Identification of ethanol-sensitive mutants
To identify ethanol-sensitive mutants, we replica-plated the complete EUROSCARF collection, consisting of 4868 deletion mutants, onto glucose-based medium containing 7% ethanol. Initially, 526 strains displayed apparently reduced growth. Estimating the difference between growth and slow growth is difficult to do by eye on plate media. Therefore, the 526 strains were first tested by re-streaking on plates containing 6% ethanol. This resulted in approximately 200 strains with reduced growth that were further tested by plating serial dilutions on 6% ethanol-containing plates. Strains that displayed a clear reduced growth phenotype were in this way unambiguously identified. These tests finally identified 14 strains that did not grow and 20 strains that showed a clear slow-growth phenotype on 6% ethanol. Independent isolates of both mating types were re-tested and showed identical phenotypes. The remaining deletion strains displayed no or only a weak growth phenotype on plates containing 6% ethanol. Figure 1 shows an example of some of the phenotypes. The 34 mutations were grouped according to gene annotation. In each group other genes with similar functions were tested on ethanol plates. This identified 12 further mutations resulting in slow growth on ethanol-containing medium. The 46 genes identified are shown in Table 1. Bold characters indicate the genes identified in the initial screen.
Figure 1. Growth phenotypes of the wild-type and representative mutants on ethanol-containing medium (6% v/v). Ten-fold serial dilutions were spotted onto YPD plates without (Y) and with (E) ethanol. Plates were incubated at 30 °C for 5 days
Download figure to PowerPoint
Table 1. List of genes whose deletion leads to growth deficiency on complex glucose-based medium containing 6% ethanol
| ||Mitochondrial function||Vacuolar function||Pkc1 pathway||Actin/tubulin folding||Histone acetylation||Miscellaneous|
|No growth||IMG1||VMA10||BEM2|| ||KAR3|
| ||VPS24||SWI4|| || ||POP2|
Five mutations are involved in mitochondrial function and only one mutant, img1, did not grow at all in the presence of 6% ethanol. Kubota et al. (2004) reported 22 genes involved in mitochondrial function that were important for growth on 11% ethanol, indicating that certain mitochondrial functions are important to resist high levels of ethanol. We tested all these mutants at 6% ethanol and found none of them to be sensitive. Remarkably, none of the mitochondrial mutants we identified is reported in the screen of Kubota et al., suggesting that mechanisms of ethanol sensitivity depend on ethanol concentration. We identified 15 mutations in genes involved in vacuolar function and vesicular transport to the vacuole. Three of these mutations are in genes specifically involved in the metabolism of phosphatidyl inositol 3,5-bisphosphate. Genes involved in vacuolar protein sorting and appearing in our screen are represented in all classes of vacuolar morphology except one, class F (Raymond et al., 1992). This strongly suggests the importance of vacuolar function and maintenance under ethanol stress. Kubota et al. lists 45 genes involved in vacuolar function that are important at high ethanol concentrations. We tested all vps mutants and all mutants listed by Kubota et al. Ten of the genes required for growth at 6% ethanol were not picked up in their screen. We identified seven mutants in the PKC1 cell integrity pathway. This pathway is involved in regulation of many responses in yeast cells, including G1–S phase transition, actin polarization/depolarization and the response to cell wall stress (Delley and Hall, 1999). Interestingly, SIT4 and BEM2, which encode proteins involved in the downregulation of signalling through the PKC1 pathway, were essential for growth of yeast on plates containing 6% ethanol. Genes SLG1 and ROM2, encoding proteins involved in upregulation of signalling, are important but not essential for growth in the presence of 6% ethanol (Table 1). We have identified GIM4 and GIM5, encoding subunits of the hetero-oligomeric co-chaperone GimC complex, as important for growth of BY4742 in the presence of 6% ethanol. This may be rationalized as follows. A subset of newly synthesized polypeptides requires folding by a chaperonin. In eukaryotes the formation of native-state tubulin and actin requires folding by the TriC/CCT complex. Efficient folding of proteins by the TriC/CCT complex requires GimC. By interactions with both folding intermediates as well as with TriC/CCT, GimC accelerates actin folding at least five-fold (Siegers et al., 2003). The F-actin dispersion occurring during ethanol stress (Kubota et al., 2004) could be lethal in mutants depleted for components of this complex.
ADA3, GCN5, SPT3 and SPT7 of the SAGA complex are also important for growth of BY4742 in the presence of 6% ethanol. The SAGA (Spt-Ada-Gcn5-histone acetyltransferase) complex is a multi-subunit complex involved in expression of ∼10% of the transcribed part of the genome. It comprises: (a) the Ada proteins (Ada1p, Ada2p, Ada3p, Gcn5p and Ada5p); (b) the Spt proteins (Spt3p, Spt7p, Spt8p and Spt20p); and (c) a number of TAFs (TATA-binding protein associated factors) in addition to the product of the essential TRA1 gene. Both the SAGA complex and TFIID deliver TBP (TATA binding protein) to promoters, leading to gene expression, and both complexes bind TAFs. Whereas TFIID is involved in expression of ∼90% of the expressed genome and is involved in expression of housekeeping genes, the SAGA-controlled genes are largely stress-induced (Huisinga and Pugh, 2004). Our data show the need for components of this complex during ethanol stress, and hence suggests the involvement of a SAGA-dependent response during ethanol stress.
We found 14 mutants that did not belong to any of the mentioned classes. Interestingly, four of these, FEN1, PLC1, ERG6 and SUR4, are involved in lipid biosynthesis, which is in addition to those involved in phosphatidyl inositol 3,5-bisphosphate synthesis (VPS34, VAC14 and FAB1). Lipid metabolism was reported to play a role in response to ethanol stress (Swan and Watson, 1999). fen1 is synthetically lethal with erg6, underlining the importance of the sterol–sphingolipid interaction (Eisenkolb et al., 2002). We tested all other viable mutants in ergosterol biosynthesis but found no strong phenotype at 6% ethanol.
Cross-sensitivity of the mutants
We tested whether the mutants we identified are sensitive to other stresses relevant for brewing conditions. The results are summarized in Table 2, and Figure 2 gives an example of the growth behaviour of three of the mutants. Only three strains showed a growth defect on 0.4 M NaCl, indicating that osmotic problems are in most cases not the reason for the growth defect on 6% ethanol. Nine of the mutants showed a growth defect on sorbic acid plates, suggesting a defect related to acid stress. Ethanol stress has been suggested to resemble temperature stress (Piper, 1995). However, only six of the mutants showed a growth defect at 37 °C, indicating only a limited overlap between ethanol and temperature stress. Several of the mutants we identified are part of the cell wall integrity pathway. Therefore, we tested all our mutants for growth on plates containing the cell wall-perturbing agent Calcofluor white. Interestingly, 15 of the mutants that showed defective growth on 6% ethanol were also sensitive to Calcofluor white. This shows that some of the mechanisms leading to ethanol sensitivity in the deletion strains are related to defects induced in the cell wall. Only rib4, vma10 and vps34 were sensitive to all five types of stress.
Figure 2. Cross-sensitivity of the wild-type and representative mutants. Ten-fold serial dilutions of the indicated strains were plated on YPD or YPD supplemented with sorbic acid (0.5 mM, pH 4.5) or Calcofluor white (50 µg/ml) and incubated for 5 days at 30 °C or 37 °C, as indicated
Download figure to PowerPoint
Table 2. Sensitivity of ethanol-sensitive deletion strains to other stress conditions
| ||0.4 M NaCl||0.5 mM sorbic acid||37 °C||Calcofluor white|
|No growth|| ||RIB4||BEM2||FAB1|
| ||RMD8|| ||VPS5|
Activation of signalling pathways
The transcription factors Msn2p and Msn4p are involved in activation of genes responding to a number of environmental stresses, such as heat shock, osmotic stress and high-ethanol stress (Martinez-Pastor et al., 1996). Deletion of MSN2 alone or MSN4 alone does not result in an ethanol-sensitive phenotype. Deletion of both genes results in a slow-growing, very stress-sensitive strain (data not shown; and Estruch and Carlson, 1993). We made use of a C-terminal GFP-MSN2 fusion protein to determine whether the MSN2/MSN4 pathway is activated under our ethanol stress conditions (Figure 3A). In exponentially growing cells in YPD medium, most fluorescence was found in the cytosol. Half an hour after addition of ethanol to a concentration of 6%, most fluorescence was found in the nucleus, indicating activation of the general stress response. Among the genes that are transcriptionally activated by this general stress pathway, we only found one to be important for growth in the presence of 6% ethanol, viz. the TPS1 gene, encoding trehalose-6-phosphate synthase. A search in the yeast genome resulted in a number of genes with two or more STRE elements in their promoters (Moskvina et al., 1998), but again, among those genes only TPS1 is present in our collection. This suggests that although a wide array of genes is upregulated during ethanol stress by the general stress response pathway, only one of them is important for growth under those conditions.
Figure 3. Activation of signalling pathways. (A) Localization of Asr1p–GFP in YPD-grown cells and cells in 6% ethanol. (B) Localization of Msn2p–GFP in YPD-grown cells and cells in 6% ethanol. (C) Total amounts of Slt2p (top) and amounts of phosphorylated Slt2p (P-Slt2p) in the indicated mutants. The conditions applied are: 1, YPD; 2, YPD 6% ethanol; 3, YPD 6% ethanol and 1 M sorbitol; 4, YPD with Calcofluor white; 5, YPD with Calcofluor white and 1 M sorbitol
Download figure to PowerPoint
The transcription factor Asr1p responds to the presence of ethanol in the growth medium but not to heat, osmotic and oxidative stresses (Betz et al., 2004). Cells deleted for ASR1 become sensitive to 1% butanol in the growth media but do not show a growth defect in the presence of 1% ethanol. We did not observe a growth defect in the presence of up to 8% ethanol (not shown). The Asr1–GFP fusion protein, however, does respond to the presence of ethanol under the conditions of our screen, by translocating to the nucleus (Figure 3B). The genes under control of Asr1p have not been identified, but either they are not essential for the response to ethanol or their regulation may involve other redundant transcription factors, as is the case for the MSN2/MSN4 pathway.
Several mutations in the cell integrity pathway have strong growth phenotypes in the presence of ethanol. Therefore, we analysed the phosphorylation of the downstream MAP kinase Slt2p. In wild-type (BY4742) cells growing exponentially in YPD, the level of Slt2p phosphorylation is low (Figure 3C, BY4742, 1). Phosphorylation can be induced in the presence of Calcofluor white (Figure 3C, BY4742, 4), which induces cell wall stress. The phosphorylation is absent when the cells are treated with Calcofluor white in the presence of 1 M sorbitol (Figure 3C, BY4742, 5). In the presence of 6% ethanol, phosphorylation of the MAP kinase is also induced, and the stress is likewise relieved by the addition of sorbitol to 1 M (Figure 3C, BY4742, 2, 3). The total expression levels of Stl2p remain the same under all conditions. The high phosphorylation levels in the presence of ethanol could be explained by an inhibitory effect of ethanol on phosphatases that might dephosphorylate Slt2p. However, since sorbitol does not prevent ethanol from entering the cells but does prevent phosphorylation, we do not find this a likely explanation. We therefore propose that the phosphorylation of Slt2p stimulated by ethanol is induced through the cell integrity pathway. This signalling pathway is already known to be involved in many stress responses, e.g. oxidative stress, heat shock, hypo-osmotic stress, actin stress and ER stress (Bonilla and Cunningham, 2003; Harrison et al., 2004; Bryan et al., 2004; Vilella et al., 2005). We decided to analyse Slt2p phosphorylation in three mutants with strong ethanol-sensitive phenotypes. In the bem2 mutant, Slt2p is phosphorylated in response to both Calcofluor white and ethanol, as in the wild-type strain. The total amount of Slt2p is also similar under these conditions (Figure 3C, bem2).
Interestingly, in the smi1 mutant the total amount of Slt2p is significantly higher than in the wild-type (Figure 3C, smi1). The presence of Calcofluor white or ethanol results in a higher phosphorylation than in the wild-type. The higher total amount of Slt2p in the smi1 strain is likely to cause the detected higher amount of phosphorylated Stl2p as compared to the wild-type. Smi1p interacts with Slt2p and is necessary for signalling through the cell integrity pathway (Martin-Yken et al., 2003).
The sit4 mutant behaves similarly to the wild-type but may have slightly lower levels of Slt2p. Slt2p phosphorylation is also low but consistently higher in the presence of ethanol or Calcofluor white (Figure 3C, sit4).
To determine the relevance of our screen for brewery fermentations, we analysed the behaviour of three of the mutants in 200 ml stirred microfermentations. These fermentations can give a first indication for the potential behaviour of yeast strains in larger-scale fermentations. Because BY4742 is not able to utilize maltose, we used complex medium supplemented with 20% glucose. Figure 4 shows the decrease in density of the medium, measured in degrees Plato, over time. The wild-type (BY4742) is able to ferment almost all the sugar in the medium in 70 h. The two mutants in the cell integrity pathway bem2 and sit4 need a substantially longer time to ferment to completeness. Surprisingly, the smi1 mutant ferments just as rapidly as the wild-type. The high expression level of Slt2p may help to overcome the defect in this strain. For two of three strains tested, increased ethanol sensitivity also resulted in impaired high-gravity fermentation. Plate assays for reduced growth on ethanol might give a fair indication for the behaviour of yeast strains under fermentation conditions.
Figure 4. Stirred high-gravity fermentations of the wild-type and indicated mutants in 200 ml complex medium initially containing 20% glucose
Download figure to PowerPoint