Low- and high-temperature chimney samples were collected during November 2014-January 2015 Dayang 35 cruise to the Longqi field by using the manned submersible Jiao Long (Figure 1, Table 1). Longqi is a large, deep-sea hydrothermal venting field with approximately 6.7 × 10⁴ m² of low-magnetization zone on the ultraslow spreading SWIR. After sampling by the robotic arm, the sample was sealed in bio-box which was prefilled with sterilized sea water to minimize contamination. The temperature of selected chimneys' venting fluid in this study was measured in a range of 13.3°C to 379°C. Immediately after the sample was brought onboard, it was stored at −20°C and later at −80°C in laboratory until further analysis. Comparisons were made to bacterial and archaeal sequence data of chimney samples collected from M vent on EPR9-10°N during the AT26-10 cruise from Dec 29, 2013 to Jan 26, 2014), and published data of LS7 vent on MAR obtained from NCBI SRA database (SRP005280).

For chimney samples, named as JL90, JL94D, JL94H, and JL95 from SWIR, and CH7 from EPR9°N (Figure S1). DNA extraction was only applied on their outer layers, approximately 1–5 mm of thickness, following the protocols described in detail previously (Zhou et al., 1996; Vengadesh et al., 2016). The 16S rRNA genes were amplified using the polymerase chain reaction (PCR) with primer sets 515F (5′-xxxxxxxGTGCCAGCMGCCGCGGTAA-3′, x region represents the tag for Miseq sequencing) and 806R (5′-GGACTACHVGGGTWTCTAAT-3′) for bacteria (De Mandal et al., 2015) and 519F (5′-xxxxxxxCAGYMGCCRCGGKAAHACC-3′) (Flores et al., 2011) and 915R (5′-RGTGCYCCCCCGCCAATTC-3′) (Pinto and Raskin, 2012) for archaea. The PCR condition for bacterial 16S rRNA genes amplification was: 95°C for 5 min; 30 cycles of 95°C for 40 s, 55°C for 40 s, 72°C for 40 s; 72°C for 10 min, and hold at 4°C. The PCR condition for archaeal 16S rRNA genes amplification was 95°C for 5 min; 35 cycles of 95°C for 40 s, 57°C for 40 s, 72°C for 1 min; 72°C for 10 min, and hold at 4°C. PCR products were purified using the Gel purification kit (Tiangen Biotech Co. Ltd., Beijing) before sending for Miseq sequencing in Personalbio Company (Shanghai). The assembled Miseq sequences were submitted into NCBI database (GenBank Accession No. PRJNA320664).

Data analysis of the 16S rRNA Miseq sequence was performed using QIIME version 1.9.1 software pipeline (Caporaso et al., 2010) and the QIIME-compatible version of the SILVA-123 database (https://www.arb-silva.de/) for template-based alignment and taxonomic assignment. Assembled reads that passed the chimera checking were clustered into de novo operational taxonomic units (OTUs) at a cut off of 97% sequence similarity. We calculated diversity indices (Shannon and chao1 evenness) using QIIME, and constructed distance matrices in PAST software package (Hammer et al., 2009) using the Bray-Curtis calculator of community membership and structure for comparison between samples. The Venn diagram and RDA analysis were carried by R Venn Diagram and Vegan package, respectively.

Mineral analysis of the chimney samples was obtained using X-Ray Fluorescence Spectrometer (XRF-1800, Shimadzu Japan), X-Ray Polycrystaline Diffractometer (D8 Advance, Bruker, German) and ICP-AES (iCAP6300, Thermo, USA). Powder XRF and XRD samples were dried overnight at 70°C and deposited on sample holders. The procedure for XRF and XRD analysis was carried out according to the methods described in details by Peng et al. (2011). Subsamples of materials used for the above analysis were collected for ICP-AES analysis; about 200 mg of this powdered and dried material from each sample was added to in a Teflon digestion-crucible and followed by a strong acid digestion method (Wong and Li, 2004).

The temperature of fluids at sampling vent sites ranged from a low temperature of 13.3°C for JL90 to a high-temperature reached 379°C at JL95. The pH of the venting fluid, the salinity and DO of surrounding deep-sea water are given in Table 1. The JL90, JL94D, and JL94H chimney samples were dominated by the elements of sulfur and iron which composed of pyrite (FeS₂). Zinc sulfide (ZnS) was a major component of these chimneys. But the JL95 chimney sample was composed of the Bementite and Birnessite according to the results of XRD (Figure S2). The concentration of manganese (3.45 × 10⁵ mg kg⁻¹) in JL95 was nearly 5 times more than the iron (6.88 × 10³ mg kg⁻¹), and much more than for other elements such as sulfur (1.68 × 10³ mg kg⁻¹) and zinc (1.54 × 10² mg kg⁻¹) (Table 1).

We targeted the V4-V5 region to characterize the archaeal communities, and V4 region for bacterial communities associated with hydrothermal deposits according to previous report (Flores et al., 2011). In total, we generated average 5.02 × 10⁴ high-quality archaeal sequences per sample (≈410 nt length) and average 2.76 × 10⁴ bacterial sequences (270–275 nt length) per sample for all detected chimneys. Alpha-diversity assessments (Chao1 index and Shannon analysis) provided comparable results showed in Figure S3 and Table 2. Additionally, the Shannon indices indicated that the bacterial diversity (ranged from 6.48 to 10.6) was greater than archaeal diversity (ranged from 4.59 to 6.83) in each sulfide sample (Table 2). The microbial community composition between chimneys samples (beta-diversity) were assessed using OTU-based metrics (Bray-Curtis) to form the neighbor-joining clustering tree with weighted UniFrac. On the whole, the JL90 and JL94D bacterial communities were grouped together, meanwhile the high-temperature vent chimney samples, JL94H and JL95 were scattered in different branch of the tree (Figure S4).

Based on the alpha-diversity of each chimney and the beta-diversity among target chimney samples in this study, we attempted to evaluate how the microbial community differences link to possible biogeographic functions.

In this study, high-throughput, DNA-based analysis of environmental samples has been applied to investigate the microbial communities of chimney samples collected from Longqi hydrothermal field. Recent studies have found archaeal DNA was poorly recovered from lower temperature, diffuse flow vents or inactive chimneys (Bourbonnais et al., 2012; Sylvan et al., 2012; Gulmann et al., 2015) and speculated that it was due to low low abundance or primer mismatch. In this study, we presented the archaeal community analysis of chimney samples including low-temperature deep-sea hydrothermal vent from Longqi field at SWIR using high-throughput Miseq sequencing. A total of 110 473 and 200 630 amplicons of bacterial and archaeal 16S rRNA genes, respectively, have been obtained, suggesting that the methods appeared to be efficient. Deep sequencing of archaea and bacteria from those chimneys revealed thousands of bacterial and archaeal lineages, the majority of which appeared in very low abundance just as presented in the previous studies (Sogin et al., 2006; Huber et al., 2007).

According to the Venn diagrams showing (Figure S6), JL94D and JL94H shared the same sampling location and also the highest percent of bacterial OTUs, 9.6%, followed by the overlap between JL90 and JL94D, 7.7%, JL90, and JL94H, 7.6%. Besides, the bacterial OTUs overlap between JL95 and JL94 or JL90 about or below 6.2%. The species composition overlap reflected the microbial resemblance among the four chimneys at Longqi field on SWIR.

Microbial community structures were clearly correlated to the environmental parameters, and among all the considered parameters the in situ temperature was the most influential one. Previous studies demonstrated that the biogeographical patterns of microbial communities were shaped in part by local fluid geochemistry in active hydrothermal vent chimneys (Flores et al., 2011), mineralogy on inactive seafloor sulfide deposits (Toner et al., 2013) and geological processes, such as eruption, on diffuse-flow vents (Gulmann et al., 2015). To evaluate the effects of linear correlation between environmental factors and microbial communities, redundancy analysis (RDA) is used in this study (Table S1). Compared to the other high-temperature vents and low-temperature vent of JL94D, the RDA analysis also confirmed the separation of JL95 from other chimneys, correlated with a significant decrease of sulfur, Fe, and Zn content, with the increase of temperature of fluid and manganese content of chimney (Figure S5). But there was no genus which was found to have a significant linear correspondence with any environmental factor. Therefore, we assumed that regional geochemical condition might have affected the structure and function of the microecosystem in the SWIR region.

Potential energy sources for deep-sea vent chemoautotrophy include reduced sulfur compounds, molecular hydrogen, reduced metals and ammonium. Classic sulfur-oxidizing bacteria have been detected as the dominant families, suggesting a strong sulfur-metabolizing potential in all tested chimney samples (Table 3). Many members within Epsilon- and Gammaproteobacteria are known to be chemoautotrophs utilizing inorganic sulfur as election donor to gain energy (Nakagawa et al., 2005; Yamamoto and Takai, 2011; Anderson et al., 2015). The deep-sea chemoautotrophic Gammaproteobacteria, possess two different sulfur-oxidization pathways including the reverse sulfate reduction and the Sox multienzyme system without SoxCD, and strictly require co-existence of reduced sulfur compounds and O₂ (Yamamoto and Takai, 2011). In this study, Gammaproteobacteria dominated in all detected chimney samples with the sulfide-oxidizing bacterium within the genera of Thiotrichaceae, Ectothiorhodospiraceae, Thiohalophilus, and Piscirickettsiaceae (Figure 4). It seemed to indicate that both the reduced sulfur compounds and O₂ were steadily supplied into the chimney habitats. Epsilonproteobacteria were known to play a significant role in carbon, nitrogen and sulfur cycling and had consistently shown to be the most numerically abundant bacteria in sediment (López-García et al., 2003), hydrothermal fluids (Huber et al., 2010), hydrothermal plumes (Nakagawa et al., 2005), and vent chimneys (Campbell et al., 2001; Opatkiewicz et al., 2009; Dahle et al., 2013). Based on our data, Epsilonproteobacterial amplicons ranged from 1.3% at high-temperature vent chimneys to 37.6% for diffusive vent chimneys. Certain sequences dominated and closely related to the known chemosynthetic, sulfur-oxidizing genera Sulfurovum and Sulfurimonas. Similar communities were also found in cool, diffusive flow at Axial Seamount on the Juan de Fuca Ridge (Akerman et al., 2013) and the biofilms growing on the chimney walls at the Loki's Castle vent field (Dahle et al., 2013). The most abundant genera of Sulfurovum and Sulfurimonas within Epsilonproteobacteria were recovered from all active chimneys in our study, and also recovered from inactive sulfides in the EPR (Sylvan et al., 2012). It is possible that these groups represent the widely distributed species at active sulfides and the survivable relict populations at inactive chimneys by oxidizing sulfide minerals.

Ammonia oxidation is the first step of nitrification, in which ammonia is first oxidized to nitrite by ammonia-oxidizing bacteria and/or archaea (AOB or AOA), then subsequently to nitrate by nitrite-oxidizing bacteria (NOB). The Thaumarchaeota had rapidly gained much attention after the discovery that some of them have been able to oxidize ammonia aerobically, providing the first example of nitrification in the Archaea and therefore extending the range of microorganisms capable of this important metabolism, which was previously thought to be restricted to a few proteobacterial lineages (Könneke et al., 2005). Therefore, we deduced that Thaumarchaeota might be the major Ammonia Oxidizing Archaea (AOA) among the recovered microorganisms at high-temperature vent chimneys JL94H and JL95 with relative abundance over 40%, also for JL90 with ~24% (Figure 2). Besides, the Ammonia Oxidizing bacteria (AOB) within Nitrosococcus were recovered and represented the abundant genus (over 1%) in JL90 and JL95. Related genus of nitrifier, the Nitrospira, and Nitrosomonas, were found in all samples and with highest abundance in JL95 (Table 3). Recently, the completely nitrifying bacterium from the genus Nitrospira was reported (Daims et al., 2015), indicating that the globally distributed nitrite oxidizers fundamentally changed the picture of nitrification and might act as key microbial communities involved in nitrogen-cycling on the high-temperature chimney JL95 and other samples at Longqi hydrothermal field.

The reduced metals (Fe, Mn, Cu, etc.,) are the endmembers in vent fluids and potential energy sources for deep-sea vent chemoautotrophs. Fe(II) is a common and often the most dominant metal. Microaerophilic Fe-oxidizing microorganisms (FeOM) colonize gradients of Fe(II) and oxygen, taking advantage of the available chemical energy. Vast communities of FeOM proliferate at deep sea hydrothermal vents, forming mineralized mats (Chan et al., 2013). The “zetaproteobacterium” Mariprofundus ferooxydans (Emerson and Moyer, 2002) from Loihi Seamount, and several Alpha- and Gammaproteobacteria strains are reported as the chemoautotrophic Fe(II) oxidizers described from deep-sea vents (Edwards et al., 2003). Mariprofundus is the sole member of the class Zetaproteobacteria in the phylum Proteobacteria. Several cultured members (JV-1, PV-1) of Fe-oxidizing Mariprofundus were isolated from deep-sea hydrothermal fields (Singer et al., 2011; Makita et al., 2017). The genus of Mariprofundus were observed as only abundant FeOM group (relative abundance over 1%) in the chimney samples JL90, JL94D, and JL94H, except in JL95 with 0.705 (Table 3), indicating that Fe-oxidizing bacteria within Mariprofundus were common at the Longqi field and probably played a major role in Fe oxidation. Mn(II) oxidation mediated by heterotrophic Bacillus species in Guaymas Basin hydrothermal plumes as reported previously (Dick et al., 2006). Those species were not recovered in any of the chimney samples collected from Longqi field at SWIR, but the genus of Roseobacter in Alphaproteobacteria were observed and inferred as the Mn(II) oxidizer.

Hydrogen has also been shown to be an important energy source in vent fluids at the Logatchev and Rainbow fields on the Mid-Atlantic Ridge (Takai et al., 2006). Chemoautotrophs with the ability to derive energy from H₂ oxidation have been isolated from various deep-sea hydrothermal fields, including Aquificales, Epsilonproteobacteria, Desulfurococcales, Methanococcales, Thermodesulfobacteriales, and Deferribacterales (Nakagawa and Takai, 2008). Analysis of tag sequences revealed members of the genera Hydrogenobactera and Persephonell within the Aquificae and Hydrogenimonas in the Epsilonproteobacteria (Table 3). These groups are likely H₂-oxidizing bacteria.

To evaluate the effects of geological and geochemical characteristic on microbial communities on the surface of active chimneys in Longqi field at SWIR, we compared the bacterial and archaeal distribution pattern with habitats of active chimney both from slow-spreading ridge of MAR and fast-spreading ridge of EPR. Overall, the microbial composition on active chimneys recovered by tag sequencing at SWIR in our libraries was different from hydrothermal vents at EPR and MAR by using cluster analysis (Figures 7, 8). The results showed that all archaeal communities from chimneys at SWIR and EPR were clustered into different branches from the high-temperature vent chimney in MAR. On the other hand, the high-temperature vent chimney LS7 from MAR was highly dominated by Epsilonproterobacteria, which quite different from the bacterial composition of other active chimneys from SWIR and EPR, might lead to be clustered into a separate branch. The bacterial communities of JL95 and CH7 from EPR were surprisingly clustered in to a group with limited Epsilonproteobacteria, but with dominant Gammaproteobacterial sulfur oxidizers. The order Xanthomonadales within Gammaproteobacteria was only dominantin JL95, which was also observed in the microbial community of deep-sea sediment (Li et al., 2015). Within archaeal communities, the culturable genus of Pyrococcus and Thermococcus were recovered frequently in molecular environmental surveys at hydrothermal vent (Edgcomb et al., 2007; Flores et al., 2011; Li et al., 2014), but were absent in our archaeal tag sequences, which might be because of the distinct geographic locations and geochemical conditions related.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.