Store-operate Ca2+ channels gate Ca2+ entry into the cytoplasm in response to the depletion of Ca2+ from endoplasmic reticulum Ca2+ stores. The major molecular components of store-operated Ca2+ entry are STIM (stromal-interacting molecule) 1 (and in some instances STIM2) that serves as the endoplasmic reticulum Ca2+ sensor, and Orai (Orai1, Orai2 and Orai3) which function as pore-forming subunits of the store-operated channel. It has been known for some time that store-operated Ca2+ entry is shut down during cell division. Recent work has revealed complex mechanisms regulating the functions and locations of both STIM1 and Orai1 in dividing cells.
Store-operated Ca2+ entry
SOCE (store-operated Ca2+ entry) is a process whereby a decrease in ER (endoplasmic reticulum) Ca2+ content activates Ca2+ entry into the cytoplasm across the PM (plasma membrane) . Despite intensive research since the initial concept was proposed , the molecular basis for this pervasive signalling pathway was not solved until large-scale genomic screens revealed the two fundamental players: STIM (stromal-interacting molecule) proteins that sense Ca2+ levels in the ER [3,4], and Orai proteins that constitute the basic pore-forming subunits of the store-operated channels [5–7]. Mammalian cells express two STIM proteins: STIM1 and STIM2 . The major Ca2+-sensing protein involved in SOCE appears to be STIM1, since its ablation results in almost complete loss of SOCE in most cell types investigated so far [9–13]. Depletion of Orai1 similarly results in substantial loss of Ca2+ entry in numerous cell types, whereas knockdown of Orai2 or Orai3 has little to no effect, leading to the conclusion that Orai1 is the predominant Orai isoform. Mice lacking STIM1 die perinatally, whereas mice lacking STIM2 may live for 4 weeks or so . Knockout of Orai1 in mice produces a severe phenotype [15,16], similar to that seen with knockout of STIM1, and mouse knockout models of the other mammalian Orais, Orai2 and Orai3, have not yet been reported.
Orai1 is primarily expressed in the PM, although a vesicular pool of variable size has been reported, as discussed below. STIM1 was initially identified as a PM protein , but is also expressed in the ER where it functions as a Ca2+ sensor to signal SOCE. The function, if any, of PM STIM1 in SOCE is unclear, but it has been suggested to play a role in another class of Ca2+ channels activated by arachidonic acid . Within the ER, STIM1 does not distribute randomly, but appears to align partially with the microtubular network, at least when overexpressed with a fluorescent tag . As discussed below, this association is not a direct one, but involves STIM1 binding to a microtubule accessory protein, EB1 (end-binding protein 1) .
The domain structure of STIM1 is shown in Figure 1. The protein traverses the limiting membrane of the ER once. The luminally directed N-terminus contains a Ca2+-binding EF-hand motif that is apparently bound to Ca2+ under resting conditions. There is also a second, somewhat degenerated, EF-hand motif downstream of the first (not shown in the Figure), which probably does not bind Ca2+, but is probably involved in the conformational change that occurs when Ca2+ dissociates, as is the case for other Ca2+-sensing EF-hand proteins . When Ca2+ decreases in the ER, for example in response to the activated formation of inositol 1,4,5-trisphosphate, Ca2+ dissociates from the N-terminus, resulting in a conformational change that permits self-association of STIM1 initially through interactions between SAM (sterile α-motif) domains . STIM1 rapidly forms dimers and subsequently higher-order oligomeric complexes . These oligomeric structures interact with moieties in the PM at close ER–PM junctions resulting in the accumulation of STIM1 in these junctions. Here, STIM1 oligomers induce the co-clustering of Orai subunits and interact with them, probably directly, to induce channel opening and Ca2+ influx [24–27]. The domain within STIM1 that is responsible for activation of Orai channels was termed SOAR (STIM–Orai-activating region) by Yuan et al. , and essentially similar findings were published by other groups shortly thereafter [26,28,29]. These clusters of STIM1 and Orai appear as distinct sites when constructs with fluorescent protein tags are expressed in cells, and the sites of concentrated STIM1 and Orai are generally referred to as puncta. The coalescence of full-length STIM1 into puncta can be readily observed with confocal microscopes. Additionally, since the formation of puncta involves the migration of STIM1 to near PM junctions, store depletion results in a substantial increase in the fluorescence signal measured by TIRF (total internal reflection) microscopy from fluorescent protein-tagged STIM1. Full-length STIM1 will form puncta in the absence of stoichiometric Orai, but Orai will not move to puncta unless there is a comparable quantity of STIM1. Experimentally, this means that overexpressed STIM1 will form observable puncta without the need for overexpressing Orai, but overexpressed Orai will not form observable puncta unless STIM1 is also overexpressed. The ability of STIM1 to form puncta independently of Orai depends upon a polybasic region at the far C-terminus of the protein , leading to speculation that this region may interact with PM polyphosphoinositides. However, strategies that deplete PM phosphatidylinositol 4,5-bisphosphate do not affect STIM1 movements nor do they alter SOCE . In the absence of the polybasic tail, STIM1 can still form puncta that depend entirely upon Orai, presumably through direct interaction with Orai; this interaction results in seemingly undiminished SOCE. Herein two major questions remain: with what does the polybasic region of STIM1 interact, and what is the role of this interaction? It is possible that STIM1 interactions with PM components help stabilize the ER–PM junctions. Consistent with this idea, overexpression of full-length STIM1 increases the size of these junctions . It is noteworthy, however, that STIM in Drosophila has a significantly shorter C-terminus, lacking a polybasic region. Yet Drosophila cells exhibit store-operated currents essentially identical with those in mammalian cells .
Domain structure of STIM1
The vast majority of studies on STIM1 cell biology have been carried out at room temperature, a condition in which STIM1 clearly acts as a depletion-dependent activator of Orai channels by relocating to ER–PM junctions. Interestingly, a recent study demonstrated that, at physiological temperature (i.e. 37°C), STIM1 forms puncta spontaneously in the absence of store depletion, yet does not activate Orai . Apparently, these spontaneous puncta form because higher temperatures favour the interaction between the polybasic tails and its cognate ligand, but inhibit interactions between the SOAR domain of STIM1 and Orai1. Further work is needed to determine whether endogenous STIM1 also exists largely in these junctional areas providing highly localized communication between Ca2+ stores and PM channels.
Ca2+ entry is lost in dividing cells
Volpi and Berlin  first reported that the sustained elevation of [Ca2+]i (intracellular Ca2+ concentration) in histamine-activated HeLa cells was lost during mitosis. Histamine stimulation of 45Ca2+ uptake was lost as well. The initial peak [Ca2+]i signal remained, however, leading to the conclusion that Ca2+ influx was specifically impaired. Interestingly, in their discussion, they consider as possible mechanisms inactivation by phosphorylation or channel internalization, both of which have turned out to be true (although it is not the channel, but STIM1 that is phosphorylated; see below). The first implication of SOCE came a few years thereafter when Preston et al.  demonstrated that thapsigargin-induced sustained Ca2+ signalling, as well as 45Ca2+ uptake, were lost during mitosis. More recently, the regulation of the current underlying SOCE, Icrac, was followed through the cell cycle in the mast cell line RBL-2H3 . Confirming the previous findings, Icrac was substantially suppressed in mitosis, and, interestingly, was significantly increased during G1/S- and S-phases.
SOCE is also diminished during meiosis, suggesting that it is a general occurrence during cell division. However, as discussed below, it is possible that the mechanisms involved in meiosis are somewhat different than in mitosis. Furthermore, the significance of Ca2+ influx suppression during cell division, whether meiosis or mitosis, is completely unknown; it will therefore be important to elucidate the underlying mechanisms in order to begin to address the impact of this regulatory process.
Mechanisms for loss of SOCE during cell division
The two most obvious molecular candidates for sites of disruption of SOCE are STIM1 and Orai1. Therefore, perhaps not surprisingly, there is evidence for modification of both molecules during cell division. Yu et al. , in studies of SOCE during meiosis in Xenopus, noted that during meiosis: (i) STIM1 fails to cluster into puncta, (ii) STIM1 becomes extensively phosphorylated, and (iii) Orai1 is internalized. The latter phenomenon depends upon constant cycling of Orai1 in and out of the PM , a process similar to that which has been reported previously for certain TRP (transient receptor potential) channels [39,40].
The inactivation of SOCE apparently results from the activation of MPF (maturation-promoting factor) [consisting of CDK1 (cyclin-dependent kinase 1) and cyclin B] . Presumably this underlies the hyperphosphorylation of STIM1 that occurs during both meiosis and mitosis [37,42]. However, there are conflicting findings with regard to the role played by phosphorylation in the inhibition of STIM1 clustering. Yu et al.  reported that mutation of multiple potential MPF phosphorylation sites to alanine did not rescue STIM1 clustering during meiosis. On the other hand, Smyth et al.  made large deletions in the far C-terminal region of STIM1 where all of the potential MPF sites are located, thereby preventing phosphorylation (detected by a phosphopeptide antibody, MPM-2 ). When co-expressed with Orai1, in mitotic cells, the truncated STIM1 relocated to near PM sites upon store depletion, and SOCE was substantially restored. In this same study, two of the sites phosphorylated in mitosis were identified: serine residues at positions 486 and 668. Interestingly, Ser668 appears to be a direct target of MPF, at least in vitro. Clearly, however, additional sites were still phosphorylated as revealed by binding of the MPM-2 antibody after mutation of these two sites to alanine. Nonetheless, this two-site mutant produced a partial rescue of peripheral localization and SOCE when co-expressed with Orai1 , further bolstering support for the role of STIM1 phosphorylation in SOCE suppression. Note that in the study by Smyth et al. , it was not possible to rescue SOCE unless the mutant STIM1 constructs were co-expressed with Orai1, both driven by strong viral promoters. This may indicate that in the mitotic cells, a substantial fraction of Orai1 is internalized, as was the case for the meiotic cells.
Why then did these two studies lead to such different conclusions regarding the consequences of STIM1 phosphorylation? It is possible that the mechanism underlying down-regulation of STIM1 function in meiosis is fundamentally different from that in mitosis. In a recent study, Yu et al.  substituted the C-terminal sequence containing all known MPM-2 phosphorylation sites with a fluorescent protein followed by the STIM1 polybasic region. This construct was fully capable of forming puncta and signalling SOCE in interphase cells, yet, despite the absence of any predicted phosphorylation sites, no puncta formed in cells in meiosis (eggs). However, this construct, unlike the unmodified STIM1 protein, was now capable of translocating to near membrane areas and interacting with PM components . Thus it seems that removal of phosphorylation sites did at least partially rescue STIM1 function in meiosis; rescue of Ca2+ entry would probably not occur, since, as these investigators showed previously, during meiosis virtually all Orai1 is internalized . The failure to form distinct and visible puncta is perhaps not a particularly reliable indicator of STIM clustering, since it probably reflects the morphological arrangement of ER–PM junctions , which may themselves be altered during meiosis.
Consequences of STIM/Orai regulation in dividing cells
Studies of the consequences of STIM and Orai regulation may have to await strategies to reverse or rescue both STIM1 function and Orai1 retention in the PM. Speculation has focused on two general ideas: the need to inactivate Ca2+ entry, and the need to prevent STIM1 association with microtubules. Note that these two possibilities may not be mutually exclusive. In order to examine the consequences of rescued Ca2+ entry on cell division, Smyth et al.  overexpressed both the C-terminal truncation and Orai1 in HEK (human embryonic kidney)-293 cells and determined the growth rate of the cells in culture. The control for this experiment was the overexpression of wild-type STIM1 and Orai1 (which does not rescue SOCE in mitosis); unfortunately, this condition resulted in substantial slowing of cell division, possibly owing to the rechannelling of cell resources by the strong viral promoters on the transfected constructs. Nonetheless, cells transfected with the truncated STIM1 grew significantly more slowly than those transfected with the wild-type protein . A clearer assessment may come when the specific sites of phosphorylation have been identified and a more subtle mutation can be tested.
Regarding STIM1 association with microtubules, it is possible that phosphorylation may play a role in regulating this interaction, although this idea has not yet been tested. The interaction of STIM1 with microtubules is not direct, but occurs through a microtubule plus-end tracking protein known as EB1 . This interaction with the growing ends of microtubules results in the comet-like movements of STIM1 seen with TIRF video imaging [13,20]. The STIM1 sequence through which STIM1 binds to EB1 is TRIP (Thr-Arg-Ile-Pro) in the C-terminus, and interestingly this site falls within the domain containing multiple MPF sites . Mutation of this TRIP sequence prevents STIM1 interaction with EB1 and microtubules , and also diminishes the ability of growing ends of microtubules to associate with and extend ER tubules . During mitosis, microtubules rearrange to form the mitotic spindle in the pro-nucleus, whereas ER is largely excluded from this region and distributes to the cell periphery. Thus it is possible that phosphorylation-dependent dissociation of STIM1 from EB1 is required during mitosis to properly segregate the ER from spindle microtubules. It is known, however, that regulation of STIM1–microtubule interaction is unlikely to contribute to the loss of SOCE since previous studies have shown that disruption of this interaction has little impact on SOCE [19,20,46].
SOCE shuts down during cell division by a complex mechanism involving redistribution and internalization of Orai channels as well as phosphorylation and inactivation of the Ca2+ sensor STIM1. Some remaining questions pertain to the signals leading to redistribution of Orai, the kinase or kinases regulating STIM1, and the relationship to STIM1–microtubule interactions. Indeed, it is still not altogether clear what is the physiological significance of this profound functional down-regulation of this widely encountered pathway of cell signalling during the critical period of cell division.
Signalling 2011: a Biochemical Society Centenary Celebration: A Biochemical Society Focused Meeting held at the University of Edinburgh, U.K., 8–10 June 2011. Organized and Edited by Nicholas Brindle (Leicester, U.K.), Simon Cook (The Babraham Institute, U.K.), Jeff McIlhinney (Oxford, U.K.), Simon Morley (University of Sussex, U.K.), Sandip Patel (University College London, U.K.), Susan Pyne (University of Strathclyde, U.K.), Colin Taylor (Cambridge, U.K.), Alan Wallace (AstraZeneca, U.K.) and Stephen Yarwood (Glasgow, U.K.).
Dr Stephen Shears read the paper and provided helpful comments before submission.
Work from our laboratories cited in the present paper was supported by the Intramural Program, National Institutes of Health, National Institute of Environmental Health Sciences.