Petrotoga mexicana sp. nov., a novel thermophilic, anaerobic and xylanolytic bacterium isolated from an oil-producing well in the Gulf of Mexico

Anaerobes have always been considered as the dominant micro-organisms of petroleum reservoirs (Magot et al., 2000). Among them, fermentative bacteria constitute an important microbial community of the oilfield environment. Thermophilic isolates are the most studied, probably because (i) most oil reservoirs occur at depths where the temperature is high and (ii) thermophiles possess thermostable enzymes of great interest in industrial processes. Stetter et al. (1993) provided evidence of the presence of Thermotoga strains, members of the order Thermotogales, in oilfields. The isolation of Thermotoga elfii, Thermotoga subterranea and Thermotoga hypogea from such ecosystems was reported shortly afterwards (Fardeau et al., 1997a; Jeanthon et al., 1995; Ravot et al., 1995a). Microbial studies on various hot oil reservoirs throughout the world have provided evidence of the importance of micro-organisms morphologically and physiologically close to other members of Thermotogales, namely Fervidobacterium, Thermosipho, Geotoga and Petrotoga (Grassia et al., 1996; Magot et al., 2000). However, no Fervidobacterium species originating from oilfield reservoirs has been characterized so far. Davey et al. (1993) first reported the presence of micro-organisms belonging to the genera Geotoga and Petrotoga in petroleum reservoirs located in Oklahoma and Texas. Three novel species were fully characterized: Geotoga petraea, Geotoga subterranea and Petrotoga miotherma. Later, Petrotoga mobilis was isolated from a North Sea oil-production well (Lien et al., 1998). Petrotoga olearia and Petrotoga sibirica were recently isolated from a continental petroleum reservoir in Western Siberia (L'Haridon et al., 2002), from which L'Haridon et al. (2001) also isolated a novel species of the genus Thermosipho. Members of the genus Petrotoga are thermophilic heterotrophs that grow in a broad range of salt concentrations (up to 10 % NaCl). During our microbial studies of oil waters originating from offshore reservoirs located in the Gulf of Mexico, we isolated a novel thermophilic strain phylogenetically related to the genus Petrotoga, which we propose to be assigned to a novel species, Petrotoga mexicana sp. nov.

Several strains were isolated from oil/water mixtures taken from production well heads (21-D) in Tabasco (Gulf of Mexico). Samples were collected in sterile plastic bottles and kept at room temperature until used. P. mobilis DSM 10674^T and P. miotherma DSM 10691^T were obtained from the DSMZ, whereas P. olearia DSM 13574^T and P. sibirica DSM 13575^T were obtained from the Institut Universitaire Européen de la Mer, Plouzané, France. Enrichment cultures were performed in 120 ml serum bottles containing 20 ml medium and inoculated with 2 ml water/oil sample. The medium contained (l⁻¹ distilled water): 1 g NH₄Cl, 0·3 g K₂HPO₄, 0·3 g KH₂PO₄, 0·1 g KCl, 0·1 g CaCl₂, 0·2 g MgSO₄.7H₂O, 0·5 g cysteine hydrochloride, 1 mg resazurin, 30 g NaCl and 10 ml trace mineral element solution (Balch et al., 1979). The pH was adjusted to 7 with 10 M KOH. The medium was boiled under a stream of O₂-free N₂ gas and cooled to room temperature. Aliquots (20 ml) were dispensed into serum bottles under a stream of N₂/CO₂ (80 : 20, v/v) and the sealed vessels were autoclaved for 45 min at 110 °C. Prior to inoculation, Na₂S.9H₂O, NaHCO₃, peptone (Difco), yeast extract and sodium thiosulfate were injected from sterile stock solutions to respective final concentrations of 0·04, 0·2, 0·5 and 0·02 % (w/v) and 20 mM. H₂ pressure in the gas phase was 2 bars. The bottles were incubated at 60 °C and three enrichment series were performed in the same culture medium before isolation. Strains were isolated by repeated use of the Hungate roll-tube technique (Hungate, 1969), with medium solidified with 0·8 % phytagel (Difco). Serial dilution in roll tubes was repeated at least twice to purify the cultures. Growth experiments were performed in duplicate, using a medium differing from the enrichment medium by the absence of peptone and hydrogen in the gas phase. In addition to the basal medium, MgSO₄.7H₂O (0·2 g l⁻¹) was replaced by MgCl₂.6H₂O (1 g l⁻¹) and the concentration of yeast extract was increased to 0·1 %.

The pH, temperature and NaCl and MgCl₂ concentration ranges for growth were determined using basal medium supplemented with 20 mM glucose. The pH of the medium in Hungate tubes was adjusted by injecting 1 M HCl, NaHCO₃ or Na₂CO₃ from 10 % sterile anaerobic stock solutions. The characterized strain was subcultured twice under the same experimental conditions before growth rates were determined. All substrates were tested at a final concentration of 20 mM except for formate (80 mM), cellulose, starch and xylan (all 10 g l⁻¹). Elemental sulfur (1 %, w/v), sulfate (20 mM), thiosulfate (20 mM), sulfite (5 mM), nitrate (10 mM) and nitrite (5 mM) were tested as electron acceptors. H₂S production was determined photometrically as colloidal CuS by using the method of Cord-Ruwisch (1985). Sensitivity to ampicillin, chloramphenicol, kanamycin, rifampicin and vancomycin was tested at 10, 25, 50 and 100 μg ml⁻¹. Controls with ethanol and DMSO (solvents used for chloramphenicol and rifampicin) were performed. Growth was monitored by measuring OD₅₈₀ and by microscopic observations. Morphological characteristics were determined with an Optiphot (Nikon) phase-contrast microscope. For electron microscopic studies, preparations were negatively stained with sodium phosphotungstate as described by Koussémon et al. (2001). Growth was monitored by measuring OD₅₈₀ using a UV/visible spectrophotometer 50 Scan (Varian). Xylanolytic activity was measured as described previously (Fardeau et al., 1997a); the assay was performed by adding dinitrosalicylic acid (Miller, 1959) and the xylose released from xylan degradation was determined from the A₅₄₀. The presence of l-alanine was determined enzymically at 340 nm as described previously (Fardeau et al., 1997a). Fermentation products were determined by HPLC using a pump (Spectra Series P100; Thermo Separation Products), an automatic sampler (Spectra Series AS100), an Aminex HPX 87H (300×7·8 mm) column (Bio-Rad), a differential refractometer detector (Spectra System RI-150) and an integrator (Azur Microsoft). An aliquot of 20 μl cell-free supernatant was injected into the column, which was maintained at 37 °C. A 2·5 mM H₂SO₄ solution was used as solvent with a flow rate of 0·6 ml min⁻¹. The effect of O₂ on growth was determined in Hungate tubes containing anaerobic basal medium supplemented with 20 mM glucose. Tubes were inoculated and 20 ml sterile air was added to the gas phase. Cultures were incubated at 55 °C under agitation (150 r.p.m.). Growth was monitored by turbidity measurements (580 nm) and by microscopic observations. Experiments were conducted in duplicate and repeated at least twice. The presence of spores was determined by microscopic observation of cultures and pasteurization tests at 80 and 90 °C for 10 and 20 min.

The G+C content of DNA was determined at the DSMZ, using HPLC as described by Mesbah et al. (1989). DNA–DNA hybridization studies were carried out by the DSMZ as described by De Ley et al. (1970), with the modifications described by Huß et al. (1983) and Escara & Hutton (1980), using a model 2600 spectrophotometer equipped with a model 2527-R thermoprogrammer and plotter (Gilford Instruments). Renaturation rates were computed with the transfer.bas program (Jahnke, 1992). The 16S rDNA of strain MET12^T was amplified by adding 1 μl cell culture to a thermocycler microtube containing 5 μl 10× Taq buffer, 0·5 μl 100 nM primers Fd1 and Rd1, 4 μl 25 mM MgCl₂.6H₂O, 0·5 μl 25 mM dNTPs, 0·5 μl Taq polymerase (5 U μl⁻¹) and 38 μl sterilized distilled water. The primers used (Fd1, 5′-AGAGTTTGATCCTGGCTCAG-3′; Rd1, 5′-AAGGAGGTGATCCAGCC-3′) were described by Winker & Woese (1991). The sample was placed in a hybrid thermal reactor thermocycler (GeneAmp PCR System 2400), denatured by heating for 3 min at 96 °C and subjected to 30 cycles of 20 s at 95 °C, 30 s at 52·9 °C and 2 min at 72 °C; this was followed by a final elongation step of 5 min at 72 °C. PCR products were cloned using the pGEM-T-easy cloning kit (Promega) according to the manufacturer's protocol. Clone libraries were screened by direct PCR amplification from a colony using the vector-specific primers SP6 (5′-ATTTAGGTGACACTATAGAA-3′) and T7 (5′-TAATACGACTCACTATAGGG-3′) and the following reaction conditions: 2 min at 96 °C, 40 cycles of 30 s at 94 °C, 1 min at 55 °C, 3 min at 72 °C, and a final extension of 10 min at 72 °C. Plasmids containing inserts of the expected length were isolated using the Wizard Plus SV Minipreps DNA purification system (Promega), according to the manufacturer's protocol. Purified plasmids were sent for sequencing to Genome Express (Grenoble, France). Sequence data were imported into the sequence editor BioEdit version 5.0.9 (Hall, 1999); base calling was examined and a contiguous consensus sequence was obtained. The full sequence was aligned using the RDP Sequence Aligner program (Maidak et al., 2001). The consensus sequence was then manually adjusted to conform to the 16S rDNA secondary structure model (Winker & Woese, 1991). A non-redundant blastn search (Altschul et al., 1997) of the full sequence through GenBank (Benson et al., 1999) identified its closest relative. Sequences used in the phylogenetic analysis were obtained from the RDP (Maidak et al., 2001) and GenBank (Benson et al., 1999). Positions of sequence and alignment ambiguity were omitted and pairwise evolutionary distances were calculated using the method of Jukes & Cantor (1969). A dendrogram was constructed using the neighbour-joining method (Saitou & Nei, 1987). Confidence in the tree topology was determined by using 100 bootstrapped trees (Felsenstein, 1985). rDNA accession numbers of reference organisms are included in Fig. 2⇓. DNA–DNA hybridizations were performed at the DSMZ as described by Koussémon et al. (2001).

Enrichment cultures were incubated at 60 °C for 1 week. Growth was regarded as positive based on optical density. Microscopic examination of the enrichment culture revealed the presence of rods with an outer, sheath-like structure (toga). Circular, transparent colonies, 2·5 mm in diameter, appeared after a 6 day incubation at 60 °C in solidified roll tubes containing phytagel. Three strains of rod-shaped bacteria were isolated that were morphologically and phylogenetically identical. Strain MET12^T was characterized further. Individual cells were 0·5–0·7 μm wide and 1·4–5·0 μm long and occurred singly, in pairs or in chains up to 30 μm long (Fig. 1⇓a). Cells were surrounded by a sheath-like structure (Fig. 1b⇓). They were motile and no spore formation was detected. Electron microscopy of cell sections exhibited a Gram-negative-type cell wall (Fig. 1c⇓). Strain MET12^T was thermophilic and grew at temperatures ranging from 25 to 65 °C, with an optimum at 55 °C. The isolate was slightly halophilic and grew in the presence of 10–200 g NaCl l⁻¹, with an optimum at 30 g l⁻¹, and 0–200 g MgCl₂.6H₂O l⁻¹ (the basal culture medium contained 30 g NaCl l⁻¹) with an optimum at 1·5 g l⁻¹. The optimum pH for growth was 6·6 and growth occurred between pH 5·6 and 8·5. The isolate did not grow under aerobic conditions. Strain MET12^T fermented d-arabinose, cellobiose, fructose, galactose, glucose, lactose, maltose, mannose, raffinose, rhamnose, ribose, starch, sucrose, xylose, xylan and pyruvate. Casamino acids and peptone were only used in the presence of thiosulfate. Glucose was fermented into acetate as the major end product, together with lactate, H₂ and CO₂. l-Alanine was produced in trace amounts. Substrates that were not used were: acetate, butyrate, formate, fumarate, lactate, malate, butanol, methanol, propanol, methylamine, trimethylamine and H₂. The xylanase activity was shown to be endocellular. The isolate used elemental sulfur, thiosulfate and sulfite but not sulfate, nitrate or nitrite as electron acceptors. Strain MET12^T could grow in the presence of 10 μg chloramphenicol, kanamycin and vancomycin ml⁻¹ (no growth was obtained at 25 μg ml⁻¹) and in the presence of 25 μg ampicillin and rifampicin ml⁻¹ (no growth was obtained at 50 μg ml⁻¹). The G+C content of strain MET12^T was 36·1 mol% (HPLC).

16S rDNA sequence analysis performed on 1195 unambiguous nucleotides revealed that strain MET12^T was a member of the genus Petrotoga (family Thermotogaceae, order Thermotogales), with P. miotherma Davey et al. 1993 (98·9 % similarity) and P. olearia L'Haridon et al. 2002 (98·6 % similarity) being its closest phylogenetic relatives. Similarities to the two other species of the genus Petrotoga were 98·1 % (P. sibirica) and 97·6 % (P. mobilis) (Fig. 2⇓). DNA–DNA hybridization studies indicated that strain MET12^T showed DNA–DNA relatedness of 57·1 % to P. mobilis, 48·9 % to P. olearia, 53·4 % to P. sibirica and 65·4±3·2 % to P. miotherma.

Moderate to extreme thermophilic anaerobes belonging to the Bacteria and Archaea inhabit oil reservoirs originating from terrestrial or marine environments (Magot et al., 2000). However, more attention has been paid so far to micro-organisms of the domain Bacteria belonging to the families Thermoanaerobiaceae and, particularly, Thermotogaceae, being well-known inhabitants of these terrestrial- and marine-subsurface ecosystems (Davey et al., 1993; Fardeau et al., 1997a; Grassia et al., 1996; L'Haridon et al., 2001, 2002; Lien et al., 1998; Magot et al., 2000; Ravot et al., 1995a; Stetter et al., 1993). The order Thermotogales and family Thermotogaceae comprise members of the genera Thermotoga, Fervidobacterium, Thermosipho, Marinitoga, Petrotoga and Geotoga (Alain et al., 2002; Reysenbach, 2001; Wery et al., 2001). Amongst these genera, strains belonging to Petrotoga and Geotoga have been isolated so far only from oil reservoirs (Davey et al., 1993; L'Haridon et al., 2002; Lien et al., 1998). Based on phylogenetic studies, we provide evidence of the existence of a novel isolate (strain MET12^T) of the genus Petrotoga that inhabits an oil reservoir in the Gulf of Mexico. This confirms the close ecological relationship between micro-organisms of this genus and the oilfield ecosystem. Strain MET12^T has P. miotherma and P. olearia as its closest phylogenetic relatives, but DNA–DNA hybridization studies (values<70 %) with both these species and with P. sibirica and P. olearia allowed its assignment to a novel bacterial species (Wayne et al., 1987). In addition, strain MET12^T possesses the highest G+C content within the genus Petrotoga (Table 1⇓).

Besides phylogenetic and genomic differences between strain MET12^T and other Petrotoga species, we also found numerous phenotypic differences. Strain MET12^T tolerates the highest NaCl concentration recorded in the genus (up to 20 %, compared with a mean of 10 % for the other species) (Table 1⇑). Strain MET12^T differs from P. olearia and P. sibirica (L'Haridon et al., 2002) by using thiosulfate as a terminal electron acceptor (Table 1⇑). It also differs from P. mobilis (Lien et al., 1998) by using raffinose, mannose and pyruvate. In contrast to P. miotherma, strain MET12^T is motile. Therefore, the combined phylogenetic, genomic and phenotypic characteristics of strain MET12^T allow us to assign it to the genus Petrotoga as the type strain of a novel species, Petrotoga mexicana sp. nov.

It is noteworthy that thiosulfate can be used as an electron acceptor and that l-alanine is a minor end product of glucose metabolism by P. mexicana. l-Alanine production together with thiosulfate reduction have been found as quite common metabolic features within members of the order Thermotogales (Ravot et al., 1995b, 1996). Because of the position of these micro-organisms in the deepest branches of the phylogenetic tree, l-alanine production from glucose fermentation has been interpreted as a remnant of an ancestral metabolism (Ravot et al., 1996). A thiosulfate-reducing activity shared by several members of this order might (i) help them when oxidizing hydrogen or amino acids such as leucine, isoleucine and valine (Fardeau et al., 1993, 1997b) or (ii) have an undesirable influence on biocorrosion due to sulfide production (Magot et al., 2000). The occurrence of Petrotoga species only in oil reservoirs so far raises questions about (i) their survival in this deep subsurface environment and (ii) their possible indigenicity to this peculiar ecosystem. Further microbial and ecological studies are needed to appreciate their impact in the oilfield environment.