Notice: Wiley Online Library will be unavailable on Saturday 30th July 2016 from 08:00-11:00 BST / 03:00-06:00 EST / 15:00-18:00 SGT for essential maintenance. Apologies for the inconvenience.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Sake is a traditional alcoholic drink in Japan, and is made from steamed rice by fermentation with Aspergillus oryzae, which saccharifies rice starch, and sake yeast, which converts the resultant glucose to ethanol. Although sake yeast strains are classified as Saccharomyces cerevisiae along with other yeast strains used in ethanol fermentation (Vaughan-Martini et al., 1998; Azumi et al., 2001), they differ from these other yeast strains in that they have many characteristics which make them suitable for sake brewing. These characteristics include a good aroma and the production of a high concentration of ethanol at low temperature (Hosaka et al., 1998). Several properties specific to sake yeast have been analysed at the gene level. For example, a mutation of ECM31 is involved in pantothenic acid auxotrophy of sake yeast Kyokai No. 7 (K7) at high temperature (Shimoi et al., 2000). AWA1 was identified as a gene involved in foam formation of mash during sake brewing (Shimoi et al., 2002; Miyashita et al., 2004; Shimizu et al., 2005). BIO6 was identified as a gene involved in biotin biosynthesis in sake yeast (Wu et al., 2005).
Sake yeast strains are usually heterothallic diploid (Akada, 2002), in contrast to wine yeast strains, which are usually homothallic. Therefore, it is likely that sake yeast strains have various heterozygous alleles on homologous chromosomes (heterozygosity) due to the accumulation of spontaneous mutations in the course of asexual vegetative growth. If this is the case, we can expect segregation of phenotypic traits in haploid strains after sporulation and concomitant meiosis of sake yeast strains. Haploids have been isolated from sake yeast and used as parental strains for intercross breeding of sake yeast strains (Kurose et al., 2000). However, a systematic analysis of many different haploid strains has not been reported to date, probably because of the very low sporulation efficiency of sake yeast (Nakazawa et al., 1992; Suizu et al., 1996).
In this study, we isolated 100 haploid strains from K7 and analysed their brewing characteristics in small-scale sake-brewing tests. K7 was originally isolated from a sake brewery in Japan and has been widely used as a prototype of sake yeast (Kasahara, 1963). Because K7 also has very low sporulation ability, we employed three screening criteria, including analyses of mating type and DNA content, to eliminate false haploid strains. We demonstrated that brewing characteristics such as fermentation ability and the production of aroma and flavour show a smooth continuous distribution in the different haploid strains. Our results suggest that K7 has multiple heterozygosities in its chromosomes and that their segregation and combination affect the brewing characteristics of K7.
Materials and methods
Yeast strains and medium
Yeast strains used in this study are summarized in Table 1. Yeast strains were routinely cultured in YPD medium (1% yeast extract, 2% peptone, 2% glucose: DIFCO, Sparks, MD, USA).
Table 1. Yeast strains used in this study
Sake yeast Kyokai No. 7, MATa/α
Brewing Society of Japan
Sake yeast Kyokai No. 7, haploid strain MATa or MATα
K7 was cultured on a YPD plate (2% agar) at 30 °C overnight. Yeast cells were transferred to a sporulation plate (2% potassium acetate, 2% agar) and incubated at 30 °C for 5–7 days. Cells were scraped off by the addition of water and collected by centrifugation. The cell suspension was treated at 65 °C for 10 min to kill vegetative cells, spread on a YPD plate and incubated at 30 °C for 2 days to allow colony formation. Colonies were randomly chosen and their ploidy was confirmed by the following three methods, as shown in Figure 1.
PCR for MAT locus
The MAT locus was analysed by PCR with the following specific primers (Huxley et al., 1990): MAT-specific primer (5′-AGTCACATCAAGATC- GTTTATGG-3′); MATa-specific primer (5′-ACTC- CACTTCAAGTAAGAGTTTG-3′); and MATα- specific primer (5′-GCACGGAATATGGGAC- TACTTCG-3′). DNA polymerase was Ex Taq (Takara Bio, Kyoto, Japan). The PCR reaction was carried out at 94 °C for 2 min, followed by 40 cycles of 94 °C for 5 s, 50 °C for 15 s and 72 °C for 10 s, followed by 72 °C for 5 min.
Flow cytometric analysis
The ploidy of the yeast strains was analysed by flow cytometry. K7 haploid candidate strains were cultured in YPD at 30 °C overnight. Cells (4.0 × 106) were harvested and fixed in 70% ethanol at 4 °C overnight. Cells were washed with 0.5 ml 50 mM sodium citrate, pH 7.5, and suspended in 0.5 ml of the same buffer. To the cell suspension, 12.5 µl 10 mg/ml RNaseA (Sigma, St. Louis, MO, USA) was added and the cells were incubated at 50 °C for 60 min. Thereafter, 25 µl 20 mg/ml proteinase K (Sigma) was added and the cells were incubated at 50 °C for 60 min. After centrifugation of the cells, 0.5 ml 50 mM sodium citrate, pH 7.5, and 4 µl 1 mg/ml propidium iodide (Sigma) were added. The cells were incubated at 4 °C overnight in the darkand analysed using a flow cytometer (Beckman Coulter Epics Elite ESP, Tokyo, Japan).
The mating test was performed according to a standard protocol (George et al., 1991). K7 haploid candidate strains, MATa tester strain (YPH499) and MATα tester strain (YPH500) were precultured in 1 ml YPD at 30 °C overnight; 100 µl each of a K7 haploid strain and a tester strain pre-culture was added to 1 ml YPD. The mixture of cells was incubated at 30 °C for 4 h. Mating of yeast cells was analysed by microscopic observation.
Small-scale sake brewing
Yeast strains were pre-cultured in 3 ml YPD at 30 °C for 24 h; 2 ml pre-cultured yeast was added to 100 ml YPD and cultured at 30 °C for 24 h. The cells (OD660 = 200 units) were harvested by centrifugation and mixed with 60 g pre-gelatinized rice (Tokushima Seiko Co., Tokushima, Japan), 23 g dry koji (Tokushima Seiko), 200 ml water and 44.5 µl 90% lactic acid. Dry koji is a culture of A. oryzae that is grown on steamed rice and then dried, and is a source of diastatic enzymes such as amylase and glucoamylase. Lactic acid was added to prevent bacterial contamination, since the fermentation set-up for sake brewing is not sterile. Sake mash was incubated at 15 °C without shaking. The sake mash was weighed every day and the CO2 output was calculated. After 20 days of culture, the mash was centrifuged and the supernatant was analysed as sake.
Analysis of sake
Ethanol concentration was measured by a contact combustion system with an alcohol densitometer (Yazaki, Tokyo, Japan). Specific gravity was measured using a density hydrometer (KEM, Kyoto, Japan). Acidity and amino acidity were measured with an electric potential difference autotitration apparatus (KEM), using the National Tax Administration Agency method (Okazaki, 1993). Volatile aromatic compounds (ethyl acetate, n-propyl alcohol, isobutyl alcohol, isoamyl acetate, isoamyl alcohol and ethyl caproate) were measured by headspace gas chromatography, using an Agilent 7694 Headspace Sampler and Agilent Technologies GC 6890N. Aliquots of samples (0.9 ml) were placed in 10 ml vials and 0.1 ml solutions of n-amyl alcohol (200 p.p.m.) and methyl caprate (5 p.p.m.), used as internal standards, were added. The vials were sealed with a silicon rubber stopper, covered with an aluminium cap and then heated at 50 °C for 30 min. Esters and higher alcohols were separated using a DB-WAX capillary column (0.32 mm i.d. × 30 m, film thickness 0.25 µm; Agilent) after autoinjection of a headspace volume of 1 ml. The following conditions were applied: injection temperature, 200 °C; oven temperature, 85 °C; detector temperature, 250 °C; and carrier gas, He 2.2 ml/min.
Principal component analysis
Principal component analysis of brewing characteristics was performed using Statistica software (StatSoft Japan Inc., Tokyo, Japan).
Isolation of K7 haploid strains
Since the sporulation efficiency of K7 is very low (Nakazawa et al., 1992; Suizu et al., 1996), we selected cells that had survived heat treatment as haploid candidates. This selection procedure yielded 1145 haploid candidates. As the diploid K7 has the MATa/α genotype, we selected 406 strains that showed MATa or MATα by PCR analysis of the MAT locus of these strains. It has been reported that diploid cells with heterozygous MAT loci transform into diploid cells with homozygous MAT loci (MATa/a or MATα/α), due to loss of heterozygosity (Hashimoto et al., 2006). To eliminate these false-positive strains, the DNA content of the selected strains was analysed by flow cytometry. Based on this analysis, 127 candidates were assessed as haploid. Finally, the selected strains were subjected to mating tests with a strain of an opposite mating type, yielding 100 haploid strains (Figure 1).
Brewing characteristics of K7 haploid strains
To analyse the brewing characteristics of 100 K7 haploid strains, we performed three small-scale sake-brewing tests with each strain. Seven of the 100 haploid strains showed very low fermentation ability, and the sake mashes were contaminated with moulds. These seven strains were therefore eliminated from the study and the brewing characteristics of the remaining 93 haploid strains were analysed by performing at least two sake-brewing tests per strain. The number of sake brewing tests performed on each strain is shown in the Supporting information (see Table S1). Of these strains, 51 were MATa and 42 were MATα. Analysis of the ethanol concentration, the CO2 output, the concentrations of ethyl acetate, n-propyl alcohol, isobutyl alcohol, isoamyl acetate, isoamyl alcohol and ethyl caproate, the specific gravity, acidity and amino acidity of sake produced from these strains are shown in Table S1 (see Supporting information) and summarized in Figure 2. All analytical values exhibited a smooth continuous distribution, indicating that these traits were controlled by multiple genes. Although 44 of the haploid strains showed a fermentation ability that was similar to the parental diploid strain, as judged by ethanol concentration, CO2 output and specific gravity, no strain was better than the parental strain for any of these parameters (Figure 2a–c). In contrast, some of the K7 haploid strains produced more aromatic components than the parental strain. One such component was isoamyl acetate, which is an important aromatic component of ginjo sake (Figure 2h). Other haploid strains produced various levels of aromatic components of sake compared to the parental strain. However, there was no strain that produced a higher level of ethyl caproate than the parental strain. Ethyl caproate is another aromatic component of ginjo sake (Figure 2i). The K7 haploid strains also showed variation in the levels of acidity and amino acidity, compared to the parental strain (Figure 2j, k). For all of the parameters analysed, there were no significant differences between MATa and MATα strains (data not shown).
Correlation of brewing characteristics
To analyse the relationship between individual brewing characteristics, we calculated the correlation coefficient between the different characteristics (Table 2). A strong correlation (p < 0.001) was observed between the concentration of ethanol and the concentration of other major components of sake, such as aromatic components, with the exception of sake acidity. This correlation indicated that strains that produce a higher concentration of ethanol also produce a higher concentration of aromatic components. Strong correlations were also observed between the concentrations of aromatic components, such as ethyl acetate, n-propyl alcohol, isobutyl alcohol, isoamyl acetate and isoamyl alcohol, that were produced by the strains, with the exception of ethyl caproate (p < 0.001). There was an especially strong correlation (r = 0.8745) between the concentration of ethyl acetate produced and that of isoamyl acetate, which are both acetate esters. Calculation of the correlation of ethyl caproate with the major aromatic components of sake showed that ethyl caproate only correlated with ethyl acetate (p < 0.01). When the correlation of acidity with the other sake components was calculated, acidity correlated only with amino acidity (p < 0.001). Amino acidity inversely correlated with all of the major sake components, with the exception of isobutyl alcohol and ethyl caproate.
Table 2. Correlation coefficients of the brewing characteristics of K7 haploid strain
To better understand the multidimensional data that was generated in the statistical analysis, we reduced the dimensions by performing principal component analysis (PCA) of the brewing characteristics (Table 3). The first two principal components (PC) explained 60.4% and 12.1%, respectively, of the total variance. PC1 strongly correlated with parameters related to fermentation ability, such as ethanol concentration, CO2 output and specific gravity, suggesting that PC1 represents fermentation ability. However, PC2 correlated only with isobutyl alcohol, ethyl caproate and acidity, suggesting that PC2 represents factors other than fermentation ability, e.g. the flavour of sake.
Table 3. First principal components of the brewing characteristics of K7 haploid strain analysed by PCA
Explained variance (%)
Total variance (%)
In this study, we isolated 100 K7 haploid strains from 1145 haploid candidates, using a random spore method and three screening criteria (Figure 1). Since haploid candidates include many false-positive strains, strict selection criteria are necessary to isolate true sake yeast haploid strains. Flow cytometric analysis of ploidy is especially crucial for the selection of true haploids, because false-positive strains such as MATa/a and MATα/α cannot be eliminated by PCR analysis of the MAT locus or by mating tests. Although there are several reports on the isolation of sake yeast haploids (Ouchi et al., 1976; Kawamura et al., 1986; Kurose et al., 2000), the ploidy of these strains was not confirmed by flow cytometric analysis. Indeed, using our screening method we found that a strain (ATCC 44 882) deposited in the American Type Culture Collection as a haploid strain derived from K7 was not haploid but in fact a MATa/a diploid strain (data not shown). Although we employed strict selection criteria to screen for haploid strains, we cannot exclude the possibility that complete or partial duplication or deletion of an individual chromosome might occur during sporulation by abnormal chromosome segregation in our haploid strains. To confirm these phenomena, all chromosomes will be examined by chromosome electrophoresis or comparative genome hybridization using DNA microarrays.
The brewing characteristics of the K7 haploid strains showed a smooth, continuous distribution, indicating that these characteristics were controlled by multiple genes that have cumulative effects on each trait. The distribution of fermentation ability, judged by parameters such as ethanol concentration, CO2 output and specific gravity, showed a smooth but skewed distribution with a peak at the higher end, and no strains showed higher values than the parental strain (Figure 2a–c). These results suggest that the parental diploid K7 has many recessive mutations that exert negative effects on fermentation ability. Also, the haploid strains showed varying degrees of decreased fermentation ability that were dependent on the combination of recessive mutations exposed in the haploid strains. In addition, we cannot exclude the possibility that an overall decrease in gene dosage brought about by haploidization affects fermentation ability as has previously been reported for a laboratory yeast strain (Takagi et al., 1983). Future diploidization experiments of the haploid strains will be helpful to elucidate this point.
In contrast to the skewed distribution of fermentation ability, the distribution of aromatic and other sake components showed a bell-shaped distribution with a single peak. In these cases some of the K7 haploid strains produced higher levels of the sake components, with the exception of ethyl caproate, compared to the parental strain (Figure 2d–i, k). In summary, these results suggest that the parental diploid K7 strain has many recessive mutations that can have either a negative or positive effect on the concentration of sake components. The specific combination of these recessive mutations determines the concentration of aromatic components and other sake components, with the exception of ethyl caproate, in the haploid strains.
We found a strong correlation between the concentration of ethanol and that of the major aromatic components, with the exception of ethyl caproate (Table 2). Higher alcohols are synthesized by decarboxylation of keto acids, which are synthesized in the amino acid biosynthetic pathway (Yoshizawa, 1966), and acetate esters are synthesized by a condensation reaction between higher alcohols and acetyl-CoA (Fujii et al., 1994). It is likely that higher ethanol production requires higher amino acid synthesis, which is necessary for the increased production of higher alcohols and their esters. We also found a very strong correlation between the concentration of ethyl acetate and that of isoamyl acetate. This correlation is consistent with the fact that these two compounds are synthesized by a similar mechanism using the same precursor (acetyl-CoA) and/or the same enzyme. In contrast, ethyl caproate is synthesized by a condensation reaction between caproic acid and ethanol, or between caproyl-CoA and ethanol through the fatty acid biosynthetic pathway in sake yeast (Kuriyama et al., 1986). One of the reasons for the low correlation between ethyl caproate and other aromatic components of sake could be that the ethyl caproate biosynthetic pathway is different from that of the other aromatic components.
Recently, the genome of the diploid yeast Candida albicans, an opportunistic pathogen of human, was sequenced (Jones et al., 2004). The authors identified numerous heterozygosities in the genome (Jones et al., 2004; Forche et al., 2004). It was also reported that the heterozygosities found in C. albicans CDR2, which encodes one member of the PDR family of the ATP-binding cassette (ABC) class of transporters, were involved in azole sensitivity (Holmes et al., 2006). Genome sequencing of the diploid K7 is now being carried out by a research consortium that includes our institute, and preliminary experiments suggest that diploid K7 has > 1000 heterozygosities in its genome (our unpublished data). However, it has recently been reported that genome tiling arrays can be used to detect single nucleotide polymorphisms (SNPs) at a genome-wide level (Gresham et al., 2006). Detection of SNPs in all the offspring of K7 by such a method would enable us to perform linkage analysis leading to identification of quantitative traits loci (QTL) of brewing characteristics. The results presented in this paper suggest that the brewing characteristics of K7 haploids are under the control of QTL and that identification of the QTL that determine the brewing characteristics of sake yeast should be of great practical use in the future breeding of sake yeasts.
Supporting information may be found in the online version of this article.
This work was supported by the Programme for the Promotion of Basic Research Activities for Innovative Biosciences (PROBRAIN).