Abstract
Menaquinone (MK) in bacterial membrane is an attractive target for the development of novel therapeutic agents. Mining the untapped chemical diversity encoded by Gram-negative bacteria presents an opportunity to identify additional MK-binding antibiotics (MBAs). By MK-binding motif searching of bioinformatically predicted linear non-ribosomal peptides from 14,298 sequenced genomes of 45 underexplored Gram-negative bacterial genera, here we identify a novel MBA structural family, including silvmeb and pseudomeb, using structure prediction-guided chemical synthesis. Both MBAs show rapid bacteriolysis by MK-dependent membrane depolarization to achieve their potent activities against a panel of Gram-positive pathogens. Furthermore, both MBAs are proven to be effective against methicillin-resistant Staphylococcus aureus in a murine peritonitis-sepsis model. Our findings suggest that MBAs are a kind of structurally diverse and still underexplored antibacterial lipodepsipeptide class. The interrogation of underexplored bacterial taxa using synthetic bioinformatic natural product methods is an appealing strategy for discovering novel biomedically relevant agents to confront the crisis of antimicrobial resistance.