C8(Chinese Academy of Medical Sciences and Peking Union Medical College)

Staphylococcus aureus (S. aureus) is one of the most concerned Gram-positive bacteria due to its resistance to the commonly used antibiotics, methicillin. To address the threat of methicillin-resistant S. aureus (MRSA), new classes of antibiotics are highly needed and artificial intelligence (AI) could accelerate antibiotic discovery. Previously, we developed a new AI-based software for the design of small-molecule antibiotics, AutoMolDesigner and had not been applied to real-world antibiotic discovery. In this study, we carried out AutoMolDesigner-based virtual screening for novel anti-S. aureus compounds, and identified a hit compound B1 (MIC: 4 μg/mL) with 4-hydroxy-2,5-dihydrothiazole as the core scaffold. From 28 derivatives synthesized and tested, we discovered compound C8 with improved antibacterial potency for MRSA (MIC: 0.5 μg/mL) and no cytotoxicity to HepG-2 and HEK293 (CC₅₀ > 50 μM). Moreover, we performed Cheminformatics-based prediction to facilitate elucidation of molecular mechanism. As a result, inhibition of S. aureus DNA gyrase by compound C8 was computationally prioritized and experimentally validated (IC₅₀: 0.320 ± 0.089 μM). The following exploration of several derivatives and two enantiomers separated from C8 further confirmed the mechanism. Lastly, we carried out comprehensive biological evaluation of compound C8 and discovered that it did not induce resistance of MRSA, and could effectively treat Galleria mellonella (G. mellonella) larvae and murine infected by MRSA in vivo. In conclusion, we have discovered 4-hydroxy-2,5-dihydrothiazoles, represented by compound C8, as a new class of gyrase-targeted antibiotics for the treatment of MRSA infection, through AI-integrated methods for drug discovery.