ctcf regulates ataxin-7 expression through promotion of a
TRANSCRIPT
brought to you by COREView metadata, citation and similar papers at core.ac.uk
provided by Elsevier - Publisher Connector
Neuron
Article
CTCF Regulates Ataxin-7 Expressionthrough Promotion of a ConvergentlyTranscribed, Antisense Noncoding RNABryce L. Sopher,1,11 Paula D. Ladd,4,11 Victor V. Pineda,1,11 Randell T. Libby,1 Susan M. Sunkin,9 James B. Hurley,2
Cortlandt P. Thienes,1 Terry Gaasterland,6,7 Galina N. Filippova,10 and Albert R. La Spada3,4,5,7,8,*1Department of Laboratory Medicine2Department of Biochemistry3Department of Medicine
University of Washington, Seattle, WA 98195, USA4Department of Pediatrics5Department of Cellular & Molecular Medicine6Scripps Institute for Oceanography7Institute for Genomic Medicine
University of California, San Diego, La Jolla, CA 92093, USA8Rady Children’s Hospital, San Diego, CA 92123, USA9Allen Institute for Brain Science, Seattle, WA 98103, USA10Fred Hutchinson Cancer Research Center, Seattle, WA 98109, USA11These authors contributed equally to this work
*Correspondence: [email protected] 10.1016/j.neuron.2011.05.027
SUMMARY
Spinocerebellar ataxia type 7 (SCA7) is a neurode-generative disorder caused by CAG/polyglutaminerepeat expansions in the ataxin-7 gene. Ataxin-7is a component of two different transcription coacti-vator complexes, and recent work indicates thatdisease protein normal function is altered in polyglut-amine neurodegeneration. Given this, we studiedhow ataxin-7 gene expression is regulated. Theataxin-7 repeat and translation start site are flankedby binding sites for CTCF, a highly conserved multi-functional transcription regulator. When we analyzedthis region, we discovered an adjacent alternativepromoter and a convergently transcribed antisensenoncoding RNA, SCAANT1. To understand howCTCF regulates ataxin-7 gene expression, we intro-duced ataxin-7 mini-genes into mice, and foundthat CTCF is required for SCAANT1 expression.Loss of SCAANT1 derepressed ataxin-7 sense tran-scription in a cis-dependent fashion and was accom-panied by chromatin remodeling. Discovery of thispathway underscores the importance of alteredepigenetic regulation for disease pathology at repeatloci exhibiting bidirectional transcription.
INTRODUCTION
Spinocerebellar ataxia type 7 (SCA7) is an inherited neurological
disorder characterized by cerebellar and retinal degeneration
(Martin et al., 1994). SCA7 is caused by a CAG/polyglutamine
(polyQ) repeat expansion in the ataxin-7 gene and is therefore
one of nine polyQ neurodegenerative disorders (La Spada and
Taylor, 2010). Included in the CAG/polyQ repeat disease cate-
gory are spinobulbar muscular atrophy (SBMA), Huntington’s
disease (HD), dentatorubral-pallidoluysian atrophy (DRPLA),
and five other forms of spinocerebellar ataxia (SCA1, 2, 3, 6,
and 17). Numerous lines of investigation in the polyQ disease
field suggest that expansion of the glutamine tract is a gain-of-
function mutation, and that the initiating event in disease patho-
genesis is transition of the polyQ expansion tract to an altered
conformation (Paulson et al., 2000; Ross, 1997). However, as
each polyQ disease displays distinct patterns of neuropathology
despite overlapping patterns of disease gene expression, it is
likely that the normal function, activities, and interactions of the
polyQ disease protein determine the cell-type specificity in
each disorder (La Spada and Taylor, 2003). Ataxin-7, the causal
protein in SCA7, contains a polyQ tract that ranges in size from
4–35 glutamines in normal individuals, but expands to 37–>400
glutamines in affected patients (David et al., 1997; Stevanin
et al., 2000). The glutamine tract is located in the amino-terminus
of ataxin-7, beginning at position #30. SCA7 is unique among
the CAG/polyQ repeat diseases, as patients with this disorder
develop a retinal degeneration phenotype, classified as a
cone-rod dystrophy (Ahn et al., 2005; To et al., 1993). To under-
stand the basis of SCA7 retinal degeneration and the reason for
the selective loss of photoreceptor cells in this disease, we, and
others, produced transgenic mice that recapitulated the
SCA7 cone-rod dystrophy phenotype and found that SCA7
retinal degeneration results from altered transcription regulation
(Helmlinger et al., 2006; La Spada et al., 2001; Yoo et al., 2003).
As the vast majority of CAG/polyQ disease proteins are well-
known transcription factors or can function as transcription
Neuron 70, 1071–1084, June 23, 2011 ª2011 Elsevier Inc. 1071
Neuron
CTCF Promotes AS ncRNA to Control atx7 Expression
co-regulators (Riley and Orr, 2006), a role for transcription dysre-
gulation in SCA7 is consistent with an emerging view of these
disorders as ‘‘transcriptionopathies’’ (La Spada and Taylor,
2003). The existence of an interaction between ataxin-7 and
a retinal transcription factor, known as CRX, suggested that
ataxin-7 is a transcription factor (La Spada et al., 2001), and
this was supported by demonstration of a functional nuclear
localization signal in ataxin-7 (Chen et al., 2004). When studies
of the yeast ortholog of ataxin-7, Sgf73, indicated that Sgf73 is
part of the SAGA complex (Sanders et al., 2002), we, and others,
found that ataxin-7 is a core component of the analogous coac-
tivator complex in mammals, known as the STAGA (Spt3-Taf9-
Ada-Gcn5-acetyltransferase) complex (Helmlinger et al., 2004;
Palhan et al., 2005). STAGA is a transcriptional coactivator
complex with histone acetyltransferase (HAT) activity (Martinez
et al., 2001). In addition to being part of the STAGA complex,
yeast Sgf73 and mammalian ataxin-7 are respectively compo-
nents of the Ubp8/USP22 deubiquitination complex (Kohler
et al., 2008; Zhao et al., 2008). While the role of altered STAGA
and USP22 deubiquitination complex function in SCA7 disease
pathogenesis is unclear, recent studies of the related polyQ
disorder SCA1 indicate that the polyQ expansion in ataxin-1
attenuates the formation and function of the Capicua transcrip-
tion factor complex, contributing to SCA1 disease pathogenesis
through a partial loss-of-function mechanism (Chen et al., 2003;
Lim et al., 2008). Hence, polyQ disease may result from an alter-
ation of normal function, combined with a gain-of-function
mechanism, and in the case of SCA7, the native protein function
of ataxin-7 appears critically important for chromatin remodeling
at the level of histone acetylation and deubiquitination.
In addition to the CAG/polyQ repeat diseases, at least three
other subclasses of repeat expansion disease are recognized:
loss-of-function repeat expansion diseases, RNA gain-of-func-
tion repeat disorders, and polyalanine gain-of-function repeat
expansion diseases (La Spada and Taylor, 2010). Although the
four repeat expansion disease subcategories are based upon
presumed differences in their mechanisms of repeat mutation
toxicity, recent studies have revealed a number of shared
genomic features for repeat disease loci, irrespective of repeat
disease category, including (1) genetic instability characterized
by a strong tendency for repeats to further expand upon germ
line transmission (Pearson et al., 2005); (2) binding sites for the
multivalent transcription regulatory factor CTCF (Ohlsson et al.,
2001), within close proximity of the repeats (Filippova et al.,
2001); and (3) bidirectional transcription typically encompassing
the repeat itself (Batra et al., 2010). These features suggest that
certain epigenetic processes and chromatin regulatory path-
ways may be shared in common between different repeat
diseases. As the SCA7 CAG repeat is the most unstable of all
the CAG/polyQ repeat loci, and the SCA7 CAG repeat is closely
flanked by two functional CTCF binding sites, the SCA7 CAG
repeat is among the repeat disease loci likely to display this
constellation of genomic features.
In light of the importance of ataxin-7 normal function for
SCA7 disease pathogenesis and potentially for global transcrip-
tion regulation, we initiated a series of studies aimed at under-
standing how ataxin-7 gene expression is regulated. The
ataxin-7 CAG repeat tract and the start site of translation are
1072 Neuron 70, 1071–1084, June 23, 2011 ª2011 Elsevier Inc.
both located in exon 3, which is flanked by two functional
CTCF binding sites (Filippova et al., 2001). CTCF is a highly
conserved 11 zinc-finger protein that mediates a variety of tran-
scription regulatory functions, including transcription activation,
transcription repression, insulator-boundary domain formation,
and genomic imprinting (Phillips and Corces, 2009). When we
analyzed the ataxin-7 repeat region, we discovered evidence
for an alternative promoter just 50 to exon 3, and identified an
antisense non-coding RNA, SCAANT1 (for spinocerebellar
ataxia-7 antisense noncoding transcript 1) that is convergently
transcribed across exon 4, exon 3, and the alternative promoter.
To understand the role of CTCF in regulating ataxin-7 transcrip-
tion, we introduced ataxin-7 minigenes, containing the ataxin-7
repeat region with a CAG repeat expansion, into transgenic
mice. Studies of these transgenic mice and of human retinoblas-
toma cell lines revealed that CTCF binding is required for
production of SCAANT1, and that loss of SCAANT1 expression
de-repressed ataxin-7 sense transcription from the alternative
promoter. Although SCAANT1 expression in trans did not reduce
ataxin-7 alternative sense promoter activity in vitro or in vivo,
convergent transcription of SCAANT1 in cis led to repression
that was accompanied by posttranslational modification of
histones. Our studies reveal a regulatory pathway that links
CTCF transactivation of antisense noncoding RNA with repres-
sion of the corresponding sense transcript. The likely contribu-
tion of this pathway to SCA7 disease pathogenesis underscores
the potential importance of altered epigenetic regulation for
disease pathology at repeat loci characterized by bidirectional
transcription.
RESULTS
Identification of Overlapping Sense and AntisenseTranscripts at the Ataxin-7 LocusThe ataxin-7 CAG/polyglutamine (polyQ) repeat is located in the
first coding exon (i.e., exon 3) and is flanked by two CTCF
binding sites (Filippova et al., 2001). To determine the basis of
ataxin-7 gene expression regulation, we surveyed ataxin-7
human genomic sequence, including exon 3 and the CAG repeat
region, with the UCSC genome browser (http://genome.ucsc.
edu). Bioinformatics analysis of this region, using FirstEF
(Davuluri et al., 2001), revealed evidence for an alternative
promoter in the intron 50 to exon 3. The presence of a strong
peak for the promoter-associated H3K4me3 modification at
this location strongly supported the existence of this alternative
promoter (see Figure S1 available online). When we analyzed
mouse ataxin-7 genomic sequence, we found that the transcrip-
tion start site (TSS) for the murine ataxin-7 gene is located in the
ataxin-7 repeat region in close proximity to the alternative
promoter predicted for the human gene. Interestingly, the previ-
ously defined human ataxin-7 TSS, which is located >40 kb 50 tothis region, annotated on the UCSC genome browser, and vali-
dated by RLM-RACE (Figure S1), is not predicted as a promoter
or TSS in mouse (http://genome.ucsc.edu). To evaluate this
prediction, we performed RLM-RACE on murine RNA samples
and identified a cluster of TSSs located 85, 100, and 255 nucle-
otides 50 to the annotated mouse ataxin-7 TSS. However, unlike
in the human, there is no distant upstream ataxin-7 TSS in mice.
A
ATG
(CAG)n
EXON 2A EXON 3 EXON 4
P2A
50 bp
> 38 kb to exon 5
>13 kb to exon 2
1Fluc
luc
luc
luc
2F
3F
4F
5F
luc
luc
luc
2F1
2F2
2F3
luc
CTCF-II CTCF-I
B
12
0
2
4
RLU
(norm
aliz
ed)
6
8
10
14
1F 2F 3F 4F 5F 2F1 2F2 2F3EV
-635 to +288
-246 to +288
-129 to +288
+79 to +288
+184 to +288
-246 to +184
-246 to +79
-246 to -129
Figure 1. The Ataxin-7 Gene Contains an Alterna-
tive Sense Promoter
(A) Diagram of the genomic region of the human ataxin-7
gene and constructs generated to map the alternative
sense promoter. Boxes represent exons, while solid lines
correspond to introns. The alternative promoter is indi-
cated as P2A and is predicted to produce a transcript that
contains an alternatively spliced intron (gray box). Various
genomic fragments, containing 10 CAG/CTG repeats,
were cloned into a firefly luciferase vector to yield the
different luciferase expression constructs shown here.
(B) Transactivation assay results for human ataxin-7
genomic fragments. The luciferase expression constructs
shown in (A) were transfected into primary astrocytes,
along with Renilla luciferase constructs for normalization,
and relative luminescence unit (RLU) measurements were
obtained. EV corresponds to empty vector, all experi-
ments were performed in quadruplicate, and errors bars =
SEM. RLU activity obtained for the vector 1F construct
corresponded to �10% of a positive control CMV-eGFP-
IRES-luciferase vector (not shown). See Figure S1.
Neuron
CTCF Promotes AS ncRNA to Control atx7 Expression
Thus, the human ataxin-7 gene contains two promoters: the
standard, upstream ‘‘P1’’ promoter and the alternative ‘P2A’
promoter (Figure S2), while the mouse ataxin-7 gene contains
only one promoter—P2A.
Although computational algorithms for identification of
promoters and TSSs are powerful approaches for defining regu-
latory regions, experimental validation is necessary. To confirm
the promoter calls, and to define the location of regulatory
elements in this region, we cloned a series of human ataxin-7
CAG10 genomic fragments into a luciferase reporter (Figure 1A)
and transfected the different ataxin-7 genomic fragment—lucif-
erase constructs into primary cerebellar astrocytes. When we
measured relative luciferase activity, we detected robust lucif-
erase transactivation for ataxin-7 genomic fragments containing
sequences just 50 to the newly discovered TSS (Figure 1B). To
define the alternative promoter, which we labeled ‘‘P2A,’’ we
performed 50 RACE and found that the alternative ataxin-7 TSS
is located at the 30 end of intron 2. Sequencing of 50 RACE clones
revealed that transcripts initiating from P2A contain a single first
exon comprising the last �400 bp of intron 2 and exon 3, or
Neuron 70, 107
instead undergo splicing to join a shorter exon
(‘‘2A’’) with exon 3 (Figure 1A).
In the course of inventorying ESTs in the
ataxin-7 repeat region, we discovered an EST
(BU569004) in antisense orientation. Analysis
of the sequence in this region revealed that
the 50 end of this EST corresponded to a NotI
restriction site, and that the 30 end of this EST
colocalized with a putative polyadenylation
signal sequence in antisense orientation (Fig-
ure 2A). As NotI digests were typically used
during the former era of EST discovery, the 50
end of this transcript was likely created during
its cloning. To identify the extent of the anti-
sense transcript, we performed strand-specific
RT-PCR and 50 RACE, and mapped the TSS to
intron 4 (Figure 2A). We named this antisense
transcript SCAANT1 for ‘‘spinocerebellar ataxia-7 antisense
noncoding transcript 1.’’ To delineate the regulatory region
responsible for transcription of SCAANT1, we cloned a series
of human ataxin-7 CAG10 genomic fragments into a luciferase
reporter construct in antisense orientation (Figure 2A) and trans-
fected the different ataxin-7 antisense genomic fragment—lucif-
erase constructs into primary cerebellar astrocytes. We noted
that a short stretch of DNA 50 to the SCAANT1 TSS was required
for transactivation, while a sizable sequence 30 to the SCAANT1
TSS was needed to achieve robust transactivation (Figure 2B).
As the two CTCF binding sites lie within the regulatory domain
mapped by the luciferase reporter assays, we derived another
set of luciferase reporter constructs, based upon our most
potent construct 2R, in which we mutated either of the CTCF
binding sites (Figure 2A). When we measured the transactivation
competence of the 2R-m2 and 2R-m1 constructs, we observed
marked reductions in luciferase activity (Figure 2B), suggesting
that CTCF binding site integrity is required for maximal
SCAANT1 expression. We also derived an ataxin-7 antisense
construct carrying a CAG92 repeat expansion (2R-exp), and
1–1084, June 23, 2011 ª2011 Elsevier Inc. 1073
A
12
0
2
4
6
8
10
2R 4R 5REV 2R-m2 2R-m1 2R-exp3R1R
RLU
(no
rmal
ized
)
B
ATG
(CAG)n
EXON 2A EXON 3 EXON 4
P2A
50 bp
> 38 kb to exon 5
>13 kb to exon 2 AAGTAAA
Not ICTCF-II CTCF-I
AAAAA
BU569004
luc
luc
luc
luc
luc
luc
X
X
(CAG)92
3R
4R
5R
2R-m2
2R-m1
2R-exp
luc
luc1R
2R-311 to +539
-175 to +539
-311 to +336
-311 to +196
-3517 to +539
‘SCAANT1’
Figure 2. Identification and Mapping of an Anti-
sense ncRNA at the Human Ataxin-7 Locus
(A) Diagram of the genomic region of the human ataxin-7
gene, showing the human ataxin-7 antisense ncRNA
(SCAANT1), and the constructs generated to map the
SCAANT1 promoter. Boxes represent exons, while solid
lines correspond to introns. The location of BU569004—
an antisense EST, a predicted polyadenylation signal
sequence, the transcriptional domain for SCAANT1, and
two CTCF binding sites are shown. Various genomic
fragments in antisense orientation, all containing 10 CAG/
CTG repeats except for 2R-exp, were cloned into a firefly
luciferase vector to produce the different expression
constructs shown here.
(B) Transactivation assay results for human ataxin-7
antisense genomic fragments. The luciferase expression
constructs shown in (A) were transfected into primary
astrocytes, along with Renilla luciferase constructs for
normalization, and relative luminescence unit (RLU)
measurements were obtained. Construct 2R-m2 contains
amutated CTCF-II binding site, construct 2R-m1 contains
a mutated CTCF-I binding site, and construct 2R-exp
contains a 92 CAG/CTG repeat. EV corresponds to empty
vector, experiments were performed in quadruplicate,
and errors bars = SEM.
See Figure S2.
Neuron
CTCF Promotes AS ncRNA to Control atx7 Expression
when we measured its transactivation competence, we docu-
mented a significant reduction in luciferase activity (Figure 2B).
An Ataxin-7 Genomic Fragment Yields a SCA7Phenotype in Transgenic Mice upon Mutationof the 30 CTCF Binding SiteThe existence of an �1.4 kb antisense noncoding transcript
overlapping a potentially strong sense promoter at the human
ataxin-7 locus suggested that their transcription regulationmight
be linked. As CTCF binding site integrity was required for
SCAANT1 transcription, we derived two ataxin-7 minigene
constructs that contain the sense P2A promoter and SCAANT1,
flanked by �5 kb of DNA 50 to this region and�8 kb of DNA 30 tothis region (Figure S2). Within this 13.5 kb human ataxin-7
genomic fragment reside two CTCF binding sites, known as
CTCF-I and CTCF-II. To understand the regulatory relationship
between SCAANT1 and ataxin-7 transcription from promoter
P2A, we introduced an 11 nucleotide substitution mutation at
the 30 CTCF-I binding site (Figure S2). The location of the muta-
1074 Neuron 70, 1071–1084, June 23, 2011 ª2011 Elsevier Inc.
tion was based upon DNA footprinting analysis,
and validation of abrogated CTCF binding was
achieved by electrophoretic mobility shift
assays, as we have shown (Libby et al.,
2008). In this way, we derived two distinct
ataxin-7 genomic fragment constructs with an
expanded CAG repeat tract: SCA7-CTCF-I-wt
and SCA7-CTCF-I-mut (Figure S2). To deter-
mine the role of CTCF binding in regulating
ataxin-7 repeat instability and transcription,
we used these constructs to produce lines of
transgenic mice. We have shown that mutation
of the CTCF-I binding site significantly dimin-
ishes CTCF occupancy in vivo in the SCA7-CTCF-I-mut mice
by ChIP analysis and found that mutation of the CTCF-I binding
site leads to increased repeat instability in the germline and
somatic tissues (Libby et al., 2008). Further studies of these
mice also revealed that SCA7-CTCF-I-mut mice become tremu-
lous, display weight loss, and develop an unsteady gait at
5–9 months of age (Movie S1). This phenotype, which is
observed in both SCA7-CTCF-I-mut transgenic lines ((1) and
(2)), progresses to become a prominent gait ataxia until the
mice die prematurely at 8–14 months of age, with the SCA7-
CTCF-I-mut-(2) line exhibiting a more rapidly progressive and
severe phenotype. In contrast, four independent lines of SCA7-
CTCF-I-wt mice did not exhibit any physical or neurological
abnormalities, and have a normal lifespan.
As SCA7-CTCF-I-mut transgenic mice develop a pronounced
ataxia, reminiscent of the gait difficulties seen in SCA7 patients
and in other lines of SCA7 transgenic mice (La Spada et al.,
2001; Yoo et al., 2003), we performed histopathology studies
and behavioral testing. SCA7 patients develop a cone-rod
Figure 3. SCA7-CTCF-I-mut Mice Develop
Retinal and Cerebellar Degeneration,
Produce polyQ-Expanded Protein and
Suffer Loss of Ataxin-7 Antisense ncRNA
Expression
(A) Confocal microscopy of retinal whole-mounts
from 11-month-old SCA7-CTCF-I-mut (mut) and
SCA7-CTCF-I-wt (wt) mice. Retinas immuno-
stained with antibodies against red/green cone
photoreceptors (red) and rod photoreceptors
(green) reveal dramatic loss of cone photorecep-
tors and more moderate, but nonetheless marked,
drop-out of rods in SCA7-CTCF-I-mut mice.
(B) Confocal microscopy of cerebellar sections
from 11-month-old SCA7-CTCF-I-mut (mut) and
SCA7-CTCF-I-wt (wt) mice. Sections immuno-
stained with antibodies against calbindin (green)
and ataxin-7 (magenta), and counterstained with
DAPI (blue) reveal marked Purkinje cell degener-
ation, loss, and protein aggregate formation
(arrows) in SCA7-CTCF-I-mut mice.
(C) Western blot analysis of cortex protein lysates
obtained from six week-old SCA7-CTCF-I-wt
mice, two independent lines of SCA7-CTCF-I-mut
mice, and a nontransgenic control (Ntg) immuno-
blotted with 1C2 antibody that detects polyQ-
expanded protein. Both lines of SCA7-CTCF-I-
mut mice display an �47 kDa band (arrow) that is absent from the Ntg and barely visible in the SCA7-CTCF-I-wt mouse. The asterisk indicates a nonspecific
background band that confirmed equivalent loading.
(D) Reciprocal expression of sense and antisense ataxin-7 RNAs in three month-old SCA7-CTCF-I-wt and SCA7-CTCF-I-mut mice. We performed strand-
specific RT-PCR, and quantified RNA expression relative to 18SRNA. Results for three sets of brain RNAs per line are shown, with ataxin-7 sense RNA expression
arbitrarily set to 1 for SCA7-CTCF-I-wt mice, and ataxin-7 antisense (SCAANT1) RNA expression arbitrarily set to 1 for SCA7-CTCF-I-mut mice. Comparable
results were obtained when we normalized to GAPDH, and absolute values indicated that ataxin-7 antisense transcript levels are �2.1-fold higher than ataxin-7
sense levels in SCA7-CTCF-I-wt mice (not shown). Assays were performed in triplicate, and error bars = SEM.
(E) In situ hybridization of SCAANT1 in threemonth-old SCA7 transgenic mice. Prominent signal in the Purkinje cell layer of SCA7-CTCF-I-wtmice (wt) is detected
with the ataxin-7 antisense riboprobe; however, minimal labeling in the Purkinje cell layer of SCA7-CTCF-I-mut mice (mut) is observed with this riboprobe.
See Figure S3.
Neuron
CTCF Promotes AS ncRNA to Control atx7 Expression
dystrophy retinal degeneration, characterized by dramatic loss
of cone photoreceptors and visual dysfunction (Ahn et al.,
2005; To et al., 1993). To determine if SCA7-CTCF-I-mut mice
recapitulate this phenotype, we immunostained retinal whole-
mounts from age-matched SCA7-CTCF-I-mut and SCA7-
CTCF-I-wt mice, and observed a marked drop-out of cone
photoreceptors in SCA7-CTCF-I-mut mice (Figure 3A). Electro-
retinogram testing corroborated this finding, as SCA7-CTCF-I-
mut mice went blind with a degradation of cone responses
ahead of rod responses (Figure S3). The visible ataxia phenotype
in affected SCA7-CTCF-I-mut mice led us to compare cerebellar
sections from age-matched SCA7-CTCF-I-mut mice and SCA7-
CTCF-I-wt mice. This analysis revealed dramatic Purkinje cell
degeneration, as well as ataxin-7 positive aggregates in Purkinje
cells in SCA7-CTCF-I-mut mice (Figure 3B). These findings
confirm that mutation of the 30 CTCF binding site, within a human
ataxin-7 minigene lacking the canonical ataxin-7 TSS at exon 1,
is sufficient to recapitulate the SCA7 phenotype in independent
lines of transgenic mice.
Recapitulation of the SCA7 phenotype in SCA7-CTCF-I-mut
mice, together with the observation of ataxin-7-positive inclu-
sions in cerebellar Purkinje cells, suggested that mutation of
the 30 CTCF binding site had resulted in the initiation of sense
transcription within the ataxin-7 minigene construct. To test
this hypothesis, we performed RT-PCR analysis on SCA7-
CTCF-I-mut mice and detected expression of the ataxin-7 first
coding exon in RNA samples from cerebellum and cortex (data
not shown). As SCA7 disease pathogenesis typically involves
production of a misfolded polyQ-expanded ataxin-7 protein,
we performed western blot analysis with the 1C2 antibody that
is specific for expanded polyQ tracts (Trottier et al., 1995) and
observed an �47 kDa protein in brain lysates from each SCA7-
CTCF-I-mut transgenic line (Figure 3C). We noted higher expres-
sion in the SCA7-CTCF-I-mut-(2) line, consistent with its more
severe phenotype. Low-level expression of the �47 kDa protein
was detected in SCA7-CTCF-I-wt mice (Figure 3C), and this
�47 kDa protein product corresponds to an open reading frame
starting at the initiator ATG codon in exon 3 and continuing
through exon 4 until the first nonsense codon in intron 4. The
production of a protein product and disease phenotype in the
SCA7-CTCF-I-mut mice is reminiscent of the R6/2 mouse model
of HD, in which a small fragment from the htt gene was intro-
duced into mice to model repeat instability, but also yielded
a truncated protein product resulting in a HD-like phenotype—
despite the fact that the construct lacked a 30 polyadenylationsite or characterized promoter (Mangiarini et al., 1996).
Neuron 70, 1071–1084, June 23, 2011 ª2011 Elsevier Inc. 1075
A
B
0
50
100
150
200
250
300
350
Cortex Cerebellum Striatum Lung Liver Kidney
Expr
essi
on re
lativ
e to
GA
PDH
Expanded
Normal
15Q
15Q
55Q
55Q
Antisense Sense
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0 15Q55Q
150Q
Atx
7 s
en
se e
xp
ress
ion
level
(rela
tive t
o G
AP
DH
)
SCA7 patient fibroblast samples
**
**
C
D
**
0
200
400
600
800
1000
1200
1400
1600
1800
2000
Cortex Cerebellum Striatum Lung Liver Kidney
Expr
essi
on re
lativ
e to
GA
PDH
*****
***
E
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0 59Q66Q
Atx
7 s
en
se e
xp
ress
ion
level
(rela
tive t
o G
AP
DH
)
SCA7 patient lymphocyte samples
**
**
2.2
2.6
2.410Q10Q
51Q
**
Figure 4. Reciprocal Ataxin-7 Sense and Antisense
Expression in Human Tissues and SCA7
(A) Results of qRT-PCR analysis of ataxin-7 sense tran-
script. Expression of ataxin-7 sense transcript was
significantly less in lung and kidney (p < 0.01 and p < 0.001
by ANOVA with post hoc Tukey test). Assays were per-
formed in triplicate, and error bars = SEM.
(B) Results of qRT-PCR analysis of ataxin-7 antisense
ncRNA SCAANT1. Expression of SCAANT1 was signifi-
cantly elevated in lung and kidney (p < 0.001 and p < 0.01
by ANOVA with post hoc Tukey test). Assays were per-
formed in triplicate, and error bars = SEM.
(C) RT-PCR analysis of SCA7 patient fibroblast RNAs.
RT-PCR amplification reveals the ataxin-7 antisense
ncRNA SCAANT1 from the normal allele in a control
fibroblast line (15Q) and SCA7 patient fibroblast line (55Q),
but not from the expanded allele for the SCA7 patient. RT-
PCR amplification of the ataxin-7 sense transcript,
however, indicates strong expression from the normal and
expanded allele in SCA7 fibroblasts.
(D) Quantification of ataxin-7 RNA levels in SCA7 patient
fibroblasts. Significant increases in ataxin-7 sense
expression were present in SCA7 patients carrying either
55Q or 150Q alleles (p < 0.01 by ANOVA). Experiments
were performed in triplicate, and error bars = SEM.
(E) Quantification of ataxin-7 RNA levels in SCA7 patient
lymphocytes. Significant increases in ataxin-7 sense
expression were present in SCA7 patients carrying 51Q,
59Q, or 66Q alleles (p < 0.01 by ANOVA). Experiments
were performed in triplicate, and error bars = SEM.
See Figure S4.
Neuron
CTCF Promotes AS ncRNA to Control atx7 Expression
Loss of SCAANT1 Expression Is Accompanied byDerepression of Sense TranscriptionOur findings indicate that mutation of the 30 CTCF binding site is
responsible for initiation of robust sense transcription in SCA7-
CTCF-mut-I mice, as SCA7-CTCF-I-wt mice carrying an
ataxin-7 genomic fragment with an intact 30 CTCF binding site
express low levels of ataxin-7 mRNA and protein. To determine
if the levels of ataxin-7 sense and antisense transcription within
the repeat region domain correlate in SCA7-CTCF-I-wt and
SCA7-CTCF-I-mut mice, we performed quantitative strand-
specific RT-PCR amplification, and detected ataxin-7 sense
and antisense transcripts in each line. We found that ataxin-7
sense transcript levels were elevated �370-fold in the brains of
SCA7-CTCF-I-mut mice compared to SCA7-CTCF-I-wt mice,
and this was accompanied by an �140-fold decrease in
SCAANT1 expression (Figure 3D). In situ hybridization analysis
confirmed robust expression of SCAANT1 in the cerebellum of
SCA7-CTCF-I-wt mice but did not detect strong SCAANT1
expression in SCA7-CTCF-I-mut mice (Figure 3E). In situ hybrid-
1076 Neuron 70, 1071–1084, June 23, 2011 ª2011 Elsevier Inc.
ization analysis indicated moderate to strong
expression of SCAANT1 in SCA7-CTCF-I-wt
mice throughout the brain (Figure S4A). Corre-
spondingly, in situ hybridization did not yield
evidence for much SCAANT1 expression in
the brain of SCA7-CTCF-I-mut mice (Fig-
ure S4B). Taken together, these findings show
that reduced SCAANT1 expression correlates
with increased P2A promoter activity, resulting in increased
sense expression of the ataxin-7 gene.
Ataxin-7 Sense Expression and SCAANT1 ExpressionAre Inversely Correlated in Human Tissues and SCA7PatientsOur studies of the SCA7-CTCF-I-wt and SCA7-CTCF-I-mutmice
suggested that expression of the ataxin-7 sense transcript
inversely correlates with expression of SCAANT1. To determine
if this reciprocal expression relationship exists in normal human
tissues, we performed qRT-PCR analysis on a panel of human
tissue RNAs.While we documented high levels of ataxin-7 sense
transcript in cortex, cerebellum, striatum, and liver, we found
much lower levels of ataxin-7 in lung and kidney (Figure 4A).
Interrogation of SCAANT1 expression levels revealed an oppo-
site pattern, as SCAANT1 was much higher in the lung and
kidney than in cortex, cerebellum, striatum, or liver (Figure 4B).
As the presence of the CAG repeat expansion decreased
SCAANT1 promoter activity in our luciferase reporter assays
Neuron
CTCF Promotes AS ncRNA to Control atx7 Expression
(Figure 2), we tested if the diametrically opposed expression of
ataxin-7 sense transcript and SCAANT1 might occur in the
context of SCA7 disease. To test this hypothesis, we performed
RT-PCR analysis of a SCA7 patient fibroblast cell line, and while
we could amplify both the normal and expanded repeat alleles
for the ataxin-7 sense transcript, we could not detect antisense
SCAANT1 transcript expression from the expanded 55Q allele
(Figure 4C). We then performed quantitative RT-PCR analysis
of ataxin-7 sense expression on fibroblasts obtained from two
SCA7 patients, one with a moderately sized disease repeat
(55Q), and one with a severely expanded repeat (150Q). With
increasing expansion size, we observed significantly increased
ataxin-7 sense transcript levels (Figure 4D), indicating that
expansion of the CAG repeat at the ataxin-7 locus yields
increased levels of ataxin-7 transcript in association with
reduced expression of SCAANT1. We also obtained a set of
peripheral blood samples from three additional SCA7 patients,
isolated RNA from their lymphocytes, and performed RT-PCR
analysis. Antisense SCAANT1 transcript expression could not
be detected from the expanded allele of the SCA7 samples,
and all three SCA7 patients exhibited significantly increased
ataxin-7 sense transcript levels (Figure 4E).
CTCF Regulates SCAANT1 Expression and Ataxin-7Alternative Sense TranscriptionAs mutation of the 30 CTCF binding site reduced the activity of
the SCAANT1 promoter while derepressing ataxin-7 sense
expression from promoter P2A, we hypothesized that CTCF
modulates ataxin-7 sense expression from this promoter by
driving the expression of SCAANT1. To test this hypothesis,
we validated two different CTCF shRNAs and derived a dual
CTCF knockdown vector. After subcloning the CTCF dual
shRNA knockdown fragment into a lentiviral construct with
a linked eGFP expression cassette, we infected human Y-79 reti-
noblastoma cells and isolated RNA from flow-sorted GFP-posi-
tive Y-79 cells. Real-time RT-PCR analysis confirmed CTCF
knockdown, and revealed a significant reduction in the expres-
sion level of SCAANT1 (Figure 5A). Significant reduction in
SCAANT1 expression was accompanied by a marked increase
in the ataxin-7 sense transcript from the P2A promoter, but not
from the previously defined ‘‘standard’’ ataxin-7 sense promoter,
located >40 kb 50 to the repeat region (Figure 5A). Although
direct, physiological comparison of the standard P1 promoter
and P2A promoter is complicated by the coexistence of
SCAANT1 transcription, analysis of ataxin-7 alternative sense
and standard sense expression at baseline in Y-79 cells revealed
only modestly (i.e., �2.5-fold) higher levels of the ataxin-7 stan-
dard sense transcript, consistent with the similar degree of
H3K4me3enrichment noted for their respective TSSs (FigureS1).
To examine the trans contribution of CTCF protein levels to the
regulation of ataxin-7 gene expression in the SCA7-CTCF-I-
mut mice, we generated CTCF heterozygous knock-out mice
(G.N.F. et al., unpublished data). ChIP analysis has indicated
that reduced CTCF occupancy at the mutated 30 CTCF binding
site occurs in the SCA7-CTCF-I-mut mice (Libby et al., 2008),
and this may account for the de-repression of ataxin-7 sense
expression from promoter P2A. To test if the effect of this cis
mutation could be compounded by reduction of CTCF expres-
sion in trans, we crossed SCA7-CTCF-I-mut mice with CTCF
heterozygous null mice, and compared the resulting SCA7-
CTCF-I-mut; CTCF+/� mice with their SCA7-CTCF-I-mut;
CTCF+/+ littermates. We confirmed reduced dosage of CTCF ex-
pression in the SCA7-CTCF-I-mut; CTCF+/� mice, and observed
significantly reduced expression of the antisense SCAANT1 tran-
script (Figure 5B). This was accompanied by increased ataxin-7
sense expression (Figure 5B), which yielded a more rapidly
progressive retinal phenotype in SCA7-CTCF-I-mut; CTCF+/�
mice (Figures 5C–5E). The worsened phenotype was also
reflected by a significantly shortened life span (Figure 5F). Hence,
decreased expression of CTCF agonized the SCA7 phenotype in
SCA7-CTCF-I-mut mice by further de-repressing ataxin-7 P2A
promoter activity. As cohesin may play a role in CTCF insulator
formation (Parelho et al., 2008; Stedman et al., 2008; Wendt
et al., 2008), we performed ChIP analysis for two cohesin
subunits in the SCA7 transgenic mice, and observed reduced
occupancy of SMC1 and SMC3 at the 30 CTCF binding site in
the cerebellum of SCA7-CTCF-I-mut mice (Figure S5), suggest-
ing that cohesin may also participate in CTCF-mediated tran-
scription regulation at the ataxin-7 locus.
SCAANT1 Mediates Repression of the CorrespondingAtaxin-7 Sense Transcript in cis
CTCF binding regulates sense and antisense transcription at the
ataxin-7 locus, and expression levels of the ataxin-7 sense tran-
script and antisense SCAANT1 message are inversely corre-
lated. Hence, a key question is whether SCAANT1 transcription
is coincident, or required for derepression of ataxin-7 sense
promoter P2A. To determine if SCAANT1 transcription is
necessary for the regulation of ataxin-7 sense expression, and
to distinguish between a cis or trans regulatory mechanism, we
developed a CMV-SCAANT1 expression construct. We then
cotransfected astrocytes with a highly active ataxin-7 genomic
fragment—luciferase reporter construct and the CMV-SCAANT1
expression construct and tested if enforced expression of
SCAANT1 would downregulate ataxin-7 sense P2A promoter
activity, but we observed no effect (Figure 6A). This in vitro
experiment was paralleled by an in vivo study in which we
crossed SCA7-CTCF-I-mut mice with SCA7-CTCF-I-wt mice,
reasoning that the dramatically elevated levels of SCAANT1 anti-
sense RNA in SCA7-CTCF-I-wt mice would reduce ataxin-7
sense expression in SCA7-CTCF-I-mut mice, if SCAANT1
regulation of ataxin-7 sense expression is occurring in trans.
However, behavioral analysis revealed further worsening in
rotarod performance in SCA7-CTCF-mut-I; SCA7-CTCF-I-wt
mice in comparison to SCA7-CTCF-I-mut mice (Figure S6).
Furthermore, SCA7-CTCF-I-mut; SCA7-CTCF-I-wt bigenic
mice displayedmore severe Purkinje cell degeneration than their
singly transgenic SCA7-CTCF-I-mut littermates (Figures 6B–6D).
Thus, greater expression of SCAANT1 in SCA7-CTCF-I-mut;
SCA7-CTCF-I-wt mice did not reduce ataxin-7 sense transcrip-
tion; instead, low levels of ataxin-7 sense protein product from
the SCA7-CTCF-I-wt mice (Figure 3C) enhanced the SCA7
phenotype in bigenic SCA7-CTCF-I-mut; SCA7-CTCF-I-wt
mice.
To test if SCAANT1 transcription regulates promoter P2A
in cis, we engineered ataxin-7 P2A-exon 3(CAG10)-exon 4
Neuron 70, 1071–1084, June 23, 2011 ª2011 Elsevier Inc. 1077
A
B
F
Weeks
Su
rviv
al ra
te
SCA7-CTCF-I-mut;CTCF +/-
SCA7-CTCF-I-mut;CTCF +/+
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
CTCF Antisense Alt. sense Standard sense
ControlLenti-CTCF-shRNA
Exp
ress
ion
leve
l (r
elat
ive
to 1
8S R
NA
)
****
*
SCA7-CTCF-I-mut; CTCF +/-
SCA7-CTCF-I-mut; CTCF +/+
CTCF Antisense
Exp
ress
ion
leve
l (r
elat
ive
to A
CT
B)
0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
Exp
ression
level (relative to
GA
PD
H)
* *
0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
*
Sense
Figure 5. CTCF Regulates Promoter Usage
for Ataxin-7 Sense and Antisense Tran-
scripts
(A) CTCF knockdown in Y-79 retinoblastoma cells
leads to decreased ataxin-7 antisense RNA
expression and increased exon 2A transcript
expression. Human Y-79 retinoblastoma cells
were infected with a lentiviral expression con-
struct containing two tandem shRNAs targeting
CTCF (lenti-CTCF-shRNA) or with empty vector
(Control), and expression levels were measured by
qPCR. Marked reductions in expression of CTCF
and ataxin-7 antisense SCAANT1 transcripts were
noted (p < 0.01 by ANOVA) andwere accompanied
by a significant increase in the ataxin-7 alternative
sense transcript (p < 0.05 by ANOVA). Knockdown
of CTCF did not affect expression of the ataxin-7
transcript initiated from a human-specific ataxin-7
promoter located >40 kb 50 to the alternative
promoter. Experiments were performed in tripli-
cate, the expression level for each transcript upon
CTCF knockdown is normalized to its respective
control, and error bars = SEM.
(B) Decreased CTCF dosage reduces SCAANT1
expression in SCA7-CTCF-I-mut mice. We
crossed SCA7-CTCF-I-mut mice with CTCF+/�
mice and derived SCA7-CTCF-I-mut mice on
a CTCF heterozygous null background (CTCF+/�)or a wild-type background (CTCF+/+). We per-
formed qRT-PCR analysis of CTCF in the brain of
three-month-old mice and confirmed reduced
expression of CTCF in SCA7-CTCF-I-mut;
CTCF+/� mice (p < .0.05 by ANOVA). When we
measured brain expression of SCAANT1 (Anti-
sense) transcript and ataxin-7 sense transcript by
qRT-PCR, we noted a marked reduction in anti-
sense expression and a significant increase in
sense expression in SCA7-CTCF-I-mut; CTCF+/�
mice (p < .0.05 by ANOVA). Experiments were
performed in triplicate, the expression level for
each transcript is normalized to the SCA7-CTCF-I-
mut; CTCF+/+ level, and error bars = SEM.
(C–E) Reduced CTCF dosage enhances SCA7
retinal degeneration in SCA7-CTCF-I-mut mice.
Confocal microscopy of retinal sections from six-
month-old nontransgenic controls (C), SCA7-
CTCF-I-mut; CTCF+/� (D), and SCA7-CTCF-I-mut;
CTCF+/+ (E), reveals more severe cone photore-
ceptor dropout in SCA7-CTCF-I-mut mice on
a CTCF heterozygous null background. Red/green pigment antibody = red; propidium iodide = cyan. OS = outer segments; ONL = outer nuclear layer; INL = inner
nuclear layer; GCL = ganglion cell layer. Scale bar corresponds to 50 mM.
(F) Kaplan-Meier survival analysis of SCA7-CTCF-I-mut; CTCF+/�mice.We recorded lifespan for cohorts ofSCA7-CTCF-I-mut; CTCF+/�mice (n = 28) andSCA7-
CTCF-I-mut; CTCF+/+ mice (n = 40) and observed a significant reduction in life span for SCA7-CTCF-I-mut mice on a CTCF heterozygous null background
(p < 0.05 by log-rank test).
See Figure S5.
Neuron
CTCF Promotes AS ncRNA to Control atx7 Expression
genomic fragment constructs with a 30 IRES-luciferase (luc) in
sense orientation, and a Renilla luciferase (Rluc) in antisense
orientation (Figure 6E). We replaced the antisense SCAANT1
promoter with a tet-regulatable element (TRE) to yield the
‘‘TRE-only’’ ataxin-7 genomic fragment construct, and then
created a second version by cloning a polyA transcription termi-
nation signal (‘‘polyA trap’’) in antisense orientation into exon 3
(Figure 6E). To confirm the integrity of the polyA trap, we trans-
fected astrocytes with either the TRE-only or TRE-polyA-trap
1078 Neuron 70, 1071–1084, June 23, 2011 ª2011 Elsevier Inc.
ataxin-7 vector, induced with doxycycline and observed marked
reduction of Rluc/luc for the TRE-polyA-trap ataxin-7 vector
(Figure 6F). When we measured ataxin-7 expression by qRT-
PCR, we observed significant derepression of ataxin-7 sense
expression in TRE-polyA-trap ataxin-7 transfected cells (Fig-
ure 6G). Hence, premature termination of SCAANT1 transcrip-
tion released repression of ataxin-7 P2A promoter activity,
indicating that SCAANT1 regulates ataxin-7 sense expression
in cis by convergent transcription.
Figure 6. The Ataxin-7 Antisense ncRNA,
SCAANT1, Regulates Ataxin-7 Alternative
Promoter Activity in cis, but Not in trans
(A) Effect of enforced expression of SCAANT1
upon transactivation competence of ataxin-7
promoter P2A. The human ataxin-7 genomic
fragment luciferase expression construct (1F; see
Figure 1A) was transfected into primary astrocytes,
along with Renilla luciferase, and a CMV expres-
sion construct containing SCAANT1 (CMV-
SCAANT1). Relative luminescence unit (RLU)
measurements were obtained. EV corresponds to
the empty luciferase vector, CMV-empty refers to
the empty CMV expression construct, all experi-
ments were performed in quadruplicate, and errors
bars = SEM.
(B–D) Effect of overexpression of SCAANT1 upon
the de-repressed ataxin-7 alternative promoter
in vivo. We crossed the SCA7-CTCF-I-mut mice
with SCA7-CTCF-I-wt mice, and compared cere-
bellar histopathology between littermates of the
following genotypes: SCA7-CTCF-I-wt (B), SCA7-
CTCF-I-mut (C), and SCA7-CTCF-I-mut; SCA7-
CTCF-I-wt (D). Here, we see the results of confocal
microscopy of cerebellar sections immunostained
with antibodies against calbindin (green) and
ataxin-7 (magenta), and counterstained with DAPI
(blue) for mice at 7 months of age. Strong calbindin
immunoreactivity is observed in SCA7-CTCF-I-wt
mice (B), while an obvious reduction in calbindin
immunoreactivity is apparent in the SCA7-CTCF-I-
mut mice (C). Calbindin immunoreactivity is further
reduced in SCA7-CTCF-I-mut; SCA7-CTCF-I-wt
bigenic mice, and this is accompanied by an
increased accumulation of aggregated ataxin-7
protein (magenta puncta) (D), indicating that
enforced in vivo expression of SCAANT1 does
not promote degradation of the complementary
ataxin-7 sense transcript. Rather, crossing with
SCA7-CTCF-I-wt mice actually worsens the
phenotype, since SCA7-CTCF-I-wt mice produce
modest levels of polyQ-expanded ataxin-7 protein
(see Figure 3C).
(E) Diagram of the ataxin-7 antisense premature
termination construct. We replaced the ataxin-7
antisense ncRNA promoter with a tet-responsive
element—minimal CMV promoter (TRE) and in-
serted Renilla luciferase (R-luc) in antisense orientation to quantify the activity of the TRE promoter. The IRES-luc segment is cloned in sense orientation to permit
normalization. The poly-A transcription termination cassette was cloned into exon 3 in antisense orientation (asterisk), and we prepared two versions of this
construct: either without the poly-A cassette (TRE-only) or with the poly-A cassette (TRE-poly-A-trap).
(F) Validation of poly-A transcription termination activity. The TRE-only construct or the TRE-poly-A-trap construct was transfected into primary astrocytes, along
with a CMV-tet-activator expression construct, and Renilla luciferase (R-luc) and luciferase (luc) activities were measured. A marked reduction in the R-luc / luc
ratio in the TRE-poly-A-trap construct was noted (p < 0.01, ANOVA).
(G) Convergent transcription is required for repression of ataxin-7 alternative promoter by the antisense ncRNA. The TRE-only construct or the TRE-poly-A-trap
construct was transfected into primary astrocytes, along with tet-activator, and ataxin-7 sense transcript levels were measured by qPCR. Ataxin-7 sense
expression was dramatically increased in cells transfected with the TRE-poly-A-trap construct (p < 0.01, ANOVA).
See Figure S6.
Neuron
CTCF Promotes AS ncRNA to Control atx7 Expression
Ataxin-7 Antisense Expression Yields RepressiveChromatin Modifications in Transgenic MiceSilencing of ataxin-7 sense P2A promoter activity by convergent
transcription of the SCAANT1 RNA raised a number of questions
as to the mechanism of repression. We hypothesized that one
possibility might be chromatin-dependent gene silencing and
proceeded to evaluate the status of covalent histone modifica-
tions known to correlate with transcription activation and repres-
sion in SCA7 CTCF-I-mut and SCA7-CTCF-I-wt mice. To do this,
we performed ChIP analysis of histone marks, and interrogated
covalent histone modification status at amplicons spanning the
ataxin-7 repeat region, including the transcription domains and
start sites for the sense P2A promoter and SCAANT1 (Figure 7A).
Quantitative PCR analysis revealed highly enriched levels of the
Neuron 70, 1071–1084, June 23, 2011 ª2011 Elsevier Inc. 1079
A
B C DA E
ATG
(CAG)n
EXON 2A EXON 3 EXON 4
P2A
50 bp
> 38 kb to exon 5
>13 kb to exon 2 AAGTAAA
CTCF-II CTCF-I
AAAAA
6
B
0
4
1
Amplicon ‘A’ Amplicon ‘B’ Amplicon ‘C’
2
3
5
Amplicon ‘D’ Amplicon ‘E’
IP /
inp
ut:
H3K
27m
e3
(no
rmal
ized
to
c-M
yc C
TC
F s
ite)
***
**
**
0
4
1
Amplicon ‘A’ Amplicon ‘B’ Amplicon ‘C’
2
3
5
Amplicon ‘D’ Amplicon ‘E’
IP /
inp
ut:
H3K
9/14
ac
(no
rmal
ized
to
c-M
yc C
TC
F s
ite) 6
*
*C
SCA7-CTCF-I-wt
SCA7-CTCF-I-mut
SCA7-CTCF-I-wt
SCA7-CTCF-I-mut
Figure 7. ChIP Analysis of Histone Modification Status at the
Ataxin-7 Alternative Promoter Region in SCA7 Transgenic Mice
(A) Diagram of the genomic region of the human ataxin-7 gene, where the
alternative promoter (P2A), first coding exon, CTCF binding sites, and anti-
sense ncRNA, SCAANT1, are located. Boxes represent exons, while solid lines
correspond to introns. The positions of the five amplicons employed in the
ChIP analysis are indicated.
(B and C) Results of ChIP analysis of histone marks in the cerebellum of SCA7
transgenic mice. ChIP was performed with antibodies against a repressive
mark (H3K27me3) and activated transcription mark (H3K9/14Ac) on cerebellar
DNAs isolated from six-month old SCA7-CTCF-I-wt and SCA7-CTCF-I-mut
mice. (B) In SCA7-CTCF-I-wt mice that express minimal amounts of ataxin-7
sense RNA from P2A, there was significant enrichment of H3K27me3 ex-
tending from the alternative promoter region into intron 3 of the ataxin-7 gene
(p < 0.01 to 0.001, ANOVA). (C) This pattern was reversed in SCA7-CTCF-I-mut
mice that display robust activation of the P2A promoter, as significant
enrichment of the H3K9/14Ac mark was present in the ataxin-7 sense P2A
region and initial transcript region (p < 0.05, ANOVA). Results represent three
independent experiments quantifying histone antibody ChIP normalized to the
c-Myc promoter region. Error bars = SEM.
Neuron
CTCF Promotes AS ncRNA to Control atx7 Expression
repressive H3K27me3 mark at amplicons 50 and 30 to the P2A
TSS and also showed low levels of the active H3K9/14ac mark
in SCA7-CTCF-I-wt mice (Figures 7B and 7C). However, ChIP
analysis revealed a reversal of this pattern in the SCA7-CTCF-
I-mut mice, as amplicons adjacent to promoter P2A exhibited
very low levels of H3K27me3 associated with elevated levels of
H3K9/14ac (Figures 7B and 7C), consistent with active transcrip-
tion at promoter P2A. Hence, our survey of covalent histone
modifications in the ataxin-7 mini-gene mice supported a role
for chromatin-dependent gene silencing by SCAANT1.
1080 Neuron 70, 1071–1084, June 23, 2011 ª2011 Elsevier Inc.
DISCUSSION
Bidirectional transcription at repeat loci is emerging as an impor-
tant theme in repeat expansion diseases, including myotonic
dystrophy 1 (DM1), spinocerebellar ataxia type 8 (SCA8), the
fragile X syndrome of mental retardation, Friedreich’s ataxia
(FRDA), Huntington’s disease (HD), and Huntington’s disease-
like 2 (HDL2) (Mirkin, 2007). At the same time, a role for CTCF
in regulating chromatin structure and transcription at such repeat
disease loci is being recognized (La Spada and Taylor, 2010). At
the SCA7 locus, the significance of CTCF for regulating repeat
instability was recently demonstrated, and shown to involve
epigenetic processes (Libby et al., 2008). In this study, we
examined the ataxin-7 repeat region where the CTCF binding
sites reside, and discovered that ataxin-7 gene expression is
governed by an antisense ncRNA transcript. This transcript,
which we named ‘‘SCAANT1,’’ appears to regulate a previously
unrecognized ataxin-7 sense promoter by convergent transcrip-
tion that overlaps the ataxin-7 repeat and the adjacent P2A
sense promoter. Our studies thus reveal a pathway for regulating
ataxin-7 gene expression at this promoter via an antisense RNA
and link CTCF transactivation of SCAANT1with repression of the
convergently transcribed sense domain (Figure 8).
Repeat tracts can greatly influence chromatin structure, espe-
cially if they are CG-rich (Wang et al., 1996). The mapping of
CTCF binding sites in close proximity to such repeats suggested
the need to insulate surrounding DNA from the potentially
untoward effects of repeat-induced changes upon chromatin
structure. CTCF is a multivalent transcription regulatory factor,
known to possess enhancer-blocking activity (Phillips and
Corces, 2009). CTCF may also prevent inactivation of gene
expression, as CTCF can restrict the spread of X-inactivation,
thereby preserving the transcriptional activity of ‘‘escape’’ genes
(Filippova et al., 2005). Previous studies of repeat disease loci
have shown that CTCF can prevent epigenetic changes associ-
ated with heterochromatin formation and gene inactivation by
constraining antisense transcription (Cho et al., 2005; De Biase
et al., 2009; Filippova et al., 2005). We evaluated the role of
CTCF in regulating ataxin-7 gene expression from an adjacent
alternative promoter (P2A) by introducing two different ataxin-
7 minigenes into mice. These minigenes are �13.5 kb ataxin-7
genomic fragments that contain the P2A promoter, the
SCAANT1 domain, the ataxin-7 start site of translation, and
the CAG repeat tract. The two minigenes were identical except
for the presence of a substitution mutation in the ‘‘SCA7-
CTCF-I-mut’’ construct at the 30 CTCF binding site. Importantly,
this substitution mutation had been shown to completely abro-
gate CTCF binding (Libby et al., 2008). The SCA7-CTCF-I-wt
construct yielded four independent lines of transgenic mice
that did not develop a phenotype, despite possessing a
CAG92 repeat tract in the fully intact ataxin-7 minigene. Instead,
two independent lines of SCA7-CTCF-I-mut mice developed
a SCA7-like phenotype, characterized by cone-rod dystrophy
retinal degeneration and cerebellar atrophy. Further studies
indicated that loss of CTCF binding results in dramatically
reduced expression of SCAANT1 in association with high-level
ataxin-7 expression from the newly discovered alternative sense
promoter. Our findings thus reveal that CTCF does regulate
(CAG)nSense
Promoter
AntisensePromoter
Ataxin-7
CTCF CTCF
(CAG)nSense
Promoter
AntisensePromoter
Ataxin-7
(CAG)nSense
Promoter
AntisensePromoter
Ataxin-7
CTCF
(CAG)nSense
Promoter
Ataxin-7
AntisensePromoter
(CAG)nSense
Promoter
Ataxin-7 AntisensePromoter
Atx7Atx7 Atx7Atx7
Atx7Atx7Atx7Atx7
Atx7Atx7
Atx7Atx7
Atx7Atx7Atx7Atx7
A B
CTCF CTCF
(CAG)nSense
Promoter
AntisensePromoter
Ataxin-7
CTCF CTCF
Atx7Atx7
Atx7Atx7
Figure 8. Model for Transcription Regula-
tion at the Ataxin-7 Locus
(A) CTCF-induced, SCAANT1-mediated, Dicer-
dependent repression of ataxin-7 sense tran-
scription. The ataxin-7 CAG repeat is flanked by
CTCF binding sites. When the CTCF binding sites
are unoccupied, the adjacent ataxin-7 promoter is
active, and covalent histone modification at the
promoter region is consistent with active tran-
scription (light green dots). When CTCF levels
increase and CTCF binds, the antisense promoter
for the noncoding RNA SCAANT1 is activated,
yielding convergent transcription. This is accom-
panied by a transition from an active chromatin
conformation to a repressed chromatin state (red
dots), silencing ataxin-7 expression from sense
promoter P2A.
(B) CAG repeat expansion or CTCF reduction can
de-repress ataxin-7 sense expression. Under
conditions of modest CTCF expression, the
ataxin-7 sense promoter is moderately active (light
green dots). However, if the CAG repeat is
expanded, or CTCF levels dwindle, the flanking
CTCF binding sites become unoccupied, resulting
in reduced activity of the SCAANT1 promoter.
Under these conditions, the sense promoter is fully
derepressed, with a chromatin conformation
consistent with highly active transcription (green
dots), yielding robust expression of ataxin-7.
Neuron
CTCF Promotes AS ncRNA to Control atx7 Expression
ataxin-7 gene expression; however, instead of preventing tran-
scription repression, CTCF supports it. Furthermore, rather
than restricting antisense expression, CTCF promotes it.
Surveys of mammalian transcriptomes are uncovering
tremendous numbers and varieties of noncoding RNAs, and
the production of antisense transcripts appears to be a pervasive
feature of the human and mouse transcriptomes (He et al., 2008;
Kapranov et al., 2007; Okazaki et al., 2002). Whenwe discovered
that SCAANT1 expression levels inversely correlate with ataxin-7
sense expression in both SCA7 transgenic mice and human
tissues, we considered the possibility that SCAANT1 might be
regulating the expression of its sense counterpart, as reciprocal
expression of sense and antisense transcripts has been reported
for a number of human andmouse genes (Katayama et al., 2005).
Indeed, at the human p15 locus, gene silencing of sense expres-
sion by an antisense RNA has been documented and can be
achieved by enforced expression of the antisense transcript
(Yu et al., 2008). We tested if SCAANT1 expression in trans can
downregulate ataxin-7 alternative sense promoter activity in
luciferase reporter assay experiments and by crossing SCA7-
CTCF-I-mut mice with SCA7-CTCF-I-wt mice, as the latter
exhibit high-level SCAANT1 expression. However, SCAANT1
transcript elevation had no effect upon ataxin-7 alternative sense
expression in vitro or in vivo. Studies of antisense transcripts in
mice and humans, as well as other eukaryotes such as yeast,
have revealed evidence for inhibition of transcription by virtue
of actual transcription interference, when RNA polymerases
moving in opposite directions collide with one another (Osato
et al., 2007; Shearwin et al., 2005). To test if SCAANT1 regulates
sense expression in cis, we engineered an ataxin-7 genomic
fragment construct with a transcription terminator positioned in
the antisense orientation, and placed antisense transcription
under the control of an inducible promoter. After validating the
efficiency of the transcription terminator, wemeasured the effect
of premature transcription termination upon SCAANT1’s ability
to repress ataxin-7 sense expression, and we noted a dramatic
derepression of sense transcription, when antisense transcrip-
tion was prematurely terminated. These results confirmed that
SCAANT1 is directly regulating ataxin-7 alternative sense
expression and demonstrated that the regulation occurs in cis.
One obvious question, based upon our results is: Why does
the cell need such a complicated pathway for adjusting ataxin-
7 expression? Ataxin-7 is a core, and likely essential, component
of different ubiquitously expressed transcription coactivator
complexes (Helmlinger et al., 2004; Palhan et al., 2005; Zhao
et al., 2008). As deletion of the ataxin-7 ortholog Sgf73 eliminates
Ubp8-mediated histone deubiquitination in yeast (Kohler et al.,
2008), and knockdown of ataxin-7 results in disassembly of the
STAGA complex in mammals (Palhan et al., 2005), tight regula-
tion of ataxin-7 expression could be a mechanism for controlling
the activity of these coactivators. CTCF is a master regulator of
transcription, and its expression cannot be significantly adjusted
without killing the cell. However, minor changes in CTCF levels,
or binding activity, could have a dramatic impact upon the tran-
scriptional activity of the cell through its regulation of ataxin-7
expression, since ataxin-7 expression alterations would be
amplified by affecting the stability and function of entire co-acti-
vator complexes. Thus, CTCF control of ataxin-7 levels could
serve as a rheostat for setting global transcription activity status
for the cell.
Another important implication of our work is its relevance to
SCA7 disease pathogenesis and repeat disease biology. As we
Neuron 70, 1071–1084, June 23, 2011 ª2011 Elsevier Inc. 1081
Neuron
CTCF Promotes AS ncRNA to Control atx7 Expression
have shown, expansion of the ataxin-7 CAG repeat tract reduces
SCAANT1 promoter activity, resulting in minimally detectable
levels of SCAANT1 from the expanded allele in SCA7 patient
fibroblasts. This reduction in SCAANT1 expression derepresses
the ataxin-7 alternative promoter, and significantly boosts the
level of ataxin-7, creating a feed forward effect that agonizes
the SCA7 disease pathway by favoring increased production of
mutant ataxin-7 protein (Figure 8). Although we cannot exclude
a role for altered transcript stability in this process, a survey of
histone posttranslational modifications in our SCA7 transgenic
mice revealed repressive chromatin modifications in the alterna-
tive promoter when SCAANT1 transcription is robust, indicating
that transcriptional activity is likely more important than tran-
script stability in controlling ataxin-7 sense expression. As bidi-
rectional transcription in association with CTCF binding is
emerging as a common feature of repeat disease loci, our find-
ings could be applicable at other repeat disease loci, including
loci that have not been carefully screened for antisense tran-
scripts. Furthermore, the existence of regulatory bidirectional
transcription may offer an entry point for therapeutically
modulating ataxin-7 expression at the RNA level. In addition to
providing a window into how gene transcription is regulated,
convergent transcription may dictate which repeat loci are
subject to dramatic genetic instability, as epigenetic modifica-
tions and chromatin alterations likely lie at the heart of the insta-
bility process. Hence, regulatory bidirectional transcription at
repeat loci, coupled with CTCF-mediated epigenetic processes
in the germline, may explain why certain repeats exhibit dramatic
parent-of-origin instability, while other repeats are relatively
stable and not subject to parent-of-origin effects (La Spada,
1997; Pearson et al., 2005). Understanding the roles and regula-
tion of convergent, bidirectional transcription at repeat loci
should provide valuable clues as to how global transcriptional
processes and epigenetic pathways work.
EXPERIMENTAL PROCEDURES
Characterization of Transgenic Mice
The generation of the SCA7 transgenic constructs, and the production of the
SCA7-CTCF-I-wt and SCA7-CTCF-I-mut mice have been previously
described (Libby et al., 2008). ERG analysis, immunostaining analysis, RT-
PCR studies, western blot analysis, and behavioral studies were performed
as previously described (Garden et al., 2002; La Spada et al., 2001). All exper-
iments were approved by the University of Washington IACUC and UCSD
IACUC. Please see Supplemental Information for detailed protocols for chro-
matin immunoprecipitation and in situ hybridization.
RNA Isolation and Analysis
Cell line, patient, and mouse tissue RNAs were isolated using Trizol and
treated 2–3 times with DNase I (NEB), then precipitated for analysis. The integ-
rity of RNA was determined by nested RT-PCR of cDNA generated with or
without reverse transcriptase, using primers to at least 3 distinct genomic
regions. For RT-PCR, cDNA was generated using gene-specific primers,
which had a linker (LK) sequence, LK 50-CGACTGGAGCACGAGGACA
CTGA-30 attached to the 50 end. Otherwise cDNA was generated with random
decamers (Ambion). cDNAwas generated with 1–3 mg of RNA and Superscript
III (Invitrogen) at 50�C. PCR amplification was performed using two gene
specific primers or with a primer to the Linker sequence and a gene-specific
reverse primer for strand-specific analysis (35 cycles at 94�C for 30 s, 55�Cfor 30 s, and 68�C for 1 min). The PCR products were cloned into the pCR4-
TOPO vector (Invitrogen) and sequenced. Strand-specific RT-PCRwas essen-
1082 Neuron 70, 1071–1084, June 23, 2011 ª2011 Elsevier Inc.
tially performed as previously described (Ladd et al., 2007). Please see the
Supplemental Information for details of construct generation.
Statistical Analysis
All data were prepared for analysis with standard spread sheet software
(Microsoft Excel). Statistical analysis was done using Microsoft Excel, Prism
4.0 (Graph Pad), or the VassarStats website (http://faculty.vassar.edu/lowry/
vassarstats.html). For ANOVA, if statistical significance (p < 0.05) was
achieved, we performed posttest analysis to account for multiple compari-
sons. The level of significance (alpha) was always set at 0.05.
SUPPLEMENTAL INFORMATION
Supplemental Information includes six figures, one movie, and Supplemental
Experimental Procedures and can be found with this article online at doi:
10.1016/j.neuron.2011.05.027.
ACKNOWLEDGMENTS
The authors wish to thank A. Smith, S. Baccam, K. Takushi, J. Huang, and
D. Possin for outstanding technical assistance, and B. Ren for critical reading
of the manuscript. This work was supported by funding from the NIH
(GM59356, EY14061, and ARRA award GM59356-09-S1 to A.R.L., and
EY01730 [Vision Research Core] to UWMC). S.M.S. is supported by the Allen
Institute for Brain Science, founded by Paul G. Allen and Jody Patton.
Accepted: May 10, 2011
Published: June 22, 2011
REFERENCES
Ahn, J.K., Seo, J.M., Chung, H., and Yu, H.G. (2005). Anatomical and func-
tional characteristics in atrophic maculopathy associated with spinocerebellar
ataxia type 7. Am. J. Ophthalmol. 139, 923–925.
Batra, R., Charizanis, K., and Swanson, M.S. (2010). Partners in crime: bidirec-
tional transcription in unstable microsatellite disease. Hum. Mol. Genet. 19
(R1), R77–R82.
Chen, H.K., Fernandez-Funez, P., Acevedo, S.F., Lam, Y.C., Kaytor, M.D.,
Fernandez, M.H., Aitken, A., Skoulakis, E.M., Orr, H.T., Botas, J., and
Zoghbi, H.Y. (2003). Interaction of Akt-phosphorylated ataxin-1 with 14-3-3
mediates neurodegeneration in spinocerebellar ataxia type 1. Cell 113,
457–468.
Chen, S., Peng, G.H., Wang, X., Smith, A.C., Grote, S.K., Sopher, B.L., and
La Spada, A.R. (2004). Interference of Crx-dependent transcription by ataxin-7
involves interaction between the glutamine regions and requires the ataxin-7
carboxy-terminal region for nuclear localization. Hum. Mol. Genet. 13, 53–67.
Cho, D.H., Thienes, C.P., Mahoney, S.E., Analau, E., Filippova, G.N., and
Tapscott, S.J. (2005). Antisense transcription and heterochromatin at the
DM1 CTG repeats are constrained by CTCF. Mol. Cell 20, 483–489.
David, G., Abbas, N., Stevanin, G., Durr, A., Yvert, G., Cancel, G., Weber, C.,
Imbert, G., Saudou, F., Antoniou, E., et al. (1997). Cloning of the SCA7 gene
reveals a highly unstable CAG repeat expansion. Nat. Genet. 17, 65–70.
Davuluri, R.V., Grosse, I., and Zhang, M.Q. (2001). Computational identifica-
tion of promoters and first exons in the human genome. Nat. Genet. 29,
412–417.
De Biase, I., Chutake, Y.K., Rindler, P.M., and Bidichandani, S.I. (2009).
Epigenetic silencing in Friedreich ataxia is associated with depletion of
CTCF (CCCTC-binding factor) and antisense transcription. PLoS ONE 4,
e7914. 10.1371/journal.pone.0007914.
Filippova, G.N., Thienes, C.P., Penn, B.H., Cho, D.H., Hu, Y.J., Moore, J.M.,
Klesert, T.R., Lobanenkov, V.V., and Tapscott, S.J. (2001). CTCF-binding sites
flank CTG/CAG repeats and form a methylation-sensitive insulator at the DM1
locus. Nat. Genet. 28, 335–343.
Filippova, G.N., Cheng, M.K., Moore, J.M., Truong, J.P., Hu, Y.J., Nguyen,
D.K., Tsuchiya, K.D., and Disteche, C.M. (2005). Boundaries between
Neuron
CTCF Promotes AS ncRNA to Control atx7 Expression
chromosomal domains of X inactivation and escape bind CTCF and lack CpG
methylation during early development. Dev. Cell 8, 31–42.
Garden, G.A., Libby, R.T., Fu, Y.H., Kinoshita, Y., Huang, J., Possin, D.E.,
Smith, A.C., Martinez, R.A., Fine, G.C., Grote, S.K., et al. (2002).
Polyglutamine-expanded ataxin-7 promotes non-cell-autonomous purkinje
cell degeneration and displays proteolytic cleavage in ataxic transgenic
mice. J. Neurosci. 22, 4897–4905.
He, Y., Vogelstein, B., Velculescu, V.E., Papadopoulos, N., and Kinzler, K.W.
(2008). The antisense transcriptomes of human cells. Science 322, 1855–1857.
Helmlinger, D., Hardy, S., Sasorith, S., Klein, F., Robert, F., Weber, C., Miguet,
L., Potier, N., Van-Dorsselaer, A.,Wurtz, J.M., et al. (2004). Ataxin-7 is a subunit
of GCN5 histone acetyltransferase-containing complexes. Hum. Mol. Genet.
13, 1257–1265.
Helmlinger, D., Hardy, S., Abou-Sleymane, G., Eberlin, A., Bowman, A.B.,
Gansmuller, A., Picaud, S., Zoghbi, H.Y., Trottier, Y., Tora, L., and Devys, D.
(2006). Glutamine-expanded ataxin-7 alters TFTC/STAGA recruitment and
chromatin structure leading to photoreceptor dysfunction. PLoS Biol. 4, e67.
10.1371/journal.pbio.0040067.
Kapranov, P., Cheng, J., Dike, S., Nix, D.A., Duttagupta, R., Willingham, A.T.,
Stadler, P.F., Hertel, J., Hackermuller, J., Hofacker, I.L., et al. (2007). RNA
maps reveal new RNA classes and a possible function for pervasive transcrip-
tion. Science 316, 1484–1488.
Katayama, S., Tomaru, Y., Kasukawa, T., Waki, K., Nakanishi, M., Nakamura,
M., Nishida, H., Yap, C.C., Suzuki, M., Kawai, J., et al; RIKEN Genome
Exploration Research Group; Genome Science Group (Genome Network
Project Core Group); FANTOM Consortium. (2005). Antisense transcription
in the mammalian transcriptome. Science 309, 1564–1566.
Kohler, A., Schneider, M., Cabal, G.G., Nehrbass, U., andHurt, E. (2008). Yeast
Ataxin-7 links histone deubiquitination with gene gating and mRNA export.
Nat. Cell Biol. 10, 707–715.
La Spada, A.R. (1997). Trinucleotide repeat instability: genetic features and
molecular mechanisms. Brain Pathol. 7, 943–963.
La Spada, A.R., and Taylor, J.P. (2003). Polyglutamines placed into context.
Neuron 38, 681–684.
La Spada, A.R., and Taylor, J.P. (2010). Repeat expansion disease: progress
and puzzles in disease pathogenesis. Nat. Rev. Genet. 11, 247–258.
La Spada, A.R., Fu, Y.H., Sopher, B.L., Libby, R.T., Wang, X., Li, L.Y., Einum,
D.D., Huang, J., Possin, D.E., Smith, A.C., et al. (2001). Polyglutamine-
expanded ataxin-7 antagonizes CRX function and induces cone-rod
dystrophy in a mouse model of SCA7. Neuron 31, 913–927.
Ladd, P.D., Smith, L.E., Rabaia, N.A., Moore, J.M., Georges, S.A., Hansen,
R.S., Hagerman, R.J., Tassone, F., Tapscott, S.J., and Filippova, G.N.
(2007). An antisense transcript spanning the CGG repeat region of FMR1 is up-
regulated in premutation carriers but silenced in full mutation individuals. Hum.
Mol. Genet. 16, 3174–3187.
Libby, R.T., Hagerman, K.A., Pineda, V.V., Lau, R., Cho, D.H., Baccam, S.L.,
Axford, M.M., Cleary, J.D., Moore, J.M., Sopher, B.L., et al. (2008). CTCF
cis-regulates trinucleotide repeat instability in an epigenetic manner: a novel
basis for mutational hot spot determination. PLoS Genet. 4, e1000257.
10.1371/journal.pgen.1000257.
Lim, J., Crespo-Barreto, J., Jafar-Nejad, P., Bowman, A.B., Richman, R., Hill,
D.E., Orr, H.T., and Zoghbi, H.Y. (2008). Opposing effects of polyglutamine
expansion on native protein complexes contribute to SCA1. Nature 452,
713–718.
Mangiarini, L., Sathasivam, K., Seller, M., Cozens, B., Harper, A.,
Hetherington, C., Lawton, M., Trottier, Y., Lehrach, H., Davies, S.W., and
Bates, G.P. (1996). Exon 1 of the HD gene with an expanded CAG repeat is
sufficient to cause a progressive neurological phenotype in transgenic mice.
Cell 87, 493–506.
Martin, J.J., Van Regemorter, N., Krols, L., Brucher, J.M., de Barsy, T.,
Szliwowski, H., Evrard, P., Ceuterick, C., Tassignon, M.J., Smet-Dieleman,
H., et al. (1994). On an autosomal dominant form of retinal-cerebellar degen-
eration: an autopsy study of five patients in one family. Acta Neuropathol.
88, 277–286.
Martinez, E., Palhan, V.B., Tjernberg, A., Lymar, E.S., Gamper, A.M., Kundu,
T.K., Chait, B.T., and Roeder, R.G. (2001). Human STAGA complex is a
chromatin-acetylating transcription coactivator that interacts with pre-mRNA
splicing and DNA damage-binding factors in vivo. Mol. Cell. Biol. 21, 6782–
6795.
Mirkin, S.M. (2007). Expandable DNA repeats and human disease. Nature 447,
932–940.
Ohlsson, R., Renkawitz, R., and Lobanenkov, V. (2001). CTCF is a uniquely
versatile transcription regulator linked to epigenetics and disease. Trends
Genet. 17, 520–527.
Okazaki, Y., Furuno, M., Kasukawa, T., Adachi, J., Bono, H., Kondo, S.,
Nikaido, I., Osato, N., Saito, R., Suzuki, H., et al; FANTOM Consortium;
RIKEN Genome Exploration Research Group Phase I & II Team. (2002).
Analysis of the mouse transcriptome based on functional annotation of
60,770 full-length cDNAs. Nature 420, 563–573.
Osato, N., Suzuki, Y., Ikeo, K., and Gojobori, T. (2007). Transcriptional interfer-
ences in cis natural antisense transcripts of humans and mice. Genetics 176,
1299–1306.
Palhan, V.B., Chen, S., Peng, G.H., Tjernberg, A., Gamper, A.M., Fan, Y., Chait,
B.T., La Spada, A.R., and Roeder, R.G. (2005). Polyglutamine-expanded
ataxin-7 inhibits STAGA histone acetyltransferase activity to produce retinal
degeneration. Proc. Natl. Acad. Sci. USA 102, 8472–8477.
Parelho, V., Hadjur, S., Spivakov, M., Leleu, M., Sauer, S., Gregson, H.C.,
Jarmuz, A., Canzonetta, C., Webster, Z., Nesterova, T., et al. (2008).
Cohesins functionally associate with CTCF on mammalian chromosome
arms. Cell 132, 422–433.
Paulson, H.L., Bonini, N.M., and Roth, K.A. (2000). Polyglutamine disease and
neuronal cell death. Proc. Natl. Acad. Sci. USA 97, 12957–12958.
Pearson, C.E., Nichol Edamura, K., and Cleary, J.D. (2005). Repeat instability:
mechanisms of dynamic mutations. Nat. Rev. Genet. 6, 729–742.
Phillips, J.E., and Corces, V.G. (2009). CTCF: master weaver of the genome.
Cell 137, 1194–1211.
Riley, B.E., and Orr, H.T. (2006). Polyglutamine neurodegenerative diseases
and regulation of transcription: assembling the puzzle. Genes Dev. 20,
2183–2192.
Ross, C.A. (1997). Intranuclear neuronal inclusions: a common pathogenic
mechanism for glutamine-repeat neurodegenerative diseases? Neuron 19,
1147–1150.
Sanders, S.L., Jennings, J., Canutescu, A., Link, A.J., and Weil, P.A. (2002).
Proteomics of the eukaryotic transcription machinery: identification of proteins
associated with components of yeast TFIID by multidimensional mass spec-
trometry. Mol. Cell. Biol. 22, 4723–4738.
Shearwin, K.E., Callen, B.P., and Egan, J.B. (2005). Transcriptional interfer-
ence—a crash course. Trends Genet. 21, 339–345.
Stedman, W., Kang, H., Lin, S., Kissil, J.L., Bartolomei, M.S., and Lieberman,
P.M. (2008). Cohesins localize with CTCF at the KSHV latency control region
and at cellular c-myc and H19/Igf2 insulators. EMBO J. 27, 654–666.
Stevanin, G., Durr, A., and Brice, A. (2000). Clinical and molecular advances in
autosomal dominant cerebellar ataxias: from genotype to phenotype and
physiopathology. Eur. J. Hum. Genet. 8, 4–18.
To, K.W., Adamian, M., Jakobiec, F.A., and Berson, E.L. (1993).
Olivopontocerebellar atrophy with retinal degeneration. An electroretino-
graphic and histopathologic investigation. Ophthalmology 100, 15–23.
Trottier, Y., Lutz, Y., Stevanin, G., Imbert, G., Devys, D., Cancel, G., Saudou,
F., Weber, C., David, G., Tora, L., et al. (1995). Polyglutamine expansion as
a pathological epitope in Huntington’s disease and four dominant cerebellar
ataxias. Nature 378, 403–406.
Wang, Y.H., Gellibolian, R., Shimizu, M., Wells, R.D., and Griffith, J. (1996).
Long CCG triplet repeat blocks exclude nucleosomes: a possible mechanism
for the nature of fragile sites in chromosomes. J. Mol. Biol. 263, 511–516.
Neuron 70, 1071–1084, June 23, 2011 ª2011 Elsevier Inc. 1083
Neuron
CTCF Promotes AS ncRNA to Control atx7 Expression
Wendt, K.S., Yoshida, K., Itoh, T., Bando, M., Koch, B., Schirghuber, E.,
Tsutsumi, S., Nagae, G., Ishihara, K., Mishiro, T., et al. (2008). Cohesin medi-
ates transcriptional insulation by CCCTC-binding factor. Nature 451, 796–801.
Yoo, S.Y., Pennesi, M.E., Weeber, E.J., Xu, B., Atkinson, R., Chen, S.,
Armstrong, D.L., Wu, S.M., Sweatt, J.D., and Zoghbi, H.Y. (2003). SCA7
knockin mice model human SCA7 and reveal gradual accumulation of mutant
ataxin-7 in neurons and abnormalities in short-term plasticity. Neuron 37,
383–401.
1084 Neuron 70, 1071–1084, June 23, 2011 ª2011 Elsevier Inc.
Yu, W., Gius, D., Onyango, P., Muldoon-Jacobs, K., Karp, J., Feinberg, A.P.,
and Cui, H. (2008). Epigenetic silencing of tumour suppressor gene p15 by
its antisense RNA. Nature 451, 202–206.
Zhao, Y., Lang, G., Ito, S., Bonnet, J., Metzger, E., Sawatsubashi, S., Suzuki,
E., Le Guezennec, X., Stunnenberg, H.G., Krasnov, A., et al. (2008). A TFTC/
STAGA module mediates histone H2A and H2B deubiquitination, coactivates
nuclear receptors, and counteracts heterochromatin silencing. Mol. Cell 29,
92–101.