Transcript Slide 1

Cloning through diffraction: Goals and technologies at the
Center for High-throughput Structural Biology (CHTSB).
INTRODUCTION: The goal of the Center for High-Throughput Structural
Biology (CHTSB) is to overcome the most significant obstacles to structure
determination by focusing on technology development in areas related to
sample preparation for X-ray diffraction studies. CHTSB is a specialized
research center created as part of the NIH Protein Structure Initiative. The
center is focusing on several areas; developing yeast (Saccharomyces
cerevisiae) as a tool for structural biology, the efficient screening, optimization
and production of crystals, and technology for crystal handling and remote data
collection. The targets include transmembrane proteins and protein-protein
complexes that have traditionally been difficult to study by diffraction methods.
TECHNOLOGY DEVELOPMENTS: A summary of CHTSB technologies
describing their fit within the target to structure pipeline, and maturity.
established biochemical assays to identify whether the protein is maintained in
a native state in the presence of detergents. Genes encoding the target TMPs
are transferred via ligation-independent cloning (LIC) procedures to a series of
vectors that allow galactose-controlled expression of reading frames fused to
C-terminal His6, His10, and ZZ (IgG-binding) domains that are separated from
the reading frame by a cleavage site for rhinovirus 3C protease. Several TMP
targets expressed from these vectors have been purified via affinity
chromatography and gel filtration chromatography at levels and purities
sufficient for ongoing crystallization trials. Current efforts are focused on
overcoming bottlenecks in protein production and crystallization by introducing
the following improvements in the production pipeline: 1) improving overall
levels of cellular expression of TMPs by altering protocols for cell growth and
induction of expression; 2) increasing efficiency of cell lysis; 3) increasing the
efficiency of detergent solubilization; 4) increasing the efficiency of proteolytic
removal of affinity tags; 5) developing protocols for obtaining highly
concentrated protein preparations that do not contain high detergent
concentrations; 6) optimizing the amount of residual lipid purifying with the
TMP; 7) reducing the number of steps required for effective purification; 8)
testing the use of additives such as lipids and enzyme inhibitors to stabilize
purified TMP (White at al., 2006).
After setting up screening experiments, the outcomes are digitally recorded 1
day after the addition of protein and weekly thereafter for a period of 4 weeks
using 3 custom-designed imaging systems. Each has the capacity to hold 28
plates and image at a rate of 3 plates (4500 experiments) per hour. Images are
stored on RAID arrays and archived on offline media. Outcomes are reviewed
to identify combinations of proteins and cocktails that demonstrate a propensity
to crystallize.
Successful combinations advance through the structuredetermination pipeline to the crystal-growth optimization stage. Samples that
fail to produce any screening outcomes suitable for optimization are analyzed
for solubility data that can be used to optimize the protein formulation.
(b)
(a)
Development of Methods for Medium-Throughput Cloning, Expression,
Analysis, and Purification of Protein:Protein Complexes: To expedite
cloning and expression analysis for protein:protein complexes, Drs. Eric
Phizicky and Elizabeth Grayhack have developed a suite of vectors for use in
yeast strains with commonly found genetic markers. Each vector features
high-throughput, ligation-independent cloning and is designed to express genes
under control of the strong, regulatable galactose promoter (PGAL), thereby
permitting the cloning of genes that are otherwise toxic as well as achieving
high-level expression of the corresponding proteins.
An initial set of vectors express individual ORFs (Open Reading Frames) with a
tri-partite affinity tag on their C-terminus. This tag is comprised of a protease
3C site, an HA epitope, a His6 tag for immobilized metal ion affinity
chromatography (IMAC) purification, and the ZZ domain of protein A for
purification on IgG Sepharose. Binary protein complexes are made using two
otherwise identical vectors (one with a LEU2 marker and one with a URA3
marker), each expressing different genes of the complex.
Figure 1. A breakdown of specific areas within the target to structure pipeline that
are being developed by groups making up the CHTSB. All of these technologies are
being developed as tools for the biology, structural biology, and structural genomics
communities. The color-coding used in this chart corresponds to sections of
similarly color-coded , detailed descriptions of the technologies in the main text.
PRODUCTION - YEAST AS A TOOL
Engineering Yeast for Efficient and Robust Incorporation of
Selenomethionine into Expressed Proteins: Yeast is a useful system for
expressing protein. Its advantages include: 1) It is a eukaryotic system; 2) The
ability to produce large amounts of protein; 3) Rapid and inexpensive culturing;
4) Completeness of genetic, genomic, metabolic characterization; 5) History of
usefulness for structure determination; 6) Similarities to mammalian cells in
post-translational modifications, sub-cellular trafficking, protein folding,
biological pathways; 7) Availability of yeast strains with altered protein
degradation, unfolded protein response, post-translational modifications,
intracellular trafficking; 8) Existence of large, interactive community of yeast
laboratories; 9) Ease of genetic manipulation. However, because of the toxicity
of selenomethionine to yeast, it was difficult to achieve efficient
selenomethionine incorporation into yeast-expressed proteins. Substituting
selenomethionine for methionine is a method commonly used for
multiwavelength anomalous dispersion (MAD) phasing methods with other
expression systems.
To overcome the bottleneck of selenomethionine incorporation using a yeast
expression system the source of selenomethionine toxicity was identified, and
a protocol for selenomethionine incorporation in yeast-expressed protein was
developed. These results constitute a general solution to the problem of
effective selenomethionine incorporation into proteins expressed in yeast, and
remove a major obstacle for the use of yeast as an alternative to E. coli for
expression of proteins for structural analysis by X-ray crystallography.
Furthermore, these results suggest that efficient selenomethionine
incorporation in other organisms might be effected by the employment of
similar methods to prevent conversion of selenomethionine into the source of
the toxicity, Se-adenosylselenomethionine (Malkowski et al., 2007).
Yeast Trans Membrane Protein Production: To address the severe lack of
three-dimensional structural information for eukaryotic transmembrane proteins
(TMPs), Dr. Mark Dumont and his colleagues at the University of Rochester
are developing protocols for the expression and purification of TMPs in the
yeast Saccharomyces cerevisiae. Initial efforts have been focused on a set of
endogenous yeast TMPs that are the highest expressing reading frames in a
previously constructed genomic collection of Saccharomyces cerevisiae
expression clones. This collection targets reading frames for which there are
More recently developed vectors allow for expression of two proteins
simultaneously, using a bi-directional PGAL promoter, with different tags on
each ORF of a pair. Three such sets of vectors, containing either a LEU2
marker or a URA3 marker, have been constructed for expression of an ORF3C-HA-His6-ZZ fusion together with an untagged ORF, a His6-tagged ORF, or
a His10-tagged ORF. These vectors have three important uses. First, vectors
with different tags can be used to unequivocally demonstrate the existence of a
complex, by co-purification of ORFs using only one tag on one ORF. Second,
use of a vector with a non-cleavable His6 or His10-tag on one ORF, and a
cleavable ZZ tag on the other ORF, allows for efficient purification using two
affinity steps, each directed at one ORF of the complex. Third, use of pairs of
vectors, each with a bi-directional PGAL promoter, allows for purification of
complexes comprised of as many as four protein subunits. Expression with
these vectors can be as high as 15 mg/L in the best cases.
Several medium-throughput methods for analysis of expression and purification
of protein complexes have also developed. These employ the use of small
cultures, rapid analysis of expression using stick-and-strip binding to either IgG
Sepharose or IMAC followed by analysis of proteins on SDS-PAGE, and rapid
analysis of authentic complexes employing differentially tagged proteins.
Major ongoing research efforts in the lab are focused on the development of
methods to further enhance expression of proteins in yeast, on further
documentation of the utility of yeast for expression of exogenous proteins and
complexes of proteins, and on development of medium-throughput methods for
detection of complexes suitable for purification and subsequent structural
analysis (Phizicky et al., 2006).
CRYSTALLIZATION
High-Throughput Crystallization Screening: Purified, soluble proteins are
processed through a mature, high-throughput screening facility (located at the
Hauptman-Woodward Medical Research Institute) to identify crystallization
conditions. Syringe-based liquid-handling systems are used to prepare
microbatch-under-oil crystallization experiments in 1536-well microassay
plates. Each experiment plate holds a single macromolecule combined with
1536 chemical solutions (cocktails). Individual experiments are composed of
200 nanoliters of protein solution and 200 nanoliters of cocktail solution. A
screen of 1536 experiments is set up with 400 nL of protein solution (at a
concentration of ~10 mg/mL). The high-throughput laboratory currently screens
as many as 200 different macromolecules each month from structural
genomics centers and the general scientific community. As of March 2008, 852
investigators sent 9855 samples to the laboratory for screening, resulting in 15
million crystallization experiments. In addition to the CHTSB efforts this facility
is used by other structural genomics groups and serves the general biomedical
community (Luft et al., 2003).
Figure 2. (a) A 1536 well plate with enlargements of four crystallization hits and (b)
two of the three imaging tables. Two of the tables are at 23oC while the third is in a
temperature-controlled room to maintain the temperature used for crystallization
during imaging.
Development of a Membrane Protein Crystallization Screen based upon
Detergent Phase Boundaries: Crystallization conditions for a proteindetergent complex are often near the detergent phase boundary (Weiner
2001). Using dye-partitioning, the phase behavior of 10 detergents, 7 PEGs,
and 9 salts were sampled at both 4 and 23oC. These phase boundaries
serve as a guide to formulate detergent-specific cocktails that are less likely
to produce an immediate detergent phase separation and subsequent TMP
precipitation during crystallization screening. The cocktails target those
regions of the phase diagram that is most likely to promote favorable
interactions for crystallization.
(a)
(b)
Figure 3. (a) Dye-partitioning highlights the phase behavior of a detergent in the
presence of gradually increasing concentrations of salt and PEG. Wells that are
uniform in color are not sufficiently supersaturated to cause phase separation of the
detergent.
Wells that display a marked contrast indicate phase-separated
detergent. The crystallization cocktails designed from this data target the border
region just below the phase boundary. (b) The phase-boundary data for one
detergent in the presence of gradually increasing concentrations of a single salt and
five chemically distinct PEGs.
Silver Bullets to Promote Macromolecular
Crystallization : A fundamentally different
approach was used to develop a set of ‘silver
bullet’ cocktails. 200 small molecules were
combined to form reagent mixes containing 3
to 20 different chemicals. These cocktails are
formulated with one of two precipitating
agents; 30% PEG 3350, or 50% TacsimateTM,
both buffered at pH 7.0. The small molecules
within the cocktails promote macromolecular
contacts. The macromolecule selects a small
molecule from the reagent. This small
molecule will often appear in the crystal
structure at a location that promotes
structural stability. These cocktails are
commercially available (Hampton Research,
HR2-096) (McPherson et al., 2006; Larson et
al., 2007a,b)
Figure 4. Mellitic acid ‘grappling
hook’ with 6 carboxylate groups
in a unique bovine trypsin
structure
Optimization of Crystallization Conditions: Combinations of protein and
cocktail solutions that produced crystalline outcomes during screening are
formatted into 96 well source plates without reformulation. Solutions are
dispensed to set up a series of 16 experiment drops with a percentage
composition of decreasing protein solution and increasing cocktail solution.
Replicate experiment plates are prepared and separately incubated at different
temperatures. In cases where DVR/T fails to produce crystals of sufficient
quality for X-ray diffraction analysis, it still provides considerable insight into
the effects of temperature and chemistry on the sample’s solubility (Luft et al.,
2007).
Remote Data Collection: By working closely with crystallographers at the
to enhance the capabilities, and improve the efficiency of the crystallization
laboratory (Figure 7a-f). (a)“Slide ‘n Shoot Pro” controls imaging hardware,
with auto-centering plate definition routines. (b)“CrystalWriter”, controls optical
disc hardware to back up image data. (c)“DVR/T Setup” & “WellPacker”
creates DVR/T experiment documents and packages DVR/T results/images,
(d)“AutoSherlock” for presenting screening experiment results in chemical
space. (e) This “AutoSherlock” screenshot displays outcomes of crystallization
results plotted on a grid showing the conditions sampled and more importantly
those not. In this case PEG and pH play a major role in optimization. The
“Grid” shows parameters that should be sampled by a grid screen for
optimization. (f) An example of how populating the crystallization results in
chemical space (using AutoSherlock) can show a potential area for
crystallization occurring between a precipitate and clear conditions.
Macroscope, a software for viewing screening results and manual scoring,
has been available at no cost to users of the crystallization screening
laboratory for the past 8 years.
Hauptman-Woodward Institute (HWI), the Structural Molecular Biology (SMB)
group at the Stanford Synchrotron Radiation Laboratory (SSRL) is perfecting
the technology for remote access data collection. The SSRL SMB group
operates six beam lines (BL1-5, BL7-1, BL9-1, BL9-2, BL11-1, and BL11-3)
for macromolecular crystallography experiments. Additionally, a new undulator
station optimized for microcrystal data collection, BL12-2, is being
commissioned. The users of these beam lines have the option to collect data
remotely and, during the 2007 run period, about 70% of experiments have
been scheduled for remote access thereby saving travel time and
expenses. While remaining at their home institutions, remote users conduct
experiments by means of advanced software tools enabling network-based
systems monitoring and control. Remote users have the capability to mount,
center (Song et al., 2007), and screen samples as well as to collect, analyze,
and backup diffraction data. Automated sample mounting is accomplished
with the Stanford Auto-Mounting System (SAM). In this way, users can screen
crystals 2-3 times more rapidly with less human error. For rapid crystal
ranking and data analysis, the diffraction images collected during screening
are automatically analyzed.
(a)
(b)
(c)
Beamline Emulator: An automated set up that
provides rapid X-ray based feedback to the crystal
optimization has been constructed at HWI. The
system mimics the SSRL Stanford Automated
Mounting Systems (SAM) (Cohen et al. 2002).
Based on typical data collection times 2 cassettes of
samples, 192 in total, will be screened in the
laboratory and sorted in ~70 hours. This prescreening will guide crystal optimization (based on
diffraction quality) and maximize productive use of
beamtime at SSRL.
Image Analysis: There are over 90 million images generated by the screening
20% PEG 8K
20% PEG 4K
7
6
5
PEG 4K
83% correct
PEG 1K
(a)
(b)
(c)
(d)
Figure 6. Tip/capillary
development showing (a)
a schematic of the plastic
capillary/tip, (b) the tip
positioned within a unipuck illustrating how the
form factor is compatible
with many systems that
utilize
a
standard
magnetic cap, (c) an
early test example with a
crystal within the tip (and
an enlarged image of the
crystal) (d) diffraction
from a crystal within one
of
the
plastic
tips
cryocooled
under
pressure
(see
later
section).
6
Small
to
large
Crystals
PEG 0.4K
Clear
5
4
Bad
Bad
Figure 8. SAM system installed at HWI with Mar 345 imaging plate detector.
Small
to
large
Grid
Grid
Grid
Grid
DATA MINING
Grid
Databases and Information Systems: Data mining in all areas is under
development. Data mining on a subset of crystallization results is underway.
The aim is to expand the analysis developed on this subset to analyze to the
full archive of results.
Unknown
Figure 7. Software developed by
CHTSB includes: (a)“Slide ‘n Shoot
Pro”; (b) “CrystalWriter”; (c) “DVR/T
Setup” & “WellPacker”; and (d-f)
“AutoSherlock”.
The
first
two
developments improve efficiency in the
laboratroy,
DVR/T
Setup
and
WellPacker are used to present
optimization data and AutoSherlock is
in beta testing with the general
biomedical community to analyze their
crystallization screening results.
14000
7.00%
13091
6.15%
NESG and SGPP combined
6.01%
12000
6.00%
% hits/xpr
hits
4.91%
4.81%
5.00%
4.49%
4.00%
3.67%
3.58%
3.44%
3.08%
3.02%
2.85%
3.00%
0.03
0.025
0.02
0.015
0.01
0.005
0
8000
2.98%
6000
2.24%
2.13%
0.04
0.035
10000
4.59%
4.53%
8441
2.16%
4153
1.77%
2.00%
4000
1.55%
2466 2345
1.22%
1823
1.00%
497
0.96%
0.82%
1092
715
115
271
100
51
122
213
580
317
2000
690
227
57
1076
4
334
0
0.00%
5
6
pH
7
40%
8
9
10
20%
Concentration
PEG 8K
r(
8
PEG 8K
are developing pipette tips as vessels for crystal growth cryopreservation, and
in situ X-ray diffraction. This will improve efficiency and eliminate potential
damage caused by physical manipulation. This work is a collaboration
between SSRL, HWI, and Cornell. Our collaborators at SSRL designed a
pipette tip for compatibility with HWI liquid-handling systems and the
automated sample handling system at SSRL. Tips produced from a number of
different plastics were tested for X-ray diffraction properties (SSRL),
compatibility with robotics (HWI, SSRL) (Cohen et al. 2002), crystallization
(HWI), and cryo-compatibility (Cornell). Large scale production of tips is now
underway.
7
Bad
20% PEG 1K
Precipitate
Crystal Production, Cryopreservation, and Diffraction in Capillaries: We
8
Grid
(f)
Table 1. Performance of current classifier. Precipitated and clear drops on average
constitute 83% (1275 out of 1536) outcomes of the HT crystal screen. If we can
focus our efforts on the 17% (261 out of 1536) images that are most likely to
contain crystals, we can eliminate time-consuming manual review of the lessinteresting outcomes. By reordering the images, placing clear and precipitated
drops after the more-interesting outcomes, we can significantly improve our
efficiency and throughput.
9
20% PEG 20K
pH
98% correct
10
pH
15
1)
[1
PE 23 7
]
G
20
00
PE
0
G
80
00
PE
G
PE
40
00
G
PE
/S
G
al
t/B
10
00
uf
fe
PE
r(
HR
G
69
40
C
2)
ry
0
[2
st
38
al
98
Sc
4]
re
en
C
HR
r
HR yo
N
HR
H
uc
R
N
G
at
le
Q
rid
rix
ic
ui
k
A
Sc
ci
Sc
d
re
re
M
en
en
in
So
iS
HR
d
c
re
i
G
en
rid um
M
Sc
al
re
on
at
HR e n
e
HR
HR
P
EG
PE
G
G
G
rid
rid
/Io /LiC
S
Sc
l
n
Sc
re
HR cre
en
en
re
G
en
Am
PE
rid
G
m
Sc
on
60
re
iu
00
en
m
So
Su
HR
di
um lfa t
e
C
ry
C
hl
st
HR al S orid
e
cr
In
d e een
x
HR
Sc H T
re
(5
en
HR
18
HT
){
Al
Sa
l C [98
lt R
5oc
x
15
kt
H
36
ai
T
ls
]+
(1
C
36
ry
1)
o}
[1
-1
53
6]
(e)
Bu
ffe
(d)
Sa
lt/
laboratory that remain unclassified. A training set of images to guide softwarecontrolled classifiers was compiled.
Crystallization outcomes from 96
macromolecular samples that underwent 1536 cocktail screening (MW range
of 10 – 2000 kDa) were manually classified by 8 reviewers (each image
classified by 3 reviewers, with an even distribution of reviewers/image). Of the
147,456 images selected for this study 70, 465 were unanimously classified.
Images fell into ten main categories. The distribution of precipitates (41.7%)
and clear drops (41.2%) was nearly identical; 49 of the 96 samples had 1 or
more outcomes classified as crystals; crystals were a rare occurrence
appearing in 0.38% of the images. Machine classification is taking place at
OCI using the human-classified ‘truth data’. Feature generation that will lead to
the development of more accurate classifiers is progressing on IBM Blue
Gene, and The World Community Grid. Currently over 7,500 years of computer
processing time has been used for the distributed project.
Percentage of hits per
experiment
Figure 5. (a) Cocktails and proteins are arranged into two separate 96 well source
plates. A 96 syringe liquid –handling system is used to aspirate from the source
plate of cocktail and protein and deliver the solutions at 16 different volume ratios
into a 1536 well plate containing USP grade mineral oil. The experiment plates are
prepared in replicate for incubation at different temperatures. (b) Illustrates the
impact of temperature on the solubility of a protein sample from the structural
biology community. This is the same protein, P6306 set up with two different
cocktail solutions. Note the cocktail dependent inverse relationship of the
temperature-dependent solubility displayed by these experiments. Regions of
clear drops that border on drops having phase separation (crystals, precipitate)
are likely to be at or near metastable supersaturation, ideal for seeding.
Software Developments for Crystallization: Software has been developed
Figure 9. Early data mining results from scored crystallization trials.
DIFFRACTION DATA
SUMMARY
High Pressure Cryocooling and Heavy Atom Derivatives: Successful protein
CHTSB has been focused on technology developments. These developments
are now bearing fruit in terms of production, identification of crystallization
leads, optimization and diffraction. The tools developed feed into the
corresponding pipeline from target to structure but individually are applicable
and available to the wider general biomedical community.
crystallography typically requires that crystals be cryocooled in order to reduce
radiation damage at the data collection stage. So far, protein crystals have most
commonly been frozen by flash cryocooling at ambient pressure, which often
requires a time-consuming search for cryoprotection conditions. These
conditions have been established for all the crystallization screening cocktails in
the HWI screen (Kempkes et al., 2008). Recently, the CHTSB-affiliated team
(led by Prof. Sol Gruner) at the Cornell High Energy Synchrotron Source
(CHESS) has developed an alternative procedure, high-pressure cryocooling,
which often does not require the addition of chemical cryoprotectants. The result
of high-pressure cryocooling is very often a dramatic improvement in the quality
and resolution of the diffraction data. The Cornell team is currently investigating
the basic underlying principles and experimental parameters important for the
optimization of the high pressure cooling method (Kim et al., 2007; 2008).
Since the high-pressure method involves the use of helium gas as a
pressurizing medium, attention has been focused on its extension to diffraction
phasing by incorporating heavy noble gases such as krypton or xenon. In the
test case of Porcine Pancreatic Elastase (PPE, 240 residues, 26kDa), very high
quality diffraction was obtained by the modified high-pressure cyrocooling
method without the help of cryoprotectants. Furthermore, a single krypton site
with an occupancy of 0.31 could be used successfully for SAD phasing at 1.3Å
resolution. The Cornell team is currently working on equipment modifications
that will be compatible with a high throughput crystallography pipeline (Kim et
al., 2006).
ACKNOWLEDGEMENTS
Work supported by NIH U54 GM074899, the John R. Oishei Foundation,
Margaret L. Wendt Foundation, Erie County and the James H. Cummings
Foundation.
REFERENCES
• Cohen et al., (2003). An automated system to mount cryo-cooled protein crystals on a synchrotron beamline, using compact sample cassettes
and a small scale robot. J. Appl. Cryst. 35, 720-726.
• Collins, M.C., et al. (2007). Structural Rigidity of a Large Cavity-Containing Protein Revealed by High-Pressure Crystallography. J. Mol. Biol.
367: 752-763.
• Kempkes, et al. (2008). Glycerol Concentrations Required for the Successful Vitrification of Cocktail Conditions in a High-Throughput
Crystallization Screen. Acta Cryst. D64:287-301.
• Kim, C., et al. (2006). Solution of Protein Crystallographic Structures by High-Pressure Cryocooling and Noble-Gas Phasing. Acta Cryst. D62 :
687-694
• Kim, C., Hao, Q., and Gruner, S.M. (2007). High Pressure Cryocooling for Capillary Sample Cryoprotection and Diffraction Phasing at Long
Wavelengths. Acta Cryst. D63: 653-659.
• Kim, C.U., et al. (2008). Pressure Induced High-density Amorphous Ice in Protein Crystals. J. Appl. Cryst. 41:1-7.
• Luft et al., (2003).A deliberate approach to screening for initial crystallization conditions of biological macromolecules. J. Struct. Biol. 142, 170179.
• Luft, J.R., et al. (2007). Efficient Optimization of Crystallization Conditions by Manipulation of Drop Volume Ratio and Temperature. Protein
Science 16: 715-722.
• Larson, S.B., et al. (2007). A Novel Strategy for the Crystallization of Proteins: X-ray Diffraction Validation. Acta Cryst. D63:310-318.
• Larson, S.B., et al. (2007). A new crystal form of bovine pancreatic RNase A in complex with 2'-deoxyguanosine-5'-monophosphate. Acta Cryst
F63:728-33.
• Malkowski, M.G., et al. (2007). Blocking Conversion of Methionine to S-Adenosylmethionine in S. cerevisiae Allows Selenomethionine
Incorporation and Multiwavelength Anomalous Dispersion Phasing. Proc. Natl. Acad. Sci., USA, 104: 6678-6683.
• McPherson, A. & Cudney, R. (2006). An Alternative Strategy for Crystallizing Macromolecules. J. Structural Biology 156: 387-406.
• McPherson, A., et al. (2007). Development of an Alternative Approach to Protein Crystallization. J. Struct. Funct. Genomics. 8:193-198.
• Phizicky, E.M. & Grayhack, E.J. (2006). Proteome-Scale Analysis of Biochemical Activity. Crit. Rev. Biochem. Mol. Biol. 41:315-327.
• Song, J., et al. (2007). Diffraction-based Automated Crystal Centering. J. Synchrotron Rad. 14:191-195.
• White, M.A., et al. (2006). Characteristics Affecting Expression and Solubilization of Yeast Membrane Proteins. J. Mol. Biol. 365 : 621-636.
People
Department of
Physics/CHESS,
162 Clark Hall,
Cornell University,
Ithaca NY 14853
Building 120, SLAC,
2575 Sand Hill Road,
Menlo Park, CA 94025
Eleanor Cook
George DeTitta
Christopher Goulah
Angela Lauricella
Joseph Luft
Michael Malkowski
Raymond Nagel
Walter Pangborn
Stephen Potter
Mary Rosenblum
Meriem Said
Edward Snell
Elizabeth Snell
Max Thayer
Christina Veatch
Charles Weeks
Jennifer Wolfley
Yoshiko Kon
Maryann Mikucki
Eric Phizicky
Erin Quartley
Katrina
Robinson
Kathy Clark
Sara Connelly
Mark Dumont
Nadia Fedoriw
Elizabeth Grayhack
Joseph Chang
Aina Cohen
Tzanko Doukov
Keith Hodgson
Michael Soltis
Yi-Fan Chen
Sol Gruner
Chae Un Kim
John Day
Aaron Greenwood
Steven Larson
Yri Kuznetsov
Alexander McPherson
Department of Molecular Biology
and Biochemistry,
University of California,
560 SH, Mail Code 3900
Irvine, CA 92697
Christian Cumbaa
Igor Jurisica
Division of Signaling Biology,
Ontario Cancer Institute,
Princess Margaret Hospital,
610 University Avenue,
Toronto, ON, M5G 2M9