An inventory of early branch points in microbial phosphonate biosynthesis

Microbial phosphonate biosynthetic machinery has been identified in ~5 % of bacterial genomes and encodes natural products like fosfomycin as well as cell surface decorations. Almost all biological phosphonates originate from the rearrangement of phosphoenolpyruvate (PEP) to phosphonopyruvate (PnPy) catalysed by PEP mutase (Ppm), and PnPy is often converted to phosphonoacetaldehyde (PnAA) by PnPy decarboxylase (Ppd). Seven enzymes are known or likely to act on either PnPy or PnAA as early branch points en route to diverse biosynthetic outcomes, and these enzymes may be broadly classified into three reaction types: hydride transfer, aminotransfer, and carbon-carbon bond formation. However, the relative abundance of these branch points in microbial phosphonate biosynthesis is unknown. Also unknown is the proportion of ppm-containing gene neighbourhoods encoding new branch point enzymes and potentially novel phosphonates. In this study we computationally sorted 434 ppm-containing gene neighbourhoods based on these seven branch point enzymes. Unsurprisingly, the majority (56 %) of these pathways encode for production of the common naturally occurring compound 2-aminoethylphosphonate (AEP) or a hydroxylated derivative. The next most abundant genetically encoded intermediates were phosphonoalanine (PnAla, 9.2 %), 2-hydroxyethylphosphonate (HEP, 8.5 %), and phosphonoacetate (PnAc, 6 %). Significantly, about 13 % of the gene neighbourhoods could not be assigned to any of the seven branch points and may encode novel phosphonates. Sequence similarity network analysis revealed families of unusual gene neighbourhoods including possible production of phosphonoacrylate and phosphonofructose, the apparent biosynthetic use of the C-P lyase operon, and a virus-encoded phosphonate. Overall, these results highlight the utility of branch point inventories to identify novel gene neighbourhoods and guide future phosphonate discovery efforts.