Smooth Muscle-selective Alternatively Spliced Exon Generates Functional Variation in Cav1.2 Calcium Channels*

Voltage-gated calcium channels play a major role in many important processes including muscle contraction, neurotransmission, excitation-transcription coupling, and hormone secretion. To date, 10 calcium channel α1-subunits have been reported, of which four code for L-type calcium channels. In our previous work, we uncovered by transcript-scanning the presence of 19 alternatively spliced exons in the L-type Cav1.2 α1-subunit. Here, we report the smooth muscle-selective expression of alternatively spliced exon 9* in Cav1.2 channels found on arterial smooth muscle. Specific polyclonal antibody against exon 9* localized the intense expression of 9*-containing Cav1.2 channels on the smooth muscle wall of arteries, but the expression on cardiac muscle was low. Whole-cell patch clamp recordings of the 9*-containing Cav1.2 channels in HEK293 cells demonstrated -9 and -11-mV hyperpolarized shift in voltage-dependent activation and current-voltage relationships, respectively. The steady-state inactivation property and sensitivity to blockade by nifedipine of the ±exon 9* splice variants were, however, not significantly different. Such cell-selective expression of an alternatively spliced exon strongly indicates the customization and fine tuning of calcium channel functions through alternative splicing of the pore-forming α1-subunit. The generation of proteomic variations by alternative splicing of the calcium channel Cav1.2 α1-subunit can potentially provide a flexible mechanism for muscle or neuronal cells to respond to various physiological signals or to diseases.

␣ 1 -subunits characterized constitute the family of L-type Ca 2ϩ channels, namely Ca v 1.1, Ca v 1.2, Ca v 1.3, and Ca v 1.4.
The Ca v 1.2 Ca 2ϩ channels are distinguished by their high sensitivity to antagonists such as dihydropyridines (DHPs), 1 phenylalkylamines, and benzothiazepines and their distinctive slow inactivating I Ba currents (1). However, in response to a cytoplasmic rise in Ca 2ϩ concentration, the Ca v 1.2 L-type channels exhibit two opposing channel properties. On the one hand, the L-type channels inactivate robustly with the rise in local [Ca 2ϩ ] i , and on the other, the channels also facilitate to enhance Ca 2ϩ entry. Both these Ca 2ϩ regulatory processes are mediated by calmodulin that acts as the Ca 2ϩ sensor (2)(3)(4).
The DHP antagonists inhibit L-type calcium channels and have been used to manage a variety of cardiovascular disorders including hypertension and angina pectoris. These drugs bind with high affinity to the DHP binding site, which is composed of transmembrane segments, IS6, IIIS5, IIIS6, and IVS6 of the ␣ 1 1.2 subunit (5,6). The human ␣ 1 1.2 gene, CACNA1C, contains 55 exons, of which 19 exons are subjected to alternative splicing (7,8). The IS6 segment is encoded by mutually exclusive exons 8/8a. The inclusion of 8 instead of 8a to form IS6 conferred a higher sensitivity to DHPs (5). The exons encoding IIIS5, IIIS6, and IVS6 are, however, constitutive exons, not subjected to alternative splicing.
Other investigators employed cDNA library screening and analyses of the ␣ 1 1.2 subunit and identified the "smooth muscle" and "cardiac muscle" ␣ 1 1.2 splice combinations as differing in four regions: (i) exon 1/1a; (ii) exon 8/8a; (iii) Ϯ75 nucleotides in the I-II loop (termed exon 9* in this report); and (iv) exon 31/32. The cardiac muscle form was proposed to consist of splice combination 1a/8a/Ϫ9*/31, whereas the proposed smooth muscle splice combination was 1/8/ϩ9*/32 (9 -11). Whereas exon 8/8a determines DHP sensitivity, exon 1a containing Ca v 1.2 channels in cardiac muscle could be regulated indirectly by protein kinase C to relieve the inhibition of the channel by exon 1a (12,13). The "smooth muscle" form of Ca v 1.2 channels contains exon 8 and therefore exhibits larger inhibition by DHP (5). It is noteworthy that the cellular localization and electrophysiological and pharmacological properties of 9*-containing Ca v 1.2 channels are largely unknown. In a separate study, we did not detect any difference in the electrophysiological properties of Ca v 1.2 channels containing either exon 31 or 32 (8).
Here, we determined the cell-selective expression of exon 9* by (i) amplifying total RNA in RT-PCR experiments to enumerate the presence or absence of exon 9* in cardiac or aortic smooth muscles and (ii) raising specific polyclonal antibody against exon 9* to identify the restricted protein localization of 9*-containing Ca v 1.2 channels on arterial smooth muscle. Whole-cell electrophysiological characterization of the Ϯ9*containing Ca v 1.2 channels revealed altered electrophysiological properties, but the sensitivity to nifedipine blockade was similar. Such expression of an alternatively spliced exon in smooth muscle would contribute to the fine tuning of Ca v 1.2 channel function and influence smooth muscle contractility.
Determination of Relative expression of Ϯ9*-Ca v 1.2 Transcripts by RT-PCR-Total RNA was extracted from adult male (7-8 weeks) Wistar rat tissues or A7r5 cell line (ATCC, Manassas, VA) using the RNeasy kit (Qiagen). Reverse transcription was carried out with SUPERSCRIPT II (Invitrogen). Negative control (reactions without reverse transcriptase) was carried out in all RT-PCRs to exclude contamination. The primer pair used was as follows: (i) rat ␣ 1 1.2F (5Ј-GTCATCATCATCTATGCCATT-3Ј) and (ii) rat ␣ 1 1.2R (5Ј-GCGGCT-GAACTTGGATT-3Ј). The PCR protocol includes a denaturation step at 95°C for 5 min; 35 cycles of 95°C for 30 s, 50°C for 1 min, and 72°C for 1 min; and a final extension step at 72°C for 10 min.
Determination of Tissue Specificity of pAb9* Antibody by Western Blot-Tissues from rat aorta, brain, and heart were homogenized in lysis buffer containing 50 mM Tris, pH 8, 1 mM EDTA, and 150 mM NaCl. Due to the small quantity of rat aorta, six or seven aortas were pooled together for protein extraction. After centrifugation at 8,000 rpm for 15 min, followed by 40,000 rpm for 1 h, membrane proteins were extracted from the pellet with lysis buffer supplemented with 1% Triton X-100 for 1 h. Following centrifugation at 40,000 rpm for 1 h, 30 g of proteins in the supernatant were separated in 6% SDS-polyacrylamide gel. The protein was then transferred onto polyvinylidene difluoride membrane (Bio-Rad) using a tank transfer system (Bio-Rad). After blocking the membrane with 5% skim milk in TBS-T (20 mM Tris, pH 7.6, 137 mM NaCl, and 0.05% Tween 20), diluted primary pAb9* antibody (1:200) was added, and membrane was incubated for 2 h. A goat anti-rabbit secondary antibody (Sigma) was added, and the specific binding of the primary antibody was probed using the ECL system according to the company's protocol (catalog no. RPN2132; Amersham Biosciences).
Immunohistochemical Localization of Ϯ9*-containing Ca v 1.2 Channels-Tissues from adult male Wistar rats were cryosectioned at 8 m in thickness. The sections were fixed with 4% paraformaldehyde (pH 7) for 20 min. After washing with 0.1% Triton/phosphate-buffered saline, 4% goat serum was used for blocking over 1 h. The samples were then incubated with primary antibodies (1:100 dilution) overnight at 4°C. The primary antibodies used were either pAb9* or commercial antibod-ies such as (i) anti-␣ 1 C (Alomone Laboratories), recognizing the intracellular loop between domains II and III of Ca v 1.2 calcium channel, (ii) anti-␣ 1 D (Alomone Laboratories), and (iii) anti-␣-smooth muscle actin (Chemicon). After washing with 0.1% Triton/phosphate-buffered saline, goat anti-rabbit fluorescein isothiocyanate-conjugated or goat antimouse Texas Red-conjugated secondary antibody was added to the samples. The nuclei were labeled with propidium iodide, and the immunostainings were visualized using a laser-scanning confocal microscope (Fluoview BX61; Olympus).
To determine the specificity of staining of the pAb9* antibody to 9*-containing ␣ 1 1.2 channels, pAb9* was used either (i) after preabsorption with GST only or (ii) after preabsorption with GST followed by a second round of preabsorption with ␣ 1 1.2-GST9* (15).
Electrophysiological Recordings and Data Analysis-HEK293 cells were transiently transfected with ␣ 1 C77-WT or ␣ 1 C77-9* (1.25 g), ␤ 2a (1.25 g), and ␣ 2␦ (1.25 g) using the standard calcium phosphate transfection method (8). The ␤ 2a and ␣ 2␦ clones were provided by Dr. Terry Snutch (University of British Columbia). I Ba was recorded at room temperature using the whole-cell patch clamp technique, 48 -72 h after transfection. The external solution contained 10 mM HEPES, 140 mM tetraethylammonium methanesulfonate, 5 mM BaCl 2 (pH was adjusted to 7.4 with CsOH and osmolarity to 290 -310 with glucose). The internal solution (pipette solution) contained 138 mM Cs-MeSO 3 , 5 mM CsCl, 0.5 mM EGTA, 10 mM HEPES, 1 mM MgCl 2 , 2 mg/ml Mg-ATP, pH 7.3 (adjusted with CsOH). The osmolarities of solutions used were adjusted to between 290 and 300 mosM with glucose. The voltages are uncorrected for a Ϫ11-mV junction potential, and actual voltage can be obtained by subtracting 11 mV from the reported values. Whole-cell currents, obtained under voltage clamp with an Axopatch 200B amplifier (Axon Instruments), were filtered at 1-5 kHz and sampled at 5-50 kHz, and the series resistance was typically Ͻ5 megaohms after Ͼ70% compensation. The P/4 protocol was used to subtract online the leak and capacitive transients.
To determine the whole-cell I-V relationships, currents were recorded by holding the cell at Ϫ90 mV before stepping to various potentials from Ϫ50 mV to 50 mV (⌬V ϭ 10 mV) over 900 ms. In each cell, I Ba at all voltages was normalized to the peak current. The steady-state inactivation curves were obtained from experiments by stepping from a holding potential of Ϫ90 mV to a family of 15-s-long prepulses, followed by a 104-ms test pulse to ϩ10 mV. Each steady-state inactivation was normalized to the maximal current amplitude for comparison. G-V curves were obtained from a tail activation protocol. The cells were activated by a variable voltage family of 20-ms test pulses, and tail currents were measured after repolarization to Ϫ50 mV. Nifedipine (Sigma) was dissolved in Me 2 SO to make a 10 mM stock solution. Nifedipine at various concentrations was applied to the HEK293 cells at a test pulse of 0 mV. Inhibition ratios were calculated by comparing the currents before and after drug treatment. Values are expressed as mean Ϯ S.E. GraphPad Prism software was used for data plotting and statistical analysis. Statistical significance of differences between means was calculated with Student's t test.

RESULTS
We employed the transcript-scanning method (16 -18) to systematically search for all possible splice variations in the human Ca v 1.2 ␣ 1 -subunit in heart and brain (8). Of the 19 splice exons reported by us, exon 9* has a low expression in human adult heart. However, there are reports that exon 9* belongs to the "smooth muscle" ␣ 1 C b isoform of Ca v 1.2 channels that also included alternatively spliced exons 1, 8, and 32 that are different from the "cardiac" ␣ 1 C a isoform (9,11). Most of the reported data relied on the identification of exon 9* in a small number of isolated cDNA clones derived from tissues that presumably contained a heterogeneous mixture of cell types. As a result, exon 9* has been reported in rabbit lung, rabbit vascular and gastrointestinal smooth muscle, mouse brain, and A7r5 smooth muscle cell line (9,19,20). To resolve the notion that exon 9* may have selective expression on smooth muscle cells, we decided to first validate the molecular biological data by probing for exon 9* expression by RT-PCR in a cell line of smooth muscle origin and in fresh rat aorta and heart tissues. Next, we raised polyclonal antibody against exon 9* to determine the protein localization of 9*-containing Ca v 1.2 channels on smooth or cardiac muscles and on neurons.
Selective Expression of Exon 9*-containing ␣ 1 1.2 Channel Transcript in Smooth Muscle- Fig. 1A showed two PCR products when RT-PCR experiments were performed on total RNA isolated from rat aorta cell line A7r5 and tissues from adult rat heart and aorta. The upper band is 755 bp, whereas the lower band is 680 bp in length. DNA sequencing and analyses of both PCR products revealed that the larger PCR product contained an additional 75-nucleotide sequence that encodes exon 9* (Gen-Bank TM accession number AY323810). Delineation of the expression of exon 9* in the heart showed that all regions (left and right ventricles; left and right atriums) exhibited low levels of exon 9* with a similar expression pattern (Fig. 1A, lanes 2-5). The ␣ 1 1.2 subunit transcripts in aorta and A7r5 cell line, however, expressed high levels of exon 9* (Fig. 1A, lanes 1 and 6). These results provide strong evidence that the 9*-containing ␣ 1 1.2 transcript is expressed selectively at a high level in aortic smooth muscle but is expressed at a much lower level in cardiac muscle.
Specificity of pAb9* Antibody-To further determine the cellselective protein expression of exon 9*, we decided to raise a polyclonal antibody (pAb9*), against the 25-amino acid polypeptide encoded by exon 9*. We found a paralogous exon 9* reported in ␣ 1 1.3 in chick cochlea (21) but did not find any other paralogous sequences in other calcium channels after a search in the public data base. From a rat cochlea cDNA library, we isolated a similar ␣ 1 1.3 subunit exon 9* flanked by exons 9, 10, 12, 13, and 14. A comparison of the amino acid sequences of rat ␣ 1 1.2 exon 9* with the rat ␣ 1 1.3 exon 9* revealed a 36% amino acid sequence identity (Fig. 1B). To investigate whether the affinity-purified, GST-preabsorbed pAb9* may cross-react with ␣ 1 1.3 exon 9* polypeptide, we expressed exons 9 -14 (including 9*) of ␣ 1 1.3 as His-tagged fusion protein. The immunoblot demonstrated clearly that only ␣ 1 1.2-GST9* was positively stained, whereas exon 9* of the ␣ 1 1.3 fusion protein was not stained ( Fig. 2A). This result indicated that pAb9* did not cross-react with ␣ 1 1.3-exon 9* or with the adjacent amino acids within the I-II loop. To further demonstrate the specificity of pAb9*, Western blot analysis on rat aorta, heart, and brain , and brain (lane 3) displayed strong staining by pAb9* of aorta lysate to reveal 9*-containing Ca v 1.2 channel proteins (ϳ220 kDa) in a 6% SDS-polyacrylamide gel. Similar amounts of protein (30 g) were loaded and stained by Coomassie Blue (lane 4, aorta; lane 5, brain; lane 6, heart). C, immunofluorescent staining of ␣ 1 C77-9*-or ␣ 1 C77-WT-transfected HEK293 cells with pAb9* and anti-␣ 1 C antibodies showed membrane localization of the Ca v 1.2 channels. Inset images displayed staining of single HEK293 cells by either pAb9* or anti-␣ 1 C antibodies. protein lysates again displayed the highly selective expression of the ϳ220-kDa 9*-containing Ca v 1.2 channel protein on aorta (Fig. 2B), which mirrored the transcript expression pattern (Fig. 1A). It is, however, possible that the 9*-containing Ca v 1.2 channels were expressed at such low levels in heart and brain and were undetectable, although equal amounts of proteins were used (Fig. 2B). Since total protein lysates contain both the Ϯ9*-containing Ca v 1.2 channels, we decided to perform another test to determine pAb9* specificity on expressed Ϯ9*containing Ca v 1.2 cDNA clones in HEK293 cells. Confocal microscopy was employed to demonstrate that pAb9* specifically immunolabeled ␣ 1 C77-9*-transfected but not ␣ 1 C77-WT-transfected HEK293 cells (Fig. 2C, left panel). When the Anti-␣ 1 C antibody was used, strong staining of both the ␣ 1 C77-WT-and ␣ 1 C77-9*-transfected HEK293 cells was indicative of their expression on the cell membrane (Fig. 2C, right panel).
Selective Localization of 9*-containing Ca v 1.2 Channels on Arterial Smooth Muscles-Strong immunohistochemical stainings by the GST-preabsorbed pAb9* revealed the cellular localization of 9*-containing Ca v 1.2 channels on the walls of isolated arteries from the brain (Fig. 3, A and B). The co-localization of the pAb9* and ␣-smooth muscle actin signals provided convincing evidence that the 9*-␣ 1 1.2 channels were expressed on the smooth muscles of the arterial wall (Fig. 3A). To further characterize the expression of 9*-␣ 1 1.2 subunit on the arterial smooth muscle, serial sections of an artery on the surface of brain were used for the following experiments (Fig. 3B, a-f). The pAb9* stained strongly to detect ␣ 1 1.2 subunit protein expression on the arterial smooth muscles (Fig. 3B, a), whereas the pan-␣ 1 C antibody, anti-␣ 1 C, also produced similar patterns on smooth muscles for the Ca v 1.2 channels (Fig. 3B, b). The lack of co-localization of the propidium iodide signal that stained the nuclei (PI; red) with pAb9* or Anti-␣ 1 C staining (fluorescein isothiocyanate; green) would agree with the known transmembrane expression of Ca v 1.2 calcium channels on the plasma membrane. The pan-␣ 1 D antibody, anti-␣ 1 D, did not stain the smooth muscle, which, together with our data presented earlier, clearly argues against the possibility of crossreactivity to the ␣ 1 1.3 subunit in the immunohistochemical experiments (Fig. 3B, c). Nonetheless, anti-␣ 1 D was able to stain HEK293 cells transfected with a rat ␣ 1 D clone (data not shown). The specificity of pAb9* labeling of 9*-containing Ca v 1.2 channels on smooth muscle was again validated, since there was no staining with preimmune serum or with pAb9* that has been preabsorbed with ␣ 1 1.2-GST9* polypeptide (Fig.  3B, d and e). It should be noted that the internal elastic lamina could either be stained nonspecifically, or the signal was due to autofluorescence. Taken together, we conclude that the pAb9* antibody that we raised stained exon 9*-containing Ca v 1.2 channels specifically.
Localization of 9*-containing Ca v 1.2 Channels on Cardiac Muscle and Cerebellar Neurons-The anti-␣ 1 C antibody stained both the blood vessel and the surrounding cardiac myocytes with similar intensity in heart tissues (Supplementary Fig. S1A, lower panel). On the other hand, pAb9* produced a restricted and robust expression of 9*-containing Ca v 1.2 channels on the smooth muscle wall of the blood vessels, but in contrast, the cardiac muscle cells showed light staining (Supplementary Fig. S1A, upper panel). This finding confirmed the low level of exon 9* transcripts in heart that was revealed by RT-PCR (Fig. 1A), since the isolated heart total RNA would contain a mixture of mRNAs from cardiac and smooth muscle cells. This contamination by smooth muscle 9*-containing Ca v 1.2 transcripts would misrepresent the actual exon 9* transcript expression level in cardiac muscle.
Catterall et al. (15) reported the presence of Ca v 1.2 protein on neuronal cell bodies and proximal dendrites of dentate gyrus and hippocampus. In the cerebellum, Ca v 1.2 mRNA was reported to be expressed in the granular layer but not in Purkinje cells or the molecular layer (22,23). The data agree with our observation that the anti-␣ 1 C antibody stained the neurons in the granule layer of the cerebellum ( Supplementary Fig. S1B, lower panel), but there was a lack of staining on the Purkinje cells and the molecular layer. The overall expression pattern of Ca v 1.2 channels on the granule layer is therefore consistent with published results (22,23). Although anti-␣ 1 C stained the granular layer, the pAb9* antibody did not detect any expression of 9*-containing Ca v 1.2 channels on the neurons of the granular layer ( Supplementary Fig. S1B, upper panel). This suggests the absence of exon 9* in ␣ 1 1.2 proteins expressed on cerebellar neurons. As such, the isolation of brain ␣ 1 1.2 cDNAs containing exon 9* reported by Ma et al. (19), would probably be due to the presence of 9*-containing Ca v 1.2 transcripts from arterial smooth muscle mixed into extracted brain total RNA. Our results, taken together, pointed to a highly selective expression of exon 9*-containing Ca v 1.2 channels on arterial smooth muscle.
Electrophysiological Characterization of ␣ 1 C77-9* Revealed Hyperpolarized Shifts in I-V Relationship and Voltage-dependent Activation-So far, the characterization of the biophysical properties of the inclusion of exon 9* in the ␣ 1 1.2 subunit has been performed in the context of the "smooth muscle" ␣ 1 C b clone (9). The "smooth muscle" ␣ 1 C b clone, besides the inclusion of exon 9*, also contained 3 other splice variations in exons 1,

FIG. 3. Localization of 9*-containing Ca v 1.2 calcium channels on arteries.
A, an artery isolated from brain was co-labeled with pAb9* (green) and anti-␣-smooth muscle actin (␣-SM Actin; red). Colocalization of the 9*-containing Ca v 1.2 channels and smooth muscle ␣-actin was demonstrated in the merged image (orange). Bars, 20 m. B, serial sections of an artery on the surface of rat brain were immunostained with one of the following antibodies: pAb9* (a), anti-␣ 1 C antibody (b), anti-␣ 1 D antibody (c), preimmune serum (d), and pAb9* preabsorbed with ␣ 1 1.2 GST9* polypeptide (e). A phase-contrast image was shown in f. Autofluorescence of the arterial wall (green) clearly defines the internal elastic lamina. Nuclei were labeled with propidium iodide (red). Bars, 50 m. 8, and 32. This is in contrast to the exonic splice combination of 1a, 8a, Ϫ9*, and 31 of the "cardiac muscle" ␣ 1 C a clone (9,11). Exon 8/8a has been demonstrated to determine DHP sensitivity (5), and the protein kinase C effect disinhibits the exon 1a containing channel indirectly. However, the electrophysiological properties of Ca v 1.2 channels arising from the inclusion of exon 9* have not been fully investigated. We therefore performed whole-cell patch clamp recordings of ␣ 1 C77-WT (Ca v 1.2 wild type) and ␣ 1 C77-9* (9*-containing Ca v 1.2). The genetic make-up of ␣ 1 C77-WT is exons 1-20, 22-30, 32-44, and 46 -50, whereas ␣ 1 C77-9* contained the additional 9* exon. The current-voltage (I-V) relationships showed a large hyperpolarized shift of Ϫ11 mV in the V 0.5 of the ␣ 1 C77-9* channels (Fig. 4A, Table I, p Ͻ 0.01). The steady-state inactivation properties were not significantly different whether exon 9* was included or excluded in Ca v 1.2 channels (Fig. 4C, Table I, p Ͼ 0.05). However, there was a difference in the voltage-dependent activation properties between the ␣ 1 C77-WT and ␣ 1 C77-9* with the V act0.5 shifted in the hyperpolarized direction by ϳϪ9 mV for the ␣ 1 C77-9* channels (Fig. 4B, Table I, p Ͻ 0.05). On the whole, the presence of exon 9* hyperpolarized shifts of the I-V curve and voltage-dependent activation of Ca v 1.2 channels but did not affect their steady-state inactivation property.
Inclusion of Exon 9* Did Not Alter DHP Sensitivity to the Ca v 1.2 Channels-It has been reported that Ca v 1.2 channels on smooth muscles are more sensitive to DHP blockade than cardiac Ca v 1.2 channels. The differences in utilization of splice exons in the two variants, ␣ 1 C a and ␣ 1 C b , has been postulated to account for the differences in sensitivity to the organic antagonists observed in heterologous expression systems (5). Whereas the DHP binding site of the "smooth muscle" ␣ 1 C b clone displayed higher DHP sensitivity, it was not clear whether the presence of exon 9* in the ␣ 1 C b variant may contribute to the enhanced sensitivity of ␣ 1 C b channels to block by DHPs. By using three different concentrations of nifedipine, we did not detect any difference in the level of channel inhibition at Ϫ90 mV holding potential between the ␣ 1 C77-WT and ␣ 1 C77-9* channels when expressed in HEK293 cells with ␤ 2a and ␣ 2 ␦ auxiliary subunits (Fig. 5B, p Ͼ 0.05). Changing the holding potential to Ϫ50 mV did not alter the drug sensitivity of the two channels (Fig. 5C, p Ͼ 0.05). The ␣ 1 C77-WT channel contains exon 8 and displayed nearly 80% inhibition by 100 nM nifedipine and the presence of exon 9* in ␣ 1 C77-9* channels did not make any difference to nifedipine blockade (Fig. 5B). However, whereas the inclusion of exon 9* shifts the I-V to the hyperpolarized direction, the addition of nifedipine on ␣ 1 C77-9* channels shifted the I-V relationship back in the depolarized direction (Fig. 5A), which is reminiscent of DHP drug effect on Ca v 1.2 channels (24).
Taken together, the inclusion of 9* underlies a dramatic hyperpolarized shift in activation potential and I-V curve but spared the Ca v 1.2 channels of any change in sensitivity to block by nifedipine.

DISCUSSION
The ability of vascular smooth muscle to contract at physiological voltages requires a remarkable customization of the FIG. 4. Basic electrophysiological properties of exon 9* on Ca v 1.2 calcium channels. A, exemplary test pulses of ␣ 1 C77-9* and ␣ 1 C77-WT stepped from Ϫ90 to Ϫ50, Ϫ30, Ϫ10, 0, or 30 mV (upper panel). An ensemble of whole-cell I-V relationships was obtained for ␣ 1 C77-9* and ␣ 1 C77-WT by fitting the Boltzmann function (lower panel). Averaged I-V curves for I Ba through ␣ 1 C77-9* and ␣ 1 C77-WT channels transiently transfected in HEK293 cells were obtained by holding the cells at Ϫ90 mV and stepping to various potentials from Ϫ50 to 50 mV (⌬V ϭ 10 mV) over 900 ms. Individual I-V curves were first normalized to the respective peak current and then averaged (• ϭ  16, E ϭ 9). B, activation properties of ␣ 1 C77-9* and ␣ 1 C77-WT. The upper panel showed the exemplary tail currents evoked by repolariza-tions to Ϫ50 mV after depolarizing test pulses at Ϫ40, 0, 40, or 90 mV. The lower panel shows the ensemble of activation properties of the ␣ 1 C77-9* and ␣ 1 C77-WT channels obtained by tail activation protocol. Normalized (G/G max ) versus V curves were generated from dual-Boltzmann functions. C, steady-state inactivation. The upper panel shows exemplar current traces after 15-s conditioning depolarizing pulses evoked at Ϫ120, Ϫ90, Ϫ60, Ϫ40, or Ϫ20 mV. The lower panel shows the ensemble of steady state inactivation properties obtained by evoking a normalizing pulse to 0 mV for 30 ms followed by a variable family from Ϫ120 to 20 mV of 15-s conditioning pulses and a 104-ms test pulse to 0 mV. L-type calcium channels found in such tissue; compared with the analogous calcium channels found in the heart, the voltage dependence of activation for smooth muscle L-type channels is shifted some Ϫ15 mV in the hyperpolarizing direction (25,26). Since the tension-voltage relation of smooth muscle bears an intimate connection to the voltage dependence of activation of its L-type channels (25), the hyperpolarized activation properties of these channels are required for adequate tension to be generated by the relatively modest depolarization seen physiologically. If the voltage-dependent activation of smooth muscle L-type channels were similar to what is found in the cardiac/neuronal setting, little tension would be generated by smooth muscle. Despite the fundamental importance of the hyperpolarized activation threshold of smooth muscle L-type channels, the molecular basis for this specialization remains elusive. This unknown represents a critical deficit in our understanding of vascular biology, especially given that the underlying structural elements could represent valuable molecular targets for novel therapeutics to modulate vascular tone. Here, we report that alternative splicing of exon 9* appears to underlie the key shift in voltage activation parameters.
In this report, we demonstrated the high expression of exon 9* in the smooth muscle layer of arteries. Whole-cell patch clamp recordings revealed hyperpolarized shifts in I-V relationship and voltage-dependent activation of the 9*-containing Ca v 1.2 channels. However, the inclusion of exon 9* did not affect the sensitivity of Ca v 1.2 channels to nifedipine blockade or voltage-dependent steady-state inactivation. Nonetheless, this report provided valuable evidence on the restricted localization of 9*-containing Ca v 1.2 channels on smooth muscle, adding further support to the expression of particular alternatively spliced exons of calcium channels in a cell-selective manner (27). It also provided further evidence for phenotypic variation generated simply by employing cassette exons to produce proteomic diversity. We have also contributed to the understanding that the combination of predominant alternatively spliced exons in arterial smooth muscle contained at least exons 1, 8, Ϯ9*, and 32 (5,9,10).
Phenotypic and genotypic diversity of genes generated by the shuffling of alternatively spliced exons can be enormous (28). The degree of proteomic variations depends largely on the number of splice exons available and the combinatorial utilization of the alternatively spliced exons. What has been fascinating is the discovery of cell-specific expression of alternatively spliced exon (27) and the choice of alternatively spliced exons in response to physiological or pathological triggers (29 -33). The phenotypic diversity in ␣ 1 1.2 channels arising from alternative splicing is especially relevant to arterial smooth muscle or cardiac function, since there might not be a direct role for ␣ 1 1.3 channels in both cardiac ionotropy and smooth muscle contractility (34).
Here we have demonstrated exon 9* to be expressed in a spatially localized or restricted manner. The physiological significance of the predominant distribution of 9*-containing channels in blood vessels will require further investigations but may account at least in part for the differences in halfactivation voltages of L-type channels in smooth and cardiac muscles (25,26). It is tempting to speculate that 9*-containing Ca v 1.2 channels would activate upon slight membrane depolarization, resulting in early onset or increased frequency of constriction of blood vessels immediately after the arrival of a depolarizing pulse. One possible approach to test this hypothesis would be the use of small interfering RNA technology to knock down exon 9* expression in isolated arteries or in primary cultures of arterial myocytes to correlate with changes in vascular tone or tension. Another approach would be to take advantage of transgenic mouse technology to ablate or enhance the expression of 9*-containing Ca v 1.2 channels and to determine the physiological consequences in relation to altered vascular tone and/or the frequency or amplitude of Ca 2ϩ spark formation. One exciting idea is that the more depolarized resting membrane potential of arterial smooth muscle may more easily allow the activation of 9*-containing Ca v 1.2 channels after sympathetic stimulation or elevation of intravascular  5. A, an ensemble of current-voltage curves from ␣ 1 C77-9* treated with or without 1 nM nifedipine was obtained at Ϫ90-mV holding potential (• ϭ 14, E ϭ 6) before stepping to 0 mV. B, the relative inhibition of ␣ 1 C77-9* and ␣ 1 C77-WT at various concentrations of nifedipine remained unchanged. The cells were first held at Ϫ90 mV in nifedipine before currents were evoked at a test pulse of 0 mV. The evoked currents after nifedipine treatment were normalized to pretreated I Ba currents. No significant difference in inhibition was found between ␣ 1 C77-9* and ␣ 1 C77-WT at the three concentrations of nifidipine used (p Ͼ 0.05). C, inhibition of ␣ 1 C77-9* and ␣ 1 C77-WT by 1 nM nifidipine applied at a Ϫ50 mV holding potential was also found to be not significantly different (p Ͼ 0.05). All of the data are given as means Ϯ S.E.
pressure. The net result may lead to further depolarization of the membrane potential, giving rise to a more sustained Ca 2ϩ influx and presumably greater myogenic tone or more prolonged contraction of the blood vessels (35). Exon 9* is positioned immediately downstream of the ␣ 1 1.2 ␣-interacting domain in the I-II loop. The binding of the ␤-subunit modulates the Ca v 1.2 channel function by enhancing Ca 2ϩ currents and changing the kinetics of the activation and inactivation profiles of the Ca v 1.2 channels (36). A cross-talk between G␤␥ and ␤-subunit occurs around the ␣-interacting domain region of the P/Q-type calcium channel (37), but Ca v 1.2 or Ca v 1.3 channels lack voltage-dependent G protein coupling (38,39). Nonetheless, it remains to be tested whether the addition of 25 amino acids encoded by exon 9* may define the binding selectivity with different ␤-subunits to impact on the modulatory effects of different ␤-subunits on Ca v 1.2 channel behavior. In the Ca v 2.1 P/Q-type channel, the insertion of a single valine residue close to the ␣-interacting domain altered both G-protein and protein kinase C regulation of the channels (40) and also drastically slowed the inactivation of the channel.
Ca v 1.2 channels in cardiac and smooth muscle have been shown to be regulated, although at different levels, by protein kinase A (41). Molecular and biochemical studies indicated the location of the protein kinase A site at serine 1928 of the ␣ 1 1.2 subunit (42)(43)(44). Another putative protein kinase A site is located on exon 9*, rgTpagmldqkkgkfawfshsteth (20). However, this threonine residue in exon 9* is only present and conserved in humans and rabbits, and it is absent in rat and mouse ␣ 1 1.2 subunits (GenBank TM accession numbers AY323810 and U17869). This lack of conservation may debunk the possibility that the threonine residue is a possible protein kinase A phosphorylation site, but on the other hand, there may be species variation in protein kinase A regulation of the Ca v 1.2 channels. This hypothesis requires further experimental testing.
Whereas this report revealed that Ϯexon9* are expressed in adult aorta in comparable amounts and that exon 9* is selectively localized on the smooth muscle, it did not address the question of the predominant splice combination of the ␣ 1 1.2 subunit. One approach is to construct a library of full-length cDNAs of the ␣ 1 1.2 subunit from mRNAs isolated from cardiac and smooth muscle myocytes and probe the identity of the alternatively spliced exons at the 19 loci that have been identified previously (8). This will provide a quantitative evaluation of the predominant smooth muscle or cardiac isoforms of ␣ 1 1.2 subunit. Such channel isoforms would be immensely helpful in furthering our understanding of the pharmacological or physiological behaviors of the Ca v 1.2 channels in heart and blood vessels.