Conformation spaces of proteins and implications for protein folding.

摘要

The following three chapters characterize structural variance seen across protein conformational ensembles.; In chapter one, two approaches to measuring conformational space are taken. A method called N-cube analysis (NCA) is developed for assigning a gross size to a conformational ensemble. Using NCA, the conformational space size of villin headpiece subdomain's compact unfolded state is calculated to be 107 times larger than the native state's. In addition, Euclidean based methods for characterizing the behavior of conformational space surrounding any given structure are developed. In the limit of small displacements, the apparent number of model Euclidean dimensions is found to correspond to the true number of mechanical degrees of freedom for a particular ensemble.; In chapter two we show how protein dynamics of the scale associated with protein folding can be modeled as a diffusive search in a high dimensional space. The size of the space is determined using principles developed in Chapter 1. Our numerical approach permits simple comparison to molecular dynamics. Time dependent displacement in the diffusion model agrees with results from molecular dynamics over many orders of magnitude. We delimit the search space into ‘native’ and ‘unfolded’ regions and test different energy biases in protein folding simulations. Our results provide insight into plausible energy landscape models for protein folding.; Chapter three shows how crystal structures of HIV-1 protease, a well-studied drug target with many crystal structures containing different inhibitors and point mutations, cluster by crystal type, with apparent independence of inhibitor type and mutation. The many studies which try to explain drug resistance by structural perturbations smaller than those seen across space groups must be accepted with caution.