Interaction of Inhibitors with Muscle Phosphofructokinase*

The interaction of several inhibitors with muscle phosphofructokinase has been studied by both equilibrium binding measurements and kinetic analysis. At low concentrations of citrate a maximum of 1 mol is bound per mol of enzyme protomer. Tight binding requires MgATP and very weak binding is observed in the absence of either magnesium ion or ATP. ITP at low concentrations cannot replace ATP. In the presence of MgATP and at pH 7.0, the dissociation constant for the enzyme. citrate complex is 20 PM. At 50 pM citrate and excess magnesium ion, the concentration of ATP required to give half-maximal binding of citrate is approximately 3 pM. Both P-enolpyruvate and 3-P-glycerate compete for the binding of citrate and the estimated Ki values are 480 and 52 PM, respectively. Creatine-P, another inhibitor of muscle phosphofructokinase, does not compete with the binding of citrate. Measurement of the equilibrium binding of ATP shows that citrate, 3-P-glycerate, P-enolpyruvate, and creatine-P all increase the affinity of enzyme for MgATP with the concentration required to give an effect increasing in the order given. In kinetic studies, citrate, 3-P-glycerate and P-enolpyruvate each act synergistically with ATP to inhibit the phosphofructokinase reaction. This is indicated by the observation that

The interaction of several inhibitors with muscle phosphofructokinase has been studied by both equilibrium binding measurements and kinetic analysis. At low concentrations of citrate a maximum of 1 mol is bound per mol of enzyme protomer. Tight binding requires MgATP and very weak binding is observed in the absence of either magnesium ion or ATP. ITP at low concentrations cannot replace ATP. In the presence of MgATP and at pH 7.0, the dissociation constant for the enzyme. citrate complex is 20 PM. At 50 pM citrate and excess magnesium ion, the concentration of ATP required to give half-maximal binding of citrate is approximately 3 pM. Both P-enolpyruvate and 3-P-glycerate compete for the binding of citrate and the estimated Ki values are 480 and 52 PM, respectively. Creatine-P, another inhibitor of muscle phosphofructokinase, does not compete with the binding of citrate. Measurement of the equilibrium binding of ATP shows that citrate, 3-P-glycerate, P-enolpyruvate, and creatine-P all increase the affinity of enzyme for MgATP with the concentration required to give an effect increasing in the order given. In kinetic studies, citrate, 3-P-glycerate and P-enolpyruvate each act synergistically with ATP to inhibit the phosphofructokinase reaction. This is indicated by the observation that the three metabolites do not inhibit the enzyme with ITP as the phosphoryl donor and that they inhibit at ATP concentrations that are not themselves inhibitory. Furthermore, the sensitivity to the inhibitors increases with increasing ATP concentrations.
Striking differences in the extent of inhibition can be seen by varying the order of addition of assay components. Preincubation of the enzyme with ATP and citrate, 3-P-glycerate, or P-enolpyruvate results in greater inhibition than when the inhibitor is added after the reaction is started with fructose-6-P. Furthermore, the inhibition is reversed partially 10 to 15 min after the addition of fructose-6-P. This phenomenon is particularly striking with creatine-P as the inhibitor. Very high concentrations of this inhibitor are required to show any effect if the inhibitor is added after fructose-6-P. These effects are interpreted as reflecting slow conformational changes between an active form with high affinity for fructose-&P and an inactive, or less active, conformation that binds the inhibitors. Citrate, 3-P-glycerate, P-enolpyruvate, and creatine-P increase the rate of the phosphofructokinase at subsaturating concentrations of MgITP. The results indicate a common binding site on the enzyme for citrate, 3-P-glycerate, and P-enolpyruvate that is distinct from the ATP inhibitory site. An additional site (or sites) for creatine-P is indicated. All four inhibitors act synergistically with ATP by increasing the affinity of the enzyme for MgATP at an inhibitory site. The inhibitors appear also to increase the affinity of the catalytic nucleoside triphosphate site for substrate.
The kinetic behavior of skeletal muscle phosphofructokinase is influenced by a large number of metabolites, suggesting the possibility of many highly specific binding sites on the enzyme surface (see Ref. 1,for review). The molecular weight of the protomer is in the range of 8 to 9 x 10' g/mol (2)(3)(4)(5) and each protomer of muscle phosphofructokinase possesses a site for fructose-6-P and another site capable of binding adenine nucleotides, with cyclic AMP' being most tightly bound (6). * This investigation was supported by United States Public Health Service Grant AM11410 and a Grant-in-Aid from the American Heart Association with funds contributed in part by the Wisconsin Heart Association.
These two sites have also been described for sheep heart phosphofructokinase by Setlow and Mansour (7). A second adenine nucleotide site binds MgATP with high affinity and is thought to be an inhibitory site (8), and an additional nucleoside triphosphate site presumably represents the catalytic site (6,8). Kinetic studies have suggested at least two sites for monovalent cations (9) and a divalent cation site has been described by Mathias (10). Other effecters of enzyme activity include inorganic phosphate, fructose-1,6-P,, citrate, 3-phosphoglycerate, 2,3-bisphosphoglycerate, P-enolpyruvate, and creatine-P. Of these only citrate has been studied from the point of view of defining binding sites. The equilibrium binding studies of Lorenson and Mansour (11) concluded that citrate competes with the binding of ATP. On the other hand, on the basis of indirect measurements, we have suggested that ATP increases the affinity of the enzyme for citrate indicating a unique site for citrate (6,12). In examining molewlar models of several effecters, structural similarities in terms of charge distribution can be seen among the inhibitors; citrate, 3-P-glycerate, P-enolpyruvate, and creatine-P. The present study examines by kinetic analysis and by equilibrium binding the role of these four inhibitors in the regulation of phosphofructokinase and provides evidence that citrate, 3-Pglycerate, and P-enolpyruvate bind at a commpn site distinct from the ATP binding sites, and that all four inhibitors act synergistically with ATP. and was allowed to enter the gel. The column was eluted with more of the same solution and fractions of 0.8 to 1.0 ml were collected.
Flow rates of 15 to 20 ml/hour were used and the columns were run at ambient room temperature (22-24"). One-tenth milliliter of each fraction was counted in a scintillation spectrometer.
All experiments were checked for complete recovery of protein in the void volume, for attainment of equilibrium as indicated by the return of the base-line concentration of the citrate to its initial value after the emergence of the protein, and for good agreement between the total areas of the peak and trough in the elution profile. To test the influence of other compounds on the binding of citrate, metabolites were included with the citrate in the column buffer and in the protein sample applied to the column. In several instances, particularly when conditions were varied, recovery of enzyme activity was also monitored extensively and in no case was loss of activity detected.
The binding data are expressed relative to a protomer molecular weight of 90,000. This value was ascertained on the basis of the binding unit for fructose-6-P and cyclic AMP (6) and of the minimal molecular weight from thiol reactivity studies (3).

AND DISCUSSION
Binding of ["C]Citrate by Muscle Phosphofructokinuse-Previous studies from this laboratory have suggested that the site of interaction of citrate on skeletal muscle phosphofructokinase was distinct from the ATP inhibitory site and that indeed the affinity for citrate was increased in the presence of MgATP. Such conclusions were based upon indirect observations on the conformation of the enzyme as monitored by thiol reactivity with dithionitrobenzoic acid (8) and with fluorodinitrobenzene (12). The converse had been demonstrated directly, that is, the affinity of the enzyme for ATP was increased in the presence of citrate (6); but direct binding studies of citrate by the skeletal muscle have not been performed. Citrate binding by sheep heart phosphofructokinase has been investigated by Mansour and co-workers (1,11) who concluded ATP and citrate bind to the same site on the enzyme although in those studies the binding of citrate was always measured in the presence of low concentratiqns of ATP. In the rabbit, and in several other mammals as well, we have concluded that the skeletal muscle enzyme and the major isozyme of heart were identical (17); and thus the finding that the citrate and ATP binding were competitive was surprising in view of the studies of muscle phosphofructokinase.
We thus decided to investigate directly the binding of citrate by skeletal muscle phosphofructokinase and to determine the role of ATP in citrat,e binding. The binding studies were carried out by the Hummel and Dreyer technique (16) as described under "Experimental Procedure." A glycylglycine/glycero-P buffer was used for these studies as it was in previous binding studies (6) for the reason that complete recovery of the protein from Sephadex columns could be obtained with this buffer. With other buffer systems in the absence of added effecters, protein aggregation occurred and subsequently precipitated on the column. This phenomenon was also observed by Lorenson and Mansour (11) in studies of equilibrium binding by heart phosphofructokinase in imidazole buffer. Glycylglycine/glycero-P buffer is not completely inert in that it does influence the kinetics as well as enhance the stability of the enzyme (18). However, the enzyme in this buffer does show all of the usual kinetic properties of phosphofructokinase as indicated in earlier studies, although kinetic parameters may be shifted from what is observed in iV-tris(hydroxymethyl)methyl-2-aminomethanesulfonic acid buffer, for instance. For binding studies, it was obviously an advantage to use a buffer in which the enzyme was stable in the presence of added effecters. Fig. 1 (lower curve) shows a double reciprocal plot of the binding of citrate by muscle phosphofructokinase in the presence of 5 mM MgCl, and 20 NM ATP. A dissociation constant of approximately 20 PM is calculated from these data, and the binding of 1 mol per protomer is indicated. The presence of ATP is indeed necessary to show significant binding of citrate in agreement with previous studies on the effect of citrate on thiol reactivity (12). In the presence of magnesium ion the concentration of ATP required to enhance citrate binding is very low as indicated in Fig. 2 bound to the enzyme that was removed neither by the prior charcoal treatment nor by the gel permeation column itself, or it is an indication of the affinity of the enzyme for citrate in the absence of MgATP. Assuming that it is indeed a measure of citrate binding without the MgATP and that the binding follows a normal Henri binding isotherm, a Kd can be calculated from the relationship: . Employing a value of 0.09 mole bound (see Fig. 2), a Kd of 0.5 mM can be calculated. The effect of MgATP is thus to increase the affinity of enzyme for citrate by at least 25-fold. The data of Fig. 2 also show that at low concentrations of ATP, maximal binding of citrate requires the presence of magnesium ion. This indicates that the MgATP is more effective than free ATP in promoting citrate binding, similar to the relative abilities of the complex and the free nucleotide to promote conformational changes (12). ITP at low concentrations cannot replace ATP as indicated by Fig. 2. This is consistent with the fact that MgITP interacts very weakly with the inhibitory MgATP binding site (8) and less effectively than MgATP at the catalytic site (see results that follow). We have shown previously that at the low ATP concentrations employed here, conformational changes as measured by sulfhydryl reactivity are the result of interaction at the inhibitory site (8). The estimated half-maximal concen-tration of MgATP required to show a conformational change was 3 to 5 PM (12), very close to the value required for citrate binding. The interaction of MgATP that led to the conformational change in phosphofructokinase was shown to be at the inhibitory site by the criteria of pH dependence, specificity, and reversibility by activators of the enzyme (8). The enhanced citrate binding due to the presence of MgATP is not necessarily accompanied by these conformational changes, however, because the studies of conformation as monitored by thiol reactivity (8, 12) employed protein concentrations one-tenth to one-twentieth of those used in these binding studies. It will be shown below that at the high concentrations of protein used in the binding studies, dissociation of enzyme does not occur. Conformational changes not resulting in dissociation may occur, however. The interactions do occur, nonetheless, at similar effector concentrations for both the equilibrium binding and the conformational studies. The K, for MgATP at the inhibitory site is much lower than the K, at the catalytic site (about 50 PM) obtained from kinetic studies. It is likely that in the absence of fructose-6-P, which decreases ATP binding at the inhibitory site (8), the affinity for MgATP at the inhibitory site is greater than at the catalytic site, a situation which is obviously reversed in the presence of fructose-6-P.
Aggregation State of Enzyme in Presence of Citrate-It should be noted that no cooperativity was observed in the binding of citrate (Fig. 1). In our earlier studies (6), binding of fructose-&P and mono-and diphosphonucleosides also exhibited hyperbolic binding isotherms. Hill and Hammes (19) have noted that the binding of fructose-6-P and fructose 1,6-bisphosphate show little or no cooperativity when the enzyme is associated to aggregates larger than the tetramer but that negative cooperativity is observed with binding by phosphofructokinase dimers. At the concentrations of enzyme employed in Fig. 1 (3.1 to 9 mg/ml) it might be expected that higher aggregates would predominate. Lad et al. (20), however, have shown that high concentrations of citrate stabilize smaller aggregates, presumably dimers. It is possible that in the binding studies of Fig. 1 small aggregates are present or, to make interpretation more complex, the aggregation state varies with the varying protein and citrate concentrations that were used in the binding studies. For this reason, the enzyme was examined in the ultracentrifuge under conditions which approximate the conditions of the binding studies. In Fig. 1, protein concentrations varied from 3.1 to 9.0 mg/ml. To approximate the extremes of the binding study, the distribution of aggregates of phosphofructokinase at 3 and 9 mg/ml with and without low concentrations of citrate was studied and the results at a single time point of sedimentation are shown in Fig. 3 (Table I) suggesting a common binding site. In those experiments unlabeled 3-P-glycerate, P-enolpyruvate, or creatine-P were included at the indicated concentrations in the gel exclusion column buffer along with the radioactive citrate. In the presence of P-glycerate and P-enolpyruvate, citrate binding was reduced. That the inhibition was competitive in nature was shown only in the instance of 3-P-glycerate.
Citrate binding was measured at several concentrations in the presence of a fixed concentration of 3-P-glycerate and the resulting data are in Fig. 1. The data indicate typical competitive inhibition of citrate binding and a K, for 3-P-glycerate of 52 PM can be estimated from the increase in the slope of plot of the citrate binding data. This value is presumably equal to the K, for 3-P-glyceric acid. If one assumes that P-enolpyruvate was also binding competitively at the citrate site, the apparent K, can be calculated from the The lack of inhibition of citrate binding by creatine phosphate indicated that this effector may be binding at a different site on the enzyme. It was possible that the binding was very weak, however, and an additional competition experiment was carried out under conditions more favorable for competition by creatine phosphate. Labeled citrate was present at 1 FM and creatine phosphate at 10 mM. With no creatine phosphate present 0.058 mol of citrate/m01 of protomer was bound and this was reduced to 0.054 in the presence of creatine phosphate. A reduction of this modest magnitude, even if significant, would indicate a Kd of greater than 1 M if indeed creatine phosphate binds at the citrate binding site. Such a value is not consistent with kinetic data and it must be concluded that creatine phosphate does not interact at the citrate site. As will be shown in the following section, binding must certainly occur under the foregoing conditions, but it must occur independently of the binding of citrate.
Effect of Inhibitors on ATP Binding-It is possible that creatine phosphate binds directly at the ATP inhibitory site or that it enhances ATP binding.
Previous studies from this laboratory have indicated that citrate enhanced the binding of ATP by phosphofructokinase (6,8). If, as suggested by the data of Table I, P-enolpyruvate and 3-phosphoglyceric acid bind at the citrate site, these metabolites should also increase ATP binding. The four inhibitors, citrate, P-enolpyruvate, 3-P-glycerate, and creatine phosphate, were examined for their ability to increase the binding of ATP by the enzyme. The results are described in Table II. In this experiment a low level of ATP was used (1 PM) and increases in affinity could be detected in the presence of all four inhibitors.
The species of ATP that is binding in this instance is MgATP because MgZ+ was present at 5 mM. A possible complication in such an experiment is that an ATPase activity has been described for muscle phosphofructokinase (23). Such an activity could lead to destruction of ATP during the binding run and invalidate the data. In the previous study of the ATPase activity (23), an activity of 0.075 unit/mg was observed at pH 8.1 and 1 mM ATP. The phosphofructokinase used in the present study also showed ATPase activity, but the specific activity was slightly less than 0.001 unit/mg at pH 8.0 and 1 mM ATP as assayed by measuring ADP production coupled to the oxidation of NADH through pyruvate kinase and lactic dehydrogenase.
That this trace of apparent ATPase activity did not contribute to the destruction of ATP during the course of an equilibrium binding study was shown by the fact that more than 90% of the radioactivity associated with the pooled fractions of the trough of a binding column profile could be retained on a column (20 x 4 mm) of charcoal. Because the radioactivity was located in the y position of the ATP, ATPase activity would have liberated radioactive inorganic phosphate which will not be bound by charcoal. If these fractions from the equilibrium binding column were first heated in 1 N HCl for 5 min at 90" and then neutralized, a procedure that removes the fi-and y-phosphates of ATP, the radioactivity was not retained on a charcoal column. Hence, the charcoal column did provide an indication of ATPase activity and no such activity was observed during the binding experiments.
The results of Table II show that citrate, 3-P-glycerate, and P-enolpyruvate all increase the amount of MgATP bound by phosphofructokinase and the order of effectiveness is in the same order as the dissociation constants deduced from Fig. 1 and Table I; that is, citrate promoted MgATP binding at lower concentrations than 3-P-glycerate which in turn required lower concentrations than P-enolpyruvate. Of particular interest is the observation that creatine phosphate also increased the amount of MgATP bound. This indicates that creatine phosphate does not bind at the ATP site but that its interaction enhances the affinity of the enzyme for ATP. Creatine phosphate thus binds at a unique site on the enzyme that is independent of the citrate site and is interacting with one or more ATP sites. Both creatine phosphate and citrate apparently inhibit the enzyme by increasing the affinity of the enzyme for MgATP at the inhibitory site, but they do so as the result of interactions at sites differing from one another and from the MgATP site. As will be indicated below, the inhibitors appear to increase the affinity of the enzyme for nucleoside triphosphate not only at the inhibitory site but at the catalytic site as well.
Kinetic Studies: General Considerations-Without employing sophisticated techniques such as stop-flow, it is not possible to study the kinetics of phosphofructokinase at the enzyme concentrations employed in the earlier studies of protein conformation (8,12) or in the studies reported here of binding and of sedimentation studies. For this reason it is not possible to make direct comparisons of quantitative kinetic data. Certain qualitative questions may be considered, however. Are the actions of citrate, P-glycerate, P-enolpyruvate, and creatine-P synergistic with ATP? Is the ranking of effectiveness, in terms of concentration, identical with that seen in the binding studies? In addition, experiments are presented to show that these same inhibitors, at identical concentrations used in inhibition studies, may act as activators at low concentrations of nucleoside triphosphate, suggesting enhanced affinity at the catalytic site for phosphoryl donor.
Effect of Order of Addition of Reaction Components-Kinetic studies of phosphofructokinase usually have been complicated by lags in the initial phases of the reaction. Part, but not all of this lag can be attributed to the approach to steady state of the auxiliary enzymes. In practice one usually measures rates achieved 2 or 3 min after starting the reaction at which time the velocity is usually constant. We have observed under certain conditions, however, that the lag can be 10 to 15 min and the results observed 3 min after starting the reaction can be quite misleading.
This phonomenon is shown in Fig. 4A. The upper curve shows the potent inhibition of phosphofructokinase by 0.1 mM citrate when the enzyme was preincubated for 4 min with all components except fructose-6-P.
After the inhibition decreased, the final rate achieved was about 20 to 30% less than the assay performed without citrate (lower curue). A similar result was obtained if the citrate and fructose-6-P were added together. The middle curve shows the result of adding 0.1 mM citrate 4 min after the reaction had started. It was observed in this case that the inhibition occurred rapidly and that the extent of inhibition was similar to that observed after the reaction had started. In Fig. 4B is shown the extent of citrate inhibition when determined by the two different assay methods. The open circles describe the effect of citrate when the citrate was added before starting the reaction with fructose-6-P and the rate was determined 3 to 4 min after the reaction started. In this situation the response to increasing citrate concentrations occurred over a very narrow range and if one attempts to analyze the data by Hill-type plots, an extremely high n value is observed. The closed circles describe the'citrate effect when it was added 4 min after the reaction had started. The rate was determined subsequently 3 to 4 min after citrate addition. A Hill plot of these data gave an n value of approximately 4. The lag could also be produced by preincubation of the enzyme with 3-P-glycerate, P-enolpyruvate, or 2,3-bisphosphoglycerate, further supporting the idea that these inhibitors and citrate act in an identical manner. Preincubation with creatine-P also produced the lag, an effect which may be related to its influence on the binding of MgATP (Table II). Prolonged preincubation with MgATP did not produce the lag periods described in Fig. 4A.
There are several possible explanations for these phenomena based upon previous information on phosphofructokinase. It is possible that the products generated in the first several  FIG. 4. Influence of order of addition of reaction components on citrate inhibition of phosphofructokinase. A, in all cases, the enzyme (0.3 rg/ml) was incubated together for 4 min at pH 7.0 with 1 mM ATP and other reaction components (see "Experimental Procedure") except fructose-6-P. Decrease in A,,o is shown for the following conditions: Cl, citrate (0.1 mM) included in preincubation, reaction started by addition of fructose-6-P to a final concentration of 1 mM; 0, no citrate, reaction started with fructose-6-P, 0, reaction started with fructose-6-P, citrate added to a final concentration of 0.1 mM 4 min later. B, 0, citrate inhibition was determined from the rate obtained 3 to 4 min after the addition of fructose-6-P (1 mM) with citrate present during preincubation; 0, reaction started with fructose-6-P, citrate added 4 min later, and the extent of inhibition determined from rate 3 to 4 min after citrate addition. minutes without citrate, or in 10 to 15 min with citrate, relieve or prevent the citrate inhibition.
This does not appear to be the case because the products generated are ADP, NAD, and glycerol-3-P, and the addition of these products at concentrations that could be generated during the lag to the preincubation with citrate did not prevent the long lag period. Fructose 1,8bisphosphate, a known activator of phosphofructokinase, would also accumulate to a steady state level during the assay. El-Badry et ul. (24) have described the inhibition of phosphofructokinase by aldolase and fructose-1,6-bisphosphatase that appeared to be due to depletion of the sugar phosphate. To test for an effect on the length of lag phase in the presence of citrate, fructose 1,6-bisphosphate was added after the addition of substrate to the assay. Following a short burst in NADH oxidation due to the presence of fructose 1,6-bisphosphate the rate decreased to the rate observed in the absence of fructose 1,Bbisphosphate.
The lag was not eliminated by this treatment. It was further reasoned that if the lag were due to fructose 1,6bisphosphate accumulation, then increasing the concentration of auxiliary enzymes in the assay should prolong the lag by reducing the steady state level of fructose 1,6-bisphosphate. However, the lag phase was not prolonged by the presence of a 4-fold higher concentration of aldolase, triose phosphate isomerase, and the dehydrogenase.
Another possible explanation for this phenomenon is that it results from slow conformational changes, termed hysteresis by Frieden (25). Previous work from this laboratory has provided evidence for at least two conformational states for the enzyme (12). Citrate and ATP have higher affinity for the inhibited conformation (or conformations) and fructose-6-P and activators have higher affinity for the active conformation. Hulme and Tipton (26) have noted that the kinetics of heart phospho-fructokinase are influenced by protein concentration, an effect which they interpreted from the viewpoint of an associatingdissociating system. They further noted that citrate exacerbated the nonlinear response of the enzyme activity to protein concentration and that activators relieved the effect. The importance of the state of aggregation in phosphofructokinase activity was first suggested by Mansour (27) in studies of the heart enzyme. Earlier work with the rabbit liver enzyme from this laboratory (28) has noted variations in initial velocities depending upon the order of addition of reactants. The initial burst of reaction rate followed by a gradual decrease when reactions were started by the addition of enzyme was thought to reflect the high activity of the aggregated enzyme which decreased following dilution in the assay mixture. Ramaiah and Tejwani (29) have examined more extensively this phenomenon with liver phosphofructokinase and have concluded that the enzyme occurs in interconvertible forms. Lad et al. (20) have shown that dilution of muscle phosphofructokinase to a concentration of 0.15 mg/ml at pH 7.0 and 5O results in a depolymerization of tetramer to dimers with reaction half-time of 1.5 hours, and that the depolymerization is associated with a loss of activity as measured at pH 8.0. These workers (20) have demonstrated further that the presence of high concentrations of citrate (5 mM) greatly enhance the rate of dissociation of muscle phosphofructokinase to inactive dimers. The other inhibitor, ATP, does not produce this effect but Bock and Frieden (30) have indicated that ATP enhances the rate of pH-dependent dissociation. Presumably, the presence of low concentrations of MgATP would have permitted Lad et al. (20) to observe the citrate effect at low concentrations in their studies, based upon the results of Fig. 2 as well as previous data from this laboratory (12). Frieden (31) has shown that the activity changes that accompany association-dissociation phenomena due to pH changes are relatively slow. The recovery of activity 10 to 15 min after the addition of fructose-6-P (Fig. 4) may be the result of a slow fructose-6-Pinduced association to a new equilibrium with the accompanying recovery of activity. Lad et al. (20) demonstrated that fructose-6-P opposes the action of citrate and that mixtures of citrate and fructose-6-P stabilized aggregates that are of intermediate size.
In the case of the upper curve in Fig. 4A where citrate and ATP were incubated with the enzyme prior to the addition of fructose-6-P, one might predict that the inactive form, which is probably a depolymerized form, was stabilized. Upon addition of fructose-6-P, the achievement of the new equilibrium of active and inactive conformers was slow.
A polymerization reaction should be concentration-dependent and the effect of varying protein concentration on the length of the lag phase was examined. Assays were performed at pH 7.0 in the presence of 1 mM ATP, 1 mM fructose-6-P, and 0.1 mM citrate with phosphofructokinase concentrations in the assay varying in the range of 0.06 to 0.6 pg/ml. Under these assay conditions, lags of less than 1 min were observed at 0.6 &ml and more than 30 min at 0.06 rg/ml. Intermediate concentrations gave intermediate lag times, but the length of the time was not directly proportional to the dilution. More extensive studies over a broader range of concentrations would be necessary to determine a quantitative relationship between the degree of dilution and the length of lag. If the lag is due to a dimerization, one might predict that lag should vary as the square of the degree of dilution. It is important to note the final rate achieved after the lag varied in direct proportion to the protein concentration as did the rate achieved if the citrate was added subsequent to starting the reaction with fructose-6-P. In all of the inhibition studies to be described, the assays were performed by adding the inhibitor 4 min after the reaction was started with fructose-6-P and subsequently measuring rates 3 to 4 min after inhibitor addition. The linearity of the rate with time and protein concentration obtained with these conditions indicated that the assay was being performed under steady state conditions. Phosphofructokinase was present in the subsequent assays at a concentration of 0.3 to 0.4 pg/ml. Synergism between ATP and Citrate-ITP is an efficient substrate for phosphofructokinase but it acts only very weakly as an inhibitor.
MgITP has been shown to bind poorly to the inhibitory ATP site (12). As shown in Fig. 2, low concentrations of ITP do not promote citrate binding by the enzyme. Because low concentrations of citrate interact with phosphofructokinase only in the presence of ATP, then citrate inhibition should not be seen when ITP is used as the phosphoryl donor in the reaction. This is demonstrated by the data of Fig. 5. In Fig. 5A, the saturation curves for ATP and ITP are shown, and in Fig.  5B is described the action of citrate at three concentrations of ATP. No inhibition by citrate was observed with ITP as substrate, whereas employing concentrations of ATP that were not inhibitory alone, i.e. 0.7 to 1.5 mM, potent inhibition was observed only in the presence of citrate. In other words, citrate had no effect alone under these conditions, but was very potent when present together with ATP and the potency increased at higher levels of ATP. It should be noted that the inhibition could be relieved by increasing the concentration of fructose-6-P or by the addition of activators such as AMP.
Synergism and 3-P-Glycerate, P-Enolpyruvate, and Creatine Phosphate-The effective concentrations of P-glycerate and P-enolpyruvate were in the order observed for their ability to enhance the binding of MgATP by phosphofructokinase; that is, the required concentrations of 3-P-glycerate were slightly higher than those used with citrate, whereas P-enolpyruvate was effective only at concentrations above 1 mM. These results are shown in Fig. 6. Increasing the level of ATP increased the effectiveness of the metabolites as inhibitors of enzyme activity. As in the case of citrate, the two inhibitors had no effect with ITP as the phosphoryl donor, and the inhibition could be relieved by the addition of AMP.
Creatine-P was a very poor inhibitor of phosphofructokinase when the assay was carried out in the prescribed manner; that is. when the inhibitor is added 4 min after the addition of the substrate, fructose-6-P. For example, at 1 mM ATP and the other conditions of Fig. 6, creatine phosphate at 40 mM inhibited the enzyme less than 50%. It has been shown by a number of workers, including ourselves, that creatine-P is an effective inhibitor at quite low concentrations (1). In work from this laboratory (18), we reported that at pH 7.1 with ATP at 0.2 mM and fructose-6-P at 0.4 mM the concentration of creatine-P required to give 50% inhibition was only 1.9 mM. However, in those studies as well as those from other laboratories, the enzyme was assayed in the conventional way of starting reactions by the addition of substrate. Under the conditions of Fig. 6, creatine-P was strongly inhibitory (about 90% inhibition at 3 mM) if it was preincubated with the enzyme for several minutes and the reaction started with fructose-6-P. After 12 to 15 min however, the rate began to slowly increase and approach that of the uninhibited enzyme. In other words, this is the same phenomenon, probably one related to slow conformational changes, that was discussed previously with regard to citrate inhibition (Fig. 4). The phenomenon observed with creatine phosphate was much more pronounced in that the inhibition was practically eliminated if creatine phosphate was added several minutes after the reaction had started. It should be noted that the relative effect of changing the order of addition with 3-P-glycerate and P-enolpyruvate as inhibitors was similar to that seen with citrate, which is consistent with these two inhibitors binding to the same site that citrate occupies.
Inhibitors as Activators-In equilibrium binding studies reported earlier (6), it was noted that citrate increased the affinity of at least two and possibly three of the sites capable of binding ATP. Randle et al. (32) demonstrated in kinetic studies that citrate increased the affinity of the enzyme for ATP at the catalytic site. This is confirmed in the upper part of Fig. 7. At low concentrations of ATP, citrate acts as an activator of the reaction whereas the well documented inhibition is shown at higher ATP concentrations. Activation by citrate is even more apparent with ITP as the phosphoryl donor as shown in the lower half of Fig. 7. MgITP is bound less tightly by the enzyme and cooperativity in kinetic behavior is indicated by the sigmoid nature of the substrate versus velocity curve. Similar cooperativity can be observed with ATP as a substrate but the apparent K, for ATP is quite low, less than 20 PM at pH 7.0. With ITP as phosphoryl donor the apparent K, is close to 1 mM and, in the presence of 60 FM citrate the K, is decreased to approximately 0.4 mM. As would be expected from the presumption of a binding site that is identical with the citrate site, 3-P-glycerate, and P-enolpyruvate also increase the activity of phosphofructoki-! 5 lnhibkor ( mM) FIG. 6. Synergism between ATP and 3.P-glycerate or P-enolpyruvate. Assays at pH 7.0 and 1 mM (0) or 1.5 m&t ATP (0) with the indicated concentrations of inhibitor. The inhibitors were added 4 min .min after the reaction was started with fructose-6-P. after the addition of fructose-6-P. Fructose-6-l' was present at 1 mM.  nase in the presence of subsaturating substrate concentrations of ITP. This is described in Table III. The effective ranges of concentration are similar to those required to give inhibition at high substrate levels. Creatine-P, which also increased the binding of MgATP (Table II) but did not bind at the citrate site (Table I), increased the activity of ihe enzyme at low substrate levels.
Some General Comments-The citrate effects described herein are those requiring a high affinity site. The equilibrium binding studies indicated a Kd of 20 fiM and from Fig. 5 this concentration was inhibitory, although the extent of inhibition was dependent upon the concentration of ATP. At ATP concentrations that are inhibitory, for example, 2 mM ATP with other conditions of Fig. 5, citrate at 20 bc~ is an even more effective inhibitor.
Citrate effects at much higher concentrations may also be observed and may require a separate mode of action or an additional site. For example, under conditions where the enzyme is less sensitive to ATP inhibition, such as 9411 pH 8.0 or with high concentrations of fructose-6-P (18), citrate can inhibit by the complex formation of magnesium. The dissociation of muscle phosphofructokinase in the presence of citrate (20) has only been studied with relatively high concentrations of citrate and it would be of greater interest to study this phenomenon at citrate concentrations that are inhibitory (20 to 200 pM) in the presence of MgATP. In the equilibrium binding studies described here, citrate binding was observed under conditions where dissociation to units smaller than the tetramer was not occurring as observed in sedimentation runs in an analytical ultracentrifuge (Fig. 3). It is quite possible, however, that citrate binding results in a conformational change that may shift the equilibrium to favor dissociation. Dissociation may only occur at concentrations much lower than those of the binding study. On the other hand, although dissociation is obviously not required for citrate binding, it is possible that it may be necessary for enzyme inhibition.
The possible importance of dissociation to citrate inhibition is suggested by the time-dependent phenomena in Fig. 4 and the dependence upon protein concentration of citrate inhibition. Further support for the role of dissociation comes from the recent observation of Lad and Hammes (33) that phosphofructokinase tetramers cross-linked with a bifunctional reagent are resistant to citrate inhibition.