Lu_2025_Biotechnol.J_20_e70061

(R)-Styrene oxide is a high-value chiral intermediate in pharmaceutical and chemical industries, yet its enantioselective synthesis remains challenging. Here, we engineered an epoxide hydrolase from Spatholobus suberectus (SsEH) to address its limitations in catalytic activity and thermostability. Through a computational strategy integrating homology modeling, molecular dynamics (MD) simulations, and machine learning, we rationally designed a mutagenesis library and identified the quintuple variant SsEH-His41Arg-Thr71Val-Lys117Leu-Leu187Phe-Ser244Ala (SsEH-M5). This variant exhibited a 17.0-fold increase in catalytic activity and a 2.1-fold improvement in thermostability (half-life at 35 degreesC) compared to wild-type SsEH. Structural analysis revealed that enhanced activity stemmed from optimized substrate binding and nucleophilic attack efficiency, while additional hydrogen bonds (Arg41-Tyr216-Asp212-His38) stabilized the enzyme's architecture. In a 3500 L bioreactor, SsEH-M5 catalyzed the enantioconvergent hydrolysis of 60 g/L racemic styrene oxide using 2.5 g/L (DCW, dry cell weight) whole-cell biocatalyst, yielding (R)-styrene oxide with >99.5% enantiomeric excess (ee) and (R)-1-phenyl-1,2-ethanediol (>96.0% ee). This work highlights the synergy of computational design and experimental validation in developing robust biocatalysts for industrial-scale chiral synthesis.