ComputationalApproaches to GPCRStructure anf function

Christopher A Reynolds

Department of Biological SciencesUniversity of Essex

United KingdomEssex Campus, John Constable (original in Washington)

Acknowledgements

Bruck Tadesse (GPCR Modelling)Shabana Vohra (GPCR Modelling)Kevin Smith (docking)

David Poyner (Aston)James Barwell (Aston)Alex Conner (Warwick)Mike Woolley (Warwick)Debbie Hay (Aukland, NZ)Harriett Watkins (Aukland, NZ)

Graham Upton (Maths, Essex, Statistics)

BHF for funding (grant about to start)

Key publications: Vohra et al. (2013), Wooley et al. (2013), J. Roy. Soc. Interface

Introducing Class B GPCRs

CGRP (calcitonin gene-related peptide receptor)

Involved in vasodilationPotential target for cardiovascular drugs, migrane

Aim: understand structure, mechanism

Motivation: new drug target c.f. well-known class A GPCRs, e.g. Dopamine D2

100

90

80

70

60

50

40

30

20

10

0

Percentage Identity

Twilight zone(18 – 25% )

N-terminus

Helical domain

CGRP peptide

RAMP: receptor activity modifying protein

CLR(no function)

Functional CGRPR(CLR + RAMP)

Challenge: no structureno known motifs in commonalignment in ‘midnight zone’i.e. %ID 13%, below threshold

OtherClass BGPCRs(144Residues)

Blue – no effect oncAMP production

Orange / purple– no effect

Mutagenesis data: Class B GPCRs

CGRP(122 residues)

Class A profile (TM3)

Class B profile(TM3)

Helix alignment: novel method

Pair-wise alignments (1000s)

# o

f vote

s(s

cale

d 0

– 1

)

Alignment

Pro

duct

sco

re Pro

duct

sco

re

-8 -4 0 4 8 -8 -4 0 4 8 -8 -4 0 4 8 -8 -4 0 4 8

-8 -4 0 4 8alignment

-8 -4 0 4 8alignment

-8 -4 0 4 8alignment

The class A –class E (GCR1) – class B alignment

Potential for the best method for aligning remote homologues?Develop for class C GPCRs (e.g. mGluR5, Parkinson’s)

Systematic studies of all GPCR families (Good test cases welcome)

Red: only fits 0 alignment Blue: consistent with alignment Grey: inconclusive

Model verification: interpreting mutationsMultiple MD refined models:

2 inactive:Implicit membraneExplicit membrane

2 Active:G-protein boundPeptide bound(MD essential to get active models)

Residue function?Consistent with alignment?

DRY/YLHR...H...E

RKLHxxxN

disulfide

W4.50

FxxP

IxxL KxxK

P...W

EVxxxL

NPXXY/VAVLY

R2.43 – T6.37 ionic lock closedcf R3.50 – E6.30 open

Motifs common to class A, GCR1 and class B GPCRs

No motifs in common?

Experimental approach to ECL2 conformation

A1

T6

T288A203

TM

5

TM1TM2TM

4

(B) Verified docked pose from sequence analysis (correlated mutations)

C2-C7

(C) Built helical extension (mutagenesis); generated 100 conformations of extracellular loop 2 (go beyond sampling limits of MD).

ECL2Modeller DOPE score

Nu

mb

er o

f in

tera

ctio

ns

Computational approach to ECL2 conformation

(A) Docked CGRP1-7

(docking methodology under development, MRC funding)

Validation of ECL2 conformation?

Y277

Y278

Y287

R274

CGRP

A273L290

L291TM5

TM4

TM4 TM5

TM3

TM3 TM6

TM7

TM3

TM4

TM5TM2

TM6

TM1

T288A5

W283

D280R11

H10

I284

T6

S285

T4D3

CGRP

CGRP

(A)

(B)

(C)

TM3

TM4

H8

TM1

TM2

TM3

TM1

TM2

TM3

TM7TM6

TM4

TM5

ECL2:CLR (CGRP)CLR (AM)CRF1RGCGR

(D)

(E)

Interactionssuggest mechanisms for helix movement / activation

...to be explored further