The budding yeast Saccharomyces cerevisiae has three homologues of glutathione peroxidase (GPX1, GPX2, and GPX3). Two structural homologues of the mammalian glutathione peroxidase, Gpx2 and Gpx3, have been proven to be atypical 2-Cys peroxiredoxins, which prefer to use thioredoxin as an electron donor. Here, we show that Gpx1 is also an atypical 2-Cys peroxiredoxin, but uses glutathione and thioredoxin almost equally. We determined the redox state of Gpx1 in vivo.
Glutathione peroxidase (GPx) is a key enzyme of the antioxidant system in eukaryotic organisms, catalyzing the reduction of H2O2 and organic hydroperoxides to water and corresponding alcohols using reduced glutathione as an electron donor (for a review, see Margis, 2008). Mammalian GPxs have selenocysteine (SeCys) at their active site, although a certain type of GPx contains cysteine instead of SeCys (Ghyselinck, 1990). Non-SeCys-type GPxs have been found in plants (Criqui, 1992; Holland, 1993; Depège, 1998; Roeckel-Drevet, 1998) and yeasts (Inoue, 1999). Non-SeCys-type GPxs often use thioredoxin as an electron donor to carry out reactions (for a review, see Herbette, 2007). The budding yeast Saccharomyces cerevisiae has three homologues of glutathione peroxidase: GPX1, GPX2, and GPX3 (Inoue, 1999). Two structural homologues of mammalian GPxs, S. cerevisiae Gpx2 and Gpx3, have been found to be atypical 2-Cys peroxiredoxins (Delaunay, 2002; Tanaka, 2005). As a consequence of the peroxidase reaction of atypical 2-Cys peroxiredoxins, an intramolecular disulfide bond is formed, which is usually cleaved by thioredoxin (Hofmann, 2002). The redox status of Gpx2 and Gpx3 in vivo is regulated by thioredoxin (Delaunay, 2002; Tanaka, 2005). In this study, we determined the enzymatic properties and the redox regulation of Gpx1.
To characterize Gpx1, we expressed it in Escherichia coli cells using basically the same method as in our previous report (Tanaka, 2005). The procedures for the construction of the plasmid are described in Supporting Information, Appendix S1. Because the recombinant Gpx1 easily aggregates (Avery & Avery, 2001), the purification and characterization of Gpx1 have not succeeded. To avoid the aggregation of Gpx1, imidazole was removed from the enzyme fraction containing Gpx1 after Ni2+-column chromatography by dialyzing stepwise against a buffer containing 100, 50, 10, and 0 mM imidazole. GPx activity with the glutathione system (glutathione, glutathione reductase, and NADPH) and the thioredoxin system (thioredoxin, thioredoxin reductase, and NADPH) was measured as described previously (Tanaka, 2005). To determine the redox state of Gpx1 in yeast cells, we used the 4′-acetamido-4′-maleimidylstilbene-2,2′-disulfonic acid (AMS) (Molecular Probes) assay and site-directed mutagenesis to change the cysteine residues of Gpx1 to serine. Details are given in Appendix S1. Saccharomyces cerevisiae YPH250 (MATatrp1-Δ1 his3-Δ200 leu2-Δ1 lys2-801 ade2-101 ura3-52) and isogenic mutants (gpx1Δ∷KanMX4 and gpx1Δ∷KanMX4 trx1Δ∷HIS3 trx2Δ∷LEU2) were used for the AMS assay. Cells were cultured in synthetic dextrose (SD) medium (2% glucose and 0.67% yeast–nitrogen base without amino acids) with appropriate amino acids and bases until a stationary phase of growth to induce the expression of GPX1 (Ohdate, 2010).
The purified Gpx1 exhibited peroxidase activity using both glutathione and thioredoxin as electron donors. The kcat/Km value of Gpx1 for reducing H2O2 was 4.10 × 106 M−1 s−1 in the glutathione system and 1.19 × 106 M−1 s−1 in the thioredoxin system (Table S1). Thus, the catalytic efficiency of Gpx1 was approximately fourfold higher in the glutathione system than the thioredoxin system with H2O2 as a substrate. Meanwhile, the kcat/Km value of Gpx1 for reducing tert-butyl hydroperoxide (t-BHP) in the glutathione system (1.49 × 106 M−1 s−1) was almost the same as that in the thioredoxin system (2.16 × 106 M−1 s−1). We have reported previously that the catalytic efficiency of Gpx2 in the reduction of H2O2 was 100 times higher with the thioredoxin system than with the glutathione system (kcat/Km of 4.79 × 107 vs. 5.85 × 105 M−1 s−1). Also, the kcat/Km value for t-BHP was approximately 10 times larger with the thioredoxin system (5.89 × 106 M−1 s−1) than with the glutathione system (3.48 × 105 M−1 s−1) (Tanaka, 2005; also see Table S1). These results indicate that Gpx2 prefers to use thioredoxin for the peroxidase reaction. Delaunay (2002) reported that Gpx3 uses only thioredoxin as an electron donor for the peroxidase reaction. However, we have demonstrated in this study that Gpx1 uses glutathione and thioredoxin almost equally as electron donors.
In the peroxidase reactions of SeCys-type GPxs, selenol (Se-H) at the active site is oxidized to selenenic acid (Se-OH), and glutathione is bound to Se-OH to form Se-GS. Subsequently, Se-GS is reduced by another molecule of glutathione to revert to Se-H, and consequently, glutathione disulfide (GS-SG) is formed. Meanwhile, for non-SeCys-type GPxs, two cysteine residues are involved in the peroxidase reaction, i.e. the first cysteine residue, referred to as the peroxidatic cysteine, is attacked by peroxide and oxidized, yielding sulfenic acid (Cys-SOH). Cys-SOH is reduced by another cysteine residue, referred to as the resolving cysteine, to form a disulfide bond between the peroxidatic cysteine and resolving cysteine. This disulfide bond is then reduced by thioredoxin. Because Gpx1 does not contain SeCys, cysteine residues are expected to be involved in the reducing reaction of peroxide using glutathione or thioredoxin. We then identified the active cysteine using the AMS assay. AMS is able to modify the sulfhydryl group (-SH) irreversibly, but not sulfenic acid (-SOH) and disulfide bonds (Hermanson, 1996). Gpx1 has six cysteines (Cys36, Cys64, Cys82, Cys98, Cys144, and Cys152) (Fig. 1a), and the molecular weight of AMS is approximately 500; therefore, the apparent molecular weight of Gpx1 theoretically increases to 3000 if all cysteines are in the reduced form and modified by AMS. In addition, the oxidized protein having disulfide bonds migrates faster than the reduced protein in nonreducing sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (Tanaka, 2005). Thus, Gpx1 in the reduced form and oxidized form is distinguishable by modification with AMS, followed by nonreducing SDS-PAGE. To verify the validity of the assay, bacterially expressed Gpx1 was oxidized by treatment with H2O2 or reduced with dithiothreitol, followed by modification with AMS, and SDS-PAGE was conducted. As shown in Fig. 1b, the reduced Gpx1 (i.e. modified with AMS) migrated slowly, whereas the oxidized Gpx1 migrated faster in SDS-PAGE. Hence, we introduced the plasmid GPX1-3HA into a gpx1Δ mutant and determined the redox status of the Gpx1 protein. As shown in Fig. 1b, Gpx1 was basically in the reduced form under normal conditions, although half of the protein was oxidized when cells were treated with H2O2 or t-BHP. Next, we determined the redox status of Gpx1 in a mutant defective in thioredoxin (gpx1Δtrx1Δtrx2Δ) to evaluate the role of thioreoxin in the redox regulation of Gpx1. The oxidized form of Gpx1 was seen in cells lacking the cytosolic thioredoxin, even though the cells were not treated with peroxides. This suggests that thioredoxin plays a role in the reduction of Gpx1 in vivo. This conclusion was supported by the result that Gpx1 was almost completely oxidized following treatment with H2O2 and t-BHP in the thioredoxin-deficient mutant (Fig. 1b). On the other hand, the glutathione/glutathione disulfide (GSH/GSSG) balance may be decreased in a trx1Δtrx2Δ mutant. Thus, we treated yeast cells with diamide, which oxidizes GSH to form GSSG. We determined the redox state of Gpx1 after the treatment of yeast cells with 1.5 mM diamide for 60 min; under these conditions, the GSH/GSSG balance was decreased as we have reported previously (Kuge, 2001). As shown in Fig. 1c, Gpx1 was essentially in the reduced form. Next, we determined the redox state of Gpx1 in a glr1Δ mutant. The GRL1 gene encodes a glutathione reductase that catalyzes the reduction of GSSG to GSH, and we have reported that the disruption of GLR1 enhanced the levels of GSSG in vivo (Izawa, 1998). Under normal conditions (no treatment with diamide), Gpx1 was in the reduced form. After treatment with diamide, Gpx1 was partially oxidized (Fig. 1c). In trx1Δtrx2Δ cells, the redox state of Gpx1 was not substantially affected by treatment with diamide (Fig. 1c). Therefore, we concluded that thioredoxin plays crucial roles in regulating the redox state of Gpx1 in vivo.
Redox regulation of Gpx1. (a) Schematic diagram of the positions of cysteines in yeast GPxs. The numbers indicate the positions of cysteine residues. A line between cysteine residues represents a disulfide bond. In Gpx1, one of three disulfide bonds, between either Cys36 and Cys64, Cys36 and Cys82, or Cys36 and Cys98, may be formed if Cys64, Cys82, or Cys98 is substituted with serine. (b) His-tagged Gpx1 expressed in Escherichia coli cell (Gpx1-6His) was treated with 5 mM dithiothreitol or 1 mM H2O2 for 1 h, and precipitated with a 20% trichloroacetic acid solution. The samples were then subjected to the AMS assay. The Gpx1 protein was detected using the anti-His-tag antibody in a Western blot analysis. pRS416-Gpx1WT-3HA (Gpx1-3HA) was introduced into gpx1Δ and gpx1Δtrx1Δtrx2Δ cells, and the cells were cultured in SD medium at 28°C. After 37 h, 0.4 mM H2O2 or 0.6 mM t-BHP was added, and cells were cultured for another 1 h at 28°C. Cells of 40 A610 units of the culture were collected, washed twice with a 20% TCA solution and once with acetone, and then subjected to the AMS assay. The Gpx1 protein was detected using an anti-HA monoclonal antibody in a Western blot analysis. (d) Each cysteine-substituted mutant of GPX1 harbored in pRS416 was expressed in gpx1Δtrx1Δtrx2Δ cells, and cells were treated with 0.6 mM t-BHP for 1 h. After the treatment, the AMS assay was conducted as described in (b). We repeated each experiment three times, and obtained essentially the same results.
In a typical 2-Cys peroxiredoxin, the disulfide bond is formed between two molecules of enzyme, whereas in an atypical 2-Cys peroxiredoxin, the disulfide bond is formed intramolecularly. Because no band whose molecular weight corresponds to a dimer of Gpx1 was seen in the AMS assay, Gpx1 seems to be an atypical 2-Cys peroxiredoxin like Gpx2 and Gpx3.
Next, to identify the cysteine residues involved in the redox regulation of Gpx1, we changed each cysteine residue of Gpx1 to serine, and introduced the mutants (Gpx1C36S, Gpx1C64S, Gpx1C82S, Gpx1C98S, Gpx1C144S, and Gpx1C152S) into gpx1Δtrx1Δtrx2Δ cells. Notably, no oxidized Gpx1 appeared in cells carrying the Gpx1C36S protein even in the presence of t-BHP (Fig. 1d). This indicates that Cys36 plays a crucial role in the determination of the redox status of Gpx1 in yeast cells. Additionally, the position of this cysteine residue is well conserved among yeast GPxs, i.e. Cys37 and Cys36 are the peroxidatic cysteines in Gpx2 and Gpx3, respectively (Delaunay, 2002; Tanaka, 2005) (Fig. 1a). Together, Cys36 is thought to be the peroxidatic cysteine of Gpx1. Meanwhile, the pattern of reduced/oxidized protein of Gpx1C144S and Gpx1C152S was similar to that of Gpx1WT, suggesting that Cys144 and Cys152 are not involved in the redox regulation of Gpx1.
Regarding the resolving cysteine, Cys83 and Cys82 are responsible in Gpx2 and Gpx3, respectively (Delaunay, 2002; Tanaka, 2005). Judging from an alignment of primary structure among S. cerevisiae GPx homologues, Cys82 seems to be the resolving cysteine of Gpx1. Actually, in contrast to Gpx1WT, no oxidized band of Gpx1C82S protein was seen in the AMS assay in thioredoxin-deficient cells under normal conditions (Fig. 1d), which suggests that Cys82 is involved in the redox regulation of Gpx1. However, because the oxidized band of Gpx1C82S protein merged following the treatment of the cells with t-BHP, another cysteine residue may be involved in the formation of the intramolecular disulfide bond. A similar pattern was seen in the Gpx1C64S and Gpx1C98S mutants, and so the possibility that these two cysteines are also responsible for the formation of a disulfide bond with Cys36 cannot be excluded, i.e. one of three disulfide bonds between either Cys36 and Cys64, Cys36 and Cys82, or Cys36 and Cys98 might be formed if Cys64, Cys82, or Cys98 was substituted with serine.
Table S1. Comparison of kinetic parameters of non-SeCys-type GPx from various sources.
Appendix S1. Detailed descriptions for enzyme purification and site-directed mutagenesis of Gpx1.
(2010) Regulatory mechanism for expression of GPX1 in response to glucose starvation and Ca2+ in Saccharomyces cerevisiae: involvement of Snf1 and Ras/cAMP pathway in Ca2+ signaling. Genes Cells 15: 59–75.