Selective determinants of inositol 1,4,5-trisphosphate and adenophostin A interactions with type 1 inositol 1,4,5-trisphosphate receptors

Abstract

BACKGROUND AND PURPOSE

Adenophostin A (AdA) is a potent agonist of inositol 1,4,5-trisphosphate receptors (IP₃R). AdA shares with IP₃ the essential features of all IP₃R agonists, namely structures equivalent to the 4,5-bisphosphate and 6-hydroxyl of IP₃, but the basis of its increased affinity is unclear. Hitherto, the 2′-phosphate of AdA has been thought to provide a supra-optimal mimic of the 1-phosphate of IP₃.

EXPERIMENTAL APPROACH

We examined the structural determinants of AdA binding to type 1 IP₃R (IP₃R1). Chemical synthesis and mutational analysis of IP₃R1 were combined with ³H-IP₃ binding to full-length IP₃R1 and its N-terminal fragments, and Ca²⁺ release assays from recombinant IP₃R1 expressed in DT40 cells.

KEY RESULTS

Adenophostin A is at least 12-fold more potent than IP₃ in functional assays, and the IP₃-binding core (IBC, residues 224–604 of IP₃R1) is sufficient for this high-affinity binding of AdA. Removal of the 2′-phosphate from AdA (to give 2′-dephospho-AdA) had significantly lesser effects on its affinity for the IBC than did removal of the 1-phosphate from IP₃ (to give inositol 4,5-bisphosphate). Mutation of the only residue (R568) that interacts directly with the 1-phosphate of IP₃ decreased similarly (by ∼30-fold) the affinity for IP₃ and AdA, but mutating R504, which has been proposed to form a cation-π interaction with the adenine of AdA, more profoundly reduced the affinity of IP₃R for AdA (353-fold) than for IP₃ (13-fold).

CONCLUSIONS AND IMPLICATIONS

The 2′-phosphate of AdA is not a major determinant of its high affinity. R504 in the receptor, most likely via a cation-π interaction, contributes specifically to AdA binding.

Introduction

Receptors for inositol 1,4,5-trisphosphate (IP₃R, nomenclature follows Alexander et al., 2009) are intracellular Ca²⁺ channels. They are expressed in the membranes of the endoplasmic reticulum of most animal cells (Foskett et al., 2007) and they both initiate and propagate the Ca²⁺ signals evoked by receptors that stimulate IP₃ formation (Berridge et al., 2003). In vertebrates, three genes encode closely related subtypes of the IP₃R, which assemble into both homo- and hetero-tetrameric channels (Taylor et al., 1999). The different subtypes share many features (Foskett et al., 2007). Each subunit has a single IP₃-binding site towards the N-terminal, a large cytosolic regulatory domain, and six transmembrane domains, the last pair of which from each subunit, together with the intervening luminal loop, form the pore (Ramos-Franco et al., 1999; Taylor et al., 2004). For all IP₃R, IP₃ binding initiates the conformational changes that lead to opening of the channel. The IP₃-binding core [IBC, residues 224–604 of type 1 IP₃R (IP₃R1)] is entirely responsible for this initial recognition. The two domains (α and β) of the IBC form a clam-like structure lined with the basic residues that coordinate the phosphate groups of IP₃ (Bosanac et al., 2002) (Figure 1A). The 4,5-bisphosphate and 6-hydroxyl groups of IP₃ are important for binding to IP₃R (Potter and Lampe, 1995), and each forms extensive interactions with the IBC. The 4-phosphate forms hydrogen bonds with several residues in the β-domain, the 5-phosphate is hydrogen-bonded to residues predominantly within the α-domain and the 6-hydroxyl interacts indirectly via water with the backbone of K569 (Figure 1A). The 1-phosphate of IP₃ is not essential for binding to IP₃R, but it substantially increases the affinity of IP₃ (Nerou et al., 2001); it interacts directly only with R568 and indirectly via a water molecule with the backbone of K569 (Bosanac et al., 2002) (Figure 1A).

Figure 1.

Structures of the IBC and the ligands used. Structure of the IBC (PDB 1N4K) with the enlarged panel highlighting the residues (R504, R568 and K569) mutated in this study and their interactions with the phosphate groups of IP₃. The red spheres represent water (A). Structures of the ligands used highlighting the 1-phosphate of IP₃ and 2′-phosphate of AdA (B). AdA, adenophostin A; IBC, IP₃-binding core; IP₃, inositol 1,4,5-trisphosphate.

Adenophostins (Figure 1B) were originally isolated from Penicillium brevicompactum (Takahashi et al., 1994a,b,c;). They are potent agonists of IP₃R (Takahashi et al., 1994b,c; Hirota et al., 1995; Marchant et al., 1997a; Shuto et al., 1998; Correa et al., 2001); they are not metabolized by the enzymes that degrade IP₃ and their structures are based on a glucose, rather than a myo-inositol, ring (Figure 1B). Adenophostin A (AdA) has proven a useful tool with which to explore the properties of IP₃R (Hirota et al., 1995; Dellis et al., 2006; Marchant and Parker, 1998; Yoshida et al., 1998; Parekh et al., 2002), and it has generated considerable interest in the synthesis of novel AdA analogues (Shuto et al., 1998; Correa et al., 2001; Borissow et al., 2005; Mochizuki et al., 2006).

Inositol 1,4,5-trisphosphate and AdA are each full agonists of the IP₃R (Rossi et al., 2009). Both IP₃ and AdA bind to the IBC, and despite their structural differences, the 3",4"-bisphosphate and 2"-hydroxyl groups of AdA evidently mimic the essential 4,5-bisphosphate and 6-hydroxyl of IP₃ (Figure 1B) (Takahashi et al., 1994c; Hotoda et al., 1999; Correa et al., 2001; Rosenberg et al., 2003). These features probably account for the binding of AdA to the IBC (Rosenberg et al., 2003), but they do not explain the ability of AdA to bind to IP₃R with greater affinity than IP₃. Hitherto, a favoured suggestion is that the 2′-phosphate of AdA, which is thought to mimic the 1-phosphate of IP₃ (Figure 1B), is ‘supra-optimally’ positioned and thereby interacts more strongly with R568 and K569 than does the 1-phosphate of IP₃ (Takahashi et al., 1994c; Wilcox et al., 1995; Hotoda et al., 1999). Alternatively, the adenine of AdA may interact directly with the IP₃R (Hotoda et al., 1999; Glouchankova et al., 2000; Rosenberg et al., 2003). Such an interaction would need to be rather tolerant of changes to the adenine group because even substantial modifications to it cause only modest decreases in affinity (Correa et al., 2001; Sureshan et al., 2008). Defining the mechanisms responsible for high-affinity binding of AdA would both provide an important step towards rational development of ligands of the IP₃R with increased affinity, and contribute to resolving the mechanisms whereby IP₃ and AdA can have different effects on Ca²⁺ signalling (Rossi et al., 2010). Here, we have used synthetic analogues of IP₃ and AdA and systematic mutagenesis of the IBC to address the structural basis of the high-affinity binding of AdA to IP₃R.

Methods

Stable expression of mutant IP₃R1 in DT40 cells

Cloning of rat IP₃R1 (without the S1 splice site) into the pENTR1A vector has been reported previously (Rossi et al., 2009). The QuikChange II XL site-directed mutagenesis kit (Stratagene, La Jolla, CA, USA) was used to introduce point mutations into rat IP₃R1 using the primers (5′-3′) forward: TCACAGCAAGACTACCAGAAGAACCAGGAGTAC, and reverse: GTACTCCTGGTTCTTCTGGTAGTCTTGCTGTGA for R568Q, and forward: TTCTCTAAGCCCAACCAAGAGCGGCAGAAGCTG, and reverse: CAGCTTCTGCCGCTCTTGGTTGGGCTTAGAGAA for R504Q. The sequences of all mutant constructs were confirmed by sequencing of the full-length IP₃R. Mutated IP₃R were subcloned into the expression vector, pcDNA3.2/V5-DEST, by recombination (Invitrogen, Paisley, UK). DT40 cells stably expressing IP₃R1 and its mutants were generated by transfection of cells lacking endogenous IP₃R (Sugawara et al., 1997). DT40 cells were cultured in RPMI 1640 medium supplemented with fetal bovine serum (10%), heat-inactivated chicken serum (1%), 2-mercaptoethanol (50 µM) and glutamine (2 mM). Cells were grown in suspension at 37°C in an atmosphere of 95% air and 5% CO₂, and passaged or used for Ca²⁺ experiments when they reached a density of ∼2 × 10⁶ cells·mL⁻¹ (Tovey et al., 2006). IP₃R expression was quantified by immunoblotting using an antiserum (Ab1.5) to a peptide corresponding to residues 2733–2749 of rat IP₃R1.

Mutagenesis of N-terminal fragments of IP₃R1

N-terminal fragments of IP₃R1 (IBC, residues 224–604; NT, residues 1–604) were amplified by PCR from the rat IP₃R1 clone lacking the S1 splice site and ligated into pTrcHisA vectors for expression of N-terminally tagged His₆ fusion proteins as previously described (Rossi et al., 2009). The IBC included the S1 splice site, but the NT lacked it. The presence of the S1 splice site does not affect the equilibrium dissociation constant (K_d) of the IBC for IP₃ (data not shown). All fragments are numbered by reference to the full-length (S1⁺) rat IP₃R1 (GenBank accession number: GQ233032.1). The QuikChange II XL site-directed mutagenesis kit was used to introduce point mutations into the IBC and NT constructs in pTrcHisA vectors using the primers listed in the preceding section. The sequences of all constructs were confirmed.

Expression of fragments of IP₃R1

His₆-tagged IBC and NT fragments were expressed as described previously (Rossi et al., 2009). Briefly, constructs were transformed into E. coli strain BL21(DE3). Cells were grown in Luria-Bertani medium containing ampicillin (100 µg·mL⁻¹) at 22°C until the OD₆₀₀ reached 1.0–1.5. The culture was then induced by addition of isopropyl β-d-thiogalactoside (0.5 mM), and after 20 h at 15°C, cells were harvested and lysates were prepared in Tris/EDTA medium (TEM: 50 mM Tris, 1 mM EDTA, pH 8.3) as described (Rossi et al., 2009). Expression was detected by immunoblotting using an anti-His₆ antibody (Sigma, Poole, Dorset, UK). Proteins were cleaved from the His₆ tags by incubating bacterial lysate (6 h, 4°C) with thrombin (43 units·mg⁻¹ bacterial protein) in phosphate-buffered saline. Cleavage was monitored by immunoblotting (Rossi et al., 2009) using anti-His₆ antibody and antisera raised to peptides corresponding to residues 62–75 (Ab1) (Cardy et al., 1997) or 326–343 (the SI splice site, Ab1.1) of IP₃R1 (Rossi et al., 2009) for the NT and IBC respectively.

Purification of IP₃R from rat cerebellum

All animal care and experimental procedures complied with UK Home Office policy and with local animal regulations. Adult male Wistar rats were humanely killed by cervical dislocation and cerebella were removed, rapidly frozen in liquid nitrogen and stored at −80°C. IP₃R1 was purified from cerebella using heparin-affinity chromatography following a published protocol (Jiang et al., 2002) with some modifications (Rossi et al., 2009). Briefly, cerebella (2 g) were homogenized in homogenization medium [30 mL, HM: 1 M NaCl, 1 mM EDTA, 50 mM Tris, 1 mM benzamidine, Roche protease inhibitor cocktail (1 tablet per 25 mL), pH 8.3], and then centrifuged (100 000× g, 30 min). The pellet was solubilized in 20 mL of HM without NaCl, but supplemented with CHAPS (1.2%). After centrifugation (100 000× g, 1 h), the NaCl concentration of the supernatant was increased to 250 mM, and the supernatant was loaded onto heparin-agarose beads (5 mL). After 30 min, the beads were washed twice in glycerol-containing medium [250 mM NaCl, 50 mM Tris, 10% glycerol, 1 mM 2-mercaptoethanol, 1 mM benzamidine, 1 mM EGTA, 1% CHAPS, Roche protease inhibitor cocktail (1 tablet per 50 mL), pH 8.0]. IP₃R were eluted with elution medium (500 mM NaCl, 50 mM Tris, 10% glycerol, 1 mM 2-mercaptoethanol, 1 mM benzamidine, 1 mM EGTA, 50 mM Tris, 1% CHAPS, pH 8.0). Samples (∼100 µg protein per mL) were frozen in liquid nitrogen and stored at −80°C.

³H-IP₃ binding

Equilibrium-competition binding assays were performed as described (Rossi et al., 2009). Briefly, incubations (500 µL) at 4°C were in either TEM or cytosol-like medium [CLM: 20 mM NaCl, 140 mM KCl, 1 mM EGTA, 20 mM PIPES, 2 mM MgCl₂, 375 µM CaCl₂ (free [Ca²⁺]= 220 nM), pH 7.0] containing ³H-IP₃ (0.75–3 nM), bacterial lysate (∼1–10 µg protein) or purified IP₃R1 (∼2.5 µg), and competing ligands. For assays using full-length purified IP₃R, all media also included CHAPS (1%). Reactions were terminated after 5 min by addition of poly(ethylene glycol) 8000 (500 µL, 30%, w/v) and γ-globulin (30 µL, 25 mg·mL⁻¹), followed by centrifugation (20 000× g, 5 min). Radioactivity was determined by liquid scintillation counting. Non-specific binding, determined by addition of 10 µM IP₃, or by extrapolation of competition curves to infinite IP₃ concentration, was <10% of total binding.

Ca²⁺ release by IP₃R

A low-affinity Ca²⁺ indicator (Mag-fluo-4) was used to monitor the free [Ca²⁺] within the intracellular Ca²⁺ stores of DT40 cells (Laude et al., 2005; Tovey et al., 2006). DT40 cells stably expressing IP₃R1 or its mutants were centrifuged (650× g, 2 min) and suspended in medium containing 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, 1 mg·mL⁻¹ BSA, 0.4 mg·mL⁻¹ Pluronic F127 and 20 µM Mag-fluo-4 AM. After 1 h at 20°C, cells were suspended in Ca²⁺-free CLM supplemented with saponin (10 µg·mL⁻¹) to allow selective permeabilization of the plasma membrane. Permeabilized cells were centrifuged (650× g, 2 min), re-suspended in CLM without Mg²⁺, but supplemented with 10 µM carbonyl cyanide p-trifluoromethoxyphenylhydrazone (FCCP) to inhibit mitochondria, and 375 µM CaCl₂ to give a final free [Ca²⁺] of ∼220 nM after addition of 1.5 mM MgATP. Cells (∼5 × 10⁵ cells per well) were attached to poly-L-lysine-coated, 96-well, black-walled plates (Greiner, Stonehouse, UK). Fluorescence was recorded at 20°C using a FlexStation III plate reader (MDS Analytical Technologies, Woking, Berks, UK) with excitation and emission wavelengths of 485 nm and 520 nm respectively. MgATP (1.5 mM) was added to initiate Ca²⁺ uptake, and when the endoplasmic reticulum had loaded to steady state with Ca²⁺, IP₃, AdA or their analogues were added. Ca²⁺ release is expressed as a fraction of the ATP-dependent uptake (Tovey et al., 2006).

Data analysis

Equilibrium binding results and concentration–effect relationships were fitted to Hill equations (GraphPad Prism, version 5) from which the Hill coefficients (n_Hill), –logIC₅₀ (pIC₅₀) and –logEC₅₀ (pEC₅₀) values were obtained. For equilibrium-competition binding assays, pK_d values were calculated using the Cheng and Prusoff equation (Cheng and Prusoff, 1973). Because pEC₅₀ and pK_d values are normally distributed, these results are presented as means ± SEM from n independent experiments. For comparisons of the ratios between mean values (EC₅₀ or K_d), statistical analyses compared the differences between their log values (ΔpEC₅₀ or ΔpK_d) (Colquhoun, 1971) with the SEM calculated as follows, assuming that the population variances are the same (confirmed using an F-test) (Ott and Longnecker, 2010):

where, s_p. is the estimate of the population variance:

where, s₁ and s₂ are the sample standard deviations, and n₁ and n₂ are the sample sizes.

Although all analyses were performed using log values, for greater clarity we present some ratios as the antilogs of the means ± SEM.

Statistical analysis used anova followed by Bonferroni test for selected pairs, or unpaired Student's t-tests (GraphPad Prism, version 5). P < 0.05 was considered significant.

Materials

Protease inhibitor cocktail was from Roche (Burgess Hill, Sussex, UK). Heparin-agarose beads and sera were from Sigma (Poole, Dorset, UK). Thrombin was from GE Healthcare (Little Chalfont, Bucks, UK). CHAPS (3-[3-(cholamidopropyl)dimethylammonio]-1-propane-sulphonate) was from Helford Laboratories (Suffolk, UK). RPMI 1640 medium, Pluronic F127 and Mag-fluo-4 AM were from Invitrogen (Paisley, Scotland). ³H-IP₃ (681 GBq·mmol^-1) was from PerkinElmer (Bucks, UK). IP₃ was from Alexis Biochemicals (Nottingham, UK). AdA (Borissow et al., 2005), 2′-dephospho-AdA (Sureshan et al., 2009), furanophostin (Marwood et al., 1999) and ribophostin (Jenkins et al., 1997) were synthesized as previously described. Inositol 4,5-bisphosphate (IP₂) was synthesized by hydrogenolytic deprotection of 1d-2,3,6-tri-O-benzyl-4,5-bis(dibenzyloxyphosphoryl) myo-inositol (Desai et al., 1994). All ligands were purified by ion-exchange chromatography, fully characterized by the usual spectroscopic methods and accurately quantified by total phosphate assay. The structures of the ligands used are shown in Figure 1B. Sources of other reagents either are specified elsewhere in the methods or were previously reported (Rossi et al., 2009).

Results

Stimulation of IP₃R1 by AdA

We used full-length IP₃R1 purified from rat cerebellum for binding assays, and DT40 cells expressing only recombinant IP₃R1 to measure Ca²⁺ release from intracellular stores. Most published analyses of ³H-IP₃ binding use media similar to TEM because its high pH and/or low ionic strength reduce the K_d of IP₃R for IP₃, thereby increasing the specific binding determined with low concentrations of ³H-IP₃. At the densities that recombinant full-length IP₃R are expressed, it is impracticable to measure ³H-IP₃ binding in CLM, although it is feasible with the bacterially expressed fragments of IP₃R. To allow comparison with published work (Hirota et al., 1995; Hotoda et al., 1999; Glouchankova et al., 2000; Rossi et al., 2009) and to provide a direct comparison with our analyses of binding to IP₃R fragments, we first examined IP₃ and AdA binding to IP₃R1 in TEM.

In both binding (in TEM) and functional analyses (in CLM), AdA was ∼12- to 19-fold more potent than IP₃ (ΔpK_d= 1.27 ± 0.09 and ΔpEC₅₀ 1.09 ± 0.05) (Figure 2A and B, Table 1, Table S1). These results are consistent with many previous studies (Hirota et al., 1995; Shuto et al., 1998; Correa et al., 2001; Morris et al., 2002). We note, however, that in one series of studies (Takahashi et al., 1994a,b,c;), the K_d values for IP₃ and AdA were incorrectly calculated from the IC₅₀. The correct K_d for IP₃ and AdA calculated from the data provided are 13 nM and 0.73 nM, respectively, suggesting that in these studies too, AdA bound with about 18-fold greater affinity than IP₃, rather than the stated 100-fold difference.

Figure 2.

Interactions of AdA, IP₃, and their dephospho analogues with IP₃R and its N-terminal (NT) fragments. Equilibrium-competition binding to purified IP₃R1 using ³H-IP₃ (1.5 nM) and the indicated ligands in TEM (A). Ca²⁺ release from permeabilized DT40-IP₃R1 cells evoked by the indicated ligands (B). Equilibrium-competition binding to the NT using ³H-IP₃ (1.5 nM) and the indicated ligands in TEM (C). Equilibrium-competition binding to the IBC using ³H-IP₃ (0.75 nM) and the indicated ligands in CLM (D). Equilibrium-competition binding to the NT using ³H-IP₃ (1.5 nM) and the indicated ligands in CLM (E). The key to the symbols shown in panel A applies to all five panels (A–E). For each analysis (A–E) the K_d (from binding) or EC₅₀ (from functional assays) is shown as a ratio for IP₃ versus AdA (F). For each analysis (A–E), the K_d or EC₅₀ is shown as a ratio for the dephospho analogue relative to IP₃ or AdA (G). Results are means ± SEM, n≥ 4. DT40-IP₃R1 cells, DT40 cells stably expressing rat type 1 IP₃R. AdA, adenophostin A; CLM, cytosol-like medium; IP₃, inositol 1,4,5-trisphosphate; IP₃R, IP₃ receptor; K_d, equilibrium dissociation constant; TEM, Tris/EDTA medium.

Table 1.

Responses of IP₃R1 to IP₃, AdA and their analogues

	Ca2+ release EC50 (nM) (pEC50 ± SEM) nHill± SEM	Release (%)	Binding Kd (nM) (pKd± SEM) nHill± SEM
IP3	38.0	79 ± 3	8.5
	(7.42 ± 0.02)		(8.07 ± 0.03)
	1.1 ± 0.2		1.0 ± 0.1
IP2	5012	73 ± 7	2394
	(5.30 ± 0.08)		(5.62 ± 0.08)
	1.4 ± 0.2		0.8 ± 0.1
AdA	3.1	78 ± 3	0.46
	(8.51 ± 0.05)		(9.34 ± 0.07)
	1.4 ± 0.1		1.5 ± 0.1
2′-dephospho-AdA	160	72 ± 3	18.5
	(6.80 ± 0.02)		(7.73 ± 0.09)
	1.3 ± 0.2		0.7 ± 0.1

From experiments similar to those shown in Figure 2A and B, the effects of each analogue on Ca²⁺ release from the intracellular stores of permeabilized DT40-IP₃R1 cells and on ³H-IP₃ binding to full-length IP₃R1 (in TEM) are summarized. Mean EC₅₀ and K_d values are shown together with means ± SEM for pEC₅₀, pK_d, Hill coefficients (n_Hill) and the percentage Ca²⁺ release. Results are from at least four independent experiments, with each Ca²⁺ release assay performed with three determinations.

AdA, adenophostin A; DT40-IP₃R1 cells, DT40 cells stably expressing rat type 1 IP₃R; IP₂, inositol 4,5-bisphosphate; IP₃, inositol 1,4,5-trisphosphate; IP₃R, IP₃ receptor; K_d, equilibrium dissociation constant; n_Hill, Hill coefficient; TEM, Tris/EDTA medium.

Inositol 1,4,5-trisphosphate binding is entirely mediated by residues in the IBC (Bosanac et al., 2002) (Figure 1A). We therefore compared IP₃ and AdA binding to the isolated NT, initially in TEM. The results, consistent with a previous analysis of a slightly shorter NT fragment of IP₃R1 (residues 1–580) (Glouchankova et al., 2000), establish that the NT alone binds AdA with about 15-fold greater affinity (ΔpK_d= 1.18 ± 0.12) than IP₃ (Figure 2C, Table 2 and Table S1). To allow more direct comparisons with functional assays, we compared IP₃ and AdA binding to the isolated IBC and NT in CLM. These results confirm the expected substantial decrease in affinity for IP₃ in CLM (∼20-fold relative to TEM). More importantly, they establish that in CLM the relative affinities for AdA and IP₃ are not significantly different for the IBC and NT (ΔpK_d= 1.37 ± 0.07 for the IBC, and 1.14 ± 0.21 for the NT) (Figure 2D and E, Table 3 and Table S1). Indeed in all our assays, the relative affinities of IP₃ and AdA for binding to the full-length receptor and its fragments, and their relative potencies in functional assays are not significantly different (Figure 2F, Table S1). It is noteworthy that the n_Hill for the interactions of AdA with IP₃R consistently tend to be greater than unity (see Tables 1–4), even when the interactions are with monomeric IBC or NT (Tables 2 and 3). We and others have reported similar observations previously (reviewed in Rossi et al., 2010), although the underlying mechanism is unresolved.

Table 2.

Binding of IP₃, AdA and their analogues to the NT fragment of IP₃R1 and its mutants assayed in TEM

	Kd (nM) (pKd± SEM) nHill± SEM NT	NTR568Q	NTR504Q
IP3	2.14	41.7	35.6
	(8.67 ± 0.13)	(7.38 ± 0.07)	(7.45 ± 0.05)
	0.9 ± 0.1	0.7 ± 0.1	0.8 ± 0.1
IP2	641	974	1250
	(6.19 ± 0.09)	(6.01 ± 0.12)	(5.90 ± 0.11)
	0.8 ± 0.1	0.8 ± 0.1	0.9 ± 0.2
AdA	0.14	2.81	16.5
	(9.85 ± 0.04)	(8.55 ± 0.06)	(7.78 ± 0.06)
	1.0 ± 0.1	1.3 ± 0.2	0.9 ± 0.1
2′-dephospho-AdA	10.5	13.7	345
	(7.98 ± 0.07)	(7.86 ± 0.04)	(6.46 ± 0.08)
	1.0 ± 0.1	0.7 ± 0.1	0.8 ± 0.1
Ribophostin	1.92	ND	40.5
	(8.72 ± 0.18)		(7.39 ± 0.02)
	0.9 ± 0.2		0.9 ± 0.1
Furanophostin	1.19	ND	52.5
	(8.92 ± 0.08)		(7.29 ± 0.03)
	1.0 ± 0.1		1.0 ± 0.1

From equilibrium-competition binding assays, the K_d, pK_d and Hill coefficients (n_Hill) values for each ligand are shown for the NT and NT mutants of IP₃R1 measured in TEM. Parallel experiments with mutant NT fragments in CLM were not practicable because of the low affinity of the interactions. Results are means (K_d) or means ± SEM (pK_d and n_Hill) from 4–14 independent experiments.

AdA, adenophostin A; CLM, cytosol-like medium; IP₂, inositol 4,5-bisphosphate; IP₃, inositol 1,4,5-trisphosphate; IP₃R, IP₃ receptor; K_d, equilibrium dissociation constant; ND, not determined; n_Hill, Hill coefficient; NT, N-terminal; TEM, Tris/EDTA medium.

Table 3.

Binding of IP₃, AdA and their dephospho analogues to the NT, IBC and IBC mutants assayed in CLM

	Kd (nM) (pKd± SEM) nHill± SEM NT	IBC	IBCR568Q	IBCR504Q
IP3	47.0	7.23	271	96.8
	(7.33 ± 0.16)	(8.14 ± 0.05)	(6.57 ± 0.04)	(7.01 ± 0.02)
	0.8 ± 0.2	1.0 ± 0.1	1.6 ± 0.3	1.3 ± 0.2
IP2	3433	432	1077	624
	(5.46 ± 0.09)	(6.37 ± 0.04)	(5.97 ± 0.09)	(6.21 ± 0.06)
	0.8 ± 0.1	0.8 ± 0.1	1.1 ± 0.3	1.2 ± 0.4
AdA	3.44	0.31	9.48	149
	(8.46 ± 0.11)	(9.51 ± 0.03)	(8.02 ± 0.09)	(6.83 ± 0.04)
	1.3 ± 0.2	1.3 ± 0.3	1.3 ± 0.6	1.5 ± 0.2
2′-dephospho-AdA	77.2	8.82	26.4	308
	(7.11 ± 0.08)	(8.05 ± 0.12)	(7.58 ± 0.06)	(6.51 ± 0.08)
	0.9 ± 0.4	0.8 ± 0.1	1.1 ± 0.3	1.8 ± 0.7

From equilibrium-competition binding assays, the K_d, pK_d and Hill coefficient (n_Hill) for each ligand are shown for the NT, IBC, IBC^R568Q and IBC^R504Q fragments of IP₃R1 in CLM. Results are means (K_d) or means ± SEM (pK_d and n_Hill) from 3–6 independent experiments.

AdA, adenophostin A; CLM, cytosol-like medium; IBC, IP₃-binding core; IP₂, inositol 4,5-bisphosphate; IP₃, inositol 1,4,5-trisphosphate; IP₃R, IP₃ receptor; K_d, equilibrium dissociation constant; n_Hill, Hill coefficient; NT, N-terminal.

Table 4.

Functional responses of mutant IP₃R1

	IP3R1 EC50 (µM) (pEC50± SEM) nHill± SEM	Release (%)	IP3R1R568Q EC50 (µM) (pEC50± SEM) nHill± SEM	Release (%)	IP3R1R504Q EC50 (µM) (pEC50± SEM) nHill± SEM	Release (%)
IP3	0.038	79 ± 3	1.19	39 ± 4	0.85	53 ± 6
	(7.42 ± 0.02)		(5.92 ± 0.06)		(6.07 ± 0.04)
	1.1 ± 0.2		1.5 ± 0.3		1.0 ± 0.3
IP2	5.01	73 ± 7	8.05	53 ± 5	26	57 ± 6
	(5.30 ± 0.08)		(5.09 ± 0.07)		(4.58 ± 0.14)
	1.4 ± 0.2		1.1 ± 0.2		0.9 ± 0.2
AdA	0.003	78 ± 3	0.02	53 ± 3	0.60	61 ± 3
	(8.51 ± 0.05)		(7.70 ± 0.05)		(6.219 ± 0.002)
	1.4 ± 0.1		1.0 ± 0.2		1.3 ± 0.2
2′-dephospho-AdA	0.16	72 ± 3	0.26	43 ± 4	4.21	46 ± 1
	(6.80 ± 0.02)		(6.59 ± 0.07)		(5.38 ± 0.05)
	1.3 ± 0.2		1.7 ± 0.4		1.9 ± 0.4

From experiments similar to those shown in Figures 2B, 3E and 4F, Ca²⁺ release was measured in DT40 cells expressing only the indicated mutant IP₃R. Results show the pEC_50, EC₅₀, n_Hill and the maximal Ca²⁺ release evoked by each agonist. Results are presented as means (EC₅₀) or means ± SEM (pEC_50,n_Hill and percentage Ca²⁺ release) from 4–6 independent experiments, each with three determinations.

AdA, adenophostin A; IP₂, inositol 4,5-bisphosphate; IP₃, inositol 1,4,5-trisphosphate; IP₃R, IP₃ receptor; n_Hill, Hill coefficient.

These results, which establish that the IBC includes the structural determinants for high-affinity binding of AdA, provide our justification for using the IBC in subsequent experiments to explore the structural determinants of AdA binding. With the exception of Gly-268, the residues within the IBC that coordinate IP₃ are conserved between all three IP₃R subtypes (Bosanac et al., 2002), and the IBC from each subtype binds IP₃ with the same affinity (Iwai et al., 2006). This, together with evidence that AdA is more potent than IP₃ in cells that predominantly express each IP₃ receptor subtype (Rossi et al., 2010), suggests that the results obtained from our subsequent analysis of IP₃R1 and the IBC from IP₃R1 are probably applicable also to IP₃R2 and IP₃R3.

The 2′-phosphate is not responsible for high-affinity binding of AdA

The suggestion that the 2′-phosphate of AdA provides a supra-optimal mimic of the 1-phosphate of IP₃ (Takahashi et al., 1994c; Wilcox et al., 1995; Hotoda et al., 1999) predicts that removal of each phosphate moiety should more profoundly reduce the affinity of AdA for IP₃R relative to IP₃. We tested this prediction using synthetic IP₂ and 2′-dephospho-AdA (Figure 1B). In equilibrium-competition binding analyses with full-length IP₃R1 (in TEM), IP₂ bound with 282-fold lower affinity than IP₃ (ΔpK_d= 2.45 ± 0.10), whereas loss of the 2′-phosphate from AdA caused only a 41-fold decrease in affinity (ΔpK_d= 1.61 ± 0.11) (Figure 2A, Table 1 and Table S1). For the IBC in CLM, removal of the critical phosphate also more substantially reduced the affinity for IP₃ relative to AdA: IP₃ bound with 60-fold higher affinity than IP₂ (ΔpK_d= 1.78 ± 0.08), whereas the affinities of 2′-dephospho-AdA and AdA differed by only 29-fold (ΔpK_d= 1.46 ± 0.12) (Figure 2D, Table 3 and Table S1). Similar results were obtained with the NT in TEM and CLM (Figure 2C and E, Tables 2 and 3 and Table S1). In Ca²⁺ release assays, and consistent with the binding analyses, removal of the 1-phosphate from IP₃ more substantially reduced its potency than did removal of the 2′-phosphate from AdA (Figure 2B, Table 1 and Table S1). A previous study suggested that loss of the 2′-phosphate of AdA more profoundly affected the K_d (Takahashi et al., 1994a). However, as stated above, the authors miscalculated the K_d from the IC₅₀, and it is impossible from the data presented to estimate the correct K_d for 2′-dephospho-AdA. We note, although it is unclear whether it contributes to their reported low affinity of 2′-dephospho-AdA for IP₃R, that these authors used 2′-dephospho-AdA produced enzymatically rather than by synthesis (Takahashi et al., 1994a).

It is noteworthy that the disparity between the affinities of IP₃ and AdA and their dephospho analogues was exaggerated in TEM (Figure 2G). We have not further explored the more pronounced effect of TEM on binding of AdA and IP₃ relative to 2′-dephospho-AdA and IP₂ (Table 2). The high pH of TEM favours substantial deprotonation of the phosphate groups in all the ligands (Felemez et al., 1999), perhaps thereby enhancing their binding to the IBC. The larger effect of TEM on binding of the trisphosphate ligands (IP₃ and AdA) may reflect a greater effect of the different media (pH, ionic strength, counter-ions) on the ionization states of these ligands relative to the bisphosphate ligands. The results do, however, highlight the necessity to examine ligand binding to IP₃R in medium resembling that used for functional analyses (e.g. CLM) if the functional and binding analyses are to be compared reliably.

Our demonstration that removal of the 1-phosphate from IP₃ reduces both affinity and potency significantly more than does removal of the 2′-phosphate from AdA (Figure 2G) is inconsistent with the notion that the high affinity of AdA results from its 2′-phosphate providing a supra-optimal mimic of the 1-phosphate of IP₃ (Takahashi et al., 1994c).

Contributions of R568 to AdA and IP₃ binding

R568 within the α-domain of the IBC is the only residue to interact directly with the 1-phosphate of IP₃, and it does so via two H-bonds with its side chain (Figure 1A) (Bosanac et al., 2002). We mutated R568 to Q and examined IP₃ and AdA binding to the mutant IBC, and Ca²⁺ release via the mutant full-length IP₃R.

The affinities of IP₃ and AdA for the IBC were similarly reduced (by 37-fold and 31-fold respectively) by the R568Q mutation (ΔpK_d= 1.57 ± 0.07 for IP₃, and 1.49 ± 0.11 for AdA), whereas the affinities of IP₂ and 2′-dephospho-AdA were minimally affected (ΔpK_d= 0.40 ± 0.14 and 0.48 ± 0.12 for IP₂ and 2′-dephospho-AdA respectively) (Figure 3A and B, Table 3, Table S2). Similar results were obtained with NT^R568Q in TEM (Figure 3C, Table 2). These results suggest that R568 recognizes the 1-phosphate of IP₃ and the 2′-phosphate of AdA similarly, and so lends further support to the view that the latter at least partially mimics the 1-phosphate of IP₃ (Figure 1B). But why, if IP₃ and AdA interact similarly with R568, should removal of the 1-phosphate from IP₃ increase its K_d more than does removal of the 2′-phosphate from AdA (Tables 1 and 2)? A likely explanation is that the indirect interaction of the 1-phosphate of IP₃ with the backbone of K569 (Figure 1A) (Bosanac et al., 2002) is stronger than the equivalent interaction with the 2′-phosphate of AdA. Unfortunately, mutagenesis using naturally occurring amino acids cannot be used to dissect the role of this backbone interaction. Introduction of a non-natural amino acid, for example replacing K569 with an α-hydroxyl acid residue to replace the backbone NH with O, might address the question (Yang et al., 2004). But the nonsense suppression techniques required to insert the non-natural residue are not presently available to us.

Figure 3.

R568 does not selectively enhance AdA binding. Equilibrium-competition binding to IBC^R568Q using ³H-IP₃ (1.5 nM) and the indicated ligands in CLM (A). Relative affinities (K_d) of ligands for the IBC^R568Q and IBC (B). Equilibrium-competition binding to the NT^R568Q using ³H-IP₃ (1.5 nM) and the indicated ligands in TEM (C). Representative immunoblot (with anti-IP₃R1 antibody, Ab1.5, top panel; and β-adaptin, bottom panel) for DT40-KO cells (KO) and DT40 cells expressing IP₃R1 or the indicated mutants (10⁵ cells per lane). Molecular weight markers (kDa) are shown. The blot is typical of six similar blots. IP₃R expression (corrected for β-adaptin loading) is shown for each mutant relative to DT40-IP₃R1 cells (%, means ± SEM) (D). Ca²⁺ release from permeabilized DT40-IP₃R1^R568Q cells evoked by the indicated ligands (E). Comparison of Ca²⁺ release for each ligand in normal and mutant IP₃R1^R568Q (F). Results are means ± SEM, n≥ 4. AdA, adenophostin A; CLM, cytosol-like medium; IBC, IP₃-binding core; IP₃, inositol 1,4,5-trisphosphate; IP₃R, IP₃ receptor; K_d, equilibrium dissociation constant; NT, N-terminal; TEM, Tris/EDTA medium.

Functional assays of DT40 cells expressing mutant IP₃R do not allow absolute sensitivities to agonists or the size of the agonist-sensitive Ca²⁺ pool to be precisely compared between cell lines because it is impossible to establish stable cell lines with identical levels of IP₃R expression (Figure 3D). The problem is less severe than might have been anticipated because even a substantial increase in IP₃R expression in DT40 cells (>20-fold) caused the sensitivity to IP₃ to increase by less than twofold (Dellis et al., 2006). The mutations we have studied caused much larger changes in IP₃ sensitivity (∼30-fold, Table 4). Furthermore, levels of IP₃R expression in the stable DT40 cell lines expressing mutant IP₃R differed by no more than 3.5-fold from the line expressing wild-type IP₃R1 (Figure 3D).

In functional assays, using DT40 cells expressing only IP₃R1^R568Q, IP₃ and AdA were less potent than in cells expressing wild-type IP₃R1: the potencies of IP₃ and AdA were reduced by 31-fold (ΔpEC₅₀= 1.50 ± 0.06) and sixfold (ΔpEC₅₀= 0.81 ± 0.07) respectively. By contrast, the potencies of IP₂ and 2′-dephospho-AdA were only very modestly reduced (1.6-fold, ΔpEC₅₀= 0.21 ± 0.11 for both ligands) (Figure 3E and F, Table 4 and Table S3). The lesser Ca²⁺ release with maximally effective concentrations of all four agonists in DT40-IP₃R^R568Q cells (compare Figures 2B and 3E) is probably attributable to the reduced level of expression of IP₃R in the mutant cell line (Figure 3D). The selective effect of the R568Q mutation on AdA and IP₃, but not the dephospho analogues, is consistent with our analyses of ligand binding (Table 3). But the significantly lesser effect of the mutation on the functional responses to AdA was unexpected because hitherto the interactions between R568 have appeared similar for the 1-phosphate of IP₃ and the 2′-phosphate of AdA. These results suggest that disruption of the interaction between R568 and the critical phosphates may selectively reduce the efficacy of IP₃. That conclusion would be consistent with evidence that inositol 2,4,5-trisphosphate is a partial agonist of IP₃R (Marchant et al., 1997b).

Selective interaction of AdA with R504

Our previous attempts to predict the binding mode of AdA to the IBC using molecular docking suggested that, in addition to possible interactions with R568, the 2′-phosphate might also interact with the amide NH of K569 and the side chain of R504 (Rosenberg et al., 2003). But our present results (Figure 2) suggest that for AdA, the 2′-phosphate is not a major determinant of its high affinity. One of the possible binding modes also suggested a cation-π interaction between the adenine of AdA and the guanidinium side chain of R504 (Rosenberg et al., 2003) (Figure 4A). Many analyses of synthetic AdA analogues lacking the adenine moiety suggest that the adenine or another aromatic moiety is an important determinant of the high-affinity binding of AdA. The AdA analogues lacking adenine, which retain a phosphate group equivalent to the 2′-phosphate of AdA, typically have K_d values similar to that for IP₃ (Table 2) (Jenkins et al., 1997; Tatani et al., 1998; Hotoda et al., 1999; Marwood et al., 2000; Correa et al., 2001). R504 is one of several residues to form a hydrogen bond with the 5-phosphate of IP₃, via a bridging water molecule (Bosanac et al., 2002), and probably also with the equivalent 3"-phosphate of AdA (Rosenberg et al., 2003) (Figure 1B). Mutation of R504 inhibits IP₃ binding (Yoshikawa et al., 1996) and is likewise expected to disrupt interaction of the 3"-phosphate of AdA. But if the proposed cation-π interaction is important for AdA binding, mutation of R504 might additionally reduce AdA binding by disrupting interactions with its adenine ring. The subsequent experiments were designed to test this hypothesis.

Figure 4.

Selective interaction of AdA with R504. Model of AdA binding to the IBC highlighting a possible cation-π interaction between the adenine ring of AdA and R504 of the IBC (Rosenberg et al., 2003) (A). Equilibrium-competition binding to IBC^R504Q using ³H-IP₃ (3 nM) and the indicated ligands in CLM (B). Comparison of the binding of each ligand to normal and mutant IBC (C). Equilibrium-competition binding to NT^R504Q using ³H-IP₃ (1.5 nM) and the indicated ligands in TEM (D). Structure of the IBC (PDB 1N4K) highlighting the likely interactions of R504 with the phosphate groups of IP₃ (see text for further details). The red spheres represent water (E). Ca²⁺ release from permeabilized IP₃R1^R504Q cells evoked by the indicated ligands (F). Comparison of Ca²⁺ release for each ligand via normal and mutant IP₃R (G). Results (B–D and F–G) are means ± SEM, n≥ 4. AdA, adenophostin A; CLM, cytosol-like medium; IBC, IP₃-binding core; IP₃, inositol 1,4,5-trisphosphate; IP₃R, IP₃ receptor; NT, N-terminal; TEM, Tris/EDTA medium.

As expected, because each ligand has a phosphate equivalent to the 5-phosphate of IP₃, mutation of R504 to Q significantly reduced the affinity of both IP₃ and AdA for the IBC (Figure 4B and C, Table 3 and Table S3). However, whereas the affinity of IP₃ was reduced by 13-fold (ΔpK_d= 1.13 ± 0.06), the affinity of AdA was reduced by 353-fold (ΔpK_d= 2.55 ± 0.05) (Figure 4B and C, Table 3 and Table S2). Binding of the dephospho analogues was less affected by the R504Q mutation: the decrease in affinity was 1.4-fold (pΔ= 0.16 ± 0.07) for IP₂ and 35-fold (ΔpK_d= 1.54 ± 0.14) for 2′-dephospho-AdA. Similar results were obtained with NT^R504Q in TEM (Figure 4D, Table 2).

A further prediction arising from our suggestion that a cation-π interaction involving R504 and the adenine moiety contributes to high-affinity binding of AdA is that adenophostin analogues lacking the adenine moiety should be less affected by mutation of R504. The results shown in Table 2 confirm this prediction for two such analogues, furanophostin and ribophostin (Figure 1B). Mutation of R504 to Q reduced the affinity of the NT for furanophostin by 44-fold (ΔpK_d= 1.63 ± 0.10) and for ribophostin by 21-fold (ΔpK_d= 1.33 ± 0.24).

Comparison of the effects of the R504Q mutation on each pair of ligands (i.e. IP₃ vs. AdA, and IP₂ vs. dephospho-AdA) indicates that for each there was a 25-fold greater decrease in the affinity for the AdA analogues (Figure 4C, Table 3 and Table S2). But the lesser effects of the mutation on both dephospho analogues is intriguing because R504 has not been reported to interact with the 1-phosphate of IP₃ (Bosanac et al., 2002). However, close inspection of the crystal structure of the IP₃-bound IBC also reveals a possible indirect interaction, via water, of R504 with the 1-phosphate (Figure 4E), although this was not discussed in the original report (Bosanac et al., 2002). It is not clear whether a similar interaction could also exist for AdA, but the possibility that both IP₃ and AdA interact with R504 via their 1- and 2′-phosphates, respectively, is appealing because it would explain the lesser effects of the R504Q mutation on the dephospho analogues (Figure 4C).

Functional analyses with DT40 cells expressing IP₃R1^R504Q confirmed the results with binding. The sensitivity of the mutant IP₃R1^R504Q was significantly decreased for all ligands, although again the effect on the dephospho analogues was more modest than that on IP₃ and AdA (Figure 4F and G, Table 4 and Table S3). Most importantly, whereas the EC₅₀ for IP₃ was reduced by 23-fold (ΔpEC₅₀= 1.35 ± 0.17), that for AdA was reduced by 196-fold (ΔpEC₅₀= 2.29 ± 0.06). These ΔpEC₅₀ values for IP₃ and AdA are significantly different. Together, the binding and functional analyses establish that R504 is more important for binding of AdA than for IP₃. The selective effect of R504 on AdA binding does not result from interaction with the 2′-phosphate of AdA, but is instead likely to reflect the contribution of a cation-π interaction between the adenine of AdA and the guanidinium side chain of R504 (Figure 4A).

Discussion

Inositol 1,4,5-trisphosphate and AdA are full agonists of IP₃R (Rossi et al., 2009), but the latter binds with substantially greater affinity than does IP₃ (Takahashi et al., 1994b; Correa et al., 2001) (Figure 2F). We have established that the IBC, to which IP₃ binds to initiate activation of the IP₃R, is alone capable of binding AdA with ∼20-fold greater affinity than IP₃ (Table 3). Presently available evidence suggests that once AdA or IP₃ has bound to the IBC, each causes indistinguishable activation of the IP₃R (Rossi et al., 2009). The increased potency of AdA, relative to IP₃, in evoking Ca²⁺ release via IP₃R is therefore likely to be entirely attributable to the stronger interactions between AdA and the IBC. This conclusion provided the impetus for resolving the interactions between AdA and the IBC that mediate its high-affinity binding.

The 4"- and 3"-phosphates of AdA mimic the essential 4,5-bisphosphate moiety of IP₃ (Figure 1B) and the 2′-phosphate of AdA has been thought to at least partially mimic the 1-phosphate of IP₃ (Takahashi et al., 1994b). None of our present results challenges this interpretation. Hitherto, an appealing explanation for the high affinity of AdA has been the suggestion that its 2′-phosphate is better placed than the 1-phosphate of IP₃ to interact with residues in the IBC (Takahashi et al., 1994c; Wilcox et al., 1995). Our results establish that this is not the basis of the high-affinity binding of AdA. First, removal of the 2′-phosphate from AdA has significantly less effect on its activity than does removal of the 1-phosphate from IP₃ (Figure 2G, Tables 1–3 and Table S1). Second, mutation of R568, one of the key residues with which the 1-phosphate of IP₃ interacts (Bosanac et al., 2002), similarly reduces the affinity of the IP₃R for IP₃ and AdA, while minimally affecting binding of IP₂ or 2′-dephospho-AdA (Figure 3A and B, Table 3 and Table S1). These results establish that the 2′-phosphate of AdA partially mimics the 1-phosphate of IP₃, but the latter probably interacts more strongly with the backbone of K569. We conclude that the high affinity of AdA for IP₃R does not result from its 2′-phosphate behaving as a supra-optimally positioned mimic of the 1-phosphate of IP₃ (Figure 5).

Figure 5.

Differential interactions of IP₃ and AdA with the IBC. Interactions of the phosphate groups of IP₃ with the residues mutated in this study are highlighted (Bosanac et al., 2002). Predicted interactions of the phosphate groups of AdA with the same residues, and the proposed cation-π interaction between the adenine moiety and R504 are shown. AdA, adenophostin A; IBC, IP₃-binding core; IP₃, inositol 1,4,5-trisphosphate.

Considerable evidence suggests that the adenine group of AdA contributes significantly to its high-affinity binding (Tatani et al., 1998; Hotoda et al., 1999; Marwood et al., 2000; Correa et al., 2001; Rosenberg et al., 2003) and our molecular modelling has suggested that this might result from a cation-π interaction between the adenine and the guanidinium side chain of R504 (Rosenberg et al., 2003) (Figure 4A). Our present results are consistent with this suggestion. Because R504 interacts with the 5-phosphate of IP₃ (Figure 1A) and, almost certainly, in similar fashion with the equivalent 3"-phosphate of AdA (Figures 1B and 5), mutation of this residue (R504Q) significantly decreased the affinity of the IP₃R for both ligands. But more importantly, the effects, in both functional and binding assays, were significantly greater for AdA than for IP₃ (Figure 4D and G, Tables 2–4, Tables S2 and S3). These results establish the greater importance of R504 for AdA binding, which would be consistent with AdA, but not IP₃, forming a cation-π interaction with this residue (Figure 5). We conclude that the high affinity of AdA for IP₃R is not due to its 2′-phosphate, and that AdA interacts more strongly than IP₃ with R504, most likely reflecting a cation-π interaction between the adenine group and R504. This interaction may provide opportunities for synthesis of less polar ligands of IP₃R (Sureshan et al., 2009).

Glossary

Abbreviations

AdA

adenophostin A

CLM

cytosol-like medium

IBC

IP₃-binding core

IP₂

inositol 4,5-bisphosphate

IP₃

inositol 1,4,5-trisphosphate

IP₃R

IP₃ receptor

K_d

equilibrium dissociation constant

n_Hill

Hill coefficient

NT

N-terminal

TEM

Tris/EDTA medium

Conflicts of interest

None.

Supporting information

Additional Supporting Information may be found in the online version of this article:

Table S1 Relative potencies and affinities of IP₃ and AdA analogues interacting with native IP3R1 and its N-terminal fragments. The table shows the relative effectiveness for each pair of ligands in evoking Ca²⁺ release from DT40-IP3R1 cells (ΔpEC₅₀) and in binding (in TEM) to full-length IP₃R1 and NT, and binding (in CLM) to the NT or IBC (ΔpK_d). Results are shown as means ± SEM, from at least three independent experiments. The data from which these values are derived are shown in Figure 2.

Table S2 Relative affinities of IP₃ and AdA analogues interacting with wild-type and mutant IBC. For each ligand, the relative affinity (ΔpK_d) is shown for mutant and wild-type IBC (in CLM). Results are shown as means ± SEM, from at least three independent experiments. The data from which these values are derived are shown in Figures 2, 3 and 4.

Table S3 Relative potencies of IP₃ and AdA analogues interacting with mutant IP₃R1. For each ligand the relative potency in evoking Ca²⁺ release (ΔpEC₅₀) is shown for wild-type and mutant IP3R. Results are shown as means ± SEM, from at least three independent experiments. The data from which these values are derived are shown in Figures 2, 3 and 4.

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.