Abstract

This paper demonstrates the first microbiological sampling of the Outokumpu deep borehole (2516 m deep) aiming at characterizing the bacterial community composition and diversity of sulphate-reducing bacteria (SRB) in Finnish crystalline bedrock aquifers. Sampling was performed using a 1500-m-long pressure-tight tube that provided 15 subsamples, each corresponding to a 100-m section down the borehole. Microbial density measurements, as well as community fingerprinting with 16S rRNA gene-based denaturing gradient gel electrophoresis, demonstrated that microbial communities in the borehole water varied as a function of sampling depth. In the upper part of the borehole, bacteria affiliated to the family Comamonadaceae dominated the bacterial community. Further down the borehole, bacteria affiliated to the class Firmicutes became more prominent and, according to 16S rRNA gene clone libraries, dominated the bacterial community at 1400–1500 m. In addition, the largest number of bacterial classes was observed at 1400–1500 m. The dsrB genes detected in the upper part of the borehole were more similar to the dsrB genes of cultured SRBs, such as the genus Desulfotomaculum, whereas in the deeper parts of the borehole, the dsrB genes were more closely related to the uncultured bacteria that have been detected earlier in deep earth crust aquifers.

Introduction

It has been estimated that as much as 50% of the planet's total biomass is present on the oceanic floor, as well as in the fluid-filled pores and fractures of continental sedimentary and crystalline rocks (Whitman et al., 1998; Onstott et al., 1999). These habitats include a wide range of microorganisms capable of living under diverse, often extreme environmental conditions (Whitman et al., 1998; Horsfield et al., 2007). Currently, the upper temperature limit of deep subsurface microorganisms is considered to be ∼121 °C (Kashefi & Lovley, 2003). Microorganisms have been reported to be active at very high pressures of up to 80 MPa (Kato et al., 1998) and even at 1.6 GPa (Sharma et al., 2002), which would already correspond to pressures prevailing in the upper mantle at depths of about 60 km. With increasing depth, temperature is the major limiting factor of microbial life in continental crust, where geothermal gradients of 10–30 K km−1 typically prevail (Pollack et al., 1993). On this account, life should be possible to depths of several kilometres in the continental crust, if other environmental conditions, such as the availability of electron acceptors and donors, prevailing hydrogeochemical conditions (fluid and gas composition and concentration) and the geological properties of the bedrock are favourable (Hallbeck & Pedersen, 2008). The terrestrial deep subsurface biosphere has been studied worldwide, and is known to consist of diverse bacterial and archaeal microbial communities, in addition to viruses (Onstott et al., 1999; Pedersen, 2000; Takai et al., 2001; Lin et al., 2005; Zhang et al., 2005; Edwards et al., 2006; Fredrickson & Balkwill, 2006; Gihring et al., 2006; Le Romancer, 2007; Chivian et al., 2008; Kyle et al., 2008; Sahl et al., 2008).

Life in the deep subsurface may be at least partially dependent on the flux of organic carbon and energy from the surface, and would therefore ultimately be driven by photosynthetic processes. In addition, sedimentary rocks rich in organic matter and petroleum with their metamorphic equivalents, such as graphite and sulphide-bearing schists, also contain small amounts of organic compounds probably feeding deep ecosystems (Price & DeWitt, 2001; Schwab et al., 2005; Fredrickson & Balkwill, 2006). On the other hand, there is accumulating evidence that microbial life deep in the crystalline rocks could derive its energy from autotrophic processes independent of photosynthesis (Stevens, 1997; Stevens & McKinley, 2000; Chapelle et al., 2002). It has been suggested that the deep microbial ecosystems in rocks with no available organic carbon could utilize hydrogen as the main energy source (Stevens, 1997; Pedersen, 1999, 2000; Chapelle et al., 2002; Chivian et al., 2008). Hydrogen can be derived from a number of abiotic geological processes including the radiolytic decomposition of water (Hoffman, 1992; Lin et al., 2005; Pratt et al., 2006; Lefticariu et al., 2010), water–rock interaction at low temperatures (Coveney et al., 1987; Stevens & McKinley, 2000), volcanic activity (Pedersen, 2000) or simply by diffusion to the substrate from deeper levels (Pedersen, 1997).

Fluids deep in the crystalline bedrock are often saline [total dissolved solids (TDS), values in the range of 10–250 g L−1] and anoxic with negative Eh potentials (Fritz & Frape, 1987). Deep saline waters are also known to contain relatively high concentrations of dissolved methane, and according to stable isotope studies, both abiotic and biogenic methane exist in deep saline waters (Sherwood Lollar et al., 1993a, b, 2006, 2008; Stotler et al., 2010). However, the interactions between deep microbial communities and deep gases are not well understood. The isotope compositions of hydrogen, oxygen, strontium and chlorine indicate that they have very long residence times in bedrock and are not involved in meteoric near-surface water circulation (Frape et al., 1984; Nurmi et al., 1988; Smalley & Blomqvist, 1988; Lippmann et al., 2003; Sherwood Lollar et al., 2006, 2008; Stotler et al., 2010). Consequently, deep saline water aquifers can be considered potential environments in which indigenous, self-sustaining microbial communities may be found.

The present study focuses on deep subsurface microbial life in a deep research borehole, the Outokumpu Deep Drill Hole, in eastern Finland, located at latitude 62.71740°N, longitude 29.06528°E. The hole was drilled by the Outokumpu Deep Drilling Project (2004–2010) of the Geological Survey of Finland into a Palaeoproterozoic sequence of rocks consisting of metasediments, ophiolite-derived altered ultramafic rocks and pegmatitic granite (Kukkonen, 2009). The borehole attains a final depth of 2516 m, and because it is cased only to a depth of 39 m below surface, it provides a unique access to deep levels in the crystalline bedrock. One of the main aims of the Outokumpu Deep Drilling Project is to improve the understanding of the composition and origin of saline fluids and gases in the crystalline rocks of Outokumpu, as well as to study the deep biosphere. The hydrogeological studies of the Outokumpu drill hole are summarized by (Ahonen et al. 2010, 2011).

During summer 2007, the first microbiological sampling was performed at the Outokumpu deep borehole using a tube sampling technique, which has been applied earlier for sampling fluid and gas in hydrogeochemical studies of crystalline bedrock (Nurmi & Kukkonen, 1986; Nurmi et al., 1988; Smalley & Blomqvist, 1988; Blomqvist, 1999). The present sampling extended to 1500 m depth. Our purpose was to evaluate the feasibility of the tube sampling technique for microbiological sampling, and to characterize the microbial communities as a function of depth and changing geochemical conditions along the borehole.

Materials and methods

Site description and drilling the deep borehole

The Outokumpu deep hole was drilled into the Palaeoproterozoic (c. 1.95–1.91 Ga) Outokumpu formation, eastern Finland (Säntti et al., 2006; Peltonen et al., 2008). The formation hosts significant polymetallic Cu–Co–Zn–Ni–Ag–Au sulphide deposits that were mined from 1911 to 1988. Before deformation and metamorphism, the formation was originally a sequence of sea bottom sediments and oceanic lithosphere (the Outokumpu ophiolite). Organic material originating from ancient deposits nowadays forms graphite- and sulphide-rich layers (black schist) in the metasedimentary rocks, as well as in the contact zones of the altered ophiolite-derived ultramafic rocks.

The deep research borehole in Outokumpu was drilled during 2004–2005 in order to study deep structures and seismic reflectors within the Outokumpu formation (Kukkonen et al., 2006; Kukkonen, 2009). Rotary drilling with steel-tooth drilling bits was used for continuous coring. The rock types encountered, as well as the measured and interpreted water-conducting fractures are shown in Fig. 1. Drilling fluid required to cool the drill bit and to remove cutting material was taken from the municipal water line and labelled with sodium fluorescein (500 mg m−3). The water is derived from a shallow Quaternary aquifer. Drilling fluid was circulated and replenished daily in order to replace losses of water in the bedrock fractures. The borehole deviates from vertical by a maximum of only 9° and extends downward to a depth of 2516 m from the surface. The diameter of the borehole is 22 cm, except the uppermost 39 m, where a steel casing (inner diameter 32.5 cm) was installed due to the presence of a 33 m thick layer of unconsolidated Quaternary sediment on top of the crystalline bedrock. The lower end of the casing and the uppermost fractures of the bedrock were cemented in order to isolate water of the deep borehole from the near-surface groundwater. The lack of a casing deeper in the hole allows direct in situ access to the bedrock, thereby facilitating the sampling of fluid and microorganisms.

1

In situ temperature log, EC of the Outokumpu borehole water determined in tube sampling (September 2007) and the geophysical down-holes logs (April 2006 and September 2006) (the first three columns on the left). In the centre, the ‘permeable zones’ represent potential hydraulically permeable fracture zones determined on the basis of the geophysical logging data (see text). ‘Hydraulic tests’ (second column from the right) indicate the test sections of one-packer drill-stem test intervals and the hydraulic permeability values obtained. The borehole lithology (first column from the right) is shown with blue and gray shades to indicate metasediments, green and orange shades to indicate ophiolite-derived rocks and a pink shade to indicate pegmatitic granite.

1

In situ temperature log, EC of the Outokumpu borehole water determined in tube sampling (September 2007) and the geophysical down-holes logs (April 2006 and September 2006) (the first three columns on the left). In the centre, the ‘permeable zones’ represent potential hydraulically permeable fracture zones determined on the basis of the geophysical logging data (see text). ‘Hydraulic tests’ (second column from the right) indicate the test sections of one-packer drill-stem test intervals and the hydraulic permeability values obtained. The borehole lithology (first column from the right) is shown with blue and gray shades to indicate metasediments, green and orange shades to indicate ophiolite-derived rocks and a pink shade to indicate pegmatitic granite.

At the end of drilling (February 2005), the borehole was flushed with fresh municipal tap water. Tap water of the municipal water line is fresh, electrical conductivity (EC) is about 0.1 mS cm−1, with Ca, Mg, HCO3 and SO4 as the main components, and TDS of about 0.1 g L−1. After drilling, the EC (and temperature) in situ was logged three times, once in 2005, only 5 days after the end of drilling and flushing by the geophysical logging team of the NEDRA, Russia, and twice in 2006 by the operational support group of International Scientific Continental Drilling Program (ICDP). The results of these logs are shown in Fig. 1. In addition to the EC logs, the data set was supported by other geophysical logs (Tarvainen, 2006) and the geological log of the core (Västi, 2005) (Fig. 1).

Tube sampling

The tube sampling method (Nurmi & Kukkonen, 1986) allows the retrieval of a continuous water column from borehole water, and it has been applied in hydrogeochemical fluid and gas sampling in deep slim holes drilled for ore exploration (Nurmi et al., 1988; Smalley & Blomqvist, 1988). The polyamide tube consisted of 100-m-long sections, having outer and inner diameters of 12 and 9 mm, respectively. Each 100-m section yields about 6.4 L of fluid. Tubes were connected to each other by ball valves and a back-pressure valve was fitted at the lower end of a 1500-m long string of tubing, which was slowly lowered into the water-filled borehole. To avoid atmospheric contamination, the tube was filled with argon gas before being inserted into the borehole. The tube automatically filled with borehole fluid and, while being drawn back to the surface, the back-pressure valve retained the fluid in the tube. During the extraction of the tube, the ball valves were closed when they emerged at the surface level, thereby providing 100-m-long pressurized samples of the borehole fluid. The tubing was factory clean. In addition, the valves and connections, as well as the tools used to connect the sections of tubing, were sterilized separately. During sampling, all the working phases were performed using the best practices to avoid contamination from soil.

Sampling and analyses of gases and geochemistry

Groundwater samples for gas and geochemistry analysis were taken at 100 m intervals down to a depth of 1500 m. Separate samples were taken for gas analysis by evacuating the gas phase through a needle in the head space flasks, which were prefilled with borehole water. Although this method does not allow the quantitative determination of the gas volume of the total fluid volume, the composition of the released gas phase can be determined. The gas composition of the samples was analysed by Ramboll Analytics, Finland. The gas sample was taken from the gas phase in septum containers by means of a gas-tight glass syringe and injected into the gas chromatograph using vacuum (Agilent HP6890, equipped with TCD and FID detectors). The gases analysed were Ar, CO2, He, H2, N2, O2, methane, ethane, ethene, ethyne, propane and propene. EC and pH were determined in the field laboratory directly on the groundwater samples retrieved from the tubes using a field analyser (WTW GmbH, Germany), and also repeated later in the laboratory (Labtium Oy, Finland) in connection with the other chemical analyses.

Following the gas sampling, water samples were taken for different analyses. From each sampling section, two water subsamples were taken for chemical analyses: a 100 mL sample was filtered (Millipore, 0.45 μm) and acidified (ultrapure HNO3) to determine the main cations and trace components by ICP-MS/AES (Thermo Jarrell Ash Corp.), an ICP-AES dual detector system, IRIS Advantage, method based on international standards ISO 17294-2 and ISO 11885. The sample (250 mL) was analysed using ion chromatography (Dionex DX 120) to determine anions by international standard ISO 10304-1. Alkalinity was determined using titrimetry (Mettler Toledo DL 70) and phosphate was determined using the spectrophotometric method (Shimadzu UV-150-02) using ammonium molybdate complexation (ISO 6878). Chemical determinations were conducted by Labtium Oy. The amount of TDS was calculated as the sum of all analysed components.

Sampling for microbiological analyses

Groundwater samples for microbial analysis were also taken from every 100 m interval down to 1500 m depth. Because of the high pressure inside the tubes, the groundwater samples for DNA analysis (volumes 100 and 500 mL) were filtered directly from the sampling tubes through the Sterivex filters (Sterivex GP 0.22 μm, Millipore, Espoo, Finland) connected by a piece of plastic tube to the sampling tube. The Sterivex filters were immediately transferred into the Falcon tubes (Corning, Mexico) and frozen in dry ice. The samples were maintained at −80 °C until analysed. Water samples for microscopy were taken and placed into sterile 100-mL head-space flasks that had been sterilized and made anaerobic by nitrogen gasification and closed with 20-mm butyl rubber caps (Bellco Glass Inc.) and aluminium crimp seals. The samples were stored at +4 °C until the microbial densities were measured after a couple of days.

Microbial analysis

The number of microorganisms in the water samples was determined by staining with a BacLight Bacterial Viability Kit (Molecular Probes). Five millilitres of sample was stained according to the manufacturer's protocol and filtered through polycarbonate Isopore Membrane filters (0.2 μm GTBP, Millipore, Ireland) with a Millipore 1225 Sampling Manifold (Millipore) using low vacuum. The filters were examined using an epifluorescence microscope (Olympus BX60, Olympus Optical Ltd., Tokyo, Japan) under UV light using × 100 magnification. Cells were counted in 20 randomly located fields and recorded using a microscope- and computer-connected camera. The number of cells in the sample was then calculated on the basis of the magnification factor, filtered volume and the surface area of the filter. Because 4′,6-diamidino-2-phenylindole staining could not be used due to precipitation problems possibly due to the high salt concentration of the borehole water, the sum of dead and viable cells was used as an estimate of the microbial cell density.

Total DNA isolation was performed directly from the frozen Sterivex filter units. The filter units were opened aseptically using sterile, side-cutting pliers and the filters were cut into pieces with a sterile surgical blade. Total DNA was extracted from the filter pieces using the PowerSoil DNA Isolation kit (MoBio Laboratories Inc., Solana Beach, CA) in accordance with the manufacturer's protocol, except that all the reaction mixes were collected in each step in order to avoid DNA loss. Negative DNA extraction controls were performed by extracting DNA using the same protocol from a clean Sterivex filter.

The bacterial community composition down the borehole was studied by denaturing gradient gel electrophoresis (DGGE). A 193-bp fragment covering the V3 variable region of 16S rRNA gene was PCR amplified with the P2 and P3 primers described by Muyzer et al. (1993). The amplicons were separated on 8% acrylamide and 20–65% denaturing gradient at 60 V and 60 °C for 18 h, and visualized with SYBR Green I staining. The banding profiles were analysed using the bionumerics software version 4.6 (Applied Maths). The similarities between banding profiles were calculated using Dice's coefficient of similarity and a dendrogram was constructed using the unweighted pair-group method with an arithmetic averageclustering algorithm with 0.5% optimization. The significance of the clusters was determined by means of the cluster cut-off method. The most prominent bands were excised from the DGGE gel, reamplified using the P2 and P3 primers, and then purified using the Qiaquick PCR Purification kit (Qiagen, Germany). Sequencing was performed from both ends of the amplicon using a BigDye® Terminator v3.1 Cycle Sequencing kit (Applied Biosystems, California) in an ABI Prism 310 Genetic Analyzer (Applied Biosystems). The sequences were corrected manually using chromas software (version 2.13, Technelysium).

Bacterial 16S rRNA gene clone libraries were prepared from three depths: 1–100, 900–1000 and 1400–1500 m. Amplification of the 16S rRNA gene was performed with primers U968f-GC-f and U1401-r, which amplify a 473-bp-long fragment containing the V6, V7 and V8 variable sites of the 16s rRNA gene (Nübel et al., 1996). The PCR products were purified using the Qiaquick PCR Purification kit (Qiagen) and cloned to chemically competent DH5αEscherichia coli cells using the TOPO TA cloning kit (Invitrogen, Carlsbad, CA) according to the manufacturer's protocol. Ninety-six clones were randomly selected for sequencing with the kit-supplied M13f primer (Invitrogen). The clones were sequenced using the Sanger sequencing method. The sequences were manually proofread using the chromas sequence analysis program (version 2.13, Technelysium).

Sulphate-reducing bacterial (SRB) diversity was investigated by PCR-DGGE analysis of the sulphite reductase gene (dsrB) fragment. A 470-bp fragment of the dsrB gene was PCR amplified using the primers dsr4R and 2060F+GC, as described by Foti et al. (2007). Amplification was carried out in mixtures containing 1 × Dynazyme II buffer (10 mM Tris-HCl, pH 8.8, 1.5 mM MgCl2, 50 mM KCl and 1% Triton-X-100), 0.2 mM of each dNTP, 50 pmol of each primer and 1 U Dynazyme II DNA polymerase (Finnzymes, Espoo, Finland) with a 5-min initial denaturation at 94 °C, followed by 35 cycles of 1 min at 94 °C, 1 min at 55 °C and 1 min at 72 °C and a final elongation at 72 °C for 20 min. The PCR products were resolved by DGGE on 8% acrylamide and 40–70% denaturing gradient at 85 V and 60 °C for 20 h. The bands were visualized with SYBR Green I staining. Cluster analysis of the banding profiles, band sequencing and sequence analysis were performed as described above.

Phylogenetic analysis of the 16S rRNA gene and dsrB gene fragments excised from DGGE gels was performed using the basic local alignment search tool (blast) algorithm and Nucleotide Collection database available at the National Center for Biotechnology Information using default settings (Altschul et al., 1997).

The 16S rRNA gene sequences obtained from the bacterial 16S gene clone libraries were analysed with arb software package version 03_08_22 (Ludwig et al., 2004) using the SILVA 96 release of the SSU Ref database. The sequences obtained from the 16S rRNA gene clone libraries were automatically aligned using the fast aligner tool of arb, and the alignments were then checked and corrected manually. The aligned 16S rRNA gene fragments were first added to the SSU Ref guide tree using parsimony. Based on the similarities observed in the resulting tree, more group-specific trees were then calculated using full-length sequences that showed the highest similarity to the studied 16S gene fragments in the SILVA database. All group-specific trees were calculated using three different methods: maximum-likelihood analysis, maximum-parsimony analysis and neighbour-joining analysis. The maximum-likelihood analysis was performed using the fastdnaml algorithm and a 50% frequency filter, the maximum-parsimony analysis using the phylip dnapars algorithm and a 50% frequency filter, and the neighbour-joining trees using Jukes–Cantor correction and a 50% frequency filter. Gene sequences obtained from the clone libraries were added to the trees by parsimony using a 50% frequency filter. The significance of the results was estimated manually by comparing the tree topologies obtained by the three methods.

Results

Hydrogeology and hydrochemistry of the deep borehole

The first in situ EC-logging in February 2005 (7 days after the end of drilling and flushing with fresh water) shows that the conductivity had already increased to 2–4 mS cm−1, which implies up to a 20-fold increase in the TDS values. The spikes of the conductivity log of February 2005 show that inflow of more saline fluids was taking place from numerous fractures along the borehole (Fig. 1). Furthermore, logging carried out twice in 2006 shows that the situation had changed in the sense that saline formation fluid discharge into the hole was dominated only by a few fracture zones (Fig. 1). The most important fracture systems are located at depths of about 967, 1400, 1460, 1600, 1720 and 2250 m and 2300 m. In the upper part of the hole, fluids from a fracture zone at about 967 m with moderate hydraulic permeability seem to dominate the EC profile. Thus, within a year after drilling, the contribution of the fresh flushing water had disappeared and EC had reached 15–19 mS cm−1 (corresponding to moderate salinity with TDS 9–11 g L−1) at 200–1500 m depth. Microbiological sampling was performed down to a depth of 1500 m, where the ambient temperature is about 20 °C.

At the time of microbial sampling in 2007, the water in the uppermost 1500 m of the borehole was saline, with an EC of about 14–19 mS cm−1. The EC values measured in the tube water samples correspond to the overall pattern of the in situ measurements, but all details present in the down-hole logs cannot be observed in tube samples due to smoothing of the conductivity variations within each 100 m sampling interval (Fig. 1). The pH values measured in the field were about 8–9, but the later laboratory determinations indicated distinctly lower pH values down to pH 6 at 1500 m.

The chemical data are presented in Table 1. Calcium, sodium and chloride were the predominant ions, and the variation in their concentrations correlated with the EC profile of the water. TDS varied from 7.7 g L−1 at a depth of 100 m to 11.5 g L−1 at 1400–1500 m. In situ EC measurements (Fig. 1) indicate a small, but persistent, local maximum in salinity over the depth range of about 800–1000 m, which clearly reflects the chemistry of the water discharging from the fracture zone at a depth of 967 m. This zone was detected in the hydrogeological packer test performed during drilling over the depth range from 957 to 997 m. Comparison with the results of sampling at a depth of 957–997 m during drilling indicates that the borehole water sampled in 2007 was slightly less saline than the water discharging from the fracture, which is possibly caused by mixing and dilution with deeper, less saline water. The similarity in water chemistry over the depth range of about 400–1000 m supports the conclusion that fracture water discharging from a depth of 967 m is flowing upwards in the borehole. There is probably also flow downwards, but here, the fluid seems to be mixing with fluids originating from the fracture systems within and in the vicinity of the ophiolite-derived rocks.

1

Chemical composition of the Outokumpu deep borehole water

Depth (m) Ca (mg L-1K (mg L-1Mg (mg L-1Na (mg L-1Si (mg L-1Sr (mg L-1S (mg L-1Br (mg L-1Cl (mg L-1Alkalinity (mmol L-1
100 1530 29.7 18.4 1320 13.6 50.8 31.3 4550 0.45 
200 1650 28.5 16.9 1370 0.94 14.4 52.3 33.7 5060 0.35 
300 1860 26 17.6 1520 0.86 15.9 52.5 39.4 5530 0.26 
400 2070 24.2 15.8 1650 0.94 17.4 50.4 41.9 6200 0.2 
500 2140 23.6 15.4 1690 1.05 17.9 52.6 42.7 6600 0.18 
600 2200 23.1 15.1 1730 1.22 18.2 51.8 44.8 6700 0.22 
700 2260 22.3 15.4 1770 1.77 18.7 51.8 44.7 6720 0.23 
800 2300 22 15.8 1790 2.15 18.9 50.4 46.1 6870 0.22 
900 2300 22 16.5 1810 2.43 19 53.3 44.5 6810 0.23 
1000 2220 23 17 1760 2.66 18.2 53.6 42.8 6550 0.24 
1100 1910 30.2 34.3 1570 2.35 15.8 52.1 38.2 5680 0.25 
1200 1730 39.7 57.1 1370 1.68 14.3 47 33.1 5220 0.26 
1300 1650 49.9 93 1150 1.02 13.2 49.7 <10 4830 0.18 
1400 1990 64 152 1070 0.47 15.3 54.1 15.2 6180 0.18 
1500 2630 63.8 90.9 1330 0.49 19.1 53 15.3 7040 0.21 
Depth (m) Ca (mg L-1K (mg L-1Mg (mg L-1Na (mg L-1Si (mg L-1Sr (mg L-1S (mg L-1Br (mg L-1Cl (mg L-1Alkalinity (mmol L-1
100 1530 29.7 18.4 1320 13.6 50.8 31.3 4550 0.45 
200 1650 28.5 16.9 1370 0.94 14.4 52.3 33.7 5060 0.35 
300 1860 26 17.6 1520 0.86 15.9 52.5 39.4 5530 0.26 
400 2070 24.2 15.8 1650 0.94 17.4 50.4 41.9 6200 0.2 
500 2140 23.6 15.4 1690 1.05 17.9 52.6 42.7 6600 0.18 
600 2200 23.1 15.1 1730 1.22 18.2 51.8 44.8 6700 0.22 
700 2260 22.3 15.4 1770 1.77 18.7 51.8 44.7 6720 0.23 
800 2300 22 15.8 1790 2.15 18.9 50.4 46.1 6870 0.22 
900 2300 22 16.5 1810 2.43 19 53.3 44.5 6810 0.23 
1000 2220 23 17 1760 2.66 18.2 53.6 42.8 6550 0.24 
1100 1910 30.2 34.3 1570 2.35 15.8 52.1 38.2 5680 0.25 
1200 1730 39.7 57.1 1370 1.68 14.3 47 33.1 5220 0.26 
1300 1650 49.9 93 1150 1.02 13.2 49.7 <10 4830 0.18 
1400 1990 64 152 1070 0.47 15.3 54.1 15.2 6180 0.18 
1500 2630 63.8 90.9 1330 0.49 19.1 53 15.3 7040 0.21 

The Outokumpu ophiolite rock sequence at depths of 1300–1500 m (Fig. 1) contains abundant Mg-bearing minerals (Mg serpentine, Mg, Fe carbonates). Similarly, the magnesium concentration in the borehole water over the 1100–1500 m depth range is distinctly higher than that of water sampled higher up in the borehole (about 150 vs. 15 mg L−1, Table 1), which indicates that water–rock interactions are affecting the water chemistry. Bromide was the second most common anion after chloride in the saline water, and exceeded the bicarbonate concentration (35–45 mg L−1, i.e. about 0.4 mM bromide vs. about 0.25 mM HCO3). The Br/Cl ratio was about 0.0065 throughout the upper part of the borehole, but decreased to about 0.002 over the depth range from 1300 to 1500 m. This also indicates that different types of processes affect the fluid salinity within the Outokumpu formation. The sulphate concentration was below the quantification limit (10 mg L−1 due to the high salinity) in all samples. However, the total dissolved sulphur concentration of the filtered, acidified water samples determined by ICP-MS was about 50 mg L−1 in all the samples down to 1500 m in the borehole (Table 1).

The saline water in the Outokumpu deep borehole contains relatively high amounts of dissolved gases (Table 2). The total gas volume in the water samples taken during drilling (at a depth of 957–997 m) was 900 mL L−1 (NTP). Methane and nitrogen were the main components (61 and 32 vol%, respectively). Gas phase analyses of the tube water samples taken in 2007 show similarity to the sample taken during drilling, but the tube water samples contained substantial amounts of oxygen. The presence of oxygen, being an atmospheric gas, indicates the possibility of contamination from various sources: (1) air diffusing into the open borehole, (2) gas residues in the drilling and flushing fluids or (3) contamination of the gas samples during transfer from the tube to the septum vials and subsequently to the gas chromatograph. Diffusion down the borehole would result in a decrease in oxygen concentrations with increasing depth. Oxygen derived from the drilling fluid and/or sample treatment in the field is therefore clearly the main source of air contamination.

2

Gas composition at different depths in the deep borehole

Depth (m) N2 O2 CO2 He Ar H2 CH4 C2H6 C3H8 C2H4 C3H6 Total 
100 33 3.9 <0.003 1.5 0.74 0.024 58 0.52 0.01 <0.001 <0.001 97.7 
200 56 12 0.21 1.3 0.62 0.007 22 0.22 0.004 <0.001 <0.001 92.2 
400 66 17 0.41 0.5 0.79 0.005 13 0.13 0.003 <0.001 <0.001 97.4 
600 45 9.8 <0.003 0.35 0.68 <0.003 41 0.47 0.01 <0.001 <0.001 97.3 
957–997 32.4 1.7 0.043 1.25 0.26 0.01 61 0.85 0.021 0.0001 0.0001 97.5 
800 38 6.9 <0.003 1.1 <0.003 50 0.51 0.011 <0.001 <0.001 97.5 
1000 40 4.5 0.019 3.1 0.3 <0.003 49 0.37 0.01 <0.001 <0.001 97.3 
1200 52 10 0.034 0.61 0.6 0.003 33 0.29 0.008 <0.001 <0.001 96.5 
1400 38 6.1 <0.003 3.7 0.7 0.02 48 0.44 0.015 0.001 <0.001 97 
Depth (m) N2 O2 CO2 He Ar H2 CH4 C2H6 C3H8 C2H4 C3H6 Total 
100 33 3.9 <0.003 1.5 0.74 0.024 58 0.52 0.01 <0.001 <0.001 97.7 
200 56 12 0.21 1.3 0.62 0.007 22 0.22 0.004 <0.001 <0.001 92.2 
400 66 17 0.41 0.5 0.79 0.005 13 0.13 0.003 <0.001 <0.001 97.4 
600 45 9.8 <0.003 0.35 0.68 <0.003 41 0.47 0.01 <0.001 <0.001 97.3 
957–997 32.4 1.7 0.043 1.25 0.26 0.01 61 0.85 0.021 0.0001 0.0001 97.5 
800 38 6.9 <0.003 1.1 <0.003 50 0.51 0.011 <0.001 <0.001 97.5 
1000 40 4.5 0.019 3.1 0.3 <0.003 49 0.37 0.01 <0.001 <0.001 97.3 
1200 52 10 0.034 0.61 0.6 0.003 33 0.29 0.008 <0.001 <0.001 96.5 
1400 38 6.1 <0.003 3.7 0.7 0.02 48 0.44 0.015 0.001 <0.001 97 

Sample depth 957–997 refers to sampling during drilling in 2004; the other samples are taken from the tube water profile in 2007. Depth scale indicates the lower end of each 100 tube section.

The main components of the gas phase were methane and nitrogen, with average concentrations of 39% and 46%, respectively. However, the air contamination evidently increases the proportion of nitrogen and decreases the proportion of methane in these samples. Most of the samples contain detectable amounts (>0.003%) of hydrogen. Hydrogen in the sample from the uppermost part of the borehole may be derived from corrosion of the iron casing, as observed earlier during the Palmottu natural analogue study (Cera et al., 2002). On the other hand, the hydrogen concentration (0.003 vol%) at a depth of about 1400 m, where the ophiolite sequence occurs, may indicate water–rock interactions as the source of hydrogen. Small concentrations of ethane and propane were detected in all the samples.

Microbial density

The number of microorganisms according to the epifluorescence microscopic determinations was 105 cells mL−1 at the depth range of 100–1100 m (Fig. 2). Below 1300 m, the number of cells started to decrease, and was 104 cells mL−1 at a depth of 1500 m.

2

Microbial cell density in the water samples taken from the Outokumpu deep borehole.

2

Microbial cell density in the water samples taken from the Outokumpu deep borehole.

Bacterial diversity in the borehole

Although most of the 16 sRNA DGGE bands were present at more than one sampling depth, cluster analysis showed that the banding profiles were most similar at depths of 100–300, 300–1300 and 1300–1500 m, with 95%, 92% and 95% similarity, respectively (Fig. 3). Banding profiles of microbial communities at 0–100 m were also more similar to those present at 100–300 m (88% similarity) than those further down in the borehole (84% similarity). The similarity of the 16S rRNA gene sequences separated on the DGGE gel to the sequences previously isolated or uncultured, but classified bacteria is shown in Table 3. The two 16S rRNA gene fragments that represented the most dominant bands in the DGGE gels (bands 1 and 2 in Fig. 3) at all sampling depths were closely related to bacteria in the Hydrogenophaga and Fusibacter genera. The other sequenced 16S rRNA gene fragments were detected in samples taken at depths >300 m. The 16S rRNA gene fragments (bands 5 and 6 in Fig. 3) corresponding to bacteria in the Desulfotomaculum and Erysipelothrix genera became more predominant in samples taken at depths >300 m, and the 16S rRNA gene fragments (bands 3 and 4 in Fig. 3) similar to uncultured Clostridiaceae bacterium in samples from deeper than 600 m. The 16S rRNA gene fragment (band 4 in Fig. 3) similar to Candidatus Desulforudis audaxviator was detected only in samples taken over the 1200–1500 m depth range.

3

DGGE profiles of the bacterial communities at different depths in the deep borehole as determined by 16S rRNA gene DGGE. The similarity of the banding profiles was calculated by means of Dice's coefficient of similarity using 0.5% optimization. Clustering was performed using the unweighted pair-group method with an arithmetic average clustering algorithm. Clusters determined as significant using the cluster cut-off method are indicated with black lines. The percentage values show the similarity between the clusters. Cut and sequenced DGGE bands are indicated by numbers. The phylogenetic affiliation of the sequences is presented in Table 3.

3

DGGE profiles of the bacterial communities at different depths in the deep borehole as determined by 16S rRNA gene DGGE. The similarity of the banding profiles was calculated by means of Dice's coefficient of similarity using 0.5% optimization. Clustering was performed using the unweighted pair-group method with an arithmetic average clustering algorithm. Clusters determined as significant using the cluster cut-off method are indicated with black lines. The percentage values show the similarity between the clusters. Cut and sequenced DGGE bands are indicated by numbers. The phylogenetic affiliation of the sequences is presented in Table 3.

3

Phylogenetic affiliation of the 16S rRNA gene fragments PCR amplified from the Outokumpu deep borehole water samples and separated on DGGE gel

Band Class Order Family Genus Strain Accession no. Coverage (bp) Similarity (%) E 
Betaproteobacteria Burkholeriales Comamonadaceae Hydrogenophaga Hydrogenophaga sp. Pd1 AM980998 161 99 10−74 
Clostridia Clostridiales Clostridiales Family XII. Incertae sedis Fusibacter Fusibacter sp. VNs02 FJ168472 154 96 10−64 
Clostridia Clostridiales Clostridiaceae  Uncultured Clostridiaceae bacterium DQ191816 142 94 10−94 
Clostridia Clostridiales Peptococcaceae  Candidatus Desulforudis audaxviator CP000860 154 91 10−54 
Clostridia Clostridiales Peptococcaceae Desulfotomaculum Desulfotomaculum sp. 175 AF295656 145 94 10−56 
Mollicutes Anaeroplasmatales Erysipelotrichidae Erysipelotrix Erysipelotrix sp. Oita548 EF494749 151 90 10−51 
Band Class Order Family Genus Strain Accession no. Coverage (bp) Similarity (%) E 
Betaproteobacteria Burkholeriales Comamonadaceae Hydrogenophaga Hydrogenophaga sp. Pd1 AM980998 161 99 10−74 
Clostridia Clostridiales Clostridiales Family XII. Incertae sedis Fusibacter Fusibacter sp. VNs02 FJ168472 154 96 10−64 
Clostridia Clostridiales Clostridiaceae  Uncultured Clostridiaceae bacterium DQ191816 142 94 10−94 
Clostridia Clostridiales Peptococcaceae  Candidatus Desulforudis audaxviator CP000860 154 91 10−54 
Clostridia Clostridiales Peptococcaceae Desulfotomaculum Desulfotomaculum sp. 175 AF295656 145 94 10−56 
Mollicutes Anaeroplasmatales Erysipelotrichidae Erysipelotrix Erysipelotrix sp. Oita548 EF494749 151 90 10−51 

Only bacterial isolates or classified uncultured bacteria are presented. Hits to uncultured and unclassified bacteria are shown in Table S1.

In general, the 16S rRNA gene sequences obtained from the deeper samples shared lower similarity to bacteria classified in the databases than the bands present in the samples taken closer to the surface. Especially in the case of gene fragments that were most effectively amplified from the water samples taken below 300 m, the sequence similarity to previously classified bacteria was 94% or less. According to the blast search, these sequences were the most similar to the 16S rRNA genes of uncultured bacteria detected in water samples taken from deep surface environments and alkaline or other aquatic environments, such as hot springs and drinking water (Supporting information, Table S1). However, due to the small number of bands sequenced from DGGE profiles, it was possible to evaluate only the major species.

Clone libraries

Ninety-six clones were sequenced from each 16S rRNA gene library prepared from the water samples taken at depths of 0–100, 900–1000 or 1400–1500 m. Based on phylogenetic analysis of the PCR amplified 16S rRNA gene fragments (470 bp), 60% of the 83 good-quality sequences produced from the 0–100 m sample were affiliated to Betaproteobacteria (Fig. 4a). These clones were most closely related to the Hydrogenophaga (7 clones, 8.4%) and Curvibacter (4 clones, 4.8%) genera, as well as to uncultured subsurface and groundwater bacteria in the Comamonadaceae family (39 clones, 47%) (Fig. S1a). 38.6% of the clones were affiliated to class Clostridia, including members of the families of Thermoanaerobacterales (2.4%), Clostridia Family XI (3.6%) and XII (9.6%), Clostridiaceae (7.2%), Ruminococcaceae (3.6%) and Peptococcaceae (12%). One clone (1.2%) was related to the class of Alphaproteobacteria.

4

Relative distribution of bacterial orders in the 16S rRNA gene clone libraries prepared from the deep borehole water samples taken at depths of (a) 0–100 m, (b) 900–1000 m and (c) 1400–1500 m. The phylogenetic distribution of the cloned 16S rRNA gene fragments was determined using the SILVA SSU Ref database and arb software package.

4

Relative distribution of bacterial orders in the 16S rRNA gene clone libraries prepared from the deep borehole water samples taken at depths of (a) 0–100 m, (b) 900–1000 m and (c) 1400–1500 m. The phylogenetic distribution of the cloned 16S rRNA gene fragments was determined using the SILVA SSU Ref database and arb software package.

On the other hand, at a depth of 900–1000 m, 51 (71.8%) of the 71 good-quality sequences were most closely related to members of Clostridia (Fig. 4b). The proportion of Betaproteobacteria fell to 14.1%, with only four clones (5.6%) related to genus Hydrogenophaga and six clones (8.5%) to uncultured bacteria detected in subsurface water and alkaline groundwater (Fig. S1a). Among Clostridia, the clones were affiliated to several families (Fig. S1b). In Clostridia Family XII, 25 clones (35.2%) were most similar to Fusibacter genus and 16S rRNA gene sequences obtained from deep coal seam groundwater, a gold mine and oil reservoirs. In Clostridiaceae, four clones (5.6%) were most closely related to the Geosporobacter genus and uncultured bacteria in aquifer soil, subsurface water, alkaline groundwater, and nine clones (12.7%) were most closely related to Clostridium species and uncultured bacteria in animal faecal contents, anaerobic bioreactors and sediments. Seven sequences (9.9%) were also affiliated to the Peptococcaceae family, and six (8.5%) to Clostridia Family XI. Among the Peptococcaceae, the clones were grouped close to Desulfitibacter, Syntrophobotulus and Desulfosporinus species as well as uncultured bacteria detected in alkaline groundwater, deep subsurface water and sub-permafrost and mine fracture waters, while the clones in Clostridia Family XI were most similar to Dethiosulfatibacter genus and uncultured bacteria in deep subsurface groundwater and bioreactors. Five clones (7.0%) were also affiliated to Erysipelotrichi and close to uncultured bacteria detected in aquifer soil and sludge (Fig. S1c). Other bacterial families included Acholeplasmataceae in the class of Mollicutes (1.4%), Rhodobacteraceae (1.4%) and Hyphomonadaceae (2.8%) in the class of Alphaproteobacteria and Flexibacteraceae (1.4%) in the class of Sphingobacteria (Fig. 4b).

The highest number of bacterial classes occurred at the 1400–1500 m depth range (Fig. 4c). At this depth, 76.5% of the 81 16S rRNA gene sequences were affiliated to class Clostridia (Fig. S1b). In Clostridia Family XI, 11 (13.6%) of the cloned 16S rRNA gene fragments were grouped with members of the Desulfatibacter genus, as well as uncultured bacteria from deep subsurface groundwater. In Clostridiaceae family, 10 clones (12.3%) were most similar to Clostridia, Geosporobacter and Oxobacter species and uncultured bacteria from bioreactors and a mine. In Clostridia Family XII, 18 clones (22.2%) were grouped with Fusibacter species, as well as with several uncultured bacteria from oil reservoirs and deep groundwater. In the family of Peptococcaceae, on the other hand, a total of 19 cloned 16S rRNA gene fragments (23.5%) were most similar to Desulfosporinus, Desulfotomaculum and Syntrophobotulus species, as well as to uncultured bacteria from deep subsurface groundwater, alkaline groundwater and soil. Three clones (3.7%) were also affiliated to the Natranaerobiaceae family, but were most closely related to uncultured bacteria detected in the deep biosphere and one clone (1.2%) within the Ruminococcacea family close to the Acidaminobacter genus (Fig. S1c). The proportion of betaproteobacterial sequences in the clone library was 9.9%, with two clones (2.5%) closely related to Curvibacter species, two (2.5%) to Hydrogenophaga species and four (4.9%) to uncultured bacteria in subsurface water and alkaline groundwater (Fig. S1a). No clones related to Sphingobacteriales or Rhodospirales were detected. In contrast, a total of four clones (4.9%) were affiliated to class Actinobacteria in the families of Microbacteriaceae, Cellulomonadaceae, Acidimicrobiaceae and Coriobacteriaceae (Fig. S1d). Two clones (2.5%) were closely related to uncultured bacteria in aquifer soil and phototrophic sludge in the class of Erysipelotrichi (Fig. S1c). In the class of Deltaproteobacteria, one clone (1.2%) was closely related to Desulfovibrio species and uncultured bacteria in deep subsurface groundwater, and two clones (2.5%) to Desulfuromonas species and uncultured bacteria in the Lupin gold mine fracture at a depth of 1130 m. One clone (1.2%) was also affiliated to class Chlamydiae close to Simkaniaceae and Parachlamydiaceae.

SRB

Sulphate reducers were detected at all depths in the borehole (Fig. 5). Similar to the 16S rRNA gene DGGE profiles, the dsrB profiles also clustered according to the depth profile. The first cluster contained samples obtained from the two uppermost depth ranges, 0–100 and 100–200 m, sharing 86% similarity. The second cluster included samples ranging from depths of 200 to 700 m (89% similarity), and the third cluster samples from 700 to 900 m and 1000 to 1200 m (88% similarity). The sample from 900 to 1000 m clustered in the fourth cluster with the samples from 1200 to 1500 m, and all shared 100% similarity.

5

Community composition of SRB along the depth profile in the deep borehole as determined by DGGE analysis of dsrB gene fragments PCR amplified from the borehole water samples. Clustering analysis was performed with Dice's coefficient of similarity and the unweighted pair-group method with an arithmetic average clustering algorithm using 0.5% optimization. Significant clusters (black lines) were determined using the cluster cut-off method. The scale bar shows the percentage similarity between clusters. The phylogenetic affiliation of the numbered dsrB gene fragments (1–9) is shown in Table 4.

5

Community composition of SRB along the depth profile in the deep borehole as determined by DGGE analysis of dsrB gene fragments PCR amplified from the borehole water samples. Clustering analysis was performed with Dice's coefficient of similarity and the unweighted pair-group method with an arithmetic average clustering algorithm using 0.5% optimization. Significant clusters (black lines) were determined using the cluster cut-off method. The scale bar shows the percentage similarity between clusters. The phylogenetic affiliation of the numbered dsrB gene fragments (1–9) is shown in Table 4.

All the dsrB gene fragments cut from the DGGE gel were affiliated to two phyla: Firmicutes and Proteobacteria (Table 4). Although PCR-DGGE is not a quantitative method, an increasing trend in the number of species related to for example Desulfotomaculum sp. and Desulfosporosinus sp. at deeper depths was observed. In addition, the similarity of the sequences to the previously identified species was, in most cases, <88%. However, higher scores were produced for dsrB gene fragments obtained from deep subsurface and aquatic environments such as sediment and groundwater (Table 2).

4

Phylogenetic affiliation of dsrB gene fragments obtained from the PCR-DGGE analysis of the Outokumpu deep borehole water samples

Band Class Order Family Genus Species Accession no. Coverage (bp) Similarity (%) E 
Deltaproteobacteria Desulfovibrionales Desulfomicrobiaceae Desulfobacterium Desulfobacterium macestii AB061533 356 88 10−124 
Deltaproteobacteria Desulfovibrionales Desulfovibrionaceae Desulfovibrio Desulfovibrio aerotolerans AY749039 264 74 10−39 
Deltaproteobacteria Desulfovibrionales Desulfovibrionaceae Desulfovibrio Desulfovibrio vulgaris ssp. vulgaris DP4 CP000527 340 100 10−173 
Deltaproteobacteria Desulfobacterales Desulfobacteraceae Desulfatibacillum Desulfatibacillum alkenivorans AK-01 CP001322 342 79 10−78 
Clostridia Clostridiales Peptococcaceae Desulfotomaculum Desulfotomaculum putei AF273032 281 77 10−54 
Deltaproteobacteria Desulfovibrionales Desulfomicrobiaceae Desulfomicrobium Desulfomicrobium sp. ADR28 AM493693 355 95 10−159 
Clostridia Clostridiales Peptococcaceae Desulfotomaculum Desulfotomaculum thermocisternum AF074396 330 73 10−49 
Clostridia Clostridiales Peptococcaceae Desulfosporinus Desulfosporinus sp. JG32A AY787791 321 78 10−69 
Clostridia Clostridiales Peptococcaceae Desulfosporinus Desulfosporinus sp. JG32A AY787791 331 82 10−92 
Band Class Order Family Genus Species Accession no. Coverage (bp) Similarity (%) E 
Deltaproteobacteria Desulfovibrionales Desulfomicrobiaceae Desulfobacterium Desulfobacterium macestii AB061533 356 88 10−124 
Deltaproteobacteria Desulfovibrionales Desulfovibrionaceae Desulfovibrio Desulfovibrio aerotolerans AY749039 264 74 10−39 
Deltaproteobacteria Desulfovibrionales Desulfovibrionaceae Desulfovibrio Desulfovibrio vulgaris ssp. vulgaris DP4 CP000527 340 100 10−173 
Deltaproteobacteria Desulfobacterales Desulfobacteraceae Desulfatibacillum Desulfatibacillum alkenivorans AK-01 CP001322 342 79 10−78 
Clostridia Clostridiales Peptococcaceae Desulfotomaculum Desulfotomaculum putei AF273032 281 77 10−54 
Deltaproteobacteria Desulfovibrionales Desulfomicrobiaceae Desulfomicrobium Desulfomicrobium sp. ADR28 AM493693 355 95 10−159 
Clostridia Clostridiales Peptococcaceae Desulfotomaculum Desulfotomaculum thermocisternum AF074396 330 73 10−49 
Clostridia Clostridiales Peptococcaceae Desulfosporinus Desulfosporinus sp. JG32A AY787791 321 78 10−69 
Clostridia Clostridiales Peptococcaceae Desulfosporinus Desulfosporinus sp. JG32A AY787791 331 82 10−92 

Hits to uncultured and unclassified bacteria are shown in Table S2.

Discussion

This study presents the results of the first microbiological tube sampling and characterization of microbial communities down to a depth of 1500 m in the 2516-m-deep Outokumpu borehole.

The tube sampling technique has been used earlier to study the geochemistry and gas composition of the deep saline methane-rich waters of the Fennoscandian Shield (Nurmi et al., 1988). The advantage of this sampling technique is that samples can be obtained from every 100 m down the borehole water column, thereby providing information about the microbial communities as a function of depth, temperature, pressure, gas content and geochemistry. The major challenges in tube sampling are related to mixing of water during sampling and possible contamination. In order to avoid contamination during sampling, a factory-clean new plastic tube and sterilized fittings were used. In addition, all the sampling phases were performed using clean gloves and plastic covers to avoid contamination from soil under field conditions.

Survival and adaptation of surface microorganisms accidentally entering the deep subsurface is unlikely due to the high pressure that increases from 1 to 150 bars at a depth of 1500 m, and the low nutrient concentrations and high salinity, with TDS ranging from 7.7 g L−1 at a depth of 100 m to 11.5 g L−1 at 1500 m.

Another critical issue for tube sampling is how well the borehole water represents water in the bedrock fractures. This depends on the prevailing flow condition of the hole and is highly site-dependent. Based on hydrogeological studies (Ahonen et al., 2011), the salinity has increased because drilling as a result of the bedrock fractures continuously discharging saline water into the borehole (Fig. 1). This means that the geochemical conditions in the borehole will gradually approach a quasi-steady-state equilibrium condition after drilling depending on the temporal development of hydraulic pressure differences between the penetrated fracture aquifers and the water column in the hole. Blomqvist et al. (1989) studied the borehole fluid salinities in the Fennoscandian Shield by sampling several locations, including the Outokumpu area, and found that the total salinity of deep waters typically increases with increasing depth, while the sulphate concentration is often low or decreases on moving downwards, indicating sulphate reduction. In addition, the fracture zone at a depth of 967 m in the Outokumpu hole is contributing methane-rich water that is gradually progressing up and down in the water column (Table 2). However, the salinity variations, local changes in geochemistry and the changes in microbial communities at several depths (Fig. 3) and the correlation of these with the locations of fracture zones and the layer of ophiolite-derived rock types (Fig. 1) confirm that the borehole water is not completely mixing, but is forming clearly discernible zones containing specific microbial communities and geochemistry. The numbers of microorganisms decreased as a function of depth, and were in agreement with previously reported cell counts in deep fracture waters of the Fennoscandian Shield (Pedersen, 2001; Hallbeck & Pedersen, 2008). The microbiological data also demonstrate the success of the tube sampling technique as a tool for an overall survey of the microbial life forms in an uncased borehole.

Characterization of the bacterial communities down the borehole with 16S rRNA gene and dsrB gene-targeting PCR-DGGE analysis demonstrated that the microbial communities varied as a function of depth (Fig. 3). The DGGE profiles of the microbial communities were clustered into three groups: depths of 100–300, 300–1300 and 1300–1500 m. Also, the 16S rRNA gene clone library analysis of three specific depths (0–100, 900–1000 and 1400–1500 m) demonstrated that Betaproteobacteria, including the family Comamonadaceae for example Hydrogenophaga, Curvibacter and uncultured groundwater microorganisms, dominated in the near surface waters. Hydrogenophaga sp. are aerobic chemolithotrophic microorganisms capable of fixing CO2 under aerobic conditions and utilizing H2 as an energy source (Yoon et al., 2008). Serpentinites in the ophiolite-derived rocks at 1300–1500 m depth may be a source of hydrogen (Coveney et al., 1987; Stevens & McKinley, 2000). Elevated amounts of hydrogen were detected in the upper part of the borehole, but there, it may be related to corrosion of the iron casing, which also produces hydrogen (Cera et al., 2002). At depths of 900–1000 and 1400–1500 m, the proportion of bacteria affiliated to the family Comamonadaceae decreased and bacteria most closely related to more strictly anaerobic Firmicutes became more dominant, Clostridia being the most dominant genus. The dominance of Clostridia increased with increasing depth, indicating their major role in deep subsurface environment, the genus Fusibacter being one of the major ones.

The highest number of classified bacteria (Clostridia, Fusibacter, Peptococcaceae Natranaerobiaceae) occurred at 1400–1500 m depth (Fig. 4c), which correlates with the ophiolite-derived altered rock types (mainly serpentinite and diopside-tremolite rock). At this depth, elevated hydrogen and Mg2+ concentrations were also observed (Table 1). The reason for the higher diversity of microorganisms at 1400–1500 m may be due to the mineralogical and geochemical properties of the rock and potentially due to the availability of hydrogen that can be produced from iron silicates with abiotic water–rock reactions (Coveney et al., 1987; Stevens & McKinley, 2000).

SRB were detected at all analysed depths, and clustering of groundwater samples based on dsrB-DGGE profiles was also observed similarly to 16S rRNA gene-DGGE profiles (Figs 3 and 5). Their species diversity changed according to the sampling depth. SRB with dsrB-genes most closely similar to Firmicutes and Proteobacteria were detected in all the samples from the surface down to 1500 m. The SRB species detected deeper in the borehole were mostly unknown species, while those closer to the surface were more typically cultured species. Several studies indicate the important role of SRBs in the deep biosphere. In deep aquifers in the Jura Mountains (France and Switzerland), Firmicutes and Deltaproteobacteria have been reported to be present as the major population (Basso et al., 2009). Similar results have been reported by Fry et al. (1997). In addition, Shimizu et al. (2007) reported that microorganisms in groundwater collected from a depth of 900 m in Japan were composed of 40%Firmicutes and 27%Deltaproteobacteria. The role of SRBs also seems to be important in our study of the Outokumpu deep borehole.

Major changes in bacterial diversity were found to be associated with the changes in geochemistry, geophysics and sampling depth. However, more data are needed on the rock–microbial interactions that seem to be associated with biodiversity. In this study, we found a high diversity of species in the Outokumpu deep borehole. In contrast, microbial diversity may also decrease with depth: 45% of the microbial species in deep granitic fracture waters at 1000 m in drill holes of a Colorado molybdenum mine were classified in the Ralstoniaceae (Sahl et al., 2008). This means that, under some specific environmental conditions, only a few microbial groups can successfully survive and will dominate the microbial communities. Deep saline waters in the fractures of crystalline rock are evidently a harsh environment for life. Even when adapted to high pressure and salinity, microbial communities must have an available energy source. Hydrogen, together with the geochemically common sulphur, iron and carbon compounds, can provide a physicochemically plausible environment for deep microbial populations. On the other hand, nutritional limitations and the restricted availability of energy lead to prolonged life cycles and, finally, to a decreasing diversity of the population at great depths, as demonstrated by Chivian et al. (2008).

Conclusions

To our knowledge, this is the first effort to study subsurface microorganisms in the groundwater of the low-porosity crystalline bedrock by means of the tube sampling method, which allows a continuous survey of changes in the microbial communities as a function of depth. We demonstrated that microbial diversities vary as a function of depth in the Outokumpu deep borehole, and that there are connections between the microbial communities and geochemical composition of water and the flow conditions between the borehole and bedrock. A most important result is the existence of deep subsurface microbial life in saline, low-nutrient groundwater of crystalline rock. Although there is a general decrease in cell density with depth, the diversity of species was not decreasing with depth, but rather at 1500 m, an even higher diversity was observed than at shallow depths.

Acknowledgements

Marjo Öster is acknowledged for her skilful technical analysis. Dr Gerard Muyzer from the University of Delft is thanked for his suggestions concerning sampling and preservation of the samples. English revision was done by our colleague and long-time friend John Derome, who sadly passed away only a few days after the revision. KYT, the Finnish research programme on nuclear waste management, is acknowledged for financing the GEOMOL project.

References

Ahonen
L
Kietäväinen
R
Kortelainen
N
Kukkonen
IT
Pullinen
A
Toppi
T
Bomberg
M
Itävaara
M
Nyyssönen
M
(
2011
)
Hydrogeological characteristics of the Outokumpu Deep Drill Hole
.
Outokumpu Deep Drilling Project 2004–2010. Geological Survey of Finland, Special Paper
  (
Kukkonen
IT
, ed), in press.
Ahonen
L
Kukkonen
I
Toppi
T
Nyyssönen
M
Bomberg
M
Nousiainen
A
Itävaara
M
(
2010
)
Deep life and gases in the Outokumpu deep borehole: base line information for nuclear waste disposal in crystalline rock
.
Scientific Basis for Nuclear Waste Management XXXIV. Materials Research Society Symposium Proceedings, 1265-AA09-09
  (
Smith
KL
Kroeker
S
Uberuaga
B
Whittle
KR
, eds), pp.
197
202
.
Materials Research Society
,
Pittsburgh, PA
.
Altschul
SF
Madden
LL
Schaffer
AA
Zhang
J
Zhang
Z
Miller
W
Lipman
DJ
(
1997
)
Gapped BLAST and PSI-BLAST: a new generation of protein database search programs
.
Nucleic Acids Res
 
25
:
3389
3402
.
Basso
O
Lascourreges
JF
Le Borgne
F
Le Goff
C
Magot
M
(
2009
)
Characterization by culture and molecular analysis of the microbial diversity of a deep subsurface gas storage aquifer
.
Res Microbiol
 
160
:
107
116
.
Blomqvist
R
(
1999
) Hydrogeochemistry of deep groundwaters in the central part of the Fennoscandian Shield, Geological Survey of Finland, Nuclear Waste Disposal Research. Report YST-101.
Blomqvist
R
Lahermo
P
Lahtinen
R
Halonen
S
(
1989
)
Geochemical profiles of deep groundwater in Precambrian bedrock in Finland
.
Proceedings of Exploration ‘87: Third Decennial International Conference on Geophysical and Geochemical Exploration for Minerals and Groundwater. Ontario Geological Survey Special
 , Vol.
3
(
Garland
GD
, ed.), pp.
746
757
.
Ministry of Northern Development and Mines
,
Sudbury
.
Cera
E
Ahonen
L
Rollin
C
Bruno
J
Kaija
J
Blomqvist
R
(
2002
)
Redox processes at the Palmottu uranium deposit
.
Eighth EC Natural Analogue Working Group Meeting: Proceedings of an International Workshop held in Strasbourg, France from 23 to 25 March 1999
  (
van Maravic
H
Alexander
WR
, eds), pp.
183
200
.
Office for Official Publications of the European Communities
,
Luxembourg
.
Chapelle
FH
O'Neill
K
Bradley
PM
Methe
BA
Ciufo
SA
Knobel
LL
Lovley
DR
(
2002
)
A hydrogen-based subsurface microbial community dominated by methanogens
.
Nature
 
415
:
312
315
.
Chivian
D
Brodie
EL
Alm
EJ
et al.  . (
2008
)
Environmental genomics reveals a single-species ecosystem deep within Earth
.
Science
 
322
:
275
278
.
Coveney
RM
Jr
Goebel
ED
Zeller
EJ
Dreschhoff
GAM
Angino
EE
(
1987
)
Serpentinization and the origin of hydrogen gas in Kansas
.
Am Assoc Petr Geol B
 
71
:
39
48
.
Edwards
RA
Rodriguez-Brito
B
Wegley
L
Haynes
M
Breitbart
M
Peterson
DM
Saar
MO
Alexander
S
Alexander
EC
Jr
Rohwer
F
(
2006
)
Using pyrosequencing to shed light on deep mine microbial ecology
.
BMC Genomics
 
7
:
57
.
Foti
M
Sorokin
DY
Lomans
B
Mussman
M
Zacharova
ERE
Pimennov
N
Kuehnen
JG
Muyzer
G
(
2007
)
Diversity, activity and abundance of sulphate-reducing bacteria in saline and hypersaline soda lakes
.
Appl Environ Micob
 
73
:
2093
2100
.
Frape
SK
Fritz
P
McNutt
RH
(
1984
)
Water-rock interaction and chemistry of groundwaters from the Canadian Shield
.
Geochim Cosmochim Ac
 
48
:
1617
1627
.
Fredrickson
JK
Balkwill
DL
(
2006
)
Geomicrobial processes and biodiversity in the deep terrestrial subsurface
.
Geomicrobiol J
 
23
:
345
356
.
Fritz
P
Frape
SK
(eds) (
1987
)
Saline Water and Gases in Crystalline Rocks. Geological Association of Canada Special Paper 33
 .
Fry
NK
Fredrickson
JK
Fishbain
S
Wagner
M
Stahl
DA
(
1997
)
Population structure of microbial communities associated with two deep, anaerobic, alkaline aquifers
.
Appl Environ Microb
 
63
:
1498
1504
.
Gihring
TM
Moser
DP
Lin
LH
et al.  . (
2006
)
The distribution of microbial taxa in the subsurface water of the Kalahari shield, South Africa
.
Geomicrobiol J
 
23
:
415
430
.
Hallbeck
L
Pedersen
K
(
2008
)
Characterization of microbial processes in deep aquifers of the Fennoscandian shield
.
Appl Geochem
 
23
:
1796
1819
.
Hoffman
BA
(
1992
)
Isolated reduction phenomena in red-beds: a result of porewater radiolysis?
Water–Rock Interaction: proceedings of the 7th International Symposium on Water–Rock Interaction – WRI-7
  (
Kharaka
YK
Maest
AS
, eds), pp.
503
506
.
Balkema
,
Rotterdam
.
Horsfield
B
Kieft
TL
the GeoBiosphere Group
(
2007
)
The GeoBiosphere
.
Continental Scientific Drilling, A Decade of Progress, and Challenges for the Future
  (
Harms
U
Koeberl
C
Zoback
MD
, eds), pp.
163
211
.
Springer
,
Berlin
.
Kashefi
K
Lovley
DR
(
2003
)
Extending the upper temperature limit for life
.
Science
 
301
:
934
.
Kato
C
Li
L
Nogi
Y
Nakamura
Y
Tamaoka
J
Horikoshi
K
(
1998
)
Extremely barophilic bacteria isolated from the Mariana Trench, Challenger Deep, at a depth of 11,000 metres
.
Appl Environ Microb
 
64
:
1510
1513
.
Kukkonen
IT
(ed) (
2009
) Outokumpu Deep Drilling Project, Third International Workshop, Espoo, Finland, November 12–13 2009, Programme and Abstracts. Geological Survey of Finland, Southern Finland Office, Marine Geology and Geophysics, Report Q10.2/2009/61.
Kukkonen
IT
Heikkinen
P
Ekdahl
E
Hjelt
S-E
Yliniemi
J
Jalkanen
E
FIRE Working Group
(
2006
)
Acquisition and geophysical characteristics of reflection seismic data on FIRE transects, Fennoscandian Shield
.
Finnish Reflection Experiment 2001–2005, Special Paper 43
  (
Kukkonen
IT
Lahtinen
R
, eds), pp.
13
43
.
Geological Survey of Finland
,
Espoo
.
Kyle
JE
Eydal
HSC
Ferris
FG
Pedersen
K
(
2008
)
Short communication. Viruses in granitic groundwater from 69 to 450 m depth of the Äspö hard rock laboratory, Sweden
.
ISME J
 
2
:
571
574
.
Lefticariu
L
Pratt
LM
LaVerne
JA
Schimmelmann
A
(
2010
)
Anoxic pyrite oxidation by water radiolysis products – A potential source of biosustaining energy
.
Earth Planet Sc Lett
 
292
:
57
67
.
Le Romancer
M
Gaillard
M
Gesling
C
Prieur
D
(
2007
)
Viruses in extreme environments
.
Rev Environ Sci Biotechnol
 
6
:
17
31
.
Lin
L-H
Slater
GF
Sherwood Lollar
B
Lacrampe-Couloume
G
Onstott
TC
(
2005
)
The yield and isotopic composition of radiolytic H2, a potential energy source for the deep subsurface biosphere
.
Geochim Cosmochim Acta
 
69
:
893
903
.
Lippmann
J
Stute
M
Torgersen
T
Moser
DP
Hall
JA
Lin
L
Borcik
M
Bellamy
RES
Onstott
TC
(
2003
)
Dating ultra-deep mine waters with noble gases and 36Cl, Witwatersrand Basin, South Africa
.
Geochim Cosmochim Acta
 
67
:
4597
4619
.
Ludwig
W
Strunk
O
Westram
R
et al.  . (
2004
)
ARB: a software environment for sequence data
.
Nucleic Acids Res
 
32
:
1363
1371
.
Muyzer
G
De Waal
EC
Uitterlinden
AG
(
1993
)
Profiling of complex microbial populations by denaturing gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA
.
Appl Environ Microb
 
59
:
695
700
.
Nübel
U
Engelen
B
Felske
A
Snaidr
J
Wieshuber
A
Amann
RI
Ludwig
W
Backhaus
H
(
1996
)
Sequence heterogeneities of genes encoding 16S rRNAs in Paenibacillus polymyxa detected by temperature gradient gel electrophoresis
.
J Bacteriol
 
178
:
5636
5643
.
Nurmi
PA
Kukkonen
IT
(
1986
)
A new technique for sampling water and gas from deep drill holes
.
Can J Earth Sci
 
23
:
1450
1454
.
Nurmi
PA
Kukkonen
IT
Lahermo
PW
(
1988
)
Geochemistry and origin of saline groundwaters in the Fennoscandian Shield
.
Appl Geochem
 
3
:
185
203
.
Onstott
TC
Phelps
TJ
Kieft
T
Colwell
FS
Balkwill
DL
Fredrickson
JK
Brockman
FJ
(
1999
)
A global perspective on the microbial abundance and activity in the deep subsurface
.
Enigmatic Microorganisms and Life in Extreme Environments
  (
Seckbach
J
, ed), pp.
489
501
.
Kluwer
,
Dordrecht
.
Pedersen
K
(
1997
)
Microbial life in granitic rock
.
FEMS Microbiol Rev
 
20
:
399
414
.
Pedersen
K
(
1999
)
Evidence for a hydrogen-driven, intra-terrestrial biosphere in deep granitic rock aquifers
.
Microbial Biosystems: New Frontiers. Proceedings of the 8th International Symposium on Microbial Ecology
  (
Bell
CR
Brylinsky
M
Johnson-Green
P
, eds), pp.
1
7
.
Atlantic Canada Society for Microbial Ecology
,
Halifax
.
Pedersen
K
(
2000
)
Exploration of deep intraterrestrial microbial life: current perspectives
.
FEMS Microbiol Lett
 
185
:
9
16
.
Pedersen
K
(
2001
)
The diversity and activity of microorganisms in deep igneous rock aquifers of the Fennoscandian Shield
.
Subsurface Microbiology and Biogeochemistry
  (
Fredrickson
JK
Fletcher
M
, eds), pp.
7
139
.
Wiley-Liss
,
New York
.
Peltonen
P
Kontinen
A
Huhma
H
Kuronen
U
(
2008
)
Outokumpu revisited: new mineral deposit model for the mantle peridotite-associated Cu–Co–Zn–Ni–Ag–Au sulphide deposits
.
Ore Geol Rev
 
33
:
559
617
.
Pollack
HN
Hurter
SJ
Johnson
JR
(
1993
)
Heat flow from the Earth's interior: analysis of the global data set
.
Rev Geophys
 
31
:
267
280
.
Pratt
LM
Lefticariu
L
Ripley
EM
Onstott
TC
(
2006
)
Radiolysis of water as a source of bioavailable chemical energy
.
Geochim Cosmochim Acta
 
70
:
A503
.
Price
LC
DeWitt
E
(
2001
)
Evidence and characteristics of hydrolytic disproportionation of organic matter during metasomatic processes
.
Geochim Cosmochim Acta
 
65
:
3791
3826
.
Sahl
JW
Schmidt
R
Swanner
ED
Mandernack
KW
Templeton
AS
Kieft
TL
Smith
RL
Sanford
CRL
Mitton
JB
Spear
JR
(
2008
)
Subsurface microbial diversity in deep-granitic-fracture water in Colorado
.
Appl Environ Microb
 
74
:
143
152
.
Säntti
J
Kontinen
A
Sorjonen-Ward
P
Johanson
B
Pakkanen
L
(
2006
)
Metamorphism and chromite in serpetinized and carbonate-silica-altered peridotites of the Paleoproterozoic Outokumpu-Jormua ophiolite belt, eastern Finland
.
Int Geol Rev
 
48
:
494
546
.
Schwab
V
Spangenberg
JE
Grimalt
JO
(
2005
)
Chemical and carbon isotopic evolution of hydrocarbons during prograde metamorphism from 100 °C to 550 °C: case study in the Liassic black shale formation of Central Swiss Alps
.
Geochim Cosmochim Acta
 
69
:
1825
1840
.
Sharma
A
Scott
JH
Cody
GD
Fogel
ML
Hazen
RM
Hemley
RJ
Huntress
WT
(
2002
)
Microbial activity at Gigapascal prssures
.
Science
 
295
:
1514
1516
.
Sherwood Lollar
B
Frape
SK
Fritz
P
Macko
SA
Welhan
JA
Blomqvist
R
Lahermo
PW
(
1993a
)
Evidence for bacterially generated hydrocarbon gas in Canadian Shield and Fennoscandian Shield rocks
.
Geochim Cosmochim Acta
 
57
:
5073
5085
.
Sherwood Lollar
B
Frape
SK
Weise
SM
Fritz
P
Macko
SA
Welhan
JA
(
1993b
)
Abiogenic methanogenesis in crystalline rocks
.
Geochim Cosmochim Acta
 
57
:
5087
5097
.
Sherwood Lollar
B
Lacrampe-Couloume
G
Slater
GF
Ward
J
Moser
DP
Gihring
TM
Lin
L-H
Onstott
TC
(
2006
)
Unravelling abiogenic and biogenic sources of methane in the Earth's deep subsurface
.
Chem Geol
 
226
:
328
339
.
Sherwood Lollar
B
Lacrampe-Couloume
G
Voglesonger
K
Onstott
TC
Pratt
LM
Slater
GF
(
2008
)
Isotopic signatures of CH4 and higher hydrocarbon gases from Precambrian Shield sites: a model for abiogenic polymerization of hydrocarbons
.
Geochim Cosmochim Acta
 
72
:
4778
4795
.
Shimizu
S
Akiyama
NT
Fujioka
M
Nako
M
Ishijima
Y
(
2007
)
Molecular characterization of microbial communities in deep coal seam groundwater of northern Japan
.
Geobiology
 
4
:
423
433
.
Smalley
PC
Blomqvist
R
(
1988
)
An isotopic cross section through stratified saline groundwater, Outokumpu, Finland
.
Chem Geol
 
70
:
165
.
Stevens
T
(
1997
)
Lithoautotrophy in the subsurface
.
FEMS Microbiol Rev
 
20
:
327
337
.
Stevens
TO
McKinley
JP
(
2000
)
Abiotic controls on H2 production from basalt-water reactions and implications for aquifer bioochemistry
.
Environ Sci Technol
 
34
:
826
831
.
Stotler
RL
Frape
SK
Ahonen
L
et al.  . (
2010
)
Origin and stability of a permafrost methane hydrate occurrrence in the Canadian Shield
.
Earth Planet Sci Lett
 
296
:
384
394
.
Takai
K
Moser
DP
DeFlaun
M
Onstott
TC
Fredrickson
JK
(
2001
)
Archaeal diversity in waters from deep South African Gold Mines
.
Appl Environ Microb
 
67
:
5750
5760
.
Tarvainen
A-M
(
2006
) Identification of water-bearing structures in the Outokumpu Deep Drill hole by geophysical well logging. Master's thesis (technology), Helsinki University of Technology, Department of Civil and Environmental Engineering (in Finnish with English abstract).
Västi
K
(
2005
) Outokumpu deep drill core: geological report. Geological Survey of Finland, Report M 52.5/4222/04/R2500, 24 March 2006.
Wanger
G
Southam
G
Onstott
TC
(
2006
)
Structural and chemical characterization of a natural fracture surface from 2.8 kilometres below land surface: biofilms in the deep subsurface
.
Geomicrobiol J
 
23
:
443
452
.
Whitman
B
Coleman
DC
Wiebe
WJ
(
1998
)
Prokaryotes: the unseen majority
.
P Natl Acad Sci USA
 
95
:
6578
6583
.
Yoon
KS
Tsukada
N
Sakai
Y
Ishii
M
Igarashi
Y
Nishihara
H
(
2008
)
Isolation and characterization of a new facultatively autotrophic hydrogen-oxidizing Betaproteobacterium, Hydrogenophaga sp. AH-24
.
FEMS Microbiol Lett
 
278
:
94
100
.
Zengler
K
Richnow
HN
Rossello-Mora
R
Michaelis
W
Widdel
F
(
1999
)
Methane formation from long-chain alkanes by anaerobic microorganisms
.
Nature
 
401
:
266
269
.
Zhang
G
Dong
H
Xu
Z
Zhao
D
Zhang
C
(
2005
)
Microbial diversity in ultra-high-pressure rocks and fluids from the Chinese Continental Scientific Drilling Project in China
.
Appl Environ Microb
 
71
:
3213
3227
.
Zhang
Y
Zhong
F
Xia
S
Wang
X
Li
J
(
2009
)
Autohydrogenotrophic denitrification of drinking water using polyvinyl chloride hollow fiber membrane biofilm reactor
.
J Hazard Mater
 
170
:
203
209
.

Supporting Information

Please note: Wiley-Blackwell is not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.