Parental yeast strain BY4743 (MATa/α his3Δ1/his3Δ1 leu2Δ0/leu2Δ0 LYS2/lys2Δ0 met15Δ0/MET15 ura3Δ0/ura3Δ0) and six derived homozygous mutants (Δmsn2, Δmsn4, Δgcn4, Δire1, Δhac1, and ΔYap1p), all obtained from the EUROSCARF collection, were used for the ethanol tolerance assay done in 96-well plates. Haploid strain BY4741 (MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0) was used in the acquired stress resistance experiments, the ultrastructural changes under ethanol stress and the qRT-PCR assays. The YPL004 strain (BY4741 Kar2-sfGFP::HIS; UPR-mCherry::URA) (Lajoie et al., 2012) was kindly provided by Dr. Erik L. Snapp. The BY4741 strains that expressed GFP-Atg8, ero-GFP and Sec63-GFP reporters were transformed with the pRS316-GFPAtg8 (ATG8 promoter, CEN, URA3), pPM28 (TDH3 promoter, CEN, URA3) and pPS1622 plasmids (SEC63 promoter, CEN, URA3), respectively (Prinz et al., 2000; Suzuki et al., 2001; Merksamer et al., 2008), by the LiAc/SS carrier DNA/PEG method (Gietz and Woods, 2002).

The basal growth media selected for the experiments contains 2% glucose in standard synthetic complete (SC) medium (Formedium™), SD-ura medium and SD-ura-his without inositol. Media were modified whenever necessary with DTT (1–2 mM) or ethanol [0, 6, 8, and 10% (v/v)]. Cultures were incubated with agitation at 28°C. All the experiments were carried out with three replicates.

An overnight BY4741 preculture grown in SC medium at 28°C was split into three cultures: two received a single dose of mild stress (0.1 and 0.2 μg/mL tunicamycin), and the other served as a control to which no stress was added. Each culture was grown in a total volume of 30 mL of SC. Once in the exponential phase [an optical density (OD₆₀₀)~0.4], cells were collected, supernatant was removed and cells were resuspended to OD₆₀₀ of 0.1 in 96-well plates that contained SC or SC plus 6, 8, and 10% ethanol. Yeast growth curves were monitored in a SPECTROstar Omega instrument (BMG Labtech, Offenburg, Germany) until yeast cells reached the stationary phase. Six replicates were used per condition. The areas under the OD₆₀₀-time curves, linearly related to both the maximum population level and maximum specific growth rate, were calculated by integration of the curves using OriginPro 7.5 software (OriginLab Corporation, Northampton, USA) (Arroyo-López et al., 2010).

The BY4743 cells that contained the multicopy pMCZ-Y plasmid (Mori et al., 1996) were grown overnight in SD-ura medium and were allowed to reach the exponential phase (OD₆₀₀ ~0.4). Then 6% ethanol (v/v) and 2 mM DTT were added separately to a final volume of 150 mL to induce LacZ gene expression. Yeast cells were harvested 1, 2, and 4 h after stress exposure. Preparation of cell extracts and the β-galactosidase activity assay were carried out as previously described (Inoue et al., 1998). Protein concentration was measured with the Bio-Rad Laboratories protein assay according to the manufacturer's instructions. Units are defined as 1000 × A₄₃₀/(c·t·v), where c is the protein concentration in mg/ml; t is the reaction time, and v is the volume of extract in μl. Statistical significance in triplicates was determined by the Student's t-test.

A RT-PCR assay was done in order to identify the activation of HAC1 by splicing under ethanol stress in BY4741. RNA isolation was performed using commercial kit Nucleospin RNA (Macherey-Nagel) according to manufacturer's instructions. The concentration and purity of RNA was determined in a spectrophotometer (NanoDrop ND1000) and RNA integrity was verified by agarose gel 1%. First strand cDNA were generated by EuroScript/RNase inhibitor mix (Eurogentec) with a mixture of oligo dT₁₅ VN, random nonamer primers and 400 ng of total RNA. The product of the reverse transcription reaction was used for amplification of the bands belonging to active/inactive HAC1 (HAC1ⁱ and HAC1^u, respectively) with the pair of primers AGGAAAAGGAACAGCGAAGG and TTCAAATGAATTCAAACCTGACT, according to previous studies (Kumar et al., 2006). Reactions were subjected to 25 PCR cycles of 94°C for 30 s, 54°C for 30 s and 72°C for 60 s. The PCR products were resolved on a 2% (w/v) agarose gel and visualized using a CCD COHU camera and an image analysis software (Scion Image for Windows).

We analyzed the growth of the BY4743 mutant strains for genes MSN2, MSN4, GCN4, IRE1, HAC1, and YAP1 and the wild-type strain in the SC medium modified with 8 and 10% ethanol (v/v). Growth was monitored at 600 nm every 30 min in 96-well plates for 72 h. The areas under the OD₆₀₀-time curves were calculated as indicated above. Dunnett's test for multiple comparisons was used to detect samples whose fractional area significantly differed from the wild-type strain. Statistical tests were performed by GraphPad Prism, version 5.0, for Windows (GraphPad Software, San Diego California USA).

BY4741 cells were precultured overnight in SC medium at 28°C and transferred to a main culture to the log phase. Then cultures were collected and divided into three 250 mL flasks that contained SC medium with 8% ethanol, 1 mM DTT or the control conditions (no stressors), which were incubated at 28°C and 150 rpm. 10 OD units of cells were collected by centrifugation 0, 3, and 6 h after stressor addition. A cell pellet was pre-fixed with 2.5% glutaraldehyde in 0.1 mol/L PBS (pH 7.2–7.4). Fixed samples were washed 3 times with 0.1 mol/L PBS, and post-fixed with 1% osmium tetroxide (OsO4) at 4°C for 1 h. After dehydration through an ethanol gradient elution and embedment in LR-White resin, the embedded samples were sectioned with a diamond knife (DiATOME) in a Leica EM UC6 (Leica Microsystems). Next 70-nm-Ultrathin sections were stained with lead citrate, viewed and photographed with a JEOL JEM-1010 TEM (JEOL Ltd., Japan). ER length was analyzed from 30 independent images that belonged to 30 different cells for each time and condition, as described by Bernales et al. (2006) using ImageJ according to Rønn et al. (2000). Dunnett's test for multiple comparisons was performed to compare the ER length for each condition and time using a 99% confidence interval. To avoid bias, data were normalized to the cutting area of the cell.

The vacuolar area was also analyzed in the same cells. For this purpose, we added the area of the vacuoles from each image, normalized according to the cutting area, and we finally divided that area by the number of vacuoles in the image to obtain the percentage occupied by each vacuole in the cell.

The Ero-GFP reporter (pPM28) was introduced into BY4741 cells by the LiAc/SS carrier DNA/PEG method and transformants were selected in SD-ura medium. Cells were cultured in SD-ura medium to the log phase, harvested by centrifugation, and transferred to the SD-ura fresh medium modified with 8% ethanol, 1 mM DTT, or not modified. Each condition was done in triplicate. GFP fluorescence at 405 nm and 488 nm was measured by flow cytometry in a BD FACSVerse™ flow cytometer (BD Biosciences). The eroGFP ratios, defined as the ratio of fluorescence from excitation at 488 vs. 405 nm, and expressed in log₂ space and normalized to the untreated case, were plotted.

The YPL004 strain was grown overnight in SD-ura-his medium and was allowed to reach the exponential phase. Cells were washed 3 times with ultrapure water to remove residual inositol. Then the culture was divided into sterile centrifuge tubes, pelleted, and incubated with the different SD-ura-his media without inositol, and modified with 1 mM DTT and 8% (v/v) ethanol. Cells were grown at 28°C, sampled every 2h. Cells from samples were pelleted and frozen in liquid nitrogen until used. Fluorescence was measured by an LSR Fortessa flow cytometer (BD Biosciences) equipped with a 488-nm laser with a 525/50 bandpass filter for sfGFP and a 561-nm laser with a 610/20 bandpass filter for mCherry, and the FACSDIVA software to compile. fcs files. Files were analyzed by FloJo (Tree Star Ashland, OR, USA). The median fluorescence intensities (MFI) were calculated for each channel and normalized to zero time for each condition. Biological triplicates were performed in all cases.

The BY4741 strain was grown overnight in SC medium at 28°C and transferred to fresh SC medium until the log phase was reached (OD₆₀₀ ~0.4). Then the culture was divided into two flasks, the second with ethanol to obtain a final concentration of 8% (v/v). Each flask was divided into three flasks to obtain three replicates. Next 50 mL sample were taken at different times. Cells were quickly harvested by centrifugation, washed and frozen in liquid N₂. RNA was extracted using the commercial kit NucleoSpin® RNA (Macherey-Nagel) according to the manufacturer's instructions. Its integrity was verified by gel electrophoresis and NanoDrop Spectrophotometer ND-100. Reverse transcription reactions were carried out using EuroScript Reverse Transcriptase (EuroGentec, Belgium) following the manufacturer's protocol.

The pair of primers used for qRT-PCR was GCAACTGCTACAAACTGGGC-TCCCGCGAAAAAGGCAAATG for ITR1 and TTTATGCCCTCGGTATCGGC-GCAGTAGCGTAGGATGTCCC for ITR2. LightCycler® 480 SYBR Green I Master (Roche) was applied for each qRT-PCR reaction and a PCR was run in the LightCycler® 480 real-time PCR system. All the samples were processed for the melting curve analysis, amplification efficiency and DNA concentration determinations. A mixture of all the samples and serial dilutions (10⁻¹ to 10⁻⁵) was used as the standard curve. Two different constitutive reference genes were used (ACT1 and RDN18-1) to normalize the amount of mRNA. These genes showed excellent uniformity at the expression levels under ethanol stress (Trotter et al., 2002; Penacho et al., 2012).

To study the relation between UPR activation and ethanol stress we performed a cross-protection assay. In the case that the UPR is activated after ethanol stress, cells with previous UPR activation should respond better to ethanol stressed cells than untreated cells. Thus, we evaluated the growth of BY4741 strain in SC medium with 6, 8, and 10% (v/v) ethanol after pre-exposure to a tunicamycin stress (Figure 1). We have used low tunicamycin levels in order to avoid indirect general stress response effects observed at high levels (Pincus et al., 2014). Control experiments without tunicamycin and without ethanol were also performed.

Our results showed that the doubling time in the cells pre-exposed to 0.2 μg/mL of tunicamycin was significantly lower than in non-pre-exposed cells under ethanol stress (Supplementary Table S1). This difference was more pronounced under 8 and 10% ethanol, which indicates a growth advantage to cope with this stress. For the cells grown in tunicamycin, and subsequently in SC medium without ethanol (control medium), doubling times were similar to the non-pre-exposed ones. Therefore, tunicamycin pre-exposure did not generate negative effects on growth when cells were passed to fresh SC medium.

We also observed a longer lag period in the non-pre-exposed cells compared to the pre-exposed cells in SC medium. Contrarily to what we thought, a scenario in which pre-exposed cells should have higher lag periods due to energetic costs derivate of an activated UPR response in an environment without stress, it seems that once the cell activates this response its growth improves if it is passed to a new culture medium (Figure 1). As tunicamycin pre-exposure generated a shorter lag phase in the control medium, we calculated the fractional area of each sample in order to normalize data, and to ensure that better growth in ethanol after tunicamycin pre-exposure was not only due to this strain growing better than the non-pre-exposed ones. The results confirmed the previous results and showed increased fractional area in samples with increasing amounts of tunicamycin used in the pre-exposure (Figure 1D). Therefore, it seemed that, somehow, UPR activation prepared the cell to face subsequent ethanol stress and to facilitate their entry into the exponential phase.

To confirm with alternative methods, that ethanol stress induces the UPR response, not previously described in the literature, we transformed the wild type BY4743 yeast strain with pMCZ-Y (UPRE-CYC1-lacZ gene carried on a multicopy vector) that was designed as a UPR reporter (Mori et al., 1996). Assays to measure β-galactosidase activity at this 22-base pair UPRE (UPR element), a sequence were Hac1p binds to activate UPR, were performed to determine the UPR induction levels in the presence or absence of ethanol at different post-treatment time points (Figure 2A). Control experiments without ethanol addition were also performed. As expected, increased β-galactosidase activity was seen after adding 2 mM DTT, an agent that causes protein misfolding selectively in the ER. Interestingly in the extracts from the ethanol-stressed cells, β-galactosidase activity significantly increased at the first time point, and then remained constant after 4 h of stress compared with the untreated cells. An important aspect to consider is that the β-galactosidase activity in ethanol treated cells was 1.95 times higher than in the cells treated with DTT after 1 h of treatment. This fact suggests that ethanol not only induced the UPR response, but this activation was also triggered rapidly and forcefully after ethanol addition. The differences in the induction pattern are probably related to growth since ethanol stress reduced growth and DTT stressed cells still continue growing (results not shown). The samples grown without ethanol did not show any noticeable β-galactosidase activity.

To obtain a measure of the UPR response in single cells, we used a plasmid reporter, called UPR-mCherry (Merksamer et al., 2008), which encodes the red fluorescent protein mCherry, and is driven by a minimal CYC1 promoter and four tandem UPR elements that the UPR transcription factor Hac1p binds. The results after 6 h of growth in 8% ethanol in SD-ura medium showed clear UPR activation compared to the control medium, as observed with DTT 2 mM (Figure 2B).

A qualitative profile of the UPR activation was done monitoring HAC1 splicing, a clear molecular proof of UPR activation, in BY4741 strain grown in SC medium in presence and absence of 8% ethanol (Figure 2C). As shown, UPR activation in yeast under ethanol occurs very quickly, since the splicing of HAC1 mRNA and the active form of the transcription factor (HAC1ⁱ) took place a few minutes after the ethanol shock and was decreasing in intensity until basal levels after 4 h. On the other hand, unstressed cells showed a low basal level over time. Is noteworthy that, although we did not observe the splicing of HAC1 beyond 4 h of stress, the β-galactosidase activity remained constant probably due to the residual presence of the enzyme in the cell.

To elucidate the relationship between ethanol tolerance and the distinct transcription factors and key components (Msn2p/Msn4p, related with the general stress response; Yap1p, related with the oxidative stress response; Hac1p, related to UPR; Ire1p, UPR transmembrane sensor protein; and Gcn4p, required for the induction of a majority of UPR target genes during ER stress), the BY4743 diploid strain and its mutants derivatives Δmsn2, Δmsn4, Δgcn4, Δire1, Δhac1, and Δyap1 were grown in the SC medium modified with 8% an 10% (v/v) of ethanol. Control experiments without ethanol addition were also performed. Results shown as Δhac1 was the strain with the lowest growth, its fractional area was reduced by 43% and 62% compared to wild type in the presence of 8 and 10% ethanol, respectively (Figure 3).

Surprisingly, we observed no significant differences in ethanol tolerance at 8% between the wild-type and Δyap1 strains. Given that YAP1 encodes the transcription factor needed for the response to oxidative stress, and ethanol generates this kind of stress; we expected a decrease in the fractional area at 8% ethanol; although we observed a decrease of 30% of its fractional area at 10% ethanol relative to BY4743. The growth difference between the Δmsn2 and Δmsn4 mutants at 10% ethanol was noteworthy, and emphasized the importance of Msn4p in ethanol tolerance regarding its homolog Msn2p.

One way of activating the UPR pathway, regardless of the presence of unfolded proteins in cells, is lack of inositol (Nikawa and Yamashita, 1992; Promlek et al., 2011; Lajoie et al., 2012), a precursor in the synthesis of phosphatidylinositol (PI) (Greenberg and Lopes, 1996). Thus, lipid imbalance can lead to UPR activation independently of unfolded proteins accumulation (Pineau et al., 2009; Promlek et al., 2011; Lajoie et al., 2012). We want to determine if ethanol could lead to a massive outflow of inositol from the cell, which could activate the UPR response, due to the damage generated in the plasma membrane, and to the inositol transporters, Itr1p and Itr2p.

To verify this idea, a modified BY4741 strain (Kar2-sfGFP::HIS; UPR-mCherry::URA), called YPL004 (Lajoie et al., 2012), was grown under ethanol in the SD-ura-his medium without inositol to study both UPR activation as the expression of UPR target Kar2p (Figure 6). Control experiments with inositol were also performed. If UPR activation by ethanol was due solely to loss of inositol in the cell through a compromised plasma membrane, we expected to observe a faster/higher activation response in the ethanol-stressed cells in a medium without inositol than in the cells grown only in a medium without inositol due to the rapid loss of their endogenous inositol pool through the membrane.

Surprisingly, presence of ethanol decreased the UPR signal, with values of 1.83 and 1.64 after 4 and 6 h of stress, respectively, which were lower than the medium without inositol alone. This activation continued to increase after 8 h of stress (Figure 6A). In the medium without inositol, UPR activation decreased after 6 h, probably due to a later accumulation of inositol by INO1 gene expression. The UPR target Kar2p signal displayed a similar behavior (Figure 6B). Our results indicated that UPR triggered under ethanol stress was not due to decrease in the concentration of intracellular inositol by plasma membrane damage.

Furthermore, a qRT-PCR analysis in ethanol stressed and unstressed cells was done to obtain a relative quantification of ITR1 gene expression, which encodes an high-affinity inositol transporter, whose mRNA level increases under inositol depletion and is repressed by inositol (Nikawa et al., 1993; Azab et al., 2007). If inositol was actually lacking in the cell due to damage caused by ethanol, we expected to see an overexpression of the ITR1 gene. Instead, presence of inositol led to the inactivation of transporter activity and reduced ITR1 transcription (Nikawa et al., 1993; Lai et al., 1995). The ITR2 gene was used as a control because it has a constitutive expression regardless of the intracellular inositol concentration (Figure 7; Nikawa et al., 1993; Azab et al., 2007). After measuring the relative expression of ITR1 and ITR2 in the SC medium, either modified or not with 8% (v/v) ethanol in BY4741 cells, our results did not show any ITR1 overexpression. Once again, this suggested that presence of ethanol did not contribute to intracellular inositol depletion.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.