A Role for Pyridoxal Phosphate in the Control of Dephosphorylation Phosphorylase cz*

The control of the dephosphorylation of phosphoryl- ase a by glucose or caffeine, but not by AMP requires the coenzyme, pyridoxal phosphate. The dephospho- rylation of apophosphorylase a and reconstituted phos- phorylase a derivatives prepared from analogs with less than two ionizable hydrogens at the 5’ position cannot be activated by glucose or caffeine. No dissociation of these inactive forms could be detected in the presence of glucose or caffeine in the ultracentrifuge, indicating that the coenzyme has an important effect on the binding, or structural changes, or both, induced by glucose or caffeine. The dephosphorylation of active phosphorylase a derivatives, pyridoxal-phosphorylase a in the presence of inorganic phosphite, or enzyme forms generated with 5’ side chains containing phos- phonate groups can be activated by glucose or caffeine. of phosphorylase a reconstituted with a-methylpyridoxal phosphate can be stimulated by caffeine but not by glucose. The phosphonate groups of the analogs have higher

From the Department of Biochemistry and Biophysics, Iowa Stnte Unicersity, Ames, Iowa 50011 The control of the dephosphorylation of phosphorylase a by glucose or caffeine, but not by AMP requires the coenzyme, pyridoxal phosphate.
The dephosphorylation of apophosphorylase a and reconstituted phosphorylase a derivatives prepared from analogs with less than two ionizable hydrogens at the 5' position cannot be activated by glucose or caffeine. No dissociation of these inactive forms could be detected in the presence of glucose or caffeine in the ultracentrifuge, indicating that the coenzyme has an important effect on the binding, or structural changes, or both, induced by glucose or caffeine. The dephosphorylation of active phosphorylase a derivatives, pyridoxal-phosphorylase a in the presence of inorganic phosphite, or enzyme forms generated with 5' side chains containing phosphonate groups can be activated by glucose or caffeine. Dephosphorylation of phosphorylase a reconstituted with a-methylpyridoxal phosphate can be stimulated by caffeine but not by glucose. The phosphonate groups of the analogs have higher p& values than does the phosphoryl portion of pyridoxal phosphate; phosphorylase reconstituted with these derivatives requires a higher pH for maximum stimulation by glucose or caffeine than does the native phosphorylase a. Ultracentrifugal experiments on active phosphorylase a derivatives do not show an exact correlation between effects of glucose or caffeine on the dimer-tetramer forms of phosphorylase a and activation. The results suggest that the ionic state of the 5' position of pyridoxal phosphate is particularly important for the regulation of the dephosphorylation of phosphorylase a by glucose or caffeine.
All a-glucan phosphdrylases contain firmly bound pyridoxal phosphate and require this coenzyme for catalytic activity (l-3). The exact role of pyridoxal phosphate in catalysis is not known, but clearly it does not function as it does in other vitamin B6-containing enzymes (3,4). Amino acid sequence analysis of pyridoxal phosphate binding sites of phosphorylases from rabbit, potato, yeast, and Escherichia coli (5)(6)(7)(8) show that this region of the protein is highly conserved, a fact that is consistent with an involvement of the coenzyme in an important function. One approach taken to evaluate the role of the coenzyme is to study the properties of the apoenzyme and phosphorylases reconstituted with various pyridoxal phosphate analogs. Physical studies have shown that the structures of apophosphorylases b and a are different from their holoenzymes (g-11), and although major structural changes occur upon the binding of various pyridoxal phosphate analogs (3, lo), differences still exist in the conformation of these reconstituted phosphorylases from the native enzyme (11). Graves et al. (12,13) used enzymes to probe conformational differences in apophosphorylase, pyridoxal-reconstituted phosphorylase, and native phosphorylase and concluded that the coenzyme affects enzymatic reactions involving the NHZ-terminal domain, e.g. phosphorylation and dephosphorylation and limited proteolysis by trypsin. Results from measurements of catalytic activity of various reconstituted phosphorylases suggest that the phosphoryl group of the coenzyme is particularly important, and suggestions have been made that it might participate in a proton shuttle (11,14,15).
Recent studies using x-ray crystallography (16, 17) and pyridoxal-reconstituted phosphorylase (18) show that the coenzyme is in close proximity to the binding site for the substrate, glucose l-phosphate. X-ray crystallography has also shown that a nucleoside binding site is adjacent to the glucose or glucose l-phosphate site (19) and is some 30 8, away from the AMP binding site, which in turn is near the phosphorylatable serine (20). Because our earlier results (12, 13) suggested that the influence of effecters on dephosphorylation of phosphorylase a depended on the coenzyme, and because information is now available about effector binding sites in phosphorylase, we have made a comprehensive study of the action of certain effecters. Our results with phosphorylases reconstituted with different pyridoxal phosphate analogs suggest that the phosphoryl group of the coenzyme is important for effects of glucose and caffeine on phosphorylase a structure and for dephosphorylation.
Effects of AMP on dephosphoryl-  Fig. 1 because this concentration gave the maximum effect. Three different protein substrate concentrations (3, 1, 0.2 mg/ml) were also tried because the quaternary structure of phosphorylase a changes with enzyme concentration, i.e. dissociation of a tetramer to a dimer occurs upon dilution (36). The percentage change in rate of dephosphorylation caused by the effecters, however, was found independent of protein concentration. Fig. lb indicates that the rate of dephosphorylation of apophosphorylase a is not affected by caffeine, but is still inhibited 70% by 3 InM AMP. The same results are seen for the dephosphorylation of pyridoxal-reconstituted phosphorylase a as shown in Fig. lc. If 6 mM phosphite is included in the test solution, the rate for dephosphorylation of pyridoxal-reconstituted phosphorylase a is increased H-fold by 3 mM caffeine. It was suggested previously that phosphite binds noncovalently at the site normally occupied by the 5'-phosphate of pyridoxal phosphate in phosphorylase a (13, 18). Thus it seems that the coenzyme is necessary for the response to caffeine and that the phosphoryl portion of the coenzyme is particularly important. A shift of the dimer-tetramer equilibrium of phosphorylase a toward the dimer by caffeine has been reported by Bot et al. (37). With the conditions used for dephosphorylation, phosphorylase a is dissociated by 3 mM caffeine to a form of  the enzyme with a sedimentation constant characteristic of a dimer species (Table I). Apophosphorylase a is present seemingly as a mixture of monomer, dimer, and tetramer (10). Caffeine at 3 mM did not change the ultracentrifuge pattern of apophosphorylase a. No dissociation of apophosphorylase a could be detected even at 30 mM caffeine. Pyridoxal-phosphorylase a is not dissociated to any appreciable amount by 3 mM caffeine, but it is fully dissociated to the dimer by 30 mM caffeine (Table I). The dephosphorylation of pyridoxalphosphorylase a, however, is not activated by 30 mM caffeine. Pyridoxal-phosphorylase a in the presence of 6 mM phosphite was not dissociated by 3 mM caffeine, although as previously mentioned the phosphatase reaction is stimulated 1.8-fold under these conditions. The ultracentrifugal results indicate that caffeine can bind to pyridoxal-phosphorylase, yet the structural change necessary for enhancement of the phosphatase reaction does not occur. In the presence of phosphite, the pyridoxal-reconstituted enzyme shows a stimulation of the phosphatase reaction, although this stimulation cannot be correlated with changes in the dimer-tetramer equilibrium. We will provide further evidence to show that dissociation of phosphorylase a is not sufficient to explain activation.
Because our present results and an earlier study (13) suggested that the phosphoryl portion of the coenzyme is important for response to caffeine and glucose, we examined carefully the effects of modification of the 5' position of the coenzyme on phosphorylase a structure and enzymatic dephosphorylation.
Seven pyridoxal phosphate analogs were used to reconstitute apophosphorylase a. The rates of dephosphorylation of these enzymes under various conditions are listed in Table II. The highest rate of dephosphorylation in the absence of effecters is seen with native phosphorylase a. Although kinetic parameters have not been established (K,,, and V,,,,, values) for the different reconstituted phosphorylases, it is clear that all phosphorylases reconstituted with analogs are poorer substrates. Also shown in Table II are activities of the various phosphorylases.
These had been reported earlier with phosphorylase b (3,10,38). None of the enzyme forms without activity could be stimulated by glucose or caffeine in the enzymatic dephosphorylation test. All active forms except phosphorylase a reconstituted with a-methylpyridoxal phosphate were stimulated by caffeine and glucose. With the a-methylpyridoxal phosphate-reconstituted enzyme, no stimulation of the dephosphorylation was seen by glucose at the concentration used, although stimulation by caffeine was similar to that of the native enzyme.
From '"P-NMR studies of phosphorylase (39) and the effects of phosphate analogs on the activity of pyridoxal-reconstituted phosphorylase (18), it was suggested that the state of ionization of the phosphoryl portion of the pyridoxal phosphate is important for catalysis. All the pyridoxal phosphate analogs used here, that gave activity upon reconstitution with apophosphorylase a, have higher pK,, values for the phosphoryl portion than does pyridoxal phosphate. ThepK,,, values are 6.20 (40), 6.76 (30), 7.0 to 7.35 (38),' and 7.i' for pyridoxal phosphate, phosphonomethyl, phosphonoethyl, and phosphonoethenyl analogs of pyridoxal phosphate, respectively, Vidgoff et al. (38) showed earlier that the pH optimum was shifted to a higher pH for phosphorylase reconstituted with the phosplionoethyl analog of pyridoxal phosphate. The effect of pH on the activation of enzymatic dephosphorylation by caffeine was studied to determine whether the higher pK,, values of coenzyme analogs also caused a shift in the pH curve. As shown in Fig. 2

Pyridoxal
Phosphate and Dephosphorylation of Phosphorylase a of activation to the pyridoxal phosphate reconstituted phosphorylase a from pH 6.0 to 6.4 and gradually loses the effect at higher pH. For phosphorylase a reconstituted with phosphonoethyl or phosphonoethenyl analogs of pyridoxal phosphate, the maximum stimulation by caffeine is around pH 7.6 and pH 8.2, respectively. Thus the results show generally that a shift in the pH optimum to the alkaline side occurs with increasing pK,, values.
Sedimentation velocity measurements were done on all the phosphorylases reconstituted with pyridoxal phosphate analogs reported in Table II to determine whether activation was correlated with the dissociation of phosphorylase a. All of the analog-reconstituted phosphorylases that did not show any response to caffeine or glucose in the enzymatic dephosphorylation were also not dissociated by these effecters. These forms include phosphorylase a reconstituted with pyridoxal phosphate methyl ester, deoxypyridoxal, or the carboxyethenyl analog of pyridoxal phosphate. All of these proteins have s~",~( values of about 13 S with or without 3 mM caffeine or 10 mM glucose. Ultracentrifugation also was done in high salt solutions since it was shown previously that dissociation of 6.0 6.4 6.8 7. The dephosphorylation of ""1'-phosphorylase a reconstituted with ( ) pyridoxal phosphate, (0) phosphonoethyl, (0) phosphonoethenyl, (A) phosphonomethyl analogs of pyridoxal phosphate at different pH values in the presence of 3 mM caffeine is shown. Procedure is same as that described in the legend for Fig. 1, except that 100 mM imidazole, 100 mM Tris, 30 mM mercaptoethanol buffer is used. The activation by 2 mM caffeine at different pH values is plotted against pH.
native phosphorylase a occurs in the presence of NaCl (41). As shown in Fig. 3, the phosphorylase a reconstituted with the carboxyethenyl analog was not dissociated by 2.5 M NaCl as compared with the native phosphorylase a. The former sedimented as a single component with an ~20,~~~ = 13 S whereas the native enzyme showed a broad peak characteristic of the equilibrium between dimeric and tetrameric forms. The sedimentation pattern of phosphorylase a reconstituted with pyridoxal phosphate methyl ester in 2.5 M NaCl showed some asymmetry indicating some dissociation but far less than that of the native enzyme (data not shown). These results suggest that the 5' side chain group of the coenzyme has an important influence on phosphorylase a structure.
The sedimentation values for reconstituted phosphorylases whose dephosphorylations were stimulated by caffeine or glucose are listed in Table III. Caffeine at 3 mM dissociates phosphorylase a reconstituted with the phosphonoethyl analog of pyridoxal phosphate to the same extent as that of native phosphorylase a (Table III, Fig. 4a). Phosphorylase a reconstituted with the phosphonoethenyl analog is also dissociated but less effectively (Fig. 4~). Caffeine at 3 mM, however, only dissociates slightly the a-methylpyridoxal phosphate-reconstituted phosphorylase a as revealed by a slight broadening of the peak and a change in sedimentation constant, spO,,,, from 13.0 to 12.3 S (results not shown). For phosphorylase a reconstituted with the phosphonomethyl analog, caffeine at 3 mM does not show any shift from tetramer to dimer (Table III). Glucose at 10 mM, on the other hand, cannot dissociate any of the reconstituted phosphorylases in Table III to the extent observed for phosphorylase a reconstituted with pyridoxal phosphate.
The sedimentation pattern of phosphorylase a reconstituted with the phosphonoethyl analog is broadened (4b) but little or no effect is seen with phosphorylase a reconstituted with the phosphonoethenyl analog (4d). Also glucose has little or no effect on the ultracentrifugal characteristics of phosphorylases reconstituted with a-methylpyridoxal phosphate or the phosphonomethyl analog. These results and those of Table II suggest that the extent of activation of the phosphatase reaction and the dissociation are not exactly correlated.
One of the possible reasons that glucose at 10 mM did not dissociate reconstituted phosphorylases with the phosphonoethenyl or phosphonoethyl analogs is that these phosphorylases do not have as high an affinity for glucose as does the native phosphorylase a. As shown in Table IV   Top, pyridoxal phosphate-reconstituted phosphorylase a with 2.5 M NaCl; bottom, carboxyethenyl analogreconstituted phosphorylase n with 2.5 M NaCl. Protein concentrations were 3 mg/ml in 50 mM Tris, 30 mM mercaptoethanol, pH 7.5, at 16°C. The picture was taken 48 min after attaining full speed of 5",000 rpm. activators, glucose (43)(44)(45) and caffeine (35), and the inhibitor, AMP (35), affect dephosphorylation by binding to the substrate, phosphorylase a. Our earlier results suggested that the coenzyme, pyridoxal phosphate was important for the effect of glucose (13). This study shows that the coenzyme is also important for activation by caffeine, but it is not essential for AMP inhibition.
Because AMP binds extremely well to apophosphorylase (46), it is not surprising that it affects enzymatic dephosphorylation. AMP binds some 30 A from the coenzyme site near the phosphorylatable serine while glucose and caffeine bind to two distinct sites in close proximity to the pyridoxal phosphate site (47). Conceivably the coenzyme could be important for the structural integrity of the two sites near it, and the lack of effect of glucose and caffeine on enzymatic dephosphorylation and on the ultracentrifugal characteristics of apophosphorylase a could be due to poor binding of these effectors.  We have found that the phosphoryl group of the coenzyme is particularly important for the effects of glucose and caffeine. This was first indicated by the fact that pyridoxal, itself, was not sufficient to give an enzyme form whose dephosphorylation was activated by glucose (13) or by caffeine. However, both glucose and caffeine can affect pyridoxal phosphorylase in the presence of inorganic phosphite, a phosphate analog believed to bind at the site where the phosphate group of the natural coenzyme resides (18). Second, only phosphorylases reconstituted with analogs containing two ionizable hydrogens on the 5' side chain showed activation of the phosphatase reaction by glucose or caffeine. In addition, phosphorylase reconstituted with the carboxyethenyl analog of pyridoxal phosphate or pyridoxal phosphate methyl ester showed an altered response to NaCl, an agent that readily promotes dissociation of the tetramer of phosphorylase a to the dimer (41), indicating that important structural differences exist between the native enzyme and these phosphorylases.
The charge state of the phosphoryl function is likely important for response to glucose or caffeine. Analogs containing phosphonate functions possess higher pK,, values than does the phosphate group of pyridoxal phosphate and require a higher pH value for maximum response to effecters. These results suggest that a dianionic form of the coenzyme is important for control of dephosphorylation, as was suggested earlier for phosphorylase activity (39). higher than those of the native enzyme especially that of the phosphorylase a reconstituted with the phosphonoethenyl analog which is 5-fold higher. K, values for caffeine are elevated 2-fold for these reconstituted phosphorylases. These results suggest that the 5' side chain function of the coenzyme affects the structural integrity of the binding sites for glucose and caffeine and could explain the differences seen in the ultracentrifuge for pyridoxal phosphate-reconstituted and analog-reconstituted phosphorylases in the presence of glucose.

DISCUSSION
A relationship seems to exist between the activity of glycogen phosphorylase and the response to glucose or caffeine. All analogs that are inactive show no response to glucose or caffeine. There is no information available that shows whether glucose or caffeine can bind to these inactive enzyme forms. All active derivatives show a response to 10 mM glucose or 3 InM caffeine except for phosphorylase a reconstituted with a-methylpyridoxal phosphate. Differences exist in response to these effecters (Table IV), and a study is presently being done to determine the exact relationship between binding and activation. Our results suggest that the involvement of pyridoxal phosphate in catalysis is related to its role in the enzymatic dephosphorylation of phosphorylase a in response to glucose and caffeine. These two events could occur through a common mechanism, perhaps through a proton shuttle, a conformational event, or another process. Some phosphorylases do not undergo covalent modification, and the coenzyme is evidently necessary just for activity (48,49). Animal phosphorylases, however, require pyridoxal phosphate for activity as well as for the control of enzymatic covalent modification as shown in this study. An important control mechanism in glycogen degradation Caffeine or glucose is known to cause a dissociation of is the enzymatic dephosphorylation of phosphorylase a. The muscle phosphorylase a from a tetramer to a dimer (37, 50). Pyridoxal Phosphate and Dephosphorylation of Phosphorylase a The relationship between dissociation and activation of the phosphatase reaction has been studied by several investigators (37,43,45,51,52). Bot and his associates (37,51,52)  phorylation of this phosphonate analog under this same condition, pH 7.5, is activated 6-fold, but native phosphorylase is 13. activated only 2-fold. Our results suggest that other factors besides dissociation of phosphorylase a are necessary to explain activation of the phosphatase reaction. Madsen et al. (47) have shown that binding of AMP to phosphorylase a 14.
induces a conformational change in the NHZ-terminal portion of the enzyme which moves inward toward a fold in the 15. enzyme, making the serine-14-phosphate unavailable to phosphatase. They also suggested that glucose or caffeine activa- 16. tion of enzymatic dephosphorylation occurs through a conformational change that makes the seryl phosphate more acces-17.
sible to the phosphatase.

18.
The extensive study of Hers and his associates has demonstrated that glycogen metabolism in uiuo is altered by glucose 19. through its interaction with phosphorylase a (53)(54)(55)(56). This binding promotes dephosphorylation of phosphorylase a, 20 which is then followed by the dephosphorylation of glycogen synthase D and glycogen synthesis. Kasvinsky et al. (19) 21 showed that synergistic regulation of phosphorylase a activity 22 occurs with glucose and caffeine and suggested that a com-23 bined action involving these two sites might be important in 24. the physiological regulation of phosphatase activity. Miller et al. (57) showed that control of glycogen synthase and phos- of dephosphorylation of phosphorylase a by glucose or caf-28. feine is dependent upon pyridoxal phosphate. An alteration in 29. the response to glucose or caffeine under some physiological 30. conditions, might be due to a change in the coenzyme state of 31.

phosphorylase.
Possibly some apoenzyme is present under 32. different physiological conditions although early studies did 33 not detect apophosphorylase in a vitamin B,; deficiency state (58). Our studies show that enzymatic dephosphorylation in 34 response to glucose or caffeine is sensitive to pH. Perhaps the effect of pH might be due to changes in the ionic state of the 35 coenzyme in phosphorylase.
However, '"P-NMR spectroscopy 36 did not show any appreciable change in the chemical shift of 37 phosphorylase b as a function of pH, but not much information is available about phosphorylase a (39