Two Ca2+-dependent ATPases in THE PREVIOUSLY PURIFIED (Ca2’-M$’)-ATPase Rat Liver Plasma Membrane IS NOT A Ca2+-PUMP BUT AN ECTO-ATPase*

We have shown that the rat liver plasma membrane has at least two (Ca’+-M&+)-ATPases. One of them has the properties of a plasma membrane Ca2+-pump (Lin, S.-H. (1985) J. Biol. Chem. 260, 7850-7856); the other one, which we have purified (Lin, S.-H., and Fain, J. N. (1984) J. Biol. Chem. 259, 3016-3020) and characterized (Lin, S.-H. (1985) J. Biol. Chem. 260,10976-10980) has no established function. In this study we present evidence that the purified (Ca’+-M8’)-ATPase is a plasma membrane ecto-ATPase. In hepatocytes in primary culture, we can detect Ca’+-ATPase and M8+-ATPase activities by addition of ATP to the intact cells. The external localization of the active site of the ATPase was confirmed by the observation that the Ca2+-ATPase and M8+-ATPase activities were the same for intact cells, saponin- treated cells, and cell homogenates. Less than 14% of total intracellular lactate dehydrogenase, a cytosolic enzyme, was released during a 30-min incubation of the hepatocytes with 2 mM ATP. This indicates that the hepatocytes This paper reports the presence of an ecto-Ca'+- and M e stimulated ATPase in hepatocytes in primary culture. The external localization of the nucleotide-hydrolyzing site is supported by the observation that the Ca'+-ATPase and M%+-ATPase activities are the same for both intact cells, saponin-treated cells, and homogenized cells. The properties of this ecto-ATPase suggest that the previously purified high affinity (Ca'+-Mp)-ATPase of rat liver plasma membrane is the ecto-ATPase. First, the nucleotide specificity of the ecto-ATPase is the same as that of the purified enzyme; both of them are able to hydrolyze ATP, GTP, UTP, CTP, ADP, and GDP to a similar extent. Second, the ecto-ATPase activity can be activated by either Ca2+ or M%+, and the effects of Ca'+ and Mg2+ on this ecto-ATPase activity are not additive, indicating that both CaZ+- and Me-ATPase activities reside on the same enzyme. This is consistent with the property of the purified enzyme in which a nonadditive effect of Ca2+ and M P on the enzymatic activity was observed. Third, the ecto-ATPase, like the purified enzyme, is not affected by oligomycin, vanadate, N-ethylmaleimide, andp-chloromercuriben-zoate. Furthermore, the activities of both the ecto-ATPase and purified ATPase are quite insensitive to protease treatments. Consistent with this conclusion is the recent finding by molecular cloning and sequencing of the gene for the (Ca'+-Mg2+)-ATPase that in this protein there is one hydrophobic segment C-terminal end of the

The cytosolic free calcium concentration of hepatocytes is in the range of 0.1 to 0.2 PM (Murphy et al., 1980). It is proposed that part of the Ca2+ gradient is maintained by a high affinity ATP-dependent Ca2+ transporter localized in the plasma membrane. The Ca2+ pumps of human erythrocyte membrane and rat heart sarcolemma have been characterized and purified (Niggli et al., 1979;Caroni and Carafoli, 1981). In those tissues the plasma membrane Ca2+ pumps, like the muscle sarcoplasmic reticulum Ca2+ pump, possess ATPase activity which can be activated by Ca2+ in the presence of Mg2'. However, no Ca2+-stimulated ATPase activity could be found in rat liver plasma membrane under similar conditions. As a result, a high affinity Ca2+-stimulated ATPase activity which was observed in the absence of exogenously added Mg2' was thought to be the enzyme responsible for hepatocyte plasma membrane Ca2+ transport (Lotersztajn et al., 1981(Lotersztajn et al., , 1984. We have purified the high affinity Ca2+-stimulated ATPase (Lin and Fain, 1984). The properties of the high affinity (Ca2+-Mg2')-ATPase are different from those of the plasma membrane Ca2+ pump studied by reconstituting liver plasma membrane proteins into artificial liposomes (Lin, 1985a(Lin, , 1985b. Further characterization of this ATPase demonstrated that this enzyme can be activated by either Ca2+ or Mg+; and it was also shown that this enzyme has broad nucleotide specificity, and its activity is not inhibited by inhibitors of known ion transporters (Lin, 1985b).
Since the high affinity (Ca2+-Mg2')-ATPase is not the liver plasma membrane Ca2+ pump, its physiological function remains unknown. Recently, Charest et al. (1985) reported that stimulation of isolated hepatocytes with ATP or ADP induced a rapid but transient increase of free cytosolic Ca2+ concentration indicating the existence of P2-purinergic receptor(s) in the hepatocyte plasma membranes. Further study also showed that the transient response was probably due to rapid hydrolysis of the ATP by a plasma membrane ecto-ATPase. As part of the effort to study the function of the (Ca2+-M$+)-ATPase, we tested the possibility that this enzyme is a plasma membrane ecto-enzyme with its nucleotide hydrolyzing site facing the outside of the cell. In this communication, we present evidence that Ca2+-ATPase and MF-ATPase activity can be detected by addition of ATP to the outside of intact hepatocytes. The properties of this enzyme indicate that it is the same enzyme that we previously purified (Lin and Fain, Ecto-(Ca2+-Mg2+)-nucleotidme of Hepatocytes 1984) and characterized (Lin, 1985b). Our finding that the (Ca2+-Mg2+)-ATPase hydrolyzes extracellular ATP and ADP suggests that it may play a role in terminating the effect of ATP and ADP on hepatocyte Ca2+ mobilization.
Assay of ATPase Activities in Suspensions and Homogenutes of Hepatocytes-Hepatocytes were isolated from Sprague-Dawley rats (Charles River, CD strain) by collagenase digestion, as described by Seglen (1976). At the beginning of each experiment, the hepatocytes were washed twice with buffer A which contained 120 mM NaCl, 5 mM KCl, 20 mM Hepes/Tris (pH 7.4), and 2 mM EGTA. The hepatocytes were resuspended in buffer A, and half of them were homogenized with a Polytron homogenizer (Brinkmann Instruments, Westbury, NY) at speed 5 for 2 s. Aliquots of the hepatocytes or the homogenates were added to tubes containing a small amount of water (for basal ATPase activity measurement) or Ca" solution to give a final concentration of 2 mM (for Ca2+-ATPase activity), or M$+ solution to give a final concentration of 1 mM (for M%+-ATPase activity). After incubation of cells or homogenates at 37 "C for 10 min, the assay was started by addition of ATP to a final concentration of 2 mM. The reaction was stopped by adding SDS to a final concentration of 2%. The ATPase activity was determined by measuring the inorganic phosphate released as described by Ames (1966), except that the time for color development was 20 min at 37 "C instead of 1 h at 37 "C.
Cell Culture-Primary hepatocyte cultures were prepared from 200-250 g male Sprague-Dawley rats (Holtzman; Madison, WI) as described by Russell et al. (1984), except that after a 15-min attachment at 37 "C, the plating medium containing calf serum was replaced with 1.5 ml of Williams' Medium E containing 150 nM insulin. The hepatocytes in collagen-coated dishes (3.75 X 10' cells in 1.5 ml of Williams' Medium E) were incubated at 37 "C for 24-48 h before use.
Assay of ATPase Activities of Hepatocytes in Primary Culture and of Their Homogenates-At the beginning of each experiment, the hepatocytes in culture dishes were rinsed twice (2 ml each time) with buffer A (for basal ATPase activity measurement), or buffer A plus 2 mM Ca2+ (for Ca2+-ATPase activity measurement), or buffer A plus 1 mM M%+ (for M%+-ATPase activity measurement). For the measurement of ATPase activity in intact cells, 2 ml of each solution was added, and ATPase assays were started by addition of ATP to each dish to a final concentration of 2 mM. After incubation of cells at room temperature with ATP for different periods of time as indicated, aliquots of the incubation medium (200 pl) were taken and added to 0.3 ml of 10% SDS. The inorganic phosphate released was determined as described above.
For the measurement of ATPase activities in cell homogenates, the hepatocytes in culture dishes were rinsed twice with buffer A containing Ca2+, M%+, or nothing as described above. One ml of each solution was added to each dish, and the cells were scraped off dishes with a rubber policeman. Another 1 ml of each solution was added to rinse the remaining cells out of the dish. Both fractions were combined and frozen in liquid nitrogen. The cells were then disrupted by thawing at 37 'C followed by sonication for 1 min (Heat Systems, Ultrasonics, Inc., Farmingdale, NY) at setting 5. The ATPase assays were started by addition of ATP, and the inorganic phosphate released was determined as described above.
Since the ATPase activities of the intact cells were measured by removing aliquots of cell medium at different times, the volume of cell medium decreased without a proportional loss of cells. As a result, the inorganic phosphate measured for each different time point in the intact cell experiments was adjusted by this factor. For example, in Fig. 1A, each culture dish had 2 ml of cell medium, and 200 pl of cell medium was taken for inorganic phosphate determination at 0, 5, 10, 15, and 30 min. The amount of inorganic phosphate present at 0 min was used for background subtraction. The total inorganic phosphate released up to 5 min is simply the "measured inorganic phosphate" (Ms) multiplied with a volume factor (1.8/0.2), since only ' The abbreviations used are: AMP-PNP, 5'-adenylyl-P,y-imidodiphosphate; EGTA, ethylene bis(oxyethylenenitri1o)tetraacetic acid; Hepes, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid; SDS, sodium dodecyl sulfate. 0.2 ml out of a total of 1.8 ml of cell medium was measured. The total inorganic phosphate released in up to 10 min is the "measured inorganic phosphate" (MI,,) adjusted for a volume factor (1.6/0.2) plus the amount withdrawn for the previous measurement, i.e. M6. By the same token, M6 and Mlo should be added back to the measurement at 15 min after the adjustment was made for the volume factor. The ATPase activities of homogenates were measured on 0.2 ml aliquots of the cell homogenate withdrawn at different periods of time. Only a volume factor of 10 is required for the ATPase activity calculation in these cases.
Affinities for Ca2+ and Mg2+-The affinity for Ca2+ of the ecto-CaZ+-ATPase activity was determined by incubating the hepatocytes in primary culture with buffer A containing different amounts of Ca2+. The reactions were started by addition of ATP to a final concentration of 2 mM. The free Ca" concentrations were calculated with the association constant of KC~EGTA = 4.28 X lo7 M" and KCATP = 7.99 X lo3 M-' . The amount of M e contaminating the assay mixtures was 1 p~ as determined by atomic absorption. In the presence of 2 mM ATP, this amount of M%+ gives less than 0.02 p~ free M e .
The affinity for M%+ of the ecto-Me-ATPase activity was determined by incubating the hepatocytes in primary culture with buffer A containing different amounts of Mg+. The reactions were started by addition of ATP to a final concentration of 2 mM. The free M%+ concentrations were calculated with the association constant of Assay of Lactate Dehydrogenose Activity-The lactate dehydrogenase activity was assayed by addition of 100 p1 of cell medium or homogenate to 2 ml of solution containing 6.6 mM NADH, 0.2 M Tris-HC1, pH 7.4. One ml of the mixture was pipetted into each cuvette containing 33 pl of distilled water or 33 pl of 30 mM sodium pyruvate. The difference in absorbance at 340 nm was recorded continuously in a Kontron Uvikon 810 spectrophotometer. One unit represents the oxidation of 1 pmol of NADH/min at 25 "C.
Other Procedures-Liver plasma membranes were isolated from the livers of male Sprague-Dawley rats according to Neville's procedure up to step 10 (Neville, 1968). The membrane-bound (Ca*+-M%+)-ATPase was solubilized and purified from plasma membrane according to Lin and Fain (1984).

RESULTS
Ca2+-ATPase and Mg2+-ATPase Activities of Intact and Disrupted Hepatocytes-In the initial experiments, hepatocytes freshly prepared by liver perfusion and collagenase digestion were used. Although Ca2+-stimulated and Mg+stimulated ATPase activities could be detected by incubating intact hepatocytes with ATP, the lactate dehydrogenase activity of the cell supernatant was about 50% of that of the homogenate, indicating that freshly isolated hepatocytes were very leaky. In order to unambiguously demonstrate the sideness of this plasma membrane enzyme, a hepatocyte preparation with minimum membrane leakage was required. Therefore, hepatocytes in primary culture were used in later experiments. Hepatocytes in primary culture were prepared by collagenase perfusion of rat livers. The method employed yielded a cell suspension containing greater than 95% hepatocytes, which were then plated on collagen-coated dishes. Under this condition, viable hepatocytes attach to the dish and the majority of damaged cells are washed off the dish at the beginning of the experiments. These results indicate that the hepatocytes in primary culture were still intact after 30 min of incubation with 2 mM ATP, and the Ca2+-ATPase and Me-ATPase activities measured in the whole cell preparation were due to cell surface ATPase activities.
In some experiments, the Mp-ATPase activity in the disrupted cell preparations was much higher than that of intact cells (about 4-fold higher), although the Ca2+-ATPase activity was the same for both intact and disrupted cell preparations (data not shown). The higher M$+-ATPase activity in the disrupted cell preparations probably is mitochondrial ATPase which became accessible after disrupting the cells by freeze-thaw and sonication. This view is supported by the observation that when oligomycin (50 pg/ml) was included in the assay media, the Mg+-ATPase activities of both intact cells and disrupted cells became the same (data not shown). These results also indicate that the ecto-M$+-ATPase activity is different from mitochondrial ATPase which is oligomycin-sensitive.
One might suppose that the M$+-and Ca2+-ATPase activities observed with intact cells might derive from a small population of dead cells with permeable plasma membrane. In this case, the observation that Ca2+-ATPase and M e -ATPase activities are the same for both intact and sonically disrupted cells could be inconclusive if sonication were to generate membrane vesicles with the same orientation as in intact cells and give no further exposure of the inner surface of the plasma membrane to the substrate, ATP. To test this hypothesis, hepatocytes in primary culture were also disrupted by addition of 40 pg/ml saponin, which is known to permeabilize hepatocyte plasma membranes without disrupting mitochondria and endoplasmic reticulum (Joseph et al., 1984). As shown in Fig. 2, the Ca2+-ATPase and Me-ATPase activities in intact, saponin-treated, and homogenized cells are about the same. The lactate dehydrogenase activities at the end of a 60-min incubation were 0.10,1.49, and 0.80 units/ lo6 cells for intact, saponin-treated, and homogenized cells, respectively. This indicates that, under the conditions used for ATPase assays, no further Ca2+-ATPase and M e -ATPase can be detected by permeabilizing the cell.
Nucleotide Specificity of Ecto-ATPase Activities-In order to test the possibility that the ecto-Ca2+-ATPase and M$+-ATPase may be the same protein we previously purified (Lin and Fain, 1984) and characterized (Lin, 1985b), several properties of the previously characterized high affinity (Ca2+-M<)-nucleotidase were determined with the enzyme in intact hepatocytes. The nucleotide specificity of the ecto-ATPase activities is shown in Table I. Both Ca2+-and Mg+stimulated activities have broad substrate specificities. The relative nucleotide-hydrolyzing rates are about the same as that of the purified (Ca2+-Me)-ATPase (Lin, 1985b) except with AMP and AMP-PNP. The hydrolysis of AMP may be due to the presence of 5'-nucleotidase, which is known to be a hepatocytye ecto-enzyme, in the intact cell. The hydrolysis of AMP-PNP in the presence of Ca2+ may be due to the sequential hydrolysis of AMP-PNP by the Ca2+-dependent adenosine triphosphate pyrophosphohydrolase activity (Flodgaard and Torp-Pedersen, 1978) and the 5"nucleotidase.
Effect of Proteases on Ecto-ATPase and Purified ATPase Activities-Since the ATP-hydrolyzing site of the ecto-ATPase is extracellular, it was interesting to see whether proteolysis would destroy the ATPase activity from the outside of the cells. As shown in Table 11, the ecto-ATPase activity of hepatocytes in primary culture is not sensitive to B, the hepatocytes in primary culture were incubated with buffer A plus 50 pg/ml oligomycin, 40 pg/ml saponin, and 2 mM Ca2+ (O), or buffer A plus 50 pg/ml oligomycin, 40 pg/ml saponin, and 1 mM MgZf (0). The ATPase activities were determined as for intact cells. C, the hepatocytes were removed from culture dishes and homogenized as described under "Experimental Procedures" with buffer A plus 50 pg/ ml oligomycin and 2 mM Ca2+ (O), or with buffer A plus 50 pg/ml oligomycin and 1 mM MgZf (0). The ATPase activities were measured as described under "Experimental Procedures.''

TABLE I Substrate specificity of ecto-Ca2+-ATPase and Mg2f-ATPase
The ecto-Ca2+-ATPase activity was measured by incubating hepatocytes in primary culture with buffer A plus 2 mM Ca2+. The ecto-Me-ATPase activity was measured by incubating hepatocytes in primary culture with buffer A plus 1 mM M e . The reactions were started by adding different nucleotides to a final concentration of 2 mM at room temperature, and the incubation time was 30 min.

TABLE I1 Effect of protease treatments on hepatocyte ecto-Ca2+-ATPase activity and on purified (Ca2+-M&ATPase activity
The effect of protease treatments on ecto-Ca2+-ATPase activity was measured by incubating hepatocytes in primary culture with buffer A plus 2 mM Ca2+ and various amounts of trypsin, chymotrypsin, or papain for 1 h at room temperature. The ATPase assays were started by adding ATP to a final concentration of 2 mM and further incubated for 30 min at room temperature. The effect of protease treatments on purified (Ca2+-Me)-ATPase activity was performed in a similar condition as described above, except that 2 pg of purified (Ca2+-MgZ+)-ATPase was used and 0.2 mg/ml C12E9 was included in the reaction mixtures. The results represent the average of duplicate determinations. trypsin, chymotrypsin, or papain treatment. Although treatment of the hepatocytes in primary culture with proteases causes dissociation of the cells from the collagen-coated culture dishes, less than 25% of the ecto-Ca2+-ATPase activity was lost with 50 kg/ml trypsin, chymotrypsin, or papain for 90 min at room temperature. When the purified high affinity (Ca2+-Mg2')-ATPase was treated with proteases under the same conditions, the purified Ca2+-ATPase activity was similarly insensitive to protease treatment (Table 11). Furthermore, after trypsin treatment of the purified enzyme, its size was not changed upon examination by SDS-polyacrylamide gel electrophoresis (data not shown). It is possible that the nucleotide-hydrolyzing site of the ecto-ATPase might be released from the plasma membrane by the proteases. To test this hypothesis, purified liver plasma membranes were treated with 50 Fg/ml trypsin, chymotrypsin, or papain, respectively, at 37 "C for 1 h. The membranes were then spun down with an Airfuge for 5 min, and the Ca2+-ATPase activities were measured in both supernatant and pellet fractions. No significant amount of Ca2+-ATPase activity was detected in the supernatant fraction (data not shown), suggesting that the ecto-ATPase enzyme was still membranebound after the protease treatment.
Effect of CaZ+ and Mg2+ on Ecto-ATPase Actiuity-The high affinity (Ca2+-M%+)-ATPase of rat liver plasma membrane can be activated by either Ca2+ or M%+ (Lin, 1985b). Addition of both Ca2+ and M e to the ATPase assay medium gave the same ATPase activity as M e alone in both whole cell and disrupted cell preparations (Fig. 1, A and B ) . The nonadditive effect of Ca2+ and Mg2' indicates that both activities are from the same enzyme.
As shown in Fig. 3A, in the absence of M g + (the free Mg+ concentration was less than 0.02 p~) the ecto-ATPase activity of intact hepatocytes is stimulated by Ca2+ in a concentrationdependent fashion. The Ca2+ concentration dependence curve is best fitted by one component which has the affinity for Ca2+ of 5.2 PM. In the absence of added Ca2+, the ecto-ATPase activity was stimulated by M P , with K d for M P of around 5 p~ free M e (Fig. 3B). DISCUSSION This paper reports the presence of an ecto-Ca'+-and M estimulated ATPase in hepatocytes in primary culture. The external localization of the nucleotide-hydrolyzing site is supported by the observation that the Ca'+-ATPase and M%+-ATPase activities are the same for both intact cells, saponintreated cells, and homogenized cells. The properties of this ecto-ATPase suggest that the previously purified high affinity (Ca'+-Mp)-ATPase of rat liver plasma membrane is the ecto-ATPase. First, the nucleotide specificity of the ecto-ATPase is the same as that of the purified enzyme; both of them are able to hydrolyze ATP, GTP, UTP, CTP, ADP, and GDP to a similar extent. Second, the ecto-ATPase activity can be activated by either Ca2+ or M%+, and the effects of Ca'+ and Mg2+ on this ecto-ATPase activity are not additive, indicating that both CaZ+-and Me-ATPase activities reside on the same enzyme. This is consistent with the property of the purified enzyme in which a nonadditive effect of Ca2+ and M P on the enzymatic activity was observed. Third, the ecto-ATPase, like the purified enzyme, is not affected by oligomycin, vanadate, N-ethylmaleimide, andp-chloromercuribenzoate. Furthermore, the activities of both the ecto-ATPase and purified ATPase are quite insensitive to protease treatments. Consistent with this conclusion is the recent finding by molecular cloning and sequencing of the gene for the (Ca'+-Mg2+)-ATPase that in this protein there is one hydrophobic segment which is localized near the C-terminal end of the protein and that there are many putative N-glycosylation sites in the rest of the protein.' The molecular arrangement of this protein is similar to that of other membrane ectoenzymes which have been studied, i.e. alkaline phosphatase (Millan, 1986;Kam et al., 1985;Henthorn et al., 1986;Ovitt et al., 1986;Berger et al., 1987) and y-glutamyltransferase (Laperche et al., 1986).
One approach to determine the sideness of a membrane protein is to use specific antibodies. If the enzymatic reaction can be inhibited in the intact cell by an antibody specific for the enzyme, then one can conclude that the enzyme has its active site located on the outside of the cell. In order to use such an approach, an antibody against purified (Ca2+-Mg2+)-ATPase was prepared by inoculating purified protein into the popliteal lymph nodes of a rabbit (Sigel et al., 1983). The antiserum reacts with the purified (Ca"-M$+)-ATPase by immunoblotting (Towbin et al., 1979) and immunoprecipitation. By indirect immunofluorescence, the antibodies recognize a protein localized in the bile canalicular domain of hepatocytes.' The antiserum binds to the detergent (polyoxyethylene 9-lauryl ether)-solubilized, enzymatically active form of the (Ca2+-Mg2+)-ATPase and depletes the (Ca'+-M$+)-ATPase activity from solution when the antibodyenzyme complex was precipitated by protein A-Sepharose. The (Ca'+-M$+)-ATPase activity was found to be associated with the precipitates. However, the antiserum inhibits the (Ca'+-MgZ+)-ATPase activity weakly, indicating that the antibodies may not recognize the nucleotide-hydrolyzing site of the enzyme. A series of monoclonal antibodies against the purified ecto-ATPase was also prepared, and none of the monoclonal antibodies inhibited the ATPase activity.
The insensitivity of the ecto-ATPase activity to protease treatments is quite surprising. Besides trypsin, chymotrypsin, and papain, other proteases (subtilisin, elastase, Staphylococ- cus aureus V8 protease, thermolysin, submaxillary protease, pronase, bromelain, and Streptomyces griseus protease) were also tested for their effect on the Ca'+-ATPase and M$+-ATPase activity. In each case, there was no significant loss (less than 20%) of Ca'+-ATPase or Me-ATPase activity. The high resistance of the ecto-ATPase activity to proteolysis may be a protective mechanism of the enzyme against extracellular proteases. Since the ecto-ATPase is a glycoprotein (Lin and Fain, 1984), it is possible that the insensitivity of the active site to proteolysis is due to the presence of carbohydrate on the protein. The nucleotide sequence of the cDNA for the gene of the ecto-ATPase shows that there are more than 15 potential asparagine glycosylation sites in the ecto-ATPase protein. ' The ecto-ATPase of liver plasma membranes was first observed as a high affinity (Ca2+-M$+)-ATPase with a dissociation constant, K d , for ca'+ in the range of 0.01 to 0.2 pM (Lotersztajn et al., 1981;Iwasa et al., 1982). The values of K d of the detergent-solubilized and purified enzyme are 0.09 and 0.16 p~ for Ca'+ and MgZ+, respectively (Lin, 198513). In hepatocytes in primary culture, however, the K d values for Ca2+ and M$+ are 5.2 and 5.0 pM, respectively, which are higher than those of the purified enzyme. The factor(s) which gives such a difference between the purified enzyme and the enzyme in intact cells is not clear. It has been reported that there is an inhibitor (Lotersztajn and Pecker, 1982;Lotersztajn et al., 1985) and an activator (Lotersztajn et al., 1981) for the high affinity (Ca'+-MP)-ATPase of rat liver plasma membranes. Whether these two putative regulators are involved in the change of affinity for Ca'+ is unknown. Also it is possible that a GTP-binding protein may be involved in such a phenomenon. Consistent with this hypothesis is the recent finding by Lotersztajn et al. (1987) that the effect of glucagon on the (Ca'+-Me)-ATPase activity is mediated by a GTP-binding protein as revealed by the sensitivity of the effect to cholera toxin. Furthermore, the change of affinity for Ca'+ may be due to the difference between the oxidationreduction status of the intracellular medium and the preparation medium. Finally, the local Ca'+ and M%+ concentrations for the enzyme may be quite different between the intact cells and the solubilized enzyme. Results from molecular cloning and sequencing of the gene for this ecto-ATPase showed that the C terminus of the postulated sequence for the ecto-ATPase contains a unique CAMP-dependent serine phosphorylation consensus sequence (Lys-Arg-X-X-Ser) (Krebs and Beavo, 1979).' Whether phosphorylation or dephosphorylation of the ecto-ATPase is involved in changing the affinity of this enzyme for Ca'+ and Mg2+ is presently under investigation. As a result of the unexpected lower affinity for Ca'+ in intact cells, the concentration of Ca'+ used in the measurement of Ca2+-ATPase activity, i.e. 1.6 p~ free [ca2+], is lower than the K d of the enzyme for Ca'+ in intact cells, while the concentration of M P used, i.e. 25 p~ free [MgZ+], is close to the concentration for maximal activation of Mg2"ATPase activity. This accounts for the observation that the M%+-ATPase activity is higher than the Ca'+-ATPase activity in several studies. In fact, in the presence of saturating amounts of Ca'+ ( i e . 100 p~ free Ca2+) or Mg2+ ( i e . 200 p~ free M e ) , the Ca2+-ATPase and MP-ATPase activities are the same (data not shown).
Most tissues contain ecto-ATPase activity which can be detected as a Ca2+-stimulated ATPase activity. This ecto-ATPase activity on the plasma membrane is higher than Ca'+-pump ATPase activity. In rat liver plasma membrane, the ATP-hydrolyzing activity of the ecto-ATPase is about 10 times that of the Ca'+-pump ATPase (Lin, 1985a;Lin and Fain, 1984). And in rat corpus luteum, a 1000-fold difference between the transport rate and the rate of the Ca2+-ATPase was observed (Minami and Penniston, 1987). As a result, there is great confusion in deducing the significance of plasma membrane Ca2+-stimulated ATPase activity. In the plasma membrane of several tissues, i.e. rat liver (Lin, 1985b), rat corpus luteum (Minami and Penniston, 1987), rat kidney (Ghijsen et al., 1984) and intestinal basolateral membrane (Moy et al., 1986), and neutrophil plasma membranes (Ochs and Reed, 1984), two different Ca2+-ATPase activities have been reported. In a recent report by Pavoine et al. (1987), it was claimed that the high affinity (Ca2+-Me)-ATPase of liver plasma membrane is a Ca2+ pump by reconstituting partially purified (Ca2+-Mg2+)-ATPase into artificial liposomes and demonstrating a small amount of Ca2+ transport activity in such a preparation. The (Caz'-M8+)-ATPase preparation in that study, however, was not highly purified, since the specific activity of the enzyme was only 20-fold greater than that of plasma membrane (Lotersztajn et al., 1981). In our previous study, a 300-fold purification was obtained (Lin and Fain, 1984). As a result, it is possible that the Ca2+pumping activity observed in the partially purified preparation may be due to contamination of the preparation with the Ca2+-pump protein. Furthermore, the nucleotide specificity and vanadate sensitivity of the transport system, two critical criteria which distinguish the Ca2+ pump from the (Ca2+-Me)-ATPase, were not reported in that study. Therefore, the claim that the (Ca2+-Mg2+)-ATPase is a Ca2+ pump is dubious. The function of the (Ca2+-Me)-ATPase which is not a calcium pump is not known.
Similar ecto-(Ca2+-Mg2+)-ATPase activities have been found in several other tissues (for references see Lin, 1985b). Several observations (Charest et al., 1985;Dubyak and De Young, 1985) suggest that possible roles for the ecto-ATPase may be to terminate the effect of ATP on the cells or to participate in the ATP effect via its phosphatase activity. It was also interesting to find that the Pz-purinergic effect has broad nucleotide specificity (Dubyak and De Young, 1985;Buxton et al., 1986;Okajima et al., 1987) as does the nucleotide-hydrolyzing activity of the ecto-ATPase. These correlations between the properties of the Pz-purinergic effect and the ecto-ATPase activity raise the possibility that the ecto-ATPase protein may be the Pz-purinergic receptor. In the 5774 mouse macrophage cell line, Steinberg and Silverstein (1987) showed that ecto-ATPase does not mediate the effects of ATP on these cells. In hepatocytes, this possibility is presently under investigation. In this report, we show that the major (Ca2+-Mp)-ATPase in the plasma membrane is a membrane ecto-ATPase and suggest that its function may be in the regulation of Pz-purinergic receptor function. This result should be helpful in clarifying the confusion about the plasma membrane (Ca2+-M2+)-ATPases. , . , Towbin, H. T., Staehelin, T., and Gordon, J. (1979) Proc. Natl. Acad.