A Smooth Muscle Cav1.2 Calcium Channel Splice Variant Underlies Hyperpolarized Window Current and Enhanced State-dependent Inhibition by Nifedipine*

Native smooth muscle L-type Cav1.2 calcium channels have been shown to support a fraction of Ca2+ currents with a window current that is close to resting potential. The smooth muscle L-type Ca2+ channels are also more susceptible to inhibition by dihydropyridines (DHPs) than the cardiac channels. It was hypothesized that smooth muscle Cav1.2 channels exhibiting hyperpolarized shift in steady-state inactivation would contribute to larger inhibition by DHP, in addition to structural differences of the channels generated by alternative splicing that modulate DHP sensitivities. In addition, it has also been shown that alternative splicing modulates DHP sensitivities by generating structural differences in the Cav1.2 channels. Here, we report a smooth muscle L-type Cav1.2 calcium channel splice variant, Cav1.2SM (1/8/9*/32/Δ33), that when expressed in HEK 293 cells display hyperpolarized shifts for steady-state inactivation and activation potentials when compared with the established Cav1.2b clone (1/8/9*/32/33). This variant activates from more negative potentials and generates a window current closer to resting membrane potential. We also identified the predominant cardiac isoform Cav1.2CM clone (1a/8a/Δ9*/32/33) that is different from the established Cav1.2a (1a/8a/Δ9*/31/33). Importantly, Cav1.2SM channels were shown to be more sensitive to nifedipine blockade than Cav1.2b and cardiac Cav1.2CM channels when currents were recorded in either 5 mm Ba2+ or 1.8 mm Ca2+ external solutions. This is the first time that a smooth muscle Cav1.2 splice variant has been identified functionally to possess biophysical property that can be linked to enhanced state-dependent block by DHP.

However, this subpopulation of Ca v 1.2 calcium channels has not been identified. Of the four subunits (␣ 1 , ␤, ␣2/␦, and ␥) that composed the calcium channels, the ␣ 1 -subunit not only forms the aqueous pore but is also the site for binding of organic agonists or antagonists (5,6). The Ca v 1.2 L-type calcium channels are localized mainly in the brain, cardiac, and smooth muscles and a large number of alternative splicing sites of the ␣ 1 -subunit have been reported in various tissues (7)(8)(9). Changes in the expression levels or mutations of alternatively spliced exons of the Ca v 1.2 calcium channel have also been identified in development, heart failure, myocardial infarction, and congenital heart disease (10 -14). Importantly, alternative splicing serves to diversify Ca v 1.2 calcium channel biophysical properties and as such it is reasonable to test our first hypothesis as to whether a Ca v 1.2 splice variant may possess properties that are more similar to the subpopulation of channels that exhibit hyperpolarized window current.
L-type calcium channel blockers, 1,4-dihydropyridines (DHPs), 2 phenylalkylamines, and benzothiazepines, are being used in the management of cardiovascular diseases (5,15,16). Two mechanisms have been suggested for the observed vascular selectivity of DHP inhibition over cardiac Ca v 1.2 channels. One mechanism was the more depolarized resting membrane potential of smooth muscles (17) and the second was the difference in voltage dependence of DHP modulation arising from tissue-specific alternative splicing of mutually exclusive exons 8 and 8a that code for IS6 in smooth and cardiac muscle Ca v 1.2 channels (18). Characterization of cDNA clones of Ca v 1.2 channels isolated from smooth and cardiac muscles segregated two clones as the cardiac muscle form Ca v 1.2a (␣ 1Ca ) channel and smooth muscle form Ca v 1.2b (␣ 1Cb ) channel (19,20). Both isoforms were reported to have minor differences at four alternative splicing loci amounting to ϳ5% of their amino acid sequence. The four splice sites are: exon 1/1a in the N terminus, exon 8/8a in the transmembrane segment IS6, exon 31/32 in the transmembrane segment IVS3, and exon 9* in the cytoplasmic loop connecting domains I and II. Ca v 1.2a was characterized to contain alternatively spliced exons 1a/8a/⌬9*/31/33 (20), whereas Ca v 1.2b contained alternatively spliced exons 1/8/9*/ 32/33 (19). Notably, both channels share similar activation and inactivation properties (21,22). However, Ca v 1.2b channels are more sensitive to nisoldipine due to the presence of the alternatively spliced exon 8 within the IS6 segment (18,21). Importantly, these previous works reported the lack of linkage between the gating properties of Ca v 1.2 smooth and cardiac muscle splice variants with vascular specificity for DHP inhibition (18,21). We have previously shown by systematic transcript scanning that the Ca v 1.2 subunit is subject to extensive alternative splicing generating potentially a large number of splice combinations (8). For our second hypothesis, we tested whether a smooth muscle Ca v 1.2 splice variant exists that possesses a gating property that will link its steady-state inactivation property to the potency of inhibition by DHP as blockade by DHP is state-dependent.
In the present study, we report a smooth muscle Ca v 1.2 splice variant (Ca v 1.2SM) that differs from the notable smooth muscle Ca v 1.2b channel that lacks exon 33. Here we showed that the Ca v 1.2SM calcium channels exhibited a more hyperpolarized Ca 2ϩ window current that supports our first hypothesis. Interestingly, the same Ca v 1.2SM splice variant also possesses electrophysiological and pharmacological properties that support our second hypothesis to link altered biophysical property of a smooth muscle Ca v 1.2 channel splice variant to the potency of DHP inhibition. Taken together, we showed that unexpectedly a single smooth muscle Ca v 1.2 splice variant, Ca v 1.2SM, underlies both a hyperpolarized window current that is reminiscent of the window current of a subpopulation of channels observed in smooth muscles (2)(3)(4) and a more potent inhibition by nifedipine in correlation with a hyperpolarized shift in steady-state inactivation potentials.
RT-PCR, Colony Screening, and Restriction Enzyme Digestion-Young adult male Wistar rats (150 -200 g) were sacrificed by CO 2 followed by subsequent cervical dislocation. All animal experimentations were conducted according to IACUC guidelines and all procedures have been approved by the University's Animal Ethics Committee. Human radial arteries were obtained, with informed consent, from patients who underwent coronary artery bypass operations in the National Heart Center of Singapore. The work was approved by the IRB com-mittees of the National Heart Center of Singapore, Singapore General Hospital, and the National University of Singapore.
Total RNA from thoracic aorta, cerebral basilar arteries, mesentery arteries, and hearts were extracted with RNAeasy kits (Qiagen). Reverse transcription for first strand synthesis was carried out using the SUPERSCRIPT TM II RNase H Reverse Transcriptase (Invitrogen) and 18-mer oligo(dT) primer. The oligonucleotide primers used for amplifying rat Ca v 1.2 amplicon flanking exon 33 were: GCCTCTTCACGGTGGAG (forward) and TCCCAATCACTGCATAGATAA (reverse). Primers used to amplify similar regions in human radial arteries were: AACACCATCTGCCTGGCCATG (forward) and AGGTCTGAAAGTTGTTGTTCCGG (reverse). The PCR protocol includes a denaturation step at 95°C for 5 min; 35 cycles of 95°C for 30 s, 52°C for 45 s, and 72°C for 1 min; and a final extension step at 72°C for 10 min. After electrophoresis in 2% agarose gel, the PCR products were excised, extracted, and purified with a Qiagen kit. The PCR amplicons were further cloned into pGEM-T Easy vector (Promega) and transformed into DH10B Escherichia coli cells. White colonies were selected and grown in 96-well plates. Colony PCRs with the same set of primers and conditions were performed to identify the presence or absence of exon 33 in each clone. Restriction enzyme digestion was used to investigate the expression of mutually exclusive exons 31 and 32, which have similar size. NsiI digestion was carried out in a 30-l reaction volume at 37°C for 5 h.
Whole Cell Electrophysiological Recordings and Data Analysis-HEK 293 cells in 35-mm dishes were transiently transfected with Ca v 1.2b, Ca v 1.2CM, or Ca v 1.2SM constructs (1.25 g), together with ␤ 2a (1.25 g) and ␣ 2 ␦ (1.25 g) using the calcium phosphate transfection method (8). The ␤ 2a and ␣ 2 ␦ clones were provided by Dr. Terrance Snutch (University of British Columbia). After 48 -72 h, I Ba was recorded at room temperature (23°C) using the whole cell patch clamp technique. The external solution contained (in mM) 10 HEPES, 140 tetraethylammonium methanesulfonate, 5 BaCl 2 , or 1.8 CaCl 2 (pH was adjusted to 7.4 with CsOH and osmolarity to 290 -310 with glucose). The internal solution (pipette solution) contained (in mM) 138 Cs-MeSO 3 , 5 CsCl, 0.5 EGTA, 10 HEPES, 1 MgCl 2 , 2 mg/ml Mg-ATP, pH 7.3 (adjusted with CsOH). Glucose was used to adjust the osmolarities of solutions to between 290 and 330 mOsm. 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 to 50 mV over 900 ms. In each cell, I Ba at all voltages was normalized to the peak current. The I-V curve was fitted with the equation: where G max is the maximum conductance; E rev is the reversal potential; V 1/2 is the half-activation potential; and k is the slope. The steady-state inactivation curves were obtained from experiments by stepping from a holding poten-tial of Ϫ90 mV to a 30-ms normalizing pulse to 10 mV followed by a family of 15-s long prepulses from Ϫ120 to 20 mV. A 104-ms test pulse to ϩ10 mV was recorded finally. Each test pulse was normalized to the maximal current amplitude of the normalizing pulse. The steady-state inactivation data were fitted with a single Boltzmann equation: I relative ϭ I min ϩ (I max Ϫ I min )/(1 ϩ exp((V 1/2 Ϫ V)/k), where I relative is the normalized current; V 1/2 is the potential for half-inactivation, and k is the slope value. G-V curves were obtained from a tail activation protocol. The cells were activated by a 20-ms test pulse of variable voltage family from Ϫ60 to 120 mV, and tail currents were measured after repolarization to Ϫ50 mV for 10 ms. The tail currents were normalized to the peak currents before fitting with a dual Boltzmann equation: where G is the tail current and G max is the peak tail current, F low is the fraction of low threshold component; V 1/2,low , V 1/2,high , k low , and k high are the half-activation potentials and slope factors for the low and high threshold components and V 1/2act was calculated when G ϭ 0.5G max . values for voltage dependence inactivation (in Ba 2ϩ solution) were calculated from a single exponential equation: I ϭ I min ϩ I max (exp(Ϫt/)), where I min is the steady-state amplitude of the current, I max is the initial current, and is the time constant. Recovery from inactivation of I Ba was investigated using a standard two-pulse protocol. Fractional recovery of peak I Ba were plotted against ⌬T. Inactivation time constants () were determined from the double exponential equation: where Y is the fraction of recovery, A 1 and A 2 are the maximum values of the fast and slow component, and f and s are the time constants, respectively.
Nifedipine (Sigma) was dissolved in Me 2 SO to make a 10 mM stock solution and stored at Ϫ20°C. Nifedipine at 0.1-1000 nM concentrations was freshly prepared in bath solution from stock and then applied to the HEK 293 cells. All nifedipine solutions were stored in the dark. The cells were held at either Ϫ50 or Ϫ90 mV and a test pulse of 100 ms at 0 mV was applied with an inter pulse interval of 30 s. Inhibition ratios were calculated by comparing the peak currents before and after drug treatment. All experiments involving nifedipine were conducted as much as possible in the absence of light.
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 or one-way ANOVA. p Ͻ 0.05 is considered significant.

Distribution of Exon 33 in Arteries and Hearts-The
Ca v 1.2 calcium channels in cardiac and smooth muscles have been shown to differ at four alternatively spliced sites: exon 1/1a, exon 8/8a, exon 9*, and exon 31/32 ( Fig. 1A) (19,20). Exon 33 forms part of the IVS3-S4 linker region and in our previous study we showed that the exclusion of exon 33 could shift the activation potential of Ca v 1.2 channels to more hyperpolarized direction (8,24). In this paper, we investigated whether exon 33 Water control (Ϫ) was also shown. C, colony PCR screening of cloned amplicons amplified from mRNA isolated from rat heart (upper panel) shows the presence of exon 33 in the majority of colonies. Lower panel shows results from aorta where clones that excluded exon 33 (⌬33) displayed PCR products of smaller size. The identities of ϩ/⌬33 were confirmed by DNA sequencing. D, summary of expression of ⌬33 in rat LV (N rat ϭ 8, N clone ϭ 585), RV (N rat ϭ 3, N clone ϭ 240), LA (N rat ϭ 5, N clone ϭ 306), and RA (N rat ϭ 3, N clone ϭ 190) as compared with rat aorta (N rat ϭ 5, N clone ϭ 498). E, expression of ⌬33 in rat mesenteric artery (Mes, n ϭ 3 rats), rat cerebral basilar artery (Cer, n ϭ 3 rats), and human radial artery (Rad, n ϭ 5 samples) as compared with rat heart (Hrt, n ϭ 17 rats). *, p Ͻ 0.01; **, p Ͻ 0.001 (one-way ANOVA and Newman-Keuls test). NOVEMBER 30, 2007 • VOLUME 282 • NUMBER 48 may demonstrate selective expression in smooth and cardiac muscles.

Novel Smooth Muscle Ca v 1.2 Splice Variant
The expression of exon 33 was assessed by a RT-PCR colony screening method (8,12,25) that provides quantification of expression that may differ at levels not easily determined by real-time PCR. PCR amplicons spanning exons 30 -35 appeared as single bands on 1.8% agarose gels for rat aorta and heart tissues (Fig. 1B). To differentiate ϩ/⌬exon 33 in the PCR products, which differed by 33 bp, the PCR products were cloned into pGEM-T Easy vector, transformed into DH10B bacteria, and PCR colony screened for size differences on 1.8% agarose gels. This method provided a means to quantify the level of expression of Ca v 1.2 transcripts, isolated from aorta and heart, with the inclusion or exclusion of exon 33. Random clones were selected for DNA sequencing analysis to validate the correlation of PCR product size with the ϩ/⌬exon 33 genotype in the PCR amplicon. Around 60 -120 colonies were screened in each rat heart or artery tissue. A representative agarose gel result of a colony PCR screen was shown in Fig. 1C. Screening of heart Ca v 1.2 channels showed that the colonies expressed almost exclusively the inclusion of exon 33 (Fig. 1C, upper panel), whereas data from aorta tissue demonstrated that a number of Ca v 1.2 channels excluded exon 33 (Fig. 1C, lower panel). DNA sequencing confirmed the PCR product of smaller size lacked exon 33, whereas the PCR product of larger size contained exon 33. To further delineate the expression of exon 33 in different regions of the heart and in aorta, we screened PCR products obtained from mRNAs isolated from 3 to 8 rats. Indeed, ϳ32% of the Ca v 1.2 channels in aorta (n ϭ 5 rats) excluded exon 33, whereas the distribution of ⌬33 in heart (n ϭ 3-8 rats) varies from 2% in the right ventricle to 5% in the left atrium (Fig. 1D). The total number of colonies screened varied from 190 to 585 for each heart sample or aorta. The expression of ⌬33 is significantly different in all four regions of the heart when compared with aorta (p Ͻ 0.001, one-way ANOVA and Newman-Keuls test). However, there is no significant difference in the expression of ⌬33 among the various regions of the heart (p ϭ 0.87, one-way ANOVA). As the difference in expression of ⌬33 was ϳ30%, it would be difficult to detect such minor changes consistently by the real-time PCR method. The colony screening method, however, is more sensitive to detect small variations in the expression of an alternatively spliced exon.
To test whether ⌬33 expresses in smaller arteries, we harvested rat cerebral basilar arteries and mesentery arteries. A similar colony screening method was used to quantify ⌬33 expression. Mesenteric arteries (n ϭ 3 rats) displayed 17%, whereas basilar arteries (n ϭ 3 rats) expressed 36% of ⌬33Ca v 1.2 channels. Both results were significantly different from the levels expressed in rat heart (Fig. 1E, p Ͻ 0.01, one-way ANOVA and Newman-Keuls test). We further investigated the expression of ⌬33 in human arteries. Of 5 human radial arteries obtained from patients who underwent coronary artery bypass, the average expression of ⌬33 was 24%, which is still significantly higher than in rat heart (Fig. 1E, p Ͻ 0.01, oneway ANOVA and Newman-Keuls test). These results indicate that ⌬33 expression appears widely in various rat and human arteries.
The well known Ca v 1.2b channels that contain exons 1/8/ 9*/32 could now be subdivided in a more defined manner into two subgroups: one group contained exon 33, whereas a minor group excluded exon 33 (Fig. 1A). In our previous work we have identified the smooth muscle Ca v 1.2 channel, genotyped as 1/8/ 9*/32/⌬33, named Ca v 1.2SM in this paper to represent ϳ10% of the total transcripts (26).
Ca v 1.2SM Channels Underlie a Hyperpolarized Window Current-To avoid Ca 2ϩ -dependent effects and also for a better comparison with previous reports on Ca v 1.2a and Ca v 1.2b channel properties, we first performed electrophysiological recordings using 5 mM Ba 2ϩ as the charge carrier (18,(21)(22)(23). WefoundthattheCa v 1.2SMsplicevariantshowedfastervoltagedependent inactivation during depolarization to V max than Ca v 1.2b channels ( Fig. 2A) and Ba 2ϩ influx was significantly reduced, during 0.9-s pulses to V max , by about 15% (Fig. 2B). It is known that an increased Ba 2ϩ concentration can accelerate current inactivation. We carefully investigated the voltage-dependent inactivation with matched peak current and found the faster inactivation is still pronounced in Ca v 1.2SM. We further calculated the time constant ( values) with a single exponential equation. The values from Ca v 1.2SM channels were significantly lower at 0 and 10 mV (Fig. 2C, p 0 mV Ͻ 0.01 and p 10 mV Ͻ 0.05, unpaired t test). It is very likely that the faster inactivation is due to the exclusion of exon 33 in the domain IV S3-S4 loop as the only difference between Ca v 1.2SM and Ca v 1.2b is the exclusion of exon 33 in Ca v 1.2SM. To further determine differences in biophysical properties of the two smooth muscle Ca v 1.2 splice variants, we examined their activation and inactivation potentials by functionally expressing the splice variants in HEK 293 cells (Fig. 2D).
Our whole cell patch clamp electrophysiological recordings showed that Ca v 1.2SM channels activated and inactivated more negatively than the Ca v 1.2b channels (Fig. 2D). The V 1/2 of steady-state inactivation significantly shifted by Ϫ16.3 mV, whereas the V 1/2 of activation potential shifted by Ϫ11.3 mV ( Fig. 2D and Table 1, p Ͻ 0.0001, unpaired t test). The slope factors between the two channels are similar in steady-state inactivation and the lower component of G-V curves (K low ). The slope factor of Ca v 1.2SM on the higher component of G-V curves (K high ) is larger than Ca v 1.2b channels with bigger standard error. The shift of both the activation and inactivation curves of Ca v 1.2SM channels generates a Ba 2ϩ window current that peaked at Ϫ29.9 mV, compared with the peak at Ϫ20.8 mV for Ca v 1.2b channels. This 9 mV left-shifted window current of Ca v 1.2 SM channel would permit the channels to maintain opening at hyperpolarized membrane potentials closer to resting potential than Ca v 1.2b channels. To test the activation and inactivation properties of both channels under physiological conditions, we performed experiments using 1.8 mM Ca 2ϩ as the charge carrier. Although generally smaller currents were recorded in 1.8 mM Ca 2ϩ bath solution, a similar left shift of window current was observed ( Fig. 2E 1.38, n ϭ 4; p ϭ 0.0003, unpaired t test). To more carefully examine the Ca 2ϩ window current, we calculated the window current by multiplying the activation conductance (d ∞ ) curve by the inactivation curve (f ∞ ) as reported previously by Chemin et al. (27). Current density from HEK 293-transfected cells was not included in the analysis. Calculated window currents from the 30% of Ca v 1.2SM and 70% of Ca v 1.2b channels were shown in Fig. 2F. The window current of Ca v 1.2SM channels starts at Ϫ60.4 mV and peaked at Ϫ45.9 mV. The window current of Ca v 1.2b channels starts at Ϫ55.5 mV and peaked at Ϫ41.07 mV. The combined window current of both channels represents around 1.5% of peak current. Taking into consideration that the 11 mV junction potential that was not corrected, the Ca v 1.2SM channels (and some of Ca v 1.2b channels) could be activated more readily close to the resting membrane potential of smooth muscles (between Ϫ75 to Ϫ60 mV in vitro, or Ϫ50 to Ϫ40 mV in vivo) (28) to maintain vasotone.
To further determine whether mechanistically the Ca v 1.2SM channels may be more likely to open during high frequency stimulation, we determined the rate of I Ba recovery of the channels from inactivation. A typical two-pulse protocol was used with a prepulse of 2-s duration chosen as in previous reports to better inactivate the channels (22,29,30). During 2-s prepulses of 0 mV from a holding potential of Ϫ90 mV, currents were reduced to around 30% (Ca v 1.2SM, 28 Ϯ 2%, n ϭ 20; Ca v 1.2b, 32 Ϯ 2.3%, n ϭ 13; p ϭ 0.19, unpaired t test). A subsequent 40-mS test pulse was applied at variable time intervals (⌬T) after prepulses. After a second pulse to 0 mV, over 80% of all available channels recovered from the inactivated state within 4 s. We found that there was no difference in the recovery rate between the two smooth muscle Ca v 1.2 splice variants (Fig. 2G).

Exon 32 Expression Is Predominant in Cardiac
Ca v 1.2 Channels-Exons 31 and 32 are mutually exclusive exons forming part of the IVS3 region (Fig. 1A). It was suggested that the cardiac isoform contained exon 31, whereas the smooth muscle isoform contained exon 32 (19,20). Alternative splicing at this locus was demonstrated to influence DHP sensitivity slightly (31). However, we and others have reported that exon 32 is predominant in adult heart from various species (8,10,14,32), and exon 31 was expressed at higher levels only in neonates (10). To further show that exon 32 is the predominant form in adult rat heart, we used restriction enzyme digestion by NsiI to identify the presence of exon 32 in the PCR products amplified from rat heart mRNA across exons 30 -35. Digestion of the 382-bp product by NsiI that yielded two smaller fragments of The current traces were shown for determining activation property evoked at Ϫ60, Ϫ20, 20, 60, and 100 mV (left) and for inactivation property evoked at Ϫ100, Ϫ60, Ϫ20, and 20 mV (right). Plots of steady-state inactivation (f ϱ ) and activation (d ϱ ) curves of Ca v 1.2SM (gray) and Ca v 1.2b (black) channels derived from recordings performed in 5 mM Ba 2ϩ . E, plots of steadystate inactivation (f ϱ ) and activation (d ϱ ) curves of Ca v 1.2SM (gray) and Ca v 1.2b (black) channels derived from recordings in bath solution containing 1.8 mM Ca 2ϩ . F, plots of calculated calcium window currents of Ca v 1.2SM (circles) and Ca v 1.2b (filled circles) channels, as well as the combined window current (line). G, recovery from inactivation of I Ba was investigated using a standard two-pulse protocol. Fractional recovery of peak I Ba was plotted against ⌬T. NOVEMBER 30, 2007 • VOLUME 282 • NUMBER 48 314 and 68 bp would indicate the presence of exon 32, whereas the undigested 382-bp product would contain exon 31 instead (Fig. 3A). The results of this experiment clearly showed that heart Ca v 1.2 channels expressed more abundantly exon 32 than exon 31 (Fig. 3B). We therefore constructed a new cardiac isoform of Ca v 1.2 channel named Ca v 1.2CM (1a/8a/⌬9*/32/33), which is a better representation of the adult cardiac isoform of Ca v 1.2 channels than the Ca v 1.2a channels (Fig. 3C).

Novel Smooth Muscle Ca v 1.2 Splice Variant
Steady-state Inactivation of Ca v 1.2SM Is More Hyperpolarized Shifted Than Ca v 1.2CM Channels-As the biophysical properties of Ca v 1.2SM and Ca v 1.2CM channels have not been evaluated together, we first compared their current-voltage (I-V) relationships by stepping to a family of test potentials from a holding potential of Ϫ90 mV. The current-voltage (I-V) curves showed no statistical difference between Ca v 1.2CM and Ca v 1.2SM channels (Fig. 4B and Table 2, Ca v 1.2CM, V 0.5 ϭ Ϫ27.49 Ϯ 1.32, n ϭ 9 cells; Ca v 1.2SM, V 0.5 ϭ Ϫ24.68 Ϯ 1.09, n ϭ 10 cells; p ϭ 0.1164, unpaired t test). From the sample traces of I-V relations shown in Fig. 4A, we found that the voltage-dependent inactivation of Ca v 1.2SM was faster than Ca v 1.2CM. To characterize the voltage-dependent inactivation of both channels, we compared the time-dependent inactivation of V max (Fig. 4C). The inactivation rate of Ca v 1.2SM is faster than Ca v 1.2CM in 0.3, 0.6, and 0.9 s, p 0.3 s ϭ 0.0007, p 0.6 s ϭ 0.0028, p 0.9 s ϭ 0.0074, unpaired Student's t test (Fig. 4D). At 0 mV, the values of Ca v 1.2SM channels are significantly smaller than Ca v 1.2CM channels (Fig. 4E) after fitting with a single exponential equation. To more accurately investigate activation potentials we employed the tail protocols and again found that there was no significant difference between the V 1/2 act of both channels ( Fig. 4F and Table 2, p ϭ 0.4728, unpaired t test). Both I-V relationships and tail currents clearly showed the Ca v 1.2SM and Ca v 1.2CM shared similar activation properties. However, when we assessed the voltage-dependent steady-state inactivation properties of both channels by providing a 15-s conditioning prepulse before stepping to test pulse at 10 mV, we detected an approximate Ϫ16 mV hyperpolarized shift of V 1/2 inact for Ca v 1.2SM over Ca v 1.2CM channels ( Fig. 4F and Table 2, p Ͻ 0.0001, unpaired t test). Thus the Ca v 1.2SM channels own a smaller window current than Ca v 1.2CM.
Biophysical Property of Ca v 1.2SM Channels Linked to Enhanced Inhibition by Nifedipine-To further test our hypothesis that altered steady-state inactivation alone in a smooth muscle Ca v 1.2 splice variant could also modulate sensitivity to DHP, nifedipine was used to block both smooth muscle splice variants, Ca v 1.2SM and Ca v 1.2b, and cardiac Ca v 1.2CM channels at various concentrations. As nifepidine blockade is state-dependent, we performed the whole cell electrophysiological recordings at two holding potentials (V h ϭ Ϫ90 or Ϫ50 mV) in bath solution containing 5 mM Ba 2ϩ . The two holding potentials were chosen to better reflect the pharmacological responses of the channels held close to the reported resting potentials of smooth and cardiac muscles. We also performed experiments using more physiological conditions with 1.8 mM Ca 2ϩ as the charge carrier. However, the current was too small when the cells were held at Ϫ50 mV. Experiments with 1.8 mM Ca 2ϩ were therefore only conducted in the holding potential of Ϫ90 mV.
Step depolarizations to 0 mV were applied every 30 s and Fig. 5A illustrates the diary plot of nifedipine inhibition. The cells were viable as after washing out the nifedipine, substantial recovery of I Ba was observed (Ͼ85% recovery). Exemplary current traces from three channels, in the absence and presence of 10 nM nifedipine, are shown in Fig. 5B. All three channels were more sensitive to nifedipine under depolarized holding potentials. When the Ca v 1.2 splice variants were characterized pharmacologically under the same holding potentials, we found that Ca v 1.2SM channels were the most potently inhibited by nifedipine, followed by Ca v 1.2b channels, whereas the Ca v 1.2CM channels exhibited the least inhibition. The dose-response curves fitted with the Hill equation clearly indicated that nifedipine inhibited Ca v 1.2SM channels between 5.4 and 6.4 times more potently than Ca v 1.2b and Ca v 1.2CM channels at the holding potential of Ϫ90 mV and between 4.7 and 9.4 times at Ϫ50 mV, respec-  tively, in bath solution containing 5 mM Ba 2ϩ (Fig. 5, C and D, and Table 3). When 1.8 mM Ca 2ϩ was used as charged carrier, Ca v 1.2SM was inhibited 2.4 and 9.3 times more potently by nifedipine than Ca v 1.2b and Ca v 1.2CM channels (Fig. 5E and Table 3). The Ca v 1.2SM splice variant differs from the Ca v 1.2b channel in that it lacks exon 33, but nonetheless, the Ca v 1.2SM splice variant exhibited a hyperpolarized shift in steady-sate inactivation property that was linked to enhanced sensitivity to nifedipine. These results clearly affirm the state-dependent block of nifedipine but more importantly, it provides another mechanism to explain the vascular selectivity of DHP inhibition of Ca v 1.2 channels in the cardiovascular system. Taken together, we have demonstrated the existence of a subpopulation of smooth muscle Ca v 1.2SM splice variant that activated and inactivated close to resting potential and that were blocked more potently by nifedipine because of a hyperpolarized shift in steady-state inactivation potential.

DISCUSSION
In this report, we have characterized functionally a Ca v 1.2 splice variant that possesses electrophysiological and pharmacological properties that at least in part addressed two puzzles about smooth muscle L-type calcium channels. One is the presence of a subpopulation of DHP-sensitive Ca v 1.2 channels that open within a narrow window current that is close to the resting potential in smooth muscles. The second is the absence of reported Ca v 1.2 splice variants with biophysical properties that linked the inactivation state to block by DHP. We hypothesized that certain uncharacterized or novel smooth muscle Ca v 1.2 splice variants would underlie these two important vascular Ca v 1.2 channel properties. Unexpectedly, we found that a single Ca v 1.2SM splice variant exhibited a hyperpolarized shift in Ca 2ϩ window current and was also more potently inhibited by nifedipine correlating with a left-shift in its steady-state inactivation potential. Given the characteristics of the Ca v 1.2SM channels, it is not unreasonable to speculate that this splice variant could contribute to the basal level of Ca 2ϩ influx near resting potential and should therefore serve a critical purpose to underlie basal contractility in myocyte that is essential for vasotone and blood flow (3,4).
The smooth muscle Ca v 1.2b splice variant was shown to exhibit a window current that is similar to the cardiac Ca v 1.2 channel isoform (21). For the generation of vascular tone, a smooth muscle Ca v 1.2 channel splice variant that activates at more hyperpolarized potential is required as increase in smooth muscle tension is sensitive to changes in membrane potential, where the relationship between vascular tone and V m is steep in the potential range of Ϫ65 and Ϫ40 mV (28,33). As such, the Ca v 1.2SM splice variant will play a critical role in being responsive to activation close to the resting potential and upon further membrane depolarization, other splice variants can then be opened to contribute to the overall window current of smooth muscle Ca v 1.2 channels. It is of interest that the window currents of both Ca v 1.2SM and Ca v 1.2b channels are rather small and they peak around Ϫ45.9 and Ϫ41.1 mV. This ϳ5 mV difference is important given that the open probability of the channels or the tension of the muscle changes e-fold for 6 -8 mV change over the physiological membrane potential negative to Ϫ40 mV (33). Therefore, by itself the Ca v 1.2SM channels do not completely constitute the window current of native smooth muscles, but it should play an important role to contribute to the more hyperpolarized part of the overall win-  NOVEMBER 30, 2007 • VOLUME 282 • NUMBER 48 dow current and should be important to maintain steady-state resting vascular tone (3).

Novel Smooth Muscle Ca v 1.2 Splice Variant
The molecular structure of Ca v 1.2 channels in cardiac and arterial smooth muscles are of particular interest as they are the targets of calcium channel blockers treating a variety of cardiovascular diseases. Previous reports identified two major isoforms of Ca v 1.2 channels generated through the differential utilization of alternatively spliced exons (19,20). The Ca v 1.2a (cardiac form) and Ca v 1.2b (smooth muscle form) are different at 4 sites, namely exon 1/1a, exon 9*, exon 8/8a, and exon 31/32. Whereas the utilization of exon 1 or 1a is regulated by different promoters (34 -36), many reports found that exon 32 instead of exon 31 is predominantly expressed in adult heart (8,10,14,32). Thus a more accurate genotype of the predominant cardiac Ca v 1.2CM channel should contain alternatively spliced exons 1a/8a/⌬9*/32 instead of 1a/8a/⌬9*/31. This isoform was found only in cardiac myocytes but not found in rat aorta (26). Nonetheless, from our previous report, the electrophysiological   properties of exon 31 or exon 32 containing Ca v 1.2 channels were similar. Whereas we have detected use of the alternate donor site for exon 32, we have not been able to discover similar alternative splicing at the donor splice junction of exon 31 (8).
Whether this is the only functional difference arising from alternative splicing at the exon 31/32 locus is still unknown.
In this report, we found another alternatively spliced exon 33, which forms part of the extracellular IVS3-4 linker, to possess different expression patterns in heart and arteries. Approximately 30% of Ca v 1.2 channels within rat aorta lacked exon 33, whereas over 95% of rat heart Ca v 1.2 calcium channels do express exon 33. Thus the arterial smooth muscles contained at least two isoforms of Ca v 1.2 calcium channels: one is the traditional Ca v 1.2b, and the other is Ca v 1.2SM. Importantly, we found that beside aorta, other arteries with smaller lumen, like rat mesenteric and cerebral arteries and human radial arteries, contained variable but significant amounts of Ca v 1.2 channels lacking exon 33. The relevance of Ca v 1.2SM channels to smooth muscle function in arteries is therefore not restricted to the aorta but will be of more general significance in the vascular system. Different expression levels of exon 33 have been detected by RT-PCR by Feron et al. (37) in rat heart and rat smooth muscle cell line A7r5, where 7% of Ca v 1.2 channels in rat heart and 29% in A7r5 lacked exon 33. This expression pattern is similar to what we found in rat heart and aorta. In our previous report (26), we have shown that Ca v 1.2SM channels represented ϳ10% of the total full-length Ca v 1.2 channel transcripts detected in aortic smooth muscles, whereas no Ca v 1.2CM splice variants were detected. To determine how such a small subpopulation of Ca v 1.2SM channels will contribute sufficiently to vasotone will require more work in the future. It is, however, noteworthy that in patients suffering from Timothy's syndrome, a mutation found in exon 8a of the Ca v 1.2 gene, which accounted for 11.5% of total Ca v 1.2 channels, die from arrhythmias and sudden cardiac death (12). One possible genetic approach is to develop a mouse that is deleted of exon 33 in the genome and the prediction is the generation of a larger number of Ca v 1.2SM channels and therefore the development of increased vasotone and blood pressure.
Previous studies on Ca v 1.2a and Ca v 1.2b found no difference in the channel gating properties such as activation and steadystate inactivation potentials (18,21,22). In contrast, Ca v 1.2SM channels displayed a prominent hyperpolarized shift in steadystate inactivation potentials compared with Ca v 1.2CM and Ca v 1.2b channels. Furthermore, pharmacological study of the three channels indicated that nifedipine blocked them in a voltage-dependent manner. All channels were more sensitive to nifedipine when the holding potential was at Ϫ50 mV compared with Ϫ90 mV. It is reasonable as DHPs tend to bind inactivated channels rather than closed or open channels and at Ϫ50 mV holding potential, more channels are in the inactivated state. When we compared Ca v 1.2SM with Ca v 1.2CM channels, we found that Ca v 1.2SM channels are much more sensitive to nifedipine blockade. Even at Ϫ90 mV HP, the IC 50 of Ca v 1.2SM is less than that of Ca v 1.2CM at Ϫ50 mV HP. Hence, from previous studies, the first mechanism to explain the ratio of vascular:cardiac DHP sensitivity was the more depolarized resting potentials of smooth muscles (17,38). The first mechanism exploits the more depolarized membrane potential in smooth muscle membrane (17). The Ca v 1.2b channels would therefore be more likely to be shifted to the inactivated states after activation. As DHPs bind preferentially to the inactivated Ca v 1.2b channels (39 -41), DHPs thus showed a high vascular selection in native smooth muscles (5,42,43). The second mechanism operates via the alternative splicing of the domain IS6 segment, which is encoded by mutually exclusive exons 8/8a that influenced the inhibitory effects of DHPs on the channels in a voltagedependent manner. As such the different localizations and structures of cardiac and smooth muscle Ca v 1.2 channels determine their DHP sensitivities (18,21,23,31). However, with the different inactivation properties of Ca v 1.2SM channels, we now provide a third mechanism to explain tissue selectivity for DHP inhibition. Here, for the third mechanism, Ca v 1.2SM channel demonstrated a large hyperpolarized shift in voltage-dependent inactivation and this gating property confers on the smooth muscle channel variant enhanced sensitivity to nifedipine, thus linking biophysical property to potency of nifedipine block.
To test the precise role of the Ca v 1.2SM channel splice variant, specific knock-down of these variants for biological assessments of vascular tone or blood pressure are required. Nonetheless, in the management of hypertension with DHP, it is conceivable that Ca v 1.2SM channels can be targeted first to block their activities resulting in decreased vessel tone, followed by the blockade of other smooth muscle Ca v 1.2 channel splice variants that activate at more depolarized potentials. A selective inhibition of the Ca v 1.2SM channels in native smooth muscles that is efficacious in the management of hypertension or hypertension-related vascular disorders will certainly be of clinical relevance and interest.