Catalytic Activity of the α3β3γ Complex of F1-ATPase without Noncatalytic Nucleotide Binding Site

A mutant α3β3γ complex of F1-ATPase from thermophilic Bacillus PS3 was generated in which noncatalytic nucleotide binding sites lost their ability to bind nucleotides. It hydrolyzed ATP at an initial rate with cooperative kinetics (Km(1), 4 μM; Km(2), 135 μM) similar to the wild-type complex. However, the initial rate decayed rapidly to an inactivated form. Since the inactivated mutant complex contained 1.5 mol of ADP/mol of complex, this inactivation seemed to be caused by entrapping inhibitory MgADP in a catalytic site. Indeed, the mutant complex was nearly completely inactivated by a 10 min prior incubation with equimolar MgADP. Analysis of the progress of inactivation after initiation of ATP hydrolysis as a function of ATP concentration indicated that the inactivation was optimal at ATP concentrations in the range of Km(1). In the presence of ATP, the wild-type complex dissociated the inhibitory [3H]ADP preloaded onto a catalytic site whereas the mutant complex did not. Lauryl dimethylamineoxide promoted release of preloaded inhibitory [3H]ADP in an ATP-dependent manner and partly restored the activity of the inactivated mutant complex. Addition of ATP promoted single-site hydrolysis of 2′,3′-O-(2,4,6-trinitrophenyl)-ATP preloaded at a single catalytic site of the mutant complex. These results indicate that intact noncatalytic sites are essential for continuous catalytic turnover of the F1-ATPase but are not essential for catalytic cooperativity of F1-ATPase observed at ATP concentrations below ~300 μM.

F 1 -ATPase is the extrinsic membrane sector of H ϩ -ATP synthase and comprises five different subunits in a stoichiometry of ␣ 3 ␤ 3 ␥ 1 ␦ 1 ⑀ 1 (1). According to the crystal structure of bovine heart mitochondrial F 1 (MF 1 ) 1 (2), the ␣ and ␤ subunits are arranged alternately to a form hexagonal ␣ 3 ␤ 3 . The six nucleotide binding sites are located at different interfaces between the ␣ and ␤ subunits. The three catalytic sites are mainly on the ␤ subunits, whereas the three other sites called noncatalytic nucleotide binding sites are mainly on the ␣ subunits. The overall structural topologies of the catalytic and noncatalytic sites are very similar to each other and both sites contain the two sequences known as the Walker motif A and B, which are commonly found in many nucleotide-binding proteins (3). Motif A, which is also called as P-loop, has the consensus sequence, GXXXXGK(T/S), and motif B consists of a stretch of four consecutive hydrophobic residues followed by Asp.
The function of the noncatalytic site is obscure. However, recent studies suggest that the F 1 -ATPase is prone to develop turnover-dependent inactivation and the noncatalytic sites play a role in relieving the inactivation (4,5). When nucleotidedepleted MF 1 or F 1 -ATPase from the thermophilic Bacillus PS3 (TF 1 ) hydrolyzes relatively low concentration of ATP, three kinetic phases are often observed in the presence of an ATP regenerating system. An initial burst rapidly decelerates to an intermediate rate that, in turn, gradually accelerates to a final steady-state rate. It has been postulated that transition from the initial phase to the intermediate phase is caused by turnoverdependent entrapment of inhibitory MgADP in a catalytic site (6 -8), and transition from the intermediate phase to the final phase reflects slow binding of ATP to the noncatalytic sites, which promotes dissociation of inhibitory MgADP from the affected catalytic site. After prior loading of a catalytic site of MF 1 (4,9,10), TF 1 (5,11), and chloroplast F 1 -ATPase (12) with MgADP, the enzymes hydrolyze ATP with extended lag. The observation that the binding of ATP to noncatalytic sites stimulates ATPase activity was also reported (13)(14)(15)(16)(17). All these kinetic features are observed with the ␣ 3 ␤ 3 ␥ complex of TF 1 (18). The ␣ 3 ␤ 3 ␥ complex of TF 1 containing ␣ subunits with a mutation in the Walker motif B, ␣-D261N, dissociates inhibitory MgADP only slowly even in the presence of ATP and the transition from the intermediate phase to the final phase almost disappeared, exhibiting a low final rate of ATP hydrolysis, about 30% of that of the wild-type complex (18). Conversely, the ␣ 3 ␤ 3 ␥ complex of TF 1 containing ␤ subunits with a mutation in the Walker motif A, ␤-T165S, efficiently dissociates inhibitory MgADP and exhibits a severalfold higher final rate of ATP hydrolysis than that of the wild-type complex (19). These results suggest that F 1 -ATPase in the inactivated state with inhibitory MgADP in a catalytic site is reactivated by ATP binding to noncatalytic sites. However, important unanswered questions remain. For instance, does enzyme containing inhibitory MgADP in a single catalytic site have weak residual ATPase activity or is it completely inactive? Is release of inhibitory MgADP totally dependent on ATP binding to noncatalytic sites or is there slow release of inhibitory MgADP that is independent of ATP binding to noncatalytic sites?
The role of noncatalytic site in the cooperative kinetics of F 1 -ATPase also remains to be clarified. F 1 -ATPase exhibits negative cooperativity characterized with two or three apparent K m values which are 1-30 M, 100 -300 M, and above 400 M (20 -24). This apparent negative cooperativity is observed also for the membrane-bound enzyme (25) and proton translocation (26). Slow binding of ATP to noncatalytic sites can explain apparent negative cooperativity at relatively high concentration of ATP represented by the highest K m value (4). Weber et al. reported a single K m value for the mutant F 1 -ATPase from Escherichia coli (EF 1 ) with mutations ␣-D261N/ ␣-R365W in which nucleotide binding to noncatalytic sites was greatly diminished (27). They stated that the kinetics of this mutant showed no deviation from simple monophasic Michaelis-Menten kinetics. However, they assayed ATPase activity by measuring P i release, which is not suitable to monitor fluctuation in rate during assay. Therefore, as stated in their paper, they did not scrutinize the kinetic behavior at very low substrate concentrations, which is necessary to detect a K m at 1-30 M. Similarly, Yohda et al. reported that ␣(D261N) 3 ␤ 3 ␥ complex of TF 1 did not exhibit cooperativity, but again they examined kinetics only above 20 M ATP (28). Therefore, the effect of noncatalytic sites on cooperative kinetics of F 1 -ATPase remains unsettled.
Since the covalent modification of the noncatalytic sites with 5Ј-p-fluorosulfonylbenzoyladenosine inactivates ATPase activity completely (29), it is even possible to argue that noncatalytic sites are essential for the activity of F 1 -ATPase, although their participation in catalysis is indirect. The mutants reported so far whose noncatalytic sites are impaired, namely EF 1 (␣-D261N/␣-R365W) and ␣(D261N) 3 ␤ 3 ␥ complex of TF 1 , have considerable ATPase activity (18,27,28). Especially, the EF 1 mutant showed ATPase activity even under the condition where noncatalytic sites were supposed to be empty and Weber et al. concluded that occupancy of the noncatalytic sites by adenine nucleotides was not required for catalysis. However, ambiguity remains because both EF 1 (␣-D261N/␣-R365W) and ␣(D261N) 3 ␤ 3 ␥ complex of TF 1 have the ability to bind nucleotide to the noncatalytic sites even though the affinity is decreased.
To obtain more discriminating data on the role of noncatalytic sites, it is necessary to characterize a mutant F 1 -ATPase that completely lacks the ability to bind nucleotides to noncatalytic sites. To generate such a mutant, we have replaced four amino acid residues in Walker motif A and B sequences of the TF 1 -␣ subunit by Ala residues, and analyzed nucleotide binding properties and ATP hydrolysis catalyzed by the ␣ 3 ␤ 3 ␥ complex containing the mutated ␣ subunits under a wide range of ATP concentration. Comparison of this mutant ␣ 3 ␤ 3 ␥ (⌬NC) complex and the wild-type ␣ 3 ␤ 3 ␥ complex has revealed the essential role of noncatalytic sites in steady-state catalytic turnover of F 1 -ATPases.
Analytical Methods-Protein concentrations of TF 1 and ␣ 3 ␤ 3 ␥ complexes were determined by measurement of absorbance at 280 nm using the factor 0.45 of absorbance for 1 mg/ml of protein. ATPase activity was measured at 25°C in the presence of an ATP regenerating system. An assay mixture contained 50 mM Tris-Cl (pH 8.0), 100 mM KCl, indicated concentrations of ATP, 2.5 mM phosphoenolpyruvate, 50 g/ml pyruvate kinase (rabbit muscle), 50 g/ml lactate dehydrogenase (pig muscle), and 0.2 mM NADH. Pyruvate kinase and lactate dehydrogenase (Boehringer Japan, Tokyo) were diluted from solution in glycerol. These amounts of auxiliary enzymes were confirmed to be sufficient for the rapid ATP hydrolysis in the initial burst phase of catalysis. MgCl 2 concentration was maintained at 2 mM excess over that of ATP in the assay mixture. Typically, the reaction was initiated by addition of the enzyme to 2 ml of the assay mixture and the rate of ATP hydrolysis was monitored as the rate of oxidation of NADH determined by the absorbance decrease at 340 nm. The data were stored in an on-line computer for further analyses. We attached a device on the photometer lid that enabled us to start the reaction by injecting the enzyme solution without opening it. The spectrophotometer was equipped with a small stirrer to ensure rapid mixing. We confirmed that the maximum dead time of measurement was below 4 s after the start of the reaction and the data from 5 s to 20 s were usually used for analysis. The initial rates were obtained from exponential extrapolation of the experimental data between 5 and 20 s to time zero. One unit of activity was defined as the activity that hydrolyzed 1 mol of ATP/min. Assessment of nucleotide binding to catalytic and noncatalytic sites of the complexes by photoaffinity labeling with 2-N 3 -[ 3 H]AT(D)P was carried out as described previously (37). Briefly, the solution (100 l) containing 1.5 mg of the wild-type or ⌬NC ␣ 3 ␤ 3 ␥ complex, 150 M 2-N 3 -[ 3 H]AT(D)P (2700 cpm/ nmol), 2 mM MgCl 2 , 100 M EDTA, and 50 mM Tris-Cl (pH 7.5), was irradiated for 40 min at room temperature with a Minerallight, and digested by L-1-tosylamido-2-phenylethyl chloromethyl ketone-treatedtrypsin after denaturation of the proteins and removal of unbound nucleotides. An aliquot of the digested solution was injected into a C4 reversed-phase HPLC column and developed with a gradient of CH 3 CN in 0.1% HCl as follows: 0 -10 min, 0%; 10 -100 min, 0 -24%; 100 -115 min, 24 -48%, 115-120 min, 48 -80%. Fractions, 1 ml each, were collected and radioactivity of each fraction was measured. Difference spectra induced by the interaction between TNP-ATP and the proteins and single-site hydrolysis of TNP-ATP were measured according to previous papers (38 -40). The bound nucleotide content of the enzyme complex was determined after separating free nucleotide from enzyme-bound nucleotides by centrifuge elution using a 1-ml column of Sephadex G-50, extracted with perchloric acid, and analyzed by HPLC (38). Release of [ 3 H]ADP from the complexes was monitored according to the methods described by Jault et al. (18).

Generation of the Stable
The crystal structure of MF 1 shows that, in the noncatalytic nucleotide binding site, Lys and Thr in the Walker motif A and Asp in the motif B of the ␣ subunit lie close to the terminal phosphate and Mg 2ϩ of bound Mg-AMP-P(NH)P, a substrate analogue (2). In addition, the conserved Asp just adjacent to the Asp of the motif B sequence of the ␣ subunit also contributes to the noncatalytic nucleotide binding site. The residues of TF 1 -␣ subunit equivalent to the above residues of MF 1 -␣ subunit are ␣-Lys-175, ␣-Thr-176, ␣-Asp-261, and ␣-Asp-262. Therefore, we replaced these residues of TF 1 -␣ subunit by Ala residues. Four mutations, ␣-Lys-175 3 Ala, ␣-Thr-176 3 Ala, ␣-Asp-261 3 Ala, and ␣-Asp-262 3 Ala, were simultaneously introduced into the ␣ subunit gene on an expression plasmid for ␣ 3 ␤ 3 ␥ complex. Although it was reported that the mutations at ␣-Lys-175 impaired assembly of subunit complexes (27,28,41), the mutant ␣ subunit constructed here assembled normally into ␣ 3 ␤ 3 ␥ complex with the ␤ and ␥ subunits. The ⌬NC complex was stable and purified to homogeneity by the same method used for the purification of the wild-type complex including the incubation at 60°C for 30 min.
Nucleotide Binding Properties of the ⌬NC ␣ 3 ␤ 3 ␥ Complex-To assess binding of adenine nucleotides to the noncatalytic sites, tryptic digests from the wild-type and ⌬NC ␣ 3 ␤ 3 ␥ complexes which were photolabeled with 2-N 3 -[ 3 H]ATP in the presence of Mg 2ϩ were analyzed. The profiles of tryptic peptides resolved by reversed-phase HPLC are shown in Fig. 1. Elutions were carried out under the same conditions reported previously in which assignment of radioactive peaks was established (42). A radioactive peak eluted at around 78 min contains the tryptic peptide with ␤-Tyr-364 derivatized, which is a part of the noncatalytic site, and peaks eluted between 90 and 100 min contain the tryptic peptides with ␤-Tyr-341 derivatized, which is a part of the catalytic site (43). It has been shown that the tryptic peptides derived from the catalytic site are often eluted as two (or more) peaks as shown in Fig. 1A because of the heterogeneity arising from hydrolysis of ATP tethered to ␤-Tyr-341 (35,37). When the elution profile of the ⌬NC complex (Fig. 1B) is compared with that of the wild-type ␣ 3 ␤ 3 ␥ complex (Fig. 1A), it is obvious that the former does not have a peak at around 78 min (shown by an arrow). The experiments with 2-N 3 -[ 3 H]ATP in the absence of Mg 2ϩ , 2-N 3 -[ 3 H]ADP in the absence and presence of Mg 2ϩ gave the same results; there was no radioactive peak at the position corresponding to a peptide derived from noncatalytic sites whereas a peak corresponding to a peptide from catalytic site was always detected (data not shown). These results show that 2-N 3adenine nucleotides cannot bind to the noncatalytic sites of the ⌬NC complex.
The binding of nucleotide was further examined by TNP-ATP-induced difference spectra (Fig. 2). The difference absorption spectra induced by binding of TNP-AT(D)P to the isolated wild-type ␣ or ␤ subunit are significantly different from each other. A trough at 450 nm and a peak at 510 nm were observed for the ␣ subunit, whereas a trough at 395 nm and a peak at around 420 nm were observed for the ␤ subunit (Fig. 2, uppermost and lowermost traces) (38). Therefore, it is possible to determine the subunit localization of the TNP-AT(D)P binding site of the complex by this means. It has been shown that the wild-type ␣ 3 ␤ 3 ␥ complex binds TNP-ADP preferentially to a single high affinity catalytic site on the ␤ subunit until molar ratio of TNP-ADP to F 1 -ATPase is 1.25:1. After this site is filled, the second site occupied by TNP-ADP is a noncatalytic binding site on the ␣ subunit (40). Similar binding characteristics were observed for TNP-ATP and, for example, a large contribution by the ␣ subunit specific difference spectrum is obvious in the difference spectrum at a 4:1 TNP-ATP⅐wild-type ␣ 3 ␤ 3 ␥ complex molar ratio (Fig. 2). In contrast, the ⌬NC complex showed spectra typical for the ␤ subunit even at a 6:1 TNP-ATP⅐complex molar ratio (Fig. 2). This indicates that TNP-ATP binds exclusively to the catalytic sites on the ␤ subunits of the ⌬NC complex and that noncatalytic sites on the ␣ subunit is unable to bind TNP-ATP. Based on these results, we conclude that the noncatalytic sites of the ⌬NC complex do not bind adenine nucleotides.
ATPase Activity of the ⌬NC ␣ 3 ␤ 3 ␥ Complex- Fig. 3 shows time courses of ATP hydrolysis by the wild-type and ⌬NC ␣ 3 ␤ 3 ␥ complexes in the presence of an ATP regenerating system. The wild-type complex hydrolyzed 20 M ATP in three phases (trace b), an initial burst decelerated to an intermediate phase that then accelerated to a final state (18). At 2 mM ATP, the tran- Difference spectra induced by the interaction between TNP-ATP and the ⌬NC ␣ 3 ␤ 3 ␥ complex or the isolated subunits of TF 1 . TNP-ATP was added to 2.0 M ⌬NC complex at indicated molar ratios. Likewise, 10 M TNP-ATP was added to 2.0 M isolated wildtype ␣ or ␤ subunit. All the solutions contained 2 mM MgCl 2 , 100 M EDTA, and 50 mM Tris-Cl (pH 7.5). Difference spectra were measured 5 min after mixing the components. A difference spectrum induced by the interaction between TNP-ATP and the wild-type ␣ 3 ␤ 3 ␥ complex at a molar ratio 4:1 (TNP-ATP complex) was also shown for comparison. Other experimental conditions are described under "Experimental Procedures." sition from the intermediate phase to the final phase was not seen and it appeared to proceed in two phases, an initial burst phase and a following decelerated constant phase (trace d). The ⌬NC complex also hydrolyzed 20 M and 2 mM ATP with an initial burst (traces f and i). The rates of the initial bursts by ⌬NC complex at both concentrations were very similar to those observed for the wild-type complex which are shown by overlaid dotted lines (traces g and j). However, the initial burst of the ⌬NC complex rapidly decelerated and hydrolysis stopped in a short period. We analyzed the kinetics of initial rates of the burst phase of the ⌬NC complex at a wide range of ATP concentrations (1-2000 M) and compared them to those of the wild-type complex (Fig. 4). Although there were some differences at high ATP concentrations, the ⌬NC complex hydrolyzed ATP in a similar manner to the wild-type complex. As shown in the inset of Fig. 4, the Eadie-Hofstee plots of the initial rates of ATP hydrolysis by the wild-type and ⌬NC complexes are concave downward, indicating that ATP hydrolysis by the ⌬NC complex, like the wild-type complex, exhibits negative cooperativity. Apparent kinetic parameters are calculated by a nonlinear regression curve-fitting (24) and summarized in Table I. At least two sets of K m and V max are necessary to simulate the experimental data for both wild-type and ⌬NC complexes. The values of K m(1) and V max(1) obtained for the ⌬NC complex are almost the same as K m (1) and V max(1) of the wild-type complex. K m (2) and V max(2) of the ⌬NC complex are also close (about 70%) to those of the wild-type complex. Thus, the ␣ 3 ␤ 3 ␥ complex without functional noncatalytic sites exhibits cooperative kinetics which are very similar to those of the ␣ 3 ␤ 3 ␥ complex with intact noncatalytic sites.
Inactivation of the ⌬NC ␣ 3 ␤ 3 ␥ Complex-Although the initial rate of ATP hydrolysis by the ⌬NC complex obeys similar kinetics to those of the wild-type complex, it is rapidly inactivated during catalytic turnover and hydrolysis stops completely in a short period as described above (Fig. 3, traces f and  i). The dependence of the rate of inactivation on ATP concentration is shown in Fig. 5 (A and B). The time course of inactivation was exponential curve, and the first order rate constants of inactivation were obtained at various ATP concentrations. As shown in Fig. 5C, the rate constants of inactivation exhibit monophasic dependence on ATP concentration. The ATP concentration that gave a half-maximal rate of inactivation (apparent K d ) was 5 M. This value agrees well with the K m(1) (4 M) obtained from analysis of the rate in the initial burst phase. The maximal rate of inactivation was 0.33 s Ϫ1 . Combining the rate of inactivation and the rate of ATP hydrolysis, we can calculate the average number of catalytic turnovers required for the inactivation at each ATP concentration. For example, at 0.53, 1.1, 4.4, 11, 22, and 33 M ATP, 17, 31, 63, 80, 97, and 105 turnovers are required for the inactivation, respectively. Therefore, inactivation is not simply proportional to turnover number. When ATP concentration is low, the enzyme is inactivated in relatively few turnovers. More turnovers are required for inactivation as ATP concentrations increased. Accordingly, the size of initial burst is smaller when  Table I. Open circle, wild-type complex; closed circle, ⌬NC complex. Other experimental conditions were described under "Experimental Procedures."

TABLE I Kinetic parameters of the initial rates of ATP hydrolysis by the wild-type
ATP concentration is low than when it is high as seen in Fig. 5  (A and B).

Entrapment of Inhibitory MgADP in a Catalytic
Site-To analyze bound nucleotides of the inactivated ⌬NC complex, the ⌬NC complex was inactivated by a 20-min incubation with 20 M ATP in the presence of 2 mM MgCl 2 and free nucleotides were removed by passing through the Sephadex G-50 centrifuge column. The complex recovered in the effluent did not have ATPase activity, ensuring that it was still in an inactivated state. Analysis of the bound nucleotides revealed that the inactivated ⌬NC complex contained 1.5 mol of ADP/mol of the complex, while a control ⌬NC complex that was not previously exposed to nucleotide contained none (Ͻ0.05 mol/mol). The role of the bound ADP in the inactivated complex was further investigated after incubating the complex with MgADP. As reported previously (5), the wild-type complex with preloaded MgADP by a prior incubation with equimolar MgADP for 5 min showed attenuated initial rates of hydrolysis of 20 M and 2 mM ATP rather than initial burst, that accelerated gradually to a constant phase (Fig. 3, traces a and c). In contrast, the ⌬NC complex with preloaded MgADP could not hydrolyze ATP at all (Fig. 3, traces e and h). Reactivation did not occur even after long incubation (5 h) of the assay mixture. To determine the stoichiometry of inhibitory MgADP bound, the ⌬NC complex was preincubated with MgADP at various molar ratios to the complex for 10 min, and the residual ATPase activities at the initial burst phase were measured in the presence of 2 mM ATP. As shown in Fig. 6, the extent of inactivation of the ⌬NC complex was almost proportional to the amount of added MgADP until the concentration of MgADP reached to a 1:1 molar ratio to the complex where nearly complete inactivation was observed. Since noncatalytic sites in the ⌬NC complex cannot bind MgADP, we conclude that the ⌬NC complex is completely inactivated by entrapping MgADP in a single catalytic site.
Effect of LDAO-It has been shown in previous studies that the ATPase activities of TF 1 and the wild-type ␣ 3 ␤ 3 ␥ complex are stimulated significantly by the neutral detergent LDAO (5,18,44). The stimulation of ATPase activity by LDAO is thought to promote release of inhibitory MgADP from a catalytic site in the presence of ATP (19). When LDAO was present in the reaction mixture, the wild-type complex hydrolyzed 2 mM ATP at a constant rate from the beginning. The constant rate in the presence of LDAO was very close to the rate in the initial burst in the absence of LDAO (Fig. 7A). It appears that LDAO does not affect the rate of ATP hydrolysis in the initial burst, but rather allows the initial burst to continue linearly without deceleration. Similar to the wild-type complex, LDAO had little effect on the initial burst of the ⌬NC complex (Fig. 7B). However, different from the wild-type complex, the initial burst decelerated even in the presence of LDAO and reached a slow, final rate. When LDAO was added to the ⌬NC complex, which was previously inactivated by aging in the assay mixture or by prior incubation with MgADP, the same slow rate was restored (Fig. 7C). The above experiments were performed at 2 mM ATP. However, a similar effect of LDAO on the ⌬NC complex was observed when it hydrolyzed 20 M ATP (data not shown). The stimulating effect of LDAO on the final steady-state hydrolysis was saturated at 0.04% LDAO for the wild-type complex and was nearly (but not completely) saturated at 0.15% LDAO for the ⌬NC complex (Fig. 7D). If the mechanism of action of LDAO is only to amplify the conformational signal generated by the ATP binding to noncatalytic sites that causes release of inhibitory MgADP, then LDAO would not have an effect on the ⌬NC complex. Probably, LDAO has a direct effect on the catalytic site occupied by inhibitory ADP. This effect might be small because the extent of activity restored by LDAO in the case of the ⌬NC complex is much smaller than that of the wild-type complex as described above.
Release of Inhibitory MgADP-The effect of ATP and LDAO on the release of preloaded, inhibitory [ 3 H]ADP from the wildtype and ⌬NC complexes was examined (Fig. 8). For the wildtype complex, release of preloaded [ 3 H]ADP was promoted by ATP but not by LDAO (Fig. 8A). However, when both ATP and LDAO were present in the solution, the release was greatly enhanced and most When LDAO and ATP are present, inhibitory MgADP is released slowly and an equilibrium is established with a small fraction of the complex free of inhibitory MgADP resulting in partial restoration of activity.
Single-site TNP-ATP Hydrolysis and Chase-promotion-TF 1 (38) and the ␣ 3 ␤ 3 ␥ complex (40) hydrolyze TNP-ATP slowly when TNP-ATP is added to the enzyme in a substoichiometric molar ratio. Slow hydrolysis of TNP-ATP is greatly accelerated by chase-promotion with ATP. It has been suggested that the ATP binding site responsible for chase-promotion is the second catalytic site to be filled (39,45). Fig. 9 illustrates time courses of TNP-ATP hydrolysis by the wild-type and ⌬NC complexes. Similar to the wild-type complex, the ⌬NC complex slowly hydrolyzed substoichiometric TNP-ATP and chase-promotion with ATP accelerated the hydrolysis of the TNP-ATP. From this result we conclude that participation of noncatalytic sites is not necessary for cooperativity between two catalytic sites.

Noncatalytic Sites Are Essential for Continuous Catalytic
Turnover-This work provides solid support for the view that entrapping inhibitory MgADP at a catalytic site, either during incubation with MgADP or during turnover under assay conditions, causes inactivation of F 1 -ATPase. In addition, it is now clear that enzyme that retains inhibitory MgADP at a single catalytic site is completely inactive in ATP hydrolysis (Figs. 3 and 6). Differing from the wild-type enzyme, inhibitory MgADP is not released from the ⌬NC complex even in the presence of ATP (Fig. 8). Since catalytic sites in the ⌬NC complex are as intact and available for ATP binding as those of the wild-type complex, the failure of ATP to promote release of inhibitory MgADP from a catalytic site can only be attributed to the lack of ability of noncatalytic sites to bind ATP. Thus, it is concluded that ATP binding to noncatalytic sites is essential for continuous catalytic turnover. This may have physiological importance since Richard et al. suggested that ATP synthesis by H ϩ -ATP synthase from thermophilic Bacillus PS3 was stimulated when the noncatalytic sites were occupied by ATP (46).
Noncatalytic Sites Are Not Essential for Cooperative Kinetics of the F 1 -ATPase-Comparison of the rates of ATP hydrolysis in the initial burst by the wild-type and ⌬NC complexes revealed that both enzymes obey very similar cooperative kinetics (Fig. 4, Table I). Both the K m value and the V max value of the ⌬NC complex are almost the same or close to the corresponding values of the wild-type complex. In addition, the ⌬NC complex can catalyze single-site hydrolysis and chase-promotion using TNP-ATP as a substrate (Fig. 9). The remarkably similar kinetics of the ⌬NC complex and the wild-type complex strongly indicates that the catalytic sites of the ⌬NC complex are intact and behave in a similar manner to the wild-type complex. In other words, noncatalytic sites contribute little, if anything, to the cooperative kinetics of the ␣ 3 ␤ 3 ␥ complex. Thus, cooperative features of ATP hydrolysis by F 1 -ATPase characterized with two sets of parameters, K m(1) ϭ 1-30 M and K m(2) ϭ 100 -300 M, reflect catalytic site to catalytic site cooperativity. The apparent K m (2) of 140 M observed here agrees well with the K m for proton translocation which was membrane potential independent (47). Cooperativity observed at high ATP concentration (Ͼ400 M) at steady-state catalysis is a phenomenon attributed to slow nucleotide binding to noncatalytic sites (4).
Simultaneous Occupation of Two Catalytic Sites Promotes Entrapment of Inhibitory MgADP at a Catalytic Site-Interestingly, the rate of the progressive inactivation of the ⌬NC complex during turnover shows a hyperbolic dependence on ATP concentration and exhibits an apparent K d of 5 M (Fig. 5). This K d value corresponds to K m(1) (4 M) obtained from initial rate analysis. This K m is thought to reflect a catalytic cycle operating when two catalytic sites are occupied, so-called bi-site catalysis (22). Owing to the very low K d and k cat values for ATP hydrolysis when only one catalytic site is occupied, single site catalysis is not amenable to steady-state kinetic analysis. This means that occupancy of two catalytic sites promotes transition from an active to an inactive enzyme. On the other hand, we observed that loading a catalytic site with exogenous MgADP is sufficient to inactivate the ⌬NC complex completely (Fig. 6). This apparent contradiction can be accommodated as follows. The formation of inactivated MgADP⅐TF 1 from TF 1 and MgADP is a slow process (2 M Ϫ1 s Ϫ1 ) (11) and this is also the case for the ␣ 3 ␤ 3 ␥ complex. The rate-limiting step is most likely to be the isomerization from a transient, active MgADP⅐enzyme complex into the stable, inactive MgADP⅐enzyme complex (6,19). If ATP hydrolysis operating with two catalytic sites can facilitate the isomerization, ATP hydrolysis characterized by K m (1) becomes an apparently responsible step for the generation of the inactive complex. After isomerization, MgADP at one of the two catalytic sites might be released during gel filtration procedures to analyze bound nucleotide. In this mechanism, cooperative interaction between two catalytic sites is assumed to accelerate not only catalysis but also generation of inactive species of the enzyme. It is interesting to note that the inhibition of F 1 -ATPase by azide also progresses during turnover and the rate of progress is dependent on the ATP binding (and/or hydrolysis) characterized by the K m of about 10 M (48), a value similar to the apparent K d above mentioned. Evidence suggests that azide stabilizes the inhibitory MgADP⅐F 1 -ATPase complex (7,19). Azide inhibition and inactivation of the ⌬NC complex during turnover probably operate by a similar mechanism.