The juvenile myoclonic epilepsy mutant of the calcium channel β(4) subunit displays normal nuclear targeting in nerve and muscle cells.

Voltage-gated calcium channels regulate gene expression by controlling calcium entry through the plasma membrane and by direct interactions of channel fragments and auxiliary β subunits with promoters and the epigenetic machinery in the nucleus. Mutations of the calcium channel β(4) subunit gene (CACNB4) cause juvenile myoclonic epilepsy in humans and ataxia and epileptic seizures in mice. Recently a model has been proposed according to which failed nuclear translocation of the truncated β(4) subunit R482X mutation resulted in altered transcriptional regulation and consequently in neurological disease. Here we examined the nuclear targeting properties of the truncated β(4b(1–481)) subunit in tsA-201 cells, skeletal myotubes, and in hippocampal neurons. Contrary to expectation, nuclear targeting of β(4b(1–481)) was not reduced compared with full-length β(4b) in any one of the three cell systems. These findings oppose an essential role of the β(4) distal C-terminus in nuclear targeting and challenge the idea that the nuclear function of calcium channel β(4) subunits is critically involved in the etiology of epilepsy and ataxia in patients and mouse models with mutations in the CACNB4 gene.

Introduction

The auxiliary β subunits of voltage-gated calcium channels promote membrane expression and modulate the gating properties of CaV1 and CaV2 calcium channels. In humans four genes encode CaVβ subunits and abundant alternative splicing further increases the molecular heterogeneity of the β subunit family. Coexpression studies demonstrated that all β isoforms promote membrane expression of any CaV1 and CaV2 channel isoform and modulate their gating properties in a similar way.

These highly promiscuous isoform interactions generate a considerable functional redundancy of β subunits. Consequently, loss-of-function mutations and knockouts of β subunit genes caused a disease phenotype primarily in those tissues that express exclusively the mutated β isoform.- In contrast, in brain, where all four β subunit genes are expressed, phenotypes were mild or non-existent, most likely because the channel function of the mutated β isoform was compensated by other β isoforms.- However, there is one notable exception: spontaneous mutations of the β4 subunit lead to idiopathic generalized epilepsy and episodic ataxia in humans and in mice., In cerebellum β4 and α2δ-2 are the major subunit partners of the P/Q-type (CaV2.1) calcium channel. Interestingly, mutations of all three subunit isoforms (CaV2.1, α2δ-2 and β4) result in an epileptic and ataxic phenotype.,, This is consistent with the notion that a deficiency of the P/Q-type channel function causes the neurological disease in β4 mutants.

Recently, we and others discovered that specific β4 subunit isoforms can also accumulate in the nucleus.- This unexpected finding suggested a role of β4 in channel-independent cell functions. Furthermore, in excitable cells nuclear export of β4b was shown to be activity-dependent. We demonstrated that in skeletal myotubes and in hippocampal neurons β4b accumulated in the nuclei during early development and in electrically quiescent cells and that it was rapidly exported in response to depolarization., Because a truncated β4c isoform has previously been shown to interact with the nuclear protein HP1γ, a possible function in gene regulation has been suggested., This calcium channel-independent nuclear function provided an alternative explanation for the etiology of the severe neurological phenotype of β4 mutations. If mutated β4 subunits differ from wildtype β4 isoforms with regard to their nuclear targeting properties or their ability to interact with nuclear proteins, then the loss of the nuclear function of β4 may cause the neurological deficits observed in human patients and in mouse models with mutations in the β4 gene. Consistent with this idea, Tadmouri et al. reported that the ataxia mutation R482X resulted in a C-terminally truncated β4 protein, which failed to be targeted into the nucleus and consequently did not interact with the regulatory protein complex shown to repress tyrosine hydroxylase expression. Moreover, in a follow-up study the same group showed that heterologous expression of the full-length and truncated β4b isoform in HEK293 cells resulted in differential gene expression.

Because we recently could demonstrate that in neurons only those β4 splice variants capable of targeting to the nucleus also regulated genes, we set out to examine the nuclear targeting function of the truncated ataxia mutant in nerve and muscle cells. Unexpectedly, however, wildtype and truncated β4b variants displayed identical nuclear targeting properties. Thus, the data presented here contradict the previous report and challenge the idea that differences in transcriptional regulation due to differential nuclear targeting of the wildtype and mutated β4b subunits may account for the neurological phenotypes in humans and mice with mutations in the CACNB4 gene.

Results

Similar incorporation into calcium channel complexes and nuclear targeting of the full-length and truncated β4b subunits in skeletal myotubes

In order to analyze the nuclear targeting properties of the wildtype and ataxia mutant of the β4 subunit in muscle and nerve cells, we generated a truncated β4b construct lacking the 39 C-terminal residues (β4b(1–481)), (Fig. 1A). Both the wildtype β4b and truncated β4b(1–481) were V5-tagged at the C-terminus to enable specific localization of the heterologous β4 subunits in neurons expressing also endogenous β4. Extensive previous analysis, demonstrated that the V5-tagged β4 subunits can functionally interact with CaV channels in the membrane and show normal nuclear targeting properties when expressed in muscle or nerve cells. Western blot analysis of the full-length β4b-V5 and the truncated β4b(1–481)-V5 constructs confirms that the two β4 subunits express as intact proteins of the expected size (Fig. 1G).

Figure 1. Nuclear targeting of V5-tagged wildtype and mutant β4b subunits in dysgenic myotubes. (A) Domain structure of the full-length β4b-V5 and truncated β4b(1–481)-V5 subunits. Colored symbols indicate positions of antibody epitopes and numbers above indicate amino acid positions at domain borders and truncation site. (B) Representative double-immunofluorescence images of myotubes transfected with β1a-V5, β4b-V5, and β4b(1–481)-V5 together with GFP-CaV1.1, labeled with anti-GFP and anti-V5 (left) or anti-β (right). (C) Fraction of myotubes showing nuclear targeting, transfected and labeled as in (B) (β4b, β4b(1–481): n = 3; anti-V5 n = 120, anti-β n = 150). (D) Double-immunofluorescence images of myotubes transfected with β1a-V5, β4b-V5, and β4b(1–481)-V5 together with GFP-CaV1.2, labeled with anti-GFP and anti-V5 (left) or anti-β (right). (E) Fraction of myotubes showing nuclear targeting, transfected and labeled as in D (β4b, β4b(1–481): n = 4; anti-V5 n = 150, anti-β n = 210). Note that all β subunits co-cluster with the CaV1 channels throughout the myotubes, but only the two β4b subunit constructs accumulate in the nuclei. (F) Nucleus/cytoplasm ratios of myotubes labeled with anti-V5. ANOVA for β subunits co-expressed with GFP-CaV1.1 (β4b, β4b(1–481): n = 3, n = 60; β1a: n = 1, n = 20): F(2,137) = 23.8 P = 1.3 e−09; ANOVA for β subunits co-expressed with GFP-CaV1.2 (β4b, β4b(1–481): n = 4, n = 80; β1a: n = 1, n = 20): F(2,177) = 1.7; P = 2.6 e−05 (P values in the figure are for post-hoc analysis; ***P < 0.001). Scale bars, 10µm. (G) western blot analysis of β4b-V5 and β4b(1–481)-V5 in dysgenic myotubes with anti-V5 and anti-β4 antibodies reveals that both proteins are expressed at similar levels and at the expected size, 15 s exposure (n = 4).

First we expressed the full-length β4b-V5 and truncated β4b(1–481)-V5 subunits in dysgenic myotubes. These muscle cells lack the endogenous CaV1.1 channel but otherwise express the full complement of calcium signaling proteins, including the auxiliary CaV β1a and α2δ-1 subunits and the ryanodine receptor. Therefore transfection with CaV subunits reconstitutes the calcium channel in dysgenic myotubes and incorporates the heterologous β subunits into the functional excitation-contraction coupling apparatus. Morphologically this is seen as co-clusters of the heterologously expressed CaV α1 and β subunits in peripheral couplings and developing triads. When the β subunit constructs (β1a-V5, β4b-V5, β4b(1–481)-V5) were coexpressed with the pore-forming subunit GFP-CaV1.1 and immunolabeled with antibodies against GFP and the V5 tag or the β1 or β4 proteins, the calcium channel subunits co-clustered at the cell surface (Fig. 1B). Qualitatively, co-clustering of β4b-V5 and CaV1.1 was not different from that of β1a-V5 and CaV1.1, although quantitatively co-clustering of the native skeletal muscle subunit partners (β1a-V5 and CaV1.1) was more robust than that of the heterologous pair (β4b-V5 and CaV1.1). Nevertheless, co-clustering with CaV1.1 confirms previous findings showing that β4b can interact with CaV1.1, and it further demonstrates that the C-terminal truncation does not perturb the interaction of β4b(1–481)-V5 with the skeletal muscle calcium channel complex. Consistent with previous findings,, normal incorporation of CaV subunits in triads and peripheral couplings was not limited to the skeletal muscle CaV1.1 channel. Also co-expression of the β4b-V5 constructs and GFP-CaV1.2 resulted in the typical clustered distribution of both channel subunits in dysgenic myotubes (Fig. 1D) and the truncated β4b(1–481)-V5 isoform was as efficiently incorporated into the calcium channel complex as the full-length β4b-V5.

In addition to its incorporation into the channel complexes at the membrane β4b-V5 accumulated in the nuclei of the dysgenic myotubes (Fig. 1B). As previously shown this nuclear targeting was specific to the β4b-V5 isoform and rarely observed with β1a-V5. Unexpectedly however, the truncated β4b(1–481)-V5 construct also accumulated in the nuclei of the myotubes. Co-clustering with the CaV1.1 channel and the accumulation of both β4b-V5 and β4b(1–481)-V5 in the nuclei was observed with the V5 tag antibody (Fig. 1B, left panels) as well as with the β4 antibody (Fig. 1B, right panels). The prevalence of nuclear targeting was quantified by assessing the fraction of transfected differentiated myotubes showing nuclear V5 or β4 staining (Fig. 1C; Table 1). Whereas no myotubes with nuclear targeting of β1a-V5 were observed, the β4b-V5 constructs were targeted into the nuclei of approximately 80% of the myotubes when co-expressed with CaV1.1, both when stained with the V5 or with the β4 antibody. Most importantly, the truncated β4b(1–481)-V5 was as frequently found in the nuclei as the full-length β4b construct. To determine whether the extent of nuclear targeting differed between full-length β4b and the truncated β4b(1–481)-V5, we analyzed the nucleus to cytoplasm ratio of the anti-V5 labeled constructs (Fig. 1F). The nucleus/cytoplasm ratio of the control β1a-V5 staining was near 1, owing to the uniform distribution of β1a-V5 clusters throughout the myotubes. In contrast, the nucleus/cytoplasm ratios of both β4b-V5 constructs were above 2, reflecting their strong nuclear staining. The nucleus/cytoplasm ratio of the truncated β4b(1–481)-V5 was not statistically different from that of the full-length β4b-V5 (Table 1).

Table 1. Nuclear Targeting

Dysgenic myotubes
	β1a-V5						β4b-V5						β4b(1–481)-V5						t test
	anti-V5			anti-β1			anti-V5			anti-β4			anti-V5			anti-β4			anti-V5	anti-β4
+ GFP-CaV1.1	0,0%			0,0%			87,2	±	7,8%	81,7	±	9,3%	81,1	±	6,2%	79,4	±	10,0%	P = 0.57	P = 0.88
+ GFP-CaV1.2	0,0%			0,0%			62,1	±	9,8%	59,6	±	10,4%	69,6	±	9,5%	63,8	±	10,5%	P = 0.89	P = 0.79
							β4b						β4b(1−481)						t test
										anti-β4						anti-β4				anti-β4
+ GFP-CaV1.1										90,6	±	1,1%				90,8	±	2,1%		P = 0.91
+ GFP-CaV1.2										74,6	±	3,7%				78,8	±	2,8%		P = 0.41
tsA-201
	β1a-V5						β4b-V5						β4b(1–481)-V5						t test
	anti-V5			anti-β1			anti-V5			anti-β4			anti-V5			anti-β4			anti-V5	anti-b4
w/o	0	±	0,0%	0	±	0,0%	98,9	±	0,6%	98,9	±	1,1%	99,8	±	0,7%	94,4	±	2,2%	P = 0.79	P = 0.15
+ GFP-CaV1.2/α2δ-1	0	±	0,0%	0	±	0,0%	60,6	±	1,1%	64,4	±	4,8%	65,6	±	10,1%	63,3	±	8,4%	P = 0.65	P = 0.91

t test between β4b-V5 and β4b(1–481)-V5, β1a-V5 is shown as comparison. For n values refer to the legends of Figure 1 (dysgenic myotubes) and Figure 3 (tsA-201)

Figure 3. Nuclear targeting of full-length β4b and truncated β4b(1–481) subunits in hippocampal neurons differentiating in culture. (A) Cultured hippocampal neurons were transfected at DIV 0 (4h after plating) with either β4b-V5 or β4b(1–481)-V5, fixed and immunolabeled with an antibody against the C-terminal V5 epitope at 1, 2, 3, 5, 14 and 21 d in culture (DIV). At DIV21 one set of cultures was treated with 1µM TTX over night to block spontaneous electrical activity. Scale bar, 10µm. (B) Nucleus/cytoplasm ratio of cultures shown in (A); including the TTX-treated DIV21 neurons; n = 5, n = 15–21 (*** = P < 0.001, unpaired t test). (C) DIV 21 hippocampal neurons double-labeled with anti-V5 and anti-CaV2.1 show similar distribution of β4b and β4b(1–481) partially overlapping with synaptic CaV2.1 clusters. Scale bars, 10µm.

To examine whether these targeting properties depended on the coexpressed α1 subunit, the experiments were repeated with β1a-V5, β4b-V5, and β4b(1–481)-V5 coexpressed with the cardiac/neuronal CaV1.2 channel isoform. Figure 1D shows that β4b-V5 and β4b(1–481)-V5, but not β1a-V5, were targeted into the nuclei of the myotubes. This was equally seen when labeled with the V5 or with the β antibody. Also counting the frequency of myotubes with nuclear targeting and analyzing the nucleus/cytoplasm ratios failed to detect any significant difference in the nuclear targeting of β4b-V5 and β4b(1–481)-V5 (Fig. 1E and F; Tables 1 and 2). Together these results demonstrate that—when co-expressed with L-type calcium channel α1 subunits in dysgenic myotubes—the truncated ataxia mutant β4b(1–481)-V5 was efficiently targeted into the nuclei of as many cells as the full-length β4b-V5 subunit.

Table 2. Nucleus/Cytoplasm ratio

Dysgenic myotubes
	β1a-V5			β4b-V5			β4b(1–481)-V5			ANOVA*
+ GFP-CaV1.1	0.95	±	0.03	2.25	±	0.11	2.42	±	0.13	P < 0.001
+ GFP-CaV1.2	1.04	±	0.03	2.03	±	0.11	2.17	±	0.12	P < 0.001
Hippocampal neurons
				β4b-V5			β4b(1–481)-V5			t test
DIV1				1.55	±	0.03	1.42	±	0.03	P = 0.66
DIV2				1.47	±	0.02	1.52	±	0.02	P = 0.21
DIV3				1.36	±	0.01	1.44	±	0.01	P = 0.79
DIV5				0.8	±	0.02	0.75	±	0.01	P = 0.12
DIV14				0.72	±	0.01	0.78	±	0.02	P = 0.84
DIV21				0.81	±	0.01	0.72	±	0.01	P = 0.33
DIV21 + TTX				1.54	±	0.03	1.57	±	0.03	P = 0.77

For ANOVA paramenters and n values refer to the legends of Figure 1 (dysgenic myotubes) and Figure 2 (hippocampal neurons).

Figure 2. Nuclear targeting of untagged wildtype and mutant β4b subunits in dysgenic myotubes. (A) Domain structure of the full-length β4b and truncated β4b(1–481) subunits. Colored symbols indicate positions of antibody epitopes and numbers above indicate amino acid positions at domain borders and truncation site. (B) Representative double-immunofluorescence images of myotubes transfected with β4b, and β4b(1–481) together with GFP-CaV1.1 (left) or GFP-CaV1.2 (right), labeled with anti-GFP and anti-β4. C: Fraction of myotubes showing nuclear targeting, transfected and labeled as in (B) (β4, β4(1–481) with GFP-CaV1.1: n = 4 n = 360; (β4, β4(1–481) with GFP-CaV1.2: n = 3 n = 240).

To exclude the possibility that the C-terminal V5 tag affected the nuclear targeting properties of the heterologous β4b subunits, we generated two corresponding β4b constructs (β4b and β4b(1–481)) without the V5 tag (Fig. 2A). When expressed in dysgenic myotubes together with either GFP-CaV1.1 or GFP-CaV1.2 and immunolabeled with the anti-β4 antibody, both β4b and β4b(1–481) were observed in co-clusters with the CaV1 subunits, confirming their expected association with the calcium channels in the membrane, and both β4b and β4b(1–481) were concentrated in the nuclei (Fig. 2B). Semiquantitative analysis showed that the untagged β4b subunits were targeted into the nuclei as efficiently as the V5-tagged versions (compare Figure 2C with Figure 1C and E). Again, there were no statistically significant differences in number of myotubes showing nuclear targeting between the full-length β4b and the truncated β4b(1–481) isoform (Table 1). These, experiments clearly demonstrate that fusing a V5 antibody-tag to the C-terminus of neither the full-length β4b nor to the truncated C-terminus of β4b(1–481) alters their nuclear targeting properties.

Nuclear targeting of the full-length β4b-V5 and the truncated β4b(1–481)-V5 subunits in cultured hippocampal neurons

Because we did not detect reduced nuclear targeting properties of the truncated β4b(1–481)-V5 subunit in the skeletal myotubes, we decided to directly compare nuclear targeting of β4b-V5 and β4b(1–481)-V5 in cultured hippocampal neurons. These neurons express all β isoforms in pre- and post-synaptic compartments throughout the neurons., In addition the β4b isoform is specifically targeted into the nuclei of young (DIV1–4) and electrically silenced differentiated neurons (DIV17)., Here we transfected hippocampal neurons with β4b-V5 and β4b(1–481)-V5 and immunolabeled them with anti-V5 to specifically detect the recombinant β4b-V5 constructs. Both β4b-V5 subunits were localized in a punctate distribution pattern in the soma and throughout the neuronal processes (Fig. 3C), indicative of their incorporation in pre- and postsynaptic calcium channel complexes. In neurons at DIV1, 2, and 3 β4b-V5 as well as β4b(1–481)-V5 also labeled the neuronal nuclei, whereas at later developmental stages (DIV5, 14, and 21) the nuclei were devoid of β4b staining (Fig. 3A). Measuring the fluorescent staining intensity of the nucleus and cytoplasm of neurons and calculating the nucleus/cytoplasm ratio showed a high ratio during the first three days in culture followed by a rapid decline and continued low nucleus/cytoplasm ratio from DIV5 onward (Fig. 3B). Importantly, the nuclear targeting at the early developmental stage as well as the lack thereof in differentiated neurons was identical for β4b-V5 and β4b(1–481)-V5 (Table 2).

Because previously we detected that nuclear export of β4b in differentiated neurons was activity dependent, we examined whether this was also the case for the truncated β4b(1–481)-V5 subunit. Therefore we blocked spontaneous electric activity in three weeks old (DIV21) hippocampal neurons by overnight incubation with 1 µM TTX just prior to fixation and immunolabeling. As shown in Figure 3A and B, TTX treatment restored nuclear targeting of both β4b-V5 and β4b(1–481)-V5 to the same levels as observed in the young neurons. Thus, the full-length β4b-V5 and the truncated β4b(1–481)-V5 subunits do not differ with respect to their nuclear targeting properties in neurons. Both accumulate in the nuclei of young, presumably electrically silent hippocampal neurons, and in differentiated neurons when electrical activity is blocked.

Similar nuclear targeting properties of the full-length β4b-V5 and the truncated β4b(1–481)-V5 subunits in tsA-201 cells

So far our results demonstrate that truncation of the C-terminus of β4b does not interfere with its nuclear targeting properties in muscle and nerve cells. Because previous studies reported a failure of nuclear targeting of β4b(1–481) in CHO and HEK293 cells, we next examined the possibility that this failed nuclear targeting of β4b(1–481) might be particular to non-excitable cells. Therefore we also analyzed the nuclear targeting properties in tsA-201 cells transfected with β4b-V5 and β4b(1–481)-V5 alone and in combination with GFP-CaV1.2 and α2δ-1 subunits. As above, β1a-V5 was used as control and all conditions were immunolabeled and analyzed with anti-V5 as well as with specific β1 and β4 antibodies. When the β subunits were expressed alone, β1a-V5 was localized in the cytoplasm but not in the nucleus. In contrast, both β4b-V5 and the truncated β4b(1–481)-V5 accumulated in the nuclei of tsA-201 cells (Fig. 4A). When coexpressed with GFP-CaV1.2, the α1 and β subunits formed co-aggregates in the cell periphery (Fig. 4B), indicative of expression of the channel complexes in the plasma membrane. In addition, the β4b-V5 subunits, but not β1a-V5 or GFP-CaV1.2, also accumulated in the nuclei. Again, no differences in the membrane and nuclear distribution patterns of β4b-V5 and β4b(1–481)-V5 were observed (Fig. 4B).

Figure 4. Nuclear targeting of full-length β4b and truncated β4b(1–481) subunits in tsA-201 cells. tsA-201 cells were transfected with β1a-V5, β4b-V5 and β4b(1–481)-V5 alone, or in combination with GFP-CaV1.2/α2δ-1 and immunolabeled with anti-GFP and anti-V5 or anti-β. (A) Representative immunofluorescence images of β subunits expressed alone; and (B), of β subunits expressed together with GFP-CaV1.2/α2δ-1. C: Fraction of cells showing nuclear targeting when β subunits were expressed alone (n = 3; anti-V5 n = 180, anti-β n = 90); and D, when β subunits were expressed together with CaV1.2/α2δ-1 (n = 3; anti-V5 n = 180, anti-β n = 90). Note that the β subunits are cytoplasmic in the absence of an α1 subunit, but co-aggregate in the plasma membrane in the presence of CaV1.2. Nuclear staining is reduced in the presence of CaV1.2/α2δ-1, but still equally abundant with both β4b constructs. Scale bars, 10µm.

Semiquantitative analysis confirmed that a similar fraction of transfected tsA-201 cells showed nuclear targeting of β4b-V5 and β4b(1–481)-V5 (Fig. 4C and D). Whereas no cells were detected where nuclear staining of β1a-V5 was above cytoplasmic staining levels, full-length and truncated β4b-V5 subunits were concentrated in almost all nuclei of tsA-201 when expressed alone. When co-transfected with CaV1.2 and α2δ-1 the fraction of tsA-201 cells with nuclear targeting was reduced to approximately 60% of tsA-201 cells, but again there was no difference between cells transfected with β4b-V5 or β4b(1–481)-V5. In both conditions the prevalence of nuclear staining appeared somewhat lower when labeled with the β4 antibody (Table 1), most likely due to reduced sensitivity of the β antibodies compared with anti-V5. In total this analysis demonstrates that β4b subunits are specifically targeted into the nuclei; that this nuclear targeting was independent of co-expression of CaV1.2; and that truncation of the C-terminus did not reduce the targeting efficiency of β4b(1–481)-V5.

Discussion

The premature-termination mutation R482X of the CACNB4 gene gives rise to a calcium channel β4 protein lacking the 39 C-terminal amino acids. In humans this mutation has been linked with juvenile myoclonic epilepsy. When coexpressed with CaV2.1 in Xenopus oocytes the truncated β4b subunit resulted in calcium currents with slightly increased current amplitudes and an accelerated fast time constant of inactivation. This demonstrated that the R482X mutant β4 subunit normally associated with the pore-forming calcium channel CaV2.1, facilitated its incorporation in the plasma membrane and modulated its gating properties. This interpretation is corroborated by our present findings, where we consistently observed co-clustering of the truncated β4b(1–481) subunits with CaV1.1 and CaV1.2 in skeletal muscle triads and peripheral junctions, co-aggregation of β4b(1–481)-V5 with CaV1.2 in the plasma membrane of tsA-201 cells, and clustering of β4b(1–481)-V5 throughout the axons and dendrites of hippocampal neurons. Thus, despite the truncation of the C-terminus tagged and untagged β4b(1–481) subunits can associate with L-type and non-L-type calcium channels and appears to be normally incorporated into native calcium channel complexes in skeletal muscle cells and neurons.

This raises the question as to whether the modest functional differences between calcium channels containing the wildtype or truncated β4b subunits can be responsible for the neuronal disease phenotype. In fact, lethargic mice—which carry a mutation in the Cacnb4 gene resulting in the total lack of the β4 proteins—as well as mice with loss-of-function mutations of the primary calcium channel partner of β4 in cerebellum, CaV2.1, develop similar ataxic and epileptic phenotypes.,- These similarities of phenotypes are consistent with a synaptic defect being the primary cause of the ataxia and epilepsy also in β4b mutants. However, a recent study demonstrated that ablation of CaV2.1 function specifically in cerebellar granule cell synapses did not generate ataxia and epilepsy. In comparison to the severe impairment of synaptic function in CaV2.1 knockouts the expected effects from the β4 R482X mutation would be rather mild. In particular since there may be functional compensation by other β subunit isoforms expressed in the cerebellum., Consequently, other, ideally unique properties of β4 subunits may be the primary cause of the neurological phenotype.

One such unique property of the β4 subunit is its ability to accumulate in the cell nucleus.- Because β4 interacts with nuclear proteins involved in epigenetic regulation of genes,,, altered gene regulation might play a role in the etiology of epilepsy in patients with the R482X mutation. Indeed, the ability of β4 to regulate genes in neurons depends on the nuclear targeting properties of its splice variants and heterologous expression of full-length β4b and truncated β4b(1–481) resulted in differential gene regulation in HEK293 cells. Thus, a model has been suggested, according to which β4b forms a complex with B56δ/PP2A that is translocated into the nucleus where, in combination with HP1γ, it modifies histone H3 and consequently transcriptional regulation., Most importantly for the issue addressed in the present study, Tadmouri et al. asserted that “the formation, as well as the nuclear translocation, of the β4/B56δ/PP2A complex is totally impaired by the premature R482X mutation of β4.”

Our results presented here do not support this notion. In contrast, we examined the nuclear targeting of the truncated β4b(1–481) subunit in three different cell models and found no difference of its nuclear targeting properties compared with full-length β4b. Normal nuclear targeting of β4b(1–481) was observed in differentiated excitable cells (myotubes and hippocampal neurons), in which both β4b constructs were also incorporated in native calcium signaling complexes (triads and synapses). But also when expressed in tsA-201 cells, with and without the CaV1.2 α1 subunit, the truncated β4b(1–481) construct was targeted to the nucleus. Furthermore, in myotubes nuclear targeting of β4b(1–481) was similarly observed with V5-tagged and untagged constructs. This excludes the possibility that the differences between our findings and those of Tadmouri et al. resulted from differences in the used β4b(1–481) constructs or from distinct nuclear targeting mechanisms in different cell types.

Their data indicate the importance for nuclear targeting of C-terminal residues and of intramolecular SH3/GK interactions. In contrast, the equal nuclear targeting of full-length β4b and truncated β4b(1–481) subunits observed here indicates that other variable regions of β4b determine its nuclear targeting properties. Previously we demonstrated the importance of N-terminal residues of β4b for isoform-specific nuclear targeting. This was further corroborated by our recent discovery of a new splice variant (β4e), which essentially lacks the variable N-terminus and displays no nuclear targeting. The localization of β4b subunits in the nucleus is further regulated by a CRM1-dependent nuclear export mechanism. Consistent with the existence of an activity-dependent β4b nuclear export mechanism, we observed β4b nuclear targeting in young and electrically silent, but not in differentiated hippocampal neurons., Moreover we showed that nuclear localization of β4b was lost upon KCl depolarization in myotubes and increased after blocking electrical activity with TTX in myotubes and in hippocampal neurons., In the present study we extended these observations to the truncated β4b(1–481), which in response to TTX treatment accumulated in the nuclei of differentiated hippocampal neurons as potently as full-length β4b.

Taken together, our present results clearly demonstrate that nuclear targeting and nuclear export properties of β4b and β4b(1–481) are indistinguishable, both in excitable cells and in heterologous expression systems. These findings contest a role of the R482X epilepsy mutation in perturbing nuclear targeting of β4 and they raise serious concerns about the effects of the mutation on gene regulation. Nevertheless, it is important to note that even when nuclear targeting of β4b(1–481) remained intact, this does not exclude the possibility that within the nucleus the interactions of the C-terminally truncated β4b(1–481) with B56γ/PP2A and HP1γ might be perturbed. Whereas this possibility would preclude a model according to which complex formation of β4 and B56γ is a prerequisite for nuclear translocation, it would still be consistent with many of the biochemical data of Tadmouri et al. (2012) as well as with the observation that β4b and β4b(1–481) differentially regulate genes in HEK293 cells.

Clearly the function of the β4b subunit in the nucleus is still far from being understood. Additional experiments will be necessary to resolve the conflicting findings as well as to settle the important problem as to whether the nuclear function of calcium channel β4 subunits is critically involved in the etiology of epilepsy and ataxia in patients and mouse models with mutations in the CACNB4 gene.

Materials and Methods

Expression plasmids

Cloning procedures were previously described for: GFP-CaV1.1 (NM_001101720) and GFP-CaV1.2 (X15539), pβA-β1a-V5 (M25514), and pβA-β4b-V5 (L02315), α2δ-1 (NM_001082276), pβA-β4b (L02315). To construct pβA-β4b(1–481)-V5 the pβA-β4b-V5 (L02315) was used as a template, the deletions of amino acids 482–519 was introduced by SOE-PCR. Briefly, the 3′ cDNA sequence coding for the C-terminus of β4b was PCR amplified with overlapping mutagenesis primers in separate PCR reactions using pbA-β4b-V5 (L02315) as template. Further the two separate PCR products were then used as templates for a final PCR reaction with flanking primers to connect the nucleotide sequences. This fragment was then BglII/SalI digested and cloned into the respective sites of pβA-β4b-V5 (L02315) yielding pβA-β4b(1–481)-V5. To construct pβA-β4b(1–481), pβA-β4b-V5 (L02315) was used as a template and the 3′ cDNA sequence coding for the C-terminus of β4b was PCR amplified with a modified reverse primer introducing a stop codon after residue 481. The PCR fragment was then EcoRV/XbaI digested and cloned into the respective sites of pβA-β4b-V5, yielding pβA-β4b(1–481). Note that to be consistent with published literature,, we named the truncated β4b construct β4b(1–481). However in the Cacnb4 gene (L02315) the R-to-X mutation occurs at amino acid position 481, and not at position 482, as previously described, so the truncated constructs actually end with amino acid 480.

Myotube cell culture and transfection

Myotubes of the homozygous dysgenic (mdg/mdg) cell line GLT were cultured as previously described. At the onset of myoblast fusion, GLT cell cultures were transfected with plasmids coding for the calcium channel subunits using FuGeneHD transfection reagent (Promega) according to the manufacturer's instructions. A total of 1 μg of plasmid DNA was used per 30 mm culture dish.

Hippocampal cultures

Low-density cultures of hippocampal neurons were prepared from 17 d-old embryonic BALB/c mice of either sex as described previously.- Neurons were plated on poly-L-lysine-coated glass coverslips in 60-mm culture dishes at a density of ~3500 cells/cm2. After plating, cells were allowed to attach for 3–4 h before transferring the coverslips neuron-side-down into a 60-mm culture dish with a glial feeder layer. For maintenance, the neurons and glial feeder layer were cultured in serum-free neurobasal medium (Invitrogen) supplemented with glutamax and B27 supplements (Invitrogen). Ara-C (5 μM) was added 3 d after plating and once a week 1/3 of medium was removed and replaced with fresh maintenance medium.

Transfection of hippocampal neurons

Cultured hippocampal neurons were transfected with pβA-β4b-V5 and pβA-β4b(1–481)-V5 constructs immediately after plating for 4h using Lipofectamine 2000-mediated transfection reagent (Invitrogen) as previously described a total amount of 0.05 μg DNA was used per each condition. Transfected neurons were used for experiments from DIV 1 onwards.

tsA-201 cell culture and transfection

tsA-201 cells were cultured in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 0.44 M NaHCO3, 10% fetal calf serum (Gibco, 500–064), 2 mM glutamine (Sigma, G753) penicillin (10 units/ml) and streptomycin (10 µg/ml) and maintained at 37 °C in a humidified environment with 5% CO2. Cells were grown and transiently transfected when they reached about 80% of confluency with FuGeneHD transfection reagent (Roche Diagnostics) according to the manufacturer's instructions. A total of 0.25 μg of plasmid DNA was used per 30 mm culture dish. Cells were replated 24h after transfection onto 13 mm poly-l-lysine coated coverslips and kept at 30 °C, 5% CO2 for 24h prior to fixation.

Immunocytochemistry and microscopy

Cells were immunostained as described in for myotubes and in for neurons. Briefly, cells were fixed in 4% paraformaldehye/4% sucrose in PBS (pF) at room temperature for 20 min and incubated in 5% normal goat serum in PBS containing 0.2% bovine serum albumin (BSA) and 0.2% Triton X-100 (PBS/BSA/Triton) for 30 min. Primary antibodies; mouse monoclonal anti-β1 (1:2000) and anti-β4 (1:500) (both NeuroMab, UC Davis/NIH NeuroMab Facility), mouse monoclonal anti-V5 (1:400; Invitrogen), polyclonal anti-GFP (1:10000; Molecular Probes, Eugene, OR, USA) were applied in PBS/BSA/Triton for 4 h at RT, washed in PBS and then stained with goat anti-rabbit Alexa 488 and/or goat anti-mouse Alexa 594 (1:4,000, Molecular Probes) for 1 h at RT. After staining coverslips were washed and mounted in Vectashield to avoid photo bleaching. Preparations were analyzed on an AxioImager microscope (Carl Zeiss, Inc.,) using a 63x 1.4 NA objective. 14-bit images were recorded with a cooled CCD camera (SPOT or INSIGHT; Diagnostic Instruments, Stirling Heights, MI, USA) and Metaview image processing software (Universal Imaging, Corp., West Chester, PA, USA). Figures were arranged in Adobe Photoshop CS6 (Adobe Systems Inc.,) and where necessary contrast, black level and gamma were adjusted to optimally display the labeling patterns.

Nuclear targeting analysis in myotubes and tsA-201 cells

Cultures labeled with anti-GFP and anti-V5 or anti-β were systematically screened for transfected, well differentiated myotubes or tsA-201 cells based on the GFP-CaV1 staining (green channel) and nuclear staining of the β subunits was analyzed after switching to the red filter channel. Nuclear targeting was rated positive, when the fluorescence intensity of any nuclei in the myotube was above that of the cytoplasm. The degree of nuclear targeting in dysgenic myotubes was determined by calculating the nucleus/cytoplasm ratio of the anti-V5 fluorescence intensity, after background subtraction using Metamorph software. Results are expressed as mean ± s.e. All data were organized in MS Excel and analyzed using ANOVA with Tukey post-hoc analysis in Excel with Daniel’s XL toolbox.

Nuclear targeting analysis in neurons

The degree of nuclear targeting in cultured hippocampal neurons was determined by calculating the nucleus/cytoplasm ratio of the anti-β4 fluorescence intensity, the analysis was performed by a semi-automated procedure using a custom programmed Metamorph Macro journal as described in.

Statistical analysis

Results are expressed as means ± SEM except where otherwise indicated. Data were organized and analyzed in Excel and GraphPad

Western blot

DIV 7 GLTs expressing pβA-β4b-V5 or pβA-β4b(1–481)-V5 were trypsinized, centrifuged, resuspended and lysed in RIPA buffer (50 mM TRIS-HCl, pH 8; 150 mM NaCl2; 10 mM NaF; 0.5 mM EDTA; 0.10% SDS; 10% glycerol; 1% igepal; 1x Protease Inhibitor Complete cocktail (Roche)) with a pestle and left on ice for 30 min. The lysates were then purified by centrifugation (4,000 g, 10 min, 4 °C). Protein concentrations were determined using a BCA assay (Thermo Scientific) according to manifacturer instructions. Thirty micrograms of protein were separated by SDS-PAGE (10%) at 196 V and 40 mA for 60 min and transferred to a PVDF membrane at 25 V and 100 mA for 3 h at 4 °C with a semidry-blotting system (Roth). The blot was incubated with mouse anti-V5 (1:5000; Invitrogen) or mouse anti-β4 (1:10,000; Neuromab) antibodies overnight at 4 °C and successively with HRP-conjugated secondary antibody (1:5000; Pierce) for 1 h at room temperature. The chemiluminescent signal was detected with ECL Supersignal West Pico kit (Thermo Scientific) and visualized with ImageQuant LAS 4000.