The plasma membrane oxidase NtrbohD is responsible for AOS production in elicited tobacco cells

Authors


*For correspondence (fax +33 0380 69 32 75; e-mail simon@epoisses.inra.fr).
† These authors contributed equally to this work.

Summary

A cDNA encoding a protein, NtrbohD, located on the plasma membrane and homologue to the flavocytochrome of the neutrophil NADPH oxidase, was cloned in tobacco. The corresponding mRNA was accumulated when tobacco leaves and cells were treated with the fungal elicitor cryptogein. After elicitation with cryptogein, tobacco cells transformed with antisense constructs of NtrbohD showed the same extracellular alkalinization as the control, but no longer produced active oxygen species (AOS). This work represents the first demonstration of the function of a homologue of gp91–phox in AOS production in elicited tobacco cells.

Introduction

Plants defend themselves against pathogens by triggering a wide range of mechanisms, including the hypersensitive response (HR), which leads to cell death lesions at the infection sites, thus limiting pathogen growth to a restricted area of the plant (Dangl, 1998; Dangl and Holub, 1997; Dangl et al., 1996; Keen, 1986; Morel and Dangl, 1997). A characteristic feature of the HR is the rapid and intense production of active oxygen species (AOS), mainly superoxide anion (O2− .) and hydrogen peroxide (H2O2). Because the generation of these AOS is believed to contribute to several disease resistance strategies (direct antimicrobial activity, cross-linking of cell wall proteins, induction of defence related genes, cell death, etc.), the mechanism and regulation of their biosynthesis have been extensively studied (for reviews see Bolwell, 1999; Bolwell and Wojtaszek, 1997; Lamb and Dixon, 1997; Wojtaszek, 1997). In particular, identification of the enzyme(s) responsible for the production of AOS during plant defence, is an intense area of research. Although the source of AOS in plants has been controversial, several lines of evidence support the involvement of an enzyme similar to the mammalian NADPH oxidase: (i) the oxidative burst in challenged plant cells resembles that of human neutrophils since it is rapid, intense, transitory, and induced by a pathogen (ii) the oxidative burst in plants is inhibited by an irreversible inhibitor of the mammalian NADPH oxidase, diphenylene iodonium (DPI) (Auh and Murphy, 1995; Simon-Plas et al., 1997). The enzymatic complex found in mammalian neutrophils has two components located in the plasma membrane (gp91–phox and p22–phox) and becomes active when several cytosolic proteins (p47–phox, p67–phox and the small G protein Rac) join these membranar components (Heyworth et al., 1999). Genes that encode gp91–phox homologues have been identified in Arabidopsis, rice, tomato and potato (Amicucci et al., 1999; Groom et al., 1996; Keller et al., 1998; Torres et al., 1998; Yoshioka et al., 2001).

Elicitins, a family of low molecular weight proteins secreted by many species of the oomycete Phytophthora (Ricci et al., 1986), induce a hypersensitive-like response in tobacco (Ricci et al., 1993), Raphanus (Keizer et al., 1998) and several Brassica species (Bonnet et al., 1996; Roussel et al., 1999) and have been recently identified as a new class of sterol carrier proteins (Mikes et al., 1998; Vauthrin et al., 1999). In order to understand the biochemical processes triggered by elicitins, the effects of cryptogein on tobacco cell suspensions have been studied for several years (for review see Ponchet et al., 1999). The early events detected in cells following elicitin treatment include binding of the elicitor to a high affinity site on the plasma membrane (Wendehenne et al., 1995), alkalinization of the extracellular medium, efflux of potassium and chloride (Blein et al., 1991), a fast influx of calcium (Tavernier et al., 1995), a transient production of active oxygen species (AOS) (Rustérucci et al., 1996; Simon-Plas et al., 1997) and an activation of MAPK homologues (Lebrun-Garcia et al., 1998; Du Zhang and Klessig, 1998).

The aim of this work was to determine the origin of AOS production induced in tobacco plant or cells by cryptogein. We first cloned a tobacco homologue of gp91–phox, NtrbohD (Nicotianatabacumrespiratory burst oxidase homologue) and we studied its expression in various plant tissues and in cultured cells treated or untreated with cryptogein. The function of NtrbohD was assessed using transgenic cells knocked-out for NtrbohD expression by an antisense strategy and the location of the protein on the plasma membrane was demonstrated.

Results

Cloning and expression of NtrbohD, a tobacco gp91–phox homologue

Using degenerate oligonucleotide primers (described in the Methods section) deduced from consensus sequence regions from human gp91–phox (Acc N° p04839), rice (Acc N° × 93301) and Arabidopsis (Acc N°AF055357) homologues, RT-PCR experiments were performed with mRNA from tobacco leaves and yielded a product of 1 kb, which was used as a probe to screen a cDNA library prepared from elicited tobacco leaves. This strategy led to the identification of 3 clones corresponding to the same cDNA of 3338 bp, and coding for a protein of 939 amino acids (NtrbohD) (Figure 1). NtrbohD displays a strong homology with some gp91–phox plant homologues cloned previously: 72% identity with AtrbohD and 63% identity with OsrbohA (Groom et al., 1996; Torres et al., 1998). The homology of the whole sequence with human gp91–phox is less pronounced (38% identity). However, as it has been described for the different Atrboh sequences, some characteristic features are conserved: the number and position of the putative membrane spanning helixes, part of the C-terminus domain that contains the motifs responsible for the binding of FAD and NADPH cofactors, in particular the four histidines postulated to bind the two hemes (Davis et al., 1998) (Figure 1). NtrbohD has a hydrophilic N-terminal extension with one Ca2+-binding motif of the EF-hand loop type already observed in Atrboh A-F (Keller et al., 1998; Torres et al., 1998) and typical of plant gp91–phox homologues. This domain (aa 258–270), is the unique site exhibiting a complete identity with the canonical EF-hand domain described by Kretsinger (Kretsinger, 1996). Another domain of the protein (aa 302–314) which displays a similarity of 74% with this canonical EF-hand domain is not indicated in Figure 1.

Figure 1.

Amino-acid sequence alignments of gp91–phox and respiratory burst oxidase homologues (rboh) from Arabidopsis (AtrbohD), rice (OsrbohA) and tobacco (NtrbohD).

The predicted transmembranar hydrophobic regions are enclosed in rectangles, and putative heme-co-ordinating histidines located in these regions are indicated by an asterisk. Nucleotide binding sites of gp91–phox are underlined ((a) potential FAD- binding site, (b and c) potential NADPH binding domains) as well as those of p47–phox binding sites (1–4) and the EF-hand motif is in bold and highlighted.

The expression of the gene encoding the NtrbohD protein was examined in different tobacco plant tissues and cells by Northern blot analysis. On gel blots of total RNA from tobacco roots a single 3.8 kb transcript hybridized with the NtrbohD cDNA probe (Figure 2a) whereas no hybridization was observed on blots of total RNA from flower, pollen, apex and leaves, and only a very weak hybridization was observed with total RNA from seedlings and cultured tobacco cells (Figure 2a). The discrepancy between the size of NtrbohD cDNA (3338 bp) and the apparent size of its transcript suggests that parts of the 5′- and/or 3′-untranslated regions from the cDNA are missing.

Figure 2.

Expression of the NtrbohD gene in different tissues of tobacco and during elicitation by cryptogein.

Total RNAs (10 µg) were analysed by Northern blot, probed with a 32P-labelled restriction fragment corresponding to the 3′ coding and UTR region of NtrbohD. The autoradiograph of the hybridized blot is shown (a) above the corresponding gels stained with ethidium bromide (b). Total RNAs were extracted from the indicated tissues. L1, L11, L16: 1st, 11th, 16th leaves from the top of mature tobacco plants; Flower: floral tissue without stamen, pistil, stigma, anther and pollen; rRNA: ribosomal RNA from L1 leaves without poly A+; L11 + cry: L11 leaves treated by cryptogein during 0, 7, 24, 28, 29, and 30 h (10 µl of 0.1 mg ml−1 cryptogein was applied to the petiole of excised tobacco leaves), necroses of the HR type (hypersensitive like) appeared after 11 h; T0 to T9: tobacco cells (var. Xanthi) treated with 20 nm of cryptogein during 0, 10, 20, 30, 40, 50, 60, 110, 190 and 240 min and which produce an oxidative burst and an alkalinization of the extracellular medium (C); C, T, C+stau and T+stau: control and treated tobacco cells with 10 nm of cryptogein during 30 min (at the maximal rate of H2O2 production with Xanthi cells) without and after addition of 2 µm staurosporine (stau).

Tobacco cells (N. tabacum cv. Xanthi) treated with cryptogein (20 nm), exhibited a rapid and intense production of AOS and an extracellular alkalinization (Figure 2c). Blots of total RNA from cells sampled at various times after elicitation, showed a strong accumulation of the NtrbohD transcript, which began a few minutes after cryptogein addition (at the onset of AOS production) and seemed to decrease after 1 h, concomitantly with the decrease of AOS (Figure 2a). When the protein kinase inhibitor staurosporine was added to the cell suspension 2 min before cryptogein treatment, the induction of the NtrbohD transcript after a 90-min treatment period was no longer observed (Figure 2a). When excised tobacco leaves were treated with 1 µg cryptogein, the induction of the NtrbohD transcript started after 7 h and increased until 29 h (Figure 2a), the hypersensitive-like necrosis appearing 11 h after the treatment.

Function of NtrbohD

The function of NtrbohD was examined by using transgenic BY2 tobacco cells (N. tabacum cv. Bright Yellow 2) transformed either with an antisense construct of NtrbohD cDNA (‘gp’ cells) or with the empty vector (pky cells). Ten independent transgenic calli were obtained and four cell lines were subcultured.

We investigated the expression of the native NtrbohD gene and of the antisense transgene in these lines by Northern-blot analysis (Figure 3). In the pkyC control cells (transformed with the empty vector), no hybridization with a NtrbohD cDNA probe was observed, whereas a strong induction occurred in the pkyT cells treated with 50 nm cryptogein at the maximal rate of AOS production (Figure 3), as previously described for the wild type cells. Conversely, no hybridization with a NtrbohD cDNA probe was observed, either with control or treated gp3C/gp3T transgenic cells (Figure 3). In this particular transgenic line, mRNA from the NtrbohD native gene and its antisense transgene were no longer accumulated, probably due to a post-transcriptional inactivation process related to RNA interference. This process known as co-suppression for sense gene and transgene occurs very often in transgenic plant tissues (Elmayan and Vaucheret, 1996; Mourrain et al., 2000). The untreated NtrbohD antisense transgenic cell lines gp1, 2, and 4 accumulated a transcript most probably corresponding to the antisense transcript since the native gene is not expressed in untreated cells. The range of separation of the gel electrophoresis used for Northern analysis do not allow to distinguish accurately between the sense transcript of 3.8 kb and the antisense transcript with an expected size of 3.33 kb (corresponding to the 2.63 kb fragment of the NtrbohD coding region introduced in antisense in the transgene construct and to the 0.7 kb of the rbcS terminator region) (Figure 3). However, the similar hybridization observed in untreated and in elicited cells suggests that the sense mRNA is not (or very weakly) induced after elicitation. In these transgenic lines exhibiting a strong antisense transgene expression, the antisense mRNA products may interfere with the translation of the sense native mRNA by a real antisense mechanism of pairing between the two complementary mRNA without necessary (or obvious) degradation of the RNA. In these three independent transgenic lines, sense mRNA degradation could be triggered only after induction of the native gene by elicitor and might lead to a less efficient knockout of the NtrbohD gene compared to the gp3 line.

Figure 3.

Expression of the NtrbohD gene in tobacco cells transformed with antisense construct of NtrbohD after elicitation by cryptogein.

Total RNAs were extracted from the transgenic tobacco cell line pky transformed with the empty binary vector pKY and from lines gp1, gp2, gp3 and gp4 transformed with the antisense construct of NtrbohD cDNA. pkyC and gpC: control cell lines; pkyT and gpT: cells treated with 50 nm cryptogein during 80–100 min (which correspond to the maximal rate of AOS production with BY-2 cells). Total RNAs (10 µg) were analysed by Northern blot, probed with a 32P-labelled restriction fragment corresponding to the NtrbohD 3′ coding region and UTR. The autoradiograph of the hybridized blot is shown above the corresponding gels stained with ethidium bromide.

RT-PCR analysis of distinct sense and antisense NtrbohD expression confirms these results (Figure 4). Antisense transcripts of NtrbohD are not detectable in pky line, are very weakly accumulated in the co-suppressed gp3 line and are strongly accumulated in cell lines gp1 and 2, and to a lesser extent in gp4. This accumulation of NtrbohD antisense transcript observed in all the gp lines seems to be decreased following the treatment by cryptogein. Conversely, sense NtrbohD mRNA are strongly accumulated in elicited pkyT line, are not detectable in elicited gp3T line, and are weakly accumulated in elicited gp1T, gp2T and gp4T lines.

Figure 4.

RT-PCR analysis of sense and antisense expression of NtrbohD in transgenic tobacco cells.

RT-PCRs were performed using 1 µg of total RNA extracted from the transgenic tobacco cell line pky transformed with the empty binary vector pKY and from lines gp1, gp2, gp3 and gp4 transformed with the antisense construct of NtrbohD cDNA. pkyC and gpC: control cell lines; pkyT and gpT: cells treated with 50 nm cryptogein during 80–100 min. Sense NtrbohD specific fragment of 2.96 kb was amplified using primers complementary to the 5′- and 3′-UTR region of the endogene NtrbohD, since NtrbohD fragment used for antisense construct is truncated in the 5′-coding region. Truncated antisense NtrbohD specific fragment of 2,2 kb was amplified using one primer complementary to the coding region between 2721 and 2740 bp after the start codon of NtrbohD and one primer complementary to the 3′ UTR of the rbcS terminator present in the transgene construct and not in the sense endogene NtrbohD. The gene for GAPDH was used as a constitutively expressed RT-PCR control. The number of PCR cycles used was 30 for GAPDH and antisense transgene NtrbohD and 35 for sense endogene NtrbohD. Since different numbers of PCR cycles were used for RT-PCR analysis of sense and antisense NtrbohD expression, the results do not reflect the relative abundances of these transcripts in the tissues tested. Ladder: 1 kb DNA ladder (BRL).

The production of H2O2 following a cryptogein treatment was investigated in these transgenic cell lines. Upon elicitation with 50 nm cryptogein, while pky cell lines (transformed with the empty vector) exhibited an important production of H2O2, the NtrbohD antisense cell lines gp1, gp2, gp4 accumulated only 2–6% of this amount of AOS and no hydrogen peroxide could be detected in the elicited cell line gp3 (Figure 5) even after several hours of treatment (data not shown). Moreover, the residual oxidase activity observed in the elicited cell lines gp1T, gp2T and gp4T, as well as the AOS production in elicited pkyT cells, was totally inhibited by 5 µm of the flavocytochrome inhibitor, DPI (Figure 5). We checked for each transgenic line that the peroxidase activity of cells was not affected by the transformation, so that the discrepancies observed in AOS accumulation are due to a modification in their production and not in their detection (data not shown).

Figure 5.

H2O2 production in transgenic tobacco cells transformed with the empty vector and the antisense construct of NtrbohD.

Transgenic lines of BY2 tobacco cells were treated with 50 nm of cryptogein. When indicated, 5 µm DPI was added 5 min before cryptogein treatment. Every 10 min, H2O2 production was measured. Results represent the mean of three independent experiments. pky: cell line transformed with the empty binary vector pKY; gp1, gp2, gp3, gp4: independent cell lines transformed with the antisense construct of NtrbohD cDNA; C: control cells; T: treated cells by cryptogein. Total H2O2 production measured during 100 min of treatment by cryptogein was summed and was expressed in percentage of H2O2 production in pkyT cell line.

This demonstrates that NtrbohD is implied in AOS production during the elicitation process since the level of AOS production is correlated with the level of the sense NtrbohD mRNA accumulation.

We decided to focus our further studies on the transgenic line gp3, which was the most affected in NtrbohD expression and in AOS production during elicitation.

Extracellular alkalinization and AOS production

The addition of 50 nm cryptogein to the pky cell line transformed with the empty vector, led to an alkalinization of the extracellular medium of 0.7 ± 0.25 pH unit after 100 min of treatment (Figure 6). The same result (an increase of extracellular pH of 0.7 ± 0.1 unit) was obtained when the transgenic line gp3 was elicited with 50 nm cryptogein for 100 min (Figure 6). This demonstrates that the extracellular alkalinization observed in elicited cells is not consecutive to the release of active oxygen species and that the inhibition of NtrbohD oxidase activity does not interfere with the cell pH modification observed during elicitation.

Figure 6.

Extracellular pH changes in transgenic cell lines pky and gp3 transformed, respectively, with the empty vector and the antisense construct of NtrbohD.

Transgenic lines of BY2 tobacco cells were treated with 50 nm of cryptogein. Every 10 min, extracellular pH changes were measured concomitantly with AOS production measurement (Figure 5). Results represent the mean of three independent experiments. pky: cell line transformed with the empty binary vector pKY; gp3: cell line 3 transformed with the antisense construct of NtrbohD cDNA; C: control cells; T: cells treated with cryptogein. The extracellular ΔpH measured after 100 min of treatment by cryptogein was expressed in percentage of ΔpH obtained with pkyT cells (0.7 ± 0.25 unit pH).

NtrbohD protein accumulation and subcellular localization

Tobacco cells transformed either with an antisense construct of NtrbohD cDNA (gp3 line) or with the vector alone (pky line) were treated with 50 nm cryptogein. At a time corresponding to the maximal rate of AOS production, different subcellular fractions corresponding to microsomes, plasma membranes and soluble proteins were isolated and analysed by western-blotting with an antibody raised against amino acids 138–152 and 784–798 of the NtrbohD protein. A band of approximately 105 kDa was weakly detected in the microsomal fraction and strongly revealed in the plasma membranes of elicited pkyT cells, whereas no signal was visible in the soluble fraction of the same cells (Figure 7). Conversely, no band was immunodetected in any fraction isolated from elicited gp3T cells transformed with antisense constructs of NtrbohD cDNA (Figure 7). This is in agreement with the results obtained by the Northern blotting and RT-PCR experiments presented in Figures 3 and 4, since this particular line is no longer able to accumulate sense mRNA of NtrbohD. Moreover, this demonstrates that NtrbohD is an intrinsic plasma membrane protein.

Figure 7.

Accumulation, size and cellular localisation of NtrbohD protein in elicited transgenic cell lines pky and gp3.

Proteins were extracted from the transgenic elicited cell lines pkyT and gp3T after 80–100 min of treatment by 50 nm cryptogein, at the maximal rate of AOS production, fractionated and analysed by Western blot with antibodies directed against 2 peptides of the NtrbohD protein as described in Methods. C: cytosolic fraction; M: microsomal fraction; PM: plasma membrane.

Discussion

Cloning and expression of NtrbohD in tobacco cells and leaves

Studies of Torres et al. (1998) indicated that Arabidopsis thaliana carries at least six gp91–phox homologues. As our purpose was to test the involvement of a gp91–phox homologue in AOS production during the elicitation process, we screened a cDNA library of elicited tobacco leaves. This strategy resulted in the isolation of only one cDNA in tobacco. As studies on A. thaliana indicated that rboh transcripts were not abundant in healthy tissues, this peculiarity of our library probably led to an over-representation of a particular clone involved in plant defence mechanisms. Further analyses are required to determine the number of rboh genes present in tobacco. It has already been established (Langebartels et al., 2001) that at least two isoforms of rboh are present in tobacco plants but there may be others whose expression patterns could be investigated. The analysis by Northern blots of the expression of NtrbohD revealed the presence of an abundant transcript in roots, but did not allow its detection in flowers, pollen, apex or healthy leaves. This pattern of expression is similar to one previously observed with AtrbohD (Keller et al., 1998; Torres et al., 1998), the gene of A. thaliana which presents the strongest homology with NtrbohD. Interestingly, the treatment of either tobacco leaves or cells with the fungal elicitor cryptogein, led to an important accumulation of NtrbohD transcripts: in leaves this induction occurs before the appearance of necrosis and, in suspension cells, this induction coincides with the onset of AOS production and tends to decrease concomitantly with the level of AOS detected. Such an accumulation of AtrbohD transcripts upon elicitation of Arabidopsis cells has already been reported (Desikan et al., 1998). The potato StrbohB gene related to AtrbohD is also induced during elicitation and this induction is kinase dependent (Yoshioka et al., 2001) as it seems to be the case for NtrbohD since its induction by cryptogein is sensitive to the protein kinase inhibitor staurosporine (Figure 2a). This pattern of expression and the homology with the neutrophil flavocytochrome gp91–phox is in favour of the involvement of NtrbohD in AOS production occurring during elicitation.

Demonstration of the function of NtrbohD

A clear demonstration of this was obtained with transgenic tobacco cells transformed with an antisense construct of NtrbohD in which the absence of induction of native sense gene leads to an important inhibition of H2O2 production. In one of these lines, gp3, presenting a total inactivation of the NtrbohD gene expression, neither native NtrbohD transcripts, nor NtrbohD protein nor AOS production could be detected following cryptogein treatment. This is, to our knowledge, the first report of a mechanism of gene silencing occurring in cultured plant cells. It is interesting to note that this silenced state is maintained in the gp3 line even after 40 weeks of subculture. Although it has been evidenced that other systems such as oxalate oxidase (Zhou et al., 1998) or cell-wall bound peroxidases (Martinez et al., 1998) could participate in H2O2 production during some plant-microorganisms interactions, an homologue of the neutrophil flavocytochrome is entirely responsible for the production of hydrogen peroxide consecutive to the treatment of tobacco cells with cryptogein.

It is worth noting that the alkalinization of the extracellular medium, typically observed after cryptogein treatment, was not affected by the inactivation of the NtrbohD gene expression. This indicates that this inactivation affects the AOS production in a specific manner and confirms that the increase of the extracellular pH does not result from the dismutation of superoxide anions generated by the flavocytochrome, as previously proposed (Simon-Plas et al., 1997). The use of an antibody directed against the NtrbohD protein allowed us to establish without ambiguity the location of the protein on the plasma membrane of elicited tobacco cells, in accordance with the plasma membrane localization of RbohA (Keller et al., 1998). The fact that NtrbohD is an intrinsic plasma membrane protein is consistent with the extracellular production of superoxide anions, inhibited by DPI, detected in tobacco cells after a cryptogein treatment (J-L Montillet, personal communication).

Thus, it can reasonably be assumed that NtrbohD shares with gp91–phox the function of triggering an oxidative burst in the context of a defence reaction. This is to our knowledge the first demonstration of a direct implication of a homologue of gp91–phox in the plant oxidative burst occurring during elicitation of a plant cell suspension. This opens a very exciting field of investigations concerning the exact number and the precise roles of the different isoforms of Ntrboh in the physiology of the plant. Regarding this later point, the very recent work of Torres et al. (2002) is of particular interest. They demonstrated, using mutants in Atrboh D and F genes, that both are required for full AOS production during incompatible interactions, but also that each of them has a different role in the defence of the plant. Indeed, atrbohD mutation eliminated the majority of total AOS production, whereas atrbohF mutation exhibited the strongest effect on cell death. Moreover, this work emphasizes the complexity of the plant defence mechanisms and provide essential clues concerning the relationship between HR and AOS, SAR and HR or SAR and AOS.

Conclusions

NtrbohD shares with some gp91–phox plant homologues previously cloned in rice (Groom et al., 1996), Arabidopsis (Desikan et al., 1998; Keller et al., 1998; Torres et al., 1998), tomato (Amicucci et al., 1999) and potato (Yoshioka et al., 2001), some characteristic features of the neutrophil flavocytochrome, but also has a hydrophilic N-terminal extension with a Ca2+-binding motif, typical of plant gp91–phox homologues (Amicucci et al., 1999; Keller et al., 1998; Torres et al., 1998). If some characteristic features are conserved between these plant and animal proteins, the similarity of their regulation remains questionable. Xing et al. (1997) reported that treatment of tomato cells with race-specific elicitors induced the translocation of proteins detected, respectively, by antip67phox, antip47phox and anti-Rac2 antibodies, from the cytosol to the plasma membrane. Desikan et al. (1996) detected proteins in Arabidopsis cells immunologically related to p47phox and p67phox and showed that human neutrophil components were able to co-operate with Arabidopsis membranes to weakly generate superoxide. However, several lines of evidence may question the conclusions raised from these studies: (i) to date, no direct evidence for sequences presenting significant homology with p22phox, p47phox or p67phox exists in the entire Arabidopsis genome database; (ii) attempts to detect clones of putative subunits of a plant NADPH oxidase by expression screening using antibodies directed against subunits of the neutrophil oxidase led to the isolation of unrelated proteins (Tenhaken and Rubel, 1998; Kieffer and Elmayan, unpublished results); (iii) the sites known in neutrophil gp91–phox to interact with p47phox (indicated in Figure 1) are not well conserved in AtrbohD and NtrbohD; and (iv) use of mammal p47phox and p67phox as bait in an heterologous 2-hybrid screen with a tobacco cDNA library did not lead to the isolation of any tobacco clones able to interact with these subunits of neutrophil NADPH oxidase (Kieffer and Elmayan, unpublished results). Moreover, a novel flavoprotein involved in the thyroid NADPH oxidase (p138Tox) has been recently purified (Dupuy et al., 1999). Its sequence contained an extended N-terminal domain comprising two EF-hand motifs and three of the four sequences implicated in the interaction of gp91phox with p47phox were missing. The authors suggested that this flavoprotein could constitute the sole component of the thyroid NADPH oxidase. Very recent experiments describing an in-gel production of superoxide by tomato proteins cross-reacting with an antiserum directed against the tomato rboh (Sagi and Fluhr, 2001) also indicate that the plant oxidase could produce O2− . in the absence of additional cytosolic components. This is in favour of a distinct regulation of the plant and animal oxidases.

In this context we are currently performing a 2-hybrid screen using the N-terminal and C-terminal part of NtrbohD in order to find the specific partners of this plant oxidase, acting in the signal transduction cascade leading to AOS production.

Experimental procedures

Plant material

Tobacco plants and cell suspensions (Nicotiana tabacum cv. Xanthi) were grown as previously described (Milat et al., 1991; Simon-Plas et al., 1997). Tobacco BY2 cells (N. tabacum cv. Bright Yellow 2) were grown in MS medium, pH 5.6, containing MS salt (Murashige and Skoog, 1962), 1 mg l−1 thiamine-HCl, 0.2 mg l−1 2,4 dichlorophenylacetic acid, 100 mg l−1 mio-inositol, 30 g l−1 saccharose, 200 mg l−1 KH2PO4 and 2 g l−1 MES. Cells were maintained by weekly dilution (2 : 80) into fresh medium. Cryptogein was purified and treatments were performed as previously described (Milat et al., 1991; Simon-Plas et al., 1997). Cryptogein (10 µl of 0.1 mg ml−1) was applied to the petiole of excised tobacco leaves (L11: 11th leaf from the top of a mature plant).

Tobacco gp91–phox homologue cloning strategy

Known amino acid sequences of gp91–phox from human (p04839) and homologues from rice (×93301) and A. thaliana (AF055357) were used to identify consensus regions and design degenerate primers subsequently used for RT-PCR. Primer sequences for PCR were: forward on amino acid consensus sequence PVCRNT, 5′-CCNGTNTGYMGNAAYAC-3′; reverse on amino acid consensus sequence IGVFYCG, 5′CCRCARTARAANACNCCDAT-3′. Primer sequences for nested PCR were: nested forward on amino acid consensus sequence FWYSHH, 5′-GCNTGGTAYMSNCAYCA-3′; nested reverse on amino acid consensus sequence FARPNW, 5′-CCARTTNGGNCKNGCRAA-3′. PCR and nested PCR were performed on cDNA obtained by reverse transcription of mRNA extracted from young tobacco leaves with an annealing temperature of 55°C. A PCR product of the predicted size of 1 kb was cloned into pGEMTeasy (Promega Corporation, Madison, WI, USA) and sequenced. This 1 kb fragment homologue of gp91–phox was used as a probe to screen a tobacco cDNA library made from tobacco leaves elicited with Pseudomonas solanacearum (Czernic et al., 1996; a gift of Dr Y. Marco) in the bacteriophage lambda zap Express (Stratagene Europe, Amsterdam, The Netherlands). The 3 positive clones inserted into pBK-CMV phagemid after excision share the same nucleotide sequence and the largest clone of 3338 bp (AJ309006) encodes a full-length protein of 939 amino acids.

DNA sequencing and analysis

Both strands of NtrbohD cDNA were sequenced by Genome Express Corporation (Meylan, France). DNA sequence data were analysed using the GCG package (University of Wisconsin Genetics Computer Group, Madison, WI, USA), and the deduced amino acid sequences were compared with sequences found in the current databases (GenBank, Swiss Protein, EMBL, etc.) using the Blast network service of the National Center for Biotechnology Information (NCBI), and the programs Bestfit, Pileup and Pretty (for alignment and consensus).

RNA isolation and blot hybridization

Total RNA from tobacco cells (BY2 and Xanthi) and from different tissues of mature tobacco plants and seedlings were extracted and purified, followed by a CsCl gradient according to Sambrook et al. (1989). Samples of total RNA (10 µg) were analysed by electrophoresis in 8% formaldehyde, 1.5% agarose gels, and transferred to Pall biodyne B membranes according to the manufacturer's instructions. A fragment of 0.86 kb corresponding to part of the NtrbohD 3′ coding and untranslating region, was labelled to 108 c.p.m. µg−1 using random priming and 32P-dCTP (Amersham Pharmacia Biotech, Orsay, France). Labelling, hybridization and washes were performed essentially as described by Sambrook et al. (1989).

Antisense gene construction and cell transformation

An SstI/XhoI fragment of 2633 bp from NtrbohD cDNA, truncated in the 5′-coding region, was inserted in antisense into pKYLX71–35S2, a plant transformation vector allowing inserts to be cloned between a 35S promoter with a duplicated enhancer and rbcS terminator region of 700 bp (Elmayan and Tepfer, 1994).

The resulting plasmids, introduced by triparental mating into a disarmed strain Agrobacterium tumefaciens, C58C1 (pMP90) (Koncz and Schell, 1986), were used to transform BY2 cells. Two ml of a 3-day-old BY2 culture were co-cultivated with 50 µl of each Agrobacterium culture (OD600 0.3) in Petri dishes in the dark for 2–3 days at 26°C. Cells were washed three times in 10 ml of fresh medium by centrifugation at 50 g, 3 min, and plated onto agar-MS medium containing 100 mg l−1 kanamycin and 500 mg l−1 cefotaxime. Transformed microcalli were propagated during 4–5 subcultures in 10 ml MS liquid medium containing the selection agent. These subcultures were then diluted weekly (2 ml in 80 ml of MS liquid medium without the selection agent). Ten independent transgenic calli were obtained and four cell lines were sub cultured.

RT-PCR expression analysis

To analyse the expression of sense and antisense messenger of NtrbohD endogene and transgene RT-PCR method was employed. As ubiquitously expressed marker genes the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) was chosen. The degenerate primers GAPDHf (5′-ATTARGATCGGAATYAACGG-3′) and GAPDHr (5′-GTAACCCCAYTCRTTGTCRTA-3′) were used to amplify a GAPDH fragment of 950 bp between the conserved motifs I(R/K)IGING and YDNEWGY, respectively.

The sense NtrbohD specific fragment of 2.96 kb was amplified using primers complementary to the 5′- and 3′-UTR region of the endogene NtrbohD (since NtrbohD fragment used for antisense construct is truncated in the 5′-coding region); 5′gp91: 5′-TTCAAGAATTCGGGTTCC-3′; 3′gp91: 5′-GCTCTGCTTAATGG TCT-3′.

The truncated antisense NtrbohD specific fragment of 2.2 kb was amplified using one primer (T7-1) complementary to the coding region between 2721 and 2740 bp after the start codon of NtrbohD and one primer (rbcS) complementary to the 3′ UTR of the rbcS terminator present in the transgene construct and not in the sense endogene NtrbohD; T7-1: 5′-CCCAATAGAAAT ATGCTCTCC-3′; rbcS: 5′-CTCTTCTCCATCCATTTCC-3′.

First strand cDNAs were made from 1 µg total RNA extracted from tobacco cells as described above, using first strand cDNA synthesis kit (Roche Diagnostics, Mannheim, Germany), according to the manufacturers' protocol. PCRs were performed using 1 µl of each cDNA diluted one in four. The PCR thermocycle profile was 94°C 1 min (initial denaturation), followed by 30 cycles of 94°C for 30 sec, 56°C for 30 sec and 72°C for 2 min with primers GAPDHf/GAPDHr and rbcS/T7-1, or 35 cycles of 94°C for 30 sec, 58°C for 30 sec and 72°C for 2 min with primers 5′gp91/3′gp91, and one cycle at 72°C for 10 min (for final elongation). PCR products were separated on 1.2% agarose gels.

AOS production and extracellular pH modification

Cells were harvested 6 days after subculture, filtered, resuspended (1 g for 10 ml) in a 2-mm MES buffer pH 5.90, containing 175 mm mannitol, 0.5 mm CaCl2 and 0.5 mm K2SO4. After a 3-h equilibration period on a rotary shaker (150 r.p.m) at 25°C, cells were treated with cryptogein and/or chemicals as indicated in the legends of the figures. The production of H2O2 was measured by chemiluminescence using luminol and a luminometer (BCL book, Berthold, la Garenne Colombe, France). Every 10 min, a 250-µl aliquot of the cell suspension was added to 50 µl of 0.3 mm luminol and 300 µl of the assay buffer (175 mm mannitol, 0.5 mMCaCl2, 0.5 mm K2SO4 and 50 mm MES pH 6.5). Extracellular pH modifications were monitored using a Radiometer pH meter.

Cell fractionation

As described above, changes in pH and AOS production were monitored on tobacco cell suspension elicited or not with cryptogein. At the time corresponding to the maximal rate of AOS production, cells were collected by filtration, frozen in liquid N2 and homogenized in grinding medium (50 mm Tris-MES pH 8.0, 500 mm sucrose, 20 mm EDTA, 10 mm DTT, 1 mm PMSF). The homogenate was centrifuged at 16 000 g for 20 min. After centrifugation, the supernatants were collected, filtered through two successive screens (63 and 38 µm) and centrifuged at 96 000 g for 35 min. An aliquot of the resulting membrane pellet (microsomal fraction) was resuspended in the storage buffer (10 mm Tris-MES pH 7.3, 250 mm sucrose, 1 mm EDTA, 10 mm DTT, 1 mm PMSF, 20% glycerol) and stored at − 80°C, whereas the rest of this microsomal fraction was purified by partitioning in an aqueous 2-phase system to obtain the plasma membrane fraction (Larsson et al., 1994). The proteins present in the supernatant of the microsomal fraction were recovered to obtain the cytosolic fraction by an overnight precipitation in 40% of a saturated ammonium sulphate solution, and a subsequent centrifugation at 30 000 g for 30 min.

Western-blotting experiments

A polyclonal rabbit serum raised against amino acids 138–152 (CLNKRPIPTGRFDRNK) and 784–798 (IAKNKGNKSGSASGGC) of the NtrbohD protein, coupled to KLH, was obtained from EUROGENTEC (4102 Seraing-Belgium). IgG were then purified according to Hardie and van Regenmortel (1977) and resuspended in distilled water at a concentration of 2 mg ml−1. Samples of 25 µg protein of each subcellular fraction were diluted with one volume of Laemmli buffer 2× and loaded on a 8% SDS polyacrylamide gel. After electrophoresis separation (1 h, 40 mA) protein fractions were electroblotted onto nitrocellulose membrane (20 min, 15 V, Trans-Blot SD, semi-dry transfer cell, Bio-Rad, Hercules, CA, USA). Probing and detection of Western-blots were performed as described in the ECL Western-Blotting detection kit (Amersham). Primary antibodies were used at a dilution of 1 : 1000 in TBS-Tween (2 mm Tris, 15 mm NaCl, Tween 20 00.5%, pH 7.6). The horseradish peroxidase antirabbit IgG antibody (Bio-Rad) was used at a 1 : 25 000 dilution in TBS-Tween. The revelation was performed using an exposition time of 15 sec.

Acknowledgements

This work was supported by a grant from the Conseil Régional de Bourgogne. We thank Dr Yves Marco for the gift of the tobacco cDNA library, Dr Marie-Claire Dagher and Dr Marie-Josephine Farmer for fruitful discussion and critical reading of the manuscript.

Accession number: AJ309006

Ancillary