Mass spectrometry-guided genome mining of peptidic and glycosylated microbial natural products

Scientific progress in organic synthesis, biochemistry and biology and cures to many infectious diseases and cancer rely on discovery of microbial natural products and their biosynthetic pathways. ̀Omics' approaches such as genome mining have opened new opportunities for natural product discovery within the last decade as ̃90% of pathways in microbial genomes are uncharacterized in their products. Genome mining for natural product discovery can be defined as the connection of a natural product (chemotype) with its biosynthetic genes (genotype) by applied biosynthetic knowledge. Traditional genome mining approaches are in silico-guided approaches in which the isolation of a new natural product is guided by bioinformatic predictions from a target cryptic gene cluster. The problem of in silico-guided genome mining in natural product discovery is its low-throughput rate as only one pathway is characterized per experiment. In this dissertation, mass spectrometry (MS)-guided genome mining approaches are introduced which rapidly connect a natural product with its biosynthetic genes by matching de novo tandem MS structures of biosynthetic building blocks such as amino acids and sugars to metabolite structures predicted from microbial genomes. As MS guided genome mining starts at the chemotype level by e.g. liquid chromatography-tandem mass spectrometry analysis of a microbial extract and subsequently connects putative natural products with their gene clusters, it has the potential for automation. In Chapter 2, peptidogenomics is introduced as a MS-guided genome mining approach for characterization of ribosomal and nonribosomal peptide chemotypes and their corresponding genotypes. Peptidogenomics characterized ten new peptide chemo- and genotypes from Streptomyces cultures including lanthipeptides, lassopeptides, linaridins and lipopeptides. In Chapter 3, a combination of imaging mass spectrometry, tandem MS and genome mining characterized the biosynthetic pathway of the didemnin anti-cancer agents in the marine alpha-proteobacterium Tistrella mobilis. In Chapter 4, glycogenomics is introduced as a MS-guided genome mining approach to connect chemo- and genotypes of glycosylated natural products. Glycogenomics enabled the discovery of putative arenimycin B, a glycosylated aromatic polyketide from the marine actinobacterium Salinispora arenicola and its biosynthetic pathway. In Chapter 5, bioactivity-guided genome mining combined with genetic knockouts and glycogenomics characterized the biosynthetic gene cluster of the lomaiviticins anti cancer agents in the marine actinobacterium Salinispora tropic