Enhancement of the Initial Rate of Ethanol Fermentation Due to Dysfunction of Yeast Stress Response Components Msn2p and/or Msn4p

Abstract

Sake yeasts (strains of Saccharomyces cerevisiae) produce high concentrations of ethanol in sake fermentation. To investigate the molecular mechanisms underlying this brewing property, we compared gene expression of sake and laboratory yeasts in sake mash. DNA microarray and reporter gene analyses revealed defects of sake yeasts in environmental stress responses mediated by transcription factors Msn2p and/or Msn4p (Msn2/4p) and stress response elements (STRE). Furthermore, we found that dysfunction of MSN2 and/or MSN4 contributes to the higher initial rate of ethanol fermentation in both sake and laboratory yeasts. These results provide novel insights into yeast stress responses as major impediments of effective ethanol fermentation.

During sake brewing, rice starch is saccharified by enzymes produced by koji mold (Aspergillus oryzae), and the resultant glucose is fermented by sake yeast, which produces ethanol to concentrations reaching approximately 20% (vol/vol), the highest level among nondistilled alcoholic beverages. Sake yeasts are strains of Saccharomyces cerevisiae and produce much more ethanol than laboratory strains of S. cerevisiae in sake mash. One of the reasons for the high ethanol production of sake yeast is its high fermentation rate (28). This property is a critical prerequisite for sake yeast strains because rapid and high-level ethanol accumulation leads to shortening fermentation periods, as well as preventing growth of unwanted microorganisms during sake brewing, in which open fermentation tanks are usually used. Therefore, yeast strains with higher fermentation rates have historically been selected as sake yeasts. In particular, modern sake yeast strains isolated within the last 80 years (also referred to as the “K7 group” [3]), represented by the most popular sake yeast, Kyokai no. 7 (K7), show high fermentation rates and are superior in flavor and aroma production. To explore the molecular mechanisms responsible for the high fermentation rate of modern sake yeast, we previously performed DNA microarray analyses of sake yeast strain Kyokai no. 701 (K701) (genetically almost identical to K7, but with a nonfoaming phenotype [23]) in fermenting sake mash (29). In this study, we compared these microarray data with data obtained from an experiment involving the laboratory yeast strain X2180 under identical fermentation conditions in an attempt to identify the underlying factors affecting ethanol production by sake yeast.

Yeast stress responses have been considered to be keys to the improvement of ethanol fermentation. A number of recent studies on industrial bioethanol production focus on modification of ethanol productivity through intensification of stress response machineries (20, 30). Msn2p and Msn4p (Msn2/4p) are functionally redundant transcription factors that have been the best characterized among regulators of environmental stress responses (11, 15, 21, 25). In response to various stresses, these factors migrate into the nucleus and bind to stress response elements (STRE) (CCCCT or AGGGG) within the promoters of stress-induced genes. The zinc finger DNA binding domains of Msn2/4p located at their C termini are essential for recognition of STRE sequences (21, 25). Interaction of Msn2/4p and STRE leads to global transcriptional activation of a large set of stress-responsive genes (8, 13), including those related to oxidative stress defense (e.g., CTT1 and SOD2) (5, 15, 21, 25), carbohydrate metabolism (e.g., TPS1, TPS2, GSY2, ALD2, ALD3, and TKL2) (5, 10, 17, 22, 25), and protein folding chaperones (e.g., HSP12, HSP104, SSA3, and SSA4) (1, 5, 15, 17, 21, 25). Consistently, a Δmsn2 Δmsn4 double mutant displays severe defects in stress-protective gene expression and thus exhibits pleiotropic stress sensitivity (11, 21). Here, we report an unexpected finding that Msn2/4p-mediated stress responses decrease the initial rate of ethanol fermentation during sake brewing.

MATERIALS AND METHODS

Strains and plasmids.

Sake yeast strains Kyokai no. 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, and 15 (K1, K2, K3, K4, K5, K6, K7, K8, K9, K10, K11, K12, K13, K14, and K15, respectively) and K701 (a nonfoaming variant of K7) were provided by the Brewing Society of Japan. A sake yeast strain (Yabe Kozai), shochu strains (S-2 and SH-4), wine strains (EC-1118, Montrachet, and OC-2), and other alcoholic strains (Bu9-7, S-3, RIB1023, RIB1024, and 389) were provided by the National Research Institute of Brewing (Japan). Beer strains (NCYC240, NCYC1026, NCYC242, and NCYC984) were provided by the National Collection of Yeast Cultures (United Kingdom). Laboratory strains X2180, X2180-1A, Σ1278b, A364A, and W303-1A were provided by the American Type Culture Collection. The BY4743, BY4743 Δmsn2, and BY4743 Δmsn4 strains were provided by EUROSCARF (Germany).

Disruption of the MSN2 or MSN4 gene in X2180-1A was performed by a PCR-based method (14) with primers MSN2-D1 (5′-CTTTTTTTCTTTGGTTTTATTTGCTTTATTTTTTCTTTCTTTTTTCAACTTTTATTGCTCATAGAAGAACTAGATCTAAACGTACGCTGCAGGTCGAC-3′) and MSN2-D2 (5′-ATAAGCCGTAAGCTTCATAAGTCATTGAACAGAATTATCTTATGAAGAAAGATCTATCGAATTAAAAAAAATGGGGTTTAATCGATGAATTCGAGCTCG-3′) (26) or MSN4-DF (5′-ATCAGTTCGGCTTTTTTTTCTTTTCTTCTTATTAAAAACAATATACGTACGCTGCAGGTCGAC-3′) and MSN4-DR (5′-TACCGTAGCTTGTCTTGCTTTTATTTGCTTTTGACCTTATTTTTTATCGATGAATTCGAGCTCG-3′) and plasmids pFA6-kanMX4 and pAG25 (14) as templates to generate X2180-1A Δmsn2, X2180-1A Δmsn4, and X2180-1A Δmsn2 Δmsn4. The mutations were confirmed by PCR with primers MSN2-3F (5′-CCAAGAGGCTACCTTTTTTTC-3′) and MSN2-4R (5′-AGCCGTAAGCTTCATAAGTCA-3′) or MSN4-3F (5′-CGCCTTTATCAGTTCGGCTTT-3′) and MSN4-4R (5′-ATCCGAATGAAATGACCAACC-3′). The STRE-pCYC1-lacZ reporter was introduced into K701, X2180, X2180-1A, and X2180-1A Δmsn4 as previously described (26).

For construction of the pYC140-ScMSN4 and pYC140-K7MSN4 plasmids, the MSN4 gene was amplified by high-fidelity PCR using KOD-Plus version 2 (Toyobo) from X2180 or K7 genomic DNA with primers MSN4-f-2 (5′-GCGTACGCGTCGACGGTACCTGGAAGACTGTCACTGAGAAATTCG-3′) and MSN4-r-2 (5′-TCCTTAAGCTTCCGGAGCTCCGTCGTACCAATCCTTGAATGC-3′) and cloned into the KpnI-SacI site of pYC140 (16) using an In-Fusion Advantage PCR Cloning Kit (Clontech).

DNA microarray analysis.

A DNA microarray experiment with X2180 sake mash was performed as previously described (29). A brief description of the details of the microarray experiments is shown in Fig. S1A and B in the supplemental material. Raw and normalized data used for the T-profiler analysis (4) are shown in Tables S1, S2, and S3 in the supplemental material.

qRT-PCR.

Sake mash using X2180-1A, X2180-1A Δmsn2 Δmsn4, or K701 was prepared as shown in Fig. S1A and B in the supplemental material. Total RNA was isolated from yeast cells in the sake mash on day 5 using an RNeasy Mini Kit (Qiagen) and quantified with a NanoDrop ND-1000 spectrophotometer. cDNA was synthesized from 300 ng of total RNA in a final volume of 20 μl using a PrimeScript RT reagent Kit (Perfect Real Time; TaKaRa). Gene-specific quantitative real-time PCR (qRT-PCR) primers were designed with Primer3 (24) as shown in Table S4 in the supplemental material. cDNA (0.2 μl each) was used in 20 μl qRT-PCR mixtures with 0.2 μM primers and SYBR Premix Ex Taq (Perfect Real Time; TaKaRa). qRT-PCR was performed with a LightCycler (Roche), and the expression data were processed by the second-derivative maximum method of LightCycler Software version 3.5. Delta cycle threshold (ΔC_T) values were calculated by subtracting the C_T of the ACT1 gene from the C_Ts of the genes of interest. ΔΔC_T values were calculated by subtracting the ΔC_T of the X2180-1A sample from the ΔC_T of the Δmsn2 Δmsn4 and K701 samples. Fold changes were calculated using the 2^−ΔΔCT method (19).

STRE-lacZ reporter gene assay.

For the reporter gene assay of yeast cells during sake fermentation, a sake mash was prepared by mixing 80 g pregelatinized rice, 20 g dried koji (rice with A. oryzae mold), 45 μl 90% lactic acid, and 160 ml of water containing yeast cells with a final optical density at 660 nm (OD₆₆₀) of 1 (precultured overnight in YPD medium [1% yeast extract, 2% peptone, 2% glucose] at 30°C). The mash was incubated at 15°C for 20 days without shaking. On days 2, 3, 5, 8, 11, and 15, yeast cells were collected from 15 g of the sake mash by serial centrifugation, washed twice with cold distilled water, disrupted in 250 μl breaking buffer (100 mM Tris-HCl [pH 8.0], 1 mM dithiothreitol, 20% glycerol, 1 mM phenylmethylsulfonyl fluoride [PMSF], and protease inhibitor cocktail [Roche]) with glass beads, and centrifuged at 21,900 × g for 10 min. One hundred microliters of the supernatant was mixed with 300 μl cold distilled water and 400 μl 2× Z buffer and then assayed for β-galactosidase activity, as previously described (2). The expression levels of the lacZ reporter gene were normalized by the protein levels quantified using a Coomassie (Bradford) Protein Assay Kit (Thermo Scientific). For the reporter gene assay of yeast cells under ethanol stress, yeast cells were precultured overnight in YPD, further cultured in YPD at 25°C until log phase, and transferred into fresh YPD or YPD containing 8% ethanol (EtOH) for 2 h. β-Galactosidase assays were performed as previously reported (2). For the complementation test of X2180-1A Δmsn4 by introducing the pYC140-ScMSN4 or pYC140-K7MSN4 plasmid, yeast cells were precultured overnight in YPD containing 300 μg/ml hygromycin to avoid curing of the pYC140-based plasmid and further cultured in YPD at 25°C until log phase with and without a temperature upshift to 39°C for 1 h. β-Galactosidase assays were performed as previously reported (2).

Measurement of the fermentation rate.

For small-scale sake brewing tests, sake mash was prepared as described above, and the fermentation was monitored by measuring the amount of evolved carbon dioxide, which was analyzed by measuring the weight loss of the sake mash at the same time each day. For the fermentation tests in YPD medium, yeast cells were precultured in YPD medium [or YPD containing 600 μg/ml hygromycin for the K7(pYC140) and K7(pYC140-ScMSN4) strains] overnight at 30°C, inoculated into 20% glucose-containing YPD medium at a final OD₆₆₀ of 0.1, and then further incubated at 30°C for 5 days without shaking. Fermentation was monitored by measuring the volume of evolved carbon dioxide using Fermograph II (Atto).

RESULTS

Sake yeast is defective in Msn2/4p-mediated gene expression in sake fermentation.

To identify differences between the sake and laboratory yeast strains that contribute to the superior brewing properties of sake yeast, we compared the gene expression profiles of both types of yeast in sake mash (the fermentation profiles are shown in Fig. S1C in the supplemental material). Analysis of differential DNA microarray data (see Table S3 in the supplemental material) using T-profiler, which is used for scoring changes in the average expression levels of predefined groups of genes (4), revealed that the expression of genes under the control of several transcription factors that are responsible for stress responses and nutrient signaling, including Msn2/4p, Skn7p, Gcr1p, Put3p, Gcn4p, and Adr1p, was significantly lower in K701 (Table 1). This result suggests that a relationship exists between their dysfunction and the superior brewing properties of sake yeast. Comparing microarray data for K701 and X2180 at similar ethanol levels also gave the same result: K701 exhibited impaired functions of transcription factors involved in stress responses and nutrient signaling (see Table S5 in the supplemental material). Among the identified transcription factors, we focused on Msn2/4p, as they are known to be crucial for yeast stress responses and their targets were the most remarkably downregulated in the K701 sake mash.

TABLE 1.

Differential gene expression analysis of K701 and X2180 yeast cells during sake brewing using T-profiler (comparison at specific time points)^a

Status	Day 2		Day 3		Day 5		Day 8		Day 11
	Motif	t value	Motif	t value	Motif	t value	Motif	t value	Motif	t value
Upregulated in K701	PAC	4.60	PAC	3.59	MBP1	4.23	MBP1	3.90	Meiosis	4.71
					PAC	4.02	REB1	3.69	RLM1	3.79
Downregulated in K701	MSN2/4b	−5.05	MSN2/4b	−6.41	MSN2/4c	−7.02	RAP1	−4.13	NDe
	Unknown	−4.15	MSN2/4c	−4.74	MSN2/4b	−6.72
	SKN7	−4.10	GCN4	−3.76	MSN2/4d	−3.92
	MCM1	−3.95	TBP	−3.71	ADR1	−3.87
	GCR1	−3.94
	TBP	−3.86
	PUT3	−3.84
	ACE2	−3.78
	PDR3	−3.73
	MSN2/4c	−3.68

^aThe names of the groups based on cis-acting motifs that were responsible for differentially expressed genes between yeast cells isolated from sake mash made with K701 and X2180 are shown.

^bCCCCT.

^cAGGGG.

^dHRCCCYTWDT.

^eND, not detected.

Our DNA microarray results indicate that the induction of known target genes of Msn2/4p, which peaked on day 3 or 5 in X2180, was significantly decreased or nearly completely abolished in the sake yeast (Fig. 1 A), consistent with the T-profiler analyses. To confirm these results, we performed qRT-PCR experiments using the samples from the sake mash on day 5 (Fig. 1B). All the target genes examined showed significantly lower mRNA expression levels in K701 than in X2180-1A (ranging from 4.1% ± 0.5% [TKL2] to 77.1% ± 2.6% [SOD2]), in agreement with our DNA microarray analyses. Comparison between the expression levels of the X2180-1A wild type and the Δmsn2 Δmsn4 mutant (ranging from 0.8% ± 0.1% [HSP12] to 51.3% ± 4.9% [ALD2]) verified that the upregulation of the target genes in the sake mash was at least partly dependent on Msn2/4p. These results revealed that Msn2/4p-mediated induction in the sake mash was significantly impaired in K701.

FIG. 1.

Sake yeast is defective for Msn2/4p-dependent gene expression. (A) Expression levels of the known Msn2/4p target genes during sake brewing revealed by DNA microarray analysis. Expression profiles of the sake yeast K701 (closed circles) and the laboratory yeast X2180 (open squares) are shown. (B) qRT-PCR experiments with X2180-1A, X2180-1A Δmsn2 Δmsn4, and K701 cells isolated from sake mash on day 5. The relative expression levels are given as fold differences compared to the induction levels obtained for X2180-1A, using ACT1 as a reference gene. The error bars indicate standard deviations.

Moreover, we examined the expression of the STRE-pCYC1-lacZ fusion gene (26) in both K701 and X2180 during sake brewing (Fig. 2 A). Whereas X2180 exhibited upregulation of the fusion gene from day 3, which indicates that the increasing ethanol concentration was sufficiently stressful to induce activation of Msn2/4p, no significant upregulation was observed in K701. In contrast, K701 exhibited an obvious induction of STRE-driven genes, the same as X2180, under the acute 8% ethanol stress in YPD medium (Fig. 2B). Altogether, these data demonstrate that sake yeast has severe defects in stress-inducible gene expression mediated by Msn2/4p and STRE specifically during sake brewing.

FIG. 2.

The sake yeast is defective for STRE-dependent gene expression, specifically in sake mash. (A) Expression levels of STRE-pCYC1-lacZ fusion during sake brewing. The fusion protein was expressed in K701 (closed circles) or X2180 (open squares) in sake mash. (B) Acute ethanol stress response of the STRE-pCYC1-lacZ reporter fusion in X2180 and K701. Expression levels of the fusion without stress (black) or under acute ethanol stress (8% ethanol, 2 h) (hatched) are shown. The error bars indicate standard deviations.

Loss-of-function mutations in the MSN4 gene are specifically distributed in modern sake yeast strains.

To investigate the cause of the defective stress responses in sake yeast, we focused on nucleotide polymorphisms in the MSN2 and MSN4 genes (Fig. 3 A). MSN2 nucleotide sequences are well conserved between K7 and X2180, except for 3 nonsynonymous polymorphisms. MSN4 has more nonsynonymous polymorphisms. We previously identified two point mutations (T2C and C1540T) in sake yeast K7 that result in the deletion of the N and C termini, respectively, of Msn4p (26). Of these mutations, C1540T is considered a loss-of-function mutation, as it deletes the C-terminal zinc finger DNA binding motifs of Msn4p. We therefore examined the functionality of Msn4p using the X2180-1A Δmsn4 strain complemented with either functional S. cerevisiae MSN4 (ScMSN4) or double-truncated MSN4 from K7 (K7MSN4). Upon heat shock, Δmsn4 showed a significant decrease of STRE-pCYC1-lacZ fusion expression compared to wild-type cells. Although the introduction of ScMSN4 rescued this impaired induction in the Δmsn4 strain, K7MSN4 did not contribute to the STRE-dependent stress response (Fig. 3B). We can therefore conclude that K7Msn4p no longer functions as a stress-responsive transcription factor.

FIG. 3.

Modern sake yeast-specific loss-of-function mutations in the MSN4 gene. (A) Point mutation sites in the MSN2 and MSN4 genes of K7. The hatched boxes show the sites of zinc finger motifs. The white circles indicate nonsynonymous polymorphisms in the MSN2 and MSN4 genes of K7. The arrows indicate two point mutation sites, T2C and C1540T, which are responsible for truncation of K7Msn4p at the N and C termini. aa, amino acids. (B) K7Msn4p loses its molecular functions as a stress-responsive transcription factor. Expression levels of STRE-pCYC1-lacZ fusion without stress (black) or under heat shock (39°C; 1 h) (hatched) are shown. WT, wild type. The error bars indicate standard deviations. (C) Distribution of the loss-of-function mutations in MSN4 among S. cerevisiae and Saccharomyces sensu stricto strains. The black and white boxes indicate X2180- and K7-type nucleotide polymorphisms, respectively.

Next, we examined the distribution of the T2C and C1540T polymorphisms among 17 sake, 3 wine, 4 beer, 2 shochu, 5 other alcohol, and 5 laboratory yeast strains (Fig. 3C). This comparison revealed that genetically closely related modern sake yeast strains, including Kyokai no. 6, 9, 10, 11, 12, 13, 14, and 15 (3), had mutations identical to those of K7 and K701. In contrast, nearly all of the other yeast strains tested, including the classical sake yeast strains (Kyokai no. 1, 2, 3, 4, 5, and 8 and Yabe Kozai), as well as the wine, beer, and laboratory yeast strains, showed no double truncation of MSN4, as was observed for X2180. Furthermore, fungal sequence alignment analysis of sequences in the Saccharomyces Genome Database (SGD) (http://www.yeastgenome.org/) revealed that both the T2 and C1540 nucleotides are universally conserved among orthologous genes in Saccharomyces sensu stricto. These data suggest that an ancestral strain of the modern sake yeasts may have acquired loss-of-function mutations in the MSN4 gene during the selection of sake yeast with desirable brewing properties.

Expression of the functional MSN4 gene delays ethanol fermentation in sake yeast.

To analyze the relationships between defective Msn4p and the brewing properties of the sake yeast, we performed small-scale sake-brewing tests (Fig. 4 A and B). Expression of the functional MSN4 gene from a low-copy-number plasmid in K7 (K7 + ScMSN4) led to delayed fermentation in the early stages of sake brewing (Fig. 4A). From days 2 to 5, carbon dioxide generation by the K7 + ScMSN4 strain was significantly lower than that of the control strain (Fig. 4B). In addition, we examined fermentation ability in the presence of 20% glucose utilizing Fermograph II, an automated gas-monitoring instrument (D. Watanabe, T. Akao, and H. Shimoi, submitted for publication), which revealed that the K7 + ScMSN4 strain also displayed slower fermentation than the control under this condition (Fig. 4C and D). Taken together, these results demonstrate that expression of functional Msn4p negatively acts on the ethanol fermentation rate of K7 and that dysfunction of Msn4p at least partly contributes to the high ethanol fermentation rate of K7.

FIG. 4.

Expression of functional MSN4 delays ethanol fermentation in K7. (A and B) Generation of carbon dioxide gas during small-scale sake-brewing tests of K7 + vector and K7 + ScMSN4. The total amounts of carbon dioxide evolved (A) and the amounts of carbon dioxide each day (B) are shown. *, significantly lower than the control strain; **, significantly higher than the control strain (t test; P < 0.05). The error bars indicate standard deviations. (C and D) Generation of carbon dioxide gas during fermentation tests in 20% glucose-containing YPD medium using K7 + vector and K7 + ScMSN4. The total amounts of carbon dioxide evolved (C) and the amounts of carbon dioxide every 15 min (D) are shown. Representative data from three independent experiments are shown.

Loss of MSN2 and/or MSN4 enhances the initial rate of ethanol fermentation.

Based on the fermentation test results described above, we hypothesized that the loss of Msn2/4p functions might lead to improved ethanol fermentation in general yeast strains other than sake yeast. We thus performed fermentation tests using the Δmsn2, Δmsn4, and Δmsn2 Δmsn4 disruptants in a laboratory yeast background (X2180-1A; haploid) (Fig. 5). It was revealed that all three disruptant strains displayed significant increases of evolved carbon dioxide gas in the early stages of sake brewing compared to the wild-type strain, with the double mutant showing the largest increase (Fig. 5A to D). After 20 days of fermentation, the ethanol concentrations were higher and the specific gravities were lower in sake produced using the disruptants (Table 2), also indicating that the abrogation of Msn2/4p leads to improved fermentation. Since the other examined brewing characteristics were not remarkably altered in the disruptants (Table 2), the major effects caused by loss of Msn2/4p functions are considered to be improvement of the initial fermentation rate and gross ethanol production. Furthermore, the disruptants also exhibited more rapid fermentation in 20% glucose-containing YPD medium at 30°C than the wild-type strain (Fig. 5E and G). In addition, we confirmed that the deletion of MSN2 or MSN4 in another genetic background (BY4743; diploid) gave results similar to those in the X2180-1A background (Fig. 5F and H). From these results, we demonstrated that impairment of Msn2/4p functions improves fermentation efficiency in various yeast strains regardless of the genetic background or the ploidy.

FIG. 5.

Depletion of MSN2 and/or MSN4 increases the fermentation rate in laboratory strains. (A to D) Generation of carbon dioxide gas during small-scale sake-brewing tests of the wild-type (BY4743) (A), Δmsn2 (B), Δmsn4 (C), and Δmsn2 Δmsn4 (D) strains. The amounts of carbon dioxide generated each day are shown. *, significantly higher than the wild type; **, significantly lower than the wild type (t test; P < 0.05). The peak level in the wild type is indicated by the red lines. The error bars indicate standard deviations. (E to H) Generation of carbon dioxide gas during fermentation tests in 20% glucose-containing YPD medium using the wild-type (green), Δmsn2 (orange), Δmsn4 (yellow), and Δmsn2 Δmsn4 (red) strains in the X2180-1A background (E and G) or in the BY4743 background (F and H). The total amounts of carbon dioxide evolved (E and F) and the amounts of carbon dioxide evolved every 15 min (G and H) are shown. Representative data from three independent experiments are shown.

TABLE 2.

Effects of deletion of MSN2 and/or MSN4 on selected sake-brewing characteristics of laboratory yeast strain X2180-1A^a

Parameter	Valued
	Wild type	Δmsn2	Δmsn4	Δmsn2 Δmsn4
Ethanol (vol%)	15.4 ± 0.2	16.0 ± 0.1b	15.7 ± 0.1b	16.0 ± 0.2b
Specific gravity (15°C/4°C)	1.0277 ± 0.0007	1.0235 ± 0.0007c	1.0256 ± 0.0006c	1.0257 ± 0.0011c
Acidity	4.7 ± 0.1	4.9 ± 0.1b	4.8 ± 0.1	4.2 ± 0.1c
Amino acidity	2.9 ± 0.1	2.9 ± 0.1	3.0 ± 0.2	3.6 ± 0.1b
Citric acid (mg/liter)	77.0 ± 0.8	74.7 ± 1.2	75.6 ± 0.3	74.8 ± 0.8
Malic acid (mg/liter)	148.7 ± 2.4	144.8 ± 2.5	151.8 ± 5.0	140.9 ± 8.2
Isoamyl acetate (ppm)	0.34 ± 0.02	0.37 ± 0.02	0.34 ± 0.04	0.29 ± 0.04
Ethyl caproate (ppm)	0.51 ± 0.01	0.59 ± 0.13	0.73 ± 0.11b	0.81 ± 0.00b

^aAnalyzed as previously reported (18).

^bSignificantly higher than the wild type (t test; P < 0.05).

^cSignificantly lower than the wild type (t test; P < 0.05).

^dAll data are presented as the mean ± standard deviation.

DISCUSSION

Yeast stress responses and the resultant stress tolerance are generally considered to be important characteristics for effective ethanol fermentation; therefore, numerous recent studies have focused on the enhancement of ethanol tolerance to achieve high ethanol productivity (20, 27, 30). Contrary to this approach, however, our present experimental data reveal that Msn2/4p-mediated environmental stress responses act as a physiological “brake” for ethanol production in both sake and laboratory yeast strains. The yeast strains defective in Msn2/4p exhibited significant increases of the peak fermentation rate in the early stages, leading to “powerful fermentation,” a novel concept aimed at advancing brewing technologies, such as optimization of continuous-culture methods for industrial bioethanol production.

Concerning the underlying molecular mechanisms that drive powerful fermentation, we hypothesize that minimal stress responses might contribute to facilitating the initial fermentation rate. In the early stages of fermentation, yeast cells might not need to rapidly acquire stress tolerance because the ethanol concentration is not increasing acutely, nor has it reached toxic levels. Therefore, the integrity of stress-responsive machineries might be less significant in this period. In fact, several stress response transcription factors, including Msn2/4p, of sake yeast were actually inactivated in the initial stage of fermentation, as shown in Table 1. Thus, yeast stress responses might be dispensable (or rather, inhibitory) in the early stage of ethanol fermentation. This idea is also in agreement with the evidence for impaired expression of heat shock proteins in beer and wine yeast strains in practical use (6, 7) and for high ethanol production by disruption of stress-responsive ubiquitin-related genes (28).

Moreover, it is noteworthy that yeast stress response machineries negatively affect cell growth, although they are important for maintaining cellular viability under acute stresses. For instance, activation of Msn2p or its overexpression under the control of a constitutive promoter has detrimental effects on cell growth and proliferation (9, 21). Considering the costs of stress responses, yeast strains with reduced stress-responsive gene expression might have advantages in increasing the population size under early fermentation conditions with moderate stress. In addition, yeast stress responses often result in drastic changes to cellular metabolic states, which might also affect powerful fermentation. The environmental stress responses mediated by Msn2/4p activate the expression of several genes involved in carbohydrate assimilation, such as trehalose and glycogen metabolism, pentose phosphate shuttling, fatty acid metabolism, and respiration (12), which also supports the idea that excess stress responses might have inhibitory effects on the glycolysis and ethanol production efficiency of individual cells by diversifying carbon flux. Thus, the stress response components of industrial yeast strains likely have been fine tuned by artificial selection.

How are the Msn2/4p-mediated stress responses of sake yeast inactivated during sake fermentation? We found that the modern sake yeasts, including K7, had nonfuctional MSN4 alleles (Fig. 3), and expression of functional MSN4 in K7 led to decrease of the initial fermentation rate (Fig. 4), suggesting that dysfunction of Msn4p might partly contribute to the powerful fermentation of K7. We speculate that an ancestral strain of the modern sake yeasts might have lost the functions of Msn4p during the artificial selection of sake yeast with a higher fermentation rate. However, this alteration in MSN4 is not sufficient to entirely explain the powerful fermentation of K7. Msn2p is redundant with Msn4p and is largely responsible for STRE-mediated transcription (25). In our data, STRE-driven gene expression was not significantly induced during sake fermentation by K701 (Fig. 2A) but was strongly induced in K701 under acute ethanol stress, almost the same as in X2180 (Fig. 2B). These results suggest that Msn2p of K701 is functional enough to induce STRE-pCYC1-lacZ under acute ethanol stress but is not activated by chronic stresses existing in sake mash (e.g., a gradual increase of the ethanol concentration). To understand the molecular basis for this phenomenon, we assume that an upstream activator(s) of Msn2/4p that specifically responds to chronic stresses under sake-brewing conditions is completely defective in K701 (and also in other sake yeast strains). More detailed analysis of the loss-of-function mutations in the sake yeast genome would reveal the central molecular mechanism responsible for both inactivation of Msn2/4p under chronic stresses and the powerful fermentation characteristic of sake yeast.

We also note that the defects in stress response components might cause some negative consequences, for example, difficulty in long-term survival in the stressful environment due to defective stress tolerance. In fact, most K701 cells lost their viability after sake fermentation was finished (see Fig. S1 in the supplemental material). This phenotype has not been a serious problem in the selection of sake yeast strains, however, since sake mash is filtered soon after the end of fermentation before yeast cell death, and recovered yeast cells are not reused for the next round of sake brewing.

We previously reported the increase of the final ethanol concentration in the sake mash fermented by a sake yeast strain overexpressing MSN2 (27). Reduced stress response and yeast cell death of sake yeast at the end of sake brewing may explain the apparent discrepancy that both inactivation and activation of Msn2/4p result in greater fermentation ability. While the excess level of MSN2 does not have an obvious effect on the initial rate of sake fermentation, it significantly improves the fermentation rate in the latter half of sake brewing; this is probably because stress tolerance caused by Msn2p might act positively on maintaining cell viability under severe ethanol stress. In contrast, the powerful fermentation phenotype attributed to loss of MSN2 and/or MSN4 is defined by enhancement of the initial fermentation rate. As shown in Fig. 5B and C, only the peak fermentation rates are specifically improved in the Δmsn2 and/or Δmsn4 disruptants. In addition, the fermentation rate in the latter stage is significantly reduced in the Δmsn2 Δmsn4 double mutant (Fig. 5D), suggesting that this strain might be too sensitive to ethanol stress to sustain its fermentation rate under the elevated ethanol concentration. We thus speculate that Msn2/4p-mediated stress responses contribute negatively to the fermentation rate in the initial stage and positively in the latter stage of the brewing process.

In conclusion, our results suggest that the inactivation of Msn2/4p may play a role in the high initial fermentation rate of modern sake yeast strains compared to other natural and industrial yeast strains. We suggest that stress responses have not evolved to promote rapid metabolism but to protect the organism from adverse conditions and to limit its tendency to make those conditions even worse. Elucidating all regulatory mechanisms affecting ethanol productivity will aid in future development of yeast strains with desirable fermentation properties.