Biochemical evidence for functional heterogeneity of cardiac sarcoplasmic reticulum vesicles.

Two subpopulations of cardiac sarcoplasmic reticulum vesicles were resolved functionally, based on their sensitivities to the drug ryanodine. These two subpopulations of sarcoplasmic reticulum vesicles, termed ryanodine-sensitive and ryanodine-insensitive, were separated by preloading crude cardiac microsomes with Ca2+ oxalate in the presence of ATP, followed by sucrose density gradient centrifugation. Ryanodine-insensitive vesicles accumulated most of the Ca2+ oxalate during the preload, and constituted the densest subfraction recovered from the sucrose gradient. These ryanodine-insensitive vesicles exhibited the highest density of Ca2+ pumps, and accounted for 10 to 15% of the total protein in crude cardiac microsomes. Ryanodine-insensitive vesicles continued to transport substantial amounts of Ca2+ after isolation. Ryanodine-sensitive vesicles accumulated negligible Ca2+ during the preload, and were recovered from the lower density regions of the sucrose gradient. On a milligrams of protein basis, these vesicles were present in 7-fold excess over ryanodine-insensitive vesicles. Ryanodine-sensitive vesicles transported low amounts of Ca2+ under normal incubation conditions, but 3 X 10(-4) M ryanodine strikingly increased their Ca2+ uptake 5- to 10-fold. Ca2+ uptake by ryanodine-sensitive vesicles was uniquely regulated by Ca2+ ion concentration. Elevation of the ionized Ca2+ concentration from 2 to 4 microM increased Ca2+ uptake by these vesicles greater than 5-fold, but had no effect on their Ca2+-dependent ATPase activity. These ryanodine- and Ca2+ concentration-dependent effects were apparent for only ryanodine-sensitive vesicles. Sodium dodecyl sulfate polyacrylamide gel electrophoresis revealed distinct differences in polypeptide staining between ryanodine-sensitive and ryanodine-insensitive vesicles, confirming by an independent method that the two populations of vesicles were different. These data provide the first biochemical evidence for functional and structural heterogeneity of cardiac sarcoplasmic reticulum vesicles.

lyzes ATP in the presence of micromolar concentrations of Ca2+ to yield ADP and inorganic phosphate (2-4). In the early studies characterizing Ca2+ uptake by these vesicles, it was generally observed that cardiac microsomes accumulated much less Ca2+ than skeletal muscle microsomes (4). Although preparative procedures for cardiac sarcoplasmic reticulum vesicles have improved with time (3,(5)(6)(7)(8), the Ca2+-dependent ATPase activity that these vesicles exhibit is still less than that of fast skeletal muscle vesicles (3,7).
The low Ca2+ transport activities originally observed for cardiac sarcoplasmic reticulum preparations were thought to be due at least in part to the relative impurity of the fractions studied (4, 9). Sarcoplasmic reticulum vesicles actively transporting Ca2+ in the presence of oxalic acid generate Ca2+ oxalate precipitates, which are readily visualized inside the vesicles with use of the electron microscope ( 4 ) . only 5 to 10% of the vesicles formed Ca" oxalate precipitates in cardiac preparations (2,9), whereas 20 to 30% of the vesicles developed precipitates in skeletal muscle Preparations (2,(9)(10)(11). To explain these observations, Baskin and Deamer postulated that contaminating sarcolemmal vesicles devoid of significant Ca2+ transport activity might contribute a substantial portion of the total membrane mass in cardiac preparations (9). This idea has been supported by recent studies demonstrating purification of sarcolemmal vesicles from such crude cardiac microsomal preparations (12,13). The sarcolemmal content of these crude cardiac preparations, however, was only 15% or less (12,13), an amount which would not seem sufficient to dilute out appreciably the activity of the sarcoplasmic reticulum vesicles (3).
Carsten and Reedy selectively increased the density of the sarcoplasmic reticulum vesicles in crude cardiac microsomes by loading them to maximum capacity with Ca2+ oxalate, thus allowing the separation of the sarcoplasmic reticulum vesicles from the other vesicles by differential centrifugation (6). Ca2+ uptake by these purified sarcoplasmic reticulum vesicles, however, could not be further characterized because the isolated vesicles did not continue to transport Ca'+ (6). Shortly thereafter, Levitsky et al. successfully used sucrose density gradient centrifugation to purify cardiac sarcoplasmic reticulum vesicles that had been preloaded with much less Ca2+ oxalate ( 7 ) . An advantage of this latter technique was that the purified vesicles were capable of accumulating high levels of additional Ca2+ after isolation ( 7 ) . This was recently confirmed by Misselwitz et al. (13).
In all of the foregoing studies which utilized the technique of Ca2+ oxalate loading to facilitate purification of sarcoplasmic reticulum vesicles from cardiac muscle, it was evident that the purified vesicles accounted for only a small percentage of the total membrane content present in the crude preparations (6, 7 , 12,13). Most of the membrane protein was recovered in other subfractions, which were also shown to be 11809 vesicular, but these subfractions were unable to transport substantial amounts of CaZ+. Paradoxically, however, these additional subfractions did exhibit appreciable specific activities of Ca2+-dependent ATPase, the enzyme responsible for ca2+ translocation (6, 12, 13). This did not appear to be a unique property of the cardiac preparations, because similar results were also reported in purification studies that used skeletal muscle as the source of the sarcoplasmic reticulum (14-16). The explanation for the inability of these additional sarcoplasmic reticulum subfractions to accumulate Ca2+ ions was unclear. It was possible that the vesicles had become leaky or were damaged during their preparation and thus were unable to retain any Ca2+ ions that they might have transported (6, 14), or that their membranes were inverted such that any active Ca2+ transport by them would have been directed outward into the incubation medium (11,13,16).
In the present study, we suggest that these cardiac sarcoplasmic reticulum vesicles thought previously to be deficient in Ca" uptake are actually capable of transporting substantial amounts of Caz+, and furthermore, that they constitute a distinct subpopulation of sarcoplasmic reticulum vesicles in cardiac microsomes. Subfractions of cardiac sarcoplasmic reticulum vesicles were first isolated by the Ca2+ oxalate-loading technique described above, and then, additional Ca2+ transport by the subfractions was studied in detail. In agreement with the results of Levitsky et al. (7) and Misselwitz et al. (13), we found that only a fraction of the total sarcoplasmic reticulum vesicles accumulated Ca2+ oxalate during the loading step, and that these vesicles continued to transport considerable Ca2+ after they were isolated. The lighter sarcoplasmic reticulum vesicles, which did not accumulate significant Ca2+ oxalate during the loading step, had only low levels of Ca2+ transport after isolation. However, Ca2+ transport by the lighter vesicles was markedly and selectively stimulated by ryanodine, a neutral alkaloid shown recently to increase Ca2+ uptake into crude cardiac sarcoplasmic reticulum preparations (8). These ryanodine-sensitive vesicles were differentiated from the remainder of the sarcoplasmic reticulum vesicles by several different criteria. The data demonstrate that isolated cardiac sarcoplasmic reticulum vesicles are functionally heterogeneous.

EXPERIMENTAL PROCEDURES
Subfractionation of Crude Cardiac Sarcoplasmic Reticulum Vesicles-Crude cardiac sarcoplasmic reticulum vesicles were isolated as described previously (12). All operations were conducted at 4 "C unless otherwise stated. Briefly, 180 to 200 g of left ventricular tissue was divided into six portions, and each portion was homogenized three times for 30 s in 100 ml of 10 m~ NaHC03 with a Polytron PT-20 (Brinkman Instruments). Two sequential low speed centrifugations at 14,000 X gmaX for 20 min were used to remove large particles. The supernatant material from the second centrifugation was next sedimented at 45,000 X g,, for 30 min. The pellets were resuspended in 0.6 M KC1 and 30 mM histidine, pH 7.0, and sedimented again to yield the crude sarcoplasmic reticulum vesicles.
Ca2+-loading of the crude sarcoplasmic reticulum vesicles was modified from our previous report (12). The crude vesicle pellets (75 to 100 mg of protein) were resuspended in 40 ml of an ice-cold medium containing 50 mM histidine, 100 m~ KCI, 65 mM MgCh, 60 mM NazATP, 25 m~ Tris/EGTA,' 20 m~ CaCh, and 5 mM Tridoxalate (pH 7.1). In experiments in which radioactive Ca2' uptake was measured, 0.3 mCi of 45CaC1z was included. The suspension was placed in a waterbath at 37 "C to initiate rapid Ca2+ uptake, and the incubation was conducted for 10 min with 5 m~ additional Tris/oxalate added after the first 5 min of incubation. The suspension was then immediately centrifuged at 4 "C for 30 min at 105,000 X gmx. Consistent with the results of Carsten and Reedy (6), we observed a tan pellet with a white central region or button. The white central region contained I The abbreviation used is: EGTA, ethylene glycol bis(P-aminoethyl ether)-N,N,N',N'-tetraacetic acid.
the Ca2+-fded vesicles. When ATP was omitted from the incubation medium, or when the Ca2+ ionophore A23187 (3 pg/ml) was included in the incubation medium, this white part of the pellet was no longer visible.
The Ca*+-loaded vesicles were resuspended in 0.25 M sucrose containing 300 mM KC1,50 mM sodium pyrophosphate, and 100 m~ Tris (pH 7.2). This material was layered over a discontinuous sucrose gradient consisting of 7 ml each of 0.6 M, 0.8 M, 1.0 M, and 1.5 M sucrose dissolved in the same buffered solution above. Routinely, a total of 15 ml of crude vesicles was layered over three sucrose gradients. Centrifugation was for 2 h at 27,000 rpm in a Beckman SW 27 rotor. Fraction A was collected at the 0.25 ~: 0 . 6 M interface, fraction B at the 0.6 ~: 0 . 8 M interface, fraction C at the 0.8 M:LO M interface, and fraction D at the 1.0 ~: 1 . 5 M interface. Fraction E was a white pellet at the bottom of the gradient. The subfractions were diluted with 4 volumes of ice-cold H20, and then sedimented at 105,000 X gmaX for 60 min. The pellets were resuspended in 0.25 M sucrose, 10 m~ histidine, and stored frozen at -20 "C. Protein was determined by the method of Lowry et al. (17).
Assay of Ca2+ Uptuke-Ca2+ uptake by the vesicles was routinely conducted at 37 ' C in a medium consisting of 50 m~ histidine (free base), 3 mM MgC12, 3 m~ Tris/oxalate, 100 m~ KCl, 50 p~ 45CaC12, and 3 m~ Tris/ATP (18). Vesicles were preincubated for 10 min at pH 7.4 prior to initiating the reactions by adding the ATP. Addition of the ATP lowered the pH to 6.8, which remained stable throughout the incubations. This slightly alkaline preincubation was required to demonstrate some of the unique properties of the ryanodine-sensitive vesicles, but it had no effect on Ca2+ uptake by the ryanodineinsensitive vesicles. Ca2+ uptake reactions were terminated by fdtration with type HA Millipore fdters (8, 18) or with Whatman GF/C filters. Similar results were obtained with either type of filter. All Ca2+ uptake reported was ATP-dependent and was completely prevented by the Ca2+-specific ionophore A23187 (3 pg/ml). Five mM sodium azide had no effect on Caz+ uptake by any of the subfractions. In all experiments on 45Ca2+ uptake reported presently employing an added Ca2+ concentration of 50 p~, the vesicles accumulated no more than 30% of the total Ca2+. In some experiments, a Ca2+:EGTA buffer was used to maintain a constant ionized Ca2+ concentration. The concentration of EGTA was fixed at 1 I", and the CaCh concentration was varied between 0.1 and 0.9 m~. Data obtained in these experiments are reported in terms of the total added Ca2+ concentration. In addition, in Table I1 and Fig. 11 ionized Ca2+ concentrations are calculated assuming an apparent association constant of Ca" for EGTA of 1 X IO6 M" at pH 6.8 (19).
Ca2+ uptake by the vesicles was also determined by a potentiometric technique, which measured the ability of the vesicles to lower the extravesicular Ca2+ concentration. This was performed as described by Madeira (20), using a Radiometer F2112 Caz+ electrode connected to a Radiometer PHM 64 pH meter to monitor the extravesicular Ca2+.
Assay of ATPase Activity-Ca2+-dependent ATPase activities were measured at 37 "C in media identical with those above. Basal ATPase activities were determined in the same media, which had no added Ca2+ and 1 m~ Tris/EGTA. These basal activities were subtracted from the total ATPase activities, to yield the Ca*+-dependent ATPase activities. In experiments in which Ca2+ uptake and Ca2+dependent ATPase activities were measured simultaneously, aliquots were taken from the same suspension of vesicles to determine the amount of Ca2+ uptake and the total ATPase activity. An additional sample was incubated with EGTA and no added Ca2+ to determine the basal ATPase activity. Production of inorganic phosphate from ATP was measured colorimetrically (21).
(Na',K+)-ATPase activity was that activity inhibited by 1 I " ouabain. This activity was determined to assess contamination of the membrane fractions with sarcolemmal vesicles (12, 22).
The contents of the tubes were applied to Whatman GF/C filters, and the tubes were rinsed two additional times with the same solution, which was also applied to the filters. The filters were then rinsed four times with 10 ml of ice-cold solution containing 5% trichloroacetic acid, 20 m~ NaH2P04, and 2 m~ NaZATP, followed by one rinse with 5 ml of ice-cold HzO. Radioactivity remaining on the filters was determined by liquid scintillation counting. The incorporation of radioactivity inhibited by EGTA was used to determine the Caz'dependent ATPase content of the subfractions (23).
Polyacrylamide Gel Electrophoresis-The subfractions were dissolved in a solution containing 2.5% sodium dodecyl sulfate as previously described (12). Twenty-five pg of protein from each subfraction were applied to a polyacrylamide slab gel, and electrophoresis was conducted according to the method of Laemmli (24). The concentrations of acrylamide in the stacking and resolving gels were 3 and 78, respectively. Proteins were visualized with the cationic carbocyanine dye Stains-All (25), which best revealed the differences in protein banding between the subfractions.
Material~-[y-~~P]ATP was obtained from ICN Pharmaceuticals. 45CaClz was obtained from New England Nuclear. Ryanodine was purchased from the S. P. Penick Co. Stains-All was purchased from the Eastman Kodak Co. A23187 was generously provided by Dr. R. L. Hamill, Eli Lilly and Co. All other reagents were purchased from the Sigma Chemical Co.

RESULTS
Isolation of Subpopulations of Cardiac Sarcoplasmic Reticulum Vesicles-To obtain good separation of different sarcoplasmic reticulum fractions, the crude vesicles were preloaded with Ca2+ plus oxalate in the presence of ATP, and then subjected to sucrose density gradient centrifugation. Virtually all of the CaZ+ accumulated during the preload was recovered in subfraction E, the densest subfraction isolated (Table I). This fraction accounted for about 12% of the total protein recovered (Table I). When the Ca2+ ionophore A23187 was included in the Ca2+-loading medium, or when ATP was omitted from the incubation mixture, the white E vesicles were no longer visible, and no radioactive Ca'+ or protein sedimented through the 1.5 M sucrose into the E region of the gradient (data not shown). This demonstrated that the radioactive calcium recovered in the E fraction was accounted for by Ca2+ oxalate which had precipitated inside the vesicles. Loss of radioactive Ca2+ from the Caz+-loaded vesicles was minimal during the centrifugation procedures, as suggested by the fact that the Ca2+ oxalate recovered in subfraction E accounted for all of the Ca2+ oxalate accumulated by the crude   membrane vesicle fraction (Table I). Ca2+ oxalate accumulated by the crude membrane vesicle fraction during the preload was determined by either immediate filtration, or by sedimentation of the membranes with the centrifuge. In both cases, similar results were obtained. Ca2+-dependent acylphosphoproteins and Ca2+-dependent ATPase activities, which measured the presence of the Ca2+ transport enzyme (21,23), were present in all the subfractions, although the E vesicles had the highest specific activities of these two parameters (Table I). Basal ATPase activity, measured in the absence of Ca", was also present, and it was consistently the lowest in subfraction E (Table I).
To investigate why subfractions A through D exhibited low apparent levels of Ca2+ uptake, but had appreciable Ca2+dependent ATPase activities, Ca" uptake was measured by the subfractions after they had been isolated. Under control incubation conditions, subfraction E continued to accumulate substantial Ca2+, while the other subfractions exhibited only Ca" uptake (except that depicted in Fig. 12), preincubation was at pH 7.4 for 10 min, after which time reactions were initiated by adding ATP, which brought the pH to 6.8. low levels of Ca2" uptake ( Fig. 1, left). All of this additional Ca2+ uptake was completely prevented by the Ca2+ ionophore A23187 (data not shown). When the drug ryanodine was included in the incubation media, Ca2+ uptake by subfractions B through D was greatly increased (Fig. 1, right). Ca" uptake by the E subfraction was unresponsive to ryanodine. Although Ca2+ uptake by the A subfraction was also increased by ryanodine, the absolute level of Ca2+ accumulated remained well below that of the other fractions (Fig. 1, right). These A vesicles were substantially contaminated with sarcolemmal membrane fragments, as evidenced by their relatively high (Na',K+)-ATPase activity (Fig. 2), and they were not studied further.
No s i g d c a n t differences have been noted between B, C, and D vesicles with regard to their Ca2+ uptake properties in all of the additional studies to be described. For convenience we w i l l refer to the vesicles in subfractions B, C, and D as the ryanodine-sensitive vesicles. The E vesicles are the ryanodineinsensitive vesicles. We should emphasize that although ryanodine-sensitive and -insensitive vesicles accumulated similar levels of Ca2+ after isolation, ryanodine-insensitive vesicles already contained approximately 8 pmol of Ca2+/mg of protein at the start of the incubations.
Ryanodine Effects on Ca2+ Uptake by Subpopulations of Cardiac Sarcoplasmic Reticulum Vesicles-Ryanodine stimulated Ca" uptake maximally into the ryanodine-sensitive vesicles at a concentration of 300 p~ (Fig. 3, right). The concentration effectiveness of ryanodine was the same, when membrane vesicle protein concentrations were varied between 10 and 100 pg/ml (data not shown). None of the ryanodine concentrations tested had any significant effect on Ca2+ uptake by the ryanodine-insensitive vesicles (Fig. 3, left). Ryanodine stimulation of Ca2+ uptake into the sensitive vesicles was associated with no change or a slight decrease in Ca2'-dependent ATPase activity (Fig. 4). Ryanodine (300 p~) also had no effect on steady state acylphosphoprotein levels of these vesicles measured as described in Table I (data not  shown). Thus, ryanodine did not appear to increase Ca2+ uptake into the sensitive vesicles by increasing the catalytic activities of their Ca2+ pumps.
Ca2+ exchange was measured in ryanodine-sensitive vesicles by adding a trace amount of 45CaClz after active calcium transport had first equilibrated in the presence of 40CaC12 (Fig. 5). This rate of Ca'+ exchange was stimulated at least 8fold by ryanodine. In other experiments, it was determined that ryanodine also stimulated Ca2+ uptake severalfold, when the drug was added 20 min after the reactions had been initiated by the addition of ATP (data not shown). Both types of experiments suggested that Ca2+ oxalate already inside cardiac sarcoplasmic reticulum vesicles did not prevent a response to ryanodine.
Ca2+ uptake into ryanodine-sensitive and -insensitive vesicles was measured with use of a ca'+-selective electrode, which monitored the extravesicular Ca2+ concentration. The ryanodine-insensitive or E vesicles effectively lowered the extravesicular Ca2+ concentration, and this process was not affected by ryanodine (Fig. 6, circles). This demonstrated that ryanodine-insensitive vesicles did in fact accumulate additional net Ca2+, in excess of that already transported during the Ca2+-loading step. With use of the Ca'+-selective electrode it was further demonstrated that the ryanodine-sensitive vesicles also lowered the extravesicular Caz+ concentration, and that this process, in contrast, was dependent upon the presence of ryanodine (Fig. 6, triangles). We have obtained similar results for both populations of vesicles with use of the dye Arsenazo 111 to monitor extravesicular Ca2+ concentration (data not shown). Comparison of the results in Table I and Figs. 1 and 6 suggests that ryanodine-insensitive vesicles were capable of accumulating approximately 3 times more total Ca2+ oxalate than ryanodine-sensitive vesicles. Ryanodineinsensitive vesicles had been preloaded with approximately 8 pmol of Ca2+/mg of protein prior to isolation, and they accumulated approximately 4 pmol of additional Ca2+/mg of protein after isolation. Ryanodine-sensitive vesicles, on the other   or without (0) ryanodine. Assays were conducted in a medium identical with that used in Fig. 1, except that a CaZ+/EGTA buffer (0.8 m~ CaC4 plus 1 mM EGTA) was included such that the vesicles accumulated only a small fraction of the total Ca2+. hand, accumulated approximately 4 pmol of total Ca*'/mg of protein under optimal conditions in the presence of ryanodine.
Ca2+ Dependence for Ca2+ Uptake by Ryanodine-sensitive and -insensitiue Vesicles-Ryanodine-sensitive and -insensitive vesicles were differentiated by their responses to Ca2+ concentration, when active Ca2+ transport was measured (Figs. 7 and 8). For both populations of vesicles, Ca2'-dependent ATPase activities were maximal at 0.6 m~ added Ca*+ in the presence of 1 m~ EGTA (or at about 1.5 X M ionized Ca2+) (Figs. 7 and 8, right panels). Ca2+ uptake for the ryanodine-sensitive vesicles, however, was very low at this added Ca" Concentration, and could barely be detected at the lower Ca2+ concentrations tested (Fig. 7, left). Ryanodineinsensitive vesicles, on the other hand, had already achieved their maximal rates of Ca2+ transport at 0.6 m~ added Ca", and exhibited considerable levels of Ca2+ transport even at the lower Ca2+ concentrations tested (Fig. 8, left). A spontaneous release of Ca2+ from ryanodine-insensitive vesicles after 10 min of incubation, which occurred at 0.6 m~ added Ca", was blocked when the Ca2+ concentration was increased to 0.8 mM (Fig. 8, left).
Most striking was the large increase in both the rate and extent of Ca2+ transport for the ryanodine-sensitive vesicles, when the Ca2+ concentration was changed from only 0.6 to 0.8 to determine basal ATPase activity. Otherwise, the incubation mixture was identical with that used in Fig. 1.
mM in the presence of 1 m~ EGTA (Fig. 7, left). This increase in Ca2+ transport was associated with no increase in Ca2+dependent ATPase activity (Fig. 7, right), Ca2+ uptake and ATPase rates were obtained at each ionized Ca"' concentration from the data depicted in Figs. 7 and 8 ( Table 11). The initial rates of Ca2+ uptake divided by the initial rates of ATP hydrolysis were used to estimate apparent coupling ratios, or the number of Ca2+ ions effectively transported per ATP molecule hydrolyzed during the early, nearly linear phases of Ca2+ uptake. Ca2+ transport was poorly coupled to ATP hydrolysis for ryanodine-sensitive vesicles at ionized Ca2+ concentrations of 1.5 PM or below (Table 11). However, between 1.5 and 4.0 PM ionized Ca2+, a 10-fold increase in apparent Ca'+ pumping efficiency occurred. This was primarily accounted for by a large increase or "jump" in the Ca2+ uptake rate. In comparison, ATP hydrolysis was fairly well coupled to Ca2+ transport at each ionized Ca2+ concentration for the ryanodine-insensitive vesicles, and no such jump in the Ca2+ uptake rate was evident. We have repeated these experiments many times with different preparations of vesicles and have always obtained similar results. The maximum coupling ratios obtained always approached a value of one for either subpopulation of vesicles, and in no cases were transport ratios greater than one obtained.  Fig. 7, except that ryanodine was not included. Both subfractions used in Figs. 7 and 8 were isolated from the same crude membrane vesicles. In control experiments, ryanodine was shown to have no effect on Ca2+ activation of uptake and ATPase activity of E vesicles.

Ca2' transport efficiencies of ryanodine-sensitive (Sens) and ryanodine-insensitive (Insens) vesicles
Results were calculated from the data depicted in Figs. 7 and 8, which are representative of a typical experiment. Ca2+ uptake and Caz+-dependent ATPase velocities were estimated from the initial, nearly linear portions of the graphs. One m~ EGTA was present in all incubation media. Car+ coupling ratios were obtained by dividing Ca2+ uptake rates by Ca"-dependent ATPase rates.

Characterization of the Jump in Ca2+ Uptake Exhibited by Ryanodine-sensitive Vesicles-
The large increase or jump in Ca2' uptake that occurred only for ryanodine-sensitive vesicles was further investigated. Vesicles were incubated under active Ca2+ transport conditions for 20 min, which allowed for equilibration of Ca2+ uptake (Fig. 7). Ca2+ transport and &'+-dependent ATP hydrolysis that had occurred during this time were measured at several added Ca2+ concentrations, in the presence of 1 m~ EGTA (Fig. 9). Between an added Ca2+ concentration of 0.7 and 0.8 m~, Ca2+ uptake increased sharply in a typical preparation of ryanodine-sensitive vesicles (Fig. 9). This increase in Ca2+ uptake was completely prevented by the Ca2+ ionophore A23187, which demonstrated that it was not due to Ca'+ oxalate precipitating outside the vesicles. Ca2+-dependent ATP hydrolysis, measured with or without A23187, was activated over a much wider range of added CaZ+ concentration (0.1 to 0.7 m~) , and Ca" activation of this process was complete before the jump in Ca2' uptake had occurred (Fig. 9). In other experiments it was demonstrated that the jump in Ca2' uptake occurred at a similar added Ca'+ concentration when either phosphate or fluoride anions were used as Ca2+ precipitants, suggesting that a purely physical interaction between Ca2+ and oxalate was subfraction was allowed to actively transport Ca2' for 20 min, then Ca2+ uptake (circles) and Caz+-dependent ATP hydrolysis (squares) that had occurred was determined. All media contained 1 m~ EGTA and 3 X IO" M ryanodine, and the concentration of 45CaC1~ was varied as indicated along the abscissa. Basal ATPase activity was measured by incubating an additional sample in the presence of only EGTA. The filled symbols indicate samples incubated in the presence of 3 p g / d of A23187 (A23). The incubation medium used was otherwise the same as that in Fig. 1.  Fig. 9, except that the 45CaClz concentration was varied over a narrower range. Open and filled circles depict Ca2+-dependent ATPase activity (pmol Pdmg protein) and Ca2+ uptake measured in the presence of 3 X M ryanodine, respectively. Open squares depict Ca" uptake measured without ryanodine. not responsible for the large increase in ea2+ uptake. Furthermore, the jump in ea2+ uptake was not detected in crude membrane vesicles incubated in the absence of ryanodine, again suggesting that this process was a property of only ryanodine-sensitive vesicles (data not shown).
The jump in ea2' uptake was measured over a narrower range of added ea2+ concentration, with ryanodine-sensitive vesicles incubated in the presence and absence of the drug (Fig. 10). The increase in Ca2+ uptake occurred with or without ryanodine, but the total ea2+ accumulated was much greater when the incubations were conducted in the presence of ryanodine. In the experiment of Fig. 10, the jump in ea2' uptake was completed between 0.70 and 0.80 m~ added ea2+, or between the calculated ionized ea2' concentrations of about 2 and 4 p~. The steepness of the response was indicated by determining a Hill coefficient of 16 for the effect in a typical preparation of ryanodine-sensitive vesicles (Fig. 11, right).
The mean kS.E. calculated from six such experiments was 17 & 1.8. In the same preparations, the Hill coefficient for ea2+activation of ATPase activity was found to be 1.9 k 0.13 (shown for a typical preparation in Fig. 11, left), which is consistent with the results of others (26).
In all experiments on ea2+ uptake so far described, the vesicles had been preincubated at pH 7.4 for 10 min, after which time ea2' transport was initiated by adding an acidic solution of ATP, which lowered the pH of the incubation medium to 6.8. When both the preincubations and incubations were conducted at pH 6.8, the jump in ea2+ uptake noted only  . 12. pH dependence for the jump in Ca2+ uptake by ryanodine-sensitive vesicles. Incubations were conducted as described in Fig. 9 for a C subfraction of vesicles. Left, Ca2+ uptake; right, corresponding Ca2+-dependent ATPase activities. Vesicles exhibiting the jump in Ca" uptake (filled symbols) were preincubated for 10 min at pH 7.4 before the reactions were initiated by adding ATP, which lowered the pH to 6.8. Vesicles not exhibiting the jump in Caz+ (open symbols) were treated identically, except that the pH was maintained at 6.8 during both the preincubation and the incubation.
for the ryanodine-sensitive vesicles disappeared (Fig. 12, left). ea2+-activation of Ca2+-ATPase activity, on the other hand, was not affected by the slightly alkaline preincubation (Fig.  12, rzght). In other experiments, it was demonstrated that eaZ+ uptake by the ryanodine-insensitive vesicles was not affected by the pH of the preincubation. Although the jump in ea2+ uptake by ryanodine-sensitive vesicles was not present when the preincubation was conducted at pH 6.8, Ca2+ uptake itself was still stimulated 5-fold or greater by ryanodine (data not shown). ( Fig. 13). Differences in protein staining between the subfractions were most clearly revealed by the protein stain, Stains-All. Fractions B, C, and D, the ryanodine-sensitive fractions, had a dark blue, intensely staining protein band of molecular weight 55,000 (double arrow), that was not apparent in the ryanodine-insensitive (E) fraction. We have consistently observed that the intensity of this blue staining protein band increases from fractions B through D. This M , = 55,000 protein was also easily visualized in the crude membrane vesicles, confirming by an independent method that most of the protein in the crude microsomes was contributed by ryanodine-sensitive vesicles. Fraction E, on the other hand, characteristically contained a very high molecular weight protein band, which was not prominent in the other subfractions or in the crude membrane vesicles (single arrow). The pink staining protein band of molecular weight lO?,000 present in all the subfractions was in large part contributed by the Ca'+-dependent ATPase (23,27). Radioactive Ca2+-dependent acylphosphoproteins (Table I) were localized to this region of the polyacrylamide gel for all the subfractions (data not shown).

DISCUSSION
Sucrose density gradient centrifugation was used in the present study to isolate two biochemically distinct subpopulatiolis of car,diac sarcoplasmic reticulum vesicles. When crude sarcoplasmic reticulum preparations were preloaded to submaximal levels with Ca2+ oxalate, only the ryanodine-insensitive vesicles became more dense, and thus they were easily separated from the ryanodine-sensitive vesicles.
Compared to ryanodine-sensitive vesicles, ryanodine-insensitive vesicles had a higher density of Ca2+ pumps, a higher Ca2+-dependent ATPase activity, and a higher initial velocity and maximal capacity for Ca2+ transport. Ryanodine-sensitive vesicles could accumulate substantial levels of Ca2+, but only in the presence of the drug, and they exhibited a unique Ca2+ concentration-dependent and pH-dependent increase in Ca2+ uptake that was not associated with any change in Ca2+dependent ATPase activity. In other experiments, we have observed that M chlorpromazine completely inhibited Ca2+ uptake by ryanodine-sensitive vesicles, but it had no effect on Ca2+ uptake by ryanodine-insensitive vesicles2 Thus, the data above provide evidence for a functional heterogeneity of cardiac sarcoplasmic reticulum vesicles.
Different subpopulations of sarcoplasmic reticulum vesicles were substantiated by examining their polypeptide compositions after sodium dodecyl sulfate polyacrylamide gel electrophoresis. Most apparent was that ryanodine-sensitive vesicles had an intense, blue staining protein band of molecular weight 55,000, which was absent from ryanodine-insensitive vesicles. Ryanodine-insensitive vesicles had an additional protein band of very high molecular weight that was not present in ryanodine-sensitive vesicles. In other recent studies it was demonstrated that, although both populations of vesicles contain a calmodulin-dependent protein kinase activity, ryanodine-sensitive vesicles have several different protein substrates of this kinase that are absent from ryanodine-insensitive vesicles (27). Moreover, although ryanodine-sensitive vesicles have a lower Ca2+-dependent ATPase content than ryanodine-insensitive vesicles, they have more of the M, = 22,000 protein substrate of CAMP-dependent protein kinase, phospholamban (27).
Some precedent, albeit from a different tissue source, exists for the results described presently. Successful separation of subpopulations of sarcoplasmic reticulum vesicles from fast skeletal muscle have been ongoing for some time. Meissner first isolated skeletal muscle sarcoplasmic reticulum subfractions enriched in either terminal cisterna or logitudinal sarcoplasmic reticulum (28), and this work has since been extended by others (29-31). In different studies, Fairhurst noted that the inhibitory effect of ryanodine on Ca2+ transport by skeletal muscle sarcoplasmic reticulum vesicles was specific for only those membrane vesicle fractions that sedimented at relatively low g forces (32). (Ryanodine apparently has opposite effects on Ca2+ uptake by skeletal muscle and cardiac microsomes, Ref. 8). Functional heterogeneity of sarcoplasmic reticulum in relatively intact skeletal muscle fibers was recently suggested by the results of Sorenson et al. (33). Using an elegant technique with skinned skeletal muscle fibers, which relied on visualization of Ca2+ oxalate crystals deposited during active CaZ+ transport, these investigators demonstrated that individual regions of the sarcoplasmic reticulum network within each sarcomere differed in their abilities to transport Ca2+ (33). Indeed, Carsten and Reedy originally postulated that functional heterogeneity of cardiac sarcoplasmic reticulum vesicles might explain why only a fraction of the vesicles accumulated Ca2+ oxalate (6).
Although our results appear to substantiate functionally different subpopulations of cardiac sarcoplasmic reticulum vesicles, some caution is warranted in interpreting experimental results obtained with ryanodine-insensitive vesicles, which were already partially filled with Ca2+ oxalate. Use of a Ca2+selective electrode, however, confirmed that the 45Ca2+ uptake exhibited by ryanodine-insensitive vesicles reflected a real, net uptake of Ca2+ and was not simply a manifestation of Ca2+ exchange. This was also supported by the high magnitude of this Ca2+ transport, which was 2 to 3 times greater than that observed by us for crude vesicles used in previous studies (8, 18). Ryanodine stimulation of net Ca2+ uptake by these vesicles should have been detectable, therefore, if the vesicles * L. R. Jones  were capable of responding to the drug. Control experiments conducted with ryanodine-sensitive vesicles also supported this conclusion. Ca2+ oxalate precipitates already inside these sensitive vesicles did not prevent a response to ryanodine. Although Carsten and Reedy (6) observed no additional Ca2+ uptake by cardiac sarcoplasmic reticulum vesicles isolated by a Ca2+-loading technique, this may have been because the vesicles had already been filled to their maximal levels (5 pmol of Ca2+/mg of protein in the crude cardiac vesicles with 26 pmol of Ca2+/mg of protein recovered in the subsequently purified vesicles). Our results c o n f i i those of Levitsky et al. (7), who observed additional Ca2+ transport (11 pmol of Ca2+/ mg of protein) by cardiac sarcoplasmic reticulum vesicles isolated by a Ca2+-loading technique similar to that used in this study.
Since the cellular localization of ryanodine-sensitive and -insensitive vesicles was not established in the present study, it is relevant to ask if in fact both populations of vesicles originated from sarcoplasmic reticulum. This can best be addressed by stating what the vesicles did not appear to be. They did not appear to be sarcolemmal vesicles because they contained very low levels of all of the different sarcolemmal markers including (Na+,K+)-ATPase, adenylate cyclase, sialic acid, P-adrenergic receptors (12), and regulatory and catalytic subunits of type I1 CAMP-dependent protein kinase (27). Although several investigators have recently suggested that sarcolemmal vesicles can actively transport Ca2+, the levels of Ca2+ accumulated in these studies were only about 1% of those presently reported, and the Ca2+ uptake was not stimulated by oxalate (34, 35). The vesicles described presently did not appear to be mitochondria because their Ca2+ uptake was insensitive to sodium azide, and their ATPase activities were only marginally inhibited by this agent. Moreover, purified cardiac sarcolemmal membranes and mitochondria had none of the blue staining protein bands observed in the sarcoplasmic reticulum fractions after polyacrylamide gel electrophoresis and staining with Stains-AlL2 Thus it seems most likely that the Ca2+ uptake reported presently was accounted for by different subpopulations of sarcoplasmic reticulum vesicles, although it cannot be definitely excluded that a subpopulation of sarcolemmal vesicles containing none of the markers above contributed to some of the Ca" uptake.
Ratios for Ca2+ ions transported per ATP molecule hydrolyzed were substantially less than two for both populations of vesicles. Since the accepted stoichiometry is two for fast skeletal muscle preparations (26), one could argue that the cardiac vesicles were damaged during the Ca2+-loading step. However, we have obtained similar coupling ratios for crude vesicles not preloaded with Ca2+ (8, 18). Although Tada et al. have reported Ca2+ transport stoichiometries of two for cardiac preparations (36), most other investigators have obtained coupling ratios of less than one (37-40). Thus, the coupling ratios reported presently were compatible with those obtained by most of the investigators who have used cardiac preparations.
For ryanodine-sensitive vesicles it was apparent that calculated coupling ratios could be increased dramatically by either ryanodine itself (Fig. 4), or Ca2+ concentration (Fig. 7). Ryanodine, an alkaloid known to cause relaxation in intact heart (41), was previously postulated to increase net Ca2+ uptake into crude preparations of cardiac microsomes by blocking passive efflux of Ca2+ from these vesicles (8). However, such an increase in net Ca2+ uptake was also compatible with an increased number of Ca'+ ions transported per ATP molecule hydrolyzed (3). The data of Fig. 5 clearly showed that ryanodine stimulated Ca2+ exchange in ryanodine-semitive vesicles after net Ca2+ uptake had reached a steady state.
Katz and co-workers (42) and Kirchberger and Wong (43), utilizing skeletal muscle and cardiac sarcoplasmic reticulum preparations, respectively, have suggested that such experiments, conducted at the steady state, measure the rate of Ca2+ influx. If such interpretations are correct, the present results would indicate that ryanodine increased the rate of Ca2+ idlux at least %fold into ryanodine-sensitive vesicles, without changing the rate of Ca'+-dependent ATP hydrolysis. Thus, ryanodine may stimulate net Ca2+ uptake into cardiac sarcoplasmic reticulum vesicles by increasing the Ca2+-pumping efficiency, not by blocking passive Ca2+ efflux from the vesicles. A similar effect of a drug on a cation-transporting enzyme has recently been postulated for quercetin, which increases severalfold the Na+ .pumping efficiency of the (Na+,K+)-ATPase enzyme klated from Ehrlich ascites tumor cells (44).
Ryanodine-sensitive vesicles also exhibited a jump in their Ca2+ uptake that occurred over a narrow range of ionized Ca2+ concentration. Caz+ activation of this process was clearly dissociated from Ca2+ activation of ATP hydrolysis. Because the dramatic increase in Ca2+ uptake occurred without a significant change in Ca2+-dependent ATPase activity, apparent coupling ratios for Ca2+ transport increased substantially during the jump. The steepness of the response as a function of Ca" concentration further emphasized that this process was unrelated to Ca2+ activation of ATPase activity.
The jump in Ca2+ uptake exhibited by the ryanodine-sensitive vesicles required preincubation of the vesicles at a slightly alkaline pH (Fig. 12). Tate et al. have recently described an uncoupling effect of alkaline pH on Ca2+ transport by cardiac microsomes incubated in the presence of oxalate (45). In this study, an alkaline pH (7.2) was observed to reduce Ca2' uptake, but the rate of Ca" transport was returned to control levels by acidifying the pH to 6.6 suddenly (45). Ca2+dependent ATPase activities, on the other hand, were always greater at pH 7.2 (45). In the present study, for the ryanodinesensitive vesicles, it was shown that the restoration of Ca2+ uptake by acidic pH was dependent upon the ionized Ca2+ concentration. Thus, as emphasized by Rossi et al., it may be an oversimplification to view the Ca2+ pump stoichiometry as a fixed number (46). ATP concentration (46), pH (45,46), and ionized Ca" concentration (present study) have now all been reported to affect the apparent stoichiometry of the Ca2+ pump of sarcoplasmic reticulum vesicles.
Finally, our results allow some conclusions to be made about the relative contributions of the ryanodine-sensitive and -insensitive vesicles to the total Ca2+ uptake of crude cardiac sarcoplasmic reticulum preparations. In the absence of ryanodine, most of the Ca2+ uptake by crude preparations appears to be contributed by ryanodine-insensitive vesicles, even though these vesicles account for only 10 to 15% of the total protein content (Table I). Ryanodine increases Ca" uptake by such crude cardiac microsomes approximately 3fold (B), because the ryanodine-sensitive vesicles present substantially outnumber these ryanodine-insensitive vesicles (Table I, legend). This suggests that previous studies on Ca'+ transport would have unknowingly focused on the properties of ryanodine-insensitive vesicles, since most previous studies have been conducted in the absence of the drug. However, it remains possible that agents other than ryanodine, shown previously to alter Ca2+ transport, may have selectively interacted with one population of vesicles or the other. The abundance of ryanodine-sensitive vesicles in cardiac microsomal preparations suggests that these membranes may play an important role in regulating intracellular Ca2+ ion concentration in intact myocardium.