Determinants for Calmodulin Binding on Voltage-dependent Ca2+ Channels*

Calmodulin, bound to the α1subunit of the cardiac L-type calcium channel, is required for calcium-dependent inactivation of this channel. Several laboratories have suggested that the site of interaction of calmodulin with the channel is an IQ-like motif in the carboxyl-terminal region of the α1 subunit. Mutations in this IQ motif are linked to L-type Ca2+ current (I Ca) facilitation and inactivation. IQ peptides from L, P/Q, N, and R channels all bind Ca2+calmodulin but not Ca2+-free calmodulin. Another peptide representing a carboxyl-terminal sequence found only in L-type channels (designated the CB domain) binds Ca2+calmodulin and enhances Ca2+-dependent I Cafacilitation in cardiac myocytes, suggesting the CB domain is functionally important. Calmodulin blocks the binding of an antibody specific for the CB sequence to the skeletal muscle L-type Ca2+ channel, suggesting that this is a calmodulin binding site on the intact protein. The binding of the IQ and CB peptides to calmodulin appears to be competitive, signifying that the two sequences represent either independent or alternative binding sites for calmodulin rather than both sequences contributing to a single binding site.

essary for calcium-dependent inactivation of the cardiac channel. P/Q channels have also been shown to be modulated by calmodulin (10).
The IQ-like domain in the L-type channels appears to bind primarily Ca 2ϩ -CaM rather than apoCaM (5). An isoleucine to alanine mutation in this motif results in loss of Ca 2ϩ -dependent inactivation and unmasks a strong facilitation by CaM (11,12). If the isoleucine is changed to a glutamate, the effects of CaM (inactivation and facilitation) are lost (11). A double mutation of I1624A/Q1625A (Swiss Prot Q0815) produced an even more pronounced facilitation (12). These investigators concluded that Ca 2ϩ -dependent inactivation requires strong binding of CaM to the IQ motif. Peterson et al. (5) find that a mutant CaM that cannot bind Ca 2ϩ at any of the four Ca 2ϩ binding sites blocks the effects of Ca 2ϩ -CaM on the L-type Ca 2ϩ channel, suggesting that both the Ca 2ϩ -free and Ca 2ϩ -bound forms of CaM bind to this channel. However, only the Ca 2ϩ -bound form can induce inactivation (5). In particular, it appears that Ca 2ϩ binding to sites 3 and 4 of CaM are required for Ca 2ϩ -dependent inactivation (5).
Other studies suggest that the IQ motif may not be the only determinant necessary for Ca 2ϩ -dependent inactivation of the channel (7,13,14). Adams and Tanabe (15) replaced the I-II loop of the cardiac channel with the corresponding loop of the skeletal channel, which slowed Ca 2ϩ -dependent inactivation. An effect of replacing the II-III loop was also observed. The role of the EF hands of the carboxyl-terminal region of the ␣ 1 subunit of the L-type channel in Ca 2ϩ -dependent inactivation has been controversial (6,7,14,16). Bernatchez et al. (16) found that mutations in the EF hand altered Ca 2ϩ -dependent inactivation, but Zhou et al. (6) found that mutant channels containing a triple mutation that disrupted Ca 2ϩ coordinating activity was still inhibited by Ca 2ϩ . Recently, Peterson et al. (14) demonstrated that replacing four amino acids (VVTL) in the F helix of the putative EF hand of ␣ 1C with those of the ␣ 1E (MYEM) channel completely abolished Ca 2ϩ -dependent inactivation. In contrast, mutating the residues presumably involved in coordinating Ca 2ϩ reduced the inactivation only about 2-fold. These authors suggested that the EF hand plays a role in transducing the signal generated by Ca 2ϩ binding to CaM into channel inactivation. Other regions of the carboxyl tail are also likely to be involved in Ca 2ϩ -dependent inactivation. Soldatov et al. (13) found that mutating regions between the EF hand and the IQ motif (IKTEG and LLDQV) eliminated Ca 2ϩdependent inactivation. They suggested that a cooperative interaction between these two regions contributed to Ca 2ϩ -dependent inactivation. Zuhlke and Reuter (7) used deletions to show that three different domains (the putative EF hand, the IQ domain, and the two-amino acid motif, NE, at amino acids 1630 and 1631 of the cardiac L-type channel) are important for the inactivation process. Sequences within the region between amino acids 1572 and 1651 have also been suggested to be important for regulating targeting, conductance, and open probability of the channel (17).
Peterson et al. (5) compared the binding of Ca 2ϩ -CaM to the IQ-like motifs from N-, P/Q-, and R-type calcium channel ␣ subunits. In their study, the P/Q-type (␣ 1A ), R-type (␣ 1E ), and N-type (␣ 1B ) calcium channels bound Ca 2ϩ -CaM, but ␣ 1B had a much lower affinity. P/Q-type calcium channels display a small amount of calcium-dependent inactivation, whereas R and N do not. If Ca 2ϩ -CaM binding to the ␣ 1C is responsible for Ca 2ϩ -dependent inactivation, this result raises the question of the functional role of CaM when bound to R-and N-type calcium channels, since they do not undergo Ca 2ϩ -dependent inactivation.
In this study we show that the IQ domain of the L-type binds partially Ca 2ϩ -saturated calmodulin, whereas the IQ domains of P/Q, R, and N have a much lower apparent affinity for the partially Ca 2ϩ -saturated calmodulin. We also show that another sequence (CB) found in the carboxyl-terminal region of L-type channels binds partially and fully Ca 2ϩ -saturated CaM and enhances I Ca facilitation in cardiac myocytes, suggesting that CB can participate in Ca 2ϩ -dependent modulation of I Ca . This sequence is found between amino acids 1484 and 1509 in skeletal L-type (␣ 1S ) and amino acids 1627-1652 in the cardiac L-type channel (␣ 1C ). This region has recently been shown to play a role in membrane targeting of the L-type channel (18), suggesting the possibility of a role for calmodulin in this process. The corresponding regions of ␣ 1E (amino acids 1828 -1853), ␣ 1A , and ␣ 1B (1815-1840) do not bind calmodulin. This region in L-type channels has been shown by Soldatov et al. (13) to be important for Ca 2ϩ -dependent inactivation. The finding that two domains of cardiac L-type channels bind CaM and each of these binds partially Ca 2ϩ -saturated CaM suggests the possibility that both of these domains may contribute to the Ca 2ϩ -dependent modulation of L-type Ca 2ϩ channels.

MATERIALS AND METHODS
Reagents-Bovine brain calmodulin (95% pure) was purchased from Sigma, solubilized in 10 mM MOPS, 1 mM EGTA, 0.02% NaN 3 , pH 7.4, and quantified by absorption from 320 to 277 nm to obtain stock solutions of about 300 M. Peptides were synthesized at the protein lab facility at Baylor College of Medicine and diluted into 200 mM MOPS, pH 7.4, for assays. The antipeptide antibody was prepared by immunization of rabbits with a peptide representing amino acids 1484 -1509 of the skeletal L-type channel coupled to keyhole limpet hemocyanin. All electrophoresis reagents were analytical grade from Bio-Rad.
Electrophoretic Mobility Shift Assays-Calmodulin electrophoretic mobility was evaluated by non-denaturing polyacrylamide gel electrophoresis under discontinuous conditions as a modified technique described by Laemmli (19). Calmodulin in 200 M CaCl 2 was incubated with the peptides in molar ratios of peptide to calmodulin of 0.1:1, 0.5:1, 1:1, 2:1, 3:1, 5:1, 10:1. Gels evaluating the ability of mutant calmodulins to bind peptides were in molar ratios of peptide to calmodulin of 0.1:1, 1:1, 2:1, 5:1, 10:1 in 200 M CaCl 2 . The extent of the interaction was quantified by densitometer analysis of the absorbance of the uncomplexed CaM at each peptide to CaM molar ratio. Values were normalized to the absorbance in the absence of peptide.
Fluorescence Studies-Peptide (730 pmol, 2.9 M) in 100 mM MOPS, pH 7.4, was added to 730 pmol of brain calmodulin in buffers containing Ca 2ϩ ranging from 1 nM to 100 M. The solution was excited at 280 nm, and emission was detected at 330 nm. Enzyme-linked Immunoabsorbance Assay for Antibody Binding-Microtiter plates were coated with CHAPS-solubilized sarcoplasmic reticulum membranes (100 g of protein/well). The plates were then blocked with 3% bovine serum albumin and incubated in the presence of 2 M CaM. The antipeptide antibody (serial dilutions 1:100 to 1:200,000) was added, and the samples were incubated overnight at 4°C. The secondary goat anti-rabbit (dilution 1:3000) coupled to alkaline phosphatase was added, and the incubation was continued for 2 h at room temperature. After washing, the plates were developed with the alkaline phosphatase substrate, p-nitrophenyl phosphate disodium salt.
Preparation of Membranes and Purified Skeletal Muscle L-type Calcium Channel (the Dihydropyridine Receptor, DHPR)-T-tubule membranes and purified DHPR were prepared as we have previously described (20).

SDS-Polyacrylamide Gel Electrophoresis and Western
Blotting -SDS-polyacrylamide gel electrophoresis and Western blotting were performed as described previously (19,21). Samples (1-2 g of protein), corresponding to the purified L-type calcium channel were applied to a 7.5% SDS-polyacrylamide gel. One gel was stained with Coomassie Brilliant Blue, and another gel was transferred for Western blotting (21). The primary antibody used was a sequence-specific antibody to amino acids 1484 -1509 of skeletal L-type channel, and the second antibody was a goat anti-rabbit antibody coupled to horseradish peroxidase. The blots were developed using the SuperSignal chemiluminescence substrate (Pierce). For competition experiments, CaM (2 M) was incubated with the blot for 6 -8 h at 4°C before the addition of the first antibody.
Calcium Currents in Cardiac Myocytes-Voltage clamp experiments were performed in whole cell mode (22) with freshly isolated rabbit ventricular myocytes, (23). L-type Ca 2ϩ current (I Ca ) was isolated by adding Cs ϩ and tetraethylammonium chloride and reducing Na ϩ and K ϩ in the pipette and bath solutions. Elimination of the residual current by nifedipine (10 M) confirmed that the identity of active current was I Ca (not shown). I Ca was activated (0.5 Hz) by voltage command pulses from Ϫ80 to ϩ10 mV for 300 ms at 24°C. Total charge movement was determined by integrating inward I Ca during the command step using pClamp 6.2 (Axon Instruments) and expressed as a ratio of the nth to the 1st stimulated "beat." The pipette (intracellular) solution was 120.0 mM CsCl, 10 For experiments designed to eliminate CaM binding, Ca 2ϩ was omitted, EGTA was substituted with BAPTA in the pipette solution, and Ba 2ϩ was substituted for Ca 2ϩ in the bath solution. The null hypothesis was rejected for p Ͻ 0.05 using the unpaired Student's t test or analysis of variance as appropriate, and data were expressed as means Ϯ S.E.

RESULTS
Binding of CaM to Peptides from the Different Voltage-dependent Ca 2ϩ Channels-Partial sequences for the carboxylterminal tails of the different voltage-dependent Ca 2ϩ channels are shown in Fig. 1. IQ peptides corresponding to the IQ motif of the different voltage-dependent channels are underlined and italicized in this figure. The cardiac IQ motif has been shown to FIG. 1. The aligned sequences of the amino-terminal parts of carboxyl-terminal tails of the ␣ 1 subunits of the Lsk (skeletal), Lc (cardiac), P/Q-, N-, and the R-type calcium channels. The sequences were obtained from the Swiss-Prot data base (accession numbers: CCAA (brain P/Q-type, ␣ 1A ) O00555; CCAB (brain N-type, ␣ 1B ) Q00975; CCAC (cardiac L-type, ␣ 1C ) Q13936; CCAE (brain R-type, ␣ 1E ) Q15878; CCAS (skeletal, L-type, ␣ 1S ) Q13698. The CB sequences are underlined, and the IQ sequences are underlined and italicized.
contribute to CaM binding and Ca 2ϩ -dependent inactivation (5,7,9,11). However, other sequences in the carboxyl-terminal region of ␣ 1C have also been shown to play important roles in Ca 2ϩ -dependent inactivation. Soldatov et al. (13) indicated the amino acids IKTEG and LLDQV, whereas Zuhlke et al. (7) suggested that the Asn and Glu at positions 1630 and 1631 were crucial for inactivation. This region of the cardiac L-type channel is highly homologous to that of the skeletal L-type channel, with only two conservative changes (Table I, italicized  letters). To address the question of how this region of the L-type channels contributes to CaM binding, we synthesized a peptide with the sequence containing these elements. The primary interest of our current studies is the skeletal muscle protein, and therefore, we prepared a peptide matching the skeletal sequence (designated CB-L-A, Table I). Other shorter peptides synthesized (see below) had identical sequences in the cardiac and skeletal proteins.
To assess the interaction of the peptide with CaM, we examined its ability to bind to CaM on nondenaturing gels. The interaction was analyzed by measuring the absorbance of the CaM band at increasing peptide to CaM ratios in the presence of 200 M Ca 2ϩ . As shown in Fig. 2, CB-L-A binds CaM in the presence of Ca 2ϩ . A representative Coomassie-stained gel is shown in Fig. 2A, top panel, and the absorbance of the CaM band with increasing peptide for three independent experiments is shown in Fig. 2B. To determine which amino acids in this sequence are important for the interaction with CaM, we synthesized several other peptides (CB-L-B, CB-L-C, CB-L-D). The sequences of these peptides are shown in Table I. All of the peptides were able to interact with Ca 2ϩ -CaM (Fig. 2), suggesting that only the sequence LRAIIKKIWKRTSMKLL is required for CaM binding. The two shorter peptides (CB-L-C and CB-L-D) both produced multiple bands in the presence of CaM. The reason for this is not clear. It may be that, under these conditions, more than one peptide can bind to the CaM molecule, possibly by binding at each of the two CaM lobes. We chose CB-L-B, whose sequence is the same in the cardiac and skeletal muscle L-type channels and has similar affinity for CaM as CB-L-A, for subsequent studies because of its simpler gel pattern and because it was shorter and hence less expensive to prepare than CB-L-A.
A Comparison of the IQ and CB Peptides for Binding CaM-We next synthesized two sets of peptides: 1) peptides matching the IQ motifs from cardiac L-, skeletal L-and the P/Q-, N-, and R-type voltage-dependent calcium channels and 2) peptides matching the regions that most closely align with the CB region of these same channels. The sequences of these peptides are shown in Table I. We assessed and compared the interaction of these peptides (both IQ and CB) with CaM on nondenaturing gels. These data for the IQ and CB peptides are summarized in Fig. 3, A and B, respectively. We found, in agreement with Peterson et al. (5), that the IQ domains from all of the different channels bound CaM in the presence of Ca 2ϩ .  IQ peptides and CB peptides of L-type, P/Q-, N-, and R-type voltage-dependent calcium channels are shown. The qualitative results of the analyses described in Fig. 1-4 are also shown in this table. The underlined sequences have been previously shown to be required for Ca 2ϩ -dependent inactivation of the L-type channels.
Of the CB peptides, only those corresponding to L-type channels were able to bind CaM with high affinity in the presence of Ca 2ϩ . The CB-R, however, showed a small amount of binding at high molar ratios. All of the IQ peptides had a strong affinity for CaM with a relative order at high Ca 2ϩ of IQ-Lc ϭ IQ-R Ͼ IQ-P/Q Ͼ IQ-L-sk Ͼ IQ-N. In contrast, only the CB peptide from the L-type Ca 2ϩ channels had a significant affinity for CaM. Ca 2ϩ Dependence of the Interaction of the Peptides with CaM-Previous studies suggest that the L-type channel (cardiac) can bind CaM at low Ca 2ϩ (5). None of the peptides used in our study were able to bind to CaM if the gels were electrophoresed in the presence of 1 mM EGTA with no added Ca 2ϩ (data not shown). This binding site, therefore, does not represent a binding site for apoCaM. We next examined the ability of the various peptides to bind to CaMs that are mutated in the first two (B12), the second two (B34), or all four Ca 2ϩ binding sites (B1234) on CaM. All mutations in CaM involve Glu to Gln substitutions at the z positions for coordinating Ca 2ϩ , resulting in a greatly decreased affinity of the EF hand for Ca 2ϩ (25). All gels were electrophoresed in the presence of 200 M Ca 2ϩ . Consistent with the findings with apoCaM, none of the peptides could bind to the B1234 mutant (data not shown). However, as can be seen in Fig. 4, both the IQ peptide from the cardiac L-type channel and CB-L-B peptide were able to bind B12 with an affinity similar to that seen with the wild type CaM. The IQ-P/Q and IQ-R peptides could also bind B12, but to a lesser extent. The cardiac IQ peptide was also able to bind B34. None of the other peptides bound B34 as well as IQ-Lc. The mutation of either the amino-or carboxyl-terminal Ca 2ϩ binding sites in CaM greatly reduced the affinity of the IQ domains of the skeletal L-type, R-type, and N-type to bind CaM, suggesting that these sites prefer fully Ca 2ϩ saturated CaM. The other CB-L peptides (A, C, and D) bound the mutant CaMs similar to CB-L-B (data not shown).
To quantify the Ca 2ϩ dependence of the cardiac L-type CB peptide-CaM interaction we examined the change in tryptophan fluorescence of the CB peptide upon binding CaM. The emission spectra for the peptide and CaM, alone and in combination, are shown in Fig. 5A. The binding of CaM to the peptide increases the emission and results in a shift of the peak to shorter wavelengths. The Ca 2ϩ dependence of this interaction is shown in Fig. 5B. The EC 50 for the Ca 2ϩ -dependent enhancement of the CB-L-B peptide-CaM interaction is 105 Ϯ 5 nM (n ϭ 3). The absence of a tryptophan in the IQ peptides prevented this type of analysis of their Ca 2ϩ dependence.
Nondenaturing gels in which the CB-L-B peptide was added at increasing concentrations to a 2:1 mixture of IQ-Lsk (Fig.  6A) and CaM, respectively, showed that the CB-L-B peptide could apparently displace the Lsk-IQ peptide from its complex with CaM, indicating a competitive interaction of these two peptides. In contrast, a 1:1 mixture of IQ-Lc (Fig. 6B) with CaM showed very little displacement by CB-L-B. This is likely due to the higher affinity of IQ-Lc compared with IQ-Lsk for CaM. There is no band corresponding to CaM complexed simultaneously to both peptides. These data clearly show that L-IQ and CB-L-B cannot bind simultaneously to CaM. Our data support a model in which the CB and IQ regions represent either distinct or alternative binding sites rather than both sequences contributing to a single binding site.
Demonstration of the Ability of the L-type Ca 2ϩ Channel to Bind Calmodulin Close to the CB Site-To demonstrate that the full-length ␣ 1S subunit can bind calmodulin at a site close to the CB sequence, we prepared an anti-peptide antibody to peptide CB-L-B and examined the ability of calmodulin to block the binding of the antibody to the ␣ 1 subunit of the DHPR on Western blots (Fig. 7A). The skeletal protein was used for these studies because it is more abundant than the cardiac channel, and its sequence in the CB region is identical to that of the cardiac channel. Pre-incubation of the blots with calmodulin before the addition of the antibody blocked the labeling of this subunit by antibody, suggesting that the denatured ␣ 1S protein bound calmodulin at a site close to the antibody binding site. To assess the effects of calmodulin on the binding of the antibody to a channel in the presence of a nondenaturing detergent, we partially solubilized membranes in CHAPS and examined the ability of calmodulin to block the interaction with the antibody in an enzyme-linked immunoabsorbance assay (Fig. 7B). Again, calmodulin blocked the interaction of the antibody with the DHPR.

CB-L-B Alters Calcium Currents in Cardiac
Myocytes-The L-type Ca 2ϩ channel IQ domain can powerfully direct I Ca facilitation (11,12) and inactivation (5,11,12) by an unknown molecular mechanism(s). To test for possible L-type Ca 2ϩ channel regulatory actions of CB, I Ca was measured in cardiomyocytes dialyzed with CB-L-B under conditions favorable or adverse to Ca 2ϩ -CaM-dependent binding. CB-L-B enhanced I Ca facilitation compared with control cells in the presence of a physiologic intracellular Ca 2ϩ concentration (Fig. 8, a, b, and  e), but CB-L-B was without effect on I Ca during increased intracellular Ca 2ϩ buffering when extracellular Ca 2ϩ was substituted for Ba 2ϩ (Fig. 8, c, d, and e). These data support the hypothesis that CB helps to determine Ca 2ϩ -CaM-dependent regulation of L-type Ca 2ϩ currents. DISCUSSION L-type Ca 2ϩ channels show Ca 2ϩ -dependent inactivation that requires CaM. Recent work has centered on the IQ-like domain in the carboxyl terminus of the ␣ subunit in Ca 2ϩ channels as representing the CaM binding site. However, this is an incomplete picture since voltage-dependent Ca 2ϩ channels that do not show Ca 2ϩ -dependent inactivation bind CaM at their corresponding IQ-like sequences in the carboxyl terminus. This raises the question of why the binding of Ca 2ϩ -CaM to the IQ domain on the L-type channel leads to Ca 2ϩ -dependent inactivation, whereas the binding to IQ domains of the other channels does not. If the IQ domain is the primary binding site for CaM on these channels, the binding to the IQ domain of the L-type channel must lead to a secondary change that does not occur in the other channels. Recently Soldatov et al. (13), Zuhlke and Reuter (7), and Peterson et al. (14) found that determinants outside of the IQ motif are necessary for Ca 2ϩ -dependent inactivation. Soldatov et al. (13) showed that the sequences IKTEG and LLDQV are required, and Zuhlke and Reuter (7) demonstrated the requirement for the IQ domain and residues 1630 and 1631 for Ca 2ϩ -dependent inactivation (7). These sequences immediately bracket the CB sequence in L-type channels that we have shown binds CaM. Petersen et al. (14) demonstrated that the exchange of the sequence VVTL in the F helix of the putative EF hand with the corresponding sequence of the R channel abolished Ca 2ϩ -dependent inactivation. The widely spaced regions of the carboxyl tail involved in Ca 2ϩ -dependent inactivation suggest that some of these regions are involved in the CaM binding, whereas others may contribute to the regulation of the channel that occurs after CaM binds. Other portions might be crucial for the communication between these two domains.
Our data with the different IQ and CB peptides are summarized in Table I CaMs mutated in the Ca 2ϩ binding sites (B12 and B34) with an affinity comparable with that of wild type CaM. The P/Q and R channel IQ peptides bound B12 CaM and B34 CaM with reduced affinity compared with wild type CaM, and the N and L-sk IQ peptide showed no ability to bind these mutant CaMs. None of the peptides were able to bind to CaM in less than 10 nanomolar Ca 2ϩ concentrations or to a CaM mutated at all four Ca 2ϩ binding sites. These peptides do not, therefore, represent binding sites for apoCaM. The cardiac L-type channel and, to a lesser extent, the P/Q channels show Ca 2ϩ -dependent modulation. The other channels do not. The unique ability of the cardiac L-type IQ sequence to bind CaM that is not fully Ca 2ϩ saturated may allow CaM to bind to this channel under resting Ca 2ϩ conditions. Higher Ca 2ϩ concentrations could then produce Ca 2ϩ -dependent inactivation. This is the first demonstration of a property of the IQ site on the L-type channels that sets it apart from the IQ sites on the other channels.
We also found that a domain (CB) between the EF hand and the IQ domain of L-type channels was able to bind B12 CaM and wild type CaM in the presence of Ca 2ϩ . Mutations in Ca 2ϩ binding sites 3 and 4 on CaM greatly reduced the affinity of CaM for this sequence. The Ca 2ϩ dependence of the interaction of CB-L-B with CaM (EC 50 ϭ 100 nM) suggests that this interaction could also take place at resting Ca 2ϩ levels in the cell, preparing the channel for inactivation when the Ca 2ϩ reaches the appropriate levels. Half-maximal inhibition of Ca 2ϩ channel activity by Ca 2ϩ occurs at about 4 M Ca 2ϩ (26). The CB-L-B peptide also enhanced I Ca facilitation in cardiac myocytes in a Ca 2ϩ -dependent manner, suggesting that this sequence may have functional significance for regulating L-type Ca 2ϩ channels. The P/Q, R, and N channels did not have comparable sequences that bound CaM. This is the second feature that distinguishes the cardiac L-type channels from the other voltage-dependent channels. The identified CaM binding sequence is within a domain previously suggested to play important roles in regulating Ca 2ϩ -dependent inactivation (7,13), targeting (17,18,27), conductance (17), and open probability (17) of the channel. These findings suggest a role for calmodulin in regulating other aspects of L-type calcium channel function.
In addition to the peptide data, other workers have found that amino acids within and bracketing this sequence are necessary for Ca 2ϩ -dependent inactivation of the channel. Our data support these findings. We have shown that calmodulin can block the interaction of an antibody to the CB-L-B sequence with both the SDS-denatured and the CHAPS-solubilized skeletal muscle L-type channel. Although these findings support a model in which the CB sequence is a CaM binding site, we cannot rule out the possibility that CaM binding to the nearby IQ sequence sterically hinders antibody binding to the CB sequence. Mutations of this region coupled to analysis of CaM binding will be necessary to resolve this issue.
The carboxyl-terminal tail of the L-type channels has also been suggested to play a role in membrane targeting of this protein (18). The CB sequence that we demonstrate to bind CaM is within the region that has recently been suggested to play a role in membrane targeting of the L-type channels (18). CaM binding may also contribute to this process.
The IQ and CB regions of the cardiac L-type channel could either each bind a molecule of CaM or they could both contribute to the same CaM binding site. We have demonstrated that the interactions of the L-type IQ and CB peptides with CaM are competitive. Based on our findings, we propose a model in which the IQ and CB domains in the carboxyl-terminal tails of the L-type channels represent either two distinct binding sites for CaM or alternative sites for the interaction with partially and fully saturated CaM. One possibility is that only one CaM can bind to the carboxyl tail, and which site (IQ versus CB) is occupied is controlled by factors regulating the conformation of this region (for example, Ca 2ϩ binding to the EF hand).