Introduction

A wide range of microorganisms have been isolated from samples of produced water obtained from oil reservoirs. These include fermentative organisms, methanogens, manganese and iron reducers, acetogens, sulphate reducers, aerobic organisms and nitrate reducers (for reviews see Magot et al. 2000; Birkeland 2004). Detection of an organism in a sample of produced water is, however, not evidence for this organism being indigenous to the reservoir, i.e. being a natural inhabitant of the reservoir before the onset of oil exploitation. Samples from oil reservoirs are, for practical and economical reasons, normally obtained from the well-head or even at a subsequent point in the pipeline system, after the produced fluids have been transported in pipelines up to several kilometres from the reservoir. Possible sources of contamination are thus numerous (McInerney and Sublette 1997; Magot et al. 2000). Microorganisms can be introduced to the production system through drilling, well operations and damaged tubings or casings, which could lead to formation of biofilms inside the oil-transporting line, and cause a continuous supply of contaminants to the system (Basso et al. 2005). An additional source of contamination can occur in reservoirs subject to water-flooding, where contaminants from injected sea-water might be introduced directly to the formations. In many cases an indigenous nature of a detected organism can be excluded. This is the case for some mesophilic organisms detected in samples from high temperature oil wells (Magot et al. 1992; Tardy-Jacquenod et al. 1998). On the other hand, the widespread distribution of thermophilic organisms—e.g. members of the order Thermotogales and the genera Thermoanaerobacter, Thermoanaerobacterium and Desulfotomaculum—in geographically separated high-temperature oil reservoirs, is taken as evidence for the existence of a natural microbial community in these habitats (Magot et al. 2000). Members of the genera Geotoga and Petrotoga, in particular, might be considered as indigenous since they have been isolated only from oil reservoirs so far. Still, little is known about the diversity and ecophysiology of microbial communities in oil formations. The primary energy source for the microorganisms in oil reservoirs is not known. At temperatures up to about 80°C, petroleum in subsurface reservoirs is often biologically degraded, over geological timescales (Head et al. 2003). Nevertheless, organisms known to be able to degrade hydrocarbons anaerobically, are not known to be indigenous to oil reservoirs. Lacking is also general knowledge about nutrient flow, how organisms have entered the oil reservoirs and how they may have survived in the reservoirs for tens of millions of years.

Studies of the distribution of organisms in produced water from high-temperature oil reservoirs, by construction of 16S rRNA gene libraries, is so far restricted to a few Californian sulphur-rich reservoirs (Orphan et al. 2000,2003) and a long term water-flooded reservoir in China (Li et al. 2006). The Californian production waters are dominated by Proteobacteria and Firmicutes, while methanogenic organisms dominate among the archaea. Bacteria in the Chinese production water are dominated by Proteobacteria. Apparently, mesophilic organisms constitute a large fraction of organisms detected in both the Chinese and the Californian production waters. From the results it can be calculated that a total of 85% of the bacterial clones detected in the Chinese production water samples are most closely related (>92.7% identity) to mesophilic members of the genera Pseudomonas and Serratia. A total of 78% of the bacterial clones in the Californian production water samples are closely related (>94.6% identity) to mesophilic organisms within the genera Pseudomonas, Halomonas, Acinetobacter, Sphingomonas, Brevundimonas, Methylobacterium and Acidaminococcus, while 86% of the clones in the archaeal library were most closely related (>95.6% identity) to the mesophilic Methanoplanus petrolearius.

In this work we have analysed produced water from a non water-flooded high-temperature oil reservoir in the North Sea by construction of bacterial and archaeal 16S rRNA gene libraries combined with culture-based methods. Our results revealed that the produced water harbours unique microbial communities as most of the detected organisms probably represent novel species or genera, previously not detected in any other environments. Most of the detected bacteria seem to be thermophilic or moderately thermophilic fermenters, while the archaea are dominated by thermophilic methanogens. A comparison of the optimum growth temperature of the closest cultivated relatives to organisms detected in the clone libraries, suggested that most of the detected organisms, including organisms representing some of the most dominating groups, are probably not able to grow at reservoir temperature and therefore should be regarded as non indigenous. This was confirmed by cultivation experiments. Thus, our results present evidence for the view that the microbiology in produced water might only to a small extent reflect the indigenous microbiology in the respective oil formations, even when the formations are not water-flooded.

Materials and methods

Site description and sample collection

The Troll C platform is located at 60° 53′ 10.73″ N, 03° 36′ 41.41″ E. The formations are situated 1550–1600 m below the sea floor and have a temperature of 70°C. The velocity of the water transport from the formations to the platform is approximately 1–2 m per second. Produced water contained the following elements (per liter): Na (16.4 g), Ca (1.6 g), Mg (560 mg), Ba (160 mg), Fe (14 mg), Sr (300 mg), K (420 mg) and P (<0.1% (w/v)) (Malvin Sperre, personal communication). Sulfate and ammonium were present at concentrations of <2 and 110 mg/l respectively. The formations are not water-flooded.

Sterile Pyrex bottles filled with Argon gas were used to collect samples at the upper riser on the Troll C oil platform. The samples were a mixture of water produced from several wells. Lines were flushed for more than 20 min prior to sampling. The bottles were filled completely with a mixture of oil and water, and transported to our laboratory at ambient temperature.

Clone library construction

Immediately after arrival of the samples at our laboratory, cells from 3 ml of the water phase were harvested in Eppendorf tubes by centrifugation and subsequently lysed by repeated freezing at −70°C for 30 min and thawing at 30°C. No intact cells could be observed after 3 cycles of freezing and thawing. The lysate was further used as template for amplification of 16S rRNA gene sequences by PCR. The oligonucleotide primers, 21F (5-TCCGGTTGATCCTGCC-3), modified after Reysenbach and Pace 1995, and A976R (5′-YCCGGCGTTGAMTCCAATT-3′) (Reysenbach and Pace 1995) were used for amplification of archaeal 16S rDNA sequences, while primers 27f (5′-GAGTTTGATCCTGGCTCA-3′) (Rainey et al. 1992) and 1525r (5′-GAAAGGAGGAGATCCAGC-3′ (Rainey et al. 1992) where used for amplification of bacterial 16S rRNA genes. PCR was performed with 30 cycles of amplification as previously described (8). PCR products were cloned with the TOPO TA cloning kit version H using chemical transformation according to the manufacturer’s instructions (Invitrogen).

Sequencing and phylogenetic analysis

Sequencing was performed on a ABI PRISM capillary sequencer according to the protocol of the ABI Prism BigDye Terminator kit (Perkin Elmer). Selected clones in the archaeal and bacterial 16S rRNA gene libraries were sequenced with primers 21F and 27f respectively, while clones representing a bacterial or archaeal OTU were additionally sequenced with reverse 16S rRNA gene primers 1525R or A976R, respectively. PCR products obtained by amplification of DNA extracted by use of a genomic DNA extraction Kit (Sigma) from liquid cultures were sequenced with one or more of the primers 27F, 1525R or 530F (5′-GTGCCAGCMGCCGCGG-3′) (Lane 1991). Sequences were checked for putative chimeric primary structures by use of the program Pintail (Ashelford et al. 2005) in combination with use of the BLAST service (Altschul et al. 1997), where identities between various fragments of a given sequence and sequences available in the GenBank database (Benson et al. 2005) were compared. Non-chimeric sequences were grouped into OTUs comprising sequences with more than 98% similarity. One representative of each OTU was sequenced in both directions. All sequences were compared with other sequences in the GenBank database using BLAST (Altschul et al. 1997) to identify closest relatives. Phylogenetic reconstructions, based on alignments created by CLUSTALX (Thompson et al. 1997), were produced using TreeView version 1.6.6 (Page 1996) with the Jukes-Cantor distance matrix (Jukes and Cantor 1969) and the neighbour-joining algorithms (Saitou and Nei 1987). Confidence in the tree topology was determined by using 100 bootstrapped trees (Felsenstein 1985). Calculation of identity between sequences was performed by use of the program BioEdit based on pairwise alignments subsequent to manual deletion of endgaps and positions with undetermined nucleotides (‘N’s).

Growth experiments

Enrichment and growth experiments of fermentative organisms were performed using an anaerobically prepared basal medium containing the following components (per liter distilled water): 20 g NaCl, 0.9 g MgCl2 × 6H2O, 1.4 g MgSO4 × 7H2O, 0.33 g KCl, 0.25 g NH4Cl, 0.14 g CaCl2 × 2H2O, 0.45 g KH2PO4, 1.0 ml trace element solution SL-10 (Widdel et al. 1983), and 0.5 ml resazurin (0.02%). After autoclaving in a dispenser (Lien and Beeder 1997) the medium was reduced with 4 ml 0.5 M Na2S under a nitrogen atmosphere. 10 ml vitamin solution (Balch et al. 1979) was added and pH was adjusted to 6.8 with 1 M NaOH or 1 M HCl. 20 or 40 ml of medium was dispensed into 50 or 100 ml serum bottles, respectively, which were subsequently sealed with butyl rubber stoppers. Substrates were added from separate, anaerobically prepared, stock solutions. R1-medium contained yeast extract (0.1% w/v), peptone (0.25% w/v) and maltodextrin (0.25% w/v). Cas-medium contained 0.5% (w/v) Casamino-acids. The same basal medium as described above was used for enrichment of sulphate-reducing organisms and methanogenic organisms except for the addition of 4 g NaHCO3 per liter medium as well as the use of a gas phase of H2 and CO2 (80:20) for enrichments of methanogens and a gas phase of N2 and CO2 (80:20) for the enrichment of sulphate-reducing organisms.

Enrichments were incubated at 6–70°C immediately after inoculation of produced fluids (a mixture of oil and water). Growth was detected by the use of a light microscope (Nikon, Eclipse E400). Isolation of pure cultures was performed by picking colonies grown in anaerobically prepared solidified medium as described elsewhere (8).

DNA–DNA hybridization

DNA–DNA hybridization experiments were performed at DSMZ. Genomic DNA from strain TCF52B and from Thermosipho africanus was prepared using a French pressure cell (Thermo Spectronic) and was purified by chromatography on hydroxyapatite as described (Cashion et al. 1977). DNA–DNA hybridization was carried out as described (Deley et al. 1970) under consideration of the modifications described by Huss et al. (1983) using a model Cary 100 Bio UV/VIS-spectrophotometer equipped with a Peltier-thermostated 6 × 6 multicell changer and a temperature controller with in-situ temperature probe (Varian).

Nucleotide sequence accession numbers

The gene sequences determined in this study have been deposited in GenBank under accession numbers DQ647058–DQ647185.

Results

Direct microscopy

Several morphotypes of microorganisms were observed by direct observation of the production water in a phase contrast microscope. Sheathed and unsheathed rods seemed to be dominating, but coccoid cells were also frequently seen. On one occasion a spirochaet-like cell was observed. Direct counts of microorganisms in production water samples using a Thoma chamber revealed the presence of approximately 106 cells/ml which is in agreement with samples from other oil reservoirs (Orphan et al. 2000).

16S rRNA gene libraries

One bacterial and one archaeal 16S rRNA gene library were constructed using PCR and DNA extracted directly from produced water as template. A total of 99 clones from the bacterial library were sequenced in the forward direction (700–800 bp). Among these, 11 sequences were found to be chimeric and were excluded from further analysis. The remaining 88 sequences were assigned to 25 distinct phylogenetic groups which clustered among Firmicutes (48 sequences), δ- and γ-Proteobacteria (16 sequences), Thermotogales (2 sequences), Spirochaetes (5 sequences) and Bacteroidetes (17 sequences) (Table 1). The samples of produced water appear to harbour unique microbial communities as the majority of sequences (83%) had less than 95% identity to closest matches in GenBank and probably represent novel genera or species (Figs. 1 and 2). A total of 29 operational taxonomic units (OTUs) were identified. A large fraction of the OTUs (62%) in the library were represented by only one sequence, while most (67%) of the sequences were assigned to OTUs comprising five or more sequences. One representative of each OTU was selected for almost full length sequencing (1300–1500 bp). Most organisms detected in the bacterial library cluster among fermentative organisms. The only exceptions are relatives to sulphate reducers in the genera Desulfomicrobium and Desulfovibrio as well as a relative to the aerobic Agarivorans albus. More than half of the OTUs in the bacterial library comprise sequences that cluster within the Firmicutes lineage, indicating the presence of a large fraction of highly diverse Gram-positive organisms.

Table 1 Closest relatives of clones obtained from the bacterial 16S rRNA gene library
Full size table
Fig. 1
figure 1

Phylogenetic neighbour-joining trees based on 16S rRNA gene sequences. The position of the clones (TCB) and isolates (strain) obtained from the Troll formations are indicated in bold. Bootstrap values from 100 replications are shown at nodes. Only values above 96 are indicated. Bar indicates 10 substitutions per 100 bases. (A) Organisms within the Firmicutes lineage. Thermotoga maritma was used as outgroup. (B) Organisms within Thermotogales, Proteobacteria, Spirocheates and Bacteriodetes. Sulfolobus acidocaldarius was used as outgroup

Full size image
Fig. 2
figure 2

Graphic representations of the novelty of organisms detected in the bacterial library. Bars indicate the distribution of identities to closest matches in GenBank, and are binned by 0.5% identity. The continuous line indicates the cumulative number of sequences that have closest match identities of at least the value defined by the x-axis. Primary y-axis (cumulative count) shows the cumulative number of sequences. Secondary y-axis (count) shows the number of counts in each bin in the histogram

Full size image

The archaeal library was less diverse than the bacterial library and only three OTUs were identified from 22 sequences of which none were chimeric (Table 2). One of the detected organisms is closely related to fermentative organisms within the Thermococcus genus while 21 sequences were closely related to members of the genera Methanothermococcus and Methanolobus. Thus, the archaeal community appears to be highly dominated by methanogens.

Table 2 Closest relatives of clones obtained from the archeal 16S rRNA gene library
Full size table

Culture-dependent analysis

In order to reveal the biological properties of the organisms detected in the culture-independent approach, we attempted to enrich and isolate organisms based on the physiological properties of close relatives to some of the organisms detected in the clone libraries. Most effort was put into enrichments of anaerobic fermentative bacteria because this physiological group seemed to be dominating according to the molecular analysis. Enrichments in anaerobically prepared medium, where a mixture of yeast extract, dextrin and peptone served as substrates (R1 medium), yielded growth in the temperature range of 6–70°C. Strain TC3 and strain TC2 were isolated from enrichments at 18 and 6°C respectively. Cells of strain TC2 were short, highly motile rods related to the recently characterized Arcobacter halophilus (95.8% 16S rRNA gene similarity; 801 bp), a gram negative, fermentative and facultative anaerobic ε-proteobacterium (Donachie et al. 2005). Strain TC3 exhibited 98% 16S rRNA gene (1435) similarity to Alcalibacter saccharofermentans, an alcaliphilic and mesophilic organism that ferments sugars and proteinous substrates (Garnova et al. 2004). Unlike what is reported for A. saccharofermentans, strain TC3 was able to grow at pH below 7.0. Two strains, strain TC371 and strain TC451 isolated from enrichments at 37 and 45°C respectively, had identical 16S rDNA sequences with 96% identity to the 16S rRNA gene from Anaerophaga thermohalophila (1438–1440 bp). Another strain isolated from an enrichment incubated at 45°C, strain TC452, had a 16S rRNA gene highly similar to Geotoga petraea. Geotoga spp. has only been recovered from oil reservoirs so far. Enrichments with R1 medium and with cellulose or fish waste as substrates at temperatures close to reservoir temperature (60–70°C) yielded cells with a morphology similar to members of the order Thermotogales. TCEL2, obtained from enrichments at 60°C on cellulose was related to Thermotoga maritima (98.5% rRNA gene sequence identity). Strain TCF52B was isolated from an enrichment incubated at 70°C where fish waste served as the only substrate, vitamin and carbon source. This strain had a 16S rDNA gene highly similar (99.9%) to 16S rDNA from Thermosipho africanus Ob7T isolated from a marine hydrothermal spring in Djibouti, Africa (Huber et al. 1989). Interestingly, the genomes of strain TCF52B and strain Ob7T were found to be identical as determined by DNA–DNA hybridization experiments.

In enrichments where Casamino acids served as the only substrate and carbon source (Cas-medium), growth was detected in the temperature range of 37–60°C, but not at 70°C. Isolates from enrichments at 37, 45 and 60°C possessed identical partial 16S rRNA genes. One of these strains, strain Cas60314, was further characterized and found to be a representative of the novel genus and species Thermovirga lienii (8). No growth was detected in non-reduced Cas- or R1 medium with a headspace of 0–50% air at 37–70°C. Enrichments for methanogenic and sulphate-reducing organisms were negative at 60 and 70°C.

Among the isolates, only the close relatives to T. africanus and T. maritima (i.e. strain TCEL2 and strain TCF52B) could grow at reservoir temperature (70°C). T. lienii strain Cas60314T was able to grow at 60°C while for the remaining isolates (i.e. strain TC2, strain TC3, strain TC371, strain TC451 and strain TC452), no growth was detected at 60 or 70°C. As indicated in Table 1, close relatives to strain TC451 and strain Cas60314T are among the organisms most frequently detected in the bacterial 16S rRNA gene library. Close relatives of strain TCEL2 and strain TCF52B were also detected. As indicated in Table 1, strain TCEL2 is a close relative to the organism detected by sequence TCB116x in the bacterial library. The 16S rRNA gene obtained from Strain TCF52B, had 99.4% identity to sequence TCB177x.

Discussion

The objective of this study was to investigate the microbiology of a high-temperature oil-bearing formation in the North Sea. The relatively low identity (<95%) between most of the sequences in the bacterial 16S rRNA gene library and their respective closest matches among known species, indicate that produced water from the Troll formations harbours a unique microbial community comprising several novel genera or species. This was further indicated by the isolation of the novel genus and species Thermovirga lienii as well as isolation of organisms relatively distantly related (<95% 16S rRNA gene identity) to members of the Anaerophaga and Arcobacter genera.

Comparison between the culture dependent and culture independent approach

Several of the isolated strains clustered among sequences obtained from the bacterial library. Relatives of T. lienii strain Cas60314 apparently constitute a major component of the microbial community in the produced fluids. T. lienii has a 16S rRNA gene highly similar to the OTU representing sequences TCB197x, TCB168x, TCB6y and TCB8y (Table 1) while OTUs represented by TCB169x, TCB170x and TCB152x form phylogenetically distinct branches and probably represent novel genera. Another dominating group is apparently organisms clustering among strain TC451, strain TC371 and A. thermohalophila. Three OTUs representing sequences TCB123x, TCB200x and TCB130x, have one of these organisms as closest match (Table 1). Finally, close relatives to organisms isolated from enrichments incubated at reservoir temperature (70°C), were also detected in the bacterial clone library i.e. strains closely related to T. africanus or T. maritima. Negative enrichments for sulphate-reducing organisms close to reservoir temperature (60 and 70°C) is in agreement with the absence of sequences in the libraries with high degree of similarity to 16S rRNA from thermophilic sulphate reducers. The same reasoning holds for the negative enrichment of methanogens at 70°C, but at 60°C close relatives to Methanothermococcus thermolithotrophicus could have been expected to grow.

Comparison with organisms previously detected in oil-affiliated samples

Some organisms detected in the bacterial library clustered among sulphate-reducing organisms while most organisms clustered among fermentative organisms. This, together with the detection of close relatives to methanogens in the archaeal library, is consistent with previous reports on the microbiology of produced water from oil reservoirs (Orphan et al. 2000; Bonch-Osmolovskaya et al. 2003; Grabowski et al. 2005). Organisms previously detected in oil-related samples and in this study include representatives of the genera Arcobacter, Anaerophaga, Desulfovibrio, Desulfomicrobium, Thermotoga, Thermosipho, Haloanaerobium and Spirochaeta.

T. maritima strain M12597 has previously been isolated from an oil reservoir in Western Siberia (Slobodkin et al. 1999). Other members of this genus have been detected in oil reservoir samples around the world, suggesting that they belong to a natural microbial community in oil-bearing strata (Magot et al. 2000; Birkeland 2004). On the other hand, the only member of the genus Thermosipho previously isolated from an oil reservoir sample, is T. geolei (L’Haridon et al. 2001).

Recently, spirochaetes were detected in enrichment cultures inoculated with produced water from a low-temperature biodegraded oil reservoir in Canada (Grabowski et al. 2005). In the same study, mesophilic members the genera Desulfovibrio and Desulfomicrobium were, by growth experiments, found to dominate among the sulphate-reducing organisms while a bacterial 16S rRNA gene library, constructed from total DNA extracted directly from the formation fluids, was entirely composed of sequences highly similar to a 16S rRNA gene sequence from a member of the Arcobacter genus. The OTU represented by sequence TCB129x in the bacterial library, provides evidence for the presence of spirochaetes also in field water from the Troll formations. However, Spirochaeta smaragdina, isolated from an offshore oil well in Congo (Magot et al. 1997), still represents the only spirochaete isolated from the deep subsurface.

We also detected a close relative to Desulfovibrio indonesiensis (sequence TCB124x) while OTUs represented by sequences TCB4y and TCB115x are relatively distantly related to Desulfomicrobium baculatum and possibly represent novel genera of sulphate reducers. Mesophilic members of the genera Desulfovibrio and Desulfomicrobium have frequently been detected in samples from high-temperature oil reservoirs around the world, but are not believed to be indigenous in these habitats (Magot et al. 2000).

Members of the genus Haloanaerobium, one of a few genera of anaerobic fermentative bacteria that are adopted to high-salt conditions, have previously been isolated from various brines with salinities above 20% (Birkeland 2004). Organisms which are relatively distantly related to Haloanaerobium saccharolyticum (88% 16S rRNA gene identity), were detected in the bacterial library in this study (Table 1). They are possibly representatives of a novel genus within the extremely halophilic Haloanaerobiaceae family and indicate the presence of a subsurface saline.

The only described species within the genus Anaerophaga is so far A. thermohalophila, isolated from an oil separation tank (Denger et al. 2002). A. thermohalophila produces a surfactant that efficiently stabilises hexadecane/water emulsions and is therefore interesting as regards microbially enhanced oil recovery (Denger et al. 2002) and it would in this respect also bee interesting to investigate whether strain TC451 and strain TC371 share this ability.

Among archaeal organisms, members of the genus Thermococcus (Stetter et al. 1993; L’Haridon et al. 1995; Miroshnichenko et al. 2001; Bonch-Osmolovskaya et al. 2003) and members of several methanogenic genera (Orphan et al. 2000; Bonch-Osmolovskaya et al. 2003; Grabowski et al. 2005), have frequently been detected in oil reservoir samples. The strains vp183 and vp21, isolated from a high-temperature oil-well sample in California (Orphan et al. 2000) and strain ST22, isolated from produced water from the Statfjord formations in the North-Sea (Nilsen and Torsvik 1996) are all close relatives to M. thermolitotrophicus. Close relatives of this species are apparently dominating among archaeal organisms in the Troll formations (Table 2). Detection of the OTU represented by sequence TCA2 (Table 2), provides evidence for the presence of members of the Methanolobus genus in produced water from the Troll formations. No members of this strictly methylotrophic and halophilic genus have so far been isolated from oil reservoir samples.

Are the detected organisms indigenous to the Troll formations?

Great care has to be taken when inferences concerning the in situ microbiology of the Troll formations are to be drawn. First, there are a number of sources for bias in the construction of 16S rRNA gene libraries that can lead to a distorted view of the de facto distribution of microorganisms in the produced water (von Wintzingerode et al. 1997). Secondly, organisms detected in the production water sample are not necessarily indigenous in the reservoir. Although no gas or liquid is injected into the Troll formations, thereby excluding an assumingly important source of contamination, contamination cannot be excluded (Magot 2005).

One approach to assess whether a certain organism obtained from a production-water sample is indigenous to the reservoir or not, is to compare the growth characteristics with that of the in situ physico–chemical parameters of the oil reservoir in question (Magot 2005). Although growth characteristics cannot easily be ascribed to organisms based on 16S rRNA gene analyses only, it is interesting to note that most sequences in the bacterial and archael libraries display highest 16S rRNA gene identities to mesophilic or moderately thermophilic organisms with optimum temperatures for growth well below reservoir temperature (70°C). On the other hand, organisms obtained from enrichments at 70°C, were exclusively members of the order Thermotogales. Furthermore, several isolates were obtained from enrichments incubated at lower temperature, but none of these were able to grow at 70°C. Interestingly, two of these strains, strain Cas60314T (the type strain of T. lienii) and strain TC451, related to A. thermohalophila, were highly similar to some of the most dominating groups of organisms detected in the bacterial 16S rRNA gene library.

From the observed genome identity of 100% between T. africanus strain TCF52B isolated in this study and the type strain of this species isolated from a hydrothermal system in Djibuti in Africa, it seems unlikely that strain TCF52B has survived and evolved in the Troll formations since reservoir deposition and consequently that these two Thermosipho strains have been separated for tens of millions of years. Based on the apparent massive contamination of moderately thermophilic organisms in our samples, among these a close relative to Geotoga petraea, it can not be ruled out that at least some of the thermophilic isolates still have a non-indigenous origin.

Taken together, no strong conclusions about the in situ microbiology of the investigated oil-formations can be drawn. Rather, our results suggest that the produced water is heavily contaminated with non-indigenous organisms and that extreme care should be taken before claiming an indigenous origin of any isolate, even if its physiological characteristics match the in situ oil reservoir conditions. Improved sampling procedures are highly required for a better understanding of the microbiology of oil reservoirs.