Permeant calcium ion feed-through regulation of single inositol 1,4,5-trisphosphate receptor channel gating

Abstract

The ubiquitous inositol 1,4,5-trisphosphate (InsP₃) receptor (InsP₃R) Ca²⁺ release channel plays a central role in the generation and modulation of intracellular Ca²⁺ signals, and is intricately regulated by multiple mechanisms including cytoplasmic ligand (InsP₃, free Ca²⁺, free ATP⁴⁻) binding, posttranslational modifications, and interactions with cytoplasmic and endoplasmic reticulum (ER) luminal proteins. However, regulation of InsP₃R channel activity by free Ca²⁺ in the ER lumen ([Ca²⁺]_ER) remains poorly understood because of limitations of Ca²⁺ flux measurements and imaging techniques. Here, we used nuclear patch-clamp experiments in excised luminal-side-out configuration with perfusion solution exchange to study the effects of [Ca²⁺]_ER on homotetrameric rat type 3 InsP₃R channel activity. In optimal [Ca²⁺]_i and subsaturating [InsP₃], jumps of [Ca²⁺]_ER from 70 nM to 300 µM reduced channel activity significantly. This inhibition was abrogated by saturating InsP₃ but restored when [Ca²⁺]_ER was raised to 1.1 mM. In suboptimal [Ca²⁺]_i, jumps of [Ca²⁺]_ER (70 nM to 300 µM) enhanced channel activity. Thus, [Ca²⁺]_ER effects on channel activity exhibited a biphasic dependence on [Ca²⁺]_i. In addition, the effect of high [Ca²⁺]_ER was attenuated when a voltage was applied to oppose Ca²⁺ flux through the channel. These observations can be accounted for by Ca²⁺ flux driven through the open InsP₃R channel by [Ca²⁺]_ER, raising local [Ca²⁺]_i around the channel to regulate its activity through its cytoplasmic regulatory Ca²⁺-binding sites. Importantly, [Ca²⁺]_ER regulation of InsP₃R channel activity depended on cytoplasmic Ca²⁺-buffering conditions: it was more pronounced when [Ca²⁺]_i was weakly buffered but completely abolished in strong Ca²⁺-buffering conditions. With strong cytoplasmic buffering and Ca²⁺ flux sufficiently reduced by applied voltage, both activation and inhibition of InsP₃R channel gating by physiological levels of [Ca²⁺]_ER were completely abolished. Collectively, these results rule out Ca²⁺ regulation of channel activity by direct binding to the luminal aspect of the channel.

INTRODUCTION

Modulating cytoplasmic free Ca²⁺ concentration ([Ca²⁺]_i) is a universal intracellular signaling pathway that regulates numerous cellular physiological processes including apoptosis, gene expression, bioenergetics, secretion, immune responses, fertilization, muscle contraction, and synaptic transmission (Clapham, 1995; Marks, 1997; Berridge, 1998, 2003; Berridge et al., 2000; Bootman et al., 2001; Orrenius et al., 2003; Braet et al., 2004; Randriamampita and Trautmann, 2004; Cárdenas et al., 2010). Ubiquitous ER-localized inositol 1,4,5-trisphosphate (InsP₃) receptor (InsP₃R) Ca²⁺ release channels (Foskett et al., 2007) play a central role in this pathway in many cells (Taylor and Richardson, 1991; Putney and Bird, 1993; Bezprozvanny and Ehrlich, 1995; Furuichi and Mikoshiba, 1995; Patterson et al., 2004; Foskett et al., 2007; Joseph and Hajnóczky, 2007; Cárdenas et al., 2010; Foskett, 2010). InsP₃ generated in the cytoplasm in response to extracellular stimuli (Berridge, 1993) binds to and activates InsP₃R channels to release Ca²⁺ stored in the ER lumen into the cytoplasm, generating diverse local and global [Ca²⁺]_i signals (Berridge, 1993, 1997; Hagar and Ehrlich, 2000; Thrower et al., 2001; Taylor and Laude, 2002; Foskett et al., 2007). Whereas much is known regarding the intricate regulation of InsP₃R channel gating by multiple processes—binding of cytoplasmic ligands (Ca²⁺, InsP₃, and ATP⁴⁻), posttranslational modifications, interactions with proteins, clustering, differential localization (Joseph, 1996; MacKrill, 1999; Patel et al., 1999; Johenning and Ehrlich, 2002; Foskett et al., 2007; Betzenhauser et al., 2008; Kang et al., 2008; Wagner et al., 2008; Li et al., 2009; Taufiq-Ur-Rahman et al., 2009)—the regulation of InsP₃R channel activity by free Ca²⁺ in the lumen of the ER ([Ca²⁺]_ER) remains poorly understood and controversial (Irvine, 1990; Tregear et al., 1991; Ferris et al., 1992; Swillens, 1992; Kindman and Meyer, 1993; Bezprozvanny and Ehrlich, 1994; Bootman, 1994a,b; Swillens et al., 1994; Shuttleworth, 1995; Dupont and Swillens, 1996; Missiaen et al., 1996; Parys et al., 1996; Beecroft and Taylor, 1997; Caroppo et al., 2003; Dawson et al., 2003; Fraiman and Dawson, 2004; Foskett et al., 2007; McCarron et al., 2008; Yamasaki-Mann and Parker, 2011).

The main techniques for studying possible [Ca²⁺]_ER modulation of InsP₃R channel gating have been ⁴⁵Ca²⁺ flux measurements (Nunn and Taylor, 1991, 1992; Tregear et al., 1991; Missiaen et al., 1992a,b; Parys et al., 1993; Beecroft and Taylor, 1997) and fluorescence Ca²⁺ imaging (Combettes et al., 1992, 1996; Missiaen et al., 1992c; Shuttleworth, 1992; Renard-Rooney et al., 1993; Short et al., 1993; Steenbergen and Fay, 1996; Tanimura and Turner, 1996; Tanimura et al., 1998; Caroppo et al., 2003; Higo et al., 2005; McCarron et al., 2008; Yamasaki-Mann and Parker, 2011). Both approaches rely on changes in [Ca²⁺]_i or [Ca²⁺]_ER to infer channel activity and therefore cannot rigorously control both [Ca²⁺]_ER and [Ca²⁺]_i simultaneously during experiments. This has made it difficult to differentiate direct effects of [Ca²⁺]_ER on the luminal aspect of the InsP₃R from feed-through effects caused by Ca²⁺ flux through the open channel.

Electrophysiological recordings of single InsP₃R channels allow InsP₃R channel activity to be determined from currents carried by K⁺ through open channels (Foskett et al., 2007) and therefore can be performed under rigorously controlled and defined [Ca²⁺]_i and [Ca²⁺]_ER. However, only two electrophysiological studies of the effects of [Ca²⁺]_ER on InsP₃R channel activity have been reported (Bezprozvanny and Ehrlich, 1994; Thrower et al., 2000). Both used InsP₃R reconstituted in lipid bilayers and explored only a limited set of [Ca²⁺]_ER, [Ca²⁺]_i, and [InsP₃]. Insights about [Ca²⁺]_ER regulation of InsP₃R channels from these studies were limited by insufficient buffering of [Ca²⁺]_i and the use of nonphysiological concentrations of divalent cations in the luminal solutions in Bezprozvanny and Ehrlich (1994), or inappropriate Ca²⁺ buffering and nonphysiological [KCl] used in Thrower et al. (2000).

Here, we studied systematically the effects of [Ca²⁺]_ER on the activity of single homotetrameric channels of recombinant rat type 3 InsP₃R (r-InsP₃R-3) expressed in cells with no endogenous InsP₃R expression (Sugawara et al., 1997; Mak et al., 2005), using nuclear patch-clamp techniques (Mak et al., 2005; Vais et al., 2010a) that record activities of the InsP₃R channel in its native membrane milieu (Foskett et al., 2007) with ionic conditions on both sides of the channel, especially [Ca²⁺]_i and [Ca²⁺]_ER, rigorously controlled. By comparing activities (open probability [P_o]) of type 3 InsP₃R channels in excised luminal-side-out (lum-out) nuclear membrane patches exposed to different [Ca²⁺]_ER by rapid perfusion solution exchange (Vais et al., 2010a), we found that high [Ca²⁺]_ER modulated InsP₃R channel activity. However, our experiments ruled out [Ca²⁺]_ER modulation of InsP₃R channel activity through intrinsic functional Ca²⁺-binding sites on the luminal side of the channel. Instead, the experimental results were consistent with [Ca²⁺]_ER affecting InsP₃R channel gating solely through the rise in local [Ca²⁺]_i in the vicinity of the open channel generated by the [Ca²⁺]_ER-driven Ca²⁺ flux through the open channel itself, which modulates InsP₃R channel activity through functional cytoplasmic Ca²⁺-binding sites of the channel.

MATERIALS AND METHODS

Nucleus isolation and nuclear patch-clamp electrophysiology

Generation and maintenance of DT40-KO-r-InsP₃R-3 cells (mutant cells derived from chicken B cells with the endogenous genes for all three InsP₃R isoforms knocked out and then stably transfected to express recombinant r-InsP₃R-3) were described in Mak et al. (2005). Nuclear patch-clamp experiments were performed using nuclei isolated from DT40-KO-r-InsP₃R-3 cells as described previously (Mak et al., 2005). Experiments investigating InsP₃R activity under constant ligand conditions were performed in the on-nucleus configuration (Mak et al., 2007). Excised nuclear membrane patches in the lum-out configuration were obtained from isolated nuclei (Mak et al., 2007) using protocols analogous to those used to obtain inside-out excised patches in plasma membrane patch-clamp experiments. The solution around the excised nuclear membrane patch was rapidly switched multiple times using a solution-switching setup described in Mak et al. (2007).

InsP₃R channel current traces were acquired at room temperature as described previously (Mak et al., 1998), digitized at 5 kHz, and anti-aliasing filtered at 1 kHz. All applied potentials (V_app) were measured relative to the bath electrode. All on-nucleus experiments were performed at V_app = −40 mV. All lum-out experiments were performed at V_app = −30 mV unless stated otherwise.

Experimental solution composition

All experimental solutions contained 140 mM KCl and 10 mM HEPES, pH to 7.3 with KOH. Because physiological levels of free Mg²⁺ (0–3 mM) have no effects on channel activities (Mak et al., 1999), and to avoid the complicating effects of Mg²⁺ on InsP₃R channel conductance (Mak and Foskett, 1998) and free [ATP] in experimental solutions, Mg²⁺ was not added to any of the solutions used.

All experiments were performed using the same bath solution with free Ca²⁺ concentration ([Ca²⁺]_f) of 70 nM buffered by 0.5 mM BAPTA (1,2-bis(o-aminophenoxy) ethane-N,N,N′,N′-tetraacetic acid). All pipette solutions contained 0.5 mM Na₂ATP.

Pipette solutions used in on-nucleus patch-clamp experiments contained various [Ca²⁺]_f buffered by 0.5 mM Ca²⁺ chelator and [InsP₃] as specified. BAPTA was used for 20 nM < [Ca²⁺]_f < 600 nM, diBrBAPTA (5,5′-dibromo BAPTA) for 600 nM ≤ [Ca²⁺]_f ≤ 4 µM, and hydroxyethylethylenediaminetriacetic acid (HEDTA) for [Ca²⁺]_f > 4 µM. ATP contributed to Ca²⁺ buffering in solutions with [Ca²⁺]_f > 30 µM (Bers et al., 2010).

Pipette solutions used in lum-out patch-clamp experiments had [Ca²⁺]_f of either 2 µM buffered by various concentrations of diBrBAPTA or HEDTA as specified, or 55 nM buffered by various concentrations of BAPTA as specified. Four perfusion solutions were used in these experiments: one had [Ca²⁺]_f of 70 nM buffered by 0.5 mM BAPTA with no ATP; one contained 1.3 mM CaCl₂ and 1.5 mM Na₂ATP, so that [Ca²⁺]_f was buffered to 300 µM according to Max Chelator freeware; and two with no Ca²⁺ chelator, with one containing 1 mM CaCl₂ and the other containing 2 mM CaCl₂, giving [Ca²⁺]_f of 550 µM and 1.1 mM, respectively, according to activity coefficient calculations (Butler, 1968; Vais et al., 2010a).

[Ca²⁺]_f in all solutions (<100 µM) was confirmed by Ca²⁺-sensitive dye fluorimetry.

Data analysis

Because there is no Ca²⁺ flux through open InsP₃R channels in steady-state on-nucleus experiments using bath solutions containing only 70 nM [Ca²⁺]_f, there was no possibility of cross talk between channels. Furthermore, we detected no effect of channel clustering on gating of InsP₃R channels expressed in DT40-KO-r-InsP₃R-3 cells (Vais et al., 2011). Thus, multi-channel and single-channel current traces were selected for channel P_o and dwell-time analysis using QuB software (Qin et al., 2000a,b). However, because substantial Ca²⁺ flux can be driven through open InsP₃R channels when some perfusion solutions were used in excised lum-out nuclear patch-clamp experiments, only single-channel current traces from such experiments were selected for analysis to avoid complications arising from Ca²⁺ flux through one InsP₃R channel affecting gating behaviors of neighboring active channel(s). We accepted only current records long enough to allow the number of active channels observed to be accurately determined (with >99% confidence) for data analysis (Mak et al., 2001b; Ionescu et al., 2006; Vais et al., 2010b). Only single-channel current traces >15 s were used for channel P_o > 0.02. Longer traces were required for lower P_o, so only single-channel traces >1.5 min were used for P_o ≈ 0.005. Least-square fitting and statistical analysis of data were done with IGOR Pro software (WaveMetrics).

Modeling of the [Ca²⁺]_i profile in the vicinity of an InsP₃R channel

The [Ca²⁺]_i profile around an open InsP₃R channel was calculated by considering the open channel as a circular aqueous pore with a diameter of 2.5 nm (Jiang et al., 2002; Serysheva et al., 2003; Wolfram et al., 2010). Magnitudes of the Ca²⁺ current through an open channel (i_Ca) in the presence of 140 mM KCl and various [Ca²⁺]_i, [Ca²⁺]_ER, and V_app were evaluated using the general Goldman–Hodgkin–Katz current equation (Lewis, 1979; Hille, 2001):

$$i_{\text{Ca}}=P_{\text{Ca}}{z}_{\text{Ca}}^2\frac{F^2{V}_{\text{app}}}{RT}\left\{\frac{{\text{[Ca]}}_{\text{i}}-{\text{[Ca]}}_{\text{ER}}\text{exp}\left(-z_{\text{Ca}}{\mathit{FV}}_{\text{app}}/RT\right)}{1-\text{exp}\left(-z_{\text{Ca}}F{V}_{\text{app}}/RT\right)}\right\},$$ (1)

where $z_{\text{Ca}}$ (= 2) is the valence of Ca²⁺, F is the Faraday constant, R is the gas constant, T is the absolute temperature, and P_Ca is the effective channel permeability for Ca²⁺ through the InsP₃R channel measured experimentally in buffers containing 140 mM KCl (Vais et al., 2010a).

Although the analytical equation obtained by using the approximation in Neher (1986, 1998), Smith (1996), and Naraghi and Neher (1997) can provide a reasonable estimation of the steady-state [Ca²⁺]_i profile around a Ca²⁺ channel generated by Ca²⁺ flux through the channel buffered by mobile Ca²⁺-binding chelators (Liu et al., 2010; Vais et al., 2010a), it says nothing about the time scale of the evolution of the [Ca²⁺]_i profile to reach the steady state. These dynamics then affect the dissipation of the Ca²⁺ profile after the channel closes. Once a channel closes, the profile collapses first to a level that is above basal very quickly but then dissipates toward basal on a slow time scale. Therefore, to follow the time-dependent evolution of the [Ca²⁺]_i profiles around the channel as it gates, numerical modeling of Ca²⁺ diffusion around the open InsP₃R channel was used to calculate the [Ca²⁺]_i profiles in the vicinity of the channel under various [Ca²⁺]_i, [Ca²⁺]_ER, and V_app combinations.

In the simulations, [Ca²⁺]_i is controlled by spatial diffusion, with the rate equation for [Ca²⁺]_i at distance r from the channel at time t after the channel opens, C(r,t), given as

$$\frac{\partial C(r,t)}{\partial t}=D_C{\nabla }^{2}C+J+\frac{\partial b}{\partial t},$$ (2)

where J is the Ca²⁺ flux passing through the channel from the ER lumen, D_C is the diffusion coefficient of Ca²⁺ in the medium (= 800 µm²s⁻¹ [Cussler, 1997] for aqueous medium), and b is the free Ca²⁺ buffer concentration.

For mobile Ca²⁺ buffers, $\partial b/\partial t$ is given as

$$\frac{\partial b}{\partial t}=D_b{\nabla }^{2}b+k_{\text{off}}\left(B-b\right)-k_{\text{on}}Cb,$$ (3)

where B is the total Ca²⁺ buffer concentration, and k_on and k_off are the rates of Ca²⁺ binding to and dissociating from the buffer, respectively, so that k_off/k_on = K_d (dissociation constant) of the buffer. D_b is the diffusion coefficient of the mobile buffer.

K_d for diBrBAPTA, HEDTA, and BAPTA is 1.6 µM, 4.7 µM, and 180 nM, respectively. k_on for diBrBAPTA, HEDTA, and BAPTA is 450, 4.5, and 450 µM⁻¹s⁻¹, respectively (Tsien, 1980; Naraghi, 1997). D_b for diBrBAPTA, HEDTA, and BAPTA is 296, 319, and 390 µm²s⁻¹, respectively, estimated from $D_b=D_C\sqrt[3]{M_C/M_b},$ where M_b and M_C are the molecular weights of the buffer and Ca²⁺, respectively.

i_Ca in Eq. 1 is converted into J in Eq. 2 by

$$J=\{\begin{array}{ccc}{i}_{\text{Ca}}/\left(F{z}_{\text{Ca}}\delta V\right)& \text{for}& r\le r_{\text{ch}}\\ 0& \text{for}& r>r_{\text{ch}}\end{array},$$ (4)

where r_ch is the radius of the channel pore. For simplicity, the channel is considered to be embedded in an infinite membrane, opening into a semi-infinite cytoplasmic volume. Then, δV is the volume of a hemisphere with radius r_ch over the channel. Propagation of Ca²⁺ and mobile buffer (if present) was simulated throughout a homogeneous semi-infinite 3-D cytosolic space.

With spherical symmetry around the channel, the Laplacian of C and b in spherical coordinates is

$${\nabla }^{2}X(r,t)=\frac{1}{r^2}\frac{\partial }{\partial r}\left(r^2\frac{\partial X}{\partial r}\right),$$ (5)

where X = C or b.

The differential Eqs. 2 and 3 were solved implicitly with a spatial grid size of 0.625 nm using the Tridiagonal Matrix Solver software in a hemispherical volume of a large radius of 10 µm, so that the volume was effectively semi-infinite. The channel opened at t = 0 and remained open for the duration of the simulation. Calcium profiles were evolved with 0.1-µs time steps.

To simulate the collapse of the [Ca²⁺]_i profile after channel closure, the [Ca²⁺]_i profile was allowed to evolve for 50 ms after the channel opened. Then, J was set equal to 0 for all r at t = 0 as the channel closed. The [Ca²⁺]_i profile was then evolved using the same software with the same time steps for various durations.

RESULTS

Dependence of steady-state InsP₃R-3 channel activity on cytoplasmic ligands

To investigate the effects of ER luminal Ca²⁺ concentration ([Ca²⁺]_ER) on InsP₃R-3 channel gating, we first performed on-nucleus patch-clamp experiments on nuclei isolated from DT40-KO-r-InsP₃R-3 cells to characterize the gating behaviors of single recombinant homotetrameric r-InsP₃R-3 channels over a wide range of [InsP₃] and [Ca²⁺]_i in the presence of a physiological level (5 mM) of free ATP, which supports InsP₃R channel gating (Mak et al., 1999, 2001a). In a bath solution with [Ca²⁺]_f = 70 nM and no MgATP to support activity of the SERCA in the outer nuclear membrane, [Ca²⁺]_f in the perinuclear space of the isolated nuclei (equivalent to the ER lumen topologically) equilibrated with that of the bath solution so that [Ca²⁺]_ER = 70 nM in these experiments. The channel with large conductance (∼550 pS in 140 mM KCl) observed in the outer nuclear membrane of nuclei isolated from DT40-KO-r-InsP₃R-3 cells (Fig. 1) was identified as recombinant InsP₃R-3 channel by its requirement of InsP₃ for activation (Cheung et al., 2008; Vais et al., 2010a).

Figure 1.

Ligand dependence of gating of single r-InsP₃R-3 channels. V_app = −40 mV. (A) Typical single-channel on-nucleus patch-clamp current traces of InsP₃R-3 channels in suboptimal (190 nM), optimal (2 µM), and inhibitory (23 µM) [Ca²⁺]_i in the presence of saturating (10 µM) [InsP₃], demonstrating biphasic [Ca²⁺]_i dependence of InsP₃R-3 channel activity. Arrow indicates closed-channel baseline current level for these and all subsequent current traces. (B) Typical single-channel on-nucleus patch-clamp current traces in the presence of subsaturating (3 µM) [InsP₃] showing that [InsP₃] reduction has little effect on channel activity at suboptimal (190 nM) [Ca²⁺]_i but increases channel sensitivity to Ca²⁺ inhibition, so channel activity is substantially decreased at [Ca²⁺]_i = 2 µM. (C) [Ca²⁺]_i dependence of mean channel P_o in various [InsP₃] as tabulated. Error bars show the SEM in this and all subsequent figures unless stated otherwise. The number of current traces analyzed for each data point is tabulated next to the data point in the same color. Curves are empirical biphasic Hill equation fits to mean P_o data points for various [InsP₃] with the same P_max, K_act, H_act, and H_inh. The purple inset shows the dependence of the K_inh on [InsP₃]. Error bars here show the estimates of fitting errors of K_inh derived from the biphasic Hill equation fits. The curve is the empirical simple Hill equation fit of the [InsP₃] dependence. (D and E) [Ca²⁺]_i dependence of mean open and closed durations of InsP₃R channel in various [InsP₃], derived from the same experimental data used in C. Data points in the same [InsP₃] are connected with lines for clearer presentation.

In the presence of high (10 µM) [InsP₃], the InsP₃R-3 channel exhibited biphasic dependence on [Ca²⁺]_i, with channel P_o increasing as [Ca²⁺]_i increased until a maximum P_o was reached at [Ca²⁺]_i of ∼2–6 µM. Further increases in [Ca²⁺]_i inhibited channel P_o (Fig. 1, A and C).

A further increase in [InsP₃] (100 µM) did not change the gating of InsP₃R-3 appreciably, indicating that the channel was saturated by 10 µM InsP₃ in [Ca²⁺]_i between 1 and 20 µM (Fig. 1 C). Reduction in [InsP₃] below 10 µM did not affect Ca²⁺ activation of the channel significantly but substantially enhanced the sensitivity of the channel to [Ca²⁺]_i inhibition so the channel was inhibited at lower [Ca²⁺]_i (Fig. 1, B and C). This reduced both the maximum channel P_o observed and the range of [Ca²⁺]_i over which the channel gated appreciably in low [InsP₃] (notably for [InsP₃] = 1 µM in Fig. 1 C).

The [Ca²⁺]_i dependence of InsP₃R-3 channel P_o in all [InsP₃] can be well fitted simultaneously for all [InsP₃] investigated by the biphasic Hill equation (Foskett et al., 2007) (Fig. 1 C):

$$P_{\text{o}}=P_{\text{max}}{\left\{1+{\left(\raisebox{1ex}{$K_{\text{act}}$}\!\left/ \!\raisebox{-1ex}{${\left[{\text{Ca}}^{2+}\right]}_{\text{i}}$}\right.\right)}^{H_{\text{act}}}\right\}}^{-1}{\left\{1+{\left(\raisebox{1ex}{${\left[{\text{Ca}}^{2+}\right]}_{\text{i}}$}\!\left/ \!\raisebox{-1ex}{$K_{\text{inh}}$}\right.\right)}^{H_{\text{inh}}}\right\}}^{-1},$$ (7)

with four of the five parameters retaining the same values: P_max = 1, K_act = 940 nM, H_act (Hill coefficient for Ca²⁺ activation) = 1.3, and H_inh (Hill coefficient for Ca²⁺ inhibition) = 1.6. Only K_inh varies with [InsP₃] (inset in Fig. 1 C). P_max from the Hill equation fit of the P_o data is significantly greater than the maximum P_o (0.78) observed in saturating [InsP₃]. This indicates that even in saturating [InsP₃], the recombinant InsP₃R-3 channel in DT40-KO-r-InsP₃R-3 cells is not fully activated by [Ca²⁺]_i before it begins to be inhibited by [Ca²⁺]_i. Because the [Ca²⁺]_i dependence of the InsP₃R-3 channel does not exhibit a clear plateau with channel P_o staying at P_max over a broad range of [Ca²⁺]_i, K_act and K_inh are not uniquely defined by the observed [Ca²⁺]_i dependence (Foskett et al., 2007). Nevertheless, because P_max must be ≤1, the large observed maximum P_o (≈0.78) indicates that the values of K_act and K_inh derived from the biphasic Hill equation fit are reasonable indications of the apparent affinities of the activating and inhibitory cytoplasmic Ca²⁺-binding sites of the channel. Thus, the biphasic Hill equation fitting result that only K_inh depends on [InsP₃] indicates that InsP₃ modulates InsP₃R-3 gating solely by changing the sensitivity of the channel to Ca²⁺ inhibition (Mak et al., 1998; Foskett et al., 2007). Both Hill coefficients for Ca²⁺ activation and inhibition are moderately >1, suggesting that both Ca²⁺ activation and inhibition are cooperative but not strongly so.

The dependence of K_inh on [InsP₃] is well described by a simple activating Hill equation (inset in Fig. 1 C):

$$K_{\text{inh}}=K_{\text{inh}}^{\mathrm{\infty }}{\left\{1+{\left(K_{\text{InsP3}}/{\text{[InsP}}_{\text{3}}]\right)}^{H_{\text{InsP3}}}\right\}}^{-1},$$

with $K_{\text{inh}}^{\mathrm{\infty }}$ (K_inh in saturating [InsP₃]) ≈ 13 µM and K_InsP3 (half-maximal [InsP₃]) ≈ 4.5 µM. H_InsP3 (Hill coefficient for modulation of K_inh by [InsP₃]) is ∼2.3. This suggests that InsP₃ modulation of InsP₃R-3 channel activity is strongly cooperative.

These main features of ligand regulation of the activity of homotetrameric recombinant r-InsP₃R-3 channel in DT40-KO-r-InsP₃R-3 cells are highly reminiscent of those of a variety of InsP₃R channels in different cell systems examined using the same approach: endogenous Xenopus laevis type 1 InsP₃R (InsP₃R-1) channel in oocytes, recombinant r-InsP₃R-3 channel expressed in Xenopus oocytes, and endogenous insect InsP₃R channel in Sf9 cells (Foskett et al., 2007). However, the r-InsP₃R-3 channel in DT40-KO-r-InsP₃R-3 cells is significantly less sensitive to [Ca²⁺]_i and [InsP₃] activation than the other channels, with significantly higher K_act and K_InsP3. Furthermore, the efficacy of InsP₃ to activate the channel by reducing its sensitivity to inhibition by Ca²⁺_i is also lowest for r-InsP₃R-3 channel in DT40-KO-r-InsP₃R-3 cells among InsP₃R channels studied (Foskett et al., 2007), as indicated by its low $K_{\text{inh}}^{\mathrm{\infty }}$.

Regulation of r-InsP₃R-3 channel gating kinetics by cytoplasmic ligands

The [Ca²⁺]_i dependence of the mean open duration (t_o) of r-InsP₃R-3 channels expressed in DT40-KO-r-InsP₃R-3 cells loosely mirrors that of channel P_o, with t_o continuously increasing as [Ca²⁺]_i was increased from 40 nM to ≈1 µM. Within this [Ca²⁺]_i range, t_o was similar in the same [Ca²⁺]_i for all [InsP₃]. t_o then remained high for a range of [Ca²⁺]_i extending beyond the point where channel P_o started to be reduced by higher [Ca²⁺]_i. Beyond a threshold [Ca²⁺]_i, t_o started to be reduced by higher [Ca²⁺]_i. The threshold [Ca²⁺]_i was ≈20 µM in 10 µM InsP₃ and ≈6 µM in 3 µM InsP₃. Thus, the threshold [Ca²⁺]_i decreased as [InsP₃] was reduced (Fig. 1 D).

Ligand regulation of the mean channel closed duration (t_c) is more complex. In saturating [InsP₃], t_c decreased continuously as [Ca²⁺]_i was increased up to 1 µM, when it reached a minimum of ∼10 ms at [Ca²⁺]_i ≈ 900 nM, before P_o attained its maximum value. t_c remained at the minimum value as [Ca²⁺]_i was increased to 6 µM. Then, t_c increased when P_o started to be reduced by rising [Ca²⁺]_i. Unlike t_o, t_c was increased in all [Ca²⁺]_i as [InsP₃] was reduced below saturating levels, except at very low [Ca²⁺]_i (∼40 nM). As [Ca²⁺]_i was increased, t_c followed a trend that is the inverse of that of P_o, decreasing until t_c reached its minimum at the [Ca²⁺]_i at which P_o was maximal, and then increasing as P_o decreased (Fig. 1 E).

These ligand dependencies of the gating characteristics (t_o and t_c) of r-InsP₃R-3 in DT40-KO-r-InsP₃R-3 cells are markedly more complex than those of other InsP₃R channels examined (Mak et al., 1998, 2001b; Ionescu et al., 2006). For the other channels, t_o remained effectively constant over most of the ranges of [InsP₃] and [Ca²⁺]_i examined, so that once a channel opens, the duration for which it remains open is largely independent of the local [Ca²⁺]_i and [InsP₃]. Ligand modulations of P_o of those channels therefore result mostly from ligand modulations of t_c.

Effects of physiological levels of [Ca²⁺]_ER on InsP₃R-3 channel gating

To study possible modulation of InsP₃R-3 channel gating by [Ca²⁺]_ER, nuclear patch-clamp experiments in the excised lum-out configuration were performed on nuclei from DT40-KO-r-InsP₃R-3 cells. The luminal side of InsP₃R channels in the isolated nuclear membrane patches was exposed rapidly and repeatedly to solutions containing different [Ca²⁺]_f using rapid perfusion solution exchange (Vais et al., 2010a,b). To avoid the complication of Ca²⁺ moving through one active InsP₃R channel affecting the activity of neighboring active channels by raising local [Ca²⁺]_i, only single-channel current traces obtained in these lum-out nuclear patch-clamp experiments were used for analysis.

To detect possible activating or inhibitory effects of [Ca²⁺]_ER on InsP₃R channel gating, we first used a pipette solution containing subsaturating (3 µM) [InsP₃] and optimal (2 µM) [Ca²⁺]_i buffered by 0.5 mM diBrBAPTA. Active InsP₃R-3 channels were exposed alternately to [Ca²⁺]_ER of 70 nM, a sub-physiological [Ca²⁺]_ER in which no Ca²⁺ flux flowed from the luminal side of the channels to the cytoplasmic side, and 300 µM, a physiological [Ca²⁺]_ER that drives substantial Ca²⁺ flux through the channels (Vais et al., 2010a). Increasing [Ca²⁺]_ER from 70 nM to 300 µM caused a reduction in the magnitude of the current passing through the active InsP₃R-3 channels (Fig. 2 A) because Ca²⁺ acts as permeant channel blocker, reducing K⁺ conductance of the channel (Vais et al., 2010a). Returning [Ca²⁺]_ER to 70 nM restored the channel current magnitude (Fig. 2 B). Such changes in channel current size were used in all experiments to mark the time when the perfusion solution exchange was completed at the luminal side of the active InsP₃R channels.

Figure 2.

Effects of [Ca²⁺]_ER on InsP₃R-3 channel activity in various [InsP₃]. (A–C) Typical single-channel current traces from excised lum-out nuclear membrane patches recorded during a rapid switch of [Ca²⁺]_ER by perfusion solution exchange. For clarity, compositions of pipette solutions ([Ca²⁺]_i, concentration and nature of Ca²⁺ chelator used, [InsP₃]) common to all experiments presented in this figure (current traces and bar graphs) are tabulated at the top of the figure. Pipette solution composition(s) specific to each current trace is tabulated at the top of the corresponding current trace. In each current trace, V_app used is tabulated at left, color bars at the top indicate [Ca²⁺]_ER in the perfusion solutions, and blue bars at the bottom indicate segments used for evaluating the channel P_o tabulated below. In these figures, short current segments were plotted to show the InsP₃R channel gating more clearly. The mean P_o, t_o, and t_c ratios (discussed in Results and Discussion, and shown in bar graphs) were derived from current segments that are significantly longer to ensure that only single-channel nuclear patch current traces were used (see Materials and methods). (D–F) Bar graphs of mean ratios of channel P_o, t_o, and t_c, respectively, observed before and after [Ca²⁺]_ER switching between 70 nM and 300 µM for different [InsP₃]. Numbers beside the bars are the number of experiments (top) and perfusion solution switches (bottom) analyzed. ***, **, and * mark significant deviation of a ratio from unity (P < 0.005, 0.01, and 0.05, respectively; paired t test). These symbols and conventions are also used for Figs. 3–6, and 8 and 9. For clarity, bars of the same color in these figures correspond to the same set of data obtained under the same experimental conditions.

Besides the reduction in channel conductance, the jump in [Ca²⁺]_ER from 70 nM to 300 µM caused a significant decrease in channel P_o (Fig. 2 A), which was reversed when [Ca²⁺]_ER was returned to 70 nM (Fig. 2 B). This effect of [Ca²⁺]_ER on channel P_o was quantified by the ratio of the P_o observed when [Ca²⁺]_ER = 300 µM (P_o(300 µM)) to that when [Ca²⁺]_ER = 70 nM (P_o(70 nM)), evaluated immediately before and after each perfusion solution switch (Fig. 2, A and B). With [InsP₃] = 3 µM and [Ca²⁺]_i = 2 µM, the InsP₃R channel was significantly less active in [Ca²⁺]_ER = 300 µM than in [Ca²⁺]_ER = 70 nM, so the mean P_o ratio ($\langle P_{\text{o}}\left(300\mu \text{M}\right)/P_{\text{o}}\left(70\text{nM}\right)\rangle ;$ the angle brackets are used to emphasize that this is the mean of the P_o ratios, not the ratio of the mean P_o’s at different [Ca²⁺]_ER) was significantly lower than unity (Fig. 2 D, red bar). The change in [Ca²⁺]_ER also caused significant changes in both t_o and t_c, so the mean t_o and t_c ratios ($\langle t_{\text{o}}(300\mu \text{M})/t_{\text{o}}\left(70\text{nM}\right)\rangle$ and $\langle t_{\text{c}}(300\mu \text{M})/t_{\text{c}}\left(70\text{nM}\right)\rangle$, respectively) are both significantly different from unity (Fig. 2, E and F, red bars).

The observed inhibition of InsP₃R gating by elevated [Ca²⁺]_ER can be caused by luminal free Ca²⁺ (Ca²⁺_ER) binding to an inhibitory site on the luminal side of the channel. Alternatively, it could be caused by Ca²⁺ binding to cytoplasmic sites on the InsP₃R channel because of local [Ca²⁺]_i in the vicinity of the pore elevated by Ca²⁺ flux driven through the channel by the high [Ca²⁺]_ER. Because [Ca²⁺]_i in the pipette solution used (2 µM) optimally activates InsP₃R channels in 3 µM InsP₃, a rise in local [Ca²⁺]_i near the channel is predicted to reduce P_o caused by Ca²⁺ binding to the inhibitory Ca²⁺-binding sites on the cytoplasmic side of the channel (Foskett et al., 2007).

[Ca²⁺]_ER inhibition of InsP₃R channel activity is abolished by saturating [InsP₃] but restored by higher [Ca²⁺]_ER

To better characterize the observed [Ca²⁺]_ER regulation of InsP₃R channel activity, we investigated the [InsP₃] dependence of the effect of [Ca²⁺]_ER by using pipette solutions containing different [InsP₃] in our lum-out nuclear patch-clamp experiments. Inhibition of InsP₃R channel gating by raising [Ca²⁺]_ER from 70 nM to 300 µM was abrogated in the presence of saturating [InsP₃] (10 µM) (Fig. 2 C), so the mean P_o, t_o, and t_c ratios observed were not significantly different from unity (Fig. 2, D–F, gray bars). In contrast, even in saturating [InsP₃], InsP₃R channel activity was still substantially inhibited when [Ca²⁺]_ER was raised to higher (1.1 mM) levels (Fig. 3 A), with mean P_o ratio significantly less than unity (Fig. 3 B) because of both longer t_c and shorter t_o (Fig. 3, C and D). This is similar to the suppression of channel gating by [Ca²⁺]_ER jumping from 70 nM to 300 µM in subsaturating (3 µM) [InsP₃] (Fig. 2). Accordingly, if the inhibitory effect of high [Ca²⁺]_ER on channel P_o is mediated by some luminal Ca²⁺-binding site(s) on the channel, the site must be allosterically coupled to the InsP₃-binding sites on the cytoplasmic side of the channel so that channel activation by InsP₃ binding to its cytoplasmic site and channel inhibition by Ca²⁺ binding to the luminal site are mutually antagonistic. Alternatively, if the [Ca²⁺]_ER effect is mediated by the Ca²⁺ flux through the channel, the lack of effect on channel gating of [Ca²⁺]_ER jump from 70 nM to 300 µM in the presence of saturating [InsP₃] can be accounted for by the InsP₃-induced reduction in sensitivity of InsP₃R to [Ca²⁺]_i inhibition. In this scenario, a rise in local [Ca²⁺]_i caused by Ca²⁺ flux driven through the channel by [Ca²⁺]_ER of 300 µM that inhibits InsP₃R channel gating in subsaturating [InsP₃] is not sufficient to affect channel gating, as saturating [InsP₃] reduces the sensitivity of the channel to [Ca²⁺]_i inhibition. However, the higher rise in local [Ca²⁺]_i driven by a higher [Ca²⁺]_ER of 1.1 mM can still cause suppression of channel activity, even in the presence of saturating [InsP₃].

Figure 3.

Modulation of InsP₃R channel activity by various [Ca²⁺]_ER in saturating [InsP₃]. (A) A typical single-channel current trace recorded during a switch of [Ca²⁺]_ER from 70 nM to 1.1 mM. Note the substantially smaller channel current when [Ca²⁺]_ER = 1.1 mM as a result of the blocking of the channel by permeant Ca²⁺. A part of the current trace (indicated by an orange line) is shown with larger current and time scales in the inset to show the details of channel gating in [Ca²⁺]_ER = 1.1 mM. (B–D) Bar graphs of mean ratios of channel P_o, t_o, and t_c observed before and after [Ca²⁺]_ER switches between 70 nM and 300 µM or 1.1 mM, as indicated.

[Ca²⁺]_i dependence of the modulation of InsP₃R channel P_o by [Ca²⁺]_ER

To identify the mechanisms underlying the regulation of InsP₃R channel activity by [Ca²⁺]_ER, we investigated the [Ca²⁺]_i dependence of the effect. If the observed [Ca²⁺]_ER modulation of InsP₃R channel activity is mediated by cytoplasmic Ca²⁺-binding sites of the channel, [Ca²⁺]_ER modulation should be consistent with the biphasic [Ca²⁺]_i regulation of InsP₃R channel P_o (Fig. 1 C). Accordingly, when a pipette solution containing suboptimal low [Ca²⁺]_i is used instead of one with optimal [Ca²⁺]_i, the rise in local [Ca²⁺]_i caused by the Ca²⁺ flux driven through the open channel by high [Ca²⁺]_ER is expected to activate instead of inhibit channel gating. Indeed, when excised lum-out experiments were performed with saturating (10 µM) [InsP₃] and suboptimal low (55 nM) [Ca²⁺]_i in the pipette solution, the activity of the InsP₃R channel was significantly enhanced when [Ca²⁺]_ER was switched from 70 nM to 300 µM (Fig. 4 A), as reflected in a mean P_o ratio significantly larger than unity (Fig. 4 B) mainly caused by increase in t_o (Fig. 4 C) while t_c remained unaltered (Fig. 4 D). This observation is difficult to account for with the hypothesis of a luminal Ca²⁺-binding site on the InsP₃R channel regulating its activity.

Figure 4.

Effects of [Ca²⁺]_i on [Ca²⁺]_ER modulation of InsP₃R channel activity. (A) A typical single-channel lum-out patch-clamp current trace recorded during a switch of [Ca²⁺]_ER from 70 nM to 300 µM by perfusion solution exchange in suboptimal (55 nM) [Ca²⁺]_i. (B–D) Bar graphs of mean ratios of channel P_o, t_o, and t_c, respectively, observed before and after [Ca²⁺]_ER switches between 70 nM and 300 µM for different [Ca²⁺]_i. [Ca²⁺]_i was buffered to 2 µM by 0.5 mM diBrBAPTA and to 55 nM by 0.5 mM BAPTA.

[Ca²⁺]_ER modulation of InsP₃R channel P_o depends on the magnitude of the Ca²⁺ flux

To further confirm that rise in local [Ca²⁺]_i in the vicinity of the channel pore caused by feed-through Ca²⁺ flux driven by high [Ca²⁺]_ER modulates InsP₃R channel significantly by Ca²⁺ binding to cytoplasmic sites on the channel, we investigated whether channel activity would be affected if the magnitude of the Ca²⁺ flux through the InsP₃R channel was altered by changing V_app only, with [Ca²⁺]_ER kept the same. In all previous experiments, V_app = −30 mV. This V_app drove Ca²⁺ from the bath solution through the channel to the pipette solution, in the same direction as the [Ca²⁺]_f gradient when the perfusion solution contained [Ca²⁺]_ER of 300 µM. When the polarity of V_app is reversed (V_app = +30 mV), the applied V_app opposes the [Ca²⁺]_f gradient. Although this change in V_app polarity is insufficient to reverse the direction of the Ca²⁺ flux through the channel, it reduces the magnitude of the flux and therefore diminishes the rise in local [Ca²⁺]_i around the channel pore. In excised lum-out nuclear patch-clamp experiments with saturating [InsP₃] (10 µM) and suboptimal [Ca²⁺]_i (70 nM), jumps of [Ca²⁺]_ER from 70 nM to 300 µM with V_app = +30 mV still enhanced InsP₃R channel activity (Fig. 5 A) in a qualitatively similar way as in V_app = −30 mV, giving a mean P_o ratio significantly larger than unity (Fig. 5 B) solely by prolonging t_o (Fig. 5, C and D). However, the increases in channel P_o and t_o were substantially less in V_app = +30 mV than those observed with V_app = −30 mV (Fig. 5, B and C).

Figure 5.

[Ca²⁺]_ER modulation of InsP₃R-3 channel activity depends on magnitude of Ca²⁺ flux through the open-channel pore. (A) A typical single-channel lum-out nuclear patch-clamp current trace recorded during a switch of [Ca²⁺]_ER from 70 nM to 300 µM in V_app = +30 mV. Note that the change in channel current size as the result of the change in [Ca²⁺]_ER was smaller. This is because of the reduction in Ca²⁺ flux through the channel by the positive V_app. (B–D) Bar graphs of mean ratios of channel P_o, t_o, and t_c observed before and after [Ca²⁺]_ER switches between 70 nM and 300 µM in V_app = ±30 mV. ^ooo indicates statistically significant difference between the two ratios connected by the bracket (P < 0.005; unpaired t test).

[Ca²⁺]_ER modulation of InsP₃R channel P_o depends on cytoplasmic Ca²⁺-buffering conditions

Our experiments so far have demonstrated convincingly that most of the effects of [Ca²⁺]_ER on InsP₃R channel activity are mediated by the Ca²⁺ flux driven by [Ca²⁺]_ER to raise local [Ca²⁺]_i to modify InsP₃R channel activity through the cytoplasmic-activating and inhibitory Ca²⁺-binding sites on the channel. We next performed experiments to examine more closely the existence of an intrinsic functional regulatory Ca²⁺-binding site on the luminal side of the channel. We investigated the effects of cytoplasmic Ca²⁺-buffering conditions on [Ca²⁺]_ER modulation of channel activity in optimal [Ca²⁺]_i (2 µM) and subsaturating [InsP₃] (3 µM). In all previous experiments involving [Ca²⁺]_i = 2 µM, [Ca²⁺]_i in the pipette (cytoplasmic) solution was buffered by 0.5 mM of the fast Ca²⁺ chelator diBrBAPTA (Ca²⁺ binding rate k_on of ∼450 µM⁻¹s⁻¹; Naraghi, 1997). In weaker buffering conditions in which [Ca²⁺]_i was buffered to 2 µM by a low concentration (0.1 mM) of the slower Ca²⁺ chelator HEDTA (k_on of ∼4.5 µM⁻¹s⁻¹) (Naraghi, 1997), suppression of channel activity when [Ca²⁺]_ER was raised from 70 nM to 300 µM was more profound than that observed in 0.5 mM diBrBAPTA (Fig. 6 A). Accordingly, the mean P_o ratio was significantly lower under weak cytoplasmic Ca²⁺-buffering conditions than under the normally used buffering conditions, even though [Ca²⁺]_i was kept at 2 µM (Fig. 6 C). This is solely because of a significant increase in t_c (Fig. 6, D and E). In contrast, the inhibitory effects of the same [Ca²⁺]_ER jump were completely abolished when cytoplasmic Ca²⁺ was strongly buffered at 2 µM with a high concentration (5 mM) of diBrBAPTA (Fig. 6 B), giving mean P_o, t_o, and t_c ratios not significantly different from unity (Fig. 6, C–E). Control experiments demonstrated that different Ca²⁺-buffering conditions themselves had no significant effect on P_o in the absence of Ca²⁺ flux through the active channels ([Ca²⁺]_ER = 70 nM) in both on-nucleus and excised lum-out configurations (Fig. 7).

Figure 6.

Effects of cytoplasmic Ca²⁺-buffering conditions on [Ca²⁺]_ER modulation of InsP₃R channel activity. (A and B) Typical single-channel lum-out nuclear patch-clamp current traces recorded during switches of [Ca²⁺]_ER from 70 nM to 300 µM in different cytoplasmic Ca²⁺-buffering conditions. (C–E) Bar graphs of mean ratios of P_o, t_o, and t_c, respectively, observed before and after [Ca²⁺]_ER switches between 70 nM and 300 µM for different cytoplasmic Ca²⁺-buffering conditions. ^oo indicates statistically significant difference between the two ratios connected by the bracket (P < 0.01; unpaired t test). Note the logarithmic scale used for the t_c axis (in red) in E.

Figure 7.

InsP₃R channel P_o is independent of cytoplasmic Ca²⁺-buffering conditions in the absence of Ca²⁺ flux through the channel. (A–C) Typical single-channel on-nucleus patch-clamp current traces with [Ca²⁺]_i in pipette solutions buffered to 2 µM by 5 mM diBrBAPTA (A), 0.5 mM diBrBAPTA (B), or 0.1 mM HEDTA (C). [InsP₃] = 3 µM and V_app = −40 mV. Bath solution contained [Ca²⁺]_ER = 70 nM. (D) Mean InsP₃R channel P_o for [InsP₃] = 3 µM and [Ca²⁺]_i = 2 µM in various cytoplasmic Ca²⁺-buffering conditions for on-nucleus (closed bars) and excised lum-out (open bars) patch-clamp experiments. Excised lum-out patches were perfused with solution containing [Ca²⁺]_ER = 70 nM. H stands for HEDTA, and dB stands for diBrBAPTA. No statistically significant difference exists between P_o in any two of the three Ca²⁺-buffering conditions plotted for on-nucleus or lum-out experiments (ANOVA).

If [Ca²⁺]_ER modulation of channel P_o is mediated by luminal Ca²⁺-binding site(s), it is difficult to conceive how such a luminal site(s) could be sensitive to Ca²⁺ buffering on the cytoplasmic side by artificial chemical Ca²⁺ chelators that are not naturally found in vivo. Furthermore, the complete abolition of the effect of physiological [Ca²⁺]_ER (300 µM) on channel activity by sufficiently strong cytoplasmic Ca²⁺ buffering suggests that there is no luminal Ca²⁺-binding site intrinsic to the InsP₃R, activating or inhibitory, that is sensitive to 300 µM [Ca²⁺]_ER.

Complete abrogation of modulatory effects of [Ca²⁺]_ER in the physiological range on InsP₃R channel activity

To determine if [Ca²⁺]_ER in the upper limits of the physiological range could possibly inhibit InsP₃R channel activity through a luminal Ca²⁺-binding site on the channel, we looked for inhibitory effects on InsP₃R channel activity by [Ca²⁺]_ER jumps from 70 nM to 1.1 mM (higher than most observed [Ca²⁺]_ER; Button and Eidsath, 1996; Bygrave and Benedetti, 1996; Meldolesi and Pozzan, 1998; Yu and Hinkle, 2000) that were independent of the rise in [Ca²⁺]_i resulting from Ca²⁺ flux passing through the open InsP₃R channel. In excised lum-out nuclear patch-clamp experiments with V_app = +30 mV (to reduce the magnitude of the Ca²⁺ flux through the open channel) and pipette solution containing saturating (10 µM) [InsP₃] and optimal (2 µM) [Ca²⁺]_i buffered by 5 mM diBrBAPTA (to reduce the rise in [Ca²⁺]_i caused by the Ca²⁺ flux), [Ca²⁺]_ER jumps from 70 nM to 1.1 mM did not affect InsP₃R channel gating (Fig. 8 A), so that the mean P_o, t_o, and t_c ratios observed were not significantly different from unity (Fig. 8, B–D).

Figure 8.

Abrogation of inhibition of InsP₃R-3 channel activity by physiological levels of [Ca²⁺]_ER. (A) A typical single-channel current trace recorded during a switch of [Ca²⁺]_ER from 70 nM to 1.1 mM. (B–D) Bar graphs of mean ratios of channel P_o, t_o, and t_c observed before and after [Ca²⁺]_ER switches between 70 nM and 1.1 mM, showing the abrogation of inhibitory effects of 1.1 mM [Ca²⁺]_ER on InsP₃R channel activity in the presence of stronger cytoplasmic Ca²⁺ buffering (5 mM BAPTA) and smaller Ca²⁺ flux through the channel as a result of positive V_app. Labels on the x axes indicate [diBrBAPTA] in mM (top) and V_app in mV (bottom).

We also checked whether the activating effects of [Ca²⁺]_ER on InsP₃R channel activity observed in suboptimal (55 nM) [Ca²⁺]_i (Fig. 4) could be completely abrogated by a combination of positive V_app and strong cytoplasmic Ca²⁺ buffering. Unlike the inhibitory effects, [Ca²⁺]_ER jumps from 70 nM to 550 µM (maximal [Ca²⁺]_ER observed in many cell types; Button and Eidsath, 1996; Bygrave and Benedetti, 1996; Meldolesi and Pozzan, 1998; Yu and Hinkle, 2000) still enhanced InsP₃R channel activity in excised lum-out nuclear patch-clamp experiments with V_app = +30 mV and pipette solution containing saturating (10 µM) [InsP₃] and suboptimal (55 nM) [Ca²⁺]_i buffered by 10 mM BAPTA (Fig. 9 A), so that the mean P_o observed was significantly higher than unity (Fig. 9 D), mostly because of longer t_o (Fig. 9, E and F). However, increasing V_app further to +50 mV substantially diminished the activating effects of the [Ca²⁺]_ER jumps (Fig. 9 B), with a mean P_o ratio that is much lower although still significantly larger than unity (Fig. 9 D), which is caused by the t_o ratio that is significantly greater than unity (Fig. 9, E and F). At V_app = +70 mV, the activating effects of the [Ca²⁺]_ER jumps were completely abrogated so that the mean P_o, t_o, and t_c ratios were not different from unity (Fig. 9, D–F).

Figure 9.

Abrogation of activation of InsP₃R-3 channel activity by physiological levels of [Ca²⁺]_ER. (A–C) Typical single-current traces recorded during a switch of [Ca²⁺]_ER from 70 nM to 550 µM with different V_app used. (D–F) Bar graphs of mean ratios of channel P_o, t_o, and t_c observed before and after [Ca²⁺]_ER switches between 70 nM and 550 µM, showing the abrogation of activating effects of 550 µM [Ca²⁺]_ER on InsP₃R channel activity by increasing positive V_app and strong cytoplasmic Ca²⁺ buffering used. ^ooo and ^oo indicate statistically significant difference between the two ratios connected by the bracket (P < 0.005 and 0.01, respectively; unpaired t test). Note the logarithmic scale used for the t_o axis (in red) in E.

In the experiments investigating the abrogation of activating effects of [Ca²⁺]_ER changes, [Ca²⁺]_ER was increased to 550 µM rather than 1.1 mM because using high concentrations of Ca²⁺ chelator in the pipette solution did not reduce the activating effects of [Ca²⁺]_ER effectively. This is probably because the cytoplasmic activating Ca²⁺-binding site is located close to the channel pore (see Discussion). To abrogate the effects of [Ca²⁺]_ER jumps to 1.1 mM as in other experiments would require using V_app of >70 mV, which severely compromised the integrity of the gigaohm seal between the isolated nuclear membrane patch and the patch-clamp microelectrode.

Collectively, the total abrogation of both the inhibitory effects (in optimal [Ca²⁺]_i = 2 µM) and the activating effects (in suboptimal [Ca²⁺]_i = 55 nM) of physiological levels of [Ca²⁺]_ER in conditions that only affected the magnitude of the Ca²⁺ flux through the channel and the changes in [Ca²⁺]_i demonstrates that there is no regulatory Ca²⁺-binding site sensitive to physiological [Ca²⁺]_ER on the luminal side of the InsP₃R channel.

DISCUSSION

This study is the first investigation of the effects of [Ca²⁺]_ER on single-channel activity of InsP₃R using the nuclear patch-clamp approach. Previously, two electrophysiological studies explored the effects of [Ca²⁺]_ER using reconstituted cerebellar type 1 InsP₃R channels in planar lipid bilayers. In the first study (Bezprozvanny and Ehrlich, 1994), the cytoplasmic solution contained no permeant ion, with 55 mM Ba²⁺, Mg²⁺, Sr²⁺, or a combination of Ca²⁺ and Sr²⁺ in the luminal solutions used as the main charge carrier. Channel P_o was significantly reduced as [Ca²⁺]_ER was increased from 300 µM to 44 mM. It was suggested that feed-through effects of [Ca²⁺]_ER-driven Ca²⁺ flux through the channel contributed to the inhibition of InsP₃R channel activity, but the existence of a luminal inhibitory Ca²⁺ site could not be conclusively ruled out because of insufficient [Ca²⁺]_i buffering by only 1 mM EGTA. The relevance of this study was further diminished by the nonphysiological ionic compositions used, which rendered it questionable whether the magnitude of Ca²⁺ flux and therefore the resulting changes in [Ca²⁺]_i resembled those in physiological ionic conditions. In the second study (Thrower et al., 2000), 500 mM K⁺ was present in all solutions with [Ca²⁺]_i buffered to 0.2–0.3 µM by either 10 mM HEDTA or 1.7 mM BAPTA. The conclusion of the study that Ca²⁺_ER affects InsP₃R channel gating through direct interaction with the luminal face of the channel was critically undermined by technical difficulties (brief and inconsistent channel activities, multiple conductance substates), a mostly qualitative description of channel activity (no quantitative P_o or t_o-t_c analysis), and the inappropriate use of Ca²⁺ buffers (HEDTA cannot effectively buffer [Ca²⁺]_i at 0.2–0.3 µM despite the high concentrations used because of its low Ca²⁺ affinity; BAPTA has the right Ca²⁺ affinity, but only a low concentration was used).

In this study, modulation of InsP₃R channel activity by [Ca²⁺]_ER was examined under rigorously controlled [Ca²⁺]_i, [Ca²⁺]_ER, [InsP₃], and cytoplasmic Ca²⁺-buffering conditions in excised lum-out nuclear patch-clamp experiments with perfusion solution exchange. The observed dependencies of [Ca²⁺]_ER modulation of channel activity on [InsP₃] (Fig. 2), [Ca²⁺]_ER (Fig. 3), [Ca²⁺]_i (Fig. 4), V_app (Fig. 5), and cytoplasmic Ca²⁺-buffering conditions (Fig. 6), and the total abrogation of the effects of [Ca²⁺]_ER on channel activity by conditions that affect the rise in local [Ca²⁺]_i caused by the Ca²⁺ flux but not [Ca²⁺]_ER itself (Figs. 6, 8, and 9), together demonstrate that InsP₃R channel activity is regulated solely by the feed-through effects of the [Ca²⁺]_ER-driven Ca²⁺ flux through the open channel, raising [Ca²⁺]_i in the microdomain around the channel to alter P_o of the channel through its cytoplasmic activating and inhibitory Ca²⁺-binding sites, and not by direct binding of Ca²⁺ to the luminal side of the InsP₃R channel.

The observed modulation of channel activity by [Ca²⁺]_ER-driven Ca²⁺ flux through the channel itself provides insights regarding the kinetics of [Ca²⁺]_i regulation of InsP₃R channel gating. From these insights and others from previous studies of ligand regulation of InsP₃R channel gating, we develop below the concept of the effective time-averaged [Ca²⁺]_i profile around the channel caused by Ca²⁺ flux through the channel itself when it is gating. With that concept, and using channel P_o observed in the presence of various [Ca²⁺]_ER-driven Ca²⁺ fluxes under various [Ca²⁺]_i, [Ca²⁺]_ER, and [InsP₃], we then estimate the distances between the channel pore and the cytoplasmic regulatory Ca²⁺ sites.

Kinetics of fluctuations of local [Ca²⁺]_i in the vicinity of the channel pore caused by Ca²⁺ flux through the pore

In a previous study (Vais et al., 2010a), we determined that physiological levels of [Ca²⁺]_ER can drive substantial Ca²⁺ fluxes through an open InsP₃R-3 channel. Numeric simulations allow us to follow the changes in the [Ca²⁺]_i profile around the Ca²⁺-permeable InsP₃R channel during its gating. The simulations indicate that under our experimental Ca²⁺-buffering conditions, the [Ca²⁺]_i profile around an open InsP₃R channel achieves steady-state levels within 100 µs after the channel opens (Fig. 10 A). Furthermore, after a channel closes, the elevated [Ca²⁺]_i around the channel collapses rapidly and returns to the basal level within 1 ms (Fig. 10 B). Thus, the [Ca²⁺]_i profile around an InsP₃R channel fluctuates abruptly between the steady-state open- and closed-channel levels in a quasi-binary manner, with kinetics rigidly dictated by the opening and closing of the channel. This has significant implications for the kinetic properties of [Ca²⁺]_i regulation of the channel and how [Ca²⁺]_ER-driven Ca²⁺ flux through the InsP₃R channel pore can modulate the activity of the channel itself.

Figure 10.

Simulated [Ca²⁺]_i profiles around an InsP₃R channel (cytoplasmic free [Ca²⁺] at various distances from the channel pore axis) under various Ca²⁺-buffering conditions. Buffering conditions are indicated for the profiles shown. [Ca²⁺]_ER = 300 µM. Bulk [Ca²⁺]_i (far from the channel) = 2 µM. (A) [Ca²⁺]_i profiles at different t after the channel opens, plotted in different colors as indicated. (B) The channel opened continuously for 50 ms before it closed. Profiles at different t after the channel closed are plotted in different colors as indicated. Even with the weakest buffering (0.1 mM HEDTA), [Ca²⁺]_i profiles reach steady-state level within 1 ms of the channel opening, and [Ca²⁺]_i around the channel returns within 100 µs to essentially the level that existed before the channel opened.

Kinetics of cytoplasmic Ca²⁺ activation of InsP₃R channel deduced from the observed enhancement of channel activity by [Ca²⁺]_ER-driven Ca²⁺ flux

To derive insights into the kinetics of cytoplasmic Ca²⁺ activation of InsP₃R channel from the observed channel response to [Ca²⁺]_ER-driven Ca²⁺ flux through the channel itself, we first consider a hypothetical, extreme kind of response of a Ca²⁺_i-activated channel to the increase in local [Ca²⁺]_i at the single activating cytoplasmic Ca²⁺-binding site of the channel. In this extreme case, the kinetics of channel gating are rigidly dictated by the local [Ca²⁺]_i at its Ca²⁺ site. To generate this kind of response, the activating latency (interval between local [Ca²⁺]_i increasing beyond K_act and the first resulting opening of the channel, τ_act) and the deactivating latency (interval between local [Ca²⁺]_i decreasing below K_act and the last closing of the channel, τ_deact) must both be much less than the time scale of channel gating (t_o and t_c), so that the channel opens/closes practically instantaneously after the rise/drop in local [Ca²⁺]_i at its activating site. Moreover, the channel must remain open as long as its single activating Ca²⁺ site is occupied, and it must remain closed whenever the activating site is vacant, i.e., the gating status (open or closed) of the channel is deterministically dependent on the occupancy of the activating Ca²⁺ site. Given that the [Ca²⁺]_i profile around a Ca²⁺-permeable channel is rigidly dictated by the gating of the channel under our experimental conditions (as discussed in the previous section), a lone Ca²⁺_i-activated, Ca²⁺-permeable channel in the ER with its gating rigidly dictated by [Ca²⁺]_i at its activating Ca²⁺ site as described above would not be expected to be activated by Ca²⁺ flux through the channel itself. This is simply because the activating site on the channel must be already occupied when the channel is open and therefore cannot further bind Ca²⁺. Consequently, the channel cannot be affected by the local [Ca²⁺]_i elevated by the [Ca²⁺]_ER-driven Ca²⁺ flux. Conversely, when its Ca²⁺-activating site is vacant and available to bind Ca²⁺, the channel must be closed so that there is no [Ca²⁺]_ER-driven Ca²⁺ flux through that channel to activate the channel. Thus, the channel would behave as if there is no Ca²⁺ flux through the open channel.

However, in lum-out nuclear patch-clamp experiments with the pipette solution containing [Ca²⁺]_f = 55 nM and [InsP₃] = 10 µM, we observed significant sustained activation of the channel in the presence of high [Ca²⁺]_ER (Figs. 4, A and B, and 5, A and B) that can be completely accounted for by Ca²⁺ flux through the channel raising local [Ca²⁺]_i at its cytoplasmic activating Ca²⁺ site to enhance its P_o (Fig. 5 A). This observation therefore has nontrivial implications for the kinetics of cytoplasmic Ca²⁺ activation of InsP₃R channel. Most obviously, this indicates that local [Ca²⁺]_i at the cytoplasmic site does not rigidly dictate the gating of the channel. Different mechanisms can contribute to uncoupling of the gating kinetics of the channel from changes in [Ca²⁺]_i at the activating Ca²⁺ site. Most importantly, Ca²⁺ regulates InsP₃R channel activity stochastically, so Ca²⁺ binding to the activating site only induces the channel to adopt a more active kinetic conformation with a higher P_o, but does not always result in channel opening. The channel can open and close with no change in its Ca²⁺-binding status (Mak et al., 2003). Furthermore, the tetrameric structure of the channel (Foskett et al., 2007) indicates that it has multiple activating Ca²⁺ sites. Thus, the channel can open even when there are vacant activating Ca²⁺ sites on the channel available to detect the elevated local [Ca²⁺]_i and further enhance channel activity. Another significant factor determining the degree of coupling between local [Ca²⁺]_i at the activating Ca²⁺ site and the gating of the channel is the kinetics of Ca²⁺_i activation of the channel, i.e., τ_act and τ_deact. Previous cytoplasmic-side-out nuclear patch-clamp experiments with rapid perfusion exchange studying endogenous InsP₃R channels from insect Sf9 cells (Mak et al., 2007) revealed that in saturating (10 µM) [InsP₃], InsP₃R channels can respond relatively slowly to abrupt changes in [Ca²⁺]_i, with long latencies (τ_act of approximately tens of milliseconds and τ_deact of approximately a few hundreds of milliseconds). With long response latencies (relative to channel gating t_o and t_c), the channel can open and close, and local [Ca²⁺]_i around the channel jumps up and down, multiple times during the time the channel takes to respond to one change in [Ca²⁺]_i, thereby effectively uncoupling the [Ca²⁺]_i at the activating sites from the gating of the channel.

In the other extreme case, in which the gating of the channel is completely uncoupled from the local [Ca²⁺]_i at the activation sites, instead of responding to the instantaneous [Ca²⁺]_i resulting from individual channel opening and closing events, the gating of the channel depends only on the time-averaged [Ca²⁺]_i at the sites. Over a long period T (>>τ_act, τ_deact, t_o, t_c), a vacant activating Ca²⁺ site on the channel will on average be exposed to the steady-state open-channel [Ca²⁺]_i (because of the open-channel Ca²⁺ current, i_Ca, driven by the electrochemical gradient across the channel) for a period of P_oT, and to the steady-state closed-channel [Ca²⁺]_i for a period of (1–P_o)T. Thus, assuming first-order Ca²⁺ binding to the activating sites, the channel will exhibit steady-state gating kinetics similar to that of a channel with activating sites constantly exposed to a local [Ca²⁺]_i equivalent to that generated by a Ca²⁺ current of magnitude P_o i_Ca passing through the pore.

In reality, the coupling between InsP₃R channel gating and changes of local [Ca²⁺]_i is partial, lying between the two extremes of total rigid dictation of gating by [Ca²⁺]_i at the activating sites and complete decoupling with gating unrelated to instantaneous [Ca²⁺]_i at the sites. Therefore, the activating sites are effectively exposed to a [Ca²⁺]_i that is equivalent to a time-averaged Ca²⁺ current of magnitude between 0 and P_o i_Ca.

Besides deducing that InsP₃R channel gating is not rigidly dictated by [Ca²⁺]_i at the activating sites, other insights about the kinetics of Ca²⁺_i activation of InsP₃R channel can be derived from the observed activating effects of Ca²⁺-driven Ca²⁺ flux on channel gating. In the present study, in the absence of Ca²⁺ flux with low [Ca²⁺]_ER (70 nM), the InsP₃R-3 channel was observed to exhibit low P_o (∼0.02–0.05) with short t_o (∼2 ms) and long t_c (∼50 ms) in constant suboptimal [Ca²⁺]_i (55 nM), even in saturating [InsP₃] (10 µM) (Figs. 1 C, 4 A, and 9 A). Nevertheless, abrupt and sustained increases in channel P_o and t_o were observed (Figs. 4 A and 9 A) in response to the onset of Ca²⁺ flux through the channel to the cytoplasmic side resulting from [Ca²⁺]_ER being raised to physiological levels. This rapid response indicates that vacant cytoplasmic activating Ca²⁺ sites of the channel were able to capture Ca²⁺ during the first couple of brief channel-opening events after the [Ca²⁺]_ER jump, when the local [Ca²⁺]_i at the sites was raised by the Ca²⁺ flux and before the channel closed and terminated the Ca²⁺ flux. Thus, the rate of Ca²⁺ binding to the activating sites must be high, suggesting that the Ca²⁺ flux can raise the local [Ca²⁺]_i at the activating sites to a high level, and therefore the sites are probably located close to the channel pore (see further discussion below).

Another feature of the activating effects of [Ca²⁺]_ER on InsP₃R-3 channel gating is that the increase in channel P_o as a result of [Ca²⁺]_ER jumps from 70 nM to physiological levels was mostly achieved by prolonging t_o, with no significant change in t_c (Figs. 4 D, 5 E, and 9 F). It has been suggested that because cytoplasmic regulatory Ca²⁺ sites are inaccessible to [Ca²⁺]_ER when a Ca²⁺-permeable, Ca²⁺_i-regulated channel is closed, the absence of luminal Ca²⁺ site on the channel means that t_c of the channel should not depend on [Ca²⁺]_ER (Laver, 2007a,b). According to this statement, our observation that t_c was not significantly affected by [Ca²⁺]_ER is consistent with a conclusion that the InsP₃R channel has no luminal regulatory Ca²⁺ site. However, the statement is only true for a Ca²⁺_i-regulated, Ca²⁺-permeable channel whose gating is strongly dictated by [Ca²⁺]_i at its cytoplasmic regulatory Ca²⁺ sites. Because gating of the InsP₃R channel is not rigidly dictated by [Ca²⁺]_i at its cytoplasmic Ca²⁺-activating sites, the observation is better interpreted as an indication that under the experimental conditions used in Figs. 4 A, 5 A, and 9 (A and B), the conformations assumed by the channel in the presence and absence of Ca²⁺ flux have similar t_c.

Cytoplasmic inhibitory Ca²⁺ sites also experience an effective time-averaged local [Ca²⁺]_i due to [Ca²⁺]_ER-driven Ca²⁺ flux through the channel

In the extreme case where gating is rigidly dictated by [Ca²⁺]_i at the regulatory sites, the situation for Ca²⁺ inhibition is different from that for Ca²⁺ activation for a Ca²⁺_i-regulated, Ca²⁺-permeable channel. Whereas activating sites can never experience the flux through the channel (discussed above), the inhibitory sites will be vacant when the channel opens and occupied when the channel is closed. Thus, effectively, the vacant sites will always be exposed to the elevated open-channel local [Ca²⁺]_i caused by the Ca²⁺ current of magnitude i_Ca through the pore.

However, this cannot be true for the InsP₃R-3 channel. Even though our experimental observations clearly demonstrated that the channel has no luminal regulatory Ca²⁺-binding site (Fig. 8), and the rapid collapse of the [Ca²⁺]_i profile around the channel pore after the channel closes (Fig. 10) means that the cytoplasmic inhibitory Ca²⁺ sites are effectively inaccessible to [Ca²⁺]_ER when the channel is closed, channel t_c still exhibited clear dependence on [Ca²⁺]_ER when the channel was inhibited by rise in [Ca²⁺]_i caused by Ca²⁺ flux through the open channel (Figs. 2 F, 3 D, and 6 E). Therefore, the gating of the channel cannot be rigidly dictated by [Ca²⁺]_i at the inhibitory sites. Rather, the observations suggest the presence of uncoupling mechanisms: the tetrameric InsP₃R channel having multiple inhibitory sites that regulate channel P_o stochastically, with Ca²⁺ inhibition latency and the latency of channel recovery from Ca²⁺ inhibition both significantly longer than the time scale of channel gating (t_o and t_c). Such mechanisms can allow the modulation of InsP₃R channel gating by the high local [Ca²⁺]_i at the inhibitory sites established during a channel opening to extend beyond the termination of that opening and the subsequent rapid collapse of the local [Ca²⁺]_i rise.

At the other extreme, with the gating of the channel completely decoupled from the fluctuations of local [Ca²⁺]_i caused by the openings and closings of the channel, the channel should gate with kinetics similar to those associated with one of the inhibitory sites exposed to a steady local [Ca²⁺]_i equivalent to that generated by a Ca²⁺ current of magnitude P_o i_Ca passing through the pore, the same as the situation for the activating site. For realistic partial coupling between the two extremes, the inhibitory sites are effectively exposed to a time-averaged local [Ca²⁺]_i caused by a Ca²⁺ current of magnitude between P_o i_Ca and i_Ca.

In summary, the observed sustained activation and inhibition of gating by Ca²⁺ flux through an InsP₃R channel indicate that channel gating is not deterministically regulated by [Ca²⁺]_i, and that a channel can respond to the Ca²⁺ flux through itself because its activation and inhibition kinetics enable it to sense an effective steady-state local flux-driven [Ca²⁺]_i.

Estimates of the locations of functional cytoplasmic Ca²⁺-binding sites from Ca²⁺ flux–mediated modulation of InsP₃R channel activity

Our experiments demonstrate that [Ca²⁺]_ER modulates the activity of r-InsP₃R-3 channels in DT40-KO-r-InsP₃R-3 cells solely via the Ca²⁺ flux it drives through the channel that raises the local [Ca²⁺]_i at the cytoplasmic regulatory Ca²⁺-binding sites. Using the steady-state [Ca²⁺]_i dependence of the channel P_o (Fig. 1), the effects of [Ca²⁺]_ER on channel activity can provide estimates of the effective time-averaged local [Ca²⁺]_i at the cytoplasmic activating or inhibitory Ca²⁺-binding sites of the channel (Table 1). [Ca²⁺]_i profiles ([Ca²⁺]_i at various distances from the channel pore) were numerically generated (Materials and methods) for the different Ca²⁺-buffering conditions, Ca²⁺ electrochemical gradients, and cytoplasmic ligand concentrations used in our experiments (Table 1). Checking the estimates of the effective time-averaged local [Ca²⁺]_i at the regulatory sites against the appropriate simulated [Ca²⁺]_i profiles, estimates can be made of the locations of the regulatory Ca²⁺-binding sites relative to the channel pore, which is situated at the center of the channel based on the structural symmetry of the tetrameric InsP₃R channel (Foskett et al., 2007).

Table 1.

Estimations of the distances between cytoplasmic regulatory Ca²⁺-binding sites and the channel pore axis

	Experimental conditions
Representative current trace	[InsP3]	[Ca2+]ERa	Bulk [Ca2+]ib	Vapp	Cytoplasmic Ca2+ buffering	Mean Po	iCa	iCa Po	[Ca2+]i profiles	Regulatory Ca2+ site	[Ca2+]i at site	Range of distance from pore
	µM			mV			fA	fA			µM	nm
Fig. 2 (A and B)	3	300 µM	2 µM	−30	0.5 mM diBrBAPTA	0.22	230	51	Fig. 11 A	Inhibitory	5.7 µM	12 < x < 39
Fig. 2 C	10	300 µM	2 µM	−30	0.5 mM diBrBAPTA	0.7	230	161	Fig. 11 B	Inhibitory	7.2 µM	23 < x < 31
Fig. 3 A	10	1.1 mM	2 µM	−30	0.5 mM diBrBAPTA	0.46	830	380	Fig. 11 C	Inhibitory	14.3 µM	25 < x < 42
Fig. 6 A	3	300 µM	2 µM	−30	0.1 mM HEDTA	0.09	230	21	Fig. 11 D	Inhibitory	12 µM	2 < x < 21
Fig. 6 B	3	300 µM	2 µM	−30	5 mM diBrBAPTA	0.4	230	92	Fig. 11 E	Inhibitory	3.1 µM	29 < x < 44
Fig. 8 A	10	1.1 mM	2 µM	+30	5 mM diBrBAPTA	0.7	75	53	Fig. 11 F	Inhibitory	3.1 µM	22 < x < 27
Fig. 4 A	10	300 µM	55 nM	−30	0.5 mM BAPTA	0.29	230	64	Fig. 11 G	Activating	500 nM	x < 62
Fig. 5 A	10	300 µM	55 nM	+30	0.5 mM BAPTA	0.12	20	2.5	Fig. 11 H	Activating	150 nM	x < 22
Fig. 9 A	10	550 µM	55 nM	+30	10 mM BAPTA	0.25	38	10	Fig. 11 I	Activating	410 nM	x < 13
Fig. 9 B	10	550 µM	55 nM	+50	10 mM BAPTA	0.08	12	1	Fig. 11 J	Activating	100 nM	x < 14
Fig. 9 C	10	550 µM	55 nM	+70	10 mM BAPTA	0.05	3	0.15	Fig. 11 K	Activating	80 nM	x < 9

^a[Ca²⁺]_ER = free [Ca²⁺] in perfusion solution.

^bBulk [Ca²⁺]_i = [Ca²⁺]_i at large distance from the channel pore = [Ca²⁺]_{r →∞} = free [Ca²⁺] in pipette solution.

For the inhibitory Ca²⁺ sites, depending on the degree of coupling between the gating of the channel and the fluctuations in local [Ca²⁺]_i associated with each opening and closing event, the [Ca²⁺]_i profile suitable for estimating the location of the sites lies between the [Ca²⁺]_i profile generated for Ca²⁺ current = i_Ca derived from the Goldman–Hodgkin–Katz current equation (Eq. 1) (for deterministic coupling between channel gating and the local [Ca²⁺]_i at the inhibitory Ca²⁺ site), and the profile generated for Ca²⁺ current = P_o i_Ca (for completely uncoupled channel gating and local [Ca²⁺]_i at the inhibitory site). Because the exact degree of coupling between channel gating and local [Ca²⁺]_i fluctuations for the experimental conditions used are not known, we use the two [Ca²⁺]_i profiles for currents = i_Ca and P_o i_Ca to derive upper and lower limits for the distance of the inhibitory Ca²⁺ sites from the channel pore.

In this study, we made six independent measurements of the inhibitory effects on channel gating of raising local [Ca²⁺]_i around the r-InsP₃R-3 channel beyond 2 µM (optimal [Ca²⁺]_i) (shown in Figs. 2, A and C, 3 A, 6, A and B, and 8 A), each of which provides an independent estimate of the range for the distance of the inhibitory Ca²⁺-binding site from the channel pore (Fig. 11, A–F, and Table 1). The estimated upper limits of this distance range between 21 and 44 nm (Table 1), with an average of 34 ± 4 nm. The estimated lower limits range between 2 and 29 nm (Table 1), with an average of 19 ± 4 nm. These suggest that the distance from the channel pore to the inhibitory site is ∼20–30 nm.

Figure 11.

Estimated effective time-averaged [Ca²⁺]_i determining InsP₃R channel gating activity, as sensed by cytoplasmic regulatory Ca²⁺-binding sites at various distances from the channel pore in various lum-out experiments. Pipette solution [Ca²⁺]_f ([Ca²⁺]_r→∞), perfusion solution [Ca²⁺]_f ([Ca²⁺]_ER), V_app, [InsP₃], and cytoplasmic Ca²⁺-buffering conditions used in each set of experiment are tabulated in each corresponding graph. A–F are related to experiments investigating the effect of Ca²⁺ flux mediated by the cytoplasmic inhibitory Ca²⁺-binding site(s), whereas G–K are related to experiments investigating the effect mediated by the cytoplasmic activating site(s). The effective [Ca²⁺]_i that produced the observed channel P_o are marked by dotted lines and tabulated (in red for inhibitory Ca²⁺ site and in blue for activating Ca²⁺ site). Black curves are effective [Ca²⁺]_i profiles derived from Ca²⁺ flux of magnitude = P_o i_Ca. The limits for the distances between the regulatory Ca²⁺-binding site and the channel pore derived from these [Ca²⁺]_i profiles are marked by black dotted lines and tabulated in black. Green curves in A–F are effective [Ca²⁺]_i profiles derived from Ca²⁺ flux of magnitude = i_Ca. The upper limits for the distances between the inhibitory Ca²⁺ site to the channel pore derived from these [Ca²⁺]_i profiles are marked by green dotted lines and tabulated in green.

Using image reconstructions based on electron cryomicroscopy or electron microscopy with negative staining, 3-D structures of single tetrameric InsP₃R channel have been determined (Jiang et al., 2002; da Fonseca et al., 2003; Hamada et al., 2003; Serysheva et al., 2003; Sato et al., 2004; Wolfram et al., 2010; Ludtke et al., 2011). Although the details differ significantly, they generally show a large structure on the cytoplasmic side with maximum radius (r) from the channel pore axis between 10.5 and 14.2 nm, and height above the ER membrane (h) between 13.5 and 18.3 nm. Simple geometric consideration for a 3-D structure suggests that the maximum distance between the channel pore and a Ca²⁺-binding site on the channel is ∼24–32.5 nm (r + h). Thus, our estimate of the inhibitory site being 20–30 nm from the pore of the channel is not in conflict with the 3-D structures reported and suggests that the inhibitory site may be located in a part of the channel furthest from the pore.

For the activating Ca²⁺ sites, an upper limit for the distance to the channel pore can be derived from the [Ca²⁺]_i profile for current with magnitude = P_o i_Ca, which corresponds to the extreme case when the channel gating is completely decoupled from local [Ca²⁺]_i at the activating sites during channel openings and closings. However, no lower limit can be deduced for the pore-to-activating-site distance from the extreme case with channel gating rigidly dictated by local [Ca²⁺]_i at the activating sites. This is because in this case, the channel cannot be activated by [Ca²⁺]_ER-driven Ca²⁺ flux through the pore, no matter where the activating Ca²⁺ sites are.

We made five independent measurements of the activating effects on channel gating of raising local [Ca²⁺]_i around the channel above 70 nM (resting [Ca²⁺]_i) (Figs. 4 A, 5 A, and 9, A–C). These provided five independent estimates of the upper limits for the activating site to channel pore distance (Fig. 11, G–K, and Table 1), suggesting that the activating site is less than ∼9–62 nm from the channel pore, which is also consistent with the 3-D structures of the channel. Moreover, 5 mM diBrBAPTA can effectively buffer the local [Ca²⁺]_i at the inhibitory Ca²⁺-binding sites to abolish the inhibiting effect of Ca²⁺ flux driven through the channel by 1.1 mM [Ca²⁺]_ER at V_app = +30 mV (Fig. 8 A), whereas even 10 mM BAPTA cannot sufficiently buffer the local [Ca²⁺]_i at the activating Ca²⁺-binding sites to abolish the activating effect of Ca²⁺ flux driven by 0.55 mM [Ca²⁺]_ER at the same V_app (Fig. 9 A). These observations strongly suggest that the activating Ca²⁺ sites are closer to the channel pore than the inhibitory sites.

It should be pointed out that these estimates of the locations of the cytoplasmic regulatory Ca²⁺ sites relative to the channel pore are very rough because the [Ca²⁺]_i profiles around the channel were simulated without taking into consideration factors that can affect the distribution of Ca²⁺ around the channel but are not known in any detail, like the 3-D surface topology and charge distribution of the channel.

A very similar approach to that used here was applied to estimate the locations of activating and inhibitory Ca²⁺-binding sites in the RyR intracellular Ca²⁺ release channel that also exhibits [Ca²⁺]_ER-driven Ca²⁺ flux modulation of channel activity (Liu et al., 2010). However, in that study, the couplings between channel gating and the local [Ca²⁺]_i fluctuations at the regulatory sites during channel openings and closings were not taken into consideration. The distances derived from the effective time-averaged [Ca²⁺]_i profile generated from Ca²⁺ current of magnitude = P_o i_Ca was assumed to be estimates of the actual distances between the regulatory sites and the channel pore, instead of limiting values of those distances. This led to their conclusion that the inhibitory Ca²⁺ site was 1.2 ± 0.16 nm from the channel pore. This is probably an underestimation because that was actually the lower limit of the distance. The activating Ca²⁺ site to channel pore distance was calculated to be 1.7 µm, which led to a conclusion that the activating site on the open channel was shielded from the channel’s own Ca²⁺ flux. However, this value should be the upper limit of the activating Ca²⁺ site to pore distance. Accordingly, their derivation does not provide strong support for the notion that the activating site is shielded from feed-through effects of the channel’s own Ca²⁺ flux.

Limitation of the excised lum-out nuclear patch-clamp experiments

Using excised-patch lum-out nuclear patch clamping, we have demonstrated that all modulation by [Ca²⁺]_ER of the gating activity of the r-InsP₃R-3 channel can be attributed to feed-through effects causing a rise in local [Ca²⁺]_i at cytoplasmic regulatory Ca²⁺-binding sites of the channel. We found no modulatory effects on InsP₃R channel gating involving luminal Ca²⁺-binding site(s) on the channel. However, it must be noted that the luminal side of the excised nuclear membrane patches was perfused with various solutions to change [Ca²⁺]_ER in our experiments. It is possible that a luminal Ca²⁺-binding factor(s) that could mediate effects of [Ca²⁺]_ER on InsP₃R channel activity was washed off by the perfusion. The possible existence of [Ca²⁺]_ER regulation of InsP₃R channel activity mediated by factor(s) in the ER lumen loosely associated with the channel should be investigated in the future using nuclear patch-clamp experiments in which the ER luminal milieu is preserved.