Correspondence: Akihiko Yamagishi, Department of Molecular Biology, Tokyo University of Pharmacy and Life Science, 1432-1 Horinouchi, Hachioji, Tokyo 192-0392, Japan. Tel.: +81 426 76 7139; fax: +81 426 76 7145; e-mail: email@example.com
The abundance and phylogenetic diversity of the microbial community in the hydrogenetic ferromanganese crust, sandy sediment and overlying seawater were investigated using a culture-independent molecular analysis based on the 16S rRNA gene. These samples were carefully collected from the Takuyo-Daigo Seamount, located in the northwest Pacific Ocean, by a remotely operated vehicle. Based on quantitative PCR analysis, Archaea occupy a significant portion of the prokaryotic communities in the ferromanganese crust and the sediment samples, while Bacteria dominated in the seawater samples. Phylotypes belonging to Gammaproteobacteria and to Marine group I (MGI) Crenarchaeota were abundant in clone libraries constructed from the ferromanganese crust and sediment samples, while those belonging to Alphaproteobacteria were abundant in that from the seawater sample. Comparative analysis indicates that over 80% of the total phylotype richness estimates for the crust community were unique as compared with the sediment and seawater communities. Phylotypes related to Nitrosospira belonging to the Betaproteobacteria and those related to Nitrosopumilus belonging to MGI Crenarchaeota were detected in the ferromanganese crust, suggesting that these ammonia-oxidizing chemolithoautotrophs play a role as primary producers in the microbial ecosystem of hydrogenetic ferromanganese crusts that was formed as precipitates from seawater.
Ferromanganese deposits are often found at the boundary between the hydrosphere and the lithosphere in natural environments. Rocks coated with ferromanganese oxides are found on modern seafloors as ferromanganese nodules and crusts (hereafter, Mn nodules and Mn crusts) depending on their mode of occurrence (e.g. Usui & Someya, 1997; Glasby, 2006; Wang & Müller, 2009). Mn nodules and crusts mainly consist of Mn and Fe oxides, more than 30% of the total mass (Mero, 1962), and contain other economic metals, for example, Co, Ni, Cu, Zn, rare earth elements and Pt (Hein et al., 2000). Oceanic ferromanganese deposits grow extremely slowly at rates of about 1–10 mm Myr−1 as determined by radioisotope dating (Hein et al., 2000; Usui et al., 2007). Although hydrothermal ferromanganese deposits occur in areas associated with volcanic activity, hydrogenetic ferromanganese deposits are distributed widely on the deep seafloor (Rona, 2003). Considering the wide distribution of Mn nodules and crusts on the seafloor and their potential for future mineral resources (Rona, 2003), the study of microorganisms attached to the Mn nodules and crusts is important to understand the significance of the role of microorganisms in the elemental cycle between the ocean and the hydrogenetic oxides. This knowledge is likely to help us develop deep-sea mining techniques utilizing microorganisms in future (Ehrlich, 2001).
Despite the early discovery of Mn nodules and crusts on the seafloor, little is known about the microbial communities and their role in Mn nodule formation. In terrestrial environments, microbial communities on ferromanganese oxides have been reported from caves (Northup et al., 2003), soils (He et al., 2008) and freshwater sediments (Stein et al., 2001), suggesting that diverse prokaryotes are present on and/or within the ferromanganese oxides. Electron microscopic observation has shown that microorganism-like structures are present on the oceanic ferromanganese oxides (Wang et al., 2009). The presence of phylogenetically diverse bacteria in the seafloor basalt covered with thin (<200 μm) ferromanganese oxides on the East Pacific Rise has been reported (Santelli et al., 2008). However, our knowledge of the spatial distribution, diversity and abundance of microbial communities on oceanic ferromanganese oxides is still limited.
Here, we report on the abundance, diversity and composition of the microbial community of an oceanic Mn crust by a culture-independent molecular microbiological analysis. The Mn crust was carefully collected with on-site observation using a remotely operated vehicle, enabling us to investigate microorganisms on the undamaged surface of the Mn crust that is exposed to overlying seawater by molecular microbiological analysis. The Takuyo-Daigo Seamount of the sampling field is a flat-topped seamount that is located approximately 150 km southeast of Minamitorishima Island, Japan, in the northwest Pacific Ocean (Supporting Information, Fig. S1). This area is one of the oldest seafloors in the world (>150 million years, Müller et al., 2008). No age determination has been carried out on the Takuyo-Daigo Seamount, but the age of nearby seamounts is around 80 million years. This seamount has a flat-top at a depth of 810 m, elevating more than 4000 m from the abyssal seafloor of 5300 m. The Mn crusts were collected from the slope of the seamount at a water depth of 2991 m. In addition to the Mn crust, we also sampled and analyzed the overlying seawater and surrounding sandy sediment using the same methods to assess the uniqueness of the microbial communities of the oceanic Mn crust.
Materials and methods
The Mn crusts, sandy sediments and overlying seawater samples were collected on the slopes of the Takuyo-Daigo Seamount (Figs 1 and S1) at 2991 m water depth during the NT09-02 cruise (February 8–23, 2009) of the R/V Natsushima (JAMSTEC, Japan) with the remotely operated vehicle Hyper-Dolphin (JAMSTEC). The temperature, dissolved oxygen concentration and salinity of the bottom ambient seawater were 2 °C, 2.5 mL L−1 and 34.0 practical salinity units, respectively. The Mn crusts were carefully collected using a manipulator on the vehicle while observing on TV monitors. Samples of sandy sediments and seawater were collected approximately 10 m from the sampling point of the Mn crusts using a push-core and a Niskin bottle sampler, respectively. Samples from 0 to 1 cm from the top of the sediments, which were collected using a push-core sampler, were used for analysis. Although the correct thickness of the covering sediments is unknown, the thickness seemed to be <1 m judging from the depth of an iron stick inserted into sediments at the sampling area. Once retrieved on-board, the Mn crust samples were crushed using an autoclaved hammer and chisel in a clean box. Parts of the Mn crusts and sediments were transferred to a DNA/RNA-free plastic tube and stored at −80 °C until DNA extraction. One liter of the seawater sample was filtered with a 0.2-μm-pore-size polycarbonate membrane to trap the suspended particles (Advantec, Tokyo, Japan) on board and then the filter was stored in a DNA/RNA-free plastic tube at −80 °C until DNA extraction.
16S rRNA gene analysis
Analysis of the 16S rRNA genes present in the collected solid and liquid samples was performed as described previously (Kato et al., 2009c, 2010). In brief, genomic DNA was extracted from the samples using a Fast DNA kit for soil (Qbiogene, Carlsbad, CA). Partial 16S rRNA genes were amplified by PCR with the prokaryote-universal primer set, Uni515F and Uni1406R. The PCR products were cloned using a TOPO TA cloning kit (Invitrogen, CA). The nucleotide sequences of randomly selected clones were determined using M13 forward and reverse primers (Invitrogen) on an ABI PRISM 3130xl Genetic analyser (Applied Biosystems, CA). Nucleotide sequences were aligned and distance matrices were generated from alignment data sets from each clone library using arb (Ludwig et al., 2004). Clones having 97% sequence similarity or higher were treated as the same phylotype using dotur (Schloss & Handelsman, 2005). Maximum-likelihood trees were constructed using phyml (Guindon & Gascuel, 2003) with non-gap homologous positions in the alignment dataset. Bootstrap values were estimated using 100 replicates. Rarefaction analysis, the Shannon diversity index and Chao1 richness estimators were estimated using dotur based on the distance matrices generated from the alignment data sets of the clones from each clone library. Chao1 species richness estimates of shared phylotypes were calculated using sons (Schloss & Handelsman, 2006). The phylogenetic (P)-test and the UniFrac significance test were performed using UniFrac (Lozupone et al., 2006).
Quantitative PCR (Q-PCR)
Bacterial and archaeal rRNA gene copy numbers in DNA extracts from each sample were determined by Q-PCR as described previously (Kato et al., 2009b). For bacterial rRNA genes, the bacterial-specific PCR primers, Bac1369F (5′-CGGTGAATACGTTCYCGG-3′) and Prok1492R (5′-GGWTACCTTGTTACGACTT-3′), and the TaqMan probe, TM1389F (5′-CTTGTACACACCGCCCGTC-3′), were used. For archaeal rRNA genes, the archaeal PCR primers, Arc349F (5′-CCTACGGGRBGCASCAG-3′) and Arc806R (5′-GGACTACNNGGGTATCTAAT-3′), and a TaqMan probe, Arc516F (5′-TGYCAGCMGCCGCGGTAAHACVNRS-3′), were used. The purified PCR products from the 16S rRNA gene of Escherichia coli and environmental archaeal clones belonging to Marine group I (MGI) were used as the standard DNA for bacterial and archaeal analyses, respectively. All assays were performed in triplicate. Regression coefficient (r2) values of the standard curve were 0.994 and 0.999 for bacterial and archaeal analyses, respectively.
The nucleotide sequences of the phylotypes reported in this paper have been deposited in the DDBJ database under the following accession numbers: AB606678–AB606781 for Mn crust clones, AB606782–AB606782 for sediment clones and AB606863–AB606968 for overlying seawater clones.
Results and discussion
The thickness of the Mn oxides covering the basement rock was ∼20 mm (Fig. 1b; a representative image of the Mn crusts collected). The chemical composition of the Mn crust sample (0–3 mm from the surface) was determined by inductively coupled plasma-optical emission spectrometry, which yielded the following results: (wt%) 17.4% Fe, 16.0% Mn, 1.62% Ca, 0.834% Na, 0.715% Ti, 0.663% Mg, 0.661% Al, 0.389% K, 0.386% Co, 0.323% P, 0.209% Ni, 0.134% Pb, 0.118% S, 0.111% Sr. This sample also contained <0.1% Ba, V, Zn, Cu, Y, Cr and Sc as minor components. Although the chemical composition of the sediments was not determined, these sediments are likely to consist of calcareous shells of foraminifers that are generally found on the seafloor of open oceans.
Bacterial and archaeal cell densities were estimated based on the 16S rRNA gene copy numbers determined by Q-PCR (Fig. 2). In principle, the quantification of microorganisms by Q-PCR provides more reliable data than by clone library analysis (Smith & Osborn, 2009). Our estimation is based on the assumption that the genomes of bacterial and archaeal cells have on average 4.06 and 1.77 copies of the 16S rRNA gene, respectively (Lee et al., 2009). The total prokaryotic cell numbers were estimated to be 7.27 × 107 cells g−1, 1.29 × 109 cells g−1 and 8.20 × 103 cells mL−1 for the Mn crust, sediment and ambient seawater, respectively. The cell numbers of deep-sea water (>2000 m depth) are generally 0.8–2.0 × 104 cells mL−1 as shown by direct counting (Karner et al., 2001; Herndl et al., 2005; Kato et al., 2009c). Our result of the seawater from Q-PCR was within the range reported previously. Bacteria were found to be dominant in the seawater sample (98.4% of the total prokaryotic cell number; Fig. 2). In contrast, Archaea were found to be dominant in the Mn crust and sediment (65.5% and 84.7%, respectively; Fig. 2). The percentage of archaeal clones in the libraries (Fig. 3) did not quantitatively match that obtained from Q-PCR (Fig. 2) and is probably due to PCR bias. In fact, the prokaryote-universal primer set that was used does not amplify 16S rRNA genes from all Archaea (Baker et al., 2003). However, the relative abundance of archaeal clones in the libraries (17.3% for the Mn crust, 24.7% for the sediment and 5.7% for the seawater, respectively; Fig. 3) showed the same trend as the results obtained by Q-PCR (65.5%, 84.7%, 1.6%, respectively; Fig. 2): the relative abundance of archaeal clones was much higher in the Mn crust and the sediment than in the seawater. Although Archaea dominate in marine sediments (Lipp et al., 2008), Archaea are thought to be a minor component of the microbial community of seafloor basaltic rocks (Einen et al., 2008) and also that within thin ferromanganese oxides (Santelli et al., 2008). Our results suggest that Archaea occupy a significant portion of the prokaryotic communities in aged Mn crusts and sandy sediments.
The microbial communities on/within basaltic glass and rocks on the seafloor have been well studied (Fisk et al., 2003; Lysnes et al., 2004; Mason et al., 2007; Einen et al., 2008; Santelli et al., 2008); however, little is known about those on well-developed Mn crusts on the aged seafloor. For the first time, we analyzed the composition and diversity of Archaea and Bacteria on an aged Mn crust (Fig. 3 and Table 1). The archaeal clones recovered from the Mn crust were affiliated with MGI Crenarchaeota (Delong, 1992; Fuhrman et al., 1992) and with the pSL12-related group (Barns et al., 1996) (Fig. S2a). MGI includes the chemolithoautotrophic ammonia-oxidizing archaeon Nitrosopumilus maritimus (Könneke et al., 2005). The pSL12-related group may also include ammonia oxidizers as inferred by the analysis of 16S rRNA and archaeal amoA genes (Mincer et al., 2007; Kato et al., 2009b). Several microdiverse phylogenetic clusters within MGI have been defined in previous reports (Massana et al., 2000; Takai et al., 2004; Durbin & Teske, 2010). Our MGI clones recovered from the overlying seawater were affiliated with the MGI-γ (Fig. S2a). Those from the Mn crust and sediment samples were affiliated with other MGI clusters such as the α, η–κ–υ, ι and ɛ–ζ–θ clusters (Fig. S2a). In the case study of the South Pacific Gyre (Durbin & Teske, 2010), the relative abundance of the MGI-α in the archaeal clone libraries has been high in the overlying seawater and those of the MGI-η and –υ have been high in the libraries from the sediments. Although it is unclear what kinds of factors are responsible for the relative abundance of each MGI cluster among deep-sea environments, these differences may reflect differences in geography, environmental characteristics and/or experimental procedures (such as the DNA extraction methods and the PCR primers used).
Numbers in parentheses indicate 95% confidential intervals.
NP, Not possible to calculate.
In contrast to Archaea, diverse bacterial phylotypes were detected in the Mn crust, sediment and seawater samples. All analyses, i.e., Chao1 species estimates and the Shannon index (Table 1) and rarefaction curves (Fig. S3), indicated that the community diversity of Bacteria in the crust sample was comparable to or higher than that in sediment and overlying seawater. In addition, the diversity of Bacteria was higher in all samples than those of Archaea (Table 1). The bacterial diversity of the Mn crust was comparable to or higher than those of seafloor basaltic rocks reported previously (Lysnes et al., 2004; Mason et al., 2007; Santelli et al., 2008), suggesting that aged Mn crusts provide a habitat for diverse Bacteria as in basaltic rocks.
Bacterial phylotypes dominated in all libraries (75.3–94.3% of the total clone numbers; Fig. 3). The bacterial phylotypes recovered were affiliated with the following phyla or uncultured clone groups (Figs 3 and S2b–e): Acidobacteria, Actinobacteria, Bacteroidetes, ‘Caldothrix’, Chlamydiae, Chloroflexi, Nitrospirae, Planctomycetes, Proteobacteria (Alpha, Beta, Gamma and Delta subdivisions), Verrucomicrobia, BRC1, KSB, NKB19, OP11, OP3, SAR406 and SBR1093. Several phylotypes were affiliated with unclassified environmental clone groups, UBSedI to VI and UBMnI and II, as defined in the present study (Fig. S2e). Phylotypes in the Gammaproteobacteria were abundant in the clone libraries from the Mn crust and sediment samples (24.0% and 23.5% of the total clone numbers, respectively; Fig. 3). These phylotypes were related to not yet cultivated environmental clones recovered from seafloor basaltic rocks (Lysnes et al., 2004; Mason et al., 2007, 2008; Santelli et al., 2008) rather than cultured species (<95% similarity) (Fig. S2b). In contrast, phylotypes in the Alphaproteobacteria were abundant in the clone libraries from the seawater sample (44.3% of the total clone number). In particular, most of them were related to Candidatus Pelagibacter (SAR11 cluster, Rappéet al., 2002) and Sphingomonadales (Fig. S2c), groups from which members have often been recovered from deep-sea water of >1000 m water depth (García-Martínez & Rodríguez-Valera, 2000; Delong et al., 2006; Kato et al., 2009a, c).
Comparative analysis showed that the microbial community composition of the Mn crust was different from those of the sediment and overlying seawater. The differences among the three communities were supported by the UniFrac significance and P values (<0.01). To compare the microbial community composition, the shared phylotype numbers among the libraries from the crust, sediment and seawater samples were estimated using sons. The Mn crust and sediment communities shared few or no phylotypes with the seawater community (Fig. 4). The Mn crust community contained a fraction of phylotypes recovered from the sediment sample (20% of the total phylotype richness estimates of the Mn crust; Fig. 4). Thus, 80% of the total phylotypes richness estimates of the Mn crust community were unique compared with the sediment communities. In fact, unique phylotypes of the Mn crust were observed in the phylogenetic trees (Fig. S2). Several phylotypes in MGI were shared between the Mn crust and sediment, but not between the Mn crust and seawater (Fig. S2a) as described above.
Phylotypes related to the genus Nitrosospira in the Betaproteobacteria were unique in the Mn crust (Fig. S2b). Representative clone 953Mn48u has 97% similarity to the ammonia-oxidizing chemolithoautotrophic bacterium Nitrosospira multiformis (Watson et al., 1971). Phylotypes related to the family Ectothiorhodospiraceae in the Gammaproteobacteria were also unique in the library of the Mn crust (Fig. S2b). Representative clone 953Mn100u has 94% similarity to the arsenite-oxidizing chemolithoautotroph Alkalilimnicola ehrlichii (Hoeft et al., 2007) or sulfur-oxidizing chemolithoautotrophic Thioalkalivibrio species (Sorokin et al., 2001). These results suggest that putative ammonia- (or in some cases, sulfur- and/or arsenite-) oxidizing chemolithoautotrophs are present on the Mn crust surface.
The detection of the phylotypes related to ammonia-oxidizing Archaea and Bacteria in the Mn crust suggests that these putative ammonia oxidizers may play a role as primary producers in the microbial ecosystem on Mn oxides that coats old seamounts in western Pacific. Although the ammonium concentration in the open ocean is generally extremely low (<5 μM) (Rees et al., 2006; Herfort et al., 2007; Agogue et al., 2008), ammonia-oxidizing Archaea belonging to MGI Crenarchaeota can grow under these conditions using ammonium as the energy source (Martens-Habbena et al., 2009). Ammonia-oxidizing bacteria can also grow at low concentrations of ammonium (Bollmann & Laanbroek, 2001; Bollmann et al., 2002). In fact, we detected both bacterial and archaeal amoA genes, which encode the alpha subunit of the ammonia monooxygenase, from DNA extracted from the Mn crust (the data will be published elsewhere). Ammonia is the most likely chemical species to be utilized as an electron donor for microbial growth on the Mn crust. Dissolved organic carbon compounds in deep-sea water may resist microbial growth (Barber, 1968). Buried organic compounds from surface seawater may be limited on the Mn crust because little sandy sediment is formed (Fig. 1a). Accordingly, H2, CH4, H2S, Fe2+ and Mn2+ from the degradation of organic compounds by anaerobes and fermenters would be limited on the Mn crust. Fe2+ and reduced sulfides contained in basaltic rocks are thought to be energy sources for the microorganisms on the rocks (Bach & Edwards, 2003; Santelli et al., 2008), but the argument is still controversial (Templeton et al., 2009). Our data suggest that ammonia in surrounding seawater is likely to be an important energy source for sustaining the microbial ecosystem on the Mn crust. Furthermore, the presence of ammonia-oxidizing bacteria on oceanic basaltic rocks has been supported by the detection of 16S rRNA genes related to these members such as Nitrosospira (Mason et al., 2008; Santelli et al., 2008). These facts lead to the hypothesis that the ammonia oxidizers play a role in the microbial ecosystem on outcrops of the global seafloor including bare young basalts and aged Mn crusts.
One of the subjects in the study of oceanic Mn nodules and crusts is the mechanism of their creation and growth. Microorganisms may play a role in the accumulation of Mn oxides by biofilm formation on rocks on the seafloor (Wang & Müller, 2009). This notion is consistent with the detection of abundant microorganisms, both Bacteria and Archaea, within/on the Mn crust (Fig. 2). Mn-oxidizing bacteria, which are thought to play a role in Mn precipitation in the first step of the biomineralization model for Mn crusts as a bioseed (Wang & Müller, 2009), have been isolated from marine environments (Tebo et al., 2005). This model is supported by the detection of phylotypes related to the Leptothrix in the Betaproteobacteria (953Asw11u; Fig. S2b), Erythrobacter and Aurantimonas in the Alphaproteobacteria (953Asw97u and 953Asw05u; Fig. S2c) and Arthrobacter in the Actinobacteria (953Asw07u; Fig. S2d), which includes marine Mn-oxidizing bacteria (Tebo et al., 2005), from the overlying seawater, but not from the Mn crust and sediment samples. Although no phylotypes related to the known Mn- or Fe-oxidizing bacteria were detected in the Mn crust and sediment, there is a possibility that as yet uncultivated Mn- or Fe-oxidizing bacteria are hidden in the diverse phylotypes detected. Further analyses, for example, isolation and characterization of Mn- and Fe-oxidizing bacteria, quantification of their abundance and determination of rates of Mn and Fe oxidation by them are required to elucidate the significance of their role in the formation of the Mn crusts. A recent study has shown that manganese precipitation is promoted by superoxide that is produced by enzymatic activity of marine bacteria (Learman et al., 2011). This biogenic superoxide is also potentially related to the precipitation of Mn in overlying seawater and on the surface of Mn crusts.
Two common features are found between the microbial communities in the oceanic Mn crust shown in the present study and those in the freshwater Mn nodules reported by Stein et al. (2001). Firstly, many bacterial phylotypes detected in the Mn crust and nodules have low similarity (<96%) to known cultured species. Secondly, the phylotypes relatively close to Hyphomicrobium in the Alphaproteobacteria and Leptothrix in the Betaproteobacteria, both of which include Mn-oxidizing bacteria, and the phylotypes close to MGI Crenarchaeota were detected in both environments. Our phylotypes related to these members were detected in the Mn crust, sediment and/or overlying seawater (Fig. S2b and c). It is unclear how these phylotypes are distributed among the Mn nodules, surrounding sediments and overlying lake water in the freshwater environment (Stein et al., 2001). Nevertheless, phylotypes related to these genera (i.e. Hyphomicrobium and Leptothrix) may play a role in Mn accumulation on solid surfaces in marine and freshwater environments.
Although numerous studies of microbial communities in coastal sediments have been conducted, those in deep-sea sediments in open oceans that are far from lands are poorly understood. Deep-sea sediments in open oceans are nutrient-poor (i.e. oligotrophic) environments (D'hondt et al., 2004), except for hydrothermal vents and cold seep areas. Previous reports have suggested that there are diverse uncultured species on the surface of such deep-sea sediments and the relative abundances of phylotypes belonging to Gammaproteobacteria and MGI Crenarchaeota are high in these environments (Li et al., 1999; Vetriani et al., 1999; Bowman & Mccuaig, 2003; Schauer et al., 2009; Durbin & Teske, 2010). These results are compatible with our results of the sediment sample (Figs 3 and S2), suggesting that these features of the microbial community structures may be common on the surface of oligotrophic deep-sea sediments in open oceans.
We would like to thank the crew of the R/V Natsushima and the operation team of the ROV Hyper-Dolphin for their cooperation in sample collection. We would like to thank Dr Blair Thornton for providing the on-site photograph of the Mn crust and for English language editing. We would like to thank Ms Satomi Minamizawa for her technical assistant on the cruise. We are also grateful to the scientists who joined the NT09-02 cruise and to Dr Katsuhiko Suzuki and the other members of the Project TAIGA for providing valuable samples and for helpful discussions. We would like to thank two anonymous reviewers for their helpful comments. This research was funded by the Ministry of Education, Culture, Science and Technology (MEXT), Japan, through a special coordination fund (Project TAIGA: Trans-crustal Advection and In-situ biogeochemical processes of Global sub-seafloor Aquifer).