Mitochondrial Superoxide Dismutase and Yap1p Act as a Signaling Module Contributing to Ethanol Tolerance of the Yeast Saccharomyces cerevisiae

ABSTRACT

There are two superoxide dismutases in the yeast Saccharomyces cerevisiae—cytoplasmic and mitochondrial enzymes. Inactivation of the cytoplasmic enzyme, Sod1p, renders the cells sensitive to a variety of stresses, while inactivation of the mitochondrial isoform, Sod2p, typically has a weaker effect. One exception is ethanol-induced stress. Here we studied the role of Sod2p in ethanol tolerance of yeast. First, we found that repression of SOD2 prevents ethanol-induced relocalization of yeast hydrogen peroxide-sensing transcription factor Yap1p, one of the key stress resistance proteins. In agreement with this, the levels of Trx2p and Gsh1p, proteins encoded by Yap1 target genes, were decreased in the absence of Sod2p. Analysis of the ethanol sensitivities of the cells lacking Sod2p, Yap1p, or both indicated that the two proteins act in the same pathway. Moreover, preconditioning with hydrogen peroxide restored the ethanol resistance of yeast cells with repressed SOD2. Interestingly, we found that mitochondrion-to-nucleus signaling by Rtg proteins antagonizes Yap1p activation. Together, our data suggest that hydrogen peroxide produced by Sod2p activates Yap1p and thus plays a signaling role in ethanol tolerance.

IMPORTANCE Baker's yeast harbors multiple systems that ensure tolerance to high concentrations of ethanol. Still, the role of mitochondria under severe ethanol stress in yeast is not completely clear. Our study revealed a signaling function of mitochondria which contributes significantly to the ethanol tolerance of yeast cells. We found that mitochondrial superoxide dismutase Sod2p and cytoplasmic hydrogen peroxide sensor Yap1p act together as a module of the mitochondrion-to-nucleus signaling pathway. We also report cross talk between this pathway and the conventional retrograde signaling cascade activated by dysfunctional mitochondria.

INTRODUCTION

Superoxide dismutases (SODs) catalyze the conversion of superoxides into hydrogen peroxide and molecular oxygen. There are two common types of the enzyme in eukaryotic cells: Cu/Zn superoxide dismutase, which is localized in the cytoplasm, and Mn superoxide dismutase, which is localized in the mitochondrial matrix (for reviews, see references 1 and 2). In the yeast Saccharomyces cerevisiae, superoxide dismutases are encoded by two genes, cytoplasmic SOD by SOD1 and mitochondrial SOD by SOD2 (3, 4). SODs play a crucial role in the protection of the cells from various stresses (5, 6). Generally, the contribution of cytoplasmic superoxide dismutase to the total cellular antioxidant capacity is much higher than the contribution of the mitochondrial isoform (7, 8). Consistent with this, SOD1 knockdown decreases the resistance of yeast cells to many different types of stresses more severely than the knockdown of SOD2. The examples of such stresses are freeze-thaw stress (9), UV irradiation (10), Zn deficiency (11), and exposure to various toxic compounds (12–14). Moreover, the depletion of Sod1p but not of Sod2p facilitates protein carbonylation in post-diauxic phase of yeast growth (15). Carbonyl groups are usually considered a hallmark of nonenzymatic protein oxidation by reactive oxygen species (ROS) (for reviews, see references 16 and 17). However, in the case of ethanol-mediated stress, the situation is the opposite. Costa et al. (7) showed that the deletion of SOD2 strongly decreases yeast resistance to 20% ethanol, whereas the deletion of SOD1 has no significant effect.

What is the reason for high sensitivity of Δsod2 cells to ethanol? One possibility is that under ethanol stress, mitochondrial ROS are the primary source of damage. If so, compromising mitochondrial antioxidant defenses may have a strong effect on survival. Indeed, it has been shown that in wild-type yeast cells, ethanol induces mitochondrial fragmentation and promotes ROS-mediated oxidation of the fluorescent redox sensors H₂DCF-DA (2′,7′-dichlorodihydro-fluorescein diacetate) and DHR123 (dihydrorhodamine 123) (18–20). Possibly, the primary targets of ethanol-induced damage are within mitochondria or functionally rely on them. However, it should be taken into account that S. cerevisiae yeast cells can survive with nonfunctional mitochondria, in particular, without mitochondrial DNA (mtDNA) (for a review, see reference 21). Such cells, [rho⁰] cells, are much less resistant to ethanol than the wild-type [rho⁺] ones (22, 23). This suggests another, although nonexclusive, mechanism of Sod2p-mediated ethanol tolerance. Possibly, a conversion of superoxide anion, which cannot cross phospholipid membranes (24), to hydrogen peroxide, which probably can, is a step in the mitochondrion-to-nucleus signaling pathway that is required for activation of the stress response. Indeed, it has been shown earlier that treatment of yeast cells with ethanol induces nuclear accumulation of the hydrogen peroxide-sensing factor Yap1p (25, 26). Yap1p contains two signals, nuclear localization signal (NLS) and nuclear export signal (NES). In the absence of oxidative stress, Yap1p is rapidly exported from the nucleus by exportin Crm1p. Under oxidative stress (e.g., H₂O₂ treatment), thiol peroxidase Gpx3p (Hyr1p) mediates the formation of disulfide bonds in Yap1p molecules. The resulting conformational changes mask nuclear export signal, thus causing accumulation of Yap1p in the nucleus (27–29).

In this work, we provide experimental support for this hypothesis. We found a genetic interaction between YAP1 and SOD2 under ethanol stress. Moreover, we have shown that Sod2p is essential for Yap1p activation. Our results indicate that Sod2p has a signaling function, which is important for ethanol tolerance in yeasts.

RESULTS

To confirm the sensitivity of the cells lacking mitochondrial superoxide dismutase to ethanol stress, we reproduced the data of Costa and coworkers (7) using P_GAL-SOD2 and P_GAL-SOD1 strains. Under repression conditions, these strains displayed decreased resistance to the prooxidants menadione and paraquat (Fig. 1A). As expected, we found that repression of the SOD2 gene decreases the survival of yeast cells treated with 12 to 18% ethanol, whereas the effects of SOD1 repression were much less pronounced (Fig. 1B). At the same time, the sensitivities to heat shock of the strains with repressed SOD1 or SOD2 were very similar (Fig. 1C). Next, we have shown that treatment of yeast cells with 12 to 16% ethanol induced cytoplasm-to-nucleus relocalization of Yap1p transcription factor fused to green fluorescent protein (GFP) (Fig. 2A and C). Such relocalization was inhibited in the cells with repressed SOD2 (Fig. 2B and C). One could argue that the absence of functional Mn²⁺ superoxide dismutase interfered with the synthesis or maturation of Yap1-GFP. Our observation that hydrogen peroxide activates Yap1-GFP in cells with repressed SOD2 to the same extent as it does in the control strain (Fig. 2A and B) excludes this possibility. Together, these data suggest that Yap1p and Sod2p are in the same stress-activated pathway required for ethanol tolerance.

FIG 1.

High concentrations of ethanol are more toxic for cells with repressed SOD2 than for cells with repressed SOD1. Survival rates of the wild-type yeast cells (WT control) and the cells with repressed SOD1 or SOD2 after exposure to different stresses are presented as percent CFU. (A) Exposure to the prooxidant menadione (25 μM) or paraquat (15 mM) for 2 h (n = 5). The values above 100% reflect cell growth during incubation time, and the blue arrows indicate the absence of viable cells of the strain with repressed SOD1. *, P < 0.05 compared with results for the WT according to the Wilcoxon rank-sum unpaired test. (B) Ethanol stress (12% to 18% ethanol [vol/vol]) for 1 h (n ≥ 5). (C) Heat shock, 45°C (n = 3).

FIG 2.

Repression of SOD2 prevents cytoplasm-to-nucleus relocalization of Yap1p in ethanol-treated cells. Localization of Yap1-GFP in W303 wild-type yeast cells (WT control) (A) and cells with repressed SOD2 (B) exposed to 1 mM H₂O₂ (positive control) or 12 to 16% (vol/vol) ethanol stress for 15 min. Scale bar, 5 μm. (C) Quantification of results. Data are presented as averages with SEM. In total, 85 to 124 cells from 3 to 5 separate experiments performed on separate days were analyzed. *, P < 0.05 compared with results for the WT according to the Wilcoxon rank-sum unpaired test. DIC, differential interference contrast.

Possibly, conversion of superoxide into hydrogen peroxide by Sod2p increases the rate of hydrogen peroxide release from mitochondria and in this way facilitates Yap1p activation. To investigate further the interrelationship between Sod2p and Yap1p, we studied the accumulation of Yap1p targets under ethanol stress by flow cytometry. First, we compared the levels of two proteins, Trx2p and Gsh1p, encoded by Yap1p-regulated genes. We found that the level of Trx2-GFP is significantly increased in the cells treated with hydrogen peroxide (Fig. 3A and B). At the same time, the accumulation of Trx2-GFP in the cells treated with a 16% (vol/vol) ethanol pulse was not significantly different from that in the control (Fig. 3B; see Materials and Methods for experimental details). However, the repression of the SOD2 gene significantly decreased the level of Trx2-GFP under all tested conditions. A similar result was obtained with another Yap1p target, GSH1 (Fig. 3C). The levels of Trx2-GFP and Gsh1-GFP were decreased in yeast cells with repressed SOD2 treated with hydrogen peroxide (Fig. 3B and C). Possibly, the hydrogen peroxide-induced increase of Trx2-GFP and Gsh1-GFP in cells with repressed SOD2 can be partially suppressed at the levels of transcription and/or translation because of the higher sensitivity to oxidative stress. Indeed, the background levels of at least some of the antioxidant enzymes are decreased in cells with repressed SOD2 (Fig. 3B and C). Together, these data suggest that modest Yap1p activation mediated by Sod2p takes place under normal conditions and increases background levels of Yap1p targets.

FIG 3.

Repression of SOD2 decreases the levels of Yap1p target genes and ethanol tolerance. (A) Representative FACS experiment showing ethanol- and hydrogen peroxide-induced activation of Trx2-GFP synthesis; (B) quantification of results for the wild-type cells (WT control) and cells with repressed SOD2 gene (n ≥ 4); (C) quantification of Gsh1p-GFP synthesis activation by flow cytometry (n ≥ 3); (D) repression of SOD2 decreases ethanol resistance in wild-type cells but not in cells with the YAP1 gene deleted (n = 5). In panels A to C, ethanol was used at a concentration equal to 16% (vol/vol); the final concentration of hydrogen peroxide was 1 mM. Data are presented as averages with SEM. *, P < 0.05; **, P < 0.01; ***, P < 0.001, according to the Wilcoxon rank-sum unpaired test. In panels B and C, GFP levels in cells with repressed SOD2 were compared with GFP levels in WT cells. In panel D, the survival of the mutant cells was compared with the survival of untreated WT cells. a.u., arbitrary units.

SOD2-dependent relocalization of Yap1-GFP suggests that Yap1p and Sod2p are in the same pathway required for ethanol tolerance. If so, the deletion of YAP1 is expected to decrease ethanol tolerance in the control strain but not in the sod2-deficient one. Thus, we compared the survival rates of yap1 and sod2 mutant strains. We found that the Δyap1 strain showed decreased survival in the presence of 16% (vol/vol) ethanol, and the same decrease was observed in cells with repressed SOD2 (Fig. 3D). Next, we generated a double mutant, the P_GAL-SOD2 Δyap1 strain. We found that under conditions of SOD2 repression, this mutant displayed the same survival rate as the parental strains, the Δyap1 and P_GAL-SOD2 strains. Therefore, our data are in agreement with the idea that Yap1p and Sod2p act in the same pathway.

Our results raise a question of the role of mitochondrial respiration in ethanol tolerance. The mitochondrial respiratory chain is considered to be a primary source of superoxide in the matrix (30). Thus, inhibition of respiration may increase the tolerance by reducing the level of ROS-dependent damage upon ethanol treatment. At the same time, impaired respiration affects mitochondrial signaling, which can diminish this effect.

Our findings suggest that the deleterious effect of SOD2 repression is absent in respiration-deficient yeast cells. To test this, we produced a yeast strain lacking mitochondrial DNA ([rho⁰] strain). As expected, ethanol did not activate Yap1p relocalization in this strain (Fig. 4A and B). At the same time, the [rho⁰] strain with repressed mitochondrial superoxide dismutase was more resistant to ethanol than the parental one (Fig. 4C), while the [rho⁰] P_GAL-SOD1 strain showed a statistically insignificant decrease in ethanol tolerance compared to the control [rho⁺] P_GAL-SOD1 strain (Fig. 4C). In agreement with previous works (22, 23), the depletion of mitochondrial DNA in the wild-type control strains decreased the number of CFU under ethanol stress. Accumulation of reduced respiratory chain cofactors can be a source of superoxide, while the dissipation of transmembrane potential by uncouplers or inhibition of complex III with myxothiazol can decrease superoxide production by yeast mitochondria (31–33). It is expected that the effects of these compounds on ethanol tolerance are similar to those of the [rho⁰] mutation. Surprisingly, treatment of yeast cells with mitochondrial uncouplers and respiratory chain inhibitors failed to provide any protection from ethanol (Fig. 4D and E). This suggests that yeast cells without mtDNA undergo regulatory reorganization, which activates an additional pathway(s) acting to prevent ethanol toxicity.

FIG 4.

Ethanol tolerance by yeast is affected by the functional state of the mitochondrial respiratory chain. (A) Ethanol does not induce Yap1p relocalization in cells lacking mitochondrial DNA ([rho⁰]); (B) quantification of the results. Data are presented as averages with SEM. In total, 87 to 93 cells from 3 experiments done on separate days were analyzed. Ethanol was used at a concentration equal to 16% (vol/vol); the final concentration of hydrogen peroxide was 1 mM. (C) Effect of mtDNA depletion on ethanol tolerance in the wild-type strain (WT control) and the strain with repressed SOD1 or SOD2 (n = 4). Ethanol was used at a concentration equal to 16% (vol/vol). *, P < 0.05 according to the Wilcoxon rank-sum unpaired test. (D and E) Mitochondrial uncouplers {FCCP [carbonyl cyanide-4-(trifluoromethoxy)phenylhydrazone], 5 μM; pentachlorophenol [PCP] 10 μM) (D) and respiration inhibitors (myxothiazol, 7 μM; antimycin A, 50 μg/ml) (E) do not prevent ethanol toxicity in yeast (n = 4).

What could this pathway(s) be? Yeast cells harbor a signaling cascade, called the retrograde signaling (Rtg) pathway, which transmits the signals from dysfunctional mitochondria to the nuclei. The core components of the Rtg pathway are Rtg2p (pathway initiator), the negative regulator Mks1p, and transcription factors Rtg1p and Rtg3p (for a review, see reference 34). We decided to test the role of this pathway in Sod2p-dependent ethanol tolerance. We found that the deletion of RTG2 decreases the resistance of yeasts to 16% (vol/vol) ethanol in the control strain, while under the conditions of SOD2 repression, the deletion did not show any effect on survival (Fig. 5A). Thus, the combined effect of the deletion of RTG2 and the replacement of the SOD2 promoter by the P_GAL promoter, which is repressed under our experimental conditions, was lower than the sum of the two effects. Such type of genetic interaction may be due to the involvement of Rtg2p and Sod2p in the subsequent steps of the same pathway. If true, inactivation of either gene will decrease the levels of the targets of both pathways. Alternatively, this kind of interaction between two genes might take place if one of the encoded proteins opposes another and at the same time provides a comparable phenotypic manifestation. To discriminate between these two possibilities, we studied the activation of the Yap1p target TRX2 under the conditions of Rtg pathway inactivation and, vice versa, the activation of the Rtg target IDH1 in cells with repressed SOD2 repressed. We found that the deletion of the RTG2 or RTG3 gene induced an increase in Trx2-GFP levels (Fig. 5B), whereas SOD2 repression did not change the levels of Idh1-GFP (Fig. 5C). To prove that our Idh1-GFP sensor is functional, we have shown that the deletion of RTG3 decreases, and the deletion of MKS1 increases, the accumulation of Idh1-GFP in yeast cells (Fig. 5C). Moreover, we monitored Yap1-GFP localization in cells with an inactivated (Δrtg2) or activated (Δmks1) retrograde pathway. It was found that the deletion of RTG2 results in a permanent cytosol-to-nucleus relocalization of Yap1p in both the control strain and the strain with repressed sod2, whereas the deletion of MKS1 has no such effect (Fig. 5D and E). At the same time, activation of the retrograde signaling pathway by methionine-sulfoximine (2 mM) restored ethanol tolerance to the level of the parental strain (Fig. 5F) and the deletion of MKS1 gene improved the resistance of cells with repressed SOD2 to ethanol treatment (Fig. 5F). Also, the double mutation Δyap1 Δrtg2 showed an additive effect on survival (Fig. 5G). Together, these data suggest that activation of the Rtg pathway interferes with Yap1p activation and at the same time activates an alternative system(s) of ethanol tolerance.

FIG 5.

Genetic interactions between RTG2 and SOD2. (A) Repression of SOD2 decreases ethanol (16% vol/vol) resistance in wild-type cells (WT control) but not in cells with deletion of the RTG2 gene. (B) Quantification of the flow cytometry analysis for activation of Trx2-GFP synthesis in strains with deleted components of the retrograde signaling pathway (n ≥ 4). (C) Quantification of the flow cytometry analysis for activation of the Rtg pathway target Idh1-GFP in strains with deleted components of the retrograde signaling pathway or repressed SOD2 (n ≥ 6). The arrowhead indicates the fluorescence level of the Idh1-GFP-expressing strain in a Δrtg3 background, which appeared to be below the level of autofluorescence in the control cells without GFP. (D) Localization of Yap1-GFP in wild-type yeast cells (WT control) and in cells with deleted components of the RTG pathway treated with ethanol. Bar, 5 μm. (E) Quantification of the results shown in panel D. In total, 71 to 123 cells were analyzed for each condition from at least three separate experiments. (F) Activation of the retrograde signaling pathway by methionine sulfoximine (MSX, 2 mM) or by deletion of the MKS1 gene restores ethanol tolerance in cells with repressed SOD2 gene (n ≥ 5). (G) Deletion of YAP1 decreases ethanol tolerance in cells with deletion of the RTG2 gene (n = 4). In panels E and G, ethanol was used at a concentration equal to 16% (vol/vol). *, P < 0.05; **, P < 0.01, compared with results for the WT according to the Wilcoxon rank-sum unpaired test.

To obtain more direct evidence on the signaling role of mitochondrial hydrogen peroxide in ethanol tolerance, we isolated mitochondria from control cells and cells with repressed SOD2. The respiratory control index for mitochondrial preparations was 2.4 ± 0.3. In agreement with the in vivo data, we found a small increase in the hydrogen peroxide generation rate after treatment with 2.5% (vol/vol) ethanol. Mitochondria from the cells with repressed SOD2 showed a marginal decrease in hydrogen peroxide production (Fig. 6A).

FIG 6.

Regulation of Yap1p during ethanol stress and hypothetical scheme of yeast ethanol tolerance by mitochondrial retrograde regulation. (A) Generation of hydrogen peroxide measured with HRP and the Amplex red system (see Materials and Methods) by mitochondria isolated from the wild-type yeast strain (WT control) or yeast with a repressed SOD2 gene. (B) Preincubation with hydrogen peroxide (0.05 mM) restores ethanol tolerance in cells with repressed SOD2 (n = 5). Ethanol was used at a concentration equal to 16% (vol/vol). *, P < 0.05 compared with results for the WT according to the Wilcoxon rank-sum unpaired test. (C) Hypothetical scheme of mitochondrial signaling promoting ethanol tolerance in yeast. Rtg2p-dependent inactivation of Yap1p can be due either to a decrease in cytoplasmic hydrogen peroxide levels or to a direct inhibition of Yap1p.

Most importantly, we found that activation of Yap1p by externally added hydrogen peroxide can prevent the deleterious effect of SOD2 repression. P_GAL-SOD2 cells pretreated with 0.05 mM hydrogen peroxide showed the same level of ethanol tolerance as untreated control cells (Fig. 6B). Together, these experiments provide additional support for a signaling role of Sod2-produced hydrogen peroxide.

DISCUSSION

High ethanol tolerance is a remarkable feature of baker's yeast. Genetic screenings have revealed that ethanol tolerance depends on multiple cellular systems (35, 36; for a review, see reference 37). The genes responsible for ethanol tolerance are also involved in the regulation of cell wall composition (38), membrane composition (39, 40), and also vacuolar ATPase (41). At the same time, the role of mitochondria in ethanol tolerance is not clear. Depletion of mitochondrial DNA in yeasts decreases the resistance to ethanol (22). However, it has also been reported that industrial yeasts incubated in the presence of 15% ethanol accumulate respiration-deficient mutants (42). That might be due either to an indirect induction of mtDNA damage by ethanol or to the selection of specific mutants. Our data show that the loss of mitochondrial function can have a dual effect on yeast ethanol tolerance depending on the nuclear genome background (Fig. 4C). Possibly, it could be explained by the dual role of mitochondria in yeast exposed to high ethanol concentrations. On the one hand, mitochondria can be a source of reactive oxygen species under stressful conditions (18, 20). On the other hand, mitochondria are required for activation of stress response genes, including Yap1p targets. We showed that mitochondrial superoxide dismutase is required for relocalization of Yap1p in the presence of high ethanol concentrations. This observation is in agreement with the previously published data by Ma and Liu (43) showing that Yap1p targets are among the key factors required for ethanol tolerance. We reason that Sod2p-Yap1p might be regarded as the retrograde signaling cascade acting in coordination with the Rtg2p/Rtg3p-dependent retrograde signaling route (Fig. 6C). If so, it appears to be an example of the optimization of the stress response mechanism— the major source of damage under a specific stress plays a signaling role to confer tolerance to this stress. From a practical point of view, this information might be useful for improvements in ethanol tolerance in industrial strains. Possibly, activated expression of Rtg1p/Rtg3p and Yap1p targets combined with evasion of Yap1p inhibition by an active Rtg pathway can increase the tolerance.

In agreement with previous observations (26, 37), our data show that Yap1p is a component of the ethanol tolerance mechanism. At the same time, the role of mitochondria in Yap1p activation is not clear. Mitochondria harbor many enzymes with redox cofactors. These cofactors can nonenzymatically reduce molecular oxygen to superoxide anion (30). Being a charged molecule, superoxide cannot cross the inner mitochondrial membrane and thus either remains in the matrix or is oxidized by the respiratory chain. Mitochondrial superoxide dismutase converts superoxide into molecular oxygen and hydrogen peroxide. The diffusion of hydrogen peroxide across the membranes is also limited (44). For instance, aquaporins facilitate the diffusion and in this way significantly increase the sensitivity of yeasts to exogenous hydrogen peroxide (45, 46). However, the diffusion rate of hydrogen peroxide across the membranes may vary depending on lipid content. It has been shown that deletions of genes required for ergosterol biosynthesis increase the rate of hydrogen peroxide consumption by yeast cells (47). As the inner mitochondrial membrane contains a low concentration of ergosterol (48), it is likely that diffusion of hydrogen peroxide across the inner membrane is less constrained than in the case of the ergosterol-enriched plasma membrane. Accordingly, in mammalian cancer cells, the overexpressed mitochondrial superoxide dismutase can provide increased hydrogen peroxide production (49). In any case, isolated respiring mitochondria generate measurable levels of hydrogen peroxide in the medium, and the absence of superoxide dismutase slightly decreases the rate of hydrogen peroxide production (Fig. 6A). Under fermenting conditions (used in our study), the concentrations of tricarboxylic acid (TCA) cycle and respiratory enzymes are decreased (for a review, see reference 50). However, the net output of mitochondrial reactive oxygen species production can be even higher in mitochondria isolated from glucose-grown yeasts (51). This effect can be explained by the compensation of increased mitochondrial reactive oxygen species production in nonfermenting cells by the upregulated antioxidant mechanisms. It has been shown that yeast antioxidant enzymes are derepressed upon switching from glucose to glycerol (52). Nevertheless, under normal conditions, mitochondria are considered to play a role in ROS sink (53). However, the severe stresses (e.g., high concentrations of ethanol) might change the net direction of ROS flux due to inactivation of antioxidant enzymes and inhibition of the respiratory chain. In this case, mitochondrially produced hydrogen peroxide may appear in the cytoplasm and activate Yap1p. Apparently, hydrogen peroxide from other sources (e.g., dismutated from superoxide by Sod1p) also can activate the Yap1-mediated compensatory response. However, under normal conditions, this process is well controlled by cytoplasmic antioxidant systems provided that Yap1p is localized mainly in the cytoplasm (Fig. 2). Therefore, it appears that ethanol induces hydrogen peroxide-mediated signaling which is moderated by the retrograde signaling cascade activated by mitochondrial dysfunction. Thus, this indicates that the transcriptional response to ethanol-induced stress is complex and includes activation of (at least) two signaling pathways. Possibly, this complexity is needed for the optimal mobilization of cellular defense resources. Consistent with this, it has been shown earlier that in animals, mitochondrial and nonmitochondrial ROS have different effects on physiology (54).

To summarize, we have shown that mitochondrial superoxide dismutase and Yap1p act together as a module of the mitochondrion-to-nucleus signaling cascade that plays a crucial role in yeast tolerance to high concentrations of ethanol.

MATERIALS AND METHODS

Strains.

In this study, we used yeast strains of the W303 genetic background that are listed in Table 1. W303 (wild type [WT]) was used as a control. We found that the strains with deleted superoxide dismutase genes are genetically unstable and eventually lose their ability to grow on nonfermentable carbon sources. This is in agreement with an article showing that the knockout of SOD1 or SOD2 can induce DNA damage (55). For this reason, we generated strains in which genomic copies of the superoxide dismutase genes were set under the control of the conditionally regulated P_GAL promoter. To do this, we used PCR cassettes described in reference 56. To produce PCR cassettes, we used the pFA6a-His3MX6-PGAL1-VN plasmid as a template DNA and the primers listed in Table 2. These P_GAL strains were cultured on galactose-containing rich medium and transferred to the repression condition (glucose-containing rich medium) only prior to the experiments. Double mutants (Table 1) were produced by crossing the corresponding single mutants and subsequent tetrad dissection; [rho⁰] strains were produced from the corresponding [rho⁺] strains by overnight incubation in yeast extract-peptone-dextrose (YPD) medium supplemented with 20 μM ethidium bromide. The cells were then transferred to solid YPDGly medium (2% Bacto agar, 2% Bacto peptone, 1% yeast extract, 0.1% glucose, and 2% glycerol) and allowed to grow for 2 days. Small colonies were selected and transferred to YPD plates. The resulting strains were tested for their ability to grow on glycerol as a carbon source and were also stained with 4′,6-diamidino-2-phenylindole (DAPI) to confirm the absence of mitochondrial DNA. All strains are listed in Table 1 (56–58). Typically, cells were grown overnight at 30°C with rotary shaking (250 rpm) to a density of 2 × 10⁶ cells ml⁻¹ in liquid YPD medium prepared as described by Sherman (59).

TABLE 1.

Strains used in the study

Strain	Genotype	Parental strain(s) and/or reference
W303	MATa ade2-101 his3-11 trp1-1 ura3-52 can1-100 leu2-3	Laboratory of A. Hyman
W303 [rho0]a	MATa ade2-101 his3-11 trp1-1 ura3-52 can1-100 leu2-3 [rho0]	W303
PGAL-SOD1 mutantb	MATa ade2-101 his3-11 trp1-1 ura3-52 can1-100 leu2-3 PGAL-SOD1::HIS3	W303
PGAL-SOD1 [rho0] mutanta	MATa ade2-101 his3-11 trp1-1 ura3-52 can1-100 leu2-3 PGAL-SOD1::HIS3 [rho0]	PGAL-SOD1 mutant
PGAL-SOD2 mutantb	MATa ade2-101 his3-11 trp1-1 ura3-52 can1-100 leu2-3 PGAL-SOD2::HIS3	W303
PGAL-SOD2 [rho0] mutanta	MATa ade2-101 his3-11 trp1-1 ura3-52 can1-100 leu2-3 PGAL-SOD2::HIS3 [rho0]	PGAL-SOD2 mutant
PGAL-SOD2 Δrtg2 mutantc	MATa ade2-101 his3-11 trp1-1 ura3-52 can1-100 leu2-3 PGAL-SOD2::HIS3 rtg2::KANMX4	PGAL-SOD2 mutant; Δrtg2
PGAL-SOD2 Δmks1 mutantb	MATa ade2-101 his3-11 trp1-1 ura3-52 can1-100 leu2-3 PGAL-SOD2::HIS3 mks1::KANMX4	Δmks1 mutant
PGAL-SOD2 Δyap1 mutantc	MATa ade2-101 his3-11 trp1-1 ura3-52 can1-100 leu2-3 PGAL-SOD2::HIS3 yap1::KANMX4	PGAL-SOD2 mutant; Δyap1
Yap1-GFP mutantb	MATa ade2-1 trp1-1 can1-100 leu2-3,112 his3-11,15 ura3 cdc13-1 YAP1-GFP::HIS3	W303
Yap1-GFP [rho0] mutanta	MATa ade2-1 trp1-1 can1-100 leu2-3,112 his3-11,15 ura3 cdc13-1 YAP1-GFP::HIS3 [rho0]	Yap1-GFP mutant
Yap1-GFP PGAL-SOD2 mutantc	MATa ade2-1 trp1-1 can1-100 leu2-3,112 his3-11,15 ura3 cdc13-1 YAP1-GFP::HIS3 PGAL-SOD2::HIS3	Yap1-GFP mutant; PGAL-SOD2
Yap1-GFP Δrtg2 mutantc	MATa ade2-1 trp1-1 can1-100 leu2-3,112 his3-11,15 ura3 cdc13-1 YAP1-GFP::HIS3 rtg2::KANMX4	Yap1-GFP mutant; Δrtg2
Yap1-GFP PGAL-SOD2 Δrtg2 mutantc	MATa ade2-1 trp1-1 can1-100 leu2-3,112 his3-11,15 ura3 cdc13-1YAP1-GFP::HIS3 PGAL-SOD2::HIS3 rtg2::KANMX4	Yap1-GFP mutant; Δrtg2 PGAL-SOD2
Yap1-GFP Δmks1 mutantc	MATa ade2-1 trp1-1 can1-100 leu2-3,112 his3-11,15 ura3 cdc13-1 YAP1-GFP::HIS3 mks1::KANMX4	Yap1-GFP mutant; Δmks1 (Zyrina et al. [57])
Δyap1 mutantd	MATa ade2-101 his3-11 trp1-1 ura3-52 can1-100 leu2-3 yap1::KANMX4	W303
Δrtg2 mutant	MATa ade2-101 his3-11 trp1-1 ura3-52 can1-100 leu2-3 rtg2::KANMX4	Zyrina et al. (57)
Δyap1 Δrtg2 mutantc	MATa ade2-101 his3-11 trp1-1 ura3-52 can1-100 leu2-3 yap1::KANMX4 rtg2::KANMX4	Δyap1; Δrtg2
Trx2-GFP mutant	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 TRX2-GFP::HIS3	Huh et al. (58)
Trx2-GFP PGAL-SOD2 mutantc	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 TRX2-GFP::HIS3 PGAL-SOD2::HIS3	Trx2-GFP mutant; PGAL-SOD2
Trx2-GFP Δrtg2 mutantc	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 TRX2-GFP::HIS3 rtg2::KANMX4	Trx2-GFP mutant; Δrtg2
Trx2-GFP Δrtg3 mutantc	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 TRX2-GFP::HIS3 rtg3::KANMX4	Trx2-GFP mutant; Δrtg3
Trx2-GFP Δmks1 mutantc	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 TRX2-GFP::HIS3 mks1::KANMX4	Trx2-GFP mutant; Δmks1
Gsh1-GFP mutant	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 GSH1-GFP::HIS3	Huh et al. (58)
Gsh1-GFP PGAL-SOD2 mutantc	MATa his3Δ1 leu2Δ0 met15Δ0 ura3Δ0 GSH1-GFP::HIS3 PGAL-SOD2::HIS3	Gsh1-GFP mutant; PGAL-SOD2
Idh1-GFP mutant	MATa ade2-1 trp1-1 can1-100 leu2-3,112 his3-11,15 ura3 cdc13-1 IDH1-GFP::TRP1	W303 (courtesy of Iuliia Karavaeva)
Idh1-GFP PGAL-SOD2 mutantb	MATa ade2-1 trp1-1 can1-100 leu2-3,112 his3-11,15 ura3 cdc13-1 IDH1-GFP::TRP1 PGAL-SOD2::HIS3	Idh1-GFP mutant; PGAL-SOD2
Idh1-GFP Δrtg3 mutantc	MATa ade2-1 trp1-1 can1-100 leu2-3,112 his3-11,15 ura3 cdc13-1 IDH1-GFP::TRP1 rtg3::KANMX4	Idh1-GFP mutant; Δrtg3
Idh1-GFP Δmks1 mutantc	MATa ade2-1 trp1-1 can1-100 leu2-3,112 his3-11,15 ura3 cdc13-1 IDH1-GFP::TRP1 mks1::KANMX4	Idh1-GFP mutant; Δmks1

^aStrain produced by ethidium bromide-induced loss of mitochondrial DNA.

^bStrain produced by insertion of the P_GAL promoter or GFP tag using a PCR cassette as described by Longtine et al. (56).

^cStrain produced by crossing and tetrad dissection.

^dStrain produced by disruption with a PCR cassette.

TABLE 2.

Primers used in the study

Sequence	Gene, direction
TCGCCATCGCAGATATATATATAAGAAGATGGTTTTGGGCAAATGTTTAGCTGTAGAATTCGAGCTCGTTTAAAC	SOD1, forward
ACTTGACAACACCAGAGACACCGGCATCACCCTTTAACACTGCGACTGCTTGAACCATTTTGAGATCCGGGTTTT	SOD1, reverse
CCCTGATGAAGAAGCTATAACTTGTATATAAGAGCGTTGCATCCCCAAAATATACGAATTCGAGCTCGTTTAAAC	SOD2, forward
AGAGCAATGACAAACCACCCTTCTTGGTTAAATTAGCAGCTGCTGTTTTCGCGAACATTTTGAGATCCGGGTTTT	SOD2, reverse

Survival experiments.

Survival of yeast cells was measured by the CFU method. Exponentially growing cells from an overnight YPD culture were taken and exposed to different stresses (12%, 14%, 16%, or 18% [vol/vol] ethanol stress, 1 h) or heat shock (45°C, 1 h). In the case of preincubation with a low H₂O₂ concentration, yeast cells from an overnight YPD culture prior to ethanol treatment (16%, 1 h) were incubated with 0.05 mM H₂O₂ for 30 min.

To measure the numbers of CFU, the cell suspensions were transferred to solid YPD medium in a set of dilutions and incubated for 48 h at 30°C. The number of colonies was counted; 100% refers to the number of CFU in the yeast suspension at the beginning of the experiment.

Microscopy.

Fluorescence microscopy was used to determine GFP fusion protein localization, to visualize nuclei in DAPI-stained live cells, and to confirm the absence of mitochondrial DNA in [rho⁰] cells stained with DAPI. Yeast cells were visualized with an Olympus BX51 microscope with a U-MNIBA3 filter for GFP (excitation wavelength, 470 to 495 nm; dichroic mirror, 505 nm; emission wavelength, 510 to 550 nm) and a U-MNU2 filter for DAPI (excitation wavelength, 360 to 370 nm; dichroic mirror, 400 nm; emission wavelength, >420 nm). Photographs were taken with a DP30BW charge-coupled-device (CCD) camera. In the case of [rho⁰] cells, they were fixed in 70% ethanol, ethanol was replaced with water, and the cells were stained with DAPI (final concentration, 2 μg/ml) for 30 min and then washed once in water. In the case of GFP fusion protein localization, cells were exposed to 1 mM H₂O₂ or 12 to 18% (vol/vol) ethanol for 15 min, washed once in PBS, and then stained with DAPI (final concentration, 2 μg/ml) for 15 min.

Flow cytometry.

Cellular fluorescence from GFP was determined quantitatively using a CytoFLEX flow cytometer (Beckman Coulter); all flow cytometry data were analyzed with CytExpert software (Beckman Coulter). For stress induction, exponentially growing cells from overnight YPD-grown cultures were incubated with 1 mM H₂O₂ or 16% (vol/vol) ethanol for 15 min and then washed in YPD and incubated in YPD for 2 h. Control cells were incubated under the same conditions but without H₂O₂ or ethanol. Before fluorescence-activated cell sorting (FACS) analysis, the cells were resuspended in phosphate-buffered saline (PBS). At least 10,000 cells were analyzed per sample.

Isolation of mitochondria.

Mitochondria from the control and P_GAL-SOD2 cells were prepared as described previously (60). The protein concentration in mitochondria was determined with the Pierce BCA protein assay kit.

Mitochondrial respiration.

The respiration of isolated mitochondria was measured using a standard polarography technique with a Clark-type oxygen electrode (Strathkelvin Instruments 782, United Kingdom) at 25°C using DATLAB software. The incubation medium for mitochondria contained 0.6 M mannitol, 10 mM Tris-HCl, 2 mM potassium phosphate (pH 7.4), and 15 mM pyruvate-malate (4:1).

H₂O₂ generation by isolated mitochondria.

H₂O₂ generation was measured fluorimetrically by means of horseradish peroxidase (HRP) and Amplex red using a FluoroMax-3 fluorimeter system with the excitation wavelength set to 530 nm and the emission wavelength set to 590 nm. The incubation mixture contained 0.6 M mannitol, 10 mM Tris-HCl, 2 mM potassium phosphate (pH 7.4), 15 mM pyruvate-malate (4:1), 2 μM Amplex red, and 0.005 mg/ml HRP.

Statistical analysis.

All data are presented as averages with standard errors of the mean (SEM). The Wilcoxon rank-sum unpaired test was used to compare data sets from different strains or conditions with the R software package.