Introduction

The family Rhodobacteraceae is a major group in the class Alphaproteobacteria (Garrity et al. 2005) and currently it comprises 93 genera according to the NCBI database. The genus Roseovarius, a member of the family Rhodobacteraceae, was proposed by Labrenz et al. (1999) with the description of a single species, Roseovarius tolerans, which was isolated from a hypersaline lake. Ten further Roseovarius species, R. nubinhibens (González et al. 2003), R. crassostreae (Boettcher et al. 2005), R. mucosus (Biebl et al. 2005), R. aestuarii (Yoon et al. 2008), R. pacificus (Wang et al. 2009), R. halotolerans (Oh et al. 2009), R. nanhaiticus (Wang et al. 2010), R. marinus (Jung et al. 2011), R. indicus (Lai et al. 2011), R. halocynthiae (Kim et al. 2012b) and R. litoreus (Jung et al. 2012) have since been described; all of them were isolated from marine environments. The genus Pelagicola, another member of the family Rhodobacteraceae, was proposed by Kim et al. (2008). Presently, there is only one species in this genus, Pelagicola litoralis, which was isolated from coastal seawater (Kim et al. 2008).

A novel bacterial strain, designated H50T, was isolated from a bottom seawater sample collected from an amphioxus breeding zone in the coastal region of the Yellow Sea, China in November 2009. In the present study, strain H50T has been characterized using a polyphasic taxonomic approach and was found to represent a novel species belonging to the genus Roseovarius, for which we propose the name R. marisflavi sp. nov.

Materials and methods

Organism and culture conditions

A bottom seawater sample (about 10-m deep) was collected from an amphioxus breeding zone in the coastal region of the Yellow Sea (36°5′N, 120°32′E), China, two miles off shore in November 2009. The cultivation of microorganisms from the sample was carried out using a modified gel microbead (GMD) cultivation method (Zengler et al. 2002; Ji et al. 2012). The individual microcolony-forming GMD was picked out by flow cytometry through Fluorescence Activated Cell Sorting and then sorted into 96-well microtiter plates filled with marine R2A (0.5 g/L yeast extract, 0.5 g/L proteose peptone, 0.5 g/L casamino acids, 0.5 g/L dextrose, 0.5 g/L soluble starch, 0.3 g/L sodium pyruvate, 20 g Agar No. 1, 75 % sea water) medium (Suzuki et al. 1997). After 8–12 days cultivation at 16 °C, the turbid wells (OD600 ≥ 0.1) were chosen for bacterial purification on marine R2A agar plates. Individual colonies showing different traits were picked off and purified by successive streaking on fresh marine agar 2216 (MA; BD) plates at 28 °C. Stocks were kept at −80 °C in sterile 0.9 % (w/v) NaCl supplemented with 15 % (v/v) glycerol. One of the isolates, designated H50T, has been further characterized.

The type strains of phylogenetically related species, R. tolerans DSM 11457T, R. mucosus DSM 17069T and P. litoralis DSM 18290T, obtained from the DSMZ (Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH, Braunschweig, Germany), were used as reference strains and were cultured under the same conditions as H50T [MA or marine broth 2216 (MB) at 28 °C], unless otherwise specified.

Determination of 16S rRNA gene sequence and phylogenetic analysis

Genomic DNA was extracted from 48 h old cultures on MA by standard methods (Ausubel et al. 1995). The 16S rRNA gene was amplified by PCR using universal primers B8f (5′-AGAGTTTGATCCTGGCTCAG-3′) and B1510 (5′-GGTTACCTTGTTACGACTT-3′). Purified PCR products were ligated to the pMD 18-T (TaKaRa) and cloned to Escherichia coli JM109 according to the manufacturer’s instructions. Sequencing reactions were carried out using ABI BigDye 3.1 Sequencing kits (Applied BioSystems) and an automated DNA sequencer (model ABI3730; Applied BioSystems). The near-complete 16S rRNA gene sequence of strain H50T was submitted to GenBank/EMBL to search for similar sequences using the BLAST algorithm. The identification of phylogenetic neighbours and the calculation of pairwise 16S rRNA gene sequence similarities were achieved using the EzTaxon-e server (http://eztaxon-e.ezbiocloud.net/; Kim et al. 2012a). Sequences were aligned using CLUSTAL X1.8 (Thompson et al. 1997). Phylogenetic trees were constructed using the neighbour-joining (NJ), maximum-likelihood (ML) and maximum-parsimony (MP) methods with Kimura 2-state parameter model analyses implemented in the program MEGA version 5 (Tamura et al. 2007). In each case, bootstrap values were calculated based on 1,000 replicates.

Phenotypic characterization

Cell morphology was determined by transmission electron microscopy (JEOL; JEM-1200EX) after cells had been negatively stained with 1 % (w/v) phosphotungstic acid. Gram and endospore staining was carried out using the standard method (Beveridge et al. 2007). Poly-beta-hydroxybutyrate (PHB) accumulation was tested by fluorescence microscope (BH2; OLYMPUS) using Nile Blue A as a fluorescent stain (Tindall et al. 2007). Growth under anaerobic conditions was determined on MA after incubation in an anaerobic jar which was filled with nitrogen and with a packet of AneroPack-Anaero (Mitsubishi Gas Chemical Co.) at 28 °C for 30 days. Halotolerance was investigated using synthetic marine ZoBell broth [5 g Bacto peptone, 1 g yeast extract and 0.1 g FePO4 in 1 L modified artificial seawater (Lyman and Fleming 1940), in which starting Na+ was replaced by the appropriate K+] supplied with various concentrations of NaCl (0–15 %, w/v, in increments of 1 %). Tolerance of sea salts was determined using synthetic ZoBell broth (5 g Bacto peptone, 1 g yeast extract and 0.1 g FePO4 in 1 L distilled water) supplied with different concentrations (0–15 %, w/v, in increments of 1 %) of sea salts (Sigma). The temperature range for growth was determined on MA plates incubated at 0–45 °C (0, 4, 16, 22, 28, 37 and 45 °C) for 30 days. The pH range for growth was tested at pH 2–10 (in steps of 1 pH unit) in MB with citrate/phosphate buffer (pH 2.0–7.0), Tris/HCl (pH 8.0–9.0) (Breznak and Costilow 1994) or sodium carbonate/sodium bicarbonate (pH 10.0). Susceptibility to antibiotics was investigated on MA plates by using discs (Hangzhou Microbiology Reagent) containing different antibiotics.

Standard protocols (Tindall et al. 2007) were used to assess catalase and oxidase activities, and degradation of casein, starch, gelatin, lecithin and Tween 20, 40, 80. The ability to hydrolyse alginate was tested as described by Smibert and Krieg (1994). Degradation of chitin was examined on chitin agar with sterile seawater (Hsu and Lockwood 1975). DNase activity was examined by using DNase agar (Qingdao Hope Bio-technology Co., Ltd) with sterile seawater, according to the manufacturer’s instructions.

Bacteriochlorophyll (Bchl) a content of cells was determined in acetone/methanol (7:2, v/v) extracts using a TU-1810 spectrophotometer (Pgeneral) as described previously (Martens et al. 2006). Activities of constitutive enzymes, substrate oxidation, carbon source utilization and other physiological properties were determined by using the API 20NE, API 20E and Gram-negative MicroPlates (Biolog GN2), according to the manufacturer’s instructions except that sterile seawater was used to prepare the inocula.

Determination of chromosomal DNA G+C content

Genomic DNA was extracted from 48 h old cultures on MA by standard method (Ausubel et al. 1995). The G+C content of the chromosomal DNA was determined as described by Mesbah et al. (1989) using a reverse-phase HPLC.

Determination of fatty acid, polar lipids compositions and isoprenoid ubiquinones

Cell masses of H50T, R. mucosus DSM 17069T, R. tolerans DSM 11457T and P. litoralis DSM 18290T were obtained from MB after cultivation at 28 °C for 2–7 days when the bacterial communities reached the late exponential stage. Cellular fatty acids of H50T, R. tolerans DSM 11457T and P. litoralis DSM 18290T were prepared and analysed according to the standard protocol described in the Sherlock System (MIDI) (Sasser 1990) with the Sherlock version 6.0 and the included database. Polar lipids of H50T, R. tolerans DSM 11457T and P. litoralis DSM 18290T were analysed by using standard procedures, in which extracted lipids were separated by two-dimensional TLC and identified by spraying with appropriate detection reagents (Minnikin et al. 1984). Respiratory quinones of H50T were analysed by Dr B. J. Tindall of the Identification Service of the DSMZ (Braunschweig, Germany).

Results

Determination of 16S rRNA gene sequence and phylogenetic analysis

A BLASTN search using the almost-complete 16S rRNA gene sequence (1,438 bp; GenBank accession number KC900366) of strain H50T placed it among the members of family Rhodobacteraceae in the Alphaproteobacteria. Strain H50T exhibited the highest 16S rRNA gene sequence similarity to P. litoralis DSM 18290T (96.1 %), R. mucosus DSM 17069T (95.8 %) and R. tolerans DSM 11457T (95.7 %), and lower than 95.7 % to all the other related strains. In the phylogenetic tree based on the NJ algorithm (Fig. 1), Strain H50T formed a coherent cluster with two species of the genus Roseovarius, i.e. R. mucosus DSM 17069T and R. tolerans DSM 11457T (the type species of genus Roseovarius) with a high bootstrap resampling value of 77 % but occupied a separate phylogenetic branch in this cluster. However, strain H50T did not form a monophyletic cluster with either P. litoralis DSM 18290T (the only species in the genus) or other genera. Topologies of phylogenetic trees constructed using the ML (Supplementary Fig. S1) and MP (Supplementary Fig. S2) algorithms were similar to the NJ phylogenetic tree but the bootstrap resampling values were lower than 70 %. On the basis of the phylogenetic evidence, strain H50T should be classified as a novel species in the genus Roseovarius instead of genus Pelagicola.

Fig. 1
figure 1

Phylogenetic dendrogram of Roseovarius marisflavi H50T and related species based on 16S rRNA gene homology. The tree was constructed using the neighbor-joining method with Kimura 2-state parameter model analyses implemented in the program MEGA version 5. Bar 0.01 nucleotide substitutions per site

Full size image

Phenotypic characterization

Cells of strain H50T were observed to be Gram stain-negative, strictly aerobic, ovoid or short rods, approximately 0.5–1.3 μm long by 0.4–0.8 μm wide, with no flagella (Fig. 2). Cells did not form endospores. PHB was found to be accumulated. Bacteriochlorophyll a (Bchl a) was found to be produced. The halotolerance range for growth was determined to be 1–15 % NaCl (optimum 2–7 %). The sea salts range for growth is 1–11 %. The pH range for growth was determined to be 6.0–8.0 (optimum 7.0). The temperature range for growth was determined to be 0–37 °C (optimum 28–37 °C).

Fig. 2
figure 2

Transmission electron micrographs of negatively stained cells of Roseovarius marisflavi H50T. a In stationary phase, b in log phase. Bars a 100 nm, b 200 nm

Full size image

The oxidase and catalase tests were positive. Tween 20, DNA and esculin ferric citrate were degraded by strain H50T but starch, Tween 80, gelatin, casein, alginate, lecithin or chitin was not degraded. Tween 40 was weakly degraded. Arginine dihydrolase and urease tests were positive but β-galactosidase activity was negative. Acid was not observed to be produced from d-glucose. Antibiotic susceptibilities were determined as follows: susceptible to streptomycin (10 μg, 17 mm), rifampicin (5 μg, 24 mm), tobramycin (10 μg, 16 mm), trimethoprim (5 μg, 16 mm), ofloxacin (5 μg, 23 mm), amoxicillin (10 μg, 32 mm), minocycline (30 μg, 24 mm), tetracycline (30 μg, 22 mm), cefobid (75 μg, 41 mm), gentamicin (10 μg, 20 mm), ceftriaxone (30 μg, 48 mm), chloromycetin (30 μg, 22 mm) and oxacillin (1 μg, 16 mm); slightly susceptible to neomycin (30 μg, 22 mm), lincomycin (2 μg, 20 mm) and erythromycin (15 μg, 18 mm); resistant to midecamycin (30 μg, 12 mm), doxycycline (30 μg, 12 mm), amikacin (30 μg, 14 mm), ciprofloxacin (5 μg, 9 mm), clindamycin (2 μg, 11 mm), vancomycin (30 μg, 9 mm), polymyxin B (300 IU, 8 mm), penicillin (10 IU, 0 mm), ampicillin (10 μg, 0 mm), carbenicillin (100 μg, 0 mm).

The chromosomal DNA G+C content

The DNA G+C content of strain H50T was determined to be 57.7 mol%.

Determination of fatty acid, polar lipids compositions and isoprenoid ubiquinones

There were eight kinds of fatty acid detected in strain H50T (Table 1), with C18:1ω7c and C16: 0 as the major components of cellular lipid acids; and other cellular fatty acids present were identified as C14:0, C18:0, C16:1ω7c/C16:1ω6c, C19:0 cycloω8c, C18:1 2-OH and C18:1ω7c 11-methyl. The following polar lipids were found to present in H50T: phosphatidylglycerol (PG), diphosphatidylglycerol (DPG) and a lipid (L) (Supplementary Fig. S3; the polar lipid profiles for R. tolerans DSM 11457T and P. litoralis DSM 18290T are available in Fig. S4 and S5, respectively). The predominant isoprenoid quinone of H50T was determined to be Q-10, consistent with members of the genus Roseovarius.

Table 1 Cellular fatty acid content (% of total) of H50T and the related species
Full size table

Taxonomic conclusion

The cultural, physiological and biochemical characteristics of strain are given in Table 2 and the species description (see below). Strain H50T exhibited the highest 16S rRNA gene sequence similarity to members of the genera Roseovarius and the Pelagicola, but the phylogenetic evidence and chemotaxonomic data support the view that H50T represents a novel species in the genus Roseovarius instead of the genus Pelagicola. The reasons are as following: (1) In the phylogenetic trees based on the NJ, ML and MP methods, strain H50T clearly formed a coherent cluster with the type species of Roseovarius, but not the sole representative of the genus Pelagicola; (2) The presence of Bchl a and PHB of strain H50T are consistent with the description of the genus Roseovarius (Labrenz et al. 1999), whereas no Bchl a or PHB are present in P. litoralis DSM 18290T (Kim et al. 2008); (3) The temperature range (0–37 °C) and sea salts range (1–11 %) for growth of strain H50T is similar to that of R. tolerans DSM 11457T and R. mucosus DSM 17069T but obviously wider than that of P. litoralis DSM 18290T (16–28 °C and 2–6 %, respectively); (4) The DNA G+C content of strain H50T (57.7 mol%) is within the range for members of the genus Roseovarius (55.4–66 mol%) (Lai et al. 2011; Kim et al. 2012b; Jung et al. 2012) but much higher than that of P. litoralis DSM 18290T (47.0 mol%; Kim et al. 2008); (5) C18:1ω7c and C16:0 were the major fatty acid components of H50T, which is consistent with the profile of R. tolerans DSM 11457T and R. mucosus DSM 17069T. However, in P. litoralis DSM 18290T, C18:1ω7c and C10:0 3-OH were major fatty acid components, but there was no detectable C10:0 3-OH in strain H50T; (6) The two identified polar lipids (PG and DPG) in strain H50T were also found in R. tolerans DSM 11457T and R. mucosus DSM 17069T (Biebl et al. 2005) but DPG could not be found in P. litoralis DSM 18290T.

Table 2 Characteristics that differentiate H50T from related strains
Full size table

Phylogenetic analysis, phenotypic characterization, DNA G+C content and chemotaxonomic characteristics support the assignment of strain H50T to the genus Roseovarius instead of the genus Pelagicola. However, the new strain H50T also showed significantly different physiological and biochemical characteristics to the phylogenetically closest species in the genus Roseovarius: (1) Hydrolytic activities. DNases, arginine dihydrolase and urease were found to be positive in strain H50T, which is in contrast to either R. tolerans DSM 11457T or R. mucosus DSM 17069T; (2) Sole carbon sources. Strain H50T can utilize d-arabitol, d-glucuronic acid and α-ketovaleric acid as sole carbon sources, but R. tolerans DSM 11457T or R. mucosus DSM 17069T could not; (3) Polar lipids. Phosphatidylcholine (PC) and Phosphatidylethanolamine (PE) were detected in both R. tolerans DSM 11457T (Labrenz et al. 1999) and R. mucosus DSM 17069T (Biebl et al. 2005) but were not present in strain H50T. Other differential properties between H50T and the phylogenetically closely related species are given in Table 2. The different physiological and biochemical characteristics imply that strain H50T should not be allocated to the closely related species R. mucosus DSM 17069T or the type species R. tolerans DSM 11457T within the genus Roseovarius. In summary, genotypic, phenotypic, and chemotaxonomic data support the conclusion that strain H50T from an amphioxus breeding zone in the coastal region of the Yellow Sea represents a novel species of the genus Roseovarius, for which the name R. marisflavi sp. nov. is proposed.

Description of Roseovarius marisflavi sp. nov.

Roseovarius marisflavi (ma.ris.fla’vi. L. gen. neut. n. maris of the sea; L. masc. adj. flavus yellow; N.L. gen. masc. n. marisflavi of the Yellow Sea).

Cells are Gram stain-negative, strictly aerobic, ovoid or short rods, approximately 0.5–1.3 μm long by 0.4–0.8 μm wide, with no flagella. Cells do not form endospores. PHB is accumulated. Colonies are cream-coloured, uniformly round, 0.2–0.8 mm in diameter, regular, convex and smooth. The halotolerance range for growth is 1–15 % NaCl (optimum 2–7 %). The sea salts range for growth of strain H50T is 1–11 %. The temperature range for growth is 0–37 °C (optimum 28–37 °C). The pH range for growth is 6.0–8.0 (optimum 7.0). Cells contain Bchl a. Oxidase and catalase are positive. Tween 20 and DNA are degraded, but starch, Tween 80, gelatin, casein, alginate, lecithin or chitin is not degraded. Tween 40 is weakly degraded.

According to API 20NE tests: nitrate cannot be reduced to nitrite; indole is not produced; arginine dihydrogenase and urease are positive; aesculin ferric citrate is degraded; β-galactosidase is negative; d-glucose, l-arabinose, d-mannose, d-mannitol, N-acetyl-glucosamine, d-maltose, potassium gluconate, capric acid, adipic acid, malic acid, trisodium citrate and phenylacetic acid are not utilized. According to API 20E tests: lysine decarboxylase, ornithine decarboxylase, tryptophane desaminase and VP reaction are negative; H2S is not produced; acid is not produced from d-glucose, d-mannitol, inositol, d-sorbitol, l-rhamnose, d-saccharose, d-melibiose, amygdalin or l-arabinose. According to Biolog GN2 tests, d-arabitol, d-glucuronic acid, α-ketovaleric acid and d, l-lactic acid are oxidized.

C18:1ω7c and C16: 0 are the major components of the cellular lipids and other fatty acids present include C14:0, C18:0, C16:1ω7c/C16:1ω6c, C19:0 cycloω8c, C18:1 2-OH, C18:1ω7c 11-methyl. The following polar lipids are present: diphosphatidylglycerol, phosphatidylglycerol and a lipid. The predominant respiratory quinone is Q-10. The DNA G+C content of the type strain is 57.7 mol%.

The type strain, strain H50T (=CGMCC 1.10799T=JCM 17553T) was isolated from seawater collected from an amphioxus breeding zone in the coastal region of the Yellow Sea, China. The GenBank accession number for the 16S rRNA gene sequence of R. marisflavi H50T is KC900366.