Glutathionylation of cytosolic glyceraldehyde-3-phosphate dehydrogenase from the model plant Arabidopsis thaliana is reversed by both glutaredoxins and thioredoxins in vitro

Biochem. J. (2012) 445, 337–347 (Printed in Great Britain) doi:10.1042/BJ20120505 337

Glutathionylation of cytosolic glyceraldehyde-3-phosphate dehydrogenasefrom the model plant Arabidopsis thaliana is reversed by both glutaredoxinsand thioredoxins in vitroMariette BEDHOMME*†, Mattia ADAMO*, Christophe H. MARCHAND†, Jeremy COUTURIER‡, Nicolas ROUHIER‡,Stephane D. LEMAIRE†, Mirko ZAFFAGNINI*1 and Paolo TROST*1

*Laboratory of Molecular Plant Physiology, Department of Experimental Evolutionary Biology, University of Bologna, Via Irnerio 42, 40126 Bologna, Italy, †Laboratoire de BiologieMoleculaire et Cellulaire des Eucaryotes, FRE3354 Centre National de la Recherche Scientifique, Universite Pierre et Marie Curie, Institut de Biologie Physico-Chimique, 13 rue Pierre etMarie Curie, 75005 Paris, France, and ‡UMR1136 Interactions Arbres-Microorganismes, Lorraine University-INRA, IFR 110, 54506 Vandoeuvre-les-Nancy cedex, France

Plants contain both cytosolic and chloroplastic GAPDHs(glyceraldehyde-3-phosphate dehydrogenases). In Arabidopsisthaliana, cytosolic GAPDH is involved in the glycolytic pathwayand is represented by two differentially expressed isoforms(GapC1 and GapC2) that are 98% identical in amino acidsequence. In the present study we show that GapC1 is aphosphorylating NAD-specific GAPDH with enzymatic activitystrictly dependent on Cys149. Catalytic Cys149 is the only solvent-exposed cysteine of the protein and its thiol is relativelyacidic (pKa = 5.7). This property makes GapC1 sensitive tooxidation by H2O2, which appears to inhibit enzyme activityby converting the thiolate of Cys149 (–S− ) into irreversibleoxidized forms (–SO2

− and –SO3− ) via a labile sulfenate

intermediate (–SO− ). GSH (reduced glutathione) prevents thisirreversible process by reacting with Cys149 sulfenates to giverise to a mixed disulfide (Cys149–SSG), as demonstrated by

both MS and biotinylated GSH. Glutathionylated GapC1 can befully reactivated either by cytosolic glutaredoxin, via a GSH-dependent monothiol mechanism, or, less efficiently, by cytosolicthioredoxins physiologically reduced by NADPH:thioredoxinreductase. The potential relevance of these findings is discussedin the light of the multiple functions of GAPDH in eukaryoticcells (e.g. glycolysis, control of gene expression and apoptosis)that appear to be influenced by the redox state of the catalyticCys149.

Key words: Arabidopsis thaliana cytosolic isoform 1glyceraldehyde-3-phosphate dehydrogenase (AtGapC1), degluta-thionylation, glutathionylation, glycolysis, oxidative stress, redoxsignalling.

INTRODUCTION

In photosynthetic organisms ROS (reactive oxygen species)constitute a natural by-product of the oxygenic metabolismand are now recognized as key signalling molecules [1]. ROSproduction can also increase under stress conditions and triggeradaptive responses while altering the redox state of the cell[2]. Different types of redox modifications of protein thiols areinvolved in ROS-mediated signalling pathways [3]. Oxidizingconditions can induce the formation of intra- or inter-moleculardisulfide bridges between pairs of cysteine residues belongingto the same or different polypeptides. In plant cells thesethiol/disulfide equilibria appear to be chiefly controlled by anarray of Trx (thioredoxin) isoforms that are expressed in virtuallyevery cell compartment [4,5]. In chloroplasts, the redox state ofTrxs depends on the activity of photosystem I via ferredoxin:Trxreductase, whereas in other cell compartments, including thecytosol, NADPH-dependent Trx reductases play an analogousrole [4]. Single cysteine residues may also be redox-modifiedin other ways, for instance S-nitrosylated by physiological NOdonors such as GSNO (S-nitrosoglutathione) [6], or oxidizedby H2O2 or other ROS [3]. Primary oxidation of proteinthiols by H2O2 is reversible and leads to sulfenic acids (SOH),

but can be followed by further irreversible oxidations to sulfinicand sulfonic forms (SO2H and SO3H respectively) [3,7]. Littleis known about the stability of protein sulphenates in vivo, butrecently an atypical plant Trx (CDSP32) was shown to reducein vitro the sulfenic acid formed on the catalytic cysteine residueof methionine sulfoxide reductase B1 after reaction with itssubstrate (methionine sulfoxide) [8]. Glutathionylation is anothermajor modification of protein thiols and consists of the formationof a mixed disulfide with glutathione, often resulting from thereaction of cysteine sulphenates with GSH (reduced glutathione)[3,9]. Glutathionylated proteins are commonly reduced back byparticular Grxs (glutaredoxins) in a GSH-dependent catalyticcycle [3,10,11]. Although protein glutathionylation also occursunder basal conditions [3], it is particularly promoted in responseto oxidative stress conditions and can act as a mechanism ofprotection against irreversible oxidation of sensitive protein thiols,but also has a role in modulating protein-specific functions [3,12].

Trxs and Grxs are small oxidoreductases sharing the sameoverall tridimensional structure and a similar sequence motif inthe active site, WC[G/P]PC and CXX[C/S] respectively. The Trxfamily in land plants contains at least 20 members divided intoseven different classes (h, o, f, m, x, y and z), each of themcontaining one or more isoforms [4,5,13]. The Grx family is

Abbreviations used: BioGSH, biotinylated reduced glutathione; BPGA, 1,3-bisphosphoglycerate; DTNB, 5,5′-dithiobis-(2-nitrobenzoic acid); DTT,dithiothreitol; ET-1, endothelin 1; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; G3P, glyceraldehyde 3-phosphate; GR, glutathione reductase;GRS, reduced glutathione reducing system; Grx, glutaredoxin; GSH, reduced glutathione; GSSG, oxidized glutathione; IAM, iodoacetamide; LB, Luria–Bertani; MALDI–TOF, matrix-assisted laser-desorption ionization–time-of-flight; NEM, N-ethylmaleimide; NTR, NADPH:thioredoxin reductase; OPP, oxidativepentose phosphate; ROS, reactive oxygen species; t1/2 , half-time for reactivation; Trx, thioredoxin.

1 Correspondence may be addressed to either of these authors (email [email protected] or [email protected]).

c© The Authors Journal compilation c© 2012 Biochemical Society

Bio

chem

ical

Jo

urn

al

ww

w.b

ioch

emj.o

rg

http://getutopia.com/documents

338 M. Bedhomme and others

composed of approximately 30 members in land plants and isdivided into three classes, presenting different active-site motifsand often showing peculiar functional properties [3,12]. In plants,the ability to function as deglutathionylating enzymes appearsspecific to Grxs belonging to class I (or CPYC-type) withthe notable exception of class II (or CGFS-type) Grx3 fromChlamydomonas reinhardtii [14–18].

It is well established that plant Trxs have a prominent role in theredox regulation of a large number of proteins via dithiol/disulfideexchange reactions [4,5]. This function probably reflects theirspecificity towards intra- and inter-protein disulfide bonds andstrictly requires both active-site cysteine residues. However,studies proposed Trx as a catalyst of protein denitrosylationin human cells [19] and of protein deglutathionylation inSaccharomyces cerevisiae [20,21] and in Plasmodium falciparum[22]. To date, no data are available for plant Trxs.

In previous years, several studies adopted diverse methodswhich aimed at identifying proteins undergoing glutathionylationin photosynthetic organisms both in vivo and in vitro [23–26].These proteomic approaches allowed the identification of morethan 200 targets implicated in numerous cell processes, includingglycolysis. Among the putative targets involved in the glycolyticpathway, the two cytosolic isozymes of phosphorylatingGAPDH (glyceraldehyde-3-phosphate dehydrogenase), GapC1and GapC2, have been identified as glutathionylated proteinsfollowing an oxidative treatment [24]. In another study, cytosolicGAPDH from Arabidopsis was detected as the prominent targetof H2O2-dependent oxidation in protein extracts and total NADH-dependent GAPDH activity was inhibited by H2O2 and recoveredby either GSH or DTT (dithiothreitol) [27]. More recently,cytosolic GAPDH isozymes were inactivated by GSSG (oxidizedglutathione) or H2O2 plus GSH, but recovery with DTT wasonly partial [28]. Chloroplast GAPDH isoform A4 (GapA), whichhas different specificity for pyridine nucleotides and is involvedin the Calvin–Benson cycle, is also oxidized by H2O2 andglutathionylated [29].

GAPDH was also identified as a target of glutathionylationin non-photosynthetic organisms including yeast and human[30–32]. Interestingly, in animal cells glycolytic GAPDHfulfils several functions unrelated to glycolysis [33]. Thisfunctional versatility is largely based on diverse post-translational modifications (phosphorylation, acetylation andredox modifications of thiols) that deeply affect the behaviour ofthe protein. The enzymatic activity of GAPDH strictly dependson a single catalytic cysteine residue, and oxidative modificationsof this catalytic cysteine, including glutathionylation, not onlyinhibit enzyme activity, but also influence its alternativefunctions, including protection of telomeric DNA, regulationof gene expression and induction of apoptosis under stressconditions [33]. The possibility that different types of redoxregulation, particularly nitrosylation, may also trigger non-glycolytic functions of plant GAPDH has been widely speculated,but awaits experimental support.

In the present study, we have cloned, recombinantly expressedand purified one of the two phosphorylating cytosolic GAPDHisozymes (GapC1) from Arabidopsis thaliana. As expected,GapC1 is strictly specific for NADH and Cys149 is essentialfor catalysis. Although catalytic cysteine is the 155th residueof Arabidopsis GapC1, in the present study it is named Cys149 forcongruence with the first solved GAPDH structure (from Bacillusstearothermophilus, PDB code 2DBV; [34]) in which the catalyticcysteine is Cys149. Owing to the micro-environment surroundingthe active-site Cys149, this residue is much more acidic than freecysteines and is sensitive to oxidation by H2O2. GSH protectsGapC1 from irreversible oxidation by reacting with transient

Cys149 sulfenates and gives rise to a glutathionylated form (Cys149–SSG) that can be rescued by a plant cytosolic Grx. Interestingly,plant cytosolic Trxs of the h-type, physiologically reduced byNTR (NADPH:Trx reductase), could also contribute to GapC1recovery, revealing an unexpected role of Trxs in the control ofprotein glutathionylation.

EXPERIMENTAL

Materials and enzymes

PD-10 and NAP-5 columns were obtained from GE Healthcare.Biotin (EZ-link Sulfo-NHS-Biotin) and DTNB [5,5′-dithiobis-(2-nitrobenzoic acid)] were purchased from Pierce. GSH and GSSGwere purchased from either Roche Diagnostics or Sigma–Aldrich.All of the other chemicals were obtained from Sigma–Aldrich.

GR (glutathione reductase) and 3-phosphoglycerate kinasefrom S. cerevisiae and thrombin protease were purchased fromSigma–Aldrich. Recombinant wild-type Trxh1 and its active-sitemutants (Trxh1-C36S and Trxh1-C39S) from C. reinhardtii wereprepared as described previously [35]. Recombinant NTRb fromA. thaliana was purified as described previously [36]. Poplarrecombinant h-type Trxs (isoform h1, h2, h3 and h5) and GrxC1(wild-type and active-site mutants GrxC1-C31S and GrxC1-C34S) were expressed and purified as described previously [37–39].

Cloning and site-directed mutagenesis

The sequence encoding the cytosolic form of GAPDH fromA. thaliana [GapC1 cDNA At3g04120 provided by TAIR (TheArabidopsis Information Resource) Stanford, CA, U.S.A.] wascloned into pET-28a( + ) (Novagen) using the following twoprimers: 5′-TCCATATGGCTGACAAGAAGATTAGC-3′ and 5′-GCGGATCCTTAGGCCTTTGACATGTCGACG-3′ (NdeI andBamHI restriction sites underlined respectively). In thepET28/GapC1 construct, the cDNA sequence for GapC1 was inframe with an N-terminal His tag and a cleavable thrombin site.

The two cysteine residues in GapC1 at positions 149 and153 were mutated into serines by PCR site-directed mutagenesisusing the pET-28a/GapC1 as a template. The following pairs ofprimers were used according to the procedures of the QuikChangesite-directed mutagenesis kit (Stratagene): C149S-up, 5′-TC-TCCAACGCTAGCTCCACCACTAACTGCC-3′; C149S-down,5′-GCAGTTACTGGTGGAGCTAGAGTTGGAGAC-3′; C153S-up, 5′-GCTGCACCACTAACTCCCTTGCTCCCCTTGC-3′; andC153S-down, 5′-CAAGGGGAGCAAGGGACTTAGTGGTG-CAGC-3′ (the codons encoding the mutated amino acids areunderlined). All of the mutations were confirmed by DNAsequencing.

Expression and purification of recombinant proteins

For protein production, the Escherichia coli BL21(DE3) strainwas chemically transformed with the different recombinantvectors allowing expression of GapC1 and its mutants GapC1-C149S (C149S) and GapC1-C153S (C153S). Ampicillin-resistant clones were amplified in LB (Luria–Bertani) mediumsupplemented with 50 μg/ml kanamycin at 37 ◦C. Proteinexpression was induced at exponential phase by adding 100 μMIPTG (isopropyl-β-D-thiogalactopyranoside) for 16 h at 30 ◦C.The cultures were then centrifuged (4400 g for 15 min at 5◦C) andthe resulting pellets were resuspended in 20 mM Tris/HCl buffer(pH 7.9), 150 mM NaCl, 5 mM imidazole and 0.14 mM NAD+ .Cell lysis was performed by sonication (5×1 min with intervals of1 min) using a Branson Sonifier 450 apparatus (pulse intervals


Deglutathionylation of GapC1 339

Table 1 Values of specific activity and number of free thiols for wild-typeGapC1 and the GapC1-C149S and GapC1-C153S mutants

The reaction mixture for the determination of specific activities (reverse reaction) contained50 mM Tris/HCl (pH 7.5), 1 mM EDTA, 5 mM MgCl2, 3 mM 3-phosphoglycerate, 5 units/mlof 3-phosphoglycerate kinase, 2 mM ATP and 0.2 mM NADH. For more details, see theExperimental section. The number of free thiols was determined according to the DTNB procedureas described in the Experimental section. Values in parentheses represent the percentage of thespecific activity of GapC1 wild-type with a His tag. WT, wild-type.

Protein Specific activity (μmol/min/mg) Number of free thiols

GapC1 WT (with His tag) 55.5 +− 4.1 (100 %) 1.02 +− 0.07GapC1 WT (without His tag) 53.2 +− 3.7 (96 %) Not determinedGapC1-C153S 51.1 +− 2.5 (92 %) 0.99 +− 0.12GapC1-C149S 0.4 +− 0.1 (0.7 %) 0.13 +− 0.03

of 0.5 s and amplitude of 15%). The soluble and insolublefractions were separated by centrifugation for 30 min at 30000 gat 5◦C. The soluble part was then applied on to an Ni2 +

HiTrap chelating resin (His-Bind Resin, Novagen) and proteinpurification was performed according to the manufacturer’sinstructions.

The subunit molecular mass and purity of the protein wereanalysed by SDS/PAGE (12% gel) and Coomassie Blue stainingafter desalting on PD-10 columns equilibrated with 50 mMpotassium phosphate buffer (pH 7.5) (Supplementary Figure S1at http://www.BiochemJ.org/bj/445/bj4450337add.htm). Theprotein concentration (on a subunit basis) of both wild-typeGapC1 and mutants was determined spectrophotometrically usinga molar extinction coefficient at 280 nm of 40910 M− 1·cm− 1

[40]. The final yield for all recombinant proteins was very high,being approximately 40–50 mg of proteins (wild-type and bothmutants) per litre of LB medium. The resulting homogeneousprotein was stored at − 20 ◦C. For some sample proteins,the N-terminal His tag was removed by thrombin protease asdescribed previously [41]. The excised His tag was eliminatedby metal-affinity chromatography and the resulting proteins weredesalted and stored as described above. However, the activity ofrecombinant GapC1 was not affected by the His tag (Table 1)and we could not observe any functional difference between His-tagged GapC1 and GapC1 preparations in which the His tag wasremoved.

Activity assay

GapC1 activity was monitored spectrophotometrically at 340 nmand 25 ◦C by following the forward (glycolytic) or reverse(gluconeogenic) reaction assays. The reverse reaction wasmeasured in an assay mixture containing 50 mM Tris/HCl(pH 7.5), 1 mM EDTA, 5 mM MgCl2, 3 mM 3-phosphoglycerate,5 units/ml of S. cerevisiae 3-phosphoglycerate kinase, 2 mM ATPand 0.2 mM NADH, whereas the forward reaction was measuredin an assay mixture containing 50 mM Tris/HCl (pH 7.5), 1 mMG3P (glyceraldehyde 3-phosphate), 1 mM NAD+ and 10 mMsodium arsenate. The forward assay used sodium arsenate insteadof phosphate as a co-substrate to form unstable 1-arseno,3-phosphoglycerate. Degradation of the product allows a favourableequilibrium for measuring the rate of GAPDH activity in theforward direction.

Reduction of GapC1 and determination of free thiol groups

Before each treatment, wild-type GapC1 and its cysteine variantswere incubated with 10 mM DTT for 30 min at room temperature

(25◦C) for reduction. DTT was subsequently removed bydesalting on NAP-5 columns equilibrated with 50 mM Bis-Tris buffer (pH 7.0). The number of free thiols was determinedspectrophotometrically with DTNB, as described previously [42].

Inactivation of GapC1 by alkylating treatments

Reduced GapC1 (2.0 μM or 78 μg/ml final concentration) wasincubated in 50 mM sodium citrate (pH 5.0) or in 50 mM Bis-Tris(pH 7.0) or in 50 mM glycine (pH 9.0) in the presence of 0.2 mMNEM (N-ethylmaleimide) or IAM (iodoacetamide). All of thereaction mixtures contained 0.14 mM NAD+ . At the indicatedtime, an aliquot of the sample (5 μl) was withdrawn for the assayof enzyme activity.

Determination of the pK a value of the catalytic cysteine residue ofGapC1

The pH-dependence of the inactivation of GapC1 by IAM wascarried out by following a procedure described in a previous study[18]. Briefly, the reduced protein (2 μM) was incubated with orwithout IAM (0.2 mM) for 20 min in different buffers with apH range from 4 to 10. After incubation, GapC1 activity wasdetermined as just described. The residual activity expressed asa percentage of maximal activity was plotted against pH, andthe pKa value was calculated by fitting the experimental datato a derivation of Henderson–Hasselbalch equation as describedpreviously [18].

Inactivation of GapC1 by oxidant treatments

Reduced GapC1 (2.5 μM or 97 μg/ml final concentration) wasincubated in 50 mM Bis-Tris buffer (pH 7.0) supplemented with0.14 mM NAD+ in the presence of different concentrations ofGSSG, H2O2 or H2O2 plus GSH at the indicated concentrations. Atthe indicated times, an aliquot of the sample (5 μl) was withdrawnfor the assay of enzyme activity. Substrate protection wasperformed by pre-incubating (5 min) the protein in the presenceof a BPGA (1,3-bisphosphoglycerate)-generating system (3 mM3-phosphoglycerate, 5 units/ml of 3-phosphoglycerate kinase and2 mM ATP) or in the presence of 1 mM G3P. The reversibility ofGapC1 inactivation (10 min incubation with 50 μM H2O2 aloneor in the presence of 0.5 mM GSH) was assessed by measuringGapC1 activity after incubation for 10 min in the presence of20 mM DTT.

Synthesis of biotinylated glutathione

BioGSH (biotinylated GSH) was prepared as follows. EZ-LinkSulfo-NHS-Biotin, a soluble biotinylation reagent, was usedto couple biotin to the primary amino group of GSH. Thebiotinylation reagent (300 μl, 25 mM) was added to GSH (300 μl,25 mM) in 50 mM potassium phosphate buffer, pH 7.2, andthe mixture was left to derivatize for 1 h at room temperature.After incubation, unreacted biotin was quenched by adding 83 μlof 0.6 M ammonium bicarbonate buffer. The concentration ofBioGSH was determined spectrophotometrically with DTNBusing a molar extinction coefficient at 412 nm of 14150 M−1·cm−1

to calculate the free thiol content of BioGSH.

In vitro glutathionylation of GapC1 and C153S mutant using BioGSH

Reduced proteins were incubated in 50 mM Tris/HCl (pH 7.9)and 0.14 mM NAD+ in the presence of 0.5 mM BioGSH aloneor supplemented with 50 μM H2O2. After a 1 h incubation,


http://www.BiochemJ.org/bj/445/bj4450337add.htm


BioGSH-treated samples were alkylated in the presence of100 mM IAM or treated with 20 mM DTT for 30 min to assess thereversibility of the reaction. All of the treatments were performedat room temperature. Proteins were then loaded on to non-reducing SDS/PAGE and analysed by Western blotting usinga primary mouse anti-biotin antibody (1:5000 dilution; Sigma)and an anti-mouse secondary antibody coupled to peroxidase(1:10 000 dilution; Sigma). Signals were visualized by enhancedchemiluminescence, as described previously [42].

MALDI–TOF (matrix-assisted laser-desorptionionization–time-of-flight) MS

Wild-type GapC1 and GapC1-C149S mutant were treated for30 min with 50 μM H2O2 in the presence of 0.5 mM GSH toinduce glutathionylation. The MALDI-TOF mass spectrometryanalyses were performed before and after incubation with 20 mMDTT for 30 min. The samples were analysed as described in [43].

Reactivation of glutathionylated GapC1

Glutathionylated GapC1 was prepared by incubating the enzymefor 20 min in the presence of 50 μM H2O2 and 0.5 mMGSH at room temperature and then desalted on a NAP-5column equilibrated with 50 mM Tris/HCl (pH 7.5) and 1 mMEDTA. Reactivation treatments of glutathionylated GapC1 wereperformed as reported previously [14,15,42]. Briefly, the reactionmixtures contained GapC1 and BPGA in the presence of 20 mMDTT or in the presence of 0.5 mM DTT supplemented withdifferent Trx or Grx isoforms at the indicated concentrations.The reactivation of GapC1 was also carried out in the presence ofGrx isoforms supplemented with a GRS (GSH reducing system;2 mM GSH, 6 μg/ml yeast GR and 0.2 mM NADPH) or in thepresence of Trx isoforms supplemented with 0.22 μM A. thalianaNTRb and 0.2 mM NADPH. Control samples were analysedunder similar conditions by omitting Trxs or Grxs. At the indicatedtimes, aliquots (5 μl) were withdrawn in order to assay enzymeactivity.

In order to investigate the molecular mechanism involved inTrx-dependent deglutathionylation of GapC1, the reactivationof glutathionylated GapC1 was performed by incubating theenzyme in the presence of different amounts of pre-reducedChlamydomonas Trxh1 and variants Trxh1-C36S and Trxh1-C39S. Wild-type Trxh1 and its variants were reduced in thepresence of 5 mM DTT and desalted on NAP-5 columnsequilibrated with 50 mM Tris/HCl (pH 7.5) and 1 mM EDTA.After different times of incubation, the reaction samples wereeither assayed for enzymatic activity or treated with 20 mM NEMand analysed by non-reducing SDS/PAGE.

Replicates

All of the results reported are representative of at least threeindependent experiments and expressed as means +− S.D.

RESULTS

Catalytically essential Cys149 is the only solvent exposed cysteineresidue of GapC1

In A. thaliana, cytosolic phosphorylating GAPDH is representedby two isozymes, namely GapC1 (locus At3g04120) andGapC2 (locus At1g13440). Although microarray data revealobvious differences in the expression of these two genes(www.genevestigator.com), the sequence alignment indicates thatGapC1 and GapC2 are 98% identical at the protein level, and

probably share comparable biochemical properties. Alignment ofGapC1 and GapC2 with GAPDH from other organisms confirmedthe absolute conservation of the catalytic Cys149 and also Cys153

and His176 (Supplementary Figure S2 at http://www.BiochemJ.org/bj/445/bj4450337add.htm)

Wild-type GapC1 from A. thaliana was expressed in E.coli and purified to homogeneity. Purified GapC1 exhibited asubunit molecular mass of approximately 40 kDa as expected(Supplementary Figure S1), and a specific activity (reversedirection) of 55 μmol·min-1·mg-1 with NADH as a coenzyme(Table 1), whereas the activity with NADPH was negligible(0.15% of the NADH-dependent specific activity). In order tocheck the importance of catalytic Cys149 in GapC1, and to test thepossible role for Cys153, site-specific mutants C149S and C153Swere generated. In terms of specific activity, the C153S mutantbehaved like the wild-type enzyme, whereas the C149Svariant was found to be virtually inactive (Table 1). The lackof activity of the C149S mutant is in full agreement with currentmodels of GAPDH catalysis which imply the acylation of thethiolate Cys149 by the substrate as an essential step of the catalyticcycle [44]. On the other hand, Cys153 has apparently no role incatalysis, in spite of the proximity in space with catalytic Cys149.

Cys149 and Cys153 are the only two cysteine residues ofthe GapC1 (and GapC2) sequence (Supplementary Figure S2).Following extensive reduction of the protein, the accessibilityof cysteines to the thiol reactant DTNB was determined undernon-denaturing conditions. A single accessible thiol was found inboth the wild-type and C153S proteins, whereas the C149S mutanthad none (0.13 +− 0.03) (Table 1). That Cys149 is both essential forcatalysis and the only accessible cysteine of GapC1 is a propertythat can be exploited for further characterization of the protein.

Catalytic Cys149 is an acidic cysteine

The alkylating agents NEM and IAM both cause rapid andcomplete inhibition of GapC1 activity (Figures 1A and 1B).Since IAM preferentially alkylates thiolates (–S− ) rather thanprotonated thiols (–SH) [18,45], the quantitative reaction betweenIAM and GapC1 was used to determine the pH-dependency ofenzyme inhibition and thus derive a pKa value that could beattributed to catalytic Cys149. As expected, the inhibition of GapC1activity after incubation with IAM at different pH values increasedwith increasing pH (i.e. increasing deprotonation of Cys149). Byplotting the residual activity against the pH, a sigmoidal titrationcurve was obtained allowing extrapolation of a pKa value of5.65 +− 0.03 (Figure 1C). A comparable value was determinedfor B. stearothermophilus GAPDH [44].

GapC1 is irreversibly inhibited by H2O2-dependent oxidation ofCys149

From its solvent accessibility and acidic nature, catalytic Cys149

might be sensitive to oxidative modifications in the presence ofROS, particularly H2O2, which is known to react with cysteinethiolates. Indeed, incubation of GapC1 with 50 μM H2O2 causeda strong inhibition of enzyme activity (Figure 2A). The presenceof the substrates (BPGA in the reverse reaction and G3P inthe forward reaction), which covalently bind to catalytic Cys149,allowed full protection of enzyme activity (Figure 2B), inagreement with Cys149 being the target of H2O2 modification.On the other hand, the cofactor NAD+ , which was always presentin our incubation medium, did not protect GapC1 from the attackof H2O2 (Figure 2A). Reactions of H2O2 with cysteine thiolatesinclude a primary oxidation to sulfenates (–SO− ) that can befurther irreversibly oxidized to sulfinates (–SO2

− ) and sulfonates





Figure 1 Alkylation treatments of GapC1

Kinetics of inactivation of GapC1 in the presence of alkylating agents. Reduced enzyme (2 μM) was incubated with 0.2 mM NEM (A) or 0.2 mM IAM (B) at pH 5.0 (�), pH 7.0 (�) or pH 9.0 (�).Aliquots of the incubation mixtures were withdrawn at the indicated time points and the remaining NADH-dependent activity was determined. Results are means +− S.D. (n = 3) for three experiments.When error bars are not visible, they are within the symbol. (C) Determination of the pK a value of the active-site cysteine residue of GapC1. Reduced enzyme (2 μM) was incubated with 0.2 mM IAMat different pH values for 20 min at room temperature and then the NADH-dependent activity was determined. After incubation, the percentage of the remaining activity at each pH was determinedby comparing the activity of the enzyme incubated with and without IAM. Results are means +− S.D. (n = 3) for three experiments. When error bars are not visible, they are within the symbol. ThepK a value was obtained by non-linear regression using an adaptation of the Henderson–Hasselbalch equation [18]. The poor fit above pH 6 was reproducible and probably corresponds to reactionof IAM with non-thiolate groups.

Figure 2 Oxidant treatments of GapC1

(A) Time-dependent inactivation. Reduced GapC1 was incubated with Bis-Tris buffer as a control (�) or 50 μM H2O2 with (�) or without (�) 0.5 mM GSH. Aliquots of the incubation mixtures werewithdrawn at the indicated time points and the remaining NADH-dependent activity was determined. Data are represented as the mean percentage of maximal control activity +− S.D. (n = 3). Whenerror bars are not visible, they are within the symbol. (B) Substrate protection of H2O2-dependent GapC1 inactivation by BPGA or G3P. Reduced GapC1 was incubated for 10 min with 50 μM H2O2

alone (black bars) or in the presence of BPGA-generating system (3 mM 3-phosphoglycerate, 5 units/ml of 3-phosphoglycerate kinase and 2 mM ATP) or 1 mM G3P (white bars). After incubation,NAD(H)-dependent GapC1 activity (forward and reverse reaction) was determined. Data are represented as the mean percentage of maximal control activity +− S.D. (n = 3). (C) Reversibility of oxidanttreatments. Reduced GapC1 was incubated for 10 min with H2O2 alone or in the presence of 0.5 mM GSH (black bars). The reversibility of GapC1 inactivation was assessed by incubation for 10 minin the presence of 20 mM DTT (white bars). NADH-dependent GapC1 activity was then determined. Data are represented as the mean percentage of maximal control activity +− S.D. (n = 3).

(–SO3− ). Among these oxidized forms, only sulfenates can be

reduced back by DTT. In the case of GapC1, the inhibition causedby a 10 min incubation with 50 μM H2O2 was largely irreversibleas only 5% of the total protein activity could be recovered byDTT (Figure 2C), suggesting that Cys149 sulfenates were veryunstable and rapidly converted into over-oxidized forms.

Glutathione prevents over-oxidation of GapC1 by H2O2

We tested whether glutathionylation could protect GapC1 fromover-oxidation. The activity of the enzyme was inhibited duringincubation with H2O2 (50 μM) and GSH (0.5 mM) (Figure 2A).

The kinetics of inhibition was slightly slower compared withH2O2 alone, probably because of the reaction of GSH with H2O2.Interestingly however, DTT could totally recover the activityof the enzyme inhibited by H2O2 in the presence of GSH(Figure 2C), indicating a reversible modification of catalyticCys149. A reasonable interpretation of this result is that GSHreacted with short-living Cys149 sulfenates, generating a stableS-glutathionylated form that could be reduced back by DTT. Bythis mechanism, glutathionylation of GapC1 would thus preventits over-oxidation.

Cytosolic GAPDH from Arabidopsis was already shown tobe glutathionylated by GSSG, but this effect was only observedin the absence of cofactor (NAD+ ) [28]. In the present study,



Figure 3 MALDI-TOF MS analysis

Mass spectra of wild-type GapC1 (GapC1 WT) (A) and C149S mutant (B) treated with 50 μM H2O2 and 0.5 mM GSH for 30 min at room temperature were performed before and after a treatment inthe presence of 20 mM DTT (30 min). The peaks labelled ‘matrix adducts’ correspond to GAPDH with a sinapinic acid adduct. The differences between mass peaks of unmodified GapC1 WT before(38 962.4 Da) and after DTT treatment (38 947.3 Da) are within the experimental error (0.1 %) of the instrument.

we also observed that GapC1 was inhibited by GSSG, but onlywhen NAD+ was omitted from the incubation medium (resultsnot shown). Bound NAD+ could apparently protect Cys149 fromthe attack of bulky GSSG, but could not protect Cys149

from reacting with smaller molecules like H2O2 and GSH(Figure 2A). The effect of GSSG on apo-GapC1 allowedestimation of the Kox (equilibrium constant for GapC1glutathionylation) of the reaction of GapC1 with GSSG, i.e. theGSH/GSSG ratio at which 50% of the protein is inactivated (i.e.glutathionylated). As compared with proteins that can be easilyglutathionylated via thiol:disulfide exchange like the plastidialGrxS12 (Kox = 300; [18]), the Kox of GapC1 was very low(0.3 +− 0.05). On the one hand, the reaction of GapC1 withGSSG is consistent with its possible glutathionylation, but onthe other hand, it can be excluded that this reaction could haveany significant role in vivo, since it requires GSH/GSSG ratios thatare far lower than ever observed ([11] and references therein).

GapC1 is glutathionylated on Cys149 as demonstrated by MSanalyses and by BioGSH treatment

The glutathionylation of GapC1 was confirmed by usingtwo different experimental approaches, MALDI–TOF MS andimmunodetection of BioGSH. MALDI–TOF analysis was carriedout on GapC1 treated with 50 μM H2O2 and 0.5 mM GSH. Underthis condition, a clear shift of the enzyme mass of 306 Da wasobserved and could be reversed by DTT treatment (Figure 3A).This reversible mass increase is consistent with the formationof one glutathione adduct (theoretical mass shift of 305 Da) perGapC1 monomer.

Alternatively, GapC1 was incubated for 1 h with buffer orwith 0.5 mM BioGSH or 50 μM H2O2 or 0.5 mM BioGSH inthe presence of 50 μM H2O2. Following incubation, sampleswere subjected to non-reducing SDS/PAGE and either stainedfor protein or immunoblotted with anti-biotin antibodies. Noimmunoreactive signal was detected when wild-type GapC1was treated with buffer or H2O2, whereas a faint 40 kDaimmunoreactive polypeptide was detected in the sample treatedwith BioGSH alone (Figure 4). However, a strong signal occurredwhen wild-type GapC1 was treated with BioGSH and H2O2,and this signal was strongly inhibited by BPGA (Figure 4), inagreement with the concept that BPGA, by covalently binding toCys149, protects this residue from H2O2 oxidation and subsequentglutathionylation. These results confirm that the mechanismof GapC1 glutathionylation is linked to sulfenate formationon Cys149. Finally, and in agreement with previous results(Figures 2 and 3), the signal detected in the presence of H2O2

and BioGSH was completely removed following treatment withDTT (Figure 4). Both the MALDI–TOF and BioGSH methodsestablish that GapC1 is glutathionylated upon treatment with H2O2

and GSH.In order to firmly demonstrate that catalytic Cys149, and not

Cys153, which is in close proximity to the active site, is the target ofglutathionylation the C149S mutant was incubated in conditionssimilar to those used to induce the glutathionylation of wild-typeGapC1 (Figure 3). The MALDI–TOF analysis revealed no shift inthe molecular mass of the protein (Figure 3B), indicating that thismutant did not undergo glutathionylation. This was confirmedby incubation of the C149S mutant with H2O2 and BioGSHsince no biotin signal was detected by Western blot analysis(Figure 4).



Figure 4 Analysis of wild-type GapC1 and C149S mutant glutathionylationwith BioGSH

Wild-type GapC1 (GapC1 WT) and C149S mutant were incubated for 1 h with buffer in thepresence and absence of 50 μM H2O2, 0.5 mM BioGSH or 50 μM H2O2 and 0.5 mM BioGSH,with or without pre-incubation with BPGA. Proteins were resolved by non-reducing SDS/PAGEand either stained for protein or transferred on to nitrocellulose membranes for Western blottingwith anti-biotin antibodies. The Coomassie Blue staining of the gels shows equal loading foreach lane. The reversibility of the reaction was assessed by treatment with 20 mM DTT for 30 minas indicated.

Since mutant C149S was not glutathionylated, it is concludedthat catalytic Cys149 is the only target of glutathionylation inwild-type GapC1. Glutathionylation prevents over-oxidation andreversibly inhibits enzyme activity.

Grxs reactivate glutathionylated GapC1 through a monothiolmechanism

The ability of diverse plant Trxs and Grxs to reactivateglutathionylated GapC1 was investigated. Reactivation testsof glutathionylated GapC1 were conducted with a cytosolicglutaredoxin from poplar (PtGrxC1), previously demonstratedto be active in a HED (2-hydroxyethyl disulfide) assay, aclassical test to determine deglutathionylating activity [39].

Table 2 t1/2 of glutathionylated GapC1 in the presence of different Grxs andTrxs

Reactivation kinetics were performed with 10 μM Grx or Trx isoforms in the presence oftheir respective reducing system. For further details, see the Experimental section. Results arepresented as average +− S.D. at least three experiments.

Protein t 1/2 of reactivation (min)

PtGrxC1 4.70 +− 0.73PtGrxC1-C34S 1.56 +− 0.19CrTRXh1 12.3 +− 3.9PtTRXh1 12.7 +− 2.3PtTRXh2 36.7 +− 4.2PtTRXh3 20.3 +− 2.8PtTRXh5 10.1 +− 0.5

The t1/2(half-time for reactivation) of glutathionylated GapC1

was approximately 5 min when incubated with PtGrxC1 in thepresence of a GRS (GSH reducing system) in which GSH waskept fully reduced by NADPH and GR (Figure 5 and Table 2). Acontrol experiment was performed in the presence of GRS aloneand no reactivation was detected, indicating that GSH alone couldnot perform deglutathionylation.

PtGrxC1 belongs to class I Grxs (CPYC-type) and contains twocysteines (Cys31 and Cys34) in its active site. The monocysteinicPtGrxC1-C31S and PtGrxC1-C34S mutants were used as toolsto investigate the mechanism of GapC1 deglutathionylation.The C31S mutant of PtGrxC1 was completely inactive, incontrast with the PtGrxC1-C34S mutant that catalysed GapC1deglutathionylation more efficiently than wild-type PtGrxC1(Figure 5 and Table 2). These results are consistent with themonothiol mechanism with side reaction depicted in Figure 6(A).The first step of this mechanism consists of the nucleophilicattack performed by Cys31 of PtGrxC1 on the mixed disulfideof glutathionylated GapC1. This reaction releases a reduced,i.e. active, GapC1 and a glutathionylated PtGrxC1 intermediate(Cys31–SSG) (Figure 6A). In a second step, either GSH or Cys34

(the second cysteine of the PtGrxC1 active site) removes the

Figure 5 Kinetics of reactivation of glutathionylated GapC1

(A) Glutathionylated GapC1 was incubated with a GSH-regenerating system alone (2 mM GSH, 0.2 mM NADPH and 6 μg/ml glutathione reductase; �) or in the presence of wild-type PtGrxC1(�) or PtGrxC1-C31S (�) or PtGrxC1-C34S (�). Aliquots of the incubation mixtures were withdrawn at the indicated time points and the remaining NADH-dependent activity was determined.Data represent the average +− S.D. of three experiments (n = 3). (B) Glutathionylated GapC1 was incubated with the NTR system (0.2 mM NADPH and 0.22 μM A. thaliana NTRb alone (control;�) or in the presence of Chlamydomonas Trxh1 (�) or poplar Trxh1 (�) or poplar Trxh5 (�). Aliquots of the incubation mixtures were withdrawn at the indicated time points and the remainingNADH-dependent activity was determined. Data represent the average of three experiments (n = 3) and are reproducible to within +− 13 % S.D. of the mean value. (C) Glutathionylated GapC1 wasincubated for 20 min with pre-reduced Trxh1 or Trxh1-C39S at a 1:10 ratio. After incubation, NADH-dependent GapC1 activity was determined. Data represent the average +− S.D. of three experiments(n = 3).



Figure 6 Possible mechanisms of GapC1 deglutathionylation

(A) Grx-dependent deglutathionylation mechanism. The most reactive cysteine of Grx (the N-terminal active-site cysteine) performs the nucleophilic attack on the glutathione-mixed disulfide onGapC1, resulting in the release of the reduced GapC1 and the formation of a glutathionylated Grx intermediate. Subsequently, the mixed disulfide on Grx is reduced by a GSH molecule to form GSSGand reduced Grx. The attack of GSH on the glutathionylated Grx intermediate can compete with the attack performed by the C-terminal cysteine residue of the active site, i.e. side reaction (brokenlines). (B) Trx-dependent deglutathionylation mechanisms. In the first step of the reaction, the most reactive cysteine of Trx (the N-terminal active site cysteine residue) performs the nucleophilicattack on the glutathione-mixed disulfide on GapC1, resulting in the release of the reduced GapC1 and the formation of a glutathionylated Trx intermediate. Alternatively, the nucleophilic attackcan lead to the formation of an intermolecular disulfide intermediate between Trx and GapC1. In a second step, the glutathionylated Trx or the Trx-GapC1 disulfide intermediate is attacked by theC-terminal cysteine of the active site, leading to the formation of oxidized Trx and, in the latter case, also the release of reduced GapC1. The regeneration of reduced Trx is then performed by NTR inthe presence of NADPH.

glutathionyl moiety from Cys31–SSG. If this reaction is performedby GSH (the only possibility in mutant C34S), the reduced (active)glutaredoxin is rapidly regenerated and deglutathionylation ofGapC1 can proceed faster. Alternatively, Cys31–SSG mixeddisulfide may be attacked by Cys34, thus forming an intramoleculardisulfide (Cys31–SS–Cys34, oxidized PtGrxC1). This side reactioncan negatively affect the overall rate of deglutathionylation of thetarget as it requires a second GSH molecule for the regenerationof the reduced (active) glutaredoxin (Figure 6A). According tothis model, wild-type PtGrxC1 was slower than mutant C34S incatalysing the deglutathionylation of GapC1.

Similar results were described previously for other class I(CPYC-type) Grxs such as mammalian Grx1 and Grx2, yeastGrx1, Arabidopsis GrxC5 [17,45,46] and Chlamydomonas Grx1(results not shown).

A new role for Trxs as proteins involved in deglutathionylationreactions

In plants, Trxs have been shown to regulate a large number ofproteins through thiol:disulfide exchange reactions [4,5], whereasno significant deglutathionylating activity was reported [14]. Wethus tested the ability of different cytosolic h-type Trxs from bothpoplar and C. reinhardtii to catalyse GapC1 deglutathionylation.A first attempt performed with DTT as a reductant of Trxsgave contrasting results because of the fast chemical reductionof glutathionylated GapC1 by DTT alone (results not shown).However, by substituting DTT with a physiological reducingsystem composed of NADPH and A. thaliana NTRb, GapC1 was

indeed reactivated by all of the h-type Trxs tested, although atslower rates compared with PtGrxC1 (Figure 5B and Table 2). Thebest results were obtained with PtTrxh5 and PtTrxh1 from poplar,and CrTrxh1 from C. reinhardtii, which reactivated GapC1 witha t1/2

of 10–13 min at 10 μM, i.e. twice as long as that of PtGrxC1at the same concentration (Table 2).

To our knowledge, this is the first report of deglutathionylatingactivity performed by plant Trxs. Trxs contain two cysteineresidues in their active site and the N-terminal one performs thenucleophilic attack on disulfides. In principle, the mechanismof Trx-mediated deglutathionylation could either involve anintermolecular disulfide intermediate between GapC1 andTrx, or a glutathionylated Trx intermediate, similar to thedeglutathionylation mechanism catalysed by Grxs (Figure 6B).Both would be attacked by the C-terminal cysteine residue,resulting in the formation of oxidized Trx that is reducedback by NTR. In order to discriminate between these twopossible mechanisms, we employed a Trx monocysteinic mutantlacking the C-terminal active-site cysteine (CrTrxh1-C39S). Themutant Trx was unable to reactivate glutathionylated GapC1(Figure 5C), suggesting a mechanism involving formation of Trx-GapC1 heterodimers (Figure 6B). However, such dimers couldnot be detected by non-reducing SDS/PAGE analysis of theproducts of the reaction (results not shown). The latter resultwould rather favour the alternative mechanism where Trx isglutathionylated. These conflicting results are probably linked tothe inability of CrTrxh1-C39S to perform the nucleophilic attackon glutathionylated GapC1, thereby preventing a discriminationbetween the two possible mechanisms. Similarly, although two



possible mechanisms have been proposed for Trx-catalyseddenitrosylation, the exact mechanism remains to be determined[47]. Whatever the mechanism, both cysteine residues of the Trxactive site are required to complete the deglutathionylation cyclewith the mandatory involvement of NTR and NADPH. In thisrespect, Trxs are different from Grxs in that Trxs cannot performmonothiol deglutathionylation (Figure 6).

DISCUSSION

A prominent property of GAPDH from many different organismsis its sensitivity to oxidation. Not surprisingly, different GAPDHisoforms were reported to be redox-modified in different ways,and several proteomic approaches which aimed at identifyingredox-modified proteins under both physiological and stressconditions identified GAPDH as a target [22,24,26,27,32]. Thecatalytic mechanism of GAPDH requires a reactive cysteineresidue that is covalently acylated by the substrate during thecatalytic cycle [44]. The reactivity of Cys149 is determinedby the close proximity of a histidine (His176 in the prototypicalstructure of B. stearothermophilus GAPDH [34], and alsoin GapC1 of Arabidopsis in the present study) that, byattracting the thiol proton, increases the acidic nature of Cys149

(Supplementary Figure S3 at http://www.BiochemJ.org/bj/445/bj4450337add.htm). Deprotonated Cys149 is at the same timea better catalyst for the enzymatic reaction and a sensitivetarget for molecules that do react with thiolates, ROS like H2O2

being a physiologically relevant example. Obviously, any redoxmodification of Cys149 completely inhibits GAPDH activity untilthe modification is removed and the reduced state of the thiolateregenerated.

In the present study we show that the catalytic activity ofcytosolic GAPDH from Arabidopsis, similarly to other knownGAPDHs, is fully dependent on an acidic cysteine residue (Cys149)that is also the only accessible cysteine of the protein. In termsof pKa, the acidity of Cys149 was found to be almost 3 pH unitsbelow that of free cysteine [7]. H2O2 fully inhibits enzyme activity,causing a practically irreversible inactivation of the enzyme, sincea large excess of DTT could not restore the activity. Irreversibleoxidation is explained as a conversion of Cys149 thiolate into over-oxidized sulfinic (or sulfonic) forms, a reaction that would involveCys149 sulfenate (Cys149–SOH) as a labile intermediate. Sulfeniccysteines are the only oxidized forms that DTT can reduce backto thiols [3].

Glutathione concentration in the cytosol of plant cells is inthe millimolar range [48] and we show that GSH can efficientlyprevent the over-oxidation of GapC1 by quantitatively reactingwith sulfenic Cys149 to form a mixed disulfide that can be reducedby physiological reductants, either glutaredoxins or thioredoxins.Thus glutathionylation of GapC1 can be considered both as aprotective mechanism against irreversible oxidation and asa means to store inhibited GapC1 in a recoverable form. Both rolesfor glutathionylation can be envisioned within the framework ofadaptation and recovery of plant cells to stress.

Cytosolic GAPDH may be a prominent target of oxidativestress in the cytosolic fraction of plant extracts [27], andit has been shown to be glutathionylated under in vitrooxidizing conditions or under in vivo oxidative stress [24,26].Glutathionylation of cytosolic phosphorylating GAPDH mayappear as a mechanism allowing down-regulation of theglycolytic pathway that might increase flux of glucose equivalentsthrough the OPP (oxidative pentose phosphate) pathway leadingto the generation of NADPH (Supplementary Figure S4 athttp://www.BiochemJ.org/bj/445/bj4450337add.htm). Cytosolic

glucose-6-phosphate dehydrogenase isoforms, which catalysethe first reaction of the OPP pathway, are indeed insensitiveto oxidative modifications [49]. However, this possible down-regulation of the glycolytic pathway may not be relevant inplant cytosol that also contains a non-phosphorylating GAPDH(GAPN) that can bypass the GapC-catalysed reaction [50]. GAPNbelongs to the aldehyde dehydrogenase superfamily, and hasno close functional/structural relationships with phosphorylatingGAPDHs [51]. Compared with oxidation-sensitive GAPDHs,GAPN (Streptococcus GAPN at least) is highly resistant toH2O2 inactivation [52]. GAPN catalyses an irreversible reactionthat is NADP specific, whereas the reaction catalysed byGapC1 is reversible and specific for NAD(H). Interestingly, theactivity of GAPN is increased up to 2-fold under oxidativestress conditions in either wheat or maize seedlings [53].Although creating a bypass for the blocked glycolytic fluxunder oxidative stress, GAPN might thus provide an alternativesource of NADPH for the antioxidant enzymes (SupplementaryFigure S4).

GR and Trx reductases are major antioxidant enzymes in thecytosol of plant cells. By keeping the glutathione pool reduced,GR provides the reductant for the efficient deglutathionylationof GapC1 via cytosolic glutaredoxins. Alternatively, as thecapacity of GR may be limiting under oxidative stress conditions[2], we show in the present study that GapC1 may be alsorecovered by a GSH-independent system, i.e. via cytosolic Trxh(Supplementary Figure S4). This is the first report of a role of plantTrxs in protein deglutathionylation. A similar role was recentlysuggested in yeast, where Grxs may be less active than in otherorganisms for catalysing protein deglutathionylation [20]. Despitetheir structural and sequence similarities, plant Grxs and Trxshave peculiar biochemical features mainly linked to the type ofdisulfides they are able to target and to the molecular mechanisminvolved in their catalytic cycle [11]. The affinity of Grxs forglutathionylated proteins is well established [3,11]. Additionalexperiments will clarify whether the capacity of h-type Trxs tocatalyse GapC1 deglutathionylation may be extended to othercombinations of targets and Trx types, thereby further sustainingthe growing evidence for the possibility of functional overlapbetween the Grx and Trx systems in plant cells [3].

In animal cells, it is clear that GAPDH redox modifications(which invariably inhibit enzyme activity and thus have anunavoidable effect on glycolysis) can also redirect the proteinitself to new, completely unrelated, functions. Comprehensivedata have demonstrated that animal GAPDH is involved inseveral cellular processes such as receptor-mediated signalling,intracellular membrane trafficking, maintenance of DNAintegrity, transcriptional and post-transcriptional regulation,oxidative stress response and apoptosis [33]. Several of thesefunctions involve the relocalization of GAPDH to the nucleus andare triggered or regulated by redox modifications of the catalyticcysteine. For instance, migration in the nucleus of cytosolicGAPDH upon oxidative stress may result in an apoptotic responseand depend on the nitrosylation of a catalytic cysteine residue[54], although this is by no means the only mechanism of nuclearlocalization of GAPDH [33]. Recently, a chimaeric GapC–GFPconstruct was shown to localize in the nucleus of transformedA. thaliana protoplasts, although these latter results mighthave been biased by the artificially strong expression providedby the 35S promoter [28]. The active site of mammalianGAPDH was also implicated in functional interactions withpolynucleotides. For example, the active site of GAPDH can bindto the 3′-untranslated region of the mRNA of the endothelialvasoconstrictor ET-1 (endothelin 1), thereby promotingdegradation of the messenger. Interestingly, glutathionylation of






GAPDH by oxidative stress conditions inhibits binding and resultsin increased ET-1 protein levels and endothelial vasoconstriction[55]. The resulting increase in ET-1 protein level would facilitateits function as an endothelial vasoconstrictor. The similaritybetween animal and plant GAPDH active sites indeed suggeststhat GapC might also interact with polynucleotides in plant cells.

Overall, the complex interplay between Trxs, Grxs and post-translational redox modifications that affect GAPDH and maymodify its physiological role in plant cells, is only emerging.Efforts aimed at understanding these complex interconnectionswill hopefully clarify the specific functions and redundanciesof the different components of the redox signalling network.Unravelling the importance of this cross-talk will certainlyconstitute a major challenge for future studies.

AUTHOR CONTRIBUTION

Mariette Bedhomme, Mattia Adamo, Christophe Marchand, Jeremy Couturier and MirkoZaffagnini performed the experiments; Mirko Zaffagnini and Paolo Trost co-ordinated theproject; Mariette Bedhomme, Nicolas Rouhier, Stephane Lemaire, Mirko Zaffagnini andPaolo Trost wrote the paper; and all of the authors discussed the data and the paper.

FUNDING

This work was supported, in part, by the Ministero dell’Istruzione, dell’Universita e dellaRicerca [grant number PRIN2008XB774B-005 (to M.A., M.Z. and P.T.)], the ANR (AgenceNationale de la Recherche) [grant numbers 08-BLAN-0153 GLUTAPHOTO (to M.B. andS.D.L.) and ANR-07-JCJC-0121 (to N.R. and J.C.)], by the PHC (Partenariats HubertCurien) Galilee Project (to M.B., S.D.L., M.Z. and P.T.) and the Ville de Paris by a“Research in Paris” fellowship (to M.Z.).

REFERENCES

1 Mittler, R., Vanderauwera, S., Suzuki, N., Miller, G., Tognetti, V. B., Vandepoele, K.,Gollery, M., Shulaev, V. and Van Breusegem, F. (2011) ROS signaling: the new wave?Trends Plant Sci. 16, 300–309

2 Foyer, C. H. and Noctor, G. (2011) Ascorbate and glutathione: the heart of the redox hub.Plant Physiol. 155, 2–18

3 Zaffagnini, M., Bedhomme, M., Marchand, C. H., Morisse, S., Trost, P. and Lemaire, S. D.(2012) Redox regulation in photosynthetic organisms: focus on glutathionylation.Antioxid. Redox Signal. 16, 567–586

4 Buchanan, B. B. and Balmer, Y. (2005) Redox regulation: a broadening horizon. Annu.Rev. Plant Biol. 56, 187–220

5 Lemaire, S. D., Michelet, L., Zaffagnini, M., Massot, V. and Issakidis-Bourguet, E. (2007)Thioredoxins in chloroplasts. Curr. Genet. 51, 343–365

6 Besson-Bard, A., Pugin, A. and Wendehenne, D. (2008) New insights into nitric oxidesignaling in plants. Annu. Rev. Plant Biol. 59, 21–39

7 Klomsiri, C., Karplus, P. A. and Poole, L. B. (2011) Cysteine-based redox switches inenzymes. Antioxid. Redox Signal. 14, 1065–1077

8 Tarrago, L., Laugier, E., Zaffagnini, M., Marchand, C. H., Le Marechal, P., Lemaire, S. D.and Rey, P. (2010) Plant thioredoxin CDSP32 regenerates 1-cys methionine sulfoxidereductase B activity through the direct reduction of sulfenic acid. J. Biol. Chem. 285,14964–14972

9 Gao, X. H., Bedhomme, M., Veyel, D., Zaffagnini, M. and Lemaire, S. D. (2009) Methodsfor analysis of protein glutathionylation and their application to photosyntheticorganisms. Mol. Plant 2, 218–235

10 Gallogly, M. M., Starke, D. W. and Mieyal, J. J. (2009) Mechanistic and kinetic details ofcatalysis of thiol–disulfide exchange by glutaredoxins and potential mechanisms ofregulation. Antioxid. Redox Signaling 11, 1059–1081

11 Zaffagnini, M., Bedhomme, M., Lemaire, S. D. and Trost, P. (2012) The emerging roles ofprotein glutathionylation in chloroplasts. Plant Sci. 185–186, 86–96

12 Rouhier, N., Lemaire, S. D. and Jacquot, J. P. (2008) The role of glutathione inphotosynthetic organisms: emerging functions for glutaredoxins and glutathionylation.Annu. Rev. Plant Biol. 59, 143–166

13 Chibani, K., Tarrago, L., Schurmann, P., Jacquot, J. P. and Rouhier, N. (2011) Biochemicalproperties of poplar thioredoxin z. FEBS Lett. 585, 1077–1081

14 Zaffagnini, M., Michelet, L., Massot, V., Trost, P. and Lemaire, S. D. (2008) Biochemicalcharacterization of glutaredoxins from Chlamydomonas reinhardtii reveals the uniqueproperties of a chloroplastic CGFS-type glutaredoxin. J Biol. Chem. 283, 8868–8876

15 Couturier, J., Koh, C. S., Zaffagnini, M., Winger, A. M., Gualberto, J. M., Corbier, C.,Decottignies, P., Jacquot, J. P., Lemaire, S. D., Didierjean, C. and Rouhier, N. (2009)Structure–function relationship of the chloroplastic glutaredoxin S12 with an atypicalWCSYS active site. J. Biol. Chem. 284, 9299–9310

16 Gao, X. H., Zaffagnini, M., Bedhomme, M., Michelet, L., Cassier-Chauvat, C.,Decottignies, P. and Lemaire, S. D. (2010) Biochemical characterization of glutaredoxinsfrom Chlamydomonas reinhardtii: kinetics and specificity in deglutathionylationreactions. FEBS Lett. 584, 2242–2248

17 Couturier, J., Stroher, E., Albetel, A. N., Roret, T., Muthuramalingam, M., Tarrago, L.,Seidel, T., Tsan, P., Jacquot, J. P., Johnson, M. K. et al. (2011) Arabidopsis chloroplasticglutaredoxin C5 as a model to explore molecular determinants for iron-sulfur clusterbinding into glutaredoxins. J. Biol. Chem. 286, 27515–27527

18 Zaffagnini, M., Bedhomme, M., Marchand, C. H., Couturier, J. R., Gao, X. H., Rouhier, N.,Trost, P. and Lemaire, S. D. (2012) Glutaredoxin S12: unique properties for redoxsignaling. Antioxid. Redox Signaling 16, 17–32

19 Benhar, M., Forrester, M. T., Hess, D. T. and Stamler, J. S. (2008) Regulated proteindenitrosylation by cytosolic and mitochondrial thioredoxins. Science 320, 1050–1054

20 Greetham, D., Vickerstaff, J., Shenton, D., Perrone, G. G., Dawes, I. W. and Grant, C. M.(2010) Thioredoxins function as deglutathionylase enzymes in the yeast Saccharomycescerevisiae. BMC Biochem. 11, 3

21 Tan, S. X., Greetham, D., Raeth, S., Grant, C. M., Dawes, I. W. and Perrone, G. G. (2010)The thioredoxin–thioredoxin reductase system can function in vivo as an alternativesystem to reduce oxidized glutathione in Saccharomyces cerevisiae. J. Biol. Chem. 285,6118–6126

22 Kehr, S., Jortzik, E., Delahunty, C., Yates, III, J. R., Rahlfs, S. and Becker, K. (2011) ProteinS-glutathionylation in malaria parasites. Antioxid. Redox Signaling 15, 2855–2865

23 Ito, H., Iwabuchi, M. and Ogawa, K. (2003) The sugar-metabolic enzymes aldolase andtriose-phosphate isomerase are targets of glutathionylation in Arabidopsis thaliana:detection using biotinylated glutathione. Plant Cell Physiol. 44, 655–660

24 Dixon, D. P., Skipsey, M., Grundy, N. M. and Edwards, R. (2005) Stress-induced proteinS-glutathionylation in Arabidopsis. Plant Physiol. 138, 2233–2244

25 Michelet, L., Zaffagnini, M., Vanacker, H., Le Marechal, P., Marchand, C., Schroda, M.,Lemaire, S. D. and Decottignies, P. (2008) In vivo targets of S-thiolation inChlamydomonas reinhardtii. J. Biol. Chem. 283, 21571–21578

26 Zaffagnini, M., Bedhomme, M., Groni, H., Marchand, C. H., Puppo, C., Gontero, B.,Cassier-Chauvat, C., Decottignies, P. and Lemaire, S. D. (2012) Glutathionylation in thephotosynthetic model organism Chlamydomonas reinhardtii: a proteomic survey. Mol.Cell. Proteomics 11, M111.014142

27 Hancock, J. T., Henson, D., Nyirenda, M., Desikan, R., Harrison, J., Lewis, M., Hughes, J.and Neill, S. J. (2005) Proteomic identification of glyceraldehyde 3-phosphatedehydrogenase as an inhibitory target of hydrogen peroxide in Arabidopsis. PlantPhysiol. Biochem. 43, 828–835

28 Holtgrefe, S., Gohlke, J., Starmann, J., Druce, S., Klocke, S., Altmann, B., Wojtera, J.,Lindermayr, C. and Scheibe, R. (2008) Regulation of plant cytosolic glyceraldehyde3-phosphate dehydrogenase isoforms by thiol modifications. Physiol. Plant 133,211–228

29 Zaffagnini, M., Michelet, L., Marchand, C., Sparla, F., Decottignies, P., Le Marechal, P.,Miginiac-Maslow, M., Noctor, G., Trost, P. and Lemaire, S. D. (2007) Thethioredoxin-independent isoform of chloroplastic glyceraldehyde-3-phosphatedehydrogenase is selectively regulated by glutathionylation. FEBS J. 274, 212–226

30 Ravichandran, V., Seres, T., Moriguchi, T., Thomas, J. A. and Johnston, Jr, R. B. (1994)S-thiolation of glyceraldehyde-3-phosphate dehydrogenase induced by thephagocytosis-associated respiratory burst in blood monocytes. J. Biol. Chem. 269,25010–25015

31 Lind, C., Gerdes, R., Schuppe-Koistinen, I. and Cotgreave, I. A. (1998) Studies on themechanism of oxidative modification of human glyceraldehyde-3-phosphatedehydrogenase by glutathione: catalysis by glutaredoxin. Biochem. Biophys. Res.Commun. 247, 481–486

32 Shenton, D. and Grant, C. M. (2003) Protein S-thiolation targets glycolysis and proteinsynthesis in response to oxidative stress in the yeast Saccharomyces cerevisiae.Biochem. J. 374, 513–519

33 Sirover, M. A. (2011) On the functional diversity of glyceraldehyde-3-phosphatedehydrogenase: biochemical mechanisms and regulatory control. Biochim. Biophys. Acta1810, 741–751

34 Biesecker, G., Harris, J. I., Thierry, J. C., Walker, J. E. and Wonacott, A. J. (1977)Sequence and structure of D-glyceraldehyde 3-phosphate dehydrogenase from Bacillusstearothermophilus. Nature 266, 328–333

35 Goyer, A., Decottignies, P., Lemaire, S., Ruelland, E., Issakidis-Bourguet, E., Jacquot, J. P.and Miginiac-Maslow, M. (1999) The internal Cys207 of sorghum leaf NADP-malatedehydrogenase can form mixed disulphides with thioredoxin. FEBS Lett. 444, 165–169



36 Jacquot, J. P., Rivera-Madrid, R., Marinho, P., Kollarova, M., Le Marechal, P.,Miginiac-Maslow, M. and Meyer, Y. (1994) Arabidopsis thaliana NADPH thioredoxinreductase. cDNA characterization and expression of the recombinant protein inEscherichia coli. J. Mol. Biol. 235, 1357–1363

37 Gelhaye, E., Rouhier, N. and Jacquot, J. P. (2003) Evidence for a subgroup of thioredoxinh that requires GSH/Grx for its reduction. FEBS Lett. 555, 443–448

38 Gelhaye, E., Rouhier, N., Laurent, P., Sautiere, P. E., Martin, F. and Jacquot, J. P. (2002)Isolation and characterization of an extended thioredoxin h from poplar. Physiol. Plant114, 165–171

39 Rouhier, N., Unno, H., Bandyopadhyay, S., Masip, L., Kim, S. K., Hirasawa, M., Gualberto,J. M., Lattard, V., Kusunoki, M., Knaff, D. B. et al. (2007) Functional, structural, andspectroscopic characterization of a glutathione-ligated [2Fe-2S] cluster in poplarglutaredoxin C1. Proc. Natl. Acad. Sci. U.S.A. 104, 7379–7384

40 Gasteiger, E., Gattiker, A., Hoogland, C., Ivanyi, I., Appel, R. D. and Bairoch, A. (2003)ExPASy: the proteomics server for in-depth protein knowledge and analysis. NucleicAcids Res. 31, 3784–3788

41 Sparla, F., Tedeschi, G., Pupillo, P. and Trost, P. (1999) Cloning and heterologousexpression of NAD(P)H:quinone reductase of Arabidopsis thaliana, a functionalhomologue of animal DT-diaphorase. FEBS Lett. 463, 382–386

42 Bedhomme, M., Zaffagnini, M., Marchand, C. H., Gao, X. H., Moslonka-Lefebvre, M.,Michelet, L., Decottignies, P. and Lemaire, S. D. (2009) Regulation by glutathionylation ofisocitrate lyase from Chlamydomonas reinhardtii. J. Biol. Chem. 284, 36282–36291

43 Sicard-Roselli, C., Lemaire, S., Jacquot, J. P., Favaudon, V., Marchand, C. andHouee-Levin, C. (2004) Thioredoxin Ch1 of Chlamydomonas reinhardtii displays anunusual resistance toward one-electron oxidation. Eur. J. Biochem. 271, 3481–3487

44 Talfournier, F., Colloc’h, N., Mornon, J. P. and Branlant, G. (1998) Comparative study ofthe catalytic domain of phosphorylating glyceraldehyde-3-phosphate dehydrogenasesfrom bacteria and archaea via essential cysteine probes and site-directed mutagenesis.Eur. J. Biochem. 252, 447–457

45 Gallogly, M. M., Starke, D. W., Leonberg, A. K., Ospina, S. M. and Mieyal, J. J. (2008)Kinetic and mechanistic characterization and versatile catalytic properties of mammalianglutaredoxin 2: implications for intracellular roles. Biochemistry 47, 11144–11157

46 Discola, K. F., de Oliveira, M. A., Rosa Cussiol, J. R., Monteiro, G., Barcena, J. A., Porras,P., Padilla, C. A., Guimaraes, B. G. and Netto, L. E. (2009) Structural aspects of thedistinct biochemical properties of glutaredoxin 1 and glutaredoxin 2 from Saccharomycescerevisiae. J. Mol. Biol. 385, 889–901

47 Benhar, M., Forrester, M. T. and Stamler, J. S. (2009) Protein denitrosylation: enzymaticmechanisms and cellular functions. Nat. Rev. Mol. Cell Biol. 10, 721–732

48 Queval, G., Jaillard, D., Zechmann, B. and Noctor, G. (2011) Increased intracellular HOavailability preferentially drives glutathione accumulation in vacuoles and chloroplasts.Plant Cell Environ. 34, 21–32

49 Wakao, S. and Benning, C. (2005) Genome-wide analysis of glucose-6-phosphatedehydrogenases in Arabidopsis. Plant J. 41, 243–256

50 Rius, S. P., Casati, P., Iglesias, A. A. and Gomez-Casati, D. F. (2006) Characterization ofan Arabidopsis thaliana mutant lacking a cytosolic non-phosphorylatingglyceraldehyde-3-phosphate dehydrogenase. Plant Mol. Biol. 61, 945–957

51 Michels, S., Scagliarini, S., Della Seta, F., Carles, C., Riva, M., Trost, P. and Branlant, G.(1994) Arguments against a close relationship between non-phosphorylating andphosphorylating glyceraldehyde-3-phosphate dehydrogenases. FEBS Lett. 339,97–100

52 Arutyunov, D. Y. and Muronetz, V. I. (2003) The activation of glycolysis performed by thenon-phosphorylating glyceraldehyde-3-phosphate dehydrogenase in the model system.Biochem. Biophys. Res. Commun. 300, 149–154

53 Bustos, D. M., Bustamante, C. A. and Iglesias, A. A. (2008) Involvement ofnon-phosphorylating glyceraldehyde-3-phosphate dehydrogenase in response tooxidative stress. J. Plant Physiol. 165, 456–461

54 Hara, M. R., Agrawal, N., Kim, S. F., Cascio, M. B., Fujimuro, M., Ozeki, Y., Takahashi, M.,Cheah, J. H., Tankou, S. K., Hester, L. D. et al. (2005) S-nitrosylated GAPDH initiatesapoptotic cell death by nuclear translocation following Siah1 binding. Nat. Cell Biol. 7,665–674

55 Rodriguez-Pascual, F., Redondo-Horcajo, M., Magan-Marchal, N., Lagares, D.,Martinez-Ruiz, A., Kleinert, H. and Lamas, S. (2008) Glyceraldehyde-3-phosphatedehydrogenase regulates endothelin-1 expression by a novel, redox-sensitive mechanisminvolving mRNA stability. Mol. Cell. Biol. 28, 7139–7155

Received 23 March 2012/18 May 2012; accepted 21 May 2012Published as BJ Immediate Publication 21 May 2012, doi:10.1042/BJ20120505


Biochem. J. (2012) 445, 337–347 (Printed in Great Britain) doi:10.1042/BJ20120505

SUPPLEMENTARY ONLINE DATAGlutathionylation of cytosolic glyceraldehyde-3-phosphate dehydrogenasefrom the model plant Arabidopsis thaliana is reversed by both glutaredoxinsand thioredoxins in vitroMariette BEDHOMME*†, Mattia ADAMO*, Christophe H. MARCHAND†, Jeremy COUTURIER‡, Nicolas ROUHIER‡,Stephane D. LEMAIRE†, Mirko ZAFFAGNINI*1 and Paolo TROST*1

*Laboratory of Molecular Plant Physiology, Department of Experimental Evolutionary Biology, University of Bologna, Via Irnerio 42, 40126 Bologna, Italy, †Laboratoire de BiologieMoleculaire et Cellulaire des Eucaryotes, FRE3354 Centre National de la Recherche Scientifique, Universite Pierre et Marie Curie, Institut de Biologie Physico-Chimique, 13 rue Pierre etMarie Curie, 75005 Paris, France, and ‡UMR1136 Interactions Arbres-Microorganismes, Lorraine University-INRA, IFR 110, 54506 Vandoeuvre-les-Nancy cedex, France

Figure S1 SDS/PAGE of wild-type GapC1 and mutants C153S and C149S ofA. thaliana expressed in E. coli and purified to homogeneity

Sample proteins (2 μg) were separated by 12 % polyacrylamide gel and stained with CoomassieBrilliant Blue. M, marker; WT, wild-type. Molecular mass in kDa is shown.

1 Correspondence may be addressed to either of these authors (email [email protected] or [email protected]).


M. Bedhomme and others

Figure S2 Multiple alignment of GAPDH from diverse organisms

The proteins were aligned with the ClustalW2 program and corrected manually for N-terminal extensions. Residues are numbered according to the structure of B.stearothermophilus GAPDHdeposited in the Protein Data Bank (PDB code 2DBV; [34]). AtGapC1 (NCBI accession number AEE74039.1), AtGapC2 (NCBI accession number AEE29016.1) AtGapA, (NCBI accession numberAEE77191.1), B. staerothermophilus BsGAPDH (PDB code 2DBV), Homo sapiens HsGAPDH (NCBI accession number P04406.3), S. cerevisiae ScGAPDH (NCBI accession number P00358.3)and Plasmodium falciparum PfGAPDH (NCBI accession number XP_001348772.1). Cys149 and Cys153 residues are on a black background, whereas His176 is on a grey background. *, invariantresidues; ., conservation between residues with weakly similar properties; :, conservation between residues with strongly similar properties.


Deglutathionylation of GapC1

Figure S3 Three-dimensional model of A. thaliana GapC1

The model was made using Swiss-Model workspace (http://swissmodel.expasy.org/workspace/) based on the known structure of GAPDH from Oryza sativa (PDB code 3E5R). The structure wasgenerated with the Swiss-PDB viewer software and rendered with POV-Ray (http://www.povray.org). The NAD + cofactor and Cys149, Cys153 and His176 are shown. Inset: blue broken lines representthe distance (in A) between Cys149 and Cys153 and between Cys149 and His176.


http://swissmodel.expasy.org/workspace/

http://www.povray.org

M. Bedhomme and others

Figure S4 Schematic representation of glycolysis showing the NADPH-producing systems in a situation of GapC glutathionylation

Under stress conditions, GapC might undergo glutathionylation with important effects on cytosolic primary metabolism. Indeed, inhibition of GapC activity and the consequent down-regulation ofthe glycolysis pathway would promote entry of glucose equivalents into the OPP (oxidative pentose phosphate) pathway leading to the generation of NADPH (red arrows). Although inhibitionof GapC would down-regulate the glycolytic pathway, plant cells also contain a non-phosphorylating GAPDH (GAPN) that can bypass the GapC-catalysed reaction, providing an alternativesource of NADPH for the antioxidant enzymes (blue arrows). GR and NTR are major antioxidants enzymes in the cytosol of plant cells. GR, using NAPDH as an electron donor, can keep the glutathionepool reduced, providing the reductant (GSH) for the efficient deglutathionylation of GapC via cytosolic glutaredoxins. Alternatively, GapC1 may be also deglutathionylated by a GSH-independentsystem involving NADPH, NTR and cytosolic Trxh. Overall, redirection of primary metabolism in stressed plant cells would allow reinforcing the antioxidant systems and creating the conditions forrecovery (e.g. reduction/reactivation of glutathionylated proteins such as GapC1). DHAP, dihydroxyacetone phosphate; 3PGA, 3-phosphoglycerate.

Received 23 March 2012/18 May 2012; accepted 21 May 2012Published as BJ Immediate Publication 21 May 2012, doi:10.1042/BJ20120505


Glutathionylation of cytosolic glyceraldehyde-3-phosphate dehydrogenase from the model plant Arabidopsis thaliana is reversed by both glutaredoxins and thioredoxins in vitro

Documents