The thioredoxin-independent isoform of chloroplastic glyceraldehyde-3-phosphate dehydrogenase is selectively regulated by glutathionylation Mirko Zaffagnini 1,2 , Laure Michelet 2 , Christophe Marchand 3 , Francesca Sparla 1 , Paulette Decottignies 3 , Pierre Le Mare ´ chal 3 , Myroslawa Miginiac-Maslow 2 , Graham Noctor 2 , Paolo Trost 1 and Ste ´ phane D. Lemaire 2 1 Laboratory of Molecular Plant Physiology, University of Bologna, Italy 2 Institut de Biotechnologie des Plantes, UMR 8618, CNRS ⁄ Universite ´ Paris-Sud, Orsay, France 3 Institut de Biochimie et Biophysique Mole ´ culaire et Cellulaire, CNRS ⁄ Universite ´ Paris-Sud, Orsay, France Glutathione represents the major low-molecular-weight thiol in most cells. In addition to its well-established role in cellular defense against oxidative stress, gluta- thione can also promote a reversible post-translational modification, termed protein glutathionylation [1,2]. This modification consists of the formation of a mixed disulfide between glutathione and cysteine residues of proteins. In mammals, glutathionylation occurs under oxidative stress conditions and may protect cysteines from oxidation to cysteine sulfinic (-SO 2 H) or sulfonic (-SO 3 H) acids. In fact, while oxidation into sulfinic or sulfonic groups is irreversible, 2-electron reduction of glutathionylated cysteines can regenerate protein thiols. Glutathionylation has been shown to alter, either posi- tively or negatively, the activity of several proteins [3–8]. Recent proteomic approaches allowed the Keywords Arabidopsis thaliana; Calvin cycle; GAPDH; glutathionylation; oxidative stress Correspondence S. D. Lemaire, Institut de Biotechnologie des Plantes, Ba ˆ timent 630, Universite ´ Paris-Sud, F-91405 Orsay Cedex, France Fax: +33 1 69153423 Tel. +33 1 69153338 E-mail: stephane.lemaire@u-psud.fr (Received 3 October 2006, revised 25 October 2006, accepted 7 November 2006) doi:10.1111/j.1742-4658.2006.05577.x In animal cells, many proteins have been shown to undergo glutathionyla- tion under conditions of oxidative stress. By contrast, very little is known about this post-translational modification in plants. In the present work, we showed, using mass spectrometry, that the recombinant chloroplast A 4 -glyceraldehyde-3-phosphate dehydrogenase (A 4 -GAPDH) from Arabid- opsis thaliana is glutathionylated with either oxidized glutathione or reduced glutathione and H 2 O 2 . The formation of a mixed disulfide between glutathione and A 4 -GAPDH resulted in the inhibition of enzyme activity. A 4 -GAPDH was also inhibited by oxidants such as H 2 O 2 . However, the effect of glutathionylation was reversed by reductants, whereas oxidation resulted in irreversible enzyme inactivation. On the other hand, the major isoform of photosynthetic GAPDH of higher plants (i.e. the A n B n -GAPDH isozyme in either A 2 B 2 or A 8 B 8 conformation) was sensitive to oxidants but did not seem to undergo glutathionylation significantly. GAPDH cata- lysis is based on Cys149 forming a covalent intermediate with the substrate 1,3-bisphosphoglycerate. In the presence of 1,3-bisphosphoglycerate, A 4 - GAPDH was fully protected from either oxidation or glutathionylation. Site-directed mutagenesis of Cys153, the only cysteine located in close proximity to the GAPDH active-site Cys149, did not affect enzyme inhibi- tion by glutathionylation or oxidation. Catalytic Cys149 is thus suggested to be the target of both glutathionylation and thiol oxidation. Glutathiony- lation could be an important mechanism of regulation and protection of chloroplast A 4 -GAPDH from irreversible oxidation under stress. Abbreviations BPGA, 1,3-bisphosphoglyceric acid; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GSH, reduced glutathione; GSSG, oxidized glutathione; ROS, reactive oxygen species; TRX, thioredoxin. 212 FEBS Journal 274 (2007) 212–226 ª 2006 The Authors Journal compilation ª 2006 FEBS identification of many proteins undergoing such post- translational modifications in mammalian and yeast cells [9–12]. One of the prominent glutathionylated pro- teins in mammalian cells under stress is the glycolytic enzyme, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), which is inactivated by glutathionylation, pre- sumably of its active-site Cys149 [13–17]. In addition to cytosolic GAPDHs, plants also contain chloroplast GAPDH isoforms that participate in the Calvin cycle by catalyzing the reduction of 1,3-bisphos- phoglyceric acid (BPGA) to glyceraldehyde-3-phosphate using NAD(P)H as a reductant. All GAPDHs, including chloroplastic isoforms, share a common reaction mech- anism based on a highly reactive cysteine (Cys149), which is made acidic by an interaction with His176 [18]. During the catalytic cycle, the highly reactive thiolated group of Cys149 (Cys-S – ) forms a thioacylenzyme inter- mediate by nucleophilic attack on the substrate [19,20]. As a side-effect, the acidic nature of Cys149 makes it particularly prone to oxidation and to other redox modi- fications of its thiol group [15,21,22]. In glycolytic mammalian GAPDH, these modifications include S-glu- tathionylation, S-nitrosylation [15,23–26] and formation of an intrasubunit disulfide with the neighboring Cys153 [27]. However, although chloroplasts are a major site of reactive oxygen species (ROS) production, particularly under photoinhibitory conditions [28], it is not known whether chloroplast GAPDH is also subject to thiol oxi- dation or glutathionylation, or any other redox reaction affecting catalytic Cys149. The major isoform of chloroplast GAPDHs in higher plants displays the A n B n structure and is regula- ted by metabolites and thioredoxin f (TRX f), a small disulfide oxidoreductase involved in the activation of the Calvin cycle by light [29–33]. B subunits are almost identical to A subunits, except for the presence of a C-terminal extension containing a pair of cysteines, which is the target of TRX regulation [34]. Oxidized thioredoxin and low NADP(H) ⁄ NAD(H) ratios pro- mote the aggregation of fully active A 2 B 2 -GAPDH into A 8 B 8 -GAPDH hexadecamers with partially inhib- ited NADPH-dependent activity, whereas the secon- dary activity with NADH as a coenzyme is constitutively low and not regulated [35]. A second iso- form of chloroplast GAPDH is homotetrameric (A 4 - GAPDH) and TRX-insensitive [36,37], but can be regulated by the redox-sensitive peptide, CP12, which promotes the formation of a supramolecular complex with phosphoribulokinase [35,38–40]. Besides regula- tion of chloroplast GAPDH by photosynthetically reduced TRX and pyridine nucleotides, a type of regu- lation that is primarily linked to light⁄ dark conditions in the chloroplast, post-translational modifications of catalytic Cys149, such as glutathionylation, might con- stitute novel mechanisms of GAPDH regulation under oxidative stress. Compared with nonphotosynthetic organisms, very little is known about glutathionylation in plants, although recent studies have allowed the identification of a number of plant proteins undergoing glutathiony- lation [41–46]. We have recently shown that chloro- plast f-type TRXs are modified by glutathionylation, resulting in less efficient activation of TRX-sensitive GAPDH and NADP-malate dehydrogenase in the light [46]. Fructose-1,6-bisphosphate aldolase, a mem- ber of the Calvin cycle, like GAPDH, has also been reported to undergo glutathionylation [44]. In order to investigate whether chloroplast GAPDH may undergo glutathionylation and other redox modi- fications, we examined the effect of oxidized glutathi- one (GSSG) and other oxidizing molecules, including H 2 O 2 , on recombinant Arabidopsis thaliana A 4 -GAPDH and on native spinach A 2 B 2 -GAPDH and A 8 B 8 -GAP- DH. We examined the effect of these modifications on the enzyme activities, their reversibility in the presence of reductants, and the protective effect of the substrate BPGA and cofactors. Glutathionylation was investi- gated by MALDI-TOF mass spectrometry and site- directed mutagenesis. The results indicate that glutathionylation might constitute a previously unde- scribed mechanism of redox regulation and ⁄ or protection against oxidative damage of chloroplastic A 4 -GAPDH. Results Inactivation of A 4 -GAPDH by GSSG and other oxidants, and protection by substrate and cofactors Incubation of recombinant Arabidopsis A 4 -GAPDH with 5 mm GSSG resulted in a rapid decrease in enzyme activity. Indeed, the NADPH-dependent activ- ity decreased to less than 20% after 10 min of incuba- tion (Fig. 1A) and identical results were obtained with NADH as the coenzyme (data not shown). The decrease in GAPDH activity displayed a linear rela- tionship with increasing GSSG concentrations in the 0–5 mm range (Fig. 1B). A slow, reproducible loss of activity, observed when the enzyme was incubated in buffer alone or with reduced glutathione (GSH) as a control, suggested that A 4 -GAPDH underwent sponta- neous inactivation upon dilution. The exposure of A 4 -GAPDH to 1 mm H 2 O 2 , CuCl 2 , or diamide also caused a rapid loss of enzyme activity (Fig. 2). Kinet- ics of inactivation were comparable for the three M. Zaffagnini et al. Glutathionylation of chloroplastic GAPDH FEBS Journal 274 (2007) 212–226 ª 2006 The Authors Journal compilation ª 2006 FEBS 213 different oxidants and led to a complete loss of activity within 2 min of incubation. In the presence of a 10- fold lower concentration of H 2 O 2 (0.1 mm, Fig. 2), the inactivation kinetics was comparable to that obtained in the presence of 5 mm GSSG (Fig. 1). These results are consistent with the extreme sensitivity to oxidants described for glycolytic GAPDH [15,18]. In order to determine if substrate and cofactors might protect A 4 -GAPDH from glutathionylation and ⁄ or oxidation, the enzyme was incubated with GSSG or diamide in the presence of BPGA or NADPH. Both the substrate and the cofactor appeared to protect efficiently the enzyme from the two inactivat- ing treatments (Fig. 3A). In contrast, protection against the small H 2 O 2 molecule was only observed in the pres- ence of BPGA (Fig. 3B). Moreover, in the presence of BPGA, but not in the presence of NADPH, the enzyme remained totally active during the time-course of incu- bation (Fig. 3), suggesting protection from the slow spontaneous inactivation of A 4 -GAPDH observed in the control samples (Fig. 1). Similar results were obtained with NADH as a cofactor (data not shown). Glutathionylation of chloroplastic A 4 -GAPDH by GSSG is reversible In order to test the reversibility of A 4 -GAPDH inacti- vation, the enzyme was treated with 10 mm dithiothrei- tol after pre-incubation with 5 mm GSSG, 1 mm diamide or 1 mm H 2 O 2. As shown in Fig. 4, enzyme inactivation caused by 1 mm diamide or H 2 O 2 could not be reversed by dithiothreitol, suggesting that these oxidants induced irreversible oxidation, possibly affect- ing the sulfhydryl group of the active site Cys149. Irre- versibly oxidized cysteine thiols are typically converted to sulfinic (-SO 2 H) or sulfonic (-SO 3 H) acids [22]. On the other hand, inactivation caused by GSSG was par- tially (60%) reversed by dithiothreitol, although the remaining 40% of the initial activity was still irrevers- ibly lost. These results clearly indicate that oxidants, such as diamide or H 2 O 2 , led to A 4 -GAPDH inactiva- tion by irreversible oxidation, whereas GSSG, at least in part, inhibited A 4 -GAPDH by a different and reversible mechanism, possibly involving glutathionyla- tion. A B GSSG Concentration (mM) 0246 A 4 nim01retfaytivitcaHDPAG- )%(tnemtaertGSSGfo 0 20 40 60 80 100 Time (min) 0246810121416 A 4 )%()HPDAN(ytivitcAHPDAG- 0 20 40 60 80 100 Buffer or GSH GSSG Fig. 1. Inactivation of A 4 -GAPDH by GSSG. (A) Time-dependent inactivation. A 4 -GAPDH was incubated with 5 mM GSSG (open cir- cles), or with tricine buffer as a control (closed circles). Results with 5 m M GSH were identical to those of controls. (B) Concentra- tion-dependent inactivation. A 4 -GAPDH was incubated with various concentrations of GSSG for 10 min at 25 °C. Aliquots of the incuba- tion mixtures were withdrawn at the indicated time points and the remaining NADPH-dependent activity was determined. Activity is given as a percentage of the initial activity. Time (min) 0246810121416 A 4 )%()HPDAN(ytivitcAHPDAG- 0 20 40 60 80 100 Buffer 0.1 m M H 2 O 2 1mM oxidant Fig. 2. Inactivation of A 4 -GAPDH by oxidants. Time-dependent inac- tivation by 1 m M H 2 O 2 ,1mM diamide or 1 mM CuCl 2 (open squares), 0.1 m M H 2 O 2 (closed squares) or tricine buffer as the control (closed circles). Aliquots of the incubation mixtures were withdrawn at the indicated time points and the remaining NADPH- dependent activity was determined. Activity is given as a percent- age of the initial activity. Glutathionylation of chloroplastic GAPDH M. Zaffagnini et al. 214 FEBS Journal 274 (2007) 212–226 ª 2006 The Authors Journal compilation ª 2006 FEBS In order to test this hypothesis, GSSG-treated A 4 - GAPDH was analyzed by MALDI-TOF mass spectro- metry (Fig. 5). A clear shift in molecular mass was observed after GSSG treatment. The shift of the major peak, corresponding to GAPDH subunit A (theoretical molecular mass 36 346 Da), is consistent with the pres- ence of one glutathione adduct per subunit (theoretical additional mass ¼ 305 Da). The other peaks observed correspond to matrix adducts on A 4 -GAPDH. More- over, upon the addition of dithiothreitol, the molecular mass of A 4 -GAPDH shifted back to the mass of the untreated protein. These results clearly demonstrate that Arabidopsis A 4 -GAPDH can undergo glutathiony- lation, and that dithiothreitol can reverse this post- translational modification. Glutathionylation of chloroplastic A 4 -GAPDH by GSH in the presence of low concentrations of hydrogen peroxide is fully reversible Besides thiol disulfide exchange mediated by GSSG, it has been reported that protein glutathionylation can also be achieved in the presence of GSH and oxidants, conditions that are believed to promote the conversion of protein thiols into sulfenic acids which then react with GSH to give rise to mixed disulfides [47–49]. Incubation of Arabidopsis A 4 -GAPDH with 0.1 mm H 2 O 2 and 0.5 mm GSH (Fig. 6A) resulted in a rapid decrease of enzyme activity with kinetics comparable to that obtained in the presence of 0.1 mm H 2 O 2 alone (Table 1). Moreover, as in the case of 0.1 mm H 2 O 2 alone, BPGA appeared to provide full protection from inactivation, whereas almost no protection was observed in the presence of NADPH. The addition of dithiothreitol provided nearly 50% recovery of the initial activity of samples inactivated by 0.1 mm H 2 O 2 (Fig. 6B). This partial recovery, not observed after treatment with 1 mm H 2 O 2 (Fig. 4), indicated that part of the A 4 -GAPDH molecules were reversibly oxidized, probably to sulfenic acid (-SOH) or, by analogy to animal GAPDH [27], an intrasubunit disulfide with Cys153 (conserved in chloroplast GAPDH [50]) might have formed. Indeed, both sulfenic acids and disulfides are generally reduced by dithiothreitol. On the other hand, an almost complete recovery of the initial A B Time (min) 0246810121416 A 4 )%()HPDAN(ytivitcAHPDAG- 0 20 40 60 80 100 BPGA then H 2 O 2 (0.1 mM) NADPH then H 2 O 2 (0.1 mM) H 2 O 2 (0.1 mM) Time (min) 0246810121416 A 4 )%()HPDAN(ytivitcAHPDAG- 0 20 40 60 80 100 BPGA then (GSSG or Diamide) NADPH then (GSSG or Diamide) GSSG Diamide Fig. 3. Protection of A 4 -GAPDH by BPGA and NADPH. A 4 -GAPDH was incubated in the presence of (A) 5 m M GSSG or 1 mM diamide in the presence of BPGA (open triangles) or 0.2 m M NADPH (closed triangles) at 25 °C, or (B) 0.1 m M H 2 O 2 in the presence of BPGA (open diamonds) or 0.2 m M NADPH (closed diamonds). Inac- tivation by 5 m M GSSG (open circles), 1 mM diamide (closed circles), or 0.1 m M H 2 O 2 (open squares) are presented for compar- ison. NADPH-dependent activity is given as a percentage of the initial activity. A 4 )%()HPDAN(ytivitcAHPDAG- 0 20 40 60 80 100 +dithiothreitol +dithiothreitol +dithiothreitol GSSG 5 m M Diamide 1 m M H 2 O 2 1 mM Fig. 4. Reversal, by dithiothreitol, of the inactivation of pretreat- ed A 4 -GAPDH. A 4 -GAPDH was incubated with 5 mM GSSG, 1 mM diamide or 1 mM H 2 O 2 for 10 min and subsequently treated with 10 m M dithiothreitol for 10 min at 25 °C. The remaining NADPH- dependent activity was determined before (black bars) and after (white bars) treatment with dithiothreitol. Activities are given as a percentage of the initial activity measured before the inactivation treatment. M. Zaffagnini et al. Glutathionylation of chloroplastic GAPDH FEBS Journal 274 (2007) 212–226 ª 2006 The Authors Journal compilation ª 2006 FEBS 215 activity upon the addition of dithiothreitol was observed for A 4 -GAPDH samples treated with 0.1 mm H 2 O 2 and 0.5 mm GSH, either in the presence or absence of NADPH. This suggested that, under the latter condi- tions, the mechanism of inactivation might have been different a n d may involve glutathionylation. This h ypo- thesis wa s confirmed by MALDI-TOF ma ss spectro- metry, which demonstrated a 305 Da mass increase for  50% of the A 4 -GAPDH subunits in the samples trea- tedwith0.1mm H 2 O 2 and 0.5 mm GSH (Fig. 7). Hence, these results clearly d emonstrate t hat A 4 -GAPDH can undergo glutathionylation in the presence of low concentrations of H 2 O 2 and GSH. Moreover, the com- plete recovery of the initial activity after dithiothreitol treatment indicates that glutathionylation is able to protect A 4 -GAPDH from irreversible oxidation. Identification of the cysteine residue involved in the inactivation of A 4 -GAPDH The observation that BPGA fully protects GAPDH from oxidation and ⁄ or glutathionylation suggests indeed that cysteines of the active site are the targets of the redox modification. Besides Cys149, which forms a covalent thioacylenzyme with BPGA during the catalytic cycle, Cys153 is also in proximity to the active site, whereas the remaining three cysteines of Arabidopsis A 4 -GAPDH reside in different regions of the protein [51]. In theory, inactivation of A 4 -GAPDH by glutathione and ⁄ or oxidants could thus depend on either Cys149 or Cys153 being redox modified. High performance liquid chromatography and MALDI- TOF mass spectrometry analyses of tryptic peptides of A 4 -GAPDH could not clarify this ambiguity because both Cys149 and Cys153 belong to a single long pep- tide which was very weakly desorbed ⁄ ionized from the matrix. This point was thus addressed by site-directed mutagenesis of Cys153. Mutation of Cys149 was not attempted, because this residue is known to be abso- lutely essential for catalysis [20]. On the other hand, the activity of the purified recombinant C153S mutant was similar to that of the wild-type protein (data not shown), and the kinetics of inactivation in the presence of 1 mm H 2 O 2 , 0.1 mm H 2 O 2 , or 0.1 mm H 2 O 2 + 0.5 mm GSH were identical for both proteins (Fig. 8A,B and Table 1). Similarly to wild-type GAPDH, BPGA (but not NADPH) protected the mutant from irreversible oxidation by H 2 O 2 and glutathionylation by H 2 O 2 + GSH. MALDI-TOF mass spectrometry confirmed that C153S A 4 -GAPDH underwent gluta- thionylation after treatment with 0.1 mm H 2 O 2 + 0.5 mm GSH (Fig. 9). The spectra are comparable to those obtained for the wild-type enzyme, showing a similar reversion of the 305 Da shift after dithiothreitol treatment. Overall, these results rule out the possibility that Cys153 might be the target of glutathionylation or % Intensity B +GSSG dithiothreitol 36400 36353 36663 36800 37200 3800037600 36000 36400 36800 37200 3800037600 36000 % Intensity A 0 20 40 60 80 100 0 20 40 60 80 100 Mass (m/z) Mass (m/z) Fig. 5. MALDI-TOF mass spectrometry indi- cates that A 4 -GAPDH undergoes glutath- ionylation. Mass spectra of GSSG treated A 4 -GAPDH were performed before and after treatment with 10 m M dithiothreitol (20 min). After dithiothreitol treatment (A), A 4 -GAPDH was at its expected mass (36 346 Da) and shows a decrease of 310 Da in comparison with the spectrum before dithiothreitol treatment (B). Accuracy of the measurement is ± 7 Da (0.02%). Glutathionylation of chloroplastic GAPDH M. Zaffagnini et al. 216 FEBS Journal 274 (2007) 212–226 ª 2006 The Authors Journal compilation ª 2006 FEBS irreversible oxidation. Although a conformational change after glutathionylation of a cysteine far distant from the active site cannot be completely excluded, the results strongly suggest that Cys149 is the target of chloroplast GAPDH glutathionylation or irreversible oxidation. Nonetheless, Cys153 seemed to play some role in GAPDH protection against irreversible oxidation. In fact, the activity of mutant C153S treated with 0.1 mm H 2 O 2 + 0.5 m m GSH could not be completely recov- ered by adding dithiothreitol, even when glutathionyla- tion was performed in the presence of NADPH (compare Figs 8C and 6B). This indicates that under mild oxidizing conditions (0.1 mm H 2 O 2 ), mutant C153S was not completely protected by GSH and par- tially underwent irreversible oxidation. Moreover, the recovery by dithiothreitol after treatment with 0.1 mm H 2 O 2 alone was also much lower in mutant C153S than in wild-type A 4 -GAPDH (Figs 8C and 6B, respectively). It is possible that the role played by Cys153 in preventing the irreversible oxidation of Cys149 might depend on the formation of a (transient) disulfide between Cys149 and Cys153, as observed in human GAPDH [27]. A n B n -GAPDH is sensitive to oxidants, but is not significantly prone to glutathionylation Besides A 4 -GAPDH, the major chloroplast GAPDH isoform of higher plants comprises A and B subunits in a stoichiometric ratio and is regulated by thioredoxins and metabolites [35]. As attempts to produce recombin- ant Arabidopsis A n B n GAPDH were unsuccessful, we purified native A n B n GAPDH from spinach chloroplasts in order to test its sensitivity to oxidants and its ability to undergo glutathionylation. A n B n -GAPDH exists in different conformations, and experiments were conduc- ted on either fully active A 2 B 2 -GAPDH (conformation prevailing in chloroplasts in the light) or A 8 B 8 -GAPDH (‘dark’ conformation with partially inhibited NADPH- dependent activity). Although both forms were inacti- vated by H 2 O 2 treatment, with or without GSH (Fig. 10A,B), hexadecameric A 8 B 8 -GAPDH was clearly less sensitive to H 2 O 2 than A 2 B 2 -GAPDH (Table 1). Similar results were obtained with NADH as a cofactor (data not shown), indicating that treatment with H 2 O 2 did not affect the redox regulation (mediated by C-ter- minal extensions) of the enzyme, a process which is thioredoxin-dependent and specific for the NADPH- linked activity of the enzyme [34]. Furthermore, BPGA protected A 2 B 2 -GAPDH from inactivation, as in the case of A 4 -GAPDH, suggesting that catalytic Cys149 was the target of the redox modification (Fig. 10A). Pro- tection by BPGA of the A 8 B 8 isoform could not be tested because BPGA incubation is known to convert the hexadecamer into active tetramers [32]. In order to test the reversibility of A n B n inactiva- tion, the enzymes were treated with 10 mm dithiothrei- tol for 10 min after the oxidative treatment. Similarly to A 4 -GAPDH, dithiothreitol treatment of A 2 B 2 GAPDH, after incubation with 0.1 mm H 2 O 2 , allowed A 4 )%()HPDAN(ytivitcAHPDAG- 0 20 40 60 80 100 A B H 2 O 2 (0.1 mM) H 2 O 2 (0.1 mM) +GSH H 2 O 2 (0.1 mM) + GSH + NADPH +dithiothreitol +dithiothreitol +dithiothreitol Time (min) 0246810121416 A 4 )%()HPDAN(ytivitcAHPDAG- 0 20 40 60 80 100 BPGA then H 2 O 2 (0.1 mM)+GSH H 2 O 2 (0.1 mM)+GSH NADPH then H 2 O 2 (0.1 mM)+GSH Fig. 6. Inactivation of A 4 -GAPDH in the presence of H 2 O 2 and GSH. (A) Protective effect of NADPH and BPGA. A 4 -GAPDH was incubated with 0.1 m M H 2 O 2 and 0.5 mM GSH alone (open trian- gles), or in the presence of 0.2 m M NADPH (closed circles) or BPGA (closed triangles). NADPH-dependent activity is given as a percentage of the initial activity. (B) Reversal of A 4 -GAPDH inactiva- tion by dithiothreitol. A 4 -GAPDH was inactivated by incubation with 0.1 m M H 2 O 2 , alone or in the presence of 0.5 mM GSH, or 0.2 mM NADPH and 0.5 mM GSH for 10 min at 25 °C and subsequently treated with 10 m M dithiothreitol for 10 min at 25 °C. The NADPH- dependent activity was determined before (black bars) and after (white bars) treatment with dithiothreitol. Activities are given as a percentage of the initial activity measured before the inactivation treatment. M. Zaffagnini et al. Glutathionylation of chloroplastic GAPDH FEBS Journal 274 (2007) 212–226 ª 2006 The Authors Journal compilation ª 2006 FEBS 217 a partial recovery of the initial enzyme activity (Fig. 10C). However, when dithiothreitol treatment was performed after inactivation with 0.1 mm H 2 O 2 + 0.5 mm GSH, the recovery was only slightly improved. This result contrasts with the almost complete recovery observed for A 4 -GAPDH in the same conditions (Fig. 6B) and suggests that A 2 B 2 GAPDH might not be significantly glutathionylated. In the case of A 8 B 8 GAPDH, the lower sensitivity of the enzyme to oxi- dants led to a more complete recovery of the initial activity after dithiothreitol treatment of samples trea- ted either with 0.1 mm H 2 O 2 alone or in the presence of 0.5 mm GSH (Fig. 10C). Therefore, A n B n -GAPDH samples incubated with 0.1 mm H 2 O 2 + 0.5 mm GSH, with or without subsequent dithiothreitol treatment, were analyzed by MALDI-TOF mass spectrometry (Fig. 11). No significant shift of the peak correspond- ing to B subunits (theoretical mass ¼ 39 357 Da) was Table 1. Half-time inactivation (in min) of wild-type A 4 -glyceraldehyde-3-phosphate dehydrogenase (A 4 -GAPDH), C153S A 4 -GAPDH, A 2 B 2 -GAPDH and A 8 B 8 -GAPDH under different oxidative treatments. Results are presented as mean ± standard deviation (SD), representa- tive of at least three independent experiments. Because A 2 B 2 -GAPDH and A 8 B 8 -GAPDH were prepared by the addition of 0.2 mM NADP(H) or NAD(H), respectively, the kinetics in the absence of cofactors could not be determined (ND). Oxidative Treatment 0.1 m M H 2 O 2 0.1 mM H 2 O 2 + 0.2 mM NAD(P)H H 2 O 2 0.1 mM + 0.5 mM GSH 0.1 m M H 2 O 2 + 0.5 mM GSH + 0.2 m M NAD(P)H Wild-type A 4 -GAPDH half time of inactivation (min) 2.28 ± 0.37 3.59 ± 0.29 1.64 ± 0.16 3.94 ± 0.23 C153S A 4 -GAPDH half time of inactivation (min) 2.04 ± 0.21 3.70 ± 0.10 1.87 ± 0.11 3.76 ± 0.81 A 2 B 2 -GAPDH half time of inactivation (min) ND 4.75 ± 0.66 ND 3.53 ± 0.11 A 8 B 8 -GAPDH half time of inactivation (min) ND 7.88 ± 0.31 ND 9.56 ± 1.32 % Intensity 0.54363 0.44663 2.14 3 63 3560035000 36200 36800 38000 37400 Mass (m/z) 35600 36200 36800 380003740035000 Mass (m/z) 0 20 40 60 80 100 % Intensity 0 20 40 60 80 100 H 2 O 2 + GSH +dithiothreitol Fig. 7. MALDI-TOF mass spectrometry indi- cates that A 4 -GAPDH also undergoes glu- tathionylation in the presence of H 2 O 2 and GSH. Mass spectra of A 4 -GAPDH treated with 0.1 m M H 2 O 2 and 0.5 mM GSH for 1 h at 25 °C were performed before and after a 10 m M dithiothreitol treatment (30 min at 25 °C). Accuracy of the measurement is ± 7 Da (0.02%). Glutathionylation of chloroplastic GAPDH M. Zaffagnini et al. 218 FEBS Journal 274 (2007) 212–226 ª 2006 The Authors Journal compilation ª 2006 FEBS observed after the glutathionylation treatments. For A subunits (theoretical mass ¼ 36 141 Da), besides the matrix adduct peak (sinapinic acid, theoretical mass increase 205 Da), a very discrete peak, corresponding to a 305 Da mass increase, was observed after treat- ment with H 2 O 2 and GSH, but disappeared after sub- sequent treatment with dithiothreitol. These results are consistent with the recovery of activity observed after dithiothreitol treatment (Fig. 10C). Therefore, it can be concluded that the B subunits of A n B n -GAPDH do not undergo glutathionylation and that glutathionyla- tion of the A subunits is very limited. Overall, these in vitro results indicate that glutathionylation does not seem to play a significant role in the protection of A n B n GAPDH isoforms against oxidative stress. Discussion The aims of the present study were to establish whe- ther chloroplastic GAPDH isoforms undergo gluta- thionylation and thiol oxidation, and to examine the effect of these modifications on enzyme activity. The results demonstrate that Arabidopsis A 4 -GAPDH can undergo glutathionylation and that this post-transla- tional modification results in inhibition of the enzyme activity. MALDI-TOF mass spectrometry revealed the presence of one glutathione adduct per A 4 -GAPDH subunit, which could be removed by dithiothreitol with a concomitant recovery of enzyme activity. Glutath- ionylation of the protein can occur either in the pres- ence of GSSG or in the presence of H 2 O 2 and GSH. A 4 -GAPDH can also be irreversibly inactivated by oxi- dants, including H 2 O 2 . The substrate BPGA, forming a covalent intermediate with catalytic Cys149, fully protects the enzyme from either oxidation or glutath- ionylation. Mutant C153S, having no other cysteines than Cys149 in the catalytic site, was oxidized and H 2 O 2 (0.1 mM) H 2 O 2 (0.1 mM) +GSH H 2 O 2 (0.1 mM) + GSH + NADPH H 2 O 2 (1 mM) A 4 )%( )HPDAN(yti v itcAS35 1C HPDAG- 0 20 40 60 80 100 Time (min) 0246810121416 0 20 40 60 80 A 4 )%()HPDAN(ytivitcAS351CHPDAG- 100 A B C BPGA then H 2 O 2 (0.1 mM)+GSH H 2 O 2 (0.1 mM)+GSH NADPH then H 2 O 2 (0.1 mM)+GSH Time (min) 0246810121416 A 4 )%()HPDAN(ytivitcAS351CHPDAG- 0 20 40 60 80 100 Buffer NADPH then H 2 O 2 (0.1 mM) H 2 O 2 (0.1 mM) H 2 O 2 (1 mM) +dithiothreitol +dithiothreitol +dithiothreitol +dithiothreitol Fig. 8. Inactivation of the C153S A 4 -GAPDH mutant and reversal by dithiothreitol. (A) Time-dependent inactivation of C153S A 4 -GAPDH in the presence of H 2 O 2 . C153S A 4 -GAPDH was incubated with 1m M H 2 O 2 (open circles), 0.1 mM H 2 O 2 alone (open squares) or in the presence of 0.2 m M NADPH (closed squares), or in tricine buf- fer as a control (closed circles). Aliquots of the incubation mixtures were withdrawn at the indicated time points and the remaining NADPH-dependent activity was determined. Activity is given as a percentage of the initial activity. (B) Time-dependent inactivation of C153S A 4 -GAPDH in the presence of H 2 O 2 and reduced glutathione (GSH). C153S A 4 -GAPDH was incubated either with 0.1 mM H 2 O 2 and 0.5 mM GSH alone (open triangles) or in the presence of 0.2 m M NADPH (closed diamonds) or BPGA (open diamonds). Aliquots of the incubation mixtures were withdrawn at the indica- ted time points and the remaining NADPH-dependent activity was determined. Activity is given as a percentage of the initial activity. (C) Reversal of C153S A 4 -GAPDH inactivation by dithiothreitol. C153S A 4 -GAPDH was inactivated by incubation with 1 mM H 2 O 2 , 0.1 m M H 2 O 2 , alone or in the presence of 0.5 mM GSH, or 0.2 mM NADPH and 0.5 mM GSH for 10 min at 25 °C, and subsequently treated with 10 m M dithiothreitol for 10 min at 25 °C. The NADPH- dependent activity was determined before (black bars) and after (white bars) treatment with dithiothreitol. Activities are given as a percentage of the initial activity measured before the inactivation treatment. M. Zaffagnini et al. Glutathionylation of chloroplastic GAPDH FEBS Journal 274 (2007) 212–226 ª 2006 The Authors Journal compilation ª 2006 FEBS 219 glutathionylated similarly to wild-type A 4 -GAPDH. Taken together, these results strongly suggest that Arabidopsis A 4 -GAPDH can be reversibly glutathionyl- ated and irreversibly oxidized on its catalytic Cys149. Moreover, analysis of the C153S mutant also suggests that the formation of a disulfide between catalytically essential Cys149 and the neighboring Cys153 could contribute to the protection of A 4 -GAPDH from irre- versible oxidation. By contrast, the results of the present study show that A n B n GAPDH isoforms, representing the major isoforms of photosynthetic GAPDH in chloroplasts of higher plants, though being sensitive to oxidants, do not undergo significant modification by glutathionyla- tion. As A and B subunits are almost identical, except for the C-terminal extension of subunits B, it is very likely that in A n B n -GAPDH isoforms, the C-terminal extension of subunits B could partially protect catalytic Cys149 from oxidation by H 2 O 2 and effectively pre- vent the attack of glutathione. The C-terminal exten- sion bears a pair of TRX-sensitive cysteines and allows A 2 B 2 -GAPDH to associate into hexadecamers in the presence of NAD(H) [35]. The low sensitivity of A 8 B 8 - GAPDH to H 2 O 2 alone, or in the presence of GSH, is thus likely to depend on partial steric protection of active sites within the hexadecamer. Because the results presented here were obtained in vitro, the question arises as to how important such modifications are in vivo. In all the initial experiments described above, we performed A 4 -GAPDH glutath- ionylation assays in the presence of 5 mm GSSG. This corresponds to classical conditions generally used to test protein glutathionylation in vitro. However, this concentration of GSSG is significantly higher than the estimated concentration in chloroplasts. Indeed, the concentration of the glutathione pool has been estima- ted to be between 1 and 4.5 mm in the stroma [52] and GSSG only represents  10% of this pool. All these considerations suggest that glutathionylation of A 4 - GAPDH, through nonenzymatic thiol disulfide exchange mediated by GSSG, is probably of limited physio- logical significance. The low efficiency of GSSG, as a mediator of protein glutathionylation, has been repor- ted previously [8,48]. Other possible mechanisms lead- ing to glutathionylation might involve more reactive oxidized forms of glutathione, such as S-nitrosoglu- tathione and glutathione disulfide S-oxide [8], or the initial oxidation of a protein thiol that would subse- quently react with GSH [47–49,53]. Which of these mechanisms determines glutathionylation of GAPDH in vivo remains to be determined. However, the results presented here clearly show that A 4 -GAPDH is 35000 35600 36200 36800 37400 38000 + dithiothreitol Mass (m/z) % Intensity 0 20 40 60 80 100 35000 35600 36200 36324.5 36633.4 36326.0 36800 37400 38000 Mass (m/z) % Intensity 0 20 40 60 80 100 H 2 O 2 + GSH Fig. 9. MALDI-TOF mass spectrometry indicates that C153S A 4 -GAPDH undergoes glutathionylation in the presence of H 2 O 2 and GSH. Mass spectra of C153S A 4 -GAPDH treated with 0.1 m M H 2 O 2 and 0.5 mM GSH for 1 h at 25 °C were performed before and after treatment with 10 m M dithiothreitol (30 min at 25 °C). The accuracy of the measurement is ± 7 Da (0.02%). Glutathionylation of chloroplastic GAPDH M. Zaffagnini et al. 220 FEBS Journal 274 (2007) 212–226 ª 2006 The Authors Journal compilation ª 2006 FEBS glutathionylated in the presence of physiologically rele- vant concentrations of H 2 O 2 and GSH [52,54], suggest- ing a mechanism of glutathionylation based on the primary oxidation of the catalytic Cys149 followed by reaction with GSH rather than GSSG. In vivo, such a mechanism would be favored under conditions of oxi- dative stress, leading to enhanced ROS production in the chloroplast, such as exposure to high light under unfavorable conditions for photosynthetic metabolism (e.g. cold or water stress). Considering the high sensi- tivity of A 4 -GAPDH to oxidation, the formation of a mixed disulfide between GSH and the active site cys- teine oxidized to sulfenic acid would prevent its irre- versible oxidation. The results we obtained in vitro indeed confirmed that GSH-mediated glutathionylation of A 4 -GAPDH can effectively protect the enzyme from irreversible oxidation. However, besides protecting A 4 -GAPDH, glutath- ionylation might play a more general role in the regu- lation of photosynthesis under stress. TRX f plays a major role in the regulation of Calvin cycle enzymes that are mostly inactive in the dark and are activated by TRXs under illumination [55]. Glutathionylation of a conserved extra cysteine of TRX f, distinct from the two active site cysteines, strongly decreases the ability of TRX f to activate target enzymes, including GAPDH isoforms containing B subunits [46]. This suggests that, under conditions leading to protein glu- tathionylation in the chloroplast (e.g. under conditions of enhanced ROS production), the activity of TRX- dependent Calvin cycle enzymes would be decreased. In particular, A n B n -GAPDH would be down-regulated A B C Time (min) 0 2 4 6 8 10 12 14 16 A 8 B 8 )%()HPDAN(ytivitcAHPDAG- 0 20 40 60 80 100 Buffer H 2 O 2 (0.1 mM) H 2 O 2 (1 mM) H 2 O 2 (0.1 mM)+GSH Time (min) 0 2 4 6 8 10 12 14 16 A 2 B 2 )%()HPDAN(ytivitcAHPDAG- 0 20 40 60 80 100 Buffer H 2 O 2 (0.1 mM) H 2 O 2 (1 mM) BPGA then H 2 O 2 (0.1 mM)+GSH H 2 O 2 (0.1 mM)+GSH )%()HPDAN(ytivitcAHPDAG 0 20 40 60 80 100 H 2 O 2 ) Mm 1. 0( H 2 O 2 )Mm 1 .0( HSG+ H 2 O 2 )Mm 1 ( A 2 B 2 A 8 B 8 H 2 O 2 ) Mm 1. 0( H 2 O 2 )Mm 1 .0( HSG + H 2 O 2 )Mm1( +dithiothreitol +dithiothreitol +dithiothreitol +dithiothreitol +dithiothreitol +dithiothreitol Fig. 10. Inactivation of A n B n -glyceraldehyde-3-phosphate dehydroge- nase (A n B n -GAPDH) and reversal by dithiothreitol. A time-dependent inactivation of nonaggregated A n B n -GAPDH (tetrameric form) by 1m M H 2 O 2 (open circles), 0.1 mM H 2 O 2 alone (open squares), 0.1 m M H 2 O 2 and 0.5 mM GSH, in the absence (open triangles) or presence of BPGA (open diamonds), or in tricine buffer as a con- trol (closed circles). Aliquots of the incubation mixtures were withdrawn at the indicated time points, and the remaining NADPH- dependent activity was determined. Activity is given as a percent- age of the initial activity. (B) Time-dependent inactivation of aggregated A n B n -GAPDH (hexadecameric form) by 1 mM H 2 O 2 (open circles), 0.1 mM H 2 O 2 alone (open squares) or in the pres- ence of 0.5 m M GSH (open triangles), or in K-phosphate buffer as the control (closed circles). Aliquots of the incubation mixtures were withdrawn at the indicated time points and the remaining NADPH-dependent activity was determined. Activity is given as a percentage of the initial activity. Note that the specific activity of A 8 B 8 -GAPDH with NADPH as the coenzyme is about fourfold lower than that of A 2 B 2 -GAPDH. (C) Reversal of A n B n -GAPDH inactivation by dithiothreitol. Nonaggregated and aggregated forms of GAPDH were inactivated by incubation with either 1 m M H 2 O 2 or 0.1 mM H 2 O 2 , with or without 0.5 mM GSH, for 10 min at 25 °C, and sub- sequently treated with 10 m M dithiothreitol for 10 min at 25 °C. The NADPH-dependent activity was determined before (black bars) and after (white bars) treatment with dithiothreitol. Activities are given as a percentage of the initial activity measured before the inactivation treatment. M. Zaffagnini et al. Glutathionylation of chloroplastic GAPDH FEBS Journal 274 (2007) 212–226 ª 2006 The Authors Journal compilation ª 2006 FEBS 221 [...]... A4-GAPDH is stable in the presence of high levels of GSH (data not shown), suggesting that an oxidoreductase, such as a glutaredoxin, might be required for reduction of the mixed disulfide In this hypothesis, glutaredoxins would be involved in a signaling pathway contributing to the redox regulation of the Calvin cycle by controlling deglutathionylation of A4-GAPDH and other glutathionylated enzymes in the. .. Clearly, further work is required to identify the factors that determine the distribution of reductants between the various photosystem I electron acceptors and to elucidate the complex interplay between the different redox factors that regulate chloroplast metabolism In any case, reducing power is required for the deglutathionylation of proteins and recovery of their 222 42000 Fig 11 MALDI-TOF mass spectrometry... cysteine The activity of another Calvin cycle enzyme, fructose-1,6-bisphosphate aldolase, could also be inhibited by glutathionylation [44] Hence, all these considerations suggest that glutathionylation could contribute to slowing down the Calvin cycle under stress A general down-regulation of the Calvin cycle under stress would allow redistribution of the reducing power in the chloroplast by decreasing... existence of chloroplast glutaredoxins in plants [62,63], and proteomic studies have identified several potential glutaredoxin targets, including the A-subunit of GAPDH, fructose-bisphosphate aldolase and other Calvin cycle enzymes [64] Thus, glutaredoxins could participate in the regulation of the Calvin cycle by controlling deglutathionylation of A4-GAPDH and other glutathionylated enzymes in the chloroplast... nonphotosynthetic organisms, glutaredoxins, small oxidoreductases of the thioredoxin superfamily using GSH as the electron donor, appear to be key regulators of protein glutathionylation [60] Indeed, glutaredoxins are very efficient catalysts of protein deglutathionylation, although other enzymes could also participate in these reactions [6,60,61] In addition, recent genomic analyses have suggested the existence... 2006 The Authors Journal compilation ª 2006 FEBS M Zaffagnini et al Experimental procedures Reagents Glutathione was purchased from Roche Diagnostics Corporation (Indianapolis, IN, USA) All other reagents were obtained from Sigma-Aldrich (St Louis, MO, USA) Expression and purification of Arabidopsis A4-GAPDH The sequence encoding the putative mature form of the A thaliana plastidial A4-GAPDH isoform. .. (2004) Emergence of new regulatory mechanisms in the Benson-Calvin pathway via protein–protein interactions: a glyceraldehyde-3-phosphate dehydrogenase ⁄ CP12 ⁄ phosphoribulokinase complex J Exp Bot 55, 1245–1254 40 Marri L, Trost P, Pupillo P & Sparla F (2005) Reconstitution and properties of the recombinant glyceraldehyde-3-phosphate dehydrogenase ⁄ CP12 ⁄ phospho- Glutathionylation of chloroplastic. .. structure of the non-regulatory A(4) isoform of spinach chloroplast glyceraldehyde-3-phosphate dehydrogenase complexed with NADP J Mol Biol 314, 527–542 Noctor G & Foyer CH (1998) ASCORBATE and GLUTATHIONE: keeping active oxygen under control Annu Rev Plant Physiol Plant Mol Biol 49, 249–279 FEBS Journal 274 (2007) 212–226 ª 2006 The Authors Journal compilation ª 2006 FEBS 225 Glutathionylation of chloroplastic. .. J (1989) Cloning and sequence analysis of cDNAs encoding the cytosolic precursors of subunits GapA and GapB of chloroplast glyceraldehyde-3-phosphate dehydrogenase from pea and spinach Plant Mol Biol 13, 81–94 32 Trost P, Scagliarini S, Valenti V & Pupillo P (1993) Activation of spinach chloroplast glyceraldehyde 3phosphate dehydrogenase Effect of glycerate 1,3bisphosphate Planta 190, 320–326 33 Baalmann... following the recommendations of the manufacturer Replicates All the results reported are representative of at least three independent experiments and expressed as mean ± standard deviation Acknowledgements We thank Eliane Keryer for expert technical assistance This work was supported, in part, by Agence Nationale de la Recherche Grant JC05-45751 (to SDL) and in ` part by Ministero dell’Istruzione, . The thioredoxin-independent isoform of chloroplastic glyceraldehyde-3-phosphate dehydrogenase is selectively regulated by glutathionylation Mirko. irre- versible oxidation. By contrast, the results of the present study show that A n B n GAPDH isoforms, representing the major isoforms of photosynthetic GAPDH in