A systems-guided approach to discover the intracellular target of a novel evolution-drug lead

Elucidating intracellular drug targets has been a difficult problem. While machine learning analysis of omics data has been a promising approach, going from large-scale trends to specific targets remains a challenge. Here, we developed a systems-guided hierarchic workflow that utilizes metabolic and structural analysis to narrow in on specific targets suggested by statistical and machine learning analysis of metabolomics data. Utilizing a novel multi-valent DHFR-targeting antibiotic compound, CD15-3, as a case study, we first measured global metabolomics and applied statistics and machine learning to locate broad areas of metabolic perturbation under antibiotic stress. We then tested the ability of suggested compounds to rescue growth and performed metabolic modelling to identify pathways whose inhibition was consistent with growth rescue patterns. Next, we utilized protein structural similarity to further prioritize candidate drug targets within these pathways. Overexpression and in vitro activity assays of a top candidate target, HPPK (folK), showed complete recovery from drug induced growth inhibition and with microscopy. As interest in white-box machine learning methods continues to grow, this study demonstrates how established machine learning methods can be combined with mechanistic analyses to improve the resolution of drug target finding workflows.