Methane, arsenic, selenium and the origins of the DMSO reductase family

Abstract

Mononuclear molybdoenzymes of the dimethyl sulfoxide reductase (DMSOR) family catalyze a number of reactions essential to the carbon, nitrogen, sulfur, arsenic, and selenium biogeochemical cycles. These enzymes are also ancient, with many lineages likely predating the divergence of the last universal common ancestor into the Bacteria and Archaea domains. We have constructed rooted phylogenies for over 1,550 representatives of the DMSOR family using maximum likelihood methods to investigate the evolution of the arsenic biogeochemical cycle. The phylogenetic analysis provides compelling evidence that formylmethanofuran dehydrogenase B subunits, which catalyze the reduction of CO₂ to formate during hydrogenotrophic methanogenesis, constitutes the most ancient lineage. Our analysis also provides robust support for selenocysteine as the ancestral ligand for the Mo/W atom. Finally, we demonstrate that anaerobic arsenite oxidase and respiratory arsenate reductase catalytic subunits represent a more ancient lineage of DMSORs compared to aerobic arsenite oxidase catalytic subunits, which evolved from the assimilatory nitrate reductase lineage. This provides substantial support for an active arsenic biogeochemical cycle on the anoxic Archean Earth. Our work emphasizes that the use of chalcophilic elements as substrates as well as the Mo/W ligand in DMSORs has indelibly shaped the diversification of these enzymes through deep time.

Figure 1

Maximum likelihood phylogeny of 1,568 DMSOR family protein sequences. All sequences came from cultured organisms with sequenced genomes. The lineage associated with each clade is indicated in the figure. The scale bar refers to the number of amino acid substitutions per site. The bootstrap support for crucial nodes in our phylogeny is provided in text.

Figure 2

This is the same phylogeny provided in Fig. 1 with a shorthand code at each branch in the tree indicating the organism from whose genome the protein homolog was found, along with the predicted lineage of the putative DMSOR based off sequence homology to and similar operon organization with the query sequence. The key to each organism’s code is found in the Supplemental Information. The name used for each protein lineage is consistent with the rest of the text. The phylum level affiliation of each organism is indicated by the color of the text of each code name as described in the figure legend. The scale bar refers to the number of amino acid substitutions per site. Bootstrap support for all nodes is denoted in text at each respective node.

Figure 3

Sub-pruned portion of the maximum likelihood phylogeny showing the FwdB/FmdB, FdhG, and various formate dehydrogenase lineages. The clade for each lineage is indicated by the color of the branches. The text labels for each protein representative in the family is color coded by whether the Mo or W atom is coordinated by a Cys or a Sec residue. The scale bar refers to the number of amino acid substitutions per site. Bootstrap support for all nodes ≥ 60 is denoted in text at the respective node.

Figure 4

Sub-pruned portion of the maximum likelihood phylogeny showing the ArxA/ArrA lineage. The putative ArxA homologs are identified both by the text label and a lighter shade of red for the branches comprising the clade. The phylum level affiliation of each organism is indicated by the color of the text of each code name as described in the figure legend. The scale bar refers to the number of amino acid substitutions per site. Bootstrap support for all nodes ≥ 60 is denoted in text at the respective node.

Figure 5

Inferred time of divergence of lineages from the DMSOR family based on our phylogenetic analysis as well as consideration of the biogeochemistry of DMSOR substrates and the physiological role (e.g., function in aerobic respiration and O₂-dependence) of DMSOR lineages positioned over the geologic time-scale. The arsenic-containing Tumbiana stromatolites are found at ~ 2.72 Ga..