Amphritea japonica sp. nov. and Amphritea balenae sp. nov., isolated from the sediment adjacent to sperm whale carcasses off Kagoshima, Japan

Members of the genus Amphritea are described as Gram-negative, rod-shaped cells with a monopolar flagellum and have been isolated from a Bathymodiolus sp. specimen collected at the Mid-Atlantic Ridge at a depth of about 3000 m (Gärtner et al., 2008). Amphritea atlantica is the only recognized species of this genus. The genus Amphritea is related to the genera Oceanospirillum and Neptunomonas and to a symbiotic bacterial clone of Osedax (Goffredi et al., 2007). The genus Osedax (Polychaeta: Siboglinidae) has recently been discovered in whale carcass environments on the deep-sea floor (Rouse et al., 2004; Glover et al., 2005; Fujikura et al., 2006; Braby et al., 2007). Osedax specimens host symbiotic bacteria in their ovisac and root systems. Phylogenetic analysis placed these Osedax microbes within a well-supported clade of the Gammaproteobacteria that contains heterotrophic members of the order Oceanospirillales (Goffredi et al., 2007). Most genera of the order Oceanospirillales (Garrity et al., 2005) are halotolerant or halophilic and motile. These genera are aerobic, microaerophilic or facultatively anaerobic chemo-organotrophs. Sperm whale carcass ecosystems off Kagoshima, Japan, were investigated over a period of 3 years (Fujiwara et al., 2007). About 800 aerobic bacteria were isolated from sediment adjacent to the sperm whale carcasses and three of these strains were related to members of the family Oceanospirillaceae and a symbiotic bacterial clone of Osedax japonicus. In this paper, several lines of evidence are presented that indicate that these isolates represent two novel species of the genus Amphritea.

The three strains, JAMM 1866^T, JAMM 1548 and JAMM 1525^T, were collected from marine sediment adjacent to the sperm whale carcasses by the unmanned ROV Hyper Dolphin off Kagoshima, Japan (dive 457, 3 ° 20.720′ N 12 ° 59.285′ E, at a depth of 228 m; dive 465, 3 ° 20.730′ N 12 ° 59.314′ E, at a depth of 229 m) during cruise NT05-12. The sediment samples were collected with the manipulator of the ROV and placed in the sample holder of the sterilized sampler. A portion of each sample was cultivated on marine agar 2216 (MA; Difco) at 12 °C for approximately 7 days. The bacteria were maintained on MA plates or in marine broth 2216 (MB; Difco) at 20 °C and stored at −80 °C in 15 % (v/v) glycerol. Unless otherwise indicated, physiological tests were performed according to the general procedures described by Barrow & Feltham (1993) and Miyazaki et al. (2008b) with the modification that artificial seawater (ASW; 1×ASW consisting of 2.75 % NaCl, 0.07 % KCl, 0.54 % MgCl₂ . 6H₂O, 0.68 % MgSO₄ . 7H₂O, 0.14 % CaCl₂ , 2H₂O, 0.02 % NaHCO₃) was used. Growth at various temperatures (4–30 °C) was measured in MB for 7 days. Growth at various NaCl concentrations was examined in medium containing 0.5 % peptone (Difco), 0.5 % yeast extract (Difco) and 0.32 % MgSO₄ . 7H₂O, with NaCl concentrations of 0–6 % (w/v) (pH adjusted to 7.2 at 20 °C). The pH range for growth was determined in 0.5 % peptone, 0.5 % yeast extract, 0.32 % MgSO₄ . 7H₂O and 3.0 % NaCl (w/v), with the pH adjusted to 6.0, 6.5, 7.0, 7.5, 8.0 and 8.5. The test strains retained viability for about 7 days at the optimal temperature. Acid production from sugars was assessed using modified oxidization–fermentation medium (Hugh & Leifson, 1953) containing 0.5×ASW, 0.1 % (NH₄)₂SO₄, 0.1 % yeast extract (Difco), 0.1 % Tris, 1.4 % NaCl, 1 % sugar and 0.006 % bromothymol blue (pH adjusted to 7.2 at 20 °C) and incubated at the optimum temperature. Oxidase activity was determined by spreading cell pellets on oxidase test paper (Nissui Pharmaceutical). Catalase activity was determined based on bubble production in 3 % (v/v) H₂O₂ solution. Gelatinase, protease, amylase and lipase (hydrolysis of tri-n-butyrin) activities were detected on MA plates using substrate concentrations of 1 % (w/v). DNase activity was assessed using DNase test agar (Difco). Hydrogen sulfide production from thiosulfate and indole production were assessed using SIM medium (Nissui Pharmaceutical) stabs prepared with 0.5×ASW instead of water. Additional phenotypic characteristics were determined using the Biolog GN2 microbial identification system following the protocol provided by the manufacturer. Susceptibility to antimicrobial substances was examined on MA using Sensi-Discs (Becton Dickinson). Any sign of growth inhibition after 72 h incubation at 20 °C was recorded as sensitivity to the respective antimicrobial agent. The following antibiotics (Becton Dickinson) were examined (amounts per disc): ampicillin (10 μg), chloramphenicol (30 μg), erythromycin (15 μg), gentamicin (10 μg), kanamycin (30 μg), nalidixic acid (30 μg), neomycin (30 μg), novobiocin (30 μg), penicillin (10 IU), streptomycin (10 μg) and tetracycline (30 μg).

Transmission electron microscopy of negatively stained cells was conducted as described by Nogi et al. (1998). Cells of strains JAMM 1866^T and JAMM 1525^T grown on MA at optimal temperature and in the mid-exponential phase of growth were used for electron microscopic observations (JEM-1210; JEOL).

Cellular fatty acids and isoprenoid quinones were analysed. Isolated strains were cultured in MB at optimal growth temperatures. Cellular fatty acids were analysed by GLC-MS (Miyazaki et al., 2006). Isoprenoid quinones were extracted with chloroform/methanol (2 : 1) from dried cells (200 mg) and purified by TLC. The purified isoprenoid quinones were analysed using reversed-phase HPLC (Komagata & Suzuki, 1987).

Chromosomal DNA was purified using the phenol extraction method (Saito & Miura, 1963). The DNA G+C content was determined using reversed-phase HPLC (Tamaoka & Komagata, 1984). For analysis of relatedness, DNA–DNA hybridization was carried out at 44 °C for 4 h and measured fluorometrically using the method of Ezaki et al. (1989).

16S rRNA gene sequences were obtained by direct sequencing of PCR-amplified DNA as described previously (Miyazaki et al., 2008a). Sequences were aligned with the subset of 16S rRNA gene sequences obtained from GenBank using the fastaligner utility of the arb software (Ludwig et al., 2004). The phylogenetic relationships of each taxon were inferred with the neighbour-joining (NJ) and maximum-likelihood (ML) methods using paup* 4.0 beta 10 (Swofford, 1998). ML trees (Felsenstein, 1981) were inferred using the default starting parameters (NJ using a Jukes–Cantor model of evolution). NJ trees (Saitou & Nei, 1987) were inferred using the Hasegawa, Kishino and Yano (HGK85) distance. Statistical support of the branch points was tested by performing 1000 NJ and 500 ML bootstrap replications. The GenBank/DDBJ/EMBL accession numbers for the 16S rRNA gene sequences of the isolates are shown in Fig. 1⇓. Other reference sequences were obtained from GenBank.

Cultural, physiological and biochemical characteristics of the isolates are shown in Table 1⇓ and given in the species descriptions below. Physiological and biochemical characteristics of the two groups showed some differences when compared with the type strains of related species and with each other. Strains JAMM 1866^T and JAMM 1548 showed no growth at 30 °C or in 6 % NaCl and did not hydrolyse Tweens 40 and 80; their DNA G+C contents differed from those of related strains. Strain JAMM 1525^T had a bipolar or polar flagellum, did not grow at 30 °C or in 6 % NaCl, was positive for catalase (weak) and gelatinase, and did not hydrolyse Tweens 40 and 80; the DNA G+C content differed from those of related strains.

The whole-cell fatty acid compositions of the isolates and reference strains are shown in Table 2⇓. The major fatty acids of the novel isolates and A. atlantica M41^T were C_16 : 0, C_16 : 1 and C_18 : 1. Furthermore, these strains contained C_12 : 1 3-OH, a characteristic feature of the genus Amphritea. The fatty acid profile showed low levels of similarity to those of reference strains. For example, the isolates produced C_12 : 1 and C_18 : 2, but A. atlantica M41^T did not produce C_12 : 1 and the reference strains did not produce C_18 : 2.

Results of the NJ phylogenetic analyses using 16S rRNA gene sequences are shown in Fig. 1⇑. These results support the conclusions described below and further clarify the taxonomic and phylogenetic positions of the three isolated strains and the symbiotic bacterial clone R21 of Osedax within the genus Amphritea. The three strains showed high sequence similarities to the following species/genera: A. atlantica M41^T (97.7–97.8 %), genus Neptunomonas (92.6–94.3 %), genus Oceanospirillum (91.3–93.2 %) and Neptuniibacter caesariensis MED92^T (94.3–94.5 %). However, they were more closely related to the 16S rRNA gene sequence of symbiotic bacterial clone R21 of Osedax japonicus (99.0–100 %). The level of 16S rRNA gene sequence similarity between strains JAMM 1866^T and JAMM 1525^T was 99.0 %. The generally recommended and accepted criteria for delineating bacterial species state that strains with 16S rRNA gene sequence dissimilarity of greater than 3 % are considered to belong to separate species (Stackebrandt & Goebel, 1994; Stackebrandt et al., 2002). The generally recognized criteria for delineating bacterial species state that strains with a DNA–DNA relatedness of less than 70 %, as measured by hybridization, represent separate species (Wayne et al., 1987). The results of DNA–DNA hybridization analysis of the isolated strains and A. atlantica M41^T showed that there was greater than 91 % DNA–DNA relatedness between strains JAMM 1866^T and JAMM 1548, but the relatedness values between these two strains, JAMM 1525^T and A. atlantica were less than 18 % and each group was clearly separate, representing distinct species according to the recommendations of Wayne et al. (1987).

From the results of phenotypic, genotypic and phylogenetic analyses, it is evident that the three isolated strains represent two novel species within the genus Amphritea for which the names Amphritea japonica sp. nov. and Amphritea balenae sp. nov. are proposed.