Structure prediction and synthesis of a new class of macrocyclic peptide natural products

Owing to advances in genomic sequencing and bioinformatics, the breadth of natural product biosynthetic gene clusters (BGCs) has meteorically risen. This remains true for ribosomally synthesized and post-translationally modified peptides (RiPPs), where the rate of bioinformatically identifying clusters vastly outpaces characterization efforts. Uniting bioinformatics and enzymological knowledge to predict the chemical product(s) of a RiPP BGC with total chemical synthesis to obtain the natural compound is an effective platform for investigating cryptic gene clusters. Herein, we report the bioinformatic identification of a biosynthetically divergent class of RiPP bearing a subset of enzymes involved in thiopeptide biosynthesis. These natural products were predicted based on BGC architecture to undergo a formal, enzymatic [4+2]-cycloaddition with subsequent elimination of the leader peptide and water to produce a tri-substituted pyridine-based macrocycle. Bearing a pyridine similar to thiopeptides but lacking the ubiquitous thiazole heterocycles, these new RiPPs were termed pyritides. One of the predicted natural products was chemically synthesized using an 11-step synthesis. This structure was verified to be chemically identical by an orthogonal chemoenzymatic synthesis utilizing the precursor peptide and the cognate [4+2]-cycloaddition enzyme. The chemoenzymatic platform was used to synthesize a second in-cluster pyritide product as well as analogs from other bioinformatically identified pyritide BGCs. This work exemplifies complementary bioinformatic, enzymological, and synthetic techniques to characterize a structurally distinct class of RiPP natural product.