Energy-dependent Transformation of F0·F1-ATPase in Paracoccus denitrificans Plasma Membranes*

F0·F1-ATP synthase in tightly coupled inside-out vesicles derived from Paracoccus denitrificans catalyzes rapid respiration-supported ATP synthesis, whereas their ATPase activity is very low. In the present study, the conditions required to reveal the ΔμH+-generating ATP hydrolase activity of the bacterial enzyme have been elucidated. Energization of the membranes by respiration results in strong activation of the venturicidin-sensitive ATP hydrolysis, which is coupled with generation of \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \({\Delta}{\tilde{{\mu}}}_{\mathrm{H}^{+}}\) \end{document}. Partial uncoupling stimulates the proton-translocating ATP hydrolysis, whereas complete uncoupling results in inhibition of the ATPase activity. The presence of inorganic phosphate is indispensable for the steady-state turnover of the \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \({\Delta}{\tilde{{\mu}}}_{\mathrm{H}^{+}}\mathrm{-activated}\) \end{document} ATPase. The collapse of \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \({\Delta}{\tilde{{\mu}}}_{\mathrm{H}^{+}}\) \end{document} brings about rapid deactivation of the enzyme, which has been subjected to pre-energization. The rate and extent of the deactivation depend on protein concentration, i.e. the more vesicles are present in the assay mixture, the higher the rate and extent of the deactivation is seen. Sulfite and the ADP-trapping system protect ATPase against the \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \({\Delta}{\tilde{{\mu}}}_{\mathrm{H}^{+}}\) \end{document} collapse-induced deactivation, whereas phosphate delays the rate of deactivation. A low concentration of ADP (<1 μm) increases the rate of deactivation. Taken together, the results suggest that latent proton-translocating ATPase in P. denitrificans is kinetically equivalent to the previously characterized ADP(Mg2+)-inhibited, azide-trapped bovine heart mitochondrial F0·F1-ATPase (Galkin, M. A., and Vinogradov, A. D. (1999) FEBS Lett. 448, 123–126). A \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \({\Delta}{\tilde{{\mu}}}_{\mathrm{H}^{+}}\mathrm{-sensitive}\) \end{document} mechanism operates in P. denitrificans that prevents physiologically wasteful consumption of ATP by F0·F1-ATPase (synthase) complex when the latter is unable to maintain certain value of \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \({\Delta}{\tilde{{\mu}}}_{\mathrm{H}^{+}}\) \end{document}.

F 0 ⅐F 1 -ATPases (ATP synthases) are the oligomeric molecular machines that couple ATP hydrolysis (synthesis) with proton translocation across the energy-transducing membranes in mitochondria, chloroplasts, and bacteria. The structural arrangement of the subunits within F 0 ⅐F 1 complexes of various organisms is assumed to be very similar (1)(2)(3). The hydrophilic F 1 is composed of a trimer of tightly packed ␣⅐␤-subunit pairs and one copy each of the ␦-, ⑀-, and ␥-subunits (Escherichia coli nomenclature for the subunits is used). Long rod-like ␥-subunit is asymmetrically positioned in the central cavity of the almost spherical globular ␣⅐␤ trimer. F 0 component is the hydrophobic complex composed of 10 -14 transmembraneously positioned c-, one a-, and two b-subunits. The hydrophilic parts of b-subunits are bound to F 1 (to one pair of ␣⅐␤and one ␦-subunit), thus forming the peripheral stalk. The other central stalk is formed by ␥⅐⑀ complex, which interacts with c-subunit(s) arranged in a ring. F 1 bears three "catalytic" nucleotide-binding sites located on ␤-subunits, and F 0 serves as a proton-conducting path. It is generally believed that the coupling between the ATP hydrolysis (synthesis) and flow of protons across the membrane results from the consequence of the long distance conformational change: ␣⅐␤-pair-associated chemical catalysis 3 ␥⅐⑀ 3 ab 2 c n leading to rotation of the rotor (␥⅐⑀ bound to c-ring) within the stator (␣⅐␤ trimer fixed by two b-and one ␦-subunits) (4).
The kinetics of ATP hydrolysis catalyzed by the soluble F 1 or membrane-bound F 0 ⅐F 1 preparations of the enzyme (coupled or uncoupled) are very complex (5). It has been documented that the key factor for such a complexity is a formation of so-called ADP(Mg 2ϩ )-inhibited form of the enzyme originally described (6) and kinetically characterized in a number of reports published by our (7)(8)(9)(10)(11) and other groups (12)(13)(14)(15)(16). Several phenomena such as hysteresis in onset of the catalytic activity (17), slow inhibition of ATPase by Mg 2ϩ (18), activation of ATP hydrolysis by sulfite and other anions (9,19,20), and the inhibitory effect of azide (9) can be consistently explained by the kinetic scheme in which the slowly reversible interconversion between catalytically competent enzyme-ADP intermediate and its inactive "isomer" plays the central role (5). Perhaps the most intriguing property of the ADP(Mg 2ϩ )-deactivated ATPase is that being inactive in ATP hydrolysis, it is fully competent in the ATP synthase activity (21,22).
The ADP(Mg 2ϩ )/ATP-and possibly ⌬ H ϩ 1 -induced rearrangement within F 0 ⅐F 1 complex that trigger its ATP hydrolase and ATP synthase activities remains unclear. The ⑀-subunit, an endogenous inhibitor of the ATP hydrolase activity of F 1 and F 0 ⅐F 1 (23,24), seems to be the most likely candidate for the triggering function. It has been shown that two distinct domains of the ⑀-subunit can exist in different conformations interacting differently with ␥-subunit (25)(26)(27). Most recently, it has been reported that the isolated ⑀-subunit of F 1 from thermophilic Bacillus PS3 specifically binds free ATP, thus suggesting its possible role as a sensor for the cellular ATP level (28). Several schemes describing different conformations of ATP synthase and ATP hydrolase states of F 0 ⅐F 1 complex have been published (3,5,27).
Tightly coupled vesicles from Paracoccus denitrificans have been shown to catalyze only very low rates of ATP hydrolysis (29,30) while being capable of high rates of oxidative phosphorylation (31,32). The reason(s) why ATP synthase in P. denitrificans carries out apparently unidirectional catalysis is not known. An explanation of this phenomenon seems to be of importance for uncovering physiologically relevant mechanisms involved in in situ operation of the energy-transducing enzymes. Clearly, this problem cannot be solved using highly purified soluble F 1 preparations, which catalyze only uncoupled ATP hydrolysis. The membrane-bound F 0 ⅐F 1 in tightly coupled P. denitrificans vesicles is ideally suited for studies on the reversibility of the ⌬ H ϩ-supported ATP synthesis. In some respects, F 0 ⅐F 1 synthase of P. denitrificans seemed to be remarkably similar to the mitochondrial ADP(Mg 2ϩ )-inhibited enzyme form stabilized by azide (22) or to the thermophilic Bacillus PS3 mutant F 0 ⅐F 1 , which is incapable of ATP binding to the "non-catalytic" sites on ␣-subunits (33). This similarity is corroborated by recent reports demonstrating that the ATPase activity of coupled vesicles derived from P. denitrificans is susceptible to activation by sulfite and/or by energization (34,35). Analogous phenomena have been reported for chloroplasts (36), photosynthetic purple bacteria (37), cyanobacterium Synechococcus 6716 (38,39), E. coli (40), and the mitochondrial enzyme (41). In this work, we report the results of studies aimed to delineate the conditions required for activation of the latent ATP hydrolase activity of P. denitrificans F 0 ⅐F 1 . We conclude that latent proton-translocating ATPase is, indeed, kinetically equivalent to the mitochondrial ADP(Mg 2ϩ ), azidetrapped F 0 ⅐F 1 .

EXPERIMENTAL PROCEDURES
Preparation of Vesicles-P. denitrificans cells (strain Pd 1222) were grown anaerobically in the presence of succinate and nitrate (42). Particles were prepared essentially as described (42) with modifications (43). The final preparation was suspended in 0.25 M sucrose, 10 mM Tris acetate (pH 7.3), 1 mM Mg-acetate, and 0.1 mM malonate and stored in liquid nitrogen. Before the experiments, the samples were thawed and diluted to 10 mg of protein/ml in the mixture comprising of 0.25 M sucrose, 0.1 M KCl, 2.5 mM Hepes, 0.1 mM EDTA (potassium salt, pH 8.0), and 5.5 mM MgCl 2 . The respiratory control measured as the ratio of NADH oxidase in the presence and absence of uncoupler (gramicidine (0.15 g/ml) and ammonium acetate (15 mM) for different preparations varied from 3.0 to 4.5. An average preparation catalyzed oxidative phosphorylation with succinate as the substrate at the rate of 0.6 mol/min/mg of protein at 25°C (pH 8.0). The ATPase activity of the membranes as prepared and assayed in the standard reaction mixture containing no succinate was not more than 0.04 mol/min/mg of protein at 25°C (pH 8.0). The activity was not increased by prolonged (3 h) incubation of vesicles in the presence of 50 mM dithiothreitol at 20°C.
ATPase Activity-The initial rates of ATP hydrolysis were determined as pH change (44)  The sensitivity was calibrated by the addition of small aliquots of titrated 20 mM HCl to the reaction mixture. The absorbance response depended linearly on the concentration of added HCl within a small range of pH change. An increase of buffer capacity due to the release of inorganic phosphate during the reaction (no more than 100 M as compared with 2 mM initially present) was negligible. The pH-metric assay for ATPase was the method of choice because, as will be shown in "Results," under most experimental conditions, the ATPase activity is strongly time-dependent, and it can only be reliably detected using the registration that allows following the time course of the reaction directly. It is worth noting that some of the phenomena reported here could be easily overlooked if discontinuous assay such as inorganic phosphate release would be used. To validate the pH-metric assay, in some experiments, ATPase activity was followed as NADH oxidation (at 366 nm, ⑀ m M ϭ 3.3) in the standard reaction mixture containing ATP regenerating system (1.5 mM P-enolpyruvate, pyruvate kinase, and lactate dehydrogenase, 20 units/ml of each), 250 M NADH, and particles preincubated with piericidin (1.5 nmol/mg of protein) to inactivate their NADH oxidase activity. The specific venturicidin-sensitive ATPase activities as determined by either method were the same. Succinate was used as the respiratory substrate because its oxidation by oxygen does not result in pH change, which would otherwise interfere with the ATPase assay.
Oxygen Consumption-Oxygen consumption was measured amperometrically with oxygen-sensitive platinum electrode.
All fine chemicals were from Sigma. Venturicidin was a kind gift of Dr. C. Hä gerhä ll (University of Lund, Lund, Sweden). A number of protonophoric uncouplers were tested in the preliminary experiments. curves, inside-out plasma membrane vesicles of P. denitrificans (P, 140 g/ml) were added to the standard reaction mixture, and the ATPase hydrolysis registrated as H ϩ release was initiated by the addition of malonate (20 mM) where indicated. The lower curves show the synchronous change of the membrane potential as followed by Oxonol-VI response. The final concentration of S-13 was 2 M. B, inorganic phosphate was omitted where indicated. The rate of activated ATP hydrolysis was 0.12 mol/min/mg of protein, and the reaction was abolished if venturicidin (1 g/mg) was added to the assay mixture.
H ϩ /min/mg of protein) and small transitory increase of the membrane potential were seen when particles were added to the reaction mixture containing no succinate (Fig. 1A). In the presence of succinate, the potential was raised, as expected, and slight alkalinization was observed. This small variable increase of pH presumably resulted from phosphorylation of contaminating ADP, as was evident from the experiments in which a coupled ATP-regenerating system was used for the ATPase activity assay. Rapid ATP hydrolysis (0.12 g of ion H ϩ /min/mg of protein) was immediately started after the succinate oxidase activity was prevented by the addition of malonate. Only partial decrease of the potential was seen in the presence of malonate, showing that activated ATP hydrolysis generated ⌬. This was confirmed by tracing of ⌬; rapid and complete drop of the potential was seen after the addition of malonate if ATP was omitted from the reaction mixture. The steady-state rate of ATP hydrolysis by coupled vesicles seen in the presence of malonate was expected to be limited by the back pressure of ⌬ H ϩ. However, paradoxically, the addition of protonophoric uncoupler (S-13), which caused an immediate drop of the membrane potential, also caused rapid and almost complete inhibition of ATP hydrolysis.
The presence of inorganic phosphate was found to be required for the continuous steady-state ATP hydrolysis induced by malonate (after pre-energization by coupled succinate oxidation), as shown in the Fig. 1B. Only a short jump of the ATPase activity accompanied by a short transitory maintenance of the potential was seen, and both the rate of ATP hydrolysis and the membrane potential rapidly declined in the absence of phosphate.
The results shown in Fig. 1 show that latent proton-translocating ATPase can be activated by the membrane energization and complete uncoupling leads to complete inhibition of ATP hydrolysis. The latter unexpected phenomenon was further investigated in the experiments shown in Fig. 2. Gradual uncoupling was induced by increasing concentrations of S-13, and the ATPase activity and ⌬ were followed as described in Fig.  1 except that malonate addition was excluded. The energization-activated ATP hydrolysis was stimulated by low concentration of S-13, which was unable to dissipate the membrane potential generated by succinate oxidation (Fig. 2, left part). Under these conditions, the ATPase activity was high enough to support energization of the membrane, as was evident from the substantial ⌬ seen after the suspension became anaerobic. The addition of a high concentration of S-13 resulted in almost the same steady-state rate of ATP hydrolysis, which declined rapidly, after the suspension became anaerobic concomitantly with dissipation of the potential (Fig. 2, right part). Intermediate patterns of ATP hydrolysis and the membrane potential change were seen at intermediate concentrations of the uncoupler (Fig. 2, middle part). Qualitatively the same pattern was seen with other uncouples, i.e. gramicidin or alamethicin (the data are not shown).
The dependence of the steady-state ⌬ H ϩ-activated ATPase activity on uncoupler concentration measured as described in Fig. 2 is shown in Fig. 3. The bell-shaped curve suggests that the optimal value of ⌬ H ϩ that is low enough to avoid its back pressure inhibitory effect on the ATP-dependent proton flow across the membrane, and high enough to maintain the enzyme in the active state, is required to reveal full ATP hydrolytic activity. This interpretation of the bell-shaped curve was supported by the experiments in which the ⌬ H ϩ generation (succinate oxidation) was partially decreased by malonate. As predicted, less uncoupler was required for the maximal ATPase rate, and a smaller fraction of the enzyme was activated (as was evident from the lower level of the ATPase activity) when ⌬ H ϩ generation (succinate oxidation) was limited. It should be noted that no quantitative correlation between the degree of activation and absolute value of the steady-state ⌬ can be estimated from the data shown in Figs. 2 and 3 because the observed amplitude of the Oxonol VI response is only a semiquantitative indicator of ⌬.
The ⌬ H ϩ-dependent activation of ATPase was shown to be reversible; when ⌬ H ϩ was collapsed either by inhibition of respiration (excess of malonate, anaerobiosis) or by the addition of uncoupler before ATP, the enzyme activity rapidly declined. In the preliminary trials, we have experienced poor reproducibility of the specific (expressed as units per mg of protein) ⌬ H ϩ-activated ATPase activity and its deactivation time course. Eventually, it was found that both the degree of activation and the rate of deactivation were dependent on protein content in the assay mixture. Fig. 4 shows the time course of the deactivation process as a function of protein content. A substantially higher level of the specific ⌬ H ϩ-induced activity and slower rate and lower degree of the deactivation were seen in very diluted samples, whereas smaller activation and rapid and complete deactivation occurred at high protein content. This behavior is expected if the activation phenomenon is due to the ⌬ H ϩ-induced dissociation of an inhibitory ligand and the deactivation results from to the backward binding of the ligand in the second-order kinetics reaction that took place after ⌬ H ϩ collapses. ADP seemed to be the most conceivable inhibitory ligand. As shown in Fig. 5A, added ADP significantly increased the rate of deactivation. The affinity of ADP to its inhibitory binding site could not be determined because the actual content of the enzyme and enzyme-bound ADP pre-existed in the vesicles prior to ⌬ H ϩ-induced activation were not known. However, very low (less than 1 M) concentrations of added ADP were sufficient for significant acceleration of the deactivation process (Fig. 5B).
Further evidence for the crucial role of tightly bound ADP in the active/de-active ATPase transition was obtained in the experiments with the ADP-trapping system (pyruvate kinase plus phosphoenol pyruvate). The experimental setup was different from that employed in Fig. 5 because the presence of an ATP-regenerating system in the assay mixture was expected to interfere with the pH-metric measurement of ATP hydrolysis. To avoid this, the ⌬ H ϩ-induced activation and subsequent deactivation were performed in the separately preincubated samples, and the final level of ATPase activity was measured after strong dilution of the suspension in the assay as described in Fig. 6. In contrast to the ADP(Mg 2ϩ )-inhibited mitochondrial F 0 F 1 -complex (8) and in accord with our recent data on P. denitrificans preparations (35), incubation of the de-energized vesicles with pyruvate kinase and phosphoenol pyruvate did not result in activation of ATP hydrolysis. The presence of the ADP-trapping system also did not affect the degree of the ⌬ H ϩ-induced activation. However, it strongly prevented the "spontaneous" ⌬ H ϩ collapse-induced deactivation of ATPase.
It has been shown that sulfite prevents and reverses the ADP-induced inhibition of the mitochondrial (9) and chloroplast (36) F 1 -ATPases. Sulfite has also been shown to stimulate bacterial F 1 -ATPases, particularly that of P. denitrificans (34). It seemed thus logical to check whether sulfite is able to protect the ⌬ H ϩ-activated ATPase against further deactivation upon de-energization. Fig. 7A shows that this was the case and that the ⌬ H ϩ-activated ATPase remained in its active state in the presence of sulfite after the membrane energization had been cancelled. A more complex effect of inorganic phosphate on the deactivation kinetics was seen (Fig. 7B). Phosphate was shown to be indispensable for the ⌬ H ϩ-activated steady-state ATP hydrolysis (Fig. 1B). If phosphate was absent in the assay in which the deactivation process was followed, the steady-state ATPase rate was negligible. Please note that the true initial rate could not be determined at the time resolution of the assay. On the other hand, the presence of inorganic phosphate in the assay during the activation and deactivation significantly delayed the latter process.

DISCUSSION
The soluble F 1 -and membrane-bound F 0 ⅐F 1 -ATPases from all sources studied so far undergo strong, slowly reversible inhibition of their ATP hydrolytic activity when exposed to very low concentrations of ADP in the presence of Mg 2ϩ (see Refs. 5, 47, and 48 for reviews). There exist several ways to reactivate the ADP(Mg 2ϩ )-inhibited ATPase. (i) The mitochondrial enzyme is slowly reactivated (k obs ϳ0.06 min Ϫ1 ) by "irreversible" removal of ADP in pyruvate kinase reaction (8). (ii) This slow activation is accelerated by a factor of about 10 in the presence of ATP (6,7). (iii) Inorganic phosphates have been shown to reactivate the inhibited form (49) and to permit the steadystate turnover of the bacterial F 1 mutated at the nucleotidebinding site located on the ␣-subunit (50,51). (iv) Removal of Mg 2ϩ by prolonged incubation with EDTA results in activation of the inhibited form, and this reactivation is considerably accelerated by P i , free ATP, and sulfite (10). (v) In tightly coupled membranous preparation in which F 0 ⅐F 1 -ATPase is originally inactive, as in chloroplasts (36) and some bacteria Vesicles were added to the standard reaction mixture (ATP was omitted), and after 2.5 min of incubation, respiration was stopped by the addition of 20 mM malonate (zero time). ATP (2 mM) was added after malonate at the time indicated on abscissa, and the initial steady-state rate of ATP hydrolysis was measured as shown in Fig. 1. In curves 1, 2, and 3, the protein content in assay mixture was 342, 114, and 38 g/ml, respectively.
FIG. 5. Effect of ADP on deactivation of the ATPase activity. A, the deactivation process was followed at 35°C as described in the legend for Fig. 4 at the protein concentration in the assay mixture of 50 g/ml. Malonate (20 mM) and ADP (where indicated) were added at zero time simultaneously. B, the residual rates of ATP hydrolysis (20 s after the deactivation was started as described in A) were measured as a function of ADP concentration. (37)(38)(39)(40), or when it was deliberately deactivated in submitochondrial particles (41), the activity can be restored by ⌬ H ϩ. Interestingly, the ADP(Mg 2ϩ )-inhibited form in contrast to its ATPase activity shows no lag phase in the ATP synthase reaction (21,22). These observations taken together can hardly be accommodated with the widely accepted "alternating binding change" mechanism, which leaves no alternative but for nucleotide and P i to participate in the process in a way different from that required for the binding and release during the reversible catalysis (52). Accumulating evidence (5,22,33,53) suggests that ⌬ H ϩ-dependent ATP hydrolysis (synthesis) cannot easily be described in terms of "reversibly" operating proton-translocating F 0 ⅐F 1 -ATPase. P. denitrificans F 0 ⅐F 1 -ATP synthase is, perhaps, the most illustrative example of the apparent irreversibility of the electron transport linked ATP synthesis as has been demonstrated many years ago (54).
Previous studies (35) have shown that, in contrast to the mitochondrial ADP(Mg 2ϩ )-inhibited F 0 ⅐F 1 , ATP hydrolysis in P. denitrificans membranes is not activated by the prolonged incubation with phosphate, or phosphoenol pyruvate and pyruvate kinase, or EDTA (pathways of activation (iii), (i), and (iv), as depicted above). These observations might seem to rule out the ADP(Mg 2ϩ ) inhibition as a possible reason for the absence of active ATPase. However, closer inspection, as reported here, suggests that similar to the chloroplast enzyme (36), the latent ATPase in P. denitrificans is present as the ADP(Mg 2ϩ )-inhibited form. The data shown in Fig. 3 suggest that ⌬ H ϩ-induced activation results from dissociation of an inhibitory ligand, and the data depicted in Figs. 4 -6 show that this ligand is ADP. The failure of pyruvate kinase, P i , and EDTA to activate ATPase can be explained by much higher affinity of ADP to its inhibitory site as compared with that for the mitochondrial enzyme. This affinity is likely to be as high as is seen in the azide-trapped ADP(Mg 2ϩ )-inhibited mitochondrial F 0 ⅐F 1 (9). This explanation is in accord with the previously reported data of Harris et al. (30), who have shown that P. denitrificans coupling ATPase contains tightly bound nucleotides, which became exchangeable upon the membrane energization. Our data somehow agree with those recently reported by Pacheco-Moisés et al. (34). However, their speculative proposal on two conformationally different active states of the enzyme as induced by either succinate or sulfite can hardly be deduced from the activating effect of trypsin on the ATPase activity in such a complex system as vesicular preparation of the bacterial membranes. Moreover, as is shown in this report (Fig. 3), in coupled P. denitrificans vesicles, great precautions should be taken if the enzyme activity is to be quantitatively characterized, and the optimal level of ⌬ H ϩ dissipation should be determined for any particular preparation under given conditions. At present, we do not know whether sulfite-protected active ATPase in P. denitrificans vesicles is fully coupled or uncoupled, as has been demonstrated for Rhodobacter capsulatus membranes (37). We were unable to follow ⌬ response in the presence of sulfite because of non-enzymatic interaction of the latter with Oxonol VI.
We were also not able to discriminate what particular component of ⌬ H ϩ, ⌬ or ⌬pH or both, serves as the driving force for the activation of ATPase (dissociation of ADP). Our attempts to distinguish these possibilities with standard approaches such as the use of valinomycin and nigericin to transform ⌬ H ϩ into ⌬pH and ⌬, respectively, have failed, presumably because of the presence of some unidentified cation (anion) transporting system in P. denitrificans plasma membrane.
Another feature of P. denitrificans ADP(Mg 2ϩ )-de-activated ATPase, in addition to higher affinity for inhibitory ADP, which differs from that of the mitochondrial enzyme, is its susceptibility to inorganic phosphate (Fig. 1B). Although P i has been shown to decrease apparent affinity of the mitochondrial F 0 ⅐F 1 to ADP (49), it does not affect the bi-phase kinetics of the uncoupled ATP hydrolysis during the assay up to 10 -50 mM Samples marked No energization initially contained 20 mM malonate to prevent succinate oxidation. After 2 min of incubation, 20 mM malonate was added to collapse ⌬ H ϩ (Fig. 1), and 1 min later, the ATPase activity was detected. The initial rates of ATP hydrolysis were measured at 35°C as described under the "Experimental Procedures" section, except that the reaction was started by the addition of small aliquots of preincubated suspension (20 l) to the standard assay mixture containing 20 mM malonate. Note that phosphoenol pyruvate and pyruvate kinase (if present) were 100ϫ diluted in the assay mixture.
The control experiments showed that no detectable pyruvate kinase reaction proceeded in the ATPase assay samples, and neither pyruvate kinase nor phosphoenol pyruvate alone had effect on the ATPase activity.

FIG. 7. Effects of sulfite (A) and
phosphate (B) on deactivation of the ATPase activity. The deactivation rate was followed at 35°C as described in the legend for Fig. 4 at the protein concentration in the assay mixture of 50 g/ml. 2 mM potassium phosphate or 1 mM sodium sulfite were present where indicated. 2 mM potassium phosphate was present in A. The initial steady-state rates of ATP hydrolysis (5 s after the addition of ATP) were measured.
concentration. The possibility that ⌬ H ϩ-induced activation, as reported here, is due to phosphorylation of tightly bound inhibitory ADP seems to be unlikely because phosphate is not required for activation per se but only needed for the continuous steady-state turnover of the activated enzyme (Fig. 1B). Please note that the presence of tightly bound P i in the latent deactivated enzyme cannot be excluded, and its presence in bovine heart F 0 ⅐F 1 preparations has been reported (55). The requirement of a high level of ⌬ H ϩ on the coupling membrane to maintain the proton-translocating ATPase activity (Fig. 3) appears to be an important physiologically relevant regulatory feature of the bacterial F 0 ⅐F 1 complex. The electron transfer-generated ⌬ H ϩ is used for a number of processes, i.e. active transport of nutrients. If the electron transfer activity becomes temporarily limited, ATP produced by anaerobic carbohydrates break down via the Entner-Doudorff pathway can be utilized by F 0 ⅐F 1 -ATPase to maintain ⌬ H ϩ, thus providing a driving force for the uptake of necessary metabolites. If the ATPase-supported ⌬ H ϩ would drop to a certain level under such conditions, further wasteful ATP hydrolysis would be prevented via the ⌬ H ϩ-sensitive deactivation mechanism. A similar regulatory mechanism realized by small ATPase inhibitor peptides has been suggested for the eukariotic mitochondrial F 0 ⅐F 1 -ATPase complex (56 -59). Interestingly, the ratio between the active and deactivated enzyme forms is apparently poised in equilibrium with ⌬ H ϩ, thus suggesting that the energy provided by either the electron transfer reactions or ATP hydrolysis is permanently utilized not only as the thermodynamic driving force for ATP synthesis or ATP-supported ⌬ H ϩ generation but also to maintain the catalytically competent state of the enzyme.
The last point to be mentioned concerns the problem of the F 0 ⅐F 1 -ATPase (synthase) reversibility, which has been discussed elsewhere (5). The ⌬ H ϩ-dependent ATP hydrolase activity of P. denitrificans F 0 ⅐F 1 was shown to be capable of proton pumping. Since several parameters for the ATPase activation and deactivation processes have been established, it is of obvious interest to find out whether these parameters are the same for transformation between the active and de-active ATP synthase forms. Work in this direction is currently underway in our laboratory.