In the budding yeast Saccharomyces cerevisiae, arsenic detoxification involves the activation of Yap8, a member of the Yap (yeast AP-1-like) family of transcription factors, which in turn regulates ACR2 and ACR3, genes encoding an arsenate reductase and a plasma-membrane arsenite-efflux protein respectively. In addition, Yap1 is involved in the arsenic adaptation process through regulation of the expression of the vacuolar pump encoded by YCF1 (yeast cadmium factor 1 gene) and also contributing to the regulation of ACR genes. Here we show that Yap1 is also involved in the removal of ROS (reactive oxygen species) generated by arsenic compounds. Data on lipid peroxidation and intracellular oxidation indicate that deletion of YAP1 and YAP8 triggers cellular oxidation mediated by inorganic arsenic. In spite of the increased amounts of As(III) absorbed by the yap8 mutant, the enhanced transcriptional activation of the antioxidant genes such as GSH1 (γ- glutamylcysteine synthetase gene), SOD1 (superoxide dismutase 1 gene) and TRX2 (thioredoxin 2 gene) may prevent protein oxidation. In contrast, the yap1 mutant exhibits high contents of protein carbonyl groups and the GSSG/GSH ratio is severely disturbed on exposure to arsenic compounds in these cells. These results point to an additional level of Yap1 contribution to arsenic stress responses by preventing oxidative damage in cells exposed to these compounds. Transcriptional profiling revealed that genes of the functional categories related to sulphur and methionine metabolism and to the maintenance of cell redox homoeostasis are activated to mediate adaptation of the wild-type strain to 2 mM arsenate treatment.
- arsenic stress
- oxidative stress
- transcriptional regulation
- yeast AP-1-like transcription factor Yap1 gene (YAP1)
- yeast AP-1-like transcription factor Yap8 gene [YAP8; ACR1 (arsenic compound resistance protein 1 gene)
- ARR1 (arsenicalresistance protein 1 gene)]
Arsenic (As) is a highly toxic metalloid widely distributed in Nature and mostly found in drinking water. The first step of inorganic As(V) removal from the cytoplasm consists of its two-electron reduction to As(III) using glutathione as the source of reducing potential . Chronic exposure to this compound is generally associated with an increased risk of multiple cancers, vascular diseases, developmental anomalies and neurological disorders [2–4]. To counteract the deleterious effects caused by arsenic compounds, almost all living organisms have developed mechanisms to eliminate it. In the budding yeast Saccharomyces cerevisiae, resistance to arsenic is achieved through the activation of the transcriptional regulator Yap8 [yeast AP-1-like transcription factor Yap8, also called Acr1 (arsenic compound resistance protein 1)] , which in turn, induces the expression of an arsenate reductase and a plasma-membrane arsenite-efflux protein encoded by the genes ACR2 and ACR3 respectively [6-10]. In addition, the YCF1 gene product, yeast cadmium factor 1, also facilitates the vacuolar extrusion of glutathione-conjugated arsenite molecules . Although Yap8 is the main regulator of arsenic stress responses, Yap1 is also involved to a lesser extent through YCF1 activation under these conditions and contributes to the full activation of enzymes encoded by the ACR genes . Both regulators belong to the Yap family of bZIP (basic domain/leucine zipper) transcription factors, formed by eight members , which modulates the activation of specific genes in response to various stress (for a review, see ).
Arsenic toxicity and carcinogenicity in animals has been suggested to be probably due to the generation of an oxidative stress, thus provoking a deleterious effect by this metal . Yeast mutants in genes related to several mitochondrial processes, which show sensitive phenotypes to arsenic compounds, were recently identified. A total of 20 specific-As(V)- sensitive mutants were found, from which 13 genes have orthologues in humans . On the other hand, high-throughput arsenite-triggered changes in transcriptional profiling [15,16] indicate that cell antioxidant defences are up-regulated in yeast. Furthermore, other investigators have shown a dose-dependent increase in the levels of peroxidation of membrane lipids as a consequence of arsenite exposure . Yap1, the best-characterized member of the Yap family and the major regulator in oxidative stress, is involved in arsenic stress responses. These facts together led to the hypothesis that arsenic induces oxidative stress in which Yap1 plays a major role. It is indeed known that arsenite [As(III)] can react with the thiol groups of proteins, inhibiting many biological pathways, whereas the pentavalent form [As(V)] of arsenic is a phosphate analogue interfering with phosphorylation reactions . Although the toxic effect of both oxidation states, As(V) and As(III), appears to be very similar, the elimination of As(V) requires its reduction to As(III) using the redox potential of GSH and thus interfering with the GSH pool of the cell . As recently pointed by other investigators , tolerance to either arsenate or arsenite also involves specific sets of mitochondrial genes.
We decided to evaluate, using biochemical and molecular approaches, the damage caused by As(V) and As(III). We show that oxidative stress is generated as an effect of arsenic exposure in strains defective in the arsenic-extrusion machinery and in the antioxidant defence system. By measuring the GSH/GSSG ratios, we provide evidence indicating that arsenic compounds trigger the disruption of the redox equilibrium (being the homoeostasis rapidly achieved through the enhancement of GSH generation). Transcriptional profiling of the wild-type strain under exposure to arsenate reveals the induction of many Yap1-dependent genes and genes involved in sulphur metabolism. Our results show that the antioxidant defences are up-regulated in the mutant yap8, which absorbs increased amounts of arsenite, in comparison with the parental strain. Since the status of protein carbonylation is not changed in the wild-type and yap8 strains, we conclude that the activation of the antioxidant system under arsenic stress prevents the accumulation of oxidized proteins. Consistent with this notion, the yap1 mutant displays high levels of protein oxidation.
Strains, plasmids and growth conditions
The yeast strains used in the present study were: BY4741 MAT a; his3Δ1; leu2Δ0; met15Δ0; ura3Δ0 (EUROSCARF), BY4741 Δyap1 MAT a; his3Δ1; leu2Δ0; met15Δ0; ura3Δ0; YML007w::kanMX4 (EUROSCARF), BYΔyap8 MAT a; his3Δ1; leuΔ0; met15Δ; lys2Δ0; ura3Δ0; YPR199c::kanMX4 , BY Δyap1Δyap8 MAT a; his3Δ1; leuΔ0; met15Δ0; lys2Δ0; ura3Δ0; YPR199c::kanMX4; YML007w::HIS  and FT4 yap1 MAT a; ura3-52; trp1Δ63; his3-Δ200; leu2::PET56; yap1Δ . The complete coding region of YAP8 gene was deleted by the microhomology PCR method  to create the strain FT4 yap8 MAT a; ura3-52; trp1Δ63; his3-Δ200; leu2::PET56; yap8::KAN. Deletion was confirmed by PCR analysis of genomic DNA using upstream and downstream primers. To overexpress YAP8, the corresponding chromosomal region was amplified by PCR using the primers 5′-CCATTGTAGGAGAGTAACCT-3′ and 5′-CATCGAATACTCCACATCGATC-3′. The product was first cloned using the Zero Blunt® TOPO® PCR cloning Kit (Invitrogen) and the XbaI/BamHI fragment was subcloned into the 2 μ vector YEplac195 . The construct was sequenced using the ABI Prism DyeDeoxy Terminator Cycle Sequencing Kit (Applied Biosystems) and ABI Prism 373A Automatic Sequencer (PerkinElmer). The plasmid overexpressing YAP1 was available in our laboratory . The CEN plasmids expressing the myctagged YAP1 and GFP (green fluorescent protein)-tagged YAP1 and YAP8 versions were obtained from Dr M.B. Toledano's group , from Kuge and colleagues  or was available in our laboratory  respectively. Strains were grown in complete YPD [1% yeast extract, 2% (w/v) bactopeptone and 2% (w/v) glucose] or selective media [SC (synthetic complete) or SD (minimal synthetic defined): 0.67% ammonium sulfate/yeast nitrogen base without amino acids (Difco) and 2% (w/v) glucose] supplemented with the appropriate selective amino acids. Early-exponential-phase cells [attenuance (D600) 0.4–0.5] were stressed by the addition of 2 mM As(V) (Na2HAsO) or As(III) (NaAsO2) and samples were collected at the indicated time points. Phenotypic growth assays were carried out by spotting 5 μl of an early-exponential-phase sequentially diluted culture (approx. 2000–20 cells) in selective medium containing up to 2 mM Na2HAsO or NaAsO2. Growth was recorded after 2 days at 30 °C. The bacterial Escherichia coli strain XL1-Blue recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac [F′proAB lacIqZDM15 Tn10 (Tetr)] (Stratagene) was used as the host for routine cloning purposes. Standard methods were used for genetic analysis, cloning and transformation .
Intracellular oxidation, lipid peroxidation and protein carbonylation
Measurements of intracellular oxidation and lipid peroxidation were performed with 50 mg (dry weight) of mid-exponential-phase cells grown under physiological conditions or exposed to arsenate or arsenite stress. Intracellular oxidative stress generated by arsenic compounds was monitored by measuring changes in fluorescence resulting from the oxidant-sensitive probe DCF-DA (2′,7′-dichlorofluorescein diacetate) . A fresh ethanol stock solution of DCF-DA was added to the culture to a final concentration of 10 μM and cells were incubated for 15 min to allow uptake of the probe. Cells from each aliquot were cooled on ice, harvested by centrifugation at 2000 g for 5 min at 22 °C and washed twice with distilled water. The cell pellets were resuspended in 500 μl of water and lysed by vortex-mixing in the presence of 1.5 g of glass beads. The extracts obtained after centrifugation at 15 000 g for 5 min were diluted and fluorescence was measured using a Photo Technology International spectrofluorimeter set at an excitation wavelength of 504 nm and an emission wavelength of 524 nm with a slit width of 5 nm. The effect of deletions alone was measured under physiological conditions and the results under arsenic stress were expressed as the relation between the fluorescence of stressed and unstressed cells. Lipid peroxidation was determined by quantifying TBARS (thiobarbituric acid-reactive substances). After 24 h incubation with As compounds, cells were cooled on ice, harvested by centrifugation and washed twice with 20 mM Tris/HCl buffer, pH 7.4. The pellets were resuspended in 500 μl of the same buffer containing 10% (w/v) trichloroacetic acid and 1.5 g of glass beads were added. The samples were lysed using six cycles of 20 s agitation on a vortex-mixer, followed by 20 s on ice. Lipid peroxidation of cell extracts was monitored spectrophotometricaly at 532 nm in an EDTA/thiobarbituric acid/NaOH solution as described in , which determines the accumulation of TBARS in cells. The control data for each mutant strain grown under physiological conditions was expressed as the amount of MDA (malondialdehyde)/mg of cell dry weight formed and the results under arsenic stress are expressed as the relation between the values of stressed and non-stressed cells. Several authors have recently used this technique [26,27]. All these experiments were carried out at least three times, with no fewer than three replicate measurements in each experiment and the results are presented are means±S.D. To detect the presence of carbonyl groups introduced in proteins as a result of arsenic stress, an OxyBlot™ protein oxidation detection kit (Intergen) was used. The samples were analysed by immunoblotting and processed as described in , using rabbit anti-dinitrophenol antibody as the primary antibody. As a loading control, levels of the co-chaperone Sba1 were measured .
Measurements of thiols were performed by the spectrophotometric GR–DTNB [glutathione reductase–5,5´-dithiobis-(2-nitrobenzoic acid)] recycling method originally described by Griffith . The wild-type and mutant strains grown to early exponential phase were induced with 2 mM As(V), and samples collected at the indicated time points were extracted in a solution consisting of 0.1 M HCl and 1 mM EDTA, pH 1.35 . The kinetics of TNB [5-thio-2-nitrobenzoate] formation was monitored photometrically at 405 nm. GSSG concentrations were determined in the same extracts after a 30 min incubation of the supernatant with 2-vinylpyridine at room temperature (20–22 °C) to derivatize GSH . The concentration was determined by reference to a standard prepared in HCl and was expressed as nmol of glutathione/mg cell dry weight. Measurements were carried out three times, with three replicates in each experiment and results are means±S.D.
Determination of As (III) retention by atomic absorption
Analysis of the capacity of S. cerevisiae wild-type cells and Δyap1 and Δyap8 mutant versions to absorb and accumulate As(III) was determined using atomic-absorption spectrophotometry, as previously described . Arsenite, to a final concentration of 2 mM, was added to the medium containing 1 mg (dry weight) of mid-exponential-phase cells, and the culture was incubated at 28 °C in a rotating bath for 4 h and 24 h. For measuring residual As(III) present in the medium, 5 ml aliquots were centrifuged at 2000 g for 5 min at 22 °C and the supernatant was collected and subjected to atomic-absorption spectrophotometry using a PerkinElmer 3100 atomic-absorption spectrometer. As(III) absorption was calculated by determining the difference in metalloid content between the control medium without cells and the test medium containing cells. Percentages of As(III) absorption were calculated by use of the following equation:
The measurements were carried out three times, with three internal replicates and the results presented are means±S.D.
Protein extraction and immunoblot analysis
Δyap1 mutant cells transformed with the plasmid encoding c-myc–YAP1  were grown to early exponential phase and induced or not with 2 mM As(V). Samples collected at the indicated time points were harvested by centrifugation at 4 °C and the protein extracts were prepared by the trichloracetic acid-lysis method and immunoblotted as described in . To follow the kinetics of the recombinant c-Myc-Yap1 protein under arsenic stress, immunoblotting was performed with 50 μg of proteins that were probed with the 9E10 anti-c-Myc monoclonal antibody. Sba1 (p23), encoded by the SBA1 gene , was used as a loading control [29,34]. Detection was performed using an ECL® (enhanced chemiluminescence) Western-blotting reagent kit (Amersham Pharmacia).
FT4 yap8 and FT4 yap1 strains transformed with pRS encoding cp-GFP-HA-YAP8  or cp-GFP-HA-YAP1  respectively (where cp is centromeric plasmid and HA is haemagglutinin), were grown to early exponential phase and induced with either 2 mM As(V) or As(III) at the indicated time points. DAPI (4´,6-diamidino-2-phenylindole) was added as a DNA marker at a final concentration of 5 μg/ml, 5 min before microscopy. After washing with PBS, cells were resuspended in a solution of 200 mM DABCO (1,4-diazadicyclo[2.2.2]octane) in 75% (v/v) glycerol and 0.25×PBS (Sigma–Aldrich). Both we and Delaunay et al.  have shown that DABCO does not affect the localization of the GFP fusions. GFP signals were analysed in living cells with a Leica DMRXA fluorescent microscope equipped with a Roper Scientific Micro-Max cooled CCD (charge-coupled device) camera and MetaMorph software (Universal Imaging Inc.).
Northern Blot, real-time PCR and microarray analysis
RNA procedures were performed as described in . RNA was isolated from cultures that were either untreated or exposed to 2mM Na2HAsO or NaAsO2 at the indicated time points. For Northern-blot analysis, approx. 40 μg of total RNA was separated in formaldehyde gels and transferred on to nylon membranes (Hybond XL;Amersham Pharmacia Biotech). Intragenic PCR fragments of GSH1 and SNR17A/small nucleolar RNA U3 were used as probes. For real-time quantitative PCR the RNA samples were treated with DNase (TURBO DNAse-free; Ambion) according to the manufacturer's instructions. cDNAs were synthesized by reverse transcription from 0.5 μg of total RNA, using 50 pmol of (dT)15, 1 mM dNTP and 5 units of Transcriptor Reverse transcriptase as described by the manufacturer (Roche). cDNA amplification was quantitatively analysed by incorporation of SYBR Green I (LightCycler FastStart DNA Master SYBR Green I; Roche) into double-stranded DNA, according to the manufacturer's instructions, on a Roche LightCycler II Instrument, using the ACT1 (actin) gene as a loading control. The fold change was determined by the 2ΔΔCT method . The primers used were as follows: GSH1: 5′-GCTGCTGGTAAAAGAGACAATG-3′ and 5′-ACTCACATCGTTAGCCTCACAA-3′; TRX2: 5′-GGTCACTCAATTAAAATCCGCTTC-3′ and 5′-CGACGACTCTGGTAACCTCCTTAC-3′; SOD1: 5′-AGCCAACCACTGTCTCTTACGA-3′ and 5′-ACACCATTTTCGTCCGTCTTTA-3′; and ACT1: 5′-CTA TTG GTA ACG AAA GAT TCA G-3′ and 5′-CCT TAC GGA CAT CGA CAT CA-3′. For transcript profiling, total RNA was purified using the RNAeasy kit (Qiagen), followed by the RNA clean-up procedure. A 10 μg portion of RNA was used to generate labelled cDNA, which was hybridized on the DNA arrays as described in . We used arrays containing probes for most of the yeast open reading frame, obtained from the plate-forme transcriptome of the IFR36 (www.transcriptome.ens.fr). Slides were read using a Genepix 4000B scanner from Axon. The images were analysed with the Genepix pro 6.0 software. Data were normalized using global lowess followed by print tip group median from Goulphar software . Complete microarray data are available as Supplementary Table S1 at http://www.BiochemJ.org/bj/414/bj4140301add.htm. The gene-ontology analyses were performed by submitting the whole set of microarray results to the t-profiler tool  using default parameters. Redundant or meaningless functional categories were hidden (Table 1). The results presented in Table 1 and Supplementary Tables S1 and S2 (at http://www.BiochemJ.org/bj/414/bj4140301add.htm) are from three independent experiments. Only genes measured at least twice were kept for further functional analyses.
The results reported in the present study are the averages for at least three independent experiments, with three replicates in each experiment, and are expressed as the mean±S.D. Statistical differences among treatments were analysed by one-way ANOVA with Tukey's HSD (honest significant difference) multiple comparisons test (α=0.05) using STATISTICA for Windows (StatSoft Inc., Tulsa, OK, U.S.A.).
YAP8 overexpression does not alleviate yap1 sensitivity under As(V)
It has been suggested that Yap1 and Yap8 play distinct and well-defined roles in the arsenic stress response by regulating distinct sets of genes [15,16]. In order to evaluate the physiological relevance of this specificity, we performed growth complementation assays (Figure 1). The yap8 mutant reveals a severe growth-sensitive phenotype to both arsenate and arsenite, which is rescued by the overexpression of YAP8. The yap1 mutant shows a mild growth-sensitive phenotype under arsenate conditions. In contrast, under arsenite treatment this phenotype is more accentuated. The overexpression of YAP8 in the yap1 mutant partially alleviates the growth-sensitivity of this strain under arsenite treatment, a finding consistent with the notion that Yap1 and Yap8 exert specific roles in arsenic stress responses. On the other hand, YAP1 overexpression is not able to rescue the sensitive phenotype of the yap8 mutant strain either under As(V) or As(III) conditions (Figure 1). Altogether, these results suggest some level of specificity of YAP1 and YAP8 in arsenic detoxification.
Induction of antioxidant defences by arsenate
In order to characterize global changes in cells subjected to arsenate treatment, we performed microarray analysis. We therefore compared the transcriptome of cells submitted to a 30 min exposure to 2 mM of arsenate with the one of cells mock-treated with water. We made a global gene-ontology search using the t-profiler software  to identify the cellular pathways that were affected by arsenate (see Table 1 and Supplementary Tables S1 and S2). This analysis indicated that arsenate up-regulated genes involved in protein folding, sulfur and methionine metabolism (mainly target genes of the transcription factor Met4p), redox homoeostasis (including most of the target genes of Yap1) and proteasome activity (including the proteasome gene transcriptional regulator RPN4). This pattern is indeed very characteristic of the oxidative stress response and is similar to what has been described in the case of cell exposure to arsenite [15,16] or cadmium . Furthermore, the cellular pathway of response to stimulus, including the genes ACR2 and ACR3, is also shown to be induced by arsenate treatment.
Arsenic treatment generates oxidative damage in the yap1 mutant
Once having established that As(V) induces expression of genes involved in redox homoeostasis, we adopted several biochemical approaches to verify whether 2mM arsenate or arsenite treatment was associated with an oxidative environment in yeast cells. Intracellular oxidation was determined using the probe DCF-DA, which is sensitive to ROS (reactive oxygen species). Time-course experiments had been previously carried out in order to determine the maximum levels of intracellular oxidation, which occurred at 24 h of arsenic treatment (results not shown). The results in Figure 2(A) reveal that deletion of YAP1, but not of YAP8, interferes with the redox state of the cytoplasm under physiological conditions. The direct exposure of wild-type cells to either arsenate or arsenite does not lead to any detectable increase in the intracellular oxidation levels (Figure 2B), the level being identical in the absence of either YAP8 or YAP1 but increasing, however, when compared with the parental strain. This value is still more accentuated in the case of the double mutant. Furthermore, under physiological conditions, increased peroxidation levels of cellular lipids, as compared with the wild-type strain, were detected in the yap1 mutant through the formation of MDA (Figure 2C). Exposure of wild-type cells to both arsenate and arsenite stress does not cause any significant increase in lipid peroxidation compared with the unstressed cells (Figure 2D), a finding that is in good agreement with the intracellular oxidation results. In single and double mutants these levels are increased about 1.6- and 2.6-fold respectively. Taken together, these results show that arsenic compounds generate oxidative stress under conditions where the arsenate extrusion system and/or the antioxidant machinery are deficient. Protein carbonyl content is an indicator of oxidative stress and is by far the most commonly used biomarker of protein oxidation . To monitor the possible oxidative effects of arsenic treatment at the protein level, the changes in protein carbonyl status during exposure of yeast cells to both arsenate and arsenite were monitored up to 4 h. The resulting OxyBlots™ show that the amount of carbonylated proteins does not vary significantly in the wild-type strain and the mutant yap8, although in the latter the levels of oxidized proteins are slightly increased at all time points studied (Figure 3). A more severe effect was, however, observed in the yap1 strain. When these cells are treated with arsenate, the protein carbonyl content is slightly enhanced compared with the untreated cells. By contrast, exposure to arsenite caused a strong increase in the protein oxidation levels. This effect is even more severe during the first hour of treatment. Furthermore, additional carbonylated proteins, most of high molecular mass, were observed in the yap1 strain. These findings suggest that the effects of oxidative stress generated by arsenic compounds lead to protein oxidation only in the absence of yap1.
YAP1 deletion interferes with the redox equilibrium under arsenate
Our transcriptome analysis showed that arsenate up-regulates genes in the cellular pathway of sulphur and methionine metabolism. Furthermore, it has been shown that arsenite-exposed cells channel a large part of assimilated sulfur into glutathione biosynthesis . Yap1 is an important regulator of GSH1, which encodes the enzyme responsible for catalysing the condensation of cysteine on to the γ-carbon atom of glutamate in the limiting step of glutathione biosynthesis. This prompted us to evaluate how the requirement of Yap1 couples with the induction of GSH biosynthesis mediated by arsenate and arsenite. It was observed, by measuring the GSH and GSSG contents in the wild-type and mutant strains subjected to arsenate stress, that in all strains the GSH levels diminish about 2-fold during the first 1 h of treatment with 2 mM As(V) (Figure 4A). These values rise gradually over time, reaching physiological levels after 4 h. A similar pattern was observed under As(III) treatment (results not shown). The GSSG/GSH ratio increases after 1 h of treatment and reflects the decrease in GSH (compare Figures 4A and 4B). Notably, the double mutant Δyap1Δyap8 takes a longer time to recover the GSH levels of the parental strain. Our results show that the redox equilibrium is disrupted in strains bearing YAP1 deletions, although homoeostasis is rapidly achieved through the enhancement of GSH generation.
Yap1 is maintained in the induced state in long-term arsenate exposure
The biochemical assays clearly show that the wild-type strain, in contrast with the mutant yap1, does not suffer the deleterious effects of oxidative stress on exposure to inorganic arsenic. Indeed, the microarrays revealed that many of the antioxidant defences, as well as the arsenic detoxification system, were induced under As(V) treatment (see Supplementary Tables S1 and S2) to facilitate cell adaptation. In order to analyse the c-Myc–Yap1 protein levels in cells induced or not with arsenate, Westernblot analyses were performed. As Figure 5(A) shows, the levels of Yap1 were rather higher in arsenate-treated cells than in cells grown under physiological conditions. Yap1 protein levels peak at about 45 min after arsenate addition and a slight induction is maintained even up until 12 h of treatment. Furthermore, localization assays revealed that GFP–Yap1, as well as GFP–Yap8, are accumulated in the nucleus until 24 h of arsenate treatment (Figure 5B).
GSH1, SOD1 and TRX2 are highly induced under arsenic treatment
To evaluate whether Yap1 activation reflects the induction of the antioxidant cell defences such as GSH1, SOD1 and TRX2, we monitored their transcriptional activation by real-time PCR. Figure 6 reveals that, in the wild-type strain, these genes are highly induced by 2 mM arsenite or arsenate. Strikingly, transcriptional activation of GSH1, SOD1 and TRX2 is even higher in the yap8 mutant than in the wild-type strain. Under conditions of exposure to arsenate, all of these genes display a first peak of induction at 90 min, although after 24 h incubation with the metalloid the mRNA levels keep increasing in yap8 (see Figure 6A). The pattern of mRNA induction in the wild-type and yap8 strains is very similar up to 4 h exposure to arsenite. However, after this point a strong transcriptional activation of the three genes is observed only in the yap8 mutant (Figure 6B). Expression of SOD1 and TRX2 is completely abolished in the yap1 mutant (results not shown), a finding consistent with the fact that Yap1 regulates them. Some level of Yap1-independent GSH1 induction was observed (Figure 6C), suggesting that other factors might be regulating its expression under arsenic stress.
The yap8 mutant absorbs increased levels of As(III)
As Yap8 regulates As(III) detoxification, we hypothesized that the strong transcriptional activation of the antioxidant genes in the mutant yap8 was caused by arsenite accumulation. In order to verify this, we used atomic-absorption spectroscopy. After adding 2 mM arsenite to wild-type and yap8 cultures, the residual amounts of the metalloid in the supernatant at the time points indicated in Figure 7 were determined. We also performed measurements in the yap1 mutant in order to evaluate the contribution of Yap1 to arsenite detoxification. As Figure 7 shows, arsenite absorption in S. cerevisiae wild-type cells is very low (around 5%). The yap8 mutant absorbs higher levels of As(III) (8%) than those observed in the wild-type. Increased As(III) absorption in the yap8 mutant may explain why the antioxidant genes are more activated in this strain than in the wild-type (see Figure 6). Surprisingly, yap1 As(III) absorption is slightly decreased compared with the wild-type, suggesting that either the uptake is compromised in this strain, that extrusion is enhanced or that both events are occurring.
It has been accepted that, among the various modes of action for arsenic carcinogenesis in human cells, the oxidative stress relevance is the one that assumes that ROS can directly or indirectly damage DNA and proteins [13,41]. Indeed, the property of the organic arsenicals to inhibit the GSH and thioredoxin reductases, as well as to bind GSH [42,43], perturbs the redox equilibrium of the cytoplasm, leading to the accumulation of ROS. Here we report that Yap1, the major regulator of oxidative stress response in S. cerevisiae, is involved in the removal of ROS generated by arsenic compounds. Our data on lipid peroxidation, together with those on intracellular oxidation, indicate that yap1 and yap8 mutant cells are more oxidized than those of the wild-type strain upon treatment with arsenic compounds (Figure 2). The increased lipid peroxidation levels that we observed upon As(III) exposure is consistent with the notion that arsenite has the ability to release Fe(II) from its complexes with proteins, potentially stimulating the peroxidation of cellular lipids in yeast  (Scheme 1). This may, in turn, give rise to the highly reactive superoxide radical. Moreover, enhanced lipid peroxidation levels are also in agreement with the assumption that its induction by arsenite may be responsible for the toxic effect of this compound in eukaryotic cells . Both yap1 and yap8 strains exhibit high levels of lipid peroxidation and intracellular oxidation; however, yap1 is the only one displaying high contents of oxidized proteins (Figure 3). The antioxidant genes GSH1, SOD1 and TRX2 are induced in the wild-type strain their induction being even higher in the yap8 mutant (see Figures 6A and 6B). The antioxidant machinery is therefore activated by Yap1, which prevents protein damage in these strains.
Redox-inactive toxic metals such as arsenic react with GSH, the reduced form of glutathione, which is the major antioxidant reserve of the cell . We have in fact observed that, during the first 1 h of treatment with arsenate, the levels of GSH are decreased in all strains (Figure 4), with a consequent increase in the oxidized form (GSSG). As GSSG is formed, cells must, in order to maintain redox homoeostasis, induce the generation of GSH. This occurs either via its recycling through the glutathione reductase Glr1 or synthesis de novo, in which the activity of Gsh1 is essential. The transcriptomic analyses fully support the activation of both pathways in the wild-type strain (see Table 1 and Supplementary Tables S1 and S2). We show that genes grouped in functional categories related to sulfur metabolism, sulfur-amino-acid biosynthesis and metabolism, as well as GLR1, are induced when this strain is treated with 2 mM arsenate for 30 min (Table 1). Indeed, it had already been shown that arsenite-exposed cells channel a large part of assimilated sulfur into glutathione biosynthesis . In the case of yap1 the recycling pathway is impaired, since GLR1 is a Yap1 target. Interestingly, even in cells harbouring YAP1 deletion, the GSH levels upon 2 h of exposure to arsenic compounds increased, a finding consistent with the observation that GSH1 mRNA levels were not completely abolished in the yap1 mutant treated with both arsenate and arsenite (Figure 6C). Similarly, under glutathione depletion  and cadmium injury , GSH1 has been shown to be regulated by both Yap1 and Met4. Furthermore, the fact that induction of GSH1 is only partially decreased in yap1 cells treated with the superoxide-anion generator menadione  might suggest that treatment of yeast cells with arsenic compounds also leads to the formation of superoxide radicals, which, in turn, could trigger the Yap1-independent basal transcription of GSH1. Yap1 contributes, therefore, to arsenic stress responses by relieving the deleterious effects of ROS in the cells (Scheme 1) through at least the transcriptional activation of antioxidant enzymes encoded by genes such as TRX2, GSH1 and SOD1 (Figure 6). Indeed, as suggested by others , the enhanced SOD1 transcription is one of the most important factors responsible for triggering the adaptation process to mitigate the toxic effect of arsenite in eukaryotic cells. Genes involved in protein folding are also induced by arsenate stress, a phenomenon that is common to many forms of stress (see Table 1 and Supplementary Tables S1 and S2).
The absence of a significant increase in intracellular oxidation, lipid peroxidation and protein carbonylation, as well the changes in the GSSG/GSH ratio, in the wild-type cells reflects the ability of the wild-type strain to counteract the direct and/or indirect deleterious effects of the metal and to adapt to the stress condition (see Scheme 1). Phenotypic assays reveal, indeed, that the wild-type cells are able to grow in the presence of 2 mM arsenate and arsenite (Figure 1). Furthermore, YAP8 and YAP1 are both necessary to trigger the adaptation response, exerting a complementary role, though at different levels. Because YAP8 is the key regulator of arsenic stress responses by controlling the expression of the arsenite efflux protein encoded by ACR3, its absence leads to the accumulation of As(III), enhancing the generation of ROS, as suggested from the atomic-absorption, lipid-peroxidation and intracellular-oxidation assays (see Figures 2, 3 and 7 and Scheme 1).
The YAP1 gene is shown to be essential in preventing protein oxidation (Figure 3). On the other hand, Yap1 also contributes to the regulated expression of ACR2 and ACR3, through recognizing the cis-element GATTAATAATCA positioned in the divergent promoter of these genes (results not shown). It also regulates the expression of the vacuolar pump encoded by YCF1, which composes a parallel arsenite detoxification pathway by catalysing the ATP-driven uptake of As(III)–GSH conjugates into the vacuole [1,49]. It is noteworthy that the yap1 mutant absorbs lower levels of arsenite than does the wild-type strain. Assuming that arsenite vacuolar extrusion mediated by Ycf1 might be at least partially compromised in this mutant, a plausible explanation for the low values observed could be related to an impaired uptake of the metal. It was in fact shown that Hog1 kinase, which in turn modulates the uptake of As(III) dependent on the aquaglyceroporin Fps1, is activated by the metalloid . It is possible that, in the yap1 strain, this process is not fully operating.
In conclusion, we have identified an oxidative pathway dependent on Yap1 involved in the response of S. cerevisiae to arsenic compounds.
We are grateful to Professor Peter Piper (Department of Molecular Biology and Biotechnology, University of Sheffield, Sheffield, U.K.) for providing us with the anti-Sba1 antibody and to Dr. Delmo Santiago Vaitsman [COPPE, UFRJ, Rio de Janeiro, Brazil] for measuring As(III) atomic absorption. This project was financially supported by FCT/POCTI (Fundação para a Ciência e a Tecnologia/Programa Operacional Ciência, Tecnologia e Inovação), to C. R.-P. The plate-forme transcriptome IFR36 is funded by the Reseau Nationale des Génopoles (France). The LIFE Laboratory is supported by CAPES/PROCAD (Coordenação de Aperfeiçoamento do Pessoal de Nível Superior/Programa Nacional de Cooperação Acadêmica) and FAPESP (Fundação de Amparo à Pesquisa do Estado de São Paulo) (grant no. 04-10067/6). R. A. M. and C. A. were supported by FCT fellowships.
Abbreviations: Arr1, arsenical-resistance protein 1; Acr1, arsenic compound resistance protein 1; DAPI, 4′,6-diamidino-2-phenylindole; DABCO, 1,4-diazabicyclo[2.2.2]octane; DNPH, 2,4-dinitrophenylhydrazine; DCF-DA, 2′,7′-dichlorofluorescein diacetate; FCT, Fundação para a Ciência e Tecnologia; GFP, green fluorescent protein; GSH1, γ-glutamylcysteine synthetase gene; LIFE, Laboratório de Investigação de Fatores de Estresse; MDA, malondialdehyde; ROS, reactive oxygen species; SOD1, superoxide dismutase 1 gene; TBARS, thiobarbituric acid-reactive substances; TRX2, thioredoxin 2 gene; UFRJ, Universidade Federal do Rio de Janeiro; Yap1 and Yap8, yeast AP-1-like transcription factors Yap1 and Yap8; YCF1, yeast cadmium factor 1 gene
- © The Authors Journal compilation © 2008 Biochemical Society