Letter

pubs.acs.org/NanoLett

Direct Imaging of DNA Fibers: The Visage of Double Helix

Francesco Gentile,, Manola Moretti, Tania Limongi,, Andrea Falqui, Giovanni Bertoni,,
Alice Scarpellini, Stefania Santoriello, Luca Maragliano, Remo Proietti Zaccaria,
and Enzo di Fabrizio,,*

Nanostructures, Neuroscience and Brain Technologies, and Nanochemistry Departments, Istituto Italiano di Tecnologia, Via
Morego 30, 16163 Genova, Italy

BIONEM, Bio-Nanotechnology and Engineering for Medicine, Department of experimental and clinical medicine, University of
Magna Graecia Viale Europa, Germaneto, 88100 Catanzaro, Italy

IMEM-CNR, Parco Area delle Scienze 37/A, 43124 Parma, Italy

*
S Supporting Information

ABSTRACT: Direct imaging becomes important when the

knowledge at few/single molecule level is requested and where
the diraction does not allow to get structural and functional
information. Here we report on the direct imaging of double
stranded (ds) -DNA in the A conformation, obtained by
combining a novel sample preparation method based on super
hydrophobic DNA molecules self-aggregation process with
transmission electron microscopy (TEM). The experimental
breakthrough is the production of robust and highly ordered
paired DNA nanobers that allowed its direct TEM imaging
and the double helix structure revealing.

KEYWORDS: DNA bundles direct imaging, transmission electron microscopy, superhydrophobic surface, nanofabricated micropillars,
molecular dynamics simulations

T he determination of the structure of DNA represented a

important historical event and revealed its fundamental
role in biology and life science.
means that we need tools to unveil the interaction, the
structure, the state (genetic, epigenetic conditions) and the
function of the coding and noncoding content of a DNA
However, the understanding of DNA was mainly related to molecule.4,5 In this work, we aimed our eort at direct DNA
its genetic content that, we now know, represents only a small imaging methods.
percentage, about 3%, of the whole information content.1 Transmission electron microscopy is a technique that allows
The role of the noncoding content, about 97% of our overall imaging with intrinsic spatial resolution at atomic scale. When
genetic material, remained elusive up to recent time, where working with few biomolecules, there are some factors that
important discoveries were made on the relationship between worsen the nal resolution of the obtained images: (i) the poor
noncoding DNA, micro RNA, and the so-called nonfunc- phase (or absorption) contrast of atomic species constituting
tional or evolutionary relics genes.2 the molecule compared to that of the substrate, where the
New direct methods are now necessary to understand the
molecule is sitting; (ii) the tendency of the molecule to be
complex relationships between DNA, proteins, micro RNAs
rapidly damaged when investigated by a high energy electron
and transcription factors.3
Traditionally, the use of diraction in determining the beam.68 However, the situation can dramatically be improved
structure of macro molecules is related to sample preparation if the molecule is suspended and background free. Under these
optimization and to obtaining high quality crystals or bers. conditions, the background noise is removed and the spatial
Unfortunately, only in a minority of cases the crystal can be resolution is primarily dictated by the intrinsic scattering and
obtained. An appealing alternative would be to have methods absorption cross section properties of the molecule itself, the
and tools allowing direct imaging of the molecule. In this case, homogeneity and regularity of the molecular aggregate and by
the need of having an organized structure to be solved by the detector features. The backbone of DNA, formed by
diraction methods would not be required. Nevertheless, there
is a deeper reason for trying to develop direct imaging methods: Received: October 24, 2012
functional information is strictly related to the knowledge of Revised: November 20, 2012
specic epigenetic signatures at level of single molecule. This Published: November 22, 2012

2012 American Chemical Society 6453 dx.doi.org/10.1021/nl3039162 | Nano Lett. 2012, 12, 64536458
Nano Letters Letter

Figure 1. Super hydrophobic DNA molecules self-aggregation and SEM and TEM imaging. SEM images (a, b, c), sketch (d), and TEM image (e) of
-DNA bers suspended on super hydrophobic pillared devices. In detail, an overview of DNA bundles hanging on top of pillars (a) and its zoom-in
tilted view (b) and top view (c) showing the DNA bundle located exactly on top of the passing through substrate hole. The sketch (d) elucidates the
concept exploited for TEM direct imaging: the passing through etched hole allows the suspended DNA to be orthogonally crossed by the
microscope electron beam; condition sine qua non: the hole needs to be in line with the pillars in order to obtain a DNA bundle alignment permitting
the TEM imaging. The suspended and aligned DNA bundle is directly imaged through the hole by TEM (e). Scale bar in panel a = 10 m; in panels
b, c = 1 m; in panel e = 500 nm.

Figure 2. TEM image with intensity prole and corresponding FFT pitch calculation of -DNA bers. (a) DNA ber TEM image. The inset shows
higher magnication DNA ber details; the red arrows point out the 2.7 nm pitch of A double helix. The scale bar corresponds to a length of 20 nm.
In panel b, a white rectangle is superimposed, showing where the intensity prole was measured. The peaks in plot c correspond to the alternation of
bright and dark bands in the original image (b): plot c displays a two-dimensional graph where the Y-axis reports the pixel intensity integrated along
the height of the rectangle and the X-axis represents the distance measured on the rectangle. Plot d shows the FFT of the signal displayed in plot c: a
well-dened maximum is observed at 0.37 0.02 1/nm, corresponding to a frequency of 2.7 0.2 nm.

phosphates and sugars may have enough TEM contrast even modifying the DNA deposition conditions as well as the device
when few aligned DNA molecules have to be imaged. structure, thin and stable suspended DNA bers could be
In previous experiments, we were able to control the investigated by the TEM high energy electron beam. The
deposition of few DNA molecules on a silicon micropatterned details on the design and fabrication of the device are reported
device designed to mimic super hydrophobic (SH) surface.9 in Methods. In summary, by an over layer multistep lithography
Besides, as a consequence of the deposition method, suspended and reactive ion etching process, we created several passing
DNA bers were observed. In the present experiment, through holes between the pillars constituting the super
6454 dx.doi.org/10.1021/nl3039162 | Nano Lett. 2012, 12, 64536458
Nano Letters Letter

Figure 3. A-DNAs superstructures used for image simulations. (a) The smallest bundle consisting of 1 + 6 A-DNAs (a central one and a shell of 6),
the bundle was turned slightly out of the low-index zone axis to reduce the contrast in the image due to the coherence created by the superstructure
periodicity. The incident parallel electron beam is sketched by the yellow arrows. (b) The 1 + 6 bundle viewed along the y direction and (c) the
corresponding simulated TEM image at about 1 m defocus with aligned chains.

hydrophobic surface (Figure SI 13, Supporting Information), details with fundamental metrological precise determination.
whose function is to allow the free passage of the electron beam Two diverse and independent simulations approach allowed to
during TEM measurements (background free imaging). The infer that a bundle is formed by paired DNA. This means that
preparation in view of the TEM imaging was obtained by ds DNA laments are aligned with their period along the z axis
allowing the water evaporation from the original liquid sample, of the helix.
at room temperature and 50% relative humidity. The super In the rst simulation approach we obtained the TEM image
hydrophobic surface was treated in a way that the adhesion of an isolated DNA bundle (background free) by imposing that
force between it and the water was very low (friction coecient the smaller bundle whose diameter is equal to 8 nm, is formed
about 0.02) and, during the evaporation, the pinning of the by 1 + 6 ds -DNA as shown in Figure 3a (see Methods).
drop was avoided. Under this condition, after the evaporation, This bundle structure is in agreement with experimental
DNA molecules could be deposited suspended and well tense TEM images (Figure 2), both in terms of diameter and
between the pillars, and more importantly several DNA bundles periodicity. In fact, the 6 + 1 simulated structure has a bundle
resulted suspended in correspondence of the holes. The diameter very close to that really measured and shown above,
suspended DNA bundles have well reproducible diameter, and the phase contrast TEM image (Figure 3c), calculated at
between 8 and 200 nm at our salinity conditions (see the experimental conditions, well reproduces the DNA helix
Methods). In Figure 1a is reported a scanning electron pitch as observed (red arrows in the inset of Figure 2),
microscopy (SEM) overview of this superhydrophobic pillared measured and reported in Figure 2bd. Notice that also the
device, in panel b and c of Figure 1 are reported a suspended experimental and theoretical intensity prole corresponds very
DNA bundle well aligned with the substrate hole, and in Figure well: peaks and minima are shifted by half DNA pitch in the
1e is seen its low magnication TEM image. In this last picture, two sides of the bundle. In the Methods section, we report the
the suspended DNA, in substrate free conguration, can easily
TEM image simulation of more complex bundle structure
be observed.
(Figure SI 4b, Supporting Information), giving details on the
As shown in the sketch of Figure 1, the suspended DNA
relevant parameters used. It is worth notice that if we introduce
corresponding to the hole was brought under the electron
a random misalignment along z axis, larger than 2 base pairs,
beam for obtaining direct TEM images. We observed several
DNA bundles with dierent diameter, and in Figure 2 one of the periodicity is lost and a blurred image appears in the
the smallest bers is shown, whose diameter imaged by 100 simulation (Figure SI 56, Supporting Information).
keV electron beam is about 8 nm, the periodicity is clearly Figure 4 reports a sketch of the evaporation mechanism on
observed. In the inset of Figure 2a, a magnied portion of the the super hydrophobic surface, where the shear/pulling
bundle is shown. About 10 periods of DNA can be seen in generated by the radial convection ux and by the drop
details. We notice that the period measures 2.7 0.2 nm, receding during the evaporation, tend to stretch ds -DNA
corresponding to that known for ds -DNA in A conformation. laments between pillars, as long as the capillary forces, in the
In panels c of Figure 2, the ber length measurements of the perpendicular direction, push them to aggregate in bundle
white selection in b is reported. In panel d, the fast Fourier during the latest drying phase.10
transform (FFT) of the metrological plot is shown, conrming In order to further support the model of paired DNA
that the dominant spatial frequency is 2.7 nm. structure, we performed Molecular Dynamics (MD) simu-
We underline that by this sample preparation the inuence of lations (see Methods) where we accounted for water
the substrate was completely removed and, in all meaning, the evaporation during the bundle aggregation. We started with
TEM measurement can be then considered substrate free. As a ds -DNA laments at a relative distance of 1 nm and a water
signicant consequence, it allowed to get DNA bundles with volume of 100 100 100 A3. The ds -DNA molecules were
strong mechanical stability under electron beam up to 100 keV left free to move along z direction but constrained by a
accelerating energy and beam current in the pA range. Besides, harmonic oscillator potential in the orthogonal direction. This
due to their well ordered ber structure, the structural last choice was suggested by the aggregation mechanism, where
information of single ds -DNA there contained was clearly capillary forces are responsible for pushing together the ds -
imaged. All these features permitted to obtain DNA structural DNA laments during the last phase of the evaporation.
6455 dx.doi.org/10.1021/nl3039162 | Nano Lett. 2012, 12, 64536458
Nano Letters Letter

these results well agree with that observed by TEM imaging.

The whole picture that comes from these results is that both
the shear/pulling and the capillary forces, acting as a
consequence of the present sample preparation method, are
compatible with the interhelix forces responsible for the DNA
pairing. In other words, under the presented superhydrophobic
evaporation conditions, the aggregation of ds -DNA molecules
leads to an ordered ber structure along z axis direction. To
further support this view, in Figure 5c, we report the calculation
of the aggregation free energy as a function of the double helix
misalignment in z direction. The minimum energy is reached
when the alignment condition between the ds -DNA laments
is fullled with a value, about 12kBT, that is big enough for a
stable conguration during the latest phase of DNA dewetting.
We point out at this stage that, incidentally and ironically,
our sample preparation route has a strong resemblance with the
method used by Wilkins et al.11 in the historical experimental
tour de force that brought to DNA structure determination. The
pulling and drying ber preparation by Wilkins et al. was similar
to the present method, where the pulling function is played by
the convection, and drop receding shear force as well as the
drying is played by the evaporation assisted by super
hydrophobic surface. The two sample preparation methods
Figure 4. Evaporation and aggregation mechanism. The most practical
property of these super hydrophobic surfaces is a reduced friction
lead to similar results, even if the aggregation length scale
coecient (about 0.02 for the present conguration) on account of dominating in our experiment is 3 orders of magnitude smaller
which they can be conveniently used to deposit, suspend and stretch than that in Franklin experiment:12 the bers obtained by
double stranded -DNA molecules between two or more pillars. Franklin et al. had a typical diameter in the range of tens of
During the evaporation, the pinning of the drop is avoided, after the micrometers, instead in our case the ber diameter falls in the
evaporation, DNA molecules are suspended and well tense between range of tens of nanometers. We demonstrated a new
the pillars and, more importantly, many DNA bundles are in preparation method that, based on the use of nanofabricated
correspondence of the holes. In the aggregation mechanism, capillary device, through the evaporation assisted by super hydrophobic
forces are responsible for pushing together the ds DNA laments self-molecular aggregation, allows the DNA bers TEM
during the last phase of the evaporation. In particular, shear forces
imaging in background free conditions. For these main reasons,
generated by a radial convection ux tend to stretch ds DNA laments
between the pillars, while capillary forces, perpendicular to those shear after almost 60 years since the rst X-ray diraction images,13,14
forces, pack the laments into bundles during the latest drying phase. we were able to obtain, for the rst time, a clear direct and
completely background-free image of DNA double stranded in
A conformation by a simple and fast preparation method.
In Figure 5a, we report the MD aggregation mechanism and Future developments based on our achievement could be of
its evolution as a function of the starting relative distance great interest for genetic and epigenetic analysis, for the
between the DNA laments and a decreasing number of water understanding on how DNA damage and its repair aect the
molecules surrounding the bundle. In Figure 5b, the (epi-) genome over a lifetime, and how these changes impact
equilibrium separation distance of the double helix laments on age-related pathology and malignant transformation.
is shown, as obtained from the starting condition reported in Furthermore, there is room for strong further improvements,
Figure 5a, and reached after about 10 ns. The equilibrium both in terms of comprehension of the whole mechanism
distance, center to center, is equal to 2.5 nm and the external presiding to the DNA bundle formation (a possible super-
backbone distance is 0.3 nm. It is important to remark that also coiling organization cannot be completely excluded), of sample

Figure 5. Molecular dynamics simulations of DNA double helices aggregation. (a) Conformations of the paired DNA molecules at the beginning of
the simulation and after 5 ns at dierent hydration conditions; the numbers () refer to the thickness of the hydration layer. (b) Time-evolution of
the distance between the centers of mass of the dodecamers along the simulations at dierent degrees of hydration. (c) Interaction energy (potential
of mean force) between the DNA molecules in the 0.3 nm solvation layer as a function of their relative vertical displacement. High and low energy
values are highlighted with red circles, and the corresponding conformations are shown.

6456 dx.doi.org/10.1021/nl3039162 | Nano Lett. 2012, 12, 64536458

Nano Letters Letter

preparation and TEM analytical perspectives. In the latter case, pulling up or sliding a at surface upon the DNA solution
a new generation of detectors, with increased sensitivity and droplet posted on the substrate. Shear ow16 and exclusion of
higher contrast, will permit low dose imaging so that even water molecules17 at the receding meniscus perpendicular to
single DNA helix would resist to high-energy electron beam the droplet evaporation direction are likely contributing to the
damage, allowing high quality image detection at single stretching and alignment of DNA among the pillars. The DNA
nucleotide level. bundles were nally imaged by TEM with a JEOL JEM-1011
Methods. Design and Fabrication of the Devices. Double (spherical aberration 5.6 mm), equipped with a W thermionic
polished, (100), 50 m, p-type thin silicon wafers were gun. It operated at an acceleration voltage of 100 kV, and
purchased from Si-Mat (Silicon Materials, Kaufering, Ger- electrons density of 500 e/(2s). The TEM images were
many). They were cleaned with acetone and isopropanol to collected by a Gatan SC-1000 Orius Camera, equipped with a
remove possible contaminant and then etched with a 4% wet ber-optical coupled 11 Mp CCD, and using an acquisition
HF solution. The wafers were then rinsed with DI water and time of 2 s.
dried with N2. A 100 nm layer of chrome was deposited upon TEM Image Simulations. The A-DNA superstructures were
the back side of the substrates using a sputter coater (Q150TES built using Discovery Studio v3.1 (Accelrys Software Inc.). To
Quorum Technologies, Dixon, CA). Standard optical lithog- avoid boundary eects, approximately 70 bases long nucleic
raphy techniques (Suss Microtec MA6/BA6, Sunnyvale, CA) acid chains were built. The DNAs were placed at a 2.5 nm
were employed to realize a regular pattern of disks within a center to center distances, aligned along their axes in a
layer of positive resist (SPR220), that was spin-coated onto the hexagonal pattern, with ideally the same orientation. The TEM
chrome layer. Therefore, the chrome was removed from the image simulations were performed by using the xHREM v3.5
disks by exposing them for 50 s to a standard chrome etching package (HREM Research Inc.), according to the multislice
solution (ETCH 18 from OSC-Organo Spezial Chemie, method,18 in which the bundle was divided in 10 phase grating
Bitterfeld, Germany). Upon removal of the residual resist planes of atoms, at which the incident front wave of electrons is
with acetone and oxygen plasma, a Deep Reactive Ion Etching dynamically scattered and propagated. The image is calculated
(DRIE) process (SI 500 Sentech Instruments Gmbh, Berlin, taking into account the additional phase and amplitude eects
Germany) was used to etch holes passing through the substrate, due to the instrument (defocus and aberrations). The choice of
the patterned chrome layer serving as a mask. The samples a defocus (about 1 m) of the images acquired at the
were then immersed in a bath of chrome etching solution to transmission microscopes boosts the low frequencies in the
dissolve the remaining chrome. The substrate was then turned images, permitting to resolve the period of A-DNA (around 2.8
upside down, and the fabrication process was pursued on the nm), while cutting the high frequencies, so losing the
front side of the samples. A second lithography exposure step information on the base positions, but with the advantage of
was used to realize a regular hexagonal pattern of disks within a cutting noise, which will otherwise aect the in-focus images at
layer of negative resist (AZ5214 from Microchemicals GmbH, such low dose and low contrast due to the low atomic
Ulm, Germany. The disks were positioned exactly on the potentials of the constituent atoms. Further details on TEM
middle point of each subset of 6 holes. A DRIE process was image simulations are reported in Supporting Information.
therefore used, whereby the nal pillars were obtained with a Molecular Dynamics (MD) Simulations. MD simulations
height h of about 12 m, a diameter d of 10 m and a pitch of were performed using the program NAMD19 and the
30 m. Further details on SH Device fabrication are reported in CHARMM2720 force eld for DNA molecules. Two laments
Supporting Information. of sequence d(GCGAAATTTGCG)2 in A conformation were
DNA Sample Preparation, Spotting, and Electron Micros- immersed in a box of 12 331 equilibrated TIP3P21 water
copy Characterization. Linearized double stranded DNA molecules together with 76 Na+ and 32 Cl atoms, which
extracted from Lambda phage (New England Biolabs Inc., amounts to a 0.137 M solution plus excess sodium ions to
Ipswich, MA) was diluted in PBS 1 (containing 0.137 M neutralize the system. The dodecamers were aligned along the
NaCl) to a nal concentration of 50 ng/L. A 20 L droplet z-axis, placed at a relative center-to-center distance of about 2.8
was post upon the superhydrophobic substrate and let nm and zero relative vertical displacement. The full system was
dehydrate overnight at 24 C and 50% humidity in a Petri simulated with DNA molecules xed for 10 ns using periodic
dish. Samples were then checked without further preparation boundary conditions (PBC), constant temperature (298 K) and
by SEM, using a JEOL JSM-7500FA microscope equipped with constant pressure (1 atm). Electrostatic interactions were
a cold eld emission gun and working at an acceleration voltage computed with the particle-mesh Ewald method.22 Temper-
of 5 kV. The SEM imaging was carried out using the secondary ature and pressure were kept constant using Langevin dynamics
electron signal. Double stranded DNA could be suspended and the Langevin NoseHoover method23 as implemented in
between two pillars during dewetting of a DNA containing NAMD. After the cloud of ions equilibrated around the xed
droplet on a SH surface as recently shown in ref 9, where SH DNA molecules, we built dierent systems at dierent degrees
surfaces have been used to concentrate solutes and to precisely of hydration, by isolating from the full system the dodecamers,
deliver few molecules on detection spots. The dewetting the water molecules, and the sodium atoms included in shells of
droplet is depinning from the pillars posts until a pinning 1, 0.5, and 0.3 nm from the DNAs. The resulting systems were
condition due to nanostructure of substrate/drop contact simulated for 5 ns without PBC and using a 10 nm cuto for
angle,9 high solute, chemical heterogeneicity or discontinuities electrostatic interactions. The orientation of the dodecamers
on the surface are reached.15 DNA combing is a well-studied along z was kept xed by a restraint potential. The free energy
method for stretching and aligning nucleotide molecules on at for the vertical displacement of the dodecamers was computed
surfaces or on nanofabricated surface, relying on a at receding for the system with solvation shell 0.3 nm by integrating the
meniscus of solution to stretch the DNA. It could be obtained mean forces calculated along 14 dierent restrained simu-
by pulling out the substrate from the DNA solution (similar to lations, each with the laments held at dierent relative
LangmuirBlodgett preparation) with controlled velocity or by displacement by harmonic potentials.24 The DNA molecules
6457 dx.doi.org/10.1021/nl3039162 | Nano Lett. 2012, 12, 64536458
Nano Letters Letter

were maintained in the A conformation using a restraint on (10) Smith, S. B.; Finzi, L.; Bustamante, C. Science 1992, 258, 1122.
their root-mean-square deviation. Each restrained simulation (11) Wilkins, M. H. F.; Gosling, R. G.; Seeds, W. E. Nature 1951,
lasted 1 ns to obtain convergence of the mean force estimator. 167, 759.

(12) Franklin, R. E.; Gosling, R. G. Acta Crystallogr. 1953, 6, 673.

ASSOCIATED CONTENT (13) Wilkins, M. H. F.; Stokes, A. R.; Wilson, H. R. Nature 1953, 171,
738.
*
S Supporting Information
(14) Watson, J. D.; Crick, F. H. Nature 1953, 171, 737.
Additional information on design and fabrication of SH devices (15) Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagel, S.
and TEM image simulations. This material is available free of R.; Witten, T. A. Phys. Rev. E 2000, 62, 756.
charge via the Internet at http://pubs.acs.org. (16) Accardo, A.; Gentile, F.; Mecarini, F.; De Angelis, F.;

AUTHOR INFORMATION
Corresponding Author
Burghammer, M.; Di Fabrizio, E.; Riekel, C. Langmuir 2010, 26,
15057.
(17) Deegan, R. D.; Bakajin, O.; Dupont, T. F.; Huber, G.; Nagl, S.
R.; Witten, T. A. Nature 1997, 389, 827.
*E-mail: [email protected]. (18) Kirkland, E. J. Advanced Computing in Electron Microscopy;
Author Contributions Plenum, New York, 1998.
E.D.F., F.G., and M.M. conceived and designed the experi- (19) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid,
ments, F.G. and S.S. realized the superhydrophobic devices, E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K. J. Comput.
M.M., A.F., and A.S. performed the experiments and the EM Chem. 2005, 26, 1781.
imaging. G.B. and L.M. performed the transmission electron (20) Mackerell, A. D.; Feig, M.; Brooks, C. L. J. Comput. Chem. 2004,
microscopy image simulation and the molecular dynamics 25, 1400.
simulations, respectively. E.D.F, A.F., T.L., and R.P.Z. analyzed (21) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R.
W.; Klein, M. L. J. Chem. Phys. 1983, 79, 926.
the data and participated in the writing of the paper. All authors (22) Darden, T.; York, D.; Pedersen, L. J. Chem. Phys. 1993, 98,
discussed the results and commented on the manuscript. 10089.
Notes (23) Martyna, G. J.; Tobias, D. J.; Klein, M. L. J. Chem. Phys. 1994,
The authors declare no competing nancial interest. 101, 4177.

ACKNOWLEDGMENTS
We are grateful to Dr. Roberto Marotta for the several, deep
(24) Maragliano, L.; Fischer, A.; Vanden-Eijnden, E.; Ciccotti, G. J.
Chem. Phys. 2006, 125, 024106.

and exciting scientic discussions about the DNA. We also

thank the IIT Platform "Computation" for CPU time. This
work was funded under European Project SMD FP7-NMP
2800-SMALL-2 proposal no. CP-FP 229375-2 and under the
Single Molecule Activation and Computing (FOCUS) project
(Grant agreement no: 270483) funded under seventh Frame-
work Programme.

REFERENCES
(1) Salmena, L.; Poliseno, L.; Tay, Y.; Kats, L.; Pandolfi, P. P. Cell
2011, 146, 353.
(2) Biemont, C.; Vieira, C. Nature 2006, 443, 521.
(3) ENCODE Project Consortium Bernstein, B. E.; Birney, E.;
Dunham, I.; Green, E. D.; Gunter, C.; Snyder, M. Nature 2012, 489,
57.
(4) Lister, R. M.; Pelizzola, M.; Dowen, R. H.; Hawkins, R. D.; Hon,
G.; Tonti-Filippini, J. S.; Nery, J. R.; Lee, L.; Ye, Z.; Ngo, Q. M.; Edsall,
L.; Antosiewicz-Bourget, J.; Stewart, R.; Ruotti, V.; Millar, A. H.;
Thomson, J. A.; Ren, B.; Ecker, J. R. Nature 2009, 462, 315.
(5) Harris, R. A.; Wang, T.; Coarfa, C.; Nagarajan, R. P; Hong, C.;
Downey, S. L.; Johnson, B. E.; Fouse, S. D.; Delaney, A.; Zhao, Y.;
Olshen, A.; Ballinger, T.; Zhou, X.; Forsberg, K. J.; Gu, J.; Echipare, L.;
OGeen, H.; Lister, R.; Pelizzola, M.; Xi, Y.; Epstein, C. B.; Bernstein,
B. E.; Hawkins, R. D.; Ren, B.; Chung, W. Y.; Gu, H.; Bock, C.;
Gnirke, A.; Zhang, M. Q.; Haussler, D.; Ecker, J. R.; Li, W.; Farnham,
P. J.; Waterland, R. A.; Meissner, A.; Marr,a, M. A.; Hirst, M.;
Milosavljevic, A.; Costello, J. F. Nat. Biotechnol. 2010, 28, 1097.
(6) Bleloch, A.; Own, C.; Hamalainen, M.; Hershleb, J.; Kemmish,
K.; Koene, R.; Stark, H.; Stark, J.; Andregg, M.; Andregg, W. Microsc.
Microanal. 2011, 17, 1274.
(7) Cerf, A.; Alava, T.; Barton, R. A.; Craighead, H. G. Nano Lett.
2011, 11, 4232.
(8) Bell, D.; Thomas, W.; Murtagh, K.; Glover, W. Microsc.
Microanal. 2011, 17, 1276.
(9) De Angelis, F.; Gentile, F.; Mecarini, F.; Das, G.; Moretti, M.;
Candeloro, P.; Coluccio, M. L.; Cojoc, G.; Accardo, A.; Liberale, C.;
Zaccaria, R. P.; Perozziello, G.; Tirinato, L.; Toma, A.; Cuda, G.;
Cingolani, R.; Di Fabrizio, E. Nat. Photonics 2011, 5, 683.

6458 dx.doi.org/10.1021/nl3039162 | Nano Lett. 2012, 12, 64536458