Computationally driven design of small-molecule reversible covalent inhibitors for targeted therapeutics

Abstract

Reversible covalent inhibitors (RCIs) represent a promising strategy for achieving durable and selective target engagement in oncological therapy. By forming transient yet sufficiently stable covalent adducts, RCIs sustain target engagement while avoiding the cumulative toxicity associated with irreversible inhibitors. However, their rational design remains challenging due to the complex interplay between binding affinity, covalent kinetics, and resistance evolution. This review presents an integrated computational framework for RCI discovery, combining cheminformatics and machine learning-based triage with structure-based docking, molecular dynamics, free energy perturbation (FEP), and hybrid quantum mechanics/molecular mechanics (QM/MM) approaches for modeling covalent reaction mechanisms. Potency, selectivity, and resistance profile are interpreted through thermodynamic and kinetic descriptors, ΔG_bind, K_i, ΔG‡, and K_cov. Importantly, this review addresses a critical gap in current drug design strategies by integrating thermodynamic, kinetic, and resistance-aware approaches into a unified and scalable paradigm for reversible covalent inhibitor discovery.