Methods for Detecting Protein Binding Interfaces

Abstract

Protein molecules often come together in complexes in order to achieve their biological functions in the living cell. Since the three-dimensional structure and the functionality of proteins are closely related to each other, characterizing the structural and dynamical properties of protein complexes through experiments or computational modeling is important for understanding their roles in the basic biology of organisms. Certain specific regions of a protein may play a critical role in its structural, dynamical, and functional properties. A protein molecule binds to another protein or to a drug molecule through a specific site on its surface, which is commonly known as the binding interface. Prediction of binding interfaces can assist in drug design, protein engineering, protein function elucidation, molecular docking, and analyzing the networks of protein-protein interactions. Experimental detection of binding interfaces can provide a wealth of information, but is time consuming and sometimes inaccurate. Computational methods can validate and complement experimental studies in a cost-efficient way. In this chapter we present a short survey of computational methods that have been suggested over the past two decades for the detection of proteinprotein and protein-drug binding interfaces, focusing on methods that use specific amino acids as determinants of binding interfaces. Later, we describe our work in using evolutionary conservation and structural features to detect binding interfaces in proteins and guide protein-protein docking.