Large-scale transcriptome mining enables macrocyclic diversification and improved bioactivity of the stephanotic acid scaffold

Nearly 10,000 plant species are represented by RNA-seq datasets in the NCBI sequence read archive, which are difficult to search in unassembled format due to database size. Here, we optimize RNA-seq assembly to transform most of this public RNA-seq data to a searchable database for biosynthetic gene discovery. We test our transcriptome mining pipeline towards the diversification of moroidins, which are plant ribosomally-synthesized and posttranslationally-modified peptides (RiPPs) biosynthesized from copper-dependent peptide cyclases. Moroidins are bicyclic compounds with a conserved stephanotic acid scaffold, which becomes cytotoxic to non-small cell lung adenocarcinoma cells with an additional C-terminal macrocycle. We discover moroidin analogs with second ring structures diversified at the crosslink and the non-crosslinked residues including a moroidin analog from water chickweed, which exhibits higher cytotoxicity against lung adenocarcinoma cells than moroidin. Our study expands stephanotic acid-type peptides to grasses, Lowiaceae, mints, pinks, and spurges while demonstrating that large-scale transcriptome mining can broaden the medicinal chemistry toolbox for chemical and biological exploration of eukaryotic RiPP lead structures.