Plant Molecular Biology 52: 591–603, 2003.© 2003 Kluwer Academic Publishers. Printed in the Netherlands.

591

Cloning, characterization and regulation of a family of phi classglutathione transferases from wheat

Ian Cummins1, David O’Hagan2, Istvan Jablonkai3, David J. Cole4, Alain Hehn5, DanieleWerck-Reichhart5 and Robert Edwards1,∗1Crop Protection Group, School of Biological and Biomedical Sciences, University of Durham, Durham, DH1 3LE,UK (∗author for correspondence; e-mail Robert.Edwards@durham.ac.uk); 2Department of Chemistry, Universityof St. Andrews, St. Andrews, Fife, KY16 9ST, UK; 3Chemical Research Centre, Hungarian Academy of Sciences,I025 Budapest, Hungary; 4Formerly of Aventis Crop Science, Fyfield Road, Ongar, Essex, CM5 OHW, UK; 5CNRS-IBMP UPR2357, Department Plant Stress Response, Universite Louis Pasteur, 28 Rue Goethe, 67000 Strasbourg,France

Received 10 September 2002; accepted in revised form 28 January 2003

Key words: chalcones, flavonoids, glutathione transferases, herbicides, ligandin, S-isoliquirtigenin-glutathione,Triticum aestivum, xenobiotics, Zea mays

Abstract

Six phi (F) class glutathione transferases (GSTs) were cloned from bread wheat (Triticum aestivum L.) treatedwith the herbicide safener fenchlorazole ethyl and named TaGSTF1–6. Recombinant TaGSTFs were assayedfor glutathione conjugating activity towards xenobiotics including herbicides and for glutathione peroxidase(GPOX) activity. TaGSTF1, which resembled ZmGSTF1, the dominant GST in maize (Zea mays), was highlyactive in conjugating 1-chloro-2,4-dinitrobenezene (CDNB) but had low activities towards chloroacetanilide,diphenyl ether and aryloxphenoxypropionate herbicides. TaGSTF2, TaGSTF3 and TaGSTF4 all resembled thesafener-inducible ZmGSTF2, with TaGSTF2 and TaGSTF3 being highly active GPOXs and rapidly detoxifyingchloroacetanilides. TaGSTF5 resembled ZmGSTF3, having limited conjugating and GPOX activity. TaGSTF6contained both ZmGSTF1- and ZmGSTF2-like sequences but was most similar to ZmGSTF1 in detoxifying activity.The expression of TaGSTFs in wheat seedlings was enhanced upon exposure to fenchlorazole ethyl, herbicides orother chemical inducing treatments. TaGSTFs were also enhanced by treatment with the natural products caffeicacid, 7,4-dihydroxyflavone and naringenin. The CDNB-conjugating activity of TaGSTF1, and to a lesser extentTaGSTF6, was highly sensitive to inhibition by flavonoids, particularly the chalcone isoliquiritigenin. The otherTaGSTFs were much less sensitive to such inhibition. It was subsequently determined that isoliquiritigenin under-went glutathione conjugation, though this reversible reaction did not require the intervention of any TaGSTF. Thepotential importance of GSTFs and glutathione conjugation in flavonoid metabolism is discussed.

Introduction

Glutathione transferases (GSTs, EC 2.5.1.18) are adiverse group of enzymes, each composed of twosubunits, which detoxify herbicides in plants bycatalysing their conjugation with the tripeptide glu-tathione (Edwards et al., 2000). Based on sequencesimilarity and gene organisation, plant GSTs can be di-vided into at least four classes, the phylogeny of whichhas recently been comprehensively reviewed (Dixon

et al., 2002). The phi (F) class GSTFs and the tau(U) class GSTUs are the most numerous and have thegreatest detoxifying activity towards herbicides (Ed-wards and Dixon, 2000). These two classes of GSTsare only found in plants and are present in large multi-gene families in Arabidopsis (Wagner et al., 2002),maize and soybean (McGonigle et al., 2000). Manyadditional DNA sequences apparently encoding GS-TUs and GSTFs have been reported as expressed se-

592

quence tags (ESTs) or genomic sequences in a diverserange of plants (Marrs, 1996). Similarly, the presenceof GST activities toward xenobiotics in diverse lowerand higher plants suggests that these are ubiquitousenzymes in the plant kingdom (Pflugmacher et al.,2000).

GSTUs and GSTFs have additional, less well char-acterised roles in both endogenous and stress-inducedmetabolism. In the case of secondary metabolism,a GSTU in maize, termed bronze II (Bz2), and aGSTF in Petunia, termed An9, have essential andcomplementary functions in ensuring the vacuolar de-position of anthocyandin glycosides (Alfenito et al.,1998). Although it was first assumed that Bz2 andAn9 conjugate anthocyanidins with glutathione, itwas subsequently demonstrated that these GSTs actas specific flavonoid-binding proteins (Mueller et al.,2000). GSTFs and GSTUs also have important rolesin preventing oxidative stress in plants by functioningas glutathione peroxidases (GPOXs), enzymes whichuse glutathione to reduce organic hydroperoxides tothe respective monohydroxy alcohols (Edwards et al.,2000). Thus, GSTFs in both tobacco (Roxas et al.,1997) and herbicide-resistant black-grass (Cumminset al., 1999) protect plants from the oxidative injuryincurred by salt or temperature stress and herbicideinjury respectively. Similarly, a GSTU from tomatoprevents apoptosis when expressed in yeast cells ap-parently by preventing oxidative damage (Kampraniset al., 2000). GSTs also have roles in stress signallingin plants, with a GSTU having an essential role in theinduction of chalcone synthase and flavonoid accumu-lation in parsley upon exposure to UV light (Loyallet al., 2000).

To date, the best studied group of phi andtau GSTs have been those involved in herbicidemetabolism in maize (Zea mays L.). Genes encodingthree phi polypeptides termed ZmGSTF1, ZmGSTF2and ZmGSTF3 and four tau polypeptides ZmGSTU1,ZmGSTU2, ZmGSTU3 and Bz2 have been cloned,and the enzymes containing these subunits charac-terised in considerable detail (Edwards et al., 2000).A genomics EST approach has recently added to ourknowledge of the maize GSTs by cloning and express-ing an additional 9 phi and 24 tau GSTs (McGonigleet al., 2000). Although at the level of gene number GS-TUs are the most abundant class in maize (McGonigleet al., 2000), GSTFs predominated in the EST data-base and at the level of expressed protein (Edwardsand Dixon, 2000). This has important implicationsin determining herbicide selectivity, as members of

theZmGSTF family are particularly active in detoxi-fying both chloroacetanilide and thiocarbamate herbi-cides (Holt et al., 1995; Dixon et al., 1997). This isparticularly the case with isoenzymes containing theZmGSTF2 subunit, which accumulate in the foliageof maize plants after application of herbicide safen-ers, compounds which enhance herbicide tolerance incereals by increasing the expression of detoxifyingenzymes such as GSTs (Davies and Caseley, 1999).

It has recently become apparent that glutathioneconjugation and GSTs are also important in themetabolism of selective herbicides in other cereals, in-cluding hexaploid bread wheat (Triticum aestivum L.).Thus, rapid glutathione conjugation of the chloroac-etamide, dimethenamid (Riechers et al., 1997a), thearyloxyphenoxypropionate fenoxaprop-ethyl, (Thomet al., 2002) and the sulphonyl urea flupyrsulfuron-methyl (Koeppe et al., 1997) are major determinantsof the detoxification and selectivity of these herbi-cides in wheat. Significantly, both dimethenamid andfenoxaprop can only be used selectively when for-mulated with safeners and inducible GSTUs, whichappear to account for the safener-enhanced detoxi-fication of these herbicides, have been identified inwheat and its progenitor Triticum tauschii (Riech-ers et al., 1997a; Cummins et al., 1997; Pascaland Scalla, 1999). These GSTUs were subsequentlycloned and characterised and found to closely resem-ble the GSTUs in maize, based on sequence sim-ilarity, organization and safener inducibility (Thomet al., 2002). During the purification of the TaGSTUs,several polypeptides, which were immunologically re-lated to phi-class maize GSTs, were also identified inwheat shoot extracts (Cummins et al., 1997). Morerecently, safener-inducible GST polypeptides with N-terminal sequences resembling ZmGSTF2 have beenidentified in wheat (Pascal et al., 1998).

We now report the identification and cloning ofa family of phi-class GSTs in wheat. The expres-sion of these polypeptides has been studied in plantaand the respective recombinant enzymes characterisedwith respect to their detoxifying activity towards xeno-biotics including herbicides and substrates resemblingnaturally occurring toxins. We have also determinedthe activity of these enzymes as GPOXs and exam-ined their potential to bind to natural products. In thecourse of these studies we have identified that at leastone of the GSTFs has the characteristics of being aflavonoid-binding protein and that one class of nat-ural products, the chalcones, can undergo reversibleglutathione conjugation in vitro.

593

Materials and methods

Plant studies

Wheat (Triticum aestivum L. cv. Hunter) seedlingswere grown with or without treatment with the wheatsafener fenchlorazole ethyl (32.5 µM) in a growthcabinet and harvested as described previously (Cum-mins et al., 1997). For short-term induction studies,seedlings grown without chemical treatments were cutat soil level 7 days after sowing and the cut stems stoodin centrifuge tubes containing aqueous treatment solu-tions made up either directly in water or by a 1000-folddilution of stock solutions prepared in acetone. Thecuttings were then returned to the growth cabinets andat timed intervals, blotted dry, weighed and frozen inliquid N2 prior to storage at −80 ◦C.

Immunoblotting with anti-GST sera

Antisera raised in rabbits to maize ZmGSTF1-2(Dixon et al., 1998) and wheat TaGSTU1-1 (Cumminset al., 1997) were available from previous studies.For immunoblotting experiments, frozen plant tissuewas homogenised in 0.1 M Tris-HCl pH 7.5 contain-ing 2 mM EDTA, 1 mM dithiothreitol (DTT) andpolyvinylpolypyrrolidione (5% w/v). After centrifu-gation (12 000 × g, 20 min), the supernatant wasadjusted to 80% saturation with respect to (NH4)2SO4and the protein precipitate collected by centrifugation.The protein was re-suspended in 20 mM Tris-HClpH 7.5 containing 1 mM DTT and dialysed againstthe same buffer overnight prior to determining proteincontent with the BioRad dye-binding assay with γ -globulin as reference protein. Samples were adjustedto 5 mg/ml protein and 30 µg applied to a 17.5% SDS-PAGE gel (Mozer et al., 1983). After electrophoresis,polypeptides were electroblotted onto nitrocellulosemembranes (Hybond C, Amersham) and probed withthe primary antisera, followed by a secondary anti-rabbit IgG serum coupled with horseradish peroxi-dase (Cummins et al., 1999). The antibody-antigencomplexes were then visualised by enhanced chemi-luminescence assay as detailed by the manufacturer(Amersham).

Isolation of cDNAs encoding TaGSTFs

Poly(A)+ RNA was isolated from the shoots of wheatseedlings, grown for 7 days in the presence of 32.5 µMfenchlorazole ethyl and used to construct a unidi-rectional cDNA library in the λUNI-ZAP XR vector

as recommended by the manufacturer (Stratagene).Membrane lifts of 200 000 plaque-forming units wereimmunoscreened with the anti-ZmGSTF1-2 serum es-sentially as described previously (Dixon et al., 1998).Positively hybridising plaques were grouped into dif-ferent classes depending on the relative intensities ofimmunodetection and purified through two rounds ofscreening. Clones were then sequenced in both direc-tions using an ABI automated sequencer and analysedwith the BLAST program (Altschul et al., 1990).

Expression and purification of recombinant GSTs

The GST clones derived from immunoscreening wereexpressed in Escherichia coli SOLR cells as theirrespective β-galactosidase fusion proteins using thepBluescript plasmid. As required, these GST-codingsequences were also sub-cloned into the pET 11d plas-mid for expression of the native recombinant polypep-tides. Bacteria transformed with GST cDNAs werecultured for 16 h in LB medium (50 ml) containing1 mM IPTG, then harvested by centrifugation. Bac-teria were re-suspended in 100 mM Tris-HCl pH 7.5containing 2 mM EDTA, 1 mM DTT and lysedby ultrasonication. After centrifugation (10 000 × g,20 min), (NH4)2SO4 was added to 80% saturationand the protein precipitate collected by centrifugation(17 000 × g, 20 min). Re-suspended protein was thendialysed against 20 mM potassium phosphate bufferpH 7.2 containing 1 mM DTT prior to applicationto 3 ml affinity columns at a flow rate of 1 ml/min.Affinity columns were prepared by coupling eitherglutathione (Simons et al., 1981), S-hexyl-glutathione(Cummins et al., 1997), or S-sulphobromophthaleinglutathione (Mozer et al., 1983) to epoxy-activatedSepharose. In each case, after washing away un-bound protein, affinity-bound GSTs were recoveredby addition of 5 mM glutathione (glutathione and S-sulphobromophthalein glutathione matrices) or 5 mMS-hexyl glutathione to the loading buffer. The GSTswere then further purified by anion exchange chro-matography using Q-Sepharose and their purity con-firmed by SDS-PAGE as described for the recombi-nant TaGSTUs (Thom et al., 2002).

Enzyme and inhibition assays

GST activities toward herbicides were determined byHPLC to quantify reaction products, with activitiestowards other xenobiotics determined spectrophoto-metrically (Cummins et al., 1997). GPOX activ-ity was determined with cumene hydroperoxide and

594

with linoleic acid hydroperoxide (13-hydroperoxy-cis-9,trans-11-octadecadienoic acid) as substrates respec-tively (Cummins et al., 1997), the latter preparedfrom linoleic acid using soybean lipoxidase (Graffet al., 1990). Determinations of kinetic constantsfor GST activity toward glutathione and 1-chloro-2,4-dinitrobenzene (CDNB) were derived by varyingsubstrate concentrations in the range 0.25–5 mM andanalysing the data by using double reciprocal plots.For inhibition studies, flavonoids and other plant sec-ondary metabolites were obtained from Apin chemi-cals (Oxon, UK) and prepared as stock solutions inmethanol. Using CDNB and glutathione in GST as-says (total volume 1 ml), natural products were added(10 µl) over a series of concentrations and the IC50(concentration giving 50% inhibition of enzyme activ-ity) determined in each case. The type of inhibitionobserved was determined by varying the substrate con-centrations while holding the inhibitor concentrationconstant and deriving double reciprocal plots.

Screening for glutathione conjugates of naturalproducts

Crude wheat protein preparations and purified recom-binant enzymes were prepared in 50 mM Tris-HClpH 7.5 to a final specific activity of 25 nkat/mlGST activity (with CDNB as substrate) and 70 µlof each preparation incubated with 10 µl glutathione(100 mM) and 10 µl of the natural product (100 mM)dissolved in methanol. After 60 min, 5 µl of the re-action mixture was applied to a silica gel TLC plate,while the remainder was diluted 10-fold with methanolprior to analysis by HPLC. For TLC, the silica gelplates containing fluorescent indicator (Merck) weredeveloped with solvent 1 (butan-1-ol/acetic acid/water4:1:1 v/v). Plates were analysed for UV-absorbingmetabolites prior to spraying with a solution of nin-hydrin dissolved in acetone (0.3% w/v) followed byincubation at 100 ◦C for 3 min. For HPLC analy-sis, the reaction products (50 µl) were applied onto areversed-phase column (octadecyl, 250 mm × 4.6 mmi.d., 5 µm particle size, Phenomenex). Compoundswere then eluted at 0.8 ml/min with a linearly increas-ing gradient of organic solvent, starting with solventsA and B (80:20 v/v) and ending at 45 min with sol-vents A and B (40:60 v/v), with solvent A consistingof water containing 1% v/v orthophosphoric acid andsolvent B being acetonitrile. Eluting compounds weredetected by their absorbance at 287 nm.

Preparation of S-isoliquiritigenin-glutathione

Isoliquiritigenin (4,2′,4′-trihydroxchalcone, 10 mg)was dissolved in 1 ml of methanol and added to0.1 mmol glutathione dissolved in 3 ml of 25 mMpotassium phosphate buffer pH 7.5. After incubat-ing at 30 ◦C for 24 h, the reaction products weredirectly applied onto a preparative HPLC column (Par-tisil ODS 3, 150 mm × 10 mm i.d., Alltech) elutedunder the solvent conditions employed for analyticalHPLC at a flow rate 2.5 ml/min, with the phosphoricacid replaced with 0.5% v/v trifluoroacetic acid. Themajor UV-absorbing peak, eluting between 8 and 10min was collected and freeze dried to yield a whitepowder (9 mg). The compound was analysed byUV spectroscopy (UV absorbance maximum 281 nm,with a shoulder maximum of 319 nm). Fast-atombombardment mass spectrometry in the presence ofthioglycerol showed a parent mass ion of 564 Da,consistent with the formation of an addition productformed between isoliquirtigenin and glutathione. Theconjugate was then dissolved in D2O and analysed by1H NMR (400 MHz). (2H, m, CH2(glu)), 2.3 (2H,d.d, CH2 (glu)), 2.65 (2H, 2xd.d, CH2S), 3.35 (2H,d.d, C(8)H2CO), 3.7 (1H, m, CH(glu)), 3.78 (2H, s,CH2(gly)), 4.2–4.25 (1H, 2 × d.d, CH(Cys)), 4.25(1H, m, C(9)HS).

Results

Induction of GSTU and GSTF polypeptides bychemical treatments

Changes in the abundance of tau- and phi-class GSTsin wheat shoots exposed to a range of chemical treat-ments were monitored using two antisera. To deter-mine GSTUs, an antiserum was raised to the majorconstitutive TaGSTU1-1 isoenzyme (Cummins et al.,1997). This antiserum has been shown to be specific tothe 25 kDa TaGSTU1 polypeptide and does not cross-react with other GST subunits (Thom et al., 2002).For the GSTFs, an antiserum to the dominant safener-inducible maize ZmGSTF1-2 isoenzyme was used(Dixon et al., 1998). In maize extracts, this antiserumrecognises both the constitutively expressed 29 kDaZmGSTF1 subunit as well as the safener-induced27 kDa ZmGSTF2 subunit. We have also demonstratedin previous studies that this antiserum recognises adistinct set of GST polypeptides to the anti-TaGSTU1serum (Cummins et al., 1997). The wheat shoots

595

Figure 1. Induction of TaGSTs in wheat shoots exposed to chemical treatments as determined by SDS-PAGE followed by immunoblotting withan antiserum raised either to ZmGSTF1-2 (anti-phi) or the TaGSTU1-1 (anti-tau). In each case identical amounts (50 µg) of crude protein wereloaded onto the gels. Relative molecular masses of the immunodetected polypeptides are indicated. Treatments were A. Wheat seedlings weregrown for 7 days treated with either a control (C) solution of 0.1% v/v acetone in water or the safener (S) fenchlorazole ethyl (32.5 µM) dissolvedin 0.1% v/v acetone. B. Cut shoots from 7-day old wheat seedlings stood in treatment solutions of (1) water, (2) metolachlor (0.1 mM), (3)paraquat (0.5 mM), (4) 2,4-D (0.1 mM), (5) ethanol (1% v/v), (6) salicylic acid (1 mM), (7) fluorodifen (0.1 mM) and (8) glutathione (5 mM).C. Cut shoots from 10-day old wheat seedlings were stood in (1) 1% v/v acetone in water or 0.2 mM solutions in 1% v/v acetone of (2)7′,4′-dihydroxyflavone, (3) biochanin A, (4) naringenin flavanone and (5) caffeic acid.

were exposed to a range of treatments known to per-turb the expression of GSTs in maize plants (Jepsonet al., 1994; Dixonet al., 1998). Treatments includeda safener, fenchlorazole ethyl, the herbicides meto-lachlor, paraquat, 2,4-dichlorophenoxyacetic acid andfluorodifen, as well as glutathione, ethanol and sal-icylic acid. The 25 kDa polypeptide recognised by

the anti-TaGSTU1-1-serum was most strongly en-hanced by exposure to fenchlorazole ethyl, ethanol,fluorodifen or glutathione, with the other compoundshaving only a weak effect (Figure 1). With the anti-ZmGSTF1-2-serum, a 24 kDa polypeptide was recog-nised in all preparations (Figure 1). The abundanceof this polypeptide was modestly enhanced by sev-

596

eral of the treatments including salicylic acid andall four herbicides. In the fenchlorazole ethyl-treatedplants, an additional 27 kDa polypeptide was recog-nised by the anti-ZmGSTF1-2-serum, which was oflow abundance in the untreated controls (Figure 1A).This 27 kDa polypeptide was also strongly enhancedin seedlings exposed to fluorodifen, with a weakerinduction seen with the other herbicides, ethanol, orsalicylic acid (Figure 1B). These studies demonstratedthat both GSTF and GSTU polypeptides in wheatwere induced in response to treatments either with thesafener, the photobleaching herbicides paraquat andfluorodifen and ethanol, but that the other treatmentsgave selective enhancement of one class of GST only.

In addition to analysing the enhancement of wheatGSTs by agents known to induce these enzymes inmaize, it was also of interest to test the inducingactivities of natural products such as flavonoids andrelated compounds, which are known to bind to spe-cific phi- and tau-class GSTs from petunia and maize(Mueller et al., 2000). The compounds chosen forthe induction study were representative of a diverserange of phenolic metabolites, notably the flavone7,4′-dihydroxyflavone, the isoflavone biochanin A, theflavanone naringenin and the phenylpropanoid caffeicacid. Treatment with 7,4′-dihydroxyflavone, narin-genin and caffeic acid enhanced the expression of bothclasses of GST polypeptides, while biochanin A hadno noticeable effect (Figure 1C). As determined by as-saying for GST activity towards the general substrateCDNB, the enhancement in the GST polypeptides wasassociated with an increase in conjugating activityfrom 8.35 µkat per gram protein in plants treated withsolvent carrier alone, to 12.7 µkat/g after flavonoidtreatment.

Cloning of ϕ-class GSTs from wheat

A cDNA library prepared from 10-day old wheatseedlings treated with the herbicide safener fenchlo-razole ethyl was immunoscreened with the anti-ZmGSTF1-2 serum. From the membrane lifts, 8 posi-tive plaques were identified and the respective clonespurified and sequenced. Three of these cDNAs hadidentical coding sequences but differed in their 5′-untranslated regions which surprisingly contained stopcodons in each instance, even though some translationof immunoreactive protein was still clearly occurring.BLAST searches with the predicted amino acid se-quence showed these cDNAs had greatest sequencesimilarity to the maize GST ZmGSTF1 (76% iden-

tity) and they were termed TaGSTF1 clones (accessionnumber AJ440796).

A further two clones were also identified from theirpredicted sequences as phi-class GSTs. These cD-NAs had similar sequences to one another and wereboth related to the safener-inducible maize ZmGSTF2subunit. These two clones were named TaGSTF2(AJ440791) and TaGSTF3 (AJ440792) and showed53% and 71% identity in sequence to ZmGSTF2 re-spectively. In an independent study, a further mem-ber of the ZmGSTF2-like phi GSTs in wheat wasidentified from a low-stringency screen of a cDNAlibrary prepared from wheat seedlings treated withthe safener cloquintocet mexyl screened with het-erologous probes directed to cytochrome P450 mixedfunction oxidases (Cabello-Hurtado et al., 1997). ThiscDNA was termed TaGSTF4 (AJ440793) and showed61% identity to ZmGSTF2. Of the remaining threeclones identified from the immunoscreen, one cDNA,TaGSTF5 (AJ440794), showed 64% sequence iden-tity to the phi maize ZmGSTF3. The remaining lasttwo clones had identical coding sequences to one an-other, but differed in their untranslated region. Thecoding sequence of these cDNAs showed greatestsimilarity to a previously identified phi wheat GSTtermed GSTA1 (Mauch and Dudler, 1993). The cod-ing sequence of these clones was collectively termedTaGSTF6 (AJ440795) and showed similarities in dif-fering regions to the sequences of both ZmGSTF1 andZmGSTF2. Alignments of the sequences of all thedifferent TaGSTFs are shown in Figure 2.

Enzyme activities of recombinant GSTs

The phi wheat GSTs identified by immunoscreen-ing were expressed in E. coli. as their respectiveβ-galactosidase fusion proteins, while TaGSTF4 wassub-cloned directly into the pET 11d vector for expres-sion of the native protein. The recombinant GSTs werepurified from the bacterial lysates by affinity chro-matography, using glutathione, or a glutathione conju-gate linked to Sepharose, followed by anion-exchangechromatography. When analysed by SDS-PAGE, thenative TaGSTF4 ran as a single polypeptide with amolecular mass of 27 kDa (Figure 3). In contrast, allthe β-galactosidase-GST fusions, with the exceptionof TaGSTF1, which ran as a single 24 kDa polypep-tide, migrated as polypeptide doublets with molecularmasses in the range 26–31 kDa. It was concludedthat the doublet of polypeptides probably arose fromthe initiation of translation from both the AUG of

597

Figure 2. Aligned predicted amino acid sequences of TaGSTFs. TaGSTF1 (AJ440796), TaGSTF2 (AJ440791), TaGSTF3 (AJ440792),TaGSTF4 (AJ440793), TaGSTF5 (AJ440794), TaGSTF6 (AJ440795). ∗ denotes conserved residues, .denotes conservative substitutions.

the β-galactosidase N-terminal fusion as well as thatof the GST sequence, giving the larger and smallerpolypeptides of the doublet respectively (Figure 3).In each case, when the purified recombinant GSTpreparations were analysed by gel permeation chro-matography, UV-absorbing protein and GST activitytowards CDNB eluted as a single peak of ca. 50 kDa(data not shown). This confirmed that all the wheatGSTs were active as the respective dimers and that theβ-galactosidase N-terminal fusions had not affectedsubunit association. Thereafter, the functional GSTswere therefore described as the respective dimers interms of their subunit composition. For example, thedimer derived from the TaGSTF1 clone was describedas the TaGSTF1-1 recombinant enzyme.

Having established their purity, the panel of phi-class wheat GSTs were characterised with respect totheir enzyme activities towards a range of xenobi-otic and naturally occurring substrates. In the caseof the xenobiotics, the model GST substrate CDNBwas used, together with the structurally diverse herbi-cides fenoxaprop ethyl, fluorodifen and metolachlor,representing examples of aryloxyphenoxypropionate,

Figure 3. SDS-PAGE analysis of purified recombinant TaGSTFs.Recombinant GSTs were purified by a combination of affinity chro-matography and ion exchange chromatography and then analysedon 17.5% acrylamide gels. Lanes: 1, TaGSTF4; 2, TaGSTF2; 3,TaGSTF3; 4, TaGSTF1; 5, TaGSTF6; 6, TaGSTF5. The relativemigration of reference proteins of known molecular masses areindicated.

598

Table 1. Detoxifying activity of pure recombinant TaGSTFs toward xenobi-otics, including herbicides, model stress metabolites and organic hydroperox-ides.

Substrate TaGST enzyme activity (nkat per mg pure protein)

F1-1 F2-2 F3-3 F4-4 F5-5 F6-6

XenobioticsCDNB 2519 407 187 237 174 980

Fluorodifen 0.01 0.14 0.1 0.03 0.03 0.04

Fenoxaprop 0.09 0.05 0.09 0.04 0.07 0.01

Metolachlor 0.05 0.32 0.51 0.06 0.04 0.04

BITC 9.0 18.0 33.3 22.4 9.4 4.2

Stress metabolitesCrotonadehyde 6.3 7.1 5.5 6.1 4.5 5.5

Ethacrynic acid ND ND ND ND ND 1.4

Glutathione peroxidase substratesCum-OOH 3.8 50.0 90.0 9.7 3.5 16.7

Lin-OOH 1.3 63.0 236 3.2 1.8 2.6

Values refer to means of duplicate determinations. ND = not detected

diphenyl ether and chloroacetanilide chemistries re-spectively. In addition, xenobiotics were includedwhich have similarities to naturally occurring com-pounds. Benzylisothiocyanate was included as similarchemicals are released as allelochemicals from glu-cosinolate precursors in Brassica species (Bartlinget al., 1993). Crotonaldehyde and ethacrynic acidwere used as examples of α,β-unsaturated alkenals,analogous to cytotoxic stress metabolites, which ac-cumulate as a result of lipid peroxidation (Berhaneet al., 1994). In addition, the GSTs were also as-sayed for their GPOX activity toward the naturally oc-curring linoleic acid hydroperoxide (13-hydroperoxy-cis-9,trans-11-octadecadienoic acid) and syntheticcumene hydroperoxide (Graff et al., 1990).

To ensure that the N-terminal β-galactosidase fu-sions did not affect the enzyme activity of the GSTs,the coding sequence of TaGSTF1 was cloned directlyinto pET 11 to give expression of the native polypep-tide. When assayed for activity, the native recombinantTaGSTF1-1 was found to have identical activities tothose quoted for the β-galactosidase-TaGSTF1 fu-sion in Table 1. It was therefore concluded that theN-terminal fusions had a negligible effect on theGST activities and that by whatever mechanism, thepolypeptide translated from the TaGSTF1-Bluescriptvector which contained a stop codon in the interced-ing sequence between the N-terminal fusion and the

GST open reading frame was a catalytically functionalGST. TaGSTF1-1, which resembles ZmGSTF1-1 insequence, had a very similar range of activities to-ward xenobiotics to the maize enzyme (Dixon et al.,1997), being very active toward CDNB, but show-ing low activities toward all the herbicides tested.However, TaGSTF1-1 did differ from ZmGSTF1-1in having measurable GPOX activity. The ZmGSTF2homologues in wheat, TaGSTF2-2 and TaGSTF3-3,were highly active in conjugating the herbicide meto-lachlor and in reducing the organic hydroperoxides.Similarly, the safener-inducible GSTs in maize con-taining the ZmGSTF2 subunit show similar detoxify-ing activities (Holt et al., 1995; Dixon et al., 1997).However, the related TaGSTF4-4 enzyme had lowactivities toward herbicide and organic hydroperox-ide substrates, and instead resembled TaGSTF5-5,which was most similar in sequence to ZmGSTF3-3,a phi-class maize GST with low activities toward mostxenobiotic substrates (Dixon et al., 1997). TaGSTF6-6, whose subunits showed sequence similarities toboth ZmGSTF1 and ZmGSTF2, was most similarin substrate preferences to ZmGSTF1-1, having highactivities towards CDNB and low activities towardsherbicides. However, TaGSTF6-6 had higher GPOXactivity than ZmGSTF1 homologues and, uniquelyamong all the TaGSTFs isolated, catalysed the con-jugation of ethacrynic acid.

599

Figure 4. Proposed reaction for the reversible formation of S-isoliquiritigenin-glutathione.

Table 2. Kinetic and inhibition characteristics offour pure recombinant GSTFs, representing each ofthe type of GSTF sequences seen in wheat. Kineticconstants were determined with CDNB and glu-tathione. In the inhibition assays, the GSTs were as-sayed with CDNB and glutathione and the concen-tration of inhibitor required to give 50% reduction inactivity determined.

TaGSTF

F1-1 F3-3 F5-5 F6-6

Kinetic constantsKm(GSH), mM 0.18 0.16 0.64 0.31

Kcat(GSH), s−1 67.9 17.7 10.3 27.8

Km(CDNB), mM 4.43 1.07 0.34 0.74

Kcat(CDNB), s−1 321.4 34.5 7.2 44.4

Inhibitor

Inhibitors IC50 (µM)

Isoliquiritigenin 4.7 >10 7.2 4.1

Naringenin 37 >200 >200 89

Quercetin 7.1 67 78 4.4

Luteolin 16 47 47 9.6

Biochanin A 46 213 118 133

2,4-D 1500 2200 920 830

Chlorophyllin 1.2 2.2 0.8 1.4

Cyanidin-3-glucoside 42 >250 >250 >250

Inhibition of TaGSTFs by natural products and theiranalogues

Having determined the GST and GPOX activity ofthe TaGSTF family towards a range of xenobiotics,it was then of interest to determine if these enzymescould bind phenolic metabolites found in a diverserange of plants and the effect of this binding onGST activity. Based on sequence similarities withtheir counterparts in maize, representatives of the fourgroups of TaGSTFs, namely TaGSTF1, TaGSTF3,

TaGSSTF5 and TaGSTF6, were characterised withrespect to their kinetic constants with CDNB and glu-tathione as substrates (Table 2). All the enzymes hadcomparable affinities and turnover numbers for glu-tathione. TaGSTF1-1 differed from the other enzymesin having a considerably higher turnover number withCDNB. The four pure GSTs were then incubated withCDNB and glutathione in the presence of a rangeof compounds which have been reported to inhibitGSTs including flavonoids (Mueller et al., 2000),the synthetic auxin 2,4-dichlorophenoxyacetic acid(Nishihira et al., 1995) and the tetrapyrrole chloro-phyllin (Singh and Shaw, 1988). The concentrationgiving 50% inhibition of activity (IC50) was then cal-culated (Table 2). The flavonoids showed a similarranking in their ability to inhibit GST activities ofall the TaGSTFs tested with efficacies in the orderchalcone > flavones > flavanone = isoflavone =anthocyanidin glucoside. TaGSTF1-1 was markedlymore sensitive than TaGSTF3-3 and TaGSTF5-5 toinhibition by all flavonoid compounds. TaGSTF6-6showed similar sensitivities to TaGSTF1-1 to the chal-cone, flavones and flavanone but was only inhibitedby the anthocyanidin glucoside at high concentra-tions. All the TaGSTFs were insensitive to inhibitionby 2,4-D but strongly inhibited by chlorophyllin. Inall cases Lineweaver-Burke analysis suggested mixedinhibition was occurring.

Identification of S-glutathionylated flavonoids

To clarify the status of flavonoids as substrates forGST-mediated glutathione conjugation, the flavonoidstested as enzyme inhibitors (Table 2) were individuallyincubated with either glutathione alone, glutathioneand each of the six pure recombinant TaGSTFs, ora crude GST preparation from safener-treated wheatshoots. The reaction mixture was then analysed fornovel polar metabolites by a combination of TLC and

600

HPLC. In the case of naringenin flavanone, quercetin,luteolin, biochanin A and cyanidin glucoside, theprofile of UV-absorbing metabolites was unchangedover the course of the incubation, demonstrating thatno glutathione conjugation of the flavonoids had oc-curred. However, when the incubations with the chal-cone isoliquiritigenin were analysed by TLC (solvent1), in addition to the parent yellow-pigmented chal-cone (Rf = 0.87), a new polar UV-absorbing metabo-lite was observed (Rf = 0.39). This metabolite reactedpositively with ninhydrin, indicating the presence ofamino groups. When analysed by HPLC, the polarmetabolite eluted with a retention time of 10 minwhile authentic isoliquiritigenin eluted at 28.5 min.As determined by HPLC, the amount of metaboliteformed was identical in all incubations, irrespectiveof the presence or absence of any of GST activ-ity, indicating that polar product formation was notdue to enzymic catalysis. The polar metabolite waspurified by preparative HPLC and analysed by FAB-MS, giving a mass of 564 Da in positive ion mode.The metabolite was then analysed by 1H-NMR. Thespectra obtained were consistent with equimolar quan-tities of two diastereoisomers of S-glutathionylatedisoliquirtigenin, conjugated at the β-position of theα,β-unsaturated moiety (Figure 4). An interesting fea-ture of the metabolite was that, upon analysis byTLC, exposure to air resulted in the initially colourlesscompound turning bright yellow and this was asso-ciated with a shift in the absorbance maxima from281 nm to 365–375 nm. When re-analysed by HPLCand TLC, this yellow derivative co-chromatographedwith authentic isoliquiritigenin, demonstrating that theS-glutathionylation reaction was reversible (Figure 4).

Discussion

Our results demonstrate that wheat contains a familyof phi-class GSTs which in terms of their amino acidsequence and detoxifying activities toward xenobioticsand herbicides are organised in a similar manner tothat seen with the dominant family of GSTFs ex-pressed in maize (Holt et al., 1995; Dixon et al., 1997;McGonigle et al., 2000). The most abundant phi-class GSTs expressed in maize plants can be groupedinto the ZmGSTF1, ZmGSTF2 and ZmGSTF3 types.When the TaGSTFs were compared with their maizecounterparts, pairings of TaGSTF1-1 with ZmGSTF1-1, TaGSF2-2 and TaGSTF3-3 with ZmGSTF2-2, andTaGSTF5-5 with ZmGSTF3-3 could be clearly identi-

fied. The similarity in enzyme activity for each set ofrelated enzymes was particularly striking, indicatingthat the active sites of related maize and wheat GSTFsmust also be conserved. The exception to this com-monality in sequence and activity for the GSTFs of thetwo species was seen with TaGSTF4, which, thoughclosely resembling ZmGSTF2 in sequence, showedactivities more akin to ZmGSTF3. In addition, wheatcontained TaGSTF6, which was a ‘hybrid’ phi GST,derived from ZmGSTF1-like and ZmGSTF2-like se-quences respectively, but which possessed enzymeactivities most like ZmGSTF1. The results of ourcloning studies also relate well to earlier biochemi-cal studies, which identified six GSTF polypeptidesin wheat seedlings treated with the safener naph-thalic anhydride (Pascal et al., 1998). These GSTFscould be divided into two groups of three. The firstgroup was constitutively expressed and was composedof polypeptides ca. 23.2 kDa, all having identicalN-terminal sequences. The second group of polypep-tides of ca. 24.9 kDa, only accumulated upon safenerapplication and showed some variety in N-terminalsequences. Based on their N-terminal sequence andthe predicted sizes of the encoded polypeptides, theconstitutive GSTs appear to be identical to TaGSTF1,while the safener-inducible polypeptides were mostalike TaGSTF6. The studies in planta therefore con-firm that several of the GSTFs cloned are among themajor phi-class enzymes expressed in safener-treatedwheat plants.

Previous studies on herbicide-detoxifying GSTsin wheat have concentrated on the constitutive andsafener-inducible tau-class enzymes, with the majorisoenzymes purified to homogeneity (Riechers et al.,1997a, Cummins et al., 1997; Pascal and Scalla,1999). The GSTUs were subsequently cloned and therecombinant proteins characterised (Riechers et al.,1997b; Thom et al., 2002). Reviewing the data ob-tained from the biochemical and cloning studies, it isclear that the dominant proportion of the profile of her-bicide detoxifying activities seen in both untreated andsafener-treated wheat (Riechers et al., 1997a, Cum-mins et al., 1997) is contributed by the TaGSTUsrather than by the TaGSTFs. This is in contrast tomaize, where the GSTFs are more abundant than theGSTUs being responsible for the rapid metabolism ofchloroacetanilide and thiocarbamate herbicides (Holtet al., 1995; Dixon et al., 1997; Irzyk and Fuerst,1993). This difference in the overall abundance ofphi- and tau-class GSTs, allied to their respectivesubstrate preferences would account for the differ-

601

ing rates of detoxification and hence selectivity ofchloroacetanilide and aryloxyphenoxypropionate her-bicides determined in these two major cereal crops(Edwards and Dixon, 2000).

Our studies on the regulation of TaGSTUs andTaGSTFs showed that both classes of GSTs wereselectively inducible in wheat seedlings. Some treat-ments, such as the safener fenchlorazole ethyl andfluorodifen, enhanced both classes of GST. Similarly,a number of studies in wheat have separately reportedthe safener enhancement of either GSTUs (Riecherset al., 1997a; Cummins et al., 1997; Pascal and Scalla,1999; Thom et al., 2002) or GSTFs (Pascal et al.,1998). The other chemical treatments were more se-lective in their enhancement of the GST classes. Inan earlier study it was suggested that while infectionby Erisyphe graminis caused a selective induction ofGSTA1, which has greatest homology to the phi-classTaGSTF6-6, treatment with cadmium and herbicidescaused an accumulation polypeptides recognised by anantiserum raised to maize GSTUs (Mauch and Dudler,1993). The current results suggest that phi-class GSTsin wheat are responsive to xenobiotics. Both TaGSTUsand TaGSTFs were also selectively enhanced follow-ing treatment with phenolic natural products. Theinduction was specific to the phenolic used and wasnot associated with any visible phytotoxic injury tothe plants during the course of the study. In view ofthe high concentrations of treatment solutions used,the physiological relevance of such induction remainsto be seen. However, in view of the potential role ofGSTs as flavonoid-binding proteins (Mueller et al.,2000), it is perhaps significant that GSTs are of-ten enhanced under stress conditions known to causeprotective phenolic compounds to accumulate.

With respect to potential roles in endogenousmetabolism, the TaGSTFs had protective activities to-ward model substrates simulating naturally occurringtoxins of exogenous origin (benzylisothiocyanate) orderived from the oxidation of endogenous unsaturatedcompounds to alkenal derivatives (crotonaldehyde).The TaGSTFs were all highly active as GPOXs, withspecific activities far higher than those reported forthe related maize enzymes (Dixon et al., 1997, 1998).In both maize and wheat, GPOX activity was partic-ularly associated with the ZmGSTF2-like sequences.GSTFs with GPOX activity are known to be involvedin counteracting oxidative stress invoked by environ-mental change (Roxas et al., 1997), or by herbicides(Cummins et al., 1999). It is therefore likely that the

TaGSTFs, especially TaGSTF2 and TaGSTF3, havesimilar antioxidant roles in wheat.

Following up on the link between GSTs andnatural product metabolism, the possibility that theTaGSTFs were binding bioactive metabolites was in-vestigated by enzyme inhibition assays. As demon-strated with a crude GST preparation from oats (Singhand Shaw, 1988) and for recombinant ZmGSTUs(Dixon et al., 1998), the TaGSTF enzymes werevery sensitive to inhibition by the porphyrin chloro-phyllin. The significance of this porphyrin bindingin plants has yet to be determined although it hasbeen suggested that GSTs could be important in por-phyrin metabolism associated with chlorophyll andheme turnover during senescence (Singh and Shaw,1988). In contrast, all of the TaGSTFs were insensitiveto the presence of the synthetic auxin 2,4-D, which ismore a characteristic of GSTUs (Marrs, 1996). TheZmGSTF1-related proteins TaGSTF1-1 and, to a lesserextent, the related TaGSTF6-6 were sensitive to in-hibition by all the flavonoids tested, particularly theflavones and the chalcone. The other TaGSTs werefar more tolerant of these compounds. When the IC50values obtained with TaGSTF1-1 were compared withthe results obtained with ZmGSTF3 from maize andAn9 from petunia determined under similar assay con-ditions (Mueller et al., 2000), very similar levels ofinhibition were determined with the flavonoids. Inpetunia flowers, An9 is required for vacuolar depo-sition of flavonoid pigments and in transient assaysin maize kernels, both An9 and ZmGSTF3 can com-plement the Bz2 mutation (Alfenito et al., 1998). Ithas been proposed that An9 is a ‘ligandin’ bindingprotein, responsible for carrying flavonoids togetherwith glutathione to ATP-binding cassette transportersin the tonoplast membrane for active uptake into thevacuole (Mueller et al., 2000). In support of thishypothesis, it was demonstrated that both An9 andZmGSTF3-3 selectively bound a range of flavonoidswith high affinity (Mueller et al., 2000). Althoughwe have not performed equilibrium dialysis studiesto determine binding stoichiometry, the similarity inenzyme inhibition characteristics for TaGSTF1-1 andAn9 would suggest that this GST could functionas a flavonoid-binding protein in wheat. TaGSTF6-6shared similar inhibition characteristics to TaGSTF1-1with most flavonoids, but was insensitive to inhibi-tion by cyanidin glucoside its status as a ligandin ismore questionable. Interestingly, the homologue ofTaGSTF1-1 in maize, ZmGSTF1-1, was not sensitiveto inhibition by flavonoids (Mueller et al., 2000), iden-

602

tifying a fundamental difference in ‘ligandin’ functionin these related GSTs from these two species.

Under non-oxidative conditions, our studies failedto identify any evidence for the conjugation offlavanones, flavones or anthocyaninidin glucosideswith glutathione in the presence, or absence, ofany of the GST preparations tested. Interestingly,kaempferol, quercetin and luteolin have been re-ported to be S-glutathionylated in the presence ofhydrogen peroxide and horseradish peroxidase afterthe oxidative generation of the respective reactiveo-quinone (luteolin) or quinone methide (flavonol)derivatives (Galati et al., 2000). When isoliquiriti-genin (4,2′,4′-trihydroxychalcone) was incubated withglutathione at pH 7.5 the respective glutathionylatedaddition product was formed without the interven-tion of a GST (Fig. 4). 1H-NMR analysis revealedthat two diastereoisomers of the adduct had formedin equal amounts, consistent with equal attack ofthe thiol both above and below the chalcone dou-ble bond. To our knowledge, this is the first reportof a chalcone natural product forming a glutathioneconjugate, though synthetic chloro-substituted 4′-phenylchalcones have been shown to undergo an anal-ogous reaction (Miyamoto and Yamamoto, 1994).The chlorinated chalcone conjugates were potent com-petitive inhibitors of murine GSTs and it is prob-able that any S-isoliquiritigenin-glutathione formedduring the IC50 assays would contribute to the po-tent inhibition of all the TaGSTFs observed in thepresence of isoliquirtigenin. As such, if chalconeconjugates were formed in planta it would be im-portant to rapidly remove them from the cytoplasmand deposit them in the vacuole, presumably usingthe ATP-binding cassette transporter system knownto operate for the glutathione conjugates of xenobi-otics (Rea et al., 1998). It is also significant that theS-glutathioylation of chalcones appears to be read-ily reversible and has precedence with other relatedaddition reactions of flavonoid derivatives as seen inthe reversible S-glutathionylation of quercetin quinonemethide (Boersma et al., 2000). It is therefore tempt-ing to speculate that S-glutathionylation of naturalproducts is more widespread than currently appreci-ated and that such derivatisations are overlooked dueto the instability of the conjugates during their ex-traction and processing. Certainly, such conjugationreactions would have profound implications on thetransport of selected flavonoids across membranes viaconjugate transporters, forming a useful alternativemechanism to the classical deposition of glucoside and

malonylated-glucoside conjugates in the vacuole (Reaet al., 1998).

Acknowledgements

I.C. and R.E. acknowledge the support of a LINKresearch grant jointly funded by Aventis Crop Sci-ence UK and the Biotechnology and Biological Sci-ences Research Council. The support of Aventis CropsScience to A.H. is also gratefully acknowledged.

References

Alfenito, M.R., Souer, E., Goodman, C.D., Buell, R., Mol, J.,Koes, R. and Walbot, V. 1998. Functional complementation ofanthocyanin sequestration in the vacuole by widely divergentglutathione S-transferases. Plant Cell 10: 1135–1149.

Altschul, S.F., Gish, W., Miller, W., Myers, E.W. and Lippmann,D.J. 1990. Basic local alignment search tool. J. Mol. Biol. 215:403.

Bartling, D., Radzio, R., Steiner, U. and Weiler, E.W. 1993. A glu-tathione S-transferase with glutathione-peroxidase activity fromArabidopsis thaliana. Eur. J. Biochem. 216: 579–586.

Berhane, K., Widersten, M., Engström, A., Kozarich, J.W. andMannervik, B. 1994. Detoxication of base propenals and otherα,â-unsaturated aldehyde products of radical reactions and lipidperoxidation by human glutathione transferases. Proc. Natl.Acad. Sci. USA 91: 1480–1484.

Boersma, M.G., Vervoort, J., Szymusiak, H., Lemanska, K.,Tyrakowska, B., Cenas, N., Segura-Aguilar, J. and Rietjens,I.M.C.M. 2000. Regioselectivity and reversibility of the glu-tathione conjugation of quercetin quinone methide. Chem. Res.Toxicol. 13: 185–191.

Cabello-Hurtado, F., Zimmerlin, A., Rahier, A., Taton, M., DeRose,R., Nedelkina, S., Batard, Y., Durst, F., Pallet, K.E. and Werck-Reichhart, D. 1997. Cloning and functional expression in yeastof a cDNA coding for an obtusifoliol 14a-demethylase (CYP51)in wheat. Biochem. Biophys. Res. Commun. 230: 381–385.

Cummins, I., Cole, D.J. and Edwards, R. 1999. A role for glu-tathione transferases functioning as glutathione peroxidases inresistance to multiple herbicides in black-grass. Plant J. 18:285–292.

Cummins, I., Cole, D.J. and Edwards, R. 1997. Purification ofmultiple glutathione transferases involved in herbicide detoxifi-cation from wheat (Triticum aestivum L.) treated with the safenerfenchlorazole ethyl. Pestic. Biochem. Physiol. 59: 35–49.

Davies, J. and Caseley, J.C. 1999. Herbicide safeners: a review.Pestic. Sci. 55: 1043–1058.

Dixon, D., Cole, D.J. and Edwards, R. 1997. Characterisation ofmultiple glutathione transferases containing the GST I subunitwith activities toward herbicide substrates in maize (Zea mays).Pestic. Sci. 50: 72–82.

Dixon, D.P., Cole, D.J. and Edwards, R. 1998. Purification, regula-tion and cloning of a glutathione transferase (GST) from maizeresembling the auxin-inducible type-III GSTs. Plant Mol. Biol.36: 75–87.

Dixon, D.P., Lapthorne, A. and Edwards, R. 2002. Plant glutathionetransferases. Genome Biol. 3: 3004.1–3004.10.

603

Droog, F.N.J., Hooykaas, P.J.J. and van der Zaal, B.J. 1995.2,4-Dichlorophenoxyacetic acid and related chlorinated com-pounds inhibit two auxin-regulated type-III tobacco glutathioneS-transferases. Plant Physiol. 107: 1139–1146.

Edwards, R. and Dixon, D.P. 2000. The role of glutathionetransferases in herbicide metabolism. In: A.H. Cobb and R.C.Kirkwood (Eds.) Herbicides and their Mechanisms of Action,Sheffield Academic Press, Sheffield, UK, pp. 39–71.

Edwards, R., Dixon, D.P. and Walbot, V. 2000. Plant glutathioneS-transferases: enzymes with multiple functions in sickness andin health. Trends Plant Sci. 5: 193–198.

Galati, G., Moridani, M.Y., Chan, T.S. and O’Brien, P.J. 2000. Per-oxidative metabolism of apigenin and naringenin versus lutolinand quercetin: glutathione oxidation and conjugation. Free Rad.Biol. Med. 30: 370–382.

Graff, G., Anderson, L.A. and Jaques, L.W. 1990. Preparationand purification of soybean lipoxygenase-derived unsaturatedhydroperoxy and hydroxy fatty-acids and determination of mo-lar absorptivities of hydroxy fatty acids. Anal. Biochem. 188:38–47.

Holt, D.C., Lay, V.J., Clarke, E.D., Dinsmore, A., Jepson, I., Bright,S.W.J. and Greenland, A.J. 1995. Characterization of the safener-induced glutathione S-transferase isoform II from maize. Planta196: 295–302.

Irzyk, G.P. and Fuerst, E.P. 1993. Purification and characterizationof a glutathione S-transferase from benoxacor-treated maize (Zeamays). Plant Physiol. 102: 803–810.

Jepson, I., Lay, V.J., Holt, D.C., Bright, S.W.J. and Greenland, A.J.1994. Cloning and characterization of maize herbicide safener-induced cDNAs encoding subunits of glutathione S-transferaseisoforms I, II and IV. Plant Mol. Biol. 26: 1855–1866.

Kampranis, S.C., Damianova, R., Atallah, M., Toby, G., Kondi, G.,Tsichlis, P.N. and Makris, A.M. 2000. A novel plant glutathioneS-transferase/peroxidase suppresses Bax lethality in yeast. J.Biol. Chem. 275: 29207–29216.

Koeppe, M.K., Barefoot, A.C., Cotterman, C.D., Zimmerman,W.T. and Leep, D.C. 1997. Basis of selectivity of the herbicideflupyrsulfuron-methyl in wheat. Pestic. Biochem. Physiol. 59:105–117.

Loyall, L., Uchida, K., Braun, S., Furuya, M. and Frohnmeyer,H. 2000. Glutathione and a UV light induced glutathione S-transferase are involved in signaling to chalcone synthase in cellcultures. Plant Cell 12: 1939–1950.

Marrs, K.A. 1996. The functions and regulation of glutathione S-transferases in plants. Annu. Rev. Plant Physiol. Plant Mol. Biol.47: 127–158.

Mauch, F. and Dudler, R. 1993. Differential induction of dis-tinct glutathione S-transferases of wheat by xenobiotics and bypathogen attack. Plant Physiol. 102: 1193–1201.

McGonigle, B., Keeler, S.J., Lau, S.-M.C., Koeppe, M.K. andO’Keefe, D.P. 2000. A genomics approach to the comprehensiveanalysis of the glutathione S-transferase gene family in soybeanand maize. Plant Physiol. 124: 1105–1120.

Miyamoto, T. and Yamamoto, I. 1994. Glutathione conjugates asthe activated form of chalcones for glutathione S-transferaseinhibition. J. Pestic. Sci. 19: 53–58.

Mozer, T.J., Tiemeier, D.C. and Jaworski, E.G. 1983. Purifi-cation and characterisation of corn glutathione S-transferase.Biochemistry 22: 1068–1072.

Mueller, L.A., Goodman, C.D., Silady, R.A. and Walbot, V.2000. AN9, a petunia glutathione S-transferase required for

anthocyanin sequestration, is a flavonoid-binding protein. PlantPhysiol. 123: 1561–1570.

Neuefeind, T., Huber, R., Reinemer, P., Knäblein, J., Prade, L.,Mann, K., and Bieseler, B. 1997.Cloning, sequencing, crystal-lization and X-ray structure of glutathione S-transferase-III fromZea mays var. mutin: a leading enzyme in detoxification of maizeherbicides. J. Mol. Biol. 274: 577–587.

Neuefeind, T., Reinemer, P. and Bieseler, B. 1997. Plant glutathioneS-transferases and herbicide detoxification. Biol. Chem. 378:199–205.

Nishihira, J., Sakai, M., Nishi, S. and Hatanaka, Y. 1995. Identi-fication of the electrophilic substrate-binding site of glutathioneS-transferase P by photoaffinity labeling. Eur. J. Biochem. 232:106–110.

Pascal, S. and Scalla, R. 1999. Purification and characterization of asafener-induced glutathione S-transferase from wheat (Triticumaestivum). Physiol. Plant. 106: 17–27.

Pascal, S., Debrauwer, L., Ferte, M.P., Anglade, P., Rouimi, P.and Scalla, R. 1998. Analysis and characterization of glutathioneS-transferase subunits from wheat (Triticum aestivum L.). PlantSci. 134: 217–226.

Pflugmacher, S., Schröder, P. and Sandermann, H. 2000. Taxo-nomic distribution of plant glutathione S-transferases acting onxenobiotics. Phytochemistry 54: 267–273.

Ploemen, J.H.T.M., van Ommen, B., Bogaards, J.J.P. and vanBladeren, P.J. 1993. Ethacrynic acid and its glutathione conju-gate as inhibitors of glutathione S-transferases. Xenobiotica 23:913–923.

Rea, P.A., Li, Z.S., Lu, Y.P., Drozdowicz, Y.M. and Martinoia,E. 1998. From vacuolar GS-X pumps to multispecific ABCtransporters. Annu. Rev. Plant Physiol. Plant Mol. Biol. 49:727–760.

Riechers, D.E., Irzyk, G.P., Jones, S.S. and Fuerst, E.P. 1997a.Partial characterization of glutathione S-transferases from wheat(Triticum spp.) and purification of a safener-induced glutathioneS-transferase from Triticum tauschii. Plant Physiol. 114: 1461–1470.

Riechers, D.E., Irzyk, G.P., Fuerst, E.P. and Jones, S.S. 1997b.Nucleotide sequence of a cDNA encoding a safener-induced glu-tathione S-transferase (accession No. AF004358) from Triticumtauschii. Plant Physiol. 114: 1568.

Roxas, V.P., Smith, R.K., Allen, E.R. and Allen, R.D. 1997.Overexpression of glutathione S-transferase/glutathione peroxi-dase enhances the growth of transgenic tobacco seedlings duringstress. Nature Biotechnol. 15: 988–991.

Simons, P. and van der Jagt, D.L. 1981. Purification of glu-tathione S-transferases by glutathione-affinity chromatography.Meth. Enzymol. 77: 235–237.

Singh, B.R. and Shaw, R.W. 1988. Selective inhibition of oat glu-tathione S-transferase activity by tetrapyrroles. FEBS Lett. 234:379–392.

Thom, R., Cummins, I., Dixon, D.P., Edwards, R., Cole, D.J.and Lapthorn, A.J. 2002. Structure of a tau class glutathioneS-transferase from wheat active in herbicide detoxification. Bio-chemistry 41: 7008–7020.

Wagner, U., Edwards, R., Dixon, D.P. and Mauch, F. 2002. Probingthe diversity of the Arabidopsis glutathione S-transferase genefamily. Plant Mol. Biol. 49: 515–532.