Elucidating Allosteric Communications in Proteins via Computational Methods

Cellular functions are primarily facilitated by biomolecular interactions with proteins, and ligand binding synchronizes the function of a protein to the requirements of its surroundings. Consequences of ligand binding to a protein may range from subtle perturbations in the side chain conformations in the vicinity of the binding region to large-scale global conformational changes. Coupling of a change in conformation with that in activity of a protein is traditionally referred to as allostery. In the recent years, however, the conventional allostery concept has been challenged to include perturbations in dynamics of a large number of proteins even in the absence of detectable changes in their backbone structure. Although it can evidently be suggested that binding produces a signal which can propagate to distant sites of a protein to achieve the observed conformational and/or dynamical perturbations, revealing a detailed mechanism of signal propagation is still an elusive task. In order to elucidate this mechanism, the following two questions demand to be answered: i) How do different regions of the protein respond? ii) How does the protein “sense” and transmit the local perturbation? The former question, being relatively easier to handle, has been tackled with Normal Mode Analysis (NMA), Elastic Network Models (ENMs), and statistical analyses of Monte Carlo (MC) and Molecular Dynamics (MD) simulation trajectories for the last ~30 years in the literature. The latter question, on the other hand, is currently a hot research topic in research community. Allosteric signals are generally suggested to propagate through “energy transport channels” (residue networks, or signaling pathways) formed by bonded and nonbonded contacts of residues, and experimental methods, such as double-mutant analysis and NMR relaxation methods, are used to identify residues participating to these intraprotein signaling pathways. For the last 10-15 years, there has been a tremendous interest in utilizing computational techniques to elucidate allostericity in proteins. While elastic network models and molecular simulations have continued to be resourceful methods, the most important novel contributions, presumably, have come from the graph theory, perturbation methods, and the statistical coupling method. In this chapter of Frontiers in Computational Chemistry, various computational techniques used to elucidate allosteric mechanisms in proteins are to be discussed with various examples.

Keywords: Conformational change, Communication pathway, Crystal structure, Database, Elastic network model, Frequency, Graph theory, Induced fit, Information theory, Ligand binding, Molecular dynamics, Monte carlo simulation, Perturbation, Population shift¸ principal component analysis¸ protein dynamics¸ residue network, Signal propagation, Statistical coupling analysis, Web-server.