Novel Regulation of Calcium Inhibition of the Inositol 1,4,5-trisphosphate Receptor Calcium-release Channel

Abstract

The inositol 1,4,5-trisphosphate (InsP₃) receptor (InsP₃R), a Ca²⁺-release channel localized to the endoplasmic reticulum, plays a critical role in generating complex cytoplasmic Ca²⁺ signals in many cell types. Three InsP₃R isoforms are expressed in different subcellular locations, at variable relative levels with heteromultimer formation in different cell types. A proposed reason for this diversity of InsP₃R expression is that the isoforms are differentially inhibited by high cytoplasmic free Ca²⁺ concentrations ([Ca²⁺]_i), possibly due to their different interactions with calmodulin. Here, we have investigated the possible roles of calmodulin and bath [Ca²⁺] in mediating high [Ca²⁺]_i inhibition of InsP₃R gating by studying single endogenous type 1 InsP₃R channels through patch clamp electrophysiology of the outer membrane of isolated Xenopus oocyte nuclei. Neither high concentrations of a calmodulin antagonist nor overexpression of a dominant-negative Ca²⁺-insensitive mutant calmodulin affected inhibition of gating by high [Ca²⁺]_i. However, a novel, calmodulin-independent regulation of [Ca²⁺]_i inhibition of gating was revealed: whereas channels recorded from nuclei kept in the regular bathing solution with [Ca²⁺] ∼400 nM were inhibited by 290 μM [Ca²⁺]_i, exposure of the isolated nuclei to a bath solution with ultra-low [Ca²⁺] (<5 nM, for ∼300 s) before the patch-clamp experiments reversibly relieved Ca²⁺ inhibition, with channel activities observed in [Ca²⁺]_i up to 1.5 mM. Although InsP₃ activates gating by relieving high [Ca²⁺]_i inhibition, it was nevertheless still required to activate channels that lacked high [Ca²⁺]_i inhibition. Our observations suggest that high [Ca²⁺]_i inhibition of InsP₃R channel gating is not regulated by calmodulin, whereas it can be disrupted by environmental conditions experienced by the channel, raising the possibility that presence or absence of high [Ca²⁺]_i inhibition may not be an immutable property of different InsP₃R isoforms. Furthermore, these observations support an allosteric model in which Ca²⁺ inhibition of the InsP₃R is mediated by two Ca²⁺ binding sites, only one of which is sensitive to InsP₃.

INTRODUCTION

The second messenger, inositol 1,4,5-trisphosphate (InsP₃), is generated in many cell types through the hydrolysis of phosphatidylinositol 4,5-bisphosphate by membrane-bound phospholipase C activated by plasma membrane receptors responding to extracellular stimuli. InsP₃ then diffuses through the cytoplasm to bind to its receptor (InsP₃R) in the ER and activate it as a Ca²⁺ channel to release Ca²⁺ stored in the ER lumen. Modulation of the cytoplasmic free Ca²⁺ concentration ([Ca²⁺]_i) by InsP₃R-mediated Ca²⁺ release is a ubiquitous intracellular signal transduction mechanism that regulates numerous processes (Berridge, 1993).

Three isoforms of the InsP₃R, with spliced variants, have been identified (Joseph, 1996). Most mammalian cell types express multiple InsP₃R isoforms in distinct and overlapping intracellular locations with their absolute and relative expression levels regulated by gene transcription, alternative splicing and receptor degradation that differ during different stages of cell development and in response to extracellular stimuli (Taylor et al., 1999). Furthermore, formation of hetero-tetrameric channels is possible in cell types expressing more than one InsP₃R isoform (Joseph et al., 1995; Monkawa et al., 1995; Wojcikiewicz, 1995; Nucifora et al., 1996). Although this diversity of InsP₃R expression is impressive, its functional correlates and physiological implications remain unclear. Studies of the single-channel properties of the various InsP₃R isoforms have revealed that whereas their permeation and conductance properties are very similar (Mak et al., 2000; Ramos-Franco et al., 2000), their gating may be differentially inhibited by high [Ca²⁺]_i (Bezprozvanny et al., 1991; Hagar et al., 1998; Mak et al., 1998; Ramos-Franco et al., 1998a,b, 2000; Boehning et al., 2001; Mak et al., 2001a). Because high [Ca²⁺]_i inhibition of InsP₃R channel gating may be a pivotal feedback mechanism for the regulation of intracellular Ca²⁺ signaling (Taylor, 1998), it has been suggested that differential inhibition by high [Ca²⁺]_i of the different InsP₃R isoforms may generate distinct Ca²⁺ signals in different cell types with different patterns of InsP₃R isoform expression, and that this may be a reason for the diversity of InsP₃R expression (Hagar et al., 1998).

It has been suggested that high [Ca²⁺]_i inhibition of the InsP₃R is mediated by calmodulin (CaM), a ubiquitous Ca²⁺-binding protein that binds to and regulates the functions of many proteins. CaM was found to bind to the InsP₃R-1 in the presence of free Ca²⁺ to a single site in the regulatory domain (Maeda et al., 1991; Yamada et al., 1995; Hirota et al., 1999). Purified InsP₃R-1 channels lacking bound CaM were not inhibited by high [Ca²⁺]_i, whereas addition of CaM restored inhibition of channel gating by high [Ca²⁺]_i (Hirota et al., 1999; Michikawa et al., 1999). The notion that high Ca²⁺ inhibition of channel gating was mediated by CaM was reinforced by observations that the type 3 InsP₃R (InsP₃R-3) did not bind CaM (Yamada et al., 1995; Cardy and Taylor, 1998; Lin et al., 2000) and was not inhibited by high [Ca²⁺]_i (Hagar et al., 1998). Nevertheless, other data suggest that the role of CaM in high [Ca²⁺]_i inhibition of InsP₃R channel gating is far from unequivocal. Despite the absence of detectable interaction between CaM and a mutant InsP₃R-1 in which the putative CaM binding site was eliminated (Yamada et al., 1995), more recent studies have demonstrated that this mutant channel is nevertheless still inhibited by high [Ca²⁺]_i (Zhang and Joseph, 2001; Nosyreva et al., 2002). Furthermore, whereas the InsP₃R-3 lacks the CaM binding site present in the InsP₃R-1 and no interaction between InsP₃R-3 and CaM has been detected (Yamada et al., 1995; Cardy and Taylor, 1998; Lin et al., 2000), electrophysiological studies of the recombinant rat InsP₃R-3 in its native membrane environment demonstrated that it is nevertheless inhibited by high [Ca²⁺]_i (Mak et al., 2001a) with quantitative features similar to those of inhibition of the InsP₃R-1 in the same membrane (Mak et al., 1998).

Here, we investigated the possible effects of CaM on high [Ca²⁺]_i inhibition of the gating of single endogenous InsP₃R-1 channels in their native membrane environment using nuclear membrane patch clamp electrophysiology (Mak and Foskett, 1994). Our experiments do not provide evidence supporting any role for CaM in this process. However, we discovered a novel regulation of high [Ca²⁺]_i inhibition of InsP₃R-1 channel gating. Inhibition of InsP₃R-1 gating by high [Ca²⁺]_i can be reversibly abrogated by exposure of the channel to a bathing solution containing ultra-low [Ca²⁺] (<5 nM). Our observations indicate that inhibition of InsP₃R-1 channel gating by high [Ca²⁺]_i can be disrupted by environmental conditions experienced by the channel, and therefore may not be an invariant property of a specific InsP₃R isoform. Furthermore, these observations support an allosteric model in which Ca²⁺ inhibition of the InsP₃R is mediated by two Ca²⁺ binding sites, only one of which is sensitive to InsP₃.

MATERIALS AND METHODS

Heterologous Expression of Calmodulin in Xenopus Oocytes

Maintenance of Xenopus laevis and surgical extraction of ovaries were performed as described previously (Mak and Foskett, 1994, 1997, 1998). Oocytes were defolliculated as described (Jiang et al., 1998). cRNA (1 μg/μl) of rat calmodulin (CaM), either wild-type (w.t.) or a quadruple mutant (q.m.) containing a D→A mutation in each of the four EF hands so that Ca²⁺ binding in all EF hands was abolished (Xia et al., 1998; Keen et al., 1999), was synthesized in vitro from cDNA provided as a gift by Dr. John P. Adelman (Vollum Institute, Portland, OR). 23 nl of cRNA (either w.t. or q.m.) was injected into the cytoplasm of oocytes 1 d after defolliculation, as described (Mak et al., 2000). cRNA-injected and uninjected control oocytes were maintained under identical conditions in individual wells in 96-well plates containing 200 μl of ASOS (100 mM NaCl, 2 mM KCl, 1 mM MgCl₂, 1.8 mM CaCl₂, 5 mM HEPES, pH adjusted to 7.6 with NaOH; with 3 mM Na pyruvate, 100 μg/ml gentamycin, and 100 μM N-acetyl-Leu-Leu-Norleucinal; Sigma-Aldrich). 80 μl of ASOS in each well was changed daily. Nuclear patch clamp experiments and immunoprecipitations were performed 2–4 d after c-RNA injection when the expression level of exogenous CaM was stable as determined by Western analysis.

Western Analysis and Immunoprecipitation

Western analysis was performed on oocyte extracts (cRNA-injected and uninjected), as described in Mak et al. (2000), to ascertain the levels of endogenous and heterologously expressed CaM in the oocytes using a specific antibody (Upstate Biotechnology). Immunoprecipitation of InsP₃R (type 1) and CaM was performed using oocyte lysates, as described in (Mak et al., 2000), with a specific type 1 InsP₃R antibody (Joseph and Samanta, 1993; Joseph et al., 1995) and protein A agarose (GIBCO BRL), and an antibody to CaM and protein G agarose (GIBCO BRL), respectively.

Solutions for Patch Clamp Experiments

All patch clamp experiments were performed with solutions containing 140 mM KCl and 10 mM HEPES with pH adjusted to 7.1 with KOH. The free Ca²⁺ concentration ([Ca²⁺]_i) of the pipette solutions (to which the cytoplasmic side of the InsP₃R is exposed in patch-clamp experiments) was tightly controlled by buffering various amounts of added CaCl₂ (40–400 μM) with 500 μM of the high-affinity Ca²⁺ chelator, BAPTA (1,2-bis(O-aminophenoxy) ethane-N,N,N′,N′-tetraacetic acid; Molecular Probes) and 0.5 mM Na₂ATP (100 nM < [Ca²⁺]_i < 2.5 μM); or 500 μM of the low-affinity Ca²⁺ chelator, 5,5′-dibromo BAPTA (Molecular Probes) and 0.5 mM Na₂ATP (5 μM < [Ca²⁺]_i < 15 μM); or 0.5 mM Na₂ATP alone (15 μM < [Ca²⁺]_i < 300 μM). Solutions with [Ca²⁺]_i > 300 μM contained no Ca²⁺ chelator for buffering. The normal Ca²⁺ bath solution (NCaS) contained 500 μM BAPTA and 250 μM CaCl₂ (free [Ca²⁺] ≈ 400–500 nM), and the physiological Ca²⁺ bath solution (PCaS) contained 500 μM BAPTA and 70 μM CaCl₂ (free [Ca²⁺] = 48 ± 5 nM). The free [Ca²⁺] of these solutions was directly measured using Ca²⁺-selective mini-electrodes (Baudet et al., 1994). The ultra-low Ca²⁺ bath solution (ULCaS) contained 1 mM BAPTA and no added CaCl₂. The contaminating [Ca²⁺] in the solution was determined by induction-coupled plasma mass spectrometry (Mayo Medical Laboratory) to be ∼6–10 μM. Ca²⁺-selective minielectrodes were unable to determine accurately the free [Ca²⁺] in the ULCaS because of the nonlinear response of the electrode in free [Ca²⁺] < 5 nM. Free [Ca²⁺] was calculated using the Maxchelator software (C. Patton, Stanford University, Stanford, CA) to be ∼0.9–1.5 nM.

Unless specified otherwise, all pipette solutions contained a saturating concentration (10 μM) of InsP₃ (Mak and Foskett, 1994) from Molecular Probes. When specified, the pipette solutions also contained 500 μM W-7 (a CaM binding antagonist; N-(6-aminohexyl)-5-chloro-1-naphthalene-sulfonamide hydrochloride; Calbiochem), or 10 μM purified bovine CaM (Calbiochem). All reagents were used with no further purification.

Oocyte Nucleus Isolation Protocols

A stage V or VI oocyte was gently teased open mechanically in the isolation bathing solution, enabling the translucent nucleus to be isolated from the cytoplasmic material. The isolated nucleus was either directly transferred to the experimental bathing solution (protocol Nd, Ld, and Pd, Fig. 1), or it was transferred through a series of culture dishes containing 4–5 ml of incubation bath solutions (protocol L, LN, and LNL, in Fig. 1) before it was ultimately transferred to the experimental bath. The nucleus remained in each incubation bath for at least 20 min before the next transfer, to ensure that the solution in the perinuclear lumen between the outer and inner nuclear envelope had attained ionic equilibrium with the bath solution (Mak and Foskett, 1994). Approximately 20 μl of the previous bath solution accompanied the nucleus to the new bath in a transfer. The culture dish containing the nucleus in the experimental bath solution was finally moved onto the stage of the inverted microscope where patch clamp experiments were performed.

Figure 1.

Schematic diagram showing the various protocols used to isolate oocyte nuclei for nuclear patch clamp experiments.

Acquisition and Analysis of Single-Channel Patch-clamp Current Records

The isolated nucleus was gently immobilized as described previously (Mak and Foskett, 1994) so that membrane patches could be repeatedly obtained from the same region (±2 μm) of the outer nuclear membrane (Mak and Foskett, 1997). Due to abrupt termination of channel activity (Mak and Foskett, 1994, 1997), patch clamp experiments were performed in “on-nucleus” configuration to maximize the duration of channel activities recorded. To prevent contamination of the pipette solution by the bath solution (especially the Ca²⁺ chelator in the bath solution) by diffusion through the pipette tip during the time when the pipette was immersed in the bath and before giga-Ohm seal formation, a positive pressure (∼10 mmHg) was maintained inside the pipette until the pipette tip was properly positioned on the nuclear membrane. Then suction was applied in the pipette to obtain the giga-Ohm seal. All experiments were performed at room temperature with the pipette electrode at +20 mV relative to the reference bath electrode unless specifically stated otherwise. Each experiment recorded the InsP₃R channel activity at a specific [Ca²⁺]_i and [InsP₃], with no change of the pipette or bath solutions during the experiment. Data acquisition was performed as previously described (Mak et al., 1998), with currents recorded with a filtering frequency of 1 kHz and a digitizing frequency of 5 kHz.

The patch clamp current traces were analyzed using MacTac software (Bruxton) to identify channel-opening and -closing events using a 50% threshold. Current traces exhibiting one InsP₃R channel, or two InsP₃R channels determined to be identical and independently gated (Mak and Foskett, 1997), were used for channel open probability (P _o) evaluation. The number of channels in the membrane patch was assumed to be the maximum number of open channel current levels observed throughout the current record. In experimental conditions with P _o > 0.1, only current records with longer than 10 s of InsP₃R channel activities were used for determination of P _o, so there is little uncertainty in the number of channels in the current traces used. In experimental conditions with P _o < 0.1, only current records exhibiting one open channel current level with InsP₃R channel activities lasting longer than 30 s were used, to ensure that they were truly single-channel records (Mak et al., 2001a). The P _o data shown for each set of experimental conditions are the means of results from at least four separate patch-clamp experiments performed under the same conditions. Error bars indicate the SEM.

RESULTS

Lack of Effect of Calmodulin on Ca²⁺ Inhibition of InsP₃R Gating in Endoplasmic Reticulum Membrane

Previous single-channel patch-clamp studies of the endogenous Xenopus type 1 InsP₃R (X-InsP₃R-1) in its native ER membrane environment revealed a biphasic regulation by [Ca²⁺]_i of the single-channel open probability (P _o) (Mak et al., 1998, 2001b). It has been suggested that calmodulin (CaM) bound to the channel mediates inhibition of InsP₃R-1 gating by high [Ca²⁺]_i (Michikawa et al., 1999). We therefore investigated the possibility that the high [Ca²⁺]_i inhibition of X-InsP₃R-1 channel gating observed in our previous studies was mediated by CaM. Oocyte nuclei were isolated and transferred directly into an experimental bath of NCaS for patch-clamp experiments (protocol Nd in Fig. 1) . By repeated patch clamping over the surface of an isolated nucleus, regions on the outer nuclear envelope were identified in which the probability of detecting InsP₃R channel activities in membrane patches (P _d) was high (Mak and Foskett, 1997). A series of patch-clamp experiments was performed at these regions with pipette solutions (to which the cytoplasmic side of the InsP₃R was exposed) alternately containing either [Ca²⁺]_i = 755 nM, or very high [Ca²⁺]_i (290 μM) with 500 μM of W-7, a CaM binding antagonist. The former solution is one in which the channel gates with a high P _o, thereby ascertaining the presence of functional InsP₃R channels in the regions selected during the series of experiments. In contrast, the latter solution has [Ca²⁺]_i sufficiently high to inhibit InsP₃R channel gating (Mak et al., 1998). Because CaM is endogenously expressed in Xenopus oocytes (Fig. 2 , Lane A and C), we reasoned that if CaM mediated the high [Ca²⁺]_i inhibition of InsP₃R channel gating, then inclusion of 500 μM of W-7 in the pipette solution may block high [Ca²⁺]_i inhibition by interfering with CaM binding to the InsP₃R channel (Michikawa et al., 1999), making channel gating observable in the 290 μM [Ca²⁺]_i solutions. Nevertheless, no channel activity was detected in any of the five patches with 500 μM W-7 and [Ca²⁺]_i = 290 μM (Fig. 3 B), whereas InsP₃R channel activities were readily detected in five out of six patches with [Ca²⁺]_i = 755 nM (Fig. 3 A).

Figure 2.

Western blots of oocyte lysates probed with a CaM antibody (both w.t. and q.m.). Lysates from oocytes injected with CaM cRNA (w.t. or q.m.) or uninjected oocytes were used as labeled. Oocytes used for lanes A and B or C and D were from the same batches, respectively. Top arrow indicates wild-type CaM and the bottom arrow indicates the quadruple mutant CaM. The slightly faster mobility of q.m. CaM is likely a reflection of the known Ca²⁺-binding dependence of CaM mobility in gels (Xia et al., 1998).

Figure 3.

Typical current traces from nuclei in NCaS bath with pipette solutions containing 10 μM InsP₃. Arrows indicate closed channel current levels. (A and B) Uninjected oocytes were used. InsP₃R channel activity was observed with [Ca²⁺]_i of 755 nM (A, n = 3), whereas no channel activity was observed in a membrane patch obtained from the same region of the same nucleus with [Ca²⁺]_i of 290 μM and the pipette solution containing 500 μM W-7 (B, n = 5). (C and D) Oocytes injected with CaM q.m. cRNA were used. InsP₃R channel activity was observed with [Ca²⁺]_i of 2.1 μM (C, n = 4), whereas no channel activity was observed in a membrane patch obtained from the same region of the same nucleus with [Ca²⁺]_i of 290 μM (D, n = 9).

Whereas this result with W-7 is seemingly inconsistent with the hypothesis that CaM mediates Ca²⁺ inhibition of InsP₃R gating, CaM-dependent regulation of the small-conductance Ca²⁺-activated K⁺ (SK) channel gating is insensitive to W-7 and other CaM inhibitors (Xia et al., 1998). However, overexpression of a mutant CaM, in which the Ca²⁺-binding EF hand motifs were disabled, interfered with the Ca²⁺ activation of the SK channel gating by competing with the endogenous CaM for the interaction with the channels (Xia et al., 1998; Keen et al., 1999). The effects of mutant CaM expression on SK channel gating provided evidence that endogenous CaM is tightly and constitutively (even in the absence of Ca²⁺) associated with the SK channel and mediates the effects of Ca²⁺ on SK channel gating. The ability of high Ca²⁺ concentrations to inhibit InsP₃R channel gating in our in vitro electrophysiological studies can be observed for long times (up to 2 h) after isolation of the nuclei (Mak et al., 1998; Boehning et al., 2001; Mak et al., 2001b). Thus, if CaM mediates the effect of high [Ca²⁺]_i, it must remain associated with the channel in the isolated nuclei, and therefore must be tightly bound to the InsP₃R and not free to diffuse away into the large experimental bath. We therefore explored the possibility that Ca²⁺ inhibition of InsP₃R channel gating was mediated by a constitutive tight association of CaM with the channel, by examining the effects of overexpression of the Ca²⁺-insensitive quadruple mutant (q.m.) CaM on the Ca²⁺ regulation of the InsP₃R.

The q.m. CaM, which has all EF hands mutated and therefore is Ca²⁺ insensitive, was overexpressed in Xenopus oocytes by cytoplasmic microinjection of cRNA. Western analysis (n = 5) indicated that the exogenous q.m. CaM was expressed to a level that was at least an order of magnitude higher than the endogenous wild-type CaM (Xia et al., 1998; Fig. 2). Patch-clamp experiments using nuclei isolated by protocol Nd (Fig. 1) from q.m. CaM-expressing oocytes revealed that InsP₃R channel gating was still inhibited by high [Ca²⁺]_i: InsP₃R channel activities were detected in 11 out of 11 patches with pipette solutions containing [Ca²⁺]_i = 2.1 μM (Fig. 3 C), but no channel activity was detected in any of 9 patches with pipette solutions containing 290 μM [Ca²⁺]_i (Fig. 3 D). These results therefore also did not support the hypothesis that Ca²⁺ inhibition of InsP₃R channel gating is mediated by CaM.

The lack of effect of overexpression of the q.m. CaM on Ca²⁺ inhibition of gating may suggest that endogenous CaM is not normally associated with the InsP₃R. We examined the biochemical association between the InsP₃R and CaM by coimmunoprecipitation. Using lysates prepared from cRNA-injected oocytes overexpressing either w.t. or q.m. CaM (Fig. 2), immunoprecipitation of the endogenous type 1 InsP₃R with a specific antibody did not coimmunoprecipitate either w.t. or q.m. CaM (n = 4; unpublished data). In the converse experiments, immunoprecipitation of CaM with an antibody that binds to both w.t. and q.m. forms did not coimmunoprecipitate the InsP₃R (n = 4; unpublished data). These results therefore do not provide evidence of an association between CaM and the InsP₃R.

In summary, our single-channel patch clamp experiments revealed that neither high concentrations of a CaM antagonist, nor overexpression of a Ca²⁺-insensitive q.m. CaM had any effect on [Ca²⁺]_i inhibition of InsP₃R channel gating. In addition, coimmunoprecipitation failed to demonstrate an association between CaM and the InsP₃R. Thus, our investigations did not provide any evidence supporting the hypothesis that high [Ca²⁺]_i inhibition of InsP₃R gating observed in in vitro patch clamp studies is mediated by CaM. These conclusions are therefore in agreement with those reached in some other studies (Zhang and Joseph, 2001; Nosyreva et al., 2002).

Abrogation of Ca²⁺-dependent Inhibition of InsP₃R Channel Gating

Our experimental results suggested that CaM is not involved in the inhibition of InsP₃R channel gating by high [Ca²⁺]_i. However, it remained possible that a different molecule may be involved, and that conditions could be identified which would strip such a putative effector from the InsP₃R in the isolated nucleus, thereby rendering the InsP₃R insensitive to Ca²⁺ inhibition. We reasoned that the putative effector, as a sensor of [Ca²⁺]_i, might be dependent on normal [Ca²⁺]_i for its association with the InsP₃R. We therefore incubated the isolated nuclei in an ultra-low Ca²⁺ bath solution (ULCaS) before using them for nuclear patch clamp experiments to determine the Ca²⁺ dependence of the InsP₃R gating.

In the first set of experiments, nuclei were isolated by protocol L (Fig. 1) into a bath of ULCaS ([Ca²⁺] < 5 nM). In the presence of 10 μM cytoplasmic (pipette) [InsP₃] and [Ca²⁺]_i < 20 μM, gating of the InsP₃R exposed to the ULCaS was very similar to that of InsP₃R in nuclei isolated directly into NCaS by protocol Nd (Fig. 4 ; Mak et al., 1998). In both cases, channel P _o was low (<0.2) in [Ca²⁺]_i < 150 nM, it increased dramatically to 0.8 as [Ca²⁺]_i was increased from 150 nM to 1μM, and then P _o remained at the maximum level of 0.8 when [Ca²⁺]_i was further increased from 1 to 20 μM (Fig. 5) . The InsP₃R in nuclei isolated by protocol Nd were inhibited by [Ca²⁺]_i > 20 μM (Mak et al., 1998) but, remarkably, InsP₃R in nuclei isolated into ULCaS by protocol L exhibited robust channel activities in [Ca²⁺]_i as high as 1.5 mM (Fig. 4) with no decrease in channel P _o (Fig. 5). Thus, a 20-min exposure to the ULCaS containing <5 nM Ca²⁺ caused the gating of InsP₃R channel to be no longer inhibited by high [Ca²⁺]_i. All of the InsP₃R channel activities observed in the ultra-low [Ca²⁺] bath solution also terminated abruptly after ∼30 s, like those previously observed in the regular bath solution (Mak and Foskett, 1994, 1997).

Figure 4.

Typical current traces from nuclei in ULCaS bath isolated by protocol L. Arrows indicate closed channel current levels. Pipette solutions contained 10 μM InsP₃ and [Ca²⁺]_i as tabulated. The last current trace was obtained with −20 mV applied transmembrane potential. Other current traces were obtained with +20 mV applied potential.

Figure 5.

[Ca²⁺]_i dependencies of the channel P _o of the InsP₃R in oocyte nuclei isolated using various protocols (Nd, L, and LNL) and applied potentials (±20 mV) as tabulated. All pipette solutions used contained 10 μM InsP₃. The dashed curve is a simple activating Hill equation fit for the data from nuclei isolated with protocol L (large open circles). For comparison, the biphasic Hill equation fit (continuous curve) for the data points from nuclei isolated directly into NCaS bath (small filled circles) obtained in a previous study (Mak et al., 1998) are also shown. The InsP₃R channel P _o was lower in ULCaS than in NCaS at [Ca²⁺]_i ≈ 100 nM. It is possible that this reflects some intrinsic properties of the InsP₃R after exposure to the low bath [Ca²⁺]. Alternately, this may only be an artifact as a result of the movement of free Ca²⁺ ion across the open channel. With pipette [Ca²⁺]_i ≈ 100 nM, when the oocyte nucleus was in NCaS ([Ca²⁺] = 400–500 nM), the Nernst reversal potential for Ca²⁺ ions was ∼35 mV so Ca²⁺ ions moved across the open InsP₃R channel from the lumenal side to the cytoplasmic side despite an applied transmembrane voltage of 20 mV. This could cause the effective [Ca²⁺]_i at the activating Ca²⁺-binding sites on the cytoplasmic side of the channel to be higher than the free [Ca²⁺] in the bulk of the pipette solution if the Ca²⁺-binding sites are close enough to the ion conducting pore. Conversely, when the nucleus was in ULCaS ([Ca²⁺] < 5 nM), Ca²⁺ ions moved across the open InsP₃R channel in the opposite direction, down the electrical and chemical gradients, possibly lowering the effective [Ca²⁺]_i at the Ca²⁺-binding sites. In [Ca²⁺]_i < 250 nM, the mean open channel duration (<τ_o>) of the InsP₃R increases with [Ca²⁺]_i (Mak and Foskett, 1998). Therefore, if Ca²⁺ flux across the open InsP₃R channel caused the effective [Ca²⁺]_i at the activating Ca²⁺-binding sites to deviate from the free [Ca²⁺] in the bulk pipette solution, then channels in NCaS bath would have longer <τ_o> and higher channel P _o than those in ULCaS bath, as observed. On the other hand, in [Ca²⁺]_i > 300 nM, <τ_o> does not exhibit any dependence on [Ca²⁺]_i although the mean closed channel duration (<τ_c>) is still affected by [Ca²⁺]_i (Mak and Foskett, 1998). Deviation of effective [Ca²⁺] at the Ca²⁺-binding sites from the bulk free [Ca²⁺] would dissipate quickly by diffusion once the channel closed and therefore would not affect <τ_c>. Thus, there would be no difference between the observed P _o of InsP₃R in ULCaS and NCaS bath in [Ca²⁺]_i > 300 nM, as observed.

Because of the difference between the free Ca²⁺ concentration in the high [Ca²⁺]_i pipette solutions and ultra-low [Ca²⁺] bath solutions, it is possible that a potential difference may be established across the membrane and affect the high [Ca²⁺]_i inhibition of the InsP₃R and thus its P _o. We performed patch clamp experiments with −20 mV applied potential, using high [Ca²⁺]_i pipette solution ([Ca²⁺]_i = 221 μM) with nuclei isolated with protocol L. The InsP₃R channel P _d (9 out of 20 patches exhibited channel activity), gating kinetics (last current trace in Fig. 4), and P _o (Fig. 5) were not detectably different from that recorded at +20 mV (P _d = 6 out of 8 patches), indicating that the abrogation of high [Ca²⁺]_i inhibition by exposure to ULCaS is not due to simple electrostatic effects that change the membrane potential.

We previously demonstrated that the Ca²⁺ dependence of channel P _o in nuclei isolated by protocol Nd into NCaS was well fitted by a biphasic Hill equation

(1)

with maximum channel open probability (P _max) = 0.81 ± 0.02, half-maximal activating [Ca²⁺]_i (K _act) = 210 ± 20 nM, activation Hill coefficient (H _act) = 1.9 ± 0.3, half-maximal inhibitory [Ca²⁺]_i (K _inh) = 54 ± 3 μM, and inhibitory Hill coefficient (H _inh) = 3.9 ± 0.7 (Mak et al., 1998). Our new data indicated that the InsP₃R in nuclei isolated by protocol L into ULCaS exhibited no inhibition by high [Ca²⁺]_i, so that the Ca²⁺ dependence of channel P _o can be fitted by a simple activating Hill equation

(2)

with maximum open probability P _max of 0.84 ± 0.01, half-maximal activating [Ca²⁺]_i (K _act) of 280 ± 30 nM, and activation Hill coefficient (H _act) of 2.7 ± 0.3 (Fig. 5).

Nuclei isolated directly into a ULCaS bath by protocol Ld were used to determine the minimum duration of exposure to ULCaS bath required to relieve high [Ca²⁺]_i inhibition of InsP₃R gating. We found that channel activities could be detected with a pipette solution containing 10 μM InsP₃ and 290 μM [Ca²⁺]_i no earlier than 5 min after the nucleus was isolated into the ULCaS bath. Thus, the process involved in the relief of Ca²⁺ inhibition of InsP₃R channel gating by exposure of the isolated nuclei to ULCaS is a slow one, requiring a few minutes.

To determine if normal cytoplasmic [Ca²⁺] (∼50 nM) is low enough to cause the relief of high [Ca²⁺]_i inhibition of InsP₃R gating, we isolated oocyte nuclei directly in PCaS bath (protocol Pd, Fig. 1). In a series of experiments performed in areas of the nuclear membrane identified with very high P _d, using pipette solutions with 10 μM InsP₃ and 0.5 mM ATP, containing alternately 630 nM or 221 μM [Ca²⁺]_i, InsP₃R channels were observed in seven out of seven patches with 630 nM [Ca²⁺]_i, but no InsP₃R channel activity was observed in any of 11 patches with 221 μM [Ca²⁺]_i, even when the nucleus was exposed to the PCaS bath for over 160 min. Thus, the normal resting [Ca²⁺] of the cytoplasm (∼50 nM) is not sufficiently low to induce the relief of Ca²⁺ inhibition observed in the ultra-low Ca²⁺ condition.

InsP₃ Dependence of the InsP₃R in ULCaS Bath

Our previous studies (Mak et al., 1998, 2001a) revealed that InsP₃ activates gating by relieving the Ca²⁺ inhibition of the channel. InsP₃ increases K _inh, the inhibitory half-maximal [Ca²⁺]_i, with no effect on the values of the channel Ca²⁺ activation parameters (K _act, H _act) or P _max in Eq. 1. It seemed likely that this mode of InsP₃ activation cannot operate if the channel is not inhibited by high [Ca²⁺]_i as observed after the channel had been exposed to the ULCaS bath for a few minutes. We therefore examined whether InsP₃ was still required to gate the InsP₃R under conditions that abrogated Ca²⁺ inhibition of the channel.

A series of experiments was performed using nuclei isolated by protocol L into ULCaS bath, patching in regions of the nuclei identified to exhibit high P _d, with pipette solutions alternately containing either 10 μM InsP₃ and [Ca²⁺]_i = 755 nM, or no InsP₃ and [Ca²⁺]_i between 60 nM and 290 μM. Again, the former solution was used to ascertain the presence of functional InsP₃R channels in the regions of the isolated nuclei selected for our experiments for the entire duration of the series. InsP₃R channel activities were observed in 27 out of 30 membrane patches in the presence of InsP₃, but no channel activity was detected in any of the 10 patches without InsP₃ (Fig. 6 A). Therefore, even though the InsP₃R was no longer inhibited by high [Ca²⁺]_i when the nucleus was isolated into ULCaS, InsP₃ was nonetheless still necessary for channel gating.

Figure 6.

(A and B) Typical current traces from nuclei isolated with protocol L. Arrows indicate closed channel current levels. (A) The pipette solutions contained no InsP₃ and 290 μM [Ca²⁺]_i as tabulated. (B) The pipette solutions contained 10 nM InsP₃ and [Ca²⁺]_i as tabulated. (C) [Ca²⁺]_i dependence of the channel P _o of the InsP₃R in the presence of various [InsP₃] as tabulated. The number of channels used to evaluate each of the data points (n) is tabulated next to the corresponding data point. Oocyte nuclei used were isolated using protocol L. The curves are simple activating Hill equation fits (Eq. 2) with the same K _act = 280 nM and H _act = 2.7. The dashed, dotted, and continuous curves have P _max = 0.18, 0.35, and 0.84 for [InsP₃] = 10 nM, 20 nM, and 10 μM, respectively.

Because it seemed paradoxical that InsP₃ activates channel gating by modulating the ability of Ca²⁺ to inhibit the channel, and yet InsP₃R channels that exhibit no high [Ca²⁺]_i inhibition still require InsP₃ for gating, we examined the effects of subsaturating [InsP₃] on channel gating under conditions that abolish high [Ca²⁺]_i inhibition. It was shown previously that in the presence of a subsaturating concentration of InsP₃ (10–33 nM), InsP₃R channels isolated directly into NCaS (protocol Nd) were much more sensitive to Ca²⁺ inhibition than those exposed to higher [InsP₃] (Mak et al., 1998). In contrast, we observed that channels in nuclei isolated into and incubated in ULCaS (protocol L), and activated by subsaturating concentrations of InsP₃ (10–20 nM) exhibited no Ca²⁺ inhibition. Channel activities were observed in 340 μM [Ca²⁺]_i, a normally inhibiting [Ca²⁺]_i, as well as in 4.2 μM [Ca²⁺]_i (Fig. 6 B) with similar channel P _o (Fig. 6 C). Importantly, the maximum P _o observed in subsaturating [InsP₃] was lower than that observed in saturating [InsP₃] (c.f. Fig. 4, [Ca²⁺]_i = 5.5 and 340 μM, and Fig. 6 B). Within the subsaturating range, i.e., [InsP₃] <100 nM, increasing [InsP₃] affected the channel activity mainly by tuning the value of P _max in the simple activating Hill equation (Eq. 2) (Fig. 6 C), instead of affecting the Ca²⁺ inhibitory parameters (K _inh or H _inh), but not P _max in the biphasic Hill equation (Eq. 1), as normally observed in the InsP₃R channel exposed to NCaS (Mak et al., 1998). Thus, the effect of InsP₃ on the InsP₃R channel in ULCaS was dramatically different from that observed in NCaS.

Reversibility of the Regulation by Bath [Ca²⁺] of Ca²⁺ Inhibition of the InsP₃R Channel

It is possible, as we stated before, that the inhibition of InsP₃R gating by high [Ca²⁺]_i is mediated by some molecule that is tightly bound to the InsP₃R in the NCaS bath, and that dissociates from the channel in the presence of extremely low [Ca²⁺] in the ULCaS bath. Dissociation of this putative effector from the InsP₃R channel can then render the channel insensitive to inhibition by high [Ca²⁺]_i. Accordingly, after dissociation, the putative effector molecule could possibly diffuse away into the essentially infinitely large volume of the bath. If this model is correct, the loss of Ca²⁺ inhibition should be irreversible. To explore the reversibility of the loss of Ca²⁺ inhibition, we performed patch-clamp experiments on nuclei isolated from the same batch of oocytes using different isolation/incubation protocols. As described above, Ca²⁺ inhibition was abrogated when the nuclei were isolated into ULCaS bath by protocol L (Fig. 7 A). However, when the nuclei were returned to the NCaS bath for 20 min before patch clamping (protocol LN, Fig. 1), no InsP₃R channel activities were detected at [Ca²⁺]_i = 290 μM (Fig. 7 B) in any of the 11 patches obtained, even though channel gating was observed in 4 out of 5 patches using pipette solutions with [Ca²⁺]_i = 5.5 μM (Fig. 7 C). Thus, despite prior exposure to ULCaS, normal Ca²⁺ inhibition of InsP₃R channel gating was restored when the nuclei were transferred back into NCaS. This restoration of normal Ca²⁺ inhibition was in turn reversible. Reexposure of the nuclei to ULCaS (protocol LNL, Fig. 1) again eliminated normal Ca²⁺ inhibition of gating (Fig. 7 D). The InsP₃R channels in nuclei isolated by protocol LNL exhibited the same P _o (Fig. 5, filled square) as those in nuclei isolated into ULCaS by protocol L without ever being exposed to NCaS (Fig. 5, open circles). These experiments indicated, first, that abolition of Ca²⁺ inhibition of channel gating by exposure of nuclei to ultra-low bath [Ca²⁺] was fully reversible, and second, that it was affected only by the [Ca²⁺] of the bathing solution in which the patch-clamp experiments were performed, independent of the history of bath [Ca²⁺] to which the nuclei were previously exposed. These results suggest either that the sensitivity of Ca²⁺ inhibition of the InsP₃R to the bath [Ca²⁺] is an intrinsic property of the InsP₃R channel, or that it is mediated by some molecule that remains in a stable complex with the channel throughout the multiple transfers of the nucleus into various baths containing different [Ca²⁺].

Figure 7.

Typical current traces obtained from nuclei isolated using different protocols, all from the same batch of oocytes. Arrows indicate closed channel current levels. All pipette solutions contained 10 μM InsP₃. (A) InsP₃R channel activity in 290 μM [Ca²⁺]_i in nuclei isolated by protocol L into ULCaS, n = 5. (B) Absence of InsP₃R channel activity in 290 μM [Ca²⁺]_i in nuclei isolated by protocol LN, n = 11. (C) InsP₃R channel gating in 5.5 μM [Ca²⁺]_i (n = 2) in the same nucleus as used in B. (D) InsP₃R channel activity in 290 μM [Ca²⁺]_i in nuclei isolated by protocol LNL, n = 4.

Is CaM Involved in the Regulation by Bath [Ca²⁺] of Ca²⁺ Inhibition of InsP₃R Channel?

We explored the possible role of CaM in mediating the novel regulation of Ca²⁺ inhibition of InsP₃R gating by the bath [Ca²⁺]. The working hypothesis was that addition of CaM would restore normal inhibition of channel gating by high [Ca²⁺]_i after it had been relieved by exposure to the low [Ca²⁺] bath. Patch-clamp experiments were performed on nuclei isolated by protocol L into ULCaS bath, using a pipette solution containing 10 μM purified CaM with 10 μM InsP₃ and high [Ca²⁺]_i (290 μM). Nevertheless, InsP₃R channel gating was observed in the presence of CaM (Fig. 8 A) that was indistinguishable (P > 0.05, Fig. 8 E) from that observed under the same conditions without CaM (compare Fig. 4, [Ca²⁺]_i = 290 μM). Thus, addition of CaM did not reconstitute normal high [Ca²⁺]_i inhibition of channel gating.

Figure 8.

(A) Typical current traces obtained from nuclei isolated from uninjected oocytes using protocol L with pipette solution containing 10 μM InsP₃, 290 μM [Ca²⁺]_i, and 10 μM purified CaM, n = 4. (B–D) Typical current traces obtained from nuclei isolated from oocytes expressing q.m. CaM with pipette solution containing 10 μM InsP₃ and 290 μM [Ca²⁺]_i. InsP₃R channel activity was observed in 290 μM [Ca²⁺]_i in nuclei isolated by protocol L (B, n = 6) or protocol LNL (D, n = 8), but not in nuclei isolated by protocol LN (C, n = 9). Arrows indicate closed channel current levels. (E) Histogram of InsP₃R channel P _o at 10 μM InsP₃ and 290 μM [Ca²⁺]_i observed in various nuclei under experimental conditions as tabulated.

We also performed a series of patch-clamp experiments on nuclei isolated from oocytes expressing the Ca²⁺-insensitive q.m. CaM. When the nuclei were isolated by protocol L into ULCaS bath, InsP₃R channel activities were observed in high (290 μM) [Ca²⁺]_i (six of eight patches, Fig. 8 B) as frequently as in the normally “permissive” [Ca²⁺]_i between 500 nM and 5.5 μM (seven of eight patches). Furthermore, the channel P _o was the same as that observed in the channels in nuclei isolated by protocol L from uninjected oocytes (P > 0.05, Fig. 8 E). In addition, expression of the q.m. CaM had no effect on the reversibility of the low [Ca²⁺] bath effect. Thus, no InsP₃R channels were detected in any of the nine patches from nuclei isolated from mutant CaM-expressing oocytes by protocol LN (Fig. 8 C). Moreover, high [Ca²⁺]_i inhibition of the channel was still completely abrogated in nuclei isolated from mutant CaM-expressing oocytes by protocol LNL. Thus, channel activities were observed in 290 μM [Ca²⁺]_i (Fig. 8 D) with P _d (8 of 13 patches) similar (P > 0.05) to that in [Ca²⁺]_i between 500 nM and 5.5 μM (seven of eight patches). InsP₃R channel P _o in these nuclei from mutant CaM-expressing oocytes was the same as that observed in nuclei isolated from uninjected oocytes by protocol L or LNL (P from t test was >0.05, Fig. 8 E). Therefore, there were no differences between the Ca²⁺ inhibition (or lack thereof) of InsP₃R channel gating observed in nuclei isolated by various protocols from oocytes overexpressing the mutant CaM and that observed in nuclei isolated from uninjected oocytes, under all experimental conditions.

DISCUSSION

Is There a Role of CaM in Inhibition by High [Ca²⁺]_i of InsP₃R-1 Gating?

Numerous investigations have explored the interactions between the InsP₃R and CaM, but their nature, regulation, and functional effects on intracellular Ca²⁺ signaling are still far from clear. Although it was reported that CaM binding regulates high [Ca²⁺]_i inhibition of the InsP₃R-1 channel (Hirota et al., 1999; Michikawa et al., 1999), subsequent studies using microsomal fluxes or reconstituted channels in lipid bilayers have provided contradictory evidence (Zhang and Joseph, 2001; Nosyreva et al., 2002). In this study, we investigated the possible involvement of CaM in the high [Ca²⁺]_i inhibition of single-channel InsP₃R-1 gating using the nuclear patch clamp method (Mak and Foskett, 1994). This approach enables single-channel recording of endogenous and recombinant InsP₃R channels in their native membrane environment. Similar biphasic regulation by [Ca²⁺]_i of both the endogenous Xenopus type 1 channel and the recombinant rat type 3 InsP₃R channel have been observed in previous nuclear patch clamp studies (Mak et al., 1998, 2001a). In this study, we directly explored the role of CaM in high [Ca²⁺]_i inhibition of InsP₃R channel gating. We have found no evidence to support the hypothesis that inhibition of InsP₃R-1 channel activities by high [Ca²⁺]_i is mediated by direct interaction between the InsP₃R channel and CaM. First, in the presence of 500 μM W-7, a CaM antagonist that was previously reported to alleviate Ca²⁺ inhibition of InsP₃R-1 channels reconstituted into bilayers (Michikawa et al., 1999), the X-InsP₃R-1 was still inhibited by high [Ca²⁺]_i in our nuclear patch-clamping experiments (Fig. 3, A and B). Second, overexpression in oocytes of a dominant-negative, Ca²⁺-insensitive q.m. CaM did not interfere with normal Ca²⁺ inhibition of the X-InsP₃R-1 in the oocyte nuclear envelope (Fig. 3, C and D). Third, addition of CaM (10 μM) to the pipette solution did not reconstitute normal Ca²⁺ inhibition of InsP₃R channel after it was abrogated by exposure of the channel to ULCaS bath (Fig. 8 A). Fourth, overexpression of q.m. CaM in oocytes did not affect the abrogation of Ca²⁺ inhibition by exposure of the channel to ULCaS bath, nor did it affect the restoration of Ca²⁺ inhibition when the channel was placed back in NCaS bath (Fig. 8, B–D). Furthermore, coimmunoprecipitation experiments did not detect any association between CaM and InsP₃R-1 in the Xenopus oocytes. Therefore, whereas CaM may regulate intracellular Ca²⁺ signaling through other mechanisms, our experimental results, together with other recent publications (Zhang and Joseph, 2001; Nosyreva et al., 2002), indicate that it does not regulate inhibition of InsP₃R-1 channel gating by high [Ca²⁺]_i.

What then could be the mechanism of Ca²⁺ inhibition? The simplest hypothesis is that the Ca²⁺ binding sites responsible for Ca²⁺ inhibition of channel gating are contained within the structure of the InsP₃R protein itself. Many regions of the protein have been shown to bind Ca²⁺ in in vitro studies (Sienaert et al., 1996, 1997). One or more of these or as yet unidentified sites may play a role, although there are no data available that address this issue. Alternately, another molecule could perhaps be involved. The InsP₃R interacts with other proteins (Patel et al., 1999; Yang et al., 2002). Of interest, a calmodulin-like protein, CaBP1, interacts with high affinity with the ligand-binding region of the channel (Yang et al., 2002). Whereas it is highly unlikely that CaBP1 and its isoforms mediate Ca²⁺ inhibition, since they are likely neurally restricted and have been shown to stimulate channel gating (Yang et al., 2002), the identification of noncalmodulin Ca²⁺-binding protein interactions with the receptor lends credence to the notion that a Ca²⁺-binding protein could possibly be involved in mediating Ca²⁺ responses of the channel. Because Ca²⁺ inhibition of channel activity has been observed in a number of distinct experimental systems from different species, such a putative effector would need to be ubiquitously expressed and tightly associated with the channel.

A Novel Regulation of [Ca²⁺]_i Inhibition of InsP₃R-1 Channel Gating

Our investigations have revealed a novel CaM-independent regulation of the InsP₃R-1 channel: abrogation of high [Ca²⁺]_i inhibition of InsP₃R-1 channel gating by exposure of the channel to ultra-low bath [Ca²⁺] (<5 nM). The physical location of the low [Ca²⁺]_bath-sensing mechanism on the InsP₃R protein is unknown. The abrogation could possibly be caused by low [Ca²⁺] in the perinuclear space between the inner and outer nuclear envelope, to which the lumenal side of the InsP₃R-1 channel is exposed. In this case, exposure to the ultra-low bath [Ca²⁺] causes the lumenal [Ca²⁺] to fall to low levels due to uncompensated Ca²⁺ leak; and the [Ca²⁺] sensing mechanism responsible for switching high [Ca²⁺]_i inhibition of InsP₃R-1 channel on and off is located on the lumenal side of the InsP₃R channel. The existence of a Ca²⁺-binding site on the lumenal side of the InsP₃R-1 channel has been reported (Sienaert et al., 1996). Our previous studies indicated that the ionic composition of the solution in the perinuclear space of the isolated oocyte nucleus is likely to be similar to that of the bath solution (Mak and Foskett, 1997). The long lag time (∼300 s) between the isolation of the nucleus into the ultra-low [Ca²⁺] bath solution and the earliest detection of InsP₃R channel activities that could not be inhibited by high [Ca²⁺]_i may reflect the time required for the solution in the perinuclear space to become fully equilibrated with the bath solution, or the time taken for Ca²⁺ bound to the lumenal Ca²⁺-binding sites of the InsP₃R channel to dissociate from the sites after the drop in lumenal [Ca²⁺], or a combination of the two.

Alternately, the [Ca²⁺]-sensing mechanism could possibly be located on the cytoplasmic side of the channel. In this case, the long lag time (∼300 s) between exposure of the channel to ultra-low [Ca²⁺] and the abrogation of high [Ca²⁺]_i inhibition would imply that dissociation of Ca²⁺ from the sensing mechanism is slow (rate ∼0.003 s⁻¹). Although such a sensing mechanism would be exposed to high [Ca²⁺] in the pipette solution as soon as the giga-ohm seal was formed, InsP₃R channel activities were nevertheless observed for typically >10 s when the channel was exposed to [Ca²⁺]_i ∼290 μM before the activities abruptly terminated (Mak and Foskett, 1997; Boehning et al., 2001). Thus, binding of Ca²⁺ to the sensing mechanism to restore normal high [Ca²⁺]_i inhibition must also be a very slow process (rate <0.1 s⁻¹). If the [Ca²⁺] sensing mechanism is in equilibrium with the cytoplasmic solution, the forward rate constant (k _f) for Ca²⁺ dissociation from the [Ca²⁺]-sensing mechanism is ≈0.003 s⁻¹ and the reverse rate constant (k _r) is such that 0.1 s⁻¹ ≈ k _r × 290 μM. If the [Ca²⁺]-sensing mechanism is a simple Ca²⁺ binding site, then the equilibrium constant K (k _f/k _r) for Ca²⁺ dissociation from the site should then be ≈10 μM. However, abrogation of channel inhibition was not observed in our normal bath solutions that contain 300–500 nM Ca²⁺ (Mak et al., 1998), or in our physiological Ca²⁺ bath solution containing 50 nM free Ca²⁺. It could only be observed when bath [Ca²⁺] was reduced to very low levels. Thus, if the [Ca²⁺]_bath-sensing mechanism is located on the cytoplasmic side of the channel, it is likely to be a set of cooperative Ca²⁺-binding sites. Further studies are necessary to distinguish whether cytoplasmic or lumenal [Ca²⁺] is being sensed in the disruption of high [Ca²⁺]_i inhibition of the InsP₃R, and to determine the molecular mechanisms involved in that process.

Mechanism of Regulation of High [Ca²⁺]_i Inhibition of InsP₃R-1 Channel Gating by Exposure to Low [Ca²⁺] Bath

A novel allosteric model, developed in the accompanying manuscript, can account for the effect of ultra-low [Ca²⁺] bath exposure on the abrogation of high [Ca²⁺]_i inhibition as well as the effect of InsP₃ to modulate maximum channel P _o, rather than K _inh, under these conditions. In brief, this model accounts for our results by postulating the existence of two functional inhibitory Ca²⁺ binding sites associated with each monomer of the tetrameric channel. One site is only inhibitory when the channel is not liganded with InsP₃, because InsP₃ binding relieves the Ca²⁺ inhibition imposed by this site. In contrast, the properties of the other inhibitory site are not affected by InsP₃ binding. In normal physiological [Ca²⁺]_i conditions, Ca²⁺ binding to this InsP₃-insensitive site provides the observed high [Ca²⁺]_i inhibition (K _inh ∼50 μM) of the fully InsP₃-liganded channel. The ability of this InsP₃-insensitive site to be inhibitory is reversibly lost after exposure of the channel for >5 min to an ultra-low bath [Ca²⁺] (<5 nM). Thus, the observed abrogation of high [Ca²⁺]_i inhibition of channel activity in saturating [InsP₃] can be accounted for by the fact that there is no longer any functional inhibitory Ca²⁺-binding site. On the other hand, in the absence of InsP₃, the InsP₃-sensitive Ca²⁺ inhibition site is functional and keeps the channel closed. Thus, the channel still requires InsP₃ to gate open even when the InsP₃-insensitive site has been disrupted by exposure to ultra-low bath [Ca²⁺]. A detailed description of this model, which can account for these and many other features of ligand regulation of the channel observed in nuclear patch clamp experiments, is developed in the accompanying manuscript (Mak et al., 2003, in this issue).

Are Different Sensitivities to Inhibition by High [Ca²⁺]_i a Fundamental Distinguishing Feature among the InsP₃R Isoforms?

The three isoforms of InsP₃R have complicated patterns of expression in various tissues with complex regulation by various mechanisms (Taylor et al., 1999). Because the permeation and conductance properties of the InsP₃R isoforms are very similar (Mak et al., 2000; Ramos-Franco et al., 2000), differences among the isoforms in localization and channel gating and its regulation are likely to be reasons for the existence of InsP₃R diversity. A review of published single-channel studies of various InsP₃R isoforms suggests that different sensitivities to inhibition by high [Ca²⁺]_i may be one distinguishing functional feature among the various InsP₃R isoforms. Nevertheless, it is not clear whether such differences are intrinsic to the channels, or whether they are perhaps artificially generated by the different experimental protocols used for studying InsP₃R channel activity.

In the presence of ∼1 μM InsP₃, native and recombinant InsP₃R-1 channels (including various splice variants) reconstituted into lipid bilayers exhibited similar strong inhibition by [Ca²⁺]_i with half-maximal inhibitory [Ca²⁺]_i of 0.1–2 μM (Bezprozvanny et al., 1991; Ramos-Franco et al., 1998a,b; Tu et al., 2002), whereas native Xenopus and recombinant rat InsP₃R-1 channels studied in their native membrane environment using nuclear patch clamp techniques exhibited inhibition by high [Ca²⁺]_i, but with a significantly higher half-maximal inhibitory [Ca²⁺]_i of ∼50 μM (Mak et al., 1998; Boehning et al., 2001). When reconstituted into planar bilayers, Ca²⁺ inhibition of InsP₃R-1 could be alleviated by very high [InsP₃] (180 μM) (Kaftan et al., 1997; Moraru et al., 1999), whereas Ca²⁺ inhibition of InsP₃R-1 studied in the native membrane environment was not further affected by [InsP₃] once the channel was saturated with [InsP₃] >100 nM (Mak et al., 1998).

InsP₃R-2 channels reconstituted in lipid bilayers exhibited variable but low sensitivity to inhibition by high [Ca²⁺]_i, with a half-maximal [Ca²⁺]_i of ∼400 μM for recombinant InsP₃R-2 channels (Ramos-Franco et al., 2000) and >1 mM for native channels (Ramos-Franco et al., 1998b, 2000) in 1 μM InsP₃.

Native type 3 InsP₃R channels reconstituted into lipid bilayers exhibited no detectable inhibition by high [Ca²⁺]_i and its P _o remained at its maximum value (∼0.05) in [Ca²⁺]_i between 1 and 100 μM in the presence of 2 μM InsP₃ (Hagar et al., 1998). In marked contrast, recombinant r-InsP₃R-3 in the nuclear membrane of oocytes is inhibited by high [Ca²⁺]_i in an InsP₃-dependent manner very similar to that for X-InsP₃R-1 under identical experimental conditions (Mak et al., 2001a).

How can we account for such divergent results? Our studies here demonstrate that Ca²⁺ inhibition of the X-InsP₃R-1 channel in its native membrane environment can be completely, specifically and reversibly abrogated under certain experimental conditions (after exposure to a nominally Ca²⁺-free bath). Associated with this effect, the InsP₃ dependence of the channel P _o was also changed—normally, InsP₃ affects the apparent affinity of the inhibitory Ca²⁺-binding sites of the channel (Mak et al., 1998), whereas after ULCaS bath exposure, InsP₃ affects the maximum P _o observed (Fig. 6 C). Of note, this InsP₃ dependence of maximum P _o is very similar to the observed effect of InsP₃ on the P _o of InsP₃R-2 channels reconstituted into lipid bilayers (Ramos-Franco et al., 1998b). These observations raise the intriguing possibility that the observed differences in the sensitivities to Ca²⁺ inhibition of various InsP₃R isoforms may be a consequence of the different environment and/or isolation conditions to which the channels were exposed, rather than the result of differences in fundamental intrinsic characteristics of the individual isoforms. For example, the InsP₃R-1 and InsP₃R-3 channel isoforms exhibited very similar inhibition by high [Ca²⁺]_i when they are studied in a native ER membrane environment (Mak et al., 1998, 2001a), but they behaved differently in reconstitution systems. We suggest that it is worth considering the possibility that procedures employed in the isolation and reconstitution and recording of the InsP₃R-3 used in (Hagar et al., 1998) disrupted the normal high [Ca²⁺]_i inhibition of the InsP₃R-3, causing the observed lack of Ca²⁺ inhibition, in very much the same way that exposure to a ULCaS bath abrogated the high [Ca²⁺]_i inhibition of InsP₃R-1 observed in this study. Whereas the procedures used in the isolation and reconstitution and recording of InsP₃R-1 by themselves did not eliminate high [Ca²⁺]_i inhibition of the channel (Bezprozvanny et al., 1991; Ramos-Franco et al., 1998a,b; Tu et al., 2002), they may account for the ability, observed only in the reconstituted systems, of extremely high [InsP₃] to abrogate high [Ca²⁺]_i inhibition (Kaftan et al., 1997; Moraru et al., 1999). By the same token, it is possible that the very low sensitivity to high [Ca²⁺]_i inhibition of the InsP₃R-2 channel isoform reconstituted in lipid bilayers (Ramos-Franco et al., 1998b, 2000) was induced by the isolation and reconstitution and recording protocols. Obviously, these issues will need to be resolved in future studies, for example, of the Ca²⁺ responses of type 2 InsP₃R channels in the native ER membrane environment, under the same experimental conditions as those used for the types 1 and 3 InsP₃R isoforms; and of the sensitivities of Ca²⁺ inhibition of the other InsP₃R isoforms to exposure to ultra-low bath [Ca²⁺]. Nevertheless, our identification in this study of conditions that can radically alter the [Ca²⁺]_i inhibition properties of the channel suggests that careful consideration of the isolation protocols and other conditions to which InsP₃R channels are exposed before they are examined will be warranted in future studies.