Isocitrate dehydrogenase kinase/phosphatase. Kinetic characteristics of the wild-type and two mutant proteins.

Isocitrate dehydrogenase (IDH) of Escherichia coli is regulated by a bifunctional protein, IDH kinase/phosphatase. In addition to the kinase and phosphatase activities, this protein catalyzes an intrinsic ATPase reaction. The initial velocity kinetics of these activities exhibited extensive similarities. IDH kinase and phosphatase both yielded intersecting double-reciprocal plots. In addition, we observed similar values for the kinetic constants describing interactions of the kinase and phosphatase with their protein substrates and the interactions of all three activities with ATP. In contrast, while the maximum velocities of IDH kinase and IDH phosphatase were nearly equal, they were 10-fold less than the maximum velocity of the ATPase. Although the IDH phosphatase reaction required either ATP or ADP, it was not supported by the nonhydrolyzable ATP analogue 5′-adenylyl imidodiphosphate. The kinetic properties of wild-type IDH kinase/phosphatase were compared with those of two mutant derivatives of this protein. The mutations in these proteins selectively inhibit IDH phosphatase activity. Inhibition of IDH phosphatase resulted from three factors: decreases in the maximum velocities, reduced affinities for phospho-IDH, and a loss of coupling between ATP and phospho-IDH. These mutations also affected the properties of IDH kinase, increasing the maximum velocities and decreasing the affinities for ATP and phospho-IDH. The intrinsic ATPase activities also exhibited reduced affinity for ATP. These results are discussed in the context of a model which proposes that all three activities occur at the same active site.

In Escherichia coli, the Krebs cycle enzyme isocitrate dehydrogenase (IDH) 1 is regulated by phosphorylation (1,2). The phosphorylation and dephosphorylation of IDH are catalyzed by a bifunctional protein, IDH kinase/phosphatase. IDH kinase/phosphatase is expressed from a single gene, aceK, and both activities reside on the same polypeptide (3)(4)(5). The phosphatase activity of this protein has an absolute requirement for ATP or ADP (5). IDH kinase/phosphatase also catalyzes an ATPase reaction. This ATPase activity occurs in the absence of the protein substrates and is only partially inhibited when these substrates are present (6).
One approach that we have taken to examine IDH kinase/ phosphatase has been the random mutagenesis of aceK. We previously described the isolation of a class of mutant alleles whose products retain IDH kinase activity but have dramatically reduced IDH phosphatase activity. The products of these alleles have suffered single amino acid substitutions that fall within a 113-amino acid region of this 578-residue protein (7,8).
The regulation of IDH kinase/phosphatase has been under study for nearly 15 years (reviewed in Ref. 9). In contrast, relatively little attention has been given to the enzymology of this unusual protein. This paper presents the results of an initial characterization of the kinetic properties of IDH kinase/ phosphatase and of the products of two of the mutant derivatives: AceK3 (Gln 373 3 Arg) and AceK4 (Tyr 414 3 Cys) (7,8).

EXPERIMENTAL PROCEDURES
Materials-[␥-32 P]ATP was obtained from DuPont NEN and further purified by the method of Axelson et al. (10). Adenylyl imidodiphosphate was purchased from Sigma and purified by the method of Olson et al. 2 All reagents were of the purest grades available.
Plasmid Construction-Plasmids pDL9, pDL25, and pDL30 express the aceK ϩ , aceK3, and aceK4 genes, respectively, from the tac promoter (7). To maintain plasmid selection during high-density fermentation, derivatives of these plasmids were constructed which contain the tet r gene for selection. Briefly, the 1.6-kilobase SspI-AvaI fragment of plasmid pBR322 containing the tet r gene was isolated, filled in with Klenow, and BglII linkers were attached. The resulting fragment was digested with BglII and then ligated into the BamHl site of pDL9, pDL25, and pDL30 generating the plasmids pEKL1, pEKL5, and pEKL3, respectively.
Protein Purification-IDH was overexpressed in E. coli strain Y1090 harboring the multicopy plasmid pTK509 which carries the gene encoding IDH (4). IDH was purified by a modification of the method of Garnak and Reeves (11).
Phospho-IDH was prepared in vitro using purified IDH and wild-type IDH kinase/phosphatase. Phospho-IDH was then separated from IDH kinase/phosphatase by chromatography on an Affi-Gel blue column (12).
Wild-type IDH kinase/phosphatase was purified from E. coli strain ST2010R harboring the plasmid pEKL1. The mutant AceK3 and AceK4 proteins were purified from the E. coli strain SL1R (⌬aceK::Kan r ) (13) harboring either plasmid pEKL5 or pEKL3, respectively. All strains were grown at 37°C in L broth (1% tryptone, 1% NaCl, 0.5% yeast extract) with 40 g/ml tetracycline in a 3.5-liter high-density fermentor. Expression of the aceK genes was induced with 3 mM isopropyl-␤-Dthiogalactopyranoside during the mid-log phase of growth and then the strains were grown for an additional 3-4 h. The cells were harvested by centrifugation and stored at Ϫ80°C. The purification of IDH kinase/ phosphatase has been described previously (5).
Protein concentrations were determined by the method of Lowry et al. (14) using bovine serum albumin as the standard.
Enzyme Activity Determinations-The activity of IDH phosphatase was measured by following the release of [ 32 P]phosphate from [ 32 P]phospho-IDH as described previously (12). The standard reaction (50 l) contained 25 mM MOPS (pH 7.5), 100 mM NaCl, 5 mM MgCl 2 , 0.5 * This work was supported by National Institutes of Health Grant DK40486.1. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. ‡ Present address: Dept. of Plant Pathology, University of California, Berkeley, CA 94720.
The activity of IDH kinase was determined by monitoring the transfer of [ 32 P]phosphate from [␥-32 P]ATP to IDH as described previously (12). The standard reaction mixture (50 l) contained 25 mM MOPS (pH 7.5), 100 mM NaCl, 5 mM MgCl 2 , 0.5 mM EDTA, 1 mM dithiothreitol, 0.2 mg/ml bovine serum albumin, and variable concentrations of Mg⅐[␥-32 P]ATP (ϳ3 ϫ 10 6 cpm/assay) and IDH. The reactions were initiated by the addition of 5.4 ng of wild-type, 1.0 ng of AceK3, or 1.8 ng of AceK4 IDH kinase/phosphatase and then incubated for 20 min at 37°C. The production of [ 32 P]phospho-IDH was monitored by precipitating the protein onto filter paper and measuring the amount of radioactivity.
The bifunctional nature of IDH kinase/phosphatase can interfere with these assays. For example, [ 32 P]phospho-IDH produced by IDH kinase could be lost due to dephosphorylation by IDH phosphatase. Control experiments established that product loss in the kinase assay was not significant under the conditions that we employed. In addition, phosphorylation of dephospho-IDH produced in the IDH phosphatase assay could dilute the specific activity of the [ 32 P]phospho-IDH used as the substrate. However, this did not interfere with our results, because substrate conversion was always less than 20%.
ATPase activity was measured using the method of Bais (15). The reactions (50 l) contained 25 mM MOPS (pH 7.5), 100 mM NaCl, 5 mM MgCl 2 , 0.5 mM EDTA, 1 mM dithiothreitol, 0.2 mg/ml bovine serum albumin, and variable concentrations of Mg⅐[␥-32 P]ATP (ϳ5 ϫ 10 5 cpm/ assay). The reactions were initiated by the addition of 140 ng of wildtype, 260 ng of AceK3, or 920 ng of AceK4 IDH kinase/phosphatase and then incubated at 37°C. The amount of [ 32 P]phosphate released was determined following absorption of the unreacted ATP with activated charcoal.

Initial Velocity Kinetics of Wild-type IDH Phosphatase-
Measurements of the initial velocity of wild-type IDH phosphatase produced intersecting patterns on double-reciprocal plots (Fig. 1A). Such plots are typical of a sequential mechanism (i.e. ordered or random). Sequential mechanisms can be described by Equation 1, where A and B are the substrate concentrations, v is the initial velocity, V max is the maximum velocity, K a is the apparent dissociation constant for A, and K ma and K mb are the Michaelis constants for A and B, respectively (16). Estimates of the kinetic parameters were determined from Eq. 1 using the computer program SEQUENL developed by Cleland (17). The results are summarized in Table I. The apparent dissociation constant of wild-type IDH phosphatase for ATP (K ATP ϭ 98 M) was much greater than the corresponding Michaelis constant (6 M). A similar relationship was observed for the kinetic constants for phospho-IDH (6 and 0.4 M, respectively). The ratio of these constants, ␣ ϭ K/K m , provides a measure of the apparent substrate interaction, where ␣ ϭ 1 indicates independent binding. The ␣-value of ϳ17 for wild-type IDH phosphatase suggests that substrate binding is coupled (i.e. the binding of one substrate enhances the binding of the other substrate). Initial Velocity Kinetics of the Mutant IDH Phosphatases-Like wild-type IDH phosphatase, the residual IDH phosphatase of the AceK3 protein produced intersecting double-reciprocal plots (Fig. 1B). Similar patterns were observed with the AceK4 IDH phosphatase (data not shown). However, the single amino acid substitutions in the mutant proteins have produced striking changes in the kinetic properties of the residual IDH phosphatase activities.
The maximal velocities of both mutant IDH phosphatases were decreased relative to the wild-type enzyme. The V max of the AceK3 phosphatase was decreased by a factor of 2.5, while the AceK4 phosphatase had suffered a 5-fold reduction in V max ( Table I).
The aceK3 and aceK4 mutations have also produced dramatic changes in the kinetic behavior of IDH phosphatase with respect to ATP (Table I). The K m,ATP of AceK3 IDH phosphatase increased 9-fold, while the value of K ATP decreased by a factor of 2 when compared with that of the wild-type enzyme. The AceK4 product exhibited increases in the K m,ATP and K ATP values of 20-and 3-fold, respectively.
The mutant IDH phosphatases also exhibited dramatic changes in the kinetic parameters for phospho-IDH (Table I).
The K m,IDH-P and K IDH-P values of AceK3 IDH phosphatase were increased 52-and 3-fold, respectively, relative to those of the wild-type protein, while the AceK4 protein showed a 48-fold increase in K m,IDH-P and 7-fold increase in K IDH-P .
The coupling between ATP and phospho-IDH that we observed for wild-type IDH phosphatase has also been effected by these mutations. The AceK4 protein exhibited a small residual coupling, while the AceK3 protein exhibited no coupling at all.
Initial Velocity Kinetics of the IDH Kinases-Like IDH phosphatase, wild-type IDH kinase exhibited intersecting patterns in double-reciprocal plots ( Fig. 2A). We observed strong coupling between ATP and dephospho-IDH, which is also similar to the behavior of wild-type IDH phosphatase (Table I).
The initial velocity patterns of the AceK3 (Fig. 2B) and AceK4 (data not shown) IDH kinases resembled those of the wild-type kinase. The mutant kinases retained affinity for dephospho-IDH and coupling between this substrate and ATP  (Table I). However, the maximum velocities were increased by 5-fold for both proteins. In addition, the affinities of these kinases for ATP were decreased by factors of 3 and 30 for AceK3 and AceK4, respectively.
The mutations in aceK3 and aceK4 have had strikingly different effects on the kinetic characteristics of IDH kinase and IDH phosphatase. The maximum velocities of the kinases have increased, while those of the phosphatase have decreased. The kinase activities retained their affinity for the protein substrate (dephospho-IDH), while the residual phosphatase activities exhibited reduced affinity for its protein substrate (phos-pho-IDH). Finally, the AceK3 and AceK4 IDH kinases retained positive coupling between ATP and the protein substrate while their phosphatase activities did not.
Inhibition of Wild-type and Mutant IDH Kinase by Phospho-IDH-Since the aceK3 and aceK4 mutations alter the affinity of IDH phosphatase for phospho-IDH, we also examined the effect of these mutations on the affinity of IDH kinase for phospho-IDH. The affinities for phospho-IDH were estimated from product inhibition studies.
The bifunctional nature of wild-type IDH kinase/phosphatase posed inherent complications for analysis of the inhibition of IDH kinase by phospho-IDH. For example, the sequential dephosphorylation and rephosphorylation of the added phospho-IDH could have resulted in the incorporation of 32 P into the phospho-IDH pool. To evaluate this source of interference, we performed IDH kinase assays in the absence of dephospho-IDH (Fig. 3A, solid circles). We observed a low level of 32 P incorporation into protein, and this effect was saturated at the lowest concentration of phospho-IDH, which we employed. The precise mechanism responsible for 32 P incorporation into IDH under these conditions is uncertain, but this effect must be considered when evaluating the ability of phospho-IDH to act as an inhibitor of IDH kinase.
Wild-type IDH kinase was inhibited by phospho-IDH (Fig.  3A, open circles). The precise value of the inhibition constant cannot be determined since 32 P incorporation into the phospho-IDH pool contributed to the observed IDH kinase activity (see above). However, we were able to calculate a range for this constant by subtracting or ignoring the incorporation observed in the absence of dephospho-IDH ( Fig. 3B and Table I). The range of this constant, 8 -15 M, is negligible compared with the effects that the aceK3 and aceK4 mutations had on the affinities of the mutant IDH kinases for phospho-IDH (see below).
We have also examined inhibition of the mutant IDH kinases by phospho-IDH. In contrast with wild-type IDH kinase/phosphatase, label was not incorporated into IDH in the absence of dephospho-IDH. (This was the expected result since the AceK3 and AceK4 proteins retain very little IDH phosphatase activity.) To aid the comparison of wild-type and mutant IDH kinases, we employed concentrations of ATP and dephospho-IDH, which were in the psuedo-first order range for all three enzymes. The mutant activities exhibited dramatic increases in the apparent K i values for phospho-IDH relative to the wildtype protein (Table I). These values were at least 20-fold higher for the AceK3 and AceK4 kinases. The effective K m,IDH-P of the IDH phosphatases were increased to similar extents at this concentration of ATP: 30-fold for AceK3 and 47-fold for AceK4.

Initial Velocity Kinetics of the Wild-type and Mutant ATPase
Activities-In addition to the IDH kinase and IDH phosphatase reactions, IDH kinase/phosphatase exhibits an intrinsic ATPase activity (6). The kinetic constants of the wild-type and mutant ATPase activities were estimated from double reciprocal plots (Table I). The maximum velocities of all three proteins greatly exceeded the corresponding values for the kinase and phosphatase activities. Although the V max values of the mutant and wild-type proteins were similar, the K m,ATP of the AceK3 and AceK4 ATPase activities are increased 3-and 12-fold, respectively.
A Nonhydrolyzable ATP Analogue Does Not Support the IDH Phosphatase Reaction-IDH phosphatase exhibits an absolute requirement for either ATP or ADP (5). Our working model proposes that the actual substrate for IDH phosphatase is ADP but that ATP can be converted to ADP by the ATPase activity (see "Discussion"). To test this hypothesis, we examined the ability of a nonhydrolyzable analogue of ATP, 5Ј adenylyl imidodiphosphate (AMPP(NH)P), to support the wild-type IDH phosphatase reaction. This nucleotide clearly binds to IDH phosphatase, since it inhibited this reaction in the presence of ATP (Table II). However, in the absence of ATP, AMPP(NH)P did not support IDH phosphatase activity. This observation is consistent with the proposal that ATP must be converted to ADP before it can participate in the IDH phosphatase reaction. DISCUSSION IDH kinase/phosphatase possesses three activities: IDH kinase, IDH phosphatase, and an ATPase. The initial velocity behaviors of these activities were similar for the wild-type protein. For example, the apparent dissociation constants for the protein substrates for IDH kinase (K IDH ) and IDH phosphatase (K IDH-P ) were 2.4 and 6 M. The apparent dissociation constant for ATP of the kinase and phosphatase and the K m of the ATPase activity were also similar. The kinase and phosphatase exhibited strong coupling (␣) between ATP and the protein substrates. In contrast, although the maximum velocities of IDH kinase and phosphatase were nearly identical, the V max of the ATPase was about an order of magnitude greater. The significance of this difference is discussed below.
We previously reported the isolation of mutant derivatives of IDH kinase/phosphatase (7,8). These proteins retained IDH kinase activity but had suffered striking reductions in IDH phosphatase activity. Our kinetic analyses provide insights into the mechanism responsible for this selective loss of IDH phosphatase activity of two of these proteins: AceK3 (Gln 373 3 Arg) and AceK4 (Tyr 414 3 Cys). Three factors have contributed to this loss: decreases in the maximum velocities, reduced affinities for the protein substrate (phospho-IDH), and a loss of coupling between ATP and phospho-IDH. These proteins retained IDH kinase activity, because binding of dephospho-IDH and coupling between ATP and dephospho-IDH were unaffected. In addition, the decrease in the affinity of these kinase activities for ATP was offset by an increase in the maximum velocities.
We have developed a working model that describes IDH kinase/phosphatase. This model proposes that the kinase, phosphatase, and ATPase reactions occur in the same active site. This active site can exist in two conformations, one favoring IDH kinase and the other favoring IDH phosphatase. The phosphatase reaction results from the kinase back reaction tightly coupled to the ATPase reaction (Fig. 4). Although this model is speculative, it is supported by a variety of observations.
The model explains the unusual nucleotide requirement of IDH phosphatase: this activity has an absolute dependence on either ATP or ADP. In contrast, this activity is not supported by a nonhydrolyzable analogue of ATP. Our model proposes that ADP acts as a transient phosphate acceptor. A similar role is played by the active site serine of E. coli alkaline phosphatase (18).
This model also explains the unusual ATPase activity of IDH kinase/phosphatase. Weak ATPase activities are often associated with kinases in the absence of the normal phosphate acceptor. In contrast, the V max of the ATPase activity of IDH kinase/phosphatase is almost an order of magnitude greater than that of either the kinase or phosphatase (Table I). Fur- FIG. 3. Product inhibition of IDH kinase activity by phospho-IDH. A, wild-type IDH kinase was assayed as described under "Experimental Procedures" using 100 M Mg⅐ATP and the indicated concentration of phospho-IDH. The assays were performed in the presence (E) and absence (q) of 0.5 M dephospho-IDH. B, Dixon plots of product inhibition of wild-type and mutant IDH kinase activity by phospho-IDH. The kinase assay was performed as described under "Experimental Procedures" using 100 M Mg⅐ATP and 0.5 M IDH. Inhibition of the wild-type (E), AceK3 (q), and AceK4 (Ⅺ) IDH kinase activities was measured as a function of varied phospho-IDH concentration. Parallel assays were performed without dephospho-IDH for the wild-type enzyme, and the observed 32 P incorporation was subtracted from that observed with dephospho-IDH.  (19). Mutation of the "invariant" lysine of this motif eliminated both IDH kinase and IDH phosphatase without destabilizing the protein (13). The results of affinity labeling experiments are also consistent with a single ATP binding site: azido-ATP produced parallel inhibition of IDH kinase and IDH phosphatase yet labeled a single peptide (20).
Our working model provides a framework for interpretation of the kinetic characteristics of the mutant proteins. We suspect that the mutations in these proteins have shifted the conformational equilibrium toward the kinase conformation at the expense of IDH phosphatase. However, it should be noted that this is not the only model that could explain the results presented in this paper.
The IDH kinase and IDH phosphatase activities of the mutant protein exhibited reciprocal changes in their maximum velocities: the V max of the kinase was increased while that of the phosphatase was decreased. This might have resulted from a shift in the conformational equilibrium toward the kinase form.
The IDH kinase and phosphatase activities exhibited a parallel loss of affinity for phospho-IDH. This observation tends to favor a single binding site for phospho-IDH, a feature of our model.
Effects of these mutations on the K ATP values for IDH kinase and IDH phosphatase and on the K m,ATP value for the ATPase are more ambiguous. In the simplest case, these values should be roughly equal if all three activities occur at the same active site. These values were similar for the wild-type protein, consistent with the evidence that favors a single ATP binding site (see above). However, while the values for the kinase and ATPase agreed for the mutant proteins, the value of K ATP for the IDH phosphatase activities did not. If our model is correct, this discrepancy might reflect differences in the properties of the kinase and phosphatase conformations of the mutant proteins. However, this issue remains to be resolved. Why would IDH kinase/phosphatase catalyze the phosphorylation and dephosphorylation of IDH at the same active site? Such an arrangement would allow these reactions to employ some of the same catalytic groups. According to our model, IDH kinase/phosphatase employs the kinase back reaction as part of the IDH phosphatase reaction. In addition, the protein substrates of IDH kinase and phosphatase are nearly identical, differing only in the phosphorylation of Ser 113 . Catalysis of both reactions at the same site would allow many of the same amino acids to be used for substrate recognition. The ultimate resolution of these issues will require the crystal structure of IDH kinase/phosphatase complexed with its substrates.