Evolutionary history and origins of Dsr-mediated sulfur oxidation

Abstract

Microorganisms play vital roles in sulfur cycling through the oxidation of elemental sulfur and reduction of sulfite. These metabolisms are catalyzed by dissimilatory sulfite reductases (Dsr) functioning in either the reductive or reverse, oxidative direction. Dsr-mediated sulfite reduction is an ancient metabolism proposed to have fueled energy metabolism in some of Earth's earliest microorganisms, whereas sulfur oxidation is believed to have evolved later in association with the widespread availability of oxygen on Earth. Organisms are generally believed to carry out either the reductive or oxidative pathway, yet organisms from diverse phyla have been discovered with gene combinations that implicate them in both pathways. A comprehensive investigation into the metabolisms of these phyla regarding Dsr is currently lacking. Here, we selected one of these phyla, the metabolically versatile candidate phylum SAR324, to study the ecology and evolution of Dsr-mediated metabolism. We confirmed that diverse SAR324 encode genes associated with reductive Dsr, oxidative Dsr, or both. Comparative analyses with other Dsr-encoding bacterial and archaeal phyla revealed that organisms encoding both reductive and oxidative Dsr proteins are constrained to a few phyla. Further, DsrAB sequences from genomes belonging to these phyla are phylogenetically positioned at the interface between well-defined oxidative and reductive bacterial clades. The phylogenetic context and dsr gene content in these organisms points to an evolutionary transition event that ultimately gave way to oxidative Dsr-mediated metabolism. Together, this research suggests that SAR324 and other phyla with mixed dsr gene content are associated with the evolution and origins of Dsr-mediated sulfur oxidation.

Keywords: SAR324; dissimilatory sulfite reductase; metagenomics; sulfur oxidation.

Figure 1

Simplified schematic representation of sulfate reduction and Dsr-mediated sulfur transformations. A) the sulfate reduction pathway (to sulfide) is depicted, with key enzymes/complexes involved at each step. The reaction highlighted in the manuscript, the reduction of sulfite to sulfide via Dsr, is enclosed within the dashed box. B) Overview of the forward and reverse Dsr-mediated reactions. Enzymes typically involved in each reaction, as discussed in text, are indicated. Figure generated with BioRender.com.

Figure 2

Dsr is present in globally distributed SAR324. A) Geographic distribution of Dsr-encoding SAR324 genomes correlated to the ecosystem type they were retrieved from. Non-Dsr encoding SAR324 genomes are included on the map and marked by smaller solid circles. B) Phylogenetic tree, inferred with maximum likelihood, of 145 SAR324 genomes generated via concatenation of 16 ribosomal protein subunits. The background colors of text in the phylogeny denote SAR324 clades I-XI. Bootstrap support ≥90 is annotated with circles on the branches of the phylogeny. Yellow squares denote presence or absence of dsrAB genes in SAR324 genomes. Clades/subclades in which dsrAB was not found were collapsed. The phylogeny is rooted to genomes belonging to the phyla Aquificota.

Figure 3

Analysis of SAR324 dsr gene content reveals distinct groupings. Visualization of dsr genomic organization in 31 SAR324 genomes. Slanted lines indicate the end of a scaffold, whereas dark, unlabled boxes denote genes encoding for proteins which were not identified in our HMM search. SAR324 genome identifiers are included to the left of each structure. In some cases, duplicate dsr genes were found in the same genome and omitted from the figure. Full gene content information is provided (Supplementary Table 5). Figure generated with BioRender.com.

Figure 4

DsrAB phylogenetic tree reveals select genomes that bridge the reductive and oxidative types. A) Concatenated DsrAB phylogenetic tree inferred with maximum likelihood. Branches are labeled with the genome identifier followed by the phylum, separated by an underscore. Multiple DsrAB copies within the same genome are indicated by “_2” or “_3” after the genome identifier. Dark yellow squares indicate genomes also encoding DsrD, blue squares denote those encoding DsrE, and red squares indicate those encoding DsrL. SOMs, SRMs, and SAR324 clades are highlighted in various colors. The phylogeny was rooted to archaeal (Thermoproteota) genomes. B) Zoomed in portion of the phylogenetic tree annotated with nodes at which various evolutionary events, as described in text, occurred. Bootstrap support ≥90 is annotated with circles on branches of the phylogenetic tree. C) Schematic of the proposed evolutionary scenario from which three different groups of bacteria with varied gene content and metabolic capabilities arose from an evolutionary transition event. Nodes, labeled on the figure, correspond to nodes annotated in panel B. All proteins/complexes are shown as monomers for simplicity. Figure generated with BioRender.com.

Figure 5

DsrL phylogenetic trees demonstrates genomic clustering based on dsr genome content. A) Circular DsrL phylogenetic tree inferred by maximum likelihood. The tree is annotated with the three distinct groups described (Fig. 4C). The phylogeny was rooted using the iTOL default root. Bootstrap support ≥90 is indicated with circles on the branches of the phylogeny B) the same phylogeny as in panel a but presented as an unrooted tree. The unrooted DsrL phylogenetic tree highlights the clustering of different phyla. For both trees, branches are labeled with the genome identifier followed by the phylum, separated by an underscore. Multiple DsrL copies within the same genome are indicated by “_2” after the genome identifier.

Figure 6

Comparison of SAR324 16 ribosomal protein subunit (RP16) phylogeny to DsrL and DsrEFH phylogenies suggests horizontal gene transfer of dsrEFH. A) Comparison of SAR324 RP16 phylogeny to SAR324 DsrL phylogeny. B) Comparison of SAR324 RP16 phylogeny to SAR324 DsrEFH phylogeny. Phylogenetic trees were inferred with maximum likelihood. Bootstrap support ≥90 is shown with circles on the branches of the phylogeny. Trees were manually rooted to the same genomes for clarity.

Figure 7

Genomes from the transitionary phyla encode DsrR. DsrR phylogenetic tree inferred with maximum likelihood. Tree demonstrates organisms encoding DsrR are restricted to four phyla and dominated by Pseudomonadota. The number of genomes is shown in parentheses next to collapsed clades. Bootstrap support ≥90 is shown with circles on the branches of the phylogeny. The phylogeny is rooted to HesB protein sequences, which have homologous domains to DsrR (95).