Inositol 1,4,5-trisphosphate Receptors in the Endoplasmic Reticulum: a Single-channel Point of View

Abstract

As an intracellular Ca²⁺ release channel at the endoplasmic reticulum membrane, the ubiquitous inositol 1,4,5-trisphosphate (InsP₃) receptor (InsP₃R) plays a crucial role in the generation, propagation and regulation of intracellular Ca²⁺ signals that regulate numerous physiological and pathophysiological processes. This review provides a concise account of the fundamental single-channel properties of the InsP₃R channel: its conductance properties and its regulation by InsP₃ and Ca²⁺, its physiological ligands, studied using nuclear patch clamp electrophysiology.

1. Introduction

The ubiquitously expressed inositol 1,4,5-trisphosphate (InsP₃) receptor (InsP₃R) is mostly localized to the endoplasmic reticulum (ER) membrane [1], where it functions as an ion channel to release Ca²⁺ stored in the ER lumen (Ca²⁺_ER) into the cytoplasm to raise the cytoplasmic free Ca²⁺ concentration ([Ca²⁺]_i) when it is activated by its physiological ligand, InsP₃, generated in the cytoplasm as part of a signal cascade resulting from activation of specific plasma membrane receptors by various extracellular stimuli [2]. InsP₃R thus plays a crucial role in the generation, propagation and regulation of cytoplasmic Ca²⁺ (Ca²⁺_i) signals that regulate numerous physiological and pathophysiological processes. It is therefore not surprising that it has been the subject of intense study ever since its identification [3]. Many aspects of the InsP₃R channel have been extensively reviewed recently, including its molecular structure [4–7] and its regulation by phosphorylation [8], redox reagents [9, 10] and ATP [8]. Thus, this short review will focus on the more fundamental properties of the single InsP₃R channel: its ion conductance properties and its regulation by its physiological ligands—InsP₃ and Ca²⁺, especially those reported since our previous reviews of single InsP₃R channel properties [11, 12].

Studies of individual InsP₃R channels began with the reconstitution of isolated and purified functional InsP₃R channels in artificial planar lipid bilayers [13], which allows the reconstituted channel(s) to be studied with known and rigorously controlled ionic and ligand conditions on both its cytoplasmic and luminal sides [14]. Because of the intracellular localization of the InsP₃R, application of patch clamp electrophysiology to investigate behaviors of the InsP₃R channel in a more native membrane environment was not feasible until isolated nuclei were used as surrogate for the ER [15, 16] because of the continuity of the outer nuclear membrane with the ER [17]. Since isolated nuclei with intact outer nuclear membranes can be obtained with high success rates [18], nuclear patch clamping in the “on-nucleus” configuration (Fig. 1A) preserves the protein environment on the luminal side of the recorded InsP₃R channel(s) while maintaining rigorous control of ligand and ionic conditions on both sides of the channel(s) [19]. Combining rapid perfusion techniques with nuclear patch clamping in luminal-side-out (lum-out) (Fig. 1B) or cytoplasmic-side-out (cyto-out) (Fig. 1C) configurations allows rapid (~ ms), repeated (tens of times in one experiment) and reversible exchange of the bath solution to study dynamic responses of InsP₃R channel activity to abrupt changes in InsP₃ and Ca²⁺ concentrations on either side of the channel, as well as to compare the gating and conductance properties of the same channels in the same isolated membrane patches under different ionic environments [19]. Performing nuclear patch clamping on an isolated nucleus with its outer nuclear membrane stripped by chemical treatment [20, 21] (Fig. 1D) can achieve the nucleoplasmic-side-out (nucleo-out) configuration (Fig. 1E) to study InsP₃R localized to the inner nuclear membrane [20, 22].

Fig. 1. Different configurations for nuclear patch clamping experiments.

Schematic diagrams illustrating the orientation of InsP₃R channels in nuclear membrane patches in various configurations of nuclear patch-clamp experiments. (A) On-nucleus configuration with outer nuclear membrane intact, (B) excised luminal-side-out configuration, (C) excised cytoplasmic-side-out configuration, (D) on-nucleus configuration with outer nuclear membrane removed, (E) excised nucleoplasmic-side-out configuration. (F) Diagram showing how the two aspects of the InsP₃R channel are represented in this figure. Figure modified from [19].

Single-channel properties of the InsP₃R channel have also been studied by applying whole-cell patch clamp techniques to chicken lymphocyte DT40 cells [23–29], in which InsP₃R channels are localized to the plasma membrane at very low density (< 5 channels/cell) [24]. However, the lipid and protein environments around these InsP₃R channels in the plasma membrane are different from those around the ER channels, and conductance and ligand regulation of InsP₃R channels in the two locations were significantly different [19]. Strictly speaking, InsP₃R channels localized to the plasma membrane are acting as plasma membrane Ca²⁺ entry channels rather than intracellular Ca²⁺ release channels [23].

Ca²⁺ signals generated by InsP₃R channels located in the ER near the plasma membrane in intact mammalian cells can also be studied using total internal reflection fluorescence (TIRF) microscopy [30]. By using a fast [Ca²⁺] indicator dye and loading Ca²⁺ buffer (EGTA) into cells to rapidly sequester Ca²⁺ released by active InsP₃R channels, spatial and temporal resolution of observed Ca²⁺ signals were sufficiently improved so that Ca²⁺ release by single InsP₃R channels can be imaged. With such “optical patch-clamping”, many individual channels can be monitored in their native environment in intact cells, thereby preserving the interaction between an active InsP₃R channel with neighboring channels mediated by Ca²⁺ released by the active channel. However, to date, the temporal resolution and signal-to-noise ratio of electrophysiological patch clamping are still substantially superior to those of optical patch clamping. Thus, optical and electrophysiological patch clamping are mutually complimentary techniques to study InsP₃R-mediated Ca²⁺ signals.

Another factor besides the intracellular location of the InsP₃R that complicates the study of InsP₃R channels is the primary amino acid sequence diversity of the channels. Vertebrates have three isoforms of InsP₃R: types 1 (InsP₃R-1), 2 (InsP₃R-2) and 3 (InsP₃R-3) that are encoded by three separate genes and are ~ 60–80% homologous. Invertebrates have only one InsP₃R that is most closely related to InsP₃R-1. Alternate splicing further enhances diversity of InsP₃R. InsP₃R-1 has three major splice regions (S1, S2 and S3), and a few minor ones. Mammalian neuronal cells express the S2+ variant while other peripheral cells express mainly the S2− form. InsP₃R-2 has at least one splice region. The invertebrate InsP₃R also has multiple splice variants. Given that most vertebrate cells express multiple InsP₃R isoforms in various levels, and InsP₃R can form homo- and heterotetrameric channels, the diversity at the InsP₃R channel level can be impressive indeed (see [11] and references therein).

To avoid possible complications arising from heterotetrameric channels, initial studies of InsP₃R channel used cells or tissues that were known to express predominantly one InsP₃R isoform [31], like cerebellum [32] for InsP₃R-1 S2+, Xenopus oocytes [15, 16] for InsP₃R-1 S2− [33], ventricular cardiac myocytes [34] for InsP₃R-2, RIN-5F insolinoma cells [35] for InsP₃R-3, and insect Spodoptera frugiperda Sf9 cells [36] for invertebrate InsP₃R. Nevertheless, the possibility that these cells may express different splice variants of the same InsP₃R isoforms cannot be ruled out. To address that concern, recombinant InsP₃R were over-expressed in cells determined to have low expression levels of endogenous InsP₃R, like Xenopus oocytes [37], COS-1 [38], COS-7 [39] and Sf9 cells [40]. However, all of those cells still express endogenous InsP₃R that can oligomerize with the recombinant InsP₃R to form contaminating heterotetrameric channels. A mutant cell line (DT40-3KO) derived from a chicken B cell line was finally generated with all three endogenous InsP₃R genes disrupted [41]. To date, it is still the only cell line available that has truly no endogenous InsP₃R background. This cell line has been used to generate cells stably expressing a single recombinant InsP₃R isoform (mutant or wild type) for single-channel studies of homotetrameric channels [42].

2. Conduction properties of InsP₃R

The conduction properties of a channel are one kind of information about protein ion channels that can only be obtained through single-channel studies. Such data in sufficient quantity can be used to inform the modeling of the structure of the channel pore [43, 44]. Practically, conduction characteristics, especially the channel conductance, are used to confirm the identity of the channel observed, especially in nuclear patch clamp experiments in on-nucleus or lum-out configurations, for which verifying the InsP₃ dependence of the observed channels by changing [InsP₃] in the cytoplasmic (pipette) solution is difficult.

2.1. Monovalent cation conductance

The most relevant channel conductance value is that observed when the channel is in physiological ionic conditions: symmetric [K⁺] of ~140 mM. This was not measured in planar lipid bilayer experiments [45, 46], probably due to the presence of contaminating channels in the reconstituted preparation. In nuclear patch clamp experiments in the absence of divalent cations, InsP₃R channels have linear, ohmic current-voltage (I–V) relation over a broad range of applied voltages (± 60 mV). Conductance values given below were measured for InsP₃R channels in outer nuclear membrane in symmetric 140 mM KCl unless stated otherwise.

The endogenous InsP₃R-1 S2− channel in Xenopus oocytes has a conductance of 370 ± 5 pS [47]. The endogenous rat InsP₃R-1 S2+ channel in the inner nuclear membrane of Purkinje neurons has a conductance of 356 ± 4 pS in 150 mM KCl [20]. Comparable conductance values of 366 ± 5 pS and 378 ± 11 pS (in 150 mM KCl) were observed for endogenous rat InsP₃R-1 S2+ channels in the inner nuclear membrane of CA1 pyramidal neurons and dentate gyrus granule neurons, respectively [22]. The endogenous InsP₃R-1 S2+ channel in mouse embryonic cortical neurons has a conductance of 375 pS [48]. Recombinant rat InsP₃R-1 S2− and S2+ channels expressed in COS-7 cells have conductance values of 389 ± 5 pS and 369 ± 6 pS, respectively [39]. Recombinant rat InsP₃R-1 S2+ channel expressed in DT40-3KO cells exhibits a conductance of 373 ± 2 pS [49].

A conductance of 375 ± 5 pS was observed for recombinant InsP₃R-2 channel expressed in DT40-3KO cells [50], identical to that for InsP₃R-1 S2+ channel measured by the same lab [49].

A conductance of 358 ± 8 pS was observed for InsP₃R channels in Xenopus oocytes with low expression levels of endogenous InsP₃R-1 injected with recombinant rat InsP₃R-3 cRNA [37]

These conductance values measured by different labs from endogenous and recombinant channels of different InsP₃R isoforms in multiple cell types were in remarkable agreement, suggesting that the high degree of sequence homology in the pore-forming regions (the pore helix, the selectivity filter together with transmembrane helices 5 and 6) among the three vertebrate InsP₃R isoforms [51] plays a crucial role in determining the channel conductance.

However, the picture is not so simple. Our lab that recorded the conductance of InsP₃R-3 channel expressed by cRNA injection in Xenopus oocytes also reported a significantly bigger conductance of 545 ± 7 pS for recombinant rat InsP₃R-3 channel expressed in DT40-3KO cells [52]. Furthermore, another lab consistently reported two significantly lower conductance values of ~ 200 pS and ~ 120 pS for both recombinant rat InsP₃R-1 S2+ [23, 53] and rat InsP₃R-3 channels [54, 55] expressed in DT40-3KO cells. It was unlikely that the lower conductance values arose from substates occasionally observed in InsP₃R channels [15, 52] since channels in the same nuclear membrane patches always exhibited the same conductance, and transitions from one conductance to another were never observed in single-channel membrane patches [54]. Other labs studying the same homotetrameric InsP3R channels in the same DT40-3KO system have not reported such dual conductance values for the same kind of InsP3R channel, and there is as yet no satisfactory explanation for the phenomenon.

Interestingly, a similar low conductance of ~ 200 pS was observed by two different labs using whole-cell patch clamping for plasma membrane-localized InsP₃R channels formed by endogenous chicken InsP₃R in wild-type DT40 cells [23]; and recombinant rat InsP₃R-1 S2+ [23, 25], rat InsP₃R-1 S2− [27] and mouse InsP₃R-2 [26] expressed in DT40-3KO cells.

These diverse conductance values for InsP₃R channels suggest that besides the primary sequence of the pore-forming regions of the InsP₃R, other factors, like the membrane environment around the channel, can also substantially affect the conductance of InsP₃R channels.

The conductance of 475 pS [48] for endogenous InsP₃R channels in human B lymphoblasts [48] is probably that of a heterotetrameric InsP₃R channel since it is different from any of the conductance values reported for homotetrameric channels.

The invertebrate InsP₃R is most closely related to InsP₃R-1 in its sequence. Endogenous invertebrate InsP₃R from Sf9 cells has a conductance of 477 ± 3 pS [36]. It is not sure if the similarity of this conductance value to that of the lymphoblast channel is coincidence or not.

2.2. Channel conductance in the presence of divalent cations

Localized to the ER membrane, InsP₃R channels are exposed to symmetric ~ mM Mg²⁺ and [Ca²⁺] of a few hundred μM on their luminal side. In such ionic conditions, the I-V relation of the channel becomes nonlinear with significantly lower slope conductance values [15, 20, 22, 37, 47, 52]. This indicates that divalent cations are permeant ion blockers of the InsP₃R channels due to the presence of high affinity divalent cation binding site(s) in the channel pore. Although divalent cations can pass through the channel pore, they bind strongly to some site(s) in the pore and interfere with the flow of K⁺ through the channel.

For InsP₃R-3 channels expressed in DT40-3KO cells, divalent cations can only block the InsP₃R channel partially and Mg²⁺ blocks the channel to the same extent whether it is on the cytoplasmic side or luminal side of the channel, suggesting that the channel has one unique saturable divalent cation binding site in the ion permeation pathway that is equally accessible from either side of the channel. This site has a substantially higher affinity for Ca²⁺ over Mg²⁺ since the half-maximal blocking concentration for Mg²⁺ is about twice that for Ca²⁺. However, at saturating concentrations, both Ca²⁺ and Mg²⁺ block the channel to the same extent [52].

Permeability ratios of various ions through different InsP₃R channels can be evaluated from reversal potentials measured in asymmetric ionic conditions using the general Goldman-Hodgkin-Katz equation [52] (tabulated in Table 1). The permeability sequence of P_Ca > P_Ba > P_Mg > P_K > P_Cl is generally preserved for all InsP₃R channels in nuclear envelope studied. The ratios show that even though the InsP₃R is a Ca²⁺-release channel, it is relatively non-selective for Ca²⁺ compared with plasma membrane Ca²⁺ channels. For a cation channel, it even has rather high permeability for Cl⁻. The high permeability of InsP₃R channels for K⁺ and Cl⁻ suggests that Ca²⁺ efflux from the ER lumen through an open InsP₃R channel to the cytoplasm during an InsP₃-induced Ca²⁺ release event can be sufficiently balanced by countercurrents of K⁺ and Cl⁻through the same open channel to maintain electrical neutrality of the ER lumen [56].

Table 1.

Permeability ratios of InsP₃R channels in nuclear envelope

InsP3R channel	Permeability ratios									references
	PCa		PBa		PMg		PK		PCl
Xenopus InsP3R-1a	7.6	:	4.3	:	3.2	:	1	:	0.23	[11]
Rat InsP3R-1a	6.6	:	6.5	:			1			[22]
Sf9 InsP3Ra	10	:			6.8	:	1	:	0.22	[36]
Rat InsP3R-1b	4.3	:					1	:	0.07	[39]
Rat InsP3R-3b	15.2	:	11.8	:	10.2	:	1	:	0.27	[52]

^aendogenous channels

^brecombinant channels

2.3. Unitary Ca²⁺ current through the InsP₃R channel in physiological ionic conditions

Since the fluxes of Ca²⁺ and K⁺ through the InsP₃R channel are not independent, the magnitude of the Ca²⁺ flux through the channel under physiological ionic conditions cannot be evaluated using the Goldman-Hodgkin-Katz equation and must be determined experimentally. Despite the large K⁺ conductance of InsP₃R channels, the relatively low Ca²⁺ permeability of the channels and the low (~ 0.5 mM) [Ca²⁺] gradient across the ER membrane suggest that the Ca²⁺ current driven through a single open InsP₃R channel by the ER [Ca²⁺] gradient (unitary Ca²⁺ current) is probably < 200 fA. Such signals are less than the noise (instrumental and thermal) observed in nuclear patch clamp experiments.

To overcome this obstacle to measure i_Ca [52], instead of trying to measure the unitary Ca²⁺ current (i_Ca) driven by a physiological [Ca²⁺] gradient at constant zero applied voltage (V_app), a voltage ramp was applied so that for part of the ramp (magenta bars at top of Fig. 2A and C), the open channel current (I_open, blue data points in Fig. 2A and C) could be clearly distinguished from the closed channel leak current (I_closed, pink data points in Fig. 2A and C). I_open and I_closed data points were manually selected and linear fitted (purple and red lines in Fig. 2). To measure i_Ca with high accuracy, any offset voltage in the instrument must be compensated precisely. This was done for each experiment by applying rapid perfusion techniques to excised lum-out outer nuclear membrane patches. With 140 mM KCl and no Ca²⁺ in both the pipette and bath solutions, no current would pass through the non-selective leak or the open Ca²⁺-selective InsP₃R channel when there was no electrical potential gradient across the isolated nuclear membrane. Therefore the intersection of the lines fitted to I_open and I_closed (indicated by orange arrowhead in Fig. 2B) marked the zero current level for that experiment (black dotted line in Fig. 2B). When the bath solution was changed to one containing physiological levels of free Ca²⁺ on the luminal side of the channel (Ca²⁺_ER) besides 140 mM KCl, the non-selective I_closed should still be zero when there was no electrical potential gradient across the membrane (orange arrowhead in Fig. 2D). The current passing through the open InsP₃R channel at that point was therefore i_Ca (green double arrow in Fig. 2D).

Fig. 2. Measurement of unitary Ca²⁺ current through a type 3 InsP₃R channel in physiological ionic conditions.

(A) Linear fits (purple and red lines) to selected open-channel current (I_open) and closed-channel leak current (I_closed) vs. V_app data (blue and pink points, respectively) from a single-channel nuclear patch clamp experiment in excised lum-out configuration under symmetric ionic conditions with pipette and perfusion solutions containing 140 mM KCl, 0.5 mM MgCl₂, and 3 μM free [Ca²⁺]. Pipette solution contained 2 μM InsP₃. (B) Fitted I_open-V_app and I_closed-V_app lines in I-V_app region marked by the black rectangle in (a). Zero-current level (black dotted line) established at the intersection of the I-V_app fits at V_app = 0 (marked by orange arrowhead). (C) Linear fits to I_open and I_closed data from V_app ramps recorded for the same membrane patch under asymmetric ionic conditions with perfusion solution containing 2 mM CaCl₂. Color and symbol conventions, I and V_app ranges are the same as in (a). (D) Fitted I–V lines in the same I–V region and with the same zero-current level as (b). V_app = 0 (marked by orange arrowhead) at the intersection of the I_closed-V_app line and zero-current level. Unitary Ca²⁺ current is I_open at V_app = 0 (marked by green double arrow). Figure modified from [52].

For homotetrameric InsP₃R-3 expressed in DT40-3KO cells, i_Ca is linearly related to the free Ca²⁺ concentration gradient across the InsP₃R channel for physiological levels of free [Ca²⁺] in the ER lumen ([Ca²⁺]_ER), with a slope of 0.30 ± 0.02 pA/mM. The linear relation is not surprising since in the absence of electrical potential gradient, the [Ca²⁺] gradient across the ER membrane provides the only driving force to move Ca²⁺ across the channel. Interestingly, the magnitude of i_Ca is not affected by physiological levels of free [Mg²⁺] or free [Cl⁻] [52]. This may be due to the greater affinity for Ca²⁺ over Mg²⁺ of the divalent cation-binding site in the channel pore.

For a non-depleted ER Ca²⁺ store with [Ca²⁺]_ER ~ 0.5 mM, i_Ca ~ 0.15 pA [52]. This agrees reasonably well with estimation of Ca²⁺ flux (~ 0.1 pA) observed for individual InsP₃R channels in intact cells by optical patch clamping [30]. This is a substantial Ca²⁺ flux that can raise the [Ca²⁺]_i in the vicinity of an open InsP₃R channel to produce positive feed-through regulation of the gating properties of the open channel itself ([57] and section 3.3 below). Furthermore, Ca²⁺ flux through a single InsP₃R channel sufficiently activated by InsP₃ can recruit neighboring InsP₃R channels in a cluster through calcium-induced calcium release (CICR) into coordinated opening [58, 59] to generate elementary Ca²⁺ release events (puffs) [60].

3. Steady-state regulation of InsP₃R channel by InsP₃ and Ca²⁺_i

3.1. Regulation of InsP₃R channel gating by Ca²⁺_i

The biphasic regulation of InsP₃R channel activity (open probability, P_o) by Ca²⁺_i was the first regulatory mechanism for the channel explored after discovery of its activation by InsP₃ [32]. Even in saturating levels of [InsP₃] (see below), InsP₃R P_o is very low at resting (~ 50 nM) [Ca²⁺]_i. Channel P_o increases as [Ca²⁺]_i is raised, up to a maximum activating [Ca²⁺]_i. Beyond a minimum inhibitory [Ca²⁺]_i, increases in [Ca²⁺]_i reduce channel P_o (Fig. 3A). The biphasic regulation by Ca²⁺_i suggests that the InsP₃R has at least two functional cytoplasmic Ca²⁺-binding sites: a high-affinity activating one and a low-affinity inhibitory one. If the Ca²⁺ affinity of the activating site is substantially higher than that of the inhibitory site, then the activating site can be saturated at some [Ca²⁺]_i before the inhibitory site starts to bind Ca²⁺ appreciably, and the Ca²⁺-dependence curve of the channel has a plateau region between the maximum activating [Ca²⁺]_i and the minimum inhibitory [Ca²⁺]_i. If the Ca²⁺ affinity of the activating site is similar to or lower than that of the inhibitory site, then Ca²⁺ binding to the inhibitory site can occur even at [Ca²⁺]_i that is sub-saturating for the activating site, then the maximum activating [Ca²⁺]_i and the minimum inhibitory [Ca²⁺]_i coincide, and the Ca²⁺-dependence curve has a distinct peak.

Fig. 3. Ligand dependencies of gating of homotetrameric recombinant InsP₃R channels of various isoforms expressed in DT40-3KO cells.

(A) Typical single-channel on-nucleus patch-clamp current traces of InsP₃R-3 channels in sub-optimal (190 nM), optimal (2μM) and inhibitory (23 μM) [Ca²⁺]_i in the presence of saturating (10 μM) [InsP₃]. V_app = –40 mV. Arrow indicates closed channel baseline current level for these and all subsequent current traces. (B–D) [Ca²⁺]_i-dependencies of mean channel P_o in various [InsP₃] and [ATP⁴⁻] as tabulated, for rat InsP₃R-1, mouse InsP₃R-2 and rat InsP₃R-3 channels, respectively. Curves are fitted to mean P_o data points using the empirical biphasic Hill equation with 5 parameters [11]. Data in (b) and (c) are from [50]. (a) and (d) are modified from [57].

This regulation has been studied for both endogenous and recombinant channels of all three InsP₃R isoforms in many different cell types using both reconstitution into artificial lipid bilayer and nuclear patch clamping techniques. A remarkably wide variety of Ca²⁺ dependencies were observed, with both plateau- and peak-shaped dependence curves [11, 12]. In particular, the same recombinant InsP₃R channel expressed in different cell types can display very different Ca²⁺ dependencies even when studied with the same nuclear patch clamping technique. Rat InsP₃R-3 channels expressed in DT40-3KO cells [57] are more than ten-fold less sensitive to Ca²⁺ activation than those expressed in Xenopus oocytes [61]. Rat InsP₃R-1 S2+ channels expressed in DT40-3KO cells [50] are activated significantly more abruptly within a narrower range of [Ca²⁺]_i (more cooperatively) than those expressed in COS-7 cells [39]. This suggests that Ca²⁺ regulation of InsP₃R channel activity can be quite labile, affected by various cytoplasmic factors and conditions besides the primary sequence of the channels, thereby allowing the magnitude and even the nature of the InsP₃R-mediated intracellular Ca²⁺ signals to be delicately tuned by multiple regulatory factors besides the primary InsP₃R ligands. The inhibitory site of the endogenous Xenopus InsP₃R-1 can even be rendered completely insensitive to Ca²⁺ irreversibly by temporary exposure to non-physiological low [Ca²⁺]_i (< 10 nM) [62]. Such environmental disruption of the inhibitory Ca²⁺ site is probably the reason behind reports of lack of Ca²⁺ inhibition for some InsP₃R channels [34, 35, 46].

Another known factor that modifies Ca²⁺ sensitivities of InsP₃R is ATP⁴⁻, which allosterically enhances the sensitivity to Ca²⁺ activation of endogenous Xenopus InsP₃R-1 channel [11], recombinant rat InsP₃R-3 channel expressed in Xenopus oocytes [63], and recombinant InsP₃R-1 channel expressed in DT40-3KO cells [50]. In recombinant InsP₃R-1 [50] and InsP₃R-3 [11] channels, ATP⁴⁻ also increases the sensitivity of the InsP₃R channels to Ca²⁺ inhibition, thus left-shifting the Ca²⁺-dependence curve. In contrast, ATP⁴⁻ does not affect the sensitivities of recombinant mouse InsP₃R-2 channels expressed in DT40-3KO cells to Ca²⁺ activation or inhibition [50]. Besides sensitivity to Ca²⁺ regulation, ATP⁴⁻ also modifies the cooperativity of Ca²⁺ activation of InsP₃R channels, but does not affect the cooperativity of their inhibition by Ca²⁺ [50, 63]. It is somewhat paradoxical that ATP⁴⁻ reduces the cooperativity of Ca²⁺ activation of rat InsP₃R-1 channels expressed in DT40-3KO cells [50] and rat InsP₃R-3 channels expressed in Xenopus oocytes [11] in a similar way, but does not affect the Ca²⁺ activation cooperativity of native InsP₃R-1 channels in Xenopus oocytes [11].

To control for various factors that affect Ca²⁺ regulation of InsP₃R channels, recombinant channels in DT40-3KO cells expressing the three isoforms individually should be studied under the same conditions (especially [ATP⁴⁻]) using the same experimental method. Although these experiments have been done [50, 57], [ATP⁴⁻] used in those experiments were not the same, complicating comparisons of Ca²⁺ regulation of homotetrameric channels of different InsP₃R isoforms. Fortunately, Ca²⁺ regulation of InsP₃R-2 channel is not sensitive to [ATP⁴⁻] [50] so it can be used as the reference to be compared with the channels of the other isoforms (Fig. 3B–D). In saturating [InsP₃] (see below), InsP₃R-3 channels are over 10-fold less sensitive to Ca²⁺ activation than InsP₃R-2, and about 4-fold less sensitive to Ca²⁺ inhibition (Fig. 3C and D). Ca²⁺ activation of InsP₃R-3 is also significantly less cooperative than that of InsP₃R-2. Cooperativity of Ca²⁺ inhibition is similar for the two isoforms. In 5 mM ATP⁴⁻, sensitivities to both Ca²⁺ activation and inhibition are similar for InsP₃R-2 and InsP₃R-1, but both Ca²⁺ activation and inhibition of InsP₃R-2 are more cooperative than those of InsP₃R-1 (Fig. 3B and C). In low 0.1 mM ATP⁴⁻, InsP₃R-1 channel is about 4-fold less sensitive to Ca²⁺ activation and nearly 15-fold less sensitive to Ca²⁺ inhibition than InsP₃R-2. However, Ca²⁺ activation of InsP₃R-1 channel is substantially more cooperative than that of InsP₃R-2, whereas its Ca²⁺ inhibition remains less cooperative than that of InsP₃R-2.

3.2. InsP₃ regulation of InsP₃R channel gating

InsP₃ is the natural ligand of InsP₃R channels [2] that activates the channel saturably. Three different forms of InsP₃ modulation of InsP₃R channel gating have been reported.

The first form reported is through tuning the sensitivity of the channel to Ca²⁺ inhibition. This has been observed in endogenous InsP₃R-1 S2− channels in Xenopus oocytes, endogenous Sf9 InsP₃R, recombinant rat InsP₃R-3 expressed in Xenopus oocytes [11] and DT40-3KO cells [57] studied using nuclear patch clamping (Fig. 3D), as well as in endogenous rat InsP₃R-1 channels reconstituted into lipid bilayers [64]. Even at saturating 10 μM InsP₃ (found to be saturating in all nuclear patch clamp studies of InsP₃R channels), Ca²⁺ at high enough concentrations still inhibits InsP₃R channels. When [InsP₃] is reduced to sub-saturating levels, reduction of [InsP₃] increases the sensitivity of the channel to Ca²⁺ inhibition, thereby decreasing the range of [Ca²⁺]_i within which InsP₃R channel activity can be observed. At low enough [InsP₃], the channel becomes more sensitive to Ca²⁺ inhibition than Ca²⁺ activation, and the channel starts to be inhibited before it is fully activated so the maximum channel P_o is decreased [11, 57]. In this kind of InsP₃ regulation, only the sensitivity of the channel to Ca²⁺ inhibition is modulated. The sensitivity of the channel to Ca²⁺ activation and the levels of cooperativity of the activation and inhibition are not affected [11, 57].

Another form of InsP₃ regulation of InsP₃R channel P_o is by changing the maximum channel activity level. This was first reported for endogenous InsP₃R-1 S2+ channels in the inner nuclear membrane of rat Purkinje neurons [20], and was also observed in recombinant rat InsP₃R-1 and mouse InsP₃R-2 channels expressed in DT40-3KO cells [50] (Fig. 3B and C). For these channels, as [InsP₃] increases through the sub-saturating range, the maximum channel activity level is increased without altering the sensitivity or cooperativity of Ca²⁺ activation and inhibition. Thus the shape of the Ca²⁺ dependence curve remains essentially unchanged.

In the above mentioned forms of InsP₃ and Ca²⁺ regulation of InsP₃R channel activity, changes in channel P_o were brought about mainly by changes in the mean duration of channel closings (t_c), which can vary over several orders of magnitude in response to changes in [InsP₃] and [Ca²⁺]_i, while the mean duration of channel openings (t_o) remain relatively unaffected by ligand concentrations[36, 50, 57, 61, 65]. This suggests that once an InsP₃R channel opens, it will stay open for a stereotypical period of time (t_o), during which it remains insensitive to changes in [InsP₃] and [Ca²⁺]_i, before it closes. This can limit the extent of the negative feed-through effect of the Ca²⁺ fluxing through an open InsP₃R channel to terminate elementary Ca²⁺ release events.

The last form of InsP₃ regulation of InsP₃R channel P_o is a complex and indirect one. A study of recombinant rat InsP₃R-3 and InsP₃R-1 channels expressed in DT40-3KO cells using nuclear patch clamp techniques [54, 55] reported that exposure to InsP₃ induces rapid clustering of InsP₃R channels that are normally randomly distributed in the outer nuclear membrane without association with other InsP₃R channels. This clustering alters the gating behavior of the channels. Clustered channels have lower sensitivity to InsP₃ activation. In saturating [InsP₃] but sub-optimal [Ca²⁺]_i, the clustered channels gate independently with identical kinetics, but with P_o half that of lone channels. Furthermore, the reduction in channel P_o is mainly due to reduction in channel t_o. In contrast, under optimal ligand conditions (saturating [InsP₃] and optimal [Ca²⁺]_i), clustered channels gate with the same overall P_o as lone channels, but exhibit coupled gating, i.e., in membrane patches with two active channels, the probabilities that the channels were both open or both closed are higher than predicted for two identical and independent channels. This regulation is controversial. Nuclear patch clamp studies in multiple systems, including recombinant InsP₃R-3 channels expressed in DT40-3KO cells, have demonstrated that InsP₃R channels in the outer nuclear membrane form clusters before exposure to InsP₃ [36, 37, 66, 67]. A study of InsP₃R-mediated elementary Ca²⁺ release events (Ca²⁺ puffs) in intact cells by TIRF microscopy revealed that those events arise from InsP₃R channel clusters established before exposure to InsP₃, and that the size of the channel clusters does not increase after exposure to InsP₃ [68]. Moreover, single-molecule tracking of InsP₃R revealed no short-term change in the motility or clustering pattern of InsP₃R following activation of the InsP₃-Ca²⁺ signaling pathway [69]. Importantly, channel gating analyses revealed no statistical difference between channel P_o in single- and multi-channel records obtained under sub-optimal ligand conditions, and no significant coupled gating behavior in two-channel records obtained under optimal or suboptimal ligand conditions for endogenous channels in Xenopus oocytes and Sf9 cells, and recombinant rat InsP₃R-3 channels expressed in Xenopus oocytes and DT40-3KO cells [67]. Thus, whether InsP₃-induced rapid clustering of InsP₃R channels that leads to modification of channel gating is a real or universal phenomenon is very much open to question.

Now that homotetrameric channels of the three InsP₃R isoforms have been studied in the same cell system (DT40-3KO), the sensitivity of those channels to InsP₃ activation can be compared properly. Comparing the values of the half-maximal [InsP₃] (EC₅₀) is complicated by the fact that those InsP₃R channels whose sensitivity to Ca²⁺ inhibition is tuned by [InsP₃], like the homotetrameric recombinant InsP₃R-3 channel expressed in DT40-3KO cells (Fig. 3D), have different EC₅₀ at different [Ca²⁺]_i. Low [InsP₃] (< 1 μM) is sufficient to activate the channel at low [Ca²⁺]_i (< 1 μM), but a much higher [InsP₃] (3–10 μM) is required to activate the channel at high [Ca²⁺]_i (> 10 μM) (Fig. 3D). It is much simpler to compare the maximum P_o that can be observed at the same sub-saturating [InsP₃] for those channels, as suggested in [50]. In 1 μM (sub-saturating) InsP₃, the maximum observed P_o are 0.27, 0.51 and 0.17 for InsP₃R-1, InsP₃R-2 and InsP₃R-3 channels, respectively. Thus, the InsP₃ sensitivity sequence is InsP₃R-2 > InsP₃R-1 > InsP₃R-3, which agrees with the sequence reported for InsP₃-binding affinities of the channels [70].

Like the sensitivity of the channels to [Ca²⁺]_i regulation, sensitivity of InsP₃R channels to InsP₃ activation is affected by factors other than the sequence of the InsP₃R. Although recombinant InsP₃R-3 channel has the lowest InsP₃ sensitivity among the three isoforms expressed in DT40-3KO cells, the same recombinant channel expressed in Xenopus oocytes has a maximum P_o of 0.77 in 33 nM InsP₃ [61], indicating a much higher InsP₃ sensitivity than any of the InsP₃R channels expressed in DT40-3KO cells. In fact, the native InsP₃R-1 channel in Xenopus oocytes has a similar InsP₃ sensitivity with maximum P_o of 0.77 in 33 nM InsP₃, suggesting that the unknown factor(s) in Xenopus oocytes that enhances the InsP₃ sensitivity of InsP₃R channels is the dominant effect on channel InsP₃ sensitivity.

3.3. Regulation of InsP₃R channel gating by [Ca²⁺]_ER

Whereas regulations of InsP₃R channel activities by [InsP₃] and [Ca²⁺]_i have been well studied, effects of [Ca²⁺]_ER on InsP₃R channels remain controversial. In a recent study [57], the rapid perfusion technique was applied in nuclear patch clamp experiments to examine the behaviors of the same InsP₃R channels in the same isolated nuclear membrane patches in different [Ca²⁺]_ER. When [Ca²⁺]_ER was switched from 70 nM, a level at which no Ca²⁺ flux was driven through the InsP₃R channel from the luminal side, to 0.3–1.1 mM, levels that generate significant Ca²⁺ flux, activities of the InsP₃R channels were indeed modified (Fig. 4A). However, the nature (activating or inhibitory) and magnitude of the modulation depended not only on [Ca²⁺]_ER (Fig. 4A vs. B), but also on cytoplasmic conditions, including [InsP₃] (Fig. 4B vs. C), [Ca²⁺]_i (Fig. 4B vs. D) and Ca²⁺ buffering capacity (Fig. 4C vs. E). Inhibition and activation of InsP₃R channel by [Ca²⁺]_ER were even completely abrogated by strong Ca²⁺ buffering on the cytoplasmic side together with V_app that opposed the [Ca²⁺] gradient to reduce the Ca²⁺ flux through the channel (Fig. 4F and G).

Fig. 4. Effects of [Ca²⁺]_ER on the gating of homotetrameric InsP₃R-3 channels expressed in DT40-3KO cells.

Typical current traces from nuclear patch clamp experiments in lum-out configuration under experimental conditions ([InsP₃], [Ca²⁺]_i, concentrations of Ca²⁺ chelator in pipette solutions solution and V_app) tabulated at top. Color bars at top indicate changes in [Ca²⁺]_ER achieved by rapid perfusion techniques. Blue bars at bottom mark the part of the current traces used to evaluate the channel P_o tabulated below. Figures are modified from [57]. For (G), V_app should be +70 mM.

Taken together, these observations indicate that Ca²⁺ in the ER lumen modifies InsP₃R channel P_o not by direct binding to some regulatory site(s) on the luminal side of the InsP₃R channel, but by a feed-through effect. Ca²⁺ driven by high [Ca²⁺]_ER through the open InsP₃R channel raises concentration of [Ca²⁺]_i in the vicinity of the channel and binds to the cytoplasmic activating and inhibitory Ca²⁺ sites to modulate the channel activity [57].

3.4. Graded recruitment of InsP₃R channels by favorable ligand conditions

The amount of Ca²⁺ flux (J) through a patch of ER membrane depends on three factors: the unitary Ca²⁺ current passing through an open InsP₃R channel (i_Ca), the activity level of the InsP₃R channels (P_o), and the total number of active channels in the membrane patch (N_A). All three factors can be individually measured using nuclear patch clamp electrophysiology. Specifically, N_A in an isolated membrane patch is taken to be the maximum number of open-channel current levels observed. To avoid undercounting N_A, and therefore overestimating P_o, only current records that lasted long enough to ensure a high degree of accuracy (>95% confidence) in determining N_A were used for N_A and P_o evaluation [36, 71]. By counting N_A in tens of membrane patches for each of a dozen different combinations of [InsP₃] and [Ca²⁺]_i in a nuclear patch clamp study of the endogenous Sf9 InsP₃R channel, it was demonstrated that on average, N_A observed in the presence of favorable ligand conditions (optimal [Ca²⁺]_i, saturating [InsP₃] or both) was significantly higher than that observed in sub-optimal or inhibitory [Ca²⁺]_i, or sub-saturating [InsP₃] [36]. However, in that study, it was not certain whether factors other than ligand concentrations that can affect N_A, like the tip size of the patch clamp microelectrodes used and the topology of the membrane patch isolated at the tip of the microelectrodes, were sufficiently controlled. Modulation of N_A by ligand concentrations was more convincingly demonstrated in a later study that counted N_A observed in the same isolated nuclear membrane patch when its cytoplasmic side was exposed to various ligand concentrations. This was done by performing rapid perfusion exchange of the bath solution in nuclear patch clamp experiments in the cyto-out configuration [71]. N_A increased incrementally as [Ca²⁺]_i was increased from resting through sub-optimal to optimal levels in saturating [InsP₃]. Similarly, more channels were active in saturating [InsP₃] than in sub-saturating [InsP₃] in the same membrane patch. This was observed for the endogenous InsP₃R channels in Sf9 cells, which have only one InsP₃R gene. More important, this was also observed for homotetrameric recombinant InsP₃R-3 channels expressed in DT40-3KO cells, suggesting that even a homogeneous population of InsP₃R channels can exhibit heterogeneous sensitivity to Ca²⁺_i and InsP₃ activation. Graded recruitment of InsP₃R channel by ligands was completely abrogated by treatment with reducing reagents and restored by treatment with oxidizing reagents, suggesting that the heterogeneous ligand sensitivity is regulated largely by post-transcriptional redox modulation of the channels.

In a population of InsP₃R channels with heterogeneous sensitivities to ligands, those channels with higher ligand sensitivity may act as “hot spots” to initiate fundamental Ca²⁺-release events (blips) [72] as [InsP₃] starts to rise in response to extracellular stimuli. The rise in [Ca²⁺]_i not only enhances the activities of the InsP₃R channels that are already activated, but also recruits more InsP₃R channels with lower [Ca²⁺]_i sensitivities to further amplify the magnitude of the Ca²⁺-induced Ca²⁺ release to propagate the Ca²⁺ signal. Thus, graded recruitment of InsP₃R channels with favorable ligand conditions can be an important mechanism in both the generation and propagation of intracellular Ca²⁺ signals.

3.5. Ligand regulation of modal gating behaviors of InsP₃R channels

In some experimental systems for nuclear patch clamping, like the native Sf9 InsP₃R and recombinant InsP₃R expressed in DT40-3KO cells, the activity of individual InsP₃R channels can be stably recorded for extensive periods (tens of minutes) [71, 73]. Examination of such long single-channel records of Sf9 InsP₃R revealed that even under constant ligand conditions ([InsP₃] and [Ca²⁺]_i), channels exhibited relatively occasional (one to two orders of magnitude less frequent than channel gating), spontaneous and abrupt transitions among three distinct gating patterns (modes) (Fig. 5A): a low-activity (L) mode with long channel closings and brief channel openings (P_o = 0.007), an intermediate (I) mode with rapid channel gating kinetics (P_o = 0.24), and a high-activity (H) mode with prolonged bursts of channel activity (P_o = 0.85)[73]. These modes were observed under all ligand conditions and remarkably, channel gating kinetics within a mode were not significantly altered by [Ca²⁺]_i and [InsP₃] (Fig. 5B). Thus, the graded dependencies of InsP₃R channel P_o on [Ca²⁺]_i and [InsP₃] of various InsP₃R channels (Fig. 3B–D) is mostly the result of continuous shifting of the relative probabilities of the channel being in these modes, especially the L and H modes (Fig. 5C) as the ligand concentrations vary. Interestingly, only the mean duration of the H mode was significantly affected by ligand concentrations, while those of the L and I modes remained within a narrow range for the [InsP₃] and [Ca²⁺]_i examined. The ligand independence of the long mean duration of the L mode (〈τ^L〉) suggests that once a channel enters the L mode, it becomes insensitive to [InsP₃] and [Ca²⁺]_i, and will remain in the L mode with very low for a stereotypical long duration ~〈τ^L〉. This may be one of the mechanisms to terminate an InsP₃R-mediated Ca²⁺ release event (puff), especially when the local cytoplasmic ligand conditions are not optimal so the channels have high probability of entering the L mode.

Fig. 5. Modal gating behaviors of endogenous insect Sf9 InsP₃R.

(A) Black traces are baseline-subtracted current records obtained in saturating 10 μM InsP₃, with sub-optimal (100 nM), optimal (1 μM) and inhibitory (89 μM) [Ca²⁺]_i, as tabulated. Blue and purple traces are idealized gating records before and after burst analysis, respectively, derived from the corresponding current records. Color bars at bottom indicate the gating mode the InsP₃R channel was in during the shown records (red, orange and green for L, I and H modes, respectively). (B) Channel open probability in each of the three modes (P_o^M), (C) probability of the channel being in each of the three modes (π^M), and (D) mean durations 〈τ^M〉 of each of the three modes observed are plotted against the ligand concentrations in which they were observed. Note the non-linear y-axis used in (B) to better show the low P_o^L values for the L mode, and the logarithmic scale used in (D). In (B-D), the top x-axis labels indicate the cytoplasmic free calcium concentrations, the bottom labels indicate the cytoplasmic InsP₃ concentrations. Figure modified from [12] and [73].

Subsequently, such modal behaviors have also been observed for endogenous InsP₃R channels from human lymphoblasts, mouse embryonic fibroblasts and mouse primary embryonic cortical neurons [48]. In these channels, modal switching is not only the major mechanism for physiological ligand regulation of InsP₃R channel activities, but is also the mechanism behind the pathogenic dysregulation of InsP₃R channel by familial Alzheimer’s disease-linked mutant presenilins. Interaction between the mutant presenilins and InsP₃R channels greatly increases the propensity of InsP₃R channels being in the H mode, thereby causing exaggerated Ca²⁺ release through those channels [48].

Recently, modal gating was observed in recombinant rat InsP₃R-1 S2+ and mouse InsP₃R-2 expressed in DT40-3KO cells [50]. Both channels exhibit two gating modes: a “park” mode in which the channels only open briefly and very occasionally (P_o < 0.01) and a “drive” mode with bursting channel activities (P_o = 0.64) [74]. Within a mode, channel gating kinetics are practically the same in all [Ca²⁺]_i, [InsP₃] and [ATP⁴⁻]. In InsP₃R-1 channel, more favorable [Ca²⁺]_i both prolongs the drive mode duration and shortens the park mode duration; more favorable [InsP₃] mainly shortens the park mode duration and lengthens the drive mode to a smaller extent; and increase in [ATP⁴⁻] shifts the effects of [Ca²⁺]_i on the modal durations towards lower [Ca²⁺]_i levels. Ligand regulation of model behaviors of InsP₃R-2 channel is more complicated. In saturating [InsP₃], [ATP⁴⁻] does not affect durations of the modes and favorable [Ca²⁺]_i lengthens drive mode duration and shortens park mode duration as in InsP₃R-1. At low [InsP₃] and saturating [ATP⁴⁻], more favorable [Ca²⁺]_i only shorten the park mode duration but do not change the drive mode duration. When both [InsP₃] and [ATP⁴⁻] are low, the channel has low P_o with long park mode duration and short drive mode duration over most [Ca²⁺]_i; and drive mode duration is further reduced and park mode duration is further lengthened at extreme [Ca²⁺]_i. The insensitivity of the mean duration of the L/park mode to ligand conditions observed in the insect Sf9 InsP₃R channel was not observed in the mammalian homotetrameric InsP₃R-1 or InsP₃R-2 channels. The durations of both the park and drive modes are affected by [InsP₃], [Ca²⁺]_i and [ATP⁴⁻] in most ligand conditions. Thus, unlike the case for Sf9 InsP₃R channels, modal gating kinetics probably does not play a significant role in the termination of InsP₃R-mediated intracellular Ca²⁺ release events involving these channels.

Because of the consistency of the P_o of the InsP₃R channel within a gating mode over a broad range of ligand conditions (Fig. 5B and [50]), modal gating behavior of the InsP₃R channels offers a way to greatly simplify the modeling of Ca²⁺ release through these channels. The gating kinetics (opening and closing) of the channels within a mode can be ignored. An InsP₃R channel in one gating mode can be considered as a channel in a conductance substate with Ca²⁺ conductance of (i_Ca)(P_o^M), where P_o^M is the P_o of the channel in that gating mode. The kinetics of the transition of the channel among the conductance substates is the kinetics of the modal transitions of the channel, which is significantly slower than the gating kinetics of the channel and therefore simpler to simulate.

3.6. Ligand regulations of heterotetrameric InsP₃R channels of known isoform composition

Although most cell types examined express, at various levels, multiple isoforms of InsP₃R [31, 75], which can form heterotetrameric channels [11], there has been no experimental technique to study the properties of individual heterotetrameric InsP₃R channels with known isoform composition until recently. The exciting technological break-through of expressing concatenated InsP₃R dimers in DT40-3KO cells to form properly localized and functional InsP₃R channels enables the study of the properties of heterotetrameric InsP₃R channels of a specific isoform composition [76]. In the pioneering study, on-nucleus patch clamp experiments demonstrated that the conductance of InsP₃R channels formed by concatenated InsP₃R-1:InsP₃R-1 (R1R1) dimers is identical to that of homotetrameric InsP₃R-1 channels. Furthermore, the gating and modal behaviors, and their regulation by [InsP₃] and [ATP⁴⁻] at 200 nM [Ca²⁺]_i of channels formed by R1R1 dimers resembled those of homotetrameric InsP₃R-1 channels, and those of channels formed by concatenated R2R2 dimers resembled those of homotetrameric InsP₃R-2 channels. Remarkably, those single-channel properties of channels formed by concatenated R1R2 dimers were indistinguishable from those of homotetrameric InsP₃R-2 channels, suggesting that InsP₃R-2 dominates over InsP₃R-1 in determining the gating properties of a heterotetrameric channel formed by the two. Characterizing the conductance and gating properties of heterotetrameric InsP₃R channels with known isoform composition can provide valuable insights into the physiological reasons behind the sequence diversity of the channel and the intricate regulation of the expression levels of InsP₃R with different sequences [11].

4. Dynamic responses of InsP₃R channel to rapid changes in ligand concentrations

Although studying InsP₃R gating properties in constant [InsP₃] and [Ca²⁺]_i has provided much insight into ligand regulation of InsP₃R channels, [Ca²⁺]_i, and to a lesser extent [InsP₃], change dynamically during Ca²⁺ signaling events. Therefore, to get a better understanding of how InsP₃R channels function and contribute to intracellular Ca²⁺ signaling, it is essential to study the behaviors of InsP₃R channels under non-steady ligand conditions.

The kinetic responses of single Sf9 InsP₃R channels to abrupt changes in ligand concentrations were examined with millisecond resolution by applying rapid perfusion techniques to cyto-out nuclear membrane patches [77]. This study provided the first and, to date, the only measurements of the lag times of InsP₃ activation and deactivation (Fig. 6A and B), Ca²⁺ activation and deactivation (Fig. 6C and D), activation and deactivation by simultaneous changes in [Ca²⁺]_i and [InsP₃] (Fig. 6E and F), Ca²⁺ inhibition and recovery from inhibition (Fig. 6G and I), and the durations of fundamental single-channel Ca²⁺-induced Ca²⁺ release events (Fig. 6J). While the majority of [Ca²⁺]_i jumps from < 10 nM to 300 μM in 10 μM InsP₃ activated Ca²⁺ release bursts (Fig. 6J), a non-trivial fraction of the jumps failed to activate any burst due to Ca²⁺ binding to the inhibitory sites of the channel before the activating sites got occupied. This showed that Ca²⁺ binding to those sites are non-sequential. Similarly, in some of the [Ca²⁺]_i drops from 300 μM to < 10 nM, the channel recovered from Ca²⁺ inhibition before Ca²⁺ unbound from its activating site, generating a brief burst of channel activity (Fig. 6K), showing Ca²⁺ unbinding from the activating and inhibitory sites are also non-sequential. The study also clearly demonstrated the absence of long-term adaptation [78] behaviors in InsP₃R channels following changes in [InsP₃] or [Ca²⁺]_i (Fig. 6L and M, respectively). The surprisingly long times InsP₃R channels took to recovery from Ca²⁺ inhibition (Fig. 6I) suggest that this may contribute significantly to the refractory interval between consecutive Ca²⁺ release from the same site in the ER [79]. The lag times for channel activation by [InsP₃] jumps from 0 to 100 μM were only modestly shorter than those for [InsP₃] jumps from 0 to 10 μM (Fig. 6H vs. A), indicating that InsP₃ binding to the channel, and the subsequent conformation changes that lead to the channel arriving at its open conformation(s) both contribute significantly to the InsP₃ activation lag times. In contrast, the lag times for channel activation by [Ca²⁺]_i jumps from < 10 nM to 300 μM were substantially shorter than those for [Ca²⁺]_i jumps from < 10 nM to 2 μM (Fig. 6J vs. C), suggesting that Ca²⁺ binding to the channel is the major contribution to the Ca²⁺ activation lag times. Furthermore, the study revealed an unexpected cooperativity between Ca²⁺ and InsP₃ modulation of the channel: in the absence of InsP₃, Ca²⁺ binding to an InsP₃R channel stabilizes it in a closed conformation so that a jump in [InsP₃] in the constant presence of high (2 μM) [Ca²⁺]_i takes significantly longer to activate the channel than when [InsP₃] and [Ca²⁺]_i are increased simultaneously (Fig. 6A vs. E). On the other hand, in the presence of InsP₃, Ca²⁺ binding to an InsP₃R channel stabilizes it in an open conformation so that decreases in [InsP₃] in the constant presence of 2 μM [Ca²⁺]_i take longer to deactivate the channel than when both [InsP₃] and [Ca²⁺]_i are dropped simultaneously (Fig. 6B vs. F). The mean duration of the L mode of the Sf9 channel (Fig. 5D) is substantially longer than the lag times of activation of the channel by abrupt jumps in [InsP₃] (Fig. 6A), or in [Ca²⁺]_i (Fig. 6C), or in both [Ca²⁺]_i and [InsP₃] (Fig. 6E). This suggests that the channel has a complex kinetic gating scheme with the channel in distinct closed kinetic states depending on the occupancies of its activating Ca²⁺ and InsP₃ binding sites. These different closed kinetic states are connected to different open kinetic states, as described in [80]. Thus, when the channel is activated by jumps in [InsP₃] or [Ca²⁺]_i, it is exiting from closed kinetic states that are different from the closed kinetic state(s) involved when the channel transits from the L to the I or H modes, with different transition kinetics.

Fig. 6. Dynamic responses of endogenous insect Sf9 InsP₃R to abrupt changes in [Ca²⁺]_i and [InsP₃] observed in cyto-out nuclear patch clamp experiments.

Color bars at top indicate changes in [Ca²⁺]_i and [InsP₃] achieved by rapid exchange of bath solutions. Times when such exchanges occurred were marked by changes in closed-channel current levels due to difference in [KCl] (100 mM vs. 70 mM) in the bath solutions used. Gray bars at bottom indicate the lag times between ligand concentration switches and the resulting changes in InsP₃R channel gating behaviors. Two arrows indicate the closed-channel current levels before and after the bath solution switches for each current traces. In (J), the dark gray bar indicates the Ca²⁺ activation lag time and the light gray bar indicates the Ca²⁺ inhibition lag time. Note that a different time scale is used for (L) and (M). Figure modified from [77].

5. Moving forward

The InsP₃R channel is unusual in that even its fundamental properties: single-channel conductance and regulation by InsP₃ and Ca²⁺, are apparently very mutable. Not only are the sensitivities and degrees of cooperativity of ligand regulation different among channels of different isoforms, but also the ways InsP₃ regulates those different channels are fundamentally different. Even the conductance and the sensitivity to InsP₃ activation of the same recombinant homotetrameric channel expressed in different cells can be radically different. Therefore due caution should be exercised when attempting to extrapolate single InsP₃R channel properties observed in one experimental system to another system. It is advisable to characterize the basic properties: conductance and [Ca²⁺]_i and [InsP₃] dependencies, of InsP₃R channels in a new experimental system, using buffer solutions with commonly used physiological compositions (140–150 mM KCl, 0.5 mM ATP⁴⁻ and pH 7.3) for easier comparison with previous studies.

The recent development of the capability of expressing concatenated InsP₃R dimers to form functional InsP₃R channels in DT40-3KO cells [76] offers the opportunity to study heterotetrameric InsP₃R channels with known isoform compositions. That study also demonstrated that concatenated InsP₃R tetramers could also be expressed in DT40-3KO cells, though in significantly lower levels than concatenated dimers. If the expression of concatenated tetramers can be enhanced, then properties of channels formed by all three InsP₃R isoforms, like R1R2R3R1, and channels with asymmetric isoform compositions, like R1R1R1R2, can be studied and compared with properties of native InsP₃R channels in various cells to shed light on the physiological significance of the molecular diversity of the InsP₃R.

Many computational kinetic models have been developed over the years to account for the observed InsP₃ and Ca²⁺_i regulation of single InsP₃R channel gating ([81] and references therein). These kinetic models were then used to simulate the generation and propagation of more complex intracellular Ca²⁺ signals (blips, puffs and waves) by organized clusters of InsP₃R channels [72]. Even though recent experiments demonstrated that modulation of modal switching is the main mechanism for steady-state ligand regulation of InsP₃R channel activities [50, 73], only a few models have been developed to account for InsP₃R modal behaviors [74, 80]. Furthermore, the roles of modal behaviors in the dynamic responses of InsP₃R channel to changes in [InsP₃] and more importantly [Ca²⁺]_i have barely been investigated [80]. Much work needs to be done to develop kinetic models that can adequately describe the complex gating behaviors of single InsP₃R channels to enable proper simulation of organized gating of multiple InsP₃ channels to generate local and global Ca²⁺ release events.

Highlights.

• InsP₃R single-channel properties studied by nuclear patch clamping are reviewed

• We examine InsP₃R K⁺ conductance and unitary Ca²⁺ current in physiological solutions

• Steady-state Ca²⁺ and InsP₃ dependencies of channels of various isoforms are compared

• Ligand-regulated modal gating and graded recruitment of InsP₃R channels are examined

• Responses of InsP₃R channels to abrupt ligand concentration changes are presented