A Retrograde Signal from Calsequestrin for the Regulation of Store-operated Ca2+ Entry in Skeletal Muscle*

Calsequestrin (CSQ) is a high capacity Ca2+-binding protein present in the lumen of sarcoplasmic reticulum (SR) in striated muscle cells and has been shown to regulate the ryanodine receptor Ca2+ release channel activity through interaction with other proteins present in the SR. Here we show that overexpression of wild-type CSQ or a CSQ mutant lacking the junction binding region (amino acids 86–191; Δjunc-CSQ) in mouse skeletal C2C12 myotube enhanced caffeine- and voltage-induced Ca2+ release by increasing the Ca2+ load in SR, whereas overexpression of a mutant CSQ lacking a Ca2+ binding, aspartate-rich domain (amino acids 352–367; Δasp-CSQ) showed the opposite effects. Depletion of SR Ca2+ by thapsigargin initiated store-operated Ca2+ entry (SOCE) in C2C12 myotubes. A large component of SOCE was inhibited by overexpression of wild-type CSQ or Δjunc-CSQ, whereas myotubes transfected with Δasp-CSQ exhibited normal function of SOCE. These results indicate that the aspartate-rich segment of CSQ, under conditions of overexpression, can sustain structural interactions that interfere with the SOCE mechanism. Such retrograde activation mechanisms are possibly taking place at the junctional site of the SR.

Calsequestrin (CSQ) 1 is a sarcoplasmic reticulum (SR) resident protein in muscle cells whose primary known function is to buffer Ca 2ϩ in the lumen of SR. It binds Ca 2ϩ with high capacity (40 -50 Ca 2ϩ /CSQ) and moderate affinity (K d ϳ1 mM) (1). Recent studies have shown, however, that CSQ participates in the active Ca 2ϩ release process from SR not simply by being a passive Ca 2ϩ storage protein but also by actively modulating the function of the ryanodine receptor (RyR), the primary SR Ca 2ϩ release channel involved in excitation-contraction coupling (2)(3)(4)(5)(6). The carboxyl terminus of CSQ contains an aspartate-rich region (amino acids 354 -367) (7,8), which functions as a major Ca 2ϩ binding motif (9) and also interacts with triadin or junctin, proteins of the SR membrane complexed to RyR with unclear roles in the operation of excitation-contraction coupling. A different region of CSQ (amino acids 86 -191) has been suggested previously to bind to junctin and triadin (6,17). The functional significance of these CSQ regions in muscle Ca 2ϩ signaling has not been examined.
The internal Ca 2ϩ store of muscle cells, located in the SR, has a limited capacity; it must be replenished regularly through the entry of Ca 2ϩ from the external environment. Depletion of SR Ca 2ϩ stores, following activation of RyR or other Ca 2ϩ release mechanisms, triggers Ca 2ϩ entry from the external environment through a process known as capacitative Ca 2ϩ entry via activation of store-operated Ca 2ϩ channels (SOC) located in the cell surface membrane (10,13,14). Research into the molecular and cellular function of store-operated Ca 2ϩ entry (SOCE) has been carried out primarily in non-excitable cells (i.e. lymphocytes, mast cells, etc.) and to some extent in smooth muscle cells (11,12). Recently, Kurebayashi and Ogawa (13) presented the first functional evidence for the existence of SOC in skeletal muscle. We have extended their observations and shown that activation of SOC in skeletal muscle is coupled to retrograde signaling via conformational changes in the RyR (14).
In this study, we test the hypothesis that the RyR might receive information on the state of SR Ca 2ϩ depletion via a direct retrograde signal from CSQ and thereby modulate both RyR-mediated Ca 2ϩ release and RyR-mediated SOCE. We found that overexpression of CSQ not only enhances active Ca 2ϩ release through the RyR but also suppresses SOCE. Deletion of the Ca 2ϩ binding, aspartate-rich region of CSQ in these overexpression experiments resulted in reversal of the suppression of SOCE by wt-CSQ. Our data suggest that modulation of the RyR complex by CSQ from the luminal side of the SR could play a major role in regulating Ca 2ϩ homeostasis in muscle cells and begin to define regions of CSQ that differentially interact with the RyR complex.

EXPERIMENTAL PROCEDURES
Cell Culture-C2C12 myoblasts derived from mouse skeletal muscle were grown in DMEM supplemented with 10% fetal bovine serum and 1% penicillin/streptomycin, as described by Shin et al. (15). Differentiation of myoblasts into myotubes was induced by changing the culture medium to DMEM supplemented with 2% horse serum (HS) and 1% penicillin/streptomycin. Experiments were performed on C2C12 myotubes expressing RyR, i.e. from the fifth day of culture in HS-DMEM, when it was possible to select myotubes having mature skeletal-type excitation-contraction coupling.
Cloning and Gene Transfection-The wt-CSQ cDNA from rabbit skeletal muscle and two deletion mutants of CSQ, ⌬junc-CSQ and ⌬asp-CSQ, were originally cloned into the pCDNA-HA 3.1 vector. For functional studies with C2C12 cells, the CSQ cDNAs were subcloned from pcDNA3.1-HA to pCMS-EGFP to create pCMS-EGFP(wt-CSQ), pCMS-EGFP(⌬asp-CSQ), and pCMS-EGFP(⌬junc-CSQ). The pCMS-EGFP plasmid contains two separate promoters that drive the transcription of green fluorescent protein (GFP, under the SV40 promoter) and the gene of interest (i.e. wt-CSQ or its mutants, under the CMV promoter) (16), thereby providing a convenient way of selecting transfected cells using GFP fluorescence. pCMS-EGFP vector alone or vector containing wt-CSQ, ⌬asp-CSQ, or ⌬junc-CSQ cDNAs were transfected into proliferating myoblasts using LipofectAMINE plus TM reagent according to the manufacturer's instructions. The culture medium was changed to HS-DMEM to allow differentiation of myoblasts into myotubes 12 h after transfection.
Immunocytochemistry-Five days after culturing in HS-DMEM medium, the C2C12 myotubes growing on coverslips and transfected with pcDNA3.1-HA plasmids containing wt-CSQ, ⌬asp-CSQ, or ⌬junc-CSQ were fixed with 4% paraformaldehyde and permeabilized with 0.5% Triton X-100 in phosphate-buffered saline (137 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , 2 mM KH 2 PO 4 , pH 7.2) for 5 min. The cells were then incubated for 30 min with primary monoclonal antibody against HA or polyclonal antibody against skeletal CSQ for detecting exogenous HA-CSQ fusion proteins or endogenous CSQ protein. The cells were washed four times with 0.1% Triton X-100, followed by incubation with rhodamine-conjugated secondary antibody for 30 min in phosphate-buffered saline containing 1% bovine serum albumin. For detection of endogenous RyR, the cells were incubated for 30 min with primary polyclonal antibody against RyR and treated with fluorescein-conjugated secondary antibody. The coverslips were then mounted with 90% glycerol and 0.1% O-phenylenediamine in phosphate-buffered saline. Immunofluorescence was analyzed under a Leica DMRBE microscope (Heidelberg, Germany) equipped with a ϫ100 objective and filters for epifluorescence. Wild-type and CSQ mutant protein expression was demonstrated by Western blot following transient transfection in Chinese hamster ovary (CHO) cells, rather than C2C12 cells, because of the low efficiency of transfection in the latter (see Fig. 1b). The expressed CSQ protein was probed with polyclonal anti-CSQ antibody. The protein-antibody complex was blotted with a horseradish peroxidaselinked secondary antibody, and the signal was detected on Eastman Kodak Co. films using a chemiluminescent kit (Pierce, Rockford, IL).
Single Cell Ca 2ϩ Measurement-The detailed procedure has been described elsewhere (15). Briefly, C2C12 myotubes were loaded with Fura-2/AM fluorescent Ca 2ϩ indicator. Individual myotubes expressing exogenous CSQ were selected by the presence of GFP fluorescence, as described above. The changes in intracellular Ca 2ϩ in single live cells was measured following exposure to 10 mM caffeine or 1 M thapsigargin (Tg), with no [Ca 2ϩ ] o present in the bath solution (Ca 2ϩ -free balanced salt solution containing 140 mM NaCl, 2.8 mM KCl, 2 mM MgCl 2 , 10 mM HEPES, pH 7.2, 0.5 mM EGTA).
Mn 2ϩ Quenching Assay of Store-operated Ca 2ϩ Entry-The detailed procedure has been described elsewhere (14). Briefly, to measure Mn 2ϩ influx rate through SOC, 0.5 mM Mn 2ϩ was added to the extracellular medium after Tg-induced SR Ca 2ϩ depletion with or without the buffering of cytosolic Ca 2ϩ by 50 M BAPTA-AM. The Mn 2ϩ quenching of Fura-2 fluorescence was measured at the Ca 2ϩ -independent wavelength of Fura-2 excitation (360 nm). The decay of Fura-2 fluorescence upon Mn 2ϩ addition was expressed as percent decrease in Fura-2 fluorescence per unit time (initial fluorescence ϭ 100%).
Statistical Analysis-Values are means Ϯ S.E. Significance was determined by Student's t test or analysis of variance. A value of p Ͻ 0.05 was used as criterion for statistical significance.

Localization of Exogenous Wild-type and Mutant CSQ in SR
of C2C12 Cells-CSQ contains a putative junctin-binding region (amino acids 86 -191; junc), as well as the Ca 2ϩ -binding aspartate-rich region (amino acids 354 -367; asp) (7,17). To examine the function of junc and asp regions of CSQ, two deletion mutants, ⌬junc-CSQ and ⌬asp-CSQ, were generated using the PCR-based method for expression and functional studies in C2C12 cells (9). To distinguish the subcellular distribution of endogenous CSQ from expressed exogenous CSQ, wt and the CSQ mutants were expressed as HA-CSQ fusion proteins in differentiated C2C12 myotubes. Subcellular localization of HA-tagged proteins was performed by immunostaining with monoclonal antibody against HA. These experiment revealed a perinuclear distribution of HA-⌬asp-CSQ expressed in C2C12 myotubes (Fig. 1b), in a pattern that is indistinguishable from that of endogenous CSQ present in the SR detected by polyclonal anti-CSQ (Fig. 1a). The subcellular distributions of HA-wt-CSQ and HA-⌬junc-CSQ were similar to that of endogenous CSQ, indicating that both exogenously expressed proteins were also localized to the SR (data not shown). This was further confirmed by co-localization studies with polyclonal anti-RyR, as shown in the lower panels of Fig.  1, a and b. Clearly, the patterns of RyR distribution are virtually identical to those of wt and mutant CSQ expressed in C2C12 cells.
To confirm that the translational products of the various CSQ cDNAs were indeed CSQ, Western blots were performed on SDS-PAGE separated proteins derived from CHO cells transiently transfected with the wt-CSQ, ⌬asp-CSQ, and ⌬junc- CSQ cDNAs. With LipofectAMINE-mediated gene transfection, CHO cells have higher transfection efficiency than C2C12 cells (30 -60% for CHO versus 3-6% for C2C12), making it easier to detect expressed CSQ proteins. As shown in Fig. 1c, proteins of the predicted molecular masses are identified by anti-CSQ antibody. As with the C2C12 cells, immunostaining studies of these CHO cells also indicated that the expressed CSQ proteins were localized in the ER (not shown).
Differential Effects of wt-CSQ and ⌬asp-CSQ on Intracellular Ca 2ϩ Release in Skeletal Muscle-Insertion of the various CSQ cDNAs into another eukaryotic expression vector, pCMS-EGFP, enabled selection of significantly transfected C2C12 cells using GFP fluorescence. The pCMS-EGFP plasmid expresses 1:1 ratio of GFP and CSQ under the control of two independent promoters (15,16). Individual C2C12 myotubes exhibiting similar levels of GFP fluorescence, and therefore most likely similar level of exogenous CSQ proteins, were selected for functional studies with caffeine-induced Ca 2ϩ release measurements. As shown in Fig. 2a, application of 10 mM caffeine resulted in Ca 2ϩ release from SR in myotubes transfected with GFP alone (control). The peak amplitude of caffeine-induced Ca 2ϩ release in myotubes transfected with wt-CSQ was ϳ1.7-fold higher than cells transfected with GFP alone (⌬F 340 /F 380 ϭ 0.80 Ϯ 0.02, n ϭ 13, GFP; 1.38 Ϯ 0.03, n ϭ 11, wt-CSQ) (Fig. 2c). In contrast, expression of ⌬asp-CSQ in C2C12 cells significantly reduced caffeine-induced Ca 2ϩ release (0.41 Ϯ 0.03, n ϭ 14). Myotubes transfected with ⌬junc-CSQ, on the other hand, showed similar enhancement of the amplitude of the caffeine-induced Ca 2ϩ release transient (1.23 Ϯ 0.03, n ϭ 14) as wt-CSQ (Fig. 2c).
A simple explanation for the enhancement of caffeine-induced Ca 2ϩ release in C2C12 myotubes overexpressing wt-CSQ is that this phenomenon likely reflects release from a concomitantly increased SR Ca 2ϩ store. To determine whether the SR Ca 2ϩ store was indeed increased, we treated cells with A23187, a Ca 2ϩ ionophore that will release the entire intracellular Ca 2ϩ store and allow its quantitation (18). As shown in Fig. 2e, the A23187-releasable Ca 2ϩ pool was significantly larger in myotubes transfected with wt-CSQ and ⌬junc-CSQ than those transfected with GFP alone (0.87 Ϯ 0.12, n ϭ 6, GFP; 1.43 Ϯ 0.13, n ϭ 8, wt-CSQ; 1.35 Ϯ 0.11, n ϭ 6, ⌬junc-CSQ), whereas cells transfected with ⌬asp-CSQ contained an A23187-releasable Ca 2ϩ pool that was statistically identical to GFP controls (0.83 Ϯ 0.15, n ϭ 6) (Fig. 2e).
We then endeavored to determine whether the effects of transfected CSQ proteins on depolarization-induced Ca 2ϩ release would parallel the results seen with caffeine-induced Ca 2ϩ release. Changing the extracellular KCl concentration from 2.8 to 10 mM led to depolarization of the cell surface membrane and induced the release of Ca 2ϩ from the SR in C2C12 cells. As shown in Fig. 2, b and d, the peak amplitude of depolarization-induced Ca 2ϩ release in cells overexpressing wt-CSQ and ⌬junc-CSQ was again ϳ1.4 -1.6-fold higher than that of GFP controls (0.74 Ϯ 0.01, n ϭ 10, GFP; 1.20 Ϯ 0.03, n ϭ 8, wt-CSQ; 1.02 Ϯ 0.01, n ϭ 10, ⌬junc-CSQ), whereas overexpression of ⌬asp-CSQ led to a significantly decreased depolarization-induced Ca 2ϩ release (0.46 Ϯ 0.03, n ϭ 11). These results parallel the aggregate caffeine-induced Ca 2ϩ release data shown in Fig. 2c. These data suggest that ⌬asp-CSQ may either directly suppress RyR channel activity or reduce the efficiency of signal transduction from the dihydropyridine receptor to the RyR. Theoretically, ⌬asp-CSQ could suppress SR Ca 2ϩ release by reducing the Ca 2ϩ buffering capacity of the SR. The aggregate Ca 2ϩ store data presented in Fig. 2e, however, demonstrate that the Ca 2ϩ store of the GFP control cells and the ⌬asp-CSQ cells are equivalent.
Overexpression of functional CSQ, wt or mutant, in C2C12 cells should inevitably increase the Ca 2ϩ buffering capacity of the SR and thereby alter the duration of passive Ca 2ϩ movement across the SR membrane through as yet undescribed leak pathways. The kinetics of decay of this passive myoplasmic Ca 2ϩ signal reflects a competition between continuing Ca 2ϩ leak from the SR store and the removal of myoplasmic Ca 2ϩ to the external environment by various plasma membrane-based mechanisms. One would expect that cells containing an ele-vated SR Ca 2ϩ store would have a longer kinetic decay of the myoplasmic Ca 2ϩ signal. To test this possibility, Tg, a potent inhibitor of SR Ca 2ϩ -ATPase (19), was used to block the Ca 2ϩ uptake function of the SR membrane, allowing for depletion of the luminal Ca 2ϩ store via SR Ca 2ϩ leak pathways. Our results, shown in Fig. 3A, demonstrate that the peak amplitudes of Tg-induced increases in myoplasmic [Ca 2ϩ ] i were comparable among the GFP control and those overexpressing wt-CSQ, ⌬asp-CSQ, and ⌬junc-CSQ. As predicted, however, the decay phase of Ca 2ϩ transients in myotubes overexpressing wt-CSQ or ⌬junc-CSQ, shown in Fig. 2e to contain greater Ca 2ϩ stores than control or ⌬asp-CSQ cells, were significantly longer (t1 ⁄2 ϭ 167 Ϯ 11 s, n ϭ 9, GFP; 332 Ϯ 11 s, n ϭ 10, wt-CSQ; 295 Ϯ 28 s, n ϭ 8, ⌬junc-CSQ) (Fig. 3b). Importantly, the decay pattern of Tg-induced Ca 2ϩ transients in cells overexpressing ⌬asp-CSQ was similar to the GFP control (t1 ⁄2 ϭ 161 Ϯ 8 s, n ϭ 9). These results suggest that removal of the asp-rich region significantly reduces Ca 2ϩ buffering capacity of endogenous CSQ, or to the exclusion of endogenous CSQ, does not participate in Ca 2ϩ buffering of the SR. These results are consistent with our previous finding that the asp-rich region contains a major Ca 2ϩ binding motif (9).
Inhibition of Store-operated Ca 2ϩ Entry in Skeletal Muscle by wt-CSQ-We have shown recently (14) that depletion of SR Ca 2ϩ storage leads to activation of SOCE in skeletal muscle. The activation of SOC in skeletal muscle appears to be coupled to conformational changes of RyR. Because our data above Changes of intracellular Ca 2ϩ in myotubes transfected with GFP alone (control), wt-CSQ, ⌬asp-CSQ, or ⌬junc-CSQ were monitored. Each trace is a representative of eight to ten independent experiments. b, the t1 ⁄2 of Tg-induced Ca 2ϩ efflux from SR are summarized as means Ϯ S.E. c, the relative SOCE values after Tg-induced SR Ca 2ϩ depletion are shown as means Ϯ S.E.
suggest that the Ca 2ϩ store, as determined by the functional Ca 2ϩ binding capacity of CSQ, determines the degree of Ca 2ϩ release via the RyR, we asked whether this Ca 2ϩ store could affect SOCE in this system. Sustained treatment of C2C12 myotubes with 1 M Tg in Ca 2ϩ -free medium resulted in complete depletion of SR Ca 2ϩ . Addition of 2 mM Ca 2ϩ to the bath solution after the myoplasmic Ca 2ϩ signal had returned to baseline triggered SOCE in these Ca 2ϩ -depleted cells (Fig. 3a). The degree of SOCE in myotubes transfected with ⌬asp-CSQ was similar to GFP control cells (Fig. 3, a and c). Strikingly, overexpression of wt-CSQ and ⌬junc-CSQ in the presence of a Tg-depleted SR Ca 2ϩ store resulted in significant inhibition of SOCE (1.03 Ϯ 0.07, n ϭ 9, GFP; 0.51 Ϯ 0.04, n ϭ 10, wt-CSQ; 0.56 Ϯ 0.06, n ϭ 8, ⌬junc-CSQ) (Fig. 3c).
The total SOCE measured in these experiments is likely the result of a summation of competing processes, SR Ca 2ϩ uptake and release and surface membrane Ca 2ϩ extrusion and influx. To isolate the measurement of SOC-mediated Ca 2ϩ influx, we used the technique of Mn 2ϩ quenching of the Fura-2 fluorescence (14). Mn 2ϩ is known to be able to permeate into cells via SOC but is impervious to surface membrane extrusion processes or SR uptake by Ca 2ϩ pumps. Hence, Mn 2ϩ fluorescence quenching represents a measurement of unidirectional Ca 2ϩ flux into cells via SOC. Under resting conditions (i.e. cells with an intact SR Ca 2ϩ pool), no detectable Mn 2ϩ quenching of Fura-2 was observed (not shown). Myotubes with Tg-depleted SR Ca 2ϩ stores in a Ca 2ϩ -free medium exhibited rapid quenching of Fura-2 fluorescence upon addition of 0.5 mM Mn 2ϩ to the bath solution (Fig. 4a). Surprisingly, cells overexpressing wt-CSQ and ⌬junc-CSQ displayed significant reduction in the rate of Fura-2 fluorescence quenching even with a depleted SR Ca 2ϩ store. On average, ϳ10-fold reduction in Mn 2ϩ influx rate was observed in cells overexpressing wt-CSQ and ⌬junc-CSQ compared with control. Consistent with the results shown in Fig. 3, overexpression of ⌬asp-CSQ did not appear to affect the rate of Mn 2ϩ influx in C2C12 cells (Fig. 4a). If the presence of exogenous CSQ is merely to increase the Ca 2ϩ load of the SR, then the complete depletion of this load should give equivalent activation of SOCE and resultant Mn 2ϩ fluorescence quenching in all four of the cell preparations. Our results imply (a) that CSQ itself initiates a signal to SOCs, and (b) that the asp-rich region of the protein is likely involved in this signal transmission process.
Lack of Effect of BAPTA on SOCE in Skeletal Muscle-Studies from other investigators (33)  the SR membrane, as indicated by the complete lack of Tginduced changes in Fura-2 signal (Fig. 4b). Fifteen min after the addition of Tg, changing the bath solution from no [Ca 2ϩ ] to 2 mM [Ca 2ϩ ] resulted in measurable increases in the Fura-2 signal, indicating significant Ca 2ϩ influx across the cell surface membrane (Fig. 4b). Myotubes transfected with wt-CSQ and ⌬junc-CSQ showed slower Ca 2ϩ entry than GFP control and ⌬asp-CSQ-transfected myotubes. Direct Mn 2ϩ quenching studies of Tg-induced Ca 2ϩ -depleted and BAPTA-buffered myotubes confirmed that the changes in Fura-2 fluorescence above were because of activation of SOCE (Fig. 4c). Here, a striking reduction in the rate of Mn 2ϩ quenching was observed in myotubes overexpressing wt-CSQ and ⌬junc-CSQ, but not ⌬asp-CSQ, with 50 M BAPTA-AM present in the cytosol. These results indicate that buffering of [Ca 2ϩ ] i does not interfere with function of SOC in skeletal muscle and that the wt-CSQ-mediated inhibition of SOCE in C2C12 cells is unlikely to correlate with any changes in myoplasmic [Ca 2ϩ ] i .

DISCUSSION
Until recently, CSQ has been thought of as the SR Ca 2ϩbinding protein whose function is simply to sequester Ca 2ϩ in the vicinity of the RyR/Ca 2ϩ release channel, to maintain a store for this ion, and to facilitate its rapid release during excitation-contraction coupling in muscle cells (2)(3)(4)(5)(6). We have shown here that overexpression of wt-CSQ enhances both caffeine-and voltage-induced Ca 2ϩ release in skeletal muscle myotubes that are associated with an increased Ca 2ϩ store in the SR. A profound reduction of SOCE was observed in cells overexpressing wt-CSQ or ⌬junc-CSQ, but not ⌬asp-CSQ, in cells with depleted SR Ca 2ϩ stores and whose myoplasmic Ca 2ϩ concentrations were buffered with BAPTA. Thus, the SR Ca 2ϩ store is necessary for RyR-dependent Ca 2ϩ release, but Ca 2ϩ store per se is not the sole signal that regulates SOCE. Rather, CSQ adds a proximal signal in the regulation of SOCE in muscle cells. Our data suggests that the asp-rich region of CSQ is essential for retrograde signaling in both RyR-mediated Ca 2ϩ release and regulation of SOCE in skeletal muscle.
The enhancement of caffeine-induced Ca 2ϩ release by wt-CSQ in C2C12 myotubes is similar to that seen in cardiomyocytes isolated from transgenic mice overexpressing CSQ (20,21). In those studies, caffeine-induced Ca 2ϩ release was increased by ϳ10-fold in the CSQ transgenic mouse, paralleling the ϳ10-fold overexpression of CSQ in the heart. Because ⌬asp-CSQ did not change the total SR Ca 2ϩ store, the negative effect of this mutant on caffeine-and voltage-induced Ca 2ϩ release in skeletal muscle may reflect a reduced activity of RyR or its interaction with accessory proteins or reduced local Ca 2ϩ in the vicinity of RyR. Our data suggest that the asp-rich region of CSQ may regulate the proper functioning of the RyR, either by directly interacting with this channel or affecting other partners in the RyR/Ca 2ϩ release channel complex. Others have suggested that proper formation of a quaternary molecular complex among CSQ, triadin, junctin, and RyR plays a critical role in the active Ca 2ϩ release process across the SR membrane (6,20). Indeed, both the carboxyl-terminal-containing asp-rich and amino-terminal regions of CSQ have been suggested as necessary for forming this quaternary SR Ca 2ϩ release complex (8,22,23). Our previous studies have shown that the asp-rich region of CSQ binds Ca 2ϩ , and this region is also involved in interaction with triadin (9). Thus, it is possible that overexpression of ⌬asp-CSQ may alter the conformation of the quaternary complex and therefore cause inhibition of the RyR channel function.
A surprising and critical observation of the present study is that overexpression of wt-CSQ inhibits the function of SOCE in skeletal muscle. The CSQ-mediated inhibition of SOCE ap-pears to involve the asp-rich region of CSQ, because the inhibitory effect was only observed with wt-CSQ and ⌬junc-CSQ but not with ⌬asp-CSQ. Our studies provide the first direct evidence for regulation of SOCE, a cell surface membrane function, through the luminal side of the SR membrane. Previous studies with other cell types have suggested that the physical docking of the ER or SR with the cell surface membrane is involved in the activation of SOC, presumably through contact interaction between SOC and protein components in the ER or SR (e.g. the inositol 1,4,5-trisphosphate receptor or RyR) (10, 24 -26). Alternatively, the release of as yet undefined diffusible second messenger(s) from the intracellular organelle into the cytosol has been proposed to serve as an activator of SOC in response to depletion of intracellular Ca 2ϩ stores (27,28). Our recent studies with primary cultured skeletal muscle cells derived from different genetically engineered mouse models suggest that activation of SOC can be achieved in a graded fashion, depending on the filling state of the intracellular Ca 2ϩ stores and/or the conformational changes of RyR (14). Although the gene(s) responsible for SOC has yet to be identified, and the exact nature of signal transduction involved in the activation of SOC remains largely unknown, our data indicate that the aspartate-rich segment of calsequestrin, under conditions of overexpression, can sustain structural interactions that interfere with the SOCE mechanism. These interactions are possibly taking place at the junctional site of the SR. Previous studies suggested that cardiomyocytes overexpressing CSQ showed abnormal enlarged junctional SR structure in triad junction, resulting in alteration of calcium signaling in muscle cells (20,21). It will be interesting, therefore, to see how the absence of CSQ in a knock-out model would affect the function of SOC in skeletal muscle.
The presence of exogenously expressed CSQ in the SR lumen adds extra Ca 2ϩ buffering capacity and increases the driving force for Ca 2ϩ movement across the SR membrane. Our experiments with Tg-induced SR Ca 2ϩ store depletion and myoplasmic BAPTA Ca 2ϩ buffering have ruled out the possibility that the reduction of SOCE seen with overexpression of wt-CSQ and ⌬junc-CSQ results from an incomplete depletion of SR Ca 2ϩ stores or because of potential changes in myoplasmic [Ca 2ϩ ] i (29,30). A previous study (19) in mouse fibroblast cells showed that overexpression of calreticulin, a major Ca 2ϩ -binding protein in the ER lumen of non-muscle cells, also inhibited SOCE through a mechanism that is independent of its Ca 2ϩ binding properties. Examination of the primary amino acid sequence of calreticulin reveals that, similar to CSQ, it too contains a highly negatively charged region at its carboxyl terminus. We speculate that the conservation of this negatively charged region of the carboxyl terminus of both of these SR/ER Ca 2ϩbinding proteins supports a significant functional role for the protein. We further suggest that this region in calreticulin will be involved in regulating SOCE in non-muscle cells.
Similar to the retrograde interaction between the inositol 1,4,5-trisphosphate receptor and SOC in non-excitable cells (10,11,30), a retrograde RyR-dihydropyridine receptor interaction exists in the skeletal muscle, as revealed by reduced dihydropyridine receptor function in RyR knock-outs (31,32). Cumulative evidence also suggests that the conformational state of the RyR can regulate the function of SOC (10,14). Our data reported here provide additional evidence for a tight link between Ca 2ϩ homeostasis in SR and Ca 2ϩ permeability in the cell surface membrane. Overexpression of CSQ in skeletal muscle not only affects caffeine-and voltage-induced Ca 2ϩ release but also regulates SOCE. The aspartate residues located in the carboxyl terminus of CSQ not only constitute binding pockets for Ca 2ϩ but also can regulate the function of the surface membrane-located, store-operated Ca 2ϩ channel, likely via retrograde interaction with the junctional protein complex in the SR.