Phosphorylation and 14-3-3 binding of Arabidopsis 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase


  • Anna Kulma,

    1. MRC Protein Phosphorylation Unit, School of Life Sciences, MSI/WTB Complex, University of Dundee, Dundee DD1 5EH, Scotland, UK, and
    Search for more papers by this author
      These authors made equally important contributions to this work.
  • Dorthe Villadsen,

    1. Plant Biochemistry Laboratory, Department of Plant Biology, The Royal Veterinary and Agricultural University, 40 Thorvaldsensvej, DK-1871 Frederiksberg C, Copenhagen, Denmark
    Search for more papers by this author
      Present address: Department of Cell Biology and Physiology/Howard Hughes Medical Institute, Washington University School of Medicine, Box 8228, 660 S. Euclid Ave., St Louis, MO 63110, USA. These authors made equally important contributions to this work.
  • David G. Campbell,

    1. MRC Protein Phosphorylation Unit, School of Life Sciences, MSI/WTB Complex, University of Dundee, Dundee DD1 5EH, Scotland, UK, and
    Search for more papers by this author
  • Sarah E. M. Meek,

    1. MRC Protein Phosphorylation Unit, School of Life Sciences, MSI/WTB Complex, University of Dundee, Dundee DD1 5EH, Scotland, UK, and
    Search for more papers by this author
      These authors made equally important contributions to this work.
  • Jean E. Harthill,

    1. MRC Protein Phosphorylation Unit, School of Life Sciences, MSI/WTB Complex, University of Dundee, Dundee DD1 5EH, Scotland, UK, and
    Search for more papers by this author
  • Tom H. Nielsen,

    1. Plant Biochemistry Laboratory, Department of Plant Biology, The Royal Veterinary and Agricultural University, 40 Thorvaldsensvej, DK-1871 Frederiksberg C, Copenhagen, Denmark
    Search for more papers by this author
  • Carol MacKintosh

    Corresponding author
    1. MRC Protein Phosphorylation Unit, School of Life Sciences, MSI/WTB Complex, University of Dundee, Dundee DD1 5EH, Scotland, UK, and
      For correspondence (fax +44 1382 223778; e-mail
    Search for more papers by this author

For correspondence (fax +44 1382 223778; e-mail

Present address: Institute of Cell and Molecular Biology, University of Edinburgh, Darwin Building, King's Buildings, Edinburgh EH9 3JR, UK.

Present address: Department of Cell Biology and Physiology/Howard Hughes Medical Institute, Washington University School of Medicine, Box 8228, 660 S. Euclid Ave., St Louis, MO 63110, USA.

These authors made equally important contributions to this work.


Fructose 2,6-bisphosphate (fru-2,6-P2) is a signalling metabolite that regulates photosynthetic carbon partitioning in plants. The content of fru-2,6-P2 in Arabidopsis leaves varied in response to photosynthetic activity with an abrupt decrease at the start of the photoperiod, gradual increase through the day, and modest decrease at the start of the dark period. In Arabidopsis suspension cells, fru-2,6-P2 content increased in response to an unknown signal upon transfer to fresh culture medium. This increase was blocked by either 2-deoxyglucose or the protein phosphatase inhibitor, calyculin A, and the effects of calyculin A were counteracted by the general protein kinase inhibitor K252a. The changes in fru-2,6-P2 at the start of dark period in leaves and in the cell experiments generally paralleled changes in nitrate reductase (NR) activity. NR is inhibited by protein phosphorylation and binding to 14-3-3 proteins, raising the question of whether fructose-6-phosphate,2-kinase/fructose-2,6-bisphosphatase protein from Arabidopsis thaliana (AtF2KP), which both generates and hydrolyses fru-2,6-P2, is also regulated by phosphorylation and 14-3-3s. Consistent with this hypothesis, AtF2KP and NR from Arabidopsis cell extracts bound to a 14-3-3 column, and were eluted specifically by a synthetic 14-3-3-binding phosphopeptide (ARAApSAPA). 14-3-3s co-precipitated with recombinant glutathione S-transferase (GST)-AtF2KP that had been incubated with Arabidopsis cell extracts in the presence of Mg-ATP. 14-3-3s bound directly to GST-AtF2KP that had been phosphorylated on Ser220 (SLSASGpSFR) and Ser303 (RLVKSLpSASSF) by recombinant Arabidopsis calcium-dependent protein kinase isoform 3 (CPK3), or on Ser303 by rat liver mammalian AMP-activated protein kinase (AMPK; homologue of plant SNF-1 related protein kinases (SnRKs)) or an Arabidopsis cell extract. We have failed to find any direct effect of 14-3-3s on the F2KP activity in vitro to date. Nevertheless, our findings indicate the possibility that 14-3-3 binding to SnRK1-phosphorylated sites on NR and F2KP may regulate both nitrate assimilation and sucrose/starch partitioning in leaves.


Fructose 2,6-bisphosphate (fru-2,6-P2) is a signal metabolite that regulates the rate of interconversion between fructose 6-phosphate (fru-6-P) and fru-1,6-P2 in the cytosol of eukaryotic cells (Dihazi et al., 2001; Okar et al., 2001; Stitt, 1990). In animal and yeast cells, fru-2,6-P2 stimulates glycolysis and inhibits gluconeogenesis (Dihazi et al., 2001; Okar et al., 2001). In plants, fru-2,6-P2 was proposed to regulate synthesis of sucrose by inhibiting the cytosolic isoform of fructose-1,6-bisphosphatase (F16BPase; Stitt, 1990). Consistent with this proposal, in transgenic plants with decreased foliar fru-2,6-P2 the synthesis of sucrose is promoted, whereas an increase in the level of fru-2,6-P2 stimulates the flux towards starch (Draborg et al., 2001; Theodorou and Kruger, 2001; Truesdale et al., 1999). Fru-2,6-P2 also activates pyrophosphate: fructose-6-phosphate 1-phosphotransferase (PFP) (but apparently not 6-phosphofructo-1-kinase (F6P,1K), as in mammals). However, the role of PFP in metabolism remains to be elucidated (Stitt, 1998).

Both the synthesis and degradation of fru-2,6-P2 are catalysed by isoforms of the bifunctional enzyme 6-phosphofructo-2-kinase/fructose-2,6-bisphosphatase (termed PFK-2/FBPase-2 in animals or F2KP in plants). Isoforms have distinctive kinase/phosphatase activity ratios and different N- or C-terminal regulatory domains that determine their responsiveness to extracellular and metabolic signals (El-Maghrabi et al., 2001). For example, in mammalian liver, stimulation of gluconeogenesis and inhibition of glycolysis by the hormone glucagon is mediated by phosphorylation of Ser32 of liver PFK-2/FBPase-2 by cAMP-dependent protein kinase (PKA), which decreases the kinase and increases the bisphosphatase activity, thereby lowering fru-2,6-P2 levels. In contrast, the heart isoform has a very high PFK-2/FBPase-2 ratio. When heart cells are deprived of oxygen, heart PFK-2/FBPase-2 is phosphorylated on Ser466 by the AMP-activated protein kinase (AMPK), which activates the PFK-2, to promote anaerobic glycolysis as part of a survival response (Marsin et al., 2000). Phosphorylation of heart PFK-2/FBPase-2 on Ser466 and Ser483 in response to insulin also activates the PFK-2 (Bertrand et al., 1999; Depréet al., 1998). Recently, we found that heart PFK-2/FBPase-2 binds to 14-3-3 proteins after phosphorylation with the insulin/growth factor-stimulated protein kinase B (PKB; Pozuelo Rubio et al., 2003). 14-3-3s are highly conserved eukaryotic proteins that bind to phosphorylated sites in diverse target proteins (Tzivion and Avruch, 2002).

In Arabidopsis, a single AtF2KP gene has been identified that encodes a 92-kDa enzyme, which appears to function as a homotetramer (Villadsen et al., 2000). The AtF2KP has a conserved catalytic region (88% identity between Arabidopsis and spinach; 47% identity between Arabidopsis and human liver isoform), and a long plant-specific N-terminal extension (59% identity between Arabidopsis and spinach sequences). Studies on recombinant AtF2KP and spinach F2KP have confirmed that these enzymes have both kinase and phosphatase activities (Markham and Kruger, 2002; Villadsen and Nielsen, 2001). The N-terminal extension is not required for the catalytic activities of AtF2KP, but it influences the ratio of PFK-2 and FBPase-2 activities (Villadsen and Nielsen, 2001). The PFK-2/FBPase-2 activity ratio is also modulated allosterically by hexose phosphates, three-carbon phosphate esters and inorganic ortho- and pyrophosphate (Markham and Kruger, 2002; Villadsen and Nielsen, 2001). While spinach leaves contain a monofunctional FBPase-2 in addition to the bifunctional F2KP (MacDonald et al., 1989), no such monofunctional FBPase-2 has been observed in Arabidopsis (Draborg et al., 2001), indicating that AtF2KP provides the only route for the synthesis and degradation of fru-2,6-P2 in Arabidopsis leaves.

In a range of plant species, the level of fru-2,6-P2 in leaves has been found to vary in response to photosynthetic activity. The observed patterns might be explained in part by variations in allosteric regulators of F2KP (Stitt, 1990). However, there are also indications that reversible covalent modification (phosphorylation) resulting in changes in kinase/phosphatase activity ratios may play a role in regulating fru-2,6-P2 levels (Rowntree and Kruger, 1995; Stitt et al., 1986; Walker and Huber, 1987). Recently, diel and developmental changes in the phosphorylation status of AtF2KP in leaves have been reported (Furumoto et al., 2001), and it was demonstrated that phosphorylation was on serine and threonine residues, distinct from the phospho-histidine intermediate that forms in the phosphatase reaction mechanism. The phosphorylated AtF2KP was most prominent at the beginning of the night, and in young leaves (Furumoto et al., 2001).

Here, we report a diurnal variation in the content of fru-2,6-P2 in Arabidopsis leaves. In particular, there were abrupt decreases in fru-2,6-P2 level at the light/dark and dark/light transitions. The enzyme nitrate reductase (NR), the first enzyme in the assimilation of nitrate into amino acids, is also regulated diurnally. These findings raised the question of whether similar mechanisms underlie the regulation of NR and fru-2,6-P2/F2KP.

Daily changes in NR activity are a function of slow changes in synthesis and degradation controlled by circadian, light and nitrate signals, and rapid changes in reversible phosphorylation and binding to 14-3-3 proteins (Kaiser and Huber, 2001; Kaiser et al., 2002; Lillo et al., 2001; MacKintosh and Meek, 2001). When photosynthesis is inhibited in the dark, a signal is transmitted from the chloroplast to the cytoplasm that triggers the phosphorylation of a serine residue (Ser543 in spinach NR). Phosphorylation of Ser543 creates a phosphopeptide motif that binds to 14-3-3 proteins, and the interaction with 14-3-3s inhibits the phosphorylated NR, explaining why nitrate reduction is rapidly inhibited in the dark. When leaves are illuminated, NR is dephosphorylated, 14-3-3s dissociate and the enzyme is activated (Kaiser and Huber, 2001; Kaiser et al., 2002; MacKintosh and Meek, 2001). Formation of the inactive 14-3-3/pNR complex in spinach is promoted by Mg2+ (Kaiser et al., 2002). Therefore, phosphorylated and 14-3-3-bound NR has low activity in the presence of Mg2+ that can be increased in vitro by blocking 14-3-3 binding by competition with a synthetic 14-3-3-binding phosphopeptide such as ARAApSAPA (where pS is phosphoserine; Moorhead et al., 1996). 14-3-3s have also been implicated in regulation of plant sugar metabolism via phosphorylation-dependent binding to sucrose-phosphatase synthase and trehalose-phosphate synthase isoforms (Moorhead et al., 1999; Toroser et al., 1998).

Here, we monitor diurnal changes in levels of fru-2,6-P2 and NR activity in Arabidopsis leaves, and changes in response to sugars, sugar analogues and inhibitors of protein kinases and phosphatases in Arabidopsis suspension cells. These studies led us to find that AtF2KP binds to the phosphopeptide-binding site on 14-3-3 proteins after in vivo or in vitro phosphorylation by Arabidopsis protein kinase(s), and we have identified the phosphorylated residues.


Diurnal variation in fru-2,6-P2 metabolism

The fru-2,6-P2 content in Arabidopsis leaves showed a distinctive diurnal variation (Figure 1a). At the beginning of the light period, there was an abrupt decrease in fru-2,6-P2 level, which remained low for the next 2 h of the photoperiod. Through the last 6 h of the photoperiod, the fru-2,6-P2 level gradually increased until it reached the same level as at the start of the photoperiod. The fru-2,6-P2 level did not change in the first 5 min of the dark period, but after 15 min in the dark, there was a decrease in fru-2,6-P2, which then increased slightly and thereafter remained at a steady level for the remainder of the dark period.

Figure 1.

Diurnal variation in fru-2,6-P2 levels and NR activity in Arabidopsis leaves.

Fully developed leaves from Arabidopsis plants grown in 8 h light (approximately 100 µmol photons m−2 sec−1) and 16 h dark for 75 days (a) or 70 days (b) were harvested at the indicated times.

(a) Fru-2,6-P2 content (pmol g−1 FW). Values are means ± SD of three samples each representing two individual plants.

(b) NR activity of desalted extracts measured in the presence of Mg2+, and 100 µm control phosphopeptide, WFYpSPFLE (circles) or 100 µm of the 14-3-3-binding phosphopeptide ARAApSAPA (triangles). Values are means of duplicates that did not vary by more than 15%, and similar results were seen in a separate experiment.

Experiment A was performed in Copenhagen and experiment B in Dundee.

These variations in fru-2,6-P2 reflect in planta regulation of AtF2KP, which provides the only route to both the synthesis and degradation of fru-2,6-P2 in Arabidopsis (Draborg et al., 2001). Substrate availability and allosteric properties of AtF2KP may represent key features in this regulation, but in addition, it is possible that the Vmax or regulatory properties of the enzyme are modulated by reversible covalent modification in response to changes in metabolic conditions. When measured under optimal conditions, the activities of Arabidopsis F6P,2K or F26BPase in leaf extracts were not significantly altered during the diurnal period (not shown), suggesting that Vmax activities do not change.

Diurnal variation of NR activity in Arabidopsis leaves

In the leaves of many plants, there is a diurnal change in NR activity, which is a function of changes in the phosphorylation-dependent binding to 14-3-3 proteins, which inhibits NR, superimposed on diurnal changes in synthesis and degradation of NR protein (Kaiser and Huber, 2001; Kaiser et al., 2002; Lillo et al., 2001; MacKintosh and Meek, 2001). In Arabidopsis leaves, the NR activity (measured in the presence of Mg2+) also varied diurnally (Figure 1b). Most notably, there was an abrupt approximately threefold increase in Arabidopsis NR activity (measured in the presence of Mg2+) at the beginning of the photoperiod, a gradual apparent decrease over the course of the day, and a more abrupt decrease at the beginning of the dark period (Figure 1b).

The phosphopeptide, ARAApSAPA, binds to 14-3-3s in competition with phosphorylated NR (as does the Raf-1 phosphopeptide in Moorhead et al., 1996). NR activity (+Mg2+) in extracts of leaves harvested in the dark was low, but this enzyme was highly activated by incubation with the ARAApSAPA phosphopeptide, indicating that a large proportion of the NR in the dark period had been inhibited by (phosphorylation and) binding to 14-3-3s (Figure 1b). In contrast, NR activity (+Mg2+) from leaves harvested in the photoperiod was more active, and not activated much further by the ARAApSAPA phosphopeptide (Figure 1b), indicating that only a small proportion was phosphorylated and bound to 14-3-3s.

In summary, the sudden decrease in fru-2,6-P2 levels at the start of the photoperiod (Figure 1a) coincided with an increase in NR activity caused by a dissociation of 14-3-3 proteins (Figure 1b), while the more modest and transient decrease in fru-2,6-P2 levels at the start of the dark period (Figure 1a) coincided with a decrease in NR activity because of phosphorylation and binding to 14-3-3s (Figure 1b).

2-Deoxyglucose prevented the rises in fru-2,6-P2 and NR activity that occurred when cells were transferred to fresh medium

For further studies of the regulation of fru-2,6-P2 levels and AtF2KP, we used Arabidopsis suspension cells because they are easy to treat with metabolites and inhibitors, and because the 14-3-3 affinity binding methods (Cotelle et al., 2002; Moorhead et al., 1999) that we use with the cell systems do not work well with leaf extracts (J.E. Harthill, unpublished).

When the cells were transferred from the medium they had been growing in for 5 days into fresh medium, fru-2,6-P2 levels gradually increased fivefold during the first hour after the transfer (Figure 2a). The rise in fru-2,6-P2 could not be attributed to higher sugar concentrations in the fresh medium because the intracellular fru-2,6-P2 remained low when cells were kept in ‘old’ medium (collected from a 5-day-old cell culture) that was supplemented with 100 mm of any of the following sugars: sucrose, glucose, fructose, sorbitol, mannitol, trehalose, galactose, arabinose, mannose, 2-deoxyglucose (2-DOG), or 3-O methyl glucose (3-OMG; not shown). Furthermore, fru-2,6-P2 increased in fresh medium that lacked sugars or polyalcohols, or fresh medium with sucrose, glucose, fructose, sorbitol, or mannitol (Figure 2b). However, the increase in fru-2,6-P2 was blocked when cells were transferred to fresh medium containing either 2-DOG as sole sugar (Figure 2b), or 2-DOG plus sucrose (not shown), but only slightly blocked when cells were transferred to fresh medium containing 100 mm 3-OMG as sole sugar (Figure 2b).

Figure 2.

Effect of sugars and inhibitors of protein phosphatases and protein kinases on intracellular fru-2,6-P2 and NR activity in Arabidopsis suspension cells transferred to fresh medium.

Cells (10 ml of a 5-day-old culture) were collected by centrifugation, transferred to 10 ml of ‘old’ medium or 10 ml of fresh medium containing the sugars or inhibitors indicated, and extracted after the times indicated (a,d) or after 1 h (b,c,e,f). Each flask contained 0.7–0.9 g FW of cells, and extracts contained approximately 2.5 mg ml−1 protein after desalting. Panels (a–c) show fru-2,6-P2 content (pmol g−1 FW) expressed as means of triplicate assays ± SDs. Panels (d–f) are NR activity (mU mg−1 protein) of desalted extracts assayed in the presence of Mg2+ with either 14-3-3-binding phosphopeptide (100 µm, ARAApSAPA, hatched bars) or control phosphopeptide (100 µm, WFYpSPFLE, filled bars). Each panel shows results from a separate experiment. Similar overall trends were obtained in at least one or two other experiments for each, although the effects of K252a were sometimes less, and the activity of NR in ‘old’ medium sometimes higher, than shown here. Sugars were used at 100 mm. Where indicated, DMSO was used at 3 µl in 10 ml of cell medium, and concentrations of inhibitors were: K252a (1 µm), U0126 (10 µm), PD98059 (50 µM), quercetin (20 µm), H89 (10 µm), okadaic acid (OA, 300 nm), and calyculin A (CA, 300 nm).

The same results were obtained whether cells were left in old medium, or harvested and re-suspended back into old medium (not shown), suggesting that the fru-2,6-P2 was being affected by a chemical difference between the old and new medium, not a physical stress.

Similarly, NR activity (+Mg2+) was regulated by the ‘old/new medium signal’ (Figure 2d); NR activity in the Arabidopsis cell extracts was low whenever fru-2,6-P2 was low, and NR activity (+Mg2+) increased in new medium (compare Figure 2a,d). As was found for fru-2,6-P2 (Figure 2b), NR activity state was unaffected by the addition of sugars, except for 2-DOG, at least over the 1-h time course of this experiment (Figure 2e; data not shown). We can deduce that where NR activity was low, the NR had been inhibited by binding to 14-3-3s because it was activated by incubation with the ARAApSAPA 14-3-3-binding phosphopeptide (Figure 2d–f).

While the source of the signals that control fru-2,6-P2 and NR activity in these experiments is not yet clear, increases in fru-2,6-P2 concentration were remarkably coordinated with increases in NR activity state that were caused by dephosphorylation of NR and dissociation of 14-3-3s.

Effects of protein kinase and phosphatase inhibitors

We therefore tested whether fru-2,6-P2 levels and AtF2KP activity were affected by inhibitors of protein serine/threonine phosphatases and protein kinases. The protein phosphatase inhibitor calyculin A prevented the gradual rise in fru-2,6-P2 level that occurred after the cells were transferred into fresh medium (Figure 2c). This effect of calyculin A was dose dependent, with maximal effect at approximately 400 nm (data not shown). In contrast, another protein phosphatase inhibitor, okadaic acid, and the protein kinase inhibitors H89, PD98059, U0126 or quercetin had little or no effect (see Experimental procedures for the specificity of these compounds). The general protein kinase inhibitor K252a had no significant effect on fru-2,6-P2 when used alone, but counteracted the effect of calyculin A (Figure 2c).

Similarly, calyculin A blocked the rise in NR activity (+Mg2+) caused by transferring to new medium, and this effect of calyculin A was counteracted by K252a (Figure 2f). This finding is consistent with the known inhibition of NR by phosphorylation and 14-3-3 binding. Other protein kinase inhibitors tested had no obvious effect on NR.

The parallels between the changes in fru-2,6-P2 and changes in NR activity were striking (Figure 2f). These results indicated that fru-2,6-P2, and hence AtF2KP, were regulated directly or indirectly by reversible phosphorylation. However, as in Arabidopsis leaves, we failed to find any changes in extractable F6P,2K or F26BPase activities that were consistent with the changes in cellular fru-2,6-P2 (not shown).

AtF2KP binds to a 14-3-3 column and is eluted by competition with a 14-3-3-binding phosphopeptide

An extract of Arabidopsis cells was passed through a column of 14-3-3-Sepharose. The column was washed extensively in 0.15 m salt and with a control phosphopeptide that does not bind to 14-3-3s. Proteins that were bound to the phosphopeptide-binding site on the immobilised 14-3-3s were then eluted with the ARAApSAPA phosphopeptide that contains a 14-3-3-binding consensus sequence.

14-3-3 overlays showed that the ARAApSAPA elution pool was highly enriched in 14-3-3-binding proteins compared with the crude extract (Figure 3b). NR and AtF2KP were in the crude extract and column flowthrough, and slightly enriched in the beginning of the salt wash (Figure 3c,d). Neither protein was eluted by the non-14-3-3-binding phosphopeptide control, but both NR and F2KP were specifically eluted by the ARAApSAPA phosphopeptide (Figure 3). In addition to the approximately 92-kDa intact protein, anti-AtF2KP antibodies detected possible degradation products and a faint band that might correspond to the additional approximately 96-kDa phosphorylated form reported by Furumoto et al. (2001). The specific activity of F6P,2K was highest in the ARAApSAP elution pool (Figure 3f). In contrast, phosphoenolpyruvate (PEP) carboxylase, which is regulated by phosphorylation, but not regarded as a 14-3-3-binding protein, could not be detected in the ARAApSAPA elution pool (Figure 3e). These results were consistent with AtF2KP binding directly, or via an intermediary protein, to the phosphopeptide binding site on 14-3-3s.

Figure 3.

14-3-3 affinity chromatography of Arabidopsis cell extracts.

Clarified Arabidopsis extract was chromatographed on 14-3-3-Sepharose as described in Experimental procedures. The extract and column flow-through were analysed without being concentrated, while 10 ml of samples from the beginning and end of the 1-l wash, a mock elution with control phoshopeptide (WFYpSPFLE) and the ARAASpSAPA elution pool were collected and concentrated in Vivaspin 6 concentrators (Vivascience, Lincoln, UK) to 150 µl, of which 20 µl of each was run on SDS–PAGE using NuPage 8% Bis-Tris gels (Invitrogen, Groningen, the Netherlands) and transferred to nitrocellulose membranes (Shliecher and Schull, Dassel, Germany). Amounts of protein run on SDS–PAGE were: extract, flow-through and start of salt wash (40 µg of each); end of salt wash (protein undetectable); WFYpSPFLE pool (<1 µg); and ARSSpSAPA pool (7.5 µg in 20 µl). Blots of column fractions were analysed for protein with Ponceau S (panel (a)), digoxygenin (DIG)-14-3-3 overlay (panel (b)), Western blotting with anti-NR (panel (c)), anti-AtPFK-2 (panel (d)) and anti-PEP carboxylase (panel (e)). Column fractions were also assayed for fructose-6-P,2-kinase activity (panel (f)). DIG-14-3-3 overlays were performed as for Westerns except that DIG-labelled 14-3-3 was used in place of primary antibody, followed by a digoxygenin-horseradish peroxidase secondary antibody (Moorhead et al., 1999).

14-3-3s bind directly to glutathione S-transferase (GST)-AtF2KP that has been phosphorylated by extracts of Arabidopsis cells incubated with Mg-ATP

Generally, but with a few exceptions, 14-3-3 proteins bind to phosphorylated serine or threonine residues in defined consensus sequences (Tzivion and Avruch, 2002). To characterise the phosphorylation dependence of the interaction between 14-3-3s and AtF2KP, a GST-AtF2KP fusion gene was constructed and expressed and the protein was purified on glutathione-Sepharose. In control experiments, 14-3-3s from Arabidopsis cell extracts did not bind to glutathione-Sepharose, after incubation with GST alone or with no additions (Figure 4a). Similarly, only trace amounts of 14-3-3s bound to GST-AtF2KP after incubation in extracts without added MgATP (Figure 4b). However, several 14-3-3 isoforms were precipitated with GST-AtF2KP that had been incubated with Arabidopsis cell extracts in the presence of Mg-ATP (Figure 4b). While the proportion of 14-3-3s that bound to the GST-AtF2KP after phosphorylation in the extract (Figure 4b) was only approximately 1% of the total (the total is indicated by the arrow in Figure 4a), this might be expected given that the GST-AtF2KP would be competing for binding to 14-3-3s with many proteins in the extract that would also be phosphorylated under these conditions.

Figure 4.

Binding of 14-3-3s to GST-AtF2KP after incubation with Arabidopsis cell extracts.

(a) GST (20 µg) was incubated with Arabidopsis cell extract (indicated amounts of protein) for 30 min at 30°C in the presence or absence of MgATP, as indicated. Control incubations were performed without cell extract or without GST, as indicated. GST was precipitated with glutathione-Sepharose, and precipitates were washed and analysed for the presence of 14-3-3s by Western blotting, as described in Experimental procedures.

(b) A parallel experiment to (a), using 20 µg GST-ATF2KP in place of GST.

(c) As for (a), except that 14-3-3-binding proteins in the glutathione-Sepharose precipitates were detected by DIG-14-3-3 overlays, as described in Experimental procedures.

(d) A parallel experiment to (c), using 20 µg GST-AtF2KP in place of GST.

While these data suggested a phosphorylation-dependent binding of GST-AtF2KP to 14-3-3s, they did not indicate whether the interaction between the two proteins was direct, or which protein had to be phosphorylated for the interaction to occur. We therefore performed 14-3-3 overlays of the GST pull-downs, and found that digoxygenin-labelled 14-3-3 proteins (DIG-14-3-3s) bound directly to GST-AtF2KP that had been incubated with an Arabidopsis extract in the presence of Mg-ATP, but not in the absence of Mg-ATP (Figure 4d). The upper 14-3-3-binding protein in Figure 4(d) is full-length GST-F2KP, while the lower band at approximately 60 kDa is a C-terminally truncated form of GST-F2KP, which binds to 14-3-3s more strongly than the intact protein (Figure 4d; tryptic mass fingerprinting data not shown). The DIG-14-3-3s did not bind to the GST control under any conditions (Figure 4c). Thus, 14-3-3s bound directly to GST-AtF2KP that had been phosphorylated by protein kinase(s) in the Arabidopsis extract.

14-3-3s bind directly to GST-AtF2KP that has been phosphorylated by CDPK or mammalian AMPK, but not PKB

We next tested recombinant Arabidopsis calcium-dependent protein kinase isoform 3 (CPK3), the mammalian AMPK and mammalian PKB for their ability to phosphorylate the 14-3-3-binding site(s) on GST-AtF2KP (Figure 5), choosing these kinases for the following reasons: The CPK3 (GenBank 836944, formerly known as calcium-dependent protein kinase 6 (CDPK6), see is a CDPK that phosphorylates the 14-3-3-binding site on NR (Douglas et al., 1998). The 14-3-3-binding site on NR can also be phosphorylated by plant SNF-1 related protein kinases (SnRKs) (Douglas et al., 1997; Ikeda et al., 2000; Sugden et al., 1999; Toroser and Huber, 1998). The plant SnRKs and their mammalian relative, the AMPK, display very similar specificities towards a panel of synthetic peptide and protein substrates (for example, Dale et al., 1995; MacKintosh et al., 1992; review by Hardie and Hawley, 2001). While we have no purified plant SnRK, the mammalian AMPK was readily available. Mammalian PKB phosphorylates the 14-3-3-binding site on cardiac PFK-2 (Pozuelo Rubio et al., 2003), and 14-3-3-binding sites on a number of other proteins, including the mammalian forkhead transcription factor FKHR (FOXO1a) (e.g. Rena et al., 2001).

Figure 5.

Binding of 14-3-3s to GST-AtF2KP that has been phosphorylated with CPK3 or AMPK.

(a) GST-AtF2KP (20 µg) was incubated with 0.5 µg recombinant Arabidopsis CPK3 (6His-tagged), highly purified AMPK (10 U ml−1), Arabidopsis cell extract (10 µg), or PKB (1 U ml−1) for 30 min at 30°C in the presence or absence of MgATP, as indicated. Control incubations were performed without kinase or GST-AtF2KP, as indicated. GST-AtF2KP was precipitated with glutathione-Sepharose, and precipitates were washed and analysed for the presence of 14-3-3s and AtF2KP by Western blotting, as described in Experimental procedures.

(b) A parallel experiment to (a), using 20 µg GST-FKHR in place of GST-AtF2KP.

We found that GST-AtF2KP that was phosphorylated by an Arabidopsis cell extract, the CPK3 or the AMPK could bind 14-3-3s (Figure 5a). However, there was no binding of 14-3-3s to either unphosphorylated GST-AtF2KP or GST-AtF2KP that had been incubated with PKB. Control experiments showed that the PKB was active, because FKHR could bind 14-3-3s after phosphorylation with PKB (Figure 5b).

Mapping of the sites on AtF2KP that were phosphorylated in vitro by CPK3, AMPK and Arabidopsis cell extract

The phosphorylations of GST-AtF2KP with CPK3, AMPK and Arabidopsis cell extract were repeated using Mg[32P-γ]ATP. The 32P-labelled GST-AtF2KP was digested with trypsin, and phosphorylated peptides were isolated by HPLC and analysed by mass spectrometry and solid-phase Edman sequencing.

The HPLC profiles showed that phosphorylation of GST-AtF2KP by the Arabidopsis CPK3 generated two major radioactive phosphopeptides that were eluted at 16% acetonitrile (termed phosphopeptide P1) and 26% acetonitrile (phosphopeptide P2), and several minor peaks of which the biggest was eluted at 25% acetonitrile (termed phosphopeptide P3) (Figure 6a).

Figure 6.

Mapping the sites on GST-AtF2KP phosphorylated by Arabidopsis CPK3, mammalian AMPK, and Arabidopsis cell extract.

Reverse-phase separation with on-line radioactivity detection of 32P-labelled tryptic digest of GST-AtF2KP phosphorylated with (a) recombinant Arabidopsis CPK3, (b) mammalian AMPK, and (c) Arabidopsis cell extract. Full-length GST-AtF2KP that had been phosphorylated in vitro was digested with trypsin and chromatographed on a Vydac 218TP54 C18 column equilibrated in 0.1% (by volume) trifluoroacetic acid in water. The column was developed with a 96-ml gradient of acetonitrile in 0.1% trifluoroacetic acid (diagonal lines, where numbers 0, 30, 50 and 100 indicate percentage (v/v) acetonitile at points where the gradient was changed). The flow rate was 0.8 ml min−1, and 0.4-ml fractions were collected. For each experiment, approximately 15 000 c.p.m. was loaded onto the HPLC columns, and 32P radioactivity was counted by an on-line Berthold LD509 radioactivity detector (trace shown) and fractions were counted by Cerenkov counting.

(d–f) Samples of the 32P-labelled peptides from GST-AtF2KP phosphorylated with CPK3 (from HPLC trace A) were coupled via carboxyl groups to a Sequelon-AA membrane (Milligen) and sequenced by Edman degradation using either an Applied Biosystems 476A or 494C protein sequencer (Applied Biosystems, Warrington, UK). Panel (d) shows the release of 32P radioactivity at each cycle and the corresponding amino acid sequence for the phosphopeptide that was eluted at 16% acetonitrile (peptide P1), panel (e) is for the peak at 26% acetonitrile (peptide P2), and panel (f) is for the peak at 25% acetonitrile (peptide P3). 1172 c.p.m. (d), 1636 c.p.m. (e) and 590 c.p.m. (f) were coupled to the membranes, and following the indicated cycles of Edman sequencing, 260 c.p.m. (d), 238 c.p.m. (e) and 90 c.p.m. (f) remained attached to the membrane.

The MALDI-TOF mass spectrum of peak P1 contained ions whose masses matched with masses of the mono-phosphorylated form of peptide 214SLSASGSFR222 of AtF2KP (numbers are residues of F2KP, not including the GST) (Table 1), and those of the poorly resolved daughter ion characteristic of the loss of H3PO4 from the phosphopeptide in a post-source decay event. Residue 213 in AtF2KP is arginine, consistent with the specificity of trypsin. When subject to solid-phase Edman degradation, Cerenkov counting of the released ATZ-amino acids showed a burst of counts at cycle 7, which means that the residue corresponding to Ser220 (SLSASGpS220FR) was phosphorylated (Figure 6d).

Table 1.  Identification of tryptic peptides isolated from GST-ATF2KP phosphorylated by Arabidopsis CPK3 in vitro
HPLC peakResiduesModificationMeasured massTheoretical mass
  1. The peptides isolated from Figure 6(a) were analysed on a PerSeptive Biosystems (Framingham, MA, USA) Elite STR MALDI-TOF mass spectrometer in linear and reflective modes, using 10 mg ml−1α-cyanocinnamic acid as the matrix (Campbell and Morrice, 2002). The masses shown are monoisotopic, except for P3 (average mass). PO4, phosphate group. The residue numbers for AtF2KP (AF190739) do not include the GST.

P1AtF2KP 214-222PO4 (1)991.4224991.4250
P2AtF2KP 301-312PO4 (1)1348.61691348.6401
P3AtF2KP 301-312PO4 (2)1428.961429.40

P2 contained masses that corresponded to masses of both the mono-phosphorylated form of 301SLSASSFLIDTK312 (Table 1) (which is preceded by an arginine in the sequence of AtF2KP) and its daughter ion guaranted by loss of H3PO4. Upon solid-phase Edman sequencing, the major release of radioactivity occurred at cycle 3 (Figure 6e), consistent with phosphorylation of Ser303 of AtF2KP (301SLpS303ASSFLIDTK312).

P3 contained an ion whose mass corresponded to mass of the di-phosphorylated form of 301SLSASSFLIDTK312 (Table 1). Solid-phase sequencing resulted in two bursts of radioactivity release at cycles 3 and 6 (Figure 6f). A doubly phosphorylated peptide would be expected to be eluted before the mono-phosphorylated version of the same peptide on reverse-phase HPLC. Together, these data are consistent with phosphorylation at residues 303 and 306, 301SLpS303AspS306FLIDTK312.

The phosphorylation of GST-AtF2KP with rat liver AMPK generated a single major phosphopeptide (Figure 6b) that coincided with the P2 phosphopeptide generated by CPK3 phosphorylation (Figure 6a). This AMPK-phosphorylated peptide was confirmed by MALDI-TOF analysis and solid-phase Edman sequencing (not shown) to be the tryptic peptide phosphorylated on Ser303 (301SLpS303ASSFLIDTK312). A minor peak that eluted at the P3 position was also observed for AtF2KP phosphorylated with AMPK, but no data could be obtained from that or other minor 32P-labelled peaks.

HPLC (Figure 6c), MALDI-TOF and solid-phase sequencing (not shown) showed that phosphorylation of GST-AtF2KP with an Arabidopsis cell extract predominantly generated the mono-phosphorylated P2 peptide (301SLpS303ASSFLIDTK312), with a lesser proportion of the mono-phosphorylated P1 peptide (214SLSASGpS220FR222).

Note that in Figure 5(a), there is an upward bandshift of the truncated GST-AtF2KP after phosphorylation by CPK3, but not after phosphorylation by AMPK or Arabidopsis cell extracts. This is consistent with CPK3 phosphorylating two sites, and suggests that phosphorylation at Ser220 is responsible for the bandshift.


In this paper, we reported first on regulation of fru-2,6-P2 levels and NR activity in Arabidopsis leaves in response to changes in photosynthetic conditions during the day–night cycle (Figure 1), and in Arabidopsis suspension cultures cells in response to changing from old to new medium (Figure 2).

The diurnal variation in fru-2,6-P2 content in Arabidopsis closely resembles that in leaves of spinach (Stitt et al., 1983), potato and tobacco (Scott and Kruger, 1994). This variation has been proposed to reflect changes in the intracellular concentrations of the metabolites that allosterically regulate F2KP (Larondell et al., 1986; Markham and Kruger, 2002; Stitt, 1990; Villadsen and Nielsen, 2001). In this way, fru-2,6-P2 acts as a signal integrating the metabolic status of the individual cell. In addition, several studies have indicated that both the spinach and Arabidopsis F2KP may be regulated by reversible protein phosphorylation (Furumoto et al., 2001; Rowntree and Kruger, 1995; Stitt et al., 1986; Walker and Huber, 1987), and the data presented here supports this hypothesis.

When cells were transferred to new medium, fru-2,6-P2 levels and NR activity increased by a mechanism(s) involving dephosphorylation that is blocked by calyculin A. It is possible that some factor secreted by growing cells into the ‘old’ medium triggers a decrease in fru-2,6-P2 and NR activity. The increases in fru-2,6-P2 and NR activity were not because of recovery from sugar starvation per se because adding various sugars to either old or new medium had no effect (Figure 2; data not shown). Nevertheless, this effect may be linked in some way to hexokinase because it was blocked by 2-DOG, which is phosphorylated by hexokinase to make 2-deoxyglucose 6-phosphate (2-DOG-6-P). Whereas the usual hexokinase product, glc-6-P, is further processed by the cell, yielding energy and serving as a carbon source, 2-DOG-6-P does not enter glycolysis at a significant rate, and in fact, this metabolite is a feedback inhibitor of hexokinase (Pego et al., 1999; citations therein). In contrast, the effect of transferring to new medium was only slightly prevented by 3-O methylglucose, which cannot be phosphorylated by hexokinase but is an inhibitor of glucose uptake and phosphorylation.

The second major conclusion of this paper is that AtF2KP that has been phosphorylated by Arabidopsis protein kinase(s) can bind to 14-3-3 proteins. Moreover, the CPK3 and mammalian AMPK, two kinases that can phosphorylate NR, also phosphorylate the 14-3-3-binding site(s) on AtF2KP (Figure 5). The major sites phosphorylated in vitro by CPK3 were Ser220 (SLSASGpS220FR) and Ser303 (SLpS303ASSFLIDTK), while the AMPK phosphorylated Ser303 (Figure 5). An Arabidopsis extract phosphorylated both sites, but predominantly Ser303. Both sites are located in the N-terminal extension of AtF2KP, immediately preceding the kinase domain, which begins at around residue 346 in the AtF2KP (Accession no. AF190739). Consistent with the observed phosphorylations, the Ser303 site (VKSLpS303ASSF) conforms to the substrate recognition motif for AMPK and plant SnRKs (phi (X/basic)XX(pS/T)XXX phi, where phi is a hydrophobic residue (M, V, L, I or F) and the basic residues are R, K or H), as defined by Dale et al. (1995). By these criteria, phosphorylation by SnRK1s at Ser220 would not be expected. The Ser303 site matches the classic motif for phosphorylation by CDPKs, with a hydrophobic residue at P-5 and basic residue at P-3. While the Ser220 site is not a classical motif, it is similar to the non-classical CDPK phosphorylation site identified by Huang et al. (2001) in which peptides lacking a basic residue at P-3/P-4 were efficiently phosphorylated by CDPKs if they had a basic residue C-terminal (at P+1 or P+2) and N-terminal (P-6 to P-9) and a hydrophobic residue at P-5, that is (basic-basic-X-basic)-phi-X(4)-(pS/T)-X-basic.

Two optimal 14-3-3-binding consensus sequences RSXpSXP (mode 1) and RXXXpSXP (mode 2) have been selected by screening a peptide library (Rittinger et al., 1999), although several proteins that bind to 14-3-3s in a phosphorylation-dependent manner contain unconventional binding sequences (Jelich-Ottmann et al., 2001; Tzivion and Avruch, 2002; Pozuelo Rubio et al., 2003). The phosphorylated Ser303 site on AtF2KP (VKSLpSASSF) ranked as potential mode 1 14-3-3 binding site, with low stringency (where high stringency conforms most closely to the optimal binding sequence) using the scansite algorithm (; Rittinger et al., 1999), although we emphasise that several bone fide binding sites are not ‘optimal’ and factors other than the phosphopeptide sequence likely contribute to 14-3-3 binding of proteins. Sequences similar to the AtF2KP-Ser303 site exist in F2KP proteins from spinach, rice, maize, mangrove and potato, although the sites in these species are not so conserved that we can be absolutely confident that they would function as regulatory sites. We now aim to generate phosphospecific antibodies to track changes in phosphorylation of AtF2KP-Ser220 and AtF2KP-Ser303 during the diurnal cycle and in response to cell stimuli.

Another challenge will be to discover how phosphorylation and 14-3-3 binding of AtF2KP affect the regulatory properties and activities of AtF2KP, and cellular fru-2,6-P2 levels. The only way for the intracellular fru-2,6-P2 concentration to change is via changes in the activities of the F6P,2K or F26BPase activities of AtF2KP. However, we were unable to find any consistent changes in the kinase/phosphatase activity ratios, Km values or sensitivity to allosteric effectors under a variety of assay conditions in the presence and absence of 14-3-3s (not shown). This situation in Arabidopsis contrasts with spinach leaves, where the F6P,2K activity was shown to vary diurnally in a pattern similar to that of fru-2,6-P2, and this was proposed to reflect a regulated protein modification (Stitt et al., 1986). Proteolysis can affect sensitivity of an enzyme to regulatory phosphorylation (e.g. Baur et al., 1992; Douglas et al., 1995), but we detected no degradation of AtF2KP in crude extracts, although there were possible proteolytic fragments of AtF2KP in the elution pool from the 14-3-3 column (Figure 3). Perhaps, the answer lies in finding assay conditions that mimic the physiological concentration of some effector. For example, the regulatory effect of phosphorylation on PEP carboxykinase and spinach NR activities are only revealed in certain assay conditions (Kaiser et al., 2002; Walker et al., 2002). Another possibility is that phosphorylation/14-3-3-binding affects AtF2KP activity by affecting its subcellular location, which would not be revealed in in vitro assays.

It makes physiological sense to expect regulatory interactions between sugar/starch partitioning, and nitrate assimilation into amino acids. Carbon flux into sugar or starch or amino acids is highly regulated and determined by photosynthate availability, C/N balance, cytosolic demand for sucrose for export and respiration, and demand for amino acids (Stitt et al., 2002). It will be most interesting therefore to determine whether or not AtF2KP and NR are co-ordinately regulated by the same underlying mechanism. Consistent with this suggestion, both enzymes bound to 14-3-3s after they were phosphorylated by a plant CDPK (Douglas et al., 1998; Figures 5 and 6), or by mammalian or plant SnRKs (Douglas et al., 1997; Ikeda et al., 2000; Sugden et al., 1999; Toroser and Huber, 1998; Figures 5 and 6). Moreover, both classes of enzyme are inhibited by the general protein kinase inhibitor K252a (not shown), which affected the cellular content of fru-2,6-P2 (Figure 2) However, while a SnRK that can phosphorylate NR is reportedly activated during the light–dark transition in Brassica campestris leaves (Nakamura et al., 2002), we do not yet have a detailed knowledge of which kinase phosphorylates NR, or indeed AtF2KP, under different conditions in planta. Furthermore, there are 34 CDPKs, 8 CDPK-related kinases and 38 members of the SnRK family in Arabidopsis (although only three SnRKs are in subfamily 1, which is most closely related to the mammalian AMPK) (Cheng et al., 2002; We do not know whether all or only some of these enzymes can phosphorylate the 14-3-3-binding sites on NR and/or AtF2KP. The picture gets further complicated when we consider that NR activity increased (phosphorylation and 14-3-3 binding decreased) and fru-2,6-P2 levels decreased abruptly at the start of the photoperiod in Arabidopsis leaves (Figure 1). In contrast, both the more modest and transient decrease in fru-2,6-P2 at the start of dark period in leaves (Figure 1), and the large decreases in fru-2,6-P2 in the cell experiments (Figure 2) coincided with decreases in NR activity (increases in phosphorylation and 14-3-3 binding of NR). One possible explanation for these observations is that the phosphorylation of AtF2KP at different sites has different effects on the F6P,2K and/or F26BPase activities. In addition, the fru-2,6-P2 levels are likely to reflect photosynthetically driven changes in the cellular concentrations of allosteric regulators of AtF2KP (Markham and Kruger, 2002; Stitt, 1990; Villadsen and Nielsen, 2001). For example, increases in phosphorylated C-3 compounds, which inhibit F6P,2K, and a decrease in inorganic phosphate, which activate F6P,2K, would be expected to contribute to the sharp decrease in fru-2,6-P2 levels at the start of the photoperiod.

Finally, in an interesting synchronicity, we recently found that the human cardiac isoform of PFK-2/FBPase-2 (termed cardiac PFK-2) binds to 14-3-3 proteins when it is phosphorylated on Ser466 and Ser483 by the insulin- and growth factor-stimulated PKB (Pozuelo Rubio et al., 2003). At the moment, it is difficult to see any obvious physiological parallels between the phosphorylation, 14-3-3 binding and activation of cardiac PFK-2 in stimulation of glycolysis in human hearts, and the phosphorylation and 14-3-3 binding of Arabidopsis F2KP in the context of diurnal regulation of sugar/starch partitioning. The regulatory domains of the cardiac and Arabidopsis proteins are very different. Both the cardiac and Arabidopsis enzymes can be phosphorylated by the AMPK/SnRKs, but while AMPK/SnRK phosphorylates the 14-3-3-binding site on AtF2KP, it is PKB that phosphorylates the 14-3-3-binding site on cardiac PFK-2 (Pozuelo Rubio et al., 2003). Nevertheless, it will be interesting to see whether any further regulatory similarities between these two systems emerge.

Experimental procedures


Reagents for cell culture were purchased from Sigma (Poole, UK); PD 98059 from New England Biolabs (Beverly, MA, USA); MG132, calyculin A, K252a, Ro 318220, H89, SB 203580 and Zwittergent 3-16 from Calbiochem (Nottingham, UK); Complete protease inhibitor mixture and protein sequencing grade trypsin were from Roche Molecular Biochemicals (Roche Diagnostics Ltd., Lowes, East Sussex, UK); enhanced chemiluminescence (ECL) reagent from Amersham Pharmacia Biotech (Little Chalfont, UK). U0126 was provided by Dr Sue Cartlidge (AstraZeneca Pharmaceuticals, Macclesfield, Cheshire, UK).

Plants and cell culture

Arabidopsis plants (cv. Columbia) were grown in peat soil in controlled climate chambers under the light regimes indicated in Figure 1. Arabidopsis cell suspension cultures (cv. Erecta) were originally derived by May and Leaver (1993) and were grown in Murashige and Skoog medium with Minimal Organics (MSMO), α-naphthalene acetic acid (0.5 mg l−1) (auxin), kinetin (0.05 mg l−1; cytokinin) and 3% (w/v) sucrose (pH 5.7) in 500 ml conical flasks under continuous light (20–25 μE m−2 sec−1) at 23°C in a rotary shaker (150 r.p.m.). For cell treatments, 10 ml of cell suspension in late exponential phase (5 days old) was pelleted by centrifugation (2 min, 65 g), re-suspended in 10 ml of the medium indicated in 25-ml glass conical flasks and returned to the growth chamber with shaking. Samples were taken at the times indicated. Cells were harvested by vacuum filtration onto Miracloth (Calbiochem), washed with distilled water, frozen in liquid nitrogen in pre-weighed tubes and stored at −80°C. Where indicated, calyculin A, okadaic acid, K252a, SB 203580, PD 98059, U0126, Ro 318220 or H89 were added from stock solutions of 10 mm in DMSO.

The specificities of these inhibitors are:

  • • Calyculin A and okadaic acid inhibit the PPP family of protein serine/threonine phosphatases (except PP2B). Okadaic acid inhibits PP2A more potently than PP1 (reviewed in MacKintosh and Diplexcito, 2003).
  • • K252a is a general inhibitor of many mammalian protein kinases (Davies et al., 2000).
  • • Ro 318220, H89, K252c and quercetin are more selective inhibitors than K252a. Each compound inhibits distinct, but overlapping, subsets of mammalian protein kinases (Davies et al., 2000). H89 and quercetin have been found to block a signalling pathway in the Arabidopsis cells used here (J. Diplexcito and C. MacKintosh, unpublished).
  • • SB 203580 is a specific inhibitor of the mammalian stress-activated protein kinase 2a (SAPK2a, also known as p38; Davies et al., 2000).
  • • PD98059 and U0126 are highly specific inhibitors of upstream kinases that activate mammalian and plant MAP kinases (Davies et al., 2000; Desikan et al., 2001). The concentrations of PD 98059 and U0126 used in this study were sufficient to inhibit a calyculin A-activated MAP kinase in the same cell culture (J. Diplexcito and C. MacKintosh, unpublished).

Extraction and assay of fru-2,6-P2

Frozen cells from a 10-ml culture (around 0.9 g FW) were ground in 5 ml of ice-cold 10 mm KOH and centrifuged at 15 000 g for 5 min, or 50 mg frozen leaf tissue was ground in 2.5 ml of cold 10 mm KOH and centrifuged for 1 min at 10 000 g. Fru-2,6-P2 (in 10 µl of extract) was measured by following the activation of the enzyme pyrophosphate-dependent 6-phosphofructo-1-kinase from potato tubers (Nielsen et al., 1991; Van Schaftingen, 1984). The recovery of fru-2,6-P2 was approximately 90%. Trial experiments showed that extraction with Triton X-100 did not affect fru-2,6-P2 levels.

Extraction and assay of NR, F6P,2K and F26BPase

For NR assays, cells from 10 ml cultures (around 0.9 g FW) or approximately 0.5 g powdered, frozen leaf tissue were thawed in 1 ml of extraction buffer (50 mm MOPS-NaOH (pH 7.3), 1 mm phenylmethylsulfonyl fluoride (PMSF), 0.1% (v/v) Triton X-100, 1 µm microcystin-LR, 1 mg ml−1 bovine serum albumin (lipid-free fraction V from Sigma), Complete protease inhibitor (one tablet in 100 ml) and 50 µm MG132. The mixture was sonicated (3 × 15-sec bursts with cooling in between) and centrifuged for 20 min at 15 000 g, and supernatants were filtered through two layers of Miracloth. Extracts were rapidly desalted in PD-10 gel filtration columns (Pharmacia) or by dialysis in extraction buffer, and stored frozen at −80°C. For NR assays, 40 µg protein (approximately 20–40 µl of desalted extract) was incubated for 10 min in 100 µl of extraction buffer containing 10 mm MgCl2, 0.5 mm NADH and 2 mm NaNO3, and nitrite production was determined as described previously by Douglas et al. (1998).

The activities of F6P,2K and F26BPase were assayed by detection of the formation or degradation of fru-2,6-P2 as in MacDonald et al. (1989) and Nielsen et al. (1991), and modified by Nielsen (1992). One unit was defined as the conversion of 1 µmol of substrate per minute. The Arabidopsis leaf and cell extracts were prepared in five volumes of buffer A (50 mm MOPS-KOH (pH 7.3), 5 mm MgCl2, 1 mm EDTA, 10% (v/v) ethylene glycol, 0.1% (v/v) 2-mercaptoethanol, 5 mm benzamidine, 1 mg ml−1 antipain, 1 mg ml−1 leupeptin, 2 mm PMSF and 0.1% (v/v) Triton X-100).

14-3-3 affinity chromatography of Arabidopsis cell extracts

Arabidopsis cells (approximately 80 g of 7-day-old culture for the experiment in Figure 3) were thawed and extracted by sonication (three 20-sec bursts, interspersed with 15-sec cooling periods) into 80 ml of buffer comprising 50 mm Hepes-NaOH (pH 7.5), 50 mm NaF, 5 mm sodium pyrophosphate, 1 mm DTT, 1 mm PMSF, 1 mm benzamidine, 10 µg ml−1 leupeptin, one tablet of Complete protease inhibitor (Roche), 1% (w/v) polyvinylpolypyrrolidone and 0.1% Triton X-100. Extracts were filtered through two layers of Miracloth and centrifuged at 10 000 g for 60 min, and the supernatant was filtered through 0.45- and 0.2-µm filters. The clarified solution was mixed end-over-end for 1 h at 4°C with Sepharose linked to BMH1/BMH2 (the Saccharomyces cerevisiae 14-3-3 isoforms) that had been prepared as described by Moorhead et al. (1999). The mixture was poured into a column, and the flow-through was collected. The column was washed with 1 l of 50 mm Hepes-NaOH (pH 7.5), 150 mm NaCl and 1 mm DTT, and mock-eluted with a synthetic peptide that does not bind 14-3-3s (10 ml, 1 mm WFYpSPFLE for Figure 3, or ARAASAPA in other experiments), before proteins that bind to the phosphopeptide-binding site of 14-3-3s were eluted with 10 ml of 1 mm ARAApSAPA phosphopeptide.

Generation of GST fusion of AtF2KP and GST-ATF2KP pull-downs

The AtF2KP-coding region (identical to Accession no. AF190739) was amplified by PCR using pBluescript containing AtF2KP cDNA as template (Villadsen et al., 2000) and Hi Fi polymerase (Roche). (Note that entering AF190739 into the NCBI database shows another AtF2KP sequence, AAF82210 labelled ‘Identical to AF190739.’ It is not. AAF82210 has an 18-amino acid deletion between residues 248/249 and an A611D change.) The primers were GTCGACGGGTCAGGTGCATCGAAGAATACTGA and GCGGCCGCGAATTCAGTCCATGAGTTTGTAGCGTTTCTC. Gel-purified fragments were ligated into the pCR2.1 TOPO vector (Roche). The fragment encoding AtF2KP was cut using NotI and SalI, and subcloned into the pGEX-GP-3 vector (Amersham Pharmacia Biotech). The final construct encoded GST immediately followed by AtF2KP lacking the initiating methionine (hereafter termed GST-AtF2KP) and was used for expression of the protein in bacteria. The authenticity of all plasmids was confirmed by restriction analysis, and DNA sequencing performed by the DNA Sequence Service at the School of Life Sciences, University of Dundee.

The GST fusion protein was overexpressed in Escherichia coli BL21 cells by induction with isopropyl-β-d-thiogalactopyranoside (50 µm, 30°C, overnight). Cells were sonicated, lysates were centrifuged and the GST fusion protein was bound to glutathione Sepharose 4B beads (Amersham Pharmacia Biotech) according to the manufacturer's instructions. The GST protein was eluted with 20 mm glutathione, 50 mm Tris–HCl (pH 7.5) and 250 mm NaCl. The purified protein was stored at −80°C. The authenticity of the protein was confirmed by tryptic mass fingerprinting (Campbell and Morrice, 2002) and Western blotting with anti-AtF2KP (Villadsen et al., 2000) and anti-GST antibodies.

For glutathione-Sepharose pull-downs from cell extracts, purified GST-AtF2KP (20 µg) was incubated with Arabidopsis cell extract under the conditions indicated in figure legends. A 1 : 1 slurry (20 µl) of glutathione-Sepahrose beads in 50 mm Tris–HCl (pH 7.5) and 250 mm NaCl was added and incubated for 1 h with end-over-end shaking. Then, beads were washed two times with 50 mm Tris–HCl (pH 7.5), 150 mm NaCl containing 0.5% (v/v) Triton X-100, two times in 50 mm Tris–HCl (pH 7.5), 150 mm NaCl and two times with 50 mm Tris–HCl (pH 7.5). Bound proteins were eluted with SDS–PAGE sample buffer.

Protein kinases and in vitro phosphorylation

The AMPK was purified from rat liver to a specific activity of 448 U mg−1 using SAMS peptide (HMRSAMSGLHLVKRR; Dale et al., 1995) as substrate, by Kevin Green and colleagues in Grahame Hardie's group in the School of Life Sciences, University of Dundee. Arabidopsis CPK3 (6His-tagged, formerly known as CDPK6) was expressed in E. coli and purified as described by Douglas et al. (1998). The recombinant CPK3 has a lower specific activity and is less Ca2+ dependent than the native enzyme (Douglas et al., 1998). Recombinant protein kinase B alpha lacking the pleckstrin homology domain (ΔPH-PKBα-S473D) was produced by Carla Clark in the Division of Signal Transduction Therapy, School of Life Sciences, University of Dundee and activated to 1739 U mg−1 (assayed as by Walker et al., 1998) by phosphorylation with PDK1, and the PDK1 was subsequently removed as described by Walker et al. (1998).

GST-AtF2KP (20 µg) was phosphorylated with Arabidopsis cell extract (10 µg for Figure 6), CPK3, AMPK or ΔPH-PKBα for 30 min at 30°C in the presence of [32P-γ]ATP with a specific radioactivity of 196 c.p.m. pmol−1. Proteins were boiled in SDS sample buffer plus 10 mm DTT for 1 min, alkylated by adding 4-vinylpyridine to 50 mm final (5 µl ml−1) at room temperature for 30 min, run on an 10% Bis/Tris Novex pre-made gel and stained with Colloidal Coomasie R-250 (Novex, Invitrogen Ltd., Paisley, Scotland), and GST-ATF2KP was excised. The gel pieces were washed sequentially for 15 min in 0.5 ml each of water, 50% acetonitrile/water, 0.1 m ammonium bicarbonate and 50% acetonitrile plus 50 mm ammonium bicarbonate. The gel was dehydrated in 0.5 ml of 100% acetonitrile, supernatant was removed and the gel was dried in a speed-vac. The gel was swelled in 50 mm ammonium bicarbonate, 0.05% Zwittergent 3-16 containing 10 µg ml−1 of trypsin (100 µl per gel band) at 30°C overnight. An equivalent volume of acetonitrile was added with shaking for 15 min. Gel debris was removed by spinning through an Ultrafree-MC cartridge, and the liquid was dried in a speed-vac. Peptides were separated by HPLC, as described in the legend to Figure 6. Isolated 32P-labelled phosphopeptide fractions were analysed by MALDI-TOF mass spectrometry, and sites of phosphorylation were determined by solid-phase Edman degradation of the peptide, as described by Campbell and Morrice (2002).


A visit to Dundee by D.V. to perform some of the experiments described in this paper was sponsored by the Royal Veterinary and Agricultural University, Denmark. Research in C.M.'s group is funded by the BBSRC (UK) and MRC (UK), and T.H.N.'s laboratory is funded by The Danish National Research Foundation, Centre for Molecular Plant Physiology. We thank Grahame Hardie, School of Life Sciences, University of Dundee, for the purified preparation of rat liver AMPK.