A High Affinity Ca 2 +-stimulated and M $ +-dependent ATPase in Rat Corpus Luteum Plasma Membrane Fractions

Plasma membrane fractions from rat corpus luteum contain two kinds of Ca2+-stimulated ATPase, one having a high affinity for Ca2+, the other a low affinity for Ca2+. The high affinity ATPase had a specific Ca2+ requirement with a K I / Z of 0.2 to 0.3 PM; it had a V, of 105 nmol min" mg" and distributed, upon subcellular fractionation, with recognized plasma membrane enzymes. The properties of this enzyme indicate that it is a Ca2+ extrusion pump. The low affinity pump (KlI2 for Caz+, about 15 PM) was nonspecific, being stimulated equally well by Ca2+ or M&+; its function is unknown. Although the high affinity ATPase resembled the erythrocyte Ca2+-pumping ATPase in the properties mentioned above, it differed in that it failed to respond to Mg2+ or calmodulin. The lack of response to M&+ was due to the enzyme's retention of endogenous M&+; it did, after incubation with chelators, show a M&+ requirement. However, we were unable to show any effect of added calmodulin or trifluoperazine. This failure may be related to the high content of tightly bound calmodulin in these membranes. Much of this calmodulin could not be extracted even by washing with 1 mM EGTA and/or 0.1% (w/v) Triton X-100. This enzyme, the erythrocyte enzyme, and the adipocyte plasma membrane Ca2+ ATPase all belong to the class of Ca2+ ATPases with plasma membrane distribution and high affinity for Ca2+, indicating that they are Ca2+ extrusion pumps. However, the data indicate that tissue-specific differences exist within this class, with the enzyme from adipocytes and rat corpus luteum belonging to a subclass in which the requirement for Mg2+ and any response to calmodulin are difficult to demonstrate.

well characterized in the case of red blood cell plasma membranes (1,2), where Ca"+"g2+ ATPase has been purified to homogenity (3) and reconstituted into functionally active vesicles (4). Similar high affinity Ca"+"g2+ ATPase have also been identified in plasma membranes from brain (5, 6), adipocytes (7), pancreatic islets (8), and sperm (9). By analogy to erythrocytes, these enzymes have been suggested to be responsible for a calcium pump in the above mentioned tissues. Another mechanism utilized by the cell to pump calcium out of the cell is via Ca:Na exchange (10, 11) as described in the case of excitable tissue.
In the female reproductive system, no concrete information is available on the role of calcium during the various metabolic events, and it is not known whether a high affinity Ca"+"g2+ ATPase is present in female ovarian tissue.
In a series of papers, Bramley and Ryan (12)(13)(14)(15) have described the properties of surface membrane fractions, isolated from the rat ovaries at various stages of the reproductive cycle, as modified by injecting the appropriate hormones. For example, fractionation of granulosa cells yields a plasma membrane fraction (d. 1.16 to 1.18) which is enriched in Mg2+dependent ATPase, 5'-nucleotidase, and much of the adenylate cyclase of the ovary. The adenylate cyclase activity of these membranes was always accompanied by a low level of gonadotropin binding activity. Luteinization of the follicle, following a hCG injection, results in a marked increase in the microvillous region of the luteal cells. When heavily luteinized ovaries are fractionated, two surface membrane fractions can be isolated. A light membrane fraction (membrane from microvillous surface, d. 1.12 to 1.13), which consists of small vesicles, is enriched in hCG binding, 5'-nucleotidase, Mg2+dependent ATPase, and (Na + K)-ATPase activities, but possesses little adenylate cyclase. In contrast, the heavy membrane fraction (basolateral surface membrane, d. 1.16 to 1.18), which consists of large membrane sheets with junctional complexes, is enriched in the basal, hCG, and fluoride-stimulated adenylate cyclase, but not enriched in other plasma membrane markers. Thus, heavy membranes contain both the adenylate cyclase as well as a low level of gonadotropin binding activity, while light membranes have considerable hCG binding activity with little adenylate cyclase.
In this paper, we report the presence and some properties of a high affinity Ca'+-stimulated and Mg'+-dependent ATPase in the two plasma membrane fractions isolated from rat corpus luteum.' maximal velocity; PMSG, pregnant mare serum gonadotropin; hCG, human chorionic gonadotropin; DCCD, dicyclohexyl carbodiimide; d., buoyant density g/cm."; SER, smooth endoplasmic reticulum.
' A preliminary report of this data was presented as a poster at The International Symposium on Calcium-binding Proteins and Calcium Function in Health and Disease, June 8 to 12, 1980, Madison, Wisconsin.

MATERIALS AND METHODS
Animals-Young, immature female Holtzman rats, 22 to 25 days old and weighing between 90 and 100 g were obtained from the Holtzman Company, Madison, WI.
Hormones-Pregnant mare serum gonadotropin was obtained from the Sigma Chemical Company, St. Louis, MO. Human chorionic gonadotropin was purchased from Ayrest, New York, NY.
Chemicals-Carrier-free [y-'"P]ATP was obtained from New England Nuclear, Boston, MA. Chemicals used in enzymic assays were obtained from the Sigma Chemical Company, St.
Louis, MO. All other chemicals were of reagent grade.
Treatment of Animals-Immature, 22to 25-day-old female Holtzman rats were injected subcutaneously with 50 IU of PMSG in 0.5 ml of 0.9%' NaCI, followed 63 h later with 50 IU of hCG, also in 0.5 ml of 0.9%' NaCI. Rats were sacrificed by cervical dislocation on the eighth day after hCG injection, and ovaries were quickly removed. Injection of hCG, after priming with PMSG, results in extensive luteinization of ovaries as described before (12,13). Z.solation of Light and Hemy Plasma Membranes from PMSG and hCG Primed Ocaries-Luteinized ovaries were fractionated essentially as described before (13,14). Eight days after the injection of hCG, rats were sacrificed by cervical dislocation, ovaries were removed and freed of fat and connective tissues, blotted dry, weighed, minced with a scissor and homogenized in 20 volumes of 0.3 M sucrose in 10 mM Tris-HCI buffer, pH 7.4, containing 1 mM EDTA, using eight complete strokes of a loose Dounce homogenizer. EDTA was omitted when membranes were isolated for determination of calmodulin contents. After filtering through four layers of wet cheesecloth, homogenate was subjected to differential centrifugation. A systematic outline of this procedure is described in Reference 15. After each primary centrifugation, sediment was gently suspended in 0.3 M sucrose in 10 mM Tris-HCI buffer, pH 7.4, containing I mM EDTA. The resuspended pellet and the supernatant were then centrifuged at the same speed and for the same time again for maximal separation of each organelle and minimal contamination by the others. Thus, 800 X g average (10 min) and 20,000 X g average (20 min) pellets were prepared and used for isolation of heavy and light plasma membrane fractions, respectively.
Density Gradient Centrifugation-All the sucrose solutions were prepared in 10 mM Tris-HCI buffer, pH 7.4, with or without 1 mM EDTA and their concentration adjusted exactly prior to use with an Abbe refractometer.
Healy Membranes-The 800 X g pellet suspended in 20% sucrose was layered on the top of a discontinuous sucrose gradient containing 5 ml of 30%, 8 ml of 36%, 8 ml of 40R, and 5 ml of 50%' sucrose solutions. T h e gradients were centrifuged at 63,000 X g average for 60 min in a Beckman SW27 rotor. Materials accumulating at the interface between 30% and 36'? sucrose were collected, diluted four times with 40 mM Tris-HCI buffer, pH 7.4, and sedimented a t 10,000 X g average for 10 min. The final pellet was gently resuspended thoroughly in 0.3 M sucrose in 10 mM Tris-HCI buffer, pH 7.4, and stored in different aliquots in a liquid nitrogen bath.
Light Membranes-The 20,000 X g pellet suspended in 207 sucrose was layered on the top of a continuous gradient, prepared using 12 ml of 20% and 12 ml of 55% sucrose solutions, according to the method of Stone (16). After the centrifugation a t 63,000 X g average for 4 h in a SW27 rotor, the gradient was fractionated from the top of the tube using a Buchler-Searle auto-densi flow gradient fractionator equipped with a meniscus-sensitive probe. Sucrose concentrations were monitored by a refractometer, and fractions corresponding to 27 to 32't sucrose were pooled, diluted four times with 40 mM Tris-HCI buffer, pH 7.4, and pelleted at 100,000 X p average for 30 min. The final pellet was gently resuspended thoroughly in 10 mM Tris-HCI, pH 7.4, and stored in different aliquots in a liquid nitrogen bath. and Doty (17). In some experiments. 2 mM EGTA was replaced by 2 MgCl' (20 p~) . calmodulin. and chlorpromazine or its derivative were also present. EGTA and CDTA solutions were standardized by titrating with a standard CaCL solution (Radiometer, Copenhagen) and using a calcium-sensitive electrode. CaCL solutions used were checked for their exact concentration using standard EGTA solution and a calcium-sensitive electrode.
Calcium-stimulated ATPase activity was expressed as nanornoles of ATP split per min per mg of membrane protein used.
Lou, Affinity Ca" or Mg2+ ATPase Assay-The low affinity enzyme was measured using buffer and pH conditions similar to those described by Shami and Radde (18) for placental membranes. A reaction mixture of 0.5 ml contained 20 mM Tris-HCI, pH 8.4, 0.1 mM ouabain, 0 to 10 mM MgC1, or CaCli (free Mi2+, 0 to 5.02 mM, free Ca", 0 to 5.01 mM, at 1 mM total divalent cation concentration: free c a 2 + , 5.0 pM and free Mg", 3.0 pM), 5 mM [ y "1'1A1'P (0.1 nCi/nmol), and membrane proteins, 20 to 80 pg. Incubations were done at 37°C for 30-min periods. The procedure to monitor the release of was the same as that used in the assay of high affinity enzyme. Divalent cation-stimulated activity was obtained by subtracting blanks (no protein) and was expressed as nanomoles of ATP split per min per mg of protein used.
Other Enzymes-5'-Nucleotidase was measured by the method described by Bramley and Ryan (12), except released inorganic phosphate was monitored by extraction into the organic phase followed by the colorimetric assay using stannous chloride as the reducing agent (18).
Washing of Membranes with Triton X-100 and/or EGTA-Two different washing procedures were used for this purpose. In one, membranes were washed in the cold (4°C) with 1 mM EGTA, followed by 0.1 mM EGTA and at last with buffer (IO mM TES-TEA, pH 7.4) alone. In the second procedure, solutions employed in order of successive washings were 0.02'% (w/v) Triton X-100, 1 n m EGTA and buffer (as above). Membranes were centrifuged at 100,000 X g average for 20 min after each washing.
Calmodulin Assay-Calmodulin contents of membranes were measured by the method of Wang and Desai (19). This method is based on the ability of calmodulin to stimulate the activity of calmodulin-requiring phosphodiesterase. Phosphodiesterase activity was measured by dephosphorylating the 5'-AMP produced with 5"nucleotidase; the inorganic,phosphate released was determined by the method described in the ATPase assay. The membrane fractions for calmodulin determination were heated in a boiling water bath for 3 min, immediately transferred to ice, and the precipitated protein was removed by centrifugation at 100,000 X g average for 15 min. The supernatant containing released calmodulin was used in the phosphodiesterase assay. The reaction mixture contained 40 mM Tris-HCI. pH 7.5, 40 mM imidazole, 3 nlM magnesium acetate, 0 . 1 mM CalCiuItl chloride, 1.2 mM 3',5'-cyclic AMP, 0.2 unit of 5'-nucleotidase, and 20 pg of calmodulin requiring phosphodiesterase in a total volume of 0.5 ml.
Protein was determined by the method of Lowry et al. (20), using bovine serum albumin as a standard. Free calcium ion concentrations were calculated by means of a computer program (21) which took into account all complexes involving magnesium, calcium, EGTA, CIITA, and ATP. The calcium ion-sensitive electrode was standardized against optimally buffered solutions of EIITA and calcium chloride in which Cad+ was varied by setting the pH at different values.

Identification of the Calcium-dependent ATPase Activity
In the initial experiments, where 6 mM MgCL was used, a n increase in activity on adding Cay+ was not detected because of the high level of low specificity Cao+ or Mg" ATPase activity present in these membrane fractions (12). A Ca''-Mg'+ ATPase activity was evident, when low or no MgClr was employed in the assay mixture. From the results shown in There was no significant difference in Ca2+-dependent ATPase activity or its profile (specific activity versus calcium concentration) if 50 p~ MgC12 (total) was included in the reaction mixture for the ATPase assay. Even in the presence of 50 PM total big2+, the free Mg'+ concentration was zero over most of the range of ea'+ concentrations shown in Fig. 1, except in the last part of the low affinity curve where it did not exceed 4 PM. Note that the free Mg2+ concentration we are talking about here is calculated presuming that the Mg2+ content of membranes is zero. Later we will show that this apparent lack of any Mg2+ requirement occurs because enough Mg2+ is present in these plasma membranes to satisfy the requirement of high affinity Ca"-dependent ATPase.
Low Affinity Divalent Metal Ion (Ca2+ or Mg") Dependent ATPase Fig. 2 shows that calcium and magnesium were equally effective in stimulating the low affinity ATPase in the rat corpus luteum plasma membranes. Here also the V,,,, value for the light membranes was higher than that of the heavy membranes. The Ca2+ concentrations required for half-maximum activity were about 1.4 and 1.0 mM (total) for light and heavy membrane fractions, respectively. Magnesium appeared to be slightly more effective than calcium. The concentrations of Mg'+ required for half-maximum activity were 1.0 and 0.8 mM (total) for light and heavy membrane fractions, respectively. In the membrane fractions, the presence of this Ca"-or Mg"-dependent ATPase activity made the quantitation of high affinity Ca2+-ATPase data more difficult. No further attempts were made to characterize this low affinity enzyme.

Subcellular Localization of the High Affinity Ca2+-Mgz+ ATPase
The fractionation procedure to isolate light and heavy plasma membranes used here has been published previously (12,13), and their enzymatic characterization has been done in detail (12)(13)(14). From the above studies it was found that the marker enzymes for light membranes are 5'-nucleotidase, Na-K ATPase, and '"I-hCG binding activity while heavy membranes are not rich in the above markers. The adenylate cyclase activity was found to be enriched in the heavy membranes but not in the light membranes. Thus, the heavy membrane fraction can be classified as a surface membrane fraction primarily on the basis of its adenylate cyclase enrichment over the homogenate. This fraction also contains a low level of gonadotropin receptor activity, which can stimulate adenylate cyclase in the presence of hCG (13,14). In Table I data are shown to compare the distribution of Ca2+-Mg2+ ATPase and 5'-nucleotidase during the fractionation of the PMSG and hCG primed rat ovarian homogenate. The per cent recoveries of Ca2+-Mg2+ ATPase and 5-nucleotidase in the heavy membranes were found to be 4.69 and 4.01, respectively, from the whole homogenate, and enrichment of specific activities over that of homogenate was 1.65-and 1.38-fold, respectively. In the light membrane, Ca2+-Mg2+ ATPase and 5'-nucleotidase were purified 4.34-and 4.52-fold, respectively, when compared to whole homogenate. The per cent recoveries of the above enzymes in this fraction were 5.35 and 5.56, respectively. Thus, the two enzymes were purified to the same extent in the light membrane fraction, while neither was purified significantly in the heavy membrane fraction. Table   I1 compares the specific activities and fold purifications of various enzymes of two membrane fractions.

Mitochondrial and SER Enzyme Activities in the Plasma Membrane Fractions
Mitochondria-NADH cytochrome c reductase has been used as a mitochondrial marker by Bramley and Ryan in the rat luteal subcellular fractions (12). Based on the marker enzymic analysis for the major subcellular organelles and electron microscopic examination of the fractions, these authors have predicted that the mitochondrial contaminations of light and heavy membranes are in the range of 1 to 2% and 8 to 1676, respectively..

Comparison of Ca2+-Mg2+ ATPase and 5'-nucleotidase distribution in subcellular fractions of rut corpus luteum homogenate
The data shown are mean f % range of two membrane preparations. Each sample is assayed for enzymic activity in triplicate. Enzyme assays are done as described under "Materials and Methods." Free Ca" concentration used in ATPase assay is 0.29 p~. This relatively low activity of the high Ca'+ affinity ATPase.
[Ca"] was chosen to avoid any contribution of low Ca" affinity ATPase to the assay, but it gave an activity of only about half the maximal

Magnesium Requirements of the High Affinity ea2+
ATPase The results indicate (Fig. 1) that the addition of external magnesium (50 ~L M total) does not affect the Ca2+-Mg'+ A T Pase activity appreciably. The high affinity Ca"-Mg" ATPases in the red blood cell (3,4) and the brain ( 5 ) require magnesium for the expression of their activity. In the case of Ca" ATPase of the rat corpus luteum membranes, there are two possibilities. Either this enzyme does not require magnesium or the magnesium content of the membrane (isolated, even in the presence of 1 mM EDTA) is enough to satisfy the Mg'+ requirement for the expression of the enzyme activity. T o t e s t the above possibilities, the enzyme was assayed using the Ca -EGTA and Ca. CDTA systems. EGTA chelates Ca" more effectively than Mg'+ while CDTA is equally effective in chelating both the divalent cations. Thus, when EGTA is present in the reaction mixture free Mg'+ is available. In contrast when CDTA is used all the Mg'+ can be completely complexed. Fig. 3 shows the results of an experiment in which the specific activities of Ca'+-Mg'+ ATPase were determined at various concentrations of free Ca", using the C a -E G T A and Ca.CDTA systems. The specific activity of Ca"-Mg" ATPase was considerably reduced when free Mg2' was che- ase activity using Ca.EGTA system where free MgL+ is available; A, 0, represents enzymatic activity under conditions when free Mg'+ is complexed. Bar represents % range of the mean. Ca'+ concentration plotted on abscissa is free Ca'+ concentration measured by a calcium electrode. lated by CDTA. At concentrations of free Ca2+ higher than 2 p~, the enzyme activity approached similar values, whether the free Mg2+ was available or complexed. This was because at such calcium concentrations, the low affinity Ca" or Mg" ATPase was also expressed.
Another proof that Mg'+ is required for the activity of high affinity calcium-stimulated ATPase came from the experiment in which the amount of free Mg2+ was increased under the conditions when either the amount of free Ca2+ present in the system was zero or was fixed to 0.56 p~. T o achieve this, two systems were used, Ca-Mg. CDTA where free Ca'" was kept constant at 0.56 p~ and Mg.CDTA, where all the Ca2+ was complexed. In both of the above systems, free Mg" was increased and the specific activity of the enzyme measured.
When the free Mg'+ was increased from zero to 200 PM and the free Ca" was kept at zero, light membranes did not show any ATPase activity and the heavy membranes showed minimal activity (Fig. 4). Even when free Ca" was 0.56 PM and free Mg'+ was less than 100 p~, no significant activity was detected in either fraction. Only at Mg2+ concentrations above 100 p~ was a significant activity evident. No attempt was made to find the amount of Mg2+ required to saturate the enzyme. The experiment proves clearly that Mg'+ is required for the enzyme activity, but Mg" alone is not enough. In the  Ca . Mg . CDTA system, free Ca2+ concentration was kept constant using a Ca'+-sensitive electrode.

Effect of Various Agents on Ca2+-Mg2+ ATPase Activity
The effect of various agents on the Ca2+-Mg'+ ATPase was tested in order to compare its properties to those of the well studied and characterized Ca2+-Mg2+ ATPase of red blood cell ghosts. Including 20 mM KCl, 20 mM NaCl and excluding 0.1 mM ouabain had no significant effect on the enzyme activity observed (Table 111). There was no significant inhibition of Ca2+-Mg2+ ATPase when sodium azide (20 mM), oligomycin (0.2 mM), and DCCD (0.1 mM) were included in the assay mixture individually (Table 111). These agents are specific inhibitors of the mitochondrial Ca" ATPase. Table IV compares the calmodulin contents of two membrane fractions from the rat corpus luteum. Both fractions contain approximately equal amounts of calmodulin per mg of protein, values which are not too different from the value observed for red blood cell ghosts from our l a b~r a t o r y .~ T h e synaptic plasma membranes and lubrol-extracted microsomes from brain contain several times more calmodulin than the corpus luteum rat membrane fraction^.^ However, treatment of these membranes with a low concentration of Triton X-100 and/or EGTA did not result in the drastic decrease in the calmodulin contents (Table V) which has been observed with red blood cell ghosts and brain rnicrosome~.~ Although treatment of heavy membranes with 0.02% Triton X-100 followed by 1 mM EGTA and washing reduced the calmodulin contents ' Unpublished observations from this laboratory. by 50% we failed to see a stimulation of Ca" ATPase activity by externally added calmodulin. Table V shows the results of an experiment in which calmodulin was added in the Ca" ATPase assay reaction mixture and enzyme activity was measured. Statistical analysis of the results shows that calmodulin stimulation of Ca2+ ATPase is not significant ( p > 0.05). The experiment shown is typical and in the majority of experiments, where increasing amounts of calmodulin were added (up to 11 pg) at the various fixed concentrations of free Ca2+ (up to 1.73 p~) , no significant stimulation of the Ca'+-Mg2+ ATPase was observed. This was true of plasma membranes isolated in the presence of 1 mM EGTA and those preparations where membrane fractions are washed with Triton X-100 and/or EGTA after preparation as in Table IV.

Effect of Phenothiazine Drugs on CaZ+ ATPase
Phenothiazine drugs have been shown to inhibit the calmodulin stimulation of phosphodiesterase (22) and high affinity Ca" ATPase (23). This effect is thought to be because of the binding of these drugs to calmodulin. Since calmodulin in the rat luteal plasma membranes is not easily dissociated, we studied the effect of phenothiazine drugs (trifluoperazine and chlorpromazine) on the activity of Ca" ATPase in a further attempt to demonstrate its calmodulin sensitivity. Trifluoperazine (see Table VI) up to a concentration of 50 PM caused no

Calmodulin contents of rat luteal plasma membranes
The data shown are mean f % range of two membrane preparations. Calmodulin contents were quantitated by measuring the activation of calmodulin-free phosphodiesterase, as described under "Materials and Methods." Each assay was done in triplicate.

Effect of calmodulin on Ca2+-Mg2+ ATPase ofplasma membranes fractionated without and with I mM EGTA
Data are mean f S.D. ( N = 6). Plasma membranes were isolated with and without EGTA present in homogenization buffer and sucrose solutions and checked for CaY+-Mg2+ ATPase activity the same day. EGTA in parenthesis indicate membranes were isolated in the presence of 1 mM EGTA. Assay conditions were the same as described under "Materials and Methods" except 20 p~ MgCl, (total), 20 mM NaC1, 50 mM KC1 were also present.  Table VI1 shows the results when 500 ~L M chlorpromazine was used in the assay mixture. Inhibition of Ca2+ ATPase of light plasma membrane even with this high concentration of chlorpromazine is only in the vicinity of 24%. Addition of 50 pg of calmodulin could not restore the activity of the chlorpromazine-inhibited enzyme to the original levels as shown in the data in Table VII.

DISCUSSION
The data indicate the presence of the Ca"+"g'+ ATPase with a high affinity for calcium in the two plasma membrane fractions isolated from the rat corpus luteum. In the light membrane, 5'-nucleotidase and Ca"-Mg" ATPase copurify to the extent of 4.5-and 4.3-fold, respectively, over the homogenate, suggesting that the Ca2+-Mg2+ ATPase studied in this fraction is of plasma membrane origin. In the heavy membrane fractions, neither 5"nucleotidase nor Ca"-Mg' ATPase were enriched significantly when compared to the homogenate (Table I).
When granulosa cells (the precursors of luteal cells) are fractionated, a surface membrane fraction accumulates in a density gradient region of 1.16 to 1.18 g/cm:' (15). This fraction is enriched in Mg2+-dependent ATPase, 5'-nucleotidase, and much of the adenylate cyclase of the ovary (15). Luteinization of the follicle following a hCG injection causes differentiation, with the appearance of luteal cells, containing a large microvillous region. When luteinized ovaries are subjected to fractionation, an additional surface membrane can be isolated at a density region of 1.12 to 1.13 g/cm:' (called light membrane fraction because of its lower buoyant density). Since the light membrane fraction is obtained as a result of extensive luteinization of the ovaries (appearance of the microvillous region) this is presumed to derive from the microvillous region of the luteal cells. The luteinized ovaries also yield a surface membrane fraction at a buoyant density of 1.16 to 1.18 g/cm" (heavy membrane) similar to that of the granulosa cells.
Light membranes from the luteinized ovaries are enriched in several plasma membrane markers, but not in adenylate cyclase. Heavy membranes (d. 1.16 to 1.18) from luteinized ovaries are not enriched in most plasma membrane markers but are enriched in adenylate cyclase. Since luteal cells are transformed granulosa cells it is unlikely that the granulosa cell plasma membranes (d. 1.16 to 1.18) will completely disappear as a result of luteinization of the follicles. Therefore, though the heavy membrane fraction from the luteal cells lacks enrichment of the surface membrane markers, it has been suggested to be a plasma membrane fraction from the basolateral surface of the luteal cells.
This inference is further supported by the fact that digitonin had differential effects on the light and heavy membranes (24). In the above study (24) it was found that the buoyant density of luteal cell light membrane as marked by [ '"I]iodo-hCG binding, Mg2+ ATPase, and 5"nucleotidase was highly perturbable by digitonin (A density > 0.05). Therefore, if the low level of 5'-nucleotidase activity found in the heavy membranes is because of the contamination of this fraction by light membranes, one would expect it to be as perturbable as the light membrane 5'-nucleotidase. However, it was found that the buoyant density of luteal cell heavy membrane fraction, as marked by adenylate cyclase, Mg" ATPase, and 5"nucleotidase was not significantly perturbed by digitonin (24). Thus, it is reasonable to infer that the low level of 5'-nucleotidase detected in the heavy membrane truly belongs to this subcellular fraction. Similarly, it may also be possible that the low level of Ca2+ ATPase in heavy membranes (although not enriched over homogenate) is truly a part of the basolateral surface of rat luteal cells.
The calcium concentration for half-maximal activity of the high affinity ATPase is in the range of 0.25 p~, making it possible for it to be of physiological significance in removing intracellular Ca".
A low affinity Ca2+ or Mg2+ ATPase is also present in these membrane fractions (Figs. 1 and 2). Such a low affinity divalent metal ion (Ca" or Mg") ATPase has been shown to be present in various tissues including placenta (18), kidney (25), liver (26), and intestine (27). The presence of such a low affinity enzyme made the quantitation of the high affinity Ca'+-Mg'+ ATPase difficult. It was possible to measure the Ca2+-Mg2+ ATPase with a high affinity for calcium only when a zero or low concentration of magnesium was employed in the reaction mixture.
The low contamination of smooth endoplasmic reticulum and mitochondrial enzymes suggest,s that the Ca2+-Mg2+ ATPase we are studying comes from plasma membrane. The effect of inhibitors of the mitochondrial ATPase, sodium azide, oligomycin, and DCCD were studied to further check mitochondrial contamination of the plasma membranes obtained. In fact, no significant effect of these compounds was found, and this indicated that the CaZ+-Mg'+ ATPase studied here was not because of mitochondrial contamination in the plasma membranes. The high affinity Ca"-Mg" ATPase from the endoplasmic reticulum of adipocytes has a requirement for Mg2+ and K' (28). KC1 (20 mM) has no effect on the rat corpus luteum Ca2+-Mg2+ ATPase suggesting that it is different from the adipocyte endoplasmic reticulum enzyme. Although from the data on the marker enzyme of SER it is not possible to absolutely rule out that the Ca"-Mg' ATPase studied here could not be of the SER origin, yet copurification of Ca"-Mi2+ ATPase and 5'-nucleotidase suggests that enzyme is present in the plasma membranes.
The enzyme in two rat corpus luteum membranes is present at a considerably higher specific activity than the red cell ghost and has a V,,,, of 110 and 55 nmol min" mg", respectively, for the light and the heavy membranes, as compared to 20 nmol min" mg" for the red blood cell ghost." All three membrane preparations show similar affinity for calcium (Ca,,? in the range of 0.2 to 0.3 pM).
The rat corpus luteum enzyme is different from the red blood cell ghosts enzyme in the respect that enough magnesium is present in the corpus luteum plasma membranes to satisfy the magnesium requirement of the enzyme. Special conditions were needed, like chelation of magnesium with CDTA, to show the requirement for magnesium. CDTA has an equal affinity for calcium and the magnesium while EGTA has a greater affinity for calcium than magnesium. Therefore, it was possible to use a low concentration (2 mM) of CDTA to control the free Ca2+ while keeping free magnesium very low. From the experiments in Figs. 2 and 3, we confirmed that Ca"-Mg"+ ATPase from the rat corpus luteum requires Mg'+ for activity and that Mi'+ alone is not enough for the expression of the activity. By analogy to the red blood cell enzyme, Ca'+-Mg'+ ATPase of the rat corpus luteum may be responsible for a calcium-pumping mechanism in these cells. The corpus luteum enzyme was found to be very similar to the properties of the adipocyte plasma membrane Caz+-Mg2+ ATPase ( 7 ) which shows a similar type of magnesium requirement.
Light and heavy plasma membranes from the corpus luteum were found to contain an appreciable amount of calmodulin (nanograms per mg of protein, light 221 f 13, heavy 216 f 0, mean f l / 2 range of two membrane preparations) ( Table V). These contents could not be reduced drastically by washing with the 1 mM EGTA and/or 0.01% Triton X-100. Such washed or unwashed membranes did not show calmodulin stimulation in the Ca"-Mg" ATPase assay (Table VI). Thus, it appears that calmodulin is more tightly bound to rat luteal cell plasma membranes than is the case with red blood cells or brain synaptic plasma membranes ( 5 ) . From the experiments where calmodulin and/or phenothiazine drugs were included in the Ca"-Mg" ATPase assay, we cannot draw any conclusions about the calmodulin sensitivity of the rat corpus luteum Ca"-Mg" ATPase, under the conditions of assay and isolation of membrane we employed.
No concrete information is available about the role of calcium in the corpus luteum function and/or reproductive cycle of the female. The calcium content of cumulus-enclosed oocytes of the rat was found to increase after injection of PMSG reaching a maximum 55 h later, when ovulation occurs (29). It is known that corpus luteum secretes progesterone under the influence of luteinizing hormones by activating the membrane-bound adenylate cyclase. The resulting cyclic AMP is known to stimulate the protein kinase activity (30). No definite role is known for Ca'+ in regulation of corpus luteum but in view of the large changes in intracellular Ca" it is probable that Ca2+ is an important intracellular messenger in this system.
We found strong similarities in the properties of Ca2+-Mg2+ ATPases of plasma membranes from the rat corpus luteum (as reported in this paper) and adipocytes ( 7 ) . In both cases, the same type of special effort was needed to show the magnesium requirement. In adipocyte plasma membranes, any response of the Ca'+-stimulated ATPase to calmodulin has been difficult to demonstrate,5 but calmodulin has been shown to stimulate the transport of Ca4' in the plasma membrane vesicles of adipocytes (31). In our system, no stimulatory effect of calmodulin on the Ca"-Mg" ATPase was found, and the effect of calmodulin on Ca"' transport was not investigated.
Like the Ca2+ ATPases from erythrocytes and brain, the corpus luteum and adipocyte enzymes are plasma membrane enzymes with a high affinity for Ca'+, suggesting that all of these enzymes are Ca2+ extrusion pumps. However, the differences mentioned above indicate that tissue-specific differences exist which distinguish the adipocyte and luteal enzymes from those of the erythrocyte and brain.