Decoding Allosteric Communication Pathways in Cyclophilin A with a Comparative Analysis of Perturbed Conformational Ensembles.

Conformational dynamics plays the key role in allosteric regulation of enzymes. Despite numerous experimental and computational efforts, the mechanism of how dynamics couple enzymatic function is poorly understood. Here, we introduce a new approach to exploring the dynamics-function relationship combining computational mutagenesis, microsecond-long molecular dynamics simulations, and side-chain torsion angle analyses. We apply our approach to elucidate the allosteric mechanism in cyclophilin A (CypA), a peptidyl-prolyl cis-trans isomerase known to participate in diverse biological processes and be associated with many diseases including cancer. Multiple single mutations are performed in CypA at previously discovered hotspot residues distal from the active site, and residues displaying significant dynamical changes upon mutations are then identified. The mutation-responsive residues delineate three distinct pathways potentially mediating allosteric communications between distal sites: two pathways resemble the allosteric networks identified in a recent experimental study, whereas the third represents a novel pathway. A residue-residue contact analysis is also performed to complement the findings. Furthermore, a recently developed difference contact network analysis is employed to explain mutation-specific allosteric effects. Our results suggest that comparing multiple conformational ensembles generated under various mutational conditions is a powerful tool to gain novel insights into enzymatic functions that are difficult to obtain through examining a single system such as the wild-type. Our approach is easy to extend for other systems. The results can also be utilized to facilitate the design of potent therapeutics targeting CypA.

Enhanced molecular dynamics sampling of drug target conformations.

Computational docking and virtual screening are two main important methods employed in structure-based drug design. Unlike the traditional approach that allows docking of a flexible ligand against a handful of receptor structures, receptor flexibility has now been appreciated and increasingly incorporated in computer-aided docking. Using a diverse set of receptor conformations increases the chances of finding potential drugs and inhibitors. Molecular dynamics (MD) is greatly useful to generate various receptor conformations. However, the diversity of the structures of the receptor, which is usually much larger than the ligand, depends on the sampling efficiency of MD. Enhanced sampling methods based on accelerated molecular dynamics (aMD) can alleviate the sampling limitation of conventional MD and aid in representation of the phase space to a much greater extent. RaMD-db, a variant of aMD that applies boost potential to the rotatable dihedrals and non-bonded diffusive degrees of freedom has been proven to reproduce the equilibrium properties more accurately and efficiently than aMD. Here, we discuss recent advances in the aMD methodology and review the applicability of RaMD-db as an enhanced sampling method. RaMD-db is shown to be able to generate a broad distribution of structures of a drug target, Cyclophilin A. These structures that have never been observed previously in very long conventional MD can be further used for structure-based computer-aided drug discovery, and docking, and thus, in the identification and design of potential novel inhibitors.