Local electrostatics governs tryptophan oxidation and enables rational stability engineering in antibodies.

Oxidation of tryptophan (Trp) residues in therapeutic antibody complementarity-determining regions (CDRs) can impair binding affinity, stability, and developability, yet the molecular determinants of site-specific susceptibility remain incompletely understood. Here, we present a structure-informed machine learning framework trained on Trp oxidation profiles from 187 monoclonal antibodies to identify the physicochemical drivers of oxidation risk. Beyond solvent accessibility, the dominant known predictor, we identify local residue-level electrostatic potential (Epot) as a strong independent modulator: negatively charged microenvironments markedly increase oxidation susceptibility. A two-parameter model combining only solvent accessibility and Epot achieves 79% classification accuracy, approaching the 84% performance of full-feature ensemble models, and correctly classifies all 10 CDR Trp sites in a blind validation panel of eight clinical-stage IgG1 antibodies. Leveraging the long-range nature of electrostatic effects, we show that targeted distal charge-altering mutations, without direct modification of the oxidation-prone Trp, reduce oxidation rates by approximately 50% in four of five re-engineered antibodies, with binding affinity preserved in two. In a clinically relevant anti-CD33 antibody where the critical CDR Trp oxidizes at 97% and Trp-to-Phe substitution abolishes binding, iterative electrostatic optimization yielded variants with up to ~50% oxidation reduction, including one single-point-mutation variant achieving 27% oxidation reduction at only 1.4-fold affinity cost, substantially outperforming direct Trp substitution. Redox replica-exchange molecular dynamics simulations provide mechanistic support, revealing an 88 mV increase in Trp reduction potential upon distal charge-reversing mutations. Together, these results establish local electrostatics as a predictive, tunable, and mechanistically grounded handle for rational antibody engineering, and provide a practical framework for simultaneously optimizing oxidation stability and antigen-binding function.