INTRODUCTION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Most aromatic rings found in natural products ultimately derive from chorismate, a product of the shikimate (prechorismate) pathway. Chorismate is synthesized in seven steps from erythrose 4-phosphate and phosphoenolpyruvate. This pathway is restricted to plants, fungi, and bacteria, which renders the respective enzymes potential targets for herbicides and antibiotics (for review, see Bentley, 1990; Herrmann, 1995; Schmid and Amrhein, 1995, 1999). In plants the third and fourth steps of the pathway are catalyzed by the bifunctional enzyme 3-dehydroquinate dehydratase- (DHQase, EC 4.2.1.10) shikimate:NADP oxidoreductase (SORase, EC 1.1.1.25). Its enzymatic activities have been reported from several plant species and the respective proteins from pea and tobacco (Nicotiana sylvestris) have been purified to homogeneity (Mousdale et al., 1987; Bonner and Jensen, 1994; Deka et al., 1994). So far, only partial cDNA sequences encoding DHQase-SORase from tobacco (Bonner and Jensen, 1994), pea (Deka et al., 1994), and soybean (AW201059) have been reported.

With the exception of DHQase-SORase-specific cDNAs, full-length cDNAs corresponding to all other genes of the prechorismate pathway in tomato (Lycopersicon esculentum) have previously been isolated (Gasser et al., 1988; Schmid et al., 1992; Görlach et al., 1993a, 1993b; Bischoff et al., 1996). Here we report the cloning and characterization of DHQase-SORase-specific cDNAs and the isolation of the corresponding gene.

Without exception, the genes of prechorismate pathway enzymes from higher plants contain sequences encoding N-terminal plastid-specific transit peptides (Schmid and Amrhein, 1999). As the previously isolated partial cDNAs encoding DHQase-SORase from tobacco and pea did not encompass this crucial region, we made a special effort to obtain full-length cDNAs, as well as to analyze the corresponding gene(s).

In tomato, the organ-specific expression of the genes encoding enzymes of the prechorismate pathway has been analyzed in great detail (Görlach et al., 1994; Bischoff et al., 1996). It is interesting that the expression patterns were not identical for all prechorismate pathway genes, but rather, three distinct patterns were uncovered. For the 3-deoxy-D-arabino-heptulosonate 7-phosphate (DAHP) synthase gene LeDHS1, a unique expression pattern was observed with comparable transcript levels in all organs analyzed. The second pattern, shared by the DAHP synthase gene LeDHS2, the DHQ synthase gene and as we show here the DHQase-SORase gene, is characterized by the highest relative abundance of transcripts in roots, lower transcript levels in stems, flowers, and cotyledons, and still lower levels in leaves. The third pattern, with the highest relative abundance in flowers and roots, lower levels in stems and the lowest levels in leaves and cotyledons, is common for the genes encoding shikimate kinase (SK), 5-enolpyruvylshikimate 3-phosphate (EPSP) synthase and the two chorismate synthases (CS; LeCS1 and LeCS2).

Stress-induced expression of the genes encoding enzymes of the prechorismate pathway has so far mainly been analyzed in the context of plant-pathogen interactions. A comprehensive analysis concerning the elicitor inducibility of these genes has been done with cultured tomato cells (Görlach et al., 1995; Bischoff et al., 1996). Results presented here complete these studies and show that genes for each step of the prechorismate pathway are elicitor-inducible.

	RESULTS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Characterization of DHQase-SORase-Specific cDNA Clones and of the Corresponding Gene

A fragment of the tomato DHQase-SORase generated as outlined in "Material and Methods" was amplified by PCR and used as probe to screen a cDNA library from tomato flowers (Bischoff et al., 1996) and plasmids were obtained from positive clones by in vivo excision. Restriction analyses revealed four distinct clones and the 5' and 3' ends of these clones were determined. The sequences of three clones (designated cDHQase-SORase1) were identical as far as they overlapped. Sequencing from the 5' end identified the clone with the longest insert and sequencing from the 3' end identified three different polyadenylation sites (Fig. 1). A fourth clone contained an additional 57-bp sequence within the coding region (designated cDHQase-SORase2), but was otherwise identical to the other three cDNAs. Both strands of the cDHQase-SORase1 clone with the longest insert and the cDHQase-SORase2 clone, respectively, were sequenced completely (Fig. 1). The deduced amino acid sequences indicated that the cDNAs encoded DHQase-SORases. Sequence comparison with DHQase-SORase from tobacco (Bonner and Jensen, 1994) showed that these clones did not contain the complete coding sequences. Therefore, 5'-RACE was performed and the sequence of the amplification product was determined (Fig. 1); the complete sequence is deposited in GenBank (accession no. AF033194). As far as they overlapped, the sequences of the RACE product and of the cDNAs were identical. A stop codon 5' of the first ATG (Fig. 1) in frame with the coding region suggested that the RACE fragment contained the beginning of the translated region. Therefore, the combined sequences of the RACE fragment and of the cDNAs must comprise the entire sequence coding for the tomato DHQase-SORase. A comparison with known sequences of microbial monofunctional DHQases and SORases revealed that the DHQase and the SORase domains reside in the N- and C-terminal parts of the bifunctional tomato enzyme, respectively, and thus, the polypeptide has the same molecular organization as the two other plant DHQase-SORases that have previously been analyzed (Bonner and Jensen, 1994; Deka et al., 1994).

View larger version (57K): [in this window] [in a new window]   Figure 1.   Nucleotide and deduced amino acid sequences of a cDNA encoding DHQase-SORase from tomato. The nucleotide sequence shown in italics (bp 1-88) was obtained by 5'-RACE. The underlined nucleotide sequence was missing in some of the analyzed cDNA clones. Horizontal arrows mark polyadenylation sites of different cDNA clones with shorter 3'-untranslated regions. In the 5'-untranslated region, the stop codon in frame with the coding region is indicated by an asterisk. The proposed cleavage site of the plastid-specific transit peptide is indicated by a vertical arrow. The tomato sequence similar to the N terminus of mature tobacco DHQase-SORase (GEAMTR, Bonner and Jensen, 1994) is shown in bold face. Double-line arrows indicate the oligonucleotide primers that were used in the reverse transcriptase (RT)-PCR experiment.

To address the question of the origin of the additional 57-bp segment within the coding region of one of the cDNA clones (cDHQase-SORase2), the corresponding gene (LeDHQase-SORase) was isolated on two overlapping  clones (Fig. 2B), and a region comprising about 23 kb was sequenced (accession no. AF034411). The DHQase-SORase gene consists of 12 exons and 11 introns, some of which are rather long (up to 5 kb). As far as they overlap, the sequences of the gene and of the two types of cDNAs are identical. The 57-bp region that is missing in cDHQase-SORase1 is identical to exon 3. To demonstrate that the transcript of LeDHQase-SORase2 is actually present within the mRNA pool we performed RT-PCR using total RNA from different tissues of tomato plants as template. Using a pair of primers (indicated in Fig. 1) corresponding to two regions within exons 2 and 3, respectively, specific PCR products were generated, showing that a mRNA that retained exon 3 is present in tomato roots, cotyledons, stems, leaves, and flowers (Fig. 3). To determine the relative abundance of the two transcripts real-time RT-PCR was performed on total RNA isolated from tomato tissues. In two experiments the DHQase-SORase1 transcript level was found to exceed that of the exon 3-containing transcript by factors of 2⁵ in leaves, and 2⁶ in roots and flowers, respectively. These results demonstrate that the two types of cDNAs represent differentially spliced transcripts originating from a single gene. Thus, the primary transcript appears to be processed by joining all 12 exons together, creating a transcript represented by cDHQase-SORase2, or by joining exon 2 directly to exon 4, i.e. skipping exon 3, by splicing out a large intron comprised of intron 2, exon 3, and intron 3, thereby creating a transcript represented by cDHQase-SORasel1. About 3 kb 5' of the translation start site of DHQase-SORase, the 3' end of a superoxide dismutase gene (LeSodCc2, accession no. X877372) was identified (Fig. 2B).

View larger version (12K): [in this window] [in a new window]   Figure 2.   Southern-blot analysis of chromosomal tomato DNA (A) and schematic representation of the tomato DHQase-SORase gene (B). A, High-Mr DNA was digested with the restriction enzymes BamHI, EcoRI, HindIII, or PstI and subjected to Southern-blot analysis using the complete cDNA cDHQase-SORase1 as probe. A 1-kb ladder (Gibco-BRL, Cleveland) was used as size marker. B, The subcloned fragments of two genomic phage clones (gLe6/1 and Le11/2) are indicated with their respective restriction sites for BamHI (B), EcoRI (E), HindIII (H), and PstI (P) and shown in the top. In the bottom, the structure of the tomato DHQase-SORase gene (LeDHQase-SORase) is shown. Boxes represent exons and translated regions are indicated by black boxes. The exons of the DHQase-SORase gene are numbered from 1 to 12. Exons VI and VII from the LeSodCc2 encoding a superoxide dismutase C-terminal domain are located directly upstream of the DHQase-SORase gene.

View larger version (61K): [in this window] [in a new window]   Figure 3.   Detection of the LeDHQase-SORase2 transcript by RT-PCR. A single-stranded cDNA was generated by reverse transcription from tomato total RNA (2 µg) using an oligonucleotide complementary to exon 4 as a primer. The cDNA was used as template in the PCR using a pair of primers corresponding to 20 nucleotides of exons 2 and 3, respectively (lanes 3, R, C, S, L, and F). Control reactions in lanes 1 and 2 lacked the 5'- and 3'-PCR primers, respectively. In the control reactions shown in lanes 4 and 5, 100 ng of tomato genomic DNA and 1 ng of the LeDHQase-SORase2 cDNA were used as PCR templates, respectively. Tomato total RNA was isolated from roots (R), cotyledons (C), stems (S), leaves (3,L), and flowers (F). A 100-bp DNA ladder was used as size marker (lane 6).

The remote possibility of cDHQase-SORase1 being derived from a second, almost identical gene was eliminated by Southern-blot analysis. All fragments that hybridized to the cDNA cDHQase-SORase2 used as radiolabeled probe could be assigned to restriction fragments strictly determined by the sequence of the DHQase-SORase gene (Fig. 2A), indicating that no other closely related gene exists in the tomato genome. However, restriction with PstI yielded three fragments that hybridized to the radiolabeled probe during Southern-blot analysis, whereas four fragments would have been predicted by analysis of the genomic sequence. It is possible that the band at 6 kb is actually a doublet of two fragments, comprising the larger PstI fragment of gLe1172 and the fragment that has not been sequenced in its entirety at the 3' end of the gene (Fig. 2, A and B).

Complementation Assays

The enzymatic activities of the two tomato DHQase-SORase isozymes were tested upon their expression in Escherichia coli strains deficient in the DHQase (strain AB1360, aroD 362) or the SORase (strain AB2834, aroE 353) activities. These strains were transformed with constructs containing the complete coding regions for one of the two putative isozymes in the vector pBluescript SK(-). Expression of the DHQase-SORase isozymes in the two E. coli mutant strains was confirmed by western-blot analysis of bacterial extracts using an antiserum directed against a DHQase-SORase-GST fusion protein (see "Material and Methods"). The expression levels of the two isozymes in the two strains were found to be similar (Fig. 4A). Ampicillin-resistant colonies were selected on rich medium, and then plated out on rich or minimal medium (Fig. 4B). On rich medium, all transformed E. coli strains were able to grow, whereas on minimal medium only the strains expressing DHQase-SORase1 (corresponding to cDHQase-SORase1) were able to grow. The E. coli strains harboring a control plasmid coding for an unrelated protein (DHQ-synthase) or the vector directing the expression of DHQase-SORase2 were not able to grow. Thus, the DHQase-SORase1 protein exhibits both enzymatic activities in E. coli, which provides ultimate proof that the isolated cDNA encodes a DHQase-SORase. DHQase-SORase2 was apparently not active in E. coli even though a protein of the expected size was detected on western blots (Fig. 4A). Thus, the presence of the additional 19 amino acids in the DHQase portion of DHQase-SORase2 must also have affected the SORase activity.

View larger version (32K): [in this window] [in a new window]   Figure 4.   Expression of DHQase-SORase1 and 2 in E. coli. A, Western-blot analysis of expression. Crude bacterial culture extracts of strains AB1360 (DHQase-deficient) and AB2834 (SORase-deficient) carrying plasmids directing the expression of DHQase-SORase2 (A), DHQase-SORase1 (B), or an unrelated protein (DHQ-synthase, C) were analyzed on protein gel blots using a polyclonal antiserum raised against tomato DHQase-SORase2 expressed in E. coli. The band at 30 kD represents an endogenous E. coli protein and indicates equal loading of the gel. B, Complementation of E. coli strains deficient for the DHQase or SORase activity, respectively. Plasmids directing the expression of DHQase-SORase2 (A), DHQase-SORase1 (B) from tomato, or an unrelated protein (DHQ-synthase, C) were used to complement the E. coli strains AB1360 (DHQase-deficient) and AB2834 (SORase-deficient). Cells were plated on rich (1) or minimal (2) medium, respectively, containing ampicillin (100 mg/L).

Subcellular Localization of the DHQase-SORase

All enzymes of the prechorismate pathway analyzed so far were found to be synthesized as precursors with N-terminal plastid-specific transit peptides (Schmid and Amrhein, 1998). Yet the N terminus of the deduced amino acid sequence of the tomato DHQase-SORase (Fig. 1) did not exhibit the typical features of such transit peptides (Keegstra et al., 1989), which suggested that additional sequences coding for an extended transit peptide may reside upstream of the first ATG identified in the DHQase-SORase sequence. This possibility, however, can be ruled out for several reasons. As mentioned above, 5' of the first ATG, a stop codon was found to be located in frame with the coding sequence, which strongly indicated that the 5'-RACE product (Fig. 1) contained the translation start site. The identity of the stop codon in the sequence of the RACE product was confirmed in the genomic sequence. Furthermore, no sequences capable of encoding a transit peptide could be identified in the genomic sequence. As the genomic clone comprised the entire sequence up to the 3' end of the next gene (Fig. 2A), it is highly unlikely that any part of the DHQase-SORase gene was missing. This conclusion was supported by the result of a northern-blot analysis with RNA from roots and flowers of tomato, which demonstrated that there is no transcribed region 5' of the designated first exon (Fig. 5). When the blot was probed with a fragment covering parts of the first intron and the second exon, signals of the expected size were detected, but when probed with a 2-kb fragment covering the region 5' of the first exon, no signal was detected.

View larger version (35K): [in this window] [in a new window]   Figure 5.   Northern-blot analysis with probes corresponding to transcribed and 5'-untranscribed regions of the tomato DHQase-SORase gene. A, Total RNA from tomato roots (R) and flowers (F) was subjected to northern-blot analysis using radiolabeled probes corresponding to the transcribed (1) and 5'-untranscribed (2) regions, respectively, of the DHQase-SORase gene. RNAs of different lengths (Gibco-BRL) were used as size markers (nucleotides × 103). B, Schematic representation of the probes used for the northern-blot analysis. The exons are numbered as in Figure 2.

We compared the N-terminal sequence deduced from the tomato cDNA with that determined by N-terminal sequencing of the mature DHQase-SORase protein isolated from tobacco (Bonner and Jensen, 1994). The N-terminal sequence of mature tobacco DHQase-SORase (GEAMRKN) was found to be highly similar to the deduced sequence of the tomato enzyme starting at amino acid 14 (GEAMTRN; compare with Fig. 1). Therefore, the precursor of DHQase-SORase appears to comprise an N-terminal extension of only 13 amino acids, which may function in chloroplast targeting.

To confirm the plastidic localization of the tomato DHQase-SORase in planta, a cell fractionation analysis was done (Fig. 6). Antiserum raised against the tomato DHQase-SORase immunodecorated a protein of the expected size in subcellular fractions of tomato leaves (Fig. 6), whereas the preimmune serum did not (data not shown). Most of the material reacting with the antiserum was detected in the stroma fraction, clearly indicating that the protein is localized in plastids. The low amount of immunoreactive protein in the fraction designed "cytosolic" was due to contamination by broken plastids, as indicated by the observation that an antiserum against the large subunit of RUBISCO revealed a contamination of the cytosolic fraction with this stroma protein in comparable magnitude (Fig. 6).

View larger version (40K): [in this window] [in a new window]   Figure 6.   Subcellular localization of the tomato DHQase-SORase. Western-blot analyses (SDS-PAGE, 10% [w/v] acrylamide) were performed with subcellular fractions obtained from tomato seedlings. A, Western blot immunodecorated with an antiserum raised against the tomato DHQase-SORase2. B, Western-blot analysis using an antiserum against the large subunit of the RUBISCO from pea. C, Proteins stained with amido black. The "low range" markers (Bio-Rad, Hercules, CA) were used as size markers (kD).

Expression Profiles of the Tomato DHQase-SORase Gene

The organ-specific expression of LeDHQase-SORase was analyzed with a dot-blot assay using RNA isolated from roots, cotyledons, stems, leaves, and flowers of tomato plants (Fig. 7A). Steady-state levels of DHQase-SORase-specific transcripts were highest in roots, lower in cotyledons, stems, and flowers, and lowest in leaves.

View larger version (9K): [in this window] [in a new window]   Figure 7.   Expression profiles of the tomato DHQase-SORase gene. The relative abundance of DHQase-SORase-specific transcripts was determined by a dot-blot assay, and the signals were quantified with a PhosphorImager and normalized to the transcript level observed in leaves. The cDNA DHQase-SORase2 was used as radiolabeled probe. A, Profile of organ-specific expression. B, Elicitor-induced expression. Cultured tomato cells were incubated with () or without () a fungal elicitor for the time periods indicated.

The elicitor inducibility of LeDHQase-SORase was tested in a dot-blot assay with RNA from cultured tomato cells that had been treated for different periods of time with a fungal elicitor (Fig. 7B). After 5 h, the abundance of DHQase-SORase-specific transcripts reached a maximum, which was about eight times higher than the level in uninduced cells.

	DISCUSSION

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

The analysis of DHQase-SORase-specific cDNAs from tomato revealed two distinct classes of clones that were shown to be derived from alternatively spliced transcripts differing in the presence or absence of the 57 nucleotides comprising the third exon (Figs. 1 and 2). Without exception, the intron-exon boundaries of the corresponding gene clearly follow the GT/AG rule (Table IA; Aebi et al., 1986) and are all closely related to the consensus for dicot plants (Table IA; Simpson and Filipowicz, 1996). As in mammals and yeast, splice-site recognition in plants requires a number of cis- and trans-acting elements. The selection of the 5'-splice site, for example, requires complementarity between the sequence at the splice site and the 5' end of the U1 small nuclear ribonucleoprotein particle, a component of the spliceosome (Simpson and Filipowicz, 1996; Eperon et al., 2000 and refs. therein). Sequence complementarity, however, is not sufficient for splice-site selection, which can be affected also by trans-acting proteins (Eperon et al., 2000) or secondary structural elements (Liu et al., 1995). As compared with mammals and yeast, a distinguishing feature in higher plant pre-mRNA splicing is the requirement for A/U or U-rich intron sequences. A/U-rich sequences are required for efficient splice-site recognition downstream of the 5'-splice site and upstream of the 3'-splice site, respectively (Simpson and Filipowicz, 1996). We analyzed the DHQase-SORase gene and compared the A/T contents and the T contents for stretches of 50 nucleotides 5' and 3' of each of the respective splice sites (Table IB). The A/T content of intron 2 versus that of exon 3 differs by only 2% and the T content by only 6%. There are two other intron/exon pairs differing by only 2% in their A/T contents (intron 4 versus exon 5, and intron 11 versus exon 12). These two regions, however, differ substantially in their T contents (20% and 14%, respectively). The small differences between the A/T contents and T contents in the intron 2 versus exon 3 regions may explain, at least in part, why exon 3 is not consistently recognized by the splicing machinery, thus giving rise to the two classes of transcripts. Exon skipping has been observed in several Arabidopsis mutants, giving rise to malfunctioning proteins (Brown, 1996; Simpson et al., 1998), but skipping of a potential exon as a prerequisite for the production of a functional protein appears to be a novel observation. As DHQase-SORase2, i.e. the protein containing the 19 amino acids encoded by exon 3, did not complement the two E. coli mutants, one must conclude that the presence of this stretch precludes both enzymatic activities. The apparent lack of enzymatic activity and the low abundance of DHQase-SORase2 transcript indicate that this protein may not be functionally relevant in planta.

View this table: [in this window] [in a new window]   Table I.   Intron-exon boundaries of the LeDHQase-SORase gene A, 5' and 3' splice site junctions of the LeDHQase-SORase introns. Sequences of splice site junctions of LeDHQase-SORase are aligned with respect to the GT/AG rule (Aebi et al., 1986). Bases corresponding to the consensus sequence for 5' and 3' splice sites of introns of dicot plants (Simpson and Filipowicz, 1996) are highlighted. B, A/T contents and T contents for stretches of 50 nucleotides 5' and 3' of each of the respective splice sites of the LeDHQase-SORase gene.

All available sequence information indicates the presence of a very short N-terminal extension in the tomato DHQase-SORase sequence that is absent from the mature, plastid-localized enzyme, as shown by N-terminal sequence analysis (Bonner and Jensen, 1994). Evidence for the existence of a transit peptide has been provided by the N-terminal sequencing of the mature tobacco protein (Bonner and Jensen, 1994). Furthermore, subcellular localization studies (Fig. 6; Feierabend and Brassel, 1977; Weeden and Gottlieb, 1980; Fiedler and Schultz, 1985) indicated that DHQase-SORase activity is localized in plastids. Yet the sequences of the putative DHQase-SORase transit peptides do not show any of the characteristics commonly found in transit peptides (Keegstra et al., 1989), nor do the putative processing sites resemble the consensus sequence for such sites (Gavel and von Heijne, 1990). In a similar manner, for betaine aldehyde dehydrogenase of spinach and sugar beet, targeting to the chloroplast has been demonstrated in transgenic tobacco plants, although these proteins contain very short transit peptides of at most eight amino acids (Rathinasabapathi et al., 1994). It is not clear at this time for betaine aldehyde dehydrogenase or for DHQase-SORase whether the short amino-terminal extension is sufficient for chloroplast import or whether additional targeting information resides within the mature polypeptides. It is clear that to resolve this question, the in vivo targeting of the DHQase-SORase needs to be investigated, e.g. by transient expression of suitable fusion constructs in mesophyll protoplasts. The lack of a typical plastid-specific transit peptide in the DHQase-SORase of tomato and tobacco may well be a specific feature of solanaceous species since the recently deposited corresponding genomic sequence of Arabidopsis (AAF08579) suggests the presence of a conventional transit peptide.

The organ-specific expression pattern of LeDHQase-SORase (Fig. 7A) is very similar to that obtained for one of the two DAHP synthase genes (LeDHS2) and the DHQ synthase gene, and very different from the pattern for LeDHS1 and those common for the SK, EPSP synthase, and the two CS genes (Görlach et al., 1993; Bischoff et al., 1996). Assuming that the abundance of the different transcripts reflects the levels of the corresponding enzymatic activities, the three distinct expression patterns seem to indicate that there exist three distinct functional modules of the prechorismate pathway. The first one (LeDHS1) provides a constant flux of erythrose 4-phosphate and phosphoenolpyruvate into the pathway. A second module comprised of LeDHS2, DHQ synthase, and DHQase-SORase may be responsible for the synthesis of pathway intermediates that are utilized for the synthesis of compounds in branch pathways on the one hand (e.g. quinate or depsides such as chlorogenic acid) and for the synthesis of chorismate on the other hand. The last module, consisting of SK, EPSP synthase, and the two CS isozymes, would be solely responsible for the biosynthesis of chorismate.

The observation that DHQase-SORase-specific transcripts accumulate in tomato cells exposed to a fungal elicitor (Fig. 7B) complemented previous results obtained in our laboratory, i.e. that genes encoding enzymes of the prechorismate pathway are induced after pathogen attack (Görlach et al., 1995; Bischoff et al., 1996). Taken together, these results suggest that the enhanced demand for chorismate, most of it presumably utilized for the synthesis of Phe, appears to be met by an enhanced transcription of the genes encoding enzymes of the prechorismate pathway. Furthermore, the finding that the transcripts for all these enzymes are induced under these conditions indicates that no storage pools exist for any of the pathway intermediates. This in turn raises the question concerning the availability of erythrose 4-phosphate and phosphoenolpyruvate. Our comprehensive analysis of the prechorismate pathway in tomato strongly suggests that the regulation of this pathway is tightly linked to the regulation of other primary, but also secondary, pathways and with advanced tools available such as the chip technology, it will be interesting to analyze the full complement of these metabolic perturbations.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION RESULTS DISCUSSION MATERIALS AND METHODS LITERATURE CITED

Basic molecular techniques were adopted from Ausubel et al. (1994) and Sambrook et al. (1989).

cDNA Libraries

The original tomato (Lycopersicon esculentum; Schmid et al., 1992; Bischoff et al., 1996) and tobacco (Nicotiana sylvestris) cDNA libraries were constructed in the expression vector ZAP (Stratagene, La Jolla, CA) according to the manufacturer's instructions. These libraries were synthesized from poly(A)⁺ RNA isolated from leaves or flowers. Recombinant pBluescript cDNA phagemids were excised in vivo from the ZAP library according to the recommended procedure (Stratagene). Ampicillin-resistant bacterial colonies containing the recombinant phagemids were selected on Luria-Bertani broth-plates, washed off the plates, and plasmid DNA was isolated. These pools were used as template in the PCR.

Generation of a DHQase-SORase-Specific PCR Fragment

Plasmid DNA (150 ng) from the tomato flower cDNA library was used in a 50-µL PCR containing 0.5 µM degenerate primer d1 (GG[GATC]GC[GATC]GA[CT][CT]T[GATC]GT[GATC]GA) corresponding to amino acids 57 to 62 of the tobacco DHQase-DHQase sequence (Bonner and Jensen, 1994), primer d4 (TT[GA]TG[GATC]GG[GAT]AT[GATC]GT[GATC]AC) corresponding to amino acids 318 to 323 of the same sequence, 200 µMdNTPs, 2.5 mM MgCl₂, and 5 units of Stoffel-fragment in the appropriate buffer (Perkin-Elmer, Foster City, CA). The mixture was overlayered with mineral oil and subjected to 30 reaction cycles using 95°C/45 s for denaturation, 45°C/60 s for annealing, and 72°C/60 s for elongation. An aliquot (1 µL) of the reaction mixture was directly used as template for a second PCR identical to the first one, except that primer d2 (CC[GATC]AC[GATC]TGGGA[GA]GG[GATC]GG) corresponding to amino acids 94 to 99 of the tobacco sequence and primer d3 (GG[CT]TT[GATC]CC[GAT]AT[GAT]AT[GATC]CC) corresponding to amino acids 265 to 270 of the same sequence were used. The resulting products were separated on a 0.8% (w/v) agarose gel and the fragment of the expected size (520 bp) was isolated, subcloned into pBluescriptSK(-) (Stratagene), and its sequence was determined according to Sanger et al. (1977).

Isolation of Tomato DHQase-SORase-Specific cDNA Clones

Tomato cDNA libraries (10⁶ phage each) from flowers and leaves were screened in duplicate using the ³²P-labeled DHQase-SORase-specific PCR fragment as probe under stringent conditions (0.1× SSC and 65°C). Positive clones were plaque-purified and excised in vivo according to the manufacturer's protocol (Stratagene).

RACE

The amplification of 5' ends of cDNA was performed with a 5'-/3'-RACE kit (Roche Diagnostics, Rotkreuz, Switzerland) according to the manufacturer's instructions with the addition of an initial denaturation step (10 min at 65°C). The tomato DHQase-SORase-specific primer was complementary to the nucleotide sequence at position 301 to 321 in Figure 1. The primer for the tobacco RACE was complementary to the sequence between position 342 and 359 of the N. tabacum DHQase-SORase cDNA (Bonner and Jensen, 1994). The resulting products were separated on a 1.2% (w/v) agarose gel, subcloned, and sequenced.

RT-PCR

An oligonucleotide primer complementary to 20 nucleotides of exon 4 (5'-GTAGGGCATTGTTGAACTCG-3', 500 nM, indicated in Fig. 1) was hybridized to 2 µg of tomato total RNA at 52°C for 5 min. Reverse transcription was performed for 45 min at 52°C using 12 U avian myelobastosis virus-reverse transcriptase (Roche Diagnostics). After heat inactivation (10 min at 65°C), the reaction products were purified using the High Pure PCR product purification system (Roche Diagnostics) according to the manufacturer's instructions. One-tenth of the reaction products was used as template in the PCR using oligonucleotides (indicated in Fig. 1) corresponding to exon 2 (5'-TCTTGTGGA- GGTTCGAGTGG-3') and exon 3 (5'-ACCGTATTAACAG- TATCCCC-3') as 5' and 3' primers, respectively. Forty cycles of amplification (94°C/45 s; 55°C/45 s; and 72°C/2 min) were performed in a thermal cycler (Cetus, Perkin Elmer) using 2.5 units Taq polymerase (Roche Diagnostics). One-tenth of the reaction products was analyzed on a 2% (w/v) agarose gel using a 100-bp DNA ladder (Gibco-BRL) as a size marker.

Real-Time PCR

Two milligrams of total tomato RNA were reverse transcribed using 10 units of avian myelobastosis virus-reverse transcriptase (Promega, Madison, WI) and a primer located in exon 4 (GTAGGGCATTGTTGAACTCG, corresponding to nucleotides 499-518 of accession no. AF033194) during 30 min at 42°C. After heat inactivation (10 min at 95°C), first-strand cDNA products were purified using the QIAquick PCR Purification System (Qiagen, Basel) according to the manufacturer's recommendations. A fraction (1/500) of the total was used as the template in real-time PCR (Taq-Man; PE Biosystems, Rotkreuz, Switzerland) using the Taq-Man Universal PCR Master Mix (PE Biosystems) and 200 nM of the Taq-Man probe labeled at the 5' end with 6-carboxyfluorescein and at the 3' end with 6-carboxy-N,N,N',N'-tetramethylrhodamine, respectively ([6-carboxyfluorescein]TCGACT-CTTTTCATCACCAGCATACTGACCA [6-carboxy-N,N,N',N'-tetramethylrhodamine], nucleotides 400-430; nos. refer to accession no. AF033194). PCR primers (900 nM each) corresponded to regions in exon 4 (reverse primer: CCCAACTCCATCGCTAATCG, nucleotides 443-462), the junction between exons 2 and 4 (DHQase-SORase1-specific forward primer: nucleotides 314-327 and 385-393), and the junction between exons 3 and 4 (DHQase-SORase2-specific forward primer: nucleotides 375-393), respectively. Taq-Man primers and probe were designed using the Primer Express Software (PE Biosystems) and were purchased from Microsynth (Balgach, Switzerland). Amplification and detection were performed with an ABI7700 real-time PCR system (PE Biosystems) using the following profile: 50°C/2 min, 95°C/10 min, followed by 40 cycles of 95°C/15 s and 63°C/1 min. The specificity of the DHQase-SORase1 and DHQase-SORase2-specific forward primers during Taq-Man PCR was confirmed using plasmids harboring the cDNAs cDHQase-SORase1 or cDHQase-SORase2 as the template.

Isolation of Genomic DNA Clones for Tomato DHQase-SORase

The genomic library, which contained partially MboI-restricted (12-23 kb) fragments of tomato (cv VFW8) genomic DNA within the BamHI cloning site of the EMBL3 vector (CLONTECH, Palo Alto, CA), was a gift of Dr. C. Ringli (University Zürich). Using the radiolabeled LeDHQase-SORase1 cDNA as a probe, 10⁶ phages were screened according to standard protocols (Sambrook et al., 1989). Twenty positive phage clones were plaque purified, and the DNA was isolated and subjected to partial sequence analysis. Two phages designated gLe6/1 and gLe11/2 were found to comprise the entire gene for tomato DHQase-SORase. The DNA inserts of these clones were subcloned into pBluescriptSK(-) and overlapping fragments were sequenced completely on both strands.

Expression Constructs and Antibody Production

Utilizing a unique HindIII restriction site (positions 256-261 in Fig. 1), the tomato RACE fragment was fused with either of the two tomato DHQase-SORase cDNAs to create constructs containing the complete coding regions in the vector pBluescriptSK(-). These constructs were used for complementation assays.

To generate sufficiently large amounts of the tomato DHQase-SORase polypeptide for antibody production, the complete coding region was fused with the coding region of glutathione-S-transferase in the vector pGEX-G (Görlach and Schmid, 1996). Crude lysates of E. coli expressing the fusion protein were separated on SDS-polyacrylamide gels and the fusion protein in polyacrylamide slices was used to raise antibodies in rabbits (BioScience, Göttingen, Germany).

Subcellular Fractionation

Intact chloroplasts were isolated and purified from 14-d-old tomato seedlings according to the method of Orozco et al. (1986). The supernatant of the first centrifugation step was designated the cytosolic fraction. Separation of chloroplasts into membranes and stroma was performed according to Keegstra and Yousif (1986). The proteins in the different fractions were separated by SDS-PAGE and were analyzed after transfer to nitrocellulose membranes using the antiserum raised against the DHQase-SORase GST fusion protein in 1,000-fold dilution.

Analysis of Data

Sequence analyses were done with the Wisconsin Sequence Analysis Package (version 8, Genetics Computer Group, Madison, WI).