Protein Design Based on the Designability of Protein Structures

Chen Zeng, PI (GWU)

Chao Tang, Co-PI (NEC)

Ned Wingreen (Princeton)

Naigong Zhang (GWU)

Luhua Lai (PKU)

NSF Support:DMR-0094176, 0313129

Protein Folds (~1000! Why so small?)

Acylphosphatase

Viral DNA-binding domain

Ribosomal protein S6

Ferredoxin

HP Lattice Model

EHH

EHP

EPP

,

( )i j i j

i j

H E r r

2.3

1

0

HH

HP

PP

E

E

E

2HH HP PP

HH PP HP

E E E

E E E

H--Hydrophobic

P--Polar(Dill)

HydrophobicitySegregation

Enumeration (HP Lattice Model) H. Li, C. Tang, N. Wingreen (NEC)

Designability

NS=1

NS=5

NS=0

NS=2

NS=1

95% sequences havedegenerate G-states

Sequences (108) Structures (105)

Top Lattice Structures

…, mm

Off-Lattice m-State ModelPark & Levitt (1995)

Fit to 1PSV By 3-State Model

crms ~ 0.8 Ao

1PSV

Enumeration of Configurations (off-lattice HP model)

ai – accessible surface area of residue i

A= ai < Ac

H = hi a’i (a’i=ai/A)

Searching for New Folds

MAQTDVILCPDTHQKAVCLEKIR EYFDCGLPSAQWCVIKNALLTFRPNMDKILYVVACQGGARASLTLRGAWETRKLHCMQPIVYTTLNGLTMNSAKPLVVTIYSQYNVHLRFDDGFDSKLACVNPMRTYEIWLETFRMLKSPQCNYIAVRIHGRYLDFGSCVNKMYHGFDAALLTRQVLPSLTQIVLTAGNYIGRGPNIPCLDIGSEFIINCAQLVRENHWGVSGLRANLAGTRVNIMPCDEWSILSLMKIHFHDISAQVYTERPQMVKRLAFRATCNMRWDPSIVYTWQFGHLCVHEWMTVINEDSAPILCWHGGLMFGNVVEERPAADIMNWGLRCSLKELTILKNETVGGAPQWYIVHNQFNAKNQKDIETRYPMKSLVSCILHIKMLMKIHYDTFREWQVNSCKLDDVSMSKALLVPQWIVRCSYTPLKWPSTCNMRWDPSIVYTWQFGHLSLVTCAKVINEDSAPILCTRSGLRLTDVNYIPPAADIAQREMRCSLTNQLVDGHETVGGAPQWYFAKRLCRGADEYLIETRYPMKSLILSTAEEKLATDIHYDAGREWQVNSFELITSVETLDLLVPQWIVRCSHTFDIPQS

N=23, r=1.9 A, =0.4 A, 10,000 configurations, 4688 clusters

o o

Histogram of Ns

Energy gap vs. Ns

Design Procedure

II. Pick Top Folds

III. Sequence Design and Verification

I. Model Computation

motif

2nac 2bnh

Design Protocol