Limitations of ligand-only approaches for predicting the reactivity of covalent inhibitors

Angus Voice, Gary Tresadern, Herman Van Vlijmen and Adrian J. Mulholland

J. Chem. Inf. Model. 2019
DOI:https://doi.org/10.1021/acs.jcim.9b00404

Abstract

Covalent inhibition has undergone a resurgence and is an important modern-day drug design and chemical biology approach. To avoid off-target interactions, and to fine tune reactivity, the ability to accurately predict reactivity is vitally important for the design and development of safer and more effective covalent drugs. Several ligand-only metrics have been proposed that promise quick and simple ways of determining covalent reactivity. In particular, we examine proton affinity and reaction energies calculated with the density functional B3LYP-D3/6-311+G**//B3LYP-D3/6-31G* method to assess the reactivity of a series of ,-unsaturated carbonyl compounds that form covalent adducts with cysteine. We demonstrate that, whilst these metrics correlate well with experiment for a diverse range of covalent fragments, these approaches fail for predicting the reactivity of drug-like compounds. We conclude that ligand-only metrics such as proton affinity and reaction energies do not capture determinants of reactivity in situ and fail to account for important factors such as conformation, solvation and intra-molecular interactions.