Differential regulation of the InsP₃ receptor type-1 and -2 single channel properties by InsP₃, Ca²⁺ and ATP

Abstract

An elevation of intracellular Ca²⁺ levels as a result of InsP₃ receptor (InsP₃R) activity represents a ubiquitous signalling pathway controlling a wide variety of cellular events. InsP₃R activity is tightly controlled by the levels of the primary ligands, InsP₃, Ca²⁺ and ATP. Importantly, InsP₃Rs are regulated by in a biphasic manner. Ca²⁺ release through all InsP₃R family members is also modulated dramatically by ATP, albeit with sub-type-specific properties. To ascertain if a common mechanism can account for ATP and Ca²⁺ regulation of these InsP₃R family members, we examined the effects of [ATP] on the Ca²⁺ dependency of rat InsP₃R-1 (rInsP₃R-1) and mouse InsP₃R-2 (mInsP₃R-2) activity expressed in DT40-3KO cells. We used the on-nucleus patch clamp recording technique with various [ATP], [InsP₃] and [Ca²⁺] in the patch pipette and measured single InsP₃R channel activity in stably transfected DT40 cells. Under identical conditions, at saturating [InsP₃] and [ATP], the activity of rInsP₃R-1 and mInsP₃R-2 was essentially identical in terms of single channel conductance, maximal achievable open probability (P_o) and the [Ca²⁺] required for activation and inhibition of activity. However, in contrast to rInsP₃R-1 at saturating [InsP₃], the activity of mInsP₃R-2 was unaffected by [ATP]. At lower [InsP₃], ATP had dramatic effects on mInsP₃R-2 P_o, but unlike the rInsP₃R-1, this did not occur by altering the relative Ca²⁺ dependency, but by simply increasing the maximally achievable P_o at a particular [InsP₃] and [Ca²⁺]. [InsP₃] did not alter the biphasic regulation of activity by Ca²⁺ in either rInsP₃R-1 or mInsP₃R-2. Analysis of the single channel kinetics indicated that Ca²⁺ and ATP modulate the P_o predominately by facilitating extended bursting activity of the channel but the underlying biophysical mechanism appears to be distinct for each receptor. Subtype-specific regulation of InsP₃R channel activity probably contributes to the fidelity of Ca²⁺ signalling in cells expressing these receptor subtypes.

Key points

• Three family members of inositol 1,4,5-trisphosphate receptors (InsP₃Rs) represent ubiquitously expressed intracellular Ca²⁺ release channels. The activity of the channels is regulated in a concerted and inter-related fashion by InsP₃, Ca²⁺ and ATP. It is not established whether each isoform is regulated by these ligands in an identical fashion. In this study we directly compare the single channel activity of mammalian InsP₃R-1 and InsP₃R-2 expressed in isolation in the presence of these ligands.

• An increase in activity for each isoform was mediated by a transition from a quiescent, ‘parked’ state to a ‘drive’ mode characterized by bursting activity. Ligands did not affect the single channel activity during these bursts but instead modulate the extent of bursting activity.

• Kinetic analysis revealed that the regulation of the transition to bursting activity by Ca²⁺ and ATP occurred by different mechanisms in InsP₃R-1 vs. InsP₃R-2: although the activity of both channels was biphasically regulated by Ca²⁺ and changes in [InsP₃] did not alter this relationship, elevated ATP increased the Ca²⁺ sensitivity of InsP₃R-1 activity without increasing the maximal achievable open probability (P_o) of the channel. In contrast, ATP simply increased the maximal achievable P_o without altering Ca²⁺ sensitivity of InsP₃R-2.

• The differing modes of regulation of InsP₃R-1 and InsP₃R-2 probably markedly influence the characteristics of intracellular Ca²⁺ signals observed in cells in which these isoforms are expressed.

Introduction

An elevation of intracellular Ca²⁺ levels following opening of inositol 1,4,5-trisphosphate receptors represents a ubiquitous signalling system engaged by a diverse array of hormones, neurotransmitters and growth factors (Berridge, 1993; Berridge et al. 2003). These Ca²⁺ signals ultimately control an expansive repertoire of physiological processes. The InsP₃R are a family of large (∼300 Da) tetrameric cation-selective channels predominately localized to the endoplasmic reticulum (Patel et al. 1999; Patterson et al. 2004; Yule et al. 2010). Three genes encode the InsP₃R family proteins, named InsP₃R-1, InsP₃R-2 and InsP₃R-3, with additional diversity at the protein level occurring as a result of alternate splicing of both the InsP₃R-1 and InsP₃R-2 genes (Furuichi et al. 1989; Danoff et al. 1991; Sudhof et al. 1991; Blondel et al. 1993). The general domain structure of InsP₃R has been established. Two major regions of high sequence homology exist between the subtypes; first, the binding site for InsP₃ in the N-terminus (Mignery & Sudhof, 1990; Yoshikawa et al. 1996) and second, a region in the C-terminus which constitutes the ion-conducting pore and determinants of tetramer formation (Michikawa et al. 1994; Boehning et al. 2001). The third domain consists of the intervening ∼1700 amino acids between the InsP₃ binding core and ion conducting pore and exhibits more sequence variability between subtypes. This region is named the regulatory and coupling domain, and modulation of release by factors such as Ca²⁺, ATP, phosphorylation and by some protein binding partners is thought to occur here (Patel et al. 1999; Patterson et al. 2004; Yule et al. 2010). Numerous studies have demonstrated that interaction of these factors with InsP₃R markedly influences release, and this often occurs in an InsP₃R subtype-specific manner (Patel et al. 1999; Patterson et al. 2004; Yule et al. 2010).

The fundamental properties of InsP₃R have largely been established based on the template of the InsP₃R-1. Substantial mechanistic information has been obtained by using single channel measurements in the so-called ‘on-nucleus’ configuration of the patch clamp technique (Stehno-Bittel et al. 1995; Mak et al. 1999, 2000, 2001a,b, 2003; Foskett et al. 2007). Primarily these studies have been performed using nuclei prepared from Xenopus laevis oocytes. This system was originally thought to express only a single InsP₃R isoform which was termed the ‘xInsP₃R’. The xInsP₃R exhibits a primary structure most closely resembling mammalian InsP₃R-1 (Parys et al. 1992). Like mammals, the Xenopus genome contains three genes encoding InsP₃R and thus it is likely that xInsP₃R represents the Xenopus InsP₃R-1 ortholog. Notably, a recent study has indicated that mRNA representing the xInsP₃R-2 and xInsP₃R-3 message can also be detected in Xenopus oocytes (Zhang et al. 2007). A series of studies have investigated how the principle ligands, InsP₃, Ca²⁺ and ATP, regulate xInsP₃R activity (Mak et al. 1999, 2001b; Foskett et al. 2007). These data have confirmed that xInsP₃Rs are regulated in a biphasic manner by cytosolic Ca²⁺ (Parys et al. 1992) and have thus further validated earlier Ca²⁺ flux and bilayer studies performed in mammalian systems (Iino, 1990; Bezprozvanny et al. 1991; Finch et al. 1991). In addition, these data suggest that InsP₃ binding primarily regulates channel activity by tuning the degree of Ca²⁺ inhibition of the channel (Mak et al. 1998, 2001b). In turn, the regulation of xInsP₃R P_o by Ca²⁺ itself is modulated by cytosolic ATP, such that increasing ATP shifts the activation of Ca²⁺ release to lower cytosolic Ca²⁺ concentrations (Mak et al. 1999, 2001a). Further studies conducted in oocytes expressing recombinant mammalian InsP₃R-3 indicated a similar mechanism could account for the regulation of InsP₃R-3 activity by ATP and Ca²⁺, albeit with subtle subtype-specific properties.

Although a consensus in the published literature suggests that all InsP₃Rs are modulated by Ca²⁺ and ATP, significant, subtype-specific differences have been reported. Indeed, initial reports suggested that InsP₃R-2 (Perez et al. 1997; Ramos-Franco et al. 2000) and InsP₃R-3 (Hagar et al. 1998) activity reconstituted in bilayers was only stimulated by increasing Ca²⁺ and not subject to inhibition at higher concentrations. Subsequent Ca²⁺ flux experiments have, however, reported biphasic regulation by Ca²⁺ attributed to expression of both InsP₃R-2 and InsP₃R-3 (Marchant et al. 1997; Boehning & Joseph, 2000; Mak et al. 2001a; Tu et al. 2005a). Modulation by cytosolic ATP also occurs in a subtype-specific manner (Tu et al. 2005b) and appears to determine the difference in Ca²⁺ sensitivity reported for xInsP₃R and mammalian InsP₃R-3 (Mak et al. 2001a). Our laboratory has also documented differences in the ATP sensitivity, the range of [InsP₃] subject to this regulation and the structural motif in the channel required for regulation of InsP₃R family members (Betzenhauser et al. 2008, 2009; Yule et al. 2010). Specifically, we have reported that InsP₃R-2 exhibits ∼3-fold greater ATP sensitivity than InsP₃R-1, is regulated only at sub-maximal [InsP₃] and, in marked contrast to other family members, this modulation is mediated by binding to a Walker A type glycine-rich nucleotide binding site (Yule et al. 2010).

While comparisons of the properties of InsP₃R isoforms have been made using imaging experiments and receptors reconstituted into bilayers (Miyakawa et al. 1999; Tu et al. 2003, 2005b), establishing a general model for InsP₃R channel gating is complicated by the relative lack of electrophysiological data documenting the fundamental biophysical properties of mammalian InsP₃R family members in native membranes. This particularly applies to InsP₃R-2, where there is only modest published patch clamp data of the channel in an intracellular membrane (Li et al. 2007; Tovey et al. 2010) and no data regarding how Ca²⁺ and ATP regulate channel gating. Given these observations, it is difficult to predict if the current model proposed for the regulation of InsP₃R-1 by InsP₃, ATP and Ca²⁺, largely based on data from xInsP₃R or insect InsP₃R in SF9 cells (Ionescu et al. 2007), can generally account for the regulation of gating of all InsP₃R family members.

In this study, we used ‘on-nucleus’ patch clamp to investigate the single channel properties and mode of regulation of unambiguously homotetrameric mInsP₃R-2 and rInsP₃R-1 following stable expression in a null background. This report provides the first comprehensive assessment of InsP₃R-2 activity following regulation by InsP₃, Ca²⁺ and ATP, and shows that under identical conditions, that like rInsP₃R-1, the P_o of mInsP₃R-2 is biphasically regulated by Ca²⁺ at various InsP₃ concentrations. However, in contrast to rInsP₃R-1, the relative Ca²⁺ sensitivity of mInsP₃R-2 is not altered by ATP, which functions to simply increase the maximally achievable P_o at a particular [InsP₃] and [Ca²⁺]. Our data also suggest that modulation of InsP₃R-1/2 activity by InsP₃ does not occur by modulation of the inhibitory effect of Ca²⁺ in this system. Further, analysis of the single channel kinetics indicates that Ca²⁺ and ATP modulate the P_o predominately by facilitating extended ‘bursting’ activity of the channel as recently reported for insect InsP₃Rs (Ionescu et al. 2007). Notably, these ligands do not significantly alter channel activity during this mode of activity. These data support the idea that InsP₃R essentially exists in two states; a long-lived closed state where the channel is essentially ‘parked’ with only very rare visits to an open state and that ligands facilitate the transition from the ‘parked’ state into a ‘drive’ mode represented by periods of bursting activity.

Methods

Creation of stable InsP₃R-1- and -2 -expressing DT40-3KO cell lines

cDNA encoding rInsP₃R-1 and mInsP₃R-2 were linearized with Nru I and introduced into DT40-3KO cells by nucleofection using solution T and program B23 as per the manufacturer's instructions (Lonza, Basel, Switzerland) as previously described (Betzenhauser et al. 2008). The cells were incubated in growth medium for 24 h prior to exposure to selection medium containing 2 mg ml⁻¹ Geneticin (Invitrogen, Carlsbad, CA, USA). Cells were then seeded into 96-well tissue culture plates at ∼1000 cells per well and incubated in selection medium for at least 7 days. Wells exhibiting growth after the selection period were picked for expansion. Several clones of DT40 cells were generated expressing both rInsP₃R-1 and mInsP₃R-2. Expression levels were documented by Western blotting with specific antisera (Betzenhauser et al. 2008, 2009) and functionality was confirmed by the ability to elicit Ca²⁺ signals following exposure to protease-activated receptor agonists. Single channel activity was indistinguishable in nuclei prepared from the various InsP₃R-1 and InsP₃R-2 clones.

Preparation of DT40 cell nuclei

Isolated DT40 nuclei were prepared by homogenization as previously described (Betzenhauser et al. 2009). The homogenization buffer (HB) contained 250 mm sucrose, 150 mm KCl, 3 μm 2-mercaptoethanol BME, 10 mm Tris, 1 mm Phenylmethanesulphonylfluoride PMSF, pH 7.5 with a complete protease inhibitor tablet (Roche, USA). Cells were washed and resuspended in HB prior to nuclear isolation using a RZR 2021 homogenizer (Heidolph Instruments, Germany) with 25 strokes at 1200 rpm. A 3 μl aliquot of nuclear suspension was placed in 3 ml of bath solution (BS) which contained 140 mm KCl, 10 mm Hepes, 500 μm BAPTA and 246 nm free Ca²⁺, pH 7.1. Nuclei were allowed to adhere to a plastic culture dish for 10 min prior to patching.

On-nuclei patch clamp experiments

Single InsP₃R channel potassium currents (i_k) were measured in the on-nucleus patch clamp configuration using pCLAMP 9 and an Axopatch 200B amplifier (Molecular Devices, Sunnydale, CA, USA) as previously described (Betzenhauser et al. 2009). Pipette solution contained 140 mm KCl, 10 mm Hepes, with varying concentrations of InsP₃, ATP, BAPTA and free Ca²⁺. Free [Ca²⁺] was calculated using Max Chelator freeware and verified fluorometrically. Traces were consecutive 3 s sweeps recorded at −100 mV, sampled at 20 kHz and filtered at 5 kHz. A minimum of 15 s of recordings were considered for data analyses. The data are representative of between 6 and 12 experiments for each condition presented. Pipette resistances were typically 20 MΩ and seal resistances were >5 GΩ.

Data analysis

Single channel openings were detected by half-threshold crossing criteria using the event detection protocol in Clampfit 9. We assumed that the number of channels in any particular nuclear patch is represented by the maximum number of discrete stacked events observed during the experiment. Even at low P_o, stacking events were evident (data not shown). Only patches with one apparent channel were considered for analyses. Probability of opening (P_o), unitary current (i_k), open and closed times, and burst analyses were calculated using Clampfit 9 and Origin 6 software (Origin Lab, Northampton, MA, USA). All-points current amplitude histograms were generated from the current records and fitted with a normal Gaussian probability distribution function. The coefficient of determination (R²) for every fit was >0.95. The P_o was calculated using the multimodal distribution for the open and closed current levels. Channel dwell time constants for the open and closed states were determined from exponential fits of all-points histograms of open and closed times. The threshold for an open event was set at 50% of the maximum open current and events shorter than 0.1 ms were ignored. A ‘burst’ was defined as a period of channel opening following a period of no channel activity which was greater than five times the mean closed time (0.2 ms) within a burst. No closed times shorter than 1 ms were therefore included in the burst analysis. The majority of the data presented were recorded and analysed at a holding potential of −100 mV in order to maximize the magnitude of the current. Similar analysis was performed on both rInsP₃R-1 and mInsP₃R-2 with a pipette solution containing 10 μm InsP₃, 5 mm ATP and 200 nm Ca²⁺ at +60 mV. Mean open (0.35 ± 0.001 ms) and closed times (0.2 ± 0.001 ms) within bursts, time constants for burst lengths (89.2 ± 10.6 and 85.3 ± 3.4 ms; InsP₃R-1 and InsP₃R-2, respectively) and intervals (0.9 ± 0.1 and 2.1 ± 0.3 ms; InsP₃R-1 and InsP₃R-2, respectively, together with P_o (0.7 ± 0.1) were essentially identical to those obtained at −100 mV.

The slope conductances were determined from the linear fits of the current–voltage relationships with the equation:

where g is unitary conductance, i_k is unitary current, V is voltage and V_k is the reversal potential. Ca²⁺ dependency curves were fitted separately for activation and inhibition with the logistic equation:

where A₁ and A₂ are asymptotes, X is the concentration of Ca²⁺, X₀ is the half-maximal concentration and p is the slope related to the Hill coefficient. Equation parameters were estimated using a non-linear, least-squares algorithm. Two-tailed heteroscedastic t tests with P values < 0.05 were considered to have statistical significance.

Results

InsP₃-evoked single channel currents from nuclei of DT40 cells stably expressing InsP₃R displayed bursts of channel activity

Since only very limited data are available which allows the direct comparison of mammalian InsP₃R-1 and InsP₃R-2 under identical conditions, we utilized the nuclear membrane patch clamp technique (Mak & Foskett, 1994; Stehno-Bittel et al. 1995; Foskett et al. 2007) to directly evaluate the single channel conductance and effects of and ATP on gating of rInsP₃R-1 and mInsP₃R-2. In this configuration, high-resistance seals were made between the pipette and the outer nuclear membrane. Nuclei were isolated from DT40 cell lines stably expressing InsP₃R as described in the Experimental procedures section. InsP₃R single channel openings were readily detected in these preparations in the presence of concentrations of InsP₃ (10 μm), ATP (5 mm) and Ca²⁺ (200 nm). These conditions were found to be optimum for channel activity and increasing the [InsP₃] (to 100 μm) or [ATP] further did not augment P_o. Figure 1 shows representative traces of channel openings from rInsP₃R-1 (Fig. 1A) and mInsP₃R-2 (Fig. 1B) at various pipette holding potentials. Figure 1C shows the current–voltage relationship derived from these records. The slope conductance was essentially identical for each InsP₃R (373 ± 2 pS and 375 ± 5 pS for rInsP₃R-1 and mInsP₃R-2, respectively) and similar to that previously reported for the InsP₃R-1 in this environment and in nuclei isolated from Purkinje neurons (Marchenko et al. 2005; Foskett et al. 2007). The activity recorded for each channel was characterized by extended periods in which the channel transitioned rapidly between the open and closed state. These ‘bursts’ of activity were separated by long-lived interburst closed times. Single channel openings (i.e not in bursts) were seldom observed. Channel openings never occurred in the presence of the InsP₃R antagonist heparin (100 μg ml⁻¹, dialysed for 30 s prior to recording), when InsP₃ was excluded from the pipette solution, or in nuclei prepared from DT40-3KO cells where InsP₃Rs are absent (Betzenhauser et al. 2009). In total, these data indicate that the recordings result from InsP₃R activity.

Figure 1. Single Channel recordings of rInsP₃R-1 and mInsP₃R-2.

A, representative single channel records of rInsP₃R-1 activity upon stimulation by 10 μm InsP₃ and 5 mm ATP and 200 nm Ca²⁺ recorded at the indicated pipette holding potential. B, representative single channel records of mInsP₃R-2 activity under identical conditions as in A. C, shows the current vs. voltage relationship for both channels and yields essentially identical slope conductances for rInsP₃R-1 (373 ± 2 pS) and mInsP₃R-2 (375 ± 5 pS).

Ca²⁺ and ATP dependence of rInsP₃R-1 activity

A characteristic property of InsP₃R is that channel openings are biphasically dependent on cytoplasmic [Ca²⁺]; increasing [Ca²⁺] from basal cytosolic values initially augments P_o while higher [Ca²⁺] attenuates channel activity. Mak et al. further demonstrated that ATP shifts the biphasic Ca²⁺ dependency of xInsP₃R, such that both activation and inhibition occurred at lower [Ca²⁺] (Mak et al. 1998, 1999, 2001a). Experiments were next performed to determine if ATP has the same effect on mammalian InsP₃R-1 expressed in the nuclei of stably transfected DT40 cells. Single channel data were initially collected with a pipette solution containing 10 μm InsP₃ (‘high’ InsP₃), various [Ca²⁺] values and either low (100 μm) or saturating (5 mm) ATP (Betzenhauser et al. 2009). The recordings in Fig. 2Aa demonstrate that in the presence of 100 μm ATP channel activity was augmented as the pipette [Ca²⁺] increased from 200 nm and achieved a maximum P_o of ∼0.71 ± 0.1 in the presence of 1 μm Ca²⁺. Activity subsequently decreased at higher [Ca²⁺]. In the presence of 5 mm ATP (Fig. 2Ab), channel activity also increased as the [Ca²⁺] was raised, but under these conditions a similar maximum P_o (0.65 ± 0.05) was achieved at lower [Ca²⁺] (∼200 nm Ca²⁺). Channel activity subsequently decreased at Ca²⁺ levels above 1 μm. Figure 2B shows pooled data from these experiments where P_o was plotted against [Ca²⁺]_i for rInsP₃R-1 in the presence of 10 μm InsP₃. Fits of these data show that in the presence of 5 mm ATP, the EC₅₀ for Ca²⁺ activation was 57.2 ± 6 nm (P= 3.1 ± 0.5) and the IC₅₀ for inhibition by Ca²⁺ was 4.0 ± 0.7 μm (P= 3.3 ± 0.7). The steepness of the slope is suggestive of a cooperative process and results in a small change in [Ca²⁺] evoking a relatively large increase in P_o. Decreasing [ATP] shifted the dependence to higher [Ca²⁺]_i for both the activation to 347.8 ± 45.7 nm and inhibition to 24.9 ± 1.9 μm without statistically changing the slope for activation or inhibition (P= 3.7 ± 0.7 and 3.4 ± 0.5, activation and inhibition respectively).

Figure 2. Effect of Ca²⁺ and ATP on rInsP₃R-1 single channel activity.

A, shows representative single channel recordings evoked by ‘high’ InsP₃ (10 μm) at various [Ca²⁺] with either 100 μm ATP (Aa) or 5 mm ATP (Ab). In both cases channel activity initially increased to reach a maximal P_o and then subsequently decreased, although higher [ATP] shifts the activation and inhibition to lower Ca²⁺ concentrations. B, shows pooled data plotted as P_o vs.[Ca²⁺] for rInsP₃R-1 with 10 μm InsP₃ and either 5 mm (filled black squares) or 100 μm ATP (open black squares) and fit as described in the Methods. C, shows representative single channel recordings evoked by ‘low’ InsP₃ (1 μm) at various [Ca²⁺] with either 100 μm ATP (Ca) or 5 mm ATP (Cb). The maximal achievable P_o at each [Ca²⁺] was lower, however increasing the [Ca²⁺] had a similar biphasic effect on P_o which was shifted by increasing [ATP]. D, shows pooled data plotted as P_o vs.[Ca²⁺] for rInsP₃R-1 with 1 μm InsP₃ and either 5 mm (filled black squares) or 100 μm ATP (open black squares) and fit as described in the Methods. E, shows a single sweep on an expanded scale for channel activity evoked by 1 μm InsP₃, 5 mm ATP at either 50 nm (left) or 200 nm Ca²⁺ (right).

At 1 μm InsP₃ (‘low’ InsP₃) a biphasic dependency on [Ca²⁺] was also readily demonstrated as shown in Fig. 2C and pooled data in Fig. 2D. However, the maximal achievable P_o was decreased to 0.27 in the presence of 5 mm ATP. The EC₅₀ value of 65.6 ± 8 nm and an IC₅₀ of 3.6 ± 1.8 μm were not statistically significantly different from data obtained at high InsP₃ and 5 mm ATP. The dependency was also shifted at lower [ATP] in a quantitatively similar manner to that seen with high InsP₃ (EC₅₀= 333.5 ± 31.2 nm, IC₅₀= 24.2 ± 2.1 μm), again with no change in the slope. These data indicate that ATP tunes the Ca²⁺ sensitivity of mammalian InsP₃R-1 in a qualitatively similar fashion to that reported for the xInsP₃R (Mak et al. 1998).

Although increasing the [InsP₃] from 1 to 10 μm approximately doubled the P_o, no alteration in the degree of Ca²⁺ inhibition was observed (IC₅₀∼24 μm; 5 mm ATP for both 1 and 10 μm InsP₃). These data suggest in contrast to xInsP₃R, that InsP₃ does not gate mammalian InsP₃R expressed in DT40 cells by tuning Ca²⁺ inhibition. A similar conclusion was reported previously for endogenous InsP₃R studied in isolated Purkinje cell neurons (Marchenko et al. 2005). Figure 2E shows a representative example of the effect of increasing Ca²⁺ on rInsP₃R-1 channel activity at low InsP₃ and illustrates the increase in the period of bursting activity.

Ca²⁺ and ATP dependence of mInsP₃R-2 activity

The mInsP₃R-2 displays a greater sensitivity to ATP than rInsP₃R-1 in Ca²⁺ release assays (Betzenhauser et al. 2008, 2009). In the context of InsP₃R-2, 100 μm ATP exerts a maximal effect and, consistent with effects on Ca²⁺ release, was shown to similarly maximally augment P_o in this study (triangles in Fig. 3D). Therefore, the [ATP] chosen to represent low ATP in experiments studying mInsP₃R-2 was reduced to 10 μm. The representative recordings in Fig. 3A obtained at high [InsP₃] (10 μm) show that in a similar fashion to rInsP₃R-1, mInsP₃R-2 activity was initially enhanced by increasing [Ca²⁺] to reach a maximal P_o of 0.7 ± 0.1 at ∼200 nm Ca²⁺ and then activity subsequently decreased at [Ca²⁺] greater than 1 μm. Fits of the pooled data (Fig. 3B) gave an EC₅₀ for Ca²⁺ activation of 58.3 ± 4 nm and IC₅₀ for inhibition of 3.7 ± 0.3 μm. These values were indistinguishable from those obtained for rInsP₃R-1 in this study (see Fig. 2B). Consistent with previous reports, at high [InsP₃] the biphasic regulation of mInsP₃R-2 activity was not altered by increasing [ATP] from 10 μm to 5 mm (compare Fig. 3Aa and Fig. 3Ab and pooled data in Fig. 3B).

Figure 3. Effect of Ca²⁺ and ATP on mInsP₃R-2 single channel activity.

A, shows representative single channel recordings evoked by ‘high’ InsP₃ (10 μm) at various [Ca²⁺] with either 10 μm ATP (Aa) or 5 mm ATP (Ab). In a similar fashion to rInsP₃R-1, the single channel activity of mInsP₃R-2 was biphasically modulated by [Ca²⁺], however [ATP] had no effect on this relationship at this [InsP₃]. B, shows pooled data and fits of the data, plotted as P_o vs.[Ca²⁺] for mInsP₃R-2 with 1 μm InsP₃ and either 5 mm (filled black squares) or 10 μm ATP (open black squares) C, shows representative single channel recordings evoked by ‘low’ InsP₃ (1 μm) at various [Ca²⁺] with either 10 μm ATP (Ca) or 5 mm ATP (Cb). The maximal achievable P_o at each [Ca²⁺] was lower, however again increasing the [Ca²⁺] had a similar biphasic effect on P_o. In contrast to rInsP₃R-1, increasing ATP did not shift the Ca²⁺ dependency but simply dramatically increased the maximally achievable P_o at each [Ca²⁺]. D, shows pooled data and fits of the data, plotted as P_o vs.[Ca²⁺] for mInsP₃R-2 with 1 μm InsP₃ and either 5 mm (filled black squares) or 10 μm ATP (open black squares). Open triangles represent data obtained with 100 μm ATP. Open circles are data obtained from a mutant InsP3R which is deficient in ATP binding. E, shows a single sweep on an expanded scale for channel activity evoked by 1 μm InsP₃, 5 mm ATP at either 50 nm (left) or 200 nm Ca²⁺ (right).

Biphasic regulation of mInsP₃R-2 channel activity by Ca²⁺ was also seen in cells exposed to low [InsP₃] (Fig. 3C and D); similar to high [InsP₃], peak activity occurred ∼200 nm Ca²⁺ and subsequently decreased at greater than 1 μm Ca²⁺ (Fig. 3C and D). Under these conditions, changing the [ATP] from 10 μm to 5 mm had striking effects on the P_o without altering the relative Ca²⁺ sensitivity of the mInsP₃R-2 activity. Specifically, the EC₅₀ for activation (68.5 ± 8 nm) and IC₅₀ for inhibition (3.4 ± 0.7 μm) for 10 μm ATP were not statistically different to 5 mm ATP or indeed to higher [InsP₃]; however, the maximally achievable P_o was reduced to 0.05 ± 0.03 from 0.51 ± 0.03; 10 μm vs. 5 mm ATP, respectively (compare Fig. 3Ca and Fig. 3Cb and pooled data in Fig. 3D). The effect of ATP to increase P_o at activating [Ca²⁺] was absent in cells expressing a mutant InsP₃R-2 with no functional ATP binding site, the so called “ATPB” glycine rich binding motif (Betzenhauser et al. 2008) (Fig. 3D). Again, although dramatic effects on mInsP₃R-2 P_o were observed by altering [InsP₃], this did not appear to occur as a function of changing the degree of Ca²⁺ inhibition. Fig. 3E shows a representative example of the effect of increasing Ca²⁺ on mInsP₃R-2 channel activity at low InsP₃ on an expanded scale and illustrates the increase in bursting activity manifested as an increase in the number of bursts of relatively constant duration. In total, these data indicate that while rInsP₃R-1 and mInsP₃R-2 activity are both biphasically regulated by cytosolic Ca²⁺, that regulation of channel opening by ATP for the individual receptors occurs by fundamentally different mechanisms. A summary of the effects of ATP and Ca²⁺ regulation and maximal P_o on rInsP₃R-1 and mInsP₃R-2 are shown in Fig. 4.

Figure 4. Summary of the major differences observed.

In all panels low ATP = 100 μm or 10 μm for rInsP₃R-1 and mInsP₃R-2, respectively, High ATP is 5 mm in both cases. In A, a histogram shows the EC₅₀ for Ca²⁺ activation of channel activity for the conditions indicated. ATP shifts the EC₅₀ only for rInsP₃R-1. In B, a histogram shows the IC₅₀ for Ca²⁺ inhibition of channel activity for the conditions indicated. ATP shifts the EC₅₀ only for rInsP₃R-1. In C, a histogram depicts the maximal achievable P_o for each receptor when exposed to optimum [Ca²⁺] (rInsP₃R-1; 1 μm in low ATP, 200 nm Ca²⁺ in high ATP; mInsP₃R-2 200 nm Ca²⁺) at the indicated [ATP]. The maximally achievable P_o is ∼2 fold greater for mInsP₃R-2 under optimum conditions (compare high ATP bars). In addition ATP markedly increases the P_o of the mInsP₃R-2.

Analysis of the effects of InsP₃, Ca²⁺ and ATP on InsP₃R single channel kinetics

In order to gain insight into mechanisms underlying the differences in regulation of rInsP₃R-1 and mInsP₃R-2, an analysis of the steady-state open and closed dwell times was performed during periods of bursting activity. Histograms were generated for all InsP₃, ATP and concentrations described in Figs 2 and 3, for both rInsP₃R-1 and mInsP₃R-2. Each condition was well fitted with single exponential functions. Remarkably, the time constants for the open times (τ_open) and the closed times (τ_closed) within the bursts were essentially identical at each InsP₃, Ca²⁺ or ATP concentration tested (τ_open, 0.3 ± 0.01 ms and τ_closed, 0.2 ± 0.02 ms). Examples of these histograms for low and high activity for both mInsP₃R-2 and rInsP₃R-1 are shown in Supplemental Figs S1 and S2. These examples clearly show a biexponential distribution of closed times but only a single population of open times (mean open time 0.35 ± 0.001 ms; mean closed₁ time, 0.2 ± 0.001 ms). The increased channel activity is associated with a decrease in the time constant for the longer duration closed time (closed₂, representing the interburst interval) with no associated change in the open time constant. This analysis clearly demonstrates that during ‘bursting’, InsP₃R activity is practically insensitive to these ligands. The relative lack of variation in the intraburst kinetic parameters is presumably the result of analysis of an extremely large number of stochastic channel openings and closings (10⁵–10⁶ events) for each experimental condition. In total, these data lead to the conclusion that increasing InsP₃R P_o is probably determined by ligand-dependent facilitation of the bursting or ‘drive’ mode of the channel.

Ligands increase rInsP₃R-1 P_o by altering the burst lengths and the interburst intervals

Because open and closed times during bursting activity were unaffected by InsP₃, Ca²⁺ or ATP, we analysed the influence of these ligands on the length of time the channel spends in the bursting mode and the quiescent periods between bursts. Time constants describing the bursts and interburst intervals were derived from the single exponential fits of the distribution of events under conditions detailed in Fig. 2. Figure 5 shows the time constants for experiments measuring rInsP₃R-1 activity plotted vs.[Ca²⁺]_i. In the presence of 5 mm ATP, as pipette [Ca²⁺] was elevated, the burst lengths increased and this was mirrored by a reciprocal decrease in interburst intervals (Fig. 5A and C). This trend was reversed at higher, inhibitory [Ca²⁺] and occurred in the presence of both 10 μm InsP₃ (Fig. 5A) or 1 μm InsP₃ (Fig. 5C). However, at the lower [InsP₃], the maximal burst lengths were shorter and the maximal interburst intervals were longer than seen with higher [InsP₃]. A decrease in [ATP] resulted in the Ca²⁺- induced lengthening of bursts and corresponding decrease in interburst intervals occurring at higher [Ca²⁺]_i consistent with the effect of ATP to shift the Ca²⁺ sensitivity (compare Fig. 5A and B with Fig. 5C and D). These data therefore also indicate that ATP binding facilitates an increase in P_o by augmenting bursting activity.

Figure 5. Analysis of rInsP₃R-1 burst kinetics.

Burst analyses were performed for rInsP₃R-1 as detailed in Methods. Histograms were constructed of burst lengths and interburst intervals for all conditions and were well fit with single exponential functions. The time constants for these fits were then plotted vs.[Ca²⁺]_i. In A, Burst length time constants (black) and interburst interval time constants (red) plotted for 10 μm InsP₃ and 5 mm ATP. Increasing Ca²⁺ initially decreases the intraburst time constant and increases the burst length time constant. This relationship is reversed at inhibitory [Ca²⁺]. In B, decreasing ATP to 100 μm results in this relationship shifting to higher [Ca²⁺]. Burst length time constants (black) and interburst interval time constants (red) are plotted for 1 μm InsP₃ and both 5 mm (C) and 100 μm ATP (D) and illustrate a similar biphasic [Ca²⁺] dependency which is shifted by ATP.

InsP₃ concentration determines the mechanism leading to bursting in mInsP₃R-2

A similar analysis was performed for channel activity recordings from mInsP₃R-2. In the presence of high [InsP₃], increasing pipette [Ca²⁺] resulted in a reciprocal increase in burst length duration and decrease in interburst intervals as P_o increased and subsequently reversed at inhibitory [Ca²⁺] (Fig. 6A and B). This occurred in essentially an identical manner to rInsP₃R-1 (Fig. 5) but was, as expected, unaffected by [ATP] (compare Fig. 6A and B). However, in marked contrast to rInsP₃R-1 (and mInsP₃R-2 at high [InsP₃]), at low [InsP₃], modulation by P_o as a function of [Ca²⁺] occurred solely by altering the interburst interval without affecting the mean burst length which averaged 6.3 ms in duration (Fig. 6C and D; compare Fig. 2E and Fig. 3E for examples on an expanded time scale) over a 3 order of magnitude change in [Ca²⁺]. An increase in InsP₃ binding therefore appears to switch the channel between these gating modes. In addition, ATP dramatically enhanced the shortening of the interburst interval at low InsP₃ to markedly increase the maximally achievable P_o at a particular [Ca²⁺] (compare Fig. 6C and D). In total, this analysis reveals that regulation of P_o by Ca²⁺ and ATP at low InsP₃ occurs through different mechanisms in rInsP₃R-1 vs. mInsP₃R-2.

Figure 6. Analysis of mInsP₃R-2 burst kinetics.

Histograms were constructed of burst lengths and interburst intervals for all conditions and were fitted with single exponential equations. The time constants for these fits were then plotted against [Ca²⁺]_i. Burst length time constants (black) and interburst interval time constants (red) plotted for 10 μm InsP₃ and both 5 mm (A) and 10 μm ATP (B). In an identical fashion to rInsP₃R-1, modulation of P_o by [Ca²⁺] is achieved by a reciprocal change in interburst interval and burst length. Burst length time constants (black) and interburst interval time constants (red) plotted for 1 μm InsP₃ and both 5 mm (C) and 10 μm ATP (D). At low [InsP₃], Ca²⁺ only regulated the interburst intervals without altering burst length. ATP dramatically increases P_o by further shortening the interburst intervals.

Discussion

Similarities in some properties of the rInsP₃R-1 and mInsP₃R-2

In this study we have directly compared the single channel properties and gating of mammalian InsP₃R-1 and InsP₃R-2 expressed in the outer leaflet of the nuclear membrane of DT-40-3KO cells under identical conditions. We report that some fundamental properties of the individual channels are identical, but also that mechanisms governing the gating of the two channels by important modulators are distinctly different. For example, the single channel conductance of ∼375 pS for each channel is indistinguishable with K⁺ as the charge carrying ion under symmetrical ionic conditions. While these data represent one of the few reports where a direct comparison of conductance is possible, given that the putative pore regions in each InsP₃R are identical, it is not unexpected and is consistent with an earlier bilayer study (Tu et al. 2005b).

The single channel activity of the mInsP₃R-2, like rInsP₃R-1, was also biphasically dependent on intracellular [Ca²⁺]. Biphasic regulation of InsP₃R by Ca²⁺ plays a primary role in controlling InsP₃R channel activity and has also been proposed to contribute to the great spatial and temporal diversity of cellular Ca²⁺ signals observed in cells. At saturating [InsP₃] and [ATP] the activity of rInsP₃R-1 and mInsP₃R-2 was essentially identical in terms of maximal P_o and the [Ca²⁺] required for activation and inhibition. Ca²⁺ activation of channel activity occurred at a [Ca²⁺] close to that reported under resting conditions (50–200 nm). This property might be predicted to result in rapid Ca²⁺ release on InsP₃ binding and would be ideally suited to mediate the initiation of signals at specialized ‘trigger zones’ (Kasai et al. 1993; Nathanson et al. 1994; Yule et al. 1997). At [Ca²⁺] above 1 μm, InsP₃R activity decreased, resulting in a narrow window in which maximal activity was observed. Physiologically, a reduction in channel activity as a consequence of initial release might contribute to spatial and temporally limiting Ca²⁺ release events and contribute to the generation of localized signals and Ca²⁺ oscillations (Bootman et al. 1997; Marchant et al. 1999; Berridge et al. 2000).

Notably, however, while increasing InsP₃ had marked effects on P_o for both InsP₃R-1 and -2, these changes were not accomplished by altering the biphasic modulation of channel activity by Ca²⁺. These data are therefore not consistent with a generalized model whereby InsP₃ binding primarily gates InsP₃R activity by modulating the degree of Ca²⁺ inhibition (Mak et al. 1998, 1999; Foskett et al. 2007) and would suggest that Ca²⁺ binding simply allosterically modulates activity to a given extent at a particular [InsP₃]. A reduction in the extent of Ca²⁺ inhibition as [InsP₃] increases has been noted in a number of systems and investigated extensively for InsP₃R expressed in Xenopus oocyte nuclei (Kaftan et al. 1997; Mak et al. 1998, 2001b; Ionescu et al. 2006). Nevertheless, other studies, similar to our current data, have suggested that this may not be a universal phenomenon (Tang et al. 2003; Marchenko et al. 2005). For example, biphasic regulation by Ca²⁺ of InsP₃R channel activity (presumably InsP₃R-1) recorded from nuclei isolated from rat Purkinje cells was unaffected by changes in [InsP₃] from threshold to saturating which resulted in a >10-fold increase in channel activity (Marchenko et al. 2005). We presently do not have a satisfactory explanation for these differences; however, InsP₃Rs expressed in Xenopus oocytes exhibit some characteristics that appear not to be universal. For example, xInsP₃Rs display an exquisite sensitivity to InsP₃ and an extremely narrow concentration range where InsP₃ influences activity (Foskett et al. 2007). This property differs from the current study of mammalian receptors expressed in DT40 cells as well as other studies of endogenous InsP₃R activity including in Purkinje neurons and SF-9 cells which are markedly less sensitive to InsP₃ (Marchenko et al. 2005; Ionescu et al. 2006). It is therefore possible that species differences in InsP₃R primary structure or the complement of accessory factors present in each experimental system may contribute to these discrepancies.

Subtype-specific properties of rInsP₃R-1 and mInsP₃R-2

ATP modulation is thought to allosterically regulate InsP₃R to facilitate opening of all InsP₃R family members without altering the affinity for InsP₃ binding per se (Ferris et al. 1990; Foskett et al. 2007; Yule et al. 2010). This form of regulation of channel activity is thought to be important both under physiological conditions, potentially coupling the metabolic status of the cell to the degree of Ca²⁺ release, and under pathological conditions where the levels of ATP may fall precipitously. Our studies clearly demonstrate that cellular ATP levels have profound effects on the gating of both rInsP₃R-1 and mInsP₃R-2. However, clear distinctions exist that provide distinguishing features for each subtype. First, while rInsP₃R-1 is regulated by ATP at all [InsP₃], the activity of mInsP₃R-2 is unaffected by ATP when the channel is exposed to saturating InsP₃ levels. This is in agreement with previous studies which reported that both the single channel activity of plasma membrane mInsP₃R-2, or rInsP₃R-2 reconstituted into bilayers, together with InsP₃-induced Ca²⁺ flux in permeabilized cells was only regulated by ATP at sub-maximal [InsP₃] (Tu et al. 2005b; Betzenhauser et al. 2008; Park et al. 2008; Yule et al. 2010). Since evidence suggests that when InsP₃R-2 is expressed its characteristics are dominant (Miyakawa et al. 1999; Park et al. 2008), this property may allow Ca²⁺ release in cells which express InsP₃R-2 to be effectively insensitive to metabolic status when experiencing intense stimulation. These differences in susceptibility to ATP regulation are consistent with the effects of ATP being mediated through distinct binding sites in each receptor subtype. Indeed, mutagenesis studies indicate that ATP binding occurs at the so-called ‘ATPB’ site, a glycine-rich motif between amino acid residues 1969–1974 in InsP₃R-2, but at a distinct, and as yet undetermined site, in rInsP₃R-1 (Betzenhauser et al. 2008, 2009). The lack of effect of ATP under these conditions may be explained by InsP₃R-2 when fully saturated with InsP₃ favouring a conformation that occludes ATP binding to the ATPB nucleotide binding site. Alternatively, binding of ATP to this conformation may not be communicated to gating of the pore.

In contrast to exposure to saturating [InsP₃], the activity of rInsP₃R-1 and mInsP₃R-2 was profoundly different when challenged with lower [InsP₃]. Most obviously, the maximal P_o at 1 μm InsP₃ and optimum Ca²⁺ and ATP was ∼2-fold higher in mInsP₃R-2 (0.27 ± 0.02 vs. 0.51 ± 0.03; rInsP₃R-1 and mInsP₃R-2, respectively) and is reflective of the well-documented rank order for InsP₃ efficacy of family members (Iwai et al. 2007). Notably, raising [ATP] had profound differences on the individual channel gating. In the case of rInsP₃R-1, raising the [ATP] effectively sensitized the channel to Ca²⁺ such that both Ca²⁺ activation and inhibition occur at lower [Ca²⁺] without altering the maximally achievable P_o at a given [InsP₃]. A quantitatively similar effect was observed at both the InsP₃ concentrations tested in this study, which occurred over a ∼2.5-fold change in maximal P_o. These data are entirely consistent with previous reports from Mak and colleagues documenting the effects of ATP on Ca²⁺ activation of xInsP₃R and InsP₃R-3 expressed in oocytes (Mak et al. 1998, 1999, 2001a,b). In marked contrast, at sub-saturating [InsP₃], ATP did not alter the EC₅₀ for Ca²⁺ activation or inhibition for mInsP₃R-2, but instead profoundly increased the maximal P_o at a particular [Ca²⁺]. For example, at optimal [Ca²⁺] of 200 nm, increasing the [ATP] from 10 μm to 5 mm resulted in a ∼10-fold increase in P_o from 0.05 ± 0.03 to 0.51 ± 0.03. In comparison, with identical [Ca²⁺] and [InsP₃], increasing ATP only resulted in a ∼2-fold increase in P_o of rInsP₃R-1. ATP therefore appears to impart high gain to Ca²⁺ release through InsP₃R-2 when exposed to low, presumably physiological [InsP₃]. A question remains as to why ATP appears to facilitate channel opening only at sub-saturating [InsP₃]. While we have no definitive evidence to support this speculation, it is possible that differential effects arise because binding of ATP only occurs to individual monomers of the InsP₃R-2 that are not concurrently bound to InsP₃.

Kinetics of rInsP₃R-1 and mInsP₃R-2 channel gating

The kinetics of rInsP₃R-1 and mInsP₃R-2 channel gating was monitored over a large range of [Ca²⁺] together with InsP₃ and ATP levels representing saturating and sub-maximal ligand concentrations. Increasing channel activity for both rInsP₃R-1 and mInsP₃R-2 was characterized by progressively longer periods of repetitive activity marked by rapid channel openings (mean open time, 0.35 ± 0.01 ms) and closings (mean closed time, 0.17 ± 0.01 ms) occurring in bursts. This mode of gating has been observed in a number of studies and modulation of the bursting period is correlated with conditions that increase the overall channel P_o (Ionescu et al. 2007; Wagner et al. 2008; Taufiq Ur et al. 2009). Variations in InsP₃, Ca²⁺ or ATP did not alter the time constant for channel opening or closing for either channel subtype within the burst. Notably, these data indicate that the channel is essentially independent of these ligands when in bursting or ‘drive’ mode and suggest that the channel is effectively impervious to its local environment when in this mode. One implication of these data is that this property would be predicted to amplify Ca²⁺ release events at sites of active bursting InsP₃R as these channels would not be susceptible to locally high, inhibitory Ca²⁺ levels. Our single channel data are consistent with Ca²⁺ flux experiments performed in permeabilized hepatocytes which demonstrated that, following InsP₃ binding, Ca²⁺ release was insensitive to Ca²⁺ inhibition (Adkins & Taylor, 1999). While most previous studies have not implicitly analysed activity solely within bursts, these observations are nevertheless consistent with prior reports indicating that the integrated (i.e. bursting and non-bursting activity) mean open time of rInsP₃R-1 is relatively unaffected by ligands (Mak et al. 1998, 2001b; Ionescu et al. 2006).

In contrast to the lack of effect on channel kinetics within bursts, ligands dramatically altered the extent of bursting activity of both rInsP₃R-1 and mInsP₃R-2. This mechanism for increasing overall P_o is reminiscent of the gating of other channels including CFTR (Vergani et al. 2003) and K_ATP channels (Proks et al. 2001). At saturating [InsP₃], increasing Ca²⁺ resulted in a reciprocal increase in burst length and decrease in interburst interval to result in the maximum P_o occurring at optimum [Ca²⁺]. These processes were reversed as the channel activity was attenuated at high [Ca²⁺]. Consistent with the body of data, ATP shifted the Ca²⁺ dependency only for the rInsP₃R-1. A comprehensive kinetic model of these data is currently being constructed; however, the activity can be minimally described by a single open and closed state during bursting activity and a longer-lived closed state which represents the ‘parked’ state corresponding to the interburst intervals. In essence, in the presence of saturating InsP₃, Ca²⁺ acts to enhance the period that the channel spends in drive mode by functioning to both facilitate the transition out of the long-lived closed state while concomitantly decreasing the likelihood that the channel re-enters the quiescent parked state. This scheme can similarly account for regulation of rInsP₃R-1 and mInsP₃R-2 at high [InsP₃] except that ATP shifts the Ca²⁺ dependency of the transitions only for rInsP₃R-1.

An essentially identical scheme can account for activity of rInsP₃R-1 at lower [InsP₃]; however, the gating of mInsP₃R-2 under these conditions is clearly different. At low [InsP₃], Ca²⁺ does not influence the length of time the channel remains in the bursting mode, and thus by inference the transition into drive mode, but simply decreases the likelihood that the channel remains in the parked state. In contrast to rInsP₃R-1, ATP then appears to dramatically increase P_o solely by markedly influencing the time the channel spends parked.

In summary, we have investigated the single channel properties of rInsP₃R-1 and mInsP₃R-2 under identical conditions. We demonstrate that although the channels share fundamental properties, marked differences underlie the gating mechanism for each individual receptor. These properties, including modulation by ATP and Ca²⁺, are likely to contribute in a dynamic fashion to the characteristics of Ca²⁺ signals observed in cells expressing these isoforms.

Glossary

InsP₃R

InsP₃ receptor

rInsP₃R-1

rat type-1 InsP₃ receptor

mInsP₃R-2

mouse type-2 InsP₃ receptor

xInsP₃R

Xenopus leavis oocyte InsP₃ receptor

P_o

open probability

Author contributions

The work was performed in Department of Pharmacology and Physiology at the University of Rochester. L.E.W. collected the data, analysed the data and generated figures. D.I.Y. was responsible for the conception, design and analysis of the experiments together with making figures and drafting the manuscript. Both authors approved the final version.

Supplementary material

Supplemental Figure S1

Supplemental Figure S2