Microbial community diversity in seafloor basalt from the Arctic spreading ridges


*Corresponding author. Tel.: +47-55588187; fax: +47-55589671, E-mail address: kristine.lysnes@bio.uib.no


Microbial communities inhabiting recent (leqslant R: less-than-or-eq, slant1 million years old; Ma) seafloor basalts from the Arctic spreading ridges were analyzed using traditional enrichment culturing methods in combination with culture-independent molecular phylogenetic techniques. Fragments of 16S rDNA were amplified from the basalt samples by polymerase chain reaction, and fingerprints of the bacterial and archaeal communities were generated using denaturing gradient gel electrophoresis. This analysis indicates a substantial degree of complexity in the samples studied, showing 20–40 dominating bands per profile for the bacterial assemblages. For the archaeal assemblages, a much lower number of bands (6–12) were detected. The phylogenetic affiliations of the predominant electrophoretic bands were inferred by performing a comparative 16S rRNA gene sequence analysis. Sequences obtained from basalts affiliated with eight main phylogenetic groups of Bacteria, but were limited to only one group of the Archaea. The most frequently retrieved bacterial sequences affiliated with the γ-proteobacteria, α-proteobacteria, Chloroflexi, Firmicutes, and Actinobacteria. The archaeal sequences were restricted to the marine Group 1: Crenarchaeota. Our results indicate that the basalt harbors a distinctive microbial community, as the majority of the sequences differed from those retrieved from the surrounding seawater as well as from sequences previously reported from seawater and deep-sea sediments. Most of the sequences did not match precisely any sequences in the database, indicating that the indigenous Arctic ridge basalt microbial community is yet uncharacterized. Results from enrichment cultures showed that autolithotrophic methanogens and iron reducing bacteria were present in the seafloor basalts. We suggest that microbial catalyzed cycling of iron may be important in low-temperature alteration of ocean crust basalt. The phylogenetic and physiological diversity of the seafloor basalt microorganisms differed from those previously reported from deep-sea hydrothermal systems.


Bacterial communities and processes in the subsurface have received increasing scientific attention over the last decade. The deep biosphere is estimated to contain a biomass in the same order of magnitude as that of the surface of this planet [1,2], with the major fraction of this biomass residing in the marine subsurface. Microbial surveys of ocean sediments have shown that high microbial diversities and large bacterial populations are present in marine sediment deposits[3]. Microorganisms have also been shown to inhabit the upper volcanic layer (∼500 m) of the oceanic crust. The presence of microbial DNA in characteristic alteration textures in the glassy margins of the basaltic lava, as well as biological carbon isotope signatures of disseminated carbonate, indicate that microorganisms participate in the alteration of the glass and the chemical exchange between oceanic crust and seawater [4–8]. As the phylogenetic and physiological diversity of these microorganisms is still largely unknown, the mechanisms for the suggested microbial influenced dissolution and alteration are not known.

Several studies of deep-sea hydrothermal vents have noted distinctive differences in the microbial communities inhibiting the hydrothermal fluids and the background seawater, indicating that vent fluids transport microorganisms from the subsurface to the seafloor [9–12]. In the recent studies of diffuse flow vent fluids (<50 °C) following a volcanic eruption on the Juan de Fuca ridge [11,12], the most frequently retrieved DNA sequences were affiliated with the ɛproteobacteria[11]. In the study by Cowen et al.[13] of 65 °C subseafloor hydrothermal fluids from 3 Ma sediment covered crust on the flank of the Juan de Fuca ridge, most of the retrieved DNA clones were related to the ammonia-producing bacterium Ammonifex degensii. The reported data indicate the presence of a diverse bacterial and archaeal community in these hydrothermal regions of the ocean crust, where the most important microbial processes appear to be reduction and oxidation of sulfur and nitrogen compounds [11,13].

In contrast, psychrophilic iron oxidizers were detected and isolated from low-temperature (∼4 °C) seafloor habitats of sulfide minerals and metalliferous sediments at the Juan de Fuca ridge[14]. Two groups of iron-oxidizing bacteria were isolated; one group was related to the γ-proteobacterium Marinobacter aquaeolei and the other to the α-proteobacterium Hyphomonas jannaschiana. The iron oxidizing bacteria were isolated from surfaces of weathered rock, and data suggest that these microorganisms participate in rock weathering at the seafloor. The microbial populations found in these low-temperature habitats differed from those of hydrothermal habitats from the same ridge [11–13].

Evidence for a diverse microbial community and microbial iron oxidation in young (<1 Ma) non-hydrothermal seafloor basalts from the Knipovich ridge has also recently been reported[15]. In this study, the presence of numerous endolithic microbial cells of various morphologies (e.g. cocci, rods, filaments, star-shaped cells, and twisted stalks resembling those of Gallionella ferruginea) was revealed by scanning electron microscopy (SEM). 16S rRNA gene amplification, denaturing gradient gel electrophoresis (DGGE) and sequencing analysis showed that the microbial community comprised the phylogenetic groups Bacteroidetes, γ- and ɛproteobacteria, and marine Group 1: Crenarchaeota. Even though cell morphologies resembling Gallionella were observed in these samples, G. ferruginea was not detected by molecular biology methods. However, a later study of basalts collected from shallow areas around Jan Mayen retrieved 16S rDNA sequences matching Gallionella[16]. The bacterial populations appeared to be characteristic and unique for the rock environment and differed from those found in associated sediments and seawater[15]. Investigation of 14–28 Ma non-hydrothermal subseafloor basalt from the northern flank of the Southeast Indian ridge also showed the presence of microbial populations characteristic for the rock environment, with Actinobacteria, Chloroflexi, Bacteroidetes, Firmicutes, and β- and γ-proteobacteria as the dominating phylogenetic groups[17]. Enrichment studies showed that iron reducing bacteria and autolithotrophic methanogens were present.

SEM observations of microbial cells in seafloor basalts from different ridges suggest that endolithic microbial growth is a general feature of ocean ridges [15,18,19]. From analysis of the organic carbon content in recent lava flows, a biomass of ∼106 cells/cm3 seafloor basalt has been estimated[19]. With time, the colonized seafloor basalt becomes buried by younger lava flows and sediments and brought to deeper levels of the oceanic crust. Consequently, a significant proportion of the alteration observed in subsurface lavas may have developed before burial[18]. More knowledge of microbial phylogeny and processes in young ocean ridge basalt is thus important when studying the deep biosphere and evaluating the function and impact this has on the alteration of basalt.

In the present study, samples of recent, non-hydrothermal seafloor basalt from various depths in the axial valleys of the Kolbeinsey, Mohns, and Knipovich ridges were investigated. The phylogenetic and functional diversity of microorganisms were studied by PCR-based fingerprinting methods in combination with enrichment techniques. Together with physical parameters of the samples and their original habitats, these results were used to draw attention to biogeochemical and microbial processes involved in basalt alteration. Samples of bottom and surface seawater, as well as one sample of basalt-associated sediment, were analyzed for comparison.

2Materials and methods

2.1Sample collection and preparation

During three cruises with R/V Håkon Mosby to the Norwegian-Greenland Sea in July–August 1999, August–September 2000, and July–August 2001, a total of 22 samples of recent basalt flows were collected from the Kolbeinsey, Mohns, and Knipovich ridges (Fig. 1, Table 1). Sampling was done either by dredging or by use of a remote operated vehicle (ROV). In addition, one sample of sediment attached to dredged basalt, four bottom seawater samples and two surface seawater samples were collected for comparison and for control of contamination (Table 1). The bottom seawater was sampled by Niskin bottles attached to a CTD (conductivity, temperature, density) rosette, whereas the surface seawater was sampled with a water hose.

Figure 1.

Location of sampling sites on the Arctic spreading ridges.

Table 1.  Location and depth (meters below sea level: mbsl) of sampling sites and character of samples
SampleSample modeLocationRidgeDepth (mbsl)Sample characterAge (Ma)DNA seq.
  1. For the dredged samples, the coordinates and depths given are the start and end point of dredging. “DNA seq.” indicates that native DNA sequences (Figs. 4 and 5) are presented for Archaea (a) and Bacteria (b), respectively.

SM99-1RROVN69° 06.87′–W16° 04.46′Kolbeinsey861Glassy basalt0.1 
SM99-2RROVN68° 56.68′–W17° 13.00′Kolbeinsey1212Glassy basalt0.1B
SM99-5RROVN68° 56.89′–W17° 11.70′Kolbeinsey1191Glassy basalt0.1 
SM99-6RROVN68° 56.89′–W17° 11.70′Kolbeinsey1191Glassy basalt0.1A + B
SM99-7RROVN68° 23.00′–W18° 00.00′Kolbeinsey730Glassy basalt0.1 
SM00-2DDredgedN72° 19.90′–E1° 29.57′–N72° 19.75′–E1° 29.87′Mohns2291–2459Glassy basalt0.1A + B
SM00-4DDredgedN72° 40.14′–E2° 53.70′–N72° 40.26′–E2° 53.20′Mohns1216–1404Basalt0.1B
SM00-5RROVN72° 40.95′–E2° 50.83′Mohns716Basalt1B
SM00-6DDredgedN72° 38.12′–E2° 42.00′–N72° 38.35′–E2° 41.68′Mohns2090–2201Glassy basalt0.1B
SM00-7DDredgedN72° 39.33′–E2° 40.87′–N72° 39.44′–E2° 40.86′Mohns920–950Basalt0.1 
SM00-10DDredgedN72° 40.69′–E2° 42.41′–N72° 40.41′–E2° 42.53′Mohns944–1600Glassy basalt0.1B
SM00-17DDredgedN72° 29.03′–E2° 40.19′–N72° 28.81′–E239.68′Mohns2201–2263Glassy basalt0.1A + B
SM00-18DDredgedN72° 26.04′–E2° 31.35′–N72° 2.85′–E232.03′Mohns2562–2598Glassy basalt0.1A + B
SM00-19DDredgedN72° 29.95′–E2° 45.70′–N72° 30.05′–E244.90′Mohns2523–2751Glassy basalt0.1B
SM00-20DDredgedN72° 31.74′–E2° 32.30′–N72° 31.84′–E229.99′Mohns2475–2681Glassy basalt0.1A + B
SM00-22DDredgedN72° 28.71′–E2° 26.59′–N72° 29.48′–E226.07′Mohns2240–2716Glassy basalt0.1B
SM00-25DDredgedN72° 23.08′–E1° 43.24′–N72° 23.24′–E142.04′Mohns2506–2695Glassy basalt0.1B
SM00-27DDredgedN72° 16.52′–E1° 19′–N72° 16.51′–E120.81′Mohns2642–2853Glassy basalt0.1B
SM00-47DDredgedN72° 44.53′–E3° 56.31′–N72° 44.92′–E357.79′Mohns2500–2700Glassy basalt0.1B
SM00-52DDredgedN76° 47.68′–E7° 22.60′–N76° 48.42′–E722.70′Knipovich3204–3385Glassy basalt2 × 10 −5B
SM00-54DDredgedN76° 48.12′–E7° 24.62′–N76° 48.92′–E7° 24.74′Knipovich3132–3152Glassy basalt2 × 10−5B
SM00-55DDredgedN76° 48.20′–E7° 25.01′–N76° 49.88′–E7° 25.28′Knipovich3150–3344Glassy basalt2 × 10−5B
SM00-2DSDredgedN72° 19.90′–E1° 29.57′–N72° 19.75′–E1° 29.87′Mohns2291–2459Sediment A + B
SM00-50SWBottleN72° 50.99′–E4° 15.86′Mohns2326Seawater A + B
SM00-65SWBottleN76° 57.03′–E7° 15.04′Knipovich3186Seawater A + B
SM00-SSWWater hose Knipovich0Surface seawater  
SM01-8SWBottleN71° 09.89′–W7° 51.09′Mohns950 mSeawater A + B
SM01-45SWBottleN71° 10.33′–W7° 52.87′Mohns650 mSeawater A + B
SM01-SSWWater hose Mohns0Surface seawater B

The water depth at the sampling sites ranged from 716 to 3385 m, and the ambient bottom seawater temperature between –0.4 and –0.8 °C, respectively. Based on the very low degree of alteration, the samples from the rift valley of Knipovich ridge were estimated to be less than 20 years old. The oldest sample (SM00-5R) was collected from a seamount north of the rift valley of the Mohns ridge, and is estimated, based on paleomagnetic data, to be around 1–2 Ma. The other samples were all collected from relatively young volcanoes within the neovolcanic zones of the Mohns and Kolbeinsey ridges, and are roughly estimated to be less than 100,000 year, since the volcanoes were yet not covered by deep-sea sediments.

Visual inspection by cameras attached to the ROV showed that the lava flows were covered by a thin layer of sediment. At shallow sites (leqslant R: less-than-or-eq, slant1200 m below sea level; mbsl) the basalt samples were colonized by a sessile macrofauna. At deeper sites, only few of the basalt samples were colonized by a macrofauna, and samples without macrofauna were chosen from these sites.

The samples were processed immediately after collection. Basalt was split with hammer and chisel and crushed in a mortar. Fragments of glassy margins with visible alteration on the surface and along fractures and cracks were preferred. All tools for rock crushing, as well as aluminum foil covering the workbench, were flame-sterilized. Two alternative applications for extracting cell material for the molecular phylogenetic techniques were used: (1) cells in the basalt samples were extracted by rinsing cracks- and fracture surfaces with sterile water; (2) crushed basalt or sediment were mixed thoroughly with phosphate buffer saline (PBS: 137 mM NaCl, 2.7 mM KCl, 4.3 mM Na2HPO4, 1.4 mM KH2PO4 (pH 7.3)) in sterile 50-ml plastic centrifuge tubes. After sedimentation of the visible solid particles, 5–10 ml of the supernatant was centrifuged to achieve a cell pellet. Seawater samples, also 5–10 ml, were centrifuged directly. The cell suspensions and cell pellets were frozen (−20 °C) for later onshore DNA amplification.

2.216S rRNA gene amplification, DGGE and sequencing

PCR amplification of the hyper variable V3-region of 16S rDNA using HotStar Taq DNA Polymerase (Qiagen, Germany) and the bacterial primers PRBA338f and PRUN518r[20], the first with a 40-nucleotide long GC clamp[21], was performed in a GeneAmp2400 thermal cycler (Perkin–Elmer Applied Biosystems, USA). The reaction mixture was prepared as recommended by the manufacturer[22]. The PCR program was: 95 °C, 15 min; 30 (up to 40 for 1999 samples) cycles of denaturation at 94 °C, 1 min; annealing 55 °C, 30 s; extension 72 °C, 1 min; and a final extension 72 °C, 10 min. The sample material from the 1999-cruise was limited, so to achieve sufficient PCR product for further analysis up to 40 cycles of amplification had to be run for these samples.

Amplification of archaeal 16S rDNA V3-region was a nested PCR protocol modified from the method of Øvreås et al.[20], with the primers mPRA46f (5′-C/TTA AGC CAT GC/TA/G AGT-3′) and mPREA1100r (5′-C/G/TGG GTC TCG CTC GTT A/GCC-3′) used for the outer reaction and the primers mPARCH340f (5′-TAC/T GGG GC/TG CAC/G CAG-3′) and PARCH519r[20], the first with a 40-nucleotide long GC clamp[21], for the inner reaction. The reaction mixture used was as for the bacterial DNA amplification. In addition, 5% acetamide was added to the mixture for the outer PCR reaction. The PCR program used for both reactions was: 95 °C, 15 min; 35 cycles of denaturation at 94 °C, 30 s; annealing at 53 °C for the outer reaction and 54 °C for the inner reaction, 30 s; extension 72 °C, 1 min; and a final extension 72 °C, 10 min. In order to maintain an appropriate control reaction during the amplification process, several controls were processed to avoid confusion with false-positive or false-negative results. DNA from Archaeoglobus fulgidus, Methanococcus voltae, and Halobacterium sp. was used as negative controls for bacterial primers and positive controls for archaeal primers, and DNA from Escherichia coli as negative controls for archaeal primers and positive controls for bacterial primers. A non-template control (added distilled water instead of template) was always included, as control for contamination from air or PCR reagents.

Amplification products were verified by visualization on 1.5% agarose gels stained with ethidium bromide.

The amplified products of microbial community were further analyzed on 1 mm thick, 8% polyacrylamide gels, as described previously[20]. The linear gradient of urea and formamide ranged from 25% to 65%. Electrophoresis was carried out at 65 °C, 70 V, and 18 h in 0.5X TAE buffer. After electrophoresis, the gels were stained with SYBR Gold (Molecular Probes, Eugene, OR) in 1X TAE buffer for 30 min and photographed. Several replicate PCR amplifications and DGGE gels were run in order to consider the reproducibility of the community profiles, which were all highly reproducible.

Images of DGGE gels were analyzed using the Gel2k image-analyzing software program developed by Svein Norland (Department of Microbiology, University of Bergen). The program was used to recognize the number of visible DNA bands.

DNA fragments to be sequenced were excised from the gel and processed as described earlier[20]. The sequences were analyzed using the BLAST tool[23] at the National Center for Biotechnology Information (NCBI). Sequences with a low degree of similarity to the NCBI database sequences were investigated for possible chimera using the CHECK_CHIMERA program[24], which is available through the Ribosomal Database Project (RDP). Evolutionary distances of the DNA sequences were calculated using Clustal_X[25] and phylogenetic trees were constructed using the neighbor-joining algorithm. In order to simplify the graphic presentation of the results, identical or near-identical (geqslant R: gt-or-equal, slanted98% similarity) sequences are presented as one operational taxonomic unit (OTU) in the phylogenetic trees. Pair-wise comparisons of similar DNA partial sequences were performed using the BLAST 2 Sequences tool[26] at the NCBI.

The sequences reported in this paper have been deposited in the GenBank database under accession numbers AY463804 to AY463909, AY505431 to AY505438, AY505441 to AY505443, and AY505445 to AY505455.

2.3Primary enrichment cultures

Subsets of different media were used in order to enrich for microorganisms participating in the cycling of iron (FePPi medium[27]; Fe-reducer medium[28]; Fe-TSB medium[29]), manganese (PYGV[30]; PC medium[30]), sulfur (thiosulphate medium[31]; synthetic seawater medium designated “W20” modified from[32] as described by[33]; PM1[34]), and methane (methanogenic medium 2[35]; NMS[36]). In addition, natural seawater enrichment medium, with and without the addition of an organic carbon source (formate, methanol, ethanol, acetate, lactate, butyrate, succinate, or caproate), in concentrations of 10–20 mM in the final medium, was prepared with 0.2 μm filtered and autoclaved aged seawater buffered with 30 ml 1 M sodium bicarbonate (NaHCO3) per liter medium. For enrichment of anaerobes, media were flushed with a mixture of N2:CO2 (90:10). Enrichment medium (30 ml) was added to each 50 ml serum bottle with ∼20 ml headspace, except for media amended with H2 or CH4, which were added to 100 ml bottles with ∼70 ml headspace.

Approximately 1 g of crushed basalt or sediment was added to each bottle containing enrichment medium. For anaerobic media, this inoculation was performed inside a disposable glove bag (Glove Bag model S30-20, Instruments for Research and Industry Inc., Cheltenham, PA, USA) flushed and filled with N2. Seawater samples were introduced into the bottles in 3 ml portions using a syringe. A total of 171 basalt enrichment cultures were inoculated onboard the ship. Identical culture media were inoculated with either sediment or seawater. A total of 35 seawater enrichments and 18 sediment enrichments were prepared. The enrichment cultures were incubated at 4 °C. Cultures were examined on shore by phase contrast microscopy after 1, 2, and 4 months' incubation in order to determine occurrence of growth. From cultures with observed growth, a subsample was centrifuged and the cell pellet frozen for 16S rRNA gene amplification, DGGE and sequencing.

Sulfide was measured in enrichment cultures after 1 month incubation according to the method described by Cord–Ruwisch[37]. Reduction of ferric iron (Fe3+) to ferrous iron (Fe2+) was observed as a color-change from reddish brown to gray after 1, 2, and 4 months. Methane in the gas phase of the methanogenic enrichment cultures was measured after 4 months of incubation using a gas chromatograph (Hewlett Packard 6890) with helium as carrier gas, as described by the manufacturer[38]. The gas chromatograph was equipped with a thermal conductivity detector (TCD) and a HaySep R packed column.


3.116S rDNA amplification and DGGE community profiles

A total of 22 basalt samples, one sediment sample and six water samples were processed for microbial community analyses. DNA was successfully amplified using bacterial primers from all basalt, sediment and seawater samples (Table 1), with the exception of one basalt sample (SM99-7R). Due to complexity of banding patterns on bacterial DGGE gels, excised bands from the basalt samples SM99-1R, -5R, and SM00-7D, and the surface seawater sample SM00-SSW did not result in clean 16S rDNA sequences. Archaeal primers amplified DNA from six of the 22 basalt samples (SM99-6R, SM00-2D, -17D, -18D, -20D, and -54D), the sediment sample, and the four bottom seawater samples, but we were unable to retrieve any archaeal PCR products from the two surface seawater samples and the remaining basalt samples. For amplification of archaeal DNA, visible PCR products appeared only after the second round of the nested PCR procedure. All excised bands from the archaeal DGGE gel resulted in clean sequences. No visible PCR products were detected in any of the blanks, either with bacterial or archaeal primers, containing distilled water instead of DNA template.

Microbial community profiles for bacterial and archaeal populations inhabiting the basalt, sediment, and seawater samples were obtained as banding patterns on DGGE gels (Fig. 2). The bacterial DNA banding patterns obtained for the basalt samples were complex and 20-40 bands were generally detected, using the Gel2k image-analyzing software program. For the sediment and seawater samples, 17 and 21–28 bands were detected, respectively. Pronounced changes in relative brightness of DGGE bands were observed between the different samples. Fragments of the 16S rDNA amplified from different microbial communities showed varying degree of sharpness on the DGGE gels. Most of the DGGE profiles appeared to have bands in common, but no banding patterns were identical to each other, suggesting that the basalts were inhabited by similar, but not identical, bacterial populations. Most basalt samples were observed to have 1–2 dominant bands (e.g., bands 119 and 121, Fig. 2(a)). Sediment and seawater samples were observed to have between 1 and 6 dominant bands.

Figure 2.

(a) DGGE banding patterns of a selection of samples (one surface seawater, three basalts) of PCR-amplified partial 16S rDNA with specific primers for Bacteria and (b) Archaea (all samples with a PCR product). Sample number is indicated for each lane. M, marker.

The archaeal DNA band patterns were unique for each individual sample, except for the four bottom seawater samples that showed pair wise (SM00-50SW and -65SW; SM01-45SW and -8SW) similar DNA profiles to each other, as shown in Fig. 2(b). The PCR amplifications were composed of 6–12 resolvable bands on the gel, and the bands were mainly positioned in the middle of the gel.

3.2Phylogenetic diversity of bacteria

Sequencing analysis of the individual bacterial DGGE fragments were carried out on the 18 environmental basalt samples SM99-2R, -6R, SM00-2D, -4D, -5R, -6D, -10D, -17D, -18D, -19D, -20D, -22D, -25D, -27D, -47D, -52D, -54D, -55D, the sediment sample SM00-2DS, the four bottom seawater samples SM00-50SW, -65SW, SM01-8SW, -45SW, and the surface seawater sample SM01–SSW. The partial DNA sequences retrieved from basalt microbial communities (a total of 65 sequences) were phylogenetically affiliated with eight main groups of the domain Bacteria: Firmicutes, Chloroflexi, Actinobacteria, Bacteroidetes, and the α-, γ-, δ-, and ɛproteobacteria (Figs. 3,4). In addition, 1 sequence (SM00-19D-300N, Fig. 4(a)) could not be assigned to any known phylum. The closest match to this sequence was a candidate division SBR1093[39], with 90% sequence similarity. The SM00-19D-300N sequence was not found to be a chimera, using the CHECK_CHIMERA program.

Figure 3.

Composition of the bacterial populations in basalt samples of different crustal age based on distribution of 16S rRNA gene sequences into different main bacterial divisions. The composition of the background seawater is shown as comparison. The figure is drawn based on 3 samples, 7 sequences for the 20 Yr basalts; 14 samples, 52 sequences for the 0,1 Ma basalts; 1 sample, 6 sequences for the 1 Ma basalts; and 5 samples, 19 sequences for the background seawater.

Figure 4.

Figure 4.

Phylogenetic tree based on partial 16S RNA gene sequences from bacteria inhabiting environmental basalt (in bold, sequence names ends with N), sediment, and seawater samples compared to reference strain sequences from GenBank. Reference sequences were cut to the same length as the sample sequences. The scale bar corresponds to 0.1 change per nucleotide, and GenBank accession numbers are in parenthesis. OTU1N = SM00-2D, -25D, -47D; OTU2N = SM00-4D, -5R, -6D, -18D, -20D; OTU3N = SM00-10D, -19D; OTU4N = SM00-20D, -55D; OTU5N = SM00-10D, -20D, -22D, -47D, -55D; OTU6N = SM00-2D, -22D; OTU7N = SM99-2R, SM00-2D, -27D. α- and γ-proteobacteria are presented in a separate phylogenetic tree (b).

Figure 4.

Figure 4.

Phylogenetic tree based on partial 16S RNA gene sequences from bacteria inhabiting environmental basalt (in bold, sequence names ends with N), sediment, and seawater samples compared to reference strain sequences from GenBank. Reference sequences were cut to the same length as the sample sequences. The scale bar corresponds to 0.1 change per nucleotide, and GenBank accession numbers are in parenthesis. OTU1N = SM00-2D, -25D, -47D; OTU2N = SM00-4D, -5R, -6D, -18D, -20D; OTU3N = SM00-10D, -19D; OTU4N = SM00-20D, -55D; OTU5N = SM00-10D, -20D, -22D, -47D, -55D; OTU6N = SM00-2D, -22D; OTU7N = SM99-2R, SM00-2D, -27D. α- and γ-proteobacteria are presented in a separate phylogenetic tree (b).

A summary of the phylogenetic data from basalt, divided into different crustal age (∼20 Yr, 0.1 Ma, and 1 Ma) together with data retrieved from the seawater samples is shown in Fig. 3. The most frequently retrieved 16S rRNA gene sequences from basalts grouped within γ-proteobacteria (18 sequences), α-proteobacteria (11 sequences), Chloroflexi (11 sequences), Firmicutes (10 sequences), and Actinobacteria (9 sequences). The γ-proteobacteria and the Firmicutes were the only phylogenetic groups that were detected in basalt from all three age groups. Sequences affiliated with α-proteobacteria and Bacteroidetes (3 sequences) were found in 20 Yr and 0.1 Ma basalts, but not in the oldest basalt. Actinobacteria were found in 0.1 and 1 Ma basalt, but not in the youngest basalt. Chloroflexi, ɛproteobacteria (1 sequence) and γ-proteobacteria (1 sequence) were only obtained from the 0.1 Ma basalts. The sequences retrieved from the basalt samples show distinct differences in phylogenetic affiliation compared to those derived from seawater. In the seawater samples, two of the six different phylogenetic groups documented, Verrucomicrobia and the plastids, were not present in basalts. Two sequences were closely related to phytoplankton plastids derived from photosynthetic organism, both most closely related to the plastid of the Haptophyceae Ochrosophaera neapolitana. Also in seawater, the abundance of δ-proteobacteria was high, whereas only one sequence affiliating with this group was retrieved from basalt.

Fig. 4 shows the phylogenetic position among the bacterial sequences from basalt and seawater. The most abundant sequences retrieved in this study belonged to the γ-proteobacteria (Fig. 4(b)). Most sites (12 of 18 sites) studied had representatives from this group, and a high heterogeneity of sequences was detected. One sequence (SM00-52D-119N), which was recovered from a dominant band (Fig. 2(a)), showed phylogenetic affiliation to the marine oligotrophic bacterium Hyphomicrobium indicum (98%). One sequence from one of the youngest lava flows (SM00-52D-121N, corresponding to dominant band 121 in Fig. 2(a)) was similar to a sequence from the iron reducer Shewanella frigidimarina (95%). Three sequences clustered together with Marinobacter sp. Five sequences from the young basalt (OTU5N), and one sequence from the 0.1 Ma basalt grouped together with Acinetobacter junii and an uncultured γ-proteobacteria. One sequence from the 1 Ma basalt was affiliated to Pseudomonas stutzeri, whereas several of the sequences retrieved from seawater were affiliated to Pseudoalteromonas. One sequence from the 0.1 Ma basalt showed phylogenetic affiliation to a novel moderately halophilic bacterium, Salinisphaera shabanense, isolated from the Shabana Deep, which represent a new deeply branching lineage within the γ-proteobacteria. Also among the α-proteobacteria several different sequences were retrieved (Fig. 4(b)). However, the 11 sequences retrieved originate from only five different sites. Three sequences (OTU6N (2 sequences) and SM00-22D-295N) showed phylogenetic affiliation to the marine oligotrophic prosthecate bacteria Brevundimonas sp. (100%), and Hyphomonas oceanitis (100%), respectively.

The Chloroflexi sequences from this study were all from the 0.1 Ma basalts (Fig. 4(a)). Almost half of the sequences (5 of 11) affiliating with Chloroflexi were from one sample (SM00-10D) from the Mohns ridge. There was a high heterogeneity of sequences within this group and, with the exception of one sequence resembling Dehalococcoides sp., all retrieved sequences resembled previously uncultured strains. Within Firmicutes, there was a much lower heterogeneity, even though this group contained representatives of all three age groups (20 Yr, 0.1 and 1 Ma basalts) and all 10 sequences were from different samples. One sequence affiliated with the psychrophilic bacterium Clostridium estertheticum and the other nine sequences matched an uncultured bacterium (AF365631), which was sampled from coral microfauna. The Actinobacteria sequences retrieved comprised nine basalt sequences and two seawater sequences grouping into roughly four different clusters: one cluster of microbial populations from the 0.1 and 1 Ma basalts affiliating with uncultured bacteria from a metal contaminated soil (AY124390) and a subseafloor basalt environment (AY129940); one cluster of two seawater sequences affiliating with Actinobacterium K20-72, also a metal-contaminated soil clone; one cluster of 0.1 and 1 Ma basalts affiliating with the soil bacterium Arthrobacter globiformis; and one cluster of the 0.1 Ma sample SM00-20D-480N affiliating with Propionibacterium sp., isolated from moderately acidic mine drainage waters. The Bacteroidetes group was divided into one cluster of three 20 Yr and 0.1 Ma basalt sequences and one seawater sequence affiliated with Flavobacterium sp., isolated from the northern Baltic Sea, and one cluster of three seawater sequences affiliated with Polaribacter irgensii, previously found in Arctic and Antarctic sea ice. One sequence from the 0.1 Ma basalt belonged to ɛproteobacteria, and showed highest phylogenetic affiliation to a sequence recently recovered from a study by Huber et al.[11] on bacterial diversity in a subsurface habitat following a deep-sea volcanic eruption. The δ-proteobacteria sequences from this study were mainly found in the seawater samples, but also one sequence from the 0.1 Ma basalt was retrieved. All δ-proteobacteria seawater sequences were affiliated to a sequence from a study from the Arctic Ocean, whereas the sequence from the basalt was most similar to a psychrophilic sulfate reducing isolate from marine Arctic sediments. Sequences affiliated to plastids and Verrucomicrobia were only retrieved from seawater samples.

The bacterial DNA sequences retrieved from bottom seawater (16 sequences), surface seawater (3 sequences), and sediment (1 sequence) differed from those retrieved from basalt samples (Fig. 4). No sequences retrieved from seawater were more than 89% similar to sequences from basalt communities.

3.3Phylogenetic diversity of Archaea

Sequencing of archaeal 16S rRNA gene DGGE band fragments were carried out on the 0.1 Ma basalt samples (SM99-6R, SM00-2D, -17D, -18D, -20D, -54D), the sediment sample SM00-2DS, and the bottom seawater samples SM00-50SW, -65SW, SM01-8SW, and -45SW. All archaeal sequences retrieved (21 from basalt, 3 from sediment, and 12 from seawater) belonged to the non-thermophilic marine Group 1: Crenarchaeota (Fig. 5). The archaeal DNA sequences from basalt, sediment, and seawater samples were similar, but not identical.

Figure 5.

Phylogenetic tree based on partial 16S rRNA gene sequences from Archaea inhabiting environmental basalt samples (in bold, sequence names ends with N), sediment sample, seawater samples, and reference strain sequences obtained from GenBank. Reference sequences were cut to the same length as the sample sequences. The scale bar corresponds to 0.1 change per nucleotide, and GenBank accession numbers are in parenthesis.

Seven sequences from 0.1 Ma basalt and tree sequences retrieved from sediment clustered together with an uncultured archaeon from the 11,000 m deep Mariana trench (D87350) and an uncultured archaeon from a deep-sea carbonate crust. Overall the archaeal sequences resembled uncultured strains previously found in seawater [40–42], deep-sea sediments[43], and deep-sea carbonate crusts[44].

3.4Primary enrichment cultures

After 4 months of incubation, growth was observed in 82 of the 171 enrichment cultures inoculated with basalt. In addition, growth was observed in 24 of the 35 cultures inoculated with seawater and in 14 of the 18 cultures inoculated with sediment. Enrichments with observed growth were screened using PCR-DGGE analysis.

Reduction of ferric iron (Fe3+) to ferrous iron (Fe2+) was observed in 4 of the 22 basalt enrichment cultures with “Fe-TSB” medium (SM00-2D, -7D, -47D, and -54D). This medium was amended with organic compounds as sources of carbon and electron donor. Enrichments SM00-7D, -47D, and -54D had a dominant band in common, which was not seen in the SM00-2D culture. Sequencing of this band position from SM00-7D resulted in a sequence forming part of OTU11E and affiliating with Shewanella frigidimarina (100% identity) and the environmental sequence SM00-52D-121N (Fig. 6). The SM00-47D and -54D bands in this position were not successfully sequenced, because we were unable to re-amplify the excised DNA. The SM00-7D culture had an additional dominant band, not seen in the other cultures. Sequencing analysis of this unique band (SM00-7D) showed that the DNA sequence also resembled the iron-reducer Shewanella frigidimarina, but at a lower similarity (99%). Iron reduction was not detected in any of the enrichments inoculated with seawater (eight cultures) or sediment (five cultures), nor in any enrichment cultures with “FePPI” or “Fe-reducer” medium, either with H2 or organic compounds as electron donors.

Figure 6.

Phylogenetic tree of sequences from enriched bacteria (in bold italic, sequence names ends with E) compared to environmental sequences (in bold, see Fig. 4) from basalt. Only those enrichment culture sequences displaying at least 92% similarity to environmental sequences are included. Reference strains are from GenBank. The scale bar corresponds to 0.1 change per nucleotide, and GenBank accession numbers are in parenthesis. OTU8E = SM00-47D, -52D; OTU9E = SM00-27D, -47D, -52D, -54D; OTU10E = SM00-27D, -52D; OTU11E = SM99-5R, -7R, SM00-7D, -18D, -22D, -27D, -52D, -54D; OTU12E = SM00-18D, -47D, -52D, -54D; OTU13E = SM99-5R, -6R.

Sulfide production, indicating the presence of sulfate-reducing bacteria (SRB), was not detected in any of the 24 basalt enrichment cultures with “W20” medium. In contrast, sulfide was produced in corresponding enrichments inoculated with bottom and surface seawater (6 out of 8) and sediment (2 out of 4).

Methane was produced in two out of seven basalt enrichment cultures with methanogenic media and H2 and CO2 as energy- and carbon sources. These methane-producing cultures originated from the basalt samples SM00-2D and -18D, but archaeal 16S rRNA gene sequences retrieved from these cultures did not match sequences from any known methanogens. Methane production was not detected in the eight enrichment cultures inoculated with basalt and amended with acetate, lactate, or trimethylamine, nor was it detected in corresponding enrichment cultures inoculated with bottom and surface seawater (three cultures) or sediment (four cultures). The archaeal DGGE profiles in the two basalt cultures where methane production was detected differed (results not shown). In the SM00-2D DGGE pattern, only three bands were detected, whereas in the SM00-18D pattern, seven bands were seen. One band was present in both profiles, but sequencing of these bands (SM00-2D-857 and –18D-853) showed that they differed (Fig. 7).

Figure 7.

Phylogenetic tree of archaeal partial sequences from enrichment cultures (in bold italic, sequence names ends with E) compared to environmental sequences (in bold, see Fig. 5) from basalt. Only those enrichment culture sequences displaying at least 92% sequence similarity to native sequences are included. Reference strains are from GenBank. The scale bar corresponds to 0.1 change per nucleotide, and GenBank accession numbers are in parenthesis.

The bacterial DNA sequences from enrichment cultures inoculated with basalt (a total of 136 sequences) affiliated with the same eight main phylogenetic groups of the domain Bacteria that were present in the native basalt samples. The major fraction of the sequences from enrichment cultures grouped within the γ-proteobacteria (114 sequences). Twenty-nine sequences retrieved from basalt enrichment cultures were at least 92% similar to environmental basalt sequences (Fig. 6). Sequences from environmental and enriched basalt communities that were 98–100% similar were only found within the Firmicutes resembling Clostridium estertheticum (SM00-17D-287N and SM99-2R-2E and OTU 8E)[45]. All other sequences were less than 98% similar.

Chloroflexi, one of the most frequently retrieved phylogenetic groups from environmental samples, was only represented by one sequence from enrichment cultures (sequence SM00-2D-18E) resembling the five environmental sequences SM00-25D-351N, -25D-353N, and OTU7N (containing three sequences). Three sequences, all from enrichment cultures inoculated with basalt from the Kolbeinsey ridge, resembled the oligotrophic bacterium Hyphomicrobium indicum and the environmental sequence SM00-52D-119N from the Knipovich ridge (Fig. 6). 12 sequences obtained from basalt enrichment cultures (with “Fe-TSB”, “Ppi”, “Fe-red”, “PYGV”, and “W20” media) resembled the iron reducer Shewanella frigidimarina and the environmental basalt sequence SM00-52D-121N (Fig. 6). Iron reduction was, however, only observed in one of these cultures, as mention above. Manganese oxidation and reduction was not measured in the “PYGV” cultures. Within the main groups Actinobacteria and δ-proteobacteria, the sequences retrieved from enrichment cultures did not match any environmental sequences.

Bacterial sequences retrieved from sediment and seawater enrichment cultures (a total of 39 sequences) affiliated with six of the same phylogenetic groups retrieved from basalt enrichment cultures: Firmicutes, Actinobacteria, Bacteroidetes, and the α-proteobacteria, γ-proteobacteria, and ɛproteobacteria (results not shown). Only one sequence from a sediment enrichment was identical to a basalt enrichment sequence (SM99-2R-2E), grouping with Clostridium estertheticum (phylum Firmicutes, Fig. 6).

All archaeal sequences from enrichment cultures were between 92% and 98% similar to the sequences retrieved directly from the basalt, except for one sequence obtained from an enrichment culture (SM00-2D-857E), which was 99% similar to an environmental sequence (SM00-18D-837N). These sequences, although obtained from different samples, came from the same ridge and age group (0.1 Ma). All archaeal sequences retrieved from basalt enrichment cultures were obtained from methanogenic media in which methane production was detected (Fig. 7). Archaeal sequences were not retrieved from any sediment or seawater enrichment cultures.


4.1Bacterial diversity

DGGE analysis showed that the bacterial DNA banding patterns obtained for the basalt samples was complex (20–40 distinguishable bands), indicating a relatively high diversity. This was higher than for seawater samples, which had 21–28 bands. Only the clearly visual bands were excised and sequenced, which gives an indication of the most abundant part of population, but not the absolute diversity (Figs. 3,4). Several DNA bands from different profiles appeared to migrate the same distance on the gel. Sequencing, however, often showed that the sequences were related but not identical. There are two possible explanations for this: (1) closely related but non-identical sequences migrated to the same distance on the gel (microheterogeneity), or (2) the bands with the same position contained originally identical DNA fragments which have accumulated ambiguous nucleotide positions or misreads during reamplification and sequencing.

The diversity of sequences retrieved from basalts was dominated by Proteobacteria (Figs. 3,4), primarily γ-proteobacteria. γ-proteobacteria and Firmicutes were found in most samples in all three age groups. The 18 γ-proteobacterial sequences originated from 12 different samples, whereas the 10 Firmicute sequences originated from 10 different samples. Sequences affiliating with α-proteobacteria, Chloroflexi and Actinobacteria were also repeatedly found, but with a more inconsistent distribution, as these divisions were abundant in a few samples, but completely absent in others. Chloroflexi, which was only found in 0.1 Ma basalts, has previously been detected in seawater[46], non-hydrothermal subsurface basalt[17], and sediment deposits[47]. Actinobacteria was absent in the youngest basalts but showed an increased abundance with age, and appeared to be the dominant group in the oldest (1 Ma) basalt. Actinobacteria are gram-positive bacteria that are well represented by cultivated organisms, and have been reported in subsurface environments[48]. However, it is not possible to draw any distinct conclusions regarding the timescale, as these data reflect sequences obtained from only one microbial community from the oldest sample.

Most DNA sequences from environmental basalt did not match precisely any sequences in the database (Figs. 4,5). One sequence (SM00-19D-300N) did not affiliate with any known phyla, and could belong to an uncharacterized candidate division. This suggests that most of the microorganisms in Arctic ridge basalts are previously uncharacterized. Exceptions were: (i) the sequence SM00-20D-480N, which was identical to the V3 region of Propionibacterium sp. WJ6, previously identified from mine waters[49], (ii) OTU6N, which was identical to the aquatic heterotrophic bacterium Brevundimonas sp.[50], and (iii) SM00-22D-295N, which was identical to Hyphomonas oceanitis[51]. Both Brevundimonas and Hyphomonas are able to attach to solid surfaces and are typical oligotrophic microorganisms. Hyphomonas are often the first to colonize surfaces and initiate biofilm formation[52].

Some of the bacterial DNA sequences resembled those of previously detected microorganisms associated with sponges, corals, or vestimentiferans (Pogonophora) (Fig. 4). These sequences were from samples collected at different depths on all three ridges, so no correlation between depth of samples and sequences resembling macrofauna-associated microorganisms could be detected.

The sequences retrieved from basalt samples grouped into nine phylogenetic main groups; Firmicutes, Chloroflexi, Actinobacteria, Bacteroidetes, α-, γ-, δ-, and ɛproteobacteria, and Crenarchaeota (Figs. 4,5). In comparison, previous studies of non-hydrothermal seafloor[15] and subseafloor[17] basalt reported four and six phylogenetic main groups, respectively. These previously reported phylogenetic groups were among the nine found in this study, except the β-proteobacteria, which were only retrieved from the older subsurface basalt[17]. Some sequences retrieved in the current study matched (93–100%) sequences previously retrieved from similar environments. Within γ-proteobacteria, Chloroflexi, and Actinobacteria, sequences retrieved in the current study matched nine different reference strains previously obtained from a similar study of subsurface basalt from the Southeast Indian ridge (marked with ∗ in Fig. 4)[17]. The uncultured γ-proteobacterium ODP-155B-597 was 93% similar to SM00-54D-12N, whereas the γ-proteobacterium ODP-1162-327 was identical to SM00-5R-460N (Fig. 4(b)). Within γ- and ɛproteobacteria, our sequences matched two reference strains previously obtained from a diffuse vent on the Juan de Fuca ridge (marked with * in Fig. 4)[11]. This suggests that some populations might be common to low-temperature seafloor environments.

In contrast, the sequences retrieved from basalt communities differ from those obtained from seawater, as shown in Figs. 3 and 4. Seawater was dominated by γ- and δ-proteobacteria. Only one sequence affiliated with the δ-proteobacteria was retrieved from basalt, and this sequence did not resemble the sequences from seawater. Several of the sequences retrieved from seawater were affiliated to Pseudoalteromonas, which is a common organism found in marine environments. Seawater also contained Verrucomicrobia and plastids, which were not detected in basalts. Verrucomicrobia is a relatively new division of bacteria, and this group is represented by few isolates[48], however, culture independent methods have shown that this group of organisms is widely distributed in the marine and terrestrial environments. Recent studies involving a substantial molecular survey of seawater also show different microbial diversity from that found in basalt in the current study, where more than 80% of the rRNA genes were affiliated to α-, β– and γ-proteobacteria[46]. Bacterial community composition and processes in sediments also differ from the basalt communities [3,47]. In a study on microbial phylogeny in sediments, populations mainly differing from those found in this current study was reported[47]. Exceptions were sequences resembling Acinetobacter junii and Pseudomonas stutzeri, but these sequences are marked as laboratory contaminants in[47]. If that is so, the similarity has nothing to do with the environments from which the samples were retrieved. Some degree of similarity was also found between the Chloroflexi isolate ODP1176A1H3z_8_B (accession number AY191335)[47] and the sequences SM99-2R-12N (90% similar), SM00-10D-277N (90% similar), and SM00-10D-264N (88% similar). Other subsurface Chloroflexi resembling sequences retrieved in the current study includes the sequences H1.2.f (AF005747) and H1.43.f (AF005749) from a subsurface paleosol[53], which were 92% similar to the sequences SM00-10D-280N and SM00-10D-277N, respectively. Microbial communities from hydrothermal areas differ from those found in this study[54]. Sequences obtained from basalt associated hydrothermal fluids from the Juan de Fuca ridge [11,13] differ from the sequences retrieved from the cold seafloor basalts of the current study. Exceptions were two sequences from diffuse vent fluids (<50 °C)[11] that were similar to the γ-proteobacterial sequences affiliating with A. junii and of the only ɛproteobacterial sequence retrieved in this study, respectively. Also, the 65 °C hydrothermal fluid[13] differed more from the non-hydrothermal basalt of this study than the >50 °C diffuse vent fluid[11], indicating that temperature is an important factor controlling the microbial community. In low-temperature areas of the Juan de Fuca ridge, psychrophilic iron oxidizers have recently been isolated from metalliferous deposits[14]. The two identified groups of iron oxidizers were most closely related to Marinobacter and Hyphomonas. Sequences affiliating with Marinobacter and Hyphomonas were also retrieved during the current study, but our sequences could not be compared to those of[14] due to sequencing of different regions of the 16S rRNA gene.

4.2Archaeal diversity

During the amplification of archaeal DNA, visible PCR products appeared only after the second round of the nested PCR procedure, indicating a much lower abundance of Archaea compared to Bacteria. Also, the DGGE analysis showed less complex patterns, with 6–12 distinguishable bands.

All archaeal sequences obtained during this study affiliated with the marine Group 1: Crenarchaeota. None of the sequences resembled any previously characterized species. The lack of other archaeal groups beside the marine group 1 Crenarchaeota may be due to the low amount of archaeal DNA in the samples, leaving part of the archaeal diversity under the detection limit of the method. All archaeal sequences obtained from environmental samples were similar, however, the archaeal sequences were short, which will affect the percentage similarity.

The Crenarchaeota is a well defined branch of the archaeal domain, which is obvious both from sequence data as well as biochemical investigations. The Crenarchaeota was long considered to be only present at high temperatures, and all cultured representatives of the Crenarchaeota are originated from extremely hot environments, including hydrothermal vents and geothermal springs[55]. Revealing the widespread diversity of Crenarchaeota in non-extreme habitats is one of the remarkable findings of culture-independent surveys. Group 1 Crenarchaeota is the most widely distributed and abundant form of all known Archaea, as they occupy several different habitats and ecological niches. In a study of Archaea in the mesopelagic zone of the Pacific Ocean, these Crenarchaeota presented one of the oceans single most abundant cell type[56]. Thorseth et al.[15] also found the Group 1 Crenarchaeota to be the dominant Archaea of cold seafloor basalt environments at the Knipovich ridge.

4.3Physiology and microbial processes in seafloor basalt

One main objective of this study was to grow microbial strains detected in the environmental samples in order to map microbial processes involved in element fluxes and biogeochemical cycles in this seafloor basalt environment. In most cases, the microorganisms detected in environmental samples were not successfully grown with the enrichment media and conditions used in this study. Exceptions were sequences from environmental and enriched basalt, found within the (1) Firmicutes resembling Clostridium psychrophilum, which were 98–100% similar to each other; and (2) Crenarchaeota, which were 99% similar to each other, and therefore could belong to the same species. All other sequences from enrichment cultures were less than 98% similar to any environmental sequence, and thus probably belonged to different species. This shows that the enriched microorganisms comprise a minor part of the total in situ community.

Within Chloroflexi, one of the most abundant phylogenetic groups present in the environmental samples (Fig. 4), only one sequence could be obtained from enrichment cultures (Fig. 6). This sequence was 96% similar to environmental sequences.

Observations of iron and manganese oxyhydroxides on the surfaces and along fractures in the basalt samples collected, as well as observations of iron reduction in iron-reducing bacterial (IRB) cultures and retrieval of sequences resembling the iron reducer Shewanella, suggest that microbial catalyzed reduction of iron, and probably also manganese, is important in non-hydrothermal seafloor basalt environments. Sequences resembling those of the iron reducer Shewanella frigidimarina were retrieved from both environmental and enriched basalt samples (Fig. 6). Shewanella resembling sequences were only retrieved from one of the four positive IRB cultures, whereas DGGE pattern observations indicated that Shewanella was also present in two more cultures. In the fourth culture, however, no indication of Shewanella was seen either by DGGE or DNA sequencing, indicating that unknown iron-reducers were present in this culture. In addition, sequences matching Shewanella were retrieved from enrichments with the media “Ppi”, “Fe-red” (both designed for iron reducers), “PYGV” (designed for manganese oxidizers), and “W20” (designed for sulfate reducers). Indeed, Shewanella is known to reduce manganese and sulfate, as well as iron. The detection of sequences resembling Shewanella in media for manganese oxidizers, as well as the lack of observation of iron reduction in the “Ppi” and “Fe-red” media, could be due to anaerobic niches in the inoculum or to the cultures also containing metal oxidizers. Sequences resembling Shewanella were obtained from enrichment cultures inoculated with basalt sampled from lavas of different depths and ages, including the very young samples from the Knipovich ridge.

Iron oxidation is a chemically spontaneous process in the presence of oxygen at near-neutral pH, but can be biologically catalyzed under microaerophilic conditions. In a previous study of seafloor lavas from the Knipovich ridge, SEM analysis revealed stalks resembling those produced by the iron oxidizer Gallionella ferruginea in the youngest and least altered sample[15]. Also, sequences matching G. ferruginea have been retrieved from young basalt flows collected from shallow areas around Jan Mayen, indicating that Gallionella is an inhabitant of Arctic ridge basalt[16]. The presence of the iron oxidizing bacteria Gallionella and iron reducers, such as Shewanella, in young seafloor basalts indicates that microbial redox cycling of iron starts soon after the formation of the lava flows and may be important in the weathering of ocean crust basalt.

Methane production was detected in two out of seven cultures on methanogenic media inoculated with basalt, both from 0.1 Ma samples collected from deep sites on the Mohns ridge. Although the primers were designed to target all known methanogens, methanogenic Archaea were not detected in the DGGE analysis. Possible explanations for this could be that unknown methanogens with different primer homology regions were present; or that the number of methanogens was under the detection level for the applied method. Bidle et al.[57] also experienced the difficulty of retrieving sequences affiliating with methanogens using 16S rRNA gene targeting primers. With these primers they only retrieved Group 1 Crenarchaeaota, whereas using specific primers targeting enzymes involved in methanogenesis resulted in sequences belonging to methanogens. In another study, lithoautotrophic methanogens and heterotrophic IRB from subseafloor basalt collected from the northern flank of the Southeast Indian ridge were successfully enriched[17]. This indicates that both methanogens and IRB are present in non-hydrothermal ocean crust basalt, and that lithoautotrophic methanogens could be important primary producers in this environment. It has been suggested that the deep biosphere has a hydrogen and carbon dioxide dependent primary production [1,58]. The hydrogen source for methanogenesis in the seafloor basalt studied here is unknown, and could be either a result of inorganic (fluid–rock interactions) or microbial processes. Geochemically produced hydrogen is, however, only reported to be produced at high-temperature fluid–rock reactions associated with volcanic eruptions and serpentinization [59,60].

Sulfate reducing bacteria were not detected in basalt samples, either in DNA analysis of environmental samples or in enrichment studies. Neither was any sulfide minerals detected in the basalts by petrographical examination, indicating that sulfate reduction is not a dominant process. We were only able to grow a minor part of the total microbial diversity in environmental basalt, so there could be sulfate-reducing bacteria in the uncultivable fraction. Sulfate reduction was, however, common in sediment and seawater enrichments. The failure of detecting SRB in basalts thus supports the view that basalt samples were not heavily contaminated with microorganisms from seawater or sediment.

4.4Evaluation of possible contamination

As the basalt samples were transported through the seawater column, indigenous seawater microbes could easily attach to the basalt. In order to evaluate this potential contamination, surface seawater samples as well as bottom seawater samples were included in the analysis. The phylogenetic analysis showed that the microbial community in the seawater samples differed from that of the basalts. The two phylogenetic groups Verrucomicrobia and plastids derived from photosynthetic organisms in the seawater were absent in the basalt samples. Also the fraction of the δ-proteobacteria was much higher in the seawater samples. The quick handling of samples and the use of methods only analyzing the dominant part of the populations may explain the lack of extensive seawater contaminants, which the rock samples were exposed to during dredging.

As only one sediment sample was analyzed, little can be concluded about possible sediment contamination. One sequence from an enrichment culture inoculated with sediment was identical to environmental sequences retrieved from basalt. This sequence similarity between strains from basalt and sediment could, however, be due to the habitats being in close contact and forming a continuous habitat, rather than contamination. Most sequences retrieved during this study were unrelated to previously characterized bacteria from sediment and seawater, and are good candidates for indigenous seafloor basalt inhabiting bacteria.

Another source of contamination is amplification of contaminant DNA instead of, or in addition to, DNA extracted from the samples. As the basalts contain low amount of biomass and DNA, they are susceptible to amplification of contaminant DNA. Yet another possible source of contamination is the laboratory water supply system. Even though blanks were routinely used for DNA amplification during this study, some candidates that have been reported as typical laboratory contaminants were found. This applies especially to sequences affiliating with the γ-proteobacteria Acinetobacter junii (6 sequences) and Pseudomonas stutzeri (1 sequence) (Fig. 4(b)). A. junii and P. stutzeri are common constituents of laboratory water supply [61–63] and have repeatedly been found to contaminate low-biomass samples [2,61]. We can therefore not exclude the possibility that some of the γ-proteobacterial sequences might be derived from laboratory contaminants.


The aim of this study was to describe the microbial diversity in seafloor basalts and to identify possible microbial groups unique to these environments. By culturing approaches, we wanted to gain an improved understanding of the physiology of the organisms present in deep sea basalts and also determine if there is a correlation between the diversity of microorganisms retrieved from these basalt samples and geochemical characteristics of the environments. The most frequently retrieved sequences from the basalt communities affiliated with the Chloroflexi, Firmicutes, Actinobacteria, and the γ- and α-proteobacteria, suggesting that members of these phylogenetic groups dominate this habitat. Our results show that distinct microorganisms, different from those observed in deep seawater, inhabit seafloor basalt. Some of the phylotypes retrieved during this study were closely related to marine bacteria, especially within the γ-proteobacteria and Marine Group 1: Crenarchaeota. Other sequences showed less than 90% sequence similarity to any previously retrieved microorganisms or DNA sequences, especially within the Chloroflexi. The original source habitat for the basalt microbial community is probably seawater, as the bacteria could not have originated from the extremely hot molten lava. However, none of the sequences retrieved from seawater during this study were more than 89% similar to sequences from basalt communities. Also, the phylogenetic and physiological diversity of the non-hydrothermal seafloor basalt microorganisms differ from those previously found at hydrothermal regions [11–13,54], seawater[46], and sediments [3,47]. Detection of DNA sequences similar to those of previous microbial surveys of seafloor and subseafloor basalt indicates that non-hydrothermal ocean crust in general could be inhabited by related microbial populations. Parts of the cultivable fraction of microorganisms participate in the reduction of iron and in the production of methane. Evidence of oxidation and reduction of iron in very young basalts indicate that biologically catalyzed cycling of iron initiates in the oceanic crust relatively quickly after its formation. Reduction of sulfate was not detected in basalt cultures, but was common in enrichments inoculated with sediment and bottom and surface seawater. The majority of bacterial sequences retrieved from basalt samples and enrichment cultures showed no close relation to cultured relatives. Further work to explore these unique microorganisms is thus urgently needed.


This work was supported by the Norwegian Research Council (NFR) through the “SUBMAR” program (128418/431). We thank Svein Norland for providing the software used for DGGE gel analysis. We also thank two anonymous reviewers for helpful and constructive comments on the manuscript.