Anion dependence of Ca2+ transport and (Ca2+ + K+)-stimulated Mg2+-dependent transport ATPase in rat pancreatic endoplasmic reticulum.

Anion dependence of (Ca2+ + K+)-stimulated Mg2+-dependent transport ATPase and its phosphorylated intermediate have been characterized in both "intact" and "broken" vesicles from endoplasmic reticulum of rat pancreatic acinar cells using adenosine 5'-[gamma-32P] triphosphate ([gamma-32P]ATP). In intact vesicles (Ca2+ + K+)-Mg2+-ATPase activity was higher in the presence of Cl- or Br- as compared to NO3-, SCN-, cyclamate-, SO4(2-) or SO3(2-). Incorporation of 32P from [gamma-32P]ATP into the 100-kDa intermediate of this Ca2+ATPase was also higher in the presence of Cl-, Br-, NO3- or SCN- as compared to cyclamate-, SO4(2-) or SO3(2-). When the membrane permeability barrier to anions was abolished by breaking vesicle membrane with the detergent Triton X-100 (0.015%) (Ca2+ + K+)-Mg2+ATPase activity in the presence of weakly permeant anions, such as SO4(2-) and cyclamate-, increased to the level obtained with Cl-. However, 32P incorporation into 100-kDa protein was still higher in the presence of Cl- as compared to cyclamate-, indicating a direct effect of Cl- on the Ca2+ATPase molecule. The anion transport blocker 4,4-diisothiocyanostilbene-2,2-disulfonate (DIDS) inhibited (Ca2+ + K+)-Mg2+ATPase activity to about 10% of the Cl- stimulation level, irrespective of the sort of anions present in both intact and broken vesicles. This indicates a direct effect of DIDS on (Ca2+ + K+)-Mg2+ATPase. K+ ionophore valinomycin influenced (Ca2+ + K+)-Mg2+ATPase activity according to the actual K+ gradient: Ko+ greater than Ki+ caused inhibition, Ko+ less than Ki+ caused stimulation. From these results we conclude that Ca2+ transport into endoplasmic reticulum is coupled to ion movements which must occur to maintain electroneutrality.

port ATPase (7,8). Further characterization of this enzyme in different steps of its turnover cycle showed formation of an acid-stable 100-kDa phosphoprotein in the presence of Ca2+ in the micromolar concentration range. For dephosphorylation Mg2+ and monovalent cations such as K+ or Na+ were necessary, the Na+ being less effective than K+ (8). We have therefore termed this enzyme (Ca2+ + K+)-stimulated M$+-dependent ATPase (8).
Since Ca2+ uptake into vesicles from rough endoplasmic reticulum was also anion-dependent (7), it was the aim of the present study to further characterize the properties of Ca2+ transport by differentiating between the effect of anions on both Ca2+ATPase and Ca2+-dependent phosphorylation of its intermediate in "intact" and "broken" vesicles. The present study shows that (Ca2+ + K+)-MgZ+ATPase-promoted Ca2+ transport into intact ER vesicles was stimulated in the presence of permeant anions, whereas in broken vesicles anion stimulation sequence was abolished. An electrical membrane potential (vesicle inside-negative) created by a K+ gradient over the membrane in the presence of valinomycin stimulated Ca2+ATPase activity, whereas a vesicle inside-positive membrane potential inhibited it. We therefore conclude that Ca2+ transport into pancreatic endoplasmic reticulum is coupled to ion movements which must occur to maintain electroneutrality. Ca2+ transport is either electrogenic and permeant anions are necessary for charge compensation of (Ca2+ + K+)-Mg2'ATPase-promoted Ca2+ transport or charge compensation is necessary for electrogenic transport of another cation into ER that then serves as counterion for a neutral Ca2+ cation countertransport.

EXPERIMENTAL PROCEDURES
Materials-All reagents were of analytical grade. EGTA, EDTA, ATP (as Tris, Mg, or K, salt), the protease inhibitor benzamidine, pyronin Y, carbonic anhydrase, egg albumin, bovine albumin, phosphorylase b, P-galactosidase, myosin, FeS04, citric acid, and lithium dodecyl sulfate were bought from Sigma. Collagenase Worthington previously (9). Briefly, pancreatic tissue from six male Wistar rats (200-250 g), which had fasted overnight was digested for 15 min at 37 "C in a collagenase (150 units/ml) containing Krebs-Ringer-Hepes solution, followed by a washing step with a 2 mmol/l EDTA containing Krebs-Ringer-Hepes solution for 10 min. Single cells were then obtained by a second collagenase digestion (225 units/ml) for 60 min.
Rough endoplasmic reticulum vesicles were prepared as described recently (8). Briefly, cells were washed twice after isolation in a mannitol buffer containing (in millimoles/liter) 280 mannitol, 10 Hepes, 1 benzamidine, pH 7.0, adjusted with Tris. Cells were then homogenized in 18 ml of mannitol buffer using a tight-fitting Teflonglass Potter-Elvehjem homogenizer by 50 strokes at 900 rpm. The homogenate was centrifuged for 15 min at 11,000 X g and the resulting supernatant for 15 min at 27,000 X g. The resulting pellet was resuspended in sucrose buffer containing (in millimoles/liter) 280 sucrose, 18 Hepes, 1 benzamidine, pH 7.0, adjusted with Tris. The protein concentration of 5 mg/ml was adjusted using the protein measurement of Bradford (10). Vesicles were stored in liquid nitrogen for a maximum of 14 days.
Measurement of 45Ca2+ Uptake-Calcium uptake was measured using "Ca2+. Fifty to 100 pg of rough endoplasmic reticulum membrane protein were preincubated for 20 min at 25 "C in 500 pl of an incubation medium containing basically (in millimoles/liter) 130 KCl, 30 Hepes, 0.01 antimycin A, 0.05 oligomycin, pH 7.0, adjusted with Tris. The amount of radioactivity varied from 4 to 12 pCi/ml according to the desired total Ca2+ concentration. Uptake was initiated by adding Tris-ATP to a final concentration of 1 mmol/l. At given time points triplicate samples were filtered rapidly through cellulose nitrate filters with a pore size of 0.65 p m (Satorius, Gottingen, F. R. G.), which had been presoaked in isotonic KC1 solution. Filters were washed with 4 ml of an ice-cold solution containing (in millimoles/ liter) 140 KCI, 10 Hepes, 1 MgC12, pH 7.0, adjusted with KOH. The radioactivity was quantitated using Rotiszint 22X scintillator (Roth, Karlsruhe, F. R. G.) in a Mark I11 Liquid Scintillation System, Model 6880 Searle Analytic Inc., Des Plaines, IL. The values for ATP-driven Ca2+ transport into vesicles were calculated as the difference between Ca2+ content in the presence and absence of ATP in all experiments. Free Ca2+ and M e concentrations were calculated with a computer program using the true proton, Ca2+, and M e dissociation constants for ATP, EDTA, and EGTA as described previously (8, 11).
Phosphorylation Procedure-The phosphorylation reaction was carried out as described previously (8) for 20 s at 4 "C or at room temperature (22-25 "C) and was started by addition of ATP solution containing 0.1 mmol/l Tris-ATP and [y-32P]ATP (10 pCi, 5 pmol/l final concentration) to the incubation medium. The reaction mixture contained 100 pg of 27,000 X g pellet protein in 200 pl of incubation medium with 130 mmol/l potassium salt of Br-, C1-, NO;, SCN-, or cyclamate-. When divalent anions were tested, incubation medium contained 65 mmol/l potassium2 salt of SO:or SO:and 65 mmol/l mannitol to adjust osmolarity. Further additions were 0.01 mmol/l antimycin A, 0.01 mmol/l oligomycin, 5 mmol/l sodium azide, 1 mmol/l benzamidine, 28 mmol/l sucrose, 18 mmol/l Hepes/Tris (pH 7.0), 3 mmol/l EDTA, 3.7 mmol/l total magnesium, and 0.3 mmol/l total calcium corresponding to 1 mmol/l free Mg2+ and 1 pmol/l free Ca", respectively. In "Ca2+-free" media, 3 rnmol/l EGTA without added calcium was used. Total magnesium concentration was 1.1 mmol/l (corresponding to 1 mmol/l free M$+) under these conditions. The reaction was terminated by addition of ice-cold stop solution containing 10% trichloroacetic acid, 10 mmol/l KH2P04, and 1 mmol/ 1 ATP. The samples were kept on ice for 10 min and were then centrifuged for 5 min at 2,250 X g at 4 "C. After aspiration of the supernatant, the pellets were washed once with 1 ml of 50 mmol/l KH2POl/H3PO, (pH 2.0) solution and centrifuged again. The final pellets were dissolved in 50 pl of a solution containing (in millimoles/ liter) I sucrose, 9.2 citric acid, 1.2 phosphoric acid, 1.2 Tris, 18 lithium dodecyl sulfate (LDS), 2.4% (v/v) mercaptoethanol, and 40 pg/ml pyronin Y (pH 4.0) for LDS-polyacrylamide gel electrophoresis (12).
Electrophoresis was performed at 4 "C at 40 mA/gel for 3-4 h. When electrophoresis was finished, the gel was soaked in 1% glycerol for 5 min and dried on filter paper (LKB-Producter AB, Bromma, Sweden) using a slab gel dryer unit (LKB).
Autoradiography and Determinution of 32P Incorporation into 100-kDa Phosphoprotein-For autoradiography of 32P-labeled proteins, Kodak X-Omat AR-5 films were exposed to dried gels at -20 "C for 3 h. Quantitative measurement of 32P incorporation was carried out with an Automatic TLC Linear Analyzer System (Berthold, Wildbad, F. R. G.). Additionally, phosphorylated protein bands on the dried gel were excised according to the superimposed autoradiogram, and radioactivity of 32P incorporated into protein was counted with 4 ml of scintillator in a liquid scintillation counter (Mark 111).
Assay for Determination of ATPase Activity-ATPase activity was determined in parallel with the assay for phosphorylation by measuring 32Pi liberated from [T-~'P]ATP during the reaction according to the method of Bais (13). After termination of the phosphorylation reaction and subsequent centrifugation, a 20-4 aliquot of the supernatant was mixed with 500 p1 of active charcoal solution (125 mg/ml 1 N HCI). The sample was centrifuged at 2500 X g for 10 min at 4 "C. 100-pl aliquots of the resulting supernatant were mixed with 4 ml of scintillator Rotiszint 22X, and 32Pi liberated was counted in a liquid scintillation counter (Mark 111). The radioactivity of a control sample obtained in the absence of membrane protein was subtracted from each sample. The Ca2+ATPase activity was estimated by subtracting the ATPase activity in the absence of Ca2+ from that in the presence of Ca2+. Free Ca2+ and M2+ concentrations were adjusted with EDTA or EGTA and calculated as indicated above under "Measurement of "Ca2+ Uptake." When oxalate was used, its true proton, Ca2+, and M$+ dissociation constants were included for calculation of free Ca2+ and M$+ concentrations.

Effect of Anions on (Ca2+ + K+)-stimulated M$+-dependent
ATPase Activity and 32P Incorporation into 100-kDa Protein in Intact Vesicles-When t h e 27,000 X g pellet protein with purified endoplasmic reticulum from rat pancreatic acini was incubated in the presence of different anions, an aniondependent 32P incorporation into a 100-kDa phosphoprotein could be demonstrated using LDS-polyacrylamide gel electrophoresis at acidic pH and autoradiography. The autoradiogram in Fig. 1 illustrates the effect of anions on 32P incorporation into a 100-kDa protein in the presence and absence of Ca2+ in the incubation medium. It can be seen that Ca2+dependent 32P incorporation was increased in the presence of C1-, NO,, and SCNas compared to cyclamate-and SO:-. A second band of about 115 kDA showing Ca2+-independent phosphoprotein formation could be a Mg2+ATPase, whose origin, however, is not known. As shown in Fig. 2b and Table I (p < 0.01), whereas no difference was found for C1-as compared to Br-. Similarly, as shown in Fig. 2c and Table 11, Caz+-dependent R2P incorporation into 100-kDa protein was anion-dependent.
The differences between phosphoprotein formation in the presence of C1-and cyclamate-, and between C1-and SO:-, were significant ( p < 0.001, p < 0.001).
Inspection of the data for ATPase activity in the absence of Ca2+ shows anion-dependent MeATPase activity with a significant increase in the presence of Bror C1-as compared to cyclamate-( p < 0.05) and of SO:as compared to cyclamate-( p < 0.05) ( Table I). "P incorporation into the 100-kDa protein in the absence of Ca2+ was only significantly different for SO:as compared to cyclamate-( p < 0.05) but not between Bror C1-as compared to cyclamate-or SO:- (Table 11). It therefore appears that the 100-kDa phosphoprotein as described here is the intermediate of the Ca2+ATPase and not of an "anion-stimulated" Ca2+-independent ATPase present in pancreatic ER (Table I) whose nature is not yet known.
Since the Ca2+ATPase test was terminated after 20 s, it had to be verified that the reaction kinetics in the presence of different anions was within a linear range. Fig. 3 shows that the time course of Ca2+ATPase activity is linear for both C1-and SO:--dependent ATPase activity up to 40 s at least. It therefore appears that the differences in Ca2+ ATPase activities in the presence of C1-or SO:are not due to changes in rates of ATPase activities within 20 s, during which period the measurement had been performed.
Anion Dependence of (Ca2+ + K+)-MpATPaseActivity and "P Incorporation into 100-kDa Protein in Broken as Compared to Intact Membrane Vesicles-If Ca2+ transport into membrane vesicles is electrogenic and the vesicle membrane has a relatively low conductance for other ions, the rate of Ca2+ uptake as well as of Ca2+ATPase activity in intact vesicles should depend on accompanying anions which easily penetrate this membrane. If the membrane barrier is abolished, weakly permeant anions should no longer be rate-limiting. We have therefore compared the effect of some permeant and weakly permeant anions on Ca2+ATPase activity in the absence and presence of the detergent Triton X-100 at a concentration (0.015%) which breaks the membrane and completely inhibits 4sCa2+ uptake but not Ca2+ATPase activity as measured in the presence of C1- (7).
As shown in Table I, at 22 "C Ca2+ATPase activity in the presence of C1-or Brwas similar with or without Triton X-100. However, Ca2+-dependent ATPase activity in the presence of weakly permeant anions, such as cyclamate-and SO:-, increased with Triton X-100 to the level of C1-or Br-stimulated Ca2'ATPase activity. This effect was significant for cyclamate-( p < 0.01) and SO:-( p < 0.01). In Triton X-100-treated vesicles, Ca2+-dependent ATPase activity was not significantly different in the presence of cyclamate-and SO:as compared to C1-, and the sequence of anions obtained in intact vesicles was abolished. In broken vesicles, Ca2+independent ATPase activity (ATPase in the presence of EGTA) was significantly different only between Brand cyclamate-.
In the presence of Triton X-100 Ca2+-dependent "P incorporation into 100-kDa protein seemed to be higher as compared to untreated vesicles for all anions tested. A significant difference between intact and broken vesicles was only obtained however for cyclamate-(1.31 & 0.13 (n = 12) pmol of "P incorporated/mg of protein and 2.18 +. 0.43 (n = 4), p < Effect of Anion Transport Blocker DIDS-Previous experiments have shown that 4sCa2+ uptake into permeabilized TABLE I Effect of different anions on ATPase activity in intact and broken membrane vesicles from pancreatic endoplasmic reticulum Vesicles were preincubated for 1 min at room temperature (22 "C) or at 4 "C in the absence or presence of Triton X-100 (0.015%), without or with DIDS (5 X mol/l), and with indicated potassium salts (130 mmol/l) or potassiumz salts (65 mmol/l) in the incubation medium. Caz+-dependent ATPase activity was calculated by (2) 0.10 (2) 0.14 f 0. DIDS inhibited Ca2+ATPase activity in both intact and broken vesicles nearly completely, indicating that DIDS had a direct effect on Ca2+ATPase (Table I). The same conclusion could be drawn from data on 32P incorporation into 100-kDa intermediate (Table 11). Similarly to intact vesicles (Table 11) in broken vesicles, phosphorylation was less with DIDS as compared to controls without DIDS (data not shown). This indicates direct inhibition of the phosphorylation step in a CaZ+ATPase turnover cycle.
The inhibitory effect of DIDS on Ca2+ATPase activity was also reflected in 45Ca2+ accumulation. With increasing DIDS concentration, 45Ca2+ accumulation decreased to -50% of the control (Table 111).
(ea2+ + K+)-M$+-ATPase Activity and 32P Incorporation into 100-kDa Protein at 4 "C-Since dephosphorylation, but not phosphorylation, is inhibited a t 4 "C (8) the effect of anions on either step could be differentiated. Ca2+ATPase activity was nearly abolished a t 4 "C (Table I). However, phosphorylation of Ca2+ATPase intermediate was still pres-ent. At 4 "C, Ca'+-dependent 32P incorporation was higher in the presence of C1-as compared to cyclamate-and SO?- (Table 11). In one experiment the rate of 32P incorporation was the same for C1as compared to SO:-(0.1 pmol/mg of protein X s ) for the initial 5 s. The level of formed phosphoprotein then remained constant (0.5 pmoljmg of protein) in the presence of SO:for the subsequent 15 s of observation, whereas it increased during this time to 1.2 pmol/mg of protein in the presence of C1-. Ca2+ATPase activity was abolished during these 20 s of observation at 4 "C (data not shown). This icdicates that anions affect the phosphorylation but not the dephosphorylation step of the Ca2+ATPase intermediate.
Effect of K+ Ionophore Valinomycin on (Ca2+ + K+)-Mg2+-ATPase Activity-Valinomycin increases K+ conductance in artificial lipid bilayers and biological membranes (14-16). This property has been used to establish electrical membrane potentials in the presence of preformed K+ gradients (16). At higher K+ concentrations outside as compared to the inside of ER vesicles, an electrical potential difference, inside positive, is generated in the presence of valinomycin. This potential difference should be higher in the presence of weakly permeant anions such as cyclamate-or SO:-. If Ca2+ transport into intact ER vesicles was electrogenic, Ca'+ATPase activity should be inhibited under this condition. Fig. 4 (upper panel) shows anion-dependent decrease in Ca2+ATPase activity.

Effect of different anions on 32P incorporation into 100-kDaprotein in intact membrane vesicles from pancreatic endoplasmic reticulum
Vesicles were preincubated for 1 min at room temperature (22 "C) or at 4 "C without or with DIDS (5 X mol/l), and with indicated potassium salts (130 mmol/l) or potassiumz salts (65 mmol/l) in the incubation medium. Phosphorylation and gel electrophoresis were performed as described under "Experimental Procedures." Caz+-dependent 32P incorporation was calculated by subtraction of 32P incorporation in the absence of Ca2+ from the values obtained in the presence of Caz+. Values are means f S.E. Numbers of experiments are given in parentheses.

Effect of DIDS on Ca2+ accumulation
The experiments were performed as described under "Experimental Procedures." The free Caz+ and M%+ concentrations were 3 pmol/l and 0.3 mmol/l, respectively. Ca2+ accumulation after 20 min in the presence of regular KC1 buffer (see "Experimental Procedures") without addition of DIDS was set as 100% and was equivalent to a specific uptake of 4.2 f 0.8 nmol/mg of protein. The values are means rf: S.E. of three preparations for Ca2+ accumulation measurements. The inhibition of Ca2+ accumulation by different DIDS concentrations was statistically significant with p < 0.05, p < 0.01, and p < 0.001 for 5 X mol/l, mol/l, and 5 X mol/] DIDS concentration, respectively, as calculated by Student's t test for unpaired samples. This effect of anions is even increased in the presence of an inside positive electrical membrane potential. When vesicles were preloaded with K' ions and transferred to a low K' medium, a transmembranous potential gradient, inside negative, is generated. As shown in Fig. 4 (lower panel) Ca*+ATPase activity was increased under these conditions. This increase was only seen when preloading K' medium also contained valinomycin and the permeant anion oxalate. It therefore can be concluded that vesicles were not loaded with K' in the absence of the K+ ionophore valinomycin. Preincubation of vesicles in a high K' medium in the presence of oxalate but without valinomycin did not show stimulation of Ca2+ATPase activity as compared to preincubation without both oxalate and valinomycin. It therefore appears that oxalate itself did not effect Ca2+ATPase activity and that the ER membrane is tight for K' ions (data on the effect of preloading without valinomycin and with or without oxalate are not shown).

DISCUSSION
In a recent study we have demonstrated the presence of a (ea2+ + K+)-stimulated M$+-dependent ATPase with a phosphorylated intermediate of 100 kDa in the rough endoplasmic reticulum of the exocrine pancreas (8). Together with our previous work (5-7) this study (8) suggested that the MFATP-dependent ea2+ uptake into this organelle is promoted by a ea2+-stimulated Me-dependent ATPase. ea2+ transport into rough endoplasmic reticulum membrane vesicles was cation-and anion-dependent (7). Whereas K' or Na' ions were necessary for ea2' transport ( 7 ) and CaZ+ATPase activity by facilitating dephosphorylation of the CaZ+ATPase intermediate (8), the role of anions on ea2' transport of pancreatic rough endoplasmic reticulum has not yet been defined. Dependence of Mg'ATP-dependent 45Ca2+ uptake on anions in the order C1-> Br-> SO:-7 NO, > I-> cyclamate-> SCN-(7) suggested electrogenicity of eaz+ uptake facilitated by permeation of accompanying anions through a hypothetical Cl-conductance pathway. On the other hand, a direct effect of these anions on the Ca*+ATPase intermediate could be possible.
In order to decide between both possibilities, we have investigated in this study the dependence of both ATPase activity and phosphoprotein formation on different anions and the effect of the anion transport blocker DIDS. Since we wanted to differentiate between direct effects of anions on Ca2'ATPase and indirect effects due to different membrane permeabilities for these anions by which electrogenic cation Kz oxalate (10 mmol/l). After the preincubation period, phosphorylation reaction was started by adding 20 pl of preincubated vesicles to 180 p1 of the incubation medium containing KZ oxalate (10 mmol/l) and valinomycin M) with KC1 (115 mmol/l) or &SO4 (57.5 mmol/l) (K: = KT) or without the K' salts ( I C < KT), respectively. CaZ+ATPase activity was measured as described under "Experimental Procedures." Columns represent Ca2+(Mg2+)-ATPase activity (nmol/mg of protein X 20 s) which was calculated by subtraction of 32P liberation in the absence of Ca2+ from the values obtained in the presence of Ca2+. Asterisks indicate significant changes against control values, as calculated by Student's t test for paired samples (*, p < 0.05; **, p < 0.01). Vertical bars represent S.E. of 4-6 experiments. transport into closed rough endoplasmic reticulum membrane vesicles could be influenced, we compared effects of different anions for Ca2'ATPase activity and 32P incorporation into 100-kDa protein in both intact and broken vesicles.
Evidence for Anion Movements for Maintenance of Electroneutrality in Ca2+ Transport-Halogenes, such as C1-and Br-, facilitate 45Ca2' uptake into membrane vesicles from rough endoplasmic reticulum (7) and, as shown in Figs. 1 and 2, enhance Ca2+ATPase activity as well as 32P incorporation into the 100-kDa intermediate of intact membrane vesicles. Furthermore, higher Ca2+ATPase activity in the presence of C1-or Br-as compared to cyclamate-and SO:ions which do not easily move through membranes could be observed. In order to examine whether this preference of C1-and Br-for Ca2+ATPase stimulation might be due to anion conductance pathways in the membrane, which can be easily passed by C1and Br-but less by cyclamate-and SO:-, we have opened vesicles by treatment with the detergent Triton X-100. When Ca2'ATPase activity was compared with and without Triton X-100 treatment, preference of C1-and Br-for stimulation of Ca2'ATPase disappeared, and weakly permeant anions were no longer rate-limiting. (Table I). These results are in agreement with the hypothesis that CaZ+ATPase activity is tightly coupled with passive ion movements which must occur to maintain electroneutrality.
Valinomycin increases K' conductance in artificial lipid bilayers and naturally occurring biological membranes (14-16) and has been shown to act as a mobile carrier (14,15). In biological membranes this property has been used to create membrane potentials from preformed K' gradients or to collapse membrane potentials in the presence of K+ salts (16).
Using valinomycin, Zimniak and Racker (17) gave evidence for electrogenic Ca2+ uptake in reconstituted sarcoplasmic reticulum systems. In contrast, Chiu and Haynes (18) have concluded that Ca2+ transport into sarcoplasmic vesicles is electroneutral due t o countertransport of cations, K+ being the major countertransported ion. A direct effect of valinomycin on Ca2+ATPase activity has also to be considered (16). Thus Davidson and Berman (16) described that valinomycin (200 nmol/mg), in the absence of monovalent cations, decreased sarcoplasmic ATPase activity by 30% and abolished the stimulatory effects of 150 mM KC1 or NaCl on Ca2+ATPase turnover. The authors concluded that the ionophore interacts directly with the Ca2'ATPase, independent of its K'-conductive effects on the lipid bilayer, and that this modifies the affinity and specificity of the monovalent cation site, either by direct interaction or by the formation of a valinomycin-monovalent cation-enzyme complex. In our study a direct effect of valinomycin (10"j M) on Ca2'ATPase could not be observed. This had been verified by measuring Ca2+ATPase activity in the presence and absence of valinomycin without monovalent cations, osmolarity kept a t 300 mosm/l with sucrose (data not shown). Oxalate itself also did not influence Ca2+ATPase activity. When vesicles were preincubated in a high K' medium in the presence of oxalate but without valinomycin and were then transferred to a medium without K' but with valinomycin, the stimulatory effect on Ca"ATPase activity as shown in Fig. 4, lower panel, could not be observed. We therefore conclude that pancreatic ER membrane is tight for K' and that, in the absence of valinomycin, vesicles were not preloaded with K' sufficiently as to generate a KT > K,' gradient. The reason why oxalate had to be present also in the preincubation medium to see the effect of vesicle inside to outside directed K' gradient could be due to non-ionic diffusion of the lipophilic oxalic acid into vesicles and accumulation of K' oxalate yielding higher K+ concentrations in the vesicle than in the absence of oxalate. It should be also considered that the valinomycin-K' complex can ionpair with anions and translocate ion pairs (19,20). This reaction will lead to increase in cation and anion fluxes to a greater extent than can be accounted for by the membrane conductance for these ions. These ion-pairing reactions (19,20) are quite avid of lipophilic anions, but are also expected to a smaller extent with C1-(18). Since valinomycin influences our Ca2+ATPase activity in the presence of preformed K' gradients only, showing inhibition at a vesicle inside-positive electrical membrane potential and stimulation at an insidenegative potential, we conclude that either transport of Ca2+ is electrogenic or transport of another cation is electrogenic to which Ca2+ is coupled in a countertransport. Passive movement of permeant anions serves for charge compensation of electrogenically transported cation (Fig. 4).
Effect of DIDS on Ca2+ATPase in Endoplasmic Reticulum-In sarcoplasmic reticulum, DIDS has been shown to inhibit anion pathways (21). If a C1-pathway separate from the Ca2+ATPase was present in the pancreatic ER membrane, one should expect that DIDS inhibited Ca2+ATPase activity in intact but not in broken vesicles. However, Table I shows that Ca2+ATPase activity was abolished by DIDS in the presence of either anion, irrespective of whether or not Triton X-100 was present. 32P incorporation into the 100-kDa protein was also reduced in the presence of DIDS and was even lower than that obtained with cyclamate-without DIDS treatment (Table 11). At 4 "C, at which temperature the effect on phosphorylation without dephosphorylation could be studied, DIDS also inhibited 32P incorporation (Table 11). Since DIDS reacts with NH, and SH side groups, the observation that the anion transport blocker DIDS (22, 23) inhibited Ca2+ATPase activity and 32P incorporation into 100-kDa protein in the presence of all anions tested could indicate that, in the Ca2+ATPase molecule, such side groups are necessary for the process of phosphorylation. Similarly, direct inhibition of purified and reconstituted Ca2+ATPase by DIDS was observed in erythrocytes (24). The inhibitory effect of DIDS on Ca2+ATPase activity is also reflected in 45Ca2+ accumulation. As shown in Table 111, the same concentration of DIDS that inhibited Ca2+ATPase nearly completely, reduced Mg2'ATPdependent 45Ca2' accumulation by -50% (Table 111).
Are There "Chaotropic" Effects of Anions on Endoplasmic Reticulum Calcium Pump?-Ca2+ATPase of pancreatic endoplasmic reticulum has many similarities to that of sarcoplasmic reticulum (25-35). Both Ca2+ATPases are dependent on cations and on anions in a similar way (36-43). Since the membrane of isolated sarcoplasmic reticulum vesicles is permeable to anions and small cations (44-47), the study of electrical properties of the Ca2+ATPase is difficult (17).
If C1-and Br-would stimulate electrogenic Caz+ uptake and consequently also Ca2+ATPase activity in intact ER vesicles by charge compensation, it should be expected that a lipophilic anion such as SCN-, which easily penetrates membranes, should stimulate even better than C1-. However, as shown in Table I and Fig. 2b, the effect of SCN-on Ca2+ATPase activity in intact vesicles compares with relatively impermeant anions, such as cyclamate-and SO:-, rather than with C1-or Br-. This could be explained by a direct effect of SCN-on the Ca2+ATPase molecule, due to a chaotropic action of this anion.
In sarcoplasmic reticulum Ca2+ transport is dependent on anions in a sequence (Cl-L CH3S0, L Br-> NO: > I-> SCN-(48)) which is similar to Ca2+ transport in pancreatic endoplasmic reticulum in the present study. The and Hasselbach (49) have tested large anions denoted as chaotropic on sarcoplasmic calcium pump. They found that chaotropic anions inhibited Ca2+ transport, Ca2+-dependent ATPase activity, phosphoprotein formation, and ATP binding, and that the effectiveness of this inhibiting effect increased in the order urea < C1-< NO: < SCN-< C10; < CCl3COO-. The authors came to the conclusion that the effects of anions such as SCN-, C10; and CC1,COO-could not be explained by chaotropic action as a result of interfering with hydrophobic interactions or by disruption of hydrogen bonds, and they explained the effects of anions by interference with ATPbinding sites (49). Similarly, in our studies, a direct effect on the Ca2+ATPase molecule by anions might be possible, as also indicated by the inhibitory effect of cyclamate-as compared to C1-and Bron phosphoprotein formation in both intact and broken vesicles. Mainly, however, the effects of anions are explained by different membrane permeabilities to these anions in intact vesicles. Our data, together with the effect of electrical membrane potentials on (Ca2+ + K+)-M$+ATPase activity in intact vesicles, suggest that Ca2+ transport into endoplasmic reticulum from pancreatic acinar cells is coupled to passive movements of ions which must occur to maintain electroneutrality. Whether the Ca2+ pump is electrogenic or coupled to other ions in an electroneutral way remains for further investigation.