Dependence of Ionophore-and Caffeine-induced Calcium Release from Sarcoplasmic Reticulum Vesicles on External and Internal Calcium Ion Concentrations*

The effects of the ionophore, X537A, and caffeine on ATP-dependent calcium transport by fragmented sarcoplasmic reticulum were studied in the absence (calcium storage) or presence (calcium uptake) of calcium-precipitating anions. The ionophore caused rapid calcium release after calcium storage, the final level of calcium storage being the same whether a given concentration of X537A was added prior to initiation of the reaction or after calcium storage had reached a steady state. Although 10 to 12 PM X537A caused approximately 90% inhibition of oxalate-supported calcium uptake when added prior to the start of the reaction, this ionophore concentration caused only a small calcium release when added after a calcium oxalate precipitate had formed within the vesicles, and only slight inhibition of calcium uptake velocity when added during the calcium uptake reaction. When low initial calcium loads limited calcium uptake to 0.4 pmol of calciumlmg of protein, subse- quent calcium additions in the absence of the ionophore led to renewed calcium uptake.

The effects of the ionophore, X537A, and caffeine on ATPdependent calcium transport by fragmented sarcoplasmic reticulum were studied in the absence (calcium storage) or presence (calcium uptake) of calcium-precipitating anions. The ionophore caused rapid calcium release after calcium storage, the final level of calcium storage being the same whether a given concentration of X537A was added prior to initiation of the reaction or after calcium storage had reached a steady state. Although 10 to 12 PM X537A caused approximately 90% inhibition of oxalate-supported calcium uptake when added prior to the start of the reaction, this ionophore concentration caused only a small calcium release when added after a calcium oxalate precipitate had formed within the vesicles, and only slight inhibition of calcium uptake velocity when added during the calcium uptake reaction.
When low initial calcium loads limited calcium uptake to 0.4 pmol of calciumlmg of protein, subsequent calcium additions in the absence of the ionophore led to renewed calcium uptake. Uptake of the subsequent calcium additions was not significantly inhibited by 10 to 12 FM X537A.
These phenomena are most readily understood in terms of constraints imposed by fixed Ca, (calcium ion concentration inside the vesicles) on the pump-leak situation in sarcoplasmic reticulum vesicles containing a large amount of an insoluble calcium precipitate, where most of the calcium is within the vesicles and Ca, is maintained at a relatively low level. These constraints restrict calcium loss after calcium permeability is increased because calcium release can end when the calcium pump is stimulated by the increased Ca,, (calcium concentration outside the vesicles) so as to compensate for the increased efflux rate. In contrast, an increased permeability in vesicles that have stored calcium in the absence of a calcium-precipitating ion causes a much larger portion of the internal calcium store to be released. Under these conditions calcium storage capacity is low so that release of stored calcium is less able to raise Caq, to levels where the calcium pump can compensate for the increased efflux rate The constraints imposed by anion-supported calcium uptake explain the finding that more calcium is released by X53'7A or caffeine when these agents are added at higher levels of Ca,,, and that more calcium leaves the vesicles in response to a given increase in calcium permeability at higher Cq. Although such calcium release is amplified by increased C%, the amplification is attributable to the constraints described above and does not represent a "calciumtriggered calcium release." Sarcoplasmic reticulum vesicles actively transport calcium in the presence of MgZ+ and ATP. When the ionized Ca'+ concentration in the external medium is in the micromolar range, the vesicles rapidly accumulate -100 nmol of calcium/ mg of protein (3)(4)(5). In the presence of calcium-precipitating anions such as oxalate and phosphate, which can lead to formation of calcium precipitates within the sarcoplasmic reticulum of partially disrupted muscle cells (6, 7), the vesicles take up a much larger amount of calcium (8-10). These anions, by stabilizing Ca, , 1 increase net calcium transport by maintaining Ca2+ concentration inside the vesicles at a low level, thereby reducing an inhibitory effect of high internal calcium on the transport process (11,12).
The present report examines the ability of the calciumprecipitating anions, oxalate and phosphate, to modify calcium release from sarcoplasmic reticulum vesicles that is induced by agents believed to increase membrane permeability. Although these anions have proven extremely useful in the analysis of the mechanism of calcium transport by the sarcoplasmic reticulum, the stabilization of Ca, is shown to impose ' The abbreviations used are: Ca,, calcium ion concentration inside the microsomal vesicles. calculated from the solubilitv nroduct of calcium oxalate of 2 x 10m6~' (9, 10) or of calcium phospdaie of 7.5 x lo-" M' (W. Hasselbach Release-Vesicles loaded with calcium in the presence of MgATP, but in the absence of calcium-precipitating anions, promptly released calcium when X537A was added. This X537A-induced calcium release, like that of the initial calcium storage reaction, was too rapid to be measured by the Millipore filtration technique. As previously reported (18), the quantity of calcium released was dependent on X537A concentration and did not increase significantly when the time of exposure to the ionophore was prolonged. The final level of calcium storage was the same whether the ionophore was added before, or after, calcium storage had reached steady state (Fig. 2). The velocity of ionophore-induced calcium release was much more rapid than that seen when Ca'+ concentration in the medium was lowered with EGTA ( Fig. 3 Calcium Release-Addition of 12 PM X537A to vesicles after oxalate-supported calcium uptake had reached a steady state caused a small amount of calcium to be released (Fig. 4). This small calcium release, approximately 100 nmol/mg of protein under the conditions of this experiment, was transient; after approximately 60 s the released calcium was again taken up by the microsomes. In contrast, the same concentration of X537A added to the reaction mixture prior to initiation of the calcium transport reaction caused almost complete inhibition of subsequent calcium uptake (Fig. 4). Addition of 12 PM X537A at various times during the initial uptake reaction slowed calcium uptake velocity without affecting calcium uptake capacity significantly (Fig. 5). Calcium Uptake -The ability of prior calcium uptake to reduce the effects of X537A could be shown when the ionophore was added to a reaction mixture after a small amount of oxalate-supported calcium uptake had taken place. In studies where calcium uptake was limited by the amount of added CaCl, to a maximum of 0.4 pmol of calcium/mg of protein, 12 PM X537A caused only the previously described transient calcium release (Fig. 6). Subsequent addition of WaC12 after a limited calcium uptake had reached steady state induced further calcium uptake (Fig. 6, Arrow A) in the control reaction. No significant inhibition of subsequent calcium uptake was seen when WaCl, was added to the X537A-containing reaction mixture either during (Fig. 6, Arrow B) or after (Fig. 6, Arrow C) the transient calcium release.
Reduction of Ca'+ concentration in the medium by millimolar concentrations of EGTA after oxalate-supported calcium uptake had reached steady state caused a slow calcium One such curve is shown in Fig. 9, which is based on published data for the Ca'+ dependence of calcium uptake velocity in 2.5 mM oxalate (16, see also In the case of acetyl phosphate-supported calcium uptake, the initial calcium uptake reaction was allowed to take place at 33". At t = 15 min, both reaction mixtures were brought to 25" and X537A was added to an aliquot of each at a final concentration of 10 FM (0, 0). Samples were also taken from reaction mixtures to which no ionophore was added (0, W.  Figs. 9 and lo), calcium influx and efflux rates in 2.5 mM oxalate will be 0.07 pmol/mg of proteinlmin at a Ca, of 0.1 PM, 0.3 pmol/mg of protein/min at a Ca,, of 0.5 pM, and 1.2 pmol/mg of protein/min at a Ca, of 1.5 pM. These values, which will be used in subsequent calculations (see below) are based on the calcium uptake measurements in Table I and in  Ref. 16.
The "leakiness" of these vesicles can be described by a calcium permeability coefficient, K,, that is calculated most simply from the relationship: According to Equation 1, K, is 0.0375 min', where calcium efflux rate is 0.3 pmol/mg of protein/min and Ca, is 8 PM (2.5 mM oxalate), at a Ca, of 0.5 PM.
Calcium permeability appears to be increased approximately 4-fold by 10 PM X537A as this ionophore concentration reduces calcium storage capacity in the absence of calciumprecipitating agents to approximately one-fourth of its initial level. If this value is correct, 10 PM X537A will increase K, in the above example from 0.0375 to 0.150 mini, thereby causing efflux rate at the constant Ca, (8 PM) to increase 4-fold to 1.2 pmol/mg of proteimmin.
As a result, calcium will be released from the vesicles into the surrounding medium until the increased Ca,, accelerates calcium influx via the pump and causes a new steady state to be reached when calcium influx and efflux rates again become equal. The ability of calcium release to restore a new steady state is made possible by the large amount of the calcium precipitate within the vesicles. According to previous data regarding the Ca"+ dependence of the calcium pump (Table I, 16), the new steady state in the above example, where calcium influx and efflux rates are 1.2 pmol/mg of protein/min, would be reached at a Caq, of approximately 1.5 FM. This response to X537A is depicted by the dotted arrow labeled A in Figs. 9 and 10, which predicts that the new steady state will be reached after calcium release from the vesicles amounts to 1.0 pmol of calcium/liter of the reaction mixture raises Ca, to 1.5 PM. This calcium release, expressed in terms of the concentration of vesicles, would be approximately 100 nmol/mg of protein at a protein concentration of 10 pg/ml.
If the same 4-fold increase in calcium efllux is produced by the addition of 10 pM X537A at a lower initial Ca,, of 0.1 pM, where at steady state both calcium influx and calcium efflux rates are 0.07 pmol/mg of proteinlmin (Table I, 16), the new efflux rate will be increased 4-fold to 0.28 pmol/mg of protein/ min. This corresponds to a rise in K, from 0.00875 to 0.035 min'. Under these conditions, the new steady state reached when calcium release from the vesicles stimulates the calcium pump will occur when both calcium influx and efflux rates are 0.28 pmol/mg of protein/ml, i.e. when Ca,, rises to 0.4 pM (dotted arrow B in Figs. 1 and 2). To achieve this higher level of Ca,,, net calcium efflux from the vesicles must be only 0.3 pmol/liter of the reaction mixture (30 nmol/mg of protein at a protein concentration of 10 pg/ml) as depicted by the dotted arrow labeled B in Figs. 9 and 10. The initial level of Ca,, thus determines the amount of calcium that must be released from the vesicles after a given increase in permeability.
The higher the initial Ca,, the more calcium must be released from the internal calcium precipitate into the medium outside the vesicles to restore a steady state through stimulation of the calcium pump.
The amount of calcium release will depend on Ca+ as long as the vesicles contain most of the calcium in the reaction mixture and the Ca'+ concentration gradient remains below the maximum that can be achieved by the ATP-dependent calcium pump. If Ca, is sufficiently high or Ca, so low that this gradient approaches its maximum of 3000 to 4000 (10,23), then the hyperbolic relationships shown in Fig. 10 will not hold as, for example, in the case of calcium storage in the absence of the calcium-precipitating anions. Furthermore, a new steady state can be restored as shown in Fig. 10 only when the calcium store in the vesicles is sufficient to raise Ca,, to levels which allow calcium influx rate via the calcium pump to equal the new, higher rate of calcium efflux. This will not usually be true for calcium storage in the absence of oxalate or phosphate, where only a small amount of calcium is retained in the vesicles.
Predicted Effects of Cai on Ionophore-induced Calcium Release -The properties of the pump-leak system operating under the constraints described above also predict that the amount of X537A-induced calcium release increases with increasing Cai. If, for example, the 4-fold increase in calcium permeability described above is brought about by the addition of the ionophore where the initial value of K,, in Equation 1 is 0.00875 and Ca is 20 ,UM (1 mM oxalate), the response will be altered from that shown by the dotted arrow B in Fig. 10 (which describes the response at a Ca of 8 PM) to that shown by the dotted arrow B'. If the rate of calcium influx was independent of Ca, the initial steady state in the absence of the ionophore would be reached where calcium efflux and calcium influx rates are 0.175 pmol/mg of proteimmin (0.00875 min' times 20 pM Ca,). Increasing Ca, in this range has a slight inhibitory effect on the calcium pump (11,12), so that an increase in Ca, from 8 to 20 PM will inhibit calcium influx rate about 20% (11). Steady state rates of calcium influx and cal-

Calcium
Release from Sarcoplasmic Reticulum cium efflux at any level of Ca, will, therefore, be lower where of this preparation to take up calcium. At the steady state of Ca is 20 pM than where Cq is 8 pM. An initial Ca, of 0.3 PM at calcium uptake by these vesicles, Ca, at the time of X537A this steady state has been calculated from the data in Table I addition ranged between 2.3 and 3.1 PM for the four reactions and Ref. 16, which reflects the 20% inhibition of calcium shown in Fig. 11A. Increasing X537A concentration from 5 to uptake rate when Ca, is increased from 8 to 20 pM. Under the 40 pM increased the amount of calcium released, as shown by latter conditions, a 4-fold increase in K, will increase calcium the shaded area in the figure. At the lower ionophore concenefflux rate from 0.175 to 0.7 pmol/mg of proteimmin.
Accord-trations, net calcium release was transient, whereas higher ing to the data in Table I and Ref. 16, modified to take into concentrations of X537A caused calcium release to be susaccount the 20% inhibition of the calcium pump at a higher tained. When this membrane preparation was examined un-Ca, the new steady state in which calcium influx rate equals der identical conditions at a higher protein concentration of the higher rate of calcium efflux would occur at a Ca,, of 13.7 pg/ml, the maximal possible calcium uptake of 3.65 pmol/ approximately 1.1 PM (dotted arrow B' in Fig. 10). The calmg of protein was less than the capacity of these vesicles, cium release needed to achieve this higher Ca,,, 0.8 ~mol/liter, which was approximately 5 pmol/mg of protein (Fig. 11A). As is more than twice that brought about by the same increase in a result, calcium uptake reached a steady state when Ca,, fell K, at the lower level of Ca, (dotted arrow B in Fig. 10). The to 0.11 pM in the four reactions shown in Fig. 11B. Under initial level of Ca, thus determines the amount of calcium that these conditions, increasing X537A concentration from 5 to 40 must be released from the vesicles after a given increase in pM again induced a calcium release that was small and trancalcium permeability.
The higher the initial Czq, the more sient at the lower ionophore concentrations, and larger and calcium will be released.
sustained when X537A concentrations exceeded approxi-The foregoing analysis does not explain an effect of increas-mately 20 pM. At all X537A concentrations, however, the ing Ca, to augment calcium release following a given increase amount of calcium released initially was less in the studies in calcium permeability at any initial level of Ca,,. It has been carried out at the higher protein concentration. The effect of shown, however, that the Ca*+ dependence of initial calcium protein concentration cannot be attributed solely to the changuptake velocity varies with oxalate concentration such that ing X537A to protein ratio as it was found that the inhibitory the apparent Kc, decreases when the oxalate level is decreased effects of 4 pM X537A on the initial rate of oxalate-supported (24). This effect would reduce the ability of higher levels of Ca, calcium uptake decreased by only 25% when protein concento stimulate the calcium pump following a given increase in tration was increased lo-fold, from 9 to 90 pg/ml. The quanticalcium permeability in the presence of 1 mM oxalate, comtative differences between the effects of X537A seen in Fig. 11, pared to that in 2.5 mM oxalate. The resulting alteration in the A and B, are, instead, attributable primarily to differences in curve shown in Fig. 9 to one showing evidence of saturation by the levels of Ca, at the time of X537A addition (see above). Ca, in the micromolar range, coupled with the inhibition of

Effect of Ca,, on X537A-induced Calcium
Release -Comparthe calcium pump at higher levels of Ca, (see above), would ison of the shaded areas in Fig. 11, A and B, both of which reduce the ability of the calcium released by an increase in utilize similar scales, shows that the total amount of calcium calcium permeability to stimulate calcium influx. In this way, released initially at the higher protein concentration (Fig.  a given increase in calcium permeability at a higher level of 11B) was much less than at the lower protein concentration Ca, would cause a greater calcium release than the same (Fig. 11A). The possibility that the decreased amount of calincrease in calcium permeability at low Ca at any given level cium released at the higher protein concentration reflected a of Ca(]. reduction in X537Alprotein ratio was excluded by comparing The response to agents such as X537A after calcium storage shown in Fig. 11, A andB. The data in Fig. 12 show that at the differs from that described above in the case of oxalate-sup-higher protein concentration of 18 pg/ml, the amount of calcium release induced by 5 PM X537A was approximately 15 ported calcium uptake. In the latter, the ability of internal nmol/mg of protein after 30 s, whereas doubling the X537A calcium stores to increase Ca,, can allow stimulation of the concentration to 10 PM increased the amount of calcium recalcium pump to restore a new steady state after only a small leased to only approximately 25 nmol/mg in 30 s. At the lower amount of calcium is released. In the case of calcium storage, protein concentration of 9 pg/ml, on the other hand, 5 PM however, the calcium content of the vesicles is much less and X537A caused a release of approximately 80 nmol of calcium/ the levels of Ca, are much greater at any initial level of Cs,. mg after 30 s. In the experiment shown in Fig. 12, Ca,, at the For this reason, the ability of the vesicles to compensate for an time of X537A addition was 0.11 PM at the higher protein increased calcium et&x rate by increasing calcium influx via concentration, and 1.7 PM at the lower protein concentration. the ATP-dependent calcium pump through increased Ca,, is This experiment, like that shown in Fig. 11 effects of release on Ca, could also be shown when Ca, was lowered by varying X537A concentration on the release of calcium after EGTA immediately before addition of the ionophore. Low calcium uptake in 2.5 mM oxalate had reached a steady state concentrations of EGTA added to the reaction mixture 30 s were examined at two protein concentrations (Fig. 11). At the prior to the transfer of an aliquot to the tube containing the lower protein concentration of 7.6 pg/ml (Fig. HA), the total ionophore significantly reduced the extent of X537A-induced amount of added 4"CaC1, (50 PM) allowed a maximum calcium calcium release (Fig. 13). As was the case when Ca, was uptake of 6.6 pmol/mg of protein, which exceeded the capacity reduced by decreasing the total calcium/protein ratio, a de- FIG. 11. Dependence of calcium release on X537A concentration. A single preparation was used to study calcium release at protein concentrations of 7.6 pg/ml (A) and 13.7 kg/ml (B). Calcium uptake was initiated by addition of WaCl, to a final concentration of 50 pM in the presence of 2.5 mM oxalate at t = 0 as described under "Materials and Methods." At t = 8 min, a 5-ml aliquot of the reaction mixture was transferred to a tube containing 25 ~1 of an ethanolic crease in ionophore-induced calcium release occurred even though the Ca"+ concentration gradient across the vesicles, CaJC~, was increased due to the lowered Ch. (The level of Ca in the experiments shown in Figs. 11 and 12 was held at 8 pM by the 2.5 mM oxalate, and in Fig. 13 at 4 pM by the 5 mM oxalate.) The findings in all of 12 consecutive experiments where calcium release was induced with 10 PM X537A after calcium uptake in 2.5 or 5.0 mM oxalate had reached a steady state are plotted in Fig. 14. An inverse relationship between Ca,, and the amount of calcium released is clearly seen. The line in Fig. 14, which was calculated by assuming that the ionophore caused a 4-fold increase in calcium permeability, as described earlier, illustrates the general correspondence of the extent of net calcium release predicted by the analysis presented earlier in this report to the data obtained in studies of the effects of 10 PM X537A on oxalate-supported calcium uptake. The agreement between measured calcium release and that predicted by this analysis is reasonably close, even though the data in Fig. 14 include measurements carried out over a range of protein concentrations at both 2.5 and 5.0 mM oxalate. Calcium uptake velocity is minimally influenced by changing oxalate concentration in this range (11). Effects of Cai on X537A-induced Calcium Release-The The difference between the measured levels of calcium content where the control levels were greater than those in the presence of the ionophore are shaded.
concentration of calcium within the sarcoplasmic vesicles can be varied by allowing calcium uptake to proceed in the presence of various concentrations of oxalate or phosphate. In the present studies, the levels of Ca, that would be reached in the presence of various concentrations of the calcium-precipitating anions were first estimated in a series of pilot runs that allowed more complex experiments to be carried out in which the effects of different levels of Ca could be compared at similar Ca, concentrations.
Typical experimental findinks are shown in Fig. 15 where Ca, at the time of X537A addition varied between 0.1 and 1.8 PM. This example was chosen because it illustrates the consistent finding that the amount of calcium released after X53'7A addition increased with increasing Cq in a manner that is independent of the effects of Ca, described above.
The dependence of X537A-induced calcium release on C% is plotted as the observed response in Ca, after X537A addition following calcium uptake in 10 to 50 mM phosphate in Fig. 16. This experiment, in which C% ranges between 750 /AM (at 10 mM phosphate) and 150 PM (at 50 mM phosphate), illustrates the typical finding that calcium release became greater as the concentration of the calcium-precipitating anion was lowered. This situation, where X537A-induced calcium release increases as the Ca2+ concentration gradient is increased by FIG. 12. Effects of varying protein and X537A concentrations on X537A-induced calcium release. Calcium release from sarcoplasmic reticulum vesicles at two protein concentrations was induced by 5 FM X537A (Panels A and C) or 10 pM X537A (Panel B). Symbols are the same as in Fig. 11. Conditions were chosen so that the X537A/ protein ratios in Panels B and C were the same, whereas the Ca, concentration at the time of X537A addition (t = 8 min) varied as shown at the left-hand side of each panel. Experimental conditions were as described under "Materials and Methods" and the legend to Fig. 11. increasing Ca, can be contrasted to that described in the preceding section where calcium release decreases when the Ca2+ concentration gradient is increased by lowering Ca+ The increasing amount of X537A-induced calcium release seen as Ca was increased (tabulated at the right in Fig. 16) caused an increase in the final, steady state, Ca, concentration reached at the conclusion of the X537A-induced calcium release. The maximum Ca*+ concentration ratios, Ca#&,, across the membrane vesicles prior to the addition of X537A for the experiment shown in Fig. 16 are tabulated in Table III, along with the calcium concentration ratios at the steady state reached after X537A addition.
The values for the concentration ratios prior to X537A addition shown in Table III are not accurate measurements of the maximal concentrating ability of these sarcoplasmic reticulum vesicles as Ca, in the control reactions continued to decrease slowly between 10 and 20 min after the reactions were started, reaching levels that gave concentration ratios up to 4800 (data not shown). Because precautions were not taken in this study to allow control reactions to reach steady state, the data in Table III (and elsewhere in the present report) cannot be used to assess the maximal concentrating ability of these preparations.
Effects ofCaffeine -It is well known that caffeine can cause calcium release from the sarcoplasmic reticulum of mammalian muscle (251, although these effects have been seen only at high caffeine concentrations and are somewhat variable. In studies of canine cardiac sarcoplasmic reticulum,Y for exam-  Calcium release was measured after X537A addition at t = 8 min for the experiment with 5 mM oxalate (0) and at t = 10 min for those with 1 mM oxalate (01 and 50 mM phosphate (A).
ple, no detectable effects of 5 mM caffeine were seen when the drug was added after oxalate-supported calcium uptake in the presence of extremely low levels of Ca,. In view of the present findings with the ionophore, X537A, the dependence of the effects of caffeine on Ca, were examined.
Caffeine was found to induce a transient calcium release  (0)  after phosphate-supported calcium uptake under conditions where Ca, was high, above 1 PM, whereas when Ca, was less than 1 PM, virtually no calcium release was seen (Fig. 17). Transient calcium release following caffeine addition was also noted in frog sarcoplasmic reticulum by Ogawa (91, who suggested that the transient nature of the caffeine effect might reflect an action of the drug to promote a "calcium-triggered calcium release." In view of the findings with X537A discussed above, however, it is much more likely that the transient calcium release following addition of caffeine to sarcoplasmic reticulum vesicles is due to the fact that the initial rate of calcium release induced by caffeine initially exceeds that at which the calcium is pumped back into the vesicles by the calcium pump. Similarly, the reported finding that caffeine reduces calcium uptake capacity less than it slows calcium uptake rate (25) is in accord with the effects of X537A reported in the present paper. Reduction of total calcium uptake capacity by caffeine at high but not low Ca, (Fig. 5 Fig. 17 of the present paper. This evidence that caffeine acts in a manner similar to X537A, an agent that increases calcium efflux, can account for the failure of caffeine to inhibit ATPase activity and its ability to reduce the coupling ratio between calcium transport and ATP hydrolysis (26). The potency of caffeine to induce calcium release was studied with protocols designed so as to magnify the effects of a leak-producing agent, i.e. where both Ca, and Ca, were relatively high. This was accomplished by examining the effects of caffeine with 50 mM phosphate as the calcium-precipitating agent, and with a slight excess of total calcium relative to protein. Under these conditions, concentrations of caffeine as low as 0.25 mM evoked significant calcium release (Fig. 18). DISCUSSION The present findings regarding the ability of low concentrations of X537A to inhibit calcium transport by vesicles derived from the sarcoplasmic reticulum are in accord with earlier studies in both skeletal (18,27,28) and cardiac (19) preparations. In view of the broad specificity of X537A (19, 27, 29), the possibility that inhibition of calcium transport is due to an interaction between the free ionophore and cations in the reaction mixture was excluded by the finding that washing of the vesicles after brief exposure to X537A does not prevent the inhibition of subsequently measured calcium transport activities (Table II). The failure of washing to reverse the X537A effects indicates that the ionophore, which is poorly soluble in aqueous solutions, remains bound to the membranes.
The inhibitory effects of X537A on calcium storage in the absence of calcium-precipitating anions are not dependent on the Ca'+ concentration in the medium and the ionophore causes no significant change in the Ca'+ concentration at which calcium storage capacity is half-maximal (Fig. 1). Similarly, the extent of inhibition by X537A of oxalate-supported calcium uptake velocity is not influenced by Ca, (Table I).
The finding that the final level of calcium storage is the same whether X537A is added prior to or after the addition of calcium (Fig. 2) is in accord with a previous report (18)  the addition of calcium differ markedly from those seen when the ionophore is added prior to initiation of the calcium uptake reaction (Fig. 4). Incubation of the vesicles in 10 to 12 pM X537A prior to the addition of calcium caused subsequent calcium uptake to be inhibited almost completely, whereas addition of the same final ionophore concentration after completion of the calcium uptake reaction caused only a transient release of a small amount of calcium as seen in Fig. 4. Similar effects have been seen in cardiac preparations (19). The effects of a calcium oxalate precipitate within the vesicles on the response of added X537A are shown in different ways in Figs. 5 and 6. In the experiment shown in Fig. 5, the ionophore was added at various times after the calcium uptake reaction had been started. Although the ionophore slows the reaction, calcium uptake continues to virtually the same level as that reached in control reactions even though this concentration of X537A, when added prior to the start of the calcium uptake reaction, completely inhibits calcium uptake (Fig. 4). Because the extent to which calcium uptake is inhibited by X537A is not influenced by Ca,, in the range between 0.24 and 1.32 pM (Table I), preservation of the ability to transport calcium in the presence of the ionophore cannot be attributed to the reduction in Ca, that occurs when the vesicles take up calcium. The effect of prior calcium oxalate precipitation within the vesicles is also seen when additional calcium is added after the vesicles have been preloaded with a small amount of calcium oxalate to approximately 15 to 20% of their total capacity (Fig. 6). These findings, that calcium uptake can continue in the presence of X537A, are not altogether unexpected as it is well established that the ionophore does not inhibit the calcium pump ATPase (18,19,28). Inhibition of the effects of X537A on calcium release after Ca, is stabilized at a low level by calcium oxalate precipitation cannot be explained on the basis that the precipitate within the vesicles is slow to dissolve, and so is retained on the Millipore filters used to measure calcium uptake, because destruction of the membranes with lipid solvents causes rapid calcium release from these calcium oxalate precipitates (30). Furthermore, prior calcium oxalate precipitation not only inhibits ionophore-induced calcium release, but also allows subsequent addition to this precipitate to take place in the presence of the ionophore.
The ability of a calcium oxalate precipitate within the vesicles to inhibit calcium release cannot be attributed to the ability of the high internal oxalate concentration to maintain internal Ca'+ concentration at a level sufficiently low so as to prevent significant calcium efflux when permeability is increased by the ionophore. Because the solubility product for calcium oxalate is approximately 2 x lo-' M' under the conditions of these studies, the Ca'+ concentration inside the vesicles will be fixed at 8 PM by 2.5 mM oxalate, so that in experiments where calcium uptake is limited by a low calcium/protein gradient (Fig. 6), over 90% of the calcium enters the vesicles to produce significant calcium concentration gradients across the membranes. Furthermore, calcium release by X537A after oxalate-supported calcium uptake is not markedly accelerated when Ca, is lowered to approximately 0.01 PM by the addition of high EGTA concentrations (Fig. 7). Modification of the response of sarcoplasmic reticulum vesicles to X537A and caffeine when calcium precipitates form within the vesicles is explained most easily by the constraints imposed by fixed Ca, in vesicles that contain large amounts of an insoluble calcium precipitate.
When Ca, is maintained constant under conditions where most of the calcium is within the vesicles, and where the calcium concentration gradient, CaJC&, is less than the maximum that can be achieved by the calcium pump, these agents cause a transient calcium release that ends when the calcium pump is stimulated by the increased Ca, so as to compensate for the increased eMux rate. Thus, where calcium stores within the vesicles are high relative to the amount of calcium outside the vesicles, only a small portion of the internal calcium is released by X537A (Fig. 4) or caffeine (Fig. 17). In contrast, when similar amounts of X537A are added after calcium storage in the absence of calciumprecipitating ions (Fig. 2), a much larger fraction of the internal calcium is released. The absolute quantity of calcium released by the ionophore after calcium storage is, however, less than after anion-supported calcium uptake, due to the lower calcium storage capacity of the vesicles (16) and the much higher Ca levels. As a result, release of even a large fraction of the calcium stored in the absence of calcium-precipitating anions cannot raise Ca,, to levels that allow the calcium pump to compensate for the increased absolute rate of calcium efflux that accompanies a given increase in calcium permeability.
The constraints described above, which are imposed by large internal calcium stores and a constant Ca,, require that more calcium be released to achieve a new steady state when cal-cium permeability is increased at high Ca,. Although X537Aand caffeine-induced calcium release are thus amplified by increased Ch, this amplification does not represent a "calcium-triggered calcium release." Instead, the amplification of calcium release by increasing Ca,, can be attributed solely to the constraints described above.
The present findings indicate that the constraints imposed upon the pump-leak situation in sarcoplasmic reticulum, and especially the amplification of calcium release by increasing Ca,,, must be considered in studies of the actions of agents that increase the calcium permeability of the membranes. Becognition of these constraints also permits a clearer understanding of spontaneous changes in the calcium permeability of these membranes (31).