INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

The release and sequestration of Ca²⁺ ions by an intracellular membrane compartment, the sarcoplasmic reticulum (SR), is essential to the process of cardiac muscle contraction and relaxation. In cardiac muscle, the influx of Ca²⁺ via a voltage-sensitive dihydropyridine receptor (DHPR)/Ca²⁺ channel (L-type) triggers the massive release of Ca²⁺ by opening SR Ca²⁺ release channels (CRCs) (for review see Wier, 1990). The CRC binds the plant alkaloid ryanodine with high affinity and specificity and hence is also known as the ryanodine receptor (for reviews see Franzini-Armstrong and Protasi, 1997; Sutko et al., 1997; Meissner, 1994). CRCs are ligand-gated channels with Ca²⁺ as a major regulator. High-affinity activating and low-affinity inactivating Ca²⁺ binding sites have been identified (Liu et al., 1998; Fruen et al., 1996; Xu et al., 1996; Laver et al., 1995; Chu et al., 1993; Zimanyi and Pessah, 1991; Meissner and Henderson, 1987). Rapid activation and inactivation by cytosolic Ca²⁺ has suggested that these sites are located on the large cytosolic foot region of the channels (Laver and Curtis, 1996; Schiefer et al., 1995; Sitsapesan et al., 1995; Gyorke and Fill, 1993). Various other endogenous effectors of CRCs have been identified including Mg²⁺, ATP, and calmodulin (Meissner, 1994).

In addition to cytosolic Ca²⁺, SR lumenal Ca²⁺ may affect CRC activity. The most direct evidence for a regulation by SR lumenal Ca²⁺ has been obtained in single channel measurements using the planar lipid bilayer technique. SR lumenal Ca²⁺ activated the skeletal muscle CRC in the presence of cytosolic ATP (Sitsapesan and Williams, 1995; Tripathy and Meissner, 1996) but no or only a modest activation was observed in the absence of ATP (Sitsapesan and Williams, 1995; Tripathy and Meissner, 1996; Herrmann-Frank and Lehmann-Horn, 1996). These results have raised the interesting possibility that skeletal CRCs have SR intralumenal Ca²⁺ binding sites that interact with cytosolic regulatory sites (Sitsapesan and Williams, 1995). An alternative suggestion has been that SR lumenal Ca²⁺ flowing through the channel regulates the skeletal muscle CRC by having access to cytosolic activation and inactivation sites (Tripathy and Meissner, 1996; Herrmann-Frank and Lehmann-Horn, 1996). In support of the latter suggestion, high concentrations of the "fast" Ca²⁺ buffer 1,2-bis(2-aminophenoxy)ethanetetraacetic acid (BAPTA) increased cytosolic ATP-activated, lumenal Ca²⁺-activated skeletal muscle channel activities. This result suggested that lumenal Ca²⁺ passing through the skeletal CRC regulates the channel by having access to "BAPTA-inaccessible" Ca²⁺ activation and "BAPTA-accessible" Ca²⁺ inactivation sites (Tripathy and Meissner, 1996).

An increase in lumenal Ca²⁺ concentration also resulted in an increase in cardiac CRC open probability. The presence of another cytosolic channel activator such as sulmazole (Sitsapesan and Williams, 1994a) or ATP (Lukyanenko et al., 1996) was required to observe activation by lumenal Ca²⁺. These results were considered to be inconsistent with the idea that lumenal Ca²⁺ ions flowing through the channel have direct access to cytosolic Ca²⁺ activation sites.

The cardiac CRC represents a classical example of a Ca²⁺-regulated Ca²⁺ release mechanism (Wier, 1990). Its regulation by Ca²⁺ and other endogenous effectors differs from that of the skeletal CRC (Franzini-Armstrong and Protasi, 1997; Sutko et al., 1997; Meissner, 1994). It is therefore conceivable that the two channel isoforms are regulated differently by lumenal Ca²⁺. To clarify the ways in which lumenal Ca²⁺ ions regulate the cardiac CRC, we have investigated their effects on single canine cardiac muscle CRCs, using the planar lipid bilayer method. Our results indicate that lumenal Ca²⁺ flowing through the channel regulates the cardiac Ca²⁺ release channel via direct feedback by binding to cytosolic Ca²⁺ activation and inactivation sites. An activation of channel activity by lumenal Ca²⁺ was observed at Mg²⁺ and ATP concentrations corresponding to those in myocardium. These results suggest that activation of cardiac CRCs by lumenal Ca²⁺ fluxes may be a physiologically relevant mechanism. A preliminary report of this work has been presented in abstract form (Xu and Meissner, 1997).

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Materials

Phospholipids were obtained from Avanti Polar Lipids (Birmingham, AL). All other chemicals were of analytical grade.

Preparation of sarcoplasmic reticulum vesicles and purification of Ca²⁺ release channels

Canine cardiac SR vesicle fractions enriched in [³H]ryanodine binding and Ca²⁺ release channel activities were prepared in the presence of protease inhibitors as described (Xu et al., 1993). The CHAPS (3-[(3-cholamidopropyl)dimethyl-ammonio]-1-propanesulfonate)-solubilized canine heart 30S Ca²⁺ release channel complex was isolated by rate density gradient centrifugation and reconstituted into proteoliposomes by removal of CHAPS by dialysis (Lee et al., 1994).

Single channel measurements

Single channel measurements were performed by fusing proteoliposomes containing the purified cardiac muscle Ca²⁺ release channel with Mueller-Rudin-type bilayers containing phosphatidylethanolamine, phosphatidylserine, and phosphatidylcholine in the ratio 5:3:2 (25 mg total phospholipid/ml n-decane) (Lee et al., 1994). The side of the bilayer to which the proteoliposomes were added was defined as the cis side. A strong dependence of channel activity on micromolar cis Ca²⁺ concentrations suggested that the cis side corresponded to the SR cytosolic side in a majority (>98%) of our recordings. The trans side of the bilayer was defined as ground. Single channels were recorded in a symmetrical KCl buffer solution (0.25 M KCl, 20 mM KHepes, pH 7.4) containing the additions indicated in the text. Electrical signals were filtered at 2 kHz, digitized at 10 kHz, and analyzed. Data acquisition and analysis were performed with a commercially available software package (pClamp 6.0.3., Axon Instruments, Burlingame, CA) using an IBM-compatible Pentium computer and 12-bit A/D-D/A converter (Digidata 1200, Axon Instruments) (Xu et al., 1996).

Determination of free Ca²⁺ concentrations

Free Ca²⁺ concentrations of >1 µM were determined with a Ca²⁺-selective electrode (World Precision Instruments, Inc., Sarasota, FL). Free Ca²⁺ concentrations of <1 µM were obtained by including in the solutions the appropriate amounts of Ca²⁺ and EGTA as determined using the stability constants and computer program published by Schoenmakers et al. (1992).

Statistics

Results are given as means ± SE. Significance of differences of data was analyzed with Student's t-test. Differences were regarded to be statistically significant at P < 0.05.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

Purified cardiac Ca²⁺ release channels reconstituted into proteoliposomes were incorporated into planar lipid bilayers and recorded in symmetrical 0.25 M KCl buffer. The use of K⁺ rather than Ca²⁺ as a current carrier avoided the buildup of a large Ca²⁺ gradient near the mouth of the channel, thus simplifying analysis of regulation of the cardiac CRC by Ca²⁺. Single channel conductance with 0.25 M K⁺ as current carrier was 770 pS (Xu et al., 1993). The effects of cytosolic and lumenal Ca²⁺ on channel activity were examined in the presence and absence of 10 mM cytosolic caffeine. Caffeine increases the apparent Ca²⁺ affinity of the Ca²⁺ activation sites (Liu et al., 1998; Zucchi and Ronca-Testoni, 1997), which allows the use of low cytosolic Ca²⁺ concentrations in testing the effects of lumenal Ca²⁺. Channels were also recorded in the presence of 5 mM cytosolic MgATP (0.7 mM free Mg²⁺) to better simulate the intracellular conditions in myocardium.

Regulation of cardiac Ca²⁺ release channels by cytosolic and lumenal Ca²⁺ in the presence of 10 mM caffeine

In Fig. 1 A, a single cardiac CRC was recorded in the presence of 10 mM cytosolic (cis) caffeine at three different cytosolic Ca²⁺ concentrations and holding potentials of 35 and +35 mV. Short, often not fully resolved channel events were observed with 0.1 µM free Ca²⁺ in the cytosolic bilayer chamber (Fig. 1 A, top traces). Elevation of cytosolic Ca²⁺ concentration to 1 µM increased channel open probability (P_o) (middle traces) at both holding potentials. In the presence of 10 µM cytosolic Ca²⁺, long open events interrupted by brief closings were observed at both holding potentials, resulting in a nearly fully activated channel (bottom traces).

View larger version (32K): [in this window] [in a new window]   FIGURE 1   Dependence of single channel activities on cytosolic [Ca2+] in the presence of 10 mM caffeine. (A) Single channel currents were recorded at 35 mV (downward deflections, left current traces) (c = closed) and +35 mV (upward deflections, right current traces) in symmetrical 0.25 M KCl, 20 mM KHepes, pH 7.4 media containing 10 mM cytosolic caffeine, 200 µM EGTA, and [Ca2+] to yield the indicated free cytosolic [Ca2+]. The trans (SR lumenal) solution contained <2 µM Ca2+. (B) Channel open probabilities (Po) were determined at 35 mV () and +35 mV () as in (A). Values are the mean ± SE of 5-12 experiments. Continuous lines were obtained assuming that the CRC possesses cooperatively interacting high-affinity Ca2+ activation and low-affinity Ca2+-inactivation sites (Scheme 1 and Eq. 1 of Liu et al., 1998). At 35 mV, Hill constants and coefficients were KHa = 1 µM, nHa = 4.5, KHi = 10 mM, and nHi = 1.6.

Fig. 1 B shows that channels in the presence of 10 mM cytosolic caffeine were half-maximally activated at ~1 µM cytosolic Ca²⁺, and half-maximally inhibited at 10 mM cytosolic Ca²⁺. In agreement with previous studies (Liu et al., 1998; Fruen et al., 1996; Xu et al., 1996; Laver et al., 1995; Zimanyi and Pessah, 1991; Meissner and Henderson, 1987), data of Fig. 1 B suggest that the cardiac CRC has both high-affinity Ca²⁺ activation and low-affinity Ca²⁺ inactivation sites. Furthermore, Fig. 1 shows that the cardiac CRC exhibits no significant voltage dependence when activated and inactivated by cytosolic Ca²⁺ in the presence of caffeine.

The CRC showed a strong voltage dependence when the lumenal instead of cytosolic Ca²⁺ concentration was elevated. In Fig. 2 A (top traces), a single cardiac CRC was initially recorded under conditions similar to those in Fig. 1 A (top traces), i.e., at a low cytosolic [Ca²⁺] (<0.01 µM) in the presence of 10 mM cytosolic caffeine. The lumenal Ca²⁺ concentration was <2 µM and the holding potentials were 50 mV and +50 mV. As in Fig. 1, brief, often not fully resolved channel events were observed at both holding potentials. An increase of lumenal Ca²⁺ concentration from <2 µM to 1 mM increased P_o >100-fold at 50 mV, but was essentially without an effect at +50 mV (middle traces).

View larger version (17K): [in this window] [in a new window]   FIGURE 2   Activation of the cardiac Ca2+ release channel by lumenal Ca2+ in the presence of 10 mM caffeine. (A) Single channel currents were recorded at 50 mV (downward deflections, left panels) and +50 mV (upward deflections, right panels) in symmetrical 0.25 M KCl, 20 mM KHepes, pH 7.4 media containing <0.01 µM free cytosolic Ca2+ (200 µM EGTA and <2 µM contaminating Ca2+) and 10 mM cytosolic caffeine. Bottom traces were obtained after the addition of 20 mM cytosolic BAPTA. SR lumenal [Ca2+] was <2 µM (top traces) and 1 mM (middle and bottom traces). Note: Negative holding potentials favor lumenal-to-cytosolic Ca2+ fluxes. (B) Dependence of Po on cytosolic and lumenal [Ca2+]. Holding potentials were 50 mV (, ) and +50 mV (, ). (C) Dependence of Po on lumenal-to-cytosolic Ca2+ fluxes. Lumenal-tocytosolic Ca2+ fluxes were calculated according to the barrier model and parameters of Tinker et al. (1992) at <0.01 µM cytosolic (closed symbols) and 4 µM cytosolic (open symbols) Ca2+ in the presence of <2 µM lumenal Ca2+ (, masked by the other symbols at the origin, ) at ±35 and ±50 mV), 1 mM lumenal Ca2+ at +65, +50, +35, +20, 20, 35-50 and 65 mV (, from left to right) and +50, +35, 35, and 50 mV (, from left to right), 3 mM lumenal Ca2+ () at 0 mV, 5 mM lumenal Ca2+ (, ) at the membrane potentials indicated for 1 mM lumenal Ca2+ except that the effects of 5 mM lumenal Ca2+ at 4 µM cytosolic Ca2+ were also determined at ± 65 mV, and in the presence of 10 mM lumenal Ca2+ at (, ) at +65, +50, +35, 35, 50 and 65 mV (from left to right). (B) and (C) Values are the mean ± SE of 3-19 experiments. (B) *Significantly different from Po at 4 µM lumenal Ca2+. (C) *Significantly different from Po at lumenal Ca2+ flux of <0.1 pA.

Fig. 2 B describes the dependence of mean P_o of minimally (<0.01 µM cytosolic Ca²⁺) and close to maximally (4 µM cytosolic Ca²⁺) activated CRCs on lumenal Ca²⁺ concentrations of 2 µM to 10 mM. For the minimally activated channels, a significant increase in channel open probability was observed at a lumenal Ca²⁺ concentration of 100 µM and holding potential of 50 mV. To obtain a similar increase in P_o at +50 mV, a lumenal Ca²⁺ concentration of 5 mM and greater was required. A different response was observed for channels that were close to maximally activated by 4 µM cytosolic Ca²⁺. In this case, an increase in lumenal Ca²⁺ concentration lowered P_o at 50 mV. No significant changes in P_o were observed at +50 mV. Data of Fig. 2 B suggest that three parameters must be taken into account to understand the way in which lumenal Ca²⁺ activates and inactivates the cardiac CRC. These are the extents to which channels are activated by cytosolic effectors such as Ca²⁺, the lumenal Ca²⁺ concentration, and the holding potential.

A negative holding potential favors, whereas a positive holding potential disfavors, the movement of cations from the SR lumenal (trans) side to the cytosolic (cis) side of the bilayer. The different Ca²⁺ activation/inactivation curves of Fig. 2 B suggest, therefore, that lumenal Ca²⁺ flowing through the open channel affects channel activity by having access to cytosolic Ca²⁺ regulatory sites. Lumenal Ca²⁺ fluxes could not be directly measured (except at 0 mV, see below) because of the presence of K⁺ as the major current carrier in our recording solutions. Lumenal-to-cytosolic Ca²⁺ fluxes were therefore calculated according to a barrier model that describes the ionic conduction of the sheep cardiac CRC (Tinker et al., 1992). Fig. 2 C shows the dependence of mean P_o of minimally (<0.01 µM cytosolic Ca²⁺) and close to maximally (4 µM cytosolic Ca²⁺) activated CRCs on the calculated lumenal Ca²⁺ fluxes. Lumenal Ca²⁺ fluxes were calculated at six holding potentials ranging from 65 mV to +65 mV and four lumenal Ca²⁺ concentrations ranging from <2 µM to 10 mM. Fig. 2 C shows that at 0.01 µM cytosolic Ca²⁺ channels were maximally activated at a lumenal-to-cytosolic Ca²⁺ flux of ~1 pA. Ca²⁺ fluxes of >10 pA appeared to be slightly inhibitory.

The SR membrane is highly permeable to K⁺ and Cl and the membrane potential across the SR membrane is therefore generally believed to be close to 0 mV (Meissner, 1983). The effects of lumenal Ca²⁺ on P_o were therefore also determined at a holding potential of 0 mV in a symmetric 0.25 M KCl buffer containing 3 mM lumenal free Ca²⁺ and a low cytosolic Ca²⁺ concentration (<0.01 µM free Ca²⁺ plus 10 mM caffeine). Under these conditions, the Ca²⁺ current could be measured directly. The measured Ca²⁺ current of 1.9 ± 0.1 pA (n = 5) was close to a calculated value of 2.1 pA. The averaged P_o of 0.53 ± 0.21 (n = 5) was close to values that yielded lumenal-to-cytosolic Ca²⁺ fluxes of ~2 pA at negative and positive holding potentials (2.1 and 1.3 pA at 20 and +20 mV and lumenal [Ca²⁺] of 1 and 5 mM, respectively).

Channels recorded at a close to maximally activating cytosolic Ca²⁺ concentration of 4 µM were not further activated by lumenal Ca²⁺ (Fig. 2 C). However, these channels were significantly inactivated at lumenal Ca²⁺ fluxes of 8 pA and greater.

An intriguing finding was that at a low cytosolic Ca²⁺ concentration lumenal Ca²⁺ fluxes were less effective than cytosolic [Ca²⁺] in activating the CRC (P_o,max = ~0.8 in Fig. 1 B vs. P_o,max of ~0.5 in Fig. 2 C). This result can be rationalized if lumenal Ca²⁺ inactivates before fully activating the release channel. We tested this idea using the "fast" complexing Ca²⁺ buffer BAPTA. Modeling studies have indicated that the free Ca²⁺ concentration near the release sites may reach values in excess of 10 mM (see Fig. 8). Because of its high association rate, BAPTA is more effective than the "slow" complexing Ca²⁺ buffer EGTA in suppressing such a rise in Ca²⁺ concentration (Stern, 1992). In the middle traces of Fig. 2 A, a single lumenal Ca²⁺-activated channel was recorded under standard conditions; that is, in the presence of <0.01 µM cytosolic Ca²⁺ and 10 mM cytosolic caffeine. Lumenal Ca²⁺ was 1 mM. Bottom traces of Fig. 2 A show that the addition of 20 mM cytosolic BAPTA increased P_o at 50 mV, but not at +50 mV. Fig. 3 B (top panel) summarizes the effects of 20 mM BAPTA on several lumenal Ca²⁺-activated single channels. At lumenal Ca²⁺ fluxes of 0.25-4 pA, 20 mM cytosolic BAPTA increased P_o. At a flux of 3 pA, a P_o value close to those observed in the presence of 0.01-1 mM cytosolic Ca²⁺ was obtained (Fig. 3 A, top panel). This result suggested that BAPTA was apparently able to prevent lumenal Ca²⁺-mediated channel inactivation by minimizing the buildup of a high inactivating Ca²⁺ concentration near the cytosolic Ca²⁺ inactivation sites. However, BAPTA did not prevent channel activation, which suggested that at a concentration of 20 mM BAPTA did not lower the Ca²⁺ concentration below a maximally activating Ca²⁺ concentration of ~5 µM (Fig. 1 B) at the cytosolic Ca²⁺ activation sites. A direct pharmacological activation of CRCs by BAPTA appeared to be unlikely because none was observed when lumenal Ca²⁺ fluxes were 0.1 pA (Fig. 3 B, top panel).

View larger version (29K): [in this window] [in a new window]   FIGURE 3   Single channel parameters of cytosolic and lumenal Ca2+-activated channels in presence of 10 mM caffeine. (A) Single channel parameters were obtained from recordings at 35 mV at indicated free cytosolic Ca2+ concentrations as described in Fig. 1. Values are the mean ± SE of 5-7 experiments. (B) Single channel parameters were obtained from recordings at holding potentials of +65, +50, +35, +20, 20, 35, 50 and 65 mV (from left to right) at lumenal Ca2+ concentration of 1 mM with () or without () 20 mM cytosolic BAPTA, as described in Fig. 2. Values are the mean ± SE of 4-7 experiments. *Significantly different from parameters in the absence of 20 mM BAPTA.

In the case of cytosolic Ca²⁺-activated CRCs, both the Ca²⁺-activating and -inactivating sites see the same [Ca²⁺]. In contrast, lumenal Ca²⁺ has access only to cytosolic regulatory sites when the channel is open. In addition, the Ca²⁺ activation and inactivation sites may see different [Ca²⁺], depending on their relative location with respect to the release site. It was therefore of interest to compare the kinetic parameters of cytosolic Ca²⁺-activated and lumenal Ca²⁺-activated channels (Fig. 3, A and B). An increase in cytosolic Ca²⁺ concentration from <0.01 µM to 100 µM increased P_o from close to zero to 0.8 by increasing the number of channel events by more than 10-fold, and the duration of mean open events by ~100-fold (Fig. 3 A). The duration of mean closed events was maximally decreased by ~10,000-fold. A further increase of cytosolic Ca²⁺ to 10 mM decreased P_o by shortening the duration of mean open events and increasing the duration of mean closed events, without having an appreciable effect on the number of channel events. In Fig. 3 B, channel parameters are plotted against the lumenal Ca²⁺ fluxes. Channels were recorded at eight holding potentials ranging from 65 to +65 mV and 1 mM lumenal Ca²⁺ and cytosolic Ca²⁺ concentration of <0.01 µM in the presence and absence of 20 mM cytosolic BAPTA. In the absence of BAPTA, lumenal Ca²⁺ fluxes were less effective than cytosolic Ca²⁺ in activating cardiac CRCs (top panels of Fig. 3, A and B). Lumenal Ca²⁺ opened and closed channels less frequently than cytosolic Ca²⁺ (second panels). In both cases, mean open times were increased as channels were maximally activated by raising cytosolic [Ca²⁺] from ~0.003 µM to 10 µM, and lumenal Ca²⁺ fluxes from 0.04 to 3 pA (third panels). However, they showed major differences in the durations of mean closed times. An increase in cytosolic Ca²⁺ from ~0.003 to 10 µM decreased the mean closed times from 10,000 ms to close to 1 ms (Fig. 3 A, bottom panel). By comparison, an increase in lumenal Ca²⁺ fluxes from 0.04 to 3 pA decreased the mean closed times by <100-fold (Fig. 3 B, bottom panel).

Cytosolic BAPTA significantly increased P_o at elevated lumenal Ca²⁺ fluxes. This increase could be accounted for mostly by an increase in mean open times (Fig. 3 B, third panel). Some changes in the number of events and mean closed times were observed as well; however, none of these was significant.

Regulation of cardiac Ca²⁺ release channel by lumenal Ca²⁺ and Mg²⁺ in the absence of caffeine

The effects of lumenal Ca²⁺ on CRCs were also investigated in the absence of caffeine. In Fig. 4 A a single channel was recorded with 10 µM and 1 mM lumenal Ca²⁺. Cytosolic Ca²⁺ was 1 µM, which was higher than in the recordings of Fig. 2 because preliminary experiments indicated that lumenal Ca²⁺ concentrations as high as 10 mM were ineffective in activating the CRC at cytosolic Ca²⁺ concentrations of <0.1 µM (not shown). At such low Ca²⁺ concentrations, channels rarely opened in the absence of caffeine. To observe appreciable channel activity in the absence of caffeine, a cytosolic Ca²⁺ concentration of 1 µM was required.

View larger version (22K): [in this window] [in a new window]   FIGURE 4   Dependence of Po on lumenal [Ca2+] in absence of caffeine. (A) Single channel currents were recorded at 35 mV (downward deflections, left panels) and +35 mV (upward deflections, right panels) in symmetrical 0.25 M KCl, 20 mM KHepes, pH 7.4 media containing 1 µM free cytosolic Ca2+ and indicated concentrations of lumenal Ca2+. (B) Single channels were recorded as in (A) in presence of indicated concentrations of cytosolic and lumenal Ca2+. Holding potentials were 35 mV (, ) and +35 mV (, ). Calculated Ca2+ fluxes at 35 and +35 mV were, respectively, 0.8 and 0.1 pA (200 µM lumenal Ca2+), 3.0 and 0.2 pA (1 mM lumenal Ca2+), 6.0 and 0.7 pA (5 mM lumenal Ca2+), and 6.9 and 0.9 pA (10 mM lumenal Ca2+). Values are the mean ± SE of 8-14 experiments. *Significantly different from Po values at 10 µM lumenal Ca2+.

Fig. 4 A shows that an increase in lumenal Ca²⁺ concentration from 10 µM to 1 mM caused an ~5-fold increase in P_o at 35 mV. By comparison, an only minimal increase in channel activity was evident at +35 mV. Fig. 4 B compares the dependence of CRC activity on lumenal Ca²⁺ concentrations at cytosolic Ca²⁺ concentrations that resulted in either a minimum (1 µM Ca²⁺) or close to maximum (10 µM Ca²⁺) channel activity in the absence of caffeine. In the presence of 1 µM cytosolic Ca²⁺, lower lumenal [Ca²⁺] was required at negative than positive holding potentials to observe a significant increase in P_o (0.2 mM at 35 mV vs. 5 mM at +35 mV; corresponding lumenal Ca²⁺ fluxes were 0.8 pA and 0.7 pA). In the presence of 10 µM cytosolic Ca²⁺, P_o decreased at the negative holding potential at [Ca²⁺] 1 mM, whereas only a small (not significant) increase was obtained at +35 mV at lumenal [Ca²⁺] as high as 10 mM (corresponding Ca²⁺ fluxes were 3 pA and 0.9 pA, respectively). These results suggest that Ca²⁺-activated CRCs can be activated or inactivated in a voltage-dependent manner by lumenal Ca²⁺ in the absence of caffeine.

The inhibitory effects of lumenal Ca²⁺ on P_o of maximally activated channels were also determined at a holding potential of 0 mV in a symmetric 0.25 M KCl buffer containing 20 mM lumenal Ca²⁺ and 10 µM cytosolic Ca²⁺. The measured Ca²⁺ current of 2.7 ± 0.4 pA (n = 4) was close to a calculated value of 3.1 pA. P_o was significantly decreased by 32 ± 7% (n = 4) compared to control values obtained at ±5 mV at lumenal Ca²⁺ concentration of <2 µM (not shown). We conclude that the maximally Ca²⁺-activated CRCs can be inactivated at 0 mV by a directly measured Ca²⁺ flux in the absence of caffeine.

CRC conducts Mg²⁺ (Meissner, 1994) and cytosolic Mg²⁺ inactivates the cardiac CRC by binding to Ca²⁺ activation and inactivation sites with micromolar and millimolar affinity, respectively (Liu et al., 1998; Laver et al., 1997). We rationalized that a voltage-independent inhibition of lumenal Mg²⁺ would suggest the existence of Mg²⁺ inhibitory sites that reside on the SR lumenal site, whereas a voltage-dependent inhibition would favor the idea of an access of lumenal Mg²⁺ to the cytosolic Ca²⁺ regulatory sites. The effects of lumenal Mg²⁺ (0-50 mM) on single cytosolic Ca²⁺-activated channels were tested at holding potentials of ±5, ±35, and ±50 mV. In the presence of ~4 µM Ca²⁺ in both bilayer chambers, a strong inhibition of channel activity was observed at negative holding potentials that favored the movement of lumenal Mg²⁺ to the cytosolic side of CRC, and yielded lumenal-to-cytosolic Mg²⁺ fluxes of 2.0 pA. No appreciable inhibition was noted at positive holding potentials that disfavored the movement of lumenal Mg²⁺ to the cytosolic side of the bilayer and yielded Mg²⁺ fluxes of <1 pA (not shown). We also measured the Mg²⁺ current at 0 mV in a symmetric 0.25 M KCl solution containing 10 mM lumenal Mg²⁺. Addition of 10 mM lumenal Mg²⁺ decreased P_o to 22 ± 10% of the control at ±5 mV in the absence of Mg²⁺ (n = 5). The directly measured Mg²⁺ current of 2.1 ± 0.1 pA (n = 5) agreed well with calculated value of 2.2 pA. We conclude from these observations that lumenal Mg²⁺ fluxes affected cardiac CRC activity by having access to the channel's cytosolic Ca²⁺ regulatory sites. In frog skeletal muscle, the Mg²⁺ levels in the SR lumen near the Ca²⁺ release sites increase rather than decrease during tetanus (Somlyo et al., 1985). An in vivo regulation by lumenal-to-cytosolic Mg²⁺ fluxes appears, therefore, to be unlikely.

Regulation of cardiac Ca²⁺ release channel by cytosolic and lumenal Ca²⁺ in the presence of 5 mM MgATP

The total ATP and free Mg²⁺ concentrations in myocardium have been estimated to range from 5 to 10 mM (Koretsune et al., 1991; Hohl et al., 1992) and 0.7-1.0 mM (Murphy et al., 1989), respectively. Figs. 5 and 6 compare the voltage-dependence of cytosolic and lumenal Ca²⁺-activated channel activities recorded in the presence of 5 mM cytosolic MgATP (~0.7 mM free Mg²⁺), but in the absence of caffeine. An ~10× higher cytosolic Ca²⁺ concentration was required to half-maximally activate the cardiac CRC (K_Ha = 14.4 µM vs. 1 µM, Figs. 5 B and 1 B, respectively). As observed in the presence of caffeine (Fig. 1), no significant voltage-dependence in channel activity was noted for cardiac release channels activated by cytosolic Ca²⁺ in the presence of 5 mM cytosolic MgATP (Fig. 5, A and B).

View larger version (30K): [in this window] [in a new window]   FIGURE 5   Dependence of single channel activities on cytosolic [Ca2+] in the presence of 5 mM MgATP. (A) Single channel currents were recorded at 35 mV (downward deflections, left current traces) and +35 mV (upward deflections, right current traces) in symmetrical 0.25 M KCl, 20 mM KHepes, pH 7.4 media containing 5 mM cytosolic MgATP and indicated free cytosolic [Ca2+]. The trans (SR lumenal) solution contained 4 µM Ca2+. (B) Dependence of Po on cytosolic [Ca2+]. Po values were determined at 35 mV () and +35 mV () as in (A). Values are the mean ± SE of five experiments. Continuous lines were obtained assuming that CRC possesses cooperatively interacting high-affinity Ca2+ activation and low-affinity Ca2+-inactivation sites (Scheme 1 and Eq. 1 of Liu et al., 1998). At 35 mV, Hill constants and coefficients were KHa = 14.4 µM, nHa = 1.8, KHi = 12.6 mM, and nHi = 1.2.

View larger version (19K): [in this window] [in a new window]   FIGURE 6   Activation of the cardiac Ca2+ release channel by lumenal Ca2+ in the presence of 5 mM MgATP. (A) Single channel currents were recorded at 50 mV (downward deflections, left panels) and +50 mV (upward deflections, right panels) in symmetrical 0.25 M KCl, 20 mM KHepes, pH 7.4 media containing 10 µM free cytosolic Ca2+ and 5 mM cytosolic MgATP, and indicated concentrations of lumenal Ca2+. (B) Dependence of Po on lumenal [Ca2+] in the presence of 2-10 µM free cytosolic Ca2+ and 5 mM cytosolic MgATP. Holding potentials were 50 mV () and +50 mV (). Values are the mean ± SE of three to nine experiments. *Significantly different from Po at 4 µM lumenal Ca2+ (B).

In contrast to cytosolic Ca²⁺-activated channels, CRC activities indicate a voltage-dependence when recorded at elevated lumenal Ca²⁺ concentrations. In Fig. 6 A, a single cardiac CRC was recorded in the presence of 10 µM cytosolic Ca²⁺ and 5 mM cytosolic MgATP at lumenal Ca²⁺ concentrations of 4 µM and 200 µM. Elevation of lumenal Ca²⁺ resulted in increased channel activity at 50 mV but not +50 mV. Fig. 6 B shows that at 50 mV channels were significantly activated at lumenal [Ca²⁺] of ~200-1000 µM. Higher lumenal Ca²⁺ concentrations resulted in (not significant) inactivation of channel activities. At +50 mV, higher lumenal Ca²⁺ concentrations were required to observe an increase in channel activity; however, these were not significant.

Fig. 7, A and B compares the kinetic parameters of cytosolic and lumenal Ca²⁺-activated channel activities recorded in the presence of 5 mM cytosolic MgATP. An increase in cytosolic Ca²⁺ concentration from ~0.1 to 100 µM increased P_o from nearly zero to ~1.0. This increase could be largely accounted for by an ~100-fold increase in the number of channel events and ~1000-fold increase in mean open times (Fig. 7 A). Mean closed events were decreased by ~100-fold. A further increase of cytosolic Ca²⁺ to 10 mM decreased P_o by decreasing mean open times and by slightly increasing the duration of mean closed events, without having an appreciable effect on the number of events. In Fig. 7 B, mean P_o, number of channel events, and mean open and closed times are plotted against the lumenal Ca²⁺ fluxes. The latter were less effective in activating cardiac CRCs than cytosolic Ca²⁺ (P_o,max = ~1 at cytosolic Ca²⁺ of ~100 µM vs. ~0.15 at lumenal Ca²⁺ flux of ~3 pA). Small increases in P_o could be largely accounted for by small (significant) increases in duration of mean open times. Few, if any, changes were observed in the number of channel events and duration of mean closed events, as lumenal Ca²⁺ fluxes increased from 0.003 to 10 pA. Taken together, the data of Fig. 7, A and B suggest that lumenal-to-cytosolic Ca²⁺ fluxes can regulate the cardiac CRC in the presence of physiologically relevant concentrations of Mg²⁺ and ATP.

View larger version (27K): [in this window] [in a new window]   FIGURE 7   Single channel parameters of cytosolic and lumenal Ca2+-activated channels in the presence of 5 mM MgATP. Single channel parameters in (A) and (B) were obtained from recordings (A) at 35 mV and (B) at 50 mV at lumenal Ca2+ concentrations of 4 µM, 200 µM, 500 µM, 1 mM, 5 mM, and 10 mM (from left to right) as described in Figs. 5 and 6, respectively. Values are the mean ± SE of three experiments (A) and three to nine experiments (B). *Significantly different from lumenal Ca2+ flux of 0.003 pA (B).

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

The results of this study suggest that lumenal Ca²⁺ flowing through the open cardiac Ca²⁺ release channel can regulate the channel by having access to cytosolic activation and inactivation sites. Activation in the presence of Mg²⁺ and ATP suggests that regulation of CRC by lumenal Ca²⁺ fluxes may be physiologically relevant.

Regulation of cardiac CRC activity by cytosolic and lumenal Ca²⁺

To distinguish between the effects of SR cytosolic and lumenal Ca²⁺ on channel activity, single purified channels were recorded in symmetric KCl media at different holding potentials and with varying Ca²⁺ concentrations in the trans (SR lumenal) and cis (cytosolic) chambers of the bilayer apparatus. As previously observed for the skeletal muscle CRC (Tripathy and Meissner, 1996), a strong voltage-dependence of channel activities was observed in the presence of elevated levels of lumenal, but not cytosolic, Ca²⁺. A voltage-dependent activation by lumenal Ca²⁺ was observed in the absence of caffeine provided sufficiently high cytosolic [Ca²⁺] was used to partially open the channel, which suggested that other channel activators such as sulmazole (Sitsapesan and Williams, 1994a) or ATP (Lukyanenko et al., 1996) were not required for cardiac channel activation by lumenal Ca²⁺. In the absence of caffeine and with cytosolic [Ca²⁺] of <0.1 µM in the presence (Fig. 7) and absence (Fig. 4) of 5 mM MgATP, the cardiac CRC rarely opened. Under these recording conditions, lumenal [Ca²⁺] as high as 10 mM was not able to significantly activate the channel. In agreement with this finding, cellular SR lumenal [Ca²⁺], which is thought to be close to 1 mM, does not activate the "closed" CRC. As in cells, where Ca²⁺ ions entering the cells activates the cardiac CRC, the presence of a cytosolic activator such as Ca²⁺ or caffeine was required before an activation and inactivation of the CRC by lumenal Ca²⁺ could be observed. Lack of an activation of the "closed" CRC by lumenal Ca²⁺ argues against a low-affinity Ca²⁺ regulatory site that resides on the lumenal site of the channel.

The lumenal-to-cytosolic Ca²⁺ fluxes were calculated using a four-barrier model that describes the ionic conduction of the sheep cardiac CRC (Tinker et al., 1992). In general, barrier models are inadequate to explain ion fluxes through channels over a large range of membrane potential (Chen et al., 1997). This limitation was also pointed out by Tinker et al. (1992) who could not fit their data by a four-barrier model at potentials >±80 mV. Recently, the flow of K⁺ through cardiac CRC has been modeled by diffusion theory using a combination of the Nernst-Plank and Poisson (PNP) equations (Chen et al., 1997). The model predicts a high K⁺ concentration (~4 M) in the selectivity filter at bath concentrations as low as 25 mM, thus providing an explanation for the high conductances of the CRCs. However, in contrast to the Tinker model, the PNP model has not yet been extended to mixed solutions containing Ca²⁺. Tinker et al. (1992) measured and modeled ion conductances in bionic and mixed solutions, including Ca²⁺, Mg²⁺, and K⁺. We directly measured Ca²⁺ and Mg²⁺ currents and their effects at 0 mV in symmetric KCl solutions. Good agreement with the calculated values suggests that at the membrane potentials used in our study, the Tinker model serves as a useful "curve-fitting" tool to predict ion fluxes in mixed solutions.

Parameters determining the extent of CRC activation and inactivation by lumenal Ca²⁺

The extent of CRC activation by lumenal Ca²⁺ was dependent on the presence of Ca²⁺, MgATP, and caffeine in the cytosolic (cis) chamber of the bilayer apparatus. In agreement with observations of an increased Ca²⁺ affinity of Ca²⁺ activation sites by caffeine (Zucchi and Ronca-Testoni, 1997; Liu et al., 1998), channels could be more effectively activated at lower lumenal Ca²⁺ fluxes in the presence of caffeine (100 µM lumenal Ca²⁺ at 50 mV corresponds to lumenal Ca²⁺ flux of 0.6 pA, Fig. 2 B) than in the absence of caffeine (200 µM lumenal Ca²⁺ at 35 mV corresponds to lumenal Ca²⁺ flux of 0.8 pA, Fig. 4 B). Addition of 5 mM cytosolic MgATP increases the Hill constant of Ca²⁺ activation by cytosolic Ca²⁺ by 3-4-fold (Xu et al., 1996; Fig. 6). In reasonable agreement with this result, CRCs were activated by lower lumenal Ca²⁺ fluxes in the absence of MgATP (0.8 pA in Fig. 4 B; in Fig. 6, 200 µM lumenal Ca²⁺ at 50 mV corresponds to lumenal Ca²⁺ flux of 1.0 pA).

Lumenal Ca²⁺ fluxes lead to the buildup of a high cytosolic Ca²⁺ concentration near the release sites (Stern, 1992; Fig. 8), which raised the possibility that lumenal Ca²⁺ fluxes inactivated the channels before they could be fully activated. We tested this idea using the "fast" Ca²⁺-complexing buffer BAPTA. Because of its high association rate BAPTA can suppress the rise in Ca²⁺ concentration at locations several nanometers away from the release site (Stern, 1992; Fig. 8). Fig. 3 B (top panel) shows that 20 mM cytosolic BAPTA increased channel activities close to those observed in the presence of micromolar-to-millimolar cytosolic [Ca²⁺] (Fig. 3 A, top panel), thus supporting the idea that lumenal Ca²⁺ fluxes cannot only activate but also inactivate the cardiac CRC. Channel activation by cytosolic effectors was required to observe the effects of lumenal Ca²⁺. This finding limited the conditions that could be used to test the effects of BAPTA. Specifically, BAPTA could not be used to test the effects of lumenal Ca²⁺ fluxes in the presence of 5 mM MgATP because, in agreement with the in vivo function of the CRC, only few, if any, channel openings could be observed at cytosolic Ca²⁺ concentrations of 0.1 µM.

View larger version (21K): [in this window] [in a new window]   FIGURE 8   Cytosolic [Ca2+] profiles at 0.1, 3, and 10 pA lumenal Ca2+ fluxes in the presence 0.2 mM EGTA and absence of BAPTA, and in the presence of 20 mM BAPTA. Also shown (dotted lines) are [Ca2+] that half-maximally activated (1 µM Ca2+) and inactivated (10 mM Ca2+) cytosolic Ca2+-activated CRCs in the presence of 10 mM caffeine (Fig. 1 B). Lumenal-to-cytosolic Ca2+ fluxes were calculated using the barrier model and parameters of Tinker et al. (1992). Cytosolic Ca2+ gradients were derived according to Eq. 13 of Stern (1992) using the following constants for Ca2+ and BAPTA: kon = 1.7 × 109 M1 s1; Kd = 4 × 107 M; DCa = 3 × 106 cm2 s1; DBAPTA = DCaBAPTA = 105 cm2 s1. EGTA has an ~1000-fold lower Ca2+ on-rate constant than BAPTA. The presence of 0.2 mM EGTA therefore did not significantly affect cytosolic [Ca2+] profiles.

Kinetics of CRC activation and inactivation by cytosolic and lumenal Ca²⁺

Kinetics of cytosolic Ca²⁺-mediated channel activation and inactivation are, at least in principle, more straightforward than those by lumenal Ca²⁺ and will therefore be discussed first. At low cytosolic Ca²⁺ concentrations, channels opened infrequently and long-closed/short-open channel events predominated, resulting in a low channel open probability (Figs. 1 A and 5 A). An increase in the number of channel events and a decrease in closed mean times with increasing Ca²⁺ concentration indicated that cytosolic Ca²⁺ increased P_o by increasing the transition rates from the closed to open state(s). A second effect of increasing cytosolic [Ca²⁺] was to increase the mean open times. Ca²⁺ activated CRCs by a cooperative mechanism in the presence of caffeine and MgATP, and the increase in mean open time may have been therefore due to the cooperative binding of Ca²⁺ to the tetrameric channel complex. An increase in open times by cytosolic Ca²⁺ was also observed for the sheep cardiac CRC (Sitsapesan and Williams, 1994b). This increase was explained by assuming a Ca²⁺-dependent pathway between two open states. High Ca²⁺ concentrations inactivate the channel by binding to low-affinity sites (Liu et al., 1998; Laver et al., 1995). In our single channel recordings, 10 mM cytosolic Ca²⁺ decreased P_o by decreasing the mean open times and increasing the mean closed times, without appreciably affecting the number of single channel events. These changes suggest that Ca²⁺ binding to the Ca²⁺-inactivation sites affects both the transition rates from the open-to-closed and from the closed-to-open states, increasing the former and decreasing the latter.

According to our model, lumenal Ca²⁺ is only available to cytosolic Ca²⁺ regulatory sites when a channel opens. Ca²⁺ gradients formed by Ca²⁺ fluxes build up and dissipate in ~50 µs as channels open and close (Simon and Llinas, 1985). Accordingly, cytosolic Ca²⁺ gradients formed by lumenal Ca²⁺ fluxes likely had a lifetime that was less than that of the shortest channel events seen in the bilayers (~0.2 ms). One would then expect that the frequency of channel openings is being set mainly by the cytosolic Ca²⁺, caffeine, and MgATP concentrations, while lumenal Ca²⁺ does not noticeably affect the number of events and duration of mean closed events. In agreement with this prediction, lumenal Ca²⁺ did not significantly affect the number of channel events and duration of mean closed events. Much of the increase in P_o observed for lumenal Ca²⁺-activated channels could be accounted for by an increase in mean open times. This increase was likely due to the rapid buildup of a cytosolic Ca²⁺ gradient because a similar prolongation in mean open times was observed with increasing cytosolic [Ca²⁺] (Figs. 3 A and 7 A, third panels). We conclude that lumenal Ca²⁺ ions flowing through open channels may increase the duration of channel open events by elevating cytosolic [Ca²⁺] at Ca²⁺ activation sites. The frequency of these regulatory events is mainly set by cytosolic factors that determine the frequency of channel openings such as cytosolic Ca²⁺ or MgATP.

Location of activation and inactivation sites

Cryoelectron microscopy and image analysis have indicated that the skeletal muscle CRC consists of a large 29 × 29 × 12 nm cytosolic "foot" region and a smaller transmembrane region that extends ~7 nm toward the SR lumen and likely contains a centrally located Ca²⁺ channel pore (Radermacher et al., 1994; Serysheva et al., 1995). A very similar architecture has been deduced for the cardiac CRC (Sharma et al., 1997). The cardiac CRC is thought to have at least two classes of Ca²⁺ binding sites, a high-affinity activation and a low-affinity inactivation site. The location of these sites, however, has not been established. Although our single channel measurements cannot pinpoint the location of the Ca²⁺ regulatory sites on the large cardiac CRC complex, our data can provide tentative information with respect to their distance from the cytosolic Ca²⁺ release site. Fig. 3 B shows that lumenal Ca²⁺ fluxes of 0.1 pA did not significantly activate the cardiac CRC in the presence of 10 mM caffeine. By comparison, Ca²⁺ fluxes of 1 pA and greater caused a nearly maximum activation of channels that were recorded in the presence of 10 mM caffeine and 20 mM BAPTA. The cytosolic Ca²⁺ concentration profiles that were obtained at lumenal Ca²⁺ fluxes of 0.1 pA and 3 pA are included in Fig. 8. Also indicated in Fig. 8 is the cytosolic Ca²⁺ concentration (1 µM, dotted line) that resulted in half-maximum activation of CRCs in the presence of 10 mM caffeine (Fig. 1 B). Together these data show that lumenal Ca²⁺ fluxes as low as 0.1 pA should have been sufficient to maximally activate the CRC, even if the activation sites would have been located 30 nm away from the release site, which is more than the dimensions of the cardiac CRC. Another argument against a distance 20 nm between the Ca²⁺ activation and release sites is that 20 mM BAPTA at lumenal flux of 3 pA would have been expected to lower channel activity, which clearly was not the case. A similar paradoxical situation between the measured cytosolic Ca²⁺-activating concentrations and calculated effects of lumenal Ca²⁺ fluxes was obtained for the skeletal muscle CRC (Tripathy and Meissner, 1996). To explain the paradox, skeletal muscle cytosolic Ca²⁺ activation sites were placed within the foot region at BAPTA "inaccessible" sites. It was further suggested that these sites see a minor portion, whereas Ca²⁺ inactivation sites see a major portion of lumenal Ca²⁺. We propose a similar model for the cardiac CRC. The model suggests that lumenal Ca²⁺ fluxes increase Ca²⁺ concentrations to a lesser extent at the Ca²⁺ activation than Ca²⁺ inactivation sites, thus explaining that, as observed in the present study, Ca²⁺ inactivation sets in before the cardiac CRC can be fully activated by lumenal Ca²⁺.

The distance between the Ca²⁺ release and Ca²⁺ inactivation sites of the cardiac CRC was estimated as follows. Fig. 2 C shows that channels activated by 4 µM cytosolic Ca²⁺ in the presence of 10 mM caffeine were half-maximally inactivated at a lumenal Ca²⁺ flux of ~10 pA. This flux resulted in a half-maximally inactivating cytosolic Ca²⁺ concentration of 10 mM (Fig. 1 B) at a distance of ~3 nm from the release site (Fig. 8). Single channel measurements with the fast Ca²⁺-complexing buffer BAPTA suggest that a distance of 3 nm between the release and Ca²⁺ inactivation sites may be an upper limit. BAPTA increased channel activities at a lumenal Ca²⁺ flux of 3 pA to close a maximum value (Fig. 3 B). At a distance of 3 nm, a cytosolic [Ca²⁺] of ~3 mM is calculated (Fig. 8), which appears to be too low to cause substantial Ca²⁺ inactivation (Fig. 1 B). Higher cytosolic [Ca²⁺] exists closer to the release site (Fig. 8). However, placement of Ca²⁺ inactivation sites too close to the release site is problematic because it renders BAPTA ineffective in lowering [Ca²⁺]. According to Fig. 8, a compromise is reached at a distance of 1 nm from the release site. At this distance and at a lumenal Ca²⁺ flux of 3 pA, a cytosolic [Ca²⁺] of ~9 mM is calculated, which is lowered by 20 mM BAPTA to ~6 mM (Fig. 8). Such a decrease can account, at least in principle, for the activating effects of BAPTA (Fig. 3 B). Taken together, these results suggest that the Ca²⁺ inactivation site(s) lie(s) at a distance of 3 nm from the release site. This distance is reasonably close to the distances of 3-6 nm estimated between the two sites of the skeletal muscle CRC (Tripathy and Meissner, 1996).

Physiological implications

In mammalian ventricular muscle, clusters of Ca²⁺ release channels are located near the surface membrane and tubular infoldings (T-tubule) of the surface membrane (Franzini-Armstrong and Protasi, 1997). Immunolocalization studies suggest a co-distribution of CRCs with surface dihydropyridine receptors (Ca²⁺ channels, L-type), which provides a morphological basis for the Ca²⁺-induced Ca²⁺ release (CICR) mechanism (Carl et al., 1995). Recent studies suggest that the opening of a single L-type Ca²⁺ channel may be sufficient to evoke a localized Ca²⁺ release event ("Ca²⁺ spark") by activating one or more CRCs (Santana et al., 1996). During L-type Ca²⁺ channel opening the Ca²⁺ concentration can reach millimolar values (Langer and Peskoff, 1996), which are more than enough to activate closely apposed Ca²⁺ release channels. The present study shows that SR lumenal Ca²⁺ can contribute to the regulation of cardiac SR Ca²⁺ release via direct feedback by binding to channels that release Ca²⁺. In myocardium these events may involve more than one channel because, in addition to its own channel, lumenal Ca²⁺ fluxes may activate and inactivate closely located Ca²⁺ release channels.