Use of SAXS in ubiquitin conjugation research Ubiquitin conjugation - - PDF document

Use of SAXS in ubiquitin conjugation research

Like phosphorylation, ubiquitination changes the fate of the target protein Ubiquitin is a 76-amino acid protein Targeted to amino groups (Lys, N-term) Ubiquitin can make chains

Ubiquitin conjugation is a signalling system

Ubiquitin

Figure from Hochstrasser Nature 2009

Regulates many essential pathways Potentially interesting drug targets E1 E3 ligases Dubs

Ubiquitin conjugation is a signalling system

Ubiquitin

Figure from Hochstrasser Nature 2009

Three stories E3 ligase RNF8 Dub USP7 Ubiquitinated target PCNA

Ubiquitin conjugation and SAXS

Ubiquitin

H2A ubiquitination

• Polycomb repressive

complex 1 (Ring1B/A, Bmi1 and Mel18)

• Rnf8 and Rnf168

Transcriptional silencing DNA damage response

UB

H2A

UB UB UB UB

H2A

Crystal Structure of the RING domain of RNF8

Mattiroli et al, Cell 2012

Dimerization in Class 1 RING/Ubox dimers

N-terminal Helix RING C-terminal helix

Members of Class 1 RING dimers

Brca1/Bard1 Ring1b/Bmi1 Traf6 Traf2 Trim37 Ubox dimer CHIP

Dimer contacts primarily through four-helix bundle from flanking helices, N- and C- terminal to RING.

Class 1 Class 1

Hibbert et al., DNA repair, 2009 Huang, Hibbert et al, J Mol Biol 2011

Brca1 is the breast cancer susceptibility gene 1 Brca1/Bard1 is a RING dimer Breast cancer-associated point mutations in Brca1 are found in the dimer interface with Bard1 Only BRCA1 is active as E3 ligase

Klevit lab: Brzovic et al, NSB 2001/ JBC 2001

Dimerization in Brca1/Bard1 is important

Buchwald et al, EMBO J, 2006 Flora Groothuizen, Francesca Mattiroli

Asymmetry is common in class 1 RING dimers

Mutation of E2/E3 interface in Ring1b/Bmi1

• in Ring1b causes loss of activity
• in Bmi1 has no effect

Heterodimer Ring1b/Bmi1

• Part of Polycomb PRC1 complex
• Interface extended by N-terminal embracing arm

H2A-Ub H2A

Ring1B Bmi1 Ring1B I53A Bmi1 Ring1B Bmi1 L20A

15 30 60 120 15 30 60 120 15 30 60 120 mins TIME

• E3

Full length CHIP structure shows Class 1 dimerization fold

• Ubox has RING fold without Zinc ions

Flanking TPR domains are highly asymmetrically arranged

UBox domain Pearl Lab: Zhang et al, Mol Cell 2006

CHIP

Spectacular asymmetry in CHIP Ubox homodimer

CHIP is a chaperone protein with E3 ligase function collaborating with e.g HSP90

UBox domain Only one E2 binding site is exposed in the full length protein

CHIP

Pearl Lab: Zhang et al, Mol Cell 2006

CHIP asymmetry causes 2:1 stoichiometry with E2s

Full-length Rad18 is an asymmetric homodimer

2:1 His6 Tag Strep Tag Strep Tag Strep-Rad6 pull-down: brings down His-Rad18 but not His-Rad6

Anti-His

Coomassie

Asymmetry is conserved in Rad18

Huang, Hibbert et al, J. Mol Biol 2011

RING domains of Rad18

Asymmetry could be due to crystal packing

RNF8 is highly asymmetric in the crystal

Superposition of the RING 35o

Chain B Chain A

483 348

CHIP and Rnf8 monomers RING superimposed

The point of divergence is conserved between RNF8 and CHIP

Small angle X-ray scattering of RNF8 RING domain

0.0 0.1 0.2 0.3 0.4 0.01 0.1 1 10 100

scattering curve theoretical curve

s = 4πsin(θ)/λ log(I) measured value (expected value in brakets)

Five different concentrations (from 0.4 to 7 mg/ml) in gel-filtration buffer. Data were collected at EMBL Hamburg, data analyzed using ATSAS

Mattiroli et al, Cell 2012

Small angle X-ray scattering

Envelope from DAMAVER where 10 DAMMIF ab initio models were used. All 10 models showed an elongated shape and their NSD (normalized spatial discrepancy) values were between 0.6 and 0.74

Generated models for symmetric dimers Two times molecule A Two times molecule B Tried to fit these separately. The asymmetric dimer seems to fit best,

AA BB AB

Usp7/HAUSP

• DUB for Mdm2 and P53
• Regulates stability
• Decision making for apoptosis, cell cycle and

senescence

• DUB for PTEN and FOXO4
• Regulates cellular localization

All proteins critical in oncogenic pathways

USP7/HAUSP protein

Faesen et al, Mol Cell 2011

In U2OS cells

The C-terminal domain of USP7/HAUSP is important for activity

USP7

Against purified ubiquitinated p53

(Fernandez-Montalvan et al., 2007)

AMC USP7 AMC

+

The HUBL domain necessary for full USP7 activity

• n minimal substrate Ub-AMC

Crystallization of HUBL domain

Initial selenomethionine phased map

The HAUSP C-terminal domain has 5 Ubl domains

5 Ubl domains 2+1+2 structure USp7/Hausp Ubl domain: HUBL Faesen et al Mol Cell 2011

The di-Ubl units have similar relative arrangement

USP7CD HUBL-13 HUBL-45 Faesen et al Mol Cell 2011

USP7CD HUBL-13 HUBL-45 Faesen et al Mol Cell 2011

HUBL USP7CD-HUBL

14 nm 5.5 nm 8 nm 5 nm

HUBL domain

• elongated,
• long atom-atom distances

HUBL + catalytic domain

• long distances lost
• HUBL domain folds back onto CD

HUBL-45 binds the catalytic domain HUBL-45 activates in trans

Understanding the activation process

C-terminal 19 amino acid tail is important but not sufficien

• requires HUBL-45 for binding

HUBL-45 is sufficient for full activity

Zoom

Zoom

Point mutations block activation by HUBL-45

Over-expression in U2OS

Usp7 point mutations block activation by HUBL-45

Faesen et al, Mol Cell 2011

Model for HUBL activation of USP7

• Inactive state: HUBL-45 interaction
• Low ubiquitin affinity
• Disorganized active site
• ‘inactive’ switching loop
• Active state: Interaction with HUBL-45
• High affinity for ubiquitin
• Catalytically competent active site
• ‘Active’ switching loop

Model for HUBL activation of USP7

Questions:

• Is this equilibrium unique for USP7
• Most USPs don’t show inactive state for isolated catalytic domain
• UBP8 in the SAGA complex is activated

through Sgf11 interaction at switching loop

• Does the equilibrium exist in cells ?
• Are there proteins that regulate it ?
• Analyze the activation by GMP synthase

Van der Knaap et al (2005) Molecular Cell 17:695-707 van der Knaap (2010) Mol Cell Biol. 30:736-44 Sarkari et al, (2009), PloS Pathog. e1000624

USP7 and GMPS form a stable complex USP7 can be activated by GMPS Activation does not require GMPS catalytic activity

GMP synthase (GMPS) is an activator of USP7

Van der Knaap et al (2005) Molecular Cell 17:695-707 van der Knaap (2010) Mol Cell Biol. 30:736-44 Sarkari et al, (2009), PloS Pathog. e1000624

Drosophila USP7 and GMPS collaborate genetically Drosophila USP7 and GMPS form a stable complex GMPS activates USP7 against p53 and is absolutely required for H2b activity GMPS catalytic activity not involved Human USP7 and GMPS form complex and are active against H2b

GMP synthase (GMPS) is an activator of USP7

GMPS activates USP7

Ubiquitin affinity

• 5.5-fold increase in kcat,
• No effect on KM
• -No change in ubiquitin affinity

Substrate dependent effects likely

GMPS binds HUBL-13

SPR analysis of GMPS binding reveals

• HUBL-13 is required for binding of GMPS (Kd ~ 35 nM)
• Point mutants don’t interfere with binding
• Neither catalytic domain nor HUBL-45 has affinity for GMPS

GMPS activation requires a functional HUBL

Cys-223 Switching loop Activation peptide

GMPS binds to HUBL-13 GMPS requires HUBL-45 activation If GMPS shifts the equilibrium, GMPS should stabilize the interaction between CD and HUBL-45

USP7 switching model

Binding of HUBL to USP7CD is enhanced by GMPS

USP7 catalytic domain: Kd = 50 mM for HUBL Kd = 2 mM for HUBL in presence of GMPS

+

Conclusion:

• GMPS stabilizes interaction between HUBL-45 and catalytic domain

Model for activation of USP7

Conclusions:

• GMPS binds to HUBL-13
• GMPS stabilizes interaction between HUBL-45 and catalytic domain
• GMPS allosterically activates by stabilization of the active state

Are there other regulators

• Stabilizing the ‘on’ state
• Stabilizing the ‘off’ state

Analysis of a ubiquitinated target

PCNA promotes processivity of DNA polymerases in replication

Mono-ubiquitination of PCNA causes a switch from replicative to TLS polymerase

Methods of produce ubiquitinated PCNA

• Native PCNA-Ub

– PNAS 2005, 2006

• ‘Split’ PCNA

– Co-expression of PCNA (1-163) with Ubiquitin fused in-line with PCNA (164-261) – Nature SMB (2010)

• Intein PCNA

– Nature Chem Biol (2010)

• Click PCNA

– Incorporating unnatural amino acids – Chembiochem (2011)

How does PCNA change upon ubiquitination

• Crystal structure of ‘split’ PCNA

– Co-expression of PCNA (1-163) with Ubiquitin fused in-line with PCNA (164-261) – Ubiquitin buries its hydrophobic patch

• Saxs analysis of ‘split’ PCNA and intein-based link

– 70% ordered structure, 2 major states.

in vitro ubiquitination of PCNA with E2 enzyme UbcH5c

• No E3 required
• Fully specific for K164

Hibbert & Sixma, JBC 2012

Facilitates the study of USP1/UAF1 deconjugation

Gel filtration / MALS analysis

SAXS analysis

SAXS analysis

SAXS analysis

SAXS analysis

The ubiquitin on PCNA-Ub is flexible

NMR analysis

NMR analysis

10 20 30 40

Ub PCNA-Ub Residue number Linewidth (Hz)

NMR analysis

NMR analysis

Linewidths: Ub~20Hz, PCNA-Ub~30Hz, PCNA~60Hz

NMR analysis – Pol eta binding site on ubiquitin is accessible

Pol eta binding site Significant Chemical shifts (PCNA-Ub vs Ub)

Conclusions

SAXS assists in ubiquitin research

– Ubiquitin E3 ligase RNF8

• Confirm extended helix
• Could not distinguish between symmetric or

asymmetric states

– Deubiquitinating enzyme USP7

• In solution less extended HUBL domain
• HUBL domain folds back on catalytic domain

– Ubiquitin target PCNA

• Ubiquitin is flexible on the target

NKI

Tassos Perrakis Annette Dirac Joep Vissers Elisabetta Citterio Anitha Shanmugham Huib Ovaa

Rotterdam

Jurgen Marteijn Wim Vermeulen Jan van der Knaap Peter Verrijzer Embl-Hamburg Dmitry Svergun Beamline scientists EMBL, ESRF & SLS NKI Flora Groothuizen Dene Littler Judith Smit Danny Sahtoe

Netherlands Cancer Institute Division of Biochemistry Rad18

Rick Hibbert

Rnf8

Francesca Mattiroli

Pim van Dijk Pauline Ikpa Usp7/HAUSP

Alex Faesen

Funding: EU Rubicon, SPINE2complexes, Ubiregulators, KWF, ERC, NWO-CW