Correlations between Bond, Backbone, and Side Chain Dihedral Angles Enable Changes Among Different Dipeptide Configurations

Abstract

Despite recent improvements in computational methods for protein design, we still lack a fundamental understanding of protein structure and interactions even at the scale of individual amino acids. We therefore perform Langevin dynamics simulations of model dipeptides that include only hard-sphere and geometric constraints to gain a predictive understanding of the distributions of backbone and side chain dihedral angles in proteins of known structure. We study alanine, isoleucine, leucine, and phenylalanine dipeptides and identify correlations between bond angles, backbone, and sidechain dihedral angles as the dipeptides change from one backbone configuration to another or between different side-chain rotamers. We also compare our results for the backbone and side chain dihedral angle distributions to those obtained from the latest versions of the CHARMM and AMBER force fields. Finally, we use our hard-sphere dipeptide models to assess the change in stability of staphylococcal nuclease following a number of point mutations. These studies serve as a first step in developing the ability to quantitatively rank the energies of designed protein constructs.