Mining soil metagenomes to better understand the evolution of natural product structural diversity: pentangular polyphenols as a case study.

Abstract

Sequence-guided mining of metagenomic libraries provides a means of recovering specific natural product gene clusters of interest from the environment. In this study, we use ketosynthase gene (KS) PCR amplicon sequences (sequence tags) to explore the structural and biosynthetic diversities of pentangular polyphenols (PP). In phylogenetic analyses, eDNA-derived sequence tags often fall between closely related clades that are associated with gene clusters known to encode distinct chemotypes. We show that these common “intermediate” sequence tags are useful for guiding the discovery of not only novel bioactive metabolites but also collections of closely related gene clusters that can provide new insights into the evolution of natural product structural diversity. Gene clusters corresponding to two eDNA-derived KSβ sequence tags that reside between well-defined KSβ clades associated with the biosynthesis of (C24)-pradimicin and (C26)-xantholipin type metabolites were recovered from archived soil eDNA libraries. Heterologous expression of these gene clusters in Streptomyces albus led to the isolation of three new PPs (compounds 1–3). Calixanthomycin A (1) shows potent antiproliferative activity against HCT-116 cells, whereas arenimycins C (2) and D (3) display potent antibacterial activity. By comparing genotypes and chemotypes across all known PP gene clusters, we define four PP subfamilies, and also observe that the horizontal transfer of PP tailoring genes has likely been restricted to gene clusters that encode closely related chemical structures, suggesting that only a fraction of the “natural product-like” chemical space that can theoretically be encoded by these secondary metabolite tailoring genes has likely been sampled naturally.