Flexibly Screening for Molecules Interacting with Proteins
Volker Schnecke, Leslie A. Kuhn · Kluwer Academic Publishers eBooks · 2005
The flexibility of proteins and their ligands (molecules specifically bound by proteins) has a major influence on the ways they interact. Generally, protein molecules are thought of as primarily rigid structures, with chemically specific and somewhat flexible side chains attached to a main chain of fixed structure. The tendency to think of proteins as rigid is reinforced by the fact that X-ray crystallography, the most widely-used technique for analyzing protein structures at atomic resolution, traps the copies of a protein molecule in the crystalline lattice into a single state. However, as seen in Figure 1, rotatable single bonds in the main as well as side chains of a protein provide significant potential for flexibility (conformational change). For a number of proteins, such as the HIV protease, lysine-arginine-ornithine binding protein, and adenylate kinase, the protein is known to undergo significant conformational change upon binding its natural ligand or drugs designed to inhibit its activity. Flexibility is thus a biologically essential feature of proteins. Despite the importance and widespread interest in characterizing protein flexibility, this remains a challenge both experimentally and computationally (see papers by David Case, Ruben Abagyan, and Mark Gerstein in this volume). Our laboratory’s goal has been to develop computational methods that incorporate realistic modeling of protein flexibility into the design of new ligands for proteins. In collaboration with Jacobs and Thorpe (see accompanying paper), we have shown that graph-theoretic analysis of the covalent and hydrogen-bond networks in proteins using the FIRST algorithm provides an extremely fast way of assessing large-scale flexibility in proteins, e.g., when large, independently folded regions of the protein (domains) are attached by hinge joints, resulting in clamshell-like motion. In the present paper, we review the state of the art in template-based algorithms for analyzing protein–