Yap1 and Skn7 Control Two Specialized Oxidative Stress Response Regulons in Yeast*

Yap1 and Skn7 are two yeast transcriptional regulators that co-operate to activate thioredoxin (TRX2) and thioredoxin reductase (TRR1) in response to redox stress signals. Although they are both important for resistance to H2O2, only Yap1 is important for cadmium resistance, whereas Skn7 has a negative effect upon this response. The respective roles of Yap1 and Skn7 in the induction of defense genes by H2O2 were analyzed by two-dimensional gel electrophoresis. Yap1 controls a large oxidative stress response regulon of at least 32 proteins. Fifteen of these proteins also require the presence of Skn7 for their induction by H2O2. Although about half of the Yap1 target genes do not contain a consensus Yap1 recognition motif, the control of one such gene, TSA1, involves the binding of Yap1 and Skn7 to its promoter in vitro. The co-operative control of the oxidative stress response by Yap1 and Skn7 delineates two gene subsets. Remarkably, these two gene subsets separate antioxidant scavenging enzymes from the metabolic pathways regenerating the main cellular reducing power, glutathione and NADPH. Such a specialization may explain, at least in part, the dissociated function of Yap1 and Skn7 in H2O2 and cadmium resistance.

Yap1 and Skn7 are two yeast transcriptional regulators that co-operate to activate thioredoxin (TRX2) and thioredoxin reductase (TRR1) in response to redox stress signals. Although they are both important for resistance to H 2 O 2 , only Yap1 is important for cadmium resistance, whereas Skn7 has a negative effect upon this response. The respective roles of Yap1 and Skn7 in the induction of defense genes by H 2 O 2 were analyzed by two-dimensional gel electrophoresis. Yap1 controls a large oxidative stress response regulon of at least 32 proteins. Fifteen of these proteins also require the presence of Skn7 for their induction by H 2 O 2 . Although about half of the Yap1 target genes do not contain a consensus Yap1 recognition motif, the control of one such gene, TSA1, involves the binding of Yap1 and Skn7 to its promoter in vitro. The co-operative control of the oxidative stress response by Yap1 and Skn7 delineates two gene subsets. Remarkably, these two gene subsets separate antioxidant scavenging enzymes from the metabolic pathways regenerating the main cellular reducing power, glutathione and NADPH. Such a specialization may explain, at least in part, the dissociated function of Yap1 and Skn7 in H 2 O 2 and cadmium resistance.
Aerobic organisms have to maintain a reduced cellular redox environment in the face of the prooxidative conditions of aerobic life. The incomplete reduction of oxygen to water during respiration leads to the formation of redox-active oxygen intermediates such as the superoxide anion radical (O 2 . ), hydrogen peroxide (H 2 O 2 ), and the hydroxyl radical (for review see Refs. [1][2][3]. Redox-active oxygen intermediates are also produced during the ␤-oxidation of fatty acids by exposure to radiation, light, metals, and redox active drugs. Redox-active oxygen intermediates perturbate the cell redox status and when present in high levels can induce toxic damage to lipids, proteins, and DNA, eventually leading to cell death. Living organisms constantly sense and adapt to such redox perturbations. The exposure of the yeast Saccharomyces cerevisiae to low doses of H 2 O 2 or O 2 . -generating drugs switches on within minutes a resistance to toxic doses of these oxidants (4 -6). The adaptive response to H 2 O 2 involves a change in the expression of at least 167 proteins (7). Such a rapid and widespread genomic response suggests the existence of specific control pathways. In S. cerevisiae, the transcription factors Yap1 (8,9) and Skn7 (10,11) have been implicated in a cellular pathway that controls the oxidative stress response. Yap1 is a bZIP DNAbinding protein of the AP-1 family (12) that binds the sequence T(T/G)ACTAA termed the Yap1 response element (YRE) (9,13,14). 1 Skn7 contains a receiver domain found in the family of two-component signal transduction systems of prokaryotes and a domain similar to the DNA-binding domain of heat shock factor (Hsf1) (15,16). Skn7 is also capable of specific DNA binding, but its cognate DNA sequence has not been identified precisely (11). Strains inactivated for either one of these regulators are hypersensitive to killing by H 2 O 2 (9,10,11,17). This oxidative stress phenotype is related to the role of Yap1 and Skn7 in controlling the induction of several defense genes by H 2 O 2 . Yap1 controls the expression of GSH1 (␥-glutamylcysteine synthetase) (13), TRX2 (thioredoxin) (9), GLR1 (glutathione reductase) (18), and YCF1 (yeast cadmium factor, a glutathione S-conjugate pump) (19). More recently, Morgan showed that the induction of TRX2 and TRR1 (thioredoxin reductase) by H 2 O 2 requires a co-operation between Yap1 and Skn7 (11). YAP1 function can be activated by H 2 O 2 , diamide, and diethylmaleate (9,20), and this activation is attributed to a redox stress-imposed nuclear redistribution of the protein involving the nuclear export receptor Crm1 (Xpo1) (21,22).
Yap1 is also important in cadmium tolerance because deletion of its gene results in a cadmium-hypersensitive phenotype (19,20,23). This function is attributed to the control by Yap1 of GSH1 (13) and of YCF1 (19). In addition to its involvement in the oxidative stress response, Skn7 is implicated in the control of cell wall biosynthesis, cell cycle, and the osmotic stress response. Overexpression of SKN7 can suppress the cell wall assembly mutation kre9 (24) and the growth defect associated with a pkc1 null deletion (15). Overexpression of SKN7 also suppresses the lethality associated with loss of the G 1 transcription factors SBF and MBF (16). Skn7 is modulated by the Sln1-Ypd1 osmosensor and contributes to regulation of the HOG osmo-stress pathway (25). The involvement of Skn7 into such diverse pathways raises the question of the possible connection between these pathways.
We sought to further analyze the co-operative functions of Yap1 and Skn7 in the control of the oxidative and cadmium stress responses and found that these two regulators do not always act together in these stress responses. Although both Yap1 and Skn7 are important for resistance to H 2 O 2 , only Yap1 is important for cadmium resistance, whereas Skn7 is not only dispensable but appears to negatively affect this response. The role of Yap1 and Skn7 in the induction of defense genes by H 2 O 2 was further analyzed by two-dimensional gel electrophoresis. The data presented here identify, within a large Yap1 stress response regulon, a gene subset that also requires Skn7 for its control. This partition of the Yap1 regulon correlates with two distinct classes of defense genes. Such a specialization within the oxidative stress response may explain the dissociated function of Yap1 and Skn7 in H 2 O 2 and cadmium resistance.
Gene Disruptions and Plasmid Constructions-Standard protocols and buffers were used (29). Gene disruptions were performed by the one-step gene disruption technique (28). pSKN7 is a YEp351 plasmid carrying a 3.5-kb genomic fragment containing the entire SKN7 gene (24). pSKN7-HA is a tagged version of SKN7 in which the HA epitope was introduced at the SKN7 PstI site 28 codons downstream of the ATG (15). The skn7::TRP1 construct used to create skn7⌬-1 lacks an internal 1.5-kb PvuII-HincII fragment of the wild type SKN7 gene, which has been replaced by the TRP1 gene (24). The yap1::LEU2 construct used to create yap1⌬-1 was prepared by removing the YAP1 coding sequence from the BamHI site (ϩ186) to the KpnI site (ϩ1650) relative to the ATG and replacing it with the LEU2 gene. pYAP1 was constructed by subcloning a 2.5-kb EcoRI DNA fragment carrying the entire YAP1 gene (12) into pRS426. pMF6(X-H) is a PUC18 plasmid lacking the AccI polylinker site and carrying the same 2.5-kb EcoRI YAP1 fragment. To generate pYAP1-9Myc, a polymerase chain reaction-amplified 390-bp sequence encoding 9 Myc epitopes was first introduced in vector pMF6(X-H) into the YAP1 AccI site located two codons downstream of the ATG; The YAP1-9Myc fusion was then subcloned at the EcoRI site of pRS426 to generate pYAP1-9Myc. The functionality of the YAP1-9Myc construct was evaluated by its ability to rescue the H 2 O 2 -hypersensitive phenotype of yap1⌬-1. The TSA1-lacZ gene fusions were constructed as follows: a 4.5-kb DNA fragment spanning the entire lacZ coding sequence from the BamHI site, two codons downstream of the ATG to a KpnI site approximately 1 kb from the stop codon, was subcloned into pRS424 to generate pMT1. A NaeII to KpnI 328-bp plasmid fragment containing the ␣ peptide lacZ sequence was then removed from pMT1 to generate pMT11. TSA1 promoter fragments corresponding to Ϫ1000, Ϫ837, Ϫ403, Ϫ243, and Ϫ204 to ϩ1 relative to the ATG were amplified by polymerase chain reaction from genomic DNA and subcloned between the XhoI and BamHI sites of pMT11.
Sensitivity Assays-Patch assays were performed as follows: 10-l aliquots containing approximately 2 ϫ 10 3 cells of an overnight culture were spotted on rich broth or synthetic complete solid plates containing H 2 O 2 , t-BOOH, or cadmium sulfate at the indicated concentration. Plates were monitored after 3-6 days incubation at 30°C.
Northern Blot Analysis-Yeast cells from overnight culture were diluted to an A 600 of 0.01 in synthetic complete medium and incubated with shaking at 30°C until they reached an A 600 of 0.3. The cells were then aliquoted and incubated in the absence or in the presence of H 2 O 2 (0.2 mM) for 20 min. Total RNA was prepared by the hot phenol method (30). For each condition tested, 20-g RNA samples were loaded per lane on an agarose gel containing formaldehyde, separated by electrophoresis, transferred to a nylon membrane (Bio-Rad), and hybridized with the indicated random primed (Roche Molecular Biochemicals) 32 Plabeled DNA probe. Hybridization of each blot with a small nuclear RNA U3 (SNR17A) specific 32 P-labeled DNA probe served as a RNA loading control. Pre-hybridization, hybridization, and washes were carried out as described (29). Hybridized membranes were exposed for autoradiography.
Yeast Crude Extracts and Electrophoretic Mobility Shift Assays-Yeast crude extracts used in EMSAs were prepared as follows. Cells were grown to an A 600 of 0.3 and were left untreated or were treated with H 2 O 2 (0.6 mM) for 5 min and harvested. Extracts were prepared by glass bead disruption as described previously (7), except for the use of a modified breakage buffer containing 200 mM Tris-HCl, pH 8.0, 10 mM MgCl 2 , 10% glycerol, complete mixture inhibitor (Roche Molecular Biochemicals). The DNA binding reactions were carried out in 1ϫ TC buffer (25 mM Tris-HCl, pH 7.5, 50 mM NaCl, 2 mM EDTA, pH 8, 5 mM MgCl 2 , 0.1% (v/v) CHAPS, 10% (v/v) glycerol) with 20 g of crude yeast extracts, 5-15 fmol [ 32 P]ATP-labeled probe, 1 g of poly(dI-dC) (Amersham Pharmacia Biotech) in a total volume of 20 l. The binding reaction was incubated for 5 min at room temperature and subjected to electrophoresis in a 6% polyacrylamide gel (acrylamide/N,NЈ-methylenebisacrylamide weight ratio, 27.5:1) in 45 mM Tris/45 mM boric acid/1 mM EDTA for 1 h at 200 V.
Measurement of the H 2 O 2 Response by Two-dimensional Gel Electrophoresis-Measurement of the H 2 O 2 response was carried out as described previously (7). Basically, mid-log cells (A 600 ϭ 0.3) were exposed or not to H 2 O 2 (0.2 mM) for 15 min, pulse labeled with [ 35 S]methionine for another 15 min, and harvested. An equal aliquot of 3 H-labeled cells was mixed to the 35 S-labeled cells and served as an internal protein concentration standard for each two-dimensional gel spot. Cell mixtures were extracted and subjected to comparative two-dimensional gel analysis. Accordingly, the previously identified 71 proteins of the H 2 O 2 stimulon were analyzed in wild type and isogenic yap1⌬-1, and skn7⌬-1. Computer Search of Yap1 and Skn7 Upstream Recognition Sequences-Specific DNA sequences were searched within 1 kilobase from the initiation codon of identified genes with the Saccharomyces Genome Data Base package. Searches were done with the following query sequences: T(T/G)ACTAA, which corresponds to the known Yap1 recognition sequences (YRE) (9,13,14), and CAGCAGCCGAAAAGA, which correspond to a 23-bp TRX2 promoter sequence capable of binding Skn7 in vitro (11).

Yap1 and Skn7 Have Distinctive Roles in the Oxidative and
Metal Stress Responses-Although strains carrying deletions of YAP1 (yap1⌬-1)or SKN7 (skn7⌬-1) are both hypersensitive to killing by H 2 O 2 (9 -11), skn7⌬-1 can tolerate higher concentrations of H 2 O 2 (Fig. 1A). The same phenotypic profile is seen in the tolerance of t-BOOH (Fig. 1B). Another functional distinction is the ability of YAP1 overexpression to partially rescue the skn7⌬-1 peroxide-hypersensitive phenotype, whereas SKN7 overexpression has no effect in yap1⌬-1 (Fig. 1B). Analysis of the cadmium tolerance further demonstrates the distinctive roles of Yap1 and Skn7 in controlling stress responses (Fig. 1C). Whereas yap1⌬-1 is hypersensitive to cadmium (19,20,23), skn7⌬-1 is significantly more resistant than the wild type strain to this toxic metal. The double delete yap1⌬-1,skn7⌬-1 is hypersensitive to cadmium, suggesting that the skn7⌬-1 cadmium hyperresistance phenotype is dependent upon YAP1. Therefore, Yap1 and Skn7 have distinctive roles in peroxide stress tolerance and opposite effects upon cadmium tolerance, with Yap1 acting positively and Skn7 acting negatively. A negative role for Skn7 is also suggested by the decreased cadmium tolerance observed upon overexpression of SKN7 (not shown). These results prompted us to evaluate the role of Yap1 and Skn7 in the control of known H 2 O 2 -inducible target genes.

Yap1 and Skn7 Do Not Always Function Together in the Activation of H 2 O 2 -inducible Target Genes-Yap1 and Skn7
co-operate to activate TRX2 and TRR1 in response to H 2 O 2 (11). We thus evaluated by Northern blot the respective roles of Yap1 and Skn7 in the control of three other known Yap1 targets, SSA1 (31), GSH1 (13), and GLR1 (18), and of four other H 2 O 2 -inducible genes, TSA1 (TSA or peroxiredoxin), AHP1 (an alkyl hydroperoxide reductase) (32), CCP1 (cytochrome c peroxidase), and HSP82 (7) (Fig. 2). The genes analyzed were all potently induced by H 2 O 2 in wild type cells, and this induction was abolished in yap1⌬-1 (Fig. 2, A and B). In skn7⌬-1 cells, the induction by H 2 O 2 was also abolished for TSA1, CCP1, TRR1, HSP82, and SSA1 and significantly diminished for TRX2 and AHP1 ( Fig. 2A). In contrast, induction of GLR1 and GSH1 by H 2 O 2 was actually stronger in skn7⌬-1 than in wild type cells (Fig. 2B). However, in yap1⌬-1,skn7⌬-1 double null cells, the induction of GSH1 by H 2 O 2 was totally abolished, demonstrating that the H 2 O 2 -superinduced levels seen in skn7⌬-1 cells are dependent upon YAP1. Therefore, Yap1 and Skn7 co-operate in the control of several H 2 O 2 target genes but have opposite effects in the control of other H 2 O 2 target genes, with Yap1 acting positively and Skn7 acting negatively. These results may explain, at least in part, the opposite functions of Yap1 and Skn7 in cadmium tolerance.
Yap1 Controls a Large Regulon of Thirty-two H 2 O 2 -inducible Proteins-To further dissect the intricate functions of Yap1 and Skn7 in the control of the oxidative and metal stress responses, we searched for other target genes by comparative two-dimensional gel electrophoresis of total soluble yeast proteins. We recently identified with this method 71 proteins whose synthesis rate is significantly increased minutes after exposure to H 2 O 2 (7). We sought to identify among these proteins those whose induction by H 2 O 2 would be lost or significantly diminished in yap1⌬-1. Exponentially growing wild type and yap1⌬-1 cells were pulse-labeled after exposure to H 2 O 2 for 15 min and then subjected to two-dimensional gel electrophoresis (Fig. 3, A-C). Uninduced and H 2 O 2 -induced synthesis rate indexes of the 71 H 2 O 2 targets were calculated in yap1⌬-1 and divided by those of the wild type strain (yap1⌬-1/WT). 31 proteins with a yap1⌬-1/WT-induced synthesis rate index ratio equal to or below the value of 0.6 were considered as dependent upon Yap1 for their induction by H 2 O 2 . Their synthesis rate indexes are represented in Fig. 4. These proteins were sorted into functional classes (Table I). The Yap1 regulon includes most of the oxidant scavenging enzymes. These are, in addition to those mentioned above, cytosolic catalase (Ctt1p), copper/zinc and manganese superoxide dismutases (Sod1p and Sod2p), YDR453Cp, and YOL151Wp. YDR453Wp is an AhpC/TSA family member, and YOL151Wp is similar to plant NADPH isoflavonoid reductases shown to rescue the diamide hypersensitiv-  skn7⌬-1 (⌬skn7), and yap1⌬-1,skn7⌬-1 (⌬yap1⌬skn7) strains were compared by the patch assay for their ability to grow on rich broth solid medium containing H 2 O 2 at the indicated concentration. For each strain, 10 l of an overnight culture (A 600 ϭ 2-3) diluted to 2 ϫ 10 3 cells were spotted on tester plates. B, wild type and isogenic yap1⌬-1 (⌬yap1) and skn7⌬-1 (⌬skn7) not carrying (WT, ⌬yap1, ⌬skn7) or carrying an episomal plasmid containing either YAP1 (pYAP1) or SKN7 (pSKN7) were compared by the patch assay for their ability to grow on rich broth solid medium containing t-BOOH at the indicated concentration. C, wild type (WT) and isogenic yap1⌬-1 (⌬yap1), skn7⌬-1 (⌬skn7), and yap1⌬-1,skn7⌬-1 (⌬yap1⌬skn7) strains were compared by the patch assay for their ability to grow on synthetic complete solid medium containing cadmium sulfate at the indicated concentration. 3) wild type (WT) and isogenic yap1⌬-1 (⌬yap1) and skn7⌬-1 (⌬skn7) cells, which had not been treated or were treated for 20 min with H 2 O 2 (0.2 mM) as indicated. The resulting RNA were subjected to Northern blot probed for the indicated genes and for U3 as a loading control, as described under "Materials and Methods." B, Northern blot analysis of Yap1-dependent and Skn7-independent target genes (same as in A). C, total RNA was isolated from a wild type and isogenic yap1⌬-1,skn7⌬-1 (⌬yap1⌬skn7) strains (same as in A and B). ity phenotype of a yap1 null strain (33). The Yap1 regulon also includes several carbohydrate metabolism enzymes, a few heat shock proteins and proteases, amino acid metabolism enzymes, and other unclassified or unknown proteins.
A Large Subset of the YAP1 Regulon Is Dependent upon Skn7-Skn7 target genes were similarly identified by comparative two-dimensional gel electrophoresis (Fig. 3, A-C). Thirteen proteins with a skn7⌬-1 to wild type (skn7⌬-1/WT)-induced synthesis rate index ratio equal to or below 0.6 were considered as dependent upon Skn7 for their induction by H 2 O 2 (Table I). Their synthesis rates indexes are represented in Fig.  4. Most of these proteins were also identified as Yap1 target genes. However, two of them were not identified as Yap1 targets. Conversely, several Yap1 target genes were still normally induced or even superinduced in the skn7 null strain (Fig. 4).
Hence, Skn7 appears to partition the Yap1 regulon into three distinct gene subsets. The first subset consists of 13 Yap1-and Skn7-dependent proteins. It includes most of the Yap1-dependent oxidant scavenging enzymes and a few other activities. The second subset contains 19 Yap1-dependent but Skn7-independent proteins and includes Glr1p, Zwf1p, Tal1p, Cys3p, and, as shown by Northern blot, GSH1. The third subset consist of two Skn7-dependent but Yap1-independent targets.
There was an excellent correlation between protein and mRNA levels for most of the genes analyzed. However, for Pgm2p and Dnm1p, in mutant strains mRNA levels appeared normal despite altered protein synthesis rates (not shown).
These discrepancies between mRNA and protein levels may indicate the existence of post-transcriptional control mechanisms.
Identification of Upstream Regulatory Sequences for Yap1 and Skn7-Approximately half of the Yap1 target genes contain a YRE motif within 1 kb upstream from their translation start, whereas the other half do not contain such a motif (Table  I) (see "Materials and Methods"). We thus analyzed one such YRE negative target, the TSA1 gene, for the presence of a Yap1/Skn7-responsive element. A translational fusion between a 1-kb region upstream of the TSA1 translational start and the bacterial lacZ gene was created, and its expression was analyzed by Northern blot (Fig. 5A). This promoter-lacZ fusion was inducible by H 2 O 2 in wild type but not in yap1⌬-1 or in skn7⌬-1 cells, demonstrating that it recapitulates the regulation of the TSA1 gene. A 5Ј deletion analysis showed that a minimal promoter sequence spanning 204 bp from the translational start (Ϫ204/ϩ1) retained the H 2 O 2 inducibility of lacZ gene expression in wild type cells (Fig. 5B). Furthermore, this induction was abolished in yap1⌬-1 and diminished in skn7⌬-1. A higher constitutive expression from this minimal promoter fragment was also observed. These data suggest that the Ϫ204/ϩ1 minimal promoter fragment retains the Yap1-and Skn7-dependent H 2 O 2 induction of the TSA1 gene but may have lost a repression mechanism imposed on the full TSA1 promoter.
To examine whether Yap1 and Skn7 affected the expression of the TSA1 gene by direct binding to its promoter, we performed EMSA with a probe corresponding to the TSA1 promoter Ϫ204/ϩ1 fragment. A H 2 O 2 -inducible slow migrating binding complex (complex a) was seen with wild type but not with yap1⌬-1,skn7⌬-1 cell extracts (Fig. 6A) or with yap1⌬-1 or skn7⌬-1 cell extracts (not shown). The same but more intense shifted complex was already seen prior exposure to H 2 O 2 with YAP1 and SKN7 overexpressing cell extracts, suggesting that it contains Yap1 and Skn7 (Fig. 6B, lanes 7 and 8). Note that H 2 O 2 treatment resulted in a slight change in the mobility of complex a. The presence of Yap1 and Skn7 in complex a could be demonstrated with the use of specific monoclonal antibodies (mAb). When the EMSAs were performed with extracts expressing a Myc epitope-tagged version of Yap1, adding the anti-Myc mAb to the binding reaction resulted in the decrease and supershift of complex a (Fig. 6C, lanes 1 and 2). Similarly, adding the anti-HA mAb to EMSAs performed with extracts expressing a HA epitope-tagged version of Skn7 (Fig. 6D, lanes  1 and 2) or to extracts overexpressing both Yap1 and the HA-tagged Skn7 (lanes 5 and 6) resulted in the complete disappearance of complex a.
Extracts from skn7⌬-1 cells overexpressing Yap1 did not generate complex a, but instead two faster migrating complexes (complexes b and c) (Fig. 6B, lanes 3 and 4). These two complexes were also seen, in addition to complex a, with extracts from cells overexpressing both Yap1 and Skn7 (lanes 7 and 8). We conclude that they both contain Yap1 because they could be supershifted with anti-Myc mAb (Fig. 6C, lanes 3 and  4). Extracts from yap1⌬-1 cells overexpressing Skn7 also did not result in the formation of complex a but instead resulted in the formation of a faster migrating complex (complex d) (Fig.  6B, lanes 5 and 6). Complex d contains Skn7, as demonstrated by its disappearance when anti-HA mAb are added to the binding reaction (Fig. 6C, lanes 3 and 4). We conclude that both Skn7 (complex d) and Yap1 (complex b and c) can bind directly and independently to the TSA1 promoter fragment and can also form a ternary complex (complex a) with this fragment. Identical EMSA were observed with a 400-bp-long TSA1 promoter fragment (Ϫ400/ϩ1) and with a probe corresponding to the TRX2 promoter (Ϫ300/ϩ1) (not shown).

DISCUSSION
The respective contributions of the Yap1 and Skn7 transcription factors to the control of the S. cerevisiae response to oxidative stress and metal toxicity were analyzed. Yap1 controls a large regulon of at least 32 proteins, accounting for approximately half of the 71 identified proteins of the H 2 O 2 stimulon (7). As shown for TRX2 and TRR1 (11), 15 of these proteins also require the presence of both Yap1 and Skn7 for their induction by H 2 O 2 . Hence, two subsets of the Yap1 regulon can be delineated. Interestingly, these two gene subsets distinguish the antioxidant scavenging enzymes from the metabolic pathways that regenerate the main cellular reducing power, GSH and NADPH. These two gene subsets appear to carry specialized defense functions as suggested by the fact that although both Yap1 and Skn7 are essential for resistance to peroxides (H 2 O 2 and t-BOOH, a synthetic alkyl hydroperoxide), only Yap1 is required for resistance to cadmium.
The Genes Required for Resistance to Peroxides Are Specified in the Entire Yap1 Regulon-The Skn7-dependent subset of the Yap1 regulon specifies genes important for peroxide tolerance. However, the more severe H 2 O 2 -hypersensitive phenotype of yap1⌬-1 versus skn7⌬-1 indicates that, in addition to the Skn7dependent gene subset, activities of the Skn7-independent  Yap1 regulon play a role in peroxide tolerance. The partition of the Yap1 regulon by Skn7 correlates with two distinct classes of defense activities: (i) The Skn7-dependent gene subset includes most of the known yeast redox-active oxygen intermediate scavenging activities. These are cytosolic catalase (CTT1), cytochrome c peroxidase (CCP1), Tsa1p (TSA1) or peroxiredoxin (34) and its homologue YDR453Cp (7), an alkyl hydroperoxide reductase (AHP1) also related to the AhpC/TSA family (32), thioredoxin 2 (TRX2), thioredoxin reductase (TRR1), copper/zinc (SOD1), and manganese (SOD2) superoxide dismutases. Therefore, three yeast proteins of the AhpC/TSA peroxiredoxin family, which are believed to reduce H 2 O 2 with electrons donated by NADPH via thioredoxin (TRX2) and thioredoxin reductase (TRR1) (32,35,36), together with their essential cofactors Trx2p and Trr1p, are coordinately controlled by the same H 2 O 2 -inducible control mechanism. Consistent with their H 2 O 2 inducibility, the participation of CTT1, TRX2, TSA1, AHP1, and TRR1, in yeast peroxide protection is demonstrated by the increased sensitivity of the corresponding null mutants to killing by these oxidants (9,32,(35)(36)(37)(38) (40,41). (ii) The Skn7-independent gene subset comprises several activities of glutathione and pentose phosphate pathways. Cystathionine ␥ lyase (CYS3) generates cysteine, a precursor of glutathione (GSH), from cystathionine. ␥-Glutamyl cysteine synthase (GSH1) is the rate-limiting enzyme in GSH biosynthesis, and NADPH-dependent glutathione reductase (GLR1) acts to recycle oxidized GSH (GSSG). Glucose-6-phosphate dehydrogenase (ZWF1) regulates the carbon flow through the pentose phosphate pathway by catalyzing its first oxidative step, and transaldolase (TAL1) catalyzes the interconversion of pentose phosphates that have been generated in the oxidative part of this pathway (42). A set of four other proteins of this gene subset could be also involved in the maintenance of the pyrimidine redox balance. These are a putative NADPH dehydrogenase (Oye3p), two proteins with similarity to alcohol dehydrogenase (YMR318C and YNL134C), and YDR032C, which encodes a member of a new family of flavodoxin-like proteins (43). The GSH pathway is most likely involved in peroxide resistance by the reduction of peroxides by glutathione peroxidase using the pentose phosphate pathway as the primary source of NADPH, the electron donor system for glutathione and thioredoxin reductases. The importance of GSH1, GLR1, ZWF1, and TAL1 in the yeast cellular tolerance of H 2 O 2 is demonstrated by the increased sensitivity of their null mutants to this oxidant (18, 44 -48). Other activities are also part of this gene subset (Table I). Their role in the tolerance of H 2 O 2 has been recently discussed (7). The yeast peroxide stress response is different from that of Escherichia coli and Salmonella tiphymurium, which involves a H 2 O 2 -inducible regulon controlled by the regulator OxyR (49). The OxyR-controlled regulon includes katG (hydroperoxidase I), ahpCF (an alkyl hydroperoxide reductase of the AhpC/ TSA family), dps (a nonspecific DNA-binding protein), gorA (glutathione reductase), and grxA (glutaredoxin) (50,51). However, in contrast to the Yap1 regulon, the OxyR regulon excludes both sodA (manganese superoxide dismutase) and zwf (glucose 6 phosphate dehydrogenase), which are part of the O 2 . -inducible regulon under control of the SoxR/S regulator (52). The Genes Required for Cadmium Tolerance Are Mostly Confined in the Skn7-independent Subset of the Yap1 Regulon-GSH represents a first line of defense against cadmium toxicity (53). GSH acts through a cadmium detoxification pathway that involves the yeast cadmium factor Ycf1p, a vacuolar GSH Sconjugate pump of the ABC superfamily of transporters (54 -56). Interestingly, YCF1 is also a target of Yap1 (19). In this pathway, cadmium is conjugated to GSH, and the GS-cadmium conjugates are transported to the vacuole by Ycf1p. Glutathione reductase (GLR1) probably acts as a cadmium tolerance factor by recycling oxidized GSH, which accumulates upon cadmium exposure (39) and is known to inhibit Ycf1p (55). The role of GSH1 (39), GLR1, and ZWF1 in the yeast tolerance of cadmium is demonstrated by the increased sensitivity of their respective null mutants to this toxic metal (Spector, unpublished observations). A few activities of the Skn7-dependent gene subset also play a role in cadmium tolerance. These are alkyl hydroperoxide reductase (AHP1) and thioredoxin reductase (TRR1), as shown by the cadmium-hypersensitive phenotype associated with deletion of their gene (32) (Spector, unpublished observations). However, neither peroxiredoxin (TSA1), nor thioredoxin 2 (TRX2) are important in the toler-  1, 3, 5, and 7) and H 2 O 2 -induced (lanes 2, 4, 6, and 8) wild type (lanes 1 and 2), skn7⌬-1 carrying pYAP1 (lanes 3 and 4), yap1⌬-1 carrying pSKN7 (lanes 5 and 6), and wild type cells carrying both pYAP1 and pSKN7 (lanes 7 and 8). Complexes sensitive to the presence or the absence of Yap1 are indicated with arrows as a, b, and c. Complexes sensitive to the presence or the absence of Skn7 are indicated with arrows as a and d. C, EMSA performed with extracts prepared from H 2 O 2 -induced wild type (lanes 1 and 2) and skn7⌬-1 carrying pYAP1-9Myc (lanes 3 and 4). The 9E10 anti-Myc monoclonal antibody was added to the binding reaction in lanes 2 and 4. D, EMSA performed with extracts prepared from H 2 O 2 -induced wild type (lanes 1 and 2), yap1⌬-1 carrying pSKN7-HA (lanes 3 and 4) and wild type cells carrying both pYAP1 and pSKN7-HA. The 12CA5 anti-HA monoclonal antibody was added to the binding reaction in lanes 2, 4, and 6. ance of this toxic metal (not shown).
A Co-operative Gene Control Exerted by Yap1 and Skn7-As shown previously (11), Yap1 and Skn7 exert a dual control on the expression of oxidative stress defense genes. However, the data presented here also show that these two regulators do not always function together. In addition, Skn7 is not only dispensable but has a negative effect upon cadmium tolerance and upon the expression of several Yap1 target genes such as GSH1 and GLR1. This may suggest that Skn7 acts as a repressor for the Skn7-independent Yap1 gene subset. Alternatively, the defective expression of several oxidant scavenging enzymes in skn7⌬-1 could create cellular pro-oxidative conditions resulting in a sustained activation of Yap1. It is also possible that, as recently proposed for OxyR (51), one of the Skn7 target genes participates in a negative autoregulatory loop interrupted in skn7⌬-1, thus leading to a deregulation of Yap1 activity.
Approximately half of the genes of the Yap1 regulon do not contain a YRE. Still, one such gene, TSA1, carries a Yap1-and Skn7-responsive element within a 200-bp proximal region of its promoter and directly binds Yap1 and Skn7. This strongly suggests that TSA1 and very likely most of the other YRE negative genes are directly controlled by Yap1 and Skn7. They also indicate the existence of another Yap1 DNA recognition sequence in addition to the YRE. Efforts are being made to characterize this sequence, as well as the sequence recognized by Skn7.
Conclusions-The H 2 O 2 -inducible Yap1 transcription factor controls a major yeast stress response regulon, which is partitioned into two distinct classes of defense genes by the differential requirement for the Skn7 transcription factor. Such an organization may provide the cell with the ability to use two overlapping sets of genes in distinct stress responses. However, like H 2 O 2 , cadmium can induce both gene subsets of the Yap1controlled regulon (not shown). Conditions under which Yap1 is active and Skn7 is inactive would result in the differential induction of these two gene subsets. The function of Skn7 can be modulated by phosphorylation of the conserved aspartate of the receiver domain (15,16,25). However, although important for the contribution of Skn7 to cell cycle control, cell wall metabolism, and modulation of the osmo-stress HOG pathway, this aspartate is dispensable in the oxidative stress response (10,11). Another post-transcriptional modification might exist that affects the ability of Skn7 to co-operate with Yap1.