Evidence for hydrogen oxidation and metabolic plasticity in widespread deep-sea sulfur-oxidizing bacteria

Abstract

Hydrothermal vents are a well-known source of energy that powers chemosynthesis in the deep sea. Recent work suggests that microbial chemosynthesis is also surprisingly pervasive throughout the dark oceans, serving as a significant CO(2) sink even at sites far removed from vents. Ammonia and sulfur have been identified as potential electron donors for this chemosynthesis, but they do not fully account for measured rates of dark primary production in the pelagic water column. Here we use metagenomic and metatranscriptomic analyses to show that deep-sea populations of the SUP05 group of uncultured sulfur-oxidizing Gammaproteobacteria, which are abundant in widespread and diverse marine environments, contain and highly express genes encoding group 1 Ni, Fe hydrogenase enzymes for H(2) oxidation. Reconstruction of near-complete genomes of two cooccurring SUP05 populations in hydrothermal plumes and deep waters of the Gulf of California enabled detailed population-specific metatranscriptomic analyses, revealing dynamic patterns of gene content and transcript abundance. SUP05 transcripts for genes involved in H(2) and sulfur oxidation are most abundant in hydrothermal plumes where these electron donors are enriched. In contrast, a second hydrogenase has more abundant transcripts in background deep-sea samples. Coupled with results from a bioenergetic model that suggest that H(2) oxidation can contribute significantly to the SUP05 energy budget, these findings reveal the potential importance of H(2) as a key energy source in the deep ocean. This study also highlights the genomic plasticity of SUP05, which enables this widely distributed group to optimize its energy metabolism (electron donor and acceptor) to local geochemical conditions.

Fig. 1.

Content and transcript abundance of genes from GB SUP05 populations and comparison with genomes of other sequenced SUP05. Nested circles from innermost to outermost represent (i–v) gene content with reference to GB-1: (i) Candidatus Vesicomyosocius okutanii; (ii) Candidatus Ruthia magnifica; (iii) Saanich Inlet OMZ SUP05; (iv) GB-2; (v) GB-1. Gaps indicate the absence of genes in comparison with other SUP05 genomes. Black lines on GB-1 denote the separation of contigs that comprise the metagenome. (vi–ix) Normalized abundance of 454 transcripts: (vi) GB-2 transcripts in background (blue); (vii) GB-2 transcripts in plume (red); (viii) GB-1 transcripts in background (blue); (ix) GB-1 transcripts in plume (red). Gray highlights on outermost circles indicate genes of interest: 1, hydrogenase operon; 2, urease operon; 3, sox operon; 4, cytochrome c oxidase complex.

Fig. 2.

Map of pathways for sulfur oxidation by GB SUP05. Inset histograms depict the gene transcript abundance for individual genes in GB-1 and GB-2. Transcript abundance is normalized for gene length and total number of reads per dataset.

Fig. 3.

(A and B) Organization and transcript abundance of GB-1 and -2 (A) and putative SUP05 (B) hydrogenase genes and comparison with closely related sequences from GenBank. Genes are colored according to normalized transcript abundance in plume and background. Arrows indicate shared genes and percent amino acid identity between predicted proteins. Dotted line in GB-2 indicates separation of contigs. (C and D) Normalized transcript abundance for genes encoding small (HydA, HupS) and large subunits (HydB, HupL) of GB-1 and -2 (C) and putative SUP05 (D) hydrogenases.

Fig. 4.

Phylogeny of group 1 membrane-bound Ni, Fe hydrogenase large subunit inferred with maximum likelihood. Bootstrap values >80 are shown. Sequences in green are from GB; sequences in red are hydrothermal vent derived and sequences in blue are from the epipelagic ocean.