Microbial Community Stratification Controlled by the Subseafloor Fluid Flow and Geothermal Gradient at the Iheya North Hydrothermal Field in the Mid-Okinawa Trough (Integrated Ocean Drilling Program Expedition 331)

Abstract

The impacts of lithologic structure and geothermal gradient on subseafloor microbial communities were investigated at a marginal site of the Iheya North hydrothermal field in the Mid-Okinawa Trough. Subsurface marine sediments composed of hemipelagic muds and volcaniclastic deposits were recovered through a depth of 151 m below the seafloor at site C0017 during Integrated Ocean Drilling Program Expedition 331. Microbial communities inferred from 16S rRNA gene clone sequencing in low-temperature hemipelagic sediments were mainly composed of members of the Chloroflexi and deep-sea archaeal group. In contrast, 16S rRNA gene sequences of marine group I Thaumarchaeota dominated the microbial phylotype communities in the coarse-grained pumiceous gravels interbedded between the hemipelagic sediments. Based on the physical properties of sediments such as temperature and permeability, the porewater chemistry, and the microbial phylotype compositions, the shift in the physical properties of the sediments is suggested to induce a potential subseafloor recharging flow of oxygenated seawater in the permeable zone, leading to the generation of variable chemical environments and microbial communities in the subseafloor habitats. In addition, the deepest section of sediments under high-temperature conditions (∼90°C) harbored the sequences of an uncultivated archaeal lineage of hot water crenarchaeotic group IV that may be associated with the high-temperature hydrothermal fluid flow. These results indicate that the subseafloor microbial community compositions and functions at the marginal site of the hydrothermal field are highly affected by the complex fluid flow structure, such as recharging seawater and underlying hydrothermal fluids, coupled with the lithologic transition of sediments.

INTRODUCTION

Numerous scientific expeditions have investigated the marine subsurface biosphere via molecular biological analyses targeting 16S rRNA and functional genes as well as by metagenomics, metatranscriptomics, microscopic analyses, metabolic activity measurements, and cultivation experiments (1–3). These previous studies demonstrated that the subseafloor biosphere is composed of a vast number of microbial cells, including uncultivated, phylogenetically diverse and physiologically unknown components. Subseafloor microbiology over the past 2 decades indicates that buried organic matter is the most important energy and carbon source in deep subsurface environments that affects microbial abundance and the community composition (2, 4, 5). Thus, organic-rich subseafloor sediments of continental margins and the eastern equatorial Pacific Ocean harbor a larger microbial biomass (6–9) than oligotrophic sediments, where an extremely low microbial cell abundance is observed (4). On the other hand, the lithologic control of subseafloor microbial community development has also been suggested (5, 10) since the physical properties of sediments, such as porosity and permeability, highly affect the subseafloor hydrogeologic structures and the spatial, energetic, and nutritional habitability of the subseafloor microbial community (11–13).

Subseafloor hydrothermal fluid flow regimes have been predicted to provide spatially expansive and physicochemically variable habitats for the phylogenetic and functional diversity of microorganisms (14–18). The mixing of high-temperature reduced hydrothermal fluids and low-temperature interstitial fluids forms a wide range of physical and chemical gradients in the subseafloor environment. Currently, the variability of 16S rRNA gene phylotype communities has been investigated in several deep-sea hydrothermal sediments (19–23). Although the hydrothermal fluid discharging zones around active hydrothermal systems have been established to often host the local recharge flows of oxygenated deep-seawater (24, 25), the relationship between microbial community development and the physicochemical conditions influenced by the hydrothermal discharging and recharging fluid flows in the subseafloor environments remains poorly understood.

We sought to determine the pattern in subseafloor microbial community development along with the lithostratigraphic transition and physicochemical gradient in a deep-sea hydrothermal system, the Iheya North Knoll in the Mid-Okinawa Trough, during the Integrated Ocean Drilling Program (IODP) expedition 331 using the D/V Chikyu (26). The drilling and coring operations at IODP site C0017 located at the margin of the hydrothermal field indicated the potential recharge flow of low-temperature seawater into the subseafloor sediments, which was likely caused by the complex hydrogeologic structure and the underlying high-temperature hydrothermal fluid flow (26). The anomalously low heat flows around site C0017 estimated from temperature gradients of surface sediments suggested the zonation of seawater recharge (27), and the downhole temperature measurement during IODP expedition 331 showed a concave downward temperature profile (Fig. 1A), which indicated the lateral flows of recharged seawater in particular lithologic layers of coarse-grained pumiceous gravels and breccias (26). The oxygenated seawater flows in the porous layers would supply relatively abundant electron acceptors to the anoxic subseafloor sedimentary habitats. The lithostratigraphic transition between hemipelagic sediments and pyroclastic deposits has been predicted by the seismic reflection signals at site C0017 (26, 28). In deeper sections of sediments at site C0017, a higher temperature gradient was found by the downhole temperature measurement during IODP expedition 331, and the temperature in the deepest part was estimated to be 90°C (26). Here, we report the variability of subseafloor microbial phylotype communities and functional gene distribution (dsrA, aprA, and amoA) in the sedimentary habitats influenced by the hydrogeologic structure and the temperature gradient near a deep-sea hydrothermal system.

FIG 1.

Temperature profile (A), lithostratigraphic transition (B), and core photograph (C) of sediment samples at IODP expedition 331 site C0017, which were originally published elsewhere (26). The black diamonds indicate in situ temperature measured by the APCT-3 temperature shoe. The gray diamond indicates the exposed minimum temperature determined by a thermoseal strip taped to the outer surface of the core liner. The lithologic description was roughly modified to show the entire sedimentary structure of the core samples. The photographs were taken from the section closest to the microbiology samples used in the present study.

MATERIALS AND METHODS

Site description and sediment sampling.

IODP expedition 331 was conducted at the Iheya North hydrothermal field in the Mid-Okinawa Trough using the D/V Chikyu in September 2010 (26). IODP site C0017 was located 1,550 m east of the hydrothermal activity center of the Iheya North field and was covered with thick terrigenous sediments, hemipelagic sediments, and pumiceous deposits (26). Coring operations retrieved sediments down to 151 m below the seafloor (mbsf). The extended shoe coring system (ESCS) was used for the coring at 95.0, 108.2, 130.1, and 141.1 mbsf, and the hydraulic piston coring system (HPCS) was used for the rest of the cores (26). Core samples collected at site C0017 were composed of pumiceous volcaniclastic gravels, breccias, and hemipelagic mud and were lithostratigraphically classified into four units (26). The upper unit I (0 to 18.5 mbsf) was predominantly composed of hemipelagic mud, and units II and III (19.1 to 36.2 and 61.1 to 78.8 mbsf, respectively) consisted of pumiceous gravel-dominant layers with minor hemipelagic mud and volcaniclastic sediment. Sections from 36.2 to 61.1 mbsf and 78.8 to 94.3 mbsf were not recovered. The deepest section (unit IV), a cored interval from 94.3 to 144.7 mbsf, was dominated by hemipelagic mud. In situ temperatures at site C0017 were measured using an advanced piston corer temperature tool (APCT-3) and thermoseal strips (Nichiyu Giken Co., Ltd., Kawagoe, Japan) (Fig. 1A).

The retrieved cores were immediately cut into 1.5-m-long sections on deck, and whole round cores (WRCs) for microbiological study (approximately 10 to 20 cm in length) were then subsampled from the short sections. The microbiological samples were obtained from the inner parts of the WRCs with a sterilized spatula and immediately stored at −80°C in heat-sealed laminated foil bags containing an oxygen scavenger. The subsamples for geochemical analyses were collected from the sections juxtaposed to the WRCs for microbiology.

Geochemical analysis.

Porewater was obtained from 10- to 20-cm-long WRCs; the total alkalinity and the ammonium, sulfate, and methane concentrations were determined previously (26). The nitrate concentration was measured by ion chromatography using a high-capacity anion exchanger (TSK-gel SAX column; Tosoh) with UV detection (LC-10Ai and SPD-10A; Shimadzu) (29, 30). The lower detection limit of the nitrate concentration was 0.3 μM, and the reproducibility was better than 10%. Subsamples for dissolved organic carbon analysis were stored frozen at −20°C in precombusted 10-ml glass vials, each with a Teflon-coated septum and screw caps. The acetate concentration and stable carbon isotopic composition (δ¹³C_acetate) were determined by isotope ratio monitoring-liquid chromatography-mass spectrometry, as previously described (31).

DNA extraction and 16S rRNA gene clone analysis.

DNA was extracted from ∼2 g of the frozen innermost parts of the WRCs using the PowerMAX Soil DNA isolation kit (MoBio Laboratories, Carlsbad, CA) according to the manufacturer's protocol, with minor modifications. A blank water sample was also used simultaneously as a negative control during the DNA extraction process. The 16S rRNA gene fragments were amplified by PCR using an universal primer set, Uni530F-907R (32), for all samples, and an archaeon-specific primer set composed of Arch_530F, Arch2_530F, Nano_530F (32), and Arc958R (33) was used for samples in which indigenous microbial populations were not detected using the universal primer set. PCR amplification with LA Taq polymerase (TaKaRa Bio, Inc., Otsu, Japan) was performed using the following cycle conditions: 40 cycles of denaturation at 96°C for 25 s, annealing at 50°C for 45 s, and extension at 72°C for 60 s. PCR amplification of the negative control for DNA extraction was used to assess experimental contamination. Cloning and sequencing of the PCR products were performed as described previously (34). Approximately 400-bp 16S rRNA gene sequences with >97% sequence identity were assigned to the same phylotype. Representative sequences were aligned using the SINA aligner (35). Phylogenetic affiliations were identified by the maximum-parsimony method using the SILVA SSU Ref 111 Database in the ARB software program (36). The same method was previously applied to the drilling fluids, and the 16S rRNA gene sequences from the WRCs displaying >97% identity with those of the drilling fluids were defined as potential contaminants (37). The 16S rRNA gene clone libraries were compared by Jackknife environment cluster analysis and principal component analysis (PCA) in the UniFrac program (http://bmf.colorado.edu/unifrac/) (38).

Archaeal amoA gene clone analysis.

The PCR amplification of amoA, encoding ammonia monooxygenase subunit A, was conducted using the primer set of Arch-amoAF and Arch-amoAR (39) and ExTaq polymerase (TaKaRa Bio Inc., Shiga, Japan) with Mg²⁺ buffer, as previously described (40). The amplification condition was 40 cycles of denaturation at 96°C for 25 s, annealing at 52°C for 30 s, and extension at 72°C for 60 s. The PCR products were cloned and sequenced as described above. Sequences presenting ≥95% identity were assigned to the same phylotype. Representative sequences were aligned with closely related amoA gene sequences deposited in public databases using the ClustalW program, and the ambiguous nucleotide positions were corrected manually. Phylogenetic affiliations were assigned based on phylogenetic trees constructed by the neighbor-joining method in the ARB software. Bootstrap analysis was performed with 1,000 replicates.

Q-PCR.

Quantitative fluorescent PCR (Q-PCR) for 16S rRNA genes was performed as described previously (41–43). Copy numbers of the 16S rRNA genes were determined using a universal primer-probe set (44), an Archaea-specific primer-probe set (44), and a Bacteria-specific primer-probe set (45). The functional genes dsrA and aprA, which encode dissimilatory sulfite reductase and adenosine 5′-phosphosulfate reductase subunit A, respectively, were quantified as described elsewhere (41, 46), using specific primer sets (47, 48). The primers and probes in the present study are summarized in Table 1 (44, 45, 47, 48). After each Q-PCR, melting curves were measured for SYBR green I assays. The sizes of the PCR products were further confirmed by gel electrophoresis. All Q-PCR assays were performed in triplicate.

TABLE 1.

Primers and probes used in the Q-PCR assays

Target gene	Primer/probe	Sequence (5′–3′)	Reference
Prokaryotic universal 16S rRNA gene	Uni340F	CCTACGGGRBGCASCAG	44
	Uni806R	GGACTACNNGGGTATCTAAT	44
	Uni516F (TaqMan probe)	TGYCAGCMGCCGCGGTAAHACVNRS	44
Archaeal 16S rRNA gene	Arch349F	GYGCASCAGKCGMGAAW	44
	Arch806R	GGACTACVSGGGTATCTAAT	44
	Arch516F (TaqMan probe)	TGYCAGCCGCCGCGGTAAHACCVGC	44
Bacterial 16S rRNA gene	331F	TCCTACGGGAGGCAGCAGT	45
	797R	GGACTACCAGGGTATCTAATCCTGTT	45
	TaqMan probe	CGTATTACCGCGGCTGCTGGCAC	45
Dissimilatory sulfite reductase (dsrA)	DSR-1F+	ACSCACTGGAAGCACGGCGG	47
	DSR-R	GGTTRKACGTGCCRMGGTG	47
Adenosine 5′-phosphosulfate reductase subunit A (aprA)	AprA-1-FW	TGGCAGATCATGATYMAYGG	48
	AprA-5-RV	GCGCCAACYGGRCCRTA	48

Nucleotide sequence accession numbers.

The nucleic acid sequences obtained in the present study have been deposited in the DDBJ/EMBL/GenBank databases under the following accession numbers: AB824899 to AB825952 (16S rRNA gene) and AB936820 to AB936831 (amoA).

RESULTS

Thermal and geochemical variation associated with sediment lithology.

At IODP expedition 331 site C0017, located 1550 m east of the activity center of the Iheya North hydrothermal field, in situ temperature measurements did not show a significant temperature change in units I and II, whereas a marked temperature increase was observed in unit III; the temperature at 69 mbsf was 25°C (Fig. 1A). At the deepest section of unit IV, the temperature increased up to 90°C.

In the unit I layer, which mainly consisted of hemipelagic sediments (Fig. 1B and C), the porewater alkalinity and ammonium concentration increased from 3.2 to 7.3 mM and from 0.02 to 0.45 mM, respectively, whereas the sulfate concentration slightly decreased as the depth increased (Fig. 2). The nitrate concentration was less than 2 μM throughout the unit. These geochemical features were indicative of a typical anaerobic sedimentary environment, where oxygen was presumably consumed by microbial respiration within the uppermost sediment. Units II and III consisted of coarse-grained porous volcaniclastic pumiceous deposits (Fig. 1B and C). The unit II layer presented an inverse trend of alkalinity and ammonium and sulfate concentrations compared to unit I (Fig. 2). Furthermore, the nitrate concentration significantly increased up to 34 μM at the bottom of the unit. The nitrate concentration and alkalinity value are similar to those in the deep seawater of the East China Sea (ca. 38 μM for nitrate and 2.5 mM for alkalinity) (49, 50). Hence, the chemical characteristics suggest that porewater with little influence of early diagenesis of infiltrated seawater exists in coarse-grained pumiceous gravels and breccias localized at 26.6 to 30.0 mbsf in unit II. Given the anomalously low thermal gradient in this area (27), a lateral flow would be caused by recharging of the oxygenated and low-temperature bottom seawater. At the bottom of this hole (unit IV), we observed no apparent chemical signature for hydrothermal fluid input in the porewater chemistry. Throughout the sediment column, the methane concentrations were quite low, mostly below 1 μM (26).

FIG 2.

Depth profile of porewater alkalinity, the ammonium, nitrate, sulfate, and acetate concentrations and the carbon isotopic composition of acetate in the core samples at site C0017. The original data regarding alkalinity, ammonium, nitrate, and sulfate were published elsewhere (26).

Acetate is a key intermediate substance of various microbial metabolic pathways in anaerobic environments (51). Biogeochemical processes via acetate in subseafloor sediments at site C0017 were examined by the stable carbon isotopic analysis of acetate in the porewater (Fig. 2). The acetate concentration highly fluctuated, possibly due to the difference in sediment lithology. High acetate concentrations of up to 80.6 μM were detected at certain depths in units I, II, and III, where the sediment is mostly hemipelagic clay. In contrast, the concentration in the layers of volcaniclastic sand ranged from 8.2 to 22.6 μM. A similar variation was also found in the total organic carbon content (26). Throughout the sediment column, the isotopic composition of acetate (δ¹³C_acetate) ranged from −37.6‰ to −32.2‰. The local maximal stable carbon isotopic compositions of porewater acetate were observed at the interface of units I and II.

Total cell counts and Q-PCR analysis.

The microbial cell abundance at site C0017 decreased logarithmically with sediment depth from 3.2 × 10⁷ cells per ml of sediment at 0.7 mbsf to less than the detection limit of approximately 6.5 × 10⁵ cells per ml of sediment at 68 mbsf (26). Similar depth profiles were obtained from the 16S rRNA gene-targeted Q-PCR analysis (Fig. 3A). The 16S rRNA gene copy number ranged from 8.8 × 10⁵ to 8.7 × 10⁷ genes g⁻¹ sediment for total prokaryotes, from 5.7 × 10⁴ to 2.7 × 10⁷ genes g⁻¹ sediment for bacteria, and from 6.4 × 10⁴ to 8.9 × 10⁶ genes g⁻¹ sediment for archaea. The highest cell count and 16S rRNA gene number were detected at a depth of 6.4 mbsf. To reveal the spatial distribution of particular physiological microbial groups, we quantified the copy numbers of functional genes dsrA and aprA for potential sulfate reducers that encoded dissimilatory sulfite reductase and adenosine 5′-phosphosulfate reductase subunit A, respectively. They were less abundant than the 16S rRNA genes but detectable in most of the samples (Fig. 3B).

FIG 3.

Total cell counts and numbers of 16S rRNA and functional genes in the subseafloor core samples at site C0017. (A) Total cell counts (open circles) and the 16S rRNA gene numbers of prokaryotes (black circles), bacteria (dark gray triangles), and archaea (light gray diamonds) quantified by Q-PCR. The total cell counts were originally reported by Takai et al. (26). (B) Numbers of functional genes dsrA (dark gray triangles) and aprA (open squares).

16S rRNA gene phylotype community.

In the shallow depths of unit I above 14.8 mbsf, the 16S rRNA gene phylotype communities were dominated by previously uncultivated sequences of typical subseafloor lineages. At 6.4 mbsf, bacterial phylotypes affiliated with phylum Chloroflexi comprised 27.1% in the 16S rRNA gene clone library (23/86 clones) (Fig. 4). Members of the deep-sea archaeal group (DSAG), also referred to as marine benthic group B (MBG-B), accounted for 26.4% (23/87 clones) at 0.7 mbsf. The 16S rRNA gene communities in unit II, except for those from the upper depth (20.1 mbsf), were markedly dominated by members of marine group I (MG-I) Thaumarchaeota and Alphaproteobacteria (Fig. 4). Archaeal sequences belonging to the miscellaneous crenarchaeotic group (MCG) were predominantly detected at depths of 63.6 and 68.1 mbsf in unit III. The data from the deep layer in unit IV showed that the 16S rRNA gene communities using a universal primer set likely reflected highly biased compositions by external contamination rather than the indigenous compositions in these deep sedimentary habitats (data not shown). Due to the markedly low microbial cell abundances (Fig. 3A), these samples would be susceptible to microbiological contamination from the drilling fluid used for the operation during IODP expedition 331 (37). Since no detectable archaeal 16S rRNA gene sequences were obtained from the drilling fluid (37), we further constructed archaeal 16S rRNA gene clone libraries in the deep sediment samples of unit IV. Notably, we found a drastic transition of the archaeal 16S rRNA gene phylotype composition in unit IV. At a depth of 95.0 mbsf, sequences of South Africa gold mine euryarchaeotic group (SAGMEG) and AK8 predominated the archaeal phylotype composition, whereas the hot water crenarchaeotic group IV (HWCGIV; also known as terrestrial hot spring crenarchaeotic group [THSCG] or UCII) were dominant at the depth of 141.1 mbsf, representing 29.0% of the clonal frequency (Fig. 4).

FIG 4.

16S rRNA gene phylotype compositions in the sediments at site C0017. The 16S rRNA gene fragments were amplified with the universal and archaeon-specific primer sets of Uni530F-907R and Arc530F-Arc958R, respectively. The numbers in parentheses indicate the number of clones.

Based on the phylogenetic distance and abundance of each 16S rRNA gene phylotype, we compared differences in the microbial phylotype composition among all samples using UniFrac analysis. The results of the cluster analysis and PCA revealed the evident compositional variability among the 11 sediment samples (Fig. 5A and B). The microbial phylotype compositions in three samples from the permeable zone in unit II were similar to one another and were significantly distinct from those in the other sample layers. The predominant phylotypes commonly found in the samples within this cluster were the members of MG-I Thaumarchaeota and Alphaproteobacteria, as described above. Similarly, the microbial phylotype compositions in the shallow sediments of unit I were also closely related to one another.

FIG 5.

Jackknife environment cluster analysis (A) and PCA (B) in the UniFrac program. The jackknife values were estimated using 100 permutations and are shown in the nodes of the dendrogram. Each axis of the PCA plot indicates the fraction of the variance in the data. The black circles, plus signs, and triangles indicate the sediment samples of units I, II, and III, respectively.

Detection of archaeal amoA genes from the permeable pumice-rich layer.

Previous studies of microbial communities in marine and soil environments showed that most of the MG-I Thaumarchaeota harbor the amoA gene, encoding ammonia monooxygenase subunit A, and can oxidize ammonia to nitrite (52). We could amplify archaeal amoA gene fragments from a permeable pumice-rich layer at a depth of 30.0 mbsf. Forty-three sequences of the archaeal amoA gene were evaluated and classified into 12 phylotypes. All of the archaeal amoA phylotypes were phylogenetically related to sequences detected in ocean waters and sediments (see Fig. S1 in the supplemental material). Co-occurrence of the MG-I 16S rRNA and archaeal amoA genes suggested the potential contribution of ammonia oxidation in the permeable layer, where the recharged seawater flows may have created the aerobic habitats.

DISCUSSION

Variation of microbial communities in the subseafloor sediments.

In this study, microbial cell populations in subsurface marine sediments at Site C0017 were quantified using both microscopic observation and Q-PCR. The 16S rRNA gene numbers showed a pattern similar to the total cell counts, indicting a high reliability of the data produced by both quantification methods (Fig. 3A). Relatively low microbial cell abundances in the subseafloor sediments at this site, compared to the abundances in other subseafloor sedimentary habitats, would be explained by the relatively low productivity of the overlying oligotrophic ocean and the distance from land (4).

In the unit I sediment, the increase of the porewater alkalinity and ammonium, the decrease of sulfate and the low nitrate concentration are considered the result of anaerobic microbial processes (Fig. 2). Furthermore, it seems likely that the anaerobic microbial activity provides the ¹³C enrichment of acetate at the interface of units I and II. One conceivable explanation would be that slight carbon isotopic fractionation occurs during fermentative acetate production (53, 54). However, judging from the fact that the maximal stable carbon isotopic compositions of acetate are associated with the local minimum concentrations of acetate, the ¹³C enrichment of acetate is a potential signature of microbial consumption as a substrate. This is supported by previous laboratory experiments with pure cultures of acetotrophic sulfate reducers, which showed that acetate was enriched in ¹³C up to 19.3‰ (55). The 16S rRNA gene phylotype communities in the Unit I sediments were dominated by the members of Chloroflexi and DSAG. Both are widely distributed in many deep-sea sediments and, in some cases, represent more than half of the 16S rRNA gene clone libraries (1). However, no representatives of marine subsurface Chloroflexi (Dehalococcoidetes) and DSAG have thus far been cultured, and thus their metabolic pathways remain elusive. Currently, a single cell genomic approach for Dehalococcoidetes from marine subsurface sediments suggests they are strictly anaerobic organotrophs or lithotrophs (56, 57). The members of DSAG are considered to be involved in the biogeochemical cycling of organic carbon, iron oxide, and/or manganese (58, 59).

The profiles of the dsrA and aprA gene copy numbers suggested that the relative abundance of sulfate-reducing bacteria was higher at shallower depths than in deeper sections (Fig. 3B), which was comparable with the findings of other sediments (41). Indeed, a current biogeochemical modeling of microbial sulfate reduction using the concentration and multiple sulfur isotopic compositions of porewater sulfate has strongly suggested the occurrence of active microbial sulfate reduction in the shallowest zone of sediments down to 20 mbsf (unit I) (60). Correspondingly, the 16S rRNA gene phylotypes affiliated with Deltaproteobacteria were detected from sediments above 74.9 mbsf (Fig. 4). These results suggest that microbial sulfate reduction may be an important metabolic function that supports subseafloor microbial production in anaerobic hemipelagic mud habitats.

The 16S rRNA gene phylotype composition drastically changed in the deeper sediments, potentially influenced by the subsurface flow of recharged seawater in the permeable pumiceous zone. The cluster analysis and PCA also indicated that the physical and/or chemical properties associated with the lithologic structure would differentiate the possible subseafloor microbial community composition responding to the lithostratigraphic transition (Fig. 5). A key component of the microbial phylotype composition in the permeable pumice-rich sediments was the predominance of the MG-I members in Thaumarchaeota (Fig. 4). Several studies have previously reported that thaumarchaeal phylotypes are frequently obtained from the aerobic and oligotrophic sediments (61) and presumably from the anoxic subsurface sediments (8, 58, 62). Although the dissolved oxygen concentration was not measured in the present study, relatively high concentrations of nitrate in the permeable samples were indicative of the presence of dissolved oxygen (63) and were distinct from the typical anaerobic sedimentary environments (Fig. 2). The chemical conditions would provide the feasibility of microbial ammonia oxidation in the permeable zone. Correspondingly, the members of MG-I Thaumarchaeota are known to possess the dissimilatory ammonia oxidation pathway (64). In fact, we successfully detected archaeal amoA genes from this zone (see Fig. S1 in the supplemental material). The Q-PCR results also indicated the distinct composition of microbial components in the permeable layer. Comparison between archaeal and total 16S rRNA gene copy numbers revealed that the archaeal 16S rRNA gene abundance accounted for 25.4 to 39.5% of the total prokaryotic 16S rRNA gene abundance in the permeable layer, whereas the archaeal population remained as 4.11 to 24.6% of the total prokaryotic 16S rRNA gene population at any other sample depth. Thus, the distinct microbial community development in the permeable layer might be caused by the dominance of potential ammonia-oxidizing MG-I Thaumarchaeota in the possible indigenous microbial communities. Although another explanation (e.g., certain Q-PCR biases due to a few mismatches of the ARCH516 TaqMan probe with the archaeal phylotypes typically found in the anoxic subseafloor sediments [3, 65]) cannot be completely excluded, the estimated 16S rRNA gene abundances of MG-I Thaumarchaeota in the specific subseafloor habitats are an order of magnitude higher than the MG-I cell abundances in the ambient bottom seawater of the Iheya North hydrothermal field (66). Therefore, the MG-I Thaumarchaeota likely represent one of the predominant indigenous microbial populations in the specific layer rather than only being contaminants from the potentially recharged seawater.

Abundant incidence of HWCGIV in high-temperature sediments.

Based on the microbial community surveys of hydrothermal mixing zones of habitats at the seafloor of the Iheya North hydrothermal field, metabolically diverse microbial communities (including psychrophilic to hyperthermophilic chemolithotrophs) were expected to be distributed abundantly and widely in the subseafloor environments beneath the hydrothermal field (67). However, the culture-dependent attempts for potentially indigenous microbial populations associated with hydrothermal activity (such as Thermococcales, Aquificales, and Epsilonproteobacteria) were unsuccessful throughout the sediments at site C0017 (26). The temperature in the deepest sample at 151 mbsf of up to 90°C was below the upper temperature limit of life (122°C) (68). The archaeal 16S rRNA gene community composition in the deepest (141 mbsf) and hottest core for the microbiological analysis was dominated by the HWCGIV and MCG (Fig. 4), whereas we found no Thermococcales-, Methanococcales-, and Archaeoglobales-related sequences throughout the sediments that were previously detected in the hydrothermal chimney structures of the Iheya North field (67). The HWCGIV have been reported in microbial habitats associated with deep-sea hydrothermal vents (21, 43). The biological thermometer estimation using the GC contents of the 16S rRNA gene sequences proposed by Kimura et al. (69) indicates that the potential optimal and maximum growth temperatures of the HWCGIV found in the present study are 74 and 83°C, respectively. These predicted values are consistent with in situ temperatures measured with thermoseal strips (Fig. 1A). Accordingly, the HWCGIV likely represent an active and indigenous population in the subseafloor sediments under the high-temperature conditions at site C0017. Nevertheless, the temperature condition does not explain why the hyperthermophilic chemolithotrophic populations, such as Thermococcales, Methanococcales, and Archaeoglobales, which can grow above 70°C (70), were absent. The porewater chemistry in the deepest sediments revealed no evident chemical input of hydrothermal fluids. Not only the temperature but also the substantial chemical fluxes from hydrothermal fluid flow may be required for hyperthermophilic chemolithotrophic microbial community development in the subseafloor environments.

Oxic fluid circulation within deep biosphere.

In hypothetical models of the total hydrothermal circulation of the Iheya North field, proposed by Kawagucci et al. (49) and Tsuji et al. (28), the hydrothermal circulation would begin with the bottom seawater recharge in the sediments along the faults of the Okinawa Trough basin far distant from the hydrothermal field. Particularly, this model assumes that the seawater recharge occur not only in the Central Valley (estimated to be 2-km square area), but also in the spatially abundant and widespread basin-filling sediments surrounding the Iheya North Knoll. During the long spatial and temporal migration in the sediments at the recharge stage, the microbially produced methane, ammonium and other compounds are likely added to the source fluids (49). However, the recharged seawater flow discovered in the present study is substantially different from such a great spatial and temporal scale of hydrothermal circulation and is spatially and temporally limited. Indeed, the porewater sulfate is partially utilized by potential functions of the indigenous subseafloor sulfate-reducers (60), but the concentration is relatively constant throughout the sediments (Fig. 2). In addition, none of the methanogen-related sequences were detected in any of the 16S rRNA gene clone libraries at site C00017 (Fig. 4). These results suggest that the sediments and porewater in core samples are less affected by the geochemical and microbiological alterations than by the large scale of seawater recharge and alteration processes. The existence of high concentrations of porewater nitrate is an important chemical signature of the oxidative (potentially aerobic) condition and the relatively fresh seawater input in the unit II layer (Fig. 2). The microbial community development pattern also suggests the drastic transition within unit II (Fig. 5). Although units II and III showed quite similar lithologic characteristics, there was a hard layer boundary (almost no core recovered) between units II and III, and the unrecovered hard layer would serve as an impermeable layer to prevent possible vertical fluid exchange (26). The microbial community compositions inferred from the 16S rRNA gene clone sequences were also significantly different between units II and III, and the sequences affiliated with MCG dominated the phylotype compositions in the unit III samples, contrasting with the MG-I Thaumarchaeota members in unit II. Thus, the highly permeable layer in unit II would provide a novel habitat of the subseafloor biosphere that has been unexplored in the previous scientific ocean drilling expeditions. The seismic reflection survey and its interpretation revealed horizontally widespread and vertically multiple distributions of porous and permeable layers in the sediments around the Iheya North hydrothermal field (28). If some of these permeable layers host the horizontal recharge flow of relatively fresh seawater, the microbial community stratification estimated in the sediments of site C0017 may be a common pattern of subseafloor microbial community development in the marginal sedimentary environment of the Iheya North field.

Conclusions.

We report here the potential microbial community stratification associated with the complex fluid flow structure, such as recharging seawater and underlying hydrothermal fluids, coupled with the lithologic transition of sediments at site C0017 in the Iheya North hydrothermal field of the Mid-Okinawa Trough. Uncultivated microbial components, which are frequently detected in subseafloor sedimentary environments, populated the shallow sections, whereas members of MG-I Thaumarchaeota dominated the 16S rRNA gene phylotype communities in the pyroclastic deposits. The sharp transition of the potential microbial community is most likely controlled by different physical properties of sediments, such as the permeability of hemipelagic muds and volcaniclastic sediments, which are further related to the hydrogeologic structure and geothermal gradient of the subseafloor environment. Our results reveal the dynamics of biogeochemical and microbiological processes in the subseafloor sediments, directly and indirectly associated with local fluid flows such as fresh seawater recharge and hydrothermal fluid discharge. A great spatial and temporal scale of hydrothermal circulation has been extensively investigated in crustal aquifers at mid-ocean ridge flanks, which also potentially supplies oxidants through the basaltic basement and has a significant role in biogeochemical cycles and crustal rock alteration (63, 71). The drilling operation during IODP expedition 331 was unsuccessful in reaching the volcanic basement, which might exist at ∼450 mbsf at site C0017 (26). However, our findings here indicate that the oxidants transported through the local seawater circulation associated with hydrothermal activity are important for generating variable chemical environments and microbial communities in the subseafloor sedimentary habitats and even in potentially deeper sediment-basement interface habitats.