Nanoliter-scale selection of optimized bioengineered peptide antibiotics that rescue mice with bacterial lung infection

Abstract

Increasing numbers of multi-drug resistant pathogens call for new chemical scaffolds, addressing novel targets, that can serve as lead structures for the development of life-saving drugs. For antibiotics, natural product-inspired molecules represent a most promising resource. Natural products evolved to high chemical complexity and occupy a chemical space different than synthetic libraries. However, clinical translation of promising natural products is often impeded by their relative inaccessibility to medicinal chemistry optimization, e.g. iterative synthesis of large series of derivatives. Here, this limitation is addressed with a randomized library of bicyclic heptapeptides based on the natural product darobactin that hits the clinically not addressed target BamA. Variants of the ribosomally synthesized and post-translationally modified peptides were generated using heterologous mutasynthesis. A parallelized screening assay is adapted in nanoliter-scale beads to test the darobactin derivatives against our sensor strain. Loss of fluorescence sorting prioritized 563 events out of the analyzed [~]500k beads. Re-testing confirmed 48 hit events, of which 40 proved to produce distinct darobactin-type molecules. Most promising structures were isolated and the growth inhibitory effects against Gram-negative pathogens validated. One of our current frontrunner compounds (i.e., darobactin B) was reinforced by the randomized screen. While microbiological investigations of the new derivatives is ongoing, darobactin B was profiled in later tier assays and compared to another promising, rationally-designed analog (i.e., darobactin B9, "D22"). Early ADMET profiling and efficacy tests in a mouse pneumonia model were performed. Darobactin B reduced bacterial load of Pseudomonas aeruginosa and Klebsiella pneumoniae by intraperitoneal, as well as intratracheal administration. Our study showcases the potential of mutasynthetic libraries for high-throughput screening and identification of functional peptides for drug lead discovery.