A novel Ca2+-induced Ca2+ release mechanism in A7r5 cells regulated by calmodulin-like proteins.

Intracellular Ca2+ release is involved in setting up Ca2+ signals in all eukaryotic cells. Here we report that an increase in free Ca2+ concentration triggered the release of up to 41 +/- 3% of the intracellular Ca2+ stores in permeabilized A7r5 (embryonic rat aorta) cells with an EC50 of 700 nm. This type of Ca2+-induced Ca2+ release (CICR) was neither mediated by inositol 1,4,5-trisphosphate receptors nor by ryanodine receptors, because it was not blocked by heparin, 2-aminoethoxydiphenyl borate, xestospongin C, ruthenium red, or ryanodine. ATP dose-dependently stimulated the CICR mechanism, whereas 10 mm MgCl2 abolished it. CICR was not affected by exogenously added calmodulin (CaM), but CaM1234, a Ca2+-insensitive CaM mutant, strongly inhibited the CICR mechanism. Other proteins of the CaM-like neuronal Ca2+-sensor protein family such as Ca2+-binding protein 1 and neuronal Ca2+ sensor-1 were equally potent for inhibiting the CICR. Removal of endogenous CaM, using a CaM-binding peptide derived from the ryanodine receptor type-1 (amino acids 3614-3643) prevented subsequent activation of the CICR mechanism. A similar CICR mechanism was also found in 16HBE14o-(human bronchial mucosa) cells. We conclude that A7r5 and 16HBE14o-cells express a novel type of CICR mechanism that is silent in normal resting conditions due to inhibition by CaM but becomes activated by a Ca2+-dependent dissociation of CaM. This CICR mechanism, which may be regulated by members of the family of neuronal Ca2+-sensor proteins, may provide an additional route for Ca2+ release that could allow amplification of small Ca2+ signals.

to cell death (1,2). Cells generate Ca 2ϩ signals through both intracellular (mainly the endoplasmic/sarcoplasmic reticulum) and extracellular Ca 2ϩ sources. Regulation of these Ca 2ϩ signals via a variety of Ca 2ϩ channels, expressed either in the plasma membrane or in the membranes of intracellular stores, is thereby essential. Ca 2ϩ fluxes from extracellular and intracellular Ca 2ϩ sources do not occur independently of each other. For example, the intracellular Ca 2ϩ store content regulates Ca 2ϩ entry from the extracellular medium via capacitative Ca 2ϩ entry (3,4), whereas Ca 2ϩ released by one channel can alter the activity of other channels. These are all well documented mechanisms whereby Ca 2ϩ can exert important effects on its own activity. The most important type of regulation is represented by the various mechanisms that may lead to the characteristic bell-shaped dependence of intracellular Ca 2ϩ channels on Ca 2ϩ itself (5)(6)(7)(8)(9). This may in principle be due to direct interaction with Ca 2ϩ or indirectly via Ca 2ϩ -sensor proteins such as calmodulin (CaM). The inositol 1,4,5-trisphosphate receptor (IP 3 R) and the ryanodine receptor (RyR) are the two major families of intracellular Ca 2ϩ release channels that have been characterized. Both types of intracellular channels are regulated in a complex way by Ca 2ϩ and CaM. CaM has been demonstrated to affect the activity of RyRs in both a stimulatory and an inhibitory manner (10,11) but not by the same mechanism for all three RyR isoforms. For the IP 3 R, CaM clearly exerts an inhibitory effect, but the precise mechanism is not yet understood (12,13). RyRs and IP 3 Rs are stimulated by small increases in [Ca 2ϩ ] c and inhibited at higher [Ca 2ϩ ] c (14 -20). Stimulation is important for the mechanism of Ca 2ϩ -induced Ca 2ϩ release (CICR), which allows amplification and regenerative propagation of intracellular Ca 2ϩ signals. CICR seems to be an operational mode of both IP 3 Rs and RyRs, and it is clearly a key feature of intracellular Ca 2ϩ signaling (21). Recent studies have emphasized the role of novel types of intracellular Ca 2ϩ release channels possibly playing an important role in intracellular Ca 2ϩ signaling (22)(23)(24)(25)(26)(27)(28). Wissing et al. (26) identified a novel CICR mechanism in permeabilized hepatocytes that responded to modest increases in [Ca 2ϩ ] c . Polycystin-2, the product of the gene mutated in type-2 autosomal dominant polycystic kidney disease and a prototypical member of a subfamily of the transient receptor potential channel superfamily (TRP), is expressed abundantly in the endoplasmic reticulum (ER) (24). It was shown recently that polycystin-2 expressed in the ER of epithelial cells is a Ca 2ϩ -activated channel that is permeable for divalent cations. Increased levels of intracellular Ca 2ϩ activated polycystin-2-mediated release of Ca 2ϩ from intracellular stores. Recent data also suggested that activation of the ER-associated vanilloid receptor 1 (VR1), a member of the TRP family, by capsaicin binding resulted in Ca 2ϩ mobilization from intracellular stores. This raises the possibility that VR1 may also function as an intracellular Ca 2ϩ release channel (27,28).
In the present study we have identified a novel CICR mechanism in permeabilized A7r5 cells, a permanent cell line derived from embryonic rat aorta. We identified a CICR mechanism that was mediated by neither the IP 3 R nor the RyR. Moreover, we found that this CICR mechanism could be inhibited by CaM 1234 , a Ca 2ϩ -insensitive CaM mutant, and by different members of the superfamily of CaM-like Ca 2ϩ -binding proteins. Our data suggest that the CICR mechanism described here may represent a novel type of release channel, which is silent at low [Ca 2ϩ ] c due to inhibition by bound apoCaM and which becomes activated by the Ca 2ϩ -dependent dissociation of CaM. This CICR mechanism may provide an additional pathway for intracellular Ca 2ϩ release and could play an important role in amplifying Ca 2ϩ signals generated by other Ca 2ϩ release channels.

EXPERIMENTAL PROCEDURES
45 Ca 2ϩ Fluxes-A7r5 cells, which are derived from embryonic rat aorta, were obtained from the American Tissue Type Culture Collection CRL 1444 (Bethesda, MD). Cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum, 3.8 mM Lglutamine, 0.9% (v/v) non-essential amino acids, 85 IU/ml penicillin, 85 g/ml streptomycin, and 20 mM HEPES (pH 7.4). For 16HBE14o-(human bronchial mucosa) and mouse embryonal fibroblast cells a mixture of Dulbecco's modified Eagle's medium and Ham's F-12 medium was used and for LLC-PK 1 cells minimal essential medium ␣ was used. 45 Ca 2ϩ fluxes were performed on saponin-permeabilized cells. The cells were seeded in 12-well clusters (Costar, MA) at a density of ϳ4 ϫ 10 4 cm Ϫ2 . Experiments were carried out on confluent monolayers of cells (3 ϫ 10 5 cells/well) between the 7th and 9th days after plating. Cells were permeabilized by incubating them for 10 min with a solution containing 120 mM KCl, 30 mM imidazole-HCl (pH 6.8), 2 mM MgCl 2 , 1 mM ATP, 1 mM EGTA, and 20 g ml Ϫ1 saponin at 25°C. The nonmitochondrial Ca 2ϩ stores were loaded for 45 min at 37°C in 120 mM KCl, 30 mM imidazole-HCl (pH 6.8), 5 mM MgCl 2 , 5 mM ATP, 0.44 mM EGTA, 10 mM NaN 3 , and 150 nM free 45 Ca 2ϩ (28 Ci ml Ϫ1 ). The cells were then washed twice with 1 ml of efflux medium containing 120 mM KCl, 30 mM imidazole-HCl (pH 6.8), 1 mM EGTA, and 10 M thapsigargin. Thapsigargin was added to block the ER Ca 2ϩ pumps during subsequent additions of Ca 2ϩ . The efflux medium was replaced every 2 min during 18 min, and the efflux was performed at 37°C. The additions of 40 Ca 2ϩ and IP 3 are indicated in the legends of the figures. Free [Ca 2ϩ ] was calculated by the Cabuf program (available at ftp.cc. kuleuven.ac.be/pub/droogmans/cabuf.zip) and based on the stability constants given by Fabiato and Fabiato (29). At the end of the experiment the 45 Ca 2ϩ remaining in the stores was released by incubation with 1 ml of a 2% SDS solution for 30 min. Ca 2ϩ release is plotted as the fractional loss, i.e. the amount of Ca 2ϩ released in 2 min divided by the total store Ca 2ϩ content at that time. The latter value was calculated by summing in retrograde order the amount of tracer remaining in the cells at the end of the efflux and the amounts of tracer collected during the successive time intervals. In experiments performed to exclude 40 Ca 2ϩ / 45 Ca 2ϩ exchange in Fig. 2, cells were loaded during 45 min in loading buffer, containing 4 mM EGTA and 680 M total CaCl 2 , resulting in 285 nM free [Ca 2ϩ ] and a specific activity for Ca 2ϩ of 28 Ci ml Ϫ1 . After 45 min, the loading buffer was replaced for 2 min by a loading buffer with an EGTA concentration of 0.76 mM and supplemented with thapsigargin, to maintain the same 40 Ca 2ϩ / 45 Ca 2ϩ ratio but resulting in an increase in free [Ca 2ϩ ] to 10 M. Efflux was then further performed in Ca 2ϩ -free efflux buffer.
Cloning of sCaBp1 and lCaBp1-Mouse CaBP1 cDNA was cloned from mouse cerebellum RNA. Poly(A) ϩ RNA from mouse cerebellum was prepared using the Micro-FastTrack kit (Invitrogen, CA). Random primed first strand cDNA was synthesized from 1 g of RNA using avian myeloblastosis virus reverse transcriptase. Reverse transcription-PCR was performed with forward primer 5Ј-GCCAGCCATATGG-GCAACTGCGTCAAGTCGCC-3Ј and reverse primer 5Ј-GCGGGCAGC-CTCGAGGCGAGACATCATCCGGAC-3Ј. The forward primer contained the site for NdeI (CATATG), and the reverse primer contained the site for XhoI (CTCGAG). PCR fragments of both isoforms, the short (sCaBp1) and long (lCaBp1) form, were then cloned into the NdeI-XhoI site of the pET21b/ϩ vector (Novagen), yielding an expression vector for a His 6 -tagged sCaBp1 and lCaBp1.
Expression and Purification of Recombinant Proteins-pET-sCaBP1 and pET-lCaBP1 were transformed in BL21 Escherichia coli cells, grown to mid-exponential phase, and induced with 0.75 mM isopropyl-1-thiol-␤-D-galactopyranoside for 4 h at 28°C. Cells were centrifuged for 10 min, and pellets were then resuspended in lysis buffer containing 50 mM NaH 2 PO 4 , pH 7.0, 300 mM NaCl, 10 mM imidazole, 1 mM ␤-mercaptoethanol, 0.8 mM benzamidine, 0.2 mM phenylmethylsulfonyl fluoride, 10 M leupeptin, 1 M pepstatin A, and 75 nM aprotinin. This cell suspension was then lysed by sonication at 20 kHz, nine times for 10 s using a probe sonicator (MSE Ltd., Crawley, Surrey, UK). 1 ml of 50% nickel-nitrilotriacetic acid (Qiagen) slurry was added to 4 ml of cleared lysate and gently mixed by shaking at 4°C for 60 min. The lysatenickel-nitrilotriacetic acid mixture was loaded on a column and washed with 2 volumes of lysis buffer supplemented with 10 mM imidazole. Finally, the recombinant protein was eluted with four times 0.5-ml elution buffer (lysis buffer containing 250 mM imidazole). The protein was eluted in the second and third elution fractions. Recombinant CaM and CaM 1234 were expressed and purified as described in a previous study (31). Recombinant CaM 1 was expressed and purified by phenyl-Sepharose chromatography in the same way as CaM. GST-NCS-1 and GST-NCS-1 E120Q were expressed and purified as described previously (32).

Increase in [Ca 2ϩ ] c Stimulates Ca 2ϩ
Release from Intracellular Stores-In A7r5 cells (embryonic smooth muscle) Ca 2ϩ release from internal stores, mainly from the ER, occurs to a large extent via production of the second messenger IP 3 . In this permeabilized cell system a maximal effective dose of IP 3 can release about 95% of the intracellular Ca 2ϩ content (33). Here, the non-mitochondrial stores of permeabilized A7r5 cells were loaded to steady state with 45 Ca 2ϩ and then incubated in a non-labeled efflux medium containing 10 M thapsigargin. The loss of Ca 2ϩ from the stores under these conditions is plotted as the fractional loss in function of time ( Fig. 1). After 10 min the cells were challenged with 1 M IP 3 (circles), as indicated by the bar. As previously documented in detail, using the same 45 Ca 2ϩ flux technique (17, 34 -37), IP 3 increased the rate of Ca 2ϩ release (Fig. 1). In the same assay, cells challenged with 3 M free 40 Ca 2ϩ (squares) also showed an increase in the rate of Ca 2ϩ release. 3 M free 40 Ca 2ϩ was able to release 25 Ϯ 2% of the stored Ca 2ϩ . The total amount of releasable Ca 2ϩ was measured by treating the cells with 5 M ionophore A23187 (triangles). This activation of Ca 2ϩ release upon elevation of the cytosolic [Ca 2ϩ ] c has also previously been observed by others (15,38), but it could not be excluded that it reflected 45 Ca 2ϩ / 40 Ca 2ϩ exchange without net transport (38).
It is indeed important to emphasize that in this type of experiment the challenge by 3 M 40 Ca 2ϩ could have caused an exchange of 45 Ca 2ϩ for 40 Ca 2ϩ . To exclude the contribution of 45 Ca 2ϩ / 40 Ca 2ϩ exchange we maintained the same 45 Ca 2ϩ / 40 Ca 2ϩ ratio during the loading and efflux phases and we changed the [EGTA] to alter the free [Ca 2ϩ ] (Fig. 2). After incubation in loading buffer during 45 min, the cells were incubated for 2 min in the same loading buffer supplemented with thapsigargin and lowered [EGTA]. This resulted in a complete inhibition of the Ca 2ϩ uptake via the sarcoplasmic/ endoplasmic-reticulum Ca 2ϩ -ATPase (SERCA) and in a rise of the free [Ca 2ϩ ] up to 10 M, while maintaining the 45 Ca 2ϩ / 40 Ca 2ϩ ratio constant. Subsequently, the cells were incubated in Ca 2ϩ -free efflux medium. The traces in Fig. 2 illustrate how the Ca 2ϩ content of the stores decreased during the 10-min incubation in the Ca 2ϩ -free efflux medium and show that the initial Ca 2ϩ content was decreased by the rise in free [Ca 2ϩ ] during the first 2 min subsequent to the loading. Cells incubated during 2 min in 10 M free [Ca 2ϩ ] medium showed a decrease in Ca 2ϩ content of 27 Ϯ 6% compared with cells that were not subjected to a [Ca 2ϩ ] rise. This finding demonstrates that a decrease in the Ca 2ϩ content was induced by 10 M free Ca 2ϩ without a change in the 45 Ca 2ϩ / 40 Ca 2ϩ ratio thereby excluding passive 45 Ca 2ϩ / 40 Ca 2ϩ exchange.
CICR Is Neither IP 3 R-nor RyR-mediated-The two major classes of intracellular Ca 2ϩ release channels are the IP 3 Rs and the RyRs. In A7r5 cells both IP 3 R1 (73%) and IP 3 R3 (26%) are expressed (39). No evidence has been found for a functional role of the RyR in A7r5 cells (33,40). IP 3 R1 and IP 3 R3 are both known to be regulated by increases in [Ca 2ϩ ] c (9, 14 -16). We therefore investigated whether the CICR described here originated from the IP 3 -sensitive stores. Permeabilized cells were loaded with 45 Ca 2ϩ in the presence or absence of a saturating dose of IP 3 (300 M). Efflux was then performed in medium without added Ca 2ϩ . After 10 min cells were incubated for 2 min with 10 M free 40 Ca 2ϩ . No CICR was observed in cells that were loaded in the presence of IP 3 (data not shown). This finding suggested that the CICR mechanism only occurred from the IP 3 -sensitive stores. Furthermore we looked whether this CICR mechanism was also restricted to the thapsigarginsensitive stores. In permeabilized A7r5 cells 92% of the total Ca 2ϩ uptake involved a thapsigargin-sensitive SERCA pump, and 8% was mediated by a thapsigargin-insensitive Ca 2ϩ -uptake mechanism (41). Cells that were loaded in the presence of 10 M thapsigargin were challenged with 10 M free 40 Ca 2ϩ . Also in this condition no CICR mechanism was observed (data not shown). Taken together, these results suggest that this CICR mode is only occurring from the thapsigargin and IP 3sensitive compartments of the ER.
Heparin, 2-aminoethoxydiphenyl borate (2-APB) and xestospongin C (XeC) are the most used antagonists of the IP 3 R. In Fig. 3 it is shown that none of these components affected the fractional loss induced by 10 M free 40 Ca 2ϩ , revealing that the IP 3 R was not involved in this mechanism. Although there is no evidence for a functional RyR in A7r5 cells, we also used antagonists of the RyR to exclude any role of the RyR in this CICR mechanism. Fig. 3 illustrates that neither ruthenium red (Ru-Red) (100 M) nor ryanodine (5 M) had any effect on the fractional loss induced by 10 M free 40 Ca 2ϩ . Ca 2ϩ release stimulated by sphingosine 1-phosphate (42) and NAADP (43) has been observed in a number of cell types. However, it is unlikely that one of these mechanisms mediated CICR in A7r5 cells, because NAADP-stimulated Ca 2ϩ release was not modulated by Ca 2ϩ (44) and no sphingosine 1-phosphate or NAADP-stimulated Ca 2ϩ release was observed in A7r5 cells under our assay conditions (data not shown).
Characteristics of the Observed CICR-To further characterize the CICR mechanism in A7r5 cells, we measured its [Ca 2ϩ ] dependence. The Ca 2ϩ release as a function of increasing free [ 40 Ca 2ϩ ] was plotted in Fig. 4a. A maximally effective free physiological range of cytosolic Ca 2ϩ levels (0.1-10 M). Fig. 4b illustrates that the CICR mechanism was controlled by the level of store loading. Ca 2ϩ stores from permeabilized A7r5 cells loaded to steady state with 45 Ca 2ϩ were incubated in Ca 2ϩ -free efflux medium, and their Ca 2ϩ content was plotted as a function of time. 40 Ca 2ϩ (10 M) was added either after 2 min (circles, full stores) or after 20 min (triangles, less filled stores). Application of 10 M free 40 Ca 2ϩ was clearly less efficient to release 45 Ca 2ϩ from less filled stores. These results indicate that the CICR mechanism was controlled by the luminal [Ca 2ϩ ]. In this respect the CICR mechanism shows the same dependence on the luminal Ca 2ϩ content as described for IP 3 -induced Ca 2ϩ release in those cells (34). Fig. 5a illustrates that Mg 2ϩ dose-dependently blocked the CICR in A7r5 cells. The EC 50 for this inhibition was 0.59 Ϯ 0.04 mM. The inhibitory effect of Mg 2ϩ was not due to the increase in osmolarity of the medium, because a similar increase in osmolarity by addition of 15 mM KCl instead of 10 mM MgCl 2 , did not inhibit the CICR (data not shown).
Furthermore, CICR was stimulated by increasing the [ATP] in the absence of Mg 2ϩ (Fig. 5b). Stimulation occurred with an EC 50 of 320 Ϯ 23 M. By adding 1 mM ATP together with 10 M free 40 Ca 2ϩ , maximal Ca 2ϩ release was increased from 27 Ϯ 4% to 41 Ϯ 3%. This indicates that in physiological conditions this CICR can release a significant fraction of the intracellular stores.
Regulation of CICR by CaM-like Proteins-CaM is a ubiquitous regulator of most if not all types of Ca 2ϩ channels, including the intracellular Ca 2ϩ release channels. We therefore investigated the effect of CaM and CaM mutants on the CICR mechanism in A7r5 cells. CaM 1234 , which is CaM rendered Ca 2ϩ -insensitive by point mutations (45), has the ability to associate with apoCaM-binding sites on Ca 2ϩ release channels (46). In this way CaM 1234 can prevent access to Ca 2ϩ /CaM effector sites, thereby eliminating Ca 2ϩ regulation via CaM as the Ca 2ϩ sensor. Recombinant CaM (10 M) or CaM 1234 (10 M) was added together with 3 M free 40 Ca 2ϩ to permeabilized cells loaded with 45 Ca 2ϩ . Fig. 6a shows that exogenously added  CaM had no effect on the Ca 2ϩ release induced by 3 M free 40 Ca 2ϩ , whereas CaM 1234 almost completely inhibited the CICR. CaM 1234 inhibited the CICR with micromolar affinity (Fig. 6b). These data obtained with CaM and CaM 1234 led us to hypothesize that Ca 2ϩ -free CaM (apoCaM) in resting conditions may be tethered to the protein responsible for the CICR. Binding of Ca 2ϩ to the tethered CaM could then dissociate or dislocate CaM, which could provoke a conformational change thereby activating Ca 2ϩ release from the intracellular stores. CaM 1234 would render the system insensitive to activation by Ca 2ϩ . To know whether CaM needed to be mutated in all four EF-hands to fulfill its inhibitory role on this CICR mechanism, we also tested CaM mutated in only one EF-hand. CaM 1 is mutated in the first EF-hand (Fig. 7a). Recombinant CaM 1 (10 M) was added together with 3 M free 40 Ca 2ϩ to permeabilized cells loaded with 45 Ca 2ϩ (Fig. 6a). CaM 1 was also able to inhibit the CICR, although not to the same extent as CaM 1234 . The EC 50 for CaM 1 inhibition was lower and inhibition was not complete (Fig. 6b). These data suggest that only wild type CaM is capable to fulfill the activation of the CICR by sensing the increase in free Ca 2ϩ , whereas mutated CaMs act as inhibitors of this mechanism.
CaM is the most ubiquitous mediator of cellular Ca 2ϩ functions, but it has also become clear in recent studies that there is a large number of other EF-hand-containing Ca 2ϩ -binding proteins belonging to the CaM superfamily. Particularly the Ca 2ϩ -binding protein (CaBP) subfamily and the neuronal Ca 2ϩ -sensor (NCS-1) subfamilies that are primarily expressed in neurons may be important for Ca 2ϩ signaling.
Both members of the CaM superfamily are small proteins (about 20 kDa) that share with CaM the basic structure of two N-terminal and two C-terminal EF-hands. However, only three of their EF-hands can bind Ca 2ϩ (Fig. 7a). Hence, we investigated whether two of these Ca 2ϩ -binding proteins, CaBP1 and NCS-1 protein, could also alter the activity of this CICR mechanism in A7r5 cells, because they both have one inactive EFhand. CICR was measured as described above. For CaBP1, both short (sCaBP1) or long (lCaBP1) isoforms were added for 2 min together with a maximally effective free [ 40 Ca 2ϩ ] of 10 M. 10 M of sCaBP1 or lCaBP1 inhibited the CICR by more than 80% (Fig. 7b). Under the same conditions 10 M GST-NCS-1 equally inhibited the CICR mechanism (Fig. 7c). GST (10 M) by itself, however, did not affect the CICR mechanism in our system, indicating a specific effect of NCS-1 (data not shown). To exclude that the remaining Ca 2ϩ -binding sites of NCS-1 could contribute to the inhibitory effect on CICR through simple Ca 2ϩ chelation, the same experiments were conducted using a mutant of NCS-1. NCS-1 E120Q , with its third EF-hand disrupted, showed impaired Ca 2ϩ -dependent conformational changes (47). This mutant was still able to inhibit the CICR mechanism to the same extent as wild type NCS-1 (Fig.  7c), thereby excluding a Ca 2ϩ chelation effect.
To test the hypothesis that the CICR mechanism is activated by a Ca 2ϩ -dependent dissociation or dislocation from an apoCaM-binding site, we performed experiments in which we trapped the endogenous CaM with a high affinity CaM-binding peptide derived from the RyR1 (amino acids 3614 -3643) (30). Fig. 8a shows that in cells incubated during the loading phase with 10 M of the CaM-binding peptide the CICR mechanism was nearly abolished. However, the RyR1 peptide had no effect on the extent of 45 Ca 2ϩ loading of the cells (data not shown). To strengthen the argument regarding the specific effects of the RyR1 peptide, CaM and CaM 1234 were re-added for a 2-min period after stripping the cells with the RyR1 peptide. Readdition of 10 M CaM almost completely restored CICR activation by 3 M free 40 Ca 2ϩ , whereas 10 M CaM 1234 was unable to restore CICR activation (Fig. 8b). Therefore, it is likely that, in permeabilized A7r5 cells, Ca 2ϩ activates a Ca 2ϩ release mechanism by binding to endogenously bound apoCaM.
CICR in Different Cell Types-To verify whether a similar CICR mechanism is also expressed in other cell types we have also screened 16HBE14o-(human bronchial mucosa), LLC-PK 1 (porcine kidney cell line), COS-1, and mouse embryonal fibroblast cells in the same conditions as described above for A7r5 cells. Permeabilized cells loaded with 45 Ca 2ϩ were challenged with 10 M free 40 Ca 2ϩ . A significant CICR response was only found in 16HBE14o-cells, although the fraction of released Ca 2ϩ was smaller (15 Ϯ 3%) than for A7r5 cells. This response in 16HBE14o-cells was also inhibited by CaM 1234 , sCaBP1, lCaBP1, and NCS-1 (data not shown), suggesting that this same CICR mechanism is not only expressed in A7r5 cells but that it could be more widespread.

DISCUSSION
Ca 2ϩ release from the intracellular stores can be triggered by either IP 3 or by a CICR mechanism. Here we report that in permeabilized A7r5 cells an increase in the free [Ca 2ϩ ] c stimulated a Ca 2ϩ release of up to 41% of the intracellular stores with an EC 50 of 700 nM and a Hill coefficient of about 2. This type of CICR mechanism was neither mediated by IP 3 Rs nor by RyRs, because it was not blocked by ruthenium red, ryanodine, heparin, 2-APB, or xestospongin C. ATP dose-dependently stimulated the CICR mechanism, whereas 10 mM MgCl 2 com-pletely abolished it. All these results suggested a novel type of CICR from the non-mitochondrial intracellular stores in permeabilized A7r5 cells. This CICR mechanism did not simply reflect passive 45 Ca 2ϩ / 40 Ca 2ϩ exchange and did not result from the SERCA pumps running in reverse, because thapsigargin was present during the efflux phase.
Recently a similar CICR pathway was identified in hepatocytes (26) suggesting that it may be more ubiquitously expressed in different cell types. We identified a similar type of CICR pathway in 16HBE14o-cells, confirming this idea. Al- though these CICR pathways appear to be quite similar, there are also striking differences between the observations made in the present study as compared with these described for hepatocytes. The CICR mechanism in hepatocytes appeared to be more sensitive, with an EC of 170 nM compared with 700 nM in the present study and was reported to be ATP-independent. This may suggest different types of transporters or at least differences in their regulation.
Polycystin-2 was recently identified as a new Ca 2ϩ release channel. Polycystin-2 is a member of the TRP channel superfamily. Polycystin-2 behaved as a Ca 2ϩ -activated, high conductance ER channel that is permeable to divalent ions and exhibited channel behavior reminiscent of RyRs and IP 3 Rs (22)(23)(24). It remains to be established if the CICR mechanism described in our study could be related to polycystin-2. The observation that LLC-PK 1 cells that endogenously express polycystin-2 did not show the CICR mechanism, however, seems to disprove this hypothesis. Another member of the TRP family, the VR1, was also recently found to act as an intracellular Ca 2ϩ release channel. Capsaicin binding to the VR1 resulted in Ca 2ϩ mobilization from the intracellular Ca 2ϩ stores, and it was found to localize with the ER (27,28). These data suggest that different members of the TRP family can act as intracellular Ca 2ϩ release channels.
The presence of a CICR mechanism could be important for the propagation and amplification of Ca 2ϩ signals initiated by other Ca 2ϩ release channels. Indeed, CICR mediated by RyRs and IP 3 Rs plays a crucial role in amplifying the Ca 2ϩ signals provided by Ca 2ϩ entry in cells such as cardiac myocytes (48), neurons (49,50), astrocytes (51), and pancreatic ␤-cells (52). For example the nature of long-term changes in synaptic activity in the hippocampus depends on whether Ca 2ϩ entry triggers CICR via RyRs or IP 3 Rs (50). It became clear that CICR is an important feature of intracellular signaling. The available data strongly suggest the presence of additional CICR pathways different from the well documented IP 3 R and RyR.
A new finding in our study is that the CICR mechanism described here was inhibited by CaM 1234 and by members of the family of CaM-like Ca 2ϩ -sensor proteins. It became clear from recent work that most of the Ca 2ϩ channels, both situated in the plasma membrane or in intracellular stores, are regulated by CaM, apoCaM, or members of the CaM superfamily. This has recently been well documented for the RyR and the IP 3 R. The skeletal-muscle Ca 2ϩ release channel, RyR1, is activated by apoCaM and inhibited by Ca 2ϩ -bound CaM (10,11,30). For the IP 3 R the functional significance of CaM is not clear (12,13). Other Ca 2ϩ channels, like the voltage-dependent Ca 2ϩ channels (53)(54)(55)(56), as well as members of the TRP family (57)(58)(59)(60)(61), have CaM-and apoCaM-binding sites. We found that the CICR mechanism described in this study is regulated by CaM. The CICR mechanism was not affected by CaM itself, but CaM 1 and CaM 1234 inhibited it. Using CaM 1234 as a negative dominant already revealed the role of CaM in K ϩ channels (62), L-type Ca 2ϩ channels (45,55), P/Q-type Ca 2ϩ channels (56), store-operated channels (63), and the RyR (10,11,30). Our data indicate the presence of an inhibitory CaM-binding site in the absence of Ca 2ϩ (apoCaM-binding site). CaM tethered to this position could then act as a Ca 2ϩ sensor, and CICR could be interpreted as a Ca 2ϩ -dependent dissociation or delocalization of CaM from its inhibitory binding site. The dominant negative effect of CaM 1234 results from its inability to perform a Ca 2ϩ -dependent interaction. Further evidence supporting this hypothesis was obtained by preincubation of permeabilized cells with a high-affinity peptide for CaM, derived from RyR1. Endogenous CaM could be trapped by this peptide, and there-fore the Ca 2ϩ sensor for the CICR mechanism would be removed. In agreement with our hypothesis the preincubation with the RyR1 peptide indeed abolished a subsequent CICR mechanism. Moreover, re-addition of CaM, but not of CaM 1234 , could restore CICR after preincubation with the RyR1 peptide. Because preincubation with the RyR1 peptide during the loading phase did not interfere with 45 Ca 2ϩ loading of the cells, stripping of CaM per se seems not to be sufficient for CICR. The data rather support a mechanism where a Ca 2ϩ -dependent delocalization of CaM to another binding site is responsible for CICR activation. Results obtained with other members of the CaBPs can also be explained by this hypothesis. These CaMlike proteins apparently all show binding affinity in the absence of Ca 2ϩ . By binding to the apoCaM-binding site they may prevent the role of CaM as a Ca 2ϩ sensor. The C termini of CaBPs are highly homologous to the corresponding region in CaM, whereas the N termini are longer and have more variation, including the myristoylation sites or alternative exons. CaBPs also have an extended 32-amino acid-long flexible central ␣-helical segment, versus 28 amino acids in CaM. These differences together with a disabled EF-hand 2 could explain the different binding characteristics as compared with CaM. Indeed, the sequential binding of the highly homologous Cterminal domain with further binding of the N-terminal domain could tether CaBPs to the effector molecules at all [Ca 2ϩ ] (64). Such changes in binding properties of CaM have also been observed when EF-hand 2 was disabled by mutations (65,66). Furthermore, NCS-1 bound in a Ca 2ϩ -independent manner to rat brain membranes (67). The more restricted expression and subcellular localization of the CaM-like Ca 2ϩ -sensor proteins could thereby provide a physiological mechanism to inhibit CICR in specific areas of neurons.
In summary, we found a novel CICR mechanism in A7r5 and 16HBE14o-cells. Although we have not yet established the molecular identity of this novel Ca 2ϩ release pathway, we found that its activation is mediated by CaM. The data suggest that CaM tethered to an inhibitory apoCaM site may act as the Ca 2ϩ sensor for activation of CICR. A possible candidate for this pathway could be a member of the TRP-channel superfamily, like the polycystin-2 channel, but there is as yet no evidence to support this. The apoCaM-binding property described here may offer a practical tool for the future identification of the transport protein involved. In addition, this novel CICR mechanism may provide an additional pathway in Ca 2ϩ release and could play an important role in amplifying Ca 2ϩ signals generated by other Ca 2ϩ release channels.