Calreticulin Modulates Capacitative Ca2+ Influx by Controlling the Extent of Inositol 1,4,5-Trisphosphate-induced Ca2+ Store Depletion*

Calreticulin (CRT) is a highly conserved Ca2+-binding protein that resides in the lumen of the endoplasmic reticulum (ER). We overexpressed CRT in Xenopusoocytes to determine how it could modulate inositol 1,4,5-trisphosphate (InsP3)-induced Ca2+ influx. Under conditions where it did not affect the spatially complex elevations in free cytosolic Ca2+ concentration ([Ca2+] i ) due to InsP3-induced Ca2+ release, overexpressed CRT decreased by 46% the Ca2+-gated Cl− current due to Ca2+influx. Deletion mutants revealed that CRT requires its high capacity Ca2+-binding domain to reduce the elevations of [Ca2+] i due to Ca2+ influx. This functional domain was also required for CRT to attenuate the InsP3-induced decline in the free Ca2+concentration within the ER lumen ([Ca2+]ER), as monitored with a “chameleon” indicator. Our data suggest that by buffering [Ca2+]ER near resting levels, CRT may prevent InsP3 from depleting the intracellular stores sufficiently to activate Ca2+ influx.

in the folding and assembly of glycoproteins (14 -17). ER CRT may also act as a repository of readily releasable Ca 2ϩ , with each mole of CRT binding as many as 20 -25 mol of Ca 2ϩ . Since Ca 2ϩ release from the ER is controlled mainly by the second messenger inositol 1,4,5-trisphosphate (InsP 3 ), CRT emerges as a potential regulator of this ubiquitous signal transduction pathway.
The InsP 3 -induced Ca 2ϩ signal has two main components, the release of Ca 2ϩ from the intracellular stores (InsP 3 -induced Ca 2ϩ release, or IICR) and the influx of Ca 2ϩ across the plasma membrane (InsP 3 -induced Ca 2ϩ influx, or IICI). IICR starts when InsP 3 binds to its intracellular receptor (InsP 3 R), a ligand-gated Ca 2ϩ channel that traverses the lipid membrane surrounding the ER. The resulting discharge of ER Ca 2ϩ into the cytosol is often spatially complex, with periodic focal release of Ca 2ϩ actively spreading to the rest of the cell through mechanisms that remain incompletely understood (18). IICI commonly follows IICR, thereby allowing cells to recharge their internal stores. Stimulation of Ca 2ϩ influx appears closely linked to the depletion of the intracellular Ca 2ϩ stores, a relationship known as capacitative Ca 2ϩ entry. We do not yet understand how Ca 2ϩ store depletion stimulates Ca 2ϩ influx. Store depletion could communicate with the plasma membrane through a diffusible messenger (19,20), through secretion-like vesicular docking (21,22), or through protein-protein interactions that may involve the InsP 3 R itself (23)(24)(25).
Even though Ca 2ϩ storage was the first function ascribed to CRT (1), we do not yet know what role, if any, CRT plays in the regulation of the InsP 3 -induced Ca 2ϩ signal. Whereas gene knock-out experiments indicate that CRT is not essential for IICR (11), overexpression of CRT has been associated with a decrease in the magnitude of IICI in mammalian cell lines (26 -28). These results have prompted the hypothesis that CRT reduces capacitative Ca 2ϩ entry by buffering the free Ca 2ϩ concentration within the ER lumen ([Ca 2ϩ ] ER ) above the levels required to trigger capacitative Ca 2ϩ influx. In this work, our objective was to put this hypothesis to a rigorous test by directly measuring [Ca 2ϩ ] ER in a widely used model cell, the Xenopus oocyte. Our data indicate that CRT requires its high capacity Ca 2ϩ -binding domain both to attenuate the InsP 3induced decrease in [Ca 2ϩ ] ER and to reduce the elevations of [Ca 2ϩ ] i due to IICI. The ability of CRT to buffer [Ca 2ϩ ] ER may thus be functionally relevant to the InsP 3 -induced Ca 2ϩ signal.

EXPERIMENTAL PROCEDURES
CRT and CRT Deletion Mutant Expression Vectors-We inserted the CRT cDNA cloned from the HL60 human leukemia cell line (29) between the PstI and SalI sites of the pMT3 plasmid vector (30). We used the polymerase chain reaction (PCR) to build CRT mutants lacking functional domains. To remove the C-domain (CRT-⌬C), we amplified * This work was supported by grants from the Department of Veterans Affairs (to S. D. and R. A. C.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
* the N and P domains of CRT (amino acids 2-315) and added a KDEL ER retrieval sequence at the C terminus (primers 1 and 2). The PCR fragment was digested with MluI and EcoRI and inserted between those sites in a modified pMT3. For the construct lacking the P-domain (CRT-⌬P), we amplified the N-domain (amino acids 2-197, primers 1 and 3) and the C-domain (amino acids 316 -401, primers 4 and 5) separately. The PCR fragments were then respectively digested with MluI/SpeI and with SpeI/EcoRI and ligated between the MluI and the EcoRI sites of pMT3. Primers sequence were as follows, from 5Ј to 3Ј: primer 1, GTCACGCGTCTGCTATCCGTGCCGCTG; primer 2, CGGA-ATTCTCAGAGTTCATCCTTGCCCAGCACGCCAAAGTTATC; primer 3, CGTACTAGTTTCCAAGGAGCCGGAC; primer 4, GTCACTAGTCT-GGATCTCTGGCAAGTCAAGTCTGG; primer 5, CGGAATTCTCATT-CGAGTCTCACAGAGAC.
cDNA-directed Protein Expression and Western Blotting-We prepared stage V-VI oocytes from albino Xenopus laevis and injected the DNA constructs into the nucleus as described previously (31,32). Oocyte microsome purification, protein solubilization, SDS-polyacrylamide gel electrophoresis, and transfer to nitrocellulose were as outlined in the past (33). For immunoblotting, we used a polyclonal antibody raised against the full-length recombinant HL60 CRT (29). We performed all our functional studies 48 -72 h after plasmid injection, when CRT expression level was found to be maximal in preliminary experiments.
To ensure that the individual cells studied electrophysiologically expressed the protein of interest, plasmid cDNA coding for CRT or CRT mutants was co-injected with pMT3-sg25 (20:1 concentration ratio), a vector coding for an enhanced green fluorescent protein (GFP). We performed electrophysiology experiments on those live cells that showed the highest GFP fluorescence.
Monitoring [Ca 2ϩ ] ER with Fluorescence Resonance Energy Transfer (FRET)-We excised the "yellow chameleon-3er or -4er" (3er or 4er) cDNAs (gifts from Dr. Roger Y. Tsien, Howard Hughes Medical Institute, University of California, San Diego, CA) from the pcDNA vector and inserted them between the NotI and EcoRI sites of pMT3. The 3er/4er Ca 2ϩ indicators contain two mutated GFP yielding cyan (CFP) or yellow (YFP) fluorescence (44). The amino acids linking YFP to CFP include the sequences for CaM and for the M13 CaM-binding peptide. Upon Ca 2ϩ binding, CaM enfolds M13 thereby creating FRET between YFP and CFP. To enable measurements of the relatively high [Ca 2ϩ ] ER , CaM sequences of 3er and 4er were modified to lower their affinity for Ca 2ϩ (44). 3er and 4er also include sequences from CRT that allow their retrieval into the ER lumen. To measure FRET, we illuminated oocytes expressing 3er/4er ( ϭ 440 Ϯ 15 nm, DeltaRam monochromator, Photon Technology International, Monmouth Junction, NJ, 455DCLP dichroic mirror) and recorded the emission intensity at two wavelengths, 485 and 535 nm, using separate photomultiplier tubes (520DCLP dichroic mirror, D485/40M and D5356/30M emission filters, Chroma Technology, Brattleboro, VT). If we could not calibrate 3er in vivo, we could nevertheless use the 535/485 fluorescence ratio to compare [Ca 2ϩ ] ER among cells submitted to identical illumination. In these experiments, the gains of the photomultiplier tubes were carefully matched, and neutral black levels were determined experimentally (45).
Immunolocalization-Oocyte cryosections (5 m thick) were incubated for 30 min first with 1/200 -1/500 dilutions of the anti-CRT antibody and then with a fluorescein isothiocyanate-conjugated antirabbit secondary antibody, and examined with a fluorescence microscope as described before (32). For ultrastructural studies, cryosections (10 m thick) were fixed, incubated for 1 h in anti-CRT antibody and then overnight in anti-rabbit antibody conjugated to 5 nm gold particles (Janssen Life Sciences Products, Piscataway, NJ), and prepared for examination under in a Hitachi 7000 electron microscope (Hitachi Ltd., Tokyo) as described previously (32). 45 Ca 2ϩ Uptake-Oocytes injected either with pMT3 alone or with pMT3-CRT were incubated in 45 Ca 2ϩ -containing solution (8 Ci/ml). 45 Ca 2ϩ -related counts increased gradually and reached a plateau at 36 -48 h, indicating equilibration across the Ca 2ϩ -binding compartments of the oocyte. After a 72-h incubation period, chosen to ensure isotopic equilibrium, intact oocytes were washed 5 times in cold solution and individually transferred to scintillation vials for counting. Alternatively, the washed oocytes were combined and homogenized, and their microsomes were purified and recovered on filter paper prior to counting, as per our previously published methodology (33).
Calcium Imaging and Confocal Microscopy-Oocytes were injected with either fluo 3 alone (250 M), Ca 2ϩ orange alone (250 M), or a mixture of fluo 3 (100 M) and fura red (300 M) (Molecular Probes). They were then stimulated with a 30-nl droplet of Ins(2,4,5)P 3 (10 M) and the resulting fluorescence visualized on a confocal microscope as described previously (46). Ca 2ϩ wave amplitude, velocity, frequency, as well as rates of rise/decrease in [Ca 2ϩ ] i at the wave fronts were determined as detailed before (32). To compare [Ca 2ϩ ] i between different cells using visible excitation wavelengths, we ratioed the simultaneously measured fluorescence signal from fluo 3 and fura red (45). The detailed validation of this technique in the oocyte has been previously published (32). Image analysis was performed using the NIH Image version 1.61 software.

RESULTS AND DISCUSSION
Overexpression of CRT in Xenopus Oocytes-On a Western blot of microsomal proteins from which we have previously purified CRT, the ER lumenal protein calsequestrin and the InsP 3 R (33, 47), an anti-CRT antibody (29), labeled an ϳ60-kDa protein (Fig. 1, 1st lane). After injection with pMT3-CRT, the density of this band increased significantly, indicating overexpression of CRT (Fig. 1, 2nd lane). Immunostaining revealed reticular fluorescence that increased from the vegetal to the animal pole of the cell in a pattern similar to that of control cells, but much brighter (Fig. 2). We observed a similar fluorescence pattern with overexpressed InsP 3 R (32). Under electron microscopy using a gold-labeled secondary antibody targeted at an anti-CRT primary antibody, the gold particles were found in association with membranous cisternae most likely representing elements of the ER (Fig. 3). There was no label associated with mitochondria or with other organelles. Given CRT's ER retrieval sequences and lack of hydrophobic regions that can stretch across lipid bilayers, our data suggest that CRT is overexpressed within the ER lumen. current reflects [Ca 2ϩ ] i just beneath the plasma membrane (48,49), where it is most likely to be affected by Ca 2ϩ influx. The assay integrates the submembranous [Ca 2ϩ ] i changes across the entire surface of the plasma membrane, thereby maximizing our ability to detect changes in the magnitude of Ca 2ϩ influx. In control cells, microinjection of InsP 3 (10 Ϫ3 M in the pipette) causes a biphasic response; there is a short initial increase in [Ca 2ϩ ] i followed by a slow increase (Fig. 4A). The fast initial component of the response, which is not affected by the removal of extracellular Ca 2ϩ , is due to the release of Ca 2ϩ from the intracellular stores (50,51). In contrast, the slow component of the response can be inhibited by decreasing the extracellular Ca 2ϩ concentration or by adding inorganic Ca 2ϩ channel blockers (Mn 2ϩ , Ni 2ϩ , and La 3ϩ ) to the bath and thus depends on Ca 2ϩ influx (50 -52). Although the magnitude of the Cl Ϫ currents due to IICR was similar to that of control cells (272 Ϯ 28 nA in CRT versus 247 Ϯ 23 nA in controls, n ϭ 31 pairs of cells), the Ca 2ϩ influx-dependent Cl Ϫ current was reduced almost in half in cells overexpressing CRT (77 Ϯ 15 nA versus 143 Ϯ 19 nA in controls, n ϭ 31 matched pairs of cells, one of which is shown in Fig. 4, p Ͻ 0.001). These results, which are consistent with those obtained in mammalian cell lines (26,27), indicate that CRT reduces the rise in [Ca 2ϩ ] i due to Ca 2ϩ influx. Direct measurements of whole cell current due to Ca 2ϩ or Ba 2ϩ entry have indicated that the Ca 2ϩ -gated Cl Ϫ current assay has a high sensitivity for measuring the magnitude of Ca 2ϩ influx (22,43 (44). At base line, the 535/485 fluorescence emission ratios in cells co-expressing 3er and CRT were similar to the ratios found in cells expressing 3er alone  These data indicate that overexpressed CRT does not measurably change the resting [Ca 2ϩ ] ER .
The ability of CRT to maintain [Ca 2ϩ ] ER near resting levels uncouples Ca 2ϩ store depletion from intracellular InsP 3 levels in a novel way; unlike the Ca 2ϩ ATPase inhibitor thapsigargin, which deplete the stores at basal InsP 3 levels, overexpressed CRT prevented store depletion despite high InsP 3 levels. Recent evidence suggests that the role played by the InsP 3 R in Ca 2ϩ influx extends beyond simply lowering [Ca 2ϩ ] ER (25). In this context, our direct measurements of [Ca 2ϩ ] ER serve as a reminder that store depletion is required to stimulate Ca 2ϩ influx and that elevated InsP 3 , by itself, is not sufficient.
Effect of CRT on IICR-The simplest mechanism by which CRT could prevent the InsP 3 -induced decrease in [Ca 2ϩ ] ER would be to inhibit IICR. However, our data suggested that this was not the case with saturating concentrations of InsP 3 ; in the cells where CRT had decreased the Ca 2ϩ influx-related Cl Ϫ current, there was no change in the initial Cl Ϫ currents due to IICR. Yet, in oocytes, CRT was previously reported to inhibit the Ca 2ϩ waves triggered by sub-maximal concentrations of Ins(1,4,5)P 3 (53). To test the effect of CRT on the spatial as-pects of IICR under experimental conditions where it could reduce Ca 2ϩ influx, we microinjected sub-maximal InsP 3 concentrations into oocytes loaded with fluorescent Ca 2ϩ indicators and visualized the resulting Ca 2ϩ waves with a confocal microscope (46,54). As shown in Fig. 6A, overexpression of CRT did not affect the amplitude, the velocity, or the frequency of the repetitive Ca 2ϩ waves. To determine whether an increase in CRT levels changes the rate at which Ca 2ϩ is released from intracellular stores, we examined the fluorescence intensity along a narrow linear path running perpendicular to the wave front; by multiplying the average slope of the increasing fluorescence values at the wave front (ѨF/Ѩ distance) by the instantaneous speed of the wave (Ѩ distance/Ѩ time), we obtained the rate at which [Ca 2ϩ ] rises at the wave front (ѨF/Ѩ time) (32). The rate of rise in fluorescence intensity over base line at the wave front was 38 Ϯ 4% s Ϫ1 for control cells and 36 Ϯ 2% s Ϫ1 for CRT-expressing cells (19 wave pairs were randomly chosen from 8 CRT/control cell pairs) suggesting a similar rate of Ca 2ϩ release at the wave front (Fig. 6A). Although our results that CRT expression did not change the amplitude of the Ca 2ϩ waves agree with those of Camacho and Lechleiter (53), we did not find that CRT decreases Ca 2ϩ waves frequency. Many reasons could explain this difference. 1) We studied our cells 2-3 days after DNA injection, whereas Camacho's group studied theirs after 5-7 days. Although we studied our cells at a time when CRT was already expressed maximally and could reduce Ca 2ϩ influx, we may have missed a time-dependent effect of CRT on the Ca 2ϩ waves. 2) We studied only those cells that had proven exogenous protein expression and that had preserved plasma membrane electrical resistance. Camacho and Lechleiter (53) did not screen individual cells for protein overexpression or for plasma membrane electrical integrity. Thus, the reported decrease in the frequency of Ca 2ϩ waves could not be unequivocally ascribed to an increase in CRT levels and may have been restricted to cells that were electrically leaky. 3) Most of the experimental data reported by Camacho and Lechleiter (53) were obtained in cells co-expressing the Ca 2ϩ -ATPase SERCA2b. In a later publication, the same group (55) reported that CRT-⌬C, which had inhibited the Ca 2ϩ waves in cells co-expressing SERCA2b, had no effect on the Ca 2ϩ waves in cells co-expressing SERCAs lacking an ERlumenal glycosylation site. Thus, their results may mainly reflect an interplay between CRT and SERCA2b (55). Our results suggest that under conditions where an increased amount of CRT within the ER can reduce the InsP 3 -induced decline in [Ca 2ϩ ] ER , CRT does not materially affect the mechanisms that initiate and propagate Ca 2ϩ waves. Thus, CRT does not appear to diminish the InsP 3 -induced decline in [Ca 2ϩ ] ER by inhibiting IICR.
Effect of CRT on Ca 2ϩ Uptake and/or Ca 2ϩ Extrusion-CRT could blunt the InsP 3 -induced decrease in [Ca 2ϩ ] ER if it accelerated the reuptake of cytosolic Ca 2ϩ into the ER lumen. To investigate this alternative hypothesis, we first estimated the rate at which [Ca 2ϩ ] i returns to base line following the passage of a Ca 2ϩ wave front using an analysis similar to that described in the preceding paragraph. We found that the rate of [Ca 2ϩ ] i decline was not accelerated in cells expressing CRT (Fig. 6B). Next, we scanned the site of an intracellular injection of CaCl 2 (1 nl, 50 mM) with a single laser line (scanning rate ϭ 500 Hz) in oocytes loaded with fluo 3. Fluorescence due to this exogenous addition of Ca 2ϩ returned to base line at a slower rate in oocytes expressing CRT (Fig. 6B). To assay more specifically Ca 2ϩ uptake into the intracellular stores, we compared the ATP-dependent 45 Ca 2ϩ uptake into microsomes isolated either from cells overexpressing CRT or from control cells. At 3 min, the time that the microsomes reach 50% of their total ATP-dependent 45 Ca 2ϩ uptake (33), we found a trend toward lower 45 Ca 2ϩ accumulation in microsomes from cells overexpressing CRT (Fig. 6B, p Ͻ 0.06). Together with the Ca 2ϩ injections The five parameters measured were (bars from left to right) as follows: 1) basal [Ca 2ϩ ] before the passage of a Ca 2ϩ wave; 2) peak [Ca 2ϩ ] at the wave front; 3) instantaneous wave velocity; 4) frequency of the repetitive waves; 5) rate of Ca 2ϩ release at the wave front. B, effect of CRT overexpression on Ca 2ϩ reuptake/extrusion. The three parameters measured were the rate at which peak [Ca 2ϩ ] i decreases following either the passage of a Ca 2ϩ wave (1st bar) or an intracellular injection of CaCl 2 (2nd bar) and the ATP-dependent 45  experiments, these last results suggest that CRT overexpression slows the rate of Ca 2ϩ uptake into filled stores. In contrast, our data with the Ca 2ϩ waves suggest that CRT does not measurably affect the rate of Ca 2ϩ reuptake into InsP 3 -depleted stores. Our results are thus compatible with a previous report of CRT inhibiting the ER Ca 2ϩ ATPase (55) but further suggest that this inhibition can be overcome when [Ca 2ϩ ] ER decreases. Overall, these data do not support our alternative hypothesis that CRT reduces the InsP 3 -induced decline in [Ca 2ϩ ] ER by accelerating Ca 2ϩ reuptake/extrusion.
CRT Increases Cellular and Microsomal Ca 2ϩ -binding Capacity-If CRT increases neither the magnitude of IICR nor the rate at which Ca 2ϩ is retrieved back into the ER, then CRT could attenuate the InsP 3 -induced decrease in [Ca 2ϩ ] ER simply by increasing the Ca 2ϩ -buffering capacity of the store; with more Ca 2ϩ to start with, more Ca 2ϩ should remain in the stores following InsP 3 stimulation. To measure the impact of CRT overexpression on cellular Ca 2ϩ content, we incubated oocytes with 45 Ca 2ϩ for 72 h, a time sufficient for the isotope to equilibrate across the Ca 2ϩ -binding compartments of the cell; 45 Ca 2ϩ -related counts were 52 Ϯ 14% higher in cells expressing CRT than they were in control cells (84 cell pairs from five different frogs, p Ͻ 0.02). Microsomes purified from groups of 10 CRT-expressing oocytes had 45 Ca 2ϩ -related counts that were 59.3 Ϯ 10% greater than controls (n ϭ 4, p Ͻ 0.007). These data indicate that CRT overexpression increases whole cell and microsomal Ca 2ϩ content and thus support the notion that CRT acts as a high capacity Ca 2ϩ buffer in vivo. Because 80% of the ionomycin-releasable Ca 2ϩ in these microsomes is InsP 3 -sensitive (33), these data further suggest that CRT levels can influence the capacity of the InsP 3 -sensitive Ca 2ϩ pools of the oocyte.
High Capacity Ca 2ϩ -binding Domain of CRT Is Required to Prevent InsP 3 -induced Store Depletion and to Reduce IICI-To probe the relationship between Ca 2ϩ -buffering properties of CRT and its ability to curb the InsP 3 -induced decrease in [Ca 2ϩ ] ER , we deleted the C-terminal domain, or C-domain (CRT-⌬C), which is responsible for the high Ca 2ϩ -binding capacity of CRT (13). To serve as a control, and to investigate the alternative possibility that CRT could regulate the InsP 3 -induced Ca 2ϩ signal by sensing a decrease in [Ca 2ϩ ] ER (53), we also made a mutant lacking the P-domain (CRT-⌬P), which contains the only high affinity Ca 2ϩ -binding site of CRT (13). Successful expression of both deletion mutants was confirmed by Western blotting (Fig. 1, 3rd and 4th lanes).
At isotopic equilibrium, whole oocytes or microsomes extracted from oocytes expressing CRT-⌬C had 45 Ca 2ϩ -related counts that were no different from controls (respectively 103 Ϯ 8% (25 cell pairs) and 91 Ϯ 6% (n ϭ 3) of controls (100%)). In contrast, oocytes or microsomes from oocytes expressing CRT-⌬P had greater 45 Ca 2ϩ counts than did controls (respectively, 148 Ϯ 13% (25 cell pairs, p Ͻ 0.004) and 174 Ϯ 14% (n ϭ 3, p Ͻ 0.01) of controls (100%)) (Fig. 7A). Thus, CRT-⌬P increased cellular and microsomal Ca 2ϩ -binding capacity to an extent similar to that observed with wild-type CRT. These results confirm that it is the C-domain that confers on CRT the ability to bind large quantities of Ca 2ϩ . Because this additional Ca 2ϩ binds to the C-domain with low affinity (K d ϳ2 mM), it should therefore be readily available for release into the cytosol upon opening of the InsP 3 R. Yet, CRT overexpression did not affect the elevations in [Ca 2ϩ ] i due to IICR. Thus, additional releasable Ca 2ϩ appears to have little impact on the magnitude of IICR in cells that already own substantial intracellular Ca 2ϩ reserves.
The InsP 3 -induced decline in the 535/485 emission ratio of 3er was smaller in cells co-expressing CRT-⌬P than it was either in cells co-expressing CRT-⌬C (2.6 Ϯ 0.9% versus 8.3 Ϯ 1.4% decline, respectively, n ϭ 11 pairs, p Ͻ 0.002) or in control cells (7.9 Ϯ 0.8%, n ϭ 11) (Fig. 7B). As shown in Fig. 7C, CRT-⌬P reduced the Cl Ϫ currents due to Ca 2ϩ influx, whereas CRT-⌬C did not; neither protein changed the magnitude of the initial Cl Ϫ current due to intracellular Ca 2ϩ release. Taken together, these data indicate that CRT relies on the C-domain both to blunt the InsP 3 -induced decline in [Ca 2ϩ ] ER and to reduce the [Ca 2ϩ ] i elevations due to IICI.
CRT Does Not Prevent Thapsigargin-induced Store Depletion from Fully Activating Ca 2ϩ Influx-If CRT acts primarily as a Ca 2ϩ buffer to reduce Ca 2ϩ influx, then a prolonged depletion of the Ca 2ϩ stores should eventually lead to a full activation of Ca 2ϩ influx. To test this hypothesis, we incubated oocytes with the Ca 2ϩ -ATPase inhibitor thapsigargin for 3 h (42), and we directly measured the resulting whole cell Ca 2ϩ influx current according to the protocol of Yao and Tsien (43). As the tracings in Fig. 8A exemplify, the magnitude of the thapsigargin-induced Ca 2ϩ influx current in cells overexpressing CRT was similar to that of controls (aggregate data for 7 cell pairs shown in Fig. 8B). Experiments with 3er-expressing oocytes confirmed that thapsigargin exposure decreases [Ca 2ϩ ] ER to a similar extent in control and in CRT-expressing cells (Fig. 8C). By demonstrating that store depletion can still fully stimulate Ca 2ϩ influx in cells overexpressing CRT, these data further support the hypothesis that CRT controls InsP 3 -induced Ca 2ϩ influx by buffering [Ca 2ϩ ] ER .
In summary, our results establish that an increase in calreticulin levels can reduce the InsP 3 -induced decline in [Ca 2ϩ ] ER , thereby confirming the proposal originally put forth by Pozzan and co-workers (26,28). The experimental results also link the high Ca 2ϩ -buffering capacity of CRT to its ability both to prevent Ca 2ϩ store depletion and to reduce Ca 2ϩ influx. Taken together, the results directly support the notion that changing levels of CRT can alter [Ca 2ϩ ] ER within ranges that are relevant to the cellular mechanisms that control InsP 3 -induced Ca 2ϩ influx.