OUP user menu

Overexpression of PDE2 or SSD1-V in Saccharomyces cerevisiae W303-1A strain renders it ethanol-tolerant

Liat Avrahami-Moyal, Sergei Braun, David Engelberg
DOI: http://dx.doi.org/10.1111/j.1567-1364.2012.00795.x 447-455 First published online: 1 June 2012


Understanding the genetic basis of the yeast ability to proliferate and ferment in the presence of restrictive concentrations of ethanol is of importance to both science and technology. In this study, we searched for genes that improve ethanol tolerance in ethanol-sensitive strains. To screen for suppressors of ethanol sensitivity, we introduced a 2µ-based genomic library, prepared from the ethanol-tolerant yeast S288C, into the ethanol-sensitive strain W303-1A. Two genomic fragments from this library rescued the ethanol sensitivity of W303-1A. One contained the PDE2 gene, which when over-expressed, conferred ethanol tolerance. Surprisingly, the effect of PDE2 was not mediated via MSN2/MSN4 transcription factors, as it was able to improve ethanol tolerance in msn2Δmsn4Δ strain. In the second genomic fragment, it was the N-terminal region of the SSD1 gene that carried the ethanol-tolerant phenotype. The SSD1-V allele of the polymorphic SSD1 gene expressed from a low-copy number plasmid also resulted in the tolerant phenotype. Both SSD1 and PDE2 seemed to improve ethanol tolerance by maintaining robustness of the yeast cell wall.

  • ethanol tolerance
  • SSD1
  • PDE2
  • 2µ-based genomic library
  • cell wall


The ability of industrial yeast to proliferate and ferment in the presence of restrictive concentrations of ethanol is one of its most important features. The genetic basis of this ability is so far unclear. Previous strategies used to identify the genetic components of ‘ethanol tolerance’ involved screening of deletion libraries for genes rendering tolerant strains sensitive to ethanol (Takahashi et al., 2001; Kubota et al., 2004; Fujita et al., 2006; van Voorst et al., 2006; Teixeira et al., 2009). These global screens identified a diverse group of factors involved in cell cycle and DNA processing, protein fate, cellular transport mechanisms, transcription, biogenesis of cellular components, vacuolar function, metabolism, signal transduction, protein synthesis, defense, etc (Fujita et al., 2006). Notwithstanding the importance of gene disruption studies, one shortcoming of such an approach is that any systematic damage to the yeast would also be detrimental to its growth under stressful conditions. Although the deletion library screens led to the identification of many genes that are important to ethanol tolerance, they are not necessarily directly involved in acquiring it.

Moreover, recent studies identified genes whose overexpression led to an increase in ethanol tolerance [MSN2 (Watanabe et al., 2009; Lewis et al., 2010); TRP1-5 (Hirasawa et al., 2007); TPS1, EDE1, and ELO1 (Lewis et al., 2010); INO1, DOG1, and HAL1 (Hong et al., 2010)], although these genes were not found to confer ethanol sensitivity in the deletion library screens.

We therefore posited that there is an advantage to searching for genes that confer tolerance on ethanol-sensitive strains, reasoning that the gain of function would be gene-autonomous. Among few such screens that employed such a notion, Yazawa et al. (2007) have identified two null mutants (ura7Δ and gal6Δ) that grew faster than the wild-type strain in medium containing 8% v/v ethanol. The survival rate of the gal6Δ strain in 10% ethanol was also higher than that of the wild-type strain. Both deletion mutants were more tolerant to zymolyase, a cell wall-degrading enzyme, indicating that the integrity of the cell wall is a prerequisite of tolerance to ethanol stress.

Ethanol tolerance is an elusive comparative property. In this work, we define ethanol tolerance ad hoc as the ability of yeast to form colonies on a rich solid medium containing at least 7% ethanol in 6 days. Within this definition, yeast strains vary in their ethanol tolerance (Fig. 1). Thus, the Japanese rice liquor producer, Sake yeast, is the most tolerant; W303-1A and BJ2168 are sensitive, while BY4741 (derived from S288C), SPI and Σ5527LH have intermediate tolerance (Fig. 1). The genetic basis of these differences is not investigated. As a tool, one could, perhaps, use the opportunity to complement the ethanol-sensitive strains with genes from the relatively ethanol-tolerant strains in the search of genes responsible for the difference between them. Here we attempted to validate this notion by challenging the sensitive W303-1A strain with a genomic library prepared from the relatively ethanol-tolerant BY4741 (derived from S288C) and cloned in a 2µ-based plasmid. We report that two genes from this library, PDE2 or a truncated SSD1-V (F6), can rescue the ethanol sensitivity of W303-1A strain.


Ethanol sensitivity of laboratory yeast strains. Spot assay was carried out as described in . Aliquots (5 µL) of 10-fold serial dilutions of the indicated strains were spotted onto plates containing YPD + 7% ethanol and cultured at 30 °C for 3 days. No differences in the apparent growth rates were found when the same strains were grown on medium that lacks ethanol (not shown).

Materials and methods

Yeast strains, media, and general methods

The Saccharomyces cerevisiae strains used in this study are listed in Table 1. Strains were grown at 30 °C on YPD medium containing (g L−1) yeast extract 10, Bactopeptone 20, and glucose 20, or on the synthetic medium YNB-Ura containing (g L−1) yeast nitrogen base 1.7, ammonium sulfate 5, glucose 20, leucine, adenine, tryptophane, uracil, histidine, lysine and methionine 0.04. Solid media were made with 3% agar. Strains were grown in liquid or solid media supplemented with ethanol as indicated in each experiment. Petri dishes with ethanol-containing solid media were sealed with parafilm to avoid evaporation.

View this table:

Yeast strains used in this study

StrainGenotype: source or reference
W303-1AMATa can1-100 ade2-1 his3-11, 15 leu2-3 trp1-1 ura3-1 Yeast, Genetic Stock Center, Berkeley, CA
BY4741MATa his3∆1 leu2∆0 met15∆0 ura3∆0 (Brachmann et al., 1998)
Kyokai 9, Sake yeastA kind gift from Dr Haruyuki Iefuji (Shobayashi et al., 2007)
BJ2168MATa leu2 trp1 ura3-52 prb1-1122 pep4-3 prc1-407 gal2, ATCC
SPIMATα his3 leu2 ura3 trp1 ade8 canr, M. Wigler (Toda et al., 1985)
∑5527LH∑1278bMATα his3 leu2 ura3 trp1(Stanhill et al., 1999)
BY4741ssd1Isogenic to BY4741 but ssd1::KanMX4, Euroscarf
W303-1Assd1Isogenic to W303-1A but ssd1::KanMX4, this study
W303msn2Δ msn4ΔIsogenic to W303-1A but msn2::HIS3 msn4::URA3, this study

The W303ssd1Δ mutant was created by disrupting the open reading frame (ORF) with the kanR marker using a PCR-based knockout strategy (Brachmann et al., 1998). Briefly, SSD1:: kanMX4 was amplified from genomic DNA of deletion strain BY4741ssd1:: kanMX4 (obtained from Euroscarf), using PCR (the primers are shown in Table 3). The PCR product was used to transform the W303-1A strain. The W303msn2Δmsn4Δ double knockout was created by disrupting the ORF of MSN2 or MSN4 using plasmids pΔBX or pZfh45-1, respectively (Estruch & Carlson, 1993). All the knockout strains were verified using PCR.

Yeast was transformed using the lithium acetate method (Gietz & Schiestl, 2007).

Plasmids and primers

The genes used in this study were inserted into pRS416, pRS426, or pRS425 (Christianson et al., 1992). Plasmids and primers are listed in Tables 2 and 3. A detailed description of plasmids construction is given in the Table 2.

View this table:

Plasmids used in this study

pRS416URA3, CEN (Christianson et al., 1992)
pRS426URA3, 2μ (Christianson et al., 1992)
pRS425LEU2, 2μ (Christianson et al., 1992)
p2μ-F2pRS426 containing the F2 genomic fragment, this study
pRS416-F2p2μ-F2 was digested with XhoI and NotI. The product, F2 genomic fragment, was cloned into pRS416
p2μ-F6pRS426 containing the genomic fragment F6, this study
pRS416-F6p2μ-F6 was digested with XhoI and NotI. The product, F6 genomic fragment, was cloned into pRS416
pRS425-F6The F6 genomic fragment was excised from p2μ-F6 using XhoI and NotI and then cloned into pRS425
pRS426-PRT1ORF of PRT1 was amplified using PCR with p2μ-F2 as a template. The PRT1-PCR product digested with SacII and cloned into pRS426
pRS426-PRT1 + PRE10The plasmid p2μ-F2 was digested with NotI and BamH1, and the product, PRT1 + PRE10, was cloned into pRS426
Yep13-PDE2ORF of PDE2 and 469 bp upstream of it cloned into Yep13 (LEU2, 2μ, Sass et al., 1986)
pRS426-HDA2ORF of HDA2 was amplified using PCR with p2μ-F2 as a template. The HDA2-PCR product digested with BamH1 and cloned into pRS426
pRS426-SSD1 + DPL1The p2μ-F6 was digested with XmaI and XhoI. The product, SSD1 + DPL1, was cloned into pRS426
pRS426-DPL1 + HDA2The p2μ-F6 was digested with PvuII. The product, DPL1 + HDA2, was cloned into pRS426
pRS416-SSD1 + DPL1The p2μ-F6 was digested with XmaI and XhoI. The product, SSD1 + DPL1, was cloned into pRS416
pRS416-VTS1 + PDE2The p2μ-F2 was digested with KpnI. The product, VTS1 + PDE2, was cloned into pRS416
pRS426-SSD1-VORF of SSD1-V allele, 399 bp upstream and 404 bp downstream, was excised from the plasmid pPL094 [a kind gift of Ted Power, (Reinke et al., 2004)], with XhoI and SacI, and cloned into pRS426
pRS416- SSD1-VORF of SSD1-V allele, 399 bp upstream and 404 bp downstream, was excised from pRS426-SSD1-V with XhoI and SacI, and cloned into pRS416
pRS426-ssd1-dORF of ssd1-d allele, 399 bp upstream and 404 bp downstream, was excised from the plasmid pPL093 [a kind gift of Ted Power (Reinke et al., 2004)], with XhoI and SacI, and cloned into pRS426
pRS316- ssd1-dpPL093 containing ssd1-d allele, 399 bp downstream and 404 bp upstream of it. A kind gift of Ted Power (Reinke et al., 2004)
View this table:

Primers used in this study

Primer nameUsed for…Primer sequence (5′ to 3′)
HDA2_Fcloning HDA2 from 2μ-F6 plasmidCGCGGATCCCCGGTATATATGGTAACBamHI restriction site is underlined
HDA2_Rcloning HDA2 from 2μ-F6 plasmidCGCGGATCCCCCTATAGTGAGTCGTATTABamHI restriction site is underlined
PRT1_FClone PRT1 from 2μ-F2 plasmidTCCCCGCGGTTTGTACGAGGAGAGSacII restriction site is underlined
PRT1_RClone PRT1 from 2μ-F2 plasmidTCCCCGCGGGACAGTGATAACGTCSacII restriction site is underlined
SSD1_KO_FSSD1 knockout. The primer position is 690 bp upstream to ssd1:: KanMX4 in BY4741ssd1GGCTGCAGGAGTTTAGATCACGAGTAT
SSD1_KO_RSSD1 knockout. The primer position is 147 bp downstream to ssd1:: KanMX4 in BY4741ssd1TTTGGATGATGCCTTACCGGAACGTGG
  • Forward primers are indicated by ‘F’, reverse primers by ‘R’.

Sequence analysis

The sequence analysis was performed using the dye-terminator sequencing chemistry, 3730xl DNA Analyzer and sequence analysis software (Applied Biosystems, Foster City, CA)

Spot assay and growth condition

Yeast cells were cultured overnight in liquid medium (YNB-Ura or YPD). The cells were diluted in the appropriate medium to OD600 = 0.2, grown for 2–5 h and harvested in early logarithmic phase (OD600 ≈ 0.4–0.6). The cells were re-suspended in a fresh growth medium to a concentration of 107 cells mL−1. Serial dilutions of the cultures were made (107, 106, 105, 104, and 103 cells mL−1). An aliquot of each dilution (5 μL) was deposited on an indicated 3% agar media and cultured at 30 °C for several days until small colonies developed. Alternatively, the dilution was made by streaking cultures of equal cell concentration on 3% agar; cell proliferation rate of various yeast isolates was compared by the observation following several days in culture. All the assays were performed with several concentrations of ethanol at least three times for each concentration; all assays were consistently reproducible.

The 2µ-based genomic library screen

The genomic library containing the 2µ-element and the URA3 gene was obtained from the laboratory of G. R. Fink of the Whitehead Institute. This library was prepared by partial digestion of S288C genomic DNA with Sau3A. All DNA fragments were cloned into BamH1 site of pRS426. This library was introduced into W303–1A cells according to Gietz & Schiestl (2007).Transformants were plated on YNB-Ura. Colonies were allowed to develop for 2 days and were then replica-plated to YPD plates supplemented with 6, 7, 8, or 9% (v/v) ethanol.

Sensitivity of yeast cell wall to zymolyase

Cell wall lysis assay was based on the study by Ovalle et al., 1998. Briefly, cells were grown overnight on indicated media at 30 °C, harvested, washed three times with deionized water, and re-suspended at OD600 = 0.5 in TE buffer (Tris/HCl, 50 mM, EDTA, 5 mM, at pH 7.5). Zymolyase 20T, 0.5 U (20 U mg−1; Seikagaku 120491) was then added. Cell suspensions were incubated at 30 °C, and their optical densities were recorded at 3 min intervals.

Results and discussion

Genomic library screen for ethanol tolerance genes

A genomic library, prepared from the ethanol-tolerant laboratory yeast S288C and cloned into a 2µ-based plasmid, was introduced into the ethanol-sensitive strain W303-1A. From a series of six transformations, which provided, c. 31 000 colonies, 26 colonies grew well on 7% and even on 9% ethanol. Plasmid loss assays and re-transfection of the plasmids, isolated from these colonies, into W303-1A cells, confirmed that 22 of these colonies were ethanol-tolerant in the plasmid-dependent manner. Sequence analysis of these plasmids showed that 20 colonies carried a plasmid containing the same genomic fragment, marked F2 (Fig. 2), and two colonies contained a different plasmid with a fragment marked F6 (Fig. 2), both capable of rescuing the sensitivity to ethanol (Fig. 3).


Schematic presentation of the F2 and F6 genomic fragments, which confer ethanol tolerance to W303-1A cells. The genes within the fragments are designated by arrows oriented at the direction of translation. Sub-fragments derived from F2 and F6 for further cloning are marked by dash lines. These sub-fragments were selected according to the availability of restrictive enzyme site for cloning.


PDE2 (a) and SSD1 (b) confer ethanol tolerance to W303-1A. Aliquots (5 µL) of 10-fold serial dilutions of W303-1A and W303-1A harboring the indicated plasmids were spotted onto plates containing YPD + 7% ethanol and cultured at 30 °C for 7 days. No differences in the apparent growth rates were found when the strains were grown on medium that lacks ethanol (not shown).

Analysis of the fragments

The F2 fragment contained the genomic region between 1 019 439 and 1 011 628 bp on chromosome XV. This region includes the whole ORFs of PDE2 and PRT1 as well as 167 bp downstream from the stop codon of PRE10 and 14 bp upstream from the stop codon of VTS1 (Fig. 2).

The p2µ-F2 plasmid was digested with convenient restriction enzymes (Table 2) producing sub-fragments containing various combinations of the genes present in the original F2 fragment (Fig. 2). Overexpression of PRT1 + PRE10 and PRT1 alone did not rescue W303-1A (Fig. 3a), while overexpression of PDE2 conferred ethanol tolerance as did the whole fragment F2 (Fig. 3a).

The second fragment, F6, contained the genomic region between 1 046 874 and 1 053 622 bp of chromosome IV. It included the whole ORF of DPL1, as well as 1002 bp of the C-terminal region of HDA2 and 2515 bp encoding the N-terminal region of the SSD1 (Fig. 2). The p2µ-F6 plasmid was also digested with convenient restriction enzymes (Table 2) producing sub-fragments containing various combinations of the genes present in the original F6 fragment (Fig. 2). Overexpression of the sub-fragment containing 2515 bp from SSD1 and a part of DPL1 (pRS426-SSD1 + DPL1) rescued W303-1A (Fig. 3b), while overexpression of either HDA2 alone (Fig. 3a) or HDA2 + DPL1 did not support growth on ethanol-containing medium (Fig. 3b). We therefore concluded that the truncated SSD1 gene is responsible for the ethanol-tolerant phenotype.

SSD1 is a polymorphic gene, which appears in laboratory strains in two known genetic variants (Sutton et al., 1991): the SSD1-V allele that gives rise to a 1251 amino acid residues-long protein and the ssd1-d allele containing stop codon that terminates the protein at the position 698 and gives rise, therefore, to a 697 amino acid residues-long protein. The ethanol-sensitive strain W303-1A possesses the ssd1-d allele; while the relatively more tolerant strain BY4741 (Fig. 1) contains the SSD1-V. In BY4741, SSD1 deletion confers ethanol sensitivity (Kubota et al., 2004; Yoshikawa et al., 2009). The fragment F6 contained the third, truncated, version of SSD1-V that encodes for a putative SSD1 variant of 838 amino acids, which we termed SSD1-V-F6. F6 fragment required multi-copy expression to rescue W303-1A, while it failed to rescue it at a low-copy expression levels (Fig. 4a). The sub-fragment containing SSD1-V-F6 and a part of DPL1 also did not rescue W303-1A in a low-copy number plasmid (Fig. 3b).


Ethanol tolerance in W303-1A: expression of SSD1-V-F6 and PDE2 from either low-copy or multiple-copy number plasmids; co-overexpression of these genes. Aliquots (5 µL) of 10-fold serial dilutions of W303-1A and W303-1A harboring the indicated plasmids were spotted onto plates containing YPD + 7% ethanol and cultured at 30 °C for 7 days. Aliquots (5 µL) of 10-fold serial dilutions of W303-1A and W303msn2Δ msn4Δ harboring the indicated plasmids were spotted onto plates containing YPD + 7% ethanol and cultured at 30 °C for 7 days. No differences in the apparent growth rates were found when the same strains were grown on medium that lacks ethanol (not shown).

The same overexpression requirement for increased ethanol tolerance characterized PDE2 or fragment F2 (Fig. 4b). Coexpression of SSD1 and PDE2 on multicopy plasmids apparently did not result in a synergistic or additive ethanol tolerance phenotype (Fig. 4a).

We wondered whether the genetic elements that we have identified are related to some specific defect in the strain chosen for experimentation, or they are more general and could confer ethanol tolerance to other strains. We tested this query on the laboratory yeast strain BJ2168. BJ2168 is as sensitive to ethanol as W303-1A strain (Fig. 1) because it lacks PEP4, a vacuolar protease (Jones, 1990), a defect known to cause ethanol sensitivity in the YPS163 background (Lewis et al., 2010). We confirmed by PCR that it contains the functional SSD1-V allele. Notwithstanding the vacuolar protease deficiency, multi-copy expression of either F6 or F2 fragments (but not their expression from a centromeric plasmid) rescued the BJ2168 strain, whereas the F6 fragment had a more pronounced effect (Fig. 5).


Overexpression of fragments containing either SSD1-V-F6 or PDE2 confers ethanol tolerance to BJ2168. Aliquots (5 µL) of 10-fold serial dilutions of BJ2168 harboring the indicated plasmid (‘CEN’- for centromeric plasmid, pRS416) were spotted onto plates containing YPD + 7% ethanol and cultured at 30 °C for 7 days. No differences in the apparent growth rates were found when the same strains were grown on medium that lacks ethanol (not shown).

Efficacy of SSD1 variants in ethanol tolerance

Toward understanding the structure–function relationship of SSD1 with respect to ethanol tolerance, we compared the ability of all three alleles, SSD1-V, ssd1-d, and SSD1-V-F6 to confer ethanol tolerance. The alleles were expressed either in a high-copy 2µ-based plasmid or in a low-copy centromeric plasmid in the original strain W303-1A, as well as in the deletion strains W303-1Assd1∆ and BY4741ssd1∆. The resulting clones were tested for growth in the presence of 7% ethanol (Fig. 6). At identical conditions, the deletion mutant, BY4741ssd1∆, expressing all three variants of SSD1 allele, demonstrated more robust growth than W303-1A, including a partial rescue with ssd1-d (Fig. 6a). Expression of the complete SSD1-V gene from the centromeric plasmid resulted in fully tolerant phenotype in all strains tested, while multi-copy plasmid gave a partial rescue in all backgrounds. Expression of F6 fragment was much less efficient in terms of ethanol tolerance, especially, at low-copy numbers. The F6 fragment, for unknown reasons, was slightly more efficient than SSD1-V-F6 derived from it. Thus, SSD1-V must be expressed at certain levels to efficiently support proliferation in the presence of ethanol. Its overexpression reduces ethanol tolerance.


Expression of the SSD1-V allele in low-copy number augments ethanol tolerance of W303-1A, W303ssd1∆, and BY4741ssd1∆ more than other SSD1 variants. Aliquots (5 µL) of 10-fold serial dilutions of BY4741ssd1∆ (a), W303-1A (b), and W303ssd1∆ (c) cells harboring the indicated vectors (‘CEN’- for centromeric plasmid, pRS416 or pRS316. ‘2μ’- for 2µ-based plasmid, pRS426) were plated onto plates containing YPD + 7% ethanol and cultured at 30 °C for 5 days (a and b) or 6 days (c). No differences in the apparent growth rates were found when the same strains were grown on medium that lacks ethanol (not shown).

Expression of either PDE2 or SSD1-V renders the cells of W303-1A more resistant to cell wall digestion

Both PDE2 and SSD1 were implicated in the maintenance of cell wall integrity function (Tomlin et al., 2000; Kaeberlein & Guarente, 2002). The defective ssd1-d allele is known to render cells susceptible to cell wall-perturbing agents, for example, Calcofluor white (Mir et al., 2009). PDE2 is reported to be involved in the maintenance of cell wall integrity, because its overexpression rescues the sorbitol dependence of cell wall-defective mutants (Tomlin et al., 2000). Its deletion in C. albicans causes cell wall defects and increases susceptibility to cell wall-disrupting agents (Jung et al., 2005). All these observations point to the importance of cell wall integrity in ethanol tolerance. Indeed, the highly ethanol-tolerant Sake yeast shows strong resistance to zymolyase, a lytic enzyme containing mainly 33-1, 3-glucanase, as well as to K1 killer toxin, whose target resides in the cell wall (Hara et al., 1976a, b).

Treatment with zymolyase in hypotonic buffer causes lysis of yeast, when its cell wall mechanical stability is compromised by the digestion of glucan fibers. The time course of digestion is thought to depend upon thickness of the cell wall (Jung et al. 2005). In our hands, the kinetics of digestion differed slightly from the kinetics described in the study by Ovalle et al. 1998. In all examined strains, we observed an initial burst of degradation during first 12 min of the assay followed by a stable degradation rate for the duration of an hour with a gradual slowdown later (Fig. 7). Unbiased observation of the results regardless of interpretation method indicated that the ethanol-tolerant Sake strain was the most stable among investigated strains, and W303-1A harboring an empty vector was the least resistant to zymolyase action. The cell wall of W303-1A expressing either PDE2 or SSD1-V was more robust than that of the W303-1A and only slightly less stable than that of the Sake. Following the study by Ovalle et al., 1998, we analyzed the kinetics after the initial stabilization of the degradation rate and until the onset of the slowdown. Raw data of three independent assays for each sample between the reaction time of 12 min and 2 h were fit to first order kinetic model resulting in the following degradation rate constants: Sake 0.128 ± 0.002 h−1; W303-1A cells harboring Yep13-PDE2 0.168 ± 0, 001 h−1; W303-1A cells harboring pRS416-SSD1-V 0.279 ± 0.002 h−1; and finally, W303-1A cells harboring empty pRS416 0.570 ± 0.003 h−1, thus confirming the unbiased observation of the results.


W303-1A cells expressing either PDE2 in a high-copy number or SSD1-V in a low-copy number are more tolerant to zymolyase than the parental W303-1A. Sake and W303-1A cells harboring empty pRS416, pRS416-SSD1-V, or Yep13-PDE2 were allowed to grow to stationary phase at 30 °C on YPD or a selective media. The cells were collected by centrifugation, washed three times with deionized water, and resuspended to an OD600 = 0.5 in TE buffer. This was followed by the addition of zymolyase at the time t = 0. The optical density, OD600, of cell suspensions was then recorded at 3-min intervals. Three independent assays were performed for each sample; all were highly reproducible (See for the statistical analysis).

Possible role of SSD1 and PDE2 in ethanol tolerance

It must be noted that among ethanol-tolerant mutants of W303-1A obtained by the selection in turbidostat, we have identified mutations that replace the stop codon in ssd1-d to a codon encoding for either one of three amino acids: lysine, tyrosine, or leucine, thus giving rise to a full length protein analogous to Ssd1-V (Avrahami-Moyal, et al., 2012). In this work, we found that an SSD1 form of intermediate length, at least partially, rescues W303ssd1Δ. The C-terminal domain of the SSD1 protein contains a putative RNA-binding site of 328 amino acids in the region between residues 689 and 1014 (Uesono et al., 1997). This region is implied in localized control of gene expression for several cell wall-related genes mediated by the binding of their mRNAs to Ssd1 (Jansen et al., 2009). The defective ssd1-d allele almost completely lacks this RNA-binding site. In contrast, the truncated SSD1-V-F6, which we have cloned from the F6 genomic fragment, has the length of 838 amino acids, including 149 C-terminal amino acids within the putative RNA-binding site. It seems reasonable to suggest that this short piece of the RNA-binding site allows the SSD1-V-F6 to function, albeit incompletely, and only when expressed in a high-copy number (Fig. 6).

PDE2 encodes for the high-affinity cAMP phosphodiesterase. Its overexpression significantly reduces cAMP levels in the cell (Sass et al., 1986). Proper regulation of cAMP levels in yeast is critical for proliferation and viability. Certain threshold cAMP level is required to allow for cell cycle to proceed (Matsumoto et al., 1982). Accumulation of cAMP renders yeast highly sensitive to various stresses, most probably, because cAMP-dependent PKA phosphorylates and, thus, inhibits the two major transcriptional activators, Msn2 and Msn4, which are responsible for transcribing hundreds of stress-related genes (MartinezPastor et al., 1996; Moskvina et al., 1998; Stanhill et al., 1999).

Following this logic, we prepared a double deletion mutant, W303msn2msn4∆, lacking both regulators. Unexpectedly, overexpression of PDE2 in this mutant enabled it to proliferate in medium containing 7% ethanol as well as it did in W303-1A (Fig. 8). Therefore, it is possible that PDE2-dependent phenotype is not mediated by the transcription factors MSN2/4, but rather by some other factors (Pedruzzi et al., 2000).


The ethanol tolerance conferred by PDE2 overexpression is not mediated via MSN2/4. Aliquots (5 µL) of 10-fold serial dilutions of W303-1A and W303msn2Δ msn4Δ cells harboring the indicated plasmids were spotted onto plates containing YPD + 7% ethanol and cultured at 30 °C for 7 days. No differences in the apparent growth rates were found when the same strains were grown on medium that lacks ethanol (not shown).

In conclusion the approach taken in this study, which was to use ethanol-sensitive laboratory strain as the platform for improving ethanol tolerance by introducing genetic elements from more tolerant strains, has been proven successful. The genes, identified here, PDE2 and SSD1, were not found by similar library screen (Hong et al., 2010). SSD1 and PDE2, together with UTH1 identified in a parallel study (Avrahami-Moyal et al., 2012), indicate the primary importance of the cell wall rigidity and thickness in improving ethanol tolerance in W303-1A and beyond. One can envision the next step of this program, when a more tolerant laboratory strain such as W303-1A expressing SSD1-V or PDE2 would be transformed with a genomic library of Sake yeast.


The funds for this work were provided by the YISSUM R&D Company of the Hebrew University.


  • Editor: Jens Nielsen


View Abstract