Discovery of a triple-site inhibitor targeting bacterial methionyl-tRNA synthetase through combined drug repurposing screening and generative AI-assisted optimization

Methionyl-tRNA synthetase (MetRS) plays an critical role in protein translation by catalyzing the attachment of l-methionine (l-Met) to its cognate tRNAMet and has long been recognized as a valuable target for antimicrobial drug development. In this study, a drug repurposing screen of a kinase inhibitor library identified AZD8186, a clinically investigated PI3Kβ modulator, as a promising inhibitor of Staphylococcus aureus MetRS (SaMetRS). The binding mode of AZD8186 to SaMetRS was elucidated through co-crystallography, and subsequent knowledge-directed ligand optimization resulted in enhanced inhibitory activity and improved synthetic accessibility. Furthermore, we developed a novel conservation-aware and interaction-guided 3D generative AI model, designated DiffDeCIG, to facilitate structure-based drug design. DiffDeCIG modified inhibitors to establish additional interactions preferentially with conserved residues within the active pocket of SaMetRS. The optimal compound, MRS-9, potentially competed with all three substrates of MetRS (ATP, l-Met and tRNAMet), and demonstrated over a 300-fold increase in inhibitory activity relative to AZD8186. Importantly, MRS-9 selectively inhibited type 1 MetRS enzymes, while minimally affecting the tested type 2 MetRSs, including the human MetRS, thereby reducing potential adverse effects. This study reveals a novel triple-site inhibitory mechanism targeting MetRS and highlights an integrated strategy that combines knowledge-directed and AI-guided approaches in drug design.