Systematic phylogenetic analyses were conducted on 110 genes/proteins involved in dissimilatory sulfur cycling processes, including sulfate reduction, sulfur oxidation, sulfur disproportionation, and organosulfur respiration. Monophyletic clades of functional orthologs were identified based on bootstrap support, biochemical verification, gene neighborhood patterns, and conserved catalytic residues. Hidden Markov Models (HMMs) were developed for these clades, optimizing both sensitivity and specificity for homology searches. These HMMs were then applied to screen the Genome Taxonomy Database (GTDB), leading to the discovery of previously unknown sulfur-cycling genes in uncultured microbial lineages. This work established phylogenetic frameworks for 110 sulfur-cycling protein families, identified 174 monophyletic clades, and expanded the known phylogenetic diversity of sulfur-cycling microorganisms by approximately 40%. A web-based database was created to compile the phylogenetic frameworks, clade-specific HMMs, and associated descriptions.
We further reconstructed the evolution of the sulfur oxidation (Sox) pathway and explored ancient biological sulfur oxidation mechanisms. Using the newly developed HMMs, genes involved in Sox, rDsr, and sHdr pathways were screened across GTDB genomes. Phylogenomic analyses of Sox system components and associated pathways were conducted, and their evolutionary development was inferred by reconciling gene trees with a dated bacterial tree of life. A machine learning approach was applied to differentiate ecological and phylogenetic drivers in the evolution of the Sox pathway. These analyses traced the origin of a truncated Sox system before the Great Oxidation Event (GOE, ~2.35 billion years ago), which later expanded to include complete thiosulfate oxidation pathways following the GOE. This expansion occurred through the acquisition of additional Sox components (SoxCD) or reverse dissimilatory sulfite reductases (rDsr).
Dating results also revealed an unresolved puzzle regarding sulfide oxidation in the Archean era, where known oxidants (e.g. oxygen and nitrate) were seemingly absent. A hypothesis was proposed that iron(III) oxides, abundant in anoxic Archean environments, may have served as electron acceptors for microbial sulfide oxidation. We validated this hypothesis by physiological experiments with a microorganism that encodes the genetic capacity to to couple sulfide oxidation with ferrihydrite reduction. Transcriptomic data suggested that this process involved a reversal of the dissimilatory sulfate reduction pathway. The electrons yielded by this reaction may have been transferred to ferrihydrite via extracellular electron transfer mechanisms, potentially mediated by multi-heme cytochromes. These findings revealed a previously unknowm microbial metabolism and support a geological scenario in which iron(III) oxides could have facilitated sulfur oxidation in anoxic Archean environments, providing new insights into ancient sulfur cycling processes on early Earth.
