Bacterial Community Associated with Organs of Shallow Hydrothermal Vent Crab Xenograpsus testudinatus near Kuishan Island, Taiwan

Shallow-water hydrothermal vents off Kueishan Island (northeastern Taiwan) provide a unique, sulfur-rich, highly acidic (pH 1.75–4.6) and variable-temperature environment. In this species-poor habitat, the crab Xenograpsus testudinatus is dominant, as it mainly feeds on zooplankton killed by sulfurous plumes. In this study, 16S ribosomal RNA gene amplicon pyrosequencing was used to investigate diversity and composition of bacteria residing in digestive gland, gill, stomach, heart, and mid-gut of X. testudinatus, as well as in surrounding seawater. Dominant bacteria were Gamma- and Epsilonproteobacteria that might be capable of autotrophic growth by oxidizing reduced sulfur compounds and are usually resident in deep-sea hydrothermal systems. Dominant bacterial OTUs in X. testudinatus had both host and potential organ specificities, consistent with a potential trophic symbiotic relationship (nutrient transfer between host and bacteria). We inferred that versatile ways to obtain nutrients may provide an adaptive advantage for X. testudinatus in this demanding environment. To our knowledge, this is the first study of bacterial communities in various organs/tissues of a crustacean in a shallow-water hydrothermal system, and as such, may be a convenient animal model for studying these systems.


Introduction
A deep-sea hydrothermal vent is one of the most extreme environments on earth, due to its poorly oxygenated, oligotrophic and toxic ecosystem [1]. In such ecosystems, chemolithotrophic bacteria are common residents [2]. Bacteria associated with host animals (e.g., Crustacea), are believed to support their hosts and enable them to adapt to their extreme environment, including high toxicity and limited nutrients [3][4].
Shallow-water hydrothermal vents are usually near active coastal or submarine volcanoes and also provide an oligotrophic and toxic environment for animals and microorganisms [1]. The crab Xenograpsus testudinatus predominates in shallow-water, sulfur-rich/highly acidic hydrothermal (pH 1.75-4.6) [13] vents near Kueishan Island, northeastern Taiwan. This crab is one of the few known vent-endemic species at depths < 200 m [13]. Unlike the deep-sea hydrothermal system, there is no chemolithoautotrophic food-web in the shallow-water hydrothermal vent off Kueishan [13]. Because the biodiversity of shallow-water hydrothermal vent is relatively low compared to deep-sea hydrothermal vents, X. testudinatus has a unique opportunistic feeding style; when the current is weak, they eat zooplankton killed by the vent's sulfurous plumes [13]. Despite some recent studies regarding nutrient acquisition by X. testudinatus, there are apparently few reports characterizing X. testudinatus-associated bacteria.
Due to differences in food-webs between shallow-water and deep-sea hydrothermal systems, Crustacea-associated bacteria in a shallow-water thermal vent might differ from those in a deep-sea hydrothermal vent. Therefore, the objective was to characterize the bacterial community in organs of X. testudinatus in the vicinity of sulphur-rich hydrothermal vents in shallow water near Taiwan.

Sampling site and sample collection
The sampling site was located near Kueishan Island (121°57'E, 24°50'N), Taiwan. The dominant species Xenograpsus testudinatus and seawater samples were collected at the hydrothermal venting area, approximately 8-20 m from the island [13]. The surface of X. testudinatus is covered by a filamentous biofilm [14]. Two crabs (one male and one female) were collected (SCUBA-diving) in depths ranging from 10 to 15 m, in May 2009. After sampling, crabs were kept in an aerated cooling box, directly transported to the laboratory, and frozen at -20°C before assaying (interval from collection to freezing was < 12 h). A seawater sample (50 ml) was also collected from the sampling place of the shallow-water hydrothermal vent. The target species X. testudinatus in this study is not listed as endangered or protected and was not collected from national parks or natural reserves in Taiwan, thus no specific permission was required for sampling.

Total DNA extraction
The entire dissection procedure was performed in a biological safety cabinet. Before dissection, crabs were washed twice with sterile seawater. All dissection instruments were sterilized over an open flame to eliminate residual DNA and washed with 75% ETOH to prevent cross-contamination. After removing upper carapaces, the digestive gland, gill, stomach, heart and midgut from each crab were excised for DNA extraction (Fig 1). Total genomic DNA was extracted according to a modified standard phenol-chloroform procedure incorporating a grinding step in liquid nitrogen to mechanically lyse cells [15]. After extraction, DNA samples of various body parts were denoted as D (digestive gland), S (stomach), H (heart), G (gill), and M (midgut). Bacterioplankton from seawater samples were filtered on cellulose acetate membranes (pore size, 0.2 μm; ADVANTEC, Tokyo, Japan). Microbial biomass was removed from membranes by washing with TE buffer (50 mM Tris-HCl, 1 mM EDTA, pH 8.0). Suspensions were collected in microtubes and total DNA was extracted by the same method as described. The DNA from seawater samples was denoted as SW. All DNA samples were stored at -20°C.

DNA tagging PCR and massively parallel pyrosequencing
The DNA tagging PCR (five cycles) was used to tag 5 0 ends of each V6-V8 amplicon [16]. For this, DNA amplified from the same parts of male and female crabs were pooled equally into a sample to provide a comprehensive bacterial community survey and ignore gender variation. Pooled lots of each tagged V6-V8 DNA samples (final DNA content~0.5 μg) were sent to Mission Biotech (Taipei, Taiwan) for massively parallel pyrosequencing using a Roche 454 GS FLX Titanium System. A total of 37157 sequences were generated and sorted. Novel bacterial community sequences were deposited in GeneBank (SRA accession PRJNA296845).

Data analysis
After trimming, chimerical reads were detected by UCHIME (http://drive5.com/uchime) [18]. For operational taxonomic units (OTU) analysis, quality-filtered reads were pooled together and analyzed with the UPARSE pipeline [19] except the chimera removal step. In UPARSE, de-replication was done with the options "-derep_prefix" and "-minsize 2" and OTUs were generated at 97% identity. Each OTU was searched (with global alignment) against the Greegenes 13_5 database to find corresponding taxonomy of the best hit using USEARCH. Alignment was calculated using MOTHUR software to define operational taxonomic units (OTUs) with a 3% cut off value in sequence dissimilarity. Defined OTUs were used to estimate the Shannon-Weaver diversity index, Chao1 estimator, and the Simpson index, as well as to construct rarefaction curves, and to calculate evenness and richness of the bacterial community. Unclassified OTUs were manually searched in NCBI using BLASTn.
Relative abundances of each classified bacterial classes in individual samples were incorporated into a matrix, with Bray-Curtis used to estimate a distance matrix. Results were presented in hierarchical clustering (CLUSTER) and non-Metric multi-Dimensional Scaling (nMDS) using Primer 6 software (PRIMER-E, Lutton, Ivybridge, UK) to analyze to relationships between bacterial communities among various samples.

Quantitative PCR
For the quantitative PCR assay, two primer sets were designed to target the dominant group OTU1 and OTU7, which were representative of Epsilonproteobacteria and Gammaproteobaceria respectively, by using Primer-BLAST tool on BLAST search (NCBI): Episilon-F (5'-CTC GTGTCGTGAGATGTTGG-3') and Episilon-R (5'-GCGAAGGCAGTCTCACTAGA-3'), Gamma-F (5'-TGTTGGGTTAAGTCCCGCAA-3') and Gamma-R (5'-CACCTTCCTCCGG TTTGTCA-3'). The hypervariable V6 region for 16S rRNA gene was used as an endogenous control, using a pair of primers: 967F (5'-CAACGCGAAGAACCTTACC-3') and 1046R (5'-CGACAGCCATGCANCACCT-3') [16]. Quantitative PCR reaction was performed in an ABI 7300 Real-Time PCR system. The PCR program consisted of 2 min at 95°C, 45 cycles of 15 s at 95°C, and followed by 30 s at 60°C, using Platinum SYBR Green qPCR SuperMix-UDG (Invitrogen, Waltham, MA, USA). A dissociation stage was performed to confirm the specificity of the product and to avoid the production of primer dimers; the parameters consisted of 15 s at 95°C, 1 min at 60°C, 15 s at 95°C, and 15 s at 60°C. For all reactions, 2μl (10 ng) of template DNA was added to a reaction of 20μl. These reactions contained 10 μl of Platinum SYBR Green qPCR SuperMix-UDG (Invitrogen), 3.8 μl of sterilized nuclease-free water, 0.4 μl each of the forward and reverse primers, 0.4μl of 1X ROX Reference Dye, and 2 μl of DNA template. Each sample was performed in duplicate.

Results
The V6 to V8 hypervariable regions of bacterial 16S ribosomal RNA genes were surveyed to determine bacterial community composition in various crab body parts/tissues (Fig 1). A total of 37157 sequences were generated; the number of V6-V8 sequences in each sample varied from 5287 to 7253 (Table 1). Rarefaction curves approached a plateau (S1 Fig), which indicated sequences adequately represented bacterial composition in crab samples. Shannon's index of crab samples ranged from 0.866 to 2.184, whereas the surrounding seawater sample was 3.065. The Chao1 index ranged from 66 to 192 in crab samples and 246 in seawater. There was much lower bacterial diversity in crab than surrounding seawater. Within crabs, Shannon index of bacterial communities was least in the stomach, but greatest in the mid-gut.
At the phylum level, bacterial composition in the seawater and crab were similar, but seawater-associated bacterial communities were more diverse than any sample of crab (Fig 2). Epsilonproteobacteria was the most dominant bacterial class, accounting for 53.94% of total sequences in seawater. In crab samples (excluding heart), Epsilonproteobacteria was also the most dominant bacterial class in stomach (80.18%), gill (85.20%), and digestive gland (76.18%). There were a total of 221 OTUs in the study. Among total OTUs, five of the top 10 OTUs affiliated to Epsilonproteobacteria (Fig 3). Four of these epsilonproteobacterial OTUs (OTU1, 3, 6, and 24) were Sulfurovum-related sequences, clustered within Marine Group 1 (Fig 4) [20]. However, only OTU4 was a Sulfurospirillum-related sequence (Fig 4). Relative abundance of OTU1 was highest in samples of crab, but relatively low in seawater. In contrast, OTU6, the OTU with the highest relative abundance in the seawater, was relatively low in crab samples. Compared to the most abundant Gammaproteobacteria OTU (OTU7), relative folds of the most abundant Epsilonproteobacteria OTU (OTU1) in each sample obtained by qPCR corresponded to relative abundance of bacterial composition obtained by sequencing, which also supported dominance of Epsilonproteobacteria (S2 Fig). In qPCR, Epsilonproteobacteria was 140.45 and 157.51 times higher than Gammaproteobacteria in stomach and digestive gland, respectively (S2 Fig), and only 1.97 in seawater and 1.08 times in mid-gut. Similarly, in relative abundance of bacterial sequences, Epsilonproteobacteria was 443.23 times higher than Gammaproteobacteria in stomach and 284.21 times higher in digestive gland respectively, but only 2.46 times in seawater and 2.10 times in mid-gut (Fig 2).
Although Gammaproteobacteria was not as dominant as Epsilonproteobacteria, it was dispersed widely in all samples, particularly in seawater (21.94%) and mid-gut (23.65%). However, further comparison revealed distinct differences between crab parts and seawater with regards to Gammaproteobacteria. The dominant gammaproteobacterial OTU (OTU7) was most closely related (99% identity) to the phylotype reported in R. exoculata [7]. This OTU was dominant in digestive gland, gill, heart, mid-gut, and seawater. Additionally, in the mid-gut, and heart, two OTUs that belonged to Vibrio were also dominant. In the stomach, there was limited Gammaproteobacteria, whereas a Methylomonas-related sequence was only detected in seawater.  Alphaproteobacteria was also not as dominant as Epsilonproteobacteria, but it was detected in every sample (Fig 2). In particular, OTU5, which was affiliated to Alphaproteobacteria (86% in sequence identity), was only retrieved in the digestive gland, whereas OTU9, which was affiliated to Roseovarius tolerans (98% in sequence identity), mainly existed in the stomach of X. testudinatus (Fig 3).
In the heart sample, OTU2, which was affiliated to the class Mollicutes based on blast against the database at NCBI (sequence identity of 89%), was most dominant (relative abundance of 79.65%; Fig 2). Although OTU2 was also present in stomach, mid-gut and digestive gland, it was present in heart, but never retrieved in seawater (Fig 3).
Based on the clustering result (Fig 5), crab-associated bacterial communities appeared to be clustered together, which was distinctly different from the seawater bacterial community. In addition, bacterial composition of gill and mid-gut were more similar to each other, whereas digestive gland and stomach not sure congregated in clustering analysis.

X. testudinatus associated Gamma-and Epsilonproteobacteria
The major bacterial group associated with inside X. testudinatus was Epsilonproteobacteria, a common bacterial group in hosts of deep sea hydrothermal systems [21]. Since there was much less diversity of bacteria in crabs than in sea water in this study, we inferred that there was potential host specificity of Epsilonproteobacteria in X. testudinatus. Furthermore, although composition of bacterial classes in both sea water and crab were dominated by Epsilonproteobacteria, the dominant epsilonproteobacterial OTUs in crab and sea water differed. Notably, OTUs of Epsilonproteobacteria had distinct preferences to host and seawater. In this regard, OTU1, which was predominant in X. testudinatus, was closely related to Epsilonproteobacterum ectosymbiont of deep sea hydrothermal vent shrimp Rimicaris exoculata (i.e., 99% in sequence identity); however, OTU6, closely related to genus Sulfurovum sp., was only dominant in seawater. Epsilonproteobacteria were widely present in other crustaceans dwelling in hydrothermal vent, including crabs (Kiwa hirsute and Shinkaia crosnieri) [8,10] and shrimp (Rimicaris exoculata) [20]. A symbiotic relationship between bacterial epibions and Rimicaris exoculata has been proposed [5][6][7], and a mutualistic relationship between R. exoculata and epibions was reported [11]. Epsilonproteobacteria in R. exoculata might be capable of autotrophic growth by oxidizing reduced sulfur compounds [11,20]. In this regard, Epsilonproteobacteria is able to assimilate inorganic carbon and directly transfer nutrients to the host R. exoculata through the shrimp integument rather than via the digestive tract [11]. In addition to Epsilonproteobacteria, there was also Thiotrichaceae, a chemolithoautotrophic sulfur-oxidizing Gammaproteobacteria frequently recovered from various deep-sea invertebrates of hydrothermal environments The bacterial community might also benefit the host by detoxification of potential metabolic inhibitors in deep-sea hydrothermal vent environments [6]. For example, by converting H 2 S to elemental sulfur by Gamma-and Epsilonproteobacteria, Rimicaris exoculata would benefit from access to less toxic products [6]. Water around shallow-water hydrothermal vents in Kueishan Island contains considerable carbon dioxide, nitrogen, oxygen, sulphur dioxide, hydrogen sulphide [24] and other metal ions, including Mg 2+ , Ca 2+ , Fe 2+ , Cu 2+ , Al 3+ and Mn 3+ [25], making this a relatively toxic environment for X. testudinatus and its prey [23]. Therefore, perhaps X. testudinatus associated Gamma-and Epsilonproteobacteria have an essential role for their host, namely detoxification. Some Epsilonproteobacteria may be autotrophic denitrifiers in the hydrothermal system [26]. Among the Epsilonproteobacteria in the deep sea hydrothermal environment, R. exoculata related epibiontic Epsilonproteobacteria have potential to produce ammonium. Therefore, Epsilonproteobacteria might uptake ammonium from the environment or host when they are free living [6]. Perhaps X. testudinatus associated Epsilonproteobacteria derive benefits from the host, similar to R. exoculata-associated Epsilonproteobacteria.
Bacterial diversity and potential organ specificity bacteria in X. testudinatus Bacterial diversity in the gill seems to be high, because gill is the first organ in direct contact with water and suspended particles. However, in this study, bacterial diversity in mid-gut was higher than that in gill. One potential reason is accumulation of trace metals in gill; concentrations of dissolved metals may be increased by the acidic environment and high temperature, accumulating in the gill of X. testudinatus by respiration [27]. Concentrations of Al, Cd, Co, Ni, Cu, and Zn were higher in gill than in exoskeleton, hepatopancreas or muscle [27]. Although the relationship between bacteria and the accumulated metals was not clear, those specific metals might influence composition of bacteria in the gill of X. testudinatus. In addition, Hu et al. [23] reported that enzymes, including trypsin-like enzymes and proteinases, in the mid-gut glands of X. testudinatus had significantly greater thermal tolerance. The microenvironment for the activity of these enzymes might also influence bacterial community structure in the mid-gut of the host.
Interestingly, unlike other parts of crabs, the predominant bacterial OTU (OTU2) in the heart was Mollicutes (in the phylum Tenericutes), not Gamma-or Epsilonproteobacteria. The class Mollicutes is classified by an absence of a cell wall. Many Mollicutes are parasites of insects, plants and crustacean [28]. In previous studies, OTUs belong to Mollicutes were present in the gut of hydrothermal shrimp Rimicaris exoculata [5] and crab X. testudinatus [29]. Because this OTU was mainly in the heart, we inferred that this was a different microenvironment from other parts of crab, and also organ specificity for X. testudinatus-related bacteria. Although it is difficult to get further taxonomic information regarding OTU2 and the potential relationship of these bacteria and host, based on heart-associated bacteria discovered in this study, we speculated that there might be a specific role for these bacteria in X. testudinatus.
Alphaproteobacterial OTUs were only detected in the stomach or digestive gland of X. testudinatus. Although further taxonomic information of OTU5 was not readily available, OTU5 which was only present in the digestive gland was close to an alphaproteobacterial sequence from the digestive gland of European shore crab Carcinus maenas [30]. Perhaps the digestive gland of X. testudinatus provided a unique environment for these bacteria. In addition, Roseobacter is usually present in macroalgae [31] and marine snow [32]. In this study, OTU9, which prevailed in the stomach of X. testudinatus, was affiliated to a strain of Roseovarius tolerans from macroalgae [33]. Perhaps these were algae-or marine snow-associated bacteria and eaten by the host.
Vibrio comprises more than 110 recognized species, of which many are pathogenic to humans and animals [34][35], although some isolates have probiotic properties [36]. Recently, the presence of virulence genes in a deep-sea hydrothermal vent Vibrio species was reported [37]. Those virulence genes were homologs of those in Vibrio species pathogenic to humans. Thus, it has been suggested that these pathogenicity genes serve a far more basic ecological role than only causation of human disease [37]. For example, in symbiosis between the squid Euprymna scolopes and luminous bacterium Vibrio fischeri [38], this bacterium helps the host by producing light in the organ, forming counterillumination to avoid predators. In the present study, these OTUs, which affiliated to Vibrio, were only present in the heart and mid-gut of X. testudinatus, and in seawater. Although the specific functions, pathogenicity, or probiotic properties of X. testudinatus associated Vibrio are unknown, the presence of Vibrio in the healthy host of shallow-water hydrothermal system in this study might provide impetus for further study regarding evolution of Vibrio and its genomic events.

Conclusions
Chemoautotrophic Gamma-and Epsilonproteobacteria were recovered from several crabs around hydrothermal vents and cold seep habitats. Although roles of Gamma-and Epsilonproteobacteria in X. testudinatus were not clear, we inferred that in these shallow-water, sulfurrich/highly acidic hydrothermal vents, X. testudinatus acquired energy not only based on opportunistic feeding (consuming dead zooplankton killed by vent's sulfurous plumes), but also from nutrient transfer from probable symbiotic Epsilonproteobacteria and/or Gammaproteobacteria.
Kueishan Island is subjected to the Kuroshio Current, one of the world's fastest currents [39]. The Kuroshio Current is low in nutrients and plankton populations, but warm with rapid flow, running from tropical Philippines through subtropical Taiwan and Okinawa to the temperate region of Japan, transferring heat from equatorial regions to higher latitudes [40][41][42][43]. Thus, due to differences between environments of deep-sea and shallow-water hydrothermal vents, the way to obtain nutrients by crustaceans in shallow-water hydrothermal vent might be more complicated than in deep-sea hydrothermal vents, including chemosynthetic and photosynthetic primary productions and a heterotrophic system. Versatile ways to obtain nutrients may confer an adaptive advantage for the crab X. testudinatus in the shallow hydrothermal vent, and account for X. testudinatus being dominant in this environment. In addition, our study revealed a potential symbiotic relationship between X. testudinatus and X. testudinatus associated Gamma-and Epsilonproteobacteria, which may be similar to the relationship between deep-sea hydrothermal vent crustaceans and Gamma-and Epsilonproteobacteria bacteria. Therefore, X. testudinatus may provide a convenient animal model for hydrothermal systems.