Protein crystallography in practice

MCB

15 Dec 2016

Daved H. Fremont fremont@pathology.wustl.edu

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 methodsEstablish 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 P214 M3 /ASU

diffraction >2.3Å14.4% Peg6K

NaCacodylate pH 7.0200mM CaCl2

Space Group P31213 M3 + 3 MCP-1/ASU

diffraction > 2.3Å18% Peg4K

NaAcetate pH 4.1100mM MgCl2

Space Group C22 M3 /ASU

diffraction >2.1Å18% Peg4K

Malic Acid/Imidazole pH 5.1

100mM CaCl2

Commercial screening kits available from http://www.hamptonresearch.com; http://www.emeraldbiostructures.com

Hanging Drop Sitting drop

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 co-factors, 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

a

b

c

α

β

γ

a

b

c

αβ γ

The unit cell

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 α=β=90o, γ=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 LaboratoryALS 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 α(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

Diffraction DataP1

SeMet-1P1

SeMet-2P1

SeMet-3P1

SeMet-4P21

Native

Wavelength (Å) 1.07813 0.97956 0.97945 0.94645 0.97945

Resolution (Å) 100-1.60 100-1.60 100-1.60 100-1.60 100-1.40

Number of sites 4 4 4 4 ---

Reflections measured (unique)

126 734(49 265)

146 273(56 587)

146 405(56 619)

149 455(57 749)

495 848(45 711)

Completeness overall (outer shell)

81.8(35.3)

93.9(80.4)

94.0(80.9)

95.8(94.8)

96.1(82.1)

I/σ (I) overall (outer shell)

22.5(7.9)

26.8(11.6)

21.9(9.8)

19.1(7.3)

23.0(3.0)

Rsym(%) overall (outer shell)

5.9(19.2)

5.8(17.0)

6.0(19.3)

6.5(28.6)

10.0(67.2)

Rcullis (20-1.6 Å) iso/ano ---/0.967 0.359/0.563 0.283/0.529 0.607/0.729 ---Phasing power iso/ano ---/0.818 4.56/3.10 6.19/3.30 0.819/2.31 ---

Resolution Number of

reflections/ Number ofprotein atoms/ Rcrystal/ Rms Deviations

Refinement Range (Å) completion solvent atoms Rfree (%) Bonds (Å) Angles (o)

P1-SeMet 20.0-1.60 (1.66-1.60)

29 053/ 97.1%(2750/95.7%))

1957/266 16.8/21.1(26.2/27.8)

0.010 1.9

P21-native 20.0-1.40 (1.45-1.40)

45 632/ 96.3%(3965/84.6%))

1957/244 17.8/20.9(26.0/26.3)

0.011 1.5

MAD phasing statistics for the AP-2 α-appendage

Electron density for the AP-2 α-appendage

Initial bones trace for the AP-2 α-appendage

Final trace for the AP-2 α-appendage

5. Build an atomic model using the electron density map

What does a good map look like?

Before computers, maps were contoured on stacked pieces of plexiglass. A “Richards box” was used to build the model.

half-silvered mirror

plexiglass stack

brass parts model

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 Overall shape of the molecule 3.5 Å Ca trace 3.0 Å Side chains 2.8 Å Carbonyl oxygens (bulges) 2.5 Å Side chain well resolved,

Peptide bond plane resolved 1.5 Å Holes in Phe, Tyr rings 0.8 Å Current limit for best protein

crystals

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 = Σ(|Fo-Fc|)/Σ(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?

Traub LM, Downs MA, Westrich JL, and Fremont DH: (1999) Crystal structure of theα-appendage of AP-2 reveals a recruitment platform for clathrin-coat assembly. Proc.Natl. Acad. Sci. U.S.A. 96:8907-8912.

6. Validation: Mapping of sequence conservation in AP-2 α-subunit appendages

Structure-Based Mutagenesis of the α-appendage

Traub LM, Downs MA, Westrich JL, and Fremont DH: (1999) Crystal structure of theα-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

Example: West Nile Virus

About 70 members, half of which are associatedwith 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

Structural proteins Non-structural proteins

5’UTR 3’UTR

Production of soluble E proteins and ectodomain fragments

Immunize mice with soluble E (25 µg x 3)

Fuse splenocytes with myeloma line

Large panels of flavivirus mAbs

Table 1. Summary of Data Collection and RefinementData Collection for West Nile Virus Envelopea

Space Group P41212Unit Cell (Å3) a=89.6 b=89.6 c=154.0Wavelength(Å) 0.90X-ray Source APS-BM 14Resolution(Å) (outer shell) 20-2.9 (3.08-2.90)Observations/Unique 14408/62790Completeness(%) 98.5 (99.5)Rsym(%) 5.7 (52.4)I/σ 16.9 (2.05)Refinement Statisticsb

Resolution(Å) (outer shell) 20-3.0 (3.19-3.00)Reflections Rwork/Rfree 11506/607#Protein Atoms/Solvent/Heterogen 3031/28/38Rwork overall(outer shell) (%) 26.2(35.6)Rfree overall(outer shell) (%) 30.8(34.1)Rmsd Bond lengths (Å)/angles(o) 0.008/1.6Rmsd Dihedral/Improper (o) 24.9/0.84Ramachandran plotMost Favored/Additional (%) 78.2/21.8Generous/Disallowed (%) 0.0/0.0Average B-values 92.0Est. Coordinate Error (Å) 0.47 aValues as defined in SCALEPACK (Otwinowski and Minor, 1997).

bValues as defined in CNS (Brunger, AT)

Structure Determination of WNV Envelope Protein

Envelope Protein and the Flavivirus virion

DIII

DI

DII

X-ray crystal structure of E

Immature

5

3 2prM Cleavage

Mature

60 trimers of prM/E heterodimers 180 E monomers

Cryo-EM model of WNV

1.E+00

1.E+01

1.E+02

1.E+03

1.E+04

1.E+05

1.E+06

1.E+07

1.E+08

1.E+09

3 3.5 4 4.5 5 5.5

PFU

/g o

f tis

sue

Day 9 - E16 Day 9 - PBS

0 5 10 15 20 25 300

20

40

60

80

100

E16 (2 mg)

E24 (2 mg)

PBS

Treatment Day 5

Days Post Infection

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

-10

-5

0

5

10

15

20

25

30

35

40

45

-50 0 50 100 150 200 250 300 350 400

RU

Response

sTime

DIII binding E16

-10

-5

0

5

10

15

20

25

30

-50 0 50 100 150 200 250 300 350 400Time s

Response

RU

DIII binding Hm-E16.3

Antibody ka (1/Ms) kD (1/s) Rmax KD (nM) Chi2

E16 1.1 x 106 0.0118 39.5 10.8 0.33Hm-E16.1 9.6 x 105 0.0201 32.8 21.0 0.16Hm-E16.2 1.0 x 106 0.0092 24.7 9.2 0.13Hm-E16.3 9.9 x 105 0.0070 24.1 7.1 0.16

Summary of Surface Plasmon Resonance (SPR) studies

Bacterial expression of WNV E Domain 3

E16 Fab by papain cleavage

mAb capture by Protein A

Elution Volume (ml)

Abs 2

80 (m

AU)

DIII-E16 Fabcomplex

DIII alone

Complex purification by size exclusion chromatography

Production and purification of DIII in complex with E16 Fab

Hybridoma expression of E16 mAb

Refolding of DIII

DIII

Structure determination of DIII-E16 complex by X-ray crystallography

Data collection for D3-E16 complex

Space Group P212121 Unit Cell (Å3) a=52.4 b=83.3 c=110.6 X-ray Source ALS Resolution(Å) (outer shell) 30-2.50 (2.59-2.50) Observations/Unique 59923/16985 Completeness(%) 97.6 (82.7) Rsym(%) 8.3 (30.6) I/ 11.3 (2.7) Atomic refinement statistics Rwork overall(outer shell) (%) 20.8(25.6) Rfree overall(outer shell) (%) 28.2(31.8) Ramachandran plot Most Favored/Additional (%) 87.5/11.9 Generous/Disallowed (%) 0.4/0.2

CH

VH

CL

VL

Structure of the DIII-E16 Fab complex

Nybakken et al, Nature 2005

DIII

DIIIE16 Fab

N-terminal region

BC Loop

DE Loop

FG Loop

L2

L1L3

H3

H2

H1

E16 Fab

VH VL

Selection of E16 specific epitope variants of DIII

Yeast library of DIII variantscreated by error prone PCR

DIII mutations at Ser306, Lys307, ThrE330 and Thr332 significantly diminish E16 binding

Pooled

DIII

mAbs

E16 staining

E -DIII

DIII yeast display mutations are centrally located at the E16 interface

DIII

ThrE332

ThrE330

LysE307

SerE306

CH

VH

CL

VL

DIII

E16 Fab

LysE307

SerE306

TrpH33

SerH95

H1 H3

ThrE332

ThrE330LysE307H2 ArgH58

AspH100

DIII N-Term

H3

DIII BC loop

DIII N-Term

WNVE DIII

E16 Fab C

C

NN

N

C

BCDEFG WNVE

DII

E53 Fab CL

CH1

VH VL

C

C

NN

1A1D-2 Fab

DV2E DIII N

C

BCDE FG

CL

CH1

VH

CL

CH1

VHVL

C

C

NN

Ser 306

Lys 306Thr 330

Thr 332

Leu 107

Gly 106

Arg 99

Pro 75Thr 76

Lys 305

Lys 307

Lys 310

Fusion loop

AB

Yeast display ≤ 4.5 Å contacts

VL

E16 Fab could potentially bind 120/180 E protein DIII sites on WNV

Zhang, et al, Nat Structural Biology, 2003Mukhopadhyay, et al, Science, 2003

Fusion Loop

E16 binding to 2- and 3-fold clustered DIIIs appears permissivewhile 5-fold clustered DIII binding appears sterically non-permissive

Combination of crystallographic and cryo-EM data - the E16 Fab/WNV complex

Cryo-EM reconstruction ofE16 Fab complex with WNV

Model of E16 Fab complex with WNV

Cryo-EM work done in collaboration with Rossmann and Kuhn groups at Purdue University

Fitting E16 Fab complex into Cryo-EM reconstruction of WNV

Cross-section of Cryo-EM reconstruction

E16 internalizes with the virus during infection of vero cells

E53 E16

DIC / Bright Field

FluorescentMerge

Alexa 488-labeled WNV mAbs and lysotracker red (acidified endosomes)

Pre bind virus + Alexa-Ab

Add to cells at 4 or 37oC

15 minutes. Fix, addLyso-trackerConfocal microscopy

E16 Fab decoration appears to trap WNV particles - a fusion intermediate?

pH 6pH 8

nucleocapsid core (~ 154Å) n ucleocapsid core (~ 158Å) outer lipid layer (~200 Å ) o uter lipid layer (~205Å) outer glycoprotein layer (~245Å) outer density layer (~340Å)

www.sciencemag.org SCIENCE VOL 314 22 DECEMBER 2006 1875

Aquaculture in Offshore Zones

THE EDITORIAL BY ROSAMOND NAYLOR,“Offshore aquaculture legislation” (8 Sept.,p. 1363), suggests that the motivation formoving aquaculture into the open ocean isthat “marine f ish farming near the shoreis limited by state regulations.” Althoughunworkable regulations may exist in a fewstates, in the larger scheme this is irrele-vant. Of the offshore aquaculture projectscurrently under way, none are occurring inthe U.S. Exclusive Economic Zone (EEZ);rather, they are happening in state waters.Even historically, only two aquacultureprojects have ever occurred in federalwaters (1).

Much of Naylor’s stated concern overoffshore aquaculture is based on historicalexperience with near-shore fish farms. Thisis in spite of years of more relevant offshore

operations that reveal little, if any, negativeimpact on the environment or local ecosys-tems (2, 3). Naylor criticizes the NationalOffshore Aquaculture Act of 2005 becauseit lacks specific environmental standards.Yet, she recommends California’s recentSustainable Oceans Act as a legislativemodel, although it is similarly silent, leavingthose details to rule-making in response tothe best available science.

Naylor criticizes the use of fishmeal asan aquaculture ingredient, ignoring the factthat industrial fisheries are well managedand would occur with or without aquacul-ture’s demand. Naylor ignores the higherefficiency of using fishmeal to feed fishcompared with its use in land-based live-stock operations (4). Also ignored is theinefficiency of using small pelagic fish inthe natural setting to feed predator fish (5).

Researchers and entrepreneurs currentlydeveloping the technologies needed for offshoreaquaculture share a vision of a well-managed

industry governed by regulations with a rationalbasis in the ecology of the oceans and the eco-nomic realities of the marketplace.

CLIFFORD A. GOUDEY

Massachusetts Institute of Technology, Cambridge, MA02139, USA.

References and Notes1. The SeaStead project a decade ago, four miles off

Massachusetts (see www.nmfs.noaa.gov/mb/sk/saltonstallken/enhancement.htm) and the recentOffshore Aquaculture Consortium experimental cageoperation 22 miles off Mississippi (see www.masgc.org/oac/).

2. See www.lib.noaa.gov/docaqua/reports_noaaresearch/hooarrprept.htm/.

3. See www.blackpearlsinc.com/PDF/hoarpi.pdf.4. See www.salmonoftheamericas.com/env_food.html.5. D. Pauly, V. Christensen, Nature 374, 255 (2002).

IN HER PROVOCATIVE EDITORIAL “OFFSHOREaquaculture legislation” (8 Sept., p. 1363),R. Naylor raises valid points regarding regu-lation of oceanic aquaculture, since it issure to grow in the future because of dwin-dling global fishery supplies. This growth is

LETTERS I BOOKS I POLICY FORUM I EDUCATION FORUM I PERSPECTIVES

1878

Generating new sciencein the classroom

How proteins connect

1880 1882

Mathematicalperspectives

LETTERSedited by Etta Kavanagh

Retraction

WE WISH TO RETRACT OUR RESEARCH ARTICLE “STRUCTURE OFMsbA from E. coli: A homolog of the multidrug resistance ATP bind-ing cassette (ABC) transporters” and both of our Reports “Structure ofthe ABC transporter MsbA in complex with ADP•vanadate andlipopolysaccharide” and “X-ray structure of the EmrE multidrug trans-porter in complex with a substrate” (1–3).

The recently reported structure of Sav1866 (4) indicated that ourMsbA structures (1, 2, 5) were incorrect in both the hand of the struc-ture and the topology. Thus, our biological interpretations based onthese inverted models for MsbA are invalid.

An in-house data reduction program introduced a change in sign foranomalous differences. This program, which was not part of a conven-tional data processing package, converted the anomalous pairs (I+ andI�) to (F� and F+), thereby introducing a sign change. As the diffrac-tion data collected for each set of MsbA crystals and for the EmrEcrystals were processed with the same program, the structures reportedin (1–3, 5, 6) had the wrong hand.

The error in the topology of the original MsbA structure was a con-sequence of the low resolution of the data as well as breaks in the elec-

tron density for the connecting loop regions. Unfortunately, the use ofthe multicopy refinement procedure still allowed us to obtain reason-able refinement values for the wrong structures.

The Protein Data Bank (PDB) files 1JSQ, 1PF4, and 1Z2R forMsbA and 1S7B and 2F2M for EmrE have been moved to the archiveof obsolete PDB entries. The MsbA and EmrE structures will berecalculated from the original data using the proper sign for the anom-alous differences, and the new C⇥ coordinates and structure factorswill be deposited.

We very sincerely regret the confusion that these papers havecaused and, in particular, subsequent research efforts that were unpro-ductive as a result of our original findings.

GEOFFREY CHANG, CHRISTOPHER B. ROTH,

CHRISTOPHER L. REYES, OWEN PORNILLOS,

YEN-JU CHEN, ANDY P. CHEN

Department of Molecular Biology, The Scripps Research Institute, La Jolla, CA 92037, USA.

References1. G. Chang, C. B. Roth, Science 293, 1793 (2001).2. C. L. Reyes, G. Chang, Science 308, 1028 (2005).3. O. Pornillos, Y.-J. Chen, A. P. Chen, G. Chang, Science 310, 1950 (2005).4. R. J. Dawson, K. P. Locher, Nature 443, 180 (2006).5. G. Chang, J. Mol. Biol. 330, 419 (2003).6. C. Ma, G. Chang, Proc. Natl. Acad. Sci. U.S.A. 101, 2852 (2004).

COMMENTARY

Published by AAAS

on

De

ce

mb

er

10

, 2

00

9

ww

w.s

cie

nce

ma

g.o

rgD

ow

nlo

ad

ed

fro

m

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 11546864Pornillos 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 16373573Reyes CL, Chang G. (2005) Structure of the ABC transporter MsbA in complex with ADP.vanadate and lipopolysaccharide. Science 308(5724):1028-31. PMID 15890884Chang 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 12823979Ma 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:

Structure of the Amino-Terminal Protein Interaction Domain of STAT-4Uwe Vinkemeier, Ismail Moarefi, * James E. Darnell Jr., John KuriyanScience 13 February 1998:Vol. 279. no. 5353, pp. 1048 - 1052

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 NH2-terminal 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 light-red 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).

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)