Properties of a Native Cation Channel Activated by Ca2+ Store Depletion in Vascular Smooth Muscle Cells*

Depletion of intracellular Ca2+ stores activates capacitative Ca2+influx in smooth muscle cells, but the native store-operated channels that mediate such influx remain unidentified. Recently we demonstrated that calcium influx factor produced by yeast and human platelets with depleted Ca2+ stores activates small conductance cation channels in excised membrane patches from vascular smooth muscle cells (SMC). Here we characterize these channels in intact cells and present evidence that they belong to the class of store-operated channels, which are activated upon passive depletion of Ca2+ stores. Application of thapsigargin (TG), an inhibitor of sarco-endoplasmic reticulum Ca2+ ATPase, to individual SMC activated single 3-pS cation channels in cell-attached membrane patches. Channels remained active when inside-out membrane patches were excised from the cells. Excision of membrane patches from resting SMC did not by itself activate the channels. Load-ing SMC with BAPTA (1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid), which slowly depletes Ca2+ stores without a rise in intracellular Ca2+, activated the same 3-pS channels in cell-attached membrane patches as well as whole cell nonselective cation currents in SMC. TG- and BAPTA-activated 3-pS channels were cation-selective but poorly discriminated among Ca2+, Sr2+, Ba2+, Na+, K+, and Cs+. Open channel probability did not change at negative membrane potentials but increased significantly at high positive potentials. Activation of 3-pS channels did not depend on intracellular Ca2+ concentration. Neither TG nor a variety of second messengers (including Ca2+, InsP3, InsP4, GTPγS, cyclic AMP, cyclic GMP, ATP, and ADP) activated 3-pS channels in inside-out membrane patches. Thus, 3-pS nonselective cation channels are present and activated by TG or BAPTA-induced depletion of intracellular Ca2+ stores in intact SMC. These native store-operated cation channels can account for capacitative Ca2+ influx in SMC and can play an important role in regulation of vascular tone.

Depletion of intracellular Ca 2؉ stores activates capacitative Ca 2؉ influx in smooth muscle cells, but the native storeoperated channels that mediate such influx remain unidentified. Recently we demonstrated that calcium influx factor produced by yeast and human platelets with depleted Ca 2؉ stores activates small conductance cation channels in excised membrane patches from vascular smooth muscle cells (SMC). Here we characterize these channels in intact cells and present evidence that they belong to the class of storeoperated channels, which are activated upon passive depletion of Ca 2؉ stores. Application of thapsigargin (TG), an inhibitor of sarco-endoplasmic reticulum Ca 2؉ ATPase, to individual SMC activated single 3-pS cation channels in cell-attached membrane patches. Channels remained active when inside-out membrane patches were excised from the cells. Excision of membrane patches from resting SMC did not by itself activate the channels. Load-ing SMC with BAPTA (1,2-bis(2-aminophenoxy)ethane-N,N,N,N-tetraacetic acid), which slowly depletes Ca 2؉ stores without a rise in intracellular Ca 2؉ , activated the same 3-pS channels in cell-attached membrane patches as well as whole cell nonselective cation currents in SMC. TG-and BAPTAactivated 3-pS channels were cation-selective but poorly discriminated among Ca 2؉ , Sr 2؉ , Ba 2؉ , Na ؉ , K ؉ , and Cs ؉ . Open channel probability did not change at negative membrane potentials but increased significantly at high positive potentials. Activation of 3-pS channels did not depend on intracellular Ca 2؉ concentration. Neither TG nor a variety of second messengers (including Ca 2؉ , InsP3, InsP4, GTP␥S, cyclic AMP, cyclic GMP, ATP, and ADP) activated 3-pS channels in inside-out membrane patches. Thus, 3-pS nonselective cation channels are present and activated by TG or BAPTA-induced depletion of intracellular Ca 2؉ stores in intact SMC. These native store-operated cation channels can account for capacitative Ca 2؉ influx in SMC and can play an important role in regulation of vascular tone.
Depletion of intracellular Ca 2ϩ stores is known to activate store-operated (or capacitative) Ca 2ϩ influx in a variety of nonexcitable cells (for review, see Refs. [1][2][3][4][5]. The idea of capacitative Ca 2ϩ entry (CCE) 1 was initially proposed for smooth muscle cells (SMC) (6). Over the past years it has been shown that CCE and contraction can be activated in SMC by passive depletion of intracellular Ca 2ϩ stores even without activation of receptor-dependent cascades (for review, see Ref. 7), although the nature of the ion channels responsible for CCE in vascular SMC remains obscure. In some nonexcitable cells, highly Ca 2ϩ -selective calcium release-activated calcium (CRAC) channels (4,8) and certain members of the diverse family of TRP channels (9,10) are thought to be responsible for CCE. However, to date the existence of neither of those has been established in SMC. Importantly, depletion of Ca 2ϩ stores was shown to trigger not only Ca 2ϩ , but also Na ϩ influx in arterial myocytes (11), which implies that store-operated channels in SMC are poorly selective for cations. In freshly isolated mouse anococcygeus SMC there are strong indications that CCE results from activation of a whole cell nonselective cation current (12,13), although in rat aortic SMC line A7r5 no currents were detected which could be associated with CCE (14,15). It is totally unclear if the same or different storeoperated channels mediate CCE in SMC from different preparations.
Here for the first time we characterize 3-pS cation channels that are activated by Ca 2ϩ store depletion in intact SMC from mouse and rabbit aorta. These channels, contrary to highly Ca 2ϩ -selective CRAC channels, are poorly selective for monoand divalent cations, and under physiological conditions they will allow both Ca 2ϩ and Na ϩ to enter SMC. Recently we found that these channels can also be activated in excised membrane patches by calcium influx factor (CIF) partially purified from human platelets or yeast with depleted Ca 2ϩ stores (16). Taken together, these data strongly support the idea that the native 3-pS channels, which we found in SMC, belong to the class of store-operated ion channels. Preliminary data have been reported in abstract form (17,18).

SMC Preparation
Four different preparations of aortic SMC were used in our experiments, and the 3-pS channel described in this paper was found to be the same in acutely dispersed and cultured SMC from mouse and rabbit aorta. Most of the experiments on characterization of the single channels and whole cell currents were done on mSMC in short term culture because they provided the most reliable model for studies of single channels, whole cell currents, and intracellular Ca 2ϩ .
killed by cervical dislocation. The thoracic aortas were rapidly removed, cleaned of connective tissues, and cut into small pieces.
Acutely Dissociated mSMC-Acutely dissociated mSMC were obtained as described previously (19). Briefly, pieces of aorta were incubated in dissociation medium (DM, in mM: 110 NaCl, 5 KCl, 10 NaHCO 3 , 0.5 NaH 2 PO 4 , 0.5 KH 2 PO 4 , 2 MgCl 2 , 10 taurine, 10 HEPES, 11 glucose, and 0.02% bovine serum albumin, pH 7.2) supplemented with 40 units/ml papain and 2 mM dithiothreitol for 15 min at 37°C with constant stirring. After incubation, pieces of aorta were rinsed twice in fresh DM and then gently triturated with a heat-polished Pasteur pipette. Isolated cells were stored at 4°C until use for up to 4 h. For experiments, 20 l of the cell suspension was placed in a 35-mm polystyrene tissue culture dish or a 0.15-mm glass-bottom chamber, and SMC were allowed to adhere to the bottom before the beginning of the experiment.
Mouse SMC in Short Term Primary Culture-Mouse SMC cultures were prepared in the following way. SMC were acutely dissociated by incubation of pieces of mouse aorta in 2 ml of Dulbecco's modified Eagle's medium containing 4 mg/ml collagenase, 1.5 mg/ml elastase, and 0.5 mg/ml trypsin inhibitor for 1 h at 37°C. During incubation, tissue was gently triturated every 15 min. Enzymatic digestion was terminated by the addition of 5 ml of fresh Dulbecco's modified Eagle's medium (free of enzymes) supplemented with 10% fetal bovine serum (FBS). The cell suspension was centrifuged, supernatant discarded, and the pellet was resuspended in 2 ml of Dulbecco's modified Eagle's medium containing 1% fetal bovine serum, 100 units/ml penicillin G, and 10 mg/ml streptomycin. The cell suspension was placed on coverslips and kept in 35-mm Petri dishes at 37°C in 5% CO 2 for 3-7 days. Under these conditions, mSMC attached to the coverslips but did not divide, had a bipolar morphology, and stained positively for ␣-actin.

Rabbit Aortic SMC (rSMC)
Male New Zealand White rabbits (2-2.5 kg) were exsanguinated after injection of 30 mg/kg sodium pentobarbital and 150 units/kg heparin. A segment of thoracic aorta was rapidly removed, cleaned of connective tissues, cut into small pieces, and rinsed in DM.
Acutely Dissociated rSMC-Acutely dissociated rSMC were obtained by incubation of the pieces of rabbit aorta in DM with 50 units/ml papain and 2 mM dithiothreitol for 30 min at 37°C in a shaking water bath as described previously (19). After incubation, they were rinsed twice in fresh DM and then gently triturated with a heat-polished Pasteur pipette. Isolated cells were stored at 4°C until use for up to 4 h.
Primary Cultured rSMC-Rabbit SMC cultures were prepared as described previously (20). Experiments on mice and rabbits were conducted in accordance with the Guide for the Care and Use of Laboratory Animals (57).

Tension Measurements
Isometric tension was measured in the rings of mouse thoracic aorta as described previously (20).

Ca 2ϩ Measurements
Cells were incubated with 2.5 M fura-2/AM ϩ 0.02% pluronic acid for 30 min at room temperature and subsequently washed for 15 min. Fluorescence (excitation at 340 and 380 nm, emission at 510 nm) was measured at 20 -22°C using an inverted microscope (Olympus IX70, Japan) equipped with a 20 ϫ fluorescence objective and dual excitation fluorescence imaging system (IonOptix Inc.). Changes in intracellular Ca 2ϩ concentration in mSMC were determined using standard methods as described earlier (19).

Electrophysiology
Single channel and whole cell currents were recorded as described recently (19) with a low noise Axopatch 200B amplifier (Axon Instruments). pCLAMP 6 software (Axon Instruments) was used for data acquisition and analysis. Data were filtered at 1 kHz and stored for later analysis. Representative traces of single channel currents were later filtered additionally at 100 -200 Hz for better visual resolution of 3-pS single channels on the figures. Experiments were conducted at 20 -22°C.
Single Channel Currents-Currents were recorded in the cell-attached and inside-out membrane patches. To improve the signal-tonoise ratio, pipettes were coated with Sylgard (Dow Corning Corp.) and polished to a resistance of 10 -20 megohms (when filled with high Na ϩ pipette solution). Single channel currents were recorded at Ϯ100, Ϯ80, and Ϯ 60 mV applied with respect to the cytosolic side of membrane.
Inward and outward currents are shown as downward and upward deflections, respectively, from the base line (labeled 0 on the figures). The amplitude of single channel currents was analyzed using all point histograms (see Fig. 2E) or amplitude histograms obtained from the event list (see Fig. 2C). Both methods gave the same single channel current amplitude for the first two current levels (Fig. 2, C and E), but we found the second method to be better for resolving infrequent channel openings in resting SMC (see Fig. 2B) or when more than two single channel current levels were observed at the peak of channel activity (see Fig. 2C). The open channel probability (NP o ) was analyzed and plotted over time to illustrate the time course of channel activity. The total apparent number of channels (N) in individual patches was estimated after their activation and was based on the maximum number of single channel current levels observed simultaneously at ϩ100 mV. In each membrane patch n varied from 3 to 5 (average N ϭ 4). Because prolonged full activation of all of the channels has not been achieved in some experiments, we cannot exclude the possibility that the actual number of the channels in each patch could be slightly higher than the apparent one. Standard bath solution contained (in mM) 140 NaCl, 2.8 KCl, 2 MgCl 2 , 5.5 glucose, 10 HEPES (pH 7.4). Standard pipette solution contained (in mM) 140 NaCl, 10 TEA, 0.2 EGTA, 10 HEPES (pH 7.4). In some experiments NaCl in the pipettes was replaced by KCl, CsCl, CaCl 2 , BaCl 2 , SrCl 2 , sodium glutamate, or NMDG-Cl (as specified in Table I). The liquid junction potential was compensated. In some experiments Ca 2ϩ (1 or 10 mM) was added to 140 mM NaCl-containing pipette or bath solutions. The pipette solutions also contained 100 M niflumic acid and 100 nM iberiotoxin to prevent activation of Cl Ca Ϫ and K Ca ϩ channels (the most abundant channels in aortic SMC), to ensure that they did not "contaminate" the recording of TG-and BAPTAactivated channels. Importantly, in control experiments 100 -200 M niflumic acid did not affect TG-induced Ca 2ϩ influx in mouse aortic SMC.
Because of the extremely small amplitude of single channel currents we were able to record and analyze them only at membrane potentials between Ϯ60 and Ϯ100 mV. For that reason it was impossible to determine experimentally and compare the exact reversal potential of single channel currents under different ion conditions and to provide the standard calculations of channel selectivity for different cations based on the shifts of the reversal potential of single channel current. Because of these technical limitations we only determined the relative cation conductivity of single channels under different ionic conditions (Table I), which was estimated from the slope conductance of inward or outward current in the presence of different cations. The slope conductance was calculated from the single channel currents at a minimum of three different membrane potentials.
Whole Cell Currents-Currents were recorded using conventional whole cell configuration. Bath solution contained (in mM) 130 NaCl, 2.
rent was monitored at holding potential of Ϫ50 mV with voltage ramp (from Ϫ100 to ϩ50 mV, 500 ms) applied every 2 s. Because whole cell currents showed a strong outward rectification at high positive potentials and because SMC loaded with BAPTA (in the absence of extracellular Ca 2ϩ ) were very fragile, whole cell current recordings at potentials higher than ϩ50 mV were generally avoided in our experiments. Leak current was not subtracted.

Drugs
Acetoxymethyl ester of BAPTA (BAPTA/AM) and fura 2-AM were from Molecular Probes Inc. (Eugene, OR). Papain was from Fluka (Bucks, Switzerland). All other drugs were from Sigma.

Statistics
The data are presented as means Ϯ S.E. with n showing the number of experiments. Statistical significance was assessed using the t test and analysis of variance. Values of p Ͻ 0.05 were considered significant.

Thapsigargin-induced Contraction and Intracellular Ca 2ϩ
Rise in Aortic SMC-5 M TG applied to the intact mouse aorta (Fig. 1A) caused a substantial contraction (610 Ϯ 22 mg), but only after the addition of 2 mM extracellular Ca 2ϩ , which is consistent with TG-induced activation of Ca 2ϩ influx. In individual SMC, TG applied in the absence of extracellular Ca 2ϩ (Fig. 1B) caused a small increase in [Ca 2ϩ ] i which can be explained by passive Ca 2ϩ release from intracellular stores leading to their depletion. Subsequent addition of 5 mM Ca 2ϩ caused a sustained rise in [Ca 2ϩ ] i , which reflects TG-induced activation of Ca 2ϩ influx (summary data are shown in Fig. 1C). Both Ca 2ϩ influx-dependent [Ca 2ϩ ] i rise and contraction were inhibited by 2-5 mM nickel, (Fig. 1).
Small Conductance Channels Activated by Thapsigargin in Aortic SMC-To define the nature of the ion channels that are responsible for TG-induced Ca 2ϩ influx, single channel currents were recorded in cell-attached membrane patches in mouse (mSMC) and rabbit (rSMC) aortic SMC. Extracellular application of 2 M TG activated small conductance channels in cell-attached membrane patches in mSMC and rSMC which were acutely dispersed or cultured. Fig. 2 shows a typical example (n ϭ 31 out of 52) of single channel activity and current amplitude in cell-attached membrane patches in mSMC (short term culture) before and after TG application with original traces of single channel outward currents (measured at ϩ100 mV applied with respect to the inside of mem-brane) at different times of the experiment. Only rare single channel openings could be detected in the resting cells. However, 40 -180 s after the application of TG, three to five single channel current levels were observed. Although the amplitude of single channel currents activated by TG was very small, it could be analyzed using either all point histograms (Fig. 2E) or event list histograms (Fig. 2C), which gave similar values of single channel current amplitude.
TG-induced activation of single channel currents was observed in control mSMC (n ϭ 10) and in mSMC with intracellular Ca 2ϩ buffered with 20 M BAPTA/AM for 20 min, (n ϭ 21). Importantly, in about 30% of SMC, BAPTA loading itself caused activation of the same channels as will be described below. TG-activated small single channel currents were observed in the absence or presence of 100 nM iberiotoxin and 100 M niflumic acid, inhibitors of K Ca ϩ and Cl Ca Ϫ channels, respectively, but no channel openings were recorded in cell-attached membrane patches when 5 mM Ni 2ϩ was present in the pipette (n ϭ 9).
It is important to emphasize that the same small conductance channels were activated by TG in acutely dispersed mSMC and rSMC as well as in rSMC in primary culture. As an example, Fig. 3A shows TG-activated inward single channel currents and their amplitude histogram in a cell-attached membrane patch in SMC acutely dispersed from rabbit aorta (n ϭ 5), which were indistinguishable from those found in mSMC (Fig. 2). Detailed characteristics of TG-activated channels in mSMC and rSMC are presented below.
Main Characteristics of TG-activated Channels in Inside-out Membrane Patches-After TG-induced activation in cell-attached patches, single channels remained active for 5-15 min even when the membrane patches were excised from SMC in Ca 2ϩ -free solution (Fig. 2D, n ϭ 4 out of 4). 2 mM LaCl 3 applied to the inside of membrane patches inhibited single channel currents (n ϭ 4). Importantly, excision of membrane patches from resting mSMC (not treated with TG or BAPTA/AM) in the presence or absence of 1 M or 1 mM CaCl 2 did not by itself activate the channels (n ϭ 54).
Single channel currents in inside-out membrane patches from TG-activated mSMC and rSMC (acutely dispersed or cultured) had identical properties, and Fig. 4 shows a typical example of single channel currents recorded in inside-out mem- brane patches with amplitude histograms at different membrane potentials. Current-voltage (I/V) relationships of single channels under different ionic conditions in inside-out membrane patches from rSMC and mSMC are shown in Fig. 3, B and C, respectively (acutely dissociated cells) and Fig. 5, A and B (cultured cells). The slope conductance in symmetrical 140 mM NaCl was 3.3 Ϯ 0.1 pS (n ϭ 4) in acutely dissociated and 3.4 Ϯ 0.2 pS (n ϭ 6) in cultured mSMC, which was similar to acutely dissociated and cultured rSMC (3.3 Ϯ 0.1 pS, n ϭ 8 and 3.2 Ϯ 0.1 pS, n ϭ 9, respectively). When Na ϩ in the pipette was replaced by 100 mM NMDG-Cl, inward single channel currents (at Ϫ100 mV) disappeared, although outward currents recorded at ϩ100 mV in the same membrane patches did not change (n ϭ 3, Fig. 5A). Replacement of Cl Ϫ with glutamate in the pipette did not affect single channel currents at both positive and negative membrane potentials (3.4 Ϯ 0.6 pS, n ϭ 3, Fig. 5A), confirming that the 3-pS channel is cation-selective. Fig. 5A shows that the conductance of 3-pS channels did not change significantly when Ca 2ϩ was added to Na ϩ -containing pipette solution at concentration of 1 mM (3.3 Ϯ 0.7 pS, n ϭ 16) or 10 mM Ca 2ϩ (2.9 Ϯ 0.8 pS, n ϭ 3), which showed that, contrary to Ca 2ϩ -selective channels, the 3-pS channel does not prefer Ca 2ϩ over Na ϩ when both cations are present. When Na ϩ in the pipette was replaced by K ϩ or Cs ϩ , the single channel inward current had similar slope conductance of 3.4 Ϯ 0.1 pS (n ϭ 3) and 3.4 Ϯ 0.2 pS (n ϭ 4), respectively, which was similar to that in Na ϩ . To test if 3-pS channels along with monovalent cations also conduct divalent cations, Na ϩ in the pipette and/or in the bath was replaced by Ca 2ϩ , Sr 2ϩ , or Ba 2ϩ (Fig. 5, A and B). When 90 mM Ca 2ϩ was used in the pipette and 90 mM Ba 2ϩ was used in the bath (with no Na ϩ present on  either side), both inward and outward currents were observed, and the slope conductance of inward (Ca 2ϩ ) current was 2.7 Ϯ 0.1 pS, and the outward (Ba 2ϩ ) current was 3.5 Ϯ 0.1 pS (Fig.  5B, n ϭ 8). When 90 mM Ca 2ϩ or Sr 2ϩ was in the pipette and 140 mM Na ϩ was in the bath, the slope conductance of inward current was 3.0 Ϯ 0.6 pS (for Ca 2ϩ , n ϭ 3) and 2.7 Ϯ 0.1 pS (for Sr 2ϩ , n ϭ 9), with no apparent change in slope conductance of outward (Na ϩ ) current. Working at the limits of resolution of single channel currents, we were not able to determine the exact reversal potential of single channel current in mixed solutions which is required for estimation of exact channel selectivity to different cations. Minor changes in conductivity of the channel in the presence of different cations (summarized in Table I) would suggest that these TG-activated 3-pS channels are cation-selective but discriminate poorly between different mono-and divalent cations, and thus under physiological conditions would allow both Ca 2ϩ and Na ϩ to enter SMC.
In inside-out membrane patches, the NP o of single channels in both mSMC (Fig. 5C) and rSMC (Fig. 5D) was similar and did not change much at negative membrane potentials (NP o Ϸ 0.2 at-60 mV) but significantly increased at high positive membrane potentials. Although only one level of single channel currents was usually seen at Ϫ100 mV, activation of three to five levels (average four, n ϭ 28) could be detected at ϩ100 mV (Fig. 4). Thus, in the physiological range of membrane potentials the P o of each individual channel activated by TG in SMC could be as low as 0.05. Thus, the same channels were activated by TG in acutely dissociated or cultured SMC from mouse and rabbit aorta, which appeared to be nonselective cation channels with about a 3-pS conductance.
Loading SMC with BAPTA Activated 3-pS Nonselective Cation Channels and Whole Cell Currents-BAPTA, a high affinity Ca 2ϩ chelator, activates CRAC currents in a variety of nonexcitable cells and is commonly used to deplete intracellular Ca 2ϩ stores without an increase in [Ca 2ϩ ] i or activation of receptordependent pathways (for review, see Refs. 4 and 21). Dialysis of mSMC with 10 mM BAPTA in the absence of extracellular Ca 2ϩ slowly activated an inward current (at Ϫ50 mV) which usually started after 40 -60 s and reached a stable plateau of 39 Ϯ 13 pA (n ϭ 6) within 2-3 min (Fig. 6, A and B). The current was inhibited by 2 mM LaCl 3 , n ϭ 5). The I/V relationship of the BAPTA-induced whole cell current (Fig. 6C) reversed at 2.5 Ϯ 1.4 mV and showed strong outward rectification, especially at high positive potentials (not shown).
When mSMC were loaded with 20 M BAPTA/AM for 20 min, small conductance channels were activated in 25 out of 72 cells, which were the same as 3-pS channels activated by TG. Fig. 6F shows an example of single channel currents and their amplitude histograms at different membrane potentials in a membrane patch excised from a BAPTA-loaded mSMC. The I/V relationship (Fig. 6E) and the increase in NP o at high positive membrane potentials (Fig. 6D) of BAPTA-activated channels were similar to the 3-pS channels activated by TG in mSMC and rSMC (see Fig. 5). In mSMC not loaded with BAPTA, spontaneous activation of 3-pS channels was observed only in 3 out of 26 cells.
Because BAPTA-activated 3-pS channels and whole cell currents were both poorly cation-selective and both showed significant outward rectification at high positive membrane potentials, we determined whether activation of 3-pS channels could indeed produce the whole cell current with characteristics of that experimentally observed in BAPTA-loaded mSMC. The amplitudes of the hypothetical whole cell current were calculated from the I/V relationship (Fig. 6D) and NP o (Fig. 6E) of single channel currents (assuming that about 1,000 membrane patches could comprise about a 2,000-m 2 surface of mSMC). As predicted, the calculated current amplitudes (Fig. 6C, open circles) were close to the amplitude of the actual whole cell current (Fig. 6C, closed diamonds) recorded in mSMC loaded with BAPTA. This result strongly suggests that 3-pS nonselective cation channels are responsible for the whole cell current activated by BAPTA in mSMC. DISCUSSION The present study provides the first insight into the nature of single ion channels that are responsible for store-operated Ca 2ϩ influx in vascular SMC. We found Ca 2ϩ -conducting nonselective cation channels of very small (3-pS) conductance which are activated after loading mSMC with BAPTA and/or treatment with TG, which are commonly used to passively deplete intracellular Ca 2ϩ stores. The very small amplitude of single channel currents activated by TG and BAPTA in mouse and rabbit aortic SMC might explain why these channels have not been described previously. Indeed, many different types of nonselective cation channels have been described in SMC (15,(22)(23)(24)(25)(26)(27)(28)(29)(30)(31)(32)(33)(34)(35)(36)(37), but the single channel properties and mechanism of activation of the 3-pS channel distinguish it from all of the channels found in SMC so far (for review, see Refs. 5, 38, and 39). Our results strongly support the possibility that novel 3-pS channels are store-operated and are responsible for TG-induced Ca 2ϩ influx in individual SMC and contraction of the aorta.
Characteristics of the 3-pS Channel in SMC-Even though the amplitude of single channel currents did not exceed 0.4 pA even at Ϯ 100 mV, the low noise patch-clamp system allowed us to resolve clearly single channel openings and to perform some analysis of their amplitude and open channel probability. TGand BAPTA-activated channels appeared to be nonselective cation channels that conducted and poorly discriminated between monovalent (Na ϩ , K ϩ , Cs ϩ ) and divalent (Ca 2ϩ , Sr 2ϩ , Ba 2ϩ ) cations. Their single channel conductance was close to 3 pS for a variety of cations tested. As a result of activation of such nonselective cation channels both Ca 2ϩ and Na ϩ are expected to enter SMC causing not only a rise in intracellular Ca 2ϩ , but also significant influx of Na ϩ . Indeed, depletion of Ca 2ϩ stores has recently been shown to activate both Ca 2ϩ and Na ϩ influx in arterial myocytes (11). Influx of Na ϩ into SMC could result in membrane depolarization that could potentially trigger activation of voltage-dependent Ca 2ϩ -selective channels that might contribute to additional Ca 2ϩ influx and SMC contraction. Single channel recording during SMC activation with TG provides some indirect evidence of TG-induced SMC depolarization. Indeed, rare openings of single channels in resting SMC result in outward currents of about 0.22 pA at ϩ100 mV applied in respect to the cytoplasmic side of cellattached membrane patches (Fig. 2, A and B). Simple calculations show that this amplitude is about what one would expect from a 3.4-pS channel if the resting membrane potential is around Ϫ40 mV (indeed we found the resting membrane potential of BAPTA-loaded SMC to be Ϫ39 Ϯ 2 mV). 2 After robust activation of the channels by TG, their current amplitude increased to about 0.31 pA (Fig. 2, A and C) which is what could be expected from a 3.4-pS channel if the membrane potential of SMC at that point is about Ϫ10 mV. Thus, the increase in the amplitude of single channel outward currents in cell-attached membrane patches is consistent with significant TG-induced depolarization of SMC. Importantly, 3-pS channels were found to be present and open rarely in resting SMC. The activation of the channels was observed within 1-3 min after the application of TG (Fig. 2), which is within the timeframe of the TG-induced onset of Ca 2ϩ influx in SMC and other cells (20,40). The NP o did not change significantly within the physiological range of membrane potentials, but it increased dramatically at high positive potentials (Fig. 5, C and D). This increase of NP o was very typical for 3-pS channels in both cell-attached and insideout membrane patches, but the mechanism underlying this phenomenon is unclear.
3-pS Nonselective Cation Channels Could Underlie BAPTAactivated Whole Cell Current-A critical "trademark" feature of store-operated channels is that they can be activated independently of intracellular Ca 2ϩ rise and major signaling cascades. Such conditions are achieved by cell dialysis with 10 -20 mM BAPTA. Indeed, BAPTA provides a fast and strong buffering of intracellular Ca 2ϩ , preventing Ca 2ϩ stores from refilling and promoting their passive depletion. Extracellular Ca 2ϩ must also be eliminated to prevent saturation of BAPTA by Ca 2ϩ entering the cell. Activation of the whole cell current under these conditions is thought to be one of the strongest pieces of evidence for the existence of store-operated channels in different types of cells (4).
Dialysis of mSMC with 10 mM BAPTA in our experiments activated a poorly selective whole cell current with a reversal potential around 0 mV (please note that Cl Ϫ equilibrium potential under our experimental conditions was Ϫ31 mV). The time course of the development of this current was similar to that observed for I CRAC in nonexcitable cells, but unlike I CRAC , the current showed strong outward, rather than inward rectification, under physiological ionic conditions (when both Ca 2ϩ and Na ϩ were present in the bath). Poor cation selectivity and pronounced outward rectification strongly resemble that of the 3-pS channels, which were also activated in mSMC as a result of BAPTA/AM loading (Fig. 6). Moreover, the I/V relationship of the inward and outward whole cell currents simulated from single 3-pS channel currents (using their NP o and I/V relationship) is similar to the BAPTA-activated whole cell current recorded in mSMC.
Thus, the BAPTA-activated current is likely to result from activation of the 3-pS nonselective cation channels found in the same mSMC. We estimated that about 5,000 channels must be present in mSMC to account for the BAPTA-activated whole cell current. Indeed, at 50 mV the average whole cell inward current (I) of about 40 pA develops after 3 min in BAPTAloaded mSMC (Fig. 6B). Assuming that the single channel current at Ϫ50 mV is ␥ ϭ 0.16 pA, and each channel is opened with P o ϭ 0.05, the minimum number of the channels which could produce the whole cell current of 40 pA will be n ϭ I/␥ P o ϭ 5,000. If the channels are homogeneously distributed in the plasma membrane, and an average mSMC has a capacitance of 20 picofarads and surface of about 2,000 m 2 , then two to three channels are expected to be present in every 1 m 2 of plasma membrane. This is in close agreement with an average of four channels which we observed in 1-2-m 2 membrane patches. Thus, 3-pS channels are likely to be responsible for the whole cell current activated during mSMC cell dialysis with BAPTA.
Native 3-pS Channels Are Not Regulated by Intracellular Ca 2ϩ -Application of TG to intact SMC causes a pronounced increase in [Ca 2ϩ ] i , and some of the channels could be activated by TG-induced intracellular Ca 2ϩ rise rather than by depletion of Ca 2ϩ stores. Several lines of evidence obtained in our experiments exclude this possibility. First, 3-pS channels were not activated directly by Ca 2ϩ when inside-out membrane patches were excised into Ca 2ϩ -containing solutions. Two different concentrations of Ca 2ϩ in the bath were used in these experiments (1 M and 1 mM) to ensure that the possibility of a bell-shaped dependence on Ca 2ϩ concentration was not overlooked. For comparison, in our recent studies (19), we demonstrated that excision of membrane patches from the same cells into Ca 2ϩcontaining solution activated Ca 2ϩ -dependent Cl Ϫ channels (in the absence of niflumic acid). These results show that contrary to Cl Ca Ϫ channels, 3-pS cation channels are not activated by Ca 2ϩ . Second, TG-induced activation of the same 3-pS cation channels was observed in both control SMC and SMC in which [Ca 2ϩ ] i was buffered with BAPTA/AM. There is always a concern that loading the cells with BAPTA/AM may not be enough to buffer Ca 2ϩ completely in certain regions of subplasmalemmal space with restricted diffusion. This problem is impossible to rule out entirely. However, we showed recently (19) that identical loading of mouse aortic SMC with BAPTA/AM completely prevents the caffeine-induced global intracellular Ca 2ϩ rise as well as activation of Cl Ca Ϫ channels, which are known to be highly sensitive and reliable sensors of Ca 2ϩ rise occurring beneath the plasma membrane. Thus, it is very unlikely that TG causes a significant Ca 2ϩ rise in BAPTA/AMloaded mSMC which could affect 3-pS channels. Third, depletion of intracellular stores with BAPTA (in the absence of extracellular Ca 2ϩ ) does not increase intracellular Ca 2ϩ , although it does activate single 3-pS nonselective cation channels and whole cell currents. Thus, it seems unlikely that Ca 2ϩ could be a natural activator of 3-pS channels.
3-pS Nonselective Cation Channels Are Likely to Be Native Store-operated Channels in SMC-The results of our studies provide several lines of evidence that in intact SMC 3-pS channels are activated by depletion of intracellular Ca 2ϩ stores rather than by other mechanisms. First, we found 3-pS channels to be activated by TG and BAPTA, which are known to cause passive depletion of Ca 2ϩ stores and activation of storeoperated channels and whole cell currents in a variety of nonexcitable cells. Second, activation of 3-pS channels and corresponding whole cell currents do not require intracellular Ca 2ϩ rise or activation of InsP 3 , or other receptor-dependent cascades and could be achieved by simple loading of SMC with the Ca 2ϩ chelator BAPTA. Third, a variety of second messengers that are known to be involved in Ca 2ϩ homeostasis of SMC (Ca 2ϩ , InsP 3 , InsP 4 , GTP␥S, cAMP, cGMP, ATP, ADP) did not activate single 3-pS channels in inside-out membrane patches, excluding their role as physiological activators of these channels.
Recently, we found that 3-pS nonselective cation channels in inside-out membrane patches from SMC are activated by CIF partially purified or bioassayed from yeast or human platelets with depleted Ca 2ϩ stores (16). This CIF is thought to be one of the possible messengers produced by endoplasmic reticulum upon depletion of Ca 2ϩ stores which can activate plasma membrane channels and cause CCE. These data strongly support the idea that 3-pS nonselective cation channels (which we showed here to be activated in intact SMC upon depletion of their stores) indeed belong to the class of store-operated channels.
Although their electrophysiological characteristics are significantly different, activation properties of 3-pS cation channels and whole cell currents resemble those of other storeoperated channels and currents in nonexcitable cells. The onset of TG-induced activation of 3-pS channels falls within the time frame of passive depletion of Ca 2ϩ stores when sarco-endoplasmic reticulum-dependent back-sequestration of Ca 2ϩ is inhibited. After being activated, 3-pS channels remain active for several minutes even after inside-out patches are excised from SMC, which is similar to the store-operated currents recorded in giant inside-out membrane patches from Xenopus oocytes (41). BAPTA loading of SMC also activates the whole cell nonselective cation current with a time course similar to that of activation of I CRAC . However, the electrophysiological characteristics of the currents in SMC are different from that known for CRAC channels (for review, see Ref. 4). For example, I CRAC in nonexcitable cells is highly Ca 2ϩ -selective and conducts Na ϩ only in the absence of Ca 2ϩ , whereas BAPTA and TG-activated single channels in SMC are poorly cation-selective, and their Na ϩ conductance does not change significantly in the presence or absence of Ca 2ϩ . Also, in the presence of Ca 2ϩ on one side of the membrane and Na ϩ on the other, the I/V relationship of 3-pS channels is linear with approximated reversal potential around zero mV. In the presence of Na ϩ and Ca 2ϩ the whole cell current in SMC reverses at 0 mV and has significant outward rectification, whereas I CRAC reverses at high positive potentials (around the Ca 2ϩ equilibrium potential) and has a very pronounced inward rectification. The whole cell currents and single 3-pS channels could be inhibited by millimolar concentrations of nickel or lanthanum (above 2 mM), whereas I CRAC is inhibited by lanthanum in the micromolar range. Finally, under the same ionic conditions (in the presence of Na ϩ and Ca 2ϩ ) the conductance of CRAC channels was estimated to be about 9 fS (42), which is 300 times less than the 3-pS conductance we observed. Thus, selectivity and conductance of 3-pS channels in SMC are clearly different from CRAC channels found in nonexcitable cells, but that does not exclude the possibility that both channels could be similarly regulated by the filling state of intracellular Ca 2ϩ stores.
The poor cation selectivity of TG-and BAPTA-activated channels in SMC resembles that of the TRP1 channel expressed in some mammalian cells (43)(44)(45), although its conductance was estimated to be around 16 pS. The possibility that 3-pS nonselective cation channels in SMC could be related to the growing family of TRP channels needs further investigation.
Physiological Relevance of Store-operated Cation Channels in Vascular SMC-The existence of store-operated Ca 2ϩ influx in vascular SMC is strongly supported by many studies that demonstrated Ca 2ϩ influx in SMC and contractions in different vessels caused by sarco-endoplasmic reticulum Ca 2ϩ ATPase inhibitors (12, 15, 35, 46 -56). Recently, we demonstrated that this pathway in vascular SMC is regulated by nitric oxide, the major endothelium-derived relaxing factor, which indirectly inhibits CCE by enhancing sarco-endoplasmic reticulum Ca 2ϩ ATPase-dependent refilling of Ca 2ϩ stores (20). This implies the physiological importance of CCE in regulation of vascular tone, but the nature of the channels involved in this storeregulated process remains obscure. Indeed, studies in a fetal rat aortic cell line (A7r5) failed to reveal single channels or cation currents activated by TG (14,15), although in freshly isolated mouse anococcygeus SMC a whole cell nonselective cation current was described which correlated with capacitative Ca 2ϩ influx (12,13).
Our studies for the first time demonstrate the presence of small conductance (3-pS) nonselective cation channels in intact SMC from mouse and rabbit aorta which are activated by depletion of intracellular Ca 2ϩ stores (with BAPTA or TG). Importantly, we found and described these channels in both acutely dispersed and cultured SMC. These new data, together with demonstration of TG-induced Ca 2ϩ influx in isolated mSMC (Fig. 1), TG-induced Ca 2ϩ influx in intact smooth muscle strips of mouse aorta (20), and TG-induced contraction of the mouse aorta (Fig. 1), strongly suggest that TG-activated 3-pS cation channels could be responsible for store-operated Ca 2ϩ influx and contraction of aortic SMC.
At the end, it is important to mention that although we found 3-pS channels in both freshly dispersed and cultured SMC, it was easier to record and study them in cultured SMC. Also, SMC in culture showed more consistent responses to TG (judged by TG-induced Ca 2ϩ influx), which was observed in 80 -90% of cells, whereas such responses in acutely dissociated cells varied from isolation to isolation, and on average TGinduced Ca 2ϩ influx was observed in only about 10% of freshly dissociated SMC. Such differences in the number of cells responding to TG most probably reflect the result of enzymatic shock which is unavoidable during the acute isolation of fresh SMC but can be reduced if SMC are allowed to recover for a few days in short term culture. Culture of the cells does not apparently change the electrophysiological properties of these 3-pS channels. This simple explanation is also supported by the fact that TG-induced Ca 2ϩ influx is present in intact smooth muscle strips of mouse aorta (20) before their exposure to enzymes. Interestingly, the same methods of acute isolation of SMC from mouse and rabbit aorta did not affect SMC responsiveness to caffeine which causes Ca 2ϩ release from the stores and activation of Ca 2ϩ -dependent Cl Ϫ channels (19) in 90% of freshly dispersed cells. The reasons for the relative partial impairment of capacitative Ca 2ϩ influx mechanism immediately after enzymatic treatment of SMC need further investigation.