Oligomeric regulation of gastric H+,K+-ATPase.

The H+,K+-ATPase of intact gastric vesicles has two Km values for ATP hydrolysis, 7 and 80 microM. Irradiation of vesicles with ultraviolet light in the presence of 1 mM ATP resulted in K+-ATPase activity that shows only the low affinity ATP binding. The irradiation stimulated or inhibited proton uptake rate compared with control vesicles at high or low ATP concentrations, respectively. The relation between proton uptake rate and K+-ATPase activity at different ATP concentrations was linear with irradiated vesicles and nonlinear with control vesicles. These results indicate that hydrolysis at the high affinity ATP binding site regulates the energy-transport coupling in negative and positive manners at high and low ATP concentrations, respectively. The complete inhibition of K+-ATPase by a specific proton pump inhibitor E3810 (rabeprazole) (2-([4-(3-methoxypropoxy)-3-methylpyridin-2-yl]methylsulf i nyl)-1H-benzimidazole sodium salt) occurred when E3810 bound to half of the alpha-subunit of H+,K+-ATPase in unirradiated vesicles at both 200 and 10 microM ATP, whereas the complete inhibition of proton uptake occurred when E3810 bound to half or a quarter of the alpha-subunit at 200 or 10 microM ATP, respectively. These results suggest that dimeric interaction between the alpha-subunits is necessary for the enzyme activity at all ATP concentrations and that dimeric or tetrameric interaction is necessary for proton transport at high or low ATP concentrations, respectively.

The H ϩ ,K ϩ -ATPase is involved in the final step of gastric acid secretion and secretes H ϩ in exchange for K ϩ (1-6). The gastric H ϩ ,K ϩ -ATPase is classified as a P-type ATPase, closely related to the Na ϩ ,K ϩ -ATPase (7). The H ϩ ,K ϩ -ATPase consists of ␣and ␤-subunits. The ␣-subunit is a catalytic subunit, and the ␤-subunit is suggested to maintain integrity of the enzyme structure and affect ATP hydrolyzing activity (8). The H ϩ ,K ϩ -ATPase has high and low K m values for ATP hydrolysis (3). Several different explanations are possible for the presence of high and low K m values (9). One possible explanation is that the oligomeric interaction of (␣␤)-protomer results in two different affinities for ATP. Diffraction pattern analysis of the electron microscopic image of two-dimensional crystals of the enzyme provided evidence that the H ϩ ,K ϩ -ATPase exists as a dimer (10) or tetramer (11) of the ␣-subunit. So far, there is no information as to whether close ␣-␣-subunit contact and interaction are necessary for the enzyme activity and active transport of proton and potassium. In Na ϩ ,K ϩ -ATPase, specific cy-toplasmic regions of the ␣-subunit have been shown to be necessary for ␣-␣ contact (12)(13)(14), and the ␣-subunit does not associate with H ϩ ,K ϩ -ATPase ␣-subunit (12,13).
In this paper, we studied the biochemical basis of two K m values for ATP hydrolysis by gastric H ϩ ,K ϩ -ATPase and found that ATP hydrolysis at the high affinity site regulated proton uptake with negative and positive cooperativity at high and low ATP concentrations, respectively. Furthermore, from measurements of the relations between ATP hydrolysis, proton transport rate and the amount of specific binding of a proton pump inhibitor E3810, 1 we found that at high ATP concentrations the functions of ATP hydrolysis and proton transport require the dimeric interaction between ␣-subunits, but at low ATP concentrations, the function of ATP hydrolysis requires the dimeric interaction, whereas that of proton uptake requires the tetrameric interaction.

MATERIALS AND METHODS
Chemicals and Drugs-E3810 (rabeprazole), [␣-methylene- 14 C] E3810 and [benzimidazole-2- 14 C]E3810 were obtained from Eisai Co. (Tokyo, Japan). The specific activity and the radiochemical purity of [␣-methylene- 14  Preparation of Hog Gastric Vesicles-Tightly sealed membrane vesicles that contain H ϩ ,K ϩ -ATPase were prepared from hog stomachs as described previously (15). Gastric vesicles in 250 mM sucrose solution were stored at Ϫ85°C and used within 1 month. Protein concentration was determined by the method of Lowry et al. (16) with bovine serum albumin as standard.
Measurement of Mg 2ϩ -activated ADP Hydrolyzing Activity-In preliminary experiments, we determined the Mg 2ϩ -activated ADP hydrolyzing activity of gastric vesicles, which was previously unknown. This activity was measured in a pyruvate kinase-lactate dehydrogenaseadenylate kinase-linked reaction. In the reaction, 2 mol of ATP are regenerated from 2 mol of ADP that are produced from 1 mol of AMP and 1 mol of ATP by adenylate kinase, and these reactions are coupled with oxidation of 2 mol of NADH (17,18). The reaction mixture containing 10 g/ml of gastric vesicles, 40 mM Tris/HCl (pH 7.40), 250 mM sucrose, 2 mM MgCl 2 , 100 M NADH, and 10 M ADP was incubated for various time at 25°C. Then, 0.8 mM phosphoenolpyruvate, 4 units/ml pyruvate kinase, 10 units/ml lactate dehydrogenase, and then 15 mM KCl were added. The decrease in the amount of NADH was measured with an Aminco DW-2C UV-visible spectrophotometer in a dual wavelength mode at 340 and 500 nm at 25°C. Then, 0.8 units/ml of adenylate kinase was added, and the decrease in the amount of NADH was measured. The initial decrease after the addition of KCl corresponds to * This work was supported in part by Grants-in-aid from the Ministry of Education, Science, and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed. Tel.: 81-764-34-2281 (ext. 2668); Fax: 81-764-34-5051; E-mail: morii@ms.toyama-mpu.ac.jp. the amount of remaining ADP, and the subsequent decrease after the addition of adenylate kinase corresponds to the amount of hydrolyzed AMP. ADP hydrolyzing activity was calibrated by the addition of a known amount of AMP to a reaction mixture. Mg 2ϩ -activated ADP hydrolyzing activity was expressed as mol of liberated inorganic phosphate/mg of protein⅐h.
Measurement of K ϩ -ATPase Activity-The K ϩ -activated ATPase activity was measured in a pyruvate kinase-lactate dehydrogenase-linked reaction where hydrolysis of ATP is coupled with oxidation of NADH (19,20). The reaction mixture contained 10 g/ml of gastric vesicles, 40 mM Tris/HCl (pH 7.40), 150 mM KCl, 10 g/ml of valinomycin, 2 mM MgCl 2 , 160 M NADH, 0.8 mM phosphoenolpyruvate, 100 M AMP, 3 units/ml pyruvate kinase, 2.75 units/ml lactate dehydrogenase, and various concentrations of ATP. When indicated, 2 g/ml of nigericin and 15 mM KCl plus 135 mM choline chloride were used in place of 10 g/ml of valinomycin and 150 mM KCl. The decrease in the amount of NADH was measured spectrophotometrically at 25°C. ATPase activity was calibrated by the addition of a known amount of ADP to a reaction mixture. Since the presence of 100 M AMP was found to prevent hydrolysis of ADP to AMP, 100 M AMP was added in the reaction mixture. When AMP was absent, ATP concentration decreased with time even in the presence of the ATP regenerating system. Mg 2ϩ -ATPase activity has been measured usually in a K ϩ -free solution, but K ϩ -free condition is not available in the coupled enzyme method because K ϩ is necessary for pyruvate kinase reaction. A K ϩ -competitive H ϩ ,K ϩ -ATPase inhibitor, SCH 28080, completely inhibited the K ϩ -ATPase activity and 38% of Mg 2ϩ -ATPase activity (21,22). Using this information, we measured the ATPase activity in the presence of 2 mM MgCl 2 , 1 mM KCl, 10 M SCH 28080, and 250 mM sucrose and in the absence of valinomycin. The remaining ATPase activity was assumed to be equal to 62% of the Mg 2ϩ -ATPase activity as described previously (23). Although this method to evaluate Mg 2ϩ -ATPase activity still contains some ambiguity, it did not introduce significant errors in estimating K ϩ -ATPase activity, because Mg 2ϩ -ATPase activity was small compared with total Mg 2ϩ plus K ϩ -ATPase activity. Resultant Mg 2ϩ -ATPase activity was comparable with the value obtained by the method of Yoda and Hokin (24). K ϩ -ATPase activity was defined as the difference between Mg 2ϩ plus K ϩ -ATPase activity and Mg 2ϩ -ATPase activity.
Measurement of Proton Transport Rate-Proton uptake into gastric vesicles was measured from quench of fluorescence of acridine orange (25). The reaction mixture contained 10 g/ml of gastric vesicles, 40 mM Tris/HCl (pH 7.40), 150 mM KCl, 10 g/ml of valinomycin, 2 mM MgCl 2 , 100 M AMP, 0.8 mM phosphoenolpyruvate, 4 units/ml pyruvate kinase, 5 M acridine orange, and various concentrations of ATP. Fluorescence of acridine orange was measured with a Hitachi 650 -10S spectrofluorometer or with a Shimadzu RF-5000 spectrofluorophotometer at 25°C (E x ϭ 495 nm, E m ϭ 525 nm). The rate of proton uptake was determined as the initial slope of fluorescence quench.
Specific Binding of E3810 to H ϩ ,K ϩ -ATPase-To evaluate the specific labeling of E3810 to the ␣-subunit of H ϩ ,K ϩ -ATPase, gastric vesicles (1 mg/ml) were incubated with 5 M [benzimidazole-2-14 C]E3810 in a solution containing 2 mM ATP, 2 mM MgCl 2 , 150 mM KCl, 10 g/ml valinomycin, and 40 mM Tris/HCl (pH 7.40) for 20 min at 25°C. The intravesicular space of gastric vesicles mimics the acidic lumen of gastric glands and was acidified when Mg-ATP was added into the vesicle solution containing valinomycin and a high concentration of K ϩ .
[benzimidazole-2-14 C]E3810 was accumulated in the acidic intravesicular space, activated by acid, and bound to a Cys residue of H ϩ ,K ϩ -ATPase from the luminal side (23). Then, the reaction was stopped by the addition of an equal volume of the electrophoresis sample buffer solution containing 30 mM Tris/HCl (pH 6.75), 4% SDS, and 20% glycerol. Immediately, SDS-PAGE of the E3810-labeled gastric vesicles was performed using Tris-glycine-SDS buffer and 7.5-15% gradient gel following the method of Laemmli (26). The SDS-solubilized sample solution was warmed to 40°C for solubilization of precipitated potassium dodecyl sulfate and applied to a well in the gel. ␤-Mercaptoethanol was omitted during electrophoresis to avoid dissociation of bound E3810 from a Cys residue of H ϩ ,K ϩ -ATPase. Then, the gel was stained with Coomassie Brilliant Blue R-250, and protein distribution was assayed with a densitometer. The gel was sliced into 2-mm-wide pieces. The sliced gel pieces were solubilized in 0.5 ml of 30% hydrogen peroxide solution in tightly capped vials for 20 h at 60°C. The amount of bound [benzimidazole-2-14 C]E3810 was determined with a liquid scintillation counter.
For determination of the ratio of bound H ϩ ,K ϩ -ATPase inhibitor, E3810, to the ␣-subunit of H ϩ ,K ϩ -ATPase, gastric vesicles were specifically labeled with [␣-methylene-14 C]E3810. For specific labeling, gastric vesicles (1 mg/ml) were incubated with 2 mM ATP in a solution containing 2 mM MgCl 2 , 150 mM KCl, 10 g/ml valinomycin, and 40 mM Tris/HCl (pH 7.40) for 20 min in the presence of various concentrations of [␣-methylene-14 C]E3810. Reaction was stopped by gel filtration through a Sephadex G-50 column equilibrated with 150 mM KCl and 40 mM Tris/HCl (pH 7.40). The gel filtration dissipated the proton gradient across the vesicle membrane, resulting in decomposition of the activated E3810, and unreacted [␣-methylene-14 C]E3810 was removed. After the redetermination of protein concentration of the filtrate by means of absorbance at 260 nm, the amount of bound [␣-methylene- 14 C]E3810/ mg of vesicle protein was determined with a liquid scintillation counter. K ϩ -ATPase and proton uptake activities of E3810-labeled vesicles were also measured at low (10 M) and high (200 M) ATP concentrations. The ␣-subunit of H ϩ ,K ϩ -ATPase comprised 48.3% of the total vesicle protein determined by densitometry of SDS-PAGE gels stained with Coomassie Brilliant Blue R-250.
UV Light Irradiation of Gastric Vesicles-A gastric vesicle solution (1 mg/ml) containing 2 mM MgCl 2 , 250 mM sucrose, and 40 mM Tris/HCl (pH 7.40) in the presence or absence of 1 mM ATP was irradiated with UV light (260 nm) for 1 h at 25°C using a Hitachi 650 -10S spectrofluorometer (150-watt Xenon lamp with 20 nm excitation slit width). The possibility of irreversible ATP binding to the enzyme was evaluated using [8-14 C]ATP under the same conditions. The UV light-irradiated gastric vesicle solution (0.1 ml) was added to 1 ml of cold stop solution containing 10% trichloroacetic acid solution and 1% sodium pyrophosphate, and precipitated enzymes were collected on a 0.45-m Millipore filter (HAWP02500). The precipitated enzymes were washed with 10 ml of cold stop solution. The amount of bound 14 C-labeled ATP was measured with a scintillation counter.

RESULTS
Mg 2ϩ -activated ADP-hydrolyzing Activity of Gastric Vesicles-To obtain the exact relation between the ATP concentration and K ϩ -ATPase activity of gastric vesicles, it is necessary to maintain constant ATP concentrations during the measurement of the enzyme activity. Although the coupled enzyme method to measure ATPase activity employs an ATP regenerating system (phosphoenolpyruvate and pyruvate kinase), we found that the ATP concentration was not constant in the gastric vesicle suspension. The Mg 2ϩ plus K ϩ -ATPase activity at 10 M ATP gradually decreased (data not shown). Preincubation of vesicles in the absence of 2 mM Mg 2ϩ for 1 h did not decrease the Mg 2ϩ plus K ϩ -ATPase activity at 10 M ATP, but preincubation in the presence of 2 mM Mg 2ϩ and absence of K ϩ for 1 h induced the same extent of decrease in the ATPase activity as that observed in the presence of Mg 2ϩ plus K ϩ . These results suggest the possibility that ADP was not completely converted to ATP, and part of ADP was further hydrolyzed to AMP in a Mg 2ϩ -dependent and K ϩ -independent manner. The hydrolysis of ADP to AMP can be detected by an adenylate kinase-catalyzed reaction linked to the coupled enzyme method where 2 mol of ADP is produced from 1 mol of AMP and 1 mol of ATP. The hydrolyzing activity of ADP to AMP by gastric vesicles was 1.45 Ϯ 0.58 mol of P i /mg⅐h (mean Ϯ S.E., n ϭ 3), which was about half of the Mg 2ϩ -ATPase activity of gastric vesicles measured at 10 M ATP. To keep the Mg 2ϩ plus K ϩ -ATPase activity constant over a long time, it was found that the addition of 100 M AMP was effective, possibly because the high concentration of AMP shifts an enzymatic reaction equilibrium between ADP and AMP toward ADP.
Here we did not further explore the mechanism of the ADP hydrolyzing activity.
[ATP] Dependence of K ϩ -ATPase Activity-K ϩ -ATPase activity was measured at various ATP concentrations (2-1,000 M) in the presence of 100 M AMP, 10 g/ml gastric vesicles, 2 mM MgCl 2 , 150 mM KCl, 10 g/ml valinomycin, and 40 mM Tris/HCl (pH 7.40). The ATP concentration was kept constant during the reaction by inclusion of phosphoenolpyruvate, pyruvate kinase, and 100 M AMP. The double-reciprocal plot between K ϩ -ATPase activity and ATP concentration yielded two K m values, 7 and 80 M ATP in the presence of proton and potassium gradients and a membrane potential (Fig. 1). However, in the presence of 2 g/ml of nigericin, which dissipated K ϩ and H ϩ gradients across the vesicle membrane, only one K m value (5 M) was observed (Fig. 1). We confirmed the existence of a potassium gradient across the vesicle membrane, in the presence of valinomycin, by inhibition of K ϩ -ATPase with SCH 28080 at 20 M, a specific reversible inhibitor that competes with K ϩ at the luminal high affinity binding site, this inhibition being reversed by the addition of 2 g/ml of nigericin (data not shown).
Specific Inhibition of the High Affinity ATP Binding by UV Light Irradiation-Irradiation of vesicles with UV light (260 nm) in the absence of ATP induced almost complete inhibition of the K ϩ -ATPase activity of gastric vesicles, whereas irradiation in the presence of 1 mM ATP produced only minor inhibition (Fig. 2). Fig. 1 shows that the high affinity ATP binding is inhibited and the low affinity ATP binding is completely pro-tected from UV light irradiation in the presence of 1 mM ATP. The values of K m for the low affinity ATP binding and V max did not change from control values. When UV light-irradiated vesicles were used, the binding of [8][9][10][11][12][13][14] C]ATP to the enzyme was not detected, indicating that UV light irradiation in the presence of ATP did not induce the irreversible binding of ATP to the high affinity ATP binding site of the enzyme. We also examined the effect of UV light irradiation on proton uptake rate into gastric vesicles. The rate of proton uptake into the UV light-irradiated vesicles was enhanced at high ATP concentrations (Ͼ100 M) but inhibited at low ATP concentrations (Ͻ20 M) when compared with control vesicles (Fig. 3). In Fig. 3A, the differences in proton uptake at 5 and 10 M ATP were significant, and those at 200-1000 M ATP were not significant (three observations). Increasing the number of observations at 500 M ATP to nine resulted in the difference becoming significant.
The relationship between K ϩ -ATPase activity and proton uptake rate was not linear in control vesicles, but it was linear in UV light-irradiated vesicles (Fig. 4), suggesting that ATP hydrolysis at the high affinity ATP binding site regulates the coupling between K ϩ -ATPase activity and proton uptake rate in negative and positive cooperative manners at high and low ATP concentrations, respectively.
Effects of AMP-PNP-We studied the effect of a nonhydrolyzing ATP analog AMP-PNP on K ϩ -ATPase activity at 200 M ATP. The presence of 10 M AMP-PNP enhanced K ϩ -ATPase activity in control vesicles but decreased it in UV light-irradiated vesicles (Table I). We discuss this AMP-PNP effect later.
The Relations between the Amount of Bound E3810 and K ϩ -ATPase Activity and between the Amount of Bound E3810 and Proton Uptake Rate-Gastric vesicles were labeled with a proton pump inhibitor [␣-methylene-14 C]E3810 according to the method described under "Materials and Methods." E3810 at 5 M inhibited about 80% of K ϩ -ATPase activity. Fig. 5 shows distribution of gastric vesicle proteins and 14 C-counts in sliced gels, indicating that E3810 bound only to the ␣-subunit of H ϩ ,K ϩ -ATPase (98.7%). The relations between the amounts of bound E3810, K ϩ -ATPase, and proton uptake rate at 200 and 10 M ATP were measured. At 200 M ATP, K ϩ -ATPase activity was completely inhibited when 2.6 nmol of E3810 bound per mg of vesicle protein, and proton uptake was completely inhibited when 2.1 nmol of E3810 bound per mg of vesicle proteins (Fig. 6). E3810 binding increased further to 8.6 nmol/mg of vesicle proteins, more than required for complete inhibition of proton uptake, probably due to nonspecific binding. At 10 M ATP, K ϩ -ATPase activity was completely inhibited when 2.5  nmol of E3810 bound per mg of vesicle proteins, whereas proton uptake was completely inhibited when 1.0 nmol of E3810 bound per mg of vesicle proteins (Fig. 7). The ratio of bound E3810/␣-subunit was calculated taking into consideration that E3810 binds only to the ␣-subunit of H ϩ ,K ϩ -ATPase, the ␣-subunit comprised 48.3% vesicle proteins and the mass of the ␣-subunit is 114 kDa; for example, 2.1 nmol/mg of proteins gives the ratio of 0.50 (ϭ 2.1 ϫ 10 Ϫ9 ϫ 1.14 ϫ 10 5 /(10 Ϫ3 ϫ 0.483)). Table II shows the ratios necessary for complete inhibitions of K ϩ -ATPase activity and proton uptake at 200 and 10 M ATP. These results suggest that binding of one molecule of E3810 can arrest the K ϩ -ATPase activity of two molecules of the ␣-subunit at the both high and low ATP concentrations, and it can arrest protein uptake of two or four molecules of the ␣-subunit at the high or low ATP concentration, respectively.

DISCUSSION
The gastric H ϩ ,K ϩ -ATPase consists of ␣and ␤-subunits. Crystallization of H ϩ ,K ϩ -ATPase has shown that two (9) or four (10) monomers comprised the crystalline unit, which may suggest the presence of ␣-␣-subunit contact and interaction, as shown in Na ϩ ,K ϩ -ATPase (12)(13)(14). So far there has been no report whether the ␣-␣-subunit interaction of H ϩ ,K ϩ -ATPase is involved in the function of the enzyme activity and the active ion transport. The present study using a specific proton pump inhibitor E3810 has demonstrated that the ratio of bound E3810/␣-subunit necessary for the complete inhibition of K ϩ -ATPase activity is about 0.6 at both high and low ATP concentrations, and the ratio necessary for the complete inhibition of proton uptake is 0.5 or 0.24 at high or low ATP concentrations, respectively (Table II). That is, when E3810 binds to one of two ␣-subunits, both subunits lose the catalytic activity at every ATP concentration. When E3810 binds to one of two ␣-subunits, both subunits lose the proton transporting activity at high ATP concentration, and when E3810 binds to one of four ␣-subunits, all four subunits lose the proton transporting activity at low ATP concentrations. These results strongly suggest that the functional unit for K ϩ -ATPase activity is the (␣␤) 2 -diprotomer at every ATP concentration, whereas the functional unit for proton transport is (␣␤) 2 -diprotomer at high ATP concentrations and the (␣␤) 4 -tetraprotomer at low ATP concentrations.
When H ϩ ,K ϩ -ATPase was irradiated by UV light in the presence of ATP, the high affinity ATP binding was inhibited, whereas the low affinity ATP binding remained intact (Fig. 1). In these irradiated vesicles, proton uptake rate was stimulated at high ATP concentrations and inhibited at low ATP concentrations compared with those of control vesicles (Fig. 3). Furthermore, control vesicles show negative or positive cooperative coupling between the K ϩ -ATPase activity and proton uptake rate at high or low ATP concentrations, respectively, whereas the UV light-irradiated vesicles show no cooperativity (Fig. 4). These results suggest that the high affinity ATP binding site has a regulatory role in addition to that of ATP hydrolysis.
The presence of AMP-PNP enhanced K ϩ -ATPase activity in control vesicles but decreased it in UV light-irradiated vesicles. We propose the following explanation. Although AMP-PNP competes with ATP and inhibits ATP hydrolysis at both ATP binding sites in control vesicles, competition at the high affinity site diminishes the negative cooperativity, resulting in a net increase of K ϩ -ATPase activity. In UV light-irradiated vesicles, which have only the low affinity ATP binding site, AMP-PNP competition with ATP inhibits K ϩ -ATPase activity.
We suggest that the presence of high and low affinity ATP binding sites in H ϩ ,K ϩ -ATPase is due to the dimeric ␣-␣ interaction shown in this study. Other explanations, however, are also possible as previously discussed for Na ϩ ,K ϩ -ATPase (9,27). For example, solubilized ␣␤ Na ϩ ,K ϩ -ATPase protomer had two different affinities for ATP (27). We have shown that H ϩ ,K ϩ -ATPase has only one ATP binding affinity (K m ϭ 5 M) when measured in the presence of nigericin, which dissipates the proton and potassium ion gradients across the vesicle membrane. This result shows that each ␣-subunit has a single ATP binding site that is involved in hydrolysis. Our explanation clearly differs from a previous one that the nucleotide binds to H ϩ ,K ϩ -ATPase with different affinities in the sequential reaction cycle, E⅐ATP⅐H ϩ and EP⅐H ϩ ⅐ADP (or EP⅐H ϩ ⅐ATP) (28).
In conclusion, we propose that the dimeric or tetrameric subunit interaction is necessary for the function of H ϩ ,K ϩ -ATPase, and two different affinities for ATP of this enzyme are due to the dimeric subunit interaction, which is sensitive to UV light irradiation and pH and/or K ϩ gradient across the vesicle membrane.