OUP user menu

Application and evaluation of denaturing gradient gel electrophoresis to analyse the yeast ecology of wine grapes

Cheunjit J. Prakitchaiwattana, Graham H. Fleet, Gillian M. Heard
DOI: http://dx.doi.org/10.1016/j.femsyr.2004.05.004 865-877 First published online: 1 September 2004


The performance of denaturing gradient gel electrophoresis (DGGE) for analysing yeasts associated with wine grapes was compared with cultural isolation on malt extract agar (MEA). After optimisation of PCR and electrophoretic conditions, the lower limit of yeast detection by PCR-DGGE was 102 cfu ml−1, although this value was affected by culture age and the relative populations of the species in mixed culture. In mixed yeast populations, PCR-DGGE detected species present at 10–100-fold less than other species but not when the ratio exceeded 100-fold. Aureobasidium pullulans was the main species isolated from immature, mature, and both damaged and undamaged grapes. It was not detected by PCR-DGGE when present at populations less than 103 cfu g−1. When approaching maturity, damaged grapes gave a predominance of Metschnikowia and Hanseniaspora species (105–107 cfu g−1), all detectable using PCR-DGGE. However, various species of Rhodotorula, Rhodosporidium and Cryptococcus were not detected by this method, even when populations were as high as 104 cfu g−1. PCR –DGGE was less sensitive than culture on MEA for determining the yeast ecology of grapes and could not reliably detect species present at populations less than 104 cfu g−1. However, this method detected a greater diversity of species than agar plating.

  • Grape
  • Yeast
  • Ecology
  • Denaturing gradient gel electrophoresis

1 Introduction

Yeasts have a defining impact on wine quality and value. They conduct the alcoholic fermentation of grape juice, thereby contributing to the basic chemical structure and individuality of wine flavour and aroma. At later stages, they can cause wine spoilage and loss of product value [1,2]. Although Saccharomyces cerevisiae and Saccharomyces bayanus are widely regarded as the principal yeasts of wine fermentation, there is increasing acceptance that other species, often referred to as indigenous, non-Saccharomyces yeasts, make important contributions to wine fermentation and character [3]. Grapes are a primary source of yeasts in wine production. Consequently, it is important to have reliable knowledge about their ecology [4].

During the past 100 years, many studies have reported the yeasts associated with wine grapes [47]. However, most of these studies give only simple qualitative descriptions of the yeasts isolated, and make little attempt to ask why certain species are predominant and what factors affect their occurrence. Answers to these questions are important in managing the indigenous yeast flora of wine grapes and optimising their potential impact on wine quality. Many studies have found that species of Hanseniaspora (anamorph Kloeckera), Metschnikowia and Candida are predominant on wine grapes at the time of harvest [4]. The principal wine yeasts, S. cerevisiae and S. bayanus, are infrequently isolated from grapes, and there is significant controversy as to their natural origin in wine production [8]. Methodology for the isolation and identification of yeasts from wine grapes has followed conventional, cultural procedures, and inherent weaknesses in this approach have probably limited progress in understanding this ecology [8].

Molecular analytical methods based on PCR-denaturing gradient gel electrophoresis (DGGE) are now being applied to the field of microbial ecology [9]. In this approach, total DNA is extracted from the ecosystem and microbial DNA is specifically amplified by PCR using particular groups of universal primers. Yeast DNA is generally amplified using primers that target the D1/D2 domain of the 26S subunit of ribosomal DNA. The DNA amplicons that are generated are then separated and detected by DGGE, which resolves individual amplicons on the basis of differences in nucleotide sequence. Individual bands are isolated from the gels and sequenced to give species identity. PCR-DGGE offers an alternative, culture-independent strategy to microbiological analysis [9], and is finding increasing application to food [10], including wine fermentations [1113]. In addition to detecting species normally found by cultural methods, it offers the prospect of detecting species that may be present in the habitat at viable but non-culturable states [14,15].

In this paper, we evaluate the application of PCR-DGGE to investigate the yeast species associated with wine grapes. Initially, some variables associated with conducting PCR-DGGE analyses are examined. Then, the performance of PCR-DGGE in the analyses of yeasts on wine grapes at different stages of maturity is compared with cultural isolation on plates of malt extract agar (MEA).

2 Materials and methods

2.1 Yeast strains and yeast cell preparation

Yeast strains used in this study are listed in Table 1 and were cultured and maintained on malt extract agar, MEA (Oxoid, Melbourne). Their identity was confirmed by sequencing the D1/D2 domain of 26S rDNA [16,17]. Yeasts were grown in 5 ml of YM broth (Oxoid, Melbourne), at 25 °C and 200 rpm for 24 h, or other periods as specified. Culture (1.0 ml) was transferred to a 1.5-ml cryogenic tube and centrifuged at 16,000g for 2 min, 4 °C, to sediment the yeast cells. Cell pellets were stored at −20 °C until extraction of DNA for analysis by PCR-DGGE.

View this table:
Table 1

Reference cultures of yeasts and their sources

Yeast strainsSources
Metschnikowia pulcherrima CBS 5833AWRIa
Kluyveromyces thermotolerans 1738AWRI
Candida stellata 1159AWRI
Pichia membranifaciens 1148AWRI
Brettanomyces bruxellensis 1102AWRI
Saccharomyces cerevisiae 1004AWRI
Saccharomyces bayanus CBS 380AWRI
Aureobasidium pullulans FRR 5400Food Science Australiab
Pichia anomala FRR 2169Food Science Australia
Kluyveromyces marxianus FRR 1586Food Science Australia
Hanseniaspora uvarum FRR 2168Food Science Australia
Issatchenkia orientalis CR1UNSWc
Cryptococcus laurentii CM2UNSW
Rhodotorula slooffiae CM3UNSW
  • aThe Australian Wine Research Institute, Urrbrae, South Australia.

  • bCSIRO: North Ryde, NSW, Australia.

  • cFood Science and Technology, University of New South Wales, Sydney.

In some experiments, cell suspensions containing a mixture of yeast species at different populations were prepared. Yeasts were grown in YM broth as described and then serially diluted with 0.1% peptone water. Samples (1.0 ml) of the appropriate species and dilution were mixed, and the cells sedimented by centrifugation (16,000g for 2 min) to give the mixed populations described in Table 2. The sedimented cells were stored at −20 °C until extraction of DNA. A similar approach was used to prepare suspensions containing mixtures of yeast cells taken from cultures after 12, 24, 48 or 96 h of incubation. The populations (cfu ml−1) of yeasts in the dilutions used to prepare the mixtures were determined by spread-plate culture on MEA.

View this table:
Table 2

Mixed populations of yeast species used to evaluate detection limits by PCR-DGGE

Yeast speciesPopulation mixtures (cfu ml−1)
H. uvarum106
R. slooffiae105105
B. bruxellensis102102103102
P. membranifaciens101101102101
A. pullulans103103106106106106105104
M. pulcherrima102102106105104103104103
C. stellata101101
  • + indicates DNA band detected by PCR-DGGE; − indicates DNA band not detected by PCR-DGGE.

2.2 Analysis of yeasts on grapes

Grapes were aseptically harvested from vineyards in the Hunter Valley, New South Wales, Australia. They were transported to the University of New South Wales, (3 h), stored overnight at 5 °C and then analysed. Samples consisted of healthy, undamaged grape bunches taken from at least five different vines and locations within the vineyard. Duplicate samples were collected. In some cases, bunches of grapes that were visibly damaged were also collected. Individual grape berries were randomly and aseptically removed from the bunches, and combined to give 50 g samples. These samples were rinsed in 450 ml of 0.1% peptone water with 0.01% Tween 80 by orbital shaking in a flask at 150 rpm for 30 min. The rinse was poured from the grapes and examined for yeasts by both culture and PCR-DGGE methods. There were only minor variations between duplicate grape samples. Data presented represent the average of the two samples, each analysed in duplicate.

Rinse samples were serially diluted in 0.1% peptone water, from which 0.1 ml was inoculated over the surface of plates of MEA. Plates were incubated at 25 °C for 4 days after which colonies were counted. Representatives of the different types of colonies were isolated and purified by restreaking onto MEA. They were identified by observation of cellular morphology and sequencing of the 26S rDNA. The remainder of the rinse (approximately 450 ml) was centrifuged at 16,000g for 15 min at 4 °C. The sedimented microbial cells were taken up in a small amount of 0.1% peptone water, transferred to a 1.5-ml cryogenic tube, centrifuged at 16,000g for 2 min at 4 °C and the cell pellet stored at −20 °C until extraction of DNA and analyses by DGGE.

One experiment examined the effect of the rinsing procedure on the recovery of yeasts from grapes. Grapes (50 g) were rinsed in 450 ml of 0.1% peptone–Tween 80 as described already, but for 10 min. The rinse solution was removed and analysed by culture and PCR-DGGE methods as previously described. The rinsed grapes were mixed with fresh 450 ml of 0.01% peptone–Tween 80 and rinsed for a further 10 min. The rinse was collected and analysed. This cycle of rinsing and analysis was repeated two more times.

2.3 DNA extraction from yeasts

DNA was extracted from pellets of yeast cells according to procedures described by Cocolin et al. [11]. Cells were resuspended in 200 μl of breaking buffer (2% Triton X-100, 1% SDS, 100 mM NaCl, 10 mM Tris, pH 8.0, 1 mM EDTA) and homogenised with 0.3 g of glass beads (0.5 mm in diameter) in a bead beater (Mini-BeadbeaterTM, Biospec product, Tasmania) at 6000 rpm for 1 min in the presence of 200 μl of phenol/chloroform/isoamyl alcohol (50:48:2). TE buffer (200 μl) (10 mM Tris, 1 mM EDTA, pH 7.6) was mixed with the disrupted cells and the suspension was centrifuged at 16,000g for 10 min, 4 °C. The supernatant was collected and 2.5 volumes of absolute ethanol were added to precipitate DNA. The DNA precipitate was sedimented by centrifuging at 16,000g for 10 min, 4 °C, washed with 70% ethanol, then resuspended in 50 μl of TE buffer.

Microbial pellets obtained from grape rinses were thawed and DNA was extracted according to the same procedure as described already, except that DNA in supernatants of the phenol–chloroform extracts was purified using a DNeasy kit (Qiagen Pty., Ltd., Canberra) according to the manufacturer's instructions. In the final purification step, DNA was totally eluted from the filter with 150 μl RE buffer supplied in the kit.

2.4 DNA amplification and primers

The D1/D2 domain of the 26S rDNA was amplified by a two-step, nested PCR. The first PCR was conducted with the forward primer NL1 (5-GCATATCAATA AGCGGAGGAAAAG-3) and reverse primer NL4 (5-GGTCCGTGTTTCAAGACGG-3) [18]. Amplification was done in a standard reaction mixture containing 10 mM Tris–HCl (pH 8.3), 20 mM KCl, 1.5 mM MgCl2, each deoxynucleotide triphosphate at a concentration of 0.2 mM, 1.25 IU of Taq polymerase, each primer at a concentration of 0.2 μM and the DNA template at a final concentration of 10 ng. PCR was run for 36 cycles with annealing at 52 °C, extension at 72 °C for 2 min, and denaturation at 94 °C for 1 min [16].

The amplicon (approximately 600 bp) from the first PCR was diluted and further amplified with a second PCR using the GC-clamp primer NL1 (5-CGCCCGCCGCGCGCGGCGGGCGGGGCGGGGGCCATATCAATAAGC-3) and forward primer LS2 (5-ATTCCCAAACAACTCGACTC-3) [11]. The conditions of this reaction were the same as those just described, except that the concentration of MgCl2 was increased to 2.25 mM. The PCR was run with an initial step at 95 °C for 5 min, and 30 cycles of denaturation at 95 °C for 1 min, annealing at 52 °C for 2 min, extension at 72 °C for 2 min, with a final extension at 72 °C for 7 min. Reagents for PCR were obtained from GeneAmp® (Perkin–Elmer), and the primers from Genset Pacific (Proligo Primer & Probe, Sydney). The reaction was conducted in a Perkin–Elmer C-960 air cooled thermal cycler (Corbett Research, NSW, Australia).

2.5 DGGE analysis

The GC-clamp PCR products were separated according to their sequences using DGGE with a DcodeTM Universal Mutation System (Bio-Rad Pty., Ltd., Sydney). PCR samples were directly applied onto 8% (w/v) polyacrylamide gels in a running buffer (1% TAE) containing 40 mM Tris–acetate, 2 mM Na2EDTA-H2O, pH 8.5. The gels were prepared with a denaturing gradient from 30–60% of urea and formamide [11] and a polyacrylamide and bis-acrylamide ratio of 37.5:1 or 19:1 (polyacrylamide mixing powder, Bio-Rad). Electrophoresis was performed at a constant voltage of 120 V for 4 h for the 37.5:1 gel, or for 6 h for the 19:1 gel with a constant temperature of 60 °C. After electrophoresis, the gels were stained in 1% TE buffer, pH 8, containing SYBR Green reconstituted according to the manufacturers directions (Molecular Probes, Eugene, Oregon) and photographed under UV translumination.

2.6 Sequence analysis of DNA bands and yeast identification

DNA bands from DGGE gels were carefully selected and excised from the gels using sterile razor blades. The pieces of gel were soaked in 20 μl of TE buffer overnight at room temperature to allow diffusion of DNA. Eluted DNA (10 μl) was amplified by PCR using NL1 and LS2 primers. The PCR products were labelled with dye terminators using the ABI Prism® BigDyeTM Terminator V3.1 Cycle Sequencing Ready Reaction Kit (PE Applied, Biosystems, Foster city, CA, USA). The labelled PCR products were purified by ethanol precipitation according to the manufacturer's instruction and sent to a commercial sequencing facility (Automated DNA Analysis Facility, UNSW). DNA base sequences were analysed by comparison with the GenBank databases of the National Center for Biotechnology Information (NCBI). Searches in GenBank with BLAST program were performed to determine the closest known relative of partial 26S rDNA sequences [19]. Yeast species were identified on the basis of their sequences (greater than 95% homology with the databases) [16,17]. Where appropriate, cell and budding morphology were examined.

3 Results

3.1 Optimisation of PCR-DGGE

The application of PCR-DGGE to the analysis of yeasts in wine and other ecosystems has so far been based on a one-step PCR amplification assay using the NL1 and LS2 primers, with the GC clamp being attached to the NL1 primer [1113]. This assay generates an amplicon of approximately 250 bp of the D1/D2 domain of the 26S rDNA. To improve the specificity and sensitivity of yeast detection, we investigated the use of a two-step, nested PCR assay. The first step using NL1 and NL4 primers generated an amplicon of about 600 bp within the D1/D2 domain of the 26S rDNA. The second step amplified a 250-bp sector within this amplicon using the NL1 and LS2 primers, with the GC clamp attached to the NL1 primer. Under comparable assay conditions, the nested PCR always gave more DNA product, as indicated by sharper, more intense DNA bands in electrophoresis gels (data not shown).

Separation and resolution of amplicons in PCR-DGGE analysis is generally done with polyacrylamide gels prepared using an acrylamide: bis-acrylamide ratio of 37.5 :1 [1113]. This ratio determines the extent of cross-linking and pore-size within the gel, and mobility of DNA fragments [2022]. Higher-molecular-weight fragments require gels with a larger pore size (e.g., less bis-acrylamide cross-linker) for effective mobility, but better resolution of the different sized fragments is obtained in gels with a smaller pore size. Since the amplicons produced from the D1/D2 domain of the 26S rDNA are relatively small (250 bp), we considered that they might be better resolved in polyacrylamide gels with a smaller pore size (e.g., acrylamide:bis-acrylamide ratio of 19:1 as used in DNA-sequencing gels). Fig. 1 shows the effect of gel cross-linking on the resolution of 26s rDNA amplicons for several reference yeasts (S. cerevisiae, S. bayanus, Metschnikowia pulcherrima and Aureobasidium pullulans) and for yeasts rinsed from the surface of wine grapes. The DNA bands obtained from the grape rinses were isolated and sequenced for identification. The main band gave sequences corresponding to A. pullulans. Sharper, more intense DNA bands were found in gels with the higher ratio (19 :1) of cross-linker. Moreover, for the grape samples, some additional bands were observed that were not found in gels prepared with the lower ratio (37.5:1). These bands were isolated and sequenced and gave the following identities: Greeneria unicola, Candida albicans, and Tremella encephala. The migration time of bands was slower in gels with the higher proportion of bis-acrylamide, and electrophoresis needed to be conducted for 6 h rather than 4 h, as commonly used for the lower-ratio gels.

Figure 1

Effect of polyacrylamide: crosslinker ratio on the resolution of 26S rDNA amplicons of yeast during DGGE (a) acrylamide cross-linker ratio 37.5:1; (b) acrylamide cross-linker ratio 19:1. Lanes: 1, Saccharomyces cerevisiae (S.c); 2, Aureobasidium pullulans (A.p); 3, Metschnikowia pulcherrima (M.p); 4, Saccharomyces bayanus (S.b); 5, yeasts rinsed from immature Shiraz grapes vineyard 1; 6, yeasts rinsed from immature Shiraz grapes, vineyard 2; 7, yeasts rinsed from mature Shiraz grapes vineyard 2. (G.u)=Greeneria unicola, (C.a)=Candida albicans, (A.sp.)=Aureobasidium sp. and (T.m)=Tremella encephala.

3.2 Detection limit of yeast populations by PCR-DGGE

The reference strains of yeasts were individually cultured to late-exponential phase (approximately 108 cfu ml−1) and then diluted in 0.1% peptone water to give final populations of approximately 101, 102, 103, 104, 106 and 107 cfu ml−1. Samples of the different dilutions were processed for DNA extraction, PCR and DGGE as described in Section 2. Dilutions with populations of 106–108 cfu ml−1 gave intensely-smeared DNA bands, while dilutions with 101–102 cfu ml−1 did not give obvious bands on DGGE gels. Dilutions containing 102–105 cfu ml−1 gave distinct bands, with more intense bands being observed for samples with 105 cfu ml−1 (data not shown). DNA preparations extracted from the dilutions containing 106–108 cfu ml−1 were diluted to give DNA concentrations of approximately 10 ng μl−1 and then used to conduct PCR-DGGE. Distinct DNA bands on DGGE gels were now observed for these samples. Since no bands were observed for populations less than 102 cfu ml−1, this represents the lower limit of detection under the conditions of our assay.

When reaction conditions were optimised, using 10 ng of template DNA (extracted from yeast populations of 108 cfu ml−1 and diluted), DGGE gave clear separation of the diversity of yeast species likely to be found on the surface of grapes (Fig. 2). Lane 15 (Fig. 2) shows the separation of DNA bands of seven yeast species from a mixed suspension. Each species was present at approximately 108 cfu ml−1 and the extracted DNA was diluted to about 10 ng μl−1 before PCR.

Figure 2

PCR-DGGE analysis of amplified 26S rDNA from reference yeast cultures. Lane: 1, Kluyveromyces thermotolerans (K.t); 2, Pichia anomala (P.a); 3, Issatchenkia orientalis (I.o); 4, Candida stellata (C.s); 5, Metschnikowia pulcherrima (M.p); 6, Aureobasidium pullulans (A.p); 7, Cryptococcus laurentii (C.l); 8, Hanseniaspora uvarum (H.u); 9, Brettanomyces bruxellensis (B.b); 10, Kluyveromyces marxianus (K.m); 11, Saccharomyces cerevisiae (S.c); 12, Saccharomyces bayanus (S.b); 13, Rhodotorula slooffiae (R.s); 14, Pichia membranifaciens (P.m); 15, cell mixture of yeasts in lanes 7–13.

The age of the yeast culture influenced the detection limit of yeast cells by PCR-DGGE. The data in Fig. 3(a), showing a detection limit of 102 cfu ml−1, was obtained for yeast cells harvested from a 24-h culture. When cells were harvested from a 96-h culture, PCR-DGGE assays could detect down to 101 cfu ml−1 (Fig. 3(b)). Although these cultures contained the same viable population (108 cfu ml−1) as the 24 h culture, they probably contained a higher population of dead cells, the DNA of which would also be detected by PCR-DGGE. This observation was obtained for several yeast species, including S. cerevisiae.

Figure 3

PCR-DGGE analysis of DNA extracted from different populations of Aureobasidium pullulans (a) cell populations from 24-h culture; (b) cell populations from 96-h culture. Lane 1, 106 cells; lane 2, 105 cells; lane 3, 104; cells; lane 4, 103 cells; lane 5, 102 cells; lane 6, 10 cells.

Because wine grapes are likely to harbour a mixture of yeast species at different relative populations, we examined the performance of PCR-DGGE in detecting individual species in cell suspensions containing various mixed populations. Table 2 shows the mixed populations of yeasts that were examined. Some corresponding DNA profiles after PCR-DGGE are shown in Fig. 4.

Figure 4

PCR-DGGE analysis of DNA extracted from mixed populations of different species of yeasts. See Table 2 for yeast populations in mixture. Lane 1, Hanseniaspora uvarum (H.u); lane 2, Rhodotorula slooffiae (R.s); lane 3, Kluyveromyces marxianus (K.m); lane 4, Saccharomyces cerevisiae (S.c); lane 5, Brettanomyces bruxellensis (B.b); lane 6, Pichia membranifaciens (P.m); lanes 7–18 mixed population A–L, respectively (from Table 2).

Mixture A contained six yeast species with populations decreasing from 106 cfu ml−1 (Hanseniaspora uvarum) to 101cfu ml−1 (Pichia membranifaciens). DGGE analysis only detected the three most prevalent species, namely, H. uvarum, Rhodotorula slooffiae, and Kluyveromyces marxianus. DNA from S. cerevisiae was not detected even though this species was present in the mixture at 103 cfu ml−1. Mixture B contained five species, decreasing in population from 105 cfu ml−1 (R. slooffiae) to 101 cfu ml−1 (P. membranifaciens). Again, DGGE detected only the three most prevalent species. In this case, S. cerevisiae (103 cfu ml−1) was detected. Mixture C contained only three species, decreasing from 104–102 cfu ml−1, and all species gave detectable DNA bands on DGGE. Mixture D was the same as mixture C except that cell population of each species was decreased by 10-fold. In this case, the species at the lowest concentration of 101 cfu ml−1 (P. membranifaciens) was not detectable.

Table 2 also shows various mixtures of S. cerevisiae, A. pullulans and M. pulcherrima. All three species were detected in mixture G that contained each species at 106 cfu ml−1. However, M. pulcherrima was not detected in mixtures H, I and J where its population had been decreased by 10–1000 times relative to other species that were kept at 106 cfu ml−1. However, M. pulcherrima at 103–104 cfu ml−1 could not be detected in mixtures (K and L) with 104–105 cfu ml−1 of S. cerevisiae and A. pullulans.

PCR-DGGE analysis was also done on mixtures of yeasts containing young and old cells. In these cases, A. pullulans was used as a source of old cells since wine grapes are known to harbour this species from the very early stages of their development until harvest time. Thus, significant populations of older cells of A. pullulans are likely to be present on grapes [4]. Cells (106 cfu ml−1) taken from a 96-h culture of A. pullulans were mixed with various populations (106, 105,104,103 cfu ml−1) of young cells from a 12-h culture of S. cerevisiae and examined by PCR-DGGE. Two bands of DNA, one for A. pullulans and one for S. cerevisiae were only recovered from the mixture containing 106 cfu ml−1 of both species (data not shown). The other mixture with lower populations of S. cerevisiae gave only the DNA band for A. pullulans. The DNA band for S. cerevisiae was absent. This result could be explained by the fact that the older culture of A. pullulans contained significant amounts of DNA from dead cells and this overloaded the PCR so that it could not amplify enough S. cerevisiae DNA when it was present at populations less than 106 cfu ml−1. When the same experiment was repeated using 106 cfu ml−1 from young cultures (24 h) of A. pullulans, DNA bands of S. cerevisiae could be detected in mixtures containing 106, 105 and 104 cfu ml−1 of this organism (data not shown). The same conclusions were found using mixtures of M. pulcherrima (old and young cultures) and S. cerevisiae.

3.3 Comparison of agar plating and PCR-DGGE methods for profiling the yeast ecology of wine grapes

Shiraz and Chardonnay grapes were sampled at different stages throughout cultivation and examined for yeasts by plating on MEA and by PCR-DGGE (Table 3 and Fig. 5(a) and (b)). Grapes at the early stage of development (peppercorn size) showed a predominance of A. pullulans, although at low populations (102–103 cfu g−1). It was detected by DGGE, but not when its population was less than 103 cfu g−1. Several Cryptococcus species were detected on MEA but at very low populations of approximately 102 cfu g−1. They were not detected by PCR-DGGE. However, PCR-DGGE analysis of the Shiraz grapes gave two DNA bands with sequences corresponding to the filamentous fungi Phialocephala scopiformis and Raciborskiomyces longisetosum. These species were not isolated on plates of MEA.

View this table:
Table 3

Yeast species associated with wine grapes as determined by plating on MEA and analysis by PCR-DGGE

Stages of grape cultivationGrape variety
Populations (cfu g−1)DGGEPopulations (cfu g−1)DGGE
Berry peppercorn size 10aAureobasidium pullulans (1.8 × 102)Aureobasidium pullulans (2.4 × 103)+
Cryptococcus laurentii (1.5 × 102)Cryptococcus oreinsis (1.5 × 102)
Cryptococcus oreinsis (1.0 × 102)
Cryptococcus magnus (1.0 × 102)
Veraison 4aAureobasidium pullulans (6.0 × 103)+Aureobasidium pullulans (2.6 × 104)+
Rhodosporidium babjevae (1.0 × 102)
Cryptococcus magnus (1.1 × 103)
Berry ripening 1aAureobasidium pullulans (1.0 × 104)+Aureobasidium pullulans (6.1 × 103)+
Rhodosporidium babjevae (2.8 × 103)
Cryptococcus victoriae (3.0 × 102)
HarvestAureobasidium pullulans (2.6 × 104)+Aureobasidium pullulans (2.6 × 104)+
Rhodotorula laryngis (2.5 × 102)
Cryptococcus victioriae (1.0 × 103)
Berry ripening (damaged) 1aMetschnikowia spp. (8.0 × 104)+Metschnikowia spp. (4.0 × 105)+
Aureobasidium pullulans (6.0 × 103)Aureobasidium pullulans (3.0 × 104)+
Rhodosporidium babjevae (1.0 × 104)Rhodosporidium babjevae (2.0 × 104)
Cryptococcus laurentii (2.0 × 104)
Harvest (damaged)Metschnikowia spp. (6.0 × 105)+Metschnikowia spp. (9.0 × 106)+
Aureobasidium pullulans (2.0 × 105)+Hanseniaspora spp. (5.0 × 106)+
Aureobasidium pullulans (2.0 × 104)+
  • aWeeks before harvest; Metschnikowia spp.=Metschnikowia pulcherrima and Metschnikowia spp.; Hanseniaspora spp.=Hanseniaspora uvarum and Hanseniaspora guilliermondii; P.s=Phialocephala scopiformis; R.l=Raciborskiomyces longisetosum; A.spp.=Aureobasidium spp.

Figure 5

PCR-DGGE analysis of microbial DNA from rinses of grape berries sampled from vineyards at different stages during cultivation. (a) Shiraz grapes; (b) Chardonnay grapes. Lane 1, peppercorn size berries; lane 2, berries during veraison; lane 3, berries seven days before harvest; lane 4, berries at harvest; lane 5, damaged berries 7 days before harvest; lane 6, damaged berries at harvest. DNA bands were isolated and sequenced. Species identities are Hanseniaspora uvarum (H.u), Hanseniaspora guilliermondii (H.g) Metschnikowia spp. (M.spp.), Metschnikowia pulcherrima (M.p), Aureobasidium pullulans (A.p), Aureobasidium spp. (A.spp.), Phialocephala scopiformis (P.s) and Raciborskiomyces longisetosm (R.l).

Aureobasidium pullulans continued to be dominant on both grape varieties until harvest. Its population increased to 103–104 cfu g−1, and was consistently detected by PCR-DGGE. Moreover, DGGE showed the presence of additional species of Aureobasidium on Chardonnay grapes that were not evident by plating on MEA. Species of Rhodotorula, Rhodosporidium and Cryptococcus were also isolated on MEA for the Chardonnay grapes. However, these populations ranged from 102 to 104 cfu g−1, and were not detected by PCR-DGGE.

Damaged grape berries, sampled at the time of harvest and a week before harvest (Table 3, Fig. 5), gave a predominance of Metschnikowia spp. (105–107 cfu g−1), and also Hanseniaspora spp. (106–107 cfu g−1) for the Chardonnay grapes. A. pullulans was also found at 104–105 cfu g−1. These three genera were detected by PCR-DGGE, where multiple bands indicated the occurrence of three species of Metschnikowia (M. pulcherrima and two unidentified species) and two species of Hanseniaspora (H. uvarum and H. guilliermondii). The diversity of species within these two genera was not evident from colony morphology on plates of MEA. Rhodosporidium and Cryptococcus species were isolated from damaged grapes (one week before harvest) at populations of 103–104 cfu ml−1, but were not detected by PCR-DGGE.

Further comparisons of cultural and PCR-DGGE methods were done using samples of undamaged mature grapes and damaged mature grapes. With undamaged Shiraz grapes, A. pullulans at 6 × 102 cfu g−1 was found by plating on MEA. On PCR-DGGE, this sample gave one band that corresponded to A. pullulans. However, damaged grapes from this sample gave M. pulcherrima (3 × 106 cfu g−1), Cr. laurentii (1 × 104 cfu g−1) and A. pullulans (1 × 103 cfu g−1) on MEA, but only Metschnikowia spp. were found by PCR-DGGE. Three DNA bands were detected, one corresponding to M. pulcherrima and two other bands that gave sequences corresponding to Metschnikowia species, which did not match descriptions in the databases (data not shown).

Our standard analytical procedure was based on rinsing the grapes for 30 min followed by examination of yeasts in this rinse fluid. We thought that this time of rinsing might increase the possibility of releasing PCR-inhibitory substances into the rinses. To investigate this possibility, samples of undamaged and damaged Chardonnay grapes were successively rinsed four times, with 10 min for each rinse. Each rinse was analysed for yeasts by plating on MEA and by PCR-DGGE (Table 4, Fig. 6).

View this table:
Table 4

Effect of rinsing on the isolation and detection of yeasts from wine grapes by cultural and PCR-DGGE methods

Rinsing operation (10 min)Grape samples
Undamaged Chardonnay berriesDamaged Chardonnay berries
Populations (cfu g−1)DGGEPopulations (cfu g−1)DGGE
1Aureobasidium pullulans (5 × 103)+Metschnikowia spp. a (8 × 106)+
Cryptococcus laurentii (4 × 102)+
Rhodotorula slooffiae (1 × 102)Hanseniaspora. spp.b (4 × 106)+
Aureobasidium pullulans (3 × 104)
Cryptococcus laurentii (1 × 104)P.s
Total population (5.5 × 103)2.1 × 106Total population (1.2 × 107)5.4 × 109
2Aureobasidium pullulans (9 × 102)+Metschnikowia spp. a (2 × 105)+
Cryptococcus laurentii (3 × 102)Hanseniaspora spp. b (9 × 104)+
Rhodotorula slooffiae (1 × 102)Aureobasidium pullulans (2 × 103)+
Cryptococcus laurentii (2 × 103)
Total population (1.4 × 103)6.3 × 105Total population (2.9 × 105)9.9 × 107
3Aureobasidium pullulans (1 × 103)+Metschnikowia spp. a (2 × 105)+
Hanseniaspora spp. b (8 × 104)+
Aureobasidium pullulans (7 × 103)+
Cryptococcus laurentii (1 × 103)
Total population (1.0 × 103)5.4 × 105Total population (2.8 × 105)8.1 × 106
4Aureobasidium pullulans (1 × 102)+Metschnikowia spp. a (2 × 104)
Hanseniaspora spp. b (1 × 103)
Aureobasidium pullulans (6 × 102)
Total population (1.0 × 102)4.5 × 103Total population (2.1 × 104)6.9 × 105
  • a Metschnikowia spp.: Metschnikowia pulcherrima and Metschnikowia spp.

  • b Hanseniaspora spp.: Hanseniaspora uvarum and Hanseniaspora guilliermondii P.s=Phialocephala scopiformis.

Figure 6

PCR-DGGE analysis of yeast populations rinsed from the surface of Chardonnay grapes. (a) Undamaged grapes (b) damaged grapes Lane 1, first rinse; lane 2, second rinse; lane 3 third rinse; lane 4, fourth rinse. DNA bands were isolated and sequenced. Species identities are Hanseniaspora uvarum (H.u), Metschnikowia spp. (M. spp.), Phialocephala scopiformis (P.s) and Aureobas pullulans (A.p).

The first rinse released about 80% of the total surface yeasts for the undamaged grapes and 96% for the damaged grapes (Table 4). The remaining yeasts were removed by subsequent rinses. As determined by culture on MEA, A. pullulans was the predominant species found in each rinse for the undamaged berries. Its population decreased from 5 × 103 cfu g−1 in the first rinse to 1 × 102 cfu g−1 in the fourth rinse. Cr. laurentii and R. slooffiae were also isolated on MEA, but lesser populations (4 × 104 and 1 × 102 cfu g−1) were found, and they were only recovered in the first and second rinses. A. pullulans was the only species detected by PCR-DGGE, but it was not detected in the fourth rinse where its population had decreased to 102 cfu g−1 (Fig. 6(a)).

Damaged grapes had much higher populations of yeasts (106–107 cfu ml−1), which were dominated by the presence of Metschnikowia and Hanseniaspora species, but notable populations of A. pullulans and Cr. laurentii were also recovered. The recovery of each species decreased with each rinsing operation (Table 4). With the exception of Cr. laurentii, all species were detected by PCR-DGGE (Table 4, Fig. 6(b)), but only in rinses 1–3. Notably, Cr. laurentii was not detected by PCR-DGGE, even though it was present at 104 cfu ml−1 in the first rinse. Overall, we could see no influence of rinsing procedures on the efficacy of the PCR-DGGE analyses.

4 Discussion

The diversity of yeast species associated with foods and beverages is usually determined by culturing homogenates of the product on plates of agar media. Yeast colonies are then enumerated, isolated and identified using standard morphological, biochemical and physiological tests, based on culture methods [23,24]. Recently, various molecular methods have been developed for faster, more definitive yeast identification [25]. Sequencing of the D1/D2 of the large-subunit 26S ribosomal DNA is now widely accepted as a standard procedure for yeast identification [16,17]. Culture on agar media still remains as the main approach for isolation of yeasts from natural habitats. However, culture-independent strategies, based on the extraction and analysis of DNA sequences are attracting increasing interest and have the advantage of detecting the occurrence of viable but non-culturable species [9,10]. It has been suggested that, in some natural habitats, as much as 90–99% of the microflora may be viable but not culturable by agar plating [2628]. PCR-DGGE has emerged as one of the more promising culture-independent molecular methods for profiling the microbial ecology of habitats [9,29], and it has been applied to food and beverage ecosystems [10,27,28,30] including wine [1113].

In this study, we evaluated the application of PCR-DGGE in profiling the yeast ecology of wine grapes by comparing its performance with cultural methods. Many variables are involved in the application of PCR-DGGE to environmental microbiology. These variables need to be recognised, optimised and managed in order to obtain reliable ecological data. The use of PCR-DGGE in food and beverage microbiology is a relatively new development, and increased understanding of the impact of these variables is required. Extraction and purification of DNA from the sample is the first step, and a significant variable in PCR-DGGE technology. We chose to extract DNA from microbial cells rinsed from the surface of grapes. Initially, we used homogenates of grapes as the starting material. Using agar culture methods, these homogenates gave yeast populations similar to those obtained in rinses (data not presented). However, no PCR amplicons and DNA bands on electrophoresis gels were obtained using homogenates as starting material and the DNA extraction and purification scheme described in Section 2. Possibly, this protocol was not adequate to produce DNA free of other plant components. Many plant constituents (e.g., polysaccharides, polyphenols) are known to inhibit PCR [31], and this may explain our observations. Also, the large amount of plant DNA relative to microbial DNA in the PCR assay could have adversely affected the kinetics and specificity of the reaction [31,32].

The conditions of the PCR assay involve many variables (e.g., reagent mixture and concentrations, buffer, pH, time, temperature). We made no attempt to investigate or optimise these conditions and followed the basic protocol reported by others [11,16]. However, we found that a two-step, nested-PCR assay improved the quality and specificity of the DNA amplicons for DGGE. The merits of a two-step, nested-PCR, (compared with a single-step PCR), have been reported elsewhere [3335]. To date, most studies on the application of DGGE to yeasts have used a single-step PCR [1113], [30]. We also noted the importance of controlling the initial concentration of template DNA. PCR assays are governed by the principles of enzyme kinetics, which include the initial concentration of substrates. Excessive template DNA can cause smearing of DNA bands on DGGE gels, as well as inhibition of PCR [31,32]. It was important to keep the initial concentration of template DNA at approximately 10 ng in the PCR assay, as has been reported generally [11,16,32]. High populations of yeast cells (e.g., greater than 107 cfu ml−1) give too much template DNA and, therefore, dilution is needed to give an acceptable PCR-DGGE outcome. The importance of this variable is generally unrecognised or not reported in the PCR-DGGE literature.

The conditions of gel electrophoresis present another suite of variables that can impact on the detection and resolution of DNA bands [2022]. We did not study all these variables. However, decreasing the pore size of the gels used for DGGE gave increased resolution and sharper, more intense bands of the DNA amplicons. Running the loaded DGGE gels at 30 V for 30 min, prior to starting the standard operating conditions, also gave improved band resolution. The use of stacking gels before the main separation gel is another factor that could be used to improve resolution [22].

The detection limit of analytical methods is an important criterion in determining the microbial ecology of foods. Spread-plate culture methods can generally detect yeast populations of 101–102 cfu ml−1 or g−1 of product. Using exponential-phase cultures of single yeast species, our optimised PCR-DGGE method could consistently detect populations as low as 102 cfu ml−1. This compares favorably with agar plating. Other studies have reported 103 cfu ml−1 as the lower detection limit for yeasts by PCR-DGGE [1113]. As mentioned previously, it is important to dilute the yeast population or extracted DNA to prevent overloading the PCR assay with template DNA. The detection limit of individual yeast species in mixed populations, however, was also determined by their relative populations. At equal populations, it is possible to detect all species in the mixture. Moreover, it is possible to detect species that are present at 10–100-fold less than others in the mixture, but not when this ratio exceeds 100-fold (Table 2). Thus, a species present in a mixture at 103 cfu ml−1 will not detected if other species are present at 106 cfu ml−1or more. Also, it may not be possible to detect a species present at 104 cfu ml−1, for example, if several other species are present at 105–106 cfu ml−1. The kinetics of the PCR assay are complex, but detection of any particular species in a mixture will depend on hybridisation of the molecules of primer DNA with its template DNA, and the affinity of the primer for homologous sequences in this DNA [33]. Species present at higher populations in the mixture will give greater amounts of template DNA, and, therefore, should have a higher probability of detection. However, it would be difficult to predict the hybridisation affinity of the primer for the different types of template DNA. Another factor would be the extraction efficiency of the template DNA, which could vary depending on the extent to which different yeast species are disrupted and on the susceptibility of their cell walls to physical disintegration. Basidiomycetous yeasts are, in general, more resistant to disruption than ascomycetous yeast [36].

Aureobasidium pullulans was the main species isolated from undamaged, immature and mature grapes. Various species of Cryptococcus, Rhodotorula and Rhodosporidium were also found, especially on Chardonnay grapes. Yeast populations on grapes were quite low, increasing from 102 to 103 cfu g−1 on immature berries to 103–106 cfu g−1 on mature berries. These population data are consistent with other ecological studies of wine grapes [4]. Using PCR-DGGE, we could not detect A. pullulans in rinses of grapes when its population was less than 103 cfu g−1. None of the Cryptococcus, Rhodotorula and Rhodosporidium species were detected by PCR-DGGE even when their populations were as high as 104 cfu g−1. As mentioned already, PCR-DGGE can detect these yeasts at populations as low as 102 cfu g−1 when pure cultures are used as the assay sample. It is possible that inhibitory substances were extracted from the grapes with the rinses, thus decreasing the efficiency of DNA amplification by PCR and, therefore, detection sensitivity. The inhibition of PCR assay by plant lipids, polysaccharides and polyphenols has been reported before [31]. Residues of pesticides that are routinely sprayed onto grapes during cultivation might also interfere with the PCR assay. The rinses contained a mixture of yeast species and, as described previously, this can decrease the detection sensitivity of individual species by PCR-DGGE. Cryptococcus, Rhodotorula and Rhodosporidium are basidiomycetous yeasts and the extraction of their DNA may have been less efficient.

Damaged grape berries had much larger yeast populations (105–107 cfu g−1) and gave a predominance of Metschnikowia and Hanseniaspora species, as well as significant populations of A. pullulans. With populations at 105–107 cfu g−1, the Metschnikowia and Hanseniaspora species were readily detected by PCR-DGGE, although there was one occasion when Metschnikowia species at 104 cfu g−1 were not detected in damaged berries. We suspect that damaged berries may give greater amounts of PCR-inhibiting substances.

There were several occasions where the presence of yeast species in grape rinses was detected by PCR-DGGE, but not found by culture on agar media. The DNA from two fungal species, Phialocephala scopiformis and Raciborskiomyces longisetosum, frequently associated with bark or wood [37], were detected in some samples. Presumably, these grape berry samples came in contact with the stem of the grapevines, or wood of the vine trellis. The occurrence and significance of these fungi on grapes have not been reported before. In addition to A. pullulans, several grape samples gave DNA bands with sequences corresponding to other species of Aureobasidium. These sequences did not match known species in the databases, but were closely related to A. pullulans. These species were not easily differentiated as colonies on agar plates. Similarly, several DNA bands were found for Metschnikowia species and two DNA bands were given by Hanseniaspora species. However, only M. pulcherrima and H. uvarum were recovered by agar culture.

We concluded that PCR-DGGE analysis is less sensitive than agar culture for determining the yeast ecology of grapes, and would not reliably detect species present at populations less than 104 cfu g−1. However, PCR-DGGE analysis offers the prospect of detecting a greater diversity of species than agar culture, and provides information in a shorter time. Consequently, both methods should be used in parallel for profiling the yeast ecology of wine grapes.

In addition to evaluating PCR-DGGE analysis, our studies revealed some new information about the yeast ecology of wine grapes. According to many previous studies, wine grapes prior to full maturity and harvest, harbor a dominance of Cryptococcus, Rhodotorula, Rhodosporidium and Sporobolomyces species [4,38]. While we frequently isolated species of Cryptococcus and Rhodosporidium, species of Aureobasidium were the predominant organisms, and persisted on the grapes up until harvest. This is a relatively novel observation that requires further investigation to determine if it is a regional or more general phenomenon, and to determine the ecological significance of Aureobasidium species on wine grapes. A. pullulans has notable antagonistic activities against other yeasts and fungi [39] and, in this context, it could influence the overall microbial ecology of wine grapes.

As grapes mature, they become more susceptible to physical damage [40,41]. Consequently, it is not uncommon to find a mixture of damaged and undamaged berries on bunches of mature grapes at the time of harvest. We were particularly careful to separate damaged and undamaged berries for yeast analysis. As reported in many previous studies, species of Hanseniaspora (Kloeckera) and, to a lesser extent, species of Metschnikowia and Candida, predominate on wine grapes at the time of harvest [4,8,38,40]. We could only confirm this observation for the analysis of damaged grape berries. Healthy, undamaged grape berries at maturity did not harbor Hanseniaspora, Metschnikowia or Candida species, and A. pullulans was the predominant organism. Some other studies have also reported an absence of Hanseniaspora and Metschnikowia species on wine grapes at the time of harvest [4,42]. The importance of berry damage in determining the yeast ecology of wine grapes has been mentioned by others [40,43] although this variable has not been systematically investigated. Our data reinforce the importance of separating damaged and undamaged berries for studies on the yeast ecology of grapes, and highlight the need for more detailed examination of this variable and its importance in wine microbiology.


The authors are grateful to the Australian Grape and Wine Research Development and Corporation (GWRDC) for providing funds for this research and the Royal Thai Government for providing a student scholarship to Cheunjit J. Prakichaiwattana.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
View Abstract