Abstract
The L-type calcium channel (LTCC) isoforms Cav1.2 and Cav1.3 display similar 1,4-dihydropyridine (DHP) binding properties and are both expressed in mammalian brain. Recent work implicates Cav1.3 channels as interesting drug targets, but no isoform-selective modulators exist. It is also unknown to what extent Cav1.1 and Cav1.4 contribute to L-type-specific DHP binding activity in brain. To address this question and to determine whether DHPs can discriminate between Cav1.2 and Cav1.3 binding pockets, we combined radioreceptor assays and quantitative polymerase chain reaction (qPCR). We bred double mutants (Cav-DM) from mice expressing mutant Cav1.2 channels [Cav1.2DHP(-/-)] lacking high affinity for DHPs and from Cav1.3 knockouts [Cav1.3(-/-)]. (+)-[3H]isradipine binding to Cav1.2DHP(-/-) and Cav-DM brains was reduced to 15.1 and 4.4% of wild type, respectively, indicating that Cav1.3 accounts for 10.7% of brain LTCCs. qPCR revealed that Cav1.1 and Cav1.4 α1 subunits comprised 0.08% of the LTCC transcripts in mouse whole brain, suggesting that they cannot account for the residual binding. Instead, this could be explained by low-affinity binding (127-fold Kd increase) to the mutated Cav1.2 channels. Inhibition of (+)-[3H]isradipine binding to Cav1.2DHP(-/-) (predominantly Cav1.3) and wild-type (predominantly Cav1.2) brain membranes by unlabeled DHPs revealed a 3- to 4-fold selectivity of nitrendipine and nifedipine for the Cav1.2 binding pocket, a finding further confirmed with heterologously expressed channels. This suggests that small differences in their binding pockets may allow development of isoform-selective modulators for LTCCs and that, because of their very low expression, Cav1.1 and Cav1.4 are unlikely to serve as drug targets to treat CNS diseases.
Depolarization-induced Ca2+ entry through voltage-gated Ca2+ channels controls a number of important physiological properties, including muscle contraction, cardiac pacemaking, hormone secretion, neurotransmitter release, and neuronal plasticity. Several subfamilies of voltage-gated Ca2+ channels with different pharmacological and biophysical properties evolved to serve these different physiological functions. Of these, the family of L-type Ca2+ channels (LTCCs) is characterized by high sensitivity to organic Ca2+ channel blockers (Catterall et al., 2005). Four LTCC isoforms exist, containing either Cav1.1, Cav1.2, Cav1.3, or Cav1.4 pore-forming α1 subunits (Catterall et al., 2005). Cav1.2 and Cav1.3 LTCCs display very similar pharmacological properties, and they exhibit an overlapping expression pattern in the cardiovascular system, endocrine cells (e.g., pancreatic β-cells), and neurons (Hell et al., 1993; Sinnegger-Brauns et al., 2004). This has impeded distinction of their respective physiological roles. We have recently generated suitable mouse models to distinguish their contribution to different physiological processes. Whereas Cav1.2 channels regulate vascular tone and cardiac inotropy, Cav1.3 channels are crucial for pacemaking in sinoatrial node cells and atrio-ventricular conduction (Platzer et al., 2000; Sinnegger-Brauns et al., 2004). In mice, fast insulin secretion is closely coupled to Cav1.2 function (Schulla et al., 2003; Sinnegger-Brauns et al., 2004), whereas inner hair cell signaling and hearing depend on Cav1.3 (Platzer et al., 2000).
The central nervous system employs Cav1.2 channels for late-phase long-term potentiation in hippocampal neurons and for spatial memory (Moosmang et al., 2005), whereas Cav1.3 channels, which can activate at more negative voltages (Platzer et al., 2000; Koschak et al., 2001; Lipscombe et al., 2004), seem to serve predominantly as modulators of neuronal spiking behavior and pacemaking (Olson et al., 2005; Chan et al., 2007).
Cav1.3-deficient mice lack amphetamine-sensitized loco-motor activity (Giordano et al., 2006) and exhibit an anti-depressant-like phenotype (Striessnig et al., 2006). Cav1.3 channels may also play an important pathophysiological role in Parkinson disease (Chan et al., 2007). Because presently available potent Ca2+ channel blockers such as the 1,4-dihydropyridines (DHPs) inhibit both Cav1.2 and Cav1.3 (Koschak et al., 2001; Helton et al., 2005), development of selective modulators of Cav1.3 channels would be required to determine whether they represent suitable targets to treat Parkinson disease, depression, and drug abuse. At present, it is unclear whether the highly conserved drug binding pockets (such as for the DHPs) of Cav1.2 and Cav1.3 channels offer enough structural heterogeneity to be exploited for that purpose. It also is still unresolved to what extent other LTCC isoforms (Cav1.1 and Cav1.4) can contribute to high-affinity drug binding in the central nervous system (Takahashi et al., 2003; Hemara-Wahanui et al., 2005) and could thus mediate pharmacological effects. These considerations raise two questions: 1) What are the expression levels of other LTCC isoforms (Cav1.1 and Cav1.4) in brain? 2) Does structural heterogeneity between the DHP binding pockets of Cav1.2 and Cav1.3 channels result in affinity differences that could be exploited for the development of isoform-selective modulators? This also raises a methodological issue, because although Cav1.3 channel complexes can be expressed for pharmacological analysis in mammalian cells, a native source of Cav1.3 channels is needed to validate data obtained with recombinant channels.
We employed a biochemical approach to address these questions, using a novel double-mutant mouse generated from previously established mouse models, radioligand binding assays, and extensive qPCR experiments.
Materials and Methods
Animals. The generation and detailed characterization of Cav1.2DHP(-/-) and Cav1.3(-/-) mice has been reported previously (Platzer et al., 2000; Sinnegger-Brauns et al., 2004). Double mutant Cav1.2DHP(-/-) × Cav1.3(-/-) mice were generated by crossing Cav1.3(-/-) male mice with Cav1.2DHP(-/-) female mice previously backcrossed into C57BL/6N background in five generations. Animals heterozygous for both mutant alleles were then crossed to obtain homozygous double mutants (referred to as Cav-DM). Tissues and RNA from 3- to 16-month old male mice or 3- to 12-month old male Sprague-Dawley rats were used for experiments. Animal handling was approved by the Bundesministerium für Bildung, Wissenschaft und Kultur (Vienna, Austria) in accordance with international laws and policies of animal welfare.
Recombinant α1 cDNAs. The following α1 subunit cDNAs were used (GenBank accession numbers given in parentheses): rabbit Cav1.2 (X15539), human Cav1.3 (EU63339; Koschak et al., 2001), and human Cav1.4 (variant of NM_005183; Koschak et al., 2003). Replacement of phenylalanine by tyrosine at position 1414 in the Cav1.4 α1 subunit (numbering according to Koschak et al., 2003) to yield Cav1.4F1414Y followed published PCR procedures (Hoda et al., 2005), using an AgeI/EcoRI cassette. The mutation was confirmed by DNA sequencing (MWG Biotech, Ebersberg, Germany).
Transfection and Cell Culture. tsA201 cells were maintained in DMEM/F-12 medium (Invitrogen, Carlsbad, CA) enriched with 10% (v/v) fetal calf serum (SEBAK, Suben, Austria) and 2 mM l-glutamine and containing 44 mM NaHCO3. Cells were incubated at 5% CO2/37°C, plated on 10-cm culture dishes, and grown to approximately 70% confluence for transfection. α1 subunits were transiently expressed together with α2δ1 (Genbank accession number NM_001082276) and β1a (M25817) subunits in pcDNA3 (Invitrogen). Expression plasmids were combined at equimolar concentrations with a total subunit cDNA amount of 10.5 μg per 10-cm culture dish. Cells were transfected by calcium phosphate precipitation of DNA according to standard protocols. Medium was refreshed 8 to 12 h after transfection. Cells were harvested after two more days of incubation.
Membrane Preparation and Immunoblot Analysis. Microsomal membranes from brain tissue and tsA201 cells were prepared as described previously (Glossmann and Ferry, 1985; Reimer et al., 2000; Sinnegger-Brauns et al., 2004). Membrane protein concentrations were determined by Lowry or Bradford assays (Lowry et al., 1951; Bradford, 1976) with bovine serum albumin as a standard. Expression of recombinant proteins was monitored using Western Blot analysis as described previously (Hoda et al., 2005).
Radioreceptor Assay. (+)-[3H]Isradipine (75-80 Ci/mmol) was purchased from PerkinElmer Life and Analytical Sciences (Vienna, Austria). Unlabeled racemic isradipine and (-)-isradipine were a gift from Novartis (Basel, Switzerland). (+)-Tetrandrine was a gift from MSD Sharp and Dohme (Rahway, NJ). (+)-cis-Diltiazem was from Gödecke (Freiburg, Germany). Mibefradil was provided by Roche (Basel, Switzerland). Azidopine was purchased from GE Healthcare (Chalfont St. Giles, Buckinghamshire, UK). Racemic nitrendipine and amlodipine as well as nifedipine were purchased from Sigma-Aldrich (Vienna, Austria). Stock solutions of all unlabeled drugs were prepared in dimethyl sulfoxide. Binding experiments were performed in binding buffer (50 mM Tris-HCl, pH 7.4, 0.1 mM phenylmethylsulfonyl fluoride, and 1 mM CaCl2) in a final assay volume of 0.5 (brain microsomes) or 1 ml (tsA201 cell microsomes), respectively. Nonspecific binding was determined in the presence of 1 μM unlabeled isradipine. After incubation for 90 to 120 min at 22°C or 60 min at 37°C (binding stimulation) in a water bath, free ligand was removed by rapid filtration of the assay mixture over GF/C Whatman glass fiber filters (Sigma-Aldrich, Vienna, Austria), which had been pretreated with 0.25% (v/v) polyethylenimine for 30 min at 22°C. Filters were washed three times with ice-cold buffer (50 mM Tris-HCl, pH 7.4) and then counted for radioactivity to quantify bound ligand.
RNA Isolation and Reverse Transcription. Animals were euthanized by CO2 exposure, and brains were excised after decapitation. Brain regions were dissected on ice with the help of a stereotaxic atlas (Franklin and Paxinos, 1997). Tissue was shock-frozen in liquid nitrogen and stored at -80 C until further processing. Total RNA was isolated using the RNAqueous-4PCR Kit (Ambion, Austin, TX; for brain tissues) or the E.Z.N.A. Total RNA Kit (Omega Bio-tek, Doraville, GA; for skeletal muscle and eye) according to the manufacturers' instructions. Total RNA was treated with DNase I (RNAqueous-4PCR Kit; Ambion), which effectively removed genomic contamination. RNA integrity was routinely checked for the presence of distinct 28S and 18S rRNA bands after loading 1 to 2 μg of RNA on a denaturating gel (formaldehyde SeaKem agarose gel, 1%). Total RNA (1 μg) was reverse-transcribed using RevertAid H Minus first-strand cDNA Synthesis Kit with random hexamer primers (MBI Fermentas, St. Leon-Rot, Germany). RNA and cDNA concentrations were determined photometrically.
Quantitative Real-Time PCR. The relative abundance of different LTCC transcripts was assessed by TaqMan quantitative PCR using a standard curve method as described previously (Koschak et al., 2007). Specific TaqMan Gene Expression Assays, designed to span exon-exon boundaries, were purchased from Applied Biosystems (Foster City, CA). The following primers (MWG Biotech) were used for PCR amplification of assay-specific fragments using whole brain cDNA as a template. Mouse: Cav1.1: forward, 5′-GTTACATGAGCTGGATCACACAG-3′; reverse, 5′-ATGAGCATTTCGATGGTGAAG-3′.CaV1.2: forward, 5′-CATCACCAACTTCGACAACTTC-3′; reverse, 5′-CAGGTAGCCTTTGAGATCTTCTTC-3′. CaV1.3: forward, 5′-ACATTCTGAACATGGTCTTCACAG-3′; reverse, 5′-AGGACTTGATGAAGGTCCACAG-3′.CaV1.4: forward, 5′-CTCTTCATCTGTGGCAACTACATC-3′; reverse, 5′-GTACCACCTTCTCCTTGGGTACTA. Rat: Cav1.1: forward, 5′-CACACAGGGTAGCATGTAAGAGG-3′; reverse, 5′-TCGATACCCATAATATTCCTCCTG-3′.Cav1.2: forward, 5′-AAGACATAGACCCTGAGAATGAGG-3′; reverse, 5′-GAAGATCACCAGCCAGTAGAAGAC-3′. Cav1.3: forward, 5′-ATTAGGTCTGAGTCAGGAGACGAG-3′; reverse, 5′-TCGTCATCTTCTAAGAAACCTTGG-3′. Cav1.4: forward, 5′-TTCTACAGTTGCACTGATGAAGC-3′; reverse, 5′-AAGGCGATGATGATGATGTAGAC-3′. The integrity of the obtained fragments was confirmed by sequencing (MWG Biotech). Fragment concentrations were determined in a Wallac 1420 Multi-label Counter (PerkinElmer Life and Analytical Sciences) using the Quant-IT PicoGreen dsDNA Reagent (Invitrogen) according to the manufacturer's instructions, and standard curve dilutions were calculated subsequently. Four to five standard curves with 10-fold serial dilutions from 107 to 10 molecules of the respective fragment were generated for each assay. Assay efficiencies were calculated from the slope of the resulting average standard curve equation. Limits of quantification in the linear range of the individual standard curves were determined by calculating the relative S.D. of the replicate values obtained for each amount of fragment. The reliable quantification range was then defined by the fragment quanta at which measurements showed relative S.D. values of less than 1%. Assay IDs, their efficiency (E), and reliable quantification range (molecules) were as follows: Mouse: Cav1.1 α1, Mm00489257_m1, E = 94.9%, 103 to 107;Cav1.2 α1, Mm00437917_m1, E = 93.4%, 102 to 107;Cav1.3 α1, Mm01209919_m1, E = 97.0%, 102 to 107;Cav1.4 α1, Mm00490443_m1, E = 94.3%, 101 to 107. Rat: Cav1.1 α1, customized (target base 2302 according to GenBank accession number L04684), E = 95.8%, 3 × 102 to 2 × 106;Cav1.2 α1, Rn00709287_m1, E = 95.6%, 102 to 107;Cav1.3 α1, Rn01453378_m1, E = 90.1%, 3 × 102 to 107;Cav1.4 α1, Rn00586734_m1, E = 98.1%, 102 to 7 × 106.
qPCR was performed in triplicate measurements using 20 ng of total RNA equivalents of cDNA and the specific TaqMan Gene Expression Assay in a final volume of 20 μl in TaqMan Universal PCR Master Mix (Applied Biosystems). Samples were obtained from at least two independent reverse transcriptions of at least two independent RNA preparations from each species and strain examined (with the exception of muscle and eye controls, where only one reverse transcription of the respective preparations was performed). β-Actin transcript monitoring served to control the quality of RNA preparations and reverse transcriptions (mouse, Mm00607939_s1; rat, Rn01412977_g1). Samples containing RNA and samples without templates served as negative controls. Analysis was performed using the ABI PRISM 7500 Sequence Detector (Applied Biosystems). The relative abundance of each of the four different LTCC RNAs was expressed as the percentage of the sum of LTCC molecules per experiment.
Data Analysis and Statistics. Data were analyzed using Origin 6.1 (OriginLab Corp., Northampton, MA) and Prism 4.03 (GraphPad, San Diego, CA) employing the statistical tests indicated in the text and figure legends. P values <0.05 were considered statistically significant. Data are expressed as means ± S.E.M.—except where indicated otherwise—for the given number of experiments.
Results
Residual LTCC Binding Activity in Cav1.3(-/-)/Cav1.2DHP(-/-) Double Mutant Mice. We have previously shown that elimination of high DHP sensitivity of Cav1.2 channels in Cav1.2DHP(-/-) mutant mice reduced (+)-[3H]isradipine binding activity to 15% of wild-type (WT) levels in brain (Sinnegger-Brauns et al., 2004). This residual activity was concluded to be associated with Cav1.3 channels, but the possibility of contribution by other LTCC isoforms has never been investigated. To quantify the contribution by LTCCs other than Cav1.2 and Cav1.3 to high-affinity DHP binding in mouse brain, we generated Cav1.3(-/-) × Cav1.2DHP(-/-) (Cav-DM) double mutant mice. In these mice, high-affinity DHP binding to Cav1.2 LTCCs is prevented by a missense mutation in Cav1.2 (Sinnegger-Brauns et al., 2004), and high-affinity binding to Cav1.3 is eliminated by Cav1.3 deficiency (Platzer et al., 2000). As a consequence, any residual binding in these mice should prompt a further biochemical analysis to determine a possible expression of other LTCCs. As expected, homozygous Cav-DM mice were viable and their phenotype was indistinguishable from that of Cav1.3(-/-) mice (Platzer et al., 2000).
In whole-brain microsomal membranes of Cav1.2DHP(-/-) mice, specific binding of the LTCC-selective high affinity probe (+)-[3H]isradipine decreased to 15.1 ± 1.9% (mean ± S.E.M., n = 5) of WT values (Fig. 1A). A significant further reduction to 4.4 ± 0.5% (p = 0.0006, unpaired t test, n = 5) of WT values was observed in Cav-DM mice. Depending on experimental conditions [membrane protein concentration, presence of (+)-cis-diltiazem, see Fig. 1 and Table 1], this residual specific binding corresponded to 100 to 700 dpm and was reproducible in three individual preparations. These data demonstrate that more than 95% of LTCC-associated DHP binding in mouse brain occurs at Cav1.2 and Cav1.3, the latter participating in 10.7% of (+)-[3H]isradipine binding activity.
To determine whether the remaining activity is linked to LTCCs or represents specific binding to other proteins, we further examined its properties. A hallmark of DHP interaction with LTCC α1-subunits is the specific stimulation by the structurally unrelated benzothiazepine (+)-cis-diltiazem through a noncompetitive interaction mechanism (Brauns et al., 1997). As illustrated in Fig. 1B, 10 μM (+)-cis-diltiazem significantly enhanced (+)-[3H]isradipine binding to Cav-DM brain membranes to 323 ± 31% (n = 3) of control. Furthermore, binding was abolished by 300 nM unlabeled racemic isradipine (corresponding to a 150 nM concentration of the active (+)-enantiomer), which indicates that the Kd of the residual binding was in the low nanomolar range. Based on these characteristics, we conclude that the residual (+)- [3H]isradipine binding in Cav-DM brains occurs at LTCCs and could therefore be associated with Cav1.1 and/or Cav1.4 channels. Alternatively, it could reflect residual (lower affinity) binding to the mutant Cav1.2 α1 subunit.
Lack of High-Affinity DHP Binding to Recombinant Cav1.4 and Cav1.4F1414Y LTCCs. To distinguish between these possibilities, we first tested whether Cav1.4 channels can be detected in radioligand binding experiments. Expression of Cav1.4 together with auxiliary subunits in tsA201 cells did not yield specific, protein-concentration-dependent high-affinity (+)-[3H]isradipine binding above background levels. Representative results obtained from three different membrane preparations are illustrated in Fig. 2B. Protein-concentration-dependent specific binding was also missing at higher (1.3-3 nM, n = 3) or lower (n = 5) (+)-[3H]isradipine concentrations. Western blot experiments revealed that the absence of Cav1.4 binding activity was not due to a failure of transient Cav1.4 α1 protein expression (Fig. 2A) (Koschak et al., 2003; Hoda et al., 2005). In comparison, membrane preparations from Cav1.2-transfected cells obtained in parallel experiments showed a robust signal with the expected binding densities (220 ± 31 fmol/mg of protein; Fig. 2B). Furthermore, specific binding could not be induced in the presence of 1, 3, and 10 μM concentrations of the nonDHP Ca2+ channel blockers (+)-cis-diltiazem, (+)-tetrandrine, and mibefradil (Table 1), which caused a robust stimulation of (+)-[3H]isradipine binding to Cav1.2 (all three drugs) and Cav1.3 LTCCs [only tetrandrine and (+)-cis-diltiazem; see legend to Table 1] expressed under identical experimental conditions. In summary, our data suggest that, contrary to other LTCCs, the expression of Cav1.4 channels cannot be quantified using radioreceptor assays, despite their sensitivity to LTCC modulators in electrophysiological studies (Koschak et al., 2003). This could be due to a slightly higher Kd (which would reduce occupancy below levels detectable under our experimental conditions) and/or to very fast binding kinetics that cause dissociation of labeled channel complexes during filter washes. Alignment of the amino acid residues critical for DHP interaction in Cav1.2 α1 subunits revealed that a tyrosine critical for DHP binding (present in Cav1.1, Cav1.2, and Cav1.3 α1-subunits) is replaced by phenylalanine in Cav1.4 α1 (Phe1414). We therefore generated the Cav1.4 mutant F1414Y to test whether a tyrosine at this position could restore high affinity binding. The exchange did not yield measurable specific DHP binding (Fig. 2B), despite efficient expression on the protein level (Fig. 2A). The failure of Cav1.4 detection by radioligand binding in our recombinant system implies that our DHP binding experiments with Cav-DM membranes do not permit predictions about Cav1.4 expression in mouse brain.
Cav1.1 and Cav1.4 RNA Transcripts Are Expressed at Very Low Levels in Mouse Brain. Because Cav1.4 quantification by radioligand binding was not feasible, we used qPCR to determine the relative magnitude of Cav1.4 and Cav1.1 α1 transcript expression compared with Cav1.2 and Cav1.3 in both mouse and rat brain. In mouse, Cav1.2 was expressed at significantly (p < 0.001, n = 9; one-way ANOVA, Bonferroni post-test) higher densities than Cav1.3 in whole-brain preparations and all brain regions investigated (Fig. 3). Cav1.3 contributed 19.4 ± 1.6% (mean, ± S.E.M.) of total LTCC α1-subunit transcripts in whole brain (Fig. 3; 7595 median number of molecules/20-ng RNA equivalent), in accordance with our biochemical estimate. The relative abundance of Cav1.3 transcripts was higher in rat than in mouse brain, where Cav1.3 significantly exceeded Cav1.2 α1 RNA expression in all regions examined (Fig. 4).
Cav1.4 comprised 0.03 ± 0.01% of L-type transcripts in mouse whole brain, which corresponds to only 3 to 30 molecules/20-ng RNA equivalent. Although the Cav1.4 transcript number was within our quantification range (>10 molecules/20 ng; see Materials and Methods) in cerebral cortex and nucleus accumbens, the relative abundance and absolute numbers were extremely low in these regions (cerebral cortex, 0.06 ± 0.02%, 20-120 copies/20 ng; nucleus accumbens, 0.03 ± 0.01%, 20-50 copies/20 ng).
Relative abundance (< 0.07%) and copy numbers (5-80, range in all brain regions) of Cav1.4 were similarly low in whole brain and all regions investigated in rat brain, where the assay was designed against a different region of the Cav1.4 gene (see Materials and Methods). However, up to 8600 copies/20 ng (rat) and 17,000 copies/20 ng (mouse) of Cav1.4 transcripts were detected in eye RNA preparations serving as positive controls (Figs. 3 and 4). This confirmed that our assay was able to efficiently detect Cav1.4 α1-subunit RNA.
Like for Cav1.4, the relative abundance of Cav1.1 transcripts was also very low in mouse brain (0.05 ± 0.01% in whole brain), with the highest contribution in mouse cerebral cortex (0.12 ± 0.05%). Relative Cav1.1 expression was higher in the rat with transcript numbers within our quantification range (i.e., 300 copies/20 ng) in cortex (500-2700 copies/20 ng, 2.39 ± 0.36% of total LTCC α1 transcripts) and nucleus accumbens (390-760 copies/20 ng, 0.62 ± 0.08%). As expected, Cav1.1 α1 RNA was heavily expressed in skeletal muscle preparations (up to 390,000 copies/20 ng in mouse and 860,000 copies/20 ng in rat), which served as positive controls (Figs. 3 and 4).
It is noteworthy that the relative transcript abundance of Cav1.1 and Cav1.4 did not change in whole-brain preparations of Cav1.3(-/-) (Cav1.4, 0.07 ± 0.01%; Cav1.1, 0.06 ± 0.01%, n = 13) and Cav1.2DHP(-/-) (Cav1.4, 0.05 ± 0.003%; Cav1.1, 0.06 ± 0.003%, n = 13) mice, suggesting that the residual DHP binding in Cav-DM whole brain is unlikely to result from compensatory expression of Cav1.1 and/or Cav1.4.
Residual Binding Activity of Cav1.2. Our results suggested that neither Cav1.1 nor Cav1.4 is expressed at levels that could sufficiently support the residual LTCC-associated binding activity detected in homozygous Cav-DM mice. Therefore, the mutated Cav1.2 α1 subunit resulting from their Cav1.2DHP(-/-) genotype remained as the most likely explanation for the residual binding. We estimated the lower affinity of the mutant Cav1.2 α1 subunit as follows: for the experiments in Fig. 1A, we employed (+)-[3H]isradipine concentrations (legend to Fig. 1) that cause similarly high occupancy of the high-affinity binding domains on both Cav1.2 and Cav1.3 LTCCs (Sinnegger-Brauns et al., 2004). From the difference in binding density between Cav1.2DHP(-/-) and Cav-DM mice, the contribution of Cav1.3 could be calculated as 10.7% of the total binding activity (see Fig. 1A). Assuming that only Cav1.2 contributes to the remaining activity, 89.3% of binding in wild-type brains should be associated with Cav1.2; 4.4% of binding were retained in the Cav-DM mice. If the latter represented lower affinity binding to the mutant Cav1.2 α1 subunit, it would correspond to a 20.3-fold decrease in occupancy. The higher Kd (Kd2) accounting for this decrease can be estimated as Kd2 = ((Kd1 + F) · 20.3) - F, where Kd1 is the Kd for the wild-type Cav1.2 channel [0.078 nM as previously determined by us in mouse brain preparations under identical experimental conditions (Sinnegger-Brauns et al., 2004)] and F is the concentration of free ligand (essentially identical to total ligand, 0.43 nM; see legend to Fig. 1). The 20.3-fold decrease in occupancy therefore translates into a 127-fold increase in Kd from 0.078 nM for the wild-type Cav1.2 channel (Sinnegger-Brauns et al., 2004) to Kd2 = 9.88 nM for the mutant channel. This calculation is in agreement with the approximately 150-fold decrease in affinity previously estimated from electrophysiological data (Sinnegger-Brauns et al., 2004).
Selectivity of Some DHP Ca2+Channel Blockers for the Cav1.2 DHP Binding Domain. As concluded from our binding data from Cav-DM mouse brain, the majority (71%; 10.7% of 15.1% remaining binding activity) of the residual high-affinity (+)-[3H]isradipine sites in the Cav1.2DHP(-/-) mutants is associated with Cav1.3. Therefore, Cav1.2DHP(-/-) brain membranes should be suitable to determine the affinity of competitive inhibitors for native Cav1.3 channels. In contrast, having excluded Cav1.1 and Cav1.4 as likely contributors to (+)-[3H]isradipine binding, the majority (89.3%) of binding in WT brain membranes should occur at Cav1.2 and thus allow determination of drug binding affinities for native Cav1.2 channels. We therefore exploited the Cav1.2DHP(-/-) model to test whether structurally different DHPs can discriminate between the native Cav1.2 and Cav1.3 binding pockets. WT and Cav1.2DHP(-/-) brain membranes were labeled with (+)-[3H]isradipine in the absence and presence of increasing concentrations of unlabeled DHPs. The resulting IC50 values were similar for (±)-isradipine [mostly reflecting activity of the more active (+)-enantiomer], its less active (-)-enantiomer [(-)-isradipine], amlodipine, and azidopine, but a 3- to 4-fold higher affinity was observed in WT brains for nifedipine and nitrendipine (Fig. 5, A and B; Table 2; both the means of IC50 values and the selectivity ratios were significantly different). This indicated that these compounds exhibit higher affinity for Cav1.2. To further confirm this finding, we repeated these experiments in membranes prepared from tsA201 cells transiently transfected with Cav1.2 and Cav1.3 α1 subunits (Fig. 5, C and D; Table 3). Again, nitrendipine and nifedipine were 3- to 4-fold selective for Cav1.2. Selectivity (albeit weaker, 2-fold or less) of amlodipine, isradipine (selectivity ratio differences), and azidopine (mean IC50 difference) was observed also in the recombinant system (Table 3). Our data demonstrate that some DHPs bind to LTCCs in an isoform-selective manner.
Discussion
Our data provide novel biochemical and pharmacological information about neuronal LTCCs in mammalian brain. By generating the double mutant strain Cav-DM from Cav1.2DHP(-/-) and Cav1.3(-/-) mice, we demonstrated the existence of a residual (+)-[3H]isradipine binding component, which not only allowed us to obtain a more precise estimate for the contribution by the Cav1.3 channel (of approximately 11%) but also prompted us to investigate possible contribution by Cav1.1 and Cav1.4 LTCCs. Using qPCR, we show that Cav1.1 and Cav1.4 are expressed at very low levels and therefore cannot represent the residual binding component in mouse brain, which we instead explain by low-affinity binding to the mutated Cav1.2 subunit in Cav1.2DHP(-/-) mice. By excluding Cav1.1 and Cav1.4 as the relevant contributors to the residual binding component, we exploited Cav1.2DHP(-/-) mouse brains (predominant expression of Cav1.3) as an assay for native Cav1.3 channels. Our data demonstrate that some DHPs such as nifedipine and nitrendipine show substantial selectivity for Cav1.2 channels, which can explain some of the Cav1.2-selectivity observed also in functional experiments (Koschak et al., 2001; Helton et al., 2005).
We have previously obtained a higher estimate for the contribution of Cav1.3 channels to high affinity (+)-[3H]isradipine binding activity (20-25%) using Cav1.3(-/-) mice (Clark et al., 2003) compared with the estimate obtained in this study. Although we found no evidence for changes of Cav1.2α1 protein expression in Cav1.3(-/-) [and Cav1.2DHP(-/-)] mice in our previous analysis (Clark et al., 2003, Sinnegger-Brauns et al., 2004), we cannot rule out the possibility that a small decrease in Cav1.2α1 protein expression resulted in an overestimation of Cav1.3 contribution in the knockouts. This again emphasizes the role of Cav1.2DHP(-/-) and Cav-DM mice as alternative models to obtain more precise estimates for the relative abundance of the Cav1.3 channel protein in mouse brain.
Our finding of very low Cav1.1 and Cav1.4 RNA expression levels is strengthened by the fact that we obtained the same result in two different species. However, our data stand in contrast to a study reporting relatively high levels of Cav1.1 expression in various human brain regions (Takahashi et al., 2003). In many of the brain regions investigated in that report, Cav1.1 transcript numbers were similar (such as in the putamen, occipital cortex, or spinal cord) or even exceeded those of Cav1.3, such as in the substantia nigra, pallidum, and caudate nucleus. We could not confirm such a high level of Cav1.1 expression in mouse and rat brain and different brain regions. In our study, the highest number of Cav1.1 transcripts was found in cerebral cortex and nucleus accumbens. In particular, we found no significant Cav1.1 expression in the caudate putamen of rat and mouse, the region reported to display the highest Cav1.1 expression level in human brain (Takahashi et al., 2003), and Cav1.3 expression by far exceeded Cav1.1 levels in all brain areas tested in our experiments. It is also noteworthy that we have observed a greater relative abundance of Cav1.3 in rat brain compared with mouse. Whether these differences represent species diversity or can be attributed to other confounding factors such as age, gender, or strain polymorphisms remains to be determined. Given the potential role of Cav1.3 LTCC activity for age-induced decreases in cognitive function (Veng et al., 2003, Herman et al., 1998) and the development of Parkinson disease (Chan et al., 2007), our finding deserves more detailed analysis in further experiments.
We found that one of the most sensitive detection methods, reversible labeling with (+)-[3H]isradipine, failed to detect Cav1.4 LTCCs after expression in tsA201 cells under conditions that reliably yield Cav1.4 α1 protein in Western blots, give rise to Cav1.4-mediated currents (Koschak et al., 2003; Hoda et al., 2005) and allow robust detection of recombinant Cav1.2 and Cav1.3 LTCCs by this radioligand (Fig. 2, Tables 1 and 3). We show that this is not due to the phenylalanine substitution of a tyrosine in position 1414 in the Cav1.4 α1 subunit, a residue contributing to the DHP sensitivity of other LTCCs and present in all other LTCC α1 subunits. In Cav1.1 channels, replacement of the homologous tyrosine by a phenylalanine results in an about 3-fold increase in the Kd for (+)-[3H]isradipine (Peterson et al., 1996). Substituting Cav1.4 Phe1414 by tyrosine did not result in detectable binding activity. It is noteworthy that there are also some other amino acid exchanges in the putative Cav1.4 DHP binding pocket that could cause a reduction in (+)-[3H]isradipine binding affinity (including proline-1037 and alanine-1087 in the IIIS5-IIIS6 pore loop). This may shift the Kd outside the detection range of our binding assay. A slightly lower affinity for the DHP binding pocket may also contribute to the lower sensitivity of Cav1.4 channel currents for isradipine compared with Cav1.3 and Cav1.2 (Koschak et al., 2003).
We demonstrated that nitrendipine and nifedipine exhibited 3- to 4-fold higher affinities for the Cav1.2 DHP binding pocket in the native membrane environment, a finding that was further confirmed employing recombinant channels. This represents the first comparison of the pharmacological properties of the DHP binding domains of native and recombinant Cav1.2 and Cav1.3 LTCC complexes. Nitrendipine and nifedipine are distinguished from the other DHPs tested by their NO2 substituent at the 3′ and 2′ position of the phenyl ring, respectively. The existence of Cav1.2 selectivity was surprising given the very high sequence similarity within the DHP binding pocket of Cav1.2 and Cav1.3. Our findings indicate that nitrendipine and nifedipine can sense differences in the molecular architectures of these two binding domains. It therefore seems likely that such differences may also allow development of molecules with higher affinity for the Cav1.3 binding pocket. This could aid in the development of Cav1.3 selective LTCC blockers, which, based on emerging knowledge about the physiological role of Cav1.3 channels (Platzer et al., 2000; Striessnig et al., 2006; Chan et al., 2007), could open novel therapeutic strategies for the treatment of depression and Parkinson disease.
Acknowledgments
We thank G. Pelster and S. Haßler for expert technical assistance, L. Glatzl and C. Bachmann for animal care, Bettina Schlick for help with qPCR, and Bernhard E. Flucher for helpful discussions.
Footnotes
-
This work was supported by the Austrian Research Fund [Grants P17159, P17807], the Tyrolean Research Fund, and the University of Innsbruck.
-
M.J.S.-B. and I.G.H. contributed equally to this work.
-
ABBREVIATIONS: LTCC, L-type Ca2+ channels; DHP, 1,4-dihydropyridine; qPCR, quantitative real-time PCR; WT, wild-type; ANOVA, analysis of variance.
-
↵1 Current affiliation: Department of Neuroscience, Faculty of Medicine, University of Geneva, Geneva, Switzerland.
- Received June 30, 2008.
- Accepted November 24, 2008.
- The American Society for Pharmacology and Experimental Therapeutics