Using genomic context to enable metabolic discovery in industrially relevant Clostridia

Abstract

Historically, the bulk of microbial biochemical discoveries were made by isolating proteins or other biological macromolecules that were responsible for a phenotype of interest, whether that was a biochemical reaction, regulatory response, or ability to synthesize a molecule of interest. These approaches, while successful, were methodologically constrained by a lack of biological DNA sequencing data. With the exponential increase in sequenced microbial genomes over the past few decades, it has become clear that the vast majority of predicted genes have no known function. This sudden glut of functionally uncharacterized genes may now allow “backward” biochemical approaches to the discovery of important microbial functions that start with predicted genes. The tendency in prokaryotes to physically co-localize sets of functionally-related genes into gene clusters within their genomes, provides an invaluable starting place for discovering new biochemical functions. This clustering provides each gene with a iii “genomic context,” loosely defined as the set of genes and regulatory elements located in, or chromosomally nearby the gene cluster in which the gene of interest occurs. Inspection of a gene of interest’s genomic context can therefore enable elucidation of its function. Theoretically, the functions of entire gene clusters can be revealed this way. Developing formal and informal approaches to exploit genomic context for to identify biochemically interesting reactions or processes would therefore greatly advance the study of microbial physiology. After providing an overview of the significance and previous uses of genomic context analysis in Chapter 1 of this Thesis, I detail the discovery of a set of microbial gene clusters (Mlp gene clusters, or MGCs) in Chapter 2. MGCs are mostly found in bacteria, but encode close homologs of the soluble coenzyme M methyltransferases, corrinoid- dependent enzymes that participate in methanogenesis in certain methanogenic archaea. In Chapter 2, I show that the overwhelming majority of MGCs are homologous to each other, suggesting a common function, or a small set of closely related functions. Analysis of the genomic contexts of more than 300 homologous MGCs may implicate these gene clusters in methyl-selenium metabolism. I observe that a methyltransferase from the industrially-used syngas fermenter Clostridium ljungdahlii can methylate selenide using methyl-cobalamin as a methyl donor, supporting this proposed function. While the bioinformatic and biochemical analyses in Chapter 2 implicate about 90% of MGCs in methyl-selenide metabolism, about 10% of MGCs are found in microbes with no known use for selenium in biosynthesis. I again use genomic context analysis to develop the hypothesis that these “selenium-independent” MGCs are involved in a methioninebiosynthetic pathway, most likely a methionine salvage pathway. In Chapter 3, I present evidence that the MGC from the industrial cellulose fermenter Hungateiclostridium (Clostridium) thermocellum encodes components of an unusual methionine biosynthetic pathway that uses a chemical logic distinct from known methionine biosynthetic pathways. Further, I use cell extract assays and sulfur source utilization growth studies to implicate the H. thermocellum MGC in a methionine salvage pathway rather than de novo methionine biosynthesis. Finally, in Chapter 4, I report the chemical structures of an elusive group of cellulose- binding natural products (yellow affinity substances or flavoaffinins) produced by H. thermocellum and other cellulolytic anaerobes. Flavoaffinins have been reported in the H. thermocellum literature at least as far back as the original isolation of the organism in 1953. These natural products had previously eluded isolation and structural characterization due to instability and low yields. Isolation of two of the flavoaffinins, followed by NMR and MS-based structural elucidation revealed a unusual hydroxy-diene- g-lactone core attached to attached by a methylene bridge to an indole head-group on one side, and by a polyene chain to an aryl group on the other. I also investigate the biosynthetic origins of this unusual natural product, identify a putative flavoaffinin biosynthetic gene cluster, and formulate a biosynthetic hypothesis to explain flavoaffinin production in H. thermocellum. In summary, the three projects detailed in this Thesis neatly illustrate various uses of genomic context analysis in studies of microbial metabolism, including exploring functionally uncharacterized gene clusters to developing testable hypotheses regarding natural product biosynthesis.