From August 2020 to April 2022, water samples were collected every 1 or 3 months (independent sample number = 10) by SCUBA divers in the vent area (24°50’02.6”N, 121°57’42.8”E; offshore the southeastern tip; 15-17 m depth) and non-vent area (24°50’29.9”N, 121°56’17.1”E; offshore the northwestern tip; 8-10 m depth) of Kueishan Island, Taiwan ( Figure 1A ). The data from each independently collected water sample was averaged from triplicate samples collected at each sampling time surrounding the hydrothermal vent crabs, but not directly above the vent opening. The crabs X. testudinatus were collected from the vent area ( Figures 1B–E ). Crabs were randomly selected for hemolymph collection (independent sample number = 9) as well from August 2020 to April 2022, and the hemolymph was carefully drawn from the coxa membrane of the 2^nd walking legs (the 3^rd pereopods) using a disposable 1 mL syringe with 24G X 1 Inch (0.55x25 mm) (Terumo Medical Corporation, Tokyo, Japan). The hemolymph was immediately used for pH measurements. To collect the bacterial community from X. testudinatus, the 5^th gills (N = 29) were sampled, and the carapace surface was applied using cotton swabs (N = 19). The gill tissues and cotton swabs were separately placed into 1.5 mL tubes and stored in liquid nitrogen on the fisheries craft. Upon arriving at the laboratory, samples were frozen at -80°C until DNA extraction.

The water quality was assessed according to temperature, pH, dissolved oxygen (DO), alkalinity, sulfide content and bicarbonate ( HC O 3 − ) content. The pH, sulfide and HC O 3 − content of crab hemolymph were compared with the corresponding measurements from water samples. The temperatures and pH values of the water samples were measured using a multi-parameter portable meter (Multiline^® Multi 3620 IDS, WTW, Weilheim, Germany) equipped with an IDS pH electrode (SenTix^® 940, WTW) on the fisheries craft. The crab hemolymph pH was determined using a portable pH meter (Profiline^® pH 3310, WTW) equipped with a glass-body micro combination pH electrode (InLab^® Micro, Mettler Toledo, Leicester, UK) immediately after hemolymph collection. The water alkalinity was spectrophotometrically assayed following a previously published method (Sarazin et al., 1999). Sulfide content was determined by a spectrophotometric method modified from (Cline, 1969), using a sulfide-sulfur specific mixed diamine reagent containing N, N-Dimethyl-p-phenylenediamine sulfate and ferric chloride in 50% hydrochloric acid. The dissolved inorganic carbon (DIC) content of water samples was analyzed in triplicate using a DIC analyzer (AS-C3, Apollo SciTech, Newark, DE, USA) with nitrogen as an inert gas, followed by quantification via a non-dispersive infra-red (NDIR) Li-7000 C O 2 / HC O 3 − analyzer (LI-COR, Lincoln, NE, USA). The hemolymph DIC content was measured following a method described previously (Allen et al., 2021). The HC O 3 − concentration of water samples was calculated according to DIC, pH, and temperature using a CO2SYS module undergone within Microsoft Excel (Pierrot et al., 2006). Dissociation constants were as provided and adjusted by two methodological studies (Mehrbach et al., 1973; Dickson and Millero, 1987). Alternatively, the hemolymph HC O 3 − concentration was calculated based on DIC and pH using the Henderson–Hasselbalch equation, with dissociation constants and solubility coefficients as previously described (Truchot, 1976).

DNA from the crab gills was extracted using the Wizard Genomic DNA Purification Kit (Promega, Madison, WI, USA), and DNA from the carapace surface (cotton part) was extracted using the QIAamp DNA Microbiome Kit (Qiagen, Hilden, Germany). Both extractions were performed in accordance with the manufacturers’ instructions. The quantity and quality of extracted DNA were verified using a NanoDrop 2000™ spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA). Afterward, the DNA samples were stored at -20°C before use.

LoopSeq™ 16S Long Read Kit (Loop Genomics, San Jose, CA, USA) was used to construct the bacterial full-length 16S rRNA gene libraries (V1-V9 region). The constructed DNA libraries were sequenced in a pair-end (2 × 150 bp) configuration on the NextSeq 550 platform (Illumina, San Diego, CA, USA) in accordance with the manufacturer’s instructions.

A Loop Genomics cloud platform (Microsoft Azure) was used to trim Illumina sequencing index and demultiplex the pair-end raw sequences for synthetic long-read reconstruction. The poor-quality reads and chimeric sequences were discarded. The remaining qualified sequences were uploaded to the CLC Genomic Workbench 20.0.1 (CLC Bio, QIAGEN, Aarhus, Denmark) for taxonomic profiling and abundance analysis based on a microbial genome database (2019-08-30, provided by CLC workbench). The total reads number of 16S rRNA gene sequences in each sequenced gill and carapace sample was shown in Table S1 (gill) and Table S2 (carapace). Taxonomic profiling was used to generate the abundance table based on the reads of the 16S rRNA sequence for each taxon. For each taxon, the number of reads in the sample was displayed along with the confidence score. The information was displayed using stacked bar charts. The Alpha diversity indices (Chao 1, Shannon and Simpson) were calculated to assess the heterogeneity of the bacterial community (species). Principal coordinate analysis (PCoA) was performed to determine the dissimilarity of bacterial compositions between the gill and carapace samples, based on the Bray-Curtis distance.

In order to predict the functions of bacteria collected from the gills and carapace surface of X. testudinatus, the taxonomic abundance table was used to retrieve functional annotations of identified prokaryotic taxa (FAPROTAX 1.2.5). FAPROTAX provides an informative database that can be used to sort prokaryotic taxa (e.g., classes, genera, or species) into presumptive functions based on studies of culture representatives (Louca et al., 2016). On the basis of the initial output, predicted functions were further classified into incorporative groups and represented as proportions of total predicted functions for each sample.

FISH was used to detect and locate the bacteria in the gills of X. testudinatus. The crab gills were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) solution and temporarily stored in a box filled with ice on the fishing craft. Afterward, the fixed gills were embedded in paraffin wax in the laboratory, and the wax sections (8 μm) were mounted on the slide and stored at 4°C. To perform FISH, dewaxing was performed with Histo-Clear (National Diagnostics, Atlanta, GA), and Histo-Clear was removed using an ethanol series (100%, 75%, 50% and 25%). After rehydrating the section with 20 mM Tris-HCl, it was treated with proteinase K (5 μg/mL) and lysozyme (5 mg/mL) in Tris-HCl to permeabilize the bacterial walls. The gill sections were hybridized with probe Sulfurovum-target EPI653 (5’-ATCTTCCCCTCCCASACTCT-3’) (Watsuji et al., 2010), universal bacteria probe EUB338 (5’-GCTGCCTCCCGTAGGAGT-3’) (Amann et al., 1990), and a customized probe NBC371 (5’-CCAGTTAGCTTCCCCGCTT-3’) designed for Sulfurovum NBC37-1 and based on the 16S rRNA gene sequence retrieved from NCBI (AP009179). Fluorescence images were obtained with an Olympus FV3000 confocal laser scanning microscope (Olympus Corporation, Tokyo, Japan).

Statistical analyses of relative bacteria abundance (species) and predicted function proportions were performed with GraphPad Prism 9.1.2 (GraphPad Software, San Diego, CA, USA), using Welch’s t-test to compare the gill and carapace groups. Statistical analyses of alpha and beta diversity were performed with CLC Genomic Workbench 20.0.1 (CLC Bio, QIAGEN). The difference in alpha diversity indices (Chao 1, Shannon, and Simpson) between the gill and carapace groups was determined using the Kruskal-Wallis test. The Bray-Curtis-based permutational analysis of variance (PERMANOVA) was used to compare the difference in bacterial species compositions between gill and carapace groups, which was visualized by PCoA. The level of significance was set at p< 0.05.

Kueishan Island is located at the end of the Okinawa Trough and possesses a shallow-water HV system on its southeastern tip (vent area; Figures 1A, B ). The water around the northwestern tip of Kueishan Island is a non-vent area that contains a coral reef ecosystem and exhibits relatively high biodiversity (Chan et al., 2016). According to our measurements, the environmental parameters of the vent area were more variable than those of the non-vent area (without vent openings). Parameters of water quality were measured in water samples collected from the vent and non-vent areas off Kueishan Island. The temperatures in both areas varied with the season. In the vent areas, water temperatures ranged from 18.6 to 28.3°C, while non-vent areas ranged from 20.2 to 27.9°C ( Figure 2A ). The dissolved oxygen (DO) ranged from 5.68 to 8.47 mg/L and from 7.16 to 8.68 mg/L in the vent and non-vent areas, respectively ( Figure 2B ). The alkalinity ranged from 1.56 to 2.43 mM and from 2.16 to 2.51 mM in the vent and non-vent areas, respectively ( Figure 2C ). Additionally, the pH, sulfide and HC O 3 − contents of environmental water were compared to those of crab hemolymph. The pH ranged from 6.19 to 8.12 in the vent area, and from 7.87 to 8.28 in the non-vent area. However, the crabs inhabiting the vent area seemed to maintain their hemolymph pH in a narrow range of 7.49 to 7.78, in spite of environmental pH fluctuations ( Figure 2D ). The vent and non-ven sulfide contents ranged from 9 to 3480 μM and from 0 to 23 μM, respectively ( Figure 2E ). Thus, the vent area had a large variation and much elevated sulfide content. In the crab hemolymph, sulfide was detected at a range of 102 to 747 μM ( Figure 2E ). As shown in Figure 2F , the HC O 3 − concentrations in the vent and non-vent areas respectively ranged from 0.76 to 2.03 mM and 1.90 to 2.21 mM. Notably, the hemolymph of crab exhibited a considerably large range of HC O 3 − (9.30 to 16.40 mM). An extremely high HC O 3 − concentration in the hemolymph is in line with the idea that HC O 3 − may be held in reserve by the vent crabs for coping with pH drops in the vent area.

The bacterial compositions were profiled and represented as bar charts in Figure 3A and Table S1 , S2 . The analysis showed that the gills and carapace surface were both dominated by four classes, including Epsilonproteobacteria, Gammaproteobacteria, Flavobacteriia, and Alphaproteobacteria. Despite sharing overall similar bacterial compositions, there were clear differences in the relative abundances of classes between the gills and carapace. The five most abundant classes in the gills were Epsilonproteobacteria (73.8%), Gammaproteobacteria (9.9%), Flavobacteriia (9.3%), Alphaproteobacteria (3.0%), and Mollicutes (1.5%); the five most abundant classes on the carapace surface were Epsilonproteobacteria (42.2%), Alphaproteobacteria (16.9%), Gammaproteobacteria, Flavobacteriia (10.6%), and Saprospiria (4.7%; Figure S1A ). At the genus level, the three most abundant genera in the gills were Sulfurovum (67.0%), Leucothrix (3.1%), and Leeuwenhoekiella (2.7%), while the five most abundant genera on the carapace surface were Sulfurovum (28.1%), Leucothrix (9.5%), Polaribacter (3.5%), Ruegeria (2.6%), and Arcobacter (2.5%; Figure S1B ).

Within the five most abundant classes, a few representative species were evaluated in terms of relative abundance. The box and whisker plot shows the mean and variation of each bacterial species ( Figure 3A ). The relative abundances of Sulfurovum sp. NBC37-1, Leewuwenhoekiella sp. MAR_2009_32, and Ornithobacterium rhinotracheale were significantly higher in the gills (67.0%, 2.7% and 1.3%, respectively) than on the carapace surface (28.0%, 0.1% and 0.3%, respectively; Figures 3A , S2 ). In contrast, on the carapace, Nitratifactor salsuginis (2.9%), Arcobacter butzleri (2.1%), Leucothrix mucor (9.5%), Hydrogenvibrio sp. Milos-T1 (1.2%), Polaribacter sp. Hel1_33_49 (1.7%), Ruegeria sp. TM1040 (1.6%), and Pseudovibrio sp. FO-BEG1 (2.5%) were more abundant than in the gills (respectively 0.7%, 0.4%, 3.1%, 0.1%, 0.2%, 0.2% and 0.1%; Figure 3A ; Figure S2 ).

Indices of alpha diversity, including Chao 1, Shannon, and Simpson, were obtained to characterize the richness and evenness of gill and carapace samples at the species level. The results showed that the carapace surface had more species (Chao 1) and higher diversity (Shannon and Simpson) compared to the gills ( Figure 3B ). Since the gills were profoundly dominated by Sulfurovum sp. NBC37-1 (67%), it is entirely reasonable that the bacterial diversity in the gills would be lower than that on the carapace surface. Additionally, according to the PCoA based on the Bray-Curtis distance, the bacterial species compositions of the gills and the carapace surface were significantly different (PERMANOVA, pseudo-F = 31.2942, p = 0.0001; Figure 4A ). In the gills of X. testudinatus, a total of 80 genera and 91 species were identified (≥ 0.1% abundance). On the carapace surface, a total of 136 genera and 167 species were identified (≥ 0.1% abundance; Figures 4B, C ). The Venn diagram analysis showed that the gills and carapace surface shared 58 genera and 64 species. There were 22 and 78 genera specific to the gills and carapace surface, respectively ( Figure 4B ). In the gills, the five most abundant specific species were Candidatus Hepatoplasma, Thiofilum, Lacinutrix, Chryseobacterium, and Thiolinea. On the carapace surface, they were Deinococcus, Cytophga, Truepera, Rhodovulum, and Rhodopirellula. In addition, 27 and 103 species were specific to the gills and carapace surface, respectively ( Figure 4C ). The five most abundant specific species in the gills were Candidatus Hepatoplasma crinochetorum, Thiofilum flexile, Lacinutrix sp. 5H-3-7-4, Thiolinea disciformis, and Methylomicrobium buryatense. On the carapace surface, Deinococcus peraridilitoris, Flavobacteria bacterium MS024-2A, Rhodovulum sulfidophilum, Deinococcus deserti, and Rubinisphaera brasiliensis were the five most abundant species ( Figure 4C ).

According to FAPROTAX analysis, several functional groups were predicted and identified based on the bacterial taxa profiles ( Figure 5 ). Only two functional groups, methylotrophy (p = 0.3287) and hydrocarbon degradation (p = 0.3401), were not significantly different between the gill and the carapace samples. In the gills, there were higher proportions of bacteria involved in sulfur oxidation, hydrogen oxidation, and denitrification. On the other hand, the community on the carapace surface had higher functional proportions of sulfur reduction, nitrogen fixation, ammonification, chemoheterotrophy, phototrophy, cellulolysis, iron respiration, fumarate respiration, fermentation, ureolysis, and parasites, pathogens or symbionts. The different bacterial functions present in the gills and the surface of the carapace provide critical information for understanding how bacterial biochemical processes may be involved with the response of the vent crab to the HV environment.

Using FISH, it was possible to spatially identify Sulfurovum-related bacteria in the gills of X. testudinatus ( Figures 6 , S3 ). The signals from Sulfurovum-targeting probe EPI653 almost completely overlapped with those from universal bacteria probe EUB338, indicating that the signals were from bacteria and further supporting our finding that the most abundant genus in gill samples is Sulfurovum ( Figure S3 ). Furthermore, the bacteria labeled with a customized probe designed for Sulfurovum sp. NBC37-1 were widely distributed in the afferent vessels of vent crab gills, and these signals also overlapped with EPI653 signals ( Figure 6 ). Moreover, our microscopy analysis revealed that the Sulfurovum-related bacteria almost occupied the cells within the afferent vessel ( Figures 6C–E ), and fluorescence signals were also observed in the epithelial principal cells and pilaster cells on the lamella ( Figures 6A, B , S3 ).

We analyzed the bacterial compositions in the gills and carapace surface of shallow-water HV crab X. testudinatus. Notably, we identified Epsilonproteobacteria as the dominant class in both gills (73.8%) and carapace (42.2%). According to culture-based characterizations and genetic analyses, Epsilonproteobacteria are associated with the utilization of reduced sulfur, essentially as chemolithoautotrophs (Takai et al., 2003; Inagaki et al., 2004; Takai et al., 2005). Epsilonproteobacteria are also known to be a major class of symbiotic bacteria in other HV endemic creatures, such as yeti crabs (Kiwa hirsute), galatheid crabs (Shinkaia crosnieri) and vent gastropods (Cyathermia naticoide) (Goffredi et al., 2008; Tsuchida et al., 2011; Zbinden et al., 2015). In aquatic animals, the gills serve as an interface between the external environment and internal body fluid, playing primary roles in respiration, osmolyte transport, nitrogenous waste excretion and acid-base regulation (Henry and Wheatly, 1992; Evans et al., 2005). Despite this broad range of functions, few studies have investigated the physiological interactions between animal hosts and symbiotic bacteria in the gills. Sulfurovum, a genus of Epsilonproteobacteria, was especially well represented in the bacterial communities of the gills (67%) and the carapace surface (28%). Furthermore, Sulfurovum were abundantly labeled in the gills of X. testudinatus by FISH. As a SOB, Sulfurovum is likely to oxidize reduced sulfur (sulfide, S^2-) into non-toxic forms (elemental sulfur S, sulfite S O 3 2 − , and sulfate S O 4 2 − ), mitigating the toxicity of hydrogen sulfide (H₂S) in hydrothermal vent fluids (Nelson et al., 1995; Akerman et al., 2013).

We identified a few bacterial classes in X. testudinatus that are also present in the ambient HV environment. These HV-associated classes include Epsilonproteobacteria, Gammaproteobacteria and Alphaproteobacteria, which have been identified in the environmental waters or fluids of either deep-sea (Adams et al., 2011; Yamamoto and Takai, 2011) or shallow-water HV systems (Giovannelli et al., 2013; Yang et al., 2016). In the sulfur-rich environment, the most abundant class was Epsilonproteobacteria, a biologically and geologically significant bacterial class (Campbell et al., 2006). In line with these findings, Epsilonproteobacteria were also found both in the surrounding water and sediments of the shallow-water HV system located off the coast of Panarea Island in Italy (Maugeri et al., 2009). Phylogenic analysis also revealed a large Sulfurovum-related cluster in the shallow-water HV fields of Milos Island in Greece (Giovannelli et al., 2013). Similar results were reported for water and sediment samples collected from the Kueishan Island vent area, which were dominated by Sulfurovum (Wang et al., 2015; Yang et al., 2016; Wang et al., 2017). According to stable isotope analysis, the bacteria samples from the vent crab’s carapace and the surrounding water did not differ in ¹³C and ¹⁵N levels, which might be potentially cognate (Wang et al., 2022). Additionally, plenty of white-colored biofilms are growing on the biotic (crab carapace, gastropod shell, algae) and abiotic surfaces (sediment, rock, or artificial fragment) in the underwater environments of Kueishan Island, which are usually fed by HV metazoans, including X. testudinatus. Since the vent crab is a benthic crustacean, its bacterial community on the carapace surface could be decided by the environments, particularly sulfur-depositing sediments. The Kueishan Island HV system is a vent area exposed to the intermittent discharge of vent fluids. We found extremely high concentrations of sulfide in the vent area, with dynamically dropping DO concentrations and a great accumulation of sulfide in the vent crab’s hemolymph, which is also observed in marine soft sediments-inhabited priapulid worms (Halicryptus spinulosus) under anoxic conditions (Oeschger and Vetter, 1992). Thus, it is reasonable that a facultative symbiont of Epsilonproteobacteria-related chemoautotrophic SOB Sulfurovum and a sulfur-resistant species X. testudinatus would be found in such sulfidic extreme environment.

In our surveys, we identified the dominant bacteria species as Sulfurovum sp. NBC37-1 in both the gills and carapace samples. This strain was previously isolated from the Iheya North field and had 99.1% similarity (16S rRNA) with a strict SOB (Sulfurovum lithotrophicum 42BKT) that was also isolated from the Okinawa Trough (Nakagawa et al., 2005a; Nakagawa et al., 2007). The high functional proportions of sulfur oxidation in our study were mainly attributable to the predominance of Sulfurovum sp. NBC37-1. According to our FISH results, the presence of Sulfurovum sp. NBC37-1 within the afferent vessel, epithelial principal cells and pilaster cells in the gills suggests an endosymbiont relationship between the SOB and the vent crab. The widespread occurrence of the SOB in X. testudinatus gills is reminiscent of findings made in the gills of vesicomyid clams (Cruaud et al., 2019), mytilid mussels (Halary et al., 2008; Szafranski et al., 2015; Duperron et al., 2016), and bresiliid shrimps (Tokuda et al., 2008) from hydrothermal vent habitats. Since the gills serve as a boundary between an aquatic animal’s internal organs and the external environment, these tissues appear to be an enclave for SOBs to carry out sulfide detoxification and organic production for the hosts (Felbeck and Somero, 1982). Likewise, the high functional proportion of hydrogen (H₂) oxidation in the gills was probably due to Sulfurovum sp. NBC37-1 with hydrogenase (Nakagawa et al., 2007). It was previously reported that H₂ can also serve as an energy source in hydrothermal vent systems by powering chemosynthetic symbioses (Petersen et al., 2011). In the denitrification process, Sulfurovum may use nitrate ( N O 3 − ) as an electron acceptor during sulfur oxidation. Then, Nitratifractor salsuginis can transfer electrons from H₂ into N O 3 − to generate nitrogen (N₂) (Inagaki et al., 2004; Nakagawa et al., 2005b; Vetriani et al., 2014). Intriguingly, the amino acid sequences of nitrate reductase (NapA) retrieved from a diffuse-flow HV area on the East Pacific Rise were related to Sulfurovum sp. NBC37-1 (44%), Arcobacter butzleri (14%) and N. salsuginis (7%), suggesting that these species may be involved in denitrification or ammonification (Smith et al., 2007; Vetriani et al., 2014). Notably, sulfur reduction and fumarate respiration were more active on the vent crab carapace surface, which had higher relative abundances of Wolinella succinogenes (another Epsilonproteobacterium). Previous studies demonstrated that W. succinogenes could oxidize H₂S or bisulfide (HS^-) coupled with fumarate respiration to form polysulfide, which is then reduced by polysulfide reductase in a periplasmic sulfur reduction process (Macy et al., 1986; Krafft et al., 1995; Hedderich et al., 1998; Kroger et al., 2002). Moreover, W. succinogenes contains a periplasmic NapA and a cytochrome c3 that are respectively involved in nitrate and nitrite reduction as an ammonification strategy, the functional proportion of which was higher on the carapace surface of the vent crab (Bokranz et al., 1983; Blackmore et al., 1986; Lovley et al., 1999; Simon, 2002; Simon et al., 2003; Strohm et al., 2007).

Among Gammaproteobacteria, Leucothrix mucor was the dominant species, as it had a higher relative abundance on the carapace surface than in the gills. Previous studies have shown that the presence of L. mucor on the surfaces of marine algae, fish eggs and aquatic invertebrates may occur as an infestation (Johnson et al., 1971; Payne et al., 2007). This filamentous heterotrophic bacterium is also known to be a sulfur oxidizer that utilizes reduced sulfur compounds as energy sources (Grabovich et al., 1999). The Gammaproteobacteria-related symbionts of vent shrimps (R. exoculate) were also dominated by L. mucor (Petersen et al., 2010). Another Gammaproteobacterium, Psychromonas ingrahamii was also more abundant on the vent crab carapace. Genomics analysis of P. ingrahamii revealed several reductases for nitrate, nitrite, sulfite, fumarate and fermentation-related enzymes, while culture studies further demonstrated that P. ingrahamii is a psychrophilic extremophile with the ability to perform nitrate reduction, fumarate utilization and fermentation (Auman et al., 2006; Riley et al., 2008; Markowitz et al., 2012). However, P. ingrahamii have mostly been isolated or identified from low-temperature marine environments, such as sea ice and polar sediments (Breezee et al., 2004; Auman et al., 2006; Canion et al., 2013). To the best of our knowledge, this is the first study to identify P. ingrahamii in a hydrothermal HV system. The Hydrogenvibrio sp. Milos-T1, which also had a higher abundance on the carapace, was previously isolated from a shallow-water HV system near the island of Milos in Greece and was closely related to Hydrogenovibrio marinus, an extremophilic hydrogen-oxidizing bacterium (Nishihara et al., 1997; Brinkhoff et al., 1999). In addition to hydrogen, H. marinus utilizes reduced sulfur compounds as electron donors, and it uses ammonium and urea as sole nitrogen sources, but not N₂, nitrate or nitrite (Nishihara et al., 1991). The fact that ureolysis is performed much more actively on the carapace may be associated with a higher abundance of Hydrogenvibrio sp. Milos-T1.

Among Alphaproteobacteria, two species were identified. Pseudovibrio sp. FO-BEG1 and Ruegeria sp. TM1040, both of which were more abundant on the carapace. It was precisely the same that Pseudovibrio sp. FO-BEG1 was shown to be capable of sulfur oxidation, denitrification and fermentation as Pseudovibrio denitrificans, a representative species of the genus (Shieh et al., 2004; Romano, 2018). There is evidence that Psedovibrio-related bacteria may exhibit symbiotic relationships with several invertebrates, such as corals, sponges and tunicates (Fukunaga et al., 2006; Riesenfeld et al., 2008; Rypien et al., 2010; Santos et al., 2010; Penesyan et al., 2011). Nevertheless, this appears to be the first study to identify Psedovibrio in a crustacean. In the case of Ruegeria sp. TM1040 (formerly Silicibacter sp.), which is typically found in the phycosphere or on the surface of dinoflagellates, genomic analysis showed no relevant genes for sulfur oxidation, denitrification or ammonification, but dimethulsulfonioproponate (DMSP) demethylase was shown to metabolize DMSP which is produced by dinoflagellates and comprises a major source of organic sulfur in the ocean (Moran et al., 2007; Belas et al., 2009).

Flavobacteriia are known as polymer degraders and are found in a wide range of environments, including seawater and sediments (Alonso et al., 2007; Kappelmann et al., 2019). The relative abundances of Leeuwenhoekiella sp. MAR_2009_132 and Ornithobacterium rhinotracheale were both higher in the gills of X. testudinatus than on the carapace surface. Among these two constituents, Leeuwenhoekiella sp. MAR_2009_132 was previously isolated near the Island Sylt in the North Sea and was closely related to Leeuwenhoekiella aequorea (Hahnke and Harder, 2013; Chen et al., 2018b). Surprisingly, O. rhinotracheale is generally known as a species that causes respiratory disease in poultry (Barbosa et al., 2019). In our literature searches, we found no studies on this species related to marine organisms. Only two studies described the detection of O. rhinotracheale in seawater samples, but each provided limited information (Valenzuela-Gonzalez et al., 2016; Porchas-Cornejo et al., 2017). On the other hand, Polaribacter spp. and Tenacibaculum maritimum were more abundant on vent crab’s carapace. Polaribacter sp. Hel_33_49 and Hel_I_88 (also named Hel1_88) were first isolated from the North Sea during a spring bloom and were reported to feed on polysaccharides and proteins with carbohydrate-active enzymes (CAZymes) and peptidases (Zerillo et al., 2013; Xing et al., 2015). In the case of T. maritimum, a prominent pathogen without known host specificity, genomic analysis revealed similar metabolism preferences (saccharide and protein degradation) to those seen in other Flavobacteriia-related bacteria, such as Polaribacter spp. (Fukunaga et al., 2006; Perez-Pascual et al., 2017). The general consensus is that Flavobacteriia subsist on abundant organic matter in the epipelagic zone, which may explain why deep-sea HV systems usually show no or very low proportions of Flavobacteriia-related bacteria in contrast to the shallow-water HV system of Kueishan Island (Goffredi, 2010; Salazar et al., 2016; Cerqueira et al., 2017; Leinberger et al., 2022). Nevertheless, Flavobacteriia were not abundant in other shallow-water HV systems (Rajasabapathy et al., 2018). Since Kueishan Island lies to the southern rifting end of the Okinawa Trough (a back-arc depression of the Ryukyu arc-trench system), the active volcanic features cause unstable hydrothermal activities to occur in the geologically “young” vent system (Konstantinou et al., 2013; Deffontaines et al., 2022). The fluids spilling out of the vents around this young volcanic island aperiodically alter the surrounding water chemistry beneath the sea surface. In addition, frequent weather and current disturbances of typhoons, monsoons and the Kuroshio may cause strong turbulent mixing in Kueishan Island HV system (Liang et al., 2003; Lebrato et al., 2019). These climate features further contribute to the establishment of the marine ecosystem near Kueishan Island, off the northeastern coast of Taiwan. The combination of climate, status as a neritic zone (providing sufficient organic matters) and presence of an extreme HV system (emitting minerals, sulfur compounds, high temperature and low pH) make Kueishan Island a truly unique natural laboratory.