Studies on the Mechanism of Oxidative Phosphorylation CATALYTIC SITE COOPERATIVITY IN ATP SYNTHESIS*

phosphorylation catalyzed by bovine heart

Fl-ATPase from bovine heart mitochondria is composed of five unlike subunits with the stoichiometric ratio of a s P a~& (1,2). The / 3 subunits alone or in combination with the a subunits carry the catalytic sites of the enzyme (1, 2). In a system containing an F1:ATP molar ratio of 3, Penefsky and co-workers (3,4) were able to demonstrate that a single catalytic site of the isolated Fl-ATPase is capable of ATP hydrolysis, albeit at an extremely slow rate ( W 4 s-'). At physiological ATP concentrations, F1 exhibits negative cooperativity with respect to [ATP] (3)(4)(5)(6)(7), and positive catalytic cooperativity in the sense that ATP binding to a second and a third site greatly enhances enzyme turnover apparently by increasing the rate of product (ADP) removal from the first site (4, 7). Consistent with the view that each Fl-ATPase molecule contains three functional and interacting catalytic sites, it has been shown that the curvilinear Eadie-Hofstee plots ( u / [ S ] uersus u ) of ATP hydrolysis by isolated bovine heart F1-ATPase represent three apparent KkTp and three associated V,,, values, the former being of the order of Site-site cooperativity in Fl-ATPase is a central feature of Boyer's proposed binding change mechanism for oxidative phosphorylation (7,(9)(10)(11). However, clear evidence of sitesite cooperativity in the direction of ATP synthesis has been lacking. Recently, Stroop and Boyer (12) have published data suggesting cooperativity in photophosphorylation and have estimated apparent KkDp values of 0.62 and 31 PM and respective Vmax values of 37 and 907 pmol. h-' (mg of chlorophyll)-'. The present communication demonstrates that the kinetics of oxidative phosphorylation catalyzed by bovine heart submitochondrial particles are consistent with negative cooperativity with respect to [ADP] and positive cooperativity in catalysis.

MATERIALS AND METHODS
The sources of chemicals used were as follows: ADP, Boehringer Mannheim; NAD and NADH, P-L Biochemicals; hexokinase, Tris, potassium succinate, and DL-P-hydroxybutyrate, Sigma; Demerol, Winthrop; CCCP; Calbiochem; and carrier-free 32P, ICN. Other chemicals were reagent grade.
Phosphorylating submitochondrial particles (SMP) were prepared from bovine heart mitochondria essentially according to Hansen and Smith (13). Heavy layer beef heart mitochondria were prepared from freshly isolated mitochondria according to Hatefi and Lester (14), suspended at 50-70 mg of protein/ml in 0.25 M sucrose, containing 10 mM Tris acetate, pH 7.5, 1.5 mM ATP, and 10 mM MgC12, frozen in liquid nitrogen, and stored frozen at -70 "C for variable lengths of time up to 6 months. The frozen mitochondrial suspension was thawed at room temperature, homogenized, adjusted to a final protein concentration of 40 mg/ml with 0.25 M sucrose, 10 mM Tris acetate, pH 7.5, and supplemented with 1 mM potassium succinate, 1.5 mM ATP, 10 mM MgCl,, and 10 mM MnC12. The suspension was subjected in 100-ml batches to sonication in a rosette cell for 1 min at 0 "C, using the Branson sonifier at maximum output and 50% pulse mode. The sonication was repeated once after allowing the samples to cool down for 3-5 min in the ice-bath. After sonication, the pH was adjusted to 7.5 with 1 N KOH and the suspension was centrifuged in a No. 30 rotor of Spinco model L ultracentrifuge for 7 min at 20,000 rpm. The supernatant layer was carefully decanted, leaving behind the loosely packed residue, and recentrifuged in a No. 40 rotor for 45 min at 40,000 rpm. The pellet was washed once in a buffer containing 0.25 M sucrose and 10 mM Tris acetate, pH 7.5, suspended by homogenization in the same buffer at 40-60 mg of protein/ml, frozen in liquid nitrogen in small aliquots, and stored at -70 "C. There was no appreciable loss of oxidative phosphorylation activity after storage for a period of at least 6 months. Protein concentration was determined by the biuret method (15) in the presence of 1 mg of deoxycholate/ml.
Oxidative phosphorylation activity was measured at 30 "C essentially as described before (16). In a final volume of 0.6 ml, the reaction mixtures at pH 7.5 contained 0.25 M sucrose, 50 mM Tris acetate, 0.5 mM EDTA, 25 mM glucose, 5 mM MgC12, 20 mM potassium phosphate (1-3 X 10l1 cpm/mol of "P), 1-1200 p~ ADP, 70 pg of hexokinase/ ml, and 50 pg of SMP/ml. When NADH was the respiratory substrate, the mixture was preincubated for 4 min at 30 "C, then the reaction was initiated by the addition of 0.5 mM NADH and terminated after 4 min (or as otherwise indicated) by the addition of 50 pl of 35% perchloric acid. In the P-hydroxybutyrate-driven reaction, 30 mM DL-@-hydroxybutyric acid and 1 mM NAD were added instead of NADH. Where indicated, Demerol in an aqueous solution or CCCP in ethanol was added to the reaction mixture prior to NADH addition. When The abbreviations used are: CCCP, carbonyl cyanide m-chlorophenylhydrazone; SMP, bovine heart submitochondrial particles.
succinate was the oxidizable substrate, 6.7 mM potassium succinate was included during preincubation (in the absence of ADP) to activate succinate dehydrogenase and ensure linearity of oxygen uptake with time. Then, oxidative phosphorylation was initiated by the addition of ADP. The rates of oxygen consumption in ng atom oxygen'min-' (mg of protein)-' in the absence and presence of 1 mM ADP were, respectively, 1000 and 1500 with NADH as the substrate (respiratory control ratio = 1.5) and 330 and 590 with succinate as the substrate (respiratory control ratio = 1.79). Esterified 32P was estimated essentially according to Pullman (17) as described by Stiggall et al. (18).
The validity of the kinetic analyses (see below) required that the oxidative phosphorylation assays conform to certain stringent conditions, which were satisfied by the following tests. (i) In all the assays, the rate of oxygen consumption was measured polarographically under the oxidative phosphorylation assay conditions in the absence and presence of variable concentrations of ADP. All oxygen uptake rates were linear during the period of time that oxidative phosphorylation measurements were made (data not shown). (ii) Regardless of the ADP concentration used, 32P esterification was linear with time, at least for 4.5 min (Fig. 1). As stated above, the reaction time employed (4 min) was within this linear range. (iii) The concentration of hexokinase employed was adequate. Even at the lowest ADP concentration used, the same values for esterified 32P were obtained when the hexokinase level was doubled or tripled (data not shown). (iv) Starting the NADH-driven reactions by addition, after 4 min of preincubation, of NADH together with variable low or high levels of ADP gave the same results in terms of 32P esterification as when ADP was preincubated for 4 min with SMP prior to NADH addition (data not shown). These results as well as the linearity of the data of Fig. 1 at low ADP concentrations eliminated the possibility of effective alteration by adenylate kinase of ADP concentration during preincubation or the course of the reaction. Fig. 2 is an Eadie-Hofstee plot of the kinetics of ATP synthesis by SMP respiring on NADH (open rectangles) or succinate (solid rectangles). The variable substrate was ADP. It is seen that both plots are curvilinear, even though with succinate as the oxidizable substrate the deviation from linearity is apparent only at ADP concentrations nearing saturation (see inset of Fig. 2 with the expanded abscissa). Analysis of the plots as before (8), for the least number of straight lines  Kinetic data are not shown in the figures. The kinetics of oxidative phosphorylation driven by @-hydroxybutyrate oxidation (Fig. 3) could also be analyzed in terms of a single straight line, but the points at the highest ADP concentrations did not fit a straight line. which when combined best fitted the experimental points, indicated two apparent' KiDP values differing by one order of magnitude and two V, , , values (Table I). Despite the fact that the relative contributions of the low and the high K,,, components to the overall plots were different for the NADHand the succinate-driven reactions (in the latter, as stated above, the contribution of the high K, component was very small), the two low K$Dp and the two high KiDp values obtained from the data of Fig. 2 were nearly the same ( Table  I). The ADP concentration range employed in these studies was 1 to 1200 p~. To avoid magnifying errors by the u / [ S ] term of the Eadie-Hofstee plots, we deliberately refrained from using ADP concentrations below 1 p~. Thus, there may still be a third, very small KiDp which would not be discerned under our experimental conditions. As pointed out above, Stroop and Boyer (12) have reported a KiDp of 0.62 p~ for photophosphorylation catalyzed by spinach chloroplasts.

RESULTS AND DISCUSSION
The results shown in Fig. 2 allow the following interpretations. (i) There are two kinds of F1-ATPase, each having its own characteristic KkDp and V, , for ATP synthesis. (ii) The catalytic sites on F1-ATPase act independently, and in the ADP concentration range of 1-1200 p~ exhibit the two K, and V, , , values shown in Table I. (iii) The catalytic sites on F1-ATPase exhibit cooperativity. The first possibility is the least likely because of lack of supportive evidence, while the last is the most probable because it has been established that the reverse reaction, i.e. ATP hydrolysis, catalyzed by isolated or membrane-bound F1-ATPase is marked by site-site cooperativity (3)(4)(5)(6)(7)(8).
It may be important at this point to discuss a question that the data of Fig. 2 may have evoked in the reader's mind, namely, why was this apparent cooperativity in oxidative phosphorylation never before observed? Two factors are clearly important for demonstration of this phenomenon. One factor is the high phosphorylation activity of the SMP used (note the apparent V,,, of 2340 nmol of ATP. min-l. mg-l in Table I; the experimentally determined rate of ATP synthesis at 1200 pM ADP was 2180 nmol. min". mg").
The other factor is the nature of the respiratory substrate and the rate of electron transfer. In oxidative phosphorylation experiments catalyzed by SMP, the respiratory substrate commonly used is either succinate or a system that generates NADH at a slow rate (e.g. P-hydroxybutyrate + NAD or alcohol dehydrogenase + ethanol + low levels of NAD). Substrate quantities of NADH are rarely employed, in part because the ATP yield * All the K , and V,, values given in this paper should be regarded as apparent K,,, and apparent V- .  FIG. 2. Eadie-Hofstee plots of the kinetics of oxidative phosphorylation catalyzed by SMP respiring on NADH (0) or succinate (W). Experimental conditions were the same as described under "Materials and Methods." u is the rate of ATP synthesis in nmol. min" (mg of SMP protein)", and [ S ] is ADP concentration (1 to 1200 p~) . This and Fig. 3 are computer printouts, and the dots represent the curves calculated from the K,,, and V,, values (see Table  I), which were obtained by computerassisted curve fitting as described elsewhere (8). The inset is a replot of the data of the succinate-driven reaction on an expanded abscissa.  (19), attenuation of the rate less and less discernible as the rate of respiration is further of respiration lowers the observed apparent K, values for attenuated. Results demonstrating this point are shown in ADP and Pi. This effect appears to be greater on the high K, Fig. 3. The inset shows the Eadie-Hofstee plot of the NADHthan on the low K, values. Consequently, the contribution of driven data of Fig: '2 (open rectangles) and a plot of data for a similar experiment in which the rate of NADH oxidation was inhibited by 80% by the addition of 2 mM Demerol to the reaction mixture (solid rectangles). Comparison of the two plots shows clearly that when the rate of NADH oxidation was attenuated the contribution of the high KhDp component to the overall plot was diminished. The main section of Fig.  3 shows in expanded form the Demerol attenuated data of the inset (solid rectangles) next to a plot obtained with 0-hydroxybutyrate + NAD as the respiratory substrate (crosses). The overall V,, in the latter case was even less than that in the presence of NADH and 2 mM Demerol, and the data fitted a single straight line up to about 100 p~ ADP and tended to deviate from linearity only at the highest ADP concentrations used. In order to obtain further support for the contention that attenuation of the rate of respiration tends to diminish the curvilinearity of the Eadie-Hofstee plots by changing the observed KEp values toward lower values, another experiment was performed in which oxidative phosphorylation was partially uncoupled by the addition of 0.58 p~ CCCP. The extent of uncoupling was so adjusted that the overall V , , was close to that obtained in the experiment in which the rate of NADH oxidation was attenuated by the addition of 2 mM Demerol. However, as we have shown elsewhere (16,19) (see also Ref. 20), uncoupling affects the observed K, values for ADP and Pi in the opposite direction, i.e. whereas attenuation of respiration decreases these K, values, partial uncoupling increases them. Thus, one would expect that in the presence of partially uncoupling concentrations of CCCP, the contribution of the low KEp component should be diminished in the Eadie-Hofstee plots, and the contribution of the high KFp component should be exaggerated. This is precisely what was observed, as seen in Fig. 3 (solid circles).
Two additional points should be discussed. (i) At the same overall V,, for ATP synthesis, the Eadie-Hofstee plot obtained with NADH as the respiratory substrate (in the presence of partially inhibitory concentration of either Demerol or rotenone, data not shown) was more curvilinear than that obtained with succinate as the energy source. The reason for this difference is under investigation. It may be pointed out, however, that the two experiments are not identical, because electrons emanating from NADH go through one additional coupling site (site 1) as compared to those from succinate. (ii) Except for the values obtained in the presence of CCCP, all the low K$Dp (7-9 p~) and the high KE' (95-120 p~) values obtained under various conditions appeared to be nearly the same (Table I). However, it should be emphasized that the only significant aspects of these numbers are their order of magnitude and the fact that the low and the high KEp values differ by at least a factor of 10. As discussed above and in detail elsewhere (16,19), the K, values measured for ADP and Pi will depend on the degree of coupling of the SMP preparation used and the electron transfer rate obtained with various respiratory substrates. The same factors appear to affect the KEp (KZ was not investigated) in photophosphorylation (21).