Corresponding author R. M. Drummond: Department of Physiology, Biomedical Imaging Group, University of Massachusetts Medical Center, 373 Plantation Street, Worcester, MA 01605, USA. Email: firstname.lastname@example.org
1The Ca2+-sensitive fluorescent indicator rhod-2 was used to measure mitochondrial [Ca2+] ([Ca2+]m) in single smooth muscle cells from the rat pulmonary artery, while simultaneously monitoring cytosolic [Ca2+] ([Ca2+]i) with fura-2.
2Application of caffeine produced an increase in [Ca2+]i and also increased [Ca2+]m. The increase in [Ca2+]m occurred after the increase in [Ca2+]i, and remained elevated for a considerable time after [Ca2+]i had returned to resting values.
3The protonophore carbonyl cyanide p-(trifluoromethoxy)phenylhydrazone (FCCP), which causes the mitochondrial membrane potential to collapse, markedly attenuated the increase in [Ca2+]m following caffeine application and also increased the half-time for recovery of [Ca2+]i to resting values.
4Activation of purinoceptors with ATP also produced increases in both [Ca2+]i and [Ca2+]m in these smooth muscle cells. In some cells, oscillations in [Ca2+]i were observed during ATP application, which produced corresponding oscillations in [Ca2+]m and membrane currents.
5This study provides direct evidence that Ca2+ release from the sarcoplasmic reticulum, either through ryanodine or inositol 1,4,5-trisphosphate (InsP3) receptors, increases both cytosolic and mitochondrial [Ca2+] in smooth muscle cells. These results have potential implications both for the role of mitochondria in Ca2+ regulation in smooth muscle, and for understanding how cellular metabolism is regulated.
The importance of Ca2+ uptake by mitochondria under pathological conditions of Ca2+ overload has long been well established. However, recently evidence has been accumulating that Ca2+ uptake by mitochondria also occurs under physiological conditions in a variety of cell types (Rizzuto et al. 1992; Hajnóczky et al. 1995). Ca2+ uptake by mitochondria appears to have two consequences. First, by buffering [Ca2+]i, mitochondria will affect the spatial and temporal aspects of cytosolic Ca2+ signalling (Duchen et al. 1990; Herrington et al. 1996; Drummond & Fay, 1996). Second, since mitochondria possess several Ca2+-sensitive dehydrogenases, mitochondrial Ca2+ uptake provides a mechanism whereby the increased energy demand brought about by stimulation can be matched to an enhancement in cellular ATP production by mitochondria (McCormack et al. 1990; Hansford, 1991; Hajnóczky et al. 1995; Robb-Gaspers et al. 1998a).
Uptake of Ca2+ by mitochondria occurs via an electrophoretic uniporter which is driven by the large electrical potential across the inner membrane, set up through proton extrusion by the respiratory chain (Gunter & Pfeiffer, 1990). The extrusion of Ca2+ out of the mitochondria mainly occurs through the electroneutral exchange of Ca2+ for 2 Na+, although a Na+-independent Ca2+ efflux mechanism may play a role in some cell types (Gunter & Pfeiffer, 1990). In cardiac muscle, increases in [Ca2+]m have been observed during electrical stimulation, using both electron probe microanalysis (Wendt-Gallitelli & Isenberg, 1991) and fluorescence spectroscopy (Miyata et al. 1991; Sheu & Jou, 1994; Trollinger et al. 1997).
In smooth muscle cells, uptake of Ca2+ by mitochondria remains a controversial issue. While studies using 45Ca2+ have indicated Ca2+ uptake by mitochondria when the extracellular [K+] is increased (Karaki & Weiss, 1981), other studies using electron probe X-ray microanalysis did not show any increase in mitochondrial [Ca2+] during a K+-induced contraction (Somlyo et al. 1979). More recently, a number of studies utilizing pharmacological approaches have provided indirect evidence of a role for mitochondria in Ca2+ regulation (Drummond & Fay, 1996; McGeown et al. 1996; Greenwood et al. 1997). However, the importance of mitochondrial Ca2+ uptake must be established from direct measurement of [Ca2+]m in living cells.
In the present study the Ca2+-sensitive fluorescent indicator rhod-2 was used to measure [Ca2+]m in isolated smooth muscle cells from the rat pulmonary artery, while [Ca2+]i was simultaneously monitored with fura-2. Release of Ca2+ from the sarcoplasmic reticulum (SR) was accomplished by two different approaches: application of caffeine, which increases the Ca2+ sensitivity of the ryanodine receptor to Ca2+, thereby promoting Ca2+-induced Ca2+ release (Itoh et al. 1981); and application of extracellular ATP, which acting through the purinoceptor causes production of inositol trisphosphate (InsP3) and release of Ca2+ from the SR (Guibert et al. 1996). Both ATP and caffeine produced an increase in [Ca2+]m, as indicated by the changes in rhod-2 fluorescence. The kinetics of the changes in [Ca2+]m differed significantly from those observed for the [Ca2+]i. Hence, this study provides direct evidence that mitochondrial Ca2+ uptake is important in the regulation of cytosolic Ca2+ in vascular smooth muscle cells.
Male Sprague-Dawley rats (150-200 g) were killed by cervical dislocation, as approved by the University of Massachusetts Medical Center Animal Care Committee, following guidelines of the US Departments of Agriculture, and Health and Human Services. The heart and lungs were removed en bloc into dissecting solution of the following composition (mM): NaCl, 119; KCl, 4.7; KH2PO4, 1.18; MgSO4, 1.17; glucose, 5.5; NaHC03, 25; Hepes, 10; pH 7.4 with NaOH. From the left lung, small diameter (∼0.7-1 mm) pulmonary arterial vessels were dissected. Smooth muscle cells were then dissociated using a modification of the procedure described by Albarwani et al. 1995. Briefly, ring segments (∼1 mm in length) were placed in dissociation solution of the following composition (mM): NaCl, 128; KCl, 5.4; KH2PO4, 0.95; Na2HPO4, 0.35; glucose, 10; sucrose, 2.9; NaHCO3, 4.16; Hepes, 10; pH 7.3 with NaOH. This solution also contained papain (1.5 mg ml−1) and DL-dithiothreitol (1 mg ml−1), and was maintained at 4°C for 60 min. The solution containing the tissue was then transferred to a water bath at 37°C for 6 min. The arterial rings were then transferred to fresh dissociation solution also containing collagenase Type III (1.5 mg ml−1) and incubated in the water bath for a further 5 min at 37°C. Thereafter, gentle trituration of the tissue with a fire polished Pasteur pipette yielded single smooth muscle cells.
Measurement of mitochondrial and cytosolic [Ca2+]
In order to measure [Ca2+]m, isolated smooth muscle cells were incubated with the Ca2+-sensitive fluorescent indicator rhod-2 AM (2 μM), for at least 1 h at room temperature. Rhod-2 has been used by several different laboratories to monitor changes in [Ca2+] within the mitochondrial matrix (Mix et al. 1994; Sheu & Jou, 1994; Rutter et al. 1996; Babcock et al. 1997; Trollinger et al. 1997). The rationale for using rhod-2 is that the net positive charge on rhod-2 AM results in significant compartmentalization of this indicator in the mitochondria. Once hydrolysed to the Ca2+-sensitive membrane-impermeable form, it becomes trapped inside the mitochondrial matrix. Due to uncertainties concerning the Kd of the indicator within this compartment and the accuracy of establishing Fmin and Fmax, rhod-2 fluorescence was not converted to [Ca2+]. [Ca2+]i was monitored with fura-2 (50 μM), which was included in the patch pipette solution. Fura-2 ratios were converted to [Ca2+]i using the method described by Grynkiewicz et al. (1985), and an assumed Ca2+-fura-2 Kd of 200 nM. Rmax, Rmin and β were determined as previously described (Becker & Fay, 1987). Fluorescence was measured using a custom built multi-wavelength microfluorimeter (Fig. 1).
Whole-cell membrane currents were recorded using thin walled borosilicate patch electrodes (3-5 MΩ) (World Precision Instruments, Sarasota, FL, USA) with an Axopatch-1D patch clamp amplifier (Axon Instruments). Following rupture of the membrane patch at least 7 min were allowed for dialysis of the cytosol before commencing any experimental protocols. This extended period of dialysis facilitated removal of any residual rhod-2 from the cytosol. Data were digitally stored on an IBM-compatible PC, at a frequency of 1 kHz after being filtered with a low-pass filter (200 Hz cut-off), for subsequent off-line analysis.
In all the experiments, cells were bathed in an extracellular solution of the following composition (mM): NaCl, 150; KCl, 5.4; MgCl2, 1.2; CaCl2, 1.8; glucose, 10; Hepes, 10; pH 7.4 with NaOH. The pipette solution contained (mM): KCl, 125; MgCl2, 4; Hepes, 10; Na2ATP, 5; GTP, 0.5; pH 7.3 adjusted with KOH. All experiments were carried out at room temperature (20-23°C).
Reagents and data analysis
Rhod-2 AM and fura-2 were obtained from Molecular Probes; collagenase Type III, papain, FCCP and all other reagents were from Sigma. FCCP was prepared as a 10 mM stock solution in DMSO. Caffeine (20 mM) and ATP (100 μM) were applied from a pressure ejection pipette, positioned about 100 μm from the cell, using a Picospritzer II (General Valve, Fairfield, NJ, USA). Data are shown as the mean ± standard error of the mean (s.e.m.), where n refers to the number of cells. Statistical tests of difference were made using Student's t tests for paired and unpaired observations. P < 0.05 was considered to be statistically significant.
Caffeine-induced changes in [Ca2+]i and [Ca2+]m
When caffeine (20 mM) was applied from a pressure ejection pipette to pulmonary artery smooth muscle cells, [Ca2+]i increased from a resting value of 151 ± 12 nM to 826 ± 54 nM (n= 23) (Fig. 2A). The peak increase in [Ca2+]i occurred 3.98 ± 0.39 s after onset of the 10 s application of caffeine. The release of Ca2+ from the SR also produced an increase in mitochondrial rhod-2 fluorescence (F/Fo), from 1 to 1.62 ± 0.08, which reached its peak value 6.71 ± 0.56 s after onset of caffeine application (Fig. 2A). The time to peak for the increase in mitochondrial rhod-2 fluorescence was significantly longer than the time to peak for the increase in [Ca2+]i (P < 0.001). This is seen more clearly in Fig. 2A, where [Ca2+]i and rhod-2 fluorescence, from the same cell shown in Fig. 1A, have been plotted on an expanded time scale. On plotting [Ca2+]ivs. rhod-2 fluorescence (Fig. 2A) it is evident that [Ca2+]i increases by several hundred nanomolar before there is any increase in [Ca2+]m. The [Ca2+]i decreased to resting values with a half-time (t½) of 6.06 ± 0.48 s, while during the same time the rhod-2 fluorescence had only decreased by 4 ± 1 %. Even at 60 s after the onset of caffeine application, when [Ca2+]i was 123 ± 13 nM, the rhod-2 fluorescence had only decreased from the peak value by 9 ± 3 %. Thus, [Ca2+]i returns to resting levels while there is very little change in [Ca2+]m.
When the membrane potential was held at -60 mV, the increase in [Ca2+]i due to caffeine application activated an inward current which had a peak value of 195 ± 28 pA. This current has previously been identified as a Ca2+ activated Cl− current (Clapp et al. 1996; Guibert et al. 1997). In a separate series of experiments, ryanodine (100 μM) completely blocked the response to caffeine (data not shown), indicating that the increase in [Ca2+]i is entirely due to release of Ca2+ from the SR.
Effect of FCCP on caffeine-induced changes in [Ca2+]i and [Ca2+]m
To show that Ca2+ is taken up by mitochondria and that rhod-2 is a reliable indicator of [Ca2+]m, the effect of the protonophore FCCP was examined. Since Ca2+ uptake by mitochondria is dependent upon the mitochondrial membrane potential, causing this potential gradient to collapse with a protonophore should impair the ability of mitochondria to take up Ca2+ from the cytosol. Thus, cells were exposed to FCCP (1 μM) before the application of caffeine. In the presence of FCCP the resting [Ca2+]i was 130 ± 29 nM. [Ca2+]i increased to 716 ± 102 nM after application of caffeine (20 mM) with a time to peak of 8.16 ± 1.01 s (n= 6) (Fig. 3A). The time to peak for the increase in [Ca2+]i was significantly longer in the presence of FCCP compared with control (P < 0.02) (Fig. 4). Rhod-2 fluorescence increased slightly from 1 to 1.17 ± 0.03 with a time to peak of 11.69 ± 0.57 s. The time to peak for the increase in rhod-2 fluorescence was significantly longer than the time to peak for the increase in [Ca2+]i (P < 0.02).
In the presence of FCCP, the t½ for cytosolic Ca2+ removal was increased from 6.06 ± 0.48 s under control conditions to 11.57 ± 2.43 s (P < 0.05) (Fig. 4). At this same time, rhod-2 fluorescence had only decreased by 3 ± 1 %. Sixty seconds after the onset of caffeine application, when the resting [Ca2+]i was 106 ± 25 nM, the rhod-2 fluorescence had decreased to its resting value of 1.01 ± 0.02.
ATP-induced changes in [Ca2+]i and [Ca2+]m
When ATP (100 μM) was applied to these smooth muscle cells, for 10 s from a pressure ejection pipette, [Ca2+]i increased from 139 ± 15 to 881 ± 65 nM with a time to peak of 5.50 ± 0.60 s (n= 11) (Fig. 5A). Mitochondrial rhod-2 fluorescence also increased from 1 to 1.72 ± 0.12 and the time to peak was 7.25 ± 0.54 s, which was significantly longer than the time to peak for the increase in [Ca2+]i (P < 0.001) (Fig. 5A and C).
When the membrane potential was clamped at -60 mV, ATP activated an inward current similar to that described by Hartley & Kozlowski (1997) for these cells, and therefore presumed to be a Ca2+-activated Cl− current. The peak value for this inward current was 207 ± 53 pA. [Ca2+]i returned to its resting value with a t½ of 5.10 ± 0.47 s, while during the same time the rhod-2 fluorescence had decreased by 2 ± 1 %. In 4 of the 11 cells, brief application of ATP was found to cause oscillations in [Ca2+]i (Guibert et al. 1996; Bakhramov et al. 1996; Hartley & Kozlowski, 1997), which produced corresponding oscillations in mitochondrial rhod-2 fluorescence (Fig. 6) and membrane currents. The oscillations in these four cells stopped approximately 60 s after application of ATP, by which time the [Ca2+]i was 127 ± 13 nM. The mitochondrial rhod-2 fluorescence only decreased by 14 ± 1 %, 60 s after the onset of ATP application.
In smooth muscle cells, following removal of a contractile stimulus, [Ca2+]i is restored to resting values by ATP-dependent Ca2+ pumps in the SR and plasma membrane (Missiaen et al. 1991), and also by the plasma membrane Na+-Ca2+ exchanger (McCarron et al. 1994). The role of the mitochondria in this process is far less certain, although recent studies utilizing pharmacological approaches have clearly implicated mitochondrial involvement (Drummond & Fay, 1996; Greenwood et al. 1997). In the present study, [Ca2+]m has been monitored with the Ca2+-sensitive fluorescent indicator rhod-2, while [Ca2+]i was simultaneously monitored with fura-2. Application of contractile agonists, which are capable of releasing Ca2+ from the SR of these smooth muscle cells, produced an increase in both [Ca2+]i and [Ca2+]m. The increase in [Ca2+]m occurred after a brief delay with respect to the increase in [Ca2+]i. Furthermore, [Ca2+]m continued to increase after [Ca2+]i had reached its peak value, and was still increasing as [Ca2+]i returned towards its resting value. The consequences of mitochondrial Ca2+ uptake in these cells will be discussed below with respect to Ca2+ homeostatic mechanisms, and as a means for regulating cellular oxidative phosphorylation.
There are several lines of evidence to support the use and efficaciousness of rhod-2 as an indicator of changes in mitochondrial Ca2+. First, rhod-2 has been used by several different laboratories to monitor [Ca2+]m during stimulation, in a variety of cell types (Mix et al. 1994; Sheu & Jou, 1994; Rutter et al. 1996; Babcock et al. 1997; Trollinger et al. 1997). While selective loading procedures have frequently been used to achieve compartmentalization of rhod-2 in the mitochondrial matrix (Trollinger et al. 1997), the present study offers an additional advantage over this approach. Specifically, by establishing the whole-cell recording configuration, any rhod-2 which has been hydrolysed in the cytosol can be dialysed away. Consequently, the rhod-2 signal is relatively free from contamination by cytosolic indicator. Second, the time courses of the observed changes in fura-2 and rhod-2 fluorescence are significantly different, consistent with these two indicators as monitors of Ca2+ in two different cellular compartments. Furthermore, if rhod-2 was compartmentalized in the SR, then the fluorescence would be expected to decrease upon stimulation rather than increase. Third, the protonophore FCCP markedly attenuated the increase in rhod-2 fluorescence normally seen following application of caffeine. The small residual response in the presence of FCCP, which was observed in the present study and in other studies using both rhod-2 (Babcock et al. 1997; David et al. 1998) and mitochondrially targeted aequorin (Rizzuto et al. 1998; Nakazaki et al. 1998), is most likely to reflect incomplete collapse of the mitochondrial potential gradient following application of FCCP. Thus, it can be concluded with reasonable certainty that rhod-2 is a reliable indicator of the changes in [Ca2+]m that occur upon stimulation.
It is only by simultaneously monitoring [Ca2+]i and [Ca2+]m at high temporal resolution that detailed information regarding the changes in [Ca2+] within the cytosol and mitochondria can be obtained. In this respect, it is evident that the increase in [Ca2+]m appears to lag behind the increase in [Ca2+]i, such that [Ca2+]i increases by several hundred nanomolar before there is any measurable increase in [Ca2+]m (see Fig. 2A). This may seem somewhat surprising, given the reported close proximity of the mitochondria to the SR in smooth muscle cells (Nixon et al. 1994), and the notion that Ca2+ release from the SR can generate microdomains with high Ca2+ close to mitochondria (Rizzuto et al. 1993, 1998; Duchen et al. 1998). However, a possible explanation for this finding is that the mitochondrial uniporter may be maintained in an inactive mode, which requires activation before Ca2+ can be taken up into the matrix (Gunter & Pfeiffer, 1990).
Mitochondrial [Ca2+] also continued to increase after [Ca2+]i had reached its peak, and even after [Ca2+]i began to decline back towards resting values, [Ca2+]m was still increasing. A similar finding has recently been observed in rat chromaffin cells during membrane depolarization, where [Ca2+]m continued to increase after Ca2+ influx had been terminated and [Ca2+]i was returning towards resting values (Babcock et al. 1997). This suggests that mitochondrial Ca2+ uptake may be playing an important role in Ca2+ removal in these smooth muscle cells. Further support for mitochondria contributing to Ca2+ removal from the cytosol is provided by the fact that FCCP significantly increased the t½ for [Ca2+]i recovery following application of caffeine. Since 5 mM ATP was included in the patch pipette, it seems unlikely that the effect of FCCP can be attributed to insufficient ATP supply to the various membrane pumps. Additional support for the importance of mitochondrial Ca2+ uptake in smooth muscle cells comes from a recent study showing that inhibition of mitochondrial Ca2+ uptake prolongs the decay time constant of Ca2+-activated Cl− currents (Greenwood et al. 1997).
Inhibition of mitochondrial Ca2+ uptake with FCCP did not lead to any significant change in resting [Ca2+]i, which is consistent with our previous findings (Drummond & Fay, 1996) and supports the view that mitochondria are not responsible for setting the resting [Ca2+]i. It should also be mentioned that FCCP significantly increased the time to peak for the caffeine-induced increase in [Ca2+]i. This finding is in agreement with a recent report that FCCP slowed the rate of bradykinin-induced Ca2+ release from the SR in BHK-21 cells (Landolfi et al. 1998). In the study by Landolfi et al. (1998) the effect of FCCP was largely overcome by loading cells with the Ca2+ chelator BAPTA, highlighting the importance of mitochondrial Ca2+ buffering in shaping the SR Ca2+ release kinetics.
For the cells used here release of Ca2+ from the SR, either through ryanodine or InsP3 receptors, appeared to produce very similar increases in [Ca2+]m, both in terms of the magnitude and kinetics of the response. An additional effect of ATP on these cells is that it produced oscillations in [Ca2+]i (see Fig. 6). While only 4 of 11 cells displayed this phenomenon following brief application of ATP, oscillations were produced in all cells studied during continuous application of ATP (data not shown). These oscillations in [Ca2+]i were associated with oscillations in [Ca2+]m and membrane current. Oscillations in [Ca2+]m have been observed in astrocytes (Jou et al. 1996) and hepatocytes (Hajnóczky et al. 1995; Robb-Gaspers et al. 1998B) where rhod-2 was used to monitor [Ca2+]m, and in a pancreatic cell line where mitochondrially targeted aequorin was used to monitor [Ca2+]m (Nakazaki et al. 1998). In the only other studies where mitochondrial and cytosolic Ca2+ were monitored in the same cell, during oscillations in [Ca2+]i, the [Ca2+] changes within these two compartments appeared to be in phase (Hajnóczky et al. 1995; Robb-Gaspers et al. 1998B). This is similar to what we are reporting in the present study, but it is not yet clear why this is the case. Given the obvious lag in the increase in [Ca2+]m, following the increase in [Ca2+]i during the initial response, one might expect that during oscillations the increase in [Ca2+]m is delayed relative to the increase in [Ca2+]i. It is conceivable that the recently discovered rapid mode of Ca2+ uptake could account for the two signals being in phase during oscillations (Sparagna et al. 1995). However, it is evident that more work is required to understand better the changes in [Ca2+]m that occur during oscillations in [Ca2+]i.
Another feature of the change in mitochondrial [Ca2+] is that it takes many minutes for Ca2+ to leave the mitochondria following the stimulus. The mechanism underlying this slow efflux is not clear at the present time. Certainly, studies on isolated mitochondria have shown that they can retain Ca2+ for several minutes following uptake into the matrix (Crompton et al. 1978). Furthermore, it has also been reported that mitochondria isolated from uterine and ileal muscle have no Na+-Ca2+ efflux mechanism (Crompton et al. 1978), and this may account for the slow time course of Ca2+ efflux observed in the present study.
In smooth muscle, the matching of ATP synthesis with ATP utilization is particularly critical to normal function because the amount of preformed phosphagen (ATP plus phosphocreatine) is small, relative to the rates of utilization concomitant with contraction. In fact, in actively contracting smooth muscle, the peak of an isometric contraction could not be attained before the cellular preformed phosphagen stores are completely depleted (Paul et al. 1984). In smooth muscle, oxidative phosphorylation is the main mechanism for supplying ATP to the contractile machinery; however, the mechanism whereby stimulation of contraction leads to enhanced ATP production remains to be established. Since mitochondria contain several dehydrogenases (pyruvate dehydrogenase, NAD+-isocitrate dehydrogenase, 2-oxoglutarate dehydrogenase) which can be activated by Ca2+ (McCormack et al. 1990; Hansford 1991), the increase in [Ca2+]m provides a potential mechanism for stimulating oxidative phosphorylation in these cells, thereby matching ATP supply with demand.
This work was supported by a Grant-In-Aid from the American Heart Association (Massachusetts Affiliate) to R. M. D. and HL 47530 from NIH. We would like to thank Drs A. Guerrero, J. J. Singer and J. V. Walsh Jr for continuous discussion and critical reading of this manuscript. I would like to dedicate this work to the memory of my mentor, Professor Fredric S. Fay, who so tragically passed away on March 18, 1997 (R. M. D.).