Broad-spectrum antibiotics can harm helpful microbes, sometimes worsening gut inflammation in conditions like Crohn’s disease. MIT and McMaster researchers developed enterololin, a targeted antibiotic that suppresses harmful Escherichia coli linked to flare-ups without disturbing the overall gut microbiome. Using DiffDock, an AI-based molecular docking model, they accelerated identification of the drug’s mechanism—binding to the LolCDE protein complex involved in bacterial lipoprotein transport—reducing the typical multi-year process to months.
The team validated DiffDock’s predictions experimentally by evolving enterololin-resistant E. coli strains and analyzing gene expression changes, confirming the target and mechanism. The AI-guided approach demonstrates a shift from purely chemical space searches to providing mechanistic insights critical for drug development. Enterololin is in early development with a startup optimizing it for human use and exploring derivatives against other resistant pathogens.
If successful in clinical trials, this targeted antibiotic could reduce Crohn’s symptoms without microbiome disruption and help combat antimicrobial resistance. The study illustrates the potential for AI to transform drug discovery by enhancing speed, reducing costs, and improving precision. Continued collaboration of AI, lab experiments, and human insight could accelerate development of new therapies for complex diseases.
👉 Pročitaj original: MIT AI News
