Protein crystallography in practice
MCB
5 Dec 2013
Daved H. Fremont
Department of Pathology and Immunology
Washington University School of Medicine
An 7-step program for protein structure
determination by x-ray crystallography
1. Produce monodisperse protein either alone or as relevant complexes
2. Grow and characterize crystals
3. Collect X-ray diffraction data
4. Solve the phase problem either experimentally or computationally
5. Build and refine an atomic model using the electron density map
6. Validation: How do you know if a crystal structure is right?
7. Develop structure-based hypothesis
1.
Produce monodisperse protein
either alone or as relevant complexes
Methods to determine protein purity, heterogeneity, and monodispersity

Gel electrophoresis (native, isoelectric focusing, and SDS-PAGE)

Size exclusion chromatography

Dynamic light scattering http://www.protein-solutions.com/

Circular Dichroism Spectroscopy http://www-structure.llnl.gov/cd/cdtutorial.htm
Characterize your protein using a number of biophysical methods
Establish the binding stoichiometry of interacting partners
2. Grow and characterize crystals
Hanging Drop vapor diffusion
Sitting drop, dialysis, or under oil
Macro-seeding or micro-seeding
Sparse matrix screening methods
Random thinking processes, talisman, and luck
The optimum conditions for crystal nucleation are not
necessarily the optimum for diffraction-quality crystal growth
Space Group P21
4 M3 /ASU
diffraction >2.3Å
14.4% Peg6K
NaCacodylate pH 7.0
200mM CaCl2
Space Group C2
2 M3 /ASU
diffraction >2.1Å
18% Peg4K
Malic Acid/Imidazole pH 5.1
100mM CaCl2
Hanging Drop
Sitting drop
Commercial screening kits available from
http://www.hamptonresearch.com;
http://www.emeraldbiostructures.com
Space Group P3121
3 M3 + 3 MCP-1/ASU
diffraction > 2.3Å
18% Peg4K
NaAcetate pH 4.1
100mM MgCl2
No Xtals?
Decrease protein heterogeneity

Remove purification tags and other artifacts of protein production

Remove carbohydrate residues or consensus sites (i.e., N-x-S/T)

Determine domain boundaries by limited proteolysis followed by mass
spectrometry or amino-terminal sequencing. Make new expression constructs
if necessary.

Think about the biochemistry of the system! Does your protein have cofactors, accessory proteins, or interacting partners to prepare as complexes?
Is their an inhibitor available? Are kinases or phosphatases available that will
allow for the preparation of a homogeneous sample?

Get a better talisman
Building a crystal
The unit cell
c
c
b b
a
b
a
b g
g
a
a
Crystal symmetries
A triclinic lattice
(no symmetry)
Crystal symmetries
Introducing a twofold axis produces
a monoclinic lattice P2
Crystal symmetries
The threefold axis generates a trigonal crystal
- but now a=b=90o, g=120o
and a = b
Crystal symmetries
We cannot fill space with a
fivefold arrangement – although
the asymmetric unit can contain
a fivefold axis (e.g. virus capsids)
These restrictions give rise to
7 crystal classes in 3 dimensions
The seven crystal classes
3. Collect X-ray diffraction data
Initiate experiments using home-source x-ray generator and detector
Determine liquid nitrogen cryo-protection conditions to reduce crystal decay
While home x-rays are sufficient for some questions, synchrotron radiation is preferred
Anywhere from one to hundreds of crystals and diffraction experiments may be required
Argonne National Laboratory Structural Biology Center beamlineID19
at the Advanced Photon Source http://www.sbc.anl.gov
3. Collect X-ray diffraction data
Lawrence Berkeley National Laboratory
ALS Beamline 4.2.2
4. Solve the phase problem either
experimentally or computationally
Structure factor equation:
By Fourier transform we can obtain the electron density.
We know the structure factor amplitudes after successful data collection.
Unfortunately, conventional x-ray diffraction doesn’t allow for direct phase measurement.
This is know as the crystallographic phase problem.
Luckily, there are a few tricks that can be used to obtain estimates of the phase a(h,k,l)
Experimental Phasing Methods
MIR - multiple isomorphous replacement - need heavy atom incorporation
 MAD - multiple anomalous dispersion- typically done with SeMet replacement
MIRAS - multiple isomorphous replacement with anomalous signal
SIRAS - single isomorphous replacement with anomalous signal
Computational Methods
MR - molecular replacement - need related structure
Direct and Ab Initio methods - not yet useful for most protein crystals
MAD phasing statistics for the AP-2 a-appendage
5. Build an atomic model using the electron density map
Electron density for the AP-2 aappendage
Initial bones trace for the AP-2 aappendage
Final trace for the AP-2 aappendage
What does a good map look
like?
plexiglass
stack
brass
parts
model
Before computers, maps were contoured
on stacked pieces of plexiglass. A
“Richards box” was used to build the
model.
halfsilvered
mirror
Low-resolution
At 4-6Å resolution, alpha helices look
like sausages.
Medium resolution
~3Å data is good enough to see the backbone
with space in between.
Holes in rings are a good thing
Seeing a hole in a tyrosine or phenylalanine ring is universally
accepted as proof of good phases. You need at least 2Å data.
The resolution of the electron-density
map and the amount of detail that can
be seen
Resolution
Structural Features Observed
5.0 Å
3.5 Å
3.0 Å
2.8 Å
2.5 Å
Overall shape of the molecule
Ca trace
Side chains
Carbonyl oxygens (bulges)
Side chain well resolved,
Peptide bond plane resolved
Holes in Phe, Tyr rings
Current limit for best protein
crystals
1.5 Å
0.8 Å
The 2.8 Å density of SrfTE
The 2.8 Å density of SrfTE could be skeletonised
and traced
6. Validation: How do you know if a crystal structure is right?
The R-factor
R = S(|Fo-Fc|)/S(Fo)
where Fo is the observed structure factor amplitude and Fc is calculated using the atomic model.
R-free
An unbiased, cross-validation of the R-factor. The R-free value is calculated with typically 5-10% of the observed reflections which are
set aside from atomic refinement calculations.
Main-chain torsions: the Ramachandran plot
Geometric Distortions in bond lengths and angles
Favorable van der Waals packing interactions
Chemical environment of individual amino acids
Location of insertion and deletion positions in related sequences
6. Validation: How do you know if a crystal structure is right?
6. Validation: Mapping of sequence conservation in AP-2 a-subunit appendages
Traub LM, Downs MA, Westrich JL, and Fremont DH: (1999) Crystal structure of the
a-appendage of AP-2 reveals a recruitment platform for clathrin-coat assembly. Proc.
Natl. Acad. Sci. U.S.A. 96:8907-8912.
7. Develop structure-based hypothesis
Structure-Based Mutagenesis of the a-appendage
Traub LM, Downs MA, Westrich JL, and Fremont DH: (1999) Crystal structure of the
a-appendage of AP-2 reveals a recruitment platform for clathrin-coat assembly. Proc.
Natl. Acad. Sci. U.S.A. 96:8907-8912.
Example: West Nile Virus
About 70 members, half of which are associated
with human disease (Yellow fever, Japanese encephalitis)
Enveloped, spherical virion, 40 - 50 nm in size
Three structural proteins: C,M (prM) and E ; seven
non-structural proteins (NS1-5)
ssRNA genome, linear, positive polarity, 11 kb,
infectious
C M
E
NS1 NS2a 2b NS3
NS4a 4b
NS5
5’UTR
Structural proteins
3’UTR
Non-structural proteins
Production of soluble E proteins and ectodomain fragments
Immunize mice with
soluble E (25 mg x 3)
Fuse splenocytes
with myeloma line
Large panels of flavivirus mAbs
Structure Determination of WNV Envelope Protein
Table 1. Summary of Data Collection and Refinement
Data Collection for West Nile Virus Envelope a
Space Group
P41212
Unit Cell (Å3)
a=89.6 b=89.6 c=154.0
Wavelength(Å)
0.90
X-ray Source
APS-BM 14
Resolution(Å) (outer shell)
20-2.9 (3.08-2.90)
Observations/Unique
14408/62790
Completeness(%)
98.5 (99.5)
Rsym(%)
5.7 (52.4)
I/s
16.9 (2.05)
Refinement Statistics b
Resolution(Å) (outer shell)
20-3.0 (3.19-3.00)
Reflections Rwork/Rfree
11506/607
#Protein Atoms/Solvent/Heterogen
3031/28/38
Rwork overall(outer shell) (%)
26.2(35.6)
Rfree overall(outer shell) (%)
30.8(34.1)
o
Rmsd Bond lengths (Å)/angles( )
0.008/1.6
Rmsd Dihedral/Improper (o)
24.9/0.84
Ramachandran plot
Most Favored/Additional (%)
78.2/21.8
Generous/Disallowed (%)
0.0/0.0
Average B-values
92.0
Est. Coordinate Error (Å)
0.47
a
Values as defined in SCALEPACK (Otwinowski and Minor, 1997).
Values as defined in CNS (Brunger, AT)
b
Envelope Protein and the Flavivirus virion
X-ray crystal structure of E
DI
DIII
DII
Immature
Cryo-EM model of WNV
Mature
5
prM Cleavage
3
60 trimers of prM/E heterodimers
2
180 E monomers
E16 is a potent neutralizing mAb with therapeutic activity against WNV in mice
Single Dose mAb at Day 5 Post-Infection
Humanized E16 binds WNV DIII with similar affinities and kinetics as E16
RU
45
40
DIII binding E16
35
30
25
Summary of Surface Plasmon Resonance (SPR) studies
20
15
Response
10
5
0
-5
-10
-50
0
50
100
150
200
250
300
350
Time
RU
30
400
s
DIII binding Hm-E16.3
25
20
15
10
Response
5
0
-5
-10
-50
0
50
100
150
200
Time
250
300
350
400
s
Antibody
E16
Hm-E16.1
Hm-E16.2
Hm-E16.3
ka (1/Ms)
1.1 x 106
9.6 x 105
1.0 x 106
9.9 x 105
kD (1/s)
0.0118
0.0201
0.0092
0.0070
Rmax
39.5
32.8
24.7
24.1
KD (nM)
10.8
21.0
9.2
7.1
Chi2
0.33
0.16
0.13
0.16
Production and purification of DIII in complex with E16 Fab
Bacterial expression
of WNV E Domain 3
Refolding of DIII
Complex purification by
size exclusion chromatography
Abs280 (mAU)
DIII
DIII-E16 Fab
complex
DIII alone
Elution Volume (ml)
Hybridoma
expression
of E16 mAb
mAb capture
by Protein A
E16 Fab
by papain
cleavage
Structure determination of DIII-E16 complex by X-ray crystallography
Structure of the DIII-E16 Fab complex
E16 Fab
CH
CL
VH
VL
VH
VL
DIII
DIII
E16 Fab
L3
DE Loop
L1
H2
BC Loop
H1
H3
N-terminal
region
L2
FG Loop
Nybakken et al, Nature 2005
Selection of E16 specific epitope variants of DIII
Yeast library of DIII variants
created by error prone PCR
E -DIII
Pooled
DIII
mAbs
E16 staining
DIII mutations at Ser306, Lys307, ThrE330 and Thr332 significantly diminish E16 binding
DIII yeast display mutations are centrally located at the E16 interface
E16 Fab
H1
TrpH33
CL
CH
H3
SerH95
VH
VL
LysE307
DIII N-Term
SerE306
DIII
H3
ThrE332
SerE306
AspH100
H2 ArgH58
LysE307
ThrE330
DIII N-Term
LysE307
ThrE332
ThrE330
DIII
DIII BC loop
E16 Fab
C
CL
C
1A1D-2 Fab
C
CL
C
CH1
VH
N
WNVE
DIII
VL
FG
VH
N
N
DE BC
DE
DV2E
DIII
N
VH
VL
N
N
BC
N
FG
AB
WNVE
DII
N
CL
C
CH1
CH1
VL
C
E53 Fab
Fusion loop
C
C
Thr 332
Thr 330
Ser 306
Lys 306
Lys 305
Arg 99
Lys 310
Gly 106
Lys 307
Yeast display
≤ 4.5 Å contacts
Thr 76
Pro 75
Leu 107
E16 Fab could potentially bind 120/180 E protein DIII sites on WNV
Fusion Loop
E16 binding to 2- and 3-fold clustered DIIIs appears permissive
while 5-fold clustered DIII binding appears sterically non-permissive
Zhang, et al, Nat Structural Biology, 2003
Mukhopadhyay, et al, Science, 2003
Combination of crystallographic and cryo-EM data - the E16 Fab/WNV complex
Model of E16 Fab
complex with WNV
Cross-section of Cryo-EM reconstruction
Cryo-EM work done in collaboration with
Rossmann and Kuhn groups at Purdue University
Cryo-EM reconstruction of
E16 Fab complex with WNV
Fitting E16 Fab complex into CryoEM reconstruction of WNV
E16 internalizes with the virus during infection of vero cells
Pre bind virus + Alexa-Ab
Add to cells at 4 or 37oC
E53
E16
15 minutes. Fix, add
Lyso-tracker
Confocal microscopy
DIC / Bright
Field
Fluorescent
Merge
Alexa 488-labeled WNV mAbs and lysotracker red (acidified endosomes)
E16 Fab decoration appears to trap WNV particles - a fusion intermediate?
pH 8
nucleocapsid core (~ 154Å)
outer lipid layer (~200 Å )
outer glycoprotein layer (~245Å)
pH 6
n ucleocapsid core (~ 158Å)
o uter lipid layer (~205Å)
outer density layer (~340Å)
When things go wrong:
Chang G, Roth CB. (2001) Structure of MsbA from E. coli: a homolog of the multidrug resistance ATP binding cassette (ABC) transporters. Science 293(5536):1793-800. PMID 11546864
Pornillos O, Chen YJ, Chen AP, Chang G. (2005) X-ray structure of the EmrE multidrug transporter in complex with a substrate. Science 310(5756):1950-3. PMID 16373573
Reyes CL, Chang G. (2005) Structure of the ABC transporter MsbA in complex with ADP.vanadate and lipopolysaccharide. Science 308(5724):1028-31. PMID 15890884
Chang G. (2003). Structure of MsbA from Vibrio cholera: a multidrug resistance ABC transporter homolog in a closed conformation. J Mol Biol 330(2):419-30. PMID 12823979
Ma C, Chang G. (2004). Structure of the multidrug resistance efflux transporter EmrE from Escherichia coli. Proc Natl Acad Sci USA 101(9):2852-7. PMID 14970332
When things go wrong:
Figure 2. Tertiary structure of the N-domain of STAT-4. (A) Overall representation of two monomers (green and gray) in the crystallographic dimer, viewed approximately orthogonal to the molecular twofold axis, which is vertical. The ring-shaped NH2terminal element is colored red in one monomer. (B) Orthogonal view of one of the N-domains shown in (A), depicting details of the architecture of the ring-shaped element. Side chains that participate in a charge-stabilized hydrogen-bond network are
shown in a ball-and-stick representation. The side chain and backbone carbonyl of buried R31 are shown in magenta. For clarity, the indole ring of the invariant residue W4 that seals off this arrangement on the proximal side is drawn with thinner bonds.
The blue sphere denotes a buried water molecule. Hydrogen bonds are indicated by dotted lines. Oxygen, nitrogen, and carbon atoms are red, blue, and yellow, respectively. Q3-N marks the position of the backbone amide group of residue Q3. The lightred segment of helix 2 highlights its 310 helical conformation. Fig. 2 and Fig. 3, B and C were created with the program RIBBONS, version 2.0 (28).
Structure of the Amino-Terminal Protein Interaction Domain of STAT-4
Uwe Vinkemeier, Ismail Moarefi, * James E. Darnell Jr., John Kuriyan
Science 13 February 1998:Vol. 279. no. 5353, pp. 1048 - 1052
Figure 1. Analysis of STAT4 dimers produced by crystallographic symmetry to identify the physiologic dimer.(a) Dimer A (produced by the
fractional transformation -Y, -X, -Z+1/6 with translation 1, 1, 1) represents the dimer implied previously22. Dimer B (produced by the
fractional transformation X, X−, -Z+5/6 with translation 0, 1, 0) represents an alternative interface recently suggested25. Highlighted
residues were targeted for mutational studies. Residues W37, T40, and E66 (magenta) are located in the dimer A interface, whereas
residues D19 and L78 (cyan) are located in the dimer B interface. (b) Surface analysis of the two dimers. According to this analysis, dimer B
is a statistically better molecular interface (as compared to dimer A) and is more likely to represent a physiologically relevant dimer.
Naruhisa Ota, Tom J Brett, Theresa L Murphy, Daved H Fremont & Kenneth M Murphy
N-domain-dependent nonphosphorylated STAT4 dimers required for cytokine-driven activation
Nature Immunology 5, 208 - 215 (2004)