Functionally Separate Intracellular Ca 2 (cid:1) Stores in Smooth Muscle*

,

a mechanism that regulates smooth muscle contractile activity, involves the participation of two receptor/channel complexes, the ryanodine receptor (RyR) and the inositol 1,4,5-trisphosphate receptor (Ins(1,4,5)P 3 R). Release from this store regulates the bulk average [Ca 2ϩ ] c both directly (1) and indirectly either via modulation of the plasmalemmal membrane potential (2) or by activation of store-operated Ca 2ϩ entry (3). The magnitude, time course, and frequency of the SR Ca 2ϩ signal depend on the functional interaction, localization, and arrangement of the Ins(1,4,5)P 3 R and RyR on the SR store(s).
Although, morphologically, the SR appears as an interconnected network of tubules (4,5), it may adopt different configurations within the cell and components may detach and reattach thereby influencing the pattern and distribution of the RyR and Ins(1,4,5)P 3 R (6, 7). In Purkinje neurons, for example, Ins(1,4,5)P 3 R-expressing regions may detach from other internal store elements (8,9). Indeed, different Ca 2ϩ concentrations have been found within the lumen of the SR (10) suggesting that discontinuities may exist within the structures surrounding the lumen itself. This provides a morphological basis for the existence of various arrangements of Ca 2ϩ stores.
Further support for the existence of two separate stores has come from studies on the response of each of the receptors to inhibitors of the SR Ca 2ϩ pump, thapsigargin and cyclopiazonic acid (CPA), each of which depletes the stores of Ca 2ϩ . Differences in the sensitivity to the pump inhibitors of the Ca 2ϩ release evoked by either caffeine or Ins(1,4,5)P 3 have been interpreted as evidence for the existence of separate stores for each receptor (5,16,22). For example, in arterial myocytes the ryanodine/caffeine-sensitive store was not sensitive to either thapsigargin or CPA, whereas the Ins(1,4,5)P 3 -sensitive Ca 2ϩ store was depleted by each (16,22). The situation has been complicated further by the proposed existence in murine blad-der smooth muscle cells (23) of three Ca 2ϩ stores, one containing RyR and Ins(1,4,5)P 3 R, another expressing only Ins(1,4,5)P 3 R, and a third containing only RyR.
Whereas the proposed arrangements of RyR and Ins(1,4,5)P 3 R may reflect the complexity of the underlying biology, differences in experimental approaches may also have contributed to the variety of views expressed. For example, caffeine, commonly used to activate RyR, also inhibits Ins(1,4,5)P 3 R (24,25). In some studies (e.g. 12, 21) caffeine remained present while an Ins(1,4,5)P 3 -generating agonist was applied. Invariably, such experiments demonstrated an inhibition of the Ins(1,4,5)P 3mediated response and the results have been taken as evidence for the existence of a common Ca 2ϩ store. Inhibition of the Ins(1,4,5)P 3 receptor by caffeine, rather than depletion of a common store, could have accounted for the absence of response to an Ins(1,4,5)P 3 -generating agonist.
Additional difficulties in the classification of SR Ca 2ϩ stores, i.e. their location and number, have followed the use of plasmalemmal agonists and the multiple, yet separate, biochemical pathways so activated. Two particular aspects of such difficulties are evident. First, when membrane currents are used as indicators of [Ca 2ϩ ] c , agonists may modify these currents independently of SR Ca 2ϩ release (e.g. 28,29). Second, regulation of the RyR and Ins(1,4,5)P 3 R by Ca 2ϩ derived from agonist activation of several different biochemical pathways may occur with misleading consequences. For example, in rabbit portal vein, depletion of the Ins(1,4,5)P 3 -sensitive store, by norepinephrine, abolished the response to caffeine (which acts on the RyR (27)), consistent with both receptors residing on a common Ca 2ϩ store. On the other hand, Ca 2ϩ released from the SR via Ins(1,4,5)P 3 R activation may have triggered a regenerative Ca 2ϩ -induced Ca 2ϩ release (CICR) at the RyR (11,28), which could have amplified the Ins(1,4,5)P 3 -evoked Ca 2ϩ transient. If so, two outcomes could be anticipated (a) the continued presence of Ins(1,4,5)P 3 could deplete the RyR-sensitive Ca 2ϩ pool; (b) depletion of the RyR-sensitive Ca 2ϩ pool would reduce the response to Ins(1,4,5)P 3 . Either of these results could be misinterpreted as support for the existence of a common Ca 2ϩ store.
Notwithstanding these difficulties, it is important to determine the arrangement of Ca 2ϩ stores in smooth muscle to help clarify the precise mechanisms of Ca 2ϩ release, a vital ingredient in our understanding of contractility. This problem has been addressed in the current investigation by seeking answers to the following questions: 1) Are Ins(1,4,5)P 3 R and RyR present on the same store or on separate stores of the SR? 2) Does Ca 2ϩ released from the Ins(1,4,5)P 3 -sensitive store trigger CICR via activation of the RyR? 3) Are there differences between the refilling mechanisms of Ins(1,4,5)P 3 -sensitive and ryanodine-sensitive intracellular Ca 2ϩ stores? In this study freshly isolated single smooth muscle cells rather than multicellular preparations were used, removing the difficulty of there being different store characteristics existing in different cells or of store reorganization, which may accompany the use of cell culture preparations. Ca 2ϩ influx was controlled under voltage clamp conditions and directly measured in this investigation. Flash photolysis of caged Ins(1,4,5)P 3 (Ins(1,4,5)P 3 ) and caffeine were each used to minimize activation of second messenger systems so that a clearer understanding of the control of Ca 2ϩ release from the receptors could be obtained. From the results of the present study it is proposed that two functionally distinct SR Ca 2ϩ stores exist in colonic myocytes; one expressing both Ins(1,4,5)P 3 R and RyR and dependent upon an external Ca 2ϩ source for replenishment and a second store containing only RyR, which can be refilled form [Ca 2ϩ ] c .

EXPERIMENTAL PROCEDURES
Cell Isolation-From male guinea pigs (500 -700 g) stunned by a blow to the head and immediately killed by exsanguination, a segment of distal colon (ϳ5 cm) was removed. The circular muscle was separated from the longitudinal layer, and single cells were prepared from the former using a two-step enzymatic process (30), stored at 4°C, and used the same day.
Membrane Current Recording-Cells were voltage-clamped in the dialyzed whole cell configuration. Currents were amplified by an Axopatch 1D (Axon Instruments, Union City, CA), low pass filtered at 500 Hz (8-pole Bessel filter, Frequency Devices, Haverhill, MA), and digitally sampled at 1.5 kHz using a Digidata interface, pCLAMP software (version 6.0.1, Axon Instruments), and Axotape (Axon Instruments) and stored for analysis. Cells were held at a membrane potential (V m ) of Ϫ70 mV unless otherwise indicated. The bathing solution contained (mM): sodium glutamate, 80; NaCl, 40; tetraethylammonium chloride, 20; Cytosolic Ca 2ϩ Concentration ([Ca 2ϩ ] c ) Measurement-[Ca 2ϩ ] c was measured using the membrane-impermeable fluo-3 (penta-ammonium salt). Fluorescence measurements were made using a microfluorometer consisting of an inverted fluorescence microscope (Nikon Diaphot) and a photomultiplier tube with a bi-alkali photocathode. Fluo-3 was excited at 488 nm (bandpass 9 nm) from a PTI Delta Scan (Photon Technology International Inc., East Sheen, London, UK) through the epi-illumination port of the microscope (using one arm of a bifurcated quartz fiber optic bundle). Excitation light was passed through a field stop diaphragm, to reduce background fluorescence, and reflected off a 505-nm long-pass dichroic mirror; emitted light was guided through a 535-nm barrier filter (bandpass 35 nm) to a photomultiplier in photon-counting mode. Longer wavelengths, from bright field illumination with a 610-nm Shott glass filter, were reflected onto a charge-coupled device camera (Sony model XC-75) mounted onto the viewing port of the Delta Scan thus allowing the cell to be monitored during experiments. Interference filters and dichroic mirrors were obtained from Glen Spectra (London, UK). To photolyze caged Ins(1,4,5)P 3 the output of a xenon flashlamp (Rapp OptoElektronic, Hamburg, Germany) was passed though a UG-5 filter to select ultraviolet light and merged into the excitation light path of the microfluorometer using the second arm of the quartz bifurcated fiber optic bundle. The nominal flash lamp energy was 57 mJ, measured at the output of the fiber optic bundle. The flash duration was about 1 ms.
Caffeine (10 mM) was applied by hydrostatic pressure (Pneumatic PicoPump PV820, World Precision Instruments, Inc., Sarasota, FL). All experiments were carried out at room temperature (18 -22°C), and drugs were applied either hydrostatically via a pipette or into the bathing solution as indicated in the text.
Data Analysis-Changes in cytosolic Ca 2ϩ were expressed as a ratio (F/F o ) of the fluorescence counts (F) relative to baseline counts before stimulation (F o ). ⌬F/F o indicates the magnitude of the change in F/F o at the peak of the evoked transient relative to the baseline ratio. Original fluorescence records were not filtered, smoothed, or averaged. Background fluorescence was not subtracted. Statistical analyses were performed using either Mann-Whitney tests (on normalized data) or paired Student's t tests (on raw data). Summarized data are shown as means Ϯ S.E. and taken to be statistically significant when p Ͻ 0.05. n indicates numbers of cells used.
Does CICR from the RyR Contribute to the InsP(1,4,5)P 3evoked Ca 2ϩ Transient?-If on the other hand two separate stores exist, i.e. one for Ins(1,4,5)P 3 R and another for RyR, release of a small amount of Ca 2ϩ from the Ins(1,4,5)P 3 -sensitive store could trigger a further, larger release of Ca 2ϩ from the separate ryanodine-sensitive store by CICR. If so, depletion of the ryanodine-sensitive store, by caffeine and ryanodine, would reduce the Ins(1,4,5)P 3 -evoked response. If Ca 2ϩ , released through the Ins(1,4,5)P 3 R, triggered CICR at the RyR, ryanodine alone would reduce Ins(1,4,5)P 3 -evoked Ca 2ϩ transients. This was not observed (Fig. 3). Ins(1,4,5)P 3 evoked reproducible increases in [Ca 2ϩ ] c of similar magnitude in the presence (50 M) and absence of ryanodine (n ϭ 5, Fig. 3, V m ϭ Ϫ70 mV). Thus Ca 2ϩ released by Ins(1,4,5)P 3 did not subsequently trigger CICR from the RyR, and this provides further evidence for the existence of a common Ca 2ϩ store. In other investigations, reduction, by ryanodine, of the Ca 2ϩ transient evoked by Ins(1,4,5)P 3 -generating agents was interpreted as evidence that Ins(1,4,5)P 3 -evoked Ca 2ϩ activates CICR at the RyR (11,28). The plasmalemma agonists used in these experiments to generate Ins(1,4,5)P 3 could also have activated other second messengers that in turn sensitized the RyR to Ca 2ϩ enabling Ins(1,4,5)P 3 -evoked Ca 2ϩ release to activate CICR at the RyR. Alternatively, Ca 2ϩ release from the SR store may activate further Ca 2ϩ release under conditions of "store overload" (31,32). Such store overload conditions could conceivably arise in some smooth muscle types, facilitating CICR.
Not All RyR Are Present on the Store, Which Contains InsP(1,4,5)P 3 R-To determine whether or not all RyR were present on the store that contained Ins(1,4,5)P 3 R, the Ins(1,4,5)P 3 -sensitive store was depleted by removal of external Ca 2ϩ and the ability of caffeine to activate the RyR and evoke a Ca 2ϩ transient was examined. Refilling of the Ins(1,4,5)P 3 -sensitive store is dependent on external Ca 2ϩ (33), and removing it reduced the response to Ins(1,4,5)P 3 to 5 Ϯ 2% of controls (n ϭ 6; p Ͻ 0.05; V m ϭ Ϫ70 mV, Fig. 5). However, after the almost complete loss of the Ins(1,4,5)P 3 -evoked transient, caffeine evoked a Ca 2ϩ transient that averaged 74 Ϯ 25% of control values (n ϭ 6; p Ͻ 0.05; Fig. 5). These results are consistent with there being a second separate Ca 2ϩ store that contains only RyR.
Refilling of the Ca 2ϩ Stores-The above results (Fig. 5) raised the possibility that the degree of dependence of the two stores on external Ca 2ϩ for Ca 2ϩ release may differ. This was examined following withdrawal of external Ca 2ϩ by investigating the refilling of the RyR-and Ins(1,4,5)P 3 -sensitive stores after either caffeine or Ins(1,4,5)P 3 . The caffeine-evoked Ca 2ϩ transient (via RyR; Fig. 6A) was reduced, on average, to some 87 Ϯ 9% of controls (n ϭ 5; p ϭ 0.5; Fig. 6, A and B). In contrast, the Ins(1,4,5)P 3 -evoked Ca 2ϩ transient (acting through Ins(1,4,5)P 3 R) was reduced to 6 Ϯ 2% of controls ( Fig. 6B; see Fig. 5). These results suggest that, unlike the situation with the Ins(1,4,5)P 3 -sensitive Ca 2ϩ store (Fig. 5), Ca 2ϩ release from  the RyR by caffeine may be recycled so that refilling is largely independent of external Ca 2ϩ .

Effects of Elevation of cAMP on InsP(1,4,5)P 3 -evoked Ca 2ϩ
Transients-Caffeine inhibits phosphodiesterase activity and so may elevate the intracellular concentration of cAMP ([cAMP] c ) (34). The persistence of the store Ca 2ϩ content, in the absence of external Ca 2ϩ , as indicated by the maintained amplitude of the caffeine-evoked Ca 2ϩ transient, could have arisen from stimulation of SERCA by an elevated [cAMP] c due to caffeine (35) rather than to a difference in the refilling mechanism. To examine this possibility, dependence of Ins(1,4,5)P 3 store refilling on external Ca 2ϩ was examined when [cAMP] c had been increased (a) by the phosphodiesterase inhibitor IBMX (500 M) and (b) by forskolin (1 M), which stimulates adenylate cyclase thereby raising [cAMP] c . In the absence of either drug, Ins(1,4,5)P 3 -evoked Ca 2ϩ transients of approximately reproducible amplitude that averaged 1.89 Ϯ 0.12 ⌬F/F o (n ϭ 6). Following incubation (10 min) with either IBMX or forskolin, Ins(1,4,5)P 3 -evoked transients of approximately reproducible amplitude (2.18 Ϯ 0.67 ⌬F/F o ; n ϭ 6; Fig. 7, for IBMX), which were not significantly different from controls. Upon removal of external Ca 2ϩ , in the continued presence of either IBMX or forskolin, repeated application of Ins(1,4,5)P 3 depleted the Ins(1,4,5)P 3 -sensitive store as evidenced by the decline in the amplitude of the Ca 2ϩ transient. With IBMX, after the fourth Ins(1,4,5)P 3 challenge, the Ca 2ϩ increase averaged 15 Ϯ 3% of the Ins(1,4,5)P 3 -evoked Ca 2ϩ transients observed in IBMX in the presence of external Ca 2ϩ (0.56 Ϯ 0.31 ⌬F/F o ; p Ͻ 0.01; n ϭ 6; Fig. 7). Qualitatively similar results were obtained with forskolin. Removal of external Ca 2ϩ again inhibited the amplitude of the Ins(1,4,5)P 3 -evoked transient significantly to 8 Ϯ 3% of controls (p Ͻ 0.05 by Mann-Whitney test; data not shown; V m ϭ Ϫ70 mV). In these same cells only 3 Ϯ 2% of the Ins(1,4,5)P 3 -evoked transient remained in the presence of forskolin (1 M) following the removal of external Ca 2ϩ (n ϭ 3; p Ͻ 0.05 by Mann-Whitney test). Together the results with IBMX and forskolin indicated that elevation of [cAMP] c is unlikely to offset the effect of external Ca 2ϩ withdrawal on store content.

DISCUSSION
Ca 2ϩ release from and uptake into internal SR Ca 2ϩ stores plays a central role in the activity of most cells, including that of excitation-contraction coupling in smooth muscle. Although it is conceded that release occurs via Ins(1,4,5)P 3 R and RyR, the interrelationship between these two receptors in their access to stored Ca 2ϩ is unclear. A diversity of store and receptor arrangements has been proposed. This study has demonstrated the presence of two Ca 2ϩ stores in colonic smooth muscle, one expressing both Ins(1,4,5)P 3 R and RyR the other only RyR. Support for this view comes from two key observations in which the stores accessed by either Ins(1,4,5)P 3 R or RyR were depleted of Ca 2ϩ . First, the store accessed by RyR was depleted of Ca 2ϩ (by repetitive activation with caffeine in the presence of ryanodine to maintain the RyR in the open state), and under these circumstances the response to Ins(1,4,5)P 3 was virtually abolished. Second, when the store accessed by Ins(1,4,5)P 3 R was depleted, by withdrawal of external Ca 2ϩ , and the Ins(1,4,5)P 3 response was effectively abolished, significantly, the Ca 2ϩ response to RyR activation by caffeine persisted. The observation that the Ins(1,4,5)P 3 -evoked Ca 2ϩ response was almost abolished after caffeine (in the presence of ryanodine) could not be attributed to inhibition of an amplification of the Ins(1,4,5)P 3 -evoked response by CICR, because Ins(1,4,5)P 3evoked Ca 2ϩ release did not activate this mechanism. Nor could the observation be explained by inactivation of the Ins(1,4,5)P 3 R by caffeine because the Ins(1,4,5)P 3 -evoked Ca 2ϩ transient had fully recovered at the time intervals between application of caffeine and Ins(1,4,5)P 3 used in the present study. Finally, the results of the present study demonstrated differences in the refilling mechanisms of the two stores. The store expressing both Ins(1,4,5)P 3 R and RyR was dependent on external Ca 2ϩ for replenishment whereas the store with only RyR was not.
The precise number and arrangement of Ca 2ϩ stores, among different cell types, is a matter of debate. In cerebellar Purkinje neurons, as in the present study, ryanodine inhibited the Ins(1,4,5)P 3 -evoked Ca 2ϩ transient, in keeping with the view that both Ins(1,4,5)P 3 R and RyR were present on a common store. However, whether or not an additional separate store with only RyR involved was not determined (38). One store with both Ins(1,4,5)P 3 R and RyR and an additional separate store, with only RyR similar to the present view, have also been proposed in vascular smooth muscle (21). Other studies had initially raised the possibility of there being two stores but with different receptor arrangements from those presently proposed. For example, from experiments in smooth muscle, of the two Ca 2ϩ stores proposed, one contained both RyR and Ins(1,4,5)P 3 R, the other only Ins(1,4,5)P 3 R (19,20). Alternatively, separate stores have been proposed for both RyR and Ins(1,4,5)P 3 receptors (e.g. 39 -41). A commonly employed method with which to study the receptors on the Ca 2ϩ stores has been to inhibit store Ca 2ϩ pumps by thapsigargin or cyclopiazonic acid (CPA) and to observe the responses to activation of Ins(1,4,5)P 3 R and RyR. In various cell types, thapsigargin and CPA each abolished Ca 2ϩ release in response to activation of either RyR or Ins(1,4,5)P 3 R but not to activation of both. This supported the idea of there being more than one store (36,37,(42)(43)(44). Studies in smooth muscle and astrocytes suggested that, despite an apparently uniform Ca 2ϩ content throughout most of the store, CPA and the RyR activator caffeine released Ca 2ϩ from seemingly separate compartments (5,22), i.e. there were in effect two stores. However, the relationship between SR Ca 2ϩ pumps on the one hand and the Ins(1,4,5)P 3 R and RyR on the other could not be presently differentiated by the use of Ca 2ϩ pump inhibitors. SR Ca 2ϩ pumps presumably on both the common Ins(1,4,5)P 3 R/RyR store and the RyR-only store were each inhibited by thapsigargin and CPA (see also 45,46). Differences in the sensitivity of the Ca 2ϩ pumps on the internal SR Ca 2ϩ store to the inhibitors (47,48), which may also vary among different tissues, were presumably responsible for these observations.

Differences in SR Luminal Ca 2ϩ Regulation of the Receptors Does Not Account for the Differences in Response to
InsP (1,4,5)P 3 and Caffeine-On the basis of the two-store system proposed, the observation that a substantial caffeineevoked Ca 2ϩ transient persisted, after depletion of the Ins(1,4,5)P 3 -sensitive store, could be explained if the opening of the Ins(1,4,5)P 3 R and RyR was each regulated differently by the SR luminal Ca 2ϩ concentration (49 -52). Thus, as luminal Ca 2ϩ decreased, the Ins(1,4,5)P 3 sensitivity to Ins(1,4,5)P 3 could have diminished and the receptor could no longer be able to release Ca 2ϩ , even though the SR itself had not been depleted. The sensitivity of the RyR to caffeine could have been affected to a lesser degree than that of the Ins(1,4,5)P 3 R by decreases in luminal Ca 2ϩ , and caffeine could have remained effective in evoking Ca 2ϩ release. However, if a single common store for both Ins(1,4,5)P 3 R and RyR existed (rather than the presently proposed two stores), luminal regulation of the receptors with the accompanying differences in sensitivity between RyR and Ins(1,4,5)P 3 R could also serve as a model in which the SR could function, in effect, as two stores. If luminal regulation was to account for the present results, a 25% decrease in the SR Ca 2ϩ content would be sufficient to prevent Ca 2ϩ release via the Ins(1,4,5)P 3 R, significantly different from previously published values (70 -95% (53)(54)(55)). This arises from the observation that some 75% of the Ca 2ϩ transient evoked by caffeine persisted when the Ins(1,4,5)P 3 -evoked transient was abolished. SR luminal Ca 2ϩ regulation of the Ins(1,4,5)P 3 R thus seems unlikely to account for the present observations. Nor is luminal regulation responsible for differences between the two stores in their dependence upon external Ca 2ϩ for replenishment. Thus the response to caffeine is maintained whereas that to Ins(1,4,5)P 3 disappears after the removal of external Ca 2ϩ . Together, these findings support the presence of two discrete SR Ca 2ϩ stores; one able to refill from [Ca 2ϩ ] c , the other from extracellular Ca 2ϩ entry.
Mechanisms of Store Refilling-Depletion of Ins(1,4,5)P 3sensitive stores in all cell types so far examined activates a store-operated Ca 2ϩ entry pathway, which is necessary for its replenishment (reviewed in Refs. 56 -58) as is depletion of the ryanodine-sensitive store, although not in all cell types (59 -63). In the present study, replenishment of the store containing RyR alone did not require external Ca 2ϩ , suggesting that plasmalemmal store-operated Ca 2ϩ entry is unnecessary for the maintenance of the Ca 2ϩ content of this store. In the store containing both RyR and Ins(1,4,5)P 3 R, on the other hand, external Ca 2ϩ entry, presumably via store-operated channels, is essential for store refilling in colonic myocytes (33). Perhaps differences in store location within the cell determines the external Ca 2ϩ dependence of the refilling mechanisms. The store containing both RyR and Ins(1,4,5)P 3 R may be positioned closer to the plasmalemma than that containing only RyR and may be functionally more closely linked to Ca 2ϩ entry via store-operated channels (63). Different isoforms of SERCA, with different Ca 2ϩ binding affinities, e.g. K 0.5 0.31 Ϯ 0.02 M and 0.17 Ϯ 0.01 M Ca 2ϩ for SERCA2a and SERCA2b, respectively (Ref. 42 for review; see also Refs. [65][66][67], may be associated with the separate stores and could explain the ability of the SR to refill in the presence of different [Ca 2ϩ ] c . Ca 2ϩ uptake into stores by SERCA2b may also be modulated by calmodulin (68), phospholamban (64,67), and calreticulin and calnexin (26) so that differential modulation of Ca 2ϩ pumps may also help to explain differences in store refilling in the absence of external Ca 2ϩ .
The present study emphasizes the complexity of the organization of the SR Ca 2ϩ stores. It proposes the existence of two in the control of Ca 2ϩ release; one containing both RYR and Ins(1,4,5)P 3 R and a second only RyR. The results also highlight the possibility of structural components within the SR, each with their unique content of receptors/channels permitting local control of the SR Ca 2ϩ signal in response to receptor modulation.