Members of the bacterial genus Sulfurimonas (phylum Campylobacterota) dominate microbial communities in marine redoxclines and are important for sulfur and nitrogen cycling. Here we used metagenomics and metabolic analyses to characterize a Sulfurimonas from the Gakkel Ridge in the Central Arctic Ocean and Southwest Indian Ridge, showing that this species is ubiquitous in non-buoyant hydrothermal plumes at Mid Ocean Ridges across the global ocean. One Sulfurimonas species, USulfurimonas pluma, was found to be globally abundant and active in cold (<0−4 °C), oxygen-saturated and hydrogen-rich hydrothermal plumes. Compared with other Sulfurimonas species, US. pluma has a reduced genome (>17%) and genomic signatures of an aerobic chemolithotrophic metabolism using hydrogen as an energy source, including acquisition of A2-type oxidase and loss of nitrate and nitrite reductases. The dominance and unique niche of US. pluma in hydrothermal plumes suggest an unappreciated biogeochemical role for Sulfurimonas in the deep ocean.
The genus Sulfurimonas belongs to the phylum Campylobacterota (former class Epsilonproteobacteria). It was originally proposed after the isolation of Sulfurimonas autotrophica from sediments collected at a deep-sea hydrothermal vent1. Since then, 12 distinct Sulfurimonas species have been isolated from oxygen-deficient environments2,3,4,5,6,7,8,9,10,11. Bassed on 16S rRNA gene sequences, this mesophilic and chemolithoautotrophic bacterial genus is ubiquitous and a dominant member of microbial communities inhabiting redoxcline environments12, including sulfidic environments at deep-sea hydrothermal vents13,14,15,16,17. The described members of the genus Sulfurimonas occupy habitats defined by moderate temperatures, elevated hydrogen sulfide concentrations and low oxygen concentrations (<40 µM) compared with the habitats of other hydrothermal Campylobacterota members (that is, Sulfuruvum16) and marine sulfur oxidizers (that is, SUP0518,19). Yet, abundant Sulfurimonas 16S rRNA gene sequences have also been reported in non-buoyant stage of hydrothermal plumes14,20,21,22,23,24. Hydrothermal plumes occur wherever hot anoxic hydrothermal fluids emitted from the seabed mix with cold oxygenated seawater. They can rise to hundreds of metres off the seafloor and disperse thousands of kilometres away from their source25. At the non-buoyant stage, hydrothermal plumes consist mostly of cold and oxygen-saturated seawater with highly dilute admixtures of hydrothermal fluid (<0.01%)25,26. For this reason, non-buoyant hydrothermal plumes have not been considered a permanent niche and habitat for Sulfurimonas. The repeated detection of Sulfurimonas sequences in such plumes was explained by passive transport from seafloor and subseafloor environments26. However, no study has directly tested whether non-buoyant plumes provide a suitable environment for growth of specific members of Sulfurimonas. The hydrothermal plumes contain substantial amounts of inorganic reduced gases (H2S, CH4 and H2) and metals (Fe, Mn, Cu, Zn and Co)27, which have considerable impact on ocean chemistry28. Thus, the identification and elucidation of the physiology of microorganisms growing in the plume are crucial to understanding the ocean’s biogeochemistry.
In this study, we investigated the distribution and function of Sulfurimonas in the hydrothermal plumes. We studied its ribotypes, genotypes and metabolism in two vent plumes of Gakkel Ridge and in one plume of the Southwest Indian Ridge (SWIR), and compared these with publicly available data from other vent plumes of Mid Ocean Ridges and other environments hosting Sulfurimonas sp. Our hypothesis is that non-buoyant hydrothermal plumes are a suitable environment for specific members of Sulfurimonas.
An uncharacterized Sulfurimonas in hydrothermal plumes
We investigated microbial community compositions in deep-sea water samples from the Gakkel Ridge and the Southwest Indian Ridge (SWIR). Campylobacterota accounted for 70−79% and 9−19% of all 16S rRNA sequences of non-buoyant hydrothermal plumes of Aurora (3,360−3,575 m depth) and Polaris (2,574−2,846 m depth) mounds29, respectively, and up to 16% of total microbial cells (Extended Data Fig. 1a). Almost all Campylobacterota-affiliated 16S rRNA sequences (>99%) in the non-buoyant hydrothermal plumes of Gakkel Ridge and in seawater from a ridge valley of the SWIR belonged to the genus Sulfurimonas (Supplementary Table 1 and Extended Data Fig. 2). In addition, more than 97% of the Sulfurimonas sequences of these three remote sites on ultraslow spreading ridges belonged to two closely related operational taxonomic units (OTU1 and OTU2), with a similarity of 99.5%. Fluorescence in situ hybridization using both a Campylobacterota-specific rRNA probe and tailored highly specific probes for the two detected Sulfurimonas OTUs confirmed these results (Extended Data Fig. 1b–f).
On the basis of environmental DNA retrieved from hydrothermal plume samples from Gakkel Ridge, we obtained two near-complete (93.6 − 99.95% completeness) and high-quality draft30 Sulfurimonas metagenome-assembled genomes (MAG-1 and MAG-2; Supplementary Table 2). These two MAGs have an average nucleotide identity (ANI) of 98.9%, confirming that the Gakkel Ridge plumes host two closely related strains of the same Sulfurimonas species. The 16S rRNA gene sequences from the MAGs are 99.5−100% identical with the dominant OTUs of the 16S rRNA gene amplicons, pointing toward the prevalence of the same Sulfurimonas strains in the Gakkel Ridge and the SWIR water samples, which are more than 15,000 km apart and in different deep-water current systems.
The Sulfurimonas species that is most closely related to the hydrothermal plume ribotype and genotype is Sulfurimonas autotrophica, with 94% 16S rRNA gene identity and an ANI of only 74.2%. S. autotrophica has been isolated from Pacific Ocean hydrothermal sediments1. Yet, phylogenetic analyses based on 16S rRNA genes assigned the plume-hosted Sulfurimonas sequences to an independent clade which included neither S. autotrophica nor Sulfurimonas sequences from other hydrothermal vent environments (Fig. 1a). Instead, our Sulfurimonas phylotypes form a well-supported clade together with sequences derived from hydrothermal plumes in the Atlantic Ocean (Mid Atlantic Ridge, Mid Cayman Ridge), in the Gulf of California (Guaymas Basin), in the Pacific Ocean (East Pacific Rise) and from an oxic subsurface aquifer (Mid Atlantic Ridge). This phylogenetic placement is supported by phylogenomic trees based on concatenated single-copy genes (SCGs) (Fig. 1c). The comparative analysis of Sulfurimonas genomes identified 7,569 clusters of orthologous genes. About 13% of these represent the core genome of this genus (Extended Data Fig. 4a). Similar results were recently reported in a more comprehensive pangenomic survey of the genus Sulfurimonas8. Altogether, these findings support the occurrence of a previously uncharacterized Sulfurimonas taxon in non-buoyant hydrothermal plumes. According to the standards for microbial uncultivated species31,32, we propose to name this uncultivated (U) taxon ‘USulfurimonas pluma’.
U S. pluma inhabits oxygen-saturated hydrothermal plumes
To identify the environmental factors shaping the habitat of the here described US. pluma taxon, we compared the hypervariable V3−V4 region of the 16S rRNA genes of Sulfurimonas obtained from the Arctic and from SWIR to those found in other environments (Extended Data Fig. 5), and assigned environmental characteristics (Supplementary Table 3). The studied 308 samples contained 1,389 Sulfurimonas oligotypes (amplicon sequence variants distinguished on the basis of information entropy determined for each nucleotide position33; Supplementary Fig. 1 and Table 4). On the basis of the presence and proportion of oligotypes within and among the environmental categories, they formed 149 ecotypes, 38 of which were specific to hydrothermal plumes (Extended Data Fig. 3). Two of these oligotypes represented between 28% and 97% of the Sulfurimonas sequences in the hydrothermal plume samples. These oligotypes were identical to the dominant OTUs in Gakkel and SWIR plumes (OTU1 and OTU2). These results confirm the existence of non-buoyant plume-specific Sulfurimonas ecotypes that are very rare in other hydrothermal environments (sequence proportions <1%; Fig. 1b). The US. pluma cluster of closely related sequences (>99.5% identity) dominated hydrothermal plumes across the ridge systems of the Central Arctic, Atlantic and Indian/Southern Oceans (Fig. 1a). The same ribotype was also found in the plume and the surrounding water column of the Guaymas Basin in the Gulf of California34, but with low proportions to the total bacterial community (Fig. 1a).
The specificity of US. pluma to hydrothermal plumes is also supported by recruitment analysis of metagenomic and metatranscriptomic reads from different hydrothermal environments. MAG-1 of US. pluma recruited preferentially metagenomic reads from hydrothermal plumes rather than from benthic hydrothermal environments (Extended Data Fig. 4b). US. pluma MAG-1 recruited transcripts from all seawater metatranscriptomes, yet the transcription of more than 60% of all its gene clusters was limited to samples from Gakkel Ridge non-buoyant plumes and from a diffusive fluid sample from Juan de Fuca Ridge (Extended Data Fig. 4b). The recruitment of few metagenomic and metatranscriptomic reads from Guaymas Basin hydrothermal plumes is consistent with previous studies reporting low proportions of Sulfurimonas sequences in the plume34,35 (Fig. 1a). As to the plume samples from Beebe vent at Mid Cayman Rise, the recorded water depth (4,900 m) suggests that water samples were collected in the deepest part of the rising plume, where the microbial community is still mostly influenced by benthic communities, yet comprises few microbes growing within the hydrothermal non-buoyant plume36,37.
U S. pluma lost typical Sulfurimonas genes
The US. pluma MAGs have a size of 1.68−1.77 Mbp, which is considerably smaller than the genomes of other Sulfurimonas strains, including S. gotlandica GD1 isolated from a pelagic redox environment (Extended Data Fig. 4a). Despite their small size, the MAGs of US. pluma are as complete as the closed genome of S. autotrophica, based both on reference SCGs and machine learning algorithm methods (Supplementary Table 2). The smaller size is partly explained by the high gene density (that is, number of genes per kbp) of US. pluma MAGs, which exceeds those of all other Sulfurimonas genomes (Extended Data Fig. 4a). This suggests partial deletion of non-coding DNA sequences, resulting in a streamlined genome. Nonetheless, the US. pluma MAGs lack several genes that code for core and apparent niche functions of isolated Sulfurimonas strains, such as for sulfide oxidation and nitrate/nitrite reduction8,12 (Table 1 and Supplementary Table 5). We investigated whether misassembly or binning errors caused the absence of these genes. None of the 16 individual genomic bins that were the basis for the consensus US. pluma MAGs contained the missing functional genes. Furthermore, contigs obtained from the individual assemblies (n = 12) and the co-assembly (n = 1) of Gakkel Ridge seawater metagenomes did not contain additional Sulfurimonas/Campylobacterota genes that did not bin to the US. pluma MAGs. The alignment of US. pluma MAGs with the S. autotrophica genome showed that only single genes or operons were missing in conserved genomic regions located in the largest contigs (>40 kpb). These results excluded that the observed genome reduction was an artefact of assembly and binning procedures.
U S. pluma is a hydrogen oxidizer
The MAGs of US pluma encode a membrane-bound (NiFe)-hydrogenase, which is the highest expressed catabolic gene in all metatranscriptomes of both Gakkel Ridge plumes (Fig. 2). (NiFe)-hydrogenases are ubiquitous in Sulfurimonas genomes, yet an expression of the encoding genes by Sulfurimonas in hydrothermal plumes has not been reported before (Supplementary Note 1). The fact that the hydrogenase was >13 to >500 times higher expressed than genes for sulfur oxidation suggests that hydrogen is a critical energy source to sustain the growth of US. pluma in the Aurora plume (Fig. 2), where it was most abundant and active (Supplementary Table 1 and Extended Data Fig. 2). Laboratory experiments with cultures of S. denitrificans also found that this species grows more efficiently when supplied with hydrogen than with thiosulfate as electron donor38, suggesting that hydrogen can be an important energy substrate for the genus Sulfurimonas.
The genome of US. pluma contains all genes for the oxidation of zero-valent sulfur (S0), or thiosulfate as typically described for the genus Sulfurimonas, but it misses the genes for sulfide:quinone reductases (sqr), a functional-core gene for this genus (Fig. 3 and Supplementary Table 5). Notably, US. pluma encodes a non-canonical flavocytochrome c sulfide dehydrogenase (fcc) that might replace sqr in sulfide oxidation (Extended Data Fig. 6 and Supplementary Note 2). At both Gakkel Ridge plumes, the expression of genes for canonical sulfur oxidation was substantially lower compared with that of membrane-bound (NiFe)-hydrogenase (Fig. 2).
In previous studies, reduced manganese and iron species were suggested to be additional potential electron donors for microorganisms in hydrothermal plumes25. In the MAGs of US. pluma, we found no evidence for known genes encoding enzymes for extracellular Fe (II) and Mn (II) oxidation, such as outer-membrane embedded cytochrome c (Cyc2 or Cyt572) and porin–cytochrome c protein complexes39,40. However, US. pluma encodes and expresses different genes for the uptake and control of the intercellular availability of iron and manganese (Fig. 3). Among them, the US. pluma genomes encode for an iron storage protein such as ferritin, which is absent in the genomes of S. gotlandica and S. autotrophica (Supplementary Table 5). Ferritin is an intracellular protein that oxidizes excess ferrous iron and can store up to 2,000 non-reactive ferric iron atoms41. During this reaction, ferretins also consume cytoplasmic O2 and remove H2O242 (Fig. 3). The high transcription of the ferritin gene in the hydrothermal plume metatranscriptomes of the Gakkel Ridge (Fig. 2) suggests a potential role in mitigating oxidative stress and in iron storage. Future studies should focus on the iron cycling pathways of US. pluma because they may be important elements of the ‘microbial Fe pump’ between hydrothermal vents and the ocean26.
Adaptations to high oxygen concentrations
US. pluma differs from cultivated Sulfurimonas strains in the lack of genes for nitrate and nitrite reductases and by coding for caa3-type (A2-type) cytochrome c oxidase (Table 1). All other Sulfurimonas sp. genomes encode only cbb3-type oxidase12, which has a high affinity for oxygen and allows growth at microaerobic conditions42. Yet, at higher oxygen concentrations (that is, >20%), the cbb3-type oxidase becomes inefficient, resulting in impaired growth9,12. In fact, the cultured Sulfurimonas strains grow optimally at an O2 concentration of 1–8%, and become inactive at O2 concentrations higher than 20%1,2,3,4,5,9,11. Moreover, previous studies found Sulfurimonas predominantly in environments subject to strong fluctuations in O2 concentrations (that is, benthic and pelagic redoxclines12; Supplementary Table 3). The cold polar waters studied here are oxygen-saturated and the diluted hydrothermal fluids do not substantially lower their oxygen contents. Hence, US. pluma is permanently exposed to high oxygen concentrations (ca. 300 µM; Supplementary Table 3). We hypothesize that the acquisition of caa3-type (A2-type) cytochrome c oxidase allows an efficient respiration of US. pluma in this fully oxic environments. This cytochrome c oxidase is present in many aerobic bacteria and it has strong homology to the mitochondrial cytochrome oxidase (A1-type)43. Of note, within the phylum of Campylobacterota, we found all four subunits of caa3-type oxidase in the genome of Sulfurovum sp. AR derived from aerobic Arctic sediments44. This oxidase has an amino acid identity of 70% to that of US. pluma. However, this caa3-type oxidase cannot be misassembled in the US. pluma MAGs because Sulfurovum sequences are rare in the Gakkel seawater (Supplementary Table 1), and the synteny analysis of contigs encoding for this enzyme points toward an acquisition by horizontal gene transfer (Supplementary Fig. 2).
The MAGs of US. pluma contain all genes of the reverse tricarboxylic acid (rTCA) cycle (Fig. 2), which is the common autotrophic carbon fixation pathway of all cultured Sulfurimonas species12. Its key genes 2-oxaglutarate:ferredoxin oxidoreductase (OOR), pyruvate:ferredoxin oxidoreductase (POR) and ATP-dependent citrate lyase (ACL) were highly expressed, suggesting the rTCA cycle as the main pathway for autotrophic growth (Fig. 3). However, carbon fixation via rTCA usually tolerates only trace amounts of oxygen because OOR and POR are oxygen sensitive45. Our study suggests that US. pluma is capable of using rTCA under fully and permanently oxic conditions. Fixing carbon via the rTCA pathway under these conditions requires specific adaptations of the genes involved as recently reported for a marine nitrite-oxidizing bacterium46. Indeed, the gene operon encoding the OOR and POR enzymes in US. pluma has five subunits instead of the four subunits present in other Sulfurimonas strains (Table 1). The phylogenetic and synteny analysis of this operon showed that the catalytic subunits of OOR (oorA) and POR (porA) are most closely related with members of the Aquificales and that the operon has the same organization as the members of this bacterial order (Fig. 4 and Extended Data Fig. 7). The Aquificales member Hydrogenobacter thermophilus encodes two OOR with either two or five subunits. Under anoxic conditions, it synthesizes the two-subunit form, but when exposed to oxygen, it switches to the five-subunit form. This strategy allows carbon fixation via the rTCA cycle even at 40% oxygen47,48. This five-subunit OOR has a lower specific activity compared with two-subunit anaerobic enzymes47, which probably requires production of large amounts of this oxygen-tolerant form49. Accordingly, the US. pluma genes encoding for OOR and POR subunits were among the most highly expressed genes in the plumes (Fig. 2).
US. pluma contains multiple oxygen-sensitive enzymes, electron carriers and enzymes that form reactive oxygen species, hence it requires solutions to reduce intracellular O2 concentrations. One of the most effective solutions is aerobic respiration. The cbb3-type and caa3-type are both highly expressed in the plumes (Fig. 2), suggesting that under sustained growth, their activity can be combined to efficiently reduce intracellular O2 concentrations and/or simultaneously use multiple electron sources for energy conservation.
Highly expressed bacterial hemerythrin-like genes may help the activity of oxidases (Fig. 3). Hemerythrins are non-heme intracellular oxygen-binding proteins. The two hemerythrin-like sequences found in genomes of US. pluma have one single hemerythrin domain, similar to those found in other Sulfurimonas species (Table 1) and other Campylobacterota and Aquificae50. Putative functions of bacterial hemerythrins include oxygen storage, oxygen detection or oxygen detoxification50. In motile microaerophilic or anaerobic bacteria, it has been proposed that hemerythrins act as an oxygen sensor for aerotaxis51,52, yet there are evidences for oxygen detoxification in microaerophilic Campylobacterota species52 and the enhancement of oxygen respiration in methane oxidizers53. US. pluma is probably motile, as indicated by the presence and expression of genes for the flagellum (Fig. 3), thus we cannot exclude a sensing role of its hemerythrins.
A further strategy to reduce intracellular O2 concentration in hydrothermal plume US. pluma is the loss of superoxide dismutases (Table 1), an enzyme that catalyses the transformation of superoxide (O2−) into molecular oxygen and hydrogen peroxide54. This is the primary antioxidant system in aerobic organisms, and it is also present in all cultivated members of Sulfurimonas (Table 1). In US. pluma, the antioxidant system includes superoxide reductase (SOR) and peroxidases (Table 1), which sequentially reduce O2− to H2O2 and then to H2O using electrons from other reduced substrates, respectively54 (Fig. 3). This system contributes to the decrease in intracellular O2 concentration and also consumes reductants in the cytosol, thereby lowering the formation of harmful reactive oxygen species54. In addition to the canonical/cytoplasmic core gene for 1Fe-SOR, US. pluma showed high transcription of a gene encoding for a 1Fe-SOR with translocation signal peptide (TAT-1Fe-SOR; Fig. 3), which is not present in any other Sulfurimonas genomes (Table 1). The presence of a TAT motif suggests a periplasmic location for this enzyme, and therefore a putative role in scavenging the periplasmic superoxide. This is further evidence for a lifestyle adapted to the oxygen-saturated non-buoyant plumes of deep hydrothermal vents and chemically similar aquatic systems.
In this comparative study of Arctic Ocean and global deep-sea microbiota, we identified an uncharacterized aerobic Sulfurimonas species inhabiting non-buoyant hydrothermal plumes and fluids, and subsurface aquifers. The genome of US. pluma shows substantial genome reduction (17−40%), including the loss of genus-specific functional genes (that is, for denitrification) and acquisition of genes allowing growth in pelagic oxygen-saturated environments.
Previous studies described SUP05 as a widespread and dominant chemolithoautotroph in hydrothermal plumes, using both sulfur compounds and hydrogen as energy source25,55. The ability of US. pluma and SUP05 to rely on the same substrates for energy to grow and their variable co-occurrence in hydrothermal plumes (Table 2) suggest that they could be competitors, as described for marine pelagic redox gradients19. Due to the low amounts of substrates (that is, tens to hundreds of nM) and the lack of a redoxcline in non-buoyant hydrothermal plumes, the mechanisms controlling niche partitioning between Sulfurimonas and SUP05 in this environment might be different from those proposed for pelagic redoxclines of the Baltic Sea19. Our results suggest hydrogen as an essential energy substrate for the growth of US. pluma in the studied hydrothermal plumes. This agrees with its dominance in other hydrogen- and metal-rich non-buoyant hydrothermal plumes (Table 2). Of note, ‘Candidatus Sulfurimonas marisnigri’ could grow at atmospheric oxygen concentrations, but only when MnO2 was supplied as a terminal electron acceptor10. It remains to be clarified whether metals and other hydrothermal compounds favour US. pluma growth and contribute to niche differentiation, and how US. pluma adapts to growth at temperatures close to the freezing point.
The global presence of a hydrogenotrophic Sulfurimonas species in transient environments such as non-buoyant hydrothermal plumes opens new paradigms in the microbial ecology of this and other aquatic habitats. So far, it has been postulated that microbes growing in the plume, such as the sulfur-oxidizing Gammaproteobacteria of the SUP05 clade and mixotrophic SAR324 deltaproteobacteria, are derived primarily from ambient seawater26,56. These microorganisms are also abundant and active in other marine pelagic environments (for example, surface and deep ocean, oxygen minimum zones), suggesting that their habitat is not exclusively hydrothermal plume but is widespread in the oceans25. Our results showed that non-buoyant hydrothermal plumes are a suitable environment for the growth of microorganisms typically inhabiting hydrothermal vents such as Sulfurimonas. We suggest that the hydrothermal plume does not act exclusively as a vector for dispersing microorganisms from benthic hydrothermal environments, but it might also support ecological connections between pelagic and seafloor/subsurface habitats. The phylogenetic analysis suggests that the US. pluma lineage could have been derived from a hydrothermal vent-associated ancestor (probably by sympatric speciation), which acquired higher oxygen tolerance and then spread across the oceans. However, it remains to be further investigated whether US. pluma originates from vent-associated environments or from background seawater. On one hand, the presence of a very similar ribotype (>99.5% 16S rRNA gene sequence similarity) in hydrothermal plumes across the globe (Fig. 1) suggests that the Sulfurimonas cluster, including US. pluma, is part of the ocean microbial seed bank, and therefore that background seawater might be the source of US. pluma. On the other hand, it may be that US. pluma enters into the hydrothermal plumes from populations living on seafloor vent-associated environments, which due to oxygen tolerance have a higher dispersal potential than benthic Sulfurimonas species, resulting in higher global connectivity17. Future studies on uncultivated Sulfurimonas species described here will be needed to verify these hypotheses, and to shed light on environmental and ecological forces that shape the connections and composition of microbial communities between different environments such as subsurface aquifers, diffusive flow and hydrothermal plumes.
Water samples were collected in the valley of the Southwest Indian Ridge (SWIR segment 10°–17° E; Atlantic sector), and at the Aurora and Polaris vent sites of the Gakkel Ridge (Central Arctic Ocean) during RV Polarstern expeditions PS81 (9 November–16 December 2013)57, PS86 (7 July–3 August 2014)58 and PS101 (9 September–23 October 2016)59, respectively. A conductivity-temperature-depth (CTD) rosette equipped with 24 12-l Niskin bottles was used to collect water samples for DNA extraction at all three sites, and for cell counts at Aurora and Polaris sites. Microbial biomass inside and outside the hydrothermal plumes at the Aurora and Polaris sites was collected by filtering in situ a large volume of seawater (ca. 200 l per sample) using in situ pumps (WTS-LV04; McLane) attached to the CTD-frame and CTD-wire and equipped with polycarbonate filters (142 mm diameter, 0.2 µm pore size; Millipore). The pumps were programmed to operate at maximum pump rate for 90 min. The hydrothermal plume signal was recorded by oxygen, sulfide, turbidity and Koichi-type redox sensors mounted on the CTD, by miniature autonomous plume recorders (MAPR supplied by the PMEL Earth-Ocean observation programme of the National Oceanic and Atmospheric Administration (NOAA)) and by custom-made sensors for temperature, redox, pH and H2S mounted on an ocean floor observation system. The typical physico-chemical signatures for hydrothermal vent plumes were not found in the water column of the SWIR segment investigated, yet turbidity and redox anomalies were recorded in bottom water60. Non-buoyant hydrothermal plumes were identified both at Aurora58 and Polaris59 locations. Both plumes contained CH4, H2 and H2S, the major difference being the high abundance of particles (that is, polymetallic sulfides) in the Aurora plume58 compared with the Polaris plume59. At these sites, seawater samples were collected in the hydrothermal plume and in the surrounding water: above the plume, below the plume, bottom water and background water (that is, seawater without physico-chemical signatures for hydrothermal plume). Seawater samples were also collected at reference stations (that is, not affected by hydrothermal plume) located 2 km (inside the ridge; ‘internal reference’) and 56 km (outside the ridge; ‘external reference’) away from the Aurora field, and 16 km (inside the ridge; ‘internal reference’) and 190 km (outside the ridge; ‘external reference’) away from the Polaris field. The stations are listed in Supplementary Table 6.
Fluorescence in situ hybridization (FISH)
Seawater samples (ca. 500 ml) were fixed with formaldehyde (2% final concentration) for 8 h at 4 °C and filtered onto 0.22 μm polycarbonate filters (47 mm diameter; Millipore). The filters were washed with sterile-filtered seawater and with 70% ethanol in MQ-water, dried and stored at −20 °C until further processing. Catalysed reporter deposition FISH (CARD-FISH) was performed according to ref. 61. For the hybridization of ribosomal RNA, the oligonucleotide EPSY914 5′–GGTCCCCGTCTATTCCTT–3′ (35% formamide62; synthetized by Biomers) was used to target members of the Campylobacterota phylum, and the oligonucleotide NON338 5’–ACTCCTACGGGAGGCAGC–3’ (synthetized by Biomers) was used as negative control (35% formamide)63. All hybridizations were conducted for 2.5 h at 46 °C. Cells were imaged at ×1,000 magnification with an epifluorescence microscope (Axiophot II Imaging, Zeiss). Cells were counter-stained with 4′,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich) and for each filter, a minimum of 800−1,000 DAPI-stained cells from 20−30 different fields of view were counted.
Oligonucleotide probes specific for US. pluma were designed on the basis of sequences retrieved through amplicon sequencing of the hypervariable V3−V4 region of the 16S rRNA gene of samples collected at the Aurora vent site and using the ARB software package64. The Probe-Match function of ARB was used to test in silico the coverage and specificity of the probes against the SILVA rRNA reference database (release 132)65 complemented with 16S rRNA gene sequences from this study. Almost all the Sulfurimonas sequences clustered into two groups (within similarity >99%), representing the Sulfurimonas OTU1 and OTU2 identified by the analysis of 16S rRNA gene amplicon sequences (described in the section ‘Illumina 16S rRNA gene sequencing’). We designed specific probes for OTU1 (SLFM-A484 5’–GCTTATTCATAGGCTACC–3’; 15% formamide) and OTU2 (SLFM-B484 5’–GCTTATTCATATGCTACC–3’; 20% formamide), both synthetized by Biomers. Due to the high similarity between these two oligonucleotides (one mismatch for G and T), each probe was used in a mix together with the other (non-labelled) probe as competitor oligonucleotide. To check the coverage and specificity of US. pluma’s probes in the environmental samples, double CARD-FISH hybridizations were carried out using the Campylobacterota probe (EPSY914) as a positive control.
DNA and RNA extraction
Seawater samples for 16S rRNA gene analysis were filtered immediately after the retrieval of the Niskin bottles. The filtration was carried out in a temperature-controlled room (2 °C) in the dark and did not exceed 1−1.5 h. Three liters of seawater were filtered onto a 0.22 μm polycarbonate filter (47 mm diameter; Millipore, Merck) during PS81 and PS86 using a vacuum pump (N 022 AN.18; KNF), and 10 liters through a 0.22 μm Sterivex filter (Millipore, Merck) during PS101 using a peristaltic pump (Masterflex; Cole Parmer). All filters were stored and transported at a temperature between −20 and −80 °C. Genomic bacterial DNA was isolated in a combined chemical and mechanical lysis procedure using the PowerWater DNA Isolation kit (MO BIO Laboratories). Before DNA isolation, the Sterivex cartridges of the 0.22 μm membranes were cracked open to place the filters in the kit-supplied bead beating tubes. The isolation was continued according to the manufacturer’s instructions, and genomic DNA was stored at −20 °C.
DNA and RNA for metagenomic and metatranscriptomic analyses, respectively, were extracted from the microbial biomass retrieved with the in situ pumps. After the recovery of the pumps, filters were immediately cut into 6 pieces, transferred to a screw-cap tube, frozen in liquid nitrogen for 5 min, and then stored and transported at −80 °C. RNA was extracted from three sections of filters in an RNase-free tube with the mirVana mRNA Isolation kit (Ambion). RNA extracts were treated with the TURBO DNA-free kit (Ambion) to remove co-extracted DNA, and purified and concentrated using the RNeasy MinElute kit (Qiagen). RNA extracted from each filter section was pooled together and stored at −80 °C. The genomic DNA was extracted from three pieces of the filter using the PowerWater DNA Isolation kit (MO BIO Laboratories) following the manufacturer’s instructions. The extracted genomic DNA from each filter was pooled and stored at −20 °C.
Illumina 16S rRNA gene sequencing
The hypervariable V3−V4 region of the bacterial 16S rRNA gene was amplified using bacterial primers S-D-Bact-0341-b-S-17 (5′-CCTACGGGNGGCWGCAG-3′) and S-D-Bact-0785-a-A-21 (5′-GACTACHVGGGTATCTAATCC-3′)66. Sequences were obtained on the Illumina MiSeq platform in a 2 × 300 bp paired-end run aiming for >50,000 reads per sample (CeBiTec), following the standard instructions of the 16S metagenomic sequencing library preparation protocol (Illumina). The workflow and scripts used in this study for the quality cleaning, merging, clustering and annotation of the sequences can be found in ref. 67. Briefly, only reverse and forward reads with quality score higher than 20 (applying a sliding window of 4) were merged, clustering of sequences into OTUs was done using the programme swarm (v2.2.2)68, and the taxonomic classification was based on the SILVA rRNA reference database (release 132)65.
RNA quantification, library preparation and sequencing were carried out at CeBiTec. Total RNA quantity and integrity were assessed using the Aligent 2100 bioanalyzer with the Prokaryote total RNA Pico assay (Agilent RNA 6000 Pico kit). Only RNA samples with an integrity number >7 were used for sequencing. The TruSeq Stranded Total RNA kit (Illumina) was used for RNA library preparation. The rRNA depletion step was omitted. Of the total RNA, 80 ng (in 5 μl volume) was combined with 13 μl of ‘Fragment, Prime and Finish mix’ for the RNA fragmentation step according to the Illumina TruSeq stranded mRNA sample preparation guide. Subsequent steps were performed as described in the sample preparation guide. The library was sequenced on a HiSeq1500 platform (Illumina) in a 1 × 150 bp single-end run generating >20 million reads per sample. The resulting reads were pre-processed, including removal of adaptors and quality trimming (slidingwindow:4:21 minlen:100) using bbduk v34 from the BBMAP package69 and Trimmomatic software v0.3570, respectively. The trimmed reads were sorted into ribosomal RNA (rRNA) and non-ribosomal RNA (non-rRNA) reads using SortMeRNA software v2.071. A random subset of 1 million rRNA reads per sample was taxonomically classified with phyloFlash software v3.0 beta 172 based on the SILVA database (release 132)65.
Oligotyping of Sulfurimonas 16S rRNA amplicon sequences
In addition to the data presented in this study, Sulfurimonas V3−V4 16S sequences were obtained from previous studies conducted in environments likely to host Sulfurimonas sp. bacteria (Supplementary Table 7). We searched the European Nucleotide Archive (ENA) and PubMed using the following search strategies: (1) The accession numbers of all amplicon datasets generated on the Illumina platform (paired-end) from ENA were extracted, filtering the results for ecological metagenomes from coastal and marine regions likely to contain Sulfurimonas sp. (that is, redoxcline environments; accessed 5 March 2020). (2) PubMed was searched using the keywords ‘Sulfurimonas, bacteria, 16S sequencing, Illumina’, filtering the results on the basis of the amplified region. (3) Further unpublished datasets were directly obtained from the authors (Supplementary Table 7).
Given the inconsistent state of the sequences obtained from public archives or the authors directly, several bioinformatic pre-processing steps were necessary to ensure a uniform data set as input for oligotyping. If required, sequences were demultiplexed and primer-clipped using cutadapt v1.9.173. Paired-end sequences were merged with PEAR v0.9.674 and quality trimmed with Trimmomatic v0.3570 using a sliding window of 4 bp with an average quality of 15. Only sequences between 380 bp and 450 bp that did not contain any ambiguous bases were retained. If minor modifications affecting at maximum 2 bp of the V3−V4 primers used here were employed in the previous studies, the resulting amplicons were trimmed to matching start and end positions, assuming that any primer bias regarding Sulfurimonas would be negligible.
Before oligotyping via minimum entropy decomposition (MED)33, non-Sulfurimonas sequences were filtered out using a blastn search against a custom Sulfurimonas database consisting of 246 curated full-length 16S Sulfurimonas sequences obtained from the SILVA reference database (v138)65 and assembled from the metagenomes presented in this study. Sequences with an alignment identity of at least 87%, an alignment length of at least 380 bp, a query start at or before position 10, a query end at or after position 380, a subject start between 280 and 380, and a subject end at or after position 690 were retained. An additional taxonomic cross-check was conducted with SINA v1.2.1175 using the full SILVA reference database (v138)65. To reduce computational requirements, sequences were first clustered with Swarm v2.2.268 and only the seed sequence of each swarm was taxonomically classified. Only sequences associated with swarms whose seed representative was classified as Sulfurimonas at a minimum alignment quality of 90 were used for oligotyping.
MED uses the information content (entropy) at each base position in a sequence alignment to partition the data set into oligotypes33. It is therefore able to resolve even single nucleotide differences between closely related sequences if these are likely to originate from a biological signal and not from random noise. Here, MED was used to identify oligotypes among the Sulfurimonas sequences generated from samples collected worldwide. MED (decompose v2.1) was run with the following parameters using default values if not otherwise indicated: minimum entropy of 0.0965 (-m), one discriminant for decomposition (-d), a minimum substantive abundance of 50 (-M) and with outlier relocation (-R). After MED, the oligotype representative sequences were again aligned against the SILVA reference database (v138)65 and only oligotypes classified as Sulfurimonas were retained for the subsequent analysis.
Before we further explored the diversity of Sulfurimonas 16S amplicons and attempted the detection of environment-specific Sulfurimonas ecotypes, we excluded samples where Sulfurimonas amplicons were likely to constitute a contamination from adjacent environments, that is, we assumed that in substrate-associated and water samples, proportions of Sulfurimonas 16S sequences lower than 0.1% and 1%, respectively, constituted a contamination. Additionally, samples with less than 1,000 sequences in total were excluded from subsequent analyses to avoid bias caused by the low sequencing depth. A total of 308 samples with 2,082,428 sequences represented in 1,389 Sulfurimonas oligotypes were retained for further analysis. A full list of samples, their total and Sulfurimonas sequence numbers, and number of Sulfurimonas oligotypes is included in Supplementary Table 4.
To identify ecotypes, samples were categorized by salinity (marine, brackish, fresh water), zone (pelagic, benthic, subsurface), water depth (coastal, deep-sea), hydrothermal vent influence and artificial environment. On the basis of these categories, the samples were grouped in 11 environments (Supplementary Table 8).
Ecotypes were defined as oligotypes occurring in at least 50% of the samples of a specific group, with 90% of the values in any of the outgroups being lower than half of the 10% quantile of the non-zero sequence proportions in the ingroup. Gamma diversity of Sulfurimonas oligotypes was assessed as rarefaction curve of oligotype numbers depending on sampling effort. For each possible number of samples, 100 samples were randomly selected to determine oligotype numbers.
Metagenome sequencing, assembly and binning
Paired-end libraries were prepared with the TruSeq DNA PCR-Free Sample Prep kit (Illumina) and sequencing of libraries was performed on a MiSeq instrument (Illumina; 2 × 300 bp paired reads) using the v3 sequencing chemistry (CeBiTec). Reads had adapters and contaminants removed with the bbduk tool of the BBMap suite (k = 27; mink = 12)69 and trimmed with Trimmomatic (slidingwindow:4:20; minlen:100; v0.35)70. Forward and reverse reads from 12 metagenomes (Supplementary Table 6) were de novo assembled following two strategies: (1) co-assembly with MEGAHIT (‘–min-contig-len 1000’; v1.1.2)76 and (2) individual assembllies with metaSPAdes (‘–meta’) within the SPAdes suite (‘-k 21,33,55,77,99,127’; v3.9.0)77. Reads were mapped and indexed using bwa (v0.7.12)78 and SAMtools (v1.5)79, and the contigs were binned with CONCOCT (v1.1.0)80. A total of 661 bins were de-replicated using dRep (v2.3.2)81 on the basis of Mash distance82 and genome-wide ANI83. Only 7% of the bins passed length (>50 kpb), completeness (>75%) and redundancy (<25%) filtering, and a total of 19 de-replicated bins (ANI > 99%) were obtained. Sulfurimonas bins were identified and refined using Anvi’o interactive interface (v6.2)84 after the Anvi’o contig database was built to calculate k-mer frequencies to identify open reading frames using Prodigal (v2.6.3)85 and single-copy genes using HMMER (v3.2.1)86, and to classify the bins on the basis of single-copy gene taxonomy of GTDB87 using DIAMOND (v0.9.14)88. Sequences of 16S rRNA genes were extracted with RNAmmer (v1.2)89. Refined Sulfurimonas bins were repeatedly re-assembled using BBmap (99% similarity) and SPAdes, removing contigs smaller than 1 kb after each re-assembly step to extend contigs and reduce the size of genome gaps. Completeness and redundancy of the final Sulfurimonas MAGs were evaluated using CheckM (v1.2.1; based on 104 bacterial single-copy genes)90, CheckM2 (v0.1.3; based on machine learning algorithm)91 and BUSCO (v5.2.2; based on 628 Campylobacterales single-copy genes)92. The number of transfer RNAs was identified using ARAGORN (v1.2.36)93. We obtained two almost complete Sulfurimonas MAGs, named MAG-1 and MAG-2 (Supplementary Table 2). These two MAGs represent consensus MAGs, which are based on 16 individual bins produced from different environmental samples. Proteins from the final Sulfurimonas MAGs were predicted and annotated using Prokka (v1.11)94. The Prokka-predicted proteins were additionally annotated with Pfam (release 30)95 and TIGRFAM (release 14)96 profiles using HMMER searches (v3.1b2)86 and by the identification of KEGG Orthology numbers with the GhostKOALA webserver97. The proteins were also assigned to clusters of orthologous groups (COGs)98 using the software COGsoft (v4.19.2012)99 and transmembrane motifs were identified using TMHMM (v2.0)100. On the basis of the various annotation tools, the annotation of proteins of specific interest was manually refined. The sequences of hydrogenases were classified using HydDB101. Iron-related genes were identified using FeGenie’s tool and database102. RedoxyBase103 and SORGOds104 were used to identify classes of peroxidase and types of superoxide reductase, respectively.
The quality-filtered short metatranscriptomic reads were mapped to annotated MAG-1 and quantified using the programme kallisto (v0.43.1)105. To compare expression levels between genes, gene expression is reported as transcripts per million (TPM)106, while the effective counts were used for the differential expression of MAG-1 genes between Aurora and Polaris plumes.
Gene prediction from metagenomic assemblies
The genes were predicted on contigs using Prodigal (v2.6.3)85 and the proteins clustered using MMseq2 (v13.45111;–min-seq-id 0.95 -c 0.95–cov-mode 1–cluster-mode 2)107. Proteins with length <50 amino acids were removed (ca. 15% of total proteins). The functional annotation of proteins was conducted using KEGG108 (release 101.0; with diamond v2.0.1588), Pfam (v35.0109; with hmmsearch HMMER v3.3.286) and NCycDB110, a curated database for nitrogen cycling genes (with diamond v2.0.1588). Taxonomic affiliation of the contigs was done with CAT (v5.2.3)111 using GTDB (release 207)87 as reference database. The metagenomic and metatranscriptomic reads were mapped to the gene sequences using bwa-mem2 (v2.2.1)112 and converted to counts per gene with htseq-count script from HTSeq (v2.0.2)113. On average, 80% and 87% of reads from Polaris and Aurora metagenomes, respectively, were retrieved by the catalogue of genes.
The backbone phylogenetic tree for 16S rRNA gene sequences was calculated using Sulfurimonas sequences from SILVA SSU r138 RefNR (n = 19), Thiomicrospira sp. CVO (original name for a Sulfurimonas denitrificans strain) from SILVA SSU r138 RefNR (U46506), Sulfurimonas crateris (MK859925), Gakkel Ridge plume Sulfurimonas MAGs (this study) and closest related 16S rRNA gene sequences to US. pluma MAGs (blastn hits filtered by identity >98 and coverage >97%: JN874148.1 and JN874176.1; GeneBank nucleotide; accessed May 2020). The sequences of Sulfuricurvum kujiense from SILVA SSU r138 RefNR (n = 3) were used as outgroup. Sequences were aligned with MAFFT using the L-INS-i method with default settings114, and the alignment was cleaned with BMGE with default setting115. Both programmes were used on the Galaxy platform116. A maximum-likelihood-based tree was constructed using W-IQ-TREE117, first searching for the best substitution model118 before evaluating branch support using 1,000 ultrafast boostrap (UFBoot) and SH-aLRT branch test replicates. Evolutionary placement algorithm (EPA) in RAxML (v8.2.4)119 was applied to add 253 partial Sulfurimonas 16S rRNA gene sequences (250−1,400 bp retrieved from GenBank nucleotide database; data accessed May 2020) to the tree without changing its topology. Further partial 16S rRNA gene sequences of Sulfurimonas sp. obtained from previous next-generation sequencing studies conducted in deep-sea hydrothermal fluids (JAH_MCR_Bv6_MCR_CTD03_08; JAH_AXV_Bv6v4_FS788; downloaded from vamps.mbl.edu) and plumes (PRJEB36848; SRP016119; PRJNA638507) were likewise added to the tree.
Phylogenetic trees for functional genes (that is, alpha subunit of 2-oxoglutarate:ferredoxin oxidoreductase, alpha subunit of pyruvate:ferredoxin oxidoreductase, beta subunit of flavocytochrome c sulfide dehydrogenase and sulfide:quinone oxidoreductase) were constructed following the workflow described for the backbone 16S rRNA gene tree, with amino acids as coding sequence and MAFFT alignment method set to ‘auto’. Sequences were obtained from GenBank nucleotide and UniProt databases (data accessed May 2020). The sulfide:quinone oxidoreductase sequence of US. pluma was added to the backbone tree with EPA. All phylogenetic trees were visualized and refined with iTOL120.
Comparative genomic analysis
We used Anvi’o v6.284 as the main tool to carry out comparative genomic analysis, and the overall workflows applied in this section can be found at http://merenlab.org/2016/11/08/pangenomics-v2/, http://merenlab.org/2017/06/07/phylogenomics/ and http://merenlab.org/data/prochlorococcus-metapangenome/.
We retrieved 26 Sulfurimonas and 2 Sulfuricurvum genomes (completeness >85% and redundancy <5%) from GenBank (accessed January 2020). Supplementary Table 9a reports information for each isolate genome and MAG. DNA and amino acid sequences of the genomes, including US. pluma MAG-1 and MAG-2, were stored in an Anvi’o’s genome database (programme ‘anvi-gen-genomes-storage’). From the genome database, we computed the pangenome to identify the gene clusters (programme ‘anvi-pan-genome’) representing sequences of one or more predicted open reading frames (Prodigal v2.6.3)85 grouped together on the basis of their homology at the translated DNA sequence level. For multiple sequences alignments, Anvi’o used MUSCLE (v3.8.1551)121, the MCL algorithm to identify clusters in amino acid sequence similarity122 and the programme ‘anvi-run-ncbi-cogs’ to annotate genes with functions by searching them against the COG database (October 2019 release)98 using blastp v2.9.0+123. ANI was computed for all Sulfurimonas species and MAGs representative for different environments (that is, hydrothermal vent and plume, marine pelagic, marine oxic aquifer, costal and terrestrial) with the anvi’o programme ‘anvi-compute-genome-similarity’.
We visualized the distribution of gene clusters across genomes with ‘anvi-display-pan’ and manually selected 258 SCGs present in all genomes. Concatenated amino acid sequences and partition file for SCG were extracted with ‘anvi-get-sequences-for-gene-clusters’ and used for phylogenomic analysis. The phylogenomic tree was built with IQ-Tree (v2.0)124, first identifying the partition substitution models125 and then constructing a consensus tree using a maximum-likelihood approach with 1,000 UFBoot branch test replicates.
Here we refer to ‘metapangenome’ as defined in ref. 126, that is, ‘the outcome of the analysis of pangenomes in conjunction with the environment where the abundance and prevalence of gene clusters and genomes are recovered through shotgun metagenomes’. We have also extended the analysis to include gene cluster expression as retrieved through shotgun metatranscriptomes. The Anvi’o pangenome was computed as described above but including only Sulfurimonas isolate genomes and US. pluma MAG-1, and using the programme ‘anvi-import-functions’ to import annotations from other databases, including Pfam95, eggNOG127, TIGRFAM96 and the script by E. Graham (https://github.com/edgraham/GhostKoalaParser) to import GhostKOALA/KEGG annotations97. To investigate the environmental distribution and gene transcription patterns of US. pluma MAG-1, metagenomic and metatranscriptomic reads from hydrothermal plumes and benthic vent environments were mapped to MAG-1 and reads recruitment information was imported to the Anvi’o pangenome. To test the specificity of read recruitments, the genomes from isolated Sulfurimonas species were also included in the metapangenomic analysis. We downloaded hydrothermal vent (that is, plume, seawater, fluid, chimney, mineral deposit and sediment; n = 63) and background seawater (n = 3) metagenomes from ENA (data accessed February 2020; Supplementary Table 9b). Hydrothermal plume (n = 2), seawater (n = 6), fluid (n = 6) and background seawater (n = 2) metatranscriptomes were downloaded from ENA (data accessed May 2020) and the metagenomics RAST server (MG-RAST; data accessed May 2020; Supplementary Table 9b). Low-quality reads were removed from metagenomes and metatranscriptomes with Trimmomatic (slidingwindow:4:20 and minlen:65; v0.35)70. The quality-filtered short metagenomic reads were mapped to the concatenated Sulfurimonas genomes and US. pluma MAG-1 with Bowtie2 (default setting and ‘–no-unal’ flag; v2.3.2)128. An Anvi’o database was created for storing contigs information (DNA sequence, GC content, tetranucleotide frequency and open reading frames) of Sulfurimonas genomes and US. pluma MAG-1 with the programme ‘anvi-gen-contigs-database’. Then the programme ‘anvi-profile’ was used to process BAM files and generate an Anvi’o profile database that contained the coverage statistics of each Sulfurimonas genome in a given metagenome, summarizing gene coverage values across metagenomes and percentage read recruitment per genome with ‘anvi-summarize’. This percentage refers to the contribution of reads recruited by a genome to total reads recruited by all genomes for each metagenome. The table was then imported to the Sulfurimonas pangenome in Anvio’o using the programme ‘anvi-import-misc-data’ as ‘layers’. The quality-filtered short metatranscriptomic reads were mapped to US. pluma MAG-1 gene clusters (extracted from Anvi’o’s Sulfurimonas pangenome with ‘anvi-get-sequences-for-gene-clusters’) with the programme kallisto (v0.43.1)105. The expression of gene clusters is reported in TPM, correcting for the effective length of the transcript. The TPM of US. pluma MAG-1 genes were imported to the Sulfurimonas pangenome using the programme ‘anvi-import-misc-data’ as ‘items’.
All statistical analyses were conducted in R using the core distribution (v3.5.2)129, the package vegan130 for ecotype analysis and the package edgeR131 for testing differences in the genes’ expression between Aurora and Polaris plumes, applying false discovery rate (FDR) criterion proposed by Benjamini and Hochberg132. Unless specified otherwise, plots were made using ggplot2133 and pooled, refined and labelled with Adobe Illustrator CS5.
Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
The sequences generated in this study and the USulfurimonas pluma genomes have been deposited in the European Nucleotide Archive (ENA) at EMBL-EBI under Bioproject PRJEB48226. All the sequences were archived using the data brokerage service of the German Federation for Biological Data (GFBio)134. The accession numbers for raw sequences and US. pluma MAGs are listed in Supplementary Table 6 and Supplementary Table 2, respectively.
Sulfurimonas V3-V4 16S rRNA gene sequences were extracted from metabarcoding studies obtained from ENA (https://www.ebi.ac.uk/ena/browser/home), and the studies and sequences accession numbers are reported in Supplementary Tables 4 and Table 7. Sulfurimonas and Sulfuricurvum kujiense full-length 16S rRNA gene sequences used in this study were obtained from SILVA RefNR database (v138; https://www.arb-silva.de) and NCBI Genebank (https://www.ncbi.nlm.nih.gov/nucleotide; JN874148.1 and JN874176.1). Partial Sulfurimonas 16S rRNA gene sequences (250−1,400 bp) were retrieved from NCBI GenBank (https://www.ncbi.nlm.nih.gov/nucleotide; data accessed May 2020), and from metabarcoding studies obtained from ENA (https://www.ebi.ac.uk/ena/browser/home) and VAMPS (vamps.mbl.edu). Sequences and studies accession numbers are reported in Fig. 1. Functional gene sequences used for phylogenetic tree analysis were obtained either from UniProt (https://www.uniprot.org/uniprotkb?query=*; data accessed May 2020) or NCBI GenBank (https://www.ncbi.nlm.nih.gov/nucleotide; data accessed May 2020), and their accessions are provided in Fig. 4 and Extended Data Fig. 6. Sulfurimonas and Sulfuricurvum genomes used in this study are available via NCBI GenBank (https://www.ncbi.nlm.nih.gov/nucleotide; data accessed January 2020) and accessions are reported in Supplementary Table 9a. Metagenomic and metranscitptomic data used in metapangenomic analysis were obtained either from ENA (https://www.ebi.ac.uk/ena/browser/home; data accessed May 2020) or MG-RAST (https://www.mg-rast.org; data accessed May 2020), and study accession number is provided in Supplementary Table 9b. Source data are provided with this paper.
Inagaki, F., Takai, K., Kobayashi, H., Nealson, K. H. & Horikoshi, K. Sulfurimonas autotrophica gen. nov., sp. nov., a novel sulfur-oxidizing e-proteobacterium isolated from hydrothermal sediments in the Mid-Okinawa Trough. Int. J. Syst. Evol. Microbiol. 53, 1801–1805 (2003).
Timmer-Ten Hoor, A. A new type of thiosulphate oxidizing, nitrate reducing microorganism: Thiomicrospira denitrificans sp. nov. Neth. J. Sea Res. 9, 344–350 (1975).
Cai, L., Shao, M. & Zhang, T. Non-contiguous finished genome sequence and description of Sulfurimonas hongkongensis sp. nov., a strictly anaerobic denitrifying, hydrogen- and sulfur-oxidizing chemolithoautotroph isolated from marine sediment. Stand. Genom. Sci. 9, 1302–1310 (2014).
Wang, S., Jiang, L., Liu, X., Yang, S. & Shao, Z. Sulfurimonas xiamenensis sp. nov. and Sulfurimonas lithotrophica sp. nov., hydrogen- and sulfur-oxidizing chemolithoautotrophs within the Epsilonproteobacteria isolated from coastal sediments, and an emended description of the genus Sulfurimonas. Int. J. Syst. Evol. Microbiol. 70, 2657–2663 (2020).
Takai, K. et al. Sulfurimonas paralvinellae sp. nov., a novel mesophilic, hydrogen- and sulfur-oxidizing chemolithoautotroph within the Epsilonproteobacteria isolated from a deep-sea hydrothermal vent polychaete nest, reclassification of Thiomicrospira denitrificans as Sulfurimonas denitrificans comb. nov. and emended description of the genus Sulfurimonas. Int. J. Syst. Evol. Microbiol. 56, 1725–1733 (2006).
Hu, Q., Wang, S., Lai, Q., Shao, Z. & Jiang, L. Sulfurimonas indica sp. nov., a hydrogen- and sulfur-oxidizing chemolithoautotroph isolated from a hydrothermal sulfide chimney in the Northwest Indian Ocean. Int. J. Syst. Evol. Microbiol. 71, 1466–5034 (2021).
Wang, S. et al. Sulfurimonas sediminis sp. nov., a novel hydrogen- and sulfur-oxidizing chemolithoautotroph isolated from a hydrothermal vent at the Longqi system, southwestern Indian ocean. Antonie Van Leeuwenhoek 114, 813–822 (2021).
Wang, S. et al. Characterization of Sulfurimonas hydrogeniphila sp. nov., a novel bacterium predominant in deep-sea hydrothermal vents and comparative genomic analyses of the genus Sulfurimonas. Front. Microbiol. 12, 626705 (2021).
Labrenz, M. et al. Sulfurimonas gotlandica sp. nov., a chemoautotrophic and psychrotolerant epsilonproteobacterium isolated from a pelagic redoxcline, and an emended description of the genus Sulfurimonas. Int. J. Syst. Evol. Microbiol. 63, 4141–4148 (2013).
Henkel, J. V. et al. Candidatus Sulfurimonas marisnigri sp. nov. and Candidatus Sulfurimonas baltica sp. nov., thiotrophic manganese oxide reducing chemolithoautotrophs of the class Campylobacteria isolated from the pelagic redoxclines of the Black Sea and the Baltic Sea. Syst. Appl. Microbiol. 44, 126155 (2021).
Ratnikova, N. M. et al. Sulfurimonas crateris sp. nov., a facultative anaerobic sulfur-oxidizing chemolithoautotrophic bacterium isolated from a terrestrial mud volcano. Int. J. Syst. Evol. Microbiol. 70, 487–492 (2020).
Han, Y. & Perner, M. The globally widespread genus Sulfurimonas: versatile energy metabolisms and adaptations to redox clines. Front. Microbiol. 6, 989 (2015).
López-garcía, P. et al. Bacterial diversity in hydrothermal sediment and epsilonproteobacterial dominance in experimental microcolonizers at the Mid-Atlantic Ridge. Environ. Microbiol. 5, 961–976 (2003).
Nakagawa, S. et al. Distribution, phylogenetic diversity and physiological characteristics of epsilon-Proteobacteria in a deep-sea hydrothermal field. Environ. Microbiol. 7, 1619–1632 (2005).
Huber, J. A. et al. Isolated communities of Epsilonproteobacteria in hydrothermal vent fluids of the Mariana Arc seamounts. FEMS Microbiol. Ecol. 73, 538–549 (2010).
Meier, D. V. et al. Niche partitioning of diverse sulfur-oxidizing bacteria at hydrothermal vents. ISME J. 11, 1545–1558 (2017).
Mino, S. et al. Endemicity of the cosmopolitan mesophilic chemolithoautotroph Sulfurimonas at deep-sea hydrothermal vents. ISME J. 11, 909–919 (2017).
Akerman, N. H., Butterfield, D. A., Huber, J. A., Huber, J. A. & Paul, J. B. Phylogenetic diversity and functional gene patterns of sulfur-oxidizing subseafloor Epsilonproteobacteria in diffuse hydrothermal vent fluids. Front. Microbiol. 4, 185 (2013).
Rogge, A., Vogts, A., Voss, M. & Labrenz, M. Success of chemolithoautotrophic SUP05 and Sulfurimonas GD17 cells in pelagic Baltic Sea redox zones is facilitated by their lifestyles as K- and r -strategists. Environ. Microbiol. 19, 2495–2506 (2017).
German, C. R. et al. Diverse styles of submarine venting on the ultraslow spreading Mid-Cayman Rise. Proc. Natl Acad. Sci. USA 107, 14020–14025 (2010).
Sylvan, J. B., Pyenson, B. C., Rouxel, O., German, C. R. & Edwards, K. J. Time-series analysis of two hydrothermal plumes at 9°50’ N East Pacific Rise reveals distinct, heterogeneous bacterial populations. Geobiology 10, 178–192 (2012).
Perner, M. et al. In situ chemistry and microbial community compositions in five deep-sea hydrothermal fluid samples from Irina II in the Logatchev field. Environ. Microbiol. 15, 1551–1560 (2013).
Haalboom, S. et al. Patterns of (trace) metals and microorganisms in the Rainbow hydrothermal vent plume at the Mid-Atlantic Ridge. Biogeosciences 17, 2499–2519 (2020).
Li, J. et al. Distribution and succession of microbial communities along the dispersal pathway of hydrothermal plumes on the Southwest Indian Ridge. Front. Mar. Sci. 7, 581381 (2020).
Dick, G. J. et al. The microbiology of deep-sea hydrothermal vent plumes: ecological and biogeographic linkages to seafloor and water column habitats. Front. Microbiol. 4, 124 (2013).
Dick, G. J. The microbiomes of deep-sea hydrothermal vents: distributed globally, shaped locally. Nat. Rev. Microbiol. 17, 271–283 (2019).
German, C. R. & Seyfried, W. E. in Treatise on Geochemistry 2nd edn (eds Holland, H. D. & Turekian, K. K.), 8, 191–233 (Elsevier, 2014).
Kadko, D., Baross, J. & Alt, J. The magnitude and global implications of hydrothermal flux. Geophys. Monogr. Ser. 91, 446–466 (1995).
German, C. R. et al. Volcanically hosted venting with indications of ultramafic influence at Aurora hydrothermal field on Gakkel Ridge. Nat. Commun. 13, 6517 (2022).
Bowers, R. M. et al. Minimum information about a single amplified genome (MISAG) and a metagenome-assembled genome (MIMAG) of bacteria and archaea. Nat. Biotechnol. 35, 725–731 (2017).
Konstantinidis, K. T., Rosselló-móra, R. & Amann, R. Uncultivated microbes in need of their own taxonomy. ISME J. 11, 2399–2406 (2017).
Murray, A. E. et al. Roadmap for naming uncultivated Archaea and Bacteria. Nat. Microbiol. 5, 987–994 (2020).
Eren, A. M. et al. Minimum entropy decomposition: unsupervised oligotyping for sensitive partitioning of high-throughput marker gene sequences. ISME J. 9, 968–979 (2015).
Dick, G. J. & Tebo, B. M. Microbial diversity and biogeochemistry of the Guaymas Basin deep-sea hydrothermal plume. Environ. Microbiol. 12, 1334–1347 (2010).
Lesniewski, R. A., Jain, S., Anantharaman, K., Schloss, P. D. & Dick, G. J. The metatranscriptome of a deep-sea hydrothermal plume is dominated by water column methanotrophs and lithotrophs. ISME J. 6, 2257–2268 (2012).
Sheik, C. S. et al. Spatially resolved sampling reveals dynamic microbial communities in rising hydrothermal plumes across a back-arc basin. ISME J. 9, 1434–1445 (2014).
Reed, D. C. et al. Predicting the response of the deep-ocean microbiome to geochemical perturbations by hydrothermal vents. ISME J. 9, 1857–1869 (2015).
Han, Y. & Perner, M. The role of hydrogen for Sulfurimonas denitrificans’ metabolism. PLoS ONE 9, 8–14 (2014).
Ilbert, M. & Bonnefoy, V. Insight into the evolution of the iron oxidation pathways. Biochim. Biophys. Acta 1827, 161–175 (2013).
Yu, H. & Leadbetter, J. R. Bacterial chemolithoautotrophy via manganese oxidation. Nature 583, 453–458 (2020).
Andrews, S. C. Iron storage in bacteria. Adv. Microb. Physiol. 40, 281–351 (1998).
Pitcher, R. S. & Watmough, N. J. The bacterial cytochrome cbb 3 oxidases. Biochim. Biophys. Acta 1655, 388–399 (2004).
Sousa, F. L. et al. The superfamily of heme–copper oxygen reductases: types and evolutionary considerations. Biochim. Biophys. Acta 1817, 629–637 (2012).
Park, B. et al. Cultivation of autotrophic ammonia-oxidizing archaea from marine sediments in coculture with sulfur-oxidizing bacteria. Appl. Environ. Microbiol. 76, 7575–7587 (2010).
Fuchs, G. Alternative pathways of carbon dioxide fixation: insights into the early evolution of life? Annu. Rev. Microbiol. 65, 631–658 (2011).
Bayer, B. et al. Metabolic versatility of the nitrite-oxidizing bacterium Nitrospira marina and its proteomic response to oxygen-limited conditions. ISME J. 15, 1025–1039 (2021).
Yamamoto, M., Arai, H., Ishii, M. & Igarashi, Y. Role of two 2-oxoglutarate: ferredoxin oxidoreductases in Hydrogenobacter thermophilus under aerobic and anaerobic conditions. FEMS Microbiol. Lett. 263, 189–193 (2006).
Yamamoto, M., Ikeda, T., Arai, H., Ishii, M. & Igarashi, Y. Carboxylation reaction catalyzed by 2-oxoglutarate:ferredoxin oxidoreductases from Hydrogenobacter thermophilus. Extremophiles 14, 79–85 (2010).
Berg, I. A. Ecological aspects of the distribution of different autotrophic CO2 fixation pathways. Appl. Environ. Microbiol. 77, 1925–1936 (2011).
French, C. E., Bell, J. M. L. & Ward, F. B. Diversity and distribution of hemerythrin-like proteins in prokaryotes. FEMS Microbiol. Lett. 279, 131–145 (2008).
Isaza, C. E., Silaghi-dumitrescu, R., Iyer, R. B., Kurtz, D. M. & Chan, M. K. Structural basis for O2 sensing by the hemerythrin-like domain of a bacterial chemotaxis protein: substrate tunnel and fluxional n terminus. Biogeochemistry 45, 9023–9031 (2006).
Kendall, J. J., Barrero-tobon, A. M., Hendrixson, D. R. & Kelly, D. J. Hemerythrins in the microaerophilic bacterium Campylobacter jejuni help protect key iron–sulphur cluster enzymes from oxidative damage. Environ. Microbiol. 16, 1105–1121 (2014).
Nariya, S. & Kalyuzhnaya, M. G. Hemerythrins enhance aerobic respiration in Methylomicrobium alcaliphilum 20Z R, a methane-consuming bacterium. FEMS Microbiol. Lett. 367, fnaa003 (2020).
Sheng, Y. et al. Superoxide dismutases and superoxide reductases. Chem. Rev. 114, 3854–3918 (2014).
Anantharaman, K., Breier, J. A., Sheik, C. S. & Dick, G. J. Evidence for hydrogen oxidation and metabolic plasticity in widespread deep-sea sulfur-oxidizing bacteria. Proc. Natl Acad. Sci. USA 110, 330–335 (2013).
Dede, B. et al. Niche differentiation of sulfur-oxidizing bacteria (SUP05) in submarine hydrothermal plumes. ISME J. 16, 1479–1490 (2022).
Schlindwein, V. (ed.) The Expedition of the Research Vessel ‘Polarstern’ to the Antarctic in 2013 (ANT-XXIX/8). Reports on polar and marine research, Bremerhaven, Alfred Wegener Institute for Polar and Marine Research, 672, 111 (2014); https://doi.org/10.2312/BzPM_0672_2014
Boetius, A. The Expedition PS86 of the Research Vessel POLARSTERN to the Arctic Ocean in 2014. Reports on polar and marine research, Bremerhaven, Alfred Wegener Institute for Polar and Marine Research, 685, 133 (2015); https://doi.org/10.2312/BzPM_0685_2015
Boetius, A. & Purser, A. The Expedition PS101 of the Research Vessel POLARSTERN to the Arctic Ocean in 2016. Reports on polar and marine research, Bremerhaven, Alfred Wegener Institute for Polar and Marine Research, 706, 230 (2017); https://doi.org/10.2312/BzPM_0706_2017
Varliero, G., Bienhold, C., Schmid, F., Boetius, A. & Molari, M. Microbial diversity and connectivity in deep-sea sediments of the South Atlantic Polar Front. Front. Microbiol. 10, 665 (2019).
Pernthaler, A., Pernthaler, J. & Amann, R. Fluorescence in situ hybridization and catalyzed reporter deposition for the identification of marine bacteria. Appl. Environ. Microbiol. 68, 3094–3101 (2002).
Alm, E. W., Oerther, D. B., Larsen, N., Stahl, D. A. & Raskin, L. The oligonucleotide probe database. Appl. Environ. Microbiol. 62, 3557–3559 (1996).
Amann, R. I. et al. Combination of 16S rRNA-targeted oligonucleotide probes with flow cytometry for analyzing mixed microbial populations. Appl. Environ. Microbiol. 56, 1919–1925 (1990).
Ludwig, W. et al. ARB: a software environment for sequence data. Nucleic Acids Res. 32, 1363–1371 (2004).
Quast, C. et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 41, 590–596 (2013).
Klindworth, A. et al. Evaluation of general 16S ribosomal RNA gene PCR primers for classical and next-generation sequencing-based diversity studies. Nucleic Acids Res. 41, e1 (2013).
Hassenrück, C., Quast, C., Rapp, J. & Buttigieg, P. Amplicon (GitHub, accessed 15 April 2019); https://github.com/chassenr/NGS/tree/master/AMPLICON
Mahé, F., Rognes, T., Quince, C., de Vargas, C. & Dunthorn, M. Swarm: robust and fast clustering method for amplicon-based studies. PeerJ 2, e593 (2014).
Bushnell, B. BBMap: A Fast, Accurate, Splice-Aware Aligner. United States (2014). https://www.osti.gov/servlets/purl/1241166
Bolger, A. M., Lohse, M. & Usadel, B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30, 2114–2120 (2014).
Kopylova, E., Noe, L. & Touzet, H. Sequence analysis SortMeRNA: fast and accurate filtering of ribosomal RNAs in metatranscriptomic data. Bioinformatics 28, 3211–3217 (2012).
Gruber-vodicka, H. R., Seah, B. K. & Pruesse, E. phyloFlash: rapid small-subunit rRNA profiling and targeted assembly from metagenomes. mSystems 5, e00920 (2020).
Martin, M. Cutadapt removes adapter sequences from high-throughput sequencing reads. EMBnet J. https://doi.org/10.14806/ej.17.1.200 (2011).
Zhang, J., Kobert, K., Fluori, T. & Stamatakis, A. PEAR: a fast and accurate Illumina Paired-End reAd mergeR. Bioinformatics 30, 614–620 (2014).
Pruesse, E., Peplies, J. & Glöckner, F. O. SINA: accurate high-throughput multiple sequence alignment of ribosomal RNA genes. Bioinformatics 28, 1823–1829 (2012).
Li, D., Liu, C., Luo, R., Sadakane, K. & Lam, T. MEGAHIT: an ultra-fast single-node solution for large and complex metagenomics assembly via succinct de Bruijn graph. Bioinformatics 31, 1674–1676 (2015).
Nurk, S., Meleshko, D., Korobeynikov, A. & Pevzner, P. A. metaSPAdes: a new versatile metagenomic assembler. Genome Res. 27, 824–834 (2017).
Li, H. & Durbin, R. Fast and accurate short read alignment with Burrows–Wheeler transform. Bioinformatics 25, 1754–1760 (2009).
Li, H. et al. The Sequence Alignment/Map format and SAMtools. Bioinformatics 25, 2078–2079 (2009).
Alneberg, J. et al. Binning metagenomic contigs by coverage and composition. Nat. Methods 11, 1144–1146 (2014).
Olm, M. R., Brown, C. T., Brooks, B. & Banfield, J. F. dRep: a tool for fast and accurate genomic comparisons that enables improved genome recovery from metagenomes through de-replication. ISME J. 11, 2864–2868 (2017).
Ondov, B. D. et al. Mash: fast genome and metagenome distance estimation using MinHash. Genome Biol. 17, 132 (2016).
Varghese, N. J. et al. Microbial species delineation using whole genome sequences. Nucleic Acids Res. 43, 6761–6771 (2015).
Eren, A. M. et al. Anvi’o: an advanced analysis and visualization platform for ‘omics data. PeerJ 3, e1319 (2015).
Hyatt, D. et al. Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinformatics 11, 119 (2010).
Wheeler, T. J. & Eddy, S. R. nhmmer: DNA homology search with profile HMMs. Bioinformatics 29, 2487–2489 (2013).
Parks, D. H. et al. A standardized bacterial taxonomy based on genome phylogeny substantially revises the tree of life. Nat. Biotechnol. 36, 996–1004 (2018).
Buchfink, B., Xie, C. & Huson, D. H. Fast and sensitive protein alignment using DIAMOND. Nat. Methods 12, 59–60 (2015).
Lagesen, K. et al. RNAmmer: consistent and rapid annotation of ribosomal RNA genes. Nucleic Acids Res. 35, 3100–3108 (2007).
Parks, D. H., Imelfort, M., Skennerton, C. T., Hugenholtz, P. & Tyson, G. W. CheckM: assessing the quality of microbial genomes recovered from isolates, single cells, and metagenomes. Genome Res. 25, 1043–1055 (2015).
Chklovski, A., Parks, D. H., Woodcroft, B. J. & Tyson, G. W. CheckM2: a rapid, scalable and accurate tool for assessing microbial genome quality using machine learning. Preprint at bioRxiv https://doi.org/10.1101/2022.07.11.499243 (2022).
Manni, M., Berkeley, M. R., Seppey, M. & Zdobnov, E. M. BUSCO: assessing genomic data quality and beyond. Curr. Protoc. 1, e323 (2021).
Laslett, D. & Canback, B. ARAGORN, a program to detect tRNA genes and tmRNA genes in nucleotide sequences. Nucleic Acids Res. 32, 11–16 (2004).
Seemann, T. Prokka: rapid prokaryotic genome annotation. Bioinformatics 30, 2068–2069 (2014).
El-Gebali, S. et al. The Pfam protein families database in 2019. Nucleic Acids Res. 47, D427–D432 (2019).
Haft, D. H. et al. TIGRFAMs: a protein family resource for the functional identification of proteins. Nucleic Acids Res. 29, 41–43 (2001).
Kanehisa, M., Sato, Y. & Morishima, K. BlastKOALA and GhostKOALA: KEGG tools for functional characterization of genome and metagenome sequences. J. Mol. Biol. 248, 726–731 (2015).
Tatusov, R. L., Galperin, M. Y., Natale, D. A. & Koonin, E. V. The COG database: a tool for genome-scale analysis of protein functions and evolution. Nucleic Acids Res. 28, 33–36 (2000).
Kristensen, D. M. et al. A low-polynomial algorithm for assembling clusters of orthologous groups from intergenomic symmetric best matches. Bioinformatics 26, 1481–1487 (2010).
Krogh, A., Larsson, B., von Heijne, G. & Sonnhammer, E. L. L. Predicting transmembrane protein topology with a hidden markov model: application to complete genomes. J. Mol. Biol. 305, 567–580 (2001).
Søndergaard, D., Pedersen, C. N. S. & Greening, C. HydDB: a web tool for hydrogenase classification and analysis. Sci. Rep. 6, 34212 (2016).
Garber, A. I. et al. FeGenie: a comprehensive tool for the identification of iron genes and iron gene neighborhoods in genome and metagenome assemblies. Front. Microbiol. 11, 37 (2020).
Passardi, F. et al. PeroxiBase: the peroxidase database. Phytochemistry 68, 1605–1611 (2007).
Lucchetti-miganeh, C., Goudenège, D., Thybert, D., Salbert, G. & Barloy-hubler, F. SORGOdb: superoxide reductase gene ontology curated database. BMC Microbiol. 11, 105 (2011).
Bray, N. L., Pimentel, H., Melsted, P. & Pachter, L. Near-optimal probabilistic RNA-seq quantification. Nat. Biotechnol. 34, 4–8 (2016).
Li, B. & Dewey, C. N. RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics 12, 323 (2011).
Steinegger, M. & Söding, J. MMseqs2 enables sensitive protein sequence searching for the analysis of massive data sets. Nat. Biotechnol. 35, 1026–1028 (2017).
Kanehisa, M., Sato, Y., Kawashima, M., Furumichi, M. & Tanabe, M. KEGG as a reference resource for gene and protein annotation. Nucleic Acids Res. 44, D457–D462 (2016).
Mistry, J. et al. Pfam: the protein families database in 2021. Nucleic Acids Res. 49, D412–D419 (2021).
Tu, Q., Lin, L., Cheng, L., Deng, Y. & He, Z. NCycDB: a curated integrative database for fast and accurate metagenomic profiling of nitrogen cycling genes. Bioinformatics 35, 1040–1048 (2019).
Li, H. et al. A cross-species alignment tool (CAT). BMC Bioinformatics 8, 349 (2007).
Vasimuddin, M., Misra, S., Li, H. & Aluru, S. Efficient architecture-aware acceleration of BWA-MEM for multicore systems. In 2019 IEEE International Parallel and Distributed Processing Symposium (IPDPS), Rio de Janeiro, Brazil, 314–324, doi: 10.1109/IPDPS.2019.00041 (2019); https://ieeexplore.ieee.org/document/8820962
Putri, G. H., Anders, S., Pyl, P. T., Pimanda, J. E. & Zanini, F. Analysing high-throughput sequencing data in Python with HTSeq 2.0. Bioinformatics 38, 2943–2945 (2022).
Katoh, K. & Standley, D. M. MAFFT multiple sequence alignment software version 7: improvements in performance and usability. Mol. Biol. Evol. 30, 772–780 (2013).
Criscuolo, A. & Gribaldo, S. BMGE (Block Mapping and Gathering with Entropy): a new software for selection of phylogenetic informative regions from multiple sequence alignments. BMC Evol. Biol. 10, 210 (2010).
Jalili, V. et al. The Galaxy platform for accessible, reproducible and collaborative biomedical analyses: 2020 update. Nucleic Acids Res. 48, 395–402 (2020).
Trifinopoulos, J., Nguyen, L.-T., von Haeseler, A. & Minh, B. Q. W-IQ-TREE: a fast online phylogenetic tool for maximum likelihood analysis. Nucleic Acids Res. 44, W232–W235 (2016).
Kalyaanamoorthy, S. et al. ModelFinder: fast model selection for accurate phylogenetic estimates. Nat. Methods 14, 587–589 (2017).
Berger, S. A., Krompass, D. & Stamatakis, A. Performance, accuracy, and web server for evolutionary placement of short sequence reads under maximum likelihood. Syst. Biol. 60, 291–302 (2011).
Letunic, I. & Bork, P. Interactive Tree Of Life (iTOL) v4: recent updates and new developments. Nucleic Acids Res. 2, W256–W259 (2019).
Edgar, R. C. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res. 32, 1792–1797 (2004).
van Dongen, S. & Abreu-goodger, C. in Bacterial Molecular Networks: Methods and Protocols, Methods in Molecular Biology (eds van Helden, J. et al.) 281–295 (Springer, 2012).
Altschup, S. F., Gish, W., Pennsylvania, T. & Park, U. Basic local alignment search tool. J. Mol. Biol. 215, 403–410 (1990).
Nguyen, L., Schmidt, H. A., Haeseler, A., Von & Minh, B. Q. IQ-TREE: a fast and effective stochastic algorithm for estimating maximum-likelihood phylogenies. Mol. Biol. Evol. 32, 268–274 (2014).
Chernomor, O., von Haeseler, A. & Minh, B. Q. Terrace aware data structure for phylogenomic inference from supermatrices. Syst. Biol. 65, 997–1008 (2016).
Delmont, T. O. & Eren, A. M. Linking pangenomes and metagenomes: the Prochlorococcus metapangenome. PeerJ 6, e4320 (2018).
Jensen, L. J. et al. eggNOG: automated construction and annotation of orthologous groups of genes. Nucleic Acids Res. 36, 250–254 (2008).
Langmead, B. & Salzberg, S. L. Fast gapped-read alignment with Bowtie 2. Nat. Methods 9, 357–360 (2012).
R Core Team. R: A Language and Environment for Statistical Computing (R Foundation for Statistical Computing, 2013).
Oksanen, J. et al. vegan: Community Ecology Package. R package version 2.6-4. https://CRAN.R-project.org/package=vegan (2022).
Robinson, M. D., Mccarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics 26, 139–140 (2010).
Reiner-Benaim, A. FDR control by the BH procedure for two-sided correlated tests with implications to gene expression data analysis. Biom. J. 49, 107–126 (2007).
Villanueva, R. A. M. & Chen, Z. J. ggplot2: elegant graphics for data analysis (2nd ed.). Meas. Interdiscip. Res. Perspect. 17, 160–167 (2019).
Diepenbroek, M. et al. Towards an integrated biodiversity and ecological research data management and archiving platform: the German Federation for the Curation of Biological Data (GFBio). In Informatik 2014 – Big Data Komplexität meistern Proc. 232 (eds Plödereder, E. et al.) 1711–1725 (Gesellschaft für Informatik, 2014).
Schmidt, K., Koschinsky, A., Garbe-Schönberg, D., de Carvalho, L. M. & Seifert, R. Geochemistry of hydrothermal fluids from the ultramafic-hosted Logatchev hydrothermal field, 15°N on the Mid-Atlantic Ridge: temporal and spatial investigation. Chem. Geol. 242, 1–21 (2007).
Perner, M. et al. The influence of ultramafic rocks on microbial communities at the Logatchev hydrothermal field, located 15 degrees N on the Mid-Atlantic Ridge. FEMS Microbiol. Ecol. 61, 97–109 (2007).
Douville, E. et al. The rainbow vent fluids (36°14’N, MAR): the influence of ultramafic rocks and phase separation on trace metal content in Mid-Atlantic Ridge hydrothermal fluids. Chem. Geol. 184, 37–48 (2002).
Ji, F. et al. Geochemistry of hydrothermal vent fluids and its implications for subsurface processes at the active Longqi hydrothermal field, Southwest Indian Ridge. Deep Sea Res. I 122, 41–47 (2017).
We thank the captain, crew and science party of RV Polarstern expeditions PS81, PS86 and PS101; H. Tegetmeyer for support with Illumina sequencing (CeBiTec laboratory; HGF-MPG Joint Research Group on Deep Sea Ecology and Technology); A. Mera and M. Philippi for helping with design and testing of the specific probes for US. pluma; M. Meiners, W. Stiens, R. Stiens, J. Barz, F. Schramm and A. Nordhausen (HGF-MPG Joint Research Group for Deep-Sea Ecology and Technology) for technical support; and the Anvi’o team and community for making microbial omics analyses accessible to every scientist. This work was funded by the Helmholtz Association (Alfred Wegener Institute Helmholtz Center for Polar and Marine Research, Bremerhaven), Max Planck Society (MPG), the ERC Advanced Investigator Grant ABYSS (294757) to A.B., and Germany’s Excellence Initiative of the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) through the Clusters of Excellence ‘The Ocean in the Earth System’ (EXC-309- 49926684) and ‘The Ocean Floor-Earth’s Uncharted Interface’ (EXC-2077–390741603).
Open access funding provided by Max Planck Society.
The authors declare no competing interests.
Peer review information
Nature Microbiology thanks Karthik Anantharaman and Mirjam Perner for their contribution to the peer review of this work.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended Data Fig. 1 Abundance of Campylobacterota and hydrothermal plume Sulfurimonas sp. cells at Arctic sites.
a, Proportion of Campylobacterota and hydrothermal plume Sulfurimonas sp. to DAPI-stained cell counts in different water types at Aurora vent and reference sites (water depth and number of samples are reported in parenthesis); b − g, fluorescence micrographs of microbes in the Aurora plume (station PS86-55) stained with DAPI (blue; b and e) and simultaneously (double CARD-FISH) with specific probes for Campylobacterota (green; c and f) and for Arctic Sulfurimonas OTU1 (red; d) and OTU2 (red; g). Panels b–g represents observations made from independent samples (n = 3). Scale bars 10 μm (b–g).
Dominant active bacterial taxa (cut-off ≥1.5 %) as detected by 16 S rRNA sequencing (cDNA). Average of sequences proportion is reported. The most dominant and active chemolithotrophs are in bold. Number of observations (n) is reported in parentheses.
Samples are ordered based on average linkage clustering of Jaccard dissimilarities (left panel). Asterisks indicate samples of hydrothermal origin. Top panel: percentage of samples an oligotype occurs in (sample coverage). Right panel: percentage of Sulfurimonas sequences of total sequences per sample. Bottom panel: Ecotypes per environmental category (brackish pelagic and benthic, as well as pelagic deep-sea and hydrothermal plume were pooled).
Pangenome analysis of Sulfurimonas isolate genomes and MAGs from different environments. Starting from the centre, the first 13 circles show the occurrence of gene clusters in a given Sulfurimonas genome. The external circle reports the gene clusters in which at least one gene was functionally annotated to COGs. The 7569 gene clusters (GCs) are organized based on their distribution across the genomes. The genomes are organized based on gene clusters they share (using Euclidian distance and Ward clustering). The barplots show some general features of each genome, the heatmap shows the ANI between genomes. b, Metapangenome analysis of Sulfurimonas isolate genomes (S. denitrificans, S. hongkongebsis, S. gotlantica, S. autotrophica) and hydrothermal plume US. pluma. Starting from the centre: the first 5 circles show the occurrence of gene clusters in a given Sulfurimonas genomes; then the following 8 circles report the expression of gene clusters of US. pluma in hydrothermal plume, seawater and fluid as inferred from metatransciptomic (MT) read recruitment (only representatives for available ridge systems are reported); finally, the external circles report the gene clusters in which at least one gene was functionally annotated (green and grey for known and unknown functions, respectively). The 4854 gene clusters (GCs) and the genomes are organized based on their distribution across the genomes and on gene clusters they share (using Euclidian distance and Ward clustering), respectively. The barplot shows the percentage of genes expressed, the heatmap shows the percentage of Sulfurimonas reads recruited by five genomes (inner circle) from metagenomes (MG; average value is reported for habitats having more than one metagenome available per ridge system; the number of metagenomes is reported in parenthesis). In both panels the dendrograms represents the hierarchical clustering of genomes based on the occurrence of gene clusters.
World map showing the geographic origin, and environment of the samples used for the analysis of Sulfurimonas oligotypes. The point size is proportional to the number of samples collected at each location. HV: hydrothermal vent.
Gene cluster organization of the five subunits of 2-oxaglutarate-ferrodoxin oxidoreductase and pyruvate-ferrodoxin oxidoreductase in the genomes of USulfurimonas pluma, Sulfurovum AR, and Aquificales’ representatives.
Phylogenetic relationships between the beta subunit of Flavocytochrome c sulfide dehydrogenase (FCC) and Sulfide:quinone oxidoreductase (SQR). The scale bar represents the expected number of changes per amino acids position. UFBoot and SH-aLRT values are based on 1000 replicates. Best substitution model: LG + R4. Sulfurimonas species are in bold.
About this article
Cite this article
Molari, M., Hassenrueck, C., Laso-Pérez, R. et al. A hydrogenotrophic Sulfurimonas is globally abundant in deep-sea oxygen-saturated hydrothermal plumes. Nat Microbiol 8, 651–665 (2023). https://doi.org/10.1038/s41564-023-01342-w