U. M. Pfitzner, Universität Hohenheim, Institut für Genetik, FG Allgemeine Virologie, Emil-Wolff-Str. 14, D-70599 Stuttgart, Germany. Fax: + 49 711459 2937, Tel.: + 49 711459 2395, E-mail: email@example.com
Tobacco pathogenesis-related protein 1a (PR-1a) is induced in plants during the hypersensitive response (HR) after exposure of plants to salicylic acid (SA) and by developmental cues. Gene activation by these diverse stimuli is mediated via an as-1-like element in the PR-1a upstream region. To further analyze the significance of this cis-acting sequence, an authentic as-1 element from the cauliflower mosaic virus 35S RNA promoter was inserted into the PR-1a promoter in place of the as-1-like motif. Reporter gene analysis in transgenic tobacco plants demonstrated that as-1 can functionally replace the as-1-like element in the PR-1a promoter in response to all stimuli. However, reporter gene induction from the as-1 carrying promoter was enhanced in response to SA compared to the wild-type promoter, and the ratio of reporter gene activities in SA treated leaf tissue to tissue exhibiting the HR increased with the as-1 promoter construct. Our findings support a model where PR-1a gene expression relies on at least two distinct signal transduction pathways initiated by SA and by a yet unknown signal produced during the HR, that promote different, albeit related, transcription complexes on the PR-1a as-1-like element. Analysis of PR-1 proteins in plants expressing salicylate hydroxylase yielded additional evidence that an HR dependent pathway leads to high level PR-1 gene induction in tobacco.
Infection of plants with pathogens usually results in a distinct host response depending on the genetic constitution of the plant and the pathogen. In an incompatible interaction, the pathogen remains localized at the primary infection sites that often are visible as necrotic local lesions on the leaves. This local defense reaction is referred to as the hypersensitive response (HR). Subsequently, the HR triggers a general resistance mechanism rendering uninfected parts of the plant less sensitive to further attack by pathogens, a phenomenon called systemic acquired resistance (SAR) . The elicitation of the HR and SAR reactions is accompanied by the coordinated induction of a heterogeneous group of proteins in the infected and uninfected leaves, commonly referred to as pathogenesis-related (PR) proteins.
PR proteins were first described in tobacco plants infected with Tobacco mosaic virus (TMV) exhibiting the HR [2,3]. By now, related proteins have been identified across the plant kingdom in both dicotyledonous and monocotyledonous species. In tobacco, seven families of PR proteins are known . Yet, the biological functions of PR proteins are not clear. It is intriguing that PR proteins are also expressed in substantial amounts in healthy plants upon the transition to flowering [5–7], suggesting that they play a role during plant development.
Although not defined clearly by their function, the expression of PR genes has served as a reliable marker for the induction of SAR [8–11]. Therefore, PR genes are also referred to as SAR genes . To identify components of the complex defense signaling pathways, several groups have studied the regulation of PR gene expression. Initially, PR-1 proteins were found to be inducible to high levels by the exogenous application of salicylic acid (SA) in healthy tobacco plants . Consistent with this finding, tissue levels of SA have been demonstrated to increase significantly locally and systemically in plants displaying the HR [13–15]. Furthermore, plants that are inhibited in SA accumulation show defects in SAR and in the expression of PR genes [16–19]. It was thus hypothesized that SA is an endogenous regulator of pathogen resistance and of PR gene expression . Subsequently, some PR genes were found to be responsive to agents other than SA. From these findings and from the study of Arabidopsis mutants exhibiting aberrant SAR expression patterns, it appears that different pathways encompassing SA or ethylene/jasmonic acid as signal molecules can lead to the induction of PR genes and SAR in plants infected by pathogens .
To dissect the functional architectures of PR gene promoters responsive to different signal molecules, reporter gene constructs were generated and tested for regulated gene expression. In addition, in vivo and in vitro studies were performed to unravel interactions of nuclear proteins with DNA sequences in the promoter regions of PR genes. By these means, diverse sequence elements and their cognate binding proteins have been identified in accordance with the view that PR genes are regulated differently. In the −906 bp promoter region of the tobacco gene encoding the acidic PR-1a protein, a duplicated TGACG motif has been shown to control reporter gene expression in transgenic plants in response to SA, to the HR, and to developmental stimuli . This element is referred to as as-1-like motif because of its relatedness to the as-1 element. The as-1 element was originally identified in the 35S RNA promoter from cauliflower mosaic virus (CaMV) , and has been shown to mediate a moderate induction of the 35S RNA promoter by SA . Transcription factors belonging to the TGA family of basic leucine zipper proteins interact in vitro with as-1[23,25], and, similarly, with the PR-1a as-1-like motif . Likewise, a related cis-acting element involved in SA responsive reporter gene expression has been identified in the homologous Arabidopsis PR-1 promoter by functional analyses in transgenic plants . The Arabidopsis cis-acting element comprises the sequence TGACG, which is directly repeated only 12 bp upstream of the identified promoter element. TGA transcription factors have been shown to bind to this motif in vitro[27–29], and an inducible in vivo footprint has revealed significant changes in DNA accessibility to the identified element upon SAR induction . In addition, overexpression of trans-dominant TGA mutant proteins, which are no longer able to bind to their target sequences, has resulted in reduced accumulation of chemically inducible PR-1 mRNAs in transgenic tobacco and Arabidopsis plants [30,31]. Thus, as-1-like elements and TGA transcription factors seem to be intimately connected with the expression of PR-1 genes in both tobacco and Arabidopsis. This association is even more supported by the finding that TGA factors from Arabidopsis, tobacco and rice have been shown to interact in vivo and in vitro with NPR1/NIM1 [27–29,32,33]. NPR1/NIM1 has been identified as a key regulator of SAR in Arabidopsis acting downstream of SA in the SAR signaling pathway [10,11,21], and analysis of npr1 mutant plants has established that efficient expression of the Arabidopsis PR-1 gene after SAR induction relies on a functional NPR1/NIM1 gene [10,11]. Furthermore, it has been demonstrated recently that Arabidopsis TGA2 acts as a transcriptional activator in transgenic plants in response to SA, and that this activity is abolished in the npr1 mutant . Taken together, substantial evidence has accumulated suggesting a prominent role for as-1-type elements and TGA transcription factors in the regulated expression of some PR genes.
Here we report that the level of SA inducible gene expression is controlled by the as-1-like element in the strong −1533 bp PR-1a promoter from tobacco. To further analyze the functional significance of as-1-like elements and their binding factors in the induction of PR genes, we have modified the PR-1a promoter to contain an authentic as-1 element. As-1 can functionally replace the as-1-like element in the PR-1a promoter. Substitution by as-1 enhances SA inducible reporter gene expression from the promoter by a factor of 3 in transgenic plants, and the ratio of inducible reporter gene expression from the as-1 containing promoter in SA treated tissue to necrotic tissue infected with TMV is increased compared to the wild-type promoter. These findings emphasize that as-1 related elements and their binding factors play crucial roles in the regulation of the tobacco PR-1a gene. Furthermore, our results demonstrate that the properties of the PR-1a promoter are subject to change by insertion of a variant as-1 element. The data are in accord with the conclusion that the tobacco PR-1a promoter is targeted by at least two distinct signal transduction pathways, elicited by SA and by a yet unknown signal produced during the HR, respectively, which mediate PR-1a gene activation through a common DNA sequence, the as-1-like element.
Construction of plasmids containing reporter genes
Recombinant DNA techniques were performed according to standard procedures .
For the construction of the PR-1 reporter gene, a 0.27 kb StyI/PstI fragment from λW38-1, comprising parts of the 5′ region and the open reading frame encoded by the W38-1 gene , was inserted into StyI/PstI cleaved pT5S/HΔP containing the PR-1a gene. The resulting plasmid was linearized with EcoRV, which cuts in the PR-1a 3′ untranslated region 56 bp 3′ to the stop codon, and a SacI linker was added to the blunt ends. The W38-1::PR-1a chimeric gene was excised from the plasmid as StyI/SacI fragment and inserted into p-1533PR1a[GUS], from which the GUS reporter gene had been removed as a StyI/SacI fragment.
The −1533as-1m4[GUS] construct was generated by the addition of the 0.64 kb HindIII/XhoI fragment from p-1533PR1a[GUS] to the 5′ end of p-906as-1m4[GUS]. The as-1m4 mutation replaces the as-1-like element occuring from positions −592 to −577 in the PR-1a promoter.
To obtain construct −906as-1[GUS], site-directed mutagenesis was employed on p-906PR1a encompassing the PR-1a sequence from −906 to +28. The DNA was primed with PR-46 (5′-TGACGTTAACTAACTAT-3′) containing an A to C nucleotide substitution with respect to the wild-type promoter (underlined). By this procedure, a unique HpaI restriction enzyme site (bold) was introduced in p-906PR1a at position −597 just upstream of the as-1-like motif (from −592 to −577). The mutant plasmid was digested with HpaI and BamHI and ligated to a 0.6 kb HpaI/BamHI fragment obtained from p-906PR1a by PCR amplification with a pUC universal primer and PR-47 (5′-CAAGCTTGTTAACTGACGTAAGGGATGACGGC CATGTTCAAGTTT-3′), which contains an authentic as-1 element as found in the CaMV 35S RNA promoter (in bold letters). The as-1 element was inserted at position −591 in the PR-1a promoter. The as-1 containing −906 bp PR-1a promoter region was added as HindIII/BamHI fragment to p0[GUS] to give plasmid −906as-1[GUS]. To obtain p-1533as-1[GUS], the 0.64 kb HindIII/XhoI fragment from p-1533PR1a[GUS] was ligated to HindIII/XhoI cleaved p-906as-1[GUS].
All plasmids generated by site-directed mutagenesis or PCR amplification of fragments were verified by DNA sequence analysis.
Construction of an expression vector containing the nahG gene from Pseudomonas putida
P. putida PpG7 containing plasmid NAH7 (DSM 4476) was obtained from the German Collection of Microorganisms and Cell Cultures, Braunschweig, and grown in mineral medium with the addition of 0.05% sodium salicylate according to the instructions of the supplier. In order to isolate the nahG gene, two primers were designed. Primer nahG-1 (5′-TTCCCGGGGCCATCACGAGTACAGCATGA-3′) is identical to the 5′ end of the nahG gene including the ATG translation start codon (bold) with an additional SmaI restriction enzyme site at its 5′ end. Primer nahG-2 (5′-AAGAGCTCGTCAACGTTGTCACCCTTG-3′) is complementary to the 3′ end of the nahG gene including the translation stop codon (bold) and carries a SacI restriction enzyme site at its 5′ end. The primers were used in standard PCR amplifications on plasmid DNA isolated from P. putida by the alkaline lysis procedure. A single reaction product of 1.3 kb was detected on ethidium bromide stained agarose gels. The PCR product was subcloned into pUC19 cleaved with SmaI and SacI and sequenced over its entire length. The nucleotide sequence determined was found to be identical to the sequence of the nahG gene described previously . For fusion of the nahG gene to the CaMV 35S RNA promoter, the 1.3 kb SmaI/SacI fragment was integrated into pBin19/35S[GUS] from which the GUS reporter gene had been removed. The resulting plasmid was designated pBin19/35S[nahG].
Generation and analysis of transgenic plants
Constructs for expression in plants were integrated into the binary vector pBin19, and Agrobacterium-mediated transformation was employed to introduce gene constructs into the genome of tobacco (Nicotiana tabacum L. cv. Samsun NN) as described previously . Primary transformants were allowed to self-fertilize, and progeny plants were selected on medium containing 400 mg·L−1 kanamycin.
For the induction of reporter gene constructs and of the endogenous PR-1 proteins, eight leaf disks of 1 cm diameter were cut from at least two different fully expanded leaves of each transgenic plant. The leaf disks were floated for 72–96 h in a Petri dish on water or on neutral solutions of 1 mm or 5 mm SA as indicated. From plants which were tested for gene induction by application of SA as well as during the HR, leaf disks for chemical treatment were collected from a single leaf which was cut from the plant just prior to virus inoculation. Infection of tobacco plants with TMV has been described previously . Eight leaf disks were collected from necrotic tissue of transgenic plants 5–7 days after virus inoculation. For the analysis of PR-1 gene expression in nahG transgenic plants, two zones of leaf material, a narrow zone including the lesion and a more distant zone, were excised from the leaves around single lesions with two different cork borers of 7 and 14 mm diameter, respectively. Leaf disks were extracted with GUS lysis buffer. Analysis of GUS reporter gene expression and immunodetection of PR-1 proteins was conducted as reported [22,37]. For monitoring the expression of the PR-1 reporter gene, protein extracts were separated by gradient gel electrophoresis. GUS activity is given in units (1 unit = 1 nmol of 4 MU·h−1 per mg protein).
To verify the correct insertion of the as-1[GUS] constructs in transgenic plants exhibiting high GUS reporter enzyme induction, genomic DNAs were isolated from transformants according to Fulton et al. . Standard PCR was performed on the genomic DNAs using PR-19, identical to the PR-1a promoter, and a GUS primer complementary to the 5′ end of the reporter gene. PCR products were cut with HpaI which cleaves only in the as-1 containing promoter regions, but not in the PR-1a wild-type promoter. Reaction products were analyzed on ethidium bromide stained agarose gels.
The −1533 bp PR-1a promoter directs high level induction of a PR-1 reporter protein in transgenic tobacco plants
Functional analysis of the tobacco PR-1a promoter by several groups had demonstrated that the 0.9 kb upstream region of the PR-1a gene is sufficient to yield regulated reporter gene expression in transgenic tobacco plants [7,39,40]. Yet, gene induction levels from the 0.9 kb PR-1a promoter are significantly lower (average fold induction of GUS reporter gene activity < 75) than induction of the endogenous PR-1a gene (≫100-fold). Previously, we have shown that a longer PR-1a upstream region, the −1533 bp promoter, is able to direct higher levels of GUS reporter gene expression in transgenic plants . To assess the strength of the −1533 bp PR-1a promoter region in direct comparison to the expression of the endogenous PR-1 genes, we have fused a PR-1 reporter gene to the −1533 bp promoter and monitored expression of the reporter protein in transgenic tobacco plants.
The PR-1 reporter gene is a chimera constructed in vitro by the replacement of a 0.27 kb StyI/PstI fragment in the PR-1a gene with the respective fragment from the W38-1 pseudogene . The open reading frame of the chimeric PR-1 gene, designated W38-1::PR-1a, is colinear with the open reading frame encoded by the PR-1a gene, and differs by only six amino acids from the PR-1a preprotein in the N-terminal moiety of the protein. Most importantly, the W38-1 pseudogene part of the W38-1::PR-1a chimeric gene contains an amino acid exchange close to the signal peptide cleavage site (Gly vs. Arg in the PR-1a preprotein at positon −2 of the signal peptide), thus giving rise to an alternatively processed mature protein in plants, which can be distinguished through its smaller size from the naturally occurring PR-1a, PR-1b and PR-1c proteins by SDS gel electrophoresis (Fig. 1 and data not shown). When put under control of the CaMV 35S RNA promoter in transgenic tobacco plants, the chimeric W38-1::PR-1a gene is equally well expressed as a similarly constructed 35S[PR-1a] transgene (Fig. 1 and data not shown). Tobacco plants transformed with the −1533PR1a[W38-1::PR-1a] construct were infected with TMV, and protein extracts from independent transformants were monitored prior to and after infection for expression of the endogenous PR-1 proteins and the W38-1::PR-1a reporter protein by immunodetection. As shown in Fig. 1, two bands were observed consistently with extracts from transformants exhibiting the HR after virus infection. The upper, more prominent band corresponds to the endogenous acidic PR-1 proteins, of which the PR-1a protein constitutes approximately 50% (data not shown). The lower band corresponds to the recombinant W38-1::PR-1a protein. Although expressed weaker than the endogenous PR-1a protein (approximately 5-fold less than PR-1a), the recombinant W38-1::PR-1a protein is clearly detected in all extracts containing considerable amounts of the endogenous PR-1 proteins, thus demonstrating that the −1533 bp PR-1a upstream region represents a promoter that encompasses important cis-acting elements for both regulated as well as high level expression of the PR-1a gene.
The as-1-like element controls the level of SA inducible gene expression from the strong −1533 bp PR-1a promoter in transgenic tobacco plants
Previously, we have shown that an as-1-like motif, located from positions −592 to −577 in the PR-1a upstream region, contributes significantly to the level of reporter gene expression from the weaker −906 bp PR-1a promoter . To analyze the significance of the as-1-like motif for expression from the strong −1533 bp PR-1a promoter, we have introduced the as-1m4 mutation, which completely destroys the as-1-like motif , in the −1533 bp PR-1a upstream region to give construct −1533as-1m4[GUS](Fig. 2A). Tobacco plants were transformed in parallel with −1533PR1a[GUS] or −1533as-1m4[GUS], and transformants were monitored for inducible reporter gene expression after floating of leaf disks on H2O or on a solution of 1 mm SA. As seen previously with the −906 bp PR-1a promoter, insertion of the as-1m4 mutation drastically reduced reporter gene expression from the strong −1533 bp promoter region (Fig. 2B). On the other hand, induction of GUS activity by SA was still preserved with the mutant promoter, albeit at a lower level (Fig. 2B), just as observed previously for the −906as-1m4 mutant PR-1a promoter . Thus, the as-1-like motif represents an important cis-acting element controlling the level of expression from both the weaker −906 and the strong −1533 bp PR-1a promoter regions in a similar way.
Replacement of the as-1-like element by as-1 enhances SA inducible and developmentally controlled reporter gene expression from the PR-1a promoter in transgenic tobacco plants
To address the question whether the PR-1a as-1-like motif is a genuine as-1-type element, possibly targeted by the same transcription factor(s) as as-1, we have introduced an authentic as-1 element in place of the as-1-like motif in the −906 and −1533 bp PR-1a promoter regions. The resulting reporter gene constructs were designated −906as-1[GUS] and −1533as-1[GUS], respectively (Fig. 3A). Tobacco plants were transformed in parallel with −906as-1[GUS], −1533as-1[GUS], or the respective wild-type promoter constructs, and transformants were monitored for reporter gene expression after treatment of leaf disks with H2O or a solution of 1 mm SA. As shown in Fig. 3B, inducibility of GUS activity was maintained with any plant containing either of the as-1[GUS] transgenes. Surprisingly, plant extracts from transformants with the as-1[GUS] constructs yielded on average threefold higher GUS activities than extracts from plants containing the respective wild-type promoter constructs (Fig. 3B). To ensure that an as-1 element in the context of the PR-1a promoter is indeed able to mediate higher average reporter gene induction in transformed plants, we have isolated genomic DNAs from four plants each exhibiting the highest GUS activities from the transformations with the −906 and −1533as-1[GUS] constructs. A 1.0-kb fragment was amplified from the isolated DNAs by PCR, and the PCR products were cleaved with HpaI (Fig. 3A). In all cases, the highest GUS activities measured in transformed plants correlated with the presence of a HpaI restriction endonuclease site occurring at position −597 only in the −906as-1 and −1533as-1 sequences, but not in the wild-type promoter regions (data not shown).
Likewise, GUS reporter gene expression was monitored in transgenic plants during normal development, as it has been shown that the PR-1a promoter responds not only to environmental, but also to developmental signals . Again, higher levels of GUS activity were observed consistently with transformants containing the as-1[GUS] transgenes and not with the wild-type promoter constructs (data not shown). Taken together, these results demonstrate that as-1 can functionally replace the as-1-like motif in the tobacco PR-1a promoter in planta.
Replacement of the as-1-like element by as-1 reduces the ratio of reporter gene activities in necrotic tissue exhibiting the HR to SA treated tissue
To analyze the effect of an as-1 containing PR-1a promoter on gene induction during the HR, the selfed progeny from two independent primary transformants with the −1533as-1[GUS] construct, 313–6 and 313–7, were selected on MS medium in the presence of kanamycin. Transformant 313–6 had displayed an intermediate reporter gene expression in the initial SA induction experiment shown in Fig. 3B, whereas transformant 313–7 had exhibited high responsiveness to SA. Resistant seedlings were transferred to soil. In parallel, seedlings containing wild-type PR-1a promoter constructs were selected. At the six-leaf stage, leaf disks were cut from the plants and floated on H2O or a solution of 1 mm SA. Immediately afterwards the same plants were infected with TMV. GUS reporter gene expression was determined in protein extracts isolated from leaf disks after 4 days of chemical treatment or 7 days after virus inoculation. As shown in Fig. 4A, reporter gene induction was considerably stronger in response to TMV infection in comparison to chemical treatment in transformants containing the −906 or the −1533 bp PR-1a wild-type promoter constructs (9.2- and 14.3-fold higher GUS activities, respectively, after TMV infection). In contrast, transformants containing the −1533as-1[GUS] transgene consistently exhibited higher levels of SA inducible GUS expression compared to reporter gene induction measured in TMV infected plants exhibiting the HR. TMV infection of transformants 313–6 and 313–7 produced only 2.7- to 4.8-fold stronger GUS activities than chemical treatment (Fig. 4A). To account for possible errors in GUS activities between individual plant extracts due to experimental variations, GUS enzyme extracts were also analyzed for the expression of the endogenous PR-1 proteins by immunodetection. In all cases, induction of the endogenous PR-1 proteins was significantly stronger in virus infected tissue than in leaf tissue subjected to SA (Fig. 4B). These results demonstrate that GUS reporter gene induction in plant lines with the −906 and −1533 bp PR-1a wild-type promoter constructs reflects induction of the endogenous PR-1 proteins by the different stimuli. Therefore, as-1 enhances gene expression from the PR-1a promoter in response to SA, but it does not seem to markedly affect PR-1a gene induction during the HR.
PR-1 proteins are induced to high levels in necrotic tissue from tobacco plants expressing the nahG gene from P. putida
It is envisaged that PR-1 gene induction in local and systemic tissues of infected plants is mediated by SA produced during the course of the HR [20,21,41]. However, our finding showing differential inducibility of an as-1 carrying PR-1a promoter in comparison to the wild-type promoter in response to SA vs. the HR, would favor an alternative explanation. In view of the results presented, it seems plausible that two distinct signal transduction pathways could be targeted by SA and by signals released during the HR, respectively, resulting in the independent activation of different transcription complexes. These complexes, in turn, could interact differentially with the wild-type and the as-1 mutant promoter regions, thus leading to differential PR-1a gene activation in response to the different stimuli. To challenge this model, we have made use of tobacco plants expressing the P. putida nahG gene which codes for salicylate hydroxylase (SAH). Previously, it has been reported that expression of the bacterial gene in tobacco leads to drastically reduced levels of endogenous SA during the plant defense response [16,19].
The nahG gene from P. putida was cloned in a plant expression vector under control of the CaMV 35S RNA promoter and several independent transgenic lines (N. t. cv. Samsun NN) were generated via Agrobacterium-mediated transformation. Primary transformants were analyzed for the activity of SAH by infection with TMV. In wild-type plants, virus infection induces a local necrotic reaction which remains restricted to the primary infection site. On the contrary, six plants out of 9 regenerants transformed with the nahG expression vector exhibited enlarged local lesions as depicted in Fig. 5. Typically, lesions had a light-brown center surrounded by two distinct necrotic regions each bounded by a thinner dark-brown margin at 14 days post-infection (Fig. 5C). In contrast, the center of local lesions on wild-type plants was surrounded by only a single necrotic region bounded by a dark-brown margin.
This phenotype has been described previously for TMV infected tobacco plants that were severely depleted from endogenous SA by the expression of the nahG gene [16,17,19]. To further demonstrate the enzymatic activity of SAH, progeny plants from primary transformants, that were no longer able to restrict lesion growth, were monitored for endogenous SA levels and for their ability to induce PR-1 proteins after treatment with SA. The amount of total SA (free SA plus glucosylated SA) reached levels of 7199 ng per gram fresh weight in TMV infected control leaves (N. t. cv. Samsun NN), which represents a 90-fold induction over SA levels in noninfected leaves (79.5 ng total SA per gram fresh weight). On the contrary, total SA levels in inoculated tissue from nahG transformants of line 201–10 (58 ng per gram fresh weight) remained even below the levels detected in noninfected control leaves. Consistent with this result, plants from lines 201–10 and 201–4 were not able to accumulate PR-1 proteins to appreciable amounts after floating of leaf disks on H2O or on solutions of 1 mm or 5 mm SA, thus demonstrating the activity of SAH in these plant lines (Fig. 6A).
Subsequently, plants were infected with TMV. After 7 and 14 days, leaf material was collected from a narrow zone including the lesion and from a more distant zone around single lesions of each plant. Whereas samples from the distant zone of control plants never contained necrotic leaf material, samples collected from the distant zone of the transformants consistently included necrotic tissue 14 days after virus infection due to lesion expansion. At 14 days post-infection, lesions on nahG expressing plants from line 201–10 had reached an average diameter of 7.3 ± 0.6 mm and were still growing (Fig. 5B,C), whereas lesions on wild-type plants remained restricted to 4.2 ± 0.5 mm. Proteins were extracted from the samples and analyzed for the expression of PR-1 proteins. Significant amounts of PR-1 proteins were detected in the samples from the narrow zone around lesions collected from tissue of nontransformed plants as well as from nahG transformants (Fig. 6B, lanes 1,3,5,7). However, PR-1 proteins were never detected in tissue further away from the center of the necrotic lesions in control plants (Fig. 6B, lanes 6 and 8), whereas high levels of PR-1 proteins had accumulated at 14 days post-infection in the samples from the distant zone around lesions in nahG expressing plants (Fig. 6B, lane 4). Therefore, although depleted from endogenous SA and thus incapable to restrict lesion growth, tobacco leaf tissue undergoing necrosis is able to support efficient induction of PR-1 proteins.
Reporter gene expression from the tobacco 0.9 kb PR-1a upstream region has been studied extensively by several groups [7,39,40]. Although clearly inducible in plants treated with SA, during the HR, and during adult stages of development, the 0.9 kb PR-1a upstream region constitutes a rather variable promoter of only intermediate strength. By using a PR-1 reporter protein, we here demonstrate that a longer upstream region, the −1533 bp PR-1a promoter, yields regulated gene expression on the same order of magnitude as observed with the endogenous PR-1a gene (Fig. 1). Likewise, GUS reporter gene expression from the −1533 bp PR-1a promoter proved to be stronger and more reliable on average than expression from the 0.9 kb promoter (Fig. 3) . On the other hand, gene expression from the −1533 bp PR-1a promoter, as expression from the short −906 bp PR-1a promoter, is controlled in a similar way by the same cis-acting sequence, the as-1-like element (Fig. 2) . Therefore, the tobacco −1533 bp PR-1a upstream sequence represents a strong and highly inducible promoter region whose transcriptional activity largely relies on the as-1-like element.
To further characterize the significance of the as-1-like element for gene expression, we have inserted an authentic as-1 element from the CaMV 35S RNA promoter in place of the as-1-like sequence in the PR-1a promoter. GUS reporter gene expression from the as-1 carrying long and shorter PR-1a upstream sequences was monitored in transgenic tobacco plants in comparison to the wild-type promoter regions in response to SA, the HR and developmental cues. Clear induction of gene expression was retained with the mutant promoter regions towards the diverse stimuli (Figs 3 and 4 and data not shown). Thus, as-1 from a viral promoter is able to fully replace the as-1-like element in the tobacco PR-1a promoter. These findings would imply that the cauliflower mosaic virus, in order to enable transcription of its genes in planta, has adopted a cis-acting element from a plant gene active during the defense response. Furthermore, our results demonstrate that different as-1 related elements can be functional in the context of the PR-1a promoter, suggesting that the same or similar factors are involved in transcription via as-1 and the as-1-like element in the complex PR-1a upstream region in planta. Our conclusion is consistent with previous reports showing that TGA transcription factors, which were originally isolated via physical interaction with as-1, bind in vitro to as-1-like elements present in the tobacco PR-1a and the Arabidopsis PR-1 upstream regions [22,27–29].
Most importantly, however, our data show that reporter gene induction from an as-1 containing PR-1a promoter is even stronger on average than reporter gene induction from the wild-type promoter in SA treated and in mature untreated plants (Fig. 3 and data not shown; threefold increase of SA responsive reporter gene expression with the as-1 upstream region). On the contrary, reporter gene expression was not markedly affected in plants with the −1533as-1[GUS] chimeric gene during the HR (Fig. 4). Therefore, the as-1 containing mutant promoter appears to be regulated differently from the wild-type promoter in response to different stimuli. To account for a differential inducibility of the PR-1a wild-type and the as-1 carrying mutant promoter regions by diverse stimuli, it seems plausible to speculate that gene expression from the PR-1a promoter occurs through different pathways activated by exogenous application of SA and by signals elicited during the HR, respectively. Given this case, different, albeit related, transcription complexes could form on the PR-1a promoter in response to varying stimuli. These complexes may include TGA factors to mediate physical association of alternating transcription complexes with the PR-1a upstream region via the as-1-like element. In this scenario, variation of the as-1-like element within the PR-1a promoter could directly affect gene expression by differential interaction of the mutated promoter region with different transcription complexes.
To seek for support for our model of distinct signal transduction pathways leading to the independent activation of the tobacco PR-1a gene by SA and by an unknown HR released signal, we have made use of transgenic tobacco plants expressing the nahG gene from P. putida. The nahG gene codes for salicylate hydroxylase (SAH), which converts SA to the inactive compound catechol. It has been shown previously that tobacco plants expressing the nahG gene are barely able to induce SA levels above the background in response to TMV infection [16,19]. Curiously, plants incapable to accumulate SA to normal levels due to nahG expression exhibit a phenotype of enlarged local lesions upon infection with avirulent pathogens, whereas wild-type plants are able to restrict lesion size more rigorously. This phenotype has been reported for nahG expressing tobacco and tomato plants after infection with TMV, and for Arabidopsis plants after infection with avirulent bacterial and fungal pathogens [16,17,19,42,43]. Our tobacco plants transformed with a 35S[nahG] chimeric gene displayed the expected phenotype of enlarged local lesions after TMV infection. Once formed, lesions kept on expanding for several weeks and typically reached diameters of up to 22 mm 8 weeks after virus inoculation on progeny plants of line 201–10 (Fig. 5 and data not shown). These results clearly indicate that the transformants, unlike wild-type plants, did not contain sufficient SA to restrict lesion growth. Further proof that the nahG transformants used in this study were severely impaired in the transmission of the SA signal came from their lack to induce PR-1 proteins to normal levels after exposure of leaf disks to high concentrations of SA (Fig. 6A). Consistently, total SA levels in TMV infected leaves from nahG transformants were shown to be even lower than the levels detected in noninfected tissue from control plants. Plants handicapped in the accumulation of SA were infected with TMV, and PR-1 proteins were monitored in the necrotic tissue induced by the spreading lesions. Whereas lesions and the accumulation of PR-1 proteins remained strictly localized in infected wild-type plants over a period of 14 days, the extension and the level of PR-1 protein synthesis increased significantly in the nahG transgenic plants in parallel with the induction of necrosis by the expanding lesions (Fig. 6B). Our finding of substantial amounts of PR-1 proteins in nahG plants is consistent with other observations. Friedrich et al.  report that SAR associated genes like PR-1 are clearly expressed in the inoculated leaves of TMV infected nahG transgenic plants exhibiting drastically reduced levels of SA. Taken together, tobacco plants, although expressing active SAH, are able to induce PR-1 proteins in tissue undergoing necrosis, demonstrating that the expression of PR-1 proteins to high levels during the HR does not depend solely on the accumulation or transmission of the signal molecule SA.
Based on our findings we propose a new model for the activation of the tobacco PR-1a gene by SA and during the HR. Foremost, our model implies at least two distinct signal transduction pathways leading independently from each other to PR-1a gene induction, an SA dependent pathway and a pathway relying on unknown signals produced during the HR (Fig. 7). As activation of the PR-1a gene strictly depends on the as-1-like element , we further suggest that the two independent signaling cascades lead to the formation of related, albeit different, transcription complexes, which are likely to include TGA factors. Thus, gene activation via distinct signal transduction events would converge on the as-1-like element in the tobacco PR-1a promoter (Fig. 7). Our model is consistent with the existence of small protein families of plant TGA factors. TGA factors from Arabidopsis and tobacco share an extremely conserved basic domain by which they are able to interact with different as-1 related target sequences in vitro[22,32,44,45]. Furthermore, using chromatin immunoprecipitation assays, it has been shown recently that TGA2 and TGA3 are recruited in vivo to the Arabidopsis PR-1 promoter in response to a stimulus induction pathway involving SA and NPR1/NIM1 . TGA factors differ, however, clearly in other biochemical properties. Some TGA factors activate transcription in yeast via their N-terminal domains, whereas others do not seem to have an intrinsic transactivation potential . Also, some TGA factors interact in the yeast two-hybrid system with NPR1/NIM1, while others fail to do so [27–29,32]. Therefore, it seems plausible that different TGA factors, although binding to the same cis-acting element, are engaged in variant transcription complexes that are induced in plants in response to different stimuli, as proposed in our model (Fig. 7).
Our model is in conflict with previous conceptions in at least one important aspect. Generally, SA accumulation in local and systemic tissues of infected plants is considered to be a product of processes induced by the HR and is thought to be responsible for PR-1 gene induction during both the HR and SAR [20,21,41]. Yet, plants transformed with the nahG gene clearly demonstrate that PR-1 protein accumulation can occur to high levels in necrotic tissue depleted from SA (Fig. 6) . However, in nahG transgenic plants, PR-1 gene expression is abolished in the uninoculated leaves of tobacco plants displaying the HR after TMV infection . Therefore, active SAH is able to fully suppress the SA dependent pathway of systemic PR-1 gene induction in tobacco, but fails to fully suppress HR dependent PR-1 gene activation (Fig. 7), although SA levels are diminished to those in noninfected wild-type plants. Apart from the results with nahG transgenic plants, there are other indications in favor of our model. In Arabidopsis plants insensitive to SA due to a mutation in the NPR1/NIM1 gene, PR gene expression is blocked in systemic tissues after induction by a pathogen. However, in local tissues, PR gene expression, including PR-1, is not affected substantially by npr1. Thus, as PR-1 gene induction is abolished nearly completely in the npr1 mutant in response to SA , local induction of PR-1 in infected tissue must be independent of SA in Arabidopsis as well. A recent observation made by Fan and Dong  is intriguingly relevant to our model. The authors found that a chimeric transcription factor consisting of a fusion between Arabidopsis TGA2 and the GAL4 DNA binding domain is able to activate a GAL4 responsive reporter gene in Arabidopsis plants after exposure to SA. Yet, when a bacterial pathogen was used to infect transgenic plants with the TGA2::GAL4 effector/reporter gene system, only little reporter gene induction was observed, although infection can cause a significant induction of the endogenous PR genes and also of a PR reporter gene.
In conclusion, based on our findings that an as-1 carrying PR-1a promoter is differentially regulated from the wild-type promoter in response to different stimuli and that PR-1 proteins are induced to high levels in necrotic tissue depleted from endogenous SA, we suggest that the tobacco PR-1a gene can be activated by at least two distinct signal transduction pathways which both rely on the as-1-like element in a similar way. Whereas one pathway is triggered by SA accumulating in noninfected tissues during the SAR, the other pathway is induced by a yet unknown signal molecule formed in tissues undergoing necrosis during the HR. The identification of factors whose expression is regulated solely by SA or by HR released signals will support our hypothesis.
We would like to thank Bernhard Roth for communicating SA levels in tobacco tissue; Ivana Glocova for help with the analysis of transgenic plants; Maren Babbick for providing photographs of nahG plants; Ingrid Priessnitz-Hohos for transformation of tobacco plants; and Jochen Grob for technical assistance. This work was supported by grants from Genzentrum München and from Fonds der Chemischen Industrie.