The bacterial sulfur cycle in expanding dysoxic and euxinic marine waters

Abstract

Dysoxic marine waters (DMW, < 1 μM oxygen) are currently expanding in volume in the oceans, which has biogeochemical, ecological and societal consequences on a global scale. In these environments, distinct bacteria drive an active sulfur cycle, which has only recently been recognized for open-ocean DMW. This review summarizes the current knowledge on these sulfur-cycling bacteria. Critical bottlenecks and questions for future research are specifically addressed. Sulfate-reducing bacteria (SRB) are core members of DMW. However, their roles are not entirely clear, and they remain largely uncultured. We found support for their remarkable diversity and taxonomic novelty by mining metagenome-assembled genomes from the Black Sea as model ecosystem. We highlight recent insights into the metabolism of key sulfur-oxidizing SUP05 and Sulfurimonas bacteria, and discuss the probable involvement of uncultivated SAR324 and BS-GSO2 bacteria in sulfur oxidation. Uncultivated Marinimicrobia bacteria with a presumed organoheterotrophic metabolism are abundant in DMW. Like SRB, they may use specific molybdoenzymes to conserve energy from the oxidation, reduction or disproportionation of sulfur cycle intermediates such as S⁰ and thiosulfate, produced from the oxidation of sulfide. We expect that tailored sampling methods and a renewed focus on cultivation will yield deeper insight into sulfur-cycling bacteria in DMW.

Fig 1

Dysoxic marine waters studied with respect to microorganisms driving the sulfur cycle. Triangles indicate locations that are permanently, seasonally or incidentally euxinic. For a list of studies per location, see Table S1. ETNP, Eastern Tropical North Pacific; ETSP, Eastern Tropical South Pacific; OMZ, oxygen‐minimum zone. [Color figure can be viewed at wileyonlinelibrary.com]

Fig 2

The dissimilatory conversions within the marine sulfur cycle. The oxidation state of the inorganic species is indicated at the left. Abiotic and assimilatory reactions are not indicated, except for the abiotic oxidation of sulfide which is illustrated by wide grey arrows. The S⁰ in DsrC‐trisulfide is considered zero‐valent (Santos et al., 2015). The sulfur atom in APS has an oxidation state of +6, and those in tetrathionate have an oxidation state of +2.5. A question mark symbol (?) shows that involvement is uncertain. The asterisk symbol (*) indicates that DsrT is required for sulfide oxidation in green sulfur bacteria (Holkenbrink et al., 2011), but is also found in SRB. Protein complexes other than DsrTMK(JOP) can also transfer electrons to DsrC to enable this reaction (Venceslau et al., 2014). The section symbol (§) indicates that the rhodanese sulfurtransferases Rhd‐TusA‐DsrE2 are also essential in the reaction mediated by this complex (Dahl, 2017). Apr, APS reductase; Asr, anaerobic sulfite reductase; Dox, thiosulfate:quinone oxidoreductase; Dsr, dissimilatory sulfite reductase; Fcc, flavocytochrome c sulfide dehydrogenase; Fsr, F₄₂₀‐dependent sulfite reductase; Hdr, heterodisulfide reductase; Otr, octaheme tetrathionate reductase; Phs, thiosulfate reductase; Psr, polysulfide reductase; Qmo, quinone‐interacting membrane‐bound oxidoreductase; Sat, sulfate adenylyltransferase; Sir, sulfite reductase; Soe, sulfite‐oxidizing enzyme; SOR, sulfur oxygenase/reductase; Sor, sulfite‐acceptor oxidoreductase; Sox, sulfur‐oxidizing multienzyme complex; Sqr, sulfide:quinone oxidoreductase; Sre, sulfur reductase; SULT, sulfotransferase; Tet, tetrathionate hydrolase; Tsd, thiosulfate dehydrogenase; Ttr, tetrathionate reductase. [Color figure can be viewed at wileyonlinelibrary.com]

Fig 3

Black Sea water column distribution of metagenome‐assembled genomes (MAGs) of sulfur‐cycling bacteria based on their genetic capacity. A. Physicochemical measurements and normalized cumulative metagenome coverage of all MAGs of putative sulfur‐oxidizing bacteria (SOB) combined, ^U P. aerophilus (NIOZ‐UU104), Marinimicrobia (NIOZ‐UU73), all dsrD genes combined and all MAGs of putative sulfate‐reducing bacteria (SRB) combined in samples of 15 different depths of the Black Sea. The oxygen, nitrite and sulfide data correspond to the PHOXY cruise of June–July 2013 (Sollai et al., 2019). The Black Sea metagenome was also constructed from samples taken during this cruise as detailed in Supporting Information Methods and Villanueva and colleagues (2020). Redox potential was measured during the 64PE408 NESSC/SIAM cruise of January–February 2016 from samples with a closely agreeing sulfide profile (Fig. S1). B,C. Relative abundances of MAGs of (B) putative SRB and (C) putative SOB and ^U P. aerophilus were based on normalized metagenome coverage. See Supporting Information Methods for details on the methodology and data processing. [Color figure can be viewed at wileyonlinelibrary.com]

Fig 4

Maximum‐likelihood phylogenetic reconstruction based on bacterial reductive DsrA proteins predicted from Black Sea MAGs and unbinned contigs (blue) and reference genomes (Anantharaman et al., 2018). Black dots indicate support of > 95% out of 1,000 ultra‐fast bootstraps. The scale bar indicates substitutions per site. See Supporting Information Methods for methodology and Data S1 for the full phylogenetic tree in Newick format. [Color figure can be viewed at wileyonlinelibrary.com]

Fig 5

A phylogenomic and genetic overview of important microbial players in the marine sulfur cycle of dysoxic marine water (DMW) environments. A. An unrooted phylogenomic maximum‐likelihood tree constructed from a concatenated alignment of 120 single‐copy household genes (Supporting Information Methods). Phylogenetic clades were identified, with numbers indicating the following lineages: 1, Campylobacterota; 2, Nitrospinae; 3, Nitrospirae; 4, Chloroflexi; 5, Bacteroidetes; 6, candidate phylum AAMBM5‐125‐24; 7, Marinimicrobia. Black dots indicate support by > 95% out of 1,000 ultra‐fast bootstraps. The scale bar indicates substitutions per site. The tree includes data from all available genomes (March 2020) from DMW sites (bold, coloured by environment following the colour code of Fig. 1) that contain dissimilatory sulfur genes, and relevant reference genomes (black). The superscript prefix ‘U’ indicates uncultured species for which a taxonomy has been proposed based on a high‐quality genome, functional annotation and environmental distribution (Konstantinidis et al., 2017) with the genome sequences as type material (Chuvochina et al., ; Murray et al., 2020; Supporting Information Protologue). B. An overview of the presence of functional genes enabling conversions of sulfur, nitrogen and oxygen, following the colour scheme of Fig. 2. Shortly, red indicates core genes of the Dsr/rDsr pathways, orange indicates dsrD, purple indicates dsrEFH and various oxidative sulfur genes, light blue indicates sox genes, light‐green indicates phs/psr/sre genes, dark‐green indicates various (potentially) reductive sulfur genes, black/dark grey indicates nitrogen genes, dark blue indicates oxygen reduction genes. The presence of the indicated functional genes or gene clusters is shown with filled circles; open circles reveal incomplete gene clusters. For ‘Rhodanese’, filled circles indicate 10 or more rhodanese domains (Supporting Information Methods). Stars distinguish the high‐quality genomes (> 80% complete, < 5% contaminated) from the medium‐quality genomes (> 50% complete, < 10% contaminated) analysed. Only two low‐quality metagenome‐assembled genomes, that is, that of Dehalococcoidia RBG_13_52_14 (35% complete, 2% contaminated) and the population genome of Gammaproteobacteria EOSA‐II composed of multiple combined single‐cell amplified genomes (63% complete, 21% contaminated), were also included. A comprehensive overview of genome origin, quality, classification, annotation and average amino acid identity (AAI) between genomes can be found in Table S2. [Color figure can be viewed at wileyonlinelibrary.com]

Fig 6

Conceptual ecophysiological model of the sulfur cycle and the involved microorganisms in DMW. Question marks and the light gray colour are used when there are indications for involvement of specific microorganisms and/or processes but definitive proof is lacking. Bacterial images were created with BioRender. [Color figure can be viewed at wileyonlinelibrary.com]