Cofactor F 420 plays critical roles in primary and secondary metabolism in a range of bacteria and archaea as a low-potential hydride transfer agent. It mediates a variety of important redox transformations involved in bacterial persistence, antibiotic biosynthesis, pro-drug activation and methanogenesis. However, the biosynthetic pathway for F 420 has not been fully elucidated: neither the enzyme that generates the putative intermediate 2-phospho- l -lactate, nor the function of the FMN-binding C-terminal domain of the γ-glutamyl ligase (FbiB) in bacteria are known. Here we present the structure of the guanylyltransferase FbiD and show that, along with its archaeal homolog CofC, it accepts phosphoenolpyruvate, rather than 2-phospho- l -lactate, as the substrate, leading to the formation of the previously uncharacterized intermediate dehydro-F 420 -0. The C-terminal domain of FbiB then utilizes FMNH 2 to reduce dehydro-F 420 -0, which produces mature F 420 species when combined with the γ-glutamyl ligase activity of the N-terminal domain. These new insights have allowed the heterologous production of F 420 from a recombinant F 420 biosynthetic pathway in Escherichia coli .

F 420 is a low-potential redox cofactor that mediates the transformations of a wide range of complex organic compounds. Considered one of the rarest cofactors in biology, F 420 is best known for its role in methanogenesis and has only been chemically identified in two phyla to date, the Euryarchaeota and Actinobacteria. In this work, we show that this cofactor is more widely distributed than previously reported. We detected the genes encoding all five known F 420 biosynthesis enzymes (cofC, cofD, cofE, cofG and cofH) in at least 653 bacterial and 173 archaeal species, including members of the dominant soil phyla Proteobacteria, Chloroflexi and Firmicutes. Metagenome datamining validated that these genes were disproportionately abundant in aerated soils compared with other ecosystems. We confirmed through high-performance liquid chromatography analysis that aerobically grown stationary-phase cultures of three bacterial species, Paracoccus denitrificans, Oligotropha carboxidovorans and Thermomicrobium roseum, synthesized F 420 , with oligoglutamate sidechains of different lengths. To understand the evolution of F 420 biosynthesis, we also analyzed the distribution, phylogeny and genetic organization of the cof genes. Our data suggest that although the F o precursor to F 420 originated in methanogens, F 420 itself was first synthesized in an ancestral actinobacterium. F 420 biosynthesis genes were then disseminated horizontally to archaea and other bacteria. Together, our findings suggest that the cofactor is more significant in aerobic bacterial metabolism and soil ecosystem composition than previously thought. The cofactor may confer several competitive advantages for aerobic soil bacteria by mediating their central metabolic processes and broadening the range of organic compounds they can synthesize, detoxify and mineralize.