β-Lactams have long been the antibiotics of choice in the fight against S. aureus infections, but resistance to these molecules has emerged, causing alarm bells. MRSA, or methicillin-resistant staphylococcus, has been a growing concern for a long time due to the so-called PBPs (Penicillin-Binding Proteins). β-Lactams are known to irreversibly acylate the active-site serine of PBPs, resulting in bacterial death. In contrast, PBP2a is refractory to inhibition by essentially all commercially available β-lactams. Below you see a generalized β-lactam structure and its PBP2 nemesis.
There have been many ways to approach the problem of creating new antibiotics, particularly non-β-lactam types (to circumvent resistance). Inevitably, these methods call for screening some large collections of molecules. I was intrigued by a paper in JACS published by Chang and co-workers from the University of Notre Dame. Looking for potential inhibitors, the authors screened 1.2 million compounds from the ZINC database against the X-ray structure of PBP2a of MRSA. The ZINC database was created by the Shoichet lab at the UCSF (Brian is now at the University of Toronto, and is also a member of the SAB of Encycle Therapeutics, a company I started in 2012). The complexes obtained using this method were scored using DOCK, Gold, FlexX, and ChemScore. Subsequently, 29 molecules were synthesized and/or purchased. The lead compound shown below was generated using a comparatively small-scale synthetic campaign, which is the main attractive feature here. This substituted oxadiazole is an exciting entity that offers a new avenue for exploring non-β-lactam inhibitors of PBPs. What is the main lesson here, you might ask? I think it is a clear and demonstrable promise of addressing important problems using modern docking algorithms. Papers such as this underscore the power of screening virtual collections.