Role of the C terminus of the α1C(CaV1.2) Subunit in Membrane Targeting of Cardiac L-type Calcium Channels*

We have previously demonstrated that formation of a complex between L-type calcium (Ca2+) channel α1C (CaV1.2) and β subunits was necessary to target the channels to the plasma membrane when expressed in tsA201 cells. In the present study, we identified a region in the C terminus of the α1C subunit that was required for membrane targeting. Using a series of C-terminal deletion mutants of the α1C subunit, a domain consisting of amino acid residues 1623–1666 (“targeting domain”) in the C terminus of the α1C subunit has been identified to be important for correct targeting of L-type Ca2+ channel complexes to the plasma membrane. Although cells expressing the wild-type α1C and β2a subunits exhibited punctate clusters of channel complexes along the plasma membrane with little intracellular staining, co-expression of deletion mutants of the α1C subunit that lack the targeting domain with the β2a subunit resulted in an intracellular localization of the channels. In addition, three other regions in the C terminus of the α1C subunit that were downstream of residues 1623–1666 were found to contribute to membrane targeting of the L-type channels. Deletion of these domains in the α1C subunit resulted in a reduction of plasma membrane-localized channels, and a concomitant increase in channels localized intracellularly. Taken together, these results have demonstrated that a targeting domain in the C terminus of the α1C subunit was required for proper plasma membrane localization of the L-type Ca2+ channels.

We have previously demonstrated that formation of a complex between L-type calcium (Ca 2؉ ) channel ␣ 1C (Ca V 1.2) and ␤ subunits was necessary to target the channels to the plasma membrane when expressed in tsA201 cells. In the present study, we identified a region in the C terminus of the ␣ 1C subunit that was required for membrane targeting. Using a series of C-terminal deletion mutants of the ␣ 1C subunit, a domain consisting of amino acid residues 1623-1666 ("targeting domain") in the C terminus of the ␣ 1C subunit has been identified to be important for correct targeting of L-type Ca 2؉ channel complexes to the plasma membrane. Although cells expressing the wild-type ␣ 1C and ␤ 2a subunits exhibited punctate clusters of channel complexes along the plasma membrane with little intracellular staining, co-expression of deletion mutants of the ␣ 1C subunit that lack the targeting domain with the ␤ 2a subunit resulted in an intracellular localization of the channels. In addition, three other regions in the C terminus of the ␣ 1C subunit that were downstream of residues 1623-1666 were found to contribute to membrane targeting of the L-type channels. Deletion of these domains in the ␣ 1C subunit resulted in a reduction of plasma membranelocalized channels, and a concomitant increase in channels localized intracellularly. Taken together, these results have demonstrated that a targeting domain in the C terminus of the ␣ 1C subunit was required for proper plasma membrane localization of the L-type Ca 2؉ channels.
Voltage-activated calcium channels are heteromeric complexes composed minimally of an ␣ 1 subunit together with accessory ␤ and ␣ 2 ␦ subunits. The ␣ 1 subunit is the channel pore-forming subunit, which contains binding sites for pharmacological agents or toxins and determines basic electrophysiological properties of different types of Ca 2ϩ channels. The accessory ␤ and ␣ 2 ␦ subunits play important roles in modulating Ca 2ϩ channel function, including modulation of voltage-dependent properties and membrane targeting of channel complexes (1)(2)(3)(4)(5)(6). The roles of the ␤ subunits are better understood than those of the ␣ 2 ␦ subunits. For example, ␤ 1a subunit null mice express greatly reduced Ca 2ϩ currents and dihydropyri-dine binding sites in skeletal muscle myotubes, suggesting that the ␤ subunit plays critical roles in maintaining the expression of the ␣ 1 subunits (7).
The molecular events responsible for targeting the subunits of voltage-activated Ca 2ϩ channels to the plasma membrane are only beginning to be understood. There are at least 10 identified genes for different ␣ 1 subunits (8), and each is predicted to contain 24 membrane-spanning domains, with the N and C-terminal domains located intracellularly (6). In early studies with the L-type ␣ 1C subunit (Ca V 1.2; Ref. 8), a surprising finding was that this protein was not targeted to the plasma membrane when it was expressed alone, but rather remained in a perinuclear location (4). All of the four known ␤ subunits are very hydrophilic and have no known transmembrane domains (9,10). With few exceptions, these subunits are localized cytoplasmically when expressed in the absence of the ␣ 1 subunits (11). One exception is the rat ␤ 2 subunit that is dually palmitoylated in its N terminus (10); this modification serves as a membrane targeting signal for this subunit (11). In contrast to what is observed when the channel subunits are expressed alone, co-expression of the ␣ 1C subunit with any ␤ subunit allows for membrane targeting of both proteins (4,12).
All ␤ subunits contain two conserved domains in the central region flanked by unique N and C termini (9,10). A region in the second conserved domain of all ␤ subunits has been identified to mediate interaction with the ␣ 1 subunits, and this region is termed the ␤-interaction domain (BID) 1 (13,14). In addition, an Src homology 3 (SH3) domain has been mapped to the first conserved domain of ␤ subunits (11). Mutations of key residues in either the BID or the SH3 domain disrupt interaction between the ␣ 1 and ␤ subunits (11,13,14). More recently, we have demonstrated that mutations in the BID or the SH3 domains of ␤ subunits also disrupt membrane targeting of the channels (12). This finding, as well as the observation that membrane targeting of the ␣ 1C can occur with any ␤ subunit (regardless of its state of palmitoylation), has led us to suggest that correct plasma membrane targeting of the calcium channel subunits requires complex formation between the ␣ 1 and ␤ subunits rather than palmitoylation of the ␤ subunit (12).
Although recent studies have addressed functional domains in the ␤ subunit that are required for targeting calcium channel complexes to the plasma membrane (11,12), less is known about the role of the ␣ 1 subunit in the membrane targeting process. Recently, an endoplasmic reticulum retention signal in the I-II loop of the ␣ 1 subunit was identified and shown to be neutralized by interaction with a ␤ subunit (15). Other studies have focused on the C terminus of the ␣ 1C subunit and shown that it is involved in multiple events regulating calcium channel functions. For example, a protein kinase A-mediated protein phosphorylation site that can modulate channel function has been mapped to residue serine 1928 in the C terminus of ␣ 1C subunit (16 -18). In addition, the C terminus of the ␣ 1C subunit contains interaction sites for the Ca 2ϩ -binding proteins sorcin (19) and calmodulin (20 -23). Ca 2ϩ -dependent inactivation and facilitation of cardiac L-type Ca 2ϩ channels may be mediated in part by association of calmodulin with the ␣ 1C subunit (20 -23). Furthermore, it has been suggested that the C terminus of the ␣ 1C subunit contains elements that are inhibitory to channel activity, as deletion of parts of the C terminus resulted in increased Ca 2ϩ current in Xenopus oocytes expressing the L-type calcium channels (24).
In the present study, we investigated the role of the C terminus of the ␣ 1C subunit in the membrane targeting of cardiac L-type Ca 2ϩ channels. By using a series of C-terminal deletion mutants of the ␣ 1C subunit, we identified a region in the C terminus of the ␣ 1C subunit that was required for membrane targeting of the L-type Ca 2ϩ channels.

EXPERIMENTAL PROCEDURES
Materials-All reagents were obtained from general sources unless otherwise stated. The large T-antigen transformed human embryonic kidney cells (tsA201) were the generous gift of Dr. Richard Horn (Thomas Jefferson University, Philadelphia, PA). The Ca 2ϩ channel subunitspecific antibodies used in this study including Card C, Card I, ␤ GEN , and anti-␤ 2 antibodies were described previously (4,25). The pCR3␣ 1C ⌬1733 expression vector was the generous gift of Dr. Roman Shirokov (Rush University, Chicago, IL). The expression plasmid of the rabbit ␤ 2a , pCDNA␤ 2a , was the gift of Dr. Franz Hofmann (Technical University of Munich, Munich, Germany). The expression plasmid pCR3␣ 1C ⌬PRD was described previously (26).
Antibody Preparation and Purification-To generate an additional ␣ 1C subunit-specific antibody, a fusion protein encoding the second intracellular loop linking the transmembrane domains II and III (amino acid residues 738 -940, termed CI2) of the ␣ 1C subunit was produced. The sequence in the CI2 region was amplified using PCR and subcloned into an expression vector pGEX-4T-1 (Amersham Pharmacia Biotech), resulting in an in-frame fusion of the CI2 residues to glutathione S-transferase (GST). The GST-CI2 fusion proteins were expressed in Escherichia coli, and purified following standard procedures. Purified GST-CI2 fusion proteins were injected into a rabbit and polyclonal antibodies were prepared at Bethyl Laboratories (Montgomery, TX).
Cell Culture and Transfection-HEK tsA201 cells were maintained in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) containing 10% fetal bovine serum (Life Technologies, Inc.) and 1% penicillin/ streptomycin at 37°C in 5% CO 2 . Transient expression of different Ca 2ϩ channel subunits in tsA201 cells was carried out using the calcium phosphate precipitation method (4).
Immunofluorescence Staining-Different combinations of Ca 2ϩ channel subunits were transfected into tsA201 cells, and immunofluorescence staining was performed ϳ36 -48 h after transfection. Prior to staining, cells were washed twice with PBS, followed by fixing in precooled (Ϫ20°C) methanol/acetone (1:1) for 5-10 min at 4°C. After washing with PBS, cells were incubated in labeling buffer (1% bovine serum albumin and 2% normal goat serum in PBS) at room temperature for 1 h to block nonspecific binding. Different primary antibodies were diluted into the labeling buffer and incubated with cells for 1-2 h at room temperature. A secondary antibody, Alexa488-conjugated goat anti-rabbit IgG (Molecular Probes), was used subsequently. Coverslips were mounted onto slides and viewed on a laser-scanning confocal microscope (Zeiss LSM-510).
Immunoprecipitation and Immunoblotting-For co-immunoprecipitation of ␣ 1 and ␤ subunits, whole cell lysates were prepared from transfected tsA201 cells in lysis buffer (50 mM Tris, pH 7.4, 5 mM EDTA, 5 mM EGTA, 0.4 M NaCl, 1% Triton X-100, and 0.1% SDS containing protease inhibitors (Ref. 18). The cell lysates were immunoprecipitated overnight with agitation at 4°C using the CI2 antibody coupled to Protein A-Ultralink resin (Pierce). Immunoprecipitates were washed with the lysis buffer three to five times and eluted with SDS sample buffer. The Card I antibody was used to detect the wild-type (WT) and mutant ␣ 1C subunits on immunoblots, and detection was with horseradish peroxidase-conjugated anti-goat IgG and enhanced chemiluminescence (ECL, Pierce). The ␤ 2 subunits that were co-immunoprecipitated with the ␣ 1C subunits were detected by the anti-␤ GEN antibody and visualized using horseradish peroxidase-conjugated anti-goat IgG and ECL (Pierce).
Intact-cell Radioligand Binding-To detect expression of functional L-type Ca 2ϩ channels, ligand binding experiments were performed using the dihydropyridine (DHP) radioligand [ 3 H]PN200-110 with intact cells transfected with different combinations of the channel subunits. The experimental procedures were as described previously (4). Scatchard analyses were performed to assess B max and K d values.
Electrophysiology -Transiently transfected tsA201 cells expressing different combinations of the Ca 2ϩ channel subunits were used for electrophysiological studies. Approximately 36 -40 h after transfection, Ba 2ϩ currents through L-type calcium channels were measured in the whole cell configuration using the patch clamp technique as described previously (26,27).
Since a functional interaction between the ␣ 1 and ␤ subunits is required for correct membrane targeting of the Ca 2ϩ channel complexes (12), we first tested whether these deletion mutants of the ␣ 1C subunit were able to associate with the ␤ subunits. The wild-type and the mutant ␣ 1C subunits were co-transfected with the rabbit ␤ 2a subunit into tsA cells, and the channel proteins were immunoprecipitated from the transfected cells using an ␣ 1C subunit-specific antibody, CI2. The immunoprecipitated proteins were analyzed using SDS-polyacrylamide gel electrophoresis and immunoblotting. The different ␣ 1C subunits were immunoprecipitated by CI2 and detected with Card I (Fig. 1B, upper panel). The molecular weight differences between the WT and individual mutant ␣ 1C subunits reflected the number of amino acids that were deleted from the protein. The ␤ 2a subunits were co-immunoprecipitated with all different mutant ␣ 1C subunits from co-transfected cells, and detected on the bottom half of the blot with the anti-␤ GEN antibody (Fig.  1B, lower panel). The expression level of different combinations of channel subunits varied somewhat among individual transfections; however, the amount of ␤ 2a subunit present in the immunoprecipitates appeared to roughly correlate with the amount of ␣ 1C subunit in the same immunoprecipitate ( Fig.  1B). This result indicated that deletion of the C-terminal regions did not interfere with ␣ 1 /␤ subunit interaction, suggesting that the overall conformation of the mutant ␣ 1C subunits was not drastically altered by the mutations. In addition, all mutants, with the exception of the ␣ 1C ⌬1623, ␣ 1C ⌬III, and ␣ 1C ⌬V, exhibited voltage-dependent barium currents that were equivalent to, or greater than, those recorded from the WT channels (Table I, Refs. 25 and 26, and data not shown). As we routinely did not observe significant currents from ␣ 1C subunits expressed alone (data not shown), and since ␣ 1C /␤ subunit interactions have been demonstrated to be critical for obtaining functional channels (6), these data further supported the contention that the mutant ␣ 1C proteins studied here interacted normally with ␤ subunits.
The C Terminus of the ␣ 1C Subunit Played a Critical Role in the Plasma Membrane Targeting of Channel Complexes-To examine whether the C-terminal region of the ␣ 1C subunit was involved in targeting of the channel to the plasma membrane, we first asked if the C-terminal truncation mutants of the ␣ 1C subunit were impaired in membrane targeting. We used the rabbit ␤ 2a subunit in membrane targeting studies, as this subunit does not undergo palmitoylation like the rat ␤ 2a subunit and consequently cannot target to the plasma membrane unless it is complexed with an ␣ 1 subunit (11). Different combinations of the mutant ␣ 1C subunits and the rabbit ␤ 2a subunit were co-expressed in tsA201 cells, and the transfected cells were immunostained with the ␣ 1C subunit-specific antibody, CI2, to reveal the expression pattern of the channel. Consistent with the results that we reported previously (4), the WT ␣ 1C subunits were localized almost exclusively at the cell surface when co-expressed with the ␤ subunits in tsA cells ( Fig. 2A), indicating that these proteins were targeted to the plasma membrane. Punctate clusters of channel complexes as detected by the CI2 antibody were observed along the cell surface ( Fig.  2A, indicated by arrows). The localization of the ␤ 2a subunits was examined using the ␤-subunit specific antibody in cotransfected cells, and the expression pattern was similar to that observed with the co-expressed WT ␣ 1C subunits (Fig. 2B) (13). Similarly, confocal images taken from cells expressing the C-terminal truncation mutant ␣ 1C ⌬2024 and ␤ 2a subunits demonstrated that the ␣ 1C ⌬2024 mutant subunits, together with the co-expressed ␤ 2a subunits, were targeted to the plasma membrane and formed punctate clusters (as indicated by arrows in Fig. 2, C and D). Among the WT ␣ 1C ␤ 2a -transfected cells, membrane staining of channel clusters was observed in 90 Ϯ 2% of the cells (Table I). A similar percentage of cells showed punctate membrane staining in the ␣ 1C ⌬2024␤ 2atransfected cells (88 Ϯ 2%, Table I). In addition to the plasma membrane staining as observed in Fig. 2, certain cells expressing the WT ␣ 1C ␤ 2a or ␣ 1C ⌬2024 subunits exhibited a small amount of intracellular staining of the channels, which might have been due to overexpression of the channel proteins in FIG. 1. Expression of the C-terminal deletion mutants of the ␣ 1C subunit and co-immunoprecipitation with the ␤ 2a subunit. A, a schematic map shows the C-terminal deletion mutants of the ␣ 1C subunit with the truncation sites labeled. B, co-immunoprecipitation of the mutant ␣ 1C subunits with the ␤ 2a subunits. Different mutant constructs of the ␣ 1C subunit were co-transfected with the rabbit ␤ 2a subunit into tsA201 cells. Whole cell lysates were prepared from the transfected cells. The channel proteins were immunoprecipitated with the ␣ 1C subunit-specific antibody, CI2, and the immunoprecipitates were separated by SDS-polyacrylamide gel electrophoresis and transferred to nitrocellulose for immunoblotting. The wild-type and mutant ␣ 1C subunits were detected on the blot using the Card I antibody, whereas the co-immunoprecipitated ␤ 2a subunits were detected using the anti-␤ GEN antibody. these cells. However, the total percentage of cells with plasma membrane staining was consistent for each of the individual WT or mutant ␣ 1C subunits analyzed (Table I). These results suggested that deletion of amino acid residues 2025-2171 from the C terminus of the ␣ 1C subunit did not affect membrane targeting of the channel complexes in tsA cells.
To further analyze a potential role for the C terminus, two other deletion mutants of the ␣ 1C subunit, ␣ 1C ⌬1733 and ␣ 1C ⌬1905, were co-expressed with the ␤ 2a subunit in tsA201 cells. Interestingly, the confocal images obtained from the ␣ 1C ⌬1905␤ 2a -or ␣ 1C ⌬1733␤ 2a -transfected cells demonstrated that these more substantial truncations of the C terminus of the ␣ 1C subunit resulted in an altered subcellular distribution pattern (Fig. 3). In contrast to the WT ␣ 1C subunit, significant amounts of the ␣ 1C ⌬1905 and ␣ 1C ⌬1733 mutant subunits were localized intracellularly, as detected by the CI2 antibody, and a smaller number of cells exhibited plasma membrane staining of the mutant ␣ 1C subunits (Fig. 3, A and C). Consistently, the co-expressed ␤ 2a subunits, detected with the anti-␤ 2 antibody, exhibited a localization pattern similar to that observed for the ␣ 1C ⌬1905 and ␣ 1C ⌬1733 mutant subunits (Fig. 3, B and D). The percentages of cells with plasma membrane staining in the ␣ 1C ⌬1905␤ 2a -and ␣ 1C ⌬1733␤ 2a -transfected cells were 69 Ϯ 2% and 40 Ϯ 5%, respectively, both of which were significantly lower than that of the wild-type channels (Table I). A decrease in the number of cells with membrane staining observed with these and other mutant ␣ 1C subunits reflected an impairment of membrane targeting of the channels (Table I and see below). However, the mutant channels that were localized to the plasma membrane were able to form punctate clusters (Fig. 3, indicated by arrows), similar to those observed in the WT ␣ 1C ␤ 2a -transfected cells.
In contrast to these C-terminal deletion mutants, further deletion of the C terminus to residue 1623 resulted in a total loss of membrane localization of the channels (Fig. 4). In cells expressing ␣ 1C ⌬1623␤ 2a , no plasma membrane staining of the mutant ␣ 1C ⌬1623 subunits was observed (Fig. 4A). A membrane outline of the cells was drawn on the phase image, and the same line was superimposed onto the immunostaining image to show lack of plasma membrane staining of the ␣ 1C ⌬1623 subunits (Fig. 4, A and B). The ␤ 2a subunit, when co-expressed with the mutant ␣ 1C ⌬1623 subunit, also failed to target to the plasma membrane (Fig. 4C). The intracellular staining pattern of the ␣ 1C ⌬1623 subunit resembled an endoplasmic reticulumlike network. Taken together, these images indicated that the distribution pattern of the channel complexes was drastically altered upon deletion of additional residues between amino acids 1623-2024 in the C terminus of the ␣ 1C subunit, suggesting that this region played an important role in membrane targeting of the Ca 2ϩ channels.
Functional Effect of a Proline-rich Domain of the ␣ 1C Subunit in Membrane Targeting of the Channels-Based on the above results, it was of particular interest to further narrow down the region within 1623-2024 that was required for membrane targeting of the channels. We divided the region between 1623 and 2024 into three separate domains: domain I, residues 1905-2024; domain II, residues 1733-1905; and domain III, residues 1623-1733. In addition, a proline-rich domain has previously been mapped to residues 1974 -2000 in the C terminus of the ␣ 1C subunit within domain I (26). PRDs have been implicated in mediating protein-protein interactions through binding to proteins containing SH3 domains (28), and the PRD of the ␣ 1C subunit binds to SH3 domains of c-Src, Grb, Hck, and the channel ␤ subunit (26). The functional impact of domains I-III and the PRD in membrane targeting was analyzed in detail as follows.
We first examined whether the PRD of ␣ 1C played any role in membrane targeting of the channels. The PRD-deletion mutant ␣ 1C ⌬PRD and ␤ 2a subunits were co-transfected into tsA201 cells, and immunofluorescence staining obtained with both Card C and CI2 antibodies was analyzed to reveal the subcellular distribution of ␣ 1C ⌬PRD. The ␣ 1C ⌬PRD mutant subunits were localized to the plasma membrane as the Card C and CI2 antibodies detected clear membrane staining (Fig. 5, A and B). A similar localization pattern was observed for the co-expressed ␤ 2a subunits (data not shown). The subcellular localization and the percentage of cells with membrane staining of ␣ 1C ⌬PRD (87 Ϯ 1%) were similar to that obtained for WT ␣ 1C mutants (Table I). Thus, deletion of only the PRD region did not affect targeting of the channels to the plasma membrane. We next examined a larger deletion mutant ␣ 1C ⌬1900 -2026 (⌬I) around the PRD region. The cells expressing the ␣ 1C ⌬I and ␤ 2a subunits were immunostained with both the Card C and CI2 antibodies to detect the distribution pattern of the ␣ 1C ⌬I mutant subunits (Fig. 5, C and D). Confocal images indicated that the ␣ 1C ⌬I mutant subunits were partially localized to the plasma membrane and formed punctate clusters (as marked by arrows; Fig. 5, C and D). However, substantial intracellular staining was observed as well. The percentage cells with membrane staining was 69 Ϯ 2%, which was significantly lower than that of the wild-type channel ( Table I). The staining pattern of ␣ 1C ⌬I was similar to that of ␣ 1C ⌬1905 (Fig. 3). These results suggested that the region surrounding the PRD contributed partially to membrane targeting of the channels.
Identification of a Critical Targeting Domain in the C Terminus of the ␣ 1C Subunit-We next analyzed two other internal deletion mutants of the ␣ 1C subunit. TsA201 cells were transfected with either ␣ 1C ⌬1733-1905 (⌬II) or ␣ 1C ⌬1623-1733 (⌬III) and the ␤ 2a subunit. Subcellular distribution of the ␣ 1C ⌬II subunits was revealed by immunostaining with the Card C and CI2 antibodies (Fig. 6, A and B). Plasma membrane TABLE I Membrane targeting of the C-terminal mutants of the ␣ 1C subunit Different ␣ 1C subunits were co-transfected with the ␤ 2a subunit into tsA201 cells. The transfected cells were subjected to immunostaining, whole cell DHP binding and electrophysiological analyses. At least three independent transfections were performed for the immunostaining study. Approximately 200 -300 cells were randomly scored for the membrane localization of the channels from each transfection. Statistical analysis was performed to compare the percentage of cells with plasma membrane staining (WT vs. individual mutant channels, and a "*" indicates p Ͻ 0.01). Two to three independent intact-cell DHP binding and patch-clamp current recording experiments were performed, except for the ␣ 1C ⌬V␤ 2a -and ␣ 1C ⌬IV␤ 2a -transfected cells (N.D., not determined). A "ϩ" indicates that saturable DHP binding or L-type Ba 2ϩ currents were observed in the corresponding transfected cells, while a "No" indicates no membrane targeting, saturable DHP binding, or no detectable L-type currents. Similar B max and K d values were observed in cells that bound DHP (data not shown). Expression of functional L-type Ba 2ϩ currents with peak current amplitude of Ն5 pA/pF is scored as "ϩ." staining was observed in the ␣ 1C ⌬II␤ 2a cells (as indicated by arrows in Fig. 6, A and B); however, intracellular staining was seen as well. The localization pattern of the co-expressed ␤ 2a subunits was similar as that of ␣ 1C ⌬II (data not shown). The percentage of cells with membrane staining was 79 Ϯ 2% (Table I). These results suggested that deletion of the region between residues 1734 and 1905 only partially affected membrane targeting of the channels. In contrast, deletion of residues 1623-1733 (domain III) drastically disrupted channel targeting (Fig. 6C). Confocal images obtained from cells expressing ␣ 1C ⌬III␤ 2a showed a complete loss of cell surface staining of the mutant channels (Fig. 6, C and D). The local-ization pattern of ␣ 1C ⌬III was similar to that of ␣ 1C ⌬1623; both mutants exhibited a total disruption of membrane targeting of the channels. Taken together, these results suggested that the region between 1623 and 1733 (domain III) of the ␣ 1C subunit was critical for targeting the channels to the plasma membrane. Interestingly, the domain III identified here to be important for membrane targeting has previously been shown to contain an IQ-like calmodulin binding motif (amino acids 1654 -1665) that has been implicated in Ca 2ϩ -dependent inactivation and facilitation of cardiac L-type channels (20 -23). To further narrow down the targeting domain, we created two additional   A and B) or the ␣ 1C -⌬1733␤ 2a subunits (C and D) were analyzed using immunofluorescence staining and confocal microscopy. The CI2 (A and C) and the anti-␤ 2 antibodies (B and D) were used to reveal the expression pattern of the mutant ␣ 1C and the ␤ 2a subunits, respectively. Punctate staining at the plasma membrane was indicated by 3. Increased intracellular staining was observed for the mutant channels. mutants, ␣ 1C ⌬IV (1668 -1733) and ␣ 1C ⌬V (1623-1666), to delete the C-or N-terminal portion of the domain III, respectively. The ␣ 1C ⌬V mutant lacked the IQ motif and 31 amino acids upstream. These two deletion mutants co-immunoprecipitated with the ␤ 2a subunit to a similar extent as the WT and all other mutant ␣ 1C subunits tested (Fig. 1B). To reveal the subcellular localization of the ␣ 1C ⌬IV and ␣ 1C ⌬V mutant subunits, the cells expressing ␣ 1C ⌬IV␤ 2a and ␣ 1C ⌬V␤ 2a were stained with the CI2 antibody, and the localization pattern of the co-expressed ␤ 2a subunits was detected using the anti-␤ 2 antibody (Fig. 7). Plasma membrane staining of the ␣ 1C ⌬IV subunit and the co-expressed ␤ 2a subunit was observed (Fig. 7,  A and C, as marked by arrows); however, a significant portion of the ␣ 1C ⌬IV and the ␤ 2a subunits were localized intracellularly. The percentage of cells with plasma membrane staining was decreased (72 Ϯ 3.5%, Table I) compared with that of the WT channels, but was similar to that of the ␣ 1C ⌬II mutant subunit. In contrast, only a very small percentage of cells expressing ␣ 1C ⌬V␤ 2a showed any plasma membrane staining (10.5 Ϯ 1.3%, Table I) and majority of the ␣ 1C ⌬V and ␤ 2a subunits were localized intracellularly (Fig. 7, C and D). These results indicated that domain V in the C terminus of the ␣ 1C subunit was critical for membrane targeting of the channels. Taken together, these results indicate that a "targeting domain" exists within the 43 amino acid stretch from 1623 to 1666 in the C terminus. In addition, deletion of other regions in the C terminus, including the region surrounding the PRD and domains II and IV, partially disrupted membrane targeting of the channels. Interestingly, the effect of domains I and II on membrane targeting appeared to be additive, as the percentage of cells that showed membrane staining for ␣ 1C ⌬1733 (lacking both domain I and II) was lower than that of ␣ 1C ⌬1905 (lacking only domain I) (Table I).
Functional Effect of Membrane Targeting of the L-type Channels-To examine the functional importance of membrane targeting of the L-type Ca 2ϩ channels, we performed intact-cell radioligand binding experiments using [ 3 H]PN200-110, a DHP antagonist. The mutant ␣ 1C subunits whose membrane targeting was completely disrupted, including ␣ 1C ⌬1623 and ␣ 1C ⌬III, exhibited no saturable binding (Table I). In cells expressing other mutant ␣ 1C subunits and the ␤ 2a subunit, DHP binding was observed and was comparable for the WT and mutant channels (Table I). Although some mutant ␣ 1C subunits showed decreased membrane staining compared with the WT channels, and concomitantly, increased intracellular staining (see above and Table I (A and B) or ␣ 1C ⌬1623-1733(⌬III)␤ 2a (C and D) were analyzed using immunostaining and confocal microscopy. The image shown in A was taken from cells stained with Card C, whereas the images shown in B and C was taken from cells stained with CI2. Punctate staining at the plasma membrane was indicated by arrows. In the phase-contrast image (D), an outline of cell plasma membrane was marked. Note lack of membrane staining of ␣ 1C ⌬III was observed along cell surface (C).  D) were analyzed using immunostaining and confocal microscopy. The images shown in A and C were taken from cells stained with CI2, whereas the images shown in B and D were taken from cells stained with the anti-␤ 2 antibody. Punctate staining at the plasma membrane was indicated by arrows. Little or no plasma membrane staining was observed for the ␣ 1C ⌬V␤ 2a channels (C and D; Table I).
resulted in loss of functional channels. For the WT and all other mutants, L-type Ba 2ϩ currents through these channels were recorded (with a mean peak current amplitude Ն 5 pA/pF, n Ն 6) from the transfected tsA cells (Table I). Taken together, these results indicated that membrane targeting was a necessary step for the formation of the functional channels at the cell surface, as currents were not observed with the ␣ 1C mutants that completely failed to target to the plasma membrane. DISCUSSION Correct targeting and translocation of multisubunit ion channels to the plasma membrane of cells is an important process that allows for the proper formation of functional channels at the cell surface. The mechanisms that allow for membrane targeting of ion channels have been investigated extensively in many recent studies (4, 12, 15, 29 -32). The assembly of different channel subunits into complexes at the early stage of protein synthesis appears to be a necessary step for proper membrane targeting of shaker K channels, G-protein activated inward rectifier K channels, and K ATP channels (29,30). This is also likely to be the case for L-type Ca 2ϩ channels (4,12), although this has not been explicitly demonstrated. However, we have previously demonstrated that a functional interaction between the ␣ 1 and ␤ subunits is a critical factor for membrane targeting of L-type Ca 2ϩ channels (12), and it is likely that complex formation occurs at the early stage of protein synthesis. Recent studies have demonstrated that ␤ subunits can neutralize an endoplasmic reticulum retention signal in the I-II loop of ␣ 1 subunits (15). Nevertheless, signals other than complex formation are required for the correct targeting of the Ltype Ca 2ϩ channels, as several of the ␣ 1C mutants studied here were able to complex with ␤ subunits, yet they were impaired in membrane targeting.
In the present study, we further analyzed the mechanism of membrane targeting of the L-type calcium channels. To address the question whether the C-terminal domain of the ␣ 1C subunit plays a role in membrane targeting, we created several C-terminal deletion mutants of the ␣ 1C subunit, and immunocytochemical experiments were performed to study the subcellular localization of the mutants. Based on the results described above, several regions in the C terminus of the ␣ 1C subunit appeared to be involved in membrane targeting of the channels to different degrees. The most important region has been mapped to residues 1623-1666, as deletion of this small targeting domain resulted in a total loss of membrane localization of the channels. The other regions (domains I, II, and IV) in the C terminus of the ␣ 1C subunit contributed partially to membrane targeting of the channels, as deletion of either of these domains resulted in a reduction of membrane-localized channels and an accumulation of intracellular channels. The results shown here provided the first explanation for failure of the previously identified "dead" mutant of the ␣ 1C subunit, ␣ 1C ⌬1623 (24) to function as a Ca 2ϩ channel, as we demonstrated here that this mutant subunit lacked the ability to target to the plasma membrane. In addition, two other dead mutants of the ␣ 1C subunit (␣ 1C ⌬V and ␣ 1C ⌬III) identified here further demonstrated the importance of C-terminal domains for proper membrane targeting to the function of L-type Ca 2ϩ channels.
There are several possible explanations for the requirement for the C-terminal domains of the ␣ 1C subunit in membrane targeting of the channels. First, the C-terminal targeting domains may be critical for protein folding and proper maturation of the ␣ 1C subunit. Previous studies have identified several deletion or truncation mutants of the cystic fibrosis transmembrane conductance regulator that lack the ability to target to the plasma membrane (33). It has been suggested that the mutant CFTRs are misfolded proteins that are degraded more rapidly than the WT proteins (33). Deletion of the C-terminal domains of the ␣ 1C subunit may alter protein conformation and result in unstable proteins that are unable to target to, or be maintained at, the plasma membrane. However, since all the ␣ 1C mutant subunits were able to associate with the ␤ 2a subunit (Fig. 1B) and be recognized by several ␣ 1C -specific antibodies, the overall conformation of the mutant ␣ 1C subunits was maintained to a certain degree.
Interestingly, although the membrane targeting of some mutant ␣ 1C subunits was partially impaired (e.g. ␣ 1C ⌬1905 or ␣ 1C ⌬1733), as less plasma membrane staining and significant intracellular staining of the channels was observed (Fig. 3), similar amounts of DHP binding were obtained for these mutants and the wild type ␣ 1C subunit (see above). This result suggested that the plasma membrane-localized, as well as the intracellularly localized, forms of these particular mutants were capable of binding the hydrophobic radioligand PN200-110. A clear parallel existed in that the mutants that could exhibit membrane targeting, also could bind DHPs and exhibit Ba 2ϩ currents, whereas those mutants that were completely unable to target to the plasma membrane were also unable to bind DHPs and mediate Ba 2ϩ currents. This suggested that there were fundamental differences in the properties of the mutants. A conceivable explanation for the significant intracellular staining observed for the "functional mutants," such as ␣ 1C ⌬1905 or ␣ 1C ⌬1733 that were also able to exhibit significant punctate staining at the plasma membrane, is that these mutants were able to target to the plasma membrane but had a diminished ability to be retained at the plasma membrane. Thus, at any given time, these mutants might exhibit significant intracellular staining due to trafficking of the channels to or from intracellular compartments. Such mutant channels would be expected to exhibit full function (DHP binding and production of currents), whereas the mutant channels that were incapable of any membrane targeting also lacked all function. Thus, the C terminus of the ␣ 1C subunit may be involved in constitutive recycling of the protein between the plasma membrane and intracellular compartments. The WT channels may be more stable at the plasma membrane and recycle at a slower rate, whereas the mutant channels may have recycled more rapidly and more intracellular localized channels were observed at a given time point. These mutants were functionally capable of making L-type channels and binding DHPs, but may have differed from WT in membrane retention. It would be of interest to learn if the dead channels that were incapable of membrane targeting might actually form functional channels if they were able to target to the plasma membrane.
Finally, the C-terminal domain may be responsible for binding to other proteins, and this association may participate in the processing and the localization of the channels to the plasma membrane. Interestingly, the targeting domain V identified here contains the "IQ" calmodulin binding site, which is located between positions 1654 and 1665 (20 -23). Thus, it is possible that an interaction between the ␣ 1C subunit and calmodulin may play a role in membrane targeting of the Ca 2ϩ channels. However, deletion of the IQ motif in a splice variant of the ␣ 1C subunit had no deleterious effect on peak currents when expressed in Xenopus oocytes (20). In several such mutants lacking the IQ motif, peak currents were greater than or equivalent to currents from wild type channels in Xenopus oocytes (20), suggesting that these mutants did not have targeting defects. In addition, neutralization of the basic residues within the IQ motif led to a loss of calmodulin binding but did not result in reduced currents in Xenopus oocytes (22). Taken together, these results suggest that the IQ motif does not play a role in membrane targeting. However, further studies are necessary to test the role of the IQ motif and upstream residues in domain V in channel targeting in mammalian cells, and to test whether the disruption of the interaction between the ␣ 1C subunit and calmodulin affects the localization of the L-type channels in mammalian cells. In addition, since the protein sequence within the targeting domain "V" of the ␣ 1C subunit is homologous among all other ␣ 1 subunits, this targeting domain identified here may be important for membrane targeting of other types of Ca 2ϩ channels as well.
In summary, we have identified domains in the C terminus of ␣ 1C subunit that are critical for membrane targeting of the L-type Ca 2ϩ channels, and targeting of the channels to the plasma membrane is required to obtain functional Ca 2ϩ channels.