Empirical estimation of the energetic contribution of individual interface residues in structures of protein-protein complexes

Abstract

We report a simple algorithm to scan interfaces in protein-protein complexes for identifying binding 'hot spots'. The change in side-chain solvent accessible area (ΔASA) of interface residues has been related to change in binding energy due to mutating interface residues to Ala (ΔΔGX→ALA) based on two criteria - hydrogen bonding across the interface and location in the interface core - both of which are major determinants in specific, high-affinity binding. These relationships are used to predict the energetic contribution of individual interface residues. The predictions are tested against 462 experimental X → ALA mutations from 28 interfaces with an average unsigned error of 1.04 kcal/mol. More than 80% of interface hot spots (with experimental ΔΔG ≥ 2 kcal/mol) could be identified as being energetically important. From the experimental values, Asp, Lys, Tyr and Trp are found to contribute most of the binding energy, burying >45 Å2 on average. The method described here would be useful to understand and interfere with protein interactions by assessing the energetic importance of individual interface residues.