Inositol 1,4,5‐trisphosphate receptors and their protein partners as signalling hubs

Abstract

Inositol 1,4,5‐trisphosphate receptors (IP₃Rs) are expressed in nearly all animal cells, where they mediate the release of Ca²⁺ from intracellular stores. The complex spatial and temporal organization of the ensuing intracellular Ca²⁺ signals allows selective regulation of diverse physiological responses. Interactions of IP₃Rs with other proteins contribute to the specificity and speed of Ca²⁺ signalling pathways, and to their capacity to integrate information from other signalling pathways. In this review, we provide a comprehensive survey of the proteins proposed to interact with IP₃Rs and the functional effects that these interactions produce. Interacting proteins can determine the activity of IP₃Rs, facilitate their regulation by multiple signalling pathways and direct the Ca²⁺ that they release to specific targets. We suggest that IP₃Rs function as signalling hubs through which diverse inputs are processed and then emerge as cytosolic Ca²⁺ signals.

Abbreviations

AC

adenylyl cyclase

B₂R

type 2 bradykinin receptor

cAMP

cyclic adenosine monophosphate

CREB

cAMP response element‐binding protein

EB3

end‐binding protein 3

ER

endoplasmic reticulum

GPCR

G protein‐coupled receptor

IBC

IP₃‐binding core

IP₃

inositol 1,4,5‐trisphosphate

IP₃R

IP₃ receptor

IRBIT

IP₃R‐binding protein released with IP₃

M₁R

type 1 muscarinic acetylcholine receptor

PKA

protein kinase A

PLC

phospholipase C

SD

suppressor domain

TMD

transmembrane domain

Introduction

Ca²⁺ signals within cells are spatially and temporally intricate, allowing them to elicit a multitude of specific downstream effects (Berridge, 2009). Inositol 1,4,5‐trisphosphate receptors (IP₃Rs), the most widely expressed class of intracellular Ca²⁺ channel, release Ca²⁺ from intracellular stores in response to binding of IP₃ and Ca²⁺ (Foskett et al. 2007; Taylor & Tovey, 2010). Dual regulation of IP₃Rs by two essential stimuli, IP₃ and Ca²⁺, is important because it endows IP₃Rs with a capacity to propagate Ca²⁺ signals regeneratively by Ca²⁺‐induced Ca²⁺ release, as Ca²⁺ released by an active IP₃R ignites the activity of adjacent IP₃Rs that have bound IP₃ (Smith & Parker, 2009). This in turn plays a key role in defining the spatial organization of IP₃‐evoked Ca²⁺ signals.

Activation of IP₃Rs is initiated by binding of IP₃ within a clam‐like structure, the IP₃‐binding core (IBC) (Bosanac et al. 2002), located near the N‐terminus of each IP₃R subunit. Binding of IP₃ causes a conformational change that rearranges the association of the IBC with the N‐terminal suppressor domain (SD). These changes are proposed to disrupt interactions between the N‐terminal regions of the four subunits of the IP₃R, leading to opening of the channel. The latter is formed by transmembrane domains (TMDs) towards the C‐terminus of each IP₃R subunit (Seo et al. 2012) (Fig. 1). It is not yet clear where binding of Ca²⁺ to the IP₃R lies within the sequence of events linking binding of IP₃ to channel gating. One possibility is that the conformational changes evoked by binding of IP₃ expose a site to which Ca²⁺ must bind before the channel can open (Marchant & Taylor, 1997; Foskett et al. 2007). However, neither the structural identity of this stimulatory Ca²⁺‐binding site, nor that of the inhibitory site through which higher concentrations of Ca²⁺ inhibit IP₃Rs have been resolved. The inhibitory site may reside on an accessory protein associated with IP₃Rs.

Figure 1. Association of proteins with IP₃Rs .

Key functional domains of a single IP₃R subunit are shown: the suppressor domain (SD), IP₃‐binding core (IBC), cytosolic regulatory domain, transmembrane domains (TMDs) and the cytosolic C‐terminus (CT). The sites to which proteins are proposed to bind are shown. Many additional proteins are thought to associate with IP₃Rs, but the binding sites have not been identified. Abbreviations and references are provided in Tables 1, 2, 3, 4.

IP₃Rs are present in almost all animal cells and some protozoa (Prole & Taylor, 2011), but there are no homologous proteins in plants (Wheeler & Brownlee, 2008) or fungi (Prole & Taylor, 2012). The genomes of vertebrates encode three subtypes of IP₃R subunit (IP₃R1–3), which can form homo‐tetrameric or hetero‐tetrameric channels (Joseph et al. 1995) with differing properties and distributions (Foskett et al. 2007; Mikoshiba, 2007). In mammalian cells, IP₃Rs have been reported to release Ca²⁺ mainly from the endoplasmic reticulum (ER) (Streb et al. 1984; Volpe et al. 1985), but the Golgi apparatus (Pinton et al. 1998) and secretory vesicles (Yoo, 2011) also respond to IP₃. Although IP₃ initiates Ca²⁺ signals by stimulating Ca²⁺ release from intracellular stores, the signals are sustained by Ca²⁺ entry across the plasma membrane. That too is indirectly regulated by IP₃, because store‐operated Ca²⁺ entry is stimulated by loss of Ca²⁺ from the ER (Parekh & Putney, 2005; Lewis, 2012). Ca²⁺ signals initiated by IP₃Rs evoke a wide variety of cellular events, ranging from embryological development (Kume et al. 1997; Uchida et al. 2010) to cellular metabolism (Cardenas et al. 2010), gluconeogenesis (Wang et al. 2012), exocrine secretion (Futatsugi et al. 2005) and neuronal function (Matsumoto et al. 1996).

Specificity within Ca²⁺ signalling pathways, or indeed any signalling pathway (Scott & Pawson, 2009; Scott et al. 2013), is achieved, in part, by the formation of macromolecular signalling complexes. Within the signalling pathways that involve phospholipase C (PLC), these complexes regulate the activity of IP₃Rs, their distribution, and their association with both the plasma membrane receptors that evoke IP₃ formation and the downstream targets of the Ca²⁺ released by IP₃Rs (Konieczny et al. 2012). The interactions of IP₃Rs with other proteins have been reviewed previously (Choe & Ehrlich, 2006; Foskett et al. 2007; Mikoshiba, 2007; Vanderheyden et al. 2009 a), but continued progress and the advent of high‐throughput proteomics methods (Havugimana et al. 2012; Rolland et al. 2014) suggest that an update is timely.

Searches of proteomic databases and published literature reveal a large number of proteins that form complexes with IP₃Rs (Tables 1, 2, 3, 4). For some of these proteins, the regions within IP₃Rs that are important for the interaction have been mapped (Fig. 1). At the outset, it is worth sounding some notes of caution regarding the reported interactions. Firstly, it is often difficult to establish that two proteins interact directly, rather than via intermediate proteins. Many of these complexes may, therefore, be formed by direct or indirect interactions of IP₃Rs with other proteins. For example, association of protein phosphatase 1 with IP₃Rs may be mediated in part by IRBIT (IP₃R‐binding protein released with IP₃), which binds directly to both proteins (Ando et al. 2014). Secondly, the interactions and their effects may depend on the cellular context, including such factors as the subtype of IP₃R, the physiological status of the IP₃R (e.g. phosphorylation), the cell type and the expression levels of the interacting proteins and IP₃Rs. Thirdly, interactions that occur in cellular lysates may be precluded within intact cells. For example, the interaction of two proteins may be prevented by their physical separation within the cell or by mutually exclusive binding of other proteins or ligands. IRBIT, for example, binds to IP₃R subunits only when they have no IP₃ bound. Lastly, some forms of experimental evidence are more discriminating than others, and it will be necessary to verify the putative interactions indicated by methods such as yeast two‐hybrid screening and mass spectrometry.

Table 1.

Proteins that form complexes with IP₃Rs and enhance their activity

Protein	References
Effective delivery of messengers
Adenylyl cyclase 6 (AC6)	Tovey et al. 2008
Bradykinin receptor B2 (B2R)	Delmas et al. 2002; Jin et al. 2013
Epidermal growth factor receptor (EGFR)	Hur et al. 2005
Erythropoietin receptor (EPO‐R)	Tong et al. 2004
Glyceraldehyde‐3‐phosphate dehydrogenase (GAPDH)	Patterson et al. 2005
Metabotropic glutamate receptor 1 (mGluR1;GRM1)	Tu et al. 1998
Phospholipase C‐β1 (PLCβ1)	Shin et al. 2000
Phospholipase C‐β4 (PLCβ4)	Nakamura et al. 2004
Phospholipase C‐γ1 (PLCγ1)	Tong et al. 2004; Yuan et al. 2005
Protease‐activated receptor 2 (PAR‐2)	Jin et al. 2013
Sensitization to IP3/Ca2 +
Bcl‐2 (B‐cell lymphoma 2)a	Chen et al. 2004; Eckenrode et al. 2010; Monaco et al. 2012; Chang et al. 2014
Bcl‐XL (B‐cell lymphoma extra large)	White et al. 2005; Eckenrode et al. 2010; Monaco et al. 2012
Chromogranin A (CGA)	Yoo & Lewis, 1998; Thrower et al. 2002
Chromogranin B (CGB; secretogranin‐1)	Yoo & Lewis, 2000; Thrower et al. 2003
Cyclin‐A	Soghoian et al. 2005
Cyclin‐B1 (CYB)	Malathi et al. 2003; Malathi et al. 2005
Cyclin‐dependent kinase 1 (CDK1)	Malathi et al. 2003; Malathi et al. 2005
Cytochrome c 1	Boehning et al. 2004
Fyn (tyrosine‐protein kinase)	Jayaraman et al. 1996; Cui et al. 2004
Glucosidase 2 subunit β (80K‐H)	Kawaai et al. 2009
Glycogen synthase kinase‐3β (GSK3β)	Gomez et al. 2016
Huntingtin‐associated protein 1 (HAP‐1)	Tang et al. 2003 b
Huntingtin (HTT) (with poly‐Q expansion, HTTexp)b	Tang et al. 2003 b
Lyn (tyrosine‐protein kinase)	Yokoyama et al. 2002
Mcl‐1 (myeloid cell leukemia‐1)	Eckenrode et al. 2010
mTOR (mammalian target of rapamycin)	Fregeau et al. 2011
Neuronal Ca2+ sensor 1 (NCS‐1)	Schlecker et al. 2006
Polo‐like kinase 1 (PLK1)	Ito et al. 2008; Vanderheyden et al. 2009 b
Presenilin‐1/Presenilin‐2 (PS‐1/PS‐2)	Cheung et al. 2008
Protein kinase A (PKA; cAMP‐dependent protein kinase)	Ferris et al. 1991; Bruce et al. 2002
Receptor of activated protein kinase C1 (RACK1)	Patterson et al. 2004
Rho‐associated protein kinase (ROCK)	Singleton & Bourguignon, 2002
TRISK 32 (cardiac triadin TRISK 32 isoform)	Olah et al. 2011
Direct activation of IP3Rs
Ca2+‐binding protein 1 (CaBP1)c	Yang et al. 2002; Li et al. 2013
CIB1 (Ca2+ and integrin‐binding protein 1; calmyrin)c	White et al. 2006
Gβγ complex	Shin et al. 2000; Zeng et al. 2003
Other
DARPP‐32 (protein phosphatase 1 regulatory subunit 1B)	Chang et al. 2014
DHHC6	Fredericks et al. 2014
EB3 (end‐binding protein 3)	Geyer et al. 2015
GRP‐78 (78 kDa glucose‐regulated protein; BiP)	Higo et al. 2010
Phosphatidylinositol trisphosphate 3‐phosphatase (PTEN)	Bononi et al. 2013
Selenoprotein K (SELK)	Fredericks et al. 2014

Data for Tables 1, 2, 3, 4 were derived from manual searches of the literature, reviews (Choe & Ehrlich, 2006; Foskett et al. 2007; Mikoshiba, 2007; Vanderheyden et al. 2009 a) and databases, including BioGRID (Chatr‐Aryamontri et al. 2015) and IntAct (Orchard et al. 2013). The nomenclature of proteins shown is consistent with the human homologues, although some data are derived from interactions of IP₃Rs and proteins from other species. ^aSome studies report sensitization of IP₃Rs by Bcl‐2, while others report inhibition. ^bHTT^exp, but not wild‐type HTT, sensitizes IP₃Rs to IP₃/Ca²⁺. ^cCaBP1 and CIB1 are also reported to inhibit IP₃Rs (see Table 2); direct activation seems to occur only transiently, and is controversial.

Table 2.

Proteins that form complexes with IP₃Rs and inhibit their activity

Protein	References
Proteins that bind reversibly and disrupt activation by IP3 and/or Ca2+
Ankyrin‐R (ANK1)	Bourguignon et al. 1993; Joseph & Samanta, 1993
Bcl‐2 (B‐cell lymphoma 2)a	Chen et al. 2004; Monaco et al. 2012; Chang et al. 2014
Ca2+‐binding protein 1 (CaBP1)b	Yang et al. 2002; Li et al. 2013
Calmodulin (CaM)	Maeda et al. 1991; Yamada et al. 1995
Carbonic anhydrase‐related protein (CARP; CA8)	Hirota et al. 2003
Caspase‐3	Hirota et al. 1999
CIB1 (Ca2+ and integrin‐binding protein 1; calmyrin)b	White et al. 2006
DANGER (IP3R‐interacting protein)	van Rossum et al. 2006
ERp44 (endoplasmic reticulum resident protein 44)	Higo et al. 2005
FKBP1A (FK506‐binding protein 1A; FKBP12)	Cameron et al. 1995 b
GIT1/GIT2 (ARF GTPase‐activating protein 1/2)	Zhang et al. 2009
IRBIT (IP3‐binding protein released with IP3)	Ando et al. 2003
K‐Ras	Sung et al. 2013
MRVI1 (IRAG; IP3R‐associated cGMP kinase substrate)	Schlossman et al. 2000
Nuclear protein localization protein 4 homologue (NPL4)	Alzayady et al. 2005
Polycystin‐1 (PC1; TRPP1)	Li et al. 2005
Proteins that post‐translationally modify IP3Rs
AKT1 (RAC‐α serine/threonine protein kinase; PKB)	Khan et al. 2006; Szado et al. 2008
Ca2+/calmodulin‐dependent protein kinase II (CaMKII)	Ferris et al. 1991; Bare et al. 2005
Calpain	Μagnusson et al. 1993; Wojcikiewicz & Oberdorf, 1996
E3 ubiquitin ligase AMFR (GP78)c	Pearce et al. 2007
E3 ubiquitin ligase RNF170c	Lu et al. 2011
Erlin‐1/Erlin‐2 (SPFH domain‐containing protein 1/2)c	Pearce et al. 2007; Pearce et al. 2009
MAPK1/MAPK3 (mitogen‐activated protein kinase 1/3)	Bai et al. 2006
Protein phosphatase 1A (PP1A)	Tang et al. 2003 a; Chang et al. 2014
Transglutaminase‐2 (TGM2)	Hamada et al. 2014
Transitional endoplasmic reticulum ATPase (p97)c	Alzayady et al. 2005
Ubiquitinc	Bokkala & Joseph, 1997; Oberdorf et al. 1999
Ubiquitin‐conjugating enzyme E2 7 (UBC7)c	Webster et al. 2003
Ubiquitin conjugation factor E4A (UFD2)c	Alzayady et al. 2005
Ubiquitin fusion degradation 1 protein (UFD1)c	Alzayady et al. 2005

^aBcl‐2 has also been reported to sensitize IP₃Rs to IP₃/Ca²⁺ (see Table 1). ^bCaBP1 and CIB1 may also cause transient activation of IP₃Rs, although this is controversial (see Table 1). ^cComponents of the proteasomal pathway.

Table 3.

Proteins that form complexes with IP₃Rs and act as downstream effectors

Protein	References
Anoctamin‐1 (ANO1, Ca2+‐activated Cl− channel)	Jin et al. 2013
Calcineurin (CN; protein phosphatase 2B)	Cameron et al. 1995 a; Chang et al. 2014
CASK (Ca2+/calmodulin‐dependent serine protein kinase)	Maximov et al. 2003
CRTC2 (CREB‐regulated transcription coactivator 2)	Wang et al. 2012
IRBIT (IP3‐binding protein released with IP3)a	Ando et al. 2003
KCa1.1 (BKCa; large conductance Ca2+‐activated K+ channel)	Zhao et al. 2010; Mound et al. 2013
Na+/Ca2+ exchanger 1 (NCX1)	Lencesova et al. 2004; Mohler et al. 2005
Orai‐1 (Ca2+ release‐activated Ca2+ channel 1)	Woodard et al. 2010; Lur et al. 2011
Plasma membrane Ca2+ ATPase (PMCA)	Shin et al. 2000; Huang et al. 2006
Protein kinase C (PKC)	Ferris et al. 1991; Rex et al. 2010
SERCA 2B/3 (sarco/endoplasmic reticulum Ca2+‐ATPase)	Redondo et al. 2008
STIM1 (stromal interaction molecule 1)	Santoso et al. 2011
TRPC1‐7 (transient receptor potential canonical channels)	Boulay et al. 1999; Mery et al. 2001; Tang et al. 2001; Yuan et al. 2003; Tong et al. 2004
VDAC1 (voltage‐dependent anion channel 1)	Szabadkai et al. 2006

^aIRBIT also inhibits IP₃Rs by occluding the IP₃‐binding site (Table 2).

Table 4.

Other proteins that form complexes with IP₃Rs

Protein	References
Cytoskeletal, scaffolding and adaptor proteins
14‐3‐3 protein zeta/delta (PKC inhibitor protein 1)	Angrand et al. 2006
α‐Actin	Sugiyama et al. 2000
Ankyrin‐B (ANK2)	Hayashi & Su, 2001; Mohler et al. 2004; Kline et al. 2008
AKAP9 (A‐kinase anchor protein 9; Yotiao)	Tu et al. 2004
BANK1 (B‐cell scaffold protein with ankyrin repeats)	Yokoyama et al. 2002
Caveolin‐1	Murata et al. 2007; Sundivakkam et al. 2009; Jin et al. 2013
Coiled‐coil domain‐containing protein 8	Hanson et al. 2014
Homer 1/2/3	Tu et al. 1998
EB1 / EB3 (end‐binding protein 1/3)a	Geyer et al. 2015
KRAP (K‐Ras‐induced actin‐interacting protein)	Fujimoto et al. 2011
LAT (linker of activated T‐cells)	deSouza et al. 2007
Myosin‐2A	Walker et al. 2002; Hours & Mery, 2010
Obscurin‐like protein 1	Hanson et al. 2014
Protein 4.1N (band 4.1‐like protein 1)	Maximov et al. 2003
SEC8 (exocyst complex component)	Shin et al. 2000
SNAP‐29 (synaptosomal‐associated protein 29)	Huttlin et al. 2013
α‐Spectrin/β‐spectrin (α/β‐fodrin)	Lencesova et al. 2004
Syntaxin 1B	Tanaka et al. 2011
Talin	Sugiyama et al. 2000
Vimentin	Dingli et al. 2012
Vinculin	Sugiyama et al. 2000
Other proteins
Anaplastic lymphoma kinase (ALK)	Crockett et al. 2004
ARHGAP1 (Rho GTPase‐activating protein 1)	Nagaraja & Kandpal, 2004
γ‐BBH (γ‐butyrobetaine dioxygenase)	Huttlin et al. 2013
Beclin‐1	Vicencio et al. 2009
BOK (Bcl‐2‐related ovarian killer protein)	Schulman et al. 2013
Calnexin	Joseph et al. 1999
CD44 antigen (heparin sulphate proteoglycan)	Singleton & Bourguignon, 2004
CEMIP (cell migration‐inducing and hyaluronan‐binding protein)	Tiwari et al. 2013
Cyclophilin D (peptidyl‐prolyl cis‐trans isomerase F)	Paillard et al. 2013
FAM19A4 (chemokine‐like protein TAFA‐4)	Huttlin et al. 2013
F‐box and leucine‐rich repeat protein 14	Huttlin et al. 2013
FGL2 (fibrinogen‐like 2)	Huttlin et al. 2013
FERM domain‐containing 1	Huttlin et al. 2013
GluRδ2 (ionotropic glutamate receptor δ2)	Nakamura et al. 2004
Golgi anti‐apoptotic protein (GAAP; Lifeguard 4; TMBIM4)	de Mattia et al. 2009
GRP‐75 (glucose‐regulated protein 75; stress‐70 protein)	Szabadkai et al. 2006
Heat shock protein 90 (HSP90)	Nguyen et al. 2009
Junctate	Treves et al. 2004
Lethal(3)malignant brain tumor‐like protein 2	Huttlin et al. 2013
Lymphoid‐restricted membrane protein (LRMP; JAW1)	Shindo et al. 2010
Na+/K+‐transporting ATPase	Mohler et al. 2005; Yuan et al. 2005
Neuronal acetylcholine receptor α3	Huttlin et al. 2013
PASK (PAS domain‐containing protein kinase)	Schlafli et al. 2011
Phospholamban	Koller et al. 2003
Polycystin‐2 (PC2; TRPP2)	Li et al. 2005
Protein kinase G1 (PKG1; cGMP‐dependent protein kinase 1)	Schlossman et al. 2000
PTPα (protein tyrosine phosphatase‐α)	Wang et al. 2009
Rab29 (Ras‐related protein Rab7L1)	Huttlin et al. 2013
Rac1 (Ras‐related C3 botulinum toxin substrate 1; TC25)	Natsvlishvili et al. 2015
RhoA	Mehta et al. 2003
Sigma 1 receptor (σ1R)	Hayashi & Su, 2001; Natsvlishvili et al. 2015
Sirtuin‐7	Tsai et al. 2012
c‐Src (proto‐oncogene tyrosine‐protein kinase Src)	Jayaraman et al. 1996; Wang et al. 2009
STARD13 (StAR‐related lipid transfer protein 13; RhoGAP)	Nagaraja & Kandpal, 2004
Syndecan‐1 (SYND1; CD138)	Maximov et al. 2003
TESPA1 (thymocyte‐expressed positive selection‐associated protein 1)	Matsuzaki et al. 2012

^aBoth EB1 and EB3 associate with IP₃Rs, but only EB3 has been shown to be required for effective Ca²⁺ signalling in endothelial cells (Table 1) (Geyer et al. 2015).

Although we focus on the ability of IP₃Rs to release Ca²⁺ from intracellular stores, IP₃Rs have additional roles. For example, binding of IP₃ is proposed to release IRBIT from the IP₃‐binding site, freeing IRBIT to regulate additional targets that include ion channels, transporters and the enzyme ribonucleotide reductase (Ando et al. 2014; Arnaoutov & Dasso, 2014). IP₃Rs may also regulate associated proteins independently of their ability to release Ca²⁺. For example, a direct interaction between IP₃Rs and TRPC (transient receptor potential canonical) channels is proposed to stimulate opening of the latter (Zhang et al. 2001). Hence, when reviewing the effects of proteins associated with IP₃Rs, we should look beyond the effects of IP₃ on cytosolic Ca²⁺ signals, to consider also consequences within the ER lumen, effects on Ca²⁺ entry, and effects unrelated to Ca²⁺ signalling. That scope is too ambitious for this short review. Instead we provide a comprehensive summary of proteins suggested to interact with IP₃Rs (Tables 1, 2, 3, 4, within which we provide most references) and then explore a few selected examples to illustrate some general features.

Signalling complexes containing IP₃Rs span entire signalling pathways

The sheer number of proteins reported to form complexes with IP₃Rs is striking and so too is their diversity, in terms of both cellular geography and function (Tables 1, 2, 3, 4). IP₃Rs form complexes with many of the proteins that link extracellular stimuli to formation of IP₃, including G protein‐coupled receptors (GPCRs), the epidermal growth factor receptor (EGFR), the erythropoietin receptor, the Gβγ complexes of G proteins, and some forms of PLC. IP₃Rs also associate with other signalling proteins linked to PLC signalling, including protein kinase C (PKC), RACK1 (receptor of activated PKC) and the phosphoinositide phosphatase PTEN. The interactions extend also to proteins from other signalling pathways, including adenylyl cyclase (AC), the small G protein K‐Ras, and the protein kinases AKT1 (RAC‐α serine/threonine protein kinase), mTOR (mammalian target of rapamycin), c‐Src and MAPK1/MAPK3 (mitogen‐activated protein kinase 1/3) (Tables 1, 2, 3, 4 and Fig. 1). Proteins that respond to the Ca²⁺ released by IP₃Rs also form complexes with IP₃Rs. These include ion channels, exchangers and pumps within the plasma membrane. It is clear that IP₃Rs reside within macromolecular complexes that both span entire signalling pathways from cell‐surface receptors to the effectors that respond to Ca²⁺, and include proteins that integrate signals from other signalling pathways.

The advantages of these signalling complexes are clear. They allow information to be directed selectively from specific extracellular stimuli to specific intracellular targets through conserved signalling pathways. Furthermore, associated proteins can integrate signals from different signalling pathways and so modulate traffic through the complex. Hence, protein complexes confer both specificity and plasticity. A third advantage is speed. Signalling pathways must be able to turn on and off quickly. Fast activation benefits from high concentrations of reactants and fast on‐rates (k ₁) for association of messengers with their targets. Rapid de‐activation requires rapid destruction or dissipation of the messenger and a fast dissociation rate (k ₋₁). By facilitating delivery of messengers at high local concentrations to their targets (e.g. IP₃ to IP₃Rs), signalling complexes contribute to both rapid activation and de‐activation, the latter because diffusion of messengers away from the site of delivery may be sufficient to allow their concentration to fall below that required for activation as soon as synthesis of the messenger ceases. Secondly, targets can have fast off‐rates (k ₋₁) with a corresponding loss of affinity (equilibrium association constant, K _A = k ₁/k ₋₁) that does not compromise their capacity to respond to high local concentrations of messenger. We suggest, then, that assembly of proteins around IP₃Rs contributes to fast and specific signalling, while providing opportunities for signal integration and plasticity.

For convenience, we consider the proteins that associate with IP₃Rs under four somewhat arbitrary (and overlapping) headings: proteins that enhance or inhibit the activity of IP₃Rs (Tables 1 and 2); proteins that respond to Ca²⁺ released by IP₃Rs (Table 3); and proteins with more general roles, including those associated with movement of IP₃Rs (Table 4).

Proteins that enhance the function of IP₃Rs

Usually, IP₃Rs open only when they have bound both IP₃ and Ca²⁺ (Foskett et al. 2007; Taylor & Tovey, 2010). Unsurprisingly, therefore, most of the proteins that associate with IP₃Rs and enhance their activity do so either by allowing more effective delivery of IP₃ and/or Ca²⁺ to IP₃Rs, or by enhancing the responsiveness of IP₃Rs to IP₃ and/or Ca²⁺ (Table 1).

The association of IP₃Rs with GPCRs, EGFR and erythropoietin receptors, with the βγ subunits of G proteins, with some isoforms of PLC, and with scaffold proteins, like Homer 1 that tethers IP₃Rs to metabotropic glutamate receptors and PLC (Tu et al. 1998), suggest mechanisms by which receptors may effectively deliver IP₃ to specific IP₃Rs. This targeted delivery of IP₃ provides two advantages: it allows rapid responses and it may allow spatially organized Ca²⁺ signals to retain an ‘imprint’ of the stimulus that evoked them. Bradykinin B₂ receptors (B₂Rs) are a well‐defined example. In sympathetic neurons, both muscarinic M₁ receptors (M₁Rs) and B₂Rs activate PLC, but only activation of B₂Rs evokes Ca²⁺ release through IP₃Rs (Delmas et al. 2002). This selectivity arises because B₂Rs, but not M₁Rs, form complexes with IP₃Rs. Rapid generation of IP₃ in response to activation of B₂Rs thereby generates relatively high concentrations of IP₃ in the vicinity of IP₃Rs, which are not achieved by the more distant M₁Rs. In this case, selective coupling between plasma membrane receptors and IP₃Rs may allow sympathetic neurons to generate different intracellular responses to pro‐inflammatory and cholinergic inputs.

Rather than enhancing the delivery of IP₃ to IP₃Rs, many other proteins sensitize IP₃Rs to prevailing concentrations of IP₃ and/or Ca²⁺ (Table 1). An example, which may play an important role in human disease, is the sensitization of IP₃Rs by mutant forms of presenilins (Cheung et al. 2008). Mutations in presenilin‐1 (PS1) and presenilin‐2 (PS2) are major causes of familial Alzheimer's disease. Although both wild‐type and mutant presenilins associate with IP₃Rs, only the disease‐causing mutant forms of PS1 and PS2 enhance the activity of IP₃Rs in response to IP₃ and Ca²⁺. The mechanism involved may be a change in the modal gating of IP₃Rs (Cheung et al. 2010). This increased activity of IP₃Rs results in enhanced release of Ca²⁺, which may lead to aberrant processing of β‐amyloid (Cheung et al. 2008), constitutive activation of cyclic AMP response element binding protein (CREB)‐mediated transcription (Muller et al. 2011), synaptic dysfunction and neuronal degeneration (Mattson, 2010).

Although activation of IP₃Rs normally requires binding of IP₃ and Ca²⁺, a few proteins have been reported to cause reversible activation of IP₃Rs directly, without the coincident presence of IP₃ and Ca²⁺ (Table 1). These include Gβγ (Zeng et al. 2003), CIB1 (White et al. 2006) and, more controversially, CaBP1 (Yang et al. 2002). The initial report on the actions of CaBP1 described an activation of Xenopus IP₃Rs in the absence of IP₃ in vitro. However, subsequent studies have demonstrated that CaBP1 inhibits Ca²⁺ release via mammalian and Xenopus IP₃Rs by stabilizing an inactive state of the IP₃R (Haynes et al. 2004; Nadif Kasri et al. 2004; White et al. 2006; Li et al. 2013). Similarly, CIB1 was reported to activate IP₃Rs in Xenopus oocytes and Sf9 insect cells in the absence of IP₃, but it too inhibits Ca²⁺ release via mammalian IP₃Rs (White et al. 2006). Uniquely, an irreversible activation of IP₃Rs appears to occur after proteolytic cleavage by caspase‐3 (Assefa et al. 2004; Nakayama et al. 2004), a process that may play a prominent role in apoptosis.

Proteins that inhibit the function of IP₃Rs

Many proteins that interact with IP₃Rs inhibit their function (Table 2). These interactions may enable rapid feedback regulation of Ca²⁺ release and provide long‐term attenuation of IP₃R activity by promoting degradation or irreversible inhibition of IP₃Rs. These mechanisms contribute to the tight regulation of IP₃R activity needed to achieve spatial and temporal organization of Ca²⁺ signals (Konieczny et al. 2012). They also provide protection from the damaging consequences of excessive increases in cytosolic free Ca²⁺ concentration (Orrenius et al. 2015) and disturbance of the other essential roles of the ER while it fulfils its role in Ca²⁺ signalling (Berridge, 2002). Proteins that inhibit IP₃Rs in a Ca²⁺‐dependent manner, like calmodulin, CaBP1, calcineurin, CaMKII and the unidentified protein(s) that may mediate the universal inhibition of IP₃Rs by Ca²⁺, are prime candidates for mediating this negative feedback. Proteins that inhibit IP₃Rs fall into two broad categories: those that bind reversibly to interfere with binding of IP₃ and/or Ca²⁺ or their links to gating; and those that cause post‐translational modifications of the IP₃R (Table 2).

IRBIT inhibits all three IP₃R subtypes by competing with IP₃ for binding to the IBC (Ando et al. 2003). IRBIT binds only when it is phosphorylated at several sites, probably because the phosphorylated residues mimic the essential phosphate groups of IP₃ (Fig. 2 A). Residue S68 is the ‘master’ phosphorylation site. When it is phosphorylated by a Ca²⁺‐dependent kinase, perhaps a Ca²⁺/calmodulin‐dependent protein kinase (CaMK), it allows casein kinase I‐mediated phosphorylation of the two residues (S71 and S74, residue numbering relates to mouse IP₃R1) that are critical for binding of IRBIT to IP₃Rs (and its other targets) (Ando et al. 2014). Dephosphorylation of S68 is catalysed by protein phosphatase 1 (PP1), which also associates with IRBIT. The competition between phospho‐IRBIT and IP₃ for occupancy of the IBC through which IP₃ initiates activation of IP₃Rs allows IRBIT to tune the sensitivity of IP₃Rs to IP₃. Hence, inhibiting expression of IRBIT, or expression of a dominant negative form (IRBIT‐S68A), allows Ca²⁺ release at lower concentrations of IP₃ (Ando et al. 2014). This tuning of IP₃R sensitivity has been demonstrated in sympathetic neurons where, as discussed earlier, M₁Rs do not associate with IP₃Rs and do not normally generate sufficient IP₃ to activate more distant IP₃Rs (Delmas et al. 2002). However, expression of the dominant negative IRBIT allows M₁Rs to evoke Ca²⁺ release through IP₃Rs (Zaika et al. 2011). Although the details are not fully resolved, the interplay between Ca²⁺ and the activation of IRBIT is intriguing because it suggests potential feedback loops that might control the sensitivity of IP₃Rs to IP₃ (Ando et al. 2014). The phosphorylation (of S68) that initiates activation of IRBIT is Ca²⁺ sensitive, deactivation of IRBIT by proteolytic cleavage within its N‐terminal may be mediated by Ca²⁺‐sensitive calpain, and IRBIT itself inhibits Ca²⁺/calmodulin‐dependent protein kinase IIα (CaMKIIα) (Kawaai et al. 2015) (Fig. 2 B).

Figure 2. IRBIT controls the sensitivity of IP₃Rs .

A, the N‐terminal region of IRBIT includes a serine‐rich domain. Phosphorylation of S68, the ‘master’ phosphorylation site, allows sequential phosphorylation of the two residues, S71 and S74, that must be phosphorylated for IRBIT to bind to IP₃Rs. Protein phosphatase 1 (PP1) bound to IRBIT dephosphorylates S68. B, phosphorylation of IRBIT (1) allows it to bind to the IBC and so compete with IP₃ for binding to the IP₃R. Phospho‐IRBIT thereby sets the sensitivity of the IP₃R to IP₃. IP₃ binding to the IBC (2) prevents IRBIT binding and initiates activation of the IP₃R. The displaced phospho‐IRBIT can regulate many additional targets, including ion channels and transporters (3). The Ca²⁺ released by active IP₃Rs may control the phosphorylation state of IRBIT, and thereby complete a feedback loop that regulates IP₃R sensitivity (4).

Post‐translational modification of IP₃Rs by associated proteins may be reversible (e.g. phosphorylation) (Betzenhauser & Yule, 2010) or irreversible (e.g. proteolysis and some covalent modifications). An example of the latter is the Ca²⁺‐dependent enzyme transglutaminase type 2 (TGM2). By covalently modifying a glutamine residue within the C‐terminal tail of IP₃R1, TGM2 causes irreversible cross‐linking of adjacent IP₃R subunits via a lysine residue and the modified glutamine. This prevents the conformational changes required for activation of IP₃Rs, and so inhibits IP₃‐evoked Ca²⁺ release (Hamada et al. 2014). The Ca²⁺ sensitivity of TGM2 may allow it to contribute to feedback control of Ca²⁺ release and to disruption of IP₃R function when dysregulation of Ca²⁺ signalling occurs in pathological conditions such as Huntington's disease (Hamada et al. 2014). Activation of IP₃Rs and the ensuing release of Ca²⁺ also trigger ubiquitination and proteasomal degradation of IP₃Rs (Pearce et al. 2009) and their cleavage by calpains (Μagnusson et al. 1993; Wojcikiewicz & Oberdorf, 1996). Hence, proteins that associate with IP₃Rs provide mechanisms that allow both acute and long‐term feedback regulation of IP₃R activity.

Downstream effectors

IP₃Rs also form complexes with proteins that are downstream effectors of IP₃R activation; most of these respond to the Ca²⁺ released by IP₃Rs (Table 3). Many of these proteins are cytosolic, but others reside within membranes that allow IP₃Rs within the ER to communicate with other intracellular organelles or the plasma membrane. The importance of this communication between organelles, mediated by junctional complexes between them, is increasingly recognized (Lam & Galione, 2013).

Hepatic gluconeogenesis, which is likely to play an important role in diabetes and obesity, is stimulated by glucagon released by the pancreas during fasting, and inhibited by insulin released when the plasma glucose concentration increases. A complex containing IP₃Rs, the Ca²⁺‐regulated protein phosphatase calcineurin, the transcriptional co‐activator of CREB‐regulated transcription CRTC2 (CREB‐coactivator C2), PKA and AKT1 coordinates gluconeogenesis (Wang et al. 2012) (Fig. 3). De‐phosphorylated CRTC2 binds to nuclear CREB and up‐regulates genes that promote gluconeogenesis. This is repressed by SIK2, a kinase that phosphorylates CRTC2. IP₃‐evoked Ca²⁺ release activates calcineurin, which de‐phosphorylates CRTC2. Glucagon receptors stimulate production of both cAMP and IP₃ (Wakelam et al. 1986; Wang et al. 2012). The cAMP activates PKA, which phosphorylates, and thereby inhibits, SIK2; and it phosphorylates IP₃Rs, sensitizing them to activation by IP₃ and Ca²⁺. IP₃Rs are also directly sensitized by cAMP (Tovey et al. 2008). Increased release of Ca²⁺ via IP₃Rs activates calcineurin, which dephosphorylates CRTC2 (Vanderheyden et al. 2009 a; Wang et al. 2012). Hence glucagon both inhibits the kinase (SIK2) and stimulates the phosphatase (calcineurin) that control phosphorylation of CRTC2. Glucagon also reduces binding of CRTC2 to IP₃Rs (Wang et al. 2012), further enhancing the nuclear translocation of dephosphorylated CRTC2. The signals evoked by insulin receptors also feed into this IP₃R complex. Insulin stimulates phosphatidylinositol 3‐kinase (PI3K) and thereby AKT1. The latter phosphorylates IP₃Rs and attenuates their activity. Hence insulin, by inhibiting IP₃Rs, opposes the actions of glucagon by restraining the activation of calcineurin and so maintains CRTC2 in its inactive phosphorylated state (Wang et al. 2012). This example illustrates some of the intricate interactions that the assembly of proteins around IP₃Rs can allow: signals from a GPCR and a receptor tyrosine kinase converge at IP₃Rs, which then integrate the inputs and transduce them into a regulation of gene expression (Fig. 3).

Figure 3. A signalling complex assembled around IP₃Rs controls gluconeogenesis .

Glucagon and insulin exert opposing effects on hepatic gluconeogenesis. Their signalling pathways converge to a protein complex assembled around IP₃Rs, the activity of which controls phosphorylation of the transcription factor CRTC2. Dephosphorylated CRTC2 translocates to the nucleus, where it associates with CREB and stimulates transcription of genes required for gluconeogenesis. SIK2 phosphorylates CRTC2, while calcineurin dephosphorylates it. Glucagon, via a GPCR, stimulates both PLC and AC. The IP₃ produced by PLC stimulates IP₃Rs. The cAMP generated by AC stimulates PKA and that promotes dephosphorylation of CRTC2 by phosphorylating both SIK2 (inhibiting its activity) and IP₃Rs, sensitizing the latter to IP₃. The larger Ca²⁺ signal then activates calcineurin. Insulin causes activation of AKT1, which phosphorylates IP₃Rs and inhibits their activity; it thereby opposes the effects of glucagon and attenuates calcineurin activity. Phosphorylation is indicated by red circles, black arrows denote stimulation and the red arrow denotes inhibition. Abbreviations and further details in the text and tables.

Proteins that determine the distribution of IP₃Rs

The subcellular distribution of IP₃Rs is an important influence on their behaviour, not least because it defines the sites at which they will release Ca²⁺, and whether they will be exposed to effective concentrations of the stimuli that activate them, IP₃ and Ca²⁺. Assembly of IP₃Rs with components of the PLC signalling pathway (see above) can ensure targeted delivery of IP₃, but Ca²⁺ is most often provided by neighbouring IP₃Rs. An important interaction, therefore, is that between IP₃Rs themselves, because their proximity to neighbours dictates whether Ca²⁺ released by an active IP₃R can ignite the activity of other IP₃Rs. Considerable evidence suggests that clustering of IP₃Rs within the plane of the ER membrane is dynamically regulated by IP₃ and/or Ca²⁺ (Tateishi et al. 2005; Rahman et al. 2009; and see references in Geyer et al. 2015), although the role of this process in shaping Ca²⁺ signals remains controversial (Smith et al. 2014). We have suggested that IP₃‐evoked clustering of IP₃Rs may contribute to the coordinated openings of IP₃Rs that underlie the small Ca²⁺ signals (‘Ca²⁺ puffs’) evoked by low stimulus intensities, by both bringing IP₃Rs together and retuning their Ca²⁺ sensitivity (Rahman et al. 2009). Head‐to‐head interactions of IP₃Rs have also been observed in electron micrographs of purified IP₃Rs (Hamada et al. 2003), between opposing ER membranes within cells (Takei et al. 1994) and between the isolated N‐terminal domains of IP₃Rs (Chavda et al. 2013). The functional significance of these interactions has not been established.

A recent study of the Ca²⁺ signals evoked by thrombin‐mediated stimulation of the protease‐activated receptor PAR‐1 in endothelial cells provides evidence that microtubules may guide IP₃Rs into the clusters within which Ca²⁺ release can most effectively recruit neighbouring IP₃Rs (Geyer et al. 2015). In lung microvascular endothelial cells, thrombin, which activates PAR‐1 by cleaving its N‐terminal, stimulates PLC and thereby evokes Ca²⁺ release through IP₃Rs. The resulting increase in cytosolic Ca²⁺ concentration contributes to disassembly of the adherens junctions that maintain the integrity of the endothelium (Komarova & Malik, 2010). These effects are attenuated when the interaction between type 3 IP₃Rs (IP₃R3) and end‐binding protein 3 (EB3) are disrupted. EB3 belongs to a family of proteins that bind to the plus‐end of growing microtubules and recruit other proteins, often via an S/TxIP motif (where x denotes any residue) (Honnappa et al. 2009). Mutation of the TxIP motif within the regulatory domain of IP₃R3 prevents its binding to EB3, attenuates thrombin‐evoked Ca²⁺ signals, and reduces both the basal clustering of IP₃R3 and the enhanced clustering evoked by thrombin. Hence, in endothelial cells, the association of IP₃R3 with EB3 and microtubules is required for both clustering of IP₃Rs and effective Ca²⁺ signalling. This suggests that clustering allows IP₃Rs to deliver Ca²⁺ more effectively to other IP₃Rs and so allows the amplification provided by Ca²⁺‐induced Ca²⁺ release (Fig. 4). We conclude that association of IP₃Rs with other proteins, components of the PLC signalling pathway or EB3, contributes to effective delivery of the two essential regulators of IP₃Rs, IP₃ and Ca²⁺, respectively.

Figure 4. EB3 is required for effective signalling by IP₃Rs in endothelial cells .

In endothelial cells, EB3 binds to a TxIP motif within the regulatory domain of IP₃R3, allowing IP₃Rs to associate with the plus‐end of microtubules. Disrupting this interaction prevents clustering of IP₃Rs and attenuates the Ca²⁺ signals evoked by thrombin, which cleaves within the N‐terminus of PAR‐1 and allows it to stimulate PLC. The evidence (Geyer et al. 2015) suggests that the EB3‐mediated interaction of IP₃R3 with microtubules is essential for the clustering of IP₃Rs that allows the Ca²⁺ released by one IP₃R to be amplified by recruitment of neighbouring IP₃Rs.

Conclusions

IP₃Rs and the Ca²⁺ they release are called upon to specifically regulate many physiological processes (Berridge, 2009), while neither perturbing the other essential roles of the ER (Berridge, 2002) nor subjecting the cell to the deleterious consequences of excessive increases in cytosolic Ca²⁺ concentration (Orrenius et al. 2015). These demands impose a need for complex regulation of IP₃Rs, much of which is achieved by assembling proteins around IP₃Rs to form signalling complexes (Konieczny et al. 2012). These complexes allow signals to be directed through conserved signalling pathways and endow the pathways with speed, integrative capacity and plasticity. The very large size of IP₃Rs relative to most other ion channels might be viewed as an evolutionary adaptation to meet this need for them to function as signalling hubs.

Advances in genomics, proteomics, antibody technologies and bioinformatics have transformed analyses of protein–protein interactions. It is now possible to interrogate these interactions on a whole‐proteome scale (Havugimana et al. 2012; Rolland et al. 2014). Bioinformatic methods can predict protein–protein interactions (Baughman et al. 2011; Kotlyar et al. 2015) and even the regions of the proteins that are involved (Gavenonis et al. 2014). These powerful technologies, and the opportunities they provide to design new therapies (Wells & McClendon, 2007), cannot displace the need for direct confirmation of the interactions and their functional significance. Together, these approaches pave the way to defining the properties and functional importance of IP₃R signalling hubs in normal physiology and disease.

Additional information

Competing interests

None declared.

Author contributions

All authors have approved the final version of the manuscript and agree to be accountable for all aspects of the work. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

This work was supported by the Biotechnology and Biological Sciences Research Council (L0000075) and the Wellcome Trust (101844).

Biographies

David Prole studied Natural Sciences at the University of Cambridge before exploring the structure and function of K⁺ channels during his PhD with Neil Marrion at the University of Bristol and HCN pacemaker channels during postdoctoral training with Gary Yellen at Harvard Medical School. After moving back to the University of Cambridge to work with Colin Taylor he held a Meres Research Associateship from St John's College and now explores the roles of ion channels in cell signalling.

Colin Taylor began his career as an insect physiologist with Mike Berridge, before moving into phosphoinositide and Ca²⁺ signalling during a postdoc with Jim Putney in Virginia. He is presently Professor of Cellular Pharmacology and a Wellcome Trust Senior Investigator in the Department of Pharmacology in Cambridge, where he continues to explore the workings of IP₃ receptors.