RAPID REFINEMENT OF PROTEIN INTERFACES INCORPORATING SOLVATION - APPLICATION TO THE DOCKING PROBLEM

A computationally tractable strategy has been developed to refine prot ein-protein interfaces that models the effects of side-chain conformat ional change, solvation and limited rigid-body movement of the subunit s. The proteins are described at the atomic level by a multiple copy r epresentation of side-chains modelled according to a rotamer library o n a fixed peptide backbone. The surrounding solvent environment is des cribed by ''soft'' sphere Langevin dipoles for water that interact wit h the protein via electrostatic, van der Waals and field-dependent hyd rophobic terms. Energy refinement is based on a two-step process in wh ich (1) a probability-based conformational matrix of the protein side- chains is refined iteratively by a mean field method. A side-chain int eracts with the protein backbone and the probability-weighted average of the surrounding protein side-chains and solvent molecules. The resu ltant protein,conformations then undergo (2) rigid-body energy minimiz ation to relax the protein interface. Steps (1) and (2) are repeated u ntil convergence of the interaction energy. The influence of refinemen t on side-chain conformation starting from unbound conformations found improvement in the RMSD of side-chains in the interface of protease-i nhibitor complexes, and shows that the method leads to an improvement in interface geometry. In terms of discriminating between docked struc tures, the refinement was applied to two classes of protein-protein co mplex: five protease-protein inhibitor and four antibody-antigen compl exes. A large number of putative docked complexes have already been ge nerated for the test systems using our rigid-body docking program, FTD OCK. They include geometries that closely resemble the crystal complex , and therefore act as a test for the refinement procedure. In the pro tease-inhibitors, geometries that resemble the crystal complex are ran ked in the top four solutions for four out of five systems when solvat ion is included in the energy function, against a background of betwee n 26 and 364 complexes in the data set. The results for the antibody-a ntigen complexes are not as encoura ging with only two of the four sys tems showing discrimination. It would appear that these results reflec t the somewhat different binding mechanism dominant in the two types o f protein-protein complex. Binding in the protease-inhibitors appears to be ''lock and key'' in nature. The fixed backbone and mobile side-c hain representation provide a good model for binding. Movements in the backbone geometry of antigens on binding represent an ''induced-fit'' and provides more of a challenge for the model. Given the limitations of the conformational sampling the ability of the energy function to discriminate between native and non-native states is encouraging. Deve lopment of the approach to include greater conformational sampling cou ld lead to a more general solution to the protein docking problem. (C) 1998 Academic Press Limited.