Tetragenococcus koreensis sp. nov., a novel rhamnolipid-producing bacterium

The genus Tetragenococcus is classified within the family ‘Enterococcaceae’, which belongs to the order ‘Lactobacillales’ in the class ‘Bacilli’ (Garrity & Holt, 2001). The genus was created by Collins et al. (1990) after reclassification of the halophilic lactic acid bacterium [Pediococcus] halophilus as Tetragenococcus halophilus. At present, the genus comprises three species: T. halophilus, Tetragenococcus muriaticus (Satomi et al., 1997) and Tetragenococcus solitarius (Ennahar & Cai, 2005).

Kimchi is a common food in Korea. It is prepared with various kinds of vegetables, spices and other ingredients, and becomes palatable through ageing under anaerobic conditions. Kimchi fermentation is initiated by various micro-organisms present in the raw materials, but is gradually dominated by lactic acid bacteria.

Strain JS^T was isolated from a traditional Korean food, kimchi, by direct plating on glucose yeast peptone (GYP)/sodium acetate/mineral salts agar at pH 7·0 (medium 243; BCCM). After 7 days incubation on GYP/sodium acetate/mineral salts agar, colonies were smooth, low convex, circular, opaque, matt and white in colour. Strain JS^T grew at 15–30 °C but not at, or above, 4 °C or at temperatures above 42 °C. The strain grew optimally at pH 9. Optimal growth was obtained in 2–5 % (w/v) NaCl. Strain JS^T did not grow under aerobic conditions on R2A agar (Difco), nutrient agar, MacConkey agar (Difco) or trypticase soy agar (Difco) at 30 °C. For maintenance of the culture and further analyses, strain JS^T was grown on MRS agar (Difco) at 30 °C, unless specified otherwise.

Cell morphology and motility were observed under a Nikon Optiphot 111675 light microscope (1000× magnification). Cell morphology was also examined by using scanning electron microscopy. The procedures used for sample fixation and preparation were as described by Chung & Hancock (2000), with some modifications. Cells were non-spore-forming, non-motile cocci occurring singly or in pairs. The cell diameter was less than 1 μm (see the Supplementary Figure in IJSEM Online) on GYP/sodium acetate/mineral salts agar. The Gram reaction, determined by the non-staining method described by Buck (1982), was positive. Oxidase activity was tested using the oxidation of 1 % p-aminodimethylaniline oxalate. Catalase activity was determined using bubble production with 3 % (v/v) hydrogen peroxide solution. Strain JS^T showed no catalase or oxidase activities.

For 16S rRNA gene sequencing, cell biomass was obtained from cultures grown in MRS broth (Difco) on a horizontal shaker (150 r.p.m.) at 30 °C. The culture was checked for purity using a light microscope prior to harvesting by centrifugation. Chromosomal DNA was extracted and purified using the DNeasy Tissue Kit (Qiagen). The 16S rRNA gene was PCR-amplified and analysed as described by Yoon et al. (1998). Complete 16S rRNA gene sequences were compiled after sequencing using SeqMan software (DNASTAR), and 16S rRNA gene sequences of the test strains were edited using the BioEdit program (Hall, 1999) and aligned using clustal_x (Thompson et al., 1997). The phylogenetic tree was constructed by using the neighbour-joining method (Saitou & Nei, 1987) in mega version 2.1 (Kumar et al., 2001).

The almost-complete sequence determined for the 16S rRNA gene of strain JS^T was a continuous stretch of 1492 nt and there were no ambiguous bases. Analysis of 16S rRNA gene sequence similarities showed that strain JS^T was most closely related to T. halophilus IAM 1676^T (97·8 %), T. solitarius DSM 5634^T (97·2 %) and T. muriaticus JCM 10006^T (95·9 %). The neighbour-joining tree showed that strain JS^T and the three Tetragenococcus species formed a coherent cluster (supported by a bootstrap resampling value of 100 %) that was adjacent to genus Melissococcus (Fig. 1⇓). This tree topology of the genus Tetragenococcus was also found in the trees generated by the maximum-parsimony algorithms (Kimura two-parameter model; Kimura, 1983).

Acid production from carbohydrates was determined as described by Hugh & Leifson (1953). The API 20NE, API 32GN, API 50 CH and API 20 Strep microtest systems were used according to the recommendations of the manufacturer (bioMérieux). Strain JS^T did not reduce nitrate to nitrite or nitrogen gas. Strain JS^T had β-glucosidase and β-galactosidase activity but protein was not hydrolysed. The differential phenotypic properties of strain JS^T and other type strains of species in the genus Tetragenococcus are summarized in Table 1⇓.

For fatty acid methyl ester analysis, cell mass for all strains was obtained after 7 days incubation at 30 °C on GYP/sodium acetate/mineral salts agar. Fatty acid methyl esters were prepared, separated and identified using the Sherlock Microbial Identification System (MIS; MIDI) (Sasser, 1990). Strain JS^T had a cellular fatty acid profile containing large amounts of saturated, unsaturated and cyclo fatty acids. The fatty acid profile (Table 2⇓) of strain JS^T comprised the dominant fatty acids C_16 : 0 (22·43 %), C_18 : 1ω7c/ω9t/ω12t (33·8 %) and C_19 : 0 cyclo ω8c (26·8 %). No significant differences in the fatty acid profiles were found between strain JS^T and other species of the genus Tetragenococcus.

The peptidoglycan structure of the cell wall was determined as described by Schleifer (1985) and Schleifer & Kandler (1972), with the modification that TLC on cellulose was used instead of paper chromatography. The total hydrolysate of the peptidoglycan contained the amino acids alanine and glutamic acid plus smaller amounts of lysine and aspartic acid. In addition, the peptide l-lys–d-Asp was detected; it is stable under the conditions used for total hydrolysis (100 °C, 4 M HCl, 16 h). Partial hydrolysis (100 °C, 4 M HCl, 45 min) resulted in the peptides l-ala–d-glu and l-lys–d-Ala. From these data it was concluded that strain JS^T possessed peptidoglycan type A4α l-lys–d-Asp (type A11.31 according to ).

For determination of the DNA G+C content, genomic DNA was extracted and purified as indicated above for 16S rRNA gene sequence analysis, and was then enzymically degraded into nucleosides as described previously (Mesbah et al., 1989; Tamaoka & Komagata, 1984). The genomic DNA G+C content of strain JS^T was 38·3 mol%, which is a little higher than those of the other three species of the genus.

For DNA–DNA hybridizations, DNA was extracted as indicated above. Hybridizations were performed fluorometrically by the method of Ezaki et al. (1989), using photobiotin-labelled DNA probes and microdilution wells, with five replications for each sample. The highest and lowest values obtained for each sample were excluded; the remaining three values were used to calculate similarity values. The DNA relatedness values quoted are the means of these three values. Strain JS^T exhibited low levels of DNA–DNA hybridization: 9·7 % with T. halophilus IAM 1676^T, 61·0 % with T. solitarius DSM 5634^T and 39·0 % with T. muriaticus JCM 10006^T.

Lactic acid production was analysed by using HPLC (Acme HPLC; Younglin Instrument Co.). The HPLC analysis was performed with an HPLC system (processor, pump model SP930D, UV detector model UV 730D, column oven, auto-sampler) from Younglin Instrument Co. This was equipped with an Aminex HPX-87H organic acid analysis column from Bio-Rad. A 20 μl aliquot of the 10⁻³ dilution of strain JS^T culture broth was injected; 0·0035 M H₂SO₄ solution was used as eluent, at 30 °C and at a flow rate of 0·7 ml min⁻¹. The lactic acid in the culture broth was detected using a UV detector (210 nm) (Younglin Instrument Co.) and compared with a standard lactic acid from Aldrich, USA. Strain JS^T produced lactic acid homofermentatively.

For the rhamnolipid production assay, HPLC was carried out using a Waters Altantis dC₁₈ reversed-phase column (150×3·9 mm, 5 μm). An aliquot (20 μl) of the strain JS^T culture broth was injected; 100 % water was used as eluent at 30 °C with a flow rate of 0·6 ml min⁻¹. Products were detected using a UV detector (204 nm), then compared with a standard rhamnolipid biosurfactant from the Korea Research Institute of Bioscience and Biotechnology (Daejeon, Republic of Korea). The surface tension of the cell-free culture broth was determined with a Thermo Cahn tensiometer with the WinDCA32 software suite (Cahn Instruments) according to the De Nöuy ring method (Zajic & Seffens, 1984). To estimate the emulsification activity, 5 ml n-hexadecane was added to 5 ml cell-free culture broth in graduated test tubes. These tubes were then vortexed for 2 min and allowed to stand at 30 °C. The emulsion stability was measured after incubation for 24 h. The emulsification index (E₂₄ %) was calculated by dividing the height of the emulsion layer by the total height of the mixture and multiplying the quotient by 100. The use of biosurfactants generally results in foaming, a decrease in surface tension, and emulsification of hydrocarbon substrates. The growth of strain JS^T on 2–5 % (w/v) NaCl on GYP/sodium acetate/mineral salts agar for 7 days lowered the surface tension to 34·9 dyne cm⁻¹, and the emulsifying activity was 61·5 %. It has been reported that rhamnolipids were produced by Pseudomonas aeruginosa (Rahman et al., 2002) from various substrates, including industrial waste (Haba et al., 2000). They are as diverse, in terms of chemical structure and properties, as the producer micro-organisms (Kosaric, 1987). Biosurfactants are biodegradable, less toxic, less allergenic and therefore less detrimental to the environment.

The overall results of the present study allowed us to assign strain JS^T to a novel species, for which we propose the name Tetragenococcus koreensis sp. nov.