A cysteine-sulfinic acid in peroxiredoxin regulates H₂O₂-sensing by the antioxidant Pap1 pathway

Abstract

The Schizosaccharomyces pombe transcription factor Pap1 regulates antioxidant-gene transcription in response to H₂O₂. Pap1 activation occurs only at low, but not elevated, H₂O₂ concentrations that instead strongly trigger the mitogen-activated protein kinase Sty1 pathway. Here, we identify the peroxiredoxin Tpx1 as the upstream activator of Pap1. We show that, at low H₂O₂ concentrations, this oxidant scavenger can transfer a redox signal to Pap1, whereas higher concentrations of the oxidant inhibit the Tpx1-Pap1 redox relay through the temporal inactivation of Tpx1 by oxidation of its catalytic cysteine to a sulfinic acid. This cysteine modification can be reversed by the sulfiredoxin Srx1, its expression in response to high doses of H₂O₂ strictly depending on active Sty1. Thus, Tpx1 oxidation to the cysteine-sulfinic acid and its reversion by Srx1 constitutes a previously uncharacterized redox switch in H₂O₂ signaling, restricting Pap1 activation within a narrow range of H₂O₂ concentrations.

Oxidative stress responses sensing an excess of hydrogen peroxide (H₂O₂) have been described during the last decade (1–3). In the yeast Schizosaccharomyces pombe, at least two independent pathways respond to increased intracellular concentrations of H₂O₂: the Pap1 and the Sty1/Spc1 pathways. Pap1 is a bZIP transcription factor, homologue of mammalian c-Jun, that, upon activation by H₂O₂, triggers a specific antioxidant gene response (4, 5). In response to H₂O₂, an intramolecular disulfide bond between distant cysteine residues is formed in Pap1, which masks the accessibility of the nuclear exporter Crm1 to the C-terminal nuclear export signal, and results in accumulation of Pap1 in the nucleus and in Pap1-dependent gene expression (6, 7). The mitogen-activated protein kinase (MAPK) Sty1, on the contrary, can be activated by different types of stresses, such as H₂O₂, heat shock, nutritional starvation or osmotic stress (8, 9). The main substrate of Sty1 is the transcription factor Atf1 (10, 11), which binds to a wide number of promoters and up-regulates their gene products in response to stress.

S. pombe seems to have these two distinct pathways to accomplish a combinational response to different concentrations of H₂O₂. Indeed, Pap1 and Sty1 have complementary roles in the oxidative stress response, Sty1 being more important for cell survival after exposure to high levels of H₂O₂ and Pap1 being essential for an adaptation response to low concentrations of the oxidant (12). Consistent with their biological roles, the Sty1 MAPK pathway preferentially responds to high doses of H₂O₂ (1 mM) (12), and Pap1 is oxidized and activated by low H₂O₂ concentrations (0.2 mM) but, at higher concentrations of the oxidant (1 mM), Pap1 activation is delayed by ≈30 min (5, 7, 13) or even longer (70 min at 5 mM; ref. 7). This late activation of Pap1 at high concentrations of H₂O₂ is not observed in cells lacking Sty1 (7), suggesting that this regulator induces the expression of one or more gene products important to restore the cell competence in Pap1 activation. These gene products are neither catalase nor glutathione peroxidase, because they, respectively, do not or are only able to very partially correct the defective Pap1 activation of Δsty1 cells when overexpressed (7).

In Saccharomyces cerevisiae activation of Yap1, the structural and functional homologue of Pap1, is not direct. Instead, the glutathione peroxidase-like enzyme Orp1/Gpx3 acts as the actual sensor of H₂O₂ and transduces this signal into a cysteine-redox cascade culminating into Yap1 oxidation (14). These data raise the question of whether there is an upstream activator of Pap1 in S. pombe and, if so, what would be the response and cysteine redox state of such an H₂O₂ sensor toward differing doses of H₂O₂.

We have analyzed the participation of the three putative peroxidases of S. pombe in H₂O₂-induced Pap1 activation. Our data shows that Tpx1 is an upstream activator of Pap1. As described for most eukaryotic peroxiredoxins (Prxs), Tpx1 suffers an oxidation of one of its cysteine residues to sulfinic acid upon high doses of H₂O₂, which leads to its temporal inactivation. As a result, this protein can transfer only the redox signal to Pap1 under low H₂O₂ stress, and this finding constitutes the basis of the concentration-dependent activation of the Pap1 pathway. Sty1 is required for Tpx1 reactivation, because synthesis of the Tpx1-reductase Srx1 depends on Sty1 upon severe H₂O₂ stress.

Methods

Yeast Strains. We used strains HM123 (h^– leu1), PN513 (h^– leu1 ura4), JA212 (h⁺ leu1), EHH108 (h^– his2 ura4 pap1::ura4^– leu1; ref. 6), EHH14 and mutant derivatives (h^– his2 ura4 pap1::ura4^– leu1 nmt::GFP-pap1::leu1⁺; ref. 6), NT224 (h^– leu1 ura4 sty1–1), AV1 (h^– leu1 ura4 sty1-1 nmt::GFP-pap1::leu⁺; ref. 7), and JM1066 (h^– leu1 ura4 atf1::ura4⁺; ref. 15). To construct S. pombe strains with specific loci deleted, we transformed strains HM123, JA212, or EHH14 with linear fragments containing ORF::kanMX6, obtained by PCR amplification with ORF-specific primers and plasmid pFA6a-kanMX6 as a template (16). We obtained strains AM004 (h^– leu1 pap1::kanMX6), AM001 (h^– leu1 sty1::kanMX6), EA37 (h^– leu1 gpx1::kanMX6), EA38 (h^– leu1 srx1::kanMX6), EA56 (h⁺ leu1 srx1::kanMX6), EA49 (h^– leu1 pmp20::kanMX6), and EA36 (h^– his2 ura4 pap1::ura4^– leu1 nmt::GFP-pap1::leu1⁺ srx1::kanMX6). To isolate cells nonexpressing Tpx1, we transformed diploid strain PN2304 (h⁺/h^– leu1/leu1 ade6-M210/ade6-M216) with a PCR-amplified tpx1::kanMX6 fragment, yielding strain EA40dip. We transformed this strain with a plasmid expressing HA-Tpx1 (p123.41x; see below). Haploid cells were isolated by standard genetic techniques (17) and spread in minimal media. Kanamycin-resistant haploid cells were isolated by replica plating. The yielding strain, EA40hap-p123.41x (h^– tpx1::kanMX6 pHA-tpx1.41x), was confirmed by Northern blot and PCR. A similar strategy was developed to isolate strain AV36hap-p123.42x (h⁺ leu1 tpx1::kanMX6 pHA-tpx1.42x).

Plasmids. The srx1, tpx1, and pap1 coding sequences were PCR amplified from an S. pombe cDNA library by using primers specific for the sulfiredoxin-, Tpx1- and Pap1-coding genes. srx1 was cloned into the nmt (no message in thiamine)-driven expression vector pRep.42x (18) to yield plasmid p120.42x (psrx1.42x). tpx1 was cloned into the S. pombe vectors pRep.41x, pRep.42x, and pRep.81x (18). A DNA fragment coding for the HA epitope was inserted between the nmt promoter and the tpx1 ORFs of those plasmids to yield p123.41x (pHA-tpx1.41x), p123.42x (pHA-tpx1.42x), and p123.81x (pHA-tpx1.81x), respectively. To construct tpx1 mutant alleles, p123.42x was used as a template for PCR reactions by using pairs of mutagenic and complementary primers containing the codon change of interest. The resulting plasmids were named p123.42x.C48S (pHA-tpx1.C48S.42x) and p123.42x.C169S (pHA-tpx1.C169S.42x). The wild-type and mutant ORFs of tpx1 were cloned, fused to the tpx1 promoter in the S. pombe plasmid pJK148 (19), to yield plasmids p145 (ptpx1′::tpx1), p145.C48S (ptpx1′::tpx1.C48S), and p145.C169S (ptpx1′::tpx1.C169S). These plasmids, once linearized, were introduced at the leu1 locus of strain AV36hap-p123.42x (h⁺ leu1 tpx1::kanMX6 pHA-tpx1.42x). A DNA fragment coding for the maltose-binding protein MBP epitope was inserted between the nmt promoter and wild-type or mutant pap1 ORFs to yield p122.41x (pMBP-pap1.41x), p122.41x.C278A (pMBP-pap1.C278A.41x), and p122.41x.C501,523A,C532T (pMBP-pap1.C501,523A,C532T.41x). We used plasmid pRep.41-sty1(HA-6His) to express a tagged-version of Sty1 (20).

Preparation of S. pombe Extracts and Immunoblot Analysis. For in vivo redox state analysis of GFP-Pap1, Pap1, MBP-Pap1, or HA-Tpx1, S. pombe cultures (5 ml) at an OD₆₀₀ of 0.4 were pelleted just after the addition of 100% trichloroacetic acid (TCA) to a final concentration of 10% and washed in 20% TCA. The pellets were lysed by vortexing after the addition of glass beads and 12.5% TCA. Cell lysates were pelleted, washed in acetone, and dried. Alkylation of free thiols was accomplished by resuspension of the pellets in 50 μl of a solution containing 75 mM iodoacetamide, 1% SDS, 100 mM Tris·HCl (pH 8), 1 mM EDTA, and incubation at 25°C for 15 min. Samples were then dephosphorylated to avoid broad bands after electrophoresis by diluting them 5-fold with calf intestinal phosphatase (CIP) buffer to a final concentration of 80 mM Tris·HCl (pH 9.5), 0.08 mM EDTA, and adding 0.04 units/μl of CIP (Roche) for 60 min at 37°C. The alkylated-dephosphorylated samples were electrophoretically separated by nonreducing (or reducing, when indicated) SDS/PAGE, and proteins were immunodetected with a polyclonal anti-Pap1 or with monoclonal anti-HA antibody (12CA5). To detect the presence of sulfinic acid in HA-Tpx1 instead of a thiol group, we used 2D PAGE analysis or Tpx1 cysteine derivatization by 4-acetamido-4′-maleimidylstilbene-2,2′-disulfonic acid (AMS). For 2D PAGE, early log-phase cultures were treated or not treated with H₂O₂. Cell processing and electrophoresis was performed as described in ref. 21. HA-Tpx1 was detected with monoclonal anti-HA antibody (12CA5). Protein extracts were derivatized by AMS and processed as described in ref. 22.

Identification of Tpx1–Pap1 Interaction. Wild-type MBP-Pap1 (from plasmid p122.41x) or empty pRep.41x and wild-type HA-Tpx1 (p123.42x) were coexpressed in Δpap1 S. pombe cells. At an OD₆₀₀ of 1, cells treated with H₂O₂ were harvested and resuspended in lysis buffer (50 mM Tris·HCl, pH 7.5/120 mM KCl/5 mM EDTA, pH 8.0/0.1% Nonidet P-40/10% glycerol/2 mM PMSF/0.16 mg/ml aprotinin/protease inhibitor mixture (Sigma)/4 mM NaF/0.1 mM sodium vanadate) and lysed by vortexing after the addition of glass beads. The protein extracts were then centrifuged to eliminate cell debris. The pull down of MBP-Pap1 and associated proteins was performed by using amylose resin as described by the manufacturer (New England Biolabs). The beads, after washing, were resuspended in SDS-sample buffer, and proteins were separated by reducing SDS/PAGE and immunoblotted with polyclonal anti-Pap1 or monoclonal anti-HA antisera.

Fluorescence Microscopy. Yeast cells expressing GFP-Pap1 were grown, loaded onto slides, and analyzed by fluorescence microscopy as described in ref. 7.

RNA Analysis. Total RNA from S. pombe cultures was obtained, processed, and transferred to membranes as described in ref. 7. Membranes were hybridized with the α-³²P-dCTP-labeled srx1 or gpx1 probes, containing the complete ORFs of the sulfiredoxin- or glutathione peroxidase-coding genes.

H₂O₂ Sensitivity Assay. S. pombe strains were grown in liquid minimal media to an OD₆₀₀ of 0.5. Cells were then diluted in water, and the indicated number of cells in 5 μl was spotted onto minimal media containing or not containing the indicated concentrations of H₂O₂. The spots were allowed to dry, and the plates were incubated at 30°C for 3–4 days.

Results

Tpx1 Is Required for H₂O₂-Activation of Pap1. Three cysteine-based peroxidase-coding sequences appear in the S. pombe genome (23): the thioredoxin peroxidase-coding genes pmp20 and tpx1, structurally related to 2-Cys Prxs, and gpx1, coding for a glutathione peroxidase homologue and structurally related to Orp1/Gpx3. Deletion of gpx1 or pmp20 did not alter cell sensitivity to H₂O₂ (Fig. 1A), and it did not affect Pap1 oxidation and its kinetics (Fig. 1B; data not shown). We failed to delete the tpx1 gene in haploid cells, suggesting that it is essential. We thus deleted one of the tpx1 alleles of a diploid strain, reintroduced in these cells a plasmid expressing HA-Tpx1 from an intermediate strength nmt promoter, and isolated haploid cells that lacked the tpx1 locus but carried the HA-Tpx1-expressing plasmid. Incubation of the cells in thiamine-containing medium blocked HA-tpx1 transcription and resulted in a delayed growth arrest, thus confirming the essential nature of tpx1.

Fig. 1.

The peroxiredoxin Tpx1 is required for Pap1 oxidation activation by H₂O₂.(A) Survival of different strains in response to H₂O₂ exposure. Strains HM123 (WT), AM004 (Δpap1), EA38 (Δsrx1), EA49 (Δpmp20), EA37 (Δgpx1), and wild-type PN513, transformed with plasmid p120.42x (psrx1.42x), were grown in minimal media (EMM) to a final OD₆₀₀ of 0.5, and the number of cells indicated at the top of the panels were spotted onto EMM plates containing or not containing H₂O₂. Plates were incubated at 30°C for 3–4 days. (B) Oxidation of Pap1 at different concentrations of H₂O₂ in different strain backgrounds. The redox state of Pap1 (with TCA extracts resolved in nonreducing SDS/PAGE followed by Western blot analysis) was determined in cells treated with H₂O₂ for the concentrations and times indicated. The strains used in this experiment were HM123 (WT), NT224 (sty1-1) and EA37 (Δgpx1). (C) A Δtpx1 strain transformed with a plasmid containing the nmt-driven HA-tpx1 gene (strain EA40hap-p123.41x) was incubated with thiamine for 5 h (thiamine-nonrepressive conditions; with HA-Tpx1) or 25 h (thiamine-repressive conditions; without HA-Txp1). Cells were left untreated or were treated with 0.2 or 1 mM H₂O₂ for the times indicated. Total HA-Tpx1 was detected in TCA extracts by reducing SDS/PAGE and Western blot analysis. Reduced and oxidized Pap1 were detected in TCA extracts by nonreducing SDS/PAGE, followed by Western blot analysis.

We analyzed the possible role of Tpx1 on Pap1 activation by monitoring its oxidation in H₂O₂-treated cells under thiamine-nonrepressive or repressive HA-Tpx1 expression conditions. In cells expressing wild-type levels of Tpx1, H₂O₂-induced Pap1 oxidation had a wild-type pattern (Fig. 1C). In contrast, in cells lacking Tpx1, Pap1 did not become oxidized in response to H₂O₂ at either 0.2 mM or 1 mM (Fig. 1C). These data demonstrate the absolute requirement of Tpx1 in Pap1 activation.

Tpx1 and Pap1 Interact in Vivo. To examine the ability of Tpx1 to interact with Pap1 in vivo, MBP-Pap1 expressed from an nmt-driven episomal plasmid, was affinity purified by using amylose resin from cells containing an nmt-driven HA-tpx1 allele. HA-Tpx1 was copurified with MBP-Pap1, either from unstressed (data not shown) or from H₂O₂-treated (Fig. 2A) cells.

Fig. 2.

Pap1 and Tpx1 interact in vivo. (A) HA-Tpx1 interacts with MBP-Pap1. Five minutes after 0.2 mM H₂O₂ treatment, cell lysates were prepared from a Δpap1 strain (EHH108) transformed with plasmid p122.41x (pMBP-pap1) or the control vector pRep.41x, and plasmid p123.41x (pHA-tpx1). MBP-Pap1, purified from cell extracts by using amylose resin, was analyzed by reducing SDS/PAGE and Western blot with anti-Pap1 or anti-HA antibodies. As a control, total extracts of both strains were also analyzed. (B) A Δtpx1 strain transformed with a plasmid containing the nmt-driven HA-tpx1 gene (strain AV36hap-p123.42x) was transformed with a linearized plasmid containing the tpx1 promoter fused to wild-type (p145) or mutated (p145.C48S or p145.C169S) tpx1 alleles. The resulting strains, with the tpx1 alleles integrated at the leu1 locus, were incubated with thiamine for 25 h to block expression of plasmid-encoded but not chromosomally encoded versions of Tpx1 to yield cells expressing wild-type Tpx1 (Tpx1 WT) or the mutant Tpx1.C48S or Tpx1.C169S forms, as indicated. Cell treatments (0.2 mM H₂O₂), reducing (with DTT) or not reducing (without DTT) SDS/PAGE, and Western blot analysis with anti-Pap1 antibodies were performed as described in Fig. 1B. The positions and mass (kDa) of markers are indicated. (C) Oxidation of wild-type and mutant forms of Pap1. Strains EHH14 (expressing wild-type GFP-Pap1, lanes 1 and 2), EHH14.C278A (expressing GFP-Pap1.C278A, lanes 3 and 4), EHH14.C501A (expressing GFP-Pap1.C501A, lanes 5 and 6), EHH14.C523A (expressing GFP-Pap1.C523A, lanes 7 and 8), EHH14. C532T (expressing GFP-Pap1. C532T, lanes 9 and 10), EHH14.C501,532A (expressing GFP-Pap1.C501,532A, lanes 11 and 12), and EHH14.C501,523A,C532T (expressing GFP-Pap1.C501,523A,C532T, lanes 13 and 14), were treated (+) or not treated (–) for 5 min with 0.2 mM H₂O₂, as indicated. Extracts were prepared and processed as described in Fig. 1B.(D) Δpap1 strain EHH108 transformed with p122.41x (pMBP-pap1) and p123.42x (pHA-tpx1) (lanes 1 and 2), p122.41x.C278A (pMBP-pap1.C278A) and p123.42x.C169S (pHA-tpx1.C169S) (lanes 3 and 4), or p122.41x.C501A,C523A,C532T (pMBP-pap1.C501A,C523A,C532T) and p123.42x.C169S (pHA-tpx1.C169S) (lanes 5 and 6) were treated (+) or not (–) with 0.2 mM H₂O₂ for 5 min. TCA cell extracts, SDS/PAGE and Western blot analysis were performed as described in Fig. 1B.

2-Cys Prxs mediate one of the principal pathways for the removal of peroxides in aerobically growing cells (24, 25). Their peroxidase function is based on two redox-active cysteines. H₂O₂ is reduced by the so-called peroxidatic cysteine that oxidizes to a sulfenic acid intermediate (Cys-SOH), which then condenses with the resolving cysteine of another subunit to form a disulfide-linked dimer, recycled by the action of thioredoxin. To determine whether Tpx1 cysteine residues are required for Pap1 oxidation, we isolated haploid cells that lacked the wild-type tpx1 locus but carried both an episomal plasmid expressing wild-type HA-Tpx1 from the nmt promoter and a chromosomally integrated wild-type tpx1 allele under control of its own promoter or mutant versions carrying substitutions of either of its catalytic cysteines, tpx1.C48S or tpx1.C169S. Incubation of these cells with thiamine blocked expression of plasmid-encoded, but not chromosomally encoded, versions of Tpx1. In cells expressing either mutant form, Pap1 oxidation to its intramolecular disulfide bond did not occur at either low (Fig. 2B Left) or high (data not shown) H₂O₂ concentrations. In addition, the presence of Tpx1.C169S had the effect of almost completely shifting Pap1 migration from its expected 70-kDa running size to several DTT-sensitive, high molecular mass complexes of ≈90–180 kDa (Fig. 2B) that were suggestive of Pap1-Tpx1 Cys-48 disulfide linkages of unspecified stoichiometries. These DTT-sensitive complexes were induced 5 min after exposure to H₂O₂ and resolved back to the 70-kDa migrating band after 30 min.

Pap1 has two cysteine-rich domains (CRDs), one located at the center of the protein (N-CRD) and the other at the C terminus (C-CRD). H₂O₂-induced Pap1 intramolecular disulfide linkage requires Cys-278 of the N-CRD and either Cys-501 or Cys-532 of the C-CRD, because mutants lacking Cys-278 or both Cys-501 and Cys-532 do not oxidize, whereas Cys-501 and Cys-532 single mutants can still oxidize partially (6, 7) (Fig. 2C). We assayed the effect of these mutations on the Pap1-Tpx1.C169S disulfide linkage. MBP-Pap1 migrated at its expected size on nonreducing SDS/PAGE and formed an intramolecular disulfide bond upon H₂O₂ treatment, whereas MBP-Pap1.C278A did not (data not shown). However, when MBP-Pap1.C278A was coexpressed with HA-Tpx1.C169S, MBP-Pap1.C278A was partially shifted upon H₂O₂ treatment to DTT-sensitive higher molecular weight complexes (Fig. 2D). These complexes were not seen with a mutant Pap1 carrying substitutions of Cys-501, Cys-523, and Cys-532. We conclude that the absolute Tpx1 requirement in H₂O₂-induced Pap1 oxidation relates to its role in transducing the H₂O₂ signal through a Tpx1-Pap1 intermolecular disulfide bond formation involving Tpx1 Cys-48 and, probably, Pap1 Cys-501 or Cys-532.

Inactivation of Tpx1 by 1 mM H₂O₂ Can Be Reversed by Srx1. In eukaryotes, 2-Cys Prxs can be temporarily inactivated during catalysis by further reaction of the Cys-SOH with hydroperoxide and formation of a cysteine-sulfinic acid (Cys-SO₂H) (26, 27), and the extent of this inactivation is in proportion with the amount of H₂O₂ present (28, 29). Cys-SO₂H in Prx is eventually reduced by the sulfiredoxin Srx1 in S. cerevisiae (22) and mammals (30) and also by mammalian sestrin Hi-95 (31). Such substrate-inactivation of 2-Cys Prxs might apply to Tpx1, thus explaining the shutoff of the Tpx1-Pap1 redox relay at elevated H₂O₂ concentrations.

We identified a putative srx1 gene in the S. pombe genome (23) (SPBC106.02c). Deletion of this gene decreased the cell tolerance to H₂O₂ to an extent comparable with that of Δpap1 cells, whereas its overexpression increased this tolerance (Fig. 1 A). We analyzed the effect of its overexpression and deletion on Tpx1/Pap1 oxidation.

We studied the oxidation of Tpx1 in vivo by using 2D PAGE as described in ref. 22. In extracts from untreated cells, Tpx1 appeared as a doublet spot, one major at the position of its expected pI and, thus, corresponding to the reduced protein and the other, minor, more acidic indicating the shift of pI imparted by the Cys-SO₂H (Fig. 3A). At 5 min after exposure to H₂O₂ (1 mM), up to 50% of the protein was shifted to the Cys-SO₂H form, whereas at 60 min, Tpx1 was back to its reduced form. Reduction of the Tpx1 Cys-SO₂H form was not observed in extracts from cells lacking Srx1 (60 min, 1 mM H₂O₂; Fig. 3A). To further demonstrate Tpx1 Cys-SO₂H formation and its Srx1-dependent reversion to the native Cys-SH form, we reacted DTT-treated lysates with AMS that alkylates cysteines in the free SH but not in the SO₂H state, increasing by 0.5 kDa the protein molecular mass per alkylated cysteine. Cells were treated with cycloheximide to block protein synthesis during analysis. In untreated wild-type cell extracts, Tpx1 migrated as a doubly AMS-modified band, indicating that the two protein thiols were free (Fig. 3B). Upon treatment of these cells with H₂O₂ at a low concentration (0.2 mM), Tpx1 still migrated as a doubly AMS-modified band, although a minor singly AMS-modified band was transiently observed after 5 min of treatment, indicating that a very small fraction of the protein had been oxidized to the Cys-SO₂H form and was then quickly reverted to the AMS-reactive Cys-SH form. In contrast, at a higher H₂O₂ concentration (1 mM), 50% of Tpx1 became oxidized to the Cys-SO₂H form for up to 45 min (Fig. 3B). As in the wild-type strain, in Δsrx1 cells treated with 0.2 mM H₂O₂, a very small fraction of Tpx1 was transiently overoxidized (data not shown), whereas upon treatment with 1 mM H₂O₂, Tpx1 accumulated in the Cys-SO₂H form, which did not return to the reduced form (Fig. 3B). We confirmed the function of Srx1 as a Tpx1-recycling sulfiredoxin by measuring the Tpx1 monomer/covalent dimer ratio by using nonreducing electrophoresis of TCA extracts: Tpx1 covalent dimer formation was not observed or only barely observed in cells lacking Srx1 and treated with 1 mM H₂O₂ (Fig. 3C). These data establish that S. pombe Srx1 is a Prx-recycling cysteinyl sulfinic reductase as in S. cerevisiae and mammals.

Fig. 3.

The sulfiredoxin Srx1 is required for Tpx1-mediated activation of Pap1 upon severe H₂O₂ stress. (A and B) Formation of cysteine sulfinic acid in Tpx1 upon severe H₂O₂ treatment. (A) 2D PAGE analysis followed by immunoblot analysis with anti-HA antibody of reduced (SH) and oxidized (SO₂H) forms of HA-Tpx1 in wild-type (Left) and Δsrx1 (Right) cells exposed to 1 mM H₂O₂ for the times indicated. (B) AMS alkylation of Tpx1. Cell cultures were treated with 0.2 or 1 mM H₂O₂, or left untreated, for the indicated times. We performed Western blot analysis of TCA extracts after in vitro alkylation of thiols (SH) but not of cysteine sulfinic acid (SO₂H) residues with AMS. (C) Tpx1 reactivation and dimer formation after severe H₂O₂ stress depends on Srx1. The redox state (monomer versus covalent dimer) of HA-Tpx1 was determined in TCA extracts of cells treated or not treated with H₂O₂ for the times indicated. We performed nonreducing SDS/PAGE and Western blot analysis by using anti-HA antibody. Strains used were wild-type HM123 (WT), EA38 (Δsrx1), and AM001 (Δsty1; only in B and C), each one of them transformed with plasmid p123.81x (pHA-tpx1) to express low levels of the fusion protein HA-Tpx1.

Srx1 Expression upon High H₂O₂ Stress Depends on Sty1. We assayed Tpx1 oxidation in cells lacking the MAPK Sty1. Surprisingly, in Δsty1 cells, Tpx1 also accumulated in the Cys-SO₂H form after exposure to 1 mM H₂O₂ (Fig. 3B), suggesting that srx1 expression depends on the Sty1 MAPK pathway. As seen in Fig. 4A, srx1 mRNA was strongly induced upon H₂O₂ treatment. The Pap1 and the Sty1-Atf1 pathways are both required for the transcriptional activation of srx1 by H₂O₂, but Sty1 is essential at high doses of the oxidant (Fig. 4A; 1 mM in Δsty1 or Δatf1 cells).

Fig. 4.

Srx1 expression depends on Sty1. (A) Northern blot analysis of srx1 expression. Total RNA from strains HM123 (WT), AM004 (Δpap1), AM001 (Δsty1), and JM1066 (Δatf1) was obtained from cultures treated with H₂O₂ for the times and concentrations indicated. (B and C) The late oxidation (B) and nuclear accumulation (C) of Pap1 after 1 mM H₂O₂ stress depends on Sty1 and Srx1. The redox state (with nonreducing SDS/PAGE followed by Western blot analysis) and cellular distribution (with fluorescence microscopy) of a GFP-Pap1 fusion protein was determined in cells treated with H₂O₂ for the concentrations and times indicated. The GFP-Pap1-expressing strains used in this experiment were EHH14 (WT), EA36 (Δsrx1), AV1 (Δsty1) and AV1 overexpressing Srx1 from plasmid p120.42x (psrx1.42x).

We tested the role of Srx1 in Pap1 oxidation and activation. Tpx1 oxidation to the Cys-SO₂H form inversely paralleled not only Tpx1-covalent dimer formation (Fig. 3C), but also the kinetics of Pap1 oxidation and activation (compare Figs. 3B, 4B, and 4C). In cells lacking Srx1, Pap1 never became oxidized upon exposure to H₂O₂ at 1 mM (Fig. 4B), nor did it accumulate in the nucleus, as shown by fluorescence microscopy with a GFP-Pap1 fusion protein (Fig. 4C). At high H₂O₂ concentrations the MAPK Sty1 is required for the oxidation activation of Pap1 (7) (Fig. 4 B and C) and as shown above for Tpx1 Cys-SO₂H reduction (Fig. 3B). Srx1 overexpression in Δsty1 restored Pap1 activation (Fig. 4 B and C) and Tpx1 Cys-SO₂H reduction (data not shown). We conclude that inactivation of the Tpx1-Pap1 redox relay at high H₂O₂ stress is restored by Sty1-dependent Srx1 expression and Tpx1 Cys-SO₂H reduction.

Discussion

We have identified Tpx1 as the upstream activator of Pap1, providing the molecular basis for the temporary shutdown of this pathway in response to elevated levels of H₂O₂ (Fig. 5). Formation of a cysteine-sulfinic acid in Tpx1 as a consequence of high doses of H₂O₂ reversibly inactivates Tpx1 and, therefore, postpones Pap1 activation until the Sty1-dependent transcriptional response is activated. This response includes Srx1, which restores the Tpx1-Pap1 redox relay and Pap1 activation by recycling oxidized Tpx1. The Sty1-dependent H₂O₂ scavenging activities Gpx1 and catalase (12) might also contribute to this regulation by decomposing the excess peroxide, thus preventing reinactivation of Tpx1.

Fig. 5.

A model for Pap1 activation in response to low (0.2 mM) vs. high (1 mM) H₂O₂ stress.

Pap1 inactivation might be a requisite step for allowing Sty1 activation by H₂O₂. We tested this hypothesis by monitoring Sty1-dependent gene expression in cells in which Pap1 was not inactivated as a result of Srx1 overexpression (see Fig. 4 B and C). In these cells, induction of the Sty1-target gene gpx1, in response to H₂O₂ at 1 mM, was dramatically impaired (Fig. 6, which is published as supporting information on the PNAS web site), verifying our hypothesis. Untimely expression of Pap1-dependent genes, which include hydroperoxide-metabolizing enzymes, may inhibit Sty1 activation by intercepting its H₂O₂ input signal. Alternatively, active (oxidized) Pap1 may directly repress Sty1-dependent genes, as shown for gpx1 (12, 32). Tpx1 has also been shown to be required for Sty1 activation by H₂O₂ (33). However, the kinetics and magnitude of Sty1 phosphorylation and Sty1-dependent transcription were not altered in Δsrx1 cells (Fig. 7, which is published as supporting information on the PNAS web site), indicating that Tpx1 oxidation to the Cys-SO₂H form does not affect its function in Sty1 activation.

Peroxiredoxin inactivation by cysteine-sulfinic acid formation has been suggested as a possible regulatory mechanism in H₂O₂ signaling in eukaryotic cells (34). In fact, it has very recently been reported that S. cerevisiae peroxiredoxin is able to regulate the transcription factor Yap1 in specific yeast strains (35). We provide here evidence that in S. pombe, cysteine-sulfinic acid formation represents a redox switch regulating the function of peroxiredoxin as an H₂O₂ sensor and redox transducer. This regulation of the Tpx1-Pap1 redox relay controls the timely expression of Pap1 target genes, which may be important for activation of the Sty1 pathway.