Aberrant E2F activation by polyglutamine expansion
of androgen receptor in SBMA neurotoxicity
Eriko Suzukia,b, Yue Zhaoa, Saya Itoa,b, Shun Sawatsubashia,b, Takuya Murataa, Takashi Furutanic, Yuko Shirodea,b,
Kaoru Yamagataa,b, Masahiko Tanabea, Shuhei Kimuraa, Takashi Uedaa, Sally Fujiyamaa,b, Jinseon Lima,
Hiroyuki Matsukawaa, Alexander P. Kouzmenkoa,b, Toshiro Aigakid, Tetsuya Tabataa, Ken-ichi Takeyamaa,
and Shigeaki Katoa,b,1
aInstitute of Molecular and Cellular Biosciences, University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-0032, Japan;bERATO, Japan Science and Technology
Agency, Kawaguchi, Saitama 332-0012, Japan;cPharmacology Research Laboratories, Drug Discovery Research, Astellas Pharma Inc., 21 Miyukigaoka,
Tsukuba, Ibaraki 305-8585, Japan; anddDepartment of Biological Sciences, Tokyo Metropolitan University, Hachioji, 1-1 Minami-osawa, Hachioji,
Tokyo, 192-0397, Japan
Edited by David J. Mangelsdorf, The University of Texas Southwestern Medical Center, Dallas, TX, and approved January 7, 2009 (received for review
October 1, 2008)
disorder caused by a polyglutamine repeat (polyQ) expansion
within the human androgen receptor (AR). Unlike other neurode-
generative diseases caused by abnormal polyQ expansion, the
onset of SBMA depends on androgen binding to mutant human
polyQ-AR proteins. This is also observed in Drosophila eyes ectopi-
cally expressing the polyQ-AR mutants. We have genetically
screened mediators of androgen-induced neurodegeneration
caused by polyQ-AR mutants in Drosophila eyes. We identified Rbf
(Retinoblastoma-family protein), the Drosophila homologue of
human Rb (Retinoblastoma protein), as a neuroprotective factor.
Androgen-dependent association of Rbf or Rb with AR was re-
markably potentiated by aberrant polyQ expansion. Such poten-
tiated Rb association appeared to attenuate recruitment of histone
deacetyltransferase 1 (HDAC1), a corepressor of E2F function.
Either overexpression of Rbf or E2F deficiency in fly eyes reduced
the neurotoxicity of the polyQ-AR mutants. Induction of E2F
function by polyQ-AR-bound androgen was suppressed by Rb in
of polyQ may potentiate innate androgen-dependent associa-
tion of AR with Rb. This appears to lead to androgen-dependent
onset of SBMA through aberrant E2F transactivation caused by
suppressed histone deacetylation.
transcriptional regulation ? neurodegenerative disease ?
motor neuron disorder caused by CAG repeat [encoding poly-
glutamine (polyQ)] expansions in the first exon of the human
androgen receptor (AR) gene (1). Characteristic movement
disorders, neuropsychiatric symptoms, and late onset of neuro-
degeneration are well described in SBMA patients carrying AR
mutations coding for abnormally expanded polyQ repeats
(polyQ-AR) (2). There are similarities to other congenital
neuronal diseases induced by polyQ repeat expansions such as
Huntington’s disease and spinocerebellar ataxia atrophy (3).
Neurodegeneration resulting from polyQ-AR has also been
experimentally demonstrated in mice and flies (4, 5). SBMA is
different from other neurodegenerative diseases in that this
disease is male-specific in mammals, presumably due to the
physiologically sufficient levels of androgens required to activate
the neurodegenerative response of polyQ-AR mutants (6, 7). As
is the case for other polyglutamine diseases, the molecular basis
of SBMA neuropathology remains elusive. However, because of
the androgen dependency of SBMA development, the neuro-
pathology is considered innate to AR function.
Most androgen actions are mediated through the AR, a
ligand-inducible transcription factor that belongs to the nuclear
-chromosome-linked spinal and bulbar muscular atrophy
hormone receptor superfamily (8). In the absence of ligand, AR
is located primarily in the cytoplasm as an inactive complex with
heat shock proteins. Upon androgen binding, AR undergoes
conformational change, translocates into the nucleus, and re-
cruits coactivator complexes for transactivation through direct
DNA binding to androgen response elements (AREs) in AR
target gene promoters (9, 10). As observed in wild-type AR,
polyQ-AR mutants reportedly translocate into the nucleus and
recruit coactivator complexes to the AREs in a ligand-
dependent manner (11).
Previously, we established an androgen-dependent SBMA
Drosophila model by ectopically expressing human polyQ-AR
mutants in eye discs (4). Androgen-induced conformational
change of polyQ-ARs and their translocation into the nucleus
appeared to dramatically increase cytotoxicity of the polyQ-AR,
leading to neuronal cell death. Because the neurodegenerative
abnormalities of this SBMA model were indistinguishable from
those in flies expressing other mutant polyQ-expanded repeats
that do not require androgen, diseases mediated by polyQ-
expanded repeats appear to share common and evolutionarily
conserved neurodegenerative mechanisms (3, 12, 13).
To identify factors required for androgen-induced neurode-
generation, we performed a genetic screen crossing polyQ-AR
transgenic flies with gmr-GAL4-driven overexpression lines. We
identified retinoblastoma-family protein (Rbf), the Drosophila
homologue of human retinoblastoma protein (Rb), to be a
repressive factor for androgen-induced neurodegeneration
caused by polyQ-AR mutants. Potentiated association of Rb
with AR by the aberrant polyQ expansion was inhibitory for
recruitment of HDAC1 corepressing E2F function. E2f defi-
ciency partially rescued the neurodegenerative rough eye phe-
notype in the polyQ-AR mutants. PolyQ-AR mutants, but not
the wild-type AR, stimulated E2F-mediated transactivation.
Thus, abnormal polyQ-AR seems to have impaired Rb transre-
pressive function. The resultant E2F hyperactivation may induce
the androgen-dependent onset of the SBMA disorders.
Rbf Overexpression Rescues PolyQ-AR-Induced Neurodegeneration.
We reported that neurodegeneration induced by the full-length
human polyQ-AR mutant [AR(Q52)] was ligand-dependent, but
Author contributions: E.S., S.S., K.Y., M.T., S. Kimura, T.U., S.F., J.L., H.M., T.T., K.T., and S.
Kato designed research; E.S., Y.Z., T.F., and Y.S. performed research; S.I. and T.A. contrib-
uted new reagents/analytic tools; E.S. and T.M. analyzed data; and E.S., A.P.K., K.T., and S.
Kato wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1To whom correspondence should be addressed. E-mail: firstname.lastname@example.org.
This article contains supporting information online at www.pnas.org/cgi/content/full/
March 10, 2009 ?
vol. 106 ?
the N-terminal domain of the polyQ-AR mutant [AR(Q52
AF-1)] was constitutively and sufficiently potent to induce
This ligand-independent neurodegeneration by AR(Q52 AF-1)
is advantageous for genetic screening to identify modifiers of
polyQ-AR mutant-dependent neurodegeneration. To under-
stand the molecular basis of polyQ-AR-induced neurodegen-
eration, we carried out genetic screening for modifiers of eye
degradation. We crossed GS lines (which misexpressed unknown
genes driven by the GAL4 protein) with UAS-AR (Q52 AF-1)
lines driven by gmr-GAL4. With gmr-GAL4, gene expression
could be specifically targeted to the eye. After screening ?2,000
fly strains, several candidate genes capable of modifying the
the phenotype, putative modifiers selected in the initial screen
were further tested by crossing with different independent
strains expressing AR(Q52) in the presence of an active andro-
gen [dihydrotestosterone (DHT)]. One of the candidate lines,
GS-5173, was found to rescue neurodegeneration by AR(Q52
AF-1) and androgen (DHT)-bound AR(Q52) (Fig. 1B).
With the Drosophila Gene Search Project (DGSP) database,
analysis of genomic sequences downstream of the P element in
the GS-5173 line led to identification of the Rbf gene. Thus, the
gmr-GAL4-driven GS-5173 line was expected to be an Rbf
overexpression mutant. We verified Rbf function by crossing a
UAS-AR(Q52 AF-1) line with an Rbf overexpression line, gmr-
Rbf. Rbf overexpression clearly prevented the rough eye pheno-
type induced by either AR(Q52 AF-1) or AR(Q52) in the
presence of DHT (Fig. 1B).
To determine the specificity of Rbf as a modifier of the
neurodegenerative phenotype, Rbf overexpression lines were
then crossed with a strain expressing the isolated polyQ127
repeat alone, the UAS-polyQ-127 line or the UAS-polyQ-ataxin3
line (SCA3trQ78 line), a Drosophila model of spinocerebellar
ataxia type 3 (12). Notably, Rbf overexpression was unable to
rescue neurodegeneration by either polyQ127 or SCA3tr-Q78
(Fig. 1 C and D). These results suggested that Rbf genetically and
selectively interacts with polyQ-AR mutants in the adjacent
regions and not through the polyQ repeats themselves in the AR
Androgens Stimulate the Interaction of PolyQ-AR Mutants with Rb.
For detailed analysis of the phenotypic effects of overexpression
of Rbf, we tested possible interactions between AR and endog-
enous Rb in human embryonic kidney 293T cells transfected
with either AR(wt) or AR(Q52) expression constructs. By
co-immunoprecipitation (Co-IP) and Western blot (WB) assays,
Rb was shown to weakly interact with AR(wt). Interaction
appeared to depended on binding of DHT but not hydroxyflu-
tamide (HF), the antagonist for AR(wt) (Fig. 2A). In contrast,
with both DHT and HF. Surprisingly, HF served as an agonist
in the association. This observation supports our previous find-
ings that HF serves as an agonist for polyQ-AR-dependent
neurodegeneration (4). To examine whether the potential asso-
ciation of AR with Rb was mediated through the expanded
polyQ repeats, a similar Co-IP experiment was performed in
293T cells expressing ataxin-3, which contains a polyQ60 repeat
similar to that of the polyQ-AR mutants. Consistent with the
inability of Rb to rescue the polyQ-expanded ataxin-3-induced
neurodegeneration phenotype (Fig. 1D), no association of Rb
with ataxin-3(Q60) was detected in the Co-IP assay with anti-Rb
antibodies (Fig. 2B). These results support our hypothesis that
Rb interacts selectively with polyQ-AR mutants.
We then performed GST pull-down assays to determine
whether the functional interaction of the polyQ-AR mutant with
Rb is direct, using recombinant proteins consisting of AR
deletion mutants and GST-fused Rb. Rb physically interacted
with the AR A/B domain containing expanded polyQ [AR(Q52)
A/B] but not with its wild-type [AR(wt) A/B] (Fig. 2C). The AR
C/D/E/F mutant reportedly associates with Rb (14). Because the
AR C domain bears a ‘‘LXCXE’’-like motif (LICGDE), a
(red) and the DNA binding domain (black) are indicated. The transactivation function 1 (AF-1) region is located within the N-terminal A/B domain, whereas the
of Rbf mutant lines rescue polyQ-AR-dependent neurodegeneration. Flies expressing gmr-GAL4 alone or with AR(Q52AF-1) or AR(Q52), GS-5173 or gmr-Rbf lines as
indicated. Genotypes: gmr-GAL4/?; UAS-AR(Q52)/? in trans to GS-5173 or gmr-Rbf. gmr-GAL4/?; UAS-AR(Q52)/? in trans to GS-5173 or gmr-Rbf in the absence or
presence of DHT. (C) Rbf overexpression does not rescue expanded polyQ127-dependent neurodegeneration. Genotypes: gmr-GAL4/?; UAS-Q127/? in trans to w or
Light microscopy (LM) and scanning electron microscopy (SEM) images of adult Drosophila eyes. Bars in (B) LM and SEM images, 100 ?m for eyes in (B–D).
Suzuki et al.PNAS ?
March 10, 2009 ?
vol. 106 ?
no. 10 ?
consensus Rb-contacting motif (15), its impact was then tested
with an AR point mutant (LIAGDE). This mutation in the
full-length AR [AR(wt) C562A] resulted in loss of the Rb
association (Fig. 2D). However, even in this mutant, the
polyQ-AR mutant [AR(Q52) C562A] was still capable of inter-
acting with Rb. Although the Rb associations with AR mutants
were weak in the GST pull-down assay, the associations were
more clearly ligand-dependent in a Co-IP experiment in 293T
cells (Fig. S1B). This discrepancy in ligand dependency between
GST pull-down and Co-IP experiments may be due to nuclear
translocation of liganded ARs in intact cells. Taken together,
these results suggest that the aberrantly potentiated association
of polyQ-AR and Rb is mediated not only through the C domain
but also through the A/B domain of the polyQ-AR mutants.
At this point, we mapped the Rb-interacting domains for ARs.
The 3 tested Rb domains seemed to weakly associate with
AR(wt), whereas the polyQ expansion [AR(Q52)] significantly
potentiated the association with the Rb A/B pocket and C
mutants in the GST pull-down assay (Fig. 2E). Similar results
were also obtained in a Co-IP experiment using 293T cells (Fig.
S1C). Because the Rb A/B pocket domain is known as the E2F
binding pocket repressing the transactivation function of E2F,
we assumed that the abnormal polyQ expansion in AR modu-
lates the E2F function via its association with the Rb A/B pocket.
Because Rb reportedly coactivates the transactivation func-
tion of androgen-bound AR (14, 16), we then tested whether Rb
serves as an AR coactivator for the polyQ-AR mutant. No
functional differences in the coactivation by Rb (Fig. 2F) or
E2F1 (Fig. S1D) were observed in AR(wt) or AR(Q52) in the
human neuroblastoma cell line, SH-SY5Y.
PolyQ-AR Counteracts Rb for E2F Repression. The Rb protein neg-
atively regulates the G1/S transition by repressing the transcrip-
tional activity of the E2F transcription factors. Rbf also represses
dE2f-dependent transcription and suppresses the phenotype
generated by ectopic expression of dE2f and dDp in the devel-
oping Drosophila eye (17) (Fig. 3A). Because Rbf overexpression
prevents polyQ-AR-induced neurodegeneration in the fly eye
similarly to dE2f/dDp1-induced neurodegeneration, we reasoned
that E2F transactivation may mediate polyQ-AR-induced neu-
rodegeneration. To test this idea, we crossed the AR(Q52 AF-1)
line with dE2f-deficient mutant lines. In 2 independent Drosoph-
ila mutant lines deficient in dE2f, E2f07172and E2fi2, the action
of AR(Q52 AF-1) in the rough eye was partially reduced when
compared with that in the control fly (Fig. 3B). Partial rescue of
the rough eye phenotype by the dE2f mutant was also prevented
by Rbf overexpression (Fig. 3C). These results indicate that
dE2f transactivation is enhanced in the Drosophila eye model
Next, we examined whether polyQ-AR influences the trans-
SH-SY5Y. An E2F-dependent reporter plasmid with a cassette
of 6 E2F binding sites derived from the E2F1 gene promoter was
cotransfected with expression vectors for AR, Rb, or both.
Considerably weaker transcriptional activation of endogenous
E2F1 was observed with AR(wt) activated by DHT (Fig. 3D).
However, AR(Q52) activated by DHT significantly enhanced
E2F1 transcriptional activity (Fig. 3D). The AR antagonist, HF,
also enhanced E2F1 transcriptional activity. Furthermore, co-
expression of Rb abrogated the enhancement of E2F1 trans-
activation by AR(Q52) (Fig. 3D).
We then studied the molecular basis of the functional asso-
ciation of the polyQ-AR mutant with E2F1. With Co-IP with
anti-E2F1 antibody (Fig. 3E), we observed DHT-induced asso-
ciation of endogeneous E2F1 with AR(Q52). The association
was undetectable when Rb was knocked down by siRNA (Fig. 3E
and Fig. S2A). Thus, it appears that Rb bridges E2F1 to the
that the AR(Q52) mutant was not coimmunoprecipitated with
an E2F1 mutant (E2F1‚C) lacking its Rb-interacting domain
(Fig. S2 B and C).
Because the polyQ-AR mutant inhibited the transrepressive
might attenuate recruitment of the Rb corepressor, HDAC1,
through association of the Rb transrepressive domain (A/B
pocket domain). We found that HDAC1’s association with Rb
was unaffected by AR(wt)-bound DHT. However, binding of
DHT to AR(Q52) indeed induced HDAC1 dissociation
Activation of E2F Target Gene Expression by PolyQ-AR Mutants.Next,
we asked whether the expression of polyQ-AR indeed leads to
expression of endogenous E2F1 target genes, both in vivo and in
vitro. In the UAS-AR(Q52) fly eye, RT-PCR analysis revealed
that the Drosophila dE2f target gene Drosophila PCNA (dPCNA)
was up-regulated severalfold (Fig. 4A). Likewise, transient ex-
pression of AR(Q52) in human SH-SY5Y cells led to up-
regulation of Cyclin E gene expression (Fig. 4B). However, such
up-regulation of E2F1 target genes was not obvious for AR(wt).
Consistent with the reporter gene expression assay data, over-
expression of both Rbf in Drosophila eyes and Rb in human
by the polyQ-AR mutant.
To explore the molecular basis of polyQ-AR function in E2F
target gene expression, the recruitment of ARs to the E2F target
gene promoters was tested by chromatin immunoprecipitation
(ChIP) analysis. DHT-induced recruitment of AR(Q52) to E2F
binding sites was observed at the dPCNA gene promoters,
together with the expected dissociation of Drosophila HDAC1
from 293T cells. (C) Association of ARs with Rb. A GST pull-down assay was
performed with purified recombinant GST-Rb protein expressed in Escherichia
coli and35S-labeled AR mutants. (D) Potentiated association of Rb with AR by
Rb equally coactivates wild-type and polyQ-AR transactivation. ARE-dependent
reporter was cotransfected with AR(wt) or AR(Q52) with or without Rb expres-
sion plasmids in SH-SY5Y cells that were treated with vehicle or 10?8M DHT.
Results are given as means ? SD for at least 3 independent experiments.
Liganded AR(Q52) interacts with Rb. (A) Co-IP of endogeneous Rb with
www.pnas.org?cgi?doi?10.1073?pnas.0809819106Suzuki et al.
(Rpd3) and hyperacetylation of histone H3, although Rbf asso-
ciation appeared unchanged (Fig. 4C). Similarly, DHT-bound
AR(Q52) was recruited to the E2F binding site of the human
Cyclin E gene promoter with dissociation of HDAC1 and
hyperacetylation of histone H3 in SH-SY5Y cells (Fig. 4D). It is
notable that, unlike AR(Q52), AR(wt) did not associate with the
Cyclin E gene promoter. Rb overexpression inhibited the asso-
ciation of AR(Q52) with the E2F1 target gene promoter
In this report, we used genetic screening for neuroprotective
factors for the neurodegenerative (rough eye) phenotype in a
Drosophila model of SBMA and identified Rbf/Rb as a can-
didate (Fig. 1). The retinoblastoma tumor suppressor Rb and
related factors, such as mammalian p107 and p130 or Dro-
sophila Rbf, are negative regulators of cell proliferation. Rb
functions at major G1checkpoints, thereby inhibiting S-phase
entry and progression. They also promote terminal differen-
tiation by inducing both cell-cycle exit and tissue-specific gene
expression (18, 19).
Consistent with the neuroprotective action of Rbf, we found
that expansion of polyQ repeats in AR mutants augmented the
physical association of Rbf and Rb with the polyQ-AR protein.
Because Rb was unable to counteract the neurodegenerative
activity of the expanded polyQ repeat on its own (Fig. 1C),
Rb/Rbf thus appears to be a selective neuroprotective factor
for SBMA among hereditary polyglutamine diseases. We
attribute the selective Rb/Rbf activity to its enhanced associ-
ation with polyQ-AR mutants. This idea is consistent with the
recent findings that abnormal polyQ expansion in spinocere-
bellar ataxia type 1 protein potentiates native protein inter-
actions, thus leading to neuropathology (20, 21).
It is widely accepted that Rb and related factors function
through their effects on the transcription of genes regulated by
the E2F proteins. E2F binding sites are found in the promoters
of many genes essential for cell-cycle progression and cell
proliferation. Most Rb regulatory functions are associated with
the repressor activities of Rb/E2F complexes located on the
promoters of cell-cycle regulatory genes such as Cyclin E and the
E2F transcription factor (22–24). Moreover, many E2F1 target
genes regulate apoptosis, development, and differentiation.
Overexpression of E2F1 in mammalian cells results in the
induction of apoptosis (25–27). Similarly, ectopic expression of
Drosophila E2f induces apoptosis in imaginal discs and, if
expressed in the eye discs, causes the rough eye phenotype (28).
Significantly, in our genetic analysis, E2f-deficient Drosophila
mutants partially reversed the polyQ-AR-induced rough eye
phenotype (Fig. 3B). Expression of Rb abrogated E2F1 trans-
activation by polyQ-AR (Fig. 3D). The reporter expression assay
data were further confirmed by observations of polyQ-AR-
induced expression of endogenous E2F target genes in Drosoph-
ila eye disc and cultured mammalian cells (Fig. 4 A and B). These
findings suggest that polyQ-AR activates endogenous E2F1
proteins in neuronal cells in a ligand-dependent manner. Al-
though it was reported that Rb coactivates AR through its
physical association (14, 16), no association of the other 8
polyQ-expanded proteins with Rb have been mentioned. In the
present study, we also could not detect the interaction of
the polyQ-expanded proteins, only polyQ-AR selectively en-
hances E2F1 transactivation via Rb.
The transrepression function of Rb/Rbf for E2F1 requires
HDAC1 to deacetylate histone for chromatin inactivation (18).
As expected, wild-type AR did not affect the function of Rb
or E2F1. However, supporting our in vitro observations that
the polyQ-AR mutant competed with HDAC1 in Rb associ-
phenotype in Drosophila eyes. Genotypes: UAS-dE2f, dDp/?; gmr-GAL4/? in trans to w or gmr-Rbf. (B) Independent dE2f -deficient mutants E2f07172and E2fi2
partially rescue polyQ-AR-induced neurodegeneration phenotype. Genotypes: gmr-GAL4 AR(Q52 AF-1)/? in trans to w or E2f07172or E2fi2. Lower show SEM
magnification images. (C) Ectopic coexpression of Rbf completely rescues AR(Q52 AF-1)-induced neurodegeneration phenotype suppressed by dE2f deficiency.
Genotypes: gmr-GAL4/?; UAS-AR(Q52 AF-1)/gmr-Rbf in trans to E2f07172or E2fi2. (A–C) Light microscopy (LM) and scanning electron microscopy (SEM) images
of adult Drosophila eyes. Bars in (A) LM and SEM images indicate 100 ?m for eyes in (A–C). (D) Human Rb abrogates agonist- and antagonist-induced
enhancement of E2F1 transactivation by AR(Q52). E2F-dependent reporter was cotransfected with hAR(wt) or hAR(Q52) and empty vector or Rb expression
to its activity in the absence of androgen. Results are given as means ? SD for at least 3 independent experiments. (E) Co-IP of endogeneous E2F1 with ARs in
293T cells. Rb was knocked down by siRNA (Fig. S2A). (F) Co-IP of endogeneous Rb with endogeneous HDAC1 in 293T cells.
Rb overexpression abrogates enhanced E2F1 transactivation by polyQ-AR. (A) Ectopic expression of Rbf rescues E2f-induced neurodegeneration
Suzuki et al.PNAS ?
March 10, 2009 ?
vol. 106 ?
no. 10 ?
ation, the polyQ-AR mutant was an overt coactivator for E2F1.
Thus, it appears that impaired HDAC1 recruitment to Rb
induced by liganded polyQ-AR mutants causes hyperfunction
of E2Fs (see Fig. S3), leading to androgen-induced develop-
ment of SBMA.
Derepression of the Rb/E2F1 transcriptional complex, or in
other words aberrant activation of E2F1, has been implicated in
diverse human neurological disorders including Alzheimer’s
disease, Parkinson’s disease, and amyotrophic lateral sclerosis
(29–31). Thus, albeit by different mechanisms, diverse neuro-
degenerative agents appear to share a significant common trait
(i.e., E2F activation).
In conclusion, we have presented evidence that abnormal
polyQ expansion in AR leads to aberrant E2F1 activation
presumably through suppression of Rb function by androgen-
induced physical interaction of Rb with polyQ-AR mutants.
Given the roles of E2F protein family members in cellular
toxicity, aberrant activation of E2F factors appears to underlie
the androgen-dependent neurodegenerative function of
polyQ-AR mutants in the development of SBMA. Our data
imply that disabling E2Fs or Rb reactivation might provide an
approach to the prevention of neurodegeneration in SBMA.
Please see SI Text for the following detailed methods: fly stocks, plasmids,
immunoprecipitation, cell culture and transfection; GST pull-down assays,
RT-PCR, and chromatin immunoprecipitation.
ACKNOWLEDGMENTS. We are grateful to I. Takada, T. Matsumoto, H.
H. Yamazaki and H. Higuchi for manuscript preparation, T. Suzuki for fly
maintenance, M. Miura for helpful discussion and providing the SH-SY5Y cell
line, and M. Hatakeyama for plasmids.
1. La Spada AR, et al. (1991) Androgen receptor gene mutations in X-linked spinal and
bulbar muscular atrophy. Nature 352:77–79.
atrophy of late onset. A sex-linked recessive trait. Neurology 18:671–680.
3. Orr HT, Zoghbi HY (2007) Trinucleotide repeat disorders. Annu Rev Neurosci 30:575–
4. Takeyama K, et al. (2002) Androgen-dependent neurodegeneration by polyglu-
tamine-expanded human androgen receptor in Drosophila. Neuron 35:855–864.
5. Katsuno M, et al. (2002) Testosterone reduction prevents phenotypic expression in
a transgenic mouse model of spinal and bulbar muscular atrophy. Neuron 35:843–
6. Riley BE, Orr HT (2006) Polyglutamine neurodegenerative diseases and regulation of
transcription: Assembling the puzzle. Genes Dev 20:2183–2192.
via degradation of androgen receptor. Nat Med 13:348–353.
Proc Natl Acad Sci USA 100:9416–9421.
9. He B, et al. (2004) Structural basis for androgen receptor interdomain and coactivator
interactions suggests a transition in nuclear receptor activation function dominance.
Mol Cell 16:425–438.
10. Rosenfeld MG, Lunyak VV, Glass CK (2006) Sensors and signals: A coactivator/
corepressor/epigenetic code for integrating signal-dependent programs of transcrip-
tional response. Genes Dev 20:1405–1428.
11. Irvine RA, et al. (2000) Inhibition of p160-mediated coactivation with increasing
androgen receptor polyglutamine length. Hum Mol Genet 9:267–274.
12. Li LB, Yu Z, Teng X, Bonini NM (2008) RNA toxicity is a component of ataxin-3
degeneration in Drosophila. Nature 453:1107–1111.
13. Pandey UB, et al. (2007) HDAC6 rescues neurodegeneration and provides an essential
link between autophagy and the UPS. Nature 447:859–863.
14. Lu J, Danielsen M (1998) Differential regulation of androgen and glucocorticoid
receptors by retinoblastoma protein. J Biol Chem 273:31528–31533.
15. Morris EJ, Dyson NJ (2001) Retinoblastoma protein partners. Adv Cancer Res 82:1–54.
16. Yeh S, et al. (1998) Retinoblastoma, a tumor suppressor, is a coactivator for the
androgen receptor in human prostate cancer DU145 cells. Biochem Biophys Res
17. Du W, Vidal M, Xie JE, Dyson N (1996) RBF, a novel RB-related gene that regulates E2F
activity and interacts with cyclin E in Drosophila. Genes Dev 10:1206–1218.
20. Lim J, et al. (2008) Opposing effects of polyglutamine expansion on native protein
complexes contribute to SCA1. Nature 452:713–718.
21. Lam YC, et al. (2006) ATAXIN-1 interacts with the repressor Capicua in its native
complex to cause SCA1 neuropathology. Cell 127:1335–1347.
22. Attwooll C, Lazzerini Denchi E, Helin K (2004) The E2F family: Specific functions and
overlapping interests. EMBO J 23:4709–4716.
23. Frolov MV, Dyson NJ (2004) Molecular mechanisms of E2F-dependent activation and
pRB-mediated repression. J Cell Sci 117:2173–2181.
24. Rowland BD, Bernards R (2006) Re-evaluating cell-cycle regulation by E2Fs. Cell
25. Nahle Z, et al. (2002) Direct coupling of the cell cycle and cell death machinery by E2F.
Nat Cell Biol 4:859–864.
26. Chau BN, Wang JY (2003) Coordinated regulation of life and death by RB. Nat Rev
27. Black EP, et al. (2005) Distinctions in the specificity of E2F function revealed by gene
expression signatures. Proc Natl Acad Sci USA 102:15948–15953.
28. Asano M, Nevins JR, Wharton RP (1996) Ectopic E2F expression induces S phase and
apoptosis in Drosophila imaginal discs. Genes Dev 10:1422–1432.
29. Liu DX, Greene LA (2001) Neuronal apoptosis at the G1/S cell cycle checkpoint. Cell
Tissue Res 305:217–228.
30. Khurana V, et al. (2006) TOR-mediated cell-cycle activation causes neurodegeneration
in a Drosophila tauopathy model. Curr Biol 16:230–241.
31. Hoglinger GU, et al. (2007) The pRb/E2F cell-cycle pathway mediates cell death in
Parkinson’s disease. Proc Natl Acad Sci USA 104:3585–3590.
Drosophila in the absence or the presence of ligand. Genotypes: gmr-GAL4/?;
target gene, human Cyclin E (hCyclin E) expression in SH-SY5Y cells transfected
with AR(wt) or AR(Q52) expression plasmids and treated with vehicle or 10?8M
DHT. (A, B) Graphs show fold induction of the indicated mRNAs compared with
their levels in the absence of ligand. Results are given as means ? SD for at least
3 independent experiments. (C) ChIP analysis of dE2F1, Rbf, and ARs, acetylated
histone H3 (AcH3), and Rpd3 at the E2F response elements of the dE2f target
gene, dPCNA, promoter in Drosophila in the presence of DHT. Genotypes: gmr-
GAL4/?; UAS-AR(wt)/? or gmr-GAL4/?; UAS-AR(Q52)/?. (D) ChIP analysis of
E2F1, Rb, ARs, AcH3, and HDAC1 at the E2F response elements in the promoters
of endogenous human E2F1 target gene, hCyclin E, in SH-SY5Y cells transfected
with AR(wt) or AR(Q52) together with empty vector or Rb expression plasmids
and treated with vehicle or 10?8M DHT.
PolyQ-AR activates E2F1 target gene expression in a ligand-dependent
www.pnas.org?cgi?doi?10.1073?pnas.0809819106Suzuki et al.
Correction Download full-text
Correction for “Aberrant E2F activation by polyglutamine expan-
sionof androgen receptor in SBMA neurotoxicity,” by Eriko Suzuki,
Yue Zhao, Saya Ito, Shun Sawatsubashi, Takuya Murata, Takashi
Kimura, Takashi Ueda, Sally Fujiyama, Jinseon Lim, Hiroyuki
Matsukawa, Alexander P. Kouzmenko, Toshiro Aigaki, Tetsuya
Tabata, Ken-ichi Takeyama, and Shigeaki Kato, which appeared in
issue 10, March 10, 2009, of Proc Natl Acad Sci USA (106:3818–
The authors note that Fig. 4 appeared incorrectly. The cor-
rected figure and its legend appear below. These errors do not
affect the conclusions of the article.
Relative mRNA level
Relative mRNA level
(Q52) Drosophila in the absence or the presence of ligand. Genotypes: gmr-GAL4/+; UAS-AR(wt)/+ and gmr-GAL4/+; UAS-AR(Q52)/+. (B) RT-PCR of the human
E2F1 target gene, human Cyclin E (hCyclin E) expression in SH-SY5Y cells transfected with AR(wt) or AR(Q52) expression plasmids and treated with vehicle or
10−8M DHT. (A, B) Graphs show fold induction of the indicated mRNAs compared with their levels in the absence of ligand. Results are given as means ± SD
for at least 3 independent experiments. (C) ChIP analysis of dE2F1, Rbf, and ARs, acetylated histone H3 (AcH3), and Rpd3 at the E2F response elements of the
dE2f target gene, dPCNA, promoter in Drosophila in the presence of DHT. Genotypes: gmr-GAL4/+; UAS-AR(wt)/+ or gmr-GAL4/+; UAS-AR(Q52)/+. (D) ChIP
analysis of E2F1, Rb, ARs, AcH3, and HDAC1 at the E2F response elements in the promoters of endogenous human E2F1 target gene, hCyclin E, in SH-SY5Y cells
transfected with AR(wt) or AR(Q52) together with empty vector or Rb expression plasmids and treated with vehicle or 10−8M DHT.
PolyQ-AR activates E2F1 target gene expression in a ligand-dependent manner. (A) RT-PCR of dE2f target gene, dPCNA, expression in AR(wt) or AR
| October 2, 2012
| vol. 109
| no. 40