Inositol 1,4,5-trisphosphate receptor subtype-specific regulation of calcium oscillations.

Oscillatory fluctuations in the cytosolic concentration of free calcium ions (Ca(2+)) are considered a ubiquitous mechanism for controlling multiple cellular processes. Inositol 1,4,5-trisphosphate (IP(3)) receptors (IP(3)R) are intracellular Ca(2+) release channels that mediate Ca(2+) release from endoplasmic reticulum (ER) Ca(2+) stores. The three IP(3)R subtypes described so far exhibit differential structural, biophysical, and biochemical properties. Subtype specific regulation of IP(3)R by the endogenous modulators IP(3), Ca(2+), protein kinases and associated proteins have been thoroughly examined. In this article we will review the contribution of each IP(3)R subtype in shaping cytosolic Ca(2+) oscillations.

Introduction

Intracellular calcium (Ca2+) dynamics play pivotal roles in numerous physiological processes, including fertilization, cell proliferation and differentiation, apoptosis, embryonic development, secretion, muscle contraction, immunity, brain function, chemical senses, and light transduction [1, 2]. Two main Ca2+ mobilizing systems co-exist in the cell: Ca2+ influx from the extracellular medium and Ca2+ release from internal stores. The inositol 1,4,5-trisphosphate (IP3) receptor (IP3R) is a tetrameric intracellular IP3-gated Ca2+ release channel that is predominantly located on the membrane of the endoplasmic reticulum (ER). It is present in almost all cell types and plays a crucial role in converting extracellular stimuli into intracellular signals [1, 3]. Upon extracellular stimulation by various agonists, such as hormones, growth factors, neurotransmitters, neurotrophins, odorants, and light, Phospholipase-C (PLC) is activated and phosphatidylinositol 4,5-bisphosphate is hydrolysated, generating IP3. IP3 binds to the IP3R, leading to the IP3-induced Ca2+ release (IICR) from the ER. Thirty years ago, Mikoshiba et al. found a mutant mouse with deficient Purkinje cells that had very low expression of P400, a glycoprotein that was later uncovered as one of the IP3R subtypes (IP3R1) [4, 5]. In 1989, Mikoshiba and co-workers were the first group to reveal that IP3R is a trans-membrane protein and determine the primary sequence of IP3R1, at the time the second largest molecule successfully cloned [6]. So far, three IP3R subtypes (IP3R1, IP3R2, IP3R3) as well as alternative splicing variants of IP3R1 and IP3R2 have been identified and cloned in mammals [1]. The expression patterns of the three subtypes are distinct but overlapping, and most cells express more than one subtype [7-9]. The three IP3R subtypes share 65–85% homology and can be separated into five functional domains [1, 10–12]. The NH2-terminal region contains a ligand coupling/suppressor domain, which suppresses IP3-binding activity and determines different IP3-binding affinity for each subtype [13], and an IP3-binding core domain that is the minimum region required for specific IP3 binding [14]. The ligand coupling/suppressor domain and the IP3-binding core are often referred to as the IP3-binding domain. Besides the IP3-binding domain is the internal coupling domain, which confers regulation by various intracellular modulators (Ca2+, calmodulin (CaM), ATP) and phosphorylation by several protein kinases (cAMP-dependent protein kinase (PKA), protein kinase C (PKC), cGMP-dependent protein kinase, Ca2+/CaM-dependent protein kinase II (CamKII), and tyrosine kinase) [3]. The COOH-terminal region has a six membrane-spanning channel domain and a short cytoplasmic COOH-terminal tail (CTT), called the “gatekeeper domain”, which is critical for IP3R channel opening [10]. The Ca2+ release activity of the IP3R channel is therefore regulated by many intracellular modulators (IP3, Ca2+, ATP, CaM), protein kinases, and IP3R-binding proteins [1, 3], leading to various spatiotemporal cytosolic Ca2+ patterns and diverse cellular responses [1, 2]. The relatively low homology in the three IP3R subtypes may underlie subtype-specific properties, that will affect Ca2+ signaling and in particular the spatiotemporal features of Ca2+ responses.

A prolonged elevation in the cytosolic Ca2+ concentration is considered toxic to the cell and in some cases may result in cell death. However, the cell can protect itself by temporally limiting the cytosolic Ca2+ elevation, often resulting in one of the most delicate patterns of Ca2+ signals, that being the oscillatory change in the cytosolic Ca2+ concentration, or Ca2+ oscillations [15-17]. Extensive studies over the past 30 years have revealed that cytosolic Ca2+ oscillations are ubiquitous and diverse cellular signals that control multiple processes in the cell. With cytosolic Ca2+ oscillations, cells not only avoid deleterious effects of sustained cytosolic Ca2+ concentrations, but also send out information encoded in the frequency and/or the amplitude of the oscillations to modulate cellular activity [15]. This review focuses on the separate role of the IP3R subtypes in generating Ca2+ oscillations and on the molecular mechanisms responsible for the specific role of each subtype in regulating this ubiquitous signal.

A General Mechanism Generating Ca2+ Oscillations Based on Regulation of IP3R

Many studies have indicated that IP3R is involved in generating cytosolic Ca2+ oscillations [15-17]. For instance, the FGF-induced Ca2+ oscillations in mice fibroblasts are inhibited by an IP3R antagonist [18]. Ca2+ oscillations are thought to arise due to periodic release of Ca2+ from intracellular Ca2+ stores via IP3R [19]. Early studies using reconstituted IP3R in lipid bilayers have indicated that Ca2+ can both activate and inhibit IP3R [20, 21]. The IP3R is activated at low cytosolic Ca2+ concentrations, elevating the cytosolic Ca2+ concentration through a process often referred to as Ca2+-induced Ca2+ release (CICR). High cytosolic Ca2+ concentration can instead inhibit IP3R, leading to a decrease in intracellular Ca2+ release. In vivo, the binding of IP3 together with fluctuating cytosolic Ca2+ concentrations can trigger successive cycles of IP3R activation and inhibition, which result in cytosolic Ca2+ oscillations. Accordingly, Ca2+ oscillations can be produced by application of IP3 to permeabilized hepatocytes [22] and blowfly salivary gland cells [23] and by injecting IP3 analogs into fertilized ascidians eggs [24]. Moreover, DT40 cells expressing a mutant IP3R with reduced sensitivity to Ca2+ do not exhibit Ca2+ oscillations upon application of cross-linked B-cell receptors [25]. Finally, thimerosal, which sensitizes IP3R for lower IP3 levels, potentiates IP3-induced Ca2+ oscillations in sea urchin eggs [26]. These data, together with mathematical models [27, 28], have confirmed that the cross-talk between Ca2+ and IP3 in regulating the IP3R is critical for generating Ca2+ oscillations. However, in Madin-Darby canine kidney epithelial cells [29] and Chinese hamster ovary cells [30-32], each peak of the oscillatory Ca2+ signal is preceded by elevated IP3, as measured by means of a pleckstrin homology domain of PLC-δ1 tagged with a fluorescent protein indicator. Therefore, it has been proposed that dynamic IP3 production may produce cytosolic Ca2+ oscillations. Nevertheless, other studies using different cells and methods reported opposite conclusions [33-35]. For example, expression of an IP3 binding domain of IP3R1 together with two different fluorescent proteins in HeLa cells does not reveal fluctuations in the intracellular IP3 concentration during Ca2+ oscillations [33].

Subtype Specificity of Ca2+ Oscillations

Numerous studies using cells endogenously or exogenously expressing single or combined IP3R subtypes indicate that the subtle distinctions in the properties of each subtype contribute differently to the regulation of cytosolic Ca2+ oscillations [3].

Miyakawa et al. [36] first described IP3R subtype-specific Ca2+ oscillations using genetically engineered B cells that express either single or combined IP3R subtypes. They found that Ca2+-signaling patterns depend on the expression levels of IP3R subtypes, probably because of their specific response to endogenous modulators, such as IP3, Ca2+ and ATP. IP3R2 is the most sensitive to IP3 and is required for robust, long lasting, and regular Ca2+ oscillations that occur upon activation of B-cell receptors. IP3R1 mediates less regular Ca2+ oscillations. IP3R3 is the least sensitive to IP3 as well as Ca2+ and generates only monophasic Ca2+ transients. Morel et al. [37] examined the roles of IP3R1 and IP3R2 in Ca2+ oscillations using vascular myocytes and found that acetylcholine induces Ca2+ oscillations in cells expressing both subtypes, and fails to do so in cells expressing only IP3R1. The oscillations are inhibited by intracellular infusion of heparin, anti-IP3R2 antibody or antisense oligonucleotides targeting IP3R2, suggesting that the IP3R2 subtype is required for acetylcholine-induced Ca2+ oscillations in vascular myocytes. Using HeLa cells, which express comparable amounts of IP3R1 and IP3R3, Mikoshiba and co-workers showed that knockdown of IP3R1 terminates Ca2+ oscillations, whereas knockdown of IP3R3 leads to more robust and long lasting Ca2+ oscillations [38]. These IP3R3 knockdown effects were similar in COS-7 cells that predominantly express IP3R3, suggesting that IP3R3 functions as an anti Ca2+-oscillatory unit. Almirza et al. reported similar results using normal kidney fibroblasts, which expresses IP3R1 and IP3R3 [39]. When IP3R1 or IP3R3 are knocked-down, the frequency of prostaglandin F2α-induced Ca2+ oscillations is significantly decreased or increased, respectively. In NIH-3T3 cells, which predominantly express IP3R2 and IP3R3, ATP activates Ca2+ oscillations [40]. Ca2+ oscillations were induced by application of carbachol in AR4-2J cells, which predominantly expresses IP3R2, and in HEK293A cells in which both IP3R1 and IP3R3 were knocked-down [41]. The contribution of IP3R2 to Ca2+ oscillations is further confirmed by the fact that IP3-dependent Ca2+ oscillations were abolished in osteoclasts of IP3R2 knockout mice [42]. In rat insulinoma RINm5F cells, which almost exclusively express IP3R3, application of carbachol or EGF, two agonists that activate PLC through different receptors, or application of IP3 to permeabilized cells, elicit transient Ca2+ release and does not induce Ca2+ oscillations [43]. Several reports, including mathematical modeling studies, have indicated that the specific intracellular localization of the IP3R is crucial for the generation of Ca2+ oscillations [44-46]. For instance, Kim et al. [45] found that HL-1 cells derived from mouse cardiac myocytes express both IP3R1 and IP3R2. IP3R1 is expressed diffusely in the perinucleus and IP3R2 is expressed in the cytosol with a punctuated distribution. Both application of ATP to intact cells and direct introduction of IP3 into permeabilized cells evoke IP3-dependent transient intracellular Ca2+ release accompanied by Ca2+ oscillations. The magnitude of Ca2+ oscillations is significantly larger in the cytosol than in the nucleus, while the monophasic Ca2+ transient is more pronounced in the nucleus. These results suggest that subtype specificity as well as specific localization of the IP3R contribute to distinct local Ca2+ signaling. Altogether, these data suggest that IP3R1 and IP3R2, in particular IP3R2, crucially contribute in generating Ca2+ oscillations, whereas IP3R3 is an anti-oscillatory unit. Nevertheless, in A7r5 cells derived from rat embryonic thoracic aorta muscle cells, which express IP3R1 and IP3R3, knockdown of IP3R1 only reduces the frequency of arginine vasopressin-induced Ca2+ oscillations without affecting the number of cells exhibiting Ca2+ oscillations [47]. Moreover, both acetylcholine and cholecystokinin octapeptide activate IP3R2- and IP3R3-dependent Ca2+ oscillations in pancreatic acinar cells. However, unlike IP3R2-dependent oscillations, the amplitude of IP3R3-dependent oscillations decreases throughout the stimulation [48]. The IP3R subtype-specific Ca2+ oscillations are summarized in Table 1.

Table 1

The occurrence of Ca2+ oscillations and the expression of the different IP3R subtypes

Cell type	IP3R1	IP3R2	IP3R3	Activator	Ca2+ oscillations	Reference
DT40	+	−	−	B cell receptor	↑	[36]
DT40	−	+	−	B cell receptor	↑	[36]
DT40	−	−	+	B cell receptor	↓	[36]
Vascular myocytes	+	+	−	Acetylcholine	↑	[37]
Vascular myocytes	+	−	−	Acetylcholine	↓	[37]
HeLa	++	+	++ (kd)	ATP	↑	[38]
HeLa	++ (kd)	+	++	ATP	↓	[38]
COS-7	+	+	++ (kd)	ATP	↑	[38]
NRK	+	−	+ (kd)	Prostaglandin F 2α	↑	[39]
NRK	+ (kd)	−	+	Prostaglandin F 2α	↓	[39]
NIH-3T3	+	++	++	ATP	↑	[40]
AR4-2 J	+	++	+	IGF-1	↑	[41]
HEK293A	+ (kd)	+	+ (kd)	Carbachol	↑	[41]
Osteoclasts	+	+	+	RANKL	↑	[42]
Osteoclasts	+	−	+	RANKL	↓	[42]
Osteoclasts	+	−	−	RANKL	↓	[42]
RINm5F	±	±	++	Carbachol, EGF, IP3	↓	[43]
HL-1	+	+	−	ATP	↑	[45]
A7r5	+ (kd)	−	+	Arginine vasopressin	↓	[47]
Pancreatic acinar cells	+	+	+	Acetylcholine, CCK-OP	↑	[48]
Pancreatic acinar cells	+	−	+	Acetylcholine, CCK-OP	↑	[48]

+ high expression, − low expression, and kd, knock down or low expression

The IP3R exists as a homo- or hetero-tetrameric complex to form a functional Ca2+ release channel [49-51]. The influence of homo- or hetero-tetrameric channels on intracellular Ca2+ oscillations has been investigated. Studies on genetically engineered DT40 cells that express a single IP3R subtype and therefore a homo-tetrameric receptor demonstrate Ca2+ oscillations [36]. Cells with all subtypes, which should at least partially express hetero-tetrameric IP3Rs, also exhibit Ca2+ oscillations [37–43, 45, 47]. Taken together these data suggest that both homo- and hetero-tetrameric IP3Rs can generate intracellular Ca2+ oscillations.

In conclusion, it appears that IP3R subtype-specific expression crucially shapes cytosolic Ca2+ signaling patterns. IP3R2 is the main pro-oscillatory subtype, whereas IP3R1 can induce a transient Ca2+ signal or an oscillatory Ca2+ signal. IP3R3 mainly shows an anti-oscillatory behavior, but could underlie short-term oscillations depending on the cell type and stimulus. Further characterization of homo- and hetero-tetrameric IP3R-dependent Ca2+ oscillations are needed for fully understanding the intricacies of each IP3R subunit in shaping Ca2+ oscillations.

Subtype Specificity of IP3-Binding Affinity to IP3R

As summarized earlier, cytosolic Ca2+ oscillations are IP3R subtype-dependent. IP3 and Ca2+ are the two key modulators of IP3R and the distinct subtype properties determine the diverse regulatory effects. Each subtype has different IP3 binding affinity. Sudhof et al. were first to report, using an equilibrium IP3 binding assay, that the order of IP3-binding affinity was IP3R2 > IP3R1 > IP3R3 [9, 52]. Applying the same method, Wojcikiewicz et al. [53] and Nerou et al. [54] later claimed a different order, IP3R1 > IP3R2 > IP3R3. Mikoshiba and co-workers performed a detailed molecular analysis of the IP3 binding affinity of all three subtypes [11, 13]. They found that the IP3-binding affinities of purified IP3-binding domains are close to the intrinsic IP3-binding affinity of all three IP3R subtypes, and describe the following order IP3R2 > IP3R1 > IP3R3. They also showed that IP3-binding core fragments, which do not contain the ligand coupling/suppressor domain, display an almost identical IP3-binding affinity for all three subtypes. By a serious and compelling molecular analysis, they concluded that the functional diversity in ligand sensitivity among IP3R subtypes arises from structural differences in the ligand coupling/suppressor domain, which attenuate the IP3-binding affinity of the IP3-binding core domain through an intramolecular mechanism. Tu et al. recorded single-channel activities of the recombinant IP3R1, IP3R2, and IP3R3 reconstituted into planar lipid bilayers [55]. This report had a similar conclusion with IP3R2 showing the highest apparent IP3-affinity, followed by IP3R1, and then by IP3R3.

Differences amongst IP3R subtypes in terms of IP3-binding affinities do not reflect intrinsic differences in the properties of the channels to regulate Ca2+ oscillations. Instead differences in the state of phosphorylation and/or association with interacting proteins exist. Nevertheless, IP3R1 and IP3R2 are most sensitive to IP3, a property that could contribute in their function as Ca2+ oscillatory unit. The exact contribution of subtype specific IP3-binding affinities on Ca2+ oscillations remains to be further investigated.

Subtype Specificity of Ca2+ Inhibition and Induction

As mentioned earlier, repeated activation and inhibition of IICR by fluctuating cytosolic Ca2+ levels have been proposed as central molecular mechanisms for IP3R-dependent Ca2+ oscillations [56]. Several stimulatory and inhibitory Ca2+ binding sites on the IP3R have been identified and characterized. For instance, two sites are localized in the IP3 binding core and another site is located close to the transmembrane domain [57], exemplifying the complex synergy between IP3 and Ca2+ in the regulation of the IICR [12]. Ca2+ regulation of IP3R activity may result in changed IP3 binding affinity, alteration of channel open probability, or indirect influence on IP3R associated proteins, such as the CaM. Interestingly, this can occur specifically on one IP3R subtype, making Ca2+ regulation of IICR one of the major mechanisms to produce versatile signals, as confirmed by mathematical modeling studies [58].

The complex regulation of the IP3R subtypes′ activity by Ca2+ has been recently reviewed in detail [3] and we will therefore mainly focus on how Ca2+ itself modulates Ca2+ oscillations. Everyone in the field agrees that all three subtypes are activated by Ca2+. Inhibition of the IP3R by Ca2+, however, is more controversial. In single channel studies, each subtype is inhibited by high Ca2+ concentrations, even though the threshold and speed of inhibition differs [3]. Moreover, Ca2+ inhibition of IICR sometimes depends on the addition of an extra factor, for example ATP for IP3R3 [55]. Therefore, all three subtypes can potentially support Ca2+ oscillations based on the model described previously, where concerted actions of IP3 and Ca2+ stimulates IP3R. Accordingly, IP3R1-, IP3R2-, and IP3R3-dependent Ca2+ oscillations have been observed, although IP3R3-dependent Ca2+ oscillations are less likely to occur and are also less frequently observed (see previous sections).

In most cases, cells express more than one IP3R subtype. Interestingly, when several IP3R subtypes are expressed, one of them becomes dominant regarding Ca2+ regulation of IICR [36]. This result also calls for caution when drawing conclusions on the subtype specificity of Ca2+ signaling, since expression of even a small amount of one subtype could critically affect the Ca2+ signaling pattern [59].

Taken together, Ca2+ activation and inhibition properties of IP3R1 and IP3R2 make them likely to support Ca2+ oscillations in physiological conditions [60], whereas specific cellular circumstances are required for activation of IP3R3-dependent Ca2+ oscillations.

Subtype Specificity of Phosphorylation of IP3R

Phosphorylation of the IP3R is involved in many Ca2+ signaling pathways [61] and the different subtypes are interacting with protein kinases and phosphatases differently [62]. Many of the phosphorylation sites are subtype-specific, increasing the diversity in regulatory fine tuning of Ca2+ oscillations. The functional consequences of these regulatory modifications are only partially understood, and in some cases remain controversial. Therefore we will here focus on those protein kinases known to modulate Ca2+ oscillations through phosphorylation of IP3R.

PKA-dependent phosphorylation of IP3R has been demonstrated extensively. Phosphomimetic mutations of IP3R1 expressed in DT40 cells showed that PKA-mediated phosphorylation decreases the threshold for Ca2+ oscillations, without affecting the amplitude or frequency [63]. PKA phosphorylates two distinct sites in IP3R1 internal coupling domain (S1588 and S1755) [64]. Although these sites are not conserved in IP3R2 and IP3R3, PKA-dependent phosphorylation of these subtypes has been demonstrated [65]. In parotid acinar cells [66] and the pancreatic AR4-2J cell line [67], PKA directly phosphorylates IP3R2, dramatically potentiating Ca2+ release. Interestingly, raising cAMP during sub-threshold agonist stimulation resulted in an oscillatory Ca2+ signal, while raising cAMP during an Ca2+ oscillation converted the response into a peak and plateau-like signal [66], probably because of a shift in the concentration dependency in IICR. CaMKII has been proposed to be involved in the control of the Ca2+-dependent regulation of IICR and in the occurrence of Ca2+ oscillations [68]. The most extensive information regarding CaMKII regulation of IP3R is derived from studies performed on IP3R2 [69, 70], which is the predominant subtype in cardiac ventricular myocytes. CaMKII-dependent phosphorylation significantly decreased the open probability of IP3R2 in lipid bilayers, which suggests a Ca2+-dependent negative feedback mechanism on IP3R2 activity in the cardiomyocyte nuclear envelope [71]. This may also result in a Ca2+-dependent inhibitory loop of Ca2+ oscillations [72]. Functional effects of PKC-mediated phosphorylation of the IP3R were first studied in isolated rat liver nuclei [73]. PKC-mediated phosphorylation of IP3R1 in vitro is in addition regulated by Ca2+ and CaM [74]. As both Ca2+ and CaM inhibit the PKC-mediated phosphorylation of IP3R1, it is possible that this process may contribute to the negative slope of the Ca2+-dependent bell-shaped regulation of IP3R by Ca2+, consequently affecting Ca2+ oscillations. Recent demonstrations suggest a role for PKC-mediated phosphorylation of IP3R2 [75] and IP3R3 [43]. These reports show that when IP3R2 or IP3R3 are phosphorylated by PKC, IP3-dependent Ca2+ oscillations are decreased in cells expressing only those subtypes. Thus, PKC may act as a subtype specific regulator of IP3R-mediated cytosolic Ca2+ oscillations. These differences are not unexpected since IP3R subtypes possess different potential phosphorylation sites [43, 76]. How phosphorylation of IP3R subtypes by distinct protein kinases affect Ca2+ oscillations are summarized in Table 2.

Table 2

The IP3R subtype specificity of protein kinases and IP3R-associated proteins and their modulating effects on Ca2+ oscillations

Effect	Stimulatory			Inhibitory
Effectors	IP3R1	IP3R2	IP3R3	IP3R1	IP3R2	IP3R3
PKC	[73, 74]	–	–	–	[75]	[43]
PKA	[63]	[67]	–	–	[67]	–
CaMKII	–	–	–	[68][72]	–	–
CaBP1	–	–	–	[82]	–	[82]
Na,K-ATPase	[83, 84]	[83, 84]	[83, 84]	–	–	–
IRBIT	–	–	–	[85]	–	–
AKAP9	[87]	–	–	–	–	–
PS1/PS2	[88, 89]	[88, 89]	[88, 89]	–	–	–
ERp44	–	–	–	[90]	–	–
GRP78	[91]	–	–	–	–	–
Bcl-2	[78, 92, 94, 96]	[78, 92, 94, 96]	[78, 92, 94, 96]	[97]	[97]	[97]
Bcl-XL	[93, 95]	[93, 95]	[93, 95]	–	–	–
Mcl-1	[78]	[78]	[78]	–	–	–
Cytochrome C	[98, 99]	–	–	–	–	–
GIT1/GIT2	–	–	–	[81]	[81]	[81]
NCS-1	[80, 100]	[80, 100]	[80, 100]	–	–	–

The subtype specific regulation of IP3R by phosphorylation and its relation to Ca2+ oscillations are not fully understood. These processes are likely to be dependent on specific IP3R subtypes expression levels and protein kinases activation, and need to be further investigated.

Regulation of IP3R Activity by Accessory Proteins

About forty proteins have been reported to interact with IP3R, most of which modulate IP3R channel activity [1, 3, 77–81]. There is a lack of data regarding IP3R2 and IP3R3 specific binding proteins since most of these proteins are identified by co-immunoprecipitation studies with one or two IP3R subtypes or using IP3R1 probes. Few reports show that some of these associated proteins modulate Ca2+ oscillations differently. Therefore, we summarize here the proteins that bind to IP3R and modulate Ca2+ oscillations, whether they bind to a specific IP3R subtype or not (Table 2; Fig. 1).

Fig. 1

Cartoon illustrating the three IP3R subtypes (IP3R1, IP3R2, and IP3R3) and related protein kinases and interacting proteins involved in the regulation of cytosolic Ca2+ oscillations

CaBP1, one of the neuronal Ca2+ binding proteins, was co-immunoprecipitated with IP3R1 and IP3R3 [82]. The CaBP1-binding site was mapped in the ligand coupling/suppressor domain of IP3R1. This interaction functionally inhibits IP3-dependent Ca2+ oscillations in COS-7 cells expressing CaBP1, in permeabilized COS-7 cells exposed to recombinant CaBP1, and in Xenopus oocytes injected with recombinant CaBP1.

Na,K-ATPase, a plasma membrane ion pump, directly binds to the IP3 binding-domain of all three IP3R subtypes through its NH2-terminal tail [83, 84]. In the presence of ouabain, Na,K-ATPase triggers IP3-dependent Ca2+ oscillations in COS-7 cells and in primary culture of rat renal proximal tubule cells. Overexpression of a peptide corresponding to the wild type NH2-terminal tail of Na,K-ATPase decreased the number of cells exhibiting Ca2+ oscillations, an effect not observed when a mutant type that does not bind to IP3R was used.

IRBIT was identified to bind to the IP3 binding core of IP3R1 [85]. This interaction suppresses the activation of IP3R by regulating the IP3 sensitivity of IP3R1. Knockdown of IRBIT in HeLa cells increases ATP-induced cytosolic Ca2+ oscillations.

AKAP9, one of the neuronal PKA-anchoring adaptor proteins, binds to the leucine/isoleucine zipper (LIZ) motif in the internal coupling domain of IP3R1 [86]. Expression of a 36-residues LIZ fragment, which can disrupt the IP3R1-AKAP9 association, reduces the frequency of Ca2+ oscillations induced by application of dopamine in primary culture of medium spiny neuron [87].

Presenilins (PS), including PS1 and PS2, are proteins bound to the gamma-secretase protease complex. Mutations in the genes encoding PS1 and PS2 are the major cause of familial Alzheimer’s disease (FAD). Wildtype and FAD-mutants of PS1 and PS2 have been co-immunoprecipitated with IP3R1 and IP3R3 [88, 89]. These interactions exert profound stimulatory effects on the IP3R gating activity. Mutated PSs were demonstrated to increase frequency of both spontaneous Ca2+ oscillations and Ca2+ oscillations triggered by cross-linking the B cell receptor with IgM antibody in both DT40 cells and FAD patient B cells.

ERp44 is an ER lumenal protein of the thioredoxin family. Depending on the oxidative status in the ER lumen, it can interact directly with the third IP3R1 lumenal loop and inhibit its activity [90]. Knockdown of ERp44 in HeLa cells increases ATP-triggered cytosolic Ca2+ oscillations.

GRP78, another ER lumenal protein, also interacts with the third lumenal loop of the IP3R1 [91]. In contrast to ERp44, GRP78 enhances IP3R1 channel activity. Knockdown of GRP78 in HeLa cells decreases ATP-triggered Ca2+ oscillations, which is restored by re-expression of the protein.

Bcl-2, Bcl-XL, and Mcl-1, three anti-apoptotic proteins that belong to Bcl-2 family, have been reported to bind to the CTT and/or the internal coupling domain of all three IP3R subtypes [78, 92–95]. Bcl-2 enhances IP3-mediated Ca2+ oscillations induced by T cell receptor activation in WEHI7.2 cells, Jurkat cells, and wild type DT40 cells [78, 92, 94, 96], whereas Ca2+ oscillations induced by serum withdrawal in NIH-3T3 murine fibroblasts are dampened [97]. Expression of Bcl-XL in wild type DT40 cells or in DT40 cells engineered to express each IP3R subtype increases the number of the cells exhibiting Ca2+ oscillations as well as the oscillatory frequency [93, 95]. Interaction of Mcl-1 with IP3R increases the number of DT40 cells exhibiting anti-B cell receptor antibody induced Ca2+ oscillations [78]. Bcl-2 and Mcl-1 also increase the number of cells exhibiting Ca2+ oscillations and the amplitude and/or the frequency of spontaneous Ca2+ oscillations in DT 40 cells [78].

Cytochrome C, one of the key components of the apoptotic cascade, was found to selectively and directly bind to IP3R1 CTT during early apoptosis via a cluster of glutamic acid residues (binding to IP3R2 and IP3R3 were not confirmed), resulting in staurosporine-induced sustained Ca2+ oscillations [98, 99].

G-protein-coupled receptor kinase-interacting proteins (GIT), including GIT1 and GIT2, bind to the CTT of all three IP3R subtypes, but have stronger binding affinity to IP3R2 (more than 10- and 20-fold as compared to IP3R1 and IP3R3, respectively), and inhibit IICR [81]. Knockdown of GIT proteins in HeLa or COS-7 cells increases the number of cells exhibiting Ca2+ oscillations.

Neuronal Ca2+ sensor 1 (NCS-1), a Ca2+ binding protein whose expression could be enhanced by application of Taxol, a natural product for the treatment of solid tumors, was co-immunoprecipitated with all three subtypes of IP3R [80, 100]. The NCS-1-IP3R interaction increases the number of cells exhibiting IP3R-dependent Ca2+ oscillations in SH-SY5Y human neuroblastoma cells [100] and the frequency of spontaneous Ca2+ oscillations in rat ventricular cardiomyocytes [80].

The diversity in distribution of associated proteins and/or IP3R subtypes is essential for the versatility of IP3R subtype-dependent Ca2+ oscillations in different cell types. More information, however, is required for determining the individual role of each separate subtype in modulating cytosolic Ca2+ oscillations.

Conclusion and Future Directions

It is evident that the different IP3R subtypes are regulated by a large number of cellular mechanisms that varies in a cell type-specific manner. In this review we have focused on IP3R subtype-specific modulation of Ca2+ oscillations. Ca2+ oscillations are repetitive increases in the cytosolic Ca2+ concentration that are used by the cell to convey information within or between cells. The oscillatory Ca2+ signal is known to be initiated at the onset of fertilization [101-103] and to continue throughout life to control a vast array of cellular processes as diverse as proliferation, differentiation, development, learning and memory, contraction, secretion, and cell death [1, 15]. Altered intracellular Ca2+ signaling has been linked to many diseases, such as Hungtington’s, Alzheimer’s and Parkinson’s diseases, amyotrophic lateral sclerosis, schizophrenia, spinocerebellar ataxias, heart failure, polycystic kidney disease, and human immunodeficiency virus infection [104-107]. It is therefore essential to determine the molecular mechanisms involved in the generation of intracellular Ca2+ oscillations. Additionally, Ca2+ oscillations are known to encode information in their frequency and amplitude to activate various specific downstream targets [15-17]. Efforts to understand the nature of these “cellular radio signals” started at the same time as Ca2+ oscillations were discovered and have resulted in a large number of publications [16–18, 22, 28, 30, 32, 35, 37, 39, 42, 47, 56, 60, 83, 100–102], most of which is cell type- and agonist-specific. To determine the associations between (1) stimulus, (2) Ca2+ oscillation, and (3) activation of a specific downstream cellular process, future studies will have to consider the molecular partners involved in each step. The recent rapid development of sophisticated molecular and genetic tools, such as small interfering RNA [108] and optogenetics [109], will surely advance our future knowledge about IP3R subtype-specific regulation of Ca2+ oscillations.