Altered Expression and Assembly of N-type Calcium Channel α1B and β Subunits in Epileptic lethargic(lh/lh) Mouse*

Voltage-dependent calcium channels (VDCC) are multisubunit complexes whose expression and targeting require the assembly of the pore-forming α1 with auxiliary β and α2/δ subunits. The developmentally regulated expression and differential assembly of β isoforms with the α1B subunit to form N-type VDCC suggested a unique role for the β4 isoform in VDCC maturation (Vance, C. L., Begg, C. M., Lee, W.-L., Haase, H., Copeland, T. D., and McEnery, M. W. (1998) J.Biol. Chem. 273, 14495–14502). The focus of this study is the expression and assembly of α1B and β isoforms in the epileptic mouse, lethargic(lh/lh), a mutant anticipated to produce a truncated β4 subunit (Burgess, D. L., Jones, J. M., Meisler, M. H., and Noebels, J. L. (1997) Cell 88, 385–392). In this report, we demonstrate that neither full-length nor truncated β4 protein is expressed in lh/lh mice. The absence of β4 inlh/lh mice is associated with decreased expression of N-type VDCC in forebrain and cerebellum. The most surprising characteristic of the lh/lh mouse is increased expression of β1b protein. This result suggests a previously unidentified cellular mechanism wherein expression of the total pool of available β subunits is under tight metabolic regulation. As a consequence of increased β1b expression, the β1b is increased in its incorporation into α1B/β complexes relative to wild type. Thus, in striking similarity to the population of N-type VDCC present in immature rat brain, the population of N-type VDCC present in adultlh/lh mice is characterized by the absence of β4 with increased β1b expression and assembly into N-type VDCC. It is intriguing to speculate that the increased excitability and susceptibility to seizures observed in the lh/lhmouse arises from the inappropriate expression of an immature population of N-type VDCC throughout neuronal development.

There has been continued effort to determine the molecular origin of epileptic seizures with the objective of identifying new therapeutic strategies (1). Recently, attention has been directed to epileptic strains of animals that exhibit absence seizures with electrographic firing patterns, onset, localization, and drug response similar to humans (2). Genetic analysis of the mouse strains tottering (tg) and leaner (ln) (3,4) has identified mutations in the ␣ 1A subunit, which, upon assembly with ␤ and ␣ 2 /␦ subunits (5), constitutes P/Q-type voltage-dependent calcium channels (VDCC). 1 Mutations in the human ␣ 1A gene, however, do not appear to be the locus of common idiopathic generalized epilepsy (6). It is important to consider that although mutations in ␣ 1 have been demonstrated to alter the biophysical properties of the VDCC (7), in vitro recombinant studies have reported modification of VDCC properties that are a consequence of differential association of ␣ 1 with specific ␤ subunit isoforms (8).
These observations are significant in light of the recent report that mutation of the ␤4 subunit is the molecular defect in the epileptic mouse strain lethargic (lh/lh) (9). The phenotype of homozygous lh/lh mice includes absence seizures, instability of gait, and convulsions (10,11). In contrast to the calcium channelopathies that underlie human spinocerebellar ataxia (SCA6) (12,13) and leaner (3,4), the cerebellum of the lh/lh mouse is structurally normal (10). Importantly, the lh gene is anticipated to produce a truncated ␤4 protein that does not possess a consensus ␣ 1 binding domain that mediates ␣ 1 /␤ interaction (14), suggesting that a defect in VDCC assembly underlies the pathogenesis of the lh/lh phenotype.
There is little information available on the mechanisms that regulate the level of expression of ␤ isoforms and their assembly with ␣ 1 . Assembly of N-type VDCC subunits has been analyzed in several developing and differentiating systems (15). During IMR32 cell differentiation, ␤1b was up-regulated and increased in parallel with the expression of ␣ 1B (16). Expression of ␤ isoforms is also highly regulated during rat brain ontogeny, with ␤1b increasing approximately 3-fold and ␤4 increasing 10-fold during the interval between postnatal day 2 (P2) and adult (17). Postnatal assembly of ␤ isoforms with ␣ 1B to form N-type VDCC indicated differential association of ␤ isoforms in immature versus mature forebrain homogenates. ␤1b was the predominant ␤ detected in assembled immature N-type VDCC at P2 (17). ␤4 was not detected as a component of immature N-type VDCC and was incorporated into mature N-type VDCC with a time course that paralleled its expression (17). Thus, differences in the ␤ component of the N-type VDCC defined both an immature and a mature population of N-type VDCC (17). Although the significance of ␤ heterogeneity has not been fully explored in vivo, these developmental studies suggested a unique role of ␤4 in N-type VDCC maturation. The lh/lh mouse offers the opportunity to study patterns of ␤ isoform expression that occur in response to abnormal expression of ␤4. The focus of this study is the level of ␣ 1B and ␤ subunit expression and assembly of N-type VDCC in the lh/lh mouse with emphasis upon identifying possible compensatory mechanisms that occur from altered ␤4 expression.

EXPERIMENTAL PROCEDURES
Lethargic (B6EiC3H-a/A-lh) and wild-type mice (strain B6EiC3H) were obtained from Jackson Labs. All reagents were obtained from sources previously cited (17). Adult mice were euthanized in accordance with accepted university guidelines, and the brains were removed and immediately placed in 50 mM Hepes, pH 7.4, 1 mM EGTA plus protease inhibitor mixture (17). The tissues were homogenized with a Polytron homogenizer for 10 s and centrifuged at 18,000 rpm (48,000 ϫ g) for 15 min. The membranes were resuspended in 50 mM Hepes, pH 7.4, plus protease inhibitors at a resulting protein concentration of 50 mg/ml. The N-type VDCC was solubilized from forebrain and cerebellar membranes of wild-type and lh/lh mice as described previously (18). For Western blot analysis, all homogenates were stored at 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. The production of anti-peptide polyclonal antibodies to VDCC subunit epitopes has been described previously (16,17,19,20). Methods for 125 I-CTX binding, Scatchard analysis (21,22), quantitative Western blot analysis using 125 I-goat anti-rabbit IgG, immunoprecipitation of N-type VDCC, and all other general methods have been described in detail (17,19). The results are expressed as mean Ϯ S.D. Statistical analysis was performed by a paired t test or Mann-Whitney Rank Sum test. p values less than 0.05 were considered significant.

RESULTS AND DISCUSSION
␤4 Isoform Is Not Expressed in lh/lh Mice-The lh gene mutation was anticipated to lead to truncation of ␤4 to an N-terminal fragment predicted to have a mass of 21 kDa (9). The lh mRNA was detected at levels 20% of the wild-type ␤4 message (9), suggesting the possibility that the lh gene product may be expressed in lh/lh mice. Using ␤4-specific antibodies, we probed forebrain and cerebellar homogenates from lh/lh mice to evaluate the level of expression of full-length ␤4. In contrast to wild-type mice where we detected full-length ␤4 (62 kDa), there was no ␤4 detected in either forebrain or cerebellum from lh/lh mice (Fig. 1). To further investigate expression of the lh gene product, we used an antibody (Ab CW24) raised to amino acids 53-70 in the ␤4 which are also present in all ␤ isoforms (16,17). Ab CW24 identified two populations of high (␤1b and ␤2) and low (␤3 and ␤4) molecular weight ␤ isoforms previously characterized in rat brain (17). The relative intensity of these bands clearly differed among the lh/lh versus wild-type mouse samples (Fig. 1). However, with the exception of 42-40-kDa proteolytic ␤ fragments, there were no detectable Ab CW24-immunoreactive proteins that could be attributed to the predicted 21-kDa product of the lh gene in either lh/lh mouse forebrain or cerebellum. These results suggest that the lh mutation causes a complete loss of ␤4 protein.
The Pool of Available ␤ Subunits Is Decreased in lh/lh Mice-To investigate the pool of available ␤ isoforms in forebrain and cerebellum of lh/lh and wild-type samples, the level of expression of all ␤ isoforms was quantified using a panspecific anti-␤ antibody (Ab CW24) and a panel of ␤ isoformspecific antibodies. There are regional differences in expression of ␤ isoforms with increased expression of all ␤ isoforms (with the exception of the ␤4) in forebrain samples. Significantly, we observed differences in expression among specific ␤ isoforms in lh/lh mice compared with wild-type mice. The level of expression of all ␤ isoforms as detected by the anti-␤ pan-specific antibody is lower in lh/lh forebrain and cerebellum than in wild-type samples (Fig. 2), indicating that the level of total ␤ isoforms is not maintained in the lh/lh samples. In both lh/lh forebrain (p Ͻ 0.001) and cerebellum (p Ͻ 0.05), the level of expression of ␤1b was increased compared with wild-type mice (Fig. 2). In forebrain, the increase in ␤1b expression was greater than 50%. These results are consistent with our previous characterization of ␤1b as an inducible and regulated protein (16,17). In contrast, differences in the levels of expression of ␤2 and ␤3 in lh/lh versus wild-type mice were not statistically significant in either forebrain or cerebellar samples.
Decreased Expression of N-type VDCC and ␣ 1B in lh/lh Compared with Wild-type Mice-The density of N-type VDCC has been previously shown to be higher in forebrain versus cerebellar samples (21,23,24). Furthermore, ␤4 is the predominant were resolved by SDS-PAGE, transferred to nitrocellulose, and incubated with affinity-purified antibodies pan-specific for all ␤ and isoform-specific antibodies to ␤1b, ␤2, ␤3, and ␤4. The amount of ␤ was quantified using 125 I-IgG. Results obtained were from duplicate blots representing n ϭ 3 wild-type (Ⅺ) and 3 lh/lh (f) animals for forebrain samples and n ϭ 2 wild-type and 3 lh/lh animals for cerebellar samples; **, p Ͻ 0.001 and *, p Ͻ 0.05 as determined by a paired t test.
isoform associated with VDCC from cerebellum, and ␤3 is the predominant isoform associated with VDCC from forebrain (5,17,20,25). Therefore, these patterns of VDCC subunit expression suggested regional differences in acquisition of functional N-type VDCC in cerebellum and forebrain from lh/lh versus wild-type mice. Using 125 I-CTX radioligand binding assays and Scatchard analyses (18) (Fig. 3), we observed a significant decrease (p Ͻ 0.05) in expression of N-type VDCC in lh/lh forebrain (1.49 Ϯ 0.41 pmol/mg) compared with wild-type forebrain (2.70 Ϯ 0.63 pmol/mg). There was a single 125 I-CTX binding site detected in the forebrain samples with K d values of approximately 28 pM for both the lh/lh and wild-type samples. The level of ␣ 1B expressed in forebrain samples was also quantified (Fig. 3) to examine possible discrepancies between expression of 125 I-CTX binding sites and ␣ 1B protein (17). Despite the decrease in 125 I-CTX binding sites in lh/lh forebrain, similar levels of ␣ 1B protein are expressed in forebrain of lh/lh and wild-type mice. These data strongly suggest that expression of ␣ 1B protein in forebrains of lh/lh mice is maintained at wildtype levels, while the assembly of ␣ 1B into a complex that can support 125 I-CTX binding is compromised. The decreased expression of 125 I-CTX binding sites in lh/lh forebrain (Fig. 3) may reflect the decreased availability of ␤ (Fig. 2) required to traffic ␣ 1B to the plasma membrane (26).
We also observed decreased expression (p Ͻ 0.05) of 125 I-CTX binding sites in lh/lh cerebellum (0.32 Ϯ 0.05 pmol/mg) compared with wild-type cerebellum (0.53 Ϯ 0.10 pmol/mg). How-ever, in contrast to the forebrain samples, radioligand binding experiments detected two 125 I-CTX binding sites in cerebellum. The high affinity site (K d values of approximately 70 and 67 pM for the lh/lh and wild-type cerebellar samples, respectively) is characteristic of the N-type VDCC. The low affinity site for 125 I-CTX detected in cerebellar samples is likely because of the low affinity binding of 125 I-CTX for the P/Q-type VDCC (5) and was not pursued further in these studies. In contrast to lh/lh forebrain, decreased ␣ 1B protein is expressed in lh/lh mouse cerebellum, suggesting that expression of ␣ 1B protein is not maintained at wild-type levels. It seems reasonable to consider that the loss of the ␤4 from lh/lh cerebellum cannot be entirely compensated despite the increased level of expression of ␤1b (Fig. 2). The decreased expression of N-type VDCC or altered expression of other VDCC in the cerebellum of the lh/lh mouse may be the molecular basis of ataxia associated with the lh/lh phenotype. It should be stated that the level of expression of functional N-type VDCC in sympathetic neurons was also decreased in "␤3 knock-out mice" (27). However, in contrast to the lh/lh mouse, the "␤3 knock-out" mouse is phenotypically normal (27). The expression of other ␤ isoforms in response to the elimination of ␤3 has not yet been reported.
Increased Incorporation of ␤1b into N-type VDCC of the lh/lh Mouse-To determine the structural consequences of abnormal ␤ isoform expression in lh/lh mice upon N-type VDCC assembly, the endogenous ␣ 1B /␤ subunit complexes were evaluated in immunoprecipitation assays using anti-␣ 1B , anti-␤ generic (Ab CW24), and ␤ isoform-specific antibodies (17). The assay conditions were defined such that the pan-specific anti-␤ antibody immunoprecipitated a similar fraction of N-type VDCC in all FIG. 3. Differential expression of ␣ 1B and N-type VDCC in wildtype and lh/lh mouse brain. Forebrains (FB) and cerebella (CB) from lh/lh and wild-type mice were removed, resuspended in 50 mM HEPES, 1 mM EGTA, and protease inhibitors, and homogenized. The samples (150 g/lane) were resolved by SDS-PAGE, transferred to nitrocellulose, and incubated with affinity-purified antibodies to ␣ 1B subunit (Ab CW14). The amount of ␣ 1B was quantified using 125 I-IgG. The results obtained were from duplicate blots representing n ϭ 3 wild-type (Ⅺ) and 3 lh/lh (f) mice for forebrain samples and n ϭ 2 wild-type and 3 lh/lh mice for cerebellar samples; **, p Ͻ 0.001, as determined by a paired t test. Scatchard plot analysis of 125 I-CTX binding to N-type VDCC in wild-type and lh/lh mouse forebrain and cerebellum was carried out as described (17,18). The amount of protein per assay was: lh/lh forebrain, 3-5 g; wild-type forebrain, 2-4 g; lh/lh cerebellum, 11-20 g; and wild-type cerebellum, 9 -20 g. The following K d values were calculated: lh/lh forebrain, 27.8 (Ϯ 7.6) pM; wild-type forebrain, 28.4 (Ϯ 15.3) pM; lh/lh cerebellum, 75.6 (Ϯ 10.0) pM; and wild-type cerebellum, 66.6 (Ϯ 5.0) pM. The results obtained were from tissue samples representing n ϭ 4 wild-type and 3 lh/lh animals for forebrain samples and n ϭ 3 wild-type and 3 lh/lh animals for cerebellar samples; *, p Ͻ 0.05, as determined by a paired t test.
FIG. 4. Altered ␣ 1B /␤ subunit assembly leads to different populations of N-type VDCC in wild-type and lh/lh brain. Forebrains (FB) and cerebella (CB) from lh/lh (f) and wild-type (Ⅺ) mice were solubilized with 0.75% CHAPS incubated with 125 I-CTX (9000 -12,000 cpm/assay) for 30 min. Affinity-purified antibodies to the ␣ 1B (Ab CW14), pan-specific anti-␤ antibodies, and isoform-specific antibodies to ␤1b, ␤2, ␤3, and ␤4 (25 g/assay) were added for 1 h (17). Protein A-Sepharose 4B was added with constant mixing. 125 I-CTX bound to the immunoprecipitated N-type VDCC was recovered in protein A pellets, counted, and normalized to the fraction of 125 I-CTX immunoprecipitated by Ab CW14 (17,19). The approximate amount of total 125 I-CTX binding/sample prior to immunoprecipitation was in the range of 1000 -1600 specific counts/min of 125 I-CTX for the cerebellar samples and 2500 -8000 specific counts/min of 125 I-CTX for the forebrain samples. Results obtained were from n ϭ 4 wild-type and 4 lh/lh animals carried out in duplicate; *, p Ͻ 0.05 and **, p Ͻ 0.001 as determined by a paired t test. samples (Fig. 4). The relative contribution of ␤ isoforms to the N-type VDCC present in forebrain and cerebellum of lh/lh is clearly altered compared with the wild-type mice (Fig. 4). As anticipated from the lack of detectable ␤4 in forebrains from lh/lh mice (Fig. 1), the disparity in the association of ␤4 with N-type VDCC in lh/lh versus wild-type mice was quite dramatic, as antibodies to ␤4 immunoprecipitated less than 10% of the total N-type VDCC solubilized from the cerebellum of lh/lh mouse (Fig. 4). With regard to the N-type VDCC extracted from forebrain, the fraction of N-type VDCC associated with ␤1b was statistically increased in lh/lh versus wild type mice (Fig. 4). In contrast, neither the association of ␤2 nor ␤3 with the N-type VDCC was affected in forebrains of lh/lh versus wild-type mice (Fig. 4). However, although ␤1b was increased in expression in lh/lh cerebellum (Fig. 2), there was no statistically significant increase in incorporation of any ␤ isoform into cerebellar Ntype VDCC (Fig. 4). It is reasonable to suggest that downregulation of ␣ 1B expression (Fig. 3), rather than increased incorporation of ␤1b into assembled N-type VDCC, is the primary mechanism of compensation in lh/lh cerebellum. Additional studies are required to determine whether the compensatory mechanisms that alter ␤ subunit composition of the N-type VDCC in the lh/lh mouse influence the expression and assembly of other VDCC.
Although the specific biophysical properties derived from the population of N-type VDCC present in wild-type and lh/lh forebrain have yet to be determined, it is interesting to note that ␤1b and ␤4 have similar effects upon closed state inactivation of recombinant N-type VDCC (8). Similar kinetic effects of ␤1b and ␤4 suggest tolerance of the ␤1b assembled into N-type VDCC, and this atypical channel composition may explain the absence of the neurodegeneration frequently observed in other epileptic mouse strains (3,4).
These results are the first to indicate that assembly of the high voltage-activated N-type VDCC is altered in the lh/lh mouse. These findings do not exclude the possibility that the expression of other high voltage-activated VDCC is also effected as ␤4 is associated with mature L-type, N-type, and P/Q-type VDCC (17,20,25). However, it is interesting to point out that although low voltage-activated T-type channels have been implicated in the initiation of thalamic seizures in absence epilepsies (28,29), the T-type ␣ 1G and ␣ 1H isoforms do not contain consensus ␤ binding domains (14,30), suggesting that T-type VDCC expression, unlike the high voltage-activated VDCC, may not be directly regulated by ␤ subunits. Differential modulation of the N-type VDCC by protein kinases in lh/lh mice is another property that may result from the assembly of ␤1b in place of ␤4. The ␤1b (31) contains consensus sites for protein kinase A modification; conversely in the ␤4, the protein kinase A consensus sites are absent (32). Thus, the inappropriate inclusion of ␤1b into the N-type VDCC complex in the lh/lh mouse in lieu of ␤4 may alter protein kinase-mediated modulation of the channel and thus effect calcium entry and calcium-dependent signaling.
␤ Subunit Composition of N-type VDCC Expressed in lh/lh Mice Resembles N-type VDCC Population of Immature (P2) Neurons-In earlier studies, ␤4 was discriminated from the other ␤ isoforms by virtue of its striking increase in expression during development (17,33). The importance of ␤4 to neuronal functioning is reflected in the epileptic and ataxic phenotype of the lh/lh mice, which stands in contrast to the "␤3 knock-out" mouse that is phenotypically normal (27). The phenotype of lh/lh mice is evident at postnatal day 15 (10,11), which is consistent with the loss of ␤4 that is normally increased in expression after P7 in developing rat brain (17). The question now arises as to whether the phenotype of lh/lh mice arises primarily because of the loss of ␤4 or as a result of the increased fractional contribution of ␤1b to N-type VDCC complexes. Our recent report that identifies ␤4 as a marker for N-type VDCC maturation unifies these two possibilities (17). The increased fractional contribution of ␤1b to N-type VDCC complexes and the absence of ␤4 assembled into adult lh/lh N-type VDCC result in a population of N-type VDCC that is strikingly similar to immature (P2) N-type VDCC in ␤ subunit composition (17). We propose that the mechanism that promotes absence seizures in lh/lh mice, a form of epilepsy more commonly associated with immature brain (28), may be a consequence of prolonged and inappropriate expression of immature ␣ 1 /␤ complexes.