Correspondence: Marilena Budroni, Dipartimento di Scienze Ambientali Agrarie e Biotecnologie Agroalimentari, Sezione di Microbiologia generale ed applicata, Facoltà di Agraria, Università degli Studi di Sassari, Viale Italia 39, 07100 Sassari, Italy. Tel.: +39 079 229 314; fax: +39 079 229 370; e-mail: firstname.lastname@example.org
A strain of Saccharomyces cerevisiae having desirable winemaking properties and high spore viability was bred from a semi-homothallic parent strain with similar winemaking properties but that produced sixfold fewer viable spores. Because the parent was homozygous for HO and for the MATa allele at both silent HMR and HML loci, it produced two MATa and two nonmating progeny per ascus. To obtain a segregant able to mate with the stable MATa progeny, a strain of the nonmating progeny, previously subjected to HO distruption with a KanMX4 cassette, was used. The resultant MATαho::KanMX4 transformant was mated to a MATa HO segregant and the diploid produced was sporulated to allow the isolation of a semi-homothallic diploid segregant designated 2D that lacked the KanMX4-disrupted HO allele as confirmed by sequence analysis. Genetic analysis indicated greater homozygosity in 2D than in the parent as assessed by PCR at five loci. The sugar consumption profiles of both 2D and the parent in grape juice fermentations were the same. Acetaldehyde levels and postfermentation biofilm formation were higher in 2D than in the parent. Because 2D has acceptable winemaking characteristics but produces significantly more viable spores than the parent strain, it will be useful in future breeding efforts.
The search for new wine yeast starter cultures generally relies on the isolation of wild yeasts from grapes, musts, or wines, characterization of the isolates, and selection of the most suitable ones. Currently, the limits of this procedure have been attained for most desirable traits (Cebollero et al., 2007). This has led wine microbiologists to look for alternative strategies, based on classical genetic methods (mating, spheroplast fusion, or cytoduction) and/or recombinant DNA technology, for the improvement of specific enological properties (Pretorius, 2000). The utilization of classical genetic methods is limited by the complexity and instability of wine yeast genome (Ibeas & Jimenez, 1996; Pretorius, 2000; Ramirez et al., 2004). Most Saccharomyces cerevisiae wine strains have a high propensity for genetic instability (Pretorius, 2000), meaning that their winemaking characteristics and fitness may change over a short period of time. These modifications are caused by mitotic and meiotic recombinations, which may occur during fermentation (Mortimer et al., 1994; Aguilera et al., 2000; Johnston et al., 2000; Mortimer, 2000; Puig & Perez-Ortin, 2000; Fernandez-Espinar et al., 2001; Marullo et al., 2004, Ramirez et al., 2004). Furthermore, the genetic stability is related to ploidy (Galitski et al., 1999) and to the degree of heterozygosity. Most yeast strains are diploid, some are aneuploid, and some polyploid (Pretorius, 2000), and homothallism represents the most common life cycle among wine strains (Johnston et al., 2000). However, some S. cerevisiae flor strains are characterized by a semi-homothallic life cycle (Budroni et al., 2000). Semi-homothallic yeasts produce two mating and two nonmating meiotic derivatives per ascus (Budroni et al., 2005). It has been demonstrated that the semi-homothallic behavior often relies on the presence of the same information encoded by Ya (or Yα) at both silent loci HML and HMR. This produces a futile interconversion in the mating spores, as the same information already present at the MAT locus is transposed (Santa Maria & Vidal, 1970; Pirino et al., 2004). Consequently, the two mating spores are unable to auto-diploidize and therefore generate populations characterized by stable ploidy (Budroni et al., 2005). The mating derivatives can be crossed with heterothallic or semi-homothallic strains of the opposite mating type for the selection of desirable phenotypes. However, semi-homothallic strains may not be easily amenable to genetic improvement with classical genetic methods because of their poor spore viability (Budroni et al., 2005). In the present study, we took advantage of the semi-homothallic life cycle in order to obtain a homozygous S. cerevisiae flor strain characterized by genetic stability and high spore viability that will be useful for further breeding.
Materials and methods
The S. cerevisiae strains used in this study belong to the culture collection of Dipartimento di Scienze Ambientali Agrarie e Biotecnologie Agroalimentari, (DISAABA) Università di Sassari (Table 1).
Table 1. Saccharomyces cerevisiae strains
The MI1C and Δ3B strains are considered heterothallic due to their mating ability.
Δ3B and MI1C were precultured in YPD (2% glucose, 2% peptone, and 1% yeast extract) at 25 °C for 48 h. After centrifugation (5 min at 1200 g), the cell pellet was resuspended in 2 mL of sterile water, and 1 mL aliquots were inoculated into 30 mL of 0.67% yeast nitrogen base (YNB) (Sigma-Chemical Co., St Louis, MO)+4% ethanol (Zara et al., 2002) or 30 mL of Vernaccia must (19% glucose+fructose; pH 4.3) in 250 mL flasks for performing cell–cell crosses. After static incubation for 15 days at 20 °C, cells taken from the biofilm formed on the top of the liquid media were streaked on the sporulation medium K acetate (1.5% potassium acetate, 2% agar). Asci were digested with 0.8 mg mL−1 lyticase (Sigma-Chemical Co.) in 200 μL TE for 20 min at room temperature. Tetrads were dissected using an MSM micromanipulator (Singer Instrument Co. Ltd, Somerset, UK). Single spore derivatives were incubated on YPD plates for 4 days at 30 °C and replica-plated onto YPD+0.2 g L−1 G418 (Sigma-Chemical Co.) and K acetate. Only G418-sensitive cultures were subjected to phenotypic characterization.
G418-sensitive cultures were characterized on the basis of fermentation trials on a glucose-rich medium (28% glucose 1% yeast extract, 1% peptone, and 1% ammonium phosphate). Ethanol and foam production were determined as described (Zambonelli, 2003). Acetaldehyde production was measured using an Acetaldehyde Assay Kit (Megazyme, Bray, Ireland). Acetic acid production was evaluated by HPLC Waters Model 510 equipped with a Waters UV-481 detector on a Alltech column IOA-1000 ORGANIC (at 70 °C). HPLC operating conditions were H2SO4 0.01 N mobile phase, 0.4 mL min−1 flow rate, 20 μL injection volume, and detection at 210 nm.
PCR, restriction analysis, and DNA sequencing
Yeast strains were precultured in YPD at 25 °C overnight on a rotary shaker (200 r.p.m.). Total DNA was extracted as described (Philippsen et al., 1991). PCR primers HOleft (5′-CGAATAAAGTCGCGGAAAAA-3′) and HOright (5′-AATCGAAGACCCATCTGCTG-3′) were based on the S. cerevisiae S288c sequence. The HO gene was amplified using a MyCycler Thermal cycler (BioRad, Milano, Italy) in a 50-μL reaction mixture containing 1 × PCR buffer, 2 mM MgCl2, 0.2 μM of each primer, 1 U of Taq polymerase (Eppendorf, Hambourg, Germany), and 50 ng of DNA template. The reactions were run for 35 cycles as follows: 95 °C for 45 s, 51 °C for 45 s, and 72 °C for 2 min. An initial denaturation step at 95 °C for 4 min and a final elongation step at 72 °C for 10 min were carried out. HO amplicons were subjected to restriction analysis using BfaI (New England Biolabs, Frankfurt am Main, Germany) in a total volume of 20 μL at 37 °C for 2.5 h. PCR products and restriction fragments were visualized using a Chemi Doc Imaging System (BioRad) after electrophoretic separation at 6 V cm−1 for 120 min in a 1.5% agarose gel stained with ethidium bromide. Sequencing of HO alleles from strains 2D and M25 was carried out by DNA BMR Genomics s.r.l. service (http://www.bmr-genomics.it/) using the following set of primers: HOseq1 (5′-ACTTCGAATAAAGTCGCG-3′); HOseq451 (5′-TATCAGCGTCTTGCATTA-3′); HOseq901 (5′-ACAACAAAAGAGCCAGAA-3′); HOseq1352 (5′-TGACGACCAGGTCAGCTA-3′); HOseq1801 (5′-AATTGCAAGTATGTACCA-3′); and HOseq2264 (5′-GTGAATTTTATTTTAT-TAAGG-3′). The nucleotide sequences were deposited in the EMBL Nucleotide DataBase (http://www.ebi.ac.uk/embl/) and given the following Accession numbers (2D: AM921676; M25: AM921677).
Life cycle analysis
Yeast cells were precultured in YPD at 25 °C overnight on a rotary shaker (200 r.p.m.), transferred to K-acetate plates, and incubated at 25 °C. After 15 days, asci were dissected and single spore cultures were incubated on YPD plates at 25 °C for 3 days. Sporulation frequency and viability were evaluated by microscopic observation and quantified as follows: (1) sporulation frequency=no. of asci/no. of asci+no. of cells; (2) spore viability=no. of viable spores/no. of asci analyzed × 4 (Mortimer et al., 1994). Only spores derived from complete asci (four viable spores) were transferred to K-acetate plates for 15 days to assess the spo+ : spo− ratio. Cultures able to sporulate (spo+) were further dissected and single spore derivatives were transferred to K-acetate plates.
Exponential-phase yeast cells were prepared as described (Bradbury et al., 2006) and DNA content was determined using an Epics XL (Beckman Coulter) equipped with an argon laser (15 mW, excitation wavelength 488 nm). Propidium iodide fluorescence was recorded on a linear scale and the median fluorescence of the G1 peak was registered after gating the dominant cell population on forward scatter/side scatter profile (Bradbury et al., 2006). YPH501 and YPH500 were used as reference strains for 2n and n DNA contents, respectively. Typically, 15 000 cells were analyzed per sample. The results obtained are means of two independent experiments.
Amplification of AGA1, DAN4, FLO11, HSP150, and SED1 genes
Yeast strains were precultured in YPD at 25 °C overnight on a rotary shaker (200 r.p.m.). Total DNA was extracted as described (Philippsen et al., 1991). The AGA1, DAN4, SED1, and HSP150 genes were amplified by PCR as described (Marinangeli et al., 2004). FLO11 was amplified according to Zara et al. (2005). All PCR products were visualized in a Chemi Doc Imaging System (BioRad) after electrophoretic separation at 6 V cm−1 for 120 min in a 1.5% agarose gel stained with ethidium bromide.
Fermentation trials in bench-top fermentors
Yeast cells were precultured overnight in YPD at 30 °C on a rotary shaker (200 r.p.m.) and added to 1.5 L of sterile Vernaccia must (23% glucose+fructose; pH 3.7) to an initial concentration of 105 cell mL−1 in a 2.5-L bench-top fermentor (Biocontroller AD1010 connected to a Bioconsole AD1025; Applikon Italia S.r.l., Genova, Italy). Temperature, pH, and oxygen concentration were monitored throughout the fermentation. Yeast growth was determined by plate counts. Residual sugar was determined using a d-glucose d-fructose Assay Kit (Megazyme, Bray, Ireland).
Biofilm formation assays
Biofilm formation ability was evaluated both on 0.67% YNB+4% ethanol and Vernaccia wine as described (Budroni et al., 2000). Quantitative determination of biofilm formation was performed as described (Reynolds & Fink, 2001). Briefly, cells were grown overnight at 30 °C in SC medium (0.67% YNB, 2% glucose), washed once in sterile water, resuspended in 0.67% YNB+0.1% glucose, and brought to an OD600 nm=1.0. For each strain, 100 μL aliquots were transferred into 10 wells of a polystyrene plate and the cell suspensions were incubated at 30 °C for 3 h. An equal volume of 1% (w/v) crystal violet was added. After 30 min, the wells were washed with sterile water. Adherence of cells was quantified by solubilizing the retained crystal violet using 100 μL of 10% sodium dodecyl sulfate and an equal volume of sterile water after 30 min. Finally, 50 μL of the solution was transferred to a fresh polystyrene 96-well plate and biofilm formation was quantified at A570 nm. The nonadhering S. cerevisiae strain S288c was used as negative control.
Unless otherwise stated, all experiments were performed in triplicate from independent precultures. Statistical analyses of the data were performed using anova, followed by Student's t-test (two-sided) using jmp version 3.1.5 software (SAS Institute Inc.). Differences were considered to be significant if P values were <0.05.
Construction of strain 2D
M25 is a semi-homothallic wine strain of S. cerevisiae that, due to its valuable winemaking characteristics, is widely used for making Vernaccia wines. This strain is also characterized by low sporulation frequency and poor spore viability (Budroni et al., 2000). In order to obtain a strain that retains the valuable winemaking characteristics of M25, but produces more viable spores, two M25 meiotic derivatives, MI1C and Δ3B, were mated (Table 1; Fig. 1). The diploid obtained (Δ3B × MI1C) was then sporulated. Eight complete asci were dissected, which allowed isolation of 32 spores that segregated 2+ : 2− for G418 sensitivity : resistance. The 16 G418-sensitive spores were homothallic, and hence able to autodiploidize and generate homozygous diploid strains, except at the MAT locus. The diploid cultures generated by diploidization of each of the 16 G418-sensitive spores were characterized for their winemaking ability, and strain 2D was selected for further use.
The HO gene was amplified by PCR in strains 2D, M25, Δ3B, and MI1C and subjected to restriction fragment length polymorphism (RFLP) analysis. RFLP analyses of the PCR products showed identical restriction patterns in strains M25, 2D, and MI1C (Fig. 2). Sequencing of the alleles in 2D and M25 indicated complete homology and absence of the KanMX4 cassette in 2D strain (data not shown).
Genetic characterization of 2D
The sporulation frequency and spore viability of M25 and 2D are reported in Table 2. The results obtained by analyzing the segregation ratio spo+ : spo− indicated that strain 2D maintained the semi-homothallic behavior of the parental strain M25 and hence generated two MATa mating spores and two nonmating MATα spores per ascus. Strain 2D showed a dramatic increase in spore viability (Table 2). While M25 did not produce any asci with four viable spores, 35% of the asci produced by 2D contained four viable spores. Flow cytometric analysis of propidium iodide-stained cells indicated that the 2D and M25 strains had DNA contents that were not significantly different from that of the diploid strain YPH501 (P<0.05). Relative to the fluorescence value of 1.0 assigned to strain YPH501, that of the haploid strain YPH500 was 0.47±0.04 and those of 2D and M25 were 1.07±0.02 and 1.09±0.04, respectively (Fig. 3).
Table 2. Sporulation frequency and spore viability of strains M25 and 2D
Sporulation frequency (%)
Spore viability (%)
Because the 2D strain originates from diploidization of a nonmating spore, it is expected to be completely homozygous, except at the MAT locus. To confirm this, five genetic markers (AGA1, DAN4, FLO11, HSP150, and SED1) were analyzed in the 2D and M25 strains. The results indicated that whereas M25 is heterozygous for DAN4, FLO11, and SED1, confirming previously observed heterozygosity (Giordano, 2000), strain 2D is homozygous at all five loci (Fig. 4).
Fermentative behavior and biofilm-forming ability of 2D
The winemaking properties of 2D and M25 were evaluated during the fermentation of Vernaccia must. The two strains displayed almost identical growth kinetics and oxygen and sugar consumption profiles (Fig. 5). An important technological feature of flor yeasts is their ability to form a biofilm on the wine surface in half-empty barrels. Contact with air and absence of sugar triggers a switch from fermentative to oxidative metabolism, thus allowing cells to produce acetaldehyde and other volatile metabolites characteristic of the aroma of sherry-like wines. We thus compared the biofilm-forming ability of 2D with that of the parental strain M25. Strain 2D was found to be able to form a complete biofilm on the Vernaccia wine 2 or 3 days postfermentation. The biofilm (velum) formed by 2D was thicker than that produced by M25 (Fig. 6). Quantitative analyses indicated that although both strains were able to adhere to polystyrene plates, strain 2D had significantly greater adherence values (A570 nm=1.67±0.21) than M25 (A570 nm=1.32±0.15). Acetaldehyde levels measured in wine were significantly higher in samples fermented by 2D (73.50±5.30 mg L−1) than in those fermented by M25 (65±2.20 mg L−1) (P<0.05).
In the present study, we describe a breeding strategy that generated a homozygous S. cerevisiae flor strain with desirable winemaking and genetic properties useful for further breeding. We used the flor strain M25, which is currently used in Sardinian wineries for fermentation and aging of Sherry-like wines and has been characterized genetically and physiologically (Budroni et al., 2000; Giordano, 2000; Pirino et al., 2004; Budroni et al., 2005). M25 has a semi-homothallic life cycle producing two MATa and two MATα spores. The MATa spores divide mitotically, generating a genetically uniform cell line selectable for winemaking characteristics and amenable for cell-to-cell crosses. MATα spores display homothallic behavior and any breeding program based on these progeny is complicated by the need to utilize a micromanipulator just before diploidization (Budroni et al., 2000, 2005).
In a previous study, we obtained strain Δ3B (MATα, HO::KanMX4), an M25 derivative unable to diploidize, and therefore useful for sexual crosses with cell lines derived from the same strain (Pirino et al., 2004). A cross between Δ3B and MI1C (MATa) generated a heterozygous diploid (MATa/MATα; HO/HO::KanMX4 HMRa/HMRa HMLa/HMLa). Contrary to what was reported by Walker et al. (2005), spore viability always cosegregated with G418 sensitivity. Cosegregation of spo+ and G418 sensitivity allowed for selection of MATα spores harboring the wild-type HO allele, and therefore the ability to diplodize and to generate homozygous diploids (except at the MAT locus). Sequencing of the HO allele showed that the breeding strategy used for 2D did not leave any trace of the KanMX4 cassette, yielding results comparable with that of the ‘delitto perfetto’ (Storici et al., 2001).
An important feature of wine strains is homozygosity, which enhances the stability of winemaking traits (Ramirez et al., 2004). Accordingly, analysis of AGA1, DAN4, FLO11, HSP150, and SED1, through 30 consecutive generations, confirmed the hypothesized mitotic stability of the 2D genome (data not shown). The two strains, 2D and M25, harbor a comparable DNA content. Presuming that low spore viability may result from unbalanced chromosomal arrangements (Adams et al., 1992; Bidenne et al., 1992; Codon et al., 1998; Rachidi et al., 1999; Delneri et al., 2003), it is possible that the increased spore viability observed in 2D relative to the parental strain M25 is due to the elimination of a deleterious chromosomal constitution.
A limitation of the use of sexual crosses as a method for genetic improvement of wine strains of S. cerevisiae is that, rather often, the homozygous spore progeny of excellent winemaking strains performs more poorly than the parental strains (Gimeno-Alcañiz & Matallana, 2001). In contrast, the 2D strain, selected for valuable winemaking performance among diploidized derivatives of Δ3B × MI1C, displayed winemaking characteristics comparable to that of the parental strain. Because of its semi-homothallic life cycle, increased spore viability, and winemaking ability, the 2D strain can be used for further breeding. The use of such strains offers a number of advantages. First, mating derivatives can be evaluated in winemaking trials and can be characterized genetically before crossing. Second, performing crosses with heterothallic or semi-homothallic strains does not require the use of a micromanipulator. Third, the hybrids generate nonmating, diploidizing, homozygous derivatives. Fourth, this strain produces a biofilm thicker than that of M25 and therefore has potential application in the protection of wines from oxidative damage during postfermentation biological aging.
The authors thank the Center for Biotechnology Development and Biodiversity Research, University of Sassari, for financial support, and Alan T. Bakalinsky for useful discussions and for critically reviewing the manuscript.