Arabidopsis thaliana contains two phosphoenolpyruvate carboxylase kinase genes with different expression patterns

Authors

  • V. Fontaine,

    Search for more papers by this author
    • *Present address: Centre for Plant Sciences, Leeds Institute for Plant Biotechnology and Agriculture, University of Leeds, Leeds LS2 9JT, UK

  • J. Hartwell,

    1. Plant Molecular Science Group, Division of Biochemistry & Molecular Biology, Institute of Biomedical & Life Sciences, Bower Building, University of Glasgow, Glasgow G12 8QQ, UK
    Search for more papers by this author
  • G. I. Jenkins,

    1. Plant Molecular Science Group, Division of Biochemistry & Molecular Biology, Institute of Biomedical & Life Sciences, Bower Building, University of Glasgow, Glasgow G12 8QQ, UK
    Search for more papers by this author
  • H. G. Nimmo

    1. Plant Molecular Science Group, Division of Biochemistry & Molecular Biology, Institute of Biomedical & Life Sciences, Bower Building, University of Glasgow, Glasgow G12 8QQ, UK
    Search for more papers by this author

Correspondence: Professor H. G. Nimmo. Fax: +44 141330 4447; e-mail: H.G.Nimmo@bio.gla.ac.uk

Abstract

We have already reported the cloning of one phosphoenolpyruvate carboxylase kinase gene from Arabidopsis thaliana (Hartwell et al. 1999, Plant Journal 20, 333–342), hereafter termed PPCK1. A second putative phosphoenolpyruvate carboxylase kinase gene (PPCK2) was identified in the Arabidopsis genome. The corresponding cDNA was amplified from flower tissue by reverse transcriptase (RT)-polymerase chain reaction (PCR). This cDNA was transcribed and translated in vitro. The translation products possessed high phosphoenolpyruvate carboxylase kinase activity, confirming the identity of the PPCK2 gene. The expression of both phosphoenolpyruvate carboxylase kinase genes was examined by RT-PCR. PPCK1 is mainly expressed in rosette leaves whereas PPCK2 is expressed in flowers and, at a lower level, in roots and cauline leaves. Light increased the expression of PPCK1 in rosette leaves and that of PPCK2 in flowers. The expression of both genes in an Arabidopsis cell culture was increased by treatment with cycloheximide. The data suggest that the two genes may have different roles in tissue-specific regulation of phosphoenolpyruvate carboxylase.

Introduction

Phosphoenolpyruvate carboxylase (PEPc, EC 4·1.1·31) catalyses the carboxylation of phosphoenolpyruvate to form oxaloacetate and inorganic phosphate. PEPc plays a range of roles in different plant tissues. In the leaves of C4 and CAM plants, it catalyses the primary fixation of atmospheric CO2. In most nonphotosynthetic tissues and in the leaves of C3 plants, it is the major anapleurotic enzyme, allowing the replenishment of TCA cycle intermediates used in biosynthesis. It also plays roles in pH regulation and in the provision of malate in such tissues as guard cells and legume root nodules (e.g. Andreo, Gonzalez & Iglesias 1987; Chollet, Vidal & O’Leary 1996; Vidal & Chollet 1997). In keeping with this diversity of roles, PEPc is encoded by a small gene family, the members of which are expressed in different tissues (e.g. Lepiniec et al. 1994).

In all tissues, PEPc is an allosteric enzyme, inhibited by l-malate and activated by glucose 6-phosphate. Superimposed on this, it is also regulated by reversible phosphorylation of a single strictly conserved Ser residue near the N-terminus of the protein. Phosphorylation reduces the sensitivity of PEPc to malate and increases its sensitivity to glucose 6-phosphate (Chollet et al. 1996; Vidal & Chollet 1997). In CAM and C4 leaves, the phosphorylation state of PEPc is controlled largely by the activity of a Ca2+-independent protein kinase termed PEPc kinase (Vidal & Chollet 1997; Nimmo 2000). This enzyme is itself controlled at the level of expression, in response to a circadian oscillator in CAM species and to light in C4 species (Hartwell et al. 1999a, b).

A similar PEPc kinase activity has been detected in several C3 plants (Chollet et al. 1996; Vidal & Chollet 1997). The malate sensitivity and phosphorylation state of PEPc, and the activity of PEPc kinase in C3 leaves, are increased by light and are sensitive to nitrogen supply (Van Quy, Foyer & Champigny 1991; Van Quy & Champigny 1992; Manh et al. 1993; Duff & Chollet 1995; Chollet et al. 1996; Li, Zhang & Chollet 1996; Smith et al. 1996). These data support the view that phosphorylation of PEPc plays a role in the co-ordination of nitrogen assimilation, CO2 fixation and carbon partitioning by redirecting carbon flow toward the biosynthesis of amino acids (Champigny & Foyer 1992). The stimulation of PEPc kinase activity by light in C3 leaves is blocked by protein synthesis inhibitors (Duff & Chollet 1995; Li et al. 1996; Smith et al. 1996), suggesting that, as in C4 and CAM leaves, expression of PEPc kinase plays a significant role in controlling the phosphorylation state and activity of PEPc. In non-photosynthetic tissues the phosphorylation of PEPc is less well understood. In soybean root nodules PEPc kinase activity seems to be controlled by the availability of photosynthate (Zhang, Li & Chollet 1995), whereas in germinating cereal seeds PEPc becomes phosphorylated after imbibition, even though the kinase is present in dry seed (Osuna et al. 1999).

Hartwell et al. (1999a) reported the first PEPc kinase sequences, from the CAM species Kalanchoë fedtschenkoi and the model C3 plant Arabidopsis thaliana. They showed that this enzyme comprises a protein kinase catalytic domain with essentially no additions. It is a member of the Ca2+/calmodulin-regulated group of protein kinases and is closely related to the calcium-dependent protein kinase (CDPK) family in plants. However, PEPc kinase lacks the auto-inhibitory region and the EF Ca2+-binding hands of plant CDPKs. It also lacks the activatory and inhibitory phosphorylation sites that are found in some other protein kinases, consistent with the view that it is largely controlled by synthesis/degradation (Hartwell et al. 1999a; Nimmo 2000).

In this paper we report the identification of a second PEPc kinase in Arabidopsis thaliana. We term the two genes PPCK1 (reported in Hartwell et al. 1999a) and PPCK2 (reported here), encoding PEPc kinases 1 and 2 (PPCK1 and PPCK2), respectively. In view of the various roles that PEPc may play in different tissues and conditions, this finding opens the possibility that isoenzymes of PEPc kinase play tissue-specific roles in controlling the activity of PEPc. We have therefore studied the expression patterns of both PPCK genes and investigated whether their expression is controlled by light.

Materials and methods

Plant material

The Landsberg erecta ecotype of Arabidopsis thaliana was used for all experiments. Seeds were sown on compost, kept in the dark at 4 °C for 4 d and then transferred to a growth chamber under the following conditions: 14 h light at 22 °C, 10 h dark at 18 °C, relative humidity 60%. Plants were grown for 10 d in low light (photon fluence rate at the top of the plants 100 µmol m−2 s−1) and then transferred to high light (photon fluence rate at the top of the plants 700 µmol m−2 s−1) for 3–5 weeks before use.

Arabidopsis thaliana cells (May & Leaver 1993) were grown photomixotrophically in 500 mL flasks containing 200 mL culture medium (1 × Murashige–Skoog salts with minimal organics, 0·5 mg L−1α-naphthaleneacetic acid, 0·05 mg L−1 kinetin (Sigma, Poola UK), 3% (w/v) sucrose, pH 5·8). Cultures were grown at 20 °C in a continuous low fluence rate of white light (20 µmol m−2 s−1) with constant shaking (110 r.p.m.) and subcultured every week by a 1 : 10 dilution (Christie & Jenkins 1996). For experiments, cell aliquots (10 mL) were transferred aseptically to 50 mL tissue culture flasks on the third day after subculturing. Cells were shaken at 80 r.p.m. for the indicated times. Cycloheximide (Sigma) was dissolved in ethanol to yield stock solutions of 5–100 mm and added to the 10 mL cell aliquots to give the required concentrations. Controls were treated with an equivalent volume of the solvent.

Isolation of RNA and DNA

Plant material was frozen under liquid nitrogen and stored at −80 °C. It was ground under liquid nitrogen using a mortar and pestle and transferred to a sterile 2 mL Eppendorf tube. Cells from the cell culture were collected onto 3MM filter paper (Whatman, Maidstone, UK) by vacuum filtration and frozen under liquid nitrogen.

RNA from Arabidopsis thaliana leaves, flowers and cell culture was extracted using a RNA isolation kit (PUREscript RNA isolation kit; Gentra Systems, Minneapolis, USA) with the addition of a chloroform wash.

RNA from stems and roots was extracted using a protocol adapted from Kay et al. (1987). Extraction buffer (25 mm Tris-HCl pH 8, 25 mm EDTA pH 8, 75 mm NaCl, 1% sodium dodecyl sulphate (SDS), 7·8% 2-mercaptoethanol) and PIC solution [phenol-chloroform-isoamyl alcohol (25 : 24 : 1)] were added to ground plant tissue. After homogenization and centrifugation (10 min at 12 000 × g), the aqueous phase was removed and extracted with 1 volume of PIC solution. Following centrifugation (12 000 × g for 10 min), the aqueous phase was further mixed with one volume of chloroform and centrifuged at 12 000 × g for 10 min. RNA present in the aqueous phase was then precipitated overnight at 4 °C by addition of LiCl to a final concentration of 2 m. The RNA pellet was recovered by centrifugation at 12 000 × g for 10 min and washed with ice-cold 2 m LiCl. The RNA pellet was recovered by centrifugation and resuspended in diethylpyrocarbonate-treated water. It was then precipitated for 30 min at −20 °C after the addition of 0·1 volume of 3 m sodium acetate pH 5·5 and 2·5 volumes of ethanol. The resulting RNA pellet was resuspended in DEPC-treated water.

Following extraction of total RNA, a DNase treatment (DNA-free®, Ambion Huntington, UK) was used to eliminate contamination with genomic DNA.

Genomic DNA from Arabidopsis thaliana was extracted using a DNA isolation kit (PUREgene DNA isolation kit; Gentra Systems).

The quantity and purity of the DNA and RNA were determined spectrophotometrically according to the method described by Sambrook, Fritsch & Maniatis (1989). Intactness of the RNA was determined by agarose gel electrophoresis.

RT-PCR

The total RNA samples (2·5 µg) were mixed with 0·25 µm oligodT for 10 min at 70 °C and cooled at 4 °C. Reverse transcription was carried out in a reaction mixture (25 µL) containing AMV reverse transcriptase buffer (Promega, Scuthamptor, UK), 1 mm dNTPs (Promega), 1 U µL−1 RNase inhibitor (Promega) and 0·4 U µL−1 AMV reverse transcriptase (Promega). The reaction was performed at 48 °C for 45 min The enzyme was then heat-inactivated at 95 °C for 5 min and the samples used directly for PCR.

PCR reactions were performed using 2·5 µL of each cDNA sample in a reaction mixture (25 µL) containing PCR buffer (1×), 0·2 mm dNTPs, 1·5 mm MgCl2, 0·5 µm of each primer and 0·025 U µL−1 of Taq DNA polymerase (Promega). The primer sequences and the predicted product sizes are indicated in Table 1. PCR reactions with primers for PPCK1 were conducted in a programmable thermocycler (PCR Sprint; Hybaid, Ashford, UK) using 23 cycles: 94 °C for 5 min, 94 °C/30 s, 55 °C/30 s, 72 °C/1 min, and a 5 min final extension step at 72 °C. For PPCK2, the annealing temperature was 50 °C and 27 cycles were used. ACT2 primers were used as a constitutive control.

Table 1.  Primers used in RT-PCR analyses and cloning
Primer name and specificitySequence
  1. PCR product sizes: APPCK1,3 and APPCK1,5 = 627 bp; APPCK2,3 and APPCK2,5 = 613 bp; APPCK2s and APPCK2a = 971 and 1045 bp from cDNA and genomic DNA, respectively. Restriction enzyme sites are in italics.

RT-PCR
APPCK1,55′-GTT TTA CGG CGA GAC AG-3′
(PPCK1, 5′ end)
APPCK1,35′-ACA ACT CTG CTT TCT CAC ATC-3′
(PPCK1, 3′ end)
APPCK2,55′-GGT TAT CAG AAT CAG AAT CC-3′
(PPCK2, 5′ end)
APPCK2,35′-TGC TGA AAC ACA CAA CTA CG-3′
(PPCK2, 3′ end)
ACT2a5′-CTT ACA ATT TCC CGC TCT GC-3′
(ACT2, 5′ end)
ACT2s5′-GTT GGG ATG AAC CAG AAG GA-3′
(ACT2, 3′ end)
Cloning
APPCK2a5′-CGT AGG AAG CTC GAG ATT GCT GAA ACA CAC AAC TAC G3′
(PPCK2, 3′ end)
APPCK2s5′-GGA ATA TTC GCG GCC GCC GAT GAC CAG AGA ATT CGA ACT GG3′
(PPCK2, 5′ end)

After amplification, the reaction was resolved by electrophoresis on a 1% agarose gel and stained with ethidium bromide. The PCR products were then quantified by scanning densitometry.

Cloning

For cloning, PCR using the primers APPCK2s and APPCK2a (Table 1) was performed using a cDNA template obtained from Arabidopsis thaliana flower tissue and a genomic DNA template. The 971 and 1045 bp amplification products obtained were digested by NotI and XhoI, purified with the QIAquick gel extraction kit (Qiagen, Crawley, UK) and cloned using the same restriction sites in pBluescript® SK+/– (Novagen, Madison, USA). Two independent clones from genomic DNA (GPPCK2) and two independent clones from flower cDNA (FlPPCK2) were sequenced using universal primers (MWG Biotech Ebersberg, Germany).

In vitro transcription and translation

The clones FlPPCK2 and GPPCK2 were used as templates for in vitro transcription followed by translation as described by Hartwell et al. (1999a). The samples were linearized with XhoI and inserts (1 µg) were transcribed from the T3 promoter using the T3 mMESSAGE mMACHINE kit (Ambion).

Transcribed RNA 1 µg was translated in a rabbit reticulocyte lysate system (Amersham Pharmacia Biotech, UK) using Redivue 35S-methionine (1000 Ci mmol−1, Amersham Pharmacia Biotech Litte Chalfont, UK) as the labelled amino acid. Incubations were for 60 min at 37 °C.

Aliquots (5 µL) of each of the translation mixtures were mixed with 25 µL of dissociation buffer (Laemmli 1970) and boiled for 5 min They were then subjected to SDS-polyacrylamide gel electrphoresis (PAGE) on a 12·5% separating gel. Proteins in the gel were stained with Coomassie blue. RNA was omitted from control reactions. The Mr values of 35S-labelled bands were estimated from the translation products of the control RNA supplied with the lysate.

Assay of PEPc kinase activity

Aliquots (5 µL) of the translation mixtures were used to measure the PEPc kinase activity of the translated products (Hartwell et al. 1996). Samples were incubated with purified dephosphorylated PEPc from Kalanchoë fedtschenkoi and 10 µCi of [γ-32P]ATP (3000 Ci mmol−1, Amersham Pharmacia Biotech). After 30 min incubation at 30 °C, PEPc was isolated by immunoprecipitation and analysed by SDS-PAGE (Hartwell et al. 1996). The incorporation of 32P into PEPc was quantified by phosphorimaging on a Fuji Bio-Imaging Analyser (Fuji Photo Film Co. Ltd, Tokyo, Japan). Controls were carried out with aliquots from translations to which no RNA had been added.

Results

A second PEPc kinase gene in Arabidopsis thaliana

Database searches revealed the presence of a second putative PEPc kinase sequence in Arabidopsis thaliana on BAC clone T27C4 (GenBank accession number AC022287, gene T27C4·19). Detailed inspection of this sequence (version 25/01/2001) suggested that the annotation of T27C4·19 was incorrect and that PPCK2 comprised two exons with an intron from base 67176 to base 67249 of T27C4, close to the 3′ end of the coding sequence. Most Ser/Thr protein kinases contain a consensus motif His.x.aromatic.hydrophobic which is taken to mark the end of the catalytic domain (Hanks & Hunter 1995). In both PPCK1 and PPCK2 this is encoded by the small second exon.

Preliminary RT-PCR experiments showed that PPCK2 was expressed in flowers (see below). To examine both the structure of the gene and the function of its product, we designed PCR primers (Table 1) to amplify fragments of flower cDNA and genomic DNA as illustrated in Fig. 1a. The PCR reactions with these templates each generated a single band with the predicted sizes of 971 bp (flower cDNA) and 1045 bp (genomic DNA) (Fig. 1b). Both bands were cloned and sequenced. The sequences of the cDNA clone (accession number AF358915) and the genomic DNA (accession number AY040830) have been submitted to GenBank. These sequences differ only in the presence of a 74 bp intron in the genomic sequence. The genomic sequence differs from the sequence of BAC T27C4 at 12 nucleotides, of which nine are in exon 1, two are in the intron and one is in the 3′-untranslated region. In addition BAC T27C4 contains an insert of 13 nucleotides in the 3′-untranslated region relative to the sequence reported here (after base 932 of GenBank AY040830). These discrepancies probably reflect ecotype differences between Landsberg erecta (this work) and Columbia (T27C4), and cause only two changes in amino acid sequence. The deduced amino acid sequence of PPCK2 from Landsberg erecta is shown in Fig. 2, aligned with the sequence of Arabidopsis thalianaPPCK1. The predicted PPCK2 protein comprises 278 residues with an Mr of 31201. It is 65·8% identical to PPCK1, and shares the same general features, i.e. it is a protein kinase catalytic domain with minimal extensions. Thus, like the PEPc kinase of K. fedtschenkoi, one might expect both PPCK1 and PPCK2 to be regulated by synthesis/degradation.

Figure 1.

The PPCK2 gene from Arabidopsis thaliana. (a) Structure of the PPCK2 gene. Open bars represent putative exons (E) separated by one intron. The asterisk shows the position of a stop codon within the putative intron in frame with exon 1. The sequences show some of the highly conserved regions in protein kinase catalytic domains (Hanks & Hunter 1995). The filled bars indicate the primers used to amplify the genomic and cDNA clones of PPCK2. (b) Amplification of PPCK2 genomic and cDNA clones from Arabidopsis thaliana. PCR products amplified with primers APPCK2s and APPCK2a from genomic DNA (lane 2) and flower cDNA (lane 3) are indicated by the bars. Lane 1 shows a DNA ladder with band sizes shown in bp. (c) Differences between the C-terminal sequences of the open reading frames in (1) the genomic and (2) the cDNA clones of PPCK2. The underlining indicates the codon that is interrupted by the intron in the genomic sequence.

Figure 2.

Amino acid sequences of PEPc kinases 1 and 2 from Arabidopsis thaliana. The PPCK1 sequence is from ecotype Columbia (Hartwell et al. 1999a); the PPCK2 sequence is from ecotype Landsberg erecta. The sequence of PPCK2 from ecotype Columbia differs from the sequence shown here only at positions 48 (I for T) and 123 (S for R). Dots indicate identities.

The cDNA and genomic PCR products from PPCK2 were cloned behind the T3 promoter in pBluescript using the NotI and XhoI restriction sites built into the primers. Following linearization with XhoI, the inserts were transcribed and translated, and the translation products were assayed for PEPc kinase activity. Figure 3a shows that in both cases translation products of the expected size (about 31 kDa) were generated. These products should differ only at the C-terminal end as shown in Fig. 1c. However, only the translation product from the cDNA clone was able to phosphorylate PEPc (Fig. 3b). This confirms the identification of this sequence as a functional PEPc kinase. Although the sequence His.x.aromatic.hydrophobic is highly conserved in Ser/Thr protein kinases, no particular function has been ascribed to it. The data shown here demonstrate that the sequence encoded by the small second exon in PPCK2, which starts with this consensus sequence, is important for function.

Figure 3.

The PPCK2 cDNA encodes a functional PEPc kinase. Panel (a) shows phosphorimages of [35S]Met-labelled products from in vitro translation of RNA samples, separated on a 12·5% SDS polyacrylamide gel. Lane 1, no RNA; lane 2, control RNA used to provide molecular weight markers (Mr values are shown × 10−3); lane 3, RNA transcribed from the cDNA clone FlPPCK2; lane 4, RNA transcribed from the genomic clone GPPCK2. The bar marks the position of the translation products. Panel (b) shows 32P-labelled PEPc immunoprecipitated after assays of translation products for PEPc kinase, separated on an 8% SDS polyacrylamide gel. Lane 1, no RNA; lane 2, translation products from FlPPCK2 RNA; lane 3, translation products from GPPCK2 RNA. The bar marks the position of PEPc.

Expression of PEPc kinase in different tissues of Arabidopsis thaliana

We were unable to detect any transcripts from either PPCK1 or PPCK2 using homologous probes and Northern blotting, indicating that both genes are only expressed at low levels. However expression was detected in RT-PCR experiments using primers specific for the two PPCK genes. Primers for ACT2 were used as a constitutive control. Conditions were chosen to ensure that the amount of product increased with cycle number so, for each gene, relative transcript levels in different organs can be compared semiquantitatively. However, since different RT-PCR conditions were used for the two genes their relative expression in each organ cannot be compared directly.

Transcripts from PPCK1 were detectable in rosette leaves, and, to a lesser extent, in flowers and roots but only barely detectable in siliques, cauline leaves and stems (Fig. 4a). For PPCK2, transcripts were almost undetectable in rosette leaves and stems but were clearly detectable in roots and flowers and, at a lower level, in cauline leaves (Fig. 4b). Thus the two genes show different expression patterns.

Figure 4.

Expression of PPCK1 and PPCK2 in different tissues of Arabidopsis thaliana. RT-PCR products from total RNA isolated from (1) roots (2) stems (3) cauline leaves (4) flowers (5) rosette leaves and (6) siliques, using gene-specific primers for (a) PPCK1 and ACT2, and (b) PPCK2 and ACT2.

Regulation of the expression of PEPc kinase

RT-PCR with gene-specific primers was used to measure the abundance of transcripts for PPCK1 in rosette leaves and for PPCK2 in flowers in the middle of the light and dark periods of the diurnal cycle. The data (Fig. 5) show that both transcripts are light-regulated; their levels are about 2·5-fold higher in the light period than in the dark period. The light regulation of PPCK1 was observed in both young and mature rosette leaves. Under continuous light, no circadian rhythm in transcript abundance was detected for either gene (data not shown).

Figure 5.

Light-stimulated expression of PPCK1 and PPCK2 in Arabidopsis thalianaRT-PCR products from total RNA, isolated from (a) rosette leaves and (b) flowers, in the middle of the light or the dark period, using gene-specific primers for (a) PPCK1 and ACT2, and (b) PPCK2 and ACT2.

Hartwell et al. (1999b) reported that treatment with cycloheximide enhanced the light-dependent increase in PEPc kinase translatable mRNA in two different species, maize and barley. To investigate whether a similar phenomenon occurs in Arabidopsis thaliana, we used a cell culture system which expresses both PPCK genes. Addition of the protein synthesis inhibitor resulted in a large increase in the transcript abundance for both PPCK1 (3- to 5-fold compared to the control) and PPCK2 (2·5-fold compared to the control) (Fig. 6). Control experiments (not illustrated) showed that addition of ethanol alone had no affect on the expression of the PPCK genes. The increase in transcript abundance in response to cycloheximide was rapid, being maximal after 30 min, and transient; it was still detectable at 8 h after addition of cycloheximide but not at 24 h after addition (data not shown).

Figure 6.

Expression of PPCK1 and PPCK2 in Arabidopsis thaliana cells treated with cycloheximide. A 3-day-old cell suspension culture was treated with cycloheximide (CHX, 0, 10 or 20 µm). RNA was extracted 30 min and 1 h after the addition of cycloheximide. RT-PCR was performed with gene-specific primers for (a) PPCK1 and ACT2 and (b) PPCK2 and ACT2.

Discussion

The work presented here demonstrates clearly that Arabidopsis thaliana contains a second functional and expressed PEPc kinase gene, termed PPCK2. Preliminary Southern analysis suggested the existence of two PEPc kinase genes in Mesembryanthemum crystallinum (Taybi et al. 2000). Furthermore, analysis of EST databases shows that several species, such as tomato and soybean, contain two or more putative PEPc kinases. Hence it seems likely that many, if not all, plants contain a small PEPc kinase gene family. Our failure to detect multiple genes in our earlier analysis of Arabidopsis and K. fedtschenkoi (Hartwell et al. 1999a) may have resulted from the use of too high a stringency in our Southern analysis. There is biochemical evidence for the existence of two types of PEPc kinase in several plant species (e.g. Chollet et al. 1996; Vidal & Chollet 1997) with Mr values in the ranges 30–32 000 and 37–39 000. However both PPCK1 and PPCK2 encode proteins with Mr values of approximately 31 000. This work therefore throws no light on the possible origin of the Mr 37–39 000 form; Arabidopsis does not appear to contain a gene directly encoding such a form.

Both PPCK genes in Arabidopsis are expressed at a relatively low level; unlike the PEPc kinase gene in mature K. fedtschenkoi leaves (Hartwell et al. 1999a), their transcripts can be detected only by RT-PCR. The two genes show clear differences in tissue expression pattern. PPCK1 is most highly expressed in rosette leaves, whereas PPCK2 is most highly expressed in flowers and roots; PPCK2 expression in rosette leaves is very low relative to these organs (Fig. 4). Given the different metabolic roles of PEPc in different tissues and conditions, this suggests that the two PPCK genes may have somewhat different roles. However in order to assess whether there are any important functional differences between the products of the two PPCK genes, it will be necessary to test the activities of the two expressed proteins against the various isoenzymes of PEPc in Arabidopsis.

Previous work has demonstrated that light can cause increases in the activity of PEPc kinase in the leaves of such C3 species as tobacco and wheat (Duff & Chollet 1995; Li et al. 1996). This is thought to play a role in the co-ordination of N and C metabolism. Clearly, expression of PPCK1 is induced by light in rosette leaves (Fig. 5); this suggests that one of the main functions of PEPc kinase 1 may lie in this aspect of metabolic control. It remains to be seen whether the expression of PPCK1 is also sensitive to nitrogen supply. The expression of PPCK2 in flower tissue is also stimulated by light (Fig. 5). The specific tissue involved is not known, but we are currently constructing transgenic plants expressing promoter : reporter fusions to address this question with more precision. Unlike the PEPc kinase gene of K. fedtschenkoi, neither of the Arabidopsis genes seems to be under circadian control. However, the marked changes in transcript levels caused by light and by cycloheximide suggest that, as in K. fedtschenkoi (Hartwell et al. 1999a), PEPc kinase activity in Arabidopsis is controlled largely at the level of gene expression.

The light-induced increase in the level of PEPc kinase translatable mRNA in both maize and barley is enhanced by cycloheximide (Hartwell et al. 1999b). In this work we chose to test the effects of cycloheximide on the expression of the PPCK genes using a cell culture, because it expresses both genes and is well suited to a pharmacological approach. Cycloheximide caused a marked increase in the expression of both genes, in both light (Fig. 6) and darkness (not shown). This is not a general effect as other genes respond differently to treatment of the cells with cycloheximide (Christie & Jenkins 1996). Several explanations of our results can be proposed. It is clear that the PEPc kinase mRNA in K. fedtschenkoi turns over very rapidly (Hartwell et al. 1999a). If the same is true in Arabidopsis, the PPCK transcripts may be destroyed by an RNAase which itself turns over rapidly. The effect of cycloheximide could be due to the lack of synthesis of this RNAase. As we have suggested for CAM species (Borland et al. 1999; Nimmo 2000), expression of PEPc kinase in Arabidopsis and the consequent activation of PEPc might lead to changes in metabolite levels which down-regulate the kinase. By preventing synthesis of the kinase proteins, cycloheximide could lead to increased expression of the PPCK genes via an effect on metabolism. It is also possible that the addition of cycloheximide causes cellular stress which in turn activates expression of the PPCK genes. However, we did not detect any increase in the expression of the PPCK genes in either drought or high light intensity (data not shown). Further work is therefore required to define the mechanism underlying the induction of PPCK gene expression by cycloheximide.

Acknowledgments

This work was supported by BBSRC. We thank Mr J. Jardine for technical assistance.

Received 29 June 2001;received inrevised form 24 August 2001;accepted for publication 24 August 2001

Ancillary