Constructing a useful tool for characterizing amino acid conformers by means of quantum chemical and graph theory indices

Abstract

The aim of this work is to construct a tool to assist in the prediction of peptidic properties resulting from the exchange of two amino acids in a proteic chain. In the past others have used experimental properties for this purpose. However, the nature of these data sets severely limits their access to important properties pertaining to secondary structure, and hence the indices used cannot characterize different backbone conformers like α helix and β strands, or side-chain conformations like gauche+, gauche− and trans. In this study we explore the importance of backbone and side-chain angles with regard to conformer similarity measured with theoretical properties calculated in an ab initio manner. For each of the 20 genetically encoded amino acids, we studied five conformers that correspond to α helical and β strand structures, with three different side chain conformations for each, defined solely by their angles Φ, Ψ and χ₁. This methodology allowed each of the 108 conformers to be represented by a mathematical object without ambiguity. The peptidic chain was emulated using two capping models to simulate the effect of nearest neighbors. These are OHCX_aaNH₂ and AlaX_aaAla, where X_aa is the conformer of interest. We then calculated 40 ab initio quantum chemical and graph theory indices for each backbone-side-chain conformer to obtain a characterization and classification scheme. We found that: (1) while backbone structure is very important to conformer similarity, side-chain conformations do not cluster together in a top-level manner; (2) amino acids with π electrons group together independent of backbone conformation.