Computational chemistry-based structure-guiding engineering of ω-transaminases to enhance catalytic activity and enantioselectivity toward bulky ketones.

Ubrogepant is a clinically approved migraine drug, and the precise enantioselective synthesis of Ubrogepant Chiral Intermediate (UCI) is critical for its efficacy. Thus, identifying an ω-transaminase (ω-TA) with high activity and enantioselectivity toward UCI is essential. This work adopted a strategy combining computation with experiment. First, an initial ω-TA catalyzing UCI was screened from an enzyme library, termed as wild-type (WT) hereinafter. Molecular docking and dynamics simulations were used to decipher structure-activity relationships at catalytic sites. Subsequently, Rosetta-based rational enzyme redesign and structural analysis were conducted, yielding mutant M12 (W136L/W192A/G284A). By experimental assay, M12 significantly outperformed WT, with an 8.98-fold increase in Kcat and enantioselectivity (d.r.) elevated from 8:1 to >100:1. The reaction mechanism was further explored via quantum chemical cluster approach of the rate-determining step's free energy barrier, with Interaction Region Indicator (IRI) analysis elucidating steric effects on energy barrier and enantioselectivity at the catalytic sites. Furthermore, M12 was assayed for its optimal reaction conditions, and under optimal conditions, M12's performance was improved markedly, yielding >99 % conversion rate and a 4.43-fold increase of Kcat value relative to its baseline. Generally, this strategy was proved to be highly effective for engineering high-performance ω-TA, holding great potential for rational enzyme design.