Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
trends in plant science Reviews Plant glutathione S -transferases: enzymes with multiple functions in sickness and in health Robert Edwards, David P. Dixon and Virginia Walbot Glutathione S -transferases (GSTs) are abundant proteins encoded by a highly divergent, ancient gene family. Soluble GSTs form dimers, each subunit of which contains active sites that bind glutathione and hydrophobic ligands. Plant GSTs attach glutathione to electrophilic xenobiotics, which tags them for vacuolar sequestration. The role of GSTs in metabolism is unclear, although their complex regulation by environmental stimuli implies that they have important protective functions. Recent studies show that GSTs catalyse glutathione-dependent isomerizations and the reduction of toxic organic hydroperoxides. GSTs might also have non-catalytic roles as carriers for phytochemicals. C ellular survival requires the management of reactive oxygen species, endogenous phytochemicals and exogenous toxins, including diverse xenobiotics added to the environment by humans. Furthermore, many cellular processes function optimally only within a narrow redox range. Glutathione (GSH), the tripeptide g-Glu–Cys–Gly, plays a central role in the crucial processes of detoxification and redox buffering1. GSH is abundant in plants, typically exceeding 1 mM in the cytoplasm1. Plants also contain several GSH-dependent detoxifying enzymes2, most notably glutathione S-transferases (GSTs), which collectively constitute .1% of the soluble protein in maize leaves3. Individual gene analysis and genomics projects indicate that plants have 25 or more genes coding for GSTs, and that the proteins share as little as 10% amino acid identity. In both crops and weeds, GSTs detoxify electrophilic herbicides by catalysing their conjugation with GSH (Ref. 4; Fig. 1). The Sglutathionylated metabolites are tagged for vacuolar import by ATP binding cassette (ABC) transporters, which selectively transport GSH conjugates5. GSTs are composed of two subunits with molecular masses in the range of 25–27 kDa and are either homodimers of a single gene product or heterodimers of subunits encoded by different genes. Individual GST isoenzymes can selectively detoxify specific xenobiotics, with species differences in GST specificity and capacity determining herbicide selectivity4. Paradoxically, in spite of the specific and crucial roles of individual GST enzymes in xenobiotic metabolism, their roles in endogenous plant metabolism have remained an enigma. There are only a few reports of natural products [such as the phytoalexin medicarpin (Fig. 1)] being conjugated with GSH (Ref. 6) in spite of a burgeoning literature showing that the expression of specific GSTs vary markedly during plant development (cell division, senescence) and after exposure to pathogens, changes in environmental conditions and chemical treatments3. GSTs with high affinity for auxins3 and cytokinin7 have been suggested to contribute to hormone homeostasis. Indeed, the enhancement of GST expression has become a marker for plant response to stress, although the functional significance of selective GST expression is only just emerging. One role of this article is to introduce the diversity of GSTs, highlighting novel features of the plant GST family compared with other eukaryotes. The complexity and antiquity of GST genes and their functioning as homo- and heterodimeric proteins present challenges in both classification and nomenclature. However, the core of the article is a consideration of the functions of GSTs in endogenous plant metabolism. In particular that GSTs: (1) (2) (3) Catalyse conjugation reactions with natural products similar to those observed with xenobiotics. Function as binding and carrier proteins for phytochemicals between cellular compartments. Catalyse alternative GSH-dependent biotransformation reactions. GST-gene diversity: who counts as a family member? GSTs are an ancient and diverse protein family, existing as multigene families in bacteria, fungi, animals and plants (Fig. 2). The ‘classic’ definition of a GST requires the catalytic formation of GSH conjugates with xenobiotic substrates. The compound 1chloro-2,4-dinitrobenzene (CDNB) is used as a model substrate (Fig. 1) but some GSTs show little activity with this compound4. Furthermore, highly electrophilic substrates spontaneously conjugate with GSH under physiological conditions, with GSTs accelerating the reaction only modestly8. Whatever their substrate, GSTs should selectively bind GSH or natural homologues, such as homoglutathione (g-Glu–Cys–b-Ala) and hydroxymethyl glutathione (g-Glu–Cys–Ser), which are commonly found in legumes and grasses, respectively9. Affinity chromatography using GSH as a ligand is widely used to extract GSTs, although resins with GSH conjugates can be useful in the recovery of specific GSTs from cellular extracts10. Members of the GST family show the greatest sequence similarity in their GSH-binding domain; four highly conserved amino acids in this domain are a ‘signature motif’ for GSTs, but this motif is not sufficient to identify a GST. By applying the dual definition of GSH binding and catalysis of GSH conjugation with electrophiles, we can distinguish GSTs from proteins with related sequences that have evolved alternative functions. Proteins related to the GST superfamily, such as the crystallin protein in the squid lens11 and elongation factor 1-gamma proteins (Ref. 12), have lost their capacity either to bind GSH or to use the thiol in conjugation reactions. Some offshoots of the GST family currently have no assignable functions. For example, several stress-induced proteins in bacteria resemble GSTs (Ref. 13), as does the herbicide safener inducible 27-kDa In2-1 protein from maize14. Classification of plant GSTs based on sequence An accurate GST classification scheme would require knowledge of the proteins’ in vivo role(s) to name individual GST enzymes 1360 - 1385/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved. PII: S1360-1385(00)01601-0 May 2000, Vol. 5, No. 5 193 trends in plant science Reviews genes have three exons. The other large group, type III, consisted mainly of auxinOH SG OH induced GSTs, with the genes containing HO HO OH two exons. Types I and III GSTs show OH .50% sequence divergence and have now been placed in separate classes15,16. Type-II + GSH GSTs have ten exons and are much closer O O to the mammalian zeta GSTs (Ref. 2). Isoliquirtigenin Recently, a Type-IV grouping was proposed for several Arabidopsis genes that are simi(b) lar to classical mammalian theta enzymes2. HO O HO O During 1999, dozens of plant GST sequences were reported from diverse angiosperm sequencing projects (e.g. GSH maize, tomato, soybean and cotton), with SG O .25 genes for GSTs identified to date in OCH3 OCH3 HO Medicarpin the Arabidopsis genome alone. This diversity makes the catch-all theta classification (c) inappropriate. Some plant GSTs clearly GS group with specific mammalian types but R R O O + GSH there are two distinct plant-specific types2 (Fig. 2). Because the principle of GreekOH OH letter designations is widely used for nonR = H − cinnamic acid plant GSTs (Ref. 15), we suggest that a R = OH − coumaric acid new nomenclature system be adopted for (d) plant GST genes (Box 1; Table 1). ThereNO2 NO2 fore, the classification of plant GSTs should be amended to include the followO2N O2N Cl + GSH SG + HCl ing new classes: • Phi (F) – a plant-specific class replacing 1-Chloro-2,4-dinitrobenzene (CDNB) Type I. • Zeta (Z) – replacing Type II. (e) • Tau (U) – a plant-specific class replacing O O SG Type III. Cl • Theta (T) – replacing Type IV. This nomenclature is applicable to SyneN N + GSH + HCl chocystis spp., whose genome contains O O genes for four GST-like proteins without the intron–exon gene structure of higher Alachlor plants. At the predicted-amino-acid level, Trends in Plant Science the gene designated gst1 (SwissProt P74665) is a zeta GST; ORF sll1902 Fig. 1. Glutathione conjugation reactions. Glutathione (GSH) forms conjugates with the chal(P73835) shows the greatest similarity to cone isoliquirtigenin (a) and the pterocarpan phytoalexin medicarpin (b). These addition reacthe tau-class gene, maize In2-1 (P49248). tions occur spontaneously under basic conditions and can be readily reversed by reducing the The gst gene (Q55139) is a phi-class mempH. (c) The phenylpropanoids cinnamic acid and coumaric acid undergo addition reactions with GSH to form the respective conjugates. This reaction is radical driven and is catalysed ber, and ORF slr0236 (P72690) has similarby an ascorbate peroxidase. (d and e) The classic detoxification reactions catalysed by GSTs ity to the tau GSTs. We can conclude that with the model substrate 1-chloro-2,4-dinitrobenzene and the herbicide alachlor, respectively. the two plant-specific classes, tau and phi, In both cases, the glutathione substitution products are stable and the reaction irreversible. are characteristic of both simple and complex photosynthetic organisms and might be required to accommodate the metabolic properly; current schemes use amino acid sequence similarity and consequence of active oxygen generated by photosynthesis. gene organization. Although this understates the functional hetAnother interesting feature of the GST family is the genes’ erogeneity of this family, we can ask whether gene similarity fore- arrangement in repeating units on plant chromosomes, which casts functional capabilities. Using a classification system based strongly suggests that there have been numerous gene duplion immunological cross-reactivity and sequence relatedness, sol- cations. Chromosome II of Arabidopsis contains seven adjacent uble mammalian GSTs have been divided into the alpha, mu, pi, tau-class genes, and all three of the known theta-class genes in the sigma, theta and zeta classes2,15 (Fig. 2). Most nonmammalian Arabidopsis genome occur as a cluster and the two known GSTs were lumped into the highly heterogeneous theta class, zeta-class genes are arranged in tandem. which was viewed as closest to the original progenitor of all eukaryotic GSTs. However, the view that plant GSTs are a primi- Structure of plant GSTs tive subgroup is far from the truth. X-ray crystallography has shown a similar three-dimensional toThree distinct types of plant GSTs were recognized initially3. pology for three phi-class GSTs from Arabidopsis17 and maize18,19, Type I included GSTs with herbicide-detoxifying activity; these which share only 20% sequence identity. Each subunit has an active (a) 194 May 2000, Vol. 5, No. 5 trends in plant science Reviews GST detoxification Box 1. Enzyme and gene nomenclature Sigma Alpha Isomerase MAAI Pi Zeta Mu Theta GPOX detoxification Tau Phi Ligandin, isomerase, GST detoxification GPOX detoxification Hormone responsive Ligandin GST detoxification GPOX detoxification Trends in Plant Science Fig. 2. A deduced glutathione S-transferase (GST) phylogenetic tree showing the relationships between the major GST classes and the extent of sequence diversity within each class. The relationships between the classes are adapted from Ref. 2. For each class, the triangle height is proportional to the number of sequences in the SwissProt database, and the triangle width is proportional to the sequence divergence (100% minus the average % identity between each pair of amino acid sequences). The lengths of the connecting bars represent the approximate extent of divergence between the classes. To illustrate the scale, the tau class shows 59% divergence (41% average identity) and includes 44 sequences. Established functions for family members are listed next to each triangle. The theta class sequences have been limited to those closely related to the mammalian theta sequences to avoid the inclusion of highly divergent sequences placed into this class by default. site that includes a conserved GSH-binding site (G-site) located in the N-terminal domain and a C-terminal cosubstrate-binding domain (H-site), which will accommodate a diverse range of hydrophobic compounds. The G-site is specific for GSH and facilitates the formation of the catalytically active thiolate anion of GSH. In all three phi GSTs, the active sites are situated on either side of a large, open cleft formed between the subunits, allowing access to large planar and spherical molecules (Fig. 3). However, each active site interacts minimally with the adjacent subunit17–19. Crystallography has also shown that certain GSTs can bind an additional GSH molecule adjacent to the active site17, although the functional significance of this secondary binding site has received little attention. In addition to the diversity of GST genes in each species, heterodimer formation could give rise to up to (1 1 2 1 3 1…1 n) distinct heterodimers, where n is the number of GST genes of a particular class. The importance of GST dimerization in xenobiotic metabolism was shown by determining the cause of sensitivity to the herbicide alachlor in an inbred maize line20: ZmGST II (Table 1) subunits were unable to dimerize with one another to form the ZmGST II–II homodimer, which is highly active in conjugating and detoxifying alachlor (Fig. 1). Studies with coexpressed GST subunits in recombinant bacteria suggest that dimerization is spontaneous but restricted to subunits of the same class. In the case of the maize tau GSTs ZmGST V and ZmGST VI, heterodimerization was as likely as homodimerization21, We propose a new gene and enzyme nomenclature for glutathione Stransferases (GSTs) to simplify future assignments and to align designations more closely to evolutionary history. The current naming of GSTs is confusing because enzymes are named in their order of purification and genes by their order of sequencing. This produces situations in which enzyme I is encoded by gene 3. Heterodimeric proteins are given a specific name, which further obscures which genes encode the subunits. We propose that plant genes be named using a species designation, a gene-class identifier, and a number within that class. For example, XyGstZ1 indicates that the gene was isolated from species Xy (such as At for Arabidopsis thaliana, Zm for Zea mays), Z denotes zeta class and the numeral 1 indicates that it is the first gene of the zeta class from that species. At the protein level, XyGSTZ1-1 would indicate a homodimer encoded by XyGstZ1; XyGSTZ1-2 would indicate a heterodimer encoded by the XyGstZ1 and XyGstZ2 genes. For example, the maize enzyme GST V–VI (Ref. 21) would become ZmGSTU1-2. This system is compatible with the current practice for animal GSTs as well as with the plant-genetics conventions of using italics for genes and standard type for proteins. Table 1 shows the application of this nomenclature to maize and Arabidopsis GSTs. Table 1. Suggested new nomenclature applied to maize and Arabidopsis GSTs Species Zea mays Zea mays Zea mays Zea mays Zea mays Zea mays Zea mays Arabidopsis Arabidopsis Arabidopsis Arabidopsis Arabidopsis Arabidopsis Arabidopsis Arabidopsis Arabidopsis Arabidopsis Old name Proposed name ZmGST I ZmGST II ZmGST III ZmGST V ZmGST VI ZmGST VII Bronze2 AtGST10 GST1, PM239x14 GST2, PM24.1 GST3, ERD11 GST4, ERD13 GST6 GST7 GST8 GST11 GST5, 103-1a ZmGSTF1 ZmGSTF2 ZmGSTF3 ZmGSTU1 ZmGSTU2 ZmGSTU3 ZmGSTU4 AtGSTT1 AtGSTF1 AtGSTF2 AtGSTF3 AtGSTF4 AtGSTF5 AtGSTF6 AtGSTF7 AtGSTF8 AtGSTU1 whereas heterodimer formation was favoured with the maize phi GSTs ZmGST I and ZmGST II (Ref. 22). Do heterodimers perform unique roles? The available kinetic measurements, performed with xenobiotic substrates in vitro, have detected only ‘additive’ properties in heterodimers21,23. We hypothesize that, in vivo, heterodimers might uniquely recognize specific phytochemicals or exhibit unique kinetic properties essential for metabolite management. Alternatively, heterodimer formation might determine the subcellular localization or control the rates of protein turnover. GSTs and endogenous metabolism: vexing problems identifying in vivo substrates GSTs evolved long before plants were challenged to detoxify pesticides and synthetic pollutants, therefore what are the physiological roles of GSTs that were shaped by natural selection? Ironically, the May 2000, Vol. 5, No. 5 195 trends in plant science Reviews detoxified by phi GSTs from sorghum23, and tau GSTs from wheat have similar activities10. Pathogen-inducible GSTs probably also detoxify exogenous natural products, though of exogenous origin, such as phytotoxins produced by competing plants or by invading microbial pathogens. A good example of this is the detoxification of isothiocyanates, reactive compounds released from glucosinolate precursors formed in Brassica species. These allelochemicals can suppress the growth of competing plants but are excellent substrates for tau and phi GSTs in wheat and maize2,10. Although inducible GSTs might use stress metabolites as substrates, their role Fig. 3. Two views of a space-filled model of the crystal structure of the Arabidopsis phi in GSH conjugation of natural products in class glutathione S-transferase (GST) dimer AtGSTF2-2. (a) A side-on view to illustrate the healthy plants is less clear. Until recently, large active-site cleft. (b) A view into the active sites. For both views, both subunits are visthe best candidates for endogenous phytoible; the lower subunit is shaded from red at the N-terminus to yellow at the C-terminus and chemicals detoxified by GSTs were the the upper subunit is shaded from blue at the N-terminus to green at the C-terminus. Each of phenylpropanoids cinnamic acid and the two active sites is occupied by a molecule of S-hexylglutathione (white). The glucoumaric acid (Fig. 1). However, recent tathione moiety is close to the dimer interface and interacts exclusively with the N-terminal analysis has demonstrated that these conjudomain of the GST, whereas the S-hexyl moiety lies further from the dimer interface and interacts with both N-terminal and C-terminal domains of the protein. gates are formed by a radical-mediated process catalysed by an ascorbate peroxidase rather than by a GST (Ref. 28). GSTs of herbivorous animals have long been accorded the role of Anthocyanins were also proposed as likely substrates for conjugametabolizing dietary toxins, most of which are plant natural prod- tion because a GST is required for the last cytoplasmic step of ucts. Interestingly, GSTs are frequently highly abundant in crops, a biosynthesis in both maize (Bz2 gene) and petunia29 (An9 gene). In characteristic exploited in selective herbicides, where the GSTs the absence of the GST-mediated step, anthocyanin accumulates which detoxify them can be 20-fold more abundant in the crop than in the cytoplasm, suggesting that conjugate formation was a prein the competing weeds24. In an evolutionary context, the evidence requisite for vacuolar sequestration. However, to date, conjugates available in wheat and its relatives suggests that enhanced GST of non-electrophilic flavonoids have not been detected by expression is not the result of modern breeding but is instead char- high-pressure liquid chromatography (HPLC) of plant extracts acteristic of all the progenitor Triticum species used in its domesti- or enzymatic reactions30. Both the AN9 and BZ2 GSTs are cation25,26. It is possible that, because they reduce stress, a high GST flavonoid-binding proteins (L.A. Mueller and V. Walbot, uncapacity is a prerequisite for successful artificial selection for high published), and the current model proposes that these gene prodyield under diverse and adverse conditions. ucts prevent anthocyanin oxidation and cross-linking in the Do the four distinct classes of plant GSTs recognize different cytoplasm, and hence indirectly aid vacuolar sequestration. substrates? As discussed below, emerging evidence suggests that While screening other flavonoids for their ability to undergo zeta and theta GSTs in plants and animals have conserved GSH- glutathione conjugation, we have recently identified the beta-Sdependent activities that have little to do with conventional conju- glutathionyl derivative of the chalcone isoliquirtigenin (4,29, gating activity. By contrast, when assayed for GSH-conjugating 49-trihydroxychalcone) as being formed in vitro (Fig. 1). Signifiactivity towards xenobiotics, the phi and tau GSTs of maize2 and cantly, this compound was formed at physiological pH in wheat10 have a surprisingly similar spectrum of activities. If the the absence of plant proteins and would readily deconjugate to phi and tau GSTs have distinct functions, these are likely to be isoliquirtigenin and GSH under mildly acidic conditions. The evident only with endogenous substrates. GSH conjugate of the pterocarpan medicarpin can also breakdown to re-release the parent phytoalexin at acidic pH (Fig. 1). GSTs conjugating natural products This conjugate, but not the parent phytoalexin, is readily transAfter glutathionation, xenobiotic conjugates are rapidly imported ported into resealed minivacuoles by an ABC transporter in vitro6. into the vacuole before being further metabolized into a range of Intriguingly, these results suggest that GSH conjugation of elecknown sulphur-containing metabolites4. By contrast, the GSH con- trophilic natural plant products could be more widespread than jugate of caftaric acid identified in wine must and a sulphur- previously thought; a role in vacuolar targeting might have been containing metabolite of gibberellic acid, termed gibberthione, are missed because the conjugates are degraded in the acidic condiamong the few candidates for GSH conjugation of endogenous tions of most vacuoles or in the course of phytochemical analysis. plant products27. This suggests either that GSH conjugation is not an important route of detoxification for endogenous metabolites or GSTs functioning as binding proteins that the elusive conjugates are unstable and have escaped detection. A hypothesis to explain why a specific GST is needed for anthoStress-inducible GSTs might act to conjugate metabolites aris- cyanin sequestration is that these proteins act as carriers. The coning from oxidative damage. For example, cytotoxic alkenals cept of GSTs as carrier proteins or ‘ligandins’ was first proposed derived from the peroxidation of natural products, such as fatty in the early 1970s, based on the finding that a similar protein was acids, are well-defined substrates of GSTs in oxidatively stressed identified as the cellular binding factor for diverse steroids, drugs animal cells2,3. In plants, 4-hydroxynonenal, a toxic alkenal and mammalian metabolites3. Later, cloning and identification of released following oxidative damage of membranes, is actively ‘ligandin’ as a GST resulted in the same quandary currently faced 196 May 2000, Vol. 5, No. 5 trends in plant science Reviews by those studying plant GSTs: an enzyme in search of substrates for conjugation. The fact that GSTs have been identified while searching for auxin- and cytokinin-binding proteins when using photoaffinity labels3,7 suggests that these proteins might have functions in the binding and transport of plant hormones. Similarly, both phi and tau GSTs have high affinities for tetrapyrroles and porphyrin metabolites21,27. With both tetrapyrroles and plant hormones, binding inhibits GST activity toward xenobiotics, but the inhibitory ligands do not undergo conjugation. Similarly, the CDNB conjugating activity of petunia AN9, a phi GST, is inhibited by flavonols, flavones and anthocyanins29. There is ample evidence that specific plant GSTs that bind defined plant metabolites fit the definition of ligandins. GSTs catalysing GSH-dependent peroxidase and isomerase reactions By catalysing the nucleophilic attack of GSH on hydroperoxides, GSTs can catalyse the reduction of organic hydroperoxides to the less-toxic monohydroxy alcohols, the resulting sulfenic acid derivative of GSH then spontaneously forming a disulfide with another GSH molecule (Fig. 4). Tobacco seedlings overexpressing a tobacco tau GST with a high glutathione peroxidase activity are more tolerant of chilling and osmotic dehydration than wildtype plants31. A further link between GSTs functioning as glutathione peroxidases and oxidative-stress tolerance was discovered in black grass32 (Alopecurus myosuroides). Herbicide-resistant weeds that are cross-resistant to multiple classes of herbicides express a phi GST (AmGSTF1) that is a highly active glutathione peroxidase; this GST is barely detectable in herbicide-sensitive black grass. AmGSTF1 appears to contribute to herbicide resistance by preventing the accumulation of cytotoxic hydroperoxides of natural products, which can be formed either directly or indirectly as a result of injury by herbicides with diverse modes of action. Significantly, AmGSTF1 is related to cereal phi GSTs that are induced in crops following treatment with herbicide safeners, providing a connection between safener-mediated protection of crops and oxidative stress tolerance32. GSTs are also essential for the isomerization of specific metabolites (Fig. 4). The proposed mechanism involves the transient formation of a GSH adduct, spontaneous isomerization of the compound and finally the release of the isomer and GSH. An excellent example is the human zeta GST, which functions as a maleylacetoacetate isomerase (MAAI), a key enzyme in phenylalanine catabolism. MAAI catalyses the cis–trans isomerization of maleylacetoacetate to fumarylacetoacetate (Fig. 4). This reaction has been known for some time to be GSH dependent and associated with a GST; the recent complementation of an MAAIdeficient strain of Aspergillus nidulans with the human zeta GST confirming its identity as the MAAI enzyme33. This GSTmediated cis–trans isomerization reaction involves the reversible addition of GSH to the cis double bond; after rotation, GSH is eliminated and the trans isomer is formed. In view of their sequence similarity, plant zeta GSTs probably have a similar activity. The carnation (Dianthus caryophyllus) zeta genes are induced during senescence2,3, which is consistent with a role in the degradation of aromatic amino acids. GST-mediated isomerase reactions recently identified in animals include prostaglandin-H E-isomerase activity34 and the isomerization of 13-cis retinoic acid to all-trans retinoic acid35. In the latter case, the enzyme catalysing the reaction is clearly a GST, yet the reaction does not require GSH and these enzymes appear to have recruited other proteinaceous thiols to catalyse the reaction. In plants, the isomerase activity of GSTs has been demonstrated (a) R− O − OH + GSH R−OH + GS − OH GSH GS–SG O– (b) O– O O– O O– OH O O Maleylacetoacetate OH O O Fumarylacetoacetate (c) N N N N N S O S– Br GSH N SG Br O S N N GSH N Br O Trends in Plant Science Fig. 4. Examples of glutathione-S-transferase (GST)-mediated and glutathione (GSH)-dependent reactions that do not result in the formation of GSH conjugates as end products. (a) GST functioning as a glutathione peroxidase, using GSH to reduce an organic hydroperoxide (ROOH) to the monohydroxy alcohol (ROH). (b) A zeta GST catalyzing a cis–trans isomerization of maleylacetoacetate to fumarylacetoacetate. (c) GST catalyzing the isomerization of a thiadiazolidine proherbicide to the phytotoxic triazolidine, showing the proposed reaction mechanism and the formation of the GSH-conjugated intermediate. using thiadiazolidine herbicides; these are bioactivated by GSTmediated isomerization to triazolidines, which are potent inhibitors of protoporphyrinogen oxidase36. A likely mechanism for isomerization involves nucleophilic attack at the carbonyl group by GSH, a process that is activated by the GST (Fig. 4). Ring opening then allows rotation around the N–C bond, with the C5N double bond transferred to the C–S group; the nitrogen atom attacks the carbonyl group to reform the ring and eliminate the glutathione36. Given the flexible architecture of the active site of plant and mammalian GSTs, it now appears likely that other isomerase activities carried out by undefined enzymes might be catalysed by GSTs. Conclusions Recent advances in molecular genetics and biochemistry have unveiled a surprising complexity in the variety of GST genes and enzymes in plants. We have tried to highlight some tantalizing reasons for this diversity: members of an ancient gene family have been recruited to perform diverse physiological functions based around their ability to activate a readily available cellular thiol while coordinately binding a diverse range of hydrophobic ligands. GST functions have direct cytoprotective activities and they might May 2000, Vol. 5, No. 5 197 trends in plant science Reviews be essential to preserve plants during environmental stress and disease, as well as supporting normal development. We consider the frontier of research to involve matching individual GSTs with their in vivo ligands and showing which interactions result in catalysis. Because of the extreme divergence among GSTs, sequence analysis alone cannot uncover function. Perhaps even more intriguing are the hints that some GSTs might be carrier proteins for reactive plant metabolites, controlling their activity and cellular distribution. In the coming years, our ability to manipulate the expression of GSTs in plants through gene disruption and transgenic approaches will allow us to test these models of GST function in vivo. Acknowledgements Support for our research on plant GSTs is supported by grants from the US National Science Foundation to V.W. (IBN 9603927) and from the Biotechnology and Biological Sciences Research Council, Zeneca Agrochemicals and Aventis Crop Science to R.E. D.D. was supported by Aventis Crop Science. References 1 Noctor, G. and Foyer, C.H. (1998) Ascorbate and glutathione: keeping active oxygen under control. Annu. Rev. Plant Phys. Plant Mol. Biol. 49, 249–279 2 Dixon, D.P. et al. (1998) Glutathione-mediated detoxification systems in plants. Curr. Opin. Plant Biol. 1, 258–266 3 Marrs, K.A. (1996) The functions and regulation of glutathione S-transferases in plants. Annu. Rev. Plant Phys. Plant Mol. Biol. 47, 127–158 4 Cole, D.J. and Edwards, R. (2000) Secondary metabolism of agrochemicals in plants. In Metabolism of Agrochemicals in Plants (Roberts, T.R., ed.), pp. 107–154, John Wiley & Sons 5 Rea, P.A. (1999) MRP subfamily ABC transporters from plants and yeast. J. Exp. Bot. 50, 895–913 6 Li, Z-S. et al. (1997) Vacuolar uptake of the phytoalexin medicarpin by the glutathione conjugate pump. Phytochemistry 45, 689–693 7 Gonneau, J. et al. (1998) A Nicotiana plumbaginifolia protein labeled with an azido cytokinin agonist is a glutathione S-transferase. Physiol. Plant. 103, 114–124 8 Coleman, J.O.D. et al. (1997) Detoxification of xenobiotics by plants: chemical modification and vacuolar compartmentation. Trends Plant Sci. 2, 144–151 9 Noctor, G. et al. (1998) Glutathione: biosynthesis, metabolism and relationship to stress tolerance explored in transformed plants. J. Exp. Bot. 49, 623–647 10 Cummins, I. et al. (1997) Purification of multiple glutathione transferases involved in herbicide detoxification from wheat (Triticum aestivum L.) treated with the safener fenchlorazole-ethyl. Pestic. Biochem. Physiol. 59, 35–49 11 Tomarev, S.I. and Piatigorsky, J. (1996) Lens crystallins of invertebrates: diversity and recruitment from detoxification enzymes and novel proteins. Eur. J. Biochem. 235, 449–465 12 Koonin, E.V. et al. (1994) Eukaryotic translation elongation factor 1-gamma contains a glutathione transferase domain – study of a diverse, ancient protein superfamily using motif search and structural modeling. Protein Sci. 11, 2045–2054 13 Serizawa, H. and Fukuda, R. (1987) Structure of the gene for the stringent starvation protein of Escherichia coli. Nucleic Acids Res. 15, 1153–1163 14 Hershey, H.P. and Stoner, T.D. (1991) Isolation and characterization of cDNA clones for RNA species induced by substituted benzenesulfonamides in corn. Plant Mol. Biol. 17, 679–690 15 Hayes, J.D. and McLellan, L.I. (1999) Glutathione and glutathione-dependent enzymes represent a co-ordinately regulated defence against oxidative stress. Free Radic. Res. 31, 273–300 16 Droog, F. (1997) Plant glutathione S-transferases, a tale of theta and tau. J. Plant Growth Regul. 16, 95–107 17 Reinemer, P. et al. (1996) Three-dimensional structure of glutathione S-transferase from Arabidopsis thaliana at 2.2 Å resolution: structural characterization of herbicide-conjugating plant glutathione S-transferases and a novel active site architecture. J. Mol. Biol. 255, 289–309 198 May 2000, Vol. 5, No. 5 18 Neuefeind, T. et al. (1997) Crystal structure of herbicide-detoxifying maize glutathione S-transferase-I in complex with lactoylglutathione: evidence for an induced-fit mechanism. J. Mol. Biol. 274, 446–453 19 Neuefeind, T. et al. (1997) Cloning, sequencing, crystallization and X-ray structure of glutathione S-transferase III from Zea mays var. mutin: a leading enzyme in detoxification of maize herbicides. J. Mol. Biol. 274, 577–587 20 Rossini, L. et al. (1996) Characterization of glutathione S-transferase isoforms in three maize inbred lines exhibiting differential sensitivity to alachlor. Plant Physiol. 112, 1595–1600 21 Dixon, D.P. et al. (1999) Dimerization of maize glutathione transferases in recombinant bacteria. Plant Mol. Biol. 40, 997–1008 22 Sommer, A. and Böger, P. (1999) Characterization of recombinant corn glutathione S-transferase isoforms I, II, III and IV. Pestic. Biochem. Physiol. 63, 127–138 23 Gronwald, J.W. and Plaisance, K.L. (1998) Isolation and characterization of glutathione S-transferase isozymes from sorghum. Plant Physiol. 117, 877–892 24 Hatton, P.J. et al. (1999) Glutathione transferases involved in herbicide detoxification in the leaves of Setaria faberi (giant foxtail). Physiol. Plant. 105, 9–16 25 Edwards, R. and Cole, D.J. (1996) Glutathione transferases in wheat (Triticum) species with activity towards fenoxaprop-ethyl and other herbicides. Pestic. Biochem. Physiol. 54, 96–104 26 Riechers, D.E. et al. (1996) Variability of glutathione S-transferase levels and dimethenamid tolerance in safener-treated wheat and wheat relatives. Pestic. Biochem. Physiol. 56, 88–101 27 Lamoureux, G.L. and Rusness, D.G. (1993) Glutathione in the metabolism and detoxification of xenobiotics in plants. In Sulfur Nutrition and Assimilation in Higher Plants (De Kok, L.J. et al., eds), pp. 221–237, SPB Academic Publishing 28 Dean, J.V. and Devarenne, T.P. (1997) Peroxidase-mediated conjugation of glutathione to unsaturated phenylpropanoids: evidence against glutathione S-transferase involvement. Physiol. Plant. 99, 271–278 29 Alfenito, M.R. et al. (1998) Functional complementation of anthocyanin sequestration in the vacuole by widely divergent glutathione S-transferases. Plant Cell 10, 1135–1149 30 Walbot, V. et al. Do glutathione S-transferases acts as enzymes or carrier proteins for their natural substrates. In Sulphur Metabolism in Plants (Brunold, C. et al., eds), Verlag Paul Haupt, Berr, Switzerland (in press) 31 Roxas, V.P. et al. (1997) Overexpression of glutathione Stransferase/glutathione peroxidase enhances the growth of transgenic tobacco seedlings during stress. Nat. Biotechnol. 15, 988–991 32 Cummins, I. et al. (1999) A role for glutathione transferases functioning as glutathione peroxidases in resistance to multiple herbicides in black-grass. Plant J. 18, 285–292 33 Fernández-Cañón, J.M. and Peñalva, M.A. (1998) Characterization of a fungal maleylacetoacetate isomerase gene and identification of its human homologue. J. Biol. Chem. 273, 329–337 34 Meyer, D.J. et al. (1996) Purification and characterization of prostaglandin-H E-isomerase, a sigma-class glutathione S-transferase, from Ascaridia galli. Biochem. J. 13, 223–227 35 Chen, H. and Juchau, M.R. (1998) Recombinant human glutathione S-transferases catalyse enzymic isomerization of 13-cis retinoic acid to alltrans retinoic acid in vitro. Biochem. J. 336, 223–226 36 Jablonkai, I. et al. (1997) Chemical catalysis of the isomerization of a peroxidising herbicidal thiadiazolidine. In Proceedings of the Brighton Crop Protection Conference – Weeds, pp. 771–776, British Crop Protection Council, Farnham, UK Robert Edwards* and David P. Dixon are at the Dept of Biological Sciences, University of Durham, Durham, UK DH1 3LE; Virginia Walbot is at the Dept of Biological Sciences, Stanford University, Stanford, CA 94305-5020, USA. *Author for correspondence (tel 144 191 374 2428; fax 144 191 374 2417; e-mail [email protected]).