Differential Expression and Association of Calcium Channel α1B and β Subunits during Rat Brain Ontogeny*

Calcium functions as an essential second messenger during neuronal development and synapse acquisition. Voltage-dependent calcium channels (VDCC), which are critical to these processes, are heteromultimeric complexes composed of α1, α2/δ, and β subunits. β subunits function to direct the VDCC complex to the plasma membrane as well as regulate its channel properties. The importance of β to neuronal functioning was recently underscored by the identification of a truncated β4 isoform in the epileptic mouse lethargic (lh) (Burgess, D. L., Jones, J. M., Meisler, M. H., and Noebels, J. L. (1997) Cell 88, 385–392). The goal of our study was to investigate the role of individual β isoforms (β1b, β2, β3, and β4) in the assembly of N-type VDCC during rat brain development. By using quantitative Western blot analysis with anti-α1B-directed antibodies and [125I-Tyr22]ω-conotoxin GVIA (125I-CTX) radioligand binding assays, we observed that only a small fraction of the total α1B protein present in embryonic and early postnatal brain expressed high affinity125I-CTX-binding sites. These results suggested that subsequent maturation of α1B or its assembly with auxiliary subunits was required to exhibit high affinity125I-CTX binding. The temporal pattern of expression of β subunits and their assembly with α1B indicated a developmental pattern of expression of β isoforms: β1b increased 3-fold from P0 to adult, β4 increased 10-fold, and both β2 and β3 expression remained unchanged. As the β component of N-type VDCC changed during postnatal development, we were able to identify both immature and mature forms of N-type VDCC. At P2, the relative contribution of β is β1b > β3 ≫ β2, whereas at P14 and adult the distribution is β3 > β1b = β4. Although we observed no β4 associated with the α1B at P2, β4 accounted for 14 and 25% of total α1B/β subunit complexes in P14 and adult, respectively. Thus, of the β isoforms analyzed, only the β4 was assembled with the rat α1B to form N-type VDCC with a time course that paralleled its level of expression during rat brain development. These results suggest a role for the β4 isoform in the assembly and maturation of the N-type VDCC.

Ca 2ϩ channels play important roles in neuronal development. Both VDCCs 1 and calcium entry have been implicated in many processes of immature neurons including neurite outgrowth (1)(2)(3), neuronal migration (4), and axon and dendrite extension (5,6). Calcium entry through VDCCs has been shown to be essential for sculpting neuromuscular synapses (7)(8)(9)(10). Calcium channels have also been implicated in the initiation of developmental gene expression (11) and are necessary for spinal cord motoneuron differentiation in rat (12).
In vitro studies indicate that functional VDCCs are composed of three subunits: ␣ 1 , ␣ 2 /␦, and ␤. Multiple VDCCs (N-, P/Q-, L-, and R-type) are expressed in neuronal tissues (13,14). The ␣ 1 subunit, of which there are at least six genes, resides in the membrane and forms the pore of the channel (14). The ␤ subunit, a putative hydrophilic protein, binds the ␣ 1 (15) and acts to regulate channel gating and kinetics (16,17). Four ␤ subunit genes have been identified, several of which exist as splice variants (14). N-type VDCCs are localized to the plasma membrane of neuronal processes (18) and are essential for the generation of action potentials and the subsequent release of neurotransmitters (19,20). Adult N-type as well as other neuronal VDCCs are heterogeneous in their ␤ subunit component (21)(22)(23)(24). Although the in vivo process which gives rise to the heterogeneity in ␣ 1B/ ␤ subunit complexes is not understood, it is anticipated to generate N-type VDCC with different channel properties based upon in vitro co-expression studies (25,26).
Developing neurons offer the opportunity to investigate the underlying trends that contribute to subunit diversity. Several families of ligand-gated ion channels evidence developmental changes in their subunit composition (27)(28)(29). Whereas studies have identified alterations in the expression of N-type and P/Q-type VDCC during synapse formation in cultured neurons (30 -32), few reports have investigated the developmental or differentiation-dependent expression of calcium channel subunits and their assembly (33,34). Interestingly, there have been reports of N-type VDCC in developing cerebellar granule cells (35) and differentiated human neuroblastoma SH-SY5Y cells (36) which are inhibited by -conotoxin GVIA (CTX) yet display two components of inactivation. The expression of Ntype VDCC with different channel properties is a possible mechanism for controlling membrane excitability. As ␤ sub- 1 The abbreviations used are: VDCC, voltage-dependent calcium channel(s); N-type VDCC, -conotoxin-sensitive VDCC; ␣ 1B , 230-kDa subunit of the N-type VDCC; ␣ 2 /␦, 160-kDa subunit of N-type VDCC; ␤1 through ␤4,   units are known to influence the time course of channel inactivation, the diversity in N-type VDCC activity was suggested to reflect heterogeneity in the ␤ subunit component. We undertook this study to test directly the hypothesis that N-type VDCCs are differentially associated with ␤ subunits during rat brain development. The objectives of this investigation are 1) to identify possible trends in calcium channel subunit expression by evaluating changes in the expression of ␣ 1B and ␤ isoforms throughout postnatal development and 2) to identify possible "immature" and "mature" forms of the N-type VDCC by characterizing the ␤ subunit component of N-type VDCC at different stages of rat brain maturation.

EXPERIMENTAL PROCEDURES
Materials-Unless noted, all reagents were obtained from Sigma. Calpain inhibitors I and II were obtained from Calbiochem. Enhanced chemiluminescence kit (ECL) was purchased from Amersham Pharmacia Biotech; unlabeled -GVIA conotoxin was from Peninsula Laboratories; [ 125 I-Tyr 22 ]-conotoxin GVIA (specific activity 2200 Ci/mmol) and 125 I-protein A (specific activity 21.1 Ci/g) were obtained from NEN Life Science Products. 125 I-IgG was obtained from ICN (specific activity 12.9 Ci/g). Nitrocellulose membranes were obtained from Schleicher & Schuell. All goat anti-mouse secondary antibodies were from Boehringer Mannheim. Sulfolink columns were purchased from Pierce. Bovine serum albumin was from U. S. Biochemical Corp. Hepes was from Research Organics.
Preparation of Rat Homogenates and Membranes-Embryonic (E18), neonatal, and adult rats were euthanized and the brains removed and immediately placed in 50 mM Hepes, pH 7.4, 1 mM EGTA plus protease inhibitors at a protein to volume ratio of 1.3 g/25 ml. The protease inhibitors were added from stock solutions prepared as follows: phenylmethanesulfonyl fluoride (1/1000 dilution from 200 mM stock in ethanol), calpain inhibitors I and II (1/1000 dilution from 4 mg/ml stock), benzamidine (1/500 dilution of 200 mM stock), aprotinin (1/500 dilution from 1 mg/ml stock), leupeptin (1/500 dilution from 1 mg/ml stock), pepstatin (1/500 dilution from 1 mg/ml stock in Me 2 SO), and DTT (1/1000 dilution from 1 M stock). The tissue was homogenized with a Polytron for 10 s and centrifuged at 18,000 rpm (48,000 ϫ g) for 15 min. The membranes were resuspended in 5 ml of 50 mM Hepes, pH 7.4, plus protease inhibitors at a resulting protein concentration of 3-10 mg/ml. For subsequent use in Western blot analysis, all samples were stored in Ϫ20°C at concentrations of 2 mg/ml in sample buffer (5ϫ sample buffer: 325 mM Tris, pH 7.0, glycerol (25% v/v), mercaptoethanol (25% v/v), SDS (10%)) in 100-l aliquots. The samples were not freeze-thawed.
Production of Anti-peptide Polyclonal Antibodies to VDCC Subunit Epitopes-The peptide antigens were synthesized to include a unique cysteine residue to be used both in the unambiguous attachment of peptide to carrier protein (maleimide-derivatized keyhole limpet hemocyanin) and to the affinity column (Sulfolink). The peptides were synthesized, purified, and coupled to maleimide-activated keyhole limpet hemocyanin. The coupled peptide antigens were used in the production of polyclonal sera in rabbits under continued contractual agreement with Covance, Inc. The rabbits were bled twice per month (15-20 ml/bleed) and tested for production of specific antibodies after 4 weeks as described previously (34).
Anti-␣ 1B rat sequence-specific antibodies (Ab CW8) were raised to a sequence (ASTPAGGEQDRTDC corresponding to amino acid residues 863-875 in the rat cDNA) that is present in the rat brain and rat spinal cord ␣ 1B cDNA (37, 38). Anti-␣ 1B subunit antiserum (Ab CW14) (34) was raised to a sequence (EQPEDADNQRNVTRMGSQP corresponding to amino acid 1051-1069) in the rat cDNA (37) which was present in all N-type ␣ 1B subunit cDNAs cloned to date.
Preparation of Peptide Columns-The concentration of free sulfhydryl group available in the peptide sample was quantified by Ellman's assay with cysteine as the standard used in the range of 20 -200 nmol/assay. The free peptide (1.5-3 mg in 2-ml volume) was coupled to a Sulfolink column (Pierce), and excess reactive groups were coupled to cysteine according to the manufacturer's instructions. The coupling efficiency was 94 -99% as determined by both dot blotting and Ellman's assay.
Affinity Purification of Anti-peptide Antibodies-Crude antisera were diluted with 3% BSA in TBS and incubated with peptide column for 90 min at room temperature. The column was then washed with 25 ml of TBS and sequentially eluted with 3 ml of 30 mM glycine, pH 5, and 3 ml of 80 mM glycine, pH 4. The affinity purified antibodies were eluted from the column with 6 ml of 200 mM glycine, pH 2.5, with each 1 ml collected and immediately neutralized by the addition of 50 l of 1 M Tris, pH 9.5. The protein concentrations of the fractions were determined, and the peak fractions were pooled and dialyzed overnight in the cold room against 4 liters of TBS. The following morning, the sample was dialyzed for an additional 4 h, and the purified antibody was aliquoted in 50 -100-l volumes and stored at Ϫ80°C without freezethawing. The peptide column was extensively washed with TBS and stored in TBS plus 0.05% azide.
Production of Anti-peptide Polyclonal Antibodies to VDCC ␤ Subunits-Anti-␤ subunit "generic" antibodies (Ab CW24) were raised to a highly conserved sequence (CESYTSRPSDSDVSLEEDRE) present in all ␤ subunits cloned to date that is not implicated in either the binding of ␤ to the ␣ 1 subunit or the consensus sites for protein phosphorylation or ATP-binding in the ␤4 subunit (39). This peptide was coupled via a unique CYS as described above.
For the generation of ␤ isoform-specific antibodies, the peptides were synthesized with an N-terminal monochloroacetyl-glycyl extension (BioTeZ, Berlin, Germany) and coupled to keyhole limpet hemocyanin after activation with 2-iminothiolane hydrochloride as described (40). Immunization was done in New Zealand White rabbits in accordance with internationally accepted principles concerning the care and use of laboratory animals.
Methods for 125 I-CTX Binding-125 I-CTX ([ 125 I-Tyr 22 ]-conotoxin GVIA (specific activity 2000 Ci (81.4 TBq/mmol)) binding was assayed by published procedures which use filtration over PEI-soaked glass fiber filters (Whatman GF/B) to separate bound from unbound ligand in the presence of bovine serum albumin (BSA). N-type VDCC fractions were routinely screened at several protein concentrations to determine the linear range for the binding. Individual assay tubes contained 100 l of representative N-type VDCC fractions diluted into 50 mM Hepes, pH 7.4, 100 l of 1% BSA (w/v), 100 l of 125 I-CTX stock solution diluted to correspond to approximately 20,000 cpm 125 I-CTX (or approximately 4.2 fmol), 100 l of 500 nM stock solution of unlabeled CTX (Peninsula Laboratories) or 50 mM Hepes, pH 7.4 buffer in 1 ml final volume. The samples were incubated at room temperature for 30 min, filtered over 0.5% PEI-soaked glass fiber filters, and rapidly washed as described. The filters were counted for 1 min in Packard Cobra autogamma counter. Scatchard analysis of 125 I-CTX binding to membranes was carried out under similar conditions with protein assayed at dilutions that supported approximately 2,500 cpm of specific 125 I-CTX bound (1-200 g/ml per assay). The amount of protein present in each assay was as follows: adult, 2 g; P0, 25 g; and E18 rat brain, 200 g. Unlabeled CTX added was from 0.1 pM to 50 nM in the presence of constant 125 I-CTX. Data presented are mean Ϯ S.D. from three determinations done in duplicate.
Immunoprecipitation of N-type VDCC-The N-type VDCC was solubilized from P2 brain and P14 and adult rat forebrain as described previously (21) with the following modifications: the N-type VDCC from P2 brain and P14 forebrain were solubilized directly from membranes using 0.75% CHAPS. Following centrifugation, the solubilized preparations were assayed for 125 I-CTX binding. Approximately 9,000 -12,000 cpm of specific 125 I-CTX receptor activity (200 l of the solubilized preparation) was added to individual microcentrifuge tubes that contained 20,000 -40,000 cpm 125 I-CTX, 0.1% BSA, plus protease inhibitors in 50 mM Hepes, pH 7.4. Identical reactions were also carried out in the presence of 50 nM unlabeled CTX to determine nonspecific binding of 125 I-CTX. Following a 30-min incubation at room temperature, antibody was added in a total volume of 100 -200 l of TBS and left to incubate at room temperature for 1 h. After this time, 50 l of protein A-Sepharose 4B (final concentration of 0.6 mg/ml) was added to each sample and rotated in the cold room overnight. 125 I-CTX binding to the soluble fraction was determined by directly filtering 1.1 ml of the sample through 0.5% PEI-soaked glass fiber filters. The pellets were washed 3 ϫ with 1 ml of 50 mM Hepes/EGTA, and 50 l of 2ϫ sample buffer were added to the protein A beads. The samples were counted in a gamma counter for 1 min.
Quantification of VDCC Subunits by 125 I-Protein A Overlay or 125 I-Goat Anti-rabbit IgG-125 I-Protein A and 125 I-goat anti-rabbit IgG were diluted in 50 ml of 3% BSA in 1ϫ TBS. Filters previously blocked with 5% milk in 1ϫ TBS and probed with primary antibody were washed with 1ϫ TBS for 15 min and then washed two additional times for 5 min. The washed filters were incubated in 50 ml of either the 125 Iprotein A, 3% BSA solution (approximately 500 cpm/l) or the 125 I-IgG, 3% BSA solution (approximately 30 -50 cpm/l) for 2 h at room temperature with constant shaking. Following this incubation, the 125 I solution was removed, and the filters were washed 3-5 times with TBS (5 min each). The approximate wash volume was 50 ml. The filters were blotted with paper towels and exposed to film at Ϫ80°C with the aid of intensifying screens. The position of the antigen was determined relative to the exposed film, and the corresponding band on the filter was cut and counted. Slices that corresponded to nonspecific areas of the filter were also counted and subtracted from the signal. The data were obtained from multiple determinations done in duplicate.
General Methods and Data Analysis-The gels were transferred to nitrocellulose at 0.45 A for 17-22 h. The filter was incubated in 5% powdered milk in TBS ϩ 0.01% sodium azide ϩ 0.05% Tween and blocked for either 3 h at 37°C with constant shaking or overnight at 4°C in the cold room. The primary antibody was diluted in 3% BSA, 1 ϫ TBS and incubated with the filter overnight at 4°C. The filters were washed 3 times in TBS at room temperature. The secondary antibody was diluted to 1/10,000 in 3% BSA, 1ϫ TBS, and the filter was incubated for 45 min at room temperature with constant shaking. The filters were washed as before. The antigen was visualized using ECL. Membrane protein and soluble protein were measured by the Pierce BCA assay. Bovine serum albumin was used as a standard in all cases, and all samples were normalized with respect to buffer and detergent composition. Gel electrophoresis was carried out on polyacrylamide gels according to standard procedures (45). Gel electrophoresis was carried out using a 4% stacking gel and a resolving gel of appropriate porosity (see figure legends) according to standard procedures. All samples were incubated with 5ϫ SDS-PAGE sample buffer (325 mM Tris, pH 7.0, glycerol (25% v/v), mercaptoethanol (25% v/v), SDS (10%) without boiling. Staining of proteins in polyacrylamide gels was Coomassie Blue (0.05%), 50% methanol, 10% acetic acid. The results are expressed as mean Ϯ S.D. Statistical analysis was evaluated by a paired t test or one-way analysis of variance with Tukey's or Dunnett's post hoc test. p values less than 0.05 were considered significant.

Expression and Properties of the N-type ␣ 1B Subunit during Rat Brain Development
Immunological Characterization of Rat ␣ 1B -Anti-peptide antibodies to the II-III intracellular loop of the rat ␣ 1B subunit were raised to two distinct epitopes. The first antibody, Ab CW14, was raised to an epitope present in all ␣ 1B subunits cloned to date (34). The second antibody, Ab CW8, was raised to an epitope present only in the rat ␣ 1B sequence (37). Both antibodies were analyzed in parallel to characterize the structure of the endogenous ␣ 1B as it relates to the original rat ␣ 1B cDNA clone (37).
The pan specificity of Ab CW14 was verified by its detection of 230/210-kDa proteins in HEK293 cell line G1A1 stably expressing the human ␣ 1B (46, 47) and a single 230-kDa protein in rat, rabbit, and mouse forebrain samples (Fig. 1A). Both the 230/210-kDa proteins in HEK293 cell line G1A1 were determined to bind 125 I-CTX by photoaffinity labeling with derivatized 125 I-CTX (21) (data not shown). The specificity of Ab CW8 for the rat epitope was evidenced by its reaction with the 230-kDa protein expressed only in rat brain. To verify the expression of the rat ␣ 1B as the isoform expressed in representative developmental samples, we quantified the amount of ␣ 1B detected by Ab CW8 and Ab CW14 in brain homogenates of E18, P0, and adult rat. As shown in Fig. 2A, ␣ 1B detected by Ab CW14 increases from E18 to adult. The signal obtained with Ab CW8 is very similar to that obtained with Ab CW14. The signals were quantified using 125 I-protein A, and the results indicate that the epitopes targeted by Ab CW8 and Ab CW14 are equivalently expressed.
Developmental Expression of ␣ 1B Subunit in Rat Brain-The change in expression of ␣ 1B subunit presented in Fig. 2 leads us to use Ab CW14 to evaluate the level of expression of ␣ 1B in postnatal (P0 -P14) rat brain and adult forebrain homogenates. As shown in Fig. 3A, the level of expression of ␣ 1B increases during rat brain development. The results of similar Western   2. Expression of different ␣ 1B epitopes during rat brain development. Experimental conditions were identical to those represented in Fig. 1. A, tissues obtained from differently aged rats were resolved by SDS-PAGE on a 4 -17% gradient gel, transferred to nitrocellulose, and probed with antibodies Ab CW14 (at 1/200 dilution) or Ab CW8 (1/200 dilution). B, following incubation with primary antibody the blots were incubated with 125 I-protein A, washed, and exposed to film, and the data were expressed as the ratio of 125 I-protein A/band relative to the adult sample. The samples are from rat brain at embryonic day 18 (lane 1), postnatal day 0 (lane 2), and adult rat forebrain (lane 3). The concentration of protein was 100 g/lane. The total 125 Iprotein/band for adult was 1340 cpm when probed with Ab CW8 and 1250 cpm when probed with Ab CW14. Nonspecific counts (150 cpm) were subtracted from each sample. blots were quantified using 125 I-IgG and evidenced statistically significant increase in expression of ␣ 1B subunit throughout the period of postnatal development (Fig. 3B).
Characterization of N-type VDCC 125 I-CTX Binding during Three Rat Developmental Stages-We used radioligand binding assays of the peptide neurotoxin, 125 I-CTX, to determine if there were developmental differences in either the density or affinity of 125 I-CTX binding to N-type VDCC. Radioligand binding assays using 125 I-CTX were carried out on homogenates from embryonic day 18 brain (E18), newborn rat brain (P0), and adult rat forebrain. The results of the pseudo-Scatchard analysis are presented in Table I. A comparison of the B max values indicates significantly less 125 I-CTX binding at E18 (B max ϭ 0.008 Ϯ 0.002 pmol/mg) and P0 (B max ϭ 0.05 Ϯ 0.01 pmol/mg) versus adult rat forebrain (B max ϭ 1.8 Ϯ 0.8 pmol/ mg). The high affinity 125 I-CTX binding site diagnostic for the N-type VDCC is present throughout development (K d of 11.7, 21.7, and 8.3 pM for 125 I-CTX binding to adult rat forebrain, P0 rat brain and E18 rat brain, respectively).
We then used the data obtained by quantitative Western blot analysis and Scatchard analysis to investigate the fraction of ␣ 1B subunits in E18 and P0 that could support high affinity binding relative to the adult sample. This comparative analysis, which assumes that the ratio of ␣ 1B /binding sites in adult is unity, identifies differences in the ratio of ␣ 1B protein present in E18 and P0 brain as quantified by 125 I-protein A versus the density of 125 I-CTX binding sites (Table I). In E18 brain, ratio of ␣ 1B /binding sites is 25:1, whereas in P0 the ratio is 12.5:1. The discrepancy between the expression of ␣ 1B protein and 125 I-CTX binding sites suggests a population of ␣ 1B present at E18 and P0 that does not support high affinity 125 I-CTX binding.

Expression and Assembly of VDCC ␤ Subunits during Rat Brain Development
Expression of ␤ Subunit Isoforms during Rat Brain Development-The level of expression of calcium channel ␤ subunits in developing rat brain was then analyzed to evaluate possible changes in the pool of available ␤ isoforms. Thus, we used Ab CW24, an antibody raised to an epitope shown to be present in all ␤ subunits cloned to date (34), to probe a Western blot of homogenates prepared from developing rat brains. These experiments revealed two populations of ␤ subunits that could be easily resolved by SDS-PAGE as follows: ␤ subunits with apparent molecular masses of Ͼ80 kDa comprised of ␤1b and ␤2 isoforms, and smaller ␤ subunits (65 kDa) comprised of ␤3 and ␤4 isoforms (34). As shown in Fig. 4, direct comparison of these two populations of ␤ subunits indicated no statistically significant change in the level of expression of the larger ␤ subunits between P0 and adult, whereas the smaller ␤ subunits evidenced a significant 3-fold increase in expression. It is important to note that the histogram reflects the expression of the ␤1b ϩ ␤2 and ␤3 ϩ ␤4 as the individual bands could not be resolved adequately (Fig. 4, B and C).
Expression of Specific ␤ Subunit Isoforms during Development-Therefore, to determine accurately the time course specific to each ␤ subunit isoform, the experiment was repeated using ␤ isoform-specific antibodies. The changes we observed in ␤ isoform expression are indicative of the total pool of available ␤ as we carried out the analysis in whole brain homogenate rather than in plasma membrane fractions or biochemically purified preparations of VDCC. As shown in Fig. 5A, the results indicate a statistically significant increase in the expression of the ␤1b isoform detected in P0 through adult with the increase commencing at the time of cerebellar maturation (P7). The ␤2 isoform (Fig. 5B) and ␤3 isoform (Fig. 5C) were expressed at constant levels. Interestingly, there is a 10-fold increase in the level of expression in the ␤4 isoform in adult brain compared with P0 that also commences at the time of cerebellar maturation (P7) (Fig. 5D). The increase in expression of the ␤4 detected in the P7-P14 interval in these rat brain samples parallels the increased expression of the ␤4 mRNA in cerebellum as determined by in situ hybridization (48).
It is evident from these results that the reactivity of the ␤-generic antibody Ab CW24 for the smaller ␤3 subunits (␤3 ϩ ␤4) accurately reflects the sum of their expression as verified by the isoform-specific antibodies (Fig. 5C). It is interesting that the results obtained using Ab CW24 to monitor the expression of a larger ␤ subunit are seemingly discrepant when compared with the results obtained using the anti-␤1b-and anti-␤2-specific antibodies. Therefore, our observations may indicate ␤ isoform(s) that are not detected by our anti-␤1b and anti-␤2 antibodies at P0 -P7.
Heterogeneity of ␣ 1B/ ␤ Complexes during Rat Brain Development-The N-type VDCC has been previously purified from rat (21,49,50) and rabbit forebrain (51), where its density of expression is 2.5-fold higher than in cerebellum or other brain regions (52). Therefore, N-type VDCC were solubilized from P2 brain, P14, and adult rat forebrain and immunoprecipitated with the anti-␣ 1B antibody Ab CW14, and the generic anti-␤ subunit antibody Ab CW24. The N-type VDCC present at early stages of rat brain development (P2 and P14) can be quantitatively immunoprecipitated by antibodies to the ␣ 1B (Fig. 6). However, only 60 -70% of all 125 I-CTX binding can be immunoprecipitated by the generic antibody reactive toward all ␤

FIG. 3. Developmental expression of ␣ 1B subunit in rat brain.
Tissue was obtained from animals at different ages, collected into sample buffer, and resolved by SDS-PAGE upon a 6% gel. The amount of protein was 150 g/lane. The gels were transferred to nitrocellulose and probed with Ab CW14 (1/100 dilution, A) and visualized by ECL. In experiments conducted on similar samples, the transferred antigens were detected using 125 I-IgG, exposed to film, and counted (B). The data were obtained from three determinations done in duplicate and normalized to the signal obtained at P0 (179 Ϯ 115 cpm band). * denotes p values less than 0.05. subunits (Ab CW24). These data identify a statistically significant fraction of ␣ 1B -binding sites in immature brain that are not tightly associated with a ␤ subunit. In contrast, the N-type VDCC extracted from adult rat can be quantitatively immunoprecipitated by both anti-␣ 1B and Ab CW24.
We then used ␤ isoform-specific antibodies to immunopre-cipitate the 125 I-CTX-labeled N-type VDCC from P2, P14, and adult rat forebrain. As shown in Fig. 7, ␤ subunit isoforms are differentially assembled with the N-type VDCC during rat brain development. At P2, the antibody to the ␤1b subunit immunoprecipitated 37% (Ϯ5) of 125 I-CTX binding, whereas the anti-␤3 antibody and ␤2 antibodies immunoprecipitated 27% (Ϯ3) and 1% (Ϯ0.9), respectively. There was no specific immunoprecipitation of 125 I-CTX by the anti-␤4 antibody, which gave a signal comparable to control samples without receptor (Fig. 7). The sum of the total 125 I-CTX immunoprecipitated from P2 by anti-␤1b, -␤2, and -␤3 antibodies (65%) is in good agreement with the results obtained using Ab CW24 alone. At P14, we observed a shift in the contribution of the various ␤ subunits to N-type VDCC channel complexes with the ␤3 isoform accounting for 32% (Ϯ5.8) of total N-type complexes and the ␤1b accounting for 13% (Ϯ2.0). The ␤4 subunit was associated with 14% (Ϯ9.0) of total N-type VDCC at P14 versus 25% (Ϯ6.3) of total N-type VDCC in adult. The sum of the total 125 I-CTX immunoprecipitated by anti-␤1b, -␤2, -␤3, and -␤4 antibodies (59%) at P14 also parallels the results obtained using Ab CW24 alone. The contribution of ␤ subunit isoforms to the N-type VDCC solubilized from adult rat brain evidenced the following distribution: ␤1b (32 Ϯ 9.0%), ␤2 (8 Ϯ 2.8%), ␤3 (55 Ϯ 5.6%), and ␤4 (25 Ϯ 6.3%) isoforms. In comparison to the previously published findings, our results suggest a greater fractional contribution of the ␤1b and a lesser contribution of the ␤4 to the adult rat N-type VDCC (Fig. 7) and further support the observation that the adult N-type VDCC can be comprised of ␤1b, ␤3, and ␤4 isoforms (Fig. 7) (22). The sum of the total 125 I-CTX immunoprecipitated from adult brain by anti-␤1b, -␤2, -␤3, and -␤4 antibodies (120%) is similar to the amount of 125 I-CTX immunoprecipitated by either Ab CW14 or Ab CW24. DISCUSSION In the past several years, investigators have struggled to make physiological sense of the vast diversity of VDCC subunit isoforms present in neural tissues. The dramatic changes that occur in calcium conductances during neuronal maturation suggest an underlying and equally dramatic change in the density and subtype of VDCC. Indeed, in cultured cerebellar Purkinje cells, studies have demonstrated that changes in calcium conductances were critical for neuronal maturation (53). More recently, important changes were observed in the expression and differential contribution of N-type and P/Q-type VDCC during synapse formation in cultured neurons (30 -32). Similarly, changes in calcium channel currents were observed at different stages of embryogenesis (54). These studies were among the first to suggest regulation of ␣ 1 subunit expression as a possible mechanism for establishing diversity in calcium signaling during development.
In support of previous findings on the role of N-type VDCC in neuronal development, we have demonstrated an increase in the expression of ␣ 1B during rat brain development (Figs. 2 and  3) which does not correlate with the acquisition of 125 I-CTXbinding sites (Table I). The parallel increase in reactivity of Ab The amount of expressed ␣ 1B protein relative to adult rat brain was determined from the quantified Western blot analysis presented in Fig. 2.   FIG. 4. Expression of all ␤ subunit isoforms as detected by pan-specific antibody Ab CW24 during development. Tissue was obtained from animals at different ages, collected into sample buffer, and resolved by SDS-PAGE upon a 10% gel. The amount of protein was 150 g/lane. The gels were transferred to nitrocellulose and probed with Ab CW24 (1/100 dilution, A) and visualized by ECL. In experiments conducted on similar samples, the transferred antigens were detected using 125 I-IgG, exposed to film. The bands inclusive of the larger ␤ subunits ␤1 ϩ ␤2 (B) and the smaller ␤ subunits ␤3 ϩ ␤4 (C) were excised and counted in a gamma counter. The data were obtained from four determinations done in duplicate and normalized to the signal obtained at P0 (␤1b ϩ ␤2 ϭ 277 Ϯ 88 cpm band; ␤3 ϩ ␤4 ϭ 210 Ϯ 46 cpm band). * denotes p values less than 0.05. CW8 and Ab CW14 throughout development indicates that the ␣ 1B present in these samples contains two epitopes originally identified in the rat ␣ 1B cDNA (37). Functionally different isoforms of N-type VDCC have been identified in rat sympathetic ganglia (38) and embryonic (E17) tissues (55); however, these variants contain both Ab CW14 and Ab CW8 epitopes. The report of splice variants in the II-III loop of ␣ 1A subunit suggests caution in dismissing the existence of additional ␣ 1B variants. It is important to note that the region defined by Ab CW8 is coincident with a region of diversity in the ␣ 1A variants (56).
In this study we have described a population of ␣ 1B detected in embryonic and P0 brain samples that does not support high affinity 125 I-CTX binding (Table I). This property is reminis-FIG. 5. Expression of specific ␤ subunit isoforms during development. Tissue was obtained from animals at different ages, collected into sample buffer, and resolved by SDS-PAGE upon a 10% gel. The amount of protein was 150 g/lane. The gels were transferred to nitrocellulose and probed with ␤ isoform-specific antibodies (A, anti-␤1b (1/60 dilution); B, anti-␤2 antibody (1/60 dilution); C, anti-␤3 antibody (1/200 dilution); D, ␤4 antibody (1/100 dilution)). Analysis was carried out using 125 I-goat anti-rabbit IgG to detect the antigens; the filter was exposed to film and autoradiographed, and the single immunoreactive bands were excised and counted in a gamma counter. The data were normalized to the signal obtained at P0 (␤1b ϭ 325 Ϯ 176 cpm band; ␤2 ϭ 1534 Ϯ 128 cpm band; ␤3 ϭ 300 Ϯ 8 cpm band; ␤4 ϭ 31 Ϯ 17 cpm band). * denotes p values less than 0.05.  7. Differential association of ␤ isoforms with 125 I-CTXlabeled N-type VDCC during development. Solubilized N-type VDCC was prepared from P2 brain, and P14 and adult forebrain, labeled with 125 I-CTX, and immunoprecipitated with anti-␤ subunitspecific antibodies. The fractional contribution of 125 I-CTX binding recovered in the immunoprecipitated pellet by each antibody is plotted relative to the amount of total 125 I-CTX binding present in the sample prior to immunoprecipitation. Data presented are mean Ϯ S.D. from two (P2 and P14) or four (adult) determinations done in duplicate. * denotes p values less than 0.05. cent of the unassembled ␣ 1B expressed in heterologous systems in the absence of ␣ 2 /␦ and ␤ subunits (57). Also, studies on the developmental expression of the sodium channel have noted the acquisition of high affinity [ 3 H]saxitoxin binding occurring in parallel with channel assembly (58). Our results suggest that the acquisition of high affinity CTX binding during rat brain development (Table I) occurs by a mechanism that is similar to the sodium channel and reflects the conversion of the pool of unassembled ␣ 1B (present at E18 and P0) to mature ␣ 1B that are assembled with component ␣ 2 /␦ and ␤.
This is the first report to demonstrate regulation of the ␤ subunit component of a specific VDCC (N-type) during neuronal development. Previously, correlations in the localization and density of ␣ 1 and ␤ isoform mRNA in adult, embryonic, and postnatal rat brain predicted likely ␣ 1 /␤ complexes (48), but there were no conclusions reached in the case of the ␣ 1B/␤ complex. Another study concluded that there was no evidence for subunit switching in developing hippocampus as they targeted only the ␣ 1B /␤3 heteromultimers for analysis (59). In this study we have identified clear developmental trends in ␣ 1B /␤ composition from a population predominant in ␣ 1B ϩ ␤1b complexes (P2) to a population comprised of ␣ 1B ϩ ␤3 and ␣ 1B ϩ ␤4 complexes in mature rat brain.
The discrepancy between the fraction of 125 I-CTX-labeled N-type VDCC that could be identified in P2 and P14 by anti-␣ 1B antibodies versus those identified by anti-␤ antibodies (Fig. 6) suggests the presence of a ␤ isoform that is not identified by our pan-specific antibody nor recognized by ␤-specific antibodies. Alternatively, there may be some structural lability in the physical coupling of ␤ to the ␣ 1B in early development. A conserved site on the intracellular I-II loop has been identified in all ␣ 1 subunits that bind ␤ (60,61). In vitro studies have determined nanomolar affinity between all ␤ and the I-II loop interaction domain in a binding reaction that is not affected by calcium or protein kinase C phosphorylation (61). Recently, a second site that mediates ␣ 1 -␤ interaction has been identified at the C terminus of the ␣ 1E , ␣ 1A , and ␣ 1B (62,63); however, the affinity for ␤ at this site has not yet been determined. The occupancy of these two sites on ␣ 1B by ␤ adds another dimension to ␣ 1B /␤ heterogeneity.
Differential modulation of the N-type VDCC by protein kinases is another property that may result from the assembly of a specific ␤ with the ␣ 1B . The ␤1b (41) and ␤2 (25) isoforms contain consensus sites for protein kinase A modification and no consensus site for tyrosine kinases; conversely, the ␤3 isoforms contain consensus sites for tyrosine kinases and no consensus site for protein kinase A (43). In the ␤4, both protein kinase A and tyrosine kinases consensus sites are absent (39). The functional consequences of assembling different isoforms of ␤ subunit that act as substrates for different protein kinases into the N-type ␣ 1B /␤ complex suggest a possible mechanism for coupling specific intracellular signaling pathways to N-type VDCC.
The report of changes in ␣ 1 and ␤ mRNA levels during development (48) and our results (Fig. 5) suggest differential regulation of ␤ isoform expression. The effect of neurotrophic agents upon changes in calcium currents (73)(74)(75) and the specific expression of ␣ 1B and ␤ isoforms have been reported (33,34) in established neurotypic cell lines. However, these in vitro models pale in comparison to the complexity of developing rat brain. The trends identified in this study require further scrutiny at the cellular level to unravel the mechanisms that underlie both ␤ expression and its association with ␣ 1 .
We would like to address the question whether the assembly of ␤ with an ␣ 1 subunit is reflective of specific assembly processes or simply reflects the fractional contribution of a ␤ isoform relative to the pool of total available ␤. As demonstrated by our results, with the exception of the ␤4 subunit, there is no straightforward relationship that emerges between the relative contribution of ␤1b, ␤2, and ␤3 to the pool of available ␤ subunits (Fig. 5) and the contribution of that isoform to the assembled N-type VDCC (Fig. 7).
A comparison of the heterogeneity in ␣ 1B /␤ complexes through development with changes in the pool of available ␤ isoforms suggests several cellular strategies are at play which regulate ␣ 1B /␤ subunit assembly. The ␤1b subunit is detected in P2 homogenate at a fraction of its maximal adult level of expression (Fig. 5A); however, the relative amount of ␤1b associated with the P2 N-type VDCC is similar to the adult N-type VDCC (Fig. 7). These findings indicate a relative enrichment of the ␤1b in the P2 N-type VDCC complex relative to adult N-type VDCC. As previously shown, the ␤2 subunit is detected throughout rat brain development (Fig. 5B), yet antibodies to the ␤2 isoform do not immunoprecipitate 125 I-CTX binding from P2 or P14 samples. In the adult samples, less than 10% of all 125 I-CTX-labeled N-type VDCC were immunoprecipitated by anti-␤2 antibody. These findings suggest the active exclusion of the ␤2 subunit from the N-type VDCC complex. Similar to ␤2 expression, the ␤3 subunit is also expressed at a relatively constant level in the interval between P0 and adult. However, in a manner similar to the ␤4, there is a statistically significant increase in its association with the ␣ 1B during development. It is very interesting to note that the onset of increased ␤1b and ␤4 expression occurs at the beginning of a well defined period of axonal outgrowth, infiltration, and synapse formation in the rat neocortex which occurs in the first 2 weeks of postnatal life (76 -78).
Significantly, the 10-fold increase in the expression of the ␤4 isoform between P0 and adult and its parallel association with the ␣ 1B through development is in striking contrast to the other ␤ isoforms and identifies a property unique to the ␤4. Interestingly, there has been a report that demonstrates the importance of the ␤4 isoform. Analyses of the mutation that underlies the mouse lethargic phenotype (lh/lh), a model of epilepsy which does not exhibit any neurodegeneration or other neurohistological abnormalities (79), have identified an insert in the ␤4 gene that leads to a truncated gene product. Specifically, the truncation of the ␤4 subunit protein eliminates the ␣ 1 -binding domain as well as more than 60% of the protein (80). The study by Burgess et al. (80) is significant as it is the first to implicate VDCC auxiliary subunit as the basis for a neurological disease. The co-localization of the ␣ 1B and ␣ 1A with the ␤4 isoform in normal rat forebrain and cerebellum (48) and the identification of the ␤4 as a component of the adult P-type (33), N-type VDCC (22,23), and L-type VDCC (23) suggest important lines of investigation toward understanding the role of the ␤4 truncation in epileptic lh/lh mice. It will be of interest to examine if the association of the ␤4 with the other VDCC during development also occurs in parallel to its level of expression. Furthermore, the epileptic phenotype of lh/lh mice that results as a consequence of a defect in a single ␤ subunit intimates a role for the ␤4 isoform that cannot be complemented by the expression of the other ␤ isoforms. Although it is clear from our study that the ␤4 isoform is unique among ␤ subunits in its magnitude of induction and temporal pattern of expression, it would be premature to suggest that it is the absence of the ␤4 isoform per se that gives rise to the epileptic lh/lh phenotype. Alternatively, one might consider that alterations in the level of expression of full-length ␤4 in lh/lh mice may induce profound compensatory effects upon the regulation of expression of the other ␤ isoforms.