Rhodothermus profundi sp. nov., a thermophilic bacterium isolated from a deep-sea hydrothermal vent in the Pacific Ocean

At the time of writing, the assignment of the family ‘Rhodothermaceae’, phylum ‘Bacteroidetes’, is uncertain (Ludwig et al. 2008). The family contains two genera, Rhodothermus and Salinibacter. The former comprises a single species, Rhodothermus marinus. Since R. marinus DSM 4252^T was first isolated from a shallow submarine hot spring in Iceland, many additional strains have been reported from marine geothermal sites around the world (Alfredsson et al., 1988; Bjornsdottir et al., 2006). Although originating from distantly located sites, R. marinus strains have been shown to be closely related on the basis of highly similar 16S rRNA gene sequences and DNA–DNA reassociation values of 68–96 % (Moreira et al., 1996; Petursdottir et al., 2000; Silva et al., 2000). R. marinus strains are also very similar in their phenotypic characters. Most are orange-reddish-pigmented, although colourless strains have been isolated (Petursdottir et al., 2000). R. marinus is non-spore-forming, heterotrophic, slightly halophilic, thermophilic and obligately aerobic (Alfredsson et al., 1988). Therefore, its growth in marine hot springs is confined to aerobic zones of suitable temperatures. It has been shown that aerobic heterotrophic bacteria, for example of the genera Thermus and Bacillus, are present in deep-sea hydrothermal vent environments (Marteinsson et al., 1995, 1996, 1999). Thus, we investigated whether the distribution of Rhodothermus encompassed such environments. This paper reports the isolation and characterization of thermophilic bacteria isolated from a deep-sea hydrothermal vent at the East-Pacific Rise at 1 °N (12.44984 ° N 103.5681 ° W) and assigns them to a novel species within the genus Rhodothermus.

Samples were collected during the French oceanographic cruise AMISTAD in June 1999. Chimney rocks, hot fluids and invertebrates were collected by the man-operated submersible Nautile and brought to the surface in an insulated box. Samples were immediately processed for enrichment aboard the ship and for transportation to the laboratory. Various media were used for aerobic enrichments at 65 °C and some growth was observed after 1–30 days of incubation. The novel isolates were obtained from an active chimney rock sample collected at a depth of 2634 m. Rock pieces and sterile seawater were mixed together and kept at 4 °C until processing 2 weeks later in the laboratory. The mixture was inoculated into medium 160 (Degryse et al., 1978) containing 1 % NaCl and 2.5 g l⁻¹ each of tryptone and yeast extract (designated 160-1). The sample was serially diluted in 25 ml culture tubes and incubated aerobically at 65 °C for 2 weeks. As resistance of R. marinus to aminoglycoside antibiotics had been reported previously (Alfredsson et al., 1988), streptomycin (1.56 μg ml⁻¹) was added to the enrichment medium to facilitate growth of Rhodothermus-like species and to prevent overgrowth of fast-growing opportunist-like Bacillus species. Previous experiments showed that Bacillus sp. and R. marinus exhibited different levels of resistance to streptomycin of up to 0.78 μg ml⁻¹ and 12.5 μg ml⁻¹, respectively (data not shown). Growth was not observed in the dilutions, but was observed after concentrating the cultures and subsequent spreading on agar plates containing marine agar (DSMZ medium 514; Bacto marine broth, Difco 2216) diluted to half strength with water (designated MB1/2) (Marteinsson et al. 1995). Homogeneous colonies appeared on the plates and nine colourless colonies were randomly picked and purified by restreaking on the same medium at least six times. Isolates PRI 2894, PRI 2895, PRI 2896, PRI 2897, PRI 2898, PRI 2901, PRI 2902^T, PRI 2904 and PRI 2908 were further examined. R. marinus DSM 4252^T was used as a reference strain throughout this work.

For phenotypic analysis, the novel isolates were grown on agar medium MB1/2. They showed very similar phenotypic characteristics, which were consistent with their assignment to the genus Rhodothermus. They stained Gram-negative and were positive for catalase and negative for oxidase. Colonies were circular, smooth, viscous and 2–3 mm in diameter after incubation at 70 °C for 4 days. Colonies were colourless, whereas those of R. marinus 4252^T are orange–red (Alfredsson et al., 1988). Sensitivity of the strains to antibiotics was examined at 65 °C on plates using standard antibiotic discs (6 mm) from Oxoid. The inhibition zone was measured after 24 h and strains scored sensitive to an antibiotic if the zone was larger than 20 mm in diameter (Kristjansson et al., 1994). The strains were sensitive to ampicillin (10 μg), chloramphenicol (30 μg), erythromycin (15 μg), novobiocin (30 μg), penicillin G (10 IU) and rifampicin (2 μg). They were resistant to gentamicin (10 μg), kanamycin (30 μg), nalidixic acid (30 μg), polymyxin B (300 IU) and streptomycin (10 μg). Phase-contrast microscopy revealed that cells of strain PRI 2902^T occurred singly or in short chains during the exponential phase. Motility and endospores were never observed. However, motility has been observed for R. marinus DSM 4252^T. Cell morphology of strain PRI 2902^T was examined using a Leo Supra 25 scanning electron microscope. Cells appeared as irregular rods and measured approximately 0.5 μm in width and 1.5–3.5 μm in length (Supplementary Fig. S1, available in IJSEM Online). The DNA G+C content of strain PRI 2902^T was 60.9 mol%, determined at the DSMZ according to Mesbah et al. (1989). Phenotypically, the novel isolates differed from R. marinus 4252^T by characteristics represented in Table 1⇓.

Isolates PRI 2894, PRI 2895, PRI 2898, PRI 2901, PRI 2902^T, PRI 2904 and PRI 2908 were grown in liquid medium MB1/2 at several temperatures, pH and NaCl concentrations. Growth was examined as described previously (Marteinsson et al., 1995). Growth was observed at 55–80 °C, with optimum growth at 70–75 °C. All isolates grew in the presence of 1 % and up to 6 % NaCl. However, a slight difference was observed for the optimal NaCl concentration which ranged from 1 to 3 % NaCl. Growth occurred between pH 6.0 and 8.0; optimum growth occurred around pH 7.0. Growth on single carbon sources was tested as described previously (Alfredsson et al., 1985; Marteinsson et al., 1995) on plates containing MB1/2 without peptone and yeast extract. The strains utilized asparagine, aspartate, casein, Casamino acids, gelatin, glutamate, glutamine, pyruvate and starch. Anaerobic growth was examined as described previously (Marteinsson et al., 1995). The isolates could not grow anaerobically and oxygen could not be replaced by nitrate.

Growth of isolates PRI 2895, PRI 2901 and PRI 2902^T and R. marinus DSM 4252^T was examined under in situ hydrostatic pressure (25 MPa), as well as under atmospheric hydrostatic pressure (0.1 MPa). Experiments were performed in duplicate as described previously (Marteinsson et al., 1999). Growth was monitored by microscopic observation and by measuring the turbidity at 600 nm. All strains exhibited growth at both 0.1 and 25 MPa. R. marinus DSM 4252^T produced about 30 % fewer cells under hydrostatic pressure of 25 MPa. Furthermore, cell deformations were observed, which were never observed at atmospheric pressure. However, the in situ hydrostatic pressure of 25 MPa did not affect the growth rates of isolates PRI 2901 and PRI 2902^T and strain PRI 2895 produced about 35 % more cells than at atmospheric hydrostatic pressure. No cell deformations were observed for isolates PRI 2895, PRI 2901 and PRI 2902^T when grown under in situ hydrostatic pressure, although the cells appeared shorter.

Fatty acids were extracted from strain PRI 2902^T and R. marinus DSM 4252^T according to the protocol of the MIDI system. Analysis by GC was controlled by the MIS software and the peaks were automatically integrated and identified by the Microbial Identification software package (Sasser, 1990). The general fatty acid profiles were similar to those presented by Silva et al. (2000) for R. marinus strains and indicated that PRI 2902^T belonged to the genus Rhodothermus. However, the comparison was inadequate as PRI 2902^T could not grow in the medium used by Silva et al. (2000). The fatty acid profiles of PRI 2902^T and R. marinus DSM 4252^T grown in parallel in liquid medium MB1/2 at 70 °C are given in Table 2⇓. Both strains contained anteiso-C_15 : 0, iso-C_16 : 0, iso-C_17 : 0 and anteiso-C_17 : 0 as predominant fatty acids, although in different ratios. The most abundant fatty acid in PRI 2902^T was iso-C_17 : 0, whereas it was iso-C_16 : 0 in R. marinus DSM 4252^T. Furthermore, the proportion of straight-chain fatty acids was lower in PRI 2902^T. Quinone and polar lipid analyses were carried out by the Identification service and Dr B. J. Tindall, DSMZ, Braunschweig, Germany. Strain PRI 2902^T contained menaquinone MK-7 (100 %), which is the predominant respiratory quinone in R. marinus DSM 4252^T (97 %) (Tindall, 1991). The polar lipid pattern of PRI 2902^T was similar to that of R. marinus DSM 4252^T (Tindall, 1991) with diphosphatidylglycerol and phosphatidylethanolamine as major phospholipids. Phosphatidylglycerol was also detected, as well as phospholipids PL1–PL2 and glycolipids GL1–GL3. The chemotaxonomic properties of strain PRI 2902^T support its affiliation to the genus Rhodothermus.

Genomic DNA was isolated with the Master Pure DNA Purification kit (Epicentre). The 16S rRNA gene was amplified with primers F9 and R1544 as described previously (Skirnisdottir et al., 2000). Nucleotide sequences were determined using an Applied Biosystems 3730 DNA analyser and the BigDye terminator cycle sequencing kit. Primers F9 and R1544 were used for sequencing, as well as primers F338, R357, R805 and R1195 (Skirnisdottir et al., 2000). The almost complete 16S rRNA gene sequences (1509 bp), which were determined for all nine isolates, were identical. Comparison of this sequence with sequences available from public databases demonstrated that the isolates belonged to the genus Rhodothermus. The sequence showed 98.1 % similarity to the sequence of R. marinus DSM 4252^T (X80994), which was the most closely related species. This is below the threshold value of 98.7 % that was recently suggested for species definition (Stackebrandt & Ebers, 2006). The two sequences were aligned with those of related bacteria using clustal x (Thompson et al., 1997). Evolutionary distance matrices were calculated using the Kimura two-parameter model (Kimura, 1980). The 16S rRNA gene sequence of strain PRI 2902^T exhibited 85 % similarity to the sequence of the type strain of Salinibacter ruber of the family ‘Rhodothermaceae’ and 80 % similarity to the type strain of Thermonema lapsum (Hudson et al., 1989; Antón et al., 2002; Ludwig et al., 2008). DNA–DNA reassociation between strain PRI 2902^T and R. marinus DSM 4252^T was performed by the DSMZ, as described by De Ley et al. (1970) with the modifications of Huß et al. (1983); a value of 37.2 % was obtained, which clearly indicated that the isolates described here do not belong to R. marinus when a threshold value of 70 % DNA–DNA reassociation is adopted for defining species (Wayne et al., 1987).

Five of the novel isolates (PRI 2894, PRI 2895, PRI 2898, PRI 2902^T and PRI 2908) were analysed by amplified-fragment length polymorphism (AFLP) fingerprinting. Three R. marinus reference strains were included: the type strain DSM 4252^T from north-west Iceland (Alfredsson et al., 1988) and two colourless isolates, PRI 831 and PRI 1298, from south and north-east Iceland, respectively (Petursdottir et al., 2000). AFLP analysis was performed essentially as described previously (Vos et al., 1995). Genomic DNA was digested with ApaI and TaqI and ligated to double-stranded adapters. Primers A00 and T00 were used for pre-selective amplification and A01-6-FAM and T00 for selective amplification (Huys & Swings, 1999). The selective amplification products were separated on a 6 % acrylamide gel with 8 M urea. The AFLP fingerprints were visualized using the Typhoon 9400 Variable Mode Imager (Amersham). The presence of individual bands was scored manually. Clustering of the patterns was performed with NTSYSpc (Rohlf, 2008) using the Dice correlation coefficient and the UPGMA clustering algorithm (Fig. 1⇓). Each fingerprint consisted of a mean of 80±8 bands where only eight bands were common for all the strains. An identical banding pattern among the novel isolates was observed. However, a considerable genetic heterogeneity was seen between the novel isolates and the three Icelandic R. marinus reference strains (Fig. 1⇓). The former were 26 % similar to R. marinus DSM 4252^T and 22 % and 26 % similar to the colourless isolates PRI 831 and PRI 1289, respectively. For some taxonomic groups, AFLP pattern similarity above 60 % indicates different strains of the same species (Savelkoul et al., 1999; Thompson et al. 2004). Thus, the AFLP results presented here further support the species status of the isolates described.

On the basis of 16S rRNA gene phylogeny and chemotaxonomy, it is clear that the isolates described here belong to the genus Rhodothermus. DNA–DNA hybridization supports their status as a novel species. The strains can be differentiated genotypically from R. marinus by AFLP as well as phenotypically by their colony colour, fatty acid composition, their optimum and maximum temperatures for growth, response to hydrostatic pressure and sensitivity to rifampicin and erythromycin, as well as their inability to assimilate acetate, galactose, sucrose, maltose and lactose. Therefore, it is concluded that the strains examined in this study represent a novel species within the genus Rhodothermus. Because of the deep origin of the isolates, the name Rhodothermus profundi sp. nov. is proposed with PRI 2902^T (=DSM 22212^T =JCM 15944^T) as the type strain.