Phosphoenolpyruvate (PEP) carboxylase plays a number of key roles in the central metabolism of higher plants. The enzyme is regulated by reversible phosphorylation in response to a range of signals in many different plant tissues. The data discussed here illustrate several novel features of this system. The phosphorylation state of PEP carboxylase is controlled largely by the activity of PEP carboxylase kinase. This enzyme comprises a protein kinase catalytic domain with no regulatory regions. In many systems it is controlled at the level of expression. In C4 plants, expression of PEP carboxylase kinase is light-regulated and involves changes in cytosolic pH, InsP3 and Ca2+ levels. Expression of PEP carboxylase kinase in CAM plants is regulated by a circadian oscillator, perhaps via metabolite control. Some plants contain multiple PEP carboxylase kinase genes, probably with different expression patterns and roles. A newly discovered PEP carboxylase kinase inhibitor protein might facilitate the net dephosphorylation of PEP carboxylase under conditions in which flux through this enzyme is not required.
Phosphoenolpyruvate carboxylase (PEPc) (EC 126.96.36.199) catalyses the carboxylation of PEP to yield oxaloacetate and Pi. The enzyme plays a wide range of metabolic roles in higher plants (Chollet et al., 1996). It catalyses the first step in the photosynthetic assimilation of CO2 in Crassulacean acid metabolism (CAM) and C4 plants. It is the major anaplerotic enzyme in most higher plant tissues, and plays specialized roles in the formation of malate in guard cells during stomatal opening and in N2-fixing legume root nodules. PEPc is an allosteric enzyme, inhibited by malate and activated by glucose 6-phosphate. However, these effects are modulated by phosphorylation of one serine residue per subunit, which activates PEPc in vivo by reducing its sensitivity to inhibition by malate. Recent work has established PEPc as perhaps the best understood example of protein phosphorylation in the control of plant metabolism (Chollet et al., 1996; Vidal & Chollet, 1997; Nimmo, 2000).
Phosphorylation of PEPc was first identified in the CAM species Kalanchoë fedtschenkoi. CAM is a metabolic adaptation to arid environments, in which stomata are closed during much of the day and open at night (Osmond & Holtum, 1981). It is typified by a temporal separation of two phases of CO2 fixation. During the night, the primary fixation of CO2 is catalysed by PEPc. This results in the formation of malic acid, which is stored in the vacuole. During the following day, the malate is released from the vacuole and decarboxylated, and the resulting CO2 is fixed via the Calvin cycle. To combat futile cycling between phosphoenolpyruvate and malate, a mechanism must exist to permit flux through PEPc at night and reduce or eliminate it during the day. The first studies of this mechanism derived from the finding that PEPc was some 10-fold more sensitive to inhibition by malate during the light period (apparent Ki for malate 0.3 mM) than in the middle of the dark period (apparent Ki 3.0 mM) (Nimmo et al., 1984). Purification of the two forms showed that the ‘night’ form was phosphorylated (on serine residues) and the ‘day’ form was dephosphorylated (Nimmo et al., 1986). The conversions of the ‘night’ to the ‘day’ form of PEPc and vice versa coincided with the cessation and onset of malate accumulation, respectively, suggesting that these interconversions are very important in regulating the flux through PEPc (Nimmo et al., 1984). Both conversions occurred during the dark period of the normal diurnal cycle (Nimmo et al., 1984). Conversions between the two forms of the enzyme were also observed in detached leaves kept in constant environmental conditions, confirming that the phosphorylation of PEPc in CAM plants is controlled by a circadian rhythm rather than by light/dark transitions (Nimmo et al., 1987b). Indeed, it has been suggested (Nimmo, 1998) that this phosphorylation controls the amplitude of the well-studied (Wilkins, 1992) circadian rhythms of CO2 fixation in these plants.
The importance of the phosphorylation of PEPc in CAM has been confirmed in a detailed study in which Kalanchoë daigremontiana (a constitutive CAM plant) and Clusia minor (a C3-CAM intermediate) adapted to an 11-h photoperiod were assessed by gas exchange and instantaneous isotope discrimination analysis (Borland & Griffiths, 1997). This allowed the in vivo flux through PEPc to be estimated; in both species, there was a good correlation between the estimated flux and the phosphorylation state of the enzyme, as judged by its malate sensitivity. Phosphorylation of PEPc has also been reported in several other CAM species (Brulfert et al., 1986; Kluge et al., 1988; Baur et al., 1992; Weigend, 1994).
In C4 plants, PEPc must achieve rapid CO2 fixation during the day in the face of the high concentration of malate in the mesophyll tissue that is required to maintain the necessary diffusion to the bundle sheath tissue. Several groups noted that C4 PEPc could be interconverted between two kinetically distinct forms (Karabourniotis et al., 1983; Huber & Sugiyama, 1986; Doncaster & Leegood, 1987; Nimmo et al., 1987a; Jiao & Chollet, 1988); for example, illumination reduced the malate sensitivity of the enzyme some two- to threefold (Nimmo et al., 1987a). Using different approaches, two groups then showed that PEPc was phosphorylated during the day but dephosphorylated at night (Nimmo et al., 1987a; Jiao & Chollet, 1988). The importance of the light-dependent phosphorylation of C4 PEPc has been demonstrated by studies (Bakrim et al., 1993) in which phosphorylation was prevented by pretreatment of leaves with the protein synthesis inhibitor cycloheximide, which blocks induction of PEPc kinase (see below). Cycloheximide markedly reduced the phosphorylation state of PEPc (as judged by its malate sensitivity) and CO2 fixation in parallel in Sorghum leaves, but had no effect on wheat leaves over the same time period.
The phosphorylation state of a protein in vivo represents the steady-state balance between the activities of the relevant protein kinase(s) and phosphatase(s), and can be regulated by changes in the activities of either or both of the kinase(s) and phosphatase(s). The enzyme responsible for the dephosphorylation of PEPc in K. fedtschenkoi was identified as protein phosphatase 2 A (PP2A) (Carter et al., 1990). Maize PEPc is also dephosphorylated by this enzyme (McNaughton et al., 1991). The activity of PP2A did not change significantly during the normal diurnal cycle in K. fedtschenkoi, nor was it affected by illumination in maize (Carter et al., 1991; McNaughton et al., 1991). This led to the suggestion that the phosphorylation state of PEPc might be regulated largely through changes in the activity of PEPc kinase. However, in this work PP2A was assayed using either casein or rabbit phosphorylase a as a substrate. This gives a measure of total PP2A activity, but it remains possible that the PEPc phosphatase activity is a subset of PP2A in which the phosphatase catalytic subunit is complexed to a regulatory subunit which may itself be a target for control.
Parallel studies with CAM and C4 plants showed not only that the activity of PEPc kinase is controlled in an unusual way but also that it plays a key role in controlling the phosphorylation state of PEPc. The nocturnal phosphorylation of PEPc in CAM plants and the light-induced phosphorylation in C4 plants could be prevented by pretreatment with inhibitors of cytosolic protein synthesis (Carter et al., 1991; Jiao et al., 1991). Calcium-independent PEPc kinase activity was greatly increased during the night in CAM and by light in C4; moreover, these increases required protein synthesis (Echevarria et al., 1990; Carter et al., 1991; Jiao et al., 1991; McNaughton et al., 1991). Similar findings have been made more recently in other systems (Vidal & Chollet, 1997).
Collectively, these data indicated that the phosphorylation state of PEPc is largely controlled by the tissue activity of PEPc kinase. They also showed that, unlike most protein kinases which are regulated by second messengers or phosphorylation cascades, PEPc kinase is regulated by a protein synthesis step in most systems. However there is at least one exception to this general rule: the phosphorylation of PEPc in barley seed increases following imbibition but this is not prevented by cycloheximide and is not associated with an increase in PEPc kinase activity in the seed (Osuna et al., 1999). The factors controlling the phosphorylation state of PEPc in this system remain to be identified.
PEPc kinase is controlled at the level of expression
The data reported above left open various possibilities as to the nature of the protein whose synthesis was required for the appearance of kinase activity. The simplest hypothesis was that the component synthesized is PEPc kinase itself. However it was also possible that PEPc kinase was present constitutively, in an inactive form, and that the protein synthesis step represented production of a protein that activated PEPc kinase, perhaps via reversible phosphorylation. Further analysis was slow because PEPc kinase is a very low abundance protein, and cloning of its gene (from any plant species) proved very difficult.
The development of a method for measurement of the amount of translatable mRNA for the kinase led to a significant advance (Hartwell et al., 1996). The method is based on in vitro translation of isolated RNA, followed by direct assay of the translation products for kinase activity. Use of this assay showed that the level of PEPc kinase translatable mRNA in K. fedtschenkoi varied by some 20-fold over the diurnal cycle, oscillated in constant environmental conditions, and determined PEPc kinase activity. Moreover, increases in kinase translatable mRNA were blocked by inhibitors of RNA synthesis (Hartwell et al., 1996). These results strongly suggested that CAM PEPc kinase activity is regulated by circadian control of kinase gene expression.
The first PEPc kinase cDNA was cloned by adapting the assay for kinase translatable mRNA to assess its level in pools and subpools of a cDNA library from the CAM plant K. fedtschenkoi (Hartwell et al., 1999a). Successive subdivisions of the library led to isolation of a single clone that expressed a protein with very high Ca2+-independent PEPc kinase activity; this protein was shown immunologically to be the major or only PEPc kinase in leaf extracts. Peptide mapping showed that this protein kinase phosphorylated the same site on PEPc that is phosphorylated in vitro by highly purified PEPc kinase and is also phosphorylated in vivo at night (Hartwell et al., 1999a).
Sequencing of the PEPc kinase cDNA showed that PEPc kinase comprises a protein kinase catalytic domain with minimal additions (Hartwell et al., 1999a). It is a member of the CaMK group (Hanks & Hunter, 1995) of protein kinases, which comprises kinases regulated by Ca2+/calmodulin and close relatives. The enzyme is related to the higher plant calcium-dependent protein kinase (CDPK) family but lacks the autoinhibitory region and the EF hands of the CDPKs. A similar PEPc kinase cDNA has recently been identified in the inducible CAM species Mesembryanthemumcrystallinum as a salt- and dark-upregulated gene whose expression shows a circadian rhythm (Taybi et al., 2000). Interestingly, the K. fedtschenkoi PEPc kinase gene contains a single intron very close to the end of the coding sequence; the catalytic domains of Arabidopsis CDPKs contain three introns, one of which is in the same position as the PEPc kinase intron (Hartwell et al., 1999a).
In K. fedtschenkoi the expression of PEPc kinase is controlled both developmentally and by a circadian oscillator (Hartwell et al., 1999a). The abundance of kinase transcripts in leaves sampled in the middle of the dark period increases with leaf age, in parallel with the development of CAM. In mature leaves, the level of kinase transcripts is high in the middle of the dark period and very low during the day. It matches the level of kinase-translatable mRNA, and controls the activity of PEPc kinase and the phosphorylation state of PEPc. Moreover, kinase transcript abundance in mature leaves placed in constant conditions exhibits a circadian rhythm (Hartwell et al., 1999a).
No PEPc kinase gene from a C4 plant has yet been cloned. However there is strong indirect evidence that the light-induction of PEPc kinase in maize represents a direct effect on expression. The amount of PEPc kinase translatable mRNA in maize leaves was increased by illumination (Hartwell et al., 1996). This was prevented by pretreatment of the leaves with RNA synthesis inhibitors. However pretreatment with protein synthesis inhibitors had exactly the opposite effect – it actually enhanced the light-dependent increase in kinase translatable mRNA (Hartwell et al., 1999b). This rules out the possibility that the effects of light on PEPc kinase translatable mRNA are mediated by another protein which is induced by light and which increases the translatability of the kinase mRNA: if this were the case, the inhibitors would block the increase in translatable kinase mRNA rather than enhance it.
A small family of PEPc kinase genes
As noted above, PEPc plays many different roles in plant metabolism. Higher plants contain several closely related PEPc genes. For example, ‘housekeeping’, ‘root’ and ‘photosynthetic’ PEPc genes have been defined in the C4 species Sorghum (Lepiniec et al., 1993). One might therefore expect multiple PEPc kinase genes, perhaps controlled in different ways. Indeed, activity staining of renatured SDS gels has provided evidence for the existence of at least two forms of Ca2+-independent PEPc kinase; several plant species contain two PEPc kinase polypeptides with Mr values of 30–33 000 and 37–39 000 (Vidal & Chollet, 1997). A PEPc kinase gene (now termed PEPcK I) was initially identified in the model C3 species Arabidopsis thaliana by virtue of its sequence similarity to K. fedtschenkoi PEPc kinase; this is gene 13 on BAC F22O13, GenBank accession number AC003981. The protein produced by transcription and translation of a full-length cDNA does indeed have Ca2+-independent PEPc kinase activity (Hartwell et al., 1999a). It should be noted that the current (September 2000) GenBank annotation for F22O13.13 is incorrect. There is an intron from base 50934 to base 51105 inclusive, and the coding sequence extends to a stop codon at bases 51153–5. More recently, a second gene (PEPcK II) has been identified in a similar way; this is gene T27C4.19 on BAC T27C4 (GenBank accession number AC022287) (V. Fontaine, J. Hartwell, G. I. Jenkins & H. G. Nimmo, unpublished). However the annotation of this BAC sequence (version 25/01/2001) is also incorrect. There is an intron from base 67176 to base 67249 and the stop codon is bases 67291–67293. The introns in these Arabidopsis PEPc kinase genes interrupt the same codon as the K. fedtschenkoi PEPc kinase intron.
Both of the Arabidopsis PEPc kinase genes encode proteins with Mr values of approx. 31000. There is also evidence for the existence of at least two PEPc kinase genes in the inducible CAM species M. crystallinum (Taybi et al., 2000). Multiple PEPc kinase genes have been identified in tomato and soybean through analysis of expressed sequence tags (our unpublished data) but no gene encoding a 37–39 000 Mr form has yet been detected. The nature and origin of the higher Mr form of PEPc kinase is not known. It remains possible that this form results from post-translational modification of a lower Mr form, perhaps by ubiquitination as part of the degradation process.
Both Arabidopsis PEPc kinase genes are expressed at relatively low levels; unlike in mature K. fedtschenkoi leaves, kinase transcripts in Arabidopsis cannot be detected by Northern analysis. However, we have detected transcripts using semiquantitative RT-PCR with gene-specific primers. This has shown that the two genes differ in their expression patterns. PEPcK I is expressed primarily in rosette leaves whereas PEPcK II is expressed in flowers, roots and cauline leaves but not rosette leaves. The expression of PEPcK I in rosette leaves is increased by light (V. Fontaine, J. Hartwell, G. I. Jenkins & H. G. Nimmo, unpublished). Thus the expectation of multiple PEPc kinase genes with different expression patterns and functions seems to be fulfilled.
Control of PEPc kinase expression
Early work on the signalling processes involved in the light-stimulated phosphorylation of PEPc in C4 plants implicated an intercellular metabolic message and Ca2+ movements (Chollet et al., 1996). More recently, great advances have been made using mesophyll cell protoplasts of the C4 species Digitaria sanguinalis (Giglioli-Guivarc’h et al., 1996, Coursol et al., 2000). Illumination stimulated PEPc kinase and increased the phosphorylation state of PEPc provided that the protoplasts were incubated in the presence of NH4Cl or methylamine, which increase the pH of the cytosol. These weak bases could be replaced by 3-phosphoglycerate, which increases cytosolic pH in the light as it is taken up into the mesophyll chloroplasts. The light- and weak base-mediated activation of PEPc kinase was reduced by TMB-8, which blocks InsP3-gated Ca2+ channels, and the calmodulin antagonist W7 (Giglioli-Guivarc’h et al., 1996). It has now been shown that light and weak base treatment of protoplasts causes a rapid, transient increase in InsP3 content. This was blocked by U-73122 which inhibits phosphoinositide-specific phospholipase C activity. This inhibitor also blocked the increased phosphorylation of PEPc and markedly reduced the activation of PEPc kinase in the protoplasts (Coursol et al., 2000). Thus the light signal transduction chain in C4 plants is thought to include increases in cytosolic pH, InsP3 and Ca2+. However the mode of connection of this transduction chain to the expression of the PEPc kinase gene is not known.
As noted above, the expression of PEPc kinase in CAM leaves is controlled by a circadian oscillator. The nocturnal appearance of PEPc kinase activity in K. fedtschenkoi is reduced by W7 (Fig. 1). This suggests that the pathway initiated by the circadian oscillator involves a Ca2+/calmodulin–like interaction. Recent work has focused on the possibility that this circadian control can be modified by metabolic status. K. daigremontiana leaves were prevented from accumulating malate by enclosure in an atmosphere of N2 during the night. This increased the amplitude and duration of the bout of CO2 fixation by PEPc in normal air the following morning, because PEPc kinase mRNA and activity, and the phosphorylation state of PEPc, all remained high for significantly longer than in control leaves (Borland & Griffiths, 1997; Borland et al., 1999). Subjecting leaves to a temperature increase in the middle of the night caused a rapid disappearance of PEPc kinase mRNA and activity in control leaves (Borland et al., 1999). This treatment is thought to lead to efflux of malate from the vacuole (Freimert et al., 1988; Wilkins, 1992). However, in leaves which had been prevented from accumulating malate, kinase mRNA actually increased slightly. This led to the proposal that a metabolite, possibly cytosolic malate, can act as a feedback regulator of the expression of PEPc kinase and over-ride circadian control (Borland et al., 1999). Indeed, it has now been suggested that a primary target of the circadian oscillator in CAM plants may be the permeability of the tonoplast (Nimmo, 2000). This could result in a circadian oscillation in the cytosolic concentration of malate or other metabolites. The control of PEPc kinase expression could simply be a secondary effect.
The expression of PEPc kinase may be affected by metabolite control in other systems. As noted above, RNA synthesis inhibitors blocked the light-induced increase in PEPc kinase translatable mRNA in maize but protein synthesis inhibitors actually enhanced the increase (Hartwell et al., 1999b). Similar results obtained with the C3 species barley are shown in Fig. 2. Since protein synthesis inhibitors prevent the appearance of PEPc kinase activity and reduce flux through PEPc, one possible explanation is that a metabolic product of this flux acts as a feedback inhibitor of PEPc kinase expression. Alternatively, a metabolite upstream of PEPc, such as PEP, could act as a feedforward activator of kinase expression. Such a mechanism could be involved in the induction of PEPc kinase by photosynthate in legume root nodules.
Control of PEPc kinase activity
Phosphorylation of PEPc by PEPc kinase in vitro is inhibited by malate and activated by glucose 6-phosphate (Carter et al., 1991; Wang & Chollet, 1993). These effects may be mediated by binding of the effectors either to PEPc or to the kinase. It is not yet clear whether they are physiologically significant. Recent work has revealed another mechanism by which the in vivo activity of PEPc kinase may be regulated. The specific activity of PEPc kinase assayed in crude extracts increases with dilution. This is because plant tissues contain a protein which reversibly inhibits PEPc kinase (Nimmo et al., 2001). This protein is not an ATPase, a protease or a PEPc phosphatase. It does not affect the phosphorylation of PEPc by cyclic AMP-dependent protein kinase (which occurs at the site that is phosphorylated by PEPc kinase), nor does it affect other protein kinases such as casein kinase I. Thus, the inhibitor seems to interact specifically with PEPc kinase rather than with PEPc. The amounts of kinase and inhibitor in leaves were estimated following separation by hydrophobic chromatography. The amount of inhibitor in K. fedtschenkoi is sufficient to inhibit the low amount of kinase activity present during the light period and the early stages of the dark period. Similarly, the amount of inhibitor in maize is sufficient to inhibit the low amount of kinase activity present in the dark and at low light intensity. Analogous to the role of the protein inhibitor of mammalian cyclic AMP-dependent protein kinase, the function of the PEPc kinase inhibitor protein may be to inhibit the low amount of kinase present in conditions under which rapid flux through PEPc is not required.
Studies of changes in the kinetic properties of PEPc in CAM plants led to the discovery that PEPc is controlled by reversible phosphorylation. It is now clear that PEPc is controlled by phosphorylation in many different plant tissues, in response to a range of signals. The data discussed here illustrate several novel features of this system. The phosphorylation state of PEPc is mainly controlled by changes in PEPc kinase activity. This unusual enzyme comprises a protein kinase catalytic domain with no regulatory regions, and in most systems it is controlled largely at the level of expression. A newly discovered inhibitor protein may help to facilitate the net dephosphorylation of PEPc under conditions in which flux through PEPc is not required.
Work from this laboratory has been supported by the Biotechnology and Biological Sciences Research Council.