INTRODUCTION

Chloroplast ATP synthase, CF₁CF₀, couples the phosphorylation of ADP to ATP to electron transfer at the expense of the electrochemical proton gradient across the thylakoid membrane. The catalytic portion of the complex, CF₁,¹ contains the nucleotide binding sites. In solution, CF₁ catalyzes a low rate of ATP hydrolysis, but removal of the inhibitory  subunit increases the activity to a high level (1).

CF₁ likely contains six nucleotide binding sites (2), not all of which are catalytic (2). These binding sites have been partially characterized (3, 4). Two sites are dissociable. There are two noncatalytic sites that are tight binding only for ATP and only in the presence of Mg²⁺, while the two remaining sites will also bind ADP tightly and do not require Mg²⁺ for tight binding. ``Tight'' binding is used as an operational definition meaning the bound nucleotide is not removed by gel filtration or dialysis. The dissociable sites are termed ``loose.'' None of these sites binds AMP or adenosine.

Alternation of site properties between loose and tight has been observed and appears to be induced by nucleotide binding (5). One proposed model for the mechanism of catalysis (6) involves at least two catalytic sites that alternate properties. The binding of nucleotide to a loose site causes it to become a tight binding site, whereas a site containing tightly bound nucleotide becomes dissociable. Catalysis occurs when the substrate is tightly bound. Of the two sites that bind ADP tightly, one will exchange its bound ADP easily for medium TNP-ADP, while the other will not (7).

Until now, studies of both the activity of CF₁ and the exchange of tightly bound ADP or TNP-ADP from the site that exchanges easily have been done with ADP present in the tight binding site that does not exchange easily (7). The depletion of the endogenous nucleotide from CF₁ would help to elucidate the function of tightly bound nucleotide and other aspects of the mechanism of catalysis. Nucleotide depletion of the CF₁ homologues in Escherichia coli (8) and in beef heart mitochondria (9) has been achieved. These studies have shown no change in the catalytic activity of the enzyme complex following depletion. Mitochondrial F₁ was found to be less stable after nucleotide depletion (9). Nucleotide-depleted CF₁ has not been prepared previously. Procedures used to deplete the bound nucleotide contents of bacterial or mitochondrial F₁, including gel filtration of CF₁ in the presence of 50% glycerol (9), were not effective in removing ADP from CF₁.

This paper describes a reproducible method for nucleotide depletion of CF₁ that has little effect on either ATPase activity or the stability of the enzyme. The initial binding of medium nucleotide to catalytic sites has been examined in control and nucleotide-depleted CF₁. Also, the effects of tightly bound ADP on the exchange of TNP-ADP at a separate site and on Mg²⁺ inhibition of ATP hydrolysis were determined. The rate of the initial exchange of tightly bound TNP-ADP was compared with the change in the ATPase activity of CF₁-, and the rate of increase of the activity with time was found to correspond to the amount of TNP-ADP that was released. The presence of a comparable rate of increase in activity with time in the nucleotide-depleted enzyme demonstrated that both exchange and the activity are induced by some separate event. These results help to clarify the catalytic mechanism and the interactions between nucleotide binding sites in the enzyme.

EXPERIMENTAL PROCEDURES

Reduced CF₁- was prepared from market spinach by the procedure described in Shapiro and McCarty (5), with modifications described by Soteropoulos et al. (10) and Digel and McCarty (7). CF₁- depleted of its endogenous nucleotide (CF₁--NT) was prepared as follows. Before use, CF₁- was desalted by passing it consecutively through two, 3-ml Sephadex G-50 columns (11) equilibrated with 50 mM Tris-HCl, pH 8.0, 50 mM NaCl, 10 mM EDTA. Incubation of CF₁- in 10 mM EDTA for 4 h at room temperature facilitates nucleotide exchange (7). EDTA inhibits alkaline phosphatase activity and must be removed from solution by passage through two consecutive 3-ml Sephadex G-50 columns equilibrated with 50 mM Tris-HCl, pH 8.0, 50 mM NaCl (TN buffer). This and all subsequent buffers were passed through a column of Chelex 100 resin to remove residual Mg²⁺ and were stored in plastic.

CF₁- concentration was determined by the Lowry method (12) or by absorbance, using an extinction coefficient of  = 0.483 cm²/mg of CF₁ at 277 nm (3). CF₁- (15 mg) was diluted to 5 mg/ml with TN buffer, and 280 units of Stratagene calf intestinal alkaline phosphatase (catalog number 600015) were added. Samples were incubated 18-20 h at room temperature. Alkaline phosphatase was removed from the CF₁- by DEAE-cellulose chromatography at room temperature. Alkaline phosphatase was eluted from the column (0.1-ml column volume/mg of CF₁ used) with 15 ml of 50 mM Tris-HCl, 50 mM NaCl, followed by elution with 45 ml of 50 mM Tris-HCl, 100 mM NaCl. Nucleotide-depleted CF₁- was eluted by 5 ml of 50 mM Tris-HCl, 400 mM NaCl and was collected in 0.5-ml fractions.

The CF₁- fractions were assayed for alkaline phosphatase activity by following p-nitrophenyl phosphate hydrolysis at 410 nm at 25 °C. The fractions that contained most of the CF₁- were then combined and desalted by two consecutive Sephadex G-50 columns equilibrated with TN buffer. The phosphatase activity of the CF₁--NT was 1 nmol min¹ mg CF₁¹ or less.

ATPase activity of CF₁- in the presence of Ca²⁺ was assayed by the incubation of 10 µg of CF₁- for 3-5 min at 37 °C in 50 mM Tris-HCl, 5 mM ATP, and 5 mM CaCl₂. The amount of P_i produced was determined colorimetrically (13). ATPase activity of CF₁- in the presence of Mg²⁺ was measured with a coupled enzyme assay at 25 °C. CF₁ (50 µg) was added to a mixture of 1 ml total volume containing 50 mM Tris-HCl, pH 8.0, 50 mM KCl, 1 mM MgCl₂, 1 mM phospho(enol)pyruvate, 5 mM ATP, 20 units each of Sigma lactate dehydrogenase (catalog number L-1254), Sigma pyruvate kinase (catalog number P-9136), and 0.25 mM NADH. The decrease in NADH concentration was measured spectrophotometrically at 340 nm.

The cold stability of CF₁--NT was monitored by loss of ATPase activity upon incubation at 0 °C in the manner of Hightower and McCarty (14). CF₁--NT was incubated at 15 µg/ml in Chelex 100-treated 50 mM Tris-HCl, pH 8.0, at 0 and 25 °C. At various times, 200-µl aliquots were taken for the assay of Ca²⁺- ATPase activity for 3 min at 25 °C. The concentration of P_i produced was determined by the sensitive malachite green assay (15) with modifications described elsewhere (14).

The analysis of nucleotide bound to CF₁ was performed with ion-pairing high pressure liquid chromatography (HPLC). Bound nucleotides were released from CF₁ by methanol precipitation (16). The relationship between the integrated peak area and the nucleotide quantity was determined for each experiment using nucleotide standards of known concentration.

CF₁- and CF₁--NT were loaded with TNP-ADP by incubation in 0.5 mM TNP-ADP for 40 min. Excess and loosely bound TNP-ADP were removed by passage of the enzyme through two consecutive Sephadex G-50 columns. The extent of TNP-ADP loading was determined by the absorbance of tightly bound TNP-ADP at 418 nm using an extinction coefficient of 2.51 × 10⁴ mol¹ cm¹ (17) and blanked against CF₁- at the same concentration that had not been loaded with TNP-ADP.

Binding of TNP-ADP as well as exchange of tightly bound TNP-ADP were observed by measuring the TNP-ADP fluorescence. The fluorescence of TNP-ADP has been shown to increase linearly with the amount that is bound to CF₁- (10). Fluorescence measurements were made using an OLIS-modified SLM/Aminco SPF-500 spectrofluorometer. The excitation wavelength was 418 nm and the emission wavelength 560 nm. Exchange was initiated by the addition of free nucleotide into the stirred sample medium. In one experiment, exchange was initiated by stopped-flow mixing.

RESULTS

As prepared, CF₁-, stored as an (NH₄)₂SO₄ precipitate in the presence of EDTA, contains 1.2-2 mol of tightly bound ADP/mol of CF₁. Tightly bound ADP is released into the medium at a slow rate and is re-bound at a fast rate (10). Alkaline phosphatase cleaves ADP to AMP and then to adenosine. Neither AMP nor adenosine binds to CF₁. In principle, therefore, alkaline phosphatase should promote the removal of bound ADP from CF₁, simply by pulling the equilibrium toward the dissociation of bound nucleotide. Although the mechanism of alkaline phosphatase-induced release of bound ADP has not been investigated in detail, results of preliminary experiments suggest that alkaline phosphatase attacks ADP that had dissociated from CF₁. Alkaline phosphatase added to the outside of a dialysis bag containing CF₁ enhanced ADP removal (data not shown). This method has been routinely effective in producing nucleotide-depleted CF₁ with an ADP content determined by HPLC (Fig. 1) over 11 trials to be 0.094 ± 0.04 mol/mol of CF₁- compared with an untreated control content of 1.2 ± 0.2 mol of ADP/mol of CF₁-. In more recent trials, the CF₁- used had a tightly bound ADP content in the range of 1.6-1.8 mol of ADP/mol of CF₁- and the bound nucleotide content after depletion was correspondingly higher, in the range 0.18-0.32 mol/mol of CF₁-.

-.

The nucleotide-depleted enzyme was stored in 50 mM Tris-HCl, pH 8.0, with and without 20% glycerol, with and without 50% saturated ammonium sulfate, at 25 or 4 °C and, for samples without ammonium sulfate, at 80 °C. Overnight storage in 20% glycerol proved most conducive to retaining Ca²⁺ATPase activity, with storage at 25 °C and 80 °C equally effective. CF₁--NT fully retains its Ca²⁺ATPase activity for up to 3 weeks in 50 mM Tris-HCl, 20% glycerol at 80 °C.

Nucleotide depletion appears to have no effect on the steady-state rate of ATP hydrolysis activity of the enzyme. The ATPase activity of CF₁--NT was equivalent to that of CF₁- with endogenous nucleotide (Table I). p-Nitrophenylphosphatase assays revealed no significant residual alkaline phosphatase activity in the CF₁--NT. To ensure that the ATPase activity measured was that of CF₁- and not of residual alkaline phosphatase, Ca²⁺ATPase assays were also performed in the presence of tentoxin, a potent inhibitor of the ATPase activity of CF₁ but not of alkaline phosphatase. The ATPase activity of CF₁--NT was fully inhibited by µM concentrations of tentoxin.

Table I. Ca2+ATPase activity of nucleotide-depleted CF1- The Ca2+ATPase activity at 37 °C of CF1- and CF1--NT was averaged over 10 different preparations, and the activity of the same samples was assayed in the presence of 1 µM tentoxin to demonstrate the lack of alkaline phosphatase activity. Enzyme preparation Specific activity µmol/min/mg CF1- 14.7  ± 5.1 CF1--NT 16.1  ± 5.2 CF1- + tentoxin (1 µM) 0 CF1--NT + tentoxin (1 µM) 0

Nucleotide depletion also has no effect on the cold stability of the enzyme. CF₁ loses its ATPase activity upon exposure to low temperatures (0 °C), a result of at least a partial dissociation of the complex (18). Nucleotides in the medium protect from the loss of ATPase activity in the cold (18, 19). Nucleotide-depleted CF₁- lost its ATPase activity at 0 °C at a rate identical to that for CF₁- (Fig. 2). The CF₁- used contained 1.8 mol of tightly bound ADP/mol of CF₁-, and the nucleotide-depleted enzyme contained 0.20 mol of ADP tightly bound/mol of CF₁-.

- by cold inactivation.

1

1

1

1

Measurements were taken of the effect of bound ADP on the loading of the enzyme with TNP-ADP. CF₁--NT containing 0.05 mol of bound nucleotide/mol of CF₁--NT was prepared in TN buffer. Half was loaded with ADP by incubation in 5 mM ADP for 1 h. Excess ADP was removed by passage through two consecutive Sephadex G-50 centrifuge columns. TNP-ADP binding to both the nucleotide-depleted and the ADP-loaded enzyme was measured as the increase in fluorescence of TNP-ADP as it binds to the enzyme in solution (Fig. 3).

- and CF

--NT.

When 1 µM TNP-ADP was added to 0.5 µM CF₁--NT essentially all of the TNP-ADP bound to the enzyme resulting in 1.95 mol of bound TNP-ADP/mol of CF₁-. The binding was complete within 1 min. The ADP-loaded enzyme in contrast bound 1.1 mol of TNP-ADP/mol of CF₁-, at these concentrations, and binding took close to 10 min to complete. When the concentration of TNP-ADP was increased to 4 µM with 0.5 µM CF₁-, the nucleotide-depleted enzyme bound 2.5 mol of TNP-ADP/mol of CF₁-, and the ADP-loaded enzyme bound 1.6 mol of TNP-ADP/mol of CF₁- (Fig. 3A). These data suggest that a nonexchanging or slowly exchanging tight ADP binding site exists, although it is possible that the site may be made to exchange at higher concentrations of TNP-ADP.

TNP-ADP binding was studied using stopped-flow fluorescence in order to resolve the phases of binding to CF₁--NT. In this case, either CF₁--NT containing 0.20 mol of tightly bound ADP/mol of CF₁- or CF₁- which contained an endogenous 1.81 mol of tightly bound ADP/CF₁--NT were mixed in the stop-flow with TNP-ADP in TN buffer to a final concentration of 0.4 µM CF₁- and 1.6 µM TNP-ADP (Fig. 3B). The resulting scans of fluorescence intensity versus time showed two distinct phases and were fit to Equation 1:

(Eq. 1)

where F(t) is the fluorescence intensity at time t, F₁ and F₂ are the extent of the two phases, and k₁ and k₂ are their rate constants. F₀ is the base-line fluorescence of the unlabeled sample with the unbound TNP-ADP.

As can be seen in Table II, in the case of binding to CF₁- that contains endogenous nucleotide, there is one rapid phase, the extent of which corresponds well to the filling of the 0.19 mol of empty ADP tightly bound site/mol of CF₁-, and a second, much slower phase whose rate constant is a good match for those observed for exchange of tightly bound TNP-ADP (7). In the nucleotide-depleted sample, at these concentrations, a total of 1.4 sites are filled per CF₁-, and two rate constants are observed, one equivalent to the fast rate constant from the sample with endogenous ADP, and one 9 times faster. Close to one site filled at the fast rate, with the remaining 0.5 nucleotide binding at the slower rate.

Table II. The rate and extent of TNP-ADP binding to CF1- and CF1--NT Numbers represent the rate and extent of TNP-ADP binding to CF1--NT and CF1- as defined by Equation 1, when TNP-ADP or TNP-ADP together with MgCl2 are added to a final concentration of 1.6 µM in 0.4 µM enzyme.   k1   k2 s1 s1 CF1- + TNP-ADP 0.11 0.10 0.41 0.0039 CF1- + Mg2+, TNP-ADP 0.18 0.10 0.35 0.0045 CF1--NT + TNP-ADP 0.52 0.11 0.88 0.93 CF1--NT + Mg2+, TNP-ADP 0.36 0.06 1.12 0.53

The measurements were repeated, mixing the enzyme solution with a mixture of MgCl₂ and TNP-ADP for a final concentration of 1.6 µM for each. The sample of CF₁- with its endogenous nucleotide showed an increase in its slower rate constant. The rate of exchange of bound nucleotide has been shown to be faster for medium MgADP then for ADP alone (7), a further indication that the slow rate constant represents exchange of bound ADP for medium TNP-ADP. There was a drop of about half for both rate constants for the nucleotide-depleted sample showing that binding of MgTNP-ADP to either empty site is slower than binding of TNP-ADP without Mg²⁺.

It has previously been suggested that both of the tight ADP binding sites will exchange for medium ADP in the absence of Mg²⁺ (4). In order to test this hypothesis, CF₁--NT containing 0.05 mol of ADP/mol of CF₁- was loaded with TNP-ADP. This resulted in CF₁- with 1.44 mol of TNP-ADP tightly bound/mol of CF₁-. Exchange of the bound TNP-ADP was observed in the presence of 5 mM ADP and 5 mM EDTA for more than 10 min and fit accurately to Equation 2:

(Eq. 2)

Here F₁ and F₂ are the extents of a biphasic decay, and k₁ and k₂ are the rate constants, with F_min = F(t = ). The extent of exchange was calculated from the fit, and from a base line obtained by addition of 0.1% SDS to remove all bound nucleotides. Of 1.44 tightly bound TNP-ADP, 1.1 would exchange, suggesting that if the second tight ADP site does exchange, it does so only very slowly.

Mg²⁺ inhibits the ATPase activity of CF₁ and is believed to inhibit through the formation of a tightly bound Mg²⁺ADP complex (20, 21). In order to determine whether ADP was a requirement of Mg²⁺-induced inhibition, the activity as a function of time of CF₁--NT was observed at 25 °C by the coupled enzyme assay, with and without prior exposure of the enzyme to Mg²⁺. The CF₁--NT contained 0.18 mol of tightly bound nucleotide/mol of CF₁-, and the control was CF₁- which contained 1.6 mol of tightly bound nucleotide/mol of CF₁-. The activity was observed by monitoring the absorbance of NADH at 340 nm with respect to time, with the rate of change of the absorbance with respect to time giving the activity of the enzyme. The activity as a function of time fit to the empirical Equation 3:

(Eq. 3)

where A and k₁, are the extent and the rate constant, respectively, of the initial increase in activity; k₂ is the rate constant for a slow inhibition that underlies the first term, and A₀ is the activity at t = 0, when the enzyme was added to the reaction mixture. In order to avoid the error introduced by numerical differentiation, the absorbance scans were fit to the integral of Equation 3, and the activity as a function of time was calculated from the fit. Excellent fits were obtainable by this method.

As can be seen from the plots of activity versus time (Fig. 4A), incubation of the nucleotide-depleted enzyme with Mg²⁺ prior to the initiation of activity had relatively little effect, particularly in contrast to the enzyme containing its endogenous ADP. More can be learned from the rate constants presented in Table III. There is only slight variation in the rate constant of the initial increase in activity; however, the activity has a lower maximum value in the Mg²⁺-pretreated case, and the rate k₂ of the underlying inhibition was much less. In contrast, in the sample of CF₁- that retained its endogenous ADP, incubation with Mg²⁺ caused a large decrease in k₁. In this case, the sample pretreated with Mg²⁺ showed no competing inhibitory term.

--NT and CF

- with and without prior Mg

Table III. Rate constants for the change in activity with time The rate constants as defined by Equation 3 describing the change with time of the ATPase activity of CF1--NT and CF1- with and without incubation of the enzyme in 1 mM MgCl2 for 1 h prior to initiation of exchange by addition of the enzyme to the reaction mixture. k1 has units of min1, and k2 has unit of µmol of Pi produced per min2/mg CF1-. CF1--NT CF1--NT + Mg CF1- CF1- + Mg k1 6.22E-01 5.94E-01 8.85E-01 1.48E-02 k2 1.04E-03 9.51E-06 4.21E-03  4.40E-03 TNP-ADP loaded k1 5.07E-01 5.32E-01 6.17E-01 5.95E-02 k2 8.89E-04 1.47E-04 1.70E-03  1.95E-03

Inhibition of the rate constant k₁ by Mg²⁺ pretreatment is clearly dependent on the presence of tightly bound ADP. The inhibitory term k₂ is also affected by Mg²⁺, but the dependence of k₂ on ADP is less clear. The lower maximum activity and extremely low k₂ of the nucleotide-depleted sample when pretreated with Mg²⁺ suggests that this inhibitory process is mostly complete in this case relative to the sample without Mg²⁺ pretreatment. This difference in k₂ values in the nucleotide-depleted sample may result in some fashion from the remaining 0.18 mol of ADP/mol of enzyme or it may be evidence of an interaction between the enzyme and Mg²⁺ in solution that does not require the presence of bound nucleotide.

It is interesting to note that the activity of CF₁- that retained its endogenous ADP reaches a higher maximum activity then that of the nucleotide-depleted sample before falling, relatively rapidly, to the same level. The sample of CF₁- had been stored as a precipitate in 50% saturated ammonium sulfate with 5 mM EDTA for several days prior to use. This procedure has been shown to produce CF₁- effectively free of bound Mg²⁺ (7). These data suggest that tightly bound ADP actually transiently enhances the activity in the absence of bound Mg²⁺. In order to ensure that the lag in activity described by the rate constant k₁ was not the result of the coupling enzymes rather than the CF₁, ADP was added directly to the reaction mix in the absence of CF₁. No such lag was observed.

CF₁- from the same batch as the preceding experiment was loaded with ADP by incubation for 1 h in 5 mM ADP with an additional 5 mM EDTA to prevent Mg²⁺ binding. Excess and loosely bound ADP were removed by passage through two consecutive Sephadex G-50 centrifuge columns. Both it and the CF₁--NT (0.18 mol of ADP/mol of CF₁-) were loaded with TNP-ADP by incubation with a substoichiometric amount of TNP-ADP, with excess and loosely bound TNP-ADP removed by more Sephadex G-50 centrifuge columns. This resulted in a sample with 0.86 mol of TNP-ADP and 0.18 mol of ADP/mol of CF₁-, and one with 0.55 mol of TNP-ADP and 1.36 mol of ADP/CF₁-. The activities were measured as before both with and without Mg²⁺ pretreatment (Table III and Fig. 4B). Despite the presence of tightly bound TNP-ADP, there is remarkably little difference from the previous measurements. The exchange of the tightly bound TNP-ADP for medium nucleotide was observed in both samples, and 90% of the bound TNP-ADP exchanged showing that the TNP-ADP was bound mainly in the easily exchanging ADP tight-binding site. Thus, it must be the second ADP tight-binding site, the one which exchanges slowly if at all, that is responsible for the Mg²⁺-induced delay in the onset of ATPase activity. The fact that the easily exchanging site fills fastest is evidence that the fast phase of binding of TNP-ADP to CF₁--NT that was observed represents binding to the easily exchanging site.

When tightly bound ADP was removed from CF₁-, preincubation of the enzyme with Mg²⁺ no longer inhibited the rate of ATP hydrolysis. Whether tightly bound ADP had a similar effect on Mg²⁺ inhibition of the exchange of bound nucleotide for medium nucleotide was determined. CF₁--NT and CF₁- were loaded with substoichiometric amounts of TNP-ADP, and the initial rate of exchange (v₀) of tightly bound TNP-ADP for medium nucleotide was measured (Table IV). v₀ is determined by fitting the TNP-ADP fluorescence to Equation 2 and is described by the apparent first order rate constant for exchange multiplied by the extent of exchange. Exchange under catalytic conditions (both Mg²⁺ and ATP present) was triphasic and fit to a suitably expanded version of Equation 2. For the purpose of comparison v₀ values were scaled by the amount TNP-ADP bound to each sample when calculating the ratio of v₀ for the two samples.

Table IV. The effect of tightly bound ADP on the initial rates of exchange of bound TNP-ADP for medium nucleotide The units of v0 are nmol of TNP-ADP exchanged per min/mg Cf1-. The samples with prior Mg2+ incubation were incubated ~1 h in 5 mM MgCl2 prior to the addition of ADP. The samples with prior EDTA incubation were incubated ~2 h in 5 mM EDTA prior to the addition of ADP. Exchange initiated with v0 (CF1--NT + TNP-ADP) v0 (CF1-)/v0  (CF1--NT) 5 mM ADP 0.87 1.03 5 mM Mg2+, 5 mM ADP 4.9 0.20 5 mM Mg2+, 25 mM Pi, 5 mM ADP 11.7 0.51 5 mM ADP (prior Mg2+ incubation) 0.54 0.70 5 mM ADP (prior EDTA incubation) 1.2 1.00 5 mM ATP 1.54 0.81 5 mM ATP (prior Mg2+ incubation) 1.05 0.27 5 mM Mg2+, 5 mM ATP, 50 mM sulfite 32.96 0.21 5 mM ATP (prior EDTA incubation) 1.28 1.05

CF₁- and CF₁--NT were prepared in TN buffer treated with Chelex 100 resin and loaded with TNP-ADP by incubation in TN + 5 mM EDTA + TNP-ADP, where the TNP-ADP concentration was slightly less than the enzyme concentration as determined by absorbance at 277 nm. Thus loaded, the CF₁- contained 1.09 mol of ADP and 0.85 mol of TNP-ADP per mol of CF₁-. The nucleotide-depleted enzyme started with 0.16 mol of nucleotide/mol of CF₁--NT and after TNP-ADP loading contained 1.03 mol of TNP-ADP/mol of CF₁-. Samples used to study exchange with medium ATP rather than ADP were the same TNP-ADP-loaded samples used for the measurements of activity as a function of time. The fraction of tightly bound TNP-ADP that would exchange was the same for sample with and without endogenous nucleotide when initiated with the same nucleotide mix, typically around 80-90%, again showing that TNP-ADP was bound mainly at the same exchangeable site in each sample.

Tightly bound ADP affected exchange rates of bound TNP-ADP only in the presence of Mg²⁺. The variations produced in exchange rates by Mg²⁺, EDTA, sulfite, and P_i are comparable with what has been observed previously, with MgADP causing faster exchange than ADP, and MgADP + P_i, or MgATP and sulfite being faster still (7). It is interesting to note that Mg²⁺ incubation reduces the rate of exchange of the nucleotide-depleted samples both for medium ADP and ATP. It was suggested by the activity of Mg²⁺-incubated CF₁--NT that Mg²⁺ is coordinated with the enzyme prior to ADP addition. If so, the Mg²⁺ coordination combined with the very rapid binding of ADP shown in Fig. 3 may result in rapid inhibition of the Mg²⁺-pretreated enzyme in the presence of ADP. However, these samples still contained 0.18 or 0.26 mol of ADP/mol of CF₁- which may be enough by itself to account for the decrease.

The ability to measure both activity and exchange as a function of time provides an opportunity to compare them directly. CF₁- was prepared with 0.9 mol of ADP and 1.1 mol of TNP-ADP/mol of CF₁-, and from CF₁--NT a sample was prepared with 0.27 mol of nucleotide with an additional 1.3 mol of TNP-ADP/mol of CF₁-. The activity of these samples was monitored using the coupled enzyme assay at 25 °C, and the exchange of tightly bound TNP-ADP was observed by fluorescence using the same assay mix and sample volumes and concentrations as the activity assay also done at 25 °C. The only exception was that the exchange measurements were done without NADH, the fluorescence of which interferes with the determination of TNP-ADP fluorescence.

The exchange of TNP-ADP under these circumstances had three phases, one of which was faster than that which could be observed in the coupled enzyme assay, where the sample cuvette had to be mixed by inversion prior to measurement. The extent of this rapid phase as a fraction of the total observed exchange corresponds well to the earliest observed activities as a fraction of the maximum activity (A₀ and the fast phase extent of Table V). This suggests that there may be a faster, unobserved phase to the increase in the ATPase activity with time and that the actual initial activity of the enzyme is much closer to zero. The maximum rate of exchange observed of the bound TNP-ADP is the initial rate (v₀), as only the very first exchange turnover is observed. This rate is much less than the maximum ATPase activity of the enzyme, and even in the case of the TNP-ADP-loaded sample of CF₁--NT, it is only 10% of the initially observed activity (Table V).

Table V. Comparison between exchange and the increase in activity as a function of time CF1--NT and CF1- refer to the TNP-ADP-loaded enzyme. Amax has units of nmol of Pi/min/mg CF1- and v0 units of nmol of TNP-ADP exchanged per min/mg CF1-. The half-time of the fast phase of exchange was 4 s for the sample of CF1--NT and 11 s for the sample of CF1-. CF1--NT CF1- Activity Exchange Activity Exchange A0/Amax Extent, fast phase A0/Amax Extent, fast phase 0.37 0.39 0.19 0.17 Amax v0 Amax v0 290 11 330 2.2

A discrepancy between the rate of exchange of tightly bound Mg[³H]ADP and the ATPase activity has been observed before in membrane bound CF₁ (22). It was then shown that under the conditions where this exchange was observed, the initial ATPase activity was much less than the steady-state activity and that the steady-state activity was reached only after much of the tightly bound Mg[³H]ADP was released (23). Using a quenched-flow apparatus it was shown that when CF₁ reaches a steady-state ATPase activity, then the exchange rate also reaches a steady state that is comparable with the rate of ATP hydrolysis (24). As measured here it can be seen (Fig. 5) that the loss of tightly bound ADP corresponds well to the increase in ATPase activity with time. The maximum activity was considered to be the sum of A₀ and A of Equation 3. The actual maximum activity was slightly less owing to the underlying inhibition. The fact that the activity reaches its maximum when the exchange is only about 85% complete may be a result of the maximum activity being higher than the eventual steady-state activity.

DISCUSSION

Our results show that CF₁- with markedly decreased bound nucleotide may be obtained by treatment of the enzyme with alkaline phosphatase for at least 18 h. This nucleotide-depleted enzyme had rates of Ca²⁺- and Mg²⁺ATPase activity (Table I and Fig. 3) equivalent to those of CF₁- with endogenous ADP. These findings are consistent with the results obtained from nucleotide-depleted E. coli and mitochondrial F₁. The initial presence of tightly bound nucleotide was not required for catalysis in E. coli F₁, and mitochondrial F₁ exhibited no loss of ATPase activity following depletion of tightly bound nucleotide. However, bound nucleotide appears to stabilize mitochondrial F₁, as the presence of 50% glycerol is required throughout the depletion and further procedures to prevent protein denaturation (9). Our depletion procedure as well as all our assays are conducted in the absence of glycerol. Our cold inactivation data (Fig. 2) demonstrated that the absence of bound nucleotide has no significant effect on the structural stability of CF₁-. Alkaline phosphatase treatment of mitochondrial or bacterial F₁ may be an effective way in which to deplete these enzymes of bound nucleotide. If so, this method for gentle removal of bound nucleotide would be preferred over the more laborious chromatographic methods.

In mitochondrial F₁, Mg²⁺ forms a complex with bound ADP that inhibits catalytic activity (25). This has been suspected to be the case for CF₁-. Due to the unavailability of nucleotide-depleted CF₁-, it has not been proven conclusively that Mg²⁺ inhibits ATPase activity by forming a complex with bound ADP. Although Mg²⁺ binding to CF₁ has been shown to inhibit its catalytic activity (26), the necessity of bound ADP had not previously been demonstrated. Our studies of Mg²⁺ inhibition of ATPase activity (Fig. 3) of CF₁ demonstrate that the neither the presence of ADP nor of Mg²⁺ alone is sufficient to inhibit. The presence of Mg²⁺ and bound ADP combined is responsible for the observed inhibition.

Earlier studies of the exchange of tightly bound ADP with nucleotide in the medium (27) concluded that CF₁ contains one tight binding site for ADP that exchanges and whose exchange is inhibited by Mg²⁺. This site was considered to be the location of the endogenous ADP found in freshly prepared CF₁. More recent studies have shown that there are two tight ADP binding sites (4). It now seems clear that the two tight binding sites for ADP have quite different properties. One will exchange easily for medium nucleotide and the second exchanges very slowly, if at all. This is shown in two ways. CF₁- that contained endogenous tightly bound ADP would not exchange about 1 of the bound ADP for TNP-ADP, and CF₁- that contained 1.44 tightly bound TNP-ADP, but no bound ADP, exchanged only about 1 of the bound TNP-ADP. A similar result has been seen in rat liver mitochondrial F₁ that has been shown to contain two ADP binding sites, one of which exchanges with medium nucleotide and the second of which does not (28).

It is this nonexchanging, or poorly exchanging, site that appears to enable Mg²⁺ to inhibit the exchange of bound nucleotide from the exchangeable site, as well as being the principle source of Mg²⁺ inhibition of the ATPase activity. It has been suggested before (29) that an ADP binding site separate from the catalytic site is responsible for Mg²⁺-induced inhibition of the ATPase activity, although it has not yet been demonstrated that the poorly exchanging site is not catalytic. In all cases in which Mg²⁺ was present in the exchange medium, the rate of exchange of tightly bound TNP-ADP was inhibited in samples with bound ADP relative to those with little bound ADP. It is even possible that the nonexchanging ADP site is the only site at which a bound MgADP complex inhibits exchange, although this is not yet clear due both to the residual bound ADP in the preparation and the ability of TNP-ADP to bind to the nonexchanging site when it is vacant.

The exchange of TNP-ADP from the tight, easily exchanging ADP site was seen to correlate with an increase in the ATPase activity of the enzyme. The data presented in Fig. 5 and Table V might then suggest that it is the exchange of tightly bound ADP from the easily exchanging ADP tight binding site that activates the enzyme. However, it has also been seen (Fig. 4 and Table III) that nucleotide-depleted CF₁-, which has no bound nucleotide to exchange, shows the same increase in activity with time. Therefore, it must be that some separate occurrence activates the enzyme, stimulating the exchange of the tightly bound TNP-ADP at the same time as the ATPase activity. Since the activity is initiated by the addition of Mg²⁺ and ATP to the enzyme, it seems likely that the filling of an additional nucleotide binding site is responsible. CF₁- that has both of its tight ADP sites filled still shows this increase in activity over time. Thus, it would be one of the remaining sites, either the noncatalytic MgATP binding sites or the loose sites. It was shown by Milgrom et al. (30) that an increase in the ATPase activity of CF₁ is correlated with an increase in MgATP bound to the enzyme. It is also possible that the first turnover of the enzyme results in some conformational change that makes the succeeding turnovers faster.