Peroxiredoxin Ahp1 Acts as a Receptor for Alkylhydroperoxides to Induce Disulfide Bond Formation in the Cad1 Transcription Factor*

Reactive oxygen species (ROS) generated during cellular metabolism are toxic to cells. As a result, cells must be able to identify ROS as a stress signal and induce stress response pathways that protect cells from ROS toxicity. Recently, peroxiredoxin (Prx)-induced relays of disulfide bond formation have been identified in budding yeast, namely the disulfide bond formation of Yap1, a crucial transcription factor for oxidative stress response, by a specific Prx Gpx3 and by a major Prx Tsa1. Here, we show that an atypical-type Prx Ahp1 can act as a receptor for alkylhydroperoxides, resulting in activation of the Cad1 transcription factor that is homologous to Yap1. We demonstrate that Ahp1 is required for the formation of intermolecular Cad1 disulfide bond(s) in both an in vitro redox system and in cells treated with alkylhydroperoxide. Furthermore, we found that Cad1-dependent transcriptional activation of the HSP82 gene is dependent on Ahp1. Our results suggest that, although the Gpx3-Yap1 pathway contributes more strongly to resistance than the Ahp1-Cad1 pathway, the Ahp1-induced activation of Cad1 can function as a defense system against stress induced by alkylhydroperoxides, possibly including lipid peroxides. Thus, the Prx family of proteins have an important role in determining peroxide response signals and in transmitting the signals to specific target proteins by inducing disulfide bond formation.

Oxygen serves as an electron acceptor, enabling efficient production of ATP. However, oxygen can also be converted into toxic reactive oxygen species (ROS), 3 such as superoxide and hydrogen peroxide (H 2 O 2 ); ROS can damage a variety of cellular components, including proteins and unsaturated lipids (1).
Toxicity often arises from lipid peroxidation as well as production of alkoxyl/peroxyl radicals and aldehyde compounds, which are products of radical lipid reactions triggered by ROS. Additionally, ROS are important contributors to cellular metal toxicity (2).
Detection of ROS (peroxides) is an important step in the oxidative stress response. In bacteria, sensing of H 2 O 2 involves transcription factor OxyR (3)(4)(5), whereas transcription factors Yap1 and Pap1 have this role in yeast. These transcription factors activate the transcription of genes encoding enzymes that mitigate oxidative stress in response to H 2 O 2 . OxyR reacts directly with H 2 O 2 to form a sulfenic acid at Cys 199 of the transcription factor. The sulfenic acid rapidly forms a disulfide bond with Cys 208 , activating the transcription factor, which then triggers expression of the target genes. By contrast, H 2 O 2 receptors are required for activation of the Yap1 transcription factor in Saccharomyces cerevisiae (budding yeast) (6) as well as its Schizosaccharomyces pombe ortholog, Pap1 (fission yeast) (7,8). Glutathione peroxidase (Gpx)-like factor Gpx3 (also known as Orp1) is the H 2 O 2 receptor involved in Yap1 activation in budding yeast. The presence of H 2 O 2 is first detected in yeast when a catalytic cysteine of Gpx3 is directly oxidized by H 2 O 2 to form sulfenic acid. This reaction is followed by the transient formation of an intermolecular disulfide bond with a cysteine residue in the Yap1 transcription factor. A thiol-disulfide exchange reaction results in the formation of an intramolecular disulfide bond in Yap1, which triggers activity of the transcription factor. In the case of Pap1, the H 2 O 2 receptor is Tpx1 in fission yeast. Whether molecular mechanisms other than those involving specific peroxide receptors participate in H 2 O 2 sensing remains to be determined. The fact that both Gpx3 and Tpx1 have thioredoxin (Trx)-dependent peroxidase (Tpx) activity is potentially important for developing a better understanding of H 2 O 2 sensing.
Prx is a family of ubiquitous peroxidases found in species ranging from Escherichia coli to humans (9). In many cases, Prx can reduce H 2 O 2 and/or alkylhydroperoxides at the expense of electrons from NADPH through the Trx-dependent redox system (Tpx activity) (see Fig. 3A). The catalytic cysteine (Cys) residue of Prx is directly oxidized by hydroperoxides (see Fig.  1D). This oxidation event is followed by the formation of a disulfide bond linkage with a resolving Cys of the same molecule of Prx or with a resolving Cys of another Prx molecule (dimer formation). The disulfide bond can be reduced by Trx. Five Prx family proteins (Tsa1, Tsa2, Prx1, Dot5, and Ahp1) (10) and two Gpx-like proteins (Gpx2 and Gpx3) display Tpx activity in budding yeast (6,11). Differential expression, localization, and specificity of each of these Prx family proteins (12) suggest that they may have different spatial-temporal roles. By contrast, only Tpx1 and Gpx1 have Tpx activity in fission yeast (13). Multiplicity of Prx enzymes in budding yeast is analogous to that in the mammalian system, as mammalian cells express six conserved Prx family proteins. Thus, functional studies on the different Prx proteins in budding yeast may provide an understanding of the role of these proteins in a broad range of eukaryotes.
Recently, we have suggested (14,15) that, in addition to their function as peroxidases, Prx family proteins may serve as intrinsic receptors of H 2 O 2 . Additionally, Prx proteins may relay information about the presence of hydroperoxides to the independent target proteins of each Prx; however, the contribution of the Prx family members in sensing the presence of hydroperoxides is yet to be determined. Furthermore, the putative alkylhydroperoxide-sensing mechanism is entirely unresolved. It is possible that multiple sensing mechanisms exist that are determined by unique specificities of each Prx protein for different hydroperoxides and to their respective target proteins. In the present study, we addressed whether the Prx Ahp1 is responsible for hydroperoxide sensing.
Oxidation Reaction of Cad1-For the in vitro assay of the oxidation reaction of Cad1, the carboxyl-terminal region of Cad1 (nucleotides 332-1230 corresponding to amino acids 111-409), in which the amino-terminal bZIP region was deleted, was cloned between BamHI and SalI sites in pGEX-6P-2 (GE Healthcare). To construct the HA-Cad1 expression plasmid, the HA tag was inserted between the BamHI site and the amino terminus of the above Cad1 sequence. We purified the Cad1 (111-409) protein and reduced it with DTT. The oxidation process was examined as described previously (14). Oxidation was initiated by the addition of 0.4 mM tBOOH. Aliquots were removed at indicated times and mixed with N-ethylmaleimide to a final concentration of 50 mM at 30°C for 30 min.
Analysis of the Redox State of Cad1 in Cells-We added tBOOH to a log phase culture (A 600 ϭ 0.5) of cad1⌬ (BY4742 cad1⌬; Y14259) cells carrying the pRS316-cup-HA2-Cad1 plasmid in SD dropout (ϪUra). We collected cells from 10 ml of culture at the indicated time and resuspended them in 20% trichloroacetic acid. Cells were disrupted in 12.5% trichloroacetic acid, after which they were neutralized in oxygen-free conditions (17). Cell lysate was diluted with buffer containing 100 mM Tris-HCl (pH 7.5), 9 M urea, 1% SDS, 1 mM EDTA, 50 mM N-ethylmaleimide, and Complete Mini protease inhibitor mixture (Roche Diagnostics).
Other Methods-We examined the interaction between Ahp1 and Cad1 by immunoblotting after immunoprecipitation as described previously (14). Two-hybrid screening (18) and chromatin immunoprecipitation (19) were performed as described previously. Time-dependent consumption of NADPH in the Tpx (Gpx3 or Tsa1 or Ahp1)/Trx2/Trr1/ NADPH redox system was monitored as described previously (14,17). Precise methods for all experiments, including analysis of expression using quantitative PCR and the generation of anti-Ahp1 antibodies, are described in the supplemental Experimental Procedures.

RESULTS
Requirement of Ahp1 for Resistance to Alkylhydroperoxide-Ahp1 was previously identified as an antioxidant to alkylhydroperoxides (20). We compared the requirement of GPX3, TSA1, and AHP1 genes for resistance to H 2 O 2 and to the alkylhydroperoxide tBOOH (Fig. 1A). Disruption of GPX3 affected cell sensitivity to these hydroperoxide stresses, suggesting that Gpx3-induced transcriptional activation of Yap1 is an important detoxifying system in response to H 2 O 2 , tBOOH and methyl linolenic acid (an unsaturated lipid that induces lipid peroxidation stress). TSA1 encodes a major Prx family protein; Tsa1 and Ahp1 consist of 91% and 3.9% of total Tpx, respectively (14,21), and disruption of the TSA1 gene resulted in a modest increase in sensitivity to tBOOH and methyl linolenic acid. Disruption of the AHP1 gene significantly increased sensitivity only to methyl linolenic acid. Our results suggested that Ahp1 contributes to resistance more to unsaturated lipid than hydroperoxides (Fig. 1A). Next, we compared the resistance of gpx3⌬ and gpx3⌬ ahp1⌬ cells to tBOOH. We employed a low concentration of tBOOH in this case because disruption of GPX3 induced the sensitivity of yeast cells to hydroperoxides (0.05 mM; Fig. 1B). Sensitivity to tBOOH was significantly increased when GPX3 and AHP1 were simultaneously disrupted (gpx3⌬ ahp1⌬). To examine the possibility that Ahp1 could act as a tBOOH receptor and activate Yap1-dependent transcription (analogous to the mechanism by which Gpx3 responds to peroxide and activates Yap1), we compared Yap1dependent transcription in an AHP1 disruption mutant with Yap1-dependent transcription in the wild-type strain. Our findings suggest that Ahp1 is involved in reducing intracellular peroxide levels because disrupting AHP1 increased Yap1-dependent transcriptional levels in unstressed cells (Fig. 1C). By contrast, hydroperoxide-induced Yap1-dependent transcription was unaffected by disruption of the AHP1 gene. Furthermore, disruption of AHP1 did not affect basal transcription levels in gpx3⌬ cells (Fig. 1C, inset). These results suggest that Ahp1 does not function as a peroxide receptor upstream of Yap1-dependent transcription.
Cad1 Is a Candidate Target Protein of Ahp1-We pursued the possibility that Ahp1 acts as an alkylperoxide receptor and transmits the peroxide signal by introducing disulfide bond(s) in target proteins other than Yap1. Our previous study (14) demonstrated that substitution of a resolving cysteine in Tsa1 stabilizes the formation of a transient disulfide bond between a

Ahp1 as Receptor for Peroxide to Activate Cad1
catalytic cysteine of Tsa1 and Yap1. Therefore, we aimed to isolate proteins that could interact with the catalytic cysteine of an Ahp1 mutant containing a substitution of the resolving Cys at position 120 (Ahp1 C120T ) (see Fig. 1D). Using the two-hybrid system, we identified Cad1 (also known as Yap2) as a protein that could specifically interact with Cys 62 of Ahp1 (Fig. 1E). To confirm this interaction, we simultaneously expressed Ahp1 C120T and HA-tagged Cad1 in yeast cells. Using this system, we found that binding of Ahp1 C120T to Cad1 was strongly induced in response to tBOOH exposure (Fig. 1F). We observed both monomeric and dimeric Ahp1 C120T , even under nonreducing conditions in SDS-denatured samples analyzed by PAGE (lane 10). In addition, we observed the presence of slower migrating complexes, likely corresponding to Cad1 bound to Ahp1 C120T via a disulfide bond. It is possible that dimeric and monomeric Ahp1 are released from slower migrating complexes after immunoprecipitation. We also observed dimeric Ahp1 (wild-type) in Cad1 immunoprecipi-FIGURE 1. Cad1 is a possible target protein of Ahp1. A and B, both Ahp1 and Gpx3 are required for resistance to tBOOH and unsaturated lipid. Resistance of wild-type (WT) cells and mutant cells to peroxide treatment was determined by a spot assay. Log phase cultures of wild-type, GPX3 disruption cells (gpx3⌬), TSA1 disruption cells (tsa1⌬), and AHP1 disruption cells (ahp1⌬) were spotted onto YPAD plates containing the indicated level of peroxides or methyl linoleic acid. The numbers of cells spotted were 2.25 ϫ 10 5 , 4.5 ϫ 10 4 , 1.5 ϫ 10 4 , 5 ϫ 10 3 , 2.5 ϫ 10 3 , and 1.25 ϫ 10 3 . The plates were incubated at 30°C for 48 h. C, induction of the transcriptional activity of the reporter gene for Yap1 (TRX2 promoter-LacZ) in the mutant cells in response to peroxide treatment is shown. The indicated mutant cells transformed with pTRX2-LacZ were cultured to log phase and treated with 0.4 mM tBOOH and 0.5 mM H 2 O 2 for 30 min. ␤-Galactosidase activity was determined as previously described. D, cysteine residues (Cys 62 and Cys 120 ) responsible for disulfide bond formation in response to peroxide exposure are indicated. Cys 62 (-SH, -S-) is a catalytic cysteine, and Cys 120 is a resolving cysteine (-SH, -S-) (20). E, two-hybrid interaction between Cad1 and Ahp1 C120T is shown. The indicator strain carrying the indicated cysteine mutant of Ahp1 in the bait plasmid and the empty pray plasmid (Ϫ) or the Cad1 pray plasmid (Cad1) were spotted onto an SD plate containing histidine (SD) or one containing 1 mM 3-aminotriazol with no histidine (SD (ϪHis) ϩ 3AT). The plates were incubated at 30°C for 3 days. F, interaction of Cad1 and Ahp1 C120T in cells treated with tBOOH is shown. The AHP1 disruption cells expressing Cad1-HA (lanes 2 -5 and lanes 7-10), Ahp1 (lanes 2, 3, 7, and 8), and Ahp1 C120T (lanes 4, 5, 9, and 10) were treated (lanes 3, 5, 8, and 10) or not treated (lanes 1, 2, 4, 6, 7, and 9) with tBOOH. HA-Cad1 was immunoprecipitated (IP), and Cad1 and Ahp1 were detected by immunoblotting (IB) as described under "Experimental Procedures." A high contrast image of the upper region of the immunoblots using anti-Ahp1 is shown. The DTT-sensitive slower migrating complexes are indicated by arrows (Complex). G, structures of Yap1 and Cad1. Schematic illustration of Yap1 and Cad1 with homologous bZIP and cCRD regions is shown. The three disulfide bonds that form between nCRD and cCRD in Yap1 (17) are indicated by a bar (3 ϫ S-S).

tate of lysate from cells treated with tBOOH (lane 8).
We reported a similar phenomenon in the case of Tsa1-Yap1 association (14).
Requirement of Ahp1 for Cad1-dependent Activity-We investigated whether Ahp1 might induce transcriptional activity of Cad1, analogous to the mechanism whereby Gpx3 induces transcriptional activity of Yap1 in response to H 2 O 2 exposure. Interestingly, Cad1 is a Yap1 family transcription factor and contains two domains that are similar to domains in Yap1: a bZIP domain and a carboxyl-terminal cysteine-rich domain (cCRD) (22) (see Fig. 1G). Consistent with homology to the bZIP domain, Cad1 exhibits a binding specificity similar to Yap1 and can activate some of the same target genes when it is overexpressed (23). However, Cad1 can also activate indepen-dent target genes, including HSP82, an inducible HSP90 family protein (24). Based on these data, we utilized chromatin immunoprecipitation followed by PCR to examine whether Cad1 binds to the promoter region of HSP82 ( Fig. 2A). We determined that Cad1 can bind to the Ϫ35 to Ϫ233 region of the HSP82 promoter but could not bind to the Ϫ510 to Ϫ671 region. The amount of Cad1 bound to the former region increased in response to tBOOH exposure (Fig. 2B). The Cad1-bound region of the HSP82 promoter does not contain any consensus sequences for Yap1 or Cad1. However, this region does contain one heat shock factor binding consensus sequence. We did not identify the exact Cad1 binding site on the HSP82 promoter. As a control, we performed a similar assay using PCR probes corresponding to a region of the TRX2 promoter that contains two Yap1 binding sites. This experiment demonstrated that exposure to H 2 O 2 enhanced binding of Yap1. In addition, this assay showed that exposure to tBOOH increased binding of Cad1 to the same promoter region (see supplemental Fig. 1). Consistent with the binding activity of Cad1, transcriptional activation of HSP82 was induced by tBOOH exposure in an Ahp1-and Cad1-dependent manner (Fig. 2C). Together, these results suggest that Ahp1 acts as receptor for tBOOH and activates binding of Cad1 to the promoter region of HSP82.
We next examined the importance of Cad1 to Ahp1 activity in promoting resistance to tBOOH. Because disruption of neither AHP1 (Fig. 1) nor CAD1 in wild-type cells resulted in tBOOHsensitive phenotypes (data not shown), we utilized YAP1 disruption mutants as the parent strains. As shown in Fig. 2D, disruption of AHP1 and CAD1 resulted in a more sensitive phenotype than observed in parent cells (yap1⌬) derived from W303. Although AHP1 disruption resulted in a more sensitive phenotype than disruption of CAD1, we found that the phenotype of the AHP1/CAD1-double mutant was similar to that of the AHP1 disruption mutant (Fig. 2D, 50 M). These results suggest that Ahp1-dependent up-regulation of Cad1 determines the requirement of Ahp1 for resistance to tBOOH at least in the specific condition. We could not observe the above phenotype with BY4742 (yap1⌬) by disruption of CAD1 (Fig.   FIGURE 2. Ahp1 is required for Cad1-specific transcriptional induction. A, schematic illustration of the promoter region of HSP82 and the position of primers used for chromatin immunoprecipitation (ChIP) analysis is shown. B, input DNA and immunoprecipitated (IP) DNA were used as templates for PCR using either a control primer set or a primer set specific for the Cad1 binding region. C, induction of HSP82 transcription was dependent on both Ahp1 and Cad1. Quantitative PCRs were performed using RNA prepared from the indicated yeast cell line after treatment with 0.6 mM tBOOH for 20 min. D and E, both Ahp1 and Cad1 are required for maximum resistance to tBOOH. The indicated yeast cells, derived from W303 yap1::LEU2 (parent) (D) and BY4742 (yap1⌬) (E), were spotted onto YPAD plates without peroxide containing 5 and 20 M tBOOH as described in the legend to Fig. 1.  2E). Mechanism for the distinct contribution of Cad1 in these strains is currently unknown. One possibility is that W303 might have been adapted to utilize Ahp1/Cad1-dependent alkylhydroperoxide resistance because hydroperoxide-induced Yap1 is solely dependent on Tsa1 but not Gpx3 in W303 (14), and W303 is more sensitive to hydroperoxide than BY4742. These characteristics of W303 are due to a truncation mutation of Yap1-binding protein (Ybp1) (15) Ybp1 is required for Gpx3-dependent activation of Yap1 in response to hydroperoxide (25). The data also suggested that peroxidase activity Ahp1 or Ahp1-dependent activation of other target protein may be required for tBOOH resistance.

Ahp1-induced Oxidation of Cad1
in Vitro-A characteristic feature of Cad1 is a cluster of cysteine-rich regions. Specifically, the protein contains both a unique domain composed of seven cysteine residues and cCRD (Fig. 1G) similar to Yap1. The presence of this domain prompted us to investigate whether Cad1 is oxidized by the redox cycle of Ahp1 in the presence of tBOOH, as we demonstrated for Yap1 (14,17). First, we evaluated the suitability of tBOOH as a substrate for Ahp1 using the Trx reduction system (Fig. 3A; Trx2/ Trr1/NADPH). As shown in Fig. 3B, both tBOOH and H 2 O 2 efficiently oxidized Ahp1. By contrast, oxidation of Tsa1 by tBOOH was inefficient. Gpx3 was efficiently oxidized by both H 2 O 2 and tBOOH in a manner similar to Ahp1 (see supplemental Fig. 2). These results suggest that Tsa1 is primarily responsible for reduction of H 2 O 2 , whereas Gpx3 and Ahp1 are also responsible for the reduction of alkylhydroperoxide.
We next examined Cad1 disulfide bond formation by assaying its mobility in nonreducing SDS-PAGE. We successes to express a truncated Cad1 protein (bZIP region was removed) in E. coli, but not full-length Cad1 protein. Direct treatment of the truncated Cad1 with tBOOH for 30 min did not significantly affect Cad1 mobility (Fig. 3C). This was also the case in the presence of the Trx reduction system (Fig. 3D). By contrast, the combination of Ahp1 and the Trx redox system (Fig. 3, E and F) yielded a slower migrating Cad1 (Cad1 ox1 ; Fig.  3F, 0.5 min) once Ahp1 was fully oxidized (dimer formation; Fig. 3E, 0.5-30 min). Cad1 ox1 appeared to convert into a rapidly migrating Cad1 molecule (Cad1 ox2 ; Fig. 3F, 1 min) but then transform into a very slow migrating Cad1 (Cad1 ox3 ; 5-30 min). The migration patterns of these three Cad1 species were reversed by DTT treatment (ϩDTT; Fig. 3, E and F). These results suggest that Ahp1-dependent mobility shifts of Cad1 in response to tBOOH are the result of possible intramolecular disulfide bonds within a Cad1 molecule as well as possible intermolecular disulfide bonds between Cad1 molecules (homodimers).
Cad1 Is Oxidized in Yeast Cells in Response to tBOOH-We next examined the oxidation state of Cad1 in cells treated with tBOOH. As shown in Fig. 4A and supplemental Fig. 3A, Cad1 formed slower migrating complexes, corresponding to FIGURE 3. Ahp1-dependent oxidation of Cad1 by tBOOH in vitro. A, a Prx (Ahp1, Tsa1, and Gpx3 in this study) coupled to thioredoxin (Trx2), thioredoxin reductase (Trr1), and NADPH (Trx/Trr/NADPH system) can reduce hydroperoxide at the expense of high energy electrons from NADPH. B, substrate-specific peroxidase activities of Ahp1 and Tsa1 are shown. Tsa1-catalyzed (left) and Ahp1-catalyzed (right) time-dependent consumption of NADPH by 0.4 mM tBOOH and 0.5 mM H 2 O 2 were observed in the Trx/Trr/NADPH reduction system in the presence or absence (ϪTsa1, ϪAhp1) of Tsa1 and Ahp1. Absorbance of NADPH at 340 nm (A 340 ) was monitored in 1-min intervals. C and D, we treated Cad1 with 0.6 mM tBOOH alone (C) or with 0.6 mM tBOOH in the Trx/Trr/NADPH reduction system (D). E and F, oxidation reaction of Cad1 in the presence of Ahp1 by 0.4 mM tBOOH in the Trx/Trr/NADPH reduction system in vitro. Reactions at each time point (minutes) were stopped by addition of N-ethylmaleimide and fractionated by SDS-PAGE (15% for C and D, 12% for E and F). We treated the reaction mixture (30 min after addition of tBOOH) with 50 mM DTT at 100°C for 5 min (indicated as ϩDTT in D-F). The gels were stained with Coomassie Brilliant Blue (C-E). Proteins resolved by PAGE shown in E were transferred to a membrane and immunoblotted with anti-HA antibody to detect HA-Cad1 (F). The green arrows indicate the position of the reduced Cad1 (Cad1 red or Red), and the red arrows indicate oxidized Cad1 (Ox1, Ox2, and Ox3). Degradation products formed during expression and purification of bacterially expressed Cad1 are indicated as Deg. Positions of reduced Ahp1 (Ahp1 red ), oxidized Ahp1 (Ahp1 ox ), and Trr1 (Trr1) are indicated by black arrows. Positions of molecular mass markers (kDa) in C-F are indicated by arrowheads. molecular masses of 110-60 kDa, in wild-type cells within 0.5 min of tBOOH exposure. Complexes lingered for 10 min but were no longer observed after 15 min. Because the complexes were sensitive to DTT treatment (Fig. 4A, right, and supplemental Fig. 3B), it is likely that slower migrating complexes contained possible dimers of Cad1 linked by disulfide bonds. We did not observe the oxidized form of Cad1 in AHP1 disruption mutant cells (Fig. 4B and supplemental Fig. 3A). These results suggest that formation of tBOOH-dependent Cad1 disulfide bridge(s) is solely dependent on Ahp1. Although ϳ1% of total Ahp1 was dimerized (oxidized) under unstressed conditions, Ahp1 was completely oxidized 0.5 min after treatment with tBOOH (Fig.  4C). The level of Ahp1 oxidation gradually declined after 15 min, suggesting that Ahp1 was continuously oxidized and reduced but not hyperoxidized (inactivated) in the cells exposed with 0.6 mM tBOOH. Interestingly, we observed slower migrating Ahp1 bands, corresponding to molecular masses of Ͼ100 kDa, within 0.5-1 min after exposure to tBOOH. It is likely that these large molecular mass bands correspond to transient disulfide bond formation between Ahp1 and other protein molecules.
It is worth noting that the oxidation process of Cad1 in vitro (Fig. 3F) was somewhat different from that observed in cells (Fig. 4A). In cells, the possible Cad1 dimer, linked via a disulfide bond, formed directly without the need for intramolecular disulfide bonds. Our use of a truncated Cad1, devoid of the bZIP region, for the in vitro oxidation study might explain the initial induction of possible intramolecular disulfide bonds. Nevertheless, our results strongly suggest that Ahp1 is required for disulfide bond formation in Cad1 in response to tBOOH exposure.
Unsaturated Lipid Can Oxidize Ahp1-Our results suggested that Ahp1 might act as a receptor for lipid peroxide induced in cells treated with methyl linolenic acid. Thus, we examined oxidation of Ahp1 in the in vitro Trx reduction system and in yeast cells treated with methyl linolenic acid. We found that methyl linolenic acid could oxidize Ahp1 (Fig. 5A).  A and B, 15% for C) after treatment with DTT (right panels). We detected Cad1 (A and B) and Ahp1 (C) proteins by immunoblotting with an anti-HA antibody and anti-Ahp1 serum, respectively. Arrows indicate oxidized Cad1 complex (Cad1 ox -complex), reduced Cad1 (Cad1 red ), oxidized Ahp1 (Ahp1 ox ), and reduced Ahp1 (Ahp1 red ). Positions of molecular mass markers (kDa) in each panel are indicated by arrowheads. D, model of the mechanism for Ahp1-induced disulfide bond formation of Cad1. Even in an unstressed steady state, Ahp1 is continuously oxidized by direct reaction with intracellular peroxide, inducing formation of the dimer. Dimerized Ahp1 is also continuously reduced by Trx. The dimer/monomer ratio may be determined by the peroxide level and redox state of Trx. When a catalytic cysteine in the Ahp1 dimer reacts with peroxide, the resulting sulfenic acid (S-OH) attacks Cad1 to form a transient complex linked by disulfide bonds. A cysteine residue of Cad1 may attack the disulfide bond between Cad1 and Ahp1 to form a Cad1 dimer linked by a disulfide bond or an intramolecular disulfide bond in Cad1.
Ahp1 but not Tsa1, and Gpx3 was oxidized in cells in response to methyl linolenic acid (Fig. 5, B-D).

DISCUSSION
The Prx family of proteins possess the ability to reduce peroxide at the expense of their free cysteine residues, resulting in the formation of disulfide bonds or peroxidation to sulfinic acid. In addition to this defined function, our previous data indicate that a major Prx, Tsa1, can induce disulfide bridges to Yap1, implying an alternative function for Prx proteins. Here, we show that the Prx Ahp1 can catalyze formation of disulfide bonds to Cad1 in response to alkylhydroperoxide tBOOH both in vitro reconstitution system and in yeast cells. Furthermore, we show that Ahp1 is required for the up-regulation of Cad1dependent transcriptional activity. Because Ahp1 is responsible for resistance to unsaturated lipid (Fig. 1) and Ahp1 is oxidized In this case, 4 mM methyl linolenic acid was solubilized by addition of 0.1% Triton X-100. Absorbance at 340 nm due to the turbidity of methyl linolenic acid was subtracted from the absorbance of NADPH at each time point. B-D, time-dependent oxidation status of Ahp1 (B), Tsa1 (C), and Gpx3 (D) in cells treated with 4 mM methyl linolenic acid in the medium containing 0.1% Triton X-100. We prepared cell lysates and fractionated by 15% SDS-PAGE as described in the legend to Fig. 4. We detected Ahp1 (B), Tsa1 (C), and Gpx3 (D) proteins by immunoblotting with an anti-Ahp1, anti-Tsa1 (14), and anti-Gpx3 (19) sera, respectively. Graphs in B and C indicated ratios of dimer protein (oxidized) level to total protein level (monomer and dimer) of Ahp1 and Tsa1, respectively, at each time point. We could not observe the oxidized form of Gpx3, which migrates faster in SDS-PAGE (6,17).
in cells in response to unsaturated lipid (Fig. 5), Ahp1 might act as both a receptor and a peroxidase for alkylhydroperoxides, including lipid peroxides (Fig. 4D). These results bolster our hypothesis that Prx proteins can act as intrinsic receptors of hydroperoxides.
We found that Gpx3 is also reactive to alkylhydroperoxide (supplemental Fig. 2) and is required for resistance to alkylhydroperoxide and methyl linolenic acid (Fig. 1), although apparent oxidized form of Gpx3 was not observed (Fig. 5D). In addition, we demonstrated that Ahp1 is required for maximum resistance of BY4742 (gpx3⌬) cells to tBOOH (Fig. 1) and that both Cad1 and Ahp1 are required for W303 (yap1⌬) (Fig. 2). These findings suggest that both Ahp1-dependent Cad1 activation and Gpx3-dependent Yap1 activation are required for protection against alkylperoxide and lipid peroxide toxicity (Fig. 6). It is clear, however, that the Gpx3-Yap1 pathway contributes more strongly to resistance than the Ahp1-Cad1 pathway.
Similar to the mechanism of Yap1 activation, tBOOH and cadmium inhibit Cad1 nuclear export by effecting structural changes through oxidation of cysteine residues (22,26). Prior studies have also demonstrated that direct binding of cadmium to cysteine residues in cCRD is responsible for inhibiting its nuclear export signal (22). Ahp1 may be unnecessary for activation in the case of the direct binding mechanism.
This study indicates that Ahp1 should be categorized as a Prx family member that induces disulfide bond formation in its target molecules. Our results further support the hypothesis that Prx proteins are generally able to act as receptors for peroxides and as transmitters of redox signals to other sensor/transducer molecules. We have chosen to call this redox signal transduc-tion "Prx-dependent peroxide signal transmission." Current efforts are under way to identify similar systems of mammalian Prx-dependent sensing and transduction of peroxide signals.