The Vacuolar H+-ATPase of Lemon Fruits Is Regulated by Variable H+/ATP Coupling and Slip*

Lemon fruit tonoplasts, unlike those of seedling epicotyls, contain nitrate-insensitive H+-ATPase activity (Müller, M. L., Irkens-Kiesecker, U., Rubinstein, B., and Taiz, L. (1996) J. Biol. Chem. 271, 1916–1924). However, the degree of nitrate-insensitivity fluctuates during the course of the year with a seasonal frequency. Nitrate uncouples H+ pumping from ATP hydrolysis both in epicotyls and in nitrate-sensitive fruit V-ATPases. Neither bafilomycin nor oxidation cause uncoupling. The initial rate H+/ATP coupling ratios of epicotyl and the nitrate-sensitive fruit proton pumping activities are the same. However, the H+/ATP coupling ratio of the nitrate-insensitive fruit H+ pumping activity is lower than that of nitrate-sensitive and epicotyl V-ATPases. Several properties of the nitrate-insensitive H+-ATPase of the fruit indicate that it is a modified V-ATPase rather than a P-ATPase: 1) insensitivity to low concentrations of vanadate; 2) it is initially strongly uncoupled by nitrate, but regains coupling as catalysis proceeds; 3) both the nitrate-sensitive and nitrate-insensitive fruit H+-pumps have identicalK m values for MgATP, and show similar pH-dependent slip and proton leakage rates. We conclude that the ability of the juice sac V-ATPase to build up steep pH gradients involves three factors: variable coupling,i.e. the ability to regain coupling under conditions that initially induce uncoupling; a low pH-dependent slip rate; the low proton permeability of the membrane.

The vacuolar H ϩ -ATPases (V-ATPases) 1 of eukaryotic cells are large, multimeric proton pumps composed of 10 to 13 different subunits organized into a hydrophilic catalytic complex, V 1 , and a hydrophobic transmembrane H ϩ channel, V o (1). V-ATPases are structurally related to the ATP synthases, or F-ATPases, of mitochondria and chloroplasts. F-and V-ATPases exhibit many structural and functional similarities at the molecular level, including several homologous subunits which play central roles in catalysis and proton transport (2,3). Because of these similarities in structure and catalytic mechanism, F-and V-ATPases are also thought to exhibit similar mechanochemical properties (4). When operating in ATP synthesis mode, F-ATPases convert the proton-motive force present across the internal membrane of chloroplasts and mitochondria into a rotary torque used to drive the synthesis of ATP by the catalytic F 1 sector (5,6). Conversely, when F-ATPases operate in proton pumping mode, they hydrolyze ATP at their catalytic site and convert the liberated energy into a rotary torque utilized to drive proton transport across the membrane (7). V-ATPases are also thought to operate via a rotary mechanism, even though direct experimental evidence for rotation is lacking (1).
In V-and F-ATPases, the efficiency of coupling between ATP hydrolysis and proton transport represents a critical, ratelimiting factor. Both pumps have been proposed to undergo slip, or intrinsic uncoupling, under specific conditions (8 -10). During slip, ATP is hydrolyzed at the catalytic site without being coupled to H ϩ transport. Since typical V-ATPases are known to operate far from thermodynamic equilibrium, slip may represent one of the enzyme's regulatory mechanisms.
In lemon fruit juice sac cells, the vacuolar pH can reach as low as 2.2, about 3 pH units lower than in typical plant vacuoles. In fact, the trans-tonoplast pH gradient in lemon fruits is close to the calculated maximum for a V-ATPase operating at thermodynamic equilibrium (4 to 5 pH units) assuming an H ϩ /ATP stoichiometry of 2 (11,12). We have previously shown that proton pumping by tonoplast-enriched juice sac vesicles is largely insensitive to the V-ATPase inhibitors nitrate, bafilomycin A 1 , and partially sensitive to high vanadate concentrations (50 to 300 M) (13). The proton pumping activity of juice sac vesicles is also less sensitive to oxidation and N-ethylmaleimide (NEM) than that of tonoplast-enriched vesicles from seedling epicotyls. In addition, cold inactivation in the presence of nitrate, which completely inhibits proton transport in epicotyl vesicles, has little effect on the H ϩ pumping activity of juice sac vesicles, even though the treatment induces the release of similar proportions of catalytic subunits from both types of vesicles. Since V 1 dissociation treatment increased the sensitivity of the juice sac H ϩ pumping activity to vanadate, we hypothesized that a second, vanadate-sensitive proton pump, possibly possessing a lower H ϩ /ATP stoichiometry that enables it to generate a steeper pH gradient than the V-ATPase, may be present on the tonoplasts of juice sac cells (13).
After purification and reconstitution into artificial proteoliposomes, both the epicotyl and juice sac proton pumps exhibited equal sensitivities to nitrate, bafilomycin A 1 , NEM, and oxidation. Thus, the insensitivity of the juice sac H ϩ pumping activity to these inhibitors appears to depend on some component(s) of the native membrane. However, purified and reconstituted fruit V-ATPases remained partially sensitive to vanadate and exhibited only half as much slip as the epicotyl V-ATPase (14). Slip was calculated from proton pumping curves, in the absence of an electric potential gradient, by the method of Tu et al. (15). Under these conditions, slip reflects intrinsic uncoupling induced by the build-up of a pH gradient across the membrane. This method gives no indication on the H ϩ /ATP coupling ratio of the pump in the absence of a pH gradient.
H ϩ /ATP coupling ratios provide a measure of pump efficiency (16). If two pumps have the same H ϩ /ATP stoichiometry, but differ in their coupling ratios, the pump with the higher H ϩ /ATP coupling ratio would be expected to generate a steeper ⌬pH than the pump with the lower coupling ratio. Thus the coupling ratio of the lemon fruit V-ATPase might be higher than that of the epicotyl V-ATPase even if the two pumps had the same stoichiometry. To test this hypothesis, we have carried out experiments to characterize the coupling ratios of the fruit and epicotyl V-ATPases under initial rate conditions. Moreover, since the H ϩ pumping activity of the fruit is relatively insensitive to V 1 dissociation, we also determined the effect of V 1 dissociation treatment on the coupling ratio of the juice sac proton pump(s). Our results show that, in contrast to pH-dependent slip, the H ϩ /ATP coupling ratios of the juice sac and epicotyl proton pumps, determined under initial rate conditions, are the same. Whereas V 1 dissociation with nitrate induced the complete uncoupling of the epicotyl V-ATPase, only a fraction of the fruit H ϩ pumping activity was uncoupled under identical conditions. However, the kinetic properties of the juice sac proton pumps after V 1 dissociation treatment were identical to those in the control vesicles. This strongly suggests that in juice sac vesicles, the residual proton pumping activity after V 1 dissociation treatment belongs to a nitrate-resistant subpopulation of V-ATPases, rather than to a different type of proton pump.
A 20 month analysis of the nitrate sensitivity of tonoplastenriched vesicles from juice sacs showed that the proportion of nitrate resistant activity varied during the course of the year. As the proportion of uncoupled enzymes varies, the overall H ϩ /ATP coupling ratio measured in juice sac vesicles changes. The significance of this type of variable H ϩ /ATP coupling in lemon fruits is unclear, but the nitrate resistant activity may represent a specialized group of V-ATPases which, under extreme conditions, is able to maintain, if not build up, the large ⌬pH across the tonoplast of juice sacs.

EXPERIMENTAL PROCEDURES
Materials-Lemon seeds (Citrus limon L. var. Schaub Rough Lemon) for growing seedling epicotyls were generously supplied by Willits & Newcomb, Inc., Arvin, CA. Lemon fruits (var. Eureka) were harvested from a tree on the campus of the University of California, Santa Cruz. Reduced nicotinamide-adenine dinucleotide (NADH) was from Boeh-ringer Mannheim. All other chemicals were purchased from Sigma or Fisher.
Membrane Preparation-Tonoplast-enriched membranes from lemon fruit juice sacs and epicotyls were prepared as described previously (14). Fruit juice sacs were homogenized in fruit homogenization buffer (1.5 M MOPS-KOH, pH 8.5, 2.25% polyvinylpyrrolidone 40, 0.75% bovine serum albumin, 7.5 mM EDTA, 2 mM DTT, and 0.1 mM phenylmethylsulfonyl fluoride) and epicotyls were homogenized in epicotyl homogenization buffer (0.5 M MOPS-KOH, pH 8.5, 1.5% polyvinylpyrrolidone 40, 0.5% bovine serum albumin, 5 mM EDTA, 2 mM DTT, and 0.1 mM phenylmethylsulfonyl fluoride). After a first centrifugation of the homogenates for 20 min at 12,000 ϫ g, the supernatants was subjected to ultracentrifugation for 60 min at 132,000 ϫ g, and the microsomal pellets obtained were further purified on a 10/35% sucrose step gradient for 60 min at 132,000 ϫ g. The 10/35% interface containing tonoplastenriched membranes was recovered, diluted with RB (10 mM BTP-Mes, pH 7.0, 20 mM KCl, 1 mM EDTA, 2 mM DTT, and 0.1 mM phenylmethylsulfonyl fluoride), and pelleted for 20 min at 174,000 ϫ g. The tonoplast-enriched membranes were resuspended in RB at a final concentration of ϳ5 g of membrane protein/l. V 1 Dissociation-Membrane vesicles were made up to 0.3 mg of protein/ml in 2.5 ml of RB containing 5 mM ATP, 7 mM MgSO 4 , and 0 or 500 mM KNO 3 . They were incubated on ice for 1 h and centrifuged 15 min at 412,000 ϫ g. The membrane pellet was resuspended in 750 l of RB and used for activity measurements and immunoblotting. The proteins in 50 l of the resuspended pellet and 167 l of the supernatant were used for immunoblotting after being precipitated with 10% trichloroacetic acid, washed with cold acetone, lyophilized, and separated by SDS-PAGE.
Activity Assays-Proton pumping and ATP hydrolysis by tonoplastenriched vesicles were measured simultaneously by using the continuous spectrophotometric assay of Palmgren (17). In this assay, proton transport into the vesicles was followed by measuring the absorbance decrease of acridine orange at 495 nm. Simultaneously, ATP hydrolysis was measured by coupling the appearance of ADP to the oxidation of NADH and by following the NADH absorbance decrease at 340 nm. The The effect of increasing concentrations of potassium nitrate on the maximum rate of proton pumping by three different preparations of tonoplast-enriched membranes from juice sacs was measured by quinacrine fluorescence quenching. ϳ80 g of membrane protein were used in each reaction. E represents a so called nitrate-insensitive preparation; Ⅺ represents a nitrate-sensitive preparation; ⌬ a partially nitrate-sensitive preparation. After correction for mixing artifacts, the results were loaded into a graphics program and plotted out (Fig. 1). When activity rates are reported, the slopes of the initial rates of proton pumping and ATP hydrolysis were calculated and expressed in arbitrary units (a.u.). The calculated H ϩ /ATP coupling ratios were independent of the amount of membrane protein used in the assay (data not shown). We nevertheless chose to do all experiments in the presence of 20 -25 g of membrane protein, unless stated differently. When proton pumping was measured in the presence of nitrate in the reaction mixture, quinacrine fluorescence quenching was substituted for acridine orange absorbance quenching. If the effect of nitrate was to be assessed on proton pumping alone, the conditions were as described previously (13). If H ϩ /ATP coupling ratios were to be determined, quinacrine fluorescence quenching was measured in the assay mixture described above, so that H ϩ pumping and ATP hydrolysis activities were measured in the same conditions. Because the initial rate of H ϩ pumping differed from the maximum activity rate in the presence of nitrate, the latter was chosen for all activity measurements in the presence of nitrate.
Immunoblotting-The proteins of tonoplast-enriched vesicles from epicotyls and juice sacs were separated by SDS-PAGE according to Laemmli (18) in 12% polyacrylamide gels and the transfer to nitrocellulose was done as described previously (13). The blots were incubated with a primary antibody to the 70-kDa subunit of the corn V-ATPase and visualized by a peroxidase-coupled secondary antibody reaction (Vectastain ABC kit, Vector Laboratories, Burlingame, CA) (13).
Protein Concentration-Estimates of protein concentrations were done routinely with Amido Black (19).

RESULTS
Sensitivity to Nitrate-Although we had previously found that proton pumping by fruit juice sac vesicles was largely insensitive to inhibition by nitrate and bafilomycin A 1 (13) (Table I), we observed that there was some variability in the nitrate insensitivity of the lemon fruit V-ATPase from one membrane preparation to another (Fig. 2). Fig. 3A shows the nitrate sensitivity of tonoplast-enriched fruit vesicles from preparations made over a period of 20 months. All fruits were harvested at the same state of development from a single tree situated on the campus of the University of California, Santa Cruz. We found that, over the duration of the experiment, the inhibition of the proton pumping activity by 100 mM KNO 3 followed a sigmoidal pattern, ranging from 0 to 20% inhibition in the fall to 70 -80% in the spring.
In addition to the seasonal variation in nitrate sensitivity noted above, a temporal component was also observed in the kinetic study of nitrate inhibition. As shown in Fig. 3B, even in a "nitrate-insensitive" juice sac preparation, H ϩ pumping appeared to be initially inhibited by nitrate. However, after 1-2 min, the activity underwent a progressive "recovery" to reach a maximum rate comparable to that of the control. This temporal delay in the attainment of the maximum rate of proton pumping was observed under all experimental conditions used (including experiments done in the presence of acridine orange) and whether an ATP regeneration system was present or not. However, we observed that the ATP regeneration system tended to reduce the magnitude of the delay. In contrast to H ϩ pumping, ATP hydrolysis did not show a similar lag phase (data not shown). This suggests that in the presence of KNO 3 , the fruit proton pumps were initially uncoupled, but became progressively coupled as catalysis progressed.
Sensitivity to Vanadate-The lemon fruit V-ATPase was previously found to be partially sensitive to high concentrations (Ն200 M) of vanadate (13,14). We therefore tested the effects of vanadate on the initial rates of proton pumping (Fig. 4, A and B) and ATP hydrolysis (Fig. 4, C and D), and on the deduced coupling ratio (Fig. 4, E and F) of tonoplast vesicles from juice sacs and epicotyls. Curves representing two types of juice sac preparations are shown, a "nitrate sensitive" one (open circles), which was ϳ70% inhibited by 100 mM KNO 3 , and a "nitrate insensitive" preparation (open triangles), which was inhibited only ϳ30% by the same concentration of nitrate.
As shown in Fig. 4, A and C, respectively, proton pumping and ATP hydrolysis by tonoplast vesicles of epicotyls exhibited little sensitivity to vanadate in the 0 -50 M range. Thus, the deduced coupling ratio of the epicotyl V-ATPase increased only slightly between 10 and 50 M vanadate. The lack of effect of vanadate suggests that contaminating P-ATPase activity is not a significant factor in our measurements of the coupling ratio of the epicotyl V-ATPase (Fig. 4E). High concentrations of vanadate inhibited proton pumping to some extent, while ATP hydrolysis appeared to be stimulated under the same conditions. However, this latter effect appeared to be an artifact due to the oxidative properties of vanadate which, at high concentration, oxidized the NADH used in the coupled assay (data not shown).
In contrast to the results with tonoplast vesicles from epicotyls, ATP hydrolysis by lemon fruit membranes was inhibited FIG. 3. Seasonal variation of the nitrate sensitivity of tonoplast-enriched juice sac vesicles. A, quinacrine fluorescence quenching was used to measure the sensitivity of tonoplast-enriched fruit vesicle preparations to 100 mM potassium nitrate over a period of 20 months. All fruits were harvested at the same stage of development from the same tree. For each membrane preparation the H ϩ pumping activity was measured in the presence and absence of 100 mM KNO 3 and the activity of the nitrate-treated samples, expressed as a percentage of the control samples, is reported. B, initial rates of quinacrine fluorescence quenching curves of a juice sac preparation showing little sensitivity to nitrate. C, control; ϩ100 mM NO 3 , treated with potassium nitrate. by ϳ50% by 50 M vanadate (Fig. 4D). However, proton pumping was only slightly inhibited by low concentrations of vanadate in both the nitrate-sensitive and the nitrate-insensitive preparations. This suggests that a considerable amount of contaminating plasma membrane ATPase activity, presumably in the form of leaky vesicles incapable of generating a ⌬pH, may be present in the fruit tonoplast-enriched preparations. Both ATP hydrolysis and proton pumping were inhibited by 400 M vanadate to ϳ60% of their activity in the presence of 50 M vanadate. The nitrate-sensitive fruit preparation was about twice as active in proton pumping as the nitrate-insensitive activity. As a result, the coupling ratio calculated for the nitrate-sensitive preparation is about 70% higher than that of the nitrate-insensitive preparation, and is comparable to the coupling ratio of the epicotyl V-ATPase (Fig. 4F).
V 1 Dissociation-We previously reported that cold dissociation in the presence of KNO 3 caused the release of catalytic subunits from fruit vesicles without affecting the total ATP-dependent H ϩ pumping activity relative to controls (13). The only effect of nitrate treatment on the fruit H ϩ pumping activity was to increase the sensitivity to high concentrations of vanadate (200 M) from 30% inhibition in control vesicles (treated with cold an 5 mM MgATP alone) to 60 -70% in vesicles treated with cold 5 mM MgATP and 500 mM KNO 3 . Fig. 5 confirms that the amount of catalytic subunit released from fruit membranes by nitrate was comparable to the amount dissociated from epicotyl vesicles, as determined by immunoblotting. The ATPdependent proton pumping activity, measured in the absence of vanadate, is given below the immunoblots.
To determine whether the H ϩ /ATP coupling ratio of the fruit V-ATPase had been altered by the treatment with 500 mM KNO 3 , nitrate-insensitive and nitrate-sensitive fruit vesicle preparations were subjected to cold inactivation in the presence or absence of 500 mM KNO 3 . H ϩ pumping and ATP hydrolysis were measured simultaneously after washing the coldreleased vesicles with buffer. The results are shown in Fig. 6 (A  to D). For a comparison, the activity of similarly treated epicotyl vesicles is also presented (Fig. 6, E and F). In the nitrate- insensitive fruit preparation, KNO 3 treatment had little or no effect on either the H ϩ pumping or ATP hydrolysis activities (Fig. 6A) and thus, the coupling ratio of the nitrate-insensitive preparation was unaffected. However, the nitrate-treated sample was strongly inhibited by 500 M vanadate (Fig. 6B). The results obtained with the nitrate-sensitive fruit preparation confirmed these findings (Fig. 6, C and D) and showed that 50 M vanadate was sufficient to inhibit most of the proton pumping activity after KNO 3 treatment. Table II shows a comparison of the H ϩ pumping and hydrolytic activities, and the coupling ratios of the "partially nitrate-sensitive" fruit preparation used in Fig. 6, C and D. Under control conditions (non-dissociated), the vanadate-sensitive ATPase activity was largely independent of H ϩ pumping, since 50 M vanadate inhibited ATP hydrolysis by 58% and H ϩ pumping by only 16%. However, after V 1 dissociation, 50 M vanadate inhibited ATP hydrolysis and H ϩ pumping by 71 and 74%, respectively, indicating that the vanadate-sensitive ATP hydrolytic activity is coupled to proton transport. Since the amount of vanadate-sensitive H ϩ pumping activity was doubled in absolute value after V 1 dissociation (0.27 to 0.54 arbitrary units), the dissociation procedure appears to have partially transformed the vanadate-insensitive H ϩ pumping activity into a vanadate-sensitive activity. Overall, the H ϩ /ATP coupling ratio of the partially nitrate-sensitive preparation, which in the control was 2.66 a.u. in the presence of 50 M vanadate, dropped to 0.61 Ϯ 0.04 a.u. after V 1 dissociation, whether vanadate was present or not (Table II). Thus V 1 dissociation has a major uncoupling effect on the nitratesensitive H ϩ pumping activity of juice sac vesicles.
A detailed analysis of vanadate sensitivity and coupling ratios was also performed under control and V 1 dissociation conditions for a nitrate-sensitive fruit preparation (Fig. 7). In the control vesicles, 50 M vanadate inhibited proton pumping by only 27% (Fig. 7A), whereas ATP hydrolysis was inhibited by 53% by the same concentration (Fig. 7B). After V 1 dissociation, 50 and 100 M vanadate inhibited the proton pumping activity by 69 and 89%, respectively (Fig. 7A). The H ϩ /ATP coupling ratio, which in the control was ϳ1.8 a.u. in the presence of 50 M vanadate, dropped to ϳ0.62 a.u. after V 1 dissociation (Fig.  7C). At 400 M vanadate the coupling ratio of the control was ϳ2.3, while that of the V 1 dissociated membranes was ϳ0.1. Thus, high concentrations of vanadate have opposite effects on the coupling ratios of control versus V 1 dissociated vesicles, increasing the former while decreasing the latter.
ATP Kinetics-Since the residual H ϩ -ATPase activity after V 1 dissociation treatment clearly differed from the control activity in terms of vanadate sensitivity, both activities were further characterized with respect to K m and V max in a fruit preparation exhibiting ϳ40% sensitivity to 100 mM KNO 3 (Fig.  8). From the curves in Fig. 8B, it is clear that the ATP hydrolytic activity was largely unaffected by nitrate up to a substrate concentration of 2 mM ATP. Beyond 2 mM, the first-order kinetics of the control vesicles diverged from strict Michaelis-Menten kinetics, suggesting the presence of some contaminating activity (Fig. 8B). From the Hanes-Woolf linearizations of the first-order kinetics between 0 and 2 mM ATP, we calculated an identical K m of 0.16 mM ATP and an identical V max for the vesicles treated in the presence or absence of KNO 3 (Fig. 8D). In contrast to ATP hydrolysis, the V max of the proton pumping activity was reduced 50% by nitrate treatment (Fig. 8, A and  C). However, the K m of 0.20 mM ATP for the nitrate-treated sample was identical to that of the control, suggesting that the same type of enzyme was active after V 1 dissociation as before. In order to determine whether the treatment with nitrate might have left intact a population of V-ATPases with a different coupling ratio than the normal V-ATPase, we normalized the initial rates of pumping of control and nitrate-treated vesicles and compared their pH gradient at steady state (Fig. 9A). The ⌬pH built up by nitrate-treated vesicles did not appear to differ significantly from that of control vesicles. The pH-dependent slip and leakage rate constants of control and nitratetreated vesicles were calculated and normalized for the apparent proton pumping rate at any time during the ⌬pH build up. As shown in Fig. 9B, the pH-dependent slip and leakage were identical in control and nitrate-treated vesicles.
Uncoupling by Nitrate-Experiments were carried out to determine whether the epicotyl V-ATPase and the nitratesensitive component of the fruit V-ATPase exhibited the same pattern of uncoupling by nitrate. Because acridine orange was reported to dissipate pH gradients in the presence of KNO 3 (20), we used quinacrine fluorescence quenching to measure proton transport in the experiments where nitrate was present.
As shown in Fig. 10, A and B, nitrate inhibited the H ϩ pumping and ATP hydrolysis activities of both the epicotyl and the fruit tonoplast vesicle preparations, although the total in-

TABLE II
Activities and coupling ratios of fruit tonoplast vesicles treated for V 1

dissociation
The proton pumping and hydrolytic activities of a partially nitrate-sensitive juice sac preparation were determined under the conditions described in the legend to Fig. 6. Activities and coupling ratios are given in arbitrary units and the inhibition by 50 M vanadate is given in arbitrary units and as a percentage of the activity of the sample not treated with vanadate. hibition was greater for epicotyl vesicles. The average coupling ratios of two fruit and two epicotyl preparations in the presence of increasing nitrate concentrations are shown in Fig. 10C. Note that only the nitrate sensitive activity was considered. The progressive decrease in the H ϩ /ATP coupling ratio between 0 and 50 mM KNO 3 clearly indicates that in fruit and epicotyl vesicles, nitrate inhibits proton pumping to a greater extent than ATP hydrolysis, and thus uncouples the V-ATPase. If only the nitrate-sensitive portion of the total activity is considered, uncoupling of the fruit V-ATPase occurs more readily at very low concentrations of KNO 3 (Ͻ10 mM); at concentrations Ͼ20 mM, however, the epicotyl is more strongly uncoupled by nitrate than the fruit. In the presence of 50 -100 mM KNO 3 , the H ϩ /ATP coupling ratio of the fruit tonoplast vesicles was consistently about twice that of the epicotyl vesicles.
Bafilomycin A 1 -Similar experiments were carried out to measure the effect of bafilomycin A 1 on the coupling ratios of fruit and epicotyl tonoplast membrane vesicles (Fig. 11, A and  B). In preliminary experiments it was found that fruit preparations that were more nitrate-sensitive exhibited increased sensitivity to bafilomycin as well. Fig. 11, A and B, show the effects of bafilomycin on the H ϩ pumping and ATP hydrolysis activities of an epicotyl preparation and a bafilomycin-sensitive fruit tonoplast preparation, respectively. The H ϩ /ATP coupling ratios between 0 and 1 M bafilomycin are shown in Fig. 11C. Bafilomycin had no effect on the coupling ratios of either the epicotyl or the fruit V-ATPases.
Oxidation-We previously demonstrated that oxidation inhibits the proton pumping activity of the epicotyl V-ATPase, and that the oxidative inactivation is partially reversible by DTT (13). Accordingly, the effect of oxidation on the H ϩ /ATP coupling ratios of epicotyl tonoplast vesicles was examined. To avoid artifacts due to contaminating ATPases, only the bafilomycin-and NEM-sensitive ATP hydrolysis and proton pumping activities were used for the determination. Exposing epicotyl tonoplast vesicles to air at 22°C in the absence of reductant reduced both H ϩ pumping and ATP hydrolysis activities in parallel (Fig. 12A). Moreover, DTT reversal was the same in both cases. The calculated coupling ratios are shown in Fig. 12B. Oxidation had no effect on the H ϩ /ATP coupling ratios of epicotyl V-ATPases. DISCUSSION V-ATPases normally operate far from thermodynamic equilibrium and are therefore considered to be under kinetic regulation (e.g. 11). Kinetic regulation of the V-ATPase may involve inhibitors (21,22), V 1 dissociation from the membrane (23), slip induced by ⌬ H ϩ (10), and variable H ϩ /ATP coupling (24). In lemon juice sacs, the steep trans-tonoplast pH gradient implies either that the V-ATPase operates close to thermodynamic equilibrium, or that a second H ϩ -ATPase with a lower H ϩ /ATP stoichiometry than the V-ATPase is present on the membrane. For the V-ATPase to reach thermodynamic equilibrium it would have to become refractory to kinetic regulation.
We previously reported that ATP-dependent proton pumping by juice sac vesicles was unusually insensitive to nitrate and other V-ATPase inhibitors (13). Although the juice sac V-ATPase became sensitive to inhibitors after being solubilized, purified, and reconstituted into liposomes, its pH-de- pendent slip rate was still about one-half that of the epicotyl V-ATPase. We therefore proposed that the low slip rate of the juice sac V-ATPase might be one of the factors which allows the lemon fruit V-ATPase to build up its steep equilibrium pH gradient (14).
In this report we have shown that the nitrate insensitivity of proton pumping by the lemon fruit V-ATPase is not constant throughout the year, but exhibits a seasonal variation. The minimum nitrate sensitivity occurs during the winter months, and the maximum nitrate sensitivity occurs in the spring and early summer. Since the nitrate sensitivity of the ATP hydrolytic activity is relatively constant throughout the year, what varies is the coupling between H ϩ transport and ATP hydrolysis. Whether this variation is due to changes in rainfall or other environmental factors remains to be determined. Since we have shown that nitrate insensitivity was membrane-dependent, it is possible that a seasonal change in membrane lipids is driving the seasonal fluctuation in nitrate sensitivity. Accordingly, the variable nitrate sensitivity may be associated with a single proton pump, the V-ATPase, which may exist in two states: a nitrate-sensitive (normal) state and a nitrateinsensitive (altered) state. These two states make up two subpopulations of V-ATPases on the juice sac tonoplast, with the relative amounts of each varying with the seasons. We designate this situation as the "one pump/two states model" in contrast to a "two pumps" model which involves the presence of two distinct H ϩ -ATPases with different stoichiometries.
One possible way to distinguish between proton pumps with different H ϩ /ATP stoichiometries is to compare their H ϩ /ATP coupling ratios, assuming tight coupling between hydrolysis and proton transport. In the present study, we used initial activity rates to compare the H ϩ /ATP coupling ratios of the proton pumps in fruit and epicotyl tonoplast vesicles.
A comparison of nitrate-sensitive versus nitrate-insensitive fruit preparations indicates that nitrate-sensitive vesicles have higher H ϩ /ATP coupling ratios than nitrate-insensitive preparations. Since the coupling ratio of the nitrate-sensitive vesicles is comparable to that of epicotyl preparations, and assuming that the epicotyl tonoplast is energized by a V-ATPase alone, the nitrate-sensitive juice sac vesicles are energized by only one type of proton pumping ATPase: the V-ATPase. If it contained a mixture of P-and V-type ATPases, with H ϩ /ATP stoichiometries of 1 and 2, respectively (11,(25)(26)(27), one would expect the coupling ratio of the fruit preparation to be lower than that of the epicotyl.
Even though the coupling ratio of the so-called nitrate-insensitive juice sac preparations was lower than that of tonoplast vesicles from epicotyls, two observations suggest that these vesicles also bear a V-ATPase rather than another type of H ϩ -ATPase: 1) during the initial seconds of H ϩ pumping in the presence of nitrate, the nitrate-insensitive fruit preparations are temporarily uncoupled. The subsequent recovery of coupling is dependent on the presence of MgATP and may involve some type of subunit rearrangement or possibly a phosphorylation reaction. 2 2) Nitrate-insensitive juice sac preparations show kinetic properties and vanadate sensitivities similar to the more nitrate-sensitive juice sac preparations after V 1 dissociation treatment. In the latter case, as shown below, the residual H ϩ pumping activity is thought to be due to a V-ATPase, based on its K m and the simultaneously measured ATP hydrolysis kinetics, even though the calculated coupling ratio is lower than in epicotyl V-ATPases.
Moriyama and Nelson (10) have proposed that nitrate induces uncoupling of proton transport from ATP hydrolysis in V-ATPases. In tonoplast-enriched vesicles from lemon juice sacs and epicotyls, nitrate indeed had such an effect. When only the nitrate-sensitive components of H ϩ pumping and ATP hydrolysis were considered (defined as the portions of the activities inhibited by 400 mM KNO 3 ), juice sac and epicotyl preparations exhibited significant uncoupling under treatment with 1 to 100 mM KNO 3 . At 50 mM KNO 3 , the H ϩ /ATP coupling ratio of epicotyl V-ATPases had dropped an average 89% from its initial value, while the fruit enzymes' uncoupling averaged ϳ72% of their initial coupling ratio. Accordingly, a subpopulation of nitrate-sensitive V-ATPases requires 400 mM KNO 3 for inhibition while remaining resistant to uncoupling by 50 to 100 mM KNO 3 .
Although the H ϩ pumping activity of fruit vesicles becomes more sensitive to vanadate after nitrate treatment, in untreated vesicles vanadate is most effective in the 50 to 300 M range. Since P-type H ϩ -ATPases are generally inhibited by 1.0 to 10 M vanadate, it is unlikely that the vanadate sensitivity of the lemon fruit H ϩ pumping activity involves the inhibition of a P-type enzyme (14). However, an ATP hydrolytic activity that is sensitive to low concentrations of vanadate is clearly present in fruit vesicle preparations. This is best illustrated by the H ϩ /ATP coupling ratios in the presence of increasing concentrations of vanadate. For nitrate-sensitive and nitrate-insensitive juice sac preparations, the coupling ratios rise sharply in the presence of 0 to 50 M vanadate, and then 2 M. Jensen, unpublished data.

FIG. 9. ATP-dependent proton pumping by fruit vesicles pretreated for V 1 dissociation in the presence or absence of nitrate.
A, tonoplast-enriched juice sac vesicles, treated as described in the legend to Fig. 8, were normalized for similar initial rates of H ϩ pumping. 7.5 g of membrane protein for the control (----) and 15 g of the nitrate-treated vesicles (O) were allowed to reach steady state equilibrium ⌬pH (no vanadate was present in the reaction mixture). Then the pH gradient was collapsed with 0.25 M nigericin. B, the pH-dependent slip and leakage rate constants were determined as described previously (14) and corrected for the rates of proton pumping by dividing them at any time point by the derivative of a fifth order equation describing the pumping curves obtained in panel A (ϭ pumping rate).
----, control; O, nitrate-treated. remain stable up to 500 M VO 4 . This indicates that the "contaminating" activity is mainly hydrolytic, and as it can be eliminated by low concentrations of vanadate (Ͻ50 M), it could be due to a P-type ATPase. Since concentrations of vanadate between 50 and 500 M inhibited the hydrolytic and pumping activities to the same extent, i.e. the H ϩ /ATP coupling ratio remains stable, high concentrations of vanadate appear to inhibit a single H ϩ -ATPase activity which is distinct from the activity inhibited by low concentrations of vanadate. V 1 dissociation treatments in the presence of high nitrate did not significantly modify the H ϩ /ATP coupling ratio in nitrate-insensitive juice sac preparations. However, V 1 dissociation did increase the sensitivity of the proton pump to high concentrations of vanadate. In nitrate-sensitive preparations, H ϩ /ATP coupling ratios were strongly reduced by the V 1 dissociation treatment, and concomitantly, proton pumping became highly sensitive to both high and low concentrations of vanadate. In the preparations analyzed, which were initially ϳ20% sensitive to vanadate, KNO 3 treatment increased the vanadate sensitivity to 50 to 70% of the initial activity. Simultaneously, the coupling ratio dropped to 1/2 or 1/4 of its value before KNO 3 treatment. Three explanations are possible. 1) KNO 3 treatment uncouples the V-ATPases present on the membranes while simultaneously coupling a population of P-type H ϩ -ATPases which were initially uncoupled. 2) KNO 3 treatment uncouples one of two populations of V-ATPases, the remaining population being vanadate-sensitive. 3) KNO 3 uncouples one of two populations of V-ATPases and modifies the remaining V-ATPases, inducing them to become vanadate-sensitive.
While the first explanation would yield a result similar to what was observed, it would imply that nitrate treatment induces the sealing of a leaky membrane fraction associated with the putative P-type hydrolytic activity. To our knowledge, no such effect of nitrate on the formation of sealed vesicles has been reported. Thus, we consider this explanation unlikely.
In the second explanation, there are two populations of V-ATPases: classical, nitrate-sensitive V-ATPases and nitrateinsensitive, vanadate-sensitive V-ATPases. However, this explanation does not account for the overall increased vanadate sensitivity of proton pumping after V 1 dissociation treatment.
The third explanation differs from the second in that a population of V-ATPases becomes vanadate-sensitive as a result of nitrate treatment. The increase in vanadate sensitivity could be a consequence of a molecular rearrangement that allows the fruit V-ATPase to continue to pump protons in the presence of nitrate. This explanation is consistent with previous results, in which we found that a vanadate-sensitive ATPase activity, apparently associated with partially disassembled V-ATPases, co-migrated with the juice sac V-ATPase during its purification by gel filtration and anion exchange chromatography (14). Thus, according to this third explanation, after some fruit V-ATPases become damaged by KNO 3 treatment, their activity is modified so as to become more vanadate-sensitive, while the overall coupling ratio is reduced due to uncoupling of another subpopulation of V-ATPases. The same type of damage may occur in vivo during certain periods of the year, resulting in nitrate-insensitive juice sac proton pumps and a low H ϩ /ATP coupling ratio. This one pump/two states model is further supported by the finding that the K m values for MgATP were identical in control vesicles and in vesicles treated for V 1 dissociation. This strongly suggests that the same type of H ϩ -ATPase was active before and after treatment with nitrate. The fact that the measured K m values were in the 0.16 to 0.20 mM range, considerably lower than the 0.88 Ϯ 0.18 mM previously determined for the juice sac H ϩ pumping activity (13) is attributed to the presence of an ATP regenerating system in the present study, which converts even traces of ADP back into substrate, while removing a potent inhibitor of the V-ATPase. Moreover, since the equilibrium ⌬pH was unchanged by treatment with nitrate, and since the calculated pH-dependent slip and leakage rate was identical to that of the control, KNO 3 treatment did not reveal a V-ATPase with a different stoichiometry. Based on all the above and consistent with the one pump/two states model, we interpret the seasonal reduction in the H ϩ /ATP coupling ratio as representing a change in the state of the V-ATPase, rather than a fluctuation in the activity of a second type of proton pump on the tonoplast.
It has been argued that inhibition of the V-ATPase by nitrate was the result of two distinct phenomena (28,29). At high concentrations, the chaotropic properties of nitrate promote the dissociation of the V 1 sector from V o (28), while below 50 mM, nitrate was proposed to involve a different mode of action (29,30). Our present results indicate that there is no distinct lowand high-nitrate effects on the lemon V-ATPases, as from 0 to 400 mM KNO 3 , the effect of nitrate corresponds to the progressive uncoupling of V 1 from V o which ends with the physical dissociation of the V 1 sector from the membrane. Note that even after treatment of epicotyl vesicles with 500 mM KNO 3 , ϳ50% of the V 1 sectors remain attached to the membrane, even though proton pumping is completely inhibited. Thus, physical release of the V 1 sector from the membrane is not required for inhibition of proton pumping, but may be required for inhibition of ATP hydrolysis.
Dschida and Bowman (31) proposed that the low concentration effect of nitrate was due to its oxidizing properties and induced the formation of disulfide cross-links in V 1 subunits, conformational changes, and subsequent release of V 1 sectors from the membrane. Accordingly, the oxidative inactivation of proton pumping that we had observed in epicotyl vesicles may not have been due exclusively to disulfide bonding at the catalytic site (13,14,32,33), but could also have involved some oxidation-induced uncoupling of the enzyme (34). Furthermore, sulfhydryl groups of the ␥ subunit of the lettuce chloroplast F-ATPase have been shown to be involved in proton slip (35). Our results, obtained by simultaneously measuring bafilomycin-and NEM-sensitive H ϩ pumping and ATP hydrolysis at different time points during oxidative inactivation show that no H ϩ /ATP uncoupling was taking place during oxidative inactivation. Even though this does not preclude nitrate from having a specific oxidizing effect, it indicates that oxidative inactivation, as we were measuring it previously, does not uncouple the enzyme and probably takes place exclusively at the catalytic site.
Bafilomycin A 1 is thought to inhibit V-ATPases by binding to the 100-kDa subunit and/or the proteolipid of the V o sector, but its mode of action, as well as that of other macrolide antibiotics, is still unknown (36 -38). We found that, like oxidation, bafilomycin A 1 did not uncouple H ϩ pumping from ATP hydrolysis.
It is important to note that in the present study H ϩ /ATP FIG. 12. Oxidative inactivation of the H ؉ pumping and ATPase activities of the epicotyl V-ATPase. Tonoplast-enriched vesicles from epicotyls were incubated at 22°C in the absence of reductant, and their H ϩ pumping and ATP hydrolysis activities were assayed at different time intervals in the presence or absence of 50 nM bafilomycin A 1 or 200 M NEM. After ϳ4 h, 50 mM DTT were added to an aliquot of the oxidized membranes. A, activity measurements; ࡗ, H ϩ pumping activity; E, bafilomycin A 1 -sensitive ATPase activity; ⌬, NEM-sensitive ATPase activity; B, deduced coupling ratios; E, bafilomycin A 1 -sensitive measurements; q, NEM-sensitive measurements. Note: the H ϩ pumping activity was 100% sensitive to bafilomycin A 1 and NEM. coupling was measured in the absence of an electrical potential gradient and in the presence of a limited pH gradient developing during the initial few minutes of H ϩ pumping. H ϩ /ATP coupling ratios under these initial rate conditions might differ drastically from those encountered under the conditions prevailing in vivo, e.g. ⌬ ϭ ϩ20 mV; ⌬pH ϭ 2 to 4.5 pH units. According to Tu et al. (15), and our own findings (14), intrinsic uncoupling, or slip, is proportional to the pH gradient built up across the membrane. When slip is determined during the attainment of a steady state pH gradient, the fruit V-ATPase was shown to exhibit half as much slip as the epicotyl V-ATPase (14). Thus, a lower rate of slip in the presence of a pH gradient remains the most likely factor allowing the generation of a steeper ⌬pH by the fruit V-ATPase. Although the coupling ratios of the fruit and epicotyl V-ATPases measured in the absence of a pH gradient were equal, the fact that the V-ATPase of juice sacs retained some coupling in the presence of nitrate, whereas the epicotyl V-ATPase did not, may be significant biologically. Nitrate may be mimicking the effect of another stress, such as low cytosolic pH, which would tend to inactivate a normal V-ATPase. The ability of the fruit V-ATPase to adjust to adverse conditions may be a key factor in the overall regulation of vacuolar pH in lemon.