International Journal of Systematic and Evolutionary Microbiology (2002), 52, 647–655 DOI: 10.1099/ijs.0.02004-0

Leuconostoc ficulneum sp. nov., a novel bacterium isolated from a ripe fig, and reclassification of fructosus as Leuconostoc fructosum comb. nov.

1 Departamento de Andre! Antunes,1 Fred A. Rainey,2 M. Fernanda Nobre,1 Peter Schumann,3 Zoologia, Universidade de 1 4 4 Coimbra, 3004-517 Ana Margarida Ferreira, Ana Ramos, Helena Santos Coimbra, Portugal and Milton S. da Costa5 2 Department of Biological Sciences, Louisiana State University, Baton Rouge, Author for correspondence: Milton S. da Costa. Tel: j351 39 824024. Fax: j351 39 826798. LA 70803, USA e-mail: milton!ci.uc.pt 3 Deutsche Sammlung von Mikroorganismen und An isolate, designated strain FS-1T, was recovered from a ripe fig. Phylogenetic Zellkulturen, Mascheroder Weg 1b, 38124 analysis of the 16S rRNA genes and DNA–DNA reassociation values showed Braunschweig, Germany that the organism represented a novel species of the genus Leuconostoc 4 Instituto de Tecnologia closely related to Lactobacillus fructosus. The novel isolate could be Quı!mica e Biolo! gica, distinguished from the type strain of Lactobacillus fructosus by the fatty acid Universidade Nova de composition and several phenotypic and growth characteristics. In strain FS-1T, Lisboa, Rua da Quinta Grande 6, Apartado 127, 18:1 ∆9 (18:1ω9c) was present in relatively large amounts whilst, in 2780-156 Oeiras, Portugal Lactobacillus fructosus, this fatty acid was a minor component. Strain FS-1T and 5 Departamento de Lactobacillus fructosus produced acid in API 50CHL microtubes from glucose, Bioquı!mica, Universidade fructose and mannitol within 48 h, whereas only strain FS-1T also fermented de Coimbra, 3001-401 trehalose, gluconate, turanose and sucrose after 48 h. Other differences in acid Coimbra, Portugal production from carbohydrates also distinguished strain FS-1T from Lactobacillus fructosus. Both organisms were heterofermentative with fructose as a substrate and fermented glucose only in the presence of fructose, as determined by nuclear magnetic resonance studies. Strain FS-1T was - positive. On the basis of the phylogenetic analysis, DNA–DNA reassociation values, physiological and biochemical characteristics and fatty acid composition, the name Leuconostoc ficulneum is proposed for the novel species represented by strain FS-1T, and it is proposed that Lactobacillus fructosus be reclassified in the genus Leuconostoc as Leuconostoc fructosum comb. nov.

Keywords: Leuconostoc ficulneum sp. nov., lactic acid , Leuconostoc fructosum comb. nov., Lactobacillus fructosus

INTRODUCTION major or sole product of . While some species are pathogenic, or are part of the normal flora are non-spore-forming, usually of animals, lactic acid bacteria are renowned for their non-motile cocci, coccobacilli or rods that belong to importance in the fermentation of food and food the low-GjC branch of the Gram-positive Bacteria. products. The term ‘lactic acid bacteria’ is generally These organisms lack catalase, need a fermentable used to include the species of the genera Carnobac- carbohydrate for growth and produce lactic acid as the terium, Enterococcus, Lactobacillus, Lactococcus, Leu- conostoc, Oenococcus, , Streptococcus, ...... Tetragenococcus, Vagococcus and Weissella (Van- Abbreviation: NMR; nuclear magnetic resonance. damme et al., 1996). The GenBank/EMBL/DDBJ accession numbers for the 16S rRNA gene sequences described in this work are AF360736 (strain FS-1T), AF360737 Early classification relied heavily on morphology, and (Leuconostoc fructosum IFO 3516T) and AF360738 (Leuconostoc fallax DSM the leuconostocs were, for example, considered to be 20189T). more coccoid than rod-like, while the reverse was true

02004 # 2002 IUMS Printed in Great Britain 647 A. Antunes and others for the lactobacilli, although there is a considerable 20410T ) were obtained from the DSMZ and were used for overlap (Garvie, 1986). Other key characteristics of the comparative purposes. All strains were cultured in MRS leuconostocs, such as the absence of arginine dehydro- medium at 30 mC and were maintained at k70 mC in the lase, the production of (k)-lactate from glucose and same medium supplemented with 15% glycerol. even peptidoglycan composition, were not exclusive, Morphological, physiological and biochemical tests. All tests but were shared with atypical heterofermentative were performed at 30 mC unless otherwise stated. Mor- lactobacilli such as Lactobacillus fructosus, Lacto- phology, the Gram reaction and the presence of cytochrome bacillus minor and Lactobacillus viridescens (Kandler oxidase and catalase were determined after 24 h incubation & Weiss, 1986; Garvie, 1986). The analysis of small- on MRS agar as described by Smibert & Krieg (1981). subunit rRNA gene sequence data has facilitated, in Growth at different temperatures was determined in MRS recent years, the proposal of several new genera of medium after incubation at temperatures ranging from 6 to 45 mC for 5 days. Growth at different pH values was also lactic acid bacteria to include species that were observed, in the same medium, after the addition of 100 mM previously included in other polyphyletic genera. MES for pH 4n0–6n5 and 100 mM HEPES for pH 7n0–8n0. Phylogenetic studies showed that the leuconostocs did The pH of each buffer was adjusted with HCl or NaOH. not form a homogeneous group, which led to division Control media, containing each buffer adjusted to pH 7n0, into three different genera corresponding to three were used to assess possible inhibitory effects of the buffering distinct clusters of the Leuconostoc line of descent. The agents. Growth in the presence of 2, 4, 6, 7, 8 and 10% (w\v) species of the genus Weissella (Weissella paramesen- NaCl and in medium containing 30, 40 and 50% (w\v) teroides, Weissella confusa, Weissella kandleri, Weis- glucose was examined in MRS broth, supplemented with sella minor and Weissella viridescens) were formerly 2% fructose, for 5 days. The ability to hydrolyse elastin, ascribed to the genera Leuconostoc and Lactobacillus fibrin, gelatin, casein, starch and aesculin was assessed on (Collins et al., 1993). The Leuconostoc cluster sensu MRS agar, as described by Hudson et al. (1986) and Smibert & Krieg (1981), for up to 6 days The fermentation of stricto now includes Leuconostoc mesenteroides and carbohydrates was determined with the API 50CHL system closely related non-acidophilic leuconostocs (Collins (bioMe! rieux) according to the manufacturer’s instructions. et al., 1991; Dicks et al., 1993; Bjo$ rkroth et al., 2000; Results were recorded after 24 h, 48 h and 5 days incubation Kim et al., 2000). This cluster also includes the at 30 mC. The ability of the strains to ferment several peripheral species Lactobacillus fructosus (Kodama, carbohydrates was also assessed with a medium com- " 1956; Skerman et al., 1980) and Leuconostoc fallax posed of (l− ): 10n0 g proteose peptone no. 3 (Difco), 5n0g (Collins et al., 1991; Martinez-Murcia & Collins, yeast extract (Difco), 1n0 ml Tween 80, 2n0g K#HPO%, 1991). The species Leuconostoc oenos (Garvie, 1967) 5n0 g sodium acetate, 2n0 g ammonium citrate, 0n2g was also shown to be phylogenetically, as well as MgSO%;7H#O, 0n05 g MnSO%;H#O and 0n17 g bromcresol phenotypically, distinct, warranting a different generic purple (all from Sigma) at pH 6n5. This basal medium (5 ml) assignment as Oenococcus oeni (Dicks et al., 1995). was added to metal-capped tubes (15i160 mm) and steri- lized by autoclaving. Filter-sterilized carbohydrates were " During a search for aerobic bacteria inhabiting sugar- added at a final concentration of 20n0gl− to the cooled basal rich environments, we isolated a lactic acid bacterium medium. The configuration of lactic acid formed by the that grew on media containing 30% glucose and which fermentation of glucose and fructose was tested enzymically using -lactate and -lactate dehydrogenases (Boehringer was closely related to Lactobacillus fructosus. On the " Mannheim) after growth in a medium composed of (l− ): basis of phenotypic and genotypic analysis of this 5 0 g casitone (Difco), 4 0 g yeast extract (Difco), 7 0g organism, we propose the name Leuconostoc ficulneum n n n T fructose, 2n3 g glucose, 0n34 g sodium citrate, 0n6gKH#PO%, sp. nov. for the species represented by strain FS-1 .We 0n45 g KCl, 0n13 g CaCl#,3n3 mg MnSO%;H#O and 0n265 g also propose that Lactobacillus fructosus should be MgSO%;7H#O (pH 6n5). The presence of arginine dehydro- reclassified as Leuconostoc fructosum comb. nov. lase was determined in MRS broth without meat extract but containing 0n05% glucose, 0n3% arginine and 0n2% sodium citrate (Shaw & Harding, 1984). Production of dextran from METHODS sucrose was assessed on MRS agar in which glucose was Bacterial strains and growth conditions. Strain FS-1T was replaced by 5% sucrose (Hitchener et al., 1982). isolated from a ripe fig. The fig was cut open and part of the Nuclear magnetic resonance (NMR) spectroscopy of the interior was removed and blended in a mortar. Drops of the glycolytic pathways and fermentation products. Strain FS- blend were inoculated into 25 ml Thermus medium (Williams 1T, Lactobacillus fructosus and Leuconostoc mesenteroides & da Costa, 1992) supplemented with 30% (w\v) glucose were grown with aeration (180 r.p.m.) at 30 mC in MRS and then incubated with shaking for 48 h at 30 mC. Turbid medium supplemented with 2% (w\v) glucose plus 2% cultures were streaked and colonies were then isolated and (w\v) fructose, harvested in the late-exponential growth purified on plates of the same medium. Culturing in phase and washed twice with 5 mM potassium phosphate Mann–Rogosa–Sharpe (MRS) medium (Difco) was later buffer (pH 6n5). The cell pellets were resuspended in 50 mM adopted and resulted in much higher biomass yields, potassium phosphate (pH 6 5) to approximately 30 mg dry " n especially when the medium was supplemented with 2% weight ml− . fructose. The type strains of Lactobacillus fructosus (DSM 20349T ), Leuconostoc fallax (DSM 20189T), Leuconostoc In vivo NMR experiments were performed in a Bruker mesenteroides subsp. mesenteroides (DSM 20343T ), Leu- DRX500 spectrometer as described previously (Ramos & "$ conostoc mesenteroides subsp. dextranicum (DSM 20484T), Santos, 1996), using [2- C]fructose (99% isotopically en- "$ Leuconostoc pseudomesenteroides (DSM 20193T), W. par- riched; Eurisotop), [3- C]fructose (99% isotopically en- "$ amesenteroides (DSM 20288T ) and W. viridescens (DSM riched; Omicron Biochemicals) and [1- C]glucose (99%

648 International Journal of Systematic and Evolutionary Microbiology 52 Leuconostoc ficulneum sp. nov. isotopically enriched; Omicron Biochemicals) as labelled (De Ley et al., 1970). Each hybridization experiment was substrates for fermentation. Lactate and acetate were quan- executed at least twice. " tified in supernatant solutions by H-NMR as described previously (Neves et al., 1999). Mannitol and erythritol were 16S rRNA gene sequence determination and phylogenetic "$ analyses. Extraction of genomic DNA, PCR amplification quantified by C-NMR in a Bruker DRX500 spectrometer of the 16S rRNA gene and sequencing of purified PCR with a 13 µs pulse width (60 flip angle) and a repetition "$ m products were carried out as described previously (Rainey et delay of 60 5s. C-labelled methanol (Sigma) was added as n al., 1996). Purified sequence reaction products were electro- a concentration standard. The end-products from fructose phoresed with a model 310 Genetic Analyzer (Applied catabolism were also quantified by HPLC as described by Biosystems). The 16S rRNA gene sequences obtained in this Starrenburg & Hugenholtz (1991). study were aligned against previously determined bacterial Cell-wall peptidoglycan type. Purified cell-wall preparations sequences by using the ae2 editor (Maidak et al., 1999). The were obtained by the method of Schleifer & Kandler (1972). method of Jukes & Cantor (1969) was used to calculate The amino acids and peptides in cell-wall hydrolysates were evolutionary distances, and phylogenetic dendrograms were analysed by two-dimensional ascending TLC on cellulose generated using various treeing algorithms contained in the plates by using systems described previously (Schleifer &  package (Felsenstein, 1993). Kandler, 1972). The amino-terminal amino acid of the interpeptide bridge was determined by dinitrophenylation as described by Schleifer (1985). RESULTS Polar-lipid and fatty-acid composition. Cultures used for Isolation and morphological and biochemical T polar-lipid analysis were grown in 1 l Erlenmeyer flasks characteristics of strain FS-1 containing 200 ml MRS medium at 30 mC in a reciprocal water-bath shaker until the exponential phase of growth. We isolated several bacteria from sugar-rich sources Harvesting of the cultures and the extraction of lipids were that included home-made jams and jellies, honey and performed as described previously (Donato et al., 1990). The ripe fruits. Most of the organisms were low-GjC individual polar lipids were separated by one-dimensional Gram-positive bacteria that, on the basis of partial 16S TLC on silica gel G plates (0n25 mm thickness; Merck) with rRNA gene sequence analysis, were closely related to a solvent system consisting of chloroform\methanol\acetic Bacillus licheniformis. One strain was found to be a acid\water (80:12:15:4, by vol.). member of the Clavibacter–Curtobacterium lineage of the . An additional isolate, designated Cultures for fatty-acid analysis were grown on MRS- T medium plates, in sealed plastic bags submerged in a water FS-1 , was recovered from a ripe fig. This organism bath at 28 mC for 24 h. Fatty acid methyl esters were obtained clustered with the species of the genus Leuconostoc and from fresh, wet biomass by saponification, methylation and was most closely related to Lactobacillus fructosus. extraction as described previously (Kuykendall et al., 1988). Colonies of FS-1T on MRS were up to 1 mm in The fatty acid methyl esters were separated using a Hewlett Packard model 5890 GC equipped with a flame-ionization diameter, smooth and convex with a grey-white color- detector fitted with a 5% phenylmethyl silicone capillary ation. The cells were approximately 0n4–1n0 µmin width and 1n8–2n9 µm in length and stained Gram- column (0n2mmi25 m; Hewlett Packard). The carrier gas T was high-purity H , the column head pressure was 60 kPa, positive. Strain FS-1 and the type strain of Lacto- # " the septum purge was 5 ml min− , the column split ratio was bacillus fructosus had an optimum growth temperature 55:1 and the injection port temperature was 300 mC. The of about 30 mC but, unlike Lactobacillus fructosus, T temperature of the oven was programmed to go from 170 to strain FS-1 did not grow at 6 or 40 mC. The optimum −" T 270 mC at a rate of 5 mC min . Identification and quantifica- pH for growth of strain FS-1 was between 6n5 and 7n0; tion of the fatty acid methyl esters and numerical analysis of no growth was observed at pH 4n5. The specific growth the fatty acid profiles were performed by using the standard rates of the two strains were highest in MRS medium MIS library generation software (Microbial ID). Double- without added NaCl, but both organisms were haloto- bond positions of unsaturated fatty acids were determined lerant; Lactobacillus fructosus grew in medium con- by GC-MS of their dimethyl disulphide adducts (Nichols et T al., 1986) by using the GC and MS conditions described taining 8n0% NaCl, while strain FS-1 grew in medium previously (Stro$ mpl et al., 1999). containing 7n0% NaCl. Both organisms grew in medium containing 40% glucose, but did not grow in Determination of mean base composition of DNA and medium containing 50% glucose. Cytochrome oxidase DNA–DNA reassociation studies. The DNA used to de- and arginine dehydrolase tests gave negative results for termine DNA base compositions was isolated as described both strains. Catalase activity was detected in strain by Cashion et al. (1987). The GjC content of the overall T genome was determined by HPLC as described by Mesbah FS-1 , but was not detected in any of the other et al. (1989). The DNA for DNA–DNA reassociation studies organisms examined. was extracted and purified by the procedure of Marmur One of the most interesting characteristics of strain (1961). The degree of DNA reassociation was determined FS-1T and Lactobacillus fructosus was their very spectrophotometrically from initial renaturation rates ac- cording to De Ley et al. (1970). The renaturation rates were restricted fermentation patterns; both strains pro- measured in 1iSSC (0n15 M NaCl and 0n015 M trisodium duced acid from the fermentation of glucose, fructose and mannitol in API 50CHL microtubes within citrate at pH 7n0) using an Uvikon 940 spectrophotometer T (Kontron) equipped with a thermostat-controlled cuvette 24–48 h (Table 1). Strain FS-1 fermented mannitol, chamber. The optimal renaturation temperature used in trehalose and gluconate after 48 h (delayed), while a each case was calculated from the GjC content (mol%) weakly positive reaction was observed with sucrose http://ijs.sgmjournals.org 649 A. Antunes and others

Table 1. Differential carbohydrate fermentation patterns of strain FS-1T and related species ...... Taxa are identified as: 1, FS-1T;2,Lactobacillus fructosus;3,Leuconostoc fallax;4,Leuconostoc pseudomesenteroides;5,Leuconostoc mesenteroides subsp. mesenteroides;6,Leuconostoc mesenteroides subsp. dextranicum;7,W. viridescens;8,W. paramesenteroides. Results were obtained with the API 50CHL system. Results are scored as: j, positive; k, negative; (d), delayed; , weakly positive; , not determined. All strains fermented -glucose and -fructose. None of the strains fermented glycerol, erythritol, -arabinose, -xylose, ribitol, methyl β-xyloside, -sorbose, rhamnose, galactitol, inositol, methyl α--mannoside, inulin, melezitose, glycogen, xylitol, -lyxose, -tagatose, -fucose, -fucose, -arabitol, -arabitol, 2-ketogluconate (weakly positive for Leuconostoc pseudomesenteroides) or 5-ketogluconate (weakly positive for Leuconostoc pseudomesenteroides).

Substrate 1 2 3 4 5 6 7 8

Methyl α--glucoside kkj(d) jkk β-Gentiobiose kkkj  kkk Cellobiose kkkjjkkk -Mannose kkj(d) jj j -Raffinose kkkjjkkk -Turanose  kjjjkk(d) -Xylose kkkj kkk Aesculin kkjjjkkk Galactose kkkjjkk(d) Gluconate (d) kkkkkk Lactose kkkk(d) kk  -Arabinose kkkjjkkj Maltose kkkjjk(d) j Mannitol (d) j (d) k  kkk Melibiose kkkjjkk N-Acetylglucosamine kkjjjjkj Ribose kkkjjkk(d) Starch kkkkkkk(d) Sucrose  kjjj kj Trehalose (d) kkjj kj

and turanose. With the exception of W. viridescens, (69n8 p.p.m.); resonances due to fructose 6-phosphate which appeared to ferment only glucose in under 48 h (75n1 p.p.m.) and mannitol 1-phosphate (70n3 p.p.m.), and fructose, mannose and maltose after 72 h, all other intracellular intermediates in the biosynthesis of man- organisms examined fermented several carbohydrates. nitol, were detected transiently during fructose utiliza- That the fermentation of glucose was reported as tion. In addition to mannitol, acetate, lactate and negative or delayed in Lactobacillus fructosus by smaller amounts of erythritol were produced from the "$ Kodama (1963) prompted us to confirm the API metabolism of [3- C]fructose by strain FS-1T, Lacto- 50CHL results by using a conventional tube-fermen- bacillus fructosus and Leuconostoc mesenteroides "$ " tation method and in vivo NMR (see below), the results (Table 2). Analysis of end-products by C- and H- of which showed that glucose was not fermented in the NMR showed that acetate was labelled on the COOH tubes within 5 days and was not degraded during the group, as expected from the operation of the heterofer- period of the in vivo NMR experiment. The lactate mentative pathway for the dissimilation of fructose. isomer determined enzymically was (k). Mannitol was labelled on the inner CHOH group, in accordance with a catabolic pathway involving fruc- Sugar metabolism and fermentation products tose 6-phosphate and mannitol 1-phosphate as inter- mediates. In addition to unlabelled lactate derived "$ The metabolism of [3- C]fructose by strain FS-1T was from the heterofermentative pathway, a small amount monitored in vivo by NMR (Fig. 1). The resonances of lactate labelled on the COOH group was also due to the furanose form (α and β anomers at 82n6 and detected, indicating that fructose was partially meta- 76 5 p.p.m., respectively) and the β-anomer of the bolized through the Embden–Meyerhof–Parnas path- n "$ pyranose form (68n2 p.p.m.) of [3- C]fructose were way. Furthermore, a small fraction of the fructose was seen clearly in the first spectrum of the sequence. The metabolized via an erythritol-forming pathway, which main end-product of fructose catabolism was mannitol was described previously for O. oeni (Veiga-da-Cunha

650 International Journal of Systematic and Evolutionary Microbiology 52 Leuconostoc ficulneum sp. nov.

Mannitol

8·5

7·5

6·5

5·5

4·5 Time (min)

3·5 ...... Fig. 1. Time course for the utilization of 2·5 17 mM [3-13C]fructose at 30 mC by a cell suspension of strain FS-1T, as monitored by 1·5 in vivo 13C-NMR. Each spectrum represents 30 s accumulation. Peaks are identified as: 85 80 75 70 65 , [3-13C]fructose; *, β-[3-13C]fructose 6- Chemical shift (p.p.m.) phosphate; >, [4-13C]mannitol 1-phosphate.

Table 2. End-products from the metabolism of fructose ...... Values are the means of at least two independent determinations and are accurate within 8% of each of the values.

Organism Product (mM)

Mannitol Acetate Lactate Erythritol

FS-1T 9n35n45n10n1 Lactobacillus fructosus 10n96n36n20n1 Leuconostoc mesenteroides subsp. mesenteroides 9n85n05n30n3

et al., 1992). The contribution of the different pathways type strain of Leuconostoc fallax possessed an ad- was very similar in the three organisms studied: ditional aminophospholipid that was not found in any approximately 60% of the fructose was converted to of the other organisms examined (results not shown). mannitol, 30% was metabolized by the heterofermen- However, the organisms could be distinguished from tative pathway, 5% by the Embden–Meyerhof–Parnas each other by their fatty acid compositions; the major pathway and 0 5–2% of the fructose produced eryth- fatty acid of strain FS-1T and Lactobacillus fructosus "$ n ritol. [1- C]Glucose as a sole carbon source was not could not be identified by GC because two fatty acids, metabolized by strain FS-1T or Lactobacillus fructosus namely 18:1 ∆6 (18:1ω12) and 18:1 ∆11 (18:1ω7), under the experimental conditions used; interestingly, had the same retention time. However, it was possible glucose was fermented after fructose was added to the to identify the major fatty acid of strain FS-1T and cell suspensions (results not shown). Lactobacillus fructosus as 18:1 ∆11 (18:1ω7c) by GC- MS. The double-bond position of hexadecenoic acid in T Polar lipid, fatty acid and peptidoglycan composition strain FS-1 was found to be at carbon atom 9 (16:1 ∆9). Strain FS-1T could be distinguished from Lacto- The polar lipid compositions of all of the organisms bacillus fructosus because the former had high levels of examined were very similar and were dominated by 16:0 and minor amounts of 18:1 ∆9 (18:1ω9c), while three phospholipids and four glycolipids, though the 18:1 ∆9(18:1ω9c) was the second most prevalent fatty http://ijs.sgmjournals.org 651 A. Antunes and others

Table 3. Mean fatty acid composition of strains examined after growth at 28 mC ...... Taxa are identified as: 1, Leuconostoc fallax DSM 20189T;2,Leuconostoc mesenteroides subsp. dextranicum DSM 20484T;3,Leuconostoc mesenteroides subsp. mesenteroides DSM 20343T;4, Leuconostoc pseudomesenteroides DSM 20193T;5,Leuconostoc fructosus DSM 20349T; 6, FS-1T (l DSM 13613T); 7, W. paramesenteroides DSM 20288T;8,W. viridescens DSM 20410T. Values are percentages of total fatty acids.

Fatty acid 12345678

14:0 17n76n718n74n80n60n218n12n3 16:1ω7c 9n713n89n510n212n07n56n48n5 16:0 32n421n422n423n212n018n421n518n4 17:0 cyclo 1n22n3– 1n2– – 0n3– 18:1ω9c 9n210n437n69n420n98n028n535n9 18:1ω7c 4n45n18n57n253n363n43n918n7 18:0 0n60n7– 0n90n40n61n35n1 19:0 cyclo ω10c 6n2– 1n7–––19n3– 19:0 cyclo ω8c 17n539n31n642n4–––– 20:2ω6,9c 0n60n4– 0n8–––– 20:1ω9c –––––––9n6 Unknown –––––––0n8

...... Fig. 2. 16S rRNA gene sequence-based neighbour-joining phylogenetic tree showing the position of the novel isolate within the radiation of representative species of the genera Leuconostoc, Weissella and Oenococcus. Bar, 10 inferred nucleotide substitutions per 100 nucleotides. Lb., Lactobacillus; Ln., Leuconostoc. acid in Lactobacillus fructosus. In all of the other 16S rRNA gene sequence comparison, mean base organisms, 18:1ω9c was more abundant than 18:1ω7c composition of DNA and DNA–DNA reassociation (Table 3). studies The peptidoglycan of strain FS-1T contained lysine, Partial 16S rRNA gene sequences comprising 1457– alanine and glutamic acid. Alanine was the amino- 1491 nucleotides were determined for strain FS-1T and terminal amino acid of the interpeptide bridge. From the type strains of Lactobacillus fructosus and Leu- these results and from two-dimensional TLC peptide conostoc fallax. Comparison of these sequences with patterns of partial hydrolysates of cell walls (data representatives of the main lineages within the Bacteria T not shown), it was concluded that strain FS-1 has indicated that this strain fell within the low-GjC peptidoglycan type A3α (Schleifer & Kandler, 1972), Gram-positive phylum and, more specifically, within variation -Lys " -Ala. the radiation of the lactic acid bacteria. The highest

652 International Journal of Systematic and Evolutionary Microbiology 52 Leuconostoc ficulneum sp. nov.

16S rRNA gene sequence similarities were found to with the tube-fermentation method or during in vivo members of the genus Leuconostoc (92n2–94n5%), NMR experiments. The discrepancy in the results while much lower similarities, in the range 85n2–88n6%, of glucose fermentation obtained between the API were found to species of the genera Weissella, Lacto- 50CHL system on the one hand and conventional bacillus and Oenococcus. A phylogenetic tree was fermentation tests and in vivo NMR on the other is not reconstructed that included the 16S rRNA gene easily resolved, but the seemingly contradictory results sequence from each of the type strains of the 10 may be due to different conditions for fermentation. It Leuconostoc species as well as outgroup organisms is noteworthy that both organisms fermented glucose from the genera Weissella, Lactobacillus and Oeno- in the presence of fructose; a similar observation was (Fig. 2). Strain FS-1T clustered with Lacto- made by Kodama (1963) for Lactobacillus fructosus. bacillus fructosus and they shared 94n5% 16S rRNA This metabolic peculiarity could be explained by low gene sequence similarity. The clustering of FS-1T with activity of the dehydrogenases, leading to the pro- both Lactobacillus fructosus and the Leuconostoc duction of ethanol in these organisms; the consequent species is supported by high bootstrap values (Fig. 2). ‘bottleneck’ in the regeneration of NAD(P) would be The DNA GjC content was 42n6 mol% for strain FS- overcome in the presence of fructose by the operation T 1 and 43n4 mol% for Lactobacillus fructosus. The of mannitol-1-phosphate dehydrogenase. T distinct species status of strain FS-1 was demon- T strated by a mean DNA–DNA reassociation value of Lactobacillus fructosus and strain FS-1 clearly rep- 46n0% with the type strain of Lactobacillus fructosus. resent two closely related, but distinct, genomic spe- cies, in view of the low DNA–DNA reassociation values (46%) and the fact that they branch together in DISCUSSION the phylogenetic analysis. The novel species repre- sented by strain FS-1T can be distinguished from A lactic acid bacterium, belonging to a novel species of Lactobacillus fructosus and from other species of the the genus Leuconostoc, was isolated from a ripe fig genera Leuconostoc and Weissella by fatty acid com- during a search for xerophilic bacteria in sugar-rich T position, the fermentation of carbohydrates, differ- environments. The organism, designated FS-1 , was, ences in the growth-temperature range and haloto- on the basis of phylogenetic analysis of the 16S rRNA lerance. gene sequence, closely related to the species Lacto- T bacillus fructosus, a misclassified lactic acid bacterium Strain FS-1 and Lactobacillus fructosus share some that clearly falls within the genus Leuconostoc. In fact, characteristics that distinguish them from the remain- strain FS-1T shared a large number of biochemical and ing species of the genus Leuconostoc, and the 16S rDNA sequence analysis also seems to indicate that physiological characteristics with Lactobacillus fruc- T tosus, but could be distinguished from this species strain FS-1 , Lactobacillus fructosus and Leuconostoc primarily by differences in fatty acid composition, as fallax could be assigned to novel genera. However, the well as differences in fermentation patterns and growth absence of noteworthy chemotaxonomic, biochemical characteristics. Since only one strain of Lactobacillus or physiological differences argues against novel fructosus is known and we could only isolate one strain generic assignments for these organisms. Moreover, of the novel species, the different fatty acid compo- DNA–rRNA studies by Schillinger et al. (1989) reveal- sitions could be attributed to strain-specific variations ed a high degree of rRNA gene similarity between and could have no taxonomic significance. However, Lactobacillus fructosus and species of the Leuconostoc other characteristics (differences in growth-tempera- sensu stricto cluster, while Lactobacillus fructosus was ture range, halotolerance and delayed or weak fer- also found to be a member of this cluster by Collins et mentation of turanose, gluconate, sucrose and tre- al. (1991). halose in API 50CHL microtubes) distinguish strain On the basis of these findings, we propose that strain FS-1T Lactobacillus fructosus T from . Strain FS-1 is FS-1T represents a novel species of the genus Leucono- unusual in that it possesses catalase (or pseudocata- stoc, which we have named Leuconostoc ficulneum sp. lase) activity, the absence of which is one of the nov. We also propose that, on the basis of phylogenetic hallmarks of lactic acid bacteria. However, some lactic criteria and physiological, biochemical and chemo- acid bacteria have been shown to possess a catalase- taxonomic characteristics, Lactobacillus fructosus like activity (Kono & Fridovich, 1983). Unfortunately, should be assigned to the genus Leuconostoc as there appears to be no information available indicating Leuconostoc fructosum comb. nov. the taxonomic significance of catalase-type activity in some lactic acid bacteria, and we must assume that the presence of this activity cannot be considered as a Description of Leuconostoc ficulneum sp. nov. taxonomically useful characteristic for defining a species among these organisms. Leuconostoc ficulneum (fi.culhne.um. L. adj. ficulneum pertaining to figs and fig trees). Both species also possess very unusual patterns of carbohydrate fermentation, since only a few carbohy- Short, rod-shaped cells, 0n4–1n0 µm in width by 1n8– drates are fermented, among them fructose and man- 2n9 µm in length. The Gram stain is positive; cells are nitol. The fermentation of glucose was not observed not motile and spores are not formed. Colonies are http://ijs.sgmjournals.org 653 A. Antunes and others

T T small, smooth, round and grey-white. The optimum 13162 l NCIB 10784 ), was isolated from flowers in temperature for growth is approximately 30 mC; no Japan. growth occurs at 6 or 40 mC. The optimum pH is between 6 5 and 7 0; no growth occurs at pH 4 5. The n n n ACKNOWLEDGEMENTS maximum NaCl concentration for growth is 7n0% and the maximum glucose concentration for growth is This work was supported, in part, by the PRAXIS XXI 40%. The cell-wall murein is type A3α, variation - programme (PRAXIS\PCNA\BIO\46\96). We would like Lys " -Ala. The predominant cellular fatty acids are to thank Professor Hans Tru$ per for the etymology of the hexadecanoic acid (16:0) and 11,12-octadecenoic acid epithet ficulneum and Cristina Leita4 o (ITQB, Oeiras, Por- (18:1 ∆11, 18:1ω7c). Catalase activity is present but tugal) for the HPLC determinations. The authors are cytochrome oxidase is not. Arginine dehydrolase is grateful to Dr Brian J. Tindall (DSMZ) and Dr M. Nimtz (GBF, Braunschweig, Germany) for their help with GC-MS absent. Does not hydrolyse elastin, fibrin, gelatin, analysis of fatty acid derivatives. casein, starch or aesculin. Does not produce dextran from sucrose. Heterofermentative; a small amount of fructose (5%) is fermented via the Embden– REFERENCES $ Meyerhof–Parnas pathway. The major fermentation Bjorkroth, K. J., Geisen, R., Schillinger, U., Weiss, N., De Vos, P., products of fructose are mannitol, lactate and acetate. Holzapfel, W. H., Korkeala, H. J. & Vandamme, P. (2000). Charac- The lactate isomer is (k). Acid is produced from - terization of Leuconostoc gasicomitatum sp. nov., associated with glucose and -fructose in API 50CHL microtubes, the spoiled raw tomato-marinated broiler meat strips packaged under modified-atmosphere conditions. Appl Environ Microbiol 66, 3764– production of acid from gluconate, mannitol and 3772. trehalose is delayed and the fermentation of sucrose Cashion, P., Holder-Franklin, M. A., McCully, J. & Franklin, M. and -turanose is weak. Acid is not produced from (1987). A rapid method for the base ratio determination of bacterial other carbohydrates tested (Table 1). The GjC DNA. Anal Biochem 81, 461–466. T content of the DNA of strain FS-1 is 42n6 mol% (as Collins, M. D., Rodrigues, U., Ash, C., Aguirre, M., Farrow, J. A. E., determined by the HPLC method). Martinez-Murcia, A., Phillips, B. A., Williams, A. M. & Wallbanks, T T S. (1991). Phylogenetic analysis of the genus Lactobacillus and related The type strain, FS-1 (l DSM 13613 l NRRL B- lactic acid bacteria as determined by reverse transcriptase sequencing of 23447T), was isolated from a ripe fig bought in 16S rRNA. FEMS Microbiol Lett 77, 5–12. Portugal. Collins, M. D., Samelis, J., Metaxopoulos, J. & Wallbanks, S. (1993). Taxonomic studies on some leuconostoc-like organisms from fermented sausages: description of a new genus Weissella for the Description of Leuconostoc fructosum comb. nov. Leuconostoc paramesenteroides group of species. J Appl Bacteriol 75, 595–603. Basonym: Lactobacillus fructosus Kodama 1956. De Ley, J., Cattoir, H. & Reynaerts, A. (1970). The quantitative measurement of DNA hybridization from renaturation rates. Eur J The description is based on the results of Kodama Biochem 12, 133–142. (1956, 1963) and Schillinger et al. (1989) and on those Dicks, L. M. T., Fantuzzi, L., Gonzalez, F. C., du Toit, M. & Dellaglio, presented in this study. Forms short, rod-shaped cells, F. (1993). Leuconostoc argentinum sp. nov., isolated from Argentine 0n5–0n8 µm in width by 2n0–4n0 µm in length. The Gram raw . Int J Syst Bacteriol 43, 347–351. stain is positive; cells are not motile and spores are not Dicks, L. M. T., Dellaglio, F. & Collins, M. D. (1995). Proposal to formed. Colonies are small, smooth, round and grey- reclassify Leuconostoc oenos as Oenococcus oeni [corrig.] gen. nov., white. The optimum temperature for growth is ap- comb. nov. Int J Syst Bacteriol 45, 395–397. proximately 30 C; growth occurs at 6 and at 40 C, Donato, M. M., Seleiro, E. A. & da Costa, M. S. (1990). Polar lipid m m and fatty acid composition of strains of the genus Thermus. Syst Appl but no growth occurs at 45 mC. The maximum NaCl Microbiol 13, 234–239. concentration for growth is 8n0% and the maximum Felsenstein, J. (1993).  (phylogenetic inference package) version glucose concentration for growth is 40%. The cell-wall 3.5.1. Seattle: University of Washington. murein is type A3α, variation -Lys " -Ala. The Garvie, E. I. (1967). Leuconostoc oenos sp. nov. J Gen Microbiol 48, predominant fatty acids are 9,10-octadecenoic acid 431–438. (18:1∆9, 18:1ω9c) and 11,12-octadecenoic acid (18:1 Garvie, E. I. (1986). Genus Leuconostoc.InBergey’s Manual of ∆11, 18:1ω7c). Catalase activity is not detected and Systematic Bacteriology, vol. 2, pp. 1071–1075. Edited by P. H. A. cytochrome oxidase is negative. Arginine dehydrolase Sneath, N. S. Mair, M. E. Sharpe & J. G. Holt. Baltimore: Williams & is absent. Does not hydrolyse elastin, fibrin, gelatin, Wilkins. casein, starch or aesculin. Does not produce dextran Hitchener, B. J., Egan, A. F. & Rogers, P. J. (1982). Characteristics of lactic acid bacteria isolated from vacuum-packaged beef. J Appl from sucrose. Heterofermentative; a small proportion Bacteriol 52, 31–37. of the fructose is metabolized via the Embden– Hudson, J. A., Morgan, H. W. & Daniel, R. M. (1986). A numerical Meyerhof–Parnas pathway. The lactate isomer is classification of some Thermus isolates. J Gen Microbiol 132, 531–540. (k). Acid is produced from -glucose, -fructose Jukes, T. H. & Cantor, C. R. (1969). Evolution of protein molecules. In and mannitol in API 50CHL microtubes. Acid is not Mammalian Protein Metabolism, pp. 21–132. Edited by H. N. Munro. produced from other carbohydrates tested (Table 1). New York: Academic Press. The GjC content of the DNA of the type strain is Kandler, O. & Weiss, N. (1986). Genus Lactobacillus.InBergey’s 43n4 mol% (as determined by HPLC). Manual of Systematic Bacteriology, vol. 2, pp. 1209–1234. Edited by T T P. H. A. Sneath, N. S. Mair, M. E. Sharpe & J. G. Holt. Baltimore: The type strain, IFO 3516 (l DSM 20349 l ATCC Williams & Wilkins.

654 International Journal of Systematic and Evolutionary Microbiology 52 Leuconostoc ficulneum sp. nov.

Kim, J., Chun, J. & Han, H.-U. (2000). Leuconostoc kimchii sp. nov., a Schillinger, U., Holzapfel, W. & Kandler, O. (1989). Nucleic acid new species from kimchi. Int J Syst Evol Microbiol 50, 1915–1919. hybridization studies on Leuconostoc and heterofermentative lacto- Kodama, R. (1956). Studies on the nutrition of lactic acid bacteria. Part and description of Leuconostoc amelibioum sp. nov. Syst Appl IV. Lactobacillus fructosus nov. sp., a new species of lactic acid bacteria. Microbiol 12, 48–55. J Agric Chem Soc Jpn 30, 705–708 (in Japanese). Schleifer, K. H. (1985). Analysis of the chemical composition and Kodama, R. (1963). The nutrition of Lactobacillus fructosus and its primary structure of murein. Methods Microbiol 18, 123–156. application to microbiological determination of nicotinamide and Schleifer, K. H. & Kandler, O. (1972). Peptidoglycan types of bacterial fructose. Annu Rep Inst Ferment Osaka 1, 11–24 (in Japanese). cell walls and their taxonomic implications. Bacteriol Rev 36, 407–477. Kono, Y. & Fridovich, I. (1983). Isolation and characterization of the Shaw, B. G. & Harding, C. D. (1984). A numerical taxonomic study of pseudocatalase of Lactobacillus plantarum. J Biol Chem 258, 6015–6019. lactic acid bacteria from vacuum-packed beef, pork, lamb and bacon. J Kuykendall, L. D., Roy, M. A., O’Neill, J. J. & Devine, T. E. (1988). Appl Bacteriol 56, 25–40. Fatty acids, antibiotic resistance, and deoxyribonucleic acid homology Skerman, V. B. D., McGowan, V. & Sneath, P. H. A. (1980). Ap- groups of Bradyrhizobium japonicum. Int J Syst Bacteriol 38, 358–361. proved lists of bacterial names. Int J Syst Bacteriol 30, 225–420. Maidak, B. L., Cole, J. R., Parker, C. T., Jr & 11 other authors (1999). A new version of the RDP (Ribosomal Database Project). Nucleic Acids Smibert, R. M. & Krieg, N. R. (1981). General characterization. In Res 27, 171–173. Manual of Methods for General Bacteriology, pp. 409–443. Edited by P. Gerhardt, R. G. E. Murray, R. N. Costilow, E. W. Nester, W. A. Marmur, J. (1961). A procedure for the isolation of deoxyribonucleic Wood, N. R. Krieg & G. B. Phillips. Washington, DC: American acid from microorganisms. J Mol Biol 3, 208–218. Society for Microbiology. Martinez-Murcia, A. J. & Collins, M. D. (1991). A phylogenetic Starrenburg, J. C. & Hugenholtz, J. (1991). Citrate fermentation by analysis of an atypical leuconostoc: description of Leuconostoc fallax Lactococcus and Leuconostoc spp. Appl Environ Microbiol 57, 3535– sp. nov. FEMS Microbiol Lett 82, 55–60. 3540. Mesbah, M., Premachandran, U. & Whitman, W. B. (1989). Precise $ $ Strompl, C., Tindall, B. J., Jarvis, G. N., Lunsdorf, H., Moore, E. R. B. measurement of the GjC content of deoxyribonucleic acid by high- performance liquid chromatography. Int J Syst Bacteriol 39, 159–167. & Hippe, H. (1999). A re-evaluation of the of the genus Anaerovibrio, with the reclassification of Anaerovibrio glycerini as Neves, A. R., Ramos, A., Nunes, M. C., Kleerebezem, M., Hugen- Anaerosinus glycerini gen. nov., comb. nov., and Anaerovibrio burkina- In vivo holtz, J., de Vos, W. M., Almeida, J. & Santos, H. (1999). bensis as Anaeroarcus burkinensis [corrig.] gen. nov., comb. nov. Int J nuclear magnetic resonance studies of glycolytic kinetics in Lactococcus Syst Bacteriol 49, 1861–1872. lactis. Biotechnol Bioeng 64, 200–212. Vandamme, P., Pot, B., Gillis, M., De Vos, P., Kersters, K. & Swings, Nichols, P. D., Guckert, J. B. & White, D. C. (1986). Determination of J. (1996). monounsaturated fatty acid double-bond position and geometry for Polyphasic taxonomy, a consensus approach to bacterial systematics. Microbiol Rev 60, 407–438. microbial monocultures and complex consortia by capillary GC-MS of 4 their dimethyl disulphide adducts. J Microbiol Methods 5, 49–55. Veiga-da-Cunha, M., Firme, P., San Romao, M. V. & Santos, H. "$ Rainey, F. A., Ward-Rainey, N., Kroppenstedt, R. M. & Stacke- (1992). Application of C nuclear magnetic resonance to elucidate the brandt, E. (1996). The genus Nocardiopsis represents a phylogenetically unexpected biosynthesis of erythritol by Leuconostoc oenos. Appl coherent taxon and a distinct actinomycete lineage: proposal of Environ Microbiol 58, 2271–2279. Nocardiopsaceae fam. nov. Int J Syst Bacteriol 46, 1088–1092. Williams, R. A. D. & da Costa, M. S. (1992). The genus Thermus and Ramos, A. & Santos, H. (1996). Citrate\sugar co-fermentation in related microorganisms. In The , 2nd edn, pp. 3745–3753. "$ Leuconostoc oenos,a C nuclear magnetic resonance study. Appl Edited by A. Balows, H. G. Tru$ per, M. Dworkin, W. Harder & K. H. Environ Microbiol 62, 2577–2585. Schleifer. New York: Springer.

http://ijs.sgmjournals.org 655