C-termini of P/Q-type Ca21channel a1A subunits
translocate to nuclei and promote
Holly B. Kordasiewicz1,2, Randall M. Thompson1,2, H. Brent Clark2,3
and Christopher M. Gomez1,2,*
1Department of Neuroscience,2Department of Neurology and3Section of Neuropathology, University of Minnesota,
420 Delaware Street SE, Minneapolis, MN 55455, USA
Received November 12, 2005; Revised and Accepted March 24, 2006
P/Q-type voltage-gated calcium channels are regulated, in part, through the cytoplasmic C-terminus of their
a1A subunit. Genetic absence or alteration of the C-terminus leads to abnormal channel function and neuro-
logical disease. Here, we show that the terminal 60–75 kDa of the endogenous a1A C-terminus is cleaved
from the full-length protein and is present in cell nuclei. Antiserum to the C-terminus (CT-2) labels both wild-
type mouse and human Purkinje cell nuclei, but not leaner mouse cerebellum. Human embryonic kidney cells
stably expressing b3 and a2d subunits and transiently transfected with full-length human a1A contain a
75 kDa CT-2 reactive peptide in their nuclear fraction. Primary granule cells transfected with C-terminally
Green fluorescent protein (GFP)-tagged a1A exhibit GFP nuclear labeling. Nuclear translocation depends
partly on the presence of three nuclear localization signals within the C-terminus. The C-terminal fragment
bears a polyglutamine tract which, when expanded (Q33) as in spinocerebellar ataxia type 6 (SCA6), is
toxic to cells. Moreover, polyglutamine-mediated toxicity is dependent on nuclear localization. Finally, in
the absence of flanking sequence, the Q33 expansion alone does not kill cells. These results suggest a
novel processing of the P/Q-type calcium channel and a potential mechanism for the pathogenesis of SCA6.
Spinocerebellar ataxia type 6 (SCA6) is a disorder of progress-
ive cerebellar dysfunction and is one of at least three domi-
nantly inherited neurological disorders caused by mutations in
the CACNA1A gene (1,2). The CACNA1A gene encodes the
a1A subunit, the transmembrane pore-forming subunit of the
P/Q-type or CaV2.1 voltage-gated calcium channel (VGCC)
(2). Whereas the other CACNA1A disorders are associated
with simple missense, truncating or splicing mutations,
SCA6, such as Huntington’s disease (HD) (3) and other
forms of SCA (4–6), is caused by abnormal expansion of a
trinucleotide CAG repeat encoding an elongated tract of gluta-
mine residues. In SCA6, the expansion is found in exon 47 of
the CACNA1A gene, which normally contains a polymorphic
CAG repeat tract (CAG)4–18encoding 4–18 glutamines in the
C-terminus of the a1A subunit, but is expanded to the patho-
logical range of (CAG)19–33, encoding 19–33 glutamines (7,8).
VGCC are multimeric complexes composed of at least three
protein subunits: a pore-forming subunit (a1) and two auxili-
ary subunits (b and a2d). Distinct genes encode more than
10 different a1 subunits (a1A–I,S), which confer different
channel properties and are expressed in different cell types
mutations in the various VGCC genes (1,2). P/Q channels
are involved in a diverse array of cell functions including neu-
rotransmitter release, regulation of gene expression, release of
calcium from internal stores and dendritic calcium transients
(9–12). P/Q channels are highly expressed in cerebellar
neurons and localize primarily to nerve terminals, dendrites
and Purkinje cell soma (11,13).
The full-length a1A subunit contains four homologous
transmembrane repeat domains (I–IV) flanked by three intra-
cellular loops (LI–II, LII–III, LIII–IV) and cytoplasmic N- and
C-termini (9). Complementary DNA sequencing and protein
immunoblot studies with domain-specific antisera have
# The Author 2006. Published by Oxford University Press. All rights reserved.
For Permissions, please email: email@example.com
*To whom correspondence should be addressed at: Department of Neurology, AMB S237, MC2030, The University of Chicago, 5841 S. Maryland,
Chicago, IL 60637, USA. Tel: þ1 7737026390; Fax: þ1 7737025670; Email: firstname.lastname@example.org
Human Molecular Genetics, 2006, Vol. 15, No. 10
Advance Access published on April 4, 2006
by guest on May 29, 2013
demonstrated the presence of a1A polypeptides of a range of
sizes and domain compositions arising from alternative spli-
cing and possible proteolytic processing (14–16). Several
studies have suggested the presence of N-terminal fragments
of the a1A subunit that lack portions of the distal C-terminus
ranging from 40 kDa to the entire last two transmembrane
repeat domains (17–19). Conversely, a 75 kDa C-terminal
fragment has been detected in protein extracts of full-length
a1A-expressing human embryonic kidney (HEK) cells
(17,20). These observations suggest that, at least in heter-
ologous systems, the a1A C-terminus is cleaved and may
form a stable a1A polypeptide.
The a1A C-terminus participates in a number of protein–
protein interactions and plays a prominent role in modulating
channel activity (21). Thus, its alteration by proteolytic clea-
vage or genetic mutation would have significant functional
consequences. Mice homozygous for the leaner (Tgla)
mutation, which express a1A subunits lacking the distal
C-terminus, are severely ataxic, exhibit Purkinje cell degener-
ation and die if unattended at ?21 days (22).
In this study, we explored, in the cerebellum, the distri-
bution of putative endogenous C-terminal polypeptides using
an antibody we raised against an a1A C-terminal epitope.
We found that the C-terminal a1A antiserum labels Purkinje
cell nuclei and that a 60–75 kDa proteolytic fragment of the
endogenous a1A C-terminus is present in cell nuclei. Using
plasmid vectors expressing recombinant human a1A protein,
we identified three nuclear localization signals (NLSs).
Finally, because the cleaved C-terminal fragment bears a poly-
glutamine tract, which is expanded in SCA6, we tested
whether expansion of the polyglutamine tract correlated with
cell death. Cells expressing C-terminal proteins bearing
polyglutamine tracts of 33 glutamines had more than twice
the rate of cell death than those expressing unexpanded
normal tracts. Furthermore, this polyglutamine-mediated cell
death appears to be dependent on nuclear localization of the
C-terminal fragment. These results suggest that the a1A
C-terminal cleavage product may play a role in nuclear signal-
ing and in the pathogenesis of SCA6.
The C-terminus of the a1A subunit is present
in cell nuclei
To determine the subcellular distribution of the a1A subunit in
human cerebellar cortex, we used affinity-purified, anti-
peptide antibodies, specific for either the N-terminus (NT-1
antibody) or the C-terminus (CT-2 antibody) of the a1A
subunit. Surprisingly, in paraffin-embedded human cerebellar
tissue, the C-terminal antibody intensely stained Purkinje
cell nuclei (Fig. 1A, inset). CT-2 nuclear staining was not
observed in granule cells or any other cerebellar cell type.
NT-1 and CT-2 antibody also labeled the Purkinje cell mem-
brane, projections and cytoplasm as previously reported for
other a1A antibodies (13,23,24) (Fig. 1A and B). NT-1 anti-
bodies did not label cell nuclei (Fig. 1B), although they did
intensely stain Purkinje cell membranes. A punctate staining
pattern was observed at the cell membrane with both NT-1
and CT-2 antibodies (Fig. 1A and B, inset). Immunostaining
by both NT-1 and CT-2 was specifically blocked by
preincubation with their respective peptide immunogens
(Fig. 1C and D), but not with other peptides (data not
shown). An identical pattern of immunolabeling was seen in
adult mouse cerebellar sections (Fig. 1E and F).
As a further test of specificity, we compared immunostain-
ing of cerebellar sections from homozygous leaner (Tgla)
pups, which express mutant a1A subunits lacking a
C-terminus and thus the CT-2 epitope, with age-matched post-
natal day 18 (P18) control pups. CT-2 antibody failed to label
Purkinje cells from leaner cerebellum, either in the cytoplasm
or in the nucleus (Fig. 1G). NT-1 antibody stained leaner cer-
ebellum in the same pattern as control cerebellum, although
slightly less intensely (Fig. 1H and J). In control P18 cerebel-
lum, CT-2 antibody labeled nuclei more intensely than wild-
type (WT) adult nuclei, and both NT-1 and CT-2 antibody
labeled WT P18 Purkinje cell soma (Fig. 1I and J, inset).
Cell membranes and projections were stained relatively less
Figure 1. The a1A C-terminus is present in Purkinje cell nuclei in mouse and
human cerebellum. Cerebellar sections stained using immunoperoxidase
(brown) with either anti-CT-2 (A, C, E, G and I) or anti-NT-1 (B, D, F, H
and J).Bluenuclearcounterstainis hematoxylin.Paraffin-embeddedcerebellar
sections were from adult human (A and B); adult human blocked by pre-
incubation with the respective peptide (C and D), WT adult mouse (E and
F), homozygous leaner mouse (G and H) and P18 WT mouse (I and J).
Insets are enlargements of Purkinje cell staining. White arrows indicate Pur-
kinje cell nuclei. Black arrows indicate Purkinje cell projections and cell
soma. Scale bars: 75 mm.
1588Human Molecular Genetics, 2006, Vol. 15, No. 10
by guest on May 29, 2013
intensely in control P18 cerebellum than in adults. These find-
ings indicate that the C-terminus of the a1A subunit is loca-
lized to the nucleus as well as to the cell membrane where
the full-length channel resides.
To investigate whether the nuclear labeling of the a1A
C-terminal antibody can be attributed to expression and pro-
cessing of the full-length a1A subunit, we transiently trans-
fected HEK cells, stably expressing b3 and a2d (SHEKb3),
with full-length human a1A cDNA (Fig. 2A). We hypoth-
esized that the CT-2 nuclear labeling resulted from the
a1A-transfected HEK cells (20). HEK cells, although they
do not express endogenous VGCCs, are able to form func-
tional channels when a1, b and a2d are ectopically
co-expressed. Auxiliary subunits (b and a2d) are required
for proper trafficking of the a1 subunit (21); in the absence
of the b subunit a1 is retained in the endoplasmic reticulum,
thus HEK cells stably expressing b3 and a2d were employed.
Eighteen to twenty hours after transfection, CT-2 antibody
stained the nucleus and cytoplasm, whereas NT-1 antibody
solely labeled SHEKb3 cell cytoplasm (Fig. 3A).
To exclude artifacts due to antibody cross-reactivity, we
generated fusion constructs that were tagged on either the
N-terminus (GFPNH3a1A) or the C-terminus (GFPCOOHa1A)
of the full-length protein with enhanced green fluorescent
protein (EGFP) (Fig. 2A). When utilized in conjunction with
our N- and C-terminal antibodies, the tagged a1A subunits
allow for simultaneous localization of both ends of the
protein in a single cell. In individual cells expressing N-
terminally GFP-tagged a1A (GFPNH3a1A) and stained with
C-terminal antibody (CT-2), signals for the two termini were
spatially separated (Fig. 3B). The N-terminal GFP localized
solely to the cytoplasm, indicating that the N-terminal
portion of the protein resided outside the nucleus, as expected.
In contrast, in these same cells, the C-terminal antibody
stained both the cytoplasm and the nucleus, suggesting that
some of the C-terminal a1A polypeptides localize to the
cytoplasm with the N-terminus of the protein and a second
population of C-terminal a1A polypeptide localizes to the
nucleus. The converse experiment, with GFPCOOHa1A and
NT-1 antibody exhibited similar results (Fig. 3B). Quantifi-
cation of immunostain data showed that a significantly
greater percentage of a1A expressing cells exhibited nuclear
staining when stained with CT-2 antibody or transfected
with C-terminally GFP-tagged a1A than with N-terminal
markers (Fig. 3C). Similar results were obtained in the
mouse fibroblast cell line, NIH3T3 (Fig. 3D), in which
GFP-tagged a1A was co-expressed with b3 and a2d. This
observation suggests that a portion of the C-terminus of the
a1A subunit is cleaved and translocated to nuclei.
To determine whether a1A cleavage also occurs in neuronal
cells, primary cerebellar granule cells were transfected with
GFPNH3a1A or GFPCOOHa1A. In neurons transfected with
GFPNH3a1A, GFP was localized to the cytoplasm, whereas
in neurons transfected with GFPCOOHa1A, the GFP signal
was present in both the nucleus and the cytoplasm (Fig. 3E).
Finally, we generated a GFP fusion construct that expresses
only the a1A C-terminus downstream of the predicted clea-
vage site, this comprises residues 2096–2510 (GFP2096)
(Fig. 2A and B). The N-terminus of this a1A protein fragment
was tagged with GFP. This C-terminal protein also localized
to nuclei in transfected granule cells (Fig. 3E). Because of
cross-reactivity, a1A specific antibodies were unable to dis-
tinguish recombinant a1A subunits from endogenous a1A in
transfected granule cells. Nevertheless, these cell culture
data suggest that the CT-2 nuclear staining observed in
Purkinje cells arises from cleavage of the a1A C-terminus
and its translocation to the nucleus.
The a1A C-terminus is cleaved from the
Using antibody to the a1A C-terminus (CT-2), we identified
three polypeptides of molecular weights 220, 170 and
60 kDain western blots of mouse cerebellar lysates
(Fig. 4A). The 220 and 170 kDa proteins correspond to pre-
viously reported full-length a1A subunit species (16,17). To
test whether the 60 kDa polypeptide represents a C-terminal
fragment of the a1A subunit, we utilized the spontaneous
mouse mutant, leaner. Leaner mice have a splice site mutation
in the CACNA1A gene that results in aberrant splicing in the
region of the a1A C-terminus. This yields two predominant
a1A splice forms that lack the C-terminus after residue
1967 including the CT-2 epitope (22). The CT-2 reactive
220, 170 and 60 kDa polypeptides are undetectable in cerebel-
lar lysates of homozygous leaner mice (Tgla), compared with
age-matched (P18) control pups under these experimental con-
ditions (Fig. 4A), whereas the signal detected by antisera to
other a1A polypeptide splice forms and antisera to other
Figure 2. Schematic diagram of a1A constructs used in this study. (A) Full-
length a1A and various truncated forms. Amino acid numbers are displayed
on the top. Constructs denoted Qn contain either Q4, Q11 or Q33 glutamine
repeats. If the glutamine repeat size is not indicated, then Q11 was used. X
indicates approximate location of NLS mutations. Constructs tagged with
EGFP and/or an artificial NES are indicated. (B) A schematic representation
of the C-terminal construct and approximate location of motifs of interest.
Included are NLS1,2,3 and 4, the evolutionarily conserved histidine tract
(HIS) and the CT-2 antibody recognition sequence (CT-2 Epitope).
Human Molecular Genetics, 2006, Vol. 15, No. 101589
by guest on May 29, 2013
synaptic components, such as SNAP 25, synaptotagamin and
synaptophysin, was identical in leaner and age-matched
controls (data not shown).
The C-terminus of the a1A subunit contains an evolutiona-
rily conserved tract of 11 histidine residues. To verify the
identity of the 60 kDa polypeptide as the C-terminus of a1A,
we immunoprecipitated proteins from mouse cerebellar
homogenates with CT-2 antibody, separated them using SDS–
PAGE gel electrophoresis and reacted them with an antibody
to penta-HIS (anti-penta-HIS, Sigma, St Louis, MO, USA).
This reagent recognizes proteins possessing five or more
consecutive histidines. Anti-penta-HIS antibody also recog-
nized a 60 kDa species in the CT-2-precipitated lanes but not
in control lanes (Fig. 4B). The 55 kDa polypeptide present in
the experimental condition and the control without lysate
(SPA and CT-2), corresponds to IgG heavy chain.
We next tested for the presence and size of a corresponding
human a1A fragment in SHEKb3 cells transiently transfected
with the cDNA encoding the full-length human a1A subunit.
We detected a 75 kDa polypeptide, similar to that observed
in mouse brain homogenates using CT-2 antibody, in lysates
of a1A transfected SHEKb3 cells (Fig. 4C). Sequence align-
ment of the mouse a1A C-terminus containing the CT-2
epitope (MPII splice form) (25) and the known human a1A
C-terminus predicts that the human form is 10–15 kDa
larger than the mouse C-terminus.
To further confirm that the 75 kDa polypeptide corresponds
to the a1A C-terminus, we performed western blot analysis of
Figure 3. The C-terminal cleavage product localizes to cell nuclei in a1A transiently transfected stable b3 and a2d HEK293 (SHEKb3) cells, NIH3T3 cells and
transfected cerebellar granule neurons. Confocal micrographs of cells transfected with human a1A cDNA (a1A, GFPNH3a1A or GFPCOOHa1A, as indicated).
Cells were stained with CT-2 or NT-1 antibodies and the nucleic acid dye ToTo-3 (blue). Transmitted light images are provided. (A) Immunocytochemistry of
SHEKb3 cells expressing a1A and stained with CT-2 or NT-1 antibody (green) and overlaid with ToTo-3 (blue). (B) Dual labeling of SHEKb3 cells with anti-
body (red) and EGFP (green). Cells transfected with GFPNH3a1A were stained with CT-2 antibody and cells transfected with GFPCOOHa1A were stained with
NT-1 (constructs and antibodies are indicated in individual images). (C) The proportion (%) of SHEKb3 cells exhibiting nuclear labeling with N-terminal EGFP
(GFPNH3a1A), C-terminal EGFP (GFPCOOHa1A), NT-1 antibody or CT-2 antibody (n ¼ 147, 114, 73 or 119, respectively). A significantly greater percentage of
nuclei were labeled with C-terminal markers (CT-2 and C-terminal EGFP) than with N-terminal markers (NT-1 and N-terminal EGFP). EGFP (z ¼ 12.21,
P , 0.01) and antibody (z ¼ 9.86, P , 0.01). (D) Dual labeling of NIH3T3 cells co-transfected with b3, a2d and a1A (GFPNH3a1A or GFPCOOHa1A) and
stained with a1A antibodies (NT-1 or CT-2). (E) Primary cerebellar cultures expressing N-terminally GFP-tagged a1A (GFPNH3a1A), C-terminally GFP-tagged
a1A (GFPCOOHa1A) or GFP-tagged a1A C-terminus (GFP2096). Scale bars: 10 mm.
1590 Human Molecular Genetics, 2006, Vol. 15, No. 10
by guest on May 29, 2013
lysates of cells transfected with full-length a1A C-terminally
labeled with GFP (GFPCOOHa1A). In these cell lysates, a
novel 100 kDa protein was recognized by CT-2 antiserum but
the 75 kDa protein was absent (Fig. 4D). The 100 kDa band
was also reactive with GFP antibody, consistent with the pre-
sence of a 30 kDa GFP protein on the C-terminus of a1A
extending the polypeptide (Fig. 4D). Thus, the C-termini of
a1A subunits from two different species are cleaved in vivo
and when expressed as recombinant proteins in cell culture.
Finally, the identity of the 75 kDa protein as the a1A
C-terminus was further confirmed using a second antibody to
C-termini, CT-1 (24). Although the signal intensity was
weaker, CT-1 antibody also recognized the 75 kDa protein
(data not shown). These data confirm that there are two distinct
populations of CT-2 reactive a1A proteins, the full-length a1A
that contains the C-terminus and an a1A C-terminal fragment.
The a1A C-terminal fragment is present in
To determine whether C-terminal labeling of nuclei (Fig. 1)
corresponded to the 60 kD fragment in mouse homogenates
(Fig. 4), we fractionated dissociated cerebellar neurons into
nuclear and cytoplasmic compartments prior to electrophoresis
and blotting with CT-2 antibody. As expected, the mouse
60 kDa polypeptide was enriched in nuclear fractions,
whereas full-length polypeptides were nearly absent from
nuclear fractions (Fig. 5A). In SHEKb3 cells transfected with
the human a1A vector, the 75 kDa CT-2 positive peptide was
also enriched in nuclear fractions (Fig. 5B). Controls for
adequate fractionation confirmed that there was minimal
contamination between the two fractions (Fig. 5A and B).
Quantification of the endogenous mouse cerebellar 60 kDa
polypeptide revealed that 84+1.2% of the cleaved C-termini
were present in the nuclear fraction (Fig. 5C). The C-terminal
fragment made up ~19% of the total CT-2 positive proteins
(Fig. 5D). These data confirm that the 75 kDa C-terminal poly-
peptide is cleaved from the full-length protein and is present in
cell nuclei. These findings can also account for the diffuse
a1A-transfected cells (Fig. 3). Presumably, the cytoplasmic
staining observed was the intact full-length protein, found
here in the cytoplasmic fraction, whereas the nuclear staining
corresponded to the cleaved C-terminus, shown here to be
predominant in the nuclear fraction.
The a1A C-terminus contains NLSs
Analysis of the C-terminal protein sequence, using an online
software package (www.psort.org), revealed four putative
NLSs (Fig. 6A). To isolate the translocation element, we
designed constructs that express only the a1A C-terminus
downstream of the cleavage site as a fusion protein with an
N-terminal GFP tag (GFP2096) (Fig. 2A and B). In SHEKb3
cells expressing GFP2096, 83.7% +1.3 of the signal was
present in the nucleus and only 17.3% +1.3 in the cytoplasm
(Fig. 6B). As shown in Figure 6, replacement of a lysine with
an alanine in the first putative NLS of GFP2096 (DNLS1)
caused a significant decrease in nuclear staining (64.3%
+2.8, P , 0.01). Further mutation of the two arginines in
NLS1 resulted in similar staining (data not shown). Mutation
of NLS2 (DNLS2) and NLS3 (DNLS3) significantly decreased
nuclear staining to 67.0% +2.9, (P , 0.01) and 72.7% +2.8
(P , 0.01), respectively (Fig. 6). Mutation of NLS4 did not
significantly alter localization (83.9% +1.0), suggesting that
the alterations observed following mutation of NLS1, 2 or 3
were not simply due to replacement of basic residues, but
rather due to alteration of specific targeting sequences.
Finally, we mutated all three NLSs in the same GFP2096 con-
struct (DNLS1,2,3) (Fig. 2A). The triple NLS mutation signifi-
cantly decreased the nuclear staining (58.0% +2.9) compared
with the individual NLS mutants and the native GFP2096
(P , 0.01, Fig. 6). The triple NLS mutant was not significantly
different than the diffuse EGFP (63.9% +1.6) staining pattern.
Figure 4. A C-terminal fragment of a1A is present in mouse cerebellum and
in a1A-transfected SHEKb3 cells. (A) Cerebellar lysates from WT adult mice,
P18 WT mice and homozygous leaner (Tgla2/ 2) mice were analyzed by
western blot with anti-CT-2 antibody. Arrows indicate 220, 170 and 60 kDa
CT-2-positive a1A proteins. Top panel of (A) is a longer exposure of CT-2
reactive 220 and 170 peptides. Anti-GAPDH was used as a loading control.
(B) Immunoprecipitation (IP) of mouse cerebellar lysate with CT-2 antibody
and staphylococcus protein A coated beads (SPA). Precipitated proteins
were analyzed by western blot with anti-penta-HIS antibody (HIS). Arrow
indicates 60 kDa band not present in negative controls, IP without antibody
(WT) and IP without lysate (SPA). (C and D) SHEKb3 untransfected (UT),
transfected with GFPNH3a1A or GFPCOOHa1A were analyzed by western
blot with anti-CT-2 or anti-GFP. GFPNH3a1A cell lysates contained a
75 kDa CT-2-positive protein that was shifted ?30 kDa in GFPCOOHa1A
cell lysates. (E) Immunoblot comparison of stable b3 and a2d HEK cells
(SHEKb3) and untransfected HEK 293 cells with anti-b3 and a loading
control (anti-GAPDH). Molecular weights are indicated at the right of each
panel in kDa. All experiments were replicated with similar results.
Human Molecular Genetics, 2006, Vol. 15, No. 101591
by guest on May 29, 2013
This suggests that the three NLSs (NLS1–3) are necessary to
maintain the selective nuclear localization of the a1A C-
Numerous proteins contain multiple targeting signals
(26–28). In cell culture studies of several proteins containing
multiple nucleocytoplasmic transport signals, the stronger
targeting signal determines the localization of the protein
(29,30). The mutations in NLS1–3 were unable to completely
abolish nuclear localization, suggesting the presence of other
nuclear targeting motifs (29,30). To further assess the nuclear
targeting ability of the putative NLSs (NLS1–3), we attached
two well-characterized nuclear export signals (NESs) to the C-
terminus of GFP2096 (Fig. 2A). We found that the NES was
unabletosignificantly alter nuclear localization ofthe native C-
terminus (79.3% +1.7, Fig. 7A). However, when the NES was
attached to C-termini containing the three NLS mutations
(DNLS1,2,3) nuclear localization was abolished (16.5%
+1.2, Fig. 7A). Thus, although the ‘artificial’ NES was
unable to remove the C-terminus with intact NLSs from the
nucleus, NES-directed export did occur following NLS elimin-
ation. Similar results were obtained in the fibroblast cell line
NIH3T3 (GFP2096, 85.6% +1.5; NES, 75.36% +2.1;
DNLS/NES, 28.8% +2.8) (Fig. 7B). This is consistent with
recent reports of ataxin-7 containing three NLSs and a native
NES. The NES in ataxin-7 was only able to abolish nuclear
localization when two of the NLSs were removed (31).
The cleaved human a1A C-terminus contains the
Purkinje cells preferentially express a splice form of the
human a1A subunit containing an extended C-terminus
encoded by exon 47, and a small tract of glutamine residues
(polyglutamine), normally polymorphic in length from four
to 18 glutamines (8,23). Expansion of this polyglutamine
tract to the range of 19–33 glutamine residues is associated
disease, SCA6 (7,8). On the basis of the size of the proteolytic
fragment (?400 amino acids), we predicted the region of clea-
vage and hypothesized that the C-terminal cleavage product
contains the polyglutamine tract.
In western blots of cells expressing a1A subunits, contain-
ing either Q11 (a1AQ11) or Q33 (a1AQ33) repeats, the size
of the proteolytic fragment correlated with the predicted size
of the polyglutamine tract (Fig. 8A). Furthermore, the CT-2
antibody reacted in western blots with the C-terminal cleavage
product generated from full-length a1AQ11 and a1AQ33 with
relatively equal intensity (density relative to GAPDH: Q11,
0.25 + 0.06; Q33, 0.28 + 0.05) (Fig. 8B). Finally, polygluta-
mine tracts do not alter the ratio of full-length to cleaved
C-terminal a1A (Q11, 0.57 + 0.17; Q33, 0.62 + 0.10;
Fig. 8B). Similarly, expression of a1A C-terminal constructs
(GFP2096) containing either Q4, Q11 or Q33 repeats were
not significantly different from each other (density relative
to GAPDH: Q4, 2.35 + 0.21; Q11, 2.41 + 0.27; Q33,
2.78 + 0.32) (Fig. 8C and D). These findings suggest that
there is no significant difference in expression between the
unexpanded (Q4 or Q11) and expanded (Q33) cleavage
products in these experimental paradigms. This contradicts
previous findings that suggest C-terminal proteolytic frag-
ments bearing expanded polyglutamine tracts are more
stable than those with normal polyglutamine tracts (20).
Polyglutamine expansions in a1A C-termini
Previous studies have suggested that the polyglutamine-
expanded a1A in SCA6 may lead to cell death by altering
calcium channel kinetics (23,32–34). Because C-termini con-
taining normal or expanded polyglutamine tracts are cleaved
from the full-length protein, we hypothesized that SCA6
C-termini themselves are toxic, independent of channel
function. We transfected SHEKb3 cells with GFP2096 con-
taining Q4, (GFP2096Q4), Q11 (GFP2096Q11) or Q33
(GFP2096Q33) and assessed cell viability with propidium
iodide (PI) exclusion analyzed by fluorescence-activated cell
Figure 5. C-terminal cleavage product is enriched in the nuclear fraction of
mouse cerebellar lysates and a1A-transfected SHEKb3 cells. Nuclear (Nu)
and cytoplasmic (Cy) fractions of (A) mouse cerebella and (B) SHEKb3
cells transfected with GFPNH3a1A were subjected to western blotting and
probed with anti-CT-2. Anti-GAPDH (cytoplasmic marker) and anti-Histone
3 (nuclear marker) were used as fractionation controls. Molecular weights
are indicated at the right of each panel in kDa. (C) Densitometer quantification
of western blots of mouse cerebellar fractions probed with CT-2. Nuclear
localization of the a1A C-terminus (60 kDa fragment) and the full-length
a1A (full-length) is expressed as a proportion (%) of the total protein
present in the nuclear fraction. (D) Quantification of the total CT-2 identified
full-length peptides and the total 60 kDa fragment is expressed as total protein
area (OD?mm). Asterisk indicates
(P , 0.01, two-tailed, unpaired Student’s t-test, n ¼ 6 cerebella). Error bars
represent mean SE.
1592Human Molecular Genetics, 2006, Vol. 15, No. 10
by guest on May 29, 2013
sorting (FACS). In this assay, cell death is expressed as a
proportion of dead transfected cells above baseline, where
baseline is the proportion of dead untransfected cells.
Seventy-two hours post-transfection, cells transfected with
GFP2096Q33 exhibited a 2-fold greater rate of cell death
GFP2096Q11 (8.4% +1.6, P , 0.01) or GFP2096Q4 (3.0%
+1.1, P , 0.01) and significantly greater than untransfected
cells (P , 0.01) (Fig. 9A).
It has been reported that expanded glutamine tracts (Q80)
alone may be toxic to cells independent of protein context
(35). To test for a direct effect of 33 glutamine residues on
cell viability, we deleted the remainder of the a1A
C-terminus, except the glutamine tract and the flanking 30
amino acids that include NLS3 (GFPQ33), and compared
the toxicity of this polypeptide. SHEKb3 cells transfected
with GFPQ33 did not exhibit any more cell death than
untransfected cells (0.8% +1.9, Fig. 9A), consistent with
previous studies using Q35 tracts (35).
In other polyglutamine diseases, including dentatorubral-
pallidoluysian atrophy and SCA1, cell death is dependent on
those transfected with
nuclear localization of the polyglutamine containing protein
(36,37). We hypothesized that, like other polyglutamine
diseases, the toxicity of expanded (Q33) C-termini is depen-
dent on nuclear localization of the C-terminal fragment. To
test this hypothesis, we utilized the constructs DNLS1,2,3
and DNLS/NES containing Q4, Q11 or Q33 glutamines.
DNLS1,2,3 is ~50% nuclear, whereas DNLS/NES is almost
entirely excluded from the nucleus (20% nuclear) and the
native C-terminal protein (GFP2096) is nearly exclusively
localized to the nucleus (90% nuclear, Fig. 7). SHEKb3
exhibit a 37% decrease in cell death (11.7% +1.4) compared
with unmodified C-termini containing Q33 (GFP2096Q33,
P , 0.01).Cells
DNLS1,2,3, containing Q4 (4.1% +1.7) or Q11 (8.7%
+0.6), were not significantly different than their respective
unmodified counterparts (GFP2096Q4, 3.0% +1.1; Q11,
8.4% +1.6). Similarly, cell death displayed in cells expressing
DNLS/NES, containing Q4, Q11 or Q33, was not significantly
different from one another (DNLS/NESQ4, 3.6% +0.4; Q11,
2.8% +0.7; Q33, 4.2% +0.8), although DNLS/NESQ11 and
Figure 6. Mutations of three putative NLSs decrease nuclear compartmentalization of the C-terminus. (A) Diagram of the NLS sequences mutated. Bold under-
lined residues were mutated to alanine residues. Numbers correspond to location in amino acid sequence. (B) Proportion (%) of nuclear relative fluorescence in
SHEKb3 cells transfected with GFP2096 NLS mutants. Asterisk indicates statistically significant differences compared with GFP2096 (P , 0.01, two-tailed,
unpaired Student’s t-test, n ¼ 25 for each condition). Error bars represent mean SE. Statistically significant results were replicated on three separate occasions.
(C) Representative confocal images from (B) transmitted image and ToTo-3 overlay is provided. A single plane was taken through the center of each nucleus.
Scale bar: 10 mm.
Human Molecular Genetics, 2006, Vol. 15, No. 10 1593
by guest on May 29, 2013
DNLS/NESQ33 were significantly less toxic than their
unmodified counterparts, GFP2096Q11 and GFP2096Q33,
respectively(P , 0.05).These
polyglutamine-mediated cell death is dependent on localiz-
ation of the C-terminal fragment to the nucleus.
The results obtained with unmodified C-termini (GFP2096)
were replicated in transfected cerebellar granule cells
(Fig. 9B). Granule cells expressing GFP2096Q33 exhibited
significantly more death (10.4% +1.7) than unexpanded
GFP2096Q4 (5.9% +4.2, P , 0.05) or GFP2096Q11 (6.0%
+0.2, P , 0.01). Interestingly, in SHEKb3 cells, cell death
was significantly different
GFP2096Q11 and GFP2096Q4. This difference was not
observed, however, in granule cells. Similar to SHEKb3
cells, death in granule cells expressing the expanded glutamine
tract alone (GFPQ33), lacking the C-terminal sequences, was
not significantly different than baseline (1.3% +1.6).
SHEKb3 cells and granule cells transfected with any of the
GFP2096 constructs exhibited nearly 90% nuclear staining
(Fig. 9D). GFP2096Q33, GFP2096Q11 and GFP2096Q4 all
exhibited a speckled distribution pattern in both SHEKb3
and granule cells (Fig. 9D). Quantification of images
between cells expressing
confirmed that GFP2096Q4, GFP2096Q11 and GFP2096Q33
populations did not have significantly different average
139.5 + 6.1;GFP2096Q11,
135.3 + 7.2). GFPQ33 did have a significantly higher
(P , 0.05) ARFI than GFP2096Q33, Q11 or Q4 (GFPQ33,
170.33 + 14.7).
These data suggest that the C-terminus of the a1A subunit
is toxic to cells when harboring an expanded polyglutamine
tract, and as suggested in other polyglutamine disorders (4),
this toxic effect may be dependent on the protein context of
the surrounding polypeptide and nuclear localization of the
GFP2096Q33, 135.1 + 8.3;
These studies demonstrate, for the first time, that the
C-terminus of the a1A subunit of the P/Q-type VGCC is
cleaved from the full-length protein and translocated to the
nucleus. This post-translational modification is robust and
occurs in Purkinje cells of multiple species and cultured
SHEKb3 cells, NIH3T3 cells or neurons transfected with
recombinant human a1A. The C-terminal cleavage product
contains three NLSs that promote its localization to the
nucleus. Finally, the polyglutamine tract in these C-termini
is polymorphic in length and expanded forms associated
with SCA6, when localized to nuclei, are toxic to cultured
cells and neurons. These findings may indicate a novel role
for the P/Q-type calcium channel and suggest a possible
mechanism in the pathogenesis of SCA6.
Figure 7. Artificial NES is unable to alter nuclear localization of the a1A C-
terminus with native NLSs intact. (A) Proportion (%) of nuclear fluorescence
intensity in SHEKb3 cells transfected with GFP2096, GFP2096 tagged with
an artificial NES (NES) or GFP2096 with triple NLS mutations and an NES
(DNLS/NES). (B) Proportion (%) of nuclear fluorescence intensity in
NIH3T3 cells transfected with GFP2096, NES and DNLS/NES. Asterisk indi-
cates statistically significant differences compared with GFP2096 (P , 0.01,
two-tailed, unpaired Student’s t-test, n ¼ 25 for each condition). Error bars
represent mean SE. (C) Representative confocal images from (A), transmitted
image and ToTo-3 overlay are provided. Statistically significant results were
replicated on three separate occasions.
Figure 8. The a1A C-terminus contains the polyglutamine tract. (A) The a1A
C-terminal cleavage product contains the polyglutamine tract and shifts
?5 kDa in SHEKb3 cells transfected with a1AQ33 compared with
a1AQ11, as analyzed by CT-2 probed western blot. (B) Left graph represents
quantification of 75 kDa band with densitometer. Intensity is expressed as the
area (mm?OD) of 75 kDa divided by the area of GAPDH (n ¼ 6). Right graph
represents the ratio of full-length (240 kDa) peptide to cleaved (75 kDa)
peptide. (C) GFP-tagged a1A C-termini shift relative to polyglutamine expan-
sion length. (D) Quantification of tail constructs relative to GAPDH (n ¼ 6).
Error bars represent mean SE. None of the conditions tested were significantly
1594Human Molecular Genetics, 2006, Vol. 15, No. 10
by guest on May 29, 2013
Previous studies have sought to determine the effect of the
expanded polyglutamine tract in the C-terminus on P/Q
channel kinetics. Expression studies in cultured cells or
oocytes have demonstrated that the SCA6 polyglutamine expan-
sion alters the P/Q channel gating and currents (24,32,34),
although the exact mutant channel phenotype appears to be
highly dependent on splice form and experimental conditions.
Moreover, expression of full-length channels bearing expanded
polyglutamine tracts has yet to be linked to cell death. Given
the present findings that cleavage of the polyglutamine-
containing C-terminus occurs in the expression systems used
to measure channel function, this post-translational modification
may impact the interpretation of these electrophysiological
studies. Specifically, the ratio of cleaved to uncleaved a1A
subunits in previous P/Q channel recordings remains to be
Until now, the a1A subunit has been considered a solely
cytoplasmic protein (7,8,38). Expression studies at the level
of both RNA and protein suggest that the a1A C-terminus
splice form bearing the polyglutamine tract is preferentially
expressed in cerebellar Purkinje cells (13,23). In this study,
shown using CT-2 antibody, a significant proportion of the
cleaved C-terminus is present in Purkinje cell nuclei. Simi-
larly, nuclear accumulation appears to be required for cell
death. This provides a potential explanation for the selective
degeneration of Purkinje cells in SCA6, regardless of
In other polyglutamine diseases, aside from SCA2, disease
pathogenesis requires entry of the mutant protein or a proteo-
lytic fragment containing the expanded polyglutamine tract
into the nucleus (3,4,6). Mice expressing expanded ataxin-1
do not develop disease when the NLS in ataxin-1 is mutated
(37). Similarly, Drosophila expressing spinal and bulbar
muscular atrophy causing mutations in the androgen receptor
do not exhibit toxicity when the mutated receptor is tagged
with an NES (39). Finally, expanded ataxin-7 is unable to
kill cultured granule cells when its nuclear targeting signals
are compromised (31). Our results are consistent with these
mechanisms of pathogenesis. The cellular toxicity of a
C-terminal a1A polypeptide bearing the expanded polygluta-
mine tract of SCA6 and requiring nuclear localization suggests
that the mechanism of pathogenesis of SCA6 may overlap, in
part, with that of other polyglutamine diseases.
In our system, toxicity is not only dependent on nuclear
localization, but also appears to occur in a dose-dependent
manner. Expanded C-termini with only 50% nuclear localiz-
ation (DNLSQ33) exhibit significantly attenuated levels of
toxicitycompared with unmodified
C-termini (GFP2096), yet toxicity is still significantly elevated
when compared with non-expanded controls (Fig. 9). This
suggests that the amount of expanded C-terminus in the
nucleus affects the level of toxicity. This is also a plausible
explanation for why cell death has not been observed in
unstressed cells expressing full-length expanded a1A under
the time frame of these experiments, but rather requires
decades of expression in most SCA6 patients. Only cleaved
C-terminal fragments translocate to nuclei, the unprocessed
full-length protein prevents nuclear entry of the polyglutamine-
containing C-terminus by retaining it in the cytoplasm, thus
reducing the amount of C-terminus present in nuclei and
reducing toxicity. This prevention of nuclear entry may account
for the delayed time course of toxicity observed in patients and
our inability to observe cell death with full-length a1A.
Although nuclear translocation may be a common feature
for SCA6 and most other polyglutamine diseases, SCA6
differs in several ways from other members of this disease
family (40). First, proteolytic cleavage appears to be a
Figure 9. Cell death in expanded a1A C-termini-expressing cells. (A) Cell
death as determined by FACS analyzed PI exclusion of SHEKb3 cells trans-
fected with GFP-tagged a1A C-termini containing glutamine tracts of 33
(GFP2096Q33), Q11 (GFP2096Q11), Q4 (GFP2096Q4) or Q33 alone
(GFPQ33). C-termini with 50% (DNLS1,2,3) or 20% nuclear localization
(DNLS/NES) were also assessed containing Q4, Q11 or Q33 glutamines.
Percent of transfected dead cells was normalized to the percent of untrans-
fected cells dead. Asterisks indicate statistically different results from respect-
ive GFP2069Qn (P , 0.05, one-way ANOVA). Twenty thousand events were
recorded per n (n ¼ 12). Statistically significant results were replicated on
three separate occasions. (B) FACS analyzed cell death in primary granule
cells after 7 DIV. Fifteen thousand events were recorded per mouse for
GFPQ33, GFP2096Q4, GFP2096Q11 and GFP2096Q33 (n ¼ 4, 4, 5 and 6,
respectively). (C) Representative plot of FACS controls. Included are popu-
lations of dead untransfected cells, healthy untransfected cells and healthy
transfected cells. (D) Representative confocal images of SHEKb3 cells and
granule cells expressing GFP2096Q33, GFP2096Q11, GFP2096Q4 or
GFPQ33. Error bars represent mean SE. Scale bar: 10 mm.
Human Molecular Genetics, 2006, Vol. 15, No. 101595
by guest on May 29, 2013
constitutive process for WT a1A as well as expanded alleles
and may reflect a normal signaling process. Thus, unlike in
HD, SCA1 or SCA3, therapeutic measures to prevent cleavage
or translocation of the C-terminus may interfere with normal
regulatory processes. This outcome might, in fact, be predicted
by the case of the leaner mouse mutant, in which the absence
of the a1A C-terminus results in selective Purkinje cell
Secondly, the pathogenic size of the polyglutamine expan-
sion in SCA6 is small compared with other polyglutamine
diseases, and the largest identified expansion (Q33), although
toxic as part of the a1A C-terminus, is in the normal range for
all other polyglutamine diseases (3,4,6). Together with the
finding that simple Q33 (Fig. 9) or Q35 (35) polyglutamine
tracts located in the nucleus have no effect on cell viability,
our finding that over-expressed WT a1A C-termini are slightly
toxic implies that the protein context of the a1A C-terminus
may mediate the effect of the pathogenic polyglutamine
tract. Our results are consistent with other studies that show
that expression of both expanded and unexpanded a1A
C-termini result in cell death, although the expanded tail is
significantly more toxic than the unexpanded (20). It is tempt-
ing to speculate that the motifs involved in the normal nuclear
function of the a1A C-terminus may facilitate the pathogen-
icity of the a1A polyglutamine tract. This conclusion is
further supported by the observation that in SCA6 the age of
onset correlates better with the combined size of expanded
and unexpanded polyglutamine alleles than with expanded
tracts alone (41). Moreover, one study suggests that patients
with another form of polyglutamine-mediated ataxia, SCA2,
develop ataxia at an earlier age if they bear normal unex-
panded a1A polyglutamine tracts in the larger, rather than
the smaller range (42).
Paradoxically, the leaner allele of the mouse CACNA1A
mutant lacks a1A splice forms bearing the same cleaved
region of the C-terminus and as homozygotes give rise to
marked Purkinje cell degeneration (22). Because cultured cer-
ebellar neurons from homozygote leaner mice exhibit 40%
reduced P/Q current density, one interpretation of this finding
is that Purkinje cell death results from reduced VGCC currents
(43). However, mice heterozygous for other CACNA1A
mutant alleles, such as the targeted a1A disruption (44),
exhibit similar reductions in P/Q current density, but have
normal cerebellar function and morphology. Moreover, heter-
ologous expression of truncated a1A subunits, predicted by
the leaner mutation, reveals that they exhibit minimally
altered channel kinetics (43). Thus, a plausible explanation
for the pronounced cell death phenotype in leaner is that the C-
terminus plays an added role in Purkinje cell viability. Our data
illustrating pronounced nuclear C-terminal staining in WT P18
Purkinje cell nuclei, coincident with the initiation of neuronal
degeneration in leaner mice, is consistent with a role for the
C-terminus in Purkinje cell viability.
Preliminary deletion studies suggest that the cleavage site is
above residue 2044 of the C-terminus near domains known to
interact with several cytoplasmic proteins (9,21,45,46). Online
sequence analysis using databases for protease sites and other
motifs reveals a putative PEST site downstream of the
predicted region of cleavage, but does not reveal discrete pro-
tease cleavage site in this region of the C-terminus (47).
Nevertheless, PEST sites are potential calpain recognition
site (47,48). Calpain has also been identified as the protease
that cleaves the C-terminus of the L-type calcium channel
(49). However, when expressed in SHEKb3 cells neither
calpain inhibitors nor co-expressed human calpastatin, a
natural calpain inhibitor, were able to block a1A cleavage
(unpublished data). This suggests that the putative protease
was resistant or spatially isolated from the protease inhibitors.
Cleavage and nuclear translocation of a portion of several
cytoplasmic proteins is a recently recognized cellular
process, shown in at least some cases to mediate nuclear sig-
naling. For example, bAPP, MAPK, p75 and Notch are
cleaved outside the nucleus, in direct or indirect response to
membrane signaling, to generate a polypeptide that is translo-
cated to the nucleus (50,51). More importantly, it has been
reported that the C-terminus of the a1C subunit of L-type
calcium channels is cleaved (49), and recently it has been
suggested that the a1C C-terminal fragment is also present
in cell nuclei and may act in transcription regulation (52).
Finally, although there is substantial evidence to indicate
that the C-terminus of the intact WT a1A subunit has a role
at the neuronal membrane (1,9,13), the potential role of the
cleaved C-terminus in the nucleus is as yet unknown. From
a structuralstandpoint, the
domains that associate with intracellular signaling proteins
to modulate channel activity and synaptic transmission, such
as calmodulin (CaM), b4 auxiliary VGCC subunits and G
protein bg (21,45,53), but no previous studies have identified
an association with any nuclear proteins.
In addition to the four NLS sequences, online motif analysis
using a sequence analysis package, Expasy (www.expasy.org),
predicts several potential interactions and post-translational
modifications that could reveal the function of a1A in the
nucleus. Further study will be necessary to explore the role of
these and other motifs in the biological activity of cleavage,
nuclear translocation and cell death in SCA6. Identification of
the biological activity of this translocation event may provide
and mechanisms of neurodegenerative disease.
MATERIALS AND METHODS
Cell culture and transfection
HEK293 cells and NIH3T3 cells were grown in Dulbecco’s
modified Eagle’s medium (DMEM)-F12 (Invitrogen, Carlsbad,
CA, USA) supplemented with 10% fetal bovine serum (FBS)
(Sigma), 2 mM L-glutamine and 50 mg/ml gentamicin (Invitro-
gen) in 5% CO2at 378C. Cells were transfected when they
were ?60% confluent with 0.4 mg of DNA per 35 mm dish
using Effectine reagent (Quiagen, Valencia, CA, USA),
0.4 mg of DNA per 35 mm dish was used in each transfection.
Stable HEK293 cells were made using linear bi-cistronic
pBudCE4.1 vector (Invitrogen) that expresses a2d and b3 sub-
units. Cells were selected with 250 mg/ml Zeocin (Invitrogen)
and cloned. Primary cerebellar cultures were obtained from
P7 mice. Cerebella were dissociated in 0.05% trypsin DMEM
(Invitrogen) containing DNAase. Cultures were maintained
FBS, 5% horse serum, 25 mM KCl, 2 mM glutamine and
1596Human Molecular Genetics, 2006, Vol. 15, No. 10
by guest on May 29, 2013
30 mM glucose. For each condition, 5 ? 106granule cells were
transfected with 5 mg of DNA by electroporation (Nucleofector
II; Amaxa, Koelin, Germany). Granule cells were plated at a
density of 2 ? 105cells/cm2.
All clones were made from a1A VGCC subunit cDNA
(GenBank accession no. AF004884, MERCK bioscience).
All GFP clones were generated with EGFP-C2 (Clontech,
Palo Alto, CA, USA). NLS and EGFP mutations were per-
formed via PCR (Quick-Change II XL Site-Directed Mutagen-
esis Kit; Stratagene, La Jolla, CA, USA). EGFP-C2 was
mutated to a fainter GFP (mGFP) for FACS by the mutations
L64F and T65S. GFPQ33 was generated by digestion of two
Q33 flanking Kpn1 sites and ligation into EGFP-C1. NES
tags were generated from custom oligonucleotides (Sigma)
encoding an NES derived from MAPKK and HIV-Rev
Immunohistochemistry was performed as previously reported
except as modified below (24). Briefly, paraffin-embedded
sections of perfused brains were dewaxed and rehydrated,
then steamed for 20 min in antigen retrieval solution
(Reveal; Biocare Medical, Walnut Creek, CA, USA). Sections
were blocked and exposed to primary antibody both NT-1 or
CT-2 (1 mg/ml) for 12 h at 48C, washed with phosphate-
buffered saline (PBS), incubated in biotinylated anti-rabbit
(Vector Labs, Burlingame, CA) secondary, reacted with
ABC (Vector Labs), washed and then counterstained with dia-
minobenzidine (Vector Labs). Images were acquired with
Zeiss light microscope and digital SPOT camera.
Immunocytochemistry of cell cultures was performed
?18–20 h after transfection of HEK cells or NIH3T3 cells
on poly-D-lysine coated cover slips and after 7 days in vitro
of primary neurons. Cells were fixed for 20 min in paraformal-
dehyde (PFA), washed three times in Tris-buffered saline
(TBS), permeabilized with 0.01% Tween, blocked 30 min in
10% donkey serum/TBS, incubated in NT-1 or CT-2
primary antibody, washed three times in TBS, then incubated
in secondary either Texas-Red donkey anti-rabbit (Jackson
Immunoresearch, Westgrove, PA, USA) or Alexafluor 488
goat anti-rabbit (Molecular probes, Eugene, OR, USA).
Images were acquired with a single-laser confocal microscope
(MRC 1024; Olympus, Melville, NY, USA) and LazerSharp
software (BIO-RAD, Hercules, CA, USA). Analysis was per-
formed with MetaMorph software. All cells were imaged at
the same settings. Intensity was determined with the brightest
condition exhibiting little saturation (,255). Percent of
nuclear staining was obtained by dividing the ARFI of the
cytoplasm by the sum of the ARFI for both the nucleus and
cytoplasm. Data are expressed as mean + standard error
(SE). Data were compared using Student’s t-test and differ-
ences were considered significant if P , 0.05. Comparison
of populations of cells labeled with either N-terminal or
C-terminal markers (Fig. 3C) were done with a significance
test for the estimation of a population proportion.
Forty-eight hours after transfection, cell cultures were lysed
with ice-cold cell lyse buffer (MPER reagent; Pierce, Rock-
ford, IL, USA) and mouse cerebellum were homogenized in
ice-cold tissue lyse buffer (TBS, 1% Triton X-100), both
buffers were supplemented with 10 mM EDTA, 10 mM
EGTA and protease inhibitor tabs (Roche, Indianapolis, IN,
USA). Cell fractionation was performed with NE-PER
Nuclear and cytoplasmic extraction reagent kit (Pierce).
Fifty micrograms of total protein or fraction were subjected
to SDS–PAGE (8% Tris–Glycine gel, Invitrogen) and trans-
ferred to a nitrocellulose membrane. Membranes were
probed with either a1A-specific custom antibodies CT-2 or
NT-1 raisedagainst RHGRRLPNGYYAGHGAPR
MARFGDEMPARYGGGGSG (15), respectively (Sigma),
anti-b3 (Alamone, Haifa, Israel), anti-Flag-M2 (Sigma),
anti-penta-HIS (Sigma), monoclonal anti-GAPDH (Abcam,
Cambridge, MA, USA), anti-GFP Living colors monoclonal
(Clontech) or anti-Histone 3 (Cell-Signaling, Beverly, MA,
USA). This was followed by incubation with a secondary
antibody, either anti-mouse or anti-rabbit IgG horseradish
peroxidase (Amersham Biosciences, Piscataway, NJ, USA).
Affinity purification was performed as previously reported (24).
Cell death assay
FACS was performed 72 h after transfection. SHEKb3 cells
were trypsinized and incubated in 2 mg/ml PI (Molecular
Probes) for 30 min at room temperature, washed in PBS and
fixed in 1% PFA/PBS at physiological pH. Cells were sorted
on FACS (FacsCalibur, BD Biosciences, San Jose, CA,
USA) and data analysis was performed with FACS quantifi-
cation software (Cellquestpro, BD Biosciences). Twenty thou-
sand events were recorded for each n, where n is equivalent to
one well in a six-well plate. mGFP constructs were used
because EGFP overlapped with FL2 filters. The percent of
transfected dead cells was expressed as the ratio of both
GFP- and PI-positivecells
GFP-positive cells. This was normalized to untransfected by
subtracting the percent of untransfected dead cells. Percent
untransfected dead was determined by dividing the number
of PI-positive cells by the total number of cells not GFP posi-
tive. Multiple groups of FACS data were analyzed using
one-way ANOVA and both Bonferroni and Tukey post hoc
tests. Results were plotted as mean + SE and differences
were considered significant if P , 0.05.
by the totalnumberof
We thank Colleen Forester, Katie M. Wiens and Anna
McDowell for their technical assistance. We would like to
acknowledge the assistance of the Flow Cytometry Core Facil-
ity of the University of Minnesota Cancer Center, a compre-
hensive cancer center designated by the National Cancer
Institute, supported in part by P30 CA77598. This work was
supported by NIH NS38332, the National Ataxia Foundation
and the Bob Alison Ataxia Research Center.
Conflict of Interest statement. There is no conflict of interest.
Human Molecular Genetics, 2006, Vol. 15, No. 10 1597
by guest on May 29, 2013
1. Garg, M.K. and Sanchette, P.C. (1999) Ion channels and channelopathy.
J. Assoc. Physicians India, 47, 436–439.
2. Pietrobon, D. (2002) Calcium channels and channelopathies of the central
nervous system. Mol. Neurobiol., 25, 31–50.
3. Ross, C.A. (2002) Polyglutamine pathogenesis: emergence of unifying
mechanisms for Huntington’s disease and related disorders. Neuron, 35,
4. La Spada, A.R. and Taylor, J.P. (2003) Polyglutamines placed into
context. Neuron, 38, 681–684.
5. Okazawa, H. (2003) Polyglutamine diseases: a transcription disorder?
Cell. Mol. Life Sci., 60, 1427–1439.
6. Orr, H.T. (2001) Beyond the Qs in the polyglutamine diseases. Genes
Dev., 15, 925–932.
7. Riess, O., Schols, L., Bottger, H., Nolte, D., Vieira-Saecker, A.M.,
Schimming, C., Kreuz, F., Macek, M., Jr, Krebsova, A., Macek, M.S.
et al. (1997) SCA6 is caused by moderate CAG expansion in the
alpha1A-voltage-dependent calcium channel gene. Hum. Mol. Genet., 6,
8. Zhuchenko, O., Bailey, J., Bonnen, P., Ashizawa, T., Stockton, D.W.,
Amos, C., Dobyns, W.B., Subramony, S.H., Zoghbi, H.Y. and Lee, C.C.
(1997) Autosomal dominant cerebellar ataxia (SCA6) associated with
small polyglutamine expansions in the alpha 1A-voltage-dependent
calcium channel. Nat. Genet., 15, 62–69.
9. Catterall, W.A. (2000) Structure and regulation of voltage-gated Ca2þ
channels. Annu. Rev. Cell Dev. Biol., 16, 521–555.
10. Dunlap, K., Luebke, J.I. and Turner, T.J. (1995) Exocytotic Ca2þ
channels in mammalian central neurons. Trends Neurosci., 18, 89–98.
11. Kulik, A., Nakadate, K., Hagiwara, A., Fukazawa, Y., Lujan, R., Saito, H.,
Suzuki, N., Futatsugi, A., Mikoshiba, K., Frotscher, M. et al. (2004)
Immunocytochemical localization of the alpha 1A subunit of the P/
Q-type calcium channel in the rat cerebellum. Eur. J. Neurosci., 19,
12. Wheeler, D.B., Randall, A. and Tsien, R.W. (1994) Roles of N-type and
Q-type Ca2þ channels in supporting hippocampal synaptic transmission.
Science, 264, 107–111.
13. Westenbroek, R.E., Sakurai, T., Elliott, E.M., Hell, J.W., Starr, T.V.,
Snutch, T.P. and Catterall, W.A. (1995) Immunochemical identification
and subcellular distribution of the alpha 1A subunits of brain calcium
channels. J. Neurosci., 15, 6403–6418.
14. Bourinet, E., Soong, T.W., Sutton, K., Slaymaker, S., Mathews, E.,
Monteil, A., Zamponi, G.W., Nargeot, J. and Snutch, T.P. (1999) Splicing
of alpha 1A subunit gene generates phenotypic variants of P- and Q-type
calcium channels. Nat. Neurosci., 2, 407–415.
15. Sakurai, T., Westenbroek, R.E., Rettig, J., Hell, J. and Catterall, W.A.
(1996) Biochemical properties and subcellular distribution of the BI and
rbA isoforms of alpha 1A subunits of brain calcium channels. J. Cell Biol.,
16. Soong, T.W., DeMaria, C.D., Alvania, R.S., Zweifel, L.S., Liang, M.C.,
Mittman, S., Agnew, W.S. and Yue, D.T. (2002) Systematic identification
of splice variants in human P/Q-type channel alpha1(2.1) subunits:
implications for current density and Ca2þ-dependent inactivation.
J. Neurosci., 22, 10142–10152.
(1995) Immunochemical identification and differential phosphorylation of
alternatively spliced forms of the alpha 1A subunit of brain calcium
channels. J. Biol. Chem., 270, 21234–21242.
18. Scott, V.E., Felix, R., Arikkath, J. and Campbell, K.P. (1998) Evidence
for a 95 kDa short form of the alpha1A subunit associated with the
omega-conotoxin MVIIC receptor of the P/Q-type Ca2þ channels.
J. Neurosci., 18, 641–647.
19. Westenbroek, R.E., Hell, J.W., Warner, C., Dubel, S.J., Snutch, T.P. and
Catterall, W.A. (1992) Biochemical properties and subcellular distribution
of an N-type calcium channel alpha 1 subunit. Neuron, 9, 1099–1115.
20. Kubodera, T., Yokota, T., Ohwada, K., Ishikawa, K., Miura, H.,
Matsuoka, T. and Mizusawa, H. (2003) Proteolytic cleavage and cellular
toxicity of the human alpha1A calcium channel in spinocerebellar ataxia
type 6. Neurosci. Lett., 341, 74–78.
21. Walker, D. and De Waard, M. (1998) Subunit interaction sites in
voltage-dependent Ca2þ channels: role in channel function. Trends
Neurosci., 21, 148–154.
22. Fletcher, C.F., Lutz, C.M., O’Sullivan, T.N., Shaughnessy, J.D., Jr,
Hawkes, R., Frankel, W.N., Copeland, N.G. and Jenkins, N.A. (1996)
Absence epilepsy in tottering mutant mice is associated with calcium
channel defects. Cell, 87, 607–617.
23. Ishikawa, K., Fujigasaki, H., Saegusa, H., Ohwada, K., Fujita, T.,
Iwamoto, H., Komatsuzaki, Y., Toru, S., Toriyama, H., Watanabe, M.
et al. (1999) Abundant expression and cytoplasmic aggregations of
[alpha]1A voltage-dependent calcium channel protein associated with
neurodegeneration in spinocerebellar ataxia type 6. Hum. Mol. Genet., 8,
24. Restituito, S., Thompson, R.M., Eliet, J., Raike, R.S., Riedl, M.,
Charnet, P. and Gomez, C.M. (2000) The polyglutamine expansion in
spinocerebellar ataxia type 6 causes a beta subunit-specific enhanced
activation of P/Q-type calcium channels in Xenopus oocytes. J. Neurosci.,
25. Tsunemi, T., Saegusa, H., Ishikawa, K., Nagayama, S., Murakoshi, T.,
Mizusawa, H. and Tanabe, T. (2002) Novel Cav2.1 splice variants isolated
from Purkinje cells do not generate P-type Ca2þ current. J. Biol. Chem.,
26. Boulikas, T. (1994) Putative nuclear localization signals (NLS) in protein
transcription factors. J. Cell. Biochem., 55, 32–58.
27. Irie, Y., Yamagata, K., Gan, Y., Miyamoto, K., Do, E., Kuo, C.H.,
Taira, E. and Miki, N. (2000) Molecular cloning and characterization of
Amida, a novel protein which interacts with a neuron-specific immediate
early gene product arc, contains novel nuclear localization signals, and
causes cell death in cultured cells. J. Biol. Chem., 275, 2647–2653.
28. Truant, R. and Cullen, B.R. (1999) The arginine-rich domains present in
human immunodeficiency virus type 1 Tat and Rev function as direct
importin beta-dependent nuclear localization signals. Mol. Cell. Biol., 19,
29. Christophe, D., Christophe-Hobertus, C. and Pichon, B. (2000) Nuclear
targeting of proteins: how many different signals? Cell. Signal., 12, 337–341.
30. Nadler, S.G., Tritschler, D., Haffar, O.K., Blake, J., Bruce, A.G. and
Cleaveland, J.S. (1997) Differential expression and sequence-specific
interaction of karyopherin alpha with nuclear localization sequences.
J. Biol. Chem., 272, 4310–4315.
31. Taylor, J., Grote, S.K., Xia, J., Vandelft, M., Graczyk, J., Ellerby, L.M.,
La Spada, A.R. and Truant, R. (2006) Ataxin-7 can export from the
nucleus via a conserved exportin-dependent signal. J. Biol. Chem., 281,
32. Matsuyama, Z., Wakamori, M., Mori, Y., Kawakami, H., Nakamura, S.
and Imoto, K. (1999) Direct alteration of the P/Q-type Ca2þ channel
property by polyglutamine expansion in spinocerebellar ataxia 6.
J. Neurosci., 19, RC14.
33. Piedras-Renteria, E.S., Watase, K., Harata, N., Zhuchenko, O.,
Zoghbi, H.Y., Lee, C.C. and Tsien, R.W. (2001) Increased expression of
alpha 1A Ca2þ channel currents arising from expanded trinucleotide
repeats in spinocerebellar ataxia type 6. J. Neurosci., 21, 9185–9193.
34. Toru, S., Murakoshi, T., Ishikawa, K., Saegusa, H., Fujigasaki, H.,
Uchihara, T., Nagayama, S., Osanai, M., Mizusawa, H. and Tanabe, T.
(2000) Spinocerebellar ataxia type 6 mutation alters P-type calcium
channel function. J. Biol. Chem., 275, 10893–10898.
35. Moulder, K.L., Onodera, O., Burke, J.R., Strittmatter, W.J. and
Johnson, E.M., Jr. (1999) Generation of neuronal intranuclear inclusions
by polyglutamine-GFP: analysis of inclusion clearance and toxicity as a
function of polyglutamine length. J. Neurosci., 19, 705–715.
36. Nucifora, F.C., Jr, Ellerby, L.M., Wellington, C.L., Wood, J.D.,
Herring, W.J., Sawa, A., Hayden, M.R., Dawson, V.L., Dawson, T.M. and
Ross, C.A. (2003) Nuclear localization of a non-caspase truncation
product of atrophin-1, with an expanded polyglutamine repeat, increases
cellular toxicity. J. Biol. Chem., 278, 13047–13055.
37. Klement, I.A., Skinner, P.J., Kaytor, M.D., Yi, H., Hersch, S.M.,
Clark, H.B., Zoghbi, H.Y. and Orr, H.T. (1998) Ataxin-1 nuclear
localization and aggregation: role in polyglutamine-induced disease in
SCA1 transgenic mice. Cell, 95, 41–53.
38. Yabe, I., Sasaki, H., Matsuura, T., Takada, A., Wakisaka, A., Suzuki, Y.,
Fukazawa, T., Hamada, T., Oda, T., Ohnishi, A. et al. (1998) SCA6
mutation analysis in a large cohort of the Japanese patients with late-onset
pure cerebellar ataxia. J. Neurol. Sci., 156, 89–95.
39. Takeyama, K., Ito, S., Yamamoto, A., Tanimoto, H., Furutani, T.,
Kanuka, H., Miura, M., Tabata, T. and Kato, S. (2002)
Androgen-dependent neurodegeneration by polyglutamine-expanded
human androgen receptor in Drosophila. Neuron, 35, 855–864.
1598 Human Molecular Genetics, 2006, Vol. 15, No. 10
by guest on May 29, 2013
40. Huynh, D.P., Figueroa, K., Hoang, N. and Pulst, S.M. (2000) Nuclear
localization or inclusion body formation of ataxin-2 are not necessary for
SCA2 pathogenesis in mouse or human. Nat. Genet., 26, 44–50.
41. van de Warrenburg, B.P., Hendriks, H., Durr, A., van Zuijlen, M.C.,
Stevanin, G., Camuzat, A., Sinke, R.J., Brice, A. and Kremer, B.P. (2005)
Age at onset variance analysis in spinocerebellar ataxias: a study in a
Dutch-French cohort. Ann. Neurol., 57, 505–512.
42. Pulst, S.M., Santos, N., Wang, D., Yang, H., Huynh, D., Velazquez, L.
and Figueroa, K.P. (2005) Spinocerebellar ataxia type 2: polyQ repeat
variation in the CACNA1A calcium channel modifies age of onset. Brain,
43. Wakamori, M., Yamazaki, K., Matsunodaira, H., Teramoto, T., Tanaka, I.,
Niidome, T., Sawada, K., Nishizawa, Y., Sekiguchi, N., Mori, E. et al.
(1998) Single tottering mutations responsible for the neuropathic phenotype
of the P-type calcium channel. J. Biol. Chem., 273, 34857–34867.
44. Fletcher, C.F., Tottene, A., Lennon, V.A., Wilson, S.M., Dubel, S.J.,
Paylor, R., Hosford, D.A., Tessarollo, L., McEnery, M.W., Pietrobon, D.
et al. (2001) Dystonia and cerebellar atrophy in Cacna1a null mice lacking
P/Q calcium channel activity. FASEB J., 15, 1288–1290.
45. Kinoshita, M., Nukada, T., Asano, T., Mori, Y., Akaike, A., Satoh, M. and
Kaneko, S. (2001) Binding of G alpha(o) N terminus is responsible for the
voltage-resistant inhibition of alpha(1A) (P/Q-type, Ca(v)2.1) Ca(2þ)
channels. J. Biol. Chem., 276, 28731–28738.
46. Walker, D., Bichet, D., Campbell, K.P. and De Waard, M. (1998) A beta 4
isoform-specific interaction site in the carboxyl-terminal region of the
voltage-dependent Ca2þ channel alpha 1A subunit. J. Biol. Chem., 273,
47. Rechsteiner, M. (1990) PEST sequences are signals for rapid intracellular
proteolysis. Semin. Cell Biol., 1, 433–440.
48. Rechsteiner, M. and Rogers, S.W. (1996) PEST sequences and regulation
by proteolysis. Trends Biochem. Sci., 21, 267–271.
49. De Jongh, K.S., Colvin, A.A., Wang, K.K. and Catterall, W.A.
(1994) Differential proteolysis of the full-length form of the L-type
calcium channel alpha 1 subunit by calpain. J. Neurochem., 63,
50. Bland, C.E., Kimberly, P. and Rand, M.D. (2003) Notch-induced
proteolysis and nuclear localization of the Delta ligand. J. Biol. Chem.,
51. Jung, K.M., Tan, S., Landman, N., Petrova, K., Murray, S., Lewis, R.,
Kim, P.K., Kim, D.S., Ryu, S.H., Chao, M.V. et al. (2003) Regulated
intramembrane proteolysis of the p75 neurotrophin receptor modulates
its association with the TrkA receptor. J. Biol. Chem., 278,
52. Wilson, G.F., MacLean, J.M., Wilson, G.F., Isom, L.L. and
Gomez-Ospina, N. (2005) Non-conducting functions of voltage-gated ion
channels. Abstract Viewer/Itinerary Planner. Society for Neuroscience,
Washington, DC, Vol. 2005. Online sfn.scholarone.com/itin2005.
53. Liang, H., DeMaria, C.D., Erickson, M.G., Mori, M.X.,
Alseikhan, B.A. and Yue, D.T. (2003) Unified mechanisms
of Ca2þ regulation across the Ca2þ channel family. Neuron, 39,
54. Kutay, U. and Guttinger, S. (2005) Leucine-rich nuclear-export signals:
born to be weak. Trends Cell Biol., 15, 121–124.
Human Molecular Genetics, 2006, Vol. 15, No. 10 1599
by guest on May 29, 2013