Computational screening reveals class-wide CYP19A1 binding potential across per- and polyfluoroalkyl substances

Per- and polyfluoroalkyl substances (PFAS) are ubiquitous forever chemicals linked to reproductive toxicity, yet their interactions with CYP19A1 (aromatase)-the rate-limiting enzyme converting androgens to estrogens-remain poorly characterized. To address this gap, we developed a pharmacophore-based virtual screening framework to systematically assess CYP19A1 binding potential across 10,946 the United States Environmental Protection Agency (EPA) -cataloged PFAS. From the PFOS-CYP19A1 docking complex, we generated 10 pharmacophore models, of which three high-performance models were selected through rigorous validation (AUC > 0.848, MCC 0.257-0.294). Subsequent virtual screening revealed that 90.1 % of PFAS (9861 compounds) exhibited measurable CYP19A1 binding potential, with 7078 high-confidence compounds matching all three models. Furthermore, molecular dynamics simulations demonstrated that PFHxS maintained superior binding stability compared to PFOA, attributable to sulfonate groups forming stronger hydrogen bonds than carboxylates. Notably, cellular thermal shift assays experimentally confirmed direct PFHxS-CYP19A1 binding, demonstrating significant thermal stabilization. Collectively, these findings reveal that CYP19A1 binding represents a class-wide PFAS property rather than being confined to legacy compounds, thereby challenging current chemical-by-chemical regulatory paradigms and providing mechanistic support for group-based PFAS regulation in reproductive toxicology risk assessment.