrated peptide,

and KD is the dissociation constant. This technique has been used success-
fully to quantitate the binding of the HP1 chromodomain to a methylated
H3 tail peptide.10
Fluorescence anisotropy spectroscopy has become popular for studying
protein-peptide, protein-protein, and protein-DNA interactions due to the
use of relatively simple and inexpensive equipment, and the ability to pro-
vide rapid results. A series of fluorescence anisotropy experiments has
demonstrated that the chromodomains of HP1,8,11 Pdd1p and Pdd3p,29
and SuVar(3–9) (shown in Fig. 4) proteins selectively bind to H3 peptide
methylated at lysine 9. For mathematical and theoretical descriptions of
this technique, the reader is directed to more specialized references.30–32
Fluorescence anisotropy titrations are fit to the following equation to
obtain dissociation constants:

½X
A ¼ Af þ ðAb Af Þ ; (3)
KD þ ½X
29
S. D. Taverna, R. S. Coyne, and C. D. Allis, Cell 110, 701 (2002).
30
R. C. Cantor and P. R. Schimmel, in ‘‘Biophysical Chemistry. Part II: Techniques for the
Study of Biological Structure and Function,’’ p. 433. WH Freeman and Co., San Francisco,
1980.
31
D. M. Jameson and W. H. Sawyer, Methods Enzymol. 246, 283 (1995).
32
T. Heyduk, Y. Ma, H. Tang, and R. H. Ebright, Methods Enzymol. 274, 492 (1996).
where A is the measured anisotropy, Af the anisotropy of the free state, Ab

the anisotropy of the bound state, and [X] is the protein concentration. This
relationship is only true if the concentration of labeled species KD, and
operates under the assumption that the peptide can exist in only two states,
free and bound to the protein. A more complicated relationship is needed
if the system to be studied necessitates that these assumptions are not met.
Figure 4 provides examples of interactions which have been successfully
studied by fluorescence anisotropy. HP1 chromodomain interactions with
two H3 peptides are shown. The HP1 chromodomain does not interact with
the unmodified H3 peptide; instead there is strong binding to H3 peptide
with a trimethyllysine at residue 9. This technique can effectively be used
to rapidly examine predictions of binding properties based on sequence
homology. The surface of the HP1 chromodomain which interacts with
methylated H3 tail consists of residues that are conserved in SUV39H1
chromodomain, and thus we predicted that SUV39H1 also interacts with
methylated H3 tail strongly.11 Figure 4 shows that indeed the chromodo-
main of SUV39HI binds the H3 peptide with trimethylated lysine 9 with
an affinity of 25 M.
Practical Considerations
When using fluorescence anisotropy to study protein-protein inter-
actions, the smaller of the two interacting proteins should carry the fluores-
cent probe. This experimental design results in the greatest difference in
fluorescence anisotropy values between the free and bound states of the
fluorophore, and thus maximal experimental sensitivity. Ideally the larger
macromolecule is serially diluted to cover a concentration range of 100-fold
below to 100-fold above the KD of the interaction being studied. Most pro-
tein-histone tail interactions observed to date occur with an affinity of
106 M. Therefore, protein samples should range from approximately
0.1 M to 1 mM for initial studies. As outlined earlier, it is very important
to keep the concentration of the fluorescently labeled peptide at least
10-fold (preferably 100-fold) less than that of the KD. Failure to do so could
result in inaccurate affinity calculations. For protein-histone tail inter-
actions, a concentration of 100 nM has been found to be a good starting
point. Concentrations of both the labeled and unlabeled proteins can then
be fine tuned after initial experiments are conducted.
Instrumentation
We have used two different instruments (Beacon 2000 and Tecan Spec-
traFluor) for this type of measurement, and obtained comparable results.
The Beacon 2000 fluorescence polarization system by PanVera has been
used with much success in chromodomain/H3 tail interaction studies.8,11,29

This system is extremely easy to use and relatively inexpensive. The
Beacon 2000 system is capable of performing experiments at temperatures

from 6–65 , making it ideal for monitoring the affinity of interaction over
various temperatures. An alternative to the Beacon 2000 system is the
Tecan SpectraFluor instrument which has the advantage that it can read
samples in a 96-well plate format in a few minutes, therefore drastically re-
ducing measurement time. However, the Tecan instrument is not capable
of cooling below ambient temperature and is thus not suitable for evalu-
ation of binding at low temperatures. Both instruments have minimal
sample volumes of 100 l and comparable sensitivities of 1011 M
fluorescein. In addition, both instruments offer numerous wavelength
interference filters enabling many different fluorophores to be monitored.
Choice of Probes
Selection of fluorophores and coupling techniques are extensively de-
scribed elsewhere.33 Fluorescein is the most popular of the fluorescent
probes used in fluorescence anisotropy assays due to its high quantum yield
and long wavelengths of excitation and emission. The lifetime of the fluo-
rescein excited-state is on the order of 4 ns,30 making it possible to monitor
protein complexes of molecular weight <100 kDa. Fluorophores with
longer excited-state lifetimes should be selected in studies examining the
formation of complexes larger than 100 kDa. A wide range of conjugation
chemistries are available making it possible to conjugate fluorophores to
amines, thiols, alcohols, aldehydes, ketones, hydrazines, and carboxylic
acids.33 The fluorescent label should be placed at a position on the peptide
far enough away from the presumed interaction site to avoid interference
with binding. In our experience, placing the fluorescein label at four
residues away from binding segment of the peptide is sufficient to avoid
interference in binding affinity.
Protocol for N-Terminal Fluorescein Succinimidyl Ester Labeling

Synthetic H3 tail peptides were N-terminally labeled with fluorescein
using a kit from PanVera in 50-l reaction volumes. Each reaction con-
tained 100 g of peptide, 2 mM fluorescein succinimidyl ester, and
100 mM potassium phosphate, pH 7.0. The reaction was carried out at

37 for 30 min. After 30 min, 10 l of 1 M Tris-HCl, pH 8.0, was added
to quench the reaction. The fluoresceinated peptide was then separated
33
R. P. Haugland, ‘‘Handbook of Fluorescent Probes and Research Products.’’ Molecular
Probes Inc., Eugene, OR, 1996.
from the free fluorescein by gel filtration chromatography in a micro

column containing 0.5–1.0 ml of sephadex G10 resin (Pharmacia) equili-
brated with water. The labeled peptide (colored orange) elutes first from
the column and is easily distinguished from the free fluorescein (yellow).
In addition, the polarization of fractions from the gel filtration column

can be used to distinguish these species. We have found that at 25 , a
labeled peptide of 15–20 residues will have a polarization of 15 mP, while
the free fluorescein molecule shows a polarization of 20 mP.
Protocol for the Determination of the Binding Affinities of the

HP1 Chromodomain to Methylated H3 Tail Peptides by
Fluorescence Anisotropy
The purified HP1 fragment 17–76 was dialyzed into 50 mM potassium
phosphate, pH 8.0, 25 mM NaCl, and 2 mM DTT and concentrated to
500 M to 1 mM. Serial twofold dilutions of this concentrated protein
stock solution were then made to produce twelve 100-l samples with the
lowest concentration 0.1 M. The fluoresceinated H3 peptide was added
to the chromodomain samples to a final peptide concentration of 100 nM.
A second set of chromodomain samples with no peptide added acts as a
blank. Samples were centrifuged in order to avoid turbidity, which can

make anisotropy measurements inaccurate. Following equilibration to 15
in a water bath, the anisotropy of the samples was read in a Beacon 2000
fluorescence polarization system.
The beacon system reports polarization values; these were converted to
anisotropy using Eq. (4):
2 P
A¼ : (4)
3P
The data are then plotted as log chromodomain concentration versus an-
isotropy and fit to Eq. (3) by nonlinear least squares analysis using Kaleida-
graph (Synergy Software). Af, Ab, and KD are fit as multiple, floating
independent variables. For presentation purposes, anisotropy is converted
to fraction bound (Fb) using the relationship of Eq. (5):
A Af
Fb ¼ : (5)
Ab Af
Isothermal Titration Calorimetry

Isothermal titration calorimetry (ITC) is presently the only method
capable of providing a complete thermodynamic profile of a protein-
protein or protein-DNA interaction for interactions with binding affinities
on the order of 108 to 103 M. This method has been exploited to examine
the interaction of the HP1 chromodomain with the H3 tail methylated at
lysine 98,11 and the TAFII250 double bromodomain with acetylated histone
H4 tail peptides.20
Small amounts of heat are either generated or absorbed by every inter-
molecular interaction. The change in enthalpy (H), binding affinity (KD),
and stoichiometry (n) of a molecular interaction can be deduced from the
measurement of such heats of interaction. These values can then be ex-
ploited to derive the changes in entropy (S), free energy (G), and heat
capacity (Cp) of binding by the following relationships:
G ¼ nRTð ln KD Þ; (6)
G ¼ H TS; (7)
@H
Cp ¼ : (8)
T
The complete thermodynamic analysis of an interaction sets ITC apart
from all other methods. A full theoretical description of ITC has been
previously reviewed and should be sought out by interested readers.34–36
ITC analysis has been used to determine the full thermodynamic
parameters of the binding of the HP1 chromodomain to H3 tail peptides
containing either trimethyllysine or dimethyllysine at position 9.11
Figure 5 shows an example of the data obtained from such an experiment
where the peptide titrated into the chromodomain sample contained tri-
methyllysine at position 9. Table I summarizes the thermodynamics of
chromodomain interaction with H3 peptides containing dimethyllysine or
trimethyllysine at position 9 as measured by ITC.
The data clearly indicate that HP1 preferentially binds to the tri-
methyllysine containing peptide over the dimethyllysine peptide. The full
TABLE I
ITC Analysis of the Binding of the HP1 Chromodomain to Methylated H3 Peptides
H3 peptide KD (M) H (kcal/mol) N G (kcal/mol) TS (kcal/mol)
Me2K9 H3 6.9 0.2 9.47 0.04 1.02 6.78 0.01 2.69 0.05
Me3K9 H3 2.5 0.1 9.83 0.04 0.99 7.36 0.02 2.46 0.02
34
H. F. Fisher and N. Singh, Methods Enzymol. 259, 194 (1995).
35
L. Indyk and H. F. Fisher, Methods Enzymol. 295, 350 (1998).
36
R. L. Biltonen and N. Langerman, Methods Enzymol. 61, 287 (1979).
thermodynamic analysis by ITC aids in the analysis of this behavior. The

interaction with trimethyllysine is more enthalpically (H ¼ 0.4 kcal/
mol) and entropically [(TS) ¼ 0.2 kcal/mol] favorable than that with di-
methyllysine. This is due to improved cation- interactions and the van der
Waals contacts.11
Binding experiments such as the one depicted in Fig. 5 were conducted

at 16 and 25 in order to calculate the change in heat capacity upon
complex formation. Large negative changes in heat capacity upon binding
are assumed to be correlated with burial of nonpolar surfaces or large
Fig. 5. Binding of the HP1 chromodomain to a trimethylated H3 tail peptide as

measured by isothermal titration calorimetry. (A) Raw data showing the heats of injection.
(B) Integrated heats of injection with the solid line representing the fit of the data to Eq. (10).

Binding measurements were conducted at 25 in 50 mM potassium phosphate, pH 8.0, 25 mM
NaCl, and 2 mM DTT.
conformational changes upon complexation.37,38 The interaction of

the chromodomain with the H3 tail shows a small Cp of 0.17 kcal/
mol. This is consistent with lack of major conformational change in the
chromodomomain upon peptide binding.10,11
Practical Considerations
ITC measurements are extremely sensitive to differences in buffer con-
ditions; therefore, all protein and peptide solutions must be thoroughly dia-
lyzed against a suitable buffer. The two interacting species must be
dialyzed against the exact same buffer, and ideally in the same container.
This is not always possible, particularly when one species is a small peptide.
Alternatively, peptides can be exchanged into pure distilled water using a
gel filtration or desalting column before being lyophilized. The lyophilized
peptide can then be resuspended in the dialysis buffer of the protein
sample. A buffer system should be chosen that ensures both the protein
and peptide species are stable and soluble. A wide range of pH’s, buffer,
salt, and DTT concentrations are suitable for use in ITC experiments.
Phosphate buffers are particularly useful in ITC due to low ionization
enthalpy.39 All solutions must be thoroughly degassed and devoid of any
particulate matter.
The initial concentrations of the protein and peptide are critical to the
success of the experiment. The proper concentrations of macromolecules
are estimated using the following equation:
c ¼ Ka Mtot n; (9)
where c is a unit-less constant, Ka the binding association constant (Ka ¼
1/KD), Mtot the total macromolecular concentration at the beginning of
the experiment, and n is the stoichiometry of the interaction. The constant
c can be used to estimate the shape of the binding isotherm produced by
the experiment. Values of c > 5000 produce a rectangular curve with a
height equal to exactly the H of the interaction. It is not possible to accur-
ately fit the KD of an interaction when the c value is this high due to the
sharp transition of the curve at the equivalence point. In the condition of
c < 1, the curve becomes broad and featureless, making it impossible to
extract the thermodynamics of the system. Therefore, to ideally measure
the binding affinity and change in enthalpy, the c value should be above 1
and below 1000.40 However, it is not possible to accurately estimate
37
P. L. Privalov and G. I. Makhatadze, J. Mol. Biol. 213, 385 (1990).
38
G. I. Makhatadze and P. L. Privalov, J. Mol. Biol. 213, 375 (1990).
39
M. M. Pierce, C. S. Raman, and B. T. Nall, Methods 19, 213 (1999).
40
T. Wiseman, S. Williston, J. F. Brandts, and L. N. Lin, Anal. Biochem. 179, 131 (1989).
the c value for initial experiments unless the binding affinity and stoichiom-
etry are known fairly accurately. This may not be practical and several pre-
liminary experiments may be needed to obtain general estimates of the
stoichiometry and affinity.
Several control experiments should be conducted.39,41 First, the buffer
solution should be titrated into the cell containing the protein and the heat
observed. Additionally, the reverse experiment, where the peptide is ti-
trated into the buffer solution containing no protein should also be per-
formed if sufficient amounts of peptide are available. These control
experiments determine the heats of dilution of the system and the results
are substracted from the experimental data to obtain accurate heats.
Protocol for Analyzing the Thermodynamics of the Interactions of the

HP1 Chromodomain with H3 Tail Peptides Containing
Dimethyllysine 9 or Trimethyllysine 9 by ITC
All calorimetry measurements were performed using a VP-ITC instru-
ment from MicroCal (NorthHampton, MA).40 Purified HP1 17–76 was
concentrated to a volume of 3 ml yielding a concentration of 70 M. This
concentration was used because it produces a c value of 15 [Eq. (9)] as-
suming a stoichiometry of 1:1 and a Ka of 25 L/mol, where the latter was
easily estimated by fluorescence anisotropy binding studies (see earlier).
The sample was dialyzed against 3 4 L of a buffer solution of 50 mM
sodium phosphate, pH 8.0, 25 mM NaCl, and 2 mM DTT. Synthetic H3
peptides were resuspended in the HP1 dialysis buffer to a concentration
of 700 M. The reference cell was filled with dialysis buffer and the sample
cell with the 70 M HP1 17–76 solution. The H3 peptide solution was then
titrated into the sample cell with 25 injections of 10 l each. Injections were
spaced at 3-min intervals. The integrated heats of injection were fit to the
quadratic equation:
Q ¼ ðn½cd t HV0 Þ=2 1 þ ½H3 t =ðn½cd t Þ þ 1=ðnKa ½cd t Þ
n o12
ð1 þ ½H3 t =ðn½cd t Þ þ 1=ðnKa ½cd t ÞÞ2 4½H3 t =ðn½cd t Þ ; (10)
where Q is the heat evolved, n the stoichiometry of the interaction, H

the enthalpy change, [cd]t the total concentration of the chromodomain,
[H3]t the total concentration of the H3 peptide, Ka the association con-
stant, and V0 is the initial volume. H, n, and Ka are fit as floating multiple
41
R. O’Brien, B. Z. Chowdhry, and J. E. Ladbury, in ‘‘Protein-Ligand Interactions:
Hydrodynamics and Calorimetry,’’ p. 263. Oxford University Press, London, 2001.
independent variables using the analysis software Origin which is used to

normalize the heats of binding and perform a volume correction due to di-
lution before integration. A control curve representing the heat of dilution
of the buffer titrated into the chromodomain sample was subtracted from
the data before fitting. Eqs. (6) and (7) were then used to obtain G and
S for the interaction.
Acknowledgments
W.F. is a Robert Black fellow of the Damon Runyon Cancer Research Foundation. This
work was supported in part by NIH grant GM116635 to SK.
[10] Expression, Purification, and Biophysical Studies

of Chromodomain Proteins
By Peter R. Nielsen, Juliana Callaghan, Alexey G. Murzin,
Natalia V. Murzina, and Ernest D. Laue
Chromodomains1,2 are one of a number of similar domains that are in-

volved in chromatin structure.3 They can be divided into two subfamilies:
the chromodomains and the chromo shadow domains.4 The chromodomain
is found in a diverse selection of proteins, but the chromo shadow domain
is only associated with the HP1-family proteins, where it is always found C-
terminal to a chromodomain. For some of these domains it has been pos-
sible to identify an interacting partner. In these cases it is possible to use
biophysical techniques to understand how these proteins work. Alterna-
tively, one can use biophysical methods to study interactions with (or iden-
tify) less well characterized interacting partners that might bind to other
chromodomains.
In the first section we will describe the structural features of chromodo-
mains and how we can use this information to identify chromodomains
from amino acid sequences. We have employed various biophysical tech-
niques to study the chromodomain and chromo shadow domain of HP1.
Here we describe how we express and purify the individual domains as well
as the intact protein. We also describe how we characterize the purified
1
J. C. Eissenberg, Gene 275, 19 (2001).
2
R. Paro and D. Hogness, Proc. Natl. Acad. Sci. 88, 263 (1991).
3
S. Maurer-Stroh, N. J. Dickens, L. Hughes-Davies, T. Kouzarides, F. Eisenhaber, and C. P.
Ponting, Trends Biochem. Sci. 28, 69 (2003).
4
R. Aasland and A. F. Stewart, Nucl. Acids Res. 23, 3168 (1995).

independent variables using the analysis software Origin which is used to

normalize the heats of binding and perform a volume correction due to di-
lution before integration. A control curve representing the heat of dilution
of the buffer titrated into the chromodomain sample was subtracted from
the data before fitting. Eqs. (6) and (7) were then used to obtain G and
S for the interaction.
Acknowledgments
W.F. is a Robert Black fellow of the Damon Runyon Cancer Research Foundation. This
work was supported in part by NIH grant GM116635 to SK.
[10] Expression, Purification, and Biophysical Studies

of Chromodomain Proteins
By Peter R. Nielsen, Juliana Callaghan, Alexey G. Murzin,
Natalia V. Murzina, and Ernest D. Laue
Chromodomains1,2 are one of a number of similar domains that are in-

volved in chromatin structure.3 They can be divided into two subfamilies:
the chromodomains and the chromo shadow domains.4 The chromodomain
is found in a diverse selection of proteins, but the chromo shadow domain
is only associated with the HP1-family proteins, where it is always found C-
terminal to a chromodomain. For some of these domains it has been pos-
sible to identify an interacting partner. In these cases it is possible to use
biophysical techniques to understand how these proteins work. Alterna-
tively, one can use biophysical methods to study interactions with (or iden-
tify) less well characterized interacting partners that might bind to other
chromodomains.
In the first section we will describe the structural features of chromodo-
mains and how we can use this information to identify chromodomains
from amino acid sequences. We have employed various biophysical tech-
niques to study the chromodomain and chromo shadow domain of HP1.
Here we describe how we express and purify the individual domains as well
as the intact protein. We also describe how we characterize the purified
1
J. C. Eissenberg, Gene 275, 19 (2001).
2
R. Paro and D. Hogness, Proc. Natl. Acad. Sci. 88, 263 (1991).
3
S. Maurer-Stroh, N. J. Dickens, L. Hughes-Davies, T. Kouzarides, F. Eisenhaber, and C. P.
Ponting, Trends Biochem. Sci. 28, 69 (2003).
4
R. Aasland and A. F. Stewart, Nucl. Acids Res. 23, 3168 (1995).

[10] biophysical studies of chromodomain proteins 149
proteins using steady-state fluorescence spectroscopy, analytical ultracen-

trifugation (AUC), and nuclear magnetic resonance (NMR) spectroscopy.
There will not be room for a description of how we use NMR to determine
3D structures, but the procedures we use in our laboratory are the subject
of a recent review.5
Sequence Analysis and Recognition of Chromodomains

The chromodomain (chromo box) was first defined as a conserved se-
quence motif found in the Drosophila HP1 and Pc proteins.2 Since then
several chromodomain structures have been determined,6–9 all of which
show that the domain extends well beyond the most conserved residues
that defined the original chromo box used in sequence searches. The avail-
ability of these 3D structures has now allowed us to define the sequence
requirements that are needed to form the chromodomain fold.
The chromodomain structure comprises an N-terminal three-stranded
-sheet that packs against a C-terminal -helix.6–9 The conserved chromo
box sequence corresponds to the second/third strands, which are connected
by a loop of variable length (Fig. 1). The structural core and packing be-
tween the -sheet and the -helix is well conserved—it is determined by
the pattern of hydrophobic residues and by the presence of conserved
bulges in the first strand (Fig. 1B). These bulges are crucial for defining
the geometry of the interaction between the sheet and the helix, and thus
the domain’s overall 3D structure (Fig. 1A). The structural core of the
domain also includes part of the linker between the -sheet and the -helix.
Consequently, variations in the length of the linker are limited to either
one or two turns of helix, as exemplified by the chromodomain and chromo
shadow domain of HP1 (aka MOD1 or M31), respectively (Fig. 1).
Neither longer nor shorter linkers connecting the sheet to the helix can
be incorporated without disrupting the structure.
5
D. Nietlispach, H. R. Mott, K. M. Stott, P. R. Nielsen, A. Thiru, and E. D. Laue, in
‘‘Methods in Molecular Biology, Protein NMR Techniques. 2nd Edition: Structure
Determination of Protein Complexes by NMR’’ (K. A. Downing, ed.), Vol. 278, Humana
Press Inc., Totowa, NJ, (in press) (2004).
6
L. J. Ball, N. V. Murzina, R. W. Broadhurst, A. R. C. Raine, S. J. Archer, F. J. Stott,
A. G. Murzin, P. B. Singh, P. J. Domaille, and E. D. Laue, EMBO J. 16, 2473 (1997).
7
S. V. Brasher, B. O. Smith, R. H. Fogh, D. Nietlispach, A. Thiru, P. R. Nielsen,
R. W. Broadhurst, L. J. Ball, N. V. Murzina, and E. D. Laue, EMBO J. 19, 1587 (2000).
8
N. P. Cowieson, J. F. Partridge, R. C. Allshire, and P. J. McLaughlin, Curr. Biol. 10,
517 (2000).
9
D. A. Horita, A. V. Ivanova, A. S. Altieri, A. J. S. Klar, and R. A. Byrd, J. Mol. Biol. 307,
861 (2001).
Fig. 1. (continued)
The structural studies, together with a large number of biochemical ex-

periments, originally suggested that chromodomains might act as protein
interacting modules, which function by targeting different protein com-
plexes to particular sites on chromatin.6 It is now clear that a complete
chromodomain structure is essential for these functions. First, there are
functionally impaired HP1 and Pc mutants with point substitutions in the
hydrophobic core which appear to destabilize the chromodomain struc-
ture.6,10,11 Second, in the HP1 and Swi6 chromo shadow domains, the -
helix provides both the interface for dimerization as well as the binding site
for interaction with other proteins.7,8 Finally, both the HP1 and
Swi6 chromodomains have recently been shown to specifically recognize
and bind to Lys-9 methylated histone H3.12–14 The hydrophobic groove
formed between the -sheet and the -helix provides the binding-site for
this interaction,15,16 that is, the formation of this binding site requires an
intact chromodomain.
10
S. Messmer, A. Franke, and R. Paro, Genes Dev. 6, 1241 (1992).
11
J. S. Platero, T. Hartnett, and J. C. Eissenberg, EMBO J. 14, 3977 (1995).
12
M. Lachner, D. O’Carrol, S. Rea, K. Mechtler, and T. Jenuwein, Nature 410, 116 (2001).
13
14
J.-I. Nakayama, J. C. Rice, B. D. Strahl, C. D. Allis, and S. I. S. Grewal, Science 292, 110
(2001).
15
P. R. Nielsen, D. Nietlispach, H. R. Mott, J. Callaghan, A. Bannister, T. Kouzarides, A. G.
Murzin, N. V. Murzina, and E. D. Laue, Nature 416, 103 (2002).
16
S. A. Jacobs and S. Khorasanizadeh, Science 295, 2080 (2002).
Fig. 1. Sequence alignment and structures of chromodomains. (A) Superposition of the

structures of the HP1 chromodomain (black, PDB code 1apo), the HP1 chromo shadow
domain (grey, PDB code 1dz1), and the Swi6 chromo shadow domain (white, PDB code
1e0b). (B) Comparison of sequences containing the chromo box consensus motif (i)
chromodomains with known structures, (ii) selected chromodomain sequences that clearly fit
the requirements of the structure, and (iii) the putative chromodomains from MSL3, MOF1,
and RBP1. In (B) the positions of secondary structure elements are indicated by green arrows
(extended strands) and dark blue cylinders (-helices). The blue dots indicate the positions of
the conserved bulges in the first strand. Conserved hydrophobic core residues that define the
fold are shown in yellow, while those determining the borders of the turn between the -
strand and the helix are shown in green. In MSL3, MOF1, and RBP1, the grey residues
indicate absence of conserved residues important for the formation of the 3D chromodomain
fold. The chromo box used in previous sequence analysis is indicated by a black line. For each
sequence the residue and SWISS-PROT/TrEMBL access numbers are given. Residues are
color coded similarly in (A) and (B)—the residue numbers in (A) correspond to the positions
of the amino acids in the sequence alignment. Panels (A) and (B) were produced using
MOLSCRIPT [P. J. Kraulis, J. Appl. Crystallogr. 24, 946 (1991).] and ALSCRIPT [G. J.
Barton, Protein Eng. 6, 37 (1993).], respectively.
Based on this analysis we can now construct a structure-based sequence

profile (the highlighted residues in Fig. 1B), which allows us to readily
detect intact, and therefore functional, chromodomains in proteins in
which it has been predicted. As an example, consider the MSL-3 protein,
and the histone acetyl-transferase MOF, which have been recently found
to contain protein-RNA interaction modules.17 Sequence analysis origin-
ally suggested that the region of these proteins involved in RNA binding
contains a chromo box motif.4,18,19 However, on examination of the struc-
ture-based sequence profile it is clear that there are no suitable conserved
sequences flanking the chromo box in the MSL3, MOF, and RBP1 families
(Fig. 1B). In fact, the only part of the sequence that is conserved in these
families maps on to the second/third strands and their connecting loop. This
loop varies between the known chromodomains and does not determine the
fold. We conclude that the RBP1, MOF, and MSL-3 proteins cannot contain
canonical chromodomains. These results suggest that the MOF and
MSL-3 proteins bind RNA sequences via a novel RNA binding domain
that contains a conserved sequence motif found also in chromodomains.
The ability to recognize domains from amino acid sequences is clearly a
useful tool when characterizing proteins with unknown function. However,
as we demonstrate here, sequence motifs have to be used with care and
should be validated against structural and functional data when it is possible.
In the following sections, we describe the procedures that we use to ex-
press and purify chromodomains, as well as the biophysical methods we
employ to study these domains and their interactions with other peptides
and proteins.
Expression and Purification Procedures
Expression of HP1 Proteins and their Domains

Ideally for structural studies, for example, for making NMR samples,
one needs 5–10 mg of protein that is at least 90% pure and stable at room
temperature for about 2 weeks. This puts exacting requirements on the
methods employed for protein production. Optimizing the level of protein
expression is the first step. With both the chromodomain and chromo
shadow domain we had unexplained protein expression before induction
that could not be prevented, for example, by adding more glucose to
the growth media, by changing the Escherichia coli host strain, or by
17
A. Akhtar, D. Zink, and P. B. Becker, Nature 407, 405 (2000).
18
E. V. Koonin, S. Zhou, and J. C. Lucchesi, Nucleic Acids Res. 23, 4229 (1995).
19
A. Hilfiker, D. Hilfiker-Kleiner, A. Pannuti, and J. C. Lucchesi, EMBO J. 16, 2054 (1997).
overexpression of the Lac repressor. This ‘‘leaky’’ expression resulted in

eventual plasmid loss and a low yield of protein—presumably due to tox-
icity of the proteins. We found that we could not obtain the desired level
of protein expression using overnight or non-stationary liquid starter cul-
tures to inoculate growth media used for protein expression. Instead, we
chose to inoculate the growth media with a large number of colonies di-
rectly from plates. Freshly transformed cells were plated out so as to get
large numbers of isolated colonies—typically we used 4 plates per 0.5 L
of growth media. All the isolated colonies, but not the cell lawn, were used
to inoculate the growth flasks with an approximately equal number of
cells—OD600 0.1–0.3. In addition, we added glucose to 1% w/v. Protein
expression was induced with 1 mM IPTG at OD600 0.7 and the cells were
grown for another 3 h before collection.
Purification of HP1 Proteins and Their Domains

The cells were then resuspended in ice cold lysis buffer (0.5 M NaCl,
20 mM Tris of pH 7.9, and 5% (v/v) Sigma General Protease Inhibitor
Cocktail) and broken in a French Press (or using an equivalent method).
NP-10 was added to 0.5% (w/v), -mercaptoethanol (BME) to 2 mM, imid-
azole to 4 mM and the cells spun at 15,000 rpm for 30 min in a precooled
rotor to remove insoluble material. (Positively charged chromodomains
also require suitable treatment with DNAse I before centrifugation.) The
supernatant was then loaded onto a Ni-column—typically we use a Ni-
NTA resin (Qiagen, Novagen, Pierce). To make a new column, 4 ml of
50% agarose beads are placed in a plastic 20-ml column (Bio-Rad) which
already contains about 5 ml of water. Once the agarose has settled a hydro-
phobic disc is placed on the top of the gel. The column is then washed with
20 ml of water, 20 ml of buffer A (300 mM NaCl in 20 mM Tris of pH 7.9),
20 ml of 1 M imidazole in buffer A, and then 20 ml of buffer A again. This
washing procedure will remove excess Ni from the column. It is optional
for new agarose, but obligatory after recharging with Ni (see later). The
columns can be reused many times for the same protein purification, but
should be recharged with Ni using 20 ml of 0.5 M NiSO4 following use.
Sometimes it is advisable to additionally repack the column.
The supernatant is loaded to the top of a column equilibrated in buffer
A. After loading, the column is washed with 50 ml of buffer A containing
2 mM BME, 4 mM imidazole and 100 l of Sigma General Protease
Inhibitors. It is then washed with 50 ml of 20 mM imidazole in buffer A
(without the BME and protease inhibitors) and then with 25–30 ml
of 80 mM imidazole in buffer A. The later washes should be collected
separately in 5 and 25 ml aliquots and analyzed by SDS gel electrophoresis.
Most of the unwanted proteins will typically elute in the first 5 ml, but the
25-ml aliquot may contain some protein. Next, the protein is eluted with
25 ml of 0.5 M imidazole in buffer A. Analysis of the various fractions by
SDS gels is then needed to determine which should be pooled.
The eluted solutions contain Ni leached from the column, which may
cause precipitation during protein concentration. To avoid this we usually
pass the solutions through a Chelex resin column (Bio-Rad). The protein is
then concentrated to 5 ml in an Amicon 60-ml cell (Millipore) using a
Diaflo YM3 membrane (Millipore) and buffer exchanged on a PD-10
column (Pharmacia) into Factor Xa cutting buffer (100 mM NaCl,
50 mM Tris of pH 8.0). At this stage the OD280 is measured to estimate
the yield. (A280 ¼ 1.0 is 0.5 mg/ml for the chromodomain, 0.66 mg/ml
for the chromo shadow domain and 0.75 mg/ml for full length HP1.)
Factor Xa (Roche) is dissolved in the Factor Xa buffer and added to the
protein at between 1:50 and 1:100 (w/w). After addition of CaCl2 to

2 mM final, the sample is left overnight at 25 . (The cleaved His-tag can
then be removed by passing the sample through the Ni-column again if re-
quired.) At this stage the protein is usually sufficiently pure for most bio-
chemical applications. However, these proteins are not sufficiently stable
for structural studies and after a few weeks, storage, degradation bands
appear on SDS gels. We have found that the chromodomains can be safely

kept at 20 , whereas the shadow domain and full-length proteins tend to
aggregate after being frozen (even with flash freezing or in the presence of
glycerol). To further purify the protein, an ion exchange chromatography
step (MonoQ or MonoS (Pharmacia) columns) helps to improve the pro-
tein stability. After Factor Xa digestion we concentrate the protein solu-
tion to 5 ml using an Amicon 60-ml cell, prior to buffer exchange on the
PD-10 column into buffer B—10 mM Na phosphate (pH 7.5–8.0 for
MonoQ and pH 5.5–6.5 for MonoS). Alternatively, one can dialyze the pro-
tein in buffer B either overnight or for 2–3 h to reduce the salt concentra-
tion. A 1-ml MonoQ or MonoS column is washed/equilibrated with 5 ml
buffer B (see earlier), 10 ml buffer C (1 M NaCl in buffer B) and 5 ml
buffer B. The sample is loaded onto the column in buffer B and run on
an FPLC system at 0.5 ml/min, collecting 1-ml fractions. The column is
then washed with 10 ml of buffer A and a gradient applied from 0% to
60% C over 60 ml and from 60% to 100% C over 5 ml. The resulting
fractions are run on an SDS gel and pooled as appropriate. The shadow
domain elutes from the column in lower-salt concentration than the
chromodomain and full-length HP1 protein, which have higher surface
charge. Mass spectrometry of the samples should then be used to identify
whether any of the protein sample has become degraded or clipped at
either the N- or C-termini.
Sample Preparation
Protein aggregation will interfere with most protein-protein interaction
assays and will decrease the quality of NMR spectra or the likelihood
of crystallization. The full-length HP1 protein and the shadow domain
usually require gel filtration to remove any aggregated protein, whereas

the chromodomains samples do not. After prolonged storage at 4 , samples
may need to be re-run on the gel filtration column to remove further
aggregates that form. For gel filtration of NMR samples, we use a
Superdex S75 column (120-ml bed volume from Pharmacia) in 100 mM
NaCl, 20 mM phosphate of pH 5.5–8.0 at 1 ml/min over 150 ml, collecting
1-ml fractions between 35 and 125 ml. The protein fractions are
pooled, concentrated, and buffer exchanged if required. To check for
any remaining aggregation, and to carry out binding assays, we run an
analogous 2.4-ml bed volume gel filtration column using a Smart system
(Pharmacia). For NMR sample preparation, the protein-containing S75
fractions were concentrated to 0.4 ml in a 10-ml Amicon cell and buffer ex-
changed into NMR buffer (10 mM Na phosphate with pH as required by
the experiment). The A280 of the solution is measured using a quartz capil-
lary (Shimadzu) with 0.5-mm diameter (sample volume of 7–10 l) to
estimate the protein concentration and 50 l D2O, 2H-DTT to 10 mM,
2
H-EDTA to 1 mM, and 0.01% NaN3 are added to the protein sample.
Finally, the sample is made up to 0.55 ml with NMR buffer.
Making Mixed Dimer Chromo Shadow Domain Samples

The shadow domain is a symmetrical dimer and equivalent residues in
each monomer give rise to the same set of signals in the NMR spectra. Ac-
cordingly, it is not possible to distinguish between signals that arise from
NOEs between intra- and inter-molecular proton interactions at the dimer
interface. To overcome this problem we made a protein sample where one
monomer is unlabeled and the other 13C/15N-labeled. With such a sample
one can employ NMR experiments that detect interactions between two
protons, one which is attached to a 13C-nucleus and the other to a 12C-
nucleus. This allows direct and unambiguous determination of inter-mo-
lecular interactions at the dimer interface. Normally one can make such a
sample by mixing equal amounts of the differently labeled proteins and
leave them until they reach equilibrium. In the case of the shadow domain
this failed. The two forms of the shadow domain were mixed to final con-
centrations of either 0.1–1 mM or 100 nM to 1 M. The mixed samples
were left to equilibrate for 1 week. No exchange of monomers was detected
by NMR experiments (for the high-concentration range) or by Coomassie
stained pull-down assays (for the low-concentration range). In addition, an

AUC experiment showed that all of the shadow domain exists as a dimer
without detectable trace of monomers—the Kd for dimer dissociation
being below the detection range. Taking all these results together we con-
cluded that the shadow domains form very stable dimers with a low ex-
change rate and an estimated Kd of below 150 nM. These findings also
suggest that, in vivo, exchange of subunits between the different isoforms
of HP1 is unlikely.
To prepare a mixed dimer for NMR experiments we denatured a mix-
ture of equal amounts of the 12C- and 13C-labeled proteins in 6 M guani-
dium HCl (GuHCl) and then refolded the protein by subsequent dilution
to reduce the denaturant concentration. To determine the GuHCl concen-
tration where unfolding/refolding is occurring, we measured UV absorp-
tion spectra of the shadow domain at different GuHCl concentrations
and found that there was a large intensity change between 4 and 5 M
GuHCl. For denaturation, 0.5 ml of the protein solution was added to
10 ml of 6 M GuHCl in 10 mM Na phosphate of pH 8.0, 1 mM EDTA
and 1 mM DTT, and left for 10 min at room temperature. For subsequent
refolding, GuHCl solutions of decreasing concentration in the same buffer
were very slowly added (drop-by-drop—each solution taking about 10 min
to add) as follows: 10 ml of 4.5 M GuHCl, 10 ml of 4 M GuHCl, 20 ml of
3 M GuHCl, and finally 30 ml of buffer alone. The refolded protein was
then concentrated in an Amicon cell and purified by gel filtration (as
mentioned earlier).
Expression and Purification of the Histone H3 N-Terminal Peptides

Histone H3 N-terminal peptides (amino acids 1–14 or 1–17) were
expressed as GST fusions in BL21pREP4 (pREP4 is Qiagen plasmid

with an extra copy of the LacI gene) at 24 . The GST-fusion peptide was
purified by conventional procedures except that the cell lysate was treated
with DNAse I and the glutathione agarose beads were extensively washed
after binding the protein with 1 M NaCl and 0.5% NP-40 in 20 mM Tris of
pH 7.5 to remove DNA contamination. After washing, the beads were
transferred into thrombin cutting buffer (as recommended by the manufac-
turer, Novagen) and the peptide was cleaved by overnight incubation of the

bead suspension at 4 . The peptide was washed from the beads with
100 mM NaCl in 10 mM Tris of pH 8.0, loaded directly onto an octadecyl
(C18) column, and eluted with a 0–100% water/acetonitrile gradient con-
taining 0.1% trifluoroacetic acid. The eluted peptide was freeze dried and
redissolved in Na phosphate of pH 8.0.
13
Reductive Methylation and Production of C-Methyl–Labeled
Histone H3 N-Terminal Peptides
It is possible to produce peptides that are dimethylated on all the amino
groups in an unmodified substrate by reductive alkylation.20 At high pH,
and in the presence of sodium borohydride as a reducing agent, the N-ter-
minal -NH2 and the e-NH2 groups of lysine side chains will receive first
one and then another methyl group from formaldehyde. We used this
method to produce peptides that were specifically 13C-labeled in all the
methyl positions of the dimethylated amino groups. This enabled us to un-
ambiguously identify signals from the Lys-9 methyl groups in NMR spectra
of histone H3. One hundred microliters of 3.5 mM purified peptide were
mixed on ice with 900 l of freshly made 1 mg/ml NaBH4 in 200 mM NaH-
BO3 of pH 9.0 and 5 l of 20% 13C-labeled formaldehyde. The mixture
was incubated on ice for 6 min and then transferred to a new tube contain-
ing 1 mg of NaBH4 powder. A further 5 l of the formaldehyde solution
was added and the cycle was repeated eight times. The final product was
checked by mass spectrometry for uniform methylation and the absence
of cross-linking.
Chromodomain–Peptide Complex NMR Samples

Samples of the complexes between chromodomains and peptides were
prepared in the following way. For the chromodomain–histone H3 peptide
complex, the protein obtained after gel filtration was buffer exchanged as
for chromodomain samples and concentrated in an Amicon cell but this
time to 0.25 ml. The histone H3 peptide was then added from a 2-mM stock
solution to the required level. For studies of the shadow domain–CAF-1
complex, the 25-mer CAFp150 peptide is not very soluble in aqueous
buffer. This meant that the complex had to be made at low concentration.
To avoid aggregation the peptide was added to a protein solution diluted
20 times compared to the NMR sample concentration before concentrat-
ing the complex using an Amicon cell equipped with a YM1 membrane. To
both types of sample 55 l D2O was added and the sample volume adjusted
to 550 l with the appropriate phosphate buffer. Usually the samples also
contained 10 mM 2H-DTT, 1 mM 2H-EDTA, and 0.01% NaN3.
Fluorescence Spectroscopy to Study Ligand Binding

Fluorescence spectroscopy is an ideal method by which to study the
interaction of the HP1 chromodomain with histone H3.21 The interaction
is mediated via the methyl groups of lysine 9 of histone H3, which fit into a
hydrophobic pocket on the surface of the chromodomain formed by the
aromatic residues Tyr 21, Phe 45, and Trp 42. Protein fluorescence origi-
nates from the aromatic residues, Phe, Tyr, and Trp. However, the fluores-
cence of proteins containing all three aromatic amino acids is usually
dominated by tryptophan. Upon methyl lysine binding to the chromodo-
main, the microenvironment of the tryptophan changes and thus the inten-
sity of the fluorescence emission alters. This provides a mechanism,
intrinsic to the protein, by which the interaction can be studied. Steady-
state fluorescence not only identifies the occurrence of an interaction, but
can also be used to determine the affinity of the binding.
Sample Preparation
In order to perform steady-state fluorescence analysis of the chromodo-
main–histone H3 complex the protein must be present in a suitable buffer.
The buffer must maintain the protein in the correctly folded state and must
not affect the intensity of the light emitted from the tryptophan residues.
For this reason it is important to remove any imidazole left in the sample
after Ni2þ column purification. (However, it is worth noting that the pres-
ence of glutathione does not have this effect.) Thus, during the final stages
of chromodomain purification the protein was buffer exchanged into
10 mM Na phosphate of pH 7.4 and 150 mM NaCl. The solution of the
HP1 chromodomain was made up to a concentration of 3 M, as con-
firmed by amino acid analysis. The solution of the histone H3 peptide, to
which binding of the chromodomain was to be investigated, was made up
to a concentration of 1 mM within a solution of the 3 M chromodomain.
This eliminates the need to account for any dilution effect upon addition of
the peptide to the chromodomain.
Experiments
In order to perform the titrations the chromodomain solution was
placed in a 1-cm 1-cm quartz cuvette in the spectrofluorimeter (Perkin
Elmer, LS55 Luminescence Spectrometer) equipped with a thermostated
cell holder. All solutions used were previously incubated in a water bath

and all experiments were performed at 23 . After the protein solution
was transferred to the cuvette and had been inserted into the spectrofluor-
imeter, the sample was left for 10 min in order to ensure that the tempera-
ture had equilibrated. (Fluctuations in temperature are to be avoided since
they affect the thermodynamics and kinetics of the interaction, and thus the
fluorescence intensity emitted on excitation.) The sample was excited at
20
G. E. Means and R. E. Feeney, Biochemistry 7, 2192 (1968).
21
M. R. Eftink, Methods Enzymol. 278, 221 (1997).
280 nm and emission spectra were recorded between 300 and 400 nm—ex-
citation and emission slits of 5 nm, and a scan speed of 200 s were used.
Following the initial measurement of the chromodomain alone, small ali-
quots of 0.5–2.5 l of the 1 mM histone H3 peptide solution were added se-
quentially (increasing the peptide concentration by 0.5–2.5 M at a time)
to the solution of the chromodomain within the cuvette. Upon peptide ad-
dition the solution was mixed using a plastic sterile loop and allowed to
settle prior to recording the next emission spectrum. The peak fluorescence
intensity is detected at 335 nm confirming that the chromodomain is
folded (the value for tryptophan in aqueous solution is for comparison,
348 nm). During the titration the tryptophan fluorescence intensity in-
creased on addition of the histone H3 peptide. Additions of the peptide
were then continued until no change in the fluorescence intensity was
detected and saturation of the chromodomain H3 binding site had been
achieved. Figure 2 shows a typical titration series. In this way it was
Fig. 2. Tryptophan fluorescence emission spectra of a HP1 (Y21W) chromodomain on

addition of histone H3 peptide. Binding of the H3 peptide to the chromodomain gives rise to
an increase in the intensity of the fluorescence. The fluorescence intensity ceases to increase
on addition of further peptide once saturation of the binding site has been achieved.
Fig. 3. Binding curve for the interaction of histone H3 tri-methylated at Lys-9 with the
chromodomain of HP1. The Kd for the interaction is 0.65 M and has been determined using
Eq. (1) (see text).
possible to compare the interaction between the chromodomain and a

number of histone H3 peptides with different modifications. We studied
H3 peptides, each methylated by different amounts at different lysine resi-
dues. One was tri-methylated at both Lys-4 and Lys-9, whereas in the other
both residues were di-methylated. A third peptide was tri-methylated
solely at Lys-9 and a fourth tri-methylated only at Lys-27.
Data Analysis
The change in fluorescence intensity upon addition of an interacting
species provides a measure of the degree to which the environment of
the Trp residues of the protein under investigation become altered due to
the interaction. The resultant fluorescent intensities can be plotted as a
function of the concentration of the interacting species (Fig. 3), and the
Kd can then be determined for the interaction by least squares fitting using
the equation21:
Fobs ¼ FP þ ðFPL FP Þ=ð2½P tot Þð½P tot þ ½L tot þ Kd

sqrt ð½P tot þ ½L tot þ Kd Þ2 4½P tot ½L tot ; (1)
where Fobs represents the fluorescence observed, Fp represents the fluores-

cence of the free protein, FPL represents the fluorescence of the bound
species, and [P]tot and [L]tot represent the total protein and ligand
concentrations, respectively.
Results
The studies of the interaction of the chromodomain with bis-methylated
H3 peptides showed that the binding affinity of the chromodomain to the H3
peptide di-methylated at both Lys-4 and Lys-9 was very similar to that of the
analogous tri-methylated peptide. The Kd s for the chromodomain binding
to the two peptides were found to be 2.1 and 1.9 M, respectively. The Kd
for the interaction of the chromodomain with the H3 peptide tri-methylated
at only Lys-9 was found to be only slightly lower than those determined for
the interaction of the chromodomain with the bis-methylated H3 peptides,
with a Kd of 0.7 M. This implies that the methylation of Lys-4 neither con-
tributes to nor greatly interferes with the recognition of Lys-9 methylated
histone H3 by the HP1 chromodomain. The binding studies of the chromo-
domain with the H3 peptide methylated at Lys-27 showed that this peptide
binds much more weakly (Kd 12 M). Despite the sequence similarity to
the Lys-9 position, it is clear that the sequence surrounding Lys-27 does not
provide a suitable interaction partner for the chromodomain of HP1, as
was predicted from the structure.15,16
Characterizing Protein Samples by AUC

Analytical ultracentrifugation is a well-established technique for meas-
uring the hydrodynamic properties of proteins (or complexes) in solution.
The equations describing the results obtained using the ultracentrifuge rest
on a solid thermodynamic foundation that has been well described.22 Ex-
periments performed in the ultracentrifuge give information on the size
and shape of solutes in well-defined solvents. It is possible to measure ac-
curate molecular weights, characterize interactions, or to investigate com-
plex mixtures. In our laboratory we mainly use AUC for two purposes: to
investigate the oligomeric state of purified protein samples, and to confirm
complex formation and the stoichiometry of interactions. It follows that we
only analyze highly purified proteins and we normally do not carry out
sedimentation velocity experiments since we already know what the com-
ponents in the solution are. Instead, we investigate the behavior of proteins
by sedimentation equilibrium analysis.
In sedimentation equilibrium experiments the sample is spun at a ve-
locity where it is possible to achieve an equilibrium between sedimentation
in the gravitational field and diffusion in the opposing direction. When
equilibrium is attained the concentration of the solute is measured as a
22
T. Svedberg and K. O. Pedersen, in ‘‘The Ultracentrifuge’’ (R. H. Fowler and P. Kapitza
eds.), Clarendon Press, Oxford, 1940.
function of the radial position. For a single ideal solute the radial
dependence of the concentration can be described by22,23:

cðrÞ ¼ cðr0 Þ exp Mð1 vbar Þ!2 r2 r02 =ð2RTÞ þ K; (2)
where vbar is the partial specific volume of the solute, is the density of the
solvent (both of which can be calculated from the amino acid and buffer
composition, respectively, using Sednterp, Tom Laue and David B. Hayes,
University of New Hampshire),24 ! is the angular velocity, c(r) is the con-
centration at the radial position r in centimeters, M is the molecular weight
of the solute, R is the gas constant, and T is the temperature in degrees
Kelvin. K is a constant added to take account of imperfections in the op-
tical system. The first point included in the calculations is at r0, with c(r0)
being the concentration at this point and r is the distance from the center
of the rotor in centimetres. With a knowledge of vbar and it is possible
to derive M by least squares fitting. This expression can be expanded to
describe mixtures of molecules, the behavior of a system in equilibrium
between several species and also non-ideal behavior.22,25
Experimental Design
Normally we design our experiments such that it would be possible to
detect multimerization in a reliable manner. This essentially means per-
forming nine separate experiments to obtain enough data to allow the de-
termination of dissociation constants. This is done by preparing three
samples, containing different concentrations of the protein in the same
buffer, and performing equilibrium experiments at three different rotor
speeds. The average molecular mass of the samples should change with
starting concentration and with angular velocity for interacting systems,
but not if it is a single species.
When choosing the optimal concentration for the experiments there are
three main issues to be considered. The first is to make sure that it is pos-
sible to detect the protein in a concentration dependent manner. Second,
when looking at interactions it is necessary to work at concentrations
where all the species in the equilibrium are present in significant amounts.
Finally, it has to be taken into consideration that at high concentrations
proteins will exhibit non-ideal behavior.
The Beckman XL-I instrument that we use employs two different types
of detection: absorption (180–800 nm) and Rayleigh interference optics. It
23
J. W. Williams, K. E. Van Holde, R. L. Baldwin, and H. Fujita, Chem. Rev. 58, 715 (1958).
24
All software mentioned in this section can be obtained from the RASMB web site:
http://www.bbri.org/RASMB/
25
D. K. McRorie and P. J. Voelker, ‘‘Self-Associating Systems in the Analytical Ultracentri-
fuge.’’ Beckman Instruments Inc., Palo Alto, California.
is recommended that the maximal adsorption does not exceed 1.2—

although in our experience the detector appears to be linear up to an ab-
sorption of 2.0. We normally choose starting concentrations that give an
absorption at 280 nm of between 0.2 and 0.5 in a 1-cm cell. (In our hands
the signal-to-noise ratio gets too low, if we use a sample with a lower ab-
sorption than this.) For the interference optics the maximal concentrations
that can be employed are around 5 mg/ml.26
The next thing to decide is what rotor speeds to apply. The goal is to
choose the speed such that there is a good curvature on the measured con-
centration profile (large concentration difference between the meniscus
and the bottom of the cell) while ensuring that the sample does not pellet.
This is achieved by selecting rotor speeds that will give a reduced apparent
molecular weight [the term in the exponent of Eq. (2)] of between 1.5 and
5.0. When studying homo dimerization the speeds are chosen based on the
molecular weight of the monomer—making sure not to pellet the dimer.
Sample Preparation
It is essential to use samples of high purity and quality if accurate infor-
mation about size or affinity is to be obtained. Any impurities or non-spe-
cific aggregates can contribute to the signal measured in the experiments
compromising the interpretation of the data. The best results are obtained,
if the sample is subjected to gel filtration chromatography shortly before
performing the AUC experiments to remove any non-specific higher-order
aggregates.
It is important that the buffer used as the reference in each cell corre-
sponds exactly to the solvent that holds the protein of interest. This can be
achieved in a number of ways. Typically, the fractions containing the pro-
tein after gel filtration chromatography will need to be concentrated. If this
is done by filter retention, the buffer that comes through the filter can be
used for any dilutions needed as well as for the reference buffer. Alterna-
tively, the protein sample can be dialyzed extensively against the required
final buffer and this can be used as the reference. Another method, that is
recommended by Laue,27 is to use spin columns for buffer exchange.28
The choice of buffer will also affect the result of the experiment. It is
important not to introduce substances that interfere with the detection
method used to follow the concentration of the protein. Thus, DTT cannot
be used in the buffer if protein concentrations are monitored by absorption
26
G. Ralston, ‘‘Introduction to Analytical Ultracentrifugation.’’ Beckman Instruments, Palo
Alto, California, 1993.
27
T. Laue, ‘‘Short Column Sedimentation Equilibrium Analysis for Rapid Characterization of
Macromolecules in Solution.’’ Technical Information DS-835. Beckman Coulter, Palo Alto,
California, 1992.
28
R. I. Christopherson, M. E. Jones, and L. R. Finch, Anal. Biochem. 100, 184 (1979).
at 280 nm. Breakdown products of DTT absorb light at this wavelength

and even very careful calibration of sample and reference solutions cannot
remove all of the contribution to the final signal that comes from these sub-
stances (i.e., interpretation of the absorbance in terms of protein concen-
tration will not be possible). Since experiments are performed over
several days, some samples require the presence of antioxidants. In these
cases we use TCEP—Tri(2-carboxyethyl)-phosphine29 (obtained from
Pierce). In order to minimize the effects of non-ideality on the results,
buffers should also contain more than 150 mM NaCl.26
AUC Studies of a Typical Chromodomain

To investigate the oligomeric state of a chromodomain, three samples
were prepared by dilution to concentrations corresponding to an A280 value
of 0.4, 0.6, and 0.8. For each sample the left chamber in a two-sector center-
piece of an assembled cell is loaded with 100 l of the sample and the right
chamber with 110 l of the reference buffer. The samples are brought to
equilibrium at three speeds of 25,000, 30,000, and 34,000 rpm at 4 . The
radial concentration gradient of protein in the sample chamber is measured
both as the absorbance at 280 nm and by the number of fringes from
bottom to top of the cell. For each speed the samples are left for 16 h prior
to the first measurement. Another set of data is then collected after 20 h
and the two are compared to verify that there is equilibrium between sedi-
mentation and diffusion in all of the samples. Ideally the samples should,
after the data sets for the highest rotor speed have been collected, be spun
at a velocity that sediments the protein, and a further data set is collected.
This is done in order to measure the contribution to the signal from imper-
fections in the optical system in the centrifuge. In practice this can be diffi-
cult to achieve when working with small proteins since the material of the
standard centerpieces (charcoal filled Epon, Beckman) cannot tolerate the
high speeds required. It can, however, be overcome by using centerpieces
made from aluminum or titanium.
The collected data can then be analyzed by least squares fitting to the
model described by Eq. (2). The values of c(r0) and K are fitted locally for each
data set, but the apparent molecular weight is determined globally. We use the
program WinNonlin.30 Models of increasing complexity are used to describe
the data until one is found where there is good agreement between meas-
ured and calculated data. This means that we first assume that there is only
one ideal component in the sample. Based on the value obtained compared
to the calculated molecular weight of the protein under investigation the
29
J. A. Burns, J. C. Butler, J. Moran, and G. M. Whiteside, J. Org. Chem. 56, 2648 (1991).
30
M. L. Johnson, J. J. Correira, D. A. Yphantis, and H. R. Halvorson, Biophys. J. 36, 575
(1981).
Fig. 4. AUC data for a chromodomain. Three data sets recorded using the Rayleigh
interference optics at equilibrium at 30,000 rpm are shown. The starting concentrations for the
three samples were: ()A280 ¼ 0.4, (~)A280 ¼ 0.6, and (þ)A280 ¼ 0.8. The solid lines are
calculated using the parameters obtained by fitting all nine data sets simultaneously to Eq. (2) as
described in the text. For clarity only every fifth point is plotted. The residual values left when
subtracting the calculated values of the concentration from the measured values are shown in
the top half of the graph. An even distribution around zero indicates that the data are described
well by the model. The remaining data sets were recorded at 25,000 and 34,000 rpm.
next step can be decided (Fig. 4). The shape of the residuals obtained from
the difference between the measured and calculated data can also be used as
an indication of the behavior of the system under study.25 In this case the
molecular weight calculated for the chromodomain is 10,600 Da compared
to 10,149 Da as calculated based on the amino acid sequence.
Characterization of Chromodomains by NMR Spectroscopy

A detailed characterization of the binding site for ligands of the
chromodomains can be carried out using NMR spectroscopy.
In NMR spectroscopy it is possible to take advantage of the fact that the
nuclei of atoms with spin quantum number equal to 1=2, distribute between
two energy levels when placed in a magnetic field.31 The energy difference
between these two states can be measured and is characteristic of the

strength of the magnetic field of the spectrometer used for the experiment
and the type of nucleus. The local electronic environment changes the size
of the magnetic field experienced by a nucleus. This means that each atom
in a protein will have a characteristic and different value. The value of this
local contribution is called the chemical shift. Because of this it is in principle
possible to identify signals in the NMR spectrum corresponding to all the
different spin 1=2 nuclei in the protein (or complex) under investigation.
For proteins the nuclei commonly used are the proton (1H) and particular
isotopes of nitrogen (15N) and carbon (13C), respectively. Whereas 99.98%
of hydrogen is 1H only 0.37% is 15N, and 1.11% is 13C. Thus, it is necessary
to enrich the fraction of these isotopes to nearly 100% by expressing the
protein in a host that grows on media containing either one or both of these
isotopes as the sole nitrogen and carbon source, respectively.32
Characterizing Interactions
A ligand will alter the environment of the atoms in the amino acids that it
interacts with. This means that the chemical shift of these atoms will change
when the ligand binds and this binding can be followed by NMR spectros-
copy. From the theory of NMR it can be shown that these changes can be
categorized into one of three types—which one occurs depends on the dif-
ference in the chemical shift between the bound and the free form compared
to the exchange rate of the ligand. If the exchange is slower than the size of
the change in chemical shift between the free and bound form, then signals
corresponding to both the free and the bound forms can be observed and the
relative populations can be estimated by measuring the intensity of the two
signals. In this case the system is said to be in the slow exchange limit. This
will be the case for strong binding. In the fast exchange limit where the ex-
change is fast compared to the chemical shift difference, there will only be
one observable signal for a given nucleus with a chemical shift that is the
weighted average of the two forms. The intermediary exchange regime gives
rise to highly broadened signals that are difficult to analyze.
This means that, unless the system is in intermediate exchange, it is
possible to obtain dissociation constants for an interaction by adding the
ligand to the protein and following the change in chemical shift (or intensity)
of a well-resolved signal during the titration. One-dimensional proton spec-
tra can in principle be used. If it is possible to obtain isotope-labeled
samples, however, it is advantageous to record 2D heteronuclear experi-
ments such as a 2D 1H-15N HSQC.33–35 This spectrum will contain one signal
31
J. Cavanagh, W. J. Fairbrother, A. G. Palmer, III, and N. J. Skelton, ‘‘Protein NMR
Spectroscopy. Principles and Practice.’’ Academic Press, San Diego, 1996.
32
L. P. Mcintosh and F. W. Dahlquist, Q. Rev. Biopys. 23, 1 (1990).
for each backbone amide proton and its attached nitrogen nucleus in the pro-
tein. The limited number of signals, combined with the improved resolution
gained from distributing the signals into two dimensions, make these spectra
some of the most convenient to analyze. The spectrometer time needed
for each experiment can be as short as 10 min, provided the sample is
sufficiently concentrated, making it feasible to perform a detailed titration.
The value of the Kd can be measured without doing any assignment of
signals in the NMR spectrum to particular atoms in the protein. If an as-
signment is available, however, it is possible to identify the amino acids
in the protein that are likely to be involved in the interaction. This is simply
done by looking to see if a signal corresponding to a given residue changes
its chemical shifts on addition of the ligand. The chemical shift is, however,
a very sensitive marker and even small structural rearrangements upon
binding can in some cases give rise to significant changes for residues far
from the actual binding site. A good example of this is the chromo shadow
domain where the addition of the PXVXL-containing peptide ligand
breaks the symmetry and alters the structure of the dimer causing two
thirds of the backbone amide signals to shift in the NMR spectra.7 In these
cases the binding site can normally still be identified by ignoring small
changes, and, if the structure is known, by only considering candidates that
are known to be surface exposed.5
Histone H3 Binding to the HP1b Chromodomain

As part of the initial characterization of histone H3 peptide binding to
HP1 we mapped the binding site using the approach described earlier. A
1-mM sample of uniformly 15N-labeled HP1 chromodomain was prepared
in 20 mM Na phosphate of pH 7.0, with 150 mM NaCl and 10% D2O. A
2D 1H–15N HSQC spectrum was recorded on a Bruker DRX600 spectrom-
eter (Bruker, Karlsruhe) at 298 K, collecting 4 scans for each of the 150
time points in the indirect dimension. The spectral width was set to
10,000 Hz in the proton dimension and 2000 Hz in the indirect nitrogen di-
mension. Titration experiments are performed by gradually adding aliquots
of concentrated solutions of ligand to the NMR sample. In cases where the
solubility of the ligand makes it impossible to achieve very high concentra-
tions the effect of dilution of the NMR sample can be alleviated by having
the chromodomain present in the ligand stock solution at the NMR sample
concentration. Additions are performed either using a spatula that will fit
in the NMR tube or by transferring the solution to an Eppendorf tube.
33
G. Bodenhausen and D. J. Ruben, Chem. Phys. Lett. 69, 185 (1980).
34
M. Piotto, V. Saudek, and V. Sklenar, J. Biomol. NMR 2, 661 (1992).
35
V. Sklenar, M. Piotto, R. Leppik, and V. Saudek, J. Magn. Reson. Series A 102, 241 (1993).
For each addition the sample is transferred to an Eppendorf tube and

mixed with peptide solution before being transferred back into the NMR
sample tube. An experiment identical to the one described earlier was then
recorded. Most signals were in the fast exchange regime. This means that
the titration can be followed by changes in chemical shift, and the Kd
of the interaction can be obtained by fitting those chemical shift changes
to the same function (1) as is used for steady state fluorescence data. Pep-
tide solutions were prepared from lyophilized material and the concentra-
tions were determined by amino acid analysis. (Note: We find that it is
possible to get a reasonable estimate of the amount of the peptide by
weighing it prior to dissolving it in the buffer and then dividing the mass
obtained by two to account for the residual water.)
Identifying the Binding Site
Figure 5 shows the first and last point of such a titration. The spectrum
had been assigned for the free chromodomain6 and, since the exchange
regime was at the fast exchange limit, the signals for the complex could
Fig. 5. 2D 1H-15N HSQC spectra identifying the peptide binding site. Signals from the 15N-
labeled HP1 chromodomain (10–80) without (black) and with (grey) a 2:1 excess of a
bis-(Lys-4,9) di-methylated histone H3 (1–18) peptide: (NH2 -ARTK(Me)2 QTARK
(Me)2STGGKAPGG-COOH). Spectra recorded on samples where the added peptides have
bis-tri-methylated (Lys-4,9) or tri-methylated Lys-9 give similar results.
Fig. 6. Chemical shift differences for the 15N (open bars) and 1HN (filled bars) atoms of
the HP1 chromodomain, in the presence of histone H3, were determined from the spectra
in Fig. 5.
be assigned by following them as they moved when the ligand concentra-

tion was gradually increased:
obs ¼ pfree free þ pcomplex complex ;
where pfree and pcomplex are the molar fractions of free protein and protein
in the complex, respectively, obs is the measured chemical shift, free is the
chemical shift of the free form, and complex is the chemical shift in the com-
plex. These assignments were later confirmed by sequential assignment.15
This meant that it was possible to assign a value for the change in the chem-
ical shift values () for most amino acid residues in the chromodomain
(Fig. 6). As can be seen from Fig. 7 this enabled us to confirm that the
peptide binds in the region originally proposed based on the structural
similarities between the chromodomain and Sac7d.6
A similar experiment was performed using the unmodified histone H3
peptide (data not shown). From the changes in the spectrum it can be seen
that the unmodified peptide interacts with the same binding site as the
methylated species. It is clear, however, that the interaction is very weak.
Even with a 2.5-fold excess of the peptide the average chemical shift is still
only approximately halfway between the free form and the shift for
Fig. 7. The histone H3 binding site mapped onto the structure of the HP1
chromodomain using the chemical shift differences plotted in Fig. 6. The chromodomain
is shown as a ribbon cartoon with the peptide represented as a ball-and-stick model. The
peptide backbone is shown in white, whereas the side chain atoms are darker. Black regions of
the chromodomain ribbon show changes of more than 1 ppm. Lighter regions show smaller
changes. The figure was prepared using Molscript [P. J. Kraulis, J. Appl. Crystallogr. 24,
946 (1991)].
the complex with Lys-9 methylated peptides. Since the experiments

were conducted at 0.1 mM this indicates that the Kd for the interaction
is around 0.2 mM, explaining why it was not possible to detect this
interaction in pull-down assays.12,13
It was possible to characterize the binding site further. The same 2D
1
H-15N HSQC experiment was performed on a complex between the chro-
modomain and a peptide that was tri-methylated at Lys-9. In this case the
chemical shifts were compared to the values from the sample that was di-
methylated at Lys-9. Only a subset of the residues was further affected by
the presence of the extra methyl group. These were Tyr-21, Trp-42, and
residues 47–51. This suggested that the methyl lysine binding site was situ-
ated primarily in the loop between the second and third -strand. When the
structure was solved we were able to confirm that residues 21, 42, and 51
are indeed some of the main residues that interact with the methyl lysine
side chain.
[11] quantitative assays for deacetylases 171
[11] Quantitative Assays for Characterization of the

Sir2 Family of NADþ-Dependent Deacetylases
By Margie T. Borra and John M. Denu
Introduction
Within nucleosomes, histones are targets of enzyme-catalyzed post-
translational modifications such as acetylation, phosphorylation, methyla-
tion, and ubiquitination.1–4 Dynamic modification of nucleosomes regulates
processes such as DNA synthesis, DNA repair, and transcription. Among
these modifications, reversible acetylation has emerged as a major mechan-
ism in controlling these processes. Hyperacetylation has been correlated
with transcriptionally active genes, while hypoacetylation has been corre-
lated with silenced loci. Acetylation is carried out by histone acetyltrans-
ferases, which catalyzes the transfer of the acetyl group from acetyl-CoA
to the e-amino group of lysine residues.5,6 Histone/protein deacetylases
catalyze the removal of the acetyl group from the lysine residues.7,8
Histone/protein deacetylases have been classified based on their simi-
larity to the yeast enzymes: the Rpd3-like (class I), the HDA-like (class
II), and the Silent Information Regulator 2 (Sir2)-like (class III) deacety-
lases.7–9 Both class I and class II, which are conserved in many eukaryotes
and are commonly referred to as HDACs, likely catalyze the removal of
the acetate group from the acetylated lysine residue by activation of a
water molecule for direct hydrolysis,10 generating free acetate and the de-
acetylated protein (Fig. 1). Both class I and class II have been found in
large multiprotein complexes that are recruited by DNA-binding proteins
to specific-chromatin domains. Class I and class II enzymes are sensitive to
the inhibitor Trichostatin A (TSA).7,8
1
J. Wu and M. Grunstein, Trends Biochem. Sci. 25, 619 (2000).
2
K. Luger and T. J. Richmond, Curr. Opin. Genet. Dev. 8, 140 (1998).
3
A. P. Wolffe and J. J. Hayes, Nucleic Acids Res. 27, 711 (1999).
4
P. Cheung, C. D. Allis, and P. Sassone-Corsi, Cell 103, 263 (2000).
5
M. H. Kuo and C. D. Allis, Bioessays 20, 615 (1998).
6
S. Y. Roth, J. M. Denu, and C. D. Allis, Annu. Rev. Biochem. 70, 81 (2001).
7
S. G. Gray and T. J. Ekstrom, Exp. Cell Res. 262, 75 (2001).
8
C. M. Grozinger and S. L. Schreiber, Chem. Biol. 9, 3 (2002).
9
S. Khochbin, A. Verdel, C. Lemercier, and D. Seigneurin-Berny, Curr. Opin. Genet. Dev.
11, 162 (2001).
10
M. S. Finnin, J. R. Donigian, A. Cohen, V. M. Richnon, R. A. Rifkind, P. A. Marks,
R. Breslow, and N. P. Pavletich, Nature 401, 188 (1999).

Fig. 1. Comparison of the HDAC and Sir2-like reactions. (A) Class I and class II HDACs
likely activate a water molecule for direct hydrolysis of the acetyl group from the acetylated
substrate to form deacetylated protein and free acetate as products. (B) The Sir2 family
of deacetylases catalyzes the tight coupling of NADþ cleavage and substrate deacetylation
to produce nicotinamide, deacetylated product, and O-acetyl-ADP-ribose (OAADPr).
Among the deacetylases, the Sir2 family (class III) of enzymes, also re-
ferred to as sirtuins,11,12 are unique, as they are TSA-insensitive and
NADþ-dependent.8,13–17 These enzymes are conserved in many organisms
ranging from some bacteria to humans.12 The founding member of this
family, yeast Sir2, has been found in multiprotein complexes and is

required for gene silencing in yeast.18 Sir2-like enzymes tightly couple the
cleavage of NADþ and the deacetylation of a substrate to produce nico-
tinamide, deacetylated product, and the novel compound, O-acetyl-ADP-
ribose (OAADPr)16,19–22 (Fig. 1). Stoichiometric analysis of the enzymatic
reaction showed that for every mole of NADþ cleaved and substrate de-
acetylated, 1 mol of nicotinamide, deacetylated product, and OAADPr
are produced.16,19,23,24 The enzymatically derived OAADPr has been
reported to be 20 -O-acetyl-ADPribose (20 OAADPr),20–22 which through
an intramolecular transesterification reaction, equilibrates with
30 OAADPr in solution.20,21 Although recent investigations demonstrated
OAADPr bioactivity and OAADPr metabolism by specific nuclear and
cytoplasmic enzymes, the physiological functions for this metabolite
remain to be established.24,25
Although histones are considered targets for so-called histone deacetylases,
enzymes within these deacetylase families also target nonhistone proteins. Re-
cently, various acetylated proteins such as p53,26,27 the transcription factor
11
R. A. Frye, Biochem. Biophys. Res. Commun. 260, 273 (1999).
12
R. A. Frye, Biochem. Biophys. Res. Commun. 273, 793 (2000).
13
S. Imai, C. M. Armstrong, M. Kaeberlein, and L. Guarente, Nature 403, 795 (2000).
14
J. S. Smith, C. B. Brachmann, I. Celic, M. A. Kenna, S. Muhammad, V. J. Starai, J. L.
Avalos, J. C. Escalante-Semerena, C. Grubmeyer, C. Wolberger, and J. D. Boeke, Proc.
Natl. Acad. Sci. USA 97, 6658 (2000).
15
J. Landry, A. Sutton, S. T. Tafrov, R. C. Heller, J. Stebbins, L. Pillus, and R. Sternglanz,
Proc. Natl. Acad. Sci. USA 97, 5807 (2000).
16
K. G. Tanner, J. Landry, R. Sternglanz, and J. M. Denu, Proc. Natl. Acad. Sci. USA 97,
14178 (2000).
17
J. M. Denu, Trends Biochem. Sci. 28, 41 (2003).
18
D. Moazed, Mol. Cell 8, 489 (2001).
19
J. C. Tanny and D. Moazed, Proc. Natl. Acad. Sci. USA 98, 415 (2001).
20
A. A. Sauve, I. Celic, J. Avalos, H. Deng, J. D. Boeke, and V. L. Schramm, Biochemistry 40,
15456 (2001).
21
M. D. Jackson and J. M. Denu, J. Biol. Chem. 277, 18535 (2002).
22
J. H. Chang, H. C. Kim, K. Y. Hwang, J. W. Lee, S. P. Jackson, S. D. Bell, and Y. Cho,
J. Biol. Chem. 277, 34489 (2002).
23
J. Landry, J. T. Slama, and R. Sternglanz, Biochem. Biophys. Res. Commun. 278, 685
(2000).
24
M. T. Borra, F. J. O’Neill, M. D. Jackson, B. Marshall, E. Verdin, K. R. Foltz, and J. M.
Denu, J. Biol. Chem. 277, 12632 (2002).
25
L. A. Rafty, M. T. Schmidt, A. L. Perraud, A. M. Scharenberg, and J. M. Denu, J. Biol.
Chem. 277, 47114 (2002).
26
J. Luo, A. Y. Nikolaev, S. Imai, D. Chen, F. Su, A. Shiloh, L. Guarente, and W. Gu, Cell
107, 137 (2001).
27
H. Vaziri, S. K. Dessain, E. Ng Eaton, S. I. Imai, R. A. Frye, T. K. Pandita, L. Guarente,
and R. A. Weinberg, Cell 107, 149 (2001).
TAFI68,28 acetyl-CoA synthetase,29 and tubulin30 have been shown to be

substrates for some Sir2-like enzymes in vitro and/or implicated in vivo.
Differences in the chemistry of Sir2- and HDAC-catalyzed reactions
necessitate the utilization of distinct assays, as approaches commonly used
for evaluation of HDAC activity may not be applicable for characterization
of Sir2-like enzymes. In this chapter, we will briefly overview general HDAC
assays, then detail methodologies for quantitative analysis of the Sir2-
catalyzed reaction. These quantitative assays permit steady-state kinetic
analysis, which can be used for comparison of different Sir2 homologues, de-
termination of the enzyme’s kinetic mechanism, examination of substrate
specificity, and analysis of enzyme inhibitors. High-performance liquid chro-
matography (HPLC)-based, charcoal-binding and TLC-based assays are
used to monitor and quantify the substrates and/or products of the reaction.
Enzyme-catalyzed nicotinamide-NADþ exchange reaction, in which
nicotinamide condenses with an enzyme-ADP-ribose–like intermediate to
regenerate the substrate NADþ and acetylated protein, allows for assess-
ment of the catalytic step preceding the formation of the deacetylated prod-
uct and OAADPr. Given that nicotinamide is a strong product inhibitor of
the Sir2-catalyzed reaction, the outlined nicotinamide inhibition analyses
can be used to distinguish the Sir2 family of deacetylases.
General Approaches for Monitoring HDAC Activity
Detection of HDAC-Produced Acetate

Detection of the liberated acetate using ethyl acetate extraction is a
common method of monitoring HDAC activity.31–37 The reactions typic-
ally involve the incubation of the enzyme with radiolabeled-acetylated
28
V. Muth, S. Nadaud, I. Grummt, and R. Voit, EMBO J. 20, 1353 (2001).
29
V. J. Starai, I. Celic, R. N. Cole, J. D. Boeke, and J. C. Escalante-Semerena, Science 298,
2390 (2002).
30
B. J. North, B. L. Marshall, M. T. Borra, J. M. Denu, and E. Verdin, Mol. Cell 11, 437 (2003).
31
D. Kolle, G. Brosch, T. Lechner, A. Lusser, and P. Loidl, Methods 15, 323 (1998).
32
S. Imiliani, W. Fischle, C. Van Lint, Y. Al-Abed, and E. Verdin, Proc. Natl. Acad. Sci. USA
95, 2795 (1998).
33
S. J. Darkin-Rattray, A. M. Gurnett, R. W. Myers, P. M. Dulski, T. M. Crumley, J. J.
Allocco, C. Cannova, P. T. Meinke, S. L. Colletti, M. A. Bednarek, S. B. Singh, M. A.
Goetz, A. W. Dombrowski, J. D. Polishook, and D. M. Schmatz, Proc. Natl. Acad. Sci. USA
93, 13143 (1996).
34
M. J. Hendzel, G. P. Delcuve, and J. R. Davie, J. Biol. Chem. 266, 21936 (1991).
35
R. Sendra, I. Rodrigo, M. L. Salvador, and L. Franco, Plant Mol. Biol. 11, 857 (1988).
36
A. Kervabon, J. Mery, and J. Parello, FEBS Lett. 106, 93 (1979).
37
J. Taunton, C. A. Hassig, and S. L. Schreiber, Science 272, 408 (1996).
histones or peptides. The reaction is quenched by addition of acid (HCl/

acetic acid) to denature the enzyme, and the radiolabeled acetate is ex-
tracted using the organic solvent ethyl acetate. Radioactivity of the organic
phase is determined by scintillation counting. Although ethyl acetate ex-
traction has been successfully used to determine general HDAC activity,
its applicability is limited to class I and class II HDACs, in which acetate
is a direct product of the reaction (Fig. 1A). For the Sir2-like–catalyzed
reaction, the acetyl group is transferred to the ADP-ribose portion of
NADþ to generate OAADPr (Fig. 1B), which cannot be extracted in ethyl
acetate (Borra and Denu, unpublished data).
Detection of Acetylated and Deacetylated Substrates

Besides ethyl acetate extraction, general histone/protein deacetylase
activity has been analyzed by various approaches such as gel electrophor-
esis, western blot analysis, and filter-binding assays. The western blot
analysis utilizes antibodies that recognize the acetylated forms of histones
or other substrates.38,39 Reaction mixtures are loaded onto SDS-PAGE
gels and subjected to western blotting using antibodies against the acety-
lated forms of proteins. Using radiolabeled, acetylated protein/histone,
SDS-PAGE gels can also be dried and exposed to photographic film, after
which the loss of radioactivity on the substrate can be visualized.15,40 Alter-
natively, the reaction mixture can be spotted onto a Whatman P81 cation
exchange paper, which traps the radiolabeled, acetylated protein
substrate.15,40 After extensive washing, the paper is subjected to scintilla-
tion counting, and a decrease in the counts reflects the level of deacetyla-
tion.15,40 Another alternative is to quench and precipitate the proteins by
adding 20% trichloroacetic acid (TCA). Samples can be filtered through
a glass filter on a vacuum manifold. The precipitated proteins bind to
the filter, and radioactivity in the filter is determined by scintillation
counting.40
Other less utilized approaches that rely on detection of the deacetylated
protein product have also been developed and include the scintillation
proximity assay, and nonisotopic, fluorescence-based assays. The scintilla-
tion proximity assay does not rely on the separation of substrates and
products and is therefore suitable for higher throughput screening.41 For
38
Y. Zhang, G. LeRoy, H. P. Seelig, W. S. Lane, and D. Reinberg, Cell 95, 279 (1998).
39
Y. Zhang, N. Li, C. Caron, G. Matthias, D. Hess, S. Khochbin, and P. Matthias, EMBO
J. 22, 1168 (2003).
40
J. Landry and R. Sternglanz, Methods 31, 33 (2003).
41
B. Nare, J. J. Allocco, R. Kuningas, S. Galuska, R. W. Myers, M. A. Bednarek, and D. M.
Schmatz, Anal. Biochem. 267, 390 (1999).
this assay, the substrate used is a biotinylated [3H]acetyl histone H4

peptide, which was shown to bind to streptavidin-coated SPA beads
impregnated with scintillant. Binding of the [3H]acetyl peptide stimulates
the light emission via the scintillant, which can be detected using a scintil-
lation counter. Decrease in the radioactive signal indicates deacetylation
of the substrate. The fluorescence-based deacetylase assay relies on the
release of a fluorophore on deacetylation of the substrate. One fluorogenic
assay relies on the use of N-(4-methyl-7-coumarinyl)-N--(tert-butyloxy-
carbonyl)-N-V-acetyllysinamide (MAL) as a substrate.42 Upon substrate
deacetylation, the fluorogenic product is separated from the reaction
mixture and quantitated using a reversed-phase HPLC. A more recently
developed fluorogenic assay is based on ability of trypsin to cleave
deacetylated substrates.43 For this assay, the peptide substrates have
the acetylated lysine residue attached to a 4-methylcoumarin-7-amide
(MCA) moiety at the carboxy terminus. Upon substrate deacetylation,
trypsin cleaves the peptides proximal to the lysine residues, releasing the
MCA moiety, whose fluorescence can be monitored at em ¼ 460 nm
(ex ¼ 390 nm).
Although the scintillation proximity and the nonisotopic, fluorescence-
based assays allow for high-throughput analysis, the substrates (biotiny-
lated [3H]acetyl histone H4 peptide, MAL, and protein-MCA complex)
used for these assays are not commercially available and may therefore
limit their general use.
Approaches for Characterization of the Sir2 Family of Deacetylases

In this section, we will describe various quantitative approaches for
specifically monitoring the enzymatic activity of the Sir2 family of
NADþ-dependent histone/protein deacetylases. First, a method for produc-
ing and quantifying a [3H]acetylated substrate will be described. Then
quantitative methods for monitoring deacetylase activity of Sir2-like
enzymes will be discussed. Using these methods, products formed during
the Sir2-catalyzed reaction can be quantified and the initial velocity of
the reaction can be determined. These approaches can, therefore, be util-
ized for steady-state kinetic analysis. Methods that are applicable to the
class I and class II deacetylases will be indicated where appropriate. Nico-
tinamide-utilizing assays that are specific for the Sir2 family of deacetylases
will also be discussed.
42
K. Hoffmann, G. Brosch, P. Loidl, and M. Jung, Nucleic Acids Res. 27, 2057 (1999).
43
D. Wegener, F. Wirsching, D. Riester, and A. Schwienhorst, Chem. Biol. 10, 61 (2003).
Production and Quantitation of [3H]acetylated H3 Peptide

The acetylated histone H3 peptide is a good substrate for many Sir2
homologues.13,15,16,21,23,24 This section provides a detailed protocol for the
p300/CBP-associating factor (P/CAF)-catalyzed lysine-14 [3H]acetylation
of the H3 peptide, corresponding to the N-terminal residues of histone
H3.24 The reaction is carried out in the presence of 10–15 M
[2-3H3]acetyl-CoA (NEN Life Science Products), 250 M H3 peptide,

6–8 M P/CAF in 100 mM Tris, pH 7.5, with 1 mM DTT at 37 for 1 h.
The reaction is quenched by the addition of trifluoroacetic acid (TFA) to
a final concentration of 1%. The quenched reaction is injected onto a re-
versed-phase HPLC column (e.g., Vydac C18 Small Pore column, 10
250 mm, 201SP1010) to resolve the acetylated and unacetylated H3 pep-
tide, which can be monitored by measuring the absorbance at 214 nm.
The HPLC is run with 100% solvent A (0.05% TFA/H2O) for 1 min
followed by a linear gradient of 0–40% solvent B (0.02% TFA in acetoni-
trile) over 40 min. Fractions are collected, and radioactivity is determined
by scintillation counting. Fractions containing the [3H]acetylated H3 pep-
tide ([3H]AcH3) are pooled and lyophilized to dryness. The lyophilized
[3H]AcH3 can be resuspended in water or buffer of choice.
To quantify the [3H]AcH3 produced, the radioactivity of 3–5 l ali-
quots of the resuspended [3H]AcH3 is determined by scintillation counting.
Because there is a 1:1 stoichiometry between the moles of [3H]acetyl-CoA
consumed and the moles of [3H]AcH3 produced, the total moles of
[3H]AcH3 in solution can be determined by dividing the CPM of the pep-
tide aliquots by the specific activity (CPM/mol) of the [3H]acetyl-CoA
stock solution and taking into account the dilution factor.
If nonradiolabeled peptides are desired, the amount of the synthetic
acetylated peptide in the stock can be determined using several methods.
To determine the amount of peptide present in the sample, the peptide sub-
strate can be subjected to amino acid analysis, which is usually performed
after peptide synthesis when the peptide is purchased commercially. If not,
the amino acid analysis can also be performed by most university protein
core facilities. Alternatively, Sir2 enzymes, which catalyze a 1:1 stoichiom-
etry between NADþ cleaved and substrate deacetylated, can also be
employed for quantification of stock peptides. A limited amount of peptide
and a known amount of NADþ can be reacted with Sir2-like enzymes until
complete substrate deacetylation is achieved. The amount of acetyl sub-
strate present can be calculated from determining the amount of NADþ
consumed after complete deacetylation of a sub-stoichiometric amount of
acetyl substrate. If [32P]NADþ is utilized in the assay, remaining NADþ
can be quantified by measuring the radioactivity. Alternatively, remaining
NADþ can be quantified using commercially available NADþ-dependent

dehydrogenases such as pyruvate dehydrogenase44 or -ketoglutarate
dehydrogenase.45 These NADþ-dependent dehydrogenases convert NADþ
to NADH, which can be spectrophotometrically monitored at 340 nm.
Subtracting the concentration of remaining NADþ from the initial concentra-
tion yields the amount of NADþ consumed and, thus, the initial concen-
tration of the acetylated peptide substrate. A time course of deacetylation
may be necessary to ensure complete conversion of peptide substrate.
Deacetylase Assays
HPLC-based Deacetylase Assay. This assay relies on the separation of
substrates and products of the deacetylase reaction by a reversed-phase
HPLC. Quenched reaction mixtures are injected onto a C18 column and,
using a gradient of increased levels of organic solvent, substrates, products,
and enzyme can be resolved. This assay is applicable to all three classes
of deacetylases.
For characterization of most Sir2-like enzymes, monoacetylated H3 and
H4 peptides, corresponding to the 20 N-terminal residues of histone H3

and H4, can be used. The 110-l reactions are carried out at 37 in
þ
50 mM Tris (or phosphate), pH 7.5, with 1 mM DTT. NAD and acety-

lated peptide are mixed and pre-incubated in a 37 water bath for 5 min.
þ
Typical concentrations of NAD and acetylated peptide have ranged from
0.25 M to 1 mM; however, the range should be determined empirically, as
the Km values for these substrates may differ by orders of magnitude
depending on the Sir2 homologue (the kinetic parameter Km is discussed
in the Kinetic Analysis of the Sir2-Mediated Reaction section). The reaction
is initiated by the addition of enzyme. Although typical enzyme concentra-
tions ranging from 0.25 to 0.5 M and reaction times ranging from 1 to
15 min have been used, these parameters should be empirically deter-
mined. The concentration of enzyme and reaction time should be adjusted
such that less than 10–15% of the substrate is converted to products to
maintain steady-state conditions. The reaction is quenched by the addition
of TFA to a final concentration of 1%. Quenched samples are kept on ice
or stored in 20 if not immediately injected onto the HPLC column.
Samples are injected onto a reversed-phase HPLC column (e.g., a Vydac
C18 column, 4.6 250 mm, 201SP104) to resolve substrates and products.
A 100-l loop is typically used for the injections and the flow rate is set
to 1 ml/min. After injection, the system is run isocratically with solvent A
44
Y. Kim, K. G. Tanner, and J. M. Denu, Anal. Biochem. 280, 308 (2000).
45
G. A. Hunter and G. C. Ferreira, Anal. Biochem. 226, 221 (1995).
(0.05% TFA/H2O) for 1 min followed by increasing levels of solvent B

(0.02% TFA in acetonitrile). The gradient used for each assay may vary
depending on the type of peptide substrate used. For efficient separation
of reactions containing the 20-mer AcH3 N-terminal peptides, a gradient
of 0–20% B over 20 min is used. For efficient resolution of reactions con-
taining the 20-mer H4 peptides, the following gradient is used: 0–10% solv-
ent B for 4 min followed by 10–25% B for 25 min. Following each run, the
column is washed with 100% B for 3–5 column volumes followed by re-
equilibration with solvent A for 3–5 column volumes. Elution of substrates
and products is monitored by measuring the absorbance at 214 nm (to
monitor all substrates and products) or at 260 nm (to specifically monitor
nicotinamide, NADþ, and OAADPr). Using the above gradients, a good
resolution between the monoacetylated and deacetylated peptides can be
achieved. Deacetylated H3 peptide typically elutes at 16 min, acetylated
H3 at 18 min, deacetylated H4 at 15 min, and acetylated H4 at 17 min.
Substrates and products elute at the approximate percentages of solvent
B: nicotinamide at 5%, deacetylated H3 at 16%, deacetylated H4 at
13%, acetylated H3 at 18%, and acetylated H4 at 14%, NADþ at 12%,
and OAADPr at 8%. The areas of the peaks are integrated for quantifica-
tion. To calculate the percent deacetylation, the area of the deacetylated
peptide peak is compared to the combined areas of the acetylated and de-
acetylated peptide peaks. Because a known amount of acetylated peptide is
used, the percentage of the deacetylation is then used to determine the
amount of deacetylated product formed over the particular time of the
assay, to obtain an initial rate.
This HPLC-based assay is also applicable for separating radiolabeled
substrates and products. For example, [3H]acetylated histone peptides or
[14C]- or [32P]-labeled NADþ can be utilized. [14C]NADþ can be synthe-
sized using the Sir2-catalyzed nicotinamide-NADþ exchange procedure
(described later) or purchased commerically, and [32P]NADþ can be
obtained from NEN Life Science Products (800 Ci/mmol). The amounts
of the radiolabeled substrates and products can be monitored and quanti-
fied by collecting the fractions eluted from the HPLC, adding a constant
volume of each fraction into scintillation vials and determining the radio-
activity of each fraction by scintillation counting. To calculate the amounts
of product formed, the radioactivity of the product can be divided by the
total radioactivity of all the fractions collected, which corresponds to the
radioactivity of the substrate prior to the reaction. The percent product is
then multiplied by the concentration of radiolabeled substrate to obtain
the concentration of product formed. Alternatively, the total radioactivity
of the product can be divided by the specific activity (CPM/mol) of the
radiolabeled substrate to obtain the moles of product formed.
One of the limitations of this HPLC-based assay is its relatively time-

consuming nature, where each HPLC run can take approximately 1 h.
Only a limited number of injections can be performed over a typical
working day. Use of an autoinjector can greatly facilitate the analysis of a
large number of samples. Another limitation of this assay is the detector’s
inability to detect low concentrations of substrates when absorbance is
used to quantitate levels of deacetylation. With low peptide concentrations
(below 5 M) larger injection volumes (e.g., 2 ml) are required in order to
have sufficient substrate and products to accurately detect. Moreover, be-
cause the retention times may change over column useage, it is necessary
to check the elution of standards routinely.
Charcoal-binding Deacetylase Assay. The charcoal-binding assay is
based on the idea of differential binding of substrates and products to acti-
vated charcoal; acetylated and deacetylated peptides, NADþ, nicotina-
mide, OAADPr, and enzyme will bind to activated charcoal. The
charcoal-binding assay is performed using [3H]acetylated substrate. This
assay takes advantage of the fact that under high pH (9.5) and heat, the
[3H]acetyl group from [3H]OAADPr is hydrolyzed. The [3H]acetate does
not bind to charcoal and can, therefore, be separated from the rest of the
charcoal-bound substrates and products. This charcoal-binding assay
allows for hydrolysis of the pH labile ester of OAADPr and is similar in
basic principle to the modified ethyl acetate extraction method developed
by Landry and Sternglanz,40 where OAADPr hydrolysis by NaOH
precedes ethyl acetate extraction in acid.
The charcoal-binding assay is performed by reacting [3H]acetylated
peptide/protein, NADþ, and enzyme in 50 mM Tris (or phosphate) pH

7.5, with 1 mM DTT in an 80-l reaction at 37 . A control solution where
no enzyme is added should be included. To quench the reaction, 70 l of
the reaction mixture is added into an aliquot (typically 70–100 l) of char-
coal slurry (1/3 w/v of charcoal (Sigma, C-5260) and 2/3 of 2 M glycine, pH

9.5). The mixture is then heated at 95 for 1 h to hydrolyze the [3H]acetate
3
from [ H]OAADPR. The samples are centrifuged and a fix volume (typic-
ally 100 l) of the supernatant, which contains the [3H]-acetate, is trans-
ferred into another aliquot of charcoal slurry (determination of the
amount/volume needed is discussed later). The addition of the supernatant
to the second charcoal slurry is performed to remove the residual peptides
that may not have bound to the first charcoal slurry. The samples are
mixed, centrifuged, and the supernatant containing the [3H]-acetate is
transferred into a clean tube. The samples are then centrifuged to remove
residual charcoal from the solution, and a constant volume of each super-
natant is transferred into a scintillation vial containing scintillation cocktail.
The radioactivity from the supernatant corresponds to the [3H]OAADPr
produced. In order to determine the specific activity (CPM/mol) of the

acetylated peptide in the reaction, 3–5 l of the reaction mixture is placed
into a scintillation vial and 2 M glycine, pH 9.5, buffer is added up to the
same volume as the volume of supernatant. This is done to minimize differ-
ences in the quenching during scintillation counting. The CPM obtained
from the reaction mixture is divided by the moles of acetylated peptide
to obtain the specific activity (CPM/mol). Given the 1:1 stoichiometry
between the amount of substrate deacetylated and products formed, divid-
ing the CPM of the supernatant by the specific activity of the peptide and
taking into account the dilution would yield the amount of [3H]OAADPr
produced. For this assay, it is important to note the volumes of the charcoal
slurry and the volume of the supernatant transferred after the first spin to
account for the dilution.
Prior to detailed analyses, the volume of charcoal slurry needed should
be determined. The amount of charcoal to be used in the assay depends on
the concentrations of substrates employed. To determine the volume or
amount of charcoal needed for a given concentration of substrate, 80-l
control solutions are prepared with the desired concentrations of
[3H]acetylated substrate and NADþ in 50 mM Tris (or phosphate), pH
7.5, with 1 mM DTT. In these control solutions where no enzyme is added,
no [3H]OAADPR is produced. A 70-l aliquot from each control solution
is added into varying volumes or amounts of charcoal slurry. Samples are
treated as described earlier. Because no [3H]OAADPR is expected to
form, only background counts should be present in the supernatant. The
volume or amount of charcoal slurry that produces no significant radioactiv-
ity in the supernatant is desirable for this method. Significant counts in the
supernatant could be indicative of unbound [3H]acetylated substrate; the
concentrations of substrates are higher than the capacity of the charcoal.
Also, the [3H]acetylated peptide will not normally liberate free acetate
under the conditions of this assay, but the presence of free [3H]acetate
in the stock solution may be observed, in which case, the radioactivity
from the control solutions can be subtracted from the reaction samples.
Comparison of the charcoal-binding assay and the HPLC-based assay
showed that the two approaches are in excellent agreement. Charcoal-
binding and HPLC-based assays were utilized in an NADþ saturation
experiment, in which 5 M of the 20-mer AcH3 peptide and increasing
NADþ concentrations (0.25–20 M) were reacted with in the presence of
25 nM HST2. The amount of products formed was quantified as described
earlier, and initial velocity data were fitted into the Michaelis-Menten
equation to obtain the kinetic parameters Km, Vmax, and Vmax/Km (these
kinetic parameters will be discussed later). Using the charcoal-binding
assay, the Km and Vmax values are 2.5 0.1 M and 0.18 0.003 s1,
respectively. Using the HPLC-based assay, the Km and Vmax values are
2.8 0.4 M and 0.24 0.01 s1, respectively. A typical saturation curve
using the charcoal-binding assay is illustrated in Fig. 2.
TLC-based Deacetylase Assay. This assay is based on the ability to re-
solve small coenzyme metabolites using thin-layer chromatography. Anal-
yses of Sir2 activity using TLC have been previously described.15,19 Sir2
reactions are performed in the presence of radiolabeled NADþ; selection
of the label’s location depends on the desired molecule to be monitored.
To monitor the formation of nicotinamide specifically, a [14C]NADþ in
which the [14C]-label is located in the nicotinamide moiety should be used.
To monitor the formation of OAADPr or ADP-ribose, a [32P]NADþ, in
which the label is located in one of the phosphate groups, should be used.
To perform this assay, the radiolabeled NADþ and the acetylated substrate

are mixed and pre-incubated in a 37 water bath for 5 min. The reaction is
initiated by addition of the enzyme and is quenched with TFA to a final
concentration of 1%. An aliquot of 3–5 l of the reaction mixture, contain-
ing approximately 1000–10,000 CPM, is spotted onto a TLC plate (What-
man aluminum-backed silica gel plates, 250-m layer), approximately
2 cm from the bottom. Samples are spotted with at least 1 cm distance from
each other. To increase the radioactivity, it may be necessary to apply a
small volume at a time and dry the plate prior to spotting additional
volumes. The TLC plate is dried prior to placement in the chamber with
Fig. 2. NADþ saturation analysis using the charcoal-binding assay. NADþ saturation of

the Sir2-like reaction rate was performed at 37 in the presence of varying NADþ concentrations
(0.25–20 M), 25 nM HST2, a yeast Sir2 homologue, 5 M AcH3(K14), and 1 mM DTT in

50 mM Tris, pH 7.5 (at 37 ). The reaction mixture was quenched by addition of the charcoal
slurry. The product formed was quantified, and the initial velocities of the reaction at various
NADþ concentrations were plotted and fitted using the Michaelis-Menten equation.
solvent containing certain percentages of ethanol and ammonium acetate

(or any desired solvent). Typical solvents used in this assay are 80:20 or
70:30 mixture of ethanol:2.5 M ammonium acetate.15,19 After placement
of the plate into the chamber, the chamber is sealed, and the chromatog-
raphy is run until the solvent front is approximately an inch from the top
of the plate to achieve a good resolution between NADþ, nicotinamide,
and OAADPr. Once the run is completed, the TLC plate is removed from
the chamber and allowed to air dry. The plate is exposed to a phosphoima-
ging screen or photographic film, for at least 16 h. Using a densitometer or
imager, the level of radioactivity of the substrate and product spots is deter-
mined, and the amounts of substrates converted and products formed are
quantified. Using the 80:20 solvent, the following Rf values were obtained:
0.75 for nicotinamide, 0.29 for NADþ, and 0.58 for OAADPR. Using the
70:30 solvent, the following Rf values were obtained: 0.84 for nicotinamide,
0.49 for NADþ, 0.74 for OAADPR, and 0.69 for ADPR.
This TLC-based assay is fast and convenient to use and multiple
samples can be applied per plate. Another advantage of using this ap-
proach is that the reaction can be performed using various acetylated sub-
strates. Using this TLC-based assay, we have looked at the ability of
various Sir2-like enzymes to deacetylate 5-mer, 7-mer, and 9-mer versions
of an H3 peptide, containing acetylated lysine-14. This approach is also
applicable for high-throughput assays, as multiple samples can be applied
onto a single TLC plate. Because this assay relies on the detection of
radioactive NADþ, nicotinamide, and OAADPr, this assay is not
applicable for the class I and class II HDACs.
Nicotinamide-Utilizing Assays
Nicotinamide-NADþ Catalyzed Exchange Assay. This assay was first

described by Landry et al.15,23 and is specific for the Sir2 family of deacety-
lases. The assay is performed in the presence of NADþ, acetylated substrate,
enzyme, and [carbonyl-14C]nicotinamide. Sir2-like enzymes follow a se-
quential mechanism in which both NADþ and acetylated substrate have to
bind prior to any catalytic step (Borra et al., manuscript in preparation).
Upon binding of both substrates, NADþ is cleaved, nicotinamide is released,
and an ADP-ribose–like intermediate is formed. The exogenous [14C]nico-
tinamide condenses with the ADP-ribose–like intermediate and drives
the reverse reaction to form [14C]NADþ.46,47 This assay can be used for
generating [14C]NADþ (with the radioisotope on the nicotinamide moiety).
46
A. A. Sauve and V. L. Schramm, Biochemistry 42, 9249 (2003).
47
M. D. Jackson, M. T. Schmidt, N. J. Oppenheimer, and J. M. Denu, J. Biol. Chem. (in press)
(2003).
To perform the [14C]nicotinamide-NADþ exchange reaction, typical

concentrations of substrates and enzyme are as follows: 0.2 M enzyme,
300 M acetylated substrate (concentration should be close to saturating,
meaning that the concentration should be at least 5–10 the Km for the
substrate), 500 M NADþ, 30 M to 1 mM [14C]nicotinamide (approxi-

mately 10–100 CPM/pmol). The reaction is carried out at 37 in 50 mM
Tris (or phosphate) pH 7.5, with 1 mM DTT. After quenching the reaction
by the addition of TFA to a final concentration of 1%, [14C]NADþ and
[14C]nicotinamide can be resolved and quantified by spotting the samples
onto a TLC plate as described earlier. The TLC plate can then be exposed
to a phosphoimaging screen or photographic film. The amounts of
[14C]NADþ and [14C]nicotinamide can be quantified by using a densitom-
eter. Alternatively, the quenched samples can be injected and resolved by
reversed-phase HPLC. Using the typical gradients described in the HPLC-
based Deacetylase Assay section, nicotinamide and NADþ elute at approxi-
mately 5% and 12% solvent B, respectively. Radioactivity of collected
fractions can be determined by scintillation counting (Fig. 3). As illustrated
in Fig. 3, the radioactivity on the nicotinamide fraction decreases while

Fig. 3. [14C]Nicotinamide-NADþ exchange reaction. The reaction was performed at 37 in
the presence of 500 M NAD þ, 300 M AcH3(K14), 0.2 M HST2, and 50 M
[14C]nicotinamide in 50 mM Tris, pH 7.5, with 1 mM DTT for the following time points:
0 min (black bar), 1 min (dark gray bar), 2 min (light gray bar), and 3 min (white bar). The
reaction was quenched with TFA to a final concentration of 1% and injected into the
reversed-phase HPLC. Radioactivity of the fractions collected from the HPLC was
determined by scintillation counting.
radioactivity on the NADþ fraction increases over time. Fractions con-

taining [14C]NADþ and [14C]nicotinamide can be pooled separately and
lyophilized for further use.
Nicotinamide Inhibition Assays. Among the different deacetylase fam-
ilies, only the Sir2 family produces nicotinamide as one of the products.
Therefore, Sir2-like enzymes, but not the class I and class II HDACs, are
susceptible to product inhibition by nicotinamide. In reactions involving
Sir2 homologues, nicotinamide has been shown to be a noncompetitive in-
hibitor against NADþ.23,48 We have also demonstrated that nicotinamide
exhibits noncompetitive inhibition with respect to the acetylated substrate
(Borra et al., manuscript in preparation). To determine nicotinamide inhi-
bition against the acetylated substrate, the deacetylase assay is performed
in the presence of saturating concentrations (i.e., at least 5- to 10-fold
higher than the Km) of NADþ, increasing acetylated substrate concentra-
tions (ranging 5- to 10-fold lower and higher than the Km) and various
but fixed concentrations of nicotinamide. To determine nicotinamide in-
hibition against NADþ, the deacetylase assay is performed in the presence
of saturating concentrations of the acetylated substrate, increasing concen-
trations of NADþ, and various but fixed concentrations of nicotinamide.
The assay is performed in 50 mM Tris (or phosphate), pH 7.5, with 1 mM
DTT at 37 . Because the nicotinamide inhibition assay requires many
samples, the charcoal-binding and TLC-based approaches should be util-
ized as described earlier. The assays are carried out under steady-state con-
ditions, in which less than 10% of substrates are converted to products. The
amounts of products formed are quantified as described in the charcoal-
binding or TLC-based methods. The initial velocity of the reaction is deter-
mined by dividing the concentration of product formed by the reaction
time and by the concentration (or total protein amount) of the enzyme
used (initial velocity calculation is discussed in the Kinetic Analysis of the
Sir2-Mediated Reaction section later). A double reciprocal plot of 1/ve-
locity versus 1/substrate concentration is constructed to determine the type
of inhibition. Because nicotinamide is a noncompetitive inhibitor of the
Sir2 family of deacetylases, the double reciprocal plot is expected to
produce a series of lines that intersect to the left of the 1/v axis (Fig. 4). Fit-
ting the inhibition data using the KinetAsyst software (IntelliKinetics,
State College, PA) or any equivalent software package can provide the
inhibition constants Kii and Kis, which reflect changes in the Vmax and
Vmax/Km steady-state parameters, respectively.
48
K. J. Bitterman, R. M. Anderson, H. Y. Cohen, M. Latorre-Esteves, and D. A. Sinclair,
J. Biol. Chem. 277, 45099 (2002).
Fig. 4. Schematic representation of a nicotinamide inhibition reaction. The reaction is

performed in the presence of varying concentrations of either NADþ or acetylated substrate,
saturating concentration of the other substrate, and various fixed levels of nicotinamide,
typically ranging from 25 to 100 M. The initial velocities of the reactions are determined, and
a double reciprocal plot of 1/velocity versus 1/[NADþ] or 1/[acetylated substrate] should yield
a series of lines that intersect to the left of the 1/velocity axis, indicating noncompetitive
inhibition.
Kinetic Analysis of the Sir2-Mediated Reaction

The quantitative approaches described earlier allow for accurate deter-
mination of the initial velocity, and thus, are suitable for performing
steady-state kinetic analyses. To calculate the initial velocity of each reac-
tion, the concentration of product formed is divided by the reaction time
and the amount of enzyme used. A graph of initial velocity versus substrate
concentrations, which typically straddles the Km by 5-fold in either direc-
tion, can be generated and the data fitted into the Michaelis-Menten equa-
tion [Eq. (i)], using Kaleidagraph (Synergy software, Reading, PA) or an
equivalent software, to determine the kinetic parameters Vmax, Km, and
Vmax/Km
V0 ¼ ðV max SÞ=Km þ S: (1)
þ
The Vmax is the maximal velocity achieved when both NAD and acety-
lated substrate are at saturating conditions (i.e., the concentrations are at
least 5- to 10-fold higher than Km), and the Km value is the substrate con-
centration at half-maximal velocity. The Vmax/Km is a second-order rate
constant that reflects the catalytic efficiency of the enzyme at low substrate
concentration and is the best parameter for examination of substrate speci-
ficity, as this term includes both substrate binding and catalysis. For the
Sir2-like reaction, the Vmax/Km parameter includes all catalytic steps from
substrate binding up to and including the nicotinamide release step. The
kinetic parameter kcat, which reflects the maximal turnover rate of the
enzyme, can be directly obtained by dividing Vmax by the concentration
of the pure enzyme. For several Sir2-like enzymes that we have examined,
the Vmax, Km, and Vmax/Km values varied depending on the enzyme and
acetylated substrate used. We have determined Vmax values ranging from
0.2 to 0.7 s1, and depending on the enzyme-substrate pair, Km values have
ranged from <2 to 400 M. The catalytic efficiency of the enzyme also
varied depending on the acetylated substrate used, with Vmax/Km value
of 1 105 M1 s1 for good substrates and 100-fold lower for poor
substrates.
Concluding Remarks
Steady-state kinetic analyses require the accurate measurement of the
substrates consumed or products formed during the reaction. The products
of the Sir2-like reaction can be identified and quantified using the HPLC-
based, charcoal-binding, and TLC-based assays described here. These
quantitative approaches can be used to compare the reactivity of various
Sir2 homologues, to determine substrate specificity, and to explore inhibi-
tor susceptibility for Sir2-like enzymes. Though beyond the scope of this
chapter, these assays can be utilized to perform more detailed kinetic anal-
yses that may include examination of various features of the Sir2-like
chemical mechanism, identification of the critical ionizations, mutational
analysis, establishing the basic kinetic mechanism, and elucidating the
order of substrate binding and product release.49–52
Acknowledgment
We would like to acknowledge Dr. Michael D. Jackson and Manning Schmidt for helpful
discussion.
49
I. H. Segel, ‘‘Enzyme Kinetics. Behavior and Analysis of Rapid Equilibrium and Steady-
State Enzyme Systems.’’ Wiley-Interscience, New York, 1993.
50
R. A. Copeland, ‘‘Enzymes. A Practical Introduction to Structure, Mechanism, and Data
Analysis,’’ 2nd Ed., Wiley-VCH, New York, 2000.
51
A. Cornish-Bowden, ‘‘Fundamentals of Enzyme Kinetics.’’ Portland Press, London, 1995.
52
A. Fersht, ‘‘Structure and Mechanism in Protein Science. A Guide to Enzyme Catalysis and
Protein Folding.’’ Freeman, New York, 1999.
[12] Selective HAT Inhibitors as Mechanistic Tools

for Protein Acetylation
By Yujun Zheng, Paul R. Thompson, Marek Cebrat, Ling Wang,
Meghann K. Devlin, Rhoda M. Alani, and Philip A. Cole
Histone acetyltransferases (HATs) regulate gene expression by the

targeted acetylation of histones and other proteins. It has been a difficult
challenge to identify the functional effects of protein acetylation in specific
pathways. We have introduced the use of selective HAT inhibitors as
mechanistic tools to probe the catalytic features of HATs and their roles
in various cellular pathways. We describe the design, synthesis, and appli-
cations of these inhibitors in different biological contexts. Lys-CoA, a se-
lective p300/CBP inhibitor, has been particularly useful in elucidation of
the role of p300/CBP HAT activity in gene regulation. It is hoped that
HAT inhibitors might ultimately serve useful clinical roles in the treatment
of cancer and other diseases.
Histone acetyltransferases (HATs), which catalyze the transfer of
acetyl groups from acetyl-CoA to the e-amino groups of lysine residues in
proteins (Fig. 1), play a major role in the regulation of transcriptional ma-
chinery and the modulation of gene expression. While HATs were initially
thought of as enzymes that catalyze the acetylation of histones on lysine
residues, it is increasingly recognized that they act on a wide variety of pro-
tein substrates.1 These acetylation events may in some cases affect protein-
protein interactions. Protein modules such as bromodomains have been
suggested to specifically bind to acetyl-lysine–containing sequences.2 Fur-
thermore, protein acetylation is a reversible modification. At least two
types of histone deacetylases (the HDACs and sirtuins) have been dis-
covered that have led to an understanding of ‘‘chromatin signaling’’ as a
dynamic process. In many ways, parallels can be drawn with reversible pro-
tein phosphorylation in cellular signaling.2 In addition to their importance
in fundamental biology, the linkage between protein acetylation and dis-
ease, especially cancer, is being intensively studied and it may be that
HATs will represent potential therapeutic targets.
The use of small molecules to modulate cellular and in vivo systems,
sometimes called ‘‘chemical genetics,’’ has received renewed interest in
the biology community in part because of continuing improvements in
1
R. Marmorstein, Nat. Rev. Mol. Cell. Biol. 2, 422 (2001).
2
S. L. Schreiber and B. E. Bernstein, Cell 111, 771 (2002).

[12] HAT inhibitors 189
Fig. 1. Histone acetyltransferase catalyzed reaction.
synthetic chemistry as well as molecular design.3 Histone deacetylase in-

hibitors such as trapoxin, trichostatin, and butyrate have played a major
role in assessing the general role of HDACs and acetylation in a variety
of transcriptional systems.2 Indeed, the use of HDAC inhibitors preceded
the isolation of the first histone deacetylases and was critical in the molecu-
lar identification of these proteins.4,5 The current reported compounds
have not proven especially useful in distinguishing among the nearly a
dozen HDACs (Class I and Class II) and thus have shown limited value
in unravelling specific pathways.
Unlike HDACs, the molecular identification of HATs was carried out
without the benefit of selective inhibitors.6 The first selective HAT inhibi-
tory compounds were reported 3 years ago,7 4 years after the first reports of
nuclear HATs. The widely divergent structural and mechanistic features
3
R. L. Strausberg and S. L. Schreiber, Science 300, 294 (2003).
4
M. Yoshida, M. Kijima, M. Akita, and T. Beppu, J. Biol. Chem. 265, 17174 (1990).
5
6
Allis, Cell 84, 843 (1996).
7
O. D. Lau, T. K. Kundu, R. E. Soccio, S. Ait-Si-Ali, E. M. Khalil, A. Vassilev, A. P. Wolffe,
Y. Nakatani, R. G. Roeder, and P. A. Cole, Mol. Cell 5, 589 (2000).
Fig. 2. Designed peptide-CoA conjugates for HAT inhibition
of the different HAT enzymes have allowed for a high degree of specificity
to be achieved in designing HAT inhibitors without intensive efforts. The
HAT inhibitory compounds reported to date fall under the category of bi-
substrate analogs. While simply designed, these compounds have proven
quite useful in the analysis of HAT function in a wide array of in vitro
and biological systems.
Design and Synthesis

The bisubstrate analogs targeted in the first generation of HAT inhibi-
tors7 involved the covalent linkage of the peptide substrate and acetyl co-
enzyme A (acetyl CoA) substrate (Fig. 2). This is an approach that has
proven quite powerful in a number of different acetyltransferases.8 Note
that the linkage is somewhat unnatural in that the attacked carbonyl of
acetyl-CoA is spaced one methylene unit from the CoA sulfur atom. The
primary motivation for this functionality comes from its track record with
other acetyltransferases and the synthetic simplicity and stability of the
target compound. Interestingly, this spacing may in fact be beneficial for
potency. For example, in a related system the loss of the methylene group
8
W. Zheng and P. A. Cole, Curr. Med. Chem. 9, 1187 (2002).
appears to reduce potency.9 Since the nature of the interaction between the
lysine-containing peptide substrate and the HAT enzymes was not known at
the outset of these efforts, a range of peptide moieties were examined. The
amino terminal tails of histones H3 and H4 were known to be likely physio-
logic acetylation targets of the HATs PCAF and GCN5 (close homologs)
and p300 and CBP (close homologs) so these sequences were selected for
the inhibitors. The length of the recognition motifs were not known so a
range of up to 20 amino acids was tested. The original synthetic approach ex-
ploited the synthesis of the peptides on solid phase with differential protec-
tion of the lysine earmarked for CoA modification.7 In these earlier efforts
the lysine was coupled to bromoacetic acid after cleavage of the peptide
from the solid support. In more recent efforts (see Fig. 3), bromoacetylation
is done while the peptide is still immobilized on the resin and the bromoace-
tamide intermediate is then cleaved from the solid phase.10,11 For Lys-CoA,
Rink amide resin is utilized, whereas for the more complex peptide-CoA
conjugates, Wang resin is used resulting in a C-terminal carboxylic acid.
Solution-phase coupling between the bromoacetamide and CoASH is gen-
erally a clean reaction and the final product purified by reversed-phase
HPLC on a C-18 column and characterized by MS. Multimilligrams of
the desired compounds have been generated in this fashion.
Kinetic and Structural Studies

Initial determination of the potency and selectivity of these compounds
was carried out with enzyme assays involving full-length recombinant p300
and PCAF as enzymes, mixed histone substrates, and 14C-acetyl-CoA.7
Separation of the acetylated products was carried out using SDS-PAGE
and quantitation was done employing phosphorimage analysis referenced
to an internal 14C-BSA standard. The results of these inhibition experi-
ments are shown in Table I. Lys-CoA was shown to be a rather potent
p300 inhibitor (IC50 0.4 M) and selective (>100-fold) versus PCAF. In
more recent assays that utilize the amino-terminal peptides rather than
the full-length histone substrates, it has been shown that Lys-CoA shows
good selectivity versus recombinant yeast GCN5 and EsaI, with IC50s
greater than 40 M for these enzymes (Table I). Peptide-CoA conjugates
9
E. M. Khalil, J. De Angelis, M. Ishii, and P. A. Cole, Proc. Natl. Acad. Sci. USA 96,
12418 (1999).
10
A. N. Poux, M. Cebrat, C. M. Kim, P. A. Cole, and R. Marmorstein, Proc. Natl. Acad. Sci.
USA 99, 14065 (2002).
11
M. Cebrat, C. M. Kim, P. R. Thompson, M. Daugherty, and P. A. Cole, ‘‘Synthesis and
Analysis of Potential Prodrugs of Coenzyme A Analogs for the Inhibition of Histone
Acetyltransferase p300.’’ Bioorg. Med. Chem. 11, 3307–3313 (2003).
Fig. 3. Synthetic schemes for (A) Lys-CoA and (B) assorted peptide-CoA conjugates
(B, Z, J represent variable amino acids that lead to the structures in Fig. 2).
have also proved to be weak inhibitors of the EsaI homolog Mof (unpub-
lished data), consistent with the nonsequential catalytic mechanisms of this
HAT family. As expected, Lys-CoA shows high potency against the p300
homolog CBP.12 In more extensive kinetic analysis, Lys-CoA was shown
to exhibit slow, tight-binding behavior with a half-life of interaction with
p300 of 4.5 min and a Ki* of 19 nM.13 Since product release appears to be
rate-limiting for p300 HAT activity, this long half-life could be rationalized
as due to a slow conformational change.
12
A. Polesskaya, I. Naguibneva, A. Duquet, S. Ait-Si-Ali, P. Robin, A. Vervish, P. Cole, and
A. Harel-Bellan, EMBO J. 20, 6816 (2001).
13
P. R. Thompson, H. Kurooka, Y. Nakatani, and P. A. Cole, J. Biol. Chem. 276, 33721 (2001).
TABLE I
IC50 Values for Synthetic Compounds and CoASHa
p300 IC50 PCAF IC50 GCN5 IC50 EsaI IC50

Compound (M) (M) (M) (M)
CoASH 200 >20 NA NA

H3-20 NA >20 NA NA
H3-CoA-7 >30 >20 NA NA
H3-CoA-20 200 0.75b 0.5b 200c
H4-CoA-20 >10 >10 NA >100c
Lys-CoA 0.5 200 200 >100c
H3-(Me)-CoA-20 60 0.36a 0.3b NA
a
These measurements were made using mixed histones (33 g/ml) and acetyl-CoA
(10 M) as substrates, and full-length HAT enzymes as described in ref.7 unless
otherwise noted. NA: not assayed.
b
This measurement was made using recombinant PCAF and GCN5 catalytic domains with
50 M H3-20 peptide substrate and 10 M acetyl-CoA as described.10
c
These measurements were made using recombinant EsaI HAT domain [Y. Yan, N. A.
Barlev, R. H. Haley, S. L. Berger, and R. Marmorstein, Mol. Cell 6, 1195 (2000)] (a kind
gift of R. Marmorstein) and mixed histones (33 g/ml) and acetyl-CoA (10 M) as
substrates as described in ref.7
The more complex peptide-CoA conjugates such as H4-CoA-7 were

considerably weaker inhibitors of p300 than the simple Lys-CoA structure.
Since these longer peptides are better p300 substrates, bisubstrate analogs
with these longer peptide moieties might have been expected to be pre-
ferred. In later work, it was discovered that p300 most likely follows a
ping-pong kinetic mechanism.13 In ping-pong kinetic mechanisms for
two-substrate enzymes, binding of both substrates to the enzyme simultan-
eously should be unfavorable, suggesting why the more complex peptide-
CoA conjugates are weak p300 inhibitors. The structural basis for the
high-affinity of lysyl moiety of Lys-CoA is still unknown, however.
In contrast to p300, PCAF is potently (IC50 0.5 M) inhibited by the
complex H3-CoA-20 derivative but not shorter analogs.7 H3-CoA-20 thus
represents a selective inhibitory agent, and is also a weak inhibitor versus
EsaI. However, it is about equipotent at blocking the PCAF homolog
GCN5 (Table I).10 Detailed kinetic studies revealed that H3-CoA-20 shows
a Ki of 28 nM and is a linear competitive inhibitor versus acetyl-CoA and a
linear noncompetitive inhibitor versus peptide substrate.14 This was
expected since PCAF follows an ordered BiBi ternary complex kinetic
14
O. D. Lau, A. D. Courtney, A. Vassilev, L. A. Marzilli, R. J. Cotter, Y. Nakatani, and P. A.
Cole, J. Biol. Chem. 275, 21953 (2000).
mechanism.14,15 Its relatively short half-life (<1 min) for binding to PCAF
was different from the behavior of Lys-CoA and p300 and was consistent
with the fact that product release does not appear to be rate-determining
for PCAF.14,15 One interesting finding was that shorter peptide-CoA conju-
gates such as H3-CoA-7 were rather weak inhibitors of PCAF and GCN5.
This result nicely correlates with studies showing that histone H3–derived
peptides considerably shorter than 20 amino acids are weak PCAF sub-
strates even though X-ray structures of the ternary complex of GCN5
reveal that most of the key interactions are mediated by the proximal
residues near Lys-14.16,17
Based on previous work on the GNAT superfamily member serotonin
N-acetyltransferase,9 a bisubstrate analog (H3-(Me)CoA-20) containing an
isopropionyl linker was synthesized10 as an epimeric mixture and shown to
be 4-fold more potent in blocking PCAF and GCN5 compared to the ori-
ginal compound (assuming only one stereoisomer to be potent). An X-ray
structure of the H3-(Me)CoA-20/GCN5 complex revealed that it was the
S-isomer (within the isopropionyl linker) of H3-(Me)CoA-20 that was
bound.10 More interestingly, the structure showed diminished interaction
between the peptide moiety of the bisubstrate analog and the enzyme
compared to the ternary complex.10 Relatively few residues of the peptide
(4–5) made contact with GCN5 and this structure seemed paradoxical in
light of the increased potency of longer peptide-CoA conjugates. However,
site-directed mutagenesis studies supported the notion that this complex
corresponds to a late catalytic intermediate in which the peptide substrate
may be dislodged in preparation for product release.10 Thus, it is likely
that the initial encounter complex between the bisubstrate analog and
GCN5, not captured by the X-ray structure, involves more extensive
peptide-enzyme interactions.
Applications of HAT Inhibitors in Gene Expression Studies

The ultimate value of the development of selective HAT inhibitors is to
address biologically meaningful questions about HAT functions. While
genetic approaches involving embryonic knockouts, RNAi, and ‘‘dominant
negatives’’ have given insights into the roles of HAT activity in a variety of
transcriptional pathways, they have limitations. For example, knockouts
and RNAi would delete the whole of p300 and CBP protein, including their
15
K. G. Tanner, M. R. Langer, and J. M. Denu, Biochemistry 39, 11961 (2000).
16
R. C. Trievel, F. Y. Li, and R. Marmorstein, Anal. Biochem. 287, 319 (2000).
17
J. R. Rojas, R. C. Trievel, J. Zhou, Y. Mo, X. Li, S. L. Berger, C. D. Allis, and
R. Marmorstein, Nature 401, 93 (1999).
many recruiting domains, and would not distinguish physiologic effects due
specifically to the HAT domains. Furthermore, since p300 and CBP (and
PCAF and GCN5) may show overlapping function, the more technically
demanding double knockouts may be required. Dominant negative
mutants can always be difficult to interpret because of the unwanted side
effect of protein overexpression. Moreover, none of these methods provide
the temporal control of dynamic processes or reversibility attainable with
small molecule-protein inhibitors. Thus, synthetic compounds can be quite
useful in assessing the role of HAT function. The lack of cell permeability
seen in many cases18 provides a significant limitation of the synthetic in-
hibitors described in this review. Nevertheless, both Lys-CoA and H3-
CoA-20 have been used in a series of transcriptional studies as highlighted
later. Their delivery in biological experiments can be divided into three
major approaches:
1. In vitro transcription studies with simple addition to cell-free
systems
2. Microinjection into cells
3. Concomitant addition of a lipid permeabilizing agent
HAT Inhibitors in In Vitro Transcription Systems

The applications of HAT inhibitors in the in vitro setting are straight-
forward since the need for cell permeability does not come into play and
because of the precise possibility of temporal control. These studies have
typically involved assessing the role of p300/CBP in a chromatin template.
In studies by Roeder and co-workers, the compound Lys-CoA has been
most useful in establishing the role of targeted histone acetylation in tran-
scriptional coactivation of p300, distinguishing the contributions of exoge-
nous versus nuclear extracts containing p300 and CBP, and revealing the
importance of histone acetylation in initiation rather than elongation of
transcription.19,20 In these studies, it was found that Lys-CoA (10 M)
could block transcription from a chromatin template but not a naked
DNA template and that the PCAF/GCN5 selective inhibitors had only a
small effect.
In related studies, Montminy and co-workers showed that CREB-
dependent activation of chromatin transcription was blocked by Lys-CoA
18
K. Subbaramaiah, P. A. Cole, and A. J. Dannenberg, Cancer Res. 62, 2522 (2002).
19
T. K. Kundu, V. B. Palhan, Z. Y. Wang, W. An, P. A. Cole, and R. G. Roeder, Mol. Cell 6,
551 (2000).
20
W. An, V. B. Palhan, M. A. Karymov, S. H. Leuba, and R. G. Roeder, Mol. Cell 9,
811 (2002).
(10 M), also supporting the role for p300 HAT activity in this process.21
Recently, Brady’s group and Nyborg and co-workers each demonstrated
the role of p300/CBP HAT activity as critical in transcriptional activation
by the HTLV-I Tax protein using Lys-CoA in a dosage-dependent
fashion.22,23 Nyborg’s study was especially noteworthy because it demon-
strated the role of HAT activity in transcriptional activation even within
a ‘‘tailless’’ chromatin template.23 Numerous potential p300/CBP sub-
strates were found in nuclear extracts and their acetylation was blocked
by Lys-CoA. Somewhat contrasting the findings of Roeder and co-
workers,20 Nyborg’s results hint at the potential for multiple acetylation
targets in transcriptional activation, at least in certain contexts.23
Kraus and co-workers have examined the role of p300/CBP in tran-
scriptional activation by thyroid hormone and retinoic acid receptors in
an in vitro transcriptional system.24 These studies showed a clear role for
the p300/CBP HAT activity which was inhibited by Lys-CoA. Interestingly,
the potential HAT activity of steroid receptor coactivator (SRC) proteins
was not important in transcriptional activations, although the SRCs play an
important role in recruiting p300/CBP to the hormone-regulated pro-
moters. The selective PCAF/GCN5 inhibitor H3-CoA-20 was also able,
albeit to a lesser extent, to block transcriptional activation in this system,
possibly suggesting a role for the PCAF/GCN-5 HAT activity in this
process.24
HAT Inhibitors and Microinjection Delivery

While cellular microinjection is a relatively specialized technique in the
cell biology community, it has played a significant part in mechanistic stud-
ies on transcriptional activation. It, thus, has been embraced by a number
of labs examining the role of p300/CBP HAT activity in various processes.
In general, concentrated (1–10 mM) solutions of Lys-CoA are needed
since the compound is typically diluted by 100-fold upon microinjection
into the cells. By microinjection into the C. elegans gonadal arm, Shi and
co-workers have used the relative effects of the CBP inhibitor Lys-CoA
to establish a role for CBP HAT activity in C. elegans development.25 In
21
H. Asahara, B. Santoso, K. Du, P. A. Cole, I. Davidson, and M. Montminy, Mol. Cell Biol.
21, 7892 (2001).
22
H. Lu, C. A. Pise-Masison, T. M. Fletcher, R. L. Schiltz, A. Nagaich, M. Radonovich, G.
Hager, P. A. Cole, and J. N. Brady, Mol. Cell Biol. 22, 4450 (2002).
23
S. A. Georges, H. A. Giebler, P. A. Cole, K. Luger, P. J. Laybourn, and J. K. Nyborg, Mol.
Cell Biol. 23, 3392 (2003).
24
J. Li, K. C. Lee, P. A. Cole, J. Wong, and W. L. Kraus, Mol. Endocrinol. 17, 908 (2003).
25
M. Victor, Y. Bei, F. Gay, D., C. Mello, and Y. Shi, EMBO Rep. 3, 50 (2002).
contrast, H3-CoA-20 had no effect in this system. These findings were cor-
roborated by a ‘‘knock-in’’ experiment where CBP lacking a functional
HAT domain showed an identical phenotype.25
The role of p73 acetylation by p300/CBP was also probed by microin-
jection studies.26 It had been proposed that p300 (but not CBP) was re-
sponsible for acetylation of p73 in the context of DNA damage induced
by doxorubicin through a c-abl–dependent pathway. The site-specific acet-
ylation of p73 was hypothesized to be responsible for apoptosis in this set-
ting. Consistent with these proposals, microinjected Lys-CoA was able to
inhibit p73-induced apoptosis in p53 / mouse embryonic fibroblasts in
response to doxorubicin.26
Ott and co-workers27 have used Lys-CoA to investigate the role of Tat
acetylation by p300 in transcriptional activation. Microinjection of only
8 M Lys-CoA into HeLa cells caused a reduction in Tat-mediated tran-
scriptional activation of a luciferase reporter. In parallel, it was shown that
siRNA specific for p300 could abolish Tat-mediated transcriptional acti-
vation, whereas siRNA specific for CBP had no effect. Taken together,
these results argue for a specific role for p300 HAT activity in Tat-medi-
ated transcriptional activation. Connected to these findings, Ott’s group
also showed that Tat acetylation modulated its interaction with CyclinT1
and proposed that this could allow for Tat to more readily be transferred
to the elongating RNA polymerase II.
Recent studies from Wong’s group highlight the value of Lys-CoA in a
frog oocyte transcription system investigating the mechanisms of androgen
receptor and thyroid hormone receptor.28 By microinjection of Lys-CoA,
Wong demonstrated that histone acetylation was critical in transcriptional
activation and cofactor recruitment but was not key in altering DNA
supercoiling mediated by SWI/SNF.28 These studies support a sequential
mechanism for chromatin remodeling.
Intracellular Delivery of Lys-CoA with Sphingosylphosphoryl Choline

While the charges on Lys-CoA reduce its bioavailability, Harel-Bellan
and co-workers discovered that the reagent sphingosylphosphoryl choline
(SPC, previously marketed by Invitrogen as ‘‘Transport’’) could be used
to enhance its cellular uptake.12 Presumably by poking small holes in
26
A. Costanzo, P. Merlo, N. Pediconi, M. Fulco, V. Sartorelli, P. A. Cole, G. Fontemaggi, M.
Fanciulli, L. Schiltz, G. Blandino, C. Balsano, and M. Levrero, Mol. Cell 9, 175 (2002).
27
K. Kaehlcke, A. Dorr, C. Hetzer-Egger, V. Kiermer, P. Henklein, M. Schnoelzer, E. Loret,
P. A. Cole, E. Verdin, and M. Ott, ‘‘Acetylation of Tat defines a Cyclin T1-independent
step in HIV transactivation.’’ Mol. Cell 12, 167–176 (2003).
28
Z.-Q. Huang, J. Li, L. M. Sachs, P. A. Cole, and J. Wong, EMBO J. 22, 2146 (2003).
cell membranes, SPC was found to allow for intracellular p300/CBP HAT
activity to be blocked. Surprisingly, the enzyme remains blocked even after
immunoprecipitation and in vitro HAT assay, suggesting a very slow disso-
ciation of inhibitor. As expected, PCAF was not inhibited by Lys-CoA
under these conditions.
Using this approach, Harrel-Bellan and co-workers showed that p300/
CBP HAT activity was essential for terminal differentiation of muscle
cells.12 This contradicted a previous study arguing against a role for p300/
CBP HAT activity in muscle cell differentiation.29 The Lys-CoA inhibitor
was also valuable in providing temporal information related to muscle
differentiation. Thus, Lys-CoA blocked cell fusion and the expression of
late muscle specific markers like myosin heavy chain and myosin creatine
kinase.
Following Harel-Bellan’s approach, Medrano and co-workers success-
fully employed SPC to introduce Lys-CoA into melanocytes.30 These stud-
ies revealed that p300/CBP HAT activity was critical in preventing
senescence in melanocytes and may play a vital role in the immortalization
of these cells associated with malignant conversion. Thus, p300/CBP
inhibitors may have therapeutic value in pigmented cell neoplasia.
Summary
Selective and potent bisubstrate analog inhibitors are now available for
the PCAF/GCN5 and p300/CBP enzymes. The development of these in-
hibitors has provided for new insights into the mechanisms of histone acet-
yltransferases and the role of protein acetylation in gene regulation. They
have been particularly powerful in dissecting the role of p300/CBP HAT
activity in the context of its overall contributions to gene regulation. De-
spite their utility, several challenges still exist. No potent HAT inhibitors
have been reported for the EsaI family of HATs and no compounds can
distinguish between the close homologs p300 and CBP or PCAF and
GCN5. Moreover, increasing the cell permeability properties of existing
compounds is an important future direction. Ultimately, it will be of great
importance to know whether small molecule HAT inhibitors can make an
impact on the treatment of human disease.
29
T. A. McKinsey, C. L. Zhang, and E. N. Olson, Curr. Opin. Cell Biol. 14, 763 (2002).
30
D. Bandyopadhyay, N. A. Okan, E. Bales, L. Nascimento, P. A. Cole, and E. E. Medrano,
Cancer Res. 62, 6231 (2002).
[13] histone deacetylase inhibitors 199
Acknowledgments
We are grateful to our many collaborators whose names are mentioned in the references.
We thank Dr. R. Marmorstein for a gift of EsaI. We thank K. Miller for technical assistance.
This work was supported in part by the NIH and Ellison Medical Foundation. P.R.T. was
supported in part by a Canadian Institutes for Health Research post-doctoral fellowship.
[13] Histone Deacetylase Inhibitors: Assays to

Assess Effectiveness In Vitro and In Vivo
By Victoria M. Richon, Xianbo Zhou, J. Paul
Secrist, Carlos Cordon-Cardo, W. Kevin Kelly, Marija
Drobnjak, and Paul A. Marks
While it has been almost four decades since the discovery that core nu-
cleosomal histones are post-translationally modified by acetylation and
methylation,1 the publications on histone modification have increased ex-
ponentially following the cloning of histone deacetylase 1 (HDAC1), and
histone acetyltransferase A (HAT A) in 1996.2,3 It is well established that
post-translational modifications, for example, acetylation, methylation, and
ubiquitination of lysine residues, phosphorylation of serine and threonine
residues, and methylation of arginine residues of the core histone tails, play
pivotal roles in regulating cellular functions involving chromatin, such as
transcription, replication, and DNA repair. Moreover, the interplay be-
tween multiple histone tail modifications has led to the histone code hy-
pothesis4,5 which states ‘‘that distinct histone modifications, on one or
more tails, act sequentially or in combination to form a ‘‘histone code,’’
that is, read by other proteins to bring about distinct downstream events.’’4
The ability to efficiently monitor changes in histone acetylation
has become increasingly important as the roles this complex event plays
in several diseases, including cancer,6 neurodegenerative diseases,7 and
autoimmune diseases,8 have become better understood. Additionally,
1
V. G. Allfrey, R. Faulkner, and A. E. Mirsky. Proc. Natl. Acad. Sci. USA 51, 786 (1964).
2
3
Allis, Cell 84, 843 (1996).
4
5
B. M. Turner, Bioessays 22, 836 (2000).
6
P. Marks, R. A. Rifkind, V. M. Richon, R. Breslow, T. Miller, and W. K. Kelly, Nat. Rev.
Cancer 1, 194 (2001).
7
J. P. Taylor and K. H. Fischbeck, Trends Mol. Med. 8, 195 (2002).
8
N. Mishra, D. R. Brown, I. M. Olorenshaw, and G. M. Kammer, Proc. Natl. Acad. Sci. USA
98, 2628 (2001).

Acknowledgments
We are grateful to our many collaborators whose names are mentioned in the references.
We thank Dr. R. Marmorstein for a gift of EsaI. We thank K. Miller for technical assistance.
This work was supported in part by the NIH and Ellison Medical Foundation. P.R.T. was
supported in part by a Canadian Institutes for Health Research post-doctoral fellowship.
[13] Histone Deacetylase Inhibitors: Assays to

Assess Effectiveness In Vitro and In Vivo
By Victoria M. Richon, Xianbo Zhou, J. Paul
Secrist, Carlos Cordon-Cardo, W. Kevin Kelly, Marija
Drobnjak, and Paul A. Marks
While it has been almost four decades since the discovery that core nu-
cleosomal histones are post-translationally modified by acetylation and
methylation,1 the publications on histone modification have increased ex-
ponentially following the cloning of histone deacetylase 1 (HDAC1), and
histone acetyltransferase A (HAT A) in 1996.2,3 It is well established that
post-translational modifications, for example, acetylation, methylation, and
ubiquitination of lysine residues, phosphorylation of serine and threonine
residues, and methylation of arginine residues of the core histone tails, play
pivotal roles in regulating cellular functions involving chromatin, such as
transcription, replication, and DNA repair. Moreover, the interplay be-
tween multiple histone tail modifications has led to the histone code hy-
pothesis4,5 which states ‘‘that distinct histone modifications, on one or
more tails, act sequentially or in combination to form a ‘‘histone code,’’
that is, read by other proteins to bring about distinct downstream events.’’4
The ability to efficiently monitor changes in histone acetylation
has become increasingly important as the roles this complex event plays
in several diseases, including cancer,6 neurodegenerative diseases,7 and
autoimmune diseases,8 have become better understood. Additionally,
1
V. G. Allfrey, R. Faulkner, and A. E. Mirsky. Proc. Natl. Acad. Sci. USA 51, 786 (1964).
2
3
Allis, Cell 84, 843 (1996).
4
5
6
P. Marks, R. A. Rifkind, V. M. Richon, R. Breslow, T. Miller, and W. K. Kelly, Nat. Rev.
Cancer 1, 194 (2001).
7
J. P. Taylor and K. H. Fischbeck, Trends Mol. Med. 8, 195 (2002).
8
N. Mishra, D. R. Brown, I. M. Olorenshaw, and G. M. Kammer, Proc. Natl. Acad. Sci. USA
98, 2628 (2001).

inhibitors of histone deacetylases have shown efficacy in several animal

models of human disease9–11 and are currently undergoing clinical evalu-
ation as anti-cancer agents.6 These histone deacetylase inhibitors fall into
four structural classes that include the short chain fatty acids (e.g., butyrate,
valproate), hydroxamic acids (e.g., suberoylanilide hydroxamic acid or
SAHA, Trichostatin A), cyclic tetrapeptides (e.g., FK-228), and benza-
mides (e.g., CI-994 and MS-275). The agents currently under evaluation
in clinical trials span all four classes of inhibitors. Several in-depth reviews
on histone deacetylase inhibitors have recently been published.6,12,13
Studies investigating the effects of histone deacetylase inhibitors nor-
mally involve measuring changes in the histone acetylation levels induced
by their treatment. For example, samples have been obtained from cells
cultured in vitro with HDAC inhibitors, from various tissues, including
spleen, brain, liver, and tumors isolated from animals treated in vivo with
HDAC inhibitors, and from peripheral mononuclear cells, bone marrow
cells, and biopsy samples from patients participating in clinical trials with
HDAC inhibitors. Protocols for the isolation of histones and the analysis
of the cellular and tissue effects of histone deacetylase inhibitors are
described in this chapter.
Isolation of Histones
A common method of monitoring histone acetylation is through the
purification and evaluation of cellular histone extracts. The method men-
tioned later details the isolation of histones from cell suspensions. When
isolated tissues are used, the samples must first be homogenized to cell sus-
pensions. When using peripheral blood mononuclear cells or bone marrow
aspirates, the cells must first be isolated from the matrix by density gradient
centrifugation. Since the acetylation of histones is a reversible protein
modification, samples should be either processed immediately or flash frozen

at 80 for later processing. Although cell type–dependent, the typical
protein yield is approximately 10–100 g of histone extract per 107 cells.
9
J. S. Steffan, L. Bodai, J. Pallos, M. Poelman, A. McCampbell, B. L. Apostol, A. Kazantsev,
E. Schmidt, Y. Z. Zhu, M. Greenwald, R. Kurokawa, D. E. Housman, G. R. Jackson, J. L.
Marsh, and L. M. Thompson, Nature 413, 739 (2001).
10
E. Hockly, V. M. Richon, B. Woodman, D. L. Smith, X. Zhou, E. Rosa, K. Sathasivam,
S. Ghazi-Noori, A. Mahal, P. A. Lowden, J. S. Steffan, J. L. Marsh, L. M. Thompson, C. M.
Lewis, P. A. Marks, and G. P. Bates, Proc. Natl. Acad. Sci. USA 100, 2041 (2003).
11
N. Mishra, C. M. Reilly, D. R. Brown, P. Ruiz, and G. S. Gilkeson, J. Clin. Invest. 111, 539
(2003).
12
W. K. Kelly, O. A. O’Connor, and P. A. Marks, Expert Opin. Investig. Drugs 11, 1695 (2002).
13
R. W. Johnstone, Nat. Rev. Drug Discov. 1, 287 (2002).
Histone Extraction Protocol

Collect cells (2.5–10 106), wash with phosphate-buffered saline (PBS),
and centrifuge at 600g for 5 min. Discard the supernatant, resuspend the cell
pellet in ice cold lysis buffer (1 ml/107 cells) [10 mM MgCl2, 10 mM Tris-
HCl, 25 mM KCl, 1% Triton X-100, 8.6% sucrose, protease inhibitors, pH

6.5], and incubate on ice for 5 min. Centrifuge at 600g for 5 min at 4 ,
remove the supernatant, wash the nuclear pellet in TE buffer (1 ml/107
cells) [10 mM Tris-HCl, pH 7.4, 13 mM EDTA], centrifuge at 600g for
5 min at 4 , and discard the supernatant. Resuspend the pellet in ice cold
water containing 0.4 N H2SO4 (100 l/107 cells), vigorously vortex for 5 s,
and incubate on ice for 1 h—vortexing intermittently during the incubation.

Centrifuge at 10,000g for 10 min at 4 , collect the supernatant, add ice cold

acetone (1 ml/100 l of supernatant), and incubate at 20 for 1 h. Centri-

fuge at 10,000g for 10 min at 4 and carefully discard the supernatant with-
out disturbing the histone pellet. After the pellet has been air-dried,
resuspend in water (100 l/107 cells) and measure the protein concentra-
tion. Store isolated histones in 80 freezer.
Analysis of Histone Acetylation Levels

Histone acetylation can be evaluated in a variety of ways depending on
the sample to analyze and on the information desired. Described later are
protocols for three popular methods of detecting histone acetylation: dif-
ferential migration via acid-urea-triton (AUT) gel electrophoresis, western
blotting, and immunohistochemistry.
Differential Migration
Histone acetylation results in the loss of a positive charge on the acety-
lated lysine residue. This loss of charge can be visualized as a reduction in
the migration rate of the histone subtypes (H2A, H2B, H3, and H4) during
electrophoresis through an AUT slab gel. Because each histone can be
acetylated at several potential lysine residues, AUT gels display a ladder
pattern detailing the extent of acetylation for each histone subtype. AUT
gel electrophoresis is the preferred method of analysis when evaluation
of the stoichiometry of histone acetylation is desired. The protocol
described later is modified from the procedure described by Yoshida
et al.14 Due to the position and resolution of the various histone subtypes,
this AUT gel electrophoresis protocol is especially useful for visualizing the
acetylated forms of histone H4.
14
M. Yoshida, M. Kijima, M. Akita, and T. Beppu, J. Biol. Chem. 265, 17174 (1990).
AUT Protocol
The resolving gel [1 M acetic acid, 8 M urea, 0.5% Triton X-100,
45 mM NH4OH, 16% acrylamide] is prepared as follows:
1.7 ml glacial acetic acid
14.4 g urea
150 l 100% Triton X-100
0.9 ml saturated NH4OH
4.64 g acrylamide
0.16 g bis-acrylamide
Add H2O to 30 ml (low heat may facilitate dissolution)
200 l 10% ammonium persulfate
20 l TEMED

Layer exposed gel edge with H2O, let polymerize at 4 for 16 h
The stacking gel [1 M acetic acid, 8 M urea, 0.5% Triton X-100, 45 mM
NH4OH, 8.2% acrylamide] is prepared as follows:
1.7 ml glacial acetic acid
14.4 g urea
150 l 100% Triton X-100
0.9 ml saturated NH4OH
2.3 g acrylamide
0.16 g bis-acrylamide
Add H2O to 30 ml (low heat may facilitate dissolution)
200 l 10% APS
20 l TEMED
Let polymerize at room temperature 4 h

The gel should be pre-run at 170 V for 4 h at 4 in running buffer [0.2 M
glycine, 1 M acetic acid]. The terminals should be switched from conven-
tional electrophoresis so that the positive charged histones migrate toward
the cathode.
Normally, 1–20 g of histone extract should be used per lane depending
on the lane size and the method of protein staining to be used. The histone
samples are diluted 1:1 with AUT sample buffer [7.4 M urea, 1.4 M NH4OH,
10 mM dithiothreitol], incubated at room temperature for 5 min, and then
2.4 l of 1% pyronine Y dye in glacial acetic acid is added per 10 l of

sample. The samples are electrophoresed at 170 V for 20–30 h at 4 . The
protein bands are then visualized via Coomassie Blue or silver staining.
An example of the differential migration of histone H4 through an
AUT gel due to acetylation induced by SAHA is shown in Fig. 1.
Fig. 1. SAHA induces histone acetylation as visualized by Coomassie staining after AUT
gel electrophoresis. Murine erythroleukemia cells were treated with 2.5 M SAHA for 6 h.
Histone extracts were prepared and evaluated via AUT gel electrophoresis as described in
this chapter. The numbers denote the bands generated by laddering of histone H4 due to
acetylation at multiple lysine residues.
Immunological Detection
Immunological detection has become the method of choice to deter-
mine histone acetylation due to the expanding commercial availability of
antibodies that recognize key histone modifications. A plethora of anti-
bodies, both monoclonal and polyclonal, has been raised against both
native histones as well as peptides corresponding to individual acetylation
sites. The availability of these antibodies makes it possible to evaluate
global histone acetylation as well as the acetylation of specific residues on
specific histone subtypes. Detailed later are two immunological detection
methods, western blotting and immunohistochemistry, used to evaluate
histone acetylation changes.
Western Blotting Protocol

Standard western blotting procedures can be used to evaluate acetyla-
tion levels in histone extracts prepared as described earlier. Typically,
1–5 g of histone extract are subjected to standard SDS-polyacrylamide
gel electrophoresis using 15% mini-gels. After transferring to nitrocellu-
lose or PVDF, the membranes are blocked with 4% non-fat milk in
PBS for 15 min, and then incubated with anti-acetylated histone primary
Fig. 2. SAHA induces histone H3 acetylation in cultured T24 cells. T24 human
bladder carcinoma cells were treated with the indicated concentration of SAHA for 4 h.
The cells were then collected and histones extracted according to the described protocol.
Western blotting was performed using a polyclonal anti-acetylated histone H3 antibody from
Upstate (Ac-H3) or a polyclonal histone H3 antibody from Abcam (Total H3).

antibody overnight at 4 . The membrane is then washed and incubated with
an appropriate biotinylated secondary antibody at room temperature for
1 h. After extensive washing, the samples are developed using an avidin-
biotin-peroxidase system (Vector Laboratories, Burlingame, CA).
Examples of histone H3 acetylation detected by western blotting analysis
from histone samples isolated from SAHA treated cells are shown in Fig. 2.
Immunohistochemistry Protocol
Immunohistochemistry can be performed on cell or tissue preparations
and is useful for evaluating histone acetylation in the context of the whole
cells and tissues. For example, for samples made up of several cell types,
such as tissues, the only feasible way to discern relative histone acetylation
levels among the various cell types present is by immunohistochemistry.
For tissues and cells, samples can be fixed in buffered formalin and paraffin
embedded. For isolated cells, a less laborious method is to cytospin the
cells and fix to microscope slides.
Paraffin-embedded samples are cut into 5-M sections, deparaffinized
by conventional methods, quenched in 0.1% H2O2 to block endoge-
nous peroxidases, and subjected to antigen retrieval by boiling in 0.01 M
citric acid (pH 6.0) for 15 min. Isolated cell samples are prepared by fix-
ation with methanol/acetone prior to quenching in 0.1% H2O2. Both types
of samples are then blocked in 10% normal goat serum, followed by incu-
bation with rabbit-derived, anti-acetylated histone antibodies of interest.
After washing, the samples are incubated with a 1:1000 dilution of a bioti-
nylated goat anti-rabbit IgG secondary antibody. After extensive washing,
the samples are developed using an avidin-biotin-peroxidase system
Fig. 3. SAHA induces histone H3 acetylation in prostate tumor biopsies. 40

magnification of prostate tumor biopsies from a patient before (A) and after (B) treatment
with 900 mg/m2/day for 3 days with the HDAC inhibitor, SAHA. Paraffin-embedded biopsy
tissue was sectioned and stained using a 1:2500 dilution of an anti-acetylated histone H3
antibody (Upstate). Positive staining was visualized with diaminobenzidine (brown color).
Cell nuclei are counterstained with hematoxylin (blue/gray color). (See color insert.)
(Vector Laboratories, Burlingame, CA). Diaminobenzidine is used as the

final chromogen and hematoxylin is used as the nuclear counterstain.
An example of immunohistochemical staining of acetylated histone H3
from a prostate tumor biopsy of a patient treated with SAHA is shown in
Fig. 3.
Acknowledgments
We wish to thank Albert Cupo and Eddie Rosa for technical assistance.
[14] immunochemical analysis of chromatin 209
[14] Immunochemical Analysis of Chromatin

By Michael Bustin, Richard C. Robinson, and Fred K. Friedman
Introduction
Ever since antibodies were first used to study the organization of
histones in chromatin,1,2 immunochemical approaches have repeatedly
proved to be important tools for elucidating the structure and function of
chromatin. Accordingly, their use is constantly increasing, concomitant
with the growing repertoire of antibodies specific to nucleosomal compo-
nents, to modified histones, nucleic acids, nonhistone proteins, chromatin-
modifying enzyme complexes, and other components that modulate the
structure and activity of chromatin. Immunochemical reagents are espe-
cially versatile because antigen–antibody reactions occur under a relatively
wide range of conditions that do not markedly change the structure of
the chromatin fiber. These considerations are important in view of the dy-
namic nature of the chromatin fiber, the transient binding of nonhistone
and regulatory factors to chromatin, and the reversible posttranslational
modification of core histones.
Antibodies have a wide range of applications for the study of chromatin
structure and function, exemplified by the list in Table I. A plethora of immu-
nochemical protocols used to study histones, nucleosomes, chromatin, and
chromosomes has already been described3–7 and additional protocols are
presented in detail in the present volume of Methods in Enzymology. Ulti-
mately, all these procedures involve antigen–antibody reactions in the context
of chromatin. This chapter focuses on the major factors that govern these
reactions and that need to be considered when employing immunochemical
approaches to gain insights into questions related to chromatin.
Antigen–Antibody Reactions in Chromatin

The basis of all immunochemical reactions in chromatin analysis is
the binding of an IgG molecule to the nucleosome. Figure 1 presents a
computer graphics model of two antibody molecules with their ‘‘active
1
M. Bustin, Nat. New Biol. 245, 207 (1973).
2
M. Bustin, D. Goldblatt, and R. Sperling, Cell 7, 297 (1976).
3
M. Bustin, Methods Enzymol. 170, 214 (1989).
4
S. Muller and M. H. Van Regenmortel, Methods Enzymol. 170, 251 (1989).
5
V. Orlando, Trends Biochem. Sci. 25, 99 (2000).
6
D. A. White, N. D. Belyaev, and B. M. Turner, Methods 19, 417 (1999).
7
N. Suka, Y. Suka, A. A. Carmen, J. Wu, and M. Grunstein, Mol. Cell 8, 473 (2001).

210 immunochemical assays of chromatin functions [14]
TABLE I
Immunochemical Methods for Studies on Chromatin and
Chromatin-Interacting Components
Method Description
Purification Immunoaffinity purification of chromatin binding

proteins, chromatin-modifying multiprotein com-
plexes, nucleosomes containing unique proteins,
or modification in histones or DNA
Detection Western analysis

Enzyme-linked immunosorbent assays (ELISAs)
Antibody-based histology
Immunofluorescence (regular and confocal)
Demonstration of complex formation Coimmunoprecipitation
Supershift assays with specific antibodies
Confocal immunfluorescence microscopy
Intranuclear localization Immunofluorescence (regular and confocal)

Immune electron microscopy
Detection of binding to Supershift assays
nucleosomes or chromatin Chromatin immunoprecipitation (ChIP) assays
Immune sedimentation
Immune adsorption
Analysis of posttranslational ChIP assays
modification in histones Immunofluorescence
Immune electron microscopy
Western blots
Functional analysis Microinjection of antibodies into cells
Expression of antibodies in cells
sites’’ placed next to the amino-terminal tail of histone H3 and the DNA,
respectively. The model illustrates that an IgG will recognize only a
small portion of the nucleosomal surface, because the antibody-binding site
comprises an area of 500–1000 Å2,8 and the cross-sectional area of the nu-
cleosome core is about 8000 Å2. As a consequence, antibody binding is
affected by changes only within the localized region in the binding site,
and nucleosomal changes that are solely distal to the binding site do not
affect antibody interactions. The model also illustrates that several anti-
bodies can simultaneously interact with a nucleosome, provided that they
are specific to different antigenic sites. Obviously, antibody accessibility to
8
S. Jones and J. M. Thornton, Proc. Natl. Acad. Sci. USA 93, 13 (1996).
Fig. 1. Molecular model of the interaction of antibodies with a nucleosome. Structures

shown are based on coordinates from the Protein Data Bank for nucleosome core particle
(1KX5), an antibody Fc fragment (1DN2), and an antibody–peptide complex represented by
the Fab of antibody PC283 bound to a peptide derived from a hepatitis B virus surface antigen
(1KC5). Antibody heavy and light chains are blue and yellow, respectively. A peptide bound
in the antibody-binding site is shown in red, and illustrates the location and relative size of an
active site. Nucleosomal DNA is cyan, the H3 tail is black, and the histones are represented by
different colors. The model illustrates the size relationship between an antibody and a
nucleosome core. The inset illustrates the size relationship between antibody probes (in red)
and chromatin (model provided by C. Woodcock). A representative compacted and
decompacted region is denoted by a circle and square, respectively; note that antigens within
these regions exhibit differential accessibility to antibody probes.
some nucleosomal sites is reduced in the context of chromatin, which

contains regions with varying degrees of compactness (Fig. 1, inset).
The major histone antigenic determinants in chromatin are the amino
termini of the histone cores. Because these protrude beyond the periphery
of the nucleosome, they are readily available for interaction with anti-
bodies. In contrast, the globular regions of the core histones are relatively
inaccessible to antibody. Thus, antigenic determinants that are fully acces-
sible in free histones may not react with antibody while embedded in the
nucleosome, owing to steric hindrance by DNA or neighboring histones. It
has been estimated that less than 5% of the antigenic sites of free histones
are available in nucleosomes.9
9
D. Goldblatt and M. Bustin, Biochemistry 14, 1689 (1975).
The model in Fig. 1 illustrates some points to consider when planning

or interpreting immunochemical studies involving structural modifications
in the tail. The most basic consideration is that although such changes are
indeed immunodetectable, even the longest histone tail (histone H3)
cannot accommodate more than one antibody at any time. Thus, two anti-
bodies that target modifications at two different sites in the tail cannot
simultaneously bind because of steric hindrance. However, a single anti-
body may recognize two or more modifications if these are localized within
the same epitope.
Immunogens
The choice of the immunogen used to elicit antibodies is the single most
important factor that determines the utility of the resulting serum. Immu-
nization with a pure, defined immunogen ensures a specific immune response
that yields antiserum with chromatin-specific antibodies. Monoclonal ap-
proaches could, in principle, circumvent the need to use defined immuno-
gens; however, the identity of an epitope recognized by the antibodies is
difficult to define. Consequently, the production of monoclonals against
complex mixtures of nuclear extracts, or even against chromatin, failed to
produce a battery of useful antibodies. We describe some immunogens used
to elicit antisera for chromatin analysis. Protocols for preparing and using
these immunogens have been described in this series and elsewhere.3,4,10
Synthetic Histone Peptides

Synthetic peptides are presently the immunogen of choice for producing
antibodies to histones, histone variants, or posttranslationally modified his-
tone regions. The advantages of using synthetic peptides for producing
specific immunochemical reagents are obvious. First, large quantities of
defined peptides can be synthesized at relatively low cost, thereby circum-
venting the need for cumbersome biochemical purification of ‘‘native’’
immunogens. Another unique advantage of this approach is the ability to
produce antibodies to specific posttranslational modifications. The exquisite
specificity of these antibodies was the key factor in identifying the biological
importance of specific posttranslational modifications in the histone tails.
For immunization, the synthetic peptides must first be coupled to large
and highly immunogenic carriers, by approaches that have been described.11
All the procedures require affinity purification of the antibodies on resins
10
B. D. Stollar, Methods Cell Biol. 18, 105 (1978).
11
E. Harlow and D. Lane, ‘‘Antibodies.’’ Cold Spring Harbor Laboratory Press, Cold Spring
Harbor, NY, 2002.
containing either the peptide or peptide conjugated to a carrier different

from the one used for immunization. Several chapters in this volume pro-
vide details on the procedures for obtaining pure antibodies that target
specific posttranslational modification in the histone tails. Examples of
the use of these antibodies, especially for immunoprecipitation and analy-
sis of posttranslational histone modification at the single gene level, are
described below. The specificity of the antibodies is due to the fact that
the antibodies can recognize as few as 6–10 amino acids, a point that is dis-
cussed in detail below. The exquisite sensitivity of antibody to the fine
structure of chromatin is exemplified by a study using antibodies to the
peptide corresponding to the dimethylated H3 Lys-9 tail.12 Antibodies
were prepared to both this single tail peptide and to a branched peptide
consisting of four tail peptides. Only the latter antibody recognized hetero-
chromatin, presumably because of the use of an immunogen with multiple
and spatially proximate methylated Lys-9 residues.
Most commercial antibodies against nonmodified histones are elicited by
immunizing with synthetic peptides, usually against sequences derived from
the amino-terminal tails. Although these antibodies have proved useful for
chromatin analysis, an antibody against full-length, purified histones is
more suitable and produces stronger signals for certain applications.
Histone Fractions
Core histones are poor immunogens and do not elicit a significant
immune response. Serum with an adequate titer of histone-specific anti-
bodies can be elicited by immunizing with purified core histone–RNA com-
plexes.10 When eliciting antibodies to purified histones, it is extremely
important to use highly purified histones that do not contain even small
amounts of highly immunogenic contaminants that could potentially elicit
a disproportionately strong immune response.
Only a small fraction of the antibodies elicited by purified histone–
RNA complexes reacts with chromatin or nucleosomes.9 As evident from
Fig. 1, access to some of the antigenic determinants in isolated histones
is sterically hindered by the DNA or by other histones. Nevertheless, be-
cause these sera have been elicited against a defined histone, they are an
excellent source of specific antibodies. Furthermore, because the mixture
of antibodies within sera is directed to a multiplicity of antigenic deter-
minants, the signal produced in Western analysis is often stronger than
that obtained with single peptide-elicited antibodies, which target a more
localized region with perhaps limited accessibility. Sera and antibodies
12
C. Maison et al., Nat. Genet. 30, 329 (2002).
produced by immunizing with histone–RNA complexes have proved useful

for chromatin analysis. Our use of these antibodies in a variety of studies
including Western analysis, immunohistochemistry, immunofluorescence,
immune electron microscopy, supershift assays, and microinjection has
already been described.3
Nucleosomes
Antisera to nucleosomes react well with mono- and oligonucleosomes
and linker histones, weakly with DNA and purified core histones, and not
at all with known nonhistone proteins such as HMGs.13 Digestion of nu-
cleosomes with proteases results in essentially complete loss of antigenicity
whereas DNase I digestion results in partial loss of antigenicity. The inter-
pretation of the results is obvious in light of Fig. 1. In nucleosomes, the
histones are organized into octamers whose topography is distinct from
that of isolated histones. Consequently, numerous antibodies will be
elicited against epitopes composed of surface regions near histone–histone
and histone–DNA interfaces. It is well documented that double-stranded
DNA is a poor immunogen and that few antibodies would thus be directed
against histone-free DNA regions. DNase I digestion nicks the DNA but
does not break histone–DNA contacts. Because the DNA remains at-
tached to the histone octamer, those epitopes composed of histone–DNA
contacts are intact and are still recognized by the antibodies. In contrast,
trypsin digestion cleaves the histones, thereby destroying epitopes com-
posed of histones or histone–DNA surfaces. Such proteolysis would thus
be expected to greatly reduce antigenicity.
Interestingly, antisera elicited against purified nucleosomes exhibit sig-
nificant species specificity.13 The specificity could be due to the sequence
variability of the histone amino-terminal tails, to distinct patterns of post-
translational modifications, or to both effects. By using affinity chromatog-
raphy, the sera could be a source of specific antibodies. However, the
inability to define the epitopes recognized by these antibodies seriously
limits their use for chromatin studies.
Purified and Dehistonized Chromatin

Antisera elicited against chromatin or ‘‘dehistonized chromatin’’ dis-
play remarkable specificity for the immunogen and distinguish between
preparations obtained from various cell types. The sera even distinguish be-
tween chromatin preparations obtained from a specific cell type at different
developmental stages or with different proliferation rates.14 However, the
13
C. S. Tahourdin and M. Bustin, Biochemistry 19, 4387 (1980).
antigenic sites recognized by the antibodies have not been clearly identi-
fied. For dehistonized chromatin, the specificity of the sera is attributed
to tissue-specific nonhistone proteins that bind tightly to DNA whereas
for chromatin preparations, the specificity could be due either to nonhis-
tone proteins or to posttranslational modifications in the histone tails.
As mentioned above, the sera can serve as analytical tools to detect chro-
matin differences between closely related cells. However, the information
gleaned from such studies is rather descriptive and thus far has not
provided new insights into chromatin structure.
Nonhistone Proteins
Antibodies are an important tool for elucidating the mechanism
whereby nonhistone proteins modify the structure and function of chroma-
tin. The choice of immunogen is guided by the same criteria as described
above; the most important consideration is rigorous purification and char-
acterization of the protein used as immunogen. Synthetic peptides corre-
sponding to an amino acid sequence deduced from genomic information
circumvent the need for protein purification and enable the production of
antibodies to modified amino acids in these proteins.
Studies with HMG proteins are a good example of the versatility of im-
munochemical techniques for analysis of the role of nonhistone proteins in
chromatin structure and function.15,16 Thus:
. Immunofluorescence, confocal immunofluorescence, immune elec-
tron microscopy, and antibody microinjections into living cells have
been used to demonstrate that HMGN proteins are associated with
transcriptionally active chromatin.17,18
. Microinjection of antibodies to HMGB1 into Pleurodeles waltlii
oocytes suggested a role for HMGB1 in transcription from
lampbrush chromosomes19 and immunofluorescence indicated that
part of the proteins are found in the cytoplasm.20
. Mobility supershift assays with specific antibodies demonstrated that
in chromatin, nucleosomes form complexes containing two molecules
of either HMGN1 or HMGN2. Furthermore, the nucleosomes
containing HMGN proteins are clustered into domains, in which
14
L. Hnilica, ‘‘Chromosomal Nonhistone Proteins.’’ CRC Press, Boca Raton, FL, 1983.
15
Y. V. Postnikov and M. Bustin, Methods Enzymol. 304, 133 (1999).
16
Y. V. Postnikov and M. Bustin, Methods Mol. Biol. 119, 303 (1999).
17
L. Einck and M. Bustin, Proc. Nat. Acad. Sci. USA 80, 6735 (1983).
18
R. Hock, F. Wilde, U. Scheer, and M. Bustin, EMBO J. 17, 6992 (1998).
19
U. Scheer, J. Sommerville, and M. Bustin, J. Cell Sci. 40, 1 (1979).
20
M. Bustin and N. Neihart, Cell 16, 181 (1979).
each domain contains an average of six contiguous HMGN-

containing nucleosomes. This approach, described in detail else-
where, is quite informative for analysis of the organization of
nonhistones in chromatin.15
. Immunofluorescence techniques were used to demonstrate the
organization of HMGB and HMGA in mitotic and polythene chro-
mosomes, to show that during mitosis HMGN proteins are
phosphorylated and not associated with chromosomes, and to study
the intranuclear organization of acetylated HMGN proteins.
. Chromatin immunoprecipitation (ChIP), which is widely used for
analysis of histone posttranslational modification and for studies on
the binding of regulatory regions to specific genomic sequences, was
also used in studies with HMG proteins. ChIP analysis of non-
histone proteins and regulatory factors requires cross-linking the
proteins to chromatin before chromatin fragmentation. This is a
critical step in the procedure because overextensive cross-linking
may reduce antigenicity. Additional examples of the use of anti-
bodies to study the role of HMG proteins in chromatin and details
of the protocols can be found elsewhere.15 These protocols can be
easily adapted to any chromatin-binding protein of interest.
Nucleic Acids
A large repertoire of antibodies specific to nucleic acids and modified
bases is available and can be used for chromatin analysis. The fundamental
approaches and technical considerations for using these antibodies are
similar to those discussed above.
Antibodies
Preparation
Procedures that maximize antibody efficacy and minimize batch-
to-batch variation have been described.11 For studies with chromatin, it is
advantageous to prepare antibodies in two different species, especially if
immunolocalization studies are planned. In our laboratory we routinely
prepare antibodies in both rabbits and goats. Repeated bleedings allow ac-
cumulation of large quantities of sera (approximately 500 ml from rabbits
and more than 1 liter from goats). The spectrum of epitopes recognized
can vary from bleed to bleed. Thus, after ensuring that each bleed has an
adequate titer, we pool the various bleeds, characterize the pooled serum,
and then store it lyophilized. Under dry and cold conditions the lyophilized
sera can be stored for long periods (we recover full activity after 20 years of
storage). The serum is reconstituted by adding H2O to the original volume,
and then centrifuged to remove insoluble residue.
Suka et al.7 describe the production of antibodies highly specific for dis-
tinct acetylation sites in yeast histone tails. The specificity of the sera was
determined by examining yeast strains bearing specific point mutations,
an extremely stringent test. Interestingly, the sera seem to be of sufficiently
high titer and specificity to be used directly, without additional purification.
However, in most cases, antibodies need to be purified by passage through
affinity columns containing the immunogen. The content of IgG in the
serum is on the order of 100 mg/ml and the content of specific antibodies,
which varies widely, is less than 1% of the total IgG. For anti-histone and
anti-HMG proteins, we recover about 0.2 mg of affinity-pure antibodies
per 1 ml of original serum. These solutions are stable for many years when
stored frozen.
Monoclonal antibodies can be an excellent choice, as high quantities of
these highly specific reagents can be prepared. However, caution must be
exercised because these are directed against a single epitope, which may
fortuitously resemble an epitope in a heterologous antigen, leading to
cross-reaction. Sera from patients suffering from autoimmune diseases
have served as a useful source of specific reagents.21–23 However, their rou-
tine use is limited by their limited availability, low amounts, and by the
labor–intensive process for deciphering their specificity. For most studies,
the antibody of choice is usually obtained by affinity purification from sera
elicited either by ‘‘native’’ proteins or by specific peptides.
Specificity and Affinity

The binding of antibody to an antigen is a noncovalent, bimolecular re-
action that can be measured and expressed in quantitative terms. A single
bleed of polyclonal antibodies typically contains a spectrum of IgG mol-
ecules (although only a few may predominate in terms of quantity or affin-
ity), directed against multiple epitopes on the immunogen. The dissociation
constants for these IgGs can vary from 10 5/mol to 10 12/mol. These values
need to be taken into consideration during affinity purification procedures.
Normally only the highest affinity IgG molecules are recovered from
the affinity columns. If the purification procedure is especially stringent,
fewer IgG molecules directed toward a more limited number of epitopes
21
R. W. Burlingame, Clin. Lab. Med. 17, 367 (1997).
22
M. Monestier, Methods 11, 36 (1997).
23
B. L. Kotzin, J. A. Lafferty, J. P. Portanova, R. L. Rubin, and E. M. Tan, J. Immunol. 133,
2554 (1984).
would be obtained. This would yield an antibody product whose narrow

specificity might approach that of a monoclonal antibody, whose possible
cross-reactivity problems have been mentioned above. Stringent affinity
purification may thus yield higher affinity yet lower specificity.
The question of specificity is relevant to practical immunochemical ap-
plications. An antibody that reacts only with the test antigen by Western
analysis may react nonspecifically in immunofluorescence where other
cellular components with somewhat lower affinity are concentrated in an
organelle. These considerations need to be taken into account in studies
with posttranslationally modified peptides. For example, in testing nuclear
extracts, we found that several antibodies to acetylated peptides of HMGN
proteins, which reacted with acetylated HMGN proteins with apparent
high specificity, cross-reacted strongly with acetylated, but not unacety-
lated, core histones (our unpublished data). In these nuclear extracts the
amount of core histones is approximately 100-fold higher than the amount
of HMGN proteins. The weak cross-reaction with the acetyl-lysine residue
rendered these antibodies unsuitable for measuring this modification in
HMGN proteins. In the preparation of antibodies to posttranslationally
modified proteins, it is customary to remove antibodies against nonmodi-
fied peptide by affinity chromatography. In our experience, overpurifica-
tion of antibodies intended to recognize a specific modification may
result in targeting the modified residue to the exclusion of the surrounding
residues. Such antibodies may then recognize such a limited structural
target in unrelated proteins, reducing their specificity.
It is important to consider the immunochemical implications of post-
translational modifications in structural terms. For example, the amino-
terminal tail of histone H3 is shown in Fig. 2. An antibody raised against
a peptide containing phosphorylated Ser-10, and that distinguishes be-
tween the phosphorylated and nonphosphorylated protein, must recognize
unique structural features aside from the phosphorylated Ser-10, in order
to minimize cross-reactivity with other proteins that contain phosphoser-
ine. The makeup of the epitope, for example, the relative contributions
of the phosphoserine and neighboring residues, dictates the antibody affin-
ity toward proteins with a phosphoserine and/or the residue neighbors. In
any event, higher antibody concentrations can compensate for lower affin-
ity, and antibody will then bind to nonphosphorylated H3 and to phosphor-
ylated serines in other proteins. It is therefore extremely important to use
appropriate concentrations of antibodies and antigens to obtain meaning-
ful results. The most stringent control is to use point mutants that lack
the modified residues, controls that have been done with yeast. The size
of an epitope is in the range of 500–1000 Å.2,8 This can cover a span of
as few as 6–21 amino acids, depending on the surface topography of the
Fig. 2. Models of an antibody–peptide interaction, and a fragment of the amino-terminal

tail of histone H3 with posttranslational modifications. For illustrative purposes, residues 1 to
20 are arbitrarily shown in an -helical conformation. From left to right: the unmodified
fragment with the oxygen of Ser-10 denoted by cyan; the fragment phosphorylated at Ser-10,
with the phospho moiety denoted by red; the fragment phosphorylated at Ser-10 and
acetylated at Lys-9 and Lys-14 (the acetyl moiety is denoted green). The representative
antibody active site is rightmost, and is composed of regions from heavy chain (blue) and light
chain (yellow). The antigen (violet) is a peptide derived from the hepatitis B virus surface.
Note the complementarity between the residues in the binding surfaces of the antigen and the
antibody, which results in an exact fit. The distance between the dashed lines denotes the
approximate size of an epitope.
epitope and the strength of the contacts with the antibody. Changes in the
epitope contact residues can considerably change both the affinity and spe-
cificity of the IgG for its intended target. Indeed, it has been reported that
acetylation of Lys-9 and Lys-14 abolishes the ability of the anti-phospho
H3S10 antibody to recognize the phosphorylated epitope.24 The steric
models in Fig. 2 illustrate how significant changes may occur in the vicinity
of the phosphorylated residue on acetylating the neighboring lysines. Inser-
tion of a few atoms within this region may significantly disrupt critical
epitope–IgG contacts if they alter the positioning of phosphate to hinder
its accessibility to the IgG.
The steric considerations mentioned above for posttranslational
modifications also pertain to macromolecular interactions. This is exempli-
fied by the interaction of a nonhistone with chromatin, in the HMG Box–
DNA complex (Fig. 3A). It is obvious that potential epitopes proximate to
24
A. L. Clayton, S. Rose, M. J. Barratt, and L. C. Mahadevan, EMBO J. 19, 3714 (2000).
Fig. 3. Effect of macromolecular interactions on accessibility of antigenic binding sites in

chromatin. (A) Model of an HMGBox–DNA complex, based on coordinates of Sry (1J46).
Regions in the interface between protein (violet) and DNA (black and yellow) are
inaccessible to antibody. (B) Globular region of histone H1 (1GHC) positioned on the
nucleosome core. Note that complex formation blocks antibody access to binding surfaces on
histone H1 (green) and nucleosomal DNA (black and yellow) and proteins (blue).
the HMG–DNA interface are not fully accessible to antibody binding.

Similarly, nucleosomal binding of histone H1 (Fig. 3B) will block anti-
body access to epitopes derived from the macromolecular interface. Meas-
urements of antibody accessibility can thus be a useful probe to map
chromatin-binding sites.
Concluding Remarks
The exquisite specificity of antibodies and their relative ease of
preparation render them extremely useful reagents for chromatin analysis.
They allow studies of specific components in the dynamic structure of a
living cell. To take full advantage of immunochemical approaches and
avoid pitfalls, one must carefully consider the basic features of the antibody
reaction with chromatin subunits, with their components, with nonhistones
such as HMGs, and with regulatory factors that constantly change and
remodel the structure of the chromatin fiber. The most important consider-
ations are as follows: choice of immunogen, purification of antibody, and
precautions in the experimental conditions to ensure that the immuno-
chemical reaction specifically reflects the quantity and state of the intended
antigen.
[15] antibodies against histones 221
[15] Generation and Characterization of Antibodies

Directed Against Di-Modified Histones, and Comments
on Antibody and Epitope Recognition
By Peter Cheung
In eukaryotic cells, all nuclear processes that require access to the DNA
template must contend with the fact that genomic DNA is packaged into
chromatin in vivo. The fundamental unit of chromatin is the nucleosome
and the definition of which is 146 bp of DNA wrapped around a histone oc-
tamer core comprising two copies each of histones H2A, H2B, H3, and
H4.1 Histone proteins are highly positively charged due to the high content
of lysine and arginine residues. In fact, many of these residues, particularly
those on the histone N-terminal tails, are subjected to post-translational
modifications such as acetylation, methylation, and ubiquitylation.2 In ad-
dition, a number of serine residues on histones are subjected to phosphor-
ylation as well. Because of the intimate contacts between histones and
DNA, these histone modifications can alter chromatin structure and in turn
play important regulatory roles in many DNA-templated processes. It was
almost 40 years ago that Allfrey et al. proposed that histone acetylation
serves to facilitate transcription3; however, it is not until recent years that
the biological roles of the different modifications on specific histone
residues are beginning to be elucidated. Just as a few examples, acetylation
of H4 at lysines 5 and 12 was found to be associated with the histone depo-
sition process, acetylation of H3 at lysines 9 and 14 as well as H4 at lysine
16 has been correlated to gene activation, and more recently, methylation
of lysine 9 of H3 has been shown to direct formation of heterochromatin
leading to gene silencing.1,4 As research into the functional roles of histone
modification continues to grow, we are beginning to understand that there
is a complex array of modified residues on the histones. Moreover, evi-
dence suggests that some histone modifications occur in combinatorial
fashions in association with specific nuclear processes.5 Given the large
number of potential combinations of modifications that can occur at the
level of the nucleosome, this has led to the proposal that histone modi-
fications collectively may form a histone or epigenetic code that specifies
1
A. P. Wolffe, in ‘‘Chromatin: Structure and Function.’’ Academic, San Diego, 1998.
2
K. E. van Holde, in ‘‘Chromatin.’’ Springer, New York, 1988.
3
V. G. Allfrey, R. Faulkner, and A. E. Mirsky, Proc. Natl. Acad. Sci. USA 51, 786 (1964).
4
Y. Zhang and D. Reinberg, Genes Dev. 15, 2343 (2001).
5
P. Cheung, C. D. Allis, and P. Sassone-Corsi, Cell 103, 263 (2000).

distinct biological responses to the different cellular pathways that con-

verge on histones.6–8 In this context, histone modifications not only func-
tion to modulate chromatin structure but also extend the amount of
information encoded by the chromatin-bound genome.
The recent advances in our understanding of histone modifications is in
part facilitated by the development and widespread uses of site- and modifi-
cation-specific histone antibodies. Early efforts were focused on generating
acetyl histone–specific antibodies to study the links between histone acety-
lation and a variety of nuclear processes. For example, by immunoblotting
analyses, histones acetylated at specific sites are linked to histone depo-
sition, transcription activation, and cell cycle progression.9–11 The acetyl
histone antibodies are also amenable to fluorescent microscopy studies
useful for examining the nuclear localization of the modified histones.
Indeed, such studies revealed that histones on the transcriptionally silent in-
active X chromosome in female mammal cells are hypoacetylated com-
pared to those on the autosomes, thus strengthening the link between
histone acetylation and transcription activity.12,13 An anti-acetyl lysine
antibody was also used in immunoselection studies that showed that acety-
lated histones are directly associated with transcriptionally active genes.14
Refinement of this experimental approach has led to development of the
chromatin immunoprecipitation (ChIP) assay, which is now widely used
to map direct association of modified histones or specific transcription
factors to known genes and genomic regions.
Following the successes of synthesizing acetyl histone antibodies, a va-
riety of antibodies against histones phosphorylated or methylated at differ-
ent sites have also been generated. These reagents have proven to be
invaluable for dissecting the diverse biological functions associated with
these histone modifications. By far, the majority of the modified-histone
antibodies that have been made are directed against single modifications
on the histone molecules. In addition, antibodies to di-modified H3 mol-
ecules have also been developed to examine the significance of specific
combinations of histone modifications. To date, two separate di-modified
6
7
8
T. Jenuwein and C. D. Allis, Science 293, 1074 (2001).
9
R. Lin, J. W. Leone, R. G. Cook, and C. D. Allis, J. Cell Biol. 108, 1577 (1989).
10
T. R. Hebbes, A. W. Thorne, and C. Crane-Robinson, EMBO J. 7, 1395 (1988).
11
B. M. Turner, Exp. Cell Res. 182, 206 (1989).
12
P. Jeppesen and B. M. Turner, Cell 74, 281 (1993).
13
B. A. Boggs, B. Connors, R. E. Sobel, A. C. Chinault, and C. D. Allis, Chromosoma 105,
303 (1996).
14
C. Crane-Robinson, T. R. Hebbes, A. L. Clayton, and A. W. Thorne, Methods 12, 48 (1997).
H3 antibodies have been generated: one specifically recognizes H3 phos-

phorylated at serine 10 (Ser10) and acetylated at lysine 14 (Lys14), the
other specifically recognizes H3 acetylated at Lys9 and phosphorylated at
Ser10.15,16 Both antibodies have been used to show that these combinations
of phosphorylation/acetylation (Phos/Ac) occur in vivo and are induced in
association with transcriptional activation of the immediate-early genes.
These studies not only illustrate that combinations of histone modifications
can be coupled, but also that antibodies are useful tools for determining
which combinations of modifications occur in the cell. In this chapter, I will
describe the approach and methods used in generating and purifying the
polyclonal Phos Ser10/Ac Lys14 di-modified H3 antibody (hereafter
Phos/Ac H3 Ab) from rabbit serum. While this particular affinity purifica-
tion scheme may not be necessary for all antisera, the overall protocol can
serve as a general guide for the generation and purification of antibodies
against modified histones.
Description of Strategy and Method
Peptide Synthesis, Coupling to Carrier Protein, and Rabbit Immunization

Peptides synthesized with the desired modifications on specific residues
are now routinely used for generating modification- and site-specific
histone antibodies. Description of peptide synthesis, coupling to carrier
proteins, and immunization of rabbits can be found in earlier publica-
tions.9,17 To raise an antiserum against the Phos/Ac H3 epitope, we synthe-
sized a peptide corresponding to amino acid residues 7–20 of the human
H3 molecule with a phosphate group on Ser10 (the number refers to the
amino acid residue on the full-length H3 molecule) and an acetyl group
on Lys14. We have chosen this combination because H3 Ser10 phosphoryl-
ation has previously been shown to correlate with activation of the imme-
diate-early genes,18 and because H3 Lys14 is one of the preferred sites of
acetylation by transcription-associated histone acetyltransferases such as
GCN5, PCAF, and p300.19,20 In addition, peptides corresponding to the
15
P. Cheung, K. G. Tanner, W. L. Cheung, P. Sassone-Corsi, J. M. Denu, and C. D. Allis,
Mol. Cell 5, 905 (2000).
16
A. L. Clayton, S. Rose, M. J. Barratt, and L. C. Mahadevan, EMBO J. 19, 3714 (2000).
17
D. A. White, N. D. Belyaev, and B. M. Turner, Methods 19, 417 (1999).
18
L. C. Mahadevan, A. C. Willis, and M. J. Barratt, Cell 65, 775 (1991).
19
M. H. Kuo, J. E. Brownell, R. E. Sobel, T. A. Ranalli, R. G. Cook, D. G. Edmondson, S. Y.
Roth, and C. D. Allis, Nature 383, 269 (1996).
20
R. L. Schiltz, C. A. Mizzen, A. Vassilev, R. G. Cook, C. D. Allis, and Y. Nakatani, J. Biol.
Chem. 274, 1189 (1999).
unmodified, the Ser10 phosphorylated, and the Lys14 acetylated forms of

H3 were also synthesized and used as controls. The Ser10 Phos/Lys14 Ac
di-modified peptide was synthesized with a cysteine residue added to the
C-terminal end for sulfhydryl coupling purposes. This peptide was coupled
to keyhole limpet hemocyanin (KLH) as carrier and then injected into
rabbits. Sera from the injected rabbits were collected over regular intervals
and evaluated by enzyme-linked immunosorbent assay (ELISA).
ELISA
To determine whether the rabbit serum contained antibodies specific
for the di-modified H3, reactivity to the original Phos/Ac H3 peptide used
for injections, as well as to the unmodified (unmod), Ser10 phosphorylated
(Phos S10), Lys14 acetylated (Ac K14) H3 control peptides was assayed by
ELISA. A brief protocol for ELISA is as follows:
1. Dilute each peptide to the desired concentrations (ranging from 0 to
25 ng/ml) in 0.05 M sodium carbonate buffer, pH 9.6.
2. Dispense 200 l of the diluted peptide solutions into the appropriate

wells on the ELISA plate and incubate at 37 overnight.
3. Wash the wells twice with PBS-T (1 PBS, pH 7.4, 0.05% Tween
20), add 200 l PBS-T containing 1% BSA to each well, and

incubate at 37 for 1 h to block the uncoated areas of the ELISA
wells with BSA.
4. Wash the wells twice with PBS-T, add 200 l of diluted rabbit serum

(in PBS-T) to each well and incubate at 37 for 2 h. Initial tests
should be done using a range of dilution of the rabbit serum (from
1:500 to 1:2000 for this Phos/Ac H3 antiserum).
5. Wash the wells twice with PBS-T, add 200 l of diluted (1:5000)
HRP-conjugated goat anti-rabbit IgG secondary antibody to each

well, and incubate at 37 for 2 h.
6. Wash the wells twice with PBS-T, add 100 l of OPD substrate
(Sigma Fast: P-9187) per well, and let the reactions develop for
30 min.
7. At the end of the incubation period, stop reactions by adding 50 l
of 1 M H2SO4 and assay the binding of rabbit antibody to the plate-
bound test peptides by measuring the absorbance of the reactions at
492 nm using a microtiter plate reader.
Figure 1A shows a typical example of the ELISA results obtained
using a 1:1000 dilution of one of the immunized rabbit bleeds. The
results show that the rabbit serum reacted to all four peptides tested
and the same results were found with different bleeds as well as with the
[15]
antibodies against histones
Fig. 1. ELISA analyses of the crude serum from a rabbit immunized with H3 peptides phosphorylated at Ser10 and acetylated at
Lys14. (A) The crude serum was tested for reactivity to the four indicated test peptides by ELISA. Binding of antibodies to the test
peptides was measured by a colorimetric assay and quantified by absorbance at 492 nm. (B) Parallel samples of crude serum were
either mock-treated or incubated with 10 g of the indicated competing peptides prior to being tested by ELISA. The addition of
different competing peptides led to selective loss of reactivity of the antibody mixture present in the crude serum.
225
sera from a second immunized rabbit. To test whether the serum

contained one population of antibody that reacted to all four peptides or
whether there was a mixture of antibodies each reacting to the different
peptides, peptide competition assays were done by adding various amounts
of competing peptides to the serum prior to addition to the ELISA plate
wells in step 4. Figure 1B shows one example of the peptide competition
assays in which 10 g of each competing peptides was added to crude
serum aliquots prior to dilution and addition to the ELISA plates. Addition
of excess unmodified H3 competing peptide selectively removed reactivity
of the rabbit serum to the unmodified test peptide. Addition of the
Lys14 acetylated H3 peptide as competitor removed reactivity of the serum
to both the unmodified and acetylated-alone peptide. Addition of the
Ser10 phosphorylated competing peptide blocked reactivity of the serum
to the unmodified as well as singly modified test peptides but did not
diminish reactivity of the serum to the phosphorylated/acetylated test
peptide. In fact, only the original di-modified peptide used to immunize
the rabbits was able to successfully compete away recognition of the phos-
phorylated/acetylated test peptide by the serum. These results together
suggest that the crude rabbit serum contained a mixture of antibodies
that reacted to the unmodified, the singly modified as well as the di-
modified forms of H3. Therefore, affinity purification is required to isolate
the fraction of antibodies that specifically recognizes the di-modified form
of the H3.
Ammonium Sulfate Precipitation

To begin the antibody purification process, the serum was first mixed
with ammonium sulfate to preferentially precipitate the immunoglobulins.
This procedure separates the immunoglobulins from serum albumin and
plasma proteins that otherwise could interfere with the affinity purification
process. A brief description of the ammonium precipitation protocol is as
follows:
1. Prepare a stock of saturated (approximately 4.1 M) ammonium
sulfate solution, filter through Whatman paper to remove undis-
solved crystals, and pH to 7.6 before use.
2. Dilute 10 ml of serum in an equal volume of 1 PBS containing
1 mM PMSF.
3. Spin the diluted serum at 3000g for 10 min to remove particulate
matters.
4. Transfer the supernatant/serum to a clean beaker and stir solution
gently using a stir bar over a magnetic stirrer.
5. In a dropwise fashion, slowly add an equal volume of saturated

ammonium sulfate to the diluted serum. A final concentration of
50% ammonium sulfate is often used for precipitation of immuno-
globulins; however, one should start to see a white precipitate
forming when the percentage of the ammonium sulfate in the
solution reaches >30%.
6. Continue to stir the 50% ammonium sulfate/serum solution for
10 min, transfer the mixture to clean centrifuge tube, and spin at
5000g for 10 min.
7. Remove the supernatant and resuspend the precipitated pellet
in 10 ml of 1 PBS. Repeat the precipitation process (steps 4–7)
to remove any residual albumin left from the first round of
precipitation.

8. The final immunoglobulin fraction in 1 PBS can be stored at 4 or
can be directly applied to the peptide column for affinity
purification.
Affinity Purification Strategy and Protocol

Initial attempts of a subtractive approach involving serial passages of
the crude serum through peptide columns containing singly acetylated or
singly phosphorylated H3 peptides resulted in non-specific loss of the di-
modified H3-specific antibodies over the peptide columns. Also, single pas-
sage of the crude serum over a peptide column containing the di-modified
form of H3 did not effectively separate the antibody populations that rec-
ognize the mono- and di-modified forms of H3. A final scheme based on
peptide competition was devised and found to be effective for isolating
the fraction of antibodies that specifically recognizes the Phos S10/Ac
K14 H3 peptide. This strategy involves adding an excess of the singly modi-
fied (Phos S10 or Ac K14) H3 peptides to the serum to block binding of the
antibodies specific for the singly phosphorylated or singly acetylated forms
of H3 to the column-bound di-modified H3 peptides. The protocol for this
purification scheme is as follows:
1. Prepare a peptide column by coupling 1 mg of the desired peptide
(in this case the Phos S10/Ac K14 peptide) to 1 ml of SulfoLink Coupling
Gel (Pierce, cat# 20401) as per manufacturer’s instructions and transfer the
peptide-coupled gel to an Econo-Pac chromatography column (BioRad,
cat# 732-1010).
2. Wash the peptide column sequentially with:
a. 10 ml of 10 mM Tris, pH 7.5
b. 10 ml of 100 mM glycine, pH 2.5
c. 10 ml of 10 mM Tris, pH 8.8
d. 10 ml of 100 mM triethanolamine, pH 11.5
e. 10 ml of 10 mM Tris, pH 7.5
3. Add 1 mg each of the Phos S10 and Ac K14 H3 peptides to the
10 ml of precipitated immunoglobulins in PBS (see step 8 of ammonium
sulfate precipitation protocol) and incubate with occasional mixing at
room temperature for 15 min.
4. Apply the immunoglobulin plus peptide mixture (step 3) over the
Phos S10/Ac K14 peptide column (step 1) and allow mixture to drip
through the column by gravity. Collect flow through and re-apply to
column one more time.
5. Wash peptide column with 3 cycles of:
a. 10 ml 1 M NaCl, 50 mM sodium acetate, pH 4.0
b. 10 ml of 1 M NaCl, 20 mM Tris, pH 8.0
6. Elute antibodies bound to peptide column with 10 ml of 100 mM
glycine, pH 2.5. Collect the eluate in four 2.5-ml fractions in separate tubes
each containing 0.5 ml of 1 M Tris, pH 8.0.
7. Wash peptide column with 10 ml of 10 mM Tris, pH 8.8.
(One can omit steps 8 and 9 if the desired antibody is known to elute in
glycine.)
8. Elute with 10 ml of 100 mM triethanolamine, pH 11.5. Collect
eluate in a tube containing 2 ml of 1 M Tris, pH 8.0.
9. Wash peptide column with 10 ml of 10 mM Tris, pH 8.8.
10. Elute one last time with 5 ml of 100 mM glycine, pH 2.5. Collect
eluate in a tube with 1 ml of 1 M Tris, pH 8.0.
11. Test the collected fractions for antibody reactivity by ELISA (see
earlier protocol).
12. Dialyze the fractions containing the desired antibody against 1
PBS until the fractions are equilibrated with 1 PBS, divide into small

aliquots and store at 80 .
Notes About This Protocol

1. This protocol is based on a standard antibody purification scheme
that uses cycling of pH’s to elute the column-bound antibodies.21
Most antibodies are eluted by the low pH glycine elution buffer;
21
E. Harlow and D. Lane, in ‘‘Antibodies: A Laboratory Manual.’’ CSHL Press, Cold Spring
Harbor, 1988.
however, some antibodies are only eluted by high pH. Therefore, to

be thorough, steps 8 and 9 are included in this protocol.
2. It is important to collect the eluate from the first glycine elution step
(step 6) in separate small volume fractions because the profile of the
antibodies eluted changes over the course of the elution step. By
this scheme, the majority of antibodies were found to be eluted in
the first two 2.5-ml aliquots collected (elution fraction 1a and b, see
Fig. 2). Fraction 1b contained the bulk of the Phos/Ac H3-specific
antibody. An even higher specificity fraction was eluted from the
second glycine wash (step 10, elution fraction 2 in Fig. 2); however,
the antibody yield from this round of elution was significantly less.
3. The eluates were collected in tubes containing 1 M Tris, pH 8, to
immediately neutralize the extreme pH’s of the glycine or
triethanolamine buffers and minimize damage to the eluted
antibodies.
4. In this case, the purified Phos/Ac H3 antibody was eluted from the
first and second rounds of glycine elution (steps 6 and 10). No
detectable antibody was eluted by the triethanolamine buffer (data
not shown).
Verification of the Specificity of the Phos/Ac H3 Antibody

In addition to testing the purified antibody by ELISA, the specificity
of histone modification antibodies should also be tested by western blot
analyses coupled with peptide competition assays. In the case of this di-
modified H3 antibody, previous studies have shown that H3 phosphoryl-
ation in mammalian cells is induced upon epidermal growth factor (EGF)
stimulation and correlates with transcription activation of the immediate
early genes.18 Indeed, we developed this particular Phos/Ac H3 antibody
to test whether EGF also induced accumulation of the Ser10-phosphoryl-
ated and Lys14-acetylated form of H3. Western blot analyses of histones
extracted from EGF-stimulated tissue culture cells showed that the purified
antibody specifically detected a band corresponding to histone H3 and the
level of this modified form of H3 is induced upon EGF stimulation (see
Fig. 3). A similar induction profile was seen using a previously character-
ized antibody specific for the phosphorylated form of H3 (Phos H3, see
Fig. 3). To confirm that the species of H3 detected by the purified antibody
was indeed di-modified, excess amounts of the same panel of modified H3
peptides used in the ELISA tests were incubated with the purified antibody
for 15 min at room temperature prior to its use in western blot analyses. As
shown in Fig. 3, only the Phos/Ac H3 peptide blocked detection of the
H3 band by the purified antibody, whereas the phosphorylated H3
230
immunochemical assays of chromatin functions
Fig. 2. Antibody specificities of the eluate collected from the affinity purification procedure. As indicated in the protocol, fractions of the
eluate from the glycine elution steps were tested by ELISA. Fractions 1a and b were the first and second 2.5-ml eluate fractions of the first 10-ml
glycine elution step. Fraction 2 was from the second glycine elution step. These ELISA analyses show that fractions 1b and 2 were specifically
enriched for antibodies that recognize the Phos/Ac H3 epitope.
[15]
Fig. 3. Further analysis of the specificity of the purified antibody by western blots.
Histones from mock-treated or EGF-treated cells were separated on SDS polyacrylamide
gels, transferred to PVDF membranes, and probed using a Phos H3 specific antibody or with
the affinity-purified Phos/Ac H3 antibody. Various competing peptides were added to the
antibodies prior to immunoblotting to test the antibodies’ specificities.
peptide was sufficient to block recognition of the H3 band by the Phos H3

antibody. These peptide competition assays therefore further confirm the
specificity of the purified Phos/Ac H3 antibody.
Density of Histone Modifications, Epitope Recognition, and the Potential

of Masking or Occlusion
While it has been known for decades that histone proteins are subjected
to a variety of post-translational modifications, the extent and functions of
these modifications are far from being fully understood. In fact, new sites of
modifications on core and variant histones are still being identified today.
The growing catalog of these modified histone residues indicates that some
regions on the N-terminal tails of H3 and H4 are particularly rich in
modified residues. For example, within the first five amino acids of H4
(Ser-Gly-Arg-Gly-Lys), it is known that Ser1 can be phosphorylated,
Arg3 can be methylated, and Lys5 can be acetylated.2,22,23 Similarly, on
22
H. Wang, Z. Q. Huang, L. Xia, Q. Feng, H. Erdjument-Bromage, B. D. Strahl, S. D. Briggs,
C. D. Allis, J. Wong, P. Tempst, and Y. Zhang, Science 293, 853 (2001).
23
B. D. Strahl, S. D. Briggs, C. J. Brame, J. A. Caldwell, S. S. Koh, H. Ma, R. G. Cook,
J. Shabanowitz, D. F. Hunt, M. R. Stallcup, and C. D. Allis, Curr. Biol. 11, 996 (2001).
H3 from Lys9 to Lys 14 (Lys-Ser-Thr-Gly-Gly-Lys), Lys9 can be acetylated

or methylated, Ser10 can be phosphorylated, Thr11 has recently been
shown to be phosphorylated during mitosis, and Lys14 is well documented
to be acetylated in association with transcription activation.1,24 At present,
the extent of multiple modifications that occur concurrently at these modi-
fication-dense regions is not known. Based on antibody studies, we know
that H3 phosphorylated at Ser10 and acetylated at Lys14, as well as those
acetylated at Lys9 and phosphorylated at Ser10 do exist15,16; however,
other combinations of histone modifications within these regions have
not been directly tested. The fact that multiple modifications can occur
on adjacent or nearby residues raises a potential problem in using histone
modification-specific antibodies. As documented by Clayton et al., the
authors of that study generated an antibody that specifically recognizes
the H3 phosphorylated at Ser10, but recognition of the Phos H3 epitope
by that antibody is totally abolished when the H3 molecules are also acety-
lated at Lys9 or Lys14.16 Therefore, some of the histone modification-
specific antibodies may be susceptible to epitope disruption or occlusion
when additional modifications are present on adjacent or nearby residues.
ELISA tests of another antibody specific for the Ser10 phosphorylated H3
(previously described in Hendzel et al.25) showed that this antibody can
recognize peptides that are singly phosphorylated at Ser10 and those that
are phosphorylated at Ser10 and acetylated at Lys14 equally well (see
Fig. 4). However, it is unable to recognize the Phos Ser10 epitope when
the adjacent Lys9 is also acetylated. On the other hand, recognition of
the intended epitope of Phos Ser10/Ac Lys14 by the affinity-purified anti-
body described in this chapter is not affected at all by the additional acet-
ylation at Lys9. Therefore, the problem of epitope disruption varies from
antibody to antibody and from site to site. With a wide variety of histone
modification-specific antibodies currently being used by many laboratories,
it is difficult to know which ones are potentially affected by epitope occlu-
sion. It is simply not feasible to test the specificity of every one of these
antibodies against all possible combinations of modifications on surround-
ing residues. Moreover, there is no doubt that modified residues on histone
molecules are yet to be discovered and therefore cannot be accounted for.
In spite of potential limitations, histone modification-specific antibodies
remain as some of the most useful tools for chromatin research. With
better characterization of new antibodies, the use of multiple sources of
antibodies, and further analyses of histone modification clusters, we can
24
U. Preuss, G. Landsberg, and K. H. Scheidtmann, Nucleic Acids Res. 31, 878
25
M. J. Hendzel, Y. Wei, M. A. Mancini, A. Van Hooser, T. Ranalli, B. R. Brinkley, D. P.
Bazett-Jones, and C. D. Allis, Chromosoma 106, 348 (1997).
Fig. 4. The Phos H3 and Phos/Ac H3 antibodies were tested for epitope occlusion by
ELISA tests. ELISA was performed as in Fig. 1 and 2 with the addition of a new test peptide
that corresponds to tri-modified H3 (acetylated at Lys9, phosphorylated at Ser10, and
acetylated at Lys14). The additional acetyl group on Lys9 inhibited recognition of the Phos
H3 epitope by the Phos H3 antibody but had no effect on the recognition of the Phos Ser10/
Ac Lys14 epitope by the affinity-purified antibody.
minimize potential issues with epitope disruption. At minimum, by being

more aware, we can be more careful and make better judgments regarding
data interpretation.
The fact that multiple modifications within close proximity can affect
the recognition of single modifications by specific antibodies points to an
interesting phenomenon that may actually have biological significance.
Recent studies have shown that modified histone residues can act as
binding motifs and mediate histone-protein interactions. For example,
acetyl lysines have been shown to bind the bromodomain motif and H3
methylated at Lys9 can bind to the chromodomain of the heterochromatin
protein HP1.4,26 The addition of other modifications on adjacent or nearby
residues potentially can disrupt binding of nuclear factors to singly modi-
fied histones and could represent a mechanism for regulating modifica-
tion-dependent binding of nuclear proteins to chromatin.5 Indeed, recent
data suggest that binding of HP1 to Lys9 methylated H3 is precluded by
26
C. Dhalluin, J. E. Carlson, L. Zeng, C. He, A. K. Aggarwal, and M. M. Zhou, Nature 399,
491 (1999).
the phosphorylation of the adjacent Ser10 residue (C. D. Allis, personal

communication; Jaskelioff and Peterson27). Therefore, it may be useful to
generate antibodies specific for the combination of H3 methylated Lys9
and phosphorylated at Ser10 to probe for their existence and occurrences
in vivo. With more and more histone residues found to be modified,
and some of them being clustered together, the generation of antibodies
against specific combinations of histone modifications will further our
understanding of the complex code of modifications present on the histone
molecules.
Acknowledgments
I would like to thank C. David Allis for frequent discussions and scientific support.
Portion of this work was done in the Allis lab. I would also like to thank Craig Mizzen for
excellent advice on antibody purification protocols. Finally, I would like to acknowledge
Elsevier for permission to adapt some of the figures originally published in Mol. Cell 5,
905–915 (2000) for this work.
27
M. Jaskelioff and C. L. Peterson, Nat. Cell Biol. 5, 395 (2003).
[16] Generation and Characterization of

Methyl-Lysine Histone Antibodies
By Laura Perez-Burgos, Antoine H. F. M. Peters, Susanne Opravil,
Monika Kauer, Karl Mechtler, and Thomas Jenuwein
Histones are among the most conserved proteins throughout evolution

and responsible for the compaction of DNA into the higher-order polymer,
known as chromatin. In addition to their structural roles, histones play import-
ant functions in the control of gene expression by regulating access to the
underlying nucleosomal template. Histone amino-termini (tails) protrude
from the nucleosome core and are subject to a variety of post-translational
modifications, including acetylation (on lysine residues), phosphorylation
(on serine and threonine residues), methylation (on lysine and arginine
residues), ubiquitination (on lysine residues), and ADP-ribosylation (on
glutamic acid residues).1 The various histone tail modifications have been
proposed to constitute a ‘‘histone code’’ that could significantly extend
the information potential of the genetic information.2–4
1
K. E. van Holde, ‘‘Chromatin.’’ Springer, New York, 1988.

the phosphorylation of the adjacent Ser10 residue (C. D. Allis, personal

communication; Jaskelioff and Peterson27). Therefore, it may be useful to
generate antibodies specific for the combination of H3 methylated Lys9
and phosphorylated at Ser10 to probe for their existence and occurrences
in vivo. With more and more histone residues found to be modified,
and some of them being clustered together, the generation of antibodies
against specific combinations of histone modifications will further our
understanding of the complex code of modifications present on the histone
molecules.
Acknowledgments
I would like to thank C. David Allis for frequent discussions and scientific support.
Portion of this work was done in the Allis lab. I would also like to thank Craig Mizzen for
excellent advice on antibody purification protocols. Finally, I would like to acknowledge
Elsevier for permission to adapt some of the figures originally published in Mol. Cell 5,
905–915 (2000) for this work.
27
M. Jaskelioff and C. L. Peterson, Nat. Cell Biol. 5, 395 (2003).
[16] Generation and Characterization of

Methyl-Lysine Histone Antibodies
By Laura Perez-Burgos, Antoine H. F. M. Peters, Susanne Opravil,
Monika Kauer, Karl Mechtler, and Thomas Jenuwein
Histones are among the most conserved proteins throughout evolution

and responsible for the compaction of DNA into the higher-order polymer,
known as chromatin. In addition to their structural roles, histones play import-
ant functions in the control of gene expression by regulating access to the
underlying nucleosomal template. Histone amino-termini (tails) protrude
from the nucleosome core and are subject to a variety of post-translational
modifications, including acetylation (on lysine residues), phosphorylation
(on serine and threonine residues), methylation (on lysine and arginine
residues), ubiquitination (on lysine residues), and ADP-ribosylation (on
glutamic acid residues).1 The various histone tail modifications have been
proposed to constitute a ‘‘histone code’’ that could significantly extend
the information potential of the genetic information.2–4
1
K. E. van Holde, ‘‘Chromatin.’’ Springer, New York, 1988.

[16] methyl-lysine histone antibodies 235
Over the last three years, histone lysine methylation (K-methylation)

has emerged as a central modification that can discriminate chromatin tran-
sitions in most of the epigenetic paradigms. These include the formation of
silent chromatin domains at pericentric heterochromatin, decoration of the
inactive X chromosome, Polycomb- and Trithorax-mediated transcrip-
tional memory, and regulation of promoter activity at euchromatic po-
sitions.5 There are at least five methylatable lysine positions in the
histone tails of H3 (K4, K9, K27, and K36) and H4 (K20)1,6 (Fig. 1A); an-
other one exists in the histone-fold domain of H3 (K79).7–10 Of those,
H3-K4, H3-K36, and H3-K79 have been primarily associated with tran-
scriptional stimulation and an activated chromatin state, whereas H3-K9,
H3-K27, and H4-K20 are linked with gene repression and silenced/
condensed chromatin domains.11–13 Particularly for the ‘‘repressive’’ lysine
methylation marks, distinct subnuclear compartments can be visualized,
thereby allowing the analysis of the relative distribution of silenced chro-
matin regions by indirect immunofluorescence (IF). Differences in the
ratio between active and repressed chromatin states have been proposed
to correlate with the developmental potential of progenitor cells, to indi-
cate the plasticity of stem cells and to be a diagnostic marker for the
subnuclear transformation that is a frequent hallmark of tumor cells.14,15
Based on these implications and broad application in many experimental
systems, methyl-lysine histone antibodies have been in high demand, both
from research laboratories and from commercial sources. However, although
the currently available methyl-lysine histone antibodies have been useful
tools to probe distinct methylation marks, they need further development
2
3
4
T. Jenuwein and C. D. Allis, Science 293, 1074 (2001).
5
6
7
Q. Feng, H. Wang, H. H. Ng, H. Erdjument-Bromage, P. Tempst, K. Struhl, and Y. Zhang,
Curr. Biol. 12, 1052 (2002).
8
N. Lacoste, R. T. Utley, J. M. Hunter, G. G. Poirier, and J. Cote, J. Biol. Chem. 277, 30421
(2002).
9
H. H. Ng, Q. Feng, H. Wang, H. Erdjument-Bromage, P. Tempst, Y. Zhang, and K. Struhl,
Genes Dev. 16, 1518 (2002).
10
F. van Leeuwen, P. R. Gafken, and D. E. Gottschling, Cell 109, 745 (2002).
11
J. C. Rice, K. Nishioka, K. Sarma, R. Steward, D. Reinberg, and C. D. Allis, Genes Dev. 16,
2225 (2002).
12
T. Kouzarides, Curr. Opin. Genet. Dev. 12, 198 (2002).
13
A. Vaquero, A. Loyola, and D. Reinberg, ‘‘The Constantly Changing Face of Chromatin.’’
Science’s SAGE KE: http://sageke.sciencemag.org/cgi/content/full/sageke;2003/14/re4, 2003.
14
M. A. Surani, Nature 414, 122 (2001).
15
P. A. Jones and S. B. Baylin, Nat. Rev. Genet. 3, 415 (2002).
236
Fig. 1. Methyl-lysine positions in histone amino-termini and peptide design. (A) Schematic representation of
methylatable lysine positions in the histone H3 and H4 amino-termini. The H3 and H4 sequences are indicated for the
murine proteins. (B) Diagram of linear and branched peptides used to generate H3-K9 di-methylation–specific
antibodies. The amino acid (aa) numbers correspond to the positions within the H3 N-terminal tail shown in panel A.
[16]
and careful controls. This is because histone lysine methylation has the
additional complexity that lysine residues can be mono-, di-, or tri-
methylated,1,16,17 with significant differences in biological output.18,19 Second,
the H3-K9 and H3-K27 positions are embedded within the same amino acid
sequence ‘‘-ARKS-’’ (see Fig. 1A), thereby allowing for considerable
cross-reactivities of the respective methyl-lysine histone antibodies.
In this chapter, we summarize our experience in the development and
characterization of rabbit polyclonal antibodies that are directed against
the methylated H3-K9 position. We compare the specificity and subnuclear
localization of several methyl H3-K9 ‘‘specific’’ antibodies that were gener-
ated in research laboratories20–22 or are available from commercial sources
(Upstate Biotech (UBI), USA and Abcam, UK). We also provide proto-
cols for peptide design, rabbit immunizations, and quality controls of
methyl-lysine histone antibodies, followed by their in vivo characterization
using indirect IF of inter- and metaphase chromatin in wild-type (wt) and
mutant mouse cells that are deficient for the Suv39h histone methyltrans-
ferases (HMTases). Our comparative analyses indicate significant discrep-
ancies in the specificity and avidity of the available methyl-lysine histone
antibodies and highlight the need for extensive quality controls, such
that experimental data can be correctly interpreted despite the exquisite
complexity of histone lysine methylation.
Results
Peptide Design and Antibody Quality Controls

The most critical step in the generation of a specific antiserum is pep-
tide (antigen) design. Peptide antigens can differ in length (6–20 amino
acids, aa), configuration (linear versus branched peptide), and methylation
16
W. K. Paik and S. Kim, Science 174, 114 (1971).
17
J. H. Waterborg, J. Biol. Chem. 268, 4918 (1993).
18
H. Santos-Rosa, R. Schneider, A. J. Bannister, J. Sherriff, B. E. Bernstein, N. C. Emre, S. L.
Schreiber, J. Mellor, and T. Kouzarides, Nature 419, 407 (2002).
19
A. H. Peters, S. Kubicek, L. Perez-Burgos, S. Opravil, K. Mechtler, A. Kohlmaier, A. Beyer,
M. Tachibana, Y. Shinkai, and T. Jenuwein, Mol. Cell, in press.
20
A. H. Peters, D. O’Carroll, H. Scherthan, K. Mechtler, S. Sauer, C. Schofer,
K. Weipoltshammer, M. Pagani, M. Lachner, A. Kohlmaier, S. Opravil, M. Doyle, M. Sibilia,
and T. Jenuwein, Cell 107, 323 (2001).
21
B. A. Boggs, P. Cheung, E. Heard, D. L. Spector, A. C. Chinault, and C. D. Allis, Nat.
Genet. 30, 73 (2002) (published online: 10 December 2001).
22
I. G. Cowell, R. Aucott, S. K. Mahadevaiah, P. S. Burgoyne, N. Huskisson, S. Bongiorni,
G. Prantera, L. Fanti, S. Pimpinelli, R. Wu, D. M. Gilbert, W. Shi, R. Fundele, H. Morrison,
P. Jeppesen, and P. B. Singh, Chromosoma 111, 22 (2002).
state (mono-, di-, or tri-) of the lysine of interest. To illustrate how different
peptide designs can affect antibody specificity, we have compared four
rabbit polyclonal antisera raised against various H3-K9 di-methylated
peptides (see Fig. 1B). These peptides comprise a linear octamer (aa
6–13) (UBI #07-212),21 a linear 12-mer (aa 23–34) (Abcam #ab7312), a
2-branched 11-mer (aa 5–15),19 and a 4-branched hexamer (aa 6–11).20
To assess the specificity of a given antiserum, three types of quality con-
trols are available. These comprise enzyme-linked immunosorbent assays
(ELISAS), peptide spotting analyses (dot blots), and protein blots with re-
combinant and nuclear histones. Whereas ELISAS or peptide and histone
blots are used for the initial characterization, additional protein blots with
whole nuclear extracts allow the detection of possible cross-reactivities
with other epitopes that may be present on non-histone molecules. We
have developed a comprehensive peptide blot analysis with a panel of 19
linear peptides spanning the histone H3 and H4 amino-termini and which
are specific for unmodified or mono-, di-, or tri-methylated H3-K4, H3-K9,
H3-K27, H3-K36, and H4-K20. A serial dilution of these peptides (50, 10, 2,
0.4 pmol) is spotted onto a Hybond-P membrane and binding of an anti-
body is visualized by enhanced chemiluminescence. The antibody titer
and avidity are determined by scoring binding to decreasing peptide
amounts. The four antisera described in this section were each probed at
three different dilutions, ranging from 1:200 to 1:2500.
The UBI antibody #07-212 displays a high affinity toward di-methylated
H3-K9, but also shows minor cross-reactivities (i.e., only when 50 pmol
peptides are spotted) with di-methylated H3-K4 and tri-methylated H3-
K27 (see Fig. 2). This profile is slightly different from a ‘‘golden bleed’’
batch, which lacks the minor cross-reactivity toward tri-methylated H3-
K27 (D. Allis, personal communication, data not shown). The Abcam
#ab7312 antibody also has a high preference for di-methylated H3-K9
and to a minor extent cross-reacts with di-methylated H3-K36 and
H4-K20 positions, and with tri-methylated H3-K27. This is a surprising
result, since the used antigen did not even comprise the H3-K9 residue
but was designed to generate an -dimeth H3-K27 antibody. The detected
cross-reactivities illustrate the difficulty to discriminate the highly similar
H3-K9 and H3-K27 positions. For each of these commercially available
antibodies, the most specific results were obtained at dilutions 1:1000
(UBI #07-212) and 1:500 (Abcam #ab7312).
In an attempt to increase specificity and antigenicity, we developed a
novel antiserum by using a 2-branched peptide that encompasses the di-
methylated H3-K9 position by five amino acids on either side. This -2-
dimeth H3-K9 antibody (#4679) has both a high titer (as shown by its use
at a 1:2500 dilution) and is exclusively specific for di-methylated H3-K9,
Fig. 2. In vitro quality control of various H3-K9 di-methyl antibodies. Immuno-dot blots
showing reactivity of antibodies raised against linear, two- and four-branched H3-K9 di-
methylated peptide antigens (lower left corners). Antibodies were probed at various dilutions,
with the most specific reactivity and dilution shown. Dot blots contain 0, 0.4, 2, 10, and
50 pmol of linear H3 (aa 1–20; aa 19–34; aa 25–45) peptides, unmodified or mono-, di- and tri-
methylated at the K4, K9, K27, or K36 positions. In addition, a linear H4 (aa 12–31) peptide,
differentially methylated at the K20 position, was also used.
without cross-reacting with H3-K27 or any other tested position (see Fig. 2).
Using a similar strategy, we had previously designed a 4-branched peptide
spanning the di-methylated H3-K9 position by only 2–3 amino acids on
either side. The rationale for this peptide was to mimic the potential in vivo
clustering of methylated histone tails at pericentric heterochromatin, which
may be organized in a more compact nucleosome structure. The resulting
-4-dimeth H3-K9 antibody (#2233) has indeed been very useful to
detect histone methylation at pericentric heterochromatin in an Suv39h-
dependent manner.20,23 At that time, dot blot analysis with only a limited
peptide panel, including unmodified and di-methylated H3-K9 and
H3-K27 peptides, categorized this antiserum as being specific for the
di-methylated H3-K9 position (see Fig. 2).
Using the complete peptide panel in dot blot analyses, however, reveals
that the -4-dimeth H3-K9 antibody has highest affinity for tri-
methylated H3-K9 and significantly cross-reacts (i.e., with 10 pmol of
spotted peptides) with di- and tri-methylated states of all tested lysine pos-
itions (see Fig. 2), even including H1-K26 di- and tri-methylated peptides
(data not shown). The promiscuity of this antibody, which may well extend
beyond histone epitopes, can be attributed to the relatively short peptide
(only 6 aa) used to present the antigen. Despite the potential to recognize
many di- and tri-methylated lysine positions, this antiserum does not uni-
formly stain interphase chromatin but selectively decorates silenced chro-
matin regions at pericentric heterochromatin and on the inactive X
chromosome (see Figs. 3 and 4). This staining pattern has been interpreted
to be a consequence of the -4-dimeth H3-K9 antibody as having a pref-
erential affinity toward a high concentration or density of methylated lysine
residues, rather than detecting more isolated methyl epitopes.20,23 Based
on its promiscuity and potential to interact with multiple methylated lysine
positions, we would like to rename the -4-dimeth H3-K9 antibody
(# 2233) as a ‘‘multi-methyl lysine’’ antiserum.
Protocol A: Generation of an H3-K9 tri-Methyl Specific

Antiserum (#4861)
Peptide Synthesis. A 2-branched peptide with the sequence 2
(QTARK(Me)3 STGGKA)-1-K-cys comprising amino acids 5–15 of
histone H3 was synthesized in a ABI 433 Peptide Synthesizer (Applied
Biosystems).
23
C. Maison, D. Bailly, A. H. Peters, J. P. Quivy, D. Roche, A. Taddei, M. Lachner,
T. Jenuwein, and G. Almouzni, Nat. Genet. 30, 329 (2002).
Fig. 3. Immunofluorescence analysis for H3-K9 di-methylation in mouse chromatin. Indirect

IF of interphase and mitotic chromatin of wt female MEFs stained with the indicated -di-meth
H3-K9 antibodies. Heterochromatic regions of the acrocentric mouse chromosomes are under-
represented for H3-K9 di-methylation (see arrowheads in the mitotic chromosome spreads)
but enriched for H3-K9 methylation with the ‘‘multi-methyl lysine’’ antibody.20 The inactive
X chromosome (see arrow Xi) is also decorated by the ‘‘multi-methyl lysine’’ antibody.24
1. Inject protected amino acids (MultiSynTech GmbH, Germany) in

reverse order at 1 mM each, as synthesis takes place from
C-terminal to N-terminal end. The first amino acid is a cysteine
(for coupling and purification purposes) that is already bound to the
resin TCP Pepchem (Goldammer & Clausen, cat. no. PC-01-0114).
Two hundred and fifty milligrams of Pepchem of low substitution
24
A. H. Peters, J. E. Mermoud, D. O’Carroll, M. Pagani, D. Schweizer, N. Brockdorff, and
T. Jenuwein, Nat. Genet. 30, 77 (2002) (published online: 10 December 2001).
Fig. 4. Immunofluorescence analysis for H3-K9 tri-methylation on mitotic chromosomes.

Indirect IF of mitotic chromosomes of wt and Suv39h dn female MEFs stained with the
indicated antibodies under prefixed (top panels) and post-fixed (bottom panels) conditions.
Pericentric heterochromatin is significantly enriched for H3-K9 tri-methylation in an Suv39h-
dependent manner. By contrast, H3-K27 tri-methylation at the inactive X chromosome (see
arrows) is Suv39h independent. Both the ‘‘multi-methyl lysine’’20 and Cowell et al.22
antibodies decorate constitutive (pericentric) and facultative (Xi) heterochromatin. Some
Suv39h dn cells are hypo-tetraploid20,24 and therefore contain two Xi.
grade (0.16 mmol/g) are used for the entire synthesis. The peptide
synthesis is monitored by conductivity measurement.
2. The second amino acid is a DI-Fmoc-lysine-OH (Bachem,
Switzerland, cat. no. B-1610), which is used to drive the synthesis
of a 2-branched peptide via its - and e-amino groups.
3. A tri-methylated Fmoc-Lysine-(Me) 3-OH chloride (Bachem,

Switzerland, cat. no. B-2685) is incorporated at the H3-K9 position.
4. Release and deprotect the completed peptide by treating the resin
with 11 ml of 90% TFA (v/v), 0.05% tri-ethylsilan (v/v), 0.05% H2O
for 60 min. Isolate the peptide by rinsing and filtering the resin
(MultySynTech GmbH, cat. no. V100PE120) with TFA.
5. Precipitate the peptide by adding it dropwise to a 35-ml cold

solution of 60% ether (v/v), 40% heptane (v/v). Store for 2 h at 80 .
Collect the precipitate by centrifugation and wash three times with
40 ml ether. Lyophilize the peptide under vacuum.
6. Dissolve the peptide in 5 ml of 0.1% (v/v) TFA in water. Inject into
a HPLC (Applied Biosystems) and purify by reverse-phase
chromatography on a C18 column (Phenomenex, lunar 5C18) using
a 45-min, 2–45% gradient of acetonitrile. The positive fractions are
identified by mass spectroscopy in linear and reflector mode (Bruker
Reflex III MALDI-TOF).

7. Freeze the peptide at 80 and lyophilize for 2 days. The
yield obtained for the 2-branched H3-K9 tri-methyl peptide was
66 mg.
Rabbit Immunizations. Screen pre-immune sera from 6–8 rabbits by dot
blots (see later) with unmodified and methylated histone peptides. Proceed
with the 2 rabbits showing the lowest background signals.
1. Combine 3 mg of peptide with 3 mg keyhole limpet hemocyanin
(KLH) (Calbiotech, cat. no. 3744805) that has been activated through
a maleimide group (Sulfo-GMBS; Pierce, cat. no. 22324) and is
dissolved in 320 l of 250 mM NaCl and 20 mM HEPES, pH 7.3. Add
340 l 1 M boric acid, pH 6.0, and 340 l 6 M guanidine chloride.
Rinse with argon to remove residual oxygen and incubate for 1 h at
RT. To assess the coupling efficiency, combine 5 l of KLH-peptide
and 5 l of Ellman’s reagent (Sigma, cat. no. D-8130) in 1 ml of 2 M
HEPES, pH 7.3; efficient coupling is indicated by a decreased signal
at 412 nm in a UV-spectrophotometer against a peptide-only blank.

2. Ship the KLH-peptide (1 ml at 6 mg/ml) at 4 to an antibody
company (e.g., Gramsch Laboratories, Schwabhausen, Germany).
3. On day 1, 400 g of KLH-peptide (i.e., 200 g peptide) in complete
Freud’s adjuvant (CFA) are injected subcutaneously into the lymph
nodes and the neck of rabbits.
4. On day 21, the animals are boosted subcutaneously in the shoulder
and the neck with 400 g of KLH-peptide in incomplete Freud’s
adjuvant (IFA).
5. On day 35, the first bleed is withdrawn (5 ml) and analyzed by dot
blots.
6. Responding rabbits are boosted intramuscularly (rear thigh) with
200 g of KLH-peptide in PBS on day 42–45, followed by additional
boosts at subsequent timepoints.
7. On day 63, the second bleed is obtained (20 ml) and additional
bleeds (20–25 ml each) can be withdrawn every other week. The
rabbits can be maintained (more boosts are required) or exsanguin-
ated (on day 85–90; preferred procedure), yielding around 80–90 ml
of crude serum.
8. To preserve the serum, fresh 10% (w/v) sodium azide is added to a
final concentration of 0.02% (w/v). The serum is then snap-frozen in

liquid N2 and kept at 80 for long-term storage in 4-ml aliquots. A

0.5-ml working aliquot is kept at 4 .
Generation of an IgG Fraction. In most cases, it is advised to generate
an IgG fraction from the crude serum, which will improve the signal to
background ratio in IF and chromatin immunoprecipitation (ChIP) anal-
yses. Although the protein A incubation can sometimes weaken antibody
avidity, a high-titer serum will almost always gain in specificity after an
IgG fractionation.
1. To de-lipify the crude serum, mix 4 ml with an equal volume of Freon
(Aldrich, cat no. 24050-8). Vortex and centrifuge three times for

10 min at 3300 rpm at 4 . Transfer supernatant containing serum to
fresh Freon after each wash. Filter the supernatant through a 0.22-m
pore-sized filter and add 5 M NaCl to a final concentration of 1 M.
2. Apply the serum to a Protein A-Porost resin column (Applied
Biosystems) and elute antibodies in two steps, using 1.5 M MgCl2,
20 mM sodium acetate, pH 5, followed by 0.1 M glycine, pH 2.45.
3. Immediately neutralize eluted fractions with 0.1 volume of 2 M
HEPES, pH 7.9, to avoid degradation of the antibodies. Pool
positive fractions (identified by O.D. at 280 nm).

4. Dialyse the 2 ml eluate for 3 h at 4 against 2 l of PBS, followed by
an overnight incubation in fresh buffer. After dialysis, the volume
increases to 5 ml.
5. Determine antibody concentration at 200–400 nm by UV scan (for
antibody #4861, it was 1.3 mg/ml).
6. Preserve and store the dialyzed IgG fraction as described for the
crude serum.
Dot Blot Analysis. ELISA and dot blots can be used equivalently to
probe antibody specificities. We prefer dot blots, simply because the same
procedure can be extended to the analysis of nuclear proteins by western

blotting.
1. Rinse a Hybond-P membrane (cat. no. RPN303F, Amersham
Biosciences) for 10 s in methanol, wash with water, and soak in PBS
until use. Identify the correct side of the membrane (characterized
by a higher surface tension) by spotting 2 l of PBS on a corner: the
drop sits on the surface for a few seconds before it is absorbed.
2. Cover a glass or Plexiglas plate with parafilm and place a sheet of
PBS-soaked Whatman paper on top. Place the Hybond-P membrane
on the Whatman paper and allow all excess liquid to run off before
proceeding to spot. Mark the membrane with blue or black ballpoint
dots and cut the upper left corner for orientation.
3. Spot 2 l each of 0, 0.2, 1, 5, and 25 M of peptide solution in PBS,
with a minimum of 1 cm between dots; peptides are spotted on the
x-axis and serial dilutions on the y-axis. After spotting, place the
membrane between two sheets of Whatman paper and allow it to
air-dry overnight.
4. On the next day, rinse the membrane for 10 s in methanol, wash in
PBS and block for 1 h in 3% (w/v) BSA in PBS (blocking solution).
5. Incubate for 1 h at RT with the appropriate primary antibody
dilution in 3 ml of blocking buffer, by placing the membrane
face-down on a parafilm-covered glass or Plexiglas plate.
6. Wash five times for 10 min in 1 TBS-T (50 mM Tris-HCl, pH 8.0,
100 mM NaCl, 0.1% (v/v) Tween 20).
7. Dilute the peroxidase-conjugated anti-rabbit secondary antibody
(Jackson Immunoresearch Laboratories, cat. no. 111-035-144) 1:2000
in 5% (w/v) non-fat milk in PBS and incubate with the membrane
for 1 h.
8. Wash five times for 10 min in 1 TBS-T (50 mM Tris-HCl, pH 8.0,
100 mM NaCl, 0.1% (v/v) Tween 20).
9. Develop using a dual ECL system (Amersham Biosciences) for
4 min and expose to scientific imaging film (Kodak X-Omat Blue
XB-1, cat. no. 165-1454).
Using the previous protocol, we have generated a highly specific, high-
titer IgG fraction for an -2-trimeth H3-K9 (# 4861) antibody. In a
similar manner, we also succeeded to generate highly selective -2-mono-
meth H3-K9 (# 4858) and -2-dimeth H3-K9 (# 4679) antibodies. All of
these antisera show negligable (i.e., only very weak signals with 50 pmol
spotted peptides) or no cross-reactivities with other methylated lysine
positions (see Table I). These antisera have further been controlled in IF
and ChIP experiments in wt and mutant (i.e., Suv39h dn) cells.19
TABLE I
246
Summary of Specificities for Methyl-Lysine Antibodiesa
Histone H3-K9 Methyl-specific antibodies H3-K27
Peters Peters Cowell Peters

Abcam UBI et al.19 et al.20 et al.22 et al.19

Mono linear Di linear Tri linear Di linear Mono 2 Di 2 Tri 2 Di 4 Tri Tri 2
(aa 23–34) (aa 23–34) (aa 5–11) (aa 6–13) (aa 5–15) (aa 5–15) (aa 5–15) (aa 6–11) linear (aa 22–33)
Antigen ab9045 ab7312 ab8898 07-212 4858 4679 4861 2233 (aa 1–20) 6523
H3-K4 Unmodified
(aa 1–20) Mono þ
Di þ þþ
Tri þ þ
H3-K9 Unmodified
(aa 1–20) Mono þþþ þþþ
Di þþþ þþþ þþþ þþ þ
Tri þþþ þþþþ þþþ þþþþ
H3-K27 Unmodified
(aa 19–34) Mono þ
Di þ
Tri þ þ þ þ þþ þþ þþþ
H3-K36 Unmodified þ þ
(aa 25–45) Mono þ
Di þ þ þþ
Tri þ þ
H4-K20 Unmodified
(aa 12–31) Mono
Di þ þþ
Tri þ þþ þþþ þ
Fraction Affinity Affinity Affinity Affinity IgG IgG IgG IgG Serum IgG
Concentration (mg/ml) 0.35 0.38 0.09 n.a. 1.7 1.7 1.3 1 n.a. 1.1
Dilution 1:1000 1:500 1:500 1:1000 1:2500 1:2500 1:2500 1:1666 1:1500 1:1500
a
Each indicated antibody has been tested at various dilutions on peptide dot blots as shown in Fig. 2. The quality control of the -2-trimeth H3-K27 antibody (#6523) is included for
comparison. The relative affinity of an antibody for a given methylated lysine residue is determined by its reactivity toward a peptide dilution series (see Fig. 2) and indicated by the number
of þ. þþþþ represents the highest antibody affinity (detecting 0.4 pmol peptide), whereas þ indicates very low-affinity (detecting 50 pmol peptide). Most of the antibodies (ab7312, UBI
07-212, #4858, #4679, #4861, #2233, Cowell et al., and #6523) have been tested by IF of mouse inter- and metaphase chromatin (see Figs. 3 and 4). Based on these IF analyses, we observed
[16]
that minor cross-reactivities (þ) can sometimes be ignored, since they do not result in significantly altered staining patterns. However, we do not know how these minor cross-reactivities
would affect readout in ChIP experiments.
n.a.: not available.
In Vivo Characterization of H3-K9 Methylation Patterns

Although ELISA or dot blot analyses provide the first level of quality
control for methyl-lysine histone antibodies, they are by no means suffi-
cient to evaluate antibody specificities in a physiological setting. For this,
IF analyses of inter- and metaphase chromatin offer a valuable platform
to discriminate the subnuclear distribution of distinct methylation marks
at constitutive (e.g., pericentric) and facultative (e.g., the inactive X chro-
mosome) heterochromatin and in euchromatin. Particularly in mouse cells,
constitutive heterochromatin can readily be visualized at one end of all 20
acrocentric chromosomes (except the Y chromosome) by intercalation
of the fluorochrome 40 ,6-diamidino-2-phenylindole (DAPI) at pericentric
A/T-rich satellite repeats.
We have tested all four -dimeth H3-K9 antibodies described in Fig. 2
by IF on interphase and mitotic chromatin present in female mouse embry-
onic fibroblasts (MEFs) (see Fig. 3). The antibody from UBI (#07-212) and
our new 2-branched antiserum (#4679) display a rather uniform staining
of interphase chromatin (upper panel) and strongly decorate the arms of
mitotic chromosomes (middle panel). Although the Abcam antibody
(#ab7312) also broadly stains interphase chromatin, it reveals a more
punctate pattern along the chromsomal arms. None of these three different
antisera show enriched staining at DAPI-dense regions, but instead are
under-represented at pericentric heterochromatin of the mitotic chromo-
somes (see arrowheads in Fig. 3). Since all three antisera have highest
affinity for di-methylated H3-K9 (see Fig. 2 and Table I), these data suggest
H3-K9 di-methylation to be a very general mark for eukaryotic chromatin.
Moreover, since all mitotic chromosomes are labeled, these results also
indicate H3-K9 di-methylation to be present at the inactive X chromosome
(Xi),21,25 although it does not appear as a prominent mark for the Xi;
(see later).
By contrast, the ‘‘multi-methyl lysine’’ antibody (# 2233) strongly dec-
orates pericentric heterochromatin both in interphase chromatin and on
mitotic chromosomes20 and is under-represented along the chromosomal
arms. In addition, it selectively stains the Xi; (see arrows in Fig. 3).24
This surprisingly distinct staining pattern is likely to be a consequence
of the promiscuity of the ‘‘multi-methyl lysine’’ antibody, but which of
the many recognized positions (see Fig. 2) could discriminate pericentric
heterochromatin and the Xi? Since H3-K4 and H3-K36 are mainly linked
25
E. Heard, C. Rougeulle, D. Arnaud, P. Avner, C. D. Allis, and D. L. Spector, Cell 107, 727
(2001).
with transcriptional activation at euchromatic positions,26 H3-K9 tri-

methylation22 and H4-K2011,27,28 or H3-K27 methylation29,30 remain as
possible candidates.
H3-K9 and H3-K27 Tri-Methylation Can Discriminate Constitutive

and Facultative Heterochromatin
We repeated the above-mentioned IF analyses of mitotic chromosomes
from wt female MEFs with a recently developed H3-K9 tri-methyl anti-
body22 and our new -2-trimeth H3-K9 antiserum (#4861). Intriguingly,
the Cowell et al. antibody22 displayed a very similar staining pattern as
the ‘‘multi-methyl lysine’’ (#2233) antibody, with prominent signals at both
pericentric heterochromatin and the Xi (see Fig. 4, top panel). By contrast,
the -2-trimeth H3-K9 antiserum (#4861) almost exclusively decorates
pericentric regions but fails to detect the Xi. Since the Cowell et al. anti-
body shows stronger cross-reactivities with tri-methylated H3-K27 peptides
as compared to antiserum #4861 (see Table I), H3-K27 tri-methylation
could be a ‘‘diagnostic’’ mark for the Xi. We, therefore, generated another
antibody that is highly specific for the tri-methylated H3-K27 position, by
following a similar stategy as outlined earlier under protocol A. This novel
-2-trimeth H3-K27 antiserum (#6523) has successfully been used to
detect the Xi30 and does not cross-react with the methylated H3-K9 pos-
ition (see Table I). Importantly, it also does not decorate pericentric
heterochromatin (see Fig. 4, top panel), indicating that tri-methylation of
two separate lysine positions (H3-K9 versus H3-K27) can discriminate
constitutive from facultative heterochromatin in mammalian cells. Using
similar assays, an H4-K20 tri-methyl antibody (UBI, #07-463) is also
not generating broad signals, but selectively stains pericentric heterochro-
matin on mitotic chromosomes (M. Lachner, J. Bone, and T. Jenuwein,
unpublished).
26
M. Hampsey and D. Reinberg, Cell 113, 429 (2003).
27
K. Nishioka, J. C. Rice, K. Sarma, H. Erdjument-Bromage, J. Werner, Y. Wang, S. Chuikov,
P. Valenzuela, P. Tempst, R. Steward, J. T. Lis, C. D. Allis, and D. Reinberg, Mol. Cell 9,
1201 (2002).
28
J. Fang, Q. Feng, C. S. Ketel, H. Wang, R. Cao, L. Xia, H. Erdjument-Bromage, P. Tempst,
J. A. Simon, and Y. Zhang, Curr. Biol. 12, 1086 (2002).
29
K. Plath, J. Fang, S. K. Mlynarczyk-Evans, R. Cao, K. A. Worringer, H. Wang, C. C. de la
Cruz, A. P. Otte, B. Panning, and Y. Zhang, Science 300, 131 (2003).
30
J. Silva, W. Mak, I. Zvetkova, R. Appanah, T. B. Nesterova, Z. Webster, A. H. Peters,
T. Jenuwein, A. P. Otte, and N. Brockdorff, Dev. Cell 4, 481 (2003).
The Ultimate Control for Methyl-Lysine Histone Antibodies

The most stringent test for any antibody specificity is to analyze
its reactivity in a wt and mutant background. To illustrate this important
control, we repeated the IF analyses with mitotic chromosomes derived
from female MEFs that are double null (dn) for both Suv39h HMTases.20
In Suv39h dn chromosomes, both the ‘‘multi-methyl lysine’’ and the
Cowell et al. antibodies loose pericentric staining but gain decoration in
the arms. However, signals on the putative Xi remain present (see arrows
in Fig. 4). IF analyses with the highly selective H3-K9 (#4861) and H3-
K27 (#6523) tri-methylation antibodies revealed the ‘‘true’’ in vivo specifi-
city of the Suv39h HMTases, whereas in Suv39h dn cells H3-K9 tri-methyl-
ation at pericentric heterochromatin is selectively abrogated, H3-K27 tri-
methylation at the Xi is unaltered. Thus, the Suv39h HMTases are H3-K9
tri-methylating enzymes with a strong preference for pericentric hetero-
chromatin, whereas Suv39h-unrelated HMTases are responsible for
H3-K27 tri-methylation of the Xi.19,30 These results also suggest that
the less-specific ‘‘multi-methyl lysine’’ and Cowell et al. antisera contain a
significant fraction of H3-K27 tri-methylation antibodies.
All of the above-mentioned IF analyses have been performed under
fixed conditions (pre-fixed). For IF as well as for ChIP, fixation may alter
epitope presentation which could affect staining and immunoprecipitation
patterns. However, cells can also be analyzed under native conditions and
fixed after the secondary antibody has been washed off (post-fixed). To
assess the influence of fixation on epitope presentation and antibody recog-
nition, we repeated the earlier analyses in wt and Suv39h dn backgrounds
under post-fixed conditions (see Fig. 4, lower panels). For all four anti-
bodies tested, we did not observe significant differences between pre-fixed
and post-fixed conditions, with the exception that the ‘‘multi-methyl
lysine’’ and Cowell et al. antibodies have a slightly reduced potential to
decorate the Xi under more native conditions.
Protocol B: Analysis of Histone Lysine Methylation by

Immunofluorescence of Mitotic Chromosomes
Mouse cells are particularly useful for IF analyses with methyl-lysine
histone antibodies, as their nuclei contain well-defined euchromatic and
heterochromatic domains that can be distinguished with the fluorochrome
DAPI. Antibodies should preferably be tested in wt and mutant cells that
are deficient for a given HMTase with known target specificities.
1. Prepare 10 KCM solution: 1.2 M KCl, 200 mM NaCl, 100 mM

Tris, pH 8, EDTA 5 mM. Adjust pH to 7.5 and autoclave. Dilute to
1 KCM with mono Q H2O.
2. Grow cells to 70–80% confluency in a maxi dish and arrest mitotic
cells at metaphase with 0.2 g colchicine (Sigma, cat. no. C-9754)

per ml for 1–2 h at 37 .
3. Perform mitotic shake-off or trypsinization, centrifuge cells for
5 min at 1100 rpm, and resuspend in a small volume of growth
medium.
4. Add 3 volumes of RBS hypotonic solution (10 mM Tris-HCl, pH

7.5, 10 mM NaCl, 5 mM MgCl2) and incubate for 15 min at 37 .
Place cells on ice for a maximum of 1 h until centrifugation.
5. Add 0.1 volume of 3% (v/v) Tween 20 in RBS to cells just prior to
centrifugation.
6. Centrifuge 100 l of cell suspension onto Superfrost slides using a
Shandon Cytospin3 (Shandon, UK) for 10 min at 2000 rpm. To
check cell density and presence of mitotic chromosomes, analyze
one test slide by DAPI staining and fluorescence microscopy.
7. After centrifugation, transfer all slides directly in KCM-T solution
(10 mM Tris, pH 8.0, 120 mM KCl, 20 mM NaCl, 0.5 mM EDTA,
0.1% (v/v) Tween 20). Extract with KCM-T containing 0.1% (v/v)
Triton X-100 for 10 min at RT and subsequently wash the slides
three times for 10 min in KCM-T.
8. Fix in freshly prepared 2% (v/v) para-formaldehyde in PBS, pH
7.4, for 10 min at RT and subsequently wash three times for 10 min
in KCM-T.
9. Wash slides three times for 10 min at RT in KCM-BT (KCM-T
containing 0.25% (w/v) BSA).
10. Block slides at RT for 30 min with 100 l of blocking buffer
(KCM-T containing 2.5% BSA (w/v) and 10% goat serum).

11. Incubate 1 h at RT or overnight at 4 with 100 l of primary
antibody (e.g., -2-tri-methyl H3-K9 (#4861) IgG fraction,
1.3 mg/ml) using dilutions from 1:500 to 1:2000 in blocking buffer.
12. Subsequently or on the next day, wash the slides three times for
10 min in KCM-BT.
13. Add 100 l of a secondary antibody diluted 1:500 in blocking buffer
(e.g., goat -rabbit IgG, Alexa Fluor488, Molecular Probes, cat. no.
A-11034). Incubate for 1 h in the dark.
14. Wash slides three times in KCM-T, followed by a last wash in PBS.
15. Embed in Vectashield containing 1.5 g DAPI (Vector Laboratories,

cat. no. H-1200) per ml and store at 4 until analysis.
16. Images are taken with an Axioplan 2 fluorescence microscope

(Zeiss, Germany).
Immunofluorescence Labeling Under Native Conditions. Both inter-
phase chromatin and mitotic chromosome spreads can also be stained
under native conditions (post-fixation). The above-mentioned protocols
are followed as described, but the fixation step (¼ step 8) is performed after
the secondary antibody is washed off (i.e., after step 15). Before embedding
in Vectashield containing DAPI (Vector Laboratories, cat. no. H-1200),
three washes of 10 min are performed with KCM-T followed by a
PBS-only wash.
Discussion
The above-mentioned examples underscore the technically complex
analysis of the functional significance for mono-, di-, and tri-methylation
of histone lysine positions in vivo. Although it is anticipated that highly
specific antibodies for each methylatable histone lysine position will be de-
veloped that can discriminate every distinct methylation state, only the mi-
nority of the currently available methyl-lysine histone antibodies fulfill all
the quality criteria and specificity controls outlined in this chapter. It is
therefore highly advised that any methyl-lysine histone antibody should
be rigorously tested in vitro and in vivo before being used in experimental
settings. Similarly, antibody companies such as Upstate Biotechnology,
USA, and Abcam, UK, would do an invaluable service to the scientific
community if they would very stringently indicate strength and weaknesses
of several of their methyl-lysine histone antibodies and, if required, stop
commercializing less-specific antibodies or sub-optimal batches.
We have taken the difficulty to discriminate between the H3-K9 and
H3-K27 positions as a quality criterion to evaluate nine different antisera
raised against the methylated H3-K9 residue (see Table I). For each anti-
serum, dot blot analyses comprising the complete peptide panel as shown
in Fig. 2 were performed. The relative affinity of an antibody for a given
methylated lysine residue was scored using reactivity toward 0.4 pmol
peptide (highest affinity; þþþþ) up to 50 pmol peptide (very low affinity;
þ). These in vitro data have then been extended and compared to in vivo
analyses by performing IF of mouse inter- and metaphase chromatin.
Although we have not done extensive IF with all antibodies, unspecific or
perturbed staining patterns became apparent if antibodies displayed cross-
reactivities in peptide dot blots ranging from high affinity (þþþþ) to
moderate/low affinity (þþ). By contrast, minor cross-reactivities (þ) can
sometimes be ignored, as they did not result in significantly altered staining
patterns (particularly for IF on mitotic chromosomes). However, we do

not know how these minor cross-reactivities may affect read-out of ChIP
experiments.
In addition to the problems outlined earlier, there are even more tech-
nical and functional complexities in analyzing the specificity of methyl-
lysine histone antibodies. For example, differences in peptide length
appear to affect antibody reactivities in dot blots (see Chapter 17) again
highlighting the need to control antibody specificity beyond peptide spotting
or ELISA assays. Moreover, it is also largely unexplored how additional
modifications would affect epitope recognition, as for example, by the sim-
ultaneous presence of H3-K9 methylation and H3-S10 phosphorylation.
Intriguingly, every methylatable histone lysine position is neighborored
by either a serine or threonine residue (D. Allis, personal communication),
thereby allowing a phosphorylation event to potentially modulate or even
mask a methylated epitope. We have tested the reactivity of the -2-
dimeth H3-K9 antibody (#4676) toward a doubly modified (H3-K9 di-
methylation and H3-S10 phosphorylation) peptide, without observing a
reduction in epitope recognition (data not shown), and are currently
extending these analyses for the mono- and tri-methyl–specific H3-K9 anti-
sera. Similarly, it is also possible that a doubly methylated H3 amino ter-
minus (e.g., H3-K9 and H3-K27 methylation) may affect epitope
presentation and antibody reactivity. Particularly for multiple histone tail
modifications, extensive in vivo analyses are required, ideally comprising
mass spectrometry of chromatin fragments that have been immunoprecipi-
tated with the respective antibodies or which were prepared during different
stages of the cell cycle. Finally, antibody development in different host species
(e.g., rabbit, chicken, mouse, lama) will provide another important tool to
analyze the presence of multiple modification marks by colocalization or
coimmunoprecipitation experiments. The availability of highly selective
methyl-lysine histone antibodies from different species will, for example,
prove advantageous for understanding the possible combinatorial contribu-
tion of H3-K9 and H3-K27 tri-methylation during Polycomb-mediated
transcriptional memory.31–34
31
R. Cao, L. Wang, H. Wang, L. Xia, H. Erdjument-Bromage, P. Tempst, R. S. Jones, and
Y. Zhang, Science 298, 1039 (2002).
32
B. Czermin, R. Melfi, D. McCabe, V. Seitz, A. Imhof, and V. Pirrotta, Cell 111, 185 (2002).
33
J. Muller, C. M. Hart, N. J. Francis, M. L. Vargas, A. Sengupta, B. Wild, E. L. Miller, M. B.
O’Connor, R. E. Kingston, and J. A. Simon, Cell 111, 197 (2002).
34
A. Kuzmichev, K. Nishioka, H. Erdjument-Bromage, P. Tempst, and D. Reinberg, Genes
Dev. 16, 2893 (2002).
Important Considerations and Technical Advice

Based on our experience, we would like to highlight the following
steps in the generation and characterization of methyl-lysine histone
antibodies:
1. Golden Rule: obtain a high-titer, position-specific antiserum that
has extensively been characterized in in vitro and in vivo analyses
and which ideally has also been controlled in wt and mutant
backgrounds.
2. To obtain a specific antiserum, the peptide length is crucial and
should be equal or greater than 10 amino acids, with the lysine of
interest centered in the middle of the sequence.
3. When branched peptides (preferably 2-branched) are used, nearly
every rabbit produces an immune response by yielding a high-titer
antiserum; in contrast, only a fraction of immunized rabbits respond
to linear peptide antigens and sometimes weakly.
4. A high-titer, position-specific antiserum (which can be used at
dilutions 1:2000) is a hallmark of a robust antibody that can be
used with confidence in assays such as protein blot, IF, and ChIP.
5. The specificity of a crude antiserum can be further improved by
performing an IgG fractionation. Although this procedure can
sometimes weaken the avidity of the antibody (particularly for
ChIP), it increases the signal to background ratio. We recommend
that a gain in specificity should be favored over a stronger output.
6. We have observed that pre-absorbing a less-specific antiserum
against peptides with which it cross-reacts, followed by affinity
purification, considerably weakens the avidity of the antibody
without a significant improvement in specificity. We therefore
recommend that new rabbits should be immunized rather than
applying extensive purification schemes that only yield suboptimal
antibody preparations. As an alternative to elaborate purification,
peptide competition may be used in some assays.
The fast-moving histone methylation field has generated many novel in-
sights that may well underpin basic mechanisms for epigenetic control.
However, many questions remain unanswered, such as, for example,
whether differences between mono-, di-, and tri-methylation of selective
lysine residues in the histone amino-termini will be of functional impor-
tance for the organization of constitutive and facultative heterochromatin,
and for the long- and short-term regulation of promoter activity. We cur-
rently have only limited knowledge about mono-methylated states, and it
is unclear whether mono-methylation could specify distinct chromatin
regions or whether it may be required to prepare a lysine position for sub-

sequent di- and tri-methylation. By contrast, di-methylation states appear
as more general marks that can broadly decorate eukaryotic chromatin.
Intriguingly, tri-methylation states are more selectively distributed along
inter- and metaphase chromatin and would be good candidates to index
subchromosomal domains. For example, H3-K4 tri-methylation has been
correlated with the promoter regions of fully activated genes18,35 but is
excluded from pericentric heterochromatin (data not shown). Similarly,
H3-K36 methylation has also been implicated in gene control during tran-
scriptional elongation,26 although H3-K36 tri-methyl antibodies are cur-
rently not available. By contrast, tri-methylation of the H3-K9, H3-K27,
and H4-K20 positions can be associated with repressive functions, since
H3-K9 tri-methylation is highly enriched at pericentric heterochroma-
tin19,22 and H3-K27 tri-methylation is a prominent mark for the Xi.19,29,30
H4-K20 di-methylation has been observed at condensed chromatin regions
during mitosis,11,27,28 and H4-K20 tri-methylation antibodies decorate
pericentric heterochromatin (M. Lachner, J. Bone, and T. Jenuwein, un-
published). It is without doubt that the development of high-quality, pos-
ition-specific methyl-lysine histone antibodies will provide important
tools for the further decoding of the epigenetic information which is,
in part, indexed by distinct methylation states of selective lysine residues
in the histone amino-termini.
Acknowledgments
We would like to thank Upstate Biotechnology (UBI, Lake Placid, NY) (http://
www.upstatebiotech.com/) and Abcam (Cambridge, UK) (http://www.abcam.com/) for
exchange of antibodies and critical comments to the manuscript. In particular, we would
like to acknowledge Judy Nisson, Rene Rice, Mary-Ann Jelinek, Thomas Jelinek, and Jim
Bone from Upstate Biotechnology and Darren Harper from Abcam. We also acknowledge
Gramsch Laboratories (Schwabhausen, Germany) (http://www.gramsch.de/) for excellent
antibody development. We are indebted to Judd Rice and David Allis for their contribution
and help in characterizing the new series of -2-mono-di-tri H3-K9 methyl-specific
antibodies. We thank Prim Singh for contributing the ‘‘Cowell et al.’’ antibody, and Danny
Reinberg and Bryan Turner for helpful discussions. We would like to thank Mathias
Madalinski for the protocol on peptide synthesis and Ines Steinmacher for help on mass
spectrometry analysis. Research in the laboratory of T.J. is supported by the I.M.P. through
Boehringer Ingelheim and by grants from the Vienna Economy Promotion Fund, the
European Union, and the GEN-AU initiative, which is financed by the Austrian Ministry of
Education, Science, and Culture.
35
H. H. Ng, F. Robert, R. A. Young, and K. Struhl, Mol. Cell 11, 709 (2003).
[17] analyzing antibodies 255
[17] Tips in Analyzing Antibodies Directed Against

Specific Histone Tail Modifications
By Kavitha Sarma, Kenichi Nishioka, and Danny Reinberg
Histone methylation has been known to exist for over 40 years1 but the
enzymes that catalyze this reaction have remained elusive until the discovery
that Suv39H1 methylates histone H3 specifically at lysine 9.2 This discov-
ery was followed by a bevy of papers describing other methyltransferases
specific for different residues and their apparent function in vivo. Histones
are methylated at lysine as well as arginine residues. Lysines can be mono-,
di-, or tri-methylated in vivo. The sites for lysine methylation on histone
H3 are 4, 9, 27, 36, and 79 (for reviews see Zhang and Reinberg (2001),3
Turner (2002),4 Bannister et al. (2002),5 Lachner et al. (2003),6 and
Vaquero et al. (2003)7) while lysines 20 and 30 remain the sole site for
methylation on H48,9,9a and H2B,10 respectively (see Fig. 1).
Studies on the effects of these modifications on gene regulation have
been greatly facilitated by the production of antibodies ‘‘specific’’ for the
modified state. The need to carefully characterize antibodies raised against
methylated histone peptides stems from the observation by several labora-
tories that these antibodies can be promiscuous depending on several factors
such as concentration, peptide context, substrate, etc. Due to this, several
papers have been subject to scrutiny in recent months as the specificity of
the antibodies used was questionable. The results of chromatin immunopre-
cipitation (ChIP) experiments became increasingly difficult to interpret
1
K. Murray, Biochemistry 3, 10 (1964).
2
S. Rea, F. Eisenhaber, D. O’Carroll, B. D. Strahl, Z. W. Sun, M. Schmid, S. Opravil,
K. Mechtler, C. P. Ponting, C. D. Allis, and T. Jenuwein, Nature 406, 593 (2000).
3
4
B. M. Turner, Cell 111, 285 (2002).
5
A. J. Bannister, R. Schneider, and T. Kouzarides, Cell 109, 801 (2002).
6
M. Lachner, R. J. O’Sullivan, and T. Jenuwein, J. Cell Sci. 116, 2117 (2003).
7
A. Vaquero, A. Loyola, and D. Reinberg, published online at http://sageke.sciencemag.org/
cgi/content/full/sageke;2003/14/re4
8
K. Nishioka, J. C. Rice, K. Sarma, H. Erdjument-Bromage, J. Werner, Y. Wang, S. Chuikov,
P. Valenzuela, P. Tempst, R. Steward, J. T. Lis, C. D. Allis, and D. Reinberg, Mol. Cell
9, 1201 (2002).
9
J. C. Rice, K. Nishioka, K. Sarma, R. Steward, D. Reinberg, and C. D. Allis, Genes Dev. 16,
2225 (2002).
9a
J. Fang, Q. Feng, C. S. Ketel, H. Wang, R. Cao, L. Xia, H. Erdjument-Bromage, P. Tempst,
J. A. Simon, and Y. Zhang, Curr. Biol. 12, 1086 (2002).
10
K. Zhang and H. Tang, J. Chromatogr. B 783, 173 (2003).

Fig. 1. A schematic representation of the various histone lysine methylation sites and their
potential degrees of methylation identified to date in vivo.
since the antibodies that had been used in these experiments were not
specific to the residue or its modified state. The ideal method to test this
would be to perform ChIP and then analyze the products that have been
immunoprecipitated by mass spectroscopy in order to confirm that the anti-
body has reacted only with the modification of interest. Since this is not
feasible, and the experiments required are too difficult to perform, initial
characterization of the antibodies and determination of optimal concentra-
tion to be used in ChIP becomes extremely important.
In this chapter, we present several parameters to be taken into consid-
eration and some useful hints for systematic characterization of antibodies
raised against methylated histone peptides. Although we have focused
on antibodies against methylated residues: H3-K4, H3-K27, and H4-K20,
the methods and procedures described herein are applicable for any anti-
body directed against the histone tail modifications, including arginine
methylation and lysine acetylation, among other modifications.
ELISA
This is the first step in antibody characterization. Analysis of the anti-
body with enzyme-linked immunosorbent assay (ELISA) is the traditional
method for characterization. The advantage of ELISA is that of all assays
described, it is the most quantitative and sensitive and the optimal concen-
tration of the antibody to be used in experiments can be ascertained, but
see later. An example of antibody characterization using ELISA is shown
in the case of polyclonal di-/tri-methyl K27 antibody (see Fig. 2). At higher
concentrations (1:300), the antibody is able to recognize both the di- and
Fig. 2. ELISA with polyclonal antibodies against di-/tri-methyl K27 (D.R.) shows almost
equal reactivity against both di- and tri-methyl K27 peptides at 1:300 antibody dilution. But
using 1:2700 dilution, the antibody is three times more preferential to di-methyl K27
compared to tri-methyl K27.
tri-methyl K27 peptides with equal efficiency, but at lower concentrations

(1:2700 dilution), the antibody is three times more preferential to the tri-
methyl peptide than to the di-methyl peptide. The ELISA should be
followed by dot blot analyses. If monoclonal antibodies are to be screened,
it is easier to perform dot blot analysis first to isolate specific clones and then
proceed to ELISA to ascertain optimal concentrations for the antibody.
Dot Blot Analysis

Care should be taken that equimolar amounts of peptide are loaded
onto the blots and this must be quantified using Ellman’s reagent. Quanti-
fication that involves staining of the membrane is ill-advised since all
peptides do not stain equally well with Coomassie blue or Ponceau.
Serum can be checked before affinity purification by dot blots. Equimo-
lar amounts of peptide should be blotted on nitrocellulose membrane
and processed in a manner similar to that used for western blot analysis
followed by detection using enhanced chemiluminescence (ECL). Initial
characterization should be performed with unmodified, mono-, di-, and tri-
methyl peptides, preferably in the same or very similar peptide context.
This provides an initial gauge of the immune response generated by the
antigen. It has been observed that, sometimes, a di-methyl peptide causes
a strong anti-monomethyl response (see Fig. 3A and B). Once it is con-
firmed that the signal is due to a di-methyl peptide, several other di-methyl
peptides should be checked for cross-reactivity to ensure that the antibody
recognizes the di-methyl peptide in a residue-specific context and not the
Fig. 3. Dot blot analyses of various antibodies raised against methylated histone peptides.
(A) Polyclonal antibodies raised against di-methyl H3 lysine K4 peptide (Upstate 07-030)
analyzed using unmodified H3 (lane 1), mono-methyl K4 (lane 2), di-methyl K4 (lane 3), and
tri-methyl K4 (lane 4). (B) Monoclonal antibodies raised against di-methyl H4K20 peptide
(D.R.) tested with unmodified H4 (lane 1), mono-methyl K20 (lane 2), di-methyl K20 (lane 3),
and tri-methyl K20 (lane 4). (C) Polyclonal antibodies raised against the branched tri-methyl
H3-K9 peptide (T.J.) analyzed for specificity using di- and tri-methyl peptides in the context of
different methylated residues. The methylation site and status of the peptides are as indicated
above the panels and the antibody used in each experiment is indicated below each panel.
di-methyl moiety alone. Recognition of a tri-methyl moiety without residue

specificity was observed with antibodies raised against the branched tri-
methyl K9, where the antibody was able to recognize tri-methyl K9,
K27, and K20 (see Fig. 3C). An example of desirable optimal conditions
is presented in Fig. 2, where we have tested anti-di-methyl H4-K20 polyclo-
nal antibodies (Upstate 07-367). Dot blot analysis with K20 peptides with
various degrees of modification shows antibody reactivity with the di-
methyl K20 peptide exclusively (see Fig. 4A). Other di-methyl peptides
were not recognized, confirming that this antibody is specific to di-methyl
K20 (see Fig. 4A). To further strengthen this observation, competition
experiments with the H4 peptides show that the antibody is competed
out with the di-methyl peptide alone and is not affected by the mono- or
the tri-methyl peptides (see Fig. 4B). Western blot analysis showed a strong
signal with the native histone H4 polypeptide, but no reactivity with the
recombinant H4 (see Fig. 4C, left panel). The signal obtained with native
Fig. 4. Characterization of polyclonal antibodies raised against di-methyl H4-K20

antibodies (Upstate 07-367) (A) Dot blot analysis to check antibody specificity using methyl
peptides as indicated above the panel. (B) Competition dot blot performed in the presence of
1 g/ml of various peptides as indicated to the right. The di-methyl signal is lost only in the
middle panel where di-methyl H4-K20 is the competing peptide. (C) Western blot analysis of the
antibody using recombinant nucleosomes labeled r (left panel, lane 1) and native oligonucleo-
somes isolated from HeLa cells labeled N (left panel, lane 2) shows reactivity with native H4 and
not with recombinant H4. Competition with di-methyl H4-K20 peptide resulted in loss of signal
on native oligonucleosomes (compare lane 2 on left and right panels). The antibody dilution used
was the same as in the dot blot analyses and competition dot blots. (D) Immunofluorescent
staining of HeLa cells with these antibodies at 1:100 dilution show a prominent nuclear signal
(panel 1) as is expected for antibodies raised against histones. This signal is completely
obliterated when the antibody is used in the presence of di-methyl K20 peptide (panel 2).
There is no change in signal when unmodified H4 peptide (panel 3) or non-specific peptides
like di- and tri-methyl K27 peptides (panels 4 and 5) are used.
histones was competed out when the western blot was performed in the
presence of the di-methyl peptide (see Fig. 4C, right panel). Immunofluo-
rescence studies also showed localization of the antibody to the nucleus
and this was lost on addition of the di-methyl K20 peptide at the time of
staining (see Fig. 4D, panel 2). Addition of non-specific peptides, in this
case di-methyl or tri-methyl H3-K27, did not affect staining (see Fig. 4D,
panels 4 and 5). Similar results were obtained using peptides containing
mono- and tri-methyl H4-K20 (data not shown).
Some problems that have been encountered in our laboratory during
the characterization of antibodies using dot blots are as follows. While
the antibody may be extremely specific for a given antigen, when the
context of the peptide is changed, the antibody is then able to recognize
even the unmodified peptide. This was seen in the case of antibodies
against methylated H3-K9, where, in the context of residues 4–15, the anti-
bodies remained specific for di- and tri-methyl K9, but the antibodies were
found to react with unmodified H3 peptide of a longer length, that is,
residues 4–32 (see Fig. 5A and B).
It is also very important to take into consideration the context of the
peptide when characterizing antibodies against the methylated H3-K9 and
H3-K27 residues. Since both of these residues are contained within a very
similar context, that is, TKQTARK9S for K9 and TKQTAARK27S for
K27, it is very likely that the antibodies will cross-react with the different
peptides. It becomes crucial then, to check every K9 antibody prepared
for cross-reactivity with K27 peptides and vice versa. Several initial reports
regarding X chromosome inactivation in mammals focused on H3-K9 meth-
ylation as the early event in X inactivation.11,12 The paper by Heard et al.11
shows that di-methylation at K9 is present on the inactive X chromosome; this
is a noteworthy point because the antibodies used in this case were the anti-
dimethyl K9 ‘‘Golden Bunny’’ antibody from David Allis’ lab and the
anti-dimethyl K9 antibodies from Upstate Biotechnology. Both of these
have been extensively characterized and have been shown to be dimethyl
K9 specific and show no cross-reactivity with other modifications. But re-
cently it has been shown without doubt that the prominent mark for the es-
tablishment of the inactive X chromosome is tri-methylation at H3-K27.13,14
11
E. Heard, C. Rougeulle, D. Arnaud, P. Avner, C. D. Allis, and D. L. Spector, Cell 107,
727 (2001).
12
T. Jenuwein, Nat. Genet. 30, 77 (2002).
13
K. Plath, J. Fang, S. K. Mlynarczyk-Evans, R. Cao, K. A. Worringer, H. Wang, C. C. de la
Cruz, A. P. Otte, B. Panning, and Y. Zhang, Science 300, 131 (2003).
14
Fig. 5. (A) Polyclonal antibodies raised against di-methyl H3-K9 before affinity
purification showed strong reactivity with the tri-methyl K9 peptide within the same context
(lane 4) and weak reactivity with either unmodified H3 within the same context (lane 1), di-
methyl K9 (lane 3), or the longer unmodified H3 peptide (lane 5). There was no cross-
reactivity seen with mono-methyl K9, mono-, di-, or tri-methyl K27 (lanes 2, 6, 7, and 8,
respectively). (B) After affinity purification on the di-methyl H3K9 column, the antibody is
able to recognize predominantly tri-methyl K9 (lane 5) and almost no reactivity is seen with
di-methyl K9 (lane 4). A strong signal is still seen with the longer unmodified H3 (lane 2)
while the shorter peptide showed no cross-reactivity. This antibody was used at a dilution of
1:1000. (C) ELISA analysis with the same antibody after affinity purification also confirmed
that it recognizes the tri-methyl K9 modification but not di-methyl K9, K27 peptides with
different degrees of methylation or the shorter unmodified H3 peptide. (D) Polyclonal
antibodies raised against the di-methyl H3-K27 showed specific reactivity with di- and tri-
methyl K27 (top panel), but not against methyl K9 peptides (middle panel). Competition with
the tri-methyl K27 peptide completely eliminates the signal with di-methyl K27 peptide also
(bottom panel). (E) Polyclonal antibodies raised against di-/tri-methyl H3-K27 are unable to
recognize or react to an insignificant level with di- and tri-methyl K4 and K20 peptides, as
indicated above the panel.
Therefore, while it is still likely that methylation of K9 may play a role in

this process (for example, in maintenance, rather than establishment, of
X-inactivation), the major modification associated with the establishment
of inactive X chromosome is methylation of K27 catalyzed by the PRC2
complex or the enhancer of Zeste complex.15–18 This was not seen earlier,
because antibodies raised against methylated H3-K9 cross-reacted with
methylated H3-K27.
The antibody must always be checked before and after affinity purifica-
tion, since the specificity or preference for the substrate may be altered.
This was seen in the case of the di-methyl H3-K9 antibody where before
affinity purification, the antibody could recognize both the di- and tri-
methyl forms of K9, with a preference for the tri-methylated form (see
Fig. 5A), but after affinity purification using bound di-methyl K9 peptide,
the antibody was now found to be tri-methyl K9 specific and no longer
showed cross-reactivity with the di-methylated form (see Fig. 5B); although
cross-reactivity was still seen with the longer form of the H3 unmodified
peptide, amino acids 4–32. The same result was obtained when the
antibody was checked by ELISA (see Fig. 5C).
Affinity purification must be performed with the antigen that was used
for generating the immune response. In most cases, pre-clearing the anti-
body using peptides showing lower reactivity in order to enrich for a
specific antibody, seems to result in complete loss of signal. An example
of this is the case of the methyl H3-K27 antibody which recognizes both
the di- and tri-methylated forms of K27 (see Fig. 5D and E). This antibody
was affinity purified using the di-methyl K27 peptide. Theoretically, in a
competition experiment with tri-methyl peptide, only the tri-methyl signal
should be reduced without any effect on the di-methyl signal. But as shown
in Fig. 5D, the di-methylated signal is also greatly reduced, suggesting that
the epitope recognized by the antibody is common to the di-as well as tri-
methylated H3-K27 peptide. The effect observed is essentially equivalent
to having pre-cleared this antibody on a tri-methyl peptide column.
Western Blot Analysis

After a particular antibody has been analyzed by either or both of the
above-mentioned techniques, it must next be tested by western blot. This
15
Dev. 16, 2893 (2002).
16
Y. Zhang, Science 298, 1039 (2002).
17
B. Czermin, R. Melfi, D. McCabe, V. Seitz, A. Imhof, and V. Pirrotta, Cell 111, 185 (2002).
18
J. Muller, C. M. Hart, N. J. Francis, M. L. Vargas, A. Sengupta, B. Wild, E. L. Miller, M. B.
O’Connor, R. E. Kingston, and J. A. Simon, Cell 111, 197 (2002).
must be tested with recombinant and native histones run side by side on an
SDS-PAGE. A good antibody should not show any reactivity with the re-
combinant histone polypeptide (see Fig. 4C).
Sometimes, an antibody that does not exhibit reactivity with the un-
modified peptide during dot blot (Fig. 6A) or ELISA analyses will give
strong positive results with the recombinant histone polypeptide in western
blots (see Fig. 6B).
Another example is shown in Fig. 6C, in which the polyclonal antibody
generated against histone tri-methyl H4-K20 showed no reactivity with
other tri-methyl peptides during dot blot analysis. When it was then tested
by western blot analysis, it showed no reactivity with the recombinant
H4 or H3 polypeptides. But there was a strong signal seen with native H4
polypeptides and, surprisingly, with native H3 polypeptides (see Fig. 6D).
This could mean that there may be some other modification on the native
H3 that is recognized by the antibody, one that we have not tested or one
that is as yet unknown. Thus, during characterization of these antibodies,
one cannot rule out the possibility that the antibody may recognize some
modification in native histones that has not been tested or identified.
An ideal set of controls to include in the western blot characterization
is to use a series of recombinant histone polypeptides having specific
modifications that were chemically incorporated, using the technology
described by Loyola et al. in this series or that described by the Peterson
laboratory.19
Immunofluorescence
In order to test if the antibody is able to recognize its target in vivo, it
becomes essential to stain cells with this antibody. Optimal concentrations
must be determined by titration of the antibody. The knowledge that the
inactive X chromosome is enriched in H3-K27 tri-methylation facilitates
characterization of good antibodies as the inactive X signal can be identi-
fied during staining along with the loss of such a signal obtained with com-
petition using the tri-methyl peptide. For example, in Fig. 7 in which
monoclonal antibodies raised against di-/tri-methyl H3-K27 were raised,
the inactive X chromosome is detected as a strong signal in mouse embry-
onic fibroblasts (MEFs). The specificity of these antibodies in vivo was con-
firmed by competition experiments performed with methylated H3-K9
peptides. This characterization can become difficult, however, when the
exact function of the modification is unknown. One can detect heterochro-
matic or euchromatic localization depending on the modification. The most
19
M. A. Shrogen-Knaak, C. J. Fry, and C. L. Peterson, J. Biol. Chem. 278, 15744 (2003).
Fig. 6. (A) Dot blot analysis of polyclonal antibodies raised against mono-methyl H4-K20
(Abcam ab9051) shows very high specificity to the mono-methyl K20 peptide (lanes 2 and 8)
and does not recognize unmodified H4, di-methyl K20, or tri-methyl K20 peptides (lanes 1, 3,
and 4). They do not react with mono-methyl K4, K9 or K27 peptides (lanes 5, 6, and 7).
(B) Antibodies raised against mono-methyl H4-K20 (ab9051) were analyzed by western blot
analysis using recombinant nucleosomes (r) and native oligonucleosomes (N) from HeLa
cells. They recognized both unmodified and native forms of H4 equally well (lanes 1 and 2).
(C) Polyclonal antibodies raised against tri-methyl H4-K20 (D.R.) were found to be extremely
specific for the tri-methyl K20 (lane 4) and showed no reactivity with either unmodified H4, or
mono- or di-methyl K20 (lanes 1, 2, and 3, respectively) by dot blot. (D) Western blot analysis
with antibodies raised against tri-methyl H4-K20 (D.R.) shows a strong signal with the native
H4 and H3 (lane 2) but no cross-reactivity with recombinant nucleosomes (lane 1).
general protocol to characterize this is to detect nuclear staining in cells

and loss of such signal upon competition with the specific antigen, but not
with a series of peptides containing other modifications.
Immunofluorescence is also the best method to test the in vivo specific-
ity of the antibody. To test this, competition experiments must be per-
formed with the antigen, as well as the unmodified peptide within the
same context, and also with modified peptides within different contexts.
Only if the signal is lost with the antigen alone can the antibody be used
with confidence in in vivo experiments (as shown in Figs. 4D and 7). A list
of antibodies that have been tested in our laboratory by all the methods de-
scribed earlier has been summarized in Table I along with their advantages
and shortcomings for various experiments.
[17]
analyzing antibodies
Fig. 7. Immunostaining of female mouse embryonic fibroblasts (MEFs) with monoclonal di-/tri-methyl K27
antibodies. Panel 1 shows staining of MEFs with di-/tri-methyl K27 antibody with a prominent signal that correlated
with the inactive X chromosome in independent experiments (for details refer to the chapter by J. Chaumeil et al. in this
volume). Panels 2 and 3 show the result of antibody competition with the di- and tri-methyl K27 peptides where the signal
on the inactive X is lost. Competition with di- and tri-methyl K9 (panels 4 and 5) and unmodified H3 peptides of different
lengths (panels 6–8) do not affect the signal obtained with this antibody, demonstrating the specificity of these antibodies
in vivo. The concentration of peptide used for competition was 100 g/ml, as indicated.
265
TABLE I
Upstate
Antibody Biotechnology Abcam D.R. Others
H3-K4
Mono- N/A Highly specific to N/A N/A
methyl mono-methyl
K4 only
Di- Methyl H3-K4 Methyl H3-K4 N/A N/A
methyl specific but specific but
reacts with reacts with
mono- and mono- and
di-K4 di-K4
Tri- N/A Highly specific Highly specific N/A
methyl to tri-methyl to tri-methyl
K4 only K4 only
H3-K27
Mono- N/A N/A N/A N/A
methyl
Di-methyl Highly specific N/A Methyl H3-K27 N/A
to di-methyl specific but
K27 only reacts with
di- and tri-
methyl K27
Tri- N/A N/A Methyl H3-K27 Highly specific
methyl specific but to tri-methyl
reacts with K27 only
di- and tri- (Thomas
methyl K27 Jenuwein’s
laboratory)
H4-K20
Mono- N/A Mono-methyl N/A N/A
methyl K20 specific
in dot blots
but cross-reacts
with recombinant
H4 in western
Di- Highly specific N/A Specific to N/A
methyl to di-methyl methyl K20
K20 only but cross-reacts
with mono-
methyl K20
Tri- Highly specific in Shows Shows N/A
methyl dot blot analysis cross-reactivity cross-reactivity
but shows slight with di-methyl with mono- and
reactivity to K20 di- methyl K20
trimethyl K9 in
western blot
Note: Some antibodies that are now available in various companies have been marked N/A
since they were not in circulation at the time this manuscript was prepared.
Concluding Remarks
Even after antibodies have been characterized by the techniques de-
scribed herein, one can perform additional experiments to confirm the anti-
body specificity. Another system that can be used for characterization of
some of these antibodies is Saccharomyces cerevisiae. This becomes espe-
cially useful when the modification to be studied exists in yeast. A yeast
strain in which the H3 and H4 chromosomal copies have been deleted
and the only source of H3 and H4 is through a plasmid-borne copy, can
be used to introduce point mutations.20 For example, H3-K4 can be mu-
tated and histones can be extracted from this strain and used for analysis
of the methyl K4 modifications. Unfortunately, not all methyl lysine modi-
fications observed in higher eukaryotes exist in S. cerevisiae such as H3-K9,
H3-K27, and H4-K20.
It has become evident that all studies that address the functional rele-
vance of the different methylated states of histones should be undertaken
only after very careful analysis of the antibody being used. The antibodies
have to be characterized by as many methods as possible and with as many
controls as possible. The number of parameters that affect the specificity of
the antibody are too many and any result that is derived from ChIP and IF
experiments must be interpreted with extreme caution.
Materials and Methods
Antibodies and Peptides

Polyclonal antibodies were obtained from the following sources:
di-methyl H3K4 (Upstate 07-030), branched tri-methyl H3-K9 (Thomas
Jenuwein), di-methyl H4-K20 (Upstate 07-367), mono-methyl H4-K20
(Abcam ab9051). Polyclonal antibodies against di-methyl H3-K27,
di-methyl H3-K9 and tri-methyl H4-K20 were generated in D.R.’s labora-
tory. Monoclonal antibodes against di-methyl H3-K27 and di-methyl
H4-K20 were generated by Bios Chile.
The contexts of the various human histone peptides synthesized
(Global Peptide Services) were as follows: H3-K4 (amino acids 1–8), H3-
K9 (amino acids 4–15 and 4–32), H3-K27 (amino acids 22–30), and
H4-K20 (amino acids 17–28).
20
W. Zhang, J. R. Bone, D. G. Edmondson, B. M. Turner, and S. Y. Roth, EMBO J. 17, 3155
(1998).
Dot Blot and Western Blot

For dot blot, load 1 l of 0.5 mM peptide solution on nitrocellulose
membrane and dry completely. Block with 3% milk in TTBS (10 mM Tris,
200 mM NaCl, and 0.05% Tween 20, pH 7.9). Wash and incubate in
primary antibody at 1:1000 dilution (or as directed by the manufacturer)
for 2 h at room temperature. Wash with TTBS and incubate in 1:5000 dilu-
tion of secondary antibody conjugated to horseradish peroxidase (HRP)
for 40 min. Wash thoroughly three times, 10 min each time and develop
using ECL. Cover membrane with plastic wrap and expose to film for
various times, ranging from 2 s to 1 min. For western blot, load 2 g of
recombinant and native core histones onto a 15% SDS-PAGE, transfer
to nitrocellulose membrane and stain with Ponceau to ensure efficient
transfer. Proceed as described earlier for dot blot. For competition experi-
ments, add peptide to a final concentration of 1 g/ml to the primary
antibody dilution.
ELISA
Peptide solution was prepared at a concentration of 5 g/ml in PBS and
50 l were added to 96-well plates (Costar) and left overnight at room tem-
perature to coat the plates with antigen. All subsequent procedures were
performed at room temperature. The plates were then washed with PBST
(1 PBS with 0.05% Tween 20) twice, for 5 min each time and incubated in
blocking solution (5% bovine serum albumin in PBST) for 30 min. Plates
were then washed with PBST twice for 10 min each time and 200 l pri-
mary antibody was added at the following dilutions: 1:300, 1:900, 1:2700,
1:8100, 1:24300, 1:72900, and 1:218700. A non-specific polyclonal or mono-
clonal antibody was used as a control at 1:300 dilution. The plates were in-
cubated with primary antibody for 2 h. The plates were then washed three
times with PBST for 10 min each time. Fifty microliters of secondary anti-
body conjugated to alkaline phosphatase were added at a dilution of 1:5000
and incubated for 2 h. The plates were washed thoroughly with PBST three
times with vortexing the last time. Fifty microliters of developing solution
(Sigma Fast p-Nitrophenyl Phosphate tablet sets) were added to each well
and color was allowed to develop. The reaction was stopped by adding
50 l of 3 N NAOH. The plates were read at 405 nm on an ELISA plate
reader (Tecan).
Immunofluorescence

HeLa cells were grown overnight at 37 on cover slips. All procedures
after this were performed at room temperature. Cells were fixed with 3.7%
[18] histone methylation 269
formaldehyde in PBS for 10 min, washed twice in PBS, and then perm-
eabilized with 0.1% Triton X-100 in PBS for 10 min. Cells can then be pro-
cessed immediately for staining or can be stored in PBS for a few hours.
Cells were incubated with blocking solution (3% bovine serum albumin
in PBS) for 30 min, washed, and then primary antibody was added at
appropriate concentrations (1:30 to 1:200 dilution). The optimal concentra-
tion is determined by titration. After 1 h in solution containing primary
antibody, coverslips were washed in PBS and incubated with second-
ary antibody conjugated to rhodamine (Santa Cruz) at 1:100 dilution
for 20 min. Cover slips were washed thoroughly in PBS, three times for
10 min each time before mounting with Vectashield mounting medium
with DAPI.
For immunofluorescent staining of MEFs, refer to chapter by
J. Chaumeil et al. in this volume.
Acknowledgments
We thank Dr. Edith Heard for providing the figure with immunostaining of MEFs and
Dr. Lynne Vales for critical reading of the manuscript. We also thank Upstate Biotechnology
and Dr. Thomas Jenuwein for kindly providing us with antibodies for testing and members of
the Reinberg laboratory for helpful discussions. This work was supported by grants to D.R.
(NIH GM37120) and the Howard Hughes Medical School.
[18] Histone Methylation: Recognizing the Methyl Mark

By Andrew J. Bannister and Tony Kouzarides
Histone N-terminal tails are subject to a variety of covalent modifica-

tions that ultimately affect chromatin structure.1,2 One such modification
is N-methylation, which occurs on Lys (K) and Arg (R) amino acids. The
enzymes performing these modifications are the histone N-methyltrans-
ferases (HMTs) that catalyze the transfer of a methyl group from s-adeno-
sylmethionine (SAM) to the e-amino groups of lysine and/or arginine
residues within histones. Many of the methylated sites within the histone
N-termini have now been mapped (see Fig. 1).
1
S. L. Berger, Curr. Opin. Genet. Dev. 12, 142 (2002).
2
B. D. Strahl and C. D. Allis, Nature (London) 403, 41 (2000).

formaldehyde in PBS for 10 min, washed twice in PBS, and then perm-
eabilized with 0.1% Triton X-100 in PBS for 10 min. Cells can then be pro-
cessed immediately for staining or can be stored in PBS for a few hours.
Cells were incubated with blocking solution (3% bovine serum albumin
in PBS) for 30 min, washed, and then primary antibody was added at
appropriate concentrations (1:30 to 1:200 dilution). The optimal concentra-
tion is determined by titration. After 1 h in solution containing primary
antibody, coverslips were washed in PBS and incubated with second-
ary antibody conjugated to rhodamine (Santa Cruz) at 1:100 dilution
for 20 min. Cover slips were washed thoroughly in PBS, three times for
10 min each time before mounting with Vectashield mounting medium
with DAPI.
For immunofluorescent staining of MEFs, refer to chapter by
J. Chaumeil et al. in this volume.
Acknowledgments
We thank Dr. Edith Heard for providing the figure with immunostaining of MEFs and
Dr. Lynne Vales for critical reading of the manuscript. We also thank Upstate Biotechnology
and Dr. Thomas Jenuwein for kindly providing us with antibodies for testing and members of
the Reinberg laboratory for helpful discussions. This work was supported by grants to D.R.
(NIH GM37120) and the Howard Hughes Medical School.
[18] Histone Methylation: Recognizing the Methyl Mark

By Andrew J. Bannister and Tony Kouzarides
Histone N-terminal tails are subject to a variety of covalent modifica-

tions that ultimately affect chromatin structure.1,2 One such modification
is N-methylation, which occurs on Lys (K) and Arg (R) amino acids. The
enzymes performing these modifications are the histone N-methyltrans-
ferases (HMTs) that catalyze the transfer of a methyl group from s-adeno-
sylmethionine (SAM) to the e-amino groups of lysine and/or arginine
residues within histones. Many of the methylated sites within the histone
N-termini have now been mapped (see Fig. 1).
1
S. L. Berger, Curr. Opin. Genet. Dev. 12, 142 (2002).
2
B. D. Strahl and C. D. Allis, Nature (London) 403, 41 (2000).

Fig. 1. Known sites of methylation within mammalian histone H3 and histone H4

N-terminal tails. Note that lysines may be mono-, di-, or tri-methylated, whereas arginines
may be mono-, symmetric di-, or asymmetric di-methylated (see Fig. 2).
HMTs can be divided into two groups: (i) histone lysine N-methyltrans-
ferases and (ii) histone arginine N-methyltransferases.
Histone Lysine N-Methyltransferases
All of the enzymes known to methylate lysines within histone
N-terminal tails contain a conserved methyltransferase domain termed an
SET [Su(var)3-9, Enhancer-of-zeste, Trithorax] domain.3 Comparison of
various SET domain proteins has allowed them to be classified into four
subfamilies.4 Recent evidence now strongly implicates these enzymes in
human disease, especially cancer.5,6
In vivo methylated lysines may be found in the mono-, di-, or tri-
methylated state (see Fig. 2A). It is becoming clear that these states of
methylation have differing effects with respect to chromatin structure and
transcription. For instance, di-methylation of lysine 4 of histone H3
(K4H3) is found at inactive genes as well as at active genes, whereas tri-
methylated K4H3 is found predominantly at active genes. Thus, high levels
of tri-methyl K4H3 appear to ‘‘mark’’ active genes.7
The recent solving of the crystal structure of the SET7/9 methyltrans-
ferase revealed a clue as to how the different methyl states of lysine
are regulated.8 This enzyme specifically mono-methylates K4H3. It is the
3
4
T. Kouzarides, Curr. Opin. Gen. Dev. 12, 198 (2002).
5
S. Varambally, S. M. Dhanasekaran, M. Zhou, T. R. Barrette, C. Kumar-Sinha, M. G.
Sanda, D. Ghosh, K. J. Pienta, R. G. Sewalt, A. P. Otte, M. A. Rubin, and A. M.
Chinnaiyan, Nature (London) 419, 572 (2002).
6
R. Schneider, A. J. Bannister, and T. Kouzarides, Trends Biochem. Sci. 27, 396 (2002).
7
H. Santos-Rosa, R. Schneider, A. J. Bannister, J. Sherriff, B. E. Bernstein, N. C. Emre, S. L.
Schreiber, J. Mellor, and T. Kouzarides, Nature (London) 419, 407 (2002).
8
B. Xiao, C. Jing, J. R. Wilson, P. A. Walker, N. Vasisht, G. Kelly, S. Howell, I. A. Taylor,
G. M. Blackburn, and S. J. Gamblin, Nature (London) 421, 652 (2003).
Fig. 2. Chemical structures of methylated states of lysine (A) and arginine (B).
positioning of a tyrosine within the active site of the enzyme that prevents
the site accommodating di-methylated K4H3, which confers the specificity.
Indeed, if the tyrosine is mutated, the active site can now accommodate di-
methyl lysine and the mutated SET7/9 is now able to di- and tri-methylate
K4H3. Perhaps in vivo an allosteric change within the SET7/9 catalytic site
regulates this ‘‘switching’’ activity. Such a change could be introduced, for
example, by post-translational modification of SET7/9 or, alternatively,
protein binding partners may have an effect. Understanding the function
and interplay of these modifications requires the generation of specific
antibodies that are capable of distinguishing the various methylated forms
in vivo.7,9
9
T. Jenuwein, and G. Almouzni, Nat. Genet. 30, 329 (2002).
Histone Arginine N-Methyltransferases

Enzymes that methylate arginines in histones are referred to a histone
PRMTs (protein arginine methyltransferases). These enzymes are divided
into two classes, Type I and Type II, depending on whether they catalyze
assymmetric or symmetric dimethylation of arginines (see Fig. 2B). To
date, the best characterized histone PRMTs (PRMT1 and PRMT4
[CARM1]) both belong to the Type I PRMTs.10,11 Little is currently known
concerning how arginine methylation regulates chromatin structure but it is
strongly implicated, at least, because many histone PRMTs are known
transcriptional activators.12,13 For instance, at the estrogen-regulated pS2
promoter addition of hormone induces chromatin remodeling and tran-
scriptional activation.14 These events are tightly linked to the recruitment
of CARM1 to the pS2 promoter and the subsequent appearance of
R17H3 methylation.15,16
Perhaps the most powerful tools in the study of histone methylation are
site-specific antibodies capable of discriminating between the different
methylated forms. These can then be used in numerous analyses aimed at
unraveling the complexities of histone methylation. Of course, analysis of
the HMTs themselves is often desirable. Here we also describe various
methods and protocols that are useful in the study of the enzymatic activity
associated with histone methyltransferases.
Methods
Raising of Polyclonal Rabbit Antisera that Recognize

Methylated Histones
Choice of Inoculating Antigen

Generally speaking there are two possibilities concerning the choice of
antigen:
10
11
J. M. Aletta, T. R. Cimato, and M. J. Ettinger, Trends Biochem. Sci. 23, 89 (1998).
12
M. R. Stallcup, Oncogene 20, 3014 (2001).
13
D. Chen, H. Ma, H. Hong, S. S. Koh, S. M. Huang, B. T. Schurter, D. W. Aswad, and M. R.
Stallcup, Science 284, 2174 (1999).
14
G. F. Sewack, T. W. Ellis, and U Hansen, Mol. Cell. Biol. 21, 1404 (2001).
15
U. M. Bauer, S. Daujat, S. J. Nielsen, K. Nightingale, and T. Kouzarides, EMBO Rep. 3,
39 (2002).
16
S. Daujat, U.-M. Bauer, V. Shah, B. Turner, S. Berger, and T. Kouzarides, Curr. Biol. 12,
2090 (2002).
1. Generation of site-specific antibodies: If the actual site of methyla-

tion within a histone has been determined (e.g., by mass spectros-
copy), a sequence-specific peptide may be synthesized that contains
a methylated lysine or arginine at the appropriate position.
2. Generation of pan–anti-methylated amino acid antibodies: A more
general anti-methylated amino acid antibody may be desired that is
capable of recognizing a broad spectrum of proteins that contain
methylated arginines or lysines.
Synthesis of Peptides Containing Methylated Lysines or Arginines

Once the site of methylation within a histone has been determined (e.g.,
by mass spectroscopy), the investigator can chemically synthesize a peptide
containing a methylated lysine or arginine at the appropriate position.
For this approach the investigator needs to decide to what degree the lysine
or arginine is methylated (see Fig. 2). In other words, for a lysine residue is
it mono-, di-, or tri-methylated, and for an arginine residue is it mono-,
symmetric di-, or asymmetric di-methylated? Without additional informa-
tion (e.g., from mass spectroscopy data) this is a very difficult decision
for histone lysines because they exist in all three states in vivo. As a first
approach, synthesizing a peptide containing di-methyl lysine at the appro-
priate position is a good starting point. For histone arginines the situation is
perhaps a little simpler; to date all the methylated arginines in histone tails
are asymmetric di-methylated.
Another consideration is the length of the peptide to be immunized. We
typically synthesize peptides that contain three or four specific amino acids
on each side of the methylated residue. The exact length actually depends
on the results from an analysis of the sequence by the Protean epitope pre-
diction package that identifies optimal immunogenic response. However,
we have also synthesized relatively long peptides (16 amino acids) that
have been used successfully to generate specific anti-methylated histone
antibodies. We generally incorporate a Cys residue at the C-terminal end
of the peptide that is used for coupling to KHL cyanin prior to immuniza-
tion and is also used for coupling the peptide to a column matrix to aid
affinity purification of antibodies.
Peptide Synthesis
Peptides are synthesized using Fmoc chemistry17 on an Applied Biosys-
tems 9050þ or Pioneer automatic peptide synthesizer with customized
17
E. Atherton and R. C. Sheppard, eds., ‘‘Solid Phase Peptide Synthesis.’’ IRL Press (Oxford
University) 1989.
protocols. Basically, N-a-Fmoc amino acids, with side chain protection

where necessary, are coupled in the specified order to a polystyrene resin
with polyethylene glycol units attached as spacers (PEG-PS). The resin
is obtained with the C-terminal amino acid already attached via a
cleavable ester bond with a typical substitution of 0.2 mmol/g and the
peptides are synthesized from C- to N-terminus. The amino acid side
chain protections used are t-butyl for ser, thr, glu, and asp; trityl for cys,
asn, gln, and his; 2,2,5,7,8-pentamethylchroman-6-sulfonyl (pmc) for arg;
t-butoxycarbonyl for lys and trp. For peptides containing side chain meth-
ylated lysines no protecting groups are used except in the case of mono-
methyl where t-butoxycarbonyl is used. Reagents are purchased from
Bachem (UK).
A 4-fold excess of amino acid to resin substitution is used and coupling
is performed using HBTU (benzotriazol-N,N,N1,N1-tetramethyluronium
hexafluorophosphate) chemistry—1 equivalent of HBTU, 1-hydroxy-
benzotriazole (HOBt), and 2 equivalents of N-methylmorpholine are
dissolved with the protected amino acid in dimethylformamide (DMF)
and added to the DMF solvated resin where the previous amino acid
has been N-a-deprotected (Fmoc removed with 20% piperidine solution
in DMF) exposing a free amino terminus. The final Fmoc group is
removed and the resin is washed with methanol and dichloromethane
and then dried under high vacuum.
The peptide is released from the resin (500 mg) during a 2-h incubation
with a cocktail (50 ml) of trifluoroacetic acid (Tfa) 92.5%, phenol 2%, water
2%, ethanedithiol 2%, anisole 1%, and triisobutylsilane 1%. The peptide is
then concentrated and ether precipitated and collected by centrifugation,
washed three times in ether and then vacuum dried for 30 min.
Characterization of the peptide is performed using reversed-phase
HPLC on a C18 silica wide-pore column (15 cm 2.1 mm) with an elution
gradient of 10–50 acetonitrile containing 0.1% Tfa over 40 min. Further
purification (where necessary) is performed on a 25 cm 22 mm column
using the same conditions. The molecular weight of the peptide is confirmed
by mass spectrometry performed on a VG Quattro triple quadrupole
instrument with electrospray ionisation. We find that peptides with a purity
of >80% make good immunogens and work well in peptide competition
assays (see later).
The above-mentioned approach is very useful for generating site-
specific, context-dependent antibodies. However, for many applications
an investigator may require an antibody that generally recognizes methyl-
ated lysines/arginines within different amino acid contexts. One way to
generate such an antibody is to inoculate rabbits with a chemically synthe-
sized peptide that contains a methylated lysine/arginine within the context
of randomized amino acids. To do this the peptide synthesis is altered

such that the process randomly incorporates amino acids except at the
methylated position.
An alternate approach to generate a general pan-methylated lysine
antibody is to chemically methylate lysines within proteins in vitro and then
use this substrate to inoculate with. Because it contains methylated lysines
in a wide variety of sequence contexts, the immune response often gener-
ates antibodies that generally recognize methylated lysines in a variety
of contexts.
Chemical Methylation of Lysine Residues

A very efficient way of generating in vitro proteins containing methyl-
ated lysines is via reductive methylation with formaldehyde. Only lysine
residues are modified and the main product is e-N,N-di-methyllysine.18 If
radioactive formaldehyde is used (e.g., 14C-formaldehyde), then the radio-
nuclide is incorporated in the methyl groups. The use of radioactive formal-
dehyde is of course optional but it does allow the reaction efficiency to be
easily determined. Obviously, if the aim of the reaction is to generate a sub-
strate for injection in animals, then radioactive formaldehyde should not be
used. To prevent deleterious effects of formaldehyde its concentration
is kept low and borohydride is used in excess. Any protein substrate
may be methylated in this approach, though we typically methylate histone
proteins in vitro.
1. Dissolve 1 mg of purified histone H3 (Roche) in 500 ml of 200 mM
borate buffer (pH 9.0) and place on ice.
2. To the chilled histone solution add 300 g of NaBH4 and continue
to keep on ice.
3. 1 ml of 14C-formaldehyde (0.83 MBq; 56 mCi/mmol, NEN) is then
added at 5-min intervals on ice, six times in total. Subsequent to the
last addition the reaction is allowed to proceed on ice for another
15 min.
4. Dialyze the reaction products against two changes of 2 L of PBS
over a 12-h period in a Mini Dialysis Unit (Pierce, 10,000 MWCO).
The relative amount of radiolabel (i.e., CH3 groups) incorporated can
be determined by SDS-PAGE and autoradiography. Alternatively, if his-
tones have been used as the substrate, a filter binding assay may be used
(see later).
18
G. E. Means and R. E. Feeney, Biochemistry 7, 2192 (1968).
Inoculating Rabbits
Peptides containing methylated residues are usually coupled to
KLH carrier protein using standard techniques. However, if the immuno-
gen is a chemically methylated protein, then it is used directly. A typical
immunization schedule is outlined later. Before initiating the immuniza-
tion, pre-immune bleeds are taken from the rabbit to identify existing
immune responses. The first immunization (week 0) injection contains
200 g of immunogen per rabbit emulsified in complete Freund’s adjuvant
and this is then followed by two further immunizations (weeks 3 and 6) of
100 g of immunogen per rabbit in incomplete Freund’s adjuvant. Our in-
jection procedure typically involves four subcutaneous site injections per
rabbit per immunization. Two bleeds at weeks 7 and 9 are then taken,
followed by a further immunization in incomplete Freund’s adjuvant.
Two further bleeds are taken at weeks 11 and 13 and finally exsanguination
of the rabbit occurs at week 14. The blood is centrifuged to isolate the sera.
The level and specificity of immune response is then determined using an
ELISA plate assay.
ELISA Plate Assay Protocol

This assay is a relatively quick and easy way to screen bleeds for a
positive immune response.
Individual Covalink microtitre plates (Nunc) are coated with 100 l of
1 g/ml peptide solution (in PBS) per well, covered and incubated over-

night at 4 . The plates are then washed three times with 100 l PBS,
0.1% Tween 20 (PBS-T) per well ensuring rigorous expulsion of liquids
out of the wells at the end of each wash. The wells are subsequently filled
with 100 l blocking buffer (PBS-T with 5% (w/v) BSA) and incubated for
1 h at 37 before being emptied.
A 1/500 dilution of the antibody serum is made in blocking buffer. Two
hundred microliters of this dilution are placed in the first well of the plate
and 100 l are withdrawn and placed into the second well containing 100 l
PBS-T. This serial dilution is performed up to eight times to give a final
serum concentration of 1:64,000 in well 8. The plates are then covered

and incubated for 1 h at 37 before being washed three times with PBS-T
again ensuring vigorous expulsion of the liquids from the wells at the end
of each wash. One hundred microliters of goat anti-rabbit IgG alkaline
phosphatase linked antibody conjugate (1:1000 dilution in blocking buffer)
are then added to each well and the plates are covered and incubated for an

additional 1 h at 37 .
During the secondary antibody incubation development buffer is

prepared by the addition of 0.051 g MgCl2 to 53.625 ml diethanolamine.
Purified water is then added to a final volume of 450 ml. The pH is adjusted
to 9.8 with dilute HCl and finally the buffer is topped up to 500 ml with
water. This buffer may then be stored at 4 for 2–3 days. To generate sub-
strate, one tablet of p-nitrophenyl phosphate (pNPP: Sigma) is added to
5 ml of development buffer and mixed by gentle rotation.
After the secondary antibody incubation the plates are washed three
times with PBS-T and then twice in PBS. Fifty microliters of substrate
are added per well and incubated at room temperature until an appreciable
color gradient appears. The reaction is stopped by the addition of 50 l of
1 M NaOH per well. The optical density (at 405 nm) of the reaction in each
well is then determined by placing the ELISA plate into an optical
plate reader. Selected bleeds with the highest ELISA titre values are then
affinity purified.
Affinity Purification of Antibody

Peptides bound to a solid phase can be used to affinity purify anti-
bodies from total serum. In addition, they can be used to test other proteins
for their ability to specifically bind the methylated peptides. When
performing binding assays of either sort it is essential to make a control
peptide that has the same amino acid sequence as the experimental (meth-
ylated) peptide but which is itself not methylated. Antibodies or other pro-
teins binding only to the experimental peptide are the desired end
products.
We design and synthesize our peptides to contain a C-terminal Cys resi-
due such that we can use this to link them to SulfoLink coupling gel (Pierce,
cat. no. 20401) that binds specifically to free sulfhydryls. All references to
‘‘gel’’ refer to a 50% (v/v) slurry of SulfoLink bead in the appropriate
buffer.
1. The peptide is dissolved in binding buffer (50 mM Tris, 5 mM
EDTA, pH 8.5).
2. Equilibrate the SulfoLink gel in binding buffer by washing 1 ml of
gel in 50 ml buffer.
3. 1 mg of peptide from step 1 (in a volume no greater than 1 ml) is
added to 1 ml of SulfoLink gel. This suspension is slowly rotated for
1 h at room temperature.
4. The SulfoLink gel is then washed three times in 50 ml of binding
buffer.
5. Non-specific binding sites on the gel are blocked by adding 1 ml of

50 mM cysteine (in binding buffer) to 1 ml of gel and then slowly
rotating the suspension for 1 h at room temperature.
6. The gel is washed once with 50 ml 1 M NaCl and once with 50 ml
binding buffer. The peptide bound gel is then stored in binding
buffer containing 0.05% sodium azide.
The next step is to apply serum (identified in ELISA as having high
titre values) to specific immobilized peptides in order to bind out specific
antibodies. We find the following protocol usually gives good consistent
results and it may be applied to a batch purification method or to a column
procedure.
1. Spin serum at 12,000g for 10 min at room temperature and recover
the cleared serum into a fresh tube. Add EDTA to a final
concentration of 5 mM.
2. Equilibrate the peptide-bound SulfoLink gel in TBS/5 mM EDTA
(50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 5 mM EDTA).
3. Incubate 500 l of equilibrated peptide-bound SulfoLink gel with
5 ml of serum/5 mM EDTA by rotating slowly for 90 min at room
temperature.
4. The bound IgG is then washed by subjecting it to 1 M NaCl and
then eluted using 0.1 M glycine solution, pH 2.75 (2 ml glycine
solution per 0.5 ml SulfoLink). The eluted IgG (in 2 ml of
glycine solution) is immediately added to a tube containing 100 l
ice-cold 1 M Tris, pH 8.8. This elution step can be repeated one
to two times to ensure complete elution of the IgG from the peptide.
Ensure that the final pH of the neutralized antibody solution is
about pH 7.0.
5. The affinity-purified antibody is then dialyzed against three changes
of 1 l PBS, and its final concentration is determined using a BCA
protein assay employing an IgG protein standard curve.
We prefer to add BSA to the purified antibodies to a final concen-
tration of 1% (w/v) instead of adding glycerol. We have found that this
often gives lower background when used to probe western blots. For

long-term storage the antibody solution is aliquotted and stored at 80 .
For short-term storage sodium azide is added to 0.04% (w/v) final con-

centration and then stored at 4 . Another ELISA assay is then per-
formed with the antibody to ensure its functionality and to determine the
post-purification titre. It can then be further tested by western blot, etc.
Characterization of Rabbit Polyclonal Anti-Methylated

Histone Antibodies
Is My Antibody Specific?
Once an antibody has been purified, it is imperative to fully character-
ize its specificity. There are two common approaches that can be taken
to achieve this: (i) an ELISA assay or (ii) a western blot approach can
be performed in the presence or absence of varying concentrations of
different peptides. The first decision here concerns the source of substrate.
In an ELISA assay the wells are usually coated with the specific peptide
or protein that the antibody was raised against. In a western blot, the
blotted protein is typically purified histones from calf thymus (Sigma)
since these are a good source of many different methylated sites within
the histone N-terminal tails. In both the ELISA and western blot
approaches, competition peptides should be included to test the specificity
of the antibody. These competition peptides should include as wide a
range as possible of similar, but not identical, epitopes. For instance, if
the antibody to be characterized was raised against a H3 peptide (amino
acids 1–8) tri-methylated at lysine 4, the competition peptides should
include a non-methylated H3 peptide (1–8) as well as the mono-, di-, and
tri-methylated derivatives. Only the tri-methylated peptide should
block the binding of the antibody to its substrate. An example of a typical
result from a western blot approach is shown in Fig. 3. In addition to
these peptides, a number of other tri-methylated histone peptides should
also be tested to determine the absolute specificity of the antibody.
Fig. 3. Determining antibody specificity using peptide competition. Claf thymus histones
were loaded into a single wide well of a 20% SDS-PAG and resolved. They were then western
blotted to nitrocellulose. Strips were then cut from the blot and separately probed with an
anti-tri-methyl K4 of H3 antibody (available from www.abcam.com) in the presence (1 g/ml)
or absence of various peptides as indicated.
Histones that have been expressed in and purified from bacteria also
serve as excellent controls in assays designed to test the specificity of
anti-methylated histone antibodies. Because bacteria lack the methyltrans-
ferases that modify histones, the histones are unmethylated and therefore
should not be recognized by the antibodies.
Characterization by ELISA Assay

Our typical ELISA protocol was described earlier. To characterize the
specificity of a particular antibody using this approach we include peptide
competitors over a wide concentration range.
Characterization by Western Blotting

Histones are resolved in a standard 20% (w/v) SDS-PAG and then
blotted to nitrocellulose. We find that histones blot more efficiently in
a simple phosphate buffer (20 mM Na phosphate; prepare a stock
solution of 1 M Na phosphate, pH 6.7—this is a 50 stock made by mixing
approximately 1000 ml 1 M NaH2PO4 with 850 ml 1 M Na2HPO4,
the exact quantities being dependent on pH). Using a Biorad miniblot
apparatus blot for 35–40 min at 0.8 A. The nitrocellulose membrane
is then recovered from the apparatus and transferred to block solution
(1 TBS, 5% [w/v] BSA, 0.5% [v/v] Tween 20) and incubated for 1 h
at room temperature. While the blocking is taking place, the anti-
methylated histone antibody, at an appropriate concentration, is added
to block solution in a series of tubes. Competitor peptides are then
be added to the tubes in order to test the specificity of the antibody.
Typically peptides are used at a final concentration of 1 g/ml and
include a variety of methylated, as well as non-methylated, peptides
encompassing the amino acid backbone of the inoculating peptide (see
Fig. 3). Peptides and antibody mixes are incubated at room temperature
for 1 h.
The blot is probed in blocking solution containing antibody plus/minus

peptide competitor at 4 overnight. Following this incubation the blot is
briefly washed in blocking solution and then probed at room temperature
with an appropriate HRP-conjugated secondary antibody in blocking solu-
tion. After washing, bound antibody is detected using a chemoluminescent
kit, such as the ECL kit available from Amersham.
Using Anti-Methylated Histone Antibodies
Western Blots
An obvious application for the use of anti-methylated histone
antibodies is the characterization of substrates by western blot analysis.
The substrate protein may include (i) histones extracted from various
cellular sources or (ii) the methylated histone product of an in vitro
methylation assay (Fig. 4). Our protocol for this procedure is described
earlier.
Immunofluorescence
An investigator often wishes to ask whether a particular histone
methylation mark is enriched in specific structures within the genome, in
specific cells, or at specific points of the cell-cycle. One way to address
these issues is by using immunofluorescence. The main decisions here
regard which fixation and permeabilization techniques to use. The fol-
lowing is our general protocol that should serve as a reasonable starting
point.
1. Cells are grown under the appropriate condition in 25-cm2 slide
flasks (NUNC).
2. Aspirate growth medium and wash the cells twice with PBS.
3. The cells are then fixed in 2 ml of 4% (v/v) formaldehyde in PBS
for 20 min at room temperature. Alternatively, paraformaldeyde
may be used if formaldehyde fails to give reasonable results.
4. Wash cells three times (5 min each) with PBS.
5. Block/permeabilize the cells with 3% (w/v) BSA, 0.6% (v/v) Triton
X-100 in PBS for 15 min at room temperature. An alternative
detergent to consider is 0.2% (v/v) Tween 20, though we find that
this often fails to fully permeabilize the cells.
6. Wash three times (5 min each) in PBS.
7. Incubate with secondary antibody (conjugated to an appropriate
flour) in 3% (w/v) BSA in PBS for 1 h in the dark.
8. Wash once in PBS.
9. Hoechst 33258 in PBS is added to a final concentration of 1 g/ml
and incubated for 5 min at room temperature in the dark.
10. The cells are finally washed four times in PBS, each for 5 min.
11. The flask is then opened using the tool supplied by the
manufacturer and the resulting slide is mounted using standard
techniques.
Fig. 4. Immunoflourescence of mammalian cells using anti-methylated K4 of histone

H3 antibodies. (A) Indian muntjac fibroblast cell stained with anti-monomethylated K4 of
H3 antibody (green) and DAPI (red). (B) Indian muntjac fibroblast cell stained with
anti-dimethylated K4 of H3 antibody (green) and DAPI (red). (C) Indian muntjac fibroblast
cells stained with anti-trimethylated K4 of H3 antibody (green).
The slides are then visualized under a microscope, typically a confocal

microscope, capable of irradiating the mounted cells with the appropriate
wavelength light in order to stimulate the flour attached to the secondary
antibody.
Chromatin Immunoprecipitation Analysis

One of the most important questions concerning chromatin research is
where in the genome do particular covalent modifications of histones
occur. In order to answer this type of question a chromatin immunopreci-
pitation (ChIP) may be performed. Chromatin is prepared from cells of
interest that have been subjected to ‘‘gentle’’ formaldehyde cross-linking.
If the researcher is interested in a chromatin-bound factor, such as a
transcription factor, then the cross-linking is necessary to link the factor to

the chromatin. If the investigator wishes to analyze covalent modification
of a histone tail, then the cross-linking is optional, though we find it is not
deleterious to the procedure and therefore recommend that it is performed.
Once the chromatin has been prepared it is necessary to reduce it to
smaller fragments. This essentially serves two purposes: (i) the immuno-
precipitation becomes more reproducible and (ii) the ‘‘resolution’’ of
the technique is dramatically increased. Purified antibodies are then
added to the fractionated chromatin and an immunoprecipitation is
performed. Typically antibodies include those that recognize modified
histones, transcription factors, or other chromatin-associated factors.
Following the immunoprecipitation, the abundance of specific sequences
within the bound fraction is quantitatively determined.
The following ChIP protocols have been used successfully by our
laboratory to study the role of histone methylation in gene transcrip-
tion.7,15,16,19
Harvesting the Cells

Formaldehyde is added dropwise directly to the cell growth media for
a final concentration of 0.75%. These are then gently swirled for 10 min
at room temperature. The cells are then rinsed twice with cold PBS
and scraped into 5 ml of cold PBS. The cells are then pelleted by spinning
at 1000g for 2 min at room temperature. It is now important to directly
proceed to nuclease digestion/sonication. Do not freeze the cells at this
point.
Micrococcal Nuclease Digestion or Sonication?

There are two commonly used procedures for breaking high molecular
weight chromatin into smaller fragments. One uses an endonuclease, typic-
ally micrococcal nuclease, to specifically digest the linker DNA between
nucleosomes. Using this approach it is possible to derive a preparation of
predominantly mononucleosomes; for this approach begin the protocol at
step 1. An alternative approach is to use mechanical shearing by sonication
to break down the chromatin. This results in random cleavage of the DNA
and conditions should be found that give rise to an approximately 700 bp
average length of sonicated chromatin—for sonication begin the protocol
at step 9.
19
S. J. Nielsen, R. Schneider, U. M. Bauer, A. J. Bannister, A. Morrison, D. O’Carroll,
R. Firestein, M. Cleary, T. Jenuwein, R. E. Herrera, and T. Kouzarides, Nature (London)
412, 561 (2001).
1. Remove the supernatant from the pelleted cells and resuspend in

10 ml cold NI buffer (15 mM Tris-HCl, pH 7.5, 60 mM KCl,
0.5 mM DTT, 15 mM NaCl, 300 mM sucrose) containing 5 mM
MgCl2. Then add 10 ml cold NI buffer containing 5 mM MgCl2 and
1% (v/v) NP-40, and mix by inverting and leave on ice for 10 min.

2. Pellet the nuclei at 2000g for 5 min at 4 .
3. Remove the supernatant and resuspend nuclei in 5 ml cold NI
buffer containing 5 mM MgCl2 and then pellet as in step 2.
buffer containing 1 mM CaCl2, and then pellet nuclei as in step 2.
buffer containing 1 mM CaCl2 and transfer to a fresh 1.5 ml
Eppendorf tube. Remove 50 l—this serves as an undigested
control.
6. Add the nuclei from step 5 to 400 l lysis solution (1% [w/v] SDS,
50 mM Tris-HCl, pH 8.0, 20 mM EDTA) and then add 375 units of
S7 micrococcal nuclease (15 U/l in H2O) to each sample. The
chromatin is incubated with the nuclease on ice and the kinetics of
digestion are followed by analyzing the products in a 1.5% agarose
gel. Ideally, the digested chromatin should give rise to mono-
nucleosomes with a low amount of di-nucleosomes being evident in
the agarose gel.
7. The reaction is stopped by the addition of EDTA to a final
concentration of 50 mM.
8. The reaction tubes are then given a brief spin (30 s at 13,000 rpm).
9. The pellet is resuspended in 0.4–1 ml (depending on amount of
cells) of ChIP lysis buffer (1% [w/v] SDS, 10 mM EDTA, 50 mM
Tris-HCl, pH 8.0, protease inhibitor cocktail [Roche]). If sonication
is not required, proceed to step 10. Begin with this resuspension if
the sonication of chromatin is to be performed. Sonicate the
chromatin on ice and follow the digestion in 1.5% agarose gel. An
average sonicated size of about 700 bp should be aimed for.

10. Pellet the cellular debris by spinning at 13,000g for 2 min at 4 .
Then transfer the supernatant to a new tube. Remove 50 l of each
preparation and add it to 400 l lysis solution (1% SDS [w/v],
50 mM Tris-HCl, pH 8.0, 20 mM EDTA). This sample is the input.
It is used for obtaining DNA concentration for subsequent IPs (see
later) and as control in final PCRs.

11. Store the remainder of the chromatin on ice at 4 (it is stable for
approximately 1 week).
Checking the DNA Concentration of Inputs

Five microliters of proteinase K (20 mg/ml stock solution) are added
to each sample in lysis solution (from step 10, above) and it is then heated

with shaking at 65 for 4–5 h to reverse the cross-linking (the samples may
be frozen here if required). The samples are then phenol:chloroform
extracted with an equal volume of 1:1 phenol:choroform. Ten microliters
of glycogen (5 mg/ml stock solution) are added to aid precipitation and
then 1/10th volume of sodium acetate and 2 volumes of ethanol are added.
The precipitations are incubated on dry ice for 10 min and the DNA is then
pelleted by spinning at 13,000g for 10 min. The pellet is carefully washed in
70% (v/v) ethanol and then dried. Finally the DNA pellet is dissolved in
50 l TE (10 mM Tris-HCl, pH 8.0, 1 mM EDTA) and its concentration
is determined by absorbance at OD260.
Performing Chromatin Immunoprecipitations

The protocol outlined later provides a good starting point for research-
ers wishing to begin ChIPs, but it should be emphasized that a number of
steps will need to be empirically optimized for each system analyzed. Each
amplicon to be studied requires a minimum of two independent immuno-
precipitations to be performed: (1) with a specific antibody (e.g., against
di-methylated K4 of histone H3) and (2) with an unrelated irrelevant anti-
body as a control. Each ChIP reaction should contain approximately 25 g
of chromatin-bound DNA as determined earlier.
1. Samples are adjusted to the same volume with ChIP lysis buffer and
each is then diluted 1:10 with dilution buffer (20 mM Tris-HCl, pH
8.0, 150 mM NaCl, 2 mM EDTA, 1% [v/v] Triton X-100). If a high
background is a problem, it is possible to use RIPA buffer here to
help reduce it.
2. Add an appropriate amount of immunoprecipitation antibody.
The actual amount will depend on the specific antibody characteris-
tics, as well as factors (e.g., epitope abundance, accessibility, etc).
Incubate on ice for 1 h.
3. Add 20 l (50% [v/v] slurry) of protein A/G beads (pre-absorbed
with sonicated single-stranded salmon sperm DNA at 1.5 g/20 l

beads). Incubate overnight with rotation at 4 .
4. Pellet the immuno complex by spinning at 1000g for 30 s. Wash the
pellet three times with 1 ml of wash buffer (20 mM Tris-HCl, pH
8.0, 150 mM NaCl, 2 mM EDTA, 1% [v/v] Triton X-100, 0.1% [w/v]
SDS) each time spinning as earlier.
5. Wash the pellet once with 1 ml final wash buffer (20 mM Tris-HCl,
pH 8.0, 500 mM NaCl, 2 mM EDTA, 1% [v/v] Triton X-100, 0.1%
[w/v] SDS). Spin as earlier.
6. Elute the bound DNA by adding 450 l of freshly made elution
buffer (1% [w/v] SDS, 100 mM NaHCO3) and rotating on a wheel
for 15 min at room temperature. Spin down as earlier and remove
the supernatant into a fresh tube.

7. Add 5 l proteinase K (20 mg/ml) and heat with shaking at 65
for 4–5 h to reverse cross-linking (can freeze samples here if
necessary).
8. The samples are then phenol:chloroform extracted with an equal
volume of 1:1 phenol:choroform. Ten microliters of glycogen (5 mg/
ml stock solution) are added to aid precipitation and then 1/10th
volume of sodium acetate and 2 volumes of ethanol are added. The
precipitations are incubated on dry ice for 10 min and the DNA is
then pelleted by spinning at 13,000g for 10 min. The pellet is
carefully washed in 70% (v/v) ethanol and then dried. Finally the
DNA pellet is dissolved in 100 l TE (10 mM Tris-HCl) pH 8.0,

1 mM EDTA, and stored at 4 or 20 .
Analysis of Results by Quantitative PCR

The best way to analyze the results from a ChIP assay is via real-time
quantitative PCR. There are many commercially available instruments
designed to perform these reactions. In each case primers against
the DNA sequence of interest need to be designed and there are many
computer programs to help with this. At the end of the day the actual
methodology for this needs to be determined on a case-by-case basis.
Detecting Histone N-Methyltransferase Activity In Vitro
In Vitro N-Methyltransferase Assays

An investigator may wish to question whether a particular protein con-
tains N-methyltransferase activity. A common approach to address this is
to express the potential methyltransferase (assuming it has already been
cloned) in bacteria and then to purify it. Typically this can be accomplished
by expressing the protein of interest as a GST- or MBP-fusion protein.
Sometimes, however, the potential methyltransferase activity may be
present in a purified cellular extract fraction or in the immuno-pellet of
an immunoprecipitation reaction. In the former case, then the extract is
directly added to the N-methyltransferase assay. In the latter case, the
immuno-pellet is washed into N-methyltransferase buffer and the other re-

action components are then added to the tube.
A protein may be used as a substrate in this assay, though we routinely
employ core histones from calf thymus (a mixture of H2A, H2B, H3, and
H4, Roche) that contain pre-existing modifications found in vivo.
Substrates expressed in and purified from bacteria should also be con-
sidered since they are most likely devoid of pre-existing methylations.
This is particularly important if the reaction products are to be detected
by antibodies where pre-existing modifications may mask the activity
of the enzyme in question. Also it should be noted that a particular
N-methyltransferase may only methylate histones when they are presented
in nucleosomes. The amount of substrate required varies depending on
numerous factors but in a typical assay is usually between 1 and 20 g.
The following protocol has been modified from Rea et al. (2000).20
1. The potential methyltransferase is made to 26 l in N-methyltrans-
ferase buffer (50 mM Tris, pH 8.5, 20 mM KCl, 10 mM MgCl2,
2 mM -mercaptoethanol, 250 mM sucrose).
2. Two microliters of substrate are then added (e.g., 20 g of core
histones, Roche) and 2 l of S-adenosyl-l-[methyl-3H]-methionine
(7.4 102 MBq; 77 Ci/mmol, NEN) as methyl donor. Radioactive
SAM allows the reaction to be easily monitored. However,
non-radioactive SAM may be used if required.

3. The reaction is incubated at 30 for 1 h.
4. Stop the reactions by adding SDS-PAG loading buffer and boiling.
The proteins are resolved by SDS-PAGE and blotted onto PVDF
membrane. We expose the membranes to BiomaxMS (Kodak) films

using a BioMax LE intensifying screen (Kodak) at 80 for 12 h. If the
N-methylation is believed to reside in a lysine near the N-terminus of
the protein, then the radiolabeled protein band can be excised from the
PVDF membrane and sequenced by sequential Edman degradation.
The amino acid residues liberated after each round of degradation are col-
lected and counted in a liquid scintillation counter. The presence of radio-
activity within a fraction indicates that the corresponding amino acid was
methylated.
In our experience, the earlier protocol works well with a range of
K-methyltransferases. However, if R-methyltransferases are to be ana-
lyzed, we have found that it is better to perform the assay as earlier except
that PBS is used in place of N-methyltransferase buffer. We also find it
20
S. Rea, F. Eisenhaber, D. O’Carroll, B. D. Strahl, Z. W. Sun, M. Schmid, S. Opravil,
K. Mechtler, C. P. Ponting, C. D. Allis, and T. Jenuwein, Nature (London) 406, 593 (2000).
better to express the arginine methyltransferases PRMT1, and CARM1 as

GST fusions rather than MBP fusions. It is important to test these enzymes
freshly eluted from the bead and dialyzed against PBS, 10% (v/v) glycerol.
To our knowledge there are no methyltransferase inhibitors that target
specific methyltransferases. Inhibitors such as 50 -methyl-thioadenosine
inhibit most, if not all, methyltransferases.
Detection of N-Methyltransferase Activity Using a Liquid Assay

If histones are used as substrate, they can be radioactively methylated
and bound to filters for counting in a liquid scintillation counter. The main
advantage of this technique over that of SDS-PAGE/western blotting
coupled to autoradiography is that it is significantly easier and quicker to
perform. However, the disadvantage is that this method does not allow
the identification of which histone is methylated.
1. Spot 20 l of a methylation reaction onto a 2-cm2 piece of P81 paper
(Whatman).
2. Wash in 50 mM carbonate buffer, pH 9.0 (3 10 min, 100 ml each
wash) at room temperature.
3. Soak the filters in acetone for 1 min and then air-dry.
4. Scintillation counting is then performed.
Acknowledgments
We would like to thank Darren Harper, Abcam Ltd. (www.abcam.com) for suggestions
and helpful advice relating to antibody production and ELISA assays, Michael Hendzel
(Alberta University, Canada) for supplying immunofluorescence images, and Graham
Bloomberg for advice on peptide synthesis. All our methylated peptides are synthesized by
Graham Bloomberg ([email protected]).
[19] acetylation microarray 289
[19] Analysis of Genome-Wide Histone Acetylation

State and Enzyme Binding Using DNA Microarrays
By Daniel Robyr, Siavash K. Kurdistani, and Michael Grunstein
Nuclear events such as gene transcription, DNA replication, and DNA

repair must cope with local chromatin environments that may be inhibitory
to these processes. Therefore, chromatin remodeling and histone post-
translational modifications are an intrinsic component of the maintenance
and establishment of repressive or permissive chromatin states. Histone
acetylation and deacetylation play an important role in this regulation.
The small genome of Saccharomyces cerevisiae contains numerous histone
deacetylases (HDACs) including Rpd3, Hda1, Hos1, Hos2, Hos3, and Sir2.
Chromatin immunoprecipitation using highly specific antibodies raised
against individual lysine residues1 has established that several of these
enzymes have different histone specificities. For instance, Hda1 deacety-
lates histones H3 and H2B,2 whereas Rpd3 strongly affects all core histones
sites analyzed with the notable exception of H4 lysine 16.1 In contrast,
Hos2 is required for H3 and H4 deacetylation only.3 To function at DNA
regulatory elements HDAC complexes are recruited to their target
promoters by specific transcriptional repressors such as Ume6 at the
Rpd3-affected INO1 gene leading to the local deacetylation of about two
nucleosomes around the TATA-box.4,5 However, Rpd3 and Hda1 also
deacetylate large regions of chromatin, including promoters and open
reading frames (ORFs), without apparent direct recruitment by DNA
binding repressors.6 This global deacetylation mechanism is involved in
the rapid establishment of a global repressive chromatin environment once
gene transcription is turned down. Finally, Hos2 primarily deacetylates his-
tones within the coding regions of genes. Surprisingly, unlike other
HDACs that are repressors, Hos2 is required directly for gene activity.3
These observations suggest that a comprehensive understanding of HDAC
functions requires the analysis of large regions of chromatin, ideally at a
genome-wide level.
1
N. Suka, Y. Suka, A. A. Carmen, J. Wu, and M. Grunstein, Mol. Cell 8, 473 (2001).
2
J. Wu, N. Suka, M. Carlson, and M. Grunstein, Mol. Cell 7, 117 (2001).
3
A. Wang, S. K. Kurdistani, and M. Grunstein, Science 298, 1412 (2002).
4
D. Kadosh and K. Struhl, Cell 89, 365 (1997).
5
S. E. Rundlett, A. A. Carmen, N. Suka, B. M. Turner, and M. Grunstein, Nature 392, 831
(998).
6
M. Vogelauer, J. Wu, N. Suka, and M. Grunstein, Nature 408, 495 (2000).

HDAC locus specificity was initially probed by analyzing the influence

of each HDAC deletion on genome-wide transcription using DNA micro-
arrays.7,8 RPD3 disruption, however, led to more genes being repressed
than up-regulated genome-wide, most likely as a result of indirect effects
on global gene regulation. Thus, determination of HDAC function
genome-wide benefits from the additional studies of acetylation (acetyla-
tion arrays) and enzyme binding (binding arrays) as more direct tools
to assess HDAC locus specificity. In this manner, a much more comprehen-
sive analysis can determine the sites at which the enzyme binds, affects
acetylation state, and results directly in changes in gene activity.9,10 We
describe here methods for identifying the sites of enzyme action using acet-
ylation microarrays and enzyme binding. We have recently used such pro-
cedures to unravel the sites of action for Rpd3 and Hos23,9,10 and
acetylation arrays for studies on Hda1, Hos1, Hos3, and Sir2.10
Acetylation microarrays use the combination of chromatin immuno-
precipitation11 and hybridization of DNA to microarray glass slides (see
Fig. 1).12,13 Highly specific antibodies are used to immunoprecipitate for-
maldehyde-crosslinked chromatin fragments enriched for a given acety-
lated lysine residue in cell lysates obtained from a wild-type (WT) strain
and its isogenic strain disrupted for the HDAC of interest. The crosslink
is reversed after chromatin immunoprecipitation allowing DNA purifica-
tion, amplification by PCR, and DNA labeling with fluorophores. One
fluorescent dye (e.g., Cy5) is used for DNA recovered from the strain
carrying the deacetylase gene mutant, whereas a second dye (e.g., Cy3) is
used to label DNA from the WT isogenic control strain. The labeled
DNA probes from both strains are combined and hybridized onto DNA
microarray glass slides containing either intergenic regions or ORFs or
both. Glass slides are scanned for both fluorescent dyes and the normalized
7
B. E. Bernstein, J. K. Tong, and S. L. Schreiber, Proc. Natl. Acad. Sci. USA 97, 13708
(2000).
8
T. R. Hughes, M. J. Marton, A. R. Jones, C. J. Roberts, R. Stoughton, C. D. Armour, H. A.
Bennett, E. Coffey, H. Dai, Y. D. He, M. J. Kidd, A. M. King, M. R. Meyer, D. Slade, P. Y.
Lum, S. B. Stepaniants, D. D. Shoemaker, D. Gachotte, K. Chakraburtty, J. Simon, M. Bard,
and S. H. Friend, Cell 102, 109 (2000).
9
S. K. Kurdistani, D. Robyr, S. Tavazoie, and M. Grunstein, Nat. Genet. 31, 248 (2002).
10
D. Robyr, Y. Suka, I. Xenarios, S. K. Kurdistani, A. Wang, N. Suka, and M. Grunstein, Cell
109, 437 (2002).
11
A. Hecht, S. Strahl-Bolsinger, and M. Grunstein, Nature 383, 92 (1996).
12
V. R. Iyer, C. E. Horak, C. S. Scafe, D. Botstein, M. Snyder, and P. O. Brown, Nature 409,
533 (2001).
13
B. Ren, F. Robert, J. J. Wyrick, O. Aparicio, E. G. Jennings, I. Simon, J. Zeitlinger,
J. Schreiber, N. Hannett, E. Kanin, T. L. Volkert, C. J. Wilson, S. P. Bell, and R. A. Young,
Science 290, 2306 (2000).
Fig. 1. Acetylation microarrays. Chromatin fragments from crosslinked mutant cells

(rpd3) and their isogenic WT counterparts were immunoprecipitated using highly specific
antibodies raised against acetylated histone sites. DNA from enriched chromatin fragments
was purified, amplified by PCR, and labeled with a fluorophore (Cy3 or Cy5). Probes from
both sets of labeled DNA were then combined and hybridized to a DNA microarray
containing either intergenic regions, ORFs, or both. For a given region on the microarray, the
ratio of the normalized fluorescent intensities between the two probes indicates whether
the analyzed lysine residue is hypo- or hyper-acetylated in the experiment strain. Reprinted
with permission from D. Robyr and M. Grunstein, Methods 31, 83–89 (2003).
ratios of intensities between the deacetylase mutant and the WT probes

reflect changes in acetylation in the mutant strain.
Chromatin Immunoprecipitation (ChrIP or ChIP)

1. Dilute an overnight pre-culture of yeast into 50 ml YEPD medium
(2% peptone, 1% yeast extract, 2% dextrose) to A600 ¼ 0.2 and
allow the cells to reach A600 1. Histones are crosslinked to DNA
in vivo by adding formaldehyde (Fisher) to the culture to a final
concentration of 1% (w/v). The crosslinking reaction is carried out
at room temperature for 15 min with constant mild agitation and is
then quenched by adding 2.5 ml 2.5 M glycine (final concentration
125 mM).
2. Harvest and wash cells twice in ice-cold 1 PBS (140 mM NaCl,
2.5 mM KCl, 8.1 mM Na2HPO4, 1.5 mM KH2PO4, pH 7.5) by

centrifugation at 4 at 2800g (Beckman J2-HC, rotor JS-4.3). The

cell pellet can be frozen in liquid nitrogen and stored at 80 if
needed.
3. Resuspend cells in 400 l ice-cold lysis buffer (50 mM HEPES/
KOH, pH 7.5, 500 mM NaCl, 1 mM EDTA, 1% (v/v) Triton X-100,
0.1 % (v/v) SDS, and 0.1% (w/v) sodium-deoxycholate) comple-
mented with proteases inhibitors (Complete, Roche; 100 stock
prepared in 500 l H2O from 1 tablet). Transfer the resuspended
cells into siliconized tubes (1.7 ml NoStick Hydrophobic Micro-
tubes, GeneMate) and add 1 volume of acid-washed glass beads
(0.45–0.52 mm diameter, Thomas Scientific).14 Cells are lysed on an

Eppendorf shaker (model 5432) between 45 and 60 min at 4 .
Puncture the bottom of the opened tubes containing the lysed cells
with a red hot 25G1 needle (Becton Dickinson). Immediately close
the tubes, place them on a fresh collecting 1.7-ml tube and recover
the cell lysate by centrifugation (5 s in an Eppendorf tabletop
centrifuge (model 5415 C) at 14,000 rpm).
4. Shear chromatin from the cell lysate by sonication down to an
average fragment size of 500 bp using a Sonic dismembrator 550
(Fisher Scientific) with a Microtip model 419 from Misonix. Sonicate
lysate on ice with two pulses of 15 s each (magnitude setting of 4)
and a 60-s rest interval. Recover the clear cell lysate from the cell

debris (pellet) by centrifugation at 4 in an Eppendorf centrifuge

(10 min at full speed). At this point the lysate can be stored at 80
for up to a month.
14
A. Hecht, S. Strahl-Bolsinger, and M. Grunstein, Methods Mol. Biol. 119, 469 (1999).

5. Immunoprecipitate acetylated chromatin fragments overnight at 4
on a nutator incubating 50 l of cell lysate with 2–5 l of a given
antiserum (see later). Then, add a suspension of 25 l of 50% (v/v)
protein A sepharose CL-4B beads (Amersham-Pharmacia),
equilibrated in lysis buffer and incubate an additional 2 h.
6. Pellet the sepharose beads for 30 s at room temperature in an
Eppendorf centrifuge (model 5415 C) at 735g (3000 rpm). Discard
the supernatant and wash successively for 5 min the beads on a
nutator with 500 l of the following solutions: twice in lysis buffer,
once in deoxycholate buffer (10 mM Tris-HCl, pH 8.0, 0.25 M LiCl,
0.5% (v/v) NP-40, 0.5% (w/v) sodium deoxycholate and 1 mM
EDTA, pH 8.0), and once in TE, pH 8.0. Pellet the beads and
discard the supernatant between each washing step as indicated
earlier. The whole procedure is carried out at room temperature.
7. Elute the immunoprecipitated chromatin fragment from the beads

and reverse crosslink overnight at 65 with 50 l elution buffer
(50 mM Tris-HCl, pH 8.0, 10 mM EDTA, pH 8.0, and 1% (v/v) SDS).
8. Add 50 l of TE, pH 8.0, 20 g glycogen and treat with 20 g

proteinase K for 2–3 h at 55 . Finally, extract DNA with 1 volume
phenol/chloroform/isoamyl alcohol [25:24:1 (v/v)] and ethanol
precipitate. Resuspend purified DNA into 50 l TE, pH 8.0 and

store at 20 .
Troubleshooting
The antibody specificity is the most critical aspect for any chromatin im-
munoprecipitation. The polyclonal antibodies raised against individual
acetylated lysine residues were prepared in our laboratory and are avail-
able at Upstate (http://www.upstate.com). Their specificity was verified
by ELISA assay and tested by ChrIP against histones mutated for the acet-
ylatable lysine.1 The titration of the antibody amount required for an im-
munoprecipitation has to be determined experimentally for all antibodies.
Sonication conditions will determine the resolution of the chromatin
immunoprecipitation (the higher the average fragment size is, the lower
the resolution will be). Shearing efficiency will also depend on the sonicator
brand. A pilot experiment should be performed in order to find the appro-
priate sonication settings. After sonication add 80 l elution buffer (see
step 7 for recipe) to a 20-l aliquot from the cleared cell lysate (step 4)

and incubate the tube overnight at 65 . Add 100 l TE, pH 8.0, 40 g pro-

teinase K and incubate the tubes for 2–3 h at 55 as indicated earlier (step
8). Resuspend DNA pellet (Input DNA) in 20 l TE, pH 8.0, after DNA
extraction and precipitation. Treat an aliquot (10 l) with RNase A

(10 g) 30 min at 37 and analyze the average DNA fragment size on a
1.5% agarose gel. Alter accordingly the sonication settings if DNA frag-

ments are too large. Store the remaining RNase untreated DNA at 20
for later PCR amplification if needed (see later).
Double Crosslinking with Protein-Protein Crosslinking Agents

and Formaldehyde
For certain proteins, such as the HDAC Rpd3, formaldehyde cross-
linking alone is inadequate for efficient crosslinking of the enzyme to
chromatin in vivo.9 This may be due to the fact that Rpd3 is part of a
large (1 MDa) multiprotein complex and, unlike histones, may lie too
far from DNA for efficient crosslinking by formaldehyde alone. The
immunoprecipitation efficiency of Rpd3 is significantly improved when in
addition to formaldehyde, a protein-protein crosslinking agent is also
used.9 In such double crosslinking scheme, the cells are sequentially treated
first with a protein-protein crosslinking agent and then with formal-
dehyde. We have successfully used several protein-protein crosslinking
agents such as dimethyl adipimidate (DMA), dimethyl pimelimidate
(DMP), or 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) prior
to formaldehyde treatment to ChrIP the Rpd3 deacetylase. This new cross-
linking method has been described recently in more detail for various
protein-protein crosslinkers15 and is summarized as follows:
1. Grow cells as indicated earlier but do not add formaldehyde
immediately when they have reached A600 ¼ 1.
2. Pellet and wash the cells twice with 20 ml ice-cold 1 PBS in a 50-ml
Falcon tube. Prepare a fresh 15 ml solution of 10 mM DMA
(Pierce) in ice-cold 1 PBS containing 0.25% (v/v) dimethyl
sulfoxide (DMSO). Since DMA reacts with amine groups, buffers
such as Tris that contain primary amines should be avoided. Non–
amine-containing buffers such as phosphate, carbonate, and HEPES
are acceptable. Resuspend the cells in the DMA solution and incubate
on a nutator at room temperature for 45 min. Pellet and wash the cells
once again in 1 PBS and resuspend in 1 PBS containing 1% (w/v)
formaldehyde. Incubate the cells on a nutator at room temperature
for 10–11 h. This incubation time is necessary for maximal
crosslinking of Rpd3 but needs to be empirically determined for
the protein under study. Stop the reaction by adding 2.5 ml 2.5 M
glycine and wash the cells again twice in ice-cold 1 PBS.
15
S. K. Kurdistani and M. Grunstein, Methods 31, 90 (2003).
3. Cell lysate is prepared as described earlier with the exception of a

salt concentration of 150 mM NaCl in the lysis buffer (150-mM lysis
buffer). DNA enrichment during chromatin immunoprecipitation
may be further improved by partially purifying chromatin prior to
sonication. To do this, pellet chromatin by centrifuging the crude

extract for 3 min at 4 at 14,000 rpm. Discard the supernatant and
carefully resuspend the chromatin pellet in 400 l 150-mM lysis
buffer.
4. Due to the longer formaldehyde crosslinking time, increase
sonication time to two pulses of 20 s each (magnitude setting of 4)
with a 60-s rest interval for more efficient shearing of chromatin.
5. ChrIP is then carried out essentially as described earlier with the
following modifications. One hundred microliters of the lysate (WT
and rpd3) was incubated with 5-l of -Rpd3 polyclonal antibody

(Upstate) and incubated on a nutator at 4 for 2 h. After 1 h
incubation with 50-l a 50% v/v suspension of protein A beads, the
beads are transferred to a 0.45-m filter unit (Millipore Ultrafree-
MC; cat#UFC30HV00) for the washing steps (2 in lysis buffer
containing 500 mM NaCl and 2 in deoxycholate buffer, 15 min
each). After each wash, the filter unit was spun at 3000 rpm for 30 s,
and the flow through fraction was discarded. To elute chromatin off
the beads, 100 l elution buffer was added to the filter unit,

incubated at 65 for 15 min, and spun at 3000 rpm for 30 s and
repeated once more. The eluates were combined and incubated at

65 overnight. DNA is then recovered as described.
Probe Amplification by PCR

Low DNA yield after chromatin immunoprecipitation is not adequate
for immediate labeling and hybridization onto DNA microarrays, thus
requiring a DNA amplification step by PCR. This approach is adapted
from Bohlander et al.16 as described at http://www.microarrays.org/
protocols.html. A similar method was recently covered by Horak and
Snyder.17 Since DNA sequences in the chromatin immunoprecipitation
are heterogeneous, the first step in this approach will consist of the
random incorporation of degenerated oligonucleotides (nanomer) at-
tached to a linker sequence. The latter sequence will then be used for the
probe amplification by PCR.
16
S. K. Bohlander, R. Espinosa, III, M. M. Le Beau, J. D. Rowley, and M. O. Diaz, Genomics
13, 1322 (1992).
17
C. E. Horak and M. Snyder, Methods Enzymol. 350, 469 (2002).
1. Use small 0.2-ml PCR tubes and perform all reactions in a PCR
thermocycler. Add 2 l 5 Sequenase buffer (200 mM Tris-HCl, pH
7.5, 100 mM MgCl2, and 250 mM NaCl) and 1 l (40 pmol)
oligonucleotide (50 -GTTTCCCAGTCACGATCNNNNNNNNN-
30 ) to 7 l immunoprecipitated DNA. Denature DNA for 2 min at

94 , cool the tubes down to 8 , and incubate at 8 for an additional
2 min. Pause the PCR machine and add 5 l of reaction mix (1
Sequenase buffer, 0.9 mM dNTPs, 15 mM DTT, 0.75 g BSA, and
4 U Sequenase 2.0). The oligonucleotide is annealed to DNA by

slowly raising the temperature from 8 to 37 over a period of 8 min.

DNA synthesis is allowed to proceed for an extra 8 min at 37 .
Repeat once the whole denaturation-annealing-elongation process.
However, since the DNA polymerase is heat sensitive, add 4 U

Sequenase after the denaturation step at 94 . Finally, stop the
reaction by diluting the reaction with 35 l TE, pH 8.0. DNA (step 1

DNA) can be stored at 20 or used immediately for the PCR
reaction described later.
2. The following step of the DNA amplification procedure is a regular
PCR reaction using the fixed linker oligonucleotide sequence (50 -
GTTTCCCAGTCACGATC-30 ) and DNA from step 1 as a
template. Transfer an aliquot of 15 l step 1 DNA into a fresh
0.5-ml PCR tube and carry the reaction in a final volume of 100 l
(20 mM Tris-HCl, pH 8.4, 50 mM KCl, 2 mM MgCl2, 1.25
nanomoles oligonucleotide, 0.25 mM dNTPs, and 5 U recombinant

Taq polymerase [Invitrogen]). Denature DNA at 92 for 30 s,

anneal with two consecutive short 30-s steps at 40 and 50 ,

respectively, and elongate at 72 during 90 s. Repeat this cycle 24
times and allow the final elongation to proceed for another 10 min

at 72 .
3. Purify the PCR reaction through columns (QIAquick PCR
purification kit, Qiagen) as indicated by the manufacturer and elute
DNA with 50 l 10 mM Tris-HCl, pH 8.5.
4. Estimate DNA yield under UV light by spotting serial dilutions of
DNA in a Petri dish (0.5 l DNA mixed with 0.5 l EtBr (10 g/ml)
along with a standard of known concentration (2-fold serial dilution
from 100 ng/l to 6.25 ng/l) and test the average size of DNA (5 l
aliquot) after PCR on a 1.5% (w/v) agarose gel. The probe average
size is reduced from 500 bp (after sonication) down to 300 bp. This
is due to the random incorporation of the oligonucleotide into DNA
during step 1.
Troubleshooting
We routinely obtain between 3 and 5 g DNA after amplification by
PCR. It is not absolutely necessary to have such a DNA yield since the la-
beling reaction described later requires 500 ng DNA. However, if the
amount of DNA is not satisfactory, try to start the amplification procedure
with more material or scale-up the chromatin immunoprecipitation. We do
not advise increasing the number of PCR cycles since it is kept relatively
low in order to ensure amplification linearity.
The final PCR purification through columns (QIAquick PCR purifica-
tion kit, Qiagen) is critical since unincorporated oligonucleotides and
dNTPs may interfere with the subsequent labeling reaction efficiency.
For some microarray applications, it may be necessary to compare im-
munoprecipitated DNA with input DNA (i.e., study of a wild-type strain
acetylation profile instead of analyzing the changes of acetylation between
a WT and a HDAC mutant). The Input DNA is prepared from 20 l cleared
lysate as indicated earlier (see the Troubleshooting section for Chromatin
Immunoprecipitation (ChrIP or ChIP)). Dilute 150-fold an aliquot of the
Input DNA prior to starting the amplification procedure (step 1).
Klenow Labeling of the Probe and Hybridization

We favor labeling of the probe by Klenow random priming over a PCR-
based method described elsewhere.17 We have noticed that the fluorescent
labels are not efficiently incorporated into DNA using a Taq DNA poly-
merase. Moreover, our first PCR amplification step described earlier yields
enough DNA for direct labeling by Klenow random priming, thus bypass-
ing the need for a second PCR reaction. It is advised to switch the fluores-
cent dye between the two probes when repeating the experiment to correct
for the incorporation efficiency of each dye into DNA.
1. The labeling reaction makes use of reagents (random octamer and
concentrated Klenow) from the Bioprime DNA labeling system
(Gibco-BRL) as described at http://www.microarrays.org/proto-
cols.html. Amplified DNA (500 ng) is mixed with 20 l 2.5
random primer solution (125 mM Tris-HCl, pH 6.8, 12.5 mM
MgCl2, 25 mM 2-mercaptoethanol, and 750 g/ml random octa-
mers). Complement the solution with water to a final reaction
volume of 41 l. Boil the samples for 5 min and transfer the tubes on
ice immediately. Add 5 l dNTPs mix (1.2 mM dATP, 1.2 mM
dTTP, 1.2 mM dGTP, and 0.6 mM dCTP in TE, pH 8.0), 3 l of
either Cyanine 3-dCTP or Cyanine 5-dCTP (Renaissance, NEN),
and 1 l of concentrated Klenow fragment (40 U/l). Incubate the

reaction in the dark for at least 2 h at 37 .
2. The probes are purified and concentrated through a microcon-30
filter (Amicon) as follows. Combine the Cy3- and Cy5-labeled
probes and add 450 l TE, pH 8.0, along with tRNA (10 g) and
salmon sperm DNA (10 g). The addition of nucleic acid competi-
tors (tRNA and salmon sperm DNA) at this stage presents the
advantage of increasing the maximum probe volume (up to 4.6 l)
that can be subsequently mixed with the hybridization solution.
Place the column on top of a collecting tube and centrifuge at room
temperature for 7 min (Eppendorf tabletop centrifuge (model 5415
C) at 12,000 rpm). The labeled DNA (violet-purple color) should be
clearly visible at the bottom of the column. Discard the flow-
through, add another 450 l TE, pH 8.0, in the column, and
centrifuge again for 7 min. Check the remaining probe volume in the
column and centrifuge by 1-min increments until it is concentrated
down to 4.6 l or less. Invert the column and recover the combined
probes in a collecting tube by centrifugation (1 min).
3. Prepare the hybridization mix as follows: mix the concentrated
probe with 5 l 20 SSC (3 M NaCl, 0.3 M trisodium citrate, pH
7.0), 10 l 100% formamide, and 0.4 l 10% (v/v) SDS. Comple-
ment the solution with water up to a final volume of 20 l if
necessary (final concentration: 5 SSC, 50% formamide, 0.2%
SDS, 0.5 mg/ml tRNA, and 0.5 mg/ml salmon sperm DNA).

Denature the hybridization mix for 5 min at 95 . The probe can
be cooled on ice briefly (few seconds) but should not remain there
in order to prevent SDS precipitation. Pipet the hybridization
mix on the glass cover slip and cover with the microarray slide.
This is favored to the alternative (i.e., pipet the probe on the slide
and cover with the cover slip), as it reduces the formation of
bubbles. The slide is placed in a hybridization chamber (Corning)
which in turn is arranged in a humid chamber consisting of a closed
plastic box (Tupperware) containing paper soaked in water and

a little stand. Hybridization is carried out overnight at 44 in a
hybridization oven.
4. The cover slip is removed after a brief immersion in 400 ml 2 SSC
at room temperature. The slides are then washed at room
temperature for 5 min in 0.1 SSC/0.1% (v/v) SDS and twice in
0.1 SSC. All washing steps are done in staining jars with 400 ml
solution and under mild shaking. Slides are dried by centrifugation
in microtiter plate carrier for 5 min at room temperature at 500 rpm
(57g in Beckman J2-HC, rotor JS-4.3). Drainage of liquid from the
slides is helped by placing a piece of folded paper towel between the

slide holder and the microtiter plate carrier. Microarrays are
scanned for both fluorescent dyes using a GMS 418 Array Scanner
(Genetic Micro Systems).
Troubleshooting
In our hands, the labeling procedure is reliable and successful most of
the time. The color of the purified probe is the best indicator for the pro-
cedure success. The combination of the two probes should give a dark
purple color after concentration. Do not discard the flow-through if the
probe is colorless after the first 7 min centrifugation step through the
microcon-30 filter, as it is possible that the filter is defective. Simply reload
the flow-through onto another filter and centrifuge again. If the problem
persists, this is a clear indication that the labeling was not efficient and
the probe should not be used for hybridization. Similarly, a pink (Cy3-
labeling) or a blue (Cy5-labeling) color indicates that one of the probes
was not labeled properly. Change the vial of dye if you suspect the lack
of efficient labeling might stem from a bad batch. The dNTP mix is also
an important factor for labeling. We have consistently observed a reduction
in its efficiency when using the same dNTP mix stock several times. We
strongly advise to prepare it fresh shortly before the labeling reaction.
We have observed that pre-hybridization of the microarray slide is not
required when used for acetylation microarrays. Some antibodies used for
immunoprecipitation of epitope-tagged protein (i.e., anti-hemagglutinin)
appear to create unacceptable hybridization background levels. The
following pre-hybridization conditions can be applied in order to alleviate
or reduce the background problem. Incubate the slides in 3.5 SSC, 0.1%

(v/v) SDS, and 10 mg/ml BSA in a staining jar for 20 min at 44 . Rinse
the slides briefly first with water then with isopropanol. Dry the slides by
centrifugation on a microtiter plate carrier as indicated earlier (step 4).
Data Quantitation, Normalization, and Analysis

Several commercial and freely available software packages can be used
for data quantitation. Fluorescence intensities are quantified in our labora-
tory using Imagene software (version 4.1) from BioDiscovery (http://
www.biodiscovery.com). The following site contains a non-exhaustive list
of various data quantitation and analysis tools: http://genome-www5.
stanford.edu/MicroArray/SMD/restech.html. Refer to their respective
manuals to learn how to use them.
Normalization is probably one of the greatest challenges in data analy-

sis. Indeed, the dye emission intensities are not equal and probes are not
labeled similarly due mostly to the different DNA incorporation efficiency
of Cy3 and Cy5. A broadly used normalization method makes the assump-
tion that the sums of all intensities of both dyes are equal across the entire
array. In other words, any acetylation enrichment should be compensated
by a similar decrease in acetylation somewhere else in the genome. If this
were true, one could use the ratio of the total intensities between both
probes as a normalization factor. While this approach is used for expression
arrays, it is not suitable for acetylation arrays since deletion of an HDAC
will lead to an increase of general histone acetylation that will not be com-
pensated by a global decrease. Therefore, we strongly suggest the use of in-
ternal controls for normalization of acetylation microarrays. Yeast
telomeres are hypoacetylated and are not affected by most of the deacety-
lases (Rpd3, Hda1, Hos1, Hos2, and Hos3), with the notable exception of
the telomeric Sir2, which deacetylates histone H4 lysine 16.10,18 Telomeric
sequences from chromosome 6R (coordinates 269505 to 270095, also
labeled Tel6R in the intergenic array) are appropriate for a crude
pre-normalization of the data. If possible, this DNA fragment should be
spotted many times throughout the glass DNA microarray in order to be
reliably used for normalization. Once the data are pre-normalized, identify
other regions in the array (between 6 and 10 such regions) whose ratio of
intensity between the two probes is very close to Tel6R (ratio of about 1).
Confirm that the acetylation state of these newly identified regions is not
affected in your experiment using a standard PCR analysis of the original
ChrIP.14 Use these regions to finalize normalization of the entire data set.
Different antibodies have inherently different affinities for their own
epitope leading to variations in immunoprecipitation efficiency. For
instance, a highly specific antibody raised against histone H3 lysine 18 may
not be as efficient as another antibody recognizing histone H3 lysine 9.
This is particularly important when comparing data obtained for different
acetylation sites. Data from different experiments can be scaled by calcu-
lating the percentile ranking (using Microsoft Excel) of their respective
enrichment ratio. This allowed us to compare acetylation with expression
data (see Fig. 2).10
Basic data analysis does not require fancy and expensive software but
can readily be performed using Microsoft Access, which is part of the
Microsoft Office bundle. Access allows the creation and handling of data-
bases. We have, for instance, created our own database containing gene
annotation, published microarray data sets from other laboratories, and
18
N. Suka, K. Luo, and M. Grunstein, Nat. Genet. 32, 378 (2002).
Fig. 2. Histone acetylation sites correlate differently with transcription resulting from
rpd3. This figure illustrates the scaling of different acetylation data sets using percentile
ranking in order to compare them with a transcription data set. Due to inherent noise in
microarray data, absolute correlation analysis between acetylation and transcription is not
very informative. A moving average analysis that greatly reduces noise can be used to extract
general trends. The moving average (window size, 100 data point; step, 1 data point)
percentile rank of acetylation enrichment is plotted as a function of transcription increase
resulting from rpd3 (B. E. Bernstein, J. K. Tong, and S. L. Schreiber, Proc. Natl. Acad. Sci.
USA 97, 13708 (2000)). Acetylation data are plotted for H4 K5 (dark blue), H4 K12 (red), H4
K16 (orange), and H3 K18 (light blue). Control corresponds to a comparison of two sets of
probes amplified from the immunoprecipitation of acetylated H4 K12 in the WT strain and
labeled separately with Cy3 and Cy5 prior to hybridization. Data show that increased
acetylation at histone H4 K5 and K12 is associated most directly with increased gene
transcription in rpd3. H4 K16 show the poorest correlation with gene activity (RPD3
disruption has no significant effects on the status of H4 K16 acetylation) (D. Robyr, Y. Suka,
I. Xenarios, S. K. Kurdistani, A. Wang, N. Suka, and M. Grunstein, Cell 109, 437 (2002)).
Reprinted with permission from D. Robyr et al., Cell 109, 437–446 (2002).
gene functional categories as defined by MIPS (Munich Information

Center for Protein Structure at: http://mips.gsf.de/proj/yeast/CYGD/db/
pathway_index.html). Further data analysis (clustering, sequence analysis,
etc.) will need software such as GeneXPress that is freely available for
academics (http://genexpress.stanford.edu/).
Most of the data analysis software, however, was designed for ORF
arrays and not for intergenic arrays. Promoter-containing intergenic frag-
ments must first be assigned to their respective ORF, especially when com-
paring acetylation status at a promoter with the transcription level of the
ORF it regulates. Some intergenic regions are located between two diver-
gent ORFs and are by default assigned to both ORFs. Other intergenic
regions are located more than 1 kb away from any ORF. These are
considered orphans and are not assigned to any particular ORF.
DNA Intergenic Microarray Preparation

Yeast intergenic microarrays include all sequences located between
ORFs, including telomeric regions, rDNA, tRNA, centromeres, and trans-
posable elements. A full description of the method to prepare microarray
slides is not the scope of this review and was thoroughly discussed earlier.19
Rather, we will briefly comment on the method we are currently using in
our laboratory. We have amplified by PCR about 6700 intergenic regions
from yeast genomic DNA using primer pairs available at ResGen (http://
www.resgen.com/). The PCR amplification is performed in 96-well plates
and is described at http://www.microarrays.org/protocols.html. The size
of every single PCR product was checked by agarose gel electrophoresis.
It is possible to coat glass slides in the laboratory with poly-l-lysine
(http://cmgm.stanford.edu/pbrown/protocols/1_slides.html). However, we
strongly recommend the use of amino-silane–coated slides from Corning
(CMT-GAPS–coated slides) since they lead to better printing and reduced
background during hybridization. PCR fragments (100 ng/l in 3 SSC)
were transferred onto 384-well plates for printing using 16 pins from
Array-It (TeleChem International) and an arrayer built according
to specifications provided at: http://cmgm.stanford.edu/pbrown/mguide/
index.html. Printed slides are processed using succinic anhydride blocking
according to a method provided by the slide manufacturer. Such a block-
ing procedure will dramatically reduce hybridization background. The
preparation of ORF microarrays is identical. We suggest the printing of
ORF and intergenic regions on the same slide if required for your applica-
tion. Information concerning microarray arrayers and slide manufacturers
is hosted at: http://www.biologie.ens.fr/en/genetiqu/puces/links.html.
Concluding Remarks
Acetylation microarrays have proved to be extremely useful in dis-
covering new functions for the yeast HDACs, functions that would have
been much more difficult to identify using classical genetic or molecular
biology techniques.10 Although the approach described here focuses on
acetylation, it can be extended to other histone modifications as well. The
19
M. B. Eisen and P. O. Brown, Methods Enzymol. 303, 179 (1999).
Fig. 3. Binding arrays are complementary to acetylation arrays. The moving average
(window size, 100 data point; step, 1 data point) of Rpd3 enrichment (binding) over intergenic
regions is plotted as a function of percentile rank of H3 K18 acetylation in rpd3. Data sets
with (+RP) and without (RP) ribosomal protein genes are plotted as indicated. Rpd3 binds
strongly at the promoter of ribosomal protein genes in logarithmically growing cells where
these genes are highly active but under the same conditions, has little or no effect on
acetylation of these promoters (S. K. Kurdistani, D. Robyr, S. Tavazoie, and M. Grunstein,
Nat. Genet. 31, 248 (2002); D. Robyr, Y. Suka, I. Xenarios, S. K. Kurdistani, A. Wang, N. Suka,
and M. Grunstein, Cell 109, 437 (2002)) or the expression of the RP genes (B. E. Bernstein,
J. K. Tong, and S. L. Schreiber, Proc. Natl. Acad. Sci. USA 97, 13708 (2000)). Thus, the
ribosomal protein gene promoters as targets of Rpd3 are only detectable by binding arrays.
This clearly illustrates the importance of combining different types of arrays (binding,
acetylation, and expression) to fully comprehend HDAC function. Reprinted with permission
from S. K. Kurdistani et al., Nat. Gen. 31, 248–254 (2002).
key to a successful analysis depends on the availability of highly specific

antibodies and an appropriate crosslinker. We have used a similar
approach to study genome-wide binding of Rpd39 using a double-crosslink-
ing approach with a protein-protein crosslinking agent (DMA) and formal-
dehyde. The latter genome-wide binding study clearly showed that Rpd3
binds to many chromatin loci where it has no detectable effect on acetyla-
tion when deleted in logarithmically growing cells. These regions would
be hidden in acetylation microarrays under the same condition (see
Fig. 3), emphasizing the need for a combination of different types of arrays
(acetylation, binding, and expression) for full understanding of Rpd3’s
function(s).
While acetylation microarrays are extremely powerful, they have some

limitations. First, the relative low resolution of the microarrays (up to 1 kb
large fragment spotted on the microarray) may not reveal regions where
only one nucleosome is modified if nearby nucleosomes are not affected.
High resolution is obtained by analyzing a region of interest by standard
ChrIP14 or by printing custom microarrays containing 100–200 bp DNA
fragments along the regions of interest. Second, histone post-translational
modifications create interacting surface for histone binding proteins
(e.g., bromodomains of Gcn5 and Snf220) that may potentially interfere
with antibody recognition for the same site. Finally, aneuploidy between
two strains21 has to be corrected in order to ascertain that increase binding,
acetylation, or expression is not the result of gene duplication in the ana-
lyzed strain versus the control. A simple hybridization of Input genomic
DNA (obtained from the sonicated cell extract but not incubated with
the antibody) from the experiment and the control strains can indicate
potential aneuploidy problems.
The study of other histone modifications using similar approach will
in the future provide important insights not only on their genome-wide
respective patterns, but also on how they may relate and influence each
other and correlate with transcription activity, unraveling the roles of
histone modifications in chromosome functions.
20
L. Zeng and M. M. Zhou, FEBS Lett. 513, 124 (2002).
21
T. R. Hughes, C. J. Roberts, H. Dai, A. R. Jones, M. R. Meyer, D. Slade, J. Buchard,
S. Dow, T. R. Ward, M. J. Kidd, S. H. Friend, and M. J. Marton, Nat. Genet. 25, 333 (2000).
[20] Use of Chromatin Immunoprecipitation

Assays in Genome-Wide Location Analysis
of Mammalian Transcription Factors
By Bing Ren and Brian D. Dynlacht
One major challenge in the study of mammalian transcriptional regula-

tion pertains to difficulties in studying direct targets of regulatory proteins
in a physiological setting. This chapter describes methods for the use of
chromatin immunoprecipitation (ChIP) in the genome-wide identification
of binding sites in mammalian cells. This approach combines the adapta-
tion of ChIP in the enrichment of mammalian transcription factor target

While acetylation microarrays are extremely powerful, they have some

limitations. First, the relative low resolution of the microarrays (up to 1 kb
large fragment spotted on the microarray) may not reveal regions where
only one nucleosome is modified if nearby nucleosomes are not affected.
High resolution is obtained by analyzing a region of interest by standard
ChrIP14 or by printing custom microarrays containing 100–200 bp DNA
fragments along the regions of interest. Second, histone post-translational
modifications create interacting surface for histone binding proteins
(e.g., bromodomains of Gcn5 and Snf220) that may potentially interfere
with antibody recognition for the same site. Finally, aneuploidy between
two strains21 has to be corrected in order to ascertain that increase binding,
acetylation, or expression is not the result of gene duplication in the ana-
lyzed strain versus the control. A simple hybridization of Input genomic
DNA (obtained from the sonicated cell extract but not incubated with
the antibody) from the experiment and the control strains can indicate
potential aneuploidy problems.
The study of other histone modifications using similar approach will
in the future provide important insights not only on their genome-wide
respective patterns, but also on how they may relate and influence each
other and correlate with transcription activity, unraveling the roles of
histone modifications in chromosome functions.
20
L. Zeng and M. M. Zhou, FEBS Lett. 513, 124 (2002).
21
T. R. Hughes, C. J. Roberts, H. Dai, A. R. Jones, M. R. Meyer, D. Slade, J. Buchard,
S. Dow, T. R. Ward, M. J. Kidd, S. H. Friend, and M. J. Marton, Nat. Genet. 25, 333 (2000).
[20] Use of Chromatin Immunoprecipitation

Assays in Genome-Wide Location Analysis
of Mammalian Transcription Factors
By Bing Ren and Brian D. Dynlacht
One major challenge in the study of mammalian transcriptional regula-

tion pertains to difficulties in studying direct targets of regulatory proteins
in a physiological setting. This chapter describes methods for the use of
chromatin immunoprecipitation (ChIP) in the genome-wide identification
of binding sites in mammalian cells. This approach combines the adapta-
tion of ChIP in the enrichment of mammalian transcription factor target

[20] use of ChIP in location analysis 305
Fig. 1. A schematic diagram of genome-wide location analysis with mammalian

transcription factors. The method can in principle be generalized for any DNA-binding
regulatory protein.
genes with DNA microarray technology. This method will enable research-
ers to overcome limitations in the study of eukaryotic transcriptional regu-
lation by allowing the identification of direct, physiological targets of
DNA-binding proteins in an unbiased, genome-wide manner.
Procedures
Overview
Genome-wide location analysis, described in detail in this chapter, is
based on the rapid purification of specific genomic fragments associated
with a particular protein followed by detection of enriched sequences with
DNA microarray technology (see Fig. 1). Intact, living cells are treated
with formaldehyde, a cross-linking agent that results in the covalent and
reversible linkage between genomic DNA and associated proteins.1 The
cross-linked chromatin is extracted after disruption of cell and nuclear
membranes and fragmented into an average length of 500 bp by sonication.
The chromatin associated with a factor of interest is then purified through
immunoprecipitation using a specific antibody against the protein. Subse-
quently, the cross-links are reversed, and the genomic DNA in the resulting
sample is purified by successive steps to remove protein and RNA. DNA
fragments are blunt-ended by T4 DNA polymerase to allow ligation of a
universal oligonucleotide linker, and all DNA fragments in the sample
are amplified through polymerase chain reaction (PCR) in an unbiased
fashion. As a control, an aliquot of the input chromatin DNA is processed
1
using an identical procedure. The ChIP-enriched and control DNA are

then labeled with different fluorescent dyes in a random priming reaction,
and the labeled DNA samples are hybridized simultaneously to a single
DNA microarray containing genomic DNA (promoter) sequences. An
enriched genomic sequence is identified as a DNA spot in which there is
significantly more fluorescence intensity with ChIP-enriched DNA than
with the control input DNA.
This method has several important advantages over other binding site
mapping methods, particularly, computational prediction and genome
expression analysis. First, the method examines direct protein-DNA inter-
actions throughout the genome, and it therefore reveals regulation by a
transcription factor in a native setting under physiological conditions.
Second, because the method does not require prior knowledge of a tran-
scription factor’s binding specificity, or its transcriptional activity, it is an
unbiased approach and can uncover novel biological properties of a tran-
scription factor. Third, in addition to revealing the location of promoter
and enhancer binding proteins, the method can also be applied to examine
any DNA-binding regulatory protein.
To adapt the method to mammalian transcription factors, the protocol
described later has taken several factors into consideration. First, instead
of using arrays that cover the entire genome, the method employs
arrays that represent only gene promoters. This design is based on three
considerations:
1. It is not yet feasible to produce DNA microarrays that cover the
entire human genome in 1-kb segments. The effort to synthesize and
print 3 million DNA fragments is not practical for individual
laboratories.
2. Transcription factors most frequently target gene promoters. A
survey of known transcription factor binding sites in the human
genome showed that the vast majority occur within 1 kb from the
transcription start site (TSS) of genes.2
3. To make a DNA microarray that represents all human gene
promoters is both affordable and manageable by individual
laboratories. The ease of production combined with high utility of
the promoter array outweighs the potential lack of coverage for a
small number of target sites in the genome.
2
B. Ren, H. Cam, Y. Takahashi, T. Volkert, J. Terragni, R. A. Young, and B. D. Dynlacht,
Genes Dev. 16, 245 (2002).
The second factor to consider is the modification of the hybridization

conditions. The human genome contains abundant repetitive DNA
sequences. Certain repetitive DNA elements are represented by millions
of copies and can lead to significant cross-hybridization problems. To
reduce or eliminate such cross-hybridization, human Cot-1 DNA enriched
for repetitive DNA sequences is included in the hybridization reaction.
In addition, a stringent hybridization temperature is used to increase
hybridization specificity.
Solutions
Cross-linking solution: 11% formaldehyde, 0.1 M
NaCl, 1 mM Na-EDTA, 0.5 mM
Na-EGTA, 50 mM HEPES,
pH 8.0
Lysis buffer 1: 0.05 M HEPES-KOH, pH 7.5,
0.14 M NaCl, 1 mM EDTA,
10% glycerol, 0.5% NP-40,
0.25%, Triton X-100, with
protease inhibitor cocktail
(Roche Applied Science)
Lysis buffer 2: 0.2 M NaCl, 1 mM EDTA,
0.5 mM EGTA, 10 mM Tris, pH
8, protease inhibitor cocktail
Lysis buffer 3: 1 mM EDTA, 0.5 mM EGTA,
10 mM Tris-HCl, pH 8, protease
inhibitor cocktail
Proteinase K stock solution: 20 mg/ml proteinase K (Sigma)
50 mM Tris-HCl, pH 8.0, 1.5 mM
calcium acetate
RIPA buffer: 50 mM HEPES, pH 7.6, 1 mM
EDTA, 0.7% DOC, 1% NP-40,
0.5 M LiCl, protease inhibitor
cocktail
Elution buffer: 50 mM Tris, pH 8, 10 mM EDTA,
1% SDS
Hybridization buffer 1: 2.2 SSC, 0.22% SDS
Hybridization buffer 2: 70% formamide, 3 SSC, 14.3%
dextran sulfate
Cross-Linking of Cells and Fragmentation and

Immunoprecipitation of Chromatin
Formaldehyde Cross-Linking and Preparation of Chromatin

109 cells suspended in growth medium are transferred as 40-ml aliquots
to 50 ml-tubes and placed on ice for 10 min. Then 1/10 volume of cross-
linking solution is added directly to each tube. After a 10-min incubation
on ice (or at room temperature, depending upon the cell type), 1/20 volume
of 2.5 M glycine solution is added to each tube to stop the cross-linking
reaction. The cells are harvested by centrifugation at 2000g for 10 min at

4 . The cell pellets are re-suspended in cold PBS (137 mM NaCl, 2.7 mM
KCl, 10 mM Na2HPO4, 2 mM KH2PO4) and washed twice. The final cell
pellet may be snap-frozen in liquid nitrogen. Adherent cells can also
be fixed directly on plates, treated with glycine, washed with PBS, and
harvested in cold PBS with a silicone scraper.
The cell pellet is re-suspended in 30 ml of Lysis Buffer 1 by pipetting
and is mixed for 10 min at 4 on a rocking platform. After centrifugation

at 2000g for 10 min at 4 , the cell pellet is re-suspended in 24 ml Lysis
Buffer 2 by pipetting and mixed gently at room temperature for 10 min

on a rocking platform. After centrifugation at 2000g for 10 min at 4 , the
pellet is re-suspended in 10 ml of Lysis Buffer 3.
The mixture is divided into 5-ml aliquots in 15 ml-conical tubes. These
tubes are then placed into 50-ml tubes containing ice. A sonicator (Bran-
son Sonifier 450, with power setting at 5) is used to disrupt the cell and nu-
clear membranes and fragment the chromatin. To avoid foaming, the
sonicator probe tip is first immersed in the mixture followed by a 25-s,
continuous pulse of sonication. The tube is immediately placed on ice for
at least 1 min to avoid overheating the sample. Sonication is repeated until
the chromatin fragments are of the desired length. The fragment size can
be examined by agarose gel electrophoresis of 10 l of cell extract digested
with Proteinase K for 1 h. The number of sonication cycles varies with cell
type and cross-linking conditions, and pilot tests are recommended. Ten
cycles of sonication are normally sufficient to achieve the desired
(500 bp) fragment size. Finally, the chromatin solution is adjusted to
0.5% Sarkosyl (sodium lauryl sarcosine) and gently mixed for 10 min at
room temperature on a rocking platform. The chromatin solution is then
transferred to a centrifuge tube and spun for 10 min at 10,000g to remove
cell debris. The supernatant is collected for ChIP. At this stage, the concen-
tration of DNA should be around 1–2 mg/ml. The solution can be stored at

80 in 1-ml aliquots. The amount of chromatin generated here should be
sufficient for 10 immunoprecipitation reactions.
Immunoprecipitation of Chromatin
Magnetic beads (Dynal) pre-bound to polyclonal antibodies are used to
immunoprecipitate the DNA associated with the protein of interest. To
prepare the magnetic beads, 100 l of sheep anti-rabbit IgG-conjugated
Dynabeads (Dynal) are first washed three times with cold PBS containing
5 mg/ml bovine serum albumin (BSA) and then re-suspended in 5 ml of
cold PBS. Ten micrograms of rabbit polyclonal antibody are added to the

mixture and incubated overnight on a rotating platform at 4 . After collect-
ing the magnetic beads by centrifugation and washing three times with cold
PBS containing 5 mg/ml BSA, the beads are re-suspended in 100 l of
cold PBS with 5 mg/ml BSA and are then ready for immunoprecipitation.
In an Eppendorf tube, 2 mg of soluble chromatin are first adjusted to
0.1% Triton X-100, 0.1% sodium deoxycholate, and 1 mM PMSF, then
mixed with 100 l of magnetic beads pre-bound to the antibody. The mix-

ture is incubated at 4 overnight on a rotating platform. The magnetic
beads are then collected using a magnet (Dynal), and the supernatant is
removed by aspiration. To remove material non-specifically bound to the
beads, 1 ml of RIPA buffer is added to the tube, and the beads are gently
re-suspended on a rotating platform in the cold room. The magnetic beads
are again collected with the magnet and washed with RIPA buffer a total of
five times. After washing once with 1 ml of TE, the beads are collected by
centrifugation at 2000g for 3 min and re-suspended in 50 l of elution
buffer. To elute precipitated chromatin from the beads, the tubes are incu-

bated at 65 for 10 min with constant agitation then centrifuged for 30 s at
2000g. Forty microliters of supernatant are removed and mixed with 120 l

of TE with 1% SDS. This solution is incubated at 65 overnight to reverse
the cross-links. As a chromatin input control, 100 g of chromatin are
mixed with 120 l of TE containing 1% SDS in a separate tube and

incubated at 65 overnight.
Purification of DNA
After reversal of cross-links, proteins in the DNA sample are removed
by incubation with 120 l of Proteinase K solution (2% glycogen, 5%

Proteinase K stock solution, and TE) for 2 h at 37 . The sample is then ex-
tracted twice with phenol and once with 24:1 chloroform/isoamyl alcohol.
The sample is adjusted to 200 mM NaCl. After ethanol precipitation,
the DNA is dissolved in 30 l of TE containing 10 g of DNase-free

RNase A and incubated for 2 h at 37 . The DNA at this step can be further
purified with a Qiagen PCR kit (Qiagen).
Blunting, Ligation of Linkers to DNA, and Amplification by PCR

The immunoprecipitated DNA is typically present in very small quan-
tities (1–10 ng total), and it is therefore necessary to perform an amplifica-
tion step prior to labeling and DNA microarray analysis. To achieve this,
a ligation-mediated PCR (LM-PCR) procedure is used. Because the
DNA at this stage contains recessed and heterogeneous ends as a result
of the physical shearing process, it is first treated with T4 DNA polymerase
to form blunt ends. Then a linker is ligated to the DNA fragments.
The addition of this linker allows the DNA to be amplified by PCR using
a universal oligonucleotide primer.
Blunting, Ligation, and PCR

In an Eppendorf tube, the immunoprecipitated DNA (or 20 ng of con-
trol input DNA) is combined with 11 l of 10 T4 DNA polymerase buffer
(New England Biolabs), 0.5 l BSA (10 mg/ml) (New England Biolabs),
0.5 l dNTP mix (20 mM each), 0.2 l T4 DNA polymerase (3 U/l)
(New England Biolabs), and distilled H2O in a total volume of 112 l.

After a 20-min incubation at 12 , 1/10 volume of 3 M sodium acetate (pH
5.2) and 10 g of glycogen (Roche Applied Sciences) are added to the
tube, and the DNA sample is extracted with phenol:chloroform:isoamyl
alcohol (25:24:1) (Sigma) once. After ethanol precipitation, the DNA is
dissolved in 25 l of distilled H2O.
The blunt-ended DNA is mixed with 8 l of distilled H2O, 10 l of
5 ligase buffer (Invitrogen), 6.7 l of annealed oligonucleotide linkers
(oligo-1: GCGGTGACCCGGGAGATCTGAATTC, oligo-2: GAATT-
CAGATC, annealed to make a 15 M solution), 0.5 l T4 DNA ligase
(New England Biolabs), and distilled water in a total volume of 50 l.
The ligation reaction is allowed to proceed overnight at 16 . After ligation,
the DNA is purified by ethanol precipitation and dissolved in 25 l of
distilled H2O.
The ligated DNA is then mixed with 4 l of 10 ThermoPol reaction
buffer (New England Biolabs), 4.75 l distilled H2O, 5 l of 10 dNTP
mix (2.5 mM each), 1.25 l oligo-1 (40 M stock) in a final volume of

40 l in a 500-l thin-walled PCR tube. The tube is first incubated at 55
for 2 min in a thermal cycler to separate linker oligonucleotides not ligated
to the DNA, then 10 l of an enzyme mix [8 l dH2O, 1 l Taq DNA poly-
merase (5U/l), 1 l ThermalPol reaction buffer, and 0.025 units of Pfu
polymerase (Stratagene)] is added. Subsequently, the following PCR
protocol is performed:

Step 1: 72 for 5 min




Step 6: go to step 3, 22 times

Step 8: 4 indefinitely
After PCR, the DNA is purified using the Qiaquick PCR purification
kit (Qiagen) and eluted in 60 l elution buffer provided with the kit.
Labeling of DNA and Microarray Hybridization
Labeling of Amplified DNA with Fluorophores

Two hundred nanograms of DNA from the previous step is mixed with
20 l of 2.5 random primer solution (BioPrime kit, Invitrogen) and dis-
tilled H2O in a final volume of 42.5 l. The mixture is boiled for 5 min
and then cooled on ice for 5 min. Subsequently, 5 l of 10 low dCTP mix-
ture (2.5 mM each for dATP, dTTP and dGTP, and 0.6 mM for dCTP),
1.5 l of Cy5-dCTP (Amersham) or Cy3-dCTP (Amersham), and 40 units

of Klenow DNA polymerase are added to tube. The tube is incubated at 37
for 2 h. After the reaction, the labeled DNA is purified using the Qiagen
PCR kit (Qiagen).
Fabrication and Hybridization of Promoter DNA Microarrays

An array (1.5K array) containing approximately 1500 human promoter
fragments was initially designed based on NCBI human genome assembly
Build 24 (July 3, 2001 freeze). One thousand two hundred genes were
selected from the collection of RefSeq entries based on three criteria:
1. Their expression varied during the mammalian cell cycle
2. The Transcription Start Site (TSS) for each of these genes was
annotated based on full-length mRNA sequences and
3. Their biological function was well documented
In addition, approximately 200 genes that do not exhibit cell cycle peri-
odicity were also included. Oligonucleotide primers were designed to
amplify fragments with endpoints at 700 to þ300 relative to the start site
of these genes using PCR. After PCR, the amplified fragments were
purified, verified by gel electrophoresis, and spotted onto GAPSII glass
slides (Corning) using a Cartesian microarrayer (Genomic Solutions).
After UV cross-linking, the glass slides were stored in a vacuum desiccator
until use.
In a new Eppendorf tube, 2.5 g of Cy5-labeled ChIP-enriched DNA

is mixed with 2.5 g of Cy3-labeled genomic (control input) DNA and
36 g of human Cot-1 DNA (Invitrogen). DNA is precipitated with 1/10
volume of 3 M sodium acetate and 2 volumes of ethanol, and the DNA is
dissolved in 22.4 l of hybridization buffer 1. The mixture can be heated

for 10 min at 37 to facilitate re-suspension of the DNA. Twenty microli-
ters of hybridization buffer 2 are added to the mixture, and the tube is

heated at 95 for 5 min to denature DNA, then incubated at 42 for 2 min.
Four microliters of yeast tRNA (10 g/l; Sigma) and 3 l of 2% BSA
are then added to the mixture, which is then spotted onto a DNA microar-
ray slide that has been incubated with pre-hybridization solution for 40 min

at 42 . A 25 mm 60 mm cover slip is then gently placed on top of the
sample, and the hybridization is carried out in a hybridization chamber

(Corning) at 60 overnight in a water bath.
After hybridization, the microarray slide is washed once with wash

buffer 1 (2 SSC, 0.1% SDS) at 60 for 5 min in a glass slide staining dish.
This is followed by a wash with buffer 2 (0.2 SSC, 0.1% SDS) for
10 min at room temperature and three times with buffer 3 (0.2 SSC) at
room temperature. The slide is then dried by a brief spin at 1000g in a table
top centrifuge.
Microarray Analysis and Identification of In Vivo DNA-Binding Sites
Microarray Scanning and Analysis

The microarray slides are scanned using a GenePix 4000B scanner
from Axon Instruments, and each microarray image is first analyzed with
the GenePix Pro 3.0 image analysis software to derive the Cy3 and Cy5
fluorescent intensity and background noise for all the spots on the array.
The intensities for both Cy3 and Cy5 channels are first adjusted by sub-
tracting the background intensity of each spot using the formula Ichannel
Bchannel, where channel represents either Cy5 or Cy3. The intensities
are further adjusted by subtracting the median intensity of all the blank
spots on the array. If any of these values are lower than 10, they are
automatically raised to a value of 10.
A normalization factor is calculated based on the intensities of spots
that are considered good (more than 65% of pixels have intensities higher
than the background intensity plus 1 standard deviation). The median of
the intensity ratios, I635/I532, is then used to adjust the Cy3 channel intensity
to the same level as that of the Cy5 channel.
The quantitative amplification of small amounts of DNA generates
some uncertainty in values for the low-intensity spots. In order to estimate
that uncertainty and average repeated experiments with appropriate

related weights, a single array error model3 is used. The significance of a
measured ratio at a spot is defined by a statistic X, which is formulated as
a2 a1
X ¼ qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
; (1)
2
21 þ 22 þ f 2 a21 þ a22
where a1,2 are the intensities measured in the two channels for each spot,
1,2 are the uncertainties due to background subtraction, and f is a frac-
tional multiplicative error such as would come from hybridization non-
uniformities, fluctuations in the dye incorporation efficiency, and scanner
gain fluctuations. The distribution of X can be found to be close to a Gauss-
ian distribution in experiments where Cy3 and Cy5 samples are identical.
The significance of a change of magnitude X is then calculated using a
one-sided probability model as follows:
x
P ¼ 1 Erf ; (2)

where is the average of X and is the standard deviation of X. Since the
intensities are normalized, should be near 0. The Erf(x) function is
the standard normal accumulative distribution function corresponding to
standard normal curve areas.
If the Cy3 and Cy5 samples are not identical, the Gaussian distribution
can be skewed, because the ChIP can result in many DNA spots with sig-
nificantly higher intensities in one channel than the other. Since the input
DNA is always present, the intensity distribution of those non-enriched
spots can be used to obtain the parameters of X distribution. First, the
spots whose X is less than 0 are identified. These spots should be on the left
half of the Gaussian distribution. Their mirror spots are then generated
using Xþ ¼ X. The mean and standard distribution of this new X can
be calculated.
When the genomic binding sites of a protein are investigated using the
above-mentioned method, it is routine to perform several independent ex-
periments so that the results are more reliable. To combine replicate data
sets, each sample is first analyzed individually using the above-mentioned
single-array error model. The average binding ratio and associated P values
from these multiple experiments are then calculated using a weighted
averaging analysis method.3 For each spot, the uncertainty in the log(ratio)
is defined as
¼ logða2 =a1 Þ=Xnorm ; (3)
3
C. J. Roberts, B. Nelson, M. J. Marton, R. Stoughton, M. R. Meyer, H. A. Bennett, Y. D.
He, H. Dai, W. L. Walker, T. R. Hughes et al., Science 287, 873 (2000).
where a1,2 are the intensities measured in the two channels for each
spot and Xnorm is the normalized X for each spot. The weights for each spot
are then defined as
wi ¼ 1=2i (4)
The averaged log(ratio) is then calculated using the following formula
P
i¼1;n wi xi
x¼ P ; (5)
i¼1;n wi
where n is the total number of experiments, xi and wi are the log(ratio) and
weight, respectively, for a particular spot in each experiment; x is normally
distributed and the averaged P value is then calculated using Eq. (2), where
the variable for Erf( ) function is x instead. Target genes were selected on
the basis of significant P values (e.g., less than 0.002) and binding ratios
(e.g., greater than 2) in replicate experiments.
Conclusion
In summary, the method that we describe in this chapter is a direct,
high-throughput and general approach to locate protein-DNA interactions
in mammalian cells. Since most transcription factors function by binding
to DNA and regulating gene expression, this method will allow for the
rapid, specific identification of physiologically relevant target genes and
provide key insights into the molecular mechanisms through which they
function. We expect that this technique will have widespread applications
in deciphering global gene regulatory networks in mammalian cells as it
has in yeast.4 In addition to revealing the location of promoter and en-
hancer binding proteins, the method can also be applied to examine regu-
latory elements bound by the DNA replication and recombination
machinery and nucleosome modification complexes, and it should also
allow a systematic identification of sites of histone modification and
DNA methylation.5–9
4
T. I. Lee, N. J. Rinaldi, F. Robert, D. T. Odom, Z. Bar-Joseph, G. K. Gerber, N. M.
Hannett, C. T. Harbison, C. M. Thompson, I. Simon, J. Zeitlinger, E. G. Jennings, H. L.
Murray, D. B. Gordon, B. Ren, J. J. Wyrick, J. B. Tagne, T. L. Volkert, E. Fraenkel, D. K.
Gifford, and R. A. Young, Science 298, 799 (2002).
5
B. E. Bernstein, E. L. Humphrey, R. L. Erlich, R. Schneider, P. Bouman, J. S. Liu,
T. Kouzarides, and S. L. Schreiber, Proc. Natl. Acad. Sci. USA 99, 8695 (2002).
6
Y. Blat and N. Kleckner, Cell 98, 249 (1999).
7
J. L. Gerton, J. DeRisi, R. Shroff, M. Lichten, P. O. Brown, and T. D. Petes, Proc. Natl.
Acad. Sci. USA 97, 11383 (2000).
[21] screening of chromatin immunoprecipitates 315
Acknowledgments
We thank H. Cam, Y. Takahashi, T. Volkert, and members of Richard Young’s
laboratory for help with the development of techniques described in this chapter. Work in
BDD’s laboratory was supported by the American Cancer Society.
8
109, 437 (2002).
9
J. J. Wyrick, J. G. Aparicio, T. Chen, J. D. Barnett, E. G. Jennings, R. A. Young, S. P. Bell,
and O. M. Aparicio, Science 294, 2357 (2001).
[21] High-Throughput Screening of Chromatin

Immunoprecipitates Using CpG-Island Microarrays
By Matthew J. Oberley, Julissa Tsao, Patrick Yau, and
Peggy J. Farnham
Sequencing of the human genome has provided a wealth of information

that is now allowing rapid advances in our understanding of gene regula-
tion. For example, 50 exon prediction programs have provided estimates
of the number of promoters in the human genome. Davuluri et al.1 have de-
veloped a program called FirstEF and used it to predict 68,645 first exons.
Such information now allows analysis of thousands of promoter sequences
and provides a comprehensive data set for a bioinformatics approach to
finding target promoters of human transcription factors. Using such an
approach, programs have been developed to identify E2F sites in sets of
mammalian promoters.2,3 Similarly, scanning the human genome for Myc
binding sites has identified a large set of promoters that contain
Myc consensus sites.4 Follow-up analyses of the predicted E2F and Myc
target promoters using chromatin immunoprecipitation (ChIP) assays have
indicated that many of the putative targets are in fact bound by E2F or Myc
family members in living cells.2,4 Therefore, such bioinformatics ap-
proaches are very instructive and have provided investigators with a large
number of potential target genes. However, there are two major concerns
1
R. V. Davuluri, I. Grosse, and M. Q. Zhang, Nat. Genet. 29, 412 (2001).
2
A. E. Kel, O. V. Kel-Margoulis, P. J. Farnham, S. M. Bartley, E. Wingender, and M. Q.
Zhang, J. Mol. Biol. 309, 99 (2001).
3
S. Aerts, G. Thijs, B. Coessens, M. Staes, Y. Moreau, and B. De Moor, Nucleic Acids Res.
31, 1753 (2003).
4
P. C. Fernandez, S. R. Frank, L. Wang, M. Schroeder, S. Liu, J. Greene, A. Cocito, and
B. Amati, Genes Dev. 17, 1115 (2003).

Acknowledgments
We thank H. Cam, Y. Takahashi, T. Volkert, and members of Richard Young’s
laboratory for help with the development of techniques described in this chapter. Work in
BDD’s laboratory was supported by the American Cancer Society.
8
109, 437 (2002).
9
[21] High-Throughput Screening of Chromatin

Immunoprecipitates Using CpG-Island Microarrays
By Matthew J. Oberley, Julissa Tsao, Patrick Yau, and
Peggy J. Farnham
Sequencing of the human genome has provided a wealth of information

that is now allowing rapid advances in our understanding of gene regula-
tion. For example, 50 exon prediction programs have provided estimates
of the number of promoters in the human genome. Davuluri et al.1 have de-
veloped a program called FirstEF and used it to predict 68,645 first exons.
Such information now allows analysis of thousands of promoter sequences
and provides a comprehensive data set for a bioinformatics approach to
finding target promoters of human transcription factors. Using such an
approach, programs have been developed to identify E2F sites in sets of
mammalian promoters.2,3 Similarly, scanning the human genome for Myc
binding sites has identified a large set of promoters that contain
Myc consensus sites.4 Follow-up analyses of the predicted E2F and Myc
target promoters using chromatin immunoprecipitation (ChIP) assays have
indicated that many of the putative targets are in fact bound by E2F or Myc
family members in living cells.2,4 Therefore, such bioinformatics ap-
proaches are very instructive and have provided investigators with a large
number of potential target genes. However, there are two major concerns
1
R. V. Davuluri, I. Grosse, and M. Q. Zhang, Nat. Genet. 29, 412 (2001).
2
A. E. Kel, O. V. Kel-Margoulis, P. J. Farnham, S. M. Bartley, E. Wingender, and M. Q.
Zhang, J. Mol. Biol. 309, 99 (2001).
3
S. Aerts, G. Thijs, B. Coessens, M. Staes, Y. Moreau, and B. De Moor, Nucleic Acids Res.
31, 1753 (2003).
4
P. C. Fernandez, S. R. Frank, L. Wang, M. Schroeder, S. Liu, J. Greene, A. Cocito, and
B. Amati, Genes Dev. 17, 1115 (2003).

that arise when one relies solely on a bioinformatics approach. First,

many consensus sites may reside within inactive chromatin and may not
be available for interaction with a factor due to inaccessibility. For
example, computational identification of binding sites for the PIF/GMEB
factor identified sites that were bound in vitro by the factor but which were
extremely sensitive to the methylation status of the DNA.5 Thus, these
sites would not be occupied if the consensus site resided within hyper-
methylated heterochromatin. Similarly, a binding site may reside within
active (euchromatin) but be inaccessible due to specific nucleosomal pos-
itioning. We have shown that a consensus binding site for E2F that is
located in exon 2 of the Myc gene is not occupied in vivo.6 Also, overlap
of a consensus binding site with the site for a more abundant and/or higher
affinity DNA binding protein can prevent in vivo binding. For example, the
consensus Ets site in the c-Myc promoter is occupied in vivo only if the
overlapping E2F site is mutated.7 Thus, methylation, nucleosomal position-
ing, and overlapping binding sites can lead to the identification of a set of
false-positive target promoters using in silico methods. Conversely, many
true positives may be overlooked if the factor is recruited to the DNA
via a site that does not exactly match a consensus site. Recruitment via a
non-consensus site could be accomplished via protein-protein interaction;
the interaction between YY1 and E2F2 is believed to allow stable recruit-
ment of E2F2 to the cdc6 promoter.8 Such promoters would be overlooked
by identification schema that rely solely on the presence of a high-affinity
consensus site in a promoter region. Also, transcription factors such as ster-
oid hormone receptors and POU domain proteins can recognize a variety
of similar but not identical sites.9,10 The lack of a strict consensus can make
it difficult to identify true target promoters using only sequence inspection
tools. Clearly, experimental methods that can allow global analyses and
identification of binding sites for human transcription factors would
complement the bioinformatics approaches.
Over the last several years, experimental methods have been developed
to identify genomic targets of DNA binding factors. All of these approaches
utilize as a first step the ChIP assay to selectively enrich for chromatin
fragments bound by transcription factors. Investigators then use the
5
E. Burnett, J. Christensen, and P. Tattersall, J. Mol. Biol. 314, 1029 (2001).
6
A. S. Weinmann, S. M. Bartley, T. Zhang, M. Q. Zhang, and P. J. Farnham, Mol. Cell. Biol.
21, 6820 (2001).
7
T. Albert, J. Wells, J. O. Funk, A. Pullner, E. E. Raschke, G. Stelzer, M. Meisterernst, P. J.
Farnham, and D. Eick, J. Biol. Chem. 276, 20482 (2001).
8
S. Schlisio, T. Halperin, M. Vidal, and J. R. Nevins, EMBO J. 21, 5775 (2002).
9
S. Millevoi, L. Thion, G. Joseph, C. Vossen, L. Ghisolfi-Nieto, and M. Erard, Eur. J.
Biochem. 268, 781 (2001).
10
M. A. Loven, V. S. Likhite, I. Choi, and A. M. Nardulli, J. Biol. Chem. 276, 45282 (2001).
chromatin fragments to probe microarrays containing intergenic se-

quences. The list of factors that have been analyzed on a genome-wide
basis using yeast as a model system is quite large.11–17 In contrast,
fewer studies have been performed using the human genome, mostly due
to the lack of availability of appropriate microarrays. Because the human
genome (3.2 Gb) is three orders of magnitude larger than the yeast genome
(12 Mb), a set of chips containing all human intergenic regions has not yet
been created. However, several groups have begun creating their own spe-
cific human intergenic microarrays. The largest effort to date consists of a
set of oligonucleotide arrays that span the smaller chromosomes 21 and 22
at 35 bp intervals.18 This array has not been used to identify sites for DNA
binding factors; rather it has been used to assess the sum transcriptional
activity of these chromosomes. In fact, no oligonucleotide arrays have been
reported to be used in combination with chromatin IP, for either yeast or
mammalian cell systems. It is not yet clear if this is due to technical prob-
lems or just due to a lack of availability of the arrays. A recent study util-
ized a DNA microarray comprising nearly all of the non-repetitive
sequences of human chromosome 22.19 In this case, PCR fragments (rather
than oligonucleotide probes) were spotted onto the array, suggesting that
this array could be useful for ChIP studies. However, the published use
of this new array has been limited to analysis of mRNA expression profiles.
Two different approaches, both using spotted PCR fragments, have suc-
cessfully been used to identify target genes of human transcription factors.
The microarrays used in these studies differ in whether they contain
‘‘investigator-selected’’ promoter regions or ‘‘randomly chosen clones.’’
11
533 (2001).
12
J. D. Lieb, X. Liu, D. Botstein, and P. O. Brown, Nat. Genet. 28, 327 (2001).
13
J. Schreiber, N. Hannett, E. Kanin, T. L. Volkert, C. J. Wilson, S. P. Bell, and R. A. Young,
Science 290, 2306 (2000).
14
H. H. Ng, F. Robert, R. A. Young, and K. Struhl, Genes Dev. 16, 806 (2002).
15
M. Damelin, I. Simon, T. I. Moy, B. Wilson, S. Komili, P. Tempst, F. P. Roth, R. A. Young,
B. R. Cairns, and P. A. Silver, Mol. Cell. 9, 563 (2002).
16
17
T. I. Lee, N. J. Rinaldi, F. Robert, D. T. Odom, Z. Bar-Joseph, G. K. Gerber, N. M.
Hannett, C. T. Harbison, C. M. Thompson, I. Simon, J. Zeitlinger, E. G. Jennings, H. L.
Murray, D. B. Gordon, B. Ren, J. J. Wyrick, J. Tagne, T. L. Volkert, E. Fraenkel, D. K.
Gifford, and R. A. Young, Science 298, 799 (2002).
18
P. Kapranov, S. E. Cawley, J. Drenkow, S. Bekiranov, R. L. Strausberg, S. P. Fodor, and
T. R. Gingeras, Science 296, 916 (2002).
19
J. L. Rinn, G. Euskirchen, P. Bertone, R. Martone, N. M. Luscombe, S. Hartman, P. M.
Harrison, F. K. Nelson, P. Miller, M. Gerstein, S. Weissman, and M. Snyder, Genes Dev. 17,
529 (2003).
A PCR fragment microarray including the sequence spanning from 700

to þ200 of 1444 human genes has been utilized to find novel promoters
bound by E2F family members.20 Although a step in the right direction,
such an array by necessity is biased in that the particular 1444 promoters
analyzed were chosen by a set of criteria. Using this same approach, it
would be feasible to prepare PCR fragments representing the sequences
upstream of all predicted first exons in the human genome. However, this
Herculean project has not been attempted to date. As an intermediate step,
our laboratory has utilized a microarray that contains 7776 CpG island
clones.21 CpG islands tend to be found in intergenic regions and at the 50
ends of genes. Thus, use of CpG arrays allows a relatively unbiased analysis
of many thousands of human intergenic regions.
As described later, we have now developed protocols that allow these
arrays to be utilized to perform a large-scale identification of genomic
regions occupied by DNA-binding proteins under physiologically relevant,
in vivo conditions. We have successfully used these arrays to identify
genomic sites to which E2F and pRb family members are recruited,21,22
whereas others have identified Myc binding sites using these same arrays.23
As illustrated in Figs. 1 and 2. the assay subsumes six major areas:
1. ChIP
2. Generation of amplicons of the ChIPs to provide sufficient amounts
of chromatin for microarray analysis
3. Amplicon labeling
4. Preparation of CpG microarrays
5. Hybridization to spotted CpG arrays
6. Data analysis and confirmation of functional binding
Details concerning each of these areas are provided (see also Figs. 1
and 2).
Chromatin Immunoprecipitation
We have recently published a detailed protocol that we have found
will efficiently and specifically immunoprecipitate sequences associated
20
Genes Dev. 16, 245 (2002).
21
A. S. Weinmann, P. S. Yan, M. J. Oberley, T. H. Huang, and P. J. Farnham, Genes Dev. 16,
235 (2002).
22
J. Wells, P. S. Yan, M. Cechvala, T. Huang, and P. J. Farnham, Oncogene 22, 1445 (2003).
23
D. Y. L. Mao, J. Watson, P. S. Yan, D. Barsyte-Lovejoy, F. Khosravi, W. W. Wong,
P. Farnham, T. H. Huang, and L. Z. Penn, Curr. Biol. 13, 882 (2003).
Fig. 1. Preparation of samples for use in a ChIP-CpG assay (Steps I–III). Step I: cells
are crosslinked in culture (A), which creates protein-protein and protein-DNA crosslinks.
Nuclei are then isolated (B) and sonicated to create chromatin fragments of approximately
500–1000 bp. The chromatin is then immunoprecipitated with antibodies specific to the
protein of interest; an IgG control sample can also be prepared (C). After extensive washing,
the crosslinks are reversed and the DNA is purified, resulting in pools of sequences that are
enriched by virtue of their binding to the protein of interest. Chromatin is also purified from
crosslinked DNA that has not been immunoprecipitated; this serves as an input DNA control
(D). At this point, PCR reactions should be performed using a positive control (see Fig. 3A).
Step II: the pools of DNA are separately subjected to ligation-mediated PCR to generate
amplicons (E). At this point, PCR reactions should be performed using the same set of
primers as above (see Fig. 3B). Step III: the IP amplicon is labeled with Cy5, and an equal
amount of the total input amplicon is labeled with Cy3; the two will be combined to hybridize
to a CpG island array (see Fig. 2). The IgG amplicon can also be labeled with Cy5 and
hybridized with the input amplicon labeled with Cy3 as an additional control (see text).
Fig. 2. Preparation and hybridization of CpG arrays (Steps IV–VI). Step IV: CpG islands
are spotted onto glass slides and post-processed according to the manufacturer’s protocols.
Step V: the Cy5- and Cy3-labeled samples are hybridized to the CpG-island microarray in the

presence of CoT-1 DNA overnight at 60 . The slides are then washed to remove unbound
labeled sample and then dried. Step VI: the fluorescent intensity of each feature is read with
an appropriate scanner and data are imported into analysis program for further data analysis.
Features with intensities significantly higher in the ChIP-labeled samples relative to the input
sample are regarded as putative binding targets. However, if a feature also is positive on the
IgG versus input control array, it is removed from the list of targets.
with a wide variety of different types of DNA binding factors.24,25 We have

utilized many different cell types (epithelial cells, fibroblasts, blood cells,
etc.) and have also adapted them for use with mouse and human tissues.
Although minor modifications may be required to develop the optimal
set of conditions for a particular antibody and/or cell type, our published
protocol will provide a good starting point for most transcription factors.
This detailed protocol, as well as a discussion of how to prevent potential
problems that may be encountered, is also available in our web site
24
M. J. Oberley and P. J. Farnham, Meth. Enzymol. 371, 579 (2003).
25
A. S. Weinmann and P. J. Farnham, Methods 26, 37 (2002).
(http://mcardle.oncology.wisc.edu/farnham/). Due to space constraints, we

will not reiterate the entire ChIP protocol in this publication. Rather, we
have briefly described several recent modifications to the protocol which
we recommend be adopted when performing ChIPs (if a step in
the protocol is preceded by an exclamation point, this indicates that the
particular step has changed substantially from previous published protocols
from this group).
Modifications to the Previously Published ChIP Protocol
! a. Although our original protocol describes the standard approach as

proceeding to the cell lysis step directly after formaldehyde crosslinking,

we now routinely freeze the crosslinked cells and store the pellets at 70 .
Having a large number of cross-linked cells frozen in aliquots has proven
to be quite useful when troubleshooting the entire assay.
! b. Due to multiple PCR reactions in the subsequent steps, it is critical
to note that care must be taken to avoid cross-contamination of samples at
all steps of the ChIP protocol. We recommend the use of barrier tips at all
times.
! c. Another modification that we have adapted to the ChIP assay that
differs from our published protocols is the use of Qiaquick PCR
purification columns (Qiagen, 28104) in lieu of the phenol-chloroform
extraction step following proteinase K digestion. We have found this
shortens the time required to complete the assay, increases the yield of
chromatin recovered, and eliminates the need for handling and disposal of
the toxic phenol-chloroform mixture. Although we previously recom-
mended using these columns only if the samples would subsequently be
used to probe an array, we now suggest that investigators may wish to use
this alternative to phenol-chloroform extraction for standard ChIP assays.
! d. The major difference between our most recently published protocol
and how we currently design our experiments is concerned with sample
size. We originally recommended beginning with approximately 15
standard ChIP reactions and pooling the resultant precipitated chromatin
to probe an array. As described in detail later, we have now incorporated
an amplification step such that a single ChIP assay will provide sufficient
material for subsequent microarray analysis. However, as before, the most
critical step of the ChIP assay is to confirm that the assay has been
successful. After generating chromatin from the ChIP assay, do not
proceed with the creation of amplicons unless the signal obtained in a test
PCR experiment shows higher signal in the antibody samples than in the
no antibody (or pre-immune) control (see Fig. 3A).
Fig. 3. ChIP controls performed prior to hybridization. (A) A traditional ChIP assay should
be performed prior to subjecting the samples to LMPCR. In this example, ChIP was performed
on 293 cells according to published protocols25 using antibodies specific for E2F6 and RNA
polymerase II (Pol2); a control was also performed using only a secondary antibody (IgG).
After part D of Step I (see Fig. 1), PCR was performed with primers specific to the UXT or
MYC promoters and the resulting signal was plotted as a fraction of the signal obtained using
0.2% of the total input chromatin. As expected, E2F6 is highly enriched on the UXT promoter
but not the MYC promoter. Pol II occupancy is robustly detected on both promoters, in
accordance with the fact that both genes are expressed in 293 cells (data not shown). Neither
promoter is enriched by non-specific IgG alone. Therefore, these ChIP samples are
appropriate for use in subsequent LMPCR reactions. (B) A PCR reaction should be
performed after the samples are subjected to LMPCR. In this case, LMPCR was performed as
described in section (Chromatin Amplicon Labeling) on the samples used in panel (A). Ten
nanograms of each amplicon (E2F6, Pol II, IgG, and total) was then subjected to PCR with
primers specific for the UXT and MYC promoters. The enrichment of UXT sequences seen in
panel (A) by the E2F6 antibody is maintained after LMPCR, as the signal seen is at least 3-fold
higher than the starting input. The E2F6 signal is not enriched relative to the input amplicon on
the MYC promoter. These results indicate the LMPCR procedure did not introduce significant
bias into the system and that the samples are appropriate for hybridization to an array.
Generating Chromatin Amplicons

In our previously published studies, we began by pooling between 28
and 50 individual ChIP assays (using 1 107 cells/assay) to obtain enough
DNA to probe replicate arrays. Although tedious, it is generally possible to
obtain this requisite number of cells using immortalized human cell lines.
However, if multiple experimental points are to be analyzed or if primary
human cells or tissues are to be used, collecting this large number of cells
becomes onerous. Therefore, we have adapted a ligation-mediated PCR
technique (LMPCR) that is based on the method of Ren et al.13 to create
amplicons of individual IPs. We have found that as little as one-half of
one ChIP assay can be efficiently amplified to give amplicons that accur-
ately represent the starting population (see Fig. 3B). The ligation-mediated
PCR technique simply involves blunt ending the chromatin, ligating a uni-
directional double-stranded oligonucleotide linker, and PCR amplifying
the resultant DNA population. Other groups have utilized random priming
of the starting chromatin material with comparable results.12 Recently, a
promising amplification method has been developed that utilizes in vitro
transcription of the chromatin templates to create RNA amplicons.26
Because this procedure results in a linear rather than exponential amplifi-
cation, this method has the potential to reduce potential bias introduced
by PCR-based amplification schemes. Indeed, this form of sample amplifi-
cation is commonly used in microarray studies of gene expression where
mRNA samples are limited.27
To ensure that the LMPCR reaction has not resulted in a loss of differ-
ence between the experimental and control samples, we amplify the IP
samples as well as a small portion (10–100 ng) of the starting input DNA.
This amplified input DNA will be used as a reference on the CpG microar-
ray. We also commonly amplify a non-specific IgG ChIP control reaction
to demonstrate that enrichment is due to the specific antibody-epitope
interaction. Therefore, three different samples can be prepared that can
be used in two different hybridizations comparisons (i.e., IP versus input
and IgG versus input). Please note that this LMPCR amplification step is
a new addition to our protocol; to call attention to this fact, both Day 1
and Day 2 have been identified with an exclamation point; all steps in these
sections are new.
! Day 1:
1. The two unidirectional linkers oligoJW102 (50 gcggtgacccgggagatct-
gaattc 30 ) and oligoJW103 (50 gaattcagatc 30 ) are annealed by combining
6.7 l of 100 M oligoJW102, 6.7 l of 100 M oligoJW103, and 86.6 l
H2O. This mixture is boiled for 5 min in a water bath, then allowed to
slowly cool to room temperature. The annealed linker can be stored

indefinitely at 20 .
2. Because sonication of chromatin creates overhanging ends, the
chromatin is blunt ended by mixing the chromatin sample derived from an
experimental sample, an IgG sample, or 10 ng of input with the following:
11 l T4 DNA polymerase buffer, 5 l 10 BSA, 5 l 2 mM dNTPs, 1 l
T4 DNA polymerase (New England Biolabs, M0203S), and H2O to 110 l.
26
B. Bernstein and S. S. Schreiber, Meth. Ezymol. 376, 350 (2004).
27
C. C. Xiang, M. Chen, O. A. Kozhich, Q. N. Phan, J. M. Inman, Y. Chen, and M. J.
Brownstein, Biotechniques 34, 386 (2003).

This mixture is placed at 37 for 45–60 min, and then each sample is
purified with a Qiaquick PCR purification kit, according to manufacturers
protocol, eluting with 30 l of elution buffer.
3. Next, double-stranded unidirectional linker from Step 1 is ligated to
the blunted chromatin by adding 27 l of the blunted chromatin, 10.3 l
H2O, 5 l 10 T4 DNA ligase buffer, 6.7 l annealed JW102/JW103
linker, and 1.0 l T4 DNA ligase (New England Biolabs, M0202S). This

mixture is placed at 16 overnight.
! Day 2:
4. The following morning, each sample is purified with the Qiaquick
PCR purification kit, as earlier, eluting with 30 l of elution buffer.
5. The linker-ligated chromatin is then PCR amplified to create stocks
of amplicons from which virtually unlimited amounts of DNA can be
generated for use with the CpG-island microarrays. Each PCR reaction
contains 5 l 10 Taq polymerase buffer, 7 l 2 mM dNTPs, 3 l MgCl2,
6.5 l betaine, 2.5 l oligoJW102 (20 M), 1 l Taq (Promega, M1861),
and 25 l of the linker-ligated chromatin.
6. PCR is performed using the following protocol:
55 for 2 min #

72 for 5 min 1 cycle

95 for 2 min
95 for 0.5 min #

55 for 0.5 min 20 cycles
72 for 1 min

72 for 4 min

4 for 1
7. Following the PCR, each reaction is purified using the Qiaquick
PCR purification kit according to the manufacturer’s instructions, eluting
in 30 l of elution buffer.
8. The PCR protocol (Steps 5–7) is repeated until enough amplicon is

made for a master stock (1–10 g), which can be stored at 20 . After the
second 20 cycles, 5 l of the amplicon should be visible on a 1% agarose
gel, with fragments ranging from 200 to 600 bp, even if the starting
chromatin was sheared to a much larger size. If sufficient chromatin is not
produced after the second 20-cycle amplification, another round of PCR
may be performed (Steps 5–7). However, always amplify all samples to be
compared using the same number of overall cycles to avoid introducing
bias between the IP and input sample. The number of times required
to repeat the PCR protocols will depend on the quality and quantity of the
starting chromatin and the efficiency of linker ligation. After amplification,

carefully quantitate the amplicon with UV spectrophotometry and store

at 20 .
9. To prepare DNA for labeling, 10 ng of the master stock of the
amplicons of the experimental IP, the IgG IP, and the input sample are
amplified for 10 cycles, which should give 1–2 g of DNA. DNA is purified
using the Qiaquick PCR purification kit and then quantitated.
10. Using 10 ng of the experimental IP, the IgG IP, and the input
amplicons, a standard PCR reaction is performed using primers cor-
responding to a known positive control (the same as used to verify that the
original ChIP assays had been successful). Do not proceed if the positive
control is not enriched in the experimental amplicon relative to the input
amplicon (see Fig. 3B) or if the IgG amplicon shows non-specific
enrichment.
Chromatin Amplicon Labeling

For our hybridization experiments we routinely compare the enrich-
ment of the ChIP sample relative to the input chromatin. This provides sev-
eral advantages: first, this comparison gives a direct (but non-linear)
indication as to how robustly a given sequence is enriched during the ChIP
procedure. Second, by comparing the IP labeled with Cy5 to the input
chromatin labeled with Cy3, we generate a ratio that helps reduce bias intro-
duced by microarray printing errors. Finally, using two-color fluorescence
reduces the number of arrays required for each experiment.
There are two primary ways to label any given chromatin sample: direct
versus indirect. Direct labeling involves denaturing the chromatin, random
priming, and polymerizing new duplex DNA with a nucleotide analog that
has the cyanine fluorescent dye conjugated to it. Because these fluorescent
dyes are bulky, the polymerase will incorporate the conjugates at different
rates than it would a natural nucleotide. This leads to labeling bias; for
example, Cy5 tends to incorporate more readily than does Cy3. This has
to be taken into account when co-hybridizing both fluorophores onto the
same array, and the reciprocal experiment must be done to control for
these differences where the test and control samples are each labeled in
separate experiments with each dye. One can also directly label the ampli-
cons by including the cyanine dye conjugated nucleotides directly in the
PCR reaction; again because of labeling bias, this method also requires
reciprocal experiments in which the samples are labeled with the other dye.
For these reasons, our laboratory primarily utilizes an indirect means of
labeling the immunoprecipitated chromatin. Indirect means of labeling
incorporates a non-fluorescent nucleotide analog such as aminoallyl dUTP
or biotinylated dUTP, followed by conjugation of the fluorophore to the

incorporated nucleotide analog. Because these small conjugates are incor-
porated into both the test and reference amplicons equally, this method
helps to eliminate the incorporation biases that occur when directly label-
ing with Cy5-dUTP and Cy3-dUTP. The cyanine dyes can be purchased as
an NHS-ester conjugate (Amersham, RPN5661), which will covalently
bind to the aminoallyl nucleotide analogues, or as streptavidin conjugates
which form a tight complex with the biotinylated chromatin. We routinely
utilize aminoallyl labeling of our chromatin samples, followed by cyanine
dye coupling via the protocol outlined later. In addition, the aminoallyl
means of amplicon labeling has proved more cost-effective than using
one of the direct labeling methods.
Day 1
! 1. Two micrograms of the experimental IP amplicon, 2 g of the IgG
IP amplicon, and 2 g of the input amplicon are separately vacuum
desiccated to complete dryness with heat and then each is resuspended in
33 l molecular biologic-grade H2O. Then, 30 l of 2.5 random primer
buffer (Invitrogen, 18094-011) is added to each.
! 2. The chromatin is denatured by holding at 95 for 5 min on a PCR
block and then is immediately placed on ice for 3 min. The labeling
reaction is initiated by adding 7.5 l of the 10 dNTP mix (2 mM dATP,
2 mM dCTP, 2 mM dGTP, 0.35 mM dTTP), 1.8 l of 10 mM aminoallyl

dUTP (Sigma, A0410; make a 0.1-M stock in H2O and store at 20 ), and
2.5 l of high-concentration Klenow (40 U/l, Invitrogen, 18094-011) to

each, and holding at 37 for 2 h.
! 3. The excess nucleotides are removed with the Qiaquick PCR
purification kit, eluting with molecular biologic-grade H2O, as the cyanine
dyes that are subsequently coupled are sensitive to Tris base. It is very
important at this stage to remove as much of the unincorporated
aminoallyl dUTP as possible, because it will bind with the NHS-ester
cyanine dyes and reduce the labeling efficiency of the chromatin. The
samples are then dried with heat in a speed-vac.
! 4. To couple the NHS-ester cyanine dye to the chromatin, the aminoallyl-
labeled chromatin is resuspended in 4.5 l H2O, and the cyanine dye is
resuspended in 4.5 l of 0.1 M NaHCO3 (pH 9). The tubes are vortexed
and spun several times to ensure that all the samples are dissolved. Next,
the chromatin and Cy dye are combined and incubated at room
temperature in the dark for 1.5 h, agitating and microfuging the samples
every 15 min of the coupling. We typically couple the Cy5 dye to the
experimental IP (or the IgG) amplicon, and Cy3 dye to the input reference
amplicon, but these could be reversed as well. Dye swapping is not
necessary because utilizing aminoallyl dUTP incorporation alleviates

cyanine dye bias problems.
5. Thirty five microliters of 100 mM Na-acetate (pH 5.2) are added to
lower the pH of the solution to allow the chromatin to bind to the Qiaquick
columns, and then H2O is added to a total of 100 l for each sample.
6. We have found that removal of the unincorporated dye is quite
efficient with Qiaquick PCR columns, and we follow the Qiagen protocol
except that we again elute with 50 l of water instead of the EB buffer
provided with the kit. The elution is performed twice and the eluates are
combined to give a total volume of 100 l of H2O.
7. Determination of the minimum amount of labeled DNA required
for any given set of CpG microarrays needs to be empirically determined
and this depends greatly on the quality of the microarrays that are produced.
If one wants a measure of the success of the labeling reaction, the total dye
incorporation can be determined by measuring the absorbance of the entire
sample at 650 nm for Cy5 and 550 nm for Cy3. To determine the number of
picomoles of incorporated Cy5 or Cy3, the A650 reading for Cy5 and A550
reading for Cy3 are taken on a spectrophotometer after dye conjugation to
the aminoallyl-labeled amplicons. The picomoles incorporation of dye is
calculated by using the following formula:
ðA650 Þðl of solutionÞðdilution factorÞ
¼ pmol of Cy dye incorporated
ð0:25 for Cy5; or 0:15 for Cy3Þ
By determining the picomoles incorporation necessary to give high quality
data on the arrays, one can avoid wasting arrays due to poorly labeled
samples. At this stage we also measure the amount of labeled DNA. After
aminoallyl labeling of the amplicons, one can ascertain the success of
the incorporation of the nucleotide analog if the measured amount of the
DNA is amplified at least 2- to 5-fold over the starting amount. However,
we have found that starting the labeling procedure with 2 g of labeled
chromatin is the least amount of chromatin one can use that will give a
good signal/noise ratio on the array. If the arrays are not giving good
median intensities per spot (i.e., >1000), one can start with two or three
times as much amplicon. Because the cyanine dyes are easily photo-
bleached, it is best to perform all steps under minimal light exposure
conditions, both during and after the cyanine dyes have been conjugated.
The cyanine dyes are also extremely susceptible to degradation by air
humidity, so they should be stored dry with a desiccant at all times. Dry the

labeled chromatin with heat in a speed-vac and store it dry at 20 until
ready for hybridization.
Preparation of the Arrays

CpG islands, found in or near approximately 50% of human promoters,
are identified by three primary characteristics: they are more than 200 bp
long, have over 50% GC composition, and retain an observed/expected
ratio of CpG dinucleotides greater than 0.6.28 The CpG library used on
the microarrays was originally prepared in the laboratory of Bird29 and
consists of approximately 60,000 independent clones that represent CpG
islands that were not hypermethylated in human male genomic DNA. Ap-
proximately 8000 clones representing hypomethylated CpG islands were
arrayed by the Huang lab30 onto slides for high-throughput study of
DNA methylation changes in human cancer. We have successfully used
these arrays to identify genomic binding sites for E2F4, E2F1, and
Rb.21,22 However, these clones were not sequence verified and therefore
cultures of each putative positive target clone had to be grown and DNA
prepared and sequenced to determine the location on the genome of the
identified CpG island. To circumvent this limiting aspect of the CpG
arrays, a new version of the arrays is now in production. For these
new arrays, the clones are sequenced prior to spotting. This prior know-
ledge of the identity of each clone will greatly speed the data analysis
and target gene identification steps.
The CGI library29 was obtained from the United Kingdom Human
Genome Mapping Project (http://hgmp.mrc.ac.uk/geneservice/reagents/)
as frozen bacterial cultures (methylation tolerant Escherichia coli strain
XL1-Blue MRF0 ) harboring individual clones in the pGEM-5Zf() vector.
The clone set consists of approximately 12,200 independent clones stored
in 96-well plates. The entire clone set was amplified using a 96-well
format. The cultures were stabbed and transferred into a total volume of
100 l containing 4 U Taq DNA polymerase, 10 mM Tris, pH 8.4,
50 mM KCl, 1.5 mM MgCl2, 800 M dNTPs, and 0.2 pM each primer.

The PCR mixture was held at 94 for 2 min and then cycled 40 times at

94 for 30 s, 53.5 for 30 s, and 72 for 1 min, followed by holding at 4 after
the final cycle. From each reaction 2.5 l were electrophoresed on a 2%
agarose gel to check the yield and quality of PCR products. The remaining
reaction volume was filtered through Unifilter 800 filter plates (Whatman,
7700-2803) and the purified PCR products were quantified by spectro-
photometric measurements at 260 and 280 nm. Then approximately 2 g
of DNA from each clone was transferred into 384-well polyfiltronic plates
28
I. P. Ioshikhes and M. Q. Zhang, Nat. Genet. 26, 61 (2000).
29
S. H. Cross, J. A. Charlton, X. Nan, and A. P. Bird, Nat. Genet. 6, 236 (1994).
30
P. S. Yan, C. M. Chen, H. Shi, F. Rahmatpanah, S. H. Wei, and T. H. Huang, J. Nutr. 132,
2430S (2002).
(Whatman, 7701-5101) using Evolution P3 liquid-handling robotics

(Perkin Elmer). The PCR products were subsequently dried by overnight
centrifugation in vacuum dessicator and resuspended in 10 l of 3 SSC
for array printing. The CpG microarrays were printed from 32 384-well
plates containing around 0.2 g/ul of DNA in 3 SSC in the VersArray
ChipWriter Pro System (Bio-Rad, 169-0006) using a 48-pin (SMP-3 pins,
ArrayIT, product #SMP3) configuration. Each clone was spotted once
on GAPS II–coated slides (Corning, 40003). The printed arrays were pro-
cessed following manufacturer’s protocol (Corning) and kept in a dessica-
tor for later use. (Contact the Microarray Centre, University Health
Network, Toronto, ON M5G2C4 at [email protected] for information
concerning the distribution of the human CpG arrays.)
Hybridization of the Arrays

Day 1:
! 1. The CpG-island microarray is placed in the hybridization chamber
probe side up. A 24 60 mm Lifterslip (Erie Scientific, 25 60I-2-4789) is
carefully placed over the area containing the DNA probes to create a
uniform space between the array and coverslip.
! 2. The Cy5- and Cy3-labeled chromatin are each resuspended in 15 l
of 1.0 g/l CoT-1 DNA (Invitrogen, 15279-011), and the two are then
combined. CoT-1 DNA is included to bind to repeat elements in the
chromatin which helps to prevent non-specific binding to the array. The
tube should be vortexed several times to completely dissolve the chromatin
and then spun in a microfuge. Seventy microliters of the hybridization
solution (Genisphere Buffer 6, cat#100V600) are added without vortexing

to avoid bubble formation. The mixture is denatured at 95 for 2 min and

then held at 37 for 30 min to allow the CoT-1 DNA to hybridize with
repeats. The mixture is then applied to the microarray by carefully adding
the solution to one end of the array and allowing it to wick underneath the
Lifterslip. Note that the size of the Lifterslip and the final volume of the
hybridization solution may need to be adjusted to match the printed area
of the microarray.
! 3. We utilize a dual hybridization chamber (Genemachines, HYB-03)
and a water bath for the hybridization to maintain a constant temperature.
It is important that the hybridization chamber not directly contact the
bottom of the water bath because it will cause inconsistent heat transfer.
We typically use a small piece of foam to separate the chamber from the

metal bottom of the water bath. The arrays are hybridized overnight at 60
for up to 18 h.
Day 2:
11. To remove the coverslip, the microarray(s) are inverted in a glass

dish filled with 1 SSC and 0.1% SDS preheated to 50 . If no drying has
occurred overnight, the Lifterslip will immediately slide off.

12. The arrays are then agitated in 1 SSC, 0.1% SDS at 50 for 5 min,
followed by agitation in 1 SSC, 0.1% SDS at room temperature for
5 min. Finally, the arrays are washed with 0.2 SSC for 5 min at room
temperature to remove residual SDS (which autofluoresces), and then
dried by centrifugation at 600 rpm for 5 min in a 50-ml conical tube.
13. The arrays are scanned immediately using an Axon 4000B scanner
(Axon Instruments).
Analysis
The identification of positive clones is a multi-step procedure, first re-
quiring normalization of the Cy5 and Cy3 channels on the array because
of potential labeling efficiency differences, followed by clone selection,
confirmation, and ultimately, functional analysis.
1. The hybridized microarrays are analyzed using the Genepix Pro 4.0
(Axon Instruments) software package. This provides a set of raw values for
each feature on each array. To identify clones that are selectively enriched
during IP relative to the starting population, one must first normalize the
Cy5 and Cy3 channels on the array. To do so, we make the assumption that
the vast majority of loci will not be enriched in the IP relative to total; thus,
we can normalize across the entire array, by taking the ratio of the medians
for all features and normalizing them to unity.
2. After normalization, a ratio is generated that is the intensity in the
Cy5 channel minus background divided by the intensity in the Cy3 channel
minus background. Features in areas of the array that are obviously
blemished are manually flagged and excluded from the data set, as are
features with signal that have low intensities (<500). The ratio is generated
for all features that meet this criterion, and features with ratios above 2 are
isolated for further analysis.
3. Because of the complexity of the mammalian genome and the
relatively large number of steps in this procedure, false positives are
inevitable. Great care must be taken to reduce these errors to prevent
further wasting of time and resources following up on potential artifacts. It
is imperative to reduce the error rate by repeating the entire procedure at
least three times, starting with independent chromatin IPs from a different
batch of crosslinked cells each time (see Fig. 4B).
Fig. 4. Confirmation of putative targets. (A) Gene-specific PCR should be performed to

verify that hybridization signals resulting from the microarray accurately represent the
relative amounts of sequences present in the amplicons. In this case, 10 ng of each amplicon
was subjected to PCR with promoter-specific primers from genes that were positive on the
CpG-island array analysis. In our hands >90% of the genes that are positive on the arrays
demonstrate at least 2-fold enrichment in the experimental IP amplicon relative to input
amplicon. (B) Reproducible identification of the putative clones should be performed using
traditional ChIP analysis in an independent assay.25 In this example, 2 l of each IP from a
second ChIP experiment was PCR amplified using the same gene-specific primers indicated in
panel (A) and the signal was compared to 0.2% of the input chromatin. Clones 1 and 2 show
the same pattern in the original amplicon and the subsequent, independent ChIP assay.
However, clone 3 is an example of a clone that appears to be robustly enriched in panel (A)
but does not appear to be bound by E2F6 in subsequent ChIP assays. This highlights the fact
that multiple, independent analyses are required due to inherent noise generated during ChIP.
The independent validation rate varies from antibody to antibody but can be increased by
utilizing amplicons from three independent ChIPs, performing three independent hybridiza-
tions, and discarding any promoters from the data set that are not repeatedly robustly
enriched relative to the input amplicon.
4. To ensure that positive signals are not simply a consequence of non-

specific enrichment of certain loci by the ChIP procedure, a comparison of
the IgG ChIP sample to input DNA must also be performed. CpG islands
that are non-specifically enriched as a result of the procedure are then
removed from the putative target list.
5. For an initial test to demonstrate that the positive signals derived
from microarray hybridization accurately reflect the relative abundance of
sequences present in the labeled chromatin, the amplicons should be
subjected to gene-specific PCR analysis starting with 10 ng of LMPCR
experimental IP amplicon, 10 ng of the LMPCR IgG amplicon, and 10 ng
of the LMPCR input amplicon. This should be done for a selected number
(10–20) of the CpG islands that appear enriched after data analysis. This
gene-specific PCR analysis will indicate which CpG islands are enriched
significantly over input (see Fig. 4A), giving an empirical measure of the
hybridization accuracy. In our initial experiments, we employed arrays in
which the clones corresponding to the CpG islands were not sequenced
prior to spotting the arrays. Therefore, to perform the test PCRs, the clones
had to be grown as overnight cultures, plasmid DNA prepared, and then
sequenced using vector primers. Gene-specific primers for the putative
positives could then be constructed and used in the confirmation assays.
However, future experiments will be much easier due to the fact that the
new arrays will be composed of sequence-verified clones.
6. Reproducible binding is demonstrated using an independent
standard ChIP assay in which 2 l of an individual IP is examined via
PCR with gene-specific primers and compared to an IgG IP and a fraction
(0.2%) of the input chromatin (see Fig. 4B). This will identify false-positive
clones that may have been enriched independently of the antibody used;
these clones should be discarded from the putative target list. Ultimately,
an ideal experimental validation can be derived from cells with specific
gene deletions and their wild-type counterparts to show the PCR signals
are the result of the desired specific antibody-epitope interaction. For
example, Wells et al.31 identified several E2F-1–specific genes and
validated them using ChIPs in wild-type and E2F-1–knockout mouse
embryo fibroblasts.
7. The final step is to determine the biological significance of the
experimentally verified binding of the factor of interest to the identified
genomic site. For example, RNAi represents an experimental system where
one can knock-down the expression of a particular gene in a transient way.32
31
J. Wells, C. R. Graveel, S. M. Bartley, S. J. Madore, and P. J. Farnham, Proc. Natl. Acad.
Sci. USA 99, 3890 (2002).
32
Y. Zhang, Science 298, 1039 (2002).
This system allows one to perturb a normal biological system without some
of the concerns with traditional gene-ablation technology where the
knock-out of a specific gene may be compensated for during development
of the animal by upregulation of other family members. One could use
RNAi to specifically knock-down the expression of the DNA binding
factor of interest and then use RNase-protection, Northern, or traditional
microarray gene expression analysis to determine whether the lack of
binding to a putative target has a functional consequence. Shi et al.33 have
developed protocols that allow the CpG-island microarray to be probed
with mRNA that utilizes RNA ligase-mediated full-length cDNA synthesis.
With this protocol, it will be possible to examine the effect of DNA binding
factor knock-down on the resultant mRNA expression of the identified
target genes using the same CpG-island microarrays as used for the target
gene identification.
Conclusions
Studies to date employing the ChIP-chip approach using human cells
have been limited by the need to refine the methodology used for similar
studies of less complex eukaryotes. However, great strides have been made
recently in overcoming issues related to sample size and genomic complex-
ity. A major problem in studying human transcription factors is the fact
that no appropriate microarrays have been widely available that can be
used to study intergenic DNA; that is, all commercially available microar-
rays have allowed only the detection of transcribed sequences. Although
individual investigators have begun spotting promoter-specific13 or pro-
moter-enriched30 arrays, individual labs cannot serve as a source for arrays
for all the investigators who wish to employ the ChIP-chip approach. Un-
fortunately, commercial microarray companies have not yet entered into
the production of human promoter arrays. However, now that the feasibil-
ity of using such arrays has been demonstrated, greater efforts to provide
investigators with a comprehensive set of human intergenic DNA chips
must be made. The Encode (Encyclopedia of DNA Elements) project,
which has the long-term goal of identifying all functional elements in the
human genome, has chosen 30 Mb (approximately 1% of the human
genome) for preliminary analysis (see http://www.genome.gov/ Pages/
Research/ENCODE/). It can be hoped that the technologies and products
arising from such projects will allow much greater access of all investigators
to the required reagents needed for a thorough analysis of the regulatory
elements encoded in the human genome.
33
H. Shi, P. S. Yan, C. M. Chen, F. Rahmatpanah, C. Lofton-Day, C. W. Caldwell, and T. H.
Huang, Cancer Res. 62, 3214 (2002).
Acknowledgments
We would like to thank P. S. Yan and T. Huang for their generous gift of the CpG arrays.
M.J.O would like to thank A. Kirmizis for help adapting the LMPCR protocol and for stimulating
discussion. The Farnham lab would also like to thank A. Kirmizis and A. S. Weinmann for critical
readings of this manuscript. The Microarray Centre, which would like to thank J. Woodgett and
Neil Winegardin for technical input, was supported by ORDCF and Genome Canada. This work
was supported in part by Public Health Service grants CA22484, CA09135, and CA45240.
[22] Chromatin Immunoprecipitation in the Analysis

of Large Chromatin Domains Across Murine
Antigen Receptor Loci
By David N. Ciccone, Katrina B. Morshead, and
Marjorie A. Oettinger
In the nucleus of eukaryotic cells, DNA is wrapped around nucleosomes

forming a regulatable polymer called chromatin. Chromatin structure has
been shown to influence many processes, including transcription, replica-
tion, recombination, DNA repair, and chromosome segregation. The in-
timate association between DNA and histone proteins of the nucleosome
can be regulated by histone-modifying enzymes and chromatin remodeling
machines to permit or restrict the occupancy of specific regulatory se-
quences by DNA binding proteins. The chromatin fiber is a very dynamic
structure that must be opened and subsequently closed in a strict temporal
and spatial manner to ensure the proper gene expression pattern and
the developmentally appropriate accessibility of certain chromatin
domains.1 Therefore, mapping the distribution of covalently modified
histones, regulatory proteins, or DNA binding factors relative to specific
sequences within a genomic locus will ultimately provide valuable insight
into the molecular mechanisms underlying localized chromatin alterations
as well as developmentally regulated, large-scale reorganizations of
chromatin structure into active and inactive domains.
Chromatin immunoprecipitation (ChIP) is a powerful technique used
to analyze protein:DNA interactions in a native chromatin context. This
versatile method has been used with cell types from a wide range of
organisms to determine the localization of specific histone modifications
with respect to promoter regions and large chromosomal domains.2,3 In
1
W. Fischle, Y. Wang, and C. D. Allis, Curr. Opin. Cell Biol. 15, 172 (2003).
2
M. H. Kuo and C. D. Allis, Methods 19, 425 (1999).
3

Acknowledgments
We would like to thank P. S. Yan and T. Huang for their generous gift of the CpG arrays.
M.J.O would like to thank A. Kirmizis for help adapting the LMPCR protocol and for stimulating
discussion. The Farnham lab would also like to thank A. Kirmizis and A. S. Weinmann for critical
readings of this manuscript. The Microarray Centre, which would like to thank J. Woodgett and
Neil Winegardin for technical input, was supported by ORDCF and Genome Canada. This work
was supported in part by Public Health Service grants CA22484, CA09135, and CA45240.
[22] Chromatin Immunoprecipitation in the Analysis

of Large Chromatin Domains Across Murine
Antigen Receptor Loci
By David N. Ciccone, Katrina B. Morshead, and
Marjorie A. Oettinger
In the nucleus of eukaryotic cells, DNA is wrapped around nucleosomes

forming a regulatable polymer called chromatin. Chromatin structure has
been shown to influence many processes, including transcription, replica-
tion, recombination, DNA repair, and chromosome segregation. The in-
timate association between DNA and histone proteins of the nucleosome
can be regulated by histone-modifying enzymes and chromatin remodeling
machines to permit or restrict the occupancy of specific regulatory se-
quences by DNA binding proteins. The chromatin fiber is a very dynamic
structure that must be opened and subsequently closed in a strict temporal
and spatial manner to ensure the proper gene expression pattern and
the developmentally appropriate accessibility of certain chromatin
domains.1 Therefore, mapping the distribution of covalently modified
histones, regulatory proteins, or DNA binding factors relative to specific
sequences within a genomic locus will ultimately provide valuable insight
into the molecular mechanisms underlying localized chromatin alterations
as well as developmentally regulated, large-scale reorganizations of
chromatin structure into active and inactive domains.
Chromatin immunoprecipitation (ChIP) is a powerful technique used
to analyze protein:DNA interactions in a native chromatin context. This
versatile method has been used with cell types from a wide range of
organisms to determine the localization of specific histone modifications
with respect to promoter regions and large chromosomal domains.2,3 In
1
W. Fischle, Y. Wang, and C. D. Allis, Curr. Opin. Cell Biol. 15, 172 (2003).
2
3

[22] chromatin immunoprecipitation 335
conjunction with previously favored procedures such as nuclease sensitivity

mapping, ChIP also provides an additional tool to search for specific DNA
regulatory elements. For example, at the 47-kb silent mating-type locus of
the fission yeast Saccharomyces pombe, the euchromatin region was found
to be associated with high levels of histone H3–lysine 4 (Lys4) methylation.
Alternatively, the heterochromatin interval, located completely within the
euchromatic domain and defined by two flanking boundary elements, is as-
sociated with histone H3 (Lys9) methylation and Swi6 (yeast HP1 homo-
logue) localization.4 In a separate study, an inverse correlation between
acetylation and methylation of histone H3 (Lys9) was observed during
the developmental changes associated with erythropoiesis at the 53-kb
chicken -globin locus.5
The murine immunoglobulin heavy chain (IgH) locus represents an-
other example of a large, highly regulated, chromosomal domain (see
Fig. 3A for a depiction of the IgH locus). The IgH locus consists of many
variable (V ) region gene segments located upstream of diversity (D) and
joining (J) gene segments spread across a genetic domain that exceeds a
megabase of DNA. Coordination of the series of DNA rearrangement
events required to assemble an Ig receptor from component V, D, and J
segments is a process called V(D)J recombination.6 This recombination
process generates a diverse repertoire of Ig receptors through multiple
combinations of possible joining events between various germline gene
segments. A vast amount of genetic and biochemical evidence suggests that
V(D)J joining is developmentally regulated by specific alterations in chro-
matin structure that render a particular locus, or certain regions within a
locus, accessible or inaccessible to the recombinase machinery.
Here we present a ChIP protocol used to identify the histone modifica-
tion patterns at this locus. Data generated from ChIP experiments have
provided valuable insight into how specific gene segments within a locus
are made accessible to the recombination machinery while simultaneously
restricting access to other gene segments during lymphocyte development.
The ChIP method involves three basic steps:
a. In vivo formaldehyde crosslinking of intact cells
b. Micrococcal nuclease (MNase) digestion and sonication of the
chromatin to generate a pool of small chromatin fragments
c. Selective immunoprecipitation (IP) of protein:DNA complexes from
the chromatin pool with specific antibodies
4
K. Noma, C. D. Allis, and S. I. Grewal, Science 293, 1150 (2001).
5
M. D. Litt, M. Simpson, M. Gaszner, C. D. Allis, and G. Felsenfeld, Science 293, 2453 (2001).
6
C. H. Bassing, W. Swat, and F. W. Alt, Cell 109 (Suppl.), S45 (2002).
This chapter is devoted to a full, detailed description of this ChIP

method, including the subsequent analysis of the immunoprecipitated
DNA.
The basic procedures used for ChIP have been developed by and repre-
sent a fusion of contributions made by a large number of researchers. A
handful of these pioneering studies are outlined here: In 1981, whole cell
formaldehyde fixation was shown to crosslink and preserve chromatin
structure7,8; a few years later, antisera specific for RNA polymerase9 and
topoisomerase I10 were used to precipitate UV crosslinked protein:DNA
complexes; in 1988, histone antibodies were used in ChIP experiments11
and a correlation between core histone acetylation and transcriptionally
poised chromatin was observed.12 Recently, with the development and sub-
sequent commercialization of highly specific antibodies recognizing various
histone modifications, ChIP has become a very popular method for analyz-
ing chromatin structure in vivo. Although many variations of the basic
ChIP method are currently used, the backbone of our protocol relies
largely on the method reported by Kuras and Struhl13 and is described in
detail.
Procedure
1. The number of cells harvested for a single ChIP can vary
tremendously depending on the total number of DNA sites being
analyzed. The DNA we recovered from a single ChIP experiment was
used to examine the association of various histone modifications at ap-
proximately 30 loci across a large genomic region. However, the total
number of cells used per ChIP can be reduced if the analysis involves only
a few DNA sites. A transformed recombinase-deficient Pro B cell line was
chosen as the most practical line for examining the chromatin structure of
the murine IgH locus. However, this ChIP method has also been
successfully applied to additional suspension cell lines as well as two
non-adherent fibroblast cell lines.
7
V. Jackson and R. Chalkley, Proc. Natl. Acad. Sci. USA 78, 6081 (1981).
8
V. Jackson and R. Chalkley, Cell 23, 121 (1981).
9
D. S. Gilmour and J. T. Lis, Proc. Natl. Acad. Sci. USA 81, 4275 (1984).
10
D. S. Gilmour, G. Pflugfelder, J. C. Wang, and J. T. Lis, Cell 44, 401 (1986).
11
M. J. Solomon, P. L. Larsen, and A. Varshavsky, Cell 53, 937 (1988).
12
T. R. Hebbes, A. W. Thorne, and C. Crane-Robinson, EMBO J. 7, 1395 (1988).
13
L. Kuras and K. Struhl, Nature 399, 609 (1999).
Use approximately 2 108 cells for a single IP. To assess the efficiency
of the ChIP experiment, a Control IP lacking an antibody is performed in
parallel with one containing the specific antibody of interest. Therefore, for
a properly controlled experiment, approximately 4 108 cells are needed.
From a larger cell preparation, such as 8 108 cells, three independent
ChIP experiments can be performed with a single Control IP. Harvest
the cell population (collect adherent cells by trypsinization and wash the
cell pellet twice in ice-cold 1 PBS) and pellet at 640g for 4 min.
Resuspend the cell pellet thoroughly in 20 ml of culture media.
2. Fix the cells by adding formaldehyde to a final concentration of 1%
directly to the cell suspension. Allow the fixation reaction to proceed for
10 min in a fume hood at room temperature. Invert the suspension every
few minutes to mix the cells. Add 1/20 the volume of 2.5 M glycine to stop
the fixation reaction, and incubate for an additional 5 min at room
temperature. Transfer the cell suspension to an ice bucket.
3. Pellet the cells at 640g for 4 min (all centrifugations should be

performed at 4 until otherwise noted). Discard the supernatant into an
appropriate receptacle designated for formaldehyde waste. Resuspend the
cell pellet in 10 ml of 1 PBS and collect the cells by centrifugation.
Discard the supernatant and repeat the 1 PBS wash to remove all traces
of formaldehyde from the cell suspension
4. Isolate nuclei from fixed cells by washing the cell pellet three times
with 10 ml of cold lysis buffer (10 mM Tris-HCl, pH 7.5, 10 mM NaCl,
3 mM MgCl2, and 0.5% NP-40). After each wash, pellet the cells at 300g
for 5 min.
5. Resuspend the pelleted nuclei in 3 ml of MNase reaction buffer
(10 mM Tris-HCl, pH 7.5, 10 mM NaCl, 3 mM MgCl2, 1 mM CaCl2, 4%
NP-40, and 1 mM PMSF added just prior to use). Add 50 units of MNase

(Sigma) to the nuclei suspension, mix, and incubate for 10 min at 37 . The
amount of enzyme and the length of incubation should be specifically
tailored for optimum digestion when using different cell types or cell
numbers. Add EGTA to a final concentration of 3 mM to stop the

digestion reaction. At this point, save a 10-l aliquot and store at 20 .
This aliquot should be decrosslinked (described later) at the end of this
procedure and then run out on an agarose gel, providing a clear indication
of the efficiency of MNase digestion.
6. Add 60 l of 50 mM PMSF, 3 l of 2 mg/ml aprotinin, 3 l of
2 mg/ml leupeptin, 3 l of 2 mg/ml pepstatin, 300 l of 10% SDS, and
120 l of 5 M NaCl to the MNase-digested nuclei. Sonicate the nuclei to
lyse the nuclear membrane and further shear the chromatin into a sample
containing primarily 0.2–2 kb DNA fragments. The settings and conditions
used during sonication must be determined empirically and tend to vary
significantly between different sonicators. (Note: the sonication conditions

outlined here have been used successfully with different murine cell types.)
We use a Heat Systems-Ultrasonics sonicator (model W-375) set at power
level 4 and 50% duty cycle. Place the nuclei on ice and sonicate the sample
for 1 min, followed by a 1-min incubation on ice. Repeat the sonication/
incubation step. Finally, sonicate the sample one last time for 30 s. Split
the sonicated sample equally into microcentrifuge tubes and spin at

maximum speed for 10 min at 4 to remove all traces of cellular and
nuclear debris prior to IP. This spin can be repeated if some cellular
remnants remain in the supernatant. Again, save a 10-l aliquot and store

at 20 . This aliquot should also be decrosslinked (described later) and
run out on an agarose gel to determine the sonication efficiency and
ultimately the size range of the DNA fragments in the chromatin sample

used for IP. At this point the chromatin can be stored at 80 and the IP
can be performed at a later date.

7. If the sample was stored at 80 , spin down any additional precipitates

at maximum speed for 5 min at 4 . Transfer the supernatant to a 50-ml
conical tube. Add IP dilution buffer (20 mM Tris-HCl, pH 8.0, 2 mM EDTA,
1% Triton X-100, 150 mM NaCl, 1 mM PMSF, 2 g/ml aprotinin, 2 g/ml
leupeptin, and 2 g/ml pepstatin) to the tube bringing the volume up to 15 ml
(remember to always add protease inhibitors to all appropriate buffers just
before use). The chromatin sample is now pre-cleared with Protein A
Sepharose 4B resin; however, alternate resins may be needed depending
on antibody isotype. Since the resin will be used to pull down antibody-
bound chromatin fragments from the IP solution, it is helpful to pre-clear
the sample with resin, removing any chromatin fragments that non-
specifically bind to the resin prior to IP. Equilibrate 50 l resin in IP
dilution buffer and add to the diluted chromatin sample. Pre-clear the
chromatin at room temperature for 15 min while rocking. Pellet the resin
at 2000g for 2 min. Transfer the supernatant to a new 50-ml conical tube

and discard the resin. Save a 50-l Input aliquot and store at 20 . This
aliquot represents a fraction of the total chromatin sample used in the IP.
8. Split the chromatin sample into an appropriate number of 50-ml
conical tubes and bring the final volume of each tube up to 10 ml with IP
dilution buffer. One of these tubes will be used for the Control IP;
therefore, no antibody should be added. To the other tube or tubes add the
appropriate amount of specific antibody, which must also be determined
empirically by comparing the amount of DNA recovered (see DNA
Quantitation later) from a Control IP versus a specific IP and adjusting the
antibody concentration accordingly (for antibodies specific to covalently
modified histones, we typically use dilutions ranging from 1:500 to 1:2000).
The optimum antibody dilution used for ChIP experiments can vary
greatly depending on the specificity and target of the antibody. It is

generally accepted that polyclonal antibodies are better suited for ChIPs
than monoclonal antibodies, presumably due to the wide range of epitopes,
some of which may be masked in a native chromatin context, recognized
by the polyclonal antibody pool. Incubate at room temperature for 3 h
while rocking. Add 50 l of pre-equilibrated resin to all tubes and continue
rocking at room temperature for an additional 3 h.
9. Collect the resin by spinning at 2000g for 2 min (these conditions
should be used for all subsequent spins unless indicated otherwise).
Discard the supernatant. Resuspend the resin in 1 ml of IP dilution buffer
and transfer the solution to a microcentrifuge tube (or a dolphin-nosed
tube, Corning Cat. No. 3213). Spin down the resin and discard the
supernatant. The subsequent steps are designed to disrupt non-specific
binding that may have occurred during the IP by washing the resin in a
series of buffers with increasing salt concentrations. Add 1 ml of ChIP
wash buffer I (20 mM Tris-HCl, pH 8.0, 2 mM EDTA, 1% Triton X-100,
0.1% SDS, 150 mM NaCl, 1 mM PMSF), collect the resin by centrifuga-
tion and discard the supernatant. Add 1.5 ml of ChIP wash buffer I to the
resin and rock at room temperature for 5 min. Spin down the resin and
discard the supernatant. Add 1.5 ml of ChIP wash buffer II (20 mM Tris-
HCl, pH 8.0, 2 mM EDTA, 1% Triton X-100, 0.1% SDS, 500 mM NaCl,
1 mM PMSF) to the resin and rock at room temperature for 5 min. Collect
the resin by spinning and discard the supernatant. Add 1.5 ml of ChIP
wash buffer III (10 mM Tris-HCl, pH 8.0, 1 mM EDTA, 0.25 M LiCl, 1%
NP-40, 1% deoxycholic acid (sodium salt)) to the resin and rock at room
temperature for 5 min. Spin down the resin and discard the supernatant.
Add 1 ml of TE buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA) to the
resin and rock at room temperature for 1 min. Collect the resin by spinning
and discard the supernatant. Repeat the last wash with TE buffer, spin, and
discard the supernatant.
10. Add 200 l of elution buffer (25 mM Tris-HCl, pH 7.5, 10 mM
EDTA, 0.5% SDS) to the extensively washed resins from each sample, and

incubate for 15 min at 65 to dissociate the resin from the antibody-
chromatin complexes. Collect the resin by spinning at 5000g for 2 min and
transfer the supernatant from each sample to a new tube. Resuspend the
resin in 100 l of elution buffer, pellet the resin at 5000g for 2 min, and
again transfer the supernatants to the new tubes.
11. Remove the formaldehyde crosslinks by adjusting the volume of
the post-MNase, post-sonication, and Input aliquots to 300 l with elution
buffer and add SDS, as necessary to each sample for a final concentration
of 0.5%. Add Pronase (Roche Applied Science) to all samples to a final

concentration of 1.5 g/l. Incubate the samples at 42 for 1 h and then at

65 for 5 h. Add LiCl to the samples to a final concentration of 0.8 M.
To remove all traces of protein, extract the samples in an equal volume
of phenol/chloroform/isoamyl alcohol, followed by an extraction in an
equal volume of chloroform/isoamyl alcohol to remove any residual
phenol. Precipitate the DNA by adding 20 g of glycogen and 1 ml of

100% ethanol to each sample, vortex and incubate at 80 for at least 1 h.
Spin at maximum speed for 15–30 min in a microcentrifuge. Wash
the DNA pellet in 1 ml of 70% ethanol. Allow the pellet to air-dry
completely, resuspend the DNA in an appropriate volume of TE buffer,

and store at 20 .
12. At this point the DNA precipitated from the post-MNase and post-
sonication aliquots should be run out on an agarose gel to assess the
efficiencies of both treatments. To demonstrate the role of MNase in the
outlined procedure, Fig. 1 shows the DNA collected after sonication from
a ChIP experiment utilizing MNase (þMNase, lane 1) and ChIP in which
the enzyme was omitted, relying solely on sonication (MNase, lane 2). In
lane 1 there is a ladder consisting of approximately 200-bp increments
characteristic of efficient MNase digestion at nucleosome boundaries.
Since the length of the DNA fragments seen in the post-sonication aliquot
is a direct indication of the length of the chromatin fragments used in the
IP step, it is extremely important to the overall resolution of the ChIP that
the size of the DNA fragments reside between 200 bp and 2 kb (see Fig. 1,
lane 1). If this range is greatly extended beyond 2 kb, then the sonicator
settings and conditions should be adjusted. In the absence of MNase (see
Fig. 1, lane 2), some smaller DNA fragments are generated through
sonication, but a large percentage of the DNA remains as high molecular
weight fragments.
DNA Quantitation
A clear indication of the overall efficiency of the ChIP is assessed by
comparing the amount of DNA present in the IP sample to the amount
present in the Control IP. However, the amount of DNA recovered in both
of these samples is much too low for accurate quantitation using standard
procedures such as UV absorbance. The PicoGreen dsDNA Quantitation
Assay from Molecular Probes provides a rapid and sensitive technique
for quantitating DNA ranging from 25 pg/ml to 1 g/ml. Picogreen is a
cyanine dye that is essentially non-fluorescent when free in solution, but
exhibits a greater than 1000-fold enrichment in fluorescence upon binding
to DNA. The amount of DNA typically recovered in the Control IP is
approximately 20 ng. This amount should not vary drastically between in-
dependent ChIP experiments; however, the amount of DNA recovered
Fig. 1. The efficiency of MNase digestion and sonication. A representative example of

post-sonication aliquots taken during ChIP experiments performed in the presence of MNase
(þMNase) and without using MNase (MNase). The DNA collected from these aliquots
were electrophoresed into a 1% agarose gel, stained with ethidium bromide, and
photographed under UV radiation. Lane 1 contains 15 l of DNA from a 50-l þMNase
sample and lane 2 contains the same amount of DNA from a 50-l MNase sample. DNA
size markers in kilobases are indicated to the left of the gel.
in the IP sample, although enriched by at least a factor of 2 as compared to

the Control IP, can vary greatly depending on the antibody. For example,
the anti-acetyl histone H3 (Lys9 and 14) antibody (Upstate Biotechnology)
immunoprecipitates approximately 90 ng of DNA, while the anti-methyl
histone H3 (Lys9) antibody (Upstate Biotechnology) yields
approximately 150 ng of DNA from a typical ChIP experiment with ap-
proximately 200 million cells. Indeed, some antibodies that are used at
low concentrations immunoprecipitate much more DNA as compared to
other antibodies used at higher overall concentrations.
Real-Time PCR Primer Design

A majority of the antigen receptor loci of the mouse are sequenced and
available on-line. To design primers suitable for real-time PCR analysis
from these large chromosomal regions, we utilized the computer program
Oligo 4.0 (National Biosciences). Oligo 4.0 is designed to search for and
analyze oligonucleotides for PCR, sequencing, and hybridization applica-
tions. The software is very good at analyzing sequences for primer design,
calculating melting temperatures, and identifying problems, such as hairpins,
primer-dimers, temperature-mismatched pairs, etc. We follow four simple
rules when searching for primers to be used in real-time PCR analysis.
1. Avoid large discrepancies in melting temperatures between template
and primers. When the difference between template and primer

melting temperatures exceeds 35 , then the PCR yield will be less
than optimal.
2. Do not let the difference in melting temperatures between each primer

exceed 2 . Highly disparate melting temperatures between the
primers may cause the less stable primer to decrease the annealing
temperature, thus increasing the likelihood of non-specific priming.
3. Avoid primer dimers with highly negative free energies (DG) less than
5.0 kcal/mol. A DNA duplex is stable when the G is negative. If
the free energy of a primer dimer exceeds 5.0 kcal/mol, the
resulting PCR reaction is less than optimal.
4. Avoid primers with stable hairpin loops. Priming is extremely
inefficient if the hairpin loop melting temperature approaches the
annealing temperature.
Once primers have been designed to meet all aspects of the rules stated
earlier, the amplification potential of the primers should subsequently be
tested in PCR to ensure the accumulation of a single product of the appro-
priate size. Due to the highly repetitive nature of the murine IgH locus, the
accumulation of a specific product is not guaranteed even though all of
the specifications listed earlier for proper primer design have been fulfilled.
In a PCR tube add 10 pmol of each primer, 50 ng of sheared genomic DNA
(template), 100 M dNTPs, 1 PCR buffer, and 0.5 units HotStarTaq
DNA polymerase (Qiagen). Each reaction is subsequently spiked with
1 Ci 32P-dNTP, which dramatically increases the sensitivity of detecting
non-specific products and misprimings as compared to traditional ethidium
bromide staining. The PCR reactions can be run out on an acrylamide gel
and subjected to autoradiography. The ability of the primers to amplify a
single product of the appropriate size is exceedingly important for SYBR
Green real-time PCR as is discussed later. Another important feature of
the reaction is the intensity of the product as observed by autoradiography.

Through the course of evaluating hundreds of primer pairs in the linear
range of amplification we have noticed that even though a particular
primer pair amplifies a single product, the intensity may be exceedingly
low when compared to other products in the same linear range, directly
correlating with a low rate of amplification (discussed later), and thus
unable to generate useful real-time PCR data.
Real-Time Quantitative PCR Analysis

Real-time quantitative PCR analysis is a reliable method used to meas-
ure the accumulation of products generated during each cycle of the PCR
reaction.14,15 These products are directly proportional to the amount of
template DNA present in each reaction prior to the onset of amplification.
Real-time PCR has become very widely used in recent years, partially due
to the technological advances made in instrumentation and fluorescent re-
agents, which have caused market prices to drop, increasing the number of
researchers with access to this technology. This method of PCR analysis
has many applications including, but not limited to, measuring DNA copy
number and mRNA expression levels. Real-time PCR is appealing because
it is much more rapid and reliable than endpoint quantitation of PCR prod-
ucts as bands on a gel. Inhibitors of the polymerase chain reaction, inherent
to the reaction itself, include template DNA concentration, limiting PCR
reagents, and the accumulation of pyrophosphate molecules. All of these
conditions act in concert to slow the amplification process and eventually
cause the reaction to enter the plateau phase, in which the generation of
the template no longer proceeds at an exponential rate. Real-time PCR
is our method of choice for analyzing the DNA recovered from a ChIP ex-
periment because of its ability to measure PCR product formation in ‘‘real-
time,’’ at a point in which amplification is considered exponential. Only
during the exponential phase of the PCR reaction it is possible to extrapo-
late back to the starting material to determine the original amount of spe-
cific DNA template added. In addition, it is during the exponential
accumulation of product that a fluorescence signal is detected (a descrip-
tion of the most common fluorescent reagents used in real-time PCR analy-
sis is provided later). The signal corresponds to the accumulation of specific
PCR products and is immediately detected and recorded in real time by the
14
A. Giulietti, L. Overbergh, D. Valckx, B. Decallonne, R. Bouillon, and C. Mathieu,
Methods 25, 386 (2001).
15
D. G. Ginzinger, Exp. Hematol. 30, 503 (2002).
same instrument that regulates the steps involved in thermocycling (e.g.,

iCycler iQ, Bio-Rad; ABI7900, Applied Biosystems).
The three most commonly used fluorescent reagents for detecting PCR
amplification in real time are TaqMan probes, molecular beacons, and
SYBR Green I dye. A TaqMan probe is designed to hybridize to the tem-
plate amplified between two standard PCR primers. The probe is conju-
gated to a fluorescent dye at its 50 end and a quencher at the 30 end. The
fluorescent tag is quenched by fluorescence resonance energy transfer
(FRET) and does not emit a signal while the probe is intact. However,
during amplification the 50 fluorescent dye is cleaved and released by the
50 to 30 exonuclease activity of Taq polymerase. Once released from
the quenching ability of the intact probe, the fluorogenic dye is able to
emit its signal, which is recorded during the exponential phase of the reac-
tion. Molecular beacons are very similar to TaqMan probes in that they
utilize the ability of a 30 quencher to suppress the fluorescent signal nor-
mally emitted by a 50 dye by FRET. However, the quenching effect is
achieved by introducing a hairpin loop in the probe. When the beacon
hybridizes to a complementary nucleotide sequence, the template, the con-
formational change resulting from the hairpin loop structure being
converted to a linear DNA molecule provides enough distance between
the dye and quencher to allow a fluorescent signal to be emitted and
recorded. The third reagent used in real-time PCR is SYBR Green I dye,
which differs from the above-mentioned reagents in that there is no need
for a modified oligo or fluorogenic hybridization probe. The SYBR Green
I intercalating dye has an undetectable fluorescence when free in solution
but emits a signal on binding to dsDNA. We have chosen this dye because
it can be used in combination with any pair of primers, specific to any
desired target sequence, thus making it extremely versatile and the least
expensive choice. However, this dye cannot discriminate between
true DNA template and non-specific products such as primer dimers and
misprimings, unlike TaqMan and molecular beacons, which recognize
and emit a fluorogenic signal only when hybridized to the template DNA.
This is why primer design for SYBR Green–mediated real-time PCR is so
important.
In our studies, differences in the amount of specific DNA sequences en-
riched in the IP sample relative to the Input sample are determined by
SYBR Green real-time PCR analysis using the iCycler iQ from Bio-Rad.
Procedure and Calculations

1. Each PCR reaction for both the IP sample and the Input sample is
to be done in triplicate. The PCR cocktail consists of 2 ng of Input or IP
DNA, 10 pmol of each primer, and SYBR Green PCR Master Mix
(Applied Biosystems). Although a 50-l minimum reaction volume is
suggested for the Bio-Rad iCycler, we have consistently obtained
reproducible results with a 20-l reaction.
2. During amplification, the fluorescence signal intensity is plotted as a
function of the cycle number (see Fig. 2). After the reaction has finished, a
fluorescence signal cycle threshold (Ct) level must be set. The threshold
level should reside in the exponential phase of the real-time PCR reaction
and represents the point at which the Input sample can now be compared
to the IP sample. The number of cycles of each PCR reaction required to
reach the set threshold level is known as the Ct. Cycle threshold values are
directly proportional to the amount of DNA template present in each
reaction prior to amplification and provide the basis for calculating relative
fold enrichment values. The Ct values for each reaction within the
triplicate are averaged according to the following rule; if a Ct value within
the triplicate fluctuates greater than or less than 0.5 cycles in either
direction from the other two values in the triplicate (likely due to
pippetting errors), then this data point is not considered when calculating
the overall cycle threshold for that particular target sequence. Next, the
rate of amplification (R) should be calculated for each PCR reaction. Due
to the exponential nature of the PCR the theoretical optimum rate of
amplification is 2.0, but to make accurate and meaningful comparisons
between primer pairs, the differences in amplification rates should be
calculated. PCR primers that have been designed according to the rules
stated earlier typically have R values ranging between 1.85 and 2.0. These
values remain relatively constant for a particular primer pair, although on
rare occasions, variability in the rate is seen, which is likely due to
evaporation of a particular reaction from the 96-well PCR plate. The R
values for each reaction within a triplicate are averaged, except for those
individual R values that are less than 1.75 or greater than 2.25. If an R
value falls outside this bracket, then the data point is not used in the
calculation of overall primer amplification rate. The fold enrichment (see
mock calculation illustrated in Fig. 2) of a specific target sequence in the IP
sample relative to the Input sample is calculated using the following
equation:
Fold enrichment ¼ RðCtInput CtIP Þ :
Fig. 2. A graphic demonstration of the data generated by real-time PCR and the
calculations used in determining the enrichment of specific DNA sequences in an IP sample
relative to an Input sample. Fluorescence data generated in real time were plotted as a
function of the cycle number during PCR amplification of a particular target sequence from
representative IP and Input triplicate reactions. The cycle number is listed along the x-axis
and the fluorescence intensity is plotted on the y-axis. For this example, the Ct level was set to
a fluorescence of 80, which resides within the exponential phase of amplification. Extrapo-
lation of the fluorescence data from the set Ct level down to the x-axis provides Ct values for
both the IP and Input samples. Taking into account the rate of amplification (R), the fold
enrichment of a particular DNA sequence in the IP sample relative to the Input sample is
45.68.
Results and Conclusions

This ChIP protocol has an important key difference when compared to
the more widely used ChIP techniques reported in the literature. In ad-
dition to sonication, or mechanical shearing, this procedure has imple-
mented the use of MNase prior to the conventional sonication step. This
additional step increases the resolution of the ChIP method significantly
by reducing the number of large DNA fragments that may remain with so-
nication alone. This additional layer of resolution is exceedingly important
in determining the fold enrichment levels of certain histone modifications.
For example, due to the relatively high overall levels of methylated histone
H3 (Lys9) in mammalian genomes (it has been reported that approxi-
mately 20% of chicken histone H3 lysine 9 residues are methylated16),
16
J. R. Davie and V. A. Spencer, J. Cell Biochem. 32/33 (Suppl.), 141 (1999).
Fig. 3. The addition of MNase prior to sonication in the outlined ChIP method provides
the resolution needed to establish an enrichment pattern of the association of certain histone
modifications with genomic loci. (A) A schematic of the immunoglobulin heavy chain (IgH)
locus. Solid, black rectangles represent families of V segments, black vertical lines depict D
regions, and J gene segments are illustrated as hatched bars. The regulatory enhancer, E (E),
and the constant region, C3 (C), are depicted as a solid, black oval and a white rectangle,
respectively. Black dots located immediately above the schematic mark the location of each
PCR primer pair used in (B) and Fig. 4. (B) The enrichment of histone H3 (Lys9) methylation
across the IgH locus in Pro B cells as determined from multiple independent ChIP
experiments with and without the addition of MNase. Along the x-axis is the type of gene
segment amplified by each primer pair (V, D, J, E, and C). These segments are arranged from
left to right in the 50 to 30 orientation as they are located in the IgH locus. The fold enrichment
of each DNA sequence in an IP sample relative to an Input sample is shown.
Fig. 4. The enrichment of histone H3 (Lys9) methylation and H3 acetylation at lysines 9

and 14 across the IgH locus in Pro B cells. The particular amplified gene segment is provided
along the x-axis and the normalized fold enrichment values for both modifications are shown
on the y-axis. The data for each modification were normalized such that, among the gene
segments shown, the lowest fold enrichment value was set to 1 and the highest fold enrichment
value was set to 2.
the level of enrichment of this modification seen at specific DNA sequences

as calculated by real-time PCR analysis may appear slightly reduced. This
is exactly the case at the -globin locus in chicken cells, where enrichments
in H3 (Lys9) methylation rarely exceed twofold.5 Indeed, Fig. 3B empha-
sizes this point by illustrating that the conventional ChIP method, relying
only on sonication (MNase), generates an ambiguous enrichment pattern
of methylated histone H3 (Lys9) at the IgH locus in Pro B cells. However,
by combining sonication with MNase digestion (þMNase) a clear and
consistently reproducible pattern of enrichment is observed. Furthermore,
in the presence of MNase, the low level of H3 (Lys9) methylation seen
across the DJ region as well as the high level of methylation observed at
the upstream V segments correlates closely with the active and inactive
domains of the IgH locus in Pro B cells, respectively (discussed later). In
the absence of MNase, the measured levels of methylation across the IgH
locus are not coincident with any known aspects of the developmental
progression of Pro B cells, providing experimental support for the inclusion
of MNase in our ChIP method.
Due to the lack of the recombinase machinery, recombinase-deficient
Pro B cells are arrested just prior to the first stage of rearrangement,
D to J recombination. Thus, D and J segments of the IgH locus reside in
an active, accessible domain, while the V segments in these arrested cells
should remain repressed and inaccessible. Using the protocol outlined
here, we have defined these active and repressed long-range chromatin
domains by histone H3 acetylation and H3 (Lys9) methylation, respec-
tively (see Fig. 4). In conclusion, the ChIP method described in this chap-
ter provides another versatile tool, in addition to DNaseI hypersensitivity
mapping, which can be used in a number of unique cell types to identify
regulatory elements, such as promoters, enhancers, or boundary elements
as well as define the histone modification pattern across chromatin
sub-domains within a large, complex genomic locus.
Acknowledgments
We would like to acknowledge Joe Geisberg and other members of Kevin Struhl’s
laboratory for their help with the design of the ChIP protocol, PCR primers, and real-time
PCR analyses. We would also like to thank Laura Corey for her help with implementing
MNase into our protocol. We thank Mary Donohoe for her helpful discussions and critical
reading of the manuscript. This work was supported by NIH grant GM48026 to M.A.O.
[23] use of chromatin immunoprecipitation assays 349
[23] The Use of Chromatin Immunoprecipitation

Assays in Genome-Wide Analyses of
Histone Modifications
By Bradley E. Bernstein, Emily L. Humphrey,
Chih Long Liu, and Stuart L. Schreiber
Introduction
Systematic studies made possible by a variety of technological advances
are impacting many areas of biology. Of special relevance to the field of
chromatin are approaches that combine chromatin immunoprecipita-
tion (chromatin IP) with microarrays to analyze histone modifications
genome-wide in yeast.1,2 The resulting genomic maps provide a unique
global perspective on the functions of, and inter-relationships between,
different modifications.
Chromatin IP experiments use antibodies to immunoprecipitate a pro-
tein of interest and associated DNA from a solubilized chromatin prepar-
ation.3–7 Specialized antibodies can be used to enrich for DNA associated
with histones that exhibit specific post-translational modifications. A con-
siderable number of antibodies that recognize histones acetylated, methyl-
ated, or phosphorylated at specific residues have been developed, and
many are commercially available. Chromatin IP experiments typically
assess whether a particular gene, gene promoter, or genomic region is en-
riched in the IP sample relative to a whole cell extract (WCE) control. For
example, quantitative polymerase chain reaction (PCR) using pre-selected
primer pairs can be used to compare the representation of specific DNA
species in the IP and WCE. Although several regions can be queried simul-
taneously, these conventional studies are limited to a subset of genes or
regions chosen by the investigator.
Global studies have the potential to overcome selection bias by
analyzing comprehensively many or all elements in a system. Microarray
1
B. E. Bernstein, E. L. Humphrey, R. L. Erlich, R. Schneider, P. Bouman, J. S. Liu,
T. Kouzarides, and S. L. Schreiber, Proc. Natl. Acad. Sci. USA 99, 8695 (2002).
2
109, 437 (2002).
3
M. J. Solomon, P. L. Larsen, and A. Varshavsky, Cell 53, 937 (1988).
4
M. Braunstein, A. B. Rose, S. G. Holmes, C. D. Allis, and J. R. Broach, Genes Dev. 7,
592 (1993).
5
V. Orlando, H. Strutt, and R. Paro, Methods 11, 205 (1997).
6
7
A. Hecht and M. Grunstein, Methods Enzymol. 304, 399 (1999).

technology, typically used for mRNA analysis,8,9 can be adapted for the
comprehensive analysis of chromatin IP samples.10,11 These studies rely
on several innovations, including:
1. The development of molecular biology protocols to amplify and label
chromatin IP DNA samples in a sequence independent manner and
2. The production of DNA microarrays that contain the open reading
frames (ORFs) as well as the regulatory regions
Recently this approach has been used to map relative levels of histone
acetylation and methylation genome-wide in yeast, and to identify regions
deacetylated by the various histone deacetylase enzymes in yeast (see
Fig. 1).1,2
Protocols
The protocol for analyzing histone modifications genome-wide in yeast
is presented in four sections:
1. Chromatin IP to isolate DNA associated in vivo with modified
histones
2. DNA amplification and labeling
3. Microarray hybridization and data acquisition
4. Data analysis and interpretation
Chromatin IP to Isolate DNA Associated In Vivo with

Modified Histones
Solutions
YPD: Yeast extract/peptone/dextrose
PBS: Phosphate-buffered saline pH 7.4
Buffer L: 50 mM HEPES-KOH, pH 7.5, 140 mM NaCl, 1 mM
EDTA, 1% Triton X-100, 0.1% sodium deoxycholate.
For 5 ml of buffer L, 20 l of protease inhibitor cocktail
(Sigma), and one mini complete protease inhibitor tablet
(Boeringer Mannheim) are added just before use
8
J. L. DeRisi, V. R. Iyer, and P. O. Brown, Science 278, 680 (1997).
9
L. Wodicka, H. Dong, M. Mittmann, M. H. Ho, and D. J. Lockhart, Nat. Biotechnol. 15,
1359 (1997).
10
533 (2001).
11
J. Schreiber, N. Hannett, E. Kanin et al., Science 290, 2306 (2000).
Fig. 1. Steps in the genome-wide analysis of histone modifications in yeast.
Buffer W1: Buffer L with 500 mM NaCl

Buffer W2: 10 mM Tris-HCl, pH 8.0, 250 mM LiCl,
0.5% NP-40, 0.5% sodium deoxycholate,
1 mM EDTA
TE: 10 mM Tris-HCl, pH 8.0, 1 mM EDTA
Proteinase K solution: TE with 0.4 mg/ml glycogen, 1 mg/ml
proteinase K
Elution buffer: TE, pH 8.0, with 1% SDS, 150 mM NaCl,
and 5 mM DTT
DNA is isolated using the following chromatin IP protocol, adapted
from published protocols by Kuo and Allis,6 and Hecht and Grunstein.7

A 180-ml culture of yeast (Saccharomyces cerevisiae) is grown in YPD at 30
to an OD600 of 1.0. To cross-link proteins and DNA, 4.9 ml of 37% formal-
dehyde is added and the culture incubated at room temperature for 15 min
with occasional shaking. The formaldehyde is quenched by the addition of
9 ml of 2.5 M glycine, and the cells are left at room temperature for an ad-
ditional 5 min. The yeast are pelleted by centrifugation at 2000g for 5 min

at 4 , and washed twice with 180 ml ice-cold PBS. Washed cells are divided

into four aliquots, pelleted, frozen in liquid nitrogen, and stored at 80 .
To lyse cells, one aliquot (representing 45 ml of the original culture) is
resuspended in 400 l of buffer L. Five hundred microliters of acid-washed

glass beads are added and the mixture is vortexed on high for 45 min at 4 .
With the lid of the microfuge tube open, the bottom of the tube is punc-
tured with a hot 25-gauge needle and the cell lysate is spun to a new tube
(5 s full speed, cut off the top of the punctured tube before spinning). The
total volume of lysate is then adjusted to 700 l by addition of buffer
L. Chromatin is fragmented by sonication using a Branson 250 Sonifier
(4 20 s at setting 3 and 70% duty, with 20-s rests), and the insoluble

fraction precipitated by centrifugation at 14,000g for 15 min at 4 . The
supernatant contains the solubilized chromatin and is referred to as the
WCE. The size distribution of the chromatin fragments can be examined
by gel electrophoresis (see Fig. 2). A desired distribution is achieved by
varying sonication time.7
Prior to immunoprecipitation, 15 l of the WCE are set aside for a con-
trol. The remaining WCE is added to 50 l of packed protein A beads (pre-

equilibrated in buffer L) and rotated at 4 for 1 h to remove proteins that
bind nonspecifically to the beads. The supernatant is then transferred to an-
other tube that contains 20–30 l of anti-modified histone antibody and ro-

tated at 4 for 4 h. Antibody-histone complexes are precipitated by the
addition of 50 l of pre-equilibrated protein A beads and rotation at 4
for 1 h. The beads are collected by centrifugation at 3000g for 30 s and
washed successively with 1 ml of buffers L (twice), W1 (twice), W2 (twice),
and TE (twice) for 5 min each. The beads are then incubated in 125 l of

elution buffer at 65 for 10 min with frequent mixing. The beads are
pelleted by centrifugation at 10,000g for 2 min and the supernatant is
retained. After a second elution, the eluted samples are combined, overlaid

with mineral oil, and left at 65 overnight to reverse cross-links. In parallel,
15 l of WCE is combined with 235 l of elution buffer, overlaid, and

incubated at 65 overnight.
To isolate DNA, the IP sample and WCE control are incubated with an

equal volume of proteinase K solution at 37 for 2 h, extracted with 1
volume phenol (twice), 1 volume chloroform/isoamyl alcohol (25:1) (once),
and ethanol precipitated. The samples are centrifuged at 15,000g for
Fig. 2. WCE starting material and products amplified by R-PCR or IVT. Lanes 2–4 of a
2% non-denaturing agarose gel contain 200 ng nucleic acid. Lanes 1 and 5: 500 ng 100 bp
ladder (NEB). Lane 2: Yeast WCE. Lane 3: DNA amplified by R-PCR from WCE. Lane 4:
RNA amplified by IVT from WCE.

15 min at 4 , washed with cold 70% ethanol, dried, resuspended in 20 l of

TE with 10 g of RNase A, and incubated at 37 for 1 h. Prior to amplifi-
cation, the DNA is further purified with a MinElute PCR Purification
Kit (Qiagen).
DNA Amplification and Labeling

Chromatin IP experiments typically yield low nanogram quantities of
DNA. An amplification step is required to obtain the 2–4 g of material
necessary for DNA microarray analysis. Amplification of a heterogeneous
population of fragmented genomic DNA poses unique challenges. The
PCR requires a known flanking sequence for priming. Two PCR-based
protocols have been developed for the amplification of genomic DNA. In
‘‘ligation-mediated PCR,’’11 the genomic DNA is blunt-ended, and ligated
to double-stranded linkers. Ligated DNA is then amplified by PCR using
the linker sequence for priming. In random-primer PCR (R-PCR),10,12,13
12
S. K. Bohlander, R. Espinosa III, M. M. Le Beau, J. D. Rowley, and M. O. Diaz, Genomics
13, 1322 (1992).
13
J. D. Lieb, X. Liu, D. Botstein, and P. O. Brown, Nat. Genet. 28, 327 (2001).
the genomic DNA is copied twice using a primer adaptor that has a
conserved 50 sequence and a degenerate 30 sequence that anneals to the
ends of the genomic DNA. The copied DNA is amplified by PCR using
the 50 conserved sequence for priming. Both methods have been used to
amplify DNA obtained by chromatin IP.
Due to its exponential kinetics, PCR is particularly susceptible to
sequence- and length-dependent biases. Recently, a linear amplification
strategy has been applied to the amplification of genomic DNA.14 In this
procedure, a T7 promoter sequence is added to the ends of the genomic
DNA, creating a template suitable for in vitro transcription (IVT). Al-
though this IVT protocol has the added complication of working with
RNA, it preserves the species representation of the chromatin IP DNA
more effectively than the PCR methods. Here we present the R-PCR
and IVT amplification protocols as we have applied them in studies of
histone modifications.
R-PCR Amplification Protocol

Solutions
5 Sequenase 200 mM Tris, pH 7.5, 100 mM MgCl2,
buffer (Amersham): 250 mM NaCl
dNTP mix: 3 mM each of dATP, dCTP, dGTP, and
dTTP
Sequenase reaction 2.5 l 5 Sequenase buffer, 3.75 l dNTP
mix (for 2 reactions): mix 1.875 l 0.1 M DTT, 3.75 l 0.5 mg/ml
BSA, 0.75 l Sequenase 2.0 (Amersham)
Sequenase Dilution 10 mM Tris, pH 7.5, 5 mM DTT, 0.1 mM
buffer (Amersham): EDTA
50 aa-dNTP mix: 25 mM dATP, 25 mM dCTP, 25 mM dGTP,
10 mM dTTP, 15 mM aminoallyl-dUTP
PCR buffer: 500 mM KCl, 100 mM Tris, pH 8.3
Primer A: 40 pmol/l 50 -GTTTCCCAGTCACGAT-
CNNNNNNN- NN-30
Primer B: 100 pmol/l 50 -GTTTCCCAGTCA-
CGATC-30
DNA from the chromatin IP sample and the WCE control are amplified
in parallel using a variant of the Round A/Round B DNA amplification
protocol used by Iyer et al.10 and described at www.microarrays.org. This
variant protocol minimizes PCR rounds in an attempt to reduce amplifica-
tion bias. In Round A, 7 l sample, 2 l 5 Sequenase buffer, and 1 l
14
C. L. Liu, S. L. Schreiber, and B. E. Bernstein, BMC Genomics 4, 19 (2003).
Primer A are combined in a thin-walled PCR tube. In a thermal cycler, the

mixture is heated at 94 for 2 min, cooled to 10 , and held for 5 min while
5.05 l of Sequenase reaction mix is added. The reaction is ramped from

10 to 37 at a rate of 0.1 per second, held at 37 for 8 min, and then heated

at 94 for 2 min. It is then cooled to 10 and held for 5 min while 1.2 l of Se-
quenase 2.0 diluted 1:4 with Sequenase Dilution Buffer is added. Finally, the

reaction is ramped from 10 to 37 at a rate of 0.1 per second and held at 37
for 8 min. The samples are diluted with water to a final volume of 60 l.
In Round B, 15 l diluted Round A product, 8 l 25 mM MgCl2, 10 l
10 PCR buffer, 2 l 50 aa-dNTP solution, 1 l Primer B, 1 l Taq (5 U/
l), and 63 l nuclease-free water are combined in a thin-walled PCR tube.

The following PCR program is carried out for 30 cycles: 92 for 30 s, 40 for

30 s, 50 for 30 s, 72 for 1 min. The cycles are followed by a 5-min exten-

sion at 72 . PCR product is purified using a QIAquick PCR Purification Kit
(Qiagen). Amplification is verified by spectrophotometry and gel electro-
phoresis. The size distribution should approximate that of the WCE, though
we note that lower molecular weight species are inefficiently amplified by
this method (see Fig. 2). Products from 2–3 Round B PCR reactions are
combined to obtain the 8–10 g of DNA required for microarray analysis.
Equal quantities of DNA amplified from IP and WCE are used for
subsequent labeling and analysis.
IVT Amplification Protocol

Solutions
NEB buffer 3 100 mM NaCl, 50 mM Tris, pH 7.9,
(New England Biolabs): 10 mM MgCl2, 1 mM DTT
TdT buffer (Roche; note 1 M potassium cacodylate, 125 mM
different source than enzyme): Tris, pH 6.6, 1.25 mg/ml BSA
8% nucleotide tailing solution: 4.6 M dTTP, 0.4 M ddCTP
Eco Pol Buffer 10 mM Tris-HCl, pH 7.5, 5 mM
(New England Biolabs): MgCl2, 7.5 mM DTT
50 aa-dNTP mix: 25 mM dATP, 25 mM dCTP, 25 mM
dGTP, 10 mM dTTP, 15 mM ami-
noallyl-dUTP
Priming mix: 5 g/l 18–22-mer oligo dT (op-
tional), 5 g/l random primer pd
N6 (Invitrogen)
T7-A18B primer: 25 pmol/l 50 -GCATTAGCGGCC
GCGAAATTAATACGACT-
CACT ATAGGGAG(A) 18 [B]-3 0
([B] stands for any base other than A)
DNA from the chromatin IP sample and the WCE control are treated
with calf intestinal alkaline phosphatase (CIP) to remove 30 phosphate
groups prior to IVT. DNA (up to 50 ng), 1 l NEB Buffer 3, and 0.25 l
of 10 U/l CIP (NEB) are combined and H2O is added to a total
volume of 10 l. The reaction is incubated at 37 for 1 h, and cleaned up
with a MinElute Reaction Cleanup Kit (Qiagen). The kit is used after sev-
eral steps in this protocol, always according to manufacturer’s instructions
except with the elution volume increased to 20 l. Terminal transferase
(TdT) is used to add 30 tails of relatively uniform length to the ends of
the genomic DNA. Ten microliters of CIP-treated DNA, 4 l TdT Buffer,
3 l 5 mM CoCl2, 1 l 8% nucleotide solution, and 2 l TdT enzyme

(10 U/l) are combined, overlaid with mineral oil, and incubated at 37
for 20 min. The reaction is halted by the addition of 2 l 0.5 M EDTA,
pH 8.0, and subjected to the MinElute Reaction Cleanup.
T7 promoters are incorporated at the ends of the DNA fragments as
follows: 20 l tailing reaction product, 0.6 l T7-A18B primer, 5 l 10
EcoPol buffer, 2 l 5.0 mM dNTPs, and 20.4 l nuclease-free water are

combined. In a thermal cycler, the samples are incubated at 94 for 2 min,

ramped down at 1 per second to 35 , held at 35 for 2 min, ramped down

at 0.5 per second to 25 and held while 2 l 5 U/l Klenow (NEB) is
added. The reaction is then incubated at 37 for 90 min, halted by the ad-
dition of 5 l 0.5 M EDTA, pH 8, and subjected to a MinElute Reaction
Cleanup. (Note: mineral oil is not used at this stage.)
Prior to the IVT, the volume is reduced to 8 l in a vacuum centrifuge
at medium heat. IVT is carried out using the T7 Megascript Kit (Ambion).
The samples are purified using an RNeasy Mini Kit (Qiagen) per manufac-
turer’s RNA cleanup protocol, except with an additional 500-l wash with
buffer RPE. RNA is quantified by absorbance at 260 nm and visualized on
a 2% native TAE agarose gel. Yields range from 10 to 60 g. The size
distribution should approximate that of the WCE (see Fig. 2).
Equal quantities of RNA from the IP sample and WCE control (2–4
g) are reverse-transcribed to generate DNA probes. In a 14.5-l
volume, the RNA is primed by addition of 1 l priming mix, incubation

at 70 for 10 min, and incubation on ice for 10 min. Then, 6 l 5 RT
buffer, 3 l 0.1 M DTT, 3 l nuclease-free H2O, 0.8 l 50 aa-dNTP
mix, and 2.5 l 200 U/l Superscript II (Invitrogen) are added, and the
mix is incubated at 42 for 2 h. The reaction is halted by the addition of
10 l 0.5 M EDTA, pH 8.0, and the RNA is hydrolyzed by the addition

of 10 l 1 N NaOH and incubated at 65 for 15 min. The solution is neu-
tralized with 25 l 1 M HEPES, pH 7.5. The entire samples are used for
labeling and hybridization.
Probe Labeling
Amplified samples are fluorescently labeled using monofunctional
Cy5 and Cy3 fluorescent dyes as described by Carroll et al.,15 and at:
www.microarrays.org. These derivatized dyes covalently couple to the ami-
noallyl groups incorporated in Round B of R-PCR, or in the RT step
following IVT. DNA samples are diluted with 400 l of nuclease-free
water, concentrated to 70 l in a YM-30 microcon (Millipore) by centri-
fugation at 11,000g, re-diluted and re-concentrated. After a final dilution,
the samples are concentrated to a volume of 18–20 l (microcons are pre-
weighed, and final volume determined by weight). Samples are collected by
inverting the microcon into a new microfuge tube and spinning at 4000g for
5 min. One microliter of 1 M NaHCO3, pH 9, is added and the samples are
transferred to a microfuge tube containing a dried aliquot of NHS cyanine
dye (Cy5 for IP sample; Cy3 for WCE control) and incubated at room tem-
perature for 75 min in the dark. [The dye aliquots are prepared by dissolv-
ing one vial of monofunctional Cy5 or Cy3 dye (Amersham) in 20 l
DMSO, distributing the solution into 10 microfuge tubes, and drying in a
vacuum centrifuge under low heat.] Reactions are diluted with 70 l H2O
and purified using a QIAquick PCR Purification Kit.
Microarray Hybridization and Data Acquisition

Microarray production, hybridization, and data acquisition are carried
out using protocols developed by Patrick Brown and colleagues, available
on-line at www.microarrays.org and recently reviewed by Gasch.16
Comprehensive coverage of the yeast genome is achieved by hybridizing
amplified DNA to microarrays that contain all ORFs and all promoter-
containing intergenic regions (INTs). The INTs are designated in such a
way that each gene promoter can be assigned to a specific INT. DNA frag-
ments representing all ORFs and INTs can be amplified from genomic
DNA using primers from Research Genetics. The ORFs and INTs are
printed on separate poly-lysine–coated slides by pin transfer using a robotic
arrayer as described at www.microarrays.org, or using alternative protocols
described elsewhere.11,17 The inclusion of INT and ORF arrays in a global
chromatin analysis can enable an experimenter to discriminate between
promoters and coding regions of genes.1,13
15
A. S. Carroll, A. C. Bishop, J. L. DeRisi, K. M. Shokat, and E. K. O’Shea, Proc. Natl. Acad.
Sci. USA 98, 12578 (2001).
16
A. P. Gasch, Methods Enzymol. 350, 393 (2002).
17
C. E. Horak and M. Snyder, Methods Enzymol. 350, 469 (2002).
Hybridization
The competitive hybridization used in DNA microarray analyses allows
flexibility in determining which samples to compare. In this presentation of
the protocol, the relative enrichment of a particular modification is ana-
lyzed genome-wide by comparing a chromatin IP sample to its WCE con-
trol. However, it is also possible to query changes in modification patterns
resulting from gene deletion or small-molecule treatment by comparing
two chromatin IP samples obtained from different sources using the same
antibody.
Reagents and solutions:
WS1: 387 ml Milli-Q water, 12 ml 20 SSC, 1 ml 10% SDS
WS2: 1 ml of 20 SSC in 399 ml of Milli-Q water
The following is adapted from the hybridization procedure described
at www.microarrays.org. Cy5-labeled chromatin IP probe and Cy3-labeled
WCE probe are combined, and concentrated in a microcon YM-30 filter
until 50 l of solution remain on the filter. The sample is collected by
inverting the microcon into a new microfuge tube and spinning at 4000g
for 5 min. Six microliters of 20 SSC, 3 l of 10 g/l poly(A), and
0.96 l of 1 M HEPES, pH 7.0, are added and the solution is filtered with
a pre-wet Millipore 0.45- filter spun at 12,000 rpm for 2 min. Ten percent

SDS, 0.9 l, is added, the mixed probe is placed in a 100 heat block for
2 min, placed at room temperature for 10 min, and then applied to the mi-
croarray. Slides printed with ORFs and INTs are placed in Telechem hy-
bridization chambers with the arrays facing up. One clean lifterslip (Erie
Scientific) is placed on each array so that the white strips face down. Half
of the mixed probe (about 20 l) is applied to the ORF array and the other
half to the INT array. Each well of the hybridization chambers is filled

with 40 l of 3 SSC, the chambers are then sealed and placed in a 60
water bath for 12–15 h.
Four Wheaton glass tanks containing WS1 or WS2 are prepared. A
metal slide rack is placed in the first tank of each solution. The microarray
slides are removed from the hybridization chambers and turned upside
down in the first tank of WS1. Slides are tilted to allow the lifterslips to fall
off, placed in the slide rack, and plunged up and down 20 times. This
dunking is repeated in the second tank of WS1 and in the two WS2 tanks.
The slides are dried by centrifugation at 1000 rpm in a Beckman tabletop
centrifuge for 2 min.
Data Acquisition and Processing

Data acquisition and analysis are detailed in a variety of sources.16–19
Briefly, data are acquired by scanning the microarrays with a scanning laser
device (Axon Instruments) that simultaneously excites and measures fluor-
escence emission for the Cy5 and Cy3 dyes. Image analysis software
(Axon) is used to quantify fluorescent features and associate them with
the appropriate ORF or INT. The data output consists of Cy5 and Cy3 in-
tensities and deviations for each ORF or INT. The Cy5/Cy3 ratios reflect
the amount of DNA in the chromatin IP sample relative to the WCE con-
trol. The data are normalized so that the relative fluorescence signal is the
same for most spots.19 Alternatively, control samples spiked into the
probes can be used for normalization.16 To eliminate questionable meas-
urements, array features having a signal-to-noise ratio of less than a cut-
off value (e.g., 2.5) are removed from further analysis. Replicate data sets
can be merged using a weighted-averaging procedure.18
Data Analysis and Interpretation

Microarray experiments generate vast amounts of data, and it can be
challenging to convert these data into informative observations. Each his-
tone modification data set contains Cy5/Cy3 ratios for each microarray fea-
ture. These ratios reflect the degree to which histones associated with the
corresponding ORF or INT exhibit a particular modification. A high ratio
indicates a high level of modification. A ratio of around 1 indicates a modi-
fication level not significantly different from the genome average. A
low ratio suggests a low level of modification, with the caveat that some
genomic regions may be less available to IP. (To control for this potential
caveat, antibody against an invariant region of a histone can be used.) It is
possible to assign p-values that reflect the probability that the modification
level of histones associated with a particular ORF or INT is significantly
higher (or lower) than the genome average.
The data can be used as a screen to identify genes or gene promoters
associated with highly modified histones. These can be verified by conven-
tional chromatin IP and followed up by other biological means. Alterna-
tively, since the genomic location of each ORF and INT is known,
modification status can be visualized across the yeast genome using
chromosome mapping programs or plotted with respect to distance from
telomere ends. The continuity of a modification along adjacent genomic
18
C. J. Roberts, B. Nelson, M. J. Marton, R. Stoughton, M. R. Meyer, H. A. Bennett, Y. D. He,
H. Dai, W. L. Walker, T. R. Hughes et al., Science 287, 873 (2000).
19
J. Gollub, C. A. Ball, G. Binkley, J. Demeter, D. B. Finkelstein, J. M. Hebert, T. Hernandez-
Boussard, H. Jin, M. Kaloper, J. C. Matese et al., Nucleic Acids Res. 31, 94 (2003).
regions can be assessed using statistical methods, though such analyses

must account for DNA fragments in the chromatin IP experiment that span
adjacent ORFs and INTs.
Data sets that assess different modifications can be compared in several
ways. The Pearson correlation provides a single measure of similarity be-
tween the global patterns of two modifications (this can be visualized by
plotting pairwise relationships between the modifications). Alternatively,
lists of outliers (e.g., features with high ratios) can be collated and
compared. The significance of an observed overlap can be calculated using
a hypergeometric probability model.20 Two-dimensional hierarchical clus-
tering, commonly used in expression profiling, can be performed and
visualized using Cluster and Treeview.21 Clustering can be used to identify
sets of genes that exhibit similar modification patterns. These comparative
analyses can also be extended to other microarray-based data sets to iden-
tify functional relationships. This kind of cross-comparison was used to
identify associations between H3 Lys4 methylation and transcription,1
and between histone deacetylation and repression.1,2,22 Although these
analyses have revealed valuable insights into roles of histone modifications,
a realization of the full potential of the histone modification data sets
awaits implementation of more sophisticated analytical approaches.
Conclusions and Future Directions

Methods for analysis of histone modifications genome-wide in yeast
have been developed and applied toward a limited subset of site-specific
acetylation2 and methylation1 marks. Future studies will assess the poten-
tial of this systematic approach to provide more comprehensive insight into
the functions of, and inter-relationships between, the many other modifica-
tions that occur on histones in yeast. Furthermore, the development of mi-
croarrays able to query regulatory regions of the human genome23,24
should allow for global analyses of histone modifications in higher organ-
isms. The recent development of a method for linear amplification of
chromatin IP samples should facilitate this transition.
20
S. Tavazoie, J. D. Hughes, M. J. Campbell, R. J. Cho, and G. M. Church, Nat. Genet. 22, 281
(1999).
21
M. B. Eisen, P. T. Spellman, P. O. Brown, and D. Botstein, Proc. Natl. Acad. Sci. USA 95,
14863 (1998).
22
B. E. Bernstein, J. K. Tong, and S. L. Schreiber, Proc. Natl. Acad. Sci. USA 97, 13708
(2000).
23
A. S. Weinmann, P. S. Yan, M. J. Oberley, T. H. Huang, and P. J. Farnham, Genes Dev. 16,
235 (2002).
24
Genes Dev. 16, 245 (2002).
[24] sequential chromatin immunoprecipitation from animal tissues 361
[24] Sequential Chromatin Immunoprecipitation from

Animal Tissues
By Dina Chaya and Kenneth S. Zaret
The chromatin immunoprecipitation (ChIP) technique for identifying

transcription factors and histones bound to specific DNA sites in cells has
revolutionized the analysis of gene regulation. In this procedure, cells or
nuclei are exposed to formaldehyde, which crosslinks proteins bound to
each other and to DNA in native chromatin. The crosslinked chromatin
is isolated and fractionated, the proteins are immunoprecipitated and
assessed by a quantitative PCR reaction to determine the relative extent
to which the antigen is associated with the DNA. While the majority of
ChIP studies have employed populations of freely growing cells, such as
cultures of yeast or animal cell lines, many important questions in gene
regulation must be addressed with native tissues. First, it is not yet clear
that biologically relevant chromatin states, which for this context consists
of bound factors and modified histones, are accurately represented by cell
lines, although of course the latter provide excellent model systems for
study. Second, there are well-documented cases where the transcriptional
regulation of all known cell lines is markedly different from that of
the native tissue, such as for liver-specific genes in hepatic cell lines versus
native liver or cultured liver slices.1 Finally, native tissues provide the op-
portunity to investigate gene-regulatory transitions during changes in orga-
nismal physiology that are simply absent in cell lines, such as during
organogenesis or diabetes. This is particularly important now that advances
in the genetics of flies, zebra fish, frogs, and mice are such that the effects of
virtually any genetic perturbation can be tested directly in the animal.
Despite the technical obstacles regarding the handling of animal tissues
for ChIP assays, a growing number of laboratories have made advances in
the area.2,3 This chapter will present two different protocols for formalde-
hyde crosslinking of mouse tissue chromatin, followed by methods for
chromatin solubilization, purification, fragmentation, and immunoprecipi-
tation; some of these steps require special adaptations for animal tissues.
Although the methods presented here were developed for liver chromatin,
1
D. F. Clayton, A. L. Harrelson, and J. E. Darnell, Jr., Mol. Cell Biol. 5, 2623 (1985).
2
M. Parrizas et al., Mol. Cell Biol. 21, 3234 (2001).
3
D. Chaya, T. Hayamizu, M. Bustin, and K. S. Zaret, J. Biol. Chem. 24 (2001).

experience in our laboratory and others has shown that they represent a
good starting point for working with other tissue types.
Note that our method for chromatin fragmentation involves a partial
sonication step followed by a complete restriction digestion step. The re-
striction digest, with an enzyme that recognizes a frequently occurring
four-base sequence, allows the target DNA to be definitively trimmed to
a known length and therefore limits the final PCR analysis solely to the
defined sequence. This eliminates a concern arising when only sonication
is used, where the immunoprecipitated protein is crosslinked to target
sequences that are statistically but not precisely sized by the extent
of DNA fragmentation, and hence may not be bound at the DNA site of
interest.
Another novel aspect of our approach is that we developed a method to
release antigen from the initial immunoprecipitation to allow a second, se-
quential immunoprecipitation of the same sample. This allows the investi-
gator to evaluate whether two distinct proteins were bound to the same
DNA fragment within the starting population of molecules. The resolution
of this approach is enhanced by the aforementioned restriction digestion of
chromatin prior to the first immunoprecipitation.
Preparing Formaldehyde-Crosslinked Chromatin from Tissues

Here we present two methods for delivering a formaldehyde crosslink-
ing agent to nuclei within tissues. In the first, formaldehyde is delivered to
cells in the native tissue in situ; in the second, nuclei are isolated from
tissues and then treated with formaldehyde. Since it is well established that
even the more careful nuclear isolation procedures lead to loss of proteins
bound to target sequences within chromatin,4 the in situ formaldehyde
delivery method is preferred.
Formaldehyde crosslinking can alter protein epitope, and therefore it
is essential to empirically determine whether crosslinking treatment is suf-
ficient or excessive. We have found it useful to use a constant 1% formalde-
hyde concentration and vary the temperature and time of the crosslinking
reaction.
Liver Retrograde Perfusion and Chromatin Crosslinking

This procedure utilizes the existing vascular system within a tissue to
deliver formaldehyde to cells in situ. Although the procedure was adapted
from well-established perfusion protocols for intact livers, our laboratory
4
G. Pfeifer and A. D. Riggs, Genes Dev. 5, 1102 (1991).
has further adapted it for pancreas, using pancreatic ducts, and it should be
adaptable to other organ systems containing vessels or ducts. In this pro-
cedure, the animal is sacrificed and the liver is perfused with phosphate-
buffered saline (PBS) and heparin to flush out blood cells. The chromatin
is then crosslinked in vivo by perfusing formaldehyde into the liver. In
order to avoid the body temperature of the animal to drop too quickly
during these steps and thereby inhibit the effect of formaldehyde, the per-

fusion is carried out on a heat block at 30 with 30 solutions. Liver nuclei
are then crudely purified.
The day before the experiment, prepare buffers A and B (see Table I)
but without 2-mercaptoethanol, which is added just prior to use. Store

buffers at 4 overnight, along with the following which are kept in the cold
room: pipettes, beakers, 15-ml Corex tubes, rotor and adaptors, slides
and cover slips, tissue homogenization device (dounce), cheesecloth, and
funnel. For tissue dissection, prepare: ethanol, 4 pins, big and small sharp
scissors, big and small tweezers, string, cotton-tipped applications, catheter
TABLE I
Solutions for Tissue Perfusion Protocol
Final concentrations
Prepare from Buffer Buffer

stock (M) A B
HEPES pH 7.6 1 15 mM 15 mM
KCl 3 60 mM 60 mM
NaCl 4 15 mM 15 mM
EDTA 0.5 0.2 mM 0.1 mM
EGTA 0.5 0.5 mM 0.1 mM
Sucrose 2.5 0.34 M 2.1 M
2-Mercaptoethanol 2 0.15 mM 0.15 mM
Glycine 2.5 125 mM —

Formaldehyde [20% in water, stored at 20 (Fischer Scientific, ref. F79–500)]
Sonication buffer
Stock Final concentration
Tris, pH 8.1 1M 50 mM
EDTA 0.5 M 2 mM
N-Lauroylsarcosine 10% 0.5% (filtered through 0.2 m)
PMSF 50 mM 0.5 mM
Protease inhibitors 5 mg/ml Each 5 g/ml (leupeptin,
trypsin inhibitor, antipain)
(Becton Dickinson, Angiocath, ref. 381137, 20 GA, 1.88 IN, 1.1 48 mm), two
10-ml syringes, plastic wrap, paper towels, small beaker, heat block,

aluminum-wrapped cardboard, hybridization oven at 30 , and a timer.
Prepare a beaker with 6 ml solution A, keep on ice. Place the following
solutions used for perfusion and pre-warm in the 30 oven: PBS with 1%
formaldehyde, PBS with 50 ng/ml heparin. Fill respective syringes with
10 ml of each solution. Prepare the dissecting area by placing heat block
in the center of the lab bench with the aluminum-wrapped cardboard cover
on top. Layer plastic wrap and then a paper towel above the cardboard.
Have all the tools ready on the side, including 10-cm long pieces of string.
Sacrifice one mouse at a time and place it on its back on the heat block.
Pin down the four legs and cut open the abdomen posterior to anterior. Cut
through the diaphram, damaging the least possible tissue to avoid bleeding.
Gently push the intestines on the side of the animal with a cotton-tipped
applicator in order to have better access to the liver. Seal the superior vena
cava, anterior to the liver, with a double knot using small tweezers and
string. Insert the catheter in the inferior vena cava, just above the kidney
junction. Remove the needle partway and blood should flow back into
the catheter. Cut the portal vein with small scissors. Screw the syringe con-
taining the PBS/heparin onto the needle and start very slowly perfusing the
liver (10 ml total). All the lobes of the organ should rapidly blanch white.
You should also see the buffer and blood flow out of the liver through the
portal vein. If the animal is sick or too old (over 3 months), the perfusion
will be difficult. If the lobes do not fully blanch, massage gently the liver
with a cotton-tipped applicator and keep perfusing very slowly. Switch

to the formaldehyde-containing syringe that is sitting in the 30 oven and
retrograde perfuse 10 ml slowly. The perfusions should take about
5–10 min. Let the formaldehyde sit in the perfused liver for 5 min in the
animal while on the heat block at 30 . Dissect out the liver and place in a

beaker in the 30 oven for another 5 min. Place the liver in 6 ml buffer A on
ice. The cold as well as the glycine will stop the crosslinking reaction. Mince
the liver with small scissors and let sit on ice while perfusing other livers.
Repeat this retrograde perfusion procedure on one or two more animals.
In the cold room, pour minced livers into a cold douncing apparatus on
ice. Rinse beaker with 5 ml buffer A and add to dounce. Aspirate excess
buffer from dounce. Dounce mechanically 12 times. Check a sample under
a phase-contrast microscope for quantitative release of nuclei and cells and
very few lumps of tissue. Filter through 8 layers of cheesecloth prewet with
buffer A. Layer gently onto a 1.5-ml cushion of a 1:1 mixture of buffers A
and B in 15-ml Corex tubes (if the tissue lysate volume is 4–6 ml, use one
tube; if 6–12 ml, use 2 tubes). Centrifuge in fixed-angle rotor (e.g., Sorvall

SS-34) at 10,000 rpm for 10 min at 4 . Gently decant the supernatant.
Resuspend the pellets gently in 5–10 ml of solution A. Layer onto a 1.5-ml

cushion of 1:1 buffers A and B and centrifuge at 5000 rpm for 10 min at 4 .
Gently decant supernatant, resuspend pellet in 2–5 ml sonication buffer,
and follow the protocol for nuclear sonication.
Isolation of Nuclei from Liver and Chromatin Crosslinking

Nuclei are first crudely prepared from a mouse liver and then treated
with formaldehyde. The nuclear isolation protocol works for many other
tissues; the main parameter to adjust is the number of strokes in the tissue
homogenizer.
The day before the experiment, prepare buffers A, B, and C (see
Table II) but without 2-mercaptoethanol, spermine, or spermidine, which

are added just prior to use. Store at 4 overnight, along with the same ma-
terials described in the section earlier except for those required for perfu-
sion. The best way to obtain high-quality nuclei is to keep all tissues and
materials ready and cold before you begin, and to work fast. Prepare a
beaker with 12 ml of buffer A, another with a 1:1 mix of solutions A and
B, and another with 8 ml of solution A. Keep on ice.
Sacrifice one mouse 2–3 months old. Carefully dissect out the liver,
rinse in the 12 ml of buffer A, and place in the beaker with 8 ml buffer A
on ice. Use small sharp scissors to mince the liver into pieces of a few milli-
meters. Perform same operation on the next mouse. In the cold room, pour
TABLE II
Solutions for Isolating Nuclei
Final concentrations
Prepare from stock (M) Buffer A Buffer B Buffer C
HEPES, pH 7.6 1 15 mM 15 mM 15 mM
KCl 3 60 mM 60 mM 60 mM
NaCl 4 15 mM 15 mM 15 mM
EDTA 0.5 0.2 mM — —
EGTA 0.5 0.5 mM — —
Sucrose 2.5 0.34 M 2.1 M 0.34 M
2-Mercaptoethanol 2 0.15 mM 0.15 mM 0.15 mM
Spermine 0.5 0.15 mM — —
Spermidine 0.5 0.5 mM — —
MgCl2 1 — — 2 mM
Glycine 2.5 (kept at room temperature)

Formaldehyde 20% in water, stored at 20
minced livers into the cold dounce on ice. Rinse the beaker with 5 ml solu-
tion A and add to dounce, to get all tissue fragments. After the tissue
settles, aspirate excess supernatant from the dounce. Dounce 10 times with
a motor-driven pestle. Check under the microscope that you have nuclei
and cells and very few lumps of tissue. Dounce further accordingly. Filter
through 8 layers of cheesecloth prewet with solution A. Layer gently onto
a 1.5-ml cushion of a 1:1 mixture of buffers A and B in 15-ml Corex tubes
(if the tissue lysate volume is 4–6 ml, use one tube; if 6–12 ml, use 2 tubes).
Centrifuge in fixed-angle rotor (e.g., Sorvall SS-34) at 10,000 rpm for

20 min at 4 . Gently decant the supernatant. Resuspend the pellets gently
in 5–10 ml of solution C. Measure the volume. Add formaldehyde (20%) to

a final concentration of 1% and incubate for 10 min at 30 . Mix gently
every 2 or 3 min. Stop the crosslinking reaction by adding room tempera-
ture glycine to a final concentration of 125 mM and let incubate on ice for
5 min. Layer onto a 1.5-ml cushion of a 1:1 mixture of solutions A and B.

Centrifuge in fixed-angle rotor at 5000 rpm for 10 min at 4 . Gently decant
the supernatant, resuspend the pellet into 2–5 ml sonication buffer, and
follow the protocol for nuclear sonication.
Sonication and Purification of Crosslinked Chromatin

from Tissues
The crosslinked nuclei are sonicated partially in order to solubilize the
chromatin and facilitate the later step of restriction enzyme digestion. The
sonication product is then loaded on a CsCl step gradient and ultracentri-
fuged to separate the crosslinked chromatin from the cell debris. This puri-
fication step is crucial for the restriction digest and for obtaining a good
signal to noise ratio in the immunoprecipitation (IP). After the spin, frac-
tions are collected from the gradient and each is assayed for the presence of
DNA. DNA containing fractions are pooled and the chromatin is dialyzed
to TE to eliminate the CsCl and allow a treatment with RNase, which we
found necessary for quantitative restriction digestion.
Nuclear Sonication
These conditions are optimized for chromatin that has been crosslinked
within tissue and should be reduced for chromatin that has been cross-
linked within isolated nuclei. Nuclear pellets that were resuspended into
2–5 ml of sonication buffer should be incubated for 5 min at room tem-
perature, then 5 min on ice. Sonicate with a Branson Sonifier 250, mounted
with a microtip, using duty cycle constant, output #2. Pulse alternatively for
30 and 15 s, incubating 30 s on ice between pulses. Avoid bubbles and be
careful not to heat the sample during the sonication, as the heat will reverse
the crosslinking in presence of Sarkosine.
CsCl Step Gradient

Prepare CsCl step gradient solutions of 1.35 g/ml, 1.5 g/ml, and 1.75 g/ml
CsCl in 0.5% Sarcosine, 1 mM EDTA. If CsCl does not solubilize, heat the
solutions. Keep all solutions and materials at room temperature. Prepare
the step gradient in ultraclear centrifugation tubes for a SW50.1 rotor.
Gently layer 1 ml of 1.75 g/ml, then 1 ml of 1.5 g/ml, and finally the
1.35 g/ml CsCl solution to near the top of the tube. Gently layer the sonica-
tion product on top of the step gradient. Do not overload the gradient; use
at least 2 tubes for each mouse liver, hence loading no more than about
1 mg of DNA per tube. Ultracentrifuge in SW50.1 rotor at 31,000 rpm at

20 for 20 h. To collect fractions, puncture a hole in the bottom of the cen-
trifuge tube with a 25G needle and let the CsCl gradient drip into 1.5-ml
tubes. Collect 5–10 drops per fraction, 500 l each, for about 10 fractions.
Crosslinked chromatin has a density equal to 1.39 g/ml CsCl and will hence
be closer to the top of the gradient. Repeat for each gradient tube.
To determine which fractions contain the crosslinked chromatin,

remove 10-l samples from each, incubate at 68 for 2 h to reverse the

crosslinking, then add 1 l of 100 ng/l RNase and incubate 37 for 30 min.
min. Add 3 l of 1% SDS (final 0.25%) to denature the protein, add elec-
trophoretic loading buffer, and analyze the DNA by agarose gel electro-
phoreses. The samples will still contain RNA at this stage; however, the
desired fractions will yield DNA as a faint smear of higher molecular
weight, ethidium bromide–staining material. Pool the gradient fractions
containing crosslinked chromatin and measure the volume. Generally,
we pool fractions 5–8; although fractions 9 and 10 contain DNA, they are
typically less clean as seen by adherence of staining material in the wells
of the analytical gel.
Tissue Chromatin Dialysis and RNase Treatment

At this stage, the crosslinked chromatin has been purified from the cell
debris. For further analysis of the crosslinked chromatin, RNA and CsCl
should be eliminated. Measure the volume of the purified crosslinked chro-
matin and place the material in a dialysis bag (3500 M.W. cutoff) that has
been rinsed with water. Dialyze against 500 ml of 10 mM Tris (pH 7.5),

1 mM EDTA (TE buffer) 5 min at 37 . Add RNase A (10 mg/ml stock)
to a final concentration of 25 ng/l directly into the dialysis bag, incubate

37 for 20 min. Change dialysis buffer to fresh TE and dialyze overnight

at 4 . Collect chromatin, dispense into 0.5–1 ml aliquots, and store at 80 .
For subsequent quantitative analysis, it is useful to remove an aliquot of

200 l or so at this stage for DNA purification and quantification. Knowing
the concentration of DNA will help in establishing conditions for restric-
tion enzyme digestion and immunoprecipitation. The concentration of
DNA of the purified crosslinked chromatin obtained by perfusion is usually
about 0.5 mg/ml. For each immunoprecipitation condition, 50–100 g of
chromatin DNA are used.
Tissue Chromatin Fragmentation by Restriction Enzyme Digestion

Restriction enzyme digestion of the chromatin will be highly sensitive
to the extent of crosslinking. A control digest reaction can be performed
on DNA samples where the crosslinks have been reversed first, in cases
where it is suspected that protein crosslinked to the DNA is inhibiting di-
gestion, rather than the digestion reaction conditions. Typically, digestion
conditions on crosslinked, partially sonicated chromatin are as indicated
by the enzyme manufacturer. For 500 ng/l chromatin, 2 ml total volume,
add 1.5 ml H2O, 400 l of 10 digestion buffer buffer, 500 units of enzyme,

and incubate at 37 for 3 h. The concentration of the purified crosslinked
chromatin after the restriction enzyme digestion is about 250 ng/l and
can be stored in aliquots at 20 .
To ascertain complete digestion, reverse the crosslinks on aliquots of
control and enzyme-treated sonicated, crosslinked chromatin. Run the
products on an agarose minigel and perform Southern blot analysis with
a 32-P–labeled, PCR-generated probe of the region of interest. The hybrid-
ization signal should appear as a smear in the undigested control and as a
unique, appropriately sized band in the enzyme treated sample (see Fig. 2C
of Chaya et al. (2001)3).
Sequential Immunoprecipitation of Purified, Restriction-Digested

Chromatin from Tissues
Sequential immunoprecipitation requires that the primary antibody be
immobilized, so that antigen complexes can be released and subjected to a
secondary immunoprecipitation. Methods are described later. Excellent
references exist for working out conditions for immunoprecipitation.5
Conditions described here were optimized for rabbit antisera against
the transcription factor HNF3 .
5
E. Harlow and D. Lane, in ‘‘Antibodies: A Laboratory Manual.’’ Cold Spring Harbor
Laboratory, NY, 1988.
Bead Preparation and Preclearing

This protocol describes the preparation of 1.5 g of Sepharose CL-4B
beads containing protein A (Sigma). Make all solutions fresh and keep so-

lutions and beads at 4 . Add 10 ml PAS (see Table III) to 1.5 g protein A-
Sepharose beads in a 50-ml plastic tube. Rinse bottle containing original
beads twice with 10 ml PAS and add to 50-ml tube. Incubate on ice 30 min,
occassionally inverting the tube gently. Centrifuge at 1500 rpm for 5 min

at 4 in a swinging bucket rotor. Aspirate supernatant and resuspend in
30 ml cold BSB by inverting the tube gently a few times. Centrifuge at

1500 rpm for 5 min at 4 in a swinging bucket rotor and then aspirate
supernatant. Repeat this wash for a total of 5 times, and then resuspend
the beads in 1 volume of BSB containing protease inhibitors. Aliquot and

store at 4 .
Before using the protein A beads for immunoprecipitation, nonspecific
DNA binding sites need to be blocked. Centrifuge an aliquot of beads at
2500 rpm for 1 min in a microcentrifuge. Discard the supernatant, add 1
volume RIPA buffer and sonicated salmon sperm DNA to a final concen-

tration of 100 g/ml, and incubate overnight at 4 while rotating. Centri-
fuge the beads at 2500 rpm for 1 min in a microcentrifuge, discard the
supernatant, and add 1 volume RIPA buffer. Beads at 1:1 ratio in RIPA
buffer are ready for immunoprecipitation.
Primary Antibody Crosslinking to Protein A-Sepharose beads

Incubate 500 l of protein A-Sepharose beads in 1:1 RIPA buffer with
75 l of specific antiserum or preimmune serum (120–150 mg/ml) in a 12-
ml polypropylene centrifuge tube. Incubate 1 h at room temperature while
rotating. Add 10 volumes (5 ml) of 0.2 M sodium borate, pH 9. Incubate
10 min at room temperature while rotating. Centrifuge at 3000 rpm for
1 min in swinging bucket rotor. Decant and repeat wash. Resuspend beads
in 5 ml of 0.2 M sodium borate, pH 9. Remove a 20-l aliquot for subse-
quent SDS-PAGE analysis (see later). Add fresh dimethyl pimelimidate
(solid) to a final concentration of 20 mM and incubate at room tempera-
ture for 30 min while rotating. Remove a 20-l aliquot for subsequent
SDS-PAGE analysis. Stop the reaction by washing with 5 ml of 0.2 M etha-
nolamine, pH 8. Centrifuge at 3000 rpm for 1 min in swinging bucket rotor.
Discard supernatant, add 5 ml of 0.2 M ethanolamine, pH 8, and incubate
2 h at room temperature while rotating. Centrifuge at 3000 rpm for 1 min
in swinging bucket rotor. Resuspend beads in 250 l RIPA buffer. Store at

4 . Analyze the pre- and post-crosslinked samples on a 10% SDS-PAGE
gel and stain with Coomassie. The crosslinking reaction should have
prevented SDS from solubilizing antibody protein from the beads.
TABLE III
Solutions for Preparing Antibody Covalently Bound to Protein A Beads,
Chromatin Immunoprecipitation, and DNA Purification
Final concentration Amounts
PAS (30 ml)

BSA 0.3 mg/ml 900 l of 10 mg/ml stock
Methionine 0.1 M 0.45 g
(Sigma, l-methionine)
SDS 0.01% 15 l to 30 ml of 20% stock
H2O
BSB (200 ml)
NaCl 150 mM 7.5 ml of 4 M stock
Borate adjusted 10 mM 20 ml to 122.5 ml of 0.1 M stock
to pH 8 (boric acid) H2O
Protease Inhibitors, PI
(added to 20 ml of BSB)
Leupeptin, trypsin inhibitor, 5 g/ml 20 l stock 5 mg/ml
antipain mix
Benzamidine 1 mM 20 l stock 1 M
PMSF 0.5 mM 200 l stock 50 mM
RIPA buffer Stock Final concentration
Tris, pH 8 1M 50 mM
NaCl 4M 150 mM (or 500 mM for I.P. washes)
IGEPAL CA-630 100% 1% (ICN)
Deoxycholic acid (Na salt) 10% 0.5% (Sigma)
SDS 20% 0.1%
RIPA buffer 2 without Tris
NaCl 4M 300 mM
IGEPAL CA-630 100% 2% (ICN)
Deoxycholic acid (Na salt) 10% 1% (Sigma)
SDS 20% 0.2%
TNESK 5X
Tris, pH 7.5 1M 50 mM
NaCl 4M 500 mM
EDTA 0.5 M 5 mM
SDS 20% 5%
Proteinase K 20 mg/ml 0.5 mg/ml (added just before use)
For each immunoprecipitation, use about 50–100 g of chromatin
DNA. Dialyze the total amount of purified crosslinked chromatin over-

night at 4 in 3500 cut-off dialysis bags against 500 ml RIPA buffer (with-
out IGEPAL). The next morning, dialyze a few hours against RIPA buffer
containing IGEPAL, then transfer to a polypropylene tube. To block non-
specific sites on the chromatin, add 2 l preimmune serum per milligram
chromatin and 200 l of 1:1 precleared protein A:RIPA. Incubate rotating
for 3 h at 4 . Centrifuge at 2500 rpm for 1 min at 4 and save supernatant
in fresh tube. Remove an aliquot of chromatin as an ‘‘input’’ sample.
Dispense 50–100 g aliquots of supernatant chromatin and add relevant
antibodies. Examples of experimental conditions include: no antibodies,
2 l nonimmune serum (stock 130 mg/ml), 2 l specific antiserum, 40 l
preimmune serum coupled to protein A, 40 l antiserum coupled to pro-

tein A. Incubate on rotator overnight at 4 . Add 40 l 1:1 precleared pro-
tein A:RIPA for samples that are not coupled to protein A and incubate
rotating for 2 h at 4 . Centrifuge all samples at 2500 rpm for 1 min at 4 .
Save supernatant for DNA analysis. Wash beads 4 times with 20 volumes
RIPA buffer containing 500 mM NaCl; for each wash, add buffer, let sit
on ice for 10 min, mixing occasionally, centrifuge, discard supernatant.
For final wash, transfer suspended beads to a fresh tube to eliminate chro-
matin bound nonspecifically to the plastic. For a single immunoprecipita-
tion experiment, beads are finally resuspended in 1 volume of TE. For
sequential immunoprecipitation, chromatin will be eluted from the beads,
as described in the following.
Releasing Primary Chromatin Antigens from Protein

A-Antibody Beads
Add to beads 5 volumes of phosphoric acid and incubate at room
temperature for 5 min, gently mixing occasionally. Centrifuge at 2500 rpm
rpm for 1 min. Transfer supernatant to fresh tube containing 1 M Tris,
pH 8 to neutralize. Add 1 volume RIPA buffer 2 without Tris. Centrifuge
at 2500 rpm for 1 min and carefully collect supernatants without
beads into fresh tubes. Divide chromatin samples for the second, sequential
immunoprecipitations. Follow protocol ‘‘D.3’’ for immunoprecipitation.
DNA Purification from Chromatin Immunoprecipitates

Reverse crosslinking for all samples to be analyzed by incubating for

at 68 for 3 h. Digest proteins by adding 5 TNESK to a final of 1 and

incubating overnight at 37 . Extract proteins with one phenol extraction
followed by one chloroform extraction. For maximal recovery of DNA,

back-extract the phenol phase with TE buffer. Add 5 g of glycogen (stock
20 mg/ml) as a carrier to facilitate the precipitation of small amounts of
DNA. Add 0.1 volumes of 3 M Na acetate and 2.5 volumes of ethanol to
each sample and incubate at 20 overnight. Centrifuge for 5 min, decant
supernatant, gently dry and resuspend pellets in TE buffer. The DNA is
ready for analysis by PCR.
Acknowledgment
Work on this project was supported by a grant from the NIH (GM47903).
[25] Immunofluorescent Staining of Polytene

Chromosomes: Exploiting Genetic Tools
By Gena E. Stephens, Carolyn A. Craig, Yuhong Li,
Lori L. Wallrath, and Sarah C. R. Elgin
Immunostaining of Drosophila polytene chromosomes is a powerful
tool for investigating the components of chromatin on a genome-wide
scale. Techniques for immunostaining were developed in the 1970s and
have evolved since their introduction.1–4
Many of the nuclei of Drosophila undergo multiple rounds of DNA
replication without cell division during the larval stages of development,
a process known as endoreduplication. While the euchromatic regions are
copied ca. 10 times, the pericentric heterochromatin undergoes only a few
rounds of replication, and centromeric satellite DNA and the Y chromo-
some are not amplified due to their heterochromatic nature. The replicated
chromosome arms remain tightly aligned and the heterochromatic regions
fuse in a common chromocenter. The amplification of the DNA provides
additional substrate for binding proteins and allows one to obtain a map
of protein distribution on polytene chromosomes with the use of an anti-
body specific for a given protein. The largest polytene chromosomes, the
product of 10 rounds of replication, are found in the salivary glands of
1
L. M. Silver and S. C. R. Elgin, Proc. Natl. Acad. Sci. USA 73, 423 (1976).
2
C. Rodriguez-Alfageme, G. T. Rudkin, and L. H. Cohen, Proc. Natl. Acad. Sci. USA 73,
2038 (1976).
3
M. Jamrich, A. L. Greenleaf, and E. K. F. Bautz, Proc. Natl. Acad. Sci. USA 74, 2079
(1977).
4
B. E. Schwartz, J. K. Werner, and J. T. Lis, Meth. Enzymol. 376, 393 (2004).

followed by one chloroform extraction. For maximal recovery of DNA,

back-extract the phenol phase with TE buffer. Add 5 g of glycogen (stock
20 mg/ml) as a carrier to facilitate the precipitation of small amounts of
DNA. Add 0.1 volumes of 3 M Na acetate and 2.5 volumes of ethanol to
each sample and incubate at 20 overnight. Centrifuge for 5 min, decant
supernatant, gently dry and resuspend pellets in TE buffer. The DNA is
ready for analysis by PCR.
Acknowledgment
Work on this project was supported by a grant from the NIH (GM47903).
[25] Immunofluorescent Staining of Polytene

Chromosomes: Exploiting Genetic Tools
By Gena E. Stephens, Carolyn A. Craig, Yuhong Li,
Lori L. Wallrath, and Sarah C. R. Elgin
Immunostaining of Drosophila polytene chromosomes is a powerful
tool for investigating the components of chromatin on a genome-wide
scale. Techniques for immunostaining were developed in the 1970s and
have evolved since their introduction.1–4
Many of the nuclei of Drosophila undergo multiple rounds of DNA
replication without cell division during the larval stages of development,
a process known as endoreduplication. While the euchromatic regions are
copied ca. 10 times, the pericentric heterochromatin undergoes only a few
rounds of replication, and centromeric satellite DNA and the Y chromo-
some are not amplified due to their heterochromatic nature. The replicated
chromosome arms remain tightly aligned and the heterochromatic regions
fuse in a common chromocenter. The amplification of the DNA provides
additional substrate for binding proteins and allows one to obtain a map
of protein distribution on polytene chromosomes with the use of an anti-
body specific for a given protein. The largest polytene chromosomes, the
product of 10 rounds of replication, are found in the salivary glands of
1
L. M. Silver and S. C. R. Elgin, Proc. Natl. Acad. Sci. USA 73, 423 (1976).
2
C. Rodriguez-Alfageme, G. T. Rudkin, and L. H. Cohen, Proc. Natl. Acad. Sci. USA 73,
2038 (1976).
3
M. Jamrich, A. L. Greenleaf, and E. K. F. Bautz, Proc. Natl. Acad. Sci. USA 74, 2079
(1977).
4
B. E. Schwartz, J. K. Werner, and J. T. Lis, Meth. Enzymol. 376, 393 (2004).

[25] immunofluorescent staining of polytene chromosomes 373
the third instar larvae (see also Schwartz et al.4 for more information on
endoreduplication).
Differential packaging of the DNA along each chromosome arm results
in a distinct banding pattern. Some of the constrictions represent sites of
underreplication, but in most cases, the bands and the interbands differ
only in their compaction of the DNA and not in their level of polyteny.5
In 1935, Bridges drew a detailed map of each chromosome arm; more re-
cently photographic representations have been made.6 A physical map
has been established through the technique of in situ hybridization. This
technique has allowed a specific band or region within a band to be cor-
related with the location of a specific gene or repetitive sequence at a
resolution of ca. 30 kb.7
Immunological methods for studying the association of proteins with
polytene chromosomes have been used to address a variety of biological
questions. In wild-type flies, immunostaining has been used to determine
the global distribution of one or more proteins; colocalization studies have
been done to determine if a protein of interest might be in close association
or part of a multiprotein complex with other proteins.8 Immunological
staining has also been performed on a variety of Drosophila species to
monitor conservation of a chromosomal protein and detect changes in
its distribution through evolution. Much has also been gained by taking
advantage of various genetic tools. Fly lines with chromosome rearrange-
ments have been used to show when the localization of a protein reflects
local as opposed to global structural features, for example, proximity to
the chromocenter.9 Mutations in chromosomal proteins can be assessed
by immunofluorescent staining to monitor both the extent of DNA binding
and distribution of the mutant protein, and the impact on chromosome
organization and distribution of other proteins.10 Finally, Drosophila
melanogaster can be transformed using a transposable element (usually a
P-element) carrying a cloned DNA fragment, allowing experiments to
examine the protein complexes that associate with a given construct.
5
A. Sperier and P. Sperier, Nature (London) 307, 176 (1984).
6
G. Lefevre, in ‘‘The Genetics and Biology of Drosophila’’ (M. Ashburner and E. Novitski,
eds.), 1a, p. 31. Academic Press, London, 1976.
7
J. Gall and M. L. Pardue, Proc. Natl. Acad. Sci. USA 63, 378 (1969).
8
C. D. Shaffer, G. E. Stephens, B. A. Thompson, L. Funches, J. A. Bernat, C. A. Craig, and
S. C. R. Elgin, Proc. Natl. Acad. Sci. USA 99, 14332 (2002).
9
T. C. James, J. C. Eissenberg, C. Craig, V. Dietrich, A. Hobson, and S. C. R. Elgin, Eur. J.
Cell Biol. 50, 170 (1989).
10
G. Schotta, A. Ebert, V. Krauss, A. Fischer, J. Hoffman, S. Rea, T. Jenuwein, R. Dorn, and
G. Reuter, EMBO J. 21, 1121 (2002).
Such assemblies can be inherited through mitosis,11 and sometimes through

meiosis, an event termed ‘‘cellular memory.’’
Squashing and Staining Protocols

The technique presented here is used for determining the in situ distri-
bution of chromosomal proteins in polytene chromosomes using specific
antibodies. We will focus on a procedure using formaldehyde fixation to
crosslink the proteins to the DNA with subsequent squashing in acetic
acid.1,12 Other labs have published similar techniques.2–4 The procedure
to be described produces polytene chromosome spreads of good morph-
ology and preserves most bound proteins in an immunologically reactive
state. Antibodies specific for a chromosomal protein are applied to poly-
tene chromosomes, and a fluorescent or enzyme-linked secondary antibody
that is directed against the primary antibody is then incubated on the
chromosomes. The chromosomes are viewed and photographed with a
fluorescence microscope and digital camera. This technique allows for re-
producible results in determining the localization of proteins on polytene
chromosomes to the resolution of individual bands (ca. 30 kb).
Fixation of Polytene Chromosomes

Third instar larvae that have been grown with minimal crowding are the
best source of polytene chromosomes. Choose the fattest larvae that have
just crawled out of the food and up the container wall. A visual montage of
the dissection of salivary glands from third instar larvae is shown in Fig. 1.
Carefully remove larvae from the bottle with dissection forceps and wash
in Cohen buffer13 in microwells. Place the larva in a microwell with fresh
Cohen buffer and using two pairs of dissection forceps, pull apart the larva
in one motion, grasping the mouth hooks at the anterior of the larva with
one pair of forceps and the posterior of the larva with the other pair of for-
ceps. The salivary glands will most likely be removed along with the brain,
eye-antennal imaginal discs, and fat tissue as one mass. Then, dissect the
salivary glands away from the fat body and any other extraneous tissue,
while allowing the glands to remain attached to each other through their
common duct. This duct is a convenient place to grasp the glands while re-
moving any extraneous tissue and while transferring the glands to new so-
lutions. Be careful not to damage the glands while removing any tissues. If
11
G. Cavalli and R. Paro, Science 286, 955 (1999).
12
L. M. Silver, C. E. C. Wu, and S. C. R. Elgin, in ‘‘Methods in Chromosomal Protein Research’’
(G. Stein, J. Stein and L. Kleinsmith, eds.), p. 151. Academic Press, New York, 1977.
13
L. H. Cohen and B. V. Gotchel, J. Biol. Chem. 246, 1841 (1971).
[25]
immunofluorescent staining of polytene chromosomes
375
Fig. 1. Diagram of isolating, squashing, and staining salivary gland polytene chromosomes. Reprinted from: Methods
Cell Biol. 35, 203 (1991), with permission from Academic Press.
the glands are damaged, the morphology of the chromosomes will suffer. If
the adhering fat body cannot be removed at this step without damaging the
glands, it can be removed later during acetic acid fixation. The paired saliv-
ary glands will either appear slightly milky or clear and are somewhat
wishbone-shaped. Incubate the excised glands in Cohen buffer for 8–
10 min. Incubation of glands in this solution with detergent allows for the
dissolution of cytoplasmic membrane structures. Poor morphology of the
chromosomes will result if an 8- to 10-min incubation is exceeded. Next,
the glands are incubated for 3–25 min in a formaldehyde fixative. Shorter
incubations can be used but may result in incomplete fixation and hence
loss of some chromosomal proteins, while longer incubations result in
chromosomes that are difficult to spread out well. The glands are then
transferred to 45% acetic acid and incubated for 3–60 min. During this in-
cubation, any adhering fat body may be removed. This acetic acid fixation
is necessary to attain good phase morphology of the spread polytene
chromosomes. The phase morphology will suffer if the glands are left
in the acetic acid for more than 1 h. This formaldehyde fixation
technique minimizes the extraction of chromosomal proteins while
maintaining good antigenicity.1

Reagents. All stock solutions should be kept at 4 unless otherwise
indicated (see Table I).
TABLE I
Final concentration Volume/5 ml Stock solution
Cohen buffer
10 mM MgCl2 50 l 1M

25 mM sodium glycerol 3-phosphate 125 l 1 M, pH 7 ( 20 )
3 mM CaCl2 150 l 0.1 M
10 mM KH2PO4 250 l 0.2 M
0.5% NP40 250 l 10%
30 mM KCl 750 l 0.2 M

160 mM sucrose 1.00 ml 0.8 M ( 20 )
H2O 2.425 ml

(Can be kept at 4 for 2–3 days.)
Formaldehyde fixative
0.1 M NaCl 100 l 5M
2 mM KCl 50 l 0.2 M
10 mM NaH2PO4 50 l 1 M, pH 7
2% NP40 1.00 ml 10%
2% formaldehyde 270 l 37%
H2O 3.53 ml
(Must be made fresh daily. Use 37% formaldehyde stock within 6 months of purchase.)
TABLE II
Final concentration Volume/5 ml (ml) Stock solution (%)
Formaldehyde fixative squashing solution

45% acetic acid 2.25 100
H2O 2.75
(Must be made fresh daily.)
Acetic acid/formaldehyde squashing solution
45% acetic acid 2.25 100
3.7% formaldehyde 0.25 37
H2O 2.25
(Must be made fresh daily.)
Squashing without prior formaldehyde fixation may result in loss of

most of the histones and approximately 10% of the nonhistone chromo-
somal proteins. In particular, lysine-rich proteins will be extracted.1 How-
ever, depending on the protein, it may be necessary to skip the
formaldehyde fixation prior to squashing. In some instances, formaldehyde
can hide antigenic determinants and/or lead to denaturation of the protein.
Forty five percent acetic acid does not extract or perturb most nonhistone
chromosomal proteins. If omitting the 1 h formaldehyde fixation, the sali-
vary glands are fixed less harshly by excising in acetic acid/formaldehyde
squashing solution (45% acetic acid, 3.7% formaldehyde, see Table II)
and incubating for 3–40 min. This solution usually gives better preservation
of morphology and antigenicity than 45% acetic acid alone.14 Glands that
are fixed in this solution are also squashed in this solution. This ‘‘conven-
tional fixation technique’’ results in polytene chromosomes with good band
morphology and good antigenicity of many proteins, but a significant
number of proteins are extracted.
Squashing of Polytene Chromosomes

After the glands are fixed, they are now ready to be squashed. Place a
drop of 45% acetic acid into the center of a clean siliconized coverslip15
and transfer the glands to the drop with forceps. Clean a glass slide with
95% ethanol, dry it, and then position the middle of the slide so that it is
on top of the coverslip. Break open the salivary gland nuclei by picking
up the slide and moving the coverslip back and forth by gently tapping
14
G. Holmquist, Chromosoma 36, 413 (1972).
15
J. Sambrook and D. W. Russell, ‘‘Molecular Cloning: A Laboratory Manual,’’ Cold Spring
Harbor Laboratory Press, Cold Spring Harbor, NY, 2001.
the edges of the slide between the thumb and forefinger. Once the salivary
gland cells and nuclei break open, the arms of the chromosomes will spread
out. Turn the slide over so that the coverslip is now on top of the slide and
monitor the spreading of the chromosome arms by looking at them under a
phase-contrast microscope at 400. Mark the location of the squash with a
marker so that it may be easily found when looking under the fluorescent
microscope. If the chromosome arms have not spread sufficiently, place the
slide on the bench, coverslip up, and tap the coverslip with the eraser end of
a pencil. When the spreading of the chromosome arms is satisfactory, flat-
ten the chromosomes by firmly applying thumb pressure to the coverslip.
Place a folded tissue between the thumb and the coverslip so that grease
from the hands does not get on the slide. Be sure that the coverslip does
not move under pressure. If movement occurs, stretching and breaking of
the chromosome arms will result. If too much pressure is applied, the result
will be fragmented chromosomes. Also, be sure that the squash does not
dry out. If the squash is allowed to dry, subsequent staining will be poor.
Once the spread is satisfactory, quickly submerge the slide in liquid nitro-
gen using hemostatic forceps. Once bubbling has ceased, remove the slide
from the liquid nitrogen and flick off the coverslip with a razor blade. Be
sure not to scrape the slide or the chromosomes may be scraped off. Before
the specimen is allowed to thaw, immerse the slide in Tris-buffered saline
(TBS-Tween, see Table III). If the slides will not be used within 2–4 h, you

may store them in slide storage medium at 20 for up to a month. If the
squash is allowed to dry out at any time during the procedure, the result
will be poor staining of the chromosomes; the stain will outline the chromo-
somes rather than staining specific bands (see Fig. 2). The squashes are
TABLE III
Final concentration Volume Stock solution
10 TBS-Tween
0.2 M Tris-HCl 200 ml 1 M, pH 8
17% NaCl 170 g
Tween-20 (Sigma) 10 ml 100%
H2O Bring to 1 L
(Dilute 10-fold in H2O before use.)
Slide storage medium
67% glycerol 335 ml 100%
a
33% PBS 165 ml
a
One liter of phosphate-buffered saline (PBS) is made with 20 g of NaCl, 0.5 g of KCl,
0.5 g of KH2PO4, and 1.45 g of Na2HPO4.
Fig. 2. Fluorescent outlining of chromosomes. Left panel: phase-contrast. Right panel:

immunofluorescence. Adapted, with permission, from: L. M. Silver, ‘‘Methods for Analyzing
the In situ Distributions of Chromosomal Proteins by Immunofluorescence.’’ Ph.D. thesis,
Harvard University, 1977.
most vulnerable to drying out after they are flattened. After flattening, the
liquid layer is very thin. If the acetic acid is seen to recede from the edges of
the coverslip, particularly at the corners, the specimen may be too dry to be
stained well. Therefore, during the squashing procedure, perform all steps
as quickly as possible without sacrificing the quality of your specimen. A
humidifier on the bench may help if the air is dry. Results will improve with
practice. A troubleshooting guide is provided at the end of this section to
help solve any problems that are encountered.
Antibody Staining Procedure

To determine the localization pattern of chromosomal proteins on
the polytene chromosomes, the spreads are first incubated with a primary
antibody against the protein(s) of interest and then incubated with a
fluorescent or enzyme-linked secondary antibody against the primary anti-
body. Immunofluorescent labeling will be described in detail here. Other
secondary labeling techniques may also be used. The use of a fluorescently
labeled secondary antibody to localize to primary antibodies was first
achieved by Coons in the early 1940s.16,17 In this technique, 5–10 secondary
antibodies can bind to each primary antibody, greatly increasing the
original signal.18–20 The antibody staining procedure described here is

essentially as performed in Weller and Coons,21 with some minor
modifications as described by Silver and Elgin.1,22
Several criteria must be followed when using antibodies for immuno-
fluorescent staining of polytene chromosomes. First, the titer of both the
primary and the secondary antibodies must be determined empirically. If
too high a concentration of either antibody is used, extensive nonspecific
staining or the loss of specific signal might result. Too high of a dilution will
result in a loss of signal. Typical dilutions used in our lab range from 1:50 to
1:5000. Second, appropriate controls must be carried out to be sure that the
antiserum is specific for the protein of interest. Preimmune serum from
many animals will carry antibodies that react generally with the chromo-
somes and some will carry antibodies that will give a very specific banding
pattern. The background signal will often be lost upon dilution since non-
specific antibodies are usually of low titer, but it is best to screen animals by
checking the preimmune serum prior to selecting animals for immuniza-
tion. Antibodies used for staining should be tested for specific binding to
Drosophila chromosomal proteins by western blotting or immunoprecipi-
tation to be sure that the interaction is specific. Unfortunately, the antibody
may react well on western blots but not on polytene chromosomes or
vice versa. Chromosome staining with only the secondary antibody should
be done with each new batch of antibody obtained to be sure that the
secondary antibody by itself does not label the chromosomes.
To begin the staining procedure, wash the slides with the polytene
chromosome squashes three times with gentle agitation for 5 min each in
cold TBS-Tween. The antiserum or monoclonal antibody should be diluted
in TBS-Tween as determined empirically. Blocking agents may be added to
the solution to help reduce nonspecific labeling. Commonly used blocking
agents are fetal calf or goat serum (10%), 5 mg/ml -globulin, and/or 5%
nonfat milk (all final concentrations). Use of these blocking agents will
not alter antibody labeling patterns. With a tissue, dry all regions of the
slide except for the squash area. Place the slide horizontally in a humidity
chamber. The humidity chambers in our lab consist of a plastic box with a
16
A. H. Coons, H. J. Creech, and R. N. Jones, Proc. Soc. Exp. Biol. Med. 47, 200 (1941).
17
A. H. Coons, H. J. Creech, R. N. Jones, and E. Berliner, J. Immunol. 45, 159 (1942).
18
M. Goldman, in ‘‘Fluorescent Antibody Methods.’’ Academic Press, New York, 1968.
19
L. A. Sternberger, in ‘‘Immunocytochemistry.’’ Prentice-Hall, Englewood Cliffs, NJ, 1974.
20
C. A. Williams and M. W. Chase, in ‘‘Methods in Immunology and Immunocytochemistry.’’
5. Academic Press, New York, 1976.
21
T. H. Weller and A. H. Coons, Proc. Soc. Exp. Biol. Med. 86, 789 (1954).
22
L. M. Silver and S. C. R. Elgin, in ‘‘The Cell Nucleus’’ (H. Busch, ed.), 5, p. 215. Academic
Press, New York, 1978.
tight-sealing lid. Wet paper towels or cheesecloth should be placed at the

bottom of the chamber and glass rods on clay holders are used to keep
the slides perfectly horizontal and above the towels. Empty plastic pipette
tip racks can also be used to support the slides in a box. Apply 100–200 l
of diluted antiserum with or without blocking agent to the squash area
immediately after wiping the slide with the tissue and placing the slide in
the humidity chamber. Be sure that there are no bubbles in the applied
solution or some parts of the squash may not be exposed to the antibody.
Be sure that the squash is kept wet at all times. Incubate the slides with the
primary antibody for 15–120 min at room temperature.
When the primary incubation is complete, rinse the antiserum from the
slide with cold TBS-Tween in a squirt bottle. Then wash the slide in cold
TBS-Tween (with gentle agitation of the slide in a rack) three times for
5 min each. Wipe the slide with the tissue as before and place back in the
humidity chamber. Add the correct dilution of secondary antibody to the
slide in the same manner that the primary antibody was added and incu-
bate for 15–120 min at room temperature as described earlier. Typical dilu-
tions range from 1:50 to 1:5000. Keep the slides away from bright light
when incubating with the fluorescently labeled secondary antibody, or the
fluorescence will fade. Work quickly and use conditions of low light if pos-
sible. After the incubation is complete, wash the slides with cold TBS
buffer as before and wipe the slide dry except for the region of the squash.
Place one drop of mounting solution on the squash and then place a non-
siliconized coverslip over it. Mounting solution may be made as shown in
Table IV or purchased under the name of Vectashield Mounting Medium
(cat# H-1000, Vector Labs). Blot excess solution from the slide and then,
with a tissue, press the coverslip hard onto the slide without letting the cov-
erslip slide. Blot excess solution from the sides once more and then use
clear fingernail polish to seal the squash and to prevent the specimen from
drying out. It is advantageous to use clear fingernail polish in case some of
the chromosomes end up covered by the polish.
We generally use a Nikon E600 microscope with a 400 oil immersion
lens, an epifluorescence attachment, and an Optronics digital camera for
TABLE IV
Final concentration Volume/100 ml Stock solution
Mounting solution
90% glycerol 90 ml 100%
0.1 M Tris-HCl 10 ml 1 M, pH 7
0.2% n-propyl gallate 0.2 g
viewing and photography. It is helpful to scan the slide with a 10 dry lens
first to find a good squash and then move to a 400 oil immersion lens
for photographs. The images are processed in Photoshop and printed on
photo-quality glossy paper using a high-quality ink jet printer or laser
printer. Different fluorophores fade at different rates. It is advantageous
to minimize the amount of time spent viewing the chromosomes under
the fluorescent microscope prior to photography to minimize fading. It is

also advantageous to store the slides in the dark at 4 ; this keeps the slides
from drying out and the fluorophores from fading.
The methodology described thus far is for labeling one protein of inter-
est. Multiple proteins may be viewed on polytene chromosomes by incubat-
ing with the primary antibodies at the same time (mixing them together in
the diluent) or incubating sequentially with wash steps in between. The
secondary antibodies may also be added together or sequentially.
Pulverize n-propyl gallate with a mortar and pestle before weighing,
and allow it to dissolve by stirring overnight in the solution. n-Propyl
gallate inhibits the loss of fluorescence during viewing.
Troubleshooting Guide
The following is a troubleshooting guide. For each problem stated,
there is a possible explanation and an approach described to correct the
problem. (Adapted from Methods Cell Biol. 35, 214–216 (1991), with
permission from Elsevier.)
a. Small Polytene Chromosomes

1. Larvae may be crowded or may have been harvested prior to the
correct stage.
Correction: Be sure that the larvae are not crowded; harvest them
when they are their fattest and crawling up the side of the bottle.
2. Larvae may have been grown at too warm of a temperature.

Correction: Growth at 18 is optimal for wild-type flies such as
Oregon-R, but may vary for other fly lines.
b. Poor Chromosome Morphology (i.e., Indistinct Bands Under Phase

Contrast, Smeared Staining Pattern)
1. Incubation too long in Cohen buffer, insufficient incubation in
formaldehyde fixative, or incubation too long in 45% acetic acid.
Correction: Check incubation periods. Different incubation
periods may be optimal for some proteins of interest; this should
be determined empirically.
c. Polytene Chromosomes Do Not Spread Well

1. Extended incubation in formaldehyde fixative.
Correction: Incubate for 25 min or less.
2. Spreading incomplete.
Correction: Using a phase-contrast microscope, constantly moni-
tor the squash; tap the coverslip over the squash with the eraser
end of a pencil until chromosome arms are well spread.
3. Squashed too gently.
Correction: Apply more pressure in the final flattening.
d. Too Much Background Material

1. Glands were insufficiently washed in Cohen buffer.
Correction: Incubate the glands for at least 8 min.
2. Fat body and other extraneous tissues were not removed.
Correction: Remove extraneous tissues before squashing.
3. Dirty coverslips/and or slides.
Correction: Clean slides and coverslips with 95% ethanol, dry with
lens paper, handle glass at edges, and squash with a tissue between
your thumb and the coverslip.
e. Broken Polytene Chromosomes

1. Squashing may have been done with too much force, or the coverslip
may have moved during the procedure.
Correction: Squash more gently and monitor under a phase-
contrast microscope. Hold the coverslip while tapping with an
eraser on the end of a pencil to prevent sliding.
2. Insufficient incubation in formaldehyde fixative, or too long in 45%
acetic acid.
Correction: See b1.
f. Refractile Polytene Chromosomes

1. Chromosomes squashed too gently. The chromosomes are not flat
enough.
Correction: See c3.
2. Allowed chromosomes to dry.
Correction: Keep squash area wet at all times. If necessary, use a
humidifier on the bench during squashing steps if the air is very dry.
3. After liquid nitrogen freezing, the chromosomes thawed.
Correction: Immediately immerse slides in TBS after removing
the coverslip.
g. Staining Pattern Outlines the Chromosome

1. Chromosomes dried (one observes fluid retracting at the corners of
the coverslip).
Correction: See f2.
h. Stretched-Out Chromosomes
1. Coverslip moved during flattening.
Correction: See e1.
2. Insufficient formaldehyde fixation.
Correction: See b1.
i. No Intact Chromosomes in Squash Area

1. Squashed gland floated out from under coverslip.
Correction: Use less 45% acetic acid on coverslip.
2. Coverslip moved during squashing or squashed too hard.
Correction: See e1.
3. Cannot find chromosomes in squash area at high magnification.
Correction: With a waterproof marker, mark the position of the
squashed gland on the edges of the coverslip.
4. Scraped off chromosomes with razor blade.
Correction: Be sure to freeze the slide before removing
the coverslip; insert the edge of a razor blade under a corner of
the coverslip and flip off the coverslip with an upward motion.
5. Chromosomes fell off the slide after the slide was frozen.
Correction: If the slides are not to be used within 2–4 h, keep the
slides in storage medium.
6. Incomplete siliconization of the coverslip.
Correction: Review methods of siliconization.15
j. High Intensity of Staining of Debris

1. Antiserum contains antibodies that cross-react with components
other than the chromosomal protein of interest.
Correction: Preabsorb or affinity purify the antiserum.
2. No blocking agents or insufficient amount of blocking agents used.
Correction: Add 10% calf or goat serum, 5 mg/ml of -globulin,
and/or 5% nonfat dried milk to the antibody. Increase the
concentrations if needed.
3. Antibody dried onto chromosome squash.
Correction: Use the humidity chamber during antibody incubations,
be sure that the slides are level on application of antibody, and
use a sufficient amount of antibody solution (200 l) to avoid drying.
4. Titer of antibody is too high.

Correction: Decrease titer of antibody.
k. Weak or Absent Signal on Chromosomes

1. Antibody does not react to the antigen on chromosomes.
Correction: Try variations in the fixation protocol.
2. Incubation in formaldehyde is too long.
Correction: See c1.
3. Weak or faded fluorescent label.
Correction: On incubation with fluorescent conjugated antibody,
keep slides in the dark. Be sure n-propyl gallate was added to the
mounting medium.
4. Titer of antibody is too low.
Correction: Increase antibody titer.
5. Incorrect secondary antibody is being used.
Correction: Check that the secondary antibody is against the
animal that the primary antibody was derived from.
l. Chromosomes Are Present, but Not Coming into Focus or Are Blurry
1. Two coverslips may be stuck together on the slide.
Correction: Remove the upper coverslip with a razor blade.
2. Chromosomes were not squashed flat and are difficult to focus on
due to multiple focal planes.
Correction: Be sure to sufficiently flatten the chromosomes when
squashing.
3. The table that the microscope is on may be vibrating while you are
viewing the chromosomes and taking pictures.
Correction: A vibration isolation table may be needed if a suitable
stable surface cannot be found.
Applications and Results

Immunofluorescent staining can be used to analyze the global distribu-
tion of a specific protein on polytene chromosomes, as well as to compare
that of two or more proteins simultaneously in wild-type or mutant larvae.
It can also be used for comparative studies of Drosophila. Genetic stocks
with rearrangements can provide important information regarding the au-
tonomy of a protein’s association within a region of the polytene chromo-
somes. The impact of mutations in chromosomal proteins on global
organization of the chromosomes can also be examined. Immunological
staining can indicate if proteins localize correctly in the mutant background
or if they localize ectopically. Finally, we will discuss the value of

transgenic fly lines that can be used to test the effect of a specific
transcriptional regulator on gene expression.
Simultaneous Localization of Multiple Proteins on Polytene Chromosomes:

HP1, HP2, and Modified Histones
An example of the use of immunofluorescence microscopy of polytene
chromosomes to help to determine the function of an unknown protein
comes from studies of Heterochromatin Protein 1 or HP1. Monoclonal
antibodies were produced using a protein extract of D. melanogaster em-
bryonic nuclei, and the antibodies were screened by staining polytene
chromosomes as described earlier.23 One of these antibodies, C1A9, recog-
nizes a protein that is highly concentrated in the chromocenter of salivary
gland polytene chromosome squashes of third instar larvae, HP1. HP1 is
most prominently associated with the chromocenter, the telomeres, and
the small fourth chromosome; some euchromatic sites are stained, albeit
with less intensity.9 The chromocenter, as mentioned previously, is formed
from the fused centric heterochromatin of all four chromosomes. Thus, the
staining pattern suggests that the antigen is primarily associated with het-
erochromatin.24 The gene for HP1 was mapped to region 29A,23,25 where
a dominant suppressor of position effect variegation (PEV) had already
been mapped.26 PEV occurs when a gene that normally resides within
euchromatin is placed near heterochromatin; the gene is expressed in
some cells and repressed in others, giving rise to a variegated phenotype.27
This phenotype is most commonly observed in the fly eye, when the white
gene is translocated next to heterochromatin on the X chromosome. Muta-
tions within HP1 have been found to be strong suppressors of PEV.25
These results indicate that HP1 contributes to the heterochromatic struc-
ture responsible for silencing. In the last few years, it has been shown
that HP1 binds to histone H3 methylated at lysine 9, a prominent marker
of heterochromatic domains.28,29
23
T. C. James and S. C. R. Elgin, Mol. Cell. Biol. 6, 3862 (1986).
24
T. C. James, J. C. Eissenberg, C. Craig, V. Dietrich, A. Hobson, and S. C. R. Elgin, Eur. J.
Cell Biol. 50, 170 (1989).
25
J. C. Eissenberg, T. C. James, D. M. Foster-Hartnett, T. Hartnett, V. Ngan, and S. C. R.
Elgin, Proc. Natl. Acad. Sci. USA 87, 9923 (1990).
26
D. A. R. Sinclair, R. C. Mottus, and T. A. Grigliatti, Mol. Gen. Genet. 191, 326 (1983).
27
J. B. Spofford, in ‘‘The Genetics and Biology of Drosophila’’ (M. Ashburner and
E. Novitski, eds.), 2a, p. 955. Academic Press, Orlando, FL, 1976.
28
29
M. Lachner, D. O’Carroll, S. Rea, K. Mechtler, and T. Jenuwein, Nature 410, 116 (2001).
Fig. 3. Simultaneous immunolocalization of HP1 and HP2. Upper: polytene chromosomes

stained with anti-HP1 (red, left) and anti-HP2 (Ab P-6; green, right). Lower: close-up view of
the chromocenter. Single antibody signals are shown in black and white; the color merge of
HP1 signal (red) and HP2 signal (green) is shown in the center. Note the staining of the
chromocenter, in a banded pattern along the fourth chromosome (double arrowhead) and of
5–6 bands in 31B (arrow). The arrowhead identifies a euchromatic band positive for HP1 but
not HP2. Adapted, with permission, from: C. D. Shaffer, G. E. Stephens, B. A. Thompson, L.
Funches, J. A. Bernat, C. A. Craig, and S. C. R. Elgin, Proc. Natl. Acad. Sci. 99, 14332 (2002).
We are interested in other proteins that bind to HP1 and might contrib-
ute to a multiprotein complex required for heterochromatin-induced gene
silencing. Through a yeast two-hybrid screen using HP1 as bait, an HP1-
interacting protein, Heterochromatin Protein 2 (HP2), has been identified.8
Upon staining of polytene chromosomes with a polyclonal antibody pre-
pared against a peptide of HP2 from the C-terminal region, a polytene
chromosome staining pattern nearly coincident with that of HP1 is seen
(see Fig. 3). Mutations in HP2 result in suppression of PEV, suggesting that
HP2 also has a role in heterochromatin structure. The nearly coincident
staining pattern of these two proteins shows that they colocalize, thus
suggesting that they may be part of a multiprotein complex.
Methylation of lysine 9 on histone H3 has been found to recruit HP1.28,29
HP1 also interacts with the histone methyltransferase SU(VAR)3–9,10,30,31
providing a mechanism for the spreading of heterochromatin. Various
30
L. Aagaard, G. Laible, P. Selenko, M. Schmid, R. Dorn, G. Schotta, S. Kuhfittig, A. Wolf,
A. Lebersorger, P. B. Singh, G. Reuter, and T. Jenuwein, EMBO J. 18, 1923 (1999).
31
S. Rea, F. Eisenhaber, D. O’Carroll, B. D. Strahl, Z. W. Sun, M. Schmid, S. Opravil, K.
Mechtler, C. P. Ponting, C. D. Allis, and T. Jenuwein, Nature 406, 593 (2000).
Fig. 4. Immunfluorescent staining for HP1 versus histone H3 acetylated at lysine 14 on the
fourth chromosome of D. melanogaster. Panel a: HP1. Panel b: Histone H3 acetylated at
lysine 14. Notice the nonoverlapping staining patterns of HP1 and this histone modification,
suggesting different functions.
histone modifications have been found to recruit nonhistone chromosomal

protein complexes. Antibodies specific for the modified histones have
allowed us to analyze the global distribution of both the nonhistone
chromosomal proteins and the modified histones. Figure 4 shows double la-
beling using antibodies for histone H3 acetylated on lysine 14, associated
with active regions of the genome,32 versus HP1, associated with inactive
regions, on the fourth chromosome of D. melanogaster. Through double
labeling, it is evident that HP1 and histone H3 acetylated on lysine
14 have different distribution patterns and therefore are likely to have
different roles in gene regulation.
Analyzing the Distribution of Proteins on Rearranged Chromosomes and

Chromosomes from Related Species
As mentioned previously, HP1 was found prominently associated with
the fourth chromosome of D. melanogaster. This chromosome is immedi-
ately adjacent to the chromocenter. To show that the staining of HP1 on
the fourth chromosome is not due to the chromosome’s close proximity
to the chromocenter, a fly line with a translocation of a portion of the
fourth chromosome to the third chromosome was stained for HP1 associ-
ation.9 Salivary gland chromosome squashes were prepared from third
instar larvae of the stock T(3;4)f/In(3L)P. A region of approximately seven
bands of the fourth chromosome is translocated to position 65D1–2 on
chromosome 3 in this stock. C1A9 antibody staining of this stock shows
that the association of the fourth chromosome with HP1 is not dependent
on its close association with the chromocenter.
32
J. Nakayama, J. C. Rice, B. D. Strahl, C. D. Allis, and S. I. S. Grewal, Science 292, 110 (2001).
Fig. 5. Immunofluorescent staining of HP1 in three Drosophila species. Left panel:

D. melanogaster. Middle panel: D. sechellia. Right panel: D. pseudoobscura. Notice the
prevalent staining of the chromocenter and the dot chromosome.
In another line, T(2;3)ltx13, chromosome arm 2L is broken within the

-heterochromatin at 2LH37, between the light gene and the chromocen-
ter, and joined to chromosome arm 3R at map position 97D2. Thus, a por-
tion of -heterochromatin is now separated from the chromocenter by long
segments of euchromatin. Upon immunofluorescent staining with the
C1A9 antibody, this region is prominently stained, again indicating that
the association of HP1 with this heterochromatic region is autonomous
and not dependent on close association with the chromocenter.
The availability of fly stocks from other Drosophila species allows
one to check for homologues and determine their localization. An HP1
homolog has been identified in D. virilis by preparing polytene chromo-
some squashes and staining with C1A9 antibody.9 There are 40–60 million
years of evolution between D. melanogaster and D. virilis. HP1 has been
observed as a heterochromatic protein, predominantly associated with
the pericentric heterochromatin found at the chromocenter in all Drosoph-
ila species examined to date (D. pseudoobscura, D. sechellia, and others)
(see Fig. 5). Interestingly, the banded portion of the fourth chromosome
or dot chromosome is not stained by HP1 antibodies in D. virilis.9
Analyzing the Impact of Mutations in Chromosomal Proteins

Su(var)3–9 encodes a histone methyltransferase that selectively methyl-
ates lysine 9 of histone H3 and binds to heterochromatin through direct
interaction with HP1.10 SU(VAR)3–9, like HP1, has been shown to be a
dominant modifier of PEV, that is, heterochromatin-induced gene silenc-
ing.33 Immunological assays of larval salivary gland polytene chromosomes
done using a polyclonal antibody indicate that SU(VAR)3–9 associates
with the chromocenter, the fourth chromosome, and more weakly at the
telomeres and a few euchromatic sites.10 GFP antibodies also recognize a
33
G. Reuter and I. Wolff, Mol. Gen. Genet. 182, 516 (1981).
SU(VAR)3–9-EGFP fusion protein. The role of HP1 in the association

of SU(VAR)3–9 with heterochromatin was examined by staining HP1-
deficient third instar larvae that express a SU(VAR)3–9-EGFP fusion
protein. In these larvae, SU(VAR)3–9-EGFP remains associated with the
heterochromatin and the fourth chromosome, but it is also associated
generally with euchromatic regions due to methylation of lysine 9 in
histone H3 at euchromatic sites. HP1 is thus essential for maintaining the
correct localization of SU(VAR)3–9 to heterochromatin.
SU(VAR)3–9 null mutants were analyzed for the distribution of HP1
on polytene chromosomes. In these larvae, the association of HP1 with
the chromocenter was dramatically reduced. Interestingly, the association
of HP1 with the fourth chromosome was not affected, suggesting that
HP1 binding to the fourth chromosome may be governed by a different
methyltransferase activity localized to that portion of the genome.
Examining Protein Localization on Custom P-Elements

Drosophila polytene chromosomes offer the ability to monitor protein-
DNA and protein-protein interactions that participate in chromatin
packaging and gene expression in vivo. To test the effects of a specific tran-
scriptional regulator on gene expression it is often desirable to tether the
regulator upstream of a reporter gene. This can be accomplished by fusing
the transcriptional regulator to a DNA binding domain that has a known
target sequence. The target sequence is cloned upstream of a reporter gene
whose expression can be easily monitored. This methodology has been ap-
plied to the study of HP1. Full-length HP1 was fused to the DNA binding
domain of the LacI repressor (see Fig. 6, panel A). The LacI-HP1 fusion
protein was expressed in Drosophila stocks carrying a mini-white reporter
transgene with LacI binding sites positioned upstream of the transcription
start site (see Fig. 6, panel A). Expression of the reporter gene provides pig-
ment to the Drosophila eye. Association of the HP1 fusion protein with the
LacI binding sites was observed in preparations of fixed and squashed poly-
tene chromosomes (see Fig. 6, panel B) and correlated with the lack of eye
pigmentation at 25 of 26 sites tested.34 These results indicate that HP1 can
establish silent chromatin at the majority of ectopic chromosome locations.
Cytological studies on fixed and squashed polytene chromosomes can
also be used to monitor protein-protein interactions in vivo. In the case
described earlier, it is possible to determine whether HP1 associates with
known partner proteins when tethered at ectopic genomic locations. An
34
Y. Li, J. R. Danzer, P. Alvarez, A. S. Belmont, and L. L. Wallrath, Development 130,
1817 (2003).
Fig. 6. (continued )
association between the histone H3 methyltransferase SU(VAR)3–9 and

HP110 is thought to be responsible for the mechanism of spreading silent
chromatin.29 Polytene chromosomes from a stock containing tethered
HP1 and expressing the GFP tagged SU(VAR)3–9 protein were fixed,
squashed, and stained with antibodies to HP1 and GFP. The results indi-
cate that SU(VAR)3–9 is recruited to the site of tethered HP1 as well as
sites of endogenous HP1 (see Fig. 6, panels C–E).
Reporter transgenes have also been designed to help elucidate the issue
of cellular memory. The Fab-7 element is used by Drosophila Polycomb
and trithorax group proteins to maintain repressed or active gene expres-
sion, respectively, of the segmentation genes during embryogenesis.35 Once
the active or repressed state is established, the state is mitotically inherited
throughout development and to some degree through meiosis. Histone H4
hyperacetylation of this element can be inherited epigenetically through
mitosis and meiosis and can be followed by immunological staining of poly-
tene chromosomes. Fab-7 has thus been deemed a cellular memory
module.36
Conclusions
In conclusion, we have presented a method for determining the in vivo
distribution of chromosomal proteins on Drosophila polytene chromo-
somes. Some applications of this technique have also been discussed. By
combining genetic, biochemical, and molecular biology techniques
with the cytological approach, a greater understanding of the molecular
35
G. Cavalli and R. Paro, Cell 93, 505 (1998).
36
K. Ekwall, T. Olsson, B. M. Turner, and R. C. Allshire, Cell 91, 1021 (1997).
Fig. 6. Determining protein-DNA and protein-protein interactions on chromosomes

in vivo. (A) Diagram of the tethering system showing the LacI-HP1 fusion transgene
expressed from an hsp70 promoter (left) and the mini-white reporter transgene containing
upstream lacI repeats (right). (B) Polytene chromosomes from a Drosophila stock expressing
a LacI-HP1 fusion protein and carrying the mini-white reporter inserted at cytological position
93A/B. Localization of the HP1 fusion protein is detected using antibodies to the LacI DNA
binding domain (green). LacI-HP1 associates with the lacI repeats (arrow) and with sites of
endogenous HP1. (C)–(E) Polytene chromosomes from a stock expressing a LacI-HP1 fusion
protein and a SU(VAR)3-9 protein tagged with GFP. HP1 localization is detected with
antibodies to HP1 (red, panel C), SU(VAR)3-9EGFP localization is detected with antibodies
to GFP (green, panel D), and colocalization is observed in the merged image (yellow, panel
E). HP1 and SU(VAR)3-9 show colocalization over the mini-white reporter at cytological
position at 1E (arrow) and at sites of endogenous HP1.
[26] immunofluorescent labeling of DROSOPHILA polytene chromosomes 393
mechanisms of gene regulation can be gained. Few other systems offer the
well-developed genetic tools in combination with the ability to perform
cytological studies as Drosophila does. For more information on the
formation of polytene chromosomes and various other applications, see
Schwartz et al.4 this volume.
[26] Indirect Immunofluorescent Labeling of

Drosophila Polytene Chromosomes: Visualizing
Protein Interactions with Chromatin In Vivo
By Brian E. Schwartz, Janis K. Werner, and John T. Lis
Background
Drosophila melanogaster was first recognized as a valuable experimen-
tal organism 100 years ago. One of the particularly attractive features of
this model system for studies of genes and their regulation is the ‘‘giant’’
or polytene chromosomes that occur in the secretory glands of Drosophila
(as well as other dipteran flies). Historically, polytene chromosomes pro-
vided an important link between the genetic map and the physical location
of deletions, insertions, inversions, and translocations on the genetic map.
More recently, these chromosomes have provided an efficient means of
analyzing both the global distribution of particular proteins on chromo-
somes and the recruitment of particular proteins to specific chromosomal
loci that are undergoing changes in activity.
Polytene chromosomes form in cells that grow in size without dividing
during larval development. In these cells, homologous diploid pairs of
chromosomes are tightly paired and undergo successive rounds of amplifi-
cation. Each synapsed pair may replicate up to nine times forming around
1000 strands of DNA, which remain aligned and attached to each other.
This precise alignment allows differences in chromatin compaction to be
seen as a series of bands and interbands extending across the width of the
chromosome arms. Electron microscopy studies show that about 95% of
the DNA is concentrated in the bands, which are much more compacted
regions of chromatin than the interbands, which separate the bands. Al-
though the packing ratio of chromatin in polytene chromosomes varies
regionally, the average packing ratio has been estimated to be between
30 and 57, about 100 times more extended than a metaphase chromosome.
This is slightly less compact than a 30-nm chromatin fiber, which has a

mechanisms of gene regulation can be gained. Few other systems offer the
well-developed genetic tools in combination with the ability to perform
cytological studies as Drosophila does. For more information on the
formation of polytene chromosomes and various other applications, see
Schwartz et al.4 this volume.
[26] Indirect Immunofluorescent Labeling of

Drosophila Polytene Chromosomes: Visualizing
Protein Interactions with Chromatin In Vivo
By Brian E. Schwartz, Janis K. Werner, and John T. Lis
Background
Drosophila melanogaster was first recognized as a valuable experimen-
tal organism 100 years ago. One of the particularly attractive features of
this model system for studies of genes and their regulation is the ‘‘giant’’
or polytene chromosomes that occur in the secretory glands of Drosophila
(as well as other dipteran flies). Historically, polytene chromosomes pro-
vided an important link between the genetic map and the physical location
of deletions, insertions, inversions, and translocations on the genetic map.
More recently, these chromosomes have provided an efficient means of
analyzing both the global distribution of particular proteins on chromo-
somes and the recruitment of particular proteins to specific chromosomal
loci that are undergoing changes in activity.
Polytene chromosomes form in cells that grow in size without dividing
during larval development. In these cells, homologous diploid pairs of
chromosomes are tightly paired and undergo successive rounds of amplifi-
cation. Each synapsed pair may replicate up to nine times forming around
1000 strands of DNA, which remain aligned and attached to each other.
This precise alignment allows differences in chromatin compaction to be
seen as a series of bands and interbands extending across the width of the
chromosome arms. Electron microscopy studies show that about 95% of
the DNA is concentrated in the bands, which are much more compacted
regions of chromatin than the interbands, which separate the bands. Al-
though the packing ratio of chromatin in polytene chromosomes varies
regionally, the average packing ratio has been estimated to be between
30 and 57, about 100 times more extended than a metaphase chromosome.
This is slightly less compact than a 30-nm chromatin fiber, which has a

packing ratio of about 50. The amplification of aligned chromatids is

responsible for the impressive width of the polytene chromosome, which
is about 5 m, or the length of a typical metaphase chromosome. A tight
association is formed not only among the replicated chromatids but also
among the four chromosomes of the Drosophila genome at the centro-
meres of each of the chromosomes. These chromosome centromeres are
attached to each other and together comprise the chromocenter, a region
of underreplicated heterochromatin that is diffuse and lacks the distinctive
banding of the euchromatin arms.
Despite their unusual structure, polytenes function as interphase
chromosomes and are transcriptionally active and responsive to hormonal
and environmental signals. The most visible manifestations of this tran-
scriptional activity are the puffs that can form at highly active loci. Puffs
are localized regions of decondensed chromatin that, in Drosophila, occur
in a developmental stage-specific pattern. They also appear at loci har-
boring stress-responsive genes following an environmental challenge, such
as heat shock. Generally, puffing occurs at loci undergoing high levels of
transcription, and the size of the puff roughly corresponds to the length
of the transcription unit.1 Nonetheless, the processes of transcription
and puffing are not strictly coupled. For example, the drug sodium salicyl-
ate can elicit puffs at some heat shock gene loci, yet transcription of
these genes is blocked by this drug.2 This type of puffing is likely a conse-
quence of salicylate-induced DNA binding of heat shock factor to the
multiple regulatory elements associated with these loci.
Indirect immunofluorescence staining of polytenes provides a rapid way
of determining the genomic distribution of chromosomal proteins. This was
first recognized in early experiments in the laboratories of Elgin3 and
Bautz.4 These early experiments revealed with spectacular clarity the loca-
tion of RNA polymerase II and its strong association with chromosomal
puffs and interbands, and histones and their concentration in DNA-rich
chromosomal bands. More recently, antibodies have been used to investi-
gate a variety of protein interactions with chromatin in vivo. For example,
male-specific lethal (msl) proteins involved in X chromosome dosage
compensation in males stain only the X chromosome.5 Factors involved
in gene silencing are found predominantly in heterochromatin.6 Many
1
J. A. Simon, C. A. Sutton, R. B. Lobell, R. L. Glaser, and J. T. Lis, Cell 4, 805 (1985).
2
N. A. Winegarden, K. S. Wong, M. Sopta, and J. T. Westwood, J. Biol. Chem. 271, 2697 (1996).
3
L. M. Silver and S. C. Elgin, Chromosoma 68, 101 (1978).
4
U. Plagens, A. L. Greenleaf, and E. K. Bautz, Chromosoma 59, 157 (1976).
5
J. R. Bone and M. I. Kuroda, Genetics 144, 705 (1996).
6
T. C. James, J. C. Eissenberg, C. Craig, V. Dietrich, A. Hobson, and S. C. Elgin, Eur. J. Cell
Biol. 50, 170 (1989).
transcription and RNA processing factors localize to sites of highly

active transcription, the chromosomal puffs.7 A TBP-related factor, TRF1,
stains chromosomes in a manner that revealed it was used predominantly for
Pol III gene promoters.8
The interpretation of a chromosomal staining pattern generated with a
specific antibody depends on the antibody or antibodies used. Several steps
can be taken to ensure that the pattern of antibody staining represents the
true distribution of the protein used to generate the antibody. First, several
animals can be used to generate antibody, and the specificity of the anti-
bodies can be checked by probing western blots of nuclear proteins. Only
those antibodies recognizing preferentially single bands corresponding to a
peptide of the correct molecular weight should be used to probe chromo-
somes. Obviously, the more specific the western the better; however, a
knowledge of the modification states, degradation products, and different
size forms of the protein is useful in assessing the specificity of the antibody
when more than one band is observed. Additionally, antibodies can be
purified by affinity chromatography to the specific antigen to further enrich
for the specific antibody.9 Finally, the specificity of any staining pattern can
be checked by double-staining with two antibodies raised to the same pro-
tein antigen but produced in different species (e.g., rat and rabbit10). The
staining patterns of two specific antibodies should show significant overlay,
or better, complete overlap.
The ability to immunostain for two or three factors at once allows one
to compare factor distributions relative to one another. The colocalization
of two factors may be indicative of a physical interaction in vivo. The
staining pattern of the kinase cdk7, for example, largely colocalizes with
its phosphorylated product, the carboxy-terminal domain of RNA poly-
merase II.11 Similarly, the kinase P-TEFb can be seen to colocalize with
its Pol II product and both can be compared to the promoter-bound
transcription factor HSF in the triple staining shown in Fig. 1.
In addition to gaining a genome-wide perspective on protein distribu-
tion, immunostaining of polytene chromosomes can also provide higher
resolution views of single genes. Now that the sequence of the genome is
complete (http://www.fruitfly.org/), and the physical and polytene maps
are more precisely correlated, the antibody staining of a particular band
7
E. D. Andrulis, J. Werner, A. Nazarian, H. Erdjument-Bromage, P. Tempst, and J. T. Lis,
Nature 420, 837 (2002).
8
S. Takada, J. T. Lis, S. Zhou, and R. Tjian, Cell 101, 459 (2000).
9
J. T. Lis, P. Mason, J. Peng, D. H. Price, and J. Werner, Genes Dev. 14, 792 (2000).
10
E. D. Andrulis, E. Guzman, P. Doring, J. Werner, and J. T. Lis, Genes Dev. 14, 2635 (2000).
11
B. E. Schwartz, unpublished results.
Fig. 1. Triple immuno-stain. Polytene chromosomes from heat shocked larvae were
stained with antibodies against HSF (red), the cyclin T subunit of P-TEFb (blue), and the
serine 2 phosphorylated form of RNA polymerase II (green).
or interband can be interpreted with precision. The resolution depends on

the chromatin environment of the gene; decondensed chromatin allows for
better resolution than condensed chromatin. For this reason, genes residing
within the interband regions are much better suited for this type of analysis
than those found in bands. In puffs, the most extreme case of chromatin de-
condensation, resolution is maximized. Puffs are therefore ideally suited to
the study of transcription factors and allow one to resolve factors associ-
ated with gene promoter regions from downstream transcribed regions.
For instance, the activator of heat shock genes, HSF, binds to specific pro-
moter DNA elements on heat stress and triggers the recruitment and
elongation of RNA polymerase II through the body of the gene. The gene
occupancy of these two factors can be resolved within the heat shock puff
(see Fig. 2).
The resolution and sensitivity of polytene immunofluorescence can be
further enhanced with the use of various transgenic fly lines. High-
resolution studies of a single gene of interest are often aided by mapping
a single-copy transgene that is removed from its native chromatin context.
This is especially valuable if the native site contains several repeats of the
gene, or nearby genes, which can complicate the assignment of a specific
stained polytene band to a particular gene or DNA segment. Detecting a
Fig. 2. Elongating form of Pol II (green) resolves from promoter-bound HSF (red) during
heat shock. The heat shock puff from a single transgenic copy of an Hsp70-lacZ fusion gene
at chromosomal locus 9D is shown. The HSF band at the left edge of the puff marks
the promoter of the transgene, while elongating Pol II appears within the body of the puff.
new labeled band created at a transgenic site (not present in the parental
line) to which the gene or DNA segment has been moved provides
unambiguous proof of the assignment. Creating transgenic sites with
smaller and smaller segments can, in principle, pinpoint the DNA sequence
with which the protein associates.
The sensitivity of the immunofluorescent signal can be further en-
hanced by examining polymeric sequences introduced on transgenes. We
have used this approach to create a polymer of 40 copies of a 55-bp
hsp70 promoter fragment12 that vigorously recruits HSF during heat
shock.13 By co-immunostaining with antibodies against other factors, we
have identified proteins that interact with DNA-bound HSF (see Fig. 3).
Therefore, this technique not only allows the amplification of otherwise
weak immunofluorescent signals but also allows one to determine the
protein recruitment ability of defined promoter elements.
12
H. Xiao, Ph.D. Thesis, Cornell University, 1989.
13
L. S. Shopland and J. T. Lis, Chromosoma 105, 158 (1996).
Fig. 3. (A) Transgenes containing a polymer of an Hsp70 promoter fragment recruit HSF
(green) during heat shock. (B) HSF can, in turn, recruit TFIIH (red). (C) Merge.
Methods
Culturing of Drosophila for Healthy Larvae

The best source of polytene chromosomes is from the salivary glands.
To obtain the best chromosome spreads, it is very important that healthy
larvae are used. Grow animals in bottles of newly made yeast glucose

media at about 23 and make sure the bottle is not overcrowded. The fly

lines are grown at about 23 . Approximately 40 flies are put into the bottle
for 2 days. New bottles are set up every 2 days. Larvae are collected from
bottles that are 6–8 days old. At this point bottles should have light brown
pupae as well as third instar larvae on the walls of the bottle. Wandering
third instar larvae are used typically. If first or second instar larvae are
climbing the walls, the culture is too dense or a media problem exists.
Heat Shocking of Larvae

Non-heat shock: Third instar larvae are taken directly from the culture
bottle and placed in the dissection buffer and dissected immediately.
Heat shock larvae: Third instar larvae are taken from the culture bottle
and placed on a piece of plastic wrap that has a small piece of moist filter
paper on it. The plastic wrap is closed with a dialysis clamp and submerged

in a 36.5 water bath for the prescribed amount of time. At the end of
the heat shock time, very quickly remove the plastic wrap from the water
bath and move the larvae quickly from the plastic wrap into the dissection
buffer and dissect immediately.
Dissection of Salivary Glands from Larvae [Modifications from

Protocol Received from Renato Paro]
The larvae are dissected in 100 l of dissection buffer (0.7% NaCl), in a
Boerner slide. To dissect the salivary glands from the third instar larvae,
hold the larvae with one pair of forceps and then using another pair of for-
ceps pull from the mouth parts. The salivary glands need to be identified
and dissected free from all the other parts of the larvae. Fat bodies can
be removed from the salivary gland, although not all need to be removed.
Transfer the salivary glands to 40 l of fix buffery on Rain-X–treated
coverslip.{ Incubate for 10–20 min before squashing.
y
Fix Buffer: 50 l 37% paraformaldehyde (1.85 g paraformaldehyde in 5 ml H2O and 70 l 1

N KOH). Heat to 80 for 30 min. Store at 20 in aliquots:
450 l acetic acid
500 l H2O
Keeps for 2 h at room temperature
Squashing of Salivary Glands

Using a base-treated slide,x pick up the coverslip with the fix buffer and
gland on it. Using a needle probe, tap up and down on the coverslip over
the area where the salivary glands are located. It is acceptable to let the
coverslip slide around as long as it stays in the same general area on
the microscope slide. Holding an edge of the coverslip in place with one
hand, using the needle probe, make circular motions on the coverslip,
paying particular attention to the area where the salivary gland is located.
Place slide, coverslip down, on a Kimwipe and apply pressure with your
hand to remove excess moisture from the slide. Using a phase scope
equipped with a 25 lens, scan the slide to find the chromosomes. Look
for intact chromosomes that have strong banding and are well-spread,
avoiding those slides that contain severely twisted or broken chromosomes.
Put the slide and coverslip in liquid nitrogen until ready to process all
slides.
Staining of Slides
1. Remove slides from liquid nitrogen. Blow breath over coverslip.
Using a razor blade under one corner, flick the coverslip off
quickly. Using a diamond-tip pen, make small lines on the
microscope slide to indicate where coverslip was.
2. Wash:
2 10 min in PBS
1 10 min in PBS with 1% Triton.
3. Block 40–60 min in block solution (PBS with 5% non-fat milk
powder).
4. Rinse well in PBS.
5. Add primary antibodies in PBS and 1% BSA, for a total of 20 l
per slide. Put antibody solution directly over the area where the
squashes are. Cover with 22-mm2 coverslip, using lines etched in

the slide to position the coverslip. Leave overnight at 4 in humid
chamber.
{
Rain-X–treated coverslips: using a Kimwipe saturated with Rain-X, coat 22 mm2 coverslip
with Rain-X. Lay coverslip on Kimwipe to dry. Rinse coverslip with Kimwipe saturated with
water. Let it dry. Clean the coverslips with lens paper prior to using. These treated coverslips
can be stored and used later.
x
Base treated slides: dissolve 140 g NaOH in 560 ml dH2O. Add 840 ml 95% EtOH. Put
microscope slides in glass slide trays. Cover with NaOH solution. Leave shaking for 2 h. Do
4 1 h rinses with dH2O. Put in drying oven. The slides need to dry at least overnight before
using. The slides should be stored in the oven until used. Slides that are removed from the
oven and not used can be returned to the oven and used another time.
Concentration of primary antibody used varies for different antiserum

preparations. We usually use them at 1:10 to 1:200.
6. Dip microscope slide in PBS to remove coverslip.
7. Rinse in PBS.
8. Wash 3 5 min in block solution.
9. Rinse in PBS.
10. Add secondary antibody solution: antibodies in PBS, 1% BSA,
and 2% donkey goat serum. Again have 20 l per slide. Cover
with 22-mm2 coverslip, using lines etched in slide to position
coverslip. Leave for 1 h in humid chamber at room temperature.
Concentration of secondary antibodies is 1:200 to 1:500.
Secondary antibodies are from Jackson ImmunoResearch, and we
usually use RhodamineRedX and Cy2.
11. Dip microscope slide in PBS to remove coverslip.
12. Wash 10 min in ‘‘wash 300’’ (300 mM NaCl, 0.2% NP40, 0.2%
Tween 20).
13. Wash 10 min in ‘‘wash 400’’ (400 mM NaCl, 0.2% NP40, 0.2%
Tween 20).
14. Rinse in PBS.
15. Hoechst stain (4 l Hoechst in 50 ml PBS for 20 min).
16. 2 15 min washes in PBS.
17. Add about 20 l of glycerol containing n-propylgallate (2.5%) to
coverslip.
18. Tap-dry the slide and touch to the glycerol on the coverslip.
19. The coverslip can be held in place with a dab of nail polish.

20. Store slides at 20 .
Imaging of Chromosome Spread

Imaging is done on a Zeiss Axioplan Microscope. Images are taken with
either 25 or 100 lens and are processed with Openlab software.
Analysis of Chromosome Staining Patterns

The successful experiment that produces 50 intense sites of chromo-
somal labeling is a joy to view, but this joy is soon replaced by the realization
of the daunting task that remains—mapping these sites. The identifi-
cation of the sites of antibody staining on polytene chromosome, even for
the seasoned cytologist, usually requires comparison to reference photo-
graphs and drawings. A place to begin is with the photographs and maps
of the arms and description of landmarks provided by Lefevre.14 This
combination of representative photos and drawings is valuable in making
the initial assignments of labeled bands to a particular chromosomal

arm and region. Comparing the stained chromosomes to photographs of
chromosomal segments at different stages of development in Ashburner15
helps one deal with the variations in banding patterns that occur in normal
development. Defining the developmental stage of particular salivary
gland chromosomal spread early in the process can direct one’s attention
to chromosomal maps representing the same stage. Finally, Sorsa’s book
of ‘‘Chromosome Maps of Drosophila’’ compiles detailed illustrations
(drawings of Bridges and Bridges, photographic maps, and electron micro-
scopic maps) of each of the 102 chromosomal numbered divisions, and is
invaluable for high-resolution mapping.16 We find it convenient to make
Powerpoint slides of separate images and merged images (e.g., antibody
and DNA stain). Lines are drawn to each labeled band in the antibody-
stained image, and then this set of lines is copied and pasted on the
DNA-stained pattern. The DNA-stained bands at and around each line
are identified by comparing to reference maps and written onto the
Powerpoint image as one proceeds, providing a ‘‘working map.’’
Advantages and Limitations

Immunofluorescent staining of polytene chromosomes provides a
highly visual representation of transcription factor distribution in vivo.
The primary advantage of this technique is the ease and quickness with
which the distribution of a particular protein over the entire Drosophila
genome can be assessed. This procedure can be completed in 1 or 2 days,
although the detailed mapping and analysis of labeled sites can take much
longer depending on the complexity of the labeling pattern and the type of
information one needs to extract from the pattern.
The advantage of surveying the entire Drosophila genome ironically is
accompanied by the biggest limitation: resolution. While the method is
about 100 times higher in resolution than comparable analysis of meta-
phase chromosomes, the site of labeling can rarely be assigned to a specific
gene or promoter without further genome manipulations (as in the use of
transgenic lines discussed earlier). As mentioned, the best spatial reso-
lution can be obtained within puffs, yet even in the largest transgenic puffs
14
G. Lefevre, Jr., in ‘‘The Genetics and Biology of Drosophila’’ (M. Ashburner and
E. Novitski, eds.), p. 31. Academic Press, New York, 1976.
15
M. Ashburner, in ‘‘Developmental Studies on Giant Chromosomes’’ (W. Beerman, ed.),
p. 101. Springer, Berlin, 1972.
16
V. Sorsa, ‘‘Chromosome Maps of Drosophila.’’ CRC Press, 1988.
it is only possible to resolve the 50 from the 30 regions of a single gene with
certainty. This may be somewhat improved with a confocal microscope or
by taking a photographic Z-series and using software that can subtract
background fluorescence. If the goal is to map the position of a protein
within a gene with very high resolution, a more appropriate assay is
chromatin immunoprecipitation or ChIP, coupled with assays of the ChIP
DNA on microarrays. In this approach, proteins are crosslinked to chroma-
tin in living cells with formaldehyde, much as they are in polytene chromo-
some fixation. The protein of interest is immunoprecipitated. The DNA
covalently attached to the immunoprecipitated protein is purified, PCR
amplified, tagged with a fluorescent label, and used to probe a microarray
(chip) that contains DNA representing various portions of the gene of
interest. Microarrays containing spotted Drosophila cDNA libraries have
been produced (BD Biosciences, Palo Alto, CA); however, the complete
genome is not yet available.
Another limitation of polytene immunofluorescence analysis is the
challenge of quantifying immunofluorescent signals. The amount of immu-
nofluorescent signal at a particular locus can vary from squash to squash
for several reasons. First, the extent of polytenization can differ in chro-
mosomes derived from the same salivary gland. This could lead to exagger-
ated signals in over-replicated chromosomes and more modest signals in
comparatively under-replicated chromosomes or chromosomal regions.
Second, the preparation of the squash can lead to deformations such as
over-stretching and twisting of the chromosome arms—aberrations that
can in some cases significantly affect the measured amount of fluorescent
signal. Due to the nature of the squashing procedure, these deformations
almost always occur. It is therefore helpful to screen for damaged chromo-
somes with a general DNA stain such as DAPI or Hoechst (see also
Stephens et al.17 for more information on troubleshooting squashes).
Although quantifying immunofluorescent bands presents the above-
mentioned problems, it can be accomplished successfully if the chromo-
somes are carefully prepared and screened for those with crisp,
undistorted banding patterns. We have used NIH Image software to quan-
tify the intensity of signals at specified loci.13 By measuring the signals of
several squashes it is possible to make quantitative conclusions within
acceptable limits of error. Typically, these measurements are borne out
by other experimental approaches, such as ChIP analysis.
Not all chromosomal proteins show highly specific, strong-staining
patterns. While antibodies to many chromosomal proteins that have been
examined produced well-defined patterns of staining that can be readily
17
G. E. Stephens, C. A. Craig, Y. Li, L. L. Wallrath, and S. C. R. Elgin, Meth. Enzymol.
376, 372 (2004).
mapped, some do not (we estimate more than half of the proteins we tested
work well). There are those that show no staining pattern, even when mul-
tiple antibodies are used, and those that produce detectable weak staining
of specific sites and a diffuse staining of chromosomes. In some cases,
examining multiple high-quality chromosome squashes can produce slides
with sufficient signal to noise ratios to allow mapping.18 The range of re-
sponses seen with different proteins and different antibodies can be simply
explained in some cases by the quality of antibodies and abundance of a
protein on a particular site, but in other cases, other factors such as the
efficiency of fixing proteins to chromosomes, availability of the antigen to
chromosome surface, or the ability of antigenic determinants to survive fix-
ation will influence signal strength. Trying more than one slide preparation
protocol can sometimes help solve these problems.
The study of transcription in Drosophila is often aided by the use of
drugs that can inactivate or otherwise regulate the activity of specific
factors. Because the reliable uptake of such drugs by whole flies can pre-
sent dosage problems, it is worthwhile to note that explanted salivary
glands have been successfully treated with drugs in vitro. In this way, the
drug dosage can be precisely controlled and delivered directly to tissues
containing polytene chromosomes. Glands cultured in buffered media
remain physiologically active and are also capable of mounting a heat
shock response. Therefore, it is possible to inactivate or regulate a specific
transcription factor, apply a heat shock treatment or other inducement of
particular gene expression, prepare a chromosome spread, and immuno-
stain to determine the effect on transcription in vivo. This approach has
been used to show that heat shock gene transcription is sensitive to alpha
amanitin.19 The localization of other factors may be indirectly influenced
by inactivating the protein of interest, and this can also be assayed by
immuno-staining.
In summary, immunofluorescent staining of polytene chromosomes is a
reliable, fast, and convenient means of assessing the genome-wide distribu-
tion of chromatin binding proteins. The availability of transgenic flies and
the ability to treat salivary glands with drugs makes it a powerful tool to
investigate the function and dynamics of transcription and other factors
in vivo.
18
J. M. Park, J. Werner, J. M. Kim, J. T. Lis, and Y. J. Kim, Mol. Cell 8, 9 (2001).
19
J. L. Compton and B. J. McCarthy, Cell 14, 191 (1978).
[27] X-chromosome inactivation in mouse embryonic stem cells 405
[27] X-Chromosome Inactivation in Mouse

Embryonic Stem Cells: Analysis of Histone
Modifications and Transcriptional Activity Using
Immunofluorescence and FISH
By Julie Chaumeil, Ikuhiro Okamoto, and Edith Heard
X-chromosome inactivation provides a powerful model system with

which to investigate the different steps in facultative heterochromatin for-
mation. During early development, one of the two X chromosomes is tran-
scriptionally silenced in every cell of a female embryo, thereby achieving
dosage compensation between males and females for X-linked gene prod-
ucts.1 The X inactivation process2 is dependent on the action of a unique
RNA, Xist, which coats the X chromosome in cis and induces its inactiva-
tion (see Brockdroff3 for review). Once established, the inactive state of
the X chromosome is highly stable in somatic cells and is normally only re-
versed in the female germ line. The inactive X is characterized by a number
of features in addition to Xist RNA coating, such as asynchronous replica-
tion timing,4 DNA methylation of promoter regions of house-keeping
genes,5 various modifications of core histone,6–10 and association with a his-
tone variant, macroH2A.11,12 The facultative nature of the X inactivation
process allows the kinetic and functional dissection of the role that these
characteristics play in the establishment and maintenance of inactive chro-
matin. As X inactivation occurs during the earliest differentiation steps
in embryogenesis (3.5–5.5 days postcoitum in the mouse), the number of
1
M. F. Lyon, Nature, 190, 372 (1961).
2
P. Avner and E. Heard, Nat. Rev. Genet. 1, 59 (2001).
3
N. Brockdorff, Trends Genet. 18, 352 (2002).
4
N. Takagi, O. Sugawara, and M. Sasaki, Chromosoma 85, 275 (1982).
5
D. P. Norris, H. N. Brockdorff, and S. Rastan, Mamm. Genome, 1, 78 (1991).
6
7
B. A. Boggs, P. Cheung, E. Heard, D. L. Spector, A. C. Chinault, and C. D. Allis, Nat.
Genet. 30, 73 (2002).
8
T. Jenuwein, Nat. Genet. 30, 77 (2002).
9
J. Silva, W. Mak, I. Zvetkova, R. Appanah, T. B. Nestorova, Z. Webster, A. H. F. M. Peters,
10
K. Plath, J. Fang, S. Mlynarczyk-Evans, R. Cao, K. A. Worringer, H. Wang, C. C. de la
Cruz, A. Otte, B. Panning, and Y. Zhang, Science 300, 131 (2003).
11
C. Costanzi and J. R. Pehrson, Nature 393, 599 (1998).
12
B. P. Chadwick and H. F. Willard, Hum. Mol. Genet. 10, 1101 (2001).

embryonic cells available at these stages can be limiting. Embryonic stem

(ES) cells, which are pluripotent cells derived from the inner cell mass of
blastocysts, provide a useful tissue culture system for studying X inactiva-
tion. ES cells can be maintained in the undifferentiated state and differen-
tiation can be easily induced in vitro. In the case of ES cells with more than
one X chromosome, X inactivation occurs on differentiation. This system
thus allows the successive steps of X inactivation to be followed13–16 and
has been shown to parallel closely the events that occur in vivo.9,10
In this section, we outline techniques for analyzing X-inactivation kin-
etics in differentiating ES cells. We focus particularly on changing patterns
of histone modifications during X inactivation, using immunuofluorescence
combined with RNA FISH on interphase nuclei, as well as metaphase
chromosome staining combined with DNA FISH. We also describe a
protocol for assaying late replication timing of the inactive X chromosome.
The techniques and tools presented should provide a basis for defining
potential causal relationships between different events, not only during X
inactivation but also during the establishment of other patterns of gene
activity.
Background to the Early Events in the X Inactivation Process

In undifferentiated ES cells, all X chromosomes present are active and
Xist is expressed at very low levels, detectable by RNA FISH as a punctate
signal (or ‘‘dot’’) at its site of transcription (see Fig. 1A). The onset of X
inactivation requires an increase in steady-state levels of Xist from the
chromosome that will be inactivated. Thus, when differentiation of female
ES cells is induced, Xist RNA accumulates in cis over the territory of the X
chromosome in interphase nuclei and this ‘‘coating’’ can be detected by
RNA FISH as a domain equivalent to the X-chromosome territory (see
Fig. 1B). This is followed (1–2 days later) by transcriptional silencing of
X-linked genes, based on the observed disappearance of primary transcript
signals detected by RNA FISH.15,17 Using immunofluorescence combined
with RNA FISH, it has been found that X-chromosome–wide changes in
13
A. M. Keohane, L. P. O’Neill, N. D. Belyaev, J. S. Lavender, and B. M. Turner, Dev. Biol.
180, 618 (1996).
14
A. Wutz and R. A. Jaenisch, Mol. Cell 5, 695 (2000).
15
(2001).
16
J. E. Mermoud, B. Popova, A. H. F. M. Peters, T. Jenuwein, and N. Brockdorff, Curr. Biol.
12, 247 (2002).
17
J. Chaumeil, I. Okamoto, M. Guggiari, and E. Heard, Cytogenet. Res. 75, (2002).
[27]
X-chromosome inactivation in mouse embryonic stem cells
Fig. 1. Examples of immunofluorescence combined with RNA FISH on interphase nuclei. Modifications of the N-terminal histone tails
of histone H3 on the inactive X chromosome in differentiating female ES cells. In each case, nuclei are shown with typical patterns observed
in undifferentiated female ES cells (panel A) or during differentiation (panel B). Immunodetection with Alexa GAR 568 conjugated
secondary antibody (red, column 2 of each panel) was combined with Xist RNA FISH (Spectrum Green labeled probe, green, column 3 of
each panel). Three antibodies were used here, which detect di-methylation of H3 lys-4 (from Upstate Biotechnology), di-methylation of H3
lys-9 (gift from D. Allis, also available from Upstate Biotechnology), and di-/tri-methylation of H3 lys-27 (gift from D. Reinberg, see Sarma
and Reinberg, this volume). Prior to the onset of X inactivation, in undifferentiated ES cells, Xist is transcribed at a low level from both X
chromosomes and the primary transcripts can be detected as a ‘‘dot’’ at each Xist locus (green, A). At the beginning of inactivation, Xist
407
RNA starts to accumulate, over the future inactive X chromosome, and can be detected as a green domain in the nuclei (green, B). In this
the modifications status of histones, such as di- and tri-methylation and

hypoacetylation of H3 Lys-9,15,16 tri-methylation of H3 Lys-27,9,10 as well
as hypomethylation of H3 Lys-4, Lys-36, and Arg-17,15,17 are early events
in X inactivation, occurring immediately after Xist RNA coating begins.
The early appearance of these global histone changes on the X chromo-
some undergoing inactivation has been correlated with the Xist RNA-
dependent recruitment of the Polycomb group proteins, Eed, and Enxl
(Ezh2) to the X chromosome,9,10 the latter being a histone methyltransfer-
ase with mainly H3 Lys-27 and some H3 Lys-9 activity in mammalian
cells.18 Other changes, such as late replication timing,4 appear to occur
within a similar window of time,13,17 while DNA methylation of promoter
regions of housekeeping genes13 and association with the histone variant
macroH2A19 appear to be later events. Histone modifications thus repre-
sent strong candidates for the early changes responsible for transcriptional
silencing of the X chromosome. Furthermore, as histone modifications
are found on the X chromosome during mitosis, they may carry part of
the mitotically heritable epigenetic signature that renders the inactive state
clonally stable.
Cell Culture
Culture of Mouse Embryonic Fibroblasts

In order to assess the status of the fully inactive X chromosome in som-
atic cells, female mouse embryonic fibroblasts (MEFs), prepared from
13.5-day embryos (sexed by gonad inspection) can be used.20 These can
be cultured in DMEM with GlutMAX (GIBCO/Invitrogen, cat. no.
31966–021) supplemented with 10% fetal bovine serum (FBS; Invitrogen,
18
Dev. 16, 2893 (2002).
19
T. P. Rasmussen, M. A. Mastrangelo, A. Eden, J. R. Pehrson, and R. Jaenisch, J. Cell Biol.
150, 1189 (2000).
20
B. Hogan, R. Beddington, F. Constantini, and E. Lacy, ‘‘Manipulating the Mouse Embryo:
A Laboratory Manual.’’ Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY,
1994.
same window of time, some histone modifications begin to appear on the inactive X
chromosome (red, B). In merge images of immunofluorescence and RNA FISH (third
column, B), green coloration shows exclusion of the modification or yellow coloration shows
the enrichment on the Xist domain. The inactive X chromosome is depleted in H3 Lys-4
di-methylation (first row) and enriched in H3 Lys-9 di-methylation (second row) and Lys-27
di-/tri-methylation (third row) (B, arrowheads). DNA is stained with DAPI (blue, first column
of each panel).
cat. no. 10270106). It should be noted that only low passage number (no
higher than 4–5) cells should be used if characteristics of primary, somatic
cells wish to be examined.
ES Cell Culture
For detailed protocols on ES cell culture, readers are urged to consult
dedicated source of information, Hogan et al.20 and Robertson.21 For
proper maintenance of ES cell lines in their undifferentiated, pluripotent
state, a well-equipped tissue culture facility and rigorous culture conditions
are required. Some general recommendations include the use of sterile, dis-
posable plasticware or glassware that has never been exposed to detergent
and is kept separate from general laboratory supplies, as ES cells are highly
sensitive to trace levels of detergent. Only high-quality water, classified as
‘‘Type I Reagent Grade Water’’ (ASTM standard), should be used. Fetal
bovine serum should be purchased as ‘‘ES cell grade’’ or various batches
must be tested in parallel for quality control. Tissue culture grade plastic
flasks or plates, or glass coverslips (ESCO, cat. no. 9611301) are used for
ES cell culture and in all cases they have to be first coated in filter-sterilized
0.1% gelatin (Type A, Sigma cat. no. G-2500) in PBS 1X (PBS 10X, GIB-
CO, cat. no. 70013–016; sterile water, Invitrogen, 15230–089). The gelatin
solution is left in contact with the plastic for a minimum of 30 min
(it should not be allowed to dry out) and should be aspirated just prior to

addition of cell medium. We grow ES cells at 37 in 8% CO2.
ES cells can be maintained in the undifferentiated state either through
culture on mitotically inactivated feeder cells (such as mouse embryonic
fibroblasts) and/or in the presence of leukemia inhibitory factor (LIF),
depending on the cell line used. It should be noted that, for X inactivation
studies, ES cells should prefererably be cultured on male rather than
female feeder cells in order to avoid interference during analysis due to
the presence of the inactive X chromosome in female fibroblast cells. For
X inactivation kinetics studies, we have mainly used two female ES cell
lines, LF2 (a gift from Dr. Austin Smith) and PGK12.1 (a gift from Dr. Neil
Brockdorff), both of which grow on gelatin-coated flasks or plates,22 with-
out the need for feeder cells. It should be noted that similar X inactivation
kinetics have been found using feeder-dependent ES cell lines. ES cells are
maintained as an undifferentiated culture in DMEM with GlutMAX, 15%
fetal calf serum (GIBCO, cat. no. 16141–079), 104 mM 2-mercaptoethanol
21
E. J. Robertson, in ‘‘Teratocarcinomas and Embryonic Stem Cells: A Practical Approach’’
(E. J. Robertson ed.), p. 71. IRL Press, Oxford, England, 1987.
22
A. G. Smith, J. Tiss. Cult. Meth. 13, 89 (1991).
(Sigma, cat. no. M7522), and 1000 U/ml LIF (Chemicon, cat. no.
ESG1107). ES cells should be plated and maintained at relatively high
density and passaged at 70–80% confluence. This usually means passaging
every couple of days (cell doubling time can be between 8 and 22 h
depending on the ES cell line and serum used). ES cell colonies should
be monitored daily for density and also any signs of differentiation
(see refs. 20 and 21 for feeder cell–dependent ES cells and Smith22 for
non–feeder-dependent cell lines). The medium should be changed 3 h
prior to trypsinization to increase viability of the cells upon passaging or
cryopreservation. To passage, the flask or plate is rinsed with PBS, then
trypsin-EDTA is added (GIBCO, cat. no. 25200–072). After 8 min in the
incubator, an equal volume of ES cell medium is added and the cells are
dispersed into a single cell suspension by vigorous pipetting (20–40 times)
with a plugged Pasteur pipette, then centrifuged. This resuspension step
is important, as undissociated clumps of ES cells will rapidly form large
colonies and begin to differentiate before the next passage. Cells should
then be plated at 2–4 104 cells/cm2. For freezing, 1 106 cells

should be resuspended in 1 ml of FCS/10% DMSO, then put at 80 O/
N then into liquid nitrogen. The number of passages should always
be kept to a minimum for ES cells in order to avoid genetic and
karyotypic abnormalities. If possible, multiple vials should be frozen at
every passage.
ES Cell Differentiation
Differentiation of ES cells can be induced using a variety of strategies.
The most classical method involves differentiation into embryoid bodies
(EBs) by removal of feeder cells,21 LIF withdrawal, and culture of aggre-
gated (lightly trypsinized) cells in non-adherent Petri dishes (plastic dishes
used for bacterial agar plates are ideal) in DMEM supplemented with 10%
FBS. EBs have to be cultured in suspension, at least for the first 4 days,
which is the period during which X inactivation is established. In order
to perform immunofluorescence and/or RNA FISH on EBs, cytocentrifu-
gation is required, which can disrupt interphase chromatin structure and
nuclear architecture. An alternative strategy (based on ref. 22), which we
prefer as it allows cells to be grown directly on coverslips during differ-
entiation, involves the use of retinoic acid. ES cells are plated at a low
density of 104 cells/cm2 (if necessary after removal of feeder cells by ad-
sorption). Once the majority of cells have attached (this usually takes
8 h–overnight), the ES cell medium is removed, the cells are washed three
times in PBS to eliminate LIF, and differentiation medium is added. Dif-
ferentiation medium is DMEM supplemented with 10% FBS, 100 nM
all-trans-retinoic acid (RA) (Sigma, cat. no. R2625; a 103 M stock solution

is prepared in ethanol and stored at 20 ), and 104 mM 2-mercaptoetha-
nol. Addition of the latter has been found to minimize cell death. Differen-
tiation medium is changed daily and can be preceded by one or two washes
in PBS to remove debris if necessary. Changes in cell morphology can
be detected by day 2, with clearly fibroblast-like and neuronal-like cells
by day 4.
Combined Immunofluorescence and RNA FISH Analysis

Numerous methods involving a variety of fixation and permeabilization
techniques are available for performing immunofluorescence (IF), the
choice depends on the cell type, the epitope, and the antibody being used.23
However, the combination of IF with RNA FISH is not always successful
and depends on the nature of the proteins and/or transcripts being
detected. Furthermore, the conditions that preserve nuclear architecture
and chromatin structure often render RNA FISH ineffecient. Conversely,
conditions that are ideal for RNA FISH (involving extraction in cytoskele-
tal buffer24,25) often result in poor-quality immuno-detection, presumably
through loss of the protein during extraction or damage to the epitope.
Thus, the aim of a combined immunofluorescence and RNA FISH
analysis is to preserve nuclear structure and the antibody’s epitope as far
as possible but at the same time to enable penetration of the FISH probe
for adequate detection of nuclear transcripts. We have tested a number
of different methods and conditions and describe a method we have found
to be optimal for Xist RNA FISH combined with immuno-detection of
histone modifications. We prefer to perform immunofluorescence (under
RNAse-free conditions) prior to RNA FISH, as the formamide treat-
ment during the RNA FISH procedure is sometimes incompatible with
preservation of the epitopes detected by some antibodies.
Immunofluorescence
Sub-confluent fibroblasts or ES cells cultured on gelatin-coated cover-
slips are briefly washed once in PBS and then fixed in freshly prepared
3% formaldehyde (paraformaldehyde, RE/PURO cat. no. 387507, in
PBS, pH 7.2) for 10 min at RT. Following two washes in PBS at RT,
23
D. L. Spector, R. D. Goldman, and L. A. Leinwand, ‘‘Cells: A Laboratory Manual.’’ Cold
Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1998.
24
R. H. Singer, J. B. Lawrence, and C. Villnave, Biotechniques 4, 230 (1986).
25
A. M. Femino, F. S. Fay, K. Fogarty, and R. H. Singer, Science 280, 585 (1998).
permeabilization of the cells is then performed on ice, in PBS containing

0.5% Triton X-100 (ICN, cat. no. 807423), 2 mM Vanadyl Ribonucleoside
Complex (an RNase inhibitor, NEB, cat. no. S1402S) for 3–5 min. The
exact time of permeabilization depends on the antibody and cell type. It
should be noted that shorter permeabilization times (less than 3 min)
results in less efficient RNA FISH. After rinsing in PBS, the cells are incu-
bated in 1% BSA (GIBCO, cat. no. 15260–037) in PBS for 15 min, to block
non-specific sites of antibody binding. The primary antibody, diluted in
blocking buffer and 0.4 U/l RNAguard (RNase inhibitor, Amersham/
Pharmacia, cat. no. 270815–01), is then applied. The coverslips are placed
cell-side down, avoiding the formation of air bubbles, onto a drop of anti-
body solution on a sterile glass slide. The volume depends on the size of
coverslip used (we routinely use 18 18 mm coverlips and 35 l of anti-
body solution). Incubation is performed in a humid chamber at room tem-
perature for 45 min (the temperature and length of incubation can vary
between antibodies). The coverslips are then carefully removed with for-
ceps. If any resistance is encountered, the coverslip should be flooded with
PBS so that it floats in order to avoid damaging the cells. The coverslip is
quickly put back into a well containing PBS and four or more washes in
PBS, of 5 min each, are performed with rotational shaking. Blocking and
incubation in secondary antibody is then performed following the same
procedure as for the primary antibody earlier. After washing in PBS, prep-
arations can be postfixed in 3% paraformaldehyde for 10 min at RT and
rinsed twice in 2 SSC prior to RNA FISH. For immunofluorescence
alone, postfixation and SSC washes are unnecessary and the sample can
be DAPI stained, washed, and mounted for observation as described later.
Note that the sample should never be allowed to dry at any stage, as this
can lead to background problems and artifacts.
The specificity of the primary antibodies used should be verified using
western blotting. For detailed protocols on the handling of antibodies
and controls of specificity the reader is referred to Harlow and Lane.26 A
list of the antibodies directed against histone modifications that we have
used (most of which are available from Upstate Biotechnology) and their
dilutions are listed in Table I.
Secondary antibodies must be targeted against the appropriate class of
immunoglobulin of the species in which the primary was made and gener-
ated from a separate species. A control of the specificity of the secondary
antibody should always be included (i.e., an IF without prior primary anti-
body incubation). For combined IF and RNA or DNA FISH, the choice of
26
E. Harlow and D. Lane, ‘‘Antibodies: A Laboratory Manual.’’ Cold Spring Harbour
Laboratory Press, Cold Spring Harbor, NY, 1989.
TABLE I
Enriched (þ)
depleted ()
Antibody on the Xi Dilution Source
Di-me(Lys4) H3 1/300 Upstate Biotechnology

Ac(Lys9) H3 1/300 Upstate Biotechnology
Di-me(Lys9) H3 þ 1/300 Gift from D. Allis/
Upstate Biotechnology
Di-me(Arg17) H3 1/300 Upstate Biotechnology
Di-/tri-me(Lys27) H3a þ 1/50 Gift from D. Reinberg
Di-me(Lys36) H3 1/300 Upstate Biotechnology
Ac(penta) H4 1/300 Upstate Biotechnology
a
Note: The reader is referred to the chapter by Samra and Reinberg for characterization of
the anti-H3 di-/tri-methylated Lys-27 antibody.
fluorochrome to which the secondary antibody is conjugated will depend

on the fluorochrome with which the FISH probe is labeled and on the filter
sets available on the microscope. For information on the choice of fluoro-
phores the reader is referred to catalogs of manufacturers. We commonly
use Molecular Probes (e.g., Alexa Fluor 568 or 488, goat anti-rabbit or goat
anti-mouse IgGs) for combination with Spectrum Red or Spectum Green
labeled probes, respectively (see later). In the case of triple labelings, in
which two different primary antibodies are used in conjunction with
RNA FISH, or else one antibody but two different transcripts are detected
by RNA FISH, fluorophores in the infrared range of the spectrum can be
used, such as Alexa 680, if the appropriate filter sets are available (note
that these fluorophores are not visible to the human eye). In the case of a
double IF experiment, high-affinity purified secondary antibodies should
be used which minimize cross-species reactivity, and the appropriate con-
trols (i.e., each primary antibody alone with both secondaries) should
always be performed.
RNA Fluorescence In Situ Hybridization

For a general description and discussion of RNA FISH protocols, the
reader is referred to Spector et al.23 To detect the primary transcripts
of genes, genomic probes several kilobase pairs long or fluorescently
labeled oligos must be used. Probes spanning introns and exons will detect
both the processed mRNA and the primary transcript. Oligonucleotides

within intronic sequences will clearly be specific for the primary transcript.
For Xist RNA detection we have used several genomic DNA probes, span-
ning a minimum of 3 kb with success. Labeling of DNA probes to be used
for in situ detection of transcripts can be performed either by nick transla-
tion or random priming or else fluorescently tagged oligonucleotides can be
used. The latter avoids the labeling step and also enables discrimination be-
tween sense or antisense transcripts (double-stranded DNA probes will of
course detect both), but is costly. We routinely use nick translation (Vysis,
cat. no. 32-801300) with fluorescent dUTPs (Spectrum Green, Spectrum
Red, Vysis, cat no. 6J9410-6J9420 or Cy5 Amersham, cat. no. PA55026)
to label genomic probes (plasmids, lambda clones, or BACs). 1–2 g of
DNA is labeled in a 50-l reaction. Following the reaction, the size range
of the labeled DNA is always checked by electrophoresis on an agarose gel.
The optimal range size of the FISH probe is between 50 and 200 bp, short
enough to enter the nucleus and long enough to be specific. Fluorescently

labeled probes of this kind can be directly stored at 20 for a few weeks.
Prior to hybridization, 0.1 g of probe (usually 5 l of a standard nick
translation reaction of 50 l) is precipitated with 10 g of salmon sperm
(per 18 18 mm coverlip) and washed twice in 70% ethanol (to remove
unincorporated nucleotides). The dried pellet is thoroughly resuspended
in 5 l formamide (Sigma, cat. no. F5786) by vortexing and denaturation

is then performed at 75 for 7 min. Following this 5 l of a 2 hybridization
buffer is immediately added, to give a final solution of 2 SSC, 20% dex-
tran sulfate, 1 mg/ml BSA (NEB, cat. no. B9001S), 200 mM VRC. The
10 l of probe solution is deposited onto a sterile glass slide and the cover-
slip is removed from the 2 SSC solution with forceps and carefully
lowered cell-side down avoiding the formation of air bubbles. Once the
coverlip has made contact with the probe solution, it should not be moved

to avoid damaging the cells. Hybridization is overnight at 37 in a dark,
humid chamber (paper tissues soaked in 50% formamide/2 SSC). After
three washes in 50% formamide/2 SSC (adjusted to pH 7.2) and three

washes in 2 SSC at 42 , DNA is counterstained for 2 min in 2 SSC con-
taining 0.2 mg/ml DAPI, followed by a final wash in 2 SSC. Coverslips
are then mounted in glycerol-based medium (90% glycerol (Sigma, cat.
no. G5516), 0.1 PBS, 0.1% p-phenylenediamine (Sigma, cat. no. 27–
515–8), pH 9). We use a Leica DMR fluorescence microscope with a Cool
SNAP fx camera (Photometrics) and Metamorph software (Roper) for
image acquisition. An example of Xist RNA FISH combined with
immunofluorescence is shown in Fig. 1.
Immunofluorescence and DNA FISH on Metaphase

Chromosome Spreads
Immunofluorescence
The technique we describe involves IF performed on unfixed meta-
phase chromosomes and is based on a previously published protocol of
ref 13. Metaphase chromosomes are prepared from cells at 80% conflu-
ence, following a 1-h incubation with 0.1 g/ml colcemid (Roche,
295892). To maximize the collection of mitotic cells the supernatant is col-
lected and adherent cells are trypsinized, pelleted, resuspended in PBS, and
pelleted (this wash performed twice). The cells are then resuspended (105
cells/ml) carefully in 75 mM KCl, incubated for 10 min at RT and then
placed on ice for between 5 min and 1 h. Cytocentrifugation is then per-
formed (Shandon), with the time and speed of centrifugation depending
on the cells used. For ES cells and early stages of differentiation, 8 min
at 1200 rpm is usually appropriate. After air-drying, the slides are incu-
bated in freshly prepared KCM buffer (120 mM KCl, 20 mM NaCl,
10 mM Tris-HCl, pH 8, 0.5 M EDTA, 0.1% Triton X-100) for 15 min at
RT, and incubated successively with primary and secondary antibody in
KCM, 1% BSA for 40 min in a humidified dark chamber at RT (immuno-
fluorescence protocol as earlier). Several washes in KCM buffer are per-
formed. A postfixation step in 3% formaldehyde can be performed at this
stage to prevent signal loss if this is a problem. The slides should again be
washed in KCM containing DAPI (0.2 mg/ml DAPI) for 2 min, washed
once more in KCM and mounted in medium (see earlier) for visualization
on a fluorescence microscope. An example is shown in Fig. 2.
DNA Fluorescence In Situ Hybridization

Xist RNA does not associate with the mitotic X chromosome efficiently
during the early stages of ES cell differentiation (J. Chaumeil and E.
Heard, unpublished observations). Therefore, in order to identify the X
chromosome on mitotic chromosome spreads, DNA FISH using an
X-chromosome paint can be performed following immunofluorescence.
This does not of course allow the inactive X chromosome to be identified.
A double immunofluorescence of a mark known to characterize the in-
active X chromosome (such as H3 Lys-4 hypomethylation) could therefore
be used. As the DNA denaturation step can destroy the immunofluores-
cence signal in some cases, images should be recorded prior to the DNA
FISH experiment. To remove the coverslip and the mounting medium from
the sample following imaging, slides are washed 3–4 times in KCM (if the
coverslip does not fall off by itself during the first wash, it should carefully
416
Fig. 2. Example of immunofluorescence on metaphase spreads. Histone H3 modifications on the inactive X chromosome in
differentiating female ES cells. Metaphase spreads are shown with typical patterns observed in female ES cells from day 3 of differentiation
onward. A single X chromosome shows specific hypomethylation on H3 Lys-4 (red, arrowhead). X chromosomes were detected by X
chromosome DNA FISH following immunofluorescence (yellow). The distal part of the X chromosome is methylated on H3 Lys-4 and
probably corresponds to the pseudoautosamal region (asterisk). DNA is stained with DAPI (blue).
[27]
be removed using forceps before the second wash). The samples should
then be dehydrated through a series of 3-min washes in 70%, 80%, and
95% ethanol. Denaturation of the DNA is performed by plunging the slide
in a 50-ml falcon tube or coplin jar containing 70% formamide/2 SSC (pH
7.2) for 3 min at 75 . The slides should immediately be plunged into ice-
cold 70% ethanol and then dehydrated (as earlier). Denaturation of
chromosome paint probes and pre-annealing of competitor DNA as well
as in situ hybridization are performed exactly as recommended by the sup-
plier (mouse X-chromosome paint, Cambio, cat. no. 1187-XMB-02). Hy-

bridization is performed overnight at 42 . Posthybridization washes are at

45 , three times in 50% formamide/2 SSC and three times in 2 SSC.
Slides are counter-stained with DAPI, mounted, and viewed under the
fluorescence microscope (as earlier). An example is shown in Fig. 2.
Replication Timing Studies Using BrdU Incorporation and Acridine

Orange Staining
In order to determine the onset of late replication timing of the inactive
X chromosome, we have tested a number of techniques, all of which
depend on the incorporation of the thymidine analog, 5-bromo-2-deoxyur-
idine (BrdU), to detect replicated DNA. The technique that proved most
robust in our hands is in fact one of the most classic techniques to have
been used for studies of replication timing and involves acridine orange
staining.4,27,28 In this method, DNA in which BrdU has been incorporated
into DNA will stain weakly red, whereas DNA with no BrdU incorporation
will stain green. This technique has the advantage that both BrdU-negative
and BrdU-positive regions can be simultaneously visualized, allowing iden-
tification of mitoses with an overall ‘‘late’’ replicating banding pattern (i.e.,
red G-bands and green R-bands). Undifferentiated or differentiating ES
cells are incubated for 8–9 h in medium supplemented with 100 g/ml
BrdU (Sigma). Colcemid (Roche) at a concentration of 0.025 g/ml is
added to the medium for the final 0.5–1 h of incubation. The cell suspen-
sion is centrifuged for 2 min at 400g. The pellet is gently resuspended in
2 ml of freshly prepared 0.075 M KCl. The cell suspension is then incu-
bated at RT for 10 min and 3 ml of freshly prepared fixative (3:1 methanol:
acetic acid) is then added directly to the hypotonic medium and incubated
for a further 10 min. The cell suspension is then centrifuged for 5 min at
150g and the pellet is dispersed in 10 ml of fixative. After a second centri-
fugation of 2 min at 400g, the pellet is dispersed in 1–2 ml of fixative and
27
B. Dutrillaux, C. Laurent, J. Couturier, and J. Lejeune, Acad. Sc. Paris 276, 3179 (1973).
28
N. Takagi and M. Oshimura, Exp. Cell. Res. 78, 127 (1973).
the tubes are kept on ice. Clean and dry slides are placed on wet tissues
(Kimwipes, Kimberly-Clark). The fixed cell suspension is spread by drop-
ping one drop onto an ethanol-cleaned slide from a height of about
10 cm. The slides are allowed to dry very quickly. They are then stained
with Sorensen’s phosphate buffer (pH 6.8) supplemented with 0.05 mg/ml
acridine orange solution. The slides are incubated for 10–15 min in this so-
lution, rinsed immediately in running tap water for 1–2 s, then rinsed in
phosphate buffer (pH 6.8). They are mounted with a coverslip in phosphate
buffer (pH 6.8) and excess buffer is removed with blotting paper. The
metaphase spreads are then immediately examined under a fluorescence
microscope4,27,28 using a filter block for acridine orange (blue filter, exc.
490 nm, emiss. 530 nm/640 nm for single- and double-stranded DNA
detection). Whole chromosomes or chromosome segments replicated in
the presence of BrdU show a weak red fluorescence whereas, those that
have not incorporated BrdU show bright green fluorescence. The auto-
somes and the active, synchronously replicating X chromosome have
certain regions that replicate late in the S phase of the cell cycle as well
as others that replicate early. When BrdU is present only during the latter
half of S phase, an unequivocal red and green banding pattern is observed
in these chromosomes in contrast with the late-replicating inactive
X chromosome, which is stained homogeneously red. An example is shown
in Fig. 3.
Fig. 3. Detection of asynchronous replication of the X chromosomes using Acridine

Orange staining. Metaphase spreads showing the typical green/orange banding pattern
representative of BrdU incorporation in the latter half of S phase are shown. In (A) no single
orange-staining X chromosome can be detected, indicating that both X chromosomes
replicate synchronously. The first sign of a single uniformly stained orange chromosome (the
late-replicating X chromosome, see arrow) was observed at day 3 of ES cell differentiation,
where 6.6% (n > 200) of metaphases displayed this pattern. The proportion rose to almost
30% (n > 200) by day 7.
[28] X inactivation in mouse ES cells 419
Conclusion
The techniques we describe here should provide a basis for investigat-
ing the onset of X inactivation or other developmentally regulated pro-
cesses in differentiating ES cells. Using these techniques we have been
able to determine the relative order of the following events: Xist RNA
coating occurs within the first 24–48 hours of differentiation; histone H3
modifications such as hypomethylation of Lys-4, hypoacetylation of Lys-
9, and hypermethylation of Lys-9 and Lys-27 (see Fig. 1) are detectable
on the X chromosome in a proportion of interphase cells as soon as Xist
RNA accumulates. Histone H4 hypoacetylation of lysines 5, 8, 12, and 16
is also observed from this time, although with slightly slower kinetics.15,17
The appearance of these marks on the X chromosome during mitosis (see
Fig. 2) is found from days 2–3, suggesting that they represent true epige-
netic marks. On the other hand, a late-replicating X chromosome is only
detected from day 3.17 Thus, histone modifications appear to be direct con-
sequences of Xist RNA’s action, while late replication timing seems to be a
slightly later event. Using this as a basis, the relationship between these dif-
ferent events can now be defined and tested, using for example, genetically
modified ES cells or siRNA-induced silencing of candidate modifying
enzymes.
[28] X Inactivation in Mouse ES Cells:

Histone Modifications and FISH
By Barbara Panning
In mammals, X inactivation is used to equalize X-linked gene dosage

between females, with two X chromosomes, and males, with only one,
and it provides a particularly dramatic example of gene regulation by alter-
ations in chromatin structure. X inactivation can be divided into several
phases: choice, initiation, spread, and maintenance.1 At around the time
of gastrulation, when totipotent embryonic cells differentiate into more
restricted somatic lineages, one X chromosome per diploid genome is
randomly chosen to remain active and all additional X chromosomes are
altered from euchromatin to transcriptionally inactive heterochromatin.
Once an X chromosome has been inactivated, it remains silenced in all
1
K. Plath, S. K. Mlynarczyk-Evans, D. A. Nusinow, and B. Panning, Annu. Rev. Genet. 36,
233 (2002).

Conclusion
The techniques we describe here should provide a basis for investigat-
ing the onset of X inactivation or other developmentally regulated pro-
cesses in differentiating ES cells. Using these techniques we have been
able to determine the relative order of the following events: Xist RNA
coating occurs within the first 24–48 hours of differentiation; histone H3
modifications such as hypomethylation of Lys-4, hypoacetylation of Lys-
9, and hypermethylation of Lys-9 and Lys-27 (see Fig. 1) are detectable
on the X chromosome in a proportion of interphase cells as soon as Xist
RNA accumulates. Histone H4 hypoacetylation of lysines 5, 8, 12, and 16
is also observed from this time, although with slightly slower kinetics.15,17
The appearance of these marks on the X chromosome during mitosis (see
Fig. 2) is found from days 2–3, suggesting that they represent true epige-
netic marks. On the other hand, a late-replicating X chromosome is only
detected from day 3.17 Thus, histone modifications appear to be direct con-
sequences of Xist RNA’s action, while late replication timing seems to be a
slightly later event. Using this as a basis, the relationship between these dif-
ferent events can now be defined and tested, using for example, genetically
modified ES cells or siRNA-induced silencing of candidate modifying
enzymes.
[28] X Inactivation in Mouse ES Cells:

Histone Modifications and FISH
By Barbara Panning
In mammals, X inactivation is used to equalize X-linked gene dosage

between females, with two X chromosomes, and males, with only one,
and it provides a particularly dramatic example of gene regulation by alter-
ations in chromatin structure. X inactivation can be divided into several
phases: choice, initiation, spread, and maintenance.1 At around the time
of gastrulation, when totipotent embryonic cells differentiate into more
restricted somatic lineages, one X chromosome per diploid genome is
randomly chosen to remain active and all additional X chromosomes are
altered from euchromatin to transcriptionally inactive heterochromatin.
Once an X chromosome has been inactivated, it remains silenced in all
1
K. Plath, S. K. Mlynarczyk-Evans, D. A. Nusinow, and B. Panning, Annu. Rev. Genet. 36,
233 (2002).

subsequent generations. As a result, every female is a mosaic of clonal

groups of cells in which either the maternally inherited X chromosome or
the paternally inherited X chromosome is silenced. This indicates that the
repressed chromatin structure of the inactive X chromosome (Xi) is extra-
ordinarily stable, since it can be propagated through all the cell divisions
in the lifetime of an individual. Thus, the study of X inactivation addresses
two equally interesting issues in the study of gene expression. First, what
activities are used to alter chromatin structure and how are these activities
restricted to particular portions of the genome? Second, how are different
chromatin states stably maintained through cell division?
The Xist gene is necessary for X inactivation as X chromosomes bearing
Xist deletions are not silenced.2,3 Xist encodes a 17.5-kb, spliced, polyade-
nylated, non-coding RNA that is expressed exclusively from the Xi in
female somatic cells.4–6 Xist RNA is unusually stable and remains in the
nucleus where it coats the Xi.7–9 In embryonic cells that are poised to
undergo X inactivation Xist RNA can be detected only at its site of
transcription on the single active X chromosome in male cells and
from both Xa’s in female cells.10 When embryonic cells are differentiated
and X inactivation is initiated, Xist transcripts spread in cis from the
Xist gene to coat the entire X chromosome. The spread of Xist RNA
correlates temporally with the spread of silencing along the X chromosome
and occurs within a single cell cycle.11–13 The Xi in female somatic cells
is characterized by increased methylation of CpG islands,14 a decrease in
2
G. D. Penny, G. F. Kay, S. A. Sheardown, S. Rastan, and N. Brockdorff, Nature 379, 131
(1996).
3
Y. Marahrens, B. Panning, J. Dausman, W. Strauss, and R. Jaenisch, Genes Dev. 11, 156
(1997).
4
G. Borsani, R. Tonlorenzi, M. C. Simmler, L. Dandolo, D. Arnaud, V. Capra, M. Grompe,
A. Pizzuti, D. Muzny, C. Lawrence et al., Nature 351, 325 (1991).
5
N. Brockdorff, A. Ashworth, G. F. Kay, P. Cooper, S. Smith, V. M. McCabe, D. P. Norris,
G. D. Penny, D. Patel, and S. Rastan, Nature 351, 329 (1991).
6
C. J. Brown, A. Ballabio, J. L. Rupert, R. G. Lafreniere, M. Grompe, R. Tonlorenzi, and
H. F. Willard, Nature 349, 38 (1991).
7
C. J. Brown and S. E. Baldry, Somat. Cell Mol. Genet. 22, 403 (1996).
8
C. M. Clemson, J. C. Chow, C. J. Brown, and J. B. Lawrence, J. Cell Biol. 142, 13 (1996).
9
C. M. Clemson, J. A. McNeil, H. F. Willard, and J. B. Lawrence, J. Cell Biol. 132, 259
(1996).
10
B. Panning and R. Jaenisch, Genes Dev. 10, 1991 (1996).
11
B. Panning, J. Dausman, and R. Jaenisch, Cell 90, 907 (1997).
12
S. A. Sheardown, S. M. Duthie, C. M. Johnston, A. E. Newall, E. J. Formstone, R. M.
Arkell, T. B. Nesterova, G. C. Alghisi, S. Rastan, and N. Brockdorff, Cell 91, 99 (1997).
13
A. Wutz and R. Jaenisch, Mol. Cell 5, 695 (2000).
14
S. F. Wolf, D. J. Jolly, K. D. Lunnen, T. Friedmann, and B. R. Migeon, Proc. Natl. Acad.
Sci. USA 81, 2806 (1984).
abundance of acetylated isoforms of core histones,15 changes in histone H3

methylation patterns,16–19 an increase in amounts of the histone H2A vari-
ant, macroH2A1.2,20 replication late in S phase,21 and localization at
the nuclear periphery.22 X inactivation occurs in mouse embryonic stem
(ES) cells when they are induced to differentiate in culture, providing
a model system in which to assay the timing of alterations in chromatin
structure relative to the cis-spread of Xist RNA that occurs at the onset
of X inactivation. Fluorescence in situ hybridization (FISH) for Xist
RNA when combined with immunostaining provides a powerful tool to
analyze the alterations in chromatin structure that occur during X chromo-
some silencing. Figure 1 provides an example results that can be achieved
using immunostaining and FISH in differentiating female ES cells.
Differentiation of ES Cells
For X inactivation studies, differentiation of ES cells into embryoid
bodies (EBs) using hanging drop cultures is very effective. When the
resulting EBs are placed on a tissue culture substrate coated slides the
EBs attach to the slides. The cells that delaminate and migrate away from
the EBs can be analyzed at various intervals after plating to visualize
the alterations in chromatin structure that occur during X chromosome
silencing.
1. Trysinize and pellet female ES cells.
2. Resuspend the cell pellet in 1 ml of differentiation media.
Differentiation media:
DMEM
15% fetal bovine serum
Non-essential amino acids
0.1 mM -mercaptoethanol
15
16
(2001).
17
J. E. Mermoud, B. Popova, A. H. Peters, T. Jenuwein, and N. Brockdorff, Curr. Biol. 12,
247 (2002).
18
19
K. Plath, J. Fang, S. K. Mlynarczyk-Evans, R. Cao, K. A. Worringer, H. Wang, C. C.
de la Cruz, A. P. Otte, B. Panning, and Y. Zhang, Science 300, 131 (2003).
20
C. Costanzi and J. R. Pehrson, Nature 393, 599 (1998).
21
N. Takagi, Exp. Cell Res. 86, 127 (1974).
22
M. L. Barr, C. D., Acta Cytol. 6, 34 (1961).
422
Fig. 1. Immunostaining for Eed and FISH for Xist RNA in differentiating female ES cells. In a differentiating population of female ES
cells, a fraction of cells exhibit Xist RNA coating of the Xi (middle panel) in nuclei, visualized DAPI staining (left panel). In these cells Eed,
a component of a Polycomb Group complex that includes the histone methyltransferase Ezh2, is also enriched on the Xi (right panel).
Recruitment of the Eed-Ezh2 complex correlates with an increase in histone H3 lysine 27 methylation on the Xi.18,19 In the cells that do not
yet show Xist RNA coating, the levels of Eed are much higher, suggesting that abundance of Eed decreases when ES cells are differentiated.
[28]
3. Determine cell number, and dilute cells to a final concentration of

10 cells per microliter.
4. Place 100–200 10–20 l drops of diluted ES cells on the inside of a
Petri dish lid.
5. Return the lid of the Petri dish to its base; the base should be filled
with PBS to hydrate the hanging drops.
6. Incubate hanging drops for 2 days.
7. After 2 days flood the lid of the Petri dish with a small volume of
differentiation media to suspend the EBs forming in the hanging
drops.
8. Transfer the EBs to a Petri dish filled with differentiation media.
9. Incubate the EBs in suspension culture for 2 more days.
10. The EBs should now be large enough to be seen by eye. Use a P20
pipetman to remove individual EBs from the suspension culture
and plate them on commercially available slides that have tissue
culture coatings. Multichamber slides can be very useful if several
antibodies or FISH probes will be employed. The EBs will adhere
to the slide in approximately 24 h.
11. At 24-h intervals after plating fix samples for immunostaining and
FISH as described later.
Immunostaining
To simultaneously visualize protein and RNA, immunostaining is
followed by FISH. Many antisera contain RNases, resulting in loss of RNA
during the immunostaining procedure. Addition of tRNA and RNase in-
hibitors to the immunostaining buffers to suppress the RNases in the sera
preserves RNA during immunostaining.
1. Rinse slides with 1 PBS.
2. Incubate slides for 30 s in ice-cold cytoskeletal buffer.
Cytoskeletal buffer:
100 mM NaCl
300 mM sucrose
3 mM MgCl2
10 mM PIPES, pH 6.8
3. Incubate slides for 30 s in ice-cold cytoskeletal buffer þ 0.5%
Triton X-100.
4. Incubate slides for 30 s in ice-cold cytoskeletal buffer.
5. Fix in 4% paraformaldehyde/1 PBS, 10 min, room temperature.
6. Wash slides two times, 5 min each in room temperature PBS-

Tween. Slides can be stored for several days in PBS-Tween at 4
prior to immunostaining.
PBS/Tween
1 PBS
0.2% Tween 20

7. Block slides for 30 min at 37 in blocking buffer in a humid
chamber. The humid chamber should contain PBS/Tween.
Blocking buffer:
1 PBS
5% goat serum (heat inactivated)
0.2% Tween 20
0.2% fish skin gelatin
4–5 units/ml RNAsin
1 mg/ml molecular biology-grade yeast tRNA

8. Incubate with primary antibody diluted in blocking buffer, 1 h, 37
in a humid chamber. Temperature may be varied depending on the
primary antibody.
9. Wash slides three times, 5 min each with PBS/Tween.

10. Block slides for 5 min at 37 in blocking buffer in humid chamber.
11. Incubate with secondary antibody diluted in blocking buffer for

30 min, 37 in a humid chamber.
12. Wash slides three times, 5 min each with PBS/Tween. Minimize
exposure to light to limit photobleaching.
13. Fix in 2% paraformaldehyde/1 PB for 10 min at room tempera-
ture.
14. Transfer slide to 70% ethanol. Slides can be stored in 70% ethanol

at 4 for several weeks prior to carrying out FISH.
FISH
The FISH procedure involves generating a labeled probe, hybridizing
the probe to the fixed and immunostained sample, and then detection
of the labeled probe. The specific activity of RNA probes is highest, and
since the immunostaining may cause a slight reduction in the amount of
RNA in the sample, using high specific activity probe generates the best
results with FISH. All stock solutions are autoclaved to minimize likeli-
hood of RNase contamination, and all stock solutions are diluted in auto-
claved water.
Generate Template
Templates can be generated by cloning fragments into plasmids with
T3, T7, or SP6 promoters or by appending T3, T7, or SP6 promoters to
PCR products. The maximum size of an in vitro transcribed RNA that will
penetrate into a fixed, permeabilized cell efficiently is approximately 500

nucleotides. Therefore, multiple 500 nucleotide RNAs spanning the length
of the cDNA will allow for the most robust detection of any RNA.
Prepare Template
If using plasmid template, digest with appropriate restriction enzyme to
linearize. Phenol-chloroform extract, ethanol precipitate twice, and dry.
Resuspend in water at a concentration of approximately 1 mg/ml.
If using PCR product, set up PCR reaction and run fraction out on gel.
If the band is clearly visible, then about 1 l of PCR product is enough for the
in vitro transcription reaction. It does not need any further manipulation.
In Vitro Transcription
1. Mix the following at room temperature:
5 buffer 20 l
100 mM DTT 10 l
4000 units/ml RNA sin 2.5 l
2.5 mM A, G, UTP mix 20 l
10 mM CTP 1 l
10 mM bioCTP 4 l
DNA, 1 mg/ml 2 l
Polymerase 2 l
Water 38.5 l

2. Incubate at 37 for 1 h.
3. Add 1 l of 1 U/l RNase free DNase 1.
4. Remove unincorporated nucleotides using G-50 Sephadex spun
columns, prepared with TE.
5. After RNA has spun through column, run 5 l on agarose minigel to
ensure that in vitro transcription has worked. The remaining RNA
should be ethanol precipitated with 10 g of molecular biology-

grade yeast tRNA and stored at 80 or 20 under ethanol.
Prepare Probe Mix

1. Aliquot 1/20 of the biotinylated RNA produced from a 2-g
labeling reaction.
2. Add:
10 l of molecular biology-grade yeast tRNA, 10 mg/ml
15 l sheared boiled salmon sperm DNA, 10 mg/ml
10 l 3 M Na acetate
45 l water
250 l 100% ethanol
3. Mix well and pellet.

4. Wash precipitated RNA with 70%, then 100% ethanol.
5. Dry pellet in speed vacuum to remove all traces of ethanol.
6. Resuspend pellet in 25 l of ultrapure formamide (Sigma). Ensure
that the pellet is completely dissolved by using pipette tip to break
up any large pieces and vortexing vigorously.
7. Mix with 25 l of 2 hybridization buffer.
2 hybridization buffer:
1 part 20 SSC
2 parts 10 mg/ml BSA, molecular biology-grade (New England
Biolabs)
2 parts 50% dextran sulphate

Mix very well, store at 20

8. Denature by incubation at 95–100 for 5–10 min.
9. Quench on ice immediately.

10. Store at 20 .
FISH
1. Dehydrate cells on slides through an ethanol series, 2 min each in
85%, 95%, and 100% ethanol, air-dry.
2. Add probe to air-dried cells.
3. Place coverslip over probe. For a 12-mm circle coverslip use 3 l of
probe; for a 22-mm square coverslip use 10 l.

4. Incubate in a humid chamber at 37 overnight. The humid chamber
should contain 2 SSC/50% formamide.
Washing and Detection

1. Immerse slide with coverslip in 2 SSC/50% formamide. Once the
slide is wet carefully slide off the coverslip and continue washing
for 5 min.

2. Wash in 2 SSC/50% formamide, once more for 5 min each at 39 .

3. Wash in wash buffer II, three times for 5 min each at 39 .
Wash buffer II:
0.5 M NaCl
10 mM Tris, pH 7.5
0.1% Tween-20
4. Dilute RNase A to 25 g/ml in wash buffer II.
5. Put 150 l of RNase A in wash buffer II on a 24 60 mm coverslip.
6. Place slide, cell-side down over coverslip and use capillary action to
draw the coverslip to the slide.
7. Flip slide and place coverslip-side up in humid chamber filled with

wash buffer II for 30 min.

8. Wash in wash buffer II, two times for 5 min each at 39 .

9. Wash in 2 SSC/50% formamide, two times for 5 min each at 39 .

10. Wash in 2 SSC, three times for 5 min each at 39 .

11. Wash in 1 SSC, two times for 5 min at 39 .

12. Wash in 4 SSC, once for 5 min at 39 .
13. Dilute fluorochrome-conjugated avidin in detection buffer, 1:200.
Detection buffer:
1 part 20 SSC
1 part BSA, 10 mg/ml, molecular biology grade (New England
Biolabs)
3 parts water
14. Put 150 l of detection buffer with avidin fluorochrome on 24
60 mm coverslip.
15. Place slide, cell-side down, over coverslip and use capillary action
to draw coverslip up to slide.
16. Flip slide and place coverslip-side up in humid chamber filled with
4 SSC.

17. Incubate for 30 min at 37 .

18. Wash in 4 SSC, for 5 min each at 39 .

19. Wash in 4 SSCþ 0.2% Tween with DAPI, for 5 min at 39 .
DAPI:
Stock of 1 mg/ml, dissolved in dimethyl formamide. Use 2 l per
50 ml of 4 SSC.

20. Wash in 4 SSC for 5 min at 39 .
21. Add antifade mounting media. Any commercially available
antifade media is appropriate.
22. Place a 24 60 mm coverslip on the slide.
23. Blot off excess mounting media.

24. Slides may be visualized immediately or stored at 4 . If slides will
be stored for longer than a week, use nail polish to seal the
coverslip to the slide.
25. Gather images on a fluorescent microscope equipped with
appropriate filter sets and a camera.
Troubleshooting
Generally it is best to test the FISH probe on fixed material prior to
doing immunostaining in combination with FISH. Follow the first five fix-
ation steps of the immunostaining protocol. After the paraformaldehyde
fixation, transfer slides to 70% ethanol. Follow the FISH protocol. If the
FISH does not work, then troubleshooting of the probe and solutions is
necessary. Introduction of RNases is one likely cause of failure, and using
freshly prepared, autoclaved solutions can solve this problem. The amount
of labeled probe that is best for detection of different RNAs varies and
testing a range of concentrations of labeled probe can be used to identify
an optimal probe concentration.
RNA has been implicated in regulating chromatin structure at centro-
meres23–25 and in recruiting chromatin modifying enzymes to the dosage
compensated X chromosome in Drosophila,26,27 indicating that RNA
may be more generally used to regulate chromatin structure. Thus, the
techniques described in this chapter may be relevant to the study of other
examples of regulated changes in chromatin structure.
Acknowledgments
Many thanks to Hannah Cohen, Susanna Mlynarczyk-Evans, and Kathrin Plath for
critical reading of this manuscript.
23
T. Jenuwein, and G. Almouzni, Nature Genetics 30, 329 (2002).
24
T. A. Volpe, C. Kidner, I. M. Hall, G. Teng, S. I. Grewal, and R. A. Martienssen, Science
297, 1833 (2002).
25
I. M. Hall, G. D. Shankaranarayana, K. Noma, N. Ayoub, A. Cohen, and S. I. Grewal,
Science 297, 2232 (2002).
26
V. H. Meller, K. H. Wu, G. Roman, M. I. Kuroda, and R. L. Davis, Cell 88, 445 (1997).
27
H. Amrein and R. Axel, Cell 88, 459 (1997).
Author Index
Numbers in parentheses are footnote reference numbers and indicate that an author’s work is
referred to although the name is not cited in the text.
A Alt, F. W., 335

Altieri, A. S., 149
Aagaard, L., 387 Alvarez, P., 390
Aasland, R., 119, 147, 152(4) Amati, B., 315
Adams, P. D., 139 Amrein, H., 428
Adrian, M., 60 An, W., 76, 78(17), 195, 196(20)
Aerts, S., 316 Anderson, J. D., 5
Agard, D. A., 45, 47 Anderson, R. M., 185
Aggarwal, A. K., 121, 125(20), 126, 127(20), Andrulis, E. D., 395
128(20), 130(20), 137, 233 Aparicio, J. G., 314(9), 315, 317
Agrawal, R. K., 60 Aparicio, O. M., 290, 314(9), 315, 317,
Ait-Si-Ali, S., 114, 189, 191(7), 192, 322(13), 333(13), 350, 353(11), 357(11)
193(7), 197(12) Apostol, B. L., 200
Akhtar, A., 152 Appanah, R., 248, 249(30), 254(30), 260, 405,
Akita, M., 189, 201 406(9), 408(9), 421, 422(18)
Al-Abed, Y., 174 Aranda, C., 120
Alani, R. M., 188 Archer, S. J., 149, 151(6), 168(6), 169(6)
Albert, T., 316 Archer, T. K., 28
Aletta, J. M., 272 Arkell, R. M., 420
Alghisi, G. C., 420 Armour, C. D., 290
Allemand, J.-F., 76, 101(6), 102(5) Armstrong, C. M., 172(13), 173, 177(13)
Alley, S. C., 120 Arnaud, D., 247, 260, 406, 408(15), 419(15),
Allfrey, V. G., 199, 221 420, 421
Allis, C. D., 106, 108, 109(15), 111(15; 16), Arrowsmith, C. H., 123
112, 112(15; 16), 114, 114(15; 16), Asahara, H., 196
117(16; 17; 29; 30), 119, 120, 121, Ashburner, M., 402
130(23), 131, 132(8), 134(8), 140, 140(8), Ashkin, A., 74, 80(4), 82
142(8; 29), 144(8), 151, 171, 189, 194, Ashworth, A., 421
199, 221, 222, 223, 223(9), 231, 232, Asturias, F. J., 48, 49, 59(4)
232(15), 233(5), 234(2; 4), 235, 237, Aswad, D. W., 272
238(21), 247, 247(21), 248, 248(11), Atherton, E., 273
254(11; 27), 255, 260, 287, 334, 335, Aucott, R., 237, 242(22), 246(22),
348(5), 349, 351(6), 387, 388, 405, 406, 248(22), 254(22)
408(15), 419(15), 419(15), 421 Ausio, J., 20
Allocco, J. J., 174, 175 Avalos, J. L., 172(14), 173
Allshire, R. C., 131, 139, 149, 151, Avner, P., 247, 260, 405, 406, 408(15),
151(8), 157(13), 170(13), 386, 419(15), 421
387(28), 392 Axel, R., 428
Almouzni, G., 119, 240, 272, 428 Ayoub, N., 428
429
430 author index
B Belyaev, N. D., 209, 223, 406, 408(13)

Bennett, H. A., 290, 313, 359
Bailly, D., 240, 271, 428 Bennink, M. L., 62, 73, 81, 93, 94,
Balasubramanian, R., 106(11), 107 95(29), 96(29)
Baldry, S. E., 420 Bensimon, D., 76, 101(6), 102(5)
Baldwin, P. R., 39, 42(12) Benton, B. K., 120
Baldwin, R. L., 162 Beorchia, A., 33(26), 34
Bales, E., 198 Beppu, T., 189, 201
Ball, C. A., 359 Bergel, M., 112
Ball, L. J., 149, 151(6; 7), 167(7), Berger, S. L., 106, 108, 109, 109(14–16),
168(6), 169(6) 111(14–16; 19), 112(14–16; 19), 114,
Ballabio, A., 420 114(15; 16; 19), 117(16; 18), 118(19),
Ballario, P., 120, 125, 137 120, 128, 131, 193, 194, 198, 269,
Balsano, C., 197 272, 283(16)
Bandyopadhyay, D., 198 Berliner, E., 379(16; 17), 380
Bannister, A. J., 131, 132, 137(10), 140(10), Bernat, J. A., 373, 387, 387(8)
146(10), 151, 157(13), 161(15), 170(13), Bernstein, B. E., 106, 188, 189(2), 237, 270,
237, 255, 269, 270, 271(7), 283, 283(7), 271(7), 283(7), 290, 301, 303, 314, 323,
386, 387(28) 349, 354, 357(1), 360, 360(1)
Bard, M., 290 Bertone, P., 317
Bar-Joseph, Z., 314, 317 Beyer, A., 237, 238(19), 245(19), 246(19),
Barlev, N. A., 109, 111(19), 112(19), 114(19), 249(19), 254(19)
118(19), 128, 193, 198 Biltonen, R. L., 144
Barnett, J. D., 314(9), 315, 317 Binkley, G., 359
Barr, M. L., 421 Binnig, G., 73
Barratt, M. J., 219, 223, 229(18), 232(16) Bird, A. P., 328
Barrette, T. R., 270 Bishop, A. C., 357
Barsyte-Lovejoy, D., 318 Bitterman, K. J., 185
Bartley, S. M., 315, 316, 332 Bjorkholm, J. E., 74, 80(4)
Barton, G. J., 151 Bjorklund, S., 119
Bassing, C. H., 335 Blackburn, G. M., 270
Bates, G. P., 200 Blandino, G., 197
Bauer, U. M., 272, 283, 283(15; 16) Blat, Y., 314
Baumeister, W., 48 Blevins, R. A., 123, 136
Bautz, E. K. F., 372, 374(3), 394 Block, S. M., 62, 63, 65(13), 67(13), 68(13),
Bax, A., 123, 124, 125, 136, 137 69(13), 71(3), 95
Baylin, S. B., 235 Bodai, L., 200
Bazett-Jones, D. P., 232 Bodenhausen, G., 166(33), 167
Beck, S., 120 Boeke, J. D., 172(14), 173, 174
Becker, P. B., 3, 12, 12(3), 152 Boggs, B. A., 222, 237, 238(21),
Beddington, R., 408, 409(20) 247(21), 405
Bednar, J., 36 Bohlander, S. K., 295, 353
Bednarek, M. A., 174, 175 Bone, J. R., 267, 394
Bei, Y., 196, 197(25) Bongiorni, S., 237, 242(22), 246(22),
Bekiranov, S., 317 248(22), 254(22)
Bell, S. D., 173 Bonnet, N., 33(26), 34
Bell, S. P., 290, 314(9), 315, 317, Borra, M. T., 171, 173, 174, 177(24)
322(13), 333(13) Borsani, G., 420
Belmont, A. S., 390 Botstein, D., 290, 317, 323(12), 350, 353,
Belotserkovskaya, R., 120 353(10), 354(10), 357(13), 360
author index 431
Bottcher, B., 60 C
Bouillon, R., 343
Bouman, P., 314, 349, 357(1), 360(1) Caiafa, P., 76
Bourachot, B., 120 Cairns, B. R., 120, 317
Boyer, L. A., 106(12), 107 Caldwell, C. W., 333
Brachmann, C. B., 172(14), 173 Caldwell, J. A., 231
Brady, J. N., 196 Callaghan, J., 132, 137(10), 140(10), 146(10),
Brame, C. J., 231 147, 151, 161(15)
Brandts, J. F., 146, 147(40) Cam, H., 306, 318, 360
Branton, D., 32 Campbell, M. J., 360
Brasher, S. V., 149, 151(7), 167(7) Cannova, C., 174
Braunstein, M., 349 Cantor, R. C., 140, 142(30)
Breslow, R., 171, 199, 200(6) Cao, R., 248, 252, 254(28; 29), 260, 262, 332,
Briggs, S. D., 131, 132(8), 134(8), 140(8), 405, 406(10), 408(10), 421, 422(19)
142(8), 144(8), 231 Capra, V., 421
Brinkley, B. R., 232 Carlson, J. E., 121, 125(20), 126, 127(20), 128,
Broach, J. R., 349 128(20), 129(53), 130(20; 53), 137, 233
Broadhurst, R. W., 149, 151(6; 7), 167(7), Carlson, M., 289
168(6), 169(6) Carmen, A. A., 209, 217(7), 289, 293(1)
Brockdorff, N., 241, 242(24), 247(24), Caron, C., 175
248, 249(30), 254(30), 260, 405, Caron, F., 92
406, 406(9), 408(9; 16), 420, Carozza, M. J., 120
421, 422(18) Carroll, A. S., 357
Brosch, G., 174, 176 Carruthers, L. M., 17, 20(7; 8)
Brower-Toland, B., 62, 65, 65(8), 67(8), Cavagnero, S., 125
68(8), 71(8) Cavalli, G., 374, 392
Brown, C. E., 120 Cavanagh, J., 165(31), 166
Brown, C. J., 420 Cawley, S. E., 317
Brown, D. R., 199(8), 200 Cebrat, M., 188, 191, 193(10), 194(10)
Brown, P. O., 290, 302, 314, 317, 323(12), 350, Cechvala, M., 318
353, 353(10), 354(10), 357(13), 360 Celic, I., 172(14), 173, 174
Brownell, J. E., 106, 117(29), 119, 120, 189, Chadwick, B. P., 405
199, 223 Chakraburtty, K., 290
Brownstein, M. J., 323 Chalkley, R., 336
Brunger, A. T., 124, 139 Chambon, P., 119, 120(4), 121(4), 127(4)
Buchard, J., 304 Chang, J. H., 173
Bulger, M., 65(16), 66 Chang, J. J., 60
Burakoff, S. J., 123 Charlton, J. A., 328
Burgoyne, P. S., 237, 242(22), 246(22), Chase, M. W., 380
248(22), 254(22) Chatenay, D., 92
Burlingame, R. W., 217 Chaumeil, J., 405, 406, 408(17), 419(17)
Burnett, E., 316 Chaya, D., 361, 368(3)
Burns, J. A., 164 Chehab, N. H., 128
Bustamante, C., 62, 76, 82, 92 Chen, C. M., 328, 333, 333(30)
Bustin, M., 112, 209, 211, 212(3), 213(9), 214, Chen, D., 173, 273
214(3), 215, 216(15), 361, 368(3) Chen, H., 45
Butler, J. C., 164 Chen, M., 323
Buttle, K., 44 Chen, T., 314(9), 315, 317
Byrd, R. A., 149 Chen, Y., 323
Bystricky, K., 17 Cheng, A., 60
432 author index
Cheung, P., 171, 221, 223, 232(15), 233(5), Copeland, R. A., 187
237, 238(21), 247(21), 405 Cornish-Bowden, A., 187
Cheung, W. L., 223, 232(15) Corona, D. F., 12
Chin, J., 3 Correira, J. J., 164
Chinault, A. C., 222, 237, 238(21), Costanzi, C., 405, 421
247(21), 405 Costanzo, A., 197
Chinnaiyan, A. M., 270 Cote, J., 235
Chiu, W., 39, 41, 42(12) Cotter, R. J., 193, 194(14)
Cho, R. J., 360 Courtney, A. D., 193, 194(14)
Cho, Y., 173 Couturier, J., 417
Choi, I., 316 Cowan, S. W., 59
Chou, J. J., 125 Cowell, I. G., 237, 242(22), 246(22),
Chow, J. C., 420 248(22), 254(22)
Chowdhry, B. Z., 147 Cowieson, N. P., 139, 149, 151(8)
Christensen, J., 316 Coyne, R. S., 140, 142(29)
Christopherson, R. I., 163 Craig, C. A., 372, 373, 386, 386(9), 387,
Chu, S., 74, 80(4) 387(8), 389(9), 394, 403
Chua, P., 120 Crane-Robinson, C., 222, 336
Chuikov, S., 248, 254(27), 255 Creech, H. J., 379(16; 17), 380
Chung, W. H., 49, 59(4) Croce, K., 139
Church, G. M., 360 Croquette, V., 76, 101(6), 102(5)
Churchill, M. E. A., 108, 117(17) Cross, S. H., 328
Ciccone, C. N., 334 Crowley, K. A., 17, 106(12), 107
Cimato, T. R., 272 Crowther, R. A., 45, 47(21), 60
Clayton, A. L., 219, 222, 223, 232(16) Crumley, T. M., 174
Clayton, D. F., 361 Cui, Y., 62, 82, 92
Cleary, M., 283 Czermin, B., 252, 262
Clements, A., 108, 109(14), 111(14), 112(14)
Clemson, C. M., 420
D
Clore, G. M., 123, 124, 136, 139
Cluzel, P., 92 Dahlquist, F. W., 166
Cocito, A., 315 Dai, H., 290, 304, 313, 359
Coessens, B., 315 Damelin, M., 317
Coffey, E., 290 Dandolo, L., 420
Cohen, A., 171, 428 Daniel, C., 3
Cohen, H. Y., 185 Dannenberg, A. J., 195
Cohen, L. H., 372, 374, 374(2) Danzer, J. R., 390
Cole, P. A., 114, 115, 188, 189, 190, 191, Darkin-Rattray, S. J., 174
191(7), 192, 193, 193(7; 10; 13), 194(9; 10; Darnell, J. E., Jr., 361
14), 195, 196, 197, 197(12), 198 Daugherty, M., 191
Cole, R. N., 174 Daujat, S., 272, 283(15; 16)
Colletti, S. L., 174 Dausman, J., 420
Compton, J. L., 404 Davidson, I., 119, 196
Connors, B., 222 Davie, J. R., 174, 346
Constantini, F., 408, 409(20) Davis, R. L., 428
Conway, J. F., 42 Davuluri, R. V., 315
Cook, R. G., 106(13), 107, 117(29; 30), 119, Dawid, I. B., 120
222, 223, 223(9), 231 De Angelis, J., 191, 194(9)
Coons, A. H., 379(16; 17), 380 Decallonne, B., 343
Cooper, P., 420 de Grooth, B. G., 62, 93, 94, 95(29), 96(29)
author index 433
de la Cruz, C. C., 248, 254(29), 260, 405, Dynlacht, B. D., 304, 306, 318, 361
406(10), 408(10), 421, 422(19) Dyson, H. J., 125, 130(44), 137
Delaglio, F., 123, 136 Dyson, M. H., 121
DeLano, W. L., 139 Dziedzic, J. M., 74, 80(4)
Delcuve, G. P., 174
Demeler, B., 18, 19(12), 20(12), 24(12)
E
Demeter, J., 359
De Moor, B., 315 Easom, R. A., 18, 27(11)
Deng, H., 173 Eberharter, A., 106(13), 107
Denu, J. M., 106(12), 107, 114, 131, 173, 174, Ebert, A., 373, 387(10), 389(10), 392(10)
177(21; 24), 178, 194, 223, 232(15) Ebright, R. H., 140
DeRisi, J. L., 314, 350, 357 Eden, A., 408
DeRosier, D. J., 45, 47(21) Edmondson, D. G., 106, 117(29), 119, 120,
Dessain, S. K., 173 189, 199, 223, 267
Deuring, R., 120 Eftink, M. R., 157(21), 158, 160(21)
Devlin, M. K., 188 Eick, D., 316
Dhalluin, C., 121, 125(20), 126, 127(20), Eiletici, P., 125
128(20), 130(20), 137, 233 Einck, L., 30, 215
Dhanasekaran, S. M., 270 Eisen, M. B., 302, 360
Diaz, M. O., 295, 353 Eisenberg, H., 5, 255
Dickens, N. J., 147 Eisenhaber, F., 147, 287, 387
Dietrich, V., 373, 386, 386(9), 389(9), 394 Eissenberg, J. C., 131, 132(8), 134(8), 140(8),
Dignam, J. D., 20 142(8), 144(8), 147, 151, 373, 386, 386(9),
Dimitrov, S., 18 389(9), 394
Dollard, C., 120 Ekstrom, T. J., 171
Domaille, P. J., 149, 151(6), 168(6), 169(6) Ekwall, K., 392
Dombrowski, A. W., 174 Elgin, S. C. R., 372, 373, 374, 376(1), 380,
Dong, H., 350 380(1), 386, 387, 387(8), 389(9), 394, 403
Donigian, J. R., 171 Ellis, C. D., 269
Dorigo, B., 17 Ellis, T. W., 272
Doring, P., 395 Emre, N. C., 237, 270, 271(7), 283(7)
Dorn, R., 373, 387, 387(10), 389(10), 392(10) Erard, M., 316
Dorr, A., 130, 197 Erdjument-Bromage, H., 120, 231, 235, 248,
Douarin, B. L., 119, 120(4), 121(4), 127(4) 252, 254(27; 28), 255, 262, 332, 395, 408
Dow, S., 304 Erlich, R. L., 314, 349, 357(1), 360(1)
Downing, K. H., 45 Escalante-Semerena, J. C., 172(14), 173, 174
Doyle, M., 237, 238(20), 240(20), 241(20), Espinosa, R. III, 295, 353
242(20), 246(20), 247(20) Ettinger, M. J., 272
Drenkow, J., 317 Euskirchen, G., 317
Driscoll, P. C., 136 Evans, P. R., 125, 137
Du, J., 120 Ezeokonkwo, C., 48
Du, K., 196
Dubochet, J., 60
F
Duggan, L. J., 120
Dulski, P. M., 174 Fairbrother, W. J., 165(31), 166
Dunu, J. M., 172(16; 17), 173, 177(16) Fanciulli, M., 197
Duquet, A., 192, 197(12) Fang, J., 248, 254(28; 29), 260, 405, 406(10),
Duthie, S. M., 420 408(10), 421, 422(19)
Dutnall, R. N., 112 Fanti, L., 237, 242(22), 246(22),
Dutrillaux, B., 417 248(22), 254(22)
434 author index
Farnham, P. J., 315, 316, 318, 320, 322(25), Fujiyoshi, Y., 60

331(25), 332, 360 Fulco, M., 197
Farooq, A., 128, 129(53), 130(53) Funches, L., 373, 387, 387(8)
Faulkner, R., 199, 221 Fundele, R., 237, 242(22), 246(22),
Fay, F. S., 411 248(22), 254(22)
Feeney, R. E., 157(20), 158, 275 Funk, J. O., 316
Felsenfeld, G., 335, 348(5) Furie, B., 139
Femino, A. M., 411 Furie, B. C., 139
Feng, Q., 231, 235, 248, 254(28)
Fernandez, P. C., 315
G
Ferreira, G. C., 178
Fersht, A., 187 Gachotte, D., 290
Fesik, S. W., 123, 127 Gaemers, S., 125
Filetici, P., 120, 137 Gafken, P. R., 235
Finch, L. R., 163 Gall, J., 373
Finkelstein, D. B., 359 Galuska, S., 175
Finnin, M. S., 171 Gamblin, S. J., 270
Firestein, R., 283 Gasch, A. P., 357, 359(16)
Fischbeck, K. H., 199 Gaszner, M., 335, 348(5)
Fischer, A., 373, 387(10), 389(10), 392(10) Gay, F., 196, 197(25)
Fischle, W., 131, 174, 334 Gelles, J., 62, 63, 65(13), 67(13), 68(13),
Fisher, H. F., 144 69(13), 71(3), 95
Flanagan, J. F., 39, 41(11) Georgakopoulos, T., 120
Flaus, A., 65 Georgel, P. T., 17, 20(9), 33
Fletcher, C. M., 112 Georges, S. A., 196
Fletcher, T. M., 17, 18, 18(5; 6), Gerber, G. K., 314, 317
19(5; 6; 12), 20(5; 6; 12), 22, 24(12), Gerstein, M., 317
26(6), 27(5; 6), 196 Gerton, J. L., 314
Fodor, S. P., 317 Ghazi-Noori, S., 200
Fogarty, K., 411 Ghislain, L. F., 82
Fogh, R. H., 149, 151(7), 167(7) Ghisolfi-Nieto, L., 316
Foltz, K. R., 173, 177(24) Ghosh, D., 270
Fontemaggi, G., 197 Gibson, T. J., 119
Formstone, E. J., 420 Giebler, H. A., 196
Foster-Hartnett, D. M., 386 Gifford, D. K., 314, 317
Frado, L. Y., 30 Gilbert, D. M., 237, 242(22), 246(22),
Fraenkel, E., 314, 317 248(22), 254(22)
Francis, N. J., 252, 262 Gilkeson, G. S., 200
Franco, L., 174 Gilmour, D. S., 336
Frank, J., 44, 49, 51(3), 52(3), 53, 57(3), 60 Gingeras, T. R., 317
Frank, S. R., 315 Ginzinger, D. G., 343
Franke, A., 151 Giulietti, A., 343
Freeman, S. J., 139 Glaeser, R. M., 45
Friedman, F. K., 209 Glaser, R. L., 394
Friedmann, T., 420 Goetz, M. A., 174
Friend, S. H., 290, 304 Goldblatt, D., 211, 213(9)
Fry, C. J., 263 Goldman, M., 380
Frye, R. A., 172(11; 12), 173 Goldman, R. D., 411, 413(23)
Fryer, C. J., 28 Gollub, J., 359
Fujita, H., 162 Gonzalez, A., 120
author index 435
Gordon, D. B., 314, 317 Hannett, N., 290, 314, 317, 322(13), 333(13),
Gotchel, B. V., 374 350, 353(11), 357(11)
Gottschling, D. E., 235 Hansen, J. C., 17, 18, 18(5; 6), 19(5; 6; 12),
Gounalaki, N., 120 20, 20(5–9; 12), 22, 24(12), 26(6),
Grant, P. A., 106, 106(11; 13), 107, 120 27(5; 6), 33
Grassucci, R. A., 60 Hansen, U., 272
Graveel, C. R., 332 Harbison, C. T., 314, 317
Gray, S. G., 171 Harel-Bellan, A., 192, 197(12)
Green, G. R., 30 Harkiolaki, M., 139
Greene, J., 315 Harlan, J. E., 123
Greenleaf, A. L., 372, 374(3), 394 Harlow, E., 212, 216(11), 228, 368, 412
Greenwald, M., 200 Harnett, T., 151
Greve, J., 62, 93, 94, 95(29), 96(29) Harper, S., 112, 114(23)
Grewal, S. I. S., 151, 335, 388, 428 Harrelson, A. L., 361
Griesinger, C., 128 Harris, K. G., 128
Griess, G. A., 18, 27(11) Harrison, P. M., 317
Grigliatti, T. A., 386 Hart, C. M., 252, 262
Grompe, M., 420 Hartman, S., 317
Gronenborn, A. M., 123, 124, 136 Hartnett, T., 386
Gros, P., 139 Hassig, C. A., 174, 189, 199
Grosse, I., 315 Haugland, R. P., 142
Grosse-Kunstleve, R. W., 139 Hayamizu, T., 361, 368(3)
Grozinger, C. M., 171, 172(8) Hayes, J. J., 171
Grubmeyer, C., 172(14), 173 Haynes, S. R., 120
Grummt, I., 174 He, C., 121, 125(20), 126, 127(20), 128(20),
Grunstein, M., 171, 209, 217(7), 289, 290, 130(20), 137, 233
290(3), 292, 293(1), 294, 294(9), 300, He, Y. D., 128, 129(53), 130(53), 290, 313, 359
300(10; 14), 301, 302(10), 303, Heagle, A. B., 60
304(14), 314(38), 315, 349, 351(7), Heard, E., 237, 238(21), 247, 247(21), 260,
352(7), 360(2) 405, 406, 408(15; 17), 419(15; 17), 421
Grzesiek, S., 123, 136, 137 Hebbes, T. R., 222, 336
Gu, W., 128, 173 Hebert, H., 36
Guarente, L., 120, 172(13), 173, 177(13) Hebert, J. M., 359
Guggiari, M., 406, 408(17), 419(17) Hecht, A., 290, 292, 300(14), 304(14), 349,
Gurnett, A. M., 174 351(7), 352(7)
Guzman, E., 395 Heller, C., 92
Heller, R. C., 172(15), 173, 175(15), 177(15),
182(15), 183(15)
H
Hendzel, M. J., 174, 232
Hager, G. L., 28, 196 Henklein, P., 130, 197
Hahn, M., 48 Hernandez-Boussard, T., 359
Hajduk, P. J., 127 Herrera, J. E., 112
Halazonetis, T. D., 128 Herrera, R. E., 283
Haley, R. H., 109, 111(19), 112(19), 114(19), Hess, D., 175
118(19), 193, 198 Hesse, J., 36
Hall, I. M., 428 Hetzer-Egger, C., 197
Halperin, T., 316 Heyduk, T., 140
Halvorson, H. R., 164 Hilfiker, A., 152
Hamilton, S. L., 41 Hilfiker-Kleiner, D., 152
Hampsey, M., 248, 254(26) Hill, J. J., 4
436 author index
Hnilica, L., 214(14), 215 Ishii, M., 191, 194(9)

Ho, M. H., 352 Ivanova, A. V., 149
Hobson, A., 373, 386, 386(9), 389(9), 394 Iyer, V. R., 290, 317, 350, 353(10), 354(10)
Hock, R., 215
Hockly, E., 200
Hoffman, J., 373, 387(10), 389(10), 392(10)
J
Hoffmann, K., 176 Jackson, G. R., 200
Hogan, B., 408, 409(20) Jackson, M. D., 173, 177(21; 24)
Hogness, D., 147 Jackson, S. P., 173
Holmes, S. G., 349 Jackson, V., 336
Holmquist, G., 377 Jacobs, S. A., 131, 132(8), 134, 134(8),
Homo, J. C., 60 137(11), 140(8; 11), 141(11), 142(8; 11),
Hong, H., 272 144(8; 11), 145(11), 146(11), 151, 161(16)
Horak, C. E., 290, 295, 317, 350, 353(10), Jacobson, R. H., 127, 137, 144(20)
354(10), 357, 359(16) Jaenisch, R., 406, 408, 420
Horita, D. A., 149 Jambeck, P., 108, 117(17)
Horiuchi, J., 120 James, T. C., 373, 386, 386(9), 389(9), 394
Horn, P. J., 17 Jameson, D. M., 140
Horowitz, R. A., 32, 34, 35(4), 36, 45, Jamrich, M., 373, 374(3)
47, 47(20) Jantzen, B. C., 62, 67(5), 68(5)
Horowitz-Scherer, R. A., 29, 39, 41(11) Jaskelioff, M., 234
Housman, D. E., 200 Jeanmougin, F., 119, 120(4), 121(4), 127(4)
Howard, R., 114 Jennings, E. G., 290, 314, 314(9), 315, 317,
Howder, B., 125 322(13), 333(13), 350, 353(11), 357(11)
Howe, L., 120 Jenuwein, T., 131, 151, 170(12), 222, 234,
Howell, S., 270 234(4), 235, 237, 238(19; 20), 240,
Huang, S. M., 272 240(20), 241, 241(20), 242(20; 24),
Huang, T. H., 318, 328, 333, 333(30), 360 245(19), 246(19; 20), 247(20; 24), 248,
Huang, Z.-Q., 197, 231 249(19; 30), 254(19; 30), 255, 260, 270,
Hudson, B. P., 125, 130(44), 137 271, 283, 288, 373, 386, 387, 387(10; 29),
Hughes, J. D., 360 389(10), 392(10; 29), 405, 406, 406(9),
Hughes, T. R., 290, 304, 313, 359 408(9; 16), 421, 422(18), 428
Hughes-Davies, L., 147 Jeppesen, P., 222, 237, 242(22), 246(22),
Humphrey, E. L., 314, 349, 357(1), 360(1) 248(22), 254(22), 405, 421
Hunt, D. F., 231 Jiang, J. S., 139
Hunter, G. A., 178 Jin, H., 359
Hunter, J. M., 235 Jing, C., 270
Huskisson, N., 237, 242(22), 246(22), John, S., 106(13), 107
248(22), 254(22) Johnson, B. A., 123, 136
Hwang, K. Y., 173 Johnson, M. L., 164
Johnston, C. M., 420
Johnstone, R. W., 200
I Jolly, D. J., 420
Imai, S., 172(13), 173, 177(13) Jones, A. R., 290, 304
Imhof, A., 252, 262 Jones, M. E., 163
Imiliani, S., 174 Jones, P. A., 235
Indyk, L., 144 Jones, R. N., 379(16; 17), 380
Inman, J. M., 323 Jones, R. S., 252, 262, 332
Ioshikhes, I. P., 328 Jones, S., 210, 218(8)
Iou, J. Y., 59 Jones, T. A., 59, 139
author index 437
Joseph, G., 316 Kingston, R. E., 252, 262

Jung, M., 176 Kissinger, M., 120
Kjeldgaard, M., 59, 139
Kladde, M. P., 120
K
Klar, A. J. S., 149
Kadonaga, J. T., 65(16), 66 Kleckner, N., 314
Kadosh, D., 289 Klee, C. B., 125
Kaeberlein, M., 172(13), 173, 177(13) Klotz, I. M., 13, 14(14)
Kaehcke, K., 197 Klug, A., 45, 47(21), 79
Kaloper, M., 359 Kobayashi, R., 106, 120, 189, 199
Kammer, G. M., 199(8), 200 Koch, S. J., 62, 67(5), 68(5)
Kanin, E., 290, 317, 322(13), 333(13), 350, Koeck, P. J., 36
353(11), 357(11) Koh, S. S., 231, 272
Kapranov, P., 317 Kohlmaier, A., 237, 238(19; 20), 240(20),
Kaptein, R., 125 241(20), 242(20), 245(19), 246(19; 20),
Karymov, M. A., 76, 78(17), 97, 103(32), 247(20), 249(19), 254(19)
195, 196(20) Kolle, D., 174
Kauer, M., 234 Koller, T., 79
Kaufman, T. C., 120 Komili, S., 317
Kay, G. F., 420 Koonin, E. V., 152
Kay, L. E., 123, 136 Kornberg, R. D., 48, 49, 59(4), 119, 120
Kazantsev, A., 200 Koster, A. J., 45, 47(20)
Kel, A. E., 315 Kotzin, B. L., 217
Kelly, G., 270 Kouzarides, T., 131, 132, 137(10), 140(10),
Kelly, W. K., 199, 200, 200(6) 146(10), 147, 151, 157(13), 161(15),
Kel-Margoulis, O. V., 315 170(13), 235, 237, 255, 269, 270, 271(7),
Kenna, M. A., 172(14), 173 272, 283, 283(7; 15; 16), 314, 349, 357(1),
Kennison, J. A., 120 360(1), 386, 387(28)
Keohane, A. M., 406, 408(13) Kozhich, O. A., 323
Kervabon, A., 174 Kraulis, P. J., 151, 170
Ketel, C. S., 248, 254(28) Kraus, W. L., 196
Khalil, E. M., 114, 189, 191, 191(7), Krauss, V., 373, 387(10), 389(10), 392(10)
193(7), 194(9) Kreider, J. I., 18, 19(12), 20(12), 22, 24(12)
Khochbin, S., 171, 175 Kubicek, S., 237, 238(19), 245(19), 246(19),
Khorasanizadeh, S., 131, 132(8), 134, 134(8), 249(19), 254(19)
137(11), 140(8; 11), 141(11), 142(8; 11), Kuhfittig, S., 387
144(8; 11), 145(11), 146(11), 151, 161(16) Kumar-Sinha, C., 270
Khosravi, F., 318 Kundu, T. K., 114, 189, 191(7), 193(7), 195
Kidd, M. J., 290, 304 Kuningas, R., 175
Kidner, C., 428 Kuo, M. H., 108, 114, 117(17; 29), 119, 171,
Kiermer, V., 130, 197 223, 334, 349, 351(6)
Kijima, M., 189, 201 Kuras, L., 336
Kim, C. M., 191, 193(10), 194(10) Kurdistani, S. K., 289, 290, 290(3), 294,
Kim, H. C., 173 294(9), 300(10), 301, 302(10), 303,
Kim, J. M., 405 314(8), 315, 349, 360(2)
Kim, S., 237 Kuroda, M. I., 394, 428
Kim, Y., 114, 178 Kurokawa, R., 200
Kim, Y. J., 405 Kurooka, H., 115, 192, 193(13)
King, A. M., 290 Kuszewski, J., 124, 139
King, D. S., 127, 137, 144(20) Kuzmichev, A., 252, 262, 408
438 author index
L Lee, K.-M., 65(15), 66

Lee, T. I., 314, 317
Lachner, M., 131, 151, 170(12), 235, 240, 255, Lee, W., 123
270, 271, 386, 387(28), 392(28), 428 Lefevre, G., Jr., 373, 401(14), 402
Lacoste, N., 235 Leinwand, L. A., 411, 413(23)
Lacy, E., 408, 409(20) Lejeune, J., 417
Ladbury, J. E., 147 Lemercier, C., 171
Ladurner, A. G., 127, 137, 144(20) Leno, G. H., 62, 90, 94, 95(29), 96(29)
Lafferty, J. A., 217 Leone, J. W., 222, 223(9)
Lafreniere, R. G., 420 Lepault, J., 60
Laible, G., 387 Leppik, R., 166(35), 167
Lakowicz, J. R., 4 LeRoy, G., 175
Lamzin, V. S., 139 Leuba, S. H., 62, 73, 74, 76, 77(3), 78(3; 17),
Landick, R., 62, 63, 65(13), 67(13), 68(13), 93, 94, 95(29), 96(29), 97, 103(32),
69(13), 71(3), 95 195, 196(20)
Landry, J., 172(14–16), 173, 175, 175(15), Levenstein, M. E., 66
177(15; 16; 23), 180(40), 182(15), Levrero, M., 197
183(15; 23), 185(23) Lewis, C. M., 200
Landsberg, G., 232 Li, F.-Y., 112, 113(21), 114(21), 194
Lane, D., 212, 216(11), 228, 368, 412 Li, J., 131, 132(8), 134(8), 140(8), 142(8),
Lane, W. S., 175 144(8), 196, 197
Langer, M. R., 106(12), 107, 194 Li, N., 175
Langerman, N., 144 Li, S., 125
Längst, G., 3, 12, 12(3) Li, X., 108, 109(15), 111(16), 112(16),
Larent, B. C., 120 114(16), 117(16), 194
Larsen, P. L., 336, 349 Li, Y., 372, 390, 403
Laskowski, R. A., 125 Liang, R., 28
Latorre-Esteves, M., 185 Lichten, M., 314
Lau, O. D., 114, 189, 191(7), 193, Lieb, J. D., 317, 323(12), 353, 357(13)
193(7), 194(14) Likhite, V. S., 316
Laue, E. D., 132, 137(10), 140(10), 146(10), Lin, L. N., 146, 147(40)
147, 149, 151, 151(6; 7), 161(15), Lin, R., 222, 223(9)
167(5; 7), 168(6), 169(6) Lin, Y., 112
Laue, T., 163 Lindsay, S. M., 74, 77, 77(3), 78(3)
Laurent, C., 417 Liphardt, J., 62
Lavender, J. S., 406, 408(13) Lis, J. T., 62, 65(8), 67(8), 68(8), 71(8), 248,
Lavery, R., 92 254(27), 255, 336, 372, 373(4), 374(4),
Lawrence, C., 420 393, 393(4), 394, 395, 397, 403(13), 404
Lawrence, J. B., 411, 420 Litt, M. D., 335, 348(5)
Laybourn, P. J., 196 Liu, C. L., 349, 354
Le Beau, M. M., 295, 353 Liu, J. S., 314, 349, 357(1), 360(1)
Lebersorger, A., 387 Liu, L., 108, 117(18), 128
Lebovitz, R. M., 20 Liu, S., 315
Lebrun, A., 92 Liu, X., 317, 323(12), 353, 357(13)
Lechner, M., 237, 238(20), 240(20), 241(20), Lobell, R. B., 394
242(20), 246(20), 247(20) Lockhart, D. J., 350
Lechner, T., 174 Lofton-Day, C., 333
Lee, H. L., 28 Loidl, P., 174, 176
Lee, J. W., 173 Lorch, Y., 48, 49, 59(4), 119
Lee, K. C., 196 Loret, E., 197
author index 439
Losson, R., 119, 120(4), 121(4), 127(4) 117(16), 118(19), 128, 131, 188, 191, 193,
Louis, J. M., 125 193(10), 194, 194(10), 198
Loven, M. A., 316 Marsh, J. L., 200
Lowary, P. T., 5, 12(7) Marshall, B. L., 173, 174, 177(24)
Lowden, P. A., 200 Martienssen, R. A., 428
Lowe, N., 125, 137 Martinez-Yamout, M. A., 125, 130(44), 137
Loyola, A., 235, 255 Marton, M. J., 290, 304, 313, 359
Lu, H., 196 Martone, R., 317
Lucchesi, J. C., 152 Marzilli, L. A., 193, 194(14)
Ludtke, S. J., 39, 42(12) Mason, P., 395
Luger, K., 58, 59(9), 64, 65, 171, 196 Mastrangelo, M. A., 408
Lum, P. Y., 290 Matese, J. C., 359
Lunnen, K. D., 420 Mathieu, C., 343
Luo, J., 173 Matson, J. C., 124
Luo, K., 300 Matthias, G., 175
Luscombe, N. M., 317 Matthias, P., 175
Lusser, A., 174 Maurer-Stroh, S., 147
Lyon, M. F., 405 McCabe, D., 252, 262
McCabe, V. M., 420
McCampbell, A., 200
M McCarthy, B. J., 404
Ma, H., 231, 272 McDowall, A. W., 60
Ma, Y., 140 McEwen, B. F., 45
MacArthur, M. W., 125 Mcintosh, L. P., 166
Mader, A. W., 58, 59(9) McKinsey, T. A., 198
Madore, S. J., 332 McLaughlin, P. J., 139, 149, 151(8)
Mahadevaiah, S. K., 237, 242(22), 246(22), McNeil, J. A., 420
248(22), 254(22) McRorie, D. K., 162, 165(25)
Mahadevan, L. C., 121, 219, 223, Meadows, R. P., 123, 127
229(18), 232(16) Means, G. E., 157(20), 158, 276
Mahal, A., 200 Mechtler, K., 131, 151, 170(12), 234, 237,
Mahandiram, D. R., 123 238(19; 20), 240(20), 241(20), 242(20),
Maison, C., 213, 240, 271, 428 245(19), 246(19; 20), 247(20), 249(19),
Mak, W., 248, 249(30), 254(30), 260, 405, 254(19), 255, 287, 386, 387,
406(9), 408(9), 421, 422(18) 387(29), 392(29)
Makhatadze, G. I., 146 Medrano, E. E., 198
Mancini, M. A., 232 Meinke, P. T., 174
Manley, J. L., 128 Meisterernst, M., 316
Mansfield, K., 128 Melfi, R., 252, 262
Mao, D. Y. L., 318 Meller, V. H., 428
Marahrens, Y., 420 Mello, C., 196, 197(25)
Marcus, G. A., 120 Mellor, J., 237, 270, 271(7), 283(7)
Marion, D., 136 Melnyk, P., 60
Marko, J. F., 69(18), 72, 95 Menager, M., 33(26), 34
Marko, M., 44 Merlo, P., 197
Marks, P. A., 171, 199, 200, 200(6) Mermoud, J. E., 241, 242(24), 247(24), 260,
Marmorstein, R., 106, 108, 109, 109(14–16), 405, 406, 408(16), 421
111(14–16; 19), 112, 112(14–16; 19), Merril, C. R., 25
113(21), 114, 114(15; 16; 19; 21; 23), Mery, J., 174
440 author index
Messmer, S., 151 Muth, V., 174

Meyer, M. R., 290, 304, 313, 359 Muzny, D., 420
Migeon, B. R., 420 Myers, R. W., 174, 175
Miller, E. L., 252, 262 Mymryk, J. S., 28
Miller, P., 317
Miller, T., 199, 200(6)
N
Millevoi, S., 316
Minor, W., 139 Nadaud, S., 174
Mirsky, A. E., 199, 221 Nagaich, A., 196
Mishra, N., 199(8), 200 Naguibneva, I., 192, 197(12)
Miska, E. A., 131, 151, 157(13), 170(13), Nakatani, Y., 112, 114, 115, 117(30), 119, 189,
386, 387(28) 191(7), 192, 193, 193(7; 13), 194(14), 223
Mitchell, L. G., 25 Nakayama, J., 151, 388
Mitra, A. K., 60 Nall, B. T., 146, 147(39)
Mittmann, M., 350 Nan, X., 328
Miyazawa, A., 60 Nardulli, A. M., 316
Mizzen, C. A., 117(30), 119, 223 Nare, B., 175
Mlynarczyk-Evans, S. K., 248, 254(29), 260, Narlikar, G., 65(15), 66
405, 406(10), 408(10), 419, 421, 422(19) Nascimento, L., 198
Mo, Y., 108, 109(15), 111(16), 112(16), Nash, P., 119
114(16), 117(16), 194 Nasir, I., 120
Moazed, D., 173, 182(19), 183(19) Nazarian, A., 395
Monestier, M., 217 Neely, K. E., 106
Montminy, M., 196 Neihart, N., 215
Moore, J. M., 127 Nelson, B., 313, 359
Moran, J., 164 Nelson, F. K., 317
Moreau, Y., 315 Neri, D., 124
Moreno, E. T., 18, 27(11) Nesterova, T. B., 248, 249(30), 254(30), 260,
Morrison, A., 283 405, 406(9), 409(9), 420, 421, 422(18)
Morrison, H., 237, 242(22), 246(22), Neuhaus, D., 125, 137
248(22), 254(22) Nevins, J. R., 316
Morshead, K. B., 334 Newall, A. E., 420
Mott, H. R., 132, 137(10), 140(10), 146(10), Ng, H. H., 235, 254, 317
149, 151, 161(15), 167(5) Ngan, V., 386
Mottus, R. C., 386 Ng Eaton, E., 173
Moy, T. I., 317 Nield, J., 38(9), 39
Muchardt, C., 120 Nielsen, P. R., 132, 137(10), 140(10), 146(10),
Muhammad, S., 172(14), 173 147, 149, 151, 151(7), 161(15), 167(5; 7)
Mujtaba, S., 119, 128, 129(53), 130(53) Nielsen, S. J., 272, 283, 283(15)
Muller, J., 252, 262 Nietlispach, D., 132, 137(10), 140(10), 146(10),
Muller, S., 209, 212(4) 149, 151, 151(7), 161(15), 167(5; 7)
Murray, H. L., 314, 317 Nightingale, K., 272, 283(15)
Murray, K., 255 Nightingdale, K. P., 119
Murzin, A. G., 132, 137(10), 140(10), Nikolaev, A. Y., 173
146(10), 147, 149, 151, 151(6), 161(15), Nilges, M., 124, 139
168(6), 169(6) Nishioka, K., 235, 248, 248(11), 252,
Murzina, N. V., 132, 137(10), 140(10), 254(11; 27), 255, 262, 408
146(10), 147, 149, 151, 151(6; 7), 161(15), Noma, K., 335, 428
167(7), 168(6), 169(6) Norris, D. P., 405, 420
author index 441
North, B. J., 174 Paik, W. K., 237

Nusinow, D. A., 419 Palhan, V. B., 76, 78(17), 195, 196(20)
Nyborg, J. K., 196 Pallos, J., 200
Palmer, A. G. III, 165(31), 166
Pandita, T. K., 173
O Panning, B., 248, 254(29), 260, 405, 406(10),
Oberley, M. J., 315, 318, 320, 360 408(10), 419, 420, 421, 422(19)
O’Brien, R., 147 Pannu, N. S., 139
O’Carroll, D., 131, 151, 170(12), 237, 238(20), Pannuti, A., 152
240(20), 241, 241(20), 242(20; 24), Pardue, M. L., 373
246(20), 247(20; 24) 255, 260, 283, 287, Parello, J., 174
386, 387, 387(29), 392(29), 405 Park, J. M., 404
O’Connor, M. B., 252, 262 Paro, R., 147, 151, 349, 374, 392
O’Connor, O. A., 200 Parrizas, M., 361
Odom, D. T., 314, 317 Partridge, J. F., 131, 139, 149, 151, 151(8),
O’Donoghue, S., 124 157(13), 170(13), 386, 387(28)
Oettinger, M. A., 334 Patel, D., 420
Okamoto, I., 405, 406, 408(17), 419(17) Pattatucci, A. M., 120
Okan, N. A., 198 Pavletich, N. P., 171
Olejniczak, E. T., 123 Pawson, T., 119
Olorenshaw, I. M., 199(8), 200 Pedal, A., 130
Olson, E. N., 198 Pedersen, K. O., 161, 162(22)
Olsson, T., 392 Pediconi, N., 197
O’Neill, F. J., 173, 177(24) Pehrson, J. R., 405, 408, 421
O’Neill, L. P., 406, 408(13) Penczek, P., 44, 60
O’Neill, R. R., 25 Peng, J., 395
Onoa, B., 62 Penn, L. Z., 318
Opravil, S., 234, 237, 238(19; 20), 240(20), Penny, G. D., 420
241(20), 242(20), 245(19), 246(19; 20), Perez-Burgos, L., 234, 237, 238(19), 245(19),
247(20), 249(19), 254(19), 255, 287, 387 246(19), 249(19), 254(19)
Orlando, V., 209, 305, 334, 349 Perrakis, A., 139
Ornaghi, P., 125, 137 Perraud, A. L., 173
O’Shea, E. K., 357 Peters, A. H., 234, 237, 238(19; 20), 240,
Oshimura, M., 417 240(20), 241, 241(20), 242(20; 24),
O’Sullivan, R. J., 255 245(19), 246(19; 20), 247(20; 24), 248,
Ott, M., 128, 129(53), 130, 130(53), 197 249(19; 30), 254(19; 30), 260, 271, 405,
Otte, A. P., 248, 249(30), 254(29; 30), 406, 406(9), 408(9; 16), 421, 422(18), 428
260, 270, 405, 406(9), 408(9), Peterson, C. L., 17, 39, 41(11), 62, 65(8),
421, 422(18; 19) 67(8), 68(8), 71(8), 106, 106(12), 107,
Otting, G., 124 234, 263
Otwinowski, Z., 139 Petes, T. D., 314
Overbergh, L., 343 Petros, A. P., 123
Owen, D. J., 125, 137 Pfeifer, G., 362
Pfeifer, J., 123, 136
Pflugfelder, G., 336
P
Phan, Q. N., 323
Pacella, L. A., 120 Pienta, K. J., 270
Pagani, M., 237, 238(20), 240(20), 241, Pierce, M. M., 146, 147(39)
241(20), 242(20; 24), 246(20), Pillus, L., 172(15), 173, 175(15), 177(15),
247(20; 24), 260, 405 182(15), 183(15)
442 author index
Pimpinelli, S., 237, 242(22), 246(22), Ranalli, T., 106, 117(29), 119, 120, 189, 199,
248(22), 254(22) 223, 232
Piotto, M., 166(34; 35), 167 Rasband, W. S., 25
Pirrotta, V., 252, 262 Raschke, E. E., 316
Pise-Masison, C. A., 196 Rasmussen, T. P., 408
Pizzuti, A., 420 Rastan, S., 405, 420
Plagens, U., 394 Ravichandran, K. S., 123
Platero, J. S., 151 Rea, S., 131, 151, 170(12), 255, 287, 373, 386,
Plath, K., 248, 254(29), 260, 405, 406(10), 387, 387(10; 29), 389(10), 392(10; 29)
408(10), 419, 421, 422(19) Read, R. J., 139
Ploton, M., 33(26), 34 Rechsteiner, T. J., 64
Poelman, M., 200 Reddy, V., 60
Poirier, G. G., 235 Reilly, C. M., 200
Polach, K. J., 5 Reinberg, D., 106, 175, 221, 233(4), 235, 248,
Polesskaya, A., 192, 197(12) 248(11), 252, 254(11; 26; 27), 255, 262,
Polishook, J. D., 174 272, 408
Ponting, C. P., 147, 255, 287, 387 Ren, B., 290, 304, 306, 314, 317, 318, 322(13),
Pope, L. H., 93 333(13), 350, 353(11), 357(11), 360
Popova, B., 406, 408(16), 421 Ren, G., 60
Portanova, J. P., 217 Reuter, G., 373, 387, 387(10), 389,
Postnikov, Y. V., 215, 216(15) 389(10), 392(10)
Poulsen, F. M., 124 Reyes, J. C., 120
Poux, A. N., 191, 193(10), 194(10) Rice, J. C., 151, 235, 248, 248(11), 254(11; 27),
Prantera, G., 237, 242(22), 246(22), 255, 388
248(22), 254(22) Rice, L. M., 139
Pray-Grant, M. G., 106(11), 107 Richard-Foy, H., 28
Prestegard, J. H., 125 Richmond, R. K., 58, 59(9)
Preuss, U., 232 Richmond, T. J., 17, 58, 59(9), 64, 65, 171
Price, D. H., 395 Richon, V. M., 171, 199, 200, 200(6)
Privalov, P. L., 146 Riegel, A. T., 28
Prives, C., 128 Riester, D., 176
Pullner, A., 316 Rifkind, R. A., 171, 199, 200(6)
Riggs, A. D., 362
Rinaldi, N. J., 314, 317
Q
Rinn, J. L., 317
Quate, C. F., 73 Robert, C., 76
Quivy, J. P., 240, 271, 428 Robert, F., 254, 290, 314, 317, 322(13),
333(13), 350, 353(11), 357(11)
Roberts, C. J., 290, 304, 313, 359
R
Roberts, S. M., 120
Rackwitz, H.-R., 130 Robertson, E. J., 409, 410(21)
Radermacher, M., 44, 53, 60 Robin, P., 192, 197(12)
Radonovich, M., 196 Robinson, R. C., 209
Rafty, L. A., 173 Robyr, D., 289, 290, 294(9), 300(10),
Rahmatpanah, F., 328, 333, 333(30) 301, 302(10), 303, 314(8), 315,
Raine, A. R. C., 149, 151(6), 168(6), 169(6) 349, 360(2)
Ralston, G., 163, 164(26) Roche, D., 240, 271, 428
Ramakrishnan, V., 112 Rodrigo, I., 174
Raman, C. S., 146, 147(39) Rodriguez-Alfageme, C., 372, 374(2)
Ramjit, R., 97, 103(32) Roeder, G. S., 120
author index 443
Roeder, R. G., 20, 76, 78(17), 114, 128, 189, Scafe, C. S., 290, 317, 350, 353(10), 354(10)
191(7), 193(7), 195, 196(20) Schalch, T., 17
Rohrer, C., 73 Scharenberg, A. M., 173
Rojas, J. R., 108, 109(14–16), 111(14–16), Schatz, M., 41
112(14–16), 114(15; 16), 117(16), 194 Scheer, U., 215
Roman, G., 428 Scheidtmann, K. H., 232
Rosa, E., 200 Scherthan, H., 237, 238(20), 240(20), 241(20),
Rose, A. B., 349 242(20), 246(20), 247(20)
Rose, S., 121, 219, 223, 232(16) Schiltz, L., 197
Roth, F. P., 317 Schiltz, R. L., 117(30), 119, 196, 223
Roth, S. Y., 106, 117(29), 119, 120, 131, 171, Schimmel, P. R., 140, 142(30)
189, 199, 223, 267 Schleucher, J., 128
Rougeulle, C., 247, 260, 406, 408(15), Schlichter, A., 120
420(15), 421 Schlisio, S., 316
Rowley, J. D., 295, 353 Schmatz, D. M., 174, 175
Royer, C. A., 4 Schmid, M., 255, 287, 387
Ruben, D. J., 166(33), 167 Schmidt, C. F., 62
Rubin, M. A., 270 Schmidt, E., 200
Rubin, R. L., 217 Schmidt, M. T., 173
Rudkin, G. T., 372, 374(2) Schnapp, B. J., 62
Ruiz, P., 200 Schneider, R., 237, 255, 270, 271(7), 283,
Rullmann, J. A., 125 283(7), 314, 349, 357(1), 360(1)
Rundlett, S. E., 289 Schnitzer, M. J., 62, 71(3)
Rupert, J. L., 420 Schnitzler, G., 64
Ruprecht, J., 38(9), 39 Schnoelzer, M., 197
Russell, D. W., 377, 384(15) Schofer, C., 237, 238(20), 240(20), 241(20),
242(20), 246(20), 247(20)
Schotta, G., 373, 387, 387(10),
S
389(10), 392(10)
Sachs, L. M., 197 Schramm, V. L., 173
Salvador, M. L., 174 Schreiber, J., 290, 317, 322(13), 333(13), 350,
Sambrook, J., 377, 384(15) 353(11), 357(11), 360
Samori, B., 76 Schreiber, S. L., 106, 171, 172(8), 174, 188,
Sanda, M. G., 270 189, 189(2), 199, 237, 270, 271(7), 283(7),
Santoso, B., 196 290, 301, 303, 314, 349, 354,
Santos-Rosa, H., 237, 270, 271(7), 283(7) 357(1), 360(1)
Sargent, D. F., 58, 59(9) Schreiber, S. S., 323
Sarma, K., 235, 248, 248(11), 254(11; 27), 255 Schroeder, M., 315
Sartorelli, V., 197 Schubert, U., 130
Sasaki, M., 405, 408(4), 417(4), 418(4) Schultz, P., 60
Sassone-Corsi, P., 171, 221, 223, Schurter, B. T., 272
232(15), 233(5) Schwartz, B. E., 372, 373(4), 374(4), 393,
Sathasivam, K., 200 393(4), 395
Sattler, M., 123, 128 Schweizer, D., 241, 242(24), 247(24), 260, 405
Saudek, V., 166(34; 35), 167 Schwienhorst, A., 176
Sauer, S., 237, 238(20), 240(20), 241(20), Scolnick, D. M., 128
242(20), 246(20), 247(20) Scott, M. P., 120
Sauve, A. A., 173 Sedat, J. W., 45, 47
Sawyer, W. H., 140 Seelig, H. P., 175
Saxton, W. O., 48 Segel, I. H., 187
444 author index
Seigneurin-Berny, D., 171 Singh, P. B., 149, 151(6), 168(6), 169(6), 237,
Seitz, V., 252, 262 242(22), 246(22), 248(22), 254(22), 387
Selenko, P., 387 Singh, S. B., 174
Selleck, W., 106(11), 107 Sistare, F. D., 28
Sendra, R., 174 Skelton, N. J., 165(31), 166
Sengupta, A., 252, 262 Sklenar, V., 166(34; 35), 167
Senn, H., 124 Slade, D., 290, 304
Serwer, P., 17, 18, 18(5; 6), 19(5; 6), 20(5; 6), Slama, J. T., 173, 177(23), 183(23), 185(23)
26(6), 27(5; 6; 11) Smith, A. G., 409
Serycheva, I. I., 41 Smith, B. O., 149, 151(7), 167(7)
Sewack, G. F., 272 Smith, C. L., 39, 41(11), 62, 65(8), 67(8),
Sewalt, R. G., 270 68(8), 71(8)
Shabanowitz, J., 231 Smith, D. L., 200
Shaffer, C. D., 373, 387, 387(8) Smith, J. S., 172(14), 173
Shah, V., 272, 282(16) Smith, P., 97, 103(32)
Shankaranarayana, G. D., 428 Smith, S. B., 62, 82, 92, 420
Sheardown, S. A., 420 Snyder, M., 290, 295, 317, 350, 353(10),
Sheppard, R. C., 273 354(10), 357, 359(16)
Sherriff, J., 237, 270, 271(7), 283(7) Sobel, R. E., 117(29), 119, 222, 223
Shi, H., 328, 333, 333(30) Soccio, R. E., 114, 189, 191(7), 193(7)
Shi, W., 237, 242(22), 246(22), 248(22), 254(22) Solomon, M. J., 336, 349
Shi, Y., 196, 197(25) Sommerville, J., 215
Shiloh, A., 173 Sopta, M., 394
Shinkai, Y., 237, 238(19), 245(19), 246(19), Sörensen, O. W., 124
249(19), 254(19) Sörensen, P., 124
Shoemaker, D. D., 290 Sorsa, V., 402
Shokat, K. M., 357 Sousa, K., 120
Shopland, L. S., 397, 403(13) Spector, D. L., 237, 238(21), 247, 247(21),
Shroff, R., 314 260, 405, 406, 408(15), 411, 413(23),
Shrogen-Knaak, M. A., 263 419(15), 421
Shundrovsky, A., 62, 67(5), 68(5) Speicher, D. W., 112, 114(23)
Sibilia, M., 237, 238(20), 240(20), 241(20), Spellman, P. T., 360
242(20), 246(20), 247(20) Spencer, V. A., 346
Siggia, E. D., 69(18), 72, 95 Sperier, A., 373
Silva, J., 248, 249(30), 254(30), 260, 405, Sperier, P., 373
406(9), 408(9), 421, 422(18) Spofford, J. B., 386
Silver, L. M., 372, 374, 376(1), 379, 380, Staes, M., 315
380(1), 394 Stahl, S. J., 136
Silver, P. A., 317 Stallcup, M. R., 231, 272
Silverman, N., 120 Stanik, V. H., 20
Simmler, M. C., 420 Starai, V. J., 172(14), 173, 174
Simon, I., 290, 314, 317, 322(13), 333(13), 350, Stebbins, J., 172(15), 173, 175(15), 177(15),
353(11), 357(11) 182(15), 183(15)
Simon, J. A., 248, 252, 254(28), 262, 290, 394 Steffan, J. S., 200
Simons, S. S., Jr., 28 Stelzer, G., 316
Simonson, T., 139 Stepaniants, S. B., 290
Simpson, M., 335, 348(5) Stephens, G. E., 372, 373, 387, 387(8), 403
Sinclair, D. A., 185, 386 Sternberger, L. A., 380
Singer, R. H., 411 Sterner, D. E., 106, 108, 109(15), 111(15),
Singh, N., 144 112(15), 114(15), 120
author index 445
Sternglanz, R., 112, 172(15; 16), 173, 175, Tafrov, S. T., 112, 172(15), 173, 175(15),
175(15), 177(15; 16; 23), 180(40), 177(15), 182(15), 183(15)
182(15), 183(15; 23), 185(23) Tagne, J. B., 314, 317
Steven, A. C., 42 Tahourdin, C. S., 214
Steward, R., 235, 248, 248(11), Takada, S., 395
254(11; 27), 255 Takagi, N., 405, 408(4), 417, 417(4),
Stewart, A. F., 119, 147, 152(4) 418(4), 422
Stoffler, G., 50 Takahashi, Y., 306, 318, 360
Stoffler-Meilicke, M., 50 Tamkun, J. W., 3, 120
Stollar, B. D., 212 Tan, E. M., 217
Stott, F. J., 149, 151(6), 168(6), 169(6) Tan, S., 65, 106(11; 12), 107, 120
Stott, K. M., 149, 167(5) Tang, H., 140, 255
Stoughton, R., 290, 313, 359 Tanner, K. G., 114, 172(16), 173, 177(16),
Stowell, M., 60 178, 194, 223, 232(15)
Strahl, B. D., 106, 121, 130(23), 131, 151, Tanny, J. C., 173, 182(19), 183(19)
199, 222, 231, 234(2), 235, 255, 269, 287, Tao, Y., 43
387, 388 Tastan, S., 420
Strahl-Bolsinger, S., 290, 292, 300(14), Tattersall, P., 316
304(14) Taunton, J., 174, 189, 199
Strausberg, R. L., 189, 317 Tavazoie, S., 290, 294(9), 303, 360
Strauss, W., 420 Taverna, S. D., 131, 132(8), 134(8), 140,
Strick, T. R., 76, 99, 101(6), 102(5) 140(8), 142(8; 29), 144(8)
Struhl, K., 235, 254, 289, 317, 336 Taylor, I. A., 270
Strutt, H., 349 Taylor, J. P., 199
Su, F., 173 Tempst, P., 120, 231, 235, 248, 252,
Subbaramaiah, K., 195 254(27; 28), 255, 262, 317, 332, 395, 408
Sugawara, O., 405, 408(4), 417(4), 418(4) Teng, G., 428
Suka, N., 209, 217(7), 289, 290, 293(1), 300, Teplyakov, A., 139
300(10), 301, 302(10), 303, 314(8), 315, Terragni, J., 306, 318, 360
349, 360(2) Thåström, A., 5, 12
Suka, Y., 209, 217(7), 289, 290, 293(1), Thijs, G., 315
300(10), 301, 302(10), 303, 314(8), 315, Thion, L., 316
349, 360(2) Thireos, G., 120
Sun, Z. W., 255, 287, 387 Thiru, A., 149, 151(7), 167(5; 7)
Surani, M. A., 235 Thiry, M., 33(26), 34
Sutton, A., 172(15), 173, 175(15), 177(15), Thoma, F., 79
182(15), 183(15) Thomas, J. O., 131, 151, 157(13), 170(13),
Sutton, C. A., 394 386, 387(28)
Svedberg, T., 161, 162(22) Thompson, B. A., 373, 387, 387(8)
Svoboda, K., 62 Thompson, C. M., 314, 317
Swat, W., 335 Thompson, L. M., 200
Switz, N. A., 82 Thompson, P. R., 115, 188, 191, 192, 193(13)
Syntichaki, P., 120 Thorne, A. W., 222, 336
Szyperski, T., 124 Thornton, J. M., 125, 210, 218(8)
Thuman-Commike, P. A., 38(10), 39
Tinoco, I., 62
T Tischendorf, G. W., 50
Tachibana, M., 237, 238(19), 245(19), Tjian, R., 127, 137, 144(20), 395
245(19), 249(19), 254(19) Tomschik, M., 76, 97, 103(32)
Taddei, A., 240, 271, 428 Tong, J. K., 290, 301, 303, 360
446 author index
Tonlorenzi, R., 420 Victor, M., 196, 197(25)

Topalidou, I., 120 Vidal, M., 316
Travers, A. A., 125, 137 Villnave, C., 411
Trievel, R. C., 108, 109(14–16), 111(14–1 6), Viovy, J. L., 92
112, 112(14–16), 113(21), 114, Visisht, N., 270
114(15; 16; 21), 117(16), 128, 194 Voelker, P. J., 162, 165(25)
Trowsdale, J., 120 Vogelauer, M., 289
Truner, B. M., 222 Voit, R., 174
Tsao, J., 315 Volkert, T. L., 290, 306, 314, 317, 318,
Tse, C., 17, 20(7) 322(13), 333(13), 360
Tsukiyama, T., 3 Volpe, T. A., 428
Turgeon, C. L., 17 Voordouw, G., 5
Turner, B. M., 106(13), 107, 199, 209, 222, Vossen, C., 316
223, 234(3), 235, 255, 267, 272, 283(16), Vuister, G. W., 123, 124, 136
289, 392, 405, 406, 408(13), 421
Tyler, J. M., 32
W
Wade, P., 33
U
Wade, W., 123
Unwin, N., 60 Wagenknecht, T., 53
Ura, K., 64 Wagner, G., 112
Utley, R. T., 235 Walker, K. P. III, 17, 20(7)
Walker, P. A., 270
Walker, W. L., 313, 359
V
Wallrath, L. L., 372, 390, 403
Vagin, A., 139 Walz, J., 45, 47(20)
Valckx, D., 343 Wang, A., 289, 290, 290(3), 300(10), 301,
Valenzuela, P., 248, 254(27), 255 302(10), 303, 314(8), 315, 349, 360(2)
van Heel, M., 39, 41 Wang, H., 231, 235, 248, 252, 254(28; 29),
van Holde, K. E., 20, 76, 162, 221, 231(2), 234, 260, 262, 332, 405, 406(10), 408(10),
235(1), 237(1) 421, 422(19)
Van Hooser, A., 232 Wang, J. C., 336
van Leeuwen, F., 235 Wang, L., 108, 109(14; 15), 111(14; 15),
Van Lint, C., 174 112(14; 15), 114(15), 117(18), 188, 252,
Van Regenmortel, M. H., 209, 212(4) 262, 315, 332
Vaquero, A., 235, 255 Wang, M. D., 62, 63, 65, 65(8; 13), 67(5; 8; 13),
Varambally, S., 270 68(5; 8; 13), 69(13), 71(3; 8), 95
Vargas, M. L., 252, 262 Wang, Y., 248, 254(27), 255, 334
Varga-Weisz, P. D., 3 Wang, Z. Y., 195
Varshavsky, A., 336, 349 Ward, T. R., 304
Vassilev, A., 114, 117(30), 119, 189, 191(7), Warren, G. L., 139
193, 193(7), 194(14), 223 Waterborg, J. H., 237
Vaziri, H., 173 Watson, J., 318
Venkataramani, R., 108, 109(15), 111(15), Webb, W. W., 82
112(15), 114(15) Webster, Z., 248, 249(30), 254(30), 260, 405,
Verdel, A., 171 406(9), 408(9), 421, 422(18)
Verdin, E., 128, 129(53), 130, 130(53), 173, Wegener, D., 176
174, 177(24), 197 Wei, S. H., 328, 333(30)
Verschoor, A., 53 Wei, Y., 232
Vervish, A., 192, 197(12) Weinberg, R. A., 173
author index 447
Weinmann, A. S., 316, 318, 320, 322(25), Wu, C. E. C., 374

331(25), 360 Wu, J., 171, 209, 217(7), 289, 293(1)
Weipoltshammer, K., 237, 238(20), 240(20), Wu, K. H., 428
241(20), 242(20), 246(20), 247(20) Wu, R., 237, 242(22), 245(22),
Weissman, S., 317 248(22), 254(22)
Weller, T. H., 380 Wurtz, J. M., 119, 120(4), 121(4), 127(4)
Wells, J., 316, 318, 332 Wüthrich, K., 124
Werner, J., 248, 254(27), 255, 372, 373(4), Wutz, A., 406, 420
374(4), 393, 393(4), 395, 404 Wynne, S. A., 60
Westwood, J. T., 394 Wyrick, J. J., 290, 314, 314(9), 315, 317,
White, D. A., 209, 223 322(13), 333(13), 350, 353(11), 357(11)
Whiteside, G. M., 164
Widom, J., 3, 5, 12, 12(7)
Wiess, K., 119 X
Wild, B., 252, 262 Xenarios, I., 290, 300(10), 301, 302(10), 303,
Wilde, F., 215 314(8), 315, 349, 360(2)
Willard, H. F., 405, 420 Xia, L., 231, 248, 252, 254(28), 262, 332
Williams, C. A., 380 Xiang, C. C., 323
Williams, J. W., 162 Xiao, B., 270
Willis, A. C., 223, 229(18) Xiao, H., 397
Williston, S., 146, 147(40)
Wilson, B., 317
Wilson, C. J., 290, 317, 322(13), 333(13) Y
Wilson, J. R., 270
Wilson, K. S., 139 Yamazaki, T., 123
Winegarden, N. A., 394 Yan, P. S., 318, 328, 333, 333(30), 360
Wingender, E., 315 Yan, Y., 109, 111(19), 112, 112(19),
Wingfield, P. T., 136 114(19; 23), 118(19), 193, 198
Winston, F., 120, 121 Yang, G., 76
Wirsching, F., 176 Yang, J. C., 125, 137
Wiseman, T., 146, 147(40) Yang, X. J., 112
Wodicka, L., 350 Yaniv, M., 120
Wolberger, C., 172(14), 173 Yau, P., 315
Wolf, A., 387 Yeh, R. C., 62, 65(8), 67(8), 68(8), 71(8)
Wolf, S. F., 420 Yin, H., 62, 63, 65(13), 67(13), 68(13), 69(13),
Wolff, I., 389 71(3), 95
Wolffe, A. P., 64, 114, 171, 189, 191(7), Yoshida, M., 189, 201
193(7), 221, 232(1) Young, R. A., 254, 290, 304, 314, 314(9), 315,
Wong, J., 196, 197, 231 317, 318, 322(13), 333(13), 360
Wong, K. S., 394 Yphantis, D. A., 164
Wong, W. W., 318
Woodcock, C. L., 18, 29, 30, 32, 34, 35(4), 36,
39, 41(11), 45, 47, 47(20)
Z
Woodcock, H., 34, 35(4) Zaniewski, E., 28
Woodman, B., 200 Zaret, K. S., 361, 368(3)
Workman, J. L., 106, 106(13), 107, 120 Zegerman, P., 131, 151, 157(13), 170(13), 386,
Worringer, K. A., 248, 254(29), 260, 405, 387(28)
406(10), 408(10), 421, 422(19) Zeichhardt, H., 50
Wright, P. E., 125, 130(44), 137 Zeitlinger, J., 290, 314, 317, 322(13), 333(13),
Wu, C., 3 350, 353(11), 357(11)
448 author index
Zeng, L., 121, 125(20), 126, 127(20), 128, Zhou, J., 106, 108, 109(15), 111(15; 16), 112,
128(20), 129(53), 130(20; 53), 137, 112(15; 16), 114(15; 16), 117(16), 120,
233, 304 189, 194, 199
Zhang, C. L., 198 Zhou, J. X., 108, 117(17)
Zhang, H. B., 128 Zhou, M.-M., 119, 121, 123, 125(20), 126,
Zhang, K., 255 127(20), 128, 128(20), 129(53), 130,
Zhang, M. Q., 315, 316, 328 130(20; 53), 137, 233, 270, 304
Zhang, T., 316 Zhou, S., 152, 395
Zhang, W., 43, 267 Zhou, X., 200
Zhang, Y., 106, 131, 132(8), 134(8), 140(8), Zhu, G., 123, 136
142(8), 144(8), 175, 221, 231, 233(4), 235, Zhu, Y. Z., 200
248, 252, 254(28; 29), 255, 260, 262, 272, Zink, D., 152
332, 405, 406(10), 310(10), 421, 422(19) Zlatanova, J., 73, 74, 76, 77(3), 78(3), 94,
Zhatnova, J., 62 95(29), 96(29), 97, 103(32)
Zheng, W., 190 Zvetkova, I., 248, 249(30), 254(30), 260, 405,
Zheng, Y., 188 406(9), 408(9), 421, 422(18)
Subject Index
A image acquisition, 79
materials, 78
Acetylation microarray surface preparation and sample
applications, 302–304 deposition, 79
chromatin immunoprecipitation instrumentation, 76–78
cell growth and harvesting, 292 principles, 73–74
cross-linking AUT gel electrophoresis, see
double cross-linking, 294–295 Acid-urea-triton gel electrophoresis
formaldehyde, 292
immunoprecipitation reaction, 293
polymerase chain reaction, 295–297 B
sonication, 292–293
troubleshooting, 293–294 Brahma, bromodomain structure, see
DNA microarray analysis Nuclear magnetic resonance
data quantification, normalization, and Bromodomain structure, see Nuclear
analysis, 299–302 magnetic resonance
Klenow labeling of probe and
hybridization, 297–299
C
yeast intergenic microarray
preparation, 302 ChIP, see Chromatin immunoprecipitation
principles, 290–292 Chromatin, see also Histone; Nucleosome
Acid-urea-triton gel electrophoresis, histone agarose multigel electrophoresis analysis
deacetylase assay for inhibitor of structure, see Agarose multigel
characterization, 201–202 electrophoresis
AFM, see Atomic force microscopy antibodies, see also Histone antibodies
Agarose multigel electrophoresis applications, overview, 209–210
chromatin structure analysis histone antigen interactions, 209–212
composition analysis of genomic immunogen preparation
fragments, 27–28 dehistonized chromatin, 214–215
data interpretation, 28–29 high mobility group proteins, 215–216
flexibility assay, 27 histone fractions, 213–214
overview, 17–18, 20–21 nucleic acids, 216
secondary chromatin structure nucleosomes, 214
formation, 26 synthetic histone peptides, 212–213
data analysis, 25–26 preparation, 216–217
electrophoresis, 24 specificity and affinity, 217–220
equipment, 21 electron microscopy three-dimensional
gel preparation, 22, 24 structure determination
principles, 18–19 adhesion to support films, 31
Southern blot, 25 buffer components, 30
AME, see Agarose multigel electrophoresis concentration of sample, 30
Atomic force microscopy cryomicroscopy, 36
chromatin fiber imaging electron tomography
glutaraldehyde fixation, 78–79 alignment and reconstruction, 46–47
449
450 subject index
Chromatin (cont.) primary antibody coupling to protein

data collection, 45–46 A beads, 369
overview, 43–44 protein A bead preparation and
resolution determination, 47 preclearing, 369
specimen preparation, 44 sonication, 366–367
visualization and analysis, 47–48 tissue chromatin dialysis and
film versus charge-coupled device ribonuclease treatment, 367–368
camera in data acquisition, 36–37 CpG island microarray screening of
fixation, 31 chromatin immunoprecipitates for
microscope parameters, 37 promoter identification
negative staining, 34–35 advantages, 315–316
positive staining, 31–32 chromatin immunoprecipitation, 318,
purity requirements, 29–30 320–321
shadow casting, 32, 34 fluorophore labeling of DNA, 324–326
single-particle reconstruction ligation-mediated polymerase chain
accuracy of reconstruction, 41 reaction, 322–325
class average alignment and microarrays
generation, 39–40 analysis, 330, 332–333
cryomicroscopy data, 42–43 hybridization, 329–330
data preparation, 38–39 preparation, 328–329
final reconstruction parameters, overview, 316–318
40–41 prospects, 333
overview, 38 genome-wide analysis of histone
sample preparation, 38 modifications in yeast with chromatin
visualization, 41–42 immunoprecipitation and DNA
RSC chromatin remodeling complex microarrays
interactions, see RSC chromatin chromatin immunoprecipitation,
remodeling complex 350–353
single-molecule dynamics studies, see DNA amplification and labeling
Atomic force microscopy; Magnetic amplification by in vitro transcription,
tweezers; Optical trapping 355–356
Chromatin immunoprecipitation fluorescent labeling, 357
acetylation microarrays, see Acetylation random-primer polymerase chain
microarray reaction, 354–356
animal tissues DNA microarray
cesium chloride step gradient, 366 data acquisition and processing, 360
DNA purification, 371–372 hybridization, 357–358
formaldehyde cross-linking interpretation, 359–360
liver retrograde perfusion and principles, 349–350
chromatin cross-linking, prospects, 360
362–365 genome-wide transcription factor binding
nuclear isolation and chromatin site identification with chromatin
cross-linking, 365–366 immunoprecipitation and DNA
overview, 361–362 microarray screening
restriction enzyme digestion, 368 applications, 314
sequential immunoprecipitation chromatin preparation, 308
chromatin antigen release from DNA purification, 309
beads, 371 fluorophore labeling of DNA, 311
immunoprecipitation conditions, 371 formaldehyde cross-linking, 308
overview, 368 immunoprecipitation reaction, 309
subject index 451
ligation-mediated polymerase chain fluorescein labeling of

reaction, 310–311 peptides, 142–143
microarray scanning and analysis, instrumentation, 141–142
312–314 overview, 139–141
overview, 304–307 probe selection, 142
promoter microarray fabrication and fluorescence spectroscopy, histone tail
hybridization, 311–312 interaction studies
solutions, 307 overview, 139, 157–158
immunoglobulin heavy chain locus data analysis, 160
analysis of histone modification methylated peptide specificity
patterns in mice studies, 161
cell culture, 336–337 sample preparation, 158
DNA precipitation, 340 titrations, 158–160
DNA yield, 340–341 isothermal titration calorimetry,
formaldehyde fixation, 337 chromodomain–histone tail
histone acetylation and methylation interaction studies
analysis, 348 heterochromatin protein
mmunoprecipitation reaction, 338–339 1 chromodomain interactions with
nuclear isolation and micrococcal tail peptides containing
nuclease digestion, 337, 346, 348 dimethyllysine or trimethyllysine,
principles, 334–336 147–148
real-time quantitative polymerase chain practical considerations, 146–147
reaction principles, 143–146
amplification reactions, 344–345 nuclear magnetic resonance,
calculations, 345 heterochromatin protein
fluorescent probes, 344 1 chromodomain studies
primer design, 342–343 backbone dynamics studies upon ligand
principles, 343–344 binding, 137
sonication, 337–338, 346 binding site identification, 168–170
methyl-histone H3 antibodies dissociation constant
cell harvesting, 283 determination, 167
DNA quantification, 285 histone H3 tail binding mapping,
immunoprecipitation reaction, 285–286 136–137, 167–168
micrococcal nuclease digestion, 283 peptide preparation, 135–136, 156–157
principles, 282–283 protein preparation, 135, 153–154
quantitative polymerase chain protein–peptide interaction screening,
reaction, 286 132, 134, 136, 157
sonication, 283–284 requirements, 132
Chromodomain sample preparation, 155–156
analytical ultracentrifugation studies purification of heterochromatin
applications, 161 protein 1 domains
data acquisition, 164–165 nickel affinity chromatography,
experimental design, 162–163 153–154
sample preparation, 163–164 polyhistidine tag cleavage, 154
sedimentation equilibrium, 161–162 recombinant protein expression in
definition, 131 bacteria, 152–153
fluorescence anisotropy, histone tail sequence analysis and identification, 149,
interaction studies 151–152
binding affinity calculation, 143 sequence homology, 150
concentrations of analytes, 141 subfamilies, 148
452 subject index
Chromodomain (cont.) fluorophore labeling of DNA, 311

x-ray crystallography, heterochromatin formaldehyde cross-linking, 308
protein 1 chromodomain–histone immunoprecipitation reaction, 309
tail structure, 134, 137–139 ligation-mediated polymerase chain
CpG island microarray, see DNA microarray reaction, 310–311
microarray scanning and analysis,
312–314
D
overview, 304–307
DNA microarray, see also Acetylation promoter microarray fabrication and
microarray hybridization, 311–312
CpG island microarray screening of solutions, 307
chromatin immunoprecipitates for
promoter identification
E
advantages, 315–316
chromatin immunoprecipitation, Electron microscopy, see Transmission
318, 320–321 electron microscopy
fluorophore labeling of DNA, 325–327 ELISA, see Enzyme-linked immunosorbent
ligation-mediated polymerase chain assay
reaction, 322–325 Embryonic stem cell, see X chromosome
microarrays inactivation
analysis, 330, 332–333 Enzyme-linked immunosorbent assay
hybridization, 329–330 di-modified histone H3 antibodies, 224,
preparation, 328–329 226, 229
overview, 316–318 methyl-histone H3 antibodies, 238,
prospects, 333 256–257, 268, 276–277, 279–280
genome-wide analysis of histone
modifications in yeast with chromatin
F
immunoprecipitation and DNA
microarrays FISH, see Fluorescence in situ hybridization
chromatin immunoprecipitation, Fluorescence anisotropy
350–353 chromodomain–histone tail interaction
DNA amplification and labeling studies
amplification by in vitro transcription, binding affinity calculation, 143
355–356 concentrations of analytes, 141
fluorescent labeling, 357 fluorescein labeling of peptides,
random-primer polymerase chain 142–143
reaction, 354–356 instrumentation, 141–142
DNA microarray overview, 139–141
data acquisition and processing, 359 probe selection, 142
hybridization, 357–358 imitation-SWI–nucleosome interactions
interpretation, 359–360 cuvettes, 11–12
principles, 349–350 emission path filters, 7–10
prospects, 360 excitation path filter, 7, 9–10
genome-wide transcription factor binding fluorescein-labeled DNA preparation, 4
site identification with chromatin instrumentation, 5
immunoprecipitation and DNA intensity changes, 11
microarray screening light scattering artifacts, 6–7
applications, 314 nucleosome preparation, 5
chromatin preparation, 308 pH sensitivity, 11
DNA purification, 309 sample prepration, 6
subject index 453
titration studies, 12–16 sequence homology, 110

principles, 3–4 structures, 112
Fluorescence in situ hybridization catalytic reaction, 188–189
X chromosome inactivation detection, inhibitors
415, 417 delivery
Xist RNA fluorescence in situ hybridization microinjection, 196–197
early events in X inactivation, 406, 421 sphingosylphoshoryl choline delivery,
hybridization, 414, 426 197–198
probes, 414, 425–426 design and synthesis, 190–191
template preparation, 424–425 gene expression studies, 194–195
transcription, in vitro, 425 history of study, 189–190
troubleshooting, 427–428 kinetic and structural studies, 191–194
washing and detection, 426–427 transcriptional activation studies,
in vitro, 195–196
types and histone specificity, 106–107
G
Histone antibodies
GCN5, inhibitor studies, 191, 194, 198 acetyl-histone antibodies, 222
di-modified histone H3 antibodies
masking effects, 233–234
H
modification density analysis, 231–233
HAT, see Histone acetyltransferase overview, 222–223
Heterochromatin protein 1 preparation
chromodomain, see Chromodomain affinity purification, 227–229
immunofluorescence microscopy on ammonium sulfate precipitation,
polytene chromosomes, 386–388 226–227
High mobility group proteins, immunogen enzyme-linked immunosorbent assay,
preparation, 215–216 224, 226, 229
Histone peptide coupling to keyhole limpet
antibodies, see Histone antibodies hemocyanin, 224
modifications, see also specific enzymes peptide synthesis, 223
antibodies, see Histone antibodies rabbit immunization, 224
functional overview, 221–222 specificity verification, 229, 231
genome-wide analysis, see Acetylation immunogen preparation
microarray; Chromatin histone fractions, 213–214
immunoprecipitation; DNA nucleosomes, 214
microarray synthetic histone peptides, 212–213
types, 221, 234 methyl-histone H3 antibodies
Histone acetyltransferase, see also Brahma; affinity purification of antibodies,
GCN5; p300/CREB-binding protein 277–278
bromodomain structure, see Nuclear antigen design, 237–238, 240, 272–273
magnetic resonance challenges in study, 251–252
catalytic domains chemical methylation of lysines on
activity assays, 112–114 peptide antigens, 275
catalytic mechanisms, 115, 117 chromatin immunoprecipitation analysis
crystallization, 111–112 cell harvesting, 283
histone-binding specificity, 117–118 DNA quantification, 285
identification for expression in bacteria, immunoprecipitation reaction,
108–109 285–286
purification of recombinant proteins, micrococcal nuclease digestion, 283
109, 111 principles, 282–283
454 subject index
Histone antibodies (cont.) enzyme sources, 286–287

quantitative polymerase chain gel electrophoresis analysis, 287–288
reaction, 286 liquid assay, 288
sonication, 283–284 substrates, 287
dot blot analysis, 244–245, 257–258, 260, chromodomain, see Chromodomain
262, 268 histone methylation sites, 235, 237, 255,
enzyme-linked immunosorbent assay, 269–270
238, 256–257, 268, 276–277, 279–280 protein arginine methyltransferase
immunofluorescence microscopy, classes, 272
247–251, 263–264, 268–269, 281–282 SET domains, 270
methylation pattern characterization SET7/9 structure, 270–271
in vivo, 247–248 trimethylation of active genes, 270
overview, 237 HMG proteins, see High mobility
peptide synthesis, 273–275 group proteins
practical guidelines, 253–254 HP1, see Heterochromatin protein 1
rabbit immunization, 276
trimethyl-specific antibody preparation
I
dot blot analysis of specificity,
244–245 Imitation-SWI, fluorescence anisotropy
immunoglobulin G fractionation, 244 studies of nucleosome interactions
peptide synthesis, 240–243 cuvettes, 11–12
rabbit immunization, 243–244 emission path filters, 7–10
Western blot analysis, 262–263, 268, excitation path filter, 7, 9–10
279–281 fluorescein-labeled DNA preparation, 4
yeast assay of specificity, 267 instrumentation, 5
Histone deacetylase, see also Sir2 intensity changes, 11
activity assays light scattering artifacts, 6–7
acetate detection, 174–175 nucleosome preparation, 5
detection of acetylated and deacetylated pH sensitivity, 11
substrates, 175–176 principles, 3–4
bromodomain structure, see Nuclear sample prepration, 6
magnetic resonance titration studies, 12–16
classes, 171, 189 Immunofluorescence microscopy
inhibitors histone modification detection in
assays X chromosome inactivation
acid-urea-triton gel electrophoresis, antibody staining, 412–413, 423–424
201–202 early events in X inactivation, 406, 408,
histone isolation, 200–201 419, 421
immunohistochemistry, 204–205 fixation, 411
Western blot, 203–204 metaphase chromosome spreads, 415
overview, 189, 199–200 microscopy, 413
nonhistone substrates, 173–174 methyl-histone H3 antibodies, 247–251,
transcriptional influences, see Acetylation 263–264, 268–269, 282–283
microarray polytene chromosome proteins of
types and functions in yeast, 289 Drosophila
Histone H3 advantages and limitations, 402–404
antibodies, see Histone antibodies antibody specificity testing, 395
methylation sites, 235, 255, 270 antibody staining, 379–381, 400–401
Histone methyltransferase custom P-element studies, 390, 392
assays fixation, 374, 376–377
subject index 455
fly culture, 399 heterochromatin protein 1 chromodomain

genetic manipulation of flies, 373, 385, interactions with tail peptides
396–397 containing dimethyllysine or
heat shocking of larvae, 399 trimethyllysine, 147–148
image analysis, 401–402 practical considerations, 146–147
microscopy, 381–382, 401 principles, 143–146
multiple staining of heterochromatin ISWI, see Imitation-SWI
proteins and histone modifications, ITC, see Isothermal titration calorimetry
386–388
mutant fly studies, 389–390
non-melanogaster species analysis, L
388–389 Lys-CoA
overview, 372–374 histone acetyltransferase inhibition
prospects, 392–393 delivery
protein colocalization studies, 394–395 microinjection, 196–197
salivary gland dissection, 399 sphingosylphoshoryl choline delivery,
sensitivity enhancement, 396–397 197–198
spatial resolution, 395–396, 402 gene expression studies, 194–195
squashing, 377–379, 400 history of study, 189–190
troubleshooting, 382–385 kinetic and structural studies, 191–194
Immunoglobulin heavy chain locus transcriptional activation studies,
chromatin immunoprecipitation analysis in vitro, 195–196
of histone modification patterns synthesis, 190–192
in mice
cell culture, 336–337
DNA precipitation, 340 M
DNA yield, 340–341
formaldehyde fixation, 337 Magnetic tweezers
histone acetylation and methylation chromatin studies
analysis, 348 force calibration, 101–103
immunoprecipitation reaction, 338–339 materials, 96–97
nuclear isolation and micrococcal single chromatin fiber assembly and
nuclease digestion, 337, 346, 348 analysis, 103
principles, 334–336 tether length changes with chromatin
real-time quantitative polymerase chain assembly, 101–103
reaction tips, 103–105
amplification reactions, 344–345 cuvette preparation and DNA
calculations, 345 immobilization, 99–101
fluorescent probes, 344 DNA biotinylation, 99
primer design, 342–343 instrumentation, 97–99
principles, 343–344 magnetic bead localization, 101
sonication, 337–338, 346 principles, 74, 76
recombination, 335
segments, 335
N
Immunohistochemistry, histone deacetylase
assay for inhibitor characterization, NMR, see Nuclear magnetic resonance
204–205 Nuclear magnetic resonance
Isothermal titration calorimetry, bromodomain structure-function studies in
chromodomain–histone tail histone acetyltransferases
interaction studies amide exchange, 124
456 subject index
Nuclear magnetic resonance (cont.) experimental samples, 66–67

backbone assignments, 123 histones, 64–65
data acquisition for structure data analysis
determination, 123 force clamp, 71
dihedral angle restraints, 124 velocity clamp, 69, 71
ligand binding DNA elasticity determination, 68–69, 93–95
assays, 129–130 instrumentation
binding site analysis, 127–129 calibration, 68
methyl groups, 124 flow cell
NOESY distance restraints, 124 construction, 83–86
overview, 119–121 micropipette placement, 86
p300/CREB-binding protein mounting, 86
bromodomain structure, 125, 127 flow system, 86–88
protein expression and isotope labeling, optics, 81–83
121–122 overview, 63, 67–68, 80–81
protein purification, 122–123 nucleosome disruption, 68
side-chain assignments, 123 principles, 62–64, 74, 80
structure refinement, 124–125 prospects for chromatin studies, 72
heterochromatin protein 1
chromodomain studies
backbone dynamics studies upon ligand
P
binding, 137 p300/CREB-binding protein
binding site identification, 168–170 inhibitor studies, 188, 192–196, 198
dissociation constant nuclear magnetic resonance of
determination, 167 bromodomain structure and function
histone H3 tail binding mapping, amide exchange, 124
136–137, 167–168 backbone assignments, 123
peptide preparation, 135–136, 156–157 data acquisition for structure
protein preparation, 135, 153–154 determination, 123
protein–peptide interaction screening, dihedral angle restraints, 124
132, 134, 136, 157 methyl groups, 124
requirements, 132 NOESY distance restraints, 124
sample preparation, 155–156 overview, 119–121
Nucleosome protein expression and isotope labeling,
fluorescence anisotropy studies of 121–122
chromatin remodeling complex protein purification, 122–123
interactions, see Fluorescence side-chain assignments, 123
anisotropy structure overview, 125, 127
optical trapping studies of nucleosome structure refinement, 124–125
dynamics, see Optical trapping PCAF, see p300/CREB-binding protein
structure, 221 PCR, see Polymerase chain reaction
Polymerase chain reaction
acetylation microarrays, 295–297
O
ligation-mediated polymerase chain
Optical trapping reaction for genome-wide
chromatin opening event analysis, 95–96 transcription factor binding site
chromatin sample preparation identification with chromatin
assembly, 65–66, 92–93 immunoprecipitation and DNA
DNA labeling and attachment, 65, microarray screening, 310–311,
88–92 323–325
subject index 457
methyl-histone H3 antibody R
chromatin immunoprecipitation
analysis, 286 RSC chromatin remodeling complex,
random-primer polymerase chain reaction electron microscopy
for genome-wide analysis of histone nucleosome interaction imaging, 55–59
modifications in yeast with chromatin sample requirements and preparation,
immunoprecipitation and DNA 48–50
microarrays, 354–356 specimen evaluation and data collection,
real-time quantitative polymerase chain 50–51
reaction in chromatin three-dimensional structure reconstruction
immunoprecipitation analysis initial calculations, 53
amplification reactions, 344–345 interpretation, 53–55
calculations, 345 two-dimensional image analysis and
fluorescent probes, 344 classification, 51–52
primer design, 342–343 unstained sample studies, 59–62
principles, 343–344
Polytene chromosome
amplification in development, 372, 393 S
immunofluorescence microscopy of Sir2
Drosophila salivary gland activity assays
chromosome proteins
charcoal-binding assay, 180–182
advantages and limitations, 402–404 high-performance liquid
antibody specificity testing, 395 chromatography of peptides,
antibody staining, 379–381, 400–401 178–180
custom P-element studies, 390, 392
kinetic parameterr analysis, 186–187
fixation, 374, 376–377 nicotinamide assays
fly culture, 399 catalyzed exchange assay, 183–195
genetic manipulation of flies, 373, 385, inhibition assays, 185
396–397
substrate preparation, 177
heat shocking of larvae, 399 thin-layer chromatography assay,
image analysis, 401–402 182–183
microscopy, 381–382, 401
catalytic mechanism, 173
multiple staining of heterochromatin functions, 172–173
proteins and histone modifications, Southern blot, agarose multigel
386–388 electrophoresis, 25
mutant fly studies, 389–390
non-melanogaster species analysis,
388–389
T
overview, 372–374
prospects, 392–393 TEM, see Transmission electron microscopy
protein colocalization studies, Transmission electron microscopy
394–395 chromatin three-dimensional structure
salivary gland dissection, 399 determination
sensitivity enhancement, 396–397 adhesion to support films, 31
spatial resolution, 395–396, 402 buffer components, 30
squashing, 377–379, 400 concentration of sample, 30
troubleshooting, 382–385 cryomicroscopy, 36
structure, 393–394 electron tomography
Protein arginine methyltransferases, see alignment and reconstruction, 46–47
Histone methyltransferase data collection, 45–46
458 subject index
Transmission electron microscopy (cont.) cell culture

overview, 43–44 differentiation induction, 410–411,
resolution determination, 47 421, 423
specimen preparation, 44 feeder fibroblasts, 408–409
visualization and analysis, 47–48 maintenance culture, 409–410
film versus charge-coupled device histone modification detection with
camera in data acquisition, 36–37 immunofluorescence microscopy
fixation, 31 antibody staining, 412–413, 423–424
microscope parameters, 37 early events in X inactivation, 406,
negative staining, 34–35 408, 419, 421
positive staining, 31–32 fixation, 411
purity requirements, 29–30 metaphase chromosome spreads, 415
shadow casting, 32, 34 microscopy, 413
single-particle reconstruction overview, 405–406
accuracy of reconstruction, 41 replication timing studies using
class average alignment and bromodeoxyuridine incorporation
generation, 39–40 and Acridine Orange staining,
cryomicroscopy data, 42–43 417–418
data preparation, 38–39 X chromosome fluorescence in situ
final reconstruction parameters, 40–41 hybridization, 415, 417
overview, 38 Xist RNA fluorescence in situ
sample preparation, 38 hybridization
visualization, 41–42 early events in X inactivation,
RSC chromatin remodeling complex 406, 421
nucleosome interaction imaging, 55–59 hybridization, 414, 426
sample requirements and preparation, probes, 414, 425–426
48–50 template preparation, 424–425
specimen evaluation and data collection, transcription, in vitro, 425
50–51 troubleshooting, 427–428
three-dimensional structure washing and detection, 426–427
reconstruction features of inactivated chromosome, 405
initial calculations, 53 mechanism, 405
interpretation, 53–55 phases, 419
two-dimensional image analysis and Xist role, 405, 420–421
classification, 51–52 X-ray crystallography
unstained sample studies, 59–62 heterochromatin protein
1 chromodomain–histone tail
structure, 134, 137–139
W
histone acetyltransferase catalytic
Western blot domains
histone deacetylase assay for inhibitor crystallization, 111–112
characterization, 203–204 histone-binding specificity, 117–118
methyl-histone H3 antibodies, 262–263, purification of recombinant proteins,
268, 279–281 109, 111
structures, 112
X
X chromosome inactivation
embryonic stem cell studies in mice

Methods in Enzymology Chromatin and Chromatin Remodeling Enzymes

Uploaded by

Copyright:

Available Formats

Methods in Enzymology Chromatin and Chromatin Remodeling Enzymes

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Methods in Enzymology Chromatin and Chromatin Remodeling Enzymes

Uploaded by

Copyright:

Available Formats

Preface

A central challenge of the post-genomic era is to understand how the 30,000 to

protocols for nucleosome reconstitution and analysis, and cytological methods

Editors’ Note: Additional methods can be found in Methods in Enzymology,

John N. Abelson Melvin I. Simon

Sidney P. Colowick and Nathan O. Kaplan

Rhoda M. Alani (12), Department of Juliana Callaghan (10), Department of

Carolyn A. Craig (25), Biology Depart- Jeffrey C. Hansen (2), Department of

Chukwudi Ezeokonkwo (4), Department Thomas Jenuwein (16), Research Institute

Wolfgang Fischle (9), Department of W. Kevin Kelly (13), Genitourinary On-

Fred K. Friedman (14), Laboratory of Me- Sepideh Khorasanizadeh (9), Depart-

Philippe T. Georgel (2), Department of Roger D. Kornberg (4), Department of

Siavash K. Kurdistani (19), Department Shiraz Mujtaba (8), Department of Physi-

Antoine H. F. M. Peters (16), Research Paul R. Thompson (12), Department of

[1] Fluorescence Anisotropy Assays for Analysis of

Fluorescence anisotropy is a rapid, sensitive, and quantitative technique

Copyright 2004, Elsevier Inc.

Instrumentation and Technical Considerations

can in practice be so great that, when combined with a relatively strong

Excitation Path Filter

Emission Path Filters

longer wavelengths with high transmittance, with a steep rise in transmit-

Filter Set for Use with Fluorescein

Excitation interference Coherent/Ealing #35-3482 interference filter,

Figure 1 shows excitation and emission spectra for fluorescein in panel

Fig. 1. Fluorescence spectra of fluorescein compared to transmission spectra of the optical

Experimental Design for Binding Titrations

Results of Binding Experiments

binding curve as plotted in this manner is too ‘‘steep’’ to correspond to

Fig. 4. Raw anisotropy data when no binding occurs. A 1 nM solution of a 50 -fluorescein–

Fig. 5. ISWI binding to a nucleosomal DNA. A 5-nM solution of a 50 -fluorescein–labeled

Finally, in Fig. 6 we study the binding of ISWI to DNA (50 -fluorescein,

Fig. 6. ISWI titration in the presence of added nucleotide. Titration of a 1 nM 50 -

[2] Biophysical Analysis of Specific Genomic Loci

Copyright 2004, Elsevier Inc.

we describe how AME can be used as a biophysical method for character-

Agarose Multigel Electrophoresis

Extending AME to In Vivo–Assembled Chromatin

Gel (%) Buffer (ml) Agarose stock (ml)

E buffer 0.755 1.02

Assay for Secondary Chromatin Structure Formation

Gel (%) Pe range (nm)

Composition Analysis of the Genomic Fragments

rather than decondensation of the promoter chromatin into an unfolded

[3] Visualization and 3D Structure Determination

Recent progress in preparing and purifying defined sequence chromatin

Copyright 2004, Elsevier Inc.

rather than decondensation of the promoter chromatin into an unfolded

[3] Visualization and 3D Structure Determination

Recent progress in preparing and purifying defined sequence chromatin

Copyright 2004, Elsevier Inc.

Method Comments Resources needed

Positive stain Restricted to 2D Transmission EM

‘‘minichromosome’’ is readily distinguished from contaminating linear

Adhesion to Support Films

Shadow Casting after Drying from Glycerol (Fig. 1A)

3. A 2 2 cm ‘‘target’’ of freshly cleaved mica is placed about 25 cm

Negative Staining (Fig. 1B)