Towards pharmaceutical protein stabilization: DFT and statistical learning studies on non-enzymatic peptide hydrolysis degradation mechanisms

Abstract

Theoretical studies on amide hydrolysis are of interest due to their impact on the drug discovery and delivery sectors of the biopharmaceutical industry. For degradation in which biopharmaceuticals and biopolymers may degrade through multiple different mechanisms, reaction families are established through the proposed mechanisms. Due to the influence of non-bonded interactions upon reactivity, the electrostatic interactions between the biopolymer and its environment must also be quantified to obtain a robust understanding of the system. For this purpose, we conducted a study of representative model dipeptides to analyze the reaction paths of non-enzymatic amide hydrolysis to generate a self-consistent database of thermochemical and kinetic values over a range of dielectrics using Density Functional Theory and statistical thermodynamics, and was found to capture experimental trends. Additionally, our study generated statistically significant linear correlations to predict the activation energy of the gemdiol formation step of the stepwise mechanism for the overall data set, as well as a function of the modeled dielectric.