INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

UNTIL RECENTLY, mitochondrial Ca²⁺ uptake was thought to occur only when cytosolic Ca²⁺ concentration ([Ca²⁺]_i) became abnormally high. However, this traditional view is now being questioned. Several reports (7, 16, 34) provided strong evidence that mitochodria are an integral part of Ca²⁺ regulatory mechanisms in a variety of cell types. The key to the reevaluation of the role of mitochondria was the development of the two mitochondria-specific Ca²⁺-sensitive indicators recombinant aequorin and rhod 2. Rizzuto et al. (36) genetically engineered mitochondria-targeted aequorin and reported that Ca²⁺ in mitochondria is elevated during the generation of inositol 1,4,5-trisphosphate, a ubiquitous second messenger that releases Ca²⁺ from intracellular stores (35). Similarly, rhod 2, a positively charged Ca²⁺-sensitive dye that tends to accumulate into mitochondria, permitted direct detection of mitochondrial Ca²⁺ concentration changes in a wide range of cells (2, 6, 17, 31). These studies often suggested that mitochondria do not simply lower [Ca²⁺]_i but modulate other Ca²⁺ regulatory mechanisms (7, 16, 34). In particular, one recurring theme is that mitochondria seem strategically positioned near the origin of Ca²⁺ sources and thus can sense localized Ca²⁺ changes. In turn, Ca²⁺ removal to neighboring mitochondria may modulate the behavior of Ca²⁺ permeable channels through feedback mechanisms (22). The hypothesis that mitochondria are positioned close to Ca²⁺ sources is also useful in explaining why mitochondria, thought to have low affinity for Ca²⁺, can sequester Ca²⁺ in the absence of very high [Ca²⁺]_i. The complexity of the role of mitochondria, however, may mean that results obtained from one cell type may not be applicable to others where intracellular architecture or [Ca²⁺]_i raising mechanisms are different. To date, relatively few studies (6, 31) have addressed the role of mitochondrial Ca²⁺ uptake by using arterial smooth muscle cells. Furthermore, increases in [Ca²⁺]_i required to induce maximum contraction in arteries may be rather modest (29); yet to our knowledge, few studies so far examined Ca²⁺ uptake by mitochondria at a lower range of [Ca²⁺]_i. Monteith and Blaustein (31) reported that a moderate increase in [Ca²⁺]_i does not produce marked Ca²⁺ changes in the majority of mitochondria with the use of cultured rat aortic smooth muscle cells. Also, using rat pulmonary artery smooth muscle cells, Drummond and Tuft (6) reported elevation in mitochondrial Ca²⁺ when [Ca²⁺]_i was raised to ~1 µM, a concentration perhaps rarely experienced under normal conditions.

The current study examines the role of mitochondria in Ca²⁺ handling in arterial smooth muscle cells. Ca²⁺ removal was studied using voltage-clamped single arterial smooth muscle cells dialyzed with the Ca²⁺-sensitive dye fura 2. Ca²⁺ transients were evoked by Ca²⁺ influx or Ca²⁺ release, and the effect of agents thought to prevent mitochondrial Ca²⁺ uptake was studied. Also, changes in mitochondria membrane potential were detected using rhodamine 123, and effects of drugs thought to modify mitochondrial potential were investigated. Our results suggest that mitochondrial Ca²⁺ uptake is activated during Ca²⁺ influx and Ca²⁺ release. Ca²⁺ sequestration to mitochondria seems to occur even when the elevation in [Ca²⁺]_i is modest. In the presence of putative inhibitors of mitochondrial Ca²⁺ uptake, the Ca²⁺ removal rate over the lower range of [Ca²⁺]_i was reduced. Moreover, the results suggest that Ca²⁺ content in mitochondria may influence Ca²⁺ load in the sarcoplasmic reticulum. Part of this study has been presented as an abstract (26).

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell preparation. Male Sprague-Dawley rats (200-300 g) were made unconscious by exposure to an increasing concentration of CO₂, and then euthanized by exsanguination in accordance with Schedule 1 of the Animals (Scientific Procedures) Act, 1986. The femoral arteries were dissected in a salt solution containing (in mM) 137 NaCl, 0.44 NaH₂PO₄, 0.42 Na₂HPO₄, 4.17 NaHCO₃, 5.6 KCl, 1 MgCl₂, 2 CaCl₂, 10 HEPES, and 10 glucose (pH adjusted to 7.4 with NaOH) (23). Single smooth muscle cells were enzymatically dissociated in a low-Ca²⁺ salt solution containing (in mM) 80 Na glutamate, 55 NaCl, 6 KCl, 2 MgCl₂, 0.2 CaCl₂, 10 HEPES, 10 glucose, and 0.2 EDTA (pH was adjusted to 7.3 at room temperature so that at 35°C, it will be 7.4) (23). At 35°C, the femoral artery was first digested for 30 min with 1.7 mg/ml papain and 0.7 mg/ml dithioerythritol and then for 20 min with 1.7 mg/ml collagenase (type F) and 1 mg/ml hyaluronidase (type I-S). Single smooth muscle cells were obtained by triturating the artery with a fire-polished Pasteur pipette. The cell suspension was kept in a refrigerator at 2-8°C, and all experiments were carried out on the same day.

Electrophysiology and microfluorimetry. Voltage clamp was achieved using a conventional whole cell clamp technique (18). Under control conditions, the extracellular solution was composed of (in mM) 80 Na glutamate, 40 NaCl, 20 tetraethylammonium Cl, 1.1 MgCl₂, 3 CaCl₂, 10 HEPES, and 30 glucose (pH 7.4 with NaOH). The intracellular solution consisted of (in mM) 145 CsCl, 3 MgCl₂, 3 Na₂ATP, 10 HEPES, and 0.05 fura 2 pentapotassium (pH 7.2 with CsOH). Membrane currents were amplified using Axopatch 200B (Axon Instruments), filtered at 1 kHz, and sampled at 5 kHz using pCLAMP version 7 software (Axon Instruments). Stock solutions of all test drugs except caffeine were made with DMSO at 1,000-fold of final concentrations. All agents except caffeine were applied by perfusing the experimental chamber with appropriate bath solutions. An elevation in [Ca²⁺]_i was induced by Ca²⁺ influx through voltage-dependent Ca²⁺ channels or by Ca²⁺ release from the sarcoplasmic reticulum. Ca²⁺ channels were activated either by 1.8-s depolarization to 0 mV from a holding potential of 70 mV or by sustained depolarization to 30 mV. Ca²⁺ release was triggered by rapid application of 20 mM caffeine using a U tube superfusion system (9). In this case, a cell was voltage clamped at 70 mV, and the U tube with a small hole in its apex was placed in close proximity. The bath solution was continuously perfused while caffeine-containing bath solution was run through the U tube. Outflow of the U tube was made larger than inflow. In this arrangement, caffeine-containing solution does not leak out from the hole. Outflow was fed through a solenoid valve before being finally discarded. When the valve is closed, the caffeine-containing solution is promptly ejected from the hole in the U tube, producing a caffeine concentration jump. The cell remains exposed to caffeine as long as the valve is closed. When the valve is reopened, the caffeine-containing solution resumes to run to waste, whereas caffeine in the experimental chamber is washed out. The valve was controlled by pCLAMP command. [Ca²⁺]_i was measured with deltaRAM (Photon Technology International), as previously described (23). A single cell was alternately illuminated with UV light of 340 and 380 nm (bandpass 8 nm) at 100 Hz. Emission signals were detected at 510 nm (bandpass 40 nm). The dissociation constant (K_d) for fura 2 was calculated as 180 nM from in vitro calibration using intracellular solutions containing a range of known free Ca²⁺. The desired free Ca²⁺ was produced by adding required amount of CaCl₂, calculated with software (Buffer version 1.0; Dr. R. Schubert, Universität Rostock, Germany), to 10 mM EGTA-buffered pipette solution. The fluorescence ratios in the absence of added CaCl₂ (R_min) and in the presence of saturating Ca²⁺ (R_max) were determined from in vitro measurements. Conversion of ratios (R) to [Ca²⁺]_i was carried out using the equation K_d(F₀/F_s)(R  R_min)/(R_max  R), where F₀ is the count at 380 nm with R_min solution and F_s is that with R_max solution (15, 33). R_min and R_max were decreased by 15% to adjust for viscosity (33). Background signal was measured for each cell after formation of a gigaohm seal and before formation of whole cell clamp mode. To measure mitochondrial membrane potential, cells were loaded with rhodamine 123 (10 µg/ml) for 10-15 min at room temperature (8). Cells were rinsed with dye-free solution consisting of (in mM) 80 Na glutamate, 55 NaCl, 6 KCl, 2 MgCl₂, 0.2 CaCl₂, 10 HEPES, and 10 glucose (pH 7.4 with NaOH). Cells were not voltage clamped in these experiments. The fluorophore was excited at 500 nm (25 nm bandpass), and emission signal was determined at 545 nm (35 nm band pass) with a frequency of 10 Hz. Rhodamine 123 signals were expressed as the relative fluorescence intensity against the photon counts at the beginning of experiments. Data were expressed as means ± SE of n cells. When appropriate, Student's paired t-test or one-way ANOVA, followed by Tukey's test as a post hoc analysis, was used to evaluate significant difference. All experiments were carried out at room temperature.

Drugs. Fura 2 pentapotassium salt and rhodamine 123 were purchased from Molecular Probes. Cyclopiazonic acid (CPA) was obtained from Calbiochem-Novabiochem. Diazoxide was purchased from Research Biochemicals International. Papain was obtained from Worthington Biochemical. All other agents were from Sigma or BDH.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Ca²+ removal profile of rat femoral artery smooth muscle cells. To investigate Ca²⁺ regulatory mechanisms, voltage-dependent Ca²⁺ channels were activated by depolarization to 0 mV from a holding potential of 70 mV (Fig. 1A, third trace). This maneuver evoked Ca²⁺ current (Fig. 1A, second trace) and elevated [Ca²⁺]_i (Fig. 1A, first trace). On repolarization to 70 mV, increased [Ca²⁺]_i returned to the resting level, allowing the examination of Ca²⁺ clearance (23, 25). A high-order polynomial equation was fitted to the declining phase of the Ca²⁺ transient, and Ca²⁺ removal rate was calculated as the negative derivative of the fit. The result is expressed as either a function of the measured [Ca²⁺]_i (Fig. 1A, fourth trace, left) or time where time 0 is the instance of repolarization (Fig. 1A, fourth trace, right). These traces (Fig. 1A, fourth trace, left and right) exemplify a complex waveform obtained in about one-half of the control cells. In these cells, the Ca²⁺ removal was initially fast over higher [Ca²⁺]_i but decreased as [Ca²⁺]_i declined. Just before elevated [Ca²⁺]_i returned to the resting level, however, an increase in Ca²⁺ removal occurred, appearing as an upward hump in the removal profile. This shape of Ca²⁺ clearance may be termed as "peak-trough-peak," or phase 1, phase 2, and phase 3 (23, 25, 30). Phase 3, a delayed increase in Ca²⁺ removal rate, was absent (Fig. 1B, fourth trace, left and right) in about one-half of the control cells. In the control cells that lack phase 3, raised [Ca²⁺]_i seems to recover quickly (Fig. 1B, first trace). These two types of Ca²⁺ removal pattern will be further discussed later.

View larger version (28K): [in this window] [in a new window]   Fig. 1.   A: elevation in intracellular Ca2+ concentration ([Ca2+]i) (first trace) was triggered by depolarization to 0 mV from a holding membrane potential (VM) of 70 mV (third trace) due to Ca2+ influx through voltage-dependent Ca2+ channels (second trace). Ca2+ removal rate was calculated as the negative derivative of the polynomial fit to the declining phase of the Ca2+ transient (d[Ca2+]i/dt). Ca2+ removal rate is shown as either a function of the measured [Ca2+]i (fourth trace, left) or time where time 0 is the instance of repolarization (fourth trace, right). Note an upward hump (phase 3), the delayed enhancement in Ca2+ removal. B: Ca2+ transient of a control cell where raised [Ca2+]i returned to the basal level quickly (first trace). As in A, the second and the third traces show the Ca2+ current (I) and the voltage protocol, respectively. In this example, phase 3 is largely diminished (fourth trace, right and left).

Mitochondrial Ca²+ sequestration examined by diazoxide application. The importance of mitochondria in Ca²⁺ regulation has now been acknowledged in many cell types (7, 16, 34), but their role in arterial smooth muscle remains unclear. First, we sought to establish that Ca²⁺ is indeed sequestered in mitochondria. Because Ca²⁺ is sequestered in mitochondria by virtue of negative potential in mitochondrial inner membrane (7, 16, 34), the agents that depolarize mitochondria should discharge stored Ca²⁺. We used 100 µM diazoxide to depolarize mitochondria. Although its mechanism has yet to be fully established, there is evidence showing that, in cardiac myocytes, diazoxide depolarizes mitochondria and raises [Ca²⁺]_i (see DISCUSSION). When diazoxide was applied to the cells voltage clamped at 70 mV, elevation in [Ca²⁺]_i was not always observed. This may not be surprising because without prior stimulation, Ca²⁺ in mitochondria is probably not much higher than [Ca²⁺]_i (31). Therefore, the cells were subjected to a brief depolarization to evoke a Ca²⁺ transient before the application of diazoxide. The Ca²⁺ efflux from the mitochondria may be a slow process in smooth muscle cells (6), and thus Ca²⁺ in mitochondria may remain elevated even when the global [Ca²⁺]_i is already returned to the resting level (6). Figure 2 shows one such experiment. [Ca²⁺]_i was first raised by depolarization (Fig. 2, left). Subsequent diazoxide application elevated [Ca²⁺]_i, whereas the membrane potential was maintained at 70 mV (Fig. 2, right). The time gap between the left trace and the right trace in Fig. 2 was ~30 s. Diazoxide-induced elevation in [Ca²⁺]_i was observed in all cells subjected to prior voltage pulse, with an average [Ca²⁺]_i increase of 90 ± 20 nM (n = 7). The results suggest that mitochondria can in fact accumulate Ca²⁺ during Ca²⁺ influx and their Ca²⁺ remains sufficiently high even after global [Ca²⁺]_i returned to the resting level.

View larger version (11K): [in this window] [in a new window]   Fig. 2.   Application of 100 µM diazoxide triggered a Ca2+ transient. The cell was subjected to a depolarizing pulse (bottom left) to elevate [Ca2+]i (top left). After [Ca2+]i returned to resting level, diazoxide was applied (downward arrow), triggering an increase in [Ca2+]i (top right), whereas VM was maintained at 70 mV (bottom right). The time gap between the left trace and the right trace was ~30 s.

Ca²+ removal profile determined in the presence of mitochondrial depolarizing agents. When depolarized, mitochondria no longer sequester Ca²⁺ due to the diminished driving force on Ca²⁺. Thus the Ca²⁺ removal profile was examined in the presence of 100 µM diazoxide. When [Ca²⁺]_i was raised by depolarization (Fig. 3A, third trace), it returned to the resting level slowly (Fig. 3A, first trace) without the upward hump in the Ca²⁺ removal profile (Fig. 3A, fourth trace). Similar results were obtained from four other cells. The inhibition of phase 3 in the diazoxide-treated cells is somewhat different from that in the control cells without phase 3. As shown in toad stomach cells (30), phase 3 occurs in femoral artery smooth muscle cells only when [Ca²⁺]_i is elevated for long enough (23). After depolarization, the average time required to restore resting [Ca²⁺]_i in the control cells with phase 3 was 49.9 ± 7.7 s (n = 7), whereas that of the control cells without phase 3 was 18.7 ± 2.7 s (n = 6). On the other hand, the average time required to restore basal [Ca²⁺]_i in diazoxide-treated cells was 54.8 ± 10.2 s (n = 5). This seems long enough to evoke the delayed acceleration in Ca²⁺ removal, yet phase 3 was largely diminished.

View larger version (26K): [in this window] [in a new window]   Fig. 3.   A: Ca2+ removal profile of a cell treated with 100 µM diazoxide. A depolarization to 0 mV from a holding potential of 70 mV (third trace) triggered Ca2+ current (second trace) and concomitant Ca2+ transient (first trace). Elevated [Ca2+]i returned to the basal level slowly. Also, phase 3, an upward hump in the Ca2+ removal profile, is largely diminished (fourth trace, right and left). B: Ca2+ removal profile of a cell treated with 1 µM carbonyl cyanide m-chlorophenylhydrazone (CCCP). Similar results as in A are shown, with little delayed facilitation in Ca2+ removal rate (fourth trace, right and left).

Effect of mitochondrial depolarizing agent on steady-state [Ca²+]_i. The inhibitory effect of diazoxide and CCCP on phase 3 over lower [Ca²⁺]_i was surprising because mitochondria are generally thought to sequester Ca²⁺ when [Ca²⁺]_i is high (31). Because the Ca²⁺ removal rates were examined after a large depolarizing pulse, it is possible that the putative mitochondrial role in Ca²⁺ clearance may be exaggerated. Unlike neurons or cardiac myocytes, many smooth muscle cells, including those from femoral arteries, do not fire action potential. Rather, stimulation produces smaller and sustained depolarization. During continued depolarization, a new level of [Ca²⁺]_i is achieved, reflecting an altered balance between Ca²⁺ influx and Ca²⁺ removal. Previous studies (10, 24) showed that [Ca²⁺]_i increased by ~100 nM during sustained depolarization. If mitochondrial Ca²⁺ uptake is important only when [Ca²⁺]_i is raised sufficiently high, inhibition of mitochondrial Ca²⁺ removal would have little impact on steady-state [Ca²⁺]_i. On the other hand, if mitochondrial Ca²⁺ clearance is also important over lower [Ca²⁺]_i, steady-state [Ca²⁺]_i may be affected by mitochondrial Ca²⁺ uptake inhibitors.

View larger version (22K): [in this window] [in a new window]   Fig. 4.   The effect of CCCP on steady-state [Ca2+]i. VM was increased from 70 to 50 to 30 mV (bottom trace), with concomitant increase in [Ca2+]i (top trace). When 1 µM CCCP was applied (downward arrow), [Ca2+]i increased. On washout of CCCP (upward arrow), partial recovery in elevated [Ca2+]i was observed.

Effect of mitochondrial depolarizing agent on caffeine-induced Ca²+ transient. When mitochondria act as a Ca²⁺ buffer, their location within the cell may be important. For example, Park et al. (32) have reported that the origin of Ca²⁺ influences which mitochondria, perinuclear, perigranular, or subplasmalemmal, will preferentially remove Ca²⁺ in pancreatic acinar cells. In arterial smooth muscle cells, elevation in [Ca²⁺]_i may occur not only due to Ca²⁺ influx but also Ca²⁺ release from the sarcoplasmic reticulum. The results described so far show mitochondrial Ca²⁺ clearance after Ca²⁺ influx. This does not mean that mitochondrial Ca²⁺ uptake is also important when Ca²⁺ is released. Therefore, we examined the effect of CCCP on caffeine-induced Ca²⁺ transients. Caffeine is one of the standard agonists for sarcoplasmic reticulum Ca²⁺ release because it activates ryanodine receptors. While cells were being voltage clamped at 70 mV, 20 mM caffeine was applied for 1 s with the U tube superfusion system, as described in MATERIALS AND METHODS. This arrangement permitted reproducible caffeine-induced Ca²⁺ transients. Figure 5A shows one such experiment under control conditions. The solid squares below the [Ca²⁺]_i trace indicate application of 20 mM caffeine. The duration of caffeine transients was somewhat variable among cells, probably reflecting variable efficiency of the drug washout. Nonetheless, Ca²⁺ transients obtained from the same cell were reproducible. For analysis purposes (see below), we discarded the first transient and used second, third, and sometimes fourth Ca²⁺ transients because in some cells the first Ca²⁺ transient was somewhat different from the rest.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Effect of CCCP on Ca2+ transients induced by caffeine applications. A: caffeine applications evoked Ca2+ transients in the control cell. Caffeine (20 mM) was applied for 1 s at the time indicated by solid squares, producing multiple Ca2+ transients due to Ca2+ release. B: caffeine-induced Ca2+ transients before and after application of 1 µM CCCP. Ca2+ transients were triggered twice by caffeine application. Application of CCCP (downward arrow) evoked elevation in [Ca2+]i. Subsequent caffeine application induced prolonged Ca2+ transient. On the washout of CCCP (upward arrow), a decrease in resting [Ca2+]i was observed. When caffeine application was repeated, raised [Ca2+]i returned to the resting level more quickly than when CCCP was present.

Modulation of mitochondrial membrane potential. The results described above attempted to dissect mitochondrial contribution to Ca²⁺ homeostasis with the use of pharmacological interventions. CCCP and diazoxide were used to inhibit mitochondrial Ca²⁺ uptake. It was assumed that both agents prevent Ca²⁺ accumulation in the mitochondria by depolarizing the mitochondrial membrane. While this assumption is likely to be true for CCCP, a consensus regarding diazoxide actions is yet to be achieved (see DISCUSSION). Therefore, we sought evidence that both CCCP and diazoxide depolarized the mitochondrial membrane. Cells were loaded with rhodamine 123, a positively charged dye with a propensity to accumulate into mitochondria due to their negative membrane potential (8). When concentrated above threshold, quenching of the indicator occurs. On mitochondrial depolarization, dye moves out from the organelles, relieving quench, and thus increasing fluorescent signal. On the other hand, mitochondrial membrane hyperpolarization will decrease fluorescence intensity due to further concentration and thus quenching of the fluorophore. In our experiments, the fluorescent signals declined continuously, presumably due to dye bleaching and leaking. Nonetheless, changes in the fluorescence intensity were detected on application of agents. As shown in Fig. 6A, an application of CCCP (1 µM) produced an increase in fluorescent signal, suggesting depolarization of mitochondria and subsequent dye redistribution into the cytosol. This CCCP effect was reversible. Diazoxide (100 µM), on the other hand, produced a small increase in signal (Fig. 6B). This may not be surprising because the half-maximum dose of diazoxide is 65-96 µM in cardiac mitochondria (19), and thus 100 µM diazoxide is likely to cause a partial depolarization. On the other hand, application of CCCP is likely to cause near-complete depolarization of mitochondria. When a higher concentration of diazoxide (500 µM) was tested, a larger increase in fluorescence intensity was noted (Fig. 6C). We also examined the effect of oligomycin on rhodamine 123 fluorescence intensity. Unlike CCCP or diazoxide, oligomycin is expected to cause a small hyperpolarization of mitochondria (8) (see DISCUSSION). Indeed, application of 3 µM oligomycin produced a small decrease in fluorescence intensity, suggesting mitochondrial hyperpolarization (Fig. 6D). The effect of oligomycin was not reversible (8). These observations indicate that mitochondrial membrane potential is depolarized by CCCP and diazoxide and hyperpolarized by oligomycin.

View larger version (31K): [in this window] [in a new window]   Fig. 6.   Changes in mitochondria VM detected with rhodamine 123. Rhodamine 123 signals were expressed as relative fluorescence intensity against the photon counts at the beginning of experiments. Because the fluorescence signals continuously declined under our experimental condition, dotted lines were added to show the trend before the application of agents. A: application of 1 µM CCCP increased fluorescence intensity, suggesting mitochondrial membrane depolarization. B: application of 100 µM diazoxide caused a small increase in fluorescence intensity. C: a larger increase in signal was detected with an application of 500 µM diazoxide. D: application of 3 µM oligomycin caused a small decrease in fluorescence intensity, suggesting mitochondrial hyperpolarization.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The main finding of the current study is that mitochondrial Ca²⁺ removal occurs over a [Ca²⁺]_i range usually thought to be too low for its action. When mitochondrial Ca²⁺ uptake was blocked by diazoxide or CCCP, phase 3 of Ca²⁺ removal corresponding to a lower range of [Ca²⁺]_i was inhibited. Also, during sustained depolarization to 30 mV where [Ca²⁺]_i increased modestly, application of CCCP raised [Ca²⁺]_i. These findings are surprising because mitochondria are thought to sequester Ca²⁺ when [Ca²⁺]_i is high (31). The [Ca²⁺]_i corresponding to phase 3 is low, and the primary candidate for Ca²⁺ removal for this range of [Ca²⁺]_i would be ATP-fueled Ca²⁺ pumps. In fact, it was reported in a previous study (23) on femoral artery smooth muscle cells that application of CPA, an inhibitor of the sarcoplasmic reticulum Ca²⁺ pump, diminished phase 3. Likewise, application of thapsigargin or ryanodine, both inhibitors of sarcoplasmic reticulum Ca²⁺ uptake, inhibited phase 3 in rat cerebral artery smooth muscle cells (25). This raises the question as to why mitochondrial Ca²⁺ uptake appears to operate over a [Ca²⁺]_i range, where excess Ca²⁺ is thought to be removed by the sarcoplasmic reticulum. One possibility is that diazoxide and CCCP reduced phase 3 by triggering Ca²⁺-induced Ca²⁺ release (12). Ca²⁺ discharge from mitochondria may activate ryanodine receptors in the sarcoplasmic reticulum. This would render the sarcoplasmic reticulum "leaky," effectively blocking Ca²⁺ sequestration. However, this seems unlikely. Under this scenario, the inhibition of sarcoplasmic reticulum Ca²⁺ uptake would be transient. In the continuous presence of diazoxide or CCCP, mitochondria are unable to sequester Ca²⁺. Therefore, when Ca²⁺ discharge from mitochondria is completed, the sarcoplasmic reticulum should resume Ca²⁺ uptake. This was not the case. Another possibility is that mitochondria depolarization deprived the Ca²⁺ pump in the sarcoplasmic reticulum of ATP. The depolarized mitochondria may rather consume than produce ATP in an attempt to restore membrane potential (5). Furthermore, even if more than enough exogeneous ATP were available in the pipette solution, sarcoplasmic reticulum Ca²⁺ pumps may require mitochondria-generated "compartmentalized" ATP (37). To test this hypothesis, the effect of the ATP synthase inhibitor oligomycin (3 µM) was examined. Oligomycin blocks mitocondrial ATP production without depolarizing mitochondria, thus keeping the mitochondrial ability to sequester Ca²⁺ intact (5). Thus if oligomycin also blocks phase 3, this would suggest that diazoxide and CCCP indirectly inhibited the Ca²⁺ pump in the sarcoplasmic reticulum. On the other hand, if Ca²⁺ removal in the presence of oligomycin is more similar to that of the control, the effect of diazoxide and CCCP is not due to ATP depletion. When Ca²⁺ removal was examined in the presence of oligomycin, phase 3 was not inhibited. Rather, Ca²⁺ removal rate was faster in the presence of oligomycin at low [Ca²⁺]_i. Figure 7 summarizes Ca²⁺ removal rates determined in the presence of CCCP, diazoxide, CPA, oligomycin, or no agent (control) at a [Ca²⁺]_i of 75 nM. A significant difference (P < 0.01) was detected between oligomycin data and CCCP, diazoxide, or CPA data using one-way ANOVA, followed by Tukey's test for post hoc analysis. Interestingly, a statistically significant difference was not detected between the control cells and CCCP-, diazoxide-, or CPA-treated cells. This is presumably because about one-half of the control cells did not develop phase 3, as discussed earlier. The oligomycin effect was not seen when the cells were also treated with CPA (n = 3, data not shown).

View larger version (18K): [in this window] [in a new window]   Fig. 7.   Summary of Ca2+ removal rates determined under various conditions at [Ca2+]i of 75 nM. All average data were obtained from five cells except for control experiments (n = 11). One-way ANOVA was carried out and the subsequent Tukey's test detected significant differences (**P < 0.01) between CCCP and oligomycin, diazoxide, and oligomycin, and cyclopiazonic acid (CPA) and oligomycin.

The results with oligomycin argue against the hypothesis that diazoxide and CCCP inhibited phase 3 by ATP depletion. Also, the apparent facilitation of Ca²⁺ removal by oligomycin is puzzling. Perhaps the hypothesis originally suggested using cardiac myocytes (4) and then the portal vein (12) that mitochondrial Ca²⁺ content influences sarcoplasmic reticulum Ca²⁺ load may be useful. If Ca²⁺ is initially stored in mitochondria and then shuttled to the sarcoplasmic reticulum, agents depolarizing mitochondria will inhibit sarcoplasmic reticulum Ca²⁺ uptake due to reduced Ca²⁺ supply. On the other hand, oligomycin hyperpolarizes mitochondria, increases Ca²⁺ content in mitochondria, and facilitates Ca²⁺ supply to the sarcoplasmic reticulum. The hyperpolarization caused by oligomycin is expected to be small because the effect of oligomycin is due to prevention of proton reentry into mitochondria, a process normally producing a small mitochondrial depolarization. This is in fact the case, as shown in Fig. 6. Nonetheless, the small hyperpolarization by oligomycin was sufficient to significantly promote Ca²⁺ removal.

There is no doubt that more work has to be carried out for final validation of the hypothesis that the Ca²⁺ content of mitochondria influences that of the sarcoplasmic reticulum. It is possible that data presented here are biased against the Ca²⁺ removal mechanisms requiring enzymes because all experiments were carried out at room temperature. Moreover, whether or not the mitochondria sequester Ca²⁺ when [Ca²⁺]_i is raised by physiological agonists, such as adrenergic agents, needs to be addressed in a future study. Also, our study is solely based on temporal analysis of [Ca²⁺]_i that provides no spatial information. In this regard, it is interesting to note that, using HeLa and HEK-293 cells expressing genetically engineered cameleon indicators, Arnaudeau et al. (1) recently provided evidence that Ca²⁺ reuptake to the endoplasmic reticulum is facilitated by Ca²⁺ efflux from the adjacent mitochondria. The conclusion offered by Arnaudeau et al. (1), that mitochondria promote endoplasmic reticulum Ca²⁺ uptake by providing Ca²⁺, not ATP, is also consistent with our findings. At the moment, the hypothesis that some Ca²⁺ is transferred from mitochondria to sarcoplasmic reticulum in the rat femoral artery seems best to reconcile our observations.

In light of its clinical importance (see below), the mechanism of diazoxide action on mitochondria merits further discussion. Diazoxide is probably best known as an opener of ATP-sensitive K⁺ (K_ATP) channels found not only in the plasma membrane (3) but also in mitochondrial membranes (21). More importantly, it has been suggested that the opening of mitochondrial K_ATP channels may confer resistance against ischemia to the heart. If the heart is subjected to mild ischemia, it becomes profoundly resistant against subsequent and more serious ischemic episodes (14), a phenomenon known as cardiac ischemic preconditioning. Interestingly, pretreatment with K_ATP channel openers seems to mimic the protective effect of the ischemic preconditioning (14). To date, however, beyond the notion that K_ATP channel openers exert preconditioning-like effects through mitochondrial rather than salcolemmal K_ATP channels, there is little consensus regarding this subject. Diazoxide has been useful in studying preconditioning because it selectively opens mitochondrial K_ATP channels (11). Although it seems logical to assume that activation of mitochondrial K_ATP channels will depolarize mitochondria (28), a recent report (13) suggested that diazoxide depolarizes mitochondria by inhibiting respiratory chain complex II, not by opening K_ATP channels. Our results seem to support this notion that in some cell types diazoxide depolarizes mitochondria independent of mitochondrial K_ATP channel opening. In this study, diazoxide triggered elevation in [Ca²⁺]_i and inhibited the phase 3 Ca²⁺ removal when cells were dialyzed with Cs⁺. Under this condition, the opening of mitochondrial K_ATP channels seems unlikely to produce mitochondria depolarization. Likewise, Kowaltowski et al. (27) reported that replacement of K⁺ in mitochondria bathing medium with Li⁺, another poorly permeant ion to K channels, had little effect on diazoxide actions. It should be noted, however, that these authors ascribed cardioprotective effects of K_ATP channel openers to mitochondrial volume regulation, not to depolarization (27). These authors also reported few inhibitory effects of diazoxide on mitochondrial Ca²⁺ uptake (27). Our results, however, suggest that diazoxide depolarizes the mitochondrial membrane, discharges stored Ca²⁺ and modulates the Ca²⁺ removal profile. These discrepancies may be explained in part by the difference in experimental methods and materials.

One unresolved important question is how mitochondria sequester Ca²⁺ over a low-[Ca²⁺]_i range. In particular, during sustained depolarization to 30 mV, [Ca²⁺]_i rarely reaches values high enough for low-affinity Ca²⁺ removal mechanisms to be activated. One possibility is that some mitochondria may be closely located to Ca²⁺ channels, sensing higher Ca²⁺ than bulk, averaged [Ca²⁺]_i. This may, in turn, cause mitochondria to modulate Ca²⁺ influx. If Ca²⁺ near the cell membrane is cleared by the mitochondria, Ca²⁺-induced inactivation of voltage-dependent Ca²⁺ channels will be suppressed. Indeed, Ca²⁺ influx regulation by mitochondria acting as a Ca²⁺ sink has been suggested for Ca²⁺ channels in neurons (5) and Ca²⁺ release-activated Ca²⁺ channels in T lymphocytes (20).

In the present study, we provided evidence that mitochondria play an important role in Ca²⁺ regulation using rat femoral arterial smooth muscle cells. Mitochondrial Ca²⁺ clearance seems important over lower [Ca²⁺]_i range because phase 3 was inhibited with mitochondrial depolarizing agents. Moreover, steady-state [Ca²⁺]_i measured at a constant membrane potential of 30 mV was elevated on addition of CCCP. Unlike cardiac muscle, arteries do not generally fire action potentials and may require a rather small increase in [Ca²⁺]_i to cause maximum contraction (29). Mitochondrial participation in shaping [Ca²⁺]_i during sustained and modest depolarization, therefore, may have particular relevance in arterial smooth muscle cells. Also, CCCP significantly prolonged the duration of Ca²⁺ transients evoked by caffeine application, implying a mitochondrial role in Ca²⁺ homeostasis when [Ca²⁺]_i is raised by Ca²⁺ release. Although further effort is required to completely elucidate mitochondrial contribution in Ca²⁺ regulation, the results described here may offer some insight in Ca²⁺ homeostasis in arterial smooth muscle cells.