| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
|
First published online October 27, 2006; 10.1104/pp.106.089458 Plant Physiology 142:1364-1379 (2006) © 2006 American Society of Plant Biologists
Plant Glutathione Peroxidases Are Functional Peroxiredoxins Distributed in Several Subcellular Compartments and Regulated during Biotic and Abiotic Stresses1,[W]Unité Mixte de Recherche Institut National de la Recherche Agronomique-Université Henri Poincaré 1136, Interactions Arbres/Micro-organismes, Institut de Formation et de Recherche 110 Génomique, Ecophysiologie et Ecologie Fonctionnelles, Université Henri Poincaré, Faculté des Sciences, 54506 Vandoeuvre cedex, France (N.N., E.G., J.-P.J., N.R.); Commissariat à l'Energie Atomique/Cadarache, Direction des Sciences du Vivant, Département d'Ecophysiologie Végétale et de Microbiologie, Laboratoire d'Ecophysiologie Moléculaire des Plantes, 13108 Saint-Paul-lez-Durance cedex, France (V.C., P.R.); Institut de Biologie Moléculaire des Plantes, Centre National de la Recherche Scientifique, 67084 Strasbourg cedex, France (J.G.); Department of Chemistryand Biochemistry, and Center for Biotechnology and Genomics, Texas Tech University, Lubbock, Texas 79409–1061 (M.H., D.B.K.); and Institut de Biotechnologie des Plantes, Unité Mixte de Recherche 8618, Université de Paris Sud, 91405 Orsay cedex, France (E.I.)
We provide here an exhaustive overview of the glutathione (GSH) peroxidase (Gpx) family of poplar (Populus trichocarpa). Although these proteins were initially defined as GSH dependent, in fact they use only reduced thioredoxin (Trx) for their regeneration and do not react with GSH or glutaredoxin, constituting a fifth class of peroxiredoxins. The two chloroplastic Gpxs display a marked selectivity toward their electron donors, being exclusively specific for Trxs of the y type for their reduction. In contrast, poplar Gpxs are much less specific with regard to their electron-accepting substrates, reducing hydrogen peroxide and more complex hydroperoxides equally well. Site-directed mutagenesis indicates that the catalytic mechanism and the Trx-mediated recycling process involve only two (cysteine [Cys]-107 and Cys-155) of the three conserved Cys, which form a disulfide bridge with an oxidation-redox midpoint potential of –295 mV. The reduction/formation of this disulfide is detected both by a shift on sodium dodecyl sulfate-polyacrylamide gel electrophoresis or by measuring the intrinsic tryptophan fluorescence of the protein. The six genes identified coding for Gpxs are expressed in various poplar organs, and two of them are localized in the chloroplast, with one colocalizing in mitochondria, suggesting a broad distribution of Gpxs in plant cells. The abundance of some Gpxs is modified in plants subjected to environmental constraints, generally increasing during fungal infection, water deficit, and metal stress, and decreasing during photooxidative stress, showing that Gpx proteins are involved in the response to both biotic and abiotic stress conditions.
In plant cells, aerobic reactions such as photosynthesis or respiration lead to reactive oxygen species (ROS) production. These ROS, such as superoxide radicals, hydroxyl radicals, or hydrogen peroxide (H2O2), can damage biological molecules, including nucleic acids, lipids, and proteins. Rates of ROS generation and cellular ROS levels both increase greatly when plants are subjected to environmental or biotic stresses. Plants have developed several nonenzymatic and enzymatic systems to protect against oxidative damage caused by these ROS. Carotenoids, tocopherols, glutathione (GSH), and ascorbate are the major nonenzymatic antioxidant compounds (Noctor and Foyer, 1998  ). Enzymatic systems include superoxide dismutases, catalases, ascorbate peroxidases, and peroxiredoxins (Prxs), also named thioredoxin (Trx) peroxidases.
Usually, plant Prxs are classified as belonging to one of four subgroups, called 2-Cys Prx, 1-Cys Prx, type II Prx, and Prx Q. They were initially identified by sequence analysis (i.e. on the number and position of conserved Cys) and on the catalytic mechanism used for peroxide reduction (Rouhier and Jacquot, 2002
Recent studies have shown that some plant and yeast (Saccharomyces cerevisiae) GSH peroxidases (Gpx) can reduce peroxides, much more efficiently or sometimes exclusively, by using the Trx system rather than GSH as a reductant (Herbette et al., 2002
Most of the Gpxs found in animal cells are well-characterized selenium-containing enzymes. Because of the high reactivity of the active site seleno-Cys, enzymes of this family are among the most efficient antioxidant systems in animal cells (Maiorino et al., 1990
The level of Gpx mRNA from various organisms is affected by stress conditions (Criqui et al., 1992 The annotation of the first release of the poplar (Populus trichocarpa) genome indicates the presence of six complete Gpx genes. To understand why so many Trx-dependent peroxidases are present in plant cell compartments, we first focused our attention on the specificity of Gpx proteins toward different electron donors and substrates. All of them use Trx but not GSH as a donor, but most importantly, a specific interaction was found between chloroplastic Gpxs and y-type Trxs. No specificity was observed with regard to substrates. Moreover, we have demonstrated that only two of the three conserved Cys present in plant Gpxs participate in the catalytic cycle and in Trx-mediated regeneration. The subcellular localization of two Gpx isoforms was characterized by green fluorescent protein (GFP) fusions in two distinct subcellular compartments, and the gene expression and protein abundance were analyzed in various plant tissues and stress conditions. The amount of some Gpxs is modified upon pathogen infection and in response to water deficit and photooxidative and metal stresses, revealing their participation in stress responses.
Sequence and Genome Analysis Five different isoforms of Gpx, termed PtrcGpx 1 to 5 (for poplar Gpx) were initially identified in the different poplar expressed sequence tag (EST) databases available. With the recent release of the first version of the poplar genome sequence by the Department of Energy (DOE) Joint Genome Institute (JGI), a sixth isoform, very similar to PtrcGpx3, was identified, and these two isoforms were named PtrcGpx3.1 and PtrcGpx3.2. The only obvious difference between the two sequences is the presence of an N-terminal extension in PtrcGpx3.2, otherwise they are 90% identical and display a 99% functional homology in their predicted mature form. The protein sequences of all poplar Gpxs identified are shown in Figure 1 and compared with Gpxs from yeast and human. The percentage of strict identity ranges from 62% to 90% among all poplar Gpxs and from 20% to 48% versus yeast and mammalian Gpxs. The lengths of the proteins vary from 168 to 238 amino acids, depending on the presence of N-terminal extensions in some isoforms.
From this amino acid sequence comparison, it appears that poplar Gpxs are more closely related to yeast Gpxs than to human Gpxs. Indeed, three Cys (residues generally assumed to be essential for Gpx catalysis) are not only strictly conserved in all poplar Gpxs but also in all higher plant and yeast sequences (Fig. 1; data not shown). This is in clear contrast to animal Gpxs, which display either two conserved Cys residues or in some isoforms a Cys and a seleno-Cys. The difference in the number of conserved Cys could result in different reactivities for plant and yeast Gpxs compared to mammalian enzymes. The closest human homolog to plant and yeast Gpxs is Gpx4, also known as phospholipid hydroperoxide Gpx. Some residues are highly conserved in all these sequences, especially in the neighborhood of the two first conserved Cys (the respective consensus sequences are VASx[C/U]G and FPCNQF, with U being the symbol for seleno-Cys). The three residues, Gly, Gln, and Trp denoted with asterisks in Figure 1, which are involved in the catalytic mechanism of the mammalian seleno-Gpxs (Prabhakar et al., 2005 ), are also conserved in plant and yeast sequences. However, no information is available about the role of these residues in plant or yeast Gpxs.
Another interesting feature, arising from the poplar genome release, is that the gpx gene structure is very conserved in poplar. All the genes contain six exons of almost the same length, whereas the size of the intron sequences varies greatly. The major differences are found in the first exon, because it includes putative targeting sequences for some isoforms (Supplemental Table S1). Another difference is the slightly smaller size of exon 6 for PtrcGpx1 (30 nucleotides instead of 33 in the other Gpx sequences, resulting in a PtrcGpx1 sequence shortened by one amino acid in the C terminus). This organization is similar in Arabidopsis (Arabidopsis thaliana), where the eight genes also contain six exons. In Arabidopsis, the last exon is also shorter for some genes, in particular for the homologs of PtrcGpx1 (Rodriguez Milla et al., 2003 Some characteristics of poplar Gpxs are summarized in Table I . The consensus subcellular predictions obtained with different software (Predotar, TargetP, Mitoprot, and Psort) are as follows: PtrcGpx1 carries a putative plastidic transit peptide; PtrcGpx3.2, a mitochondrial or plastidic one; and PtrcGpx2 and PtrcGpx5, a peptide presumably directing the proteins into the secretory pathways, whereas PtrcGpx4 and PtrcGpx3.1 should be localized in the cytosol.
We then built a phylogenetic tree using most of the plant Gpx sequences present in the databases with an emphasis on the eight and five Gpxs identified in Arabidopsis and Oryza sativa genomes, respectively (Fig. 2 ). Although the tree was constructed using sequences devoid of N-terminal extensions, five subgroups of Gpx, which are likely to correspond to the predicted intracellular protein localization, can be defined. This means that there are probably some amino acid signatures specific for each subgroup. It also appears that some organisms possess two members of the same group. For example, Arabidopsis has two putative chloroplastic and two cytosolic Gpxs, explaining the higher gene content in this plant. In contrast, the subgroup including the predicted secreted proteins is not found in O. sativa.
Biochemical Characterization We have cloned five of the six Gpx sequences, minus their N-terminal extensions, in the pET-3 d plasmid for subsequent expression in Escherichia coli. Because of the high similarity between PtrcGpx3.1 and PtrcGpx3.2, we have only expressed the mature form of PtrcGpx3.2, which was the most abundant in terms of ESTs and also the first identified. The sizes of the mature forms that we have expressed range from 164 to 171 amino acids. All the proteins, except PtrcGpx2, were soluble. After purification to homogeneity, the yields ranged from 5 mg for PtrcGpx2 (after resolubilization from inclusion bodies) to about 100 mg of protein/L culture for the other isoforms.
As it was demonstrated earlier that Trx was a better electron donor than GSH for some Gpx-like proteins, we have tested different electron donor systems, i.e. the Trx and Grx systems or GSH alone. We have first measured the Gpx activity using a spectrophotometric test coupled to NADPH oxidation.
Except for PtrcGpx2, which was completely inactive, all the proteins (PtrcGpx1, PtrcGpx3.2, PtrcGpx4, and PtrcGpx5) catalyzed H2O2 reduction at similar rates using poplar Trx h1 as the electron donor, NADPH, and recombinant Arabidopsis type B NADPH-thioredoxin reductase (data not shown; Table II
). In contrast, no activity was found when using either GSH alone or in combination with different poplar Grxs belonging to the same subgroup, either dithiol Grx with a classical CxxC active site (Grx C1, C2, C3, and C4) or a monothiol chloroplastic Grx (Grx S12), which possesses a CxxS active site (see Rouhier et al., 2006
Because all recombinant proteins, with the exception of PtrcGpx2, were equally active in transferring electrons from reduced Trx to H2O2, only one of them, PtrcGpx3.2, was selected for more detailed study.
As PtrcGpx3.1, the cytosolic isoform, is very similar in sequence to PtrcGpx3.2, we tested in vitro various cytosolic Trxs of the h type in an attempt to find a potential physiological electron donor for this protein, using NADPH oxidation coupled to the NTR-catalyzed reduction of Trx to monitor activity. All the poplar Trx h tested (Trx h1, h3, and h5) are efficient electron donors for PtrcGpx3.2 in vitro, with apparent Km values being around 10 µM (Table II). It was previously shown that Trx h2, although it colocalizes with PtrcGpx3.2 in mitochondria, is a very poor electron donor (Gelhaye et al., 2004
Peroxide Specificity The substrate specificity of PtrcGpx3.2 was tested under steady-state conditions using Trx h1 as reductant and different types of peroxides, ranging from H2O2 to more complex molecules, such as tert-butyl hydroperoxide (tBOOH) or cumene hydroperoxide (COOH), as electron acceptors. The Km values for the different substrates and the catalytic efficiencies of the enzymes are summarized in Table II. Although slightly different (Km ranging from 239 µM to 1.41 mM), the Km values for the three substrates are all of the same order of magnitude in the millimolar range. In terms of catalytic efficiency (kcat/Km), PtrcGpx3.2 is slightly more efficient in reducing COOH (5.3 x 104 M–1 s–1) compared to H2O2 (20 x 103 M–1 s–1) and to tBOOH (6.4 x 103 M–1 s–1). We then examined the stoichiometry of the reaction catalyzed by Gpxs, by mixing known concentrations of reduced enzyme with various known concentrations of tBOOH in the absence of reductants. After completion of the reaction, the remaining peroxide content was measured using the ferrous oxidation of xylenol orange (FOX) colorimetric method. All the enzymes tested (PtrcGpx1, 3.2, and 5) can reduce 1 mol substrate per mole enzyme, indicating that Gpxs use only one or two Cys but certainly not three.
It is known that, for enzymes that use sulfenic acid chemistry for their catalysis, the stoichiometry depends on the number of Cys involved. When one or two Cys are involved, the stoichiometry of the reaction is 1 mol of enzyme oxidized per mol of peroxide reduced (Boschi Muller et al., 2000
To understand the mechanism used by PtrcGpxs for peroxide reduction and for the Trx-dependent regeneration, site-directed mutagenesis was used to generate four variants of PtrcGpx3.2 (PtrcGpx3.2 C107S, C136S, C155, or C136/155S) in which Cys residues were replaced by Sers. The activity of the mutated proteins was tested in the presence of the Trx system. As expected, the Cys 107S mutant is totally inactive. While PtrcGpx3.2 C155S or the double mutant PtrcGpx3.2 C136/155S are able to reduce peroxides in single-turnover experiments (i.e. prereduced samples of these two variants can reduce an equal amount of peroxide), they are completely inactive with the Trx-reducing system, regardless of the peroxide substrate used. This indicates that Trx is probably not able to directly reduce the sulfenic acid formed on Cys-107. In contrast, PtrcGpx3.2 C136S displays almost identical enzymatic properties to PtrcGpx3.2 (Table II). The only difference is a small modification of the Km value for peroxides that results in a decreased catalytic efficiency. These results indicate unequivocally that Cys-107 and Cys-155 are required for Trx regeneration, whereas Cys-136, albeit strictly conserved and found in a very conserved motif, does not have a catalytic role.
We examined the behavior of PtrcGpx3.2 and its various Cys/Ser variants during denaturing polyacrylamide gel electrophoresis (SDS-PAGE) under different conditions. Under reducing conditions (30 mM DTT), all the proteins migrated at their expected molecular masses, around 19 kD (Fig. 4A
, lanes 2, 4, 6, 8, and 10). Under oxidizing conditions (30 mM H2O2), the migration of PtrcGpx3.2, as well as of its C107S and C136S variants, is shifted to a lower apparent molecular mass (around 13 kD), whereas the migration of C155S and C136S/155S mutated proteins was not altered (Fig. 4A, lanes 1, 3, 5, 7, and 9). The change in the migration profile under oxidizing conditions for all the proteins in which Cys-155 has not been mutated suggests that an intramolecular disulfide bond involving Cys-155 can be formed. The disulfide, formed between Cys-107 and Cys-155 in the case of the C136S mutant and between Cys-136 and Cys-155 in the case of the C107S mutant, apparently modifies the overall structure of the proteins sufficiently to alter the migration behavior during electrophoresis. It should also be pointed out that a covalent dimer, involving Cys-107 of two subunits, can be observed for the double mutant under nonreducing conditions but not under reducing conditions, as has been previously observed for mutated proteins in which only the catalytic Cys remains (Rouhier et al., 2004
Gel-filtration experiments performed with reduced or untreated PtrcGpx3.2 showed that it consistently elutes with an apparent molecular mass of 38 kD, corresponding to a homodimer (data not shown). As the protein migrates as a monomer (around 19 kD) under both reducing and nonreducing conditions during denaturing SDS-PAGE, the dimer observed during gel filtration cannot involve an intermolecular disulfide and must be stabilized by noncovalent forces (e.g. electrostatic or hydrophobic interactions). Redox titrations of PtrcGpx3.2 using redox equilibration buffers composed either of defined mixtures of oxidized and reduced DTT or of defined mixtures of GSH plus oxidized GSH (GSSG) are most readily interpreted in terms of a single disulfide/dithiol redox couple per monomer, with a midpoint redox potential (Em) value of 295 ± 10 mV at pH 7.0 (Fig. 5 ). Redox titrations (data not shown) over a more positive range than that shown in Figure 5 (i.e. Eh values ranging from –250 mV to –80 mV), in which defined mixtures of GSH plus GSSG were used as redox equilibration buffers, demonstrated that no additional more positive dithiol/disulfide couple is present in the protein. As Cys-136 is not involved in catalysis and the protein is not present as a disulfide-linked covalent dimer, one can conclude that the value observed corresponds to an intramolecular disulfide bridge linking Cys-107 to Cys-155.
Expression of Prx in Plant Organs and in Stress Conditions We first performed an in silico analysis based on ESTs found in the DOE JGI database and found 191 ESTs coding for poplar Gpxs (Table I). The different isoforms are not equally represented among these ESTs; PtrcGpx3.2 seems to be the most frequently expressed gene, with 85 corresponding ESTs (i.e. approximately 45%). We also found 48 ESTs for PtrcGpx1, 10 for PtrcGpx2, 23 for PtrcGpx3.1, 17 for PtrcGpx4, and eight for PtrcGpx5. These results were combined with a semiquantitative reverse transcription (RT)-PCR approach. As these genes belong to the Prx family, we have studied the transcript expression pattern of all Gpxs together with the nine Prxs existing in poplar in seven different tissues (roots, young leaves, expanded leaves, fruits, stems, petioles, and stamen). For most genes, we have been able to find specific hybridizing areas. However, we were sometimes unable to distinguish some closely related sequences (i.e. Prx Q1 and Q2 or 2-Cys Prx A and B). In these two cases, we used nonspecific primers able to amplify both transcripts (Prx Q1 and Q2 or 2-Cys Prx A and B), but we were also able to define one primer to specifically amplify one gene (Prx Q1 and 2-Cys Prx A). From the difference between these two amplifications, the expression of the second gene can be deduced. For example, as the 2-Cys Prx A is only expressed in expanded leaves, we can estimate from the 2-Cys Prx A and B expression profile that 2-Cys Prx B is expressed in fruits, stems, petioles, and stamens and possibly in expanded leaves. All the genes exhibit different expression patterns, but there is at least one gene expressed in each organ tested (Fig. 6A ). These data, coupled with the in silico analysis, are indicative of a broad distribution and expression of Prxs in plants, at least at the transcript level. Nevertheless, the multiplicity of Prx genes can be explained by some specificity in the expression pattern. For example, out a total of 17 genes, only three (coding for PtrcGpx2, PtrcGpx 3.1, and PtrcPrx IIE) are expressed at detectable levels in roots under the growth conditions used. One gene (PtrcPrx IIE) encodes a plastidic protein, another (PtrcGpx3.1) appears to encode a cytosolic Gpx, and the third (PtrcGpx2) appears to encode a secreted Gpx. In contrast, all the genes are expressed in expanded leaves. Given the specificity of the two chloroplastic Gpxs (PtrcGpx1 or PtrcGpx3.2) for a Trx of the y type as an electron donor, it was also of interest to examine the expression patterns of these two poplar Trxs. Both Trx y1 and y2 were found to be expressed in all tissues tested, and thus their role cannot simply be restricted to serving as specific reductants for plastidial Gpxs.
To get a deeper view of the expression of Gpx genes, we used both microarray data available using Genevestigator and the results presented by Rodriguez Milla et al. (2003) to follow the expression pattern of the different Arabidopsis Gpx genes in the whole plant (Fig. 6B; Zimmermann et al., 2004). Based on the results of Rodriguez Milla et al. (2003), AtGpxs1 to 7 are expressed in all tissues analyzed, except AtGpx2, which does not seem to be expressed in dry seeds. AtGpx4 and 7 are the less expressed genes, as they were only detected by RT-PCR and few or no ESTs are found for these genes (Rodriguez Milla et al., 2003). Using Genevestigator, we observed that AtGpx1, AtGpx2, and AtGpx6, respectively homologous to PtrcGpx1, PtrcGpx2, and PtrcGpx3.2, were highly expressed in all organs or tissues analyzed (i.e. in young, expanded, and senescent leaves, as well as in stems, roots, and flowers), while AtGpx3, AtGpx5, AtGpx7, and AtGpx8, and especially AtGpx4, generally show a low expression level in these organs. Nevertheless, as expected for a chloroplastic protein, AtGpx1 is preferentially expressed in green tissues (young and expanded leaves, stems, and flowers). The main difference between the two sources concerns the expression of AtGpx6, which was found to be low in all organs except dry seeds by northern blot, while microarray data indicates a strong expression in all organs. Using an antibody raised against purified recombinant PtrcGpx3.2, we first checked its specificity against the recombinant proteins. It reacts with all five recombinant proteins (which incidentally all migrate at slightly different positions on SDS-PAGE) but only faintly with PtrcGpx1 and 5 and more strongly with PtrcGpx2, 3.2, and 4 (Fig. 7A ). This antibody is thus specific for Gpx as a whole, including the sixth protein, PtrcGpx3.1, which is very similar to PtrcGpx3.2 but does not discriminate between the isoforms. We could, however, differentiate at least two isoforms with two slightly different molecular masses in poplar protein extracts. The higher molecular mass band seems specific to roots, and because the signals obtained with the extracts from all the other organs tested have a similar size, they could thus correspond to a single isoform that we cannot identify for the reasons explained above (Fig. 7B).
In Arabidopsis (Fig. 7C), two bands of different sizes arising from two or more different isoforms have been detected in almost all organs, except roots. The protein(s) of smaller molecular mass is most easily detected in leaves and flowers, while in contrast the protein(s) of higher molecular mass is present in stems and roots and to a lesser extent in leaves and flowers. Using proteins extracted from purified leaf chloroplasts, only the lower band was detected (data not shown). Given the fact that AtGpx7 is weakly expressed at the transcriptional level, only the other two predicted chloroplastic proteins (i.e. AtGpx1 and AtGpx6) are likely to be detectable. Thus, we conclude that the lower band consists predominantly of AtGpx1 and/or AtGpx6. The upper band is likely to arise from the presence of AtGpx2, as it is not a chloroplastic protein and because AtGpx3, AtGpx4, AtGpx5, and AtGpx8 also seem to be weakly expressed at the transcriptional level. We next investigated the abundance of these proteins in stress situations (Fig. 8 ). Gpx abundance as a function of time was followed after infection of poplar by the agent of poplar rust, the fungus Melampsora larici-populina, and using two different isolates, a virulent one that induces a compatible reaction and an avirulent one that induces an incompatible reaction (Fig. 8A). As is the case for untreated leaf extracts, a single band is visible. During the incompatible reaction, the expression of Gpx increased greatly 2 h after infection and even more after 48 h. The pattern of expression during the compatible reaction is strikingly different, with a lower protein level observed at both 2 h after infection and then again at 24 h after infection but with levels very similar to those seen prior to infection found at all other times. Because this expression pattern is surprising, we have analyzed the expression of Trx h1, which is constitutively expressed and was never found to vary with stress, as a control (N. Rouhier, unpublished data). As expected, the Trx h1 abundance is not dependent on the pathogen infection, regardless of which fungal isolate was used.
In addition, we followed Gpx levels in young leaves from Arabidopsis or tobacco (Nicotiana tabacum) plants submitted to various abiotic stresses. After 6 d of water deficit, the intensity of the lower band decreased, while that of the upper band increased (Fig. 8B). This was even more pronounced in young tobacco leaves subjected to the same conditions with an increased intensity of the Gpx signal, which could correspond to the presence of one or more Gpx isoforms (Fig. 8B). Similar profiles were observed using expanded leaves (data not shown). In plants subjected to photooxidative stress (Fig. 8C), the amount of the two proteins varied in the opposite way; the intensity of the upper band increased slightly under stress conditions in expanded leaves, whereas the intensity of the lower band decreased after 3 or 8 d of treatment. In plants exposed to metal stresses induced by application of various concentrations of cadmium or copper for 2 and 7 d, only the amount of the lower Mr protein(s) detected is modified. It decreased at 150 µM copper after 2 d (Fig. 8E) but increased strongly after 7 d at cadmium concentrations of 50 and 75 µM (Fig. 8D) and at copper concentrations of 25, 75, and 150 µM (Fig. 8E).
PtrcGpx2 and PtrcGpx5 are predicted to be secreted but were not studied here, as we were unable to amplify the 5' end coding for their N-terminal extensions from either cDNA libraries or genomic DNA. Although Arabidopsis and O. sativa homologs possess this extension, it is not yet clear whether these extensions really belong to the cDNA or are annotation or assembly errors. PtrcGpx4 and PtrcGpx3.1 contain neither an N-terminal extension nor known C-terminal signals, suggesting that the proteins are cytosolic. To determine the subcellular localization of PtrcGpx1 and PtrcGpx3.2, the 5' portions of the open reading frame encoding only the N-terminal parts were fused to the sequence coding for a GFP. The recombinant pCK GFP S65C plasmids were then used to transiently transform tobacco cells by bombardment. In the guard cells shown here, it appears that Gpx1 is specifically targeted to chloroplasts and that Gpx3.2 is directed both to chloroplasts and mitochondria (Fig. 9 ). Similar results were obtained in mesophyll cells (data not shown).
As more and more genomes are sequenced and annotated, it is now clear that multigenic families are very common in organisms. Here, we describe the existence of six Gpx isoforms in the model tree poplar. By comparison, eight and five Gpx isoforms are present in the genome of Arabidopsis and O. sativa, respectively. The number of Gpx genes thus varies among species, probably because of gene duplication events. A continuing challenge is to understand why so many genes or proteins with similar functions are needed in plants. To answer this question, both the abundance and localization of the proteins and the substrate specificities of these enzymes were studied. As a result of our study, it is now clear that the poplar enzymes originally named Gpxs are reduced by Trxs rather than by GSH and belong to the Prx family. This large family thus includes 15 different Prxs in poplar.
The poplar Gpxs do not exhibit any difference in the reduction of H2O2 compared to complex hydroperoxides (COOH and tBOOH), as the catalytic efficiencies (kcat/Km) vary only over a 20-fold range (i.e. from 2.6 to 53 x 103 M–1 s–1). The reduction of complex peroxides is consistent with a role for Gpxs (and for Prxs in general) that complements other systems, like catalases or ascorbate peroxidases, which specifically reduce H2O2. The catalytic efficiencies of poplar Gpxs lie in the range previously described for other Prxs. Rouhier et al. (2004)
In so far as the electron donor specificity of poplar Gpxs is concerned, these enzymes seem to behave like previously characterized plant Gpx in that they can be reduced by Trxs, as earlier demonstrated for some plant Gpx (Herbette et al., 2002
The question of which Trx is the actual physiological electron donor for different Gpxs remains unsettled. Taking the Trx h family as an example, we were unable to establish a preference between Trx h1, h3, and h5 in in vitro assays. It was previously demonstrated that neither a poplar Trx h2 nor an Arabidopsis Trx o were able to serve as reductants to PtrcGpx3.2 (Gelhaye et al., 2004
Considering the putative secreted isoforms, namely PtrcGpx2 and PtrcGpx5, we found that Trx h5 is able to reduce PtrcGpx5 in a catalytic manner (data not shown). Juarez-Diaz et al. (2005)
There are at least 11 chloroplastic Trxs in Arabidopsis (NtrC, CDSP32, Trx x, Trx y1 and y2, Trx f1 and f2, Trx m1, m2, m3, and m4) and even more in poplar (E. Gelhaye, unpublished data). Although we did not test the efficiency of the bimodular Trxs (NtrC and CDSP32), it appears that only the two Trx y are able to efficiently support the activity of PtrcGpx1 and PtrcGpx3.2. These chloroplastic Trxs all have very similar Em, and so the specificities observed must arise from properties of the Trxs other than their Em values (e.g. the distributions of surface charges and of surface hydrophobic patches). In a previous study, it was found in Arabidopsis that only Trx x is able to reduce 2-Cys Prx, and Trx y is the most efficient electron donor to Prx Q (Collin et al., 2004
Using the digital northern function of Genevestigator online database (https://www.genevestigator.ethz.ch; Zimmermann et al., 2004
When starting this work, one question was to determine whether one, two, or three Cys were involved in the reactions catalyzed by plant Gpxs. Although yeast Gpx2 had been shown to possess a Trx-dependent activity, involving only the two Cys equivalent to Cys-107 and Cys-155 in poplar (Tanaka et al., 2005 Unlike the case for many other eukaryotic Prx characterized so far, we did not observe the formation of overoxidized forms of Gpx. Indeed, the stoichiometry of the reaction catalyzed by Gpxs is consistently 1, the activity measured in the presence of the Trx system is linear with time, and H2O2-treated proteins do not undergo the mass increment expected for sulfinic or sulfonic formation (data not shown). We thus propose a three-step reaction mechanism for plant Gpxs similar to the one used by Prx Q, with two Cys forming an intramolecular disulfide bridge in the oxidized state (Fig. 10 ). These steps are: (1) a nucleophilic attack of the catalytic Cys (Cys-107) on the peroxide with the release of an alcohol and the concomitant formation of a sulfenic acid; (2) an attack of the sulfenic acid by Cys-155 and formation of an intramolecular disulfide bridge between Cys-107 and Cys-155; and (3) a reduction of the disulfide bridge by Trx.
Expression and Localization of Gpx in Plants We analyzed the expression of all the Prxs in poplar organs. All the members of this family are expressed at the transcriptional level in at least one of the organs tested, with a specific expression pattern for each of them. Gpxs are likely to be present in all organs and in nearly all cell compartments, suggesting a redundancy with the Prxs identified previously. Interestingly, many isoforms are expressed in flowers or in fruits, but their precise physiological function is yet unknown. Nevertheless, to explain the need for proteins with identical functions, we can imagine that some genes could be expressed in specific conditions under stress, for example, or could be temporally regulated.
When poplar leaves are subjected to a biotic stress (infection by the rust fungus M. larici-populina), the abundance of these proteins is modified. This is in line with previous observations that the levels of Prx Q, Prx IIC, and Prx IIF but not of 2-Cys Prx are modified in this situation (Rouhier et al., 2004
Another possible explanation for the redundancy of Trx peroxidases as a whole could be linked to their subcellular distribution. Using GFP fusion, we demonstrated that two Gpxs (PtrcGpx1 and 3.2) are localized in chloroplast, one of these (PtrcGpx3.2) being also targeted to mitochondria. This dual targeting is not unusual, having already been observed for some other antioxidant proteins (GSH reductase [GR], monodehydroascorbate reductase, and ascorbate peroxidase; Chew et al., 2003 Adding the results of this investigation to the previously available data on distribution of Prxs, eight poplar Prxs are now known to be expressed in plastids (Prx IIE, 2-Cys Prx A and B, Prx Q1 and 2, and PtrcGpx1 and PtrcGpx3.2), two in mitochondria (Prx IIF and PtrcGpx3.2), three in the cytosol (Prx IIB, Prx IIC, and PtrcGpx3.1), one in the nucleus (Prx 1-Cys), and two predicted to be directed to the secretory pathways (PtrcGpx2 and PtrcGpx5), assuming that these proteins indeed contain an N-terminal extension. This situation is slightly different from Arabidopsis, in which there is only one Prx Q but an additional cytosolic Prx II and eight Gpxs instead of the six found in poplar.
Materials H2O2, tBOOH, COOH, NADPH, GSH, and GR from yeast (Saccharomyces cerevisiae) were purchased from Sigma.
RT-PCR
Western Blotting
After identification in the databases (see "Results"), the nucleotide sequences of PtrcGpx2, 3.2, and 4 were then amplified by PCR using the primers described in Supplemental Table S2 from a Populus tremula x trichocarpa cv Beaupre poplar root cDNA library (Kohler et al., 2003
The recombinant plasmids obtained were used to transform the BL21(DE3) pSBET strain of E. coli (Schenk et al., 1995 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TOP
ABSTRACT
RESULTS
), Trx m3 (
, dotted line), Trx m1 (
, dotted line), Trx m2 (
), Trx f2 (
), Trx x (x), poplar Trx h1(
), Trx y2 (
), and Trx y1 (
).
-D-thiogalactoside. For PtrcGpx3.2, the amount of soluble protein increased when bacteria were grown over