Activation of IP3 receptors requires an endogenous 1-8-14 calmodulin-binding motif

Binding of IP3 (inositol 1,4,5-trisphosphate) to the IP3-binding core (residues 224–604) of IP3Rs (IP3 receptors) initiates opening of these ubiquitous intracellular Ca2+ channels. The mechanisms are unresolved, but require conformational changes to pass through the suppressor domain (residues 1–223). A calmodulin-binding peptide derived from myosin light chain kinase uncouples these events. We identified a similar conserved 1-8-14 calmodulin-binding motif within the suppressor domain of IP3R1 and, using peptides and mutagenesis, we demonstrate that it is essential for IP3R activation, whether assessed by IP3-evoked Ca2+ release or patch-clamp recoding of nuclear IP3R. Mimetic peptides specifically inhibit activation of IP3R by uncoupling the IP3-binding core from the suppressor domain. Mutations of key hydrophobic residues within the endogenous 1-8-14 motif mimic the peptides. Our results show that an endogenous 1-8-14 motif mediates conformational changes that are essential for IP3R activation. The inhibitory effects of calmodulin and related proteins may result from disruption of this essential interaction.


INTRODUCTION
Ca 2 + channels allow most electrical and many chemical signals to be transduced into the changes in cytosolic Ca 2 + concentration that regulate almost every aspect of cellular activity [1]. Most Ca 2 + channels are also regulated by Ca 2 + , either directly or via CaM (calmodulin) [2]. This provides feedback regulation of Ca 2 + signalling and it allows Ca 2 + channels to evoke regenerative Ca 2 + signals [3]. The latter are important because they underpin the versatility of Ca 2 + as an intracellular messenger, permitting it to function either locally or globally [1].
All IP 3 Rs are inhibited by Ca 2 + -CaM [22], but neither of the two CaM-binding sites within IP 3 R1, nor a third that is created by alternative splicing [23], clearly mediates this inhibition of IP 3evoked Ca 2 + release. The central site [24] ( Figure 1A) mediates neither Ca 2 + nor CaM regulation of IP 3 R activity [16,17] and it is absent from IP 3 R3. The functional role of the split N-terminal site ( Figure 1A), one component of which may also bind CaBP1 [25], is also unclear. It has been proposed to bind CaM and thereby to inhibit IP 3 R activity, but only when Ca 2 + has bound elsewhere [26]. The evidence that CaM inhibits IP 3 R only in the presence of Ca 2 + , without CaM itself providing the Ca 2 + -sensor, is persuasive [26], but there is no compelling evidence to link this to the N-terminal CaM-binding site [27].
The links between CaM binding and function are better understood for RyRs, although the effects differ between RyR subtypes [7]. A single site on each RyR1 subunit (residues 3614-3643 in rabbit RyR1), which is conserved in all RyRs, binds the C-terminal lobe of both apo-CaM and Ca 2 + -CaM and appears to mediate the functional effects of CaM [20,28,29]. As this tethered CaM binds Ca 2 + , it migrates towards the NT (N-terminus) of the binding site and the CaM switches from activating RyR1 to inhibiting it [19]. The CaM-binding site of RyR1 also engages other CaM-like domains, notably the C-terminus of the L-type Ca 2 + channel which inhibits RyR1 activity [30], and perhaps an EF-hand-like structure within the C-terminal region of RyR1 which binds Ca 2 + and modulates Ca 2 + regulation of RyR [31]. These observations suggest that the CaM-binding domain of RyR also mediates important inter-and intra-molecular interactions, and that the complex effects of CaM and related proteins may, at least in part, result from disrupting these interactions [29,31,32]. showing the conserved hydrophobic residues of the consensus sequence in blue. Charged residues within the 1-8-14 motif are highlighted in red because the consensus motif has a net charge of + 3 to + 6. The lower panel shows the peptides used with mutated residues underlined. (C) Structure of the SD of IP 3 R1 (PDB code 1XZZ) and the equivalent region (A domain) of RyR1 (PDB code 3HSM) with the pseudo-1-8-14 motif highlighted and compared with MLCK in the structure it adopts when bound to Ca 2 + -CaM (PDB code 1QTX).
For IP 3 Rs, IP 3 binding to the IBC (IP 3 -binding core) (residues 224-604) ( Figure 1A) initiates the conformational changes that lead to opening of a pore formed by the C-terminal transmembrane domains of each of the four IP 3 R subunits [5,33]. These conformational changes pass via the N-terminal SD (suppressor domain) (residues 1-223), which is essential for IP 3 R activation. Indeed, the major conformational changes associated with IP 3 R activation appear to occur within the NT (residues 1-604) [5,33]. Although both IP 3 and Ca 2 + are required for IP 3 R activation [6,34], it is not yet clear how the conformational changes initiated by IP 3 lead to Ca 2 + binding and then to gating of the pore. It is therefore intriguing that a CaMbinding peptide derived from MLCK (myosin light chain kinase), which comprises a 1-8-14 CaM-binding sequence [35], reversibly inhibits IP 3 -evoked Ca 2 + release [36] via all three vertebrate IP 3 R subtypes. Furthermore, MLCK peptide is more potent in the presence of Ca 2 + [35]. This inhibition is entirely independent of CaM and involves interaction of MLCK peptide with the NT in a manner that requires the SD [35]. We speculate, by analogy with RyRs, that inhibition of IP 3 Rs by MLCK peptide might result from disruption of an interaction between endogenous CaM-like and CaM-binding domains within IP 3 Rs, and that, for IP 3 Rs, this interaction is essential for activation. In the present study, we explored this hypothesis further.

Site-directed mutagenesis
The NT (residues 1-604) and IBC (residues 224-604) of rat IP 3 R1 were amplified by PCR from the full-length receptor clone lacking the SI splice region (GenBank ® accession number GQ233032.1) as described previously [33]. The fragments were ligated into pTrcHis A (Invitrogen) to allow expression of N-terminally His 6 -tagged proteins. Mutagenesis of the 1-8-14 motif within the NT used the QuikChange ® II XL site-directed mutagenesis kit (Stratagene) for single mutants (F53E, L60E, Y66E and K52E) and the QuikChange ® multi-site-directed mutagenesis kit for the double mutant (F53E and Y66E). The primers used are listed in Supplementary Table S2 (at http://www.biochemj.org/bj/449/bj4490039add.htm). The same primers and conditions were used for mutagenesis of full-length IP 3 R using IP 3 R1 in the pENTR 1A vector. Full-length constructs were subcloned into pcDNA3.2/V5-DEST for expression in DT40 cells. The complete sequence of every mutant construct was verified by sequencing.

Culture and stable transfection of DT40 cells
DT40 cells in which the genes for all three IP 3 R subtypes had been disrupted (DT40-KO) [37] and DT40 cells stably expressing rat IP 3 R1 (DT40-IP 3 R1) were grown in RPMI 1640 medium supplemented with 10 % (v/v) fetal bovine serum, 1 % (v/v) heat-inactivated chicken serum, 2 mM L-glutamine and 50 μM 2-mercaptoethanol. Cells were grown in suspension in 175 cm 2 flasks at 37 • C in an atmosphere of 5 % CO 2 . They were used or passaged when they reached a density of ∼2×10 6 cells/ml. To generate stable cell lines expressing mutant IP 3 R, the mutant construct in pcDNA3.2/V5-DEST was linearized, and DT40 cells were transfected by nucleofection (Amaxa, protocol B-23). Cell lines were selected with G-418 (2 mg/ml) and screened initially by Western blotting using a peptide antiserum to IP 3 R1 [38] as described previously [33], and then using the functional assay described below.

Ca 2 + release from the intracellular stores of permeabilized cells
The free Ca 2 + concentration of the intracellular stores of permeabilized cells was measured using a low-affinity Ca 2 + indicator trapped within the endoplasmic reticulum as reported previously [39]. Briefly, DT40 cells (4×10 7 cells/ml) were suspended in HBS (Hepes-buffered saline: 135 mM NaCl, 5.9 mM KCl, 11.6 mM Hepes, 1.5 mM CaCl 2 , 11.5 mM glucose and 1.2 mM MgCl 2 , pH 7.3) containing 1 mg/ml BSA, 0.4 mg/ml Pluronic F127 and 20 μM mag-fluo-4/AM (Invitrogen). After 1 h at 20 • C in the dark with gentle shaking, cells were centrifuged at 650 g for 2 min and resuspended to 10 7 cells/ml in Ca 2 + -free CLM (cytosol-like medium) (20 mM NaCl, 140 mM KCl, 1 mM EGTA, 20 mM Pipes and 2 mM MgCl 2 , pH 7.0) containing 20 μg/ml saponin. After incubation at 37 • C with gentle shaking for 4 min, permeabilized cells were centrifuged at 650 g for 2 min and resuspended in Mg 2 + -free CLM, supplemented with CaCl 2 to give a final free Ca 2 + concentration of 220 nM. The free Ca 2 + concentration of CLM was calculated using the MaxChelator program (http://maxchelator.stanford.edu) and then measured using fluo-3 or fura-2. Cells were then washed, resuspended in Mg 2 + -free CLM containing 10 μM FCCP (carbonyl cyanide p-trifluoromethoxyphenylhydrazone) to inhibit mitochondria, and distributed into a 96-well plate (10 6 cells in 50 μl of CLM/well). After centrifugation, fluorescence from the luminal indicator was recorded using a FlexStation II platereader (Molecular Devices) equipped to allow automated additions [39]. In all experiments, the intracellular stores were allowed to load to steady-state with Ca 2 + after addition of MgATP. IP 3 was then added with thapsigargin (1 μM, to inhibit Ca 2 + reuptake). The Ca 2 + release evoked by IP 3 is expressed as a fraction of the ATP-dependent Ca 2 + uptake.

Patch-clamp recording
Currents were recorded from patches excised from the outer nuclear envelope of DT40 cells expressing recombinant rat IP 3 R1 using symmetrical caesium methanesulfonate (140 mM) as the charge-carrier. The composition of recording solutions and methods of analysis were otherwise as described previously [40].

Expression of N-terminal fragments of IP 3 R
The pTrcHis constructs were used for expression of N-terminally His 6 -tagged proteins in Escherichia coli strain BL21(DE3) cells. Before use for [ 3 H]IP 3 binding, proteins were cleaved from the His 6 tags using biotinylated thrombin (Novagen) at the engineered thrombin-cleavage site [33]. Complete cleavage was verified by Western blotting using an anti-His 6 antibody. The proteins were used for [ 3 H]IP 3 binding without further purification [33].

[ 3 H]IP 3 binding
Equilibrium-competition binding assays were performed at 4 • C for 5 min in CLM (500 μl) with a free Ca 2 + concentration of 220 nM and containing [ 3 H]IP 3 (0.75-1.5 nM), bacterial lysate (10 μg of protein for IBC and 100 μg of protein for NT) or cerebellar membranes (50 μg of protein) and competing ligands. Non-specific binding was defined by addition of 10 μM IP 3 . Bound and free [ 3 H]IP 3 were separated by centrifugation at 20 000 g for 5 min, after addition of poly(ethylene glycol) (15 % final concentration) and γ -globulin (0.75 mg) for soluble proteins. Results were analysed by fitting to a Hill equation (using GraphPad Prism) from which the IC 50 (half-maximal inhibitory concentration) and thereby the K d (equilibrium dissociation constant) were calculated [33].

Statistical analysis
For comparisons of K d , EC 50 (half-maximally effective concentration) or IC 50 values, their negative logarithms (pK d , pEC 50 and pIC 50 ; means + − S.E.M.) were used for statistical analyses. For clarity, some Figures show normalized results, but all statistical analyses were performed on the raw data using paired or unpaired Student's t tests. P < 0.05 was considered significant.

Reversible inhibition of IP 3 -evoked Ca 2 + release by an endogenous 1-8-14 peptide
A sequence within the SD of all known IP 3 Rs (residues 53-66 in rat IP 3 R1; Supplementary Figure S1 at http://www.biochemj. org/bj/449/bj4490039add.htm) includes the critical hydrophobic residues of a 1-8-14 CaM-binding motif appropriately oriented within the known structure of the SD [41] (Figures 1B and 1C) and with the required net positive charge [35]. The sequence lies within one of the two regions (residues 49-81; Figure 1A) within the NT reported to bind CaM [42] and CaBP1 [14]. A similar sequence is present within the N-terminal of all RyRs (Supplementary Figure S1). To test our hypothesis that inhibition of IP 3 R by MLCK peptide results from disruption of an essential interaction involving an endogenous 1-8-14 motif, we assessed the effects of a peptide derived from this motif (1-8-14 peptide; Figure 1B and Supplementary Table S1) on IP 3 -evoked Ca 2 + release.
The results shown in Figure 2(E) demonstrate that the effects of a maximally effective concentration of 1-8-14 peptide (3 mM) are fully reversible. These experiments, which require extensive washing of the cells between successive challenges with the peptide, confirm that the inhibition of IP 3 Rs by the 1-8-14 peptide, like that by MLCK peptide [35], does not result from dissociation of CaM from IP 3 R [36]. Our previous study demonstrated that MLCK peptide more potently inhibited IP 3 R when the cytosolic free Ca 2 + concentration was increased [35]. Similar results were obtained with 1-8-14 peptide ( Figure 2F). We conclude that 1-8-14 peptide inhibits IP 3 -evoked Ca 2 + release by binding to the IP 3 R and the inhibition is enhanced at elevated cytosolic Ca 2 + concentrations.

Inhibition of single-channel currents through IP 3 Rs by 1-8-14 peptide
In patch-clamp recordings from the nuclear envelope of DT40 cells expressing rat IP 3 R1, a maximally effective concentration of IP 3 stimulated IP 3 R activity and this was massively attenuated by the 1-8-14 peptide (3 mM) ( Figures 3A and 3B). Our results are consistent with the peptide causing a 50 % decrease in the mean channel open time (τ o ) ( Figure 3C). However, the overall channel activity (NP o ) was so low under these conditions that we cannot reliably estimate the number of active IP 3 Rs (N) within each patch. We cannot therefore entirely eliminate the possibility that each patch fortuitously included several IP 3 Rs and that their clustering caused τ o to fall from ∼10 ms to ∼5 ms as we reported previously [40]. An effect on τ o would be unusual because most regulators of IP 3 Rs affect the duration of closed states (τ c ) [6,40]. The effect of the peptide on τ o is not, however, sufficient to account for the ∼10-fold decrease in NP o ( Figure 3A), suggesting that the 1-8-14 peptide must also affect the rate of channel opening (i.e. τ c ). Because it was impossible to determine the number of active IP 3 Rs in the presence of 1-8-14 peptide (see above), we could not reliably determine τ c . The single-channel conductance (γ Cs ) was unaffected by 1-8-14 peptide: it was 214 + − 6 pS (n = 3) and 209 + − 6 pS (n = 3) for control and peptide-treated IP 3 Rs respectively ( Figure 3D).
These results establish that a peptide derived from an endogenous 1-8-14 motif within the SD of the IP 3 R is similar to MLCK peptide in causing substantial and reversible inhibition of IP 3 Rs that is independent of CaM. This conclusion is consistent with our suggestion that MLCK peptide inhibits IP 3 Rs by mimicking an endogenous 1-8-14 motif, and so perhaps 'unzipping' an interdomain interaction [43] that is essential for activation of IP 3 Rs.

1-8-14 peptide uncouples IP 3 binding from activation of IP 3 Rs
Removal of the SD increases the affinity of both full-length IP 3 Rs and the NT for IP 3 [33]. We [33] have suggested that this reflects the use of binding energy to drive conformational rearrangement of SD-IBC interfaces during the initial steps of IP 3 R activation [5,44].
1-8-14 peptide (3 mM) increased specific binding of [ 3 H]IP 3 to full-length IP 3 R1. Similar results were obtained with the NT, but IP 3 binding to the IBC was unaffected ( Figure 4A). The latter demonstrates that 1-8-14 peptide does not interact directly with either the IP 3 -binding site or with IP 3 . Neither the mutated (1-8-14 C ) nor scrambled (1-8-14 S ) peptide had any effect on IP 3 binding to the NT ( Figure 4A). These results with IP 3 R fragments expressed in E. coli, which lack CaM, also further support our conclusion that the effects of 1-8-14 peptide are entirely independent of CaM.
Comparison of the effects of 1-8-14 peptide on stimulating [ 3 H]IP 3 binding to the NT (EC 50 , 615 μM; pEC 50 , 3.21 + − 0.19) ( Figure 4B) with its inhibitory effect on IP 3 -evoked Ca 2 + release (IC 50 , 767 μM; pIC 50 , 3.1 + − 0.25) ( Figure 2C) demonstrates that each is similarly sensitive to the peptide. These results are consistent with our hypothesis that the 1-8-14 peptide disrupts an interaction between the SD and IBC that is essential for IP 3 R activation. The peptide thereby inhibits IP 3 -evoked Ca 2 + release ( Figure 2) and IP 3 R activity (Figure 3) and, by uncoupling IP 3 binding from subsequent conformational changes, it stimulates IP 3 binding (Figure 4). Subsequent experiments used mutagenesis of residues within the endogenous 1-8-14 motif to test this hypothesis further.

Mutations within the endogenous 1-8-14 sequence increase IP 3 -binding affinity
If, as we suggest, the 1-8-14 peptide disrupts an essential interaction between the endogenous 1-8-14 sequence and another domain within the NT, we might expect mutation of appropriate residues in the SD to both disrupt IP 3 R activation and increase IP 3 -binding affinity. We tested the latter prediction by examining IP 3 binding to the NT in which each of the critical (1, 8 and 14) hydrophobic/aromatic residues that are important for Ca 2 + -CaM binding to 1-8-14 motifs [45] was replaced with a charged hydrophilic residue (glutamate). The same hydrophobic residues are essential for MLCK [35] and 1-8-14 ( Figure 2C Figure 1A) were expressed in E. coli. Expression levels of the NT and its mutants were not identical ( Figure 5A), but they were each sufficient to allow the affinity for IP 3 and the effects of peptides to be determined after cleavage of the His 6 tag, but without further purification [33]. As expected, IP 3 bound to the IBC with greater affinity (17fold) than to the NT ( Figure 5B) [33,46,47], consistent with our suggestion that, in the absence of the SD, less binding energy is diverted into conformational changes [33]. Mutation of critical residues within the endogenous 1-8-14 motif significantly increased the affinity of the NT for IP 3 ( Figure 5B and Table 1), although none was as effective as complete removal of the SD. This is consistent with our observation that neither the 1-8-14 ( Figure 2) nor MLCK [35] peptide entirely inhibits IP 3 -evoked Ca 2 + release, whereas removal of the SD totally uncouples IP 3 binding from IP 3 R activation [48]. Although maximally effective concentrations of MLCK (100 μM) or 1-8-14 (3 mM) peptides similarly increased IP 3 binding to the NT, neither peptide had any effect on [ 3 H]IP 3 binding to the NT with mutations in any of the critical 1-8-14 residues (Figures 5C and 5D). Mutation of a residue immediately preceding the critical 1position of the 1-8-14 motif (K52E), which did not increase the affinity of IP 3 for the NT (Supplementary Figure S2A at http://www.biochemj.org/bj/449/bj4490039add.htm), had no effect on the responses to MLCK or 1-8-14 peptides (Figures 5C  and 5D) and neither did it affect IP 3 -evoked Ca 2 + release [33] (Supplementary Figure S2B). These results establish that mutation of critical residues within the endogenous 1-8-14 motif selectively increases IP 3 -binding affinity and these effects are non-additive with those of either MLCK or 1-8-14 peptide.  the IP 3 R [33]. Antagonists, because they need not evoke the rearrangement of the IBC and SD that initiates IP 3 R activation, may be less affected by disruption of these interactions. We therefore examined the effects of the SD and of point mutations within the endogenous 1-8-14 sequence on binding to the NT of heparin, a competitive antagonist of IP 3 [49]. The results demonstrate that, whereas removal of the SD increased the affinity of the NT for IP 3 17-fold, it caused only a 2-fold increase in the affinity for heparin. Point mutations within the endogenous 1-8-14 motif also caused larger increases in the affinity for IP 3 than for heparin ( Figures 5E and 5F, and Table 1). These results are important because they demonstrate that the effects of the SD and of mutations within the 1-8-14 sequence on ligand binding are specific for an agonist of the IP 3 R. They thereby demonstrate the importance of the 1-8-14 motif in specifically mediating activation of IP 3 Rs. and the second has mutations at both the 1-(F53E) and 14positions (Y66E) (IP 3 R FY ). As expected, a maximally effective concentration of IP 3 (10 μM) failed to stimulate Ca 2 + release from permeabilized DT40 cells lacking IP 3 R (DT40-KO cells) [37,40], but it caused release of 81 + − 1 % of the Ca 2 + stores of DT40-IP 3 R1 cells (Figures 6B and 6C). In the cell lines expressing IP 3 R with a mutated 1-8-14 motif, there was barely detectable Ca 2 + release that was not significantly different from that observed in DT40-KO cells (Figures 6B and 6C). ATP-dependent Ca 2 + uptake into the ER was similar for each cell line ( Figure 6C). We were concerned that the lower level of expression of mutant IP 3 R relative to wild-type (∼30-50 %, Figure 6A) might have contributed to the lack of detectable IP 3 -evoked Ca 2 + release. However, in another stable DT40 cell line where the IP 3 -binding site was mutated (R568Q), causing a ∼10-fold decrease in IP 3 affinity [50], IP 3 R expression (∼15 % of wild-type) was less than half that of the cell lines with mutations in the 1-8-14 motif ( Figure 6A). Nevertheless, IP 3 caused a readily detectable release of Ca 2 + from the intracellular stores of DT40-IP 3 R R568Q cells (49 + − 2 % of that detected in DT40-IP 3 R1 cells) (Figures 6B and  6C). We conclude that the lack of detectable Ca 2 + release in cells expressing IP 3 R with a mutant 1-8-14 motif is not attributable to reduced IP 3 R expression. Neither is it likely that the lack of response to IP 3 from mutant IP 3 R reflects a more global disruption of IP 3 R structure because each of the full-length mutant IP 3 Rs bound IP 3 , although, as predicted, addition of MLCK peptide increased IP 3 binding to only the wild-type IP 3 R ( Figure 6D). Furthermore, DT40 cells expressing IP 3 R1 with a mutation in an adjacent residue (DT40-IP 3 R1 K52E ) responded normally to IP 3 [33] (Supplementary Figure S2B). These results are consistent with the suggestion that mutations within the endogenous 1-8-14 motif mimic addition of exogenous MLCK peptide by uncoupling IP 3 binding from the conformational changes that lead to opening of the IP 3 R pore. Single-channel analyses provide further support for this conclusion.

Mutations within the endogenous 1-8-14 motif uncouple IP 3 binding from gating of IP 3 Rs
Yamazaki et al. [51] reported recently the functional effects of mutations within IP 3 R including some within the 1-8-14 motif (F53D and Y66A). We note, however, that some of their mutations, e.g. Y167A, which is clearly implicated in IP 3 R activation, abolished IP 3 -evoked Ca 2 + release from microsomes without affecting Ca 2 + signals evoked by activation of the BCR (B-cell receptor) in intact cells. This unexplained disparity casts some doubt over whether in these assays responses from intact cells faithfully report the activity of IP 3 R. In DT40 cells expressing an IP 3 R with five mutations that included Y66A (the 14-position of the 1-8-14 motif), activation of the BCR evoked a Ca 2 + signal, suggesting that the mutant IP 3 R was functional [51]. However, in this IP 3 R, the mutant had one hydrophobic residue replaced by another and this might not radically affect the behaviour of the 1-8-14 motif. In preliminary analyses of cells expressing IP 3 Rs in which the first position of the 1-8-14 motif was mutated (F53D), Ca 2 + signals were also observed after activation of the BCR [51]. This may reflect a limitation of the BCR-based assay (see above) or it may provide evidence for a lesser role of the 1-position in the 1-8-14 motif. We have not succeeded in establishing a DT40 cell line expressing IP 3 Rs with only this mutation, although our results do clearly show that IP 3 Rs with mutations in both the 1and 14-positions (IP 3 R FY ) are barely responsive to IP 3 ( Figure 6).

Mutation of the endogenous 1-8-14 motif attenuates IP 3 R gating without affecting single-channel conductance
In keeping with the reduced expression of IP 3 R L60E in DT40 cells ( Figure 6A), the frequency with which functional IP 3 Rs were detected in excised nuclear patches was much lower for nuclei from paired experiments with DT40-IP 3 R1 L60E cells (three of 48 patches) than from DT40-IP 3 R1 cells (five of 13 patches). In parallel analyses, functional IP 3 Rs were never detected in DT40-KO cells (none of 30 patches). The single-channel conductances (γ Cs ) of the mutant IP 3 R L60E (209 + − 8 pS) and normal IP 3 R (214 + − 6 pS) were indistinguishable ( Figure 3D), but NP o was massively decreased in the mutant (Figures 3B and 6E). Our interpretation of the latter is, as we described in our analyses of the 1-8-14 peptide, limited by our inability, when NP o is so low for IP 3 R1 L60E , to estimate reliably the number of active IP 3 R within a patch. Nevertheless, it is clear that the major effect on single-channel behaviour of mutating the endogenous 1-8-14 motif of IP 3 R1 ( Figures 3B-3D and 6E) and of adding 1-8-14 peptide to normal IP 3 R1 (Figure 3) is similar: both decrease NP o without affecting γ Cs . These results establish that mutations in the endogenous 1-8-14 motif or addition of 1-8-14 peptide uncouple ligand binding from channel gating without compromising the behaviour of the pore.

Conclusions: interactions between endogenous 1-8-14 and CaM-like motifs mediate activation of IP 3 Rs
CaM [22] or related EF-hand-containing proteins [14,25], peptides that comprise 1-8-14 CaM-binding motifs [35,36] (Figures 2-4) or disruption of a conserved endogenous 1-8-14like motif within the SD of IP 3 Rs inhibit IP 3 -evoked Ca 2 + release ( Figures 5 and 6) by massively reducing NP o of IP 3 R (Figures 3  and 6E). We conclude that an endogenous 1-8-14 motif within the SD (Figure 1) is essential for IP 3 R activation. Where it has been examined, the inhibitory proteins or peptides are more potent when Ca 2 + is bound to the IP 3 R [26,35] (Figure 2F). We therefore speculate that the endogenous 1-8-14 motif may interact with an unidentified domain that includes an EF-hand-like structure and that these interactions might be related to Ca 2 + regulation of IP 3 R (Figure 7). We suggest that competing peptides (CaM-like or 1-8-14 motifs) or mutagenesis of the endogenous 1-8-14 motif inhibit IP 3 Rs by disrupting this essential interaction in a manner IP 3 binding to the IBC initiates conformational changes that pass via the SD and lead, via regulation of Ca 2 + binding to the IP 3 R, to opening of the pore [33]. (A) An endogenous 1-8-14 motif within the SD is essential for IP 3 R activation. We speculate (upper panel) that interaction of this CaM-binding motif (red, conserved hydrophobic residues in dark blue) with an endogenous, but presently unknown, CaM-like structure (pale blue) within the NT may link IP 3 binding to Ca 2 + binding. (B) Another possibility is that IP 3 binding rearranges the 1-8-14 motif and so repositions a critical acidic residue (Glu 246 ) that may then contribute to a Ca 2 + -binding site (Ca-1) [55]. The NT without IP 3 bound (PDB code 3UJ0) [5] is shown with the IBC in grey and the SD in green to highlight Phe 53 (within the 1-8-14 motif) and Phe 223 to which it is closely apposed (yellow box), residues proposed to form the Ca-1 site (pink box) and the β-sheet that links Phe 223 to Glu 246 (cyan box). The expanded views (each rotated to show key movements) show the critical residues and the linking β-sheet before (green) and after IP 3 binding (blue, PDB code 3UJ4). The carboxy oxygen atoms in Glu 246 are shown in magenta. We speculate that separation of Phe 53 and Phe 223 when IP 3 binds is associated with twisting of the linking β-sheet and movement of Glu 246 towards three other acidic residues (Glu 425 , Asp 426 and Glu 428 ) and that they may then together form an effective Ca 2 + -binding site. similar to the 'unzipping' of interdomain interactions in RyRs [32,43,52]. The scheme is appealing because IP 3 regulates binding of Ca 2 + to IP 3 Rs and thereby leads to channel gating [34,53]. The identity of this Ca 2 + -binding site is unknown. It is, however, clear that Ca 2 + regulates IP 3 binding to the NT only when the SD is present [42], suggesting that a Ca 2 + -binding site within the NT may be regulated by interactions between the SD and IBC. One possibility is that an endogenous EF-hand-like structure might provide the Ca 2 + -binding site and that its interaction with the 1-8-14 motif links IP 3 and Ca 2 + binding ( Figure 7A). Bioinformatic analyses had suggested the presence of two possible EF-hand-like structures within the IBC [9,54], but neither is evident in high-resolution structures of the IBC [55] and NT [5,56]. Neither have we succeeded in identifying a complementary partner of the 1-8-14 motif. Another possibility is suggested by comparison of the structures of the NT with and without IP 3 bound [5,56], which reveal that Phe 53 (the first hydrophobic residue of the 1-8-14 motif) and Phe 223 are closely apposed (∼ 3.9 Å; 1 Å = 0.1 nm), but they move apart (∼ 5.3 Å) when IP 3 binds ( Figure 7B). A β-sheet links Phe 223 to Glu 246 , and the movement of Phe 223 is associated with a repositioning of an acidic residue in the β-domain of the IBC (Glu 246 ). This brings Glu 246 closer to three other acidic residues (Glu 425 , Asp 426 and Glu 428 ). The rearrangement is interesting because these four residues have been proposed to form a Ca 2 + -binding site (Ca-I) [55]. Furthermore, a peptide (residues 378-450) that includes most of these residues binds Ca 2 + , and the binding is abolished by mutation of the acidic residues [42]. A second possibility is therefore that IP 3 -evoked movement of the critical 1-8-14 motif contributes to formation of an effective Ca 2 + -binding site within the IBC by bringing a fourth acidic residue into appropriate association with three others.
We conclude that a conserved 1-8-14 motif within the SD is essential for IP 3 R activation and speculate that its interaction with either an endogenous CaM-like motif or acidic residues within the IBC may link IP 3 and Ca 2 + binding. Inhibition of IP 3 R by CaM and related proteins probably results from disruption of this essential interaction.