Lactic acid bacteria associated with wine grapes from several Australian vineyards

Transcription

1 Journal of Applied Microbiology ISSN ORIGINAL ARTICLE Lactic acid bacteria associated with wine grapes from several Australian vineyards S. Bae, G.H. Fleet and G.M. Heard Food Science and Technology, School of Chemical Engineering and Industrial Chemistry, The University of New South Wales, Sydney, New South Wales, Australia Keywords enrichment, lactic acid bacteria, PCR-DGGE, wine grapes. Correspondence Graham H. Fleet, Food Science and Technology, School of Chemical Engineering and Industrial Chemistry, The University of New South Wales, Sydney, New South Wales 2052, Australia /0521: received 15 May 2005, revised 18 September 2005 and accepted 26 September 2005 doi: /j x Abstract Aims: The detection and isolation of lactic acid bacteria by enrichment methods from wine grapes cultivated in vineyards located in New South Wales, Australia. Methods and Results: Enrichment cultures in de Man, Rogosa and Sharpe (MRS) broth, MRS + ethanol (5%), MRS broth supplemented with 15% (v/v) tomato juice (MRST), ph 5Æ5 and 3Æ5 and autoenrichment in grape juice homogenate were used to detect lactic acid bacteria on wine grapes. Bacteria were isolated from enrichment cultures by plating onto MRS and MRST agar and identified by 16S rdna sequence analysis and phenotypical methods. A molecular method, PCR-denaturing gradient gel electrophoresis (DGGE) was also used to examine the bacteria that developed in enrichment cultures. Species of Lactobacillus, Enterococcus, Lactococcus and Weissella were detected in enrichments by plating and PCR-DGGE. Other bacteria (Sporolactobacillus, Asaia, Bacillus ssp.) were also found in some enrichment cultures. The principal malolactic bacterium, Oenococcus oeni, was not isolated. Conclusions: The incidence and populations of lactic acid bacteria on wine grapes were very low. Damaged grape berries showed a greater presence of these bacteria than undamaged berries. The diversity of bacterial species isolated from the grapes was greater than those previously reported and represented both lactic acid bacteria and nonlactic acid bacteria. Some of these bacteria (i.e. Lactobacillus lindneri, Lactobacillus kunkeei) could be detrimental to wine production. Oenococcus oeni was not found on grapes, but its recovery could be obscured by overgrowth from other species. Significance and Impact of the Study: Lactic acid bacteria are significant in wine production because they conduct the malolactic fermentation and cause stuck or sluggish alcoholic fermentation and wine spoilage. This study investigates wine grapes as a potential source of these bacteria. Introduction Lactic acid bacteria have a significant role in the production of wines (Lonvaud-Funel 1999; Fleet 2001, 2003). They are responsible for conducting the malolactic fermentation (MLF) which is an important secondary reaction that occurs in many wines after alcoholic fermentation by yeasts. The MLF has important influences on wine acidity, flavour and microbiological stability (Henick-Kling 1993; Lonvaud-Funel 1995, 1999). Oenococcus oeni, formerly Leuconostoc oenos, is the main species responsible for MLF, although other species of lactic acid bacteria may also conduct this reaction. Various species of Lactobacillus, Pediococcus and Leuconostoc can cause spoilage of wine during bulk storage in the cellar and after bottling (Sponholz 1993). Growth of lactic acid bacteria in grape juice can have inhibitory effects on yeasts and cause stuck incomplete 712 ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)

2 S. Bae et al. Lactic acid bacteria on wine grapes and sluggish alcoholic fermentations (Bisson 1999; Fleet 2003). Despite their importance, few studies have specifically investigated the origin or source of lactic acid bacteria in the winemaking process. Freshly extracted grape juice or must, produced under commercial conditions, generally contains various species of Lactobacillus, Pediococcus and Leuconostoc at populations of CFU per ml (Costello et al. 1983; Fleet et al. 1984; Pardo and Zuńiga 1992; Fleet 1993; Fugelsang 1997). It is considered that these bacteria originate from the surface of the grapes or as contaminants of winery equipment that is used to process the juice or must. Surprisingly, very few studies have focused on the bacteria associated with grapes. Lactobacillus, Lactobacillus casei, Lactobacillus brevis, Lactobacillus hilgardii, Lactobacillus curvatus, Lactobacillus buchneri, Leuconostoc dextranicum and Leuconostoc mesenteroides were inconsistently isolated from several grape varieties harvested from vineyards in Spain (Sieiro et al. 1990; Suárez et al. 1994), France (Lafon-Lafourcade et al. 1983) and Germany (Weiller and Radler 1970). We have not been able to find any literature that conclusively associates the principal malolactic bacterium, O. oeni with the surfaces of immature, mature or damaged grape berries (Radler 1958). Nevertheless, this species is commonly isolated from wines (Chalfan et al. 1977; Davis et al. 1985, 1986; Wibowo et al. 1985; Edwards et al. 1991). As a part of a broad study on the yeasts and bacteria associated with wine grapes cultivated in Australia, we have rarely detected the presence of lactic acid bacteria by plate culture methods or by culture-independent methods using PCR-denaturing gradient gel electrophoresis (DGGE). In particular, we have not been able to isolate the key malolactic bacterium, O. oeni, from grapes. Possibly, these bacteria may be present on grapes at such low populations that they are not detectable by plate culture and PCR-DGGE methods. Culture enrichment methods, therefore, may be required for their isolation. In this paper, we report the application of enrichment culture to investigate the presence of lactic acid bacteria on wine grapes harvested from several Australian vineyards. PCR- DGGE was also used to examine the bacterial species that developed during enrichment culture of the grapes. Healthy, undamaged and physically damaged grape berries were analysed in this study. Materials and methods Grape samples Grape samples, collected during January March 2004, were obtained from six commercial vineyards located in the upper and lower Hunter Valley, Mudgee and Griffith regions of New South Wales, Australia. The northern region (upper and lower Hunter Valley), the northwestern region (Mudgee) and the southwestern region (Griffith) are geographically distinct from each other, being km from Sydney. Grapes of optimal maturity were hand-harvested and included five varieties of red grapes (Cabernet Sauvignon, Merlot, Pinot Noir, Shiraz, Tyrian) and three varieties of white grapes (Chardonnay, Sauvignon Blanc, Semillon). Among these varieties, Cabernet Sauvignon and Shiraz from one vineyard at the upper Hunter Valley and Cabernet Sauvignon, Shiraz and Tyrian from another at Griffith were collected on two occasions: at 30 days prior to harvest and at harvest time (full maturity). Each sample (about 2 kg) consisted of bunches of grapes that were randomly and aseptically removed from at least five different vines within the vineyard. Separate samples of undamaged and damaged grape bunches were taken for each variety at each vineyard. Damaged grapes were identified based on their physical appearance (broken skin, shrivelled, discoloured). Samples were stored at 4 C and analysed within 24 h of harvest from the vine. Sample preparation The overall strategy for this study is outlined in Fig. 1. Individual berries were randomly and aseptically removed from the bunches to give a composite sample of about 500 g. They were placed in sterile Stomacher bags and homogenized for 2 min (StomacherÒ Lab System, Norfolk, UK). Homogenate (10 ml) was used as inoculum for enrichment cultures. Enrichment cultures were conducted in the following media: (i) de Man, Rogosa and Sharpe (MRS) broth, (ii) MRS broth containing 5% (v/v) ethanol (Sigma Chemical Co., St. Louis, MO, USA), (iii) MRST 5Æ5 broth which consisted of MRS broth supplemented with 15% (v/v) tomato juice (MRST; Berri, Australia), adjusted to ph 5Æ5 and (iv) MRST 3Æ5 broth adjusted to ph 3Æ5. MRS broth was obtained from Oxoid, Melbourne, Australia. All these media were supplemented with 100 mg l )1 of cycloheximide (AppliChem GmbH, Darmstadt, Germany) to prevent the growth of yeasts and other fungi. Media were dispensed as 90 ml volumes in screw-capped containers (120 ml size, Sarstedt Australia Pty Ltd, Technology Park, SA, Australia). While MRS broth is widely used to culture and isolate lactic acid bacteria, its performance can be improved by fortification with tomato juice (Pan et al. 1982). The conditions of 5% ethanol and ph 3Æ5 were included to bias the selectivity to wine conditions. An additional enrichment culture, termed autoenrichment, was conducted and consisted of a sample (100 ml) of the ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)

3 Lactic acid bacteria on wine grapes S. Bae et al. Grapes Extracted grape juice MRS broth with 30% glycerol at )80 C. Generally, the primary isolation plates from enrichment cultures gave colonies that were homogeneous in their morphology and appearance. Consequently, only two to three representatives were selected for detailed identification. Enrichment cultures (Autoenrichment, MRS, MRS + EtOH, MRST[pH 5 5], MRST [ph 3 5]) PCR-DGGE DGGE band excision grape homogenate (with 100 mg l )1 of cycloheximide) that was incubated at 25 C. All enrichment cultures were incubated at 25 C for 10 days. Cultures were sampled for isolation of bacteria at 0 (no enrichment), 5 and 10 days. Isolation of lactic acid bacteria Isolation of lactic acid bacteria by plate culture DNA Extraction Sequencing of 16S rdna Taxonomic identification Amplification of 16S rdna Figure 1 Experimental protocol for the analysis of lactic acid bacteria on grapes. Samples of enrichment and autoenrichment cultures were diluted as required in sterile 0Æ1% bacteriological peptone (Oxoid) and aliquots (0Æ1 ml) were spread inoculated in duplicate over the surface of plates of MRS agar or MRST agar (ph 5Æ5), both supplemented with cycloheximide to a final concentration of 100 mg l )1. After incubation at 25 C for 3 7 days, colonies were differentiated and counted based on their morphology. Presumptive lactic acid bacteria were selected and restreaked on the corresponding agar medium (without cycloheximide) to obtain pure cultures. Isolates were maintained as liquid cultures in Identification of lactic acid bacteria Isolates were presumptively classified as lactic acid bacteria if they were rods or cocci that gave Gram-positive and catalase-negative reactions. They were then identified by sequence analysis of segments of the 16S ribosomal DNA. Essentially, pure cultures were grown and the DNA was extracted as described later. The extracted DNA was amplified by PCR using the universal primers flanking the region between 968 and 1401 (Escherichia coli numbering) as described by Bae et al. (2004). The amplicons were confirmed by agarose gel electrophoresis and then used directly for sequencing with ABI PRISMÒ BigDye TM Terminators v 3Æ1 Cycle Sequencing kit (Applied Biosystems, Foster City, CA, USA). The products were sequenced at the Automated DNA analysis facility, School of Biotechnology and Biomolecular Sciences, University of New South Wales. The sequences were aligned to determine the closest known relatives of 16S rdna gene fragments available from BLAST in the GenBank database (Karlin and Altschul 1990). Some isolates were subject to further phenotypical characterization, as an additional approach to confirm their identification. These isolates as well as reference strains of lactic acid bacteria were grown on MRS agar plates until prominent colonies were observed. Biomass from single colonies was mixed in sterile saline to give a suspension of CFU per ml. This suspension was used to inoculate media for the following reactions: gas formation from glucose; production of ammonia from l-arginine; mannitol formation from fructose (Pilone et al. 1991); growth at 10 C, 15 C, 37 C and 45 C in MRS broth; growth in MRS broth at ph 4Æ8 or9æ6 and growth in MRS broth containing 3Æ0 or 6Æ5% NaCl (Kandler and Weiss 1986; Schleifer 1986). Reference strains used for these tests were brevis UNSW , Lactobacillus rhamnosus UNSW , UNSW , Leuc. mesenteroides UNSW and Pediococcus pentosaceus UNSW They were obtained from the culture collection, School of Biotechnology and Biomolecular Sciences, University of New South Wales. DNA extraction DNA was extracted by procedures described previously (Lopez et al. 2003). For isolation of total DNA from either pure cultures or samples of enrichment cultures, 714 ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)

4 S. Bae et al. Lactic acid bacteria on wine grapes the cell pellet from a 2-ml culture sample (centrifuged at g, 10 min) was resuspended in 200 ll of breaking buffer [2% Triton X-100, 1% SDS, 100 mmol l )1 NaCl, 10 mmol l )1 Tris-Cl (ph 8Æ0), 1 mmol l )1 EDTA (ph 8Æ0)] and 20 ll of lysozyme (10 mg ml )1, Sigma) and incubated for 15 min at 37 C. The mixture was transferred to a microtube containing 0Æ3 g of 0Æ1 mm diameter zirconia/silica beads (BioSpec Products Inc., Bartlesville, OK, USA). The cells were homogenized twice for 40 s at a speed setting of 46 in a bead beater (Mini-BeadBeater-1 TM, BioSpec Product Inc.) in the presence of 200 ll of phenol chloroform isoamylalcohol (25 : 24 : 1). Two hundred microlitres of TE buffer (10 mmol l )1 Tris-Cl, 1 mmol l )1 EDTA, ph 8Æ0) were then added, and the bead cell mixture was centrifuged at g for 10 min at 4 C. The aqueous phase was transferred to another microcentrifuge tube, and the DNA was precipitated by the addition of 0Æ6 vol. of isopropanol and recovered by centrifugation at g for 10 min at 4 C. The DNA pellet was washed with 70% ethanol, dried and resuspended in 50 ll of sterile TE buffer. The crude DNA extracts of enrichment cultures were further purified by the Wizard DNA Purification Kit (Promega Corp., Madison, WI, USA) according to the protocol supplied by the manufacturer. Analysis of enrichment cultures by PCR-DGGE Purified DNA was amplified with primers 341f-gc/518r (Muyzer et al. 1993) and 968f-gc/1378r (Heuer et al. 1999) spanning the V3 region and V6 V9 region of the 16S ribosomal DNA, respectively. The 40-nucleotide GC-rich sequence at the 5 -end of the forward primers (341f and 968f) was attached to improve the detection of sequence variations of amplified DNA fragments by subsequent DGGE (Muyzer et al. 1993). These primers were obtained from SigmaGenosys, Castle Hill, NSW, Australia. The PCR reaction mixtures contained 10 mmol l )1 Tris HCl (ph 8Æ3), 50 mmol l )1 KCl, 0Æ2 lmol l )1 of each primer, 200 lmol l )1 of each dntp (Roche Diagnostics, Indianapolis, IN, USA), 1Æ5 mmol l )1 MgCl 2, 1Æ25 U of Gold Taq DNA Polymerase (AmpliTaq TM, Roche Molecular Systems, Branchburg, NJ, USA) and 10 ng of purified template DNA in 50 ll final volume. Amplification using the 341f-gc/518r primers was performed under the following program: initial denaturation at 95 C for 7 min, 20 cycles at 94 C for 1 min (denaturation), 65 C for 1 min (annealing) and 72 C for 1 min (extension) with a 1 C touchdown every second cycle, followed by 10 cycles with an annealing temperature of 55 C, and the final extension was conducted for 10 min at 72 C. The concentration of MgCl 2 was increased to 3Æ75 mmol l )1 for PCR with the 968f-gc/1378r. Amplification with these primers was performed as follows: initial denaturation at 95 C for 7 min, 10 cycles at 94 C for 1 min (denaturation), 60 C for 1 min (annealing) and 72 C for 2 min (extension) with a 1 C touchdown every second cycle, followed by 20 cycles with an annealing temperature of 55 C, and the final extension was conducted for 10 min at 72 C. PCR amplifications were conducted in a 9600 Thermal Cycler (Applied Biosystems). The PCR amplification products were verified by 2% agarose gel electrophoresis and then resolved by DGGE using a DCode TM Universal Mutation Detection System (Bio-Rad, Hercules, CA, USA) to determine the species composition of the enrichment cultures. Essentially, DGGE was performed as described by Muyzer et al. (1993). Eight percent polyacrylamide gels (acrylamide: N,N -methylene bisacrylamide ratio, 19 : 1, Bio-Rad) with 30 50% linear denaturing gradient (100% denaturant corresponds to 7 mol l )1 urea and 40% formamide) were used for analysing amplicons of 341f-gc/518r. Electrophoresis was performed for 30 min at 20 V and for 6 h at 120 V with a constant temperature of 60 C. Six percent polyacrylamide gels with 45 65% linear denaturing gradient were used for analysing amplicons produced with primers of 968fgc/1378r, and electrophoresis was run for 30 min at 20 V, then for 14 h at 100 V. Subsequently, gels were stained with SYBRÒ Green I (AmrescoÒ, Solon, Ohio, USA) for 20 min in 1 TAE buffer (40 mmol l )1 Tris-acetate, 1 mmol l )1 EDTA, ph 8Æ0) and photographed under UV transillumination with a Polaroid DS-34 Camera (Polaroid, Scoresby, Vic, Australia). DNA bands in DGGE gels were excised and incubated overnight in 50 ll of sterile Milli-Q water (Millipore Academic System; Millipore Australia Pty Ltd, North Ryde, NSW, Australia) at 4 C. Eluted DNA was re-amplified with corresponding primers, and the PCR products were re-evaluated by DGGE. Only products that migrated as a single band and at the same distance with respect to the original sample were then excised and amplified with corresponding primers without a GC clamp. Amplified products were purified with QIAquick PCR Purification Kit (Qiagen Pty Ltd, Clifton Hill, Victoria, Australia) and subjected to DNA sequence analysis. Interactive growth of Oenococcus oeni, Lactobacillus kunkeei and Lactobacillus lindneri To investigate interactive growth between O. oeni and Lactobacillus kunkeei, and O. oeni and Lactobacillus lindneri, mixed cultures were conducted in MRST broth at ph 3Æ5 and 5Æ5. Two strains of O. oeni were isolated by plating samples from commercial starter culture products onto MRST agar. Oenococcus oeni 1 and O. oeni 2 were isolated from Viniflora CH35 TM and Viniflora Oenos TM ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)

5 Lactic acid bacteria on wine grapes S. Bae et al. (CHR Hansen Pty Ltd, Bayswater, Victoria, Australia), respectively. Oenococcus oeni, grown in MRST broth at 25 C for 5 days, and kunkeei or, grown in MRST broth at 25 C for 24 h, were inoculated as single or mixed cultures into 200 ml of MRST broth in a 250 ml DuranÒ Schott bottle (Schott UK Ltd, Stafford, UK). The cultures were incubated at 25 C for 10 days. Samples were removed every 12 h or 24 h to determine the population of these micro-organisms by spreading plating onto MRST agar. Isolates of O. oeni, kunkeei and were also characterized for their interactions using the spot on lawn assay and the deferred antagonism assay as described previously (Bae et al. 2004). Results Isolation of lactic acid bacteria from enrichment cultures To increase the prospects of isolating lactic acid bacteria, the experimental plan included analysis of grapes representing eight different cultivars, grapes harvested from six vineyards located in four different wine producing regions and samples of undamaged and damaged grape berries. Overall, 43 batches of grapes (21 undamaged, 22 damaged) were examined. Samples from each batch were analysed by five different enrichment procedures. Isolations were prepared from each enrichment culture after 5 days of incubation (when microbial growth first became evident) and, again, after 10 days of incubation. Tables 1 4 show the species of lactic acid bacteria isolated from grapes according to this plan. The populations of the species in the enrichment cultures are given in the tables, with most being in the range CFU per ml. With a few exceptions, data from the 5-day enrichment cultures and 10-day cultures were generally similar (Table 1). Consequently, Tables 2 4 report only the 5-day data, with differences from the 10-day data given as footnotes. All grape samples were analysed by direct plate culture before enrichment (0 days), but only a few samples of damaged grapes of Chardonnay (vineyard D), Shiraz and Chardonnay (vineyard E) and Sauvignon Blanc (vineyard F) gave isolation of lactic acid bacteria (, kunkeei) at CFU per ml by this method. Lactic acid bacteria were isolated from only 9 of the 21 batches of undamaged grapes and 16 of 22 batches of damaged grapes (Tables 1 4). Thus, there were numerous batches from which no lactic acid bacteria could be cultured. In many cases, bacteria other than lactic acid bacteria were found in enrichment cultures. For example, Cabernet Sauvignon and Tyrian grapes from vineyard F of the Griffith region (Table 1) gave no lactic acid bacteria in any of the enrichment cultures, but Sporolactobacillus inulinus (Gram-positive rod, spore former, catalase negative and identified by sequencing) was isolated in MRS, MRS + EtOH and MRST ph 5Æ5 enrichment cultures at populations of CFU per ml. No lactic acid bacteria were isolated from undamaged Semillon grapes from vineyard A of the Hunter Valley region (Table 4) but all enrichment cultures, except MRST + EtOH, showed the presence of Asaia siamensis (Gram-negative rod, catalase positive and identified by sequencing). Species of Bacillus, Staphylococcus and Gluconobacter were also detected infrequently. Nonlactic acid bacteria were mostly detected in enrichment cultures where lactic acid bacteria were not present, but there were some enrichment cultures that contained both nonlactic acid bacteria and lactic acid bacteria. Overall, very few batches (2 out of 21 undamaged grapes and 1 out of 22 damaged grapes) showed no bacterial presence in enrichment cultures. Table 5 lists the different species of lactic acid bacteria that were found and their frequency of isolation from the different batches of grapes. The frequency of isolation was higher from white grape varieties (25/19, isolation/batches) than from red grape varieties (14/24), and from damaged grape samples (27/22) than from undamaged grape samples (12/21). The isolates consisted of six species of Lactobacillus, four species of Enterococcus and one species from each of Lactococcus and Weissella. Species of Leuconostoc, Pediococcus and Oenococcus were not found. The most frequently isolated species were, Enterococcus durans, Lactococcus lactis and kunkeei. The data are too limited to show any definitive relationship between species occurrence, grape variety and vineyard from where the grapes were collected. Several batches of red grapes were collected from vineyards A and F, 30 days before their scheduled harvest time. For vineyard A, associated lactic acid bacteria and grapes were Enterococcus hermanniensis (Shiraz) and Enterococcus avium, Ent. hermanniensis, (Cabernet Sauvignon), and for vineyard F, the data were Enterococcus faecium, Weissella paramesenteroides (Shiraz), Ent. faecium (Cabernet Sauvignon) and W. paramesenteroides (Tyrian). The reasons for using an array of enrichment conditions were described previously (see also Pan et al. 1982). It is evident from Tables 1 4 that, for some grape batches, the enrichment medium could determine whether or not any lactic acid bacteria were isolated at all, and the species diversity found. Autoenrichment and enrichment in MRST adjusted to ph 3Æ5 were the conditions that gave the least isolations of lactic acid bacteria (Table 6). Similarly, overall isolation frequencies were found for enrichment in MRS, MRS + EtOH and MRST ph 5Æ5, but there were examples where some species were isolated under one condition but not the other. For damaged Chardonnay grapes from vineyard C, L. lactis was isolated 716 ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)

6 S. Bae et al. Lactic acid bacteria on wine grapes Table 1 Isolation of lactic acid bacteria by enrichment culture from Cabernet Sauvignon, Merlot, Pinot Noir and Tyrian grapes Enrichment condition Grape variety Vineyard* Physical status Autoenrichment MRS MRS + EtOH MRST (ph 5Æ5) MRST (ph 3Æ5) 5 days 10 days 5 days 10 days 5 days 10 days 5 days 10 days 5 days 10 days Cabernet Sauvignon A UD D mali (6Æ )à (2Æ ) mali (2Æ ) (1Æ ) (5Æ ) (4Æ ) mali (4Æ ) (9Æ ) (2Æ ) mali (1Æ ) (6Æ ) mali (2Æ ) (2Æ ) (2Æ ) D UD D (2Æ ) (4Æ ) F UD D Merlot D UD (2Æ ) D Ent. durans (1Æ ) (1Æ ) Ent. durans (1Æ ) (4Æ ) Ent. durans (1Æ ) Ent. durans (6Æ ) (7Æ ) Pinot Noir D UD Ent. durans (1Æ ) (1Æ ) Ent. durans (3Æ ) D Tyrian F UD D mali (2Æ ) (2Æ ), Lactobacillus; Ent., Enterococcus; ), no colonies of lactic acid bacteria detected on plating media. *Vineyard location: A, Hunter Valley; D, Mudgee; F, Griffith. UD, undamaged grape berries; D, damaged grape berries. àpopulation (CFU per ml) in enrichment culture. ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)

7 Lactic acid bacteria on wine grapes S. Bae et al. Table 2 Isolation of lactic acid bacteria by enrichment culture from Shiraz grapes Enrichment condition Vineyard* Physical status Autoenrichment MRS MRS + EtOH MRST (ph 5Æ5) MRST (ph 3Æ5) 5 days 5 days 5 days 5 days 5 days A UD L. lactis (2Æ )à (3Æ ) D L. lactis (6Æ ) L. lactis (1Æ ) B UD D (2Æ ) (9Æ ) (2Æ ) C UD Ent. avium Ent. avium (1Æ ) (2Æ ) D D UD D E UD D (4Æ ), Ent. durans (7Æ ) (3Æ ), Ent. durans (1Æ ) (1Æ ) F UD D L. lactis L. lactis (1Æ ) (4Æ ) L., Lactococcus;, Lactobacillus; Ent., Enterococcus. *Vineyard location: A, B, C, Hunter Valley; D, E, Mudgee; F, Griffith. UD, undamaged grape berries; D, damaged grape berries. àpopulation (CFU per ml) in enrichment culture. Ent. avium also found in the 10-day enrichments. Ent. durans not found in the 10-day enrichments. (4Æ ) by enrichment in MRS, kunkeei was obtained in MRS + EtOH and was found in MRST ph 5Æ5 (Table 3). Damaged Chardonnay grapes from vineyard F gave Ent. durans in MRS, W. paramesenteroides in MRS + EtOH and Lactobacillus ssp. in MRST ph 5Æ5. While most samples gave only one species after enrichment, there were some samples where more than one species was found (e.g. Lactobacillus mali and in damaged Cabernet Sauvignon from vineyard A, Table 1; kunkeei and in damaged Sauvignon Blanc from vineyard F, Table 4). Identification of isolates A total of 160 strains were isolated from the enrichment cultures. All isolates were considered to be lactic acid bacteria based on their positive Gram reactions, nonmotility, absence of catalase activity and spore formation and rod or coccal shape. All 160 isolates were identified by sequencing of partial 16S ribosomal DNA and then divided into nine major groups based on sequence analysis. Representative isolates from each group were examined for key phenotypical tests as mentioned in the Materials and Methods section. Group I included 13 strains showing ovoid cells in short chains that did not produce gas from glucose and hydrolyse l-arginine. They did not grow at 45 C, which is different from the standard description (Schleifer 1986), and showed weak growth at ph 4Æ8 after incubation for 7 days. The identity of all 13 strains was determined as Ent. avium (AY442814) with 96 99% DNA sequence similarity (Schleifer 1986). Groups II, III and IV were homofermentative cocci in short chains and hydrolysed l-arginine to ammonia. In comparison with other groups, isolates of group II grew well over a range of different conditions. All 14 strains were identified as Ent. faecium (GenBank accession No. AJ420800) with 97 99% homology (Schleifer 1986). No growth at ph 4Æ8 and relatively weak growth at 45 C were detected for isolates of group III. All 14 isolates of group III were shown to be Ent. durans (AJ420801) with 97 99% homology (Schleifer 1986). Group IV included six strains that did not grow at 45 C, 718 ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)

8 S. Bae et al. Lactic acid bacteria on wine grapes Table 3 Isolation of lactic acid bacteria by enrichment culture from Chardonnay grapes Enrichment condition Vineyard* Physical status Autoenrichment MRS MRS + EtOH MRST (ph 5Æ5) MRST (ph 3Æ5) 5 days 5 days 5 days 5 days 5 days A UD Ent. avium (1Æ ) Ent. avium (1Æ ) Ent. avium (1Æ ) D C UD kunkeei (1Æ )à Ent. faecium (4Æ ) kunkeei (2Æ ) kunkeei (1Æ ) D L. lactis kunkeei (1Æ ) (3Æ ) (2Æ ) D UD Ent. durans Ent. durans Ent. durans (1Æ ) (8Æ ) (5Æ ) D (3Æ ) (2Æ ) (1Æ ) (2Æ ) (3Æ ) E UD D kunkeei (1Æ ) (1Æ ) (7Æ ), Ent. faecium (6Æ ) (1Æ ) (7Æ ) F UD W. paramesenteroides (1Æ ) D Ent. durans (7Æ ) L., Lactococcus;, Lactobacillus; Ent., Enterococcus; W., Weissella. *Vineyard location: A, C, Hunter Valley; D, E, Mudgee; F, Griffith. UD, undamaged grape berries; D, damaged grape berries. àpopulation (CFU per ml) in enrichment culture. W. paramesenteroides (1Æ ) W. paramesenteroides (1Æ ) Lactobacillus sp. (9Æ ) which is unlike the other enterococci, and showed 98 99% DNA sequence similarity to Ent. hermanniensis (Koort et al. 2004). Group V included seven strains with similarity to enterococci, but they were distinguished by their inability to grow at 6Æ5% NaCl and ph 9Æ6. They were shown to be L. lactis (AF493058) with 99% DNA sequence homology (Schleifer 1986). Group VI isolates were short rods, produced gas from glucose and gave no formation of ammonia from l-arginine and mannitol from fructose. They did not grow at 10 C or45 C. A total of 11 strains of group V were identified as W. paramesenteroides (S67831) with 97 99% sequence similarity (Cai et al. 1998). A large number of lactic acid bacteria (groups VII and VIII) were distinguished by their rod cell shape, presence of weak catalase activity, production of gas from glucose and mannitol from fructose and their inability to hydrolyse l-arginine. Group VII gave different growth patterns at 10 C and 6Æ5% NaCl compared with active growth of group VIII at these conditions. All 20 isolates of group VII were assigned as kunkeei (Y11374) with 98 99% sequence similarity (Edwards et al. 1998), and 58 isolates of group VIII were identified as (X95421 and D37784) with 97 98% DNA sequence homology (Back et al. 1996). The isolates in group IX differed from the other groups by their rod shape, homofermentative property and failure to produce mannitol from fructose. They grew well under most conditions except at 45 C and ph 9Æ6. Sequence analysis identified this group as (AL935260) with 98 99% similarity (Kandler and Weiss 1986). Analysis of lactic acid bacteria in enrichment cultures by PCR-DGGE In addition to plating on agar media, PCR-DGGE was used to detect lactic acid bacteria in several of the enrichment cultures reported in Tables 1 4. The PCR-DGGE banding patterns of pure cultures of various species of lactic acid bacteria (M1) isolated from grapes by enrichment-agar plating as well as some additional reference species (M2) have been examined and displayed in Fig. 2. Good separation of the different species was obtained. DGGE validation markers (M1 and M2) shown in Fig. 2 were generated by pooling the amplicons produced with 341f-gc/518r primers. Amplicons produced with primers ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)

9 Lactic acid bacteria on wine grapes S. Bae et al. Table 4 Isolation of lactic acid bacteria by enrichment culture from Semillon and Sauvignon Blanc grapes Enrichment condition Grape variety Vineyard* Physical status Autoenrichment MRS MRS + EtOH MRST (ph 5Æ5) MRST (ph 3Æ5) 5 days 5 days 5 days 5 days 5 days Semillon A UD D Ent. hermanniensis Ent. hermanniensis (1Æ )à (1Æ ) B UD D (3Æ ) (3Æ ) (1Æ ) F UD W. paramesenteroides (8Æ ) Sauvignon blanc D W. paramesenteroides (1Æ ) W. paramesenteroides (4Æ ) W. paramesenteroides (4Æ ) D UD D L. lactis Ent. durans Ent. durans (1Æ )** (1Æ ) (9Æ ) F D kunkeei (3Æ ) kunkeei (1Æ ), (1Æ ) L., Lactococcus;, Lactobacillus; Ent., Enterococcus; W., Weissella. *Vineyard location: A, B, Hunter Valley; D, Mudgee; F, Griffith. UD, undamaged grape berries; D, damaged grape berries. àpopulation (CFU per ml) in enrichment culture. W. paramesenteroides not found in the 10-day enrichments; kefiri isolated. Ent. faecium also found in the 10-day enrichments. **L. lactis not found in the 10-day enrichments. kunkeei (9Æ ), (1Æ ) kunkeei (4Æ ), (2Æ ) kunkeei (4Æ ), (4Æ ) Table 5 Frequency of isolation of lactic acid bacteria from undamaged and damaged wine grapes Species Undamaged grapes Damaged grapes R (12)* W (9) R (12) W (10) Ent. durans L. lactis kunkeei W. paramesenteroides Ent. faecium Ent. avium kefiri mali Lactobacillus sp Ent. hermanniensis Total Total L., Lactococcus;, Lactobacillus; Ent., Enterococcus; W., Weissella; R, red grape variety; W, white grape variety. *Number of batches. 968f-gc and 1378r were not well isolated by DGGE, even after varying the concentration of denaturant used in preparation of the electrophoresis gel (data not shown). The enrichment cultures from six batches of grapes were analysed for lactic acid bacteria by both plate culture and PCR-DGGE. For each batch, the 10-day enrichment cultures in MRS, MRS + EtOH, MRST 5Æ5 and MRST 3Æ5 were examined. Table 7 shows the species detected by each method, and Fig. 2 shows the corresponding PCR-DGGE data. For most cases, the same species of lactic acid bacteria were detected by both methods. However, some exceptions occurred. In some examples, species of lactic acid bacteria were recovered by plating but not by DGGE. In the Cabernet Sauvignon grapes (vineyard A), and mali were not detected by PCR-DGGE in the MRST 5Æ5 enrichment but were found by plating. Ent. faecium was not detected by PCR-DGGE but was found by plating in MRS + EtOH and MRST 5Æ5 enrichment of Semillon (vineyard F). In Shiraz (vineyard E) and Sauvignon Blanc (vineyard F) grapes, was not found by PCR-DGGE in the MRST 3Æ5 enrichment but was recovered by plating. In other cases, lactic acid bacteria were detected by PCR-DGGE but not by plating. For example, L. lactis was found in enrichment cultures of the Cabernet Sauvignon (vineyard A) by PCR-DGGE but not by plate culture. In several cases, multiple bands 720 ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)

10 S. Bae et al. Lactic acid bacteria on wine grapes Table 6 Frequency of isolation of lactic acid bacteria from wine grapes according to enrichment medium Species isolated Enrichment condition Autoenrichment MRS MRS + EtOH MRST (ph 5Æ5) MRST (ph 3Æ5) Total Ent. durans kunkeei L. lactis W. paramesenteroides Ent. avium mali Ent. faecium Ent. hermanniensis Lactobacillus sp kefiri Total L., Lactococcus;, Lactobacillus; Ent., Enterococcus; W., Weissella. on DGGE gels were obtained for some species. from enrichment cultures of Shiraz grapes (vineyard E) gave three bands (Fig. 2b) and from all enrichment cultures of Chardonnay grapes (vineyard D) gave nine bands (Fig. 2c). Weissella paramesenteroides from Semillon grapes (vineyard F) gave two bands only for MRST 5Æ5 enrichment culture (Fig. 2e). The nine DNA bands obtained after PCR-DGGE analysis of MRS enrichment culture of damaged Chardonnay grapes (vineyard D) (Figs 2c and 3, lane 1) were excised and sequenced. All of the bands gave sequences that were homologous (92 95%) with. After excision, each DNA band was subjected to another round of PCR- DGGE analysis, and these also gave multiple DNA bands that corresponded with patterns in the first analysis (Fig. 3, lanes 2 4). These bands were all sequenced as. However, band d, from the original analysis only gave one band after the second PCR-DGGE (Fig. 3, lane 5). The mobility of this band was the same as the original band d and corresponded to the band mobility obtained on PCR-DGGE analysis of a pure culture of. Using pure cultures of either or kunkeei, it was determined that PCR-DGGE could detect cell populations down to 10 3 CFU per ml. No DNA bands were obtained when PCR-DGGE was performed with the population of cells diluted to 10 2 CFU per ml (data not shown). A mixture of 10 5 CFU per ml of kunkeei and 10 5 CFU per ml of gave two DNA bands after PCR-DGGE analysis although the band for had weaker intensity. A mixed population of kunkeei (10 6 CFU per ml) and (10 5 CFU per ml) gave only one band with mobility of kunkeei after PCR-DGGE analysis, while a mixture of (10 6 CFU per ml) and kunkeei (10 5 CFU per ml) gave two bands, but the band of kunkeei was weak. Thus, the PCR-DGGE assay could not reliably detect one species in a mixture if its population was tenfold less than the other species. Interactive growth of O. oeni and lactic acid bacteria Failure to isolate O. oeni from the grapes by enrichment culture could mean (i) that it is not present in the habitat or (ii) that it could be present but is out-competed and overgrown by other, faster growing species during enrichment. To test this possibility, we investigated the growth behaviours of two strains of O. oeni in relation to two other bacteria, and kunkeei, that were frequently found on grapes. Both strains of O. oeni gave similar responses, so data for only one strain are reported. Figure 4 shows the growth responses of O. oeni, kunkeei and as single cultures and when grown in mixed cultures of (i) O. oeni with kunkeei and (ii) O. oeni with. As single cultures, O. oeni grew more slowly than kunkeei or lindneri. Oenococcus oeni failed to grow in mixed culture with and exhibited delayed growth in mixed culture with kunkeei of ph 5Æ5 (Fig. 4a). Interestingly, viability of the inoculated cells of O. oeni rapidly decreased to nondetectable levels on mixture with either kunkeei or. Such rapid loss of viability of O. oeni was not observed for the mixture at ph 3Æ5 (Fig. 4b). In fact, there was slight growth of O. oeni during the first 2 days, before it died off. However, it did not ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)

11 Lactic acid bacteria on wine grapes S. Bae et al. (a) (b) (c) A B M1 M M1 M M1 M2 Lp I C D E F G H Lc Lm Ll Ll Ll Ll Ll J K L (d) (e) (f) M1 M M1 M M1 M Wp Lk Wp Lkf Ll Ll Figure 2 DGGE profiles of PCR amplicons for lactic acid bacteria found in enrichment cultures of grapes. Lanes: 1, MRS enrichment; 2, MRS + EtOH enrichment; 3, MRST (ph 5.5) enrichment; 4, MRST (ph 3.5) enrichment; M1 and M2, reference cultures. A, Lactobacillus kunkeei (LK); B, Weissella paramesenteroides (Wp); C, Enterococcus avium; D, Enterococcus faecium; E, Lactobacillus kefiri (Lkf); F, Enterococcus durans; G, Lactococcus lactis (Lc); H, Lactobacillus lindneri (Ll); I, Lactobacillus UNSW (Lp); J, Lactobacillus brevis UNSW ; K, Pediococcus pentosaceus UNSW ; L, Lactobacillus rhamnosus UNSW ; Lm, mali. The subfigures (a), (b), (c), (d), (e), (f) are referred to in Table 7. recover in the mixture with but eventually re-established growth in the presence of kunkeei. The growth of either or kunkeei was not affected (cf. control) by co-culture with O. oeni (Fig. 4c,d). The interaction of O. oeni with two species of Lactobacillus was also determined by the spot on lawn assay and deferred antagonism plate assay (Bae et al. 2004). The two strains of O. oeni were not inhibited by kunkeei, but clear inhibition of the growth of O. oeni by was observed (data not shown). Discussion Grapes are a primary source of micro-organisms associated with winemaking (Fleet 2001). Despite the overall significance of lactic acid bacteria in wine production, there are only occasional reports of their isolation from grapes. Oenococcus oeni is particularly important because of its role in conducting the MLF (van Vuuren and Dicks 1993) but, to our knowledge, it has never been isolated from grapes. Its ecological origin in wine production remains a mystery but, like the principal wine yeast 722 ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)

12 S. Bae et al. Lactic acid bacteria on wine grapes Table 7 Comparison of agar plating and PCR-DGGE methods for detection of lactic acid bacteria in enrichment cultures of wine grapes* Enrichment-plating PCR-DGGE DGGE profile Grape sample Species 1à à (a) Cabernet Sauvignon, damaged, vineyard A ) + mali ) ) + + ) + ) + L. lactis ) ) ) ) + + ) (b) Shiraz, damaged, vineyard E ) (c) Chardonnay, damaged, vineyard D (d) Chardonnay, damaged, vineyard E (e) Semillon, damaged, vineyard F W. paramesenteroides ) + + ) ) kefiri + ) ) ) + ) ) ) Ent. faecium ) + + ) ) ) ) ) (f) Sauvignon Blanc, damaged, vineyard F kunkeei ) L., Lactococcus;, Lactobacillus; Ent., Enterococcus; W., Weissella. *See Tables 1 4. Profiles shown in Fig. 2. àenrichment cultures: 1, MRS; 2, MRS + 5% ethanol; 3, MRST ph 5Æ5; 4, MRST ph 3Æ5. M a b c d Figure 3 Re-evaluation of DNA bands from PCR-DGGE analysis of damaged Chardonnay grapes, vineyard D. Bands a, b, c, d from lane 1 were excised and re-examined by PCR-DGGE giving profiles in lanes 2, 3, 4 and 5, respectively. Saccharomyces cerevisiae, it probably establishes residence in the wine environment after initial introduction from grapes (Fleet et al. 2002). We have had little success in isolating lactic acid bacteria from wine grapes by agar plate culture. Instead, we found a predominance of Bacillus thuringiensis by this method, and this correlated with its widespread use as a bioinsecticide in Australian viticulture (Bae et al. 2004). Consequently, in this study, we focused on enrichment culture as an approach to determine the diversity of lactic acid bacteria associated with wine grapes and used conditions (ph 3Æ5, 5Æ5 and autoenrichment) that prevented the growth of B. thuringiensis and its potential to interfere with the recovery of lactic acid bacteria (Bae et al. 2004) Only 4 of the 43 batches examined gave isolates of lactic acid bacteria by direct plating, suggesting that initial populations were very low and less than 10 2 CFU per gram. Enrichment substantially increased the isolation frequency of lactic acid bacteria from the grapes, with damaged berries giving more isolates. Nevertheless, many samples of both undamaged and damaged berries did not give detectable lactic acid bacteria. Either they were absent in the grapes or they were overgrown by other species during enrichment. Surprisingly, several species of bacteria, other than lactic acid bacteria, were isolated from the enrichment cultures, including those with added ethanol or decreased ph as further selective conditions. The population of these bacteria after 5 10 days of enrichment was about CFU per ml. The two most frequently isolated species in this category were S. inulinus and A. siamensis. Sporolactobacillus inulinus was first described by Kitahara and Suzuki (1963) and is an unusual endospore forming species that produces lactic acid by homofermentation. It has the potential to survive at low ph and tolerate ethanol (Sanders et al. 2003). Asaia siamensis was first described by Katsura et al. (2001) and is considered within the genera of acetic acid bacteria. This is the first time these two species have been isolated from wine grapes. Their potential significance in wine production is not known and requires investigation. It is possible that the growth of these and other bacteria during enrichment suppressed the growth of O. oeni and other lactic acid bacteria, but this requires further study. ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)

13 Lactic acid bacteria on wine grapes S. Bae et al. (a) ph 5 5 (b) ph 3 5 Populations (log CFU ml 1 ) (c) ph 5 5 (d) ph 3 5 Populations (log CFU ml 1 ) Incubation time (day) Incubation time (day) Figure 4 Mixed culture of Oenococcus oeni 1 with either Lactobacillus kunkeei or Lactobacillus lindneri in MRST at ph 5.5 and ph 3.5, 25 C. (a), (b) Populations of O. oeni in single and mixed culture; (c), (d) Populations of kunkeei and. (d) O. oeni (single culture), (s) O. oeni with kunkeei, ( ) O. oeni with, ( ) kunkeei (single culture), (() kunkeei with O. oeni, ( ) (single culture), (n) with O. oeni. It may be necessary to design more highly selective enrichment media to detect the low populations of lactic acid bacteria that occur on wine grapes. It was thought that autoenrichment of grape homogenates might present more specific conditions for the growth of wine-related lactic acid bacteria, but this initiative gave only low isolation rates (Table 6). PCR-DGGE was used as an additional method to detect and identify bacteria in enrichment cultures. Generally, good agreement was obtained between data obtained by PCR-DGGE and plate culture isolation, but some inconsistencies were noted (Table 7). The limitations of PCR-DGGE analysis in ecological studies have been discussed previously (Prakitchaiwattana et al. 2004) and include different affinity of the primer DNA to template DNA for different species and the competitive influences when template DNAs are present in different relative amounts. These findings highlight the need for using a combination of different isolation and detection techniques in ecological studies. Multiple banding is another complicating factor that often arises with PCR- DGGE analysis (Fig. 3). The heterogeneity of 16S rdna because of the presence of multiple copies of the ribosomal genes has been reported to produce several bands per species in PCR-DGGE analysis (Nübel et al. 1996; Fogel et al. 1999). In addition, multiple banding patterns caused by heteroduplex bands, can be obtained during mixed-template PCR when annealing occurs between template DNAs with high sequence similarity (Ferris and Ward 1997; Wintzingerode et al. 1997). Not all nine bands in the DGGE profiles shown in Fig. 3 corresponded to 16S rrna molecules actually present in the samples, and these multiple bands could suggest the formation of chimeric molecules caused by two or more different strains within a species. The diversity of lactic acid bacteria isolated in our study was substantially different to that reported by others. and mali were the only species that corresponded with previous reports (Lafon-Lafourcade et al. 1983; Suárez et al. 1994; Rodas et al. 2003)., Ent. durans, L. lactis and kunkeei, the most frequently isolated species in our study, have not 724 ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)

14 S. Bae et al. Lactic acid bacteria on wine grapes been mentioned previously as grape isolates. The occurrence and prevalence of is noteworthy, as it has not been associated with wine production in the past. However, it has been described as a significant beer spoilage organism (Sakamoto and Konings 2003) and, therefore, could have the potential to spoil wines. Lactobacillus kunkeei was recently described as a new species and was originally isolated from wines undergoing stuck or sluggish fermentations (Edwards et al. 1998). It is an undesirable species in wine production and we report for the first time its association with wine grapes. Both and kunkeei are antagonistic towards O. oeni (particularly ), and their frequent occurrence on grapes may explain the difficulty of isolating O. oeni from this habitat (Fig. 4). Bacteriocin production by these species is probably the underlying mechanism of this inhibition (Radler 1991; Navarro et al. 2000), but this possibility requires confirmation. Such bacteriocins could have commercial application in preventing MLF in wines where this reaction is not desired. The potential use of nisin, plantaricin 423 and pediocin PD-1 for controlling MLF has been reported (Radler 1990; Nel et al. 2002; Bauer et al. 2003). Species of Enterococcus, Weissella and Lactococcus are not previously known as wine micro-organisms (Stiles and Holzapfel 1997), and probably do not grow at the higher ethanol concentrations in wine. Nevertheless, they were isolated from enrichment cultures containing 5% ethanol. The ethanol tolerance of these species requires more detailed investigation. Their association with the surface of plants and agricultural environments is well documented (Mundt et al. 1967; Sandine et al. 1972; Cai 1999; Chen et al. 2005). Consequently, their isolation from the surfaces of grapes is not unexpected. In summary, the incidence and populations of lactic acid bacteria on wine grapes were very low. Damaged grape berries showed a greater presence of these bacteria than undamaged berries. The diversity of bacterial species isolated from the grapes was greater than previously reported and represented both lactic acid bacteria and nonlactic acid bacteria. Some of these bacteria (i.e. lindneri, kunkeei) could be detrimental to wine production. Oenococcus oeni and other well-known wine bacteria within the genera of Leuconostoc and Pediococcus were not found on grapes. Either these bacteria do not commonly occur on grapes or improved methods are needed to detect their presence. Acknoweldgements This research was supported by funds from the Grape and Wine Research and Development Corporation of Australia (GWRDC). References Back, W., Bohak, I., Ehrmann, N., Ludwig, W. and Schleifer, K.H. (1996) Revival of the species Lactobacillus lindneri and the design of a species specific oligonucleotide probe. Syst Appl Microbiol 19, Bae, S., Fleet, G.H. and Heard, G.M. (2004) Occurrence and significance of Bacillus thuringiensis on wine grapes. Int J Food Microbiol 94, Bauer, R., Nel, H.A. and Dicks, L.M.T. (2003) Pediocin PD-1 as a method to control growth of Oenococcus oeni in wine. Am J Enol Vitic 54, Bisson, L.F. (1999) Stuck and sluggish fermentations. Am J Enol Vitic 50, Cai, Y. (1999) Identification and characterization of Enterococcus species isolated from forage crops and their influence on silage fermentation. J Dairy Sci 82, Cai, Y., Benno, Y., Ogawa, M., Ohmomo, S., Kumai, S. and Nakase, T. (1998) Influence of Lactobacillus spp. from an inoculant and of Weissella and Leuconostoc spp. from forage crops on silage fermentation. Appl Environ Microbiol 64, Chalfan, Y., Goldberg, I. and Mateles, R.I. (1977) Isolation and characterization of malolactic bacteria from Israeli red wines. J Food Sci 42, Chen, Y.S., Yanagida, F. and Shinohara, T. (2005) Isolation and identification of lactic acid bacteria from soil using an enrichment procedure. J Appl Microbiol 40, Costello, P.J., Morrison, G.J., Lee, T.H. and Fleet, G.H. (1983) Numbers and species of lactic acid bacteria in wine during vinification. Food Technol Aust 35, Davis, C.R., Wibowo, D., Eschenbruch, R., Lee, T.H. and Fleet, G.H. (1985) Practical implications of malolactic fermentation. A review. Am J Enol Vitic 36, Davis, C.R., Wibowo, D., Lee, T.H. and Fleet, G.H. (1986) Growth and metabolism of lactic acid bacteria during and after malolactic fermentation of wines at different ph. Appl Environ Microbiol 51, Edwards, C.G., Hagg, K.M., Collins, M.D., Hutson, R.A. and Huang, Y.C. (1998) Lactobacillus kunkeei sp. nov.: a spoilage organism associated with grape juice fermentations. J Appl Microbiol 84, Edwards, C.G., Jensen, K.A., Spayd, S.E. and Seymour, B.J. (1991) Isolation and characterization of native strains of Leuconostoc oenos from Washington State wines. Am J Enol Vitic 42, Ferris, M.J. and Ward, D.M. (1997) Seasonal distributions of dominant 16S rrna-defined populations in a hot spring microbial mat examined by denaturing gradient gel electrophoresis. Appl Environ Microbiol 63, Fleet, G.H. (1993) The microorganisms of winemaking isolation, enumeration and identification. In Wine Microbiology and Biotechnology ed. Fleet, G.H. pp Switzerland: Harwood Academic Publisher. ª 2006 The Society for Applied Microbiology, Journal of Applied Microbiology 100 (2006)