Cooperation of Yeast Peroxiredoxins Tsa1p and Tsa2p in the Cellular Defense against Oxidative and Nitrosative Stress*

Peroxiredoxins are a family of antioxidant enzymes conserved from bacteria to humans. In Saccharomyces cerevisiae, there exist five peroxiredoxins, among which Tsa2p shares striking homology with the well described Tsa1p but has not been extensively studied. Here we report on the functional characterization of yeast tsa2Δ mutants and the comparison ofTSA1 with TSA2. The tsa2Δ andtsa1Δ tsa2Δ cells grew normally under aerobic conditions. However, the tsa1Δtsa2Δ mutant yeast was more susceptible to oxidants than either tsa1Δ or tsa2Δ cells. Notably, thetsa1Δ tsa2Δ yeast was also hypersensitive to peroxynitrite and sodium nitroprusside. This phenotype was rescued by the expression of either the TSA1 or TSA2gene. The demonstration of a peroxynitrite reductase activity of Tsa2pin vitro points to a pivotal role for peroxiredoxins in the protection against nitrosative stress. In yeast cells, Tsa1p and Tsa2p exhibited comparable antioxidant activity. While the basal expression level of TSA1 was significantly higher than that ofTSA2, the transcription of TSA2 was stimulated more potently by various oxidants. In addition, TSA2 was activated in tsa1Δ cells in a Yap1p-dependent manner. Taken together, our findings implicate the cooperation of Tsa1p and Tsa2p in the cellular defense against reactive oxygen and nitrogen species.

Living organisms are constantly exposed to reactive oxygen species (ROS) 1 that are produced during metabolism or in response to external stimuli (1). In addition to ROS, reactive nitrogen species (RNS) have emerged as another source of oxidative and nitrosative stress (2,3). Both ROS and RNS have been implicated in various physiological and pathological processes including metabolism, immunity, inflammation, cell signaling, transcriptional regulation, and apoptosis (1)(2)(3). The cellular defense against oxidative and nitrosative stress is important for homeostasis and survival.
Antioxidant enzymes are important components of the cellular defense system against ROS and RNS. In addition to well documented antioxidant enzymes such as superoxide dismutase and catalase, a novel family of peroxidases, designated peroxiredoxins, has recently been characterized (4 -6). Peroxiredoxins are found in all organisms ranging from bacteria to humans. They are thought to be active peroxidases supported by thioredoxin and other electron donors (5,7). The amino acid sequences around the peroxidatic center characterized by a cysteine residue are highly conserved. The oxidation of the cysteine induced the formation of a decameric structure comprising five dimers (8,9). In a more recent study, bacterial peroxiredoxin AhpC has been shown to be peroxynitrite reductase (10), thus conferring resistance to RNS (11). It remains to be seen whether eukaryotic peroxiredoxins can generally scavenge peroxynitrite in addition to hydrogen peroxide and directly protect cells from RNS.
To date, only a limited number of enzymes in the large family of peroxiredoxins have been characterized for function. Fundamental questions as to whether and how peroxiredoxins scavenge RNS and ROS remain unanswered. Coordinated efforts using different biological systems are necessary for functional studies. One route toward understanding the physiology of peroxiredoxins is through the phenotype of null mutants in yeasts. Among the yeast peroxiredoxins, Tsa2p is highly homologous to the well described Tsa1p (5,7). However, Tsa2p has been shown to be very different from Tsa1p and other peroxiredoxins in yeast. Surprisingly, TSA2-null mutants suffered a severe growth retardation characterized by the accumulation of G 1 cells and were insensitive to oxidants (16). To shed additional light on the physiological functions of Tsa2p, we constructed and characterized tsa2⌬ and tsa1⌬ tsa2⌬ mutant yeasts. We also compared the function and regulation of TSA1 and TSA2. The tsa2⌬ and tsa1⌬ tsa2⌬ cells were indistinguishable from wild type under aerobic conditions. However, the tsa1⌬ tsa2⌬ mutant yeast grew more slowly in the presence of ROS and RNS. The antioxidant properties of Tsa1p and Tsa2p were comparable. The basal expression level of TSA1 was significantly higher, but the transcription of TSA2 was more potently activated in response to ROS or RNS and to the loss of Tsa1p protein. We also provide the first evidence for the peroxynitrite reductase activity of Tsa2p. Our findings suggest that Tsa2p cooperated with Tsa1p to protect the yeast cells from oxidative and nitrosative stress.
Construction of tsa1⌬ and tsa2⌬ Mutants-DNA fragments containing TSA1 or TSA2 open reading frame were PCR-amplified from S. cerevisiae genomic DNA (Novagen). The oligonucleotide primers are 5Ј-CGGAATTCGGTTGGCAAAGTCGGCTG (forward) and 5Ј-CCGCTC-GAGAC GGTAGTGTATCCTT (reverse) for TSA1 (expected product size: 869 bp) and 5Ј-CGGAATTCGGGCGAGGCTCTCCTTTCT (forward) and 5Ј-CCGCTCGAGCTTATCATATAT (reverse) for TSA2 (product size: 906 bp), respectively. These primers introduce EcoRI (forward) and XhoI (reverse) sites (underlined). The resulting PCR fragments were gel-purified, digested with EcoRI and XhoI, and cloned into plasmid pBluescript II SK (Stratagene). The ϩ52 to ϩ447 nucleotides from the TSA1 open reading frame were excised with SalI and HincII and replaced with the HIS3 gene from pDG201 (a gift from L. Derr). The ϩ54 to ϩ452 sequences from the TSA2 open reading frame were removed with HincII and HindIII and substituted with the LEU2 gene from plasmid pGAD424 (CLONTECH). The tsa1⌬ disruption strains were obtained by allele replacement using a one-step displacement method (30). The DNA fragment, which contains the HIS3 selectable marker flanked by the TSA1 sequence, was transformed into different strains. The tsa2⌬ disruption strains were obtained by the same approach, except that the DNA fragment containing the LEU2 selectable marker flanked by the TSA2 sequence was used. The tsa1::HIS3 and tsa2::LEU2 genotypes were verified by PCR and Southern analysis.
Construction of TSA1 and TSA2 Expression Plasmids-To construct inducible expression vectors for TSA1 and TSA2, DNA fragments harboring TSA1 and TSA2 gene were cloned into the HindIII and XhoI sites of yeast expression vector pYEUra3 (CLONTECH). To construct pTSA1 and pTSA2 plasmids, which express TSA1 and TSA2 under the control of their own promoters, DNA fragments containing, respectively, the TSA1 and TSA2 genes with ϳ1-kb promoter sequences were cloned into yeast centromere plasmid pRS413 (Stratagene). These fragments were amplified from yeast genomic DNA using primers 5Ј-CGG-AATTCATGAATCCAATTCATT (forward) and 5Ј-CCGCTCGAGACGG-TAGTGTATCCTT (reverse) for TSA1 and primers 5Ј-CAGAATTCATT-GAATGGGCGCA (forward) and 5Ј-CCGCTCGAGCTTATGATATAT (reverse) for TSA2. The primers introduce EcoRI (forward) and XhoI (reverse) sites (underlined). To achieve TSA1 promoter-driven expression of the TSA2 gene and TSA2 promoter-driven expression of the TSA1 gene, the fragments containing TSA1 and TSA2 promoters were swapped between the two constructs pTSA1 and pTSA2 to give plasmids pTSA1-2 and pTSA2-1.
Southern Blot Analyses-Chromosomal DNA was isolated from yeast cells as previously described (31). Southern blot analyses were performed with 10 g of chromosomal DNA and 32 P-labeled probes.
Northern Blot Analyses and RT-PCR-Total RNA was isolated from yeast cells as described previously (31). Northern blotting was performed with total RNA (10 g) and 32 P-labeled probes. For RT-PCR, cDNA was synthesized with the Advantage® RT-for-PCR kit (CLON-TECH). The ORC5-specific primers are 5Ј-TGTGACCACTCCGGAAG and 5Ј-GGAATGTT GGATGTGAA (expected size of product: 264 bp). The TSA1-specific primers are 5Ј-CTTGGACAAATACAAG and 5Ј-CAAAGAGTGGTTGGTG (expected size of product: 247 bp). The TSA2specific primers are 5Ј-ACTGGAAAAGTATAAA and 5Ј-ATAAGGAAT-GATTCTTA (expected size of product: 247 bp). The amounts of the templates were adjusted, and the conditions were optimized to ensure that the amplification was in the linear range.
Construction of TSA1-lacZ and TSA2-lacZ Fusions-The TSA1-lacZ plasmid contains the Ϫ1000 to ϩ50 sequences of TSA1 fused to bacterial lacZ gene. Likewise, the TSA2-lacZ fusion plasmid carries the Ϫ1000 to ϩ50 sequences of TSA2. The TSA1 and TSA2 fragments containing the promoter and the first few codons of the open reading frame were amplified from S. cerevisiae genomic DNA (Novagen) using Pfx DNA polymerase (Invitrogen). The oligonucleotide primers are 5Ј-CCGAGCTCTCGTCAAAGACACCGTC (forward) and 5Ј-CGGAATTC-ATCAATTCCAATCATT (reverse) for TSA1 and 5Ј-ACGCGTCGACA-TTCATTGAATGGGCGCAAT (forward) and 5Ј-CGGAATTCACTACGG-CGGTTTTC (reverse) for TSA2, respectively. The restriction sites for subcloning are underlined. The fragments were cloned first into pBluescript II SK and then inserted into the lacZ plasmid pEB39 (a gift from E. A. Elion). For insertion of the lacZ gene into the chromosomal TSA1 and TSA2 loci, the TSA1-lacZ and TSA2-lacZ plasmids were digested with HindIII and ClaI. The DNA fragments harboring the TSA1 and TSA2 promoters were subcloned into pLacZi (CLONTECH). The resulting plasmids were digested by MunI and BglII and integrated into the yeast genomic DNA by homologous recombination. In the TSA1-lacZ mutant yeast, the lacZ gene was driven by the authentic TSA1 promoter, and the TSA1 gene was under the control of another copy of the previously cloned TSA1 promoter (ϳ1 kb). Likewise, in the TSA2-lacZ yeast, the lacZ gene was driven by the authentic TSA2 promoter, and the TSA2 gene was under the control of another copy of the previously cloned TSA2 promoter (ϳ1 kb). The sequences of all PCR products were confirmed by DNA sequencing, and no mutation had been introduced.
ROS Detection-Intracellular redox levels were measured by fluorescence microscopy or by fluorimetry using the fluorescent dye 2Ј,7Јdichlorofluorescein diacetate. Cells were grown in YPD medium until A 600 reached 0.5. Hydrogen peroxide was added to a final concentration of 1 mM. After an additional incubation for 15 min, cells were collected by centrifugation from 2 ml of culture and then washed three times with phosphate-buffered saline (PBS). Cells were resuspended in PBS with 10 M 2Ј, 7Ј-dichlorofluorescein diacetate (Molecular Probes, Inc., Eugene, OR) and incubated at 28°C for 1 h. The dye can react specifically with hydrogen peroxide to give a highly fluorescent 2Ј,7Ј-dichlorofluorescein (DCF). Cells were collected, washed three times with PBS, mounted onto slides, and examined under a confocal fluorescence microscope (Zeiss). An argon ion laser with an emission line at 488 nm was used to excite DCF. Alternatively, the cells were disrupted by glass beads, and the supernatant was collected by centrifugation. The crude extract containing 500 g of protein was suspended in PBS, and the fluorescence was measured on a F-4500 spectrofluorimeter (Hitachi). The excitation and emission wavelengths were 488 and 520 nm, respectively.
Spot Assays for Sensitivity to Oxidants-Cells were grown in YPD medium until A 600 reached 0.5 and exposed to oxidants for 0.5 h. Cells were then diluted and plated for colony survival.
␤-Galactosidase Assay-␤-Galactosidase levels were measured with treated or untreated midlog phase cells as previously described (32,33). o-Nitrophenyl-␤-D-galactopyranoside was used as substrate. Activities were given in A 420 units/min/mg of protein.
Flow Cytometric Analysis of DNA Content-After two washes with water, cells were fixed with 70% ethanol for 12 h at 4°C and then treated with RNase A (1 mg/ml) in PBS for 30 min at 37°C. Then proteinase K (10 mg/ml) was added, and the cells were further incubated for 1 h at 55°C. Cells were stained with propidium iodide (50 g/ml) in 10 mM Tris-HCl (pH 8.0), 10 mM NaCl. Stained cells were subsequently diluted in PBS, and for each sample the DNA content in 10,000 cells was determined with a FACScan flow cytometer. Flow cytometric analysis was performed with EXPO program (EPICS).
Expression and Purification of Histidine-tagged Tsa2p-A DNA fragment containing TSA2 gene was prepared by PCR as described above and ligated into expression vector pET-28a (Novagen). Histidine-tagged Tsa2p was expressed in E. coli BL21 (DE3) and purified using procedures recommended by Novagen. The purified His-Tsa2p was reduced by 5 mM DTT. The reduced His-TSA2p was then chromatographed through a Fast-Flow desalting column (Amersham Biosciences, Inc.) to remove DTT.
Peroxynitrite Reductase Assay-The peroxynitrite-mediated oxidation of dihydrorhodamine 123 to rhodamine was followed as previously  (10). The final concentrations of peroxynitrite and dihydrorhodamine 123 were 10 and 100 M, respectively, in potassium phosphate buffer (pH 7.0) with 100 M diethylenetriaminepentaacetic acid. The rhodamine formation was quantitated by measurement of absorbance at a wavelength of 500 nm. Oxidized His-Tsa2p was prepared by the addition of 5 mM H 2 O 2 .

Molecular Evolution of Yeast
Tsa1p and Tsa2p-The yeast S. cerevisiae is a useful model for studies of eukaryotic antioxidant enzymes and redox signaling (1). Budding yeast Tsa1p is the first identified peroxiredoxin (5,34) and is one of the best-studied members in this family (5, 7, 16, 34 -36). Among the five peroxiredoxins in yeast, Tsa2p is unique in sharing striking homology with Tsa1p. Notably, 86% of the amino acid residues in Tsa1p and Tsa2p are identical, and 96% are similar. The separation of TSA2 with TSA1 probably occurred after the speciation of budding yeast, as the result of a gene duplication event. The high degree of sequence homology suggests a conservation of function. Studies of yeast Tsa1p and Tsa2p may derive novel insights into the biology of closely related peroxiredoxins in other species.
Growth Phenotypes of the ⌬tsa2 Strains-The five mutants deleted for individual peroxiredoxins have been shown to be viable (16,35). This raises two not mutually exclusive possibilities that may explain the phenotype. First, there could be functional overlap between different peroxiredoxins. Second, all peroxiredoxins might be dispensable for viability because of functional overlap between peroxiredoxins and other antioxidant enzymes. One way to explore these possibilities is through the construction of multiple peroxiredoxin-null mutants.
We also noted that the TSA2-null strain exhibited a unique phenotype characterized by severe growth retardation with the accumulation of G 1 cells (16). In addition, unlike other peroxiredoxin-null mutants, the yeast strain deleted for TSA2 was insensitive to oxidant challenge (16). A closer examination of the tsa2⌬ phenotype is required to further elucidate the underlying mechanisms.
To address the above issues and to characterize the cellular functions of yeast Tsa1p and Tsa2p, we constructed tsa2⌬ and tsa1⌬ tsa2⌬ strains by replacing an internal fragment within the coding region of TSA2 with the LEU2 marker. PCR (Fig.  1A) and Southern blotting (Fig. 1B) were performed to confirm the insertion of LEU2 into the TSA2 gene. In addition, no TSA2 mRNA was detected in tsa2⌬ or tsa1⌬ tsa2⌬ cells by Northern blotting (Fig. 1C). All three experiments consistently demonstrated the disruption of the TSA2 gene in the tsa2⌬ and tsa1⌬ tsa2⌬ strains.
The tsa1⌬, tsa2⌬, and tsa1⌬ tsa2⌬ strains were all viable, indicating that neither TSA1 nor TSA2 is essential for normal aerobic growth. In addition, both the tsa2⌬ and the tsa1⌬ tsa2⌬ strains showed wild-type growth rate in rich (YPD; Fig. 2A) or minimal (SD; data not shown) medium. The tsa2⌬ and tsa1⌬ tsa2⌬ mutants were indistinguishable from the isogenic wildtype BY4741 strain either in growth ( Fig. 2A) or in cell morphology (data not shown). Next, we performed flow cytometric analysis to assess the DNA content and to compare the cell cycle profiles of the wild-type and mutant yeasts (Fig. 2B). Again, the tsa2⌬ (panel 3) and tsa1⌬ tsa2⌬ (panel 4) strains did not show any difference from the BY4741 (panel 1) or tsa1⌬ (panel 2) cells. Notably, the distribution profiles of G 1 /S and G 2 /M cells in the four strains were very similar. In sharp contrast to a previous report (16), the G 1 peak in either the tsa2⌬ yeast or the tsa1⌬ tsa2⌬ double mutant is not higher than in the BY4741 or tsa1⌬ strain (i.e. we did not observe the accumulation of G 1 cells in the tsa2⌬ strains).
Sensitivity of tsa2⌬ Strains to Oxidants-The exponentially growing TSA2-null yeast cells have been shown to be insensitive to oxidants (16). Because the behaviors of our tsa2⌬ strains were different from the reported slow growth phenotype, we sought to reexamine the sensitivity of our tsa2⌬ strains to oxidants and the antioxidant response in these strains. First we challenged the four yeast strains (wild-type BY4741, tsa1⌬, tsa2⌬, and tsa1⌬ tsa2⌬) with H 2 O 2 and tbutylhydroperoxide (t-BHP). The sensitivity was assessed in the spot assay (Fig. 3A). Consistent with our findings from growth rate studies ( Fig. 2A) and flow cytometric analysis (Fig.  2B), the four strains grew normally in the absence of oxidant insult (Fig. 3A, lanes 1-3). However, the tsa2⌬ cells were more sensitive to peroxides than the wild type BY4741 but less sensitive than the tsa1⌬ cells (Fig. 3A, lanes 6 and 9). This implicates that Tsa2p plays a significant, albeit secondary, role in the cellular response to oxidants. Among the four isogenic strains, the double mutant tsa1⌬ tsa2⌬ is most sensitive to oxidant challenge, suggesting that Tsa1p and Tsa2p cooperate in the antioxidant defense.
The antioxidant properties of peroxiredoxins to protect cells from oxidant insult are ascribed to their ability to scavenge hydrogen peroxide (7). To follow more closely the removal of H 2 O 2 , we measured the levels of intracellular H 2 O 2 in the four strains by confocal fluorescence microscopy using the fluorescent probe 2Ј,7Ј-dichlorofluorescein diacetate (Fig. 3B). This dye has been widely used for ROS detection (17), and it reacts specifically with H 2 O 2 to produce a highly fluorescent DCF. We observed that the DCF fluorescence in all three mutants (Fig.  3B, panels 4, 6, and 8) significantly increased compared with that in the parental wild-type strain BY4741 (panel 2). The relative intensity of the fluorescence in the four strains was in the following order: tsa1⌬ tsa2⌬ Ͼ tsa1⌬ Ͼ tsa2⌬ Ͼ BY4741 (Fig. 3B). This order is generally consistent with the sequence of sensitivity to oxidants (Fig. 3A).
Sensitivity of tsa1⌬ and tsa2⌬ Strains to RNS-Bacterial peroxiredoxin AhpC has peroxynitrite reductase activity (10) and therefore confers resistance to RNS (11). Yeast Tsa1p and Tsa2p are in the same subfamily as AhpC. However, it remains to be seen whether Tsa1p, Tsa2p, and eukaryotic peroxiredoxins can protect cells from RNS. To address this question, we   FIG. 1. Disruption of TSA2 in tsa2⌬ and tsa1⌬ tsa2⌬ strains. A, PCR analysis. Targeted and isogenic wild-type TSA2 fragments in yeast were detected by PCR with specific primers. PCR products of 0.9 and 2.3 kb in size correspond to TSA2 and tsa2::LEU2, respectively. DNAs amplified from the indicated strains (lanes 3-6) were compared with the positive (ϩve control; lane 7) and negative (Ϫve control; lane 2) control DNAs amplified from plasmids. B, Southern blot analysis. The 2.5-and 4.0-kb fragments correspond to TSA2 and tsa2::LEU2, respectively. C, Northern blot analysis. The 0.9-kb mRNA is specific for TSA2. As a control for equal loading, the staining with ethidium bromide is shown in the lower panel.
tested the sensitivity of TSA1-and/or TSA2-null strains to peroxynitrite and sodium nitroprusside (SNP).
As a product of nitric oxide and superoxide, peroxynitrite (ONOO Ϫ ) is a potent oxidizing and nitrating species with mutagenic, proapoptotic, and cytotoxic activities. Peroxynitrite has been implicated in the pathogenesis of various diseases (37). SNP is a nitric oxide donor frequently used in biomedical research and in clinical practice (37). When the four yeast strains were exposed to 1 mM peroxynitrite, the tsa1⌬, tsa2⌬, and tsa1⌬ tsa2⌬ strains were significantly more susceptible than the wild-type BY4741 (Fig. 4A). Similarly, the tsa1⌬ yeast was more sensitive to 2 mM SNP than BY4741. While no difference in the sensitivity of the tsa2⌬ and the parental BY4741 strains to SNP was noted, the tsa1⌬ tsa2⌬ double mutant displayed an increased sensitivity compared with the BY4741 and tsa1⌬ strains (Fig. 4B). Collectively, these results support a model in which Tsa1p cooperates with Tsa2p in the protection against RNS. To verify the specificity of action, we performed complementation assays with TSA1 or TSA2. A centromere expression plasmid for TSA1 was transformed into the tsa1⌬, tsa2⌬, and tsa1⌬ tsa2⌬ strains (Fig. 4C). The growth rates of the four strains were indistinguishable, indicating that the expression of TSA1 driven by the TSA1 promoter fully complemented the deficiency of TSA1 and/or TSA2. In contrast, the expression of TSA2 driven by the TSA2 promoter complemented the defects in TSA1 and/or TSA2 partially (Fig. 4D). Conceivably, the incompleteness of the effect may arise from the inefficient expression and/or the lower antioxidant activity of the protein.
To shed additional light on this, we swapped the TSA1 and TSA2 promoters and constructed plasmids pTSA1-2 and pTSA2-1. The expression of Tsa2p from plasmid pTSA1-2 was regulated by the TSA1 promoter. Vice versa, the TSA2 promoter was used to drive the expression of Tsa1p from pTSA2-1. Interestingly, the tsa1⌬ tsa2⌬ strain harboring the pTSA1-2 plasmid (Fig. 4E, curve 5) was as resistant to SNP as the BY4741 yeast (curve 1), indicating that the expression of Tsa2p alone sufficiently protected cells from SNP challenge. Likewise, the pTSA2-1 plasmid conferred substantial but not full protection against SNP (Fig. 4E, curve 6). One interpretation for the partial effect is that the TSA2 promoter is less potent than the TSA1 promoter. This hypothesis is supported further by the facts that both TSA1 (Fig. 4C) and TSA1-2 (Fig. 4E, curve 5) are fully competent in the defense against SNP.
Peroxynitrite Reductase Activity of Tsa2p-In light of the ability of Tsa1p and Tsa2p to protect cells against RNS (Fig. 4), we asked whether Tsa2p might catalytically detoxify peroxynitrite. One efficient and selective method to detect peroxynitrite is through the oxidation of dihydrorhodamine 123 to rhodamine 123 (10). We expressed His-tagged Tsa2p protein in E. coli and purified it to Ͼ90% homogeneity as assessed on nonreducing and reducing PAGE gel (Fig. 5A). We incubated the purified His-Tsa2p with 5 mM DTT and then removed DTT by gel filtration. When this preparation of reduced His-Tsa2p was added to the reaction containing peroxynitrite and dihydrorhodamine 123, we observed a pronounced inhibition of rhodamine 123 formation (Fig. 5B, curve 3). Notably, neither bovine serum albumin treated with DTT in the same way as His-Tsa2p (Fig.  5B, curve 1) nor oxidized His-Tsa2p preincubated with 5 mM H 2 O 2 (curve 2) had a significant effect on peroxynitrite-mediated oxidation of dihydrorhodamine 123. Moreover, we demonstrated that the oxidation of His-Tsa2p by peroxynitrite (Fig.  5C, lanes 1 and 2) 4. Sensitivity of tsa2⌬ mutants to peroxynitrite and SNP. A, sensitivity to peroxynitrite. Cells were grown to saturation, diluted to an A 600 of 0.2, and then treated with 1 mM peroxynitrite. B, sensitivity to SNP. Cells were grown to saturation, diluted to an A 600 of 0.2, and then treated with 2 mM SNP. C, sensitivity of TSA1-transformed mutants to SNP. The tsa1⌬, tsa2⌬, and tsa1⌬ tsa2⌬ strains (curves 2-4) were transformed with a TSA1 expression plasmid driven by the TSA1 promoter. The transformants were compared with the mock-transformed BY4741 strain (curve 1). D, sensitivity of TSA2-transformed mutants to SNP. The tsa1⌬, tsa2⌬, and tsa1⌬ tsa2⌬ strains (curves 2-4) were transformed with a TSA2 expression plasmid driven by the TSA2 promoter. The transformants were compared with the mock-transformed BY4741 strain (curve 1). E, expression of Tsa1p or Tsa2p rescues the sensitivity to SNP. The tsa1⌬ tsa2⌬ strain was transformed individually with either pTSA1-2 (curve 5) or pTSA2-1 (curve 6). The expression of Tsa2p from pTSA1-2 was driven by a TSA1 promoter (Ϫ1000 to ϩ52 nucleotides). In pTSA2-1, a TSA2 promoter (Ϫ1000 to ϩ52 nucleotides) was used to control the expression of Tsa1p.
6-fold higher than that of Tsa2p (16). Above we showed that the expression of either Tsa1p or Tsa2p in yeasts sufficiently protected cells from ROS and RNS. To further characterize the antioxidant activities of Tsa1p and Tsa2p, we compared their H 2 O 2 -removing activities in vivo.
For this experiment, we constructed inducible expression plasmids for Tsa1p and Tsa2p. A GAL1 promoter was used to control the expression of Tsa1p and Tsa2p in plasmids pTSA1 and pTSA2, respectively. The tsa1⌬ tsa2⌬ double mutant was transformed individually with the empty vector, pTSA1, and pTSA2. The expression of Tsa1p and Tsa2p was induced by transferring the yeasts to a medium containing galactose. The cells were treated with H 2 O 2 , and the fluorescent dye 2Ј,7Јdichlorofluorescein diacetate was added to chase the removal of H 2 O 2 . The DCF fluorescence reflects the relative levels of intracellular ROS. From the representative fields of cells under the fluorescence microscope (Fig. 6A) and from the quantitation based on a fluorimeter (Fig. 6B), the expression of either Tsa1p or Tsa2p led to a substantial reduction of DCF fluorescence, which reflects the removal of intracellular H 2 O 2 . Tsa1p appeared to be a more potent peroxidase in this assay. For all, the H 2 O 2 -scavenging activities of the two peroxiredoxins were comparable. These data implicate that both Tsa1p and Tsa2p are active peroxidases in vivo.
Comparison of the Basal Transcriptional Activities of TSA1 and TSA2-The above data suggested that the TSA1 promoter might be stronger than the TSA2 promoter (Fig. 5, C-E). To formally compare their basal activities, we performed lacZ reporter assays. Reporter plasmids (2-based) driven by the TSA1 and TSA2 promoters (TSA1-lacZ and TSA2-lacZ) were transformed into the BY4741 strain. The ␤-galactosidase activity was assayed and compared. In this assay, the TSA1 promoter is about 3 times stronger than the TSA2 promoter (Fig.  7). Results from the semiquantitative RT-PCR analysis of the TSA1 and TSA2 transcripts in untransformed BY4741 cells lent further support to the notion that the basal transcriptional level of TSA1 is significantly higher (Fig. 7, inset).
Differential Regulation of TSA1 and TSA2 by ROS and RNS-While the basal expression levels of TSA1 and TSA2 were different (Fig. 7), both genes have been shown to be induced by H 2 O 2 and diamide (16). To investigate the transcriptional regulation of the chromosomal TSA1 and TSA2 loci, we inserted the lacZ reporter immediately downstream of the chromosomal TSA1 and TSA2 promoters. The expression of the Tsa1p/Tsa2p was rescued by simultaneously introducing an extra copy of the TSA1/TSA2 promoter immediately upstream of the coding region. In this setting, the integrated single copy lacZ reporter may better reflect the transcriptional activities of the chromosomal TSA1 and TSA2 genes.
We compared the relative ␤-galactosidase activities of the TSA1-lacZ and TSA2-lacZ strains in the presence of H 2 O 2 , t-BHP, diamide, peroxynitrite, and SNP (Fig. 8). Interestingly, the activities of the TSA1 promoter did not change substantially in response to ROS or RNS (Fig. 8A). In most cases, the increase in transcriptional level was less than 50%. The stimulation by t-BHP was less than 2-fold. By sharp contrast, the induction of TSA2 promoter by H 2 O 2 , t-BHP, diamide, and peroxynitrite was much more dramatic, ranging from 5-to 11-fold (Fig. 8B). Similar results were obtained from strains transformed with 2-based TSA1-lacZ and TSA2-lacZ plasmids (data not shown). These data provide the evidence for differential regulation of TSA1 and TSA2 genes in response to ROS and RNS.
Compensational Activation of TSA2 in tsa1⌬ Strain-To better understand the functional overlap between Tsa1p and Tsa2p, we asked whether TSA1 is activated in the tsa2⌬ strain, and vice versa. We observed that the activity of TSA1-lacZ was only slightly increased in the tsa2⌬ strain either in the absence or in the presence of H 2 O 2 (Fig. 9A). The TSA1-lacZ activity was substantially reduced in the yap1⌬ tsa2⌬ strain, implying that the basal activation of TSA1 is mediated through the redox-regulated transcription factor Yap1p. In contrast, the TSA2-lacZ activity significantly increased in the tsa1⌬ strain (Fig. 9B). We also noted that the TSA2-lacZ activity was lost almost completely in the yap1⌬ tsa1⌬ strain. One interpretation is that the Yap1p is responsible for both the basal and the induced activation of TSA2.
To verify that the lacZ reporter activity reflects the authentic TSA1 and TSA2 genes, we performed semiquantitative RT-PCR to compare the relative amounts of TSA1 and TSA2 mRNA in BY4741, tsa1⌬, and tsa2⌬ strains (Fig. 9C). Consistent with the results from the reporter assay, the TSA2 mRNA was more abundantly expressed in the tsa1⌬ strain (compare lane 2 with lane 1 and lane 5 with lane 4). This compensational activation suggests that TSA2 may play a more important role when TSA1 is compromised. DISCUSSION In this study, we used a genetic approach (Fig. 1) to characterize the function and regulation of yeast peroxiredoxins Tsa1p and Tsa2p. The tsa2⌬ and tsa1⌬ tsa2⌬ yeast strains were viable and indistinguishable from the parental wild type under aerobic conditions (Fig. 2). However, the disruption of TSA1 and/or TSA2 conferred hypersensitivity to RNS (Fig. 4) in addition to ROS (Fig. 3). In line with this, Tsa2p acts as a peroxynitrite reductase to protect against peroxynitrite-mediated oxidation in vitro (Fig. 5). The in vivo H 2 O 2 -scavenging activities of Tsa1p and Tsa2p were comparable (Fig. 6), but the basal expression level of TSA1 was significantly higher (Fig. 7).
While the transcription of TSA2 was potently activated in response to ROS/RNS (Fig. 8B) and as a result of TSA1 disruption (Fig. 9), the expression of TSA1 was less responsive to stimuli (Fig. 8A). Our findings support the model in which Tsa2p cooperated with the primary cytoplasmic peroxiredoxin Tsa1p in the cellular defense against oxidative and nitrosative stress.
Tsa2p Is a Functional Antioxidant Enzyme in Yeast-Tsa2p is closely related to the well described Tsa1p. The separation of these two peroxiredoxins probably arose from a gene duplication event after the speciation of budding yeast. A previous study (16) has described two unique characteristics of a TSA2null mutant: the insensitivity to oxidants and the induction of growth retardation presented as G 1 arrest. In addition, Tsa2p has a low thioredoxin peroxidase activity in vitro (16). In the present work, we did not observe the slow growth phenotype in tsa2⌬ strains (Fig. 2). We wondered how the particular genetic background of the tsa2⌬ strain used in Ref. 16 or changes other than the loss of TSA2 might explain the different observations. We also presented several lines of evidence to support the notion that Tsa2p is a functional antioxidant enzyme in vivo. First, the tsa2⌬ and tsa1⌬ tsa2⌬ strains are more sensitive to ROS and RNS than the parental BY4741 and tsa1⌬ strains, respectively (Figs. 3 and 4). Second, the expression of TSA2 driven by different promoters can partially or fully rescue the hypersensitivity to SNP caused by disruption of TSA1 and/or TSA2 (Fig. 4, D and E). Third, Tsa2p acts as a peroxynitrite reductase in vitro (Fig. 5). Fourth, the overexpression of TSA2 under the control of the inducible GAL1 promoter can effectively scavenge H 2 O 2 in the tsa1⌬ tsa2⌬ cells (Fig. 6). Tsa2p appears to be a more active peroxidase in vivo (Fig. 6) than in vitro (16). It remains unanswered whether electron donors other than thioredoxin can support the peroxidase and peroxynitrite reductase activities of Tsa2p and other peroxiredoxins. In this regard, cyclophilin A has recently been shown as a binding partner as well as peroxidase activator of mammalian peroxiredoxins (38). It would be of interest to see whether yeast cyclophilins might serve similar functions to support the antioxidant activities of peroxiredoxins. Last but not least, the FIG. 6. Peroxidase activity of Tsa1p and Tsa2p in vivo. A, DCF oxidation. The tsa1⌬ tsa2⌬ yeasts were transformed individually with empty vector (pYEUra3 from CLONTECH), with pTSA1, or with pTSA2. The expression of Tsa1p and Tsa2p from pTSA1 and pTSA2, respectively, was driven by an inducible GAL1 promoter. The yeasts were initially grown in SDϪUra with 2% glucose. To induce the expression of Tsa1p and Tsa2p, cells were transferred to YPG medium containing 2% galactose, cultured for an additional 2 h, treated first with 1 mM H 2 O 2 and then with 10 M 2Ј, 7Ј-dichlorofluorescein diacetate, and examined under a confocal fluorescence microscope. The DCF fluorescence (panels 1-3) and the light fields (panels 4 -6) for the same fields of cells (panels 1 and 4, 2 and 5, and 3 and 6) were shown at ϫ100 magnification. B, quantitation of DCF fluorescence. Cells were disrupted by glass beads, and the fluorescence was measured on an F-4500 spectrofluorimeter. The fluorescence intensity of the tsa1⌬ tsa2⌬ cells transformed with an empty vector was taken as 100%. Results represent the average of three independent experiments, and error bars indicate S.E. RT-PCR was performed to verify the expression of TSA1 and TSA2 genes (inset). The housekeeping gene ORC5 was used as a control.
FIG. 7. The basal expression levels of TSA1 and TSA2 genes. The BY4741 yeast was transformed individually with pTSA1-lacZ and pTSA2-lacZ. Cells were cultured in YPD medium to log phase, and ␤-galactosidase activity was measured. Results represent three independent experiments, and error bars indicate S.E. Semiquantitative RT-PCR was performed to compare the mRNA expression levels in the untransformed BY4741 strain (inset). The housekeeping gene ORC5 was used as a control. The expected sizes of the ORC5, TSA1, and TSA2 amplification products are similar (264, 247, and 247 bp). expression of TSA2 was stimulated potently by ROS and RNS (Fig. 8). Our findings argue for an important role of Tsa2p in the antioxidant defense.
Tsa1p and Tsa2p Protect Cells against RNS-The disruption of TSA1 and/or TSA2 conferred susceptibility to peroxynitrite and SNP (Fig. 4, A and B). In addition, the expression of either TSA1 or TSA2 sufficiently reversed this phenotype (Fig. 5,  C-E). Thus, we provide the first evidence that eukaryotic peroxiredoxins conferred resistance to RNS. This appears to be a biological function conserved in both prokaryotic and eukaryotic peroxiredoxins (11). In support of this, we demonstrate for the first time the peroxynitrite reductase activity of Tsa2p in vitro (Fig. 5). Thus, bacterial peroxiredoxin AphC (10), yeast Tsa2p (this study), and bovine peroxiredoxin VI (also known as 1-Cys peroxiredoxin; Ref. 39) can act as peroxynitrite reductase to directly protect cells against peroxynitrite-mediated oxidations. Since these three peroxiredoxins are from different species and different subfamilies, it is tempting to assume that most if not all peroxiredoxins might conserve the same property in the cellular defense against RNS.
Differential Expression and Cooperation of Tsa1p and Tsa2p-TSA1 and TSA2 are differentially regulated in yeasts. On one hand, the basal transcription level of TSA1 is significantly higher (Fig. 7). This implicates that Tsa1p serves as a primary or principal housekeeping antioxidant enzyme in the cellular defense against ROS and RNS. In this regard, Tsa2p is secondary and functions as an antioxidant enzyme in reserve. This model can explain the relative sensitivity of the tsa1⌬, tsa2⌬, and tsa1⌬ tsa2⌬ strains to ROS and RNS (Figs. 3 and 4).
On the other hand, the transcription of TSA2 is induced substantially in response to ROS, RNS, and the absence of TSA1 (Figs. 8 and 9). This induction indicates that Tsa2p plays a particularly important role in the adaptation to oxidative and nitrosative stress. Our findings are generally consistent with results from several genome-wide proteomic or microarray analyses (40 -42). The differential expression of Tsa1p and Tsa2p suggests that these two peroxiredoxins may fulfill their functions in different phases of the cellular response to stress.
The cooperation between Tsa1p and Tsa2p is supported by three lines of data. First, the tsa1⌬ tsa2⌬ strain is more sensitive to ROS (Fig. 3) and RNS (Fig. 4) than the tsa1⌬ strain. These results support the additive action of Tsa1p and Tsa2p. Second, the expression of TSA2 is activated in response to the loss of TSA1 (Fig. 9). Third, the induced overexpression of TSA2 alone can sufficiently complement the loss of TSA1 (Fig.  4). Taken together, the differentially expressed yeast peroxiredoxins Tsa1p and Tsa2p serve similar functions, and they cooperate with each other in the cellular defense against oxidative and nitrosative stress.