Glutathione oxidation in response to intracellular H₂O₂: Key but overlapping roles for dehydroascorbate reductases

ABSTRACT

Glutathione is a pivotal molecule in oxidative stress, during which it is potentially oxidized by several pathways linked to H₂O₂ detoxification. We have investigated the response and functional importance of 3 potential routes for glutathione oxidation pathways mediated by glutathione S-transferases (GST), glutaredoxin-dependent peroxiredoxins (PRXII), and dehydroascorbate reductases (DHAR) in Arabidopsis during oxidative stress. Loss-of-function gstU8, gstU24, gstF8, prxIIE and prxIIF mutants as well as double gstU8 gstU24, gstU8 gstF8, gstU24 gstF8, prxIIE prxIIF mutants were obtained. No mutant lines showed marked changes in their phenotype and glutathione profiles in comparison to the wild-type plants in either optimal conditions or oxidative stress triggered by catalase inhibition. By contrast, multiple loss of DHAR functions markedly decreased glutathione oxidation triggered by catalase deficiency. To assess whether this effect was mediated directly by loss of DHAR enzyme activity, or more indirectly by upregulation of other enzymes involved in glutathione and ascorbate recycling, we measured expression of glutathione reductase (GR) and expression and activity of monodehydroascorbate reductases (MDHAR). No evidence was obtained that either GRs or MDHARs were upregulated in plants lacking DHAR function. Hence, interplay between different DHARs appears to be necessary to couple ascorbate and glutathione pools and to allow glutathione-related signaling during enhanced H₂O₂ metabolism.

Redox signaling linked to reactive oxygen species such as H₂O₂ has become accepted as an important player during plant development and responses to the environment.¹ Thiol modifications are thought to be among the predominant mechanisms by which redox signaling occurs, and are controlled by components such as thioredoxins, glutaredoxins, and glutathione.^2-5 As the main non-protein thiol in cells, glutathione contributes to redox homeostasis and signaling under conditions of oxidative stress, when accumulation of the oxidized (disulfide) form (GSSG) is often observed.^6,7 Oxidation of glutathione to GSSG has been for many years thought to be coupled to ascorbate peroxidase (APX) activity through the ascorbate-glutathione pathway,⁸ either through chemical reduction of dehydroascorbate (DHA) or through DHA reductases (DHAR). However, evidence that this occurs in planta is mainly circumstantial and work over more recent years has identified several types of peroxidases that can link glutathione oxidation to H₂O₂ more directly.⁹ These include glutathione S-transferases (GSTs), some of which can act as glutathione peroxidases,¹⁰ and certain types of peroxiredoxin (PRXII) that are coupled to glutathione oxidation via glutaredoxins.¹¹

We have used loss-of-function mutants to evaluate the contribution of candidate enzymes to glutathione oxidation, within the wider objectives of understanding the complexity of H₂O₂ metabolism in plants and its coupling to oxidative signaling. Because the formerly named glutathione peroxidases are now considered to use thioredoxins as a reductant,^12,13 they are probably not major players in glutathione oxidation. Hence, we focused on available mutants for GSTs, PRXs and DHARs. These types of proteins are encoded, respectively, by about 50, 5, and 3 expressed genes in Arabidopsis. GSTs were originally considered to serve mainly in conjugation of endogenous and xenobiotic compounds detoxification until the discovery of their functions in preventing oxidative damage to cells.^10,14,15 Stress-inducible GSTs with glutathione peroxidase activity can protect plants from oxidative injury, strengthening the idea that GSTs have peroxidatic functions in addition to catalyzing the formation of GSH conjugates.^16,17 Such functions may contribute to accumulation in GSSG in conditions of stress. Similarly, the glutaredoxins that regenerate type II PRX may oxidize GSH to GSSG.^4,18

There have been very few gene-specific studies of Arabidopsis GSTs and PRXs that have reported data on effects on tissue glutathione contents. Loss of PRXIIF function slightly affected root glutathione pools in control conditions but during stress little difference from the wild-type was observed.¹⁹ We obtained T-DNA mutants for 2 type II PRX and 3 GSTs that are induced by oxidative stress⁹ (Supp. Fig. S1). RT-PCR analysis showed that the corresponding transcripts were absent in homozygotes (Supp. Fig. S2). We did not find effects of any of the mutations on rosette phenotype under optimal growth conditions (data not shown). To assess the roles of the candidate genes in GSSG accumulation triggered by oxidative stress, we used 3-aminotriazole (3-AT) to inhibit catalase, because catalase deficiency causes a well-defined oxidative stress that leads to glutathione accumulation as GSSG.⁶ 3-AT decreased catalase to very low levels while inducing ascorbate peroxidase (APX) and DHAR activities (Supp. Fig. S3). These effects were accompanied by extensive leaf bleaching, and glutathione oxidation was evident as marked accumulation of GSSG (Fig. 1). No difference was observed in bleaching or glutathione contents between the wild-type control and any of the mutants (Fig. 1). Thus, while some of these genes, such as GSTU24, are strongly induced by oxidative stress,⁹ our initial analysis provides little evidence that any of them are indispensable players in conditions where increased H₂O₂ availability is driving GSSG accumulation. This could reflect redundancy between gene functions within a particular family. As a first step to verify this point, we produced 3 double mutants from the 3 GSTs and a prxIIE prxIIF line. Like the single mutants, no effects on bleaching or glutathione oxidation were observed.

Figure 1.

Effect of 3-AT treatment on Col-0, gstU8, gstU24, gstF8, prxIIE, prxIIF, gstU8 gstU24, gstU8 gstF8, gstU24 gstF8 and prxIIE prxIIF phenotypes and leaf glutathione content. (A) Phenotype of plants 48 h after 3-aminotriazole (3-AT) treatment. Plants were grown at an irradiance of 200 µmol.m⁻².s⁻¹ in a 16h photoperiod and treated with 2 mM 3-AT after 3 weeks. Bar indicates 1 cm. (B) Bleached areas in the different genotypes as a percentage of the total rosette area. ND, not detected. Values are means ± SE of 15 plants. Bleaching was quantified using IQmaterials software. (C) Leaf contents of glutathione, measured as described previously.²⁰ White bars, reduced form. Black bars, oxidized form. Numbers above each bar indicate percentage reduction states [100 × (reduced form/total)] and are means of 3 biological replicates. Letters indicate significant difference at P< 0.05.

Clearly, further work would be required to examine a wider variety of conditions and to explore possible redundancy before the GSTs and GRX-PRXII systems can be ruled out as significant routes for glutathione oxidation. Nevertheless, our analyses of loss-of-function mutants for DHAR suggest that ascorbate regeneration is the major route leading to GSSG accumulation in response to intracellular H₂O₂.²⁰ A complete set of single, double and triple mutants carrying T-DNAs in DHAR1, DHAR2, and DHAR3 was obtained. No effect on phenotype was observed in the absence of stress. When the different dhar mutant combinations were introduced into a catalase-deficient background (cat2), the combined presence of dhar1 and dhar2 decreased GSSG and total glutathione accumulation (Fig. 2; Supp. Table S2). When all 3 DHARs were knocked out, cat2-triggered glutathione oxidation was almost completely inhibited (Fig. 2). Similar effects were observed in dhar1 dhar2 and dhar1 dhar2 dhar3 mutants using 3-AT to inhibit catalase in the same experimental conditions as those used for Fig. 1.²⁰ Interestingly, the effect on glutathione was not accompanied by significant effects on DHA or total ascorbate (Fig. 2; Supp. Table S2), suggesting that DHARs may be more important to ensure GSH oxidation than to maintain low DHA. Indeed, the downregulation of GSSG accumulation was accompanied by suppression of cat2-triggered lesions (Fig. 2), which are dependent upon both salicylic acid and glutathione.^21,22 This occurs even though H₂O₂-inducible genes such as APX1 and GSTU24 are not affected or even increased by the triple loss of DHAR function in the catalase-deficient background (Fig. 2). The major contribution to both lesion formation and glutathione oxidation triggered by catalase deficiency appears to come from DHAR1 and DHAR2, which we found to be cytosolic, with a minor but significant contribution from chloroplastic DHAR3. A key role for the cytosol in oxidative stress responses triggered by catalase deficiency is consistent with other observations.^7,23

Figure 2.

Phenotypes and accumulation of oxidants and oxidative stress marker transcripts implicate DHAR1 and DHAR2 as redundant players in glutathione oxidation in response to increased intracellular H₂O₂. The 2 triple mutants cat2 dhar1 dhar2, dhar1 dhar2 dhar3 and the quadruple mutant cat2 dhar1 dhar2 dhar3 were produced as described previously.²⁰ Plants were grown as in Fig. 1. Bars = 1 cm. Leaf contents of dehydroascorbate (DHA) and glutathione disulfide (GSSG) were measured as described previously. Values are means ± SE of 9 plants. APX1 and GSTU24 transcripts were quantified by qRT-PCR using 2 reference genes (ACTIN2 and RCE1). Data are means ± SE of 3 biological replicates and have been multiplied by 100 for ease of expression. Significant difference between mutants and Col-0 at P< 0.05 is indicated by asterisks while diamonds indicates significant difference between cat2 and triple and quadruple mutants in the cat2 background.

These observations and those we have recently reported elsewhere suggest that DHARs operate redundantly to ensure glutathione oxidation in response to intracellular H₂O₂. This is somewhat surprising because even if glutathione oxidation were entirely DHA-dependent, some might still be expected to occur independently of DHARs through the rapid chemical reaction between GSH and DHA.^8,24 Thus, it is possible that the decreased oxidation of glutathione when DHAR functions are lost is not a direct effect, but rather the result of secondary mechanisms. One possibility is that glutathione reductase (GR) is upregulated to keep glutathione more reduced. However, the triple dhar1 dhar2 dhar3 mutant does not have a marked increase in extractable GR, even in conditions of oxidative stress²⁰ and neither of the 2 Arabidopsis genes encoding this enzyme is upregulated in response to loss of DHAR function (Supp. Fig. S4). A second indirect mechanism could be marked upregulation of MDHAR activity, which could choke off supply of DHA produced during oxidative stress by competing with MDHA dismutation. Potentially, this could lead to lower accumulation of DHA and, thereby, lower accumulation of GSSG. MDHAR is encoded by 5 genes in Arabidopsis, 2 of which (MDHAR2, MDHAR3) give rise to cytosolic enzymes. As Fig. 3 shows, both these genes were markedly induced by oxidative stress in cat2, and this was accompanied by an increase in extractable MDHAR activity. No further increase in either expression or activity was detected in cat2 dhar1 dhar2 dhar3. Indeed, MDHAR activity was somewhat lower in the quadruple mutant than in cat2 (Fig. 3).

Figure 3.

Expression of the 2 cytosolic isoforms Arabidopsis genes encoding monodehydroascorbate reductase (MDHAR) and leaf MDHAR extractable activity in Col-0, dhar, and cat2 mutants. For the transcript abundance, data are expressed relative to ACT2 and RCE1, have been multiplied by 100 for ease of expression, and are means ± SE of 3 biological replicates from plants growing in conditions as described for Fig. 2. For MDHAR extractable activity, values are means ± SE of 4 plants. Significant difference between mutants and Col-0 at P< 0.05 is indicated by asterisks while diamonds indicates significant difference between cat2 and the quadruple cat2 dhar1 dhar2 dhar3.

Our observations provide further evidence that the ascorbate-glutathione pathway is a major player in removing intracellular H₂O₂, and identify key overlapping roles for DHAR in ensuring glutathione oxidation in these conditions. Other recent studies of single dhar mutants have also reported significant changes in glutathione oxidation during stress conditions.^25,26 Results shown here and elsewhere²⁰ underscore the role of DHARs in coupling H₂O₂ to glutathione oxidation during signaling while at the same time raising perspectives for further work. Questions notably include whether DHA is indeed the major physiological oxidant used by all 3 DHARs and the nature of the intracellular traffic that functionally links chloroplastic DHAR3 with cytosolic DHAR1 and DHAR2 during H₂O₂-triggered oxidation of glutathione.²⁰

Acknowledgments

We are grateful to the Salk Institute Genomic Analysis Laboratory for providing the sequence-indexed Arabidopsis T-DNA insertion mutants and to the Nottingham Arabidopsis Stock Center (UK) for supply of seed stocks. M-S.R. wishes to thank the Doctoral School ED129 “Sciences de l'Environnement d'Ile de France” for the kind award of a Bourse Ministérielle (PhD studentship). G.N. thanks the ANR (France) for financial support through the « Cynthiol » research program (grant no. ANR12–BSV6–0011).