|Home | About | Journals | Submit | Contact Us | Français|
Calcium signals mediate diverse cellular functions in immunological cells. Early studies with mast cells, then a preeminent model for studying Ca2+-dependent exocytosis, revealed several basic features of calcium signaling in non–electrically excitable cells. Subsequent studies in these and other cells further defined the basic processes such as inositol 1,4,5-trisphosphate-mediated release of Ca2+ from Ca2+ stores in the endoplasmic reticulum (ER); coupling of ER store depletion to influx of external Ca2+ through a calcium-release activated calcium (CRAC) channel now attributed to the interaction of the ER Ca2+ sensor, stromal interacting molecule-1 (STIM1), with a unique Ca2+-channel protein, Orai1/CRACM1, and subsequent uptake of excess Ca2+ into ER and mitochondria through ATP-dependent Ca2+ pumps. In addition, transient receptor potential channels and ion exchangers also contribute to the generation of calcium signals that may be global or have dynamic (e.g., waves and oscillations) and spatial resolution for specific functional readouts. This review discusses past and recent developments in this field of research, the pharmacologic agents that have assisted in these endeavors, and the mast cell as an exemplar for sorting out how calcium signals may regulate multiple outputs in a single cell.
Ringer’s formulation of his eponymous solution1 made possible the discoveries that extracellular Ca2+ was required for many cellular activities. These included anaphylactic release of histamine from chopped guinea pig lung,2 human leuko-cytes,3 rabbit basophils,4 and rat peritoneal mast cells.5 An often forgotten observation from these early studies was that Sr2+ or Ba2+ could replace Ca2+ in supporting anaphylactic histamine release from rat peritoneal mast cells.5,6 Subsequent studies with 45Ca2+ and 89Sr2+ showed that uptake of these ions by stimulated peritoneal mast cells correlated with the extent of histamine release.6,7 The uptake of 45Ca2+ was substantial and rose from ~ 5 to ~ 230 amoles/cell/min after stimulation (recalculated from published data7). The same was true for uptake of 89Sr2+.6 The concept drawn from these studies was that mast cell degranulation was dependent on influx of Ca2+ through “Ca2+ channels” that can convey Sr2+.6 The release of hista-mine from rat peritoneal mast cells, whether supported by Ca2+ or Sr2+, was effectively blocked by low concentrations of La3+.8,9
Our understanding of Ca2+ homeostasis in mast cells was further advanced by the introduction of Ca2+-specific fluorescent probes by Tsien et al.10 The initial studies with Fura-2 in the RBL-2H3 mast cell line indicated that antigen elicited an increase in concentration of cytosolic Ca2+ ([Ca2+]i) from ~ 0.1 to ~ 1 µM, which was maintained by a La3+-inhibitable influx of Ca2+, to generate an essential signal for degranulation.11 This increase correlated with the production of inositol phosphates12 of which the generation of inositol 1,4,5-trisphosphate (IP3)13 by phospholipase (PL) Cγ14 preceded the increase in [Ca2+]i. The calculated influx of Ca2+ was of similar magnitude to that calculated in the earlier studies with 45Ca2+.11 These data pointed to an influx mechanism(s) of relatively high capacity.
Later studies with calcium fluorescent probes and 45Ca2+-equilibrated RBL-2H3 cells demonstrated that Ca2+ was released from intracellular stores15–18 and that the subsequent influx of Ca2+ was associated with replenishment of an antigen/IP3-sensitive pool,19 accumulation of Ca2+ in mitochondria,20,19 and depolarization of the plasma membrane.21 Repolarization of the plasma membrane was necessary for continued influx of Ca2+ and degranulation.15,22 The fluxes of Ca2+ during this process were substantial. Intracellular 45Ca2+ increased from approximately 0.5 mM (about 500 amoles/cell) to about 2.5 mM and then decreased as the levels of cytosolic Ca2+ subsided.15 Other studies reaffirmed that the influx pathway was permeable to Sr2+ and other divalent metal ions.23 The mitochondrial pool was of high capacity and was insensitive to IP3 and thapsigargin, but was inhibited by oligomycin and antimycin A.19,20 The antigen-responsive pool, in contrast, could be depleted by IP3 and thapsigargin.19 Both pools, however, could be depleted by the Ca2+ ionophore, ionomycin. At low Ca2+ concentrations (0.1 µM), uptake of Ca2+ was largely confined to the IP3-sensitive store, but at higher concentrations (>1 µM), uptake into the mitochondrial pool predominated.19 Together, these studies suggested that antigen stimulation resulted in release of Ca2+ from an IP3- and thapsigargin-sensitive store, which was replenished by influx of extracellular Ca2+. This influx was dependent on maintenance of membrane polarity, presumably through efflux of K+. During this process, [Ca2+]i was buffered at concentrations of around 1 µM by uptake and release of Ca2+ in mitochondria. As depicted in Figure 1, these processes were counterbalanced by extrusion of Ca2+ from the cell by an ATP-dependent mechanism.15
Studies in other types of cells had identified the source of IP3-releasable Ca2+,24 as the endoplasmic reticulum (ER), the receptor for IP3 (IP3R)25 as the Ca2+-efflux channel in ER,26 and the mechanism of reuptake of Ca2+ into ER as an ATP-dependent process.24,27–30 Thapsigargin was shown to block reuptake into ER by inhibiting specifically the sarcoendoplasmic Ca2+ATPase pump (SERCA)31,32 and by doing so unmasks a spontaneous loss or “leak” of Ca2+ from the IP3-sensitive pool without stimulating phosphoinositide hydrolysis.33,34 Depletion of the ER pool by thapsigargin or IP3 inevitably results in entry of Ca2+ into the cell. This feature led to the concept, best articulated by Putney in 1986, of “capacitative calcium entry” or “store-operated calcium entry” (SOCE),35–37 a process that appeared to be relevant to mast cells.19,38 The description of a unique Ca2+-selective calcium-release activated calcium current (ICRAC) in RBL mast cells by Hoth and Penner in 1992 and 199339,40 initiated the search for an elusive ICRAC (CRAC) channel and set the tone for subsequent research on SOCE. However, the mechanism of communication between depleted ER stores and the putative CRAC channel in the plasma membrane remained unclear until recently. The properties and molecular identity of the CRAC channel and other components that are involved in the mobilization of Ca2+ are discussed in subsequent sections. Although our focus is the mast cell, our review is set in the context of IP3-mediated calcium signaling in other electrically nonexcitable cells because of the generic similarities in signaling mechanisms.
These include Ca2+ stores within the ER and, at least in some cells, nucleoplasmic reticulum (NR) to form a contiguous Ca2+ store of ~ 0.5 mM.41 The Golgi also contains high concentrations of Ca2+ (~ 0.3 mM), which is released into the adjacent cytosol.42 Both Golgi and ER contribute to the increase in [Ca2+]i on cell stimulation although at different rates and duration.43 In common with ER and NR, Golgi membranes contain the thapsigargin-sensitive SERCA, which enables reuptake of cytosolic Ca2+ against a high concentration gradient (see Section VIII.A). The Golgi membranes in addition contain the thapsigargin-insensitive secretory pathway calcium ATPase (SPCA). An inactivating mutation of the SPCA1 gene in the Hailey-Hailey skin disease indicates a prominent role for Golgi SPCA1 in regulating Ca2+ homeostasis at least in certain tissues.44 The relative contributions of Golgi and NR to the calcium signal in mast cells, or indeed any immu-nologic cell, has not been determined.
As noted earlier, the global calcium signal in mast cells is substantially attenuated and prolonged by Ca2+ uptake and release from mitochondria.20 It is thought also that uptake into mitochondria influences the spatial as well as temporal profile of the calcium signal to permit more discrete regulation of cellular events.45,46 The consequences are not only stimulation of mitochondrial energy production but also localization of the calcium signal to specific subcellular domains. Uptake is believed to occur through the mitochondrial Ca2+ uniporter (MCU) in the inner mitochondrial membrane.47 This uptake is driven by a negative membrane potential (Δψ) maintained by the respiratory chain and is accordingly impaired by the mitochondrial respiratory chain inhibitor, antimycin A, in conjunction with the ATP synthase inhibitor, oligomycin.20 MCU is also inhibited by ruthenium red analogs, Ru360 being a particularly potent inhibitor.48 MCU is activated via calmodulin when [Ca2+]i is elevated.49 Activating levels of [Ca2+]i are most likely to occur in the proximity of Ca2+ channels in ER, NR, Golgi, and plasma membranes and there is evidence that mitochondria interact with these channels to facilitate mitochondrial Ca2+ uptake and thereby restrict increases in Ca2+ to specific cellular domains.44,46 Mitochondria in the vicinity of Golgi, for example, may limit activation of Ca2+-dependent signaling molecules to those on or located around Golgi.44 Mitochondrial activity may also promote Ca2+ influx through CRAC channels by scavenging Ca2+ released from ER stores and ensuring efficient depletion of these stores.50 Impairment of mitochondrial respiration by depolarization suppresses ICRAC activity and Ca2+ entry whether stores are depleted by IP3, thapsigargin, or Adenophostin A.50 The actions of thapsigargin and Adenophostin A are described in more detail in Section IX.
It was noted 45 years ago that accumulation of Ca2+ in mitochondria requires the simultaneous but independent uptake of phosphate51 resulting in the formation of a freely dissociable calcium phosphate (Ca3(PO4)2) complex within the mitochondrial matrix.52 As long as phosphate ions remain in excess, mitochondrial Ca2+ can reach millimolar concentrations. Nevertheless, free Ca2+ concentrations in mitochondria [Ca2+]m are held within the narrow range of 0.5–2 µM regardless of the total Ca2+ load within mitochondria. The uncertainties about the nature and kinetic properties of the calcium phosphate complex are discussed in a concise review on calcium signaling and mitochondria.52 Egress of Ca2+ is mediated by a Na+/Ca2+ exchanger,53,54 now identified as NCX, which is thought to allow continuous recycling of Ca2+ across the mitochondrial membrane resulting in reciprocal changes in Ca2+ concentrations in mitochondria and cytosol.52 Ca2+ influx and efflux across the membrane of isolated mitochondria reaches rapid equilibrium over a wide range of loading conditions to buffer extramitochondrial [Ca2+] at 0.5–1.0 µM, the so-called mitochondrial “set point” for [Ca2+]i.55 The importance of NCX to mitochondrial Ca2+ homeostasis is evident in disorders such as hamster hereditary cardiomyopathy and human mitochondrial type 1 deficiency when enhanced NCX activity results in reduced mitochondrial Ca2+ levels and pathological defects.56,57 The Ca2+ antagonists, verapamil and diltiazem, inhibit NCX activity and have been tested with positive effects in these disorders.
IP3Rs are of three subtypes that are closely related in terms of function but show differential expression among tissues and within cells (see previous reviews45,58–61). IP3R subtypes are usually restricted to particular subcellular locations. RBL-2H3 mast cells express all three subtypes; IP3R1 and IP3R2 are located in ER and IP3R3 elsewhere.62 In liver HepG2 cells, ER contains IP3R2 and IP3R3, whereas NR contains only IP3R2.63 However, redistribution can occur during cell stimulation or polarization. Stimulation of RBL-2H3 cells with antigen, for example, results in rapid redistribution of IP3R2 in ER and NR from a diffuse to a clustered pattern.64 In endothelial and epithelial cell lines, IP3Rs cluster around tight junctions once cells become confluent.65,66 These patterns of redistributions are dependent on extracellular Ca2+. Other examples of the heterogeneous distribution and redistribution of IP3Rs are described in detail elsewhere,58 but it is apparent that there is no common pattern of distribution, or redistribution, of IP3Rs among different types of cells, and changes may be specific for particular functional responses.
The IP3Rs are expressed predominantly in the ER and to a lesser extent in NR and Golgi.45 In addition, a few (~ 2) IP3Rs localize in the plasma membrane of activated B cells, and because of their greater conductance (> 200 pS) than ICRAC (< 1.0 pS), they could conceivably contribute to the influx of Ca2+ during cell stimulation.67 Activation of these plasma membrane IP3Rs by IP3 requires several minutes, a delay that has been attributed to IP3R translocation to the plasma membrane.67
The IP3Rs normally exist as homo- or hetero-tetrameric structures to form functional gated Ca2+ channels that bear substantial sequence homology with the corresponding channel region of the sarcoplasmic ryanodine receptor and some structural resemblance to voltage-gated Ca2+ and K+ channels.61,68 Unlike ICRAC, IP3Rs conduct other divalent ions and Na+ but are insensitive to Gd3+ or La3+. Each IP3R subunit contains a cytosolic N-terminal ligand-binding domain, an intermediate modulatory domain, and six transmembrane-spanning helices at the C-terminus that constitute part of the tetrameric channel pore. The ligand-binding domain contains a binding pocket for IP3 as revealed by mutagenesis and crystal structure.69 Several lines of evidence indicate that the three-dimensional conformation of IP3R places the IP3-binding pocket in close proximity to the channel pore and that only subtle changes in conformation are required for channel gating (see Refs. 61 and 68 for further details). The intermediate modulatory domain is the least homologous region among the IP3Rs but contains binding sites for Ca2+, calmodulin, and other modulatory factors as outlined below.
Apart from IP3, IP3R interacts with a host of regulatory factors including Ca2+, calmodulin, ATP, other nucleotides, kinases, phosphatases, and non-enzymatic proteins that enable fine-tuning of IP3R function.58,59,61 Indeed, the recognition sites for these molecules constitute the major portion of the IP3R molecule. Ca2+ is the most important regulatory factor exhibiting stimulatory activity at low concentrations and inhibitory activity at high concentrations. Although initially a matter of controversy, it now appears that all three IP3Rs exhibit biphasic responses to Ca2+ but with different affinities in the order IP3R3 > IP3R2 > IP3R1.70 Multiple Ca2+-binding sites have been identified on IP3R, some of which may act cooperatively. However, the interactions that underlie the stimulatory and inhibitory effects of Ca2+ are still obscure although one enticing model has been proposed to account for both effects.71 The model invokes an N-terminal suppressor domain that can interact with either the IP3-binding domain or the Ca2+-gatekeeper region in a calmodulin-dependent manner according to the concentrations of Ca2+ and IP3. This model is based on reports that Ca2+ is an essential coactivator of channel opening,72,73 that a Ca2+-binding region within the IP3-binding domain acts in a coordinate manner,74 and that inhibition may involve a Ca2+-binding partner such as calmodulin.75–77 Other Ca2+-sensor proteins have been implicated as negative or positive regulators of IP3R activity but with conflicting results. In general, the precise roles of the Ca2+ sensors, including calmodulin, are unclear.
Many other proteins and small molecules are reported to interact with and regulate function of IP3Rs and it is thought that this multiplicity of interactions regulates the location of IP3R and adds precision to the way IP3Rs respond to IP3-dependent stimulants. These molecules include scaffolding proteins such as ankyrin, myosin, and Homer proteins that allow anchoring of IP3R at appropriate sites within the cell, TRPC family members (see Section V), receptors for activated C-kinase-1 (RACK1), and IP3R-binding proteins released by IP3 (IRBIT), among others.61,59 IP3Rs thus have the inherent ability to form macroprotein complexes or “signalosomes” that permit fine-tuning of Ca2+ release in a cell- and stimulus-dependent manner. However, many details of the molecular interactions require clarification and the reader is referred to recent reviews for more comprehensive accounts of IP3R.58–61,68,71
IP3Rs are also phosphorylated by multiple serine/threonine protein kinases (PKs) including PKA, PKG, PKC, and calcium/calmodulin-dependent PK. In general, these phosphorylations create positive feedback loops to regulate Ca2+ release possibly in spatially restricted areas.61 IP3R is also phosphorylated by the nonreceptor tyrosine kinases Fyn78 and Lyn,79 in T and B cells, respectively, thus facilitating release of intracellular Ca2+. Both kinases are activated on ligation of FcεRI in mast cells80 and, although not examined specifically, such a mechanism could conceivably contribute to the diminution of the calcium signal in Lyn-deficient mast cells.80
The highly Ca2+-selective current, ICRAC, that was initially characterized by Hoth and Penner in RBL mast cells, was found to be activated by depletion of intracellular Ca2+ stores by IP3, ionomycin, and chelation of cytosolic Ca2+.39,40 This current was further characterized in T lymphocytes by Zweifach and Lewis.81–83 Although ICRAC exhibited some similarities to Ca2+-selective voltage-activated currents, it differed in several respects. It was of low conductance,81 inhibited by elevated [Ca2+]i,40,82 and potentiated by extracellular Ca2+.83 Of particular note is that replacement of external Ca2+ with Sr2+ or Ba2+ resulted in a decline in ICRAC activity.83,84 The apparent abundance of ICRAC channels in T cells (possibly > 10,000/cell)81 may compensate for the slow conductance of this channel.
Nonetheless, the early patch clamp studies revealed activation of additional cation channels in mast cells and T cells. ICRAC85 and a 50 pS Ca2+/Mn2+ conducting channel activity 86 were detected in activated rat peritoneal mast cells with both channels contributing to the increase in [Ca2+]i. An additional current with characteristics of an amplified ICRAC was noted in Jurkat T cells87 and the RBL-2H3 mast cell line88 that was preactivated by the absence of external divalent metal ions and by store depletion.88 This amplified current appeared to be conducted through relatively few channels (several hundred per cell) and, like ICRAC, it was highly sensitive to inhibition by physiologic concentrations of intracellular Ca2+.88 Whether the amplified current was due to loss of ICRAC regulatory factors or to association with another protein(s) that modified ICRAC activity was unclear and the singular focus has remained on store-operated ICRAC and its putative Ca2+-specific channel, CRAC. However, CRAC defied description until recently with the discoveries of a Ca2+-specific channel protein called Orai1 and the ER Ca2+ sensor, STIM1, as described in Sections VI and VII.
At present, the exact physiological role of transient receptor potential (TRP) channels (TRPCs) in calcium signaling is still an issue of debate. Shortly after the initial description of ICRAC, Hardie and Minke89 noted that the PLC-dependent Drosophila photoreceptor Ca2+ channel, TRP, had the attributes of the putative store-operated Ca2+ channel for SOCE in vertebrate cells. This possibility was examined by many investigators after the cloning of the mammalian homologs of the TRP gene,90,91 which were later classified as canonical TRPs or TRPCs. Although it was clear that knockdown of some TRPCs diminished SOCE and when overexpressed they were activated by store depletion, the verdict was mixed. In particular, none of the TRPCs exhibited the exact characteristics of ICRAC (reviewed in Refs. 91 and 92). In contrast to ICRAC, TRP channels can conduct Sr2+ and Ba2+ and have single-channel conductances several orders of magnitude greater than ICRAC. We argue later that these properties are still relevant to the characteristics of SOCE in mast cells.
The TRPCs, of which there are seven members, are now recognized as a subset of a super-family of TRPs that include TRPC, TRPM (melastatin), TRPV (vanilloid), TRPA (ankyrin), TRPP (polycystin), and TRPML (mucolipin).93 All TRPC channels and a few members of other TRP subfamilies have been considered as SOCE channels. The TRP channel subunits consist of six transmembrane segments (S1–S6) with the N-and C-terminal regions located within the cell. A pore-forming loop exists between S5 and S6 and formation of the complete cation channel requires assembly of homo- or heteromeric complexes of four TRP subunits. TRPCs, in addition, contain a common motif, the TRP box, in the carboxy-terminal tail and several ankyrin repeats in the N-terminal region.
All TRPCs are activated as a consequence of activation of PLC.93 Some but not all TRPCs can be activated by store depletion, although which ones actually operate as store-operated channels is still controversial.94 Some are activated by diacylglycerol (i.e., TRPC2, −3, −6, and −7), whereas others are not (i.e., TRPC1, −4, and −5). Like other TRPs, TRPCs can form heteromers among themselves as, for example TRPC1 with TRPC3, TRPC4, TRPC5, or TRPC7 and TRPC3 with TRPC6 or TRPC7.95 As a consequence, current properties are significantly altered. Single-channel conductances range from ~ 15 to 75 pS and exhibit a varying selectivity toward Ca2+ over Na+ from 1.1 to 9.93
It has been suggested that TRPCs can be functionally organized into segregated signaling complexes or “channelsomes” according to composition of the tetrameric TRPC complex and the associated accessory proteins.95 As with the IP3Rs, the accessory proteins help regulate the cellular location and function of the TRPCs in a cell- and stimulus-specific manner. These proteins include those involved in cytoskeletal interactions, vesicular trafficking, scaffolding, and calcium signaling and include PLCβ, calmodulin, IP3R, immuno-phillins, STIM1, plasma membrane Ca2+ ATPase (PMCA), and the sarco/endoplasmic Ca2+ ATPase, SERCA.96,95 We shall discuss the latter proteins in subsequent sections. The ankyrin regions of TRPCs appear to account for many of the interactions of TRPCs with their associated proteins. The composition of the signaling complex is thought to determine not only the channel properties but also the cellular location and downstream functions of TRPCs. The concept is that in addition to contributing to SOCE, TRPCs like IP3Rs may regulate [Ca2+]i in discrete locations or modify the amplitude of the calcium signal in localized regions for specialized cellular functions.95 In other words, the variable subunit composition of TRPC channels and ability to interact with regulatory subunits and other channel proteins provides flexibility of the cell to use TRPCs in diverse ways.97
The TRPCs contain an IP3R consensus binding domain but the regulatory consequences have been defined only for TRPC1, TRPC2, and TRPC3. In general, IP3R positively regulates TRPC function. TRPC1 in the plasma membrane is reported to couple to IP3R in the ER,98–100 an interaction facilitated by RhoA99 or Homer,100 and thus regulate SOCE in conjunction with STIM1.98 As discussed in the next section, STIM1 acts as a Ca2+ sensor in the ER and can interact with and activate some TRPCs. The possible functional combinations of TRPC1 and other TRPCs with IP3R and STIM1 are topics of a recent review.95
The uncertainty about the exact role of TRPCs stems in part from the modest selectivity toward Ca2+, the variability in channel properties of different TRPC complexes, and the lack of correspondence to ICRAC. Many of the early studies relied on overexpression of TRPCs and the potential pitfalls of this approach are now well recognized. These include promiscuous interactions with endogenous TRPCs and other proteins or aberrant localization within the cell. As a result, the levels of expression may result in altered properties as exemplified in studies with TRPC3 where TRPC3 exhibited either store-dependent or store-independent activation when expressed at low or high levels, respectively.101 The introduction of siRNA technology and use of TRPC knockout mice circumvent these pitfalls, but even these approaches can introduce new ones such as compensatory expression of other endogenous TRPCs.102
The field of calcium signaling was galvanized recently by the identification of the ER Ca2+ sensor, STIM1, and the Ca2+-channel protein, Orai1 (also called CRACM1), as the long-sought-for components underlying ICRAC.103–107 Over-expression of both, but not individually, in a variety of cells results in large CRAC currents.108–110 Also, a mutated form of Orai1 was found to be responsible for severe combined immune deficiency with defective Ca2+ influx in T cells that could be rectified by expression of wild-type Orai1.105 Since STIM1 and Orai1 are featured topics in many contemporary reviews,92,111–114 some focused exclusively on immunological cells,115–118 we shall summarize the major findings in this and the subsequent section.
Orai1 is an atypical cation channel with four transmembrane domains.105–108,110 At present, three mammalian Orai gene products have been identified in a wide variety of cells. All three contain a proline/arginine-rich region in the N-terminus and a putative C-terminus coiled-coiled domain, with predicted probabilities that are severalfold higher for Orai2 and Orai3 than Orai1.119 An extracellular loop between the third and fourth transmembrane domains of Orai1 contains an N-glycosylation site but this does not appear to be functionally critical.120 The selectivity filter of Orai1 is linked to acidic residues in the first and third transmembrane domains and the first loop segment. However, a CRAC channel is created only after oligomerization of Orai1121,122 to form a fully functional channel of four Orai1 subunits.123,124 This was demonstrated by coexpression of preassembled Orai1 multimers of varying numbers along with STIM1. In heterologous expression systems, Orai1 can also oligomerize with Orai2 and Orai3 to create CRAC channels in the presence of STIM1, each with slightly different ion-selectivity profiles, feedback inhibition by cytosolic Ca2+, and responses to 2-aminoethoxydiphenyl borate (see Section IX.C).125 The physiological relevance of these various combinatorial arrangements are unclear but it could add flexibility to the “tool kit” that is available for regulation of calcium signaling in a given cell. All three Orai proteins can be activated by STIM2 as well as by STIM1.126 The historical sequence of key observations in these and earlier studies of calcium signaling are illustrated in Figure 2.
While it is now apparent that STIM1 and Orai1 are sufficient to reconstitute CRAC channels, Orai1 also forms complexes with some TRPCs and in conjunction with STIM1 creates SOCE channels with properties distinct from ICRAC.127–130 These interactions include TRPC1 or TRPC5 with Orai1127,129–131 and TRPC3 or TRPC6 with Orai1, Orai2, or Orai3.128 Although many of these studies were based on heterologous expression systems, the interactions were also apparent by coimunoprecipitation of endogenous proteins,127,131 by knockdown of individual endogenous components of the presumed STIM1/channel complex,129 or by introduction of neutralizing antibodies.130
There are two views as to or whether or not TRPCs are directly involved in SOCE. One view is that STIM1 and Orai1 create functional CRAC channels and are the exclusive SOCE channels. Another is that STIM1, Orai1, and TRPCs can also form SOCE channels with varying degrees of Ca2+ selectivity.131,132 In support of the latter view are the interactions just noted, that not all cells bearing Orai1 generate ICRAC 133–135 (see also Section IV), and that coexpressed TRPC1 and STIM1 are able to activate a nonselective current (ISOC) following store depletion.131 ISOC is blocked by low concentrations (~ 1 µM) of La3+ and 2-amino-ethoxydiphenyl borate and is attenuated by Orai1 siRNA or expression of inactive mutants of Orai1 to indicate a dependence on Orai1.131 One proposal is that Orai proteins (i.e., 1, 2, or 3) act as a regulatory subunit of TRPCs (i.e., 3 or 6) to confer sensitivity to STIM1 following store depletion.128 The ability of Orai proteins to interact not only among themselves but also with TRPCs would add further flexibility in the assembly of calcium channels from the calcium tool kit within cells. At the moment, the field lacks physiologic examples of how these various combinatorial arrangements are used by cells for specific purposes.
STIM1 and STIM2 were initially identified as potential tumor suppressors136,137 and only some time later as regulators of SOCE.103,104,107 Both proteins are ubiquitously distributed in cells and can cooligomerize to indicate possible functional interactions.137 Both are highly homologous with single transmembrane domains, an N-terminal helix-loop-helix (EF hand) motif and a sterile a motif (SAM) within the ER lumenal segment, and protein-interacting domains in their lumenal and cytoplasmic portions.103,104,138 The STIM proteins can sense the state of depletion of ER Ca2+ stores by virtue of the EF-SAM domains that have an affinity for Ca2+ that is appropriate for the high concentration of Ca2+ within the ER lumen.139,140 However, the STIM isoforms differ markedly in their rates of oligomerization and dissociation, which may portend differences in their regulatory roles in calcium signaling.141
STIM1 was the first isoform to be recognized as a regulator of SOCE through RNA interference screens in thapsigargin-stimulated Drosophila104,107 and HeLa cells.103 When ER Ca2+ stores are full, STIM1 is localized throughout the ER network in structures organized by the microtubule network142 but when Ca2+ stores are depleted, STIM1 oligomerizes and migrates into punctaelike structures at the cell periphery.103,143 Viewpoints differ as to whether STIM1 actually transfers to the plasma membrane following store depletion143 or merely redistributes into punctae beneath the plasma membrane144,109 to enable activation of ICRAC. In addition to acting as an ER sensor, STIM1 is reported to also regulate ICRAC from within the plasma membrane.145 Mutating the EF-hand results in constitutive localization of STIM1 in punctae and activation of ICRAC independently of store depletion.143,144 A coiled-coil domain in the C-terminus is crucial for its oligomerization as is a serine-proline-rich region for correct targeting of the STIM1 complex at the cell periphery.146,147 The process is reversed on store repletion and, paradoxically, by the myosin light chain kinase inhibitor ML-9, which acts independently of the light chain kinase.148
When located in peripheral puntae, STIM1 is placed in apposition to Orai1, resulting in Ca2+ entry.108–110 Although the mechanism of Orai activation by STIM1 is not entirely clear, functional interaction of both proteins is dependent on the coiled-coil domain in the cytoplasmic C-terminal tail of Orai1149 and two distinct C-terminal domains of STIM1150 as indicated by genetic manipulations that impact on these domains. The interaction may also occur in the context of macromolecular complexes that contain an unidentified component(s) in addition to STIM1.148,151
As noted earlier (Section VI), STIM1 also interacts with various TRPC channels, either individually or as heteromeric combinations with Orai,132,152,153 and appears to do so though electrostatic attraction between positively charged residues in the polybasic domain of STIM1 and matching negatively charged residues in conserved domains of TRPC1 or TRPC3.153 Since these domains are not required for functional coupling of STIM1 to Orai1, the coupling mechanisms for Orai and TRPC appear to be different. STIM1 has also been linked to activation of SERCA3 in the refilling of acidic Ca2+ stores in platelets.154
STIM2 has been less well studied. STIM2, when expressed by itself, inhibits SOCE,155 but when expressed with Orai1, it stimulates Ca2+ influx without store depletion.108 Expressed STIM2 localizes in ER but only migrates into punctae following store depletion if it is coexpressed with STIM1.155 An RNA interference screen has identified STIM2 as a particularly strong regulator of basal [Ca2+]i possibly acting as a feedback modulator that keeps ER and cytosolic Ca2+ concentrations “within tight limits.”138 Furthermore, STIM2 migrates to peripheral punctae in response to relatively small decreases in concentrations of ER Ca2+ and thus stimulates Ca2+ influx via Orai1.138 In contrast, STIM1 comes to the fore with more profound store depletion. This was indicated in studies of T cells and fibroblasts from mice with conditionally targeted alleles of STIM1 and STIM2.156 Deficiency of STIM1 severely impaired thapsigargin-induced SOCE in these cells, whereas deficiency in STIM2 had much less but still significant effect. Other studies based on overexpression of STIM proteins and whole-cell dialysis suggest that store-dependent and store-independent mechanisms exist for regulating coupling of STIM2 with Orai1 and the associated activation of ICRAC.126 The store-independent mechanism was unmasked by cell dialysis and was attributed to washout of a cytosolic inhibitory factor, tentatively identified as calmodulin. In this model, STIM2 is viewed as constitutively active under basal conditions and negatively regulated by elevation of [Ca2+]i through calmodulin. In this respect, the model is a variation of the feedback model noted above.
PMCA removes Ca2+ from the cytosol to the cell exterior and SERCA transfers cytosolic Ca2+ to ER, and both operate against a concentration gradient. They share many of the same structural and functional features and belong to a large family of the so-called P-type ATPases because the reaction cycle involves formation of a phosphorylated intermediate.157,158 The recent determination of the crystal structure of SERCA159 has confirmed that counterion (i.e., H+) transport is an obligatory part of the reaction mechanism.158 In addition to PMCA, Na+/Ca2+ exchangers also participate in the removal of excess Ca2+ from the cell. It is thought that they enable rapid extrusion of Ca2+ because of their high capacity. The more metabolically demanding PMCAs may then act to fine-tune the calcium signal and ensure a return of [Ca2+]i to the final basal levels.160
As noted in Section I, both uptake of Ca2+ into the endoplasmic reticulum and efflux of Ca2+ from mast cells are ATP-dependent processes. Although it is assumed by analogy with other types of cells that Ca2+-ATPase pumps participate in these processes in mast cells,161 the particular isoforms of PMCA and SERCA involved have not been determined. SERCA is probably responsible for uptake of Ca2+ into ER stores because thapsigargin effectively blocks this uptake and localizes within ER of RBL-2H3 cells.161 However, PMCAs appear to share an equal role with Na+/Ca2+ exchangers in the extrusion of Ca2+ from mast cells (see Section 2 below).
A plasma membrane Ca2+ ATP-dependent transporter was first described in erythrocyte ghosts in 1966162 and subsequently purified from several tissue sources. The purified plasma membrane transporter was found to have the same basic features whatever the source and was classified as a P-type ATPase for reasons noted above.157 Other identifying features include stimulation by direct interaction with calmodulin and inhibition by vanadate. The complete sequences of the mammalian pumps were first reported in 1988163,164 and they are now generally referred to as PMCAs. To date four PMCA isoforms, 1–4, and many of their spliced variants have been cloned.160,165 They contain 10 transmembrane regions. The second cytosolic loop between the fourth and fifth transmembrane segments contains a critical aspartate residue and an ATP-binding domain. The aspartate residue is phosphorylated during each cycle of Ca2+ transport. The C-terminal domain possesses a Ca2+/calmodulin-binding region although alternative splicing results in considerable variation in the affinity of different spliced variants for calmodulin. At low concentrations of [Ca2+]i (50 to 100 nM), the C-terminal domain interacts with the first and second cytosolic loops thus masking the ATP-binding and aspartate sites making the PMCAs inactive. This inhibitory state is reversed at higher concentrations of [Ca2+]i by the interaction of Ca2+/calmodulin with the C-terminal domain. The multiplicity of PMCA variants is thought to accommodate the specific needs of different types of cells for Ca2+.166 This is likely reflected by the differences in Ca2+ extrusion properties and tissue distribution of the various PMCA isoforms and their variants. In addition to activation via calmodulin, PMCA activity may be further modulated by other factors. These include protein kinases and acidic phospholipids, particularly the inositol phospholipids, sphingosine, and phosphatidyl serine. PMCAs also interact with other proteins, sometimes in an isoform/variant-specific manner, to allow assembly of multiprotein complexes within specialized membrane domains as described elsewhere.160,165
In addition to restoring [Ca2+]i to basal levels, PMCAs are thought to influence the dynamics of Ca2+ waves, spikes, and oscillations (see Section X.D). There is compelling evidence that PMCAs not only help shape transient increases in [Ca2+]i during brief periods of SOCE, but also help stabilize the calcium signal during sustained SOCE in part because of an ability to adapt rapidly to CRAC channel activity167 and act in concert with other components of the calcium signaling machinery.160 In cells with specialized functions, PMCAs are concentrated in locations specific for those functions, as for example, in the apical membranes of pancreatic and salivary gland cells.168,169 The presumption is that these cells possess machinery, as yet undefined, for recruitment of PMCAs within membrane microdomains,165 as might well be the case for other proteins involved in calcium signaling.
The functions of individual PMCA isoforms have not been well defined but gene deletion in mice indicates little redundancy among their functions. PMCA1 knockout mice are embryonic lethal, whereas knockout of PMCA2 and PMCA4 genes results in more specific mouse phenotypes.170 The lethality of PMCA1 deficiency and the ubiquitous distribution of this isoform suggest that it has a generic function in the regulation of cellular calcium. Other isoforms, or combination of isoforms, may have more specialized tissue-specific functions. As far as we are aware, there are no reports of the effects of these genetic deletions on mast cell function.
Three genes encode SERCA1, SERCA2, SERCA3, and their splice variants.160 The sensitivity of these proteins to thapsigargin and another inhibitor, cyclopiazonic acid, is discussed in Section IX.A. Expression of SERCA1 isoforms is restricted almost exclusively to fast-twitch skeletal muscle fibers. SERCA2a is expressed predominantly in cardiac muscle and SERCA2b is found in almost all tissues including smooth muscle. SERCA3 isoforms are found, often in conjunction with SERCA 2b, in non–muscle cells. Although SERCAs and PMCAs have common molecular and functional features, SERCAs lack an extended C-terminal tail and transport two Ca2+ ions for each ATP molecule consumed as opposed to one Ca2+ ion for PMCAs. Among the Ca2+ transporters, SERCAs have the highest affinity for removal of Ca2+ from cytosol, which probably accounts for the efficient recapture of released Ca2+ by ER from the cytosol such that several Ca2+ oscillations can still occur in the absence of external Ca2+. It has also been postulated that in circumstances where ER is in close proximity to SOCE channels, SERCA may ensure capture of Ca2+ to allow refilling of ER with minimal perturbation of [Ca2+]i.171
SERCAs appear to be regulated directly by the concentrations of free Ca2+ in ER, which is typically held at ~ 500 µM.172 The Ca2+-sensing mechanism for SERCA is not as well defined as that for PMCA. Suggested mechanisms include interactions of SERCA2b with calreticulin and calnexin,173 SERCA3 with STIM1,174 SERCA2b with presenilins,175 and Xenopus SERCA with the second messenger cyclic ADP-ribose,176 among others. Calreticulin and calnexin are ER chaperone proteins primarily recognized for their role in protein folding. Calreticulin resides within the ER lumen and contains a high-capacity Ca2+-binding domain of low affinity. Calnexin is a related transmembrane ER protein whose lumenal portion shares homology with calreticulin but lacks the high-capacity Ca2+-binding domain. Coexpression of either protein with SERCA2b, but not SERCA2a, inhibits Ca2+ oscillations in Xenopus oocytes, and certain domains of calnexin and calreticulin appeared to be critical for this inhibition.173 Overexpression of wild-type presenilins accelerates ER Ca2+ uptake and presenilin-deficient cells have a phenotype similar to SERCA knockdown cells.175 Clearly, additional studies are required to verify the physiologic significance of these interactions to gain a broader view of the situation.
In addition to the established roles of SOCE and ATP-dependent Ca2+ channels in calcium signaling, Na+/Ca2+ exchangers (NCXs) should also share this limelight. These exchangers are widely expressed177 and have been categorized as K+ independent (designated as NCX) or K+ dependent (designated as NCKX). Those of the former category are encoded by a family of three genes (NCX1, NCX2, and NCX3) and the latter by a family of five genes (NCKX1, NCKX2, NCKX3, NCKX4, and NCKX5). While both families of exchangers are bidirectional and are driven by the electrochemical Na+ and Ca2+ gradients across cell membranes, the NCKX family members are also dependent on the K+ gradient.177 Although the normal function of the Na+/Ca2+ exchangers is the extrusion of Ca2+ in exchange for Na+ when [Ca2+]i is elevated, ion exchange can be reversed by, for example, removal of external Na+ to cause influx of Ca2+.
Na+/Ca2+ exchangers are known to operate in conjunction with voltage-gated channels to regulate localized changes in [Ca2+]i 177 but they could also operate in conjunction with SERCA to limit increases in [Ca2+]i following activation of SOCE channels.178 Hints that this might be the case comes from studies of the effects of expressed Na+/Ca2+ exchangers. When activated by store depletion, the exogenous exchangers modulate localized and global changes in cellular Ca2+ levels.179,180 This also appears to be true for endogenous Na+/Ca2+ exchangers in mast cells where significant coupling of Na+ and Ca2+ plasma gradients has been observed.181,182 The effects of varying the concentrations of external Na+ on the influx of Ca2+ or Sr2+, changes in [Ca2+]i, and mast cell degranulation suggest that these events are tightly coupled.182 Whole-cell patch clamp recordings indicate that mast cells express both K+-dependent and K+-independent Na+/Ca2+ exchangers (identified as NCKX3, NCKX1, and NCX3), which may account for as much as 50% of the Ca2+ extruded from cells once [Ca2+]i reaches 200 nM.178 A further indication of the prominent role played by Na+/Ca2+ exchange in mast cells is that its reversal by omission of external Na+ significantly diminishes SOCE and increases in [Ca2+]i. In similar experiments in Jurkat T cells, however, Ca2+ was extruded predominantly through PMCA.167 Therefore, the relative contributions of Na+/Ca2+ exchange and PMCAs may vary from one cell type or experimental condition to another.
As with other channel proteins, an expanding list of interacting partners have been described as regulators of NCX activity.177 These include lipids, especially phosphatidylinositol 4,5-bisphosphate, as well as cytoskeletal, scaffolding, and signaling proteins. However, the molecular context in which the ion exchangers operate and their proximity to other Ca2+ channels is still obscure and many questions remain.
There are two sides to the coin in the debate on the etiology of Alzheimer’s disease. On the one hand, there is the well-documented formation of β-amyloid plaques; and on the other, the presumed, but not yet established, dysregulation of calcium homeostasis leading to neuronal “calcium overload” and cell death.183,184 The relationship between these two phenomena remains a matter of investigation. Nevertheless, research on either the inherited familial Alzheimer’s disease (FAD) or the more prevalent sporadic late-onset form of the disease has brought forth two proteins of relevance to calcium signaling. One is PS and the other is the newly discovered protein, CALM1.
PS is expressed as two isoforms, PS1 and PS2, and resides mainly in ER and Golgi. Both isoforms contain multiple transmembrane regions185 and undergo endoproteolytic cleavage to form N-terminal and C-terminal fragments that reassociate within a multimeric protein complex.186 In this form, the PS proteins constitute the core γ-secretase activity that cleaves, among other proteins, membrane β-amyloid precursor protein to release β-amyloid. About 40% of FAD patients express mutated versions of PS that promote formation of the highly fibrillogenic β-amyloid42 protein, which is characteristic of this disease.187 In addition, many of these PS mutations are also associated with perturbed calcium signaling, which is thought also to contribute to the pathology of Alzheimer’s disease.186,188
It was initially shown that several FAD-related PS mutations profoundly affect Ca2+ release from Ca2+ stores and in turn SOCE,189,190 but the mechanisms are still debated.183,184 Recently, the PS proteins have been proposed as candidates for the elusive Ca2+-leak protein in ER based on observations that PS1 and PS2 form Ca2+-permeable ion channels in bilayer vesicles.191,192 However, other workers detected no such ability but concluded instead that PS1 and PS2 interact with IP3R to enhance IP3-mediated Ca2+ release.193 PS proteins have also been implicated as interacting active partners with SERCA175 and the ryanodine receptor.194 The FAD-related PS1 and PS2 mutants exhibit gain of function in the case of the IP3 and ryanodine receptors193,194 and loss of function in the case of the Ca2+-leak protein,183,184 all of which theoretically would lead to increased [Ca2+]i or “calcium overload.” These and earlier studies clearly establish a role for PS proteins in regulating Ca2+ homeostasis in ER although it is premature to declare which of the above mechanisms operate under normal or pathological conditions. We also note that PS is widely expressed in tissues including mast cells (our unpublished results) but the significance of PS, and inherited mutations thereof, in calcium signaling in mast cells still requires investigation.
CALHM1 was recently identified as a risk gene for late-onset Alzheimer’s disease,195 although this is disputed,196 and was thought to encode a previously uncharacterized membrane Ca2+ channel protein.195 CALHM1 has three transmembrane domains but assumes a multimeric structure when expressed in cells and has sequence similarities with the selectivity filter of the NMDA receptor. Expressed CALHM1 generated a large Ca2+-selective conductance in the plasma membrane and enhanced restoration of basal [Ca2+]i on provision of external Ca2+ to Ca2+-deprived cells.195 The pharmacological and channel properties of CALHM1 were quite distinct from those of SOCE, voltage-gated Ca2+ channels, and ryanodine receptors as indicated by use of various pharmacologic inhibitors. In addition, the effects CALHM1 on Ca2+ flux (and the processing of β-amyloid) were impaired by a single nucleotide polymorphism in CALM1 that was found to be particularly prevalent in Alzheimer’s disease. Inevitably, novel findings such as these raise questions that need to be addressed before CALM1 can be considered as a bone fide Ca2+ channel. One question is whether or not CALHM1 forms similar Ca2+ channels in ER, the main location of expressed CALHM1,195 and if so in what orientation. Another is whether expressed CALHM1 recapitulates the exact function of the endogenous protein and if so how are these channels regulated.
Thapsigargin, a skin irritant isolated from Thapsia garganica,197 is perhaps the most widely used pharmacological probe for studies of IP3-dependent calcium mobilization because of its unique ability to block replenishment of IP3-sensitive stores through SERCA. It was found to be a potent stimulant of mast cell degranulation198–200 and, in 1985, to elevate platelet [Ca2+]i by an undefined mechanism.200 Later studies in other types of cells showed that thapsigargin caused loss of Ca2+ from intracellular stores201 without stimulating inositol phosphate formation,33,34 and did so by blocking Ca2+ uptake via SERCA31 into an IP3-sensitive pool33 resulting in spontaneous “discharge” of Ca2+ from that pool.32 Moreover, thapsigagin and receptor-mediated generation of IP3 release Ca2+ from the same pool.33 Thapsigargin acted in the same manner in RBL-2H3 cells where it was apparent that Ca2+ was released from an IP3-sensitive pool and that both antigen and thapsigargin stimulated Ca2+ influx by ostensibly similar mechanisms in which Sr2+ and other divalent metal ions competed for Ca2+ entry.19 The studies with thapsigargin provided the definitive evidence that entry of extracellular Ca2+ was directly linked to depletion of the IP3-senstive intracellular pool.202 Subsequently, thapsigargin was to prove equally valuable in the identification of the molecular components involved in SOCE.
By 1991, thapsigargin was found to inhibit all isoforms of SERCA at nanomolar concentrations in a stoichiometric manner without affecting PMCA or Na+/K+-ATPases.31 Its action could be attributed to binding of one thapsigargin molecule to each molecule of SERCA, presumably at a site that is not shared by PMCAs.203 This site was later identified as a cavity circumscribed by the third, fifth, and seventh transmembrane domains within the cytosolic side of the membrane.204 In this location, thapsigargin interferes with the necessary conformational changes that accompany the ATPase cycle.205 Thapsigargin has very high but differing affinities for all three SERCA isoforms206 and a slow off rate so that its action is virtually irreversible.31,203 Thapsigargin rapidly penetrates cells, because its sesquiterpenelike structure confers high lipid solubility, and its effects on Ca2+ are apparent in RBL-2H3 cells at concentrations as low as 1 nM.19
Cyclopiazonic acid, a mycotoxin from Aspergillus and Pencillium, was originally described as a reversible and highly selective inhibitor of the Ca2+-ATPase in skeletal207,208 and smooth209 muscle sarcoplasmic reticulum. Comparative analysis of the effects of thapsigargin and cyclopiazonic acid on expressed mutated forms of SERCA1a suggest considerable overlap in their binding domains such that both interfere with the coupling of “ATP utilization in the ATPase cytosolic region and Ca2+ binding in the membrane-bound region” of SERCA,210 although other data suggest differences in the binding domains.206 This issue is partially resolved by recent X-Ray crystallographic data with SERCA in its ADP/Ca2+ free state.211 Both inhibitors block Ca2+-ATPase activity by immobilizing the SERCA transmembrane helices albeit on different subsets of transmembrane helices (first, second, and fourth in the case of cyclopiazonic acid). Thus, unlike thapsigargin, cyclopiazonic acid blocks access of Ca2+ to the channel.
As a practical matter, cyclopiazonic acid has much lower affinity for all SERCA isoforms than thapsigargin206 and its action is reversible. Data on the effects of cyclopiazonic acid in mast cells are quite limited. It was shown to cause store depletion and activate a Ca2+-influx pathway that was permeable to both Mn2+ and Ca2+ in RBL-2H3 cells38 and, in this respect, was reminiscent of the nonselective 50 pS influx pathway detected by patch clamp techniques (see Section VI).86 Ca2+ stores in RBL-2H3 cells and BMMC can be depleted by cyclopiazonic acid as well as by thapsigargin and antigen.212 The subsequent Ca2+ influx, but not store depletion, is impaired by microtubular disrupting agents and attributed to altered cytoplasmic distribution of ER and loss of communication with putative SOCE channels in the plasma membrane.
The Adenophostins (A and B), products of Penicillium brevicompactum,213 and their synthetic derivatives214 are the most potent IP3R agonists described to date. Adenophostin A has almost 100-fold greater affinity for IP3R than IP3215 and a mechanism has been proposed to account for this greater affinity.216,217 Otherwise the interaction of Adenophostins with IP3R is indistinguishable from that of IP3 at the intracellular and single-channel levels. Adenophostin A elicits a quantal pattern of release of Ca2+,218 intracellular Ca2+ oscillations,219 and changes in electrophysiological parameters that are reminiscent of IP3.220 However, Adenophostin A diffuses through the cytosol and detaches from IP3R more slowly than IP3. Consequently, with low concentrations of Adenophostin A Ca2+ release, the associated Ca2+ oscillations and Ca2+ entry are spatially restricted to those regions that are presumed to contain sufficient Adenophostin A to activate IP3R.221 Such studies suggest that Ca2+ oscillations and entry do not extend into regions beyond the range of activated IP3Rs.
Because the Adenophostins are cell impermeant, their effects on calcium signaling have been studied by conventional patch clamp techniques with single cells, particularly the Xenopus laevis oocyte and the RBL1 mast cell line. In RBL1 cells, Adenophostin A can induce Ca2+ oscillations when Ca2+ entry into the cell is blocked by voltage clamp to indicate that the oscillations are regenerative; that is, reuptake of Ca2+ into ER replenishes stores sufficiently for another cycle of release and reuptake.219 Unlike IP3, Adenophostin activates ICRAC at low buffered concentrations of intracellular Ca2+.219,222,223 This has been attributed to reduced sensitivity to the normal Ca2+-dependent inactivation of either IP3R222 or the CRAC channel219 or alternatively to another unique action of Adenophostin.223 As noted earlier, the activation of CRAC channels by Adenophostin A, thapsigargin, or IP3 is substantially diminished by mitochondrial depolarization.50
Pharmacological antagonists of IP3R-mediated Ca2+ release are few and all have limitations.224 Heparin has been used for such purpose but it is cell impermeable and it has multiple actions including activation of ryanodine receptors. The membrane-permeant xestospongins225 and 2-APB226 were initially identified as IP3R antagonists but their mechanisms of action remain undefined and it is now apparent that 2-APB targets other entities that regulate calcium signaling.
The xestospongins (B, C, and D) are macrocyclic bis-1-oxaquinolizidines present in the sponge Xestospongia sp and are thought to be allosteric inhibitors of IP3R. Since the initial report that xestospongin C inhibited IP3-elicited Ca2+ release without affecting IP3 binding,225 xestospongin C has been used in various types of cells and tissues for this purpose. In RBL-2H3 cells, for example, xestospongin C (3–10 µM) inhibited the transient antigen-induced increase in [Ca2+]i in the absence of external Ca2+, influx of Ca2+ in the presence of external Ca2+, and degranulation.227 IP3-induced depletion of Ca2+ stores was also inhibited. However, depletion of stores by thapsigargin and the subsequent influx of external Ca2+ were unaffected. Other reports have implied that xestospongin also inhibits SERCA in several228–230 but not all tissues.231,232 This and other unresolved issues about the specificity and mechanism of action of the xestospongins limit interpretation of published experimental data. Other drawbacks include their slow action and limited availability.
2-APB was initially reported to inhibit IP3-induced Ca2+ release from stores without affecting IP3 production or binding.225 Nor did it appear to activate ryanodine receptors or voltage-operated Ca2+ entry. Subsequent studies with cell membranes, permeablized cells and intact cells pointed to inconsistent actions of 2-APB on IP3R in different types of cells and various explanations were proffered for this inconsistency. These include differences in sensitivity of the different isoforms of IP3R to 2-APB and a dependency on ambient IP3 concentrations that may outcompete 2-APB as discussed in a recent review.224
Interest in 2-APB was enhanced by the report that 2-APB blocked IP3R-mediated SOCE.233 Studies since then have shown that high concentrations of 2-APB can reliably block SOCE, independently of its effects on IP3R,234–237 by interacting with components on the exterior surface of the cells.88,235,238,239 Moreover, 2-APB blocks SOCE whether mediated by ICRAC 235,236 or TRPC channels.234,237,239–242 The latter include heterologously expressed TRPC1,240 TRPC3,234,241 TRPC5,239,242 TRPC6, and several TRPMs.239,243,244 However, the action of 2-APB is biphasic, enhancing ICRAC at low concentrations (1–5 µM) and suppressing ICRAC at high concentrations (10 µM or greater) in IP3R-defective DT40 B cells, Jurkat T B cells, and RBL mast cells.235 This biphasic action has recently been examined in heterologous expression systems, and differences in 2-APB sensitivity were noted among the Orai proteins.245–248 In one report, Orai1 was weakly activated, Orai2 was unresponsive, and Orai3 was strongly activated by 2-APB, but when coexpressed with STIM1, Orai1 and Orai2 were activated at low concentrations and inhibited at high concentrations of 2-APB.246 These stimulatory and inhibitory actions of 2-APB were attributed to a variety of actions including alteration of pore architecture, channel blockade, reversal of store-dependent polymerization of STIM1,246 or reduced accumulation of STIM in punctae near the plasma membrane,247 In all studies, 2-ABP activated Orai3 independently of STIM1 and store depletion in a manner that resulted in a nonselective cation current possibly due to enlargement of the pore size.248 In contrast to the Orai proteins, only inhibitory effects have been noted with TRPCs. 2-APB inhibits TRPC5 and TRPC6 activities in a monophasic manner to suggest that each TRPC channel is blocked by a single molecule of 2-APB.239
Other effects of 2-APB include reversible block of SERCA and “leakage” of Ca2+ from ER pools,249 release of sequestered Ca2+ from mitochondria possibly by acting on the Na+/Ca2+ exchanger,235 and blockade of TRPMs,239,243,244 all of which may be relevant to calcium signaling in mast cells. Despite these and other diverse actions of 2-APB,224,250,251 the compound can be a useful reagent when used cautiously.
The aggregation of high-affinity receptors for IgE (FcεRI) by antigen results in transphosphorylation of adjacent FcεRI subunits and the initiation of phosphorylation cascades that involve Src kinases and the tyrosine kinase Syk. One consequence is the tyrosine phosphorylation of PLCγ, production of IP3, and a store-dependent calcium signal.252 In contrast, stimulation of mast cells through G-protein-coupled receptors is linked to activation of PLCβ, although some of these receptors undergo rapid desensitization and the associated IP3 production and calcium signal are transient and insufficient to promote functional responses.253
Less clear is the role of the sphingosine kinases (SKs) in calcium signaling. The SK product, sphingosine 1-phosphate (S1P), is proposed to be equally essential for Ca2+ mobilization, degranulation,254,255 and the production of eicosanoids and cytokines256 in mast cells. It was originally postulated that IP3 and S1P are both necessary for release of Ca2+ from intracellular stores254 with S1P promoting transient release and IP3 a more sustained release coupled to Ca2+ entry.255 However, recent evidence suggests that Ca2+ entry, and not release, is the process primarily regulated by S1P.256 Potential upstream regulators of the SK/Ca2+ pathway include the Src kinases Lyn and Fyn257 as well as PLD1.255,258 In addition, activation of the SK/Ca2+ pathway, but not the PLC/Ca2+ pathway, appears to be dependent on clathrin, which is thought to facilitate transfer of the lipophilic S1P from the plasma membrane to ER.259
There are, nevertheless, several unresolved issues.260,261 Of the two isoforms of SK, SK1 was claimed to be the predominant regulator of calcium signaling and downstream events in one study255 and SK2 in another.256 In our view, the latter study is the most carefully crafted investigation and is consistent with our own unpublished results. Another issue is whether or not PLD tightly regulates one or both SKs.260 In our findings, the effects of pharmacologic inhibition or knockdown of the PLDs on calcium signaling are relatively modest (Peng, Ma, and Beaven, unpublished data). It should be noted also that S1P acts both intracellularly and extracellularly—after export from mast cells—to activate G-protein-coupled S1P receptors and cell migration.262 This could be a confounding factor in studies with mast cells because extracellular S1P can mediate increases in [Ca2+]i via PLCβ-coupled S1P receptors in other types of cells.263,264 Regardless of these issues, there is consistent evidence that S1P and related sphingolipids are critical for mast cell activation and allergic reactions.265–267
SIP has also been implicated in Ca2+ mobilization following stimulation through Fcγ and G-protein-coupled receptors in other types of cells independently of the classical PLC/IP3/IP3R-pathway.258,268–271 It is reported that S1P is generated in, and releases Ca2+ from, ER-containing vesicles.272 Also, intracellular application of S1P causes release of stored Ca2+,271 and does so independently of S1P receptors on the cell surface.273
Although the accumulated information is intriguing, the determination of the role of S1P and other sphingolipids in calcium signaling is beset with difficulties.261 Early studies, of necessity, relied on the use of the SK inhibitors DL-threo-dihydrosphingosine and N,N-dimethylsphingosine; which may not have the specificity intended. A particular problem with mast cells is the phenotypic flexibility of these cells in vivo and in culture.274 Deletion of molecules, such as SK and PLD, that perform essential housekeeping functions may well result in altered phenotype or compensatory expression of alternate isoforms of these molecules. In addition, several questions need to be resolved and some revisited to make a convincing case that production S1P has an integral, rather than indirect, role in FcεRI-mediated calcium signaling in mast cells. Despite numerous studies, the channel(s) targeted by S1P and by another possible candidate, sphingosylphosphorylcholine, has not been identified.261 The sites of production and mechanisms of delivery of these phospholipids to targets need to be clearly defined. Unlike IP3 they are not freely diffusable in the cytosol.260
The generic scheme for IP3-dependent calcium mobilization in RBL-2H3 cells is shown in Figure 3. It is unknown whether certain details of this scheme vary from one mast cell population to another because FcεRI-mediated calcium signaling has been studied almost exclusively with RBL-2H3 cells. The composition of the SOCE channel is probably a variable feature among different cell types, but we propose that TRPC5 addition to Orai1 and STIM1 participates in SOCE channel activity in RBL-2H3 cells.
As noted earlier, RBL-2H3 cells produce IP3 via PLCγ14 to induce release of Ca2+ from intracellular stores.129 This results in sustained influx of Ca2+ to maintain an elevated [Ca2+]i 11 and replenish IP3-sensitive intracellular stores19 in an ATP-dependent manner15 through a thapsigargin-sensitive pump.19 RBL-2H3 cells express all three forms of IP3R and thapsigargin-sensitive SERCA2b and SERCA3 pumps.62 The relative importance of PMCA and Na+/Ca2+ exchangers in the extrusion of Ca2+ from mast cells was described in Section VIII.B.
Influx is also associated with substantial uptake of Ca2+ into mitochondria, which then releases Ca2+ once [Ca2+]i declines to below the mitochondrial set point for Ca2+.19,20 In fact, most mitochondria in RBL-2H3 cells lie in very close proximity to ER to form ER/mitochondrial junctions275 with access to ER SERCA pumps.276 This situation suggests that ER/mitochondrial junctions may be equipped to deliver Ca2+ effectively to and from the mitochondria, enabling regulation of mitochondrial dehydrogenases.276 Indeed, IP3 induces incremental Ca2+ release from ER and Ca2+ uptake into mitochondria in permeabized RBL-2H3 cells.276 In intact RBL-2H3 cells, transfer of Ca2+ to mitochondria is most efficient during Ca2+ oscillations where the peak [Ca2+]i appears to activate the MCU Ca2+ transporter, possibly through calmodulin.277
All indications point to STIM1 and Orai1 as being essential for CRAC and SOCE channel activity in mast cells. Both proteins are absolutely required for FcεRI-mediated mast cell activation in vitro and in vivo. Orai1/CRACM278 or STIM1279 deficient mast cells exhibit defective calcium signaling, degranulation, and cytokine release. Mice, from which these cells were derived, have impaired IgE-dependent allergic responses. Yet, the known features of ICRAC and Orai proteins are not in accord with the permeability of mast cells to divalent metal ions and Ca2+ (see Section I) or the non-ICRAC currents in RBL-2H3 cells (as described Section IV). Resolution of this conundrum may come from studies of the less selective TRPCs channels, which appear to act coordinately with Orai1 and STIM1. Knockdown of endogenous TRPC5, STIM1, or Orai1 individually with inhibitory RNAs substantially reduces influx of Ca2+ as well as degranulation in RBL-2H3 cells.129 Moreover, overexpression of Orai1 with STIM1 promotes constitutive influx of Ca2+ but not of Sr2+, whereas overexpression of TRPC5 with STIM1 promotes constitutive influx of both ions. These and other data suggest that Sr2+-permeable TRPC5 acts in conjunction with Orai1 and STIM1 to allow Sr2+ and other divalent metal ions to permeate and support degranulation in mast cells129 and possibly amplify channel conduc-tance.88 Given the phenotypic diversity274 and variable expression of TRPCs among mast cell subpopulations (Ma, Iwaki, Gilfillan, and Beaven, unpublished data), additional studies are necessary to determine whether such subpopulations continue to employ TRPC5 or rely on other TRPCs for SOCE.
The communication between STIM1 and Orai1 has been examined by confocal microscopy and fluorescence resonance energy transfer (FRET) in RBL-2H3 cells that expressed fluorescent tagged versions of these molecules.280 In resting cells, STIM1 was associated with microtubules and was diffusely distributed on ER. Following depletion of ER stores by thapsigargin, STIM1 redistributed along with ER which underwent substantial rearrangement to form peripheral punctae where STIM1 in ER and Orai1 in plasma membrane were placed in close proximity. The extent of this interaction may depend on the extent of depletion of ER stores. Antigen stimulation resulted in less extensive colocalization of STIM1 and Orai1 but was more prominent when refilling of stores was blocked by Gd3+. Without Gd3+, maximal colocalization coincided with the initial spike in [Ca2]i following addition of antigen with rapid reversal on decay in [Ca2+]i and refilling of ER stores.280 These reactions may involve ionic interactions between STIM1 and Orai1 because amphiphilic molecules such as D-sphingosine and N,N-dimethylsphingosine inhibit not only IP3-mediated ICRAC 281 but also the FRET-monitored interaction of STIM1 with Orai1 and Ca2+ influx.280 Similar effects were noted by mutation of acidic residues in the cytoplasmic tail of Orai1.280 The proposed model was that the positively charged sphingosines, which flip to the cytoplasmic surface of the plasma membrane, neutralize the Orai1 acidic residues resulting in Orai1 homooligomerization and preclusion of STIM1. Although these findings may imply a potential regulatory role for sphingolipids on this aspect of SOCE, they exclude a direct role of SKs because D-sphingosine is a substrate and N,N-dimethylsphingosine is an inhibitor of SK, and yet both have the same effect.
Ca2+ influx is driven not only by the high Ca2+ concentration gradient, but also by the electrochemical gradient across the plasma membrane21 such that depolarization with high external K+ abolishes Ca2+ influx and degranulation.15 The mechanism(s) responsible for maintenance and restoration of plasma membrane polarity in mast cells is not entirely clear.282 Also, mechanisms may vary from one type of mast cell to another. For example, RBL-2H3 cells express an inwardly rectifying K+ channel, Kir2.1, which sets the membrane potential at about −80 mV,283 whereas resting primary human cell lines are electrically silent with a membrane potential of about 0 mV and no demonstrable Kir current.282 Following antigen stimulation, RBL-2H3 cells partially depolarize,21 due in part to monovalent cation influx via TRPM4 (described later in this section), and then repolarize by a mechanism attributed to K+ efflux,284 possibly through the Ca2+-activated iKCa3.1 (also known as iKCa1) potassium channel.285 In human mast cells, stimulation causes a rapid change in membrane potential from 0 to around −45 mV as a result of activation of iKCa3.1.282,286 Blockade of iKCa3.1 impairs degranulation286 and chemotaxis.287 Interestingly, the expression of functional iKCa3.1 is increased when human mast cells are cultured with stem cell factor (SCF), interleukin (IL)-6, and IL-10, which likely accounts for the enhanced reactivity of these cells as compared to their freshly isolated counterparts.286
Chloride channels have also been implicated in regulating membrane polarity and promoting Ca2+-influx and degranulation in mast cells. However, the evidence is indirect and not definitive. A number of chloride channels, including ClC-2, −3, −4, −5, and −7 as well as CFTR, have been identified electrophysiologically or by RT-PCR in rodent and human mast cell lines. The contributions of these channels to mast cell activation are unclear and the reader is referred to a more detailed discussion of this topic in a recent review.282
Nonselective Ca2+-activated cation channels, of which TRPM4 and TRPM5 are two candidates, are also thought to regulate membrane polarity.166 One such channel, TRPM4, has been identified in mouse BMMC where it promotes membrane depolarization, limits Ca2+ influx, and constrains the extent of mast cell activation.288 TRPM4 deficiency enhances the activation of mast cells in vitro and in vivo. This is manifested by increased FcεRI-mediated Ca2+ entry and release of inflammatory mediators in TRPM4−/− BMMC and increased severity of the acute phase of IgE-mediated cutaneous anaphylactic response in TRPM4−/− mice.288 TRPM4, like TRPM5, conducts Na+ and other monovalent cations but not divalent cations in mast cells.289 Consistent with the proposed role of TRPM4, it is activated by the immunosuppressive compound YM-58483, and this activation is associated with suppression of Ca2+ and cytokine production in lymphocytes.290 Inhibitors include the sulfonylurea, glibenclamide, and hydroxytricyclic compounds such as 9-phenanthrol.291 Consideration of TRPM4 as a therapeutic target in allergic disease as suggested would obviously be predicated on the extent of its functional distribution in different types of cells.
Ca2+ waves were first described in fertilized Xenopus oocytes 292 and since then in electrically excit-able293 and nonexcitable cells.294,295 Ca2+ waves are preceded by localized “Ca2+ sparks”296 or puffs and then spread through the cytosol by Ca2+-induced Ca2+ release from ER.297,298 This is usually followed by regenerative oscillations once ER stores are replenished with external Ca2+. The dynamics of these oscillations may determine the functional output of the cell.298 In RBL-2H3 cells, Ca2+ oscillations16,299,300 are temporally correlated to degranulation301,302 and are dependent on Ca2+ entry to sustain oscillations beyond the initial few oscillations.16 These oscillations are of variable amplitude and duration and are usually superimposed on a rising then decreasing baseline of elevated [Ca2]i.16 The oscillations are not dependent on transient changes in membrane potential or mitochondrial stores nor are they induced by ionomycin16 or thapsigargin.19 Together these results, in the context of studies in other types of cells, suggest that these oscillations result from IP3-induced release of Ca2+ from ER.
Increases in free Ca2+ within the cytosol are essential for many cellular activities such as proliferation, gene expression, secretion, migration, and adhesion. A particular activity may require spatial and temporal resolution of the calcium signal, a few examples of which have been described in previous sections. These and other aspects of calcium signaling have been described in numerous reviews,44,45,58,60,92,95,111–118,303–305 but the question remains as to how a multifunctional cell, such as the mast cell,274,306 directs and regulates its calcium signal to promote a particular response or subset of responses.
The mast cell has remarkedly diverse functional outputs. These include degranulation,307 formation of the eicosanoid precursor arachidonic acid,308,309 chemotaxis,310 and production of cytokines,311–313 and all to some degree are dependent on the generation of a calcium signal regardless of stimulant. Some mast cell–activating ligands such as Toll-like receptor ligands can elicit modest production of cytokines in the absence of a calcium signal but this production is markedly enhanced when mast cells are costimulated with antigen that activates the Ca2+/calcineurin/NFAT pathway.314 Under certain conditions, mast cells can be stimulated to undergo chemotaxis310,315,316 or produce cytokines317 without degranulation. Preferential release of either histamine (a secretory granule constituent) or eicosanoids has been noted with different stimulants267,318,319 and attributed to differences in the amplitude and duration of the calcium signal.319
Examples of coupling of responses to specific components of the calcium signal in mast cells are limited. Early studies indicated that influx of Ca2+ from the external medium was absolutely required for degranulation11 and that Ca2+ oscillations generated by release of Ca2+ from intracellular stores failed to support degranulation in antigen-stimulated RBL-2H3 cells.301 Instead, sustained Ca2+ oscillations or global increases of [Ca2+]i maintained by Ca2+ influx did support degranulation.309,320 A similar tight dependency on Ca2+ influx was noted for the recruitment of protein kinase C and components of the ERK/PLA2/5-lipoxygenase pathway in RBL cells activated with thapsigargin309 or receptor agonists.320 Some mast cell stimulants such as agonists of the adenosine A3 receptor321 and purinergic P2Y receptor320 elicit transient production of IP3 and release of stored Ca2+ without activating influx mechanisms, Ca2+-dependent signals, and degranulation. These studies might imply that downstream effectors are most tightly coupled to influx rather than release of Ca2+. A complicating factor in relating the calcium signal to specific responses is that Ca2+ entry and activation of ICRAC was found to be an all-or-none reaction in individual cells regardless of concentration of stimulant.320 Therefore, the typical concentration-response curves for cell populations may reflect an increasing proportion of cells responding in an all-or-none fashion.
The observations to date do not adequately explain how information is transmitted through the calcium signal to direct mast cells to respond in a particular manner to a given stimulant. This is still largely uncharted territory that needs to be examined at a level that, for example, has been described recently for fibroblast migration where “calcium flickers” guide migration down PDGF gradients.322 Our suspicion is that the details will lie in the spatiotemporal aspects of the calcium signal as outlined in the previous subsection.
We have highlighted several issues that will likely guide future research on calcium signaling, especially in regard to the mast cell. These include the temporal and spatial aspects of the calcium signaling and how these relate to specific cellular responses to agonists; how and where channel proteins and their regulatory partners might be recruited to specialized microdomains within membranes; the mechanisms of coordination of these regulatory microdomains that shape the profile and dynamics of the calcium signal; and the potential flexibility of calcium signaling as, for example, the combinatorial arrangements of Orai proteins and TRPCs that might fine-tune influx to fit the particular needs of the cell. Much has been accomplished in recent years in defining the individual components of calcium signaling, but to gain a more global perspective of the signaling process, mathematical modeling of calcium signaling in mast cells has been attempted161,323 and may be a useful start in this direction. The mast cell has proved to be a useful experimental model for studies of calcium signaling and much primary information has been gained from studies with this cell in the past 40 years. The diversity of its receptors and responses to different stimuli are an additional asset in examining how calcium signals communicate information for specific responses.
This work is supported by the intramural program of the National Heart, Lung, and Blood Institute of the National Institutes of Health.