INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

INTRACELLULAR INORGANIC PHOSPHATE concentration ([P_i]) increases to ~10 mM at the onset of hypoxia or ischemia in cardiac muscle (8, 26) and is accompanied by a rapid fall in contractility. Although other metabolic changes may occur at the same time (19, 33), the increased intracellular [P_i] is thought to be a significant cause of reduced contractility within the first 1-2 min. P_i acts directly on the contractile proteins to reduce Ca²⁺-activated force (13, 22). Furthermore, studies have shown that millimolar levels of P_i reduce the amount of Ca²⁺ available for release from the sarcoplasmic reticulum (SR) of cardiac muscle (37, 42), and these studies suggest that the inhibitory action of P_i on the SR may be an important contributor to the rapid fall in contractility. P_i may inhibit SR function by actions within the SR lumen, on SR Ca²⁺ channels, or SR Ca²⁺ pumps. Fryer et al. (12) suggest that P_i may inhibit Ca²⁺ release from skeletal muscle SR by precipitating Ca²⁺ inside the SR lumen. But similar experiments on cardiac muscle failed to find evidence that precipitation limits SR Ca²⁺ release (39). It has been shown that P_i directly activates the cardiac ryanodine receptor and may increase Ca²⁺ efflux from the SR via this route (23). P_i may also activate a Ca²⁺ efflux pathway that is independent of the ryanodine receptor (39). Direct effects of P_i on the SR Ca²⁺ pump to reduce the Ca²⁺ sensitivity of Ca²⁺ uptake was suggested by work on cardiac SR vesicles (25), but that study demonstrated that P_i-induced increase of Ca²⁺ uptake was also possible. Alternatively, P_i may reduce only the maximum uptake capacity(41). That P_i can cause reversal of skeletal and cardiac SR Ca²⁺ pumps has been demonstrated amply in SR vesicle preparations (15, 40). Under these circumstances, Ca²⁺ efflux via the Ca²⁺ pump is linked to ATP synthesis from ADP and P_i. However, although evidence exists to suggest that Ca²⁺ pump reversal may limit the maximum Ca²⁺ content of cardiac muscle SR (35), P_i-induced pump reversal has not been demonstrated directly in native SR at concentrations of P_i, ATP, Mg, and pH normally observed in the early stages of hypoxia and ischemia.

The purpose of this study was to examine the effects of P_i on Ca²⁺ fluxes across cardiac SR membrane under conditions where both Ca²⁺ precipitation by P_i and Ca²⁺ leak via the ryanodine receptor are prevented. The actions of P_i are compared with those of specific Ca²⁺ pump inhibitors and a Ca²⁺ ionophore (ionomycin). The results show that P_i decreases the SR Ca²⁺ uptake rate and activates a Ca²⁺ leak from the SR. P_i-induced Ca²⁺ leak from the SR was not present when the SR Ca²⁺ pump was inhibited by thapsigargin. Possible mechanisms linking P_i-induced Ca²⁺ pump inhibition to activation of a Ca²⁺ leak are discussed.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Cell isolation. New Zealand White rabbits (2.5-3.0 kg) were given an intravenous injection of 500 U heparin together with an overdose of pentobarbital sodium (100 mg/kg). Isolated hearts were perfused retrogradely (25 ml/min, 37°C) with a nominally Ca²⁺-free Krebs-Henseleit solution for 10 min. This was followed by perfusion for 10-17 min with recirculated Krebs-Henseleit solution supplemented with 0.6 mg/ml collagenase (type 1, Worthington Chemical), 0.1 mg/ml protease (type XIV, Sigma), and 80 µM CaCl₂. The left ventricular free wall was isolated and incubated separately for 5 min in enzyme solution containing 80 µM CaCl₂ and 4% bovine serum albumin (BSA, fraction V, Sigma). The cell suspensions obtained at the end of the incubation period were filtered into Krebs-Henseleit solution without added Ca²⁺. This procedure routinely produced a 80-90% yield of rod-shaped myocytes. Myocyte concentration was determined by use of a hemocytometer. Cells were lightly centrifuged (2 g for 1 min), and the supernatant was replaced with a mock intracellular solution. This procedure was repeated three times, and the cells were resuspended in a mock intracellular solution to give a final concentration of 5 × 10⁶ cells/ml.

Cell permeabilization and fluorescence measurements. A 0.3-ml aliquot of cell suspension was exposed to 0.1 mg/ml -escin (Sigma) and gently stirred for 1 min. The -escin was removed by centrifuging and resuspending the cells in mock intracellular solution. The cells were then placed in a cuvette, and further solutions were added to give a final volume of 1.5 mM containing 5 mM ATP, the mitochondrial inhibitors carbonyl cyanide (20 µM) and oligomycin (20 µM, Calbiochem), 10 mM oxalate (Sigma), 5 µM ruthenium red (Sigma), and 10 µM fura 2 (Molecular Probes). Cells were maintained in suspension by gentle stirring, and the fura 2 fluorescence within the cuvette was recorded at 30 Hz with a spinning wheel spectrophotometer (Cairn Research). All experiments were done at room temperature (20-22°C).

Solution composition and calibration of fura 2 fluorescence signal. A mock intracellular solution was used to mimic the intracellular environment. It had the following composition (in mM): 130 K⁺, 10 Na⁺, 1 Mg²⁺, 25 HEPES, 140 Cl, and 0.05 EGTA; pH 7.0. Solutions to measure the maximum and minimum fluorescence ratios (R_max and R_min) and the affinity constant of fura 2 contained 10 mM EGTA, 5 mM ATP, 5 µM ruthenium red, and permeabilized cells (at a concentration of 1 × 10⁶ cells/ml). The presence of cells at this concentration did not affect these values. The equilibrium concentrations of metal ions in the calibration solutions were calculated by means of a computer program with the affinity constants for H⁺, Ca²⁺, and Mg²⁺ for EGTA (taken from Ref. 36). The affinity constants used for ATP and creatine phosphate (CrP) were those quoted by Fabiato and Fabiato (9). Corrections for ionic strength, details of pH measurement, allowance for EGTA purity, and the principles of the calculations are detailed elsewhere (30). Free Mg²⁺ concentration was 0.9-1.0 mM in all solutions. Under the conditions used in this study, the apparent affinity constant of fura 2 for Ca²⁺ was 110 ± 20 nM, and the buffer value () was 14.0 ± 0.1, values close to that measured by Grynkiewicz et al. (14) and as noted by Kargacin et al. (21). As described above, ruthenium red was used to block Ca²⁺ leak via the ryanodine receptor, and initial experiments determined that the block achieved with 5 µM was complete, and higher concentrations (20 µM) provided no further block. However, the 20 µM concentration was not used in this study, because previous work suggested that at this concentration, ruthenium red reduced cardiac SR Ca²⁺ pump activity (2, 11, 20). Furthermore, ruthenium red quenches the fluorescence signal from fura 2 (20). However, at 5 µM ruthenium red, the quench appeared to affect the fluorescence at both excitation wavelengths equally, with no effect on the values of the dissociation constant K_d and . Ca²⁺ uptake measurements were made by resuspending the cells in a solution of the following composition (in mM): 120 KCl, 5 Na₂ATP, 5.4 MgCl₂, 25 HEPES, 0.05 K₂EGTA, 0.02 carbonyl cyanide m-clorophenylhydrazone, 0.02 oligomycin, 10 K₂-oxalate, 0.005 ruthenium red, and 0.01 fura 2; pH 7.0. Addition of 10 mM P_i was accompanied by 0.25 mM MgCl₂. The added Mg ensured that free Mg²⁺ levels remained between 0.9 and 1.0 mM in the mock intracellular solution.

Data recording and analysis. The fluorescence ratio signal and the individual 340- and 380-nm wavelength signals were low-pass filtered (3dB at 30 Hz) and digitized at 10 Hz for later analysis. Sections of the trace were converted to plots of [Ca²⁺] against time. Fitting of [Ca²⁺] decay curves to Eq. 3 (Fig. 1C) and linear increases were performed with Origin (Version 5.0, Microcal). SR uptake rate constant (k) and leak rate (l) are expressed with the asymptotic standard error-computed error of the parameter (a measure of the uncertainty of the parameter estimate). Where appropriate, curve fits were compared using an F-test, based on the sum-of-squares (SSQ) difference between the fitted curve and data values. Changes in k and l are calculated relative to the preceding control value. Changes in these parameters are expressed as means ± SE. t-Tests were used to compare the relative changes; P < 0.05 was considered statistically significant.

View larger version (26K): [in this window] [in a new window]   Fig. 1.   A: records of [Ca2+] against time recorded from a cuvette containing 1.5 ml 1 × 106/ml permeabilized myocytes suspended in mock intracellular solution by continuous stirring equilibrated with 10 mM oxalate. Additions of aliquots of CaCl2 are indicated above the trace as increases in total [Ca2+] within the cuvette. Thapsigargin (20 µM) was added at the point indicated. B: the record of [Ca2+] from a section of the trace shown in (A). The broken line through the decay phase indicates the best-fit single exponential curve to the record from 600 nM to the steady state. The broken line through the trace after addition of thapsigargin represents the best-fit straight line to the record 600 nM. Values of rate constant and calculated and measured leak rate are shown on the record. C: a simplified model of the sarcoplasmic reticulum (SR) with Ca2+ pump and Ca2+ leak indicated. Ca2+ chelation/precipitation within the SR lumen is also indicated (CaOx). Below the diagram are equations describing the role of Ca2+ leak and uptake in determining the rate of change of cytosolic [Ca2+] (Eq. 1) and the time course of change in [Ca2+] from a starting [Ca2+] (Ca[0]). [SS], Steady state; t, time.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Ca²+ uptake and release from oxalate-equilibrated permeabilized cardiac myocytes. Figure 1A shows a typical trace of [Ca²⁺] recorded from a cell suspension. Addition of an aliquot of CaCl₂ (10 µl of 10 mM) increases the total [Ca²⁺] within the cuvette by 67 µM and causes a rapid increase of the free [Ca²⁺] to ~0.8 µM. Over the next 10 min, the [Ca²⁺] decayed to below 100 nM. This procedure could be repeated up to five times before changes in the time course of the decay became detectable. Addition of thapsigargin (20 µM) caused an increase in the [Ca²⁺] within the cuvette, representing Ca²⁺ leak from the SR in the absence of an active Ca²⁺ uptake process. The rise of [Ca²⁺] after thapsigargin was approximately linear below 600 nM; above this value, the rate of leak decreased. On further addition of Ca²⁺, there was no evidence for SR Ca²⁺ uptake (results not shown). High concentrations of thapsigargin (20 µM) were used to ensure effective inhibition of Ca²⁺ pump activity. This thapsigargin concentration was equivalent to 20 nmol/mg total cell protein. Previous work (17) has shown complete pump inhibition at 110 nM/mg total protein. A higher concentration of thapsigargin (50 µM) produced no further increase in the rate of Ca²⁺ leak from the SR (not shown). In all cases, the rise of [Ca²⁺] after high concentrations of thapsigargin (>1 µM; n = 11) and cyclopiazonic acid (CPA) (>100 µM; n = 4) was initially linear up to 650-700 nM. At 600 nM, the [Ca²⁺] was still within the standard error of the best-fit linear relationship in all 11 experiments. This remained the case in 9 of the 11 experiments at 650 nM and in 5 of the 11 experiments at 700 nM. For this reason, 600 nM was taken as the maximum [Ca²⁺] for the subsequent pump leak modeling.

Pharmacological alterations of SR Ca²+ uptake and leak characteristics. Figure 2 shows the effects of partial inhibition of the pump on the calculated uptake and leak rate constants. Figure 2A shows two Ca²⁺ uptake curves superimposed, one before (control) and one after addition of a submaximal dose of thapsigargin (50 nM), a value that would be expected to approximately halve the rate of Ca²⁺ uptake by the SR Ca²⁺ pump (17). The rate of fall of [Ca²⁺] was slower, and the steady state [Ca²⁺] higher, in the presence of thapsigargin. Fitting this individual decay to Eq. 3 (Fig. 1C) indicates that the rate constant for Ca²⁺ uptake was approximately one-half of the control value. However, the calculated leak rate constant was not significantly different from control. Figure 2B shows curves and fitted model parameters to the decay of [Ca²⁺] before and after adding 1 µM CPA. As with thapsigargin, partial inhibition of the pump reduced the Ca²⁺ uptake rate constant by ~50% but caused no significant change in the leak rate. Figure 2C shows the effects of addition of the Ca²⁺ ionophore ionomycin (1 µM). The best-fit Ca²⁺ uptake rate constant was not altered, but the higher steady-state [Ca²⁺] was modeled by an increased leak rate. These results are consistent with the simplified model for Ca²⁺ uptake and release proposed in this study and indicate that SR Ca²⁺ uptake is normally independent of Ca²⁺ leak.

View larger version (24K): [in this window] [in a new window]   Fig. 2.   Effects of thapsigargin (A), cyclopiazonic acid (B), and ionomycin (C) on the time course of decay of [Ca2+] in oxalate-equilibrated permeabilized cardiac myocytes. In each panel, the broken lines indicate the best fit to Eq. 3 (Fig. 1C) with a starting [Ca2+] (Ca[0]) of 600 nM. The uptake rate (k) and leak (l) associated with each curve are shown.

Effects of P_i on Ca²+ uptake and leak rate constants. Figure 3A shows a continuous record of [Ca²⁺] from permeabilized cell aggregates during repetitive additions of Ca²⁺. The addition of 10 mM P_i to the cuvette solution caused a small increase in steady-state [Ca²⁺]. On subsequent addition of Ca²⁺, the rate of decay was slower, and the steady-state [Ca²⁺] was increased. Addition of 10 mM CrP caused a fall in the steady-state [Ca²⁺], and upon subsequent addition of Ca²⁺, the rate of Ca²⁺ uptake and the steady-state [Ca²⁺] were restored to values close to the control values. In Fig. 3B, left, the Ca²⁺ decay curves before and after addition of P_i are superimposed. After addition of P_i, a decrease in the Ca²⁺ uptake rate constant and an increase in the leak are required to provide an accurate fit to the Ca²⁺ decay. The sensitivity of the decay curve to the fitted parameters is shown in Fig. 3B, right. The trace is the recorded time course of Ca²⁺ uptake in 10 mM P_i. The curve lying mainly above the trace is the best fit generated with the Ca²⁺ uptake rate fixed at the control value (i.e., only the leak rate was allowed to vary). The curve lying mainly below the trace is the best fit generated with the leak rate fixed at the control value (i.e., only the Ca²⁺ uptake rate was allowed to vary). Comparison of SSQ from these curves with that from the more complex model in which both uptake rate and leak were allowed to vary (Fig. 3B, left) indicated that the more complex model fit the data significantly better (P < 0.001). This suggests that both increased leak and uptake rates are necessary to explain the effects of 10 mM P_i. Experiments were done to examine the effects of 1 mM P_i. No significant effect of this lower concentration of P_i on either uptake rate or leak was observed; the mean values of these measurements are shown below.

View larger version (32K): [in this window] [in a new window]   Fig. 3.   A: records of [Ca2+] against time recorded from a cuvette containing 1.5 ml 1 × 106/ml permeabilized myocytes suspended in mock intracellular solution (with 10 mM oxalate). Additions of aliquots of CaCl2 are indicated above the trace as increases in total [Ca2+] within the cuvette. Pi (10 mM) and creatine phosphate (CrP, 10 mM) were added at the points indicated. B, left, shows the superimposed records of the decay of [Ca2+] from 600 nM. The broken line through each Ca2+ decay indicates the best-fit single exponential curve; the associated parameters of uptake rate and leak are shown beside each curve. Sum-of-squares (SSQ) values for the fit were the following: control, 4.20 × 1015; 10 mM Pi, 5.12 × 1015. B, right, shows the decay of [Ca2+] recorded in the presence of 10 mM Pi. The two broken lines represent 1) a best-fit curve calculated by fixing the Ca2+ uptake rate at the control value (7.3 m/s) and allowing the leak rate to be adjusted (SSQ = 3.2 × 1014) and 2) a best-fit curve calculated by fixing the leak rate at the control value (1.24 nM/s) and allowing the Ca2+ uptake rate to be adjusted (SSQ = 5.6 × 1014). C: superimposed records of the decay of [Ca2+] from 600 nM for the control (), 10 mM Pi (), and 10 mM Pi/10 mM CrP (). The broken line is the best-fit curve to the latter curve; the uptake and leak rates are indicated.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Relative change in uptake rate (A) and leak rate (B) in 10 µM cyclopiazonic acid (CPA), 50 nM thapsigargin (Thaps), 10 µM ionomycin (Iono), and 1 mM and 10 mM Pi. Values are expressed as means ± SE. *Significant difference from control (P < 0.05; n = 6); ns, no significant difference.

P_i does not affect the rate of Ca²+ uptake in the presence of CrP. Figure 3, A and C, show that addition of CrP (10 mM) effectively reversed the effects of P_i on the SR. Another important aspect of this effect is shown in Fig. 5. Addition of 10 mM to the medium (in the absence of P_i) caused a significant increase in the rate constant and a decrease in steady-state [Ca²⁺], i.e., the effect of CrP was not dependent on the presence of 10 mM P_i. Of direct relevance to this study was the subsequent lack of effect of P_i on the rate of decay of [Ca²⁺] in the continued presence of CrP. This is highlighted in Fig. 5B, where the decay curves before and after P_i are superimposable. Thus P_i had no discernible effects on SR Ca²⁺ fluxes in the presence of CrP.

View larger version (25K): [in this window] [in a new window]   Fig. 5.   A: records of [Ca2+] against time recorded from a cuvette containing 1.5 ml 1 × 106/ml permeabilized myocytes suspended in mock intracellular solution (with 10 mM oxalate). Additions of aliquots of CaCl2 are indicated above the trace as increases in total [Ca2+] within the cuvette. CrP (10 mM) and Pi (10 mM) were added at the points indicated. B: the superimposed records of the decay of [Ca2+] from 600 nM.

Sensitivity of P_i-induced SR Ca²+ leak to thapsigargin. To test whether 10 mM P_i could increase the SR leak rate after Ca²⁺ pump inhibition, the effect of P_i was studied in the presence of a high concentration of thapsigargin. As shown in Fig. 6A, 10 mM P_i had no obvious effect on the background SR Ca²⁺ leak rate revealed by the addition of 20 µM thapsigargin. The best-fit straight lines through the record before and after addition of P_i have gradients that are not different. Addition of 1 µM ionomycin significantly increased the rate of loss of Ca²⁺ from the SR. On average, the background leak of Ca²⁺ in 10 mM P_i was 101 ± 1.4% (n = 4) of the control value in the presence of 20 µM thapsigargin. This suggests that 10 mM P_i induces a Ca²⁺ leak from the SR that is dependent on a functional Ca²⁺ pump. It follows that, in the continued presence of P_i, the SR pump serves as a route for Ca²⁺ leak and uptake; therefore, submaximal doses of thapsigargin should reduce leak and uptake rates (assuming the drug has equal ability to inhibit the SR Ca²⁺ pump in leak and uptake mode). To test this hypothesis, Ca²⁺ uptake and leak were assessed before and after partial pump inhibition by thapsigargin (50 nM) in the continued presence of 10 mM P_i (Fig. 6B). Under these circumstances, 50 nM thapsigargin slowed the rate of Ca²⁺ uptake by the SR, but unlike in Fig. 2A, there was only a limited increase in steady-state [Ca²⁺]. According to the model, 50 nM thapsigargin significantly decreased the rate of Ca²⁺ uptake and decreased the leak rate. On average, 50 nM thapsigargin did not change the steady-state [Ca²⁺] (103 ± 3%; n = 6), but it reduced the Ca²⁺ uptake rate constant to 68 ± 4% of the control value and leak rate to 72 ± 3% (n = 6) in the presence of 10 mM P_i.

View larger version (18K): [in this window] [in a new window]   Fig. 6.   A: records of [Ca2+] against time recorded from a cuvette containing 1.5 ml 1 × 106/ml permeabilized myocytes suspended in mock intracellular solution (with 10 mM oxalate). Bars indicate when 20 µM thapsigargin, 10 mM Pi, and 1 µM ionomycin were added at the points indicated. The broken lines are the best-fit linear regression lines to the last 120 s before and the first 120 s after addition of Pi. B: top, records of [Ca2+] against time recorded from a cuvette containing 1.5 ml 1 × 106/ml permeabilized myocytes suspended in mock intracellular solution (with 10 mM oxalate) in the continuous presence of 10 mM Pi. Additions of aliquots of CaCl2 are indicated above the trace as increases in total [Ca2+] within the cuvette. Thapsigargin (50 nM and 20 µM) was added at the points indicated. Bottom, the superimposed records of the decay of [Ca2+] from 600 nM in 5 mM ATP/10 mM Pi and after addition of 50 nM thapsigargin. The broken lines are the best-fit curves to the individual decays with the associated uptake and leak rates.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Quantification of Ca²+ uptake and leak pathways in permeabilized cardiac myocytes. The experiments described in this study were designed to investigate the direct effects of P_i on SR Ca²⁺ pump function. For this reason, the experimental conditions were arranged to minimize any indirect effects that would complicate analysis. Modulation of SR pump activity via low-affinity Ca²⁺ binding sites within the SR has been previously demonstrated in SR vesicles from cardiac and skeletal muscle (18). P_i is known to enter the SR and, under some conditions, chelate and/or precipitate luminal Ca²⁺, and in doing so, it may indirectly affect Ca²⁺ pump activity. To prevent this, these experiments were done in the presence of 10 mM oxalate. Previous studies (e.g., Ref. 5) have shown that in the presence of both oxalate and P_i, only Ca-oxalate precipitates are formed when the ratio of [oxalate] to [P_i] is <2. Ruthenium red (5 µM) ensures that Ca²⁺ leak via the ryanodine receptor was inhibited (11, 28, 29). Under these conditions, the steady-state [Ca²⁺] in this system (1.0 × 10⁶ cells/ml) was normally between 100 and 150 nM, a value that represents a balance between a background Ca²⁺ leak from the SR and a maintained Ca²⁺ uptake. Addition of aliquots of Ca²⁺ caused an initial increase in [Ca²⁺], which decreased over the subsequent 5-10 min to the same steady state. The time course of the decrease in [Ca²⁺] from 600 nM to the steady state followed a single exponential decay, suggesting that over this range of [Ca²⁺] (between 600 and 100 nM), the rate of Ca²⁺ uptake is a function of [Ca²⁺]. This is a simplification of the Ca²⁺ dependence of SR Ca²⁺ uptake anticipated on the basis of the known stoichiometry and affinity of the pump for Ca²⁺ in permeabilized rabbit myocytes [2 Ca²⁺ with an affinity of ~0.3 µM (17, 21)]. The alteration in time course of Ca²⁺ uptake by CrP suggests that this metabolite may alter the affinity and/or the stoichiometry of the SR Ca²⁺ pump and is discussed in more detail below.

Limitations of the method of analysis of SR Ca²+ flux. The model of the SR, expressed as equations in Fig. 1C, suggests that knowledge of the rate constant for Ca²⁺ uptake and the steady-state [Ca²⁺] is sufficient to estimate the Ca²⁺ leak rate. It should be emphasized that this method of analysis is valid only for the experimental conditions used in this study. The range of [Ca²⁺] studied must be in the range (between 200 and 600 nM) in which SR Ca²⁺ uptake is linear for forward transport via the SR Ca²⁺ pump. Furthermore, the [Ca²⁺] within the SR lumen must be constant; this ensures that the activity of the SR Ca²⁺ pump is not modulated by changes in [Ca²⁺] within the SR lumen. In addition to this, a constant SR [Ca²⁺] will ensure a constant passive leak, thus simplifying this aspect of the analysis. One further simplification inherent in this model is the neglect of Ca²⁺ buffering by the cells and constituents of the bathing solution. The Ca²⁺ binding properties of permeabilized cardiac myocytes have been studied in detail (16). These measurements suggest that the presence of myocytes (10⁵/ml) would bind ~0.1 µM Ca²⁺ at 600 nM [Ca²⁺]. EGTA (50 µM) and fura 2 (10 µM) provide additional buffering with minor contributions from ATP and oxalate. Under the conditions of this study, these buffers would provide an approximately constant Ca²⁺ buffer within the range between 100 and 600 nM. In terms of the calculation of Ca²⁺ fluxes, buffering would have an identical scaling effect on both uptake and leak pathways and can therefore be ignored for the current purposes.

Pharmacological alterations of SR Ca²+ uptake and leak characteristics. The validity of this model was tested by selective inhibition of the Ca²⁺ pump with CPA and thaspigargin. Both agents are specific inhibitors of sarco(endo)plasmic reticulum Ca²⁺ pumps in muscle and nonmuscle cells (4, 34). At submaximal concentrations, these agents reduced the rate constant of Ca²⁺ uptake by the SR (17). Using the uptake-leak analysis shown in Fig. 1C, the calculated leak after partial inhibition of the pump was not significantly different from control values. This suggests that, in the absence of P_i, background Ca²⁺ leak was unaffected by an ~50% reduction of Ca²⁺ pump activity. Increased Ca²⁺ leak by addition of ionomycin increased the steady-state [Ca²⁺] but did not affect the rate constant for decay. This further validates the simple model of the SR described quantitatively in Fig. 1C and indicates that changes in SR Ca²⁺ efflux that are independent of the Ca²⁺ pump do not affect the SR Ca²⁺ uptake rate constant. Moreover, the results from Figs. 1 and 2 indicate that, under control conditions, Ca²⁺ leak from the SR is independent of SR Ca²⁺ pump activity, consistent with the absence of significant pump-mediated Ca²⁺ leak in the absence of P_i.

Effects of P_i on Ca²+ uptake and leak rate constants. Previous studies using permeabilized cardiac muscle have measured Ca²⁺ release on application of caffeine (37-39, 42). Results of these experiments are difficult to interpret in terms of actions on the SR, because chelation and/or precipitation of calcium phosphate within the SR and effects of P_i on the ryanodine receptor cannot be excluded. However, these studies clearly showed that 10 mM P_i reduced the Ca²⁺ released by caffeine to ~60% of the control value. This action was accompanied by a transient release of Ca²⁺ from the SR (37, 38). The transient release was not blocked by ryanodine (39), suggesting that a release pathway other than the active ryanodine receptor was involved. However, a link between a functional Ca²⁺ pump and P_i-induced Ca²⁺ release was not established.

The effect of CrP on SR Ca²+ uptake. The effect of P_i could be reversed by the addition of 10 mM CrP (Fig. 3C). Furthermore, in the continued presence of CrP, 10 mM P_i had no significant effects on either flux process (Fig. 4). A similar dependence on CrP was described previously (38). This effect was attributed to the ability of CrP to rephosphorylate ADP via the enzyme creatine phosphokinase (CK), because inhibition of CK abolished the effect of CrP (38). There is good evidence that an SR-bound CK exists spatially close to the Ca²⁺ pump ATPase and preferentially rephosphorylates the ADP produced by the Ca²⁺ pump (32). These results suggest that the ability of P_i to affect the SR Ca²⁺ pumps requires high concentrations of ADP. The results from this study also support the view that CrP can significantly increase the Ca²⁺ uptake rate of the SR and may have significant effects on the Ca²⁺ affinity and/or stoichiometry of the pump (6, 24). In this study, background SR Ca²⁺ leak (in the absence of P_i) was unaffected by CrP, suggesting that, under control conditions, there was no pump-mediated leak from the SR.

Potential mechanism(s) of action of P_i on the SR Ca²+ pump. There are two mechanisms by which P_i may induce a Ca²⁺ pump-mediated SR Ca²⁺ leak, "pump reversal" and "pump-channel transition." It has been shown that, in SR vesicle preparations, millimolar levels of P_i can phosphorylate the SR Ca²⁺ pump directly (31). In the presence of ADP, this form of the pump can generate a Ca²⁺ efflux from the SR accompanied by the synthesis of ATP, i.e., pump reversal (15). In the present study, in the absence of CrP, ADP concentration within the cuvette would be expected to continuously increase during the experiment because of the cellular ATPases associated with permeabilized myocytes. Separate measurements of the cellular ATPase rate suggest that 1 × 10⁶ cells/ml lower the ATP concentration at ~10 nM/s (A. Pagliocca and G. L. Smith, unpublished observation). This gives a predicted concentration range of ADP from ~60 to 120 µM at the time of P_i addition. These values are similar to the intracellular values measured during hypoxia and ischemia in cardiac muscle (1, 33) and are close to optimal for pump reversal (3). However, it is unclear from previous studies whether pump reversal occurs to a significant extent in the presence of millimolar concentrations of cytosolic ATP. Cytosolic ATP can also phosphorylate the SR Ca²⁺ pump as part of the normal sequence of events in Ca²⁺ transport. Pump reversal in the presence of ATP may be possible if the concentration of the phosphorylated intermediate of the pump increases above the level normally achieved by ATP. The ability of P_i to induce SR Ca²⁺ leak and decrease the uptake rate constant of the pump may be linked. While operating in the reverse mode, the pump would be unavailable for Ca²⁺ uptake, thereby reducing the number of active Ca²⁺ pumps. This could account for the decrease in the uptake rate constant. As mentioned earlier, the absence of CrP and increased levels of ADP may also alter the stoichiometry and Ca²⁺ affinity of the fraction of pumps operating in the normal mode (18). This effect could not be distinguished from a reduced maximal rate of pumping in these studies.