JOBNAME: PROSCI 16#4 2007 PAGE: 1 OUTPUT: Wednesday March 7 19:04:08 2007

csh/PROSCI/131728/ps0626997

Efficient optimization of crystallization conditions by

manipulation of drop volume ratio and temperature

JOSEPH R. LUFT,1,2 JENNIFER R. WOLFLEY,1 MERIEM I. SAID,1 RAYMOND M. NAGEL,1

ANGELA M. LAURICELLA,1 JENNIFER L. SMITH,1 MAX H. THAYER,1
CHRISTINA K. VEATCH,1 EDWARD H. SNELL,1,2 MICHAEL G. MALKOWSKI,1,2
1,2
AND GEORGE T. DETITTA
1
Hauptman-Woodward Medical Research Institute, Buffalo, New York 14203, USA
2
Department of Structural Biology, SUNY at Buffalo, Buffalo, New York 14203, USA
(R ECEIVED December 1, 2006; F INAL R EVISION December 15, 2006; ACCEPTED December 20, 2006)

Abstract
An efficient optimization method for the crystallization of biological macromolecules has been developed
and tested. This builds on a successful high-throughput technique for the determination of initial
crystallization conditions. The optimization method takes an initial condition identified through screening
and then varies the concentration of the macromolecule, precipitant, and the growth temperature in a
systematic manner. The amount of sample and number of steps is minimized and no biochemical
reformulation is required. In the current application a robotic liquid handling system enables high-
throughput use, but the technique can easily be adapted in a nonautomated setting. This method has been
applied successfully for the rapid optimization of crystallization conditions in nine representative cases.
Keywords: protein crystallization; high throughput; optimization; temperature

The importance of obtaining a crystal for X-ray crystallo- Jancarik and Kim (1991) used a sparse matrix approach
graphic studies is difficult to overstate; it is required to reach to design a set of solutions (cocktails) to screen and
the goal of a three-dimensional structure of the target mole- identify crystallization conditions; these solutions are
cule. Producing crystals of biological macromolecules is often readily available in commercial kits. An expansive
a challenging task. For 119,721 targets entered into TargetDB collection of chemicals known to crystallize macromole-
by the worldwide structural genomics centers, only 14.1% of cules appears in the literature. It is not practical or
the purified, soluble targets produced a crystal structure feasible to exhaustively screen all of the combinations
(Berman et al. 2000). The crystallization of biological macro- of these chemicals. These cocktails incorporate a range
molecules can be considered as a two-stage process. The first of pH, chemical species, and concentrations. A sparse
stage, ‘‘screening,’’ determines chemical and physical con- matrix design is used to formulate a reasonable number of
ditions under which the sample has a propensity to crystallize. cocktails to survey the vast chemical landscape. When
The second stage, ‘‘optimization,’’ refines the chemical and even a single experiment produces a crystalline outcome,
physical parameters to produce crystals suitable for analysis the screening experiments are considered successful.
by X-ray diffraction. Optimization methods have not been Optimization makes use of information derived from the
brought to high-throughput contemporaneously with screen- screening experiments to produce crystals of sufficient size
ing (Chayen and Saridakis 2002). and quality for diffraction. Both environmental variables
(such as temperature) and chemical variables (type and
concentration of chemicals and the solution pH) are refined.
Reprint requests to: Joseph R. Luft, Hauptman-Woodward Medical It is advantageous when multiple experiments produce
Research Institute, 700 Ellicott Street Buffalo, NY 14203, USA; outcomes suitable for optimization from different cocktails.
e-mail: [email protected]; fax: (716) 898-8660.
Article published online ahead of print. Article and publication date The crystals produced from these chemically distinct
are at http://www.proteinscience.org/cgi/doi/10.1110/ps.062699707. solutions often have different physical properties, which

Protein Science (2007), 16:715–722. Published by Cold Spring Harbor Laboratory Press. Copyright Ó 2007 The Protein Society 715

ps0626997 Luft et al. ARTICLE RA

JOBNAME: PROSCI 16#4 2007 PAGE: 2 OUTPUT: Wednesday March 7 19:04:09 2007
csh/PROSCI/131728/ps0626997

Luft et al.

can be exploited to provide alternatives if difficulties are study. Crystallization outcomes are presented for eight
encountered downstream during optimization, X-ray dif- different proteins that have been optimized for crystal
fraction, or structure solution and refinement. volume using a high-throughput Drop Volume Ratio/
A widely used optimization strategy, the grid screen, Temperature (DVR/T) method. This method can be setup
arrays chemical variables identified during screening rapidly with a small volume of protein solution. There is no
experiments to refine the chemical conditions and produce need to reformulate protein or cocktail solutions from those
crystals of increased volume (Cox and Weber 1988). The used for the screening experiments. DVR/T experiments
concentration of a precipitating agent (the chemical species sample temperature simultaneously with the concentrations
that is primarily responsible for driving the macromolecule of the protein and cocktail solutions.
to a state of supersaturation) and solution pH are varied in a
regular fashion from an initial coarse screen to finer grids in
Results
a series of successive experiments. Another experimental
design, first reported to crystallize Bacillus stearothermo- Eight biological macromolecules submitted to the Center
philus tryptophan tRNA synthetase, refines crystallization for High-Throughput Structural Biology were used for this
solutions using an incomplete factorial to distribute study. These represented typical problems and ranged from
crystallization variables in a randomized and balanced 25- to 75-kDa molecular weight (Table 1; Fig. 1A–I).
manner (Carter and Carter 1979). Using stepwise multiple- Screening experiments for these samples at 23°C produced
regression analysis, scores of crystal quality are correlated outcomes ranging in quality from needles and highly
to design the next series of experiments. twinned plates to small single crystals. The effects of
Both the grid and incomplete factorial approaches changing the experiment drop composition (volume of
produce a common product. They describe the next gener- protein to crystallization cocktail) and temperature on the
ation of cocktails that will be used to optimize crystal- outcomes of these experiments are readily observed (Fig. 1).
lization conditions. Regardless of the method used to design A simultaneous, microscopic assessment of the outcomes
these solutions, time and materials must be expended to shows the incubation temperature affected all of these
prepare them. The majority of the time required to set up samples. Examples where solubility is directly (Fig. 1A)
optimization experiments is devoted to formulating and and inversely related to temperature (Fig. 1D) are observed
then producing cocktail solutions. A significant time-savings by comparing the number of clear (undersaturated) out-
and potential increase in reproducibility would be realized comes to the number of supersaturated outcomes (phase
if solutions which produced crystals in the screening separation, precipitate or crystals) for each group of 64
experiments could be directly employed for optimization experiments. The experiment drop chemistry plays an
without the need to reformulate. important role in the relationship between a protein’s
The use of temperature to control the level of super- solubility and temperature. Interestingly, the solubility
saturation in crystallization experiments is well-established dependence on temperature can be reversed by changing
for small molecules, but remains an underutilized variable the chemistry of the cocktail solution for the same protein,
for the crystallization of macromolecules (Rosenberger 1986; (Fig. 1A,B). Sample P6306 (Fig. 1A) is more soluble with an
Landsberg et al. 2006). Despite its limited use, temperature increase in temperature using a cocktail solution of 100 mM
has been shown to be an important and generally applicable Na Acetate pH 5.0, 100 mM NH4SCN, 20% (w/v) PEG
variable for the crystallization of biological macromolecules.
Batch crystallization methods, including microbatch-
under-oil (Chayen et al. 1992), combine an aliquot of Table 1. Protein solutions used for optimization
protein and cocktail solution for crystallization. The two [Protein]
solutions merge to form an experiment drop. Microbatch- Figure 1 Protein mg/mL Protein solution
under-oil differs from traditional batch methods because it A P6306 20 10 mM Tris pH 7.4, 100 mM NaCl
uses oil to containerize and to retard the rate of dehydration B P6306 20 10 mM Tris pH 7.4, 100 mM NaCl
of the aqueous experiment drop. Batch methods are in C P5687 4 20 mM HEPES pH 7.6, 100 mM
general well-characterized (Koszelak et al. 1996), can be NaCl,5% (v/v) glycerol,
1 mM DTT, 1 mM EDTA
used with small solution volumes, and can be incubated at a
D P6334 9 50 mM Tris pH 8.5, 500 mM NaCl
variety of temperatures (Chayen et al. 1990). Manipulating E P6127 35 20 mM Tris pH 9.0
the concentration of the protein and precipitating agent by F P6512 20 10 mM Tris pH 7.5, 0.02% (w/v)
varying their ratios to each other has been demonstrated to Na Azide
be an effective and efficient way to produce high-quality G P6510 10 50 mM Tris pH 7.6, 1 mM EDTA
H P6102 11 50 mM K Phosphate pH 7.4,
crystals using batch methods (Rayment 2002).
100 mM KCl
The microbatch-under-oil technique has been used for I P6893 10 20 mM Tris pH 7.6
both screening and optimization experiments during this

716 Protein Science, vol. 16

JOBNAME: PROSCI 16#4 2007 PAGE: 3 OUTPUT: Wednesday March 7 19:04:10 2007
csh/PROSCI/131728/ps0626997

Efficient optimization of crystallization

Figure 1. Outcomes of DVR/T experiments. All of the images have a well diameter (circle) of 0.9 mm. The highlighted thumbnail image is considered a
‘‘good outcome’’ and is enlarged to the right of the array. Each array of 64 images (4 rows 3 16 columns) is a single-protein and cocktail solution. The
numbers 1 through 16 correlate with the volume of protein and cocktail solution used to construct the experiment drop. A gradual progression of inversely
related volume ratios takes place in the series of images. Column 1 has the highest volume of protein and lowest volume of cocktail solution; column 16 has
the highest volume of cocktail and lowest volume of protein solution. The four different temperatures used to incubate the plates (4°C, 14°C, 23°C, and
37°C) appear along the left side of each array. All images were recorded three weeks after the experiment was set up. The protein and cocktail solutions (A–I)
are listed in Tables 1, 2, respectively.

4000. P6306 is less soluble with increasing temperature [protein] and increases the [cocktail] sufficiently to cause
(Fig. 1B) using a cocktail solution of 100 mM MOPS the dramatic shift in morphology. These trends in morphol-
pH 7.0, 100 mM NH4Br, 80% (v/v) PEG 400. In most cases ogy continue for both dendrites and plates in sequential
there is a readily identified, optimum temperature for crys- progression away from this boundary.
tallization. The optimum temperature for this group of Crystals that were judged to be of good optical quality
proteins (highlighted by the thumbnails of ‘‘good’’ outcomes) appeared across the range of drop volume ratios used for
varies and is distributed across all four temperatures. these experiments. An example where the best crystals form
Varying the volume ratio used to compose the experiment near one volume ratio extreme (Vprotein > Vcocktail) is P5687
drop can lead to a dramatic change in crystal morphology. (Fig. 1C). It should be noted that the starting concentration
An example of this effect, at a constant temperature of of P5687 was relatively low (4 mg/mL). This makes the
23°C, appears in Figure 1I. In this series of experiments, higher volume ratio of protein to cocktail producing the best
fibrous, dendrite crystals abruptly change to plate morphology. crystals somewhat intuitive. It demonstrates crystal quality
Dendrites form at a higher concentration of protein (here- can be improved by changing the volume ratio without the
after designated [protein]) and lower [cocktail] in the exper- need to reformulate (concentrate) the protein solution.
iment drop. The demarcation takes place when the cocktail The 1536 well microassay plates used to set up these
volume is held constant at 200 nL, while the protein volume experiments have narrow, conical wells (0.9 mm diameter
decreases from 250 nL (experiment number 7, 23°C) to 200 at the bottom of the well); crystal retrieval from these
nL (experiment number 8, 23°C). This decreases the plates is not easily achieved. As such, there is no X-ray

www.proteinscience.org 717
JOBNAME: PROSCI 16#4 2007 PAGE: 4 OUTPUT: Wednesday March 7 19:04:20 2007
csh/PROSCI/131728/ps0626997

Luft et al.

data to demonstrate the quality of diffraction. These initial conditions are well known; they are not dependent on
small-volume batch experiments are designed to quickly time or geometry. Figure 2 shows a simplified phase
identify optimized crystallization conditions. Scale-up diagram for crystallization. The initial crystallization con-
for X-ray diffraction analysis is the next challenge. dition is known to be at a metastable state (nucleation was
heterogeneous) or a labile state (nucleation was sponta-
neous and homogeneous). To optimize this condition the
Discussion
supersaturation has to be decreased such that nucleation
Every crystallization protocol will take a unique path still occurs while approaching the lower level of super-
through the phase diagram influencing the experiment’s saturation that is ideal for crystal growth. This promotes a
outcome (Ries-Kautt and Ducruix 1992). The DVR/T small number of crystal nuclei and prolongs the growth of
method uses exactly the same microbatch-under-oil crys- those into large crystals. The DVR/T experiments sample
tallization protocol for both screening and optimization. different volume ratios with a midpoint (equal volume of
Using the same protocol improves reproducibility and protein and cocktail solution) mimicking the initial crys-
eliminates the complication of converting crystallization tallization condition. Depending on the location of this
conditions from one method used for screening to another condition on the solubility diagram, the DVR/T path may
for optimization (Chayen 1998). The optimized condi- first sample a local maximum of supersaturation; however,
tions will need to be reproduced in a container that is it will also sample points where there is a decrease in the
better suited for retrieving the crystals; this may require supersaturation. The DVR/T technique should be centered
an increase in solution volumes. The method has another on the temperature of the initial crystallization hit (23°C
practical advantage; it makes use of the same cocktails for the experiments presented in this paper) to ensure that
for screening and optimization. This prevents batch the starting point falls in the labile or metastable zone.
differences caused by reformulation. Cocktails, espe- In a DVR/T experiment, a range of pH values will be
cially those containing PEGs, undergo chemical changes sampled. The protein and cocktail solutions typically con-
over time (Jurnak 1985). The high-throughput crystalli- tain different buffers. The [Buffer]cocktail > [Buffer]protein
zation screening experiment uses some variety of poly- so that the pH of the cocktail solution dominates the
ethylene glycol (PEG) in >1000 of the 1536 cocktail experiment drop. As the ratio of protein to cocktail
solutions. Aged PEG solutions have been reported to solutions changes, pH is effectively used as a variable
produce crystals of RNase A minor trimer that could not during the screening experiments. Fine screening of pH as
be obtained using freshly prepared solutions of PEG a single variable has been successfully used for optimi-
(Liu et al. 2002). This aging of the cocktails can cause zation (McPherson 1995). The range sampled in the
considerable difficulty when attempting to reproduce and DVR/T method will depend in part upon the type and
optimize results from the screening experiments. If the concentration of the buffers used in the cocktail and
optimization experiments are set up shortly after the protein solutions. It will also depend on the volume of
screening experiments, then using exactly the same
solutions for screening and optimization can eliminate
this aging effect. If the cocktails are not freshly prepared
prior to setting up the screening experiments, the DVR/T
approach may improve reproducibility by making use of
these same ‘‘aged’’ screening cocktails for optimization.
The DVR/T method has proven to be extremely effective
in optimizing initial hits. While simple to perform and
analyze, the optimization protocol is deceptively multi-
parametric but appropriately so for the task at hand. Carter
and Yin (1994) explored Harden-Sloane response-surface
designs to optimize crystallization conditions. They found it
was advantageous to sample the effects of protein concen-
Figure 2. Drop volume ratio effect on supersaturation. Crystals are
tration and supersaturation simultaneously. The DVR/T observed in the screening trials with a 1:1 ratio of protein to cocktail
technique implicitly samples supersaturation space. This solution (1) or (2). Holding all other variables constant, we assume this
is an effective way to detect the presence of optimal experiment falls someplace in the labile zone where spontaneous homoge-
conditions that lie well outside the range of supersaturation neous nucleation will occur. Varying the volume ratio of protein to cocktail
investigated in traditional optimization methods. solution (in a series of batch crystallization experiments) will sample
points that lie roughly along a path indicated by the arrows on the graph.
In the microbatch-under-oil method, the initial hit is a Different areas in the solubility diagram where the [protein] is higher and
logical point to start the optimization. Rather than follow- the [precipitating agent] is lower, and where the [precipitating agent] is
ing a complicated path to reach a nucleation point, the higher and the [protein] is lower will be sampled.

718 Protein Science, vol. 16

JOBNAME: PROSCI 16#4 2007 PAGE: 5 OUTPUT: Wednesday March 7 19:04:25 2007
csh/PROSCI/131728/ps0626997

Efficient optimization of crystallization

cocktail and protein solution in the experiment drop. behavior. Extracting information on the contribution of an
Using P5687 as an example, the cocktail buffer is individual parameter to the crystallization would require a
100 mM Citrate, pH 4.0 and the protein buffer is 20 mM more sophisticated optimization screen taking into account
HEPES, pH 7.6. The cocktail buffer will dominate the pH if the correlation between variables and the nonlinearity of
the solutions are combined in an equal volume ratio response to a change in any individual variable.
(50 mM Citrate, 10 mM HEPES). However, at the extreme Optimization takes place by a qualitative visual assess-
of the volume ratio (23% cocktail, 77% protein) the ment of the resulting crystals. X-ray analysis of the
protein’s buffer will significantly affect the final pH of resulting crystals is the only method that provides a
the experiment drop (23 mM Citrate, 15 mM HEPES). quantitative evaluation of the optimization. This is a logical
This sampling of pH will be a volumetric titration of the next step for the final round of optimization before X-ray
Buffercocktail and Bufferprotein. The pH of DVR/T experi- structural data collection takes place. However, as it stands
ments will also depend upon the incubation temperature, in the DVR/T method is directly applicable to neutron
particular the temperature coefficient of the buffers. This diffraction studies. Neutrons are uniquely sensitive to
effect can be significant when experiments are being protonation state or hydrogen positions, key to many
incubated in a range of temperatures from 4°C to 37°C. biological mechanisms. Typically, structures derived from
The experiments presented here show that temperature is X-ray data already exist and crystallization conditions are
a very significant variable. Studies have shown that temper- known. Neutron sources are far less intense than the
ature influences protein solubility, affecting both nucleation equivalent synchrotron X-ray source and neutrons are
and crystal growth (Giegé and Mikol 1989). It is an weakly scattered. Therefore, a key step to enabling suc-
important, and yet often ignored, variable for crystallization cessful neutron diffraction is maximizing the volume of
that will have some effect on solubility for the majority of those crystals (Snell et al. 2006). DVR/T naturally achieves
samples (Christopher et al. 1998). The effects of temper- this by providing a fine sampling around the X-ray
ature are not always obvious. Within a narrow 5°C temper- optimized conditions. As the crystal is already known to
ature range three different growth mechanisms were diffract, visual results on volume and volume trends from
reported for crystals of tRNAPHE (Ng et al. 1997). Calcu- DVR/T can be immediately used to enable neutron studies.
lations predict the existence of a highly concentrated layer Perhaps the most striking result of the method is that a
of metastable liquid–liquid protein phase separation on the small change can have a dramatic effect on the result.
surface of the growing crystal. This layer has an effect on Conditions that would otherwise not be considered as a
both the kinetics of nucleation and the growth of macro- starting point for optimization are strikingly close to the
molecular crystals and is critically dependent on temper- conditions required for producing visually perfect single
ature (Haas and Drenth 2000). Temperature induction (brief crystals. This offers a potential method to address the 86%
incubation of the crystallization experiment at an elevated of soluble proteins that do not form crystals. Using high-
temperature from subsequent growth) was required to throughput methods, the DVR/T technique could be used as a
crystallize GrpE from Thermus thermophilus (Kitano secondary hit screen. With a minimal amount of sample the
et al. 1998). Temperature change has also been used to top 24 initial hits could be run through the DVR/T method
reduce microheterogeneity (Leinala et al. 2002) and has instead of a single initial condition. The technique could also
been shown to cause conformational changes in the be used to explore an often overlooked problem in initial
structure even after crystals have formed (Dunlop et al. condition screening. A perfect crystallization condition results
2005). Samples which do not show a temperature-dependent in a large single crystal. Nucleation is a stochastic process
solubility in one condition may well be sensitive in others. with inherent variability. In the perfect case this variability
The DVR/T method does not separate specific temperature- can produce a range of zero, one, or several nucleation points
related effects but does make use of temperature during that grow into crystals. A perfect crystallization condition
optimization. The relationship between temperature and may not be identified as such, i.e., a clear drop could result.
solubility is readily observed in the results, guiding the One could replicate the number of experiments to try and
choice of a preferred crystallization condition. exclude this at the expense of sample needed. However, the
Crystallization is a complex process and the DVR/T DVR/T method provides a visual indication of the crystal-
embraces this as an optimization method. As described above lization space around a condition. A clear drop surrounded by
variables exploited by the nature of the method include a sea of drops containing small crystals can be immediately
[protein], [precipitating agent], pH, temperature, and even identified and that condition revisited. This does not preclude
the kinetics of supersaturation. All of these factors influ- missing a condition in the initial hit screen and the application
ence the outcomes of the crystallization experiments (Fig. 1). of DVR/T to that case is an area of future research.
These variables are highly correlated and not sampled in While DVR/T is a powerful method, a note of caution
fine-enough detail to gain theoretical insight but do provide needs to be sounded. The DVR/T experiments are not ideal,
an expansive empirical overview of a sample’s crystallization static batch experiments; they will dehydrate over time even

www.proteinscience.org 719
JOBNAME: PROSCI 16#4 2007 PAGE: 6 OUTPUT: Wednesday March 7 19:04:25 2007
csh/PROSCI/131728/ps0626997

Luft et al.

Materials and Methods

The samples used for the crystallization experiments were provided
by investigators through the Center for High-Throughput Structural
Biology (www.chtsb.org). These were soluble proteins of unknown
structure. A high-throughput infrastructure was used to screen for
initial crystallization conditions (Luft et al. 2003), thereby identi-
fying suitable cocktails for optimization. Eight proteins that
produced results, ranging in quality from needles and highly
twinned plates to small single crystals, were randomly selected
for the optimization experiments. The protein solutions used for
both the screening and optimization are listed (Table 1). None of
the samples had been crystallized previously. The screening facility
utilizes source/destination plate protocols with the TANGOä
liquid handling system (Matrix Technologies Corp.) to set up
Figure 3. Experiment drop composition. This graph illustrates the volumes microbatch-under-oil crystallization experiments. The experiments
of protein and cocktail solutions used to construct the experiment drops. were set up in 1536-well experiment plates (Greiner BioOne) using
Progressing from experiment 1 through 16, there is a sequential decrease in USP-grade mineral oil (Sigma). Upon completion, each experiment
the volume percent (and correlated concentration) of the protein solution, in plate contained a single-protein solution combined with an equal
the experiment drop ranging from a high of 77% (v/v) to a low of 25% volume of 1536 different crystallization cocktails under mineral
(v/v). The inverse relationship exists for the cocktail solution; its volume oil; each well held a unique experiment drop (200 nL protein + 200 nL
percent progressively increases from 23% (v/v) to 75% (v/v). The final drop cocktail). An automated imaging system recorded the experi-
volume is not held constant, ranging from 350 to 800 nL. This was done to ments’ outcomes immediately after the addition of the protein
reduce the amount of protein solution required to set up these experiments.

when covered by oil. The rate of dehydration is not

constant. Plates incubated at 4°C will dehydrate more
slowly than those stored at 37°C. The drop volume will in
part determine the impact of this dehydration. At constant
temperature, the dehydration will occur at approximately
the same rate for drops of different volumes. When the same
volume of water evaporates, the relative increase in con-
centration of nonvolatile components is roughly double for
a 350-nL compared to an 800-nL drop. This affects the rate
at which the protein solution supersaturates, an influential
variable on the outcome of crystallization experiments (Luft
and DeTitta 1997). Even in cases where DVR/T may fail to
produce crystals of sufficient quality it will provide consid-
erable insight on the effects of temperature and chemistry
on the sample’s solubility. This knowledge will aid a more
detailed approach to optimizing these recalcitrant samples.
DVR/T optimization experiments can be set up in any
laboratory that crystallizes biological macromolecules. The
method makes use of the initial hit and the biochemical
conditions associated with that hit alone. It is highly suited to
automation yet robotic techniques are not needed (but they
do speed up the process). In the case of the Center for High-
Throughput Structural Biology, a significant community of
700 researchers makes use of the high-throughput screening Figure 4. High-throughput DVR/T optimization. An entire source plate
holding as many as 96 cocktail solutions is simultaneously aspirated and
facilities. This laboratory, located at the Hauptman-Woodward dispensed into an oil-filled, 1536-well experiment plate using a 96-channel
Medical Research Institute, screens 200 samples each month liquid handling system. The same process is used to deliver protein
to identify crystallization conditions. Data tracking and image solution to the experiment plate. The 96 syringes in the liquid handling
packaging software needs to be developed and refined before system are arranged in an 8 3 12 array. The syringes are geometrically
the DVR/T method can be offered to this community of centered on every fifth well of a 1536-well plate. Each syringe dispenses
solution into a 4 3 4 array of 16 adjacent wells in the experiment plate.
researchers to optimize crystallization conditions identified Four replicate plates are prepared and separately incubated at temperatures
during screening. This will add significantly to the automa- of 4°C, 14°C, 23°C, and 37°C. Individual experiment plates hold as many
tion of the pipeline from macromolecule to structure. as 96 different proteins and cocktails at 16 different volume ratios each.

720 Protein Science, vol. 16

JOBNAME: PROSCI 16#4 2007 PAGE: 7 OUTPUT: Wednesday March 7 19:04:46 2007
csh/PROSCI/131728/ps0626997

Efficient optimization of crystallization

solution, and weekly thereafter for one month. Outcomes were Verification that the crystals were protein was accomplished
reviewed to identify those that have potential in optimization trials. either by observing color in the crystals (chromophores) or
For the optimization step, microbatch-under-oil crystallization through the addition of Izit dye to the experiment drop (Hampton
experiments were set up using a PlateMate2x2 liquid handling Research, Aliso Viejo, CA, USA). If the dye was absorbed by the
system equipped with a 96-channel syringe head (Matrix Tech- crystals, they were considered to be composed of protein.
nologies Corp.). Oil (5 mL) was delivered to each well of a 1536-
well microassay plate. Next, two separate 96-well source plates
were prepared. The first contained 25 mL aliquots of as many as Acknowledgments
96 different protein solutions; the second held aliquots of This work was supported by NIH U54 GM074899-01, the John
crystallization cocktails arrayed to match the proteins. The 96 R. Oishei Foundation, Margaret L. Wendt Foundation, and the
cocktail solutions were simultaneously aspirated from the source James H. Cummings Foundation. Special thanks to the follow-
plate (once) and dispensed into an oil-filled 1536-well experiment ing individuals for their enthusiastic and valuable collabora-
plate (16 times) at the volumes shown (Fig. 3). Each of the 96 tions: Bob Cudney (Hampton Research); Rachel Cochran
syringes dispensed one cocktail solution at volumes ranging from (Matrix Technologies); Ulrike Honisch, Guenther Knebel, and
0.15 to 0.60 mL into a 4 3 4 array in the experiment plate (Fig. 4). Bob Brino (Greiner-BioOne) and Brian Wright (Brook-Anco);
This process was repeated to prepare four replicate plates. The Daniel Degnan and Geoffrey Franks (Hauptman-Woodward
protein solutions were delivered in a volume range of 0.15–0.50 mL Medical Research Institute).
into the same four experiment plates that contained the cocktails.
These deliveries were completed in <1 h. Plates were centrifuged
at low speed to merge the cocktail and protein solutions. A single References
25-mL aliquot of protein was sufficient to prepare four replicate 16 Berman, H.M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T.N., Weissig, H.,
protein:cocktail drop volume ratio trials. The four plates were Shindyalov, I.N., and Bourne, P.E. 2000. The Protein Data Bank. Nucleic
separately incubated at temperatures of 4°C, 14°C, 23°C, and Acids Res. 28: 235–242.
37°C. These values were chosen to sample a range of temperatures Carter, C.W. Jr. and Carter, C.W. 1979. Protein crystallization using incomplete
nearing the extremes of the range for crystallization (4°C through factorial experiments. J. Biol. Chem. 254: 12219–12223.
Carter, C.W. and Yin, Y. 1994. Quantitative analysis in the characterization and
37°C); this includes values near commonly (4°C and 23°C) and optimization of protein crystal growth. Acta Crystallogr. D50: 572–590.
infrequently (14°C and 37°C) sampled temperatures (Charles et al. Charles, M., Veesler, S., and Bonnete, F. 2006. MPCD: A new interactive
2006). All of the crystallization experiments (screening and on-line crystallization data bank for screening strategies. Acta Crystallogr.
optimization) were set up and imaged weekly at 23°C. It required D62: 1311–1318.
25 min to image each plate; plates were returned to an incubator Chayen, N.E. 1998. Comparative studies of protein crystallization by vapour-
immediately after imaging. After setting the screening experi- diffusion and microbatch techniques. Acta Crystallogr. D54: 8–15.
Chayen, N.E. and Saridakis, E. 2002. Protein crystallization for genomics:
ments, protein solutions were stored at 80°C for several months Towards high-throughput optimization techniques. Acta Crystallogr.
before being used for optimization trials. The same mineral oil, D58: 921–927.
1536-well plates, protein solutions (Table 1), and cocktail sol- Chayen, N.E., Stewart, P.D.S., Maeder, D.L., and Blow, D.M. 1990. An
utions (Table 2) were used in both the screening and optimization automated system for micro-batch protein crystallization and screening.
trials. J. Appl. Crystallogr. 23: 297–302.
Chayen, N.E., Stewart, P.D.S., and Blow, D.M. 1992. Microbatch crystallization
under oil: A new technique allowing many small-volume crystallization
trials. J. Cryst. Growth 122: 176–180.
Table 2. Cocktail solutions used for optimization Christopher, G.K., Phipps, A.G., and Gray, R.J. 1998. Temperature-dependent
solubility of selected proteins. J. Cryst. Growth 191: 820–826.
Figure 1 Cocktail solution Cox, M.J. and Weber, P.C. 1988. An investigation of protein crystallization
parameters using successive automated grid searches (SAGS). J. Cryst.
A 100 mM Na Acetate pH 5.0, 100 mM NH4SCN, Growth 90: 318–324.
Dunlop, K.V., Irvin, R.T., and Hazes, B. 2005. Pros and cons of cryocrystallog-
20% (w/v) PEG 4,000
raphy: Should we also collect a room-temperature data set? Acta
B 100 mM MOPS pH 7.0, 100 mM NH4Br, 80% Crystallogr. D61: 80–87.
(v/v) PEG 400 Giegé, R. and Mikol, V. 1989. Crystallogenesis of proteins. Trends Biotechnol.
C 100 mM Na Citrate pH 4.0, 100 mM NaCl, 60% 7: 277–282.
(v/v) PEG 400 Haas, C. and Drenth, J. 2000. The interface between a protein crystal and an
D 100 mM CAPS pH 10.0, 100 mM K3PO4 20% aqueous solution and its effects on nucleation and crystal growth. J. Phys.
Chem. B 104: 368–377.
(w/v) PEG 4,000 Jancarik, J. and Kim, S.H. 1991. Sparse-matrix sampling: A screening method
E Crystal Screen HTä E12 (Hampton Research, for crystallization of proteins. J. Appl. Crystallogr. 24: 409–411.
Aliso Viejo, CA, USA), 100 mM Na Acetate Jurnak, F. 1985. Induction of elongation-factor Tu-GDP crystal polymorphism
pH 4.6, 100 mM CdCl2, 30% (v/v) PEG 400 by polyethylene-glycol contaminants. J. Mol. Biol. 185: 215–217.
F 100 mM Na Citrate pH 4.0, 100 mM (NH4)2HPO4, Kitano, K., Motohashi, K., Yoshida, M., and Miki, K. 1998. A novel approach
to crystallizing proteins with temperature-induction method: GrpE protein
40% (w/v) PEG 1,000
from Thermus thermophilus. J. Cryst. Growth 186: 456–460.
G 100 mM HEPES pH 7.0, 100 mM MgCl2, Koszelak, S., Leja, C., and McPherson, A. 1996. Crystallization of biological
20% (w/v) PEG 4,000 macromolecules from flash frozen samples on the Russian space station
H Indexä G12 (Hampton Research, Aliso Viejo, mir. Biotechnol. Bioeng. 52: 449–458.
CA, USA), 100 mM HEPES pH 7.5, Landsberg, M.J., Bond, J., Gee, C.L., Martin, J.L., and Hankamer, B. 2006. A
200 mM MgCl2, 25% (w/v) PEG 3350 method for screening the temperature dependence of three-dimensional
crystal formation. Acta Crystallogr. D62: 559–562.
I Crystal Screen HTä B6 (Hampton Research,
Leinala, E.K., Davies, P.L., and Jia, Z.C. 2002. Elevated temperature and
Aliso Viejo, CA, USA), 100 mM Na tyrosine iodination aid in the crystallization and structure determination of
Cacodylate pH 6.5, 200 mM Mg Acetate, an antifreeze protein. Acta Crystallogr. D58: 1081–1083.
20% (w/v) PEG 8,000 Liu, Y.S., Gotte, G., Libonati, M., and Eisenberg, D. 2002. Structures of the two
3D domain-swapped RNase A trimers. Protein Sci. 11: 371–380.

www.proteinscience.org 721
JOBNAME: PROSCI 16#4 2007 PAGE: 8 OUTPUT: Wednesday March 7 19:04:48 2007
csh/PROSCI/131728/ps0626997

Luft et al.

Luft, J.R. and DeTitta, G. 1997. Kinetic aspects of macromolecular crystal- Rayment, I. 2002. Small-scale batch crystallization of proteins revisited:
lization. Methods Enzymol. 276: 110–131. An underutilized way to grow large protein crystals. Structure 10:
Luft, J.R., Collins, R.J., Fehrman, N.A., Lauricella, A.M., Veatch, C.K., and 147–151.
DeTitta, G.T. 2003. A deliberate approach to screening for initial Ries-Kautt, M. and Ducruix, A. 1992. Phase diagrams. In Crystallization of
crystallization conditions of biological macromolecules. J. Struct. Biol. nucleic acids and proteins. (eds. A. Ducruix and R. Giege), pp. 195–218.
142: 170–179. IRL Press/Oxford University Press, New York.
McPherson, A. 1995. Increasing the size of microcrystals by fine sampling of Rosenberger, F. 1986. Inorganic and protein crystal growth-similarities and
pH limits. J. Appl. Crystallogr. 28: 362–365. differences. J. Cryst. Growth 76: 618–636.
Ng, J.D., Kuznetsov, Y.G., Malkin, A.J., Keith, G., Giege, R., and Snell, E.H., van der Woerd, M.J., Damon, M., Judge, R.A., Myles, D.A.A., and
McPherson, A. 1997. Visualization of RNA crystal growth by atomic force Meilleur, F. 2006. Optimizing crystal volume for neutron diffraction:
microscopy. Nucleic Acids Res. 25: 2582–2588. D-xylose isomerase. Eur. Biophys. J. 35: 621–632.

722 Protein Science, vol. 16