Effective Riemannian diffusion model for conformational dynamics of membrane transporters
Mol. dynamics (MD) simulation technique offers a dynamic picture of mol. processes with an unparalleled spatiotemporal resoln. However, functionally relevant conformational changes of proteins often occur on timescales far beyond those currently accessible to conventional MD. For instance, membrane transporters undergo large-scale conformational changes between inward- and outward-facing structures to transport mols. across the membrane, a process which takes tens of milliseconds to seconds per cycle. These conformational changes are almost entirely unknown due to the limitations assocd. with both computational and exptl. techniques. Employing a novel computational approach based on ideas from nonequil. statistical mechanics and Riemannian geometry and taking advantage of petascale computing, we have reconstructed the entire transport process of a transporter at an at. level within a multiscale framework. The underlying assumption is the existence of a low-dimensional manifold on which lie most of the relevant conformations visited during the functionally relevant structural transitions. The effective dynamics of the system in this low-dimensional space is modeled as Riemannian diffusion. This study provides a detailed description of the transport mechanism in a membrane transporter and, more generally, introduces a robust framework for the study of structure-function relationships in proteins.