Reversibility of phosphorylase kinase reaction.

(Received for publication, May 17, 1976, and in revised form, January 3, 1977) YUTAKA SHIZUTA,$ RAMJI L. KHANDELWAL,~ JAMES L. MALLER,~ JACKIE R. VANDENHEEDE,~~ AND EDWIN G. KREBS** From the Department of Biological Chemistry, School of Medicine, University of California, Davis, California 95616 Using a highly purified enzyme from rabbit skeletal mus- cle, it has been demonstrated that the glycogen phosphoryl- ase kinase reaction is reversible. In addition to the phospho- rylated protein substrate, phosphorylase a, the reverse reac- tion requires Mg’+, ADP, Ca*+, and glucose. Glucose cannot be replaced by glucose 6-phosphate or glucose l-phosphate but can be partially replaced by glycogen. The half-maximal concentration of glucose required for the reverse reaction is approximately 25 mM at pH 8.2 and 30”. Ultracentrifugation experiments have indicated that glucose exerts its effect by facilitating the formation of a dimer from the tetrameric form of phosphorylase a, the major form present under the reaction conditions used. The quaternary structure of phos- phorylase kinase itself is not influenced by the presence of glucose. These results suggest that the kinase catalyzes the reversible transfer of phosphate using as substrate only the dimeric forms of phosphorylase a and b. In contrast to what is seen for the torward reaction, the optimum pH for the reverse reaction is in the range of 6.8 to 7.6 and is not markedly changed when the kinase itself is activated by phosphorylation using the cyclic 3’:5’-AMP-dependent pro- tein kinase. Among the nucleotide substrates tested, ADP and GDP serve as the best phosphoryl group acceptors. The saturation curve of the enzyme for ADP is biphasic with an approximate so,5 value of 10 mM whereas that for phospho- rylase a gives a hyperbolic curve with an approximate

Using a highly purified enzyme from rabbit skeletal muscle, it has been demonstrated that the glycogen phosphorylase kinase reaction is reversible. In addition to the phosphorylated protein substrate, phosphorylase a, the reverse reaction requires Mg'+, ADP, Ca*+, and glucose. Glucose cannot be replaced by glucose 6-phosphate or glucose l-phosphate but can be partially replaced by glycogen. The half-maximal concentration of glucose required for the reverse reaction is approximately 25 mM at pH 8.2 and 30". Ultracentrifugation experiments have indicated that glucose exerts its effect by facilitating the formation of a dimer from the tetrameric form of phosphorylase a, the major form present under the reaction conditions used. The quaternary structure of phosphorylase kinase itself is not influenced by the presence of glucose. These results suggest that the kinase catalyzes the reversible transfer of phosphate using as substrate only the dimeric forms of phosphorylase a and b. In contrast to what is seen for the torward reaction, the optimum pH for the reverse reaction is in the range of 6.8 to 7.6 and is not markedly changed when the kinase itself is activated by phosphorylation using the cyclic 3':5'-AMP-dependent protein kinase. Among the nucleotide substrates tested, ADP and GDP serve as the best phosphoryl group acceptors. The saturation curve of the enzyme for ADP is biphasic with an approximate so,5 value of 10 mM whereas that for phosphorylase a gives a hyperbolic curve with an approximate K,,, value of 40 mglml.
Muscle phosphorylase kinase catalyzes the conversion of phosphorylase b to a resulting in the activation of the enzyme The terminal phosphate of ATP is transferred to a specific seryl residue (2) in each subunit of phosphorylase a, which has a tetrameric structure. Heretofore, no reversal of this reaction could be demonstrated by any of several means employed (1). Studies from this laboratory to determine the mechanism of action of adenosine 3':5'-monophosphate in stimulating glycogenolysis revealed that phosphorylase kinase in skeletal muscle is activated by phosphorylation catalyzed by a separate protein kinase, an enzyme that was designated as the cyclic AMP'-dependent protein kinase (3)(4)(5). The properties of this latter enzyme as well as those of phosphorylase kinase have been discussed in recent reviews (6)(7)(8). The methods for obtaining homogeneous preparations of both kinases from rabbit skeletal muscle have also been described (9,10).
It has been demonstrated that several protein kinase reactions are reversible. Rabinowitz and Lipmann (11) first reported that phosvitin kinase, which was partially purified from yeast or brain, catalyzed a reversible reaction when phosphorylated phosvitin was used as the substrate. Lerch et al. (12) purified a similar enzyme to homogeneity from bakers' yeast and also observed that it catalyzed a reversible reaction using casein or phosvitin as the substrate. Shizuta et al. (13) demonstrated that homogeneous rabbit muscle cyclic AMPdependent protein kinase catalyzed a reversible phosphorylation reaction when any of several protein substrates were used. Rosen and Erlichman (14) showed a reversible transfer of phosphate in the autophosphorylation reaction of bovine heart cyclic AMP-dependent protein kinase. In the present study, it was found that the phosphorylase kinase reaction, like the other protein kinase reactions described above, is also reversible. The probable reason for the discrepancy between the present finding and that previously reported (1) lies in the fact that only the dimeric form of phosphorylase a will serve in this process.   an E:,X:: value of 11.8 (10). The concentrations of other proteins were determined by the method of Lowry et al. (21). All other analytical methods were performed as described previously (13).

Glucose-dependent Phosphate
Transfer from ["'PJPhosphorylase a to ADP-As noted above, an early attempt to demonstrate the occurrence of a reverse reaction catalyzed by phosphorylase kinase was unsuccessful (1). In the present study, however, it was noted that when reaction mixtures were supplemented with a relatively high concentration of glucose, liberation of radioactivity from [3"P]phosphorylase a into a trichloroacetic acid-soluble form occurred in the presence of ADP (Fig. 1). This reaction proceeded linearly for 30 min. On the other hand, almost no radioactivity was released in the absence of ADP. The radioactive reaction product liberated into the trichloroacetic acid-soluble fraction was isolated by adsorption on charcoal and identified as [Y-~*PIATP as described under "Methods." More than 90% of the reaction product was adsorbed by the charcoal and there was no indica- tion to suggest that glucose was directly involved in the reaction, i.e. that any sugar/phosphate compound was being formed. These findings indicate that the phosphorylase kinase reaction is reversible under these conditions.
In the experiment shown in Fig. 2, the effect of phosphorylase kinase concentration on the liberation of radioactivity from ["'Plphosphorylase was examined in the presence and absence of ADP. As indicated in this figure, the amount of radioactivity liberated from [32Plphosphorylase a in the presence of ADP ( i .e the amount of [y -"'P]ATP formed) was linear up to an enzyme concentration of 200 pg/ml. Another set of experiments was performed to see what components were required for the reverse reaction. The results are summarized in Table I. In addition to ADP, Ca" and Mg2+ were essential for the glucose-dependent reaction, consistent with the known requirements of the enzyme in catalysis of the forward reaction (3,22). Characterization of Glucose Effect-It was found that the glucose effect in the reverse reaction became less marked when an aged preparation of phosphorylase a was employed as the substrate. Namely, in this case the reverse reaction proceeded to a considerable degree without glucose. Furthermore, procedures that might modify the conformation of phosphorylase a, such as prolonged dialysis against TGP buffer at 30-37 or freezing and thawing, facilitated the reverse reaction in the absence of glucose. These observations, together with the reports by other investigators (23,241, suggested that glucose exerts its effect by modifying the conformation of the protein substrate, phosphorylase a. The effects of various substances, including other ligands, on the reverse reaction of phosphorylase kinase were then examined using a fresh crystalline preparation of phosphorylase a. Table II shows the results of a typical experiment. It is seen that glucose was most effective and glycogen was partially effective in promoting the reverse reaction. Glucose could not be replaced by other compounds such as glucose l-phosphate or glucose 6-phosphate under the conditions employed, although there is a slight suggestion that the latter compound had some effect.
In the experiments shown in Fig. 3, the effect of glucose concentration on the reverse reaction of the kinase was examined using two different concentrations of phosphorylase a. The results indicate that the concentration of glucose giving a half-maximal effect is approximately 25 mM whereas 80 to 200 II~M glucose is required for the maximal effect.
It was determined that at a concentration of glucose causing a maximal increase in velocity of the reverse reaction, i.e. 0.1 M, phosphorylase a is apparently completely dissociated from its tetrameric to its dimeric form (Fig. 4). Thus, in the absence of glucose, the enzyme sedimented as a single somewhat asymmetric peak with an stow value of 13.1 whereas in the presence of 0.1 M glucose the sgo,u, value was 8.6. The symmetry of the peak was noticeably greater under the latter condition. These sedimentation values correspond closely to those reported for the tetrameric and dimeric forms of phosphorylase respectively. On the other hand, glucose did not induce a marked change m the s20,11' value of phosphorylase kinase.3 3 The effects of elevated temperature and high salt on the reverse reaction and on the quaternary structure of phosphorylase a were was approximately 9. The ratio of activity at pH 6.8 to activity at 8.2 was 0.06, as is characteristic for nonactivated (nonphosphorylated) phosphorylase kinase (4). When the enzyme was activated by preincubation with MgATP in the presence of the catalytic subunit of cyclic AMP-dependent protein kinase (see "Experimental Procedures"), the pH 6.8 to pH 8.2 activity ratio increased to 0.5 for the forward reaction as anticipated (3,4,15). In contrast, the optimum pH of the reverse reaction was in the range of 6.8 to 7.6 before and after the enzyme was activated by the cyclic AMP-dependent protein kinase catalytic subunit.
(The data for the activated form of the kinase are not illustrated.) As shown in Fig. 6 using three different concentrations of phosphorylase a gave biphasic curves (Fig. 7). Table III shows the results of a typical experiment in which the nucleotide substrate specificity of the enzyme for the reverse reaction was examined. The best nucleotide substrate was ADP, but GDP also served as a substrate for the reverse reaction. Other nucleotides did not serve effectively as acceptors of phosphate.

DISCUSSION
The present study has shown that phosphorylase kinase can catalyze the reversal of the phosphorylase b to phosphorylase a reaction under certain conditions. This conclusion is based on also examined, since it has been reported that these conditions facilitate the dissociation of phosphorylase a into ita dimeric form (25-2'7) as do glucose (23,241 and several other carbohydrates including glycogen (28)(29)(30). Neither conditions stimulated the reverse reaction. It is possible, however, that the high salt concentration (see Table II) may have caused almost complete inhibition of the enzyme even in the forward reaction (3). It is also possible that elevation of the temperature (to 37") did not cause appreciable dimer formation in the reaction mixture due to the presence of 10 mxe ADP. The reverse reaction occurs only in the presence of glucose when fresh crystalline preparations of the phosphoprotein substrate, phosphorylase a, are used. The reverse reaction does proceed, however, in the absence of glucose after phosphorylase a preparations have been aged or modified by any of several different procedures. Ultracentrifugation experiments showed that glucose caused the protein substrate, phosphorylase a, to dissociate into a dimeric form. However, glucose did not influence the quaternary structure of phosphorylase kinase. These findings suggest that for the reverse reaction phosphorylase kinase recognizes the dimeric form of phosphorylase a. Since association-dissociation of phosphorylase a can be achieved nonenzymatically under a certain set of conditions (23-301, it seems reasonable to conclude that the overall reaction of phosphorylase kinase as described previously (1) is composed of two partial reactions and that the apparent irreversibility of the reaction is derived from the nonenzymatic association of phosphorylase a into its tetrameric form as follows:

Reversal of Phosphorylase
It is of interest to note that the dimeric form of phosphorylase a is a better substrate than the tetrameric form in the phosphoprotein phosphatase reaction (25,26). It is also noteworthy that the dimeric form of phosphorylase a is much more active than the tetrameric form of the same protein with respect to its glycogenolytic activity (23,(28)(29)(30). The authors recognize that properties attributed to dimeric phosphorylase a, in contrast to those of the tetrameric form, may in reality be due to a particular conformational form induced by an effector such as glucose. Such a form may coincidentally have less ability to polymerize to the tetrameric form.