Endogenously Bound Calmodulin Is Essential for the Function of the Inositol 1,4,5-Trisphosphate Receptor*

Calmodulin (CaM) is a ubiquitous Ca2+ sensor protein that plays an important role in regulating a large number of Ca2+ channels, including the inositol 1,4,5-trisphosphate receptor (IP3R). Despite many efforts, the exact mechanism by which CaM regulates the IP3R still remains elusive. Here we show, using unidirectional 45Ca2+ flux experiments on permeabilized L15 fibroblasts and COS-1 cells, that endogenously bound CaM is essential for the proper activation of the IP3R. Removing endogenously bound CaM by titration with a high affinity (pm) CaM-binding peptide derived from smooth muscle myosin light-chain kinase (MLCK peptide) strongly inhibited IP3-induced Ca2+ release. This inhibition was concentration- and time-dependent. Removing endogenously bound CaM affected the maximum release capacity but not its sensitivity to IP3. A mutant peptide with a strongly reduced affinity for CaM did not affect inhibited IP3-induced Ca2+ release. Furthermore, the inhibition by the MLCK peptide was fully reversible. Re-adding exogenous CaM, but not CaM1234, reactivated the IP3R. These data suggest that, by using a specific CaM-binding peptide, we removed endogenously bound CaM from a high affinity CaM-binding site on the IP3R, and this resulted in a complete loss of the IP3R activity. Our data support a new model whereby CaM is constitutively associated with the IP3R and functions as an essential subunit for proper functioning of the IP3R.

The inositol 1,4,5 trisphosphate receptor (IP 3 R) 3 is a homo-or heterotetrameric intracellular Ca 2ϩ release channel with a monomeric molecular mass of ϳ310 kDa. In mammalian tissues, three isoforms have been identified (IP 3 R1-3). All three isoforms are structurally and functionally related (1). The major part of the protein resides in the cytosol, where binding sites for IP 3 (3)(4)(5). The positive regulation of the IP 3 R by Ca 2ϩ may be largely due to a direct binding of Ca 2ϩ to the receptor (6,7). The mechanism by which Ca 2ϩ inactivates IP 3 Rs is less clear. In that respect, much attention has already been focused on elucidating the role of calmodulin (CaM) (8).
CaM is a ubiquitous Ca 2ϩ -binding protein that plays an important role in Ca 2ϩ signaling in many cell types by modulating the activity of numerous proteins, including ion channels (9). Each IP 3 R subunit binds at least one CaM regardless of the [Ca 2ϩ ] (10,11). Both, Ca 2ϩ -dependent and -independent CaM-binding sites have been mapped to different regions of the IP 3 Rs. A Ca 2ϩ -dependent CaM-binding site was identified in the regulatory domains of IP 3 R1 and IP 3 R2 (amino acids (aa) 1564 -1595) (12), and a Ca 2ϩ -independent CaM-binding site was localized in the N-terminal parts of IP 3 R1-3 (aa 49 -81, aa 106 -121) (13). At nanomolar [Ca 2ϩ ], CaM does not affect IP 3 R function, whereas at micromolar [Ca 2ϩ ] CaM inhibits IP 3 R function. The original hypothesis therefore was that Ca 2ϩ CaM mediates the Ca 2ϩ -dependent inactivation of IP 3 Rs (14 -16). However, this hypothesis became controversial, because we have recently shown, using a CaM mutant (CaM 1234 ), that CaM is not a Ca 2ϩ sensor, as such, for the IP 3 R. We demonstrated that Ca 2ϩ -independent CaM binding to the N-terminal CaM-binding site is responsible for the CaM inhibition of IP 3 -induced Ca 2ϩ release (IICR), even though this inhibition requires Ca 2ϩ (17). This can be explained by the large conformational change that the IP 3 R undergoes in the presence of Ca 2ϩ and which may be necessary for CaM action. This interaction may provide a tonic regulation of IP 3 R activity and can explain the low sensitivity of the IP 3 R in neuronal tissues where CaM is highly expressed. The role of the Ca 2ϩ -dependent CaM-binding site in the regulatory domain, however, still remains to be elucidated.
CaM-binding sites have not merely been identified as regulatory sites but have also recently been implicated in inter-and intrasubunit interactions. For example, the CaM-binding domain of the RyR1 modulates channel activity by at least two mechanisms: 1) by direct binding of CaM and 2) by forming a bridge between two different regions on the RyR1. Peptides of the RyR1 were used to demonstrate that the CaM-binding region is indeed directly involved in intersubunit interactions between the RyR subunits (18). In the case of small conductance K ϩ channels, CaM is involved in intersubunit interactions. Functional small conductance K ϩ channels are heteromeric complexes with CaM, which is constitutively associated with the ␣-subunits in a Ca 2ϩ -independent manner (19). In this study, we investigated whether endogenously bound CaM is essential for IP 3 R functioning. To test this hypothesis, we measured IP 3 R activity in the presence of different synthetic peptides corresponding to the CaM-binding region of myosin light-chain kinase (MLCK), which has a very high affinity for CaM, and in the presence of peptides corresponding to CaM-binding sites of the IP 3 R and other channels. We found that removing endogenously bound CaM dramatically reduced the efficacy of the IP 3 R to release Ca 2ϩ . Our data support a new model whereby CaM is constitutively associated with the IP 3 R as an essential subunit for proper functioning of the IP 3 R.

MATERIALS AND METHODS
Synthetic Peptides-The synthetic peptides derived from myosin light-chain kinase (MLCK, Ac-RRKWQKTGHAVRAIGRL-NH 2 ; and MLCK control peptide, Ac-RRKEQKTGHAVRAIGRE-NH 2 ) were obtained from Calbiochem. Peptides derived from IP 3 R1 (peptide B, PPKKFRDCLFKLCPMNRYSAQKQFWKAAKPGAN; peptide E, ENR-KLLGTVIQYGNVIQLLHLKS; peptide 1564 -1595, KSHNIVQKTAL-NWRLSARNAARRDSVLAASRD), RyR1 (KSKKAVWHKLLSKQRRR-AVVACFRMTPLYN), and hTrp3 (SFNSILNQPTRTQQIMKRLIKRY-VLKAQVD) were all obtained from Sigma. Autocamtide 2-related inhibitory peptide (AIP) was obtained from Sigma. 45 Ca 2ϩ Fluxes-L15 cells were obtained by stable exogenous expression of IP 3 R1 in Lvec cells, whereas Lvec cells represent the control cells expressing the empty vector (20,21). In some experiments, COS-1 cells were 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 seventh and ninth day after plating. The cells were permeabilized by incubation 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 non-mitochondrial Ca 2ϩ stores were loaded for 45 min at 25°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 4 M thapsigargin. Thapsigargin was added to block the endoplasmic reticulum Ca 2ϩ pumps during subsequent additions of Ca 2ϩ . The efflux medium was replaced every 2 min, and the efflux was performed at 25°C. The additions of Ca 2ϩ , IP 3 and peptides are as indicated in the figures. Free [Ca 2ϩ ] was calculated by the Cabuf program (ftp.cc.kuleuven.ac.be/pub/droogmans/cabuf.zip) and based on the stability constants as previously published (22). At the end of the experiment, the 45 Ca 2ϩ remaining in the stores was released by incubation with 1 ml of a 2% sodium dodecyl sulfate solution for 30 min. Ca 2ϩ release in some experiments was plotted as fractional loss, i.e. the amount of Ca 2ϩ released in 2 min divided by the total 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.
Ca 2ϩ Measurements in Sea Urchin Egg Homogenates-Agonist-induced calcium release was measured with the fluorescent dye fluo3 (3 M) (Invitrogen) using Lytechinus pictus egg homogenates as previously described (23). Briefly, homogenates were diluted successively to 2.5% final v/v over 4 h at 17°C in Glu IM (250 mM potassium gluconate, 20 mM HEPES, and 1 mM MgCl 2 (pH 7.2)) supplemented by an ATP regeneration system (1 mM ATP, 10 units/ml creatine kinase, and 10 mM phosphocreatine) together with protease inhibitors. Aliquots of 100 l of homogenate were added to each well in 96-well plates (Greiner Bio-One) and read by a NOVOstar microplate reader (BMG Labtech, Aylesbury, UK). Additions of agonists were made in 1-l volumes. Measurements were carried out at room temperature. Values were considered statistically different when p ϭ Ͻ 0.05.
[ 3 H]IP 3 Binding-Binding studies were performed on microsomes of IP 3 R1-expressing Sf9 cells (150 g) or microsomes derived from cere-bellum (150 g). [ 3 H]IP 3 binding was performed at 0°C in 160 l of binding buffer containing 50 mM Tris-HCl (pH 7.0), 1 M free [Ca 2ϩ ], 10 mM ␤-mercaptoethanol, and 10 nM [ 3 H]IP 3 . Nonspecific binding was determined in the presence of 12.5 M unlabeled IP 3 . After 30 min of incubation, the samples were rapidly filtered through glass fiber filters. Statistical analysis was performed using the paired Student's t test. Values were considered significantly different when p ϭ Ͻ0.05.

CaM-binding Peptide Derived from MLCK Decreases the IP 3 R
Activity-A synthetic CaM-binding peptide (Ac-RRKWQKT-GHAVRAIGRL-NH 2 ) derived from the aa sequence of smooth muscle MLCK was previously characterized as an inhibitor of CaM activity. We used this peptide as a very effective inhibitor of CaM function, as the K d for CaM was measured to be very low (6 pM) (25).
The non-mitochondrial Ca 2ϩ stores of permeabilized L15 cells were first loaded to equilibrium with 45 Ca 2ϩ and then incubated in efflux medium containing 0.5 M free Ca 2ϩ . Thapsigargin (4 M) was added to the efflux medium to allow a unidirectional Ca 2ϩ efflux. Fig. 1A, filled squares, illustrates that a 2-min exposure to 1 M IP 3 accelerated the rate of Ca 2ϩ loss. At this concentration, IP 3 released 60 Ϯ 4% of the maximum releasable Ca 2ϩ measured by the addition of 5 M A23187. The IICR was much less pronounced when, prior to exposure with 1 M IP 3 , 10 M MLCK peptide was added to the efflux medium for a period of 2 min. In this condition, 1 M IP 3 released only 28 Ϯ 5% of the maximum releasable Ca 2ϩ (Fig. 1A, triangles).
Ca 2ϩ release was always measured in the absence of Mg 2ϩ -ATP. Moreover, because there were four wash steps between the loading of the stores in the presence of Mg 2ϩ -ATP and the challenge with IP 3 , all residual Mg 2ϩ -ATP should have been effectively removed. The involvement of the Ca 2ϩ -and CaM-dependent protein kinase (CaMKII) in the observed inhibition by the MLCK peptide seems therefore unlikely. In addition, we have also tested the effect of the CaMKII inhibitor AIP. KN93, which is also a common inhibitor, was not used because of its side effects on IICR (26). AIP had no effect on IICR, and the inhibition of IICR by the MLCK peptide (10 M) was not affected (Fig. 1D). Similar data were observed with KN62, another CaMKII inhibitor (data not shown). We also used FK506 to exclude the role of calcineurin. Similar to AIP, FK506 did not show any effect on IICR (Fig. 1D). These findings exclude the involvement of two important CaM-binding proteins, CaMKII and calcineurin, in the inhibition of the IP 3 R by the MLCK peptide.
Removal of Endogenously Bound CaM Causes Inhibition of IICR-It is conceivable that the observed decrease in IP 3 R activity caused by the MLCK peptide might be due to either 1) a direct effect of the peptide on the IP 3 R, 2) an indirect effect produced by the MLCK peptide binding endogenous CaM and thereby removing CaM from the IP 3 R, or 3) a combination of the two effects. In an effort to distinguish between these possibilities, we have used a mutant of the MLCK peptide (MLCK control) in which a critical tryptophan and a leucine residue were mutated to a glutamate (Ac-RRKEQKTGHAVRAIGRE-NH 2 ). These mutations strongly reduced the affinity of the peptide for CaM from 6 pM to 10 M (25). Application of this mutant peptide did not result in a decreased IICR (Fig. 1B). We used the peptide in conditions in which it does not bind CaM. As the MLCK peptide only binds Ca 2ϩ CaM, we measured IICR on permeabilized cells in efflux buffer containing 1 mM EGTA (Fig.  1C). In both cases, using MLCK peptide (Fig. 1C, triangles) or the MLCK-control peptide (circles) in the absence of Ca 2ϩ , no decrease in IICR was observed. These findings suggest that the inhibitory action of the MLCK peptide is dependent on its association with CaM.
Characteristics of the Inhibition of IICR by MLCK Peptide-We further assessed the possibility that the removal of CaM endogenously bound to the IP 3 R directly resulted in a decrease of the IP 3 R activity. In such a case, inhibition of IICR would be dependent on the concentration of the peptide as well as on the time of incubation with the peptide. A dose-response curve revealed that the MLCK peptide inhibited IICR with an IC 50 value of 6.9 Ϯ 0.7 M when the permeabilized cells were incubated for 2 min prior to activation with 1 M IP 3 (Fig. 2, squares). The amount of Ca 2ϩ released by IP 3 was measured as a percentage of the maximum releasable Ca 2ϩ by the addition of 5 M A23187.
Next, we measured the dose-response curve on permeabilized cells that were incubated for 4 (Fig. 2, circles) or 6 ( Fig. 2, triangles) min with MLCK peptide, replacing the peptide-containing efflux solution every 2 min. We observed that IICR was dramatically decreased when the peptide was added for a longer time. The IC 50 value shifted to 4.6 Ϯ 0.5 and 2.1 Ϯ 0.7 M, respectively, for the 4-and 6-min incubations. In the absence of IP 3 , we observed no effect of the peptide on the rate of efflux or on the endoplasmic reticulum Ca 2ϩ content, when the peptide was added for a longer period (Fig. 2, inset). Both the concentration and the time dependence indicate that MLCK peptide inhibits IICR by removing endogenously bound CaM from the IP 3 R. The dissociation of CaM from a high affinity binding site could be the rate-dependent step in this process. From the observation that the dissociation takes several minutes, it can be expected that the affinity is in the nanomolar or subnanomolar range.
Removal of Endogenously Bound CaM Alters the Extent of IICR but Not the Affinity of the IP 3 R for IP 3 -Previously, we have shown that adding exogenous CaM to permeabilized cells inhibited IICR (14,17). This inhibition is Ca 2ϩ -and IP 3 -dependent. Fig. 3A shows a dose-re- sponse curve of the Ca 2ϩ release as a function of [IP 3 ] in the absence or presence of 10 M CaM. Indeed, in L15 cells, 10 M CaM inhibited IICR by increasing the EC 50 value from 0.44 Ϯ 0.04 to 0.97 Ϯ 0.07 M but did not modify the extent of Ca 2ϩ release at high [IP 3 ]. This is in accordance with our previously obtained data in A7r5 cells (14). In contrast, the addition of MLCK peptide (1 or 10 M) had no effect on the EC 50 value for Ca 2ϩ release compared with the control condition. The IC 50 value was 0.48 Ϯ 0.07 and 0.46 Ϯ 0.05 M for 1 and 10 M MLCK peptide, respectively (Fig. 3B). MLCK peptide, however, clearly reduced the extent of IICR. The addition of 10 M MLCK peptide reduced the maximum IICR induced by 100 M IP 3 from 84 Ϯ 5 to 25 Ϯ 6% of the total releasable Ca 2ϩ . The addition of exogenous CaM (10 M) to 1 M MLCK peptide prevented the above effects (data not shown).
In addition, IP 3 -binding measurements were performed in the presence of 10 M of MLCK and MLCK control peptide on microsomes from Sf9 cells overexpressing IP 3 R1. As expected from the IP 3 doseresponse curve for IICR (Fig. 3B), MLCK peptides did not alter the sensitivity of the IP 3 R for IP 3 (Fig. 3C). IP 3 binding on microsomes of L15 cells was also not changed in the presence of MLCK peptide (data not shown). However, to exclude the possibility that IP 3 R overexpression in Sf9 cells would completely deplete endogenous CaM levels, leaving the bulk of the IP 3 R unassociated, we also performed IP 3 -binding experiments on microsomes derived from cerebellum, which contains high endogenous CaM levels. Interestingly, MLCK peptide increased IP 3binding in this preparation, whereas the MLCK control peptide did not (Fig. 3C). This is in agreement with our previous observations that CaM bound to the N-terminal CaM-binding domain of the IP 3 R inhibits IP 3 binding (17).
Thus, the major effect of removing endogenously bound CaM by adding MLCK peptide appears to be a reduction in the extent of IICR rather than reducing the affinity of IP 3 for the IP 3 R. This is in contrast to the effect of adding exogenous CaM to cells that were not exposed to MLCK peptide, where the major effect consists in reducing the affinity of the IP 3 R for IP 3 , without altering the extent of IICR (17).
Different CaM-binding Peptides Inhibit IICR-To confirm the idea that trapping endogenously bound CaM inhibits IICR, we reasoned that, by using other CaM-binding peptides, we should obtain similar results as those for MLCK peptide. Several CaM-binding peptides with different apparent affinities for CaM were used in a 45 Ca 2ϩ flux assay. Permeabilized L15 cells were incubated for 2 min with 10 M peptides derived from the RyR1 (aa 3614 -3643; K d , 10 nM) (27) and human  homologue of Drosophila Trp (transient receptor potential) hTrp3 (aa 764 -793; K d , 100 nM) (28) prior to activation with 1 M IP 3 . We also used the CaM-binding peptides derived from the IP 3 R1: peptide B (aa 49 -81; K d , 100 nM), peptide E (aa 106 -128; K d , 1 M) (13), peptide 1564 -1595 (aa 1564 -1595; K d , 64 nM) (12,17). When the apparent affinity of the peptides was plotted as a function of the extent of inhibition after application of 1 M IP 3 in the presence of 0.5 M Ca 2ϩ , we observed a good correlation between the affinity of the peptide for CaM and the percentage of inhibition (Fig. 4). In the absence of Ca 2ϩ in the efflux medium, RyR1 peptide and peptide E, which can bind Ca 2ϩ -free CaM, showed an inhibition of IICR of 22 Ϯ 4 and 11 Ϯ 6%, respectively, which is comparable with the inhibition in the presence of Ca 2ϩ (Fig. 4). Other peptides that do not bind CaM in the absence of Ca 2ϩ did not inhibit IICR (data not shown).
Effect of CaM-binding Peptides on Different IP 3 R Isoforms-We measured the effects of MLCK peptide in different cell lines to evaluate the effects on the different IP 3 R isoforms. The effect of 10 M MLCK peptide was measured on permeabilized Lvec cells (expressing primarily IP 3 R3) (see "Materials and Methods") and in COS-1 cells (expressing both IP 3 R2 (ϳ25%) and IP 3 R-3 (ϳ65%)). In both cell lines, the addition of 10 M MLCK for 2 min inhibited IICR with 51 Ϯ 3 and 40 Ϯ 4%, respectively (data not shown). Although the effects were clearly present in three cell lines expressing different isoforms, it is difficult to conclude whether both isoforms present in COS-1 cells equally contributed to the observed inhibition. From the data in L15 and Lvec cells, we can, however, conclude that IICR is effectively inhibited for at least IP 3 R1 and -3.
Next, we measured the effect of the peptides in sea urchin egg homogenates. Homogenates were loaded with 3 M fluo3 and induced with 2 M IP 3 or 200 nM cADPR to stimulate the IP 3 R or RyR, respectively. In the presence of 50 M MLCK peptide, IICR was reduced to 62.1 Ϯ 2.7% of the control (Fig. 5A), whereas cADPR-induced Ca 2ϩ was not affected (Fig. 5B). The control peptide did not affect release through the IP 3 R or RyR.
Reversing the Effects of the Peptide by Re-addition of CaM-We assessed the possibility of reversing the inhibition on IICR caused by the removal of CaM by re-adding exogenous CaM. We showed earlier that adding exogenous CaM inhibited IICR (Fig. 3A) in an IP 3 -dependent way. Therefore we chose to perform the experiment at an [IP 3 ] of 100 M. At this high [IP 3 ], exogenous CaM did not inhibit IICR, whereas IICR was still largely inhibited by MLCK peptide (Fig. 3). Permeabilized L15 cells were first incubated with 1 M peptide for 6 min, and efflux medium with peptide was then replaced by efflux medium with (Fig. 6, triangles) or without (Fig. 6, circles) 1 M CaM for 4 min. The cells were then challenged with 100 M IP 3 , and IICR was compared with cells that were only challenged with 100 M IP 3 (Fig. 6, squares). In the condition in which only MLCK peptide was added, IICR was inhibited by 42 Ϯ 4% compared with the control condition. In the condition in which exogenous CaM was added after the addition of the peptide, IICR was only reduced by 14 Ϯ 5%. This implicates that adding exogenous CaM after having removed CaM with MLCK-peptide partially restores the activity of the IP 3 R. The same experiment was performed with a mutant CaM, which was mutated in the 4 EF-hands, CaM 1234 (Fig. 6, inverted triangles). In contrast to CaM, CaM 1234 did not restore IICR. We further investigated whether CaM mutants, mutated in the N-terminal EF-hands (CaM 12 ) or in the C-terminal EF-hands (CaM 34 ), could restore IICR. But neither CaM 12 nor CaM 34 restored IICR. A CaM-like protein, i.e. calcium-binding protein-1 (CaBP1), which was recently shown to interact with the N-terminal CaM-binding site of the IP 3 R (24, 29 -30), could also not restore IICR (data not shown). We conclude that only fully Ca 2ϩ -loaded Ca 2ϩ CaM is able to reverse the inhibition by MLCK peptide.

DISCUSSION
CaM is the first protein to have emerged as a potential inhibitory Ca 2ϩ sensor for the IP 3 R, but the exact nature of its involvement  remains unclear. Exogenous CaM inhibits IICR in a Ca 2ϩ -dependent way in various cell types expressing different combinations of IP 3 R isoforms (14,16,31). In a previous study, we have shown that the Ca 2ϩdependent inhibition occurs through binding of CaM to the N-terminal Ca 2ϩ -independent CaM-binding site. The Ca 2ϩ dependence then was proposed to be a property of the IP 3 R itself. We suggest that CaM is endogenously bound to the IP 3 R (17).
Recently, it was also shown that CaM and CaM-binding sites are directly involved in inter-and intrasubunit interactions for RyR1 (18,32). Here we investigated the effects of removing endogenously bound CaM from the IP 3 R. Previously, very hydrophobic compounds, such as calmidazolium or W-7, were used as CaM antagonists. This approach, however, resulted in different observations and led to some confusion. W-7, calmidazolium, trifluoperazine, chlorpromazine, and fendiline caused Ca 2ϩ release from the endoplasmic reticulum (33)(34)(35), whereas it was also reported that they directly inhibited IICR without interacting with CaM (36).
We therefore tested the hypothesis that CaM is endogenously bound to the IP 3 R and might be involved in inter-or intrasubunit interactions by measuring the effects on IICR of different synthetic peptides representing CaM-binding regions. First, we used the CaM-binding region derived from MLCK, as this peptide was found to bind CaM with an extremely high affinity (K d , 6 pM) (25). The results demonstrated that adding MLCK peptide to permeabilized L15 cells dramatically decreased IICR. We showed that MLCK peptide caused a reduction in the extent of IICR with no change in the affinity of IP 3 for the IP 3 R. A direct effect of the different CaM-binding peptides on the IP 3 R is unlikely, as except for their CaM-binding abilities, they have no particular homology. Furthermore, the observation that MLCK peptide was only effective in the presence of Ca 2ϩ , whereas the Ca 2ϩ -independent RyR peptide also inhibited IICR in the absence of Ca 2ϩ , is more compatible with the removal of endogenous CaM than with binding of these peptides directly to the IP 3 R. Furthermore, using inhibitors for CaMKII and calcineurin, we also excluded the involvement of both CaM-binding proteins. Although we cannot fully exclude the effect of an as yet unknown auxiliary CaM-binding protein, we therefore concluded that removal of endogenous CaM from the IP 3 R was the most probable cause of the inhibition of IICR. This hypothesis was strengthened by reversing the inhibitory effects of the peptides by adding exogenous CaM but not CaM 1234 , CaM 12 , or CaM 34 .
Previously, CaBP1 was identified as a novel activator of the IP 3 R (29). Although this is still controversial (24,30), we have envisaged the hypothesis that CaBP1 could activate IP 3 R after complete removal of CaM from the IP 3 R. However, we did not observe any activation by CaBP1 alone or potentiation of IICR after removal of the endogenously bound CaM. This indicates that the N-terminal CaM-binding site that is also used by CaBP1 (24) is not involved in this process.
Interestingly, exogenously added CaM and removal of endogenously bound CaM inhibited IICR by a completely different mechanism. Although adding exogenous CaM inhibited IICR by reducing the affinity of the IP 3 R for IP 3 , removing endogenously bound CaM with the peptides did not alter the IC 50 value for IP 3 but dramatically decreased the maximum release capacity of the IP 3 R. This indicates that a different CaM-binding site than the N-terminal CaM-binding site must be involved. The CaM-binding site that was previously identified in the regulatory region of IP 3 R1 and IP 3 R2 (aa 1564 -1585) can also not explain the effect observed here, as this site is not conserved in IP 3 R3 (12). The inhibitory effects were observed in IP 3 R1 and IP 3 R3, because they were observed in different cell types, expressing mainly either one of these isoforms. These inhibitory effects were also observed in sea urchin egg homogenates. The effects probably therefore involve an as yet unidentified CaM-binding site on the IP 3 R.
Recently, major progress has been made in understanding the gating mechanisms of the IP 3 R (37). It becomes clear from the data that intramolecular interactions control the opening and closing of the channel. Reports demonstrated the proximity and physical interaction of the ligand-binding domain with the IP 3 R channel domain (38,39). The truncated receptor containing only the transmembrane segments, i.e. the channel domain, was shown to be constitutively active (40). In addition, studies by Ramos-Franco et al. (41) show that a truncated IP 3 R missing transmembrane segments 1-4 also forms a constitutively open channel. Different regions may therefore contain important domains to keep the channel closed. Current views suggest that IP 3 binding initiates a conformational change or perhaps a series of conformational changes that relieve the inhibitory effect of a yet unidentified segment of the regulatory region (42) and/or the transmembrane segments 1-4 (41). One could speculate that CaM is directly involved in stabilizing these intramolecular interactions or that CaM is directly involved in the gating mechanism itself. Removing this endogenous CaM would disable the IP 3 R to open or close properly. As the endoplasmic reticulum content is not affected by the CaM-binding peptides, we suggest that the removal of CaM stabilizes the IP 3 R in a closed conformation. In the case of the RyR1, disrupting the intersubunit interactions directly altered the opening properties of the channel, leading to an increase in spontaneous Ca 2ϩ sparks in frog skeletal muscle. The CaM-binding site in the RyR1 is thought to be at an intersubunit contact site, and depletion of CaM may sensitize the channel to activation by disrupting this intersubunit interaction (18,32).
Our data are most easily interpreted as depletion of endogenous CaM, because the effect is proportional to the affinity of the peptide for CaM. This depletion did not empty the internal Ca 2ϩ stores, indicating that the IP 3 R remained in a closed state. CaM seems to be required for the coupling mechanism between IP 3 binding and channel opening, perhaps by stabilizing the IP 3 -induced conformational change.
In summary, we have shown that a high affinity CaM-binding peptide strongly decreased the ability of the IP 3 R to release Ca 2ϩ . We found that stripping CaM from the IP 3 R was responsible for this inhibition. These results are consistent with a model in which CaM is not merely a regulator of the IP 3 R but is also an essential component in the proper gating of the IP 3 R.