A. Kolinski et J. Skolnick, MONTE-CARLO SIMULATIONS OF PROTEIN-FOLDING .2. APPLICATION TO PROTEIN-A, ROP, AND CRAMBIN, Proteins, 18(4), 1994, pp. 353-366
The hierarchy of lattice Monte Carlo models described in the accompany
ing paper (Kolinski, A., Skolnick, J. Monte Carlo simulations of prote
in folding. I. Lattice model and interaction scheme. Proteins 18:338-3
52, 1994) is applied to the simulation of protein folding and the pred
iction of 3-dimensional structure. Using sequence information alone, t
hree proteins have been successfully folded: the B domain of staphyloc
occal protein A, a 120 residue, monomeric version of ROP dimer, and cr
ambin. Starting from a random expanded conformation, the model protein
s fold along relatively well-defined folding pathways. These involve a
collection of early intermediates, which are followed by the final (a
nd rate-determining) transition from compact intermediates closely res
embling the molten globule state to the native-like state. The predict
ed structures are rather unique, with native-like packing of the side
chains. The accuracy of the predicted native conformations is better t
han those obtained in previous folding simulations. The best (but by n
o means atypical) folds of protein A have a coordinate rms of 2.25 ang
strom from the native Calpha trace, and the best coordinate rms from c
rambin is 3.18 angstrom. For ROP monomer, the lowest coordinate rms fr
om equivalent Calphas of ROP dimer is 3.65 angstrom. Thus, for two sim
ple helical proteins and a small alpha/beta protein, the ability to pr
edict protein structure from sequence has been demonstrated. (C) 1994
Wiley-Liss, Inc.