Abstract
The transcripts of L-type voltage-gated calcium channels (CaV) 1.3 undergo extensive alternative splicing. Alternative splicing, particularly in the C terminus, drastically modifies gating properties of the channel. However, little is known about whether alternative splicing could modulate the pharmacologic properties of CaV1.3 in a manner similar to the paralogous CaV1.2. Here we undertook the screening of different channel splice isoforms harboring splice variations in either the IS6 segment or the C terminus. Unexpectedly, while inclusion of exon 8a or 8, which code for IS6, did not alter dihydropyridine (DHP) sensitivity, distinct pharmacologic properties were observed for the various C-terminal splice isoforms. In the presence of external Ca2+, fast inactivating splice variants including CaV1.342a and CaV1.343s with intact calmodulin-IQ domain interaction showed consistently low DHP sensitivity. Interestingly, attenuation of calcium-dependent inactivation with overexpression of calmodulin34 did not enhance the sensitivity of CaV1.342a, suggesting that the low DHP sensitivity may not be a result of fast channel inactivation. Alternatively, disruption of calmodulin-IQ domain binding in the CaV1.3Δ41 and full-length CaV1.342 channels was associated with heightened DHP sensitivity. In distinct contrast to the well-known modulation of DHP blockade of CaV1.2 channels, this study has therefore uncovered a novel mechanism for modulation of the pharmacologic properties of CaV1.3 channels through posttranscriptional modification of the C terminus.
Introduction
L-Type voltage-gated calcium channels (CaV) 1.3 contribute to important physiologic roles, including neurotransmitter release (Brandt et al., 2005), neuronal excitability (McKinney et al., 2009), neural development (Hirtz et al., 2011; Satheesh et al., 2012), and pacemaking in neurons and the sino-atrial node of the heart (Pennartz et al., 2002; Mangoni et al., 2003; Chan et al., 2007; Baig et al., 2011). The CaV1.3 transcripts undergo post-transcriptional modifications including alternative splicing and A-to-I RNA editing, generating channel isoforms with different electrophysiologic or pharmacologic properties (Xu and Lipscombe, 2001; Klugbauer et al., 2002; Shen et al., 2006; Singh et al., 2008; Bock et al., 2011; Tan et al., 2011; Huang et al., 2012).
While an initial study (Xu and Lipscombe, 2001) did not reveal functional changes associated with alternative splicing in the I-II loop or IVS3-IVS4 linker regions, altered channel properties due to alternative splicing in the carboxyl terminus have been reported (Shen et al., 2006; Singh et al., 2008; Liu et al., 2010; Bock et al., 2011; Tan et al., 2011). Splice variations in the C terminus strongly influence calcium-dependent inactivation (CDI), voltage-dependence of activation, surface expression, and current density.
CDI is a negative feedback mechanism triggered by binding of Ca2+ ions to the bilobed calcium sensor, calmodulin (CaM), to initiate a series of conformational changes leading to allosteric inhibition of channel opening (Peterson et al., 1999; Zuhlke et al., 1999; Pan and Lipscombe, 2000; Pitt et al., 2001; Erickson et al., 2003; Mori et al., 2004; Dick et al., 2008; Tadross et al., 2008; Johny et al., 2013). The IQ domain serves not only as the initial binding site for Ca2+-free CaM (apoCaM), but also contributes to a tripartite complex (IQ, C-lobe of Ca2+/CaM, and EF region of channel C terminus) that triggers a C-lobe component of CDI in L-type channels (Johny et al., 2013).
The strength of CDI is modulated by splice variation in the C terminus of CaV1.3 channels (Fig. 1). CDI is weak in the full-length CaV1.342 channels (Fig. 1A), and attenuation of CDI requires the presence of both a proximal C-terminal regulatory domain (PCRD), located just downstream of IQ domain, and a distal C-terminal regulatory domain (DCRD), at the distal end (Wahl-Schott et al., 2006; Singh et al., 2008; Liu et al., 2010). The DCRD functions in an enzyme-inhibitor–like manner that is actively bound to and competes with calmodulin for the binding to the preIQ-IQ domain (Liu et al., 2010). Recently, we and others reported a novel CaV1.3Δ41 variant that has the entire exon 41 that codes for the IQ domain alternatively excluded (Fig. 1E), with the likely disruption of the high-affinity interaction between apoCaM and the carboxy terminus (Erickson et al., 2003; Van Petegem et al., 2005), and thereby abolishing CDI (Tan et al., 2011).
By comparison, selective inclusion of exon 42a instead of exon 42 yields the CaV1.342a channel isoform, featuring early termination of the C terminus, which now harbors only six amino acids immediately after exon 41 (Fig. 1C). This isoform exhibits robust CDI. However, truncation of exon 43 due to the alternative use of splice donor site within exon 43 yielded CaV1.343s channels, which terminate shortly after the PCRD domain (Fig. 1D). High-affinity binding of apoCaM to both CaV1.342a and CaV1.343s channel isoforms (Erickson et al., 2003; Liu et al., 2010) occurs in the absence of the downstream DCRD domain; consequently CDI is strong in both of these contexts (Singh et al., 2008; Tang et al., 2011).
While there is a paucity of information on the modulation of the pharmacologic properties of CaV1.3 channels via alternative splicing, the pharmacology of homologous CaV1.2 channels is, however, strongly influenced by alternative splicing. Notably, alternative splicing of the mutually exclusive exons 8a and 8, which code for the IS6 segment of the CaV1.2 channel, resulted in differential sensitivity toward dihydropyridine (DHP), with the exon 8a splice isoform having less sensitivity toward isradipine (Zuhlke et al., 1998). In addition, alternative splicing in the C terminus of CaV1.2 channels was also known to drastically affect their DHP sensitivity (Zuhlke et al., 1998). In this study, we examined sensitivities toward different DHP blockers among various CaV1.3 channel isoforms harboring splicing variations, particularly in the IS6 segment and C terminus. Unexpectedly, while inclusion of either mutually-exclusive exon 8 or 8a did not alter DHP sensitivity, C-terminal splicing produced opposing effects on nimodipine or nifedipine inhibition. This study thus identifies a novel mechanism whereby the pharmacologic properties of the CaV1.3 channels were modulated by alternative splicing in the C terminus of the channel.
Materials and Methods
Correction of G244S and A1104V Substitutions and Generation of Recombinant CaV1.3 Constructs.
Pairs of primers were used to amplify regions spanning exon 5 (forward primer: 5′-AAGAAACAGAAGGCGGGAACCACT-3′, reverse primer: 5′-GGAAGGAGTGGCTGCGGATGG-3′) and exon 27 (forward primer: 5′-AGCGCATTCTTCATTCTTAGCA-3′, reverse primer: 5′- TCCCCACGGTTACCTCATCATCAC-3′), respectively, using rat brain cDNA libraries (Marathon-Ready cDNA, catalog numbers 639412; Clontech, Mountain View, CA) as template. The polymerase chain reaction (PCR) products were gel-purified and sent for direct DNA sequencing. The PCR amplicons from the rat brain cDNA library were subcloned into pGEM-T Easy vector and transformed into DH10B Escherichia coli cells. To engineer S244G and V1104A substitutions, PCR amplicons containing G244 or A1104 were then substituted into CaV1.3A2123V, CaV1.3Δ41, and CaV1.343s (Tan et al., 2011) in addition to CaV1.342a (Xu and Lipscombe, 2001) to generate the corrected CaV1.342 CaV1.3Δ41, CaV1.343s, and CaV1.342a constructs, respectively, by restriction digestion, followed by ligation. PCR amplicons containing exon 8a were substituted into the corrected CaV1.342 to generate the CaV1.38a_42 construct.
Electrophysiological Recordings and Data Analysis.
Whole-cell patch-clamp electrophysiologic recordings were used to characterize the recombinant CaV1.3 constructs. Ba2+ current (IBa) or Ca2+ current (ICa) were recorded from transiently transfected mammalian human embryonic kidney 293 cells at room temperature according to methods described previously (Patil et al., 1998; Peterson et al., 1999). Outward IK currents were blocked by Cs+ in the internal and external solutions. Cells were transiently transfected with the respective CaV1.3 constructs with rat β2a subunit and rat α2δ subunit using the standard calcium phosphate transfection method. IBa or ICa was recorded at room temperature with 5 mM Ba2+ or Ca2+ as charge carrier, 48–72 hours after transfection. For whole-cell patch-clamp recording, the internal solution (patch-pipette solution) contained the following (in mM): 138 Cs-MeSO3, 5 CsCl, 5.0 EGTA, 10 HEPES, 1 MgCl2, 2 mg/ml Mg-ATP, pH 7.3 (adjusted with CsOH), 290 mOsm with glucose. The external solution contained the following (in mM): 10 HEPES, 140 tetraethylammonium methanesulfonate, 5 BaCl2, or 5 CaCl2 (pH adjusted to 7.4 with CsOH and osmolality to 290–310 with glucose). Pipettes of resistance 1.5–2.0 MΩ were used. Whole-cell currents, obtained under voltage clamp with an Axopatch 200B amplifier (Molecular Devices, Union City, CA), were filtered at 1–5 kHz and sampled at 5–50 kHz, and the series resistance was typically <5 MΩ after >80% compensation. A P/4 protocol was used to subtract online the leak and capacitive transients.
The protocols for recording current-voltage (I-V) and ramp current have been described previously (Shen et al., 2006). Briefly, I-V curve relationships were obtained by step depolarization from a holding potential of −90 mV to various test potentials from −60 to 50 mV over 500 milliseconds. The sweep intervals are 30 seconds, to allow fully recovery from channel inactivation. On the other hand, ramp currents were recorded by ramp voltage from −80 to 50 mV over duration of 300 milliseconds with sweep intervals of 20 seconds.
Data were acquired using the software pClamp9 (Molecular Devices), and analyzed and fitted using Graphpad Prism V software (San Diego, CA) and Microsoft (Seattle, WA) Excel. Data are expressed as mean values ± S.E.M. Statistical analysis was performed using unpaired Student’s t test. I-V curves were fitted according to Eq. 1: I = Gmax(V − Erev)/(1 + exp [(V − V1/2act)/kact]), where Gmax is the maximum conductance of the cell, Erev is the reversal potential, V1/2act is the voltage for half-maximal activation, kact is the slope of Boltzmann function. To obtain the IC50 values, the maximal current for each cell after drug treatment was compared against maximal current prior to drug treatment before being averaged. The percentage of inhibition for each cell at respective nimodipine concentration was then fit to the sigmoidal dose-response equation: Y = Mininh + (Maxinh − Mininh)/(1 + 10^[LogIC50 − X]), where LogIC50 is the logarithm of the IC50 (inhibition concentration, 50%). Maxinh is the maximal percentage of inhibition at the top plateau. Mininh is the minimal percentage of inhibition at the bottom plateau.
Drug Preparation.
Stock solutions were prepared by dissolving nimodipine (RBI, Natick, MA) and nifedipine (Sigma-Aldrich, St. Louis, MO) in dimethylsulfoxide to make a 10-mM stock solution and stored at −20°C in the dark. Respective concentrations were freshly prepared in the bath solution from stock, and perfused (1 ml/min) into the whole recording chamber by gravity during current recording. The cells selected for patch-clamp recording were positioned as close to the outlet as possible. The nimodipine and nifedipine solutions were protected from light throughout the experiment.
Results
Altered Gating Properties Due to C-Terminal Alternative Splicing of CaV1.3 Channels.
Taking hints from recent studies (Tan et al., 2011; Lieb et al., 2012) reporting several cloning errors within the original rat CaV1.342 clones, all CaV1.3 channel isoforms used in this study were corrected by S244G, V1104A, and A2123V substitutions to avoid any potential confounding effects prior to electrophysiologic and pharmacologic experiments. Reassuringly, characterization of the corrected CaV1.342 clone revealed channels with slowed CDI (Fig. 2A, top panel), due to the restoration of DCRD activity as a result of the A2123V change (Liu et al., 2010; Tan et al., 2011). Quantitatively, the rates of inactivation of the respective Ba2+ or Ca2+ currents were indicated by an r50 value, which is the ratio of current remaining after 50 milliseconds of depolarization. For CaV1.342 channels, the Ba2+ r50 values were close to 1.0 across the voltage range from −40 to 40 mV, indicating little voltage-dependent inactivation. Alternatively, CDI was reflected in the gentle decline of the Ca2+ r50 relation, which exhibited a shallow U-shaped voltage dependence (Fig. 2A, second panel). Pure CDI was quantified by the f-index, calculated from the difference in r50 measured in Ba2+ and Ca2+ traces at −10 mV (CaV1.342: f = 0.16 ± 0.01). In addition, the f-index was also measured for other time points, including 100, 300, and 500 milliseconds, as summarized in Table 3. On the other hand, the substantially right-shifted V1/2 act values of IBa (−17.43 ± 0.48 mV) and ICa (−7.41 ± 0.67 mV) as compared with the published values of −24.33 ± 0.49 mV and −12.28 ± 0.53 mV, respectively, for the A2123V change alone (Tan et al., 2011) could be attributed to the additional S244G substitution (Lieb et al., 2012).
Direct comparison of the electrophysiologic properties of CaV1.38a_42 channels containing exon 8a (Fig. 1) with CaV1.342 channels containing exon 8 revealed similar inactivation and I-V relationships (Fig. 2B; Tables 1–3). These results clearly showed that alternative splicing in the IS6 segment did not alter the basic biophysic properties of CaV1.3 channels.
In contrast, the CaV1.342a channel isoform displayed drastically faster CDI and left-shifted I-V relationships, as compared with the full-length CaV1.342 channels (Fig. 2C). Furthermore, CaV1.343s channels (which retain only the PCRD domain) similarly supported rapid CDI and negative I-V relationships, as did the CaV1.342a channels (Fig. 2D). Lastly, characterization of CaV1.3Δ41 revealed further reduction of CDI (Fig. 3D, top two panels). The residual CDI (Table 3) could be mediated through the interaction of the N-lobe of calmodulin with the N-terminal spatial Ca2+ transforming element (NSCaT domain at the N terminus of CaV1.3 channels) (Dick et al., 2008). In addition, significant left-shifted I-V relationships for both IBa and ICa of CaV1.3Δ41 isoform as compared with the full-length CaV1.342 channels were also observed (Fig. 3D, bottom two panels; Tables 1 and 2). Therefore, the S244G and V1104A substitutions did not overtly affect the predicted functional outcomes associated with C-terminal alternative splicing.
Altered DHP Sensitivity Due to Alternative Splicing of the C Terminus.
After the electrophysiologic properties of the corrected CaV1.3 channels were characterized, their pharmacologic properties were subsequently investigated. Traditionally, single square pulse protocols at different depolarizing potentials were often used to access the DHP sensitivity of L-type calcium channel. However, measuring the extent of DHP inhibition at any particular voltage could be confounded by the fact that DHP channel blockers also shift the I-V relationships. Indeed, comparison of I-V plots before and after treatment with 1.0 μM of nimodipine revealed varying extents of hyperpolarizing shifts in the V1/2 act values of the CaV1.3 channel isoforms (Supplemental Fig. 1). To complicate the issue further, inhibition of CaV1.3 channels by DHP has been shown to be voltage dependent with generally increasing sensitivity at higher depolarizing voltages (Xu and Lipscombe, 2001). Therefore, given the widely different I-V profiles displayed by the CaV1.3 splice isoforms within the study, a different approach was taken to investigate drug sensitivity, whereby an I-V protocol was first recorded before drug treatment followed by an ensemble of ramp currents during drug perfusion to monitor the extent of current inhibition. Finally, another I-V protocol was recorded after the equilibrium inhibition was reached (Fig. 3, A and B). The maximal peak current for each cell after drug treatment was compared against the maximal peak current prior to drug treatment to calculate the percentage inhibition at the respective drug concentrations. Fitting the CaV1.342 concentration-response curve with the Sigmoidal dose-response equation revealed an IC50 value of 0.67 μM for nimodipine inhibition (Fig. 3C; Table 4). Interestingly, despite the high sequence conservation of exon 8a and 8 between CaV1.2 and CaV1.3 channels (Supplemental Fig. 2), CaV1.38a_42 channels (IC50 = 0.95 μM) were inhibited to a similar extent by nimodipine, as compared with CaV1.342 channels (Fig. 4A; Supplemental Fig. 3).
As C-terminal alternative splicing gives rise to channel splice variants with distinct biophysical properties, we wondered whether DHP sensitivities of CaV1.3 channel splice variants could also be different. Given the faster inactivation kinetics of CaV1.342a and CaV1.343s channels, it would be expected that both channel variants would be more sensitive to nimodipine inhibition, as DHP molecules have been shown to bind preferentially to the inactivated channels (Bean, 1984). Unexpectedly, we found that both channel isoforms displayed much weaker sensitivity toward nimodipine as compared with the CaV1.342 channel (Fig. 4, B and C; Supplemental Figs. 4 and 5), offering the first indication that DHP sensitivity could be modulated by alternative splicing of the C terminus. The IC50 values of CaV1.342a (2.75 μM) and CaV1.343s (1.73 μM) were 4.1- and 2.6-fold larger than that of the full length CaV1.342 channels respectively (Table 4). In comparison, subsequent investigation of CaV1.3Δ41 isoform (IC50 = 0.36 μM) revealed however greater sensitivity toward nimodipine as compared with CaV1.342 channels (Fig. 4D; Supplemental Fig. 6; Table 4).
Alternatively, additional experiments were performed using a single square pulse protocol to assess both the run-down and extent of DHP inhibition at 0.5 μM nimodipine (Supplemental Fig. 13). For CaV1.342a and CaV1.343s, a 300 millisecond square pulse at 0 mV from holding potential of −90 mV was used while square pulse at 10 mV was used for CaV1.342 and CaV1.3Δ41 channels. The interval for each pulse was 30 seconds to allow for full recovery from channel inactivation. The effect of run-down was within 10% for CaV1.342a, CaV1.343s, and CaV1.342 at the 20th sweep and the effect of run-down was not significantly different comparing either CaV1.342a or CaV1.343s with CaV1.342, while approximately 10% of run-down was observed for CaV1.3Δ41 and it was significantly different from CaV1.342. On the other hand, upon treatment with 0.5 μM nimodipine, similar trend of DHP sensitivity was observed: CaV1.342a and CaV1.343s were weakly inhibited by nimodipine as 15.45 ± 1.81% and 15.1 ± 2.68% of their currents were inhibited respectively. In comparison, 36.04 ± 4.45% of the CaV1.342 current and 47.73 ± 3.18% of CaV1.3Δ41 current were inhibited. The difference among CaV1.342a, CaV1.343s and CaV1.3Δ41 and CaV1.342 was significant with P < 0.0001 (one-way analysis of variance and Bonferroni test). In addition, the difference of DHP sensitivity among the different channel splice variants could be repeatedly confirmed with 1.0 and 10.0 μM of nifedipine (Fig. 5). While CaV1.38a_42 displayed a similar extent of inhibition as CaV1.342, CaV1.342a and CaV1.343s were less sensitive, and CaV1.3Δ41 showed heightened sensitivity at both concentrations.
DHP Sensitivity Is Not Determined by the Kinetics of Channel Inactivation.
As the degree of DHP sensitivity (CaV1.3Δ41 > CaV1.342 > CaV1.343s > CaV1.342a) appeared to be inversely correlated with the rate of calcium-dependent inactivation (CaV1.342a > CaV1.343s > CaV1.342 > CaV1.3Δ41; Fig. 2; Table 3), it was suggested at first that the rate of channel inactivation could serve as a predictor of DHP sensitivity. Calcium-dependent regulation in CaV1 and CaV2 channels relies on calmodulin molecules that preassociate with the C-terminal preIQ-IQ domain. Although both lobes of calmodulin are able to mediate CDI independently in CaV1.3 channels, C-lobe regulation through the C-terminal IQ domain appeared to be dominant, as close to 90% of CDI was eliminated upon coexpression with a mutant calmodulin, CaM34, that contained a Ca2+ insensitive C-lobe (Yang et al., 2006).
To investigate the relationship between C-lobe–mediated CDI and the DHP sensitivity of the channel, we coexpressed CaV1.342a with mutant calmodulin CaM34. As expected, overexpression of CaM34 selectively isolated a slower, N-lobe component of CDI (Fig. 6A, top two panels), as corroborated by reduction of the f50 index of rapid CDI to 0.08 ± 0.02, as compared with 0.72 ± 0.01 when CaV1.342a was expressed alone. The I-V relationships for both IBa and ICa were, however, not significantly altered (Fig. 6A, bottom two panels). Interestingly, despite the substantial slowing of CDI, overexpression of CaM34 had little effect on the DHP sensitivity of CaV1.342a channels (Fig. 6B; Supplemental Fig. 7; Table 4). The data therefore suggested that the weaker DHP sensitivity observed for the CaV1.342a isoform did not result from the fast calcium-dependent inactivation. Furthermore, CaV1.342a and CaV1.343s channels were also more weakly inhibited by DHP as compared with CaV1.342 and CaV1.3Δ41 channel displayed the highest sensitivity when nimodipine sensitivity was assessed in 5 mM Ba2+ (Supplemental Fig. 8). Taken together, our data showed that binding of calmodulin to the IQ-domain, rather than calcium-dependent inactivation property, influenced DHP-sensitivity in CaV1.3 channels. Furthermore, we wondered whether differences in kinetics of voltage-dependent inactivation (VDI) could serve to predict DHP sensitivity. The Ba2+ currents were recorded using prolonged 15s voltage pulses to respective voltages that elicited maximal currents (Supplemental Fig. 12). Essentially, the onset of VDI was indeed different among CaV1.342, CaV1.342a, CaV1.343s, and CaV1.3Δ41 channel isoforms; the r500 (fraction of current remaining at 500 milliseconds) and r1000 values of CaV1.342a and CaV1.343s were significantly higher than those of the CaV1.342 and CaV1.3Δ41 channels. However, the differences between CaV1.342a and CaV1.342 at 5, 10, and 15 seconds were not significant.
Discussion
The CaV1.3 channels have been known to be much less sensitive to DHP inhibition than the paralogous CaV1.2 channels (Xu and Lipscombe, 2001). Both channels are subjected to extensive alternative splicing in a spatiotemporal manner to customize different channel properties for various physiologic roles. While the pharmacologic properties of CaV1.2 channels can be modulated by alternative splicing, little is known about CaV1.3 channels in this regard. The data here revealed that alternative splicing, specifically within the C terminus of the CaV1.3 channel, generated a spectrum of splice isoforms that exhibited varying sensitivities toward DHP, with ∼8-fold difference between the most sensitive CaV1.3Δ41 and the least sensitive CaV1.342a channels.
Interestingly, Zuhlke et al. previously reported two C-terminal splice isoforms of the CaV1.2 channel, α1C,86 and α1C,72, that were substantially more sensitive to isradipine than the wildtype, α1C,77 (Zuhlke et al., 1998). Upon close examination of their respective sequences, the amino acid sequence from 1572 to 1651, which comprises a substantial segment of preIQ and the entire IQ domain, was replaced with 81 entirely different amino acids in the α1C,86 channel, and insertion of 19 amino acids at the position of 1575 in the α1C,72 channel would disrupt the preIQ domain (Soldatov et al., 1995). Considering the result from this report and the work by Zuhlke et al. (1998), it would be reasonable to conclude that calmodulin tethering to the C-terminal preIQ-IQ domain is a conserved molecular mechanism that regulates the DHP sensitivities of both CaV1.2 and CaV1.3 channels.
Despite the seeming inverse correlation between DHP sensitivity and the calcium-dependent inactivation kinetics of the channel isoforms, attenuation of the CDI of CaV1.342a channels upon overexpression of mutant CaM34 did not enhance its sensitivity toward nimodipine, suggesting that the low DHP sensitivity of the channel is not modulated by fast calcium-dependent inactivation kinetics.
In comparison, the onset of VDI, especially within 1 second of prolonged depolarization (CaV1.3Δ41 > CaV1.342 > CaV1.343s > CaV1.342a) appeared also to correlate positively with the DHP sensitivity of CaV1.3 channels (CaV1.3Δ41 > CaV1.342 > CaV1.343s > CaV1.342a). However, the fast kinetics of VDI are not always associated with more pronounced block. Taking the α1C,86 and α1C,72 constructs as reported by Zuhlke et al. (1998) as an example, while deletion of entire IQ domain in α1C,86 resulted in substantially faster VDI and stronger DHP sensitivity, resembling the observation with CaV1.3Δ41 channel, α1C,72 construct, which displayed similar enhancement of DHP inhibition, does not express a similar increase in the VDI kinetic (Zuhlke et al., 1998). In addition, despite the dramatic effect of different β-subunits on gating properties of the L-type calcium channels (Berjukow et al., 2000; Takahashi et al., 2003), it was shown that different β-subunits did not overtly affect DHP sensitivity (Berjukow et al., 2000). Hence, although it seems that differences in the onset of VDI in the four C-terminal splice variants indeed correlated well with their respective DHP sensitivities, whether faster VDI can influence DHP sensitivity is still unresolved.
Furthermore, the V1/2,inact values of CaV1.342, CaV1.342a, CaV1.343s, and CaV1.3Δ41 channels derived from steady-state-inactivation current traces were compared in Supplemental Fig. 10. Upon plotting the V1/2,inact values against the previously measured IC50 values, an inverse correlation was obtained with an R2 value of 0.7472. However, caution should be taken when interpreting this result. Although CaV1.3Δ41 is the most sensitive to DHP, its V1/2,inact values are significantly more hyperpolarized than those of the CaV1.342 channels. In addition, previous studies were rather divided regarding if the V1/2,inact values could predict DHP sensitivity. While Berjukow et al. showed that there was no direct correlation between the V1/2,inact values and sensitivity of CaV1.2 channels to isradipine (Berjukow et al., 2000), data from Liao and colleagues, on the other hand, seemed to support this link (Liao et al., 2007). Lastly, DHP sensitivity may not be closely correlated with current density, as shown in Supplemental Fig. 11. Although the CaV1.343s splice variant only displayed current density around about one-third of CaV1.342a channels, their DHP sensitivity is consistently similar.
Alternatively, we hypothesized that strong association of calmodulin with IQ domain allosterically predisposes the channel into a state that has low DHP binding affinity (Fig. 7). As calcium-insensitive mutant calmodulin retains high binding affinity toward IQ domain (Liang et al., 2003), CaM34 has little effect on the drug sensitivity of the channels, even though it slowed CDI. Moreover, such a model could also explain the weaker sensitivity of both CaV1.342a and CaV1.343s channels, where calmodulin-IQ interaction is robust (Erickson et al., 2003; Liu et al., 2010). The stronger DHP sensitivity of CaV1.343s as compared with CaV1.342a channels could perhaps be attributed to slight alteration of calmodulin-IQ interaction with the presence of PCRD domain (Liu et al., 2010). To further strengthen our views, attenuated IQ-calmodulin interaction due to the presence of competing DCRD domain in the context of full-length CaV1.342 or complete deletion of IQ domain in CaV1.3Δ41 channels produced heightened DHP sensitivities.
Apart from the mechanistic insights garnered from these results, the presence of CaV1.3 isoforms with different DHP sensitivities has several important physiologic implications. Firstly, the presence of highly sensitive CaV1.3Δ41 or CaV1.342 isoforms suggests that it is not ideal to use low (1.0 μM) or high (10.0 μM) DHP concentrations to fully segregate CaV1.2 and CaV1.3 currents, as previously thought. Secondly, the expression of different CaV1.3 channel variants with different gating properties may implicate distinct physiologic roles that could be targeted differently with different DHP concentrations. The CaV1.3 currents have been shown to be important for regulating pacemaking activities in neurons such as substantia pars compacta and suprachiasmatic nucleus (Pennartz et al., 2002; Chan et al., 2007; Huang et al., 2012). The CaV1.342a and CaV1.343s currents that activate closer to resting membrane potential provide pace-making current that is more resistant to DHP inhibition, while the more long-lasting CaV1.3Δ41 or CaV1.342 current could be important for maintaining subsequent regular firing of neurons or tonic influx of calcium currents that are more susceptible to lower DHP concentration.
Moreover, inhibiting pace-making CaV1.3 current in the substantia pars compacta neurons has been recently suggested to be protective against the pathogenesis of Parkinson’s disease (Chan et al., 2007). However, the nonselective blockade of L-type calcium channels limits the use of DHP directly for such therapeutic purposes, as inhibition of both CaV1.2 and CaV1.3 channels could result in serious neurologic disturbances. An exciting recent study (Kang et al., 2012) reported a highly selective compound, 1-(3-chlorophenethyl)-3-cyclopentylpyrimidine-2,4,6-(1H,3H,5H)-trione, against CaV1.3 channels, and the drug is currently undergoing preclinical trials. However, it should be noticed that the use of the uncorrected CaV1.342 (Genbank accession no. AF370010) could possibly confound the findings (Kang et al., 2012). Furthermore, previous studies have shown complex post-transcriptional modification of CaV1.3 channels in rat substantia nigra neurons; close to 30% of CaV1.3 transcripts contains the mutually exclusive exon 42a. In addition, CaV1.343s or CaV1.3Δ41 transcripts exist in 44 and 4.6% of the total CaV1.3 transcripts screened in the mouse brain (Tan et al., 2011). Identification of the predominant tissue-selective CaV1.3 isoforms in the future, and subsequent examination of the effects of the newly found drug on such channel isoforms could hopefully offer a better prediction of its dosage requirement and therapeutic outcome.
Acknowledgments
The authors thank Dr. Diane Lipscombe (Brown University, Province, RI) for the rat CaV1.342 and CaV1.342a cDNA, Dr. Terry P. Snutch (University of British Columbia, Vancouver, BC, Canada) for the rat β and rat α2δ cDNA, and Dr. David Yue (Johns Hopkins University, Baltimore, MD) for the CaM34 cDNA construct. The authors also thank Dr. David Yue for invaluable advice on the manuscript.
Authorship Contributions
Participated in research design: Huang, Soong.
Conducted experiments: Huang, Yu.
Wrote or contributed to the writing of the manuscript: Huang, Soong.
Footnotes
- Received May 6, 2013.
- Accepted August 7, 2013.
The work was supported by the Singapore Biomedical Research Council [10/1/21/19/655].
↵This article has supplemental material available at molpharm.aspetjournals.org.
Abbreviations
- CaM
- calmodulin
- CaV
- voltage-gated calcium channels
- CDI
- calcium-dependent inactivation
- DCRD
- distal C-terminal regulatory domain
- DHP
- dihydropyridine
- IBa
- Ba2+ current
- ICa
- Ca2+ current
- I-V
- current-voltage
- PCR
- polymerase chain reaction
- PCRD
- proximal C-terminal regulatory domain
- VDI
- voltage-dependent inactivation
- Copyright © 2013 by The American Society for Pharmacology and Experimental Therapeutics