In this study, we sought to determine whether an in vivo assay for studying antibiotic mechanisms of action could provide insight into the activity of compounds that may inhibit multiple targets. Thus, we conducted an activity screen of 31 structural analogs of rhodanine-containing pan-assay interference compounds (PAINS). We identified nine active molecules against Escherichia coli and classified them according to their in vivo mechanisms of action. The mechanisms of action of PAINS are generally difficult to identify due to their promiscuity. However, we leveraged bacterial cytological profiling, a fluorescence microscopy technique, to study these complex mechanisms. Ultimately, we found that although some of our molecules promiscuously inhibit multiple cellular pathways, a few molecules specifically inhibit DNA replication despite structural similarity to related PAINS. A genetic analysis of resistant mutants revealed thymidylate kinase (essential for DNA synthesis) as an intracellular target of some of these rhodanine-containing antibiotics. This finding was supported by in vitro activity assays, as well as experiments utilizing a thymidylate kinase overexpression system. The analog that demonstrated the half-maximal inhibitory concentration in vitro and MIC in vivo displayed the greatest specificity for inhibition of the DNA replication pathway, despite containing a rhodamine moiety. Although it is thought that PAINS cannot be developed as antibiotics, this work showcases novel inhibitors of E. coli thymidylate kinase. Moreover, perhaps more importantly, this work highlights the utility of bacterial cytological profiling for studying the in vivo specificity of antibiotics and demonstrates that bacterial cytological profiling can identify multiple pathways that are inhibited by an individual molecule. IMPORTANCE We demonstrate that bacterial cytological profiling is a powerful tool for directing antibiotic discovery efforts because it can be used to determine the specificity of an antibiotic's in vivo mechanism of action. By assaying analogs of PAINS, molecules that are notoriously intractable and nonspecific, we (surprisingly) identify molecules with specific activity against E. coli thymidylate kinase. This suggests that structural modifications to PAINS can confer stronger inhibition by targeting a specific cellular pathway. While in vitro inhibition assays are susceptible to false-positive results (especially from PAINS), bacterial cytological profiling provides the resolution to identify molecules with specific in vivo activity.
Keywords: E. coli; PAINS; SAR; antibiotic; bacterial cytological profiling; mechanism.