Interaction of the Clathrin-coated Vesicle V-ATPase with ADP and Sodium Azide*

The kinetics of adenosine triphosphate (ATP)-dependent proton transport into clathrin-coated vesicles from bovine brain have been studied. We observe that the vacuolar proton-translocating ATPase (V-ATPase) from clathrin-coated vesicles is subject to two different types of inhibition by ADP. The first is competitive inhibition with respect to ATP, with aK i for ADP of 11 μm. The second type of inhibition occurs after preincubation of the V-ATPase in the presence of ADP and Mg2+, which results in inhibition of the initial rate of proton transport followed by reactivation over the course of several minutes. The second effect is observed at ADP concentrations as low as 0.1–0.2 μm, indicating that a high affinity inhibitory complex is formed between ADP and the V-ATPase and is only slowly dissociated after the addition of ATP. We have further investigated the effect of sodium azide, an inhibitor of the F-ATPases that has been shown to stabilize an inactive complex between ADP and the F1-F0-ATP synthase (F-ATPase). We observed that azide inhibited ATP-dependent proton transport by the purified, reconstituted V-ATPase with aK 0.5 of 0.2–0.4 mm but had no effect on ATP hydrolysis. Azide was shown not to increase the passive proton permeability of reconstituted vesicles and did not stimulate ATP hydrolysis by the reconstituted enzyme, in contrast with CCCP, which both abolished the proton gradient and stimulated hydrolysis. Thus, azide does not appear to act as a simple uncoupler of proton transport and ATP hydrolysis. Rather, azide may have some more direct effect on V-ATPase activity. Possible mechanisms by which azide could exert this effect on the V-ATPase and the contrasting effects of azide on the F- and V-ATPases are discussed.

tochondria, chloroplasts, and bacteria (9 -15), both in overall structure and in sequence homology between individual subunits. The A and B subunits of the V-ATPases and the ␣ and ␤ subunits of F 1 are all derived from a common ancestral nucleotide-binding protein (16,17), whereas the c subunit of the V-ATPases appears to have arisen by a gene duplication and fusion of the gene encoding the corresponding c subunit of F 0 (18). Also, V-and F-type ATPases may share some similarities in their catalytic cycle. For example, cooperative interaction between the nucleotide binding sites has been demonstrated for the V-ATPase (19 -22), as well as for the F-ATPase (9).
On the other hand, there are many differences between the F-and V-type ATPases, including their normal physiological functions. The primary function of F-ATPases (more properly termed ATP synthases) is ATP synthesis using the energy of the proton electrochemical gradient generated by electron transport during respiration (9 -15). In contrast, the V-AT-Pases function principally as ATP-dependent proton pumps to acidify intracellular compartments or in proton transport across the plasma membrane (1)(2)(3)(4)(5)(6)(7)(8). In addition, in vitro, the soluble F 1 domain of the F-ATPase is able to hydrolyze ATP, whereas isolated V 1 is inactive in ATP hydrolysis (1). Similarly, the isolated V 0 domain, unlike the corresponding F 0 domain, is not an open proton channel (23).
The difference between these two types of proton transport ATPase is also demonstrated by their inhibitor sensitivity: V-ATPases are specifically and potently inhibited by bafilomycin and the related compound concanamycin (24,25), whereas the F-ATPases are inhibited by oligomycin (15). Sodium azide at concentrations below 1 mM has also been shown to inhibit the F-ATPases, but only in the direction of ATP hydrolysis, not in the direction of ATP synthesis (26,27). It is widely accepted that the V-ATPases are insensitive to azide, but this conclusion is based upon studies of ATP hydrolysis (28 -30). The effect of azide on proton transport by the V-ATPases is controversial (31)(32)(33)(34).
Although there has been extensive kinetic characterization of the F-ATPases, there is relatively little information concerning the kinetics and regulation of the V-ATPases. There is considerable evidence that cells maintain different intracellular compartments at different pH values (1)(2)(3)(4)(5)(6)(7)(8), but the mechanisms employed to accomplish this remain uncertain. A change in coupling efficiency between proton transport and ATP hydrolysis has been proposed as one potential regulatory mechanism (1,4,8), but the intracellular signals that might control coupling remain uncertain.
We report here the modulation of V-ATPase activity by ADP and azide. ADP is shown to form an inactive complex with the V-ATPase in the presence of Mg 2ϩ , which is only slowly reactivated upon addition of ATP. This effect is independent of its action as a competitive inhibitor. Unlike the case of the F-ATPases, azide does not stabilize the ADP-inactivated state of the V-ATPase. Instead, azide is shown to inhibit proton transport without affecting ATP hydrolysis. This effect does not appear to be due to azide acting as a proton ionophore. Rather, azide may be having a direct effect on the V-ATPase. The possible significance of these effects of ADP and azide on V-ATPase activity are discussed below.
Purification and Reconstitution of the Clathrin-coated Vesicle V-ATPase-Clathrin-coated vesicles were prepared from calf brains and stripped of their clathrin coat, and the V-ATPase was solubilized with C 12 E 9 and isolated by glycerol density gradient sedimentation as described previously (35). The protein concentration was measured following precipitation with trichloroacetic acid as described (36). The purified V-ATPase was in some cases reconstituted into phospholipid vesicles as described previously (23,35).
ATPase Activity Assay-ATPase activity was measured at 23°C by a continuous spectrophotometric assay as described (37) and by a modified 32 P i release method (22,38). In some cases, the potassium salt of phosphoenolpyruvate and lyophilized pyruvate kinase were employed. In this case, lactate dehydrogenase that had been dialyzed against solubilization buffer containing 50 mM NaCl, 30 mM KCl, 20 mM HEPES (pH 7.0), 0.2 mM EGTA, and 10% glycerol to remove ammonium sulfate was employed.
ATP-dependent Proton Transport Assay-Proton transport was measured by ATP-dependent fluorescence quenching of ACMA (see figure legends) at 23°C using a Perkin-Elmer LS-5 spectrofluorometer as described previously (39). Assays were carried out in solubilization buffer (see above) containing 2 M dye and 0.5 M valinomycin in a total volume of 0.5 ml at room temperature. After equilibration of the vesicles with the dye for 3-5 min, MgATP was added, and the fluorescence intensity was monitored. After 5-8 min, the uncoupler CCCP was added to a final concentration of 2 M, producing an immediate reversal of fluorescence quenching. The addition of CCCP before MgATP completely prevented the initiation of fluorescence quenching by MgATP. Excitation and emission wavelengths were as follows: ACMA, 490 and 530 nm, acridine orange, 410 and 485 nm, and quinacrine, 420 and 510 nm (40). The degree of fluorescence quenching was linearly related to the amount of the vesicle protein added to the assay up to 60% quenching.
Measurement of Passive Proton Flux in Reconstituted Vesicles Containing K ϩ or Na ϩ -Phospholipid vesicles containing the reconstituted V-ATPase were used to measure passive proton flux as described previously (23). For these experiments, reconstituted vesicles were prepared in the presence of 150 mM KCl followed by replacement of the external KCl with 150 mM NaCl by dialysis. The vesicles were then diluted into buffer containing 150 mM NaCl in the presence of 20 nM valinomycin, thereby generating an interior negative membrane potential. Proton uptake under these conditions was monitored by fluorescence quenching using ACMA.
In order to measure passive proton flux in reconstituted vesicles containing an interior positive membrane potential, reconstituted vesicles were prepared in the presence of 150 mM NaCl and 1 mM of the membrane-impermeant, pH-sensitive fluorescence dye 8-hydroxypyrene-1,3,6-trisulfonic acid (HPTS) (41). Following removal of the external dye and replacement of the external NaCl with KCl using a Sephadex G-50 spin column, the vesicles were diluted into buffer containing 150 mM KCl, and valinomycin was added. The fluorescence intensity of the sample was measured by excitation at 460 nm and observation of the emission at 520 nm. In control samples, vesicles were diluted into buffer containing 150 mM NaCl.

RESULTS
Effect of ADP and Mg 2ϩ on ATP-dependent Proton Transport- Fig. 1 shows ATP-dependent proton transport into stripped clathrin-coated vesicles monitored by fluorescence quenching using the fluorescence dye ACMA. Proton transport was entirely dependent upon the addition of MgATP and completely abolished by the specific V-ATPase inhibitor concanamycin A or the uncoupler CCCP, added prior to ATP. As can be seen, no linear phase in the fluorescence change was observed, with the maximal change observed initially and the fluorescence eventually approaching a plateau. Two parameters were used to characterize ATP-dependent proton transport. The initial rate of proton transport was derived from the slope of the fluorescence curve generated during the first 30 s of the assay after the addition of ATP. After a period of 5-6 min, a time at which the maximal acidification was usually reached, the vesicles were uncoupled by the addition of 1 M CCCP to monitor the total acidification achieved.
Preincubation of the V-ATPase with Mg 2ϩ and ADP resulted in significant inhibition of the initial rate of acidification followed by a gradual reactivation of the enzyme over the course of several minutes following addition of ATP. The total acidification observed also decreased on preincubation with Mg 2ϩ and ADP relative to preincubation with Mg 2ϩ alone, although at the low concentration of ADP employed (2 M), this effect was small (see below). Preincubation with Mg 2ϩ alone did not decrease proton transport, whereas preincubation with ADP alone caused a partial decrease in both the initial rate and total acidification. However, no slow activation of proton transport was observed unless Mg 2ϩ was included with ADP in the preincubation mixture. Addition of ADP together with ATP and Mg 2ϩ resulted in the same level of inhibition observed with preincubation with ADP alone. This inhibition was characterized by determination of the ATP dependence of proton transport in the absence or presence of ADP (Fig. 2). ADP added . When Mg 2ϩ was present during the preincubation, the Mg 2ϩ concentration added with ATP was adjusted so that in all cases the final Mg 2ϩ and ATP concentrations in the assay were 1.5 and 0.5 mM, respectively. After 5-8 min, 1 M CCCP was added (at the second arrow). The proton transport activity was characterized either by the initial rate, derived from the slope of the tangent to the curve at 30 s following addition of ATP, or by the total acidification, as determined from the change in fluorescence following addition of CCCP.
directly to the assay mixture behaved as a competitive inhibitor with a K i of 11 M (the K m for ATP was 125 M).
These data demonstrate that ADP acts both as a competitive inhibitor of activity and, when preincubated with the V-ATPase in the presence of Mg 2ϩ , leads to an inactive complex that is slowly activated in the presence of ATP. To further characterize this inactive complex, the enzyme was preincubated with Mg 2ϩ in the presence of varying concentrations of ADP. As can be seen in Fig. 3, significant formation of the inactive complex was observed at ADP concentrations as low as 0.14 M. At this concentration, the fall-off in the rate of fluorescence change (observed in the control) is balanced by the activation associated with the addition of ATP. Thus, the ADP concentration required to stabilize the inactive state of the enzyme is much lower than the K i for ADP as a competitive inhibitor.
Preincubation of the V-ATPase with Mg 2ϩ and low concentrations of ADP also resulted in significant inhibition of ATP hydrolysis followed by a slow reactivation, similar to inhibition of ATP-dependent proton transport (Fig. 4).
Sodium Azide as an Inhibitor of ATP-dependent Proton Transport-It has been shown that azide inhibits the F-ATPase by stabilization of an ADP-inactivated form of the enzyme (42,43). We thus wished to determine whether azide would also stabilize the ADP-inhibited form of the V-ATPase. Fig. 5 shows that ATP-dependent proton transport monitored by quenching of ACMA was decreased in the presence of sodium azide. Inhibition by azide was also observed using other fluorescence probes (acridine orange or quinacrine) to monitor acidification (data not shown). Azide inhibited ATP-dependent proton transport in both stripped clathrin-coated vesicles and reconstituted vesicles containing the purified V-ATPase (Fig.  6). As can be seen from Fig. 5, although azide did not appear to inhibit the initial rate of proton transport, inhibition was observed within the first minutes after addition of ATP. The concentration of azide required to give 50% inhibition of total acidification was 220 M for stripped clathrin-coated vesicles or 400 M for the purified, reconstituted V-ATPase. Preincubation with ADP and Mg 2ϩ resulted in a slow activation both in the absence and presence of azide (Fig. 5).
Sodium Azide Does Not Inhibit ATP Hydrolysis by the V-ATPase-The effect of ADP and azide on ATP hydrolysis by the purified, reconstituted V-ATPase was tested using two different assays: a coupled spectrophotometric assay that contains an ATP-regenerating system that prevents the accumulation of ADP during the course of the assay, and a radioisotopic method that directly measures the release of 32 P i from [␥-32 P]ATP. As can be seen from Fig. 7, azide had no significant effect on ATP hydrolysis in either assay, whereas ADP inhibited release of 32 P i from [␥-32 P]ATP. Consistent with the latter result was the increase in 32 P i release observed on addition of the regenerating system, which should eliminate inhibition due to ADP generated during the course of the assay. Azide also caused no increase in the degree of inhibition observed in the presence of ADP. It should also be noted that addition of CCCP caused a large increase in ATP hydrolysis by the reconstituted V-ATPase and that ATP hydrolysis was completely inhibited by the specific V-ATPase inhibitor concanamycin.
To determine whether at sufficiently low concentrations CCCP might cause the same effect as azide, namely inhibition of proton transport without stimulation of ATP hydrolysis, the concentration dependence of both proton transport and ATP hydrolysis was tested for CCCP and sodium azide. As can be seen in Fig. 8, this was not the case. Stimulation of ATP hydrolysis generally paralleled inhibition of proton transport for CCCP (Fig. 8A), whereas with azide, inhibition of proton transport was observed without significant stimulation of ATP hydrolysis (Fig. 8B). Interestingly, low concentrations of CCCP actually caused a somewhat larger change in ATP hydrolysis than in proton transport. This may be due to the fact that at low CCCP concentrations, although the vesicles are made slightly more permeable to protons than in the control case, the V-ATPase is still able to keep up with the passive leakage rate and maintain nearly the same level of steady state acidification, even though ATP hydrolysis is stimulated. These results suggest that azide is not acting as a simple uncoupler of proton transport, as with CCCP.
Sodium Azide Does Not Increase Passive Proton Transport in Reconstituted Vesicles-To further test the possibility that azide is acting as a nonspecific protonophore, thereby dissipating the proton gradient generated, reconstituted vesicles loaded with 150 mM KCl were prepared and then diluted into potassium-free media containing 150 mM NaCl and valinomycin. The efflux of potassium creates an internally negative membrane potential that is capable of driving proton influx, provided that a proton conduction pathway is present. Proton uptake was measured using the fluorescence probe ACMA. As can be seen in Fig. 9, addition of 400 M sodium azide did not significantly increase the proton permeability of the reconstituted vesicles.
In contrast, addition of the proton ionophore CCCP led to a rapid fluorescence quenching corresponding to intravesicular acidification.
We also wished to determine whether azide could act as a proton ionophore in the case where a positive interior membrane potential was being generated (as occurs during ATP dependent proton transport by the V-ATPase). To test this, reconstituted vesicles were prepared in buffer containing 150 mM NaCl and no potassium. In addition, the vesicles were loaded during their formation with the membrane-impermeant, pH-sensitive fluorescence probe HPTS (41). This dye FIG. 6. Concentration dependence of sodium azide inhibition of proton transport activity. Proton transport was measured as described in Fig. 1 using stripped clathrin-coated vesicles (22 g of protein) (q) or reconstituted vesicles (2 g of protein) (E) prepared in the solubilization buffer as described under "Experimental Procedures." The vesicles were equilibrated in the presence of 1 mM MgSO 4 and the indicated concentration of sodium azide. The reaction was started by the addition of MgATP such that the final concentrations of Mg 2ϩ and ATP were 1.5 and 0.5 mM, respectively. After the fluorescence change had plateaued (6 min), 1 M CCCP was added, and the return of fluorescence was measured. The total acidification (the change in fluorescence occurring in response to CCCP) was used to characterized proton transport activity, with 100% corresponding to the fluorescence change observed in the absence of azide.
shows increased fluorescence emission at 520 nm with increasing pH. After removal of the external dye, the HPTS-loaded reconstituted vesicles were diluted into buffer containing 150 mM KCl in the presence of valinomycin. The resultant interior positive membrane potential thus provides a driving force for efflux of protons from the vesicles, which was detected by the fluorescence intensity of the trapped HPTS. Fig. 10 shows that although there is some intrinsic passive proton conductance of the reconstituted vesicles (as demonstrated by the fluorescence increase observed following addition of valinomycin), sodium azide does not significantly increase the rate of passive proton conductance. In contrast, a rapid increase in fluorescence corresponding to an increase in intravesicular pH was observed on addition of CCCP. Thus, azide does not appear to act as a proton ionophore under our conditions. Consistent with this is the observation that sodium azide does not increase ATP hydrolysis by reconstituted V-ATPase, as was observed for CCCP ( Figs. 7 and 8). DISCUSSION In this work, we observed that the V-ATPase from clathrincoated vesicles is subject to two different types of inhibition by ADP. The first is competitive inhibition with respect to ATP, with a K i for ADP of 11 M. The other occurs after preincubation of the V-ATPase in the presence of ADP and Mg 2ϩ , which results in inhibition of the initial rate of proton transport followed by reactivation of the enzyme over the course of several minutes following addition of ATP. The second effect was observed at ADP concentrations as low as 0.1-0.2 M, suggesting the involvement of a high affinity site distinct from that involved in competitive inhibition.
Competitive inhibition of the V-ATPases of kidney and osteoclasts by ADP has previously been reported with K i values of 37 and 17 M, respectively (33), whereas ADP was reported to be a noncompetitive inhibitor of the yeast V-ATPase (5), although in neither case were effects on proton transport evaluated. Allosteric inhibition by ADP of ATP hydrolysis was reported for the chromaffin granule V-ATPase (44), but this effect required much higher concentrations of ADP (100 M). Similarly, ADP was shown to increase the cooperativity of ATP binding to the chromaffin granule V-ATPase, but the concentrations of ADP required were rather high (25-200 M), and only steady state analysis was carried out, so that no effects of preincubation with ADP and Mg 2ϩ were investigated (45). Binding of [ 3 H]ADP to a high affinity site (K d ϭ ϳ70 nM) were observed for the chromaffin granule V-ATPase (46), but the effect of occupancy at this site on activity was not reported. Finally, sulfite has been suggested to activate the V-ATPase of yeast by causing release of inhibitory ADP that accumulates at high ATP concentrations during the initial phase of the reaction (47), but no direct interaction of the V-ATPase with ADP was investigated in this study. Our study thus represents the first direct demonstration that ADP (at submicromolar concentrations) in the presence of Mg 2ϩ is able to stabilize an inhibited state of the enzyme, which is only slowly activated on addition of ATP.
The observed inhibitory effect of ADP on the V-ATPase is similar to that described for the F-ATPase. Thus, preincubation of the F-ATPase with ADP and Mg 2ϩ results in a lag phase in ATP hydrolysis that has been interpreted as indicating the formation of an inhibited complex between the F-ATPase and ADP bound with high affinity at a catalytic site (42,48,49). The same complex is believed to be formed during ATP hydrolysis 32 P i release assay, as indicated by the stimulation observed in the presence of the ATP regenerating system. . In all cases, the reaction was started by the addition of reconstituted V-ATPase (2 g of protein). For the coupled spectrophotometric assay, 100% activity corresponds to a specific activity of 3.8 mol/min/mg of protein, whereas for the 32 P i release assay, 100% activity corresponds to a specific activity of 2.1 mol/min/mg of protein. As can be seen, part of the difference in specific activity measured by these two methods can be accounted for by the accumulation of ADP in the as a transient intermediate during turnover (42). ADP bound to the noncatalytic sites of the F-ATPase has also been reported to cause hysteretic inhibition of activity (50). Azide has been shown to inhibit the F-ATPase by stabilization of the ADP-inactivated complex (43,51,52). Interestingly, ATP synthesis is much less sensitive to azide inhibition than ATPase activity (26,27).
Because of the similar inhibitory effects of ADP on the V and F-ATPases, we decided to test the effects of azide on V-ATPase activity. Azide has previously been reported to not inhibit V-ATPase activity, but the principal assay used in these studies was ATP hydrolysis (28 -30). We therefore tested the effects of azide on ATP-dependent proton transport by the V-ATPase. We observed that azide does inhibit proton transport by the V- FIG. 8. Concentration dependence of effects of CCCP and azide on proton transport and ATP hydrolysis by the purified, reconstituted V-ATPase. Proton transport (E) into reconstituted vesicles was measured as described in Fig. 1. Vesicles (2 g of protein) were equilibrated in 0.7 ml solubilization buffer containing ACMA and valinomycin together with the indicated concentrations of CCCP or sodium azide. The reaction was started by addition of Mg 2ϩ and ATP such that the final concentrations were 1.5 and 0.5 mM, respectively. After the fluorescence change had plateaued, 1 M CCCP was added, and the return of fluorescence was quantitated. This measure of total acidification (change in fluorescence in response to addition of 1 M CCCP) was used to characterize proton transport activity, with 1.0 corresponding to the fluorescence change observed with no CCCP or azide added at the beginning of the reaction. ATPase activity (q) of the purified, reconstituted V-ATPase was measured using a continuous, spectrophotometric assay as described under "Experimental Procedures." Reconstituted vesicles (2 g of protein) were assayed in the presence of 0.5 mM ATP and 1.5 mM MgSO 4 in the presence of the indicated concentrations of CCCP or sodium azide. Activity is again expressed relative to reconstituted vesicles assayed in the absence of CCCP or azide, which is defined as 1.0.  ATPase, with 50% inhibition observed at 200 -400 M. In agreement with previous observations, however, azide does not have a measurable effect on ATP hydrolysis.
These results suggest that azide acts to functionally uncouple ATP hydrolysis and proton transport by the V-ATPase. One possible mechanism by which this might occur is if azide can act as a proton ionophore. We therefore tested the ability of azide to act as a proton ionophore in the presence of both an interior positive and an interior negative membrane potential. In neither case did we observe any increase in passive proton conductance in the presence of azide. This suggests that azide does not inhibit proton transport by the V-ATPase by simply dissipating the proton gradient generated, although this possibility is difficult to rule out with certainty.
A second possibility is that azide specifically uncouples proton transport from ATP hydrolysis by the V-ATPase. In fact, several lines of evidence suggest that the V-ATPase is poised to change the efficiency of coupling between hydrolysis and transport. Thus, mild proteolysis results in a complete loss of proton transport under conditions where only 50% of the ATP hydrolysis is lost (53). High concentrations of ATP (54), high concentrations of sulfite (47), and the presence of detergent (55) have also been suggested to alter the efficiency of coupling between ATP hydrolysis and proton transport. In fact, a "slip" mechanism has been proposed as a means of regulating vacuolar acidification in vivo (1,4), although the nature of the cellular factors controlling coupling efficiency remains uncertain. Azide is not simply causing a complete uncoupling of ATP hydrolysis from proton transport, however, because in this case, azide would be expected to increase ATP hydrolysis by the reconstituted V-ATPase (as is observed on addition of CCCP). No such increase in ATP hydrolysis is observed on addition of azide.
Rather, azide appears to reduce the coupling efficiency in such a way that proton transport is decreased without stimulation or inhibition of ATP hydrolysis.
An alternative possible explanation for the observed results is that azide inhibits only tightly coupled V-ATPase after formation of a pH gradient. In the case of proton transport, this would explain why azide did not inhibit the initial rate of proton transport but did inhibit total acidification, because no pH gradient would be present prior to the addition of ATP. In addition, the ATP hydrolysis observed in the absence of CCCP may represent V-ATPase that is not tightly coupled (and hence not susceptible to inhibition by azide), whereas the ATP hydrolysis observed in the presence of CCCP is obviously uncoupled (and hence not inhibited). Such a mechanism is intriguing because it suggests the possibility that the proton gradient itself may be feeding back on the V-ATPase cycle in such a way that the enzyme activity can be inhibited in response to the pH gradient generated.
The connection between the effects of ADP and azide on V-ATPase activity is uncertain. Although the enzyme is still able to form an inactive complex with ADP and Mg 2ϩ in the presence of azide, the kinetics of reactivation after addition of ATP are not greatly different in the presence or absence of azide, suggesting that azide is not stabilizing the ADP-inactivated state of the enzyme, as proposed for the F-ATPases.
The nature of the high affinity ADP site responsible for inhibition (catalytic or regulatory), the role of Mg 2ϩ and the molecular basis for the azide sensitivity of the V-ATPase remain to be determined. It has previously been shown that both rapidly and slowly exchangeable nucleotide binding sites can be detected on the V-ATPase (21,22), and it is possible that ADP causes this slowly reversible inactivation by binding to the slowly exchangeable site. Although the V-ATPase is unlikely to be exposed to ADP in the absence of ATP under intracellular conditions, the existence of such a slowly reversible inactive state of the enzyme may be of relevance to regulation of vacuolar acidification in vivo. In particular, additional information will be required concerning the possible cellular factors influencing the equilibrium between the active and ADP-inhibited states.