The Substitution of Calcium for Magnesium in H+,K+-ATPase Catalytic Cycle EVIDENCE FOR TWO ACTIONS OF DIVALENT CATIONS*

determine the role of divalent cations in the mechanism of the H+,K+-ATPase, we substituted calcium for magnesium, which the H+,K+-ATPase for phosphorylation from ATP and from POr. Calcium was chosen over other divalent cations assayed (barium and manganese) in the absence of magnesium, calcium activated ATP hydrol- ysis, generated sufficiently high levels of phosphoenzyme (573 pmol.mg“) from to study dephosphorylation, and inhibited K+-stimulated ATP hydrolysis. The Ca2+-ATPase activity of the H+,K+- ATPase was 40% of the basal Mg2+-ATPase activity. activating ATP hydrolysis hydrolysis. an Ki of by uptake in the presence of and In the presence of and K+, Ca2+ inhibited proton transport with an appar- ent affinity similar to the inhibition of the Mg2+,K+-ATPase activity. The site of calcium inhibition was on the exterior of the vesicle. These results suggest that calcium activates basal turnover and inhibits K+ stim- ulation of the H+,K+-ATPase by binding at a cytosolic divalent cation site. external Ca2+ on Mg-ATP-dependent proton transport. Assay conditions are discussed under "Experimen- tal Procedures." The K'/H' exchanger nigericin is abbreviated using NIG. Addition of nigericin reversed the signals indicating a proton gradient was present in the vesicles.

In order to determine the role of divalent cations in the reaction mechanism of the H+,K+-ATPase, we have substituted calcium for magnesium, which is required by the H+,K+-ATPase for phosphorylation from ATP and from POr. Calcium was chosen over other divalent cations assayed (barium and manganese) because in the absence of magnesium, calcium activated ATP hydrolysis, generated sufficiently high levels of phosphoenzyme (573 2 51 pmol.mg") from [-pSaP]ATP to study dephosphorylation, and inhibited K+-stimulated ATP hydrolysis. The Ca2+-ATPase activity of the H+,K+-ATPase was 40% of the basal Mg2+-ATPase activity. However, the Ca2+,K+-ATPase activity (minus the Ca2+ basal activity) was only 0.7% of the Mc,K+-ATPase, indicating that calcium could partially substitute for M e in activating ATP hydrolysis but not in K+ stimulation of ATP hydrolysis. Approximately 0.1 mM calcium inhibited 50% of the Mg2+-ATPase or MS+,K+-ATPase activities. Inhibition of Mg2+,K+-ATPase activity was not competitive with respect to K+. Inhibition by calcium of Mg2+,K+ activityp-nitrophenyl phosphatase activity was competitive with respect to Mg2+ with an apparent Ki of 0.27 mM. Proton transport measured by acridine orange uptake was not detected in the presence of Ca2+ and K+. In the presence of M S + and K+, Ca2+ inhibited proton transport with an apparent affinity similar to the inhibition of the Mg2+,K+-ATPase activity. The site of calcium inhibition was on the exterior of the vesicle. These results suggest that calcium activates basal turnover and inhibits K+ stimulation of the H+,K+-ATPase by binding at a cytosolic divalent cation site. The pseudo-first order rate constant for phosphoenzyme formation from 5 PM [y-32P]ATP was at least 22 times slower in the presence of calcium (0.015 s-I) than magnesium (>0.310 s-'). The Ca*EP (phosphoenzyme formed in the presence of Ca2+) formed dephosphorylated four to five times more slowly that the Mg*EP (phosphoenzyme formed in the presence of Mg2+) in the presence of 8 mM trans-1,2-diaminocyclohexane-N,N,N',N'-tetraacetic acid (CDTA) or 250 PM ATP. Approximately 10% of the Ca*EP formed was sensitive to a 100 mM KC1 chase compared with >85% of the Mg-EP. By comparing the transient kinetics of the phosphoenzyme formed in the presence of magnesium (Mg-EP) and calcium (Ca-EP), we found two actions of divalent cations on dephosphorylation. Dephosphorylation was three times faster with 8 mM CDTA than * 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. with 250 PM ATP, regardless of whether the EP was formed with Mg or Ca, suggesting that one action of the divalent cation, at 2 mM, is inhibition of dephosphorylation and that the divalent cation dissociates rapidly from this site. The persistence of difference in the rates of Mg*EP and Ca*EP dephosphorylation during the CDTA chase indicates that this inhibitory site is different from the divalent cation site required for catalysis. The Ca*EP also dephosphorylated at least 10-20 times more slowly than the Mg*EP in the presence of 10 mM KC1 with either 8 mM CDTA or 1 mM ATP. The inability of the Ca-EP to dephosphorylate in the presence of K+, compared with the Mg*EP, demonstrates a second action of the divalent cation: that the type of divalent cation which occupies the catalytic divalent cation site required for phosphorylation is important for the conformational transition to a K+sensitive phosphoenzyme and is distinct from the inhibitory divalent cation site previously mentioned. The slower rates of Ca-EP dephosphorylation, compared with the Mg*EP, in the presence of a chelator (with or without KCl), suggests that calcium is tightly bound to the divalent cation site of the phosphoenzyme and the occupation of this site by calcium causes slower phosphoenzyme kinetics.
Phosphorylating transport ATPases bind a variety of ligands during the course of phosphoenzyme intermediate (EP)' formation and breakdown. It has been hypothesized that binding of these ligands to phosphorylating ATPases produced changes in conformation that expose cation transport site(s) on the ATPase to the cytosolic (El) or extracellular (E2) surface K+ (3, 4). In the presence of extracytosolic K+, EzP dephosphorylates to form EzK+ at a rate that is faster than enzyme turnover (4). Under these conditions the rate-limiting step is assumed to occur after dephosphorylation but to precede phosphorylation. The dephosphorylation reaction is thought to require Mg2+ because phosphorylation from PO, and "0-PO4 medium exchange require M P (5,6).
The EzK+ form converts to the EIK+ form of the enzyme to release K+ into the cytosol, rebind H+, and to initiate a new catalytic cycle with the binding of ATP.
The biochemical data accumulated thus far, suggest a similarity in mechanism and structure for the H+,K+-ATPase and other enzyme phosphorylating ion transport ATPases, such as the Na+,K+ and sarcoplasmic reticulum Ca2+-ATPases (7-9). Four partial reactions are used here to describe the reaction cycle for H+,K+-ATPase.
Phosphoenzyme formation: Mp",H+ Conformational change from EIP to E2P: Dephosphorylation: Conformational change from Ez to El: These four reactions are common to EP ion translocating ATPases and have been elaborated for the Na+,K+ and sarcoplasmic reticulum CaZC-ATPases to account for additional partial reactions such as ion occlusion and passive ion fluxes (10, 11). The conformational changes EIP to EzP, and E2 to El are thought to be associated with the movement of ions from one side of the enzyme to the other. This paper will focus on the role of the divalent cation in the above partial reactions. The H+,K+-ATPase requires Mg2+ for phosphorylation from ATP and for phosphorylation from PO4 but there is no direct evidence that Mg2+ is required for conformational changes (3-5). To determine the MgZ+ requirement in the partial reactions of ATP hydrolysis, MgZ+ has been replaced with other divalent cations. Three criteria were used to select a probe of the H+,K+-ATPase divalent cation site. 1) The divalent cation should substitute for Mg2+ in the formation of phosphoenzyme. 2) The level of EP in the presence of the divalent cation should be comparable to the level of EP with M T .
3) The divalent cation should not substitute for Mg2+ during K+-stimulated enzyme turnover. In preliminary experiments Ba", Mn2+, and Ca2+ substituted for M$+ in the formation of phosphoenzyme (30,90, and 60% of the Mg level were, respectively) but only Ca2+ yielded high levels of EP and did not substitute for MgZ+ in K+-stimulated dephosphorylation.
In this report we describe the effect of replacing MgZ+ with Ca2+ as the divalent cation in the H+,K+-ATPase reaction cycle. We propose that Ca2+ has two types of effects on the H+,K+-ATPase. 1) Ca2+ activates ATP hydrolysis and substitutes for M$+ in the formation and dephosphorylation of the phosphoenzyme with slower kinetics. 2) Ca2+ inhibits M$+dependent hydrolysis, K+-stimulated hydrolysis of ATP or pNPP, and proton transport, and does not substitute for Mg2+ in the K+-stimulated dephosphorylation reaction of EP.

Reagents
[ Y -~~P ] A T P was purchased from Amersham Corp. and "vanadium free" &sodium ATP was obtained from Sigma. All other reagents were at least of reagent grade.
Methods Enzyme Preparation-The porcine gastric H+,K+-ATPase was prepared according to previously published methods (12). Briefly, scrapings from stomachs obtained at slaughter were homogenized in 0.25 M sucrose buffered with 4 mM Pipes-Tris, pH 7.4. The homogenate was centrifuged at 20,000 X g for 45 min with a Sorvall GSA rotor. The supernatant from the 20,000 X g spin was then spun at 100,000 X g for 1 h with a Beckman Ti-70 rotor. Pellets were resuspended immediately and placed on a 0.25 M sucrose; 0.25 M sucrose, 7% Ficoll; 30% sucrose step gradient in a Beckman Z60 zonal rotor. After 2.5 h at 59,000 rpm, the vesicles were harvested and the fraction from the 0.25 M sucrose; 0.25 M sucrose, 7% Ficoll interface (Gl) was used in these experiments. For ATPase and phosphoenzyme studies the G1 fraction was diluted 10-15 times in 2 mM Tris-C1, pH 7.4, buffer, spun at 100,000 X g for 1 h subsequently lyophilized (LG1) and stored at -80 "C. The LG1 material is considered to be freely permeable to ions since the ATPase activity is neither valinomycin or nigericin stimulated and proton transport was not detected after lyophilization. The specific activity of the LG1 preparation varied between 140-160 pmol of Pi .mg". h-'. 100% of the K+-stimulated ATP hydrolysis and 75% of the M%+-stimulated ATP hydrolysis are inhibited by SCH 28080, a specific inhibitor of the H+,K+-ATPase (17).
ATPase Assays-ATPase activity was measured at 37 "C with 5-10 pg of enzyme, and the PO4 liberated was determined by the method of Yoda and Hokin (13). Final reaction concentrations in 1-ml reaction volume were 2 mM Na3ATP, 20 mM Tris-C1, pH 7.4 or 7.2, 20 mM KC1 and 2 mM MgClZ, unless otherwise indicated. CaC1, concentrations varied from 25 p~ to 2 mM. In solutions with no added Ca", Caz+ was measured by atomic absorption, and the Caz+ concentration pNPPase Assays-pNPPase activity was measured at 37 "C with 10 pg of enzyme and the p-nitrophenol liberated was measured spectrophotometrically at 405 nm at a pH greater than 10. Final reaction concentrations in 1 ml were 6 mM pNPP, 20 mM Tris-C1, pH 7.2, 40 mM KC1 and 6 mM MgC12, unless otherwise indicated.
Phosphoenzyme formation was initiated with the addition of 100 pl of labeled ATP to a 900 or 800 pl (for dephosphorylation experiments) volume of protein preincubated with reaction solutions at room temperature for at least 10 min and then placed on ice for at least 10 min. In the dephosphorylation studies, ligands were added in a 100-p1 volume to a 900-p1 volume of protein preincubated with the labeled ATP for 15 s or 180 s for Mg.EP and Ca.EP, respectively. Reactions were stopped with the addition of a 500-pl ice-cold solution of 40% (w/v) trichloroacetic acid, 5 mm of H3PO4, and 5 mM unlabeled ATP. Trichloroacetic acid-precipitated proteins were filtered within 15 min of reaction termination using a 3 or 0.45 p M HAWP millipore filter. The filter support and test tube were washed four times with 5 ml of an ice-cold solution of 3% trichloroacetic acid and 5 mM H3P04. Filter papers were dried, dispensed into 10 ml of Amersham Corp. counting scintillate and counted using a LKB liquid scintillation counter.
Nonspecific labeling was measured as the amount of radioactive label incorporated in the presence of 4 mM CDTA, 4 mM EGTA, and 1 mM KC1. Incorporation was usually between 5-10% of the total phosphoenzyme generated with M$+ and was subtracted as a blank from the total Mg . EP or Ca . EP formed.
Transport Assays-Proton transport was measured using the weak base acridine orange as previously described (14). The pH-dependent absorbance shift at 492 nm was monitored with an Aminco DW-2 spectrophotometer at 22-24 "C. Vesicles were equilibrated at 4 "C with the appropriate ions for a minimum of 30 h. Final concentrations of dye and ions in the transport reaction media were 10 p~ acridine orange, 0.25 M sucrose, 4 mM Pipes-Tris, pH 7.4,l mM ATP, 150 mM KCl, and 1 mM MgClZ. Experimental changes in these conditions and was 3-6 pM.
internal vesicular concentrations of ions are indicated in the figure legends.
Protein Determination-Protein was measured by the method of Lowry et al. (15) using bovine serum albumin standards.
Calculations-The rate constants were calculated by a computer program provided by the Biomathematics unit of CURE. A single exponential provided a reasonably good fit of the data and the standard errors for the rate constants are given in Table 11. The data points and associated bars indicate the mean and standard error of two to six experiments. The data points in the figures are connected by an interpolation between the points.
Free Ca2+ concentrations were calculated, using values of Martell and Smith (16) for ligand dissociation constants, with a computer program kindly provided by Dr. D. D. F. Loo.  Table I. tracted) with an IC, of approximately 100 PM (Fig. 1).

Ca2+ Inhibition and Activation
In the presence of K+ and Mg+, the gastric H+,K+-ATPase demonstrates a stimulated activity (MP,K'-ATPase). Calcium completely inhibits the MP,K+-ATPase activity with (Ca2+,Mg2+-ATPase activity was subtracted) an ICw for Ca2+ of approximately 100 pM. Presumably Ca" could prevent monovalent cation stimulation by competitively displacing K+ from an activating site on the enzyme. The simplest expectation of such an interaction by Ca2+, is that the apparent Ki for Ca2+ would increase as the K+ concentration increased. This was not observed; the ICso for Ca2+ does not increase significantly over a 40-fold K+ concentration range extending from 0.5 to 20 mM (data not shown). Calcium also inhibited K+ stimulation of the MP,K+-pNPPase activity in a competitive manner with respect to M e with an apparent Ki of 0.27 mM (Fig. 1B).
In the absence of M P , with only Ca2+ present, there was no detectable K+ stimulation of either ATP or pNPP hydrolysis (Table I).
In contrast to the situation where Ca2+ inhibits the M$+activated enzyme, Ca2+ alone activates ATP hydrolysis (Ca2+-ATPase) with an activity equaling 40% of the MP-ATPase activity (Table I). Half-maximal activation of ATPase activity occurs between 0.4-0.7 mM CaC12. The ea2+-ATPase and Me-ATPase activities are also inhibited by 45 and 75%, respectively, by 60 PM SCH 28080, a specific inhibitor of the H+,K+-ATPase (17).

EP Studies: Inhibition of the Partial Reactions
Calcium Slows Phosphoenzyme Formation-The steady state hydrolysis measurements indicate that calcium can both activate and inhibit H+,K+-ATPase turnover. Kinetic analysis of the partial reactions of the gastric ATPase's phosphoenzyme further defines calcium inhibition and activation of hydrolysis. In the first step, the formation of phosphoenzyme: E + M e + ATP "-* ADP + Mg-EP h calcium forms phosphoenzyme from 5 PM ATP 22 times slower than M P (Fig. 2 and Table 11).
In the presence of 0.3 mM CDTA and in the absence of divalent cation, no phosphoenzyme formation occurs, indicating that either Caz+ or Mg+ are necessary to generate Ca2+ and M P phosphoenzyme (Ca.EP or Mg.EP, respectively).  The slower phosphorylation kinetics induced by Ca2+ was also observed by Tobin et al. (18) for the Na+,K+-ATPase. Slower phosphorylation of the H+,K+-ATPase ( k1 = 0.011 s-' f 0.001) in the presence of calcium was seen at 0.5 p~ ATP as well (19). Maximal Table 11. (dephosphorylation): Mg.EP + Hz0 -E + P + MgZ+ kz This reaction is slow relative to that found when K+ is included with Mg2+ in the reaction media (4). Dephosphorylation is measured after abolishing phosphorylation in one of two ways: 1) by diluting the labeled [y3'P]ATP 50-200-fold with the addition of unlabeled ATP or, 2) by the addition of a chelator in excess of the divalent cation concentration. Fig.  3, A and B, shows that the replacement of magnesium with calcium slows another partial reaction of the phosphoenzyme. In this case, dephosphorylation of the Ca. EP is four to five times slower than that of the Mg-EP, by either method of determining dephosphorylation. Comparison of the chelator and unlabeled ATP induced dephosphorylation rates for either the Ca2+ or M P , EPs assesses the effect of free divalent cation on the dephosphorylation rate (Fig. 3, A and B ) .
Apparently, chelator-dependent breakdown of either the Mg .
EP or the Ca-EP was faster than their corresponding unlabeled ATP-dependent breakdown ( Table II), suggesting that chelator accessible divalent cation inhibits the dephosphorylation reaction. The possibility that the slower unlabeled ATPinduced dephosphorylation rates were due to ATP inhibition of dephosphorylation is unlikely since the dephosphorylation rate did not vary as a function of concentration of unlabeled ATP added (25-1000 p~) , providing there was sufficient (250 PM) unlabeled ATP to observe dephosphorylation (data not shown). The lack of inhibition by ATP was also shown by Ray and Forte (22). Ca.EP Is ADP Insensitive-Since calcium slows phosphorylation, inhibits K+ stimulation of ATPase activity, and maintains a K+-insensitive EP (as discussed in the following section) an expected correlate would be the generation of an ADP-sensitive Ca. EP. This hypothesis was tested by forming EP in the presence of Ca2+ and subsequently adding 250 p~ ADP or 5 mM ADP plus 5 mM ATP. Since ADP addition could lead to EP dephosphorylation via two mechanisms: 1) by prevention [y3'P]ATP from forming E3'P or, 2) by reacting with previously formed C2P, the ADP induced dephosphorylation was compared with the unlabeled ATP dephosphorylation which measures the former mechanism. The Ca. EP did not dephosphorylate faster with ADP than with ATP, like the ADP-insensitive phosphoenzyme found by Rabon et al. (23) ( Table 11).
Ca,EP Is K' Insensitive-When K+ is added to the Mg. EP there is a rapid phase of dephosphorylation as a new Rate constants are for reactions on ice. Assay conditions are discussed under "Experimental Procedures" or in the figure legends. The partial reactions are divided into three groups: EP formation, dephosphorylation, and K+stimulated dephosphorylation. Some rates were determined on only 4 0 -1 5 % of the E P and although the fit of the data to a single exponential was good (3 = 0.99) these rate constants are probably underestimated being too fast to accurately determine (Figs. 2,5, A and B ) .  Fig. 2. In A the rate for Mg.EP dephosphorylation initiated with a chelator is compared with the rate initiated with excess ATP. In B the rate for Ca. EP dephosphorylation initiated with a chelator is compared with the rate initiated with excess ATP. Reactions are on ice. Assay conditions are discussed under "Experi- Table 11. mental Procedures" and the first order rate constants are found in steady state level of EP is approached (3, 4). Ca2+ prevents the initial rapid phase of K+-stimulated dephosphorylation of EP (Fig. 4A). Even after 120 s when a new steady state level of EP is reached only 10% of the Ca.EP is labile to 0.1 mM KC1 while 65% of the Mg-EP is K+ labile. Abolition of the rapid phase of K+-induced dephosphorylation of EP is more pronounced at 100 mM KC1. After 10 s only 15% of the Mg-EP remains while over 90% of the Ca EP is stable to these conditions (Fig. 4B). K+ stimulation of phosphoenzyme breakdown was also measured when phosphorylation was prevented by the addition of unlabeled ATP as shown in Fig. 5A. Under these conditions the Ca.EP dephosphorylates 10-20 times more slowly than the Mg.EP (Table 11). K' increases the Ca.EP rate of dephosphorylation measured in the presence of unlabeled ATP only 3-fold. This is in contrast to the Mg.EP where K' increases the dephosphorylation rate in the presence of unlabeled ATP greater than 20-fold.
To assess the effect of free divalent cation on K+ stimulation of EP dephosphorylation, phosphorylation from ATP was prevented by the removal of divalent cation using a chelator (Fig. 5B). Under these conditions the Ca-EP dephosphorylates at least 10-20 times more slowly than the Mg. EP (Table 11). The rate constant of dephosphorylation of the Ca. EP in the presence of a chelator is increased %fold by K+.
In contrast, K+ increases the rate constant for dephosphorylation of Mg.EP in the presence of a chelator at least 20-fold. Apparently Ca2+ blocks K+ stimulation of dephosphorylation of E2P even after chelation, suggesting either that Ca" dissociated from the phosphoenzyme, conferring slower K+-stimulated dephosphorylation kinetics or that Ca2+ tightly bound to the enzyme is responsible for the slower kinetics.

Inhibition of Proton Transport
The Sideness of Ca Inhibition-In the hydrolysis and phosphoenzyme experiments lyophilized vesicles were used, with access of Ca2+ to both the cytosolic and extracellular face of the enzyme. The predominantly inside-out orientation of the gastric G1 vesicle fraction was exploited to assess whether the Ca2+ inhibitory site is located on the lumenal or cytosolic face of the H+,K+-ATPase. The data described thus far points toward a cytosolic site of inhibition. The acridine orange uptake data in Fig. 6A show that external Ca2+ does not activate proton transport and that internal Ca2+ does not  Table 11. inhibit proton transport. Control vesicles equilibrated with 150 mM KC1 and diluted into 150 mM KC1 and 1 mM MgC12 show a rapid uptake of the lipophilic weak base acridine orange as a consequence of ATP addition. The change in absorbance indicates that protons and acridine orange were transported and trapped inside the vesicle lumen (recording A ) . Vesicles equilibrated with 150 mM KC1 and 1 mM CaC12 and diluted into 150 mM KC1 and 1 mM MgC12, displayed an acridine orange uptake response to the addition of ATP nearly identical to that of the control (recording B). The slightly slower rate of acridine orange uptake into the CaClz equilibrated vesicles is explained by external Ca2+ present from diluting the equilibrated vesicles into the assay solution. Vesicles with both internal and external KC1 and 1 mM external CaCL showed no proton transport. To ensure that CaClz was in the lumen of the vesicles, vesicles equilibrated with CaClz and KC1 were diluted into a KC1 assay solution. Addition of the Ca2+/'H+ exchanger A23187 caused proton uptake into the vesicles via the outwardly directed CaC12 gradient produced upon dilution of the vesicles into the assay solution. external Ca2+ for inhibition of proton transport indicates that the Ca2+ inhibitory site is on the cytosolic face of the enzyme.

DISCUSSION
The present investigation demonstrates the activating and inhibitory effects of Ca2+ on H+,K+-ATPase turnover and partial reactions of the ATPase.
Actiuation-The Ca2+-ATPase activity, the Ca2+-pNPPase activity, and the formation and dephosphorylation of Ca EP demonstrate that Ca2+ activates the H+,K+-ATPase (Figs. 2 and 3 and Table I). The activation of the H+,K+-ATPase by