INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Natural environments for thermophilic microorganisms are widespread on Earth's surface. The most common and accessible thermal habitats are hot springs, sulfatara, and geothermally heated soils. Thermophiles have also been found in less accessible biotopes, like terrestrial and oceanic deep-subsurface environments (2, 4, 7, 28, 31, 33, 34, 39). The existence of microorganisms in the deep terrestrial subsurface has been noted for a long time, but only in the past decade has there been an increasing interest toward exploring subterranean microbial life in deep-surface environments (27).

The subsurface environment can be divided into nonvolcanic areas that are cold and volcanic areas that are hot. Sometimes the thermal energy is visible on the surface as hot springs and sulfatara fields, and sometimes it is visible in the form of volcanic eruptions. The temperature at nonvolcanic areas increases by 25°C for each kilometer of depth; therefore, thermophilic microorganisms should thrive in the deep subsurface. A number of thermophiles and hyperthermophiles within the domains of Bacteria and Archaea have been isolated from continental and deep-sea oil reservoirs (12, 21, 28) and from other deep wells where the temperature does not exceed 113°C (27). In volcanic areas, however, heat from the mantle is transported to the surface by conduction (heat flow), by volcanic eruptions, and by water in geothermal systems. Such energy has been harnessed as geothermal energy and is used in Iceland for such applications as heating and electricity production. The temperature of the source water is generally 50 to 130°C in wells that are typically drilled to a depth of 1,500 to 2,000 m. The bottoms of the boreholes could therefore be regarded as subterranean hot springs, the boreholes could be regarded as windows into the springs, and pipelines could be regarded as extensions, allowing for access to the subterranean springs. These environments have been minimally investigated (15, 23, 33), while terrestrial surface hot springs are one of the most extensively studied natural environments for thermophiles (3, 4, 17).

Here, we report the diversity of microbes associated within the geothermal fluid of boreholes that reach about 2,000 m below the surface, with reservoir temperatures being below 100°C. Culture-dependent and culture-independent molecular phylogenetic surveys showed that some of the community members are shared with terrestrial thermal springs but that some of the diversity appears to be unique to this subsurface biosphere. We present here the existence of an indigenous thermophilic subterranean community in the subsurface in Iceland, and we suggest that nonendemic thermophiles may be disseminated in the subsurface.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Study site and sample collection. Samples were collected from five geothermal wells (Table 1) located in four separate geothermal fields in the Reykjavík region (Fig. 1) and from a reservoir whose geothermal fluid was a mixture of geothermal water transported from 50 geothermal wells (hot tap water). The fields are located in the same Quaternary rock section within the Kjalarnes caldera. All the fields appear to have a local heat source, different chemical properties, and water recharge (20). The production characteristics of the fields are similar except for the fields with wells SN-4 and SN-5, which have higher salinity in their geothermal fluids and fluctuation in temperature.

View larger version (23K): [in this window] [in a new window]   FIG. 1.   Shown are the locations of the three geothermal fields. From west to east are Seltjarnernes (boreholes SN-4 and SN-5), Laugarnes (borehole RV-5), Ellidaar (borehole RG-39) in the Reykjavík area and the fourth geothermal field in the east, and Reykir (borehole MG-18). The inset shows the location of the Reykjavík area in Iceland.

Microscopy. Samples and cultures were viewed with a Leica DM LB light microscope equipped with a phase-contrast oil immersion objective (magnification, ×100). A Petroff-Hausser chamber (depth, 0.02 mm) was used for counting cells. In situ hybridization was performed as described by Harmsen et al. (13). Fixed cells were hybridized with the ARCH915 and EUB338 fluorescein-labeled probes. Cells were stained with 0.01% (wt/vol) acridine orange (14).

Enrichments and isolation. Concentrated water samples from borehole RG-39 and hot tap water were enriched in different media for chemolithotrophic and chemorganotrophic organisms under anaerobic conditions. Samples from other boreholes were not tested. The enrichments were incubated at 65 and 85°C. Enrichments were performed in standard medium for thermophilic microorganisms (9, 13, 25) and in modified media with different pHs. The following modified medium was used: medium prepared with hot tap water and supplemented with 0.4% (wt/vol) yeast extract, 0.1% (wt/vol) peptone, 10 ml of vitamin solution per liter, 10 ml of element trace solution (1) per liter, and 5 g of S⁰ liter¹. The pH was adjusted to 9.0, and N₂ or H₂-CO² (80%/20%) was used as the gas phase in enrichments. Strain isolation was performed with serial dilutions in Hungate culture tubes.

DNA extraction and PCR amplification. DNA was extracted from the biomass obtained by filtration of the geothermal fluid, from enrichments, and from the isolated strains (24). DNA extractions and 16S rRNA gene (rDNA) PCR amplifications were performed as described before (24, 30). One microliter was used as the template in PCR amplifications with both archaeal (23 FPL and 1391R [4]) and bacterial (F9, R1544, and R805 [30]) primers.

Cloning and sequencing of 16S rDNA. The PCR products from the biomass were cloned directly by the TA cloning method by using a TOPO TA cloning kit according to the instruction of the manufacturer (Invitrogen). Plasmid DNA from single colonies was isolated and sequenced by using reverse and forward M13 vector primers and R805 on an ABI 377 DNA sequencer by using a Big Dye Terminator Cycle Sequencing Ready Reaction Kit (PE Applied Biosystems). The following specific Archaea sequencing primers were used: FPL23, 765FA, R1391, 340RA, 774RA, and R805 for sequencing more than 400 bp of the 16S rDNAs and F9, F338, F515, F1392, R357, R805, R1195, and R1544 as specific primers for the bacterial genes (30). PCR products from isolates were partially sequenced on an ABI 377 sequencer using the R805 sequencing primer.

Phylogenetic analysis. All sequences were manually aligned (400 to 500 bp) with closely related sequences obtained from the Ribosomal Database Project (22) after BLAST searches. Sequence alignments and phylogenetic analysis were performed with the ARB program (http:/www.mikro.biologie.tu-muenchen.de), with regions of sequence ambiguity being omitted. The phylogenetic trees based on three algorithms (neighbor joining, maximum parsimony, and maximum likelihood) were constructed using 300 to 400 bp and 800 to 1,390 bp of the 16S rDNA. Distance trees were constructed by using neighbor-joining algorithms with the Jukes and Cantor correction, and maximum-likelihood trees were constructed by the fastDNAml software included in the ARB package. Homologous nucleotide positions, based on the filter of the ARB database, were included in the alignment and used for the comparison analysis. The CHECK-CHIMERA program of the Ribosomal Database Project server was used for searches of chimera artifacts (22).

Nucleotide sequence accession numbers. All small-subunit rRNA sequences were deposited in the GenBank database under the following accession numbers: AF361206 (SUBT-2), AF361207 (SUBT-3), AF361208 (SUBT-10), AF361209 (SUBT-12), AF361210 (SUBT-6), AF3612011 (SUBT-14), AF361212 (SUBT-13), AF361213 (SUBT-9), AF361214 (SUBT-11), AF361215 (SUBT-7), AF361216 (SUBT-5), AF361217 (SUBT-1), and AF361218 (SUBT-4).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Sample collection. As listed in Table 1, 55 to 1,900 liters of geothermal fluid with different chemical characteristics was filtered from wells and hot tap water.

Microscopic observations. Both coccoid and rod-shaped cells were observed in samples from boreholes RG-39 and MG-18 and the hot tap water. The cells were thin rods of various lengths, from short rods up to long filaments. Some of the rods had spherical structures within them or on their ends, and some rods were bundled together as narrow pins. To estimate the original population densities of microorganisms in emergent reservoir fluid from borehole RG-39, we collected four sequential 20-liter samples and one additional 20-liter sample after 180 liters had been discarded. Cells were counted in the range of 4 × 10⁶ to 7 × 10⁶ cells liter¹ in each sample. Similar cell counts in each sample portion taken in well RG-39 could indicate constant numbers of cells dispersed by the well fluid. No cells were detected in the two geothermal field of Seltjarnarnes (SN-4 and SN-5) and Laugarnes (RV-5) (water above 100°C). Domain-specific 16S rRNA-targeted oligonucleotide probes were used to identify Bacteria and Archaea and revealed coccoid and rod-shaped cells belonging to both domains. Long rod-shaped cells or filaments were observed with the Archaea-specific probes in well RG-39 (data not shown).

Enrichments, isolations, and identification. Growth was not significant in any of the enrichments, except in medium prepared with reservoir water, complex nutrients, and elemental sulfur. Mainly coccoid cells and a few rods were observed in the enrichments at 65 and 85°C. In order to analyze the growing communities, the enrichments were mixed together before 16S rDNA analysis. The majority (97%) of the sequences were identified by BLAST search as belonging to Desulfurococcus mobilis (19, 39), and 3% belonged to Aquificales clone SRI-48 (30). Coccoid cells forming H₂S were isolated after seven sequential subcultures at 85°C and identified as D. mobilis (99% similarity by BLAST search) by 16S rDNA sequencing (450 bp). Straight rods that did not form spores or H₂S were isolated at 65°C. The rods were identified by 16S rDNA analysis as a new member of the Thermus (also called Deinococcus) group.

Phylogenetic analysis. Both bacterial and archaeal sequences were obtained from all biomass samples by PCR amplifications of 16S rDNAs. One hundred thirteen clones (73 bacterial and 40 archaeal) from borehole MG-18, 74 clones (57 bacterial and 17 archaeal) from borehole RG-39, and 76 clones (20 bacterial and 56 archaeal) from the hot tap water were successfully sequenced. The results obtained from bacterial clones are presented in Table 2 and in Fig. 2. The bacterial sequence libraries revealed seven phylogenetically distinct lineages: type sequences SUBT-1, -2, -3, -4, -5, -6, and -7. Two operational taxonomic units (OTUs) (SUBT-1 and SUBT-7) were found in all samples, and their closest database matches were the Aquificales (30) and Nitrospira (5) groups. SUBT-2, -3, and SUBT-5 were found only in hot tap water. These OTUs were most closely related to Desulfotomaculum reducens (35), to Deinococcus geothermalis (37), and to an unidentified actinomycete OPB41 (17). Borehole RG-39 contained two OTUs (SUBT-4 and SUBT-6) whose closest database matches were to OP11 and OP8 (17) and which were not found in the other samples. The results obtained from the archaeal clones are presented in Table 3 and in Fig. 3. The archaeal sequence libraries revealed seven phylogeneticly distinct lineages, type sequences SUBT-8, -9, -10, -11, -12, -13, and -14. All OTUs, except two (SUBT-8 and -9) belonged to the candidate division Crenarchaeota (3-5). Only one OTU (SUBT-9) was found in all samples, and one other (SUBT-11) was found in both boreholes. One OTU (SUBT-8) was found only in hot tap water, one (SUBT-12) was found in borehole RG-39, and three (SUBT-10, -13 and -14) were found in borehole MG-18. Although different numbers of bases (350 to 1,390) were used in the phylogenetic analysis, the topology in trees did not significantly change within the bacterial or archaeal OTUs.

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Phylogenetic relationships of bacterial 16S rRNA sequences as determined by maximum-likelihood analysis. Sequences from the subterranean hot springs are marked in bold with the abbreviation SUBT. Accession numbers of reference sequences are in parentheses. Sulfolobus acidocaldarius was used as the outgroup. The scale bar indicates the estimated number of base changes per nucleotide sequence position.

View larger version (20K): [in this window] [in a new window]   FIG. 3.   Phylogenetic relationships of archaeal 16S rRNA sequences as determined by maximum-likelihood analysis. Sequences from the subterranean hot springs are marked in bold with the abbreviation SUBT. Accession numbers of reference sequences are in parentheses. Thermotoga maritima was used as the outgroup. The scale bar indicates the estimated number of base changes per nucleotide sequence position.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

We have shown that two deep subsurface geothermal fields in Iceland with temperatures below 100°C harbor a diverse community of unknown thermophiles. In situ whole-cell hybridization with fluorescently labeled 16S rRNA-based oligonucleotide probes indicated that the archaeal and bacterial cells were still intact. From enrichment cultures it was confirmed that cells were thermophilic, growing at high temperatures. Direct 16S rDNA analysis showed that all sequences could be affiliated with thermophilic divisions and that almost all OTUs were new and have not yet been obtained in pure cultures. Despite the use of various media and enrichment conditions, growth was observed only in media with source borehole water. The cultures were dominated by the thermophilic archeon Desulfurococcus mobilis (19, 39) and a few bacterial sequences closely related to Aquificales clone SRI-48 (30). This may suggest that subsurface growth conditions are needed for growing the noncultivated subterranean cells detected in samples.

Our findings are in agreement with recent demonstrations of other deep subsurface environments such as hot subterranean biospheres in continental oil reservoir aquifers (30, 29, 32), groundwater from a borehole in granite rock (10, 27), seafloor diking eruptive events (8), and marine sediments (26). The occurrence of these thermophiles in the subsurface is thought to be due to their deposition with the original sediment and their survival over geological time. The origin of the subterranean hot spring thermophiles is at present unclear. Nevertheless, our result could suggest an indigenous biosphere. In this study we have sequenced many new and diverse OTUs that are phylogenetically distinct lineages. At least 11 type sequences out of 14 were distantly related or showed less than 95% similarity to OTUs that have been detected in surface hot springs. Furthermore, eight of these type sequences were distantly related (64 to 95%) to candidates of new divisions originated from Obsidian Pool, Yellowstone Park. If the OTUs were introduced to the geothermal water during drilling, we would have expected to detect more OTUs closely related to known subsurface OTUs. The boreholes were drilled in the late 1970s, which is a relatively short time to form so many divergent population groups from contaminants. Closely related OTUs with more than 98% similarity to each other were found in wells RG-39, MG-18, and hot tap water. Their dominance in wells varied with the in situ temperature, and their presence was correlated to some extent to the known temperature range of their closest relatives (Tables 2 and 3). Identical OTUs within wells could suggest a subsurface geographical interconnection between wells (Fig. 2 and 3). However, several type sequences were found to be site restricted to wells and were not detected in the other wells. The absence of similar or identical OTUs within wells could possibly be explained by their low abundance in wells caused by differences in water chemistry and temperature, and therefore they were not detected.

The Seltjarnarnes field (SN-4 and SN-5) had water temperatures that fluctuated between 95 to 117°C, which is in the range of the upper temperature limit for life (6). However, no intact cells or DNA could be detected in this field, suggesting that it is not possible to find deep subsurface life everywhere. The reason could be that the field is geographically isolated from the other fields and without a terrestrial subsurface interconnecting flow. The Seltjarnarnes field is surrounded on four sides by the sea, and a hydrological barrier has been suggested between the field (Fig. 1) and the other fields (36). Furthermore, the nearest geothermal field, RV-5, is a high-temperature field with temperatures above 130°C, which is lethal for any known life form and could therefore possibly function as a dispersal barrier for subsurface migrating microbes. High chloride concentrations in SN-4 and SN-5 could also have an inhibiting effect on migrating salt-sensitive terrestrial microbes.

We have shown that geothermal reservoirs are a novel, volcanic, high-temperature environment for thermophilic microorganisms that extends the known ecological habitats of such groups of organisms. Iceland is one of the world's largest subsurface hot spots; its crust is very permeable, and therefore there are ideal paths through interconnected subsurface conduits for microorganisms that are able to survive over geological time in subsurface environments. The extent of subsurface dissemination depends on dispersal barriers, such as seawater and temperature, or hydrological barriers. The results highlight a new indigenous microbial biosphere and indicate that thermophilic microorganisms disseminate in the subsurface within volcanic zones through subsurface conduits, as has already been observed with meteoric groundwater (24) and on the surface in cold aquifers (16).