Genome mining strategies for ribosomally synthesised and post-translationally modified peptides

Abstract

Genome mining is a computational method for the automatic detection and annotation of biosynthetic gene clusters (BGCs) from genomic data. This approach has been increasingly utilised in natural product (NP) discovery due to the large amount of sequencing data that is now available. Ribosomally synthesised and post-translationally modified peptides (RiPPs) are a class of structurally complex NP with diverse bioactivities. RiPPs have recently been shown to occupy a much larger expanse of genomic and chemical space than previously appreciated, indicating that annotation of RiPP BGCs in genomes may have been overlooked in the past. This review provides an overview of the genome mining tools that have been specifically developed to aid in the discovery of RiPP BGCs, which have been built from an increasing knowledgebase of RiPP structures and biosynthesis. Given these recent advances, the application of targeted genome mining has great potential to accelerate the discovery of important molecules such as antimicrobial and anticancer agents whilst increasing our understanding about how these compounds are biosynthesised in nature.

Keywords: Antibiotic; BGC, biosynthetic gene cluster; Bioinformatics; Biosynthesis; DNN, deep neural network; Genome mining; HMM, hidden Markov model; MS, mass spectrometry; NP, natural product; Natural product; ORF, open reading frame; PTM, post-translational modification; RTE, RiPP tailoring enzyme; RiPP; RiPP, Ribosomally synthesised and post-translationally modified peptide.

Graphical abstract

Fig. 1

Schematic of RiPP biosynthesis.

Fig. 2

Examples of RiPP natural products. A. Structures of a thiopeptide (thiostrepton), a recently discovered antibiotic (darobactin), a redox cofactor (pyrroloquinoline quinone, PQQ) and a thioviridamide-like molecule (prethioviridamide). B. Precursor peptides corresponding to these RiPPs, where core peptides are coloured red.

Fig. 3

Examples of RiPPs whose discovery was guided by the use of genome mining tools. The compound name, class and tool are listed alongside each structure.

Fig. 4

Overview of RiPP mining results for Streptomyces scabies 87–22. A. Genetic details of all RiPP BGCs identified by one or more tools. B. Summary of predictions made by each tool for a given RiPP BGC. Regions highlighted in red relate to predicted core peptides. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

Fig. 5

Summary of structural predictions provided for lanthipeptide BGCs by antiSMASH, RiPPMiner, PRISM and DeepRiPP. A. Summary of predictions (note: both PRISM and DeepRiPP predict multiple possible RiPP products and only the first prediction is visualised here). B. Structures of two characterised lanthipeptides whose BGCs have homology to BGC1 and BGC2.