The First Step of Gibberellin Biosynthesis in Pumpkin Is Catalyzed by at Least Two Copalyl Diphosphate Synthases Encoded by Differentially Regulated Genes

Abstract

The first step in gibberellin biosynthesis is catalyzed by copalyl diphosphate synthase (CPS) and ent-kaurene synthase. We have cloned from pumpkin (Cucurbita maxima L.) two cDNAs, CmCPS1 and CmCPS2, that each encode a CPS. Both recombinant fusion CmCPS proteins were active in vitro. CPS are translocated into plastids and processed by cleavage of transit peptides. For CmCPS1 and CmCPS2, the putative transit peptides cannot exceed the first 99 and 107 amino acids, respectively, because longer N-terminal deletions abolished activity. Levels of both CmCPS transcripts were strictly regulated in an organ-specific and developmental manner. Both transcripts were almost undetectable in leaves and were abundant in petioles. CmCPS1 transcript levels were high in young cotyledons and low in roots. In contrast, CmCPS2 transcripts were undetectable in cotyledons but present at significant levels in roots. In hypocotyls, apices, and petioles, CmCPS1 transcript levels decreased with age much more rapidly than those of CmCPS2. We speculate that CmCPS1 expression is correlated with the early stages of organ development, whereas CmCPS2 expression is correlated with subsequent growth. In contrast, C. maxima ent-kaurene synthase transcripts were detected in every organ at almost constant levels. Thus, ent-kaurene biosynthesis may be regulated through control of CPS expression.

GAs are synthesized from a common cyclic precursor, ent-kaurene (Graebe, 1987). This precursor is produced in two steps: (a) the cyclization of GGDP to CDP and (b) the cyclization of CDP to ent-kaurene. In plants these reactions are catalyzed by two distinct enzymes that have been purified separately (Duncan and West, 1981; Saito et al., 1995). Formerly, the enzymes were known as kaurene synthetase A and B (Duncan and West, 1981). At present, they are named CPS and KS, respectively (MacMillan, 1997).

Both CPS and KS are known to be localized in plastids (Railton et al., 1984; Aach et al., 1995, 1997). Early experiments in which [¹⁴C]GGDP and [³H]CDP were simultaneously supplied as the substrates to partially purified enzymes from Marah macrocarpus endosperm suggested that CPS and KS might interact (Duncan and West, 1981). In these experiments, CDP synthesized by CPS activity was utilized 13 times more efficiently for ent-kaurene production than exogenously supplied CDP. Thus, ent-kaurene biosynthesis was suggested to be catalyzed by a one-to-one CPS/KS complex in which CDP could be channeled from CPS to the KS catalytic site (Duncan and West, 1981). In the fungus Phaeosphaeria sp. L487, ent-kaurene biosynthesis is catalyzed by a single bifunctional CPS/KS enzyme (Kawaide et al., 1997); thus, the presence of two activities in one complex may be important for efficient ent-kaurene production.

ent-Kaurene and CDP are the first intermediates in GA biosynthesis from GGDP, and in pea, GA levels are directly correlated with the rate of ent-kaurene production (Moore and Coolbaugh, 1991). However, GGDP serves as a common precursor for several large families of terpenoid natural products (Graebe, 1987; Chappell, 1995; McGarvey and Croteau, 1995). Thus, the regulation of GGDP conversion into ent-kaurene is important for the partitioning of GGDP among different pathways and for the rate of GA biosynthesis (Duncan and West, 1981; Moore and Coolbaugh, 1991). Cloning of the CPS and KS genes from several species provided a way to examine the regulation of ent-kaurene biosynthesis by analyzing patterns of gene expression (for review, see Sun and Kamiya, 1997). The first KS gene was cloned from pumpkin (Cucurbita maxima L.) immature seeds, which contain high ent-kaurene synthetic activity (Yamaguchi et al., 1996). Thus, to compare expression patterns and to analyze a possible interaction between CPS and KS, we undertook the cloning of a pumpkin CPS gene.

Previously, studies of GA-deficient mutants resulted in the cloning of CPS genes from Arabidopsis (GA1; Sun and Kamiya, 1994), maize (An1; Bensen et al., 1995b), and pea (LS, Ait-Ali et al., 1997). However, analyses of mutants in some species also suggested the presence of additional CPS genes. In Arabidopsis, a mutant ga1–3 with a large deletion in the locus encoding CPS is still able to produce low amounts of GAs (Zeevaart and Talon, 1992). In a similar fashion, a maize deletion mutant with a knock-out for the An1 gene encoding a putative CPS was shown to accumulate ent-kaurene (Bensen et al., 1995b). These data indicate that several CPS genes may be expressed in a plant species.

In the present paper, for the first time to our knowledge, two CPS genes from the same plant species, CmCPS1 and CmCPS2, were cloned and characterized. Functional expression of the corresponding cDNAs demonstrated that both CmCPS proteins had enzyme activity. No interaction between the pumpkin CPS and KS proteins could be detected. The transcript levels of the two CmCPS genes and the CmKS gene were compared during seedling development and in adult plants. Our data indicate that the CmCPS genes are strictly regulated in a different organ-specific and developmental manner.

MATERIALS AND METHODS

Plant Material

Seeds of pumpkin (Cucurbita maxima L. cv Riesenmelone gelb vernetzt) were obtained from van Waveren Pflanzenzucht (Rosdorf, Germany) courtesy of Professor Jan Graebe. Pumpkin seedlings were cultivated under continuous light (220 μmol m⁻² s⁻¹) at 25°C on moist vermiculite. Adult (1-month-old) plants were grown under the same conditions. Immature seeds were harvested from field-grown plants, as described by Yamaguchi et al. (1996).

PCR Amplification of Pumpkin CPS cDNA Fragments

Degenerate primers were designed from the sequences SAYDTAWV (1F forward primer), FNGGVPN (3R reverse primer), and KHFERNG (5R reverse primer) (Ait-Ali et al., 1997). Double-stranded cDNA was synthesized from poly(A⁺) RNA isolated from cotyledons of immature pumpkin seeds, as described by Yamaguchi et al. (1996). PCR was carried out as described by Ait-Ali et al. (1997). Two PCR fragments of 0.89 kb (1F and 5R primers) and 0.64 kb (1F and 3R primers) were obtained and subcloned into the pCRII vector, as described by the manufacturer (TA Cloning, Invitrogen, San Diego, CA). We named the genes corresponding to the 0.89- and 0.64-kb PCR fragments CmCPS1 and CmCPS2, respectively.

Cloning of CmCPS1 cDNAs

Both 0.89- and 0.64-kb PCR fragments were used as the probes to screen a cDNA library prepared from immature pumpkin seeds (Yamaguchi et al., 1996). The probes were labeled using the ECL Direct Nucleic Acid Labeling and Detection System (Amersham, Japan). Plaque lifts on nylon membranes (Hybond-ECL, Amersham) were hybridized in the buffer provided by the manufacturer at 42°C. Washing was repeated three times at 42°C with a buffer containing 4 m urea, 0.5× SSC, and 0.4% (w/v) SDS, and then three times at room temperature with 2× SSC buffer.

All positive clones that were selected by library screening carried the CmCPS1 sequence. Only one of the clones had the full-length ORF; however, it also contained a putative unspliced intron. To obtain a CmCPS1 cDNA for functional expression in Escherichia coli, the coding region was amplified by PCR using double-stranded cDNA prepared from hypocotyls and petioles of pumpkin seedlings with the Gene Amplification RNA PCR Kit (Perkin-Elmer). Poly(A⁺) RNA for cDNA synthesis was enriched using magnetic resin with oligo(dT)₂₅ (Dynal, Oslo, Norway). For PCR, end-specific primer linkers were used: forward, 5′-ATATAAGTCGACATGAAAGCTCTCTCTCTCTCTCGC-3′ (the SalI site upstream of the putative first ATG codon of the CmCPS1 ORF is underlined), and reverse, 5′-ATATTAGCGGCCGCTTGACAATACAACATGGCTG-3′ (the NotI site is underlined). The amplified cDNAs were subcloned into both the expression vector pGEX-4T-3 (Pharmacia) and the pBluescript SK(−) vector (Stratagene). The obtained constructs were used to transform E. coli strain JM109.

To determine which part of the ORF is dispensable for the CPS activity, we subcloned several 5′- and 3′-truncated CmCPS1 cDNAs in the pGEX-4T-3 (Pharmacia) for functional expression in E. coli. For cloning, 6 reverse (containing a NotI site) and 15 forward (containing a SalI site) primers were designed from the CmCPS1 ORF and used for PCR reactions with the CmCPS1 cDNA as the template.

Cloning of CmCPS2 cDNAs

To isolate the full-length CmCPS2 cDNA, CmCPS2, 5′- and 3′-RACE experiments were done with the Marathon cDNA amplification kit (Clontech, Palo Alto, CA). We used cDNA fractions prepared from male pumpkin flower buds. PCR reactions were done with the specific primers designed from the sequence of the 0.64-kb fragment: forward, 5′-GGGAATCCGGAGCGACTCCCCGGCG-3′, and reverse, 5′-CGCCGGGGAGTCGCTCCGGATTCCCAGC-3′. Twelve independently obtained 5′- and 3′-RACE products were subcloned into the pCRII vector (Invitrogen) and sequenced.

Next, cloning of CmCPS2 cDNAs containing the full-length ORF was done from male flower bud cDNA by PCR with end-specific primers: forward, 5′-ATATATGAATTCCATGTCCTCCTCCTCCTCTCTCT-3′ (the EcoRI site upstream of the putative first ATG codon of the CmCPS2 ORF is underlined), and reverse, 5′-ATAATTCTCGAGACAACATGGGTGTGTGGGTAGCTA-3′ (the XhoI site is underlined). The amplified cDNAs were subcloned into the pGEX-4T-3 and the pCRII vectors. Also, three 5′-truncated CmCPS2 cDNAs were cloned into the pGEX-4T-3 for functional assay. For cloning, three forward primers containing an EcoRI site in frame with the CmCPS2 ORF were used in PCR reactions with the CmCPS2 cDNA as the template.

DNA Sequencing

DNA sequencing was done with double-stranded DNA using a dye primer cycle sequencing kit (Applied Biosystems) and an ABI373A DNA Sequencer (Applied Biosystems). For sequencing of the CmCPS1 cDNA clones obtained from the library, a series of unidirectional deletions was constructed from both ends using the Exonuclease III mung bean nuclease deletion kit (Stratagene).

Functional Analysis of the CPS Fusion Proteins

For each expression construct in pGEX-4T-3, at least five independent clones were analyzed for in vitro activity. Recombinant fusion proteins were produced by E. coli cultures incubated for 22 h at 20°C. Bacterial extracts in a CPS buffer (50 mm potassium phosphate, pH 8.0, 10% [w/v] glycerol, 2 mm DTT, and 5 mm MgCl₂) were cleared by centrifugation at 12,000g for 20 min at 4°C. Rapid assays of CPS and KS activities were carried out as described by Ait-Ali et al. (1997). Reactions were performed for 30 min at 30°C in 200 μL of the CPS buffer that contained either 1 kBq (75 GBq mmol⁻¹) of [³H]GGDP (Amersham) or 1.5 kBq (74 GBq mmol⁻¹) of [³H]CDP (a gift from Dr. T. Saito, Institute of Physical and Chemical Research, Saitama, Japan).

The products of CPS enzymatic activity were identified by full-scan GC-MS (GCQ, Finnigan MAT, San Jose, CA) as described by Kawaide et al. (1995). For analysis of CPS activity, 5 μg of GGDP was incubated with the bacterial extracts as described above. In the absence of MBP-KS, CDP was hydrolyzed to copalol as described by Sun and Kamiya (1994). Nonpolar products were purified from the reaction mixtures as described by Kawaide et al. (1995). Spectra were compared with that of the authentic ent-kaurene.

Immunodepletion and Immunoblotting

Five synthetic oligopeptides were designed from the CmCPS1 sequence, as shown in Figure 1. A CmKS oligopeptide was designed from the sequence EDDGYTSNRLMNTVK. The oligopeptides were coupled to Imject maleimide-activated keyhole limpet hemocyanin (Pierce). The anti-CPS1 and anti-KS polyclonal antibodies were custom produced by Sawady Technology (Tokyo, Japan) in rabbits against the mixture of the five CmCPS1 oligopeptides and the CmKS oligopeptide, respectively.

Figure 1.

Sequence alignment of plant CPS. The deduced CmCPS1 and CmCPS2 polypeptides are compared with the deduced pea LS, Arabidopsis GA1, and maize genes. Identical residues in two or more sequences are shown in black boxes with white lettering; biologically similar amino acids are shown by gray boxes. The terpene cyclase motif “DXDDTA” is shown with plus signs, and the motif “SAYDTAWVA” conserved among the CPS and KS enzymes is shown by asterisks below the alignment. Sequences used to design the degenerate primers for initial cloning are shown by arrows, and the sequences of synthetic oligopeptides are shown by horizontal bars above the alignment. The inverted triangle marks the limit of N-terminal deletions in CmCPS1 and CmCPS2 that still retain enzyme activity. Dashes indicate gaps introduced for optimization of the alignment.

Immunodepletion and immunoblotting were carried out as described by Harlow and Lane (1988). The endosperm extracts were prepared from immature pumpkin seeds, as described by Saito et al. (1995). Each immunoprecipitation reaction was performed using 30 μL (80 μg of total protein) of the endosperm extract in a total volume of 150 μL of the CPS buffer containing 150 mm NaCl. Proteins were precipitated by addition of the whole anti-CPS1 antiserum (in the control, this was replaced by the preimmune serum). Immune complexes were collected using 50 μL of protein A-Sepharose beads (Pharmacia). Protein A beads containing precipitated proteins were used for the rapid assay of KS activity, as described above. The remaining supernatants were concentrated 10 times using Molcut LGC (Nihon Millipore, Yonezawa, Japan) and subjected to SDS-PAGE using 9% (w/v) acrylamide gels. After the gel was blotted, the membranes (Hybond-ECL) were incubated with a 5000-fold dilution of anti-CPS1 or anti-KS antibodies. Detection was achieved using peroxidase-conjugated anti-rabbit IgG (Promega) and enhanced chemiluminescence (ECL, Amersham).

RNA Extraction and Gel-Blot Analysis

Total RNA was extracted using Trizol reagent (GIBCO-BRL) according to the manufacturer's protocol. The CmCPS and CmKS cDNAs containing the full-length ORF were used as the templates for preparation of [³²P]UTP-labeled RNA probes using the Riboprobe Combination Systems kit (Promega) as described by the manufacturer. RNA samples (40 μg per lane for total RNA or 3 μg per lane of poly(A⁺) RNA) were subjected to electrophoresis in 1% (w/v) agarose-2.2 m formaldehyde gels. After the gel was blotted, the membranes (Hybond N⁺) were hybridized for 16 h at 65°C in a buffer containing 50% (v/v) formamide, 1× Denhardt's solution, 1% (w/v) SDS, 5× SSPE, 0.1 m sodium phosphate (pH 7.0), and 0.2 mg/mL calf-liver RNA. The membranes were then washed in a 0.1× SSC and 0.1% (w/v) SDS buffer at 68°C. For control of RNA loading, membranes were rehybridized with an oligonucleotide probe complementary to 18S rRNA (Gallo-Meagher et al., 1992). The signals were visualized using a bio-imaging analyzer (BAS 2000, Fuji, Japan).

RESULTS

Cloning of Two Pumpkin CPS cDNAs

To clone pumpkin CPS sequences we used three degenerate PCR primers, designed by Ait-Ali et al. (1997), that consisted of one forward primer, 1F, and two reverse primers, 3R and 5R (Fig. 1). For reverse transcriptase-PCR, RNA fractions were isolated from cotyledons of immature pumpkin seeds. PCR reactions with each of the two different primer combinations generated a single product: a 0.89-kb cDNA fragment with primers 1F and 5R and a 0.64-kb cDNA fragment with primers 1F and 3R. Sequencing showed that the two PCR fragments were derived from different mRNA species; however, both fragments had high sequence homology with known CPS genes from other plants.

The two obtained PCR fragments were used as the probes for screening of a pumpkin cDNA library prepared from cotyledons of immature seeds, as described in Yamaguchi et al. (1996). All five cDNA clones obtained from the library corresponded to the 1F/5R PCR fragment of 0.89 kb. Among these cDNAs named CmCPS1, four were lacking the 5′ end and the full-length ORF of the fifth one was interrupted by a sequence that may have been an unspliced intron (data not shown). The deduced CmCPS1 sequence of 2.98 kb encodes an ORF of 823 amino acids followed by a 506-bp 3′-untranslated region.

To clone cDNAs corresponding to the 1F/3R PCR fragment of 0.64 kb, we used 5′- and 3′-RACE with gene-specific primers designed from the fragment sequence. RACE products of the appropriate sizes could be amplified only from RNA fractions of male flower buds. Six independent clones were analyzed for both the 3′ and 5′ ends of this sequence named CmCPS2. CmCPS2 encodes an ORF of 828 amino acids followed by a 156-bp 3′-untranslated region.

Alignment of the putative CmCPS polypeptides with other known CPS amino acid sequences is shown in Figure 1. CmCPS and CmCPS2 have 78% identity with each other and show high identity scores with the deduced CPS sequences of pea, Arabidopsis, and maize (about 55%, 50%, and 42%, respectively). Both pumpkin polypeptides carry the same conserved motifs as the other CPS, including the aspartate-rich box “DXDDTA” that is present in terpene cyclases catalyzing cyclization without removal of the diphosphate group (Sun and Kamiya, 1994). The two CmCPS have approximately 30% identity with the pumpkin KS (CmKS), and all three sequences contain a motif, “SAYDTAWVA,” that is conserved among the other plant CPS and KS polypeptides (Yamaguchi et al., 1998).

To show that the cloned CmCPS genes are present in the pumpkin genome, genomic DNA blots were hybridized with the CmCPS1 and CmCPS2 probes. Under high-stringency conditions, each probe hybridized with one or two bands of genomic DNA, and no cross-hybridization between probes could be detected (data not shown). The observed patterns were consistent with the restriction maps of the corresponding cDNAs. Thus, the pumpkin genome contains at least two different CPS genes that correspond to the cloned CmCPS1 and CmCPS2 cDNAs.

The CmCPS1 and CmCPS2 cDNAs Encode Proteins with CPS Activity

Recombinant CmCPS proteins were produced by expression in E. coli as GST-CPS fusions. The enzyme activities were first measured in vitro by conversion of labeled GGDP into ent-kaurene in the presence of the recombinant fusion of maltose-binding protein with CmKS (MBP-KS) (Ait-Ali et al., 1997). We used pumpkin endosperm extract as a positive control for the enzyme assay (Table I). Both GST-CPS1 and GST-CPS2 in the presence of MBP-KS converted labeled GGDP into a nonpolar product (presumably, ent-kaurene). To confirm the identity of the products, the n-hexane extracts from the enzymatic reactions with the unlabeled GGDP were analyzed by combined GC-MS, as described by Kawaide et al. (1997). As shown in Figure 2, the control extracts containing only GST or GST fused to a truncated CPS1 (Δ102, 102 amino acids deleted from the N terminus) did not metabolize GGDP, whereas the full-length GST-CPS1 converted all GGDP into CDP. The mixture of GST-CPS1 and MBP-KS converted all GGDP into ent-kaurene.

Table I.

ent-Kaurene synthetic activity of E. coli extracts containing GST-CPS fusion proteins of different sizes and the full-length MBP-KS

Fusion Protein	Radiolabeled ent-Kaurene
	dpm
GST (control)	33  ± 18
GST-CPS1 (full-length)	25,643  ± 3,821
GST-C-Δ86-CPS1	58  ± 12
GST-Δ98-CPS1	21,333  ± 4,741
GST-Δ99-CPS1	26,801  ± 3,656
GST-Δ100-CPS1	46  ± 9
GST-Δ101-CPS1	71  ± 7
GST-CPS2 (full-length)	16,723  ± 1,245
GST-Δ106-CPS2	14,562  ± 2,623
GST-Δ107-CPS2	14,195  ± 1,651
GST-Δ108-CPS2	39  ± 11
Pumpkin endosperm	24,490  ± 657

Recombinant fusion proteins were incubated with 1 kBq (about 50,000 dpm) of [³H]GGDP for 30 min at 30°C. Values are means ± se from three assays shown. For each expression construct, similar results were obtained with four to seven independent clones.

Figure 2.

Mass chromatograms of products after incubation of GGDP with recombinant fusion proteins. The indicated E. coli protein extracts (50 μg) were incubated with 5 μg of GGDP. Fragment ion peak with a mass-to-charge ratio (m/z) of 257 from each molecular ion (geranylgeraniol [GGol] and copalol [Col], m/z 290; ent-kaurene, m/z 272) was monitored.

Functional expression in E. coli was used to analyze which parts of the CmCPS sequences are dispensable for the enzyme activity. Six C-terminally truncated GST-CPS1 fusion proteins were prepared and all were inactive (the smallest examined deletion, C-Δ86, is shown in Table I). In contrast, N-terminal deletions within the first 99 residues (Δ98 and Δ99) did not affect the activity of GST-CPS1 (Table I). However, the deletion of the next residue I-100 (Δ100) and further deletions (Δ101) destroyed the enzyme activity. Also, three N-terminal deletions of CmCPS2, Δ106, Δ107, and Δ108, were prepared that corresponded to Δ98, Δ99, and Δ100 of CmCPS1, respectively (Fig. 1). The first two deletions, Δ106 and Δ107, did not change the activity. As expected from the homology between two CPS sequences, the third deletion of I-108 rendered the truncated GST-Δ108-CPS2 inactive (Table I).

The N termini of the CPS proteins contain transit peptides cleaved in the process of plastid import (Sun and Kamiya, 1997). The in vitro translated CmCPS1 protein was imported by isolated pea chloroplasts (data not shown) in a fashion similar to the Arabidopsis GA1 protein (Sun and Kamiya, 1994). After chloroplast uptake, the processed CmCPS1 protein was smaller than the full-length protein CmCPS1. The size of the putative CmCPS1 transit peptide was estimated at 10 kD.

Immunodepletion of Pumpkin Protein Extracts with Antibodies against CmCPS1

To study the possible complex formation between the CPS and KS proteins, polyclonal antibodies were raised against five oligopeptides designed from the CmCPS1 sequence (Fig. 1). The anti-CPS1 antibodies recognized a single 82-kD polypeptide of the putative CPS protein in the endosperm of immature pumpkin seeds but did not react with proteins from vegetative tissues (data not shown). Thus, endosperm extracts were used for immunodepletion experiments with anti-CPS1 antibodies and protein A-Sepharose beads.

As shown in Figure 3, the 82-kD CPS band remained in the supernatant after precipitation with the preimmune serum, but was depleted from the extracts by excess amounts of anti-CPS1 antibodies. However, the supernatants that were completely immunodepleted of the CPS protein still contained a putative KS protein of 81 kD recognized by anti-KS antibodies. We could not detect any KS activity in the immune complexes precipitated by anti-CPS1 antibodies (data not shown). Thus, no coprecipitation of CPS and KS proteins was observed in the extracts from immature pumpkin seeds.

Figure 3.

Immunodepletion of the pumpkin CPS protein with the anti-CPS1 antibodies. Endosperm extracts from immature pumpkin seeds (80 μg of total protein per reaction) were immunoprecipitated as follows: lane 1, by addition of 5 μL of the preimmune serum; lanes 2, 3, and 4, by addition of 1, 2, and 5 μL of the anti-CPS1 antiserum, respectively. Precipitated immune complexes were removed using protein A beads. The resultant supernatants were concentrated and subjected to SDS-PAGE and immunoblotting with either anti-CPS1 (left) or with anti-KS (right) antibodies (Ab).

The Levels of CmCPS1, CmCPS2, and CmKS Transcripts in Developing Pumpkin Seedlings

Pumpkin seedlings were collected daily from 3 to 10 d after imbibition and separated into roots (including root tips), hypocotyls, cotyledons, and apical parts containing meristems with small (less than 3 mm) leaf primordia and some remaining hypocotyl tissues. In addition, petioles and first leaves were collected from 7- and 10-d seedlings. Root tips were collected from 7-d seedlings.

Results of RNA gel-blot analyses are shown in Figure 4A. Although the nucleotide sequences of CmCPS1 and CmCPS2 shared about 80% identity, no cross-hybridization was observed between each probe and the heterologous CmCPS mRNA produced by in vitro transcription (data not shown). The CmCPS1 probe hybridized with two RNA species of about 2.7 and 1.8 kb. RNase A treatment of blots after hybridization with the CmCPS1 probe did not remove any of these signals. Also, both RNAs were detected on blots of poly(A⁺) RNA (data not shown). Thus, the two RNAs are likely to be mRNA species specifically recognized by the probe. Most probably, the higher band represented the full-length CmCPS1 transcript. The identity of the lower 1.8-kb band was not clear. This band was absent in some organs (e.g. cotyledons), and we speculate that it may be an alternatively spliced form or a specific RNA degradation product.

Figure 4.

A, Transcript levels of CmCPS1, CmCPS2, and CmKS in growing pumpkin seedlings. Seedlings were collected daily from 3 to 7 d after imbibition and divided into roots, hypocotyls, cotyledons, and apical parts. Three different blots of total RNA (40 μg/lane) were hybridized with the antisense RNA probes prepared from the coding regions of CmCPS1, CmCPS2, and CmKS cDNAs. As a loading control, the same blots were rehybridized with the labeled oligonucleotide probe recognizing 18S rRNA. Numbers on the top indicate days after imbibition. P, Petiole; Rt, root tip. Leaves were divided by size into the following three groups: S, small (less than 1.5 cm in length); M, medium (1.5–3 cm); and L, large (more than 3 cm). Arrows and numbers at the left indicate locations and sizes (in kb), respectively, of the mRNA species hybridizing with the probes. B, The growth of pumpkin seedlings from 3 to 7 d after imbibition. The length (○) and fresh weight (•) of hypocotyls and cotyledons were measured for 15 seedlings daily.

Both CmCPS1 and CmCPS2 transcripts were present in hypocotyls, petioles, and apices and undetectable in leaves. In cotyledons CmCPS1 transcript levels were very high at 3 and 4 d after imbibition. In contrast, CmCPS2 transcripts were almost undetectable in cotyledons of any age. In roots CmCPS1 transcript levels were very low, whereas CmCPS2 transcripts were observed at higher levels.

As shown in Figure 4B, for the whole period from 3 to 7 d, the hypocotyls and cotyledons underwent very rapid growth both in length and fresh weight. However, CmCPS1 transcript levels decreased below detection in 5-d cotyledons and decreased drastically in 4-d hypocotyls (Fig. 4A). Thus, in seedlings older than 4 d after imbibition, CmCPS1 expression did not correlate with organ growth. In contrast, CmCPS2 transcript levels in growing hypocotyls and roots did not change significantly from 3 to 7 d. Similar results were observed for the apical parts. Although the apices were still actively producing new leaves at 10 d after imbibition, CmCPS1 transcript levels decreased drastically, whereas CmCPS2 transcript levels decreased only slightly.

In contrast to the expression patterns observed for the CmCPS genes, CmKS transcripts were detected in all organs, and the transcript levels changed only slightly with time.

The Levels of CmCPS1, CmCPS2, and CmKS Transcripts in Different Organs of Adult Pumpkin Plants

To analyze the expression levels in adult organs, we collected 1-month pumpkin plants that had eight fully expanded true leaves. As shown in Figure 5A, plants were divided into four regions each containing leaves, petioles, and stems from two internodes. Regions were numbered from the first formed (lowest) to the last formed. At this age, first true leaves started to turn yellow and were collected as senescing material. The uppermost parts of plants, including shoot apices, youngest small leaves, and petioles, were pooled together. In addition, tendrils and male flower buds were collected from the same plants. Cotyledons of immature pumpkin seeds were obtained from a different study (Yamaguchi et al., 1996).

Figure 5.

A, Scheme showing the division of adult (1-month) pumpkin plants into four regions. Thick vertical bars show the region borders and numbers correspond to the region number. Ap, Apical part; Sc, senescing leaf. B, Transcript levels of CmCPS1, CmCPS2, and CmKS in 1-month pumpkin plants. Three different blots of total RNA (40 μg/lane) were hybridized with the antisense RNA probes prepared from the coding regions of CmCPS1, CmCPS2, and CmKS cDNAs. As loading controls, the same blots were rehybridized with the labeled oligonucleotide probe recognizing 18S rRNA. Numbers on the top indicate region numbers. T, Tendrils; Fb, male flower buds; Is, cotyledons of immature seeds. Arrows and numbers at the left indicate locations and sizes (in kb), respectively, of the mRNA species hybridizing with the probes.

Results of RNA gel-blot analyses are shown in Figure 5B. Both CmCPS transcripts were observed in young apical parts, petioles, male flower buds, and cotyledons of immature seeds. The relative amounts of CmCPS1 transcripts in immature seeds were extremely high (much higher than in the other tissues). In contrast, the relative amounts of CmCPS2 transcripts were higher in male flower buds than in immature seeds.

In leaves and tendrils, both CmCPS transcripts were almost undetectable by northern-blot analysis. However, CmCPS1 transcripts could be detected in young leaves by reverse transcriptase-PCR (data not shown). In stems high CmCPS1 transcript levels were observed mainly in the younger organs (third and fourth regions), whereas CmCPS2 transcript levels were generally very low.

In stems and petioles CmCPS1 transcript levels were much higher in the younger organs than in the older ones. In contrast, CmCPS2 transcript levels in petioles did not change as much as those of CmCPS1, were high in the second region, and were still detectable in the first region.

In contrast to the organ-specific expression of the CmCPS genes, CmKS transcripts were observed in all of the organs. CmKS expression levels did not change much from younger to older organs and could be detected even in senescing leaves.

DISCUSSION

Two Genes Encoding Functional CPS Enzymes Are Present in Pumpkin

Previous reports indicated that several plant species taxonomically distant from pumpkin may have more than one CPS catalyzing the first step of ent-kaurene biosynthesis (Sun and Kamiya, 1997). In maize a partial cDNA fragment for a putative second CPS (An2) has been cloned (Bensen et al., 1995a), but, to our knowledge, the corresponding protein has not yet been tested for the CPS activity. We have cloned from pumpkin two different cDNAs encoding enzymatically active CPS proteins. The deduced CmCPS polypeptides share 78% identity with each other and about 40% to 50% identity with the other known plant CPS proteins.

No Complex Formation Could Be Detected between the CPS and KS Enzymes

Endosperm extracts of Marah macrocarpus seeds were used for studies that suggested a possible interaction and substrate channeling between the CPS and KS enzymes (Duncan and West, 1981). However, the same authors reported that no biochemical evidence could be found for a possible heterodimer or heterooligomer of the native CPS and KS partially purified from seed endosperm extracts.

Our data indicate that CmCPS1 may be the main CPS species present in pumpkin seeds, since CmCPS1 cDNA was highly represented in the library prepared from immature seeds and the CmCPS1 transcript levels in this material were extremely high (Fig. 5). Moreover, antibodies against CmCPS1 protein recognized in seed extracts a putative CPS protein (Fig. 3). A putative KS was also detected in this material. However, immunodepletion and immunoprecipitation experiments using the anti-CPS1 antibodies failed to detect a coprecipitation of pumpkin CPS and KS (Fig. 3). Addition of the substrate (GGDP) did not facilitate interaction (M.W. Smith, unpublished results). Also, no complex formation could be detected in vitro between the GST-CPS1 and MBP-KS recombinant fusion proteins or by the in vivo yeast two-hybrid assay (M.W. Smith, unpublished results). We conclude that if the pumpkin CPS and KS do interact, they do not form a stable complex.

Putative Transit Peptides of the CmCPS Proteins

ent-Kaurene biosynthesis is known to occur in plastids; for pea and wheat, it is shown to be located in the proplastid stroma of rapidly dividing vegetative tissues (Aach et al., 1995, 1997). Both the Arabidopsis GA1 protein (Sun and Kamiya, 1994) and the CmCPS1 protein (M.W. Smith, unpublished results) were imported by isolated pea chloroplasts. During import the 10-kD transit peptides were cleaved from the CPS N termini. Functional analyses of N-truncated CmCPS1 and CmCPS2 proteins showed that the sizes of transit peptides cannot exceed the first 99 and 107 amino acids, respectively (Table I). The region of high sequence homology among CPS proteins, which begins at position 100 of CmCPS1 (108 of CmCPS2, Fig. 1), is indispensable for the CPS activity of both pumpkin proteins. Our data indicate that the calculated size of the 99-residue polypeptide, which could be truncated from the N terminus of CmCPS1 without activity loss, is close to the size of the CmCPS1 transit peptide estimated by chloroplast import experiments. However, further experiments are required to determine the precise N termini of mature CPS proteins.

The Two CmCPS Genes Are Differentially Regulated in an Organ-Specific and Developmental Manner

Previously, developmental control of CPS gene expression was shown in pea, in which the LS transcript levels were regulated during seed development (Ait-Ali et al., 1997). In Arabidopsis the GA1 gene expression patterns were analyzed using promoter-reporter fusions, and the promoter activity was shown to be restricted to specific cell and tissue types (Silverstone et al., 1997). In contrast, CmKS expression in pumpkin seedlings was not organ specific (Yamaguchi et al., 1996).

The expression patterns of the CmCPS and CmKS genes were analyzed in pumpkin plants using gel-blot analyses of total RNA. Both CmCPS genes showed organ-specific expression patterns. High levels of the two CmCPS transcripts were observed in male flower buds and cotyledons of immature seeds (Fig. 5B). Our data are consistent with the results of Silverstone et al. (1997) showing that GA1 promoter activity is high in Arabidopsis inflorescence meristem, anthers, and immature seeds. It is interesting that CmCPS1 transcript levels in cotyledons of immature seeds were much higher than in any other pumpkin organ (Fig. 5B). Also, very high CmCPS1 transcript levels were observed in cotyledons early in the development of pumpkin seedlings (Fig. 4A). It is unclear whether the CmCPS1 gene is specifically expressed in young cotyledons or whether the CmCPS1 transcripts remain from the extremely high levels found in seeds.

Among vegetative organs the levels of the two CmCPS transcripts were high in petioles but undetectable in all types of leaf material by RNA gel-blot analysis (Figs. 4A and 5B). However, GA1 promoter activity was detected in Arabidopsis leaf vascular tissues (Silverstone et al., 1997). We could detect CmCPS1 transcripts in young pumpkin leaves by reverse transcriptase-PCR (S. Yamaguchi, unpublished data). Thus, CmCPS transcript levels in pumpkin leaves may be extremely low and/or the CPS expression may be restricted to certain cell types. However, high levels of CmKS transcripts were present in leaves (Figs. 4A and 5B). Our data are consistent with the biochemical evidence showing that only KS, not combined CPS/KS, activity is found in fractions of mature chloroplasts (Railton et al., 1984). The experiments of Aach et al. (1995, 1997) showed that combined CPS/KS activity is observed only in proplastids of the growing plants. Leaves are known to produce large amounts of GGDP for biosynthesis of carotenoids and phytol. We speculate that CPS expression in leaf cells should be strictly regulated and kept at very low levels to prevent the overproduction of ent-kaurene and competition for the same substrate pool among the terpene-biosynthetic pathways.

During vegetative organ development, the transcript levels of both CmCPS genes decreased, albeit in a different manner. In the growing hypocotyls and apices of seedlings, CmCPS1 transcript levels decreased with age much faster than those of CmCPS2. Measurements showed that all seedling organs underwent very rapid growth for the whole period from 3 to 7 d (Fig. 4B). Thus, CmCPS1 gene expression was related to early organ development. We speculate that in pumpkin seedlings CmCPS1 expression may be high in dividing cells, which is similar to what was found in wheat shoots, where combined CPS/KS activity was predominantly in the meristematic tissues (Aach et al., 1997). However, even in meristematically active pumpkin shoot apices, CmCPS1 transcript levels rapidly decreased with age. In contrast, CmCPS2 transcripts were observed for a long time in apices and stayed at almost constant levels in growing roots and hypocotyls (Fig. 4A). Thus, CmCPS2 expression seemed to be related more to the growth than to the age of organs.

Although the CmCPS1 gene seems to encode a CPS enzyme highly expressed in seeds and young organs, CmCPS1 gene expression cannot be attributed to the early stages of plant ontogenesis, because significant levels of the transcripts were present in the young organs of adult plants. The different expression patterns of the two CmCPS genes suggest that they may play complementary roles in providing GA precursors required for plant growth.

The presence of at least two CPS genes may be a general feature of higher plants. Our data show that in pumpkin both CmCPS genes are much more strictly regulated than the CmKS gene. Thus, the first cyclization reaction catalyzed by CPS enzymes may be the key regulatory step in ent-kaurene biosynthesis.

Abbreviations:

CDP

copalyl diphosphate

CPS

copalyl diphosphate synthase(s)

GGDP

geranylgeranyl diphosphate

GST

glutathione S-transferase

KS

ent-kaurene synthase

MBP

maltose-binding protein

ORF

open reading frame

RACE

rapid amplification of cDNA ends