Research ArticlesChannels

Defining the stoichiometry of inositol 1,4,5-trisphosphate binding required to initiate Ca2+ release

See allHide authors and affiliations

Sci. Signal.  05 Apr 2016:
Vol. 9, Issue 422, pp. ra35
DOI: 10.1126/scisignal.aad6281

Four says, “Open!”

Intracellular calcium regulates such specialized processes as muscle contraction, neurotransmitter release, and insulin secretion and such general cellular processes as gene expression, proliferation, and cell death. The inositol 1,4,5-trisphosphate (IP3) receptor (IP3R) plays key roles in increasing intracellular calcium by functioning as a channel that releases Ca2+ from the endoplasmic reticulum. The IP3R is a tetramer with each subunit having a binding site for IP3. Alzayady et al. determined how many subunits have to bind IP3 for the channel to open by engineering versions of the receptor expressed as a single protein with different numbers of IP3-binding sites. Analysis of these concatenated receptor proteins revealed that channel activity required IP3 bound to each of the four binding sites, which ensures that cells do not discharge calcium unless the signal to do so is strong enough. A similar approach could be used to define how heterozygous mutations in IP3R subunits produce disease.


Inositol 1,4,5-trisphosphate (IP3) receptors (IP3Rs) are tetrameric intracellular Ca2+-release channels with each subunit containing a binding site for IP3 in the amino terminus. We provide evidence that four IP3 molecules are required to activate the channel under diverse conditions. Comparing the concentration-response relationship for binding and Ca2+ release suggested that IP3Rs are maximally occupied by IP3 before substantial Ca2+ release occurs. We showed that ligand binding–deficient subunits acted in a dominant-negative manner when coexpressed with wild-type monomers in the chicken immune cell line DT40-3KO, which lacks all three genes encoding IP3R subunits, and confirmed the same effect in an IP3R-null human cell line (HEK-3KO) generated by CRISPR/Cas9 technology. Using dimeric and tetrameric concatenated IP3Rs with increasing numbers of binding-deficient subunits, we addressed the obligate ligand stoichiometry. The concatenated IP3Rs with four ligand-binding sites exhibited Ca2+ release and electrophysiological properties of native IP3Rs. However, IP3 failed to activate IP3Rs assembled from concatenated dimers consisting of one binding-competent and one binding-deficient mutant subunit. Similarly, IP3Rs containing two monomers of IP3R2short, an IP3 binding–deficient splice variant, were nonfunctional. Concatenated tetramers containing only three binding-competent ligand-binding sites were nonfunctional under a wide range of activating conditions. These data provide definitive evidence that IP3-induced Ca2+ release only occurs when each IP3R monomer within the tetramer is occupied by IP3, thereby ensuring fidelity of Ca2+ release.


The modulation of intracellular calcium concentration [Ca2+]i is a signal used by all living organisms to control many cellular processes, including gene transcription, regulated secretion, proliferation, muscle contraction, fertilization, and apoptosis (1). Inositol 1,4,5-trisphosphate (IP3) receptors (IP3Rs) are ligand-gated calcium channels present in most eukaryotic species (13). There are three homologous isoforms ubiquitously distributed in mammalian tissues (denoted as IP3R1, IP3R2, and IP3R3) (4). The individual channel is assembled with a tetrameric architecture consisting of homo- or heteromeric subunits and is localized in the membranes of intracellular calcium stores, such as the endoplasmic reticulum and sarcoplasmic reticulum (3, 5). Each IP3R subunit consists of ~2700 amino acids and three functional domains: the N-terminal IP3-binding domain, an intervening modulatory domain, and the C-terminal channel domain containing six transmembrane helices and the Ca2+-permeable pore (5, 6). All known regulatory sites, including phosphorylation sites, nucleotide binding sites, and interaction motifs with protein binding partners are present in each monomer (3, 5, 7). Accordingly, each monomer contributes a ligand-binding site to the tetrameric channel.

Previous studies suggest that IP3 binding triggers inter- or intramolecular conformational changes between subunits involving interactions between the N and C termini that ultimately culminate in channel opening and Ca2+ release (811). However, fundamental questions regarding the molecular basis of channel gating remain (12, 13). For example, the number of IP3 molecules required to open the tetrameric IP3R channel is not established; indeed, the stoichiometry necessary for any regulatory input is not known. Understanding the stoichiometry of IP3R regulation is especially important because IP3R dysfunction as a result of mutations in the protein has been implicated in human pathologies, such as anhidrosis (an inability to perspire) and spinocerebellar ataxia (1416), and IP3R-associated diseases occur in the heterozygous condition in which the IP3R channel is likely formed from various combinations of normal and mutant monomers (16, 17).

Although IP3 binding to IP3R is not generally reported to be highly cooperative (18, 19), some studies have reported Ca2+ release to be highly cooperative in permeabilized cell systems (20, 21). This cooperative relationship could be a consequence of Ca2+ facilitating further release by acting as a coagonist; however, the cooperativity has been observed in experiments in which intracellular Ca2+ is buffered (20, 22). One explanation for the cooperative Ca2+ release has been that at least three ligand-binding events are required for channel opening (20, 23). Contrary to this idea, transiently expressed IP3R channels engineered with only two intact binding sites are reported to be activated by IP3 (10). Further, structural analysis by cryoelectron microscopy (cryo-EM) suggests that, as a result of interactions between the C terminus of one subunit and the N termini of neighboring subunits, IP3 binding to one subunit induces conformational changes in two adjacent subunits, and thus, this model implies the requirement for binding of less than four molecules of IP3 for channel gating (11). Nevertheless, a comprehensive molecular methodology has never been attempted to address how many IP3 molecules are necessary to activate IP3R channels. Here, we engineered concatenated IP3R, in which progressive numbers of IP3-binding sites were disrupted by either mutation or incorporation of naturally occurring binding-deficient mutants. Furthermore, we expressed these concatenated receptors in cells lacking IP3Rs, thereby avoiding issues of incorporation of endogenous subunits into the channel. With these concatenated receptors, we systematically monitored IP3R activity and addressed the stoichiometry of IP3 molecules required for optimum channel function. Our data establish that IP3-induced Ca2+ signals are only initiated after occupation of IP3R by four IP3 molecules.


The relationship between IP3 binding and IP3R activation is complex

The DT40-3KO cell line is a chicken B-lymphoid cell type that has been genetically modified to be deficient in all subunits of the IP3R (24, 25). Thus, these cells provide a system for analyzing IP3Rs with controlled subunit composition. We investigated the relationship between IP3 binding and IP3-induced Ca2+ release by comparing data obtained from competitive 3H-IP3 binding assays and unidirectional Ca2+-release assays performed in permeabilized DT40-3KO cells expressing rat IP3R1. Ca2+ release was only triggered after substantial (~82%) steady-state receptor occupancy (Fig. 1). By necessity, the 3H-IP3 binding and Ca2+-release assays were performed under somewhat different assay conditions including differences in the experimental buffers and assay endpoints. Despite this caveat, if extrapolated to dynamic conditions in intact cells responding to a natural IP3-generating stimulus, the results are consistent with Ca2+ release only being activated at any particular IP3 concentration ([IP3]) when the probability is high that most, if not all, subunits are occupied by IP3 at a given time (20). Thus, these data suggest that multiple subunits must be occupied to initiate Ca2+ release.

Fig. 1 Comparing the IP3 occupancy with IP3-induced Ca2+ release in DT40-3KO cells expressing R1.

Competitive binding assay was performed using IP3R1 immunopurified from corresponding DT40-3KO cells. IP3-evoked Ca2+ flux was assessed in permeabilized DT40-3KO cells expressing R1. Normalized values were plotted as a function of the log[IP3]. Median effective concentration (EC50) for IP3 binding is 50 ± 5 nM, and EC50 for Ca2+ release is 1.84 ± 0.3 μM. Data are shown as means ± SE of at least three independent experiments.

Next, we tested whether incorporating IP3 binding–defective subunit(s) into the tetrameric assembly impaired IP3R channel function. Positively charged residues scattered in the N terminus of IP3R subunits mediate IP3 binding (26). These early studies, which defined the IP3-binding domain, replaced key arginine and lysine residues absolutely required for IP3 binding with glutamine. The rationale being that lysine, arginine, and glutamine are amphipathic and have similar side-chain conformational entropy, and thus, the mutation, while negating the positive charge, is predicted to be less disruptive to the structural and functional integrity of the receptor (26). In mammalian IP3R family members, there are 10 conserved basic residues that are essential for binding, and 3 of which are considered critical for specific binding (R265, K508, and R511; residue numbering based on Rat IP3R1) (26). Although initial studies using IP3R1 with a single R265Q mutation reported that this mutant does not bind IP3 and thus is not activated by IP3 (10, 26), a subsequent study showed that this construct retains IP3 binding capacity (27).

To resolve this discrepancy, we generated two different IP3R1 mutants, IP3R1.R265Q (designated as R1Q) and IP3R1.R265/K508/R511Q (designated as R1QQQ), and stably transfected plasmids encoding these into IP3R-null, DT40-3KO cells. When compared to wild-type IP3R1 (designated as R1), R1QQQ retained no binding activity, whereas R1Q retained low but detectable binding (~10%) (Fig. 2A). In permeabilized cell unidirectional Ca2+-release assays, IP3 evoked robust Ca2+ release from cells expressing R1, but IP3 failed to release Ca2+ in nontransfected DT40-3KO cells or cells stably expressing R1QQQ (Fig. 2B). However, cells expressing R1Q produced a small but measurable Ca2+ signal in response to exogenous IP3, suggesting that the residual binding capacity of this mutant supported limited channel activation under these assay conditions.

Fig. 2 Evaluating the function of IP3R1 with different mutations in the ligand-binding domain expressed in DT40-3KO cells.

(A) DT40-3KO cells expressing the indicated IP3R constructs were lysed, and IP3R proteins were immunoprecipitated and used for binding assays. Specific-binding values were divided by the amounts of the corresponding immunoprecipitated proteins, and these values were normalized to that of the R1-expressing cells. Statistical analysis was carried out with one-way analysis of variance followed by Tukey post hoc test. Pooled data from three independent experiments are shown. (B) Unidirectional Ca2+ flux assays in permeabilized DT40-3KO cells stably expressing the indicated IP3R constructs and loaded with Furaptra. Data presented as means ± SE of at least three experiments. (C) Trypsin-stimulated Ca2+ release in DT40-3KO cells stably expressing the indicated IP3R constructs. The cells were loaded with Fura2-AM and stimulated with 500 nM trypsin. Representative traces are shown. (D) Quantitative analysis of the response of the indicated IP3R-expressing DT40-3KO cells to trypsin. Red histograms depict the average maximum change over basal 340:380 fluorescence ratio resulting from trypsin stimulation. Blue histograms represent percentage of the cells responding to trypsin with >0.1 change in the 340:380 fluorescence ratio. Data are presented as means ± SE. Experiments were repeated at least three times with more than 60 cells analyzed in each experiment. *Statistically significant differences (P < 0.01) as determined by Tukey post hoc test.

Using the stably transfected DT40-3KO, we investigated the ability of these IP3R1 constructs to respond to IP3-generating stimuli in intact cells. We challenged cells with the Gαq-coupled G protein (heterotrimeric guanine nucleotide–binding protein)–coupled receptor (GPCR) agonists, endothelin and trypsin. We also stimulated the B cell receptor (BCR), a tyrosine kinase receptor, by crosslinking with an antibody recognizing IgM. All three IP3-producing stimuli failed to trigger Ca2+ release in DT40-3KO and cells expressing R1Q or R1QQQ, whereas DT40-3KO cells expressing R1 showed robust responses to all three stimuli (Fig. 2, C and D, and fig. S1, A and B). The contradictory behavior of R1Q in the two assay systems is likely due to the different experimental conditions. In particular, the permeabilized cell assay exposes IP3R to uniform maximum concentrations of IP3 and adenosine 5′-triphosphate (ATP), whereas the intact cell assay relies on endogenously produced IP3 downstream of surface receptor signaling that likely results in less intense stimulation. These data clearly illustrated that the use of the single point mutation (R265Q) is inadequate for assessing ligand stoichiometry. Therefore, we used R1QQQ for the subsequent experiments.

Mutagenesis in the ligand-binding domain is unlikely to result in a global structural rearrangement of IP3R

A concern associated with functional analysis of mutated proteins is that any substitution may interfere with protein folding, resulting in unwanted, allosteric effects beyond the intended targeted disruption. To investigate whether the introduction of glutamine residues in R1QQQ markedly disrupted the overall structural integrity of IP3R1, we analyzed the available IP3R structures. A cryo-EM study at 4.7 Å shows the full-length IP3R (Fig. 3A), but the local resolution in the IP3-binding site is not sufficient to reveal individual side chains (11). Therefore, we used a crystal structure of the N-terminal region, captured in the absence of IP3 (9). The crystal structure contains two IP3R molecules in the asymmetric unit, which show identical folds and relative domain orientations upon superposition. Each of the three mutated residues (R265, K508, and R511) is surface-accessible and is not involved in domain-domain interactions (Fig. 3, B and C). To verify that the mutations are sterically allowed, we generated a homology model of the ligand-binding domain using Modeller, which satisfied spatial restraints (28). Although a high-resolution structure of the R1QQQ mutant would be required to unequivocally rule out any major conformational changes, our homology model predicted that the glutamine side chains did not produce substantial rearrangements of neighboring residues. Furthermore, the side chains of R265 and K508 make no significant interactions with other residues, and the few hydrogen bonds (Fig. 3C) are only observed in one of two molecules in the asymmetric unit. R511 makes hydrogen bonds with two neighboring residues, observed in both molecules of the asymmetric unit (Fig. 3C). However, because these residues are within the same domain, we predict that it is unlikely that disruption would result in long-range allosteric effects (Fig. 3D). In conclusion, although conformational changes and allosteric effects of the R1QQQ mutant cannot be completely ruled out, we deem this unlikely.

Fig. 3 Overview of the IP3R1 structure and IP3 binding site.

(A) Cryo-EM structure of type 1 IP3R (cytosolic view), showing the electron density map in transparent white [Electron Microscopy Data Bank (EMDB) entry 6369] (11), and the model in cartoon representation [Protein Data Bank (PDB) ID: 3JAV]. The area for which crystal structures are available is highlighted in colors, with the suppressor domain in blue and the two domains contributing to IP3 binding in green and red. The structure was solved in the absence of IP3, but its binding site in each subunit is indicated by black spheres. (B) Close-up of the 3.6 Å crystal structure of the IP3R N-terminal region in complex with IP3 (PDB ID: 3UJ0) (9). The IP3 as well as selected residues that make interactions are shown in stick representation. The residues mutated in this study are indicated in black and labeled (numbering according to rat IP3R1). (C) Close-up of the 3.0 Å crystal structure of the IP3R N-terminal region without IP3 (PDB ID: 3UJ4) (9). All side chains are shown in black, with the three residues mutated in this study in stick representations. Hydrogen bonds made with the three residues are shown as dashed lines. Only the hydrogen bonds with R511 are observed in both molecules of the asymmetric unit. (D) Homology-based model of the R1QQQ mutant, showing the same area as in (C).

R1QQQ subunits exert an apparent dominant-negative effect on IP3R1 activity

We hypothesized that if incorporation of R1QQQ subunit(s) into tetrameric IP3R attenuated IP3-induced Ca2+ release, then it would indicate a dominant-negative effect and suggest that full occupancy of IP3R monomers with IP3 is required to induce Ca2+ release. We transiently transfected wild-type IP3R1 tagged with mCherry (cherryR1) or empty vector into DT40-3KO or DT40-3KO stably expressing a high level of R1QQQ. We used the mCherry fusion construct to identify transfected cells. Because both transiently and stably transfected IP3R1 constructs are driven by an exogenous cytomegalovirus promoter, we anticipated that nascent subunits would oligomerize to form tetrameric IP3R complexes in various combinations. To establish that transiently transfected R1 oligomerized with stably expressed R1 proteins in the DT40-3KO cell system, we performed coimmunoprecipitation experiments, which confirmed that transiently expressed cherryR1 associated with stably expressed FLAG-tagged IP3R1 (Fig. 4A, inset). Although we cannot formally exclude the possibility of intermolecular interactions between adjacent homomeric populations of IP3R constructs, these data suggested that tetrameric IP3R complexes can be assembled from transiently and stably coexpressed IP3R subunits in DT40-3KO cells.

Fig. 4 Attenuation of IP3-induced Ca2+ signal in DT40-3KO cells and HEK-3KO expressing IP3 binding–deficient subunits.

(A) Ca2+ release in transiently transfected cells loaded with Fura2-AM and stimulated with 50 nM trypsin. Inset: Lysates from DT40-3KO cells expressing FLAG-tagged IP3R1 transfected with mCherry or cherryR1 were mock-treated or immunoprecipitated (IP) with antibodies recognizing FLAG antibodies and immunoblotted (Western blot: WB) for the indicated tags. (B) Quantitative analysis of the response of the indicated IP3R-expressing DT40-3KO cells from (A). Data presented as means ± SE of at least four independent experiments. *Statistically significant differences (P < 0.01). (C) Ca2+ release in transiently transfected HEK-3KO cells loaded with Fura2-AM and stimulated with 500 pM trypsin. Inset: Lysates of HEK-3KO cells expressing the indicated constructs or vector-control cells were analyzed by immunoblotting for IP3R1. Upper bands represent cherryR1 and lower band represents R1QQQ. (D) Quantitative analysis of the response of the indicated IP3R-expressing cells from (C). Data presented as means ± SE of at least four independent experiments. *Statistically significant differences (P < 0.01) as determined by Tukey post hoc test.

DT40-3KO transfected with cherryR1 supported robust Ca2+ release in response to trypsin, whereas stimulation of R1QQQ cells transfected with cherryR1 resulted in weak Ca2+ release (Fig. 4A and pooled data in Fig. 4B). These data indicated that, although transiently transfected cherryR1 assembled to form functional channels in DT40-3KO, the coexpression and incorporation of ligand binding–defective subunit(s) inhibited the formation of competent channels. Together, these findings suggested that incorporating a mutant subunit(s) impairs the tetrameric channel function, which supports the hypothesis that an IP3R with less than four intact ligand-binding sites does not function as an IP3-gated Ca2+ channel.

Although DT40-3KO cells are a useful expression system, there are caveats associated with this cell line, such as their relatively low transfection efficiency and their nonmammalian origin. Therefore, we sought to confirm our findings in a mammalian cell type. Human embryonic kidney (HEK) 293 cells are widely used, easily maintained, and readily transfected, but all three IP3R subtypes are present in these cells (fig. S2A). To generate IP3R-null HEK293 cells, we used CRISPR/Cas9 gene editing to simultaneously disrupt all three IP3R-encoding genes (29). Genotyping, Western blot analyses, and single-cell imaging confirmed the disruption of IP3R-encoding genes and the absence of functional IP3Rs (Fig. 2A, fig. S2B, and table S1).

We assessed the effect of wild-type IP3R1 and binding-defective subunit (R1QQQ) on channel function in these IP3R-deficient HEK293 cells (designated as HEK-3KO). The amount of cherryR1 was similar under both conditions (Fig. 4C, inset). As expected, cherryR1 alone expressed mediated robust Ca2+ release in response to trypsin (Fig. 4C and pooled data in Fig. 4D). Coexpression of R1QQQ with cherryR1 significantly attenuated the trypsin-induced Ca2+ response. Cotransfection would be anticipated to result in a binomial distribution of heterotetrameric assemblies of all combination of R1 and R1QQQ (30). The degree of inhibition of Ca2+ release appears inconsistent with a minority population of homomeric R1QQQ and indicates that the mutant subunit likely exhibits dominant-negative effects when present in the complex with some wild-type subunits.

IP3R1 tetramers with two IP3 binding–deficient subunits lack activity

Our results thus far indicated that a tetrameric channel with less than four intact binding sites is not an IP3-gated Ca2+ channel. However, in these experimental paradigms, it is not possible to accurately define the subunit composition of each assembled channel. To address precisely how many binding sites are required for channel activation, we generated concatenated channels with predefined subunit composition. Constructs encoding dimeric concatenated IP3R (for cartoon, see Fig. 5A) generate proteins that dimerize to form authentic IP3-gated tetrameric channels (Fig. 5B) (31, 32). To determine whether IP3R channels with two ligand binding–deficient subunits are functional, we engineered concatenated IP3R constructs in which two IP3R1 molecules are linked tail-to-head. We constructed two dimeric concatenated receptors with different configurations containing wild-type subunits and IP3 binding–deficient subunits, designated as R1R1QQQ and R1QQQR1, and stably expressed them in DT40-3KO (Fig. 6A). Permeabilized cell assays demonstrated that cells expressing R1R1QQQ did not produce any Ca2+ release in response to application of IP3, suggesting that channels containing two mutant subunits were not functional (Fig. 6B). Furthermore, consistent with the dimeric binding-deficient constructs not forming functional channels, cells expressing either R1R1QQQ or R1QQQR1 did not produce a Ca2+ signal in response to trypsin (Fig. 6C and pooled data in Fig. 6D), endothelin, or BCR stimulation of intact cells (fig. S3, A and B). Together, these findings showed that binding of two IP3 molecules is not sufficient to activate tetrameric IP3R channels. Notably, we observed Ca2+-release activity in both the permeabilized cell assay and by single-cell imaging in cells expressing tetrameric channels assembled from R1QR1 or R1R1Q dimers (fig. S4, A to D). These channels consist of two wild-type IP3R subunits and two R1Q subunits. The results are consistent with the residual binding activity of monomeric R1Q subunits (Fig. 2A).

Fig. 5 Diagram showing the construction of dimeric R1 and its oligomerization into a functional receptor.

(A) A cartoon depicting a dimeric R1 complementary DNA (cDNA) construct flanked by the two arms of the linear vector, pJAZZ mamm. The “head” subunit (cDNA coding for rat IP3R1QQQ) was modified so that it contains an Nco I site just before the start codon. The stop codon was deleted and a nucleotide sequence coding for a seven–amino acid linker was added after the IP3R1 coding sequence followed by an Age I site. The “tail” subunit (cDNA coding for rat IP3R1) was modified so that it contains an Age I site followed by a nucleotide sequence coding for a seven–amino acid linker inserted before the start codon, and a blunt end restriction site was inserted after the stop codon. The resultant construct encodes one open reading frame consisting of two IP3R subunits connected with a 14–amino acid linker. CT, C terminus; NT, N terminus. (B) A scheme showing how dimeric R1R1 molecules assemble to form a tetrameric channel.

Fig. 6 IP3R channels with two ligand binding–deficient subunits are not functional.

(A) Immunoblot shows the abundance of the indicated IP3R constructs. The number of asterisks corresponds to the number of the concatenated subunits in the IP3R1 constructs. (B) Unidirectional Ca2+ flux assays in permeabilized DT40-3KO cells stably expressing the indicated IP3R constructs. Cells were loaded with Furaptra-AM and permeabilized, and Ca2+ flux was measured. Independent experiments were repeated at least three times in which greater than 60 cells were imaged in each experiment. Data presented as means ± SE. (C) Trypsin-stimulated Ca2+ release in DT40-3KO or DT40-3KO cells stably expressing IP3R constructs as indicated. Cells were loaded with Fura2-AM and stimulated with 500 nM trypsin. Representative traces are shown. (D) Quantitative analysis of the response of the indicated IP3R-expressing DT40-3KO cells in (C). Red histograms depict the average change over the basal 340:380 fluorescence ratio resulting from trypsin stimulation. Blue histograms represent percentage of the cells responding to trypsin with >0.1 change in the 340:380 fluorescence ratio. *Tukey post hoc statistically significant differences (P < 0.01).

A functional IP3R1 channel can be generated as one concatenated molecule

To precisely control the subunit composition of the receptor without the complicating factor of oligomerization, we generated IP3R from a single concatenated polypeptide with a 14–amino acid linker between the subunits and expressed this protein in DT40-3KO cells. We confirmed by Western blotting that the wild-type version of this IP3R1 construct (R1R1R1R1) could be stably expressed in DT40-3KO cells (Fig. 7A, left). Cells expressing R1R1R1R1 exhibited robust IP3-induced Ca2+ release in the permeabilized cell assay (Fig. 7B) and agonist-induced Ca2+ signals in intact cells (Fig. 7, C and D, and fig. S5, A and B). Pivotal to interpretation of these data is to confirm that the concatenated tetrameric channel behaves in an identical fashion to IP3R assembled from wild-type subunits. To confirm that the concatenated channel generated from a single polypeptide chain had the same electrophysiological properties as IP3R assembled from wild-type monomer subunits, we performed extensive electrophysiological analyses of single-channel properties with the “on-nucleus” configuration of patch clamp. ATP and Ca2+ function as coagonists with IP3 of the IP3R and enhance the responsiveness of the channel IP3 (3, 33). Single-channel recordings of R1R1R1R1 showed that the channel was activated by IP3 and that the response to IP3 was enhanced by either increasing the concentration of ATP (Fig. 7E, left) or Ca2+ (Fig. 7E, right). Indeed, the activity of channels produced from monomers, concatenated dimers, or the single polypeptide concatenated tetramer was indistinguishable in terms of current-voltage relationship (fig. S6, A to D), regulation by Ca2+ and ATP (figs. S7 and S8), and the biophysical characteristics of gating (figs. S9 and S10), all of which define IP3R activity. Thus, channels formed from concatenated R1R1R1R1 have biophysical properties of native channels and function as bona fide IP3Rs (Table 1 and table S2). We first used this concatenated single polypeptide approach to examine the response of cells expressing an IP3R with only two IP3-binding sites (R1R1QQQR1R1QQQ). Consistent with data generated with the R1R1QQQ dimer–expressing cells, R1R1QQQR1R1QQQ expression did not support trypsin-stimulated Ca2+ signals (fig. S11).

Fig. 7 IP3R channels with one ligand binding–deficient subunits are not functional.

(A) Immunoblot shows the abundance of the indicated IP3R constructs. The number of asterisks corresponds to the number of the concatenated subunits in the IP3R1 constructs. (B) Unidirectional Ca2+ flux assays in permeabilized DT40-3KO cells stably expressing the indicated IP3R constructs. Cells were loaded with Furaptra-AM and permeabilized, and Ca2+ flux was measured. Independent experiments were repeated at least three times. Data presented as means ± SE. (C) Trypsin-stimulated Ca2+ release in DT40-3KO or DT40-3KO cells stably expressing IP3R constructs as indicated. Cells were loaded with Fura2-AM and stimulated with 500 nM trypsin. Independent experiments were repeated at least three times where greater than 30 cells were imaged in each experiment. Shown are representative traces. (D) Quantitative analysis of the response of the indicated IP3R-expressing DT40-3KO cells in (C). Red histograms depict the average change over the basal 340:380 fluorescence ratio resulting from trypsin stimulation of corresponding cells. Blue histograms represent percentage of the cells responding to trypsin with >0.1 change in the 340:380 fluorescence ratio. *Statistically significant differences (P < 0.01). (E) Single-channel recordings of DT40-3KO cells expressing the four-concatenated subunit IP3R1. Recordings were made with the on-nucleus configuration of the patch-clamp technique. Left traces show the effect of two concentrations of ATP on channel activity in response to a low (1 μM) and high (10 μM) concentrations of IP3. Right traces show effect of two concentrations of Ca2+ on channel activity in response to low and high concentrations of IP3. (F) Single-channel recordings of DT40-3KO cells expressing the four-concatenated subunit lacking one ligand-binding site using the on-nucleus configuration. Recordings with two concentrations (10 and 100 μM) of IP3 in the presence of three concentrations (200 nM, 1 μM, and 10 μM) of Ca2+ are shown.

Table 1. Biophysical properties of wild-type IP3R formed from monomeric subunits, IP3R formed from concatenated receptor subunit dimers, and single polypeptide IP3R in response to maximal and submaximal IP3.
View this table:

IP3R1 with a single IP3bindingdeficient subunit lacks activity

Using the single concatenated polypeptide channel approach, we assessed whether an IP3R lacking only a single IP3-binding site was functional. We generated a single polypeptide construct containing three wild-type subunits conjugated to one R1QQQ subunit (R1R1R1R1QQQ) and stably expressed this protein in DT40-3KO cells (Fig. 7A, right). Cells expressing this construct were unresponsive to application of IP3 in the permeabilized cell assay (Fig. 7B) and did not display any Ca2+ signal in intact cells when stimulated with trypsin (Fig. 7C and pooled data in Fig. 7D) or endothelin (fig. S12A) or in response to BCR stimulation (fig. S12B). In single-channel electrophysiological studies, varying Ca2+ and ATP with IP3 concentrations as high as 100 μM never elicited any IP3R channel activity (Fig. 7F). In total, these data suggested that a fundamental property of the IP3R is the requirement for each monomer to bind IP3 to achieve channel activation.

IP3R2 containing two binding-deficient IP3R2(short) splice variant subunits lacks activity

The necessity for IP3 binding to each subunit of the IP3R tetramer may also have physiological consequences beyond the fundamental biophysical requirement to initiate Ca2+ release. A widely distributed splice variant of mammalian IP3R2 (R2short) does not bind IP3 despite having an intact ligand-binding domain (18). This splice variant lacks amino acids 176 to 208. These amino acids are in the region termed the “suppressor domain,” which is a major modulator of IP3-binding affinity (19, 26), and channels formed by this variant are nonfunctional (18). We determined by coimmunoprecipitation that this splice variant formed heteromeric channels when expressed in HEK293 cells with other full-length isoforms (Fig. 8A), consistent with previous reports that indicate that transiently transfected IP3Rs assemble into tetrameric complexes (30). IP3 binding in permeabilized DT40-3KO cells expressing a dimer of R2R2short was reduced by ~50% when compared with dimeric R2R2 composed of wild-type IP3R2 and, comparable to binding to R2R2QQ, a construct with two amino acid mutations in IP3R2 corresponding to amino acids analogous to those involved in IP3 binding in IP3R1 (Fig. 8B). However, a channel containing two subunits of R2short or R2QQ was not functional in response to trypsin activation (Fig. 8C and pooled data in Fig. 8D). These results indicated that this widely distributed variant could incorporate into tetrameric receptors and that the result would be to dampen or even eliminate intracellular Ca2+ signals produced by IP3-stimulated release from the intracellular stores. In summary, the experiments presented here support the conclusion that four binding-competent subunits are required for IP3-mediated channel activation and demonstrate the general utility of concatenated proteins to illuminate IP3R biology.

Fig. 8 Assessment of R2short oligomerization, IP3 binding, and Ca2+ release.

(A) Immunoblot shows the abundance of the indicated IP3R proteins from transiently cotransfected HEK293 cells. Lysate proteins were immunoprecipitated with the indicated antibodies, and proteins were detected with antibodies recognizing the indicated protein or tag. Lower two panels depict lysate input. HA, hemagglutinin. (B) IP3 binding of the indicated dimeric constructs immunopurified from corresponding DT40-3KO cell lines. Binding values were divided by corresponding densitometric values obtained from parallel immunoblots, and the results were normalized to that of R2R2. (C) Trypsin-stimulated Ca2+ release in DT40-3KO or DT40-3KO cells stably expressing IP3R constructs as indicated. Cells were loaded with Fura2-AM and stimulated with 500 nM trypsin. Independent experiments were repeated at least three times where greater than 50 cells were imaged in each experiment. Shown are representative traces. (D) Quantitative analysis of the response of the indicated IP3R-expressing DT40-3KO cells in (C). Red histograms depict the average change over the basal 340:380 fluorescence ratio resulting from trypsin stimulation of corresponding cells. Blue histograms represent percentage of the cells responding to trypsin with >0.1 change in the 340:380 fluorescence ratio. *Statistically significant differences (P < 0.01) as determined by Tukey post hoc test.


A question fundamental to understanding the activation of IP3R is how the IP3-binding event in the N terminus of the IP3R subunit is transmitted to the channel pore localized in the distant C terminus (8, 10, 13, 34). Multiple intra- and intersubunit interactions revealed by biochemical and structural studies, including a current model generated from a cryo-EM structure of the entire IP3R at 4.7 Å resolution (811, 35), are likely critical to coupling IP3 binding to channel opening. Central to understanding how the interactions between various domains gate the channel is determining the obligatory stoichiometry of IP3 binding necessary for channel activation. Here, we provide several lines of evidence from assaying IP3R channel activity that indicate that IP3R channels must bind four IP3 molecules to initiate activation and Ca2+ release. First, an analysis of the relationship between IP3 binding and unidirectional Ca2+ flux is consistent with the idea that IP3R1 has to be nearly fully occupied before Ca2+ release is activated. Because Ca2+ release is not an “all-or-nothing” event, these data can be explained if the graded response to IP3 is a consequence of mass action increasing the probability that all four monomers are concurrently IP3-bound. Second, coexpression of wild-type and ligand-binding defective subunits in both HEK-3KO and DT40-3KO cells significantly attenuated the formation of functional channels as evidenced by an apparent dominant-negative effect of ligand binding–deficient subunits. Clearly, the most definitive approach implemented in this study is the use of concatenated constructs whereby a single polypeptide channel of any predefined composition can be fashioned and expressed in isolation.

Previously, we characterized the biochemical and biophysical properties of the IP3Rs assembled from dimers of concatenated pairs of subunits expressed as a single polypeptide (31, 32). Here, we extended this approach to produce channels of defined composition from all four subunits expressed as a single concatenated polypeptide. No oligomerization of individual subunits is necessary to form functional channels with this construct. These channels assembled from concatenated dimers (31) or a single concatenated polypeptide of all four subunits had the same properties as wild-type channels in terms of IP3 binding, Ca2+ release, the amplitude and frequency of evoked Ca2+ signals, and the increase in IP3 sensitivity resulting from ATP and Ca2+. Further validation presented here showed that the single-channel conductance and open probability under various conditions are like those of channels assembled from monomeric subunits. Moreover, channel open and closed times, characteristics of channel gating, and the biophysical fingerprints of IP3R were identical comparing monomeric, concatenated dimeric, or concatenated tetrameric IP3R (as summarized in Table 1 and table S2). These data indicated that linking the IP3R subunits does not constrain the normal gating of the channel and established that the concatenation of subunits is a valid technique to interrogate IP3R function. Indeed, the cryo-EM structure of IP3R1 shows a close juxtaposition of the N and C termini of the tetrameric channel, which may explain how the four subunits of this extremely large tetrameric complex can be linked without compromising channel structure and, thus, activity (11).

Taking advantage of this approach, we systematically demonstrated that channels with fewer than four competent ligand-binding sites do not form IP3R capable of opening and initiating Ca2+ release. Initially, we examined the IP3-induced activity of concatenated dimeric IP3R1 with one wild-type and one ligand binding–deficient subunit. This construct is expected to form channels with two intact binding sites. The results of three complementary experimental approaches to monitor channel activity indicated that two IP3-binding events are not sufficient for channel activation. This notion is, however, not consistent with a previous report that heterotetrameric channels assembled from R1Q and binding-competent subunits were partially functional, leading the authors to conclude that less than full occupancy by IP3 was sufficient to gate IP3R (10). Although differing experimental systems and measurements of activity were used in this study, an interpretation from our data is that Ca2+ release observed in the previous study may be a consequence of the substantial IP3 binding and Ca2+-release activity retained by the R1Q subunit and thus does not necessarily provide definitive evidence of the stoichiometry of IP3 binding. Notwithstanding this apparent discrepancy, the absence of any activity in a concatenated single polypeptide channel with three intact binding sites is again consistent with the concept that all four subunits must be bound to the ligand to initiate Ca2+ release.

IP3R activation is a complex process and involves binding of Ca2+, IP3, ATP, and perhaps many other factors, including protein association or dissociation, redox regulation, and phosphorylation status (3, 7, 36). IP3 is obligate for channel opening and Ca2+ is also required for optimal channel activation, although the exact mechanism by which these two coagonists interact is not well established (3, 12, 26). Current proposals suggest that after IP3 binding, Ca2+-regulatory sites can be engaged to enhance IP3R open probability (3, 22). Our data are important in understanding the activation of IP3R involved in the initial Ca2+ release at “trigger” zones (37). The initial IP3-induced Ca2+-release event occurs before local Ca2+ concentrations at the IP3R have increased; thus, the intracellular Ca2+ concentration near the IP3R would likely be less than the threshold needed to regulate IP3R activity. Our single-channel data obtained with a Ca2+ concentration mimicking cellular resting levels, together with the intact-cell intracellular Ca2+ concentration measurements in cells that were stimulated with agonists that increase IP3, revealed that full IP3 engagement of the IP3R is necessary for the initial activation. These data are, therefore, consistent with the idea that upon stimulation, there is no absolute requirement for Ca2+ concentrations to increase above the resting level to initially open the IP3R. Nevertheless, after the initial activation, neighboring IP3Rs will experience an increase in local Ca2+ concentration; therefore, it is possible that the IP3-binding requirement is altered under these conditions. However, we found that in permeabilized cell assays and single-channel recordings, constructs with fewer than four intact binding sites were refractory to opening over a wide range of activating IP3, Ca2+, and ATP concentrations. Although many factors may affect the kinetics of IP3R channel opening after binding of four IP3 molecules, we conclude that each of the four subunits must bind IP3 for the channel to open. IP3R activity can be modulated by a variety of factors not addressed in the current studies, including by binding partners and redox modification (7, 22, 36, 38, 39). In general, these events occur by allosteric regulation of the functional affinity for IP3 and not by altering IP3 binding directly. Thus, although our findings clearly show that all four subunits must be bound to IP3 to trigger channel activity in the conditions tested, it remains possible that the stoichiometry of IP3 activation of IP3R may be altered under particular metabolic conditions or in the presence of binding partners.

Why has IP3R evolved to require that each monomer is bound to IP3 before activation? Intracellular Ca2+ is essential for almost all cellular processes but, paradoxically, has detrimental consequences if not fine-tuned to meet the cell’s physiological needs. Thus, eukaryotic cells have evolved very intricate Ca2+-signaling cascades with multiple layers of checks and balances to maintain strict control over intracellular Ca2+ concentrations. The requirement for maximum ligand occupancy for channel activation may ensure against unwarranted and potentially deleterious increases in intracellular Ca2+ (20).

Finally, the experimental approach used in this study can be extended to address many unanswered questions pertaining to IP3R biology, including the investigation of other ligand or binding factor stoichiometries, such as ATP, Ca2+, or calmodulin (7, 22). Similarly, an increasing number of IP3R mutations have been identified in human diseases, many of which clinically manifest in heterozygous patients (1417). We predict that the use of concatenated IP3R with the defined expression of monomers will be the method of choice to examine the pathophysiological roles of IP3R channels containing various combinations of wild-type and mutant subunits.



All reagents used for SDS–polyacrylamide gel electrophoresis (SDS-PAGE) were from Bio-Rad. DNA T4 ligase and restriction enzymes were purchased from New England Biolabs. Dulbecco’s modified Eagle’s medium (4.5g/liter d-glucose), RPMI 1640 media, G418 sulfate, penicillin/streptomycin, chicken serum, and β-mercaptoethanol were obtained from Life Technologies. Fetal bovine serum was purchased from Gemini. Protein A/G agarose beads were purchased from Santa Cruz Biotechnology. Rabbit antibodies recognizing IP3R1 (CT1) raised against the 19 C-terminal amino acids of rat IP3R1 and rabbit antibodies recognizing IP3R2 raised against amino acids 320 to 338 in mouse IP3R2 were generated by Pocono Rabbit Farm and Laboratories (31, 41). Mouse monoclonal antibody against residues 22 to 230 of human IP3R3 was from BD Transduction Laboratories. DyLight 800CW secondary antibodies were from Thermo Scientific. Fura2-AM was from TEFLabs. All other chemicals were obtained from Sigma unless otherwise indicated.

Plasmid construction

All constructs used in this study were based on rat IP3R1 or mouse IP3R2 cloned in pcDNA3.1. DNA modifications were made after QuikChange mutagenesis as described before (31). All polymerase chain reaction steps were carried out using Pfu Ultra II Hotstart 2× Master Mix (Agilent). Primers used in this study were synthesized by Integrated DNA Technologies and are listed in table S3. To generate IP3R1-R265Q, F1 and R1 primers were used. To introduce K508Q and R511Q mutations into IP3R1, F2 and R2 primers were used. To make mCherry-IP3R1 fusion protein, an Nhe I site was engineered immediately after the start codon in IP3R1 using F3 and R3 primers. To subclone mCherry coding sequence from pmCherry-C1 (Clontech), an Nhe I site was introduced at the end of coding sequence of mCherry using primer pair F4 and R4. The modified pmCherry-C1 plasmid was digested with Nhe I, and the Nhe I–Nhe I fragment flanking the mCherry coding region was inserted into the Nhe I site engineered before IP3R1 coding sequence. To add a FLAG tag to the C terminus of IP3R1, F5 and R5 were used. To mutate the ligand-binding domain IP3R2, F6 and R6 were used to introduce R568Q and K569Q mutations. Finally, F7 and R7 primer pair was used to make R2short. The coding sequences and desired DNA modifications were confirmed by sequencing. Concatenated IP3R subunits were created as described before (31, 32). Briefly, to make a concatenated R1R1 or R2R2 dimer, the corresponding IP3R cDNAs in pcDNA3.1(+) vector were modified to introduce unique restriction sites and extra nucleotides encoding the seven–amino acid linker (QLNQLQT), and two cDNAs coding for IP3R subunits were then ligated tail-to-head between the two arms of pJAZZ mamm linear vector based on coliphage N15 (Lucigen). The resulting construct encodes two IP3R subunits conjugated with a 14–amino acid linker. Similarly, concatenated tetrameric IP3R1 was made by linking four subunits to form one reading frame encoding four IP3R subunits with 14–amino acid linkers separating one subunit from the next. The coding sequences and desired DNA modifications were confirmed by sequencing.

Cell culture and transfection

DT40-3KO cell line and HEK293 cells were maintained accordingly as described before (2, 31). DT40-3KO cells were transiently transfected as follows: 5 million cells were pelleted by centrifugation and washed once with phosphate-buffered saline (pH 7.3) and then electroporated with 5 μg of plasmid DNA using Amaxa Cell Nucleofector Kit T (Lonza Laboratories). Cells were supplied with fresh complete RPMI medium [RPMI 1640 medium supplemented with 1% chicken serum, 10% fetal bovine serum, penicillin (100 U/ml), streptomycin (100 μg/ml)] and allowed to recover for 24 hours and then used for experiments. For the generation of stable cell lines, 24 hours after transfection, cells were plated in medium containing G418 (2 mg/ml) in five 96-well plates. After 10 to 14 days, Western blot analyses were used to screen G418-resistant clones for expression of the desired constructs. For HEK293 cell transfection, cells were transfected with cDNA constructs using Lipofectamine 2000 following the manufacturer’s protocol. Cells were used for experiment 24 to 48 hours after transfection. Cell lysates, SDS-PAGE, and coimmunoprecipitation were carried out as described before (40).

CRISPR-mediated disruption of human IP3R complements in HEK293 cells

Single-guide RNA (sgRNA) target sites in human ITPR1, ITPR2, and ITPR3 loci were identified using (29). Oligonucleotides corresponding to these sgRNAs, designed so that they can be annealed and cloned into pX458 vector (Addgene) that was digested with Bbs I, were synthesized by Integrated DNA Technology. The pX458 vector encodes Cas9 nuclease and enhanced green fluorescent protein (EGFP) (29). Because our preliminary experiments showed that simultaneously using two sgRNAs to target each IP3R was more effective in disrupting IP3R-encoding alleles, two sgRNAs were chosen targeting the third and fourth exons in the ITPR1 gene and the first and fourth exons in ITPR2 and ITPR3. pX458-sgRNAs were transfected into HEK293 cells. Forty-eight hours after transfection, EGFP-expressing cells were sorted and grown as single cells in 96-well plates. After 3 weeks, clones were screened by Western blotting using IP3R subtype–specific antibodies. Potential clones were genotyped as follows: (i) genomic DNA was isolated and amplified using a pair of primer flanking the CRISPR target site; (ii) amplicons were cloned into pcDNA3.1(+) and sequenced; (iii) six clones were sequenced, and sequencing data indicate the introduction of indels (insertion or deletion) in both alleles.

IP3-binding assay

IP3R constructs stably expressed in DT40-3KO cells were immunoprecipitated. The binding reaction consisted of 100-μl volume containing equal amounts of immunoprecipitated proteins, 2.5 nM tritiated IP3 (3H-IP3) in the presence or absence of increasing concentrations of unlabeled IP3. Tubes were incubated for 1 hour at 4°C with mixing every 10 min. Beads were then centrifuged at 13,000g, supernatants were removed, and 500 μl of 1% of SDS was added to each tube. After 12 to 24 hours, liquid scintillation counting was used to measure bound radioactivity. Nonspecific binding was calculated as the amount of bound radioactivity in the presence of 50 μM unlabeled IP3. Specific binding is determined by subtracting nonspecific binding obtained in parallel. All values were normalized to total specific binding obtained in the absence of unlabeled IP3. The averages of normalized values from three to four experiments were used to generate best fits. For Figs. 2A and 8B, specific binding was divided by the corresponding densitometric values, obtained from parallel analyses of immunoblots of equivalent amount of corresponding immunoprecipitated protein. These values were then normalized to that of IP3R1 (Fig. 2A) and R2R2 (Fig. 8B).

Single-cell imaging

Cytosolic Ca2+ changes were measured as described previously (31). Briefly, DT40-3KO cells expressing the indicated IP3R constructs were loaded in imaging buffer (137 mM NaCl, 0.56 mM MgCl2, 4.7 mM KCl, 10 mM Hepes, 5.5 mM glucose, 1.26 mM Ca2+, and 1 mM Na2HPO4 at pH 7.4) containing 2 μM Fura2-AM and were placed on a glass coverslip mounted in a Warner chamber for 20 min at 37°C. Experiments were performed at 37°C. TILLvisION software was used for image acquisition and analyses. Experiments were repeated at least three times.

Permeabilized cell assays

Permeabilized cell assays were carried out to assess unidirectional Ca2+ flux as described before (41). Briefly, cells were washed once with imaging buffer and incubated with 20 μM Furaptra-AM at room temperature for 1 hour and then permeabilized using 40 μM β-escin. Intracellular Ca2+ stores were loaded by adding 0.650 mM CaCl2, 1.4 mM MgCl2, and 1.5 mM Mg-ATP to activate the endoplasmic reticulum–localized Ca2+-ATPase SERCA. Upon loading, SERCA was disabled by removing MgCl2. IP3Rs were then activated by addition of the indicated concentration of IP3 in the presence of 5 mM ATP. The data were fit to a single exponential function to determine the initial release rate.

Homology modeling

A homology-based model was generated for R1QQQ using version 9.16 of Modeller (28) in default mode. As a template, we used an available crystal structure of the IP3R1 N-terminal region in the absence of IP3 (PDB ID: 3UJ4) (9). This structure is resolved to a higher resolution in this domain (3.0 Å) than the cryo-EM structure of the entire IP3R1 (4.7 Å), and the amino acid side chains are clearly resolved. Structural models were generated using UCSF Chimera (42) and PyMOL (Schrödinger).

Single IP3R1 channel measurements in isolated DT40-3KO nuclei

Isolation of nuclei and on-nucleus patch clamping were described before (33). Single IP3R channel potassium currents (ik) were measured in the on-nucleus patch clamp configuration using pCLAMP 9 and an Axopatch 200B amplifier (Molecular Devices). Pipette solution contained 140 mM KCl and 10 mM Hepes, with varying concentrations of IP3, ATP, BAPTA, and free Ca2+. Free concentrations of Ca2+ were calculated using Maxchelator freeware. Traces were consecutive 3-s sweeps recorded at −100 mV, sampled at 20 kHz, and filtered at 5 kHz. A minimum of 15 s of recordings was considered for data analyses. Pipette resistances were typically 20 megaohm and seal resistances were >5 gigaohm.

Data analysis

Single-channel openings were detected by half-threshold crossing criteria using the event detection protocol in Clampfit 9. We assumed that the number of channels in any particular cell is represented by the maximum number of discrete stacked events observed during the experiment. Even at low probability of opening (Po), stacking events were evident. Only patches with one apparent channel were considered for analyses. Po, unitary current (ik), open and closed times, and burst analyses were calculated using Clampfit 9 and Origin 6 software (Origin Lab). All-point current amplitude histograms were generated from the current records and fitted with a normal Gaussian probability distribution function. The coefficient of determination (R2) for every fit was >0.95. The Po was calculated using the multimodal distribution for the open and closed current levels. Channel dwell-time constants for the open and closed states were determined from exponential fits of all-point histograms of open and closed times. The threshold for an open event was set at 50% of the maximum open current, and events shorter than 0.1 ms were ignored. A “burst” was defined as a period of channel opening after a period of no channel activity that was greater than five times the mean closed time within a burst. The slope conductances were determined from the linear fits of the current-voltage relationships with the equationg=ik/(VVk)where g is unitary conductance, ik is unitary current, V is voltage, and Vk is the reversal potential. Ca2+ dependency curves were fitted separately for activation and inhibition with the logistic equationY=[(A1A2)/(1+(X/X0)P)]+A2


Fig. S1. Assessment of the Ca2+-release activities of R1Q-expressing DT40-3KO cells in response to endothelin and BCR stimulation.

Fig. S2. Generation of HEK293 cells with IP3R-null background using CRISPR/Cas9 technology.

Fig. S3. Ca2+-release activity of DT40-3KO cells expressing R1R1QQQ or R1QQQR1 in response to endothelin and BCR stimulation.

Fig. S4. Ca2+-release activity of IP3R1 channel incorporating two R1Q subunits.

Fig. S5. Ca2+-release activity of R1R1R1R1 in response to endothelin and BCR stimulation.

Fig. S6. IP3R1 channels assembled from concatenated IP3R1 have identical single-channel properties as that of wild-type IP3R1.

Fig. S7. Modulation of single-channel properties of concatenated IP3R1 by ATP.

Fig. S8. Modulation of single-channel properties of concatenated IP3R1 by calcium.

Fig. S9. Analysis of IP3R burst kinetics.

Fig. S10. Representative current amplitude histograms for IP3R1 activity.

Fig. S11. Assessment of R1R1QQQR1R1QQQ Ca2+-release activities in response to trypsin.

Fig. S12. Ca2+-release activities of R1R1R1R1QQQ in response to endothelin and BCR stimulation.

Table S1. Genomic analysis of HEK-3KO cells.

Table S2. Biophysical characteristics of IP3R and concatenated IP3R constructs.

Table S3. List of primers used in this study.


Funding: This work was supported by the NIH through RO1 DE014756 and DE019245 (to D.I.Y.). Author contributions: K.J.A., L.W., R.C., and L.E.W.II performed research and analyzed the data. F.V.P. performed homology modeling. D.I.Y. conceived and supervised the project and analyzed the data. K.J.A., L.E.W.II, F.V.P., and D.I.Y. wrote the paper. All authors read and approved the final manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: Constructs and cell lines will be made freely available for academic use after completion of a material transfer agreement between the University of Rochester and requesting institution.
View Abstract

Navigate This Article