Ca2+ Sparks Act as Potent Regulators of Excitation-Contraction Coupling in Airway Smooth Muscle*

Ca2+ sparks are short lived and localized Ca2+ transients resulting from the opening of ryanodine receptors in sarcoplasmic reticulum. These events relax certain types of smooth muscle by activating big conductance Ca2+-activated K+ channels to produce spontaneous transient outward currents (STOCs) and the resultant closure of voltage-dependent Ca2+ channels. But in many smooth muscles from a variety of organs, Ca2+ sparks can additionally activate Ca2+-activated Cl− channels to generate spontaneous transient inward current (STICs). To date, the physiological roles of Ca2+ sparks in this latter group of smooth muscle remain elusive. Here, we show that in airway smooth muscle, Ca2+ sparks under physiological conditions, activating STOCs and STICs, induce biphasic membrane potential transients (BiMPTs), leading to membrane potential oscillations. Paradoxically, BiMPTs stabilize the membrane potential by clamping it within a negative range and prevent the generation of action potentials. Moreover, blocking either Ca2+ sparks or hyperpolarization components of BiMPTs activates voltage-dependent Ca2+ channels, resulting in an increase in global [Ca2+]i and cell contraction. Therefore, Ca2+ sparks in smooth muscle presenting both STICs and STOCs act as a stabilizer of membrane potential, and altering the balance can profoundly alter the status of excitability and contractility. These results reveal a novel mechanism underlying the control of excitability and contractility in smooth muscle.

. STOCs hyperpolarize the membrane and turn off preactivated voltage-dependent Ca 2ϩ channels (VDCCs), leading to the relaxation of smooth muscle (4,6). Knock-out of either the pore-forming BK ␣ subunit or the auxiliary BK ␤1 subunit results in elevated blood pressure or overactive bladder in mice (8 -10). A weaker coupling between Ca 2ϩ sparks and STOCs has also been linked to angiotensin II-induced hypertension (11) and diabetic retinopathy (12) in animal models.
In many smooth muscles from a variety of organs, however, Ca 2ϩ sparks, in addition to activating STOCs, turn on Ca 2ϩactivated Cl Ϫ (Cl Ca ) channels to generate spontaneous transient inward currents (STICs) (7,(13)(14)(15)(16)(17). Because E Cl in smooth muscle is less negative than resting membrane potential (RMP), the activation of STICs by Ca 2ϩ sparks is expected to depolarize the membrane, an opposite effect to that of STOCs. But the precise effect of Ca 2ϩ sparks on membrane potential in this class of smooth muscle has not been experimentally determined nor has the physiological consequence of changes in membrane potentials caused by Ca 2ϩ sparks. Therefore, a major unsolved question is the physiological function of Ca 2ϩ sparks in smooth muscle cells that possess both STOCs and STICs.
In this study, we used airway smooth muscle (ASM), a prototypical smooth muscle exhibiting STOCs and STICs (7,14,18), to explore the physiological function of Ca 2ϩ sparks by directly measuring Ca 2ϩ sparks, membrane potential, global intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ), and contractile state. We demonstrated that Ca 2ϩ sparks under physiological conditions induce biphasic membrane potential transients (BiMPTs), leading to membrane potential oscillations. BiMPTs clamp membrane potential within a negative range and prevent the generation of action potentials, thus serving as a potent mechanism to maintain a low excitability of the cells. Strikingly, the reagents that block Ca 2ϩ sparks or upset the balance of BIMPTs can depolarize the membrane and activate VDCCs, resulting in an increase in global [Ca 2ϩ ] i and contraction. Therefore, Ca 2ϩ sparks and BiMPTs function as powerful regulators of membrane excitability and contractility in smooth muscle.

EXPERIMENTAL PROCEDURES
Cell Isolation-Male Swiss Webster mice (4 -6 weeks) were euthanized with intraperitoneal injection of a lethal dose of sodium pentobarbital (50 mg kg Ϫ1 ). After each animal was unresponsive to applied stimulus, the trachea was quickly removed and placed in pre-chilled dissociation solution consisting of (in mM) the following: 136 NaCl, 5.36 KCl, 0.44 KH 2 PO 4 , 4.16 NaHCO 3 , 0.34 Na 2 HPO 4 , 5 MgCl 2 , 20 Hepes, and 10 glucose (pH 7.1). The trachea was dissected free from the surface of connective tissues and incubated in the dissociation medium with 30 units/ml papain, 0.2 mM dithiothreitol, and 0.02 mM EDTA at room temperature for 30 min. The tissue was then incubated at 32°C for another ϳ6 min with dissociation medium containing 3 units/ml collagenase 1A, 0.2 mg/ml Pronase E, 0.1 mg/ml DNase I, and 1 mg/ml bovine serum albumin. Finally, the tissue was agitated with a fire-polished wide bore glass pipette to release the cells. The isolated single cells were used on the day of isolation, and all the experiments were carried out at room temperature (22-25°C).
Patch Clamp Recording and Analysis-Conventional or perforated whole-cell voltage clamp or current clamp recording was done with an Axopatch-1D amplifier or a HEKA EPC10 amplifier. The extracellular solution contained (in mM) the following: 130 NaCl, 5.5 KCl, 2.2 CaCl 2 , 1 MgCl 2 , and 10 Hepes, pH adjusted to 7.4 with NaOH. The pipette solution contained (in mM) the following: 75 KCl, 64 potassium aspartate, 1 MgCl 2 , 3 Na 2 ATP, and 10 Hepes, pH adjusted to 7.3 with KOH. In conventional whole-cell clamp, 0.05 mM fluo-3 K 5 was included in the patch pipette. For perforated patch technique, fluo-3 K 5 was omitted, and 160 g/ml amphotericin B was added in the patch pipette. Events were analyzed by the mini analysis program. To minimize the effect of leak conductance at the seal between the patch pipette and membrane, only cells with a seal resistance greater than Ͼ1 gigohm were recorded and analyzed. The capacitance of these cells was 41.3 Ϯ 1.5 picofarads (n ϭ 27). Because of oscillating nature of the membrane potential, all-point histograms were constructed using the segments of interest, and the mode values were taken as the RMP.
Measurement and Analysis of Ca 2ϩ Sparks-Fluorescence images using fluo-3 as a calcium indicator were obtained using a custom-built wide field digital imaging system (7). Rapid imaging at 100 Hz (exposure, 3 ms) was made possible by equipping the system with a cooled high sensitivity, charge-coupled device camera developed in conjunction with the Lincoln Laboratory, Massachusetts Institute of Technology (Lexington, MA (7)). The camera was interfaced to a custom-made inverted microscope, and the cells were imaged using a ϫ60 Nikon 1.4 NA oil, giving a pixel size of 333 nm at the specimen. The 488 nm line of an argon ion laser provided fluorescence excitation, with a shutter to control exposure duration, and emission of the Ca 2ϩ indicator was monitored at wavelengths of Ͼ500 nm. Subsequent image processing and analysis were performed off line using a custom-designed software package, running on either a Silicon Graphics or Linux/PC workstation. Signal mass of Ca 2ϩ sparks was estimated according to the methodology published previously (19). The measured endogenous Ca 2ϩ buffers from the same type of the cells were used to correct for the signal mass (13).
Measurement of Global [Ca 2ϩ ] i -To monitor global cytosolic [Ca 2ϩ ] i , fura-2 fluorescence was measured using a custombuilt multiwavelength microfluorimeter (20). Briefly, the system consisted of a Zeiss IM-35 inverted microscope (Nikon ϫ40, 1⅐3NA) with a specially designed excitation path and photomultiplier tube (Thorn EMI type 9954A, Thorn EMI, Rock-FIGURE 1. Relationship between Ca 2؉ sparks and their evoked membrane potential transients. A, images of approximately one-third of a cell display the spatiotemporal evolution of a single Ca 2ϩ spark. The cell was current-clamped without injecting current. The reversal potentials for Cl Ϫ and K ϩ were set at Ϫ15 and Ϫ80 mV, respectively, in accordance with the concentration gradient of these two ions in smooth muscle under physiological conditions (28). The images were acquired at a rate of 100 Hz with an exposure time of 3 ms. Cytosolic Ca 2ϩ was measured using fluo-3 (50 M), which was introduced into the cell in the K ϩ form through the patch pipette. Changes in Ca 2ϩ concentration in the images are expressed as ⌬F/F 0 (%) and displayed on a pseudocolor scale calibrated at the right of images. away, NJ). The excitation path included a series of dichroic mirrors and a "chopper" wheel, which permitted excitation of up to four wavelengths every 20 ms. In this study, the excitation wavelengths for fura-2 were 340 and 380 nm (ϳ10-nm bandwidth), and a 560-nm (80-nm bandwidth) emission filter was used for fura-2 emission. An image mask was used to exclude the field beyond the cell, thereby reducing the background contribution to the fluorescence signal. [Ca 2ϩ ] i was monitored with fura-2 AM (1 M), after loading for 30 min at room temperature. Fura-2 ratios were converted to [Ca 2ϩ ] i using the method described by Grynkiewicz et al. (21), and an assumed K d of 200 nM for Ca 2ϩ -fura-2. R max , R min , and ␤ were determined as described previously (22).
Measurement of Cell Shortening-Myocytes were placed to a recording chamber superfused with the bath solution for patch clamp experiments at room temperature. Cells were imaged using a custom-built wide field digital imaging system (7), and their lengths were determined by a custom-written software.
Reagents and Their Application-All chemicals except fluo-3, fura-2 (Invitrogen), and ryanodine (Calbiochem) were purchased from Sigma. Agonists and antagonists were applied locally to cells via a picospritzer at a constant pressure, so that the duration of its action and concentration can be controlled with ease.
Statistics-Unless stated otherwise, data are reported as mean Ϯ S.E. Statistical analysis of difference was made with Student's paired t test, and the significance level was set at p Ͻ 0.05.

Ca 2ϩ Sparks Induce BiMPTs via Activating BK Channels and Cl Ca Channels under Physiological
Conditions-Ca 2ϩ sparks activate only STICs at E K , only STOCs at E Cl , and spontaneous transient outward and inward currents at potentials between E K and E Cl in ASM from mouse as they do in ASM from other species, and in the smooth muscle from other tissues (supplemental Fig. S1) (7,14,15,17). To explore the role of Ca 2ϩ sparks under physiological conditions, Ca 2ϩ sparks and membrane potential were simultaneously measured with a combination of high speed fluorescence imaging and conventional whole-cell current clamp technology. At rest (i.e. without injection of current), spontaneous local Ca 2ϩ transients were detected (Fig.  1A), and they were Ca 2ϩ sparks because no transient was observed in the presence of 100 M ryanodine. Estimated with signal mass (SM) methodology (19), the signal mass and peak Ca 2ϩ current underlying these Ca 2ϩ sparks (I Ca(spark ) were 244,794 Ϯ 24,818 Ca 2ϩ ions and 3.56 Ϯ 0.32 pA (n ϭ 60), respectively. The amplitudes of SM and I Ca(spark) were independent of the onset membrane potential (V on ) as determined by the events activated by Ca 2ϩ sparks (see below). Collectively, these results suggest that Ca 2ϩ underlying Ca 2ϩ sparks in phys- showing that ϳ90% of the time membrane potential is more negative than Ϫ40 mV, the potential at which L-type Ca 2ϩ channel currents can be detected in these cells. JANUARY 15, 2010 • VOLUME 285 • NUMBER 3 iological conditions are from Ca 2ϩ release from RyRs and not from Ca 2ϩ influx from outside of the cells.

Physiological Role of Ca 2؉ Sparks in Airway Myocytes
Ca 2ϩ sparks correlated temporally with BiMPTs with a hyperpolarization followed by a depolarization (Fig. 1, A and B). (At times, Ca 2ϩ sparks correlated only with either hyperpolarization transients or depolarization transients, as indicated by the events with zero value in Fig. 1C.) The V on of BiMPTs was Ϫ46.3 Ϯ 0.9 mV (n ϭ 60); and the average amplitude of the hyperpolarization phase in BiMPTs was Ϫ20.1 Ϯ 1.4 mV and that of the depolarization was 10.1 Ϯ 1.3 mV (n ϭ 60). These values are in line with those obtained in much longer recordings under perforated whole-cell configurations (see Fig. 2). Thus, a single Ca 2ϩ spark can change membrane potential by ϳ30 mV under physiological conditions.
Scatter plots of the parameters of Ca 2ϩ sparks and BiMPTs in Fig. 1C show weak correlations between SM and the hyperpolarization phase or depolarization phase of BiMPTs, and between I Ca(spark) and both phases of BiMPTs. These data indicate a great variation between the coupling of Ca 2ϩ sparks and their target channels, as suggested in the voltage clamp studies (13,23).
To determine the underlying channels for BiMPTs, cells were treated with the blockers of BK channels and Cl Ca channels, respectively. Paxilline (1 M), a BK channel blocker, abolished the hyperpolarization phases of BiMPTs ( Fig. 2A, panel i). Niflumic acid (100 M), a Cl Ca channel blocker, gradually abolished the depolarization phases of BiMPTs ( Fig. 2A, panel ii).
(Changes in the RMP in these experiments will be discussed below.) Therefore, Ca 2ϩ sparks induce BiMPTs by activating BK channels and Cl Ca channels in ASM from mouse.
BiMPTs Set the Range of Membrane Potential under Physiological Conditions-Ca 2ϩ sparks continually activated BiMPTs, resulting in the oscillation of membrane potentials with varying amplitude (Fig. 2). Given the variation between the coupling of Ca 2ϩ sparks and BiMPTs (Fig. 1C), all-points histograms of the recorded membrane potentials were used to quantify the relationship between BiMPT and RMP. Because the histogram could not fit well with a Gaussian function, the mode of the histogram was designated as RMP (Fig. 2B, panel ii). In the example in Fig. 2B, the value of RMP was Ϫ46 mV, and the histogram was skewed to the left. Overall the direction of skewness depends on RMP, i.e. the more negative the RMP, the more likely the distribution skews to the right (toward depolarization) and vice versa (Fig. 2B, panel iii). Also, in 39 cells with over 20 h of recordings, no single BiMPT or monophasic transient was observed to depolarize beyond E Cl (i.e. Ϫ15 mV) or hyperpolarize below E K (i.e. Ϫ80 mV), and membrane potential oscillates between Ϫ60 mV and Ϫ40 mV for ϳ80% of the recording time. Of these 39 cells, in 14 cells membrane transients never reach Ϫ40 mV. Finally, 1 M nifedipine, an L-type Ca 2ϩ channel blocker, did not change RMP, nor did it alter the amplitude of depolarization and hyperpolarization components in BiMPTs (n ϭ 3). The results in Fig. 2B demonstrate that Ca 2ϩ sparks and BiMPTs act to prevent the membrane from either extreme hyperpolarization or extreme depolarization, thus serving as a stabilizing mechanism of membrane potential.
The above assertion is strengthened by the experiments in Fig. 2A where both paxilline and niflumic acid changed RMP.
By blocking the hyperpolarization phase of BiMPTs, paxilline shifted RMP to a more depolarized level ( Fig. 2A, panel i, RMP: Ϫ44 Ϯ 1.8 mV in control versus Ϫ22 Ϯ 2.4 mV in the presence of paxilline, n ϭ 5, p Ͻ 0.001 with paired t test). Along with inhibiting the depolarization phase of BiMPTs, niflumic acid (100 M) caused the membrane to become more hyperpolarized ( Fig. 2A, panel ii, RMP: Ϫ44 Ϯ 3.7 mV in control versus Ϫ56 Ϯ 3.6 mV in the presence of niflumic acid, n ϭ 5, p Ͻ 0.01 with paired t test). Because niflumic acid does not activate BK channels in airway smooth muscle (14), 3 as it does in vascular smooth muscle (24), the hyperpolarization by niflumic acid is most likely mediated by its inhibition of Cl Ca channels. Therefore, both BK channels and Cl Ca channels are critical to membrane potential, and their balanced activation is required to maintain RMP at physiological conditions. Ca 2ϩ Sparks and BiMPTs Prevent Evoked Action Potential-ASM does not fire action potentials under physiological conditions, but the underlying reasons remain incompletely understood (25). In light of the stabilizing effect of Ca 2ϩ sparks on the membrane potential, we explored the hypothesis that Ca 2ϩ sparks and BiMPTs are the signals that prevent ASM from generating action potential. In the presence of Ca 2ϩ sparks and BiMPTs, short depolarizing currents with amplitude (10 -1200 3 R. ZhuGe, R. Bao, K. E. Fogarty, and L. M. Lifshitz, unpublished results. B, same depolarizing current as in A triggered an action potential when BiMPTs were blocked by ryanodine (Ry, 100 M). To prevent the inactivation of VDCCs by the sustained depolarization caused by ryanodine (Fig. 2), Ϫ50 pA was applied to reset RMP to approximately Ϫ60 mV. C, nifedipine (Nif, 1 M) blocked the evoked action potential in the presence of ryanodine, indicating that this potential is mediated by L-type VDCCs. Insets depict expanded views of the recordings marked by dotted boxes. Four cells gave similar responses.

Physiological Role of Ca 2؉ Sparks in Airway Myocytes
pA) and duration (2-20 ms) did not induce an action potential. Fig. 3A shows a typical response to a pulse of 400 pA for 5 ms. When delivered either at the valley or the peak of BiMPTs, this pulse only caused a passive potential transient. Interestingly, in the presence of ryanodine at the concentration sufficient to block Ca 2ϩ sparks and BiMPTs, the same stimulation generated an action potential after the passive potential transient (Fig. 3B). (The membrane potential was reset to around Ϫ60 mV after ryanodine but prior to stimulation, because ryanodine caused sustained depolarization, which in turn could inactivate L-type VDCCs ( Fig. 2A, panel iii) (26). Nifedipine (1 M) blocked the action potential induced by the depolarizing current, confirming that the action potential is mediated by the opening of L-type VDCCs. In conjunction with results in Fig. 2, it is reasonable to suggest that at rest Ca 2ϩ sparks and the resulting BiMPTs act as a major inhibitory mechanism of membrane excitability in ASM.
Tipping the Balance of BiMPTs Results in Membrane Depolarization-The inhibitory and stabilizing nature of Ca 2ϩ sparks and BiMPTs on membrane excitability predicts that reagents that alter these events could change membrane excitability. This prediction was confirmed in Fig. 2A where paxilline (1 M), by blocking the hyperpolarization phase of BiMPTs, depolarized membrane to a level approaching E Cl . Strikingly, ryanodine by blocking BiMPTs also depolarized the membrane close to E Cl ( Fig. 2A, panel iii; RMP: Ϫ48 Ϯ 2.3 mV in the control versus Ϫ27 Ϯ 4.5 mV in the presence of ryanodine, n ϭ 5, p Ͻ 0.001 with paired t test). These observations prompted us to determine the underlying channels for the induced depolarization. Because both reagents depolarize membrane potential close to E Cl , the involvement of Cl Ca channels was first examined. Fig. 4, A and B, shows that niflumic acid (100 M) prevented both ryanodine-and pax-
Because the depolarization caused by paxilline and ryanodine is more positive than the potential at which the L-type Ca 2ϩ channel current was detected in these cells (supplemental Fig. S2), the effect of nifedipine on the depolarization induced by these two compounds was next assessed. Fig. 4, C and D, demonstrates that nifedipine (1 M) blocked the sustained depolarization caused by ryanodine (see Fig. 2A, panel ii) and paxilline (see Fig. 2A, panel i), respectively (Ϫ47.8 Ϯ 2.9 in the control versus Ϫ47.3 Ϯ 3.2 in the presence of nifedipine and ryanodine, n ϭ 5, p Ͼ 0.05; and Ϫ47.5 Ϯ 1.7 in the control and Ϫ45.0 Ϯ 2.3 mV in the presence of nifedipine and paxilline, n ϭ 4, p Ͼ 0.05). Contrary to the effect on ryanodine, nifedipine did not affect the depolarization components induced by paxilline. These results indicate that blocking either Ca 2ϩ sparks or the hyperpolarization phase of BiMPTs can trigger a positive feedback loop between Cl Ca channels and VDCCs, resulting in a depolarization close to E Cl .
Tipping the Balance of BiMPTs Leads to an Increase in Global [Ca 2ϩ ] i and Cell Shortening-The activation of VDCCs by depolarization induced by ryanodine and paxilline suggests Ca 2ϩ sparks and BiMPTs could regulate global [Ca 2ϩ ] i . Yet, as is evident in Fig. 1, a single Ca 2ϩ spark raised Ca 2ϩ locally but exerted no effect on the global [Ca 2ϩ ] i . Therefore, we examined the accumulated effect of Ca 2ϩ sparks and BiMPTs on [Ca 2ϩ ] i by monitoring global [Ca 2ϩ ] i dynamics using fura-2, a ratiometric indicator that is well suited for quantifying [Ca 2ϩ ] i . Although these cells exhibit spontaneous Ca 2ϩ sparks at rest, the global [Ca 2ϩ ] i in the majority of cells (59 of 61) was stable with a mean value of 137 Ϯ 12 nM, and nifedipine (1 M) caused no change in resting [Ca 2ϩ ] i (Fig. 5), indicating that L-type Ca 2ϩ channels do not contribute significantly to set the resting [Ca 2ϩ ] i in these cells.
We next examined whether changes in Ca 2ϩ sparks and BiMPTs alter global [Ca 2ϩ ] i by treating cells with ryanodine and paxilline. At 100 M, ryanodine increased the [Ca 2ϩ ] i from 136 Ϯ 18 to 210 Ϯ 23 nM (p Ͻ 0.01, n ϭ 7), and paxilline (1 M) elevated it to 213 Ϯ 25 nM from 131 Ϯ 19 nM (p Ͻ 0.01, n ϭ 11) (Fig. 5). Removal of both agents restored global [Ca 2ϩ ] i to its normal resting values. Nifedipine (1 M) blocked the increase in [Ca 2ϩ ] i caused by either ryanodine or paxilline (Fig. 5B). Niflumic acid itself exerted no effect on [Ca 2ϩ ] i , but it blocked the increase in Ca 2ϩ induced by either paxilline or ryanodine (Fig.  5B). Put together, the results in Fig. 5 (27), we measured cell shortening at the single cell level in an attempt to establish a direct link between Ca 2ϩ sparks/BiMPTs and contractility. Cells used in this study are relaxed and spindle-shaped, with a length of 166 Ϯ 5 m (n ϭ 57). Fig. 6A demonstrates that ryanodine (100 M) initiated a contraction within 5 s of application and short-ened the cell by 39% within 90 s of treatment. On average, ryanodine contracted the cells by 20.4 Ϯ 3.7% (p Ͻ 0.01, n ϭ 10). Paxilline (1 M) caused a similar effect on cell shortening as ryanodine. On average, paxilline contracted the cells by 15.5 Ϯ 2.2% (p Ͻ 0.01, n ϭ 9). Both niflumic acid (100 M) and nifedipine (1 M) prevented ryanodine-and paxilline-induced contraction (Fig. 6B).

DISCUSSION
In this study we provide direct evidence that Ca 2ϩ sparks and BiMPTs play critical roles in determining the status of membrane excitability and contractility in ASM. At rest, they maintain the membrane potential within a negative range and prevent the cells from generating action potentials triggered by external stimuli, thus keeping the cells at a low level of excitability. Blocking Ca 2ϩ sparks or the hyperpolarization components of BiMPTs turns this inhibitory mode into an excitable mode by activating Cl Ca channels and VDCCs in a positive feedback manner, which in turn depolarizes the membrane, raises global [Ca 2ϩ ] i , and induces contraction.
It has long been recognized that Ca 2ϩ sparks activate Cl Ca channels and BK channels in a voltage-dependent manner in ASM (7). In smooth muscle cells, reversal potentials for Cl Ϫ and K ϩ are around Ϫ15 and Ϫ80 mV, respectively (28). Therefore, under physiological conditions, Ca 2ϩ sparks are expected to cause biphasic membrane potential transients and, moreover, due to the depolarization component of the transients, to activate L-type Ca 2ϩ VDCCs, leading to an increase in global [Ca 2ϩ ] i or even to the generation of an action potential. Evidence from this study confirms the first prediction but does not validate the second one. We found that BiMPTs dwell most of the time below the potential at which the activity of L-type VDCCs was detected. This is in line with findings that the L-type Ca 2ϩ channel blocker nifedipine neither alters RMP amplitudes and membrane potential oscillations nor does it decrease resting [Ca 2ϩ ] i . The ineffectiveness of nifedipine further indicates that Ca 2ϩ spark-induced depolarizations that transiently reach the potential where L-type VDCCs can be activated either do not activate these channels at all or activate them to such a minimal extent that Ca 2ϩ influx by them is not sufficient to alter global [Ca 2ϩ ] i . The latter could happen if the entering Ca 2ϩ is buffered by endogenous Ca 2ϩ buffers (13) or is compensated by Ca 2ϩ extrusion mechanisms in the cells (29).
Not only do Ca 2ϩ sparks not activate L-type VDCCs at a detectable level, but also they suppress the generation of VDCC-mediated action potentials by external stimuli. As shown in this study, in the presence of Ca 2ϩ sparks and BiMPTs, ASM cells do not produce this form of action potential in response to depolarizing currents. Strikingly, when Ca 2ϩ sparks and their resulting BiMPTs were abolished, these cells generate action potentials upon stimulation with the same strength. The underlying mechanisms for this effect are to be determined. A likely possibility is that the currents resulting from the opening of BK channels and Cl Ca channels are much greater than that of L-type VDCCs, so the membrane potential is dominated by the activities of BK channels and Cl Ca channels. Both BK channels and Cl Ca channels in an ASM cell are in the range of 10,000 -20,000, which could give rise to peak currents on the order of 2-5 nA (13,30,31). But the peak current for L-type VDCCs is around 50 pA, which could be accounted for by the opening of ϳ500 channels (assuming a P o of 0.4 (supplemental Fig. S2) and a unitary conductance of 3 picosiemens (26)). When membrane is depolarized to the levels more negative than E Cl , BK channels would be the dominant force to opposing the depolarizing effect of L-type VDCCs; when the membrane potential becomes less negative than E Cl , both BK current and Cl Ca current act against the depolarization caused by L-type VDCCs, thus making the cells much harder to be depolarized. Because of the overwhelming effects of BK and Cl Ca channels on the membrane potential, Ca 2ϩ sparks serve as powerful safeguard devices to prevent hyper-excitability in ASM.
The inhibitory and stabilizing nature of Ca 2ϩ sparks and BiMPTs suggest that tipping the balance between BK channels and Cl Ca channels can change the excitability and contractility in ASM. This study demonstrated two such mechanisms as follows: one is the blockage of BK channels by paxilline, and the other is the blockage of Ca 2ϩ sparks by ryanodine. Interestingly, although the first mechanism directly inhibits BK channels and the second indirectly blocks both BK channels and Cl Ca channels by stopping Ca 2ϩ sparks, both actions result in the activation of Cl Ca channels and L-type VDCCs in a positive feedback manner, leading to membrane depolarization, global [Ca 2ϩ ] i elevation, and contraction of the cells. These results suggest that BK channels and Cl Ca channels activated by Ca 2ϩ sparks exert a dominant influence on membrane potential in these cells. This is in line with the previous findings that BK channels form clusters near Ca 2ϩ spark sites, and almost all Cl Ca channels in the membrane appear to concentrate in the areas Ca 2ϩ sparks occur (13,32). This is also supported by immunocytochemical studies revealing that BK channels localize in puncta in the surface membrane in several smooth muscle types, including ASM (33,34). 4 We propose a model for the role of Ca 2ϩ sparks and BiMPTs in ASM as follows (Fig. 7). At rest, at a given moment the majority of Ca 2ϩ spark sites are quiet, although a few generate Ca 2ϩ sparks. BK channels and Cl Ca in the quiescent sites open at low P o at normal resting [Ca 2ϩ ] i , contributing to the RMP. Those in the active sites open at high P o to generate BiMPTs, resulting in membrane potential oscillations. Because the RMP and membrane potential oscillations are below the activation potential of L-type Ca 2ϩ channels for most of the time, VDCCs are not activated or activated at an undetectable level. When BK channels are blocked by paxilline, or the balanced activity of BK channels and Cl Ca channels in Ca 2ϩ spark sites is disrupted by ryanodine, Ca 2ϩ sparks continue to activate or preferentially activate Cl Ca channels, leading to a stronger depolarization of the membrane. Such depolarization reaches the potential for the activation of VDCCs, resulting in their opening and Ca 2ϩ influx. Ca 2ϩ influx via VDCCs increases global [Ca 2ϩ ] i , which in turn activates more Cl Ca channels, the membrane becomes more depolarized, and more VDCCs are consequently activated, until a new equilibrium potential, i.e. near E Cl , is reached. Finally, the increase in global [Ca 2ϩ ] i causes cells to contract. It is likely that synchronizing activation of Ca 2ϩ sparks, BIMPTs, and VDCCs in ASM could influence the contractility at the tissue and organ level under physiological conditions, a possibility that warrants further investigation.
In summary, our study reveals that in ASM Ca 2ϩ sparks exert a bidirectional effect on membrane potential and can mediate both inhibitory and excitable responses. Therefore, Ca 2ϩ sparks and their evoked currents serve as a powerful mechanism that allows ASM to adapt to diverse internal and external stimuli. A consequence of this mechanism is that any changes in the composition of the Ca 2ϩ spark signaling complex could disrupt this plasticity, leading to an alteration in contractility with possible pathological consequences in ASM.