The novel protein phosphatase PphA from Synechocystis PCC 6803 controls dephosphorylation of the signalling protein PII



The family of PII signal transduction proteins consists of one of the most highly conserved signalling proteins in nature. The cyanobacterial PII homologue transmits signals on the nitrogen and carbon status of the cells through phosphorylation of a seryl residue. Recently, we identified a protein phospha-tase 2C (PP2C) homologue from the cyanobacterium Synechocystis PCC 6803, termed PphA, to be the cellular phospho-PII (PII-P) phosphatase. In this investigation, we characterized the enzymatic properties of PphA and investigated the regulation of its catalytic activity towards PII-P. PphA dephosphorylates phosphocasein and PII-P with similar efficiency in a strictly Mg2+- or Mn2+-dependent reaction. Low-molecular-weight phosphorylated molecules are poor substrates for PphA. Its reactivity towards PII-P, but not towards phosphocasein, is inhibited by various nucleotides, suggesting that this effect is based on specific properties of the PII protein. The inhibitory effect of ATP can be strongly enhanced by the addition of 2-oxoglutarate or oxaloacetate. At low concentrations of 2-oxoglutarate, changes in the ATP levels within the physiological range affect the degree of PII-Pase inhibition, whereas at 2-oxoglutarate levels beyond 0.1 mM, inhibition is almost complete at very low ATP levels. This suggests that PII dephosphorylation is not only sensitive to 2-oxoglutarate and oxaloacetate levels, it also integrates signals from the energy charge of the cells under specific cellular conditions.


The PII signalling proteins comprise a family of highly conserved and widely distributed signal transduction proteins, being present in Eukarya (plants), Bacteria and Archaea. The PII homologues in different organisms are involved in various aspects of nitrogen control and co-ordinate cellular responses at the level of nitrogen-dependent gene expression (e.g. through the two-component system NtrBC), modulation of enzyme activity through covalent modification (glutamine synthe-tase adenylylation) or nutrient transport (e.g. nitrite/nitrate transport) (reviewed by Ninfa and Atkinson, 2000; Arcondéguy et al., 2001). PII exerts its effect by protein– protein interactions with the targets of its regulation, the PII receptors. As far as has been investigated to date, the PII proteins are composed of three identical (Forchhammer and Tandeau de Marsac, 1994; de Mel et al., 1994; Vasudevan et al., 1994) or paralogous (Forchhammer et al., 1999; van Heeswjik et al., 2000) subunits of ≈ 12.5 kDa and bind ATP and 2-oxoglutarate in a mutually dependent manner (Kamberov et al., 1995; Forchhammer and Hedler, 1997). Furthermore, PII proteins may be subject to post-translational modification at a solvent-exposed T-loop structure of the protein. In various PII proteins from proteobacteria and a Corynebacterium species, a tyrosyl residue (Tyr-51 in Escherichia coli), located at the tip of the T-loop, is modified by uridylylation (reviewed by Ninfa and Atkinson, 2000; Arcondeguy et al., 2001). The modifying enzyme, the glnD gene product, is a bifunctional uridylyltransferase/ uridyl-removing enzyme (UTase/UR), the activity of which is regulated by glutamine (Jiang et al., 1998a). In the absence of glutamine, the enzyme catalyses PII uridylylation, whereas binding of glutamine to UTase/UR activates the uridylyl-removing activity. Binding of 2-oxoglutarate to E. coli PII and uridylylation of its T-loop were shown to modulate the interaction with its receptors (Jiang et al., 1998b,c). Thus, the PII protein integrates signals of the cellular nitrogen state through glutamine-regulated uridylylation/deuridylylation of Tyr-51 and of carbon metabolism through direct 2-oxoglutarate binding.

In contrast to the post-translational modification of PII by uridylylation, it has been shown that the PII homologue from the cyanobacterium Synechococcus PCC 7942 was modified by phosphorylation of serine 49 (Forchhammer and Tandeau de Marsac, 1995a) in the vicinity of the conserved Tyr-51. Phosphorylation and dephosphorylation are catalysed by two different proteins (Irmler et al., 1997), and in vitro analysis using partially purified extracts from Synechococcus PCC 7942 cells showed that phosphorylation and dephosphorylation are controlled by ATP and 2-oxoglutarate levels. The presence of these effector molecules stimulated the PII kinase activity, whereas it inhibited phospho-PII (PII-P) dephosphorylation. Considering the metabolic design of cyanobacteria, we hypothesized that 2-oxoglutarate acts as a signalling molecule for the balance of carbon and nitrogen metabolism in these organisms (Forchhammer, 1999). In accordance with the metabolite-binding properties of the Synechococcus PCC 7942 PII protein (Forchhammer and Hedler, 1997), we argued that the regulatory effect of these effector molecules is based on their binding to the PII protein. The preliminary experiments with partially purified extracts showed that the PII-phosphatase activity is Mg2+ dependent (Irmler et al., 1997), a characteristic feature of the PP2C family of protein phosphatases (Cohen, 1994; Shi et al., 1998). The genome of the cyanobacterium Synechocystis PCC 6803 contains a PII system closely related to that of Synechococcus PCC 7942 (Garcia-Dominguez and Florencio, 1997; Hisbergues et al., 1999) and harbours eight genes coding for putative PP2C homologues (Kaneko et al., 1996; Shi et al., 1998; Zhang et al., 1998). By targeted mutagenesis of these genes, we recently identified the cellular PII-P phosphatase from Synechocystis PCC 6803, which we termed PphA (Irmler and Forchhammer, 2001). PphA is a monomeric protein of 28.5 kDa, and the entire amino acid sequence comprises the PP2C catalytic domain and lacks N-terminal or C-terminal extensions that could act as regulatory or substrate-specifying determinants. Sequence comparison with the database revealed that PphA is member of a large subgroup of bacterial PP2C homologues, most of which share the small size of PphA. Their amino acid sequences are only distantly related to those of the previously characterized eukaryotic PP2C homologues or to the PP2C homologues from Bacillus subtilis involved in antisigma factor regulation such as SpoIIE (Adler et al., 1997). Most of these novel bacterial PP2C homologues, classified as group I (Treuner-Lange et al., 2001), were identified by genome sequencing and in no case, except for PphA, has the physiological substrate been identified. This study aimed to characterize the enzymatic properties of this novel bacterial PP2C and to elucidate the regulation of its activity towards its physiological substrate, PII-P.


Reactivity of PphA towards phosphocasein

Purified recombinant PphA was reacted with 32P-labelled phosphocasein, a standard substrate for various protein phosphatases, in the presence of 5 mM MgCl2, and the release of [32P]-orthophosphate was measured in a time course experiment. As shown in Fig. 1A, a constant dephosphorylation of phosphocasein could be detected during the first 60 s of the reaction. To determine the dependence of this reaction towards divalent cations, phosphocasein dephosphorylation was assayed using an enzyme preparation that had been dialysed against a buffer devoid of divalent cations. Reactions with increasing amounts of MgCl2 or MnCl2 were carried out. Figure 1B shows that the addition of as little as 0.5 mM MnCl2 saturated the reaction, whereas 5 mM MgCl2 was required to reach a similar reactivity of PphA., To determine the apparent KM of the reaction of PphA towards phosphocasein, a set of experiments was conducted in which 50 fmol of PphA was reacted with increasing concentrations of phosphocasein (10–110 nM), and the initial velocity of [32P]-orthophosphate release was measured. A double reciprocal plot of substrate concentration versus reaction velocity revealed an apparent KM for this reaction of 0.2 μM (Fig. 1C). The Vmax of the reaction was calculated to be 0.63 pmol of phosphocasein min−1 per 50 fmol of PphA, which cor-responds to a kcat of 0.2 s−1. Thus, kcat/KM is ≈ 1 × 106, indicating a high catalytic efficiency of PphA towards phosphocasein.

Figure 1.

A. Time course of PphA-catalysed phosphocasein dephosphorylation. Reactions of 25 μl aliquots, each containing 1.5 ng of purified PphA, were started by the addition to each of 3000 c.p.m. of 32P-labelled phosphocasein. After the indicated times (s), 125 μl of ice-cold stop mix was added, and the released 32Pi was quantified by scintillation counting.

B. Mg2+/Mn2+ dependence of PphA reactivity towards phosphocasein. PphA that was dialysed free of divalent cations was incubated with [32P]-phosphocasein in the presence of increasing concentrations of MgCl2 (closed diamonds) or MnCl2 (open squares). Released 32Pi was determined after 30 s, and the activity was compared with the control reaction using untreated purified PphA (defined as 100% activity in the presence of both 5 mM MgCl2 and 2 mM MnCl2).

C. Determination of kinetic constants of PphA towards phosphocasein. Reactivity of PphA towards different concentrations of [32P]-phosphocasein (from 10 nM to 110 nM) was determined as described in (A), and the reaction velocity (nM Pi released min−1) at each phosphocasein concentration was determined. A double reciprocal plot revealed apparent KM and kcat values of this reaction of 0.2 μM and 0.2 s−1.

Reactivity towards p-nitrophenyl phosphate and phosphopeptides

To elucidate the substrate specificity of PphA further, we tested its reactivity towards artificial low-molecular-weight substrates. A widely used substrate of protein phosphatases is p-nitrophenyl phosphate (PNPP) (MacKintosh, 1993). Using our standard buffer conditions (at pH 7.4 and 5 mM MgCl2), no reactivity of PphA towards PNPP could be detected. However, when the pH was raised above 8.0, a strictly Mn2+-dependent dephosphorylation of PNPP occurred, with a pH optimum between 8.5 and 9.0 (data not shown). In contrast to the reactivity towards phosphocasein, Mg2+ could not substitute for Mn2+ in this assay. In the presence of 2 mM MnCl2 and at a pH value of 8.5, an apparent KM of 1 mM and a kcat of 1.5 min−1 were calculated.

To determine the specificity of the enzyme towards different phosphorylated amino acid residues, we used synthetic oligopeptides containing phosphoserine (peptide p-S: RRA(pS)VA), phosphothreonine (peptide p-T: RRA(pT)VA) or phosphotyrosine (peptide p-Y: DADE(pY)LIPQQG) as substrates. Table 1 summarizes the results of these experiments. Reactivity towards all three phosphopeptides could be observed in the presence of 2 mM MnCl2. Compared with phosphocasein, the phosphopeptides are poor substrates for PphA, with KM values being three orders of magnitudes higher and the overall catalytic efficiency about four orders of magnitudes lower. Of the different phosphopeptides, p-T was the best substrate for PphA, and dephosphorylation required Mg2+ or Mn2+ ions (see Fig. 2). Mn2+ is effective at concentrations about fivefold lower than that of Mg2+; a similar ion dependence could be observed for p-S (not shown). Surprisingly, we observed a Mn2+-dependent dephosphorylation of p-Y by PphA, although phosphotyrosyl residues are not usually recognized by PP2C enzymes. Mg2+ (at a concentration of 10 mM) was six times less efficient in stimulating this reaction than 2 mM Mn2+. This result was confirmed with a second phosphotyrosyl peptide (END(pY)INASL). Reactivity of PphA towards this nine-amino-acid-long peptide was lower than that towards the 11-amino-acid-long peptide p-Y. The reduction was twofold or fivefold when assayed in the presence of 2 mM Mn2+ or 10 mM Mg2+ respectively (data not shown).

Table 1. Reactivity of PphA towards oligopeptides phosphorylated at threonyl- (p-T), seryl- (p-S) or tyrosyl- (p-Y) residues.
SubstrateSpecific activitya K M (M−1)b k cat (s−1) k cat/KM (M−1 s−1)
  1. a. Specific activity is given in pmol of P i released min−1μg−1 enzyme.

  2. b. Apparent K M values were calculated from a duplicate series of dephosphorylation assays using substrate concentrations in the range from 40 μM to 400 μM. Reactions were carried out under standard assay procedures in the presence of 2 mM MnCl2.

p-T403 × 10−40.051.7 × 102
p-S571.5 × 10−40.053.3 × 102
p-Y96 × 10−40.020.3 × 102
Figure 2.

Mg2+ or Mn2+ dependence of phosphothreonyl-peptide (p-T) dephosphorylation by PphA. Dephosphorylation reactions were carried out as described in Experimental procedures in the presence of various concentrations of Mg2+ (open circles) or Mn2+ (closed triangles) using a divalent cation-free PphA preparation.

Together, these experiments suggest that Mn2+ ions broaden the substrate specificity of PphA, making it accessible to a wider range of substrates such as PNPP or phosphotyrosyl peptides.

Regulated dephosphorylation of PII-P

To determine the in vitro reactivity of PphA towards its physiological substrate PII-P, PphA was incubated with a highly phosphorylated PII protein under various conditions and, subsequently, the phosphorylation state of PII was determined by non-denaturing PAGE and immunoblot analysis. By these means, we could demonstrate Mg2+/Mn2+-dependent dephosphorylation of PII-P by PphA and estimated a turnover rate of 1 min−1 at a substrate concentration of 38 nM (Irmler and Forchhammer, 2001). By analogy with what has been found for other type 2C protein phosphatases, this activity was completely inhibited by phosphate and partially inhibited by Zn2+ and Ca2+ ions (50% inhibition with 0.5 mM Zn2+ or 5 mM Ca2+; data not shown), which compete with the Mg2+/Mn2+ binding sites in the active site (compare Das et al., 1996). In extracts of the cyanobacterium Synechococcus PCC 7942, we identified previously a PII-P phosphatase activity that was inhibited by ATP and 2-oxoglutarate in a mutually dependent manner (Irmler et al., 1997). To elucidate the regulation of PphA activity towards PII-P, we tested the effect of various effector molecules in our purified in vitro system. In preliminary experiments, we titrated the amount of PphA to obtain ≈ 90% dephosphorylation of PII-P under standard reaction conditions in the absence of effector molecules. This quantity of PphA (10 ng) was then incubated with PII-P in the presence of different nucleotides (each 1 mM), with or without the addition of 0.5 mM 2-oxoglutarate (Fig. 3A). Compared with the positive control (reaction without the addition of effector molecules; Fig. 3A, lane c), ATP, GTP and ADP caused a partial inhibition of PII-P dephosphorylation already in the absence of 2-oxoglutarate. The addition of 2-oxoglutarate strongly enhanced the inhibition of phosphatase activity by ATP, whereas it had only a small effect in combination with ADP and no effect with GTP. In combination with UTP, 2-oxoglutarate caused a slight inhibition of phosphatase activity. Note that 2-oxoglutarate alone had no effect on PII-P dephosphorylation (compare Fig. 4; in the presence of various concentrations of 2-oxoglutarate, no inhibition occurs at 0 mM ATP or ADP). These results demonstrate that 2-oxoglutarate acts in synergy with ATP to inhibit PphA reactivity towards PII-P, and are in accordance with the results of a previous study that used a partially purified phosphatase extract from Synechococcus PCC 7942 (Irmler et al., 1997). However, the inhibitory effect of various nucleotides was novel, and the different synergy of 2-oxoglutarate with ATP or ADP suggested physio-logical significance. Further experiments showed that nucleotide monophosphates did not cause any inhibition of PII-P dephosphorylation (data not shown). To investigate further the differences between ATP and ADP in inhibiting PphA reactivity towards PII-P, various combinations of 2-oxoglutarate with either ATP or ADP were incubated in the reaction assay. Reaction mixtures, containing either 1 mM ATP or 1 mM ADP, were supple-mented with increasing amounts (from 0 to 0.1 mM) of 2-oxoglutarate (Fig. 3B). 2-Oxoglutarate displayed strong synergy with ATP, leading to almost complete inhibition of PII-P dephosphorylation at increased 2-oxoglutarate concentrations, whereas synergy between 2-oxoglutarate and ADP was hardly detectable. A quantitative analysis of these data is shown in Fig. 3C.

Figure 3.

Reactivity of PphA towards PII-P in the presence of effector molecules.

A. Standard PII-P dephosphorylation assays (as described in Experimental procedures) were carried out in the presence of 1 mM ATP, GTP, CTP, UTP or ADP or as a control without nucleotides (c) and, additionally, 0.5 mM 2-oxoglutarate was added as indicated. To estimate the degree of PII-P dephosphorylation, unreacted substrate PII-P was applied to lane S.

B. Standard PII-P dephosphorylation assays containing either 1 mM ATP or 1 mM ADP, as indicated, were supplemented with various concentrations of 2-oxoglutarate: lanes 1 and 6, none; lanes 2 and 7, 10 μM; lanes 3 and 8, 25 μM; lanes 4 and 9, 50 μM; lanes 5 and 10, 100 μM 2-oxoglutarate.

C. The phosphorylation state of PII in the reactions from (B) was quantified densitometrically, and the inhibition of PphA reactivity was calculated taking the phosphorylation state in the positive control reaction (without effector molecules, lane c) as 0% and that of the substrate control (lane S) as 100% inhibition. 2-Oxoglutarate effect in the presence of ATP (closed diamonds) or ADP (closed squares) is shown.

Figure 4.

ATP or ADP dependence of PphA-catalysed PII-P dephosphorylation in the presence of various concentrations of 2-oxoglutarate. A series of PII-P dephosphorylation reactions was conducted in the presence of none (closed squares), 10 μM (closed diamonds), 25 μM (closed triangles), 100 μM (closed circles) or 500 μM (crosses) 2-oxoglutarate. In each series of experiments, increasing amounts of either ATP (A) or ADP (B) (0, 0.05 mM, 0.1 mM, 0.5 mM, 1 mM and 2.5 mM) were added to the individual reactions. Subsequent to the analysis of the phosphorylation state of PII, inhibition of PphA reactivity was determined (see Fig. 3) and plotted against the respective ATP or ADP concentration.

From these data, it can be inferred that dephosphorylation of PII, depending on the 2-oxoglutarate level in the cells, might respond to the energy status of the cells. To test this hypothesis, the effects of increasing concentrations of ATP or ADP on PII dephosphorylation in the absence or in the presence of 10 μM, 25 μM, 100 μM or 0.5 mM 2-oxoglutarate were tested. A quantitative analysis of the data is shown in Fig. 4A and B. In the presence of low levels of 2-oxoglutarate, dephosphorylation of PII-P responded to differences in ATP levels, which might be of physiological significance. However, at 0.1 mM 2-oxoglutarate, dephosphorylation of PII-P was almost completely inhibited by ATP levels that are well below the physiological range.

As the PII protein in E. coli was reported to respond not only to 2-oxoglutarate but, although with less affinity, to various other metabolites such as glutamate or oxaloacetate (Ninfa and Atkinson, 2000), we used the highly sensitive dephosphorylation assay of PII-P to test the effect of different metabolites on the Synechococcus PII protein. Assays were performed in the presence of 1 mM ATP and 0.5 mM different C-metabolites (2-oxoglutarate, oxaloacetate, malate, succinate, citrate, pyruvate, acetyl-CoA) or 5 mM glutamate or aspartate (Fig. 5). Oxaloacetate caused a significant inhibition of PII dephosphorylation, whereas the other metabolites showed no marked change in the reactivity of PphA towards PII-P, when compared with the reaction with ATP alone.

Figure 5.

Effect of various effector molecules on PphA-catalysed PII-P dephosphorylation. Standard PII-P dephosphorylation assays each containing 1 mM ATP were supplemented by the addition of 0.5 mM (1) 2-oxoglutarate, (2) succinate, (3) oxaloacetate, (4) pyruvate, (5) acetyl-CoA, (6) citrate, (7) malate or 5 mM (8) aspartate or (9) glutamate or (10) without further additions. Substrate control and positive PphA control reactions without ATP are shown (S and C respectively).

A. Immunoblot of a representative experiment.

B. Quantitative analysis of (A).

In an attempt to analyse the mechanism of the ATP effect on PII-P dephosphorylation, we tested different non-hydrolysable ATP analogues in the PphA-PII-Pase assay. In particular, adenosine-5′-[β,γ-methylene]-triphosphate (AMPPCP), adenosine-5′-[β,γ-imido]-triphosphate (AMPPNP) and adenosine-5′-[γ-thio]-triphosphate (ATP-γ-S) were used at 0.5, 1 and 2.5 mM concentration in the presence or absence of 0.5 mM 2-oxoglutarate (Fig. 6). Whereas AMPPNP and ATP-γ-S showed a similar synergy with 2-oxoglutarate as ATP, AMPPCP inhibited the reactivity of PphA towards PII-P in a completely 2-oxoglutarate-independent manner. In the absence of 2-oxoglutarate, AMPPCP inhibited much more strongly than ATP, AMPPNP or ATP-γ-S and, at a concentration of 2.5 mM, it was nearly as effective as ATP plus 2-oxoglutarate. To investigate possible reactions of ATP during the assay, we performed PII-P dephosphorylation tests in the presence of [8-14C]-ATP and separated the reaction mixtures on cellulose-PEI-TLC plates. As revealed by autoradiography, ATP was unaffected by the reaction carried out in the presence or absence of 2-oxoglutarate (data not shown), confirming that no ATP hydrolysis takes place.

Figure 6.

Effect of the non-hydrolysable ATP analogues AMPPCP, ATP-γ-S and AMPPNP on PphA-catalysed PII-P dephosphorylation as shown by immunoblot analysis. Duplicate standard PII-P dephosphorylation assays, of which one of each contained 0.5 mM 2-oxoglutarate, were supplemented with 0.5 mM, 1 mM or 2.5 mM ATP analogues, as indicated. Controls: (C) PphA control reaction in the absence of effector molecules and (S) substrate control.

A. Immunoblot of a representative experiment.

B. Quantitative analysis of (A); the lane numbers correspond to those in the immunoblot.

In further control experiments, we examined whether the different nucleotides and metabolites that affected PphA-catalysed PII-P dephosphorylation had any effect on PphA reactivity towards other substrates. None of these compounds, either added individually or in combination with 2-oxoglutarate, caused any inhibition of phosphocasein or phosphopeptide dephosphorylation (data not shown), implying that these effector molecules do not directly inhibit PphA activity but rather act via the PII protein.


This investigation has revealed that PphA displays the characteristic catalytic properties of the protein phosphatases 2C (Mackintosh, 1993; Cohen, 1994; Shi et al., 1998), in particular with respect to divalent cation dependence and substrate specificity. The exception is its unusual reactivity towards phosphotyrosyl peptides, which deserves further investigation. Comparison of PphA reactivity towards different substrates (low-molecular-weight compounds, phosphopeptides, phosphoproteins) revealed a preference for substrates in which the phosphorylated residue is embedded in a larger peptide environment, as dephosphorylation of phosphocasein is more efficient by greater than three orders of magnitude than that of various phosphopeptides.

At a substrate concentration of 38 nM PII-P (trimers), PphA dephosphorylated its physiological substrate with a turnover rate of ≈ 1 min−1 (Irmler and Forchhammer, 2001), which is very similar to the reactivity towards phosphocasein at the same substrate concentration. Assuming a similar reactivity and affinity of PphA towards PII-P as towards phosphocasein (KM of 0.2 μM), the in vivo reaction towards PII-P would occur at substrate saturation and near Vmax (0.2 s−1). This is because the cellular concentration of PII was estimated to be about 0.1% of total cellular protein (Forchhammer and Tandeau de Marsac, 1994), which corresponds to a cellular concentration in the μM range. This is consistent with the observed rapid dephosphorylation of PII-P in intact cells, which occurs within a few minutes after appropriate treatments such as ammonium shock or inhibition of CO2 fixation (Forchhammer and Tandeau de Marsac, 1995b).

This study confirms our previous suggestion that the inhibitory effect of the low-molecular-weight effectors on PII-P dephosphorylation is based on the ligand-binding properties of the PII protein rather than on direct regulation of phosphatase activity, as the catalytic activity of PphA towards various artificial substrates was not affected by any effector molecule. Sequence alignments showed that the entire PphA sequence matches the PP2C catalytic domain and lacks a putative regulatory domain (Irmler and Forchhammer, 2001). Regulation of PphA activity through the ligand-binding status of its substrate provides a mechanism in which a regulatory domain of the phosphatase is dispensable.

The PII-P dephosphorylation assay using highly purified components and a PII-P preparation that consists almost exclusively of the form PII3 allowed a highly sensitive analysis of PphA–PII-P reactivity and can thus be used to monitor the ligand-binding status of PII-P. This allowed the detection of previously unrecognized effects of various compounds on PII-P. Structural analysis of the E. coli PII homologues GlnB and GlnK revealed three ATP binding sites (Xu et al., 1998; 2001), whereas previous biochemical studies of the Synechococcus and E. coli PII proteins showed that, in the absence of 2-oxoglutarate, only one site is occupied with high affinity (Kamberov et al., 1995Forchhammer and Hedler, 1997). This implies antico-operativity between these binding sites. Binding of an ATP molecule to PII generates a high-affinity 2-oxoglutarate binding site. Conversely, the presence of 2-oxoglutarate increases the affinity for ATP binding. As shown for Synechococcus PII, the presence of 2-oxoglutarate increased the stoichiometry of ATP binding to PII, suggesting that 2-oxoglutarate decreases the antico-operativity of the ATP binding sites (Forchhammer and Hedler, 1997). Here, we showed that, in the absence of 2-oxoglutarate, a concentration of ATP that is sufficient to saturate the high-affinity binding site of the PII protein (KD = 40 μM) is not sufficient to inhibit PII-P dephosphorylation. Therefore, occupation of the high-affinity ATP binding site does not affect PII-P reactivity towards PphA. The high concentrations of ATP (in the absence of 2-oxoglutarate) required to partially inhibit PII-P dephosphorylation might suggest that all three binding sites have to be occupied to inhibit interaction with PphA. Apart from ATP, elevated concentrations of ADP and GTP also caused a partial inhibition of PII-P dephosphorylation, although ligand-binding studies using the equilibrium dialysis method revealed no high-affinity binding sites for these nucleotides. According to our suggestion, these nucleotides could bind to the ATP binding sites with low affinity, not measurable by the equilibrium dialysis technique, and thereby affect the interaction with PphA.

In the presence of 2-oxoglutarate, the concentration of ATP required to achieve PII-Pase inhibition decreases dramatically, whereas only a very weak synergy is observed with ADP and none with GTP. This is consistent with the observation that 2-oxoglutarate binding to PII specifically increases the affinity of the ATP binding sites and decreases their antico-operativity. Therefore, at elevated 2-oxoglutarate concentrations, ATP can bind more easily to the second and third binding site, which would then result in inhibition of PII-P dephosphorylation. Surprisingly, no synergy could be observed between AMPPCP and 2-oxoglutarate for PII-Pase inhibition, although AMPPCP is a closely related structural homologue of ATP. In contrast, the other non-hydrolysable ATP analogues, AMPPNP and ATP-γ-S, mimic ATP, suggesting that the lack of ATP hydrolysis is not responsible for the lack of synergism, as confirmed by further control experiments. The only difference between ATP, AMPPCP and AMPPNP is the atom between the β- and γ-phosphate. Oxygen, nitrogen or carbon display different hydrogen-bonding properties and different angles of the Pβ-C-Pγ, Pβ-N-Pγ, or Pβ-O-Pγ bonds, resulting in slightly different orientations of the γ-phosphate group. This suggests that hydrogen bonding of the oxygen atom between β- and γ-P or the stereoche-mistry of the γ-phosphate group plays a critical role in the synergistic effect exerted by the 2-oxoglutarate molecule and for the antico-operativity between the ATP binding sites. In agreement with this suggestion, the recent resolution of the structure of the PII-ATP complex from E. coli (Xu et al., 2001) showed that, within the ATP-binding pocket, the γ-phosphate group is highly co-ordinated., Taken together, this study has shed some new light on the sophisticated signal integration properties of the PII protein from cyanobacteria. In addition to 2-oxoglutarate, oxaloacetate, which is structurally related to 2-oxoglutarate, displayed a significant effect on ATP-mediated inhibition of PII-P dephosphorylation. Moreover, besides sensing the central metabolites 2-oxoglutarate (and, to some extent, oxaloacetate), PII dephosphorylation is sensitive to the energy status of the cells, as it responds to changes in ATP levels within the physiological range when the oxoglutarate concentration is low. A recent investigation of Rhodospirillum rubrum PII homologues suggested, on the basis of physiological experiments, a role for PII in sensing the energy status of the cells (Zhang et al., 2001). Thus, the nucleotide effects reported in this study might be more widespread among bacterial PII proteins and provide a mechanistic basis for the sensitivity of PII signalling pathways towards the energy status of the cells. In unicellular cyanobacteria, PII-P dephosphorylation is most efficient when both 2-oxoglutarate and ATP levels drop. Such a situation might occur when cells are subjected to a sudden ammonia treatment or to extreme CO2 limitation (Forchhammer and Tandeau de Marsac, 1995b). These conditions will drain the cellular 2-oxoglutarate pool and might affect the ATP level through activation of the highly energy-consuming carbon concentration mechanism (Kaplan et al., 1994) or uncoupling by ammonia.

Experimental procedures


Fine chemicals used in this study were purchased from Sigma-Aldrich, Carl Roth or Roche Diagnostics and were of the highest quality available. Stock solutions (200 mM) of nucleotides and other effector molecules were prepared, and the pH of the solution was determined. Acidic solutions were adjusted to pH 7.5 by titrating them with Tris base.

Preparation of PphA and PII-P

PphA was purified from a recombinant E. coli strain over-producing the pphA gene from Synechocystis PCC 6803 according to the purification protocol described recently (Irmler and Forchhammer, 2001). The resulting protein was almost homogeneous, as assessed from silver-stained SDS gels.

As Ser-49-phosphorylated PII cannot be prepared from PII0 by in vitro phosphorylation, PII-P was purified from Synechococcus PCC 7942 cells as described previously (Irmler and Forchhammer, 2001). To achieve a maximal degree of in vivo PII phosphorylation, the cells were grown under nitrogen-limiting conditions. Preparation of PII-P from 15 g of wet cell pellet resulted in ≈ 150 μg of highly pure PII-P, containing predominantly the highest phosphorylated form PII3 and small amounts of PII2. Small aliquots of the protein pre-paration were stored at –70°C.

PphA reactivity towards phosphocasein

Dephosphorylation of 32P-labelled phosphocasein was determined as described previously (Irmler et al., 1997) using freshly prepared phosphocasein that had been phosphorylated with [γ-32P]-ATP (2 mCi μmol−1; Amersham Pharmacia) according to the procedure of Kennelly et al. (1993). Each dephosphorylation reaction was carried out in a volume of 25 μl containing 1.5 ng of purified PphA and, at the indicated times, the reactions were stopped by the addition of ice-cold stop mix (Howell et al., 1996). Released 32Pi was quantified by scintillation counting.

PphA reactivity towards PNPP

PNPP phosphatase activity was assayed according to the method of Mackintosh, (1993). In a 1 ml standard assay, 2 μg of purified PphA was reacted with 5 mM PNPP in a buffer consisting of 10 mM Tris-Cl, pH 8.5, 2 mM MnCl2, 1 mM dithiothreitol (DTT) and 50 mM NaCl. The increase in absorbance at 400 nm was recorded in an Ultrospec 3000 spectrophotometer (Amersham-Pharmacia) against a blank reaction without enzyme. The specific activity of PNPP hydrolysis was calculated with a molar extinction coefficient for p-nitrophenol of 16 500 l mol−1 cm−1. For the determination of apparent KM, reactions were performed with substrate concentrations between 0.1 and 5 mM PNPP.

PphA reactivity towards phosphopeptides

Phosphopeptide dephosphorylation was measured using the non-radioactive serine/threonine or tyrosine phosphatase assay system from Promega according to the manufacturer’s protocol. In a standard assay, 2.4 μg of purified PphA was reacted with 100 μM phosphopeptide in a reaction volume of 50 μl containing 10 mM Tris-Cl, pH 7.4, 1 mM DTT, 50 mM NaCl, 5 mM benzamidine and 2 mM MnCl2, unless otherwise stated. Reactions were incubated at 30°C for 30 min, then stopped by the addition of molybdate–dye mix, and the released Pi was quantified colorimetrically. For the determination of kinetic parameters, substrate concentrations from 40 μM to 400 μM were used.

PphA reactivity towards PII-P

To assay the reactivity of PphA towards PII-P, 14 ng of purified PII-P was incubated with the indicated amount of PphA in a 10 μl reaction volume of 10 mM Tris-Cl, pH 7.4, 50 mM NaCl, 1 mM DTT, 10 μg BSA, 0,05% NP40 and 5 mM MgCl2, unless otherwise indicated. After 20 min at 37°C, the reactions were stopped by chilling on ice, followed immediately by separating the different phosphorylated forms of PII by non-denaturing PAGE and detection of PII by immunoblot analysis as described previously (Forchhammer and Tandeau de Marsac, 1994). As in the highly purified system, the PII protein was unstable during electrophoresis (the bands disappeared in a smear), the non-denaturing gel-loading buffer was modified by the addition of 1 mg ml−1 bovine serum albumin (BSA) and 10% xylitol. The phosphorylation state of PII (ratio of phosphorylated PII subunits to total PII subunits) was quantified by densitometry basically as described previously (Irmler et al., 1997); however, image analysis was carried out on a Bio-Rad Fluor S Imager together with the QUANTITYONE software. To calculate the degree of inhibition of PII-P dephosphorylation in different assays, the corresponding PII phosphorylation state was determined, and the difference from the phosphorylation state of the unreacted PII-P preparation was calculated (quantified PII-P dephosphorylation). The quantified PII-P dephosphorylation of a positive control reaction (without the addition of effector molecules) was defined as 100% activity, corresponding to 0% inhibition.

Analysis of ATP hydrolysis

Potential hydrolysis of ATP during PII-P dephosphorylation reactions was investigated by performing standard PII-Pase assays (see above) in the presence of 0.1 mM [8-14C]-ATP (10 nCi) and different amounts of 2-oxoglutarate. After 20 min incubation at 37°C, the 10 μl reactions were spotted on a thin-layer CEL 300 PEI/UV254 plate (Macherey and Nagel) and developed with 0.5 M sodium phosphate, pH 3.4. Non-radioactive standard nucleotides were co-chromatographed and visualized by UV shadowing. The plates were exposed to a Kodak Imaging Screen K (Eastman Kodak) and evaluated on a Bio-Rad Molecular Imager.


We thank Dr Sawers (Norwich, UK) for critical reading of the manuscript, Dr Treuner-Lange for helpful discussions, and Annette Heinrich for preparation of purified PphA. This work was supported by a grant from the Deutsche Forschungsgemeinschaft (Fo 195/4-1).