Interactions of antagonists with subtypes of inositol 1,4,5-trisphosphate (IP₃) receptor

Abstract

BACKGROUND AND PURPOSE

Inositol 1,4,5-trisphosphate receptors (IP₃Rs) are intracellular Ca²⁺ channels. Interactions of the commonly used antagonists of IP₃Rs with IP₃R subtypes are poorly understood.

EXPERIMENTAL APPROACH

IP₃-evoked Ca²⁺ release from permeabilized DT40 cells stably expressing single subtypes of mammalian IP₃R was measured using a luminal Ca²⁺ indicator. The effects of commonly used antagonists on IP₃-evoked Ca²⁺ release and ³H-IP₃ binding were characterized.

KEY RESULTS

Functional analyses showed that heparin was a competitive antagonist of all IP₃R subtypes with different affinities for each (IP₃R3 > IP₃R1 ≥ IP₃R2). This sequence did not match the affinities for heparin binding to the isolated N-terminal from each IP₃R subtype. 2-aminoethoxydiphenyl borate (2-APB) and high concentrations of caffeine selectively inhibited IP₃R1 without affecting IP₃ binding. Neither Xestospongin C nor Xestospongin D effectively inhibited IP₃-evoked Ca²⁺ release via any IP₃R subtype.

CONCLUSIONS AND IMPLICATIONS

Heparin competes with IP₃, but its access to the IP₃-binding core is substantially hindered by additional IP₃R residues. These interactions may contribute to its modest selectivity for IP₃R3. Practicable concentrations of caffeine and 2-APB inhibit only IP₃R1. Xestospongins do not appear to be effective antagonists of IP₃Rs.

Introduction

Inositol 1,4,5-trisphosphate receptors (IP₃R) are intracellular Ca²⁺ channels expressed in the membranes of the endoplasmic reticulum (ER) in most eukaryotic cells (Berridge, 1993; Taylor et al., 1999; Foskett et al., 2007; nomenclature follows Alexander et al., 2013). IP₃Rs are essential links between the many extracellular signals that stimulate PLC and initiation of cytosolic Ca²⁺ signals triggered by IP₃-evoked Ca²⁺ release from the ER. Three genes encode closely related IP₃R subunits in vertebrates, whereas invertebrates have only a single IP₃R gene (Taylor et al., 1999). Each of the three vertebrate IP₃R subtypes encodes a large polypeptide of about 2700 residues, and they share about 70% amino acid sequence identity (Foskett et al., 2007). Within each IP₃R subunit, IP₃ binds to a clam-like IP₃-binding core (IBC; residues 224–604 in IP₃R1) (Bosanac et al., 2002) near the N-terminus. IP₃ binding to the IBC re-orients its relationship with the associated suppressor domain (residues 1–223). That rearrangement disrupts interactions between adjacent subunits within the tetrameric IP₃R leading to gating of the Ca²⁺-permeable channel (Seo et al., 2012). This central channel of each tetrameric IP₃R is formed by transmembrane helices and their associated re-entrant loops. These pore-forming structures lie towards the C-terminal of each subunit. How IP₃-evoked re-arrangement of N-terminal domains of the IP₃R leads to opening of the pore is not yet resolved, although it is likely to be conserved in all IP₃R subtypes and broadly similar for the other major family of intracellular Ca²⁺ channels, ryanodine receptors (Seo et al., 2012).

Most cells express mixtures of IP₃R subtypes, although tissues differ in which complements of IP₃R subunits they express (Taylor et al., 1999). Furthermore, the subunits assemble into both homo-tetrameric and hetero-tetrameric structures (Wojcikiewicz and He, 1995). Although all IP₃Rs are built to a common plan and they are all regulated by IP₃ and Ca²⁺ (Foskett et al., 2007; Seo et al., 2012), the subtypes are subject to different modulatory influences (Patterson et al., 2004; Higo et al., 2005; Foskett et al., 2007; Betzenhauser et al., 2008; Wagner and Yule, 2012) and they are likely to fulfil different physiological roles (Matsumoto et al., 1996; Hattori et al., 2004; Futatsugi et al., 2005; Tovey et al., 2008; Wei et al., 2009). It is, however, difficult to disentangle the physiological roles of IP₃R subtypes in cells that typically express complex mixtures of homo- and hetero-tetrameric IP₃Rs. There are no ligands of IP₃Rs that usefully distinguish among IP₃R subtypes (Saleem et al., 2012; 2013) and nor are there effective antagonists that lack serious side effects (Michelangeli et al., 1995). Heparin (Ghosh et al., 1988), caffeine (Parker and Ivorra, 1991), 2-aminoethoxydiphenyl borate (2-APB) (Maruyama et al., 1997) and Xestospongins (Gafni et al., 1997) have all been widely used to inhibit IP₃-evoked Ca²⁺ release, but each has its limitations (see Results). Furthermore, the interactions of these antagonists with IP₃R subtypes have not been assessed. Peptides derived from myosin light-chain kinase (Nadif Kasri et al., 2006; Sun and Taylor, 2008), the N-terminal of IP₃R1 (Sun et al., 2013) or the BH4 domain of bcl-2 (Monaco et al., 2012) also inhibit IP₃-evoked Ca²⁺ release. These peptides are unlikely to provide routes to useful IP₃R antagonists because they are effective only at high concentrations and they need to be made membrane-permeable. A naturally occurring protein that inhibits IP₃ binding to IP₃R, IRBIT (Ando et al., 2003), has the same limitations as an experimental tool, and it is effective only when phosphorylated. Many other drugs inhibit IP₃-evoked Ca²⁺ release, but none of these has found widespread use (see Michelangeli et al., 1995; Bultynck et al., 2003).

In the present study, we provide the first systematic analysis of the interactions between IP₃R subtypes and each of the commonly used antagonists. We use DT40 cell lines stably expressing only a single mammalian IP₃R subtype to define the effects of these antagonists on IP₃-evoked Ca²⁺ release via each IP₃R subtype.

Methods

Measurement of IP₃-evoked Ca²⁺ release

We used DT40 cells lacking endogenous IP₃Rs (Sugawara et al., 1997), but stably expressing rat IP₃R1 (GenBank accession number GQ233032.1; Pantazaka and Taylor, 2011), mouse IP₃R2 (GU980658.1; Tovey et al., 2010) or rat IP₃R3 (GQ233031.1; Rahman et al., 2009). Cells were grown in suspension in RPMI 1640 medium supplemented with 10% FBS, 1% heat-inactivated chicken serum, 2 mM glutamine and 50 μM 2-mercaptoethanol at 37°C in humidified air containing 5% CO₂. Cells were used or passaged when they reached a density of ∼1.5 × 10⁶ cells mL⁻¹.

A low-affinity Ca²⁺ indicator trapped within the ER of permeabilized DT40 cells was used to measure IP₃-evoked Ca²⁺ release (Tovey et al., 2006; Saleem et al., 2012). Briefly, the ER was loaded with indicator by incubating cells (∼5 × 10⁷ mL⁻¹) in the dark with Mag-fluo-4AM (20 μM) in HEPES-buffered saline (HBS) containing 0.02% (v/v) Pluronic F127 for 1 h at 22°C. HBS had the following composition: 135 mM NaCl, 5.9 mM KCl, 11.6 mM HEPES, 1.5 mM CaCl₂, 11.5 mM glucose, 1.2 mM MgCl₂, pH 7.3. After permeabilization of the plasma membrane with saponin (10 μg·mL⁻¹, 4 min, 37°C) in Ca²⁺-free cytosol-like medium (CLM), permeabilized cells were washed (650× g, 2 min) and resuspended (∼10⁷ mL⁻¹) in Mg²⁺-free CLM containing carbonyl cyanide 4-(trifluoromethoxy) phenylhydrazone (FCCP, 10 μM) to inhibit mitochondria, and supplemented with CaCl₂ to give a final free [Ca²⁺] of 220 nM after addition of 1.5 mM MgATP. Ca²⁺-free CLM had the following composition: 2 mM NaCl, 140 mM KCl, 1 mM EGTA, 20 mM PIPES, 2 mM MgCl₂, pH 7.0. Permeabilized cells were then distributed into 96-well plates (50 μL, 5 × 10⁵ cells per well), centrifuged (300× g, 2 min) and used for experiments at 20°C. Addition of MgATP (1.5 mM) allowed Ca²⁺ uptake by the ER, which was monitored at intervals of ∼1 s using a FlexStation-3 plate reader (MDS Analytical Devices, Berkshire, UK; Tovey et al., 2006). After 2 min, when the ER had loaded to steady-state with Ca²⁺, IP₃ was added with CPA (10 μM) to inhibit further Ca²⁺ uptake. IP₃-evoked Ca²⁺ release was expressed as a fraction of that released by ionomycin (1 μM; Tovey et al., 2006). Similar methods were used to measure IP₃-evoked Ca²⁺ release from intact or permeabilized HEK cells (Tovey et al., 2008). The timings of antagonist additions are described in the figure legends. The affinity of each competitive antagonist (pK_D) was determined from the intercept on the abscissa of the Schild plot.

Concentration–effect relationships were fitted to Hill equations using Prism (version 5.0, GraphPad, San Diego, CA, USA), from which Hill coefficients (h), the fraction of the intracellular Ca²⁺ stores released by a maximally effective concentrations of IP₃, and pEC₅₀ values were calculated.

Expression of N-terminal fragments of IP₃ receptors

The plasmids used for bacterial expression of GST-tagged N-terminal fragments (NT, residues 1–604) of rat IP₃R1, mouse IP₃R2 and rat IP₃R3 have been described, and their coding sequences have been confirmed (Khan et al., 2013). Plasmids were transformed into BL21-CodonPlus (DE3)-RILP competent cells (Rossi and Taylor, 2011), and grown for 12 h at 37°C in 20 mL of Luria-Bertani (LB) medium containing carbecillin (50 μg·mL⁻¹). The volume of medium was then increased to 1 L, and the incubation was continued at 37°C for 3–4 h until the OD₆₀₀ reached 1–1.5. Protein expression was induced by addition of IPTG (0.5 mM) for 20 h at 15°C. Bacteria were harvested (6000× g, 5 min), washed twice with cold PBS, and the pellet was suspended (∼10⁹ cells·mL⁻¹) in 50 mL of Tris-EDTA medium (TEM: 50 mM Tris, 1 mM EDTA, pH 8.3) supplemented with 10% PopCulture, 1 mM 2-mercaptoethanol and protease inhibitor cocktail (Roche, Burgess Hill, West Sussex, UK; 1 tablet per 50 mL). After lysis by incubation with lysozyme (100 μg·mL⁻¹) and RNAse (10 μg·mL⁻¹) for 30 min on ice and then sonication (Transsonic T420 water bath sonicator, Camlab, Cambridge, UK; sonicator, 50 Hz, 30 s), the supernatant was recovered (30,000× g, 60 min, 4°C). The supernatant was mixed with glutathione Sepharose 4B beads (50:1, v/v, lysate : beads) and incubated with gentle end-over-end rotation (6 rpm) for 45 min at 4°C. The beads were then loaded onto a PD-10 column and washed twice with PBS and twice with PreScission cleavage buffer (GE Healthcare) supplemented with 1 mM DTT. The column was then incubated with 0.5 mL of PreScission cleavage buffer containing 1 mM DTT and 80 units of GST-tagged PreScission protease for 12 h at 4°C using gentle end-over-end rotation. The PreScission protease cuts an engineered cleavage site to release the NT free of its GST tag. The eluted NT (∼15 mg protein mL⁻¹) was rapidly frozen and stored at −80°C.

³H-IP₃ binding

Equilibrium competition binding assays were performed at 4°C in 500 μL of CLM (final free [Ca²⁺] = 220 nM) containing purified NT (30 μg) or cerebellar membranes (5 mg protein), ³H-IP₃ (1.5 nM) and appropriate concentrations of competing ligand. Reactions were terminated after 5 min by centrifugation (20,000× g, 5 min) for membranes, or by centrifugation after addition of poly(ethylene glycol)-8000 [30% (w/v), 500 μL] and γ-globulin (30 μL, 25 mg·mL⁻¹) for NT. The pellet was washed (500 μL of 15% PEG or CLM) and solubilized in 200 μL of CLM containing 1% (v/v) Triton-X-100 before liquid scintillation counting. Non-specific binding, whether determined by addition of 10 μM IP₃ or by extrapolation of competition curves to infinite IP₃ concentration, was <10% of total binding. Results were fitted to Hill equations using Prism, from which IC₅₀ values were calculated. K_D (equilibrium dissociation constant) and pK_D (–logK_D) values were calculated from IC₅₀ values using the Cheng and Prusoff equation (Cheng and Prusoff, 1973).

Data analysis

Statistical comparisons used pEC₅₀ (or pK_D) values. For paired comparison of the effect of an antagonist, ΔpEC₅₀ values were calculated, where ΔpEC₅₀ = . Results are expressed as means ± SEM from n independent experiments. Statistical comparisons used paired Student's t-test or anova followed by Bonferroni's test, with P < 0.05 considered significant.

Materials

Sources of many reagents were specified in earlier publications (Rossi et al., 2010a,b; Saleem et al., 2012). IP₃ was from Enzo Life Sciences (Exeter, UK). ³H-IP₃ (19.3 Ci mmol⁻¹) was from PerkinElmer (Buckinghamshire, UK). Heparin (from porcine mucosa, M_r 5000) and cyclopiazonic acid (CPA) were from Fisher Scientific (Loughborough, UK). Caffeine, 2-APB, lysozyme, RNAse, γ-globulin and poly(ethylene glycol)-8000 were from Sigma-Aldrich (Dorset, UK). Xestopongins C and D were from Calbiochem (Gibbstown, NJ, USA) or isolated and characterized as previously described (Gafni et al., 1997). PopCulture was from Novagen (Darmstadt, Germany). Simply Blue stain was from Invitrogen (Renfrewshire, Scotland). Dioxin-free isopropyl-β-D-thiogalactoside (IPTG), and Luria–Bertani agar and broth were from Formedium (Norfolk, UK). Glutathione Sepharose 4B beads and GST-tagged PreScission protease were from GE Healthcare (Buckinghamshire, UK). Carbecillin was from Melford Laboratories (Suffolk, UK). BL21-CodonPlus (DE3)-RILP competent bacteria were from Agilent Technology (Berkshire, UK).

Results

Heparin is a competitive antagonist with different affinities for IP₃ receptor subtypes

Heparin is a competitive antagonist of IP₃-evoked Ca²⁺ release (Ghosh et al., 1988), but it is membrane-impermeable and it has many additional effects. These include uncoupling of receptors from G-proteins (Willuweit and Aktories, 1988; Dasso and Taylor, 1991), stimulation of ryanodine receptors (Ehrlich et al., 1994) and inhibition of IP₃ 3-kinase (Guillemette et al., 1989). To assess the effects of heparin on each IP₃R subtype, permeabilized DT40 cells expressing each of the three IP₃R subtypes were incubated with heparin for 35 s. The effect of IP₃ on Ca²⁺ release from the intracellular stores was then assessed (Figure 1A). In permeabilized DT40-IP₃R1 cells, heparin caused parallel rightward shifts of the concentration–response relationship for IP₃-evoked Ca²⁺ release (Figure 1B). Schild plots, which had slopes of 0.95 ± 0.02 (mean ± SEM, n = 3), established that the equilibrium dissociation constant (K_D) for heparin was 4.1 μg·mL⁻¹ (pK_D = 5.39 ± 0.00) (Figure 1C). Similar results were obtained when adenophostin A (AdA), a high-affinity agonist of IP₃Rs (Rossi et al., 2010b; Saleem et al., 2013), was used to stimulate Ca²⁺ release. The Schild plots had slopes of 0.94 ± 0.03 (n = 3) and the K_D for heparin was 6.9 μg·mL⁻¹ (pK_D = 5.16 ± 0.05) (Figure 1D and E; Table 1).

Figure 1.

Heparin competitively inhibits IP₃-evoked Ca²⁺ release via type 1 IP₃ receptors. (A) Typical traces from a population of permeabilized DT40-IP₃R1 cells showing the fluorescence (RFU, relative fluorescence units) recorded from a luminal Ca²⁺ indicator after addition of MgATP (1.5 mM), heparin (400 μg·mL⁻¹, red lines; or CLM alone, black lines) and then IP₃ (1 or 100 μM). The traces show average responses from two wells in a single plate. (B) Experiments similar to those in A show concentration-dependent effects of IP₃ on Ca²⁺ release in the presence of the indicated concentrations of heparin. (C) Schild analysis of the results shown in B. (D, E) Similar analyses of the effects of heparin on AdA-evoked Ca²⁺ release via IP₃R1. Results (B–E) are means ± SEM from three experiments.

Table 1.

Effects of heparin on IP₃-evoked Ca²⁺ release and IP₃ binding

		Functional analysis		aBinding	pEC50(IP3)-pKD(heparin)
		IP3 or AdA	heparin	heparin
		pEC50	pKD	pKD
IP3R1	IP3	7.47 ± 0.02	5.39 ± 0.00	4.66	2.08 ± 0.02
IP3R1	AdA	8.35 ± 0.03	5.16 ± 0.05	–	–
IP3R2	IP3	6.82 ± 0.04	4.66 ± 0.07	4.62	2.16 ± 0.09*
IP3R3	IP3	6.66 ± 0.07	5.55 ± 0.09	5.34	1.11 ± 0.08*
IP3R3	AdA	7.71 ± 0.01	5.68 ± 0.04	–	–

From experiments similar to those shown in Figures 1 and 2, AdA or IP₃-evoked Ca²⁺ release and their sensitivity to heparin were used to determine pEC₅₀ (as M) and pK_D (as g mL⁻¹) for DT40 cells expressing IP₃R1, IP₃R2 or IP₃R3. Results are means ± SEM from three independent experiments (six for IP₃R3).

^aThe affinities for heparin determined from equilibrium-competition binding with ³H-IP₃ to Sf9 membranes expressing IP₃R1-3 are reproduced from (Nerou et al., 2001). The batch of heparin used for those binding studies was different from that used for the work reported here. The final column (derived from the results shown in Figures 1B,C and 2A–D) shows paired comparisons of pEC₅₀(IP₃) – pK_D(heparin) as a means of reporting the relative effectiveness with which heparin might be expected to block IP₃-evoked Ca²⁺ release via different IP₃R subtypes. The results suggest that IP₃R3 is likely to be substantially more susceptible to inhibition than IP₃R1 or IP₃R2.

^*Denotes a value significantly different from IP₃R1 in the final column (P < 0.05).

A similar analysis of the effects of heparin on IP₃-evoked Ca²⁺ release from permeabilized DT40-IP₃R2 cells was also consistent with competitive antagonism. The slope of the Schild plots was 0.97 ± 0.06 (n = 3) and the K_D for heparin was 22 μg·mL⁻¹ (pK_D = 4.66 ± 0.07) ( Figure 2A and B). IP₃R3 are less sensitive to IP₃ than the other subtypes (Iwai et al., 2007; Saleem et al., 2013) (Table 1). This made it difficult to add IP₃ at concentrations sufficient to achieve maximal Ca²⁺ release in the presence of heparin concentrations greater than 5 μg·mL^–1 (Figure 2C). Assuming the maximal response to IP₃ was unaffected by heparin, we used the concentrations of IP₃ that evoked release of 40% of the intracellular stores to construct Schild plots for IP₃R3. The results were consistent with competitive antagonism. The slope of the Schild plots was 1.14 ± 0.41 (n = 3) and the K_D for heparin was 2.8 μg·mL⁻¹ (pK_D = 5.55 ± 0.09) (Figure 2D and Table 1). AdA has ∼10-fold higher affinity than IP₃ for all three IP₃R subtypes (Table 1) (Rossi et al., 2010a; Saleem et al., 2013), and we have shown that the affinity of heparin for IP₃R1 is similar whether IP₃ or AdA is used to evoke Ca²⁺ release (Figure 1B–E). To obtain an independent measure of the affinity of IP₃R3 for heparin, free of the problems associated with using IP₃, we therefore repeated the Schild analysis using AdA to stimulate Ca²⁺ release. These conditions provided complete concentration–effect relationships for AdA at a wider range of heparin concentrations (Figure 2E). The Schild plots had a slope of 0.98 ± 0.04 (n = 6) and the K_D for heparin was 2.1 μg·mL⁻¹ (pK_D = 5.68 ± 0.04) (Figure 2F and Table 1). The affinity of heparin for IP₃R3 was therefore similar whether measured using IP₃ or AdA to evoke Ca²⁺ release.

Figure 2.

Heparin is a competitive antagonist with different affinities for types 2 and 3 IP₃ receptors. (A) Concentration-dependent release of Ca²⁺ by IP₃ from the intracellular stores of DT40-IP₃R2 cells in the presence of the indicated concentrations of heparin added 35 s before IP₃. (B) Schild plot of the results. (C–F) Similar analyses of DT40-IP₃R3 cells stimulated with IP₃ (C, D) or AdA (E, F). For D, where maximal attainable concentrations of IP₃ were insufficient to evoke maximal responses in the presence of the highest concentrations of heparin, the Schild plot shows dose ratios calculated from IP₃ concentrations that evoked 40% Ca²⁺ release. Results (A–F) are mean ± SEM from three experiments.

These functional analyses establish that heparin is a competitive antagonist of IP₃ at all three IP₃R subtypes, but with different affinities for each (IP₃R3 > IP₃R1 ≥ IP₃R2) (Table 1). The results are consistent with an analysis of IP₃ binding to mammalian IP₃R expressed in Sf9 cells (Nerou et al., 2001), where the pK_D values and rank order of heparin affinity (IP₃R3 > IP₃R1 ∼ IP₃R2) were similar to those from the present functional analyses (Table 1).

Heparin binding is not solely determined by its interactions with the IP₃-binding site

Activation of IP₃Rs is initiated by binding of IP₃ to the IP₃-binding core (IBC, residues 224-604 of IP₃R1) within the N-terminal region of each IP₃R subunit (see Introduction) (Seo et al., 2012). The only contacts between IP₃ and the IP₃R are mediated by residues within the IBC (Bosanac et al., 2002), but interaction of the N-terminal suppressor domain (residues 1-223) with the IBC reduces its affinity for IP₃. Hence, the IBCs from different IP₃R subtypes bind IP₃ with similar affinity, whereas the larger N-terminal regions (NT, residues 1-604) have lower affinities that differ between subtypes. The NTs bind IP₃ with two- to threefold greater affinities than those of full-length IP₃Rs, but the NTs and full-length IP₃Rs have the same rank order of affinities for IP₃ (NT2 > NT1 > NT3) (Iwai et al., 2007; Rossi et al., 2009). The results shown in Figure 3A and B, which show IP₃ binding to bacterially expressed NTs from each of the three IP₃R subtypes (NT1-3), confirm previous results. Surprisingly, however, equilibrium-competition binding of heparin to NTs in medium that matches that used to measure IP₃-evoked Ca²⁺ release was not consistent with the results obtained from functional analyses (Figure 3C). The affinity of the NT for heparin was up to 2000-fold greater than that measured in functional analyses, and the rank order of affinity for heparin was different for NTs (NT2 > NT1 > > NT3) and full-length IP₃Rs (IP₃R3 > IP₃R1 ≥ IP₃R2) (Nerou et al., 2001; Tables 1 and 2).

Figure 3.

Heparin binding is not solely determined by its interactions with the IP₃-binding core. (A) Immunoblots of purified NT1-3 (∼15 μL protein per lane) using an antiserum that recognizes a conserved sequence within all three IP₃R subtypes (residues 62–75 in rat IP₃R1). The positions of M_r markers (kDa) are shown alongside each blot. (B, C) Equilibrium-competition binding of IP₃ (B) and heparin (C) to purified NT1-3 in CLM. (D, E) Similar analyses of binding to cerebellar membranes (IP₃R1). Results (B–E) are means ± SEM from three to six experiments.

Table 2.

Heparin and IP₃ binding to N-terminal fragments of IP₃ receptor subtypes

	NT1	NT2	NT3	IP3R1
IP3	7.76 ± 0.07	8.67 ± 0.15	7.39 ± 0.08	7.13 ± 0.08
Heparin	7.42 ± 0.09	7.95 ± 0.32	6.59 ± 0.09*	5.61 ± 0.13

Equilibrium-competition binding with ³H-IP₃ was used to measure pK_D values for IP₃ (as M) and heparin (as g mL⁻¹) binding to purified NT1-3 and cerebellar membranes (IP₃R1). Results are means ± SEM from three to six experiments.

^*Denotes a significant difference from NT1 (P < 0.05) for .

IP₃R1 is the major (>99%) subtype in cerebellar membranes (Wojcikiewicz, 1995). Equilibrium-competition binding of heparin to cerebellar membranes in CLM established that the affinity of IP₃R1 for heparin (pK_D = 5.61 ± 0.13, n = 3) was similar to that derived from Schild analysis of DT40-IP₃R1 cells (pK_D = 5.39 ± 0.00, n = 3) and similar to that reported for heparin binding to IP₃R1 heterologously expressed in Sf9 cells (Nerou et al., 2001), but very different to the heparin affinity of NT1 (pK_D = 7.42 ± 0.09, n = 3) (Tables 1 and 2). These results demonstrate that the IBC is not the only determinant of competitive heparin binding to IP₃Rs and suggest either that access of heparin to the IBC is influenced by additional interactions or that heparin binding to an additional site affects IP₃R gating.

2-APB selectively inhibits Ca²⁺ release via type 1 IP₃ receptors without affecting IP₃ binding

2-APB is membrane-permeant and is often used to inhibit IP₃-evoked Ca²⁺ release (Maruyama et al., 1997; Missiaen et al., 2001; Bilmen et al., 2002), but it has many additional effects. These include modulation of store-operated Ca²⁺ entry (Goto et al., 2010) and inhibition of the sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase (SERCA) that mediates Ca²⁺ sequestration by the ER (Missiaen et al., 2001; Bilmen et al., 2002; Bultynck et al., 2003). In permeabilized DT40-IP₃R1 cells, 50 μM 2-APB had no effect on Ca²⁺ uptake by the ER, although higher concentrations reduced the steady-state Ca²⁺ content (Figure 4A and B). This is consistent with high concentrations of 2-APB causing inhibition of SERCA.

Figure 4.

2-APB selectively inhibits Ca²⁺ release via type 1 IP₃ receptors. (A) Ca²⁺ uptake into the intracellular stores of permeabilized DT40-IP₃R1 cells is shown after addition of ATP in the presence of the indicated concentrations of 2-APB. Each trace is the average from two wells in a single plate. (B) Summary results show effects of 2-APB on Ca²⁺ contents measured 180 s after addition of ATP. (C–E) Concentration-dependent effects of IP₃ on Ca²⁺ release from permeabilized DT40-IP₃R1-3 cells alone or with the indicated concentrations of 2-APB added 35 s before IP₃. (F) Binding of ³H-IP₃ (1.5 nM) to cerebellar membranes (IP₃R1), with 3 μM IP₃ (non-specific) or with 2-APB. Results (B–F) are means ± SEM from three to nine experiments. *P < 0.05, significantly different from control.

In permeabilized DT40-IP₃R1 cells, 2-APB caused a concentration-dependent inhibition of IP₃-evoked Ca²⁺ release (Figure 4C). With 50 μM 2-APB, the highest concentration that avoids inhibition of Ca²⁺ uptake, there was an approximately sevenfold decrease in IP₃ sensitivity (ΔpEC₅₀ = 0.84 ± 0.12) with no effect on the maximal response to IP₃ (Figure 4C). The same concentration of 2-APB (50 μM) had no significant effect on IP₃-evoked Ca²⁺ release from permeabilized DT40-IP₃R2 or DT40-IP₃R3 cells (Figure 4D and E). When the 2-APB concentration was increased to 100 μM, which caused some inhibition of Ca²⁺ uptake (Figure 4A and B), there was some inhibition of IP₃R3, but no effect on IP₃-evoked Ca²⁺ release via IP₃R2 (Figure 4D and E; Table 3).

Table 3.

Selective inhibition of IP₃ receptor subtypes by common antagonists

	IP3R1		IP3R2		IP3R3
	ΔpEC50 (M)	ΔMax (%)	ΔpEC50 (M)	ΔMax (%)	ΔpEC50 (M)	ΔMax (%)
Heparin, 400 μg·mL−1	1.88 ± 0.05*	−7 ± 2	ND	–	2.34 ± 0.07*	−3 ± 3
Heparin, 800 μg·mL−1	ND	–	1.49 ± 0.09*	−4 ± 2	ND	–
Caffeine, 70 mM	0.61 ± 0.07*	12 ± 4	−0.2 ± 0.07	−1 ± 0	−0.07 ± 0.08	0 ± 5
2-APB, 50 μM	0.84 ± 0.12*	0 ± 4	−0.05 ± 0.10	0 ± 4	0.02 ± 0.09	8 ± 4
Xestospongin C, 20 μM	0.21 ± 0.10*	6 ± 2*	−0.06 ± 0.04	1 ± 1	0.12 ± 0.03*	1 ± 2
Xestospongin D, 20 μM	0.26 ± 0.09*	18 ± 2*	−0.15 ± 0.05	8 ± 3*	0.21 ± 0.10*	2 ± 2

Summary of the functional analyses of antagonists on IP₃-evoked Ca²⁺ release from permeabilized DT40-IP₃R1-3 cells. The pEC₅₀ values for IP₃ and the maximal Ca²⁺ release are each expressed relative to the response evoked in paired controls without antagonist (Δ = control – response with antagonist). A positive Δ value demonstrates an inhibition of IP₃-evoked Ca²⁺ release by the antagonist. The results with Xestospongins C and D are pooled from experiments that included pre-incubation periods of 7 and 12 min (see Supporting Information Table S1). Results are means ± SEM from three to nine experiments.

^*Denotes a value significantly greater than 0 (P < 0.025, one-tailed test).

ND, not determined.

Binding of ³H-IP₃ to IP₃R1 of cerebellar membranes in CLM was unaffected by 2-APB (Figure 4F) consistent with published results (Maruyama et al., 1997; Bilmen et al., 2002). This demonstrates that inhibition of IP₃R1 by 2-APB is neither due to competition with IP₃ nor to allosteric inhibition of IP₃ binding.

Caffeine is a low-affinity antagonist of type 1 IP₃ receptors

Caffeine is another membrane-permeant antagonist of IP₃-evoked Ca²⁺ release (Parker and Ivorra, 1991; Brown et al., 1992; Bultynck et al., 2003; Laude et al., 2005), but it is effective only at high (mM) concentrations and it has many additional effects (Michelangeli et al., 1995; Taylor and Tovey, 2010). These include stimulation of ryanodine receptors, inhibition of cyclic nucleotide phosphodiesterases, competitive antagonism of adenosine receptors, and effects on the fluorescence of some Ca²⁺ indicators (Brown et al., 1992; Ehrlich et al., 1994; Michelangeli et al., 1995; McKemy et al., 2000; Taylor and Tovey, 2010). High concentrations of caffeine (10–70 mM) inhibited Ca²⁺ release via IP₃R1 (Figure 5A) without affecting ³H-IP₃ binding to cerebellar membranes (Figure 5D). The latter is consistent with published work (Brown et al., 1992). The maximal attainable concentration of caffeine (70 mM) caused an approximately fourfold decrease in IP₃ sensitivity (ΔpEC₅₀ = 0.61 ± 0.07) (Figure 5A). Caffeine had no significant effect on IP₃-evoked Ca²⁺ release via IP₃R2 or IP₃R3 (Figure 5B and C; Table 3). At the highest concentration used (70 mM), caffeine significantly reduced the Ca²⁺ content of the intracellular stores, but this inhibition was similar for DT40 cells expressing each of the IP₃R subtypes (Figure 5E). Inhibition of Ca²⁺ sequestration by the ER is unlikely, therefore, to account for the selective inhibition of IP₃-evoked Ca²⁺ release via IP₃R1 (Table 3). These results demonstrate that a high concentration of caffeine modestly, but selectively, inhibits IP₃-evoked Ca²⁺ release via IP₃R1 without affecting IP₃ binding.

Figure 5.

Caffeine is a low-affinity antagonist of type 1 IP₃R receptors. (A–C) Concentration-dependent effects of IP₃ on Ca²⁺ release from permeabilized DT40-IP₃R1-3 cells in the presence of the indicated concentrations of caffeine added 4 min before IP₃. (D) Binding of ³H-IP₃ (1.5 nM) to cerebellar membranes alone (total), with 3 μM IP₃ (non-specific) or caffeine. (E) Effect of caffeine added 2 min before ATP on the steady-state Ca²⁺ content of the intracellular stores (percentage of matched control cells) measured 90 s after addition of ATP to DT40-IP₃R1-3 cells. Results (A–E) are means ± SEM from three experiments. *P < 0.05 significantly different from control.

Xestospongins do not effectively inhibit IP₃-evoked Ca²⁺ release

Xestospongin C is membrane-permeant and was reported to inhibit IP₃-evoked Ca²⁺ release from cerebellar microsomes (IC₅₀ = 358 nM) without affecting IP₃ binding (Gafni et al., 1997). Xestospongin D is less potent. Higher concentrations of Xestospongin C (10–20 μM) were required to inhibit IP₃-evoked Ca²⁺ release in intact cells. We assessed the effects of Xestospongins C and D from different suppliers (see Materials) on Ca²⁺ release mediated by each of the three IP₃R subtypes.

Pre-incubation of permeabilized DT40 cells with Xestospongin C (5–20 μM from either source) for 5–12 min before addition of IP₃ had no significant effect on IP₃-evoked Ca²⁺ release mediated by any of the three IP₃R subtypes (Supporting Information Table S1). Figure 6A–C show IP₃-evoked Ca²⁺ release after a 5 min pre-incubation with 5 μM purified Xestospongin C (Gafni et al., 1997). It had no significant effect on either the response to IP₃ (Figure 6A–C) or the Ca²⁺ content of the stores (Figure 6D). Pooling all experiments with the highest concentration of Xestospongin C (20 μM, n = 6) revealed a statistically significant (P < 0.025, one-tailed test), but very small, inhibition of the maximal response from IP₃R1, and an even smaller increase in pEC₅₀ for IP₃R1 and IP₃R3 (Table 3 and Supporting Information Table S1).

Figure 6.

Xestospongins do not effectively inhibit IP₃ receptors. (A–C) IP₃-evoked Ca²⁺ release from permeabilized DT40-IP₃R1-3 cells is shown with or without 5 μM Xestospongin C (from Gafni et al., 1997) added 5 min before IP₃. (D) Effects of Xestospongin C (5–20 μM) added 5–12 min before ATP on the Ca²⁺ content of the intracellular stores (percentages of matched controls without Xestospongin). (E–H) Similar analyses using Xestospongin D (10 μM added 5 min before IP₃). Results (A–H) are means ± SEM from three experiments.

Similar treatments with Xestospongin D (10–20 μM from either source) for 5–12 min caused a modest, but statistically significant (P < 0.025, one-tailed test), inhibition of IP₃-evoked Ca²⁺ release via IP₃R1 (Supporting Information Table S1). Figure 6E–H show that a 5 min pre-incubation with 10 μM purified Xestospongin D (Gafni et al., 1997) had no effect on the Ca²⁺ content of the intracellular stores, but modestly inhibited IP₃-evoked Ca²⁺ release via IP₃R1 (P < 0.025, one-tailed test, Figure 6E). Pooling results with the highest concentration of Xestospongin D (20 μM, n = 6) revealed a statistically significant (P < 0.025, one-tailed test), but very small, inhibition of the maximal response from IP₃R1 and IP₃R2, and a tiny increase in the pEC₅₀ for IP₃R1 and IP₃R3 (Table 3 and Supporting Information Table S1). These small inhibitory effects of Xestospongins C and D are not sufficient to be useful, and nor are they sufficient to reliably assess whether there is any subtype-selective interaction of Xestospongins with IP₃Rs.

We also assessed the effects of Xestospongins on IP₃-evoked Ca²⁺ release from intact and permeabilized HEK cells. IP₃ caused a concentration-dependent release of Ca²⁺ from the intracellular stores of permeabilized HEK cells (Figure 7A and B). Pre-incubation of the permeabilized cells for 5 min with Xestospongin C (5 μM) or Xestospongin D (10 μM) had no effect on the Ca²⁺ content of the intracellular stores (Figure 7C) or the Ca²⁺ release evoked by IP₃ (Figure 7A and B). Carbachol, via endogenous M₃ muscarinic receptors of HEK cells, stimulates PLC and thereby IP₃-evoked Ca²⁺ release. Preincubation of HEK cells with Xestospongin C or D (10 μM) for 30 min had no significant effect on the Ca²⁺ signals evoked by any concentration of carbachol (Figure 7D). This conflicts with published results from similar experiments, where Xestospongin C (10 μM for 30 min) caused substantial, though incomplete, inhibition of carbachol-evoked Ca²⁺ signals (Kurian et al., 2009). It is, however noteworthy, in light of evidence that Xestospongins have been reported to inhibit Ca²⁺ uptake into the ER (Castonguay and Robitaille, 2002; Solovyova et al., 2002), that in the experiments from Kurian et al. HEK cells were incubated with Xestospongin for 30 min in Ca²⁺-free medium, while in our experiments extracellular free Ca²⁺ was removed immediately before stimulation with carbachol. The discrepant results may, therefore, reflect an increased loss of Ca²⁺ from intracellular stores during prolonged exposure to Xestospongin in Ca²⁺-free medium.

Figure 7.

Xestospongins do not inhibit IP₃-evoked Ca²⁺ signals in HEK cells. (A–C) Permeabilized HEK cells were incubated with Xestospongin C (5 μM) or Xestospongin D (10 μM) for 5 min before addition of IP₃. Both Xestospongins were prepared as described (Gafni et al., 1997). Results show IP₃-evoked Ca²⁺ release (A, B) or the steady-state Ca²⁺ content of the intracellular stores (C, as a percentage of matched controls without Xestospongin). (D) Concentration-dependent effects of carbachol on the increase in intracellular free Ca²⁺ concentration [Ca²⁺]_i of intact fluo-4-loaded HEK cells after treatment with Xestospongins C or D (10 μM for 30 min). pEC₅₀ (M) values for the carbachol-evoked Ca²⁺ signals were 4.99 ± 0.13, 4.92 ± 0.23 and 4.70 ± 0.11 for control cells and cells treated with Xestospongins C and D respectively. Results (A–D) are means ± SEM. from three experiments.

Discussion

Acute analyses of IP₃-evoked Ca²⁺ signalling are handicapped by lack of effective and selective antagonists (Michelangeli et al., 1995; Bultynck et al., 2003). Furthermore, the subtype-selectivity and in many cases the mechanism of action of the antagonists that are routinely used are not known. We have addressed these issues by examining the functional effects of the most widely used antagonists of IP₃R in cells expressing only a single IP₃R subtype.

Heparin is a competitive antagonist of IP₃ at cerebellar IP₃Rs (Ghosh et al., 1988), most likely because as a polyanion it may partially mimic the phosphate groups of IP₃. That is consistent with evidence that other polyanions, like decavanadate, ATP and dextran sulphate, can also competitively inhibit IP₃Rs (Bultynck et al., 2003). Our functional analyses establish that heparin is a competitive antagonist of all three IP₃R subtypes, but with modestly different affinities for each (IP₃R3 > IP₃R1 ≥ IP₃R2) (Figures 1 and 2; Table 1). The affinities of IP₃R subtypes for heparin derived from functional analyses were similar to those determined from equilibrium-competition binding to native IP₃R1 (Figure 3E) or to heterologously expressed IP₃R subtypes (Table 1). However, heparin bound to N-terminal fragments (NT) of IP₃Rs that include the IBC with an affinity that was up to 2000-fold greater than its affinity for the corresponding full-length IP₃R (Tables 1 and 2). Furthermore, the rank order of heparin affinity for IP₃R1-3 and NT1-3 was different. We conclude that heparin inhibits IP₃-evoked Ca²⁺ release by competing with IP₃, but its access to the IBC is substantially impaired in full-length IP₃Rs within native membranes. Phospholipids may contribute to the substantially lesser affinity of heparin for IP₃R in native membranes by electrostatically repelling the approach of polyanionic heparin to the membrane-bound IBC. In addition, we suggest that charged residues on the IP₃R surface may differentially influence heparin access to the IBC of each IP₃R subtype and thereby contribute to the modestly different affinities of heparin for IP₃R subtypes (Table 1). Our observations have more general significance for analyses of competitive antagonism. We have demonstrated that properties of either the receptor or its environment that are remote from the ligand-binding site may significantly affect the apparent affinity of a receptor for a competitive antagonist.

Because heparin is a competitive antagonist of IP₃ (Figures 1 and 2), its experimental utility will depend on its affinity relative to IP₃ for each IP₃R subtype. Table 1 addresses this issue by comparing measured affinities for heparin with EC₅₀ values for IP₃ as an estimate of the relative affinity of each IP₃R subtype for IP₃. The analysis indicates that within native cells, responses of IP₃R3 to IP₃ are likely to be more susceptible to inhibition by heparin than the responses mediated by other IP₃R subtypes.

Both 2-APB and caffeine selectively inhibited IP₃-evoked Ca²⁺ release via IP₃R1, without affecting IP₃ binding (Figures 4 and 5; Table 3). Higher concentrations of 2-APB caused some inhibition of IP₃R3, but this was accompanied by inhibition of ER Ca²⁺ uptake (Figure 4). The highest concentration of caffeine used (70 mM) also inhibited Ca²⁺ sequestration by the ER, but without significantly affecting the sensitivity to IP₃ of IP₃R2 or IP₃R3, or the fraction of the remaining Ca²⁺ stores released via them by a maximally effective concentration of IP₃ (Figure 5). Previous analyses of cells expressing different mixtures of native IP₃R subtypes have also suggested that IP₃R2 may be resistant to inhibition by 2-APB (Gregory et al., 2001; Hauser et al., 2001; Kukkonen et al., 2001; Bootman et al., 2002; Soulsby and Wojcikiewicz, 2002) and caffeine (Kang et al., 2010). The mechanism of action of 2-APB is unresolved, but for IP₃R1 caffeine appears to compete with ATP for the site through which ATP potentiates IP₃-evoked Ca²⁺ release (Missiaen et al., 1994; Maes et al., 2001). This mechanism appears not to explain the actions of 2-APB (Missiaen et al., 2001). ATP potentiates IP₃-evoked Ca²⁺ release via all three IP₃R subtypes (Smith et al., 1985; Mak et al., 1999; Maes et al., 2001; Tu et al., 2005; Betzenhauser et al., 2008), but the mechanisms and ATP-binding sites differ (Betzenhauser et al., 2008; 2009; Betzenhauser and Yule, 2010). Work from Yule and his colleagues suggests that IP₃R2 is most sensitive to ATP and for it, but not other IP₃R subtypes, an ATPB site within each IP₃R subunit mediates the potentiating effect of ATP (Betzenhauser and Yule, 2010). It is, therefore, tempting to speculate that the different sensitivities of IP₃R subtypes to inhibition by caffeine (Figure 5) may be related to their different modes of regulation by ATP.

Xestospongins were initially shown to inhibit IP₃-evoked Ca²⁺ release selectively (Gafni et al., 1997), and numerous subsequent analyses of their effects on intact cells are consistent with inhibition of IP₃Rs (e.g. Bishara et al., 2002; Duncan et al., 2007; Oka et al., 2002; Ozaki et al., 2002; Rosado and Sage, 2000; Schafer et al., 2001; Yuan et al., 2005), but few of these later analyses directly addressed the effects of Xestospongins on IP₃Rs (e.g. Oka et al., 2002; Ozaki et al., 2002). The latter is important because Xestospongins have additional effects that include inhibition of SERCA (De Smet et al., 1999; Castonguay and Robitaille, 2002; Solovyova et al., 2002), store-operated Ca²⁺ entry (Bishara et al., 2002), L-type Ca²⁺ channels and Ca²⁺-activated K⁺ channels (Ozaki et al., 2002), and modulation of ryanodine receptors (Ta et al., 2006). The potencies of Xestospongins also differ between studies and some reports challenge whether they effectively inhibit IP₃Rs (Solovyova et al., 2002; Duncan et al., 2007; Govindan and Taylor, 2012). We used two sources of Xestospongins C and D, a range of concentrations and incubation periods, two different cell types (see also Govindan and Taylor, 2012), and both intact and permeabilized cells. Although the Xestospongins caused some inhibition of IP₃-evoked Ca²⁺ release, none of our analyses succeeded in demonstrating that attainable (≤20 μM) concentrations of Xestospongins substantially inhibited any IP₃R subtype (Figures 6 and 7; Table 3; Supporting Information Table S1).

We conclude that none of the commonly used antagonists of IP₃Rs is free of pitfalls. Heparin is perhaps the most reliable, it is competitive with IP₃, but it is membrane-impermeant, and its binding to the IBC of IP₃Rs is influenced by more distant residues that cause it to bind with different affinity to each IP₃R subtype (Figures 3). Caffeine and 2-APB are membrane-permeant, they do not compete with IP₃, but neither achieves effective inhibition of IP₃Rs without affecting other Ca²⁺-regulating proteins, and both show selectivity for IP₃R1 (Figures 4 and 5). Xestospongins are membrane-permeant and reported to inhibit IP₃-evoked Ca²⁺ release without affecting IP₃ binding (Gafni et al., 1997), but in our hands they do not inhibit any IP₃R subtype (Figures 6 and 7).

Glossary

2-APB

2-aminoethoxydiphenyl borate

AdA

adenophostin A

CLM

cytosol-like medium

CPA

cyclopiazonic acid

ER

endoplasmic reticulum

HBS

HEPES-buffered saline

IBC

IP₃-binding core (residues 224-604)

IP₃

inositol 1,4,5-trisphosphate

IPTG

isobutyl-β-D-thiogalactoside

NT1-3

residues 1-604 of IP₃R1-3

SERCA

sarcoplasmic/endoplasmic reticulum Ca²⁺-ATPase

TEM

Tris-EDTA medium

Author contributions

HS performed and analysed experiments. TFM provided reagents. CWT and SCT supervised the project and contributed to analysis. CWT wrote the paper. All authors reviewed the paper.

Conflict of interest

None

Supporting information

Additional Supporting Information may be found in the online version of this article at the publisher's web-site:

http://dx.doi.org/10.1111/bph.12685

Table S1 Xestospongins ineffectively inhibit IP₃-evoked Ca²⁺ release.