Grant: Alliance Grants

The Erythromycin-Resistance Methyltransferase (Erm) is an enzyme involved in establishing bacterial resistance to antibiotics. Recent discovery on how the enzyme interacts with ribosomal RNA offers the possibility of selectively modulating Erm activity without targeting the catalytic site which is highly conserved in this family of enzymes. Hence, the discovery of molecules able to interfere the Erm-RNA interface opens new venues to develop chemical probes for studying the enzyme's function as well as to develop innovative compounds with future therapeutic potential.In the last 20 years, methods of virtual screening have significantly improved the possibility of efficiently discovering ligands for enzymatic sites from large chemical libraries. Unfortunately, these methods are ill-suited for unconventional targets such as protein-RNA/DNA and protein-protein interaction interfaces because they fail to recapitulate the unique physical chemistry dominating the interaction between small molecules and these sites. We propose to leverage machine learning to develop an improved, hybrid data/physics-driven virtual screening method suitable for identifying protein-RNA modulators, and to apply it to identify novel ligands for the Erm-RNA interface. The computationally predicted ligands will be experimentally assessed thanks to our collaboration with IIT Bombay.We envision that our proposal will lead to the development of innovative open-source tools for ligand discovery in Canada and worldwide and will accelerate the discovery of specific drug-like Erm inhibitors. Moreover, our tools will also be easily transferrable to other domains such as material discovery.