CALHM1 (calcium homoeostasis modulator 1), a membrane protein with similarity to NMDA (N-methyl-D-aspartate) receptor channels that localizes in the plasma membrane and the ER (endoplasmic reticulum) of neurons, has been shown to generate a plasma-membrane Ca2+ conductance and has been proposed to influence Alzheimer's disease risk. In the present study we have investigated the effects of CALHM1 on intracellular Ca2+ handling in HEK-293T [HEK (human embryonic kidney)-293 cells expressing the large T-antigen of SV40 (simian virus 40)] cells by using targeted aequorins for selective monitorization of Ca2+ transport by organelles. We find that CALHM1 increases Ca2+ leak from the ER and, more importantly, reduces ER Ca2+ uptake by decreasing both the transport capacity and the Ca2+ affinity of SERCA (sarcoplasmic/endoplasmic reticulum Ca2+-ATPase). As a result, the Ca2+ content of the ER is drastically decreased. This reduction in the Ca2+ content of the ER triggered the UPR (unfolded protein response) with induction of several ER stress markers, such as CHOP [C/EBP (CCAAT/enhancer-binding protein)-homologous protein], ERdj4, GRP78 (glucose-regulated protein of 78 kDa) and XBP1 (X-box-binding protein 1). Thus CALHM1 might provide a relevant link between Ca2+ homoeostasis disruption, ER stress and cell damage in the pathogenesis of neurodegenerative diseases
Ca2+ signalling is a key regulation pathway in animal cells. In the nervous system, Ca2+ signalling is essential for learning and memory, neurotransmitter action and synaptic activity, gene expression, excitotoxicity and cell death [1,2]. Perturbations in Ca2+ signalling have been proposed to play a role in the pathogenesis of neurodegenerative disorders, including AD (Alzheimer's disease). The proposed alterations refer both to Ca2+ entry [2–6] or to intracellular Ca2+ handling by the ER (endoplasmic reticulum) [2,7–10] and mitochondria [5,11–13]. Possible involvement of presenilins in Ca2+ homoeostasis and of their mutants in the pathogenesis of AD has received particular attention . Presenilins seem to interact and modulate both Ca2+ influx into the ER via SERCA (sarcoplasmic/endoplasmic Ca2+-ATPase) [15–17] and Ca2+ exit from the ER via interactions with the ryanodine receptors [2,18,19], inositol 1,4,5-trisphosphate receptors [15,20,21] or endogenous leak channels [10,22,23].
A novel gene called CALHM1 (calcium homoeostasis modulator 1), which is expressed preferentially in the nervous system, has been proposed to influence Ca2+ homoeostasis, amyloid β-peptide levels and AD risk . CALHM1 shares strong similarities with the selectivity filter of the NMDA (N-methyl-D-aspartate) receptor channel and is able to generate a PM (plasma membrane) Ca2+ conductance. The CALHM1 P86L polymorphism was reported to be associated with late-onset AD susceptibility . This proposal has been very controversial and a recent meta-analysis has established that the CALHM1 P86L polymorphism is not a genetic determinant of AD, but may modulate age-of-onset by interacting with the effect of the ϵ4 allele of the APOE (apolipoprotein E) gene .
Since CALHM1 localizes to the ER as well as to the PM, it has been speculated without experimental evidence that CALHM1 might increase the Ca2+ leak from the ER [18,24]. This possibility is very interesting, as the increased Ca2+ leak would lead to a decrease in the Ca2+ concentration inside the ER ([Ca2+]ER), which, in turn, could produce ER stress and the UPR (unfolded protein response) [26,27], a reaction that triggers both homoeostatic and pathophysiological mechanisms [13,28,29]. In the present study we directly address this hypothesis by selectively monitoring Ca2+ transport through the ER membrane using ER-targeted aequorin Ca2+ probes. We find that CALHM1 both increased Ca2+ leak and decreased Ca2+ pumping into the ER. As a result, [Ca2+]ER decreases and an UPR cascade is activated.
MATERIALS AND METHODS
Human CALHM1 and its P86L mutant, both cloned in pcDNA3.1 Myc–His, were a gift from Dr Philippe Marambaud (Litwin-Zucker Research Center for the Study of Alzheimer's Disease, The Feinstein Institute for Medical Research, North Shore-LIJ, Manhasset, NY 11030, U.S.A.) . In order to trace the subcellular localization of CALHM1, the CALHM1–mCherry fluorescent protein (CALHM1–Cherry) fusion protein was generated by cutting pcDNA3.1-CALHM1 with EcoRV and XhoI. Cherry was amplified by PCR with a forward primer 5′-atcGTGAGCAAGGGCGAGGAGGATAAC-3′ which introduces half a site of an EcoRV site (in lowercase italics) and a reverse primer 5′-cactcgagTCACTTGTACAGCTCGTCCAT-3′ which introduces a XhoI restriction site (in lowercase italics) at the 3′ end of Cherry. The fragment was cut with XhoI and fused in-frame at the 3′ end of CALHM1 in pcDNA3.1.
Cell culture and gene transfection
HEK-293T [HEK (human embryonic kidney)-293 cells expressing the large T-antigen of SV40 (simian virus 40)] (A.T.C.C. CRL-11268) cells were maintained in DMEM (Dulbecco's modified Eagle's medium; Invitrogen) supplemented with 10% FBS (fetal bovine serum), 2 mM L-glutamine, 100 μg/ml streptomycin, 100 units/ml penicillin and 5 μg/ml Plasmocin™ (InvivoGen) at 37 °C, under an air/5% CO2 mixture. For aequorin experiments, approximately (6–10)×104 cells were seeded on to poly-L-lysine-coated four-well plates and co-transfected with pcDNA3.1-CALHM1 and aequorin cDNA (2:1 ratio) using Lipofectamine™ 2000 (Invitrogen). The experiments were performed at 6 or 24 h after transfections. For quantitative RT (reverse transcriptase)-PCR experiments (see below), 5×105 cells were seeded on to 35 mm plates and transfected with pcDNA3.1-CALHM1 or the empty vector.
Expression of fluorescent proteins and immunofluorescence
The procedure used was as described previously . Briefly, HEK-293T cells were fixed with 4% paraformaldehyde and blocked with 10% goat serum and 0.5% Triton X-100 in TBS (Tris-buffered saline, 20 mM Tris/HCl and 134 mM NaCl, pH 7.5) for 20 min. The fixed cells were incubated overnight at 4 °C with the primary antibody diluted in 1% goat serum and 0.5% Triton X-100 in TBS. A rabbit anti-CALHM1 polyclonal antibody (from Sigma) was used at a dilution of 1:1000. After washing with 10% goat serum and 0.5% Triton X-100 in TBS, the cells were incubated [45 min at room temperature (22 °C)] with Alexa Fluor® 488- or Alexa Fluor® 568-conjugated goat anti-rabbit IgG (1:500 or 1:200 dilution respectively; Molecular Probes) in 1% goat serum and 0.5% Triton X-100 in TBS. Nuclei were stained with DAPI (4′,6-diamidino-2-phenylindole). Cells were mounted in Vectashield® (Vector) and observed with a C-Apochromat ×63 water-immersion objective [NA (numerical aperture), 1.20] in a Zeiss Axioplan Z microscope. The Zeiss ApoTome™ system was used for optical sectioning. Expression of CALHM1 was analysed by imaging the green (excitation wavelength, 490/20 nm; emission wavelength, 540/50 nm) and the red (excitation wavelength, 560/40 nm; emission wavelength, 615/45 nm) fluorescences in fixed cells.
Measurement of aequorin bioluminescence
The chimaeric GFP (green fluorescent protein)–aequorin (termed GA) localized to the cytosol [termed lucGA (luciferase–GFP–aequorin fusion protein)] or the low-Ca2+-affinity probe targeted to the ER (termed ermutGA) have been described previously [31,32]. Cells expressing lucGA were incubated for 1 h at room temperature with 1 μM coelenterazine in a standard incubation medium with the following composition: 145 mM NaCl, 5 mM KCl, 1 mM CaCl2, 1 mM MgCl2, 10 mM glucose and 10 mM sodium-Hepes (pH 7.4). For [Ca2+]ER measurements, cells expressing ermutGA  were reconstituted with 1 μM coelenterazine  by incubation for 1 h at room temperature in Ca2+-free medium containing 0.5 mM EGTA and 10 μM of the SERCA inhibitor TBH (2,5-ditert-butyl-benzohydroquinone) . The experiment was started by washing the TBH with standard medium containing 1 mM Ca2+. Aequorin photoluminescence was measured using a Cairn Research luminescence spectrophotometer as described previously . Calibrations in [Ca2+] were performed using the constant values published previously [33,36]. All of the measurements were performed at 22 °C. In the experiments using PM-permeabilized cells, perfusion was performed with intracellular-like medium with the composition 140 mM KCl, 1 mM KH2PO4, 1 mM MgCl2, 1 mM Mg-ATP, 2 mM sodium succinate and 20 mM sodium-Hepes (pH 7.0), and different Ca2+ concentrations and other additions as detailed below. First, the cells were perfused at approximately 5 ml/min with Ca2+-free medium and then the PM was permeabilized by perfusion with 60 μM digitonin in Ca2+-free (containing 2 mM EGTA) intracellular-like medium for 2 min. The solution was then switched to intracellular-like medium without digitonin and containing 20–1000 nM Ca2+ buffered with EGTA  or 20 μM Ca2+ for 3–5 min, and Ca2+ refilling of the ER was monitored from the emitted photoluminescence. Then the effects of different test solutions were investigated by perfusion in intracellular-like medium (see the Figure legends).
For Ca2+ imaging, the procedure used was as described previously [30,38,39]. Briefly, HEK-293T cells transfected with CALHM1–Cherry were loaded with 4 μM fura 2/AM (acetoxymethyl ester) (Molecular Probes) for 1 h at room temperature in standard incubation medium (see above). Cells were then washed with fresh medium and the cell-containing coverslips were mounted under the microscope (Nikon Diaphot). Test solutions were applied by continuous perfusion at 2–3 ml/min. For fluorescence measurements, cells were alternately epi-illuminated at 340 and 380 nm, and light emitted above 520 nm was recorded using a Hamamatsu Digital Camera C4742–98 handled by Simple PCI 6.6 Hamamatsu software. Consecutive frames obtained at 340 and 380 nm excitation were ratioed pixel by pixel using ImageJ software and calibrated in [Ca2+]C (cytosolic Ca2+ concentration) by comparison with fura 2 standards. Expression of CALHM1–Cherry was analysed by imaging of the fluorescence at 555 nm.
Relative expression of ER stress genes was assessed by quantitative RT-PCR. Total RNA was isolated from cells 24 h after transfection using TRIzol® reagent (Invitrogen). RNA (1–2 μg) was reverse-transcribed using M-MLV (Moloney murine leukaemia virus) reverse transcriptase (Applied Biosystems) and random hexamer as primers. Amplifications of 100 ng of cDNA were performed in triplicate in 20 μl reaction mixtures containing 200 nM primers and SYBR Green PCR master mix (Applied Biosystems). Primers used were: h-GRP78 (glucose-regulated protein of 78 kDa; h- is human)/BiP (immunoglobulin heavy-chain-binding protein) forward, 5′-CGGGCAAAGATGTCAGGAAAAG-3′ and reverse 5′-TTCTGGACGGGCTTCATAGTAGAC-3′; h-CHOP [C/EBP (CCAAT/enhancer-binding protein)-homologous protein] forward, 5′-ACCAAGGGAGAACCAGGAAACG-3′ and reverse, 5′-TCACCATTCGGTCAATCAGAGC-3′; h-ERdj4 forward, 5′-TGGTGGTTCCAGTAGACAAAGG-3′ and reverse, 5′-CTTCGTTGAGTGACAGTCCTGC-3′ [40,41] and human GAPDH (glyceraldehyde3-phosphate dehydrogenase) forward, 5′-ACACCCACTCCTCCACCTTTG-3′ and reverse, 5′-CATACCAGGAAATGAGCTTGACAA-3′. The PCR was run on a Light Cycler 480 system (Roche) and the conditions were: 95 °C for 15 min, 95 °C for 15 s, 60 °C for 30 s, 72 °C for 30 s and 72 °C for 1 min, with 40 cycles of amplification. Relative gene expression was determined using the 2−ΔΔCt method .
Results are expressed as means±S.E.M. Statistical significance was evaluated by Student's t test or one-way ANOVA using GraphPad InStat3 software.
The PM permeability to Ca2+ (PCa) was assessed from the size of the [Ca2+]C overshoots on transition from medium with no Ca2+ to medium containing 2 mM Ca2+ (Figure 1). Confirming previous reports [1,24,43], expression of CALHM1 increased the PCa. In CALHM1-transfected cells, the overshoot was increased up to 8-fold after 24 h of expression (Figure 1A) and approximately 2-fold at 6 h (Figure 1B). The effect of CALHM1 on PCa was fully blocked by nickel (Figure 1A). In the cells expressing CALHM1 for 24 h there was a clear increase in the resting [Ca2+]C (Figure 1A), probably as result of the greatly increased Ca2+ entry, which should follow the increased PCa. In our hands, the effects of wild-type CALHM1 and the P86L mutant on Ca2+ entry were indistinguishable: the Δ[Ca2+]C (mean±S.E.M.; n=4) was 3.99±0.04 for the wild-type and 3.93±0.21 for the mutant (results not shown).
CALHM1 increases Ca2+ entry
Next we compared the uptake into the ER in control and in CALHM1-transfected cells. In order to minimize non-specific deleterious effects, the experiments were performed in cells expressing CALHM1 for only 6 h. Typical results are illustrated in Figure 2(A), and averaged values are summarized in Figures 2(B)–2(D). Ca2+ uptake in the ER was faster in the CALHM1-transfected cells than in the control cells (Figures 2A and 2C); this is, most probably, attributable to a faster PM Ca2+ entry and, consequently, to a larger increase in [Ca2+]C (Figure 1). However, the ER accumulation of Ca2+ at the steady-state was significantly decreased (Figures 2A and 2B). This decrease suggests that either the Ca2+ pumping into the ER is decreased or that the ER Ca2+ leak is increased. An approximate measurement of the Ca2+ exit from the ER has been attempted in the experiments shown in Figure 2(A), by removing external Ca2+ once the ER had been refilled (but note that the decrease of [Ca2+]ER depends on both the exit of Ca2+ from the ER and the extrusion of Ca2+ from the cytosol to the extracellular medium through the PM). These measurements revealed that the Ca2+ exit from the ER was 1.8-fold faster in the CALHM1-transfected cells (Figure 2D). Since [Ca2+]ER is lower in the CALHM1-transfected cells (Figures 2A and 2B), the rate differences should be even larger. If data are expressed as the percentage [Ca2+]ER decay per s, the rates would be 0.25±0.03 for the controls compared with 0.73±0.06 for the CALHM1-transfected cells, approaching a 3-fold increase in the exit rate.
Effects of CALHM1 on the uptake of Ca2+ into the ER
ER passive permeability (leak) can be studied directly in PM-permeabilized cells with blocked Ca2+ pumping by imposing a Ca2+ concentration gradient to drive Ca2+ diffusion. In these experiments cells were co-transfected with ermutGA and either CALHM1 or the empty pcDNA3 vector. The SERCA pump was irreversibly blocked with thapsigargin. Results are shown in Figure 3. After PM permeabilization in medium containing digitonin and no Ca2+ (DIGIT.), ER uptake was started by suddenly increasing [Ca2+] to 1000 μM (Figure 3A). Ca2+ uptake was quite fast and tended to reach a [Ca2+]ER plateau of approximately 2 mM at the steady-state, with half-times of approximately 1 min in the control and 30 s in the CALHM1-transfected cells. At 3 min later, Ca2+ was removed to force Ca2+ exit and [Ca2+]ER decreased quickly, again faster in the CALHM1-transfected cells than in the control cells (Figure 3A). The mean values (±S.E.M.) of several similar experiments are shown in Figure 3(B). Both influx and efflux were approximately twice as fast in the CALHM1-transfected cells, confirming that ER leak is increased, as it was suggested by the results in the intact cells (Figure 2E). Another series of experiments using 0.5 instead of 1 mM external Ca2+ gave similar results (see the legend to Figure 3).
Passive Ca2+ uptake and release (leak) by ER in permeabilized cells
Although the increase of ER Ca2+ leak was clear and statistically significant, we thought that it was not enough to convincingly explain the decrease in [Ca2+]ER at the steady-state. The Ca2+ entry through the PM is faster in the CALHM1-expressing cells (Figure 1); this should increase [Ca2+]C and favour Ca2+ uptake into the ER (as seen in Figures 2A and 2C), which would oppose the [Ca2+]ER decrease. The other determinant of [Ca2+]ER is the active Ca2+ uptake by SERCA. We then studied Ca2+ uptake by the ER in PM-permeabilized cells, in which [Ca2+]C can be varied and controlled as required. Figure 4 summarizes the results of these experiments. Cells were co-transfected with ermutGA and either the empty pcDNA3 vector or CALHM1. After a brief permeabilization of the PM with digitonin in intracellular-like Ca2+-free medium (results not shown; see the Materials and methods section), the cells were perfused with medium containing ATP and different Ca2+ concentrations, from 0.02 to 20 μM, as shown (compare Figures 4A and 4B). The uptake was much slower in the CALHM1-transfected cells, especially at the lower Ca2+ concentrations. Figure 4(C) shows the kinetics of the rate of uptake against the Ca2+ concentrations. Both transport capacity and Ca2+ affinity were decreased. Vmax decreased from 50 to 20 μM/s and K50 increased from 0.3 to 0.9 μM. The [Ca2+]ER reached at the steady-state (Figure 4D) was affected even more by CALHM1 expression, especially near the physiological [Ca2+]C concentrations (10−7 M), where [Ca2+]ER was decreased almost 7-fold (P<0.001). It should be kept in mind that both the decreased ER uptake (Figure 4) and the increased leak (Figures 2D and 3) co-operate to decrease [Ca2+]ER.
Kinetics of Ca2+ uptake by the ER in permeabilized cells
Next we investigated the functional consequences of the reduction in [Ca2+]ER by checking whether the cells expressing CALHM1 developed ER stress. First, we wanted to investigate the role of CALHM1 in the activation of the IRE1 (inositol-requiring enzyme 1)/XBP1 (X-box-binding protein 1) pathway. Under normal conditions, the mRNA for the transcription factor XBP1 contains a premature stop codon and thus produces an immature protein with a short half-life. During ER stress, the activated IRE1 causes the splicing of XBP1 mRNA into a mature mRNA, which in turn produces the functional transcription factor. In the present study we used the ER stress fluorescent reporter XBP1–Venus , which is only expressed when IRE1-mediated splicing takes place. Thus the presence of the green fluorescence in the nucleus indicates, at the single-cell level, the activation of IRE1. In order to also monitor CALHM1 expression at the single-cell level, a CALHM1–Cherry construct was generated and characterized. First, we checked whether the chimaeric protein preserved the Ca2+-conducing ability of CALHM1 (Figures 5A and 5B). For this purpose, cells were transfected with CALHM1–Cherry, loaded with fura 2, and then subjected to a protocol including: (i) a Ca2+ overshoot, by suddenly changing the incubation medium from Ca2+-free (Ca0) to Ca2+-containing medium (10 mM; Ca10), and (ii) an ER Ca2+ release by stimulation with 100 μM ATP+100 μM cabachol (CCh) in Ca2+-free medium. Results were analysed separately in the CALHM1–Cherry-expressing cells (emitting red fluorescence) and in the control cells (not emitting red fluorescence). The cells expressing CALHM1–Cherry (dotted trace in Figure 5A) showed a much bigger Ca2+ overshoot than the control cells (continuous trace). By contrast, the release from the ER after maximal stimulation (ATP+CCh) (second [Ca2+]C peak) was smaller in the CALHM1-expressing cells, indicating that their intracellular Ca2+ stores contained less Ca2+. The mean values from several experiments are shown in Figure 5(B). The Ca2+ overshoot was 7-fold larger in the CALHM1-expressing cells than in the controls, whereas the agonist-induced release from the stores was approximately 3-fold smaller. These results are in perfect agreement with the values obtained for CALHM1 (Figures 1–3), indicating that the fusion CALHM1–Cherry is a functional protein.
CALHM1 expression induces ER stress
Activation of the ER-stress XBP1 gene was studied at the single-cell level in cells co-transfected with CALHM1–Cherry and XBP1–Venus (a gift from Dr Masayuki Miura, Department of Genetics, Graduate School of Pharmaceutical Science, The University of Tokyo and Brain Science Institute, RIKEN, Wako, Saitama, Japan). Expression of both genes was closely associated at the single-cell level, whereas the control cells expressing only Cherry fluorescent protein (without CALHM1) did not activate the XBP1 gene (Supplementary Figure S1 at http://www.BiochemJ.org/bj/437/bj4370469add.htm). Figure 5(C) summarizes the results of six different experiments. The frequency of XBP1–Venus-positive cells within the Cherry subpopulation was 60-fold larger in the cells transfected with CALHM1–Cherry than in the controls transfected with Cherry (P<0.0001), thus demonstrating that CALHM1 induces a strong ER stress.
In order to characterize better the ER-stress response, we investigated the activation of h-CHOP (h is human), h-ERdj4 and h-GRP78, three genes whose induction during UPR has been well documented [45,46]. The expression was measured by quantitative RT-PCR (see the Materials and methods section). Our results, summarized in Figure 5(D), demonstrate a strong induction of all the three genes in CALHM1-expressing cells.
The results of the present study show that CALHM1 induces Ca2+ entry through the PM in HEK-293T cells (Figure 1). This Ca2+ entry was blocked with Ni2+, increased with the time of expression and was able to produce a maintained increase in the resting [Ca2+]C. These results confirm previous reports [24,43]. Apart from its expression in the PM, a vast majority of CALHM1 was localized to the ER, where its function is unknown. It could form, as it has been speculated , a functional Ca2+ channel in the membrane of the ER. Since [Ca2+]ER is far from its electrochemical equilibrium, this channel would facilitate the leak of Ca2+ out of this organelle.
By using ER-targeted aequorin, we have been able to directly measure the Ca2+ fluxes in this compartment. By removing external Ca2+ in order to drive exit from the cells, we observed that CALHM1 also increases the exit of Ca2+ from the ER in intact cells (Figures 2A and 2D). Using PM-permeabilized cells in which ER Ca2+ pumping was blocked by thapsigargin, we found that CALHM1 increased the concentration-gradient-driven passive (leak) Ca2+ fluxes through the ER membrane (Figure 3), and that both influx and efflux were similarly increased, to approximately twice the control values (Figure 3B). The increase in the leak rates in the PM-permeabilized cells was similar to that found in the experiments with intact cells (Figure 2D). The normal pathway for ER Ca2+ leak has not been established , even though many candidates have been proposed [10,27,48]. It is not likely that CALHM1 plays an important role as a universal ER leak channel, since it is weakly expressed in non-neural tissues . The increase in Ca2+ entry and the resulting increase in [Ca2+]C in CALHM1-expressing cells (Figure 1) should result in an increase in [Ca2+]ER. An excessive filling of the ER has been proposed as part of some working hypotheses for the pathogenesis of AD [2,18]. In contrast, we find that the steady-state [Ca2+]ER level was decreased in the CALHM1-expressing cells (Figure 2). This outcome cannot only be explained by the increase in the ER leak (Figure 3) that parallels the PM leak (Figure 1) ; since ER uptake is proportional to [Ca2+]2 , it should be very strongly affected by the increase in [Ca2+]C. Therefore an increase rather than a decrease in [Ca2+]ER should be expected from a parallel increase in the leak of both the PM and the ER membrane. Using the PM-permeabilized cells we found that active ER uptake is substantially modified in the CALHM1-expressing cells, with a decrease in both Vmax and Ca2+ affinity (Figure 4C). The combined action of both increased ER leak and decreased ER pumping, decreases the steady-state [Ca2+]ER by 6–7-fold at the normal resting [Ca2+]C of 10−7 M (Figure 4D). As a result of these alterations, [Ca2+]ER should be permanently low in CALHM1-expressing cells. This decrease in [Ca2+]ER could activate store-operated Ca2+ entry , and this action could contribute to increase the PM Ca2+ entry and the resting [Ca2+]C.
The decrease in [Ca2+]ER activates the ER stress response and UPR . We found in the present study that several UPR genes were turned on in the CALHM1-expressing cells, nominally h-CHOP, h-ERdj4, h-GRP78 and XBP1 (Figure 5). ER stress first helps cell survival by triggering corrective mechanisms that halt translation and increase chaperone production; but if correction is not attained, then UPR initiates cell death programmes [28,40]. It has been proposed that UPR may be involved in the pathogenesis of neurodegenerative diseases [13,28,29,51,52]. The actions of CALHM1 described in the present paper and leading to a decrease in the Ca2+ content of the ER may provide a link between the Ca2+ hypothesis and cell death, and deserves further investigation.
cytosolic Ca2+ concentration
endoplasmic reticulum Ca2+ concentration
calcium homoeostasis modulator 1
C/EBP (CCAAT/enhancer-binding protein)-homologous protein
GFP (green fluorescent protein)–aequorin chimaeric protein
low-Ca2+-affinity ER-targeted GA
green fluorescent protein
glucose-regulated protein of 78 kDa
HEK (human embryonic kidney)-293 cells expressing the large T-antigen of SV40 (simian virus 40)
inositol-requiring enzyme 1
luciferase–GFP–aequorin fusion protein
PM permeability to Ca2+
reverse transcriptase PCR
sarcoplasmic/endoplasmic reticulum Ca2+-ATPase
unfolded protein response
X-box-binding protein 1
Sonia Gallego-Sandín performed most of the experiments. Maria Teresa Alonso and Javier García-Sancho provided conceptual input and designed the experiments. All authors participated in analysis, discussion and interpretation of results, revised the paper, and gave final approval. Javier García-Sancho put together all of the results and wrote the final form of the paper.
We thank Ms Miriam García Cubillas and Mr Jesús Fernández for technical assistance. We also thank B. Duran for assistance in quantitative RT-PCR.
This work was supported by the EU-ERA-Net programme; the Spanish Ministerio de Ciencia e Innovación [MICINN; grant numbers SAF2008-03175-E, BFU2007-60157, BFU2010-17379]; the Instituto de Salud Carlos III [grant number RD06/0010/0000]; and the Junta de Castilla y León [grant number gr175]. S.G.-S. was supported by a postdoctoral JAE contract from the Spanish National Research Council (CSIC).