The alpha3beta3gamma subcomplex of the F1-ATPase from the thermophilic bacillus PS3 with the betaT165S substitution does not entrap inhibitory MgADP in a catalytic site during turnover.

The hydrolytic properties of the mutant α3(βT165S)3γ and wild-type α3β3γ subcomplexes of TF1 have been compared. Whereas the wild-type complex hydrolyzes 50 μM ATP in three kinetic phases, the mutant complex hydrolyzes 50 μM ATP with a linear rate. After incubation with a slight excess of ADP in the presence of Mg2+, the wild-type complex hydrolyzes 2 mM ATP with a long lag. In contrast, prior incubation of the mutant complex under these conditions does not affect the kinetics of ATP hydrolysis. The ATPase activity of the wild-type complex is stimulated 4-fold by 0.1% lauryl dimethylamine oxide, whereas this concentration of lauryl dimethylamine oxide inhibits the mutant complex by 25%. Compared with the wild-type complex, the activity of the mutant complex is much less sensitive to turnover-dependent inhibition by azide. This comparison suggests that the mutant complex does not entrap substantial inhibitory MgADP in a catalytic site during turnover, which is supported by the following observations. ATP hydrolysis catalyzed by the wild-type complex is progressively inhibited by increasing concentrations of Mg2+ in the assay medium, whereas the mutant complex is insensitive to increasing concentrations of Mg2+. A Lineweaver-Burk plot constructed from rates of hydrolysis of 20-2000 μM ATP by the wild-type complex is biphasic, exhibiting apparent Km values of 30 μM and 470 μM with corresponding kcat values of 26 and 77 s−1. In contrast, a Lineweaver-Burk plot for the mutant complex is linear in this range of ATP concentration, displaying a Km of 133 μM and a kcat of 360 s−1.

The hydrolytic properties of the mutant ␣ 3 (␤T165S) 3 ␥ and wild-type ␣ 3 ␤ 3 ␥ subcomplexes of TF 1 have been compared. Whereas the wild-type complex hydrolyzes 50 M ATP in three kinetic phases, the mutant complex hydrolyzes 50 M ATP with a linear rate. After incubation with a slight excess of ADP in the presence of Mg 2؉ , the wild-type complex hydrolyzes 2 mM ATP with a long lag. In contrast, prior incubation of the mutant complex under these conditions does not affect the kinetics of ATP hydrolysis. The ATPase activity of the wild-type complex is stimulated 4-fold by 0.1% lauryl dimethylamine oxide, whereas this concentration of lauryl dimethylamine oxide inhibits the mutant complex by 25%. Compared with the wild-type complex, the activity of the mutant complex is much less sensitive to turnoverdependent inhibition by azide. This comparison suggests that the mutant complex does not entrap substantial inhibitory MgADP in a catalytic site during turnover, which is supported by the following observations. ATP hydrolysis catalyzed by the wild-type complex is progressively inhibited by increasing concentrations of Mg 2؉ in the assay medium, whereas the mutant complex is insensitive to increasing concentrations of Mg 2؉  F 0 F 1 -ATP synthases found in energy-transducing membranes couple ATP synthesis and hydrolysis to proton electrochemical gradients (1). They are composed of an integral membrane protein complex, F 0 , which mediates proton conduction, and a peripheral membrane protein complex, F 1 , which bears the catalytic sites. When separated from F 0 as a soluble complex, F 1 is composed of five different subunits in a stoichiometry of ␣ 3 ␤ 3 ␥␦⑀ and functions as an ATPase (2). The ␣ 3 ␤ 3 , ␣ 3 ␤ 3 ␥, and ␣ 3 ␤ 3 ␦ subcomplexes reconstituted from the isolated sub-units of TF 1 1 are active as ATPases (3)(4)(5)(6). Both the ␣ 3 ␤ 3 and ␣ 3 ␤ 3 ␦ subcomplexes differ from TF 1 in that they are less specific for divalent cations and are insensitive to inhibition by azide (4,5). In contrast, the catalytic characteristics of the ␣ 3 ␤ 3 ␥ subcomplex are very similar to those of TF 1 (4,7,8).
Moreover, it has been demonstrated that expression of a plasmid bearing the genes encoding the ␣, ␤, and ␥ subunits of TF 1 in an unc Ϫ strain of Escherichia coli leads to overproduction of the assembled ␣ 3 ␤ 3 ␥ complex, which can be purified in high yield (9). For these reasons, the wild-type and mutant ␣ 3 ␤ 3 ␥ complexes are valuable tools for examining structure-function relationships in F 1 -ATPases.
The F 1 -ATPases contain six nucleotide binding sites. Three of these are potentially catalytic, whereas the others, for want of a defined function, are called noncatalytic nucleotide binding sites (10,11). The 2.8-Å resolution crystal structure of MF 1 shows that catalytic sites are predominantly in ␤ subunits but also contain side chains arising from ␣ subunits. Conversely, the noncatalytic sites are mostly in ␣ subunits, with side chains from the ␤ subunit contributing to them (12). The nucleotide binding domains of the ␣ and ␤ subunits have a common topology and contain the consensus sequence GX 4 GK(T/S), known as the Walker A motif (13) or the P-loop (14). The P-loop is found in the nucleotide binding sites of many proteins including Ras p21, elongation factor-Tu, and the RecA protein (15)(16)(17)(18). The critical role of the P-loop in catalytic sites has been revealed by site-directed mutagenesis of the ␤ subunits of the E. coli and YF 1 -ATPases (19 -21). The sequence of the P-loop in the ␤ subunit of TF 1 is 158 GGAGVGKT 165 .
Kinetic analysis of F 1 -ATPases is complicated by turnoverdependent entrapment of inhibitory MgADP in a catalytic site, which occurs during ATP hydrolysis when noncatalytic sites are not saturated with ATP. Three kinetic phases are observed when MF 1 , TF 1 , and the ␣ 3 ␤ 3 ␥ subcomplex of TF 1 hydrolyze low concentrations of ATP (8,22,23). An initial burst rapidly decelerates to a slow, intermediate rate, which, in turn, progressively accelerates to the final steady-state rate. Transition from the burst to the intermediate rate, which has been well characterized in MF 1 (24,25) and CF 1 (26 -28), is caused by turnover-dependent entrapment of inhibitory MgADP in a catalytic site. Slow binding of ATP to noncatalytic sites relieves inhibition by promoting release of entrapped MgADP from the affected catalytic site. This is responsible for the transition from the intermediate rate to the final steady-state rate (8,22,23).
It has become clear that the apparent K m observed in the millimolar range when F 1 -ATPases hydrolyze 30 -5000 M ATP in the absence of activating anions or activation of the enzymes by prior loading of noncatalytic sites with PP i or ATP does not represent binding of ATP to a low affinity catalytic site but rather represents a rate acceleration caused by binding of ATP to noncatalytic sites (8,22). Therefore, the apparent negative cooperativity in this range of ATP concentration, initially reported for rat liver MF 1 by Ebel and Lardy (29) and later for bovine heart MF 1 (30) and yeast F 1 (31), is an in vitro artifact and is of little, if any, physiological significance.
There have been several reports that K m values of 1-5 M and 100 -200 M are observed when MF 1 hydrolyzes ATP (32)(33)(34)(35). Since substoichiometric ATP binds to a single catalytic site under "unisite" conditions with a K d of less than 1 nM (36), it appears that the K m of 1-5 M represents "bisite" catalysis and that the K m of 100 -200 M represents "trisite" catalysis. However, Boyer (37) has suggested that the K m of 1-5 M is also an artifact associated with entrapment of inhibitory ADP in a catalytic site and contends that the K m of 100 -200 M represents bisite catalysis and that V max rates are attained when two catalytic sites are saturated.
Substitution of the adjacent threonine on the C-terminal side of the lysine in the P-loop of the ␤ subunit of YF 1 (38) or E. coli F 1 (20) leads to enzyme with augmented ATPase activity. This substitution increases the specific activity of YF 1 3-fold, eliminates stimulation by oxyanions, and reduces sensitivity to inhibition induced by azide 25-fold (38,39). Although these characteristics suggest that the mutant YF 1 might not entrap inhibitory MgADP in a catalytic site during turnover, the kinetic analysis used by Mueller et al. (39) was not sufficiently discriminating to establish this with certainty. Rather than looking for turnover-dependent entrapment of inhibitory MgADP, they determined apparent K i values for ADP using a phosphate release assay. The K i values reported by Mueller et al. (39) are based on the effect of increasing, fixed ADP concentrations on the high K m portions of biphasic Lineweaver-Burk plots. However, for reasons stated above, the apparent K m in this concentration range does not represent ATP binding to catalytic sites but rather represents, at least for the wild-type enzyme, ATP binding to noncatalytic sites that promotes dissociation of inhibitory MgADP from a catalytic site (22). Therefore, the K i values obtained represent competition of ADP with ATP for noncatalytic sites.
This investigation was initiated to determine 1) whether the ␣ 3 (␤T165S) 3 ␥ mutant subcomplex of TF 1 does or does not entrap inhibitory MgADP in a catalytic site during turnover using appropriate kinetic analysis and 2) if it does not, whether the mutant subcomplex exhibits a K m of 1-5 M when it hydrolyzes low concentrations of ATP. The results obtained clearly show that the mutant complex does not entrap appreciable MgADP in a catalytic site during turnover and that it exhibits a K m of 1-5 M when it hydrolyzes low concentrations of ATP.

EXPERIMENTAL PROCEDURES
Materials-Biochemicals used in the assays and buffer components were purchased from Sigma. DuPont NEN supplied the [ 3 H]ADP. DCCD and Rhodamine 6G were purchased from Aldrich. LDAO was purchased from Calbiochem.
Purification of the wild-type ␣ 3 ␤ 3 ␥ and mutant ␣ 3 (␤T165S) 3 ␥ complexes was carried out according to Matsui and Yoshida (9). The complexes were stored at 4°C as precipitates in 70% saturated ammonium sulfate. After submitting the ␣ 3 ␤ 3 ␥ and ␣ 3 (␤T165S) 3 ␥ complexes to SDS-polyacrylamide gel electrophoresis, the only protein bands that were revealed on the stained gels corresponded to the ␣, ␤, and ␥ subunits in a 3:3:1 ratio. Unless indicated otherwise, stock solutions of the complexes were prepared by removing ammonium sulfate precipi-tates from mother liquor by centrifugation and dissolving the pellets in either 50 mM Tris-HCl, pH 8.0, containing 0.1 mM EDTA or 50 mM Tris-HCl, pH 7.3, containing 10 M CDTA. The dissolved complexes were then passed through 1-ml centrifuge columns of Sephadex G50, which were equilibrated with the same buffer (40). The wild-type and mutant complexes were stable at least 1 week in these buffers at 4°C. However, they were usually used for experiments on the day of preparation.
Methods-Site-directed mutagenesis was performed as described by Kunkel et al. (41). The oligonucleotide used to introduce the ␤T165S mutation was: 5Ј-GCT CTT GGA TCA GTA CTG ATT TTC CTA CG-3Ј. The oligonucleotide also contained a new site for ScaI, which allowed facile screening for the mutation. The mutation was confirmed by DNA sequencing (42) using the Sequenase 2.0 kit supplied by Amersham.
Protein concentrations were determined by the method of Bradford (43). ATPase activity was determined spectrophotometrically using an ATP regeneration system at 30°C and pH 8.0, as specified previously (44). Unless indicated otherwise, the Mg 2ϩ concentration in the assay medium was 1 mM greater than the ATP concentration.
Binding of [ 3 H]ADP to the wild-type ␣ 3 ␤ 3 ␥ and mutant ␣ 3 (␤T165S) 3 ␥ complexes was performed as described previously (22) with the use of 1-ml centrifuge columns of Sephadex G50 (40) under conditions specified in the legend of Fig. 4. Radioactivity was detected with a Packard 1600 TR counter using Ecoscint from National Diagnostics. final rate of ATP hydrolysis in trace a is 5.4 mol of ATP hydrolyzed/mg/min. In contrast, trace c shows that 50 M ATP is hydrolyzed by 2 g of the mutant complex with a slightly accelerating rate before a final steady-state rate of 30 mol of ATP hydrolyzed/mg/min is attained. After incubation with a 1.4 M excess of ADP plus Mg 2ϩ , the wild-type complex hydrolyzes 50 M ATP with an extended lag, which is illustrated by trace b. After prior incubation with ADP plus Mg 2ϩ , the mutant complex hydrolyzes 50 M ATP with a clearly perceptible lag (trace d), albeit one that is much less pronounced than that observed for the wild-type complex under the same conditions (trace b). Fig. 2 compares hydrolysis of 2 mM ATP by 5.5 g of the wild-type ␣ 3 ␤ 3 ␥ complex or 1.5 g of the mutant complex before and after incubation of each with a 1.4 M excess of ADP in the presence of 1 mM Mg 2ϩ . The effects of azide on hydrolysis of 2 mM ATP by the wild-type and mutant complexes are also compared. Traces a and c of Fig. 2A compare hydrolysis of 2 mM ATP by the wild-type complex before and after incubation with ADP plus Mg 2ϩ . As observed for intact TF 1 (23,45), trace a illustrates that the wild-type ␣ 3 ␤ 3 ␥ complex hydrolyzes 2 mM ATP with a pronounced lag. The final rate of ATP hydrolysis in trace a is 18 mol/mg/min. The lag becomes more pronounced after incubating the wild-type complex with a 1.4 M excess of ADP plus Mg 2ϩ as illustrated by trace c. However, the final rate of ATP hydrolysis in trace c is also 18 mol/mg/min. Trace b of Fig. 2A illustrates inhibition that develops during turnover when the wild-type complex hydrolyzes 2 mM ATP when 1.0 mM NaN 3 is present in the assay medium. Maximal inhibition developed within 3 min after initiating the assay. The final specific activity exhibited in the trace is about 0.45 mol of ATP hydrolyzed/mg/min. After incubating the wild-type complex with a 1.4 M excess of ADP plus Mg 2ϩ and 1.0 mM NaN 3 prior to assay, the rate of ATP hydrolysis is slowed to 0.15 mol/mg/ min, which is illustrated by trace d of Fig. 2A. Fig. 2B shows that the ATPase activity of the mutant ␣ 3 (␤T165S) 3 ␥ complex responds differently than that of the wild-type complex to prior incubation with ADP plus Mg 2ϩ and also shows that the mutant complex is much less sensitive to turnover-dependent inhibition by azide. Comparison of traces a and c of Fig. 2B shows that prior incubation of the mutant complex with ADP plus Mg 2ϩ has very little effect on the kinetics of hydrolysis of 2 mM ATP. Nearly identical lags are exhibited before a final rate of 85 mol of ATP hydrolyzed/mg/ min is attained with or without prior incubation of the mutant complex with ADP plus Mg 2ϩ . Trace b shows that only slight turnover-dependent inhibition develops in the early stage of hydrolysis of 2 mM ATP by the mutant complex in the presence of 1.0 mM NaN 3 . However, it is clear that inhibition slowly develops in the latter part of the trace. Comparison of traces d of Fig. 2, A and B, shows that the mutant complex is inhibited to a much lesser extent than the wild-type complex when it is incubated with a 1.4 M excess of ADP plus Mg 2ϩ in the presence of 1.0 mM NaN 3 prior to assay. The slowly accelerating rate in the time interval illustrated in trace d of Fig. 2B corresponds to about 6 mol of ATP hydrolyzed/mg/min. This is 40-fold greater than that catalyzed by the wild-type complex under the same conditions (trace d of Fig. 2A).

Comparison of the Effects of MgADP and Azide on the Hydrolytic Properties of the Wild-type ␣ 3 ␤ 3 ␥ and the Mutant
Comparison of the Effects of LDAO and Rhodamine 6G on the ATPase Activities of the Wild-type ␣ 3 ␤ 3 ␥ and Mutant ␣ 3 (␤T165S) 3 ␥ Complexes-It has been shown in previous studies that the ATPase activities of TF 1 and the wild-type ␣ 3 ␤ 3 ␥ complex are stimulated about 4-fold by LDAO (7,8). Therefore, the observation that hydrolysis of 2 mM ATP catalyzed by the mutant ␣ 3 (␤T 165 S) 3 ␥ complex is inhibited by about 25% in the presence of 0.06% LDAO was unexpected. This observation prompted a comparison of the dependence of the ATPase activities of the wild-type and mutant complexes on increasing concentrations of LDAO. Fig. 3A shows that the steady-state rate of hydrolysis of 2 mM ATP catalyzed by the wild-type complex (open circles) is progressively augmented by increasing concentrations of the detergent until maximal stimulation of about 4-fold is observed at 0.15% LDAO, in agreement with previously reported results (8). In contrast, LDAO inhibits rather than stimulates the steady-state rate of hydrolysis of 2 mM ATP by the mutant complex (solid circles). Maximal inhibition of about 25% is observed at approximately 0.05% LDAO.
The differential effects of rhodamine 6G on the mutant and wild-type complexes were also examined. It was previously shown that the ATPase activities of TF 1 and the ␣ 3 ␤ 3 ␥ and ␣ 3 ␤ 3 ␥␦ complexes are first stimulated and then inhibited by increasing concentrations of rhodamine 6G. Maximal stimulation is observed at about 10 M rhodamine 6G (7). When the preloaded wild-type complex was diluted into a reaction mixture containing 0.045% LDAO, dissociation of [ 3 H]ADP was greatly accelerated to a rate roughly the same as that observed for the mutant complex (solid circles). Dilution of the preloaded mutant complex into a reaction mixture containing 0.045% LDAO had virtually no effect on the rate of dissociation of [ 3 H]ADP (solid triangles). As previously reported for the wild-type complex (8), dilution of the mutant complex preloaded with Mg[ 3 H]ADP in buffer containing 0.045% LDAO in the absence of ATP did not promote dissociation of [ 3 H]ADP (data not shown).

Comparison of the Effects of Increasing Concentrations of Mg 2ϩ on Hydrolysis of 50 M or 2 mM ATP by the Wild-type ␣ 3 ␤ 3 ␥ and Mutant ␣ 3 (␤T165S) 3 ␥ Complexes-It has been
clearly shown that turnover-dependent inhibition of CF 1 and MF 1 caused by entrapment of inhibitory MgADP in a catalytic site during ATP hydrolysis increases as the concentration of free Mg 2ϩ in the assay medium is raised (46,47). Since the results presented in Figs. 1 and 2 suggest that the mutant complex does not entrap inhibitory MgADP in a catalytic site during turnover, the effects of increasing the concentration of free Mg 2ϩ on the rates of hydrolysis of 50 M and 2 mM ATP by the wild-type and mutant complexes were examined. The results of this comparison are summarized in Table I. The rates were measured between 1.0 and 2.0 min after initiation of the assays. This interval represents the rate of the intermediate phase when the wild-type complex hydrolyzes 50 M ATP. Increasing the concentration of Mg 2ϩ decreases the rate of hydrolysis of 50 M or 2.0 mM ATP by the wild-type complex. In contrast, increased Mg 2ϩ concentration has no effect on the hydrolytic activity of the mutant complex at either concentration of ATP in the assay medium.
Comparison of the Sensitivity of the ATPase Activity of Wildtype ␣ 3 ␤ 3 ␥ and Mutant ␣ 3 (␤T165S) 3 ␥ Complexes to Inactivation by DCCD-Inactivation of TF 1 with DCCD derivatizes Glu ␤190 (48) which is equivalent to Glu ␤188 of MF 1 . In the deduced crystal structure of MF 1 , the carboxylic acid side chain of Glu ␤188 is positioned appropriately to act as the general base that activates the attacking water molecule during ATP hydrolysis (12). The rate of inactivation of TF 1 by DCCD is accelerated 7-fold when ADP binds to a single catalytic site (49), whereas the addition of Mg 2ϩ to form the 1:1:1 TF 1 ⅐ADP⅐Mg complex or the addition of Mg 2ϩ directly to TF 1 decreases the rate of inactivation by DCCD. These opposing effects suggest that reactivity of DCCD with the carboxylate side chain Glu ␤190 is sensitive to conformational changes induced by binding of ligands to a catalytic site. Therefore, it was of interest to compare the inactivation of the wild-type and mutant complexes by DCCD in the presence of ADP, ADP plus Mg 2ϩ , or Mg 2ϩ alone. Table II compares the pseudo-first-order rate con- stants obtained for inactivation of the wild-type and mutant complexes with DCCD in the presence or absence of ADP and/or Mg 2ϩ . The rate of inactivation of the wild-type complex is accelerated 2.4-fold in the presence of ADP. In the presence of ADP plus Mg 2ϩ or of Mg 2ϩ alone, the rate of inactivation of the wild-type complex is decelerated by about 65%. In the absence of effectors, the rate of inactivation of the mutant complex by DCCD is only 50% of that observed for the wild-type complex under identical conditions. The rate of inactivation of the mutant complex is not affected by ADP alone. Furthermore, the rate of inactivation of the mutant complex is decelerated by only 15% of that observed for the wild-type complex in the presence of ADP plus Mg 2ϩ or Mg 2ϩ alone.
Comparison of Lineweaver-Burk Plots Obtained for the ␣ 3 ␤ 3 ␥ and ␣ 3 (␤T165S) 3 ␥ Complexes- Fig. 5A compares Lineweaver-Burk plots obtained from rates of hydrolysis (recorded between 15 and 45 s after initiating assays) of 20 -2000 M ATP catalyzed by the wild-type (solid circles) and mutant (solid squares) complexes. The double reciprocal plot for the wild-type complex is biphasic, displaying apparent negative cooperativity in this range of ATP concentration. Extrapolation of the linear portion of the plot obtained for the wild-type complex in the high range of ATP concentration reveals an apparent K m of 470 M with an associated k cat of 77 s Ϫ1 . An apparent K m of 30 M with an associated k cat of 26 s Ϫ1 is estimated from extrapolation of the phase obtained at low concentrations of ATP. In contrast, the Lineweaver-Burk plot for the mutant complex in this range of ATP concentration is linear and extrapolates to a K m of 133 M and a k cat of 360 s Ϫ1 . Fig. 5B shows that a Lineweaver-Burk plot of rates of hydrolysis (recorded between 15 and 45 s after initiating assays) of 4.0 -2000 M ATP catalyzed by the mutant complex displays negative cooperativity with the transition between high and low affinity sites occurring between 8 and 10 M ATP. DISCUSSION The major conclusion emerging from this study is that the mutant ␣ 3 (␤T165S) 3 ␥ complex, unlike the wild-type ␣ 3 ␤ 3 ␥ complex, does not entrap appreciable inhibitory MgADP in a catalytic site when it hydrolyzes ATP in the presence of a regener- by the wild-type and mutants complexes Desalted stock solutions contained 1 mg/ml of the mutant or wildtype complexes in 50 mM Tris-HCl, pH 8.0. Samples (6.0 l of the wild-type or 1.0 l of the mutant complex) were assayed in 50 mM Hepes-KOH, pH 8.0, which contained 50 M ATP and the concentrations of MgCl 2 indicated or in the same buffer, which contained 2 mM ATP and the concentrations of MgCl 2 indicated using the ATP regeneration system described previously (44). [ to 2 mM. DCCD was then added to a final concentration of 0.2 mM to initiate inactivations. Samples, 1 l each, of the inactivation mixtures were withdrawn at intervals and assayed with 2 mM ATP in the ATP regenerating system described earlier (44). First-order rate constants for the inactivations (k inact were determined from Guggenheim plots (58).

Complex
Additions  5. Comparison of Lineweaver-Burk plots obtained for the the wild-type ␣ 3 ␤ 3 ␥ and mutant ␣ 3 (␤T165S) 3 ␥ complexes. Assays were conducted as described (57)  ating system. Differences in the hydrolytic properties of the mutant and wild-type complexes that support this contention are the following. 1) The wild-type complex hydrolyzes 50 M ATP in three phases; an initial burst rapidly decelerates to a slower, intermediate phase, which, in turn, accelerates to the final steady-state rate. In contrast, the mutant complex hydrolyzes 50 M ATP with a slightly accelerating rate. The final steady-state rate of the mutant complex is 5-fold greater than that observed for the wild-type complex.
2) The rate of hydrolysis of 50 M or 2 mM ATP by the mutant complex is not affected by increasing the concentration of Mg 2ϩ in the assay medium. In contrast, the rate of hydrolysis of 50 M or 2 mM ATP catalyzed by wild-type complex is markedly suppressed by increasing the concentration of Mg 2ϩ in the assay medium.
3) The mutant complex is much less sensitive than the wild-type complex to turnover-dependent inhibition when it hydrolyzes ATP in the presence of azide. 4) A Lineweaver-Burk plot of the initial rates of hydrolysis of 20 -4000 M ATP catalyzed by the mutant complex is linear rather than biphasic as observed for the wild-type enzyme. It has been shown with MF 1 that biphasic Lineweaver-Burk plots for hydrolysis of ATP in this concentration range are caused by slow binding of ATP to noncatalytic sites, which promotes dissociation of inhibitory MgADP from a catalytic site (22).
To explain transient entrapment of inhibitory MgADP in a single catalytic site during ATP hydrolysis by MF 1 , it has been hypothesized that an active (24) or a readily activable (50) F 1 ⅐ADP⅐Mg complex is in equilibrium with an inactive F 1 *⅐ADP⅐Mg complex. It has been proposed that turnover-dependent inhibition by azide is caused by stabilization of the F 1 *⅐ADP⅐Mg complex (51,52). The postulated equilibria are illustrated in Scheme I.
The results presented here are entirely consistent with the proposed equilibria. The kinetic characterization described suggests that, in the case of the ␣ 3 (␤T165S) 3 ␥ complex, the equilibrium between the F 1 ⅐ADP⅐Mg and F 1 *⅐ADP⅐Mg complexes is predominantly in favor of the F 1 ⅐ADP⅐Mg complex. This argument is consistent with the observation that preloading a catalytic site of the mutant complex with MgADP only slightly affects the initial rate of hydrolysis of 50 M ATP and has essentially no effect on the hydrolysis of 2 mM ATP. It is also consistent with the observation that the mutant complex is much less sensitive to turnover-dependent inhibition when assayed in the presence of azide but becomes substantially inhibited when incubated with MgADP in the presence of azide prior to assay. Presumably, this occurs with slow formation of the more stable F 1 *⅐ADP⅐Mg⅐N 3 Ϫ complex at a catalytic site. Milgrom and Murataliev (50) and Bulygin and Vinogradov (53) have reported that the MF1⅐ADP⅐Mg complex slowly isomerizes to the MF 1 *⅐ADP⅐Mg complex. Although the molecular basis for the isomerization remains obscure, results presented here suggest that changes in the conformation of the affected catalytic site might be involved. Unlike the wild-type complex and TF 1 , the rate of inactivation of the ␣ 3 (␤T165S) 3 ␥ complex by DCCD is not accelerated by ADP, and compared with the wild-type complex and TF 1 , it is only slightly decelerated by Mg 2ϩ in the presence or absence of ADP. These observations suggest that the active site of the mutant complex does not undergo the same ligand-induced conformational changes that occur in the wild-type complex. In the crystal structure of MF 1 , the equivalent of Thr ␤165 of TF 1 is liganded to the Mg 2ϩ ions complexed with ADP or AMP-PNP at the catalytic sites designated ␤ DP and ␤ TP , respectively (12). A hypothetical explanation for the turnover-dependent entrapment of MgADP in a catalytic site of wild-type F 1 -ATPases is the following. In the absence of saturation of noncatalytic sites with ATP, MgP i dissociates first at low frequency during ATP hydrolysis leaving ADP bound to a catalytic site. Binding of Mg 2ϩ from the medium to the catalytic site containing bound ADP forms the F 1 ⅐ADP⅐Mg 2ϩ complex that slowly isomerizes to the inactive F 1 *⅐ADP⅐Mg complex.
Although the mutant complex does not entrap substantial inhibitory MgADP in a catalytic site during turnover, a slightly accelerating rate is exhibited when it hydrolyzes 50 M ATP, and a distinct lag is observed when it hydrolyzes 2 mM ATP. To explain this apparent dilemma, we suggest that slow binding of ATP to noncatalytic sites during turnover causes a slight activation, which is independent of promotion of dissociation of inhibitory MgADP from a catalytic site. If this is indeed the case, the accelerations attributed to the binding of ATP to noncatalytic sites of the wild-type complex would reflect dissociation of inhibitory MgADP from a catalytic site plus the acceleration observed for the mutant complex.
The linear Lineweaver-Burk plot exhibited for the initial rate of hydrolysis of 20 -2000 M ATP by the mutant complex is consistent with the finding that it does not entrap appreciable inhibitory MgADP in a catalytic site during turnover. It was shown earlier that the apparent negative cooperativity displayed on Lineweaver-Burk plots for hydrolysis of ATP in this concentration range by MF 1 is caused by slow binding of ATP to noncatalytic sites, which promotes dissociation of inhibitory MgADP from the affected catalytic site (22). The negative cooperativity exhibited on the Lineweaver-Burk plot constructed from the initial rates of hydrolysis of 4 -2000 M ATP catalyzed by the mutant complex is consistent with previously reported results (54). It was shown that an Eadie-Hofstee plot constructed from the initial rates of hydrolysis of 0.  (37). Therefore, we interpret the K m of 1-5 M with an associated k cat of about 20 s Ϫ1 to represent catalysis when only two catalytic sites are saturated with ATP and the K m of about 120 M with an associated k cat of about 350 s Ϫ1 to represent catalysis when three catalytic sites are saturated with substrate. Using a different approach, Weber et al. (55) also conclude that the ␤Y331W mutant of the E. coli F 1 -ATPase attains maximal velocity when three catalytic sites are saturated with ATP. It should be pointed out that in their correlation of quenching of tryptophan fluorescence with ATPase activity as ATP concentration was increased, Weber et al. (55) failed to detect significant "bisite catalysis." However, more recently, using the same method to correlate catalytic site occupancy with TNP-ATP hydrolysis, Weber and Senior (56) reported that the V max for TNP-ATP hydrolysis under bisite conditions is about 40% of the V max observed under trisite conditions.