CDYL bridges REST and histone methyltransferases for gene repression and suppression of cellular transformation.
ABSTRACT The neuronal gene repressor REST/NRSF recruits corepressors, including CoREST, to modify histones and repress transcription. REST also functions as a tumor suppressor, but the mechanism remains unclear. We identified chromodomain on Y-like (CDYL) as a REST corepressor that physically bridges REST and the histone methylase G9a to repress transcription. Importantly, RNAi knockdown of REST, CDYL, and G9a, but not CoREST, induced oncogenic transformation of immortalized primary human cells and derepression of the proto-oncogene TrkC. Significantly, transgenic expression of TrkC also induced transformation. This implicates CDYL-G9a, but not CoREST, in REST suppression of transformation, possibly by oncogene repression. CDYL knockdown also augments transformation in a cell culture model of cervical cancer, where loss of heterozygosity of the CDYL locus occurs. These findings demonstrate molecular strategies by which REST carries out distinct biological functions via different corepressors and provide critical insights into the role of histone-modifying complexes in regulating cellular transformation.
- SourceAvailable from: Colby Zaph[Show abstract] [Hide abstract]
ABSTRACT: Inflammatory bowel disease (IBD) pathogenesis is associated with dysregulated CD4+ Th cell responses, with intestinal homeostasis depending on the balance between IL-17-producing Th17 and Foxp3+ Tregs. Differentiation of naive T cells into Th17 and Treg subsets is associated with specific gene expression profiles; however, the contribution of epigenetic mechanisms to controlling Th17 and Treg differentiation remains unclear. Using a murine T cell transfer model of colitis, we found that T cell-intrinsic expression of the histone lysine methyltransferase G9A was required for development of pathogenic T cells and intestinal inflammation. G9A-mediated dimethylation of histone H3 lysine 9 (H3K9me2) restricted Th17 and Treg differentiation in vitro and in vivo. H3K9me2 was found at high levels in naive Th cells and was lost following Th cell activation. Loss of G9A in naive T cells was associated with increased chromatin accessibility and heightened sensitivity to TGF-β1. Pharmacological inhibition of G9A methyltransferase activity in WT T cells promoted Th17 and Treg differentiation. Our data indicate that G9A-dependent H3K9me2 is a homeostatic epigenetic checkpoint that regulates Th17 and Treg responses by limiting chromatin accessibility and TGF-β1 responsiveness, suggesting G9A as a therapeutic target for treating intestinal inflammation.The Journal of clinical investigation 03/2014; · 15.39 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: Human neurons are functional over an entire lifetime, yet the mechanisms that preserve function and protect against neurodegeneration during ageing are unknown. Here we show that induction of the repressor element 1-silencing transcription factor (REST; also known as neuron-restrictive silencer factor, NRSF) is a universal feature of normal ageing in human cortical and hippocampal neurons. REST is lost, however, in mild cognitive impairment and Alzheimer's disease. Chromatin immunoprecipitation with deep sequencing and expression analysis show that REST represses genes that promote cell death and Alzheimer's disease pathology, and induces the expression of stress response genes. Moreover, REST potently protects neurons from oxidative stress and amyloid β-protein toxicity, and conditional deletion of REST in the mouse brain leads to age-related neurodegeneration. A functional orthologue of REST, Caenorhabditis elegans SPR-4, also protects against oxidative stress and amyloid β-protein toxicity. During normal ageing, REST is induced in part by cell non-autonomous Wnt signalling. However, in Alzheimer's disease, frontotemporal dementia and dementia with Lewy bodies, REST is lost from the nucleus and appears in autophagosomes together with pathological misfolded proteins. Finally, REST levels during ageing are closely correlated with cognitive preservation and longevity. Thus, the activation state of REST may distinguish neuroprotection from neurodegeneration in the ageing brain.Nature 03/2014; 507(7493):448-54. · 38.60 Impact Factor
- [Show abstract] [Hide abstract]
ABSTRACT: We detected and characterized the binding sites of the representative Rest complex components Rest, Sin3A, and Lsd1. We compared their binding patterns in mouse embryonic stem (ES) cells and epiblast stem (EpiS) cells. We found few Rest sites unique to the EpiS cells. The ES-unique site features were distinct from those of the common sites, namely, the signal intensities were weaker, and the characteristic gene function categories differed. Our analyses showed that the Rest binding sites do not always overlap with the Sin3A and Lsd1 binding sites. The Sin3A binding pattern differed remarkably between the ES and EpiS cells and was accompanied by significant changes in acetylated-histone patterns in the surrounding regions. A series of transcriptome analyses in the same cell types unexpectedly showed that the putative target gene transcript levels were not dramatically different despite dynamic changes in the Rest complex binding patterns and chromatin statuses, which suggests that Rest is not the sole determinant of repression at its targets. Nevertheless, we identified putative Rest targets with explicitly enhanced transcription upon Rest knock-down in 143 and 60 common and ES-unique Rest target genes, respectively. Among such sites, several genes are involved in ES cell proliferation. In addition, we also found that long, intergenic non-coding RNAs were apparent Rest targets and shared similar features with the protein-coding target genes. Interestingly, such non-coding target genes showed less conservation through evolution than protein-coding targets. As a result of differences in the components and targets of the Rest complex, its functional roles may differ in ES and EpiS cells.PLoS ONE 01/2014; 9(4):e95374. · 3.73 Impact Factor
CDYL Bridges REST and Histone
Methyltransferases for Gene Repression
and Suppression of Cellular Transformation
Peter Mulligan,1Thomas F. Westbrook,2,3Matthias Ottinger,1Natalya Pavlova,2Bin Chang,4,5Eric Macia,6
Yu-Jiang Shi,1,8Jordi Barretina,7Jinsong Liu,3Peter M. Howley,1Stephen J. Elledge,2and Yang Shi1,*
1Department of Pathology
2Howard Hughes Medical Institute, Department of Genetics, Harvard Partners Center for Genetics and Genomics
Harvard Medical School, 77 Avenue Louis Pasteur, Boston, MA 02115, USA
3Department of Biochemistry and Molecular Biology, Department of Molecular and Human Genetics, Baylor College of Medicine, Houston,
TX 77030, USA
4University of Texas M.D. Anderson Cancer Center, 1515 Holcombe Boulevard, Unit 85, Houston, TX 77030-4009, USA
5Department of Pathology and Laboratory of Xinjiang Endemic and Ethnic Diseases, Shihezi University School of Medicine, Shihezi,
Xinjiang 832002, China
6Department of Cell Biology, Harvard Medical School, 200 Longwood Avenue, Boston, MA 02115, USA
7Department of Medical Oncology, Dana Farber Cancer Institute, Dana 720C, 44 Binney Street, Boston, MA 02115-6084, USA
8Present address: Division of Endocrinology, Hypertension, and Diabetes, Brigham and Women’s Hospital, 221 Longwood Avenue,
EBRC 222A, Boston, MA 02115, USA
The neuronal gene repressor REST/NRSF recruits
corepressors, including CoREST, to modify histones
and repress transcription. REST also functions as
a tumor suppressor, but the mechanism remains un-
clear. We identified chromodomain on Y-like (CDYL)
as a REST corepressor that physically bridges REST
and the histone methylase G9a to repress transcrip-
tion. Importantly, RNAi knockdown of REST, CDYL,
and G9a, but not CoREST, induced oncogenic trans-
formation of immortalized primary human cells and
derepression of the proto-oncogene TrkC. Signifi-
cantly, transgenic expression of TrkC also induced
transformation. This implicates CDYL-G9a, but not
CoREST, in REST suppression of transformation,
possibly by oncogene repression. CDYL knockdown
also augments transformation in a cell culture model
of cervical cancer, where loss of heterozygosity of
the CDYL locus occurs. These findings demonstrate
molecular strategies by which REST carries out
distinct biological functions via different corepres-
sors and provide critical insights into the role of
histone-modifying complexes in regulating cellular
Regulated gene expression is critical for normal cellular function
and plays a central role in developmental processes. This is
achieved, in large part, by the activity of transcription factors
that bind to gene regulatory elements, which recruit enzymatic
complexes that render the chromatin environment of genes
more or less favorable to transcription. Aberrations in either the
transcription factors themselves or the enzymatic complexes
that mediate their activity have been implicated in various
diseases, including cancer. However, the molecular mecha-
nisms that link particular transcription regulatory complexes to
specific biological and disease processes remain incompletely
to conserved DNA sequences, RE1 sites, to repress gene tran-
scription (Chong et al., 1995; Schoenherr and Anderson, 1995).
Recently, unexpected roles for REST in regulating embryonic
stem cell pluripotency and self-renewal and tumor suppression
have also been described (Singh et al., 2008; Westbrook et al.,
2005). However, it is unclear how this single transcription factor
regulates apparently diverse biological processes. Gene repres-
sion by REST depends on the recruitment of multiple enzymatic
corepressor complexes that modify chromatin to repress tran-
scription (Ballas and Mandel, 2005). These include mSin3A
acetylases (HDACs), and CoREST (Andres et al., 1999; Shi et al.,
thylase, LSD1 (Shi et al., 2004). The H3K9methyltransferase G9a
In the current study, we describe the purification of a chromo-
domain on Y-like (CDYL) corepressor complex that contains
REST. CDYL is a chromodomain protein that contains an un-
usual sequence homology with lipid-metabolizing enzymes of
the enoyl CoA hydratase family (Caron et al., 2003; Lahn et al.,
2002). While the chromodomain of CDYL binds to methylated
H3K9 residues on histones (Fischle et al., 2008 and unpublished
data), the function of the metabolic enzyme homology domain
remainsunclear,althoughit wasshowntobind toHDAC1(Caron
718 Molecular Cell 32, 718–726, December 5, 2008 ª2008 Elsevier Inc.
et al., 2003). Previous work in our laboratory identified CDYL as
a substoichiometric component of the CtBP corepressor super-
complex (Shi et al., 2003), and it was shown to repress heterolo-
gous gene transcription in reporter-based assays (Caron et al.,
2003), consistent with a role in gene repression. However, by
and large, the function and mechanism of CDYL are unclear.
We have identified a CDYL complex that contains REST and
the histone H3K9 methylases G9a, EuHMT1, and SETDB1,
among others. We show that CDYL bridges the interaction be-
tween REST and G9a/EuHMT1 in vitro and in vivo and functions
as a REST corepressor that facilitates G9a recruitment to REST
both CDYL and G9a in REST suppression of cellular transforma-
tion. REST/CDYL/G9a directly represses transcription of the
proto-oncogene TrkC, whose overexpression is sufficient to
cause cellular transformation, suggesting that the mechanism
of REST/CDYL/G9a transformation suppression may be linked
to their ability to repress oncogene transcription. Finally, we pro-
vide evidence that transformation suppression by CDYL may be
the genetic locus encoding CDYL is frequent and correlates with
poor prognosis (Arias-Pulido et al., 2004; Chatterjee et al., 2001;
Krul et al., 1999). Together, these data provide insights into
CDYL function and mechanism and support a specific role for
CDYL and its associated histone methylase G9a in mediating
REST tumor suppression function. These findings shed light on
molecular strategies utilized by a single eukaryotic DNA-binding
transcription factor to regulate divergent biological functions.
RESULTS AND DISCUSSION
Identification of a CDYL Repressor Complex
To gain insights into the mechanism and biological function of
CDYL, we carried out affinity purification of Flag-HA-tagged
CDYL from HeLa nuclear extracts to isolate CDYL-associated
proteins (Figure 1A, compare lanes 4 and 3). Mass spectrometry
analysis identified more than 22 associated proteins (Table S1
available online), the majority of which are involved in transcrip-
tional repression. These include histone-modifying enzymes,
such as HDAC1 and HDAC2 (Taunton et al., 1996), G9a and
EuHMT1(Ogawa etal.,2002),andtheH3K9 trimethyltransferase
SETDB1 and its associated cofactor MCAF/hAM (Wang et al.,
2003). Also identified was mesoderm induction early response 1
(MI-ER1) (Ding et al., 2003), a transcriptional corepressor that
binds HDAC1, is critical in Xenopus development, and is aber-
rantly expressed in breast cancer (Ding et al., 2003), as well as
its homolog MI-ER2, the function of which is unknown. Three
large multizinc finger proteins, KIAA1221, its homolog ZNF644,
and widely interspaced zinc fingers (WIZ) (Matsumoto et al.,
1998), were also copurified with CDYL. WIZ associates with
G9a/EuHMT1, directly binds CtBP, and represses E-cadherin
gene expression (Ueda et al., 2006), while the functions of
KIAA1221 and ZNF644 are unknown. We also identified the
sequence-specific DNA-binding factor RE1-binding silencer of
transcription (REST)/neuronal-specific silencing factor (NRSF)
(Chong et al., 1995; Schoenherr and Anderson, 1995). REST is
a silencer of neuronal gene transcription in nonneuronal cells
and a master regulator of neuronal differentiation. Importantly,
a recent report identified REST as a tumor suppressor (West-
brook et al., 2005), but the underlying mechanism is unclear.
proteins formsingle or discrete subcomplexes.Glycerol gradient
tion of purified Flag-HA-CDYL exists in the free form (Figures 1B
appear to exist. A slower sedimenting complex, which peaks in
faster sedimenting complex peaks in fraction 9 and contains
REST, WIZ, and all three histone methyltransferases identified
in the purification (Figures 1B and 1C). This suggests that
CDYL, REST, and histone methyltransferases are likely to be in
the same protein complex. We carried out coimmunoprecipita-
tion to further examine this possibility. Importantly, both CDYL
pare lane 2 to lane 3), further supporting the idea that REST,
CDYL, and G9a are components of the same protein complex.
CDYL Bridges the Interaction between REST and G9a
REST directly interacts with both the mSin3A and CoREST
corepressors for transcriptional repression (Andres et al., 1999;
Grimes et al., 2000; Roopra et al., 2000). A recent study also
rectly or indirectly interacts with G9a (Roopra et al., 2004).
Interestingly, CoREST appears to be dispensable for G9a
recruitment (Roopra et al., 2004). To determine whether CDYL
ried out in vitro protein interaction assays. We found that purified
combinant Flag-REST (Figure 2A; Figure S1). Purified recombi-
nant G9a and EuHMT1 also interacted robustly with GST-CDYL
invitro (Figure2A).Wethenusedaninvitroprotein interactionas-
say (diagrammed in Figure 2B) to ask whether CDYL functions to
bridge the interaction between REST and G9a. Purified HA-REST
of GST-CDYL. The ability of HA-REST to interact with Flag-G9a
was investigated byimmunoprecipitating HA-REST and assaying
for coimmunoprecipitation of Flag-G9a. As shown in Figure 2C,
purified REST and G9a do not interact in vitro (lane 5). However,
GST-CDYL, but not GST alone or CoREST, restored binding of
G9a (and EuHMT; data not shown) to REST (Figure 2C, compare
lane 6 with 7 and 8). These data strongly suggest that REST re-
quires CDYL to bridge the interaction with G9a/EuHMT1.
We then investigated whether CDYL mediates the interaction
between REST and G9a in vivo. Although REST antibodies
immunoprecipitated comparable amounts of REST from both
control (U6) and shCDYL nuclear extracts (Figure 2D and data
not shown), coimmunoprecipitation of G9a was significantly di-
minished in the shCDYL extracts (Figure 2D, compare lanes 2
and 5). This suggests that CDYL also mediates the interaction
between endogenous REST and G9a in vivo.
Repression of REST Target Genes by CDYL Correlates
with G9a Recruitment and H3K9 Dimethylation
We next asked whether the ability of CDYL to bridge the interac-
tion between REST and G9a is important for gene repression
CDYL-G9a Regulation of REST Activity
Molecular Cell 32, 718–726, December 5, 2008 ª2008 Elsevier Inc. 719
Figure 1. CDYL Forms a Multiprotein Complex with REST and Histone H3K9 Methyltransferases
(A) Human CDYL tagged with both Flag and HA epitopes was sequentially immunoprecipitated from transduced HeLa nuclear extracts using Flag and HA
antibody resins (lanes 2 and 4). Mock-transduced HeLa cells were used as a control (lanes 1 and 3). CDYL-associated proteins were resolved by SDS-PAGE
and identified by mass spectrometry, as indicated on the silver stained gel shown.
(B) Glycerol gradient ultracentrifugation was used to resolve CDYL-associated proteins into differentially sedimenting multiprotein complexes. Fractions were
collected from the bottom of the gradient and analyzed by SDS-PAGE. Visualization of protein bands by silver staining revealed the presence of at least two
CDYL subcomplexes, represented by fractions F15 and F9, respectively. The positions of CDYL and major components of each subcomplex are indicated.
(C) Immunoblotting of glycerol gradient fractions was used to confirm the identity of major components present in the two CDYL subcomplexes.
(D) Antibodies against REST (lane 2), but not control IgG (lane 3), coimmunoprecipitated G9a and CDYL in the Flag-purified CDYL complex.
CDYL-G9a Regulation of REST Activity
720 Molecular Cell 32, 718–726, December 5, 2008 ª2008 Elsevier Inc.
in vivo. Quantitative RT-PCR showed that stable RNAi knock-
down of CDYL derepressed several REST target genes in
TLM-HMEC (Figures 2E and 2F). Chromatin immunoprecipita-
tion (ChIP) revealed that the genomic regions inclusive of the
RE1 sites at the TrkC and NPTXR genes were occupied by
REST, CDYL, and G9a and enriched in H3K9 dimethylation rela-
tive to the control GAPDH promoter (Figure S2). Significantly,
RNAi knockdown of CDYL also resulted in a diminished occu-
pancy of CDYL, G9a, and H3K9 dimethylation around the RE1
sites at both the TrkC and NPTXR genes (Figure 2G). The mod-
erate loss of occupancy by these factors may be due to the
incomplete knockdown of CDYL in TLM-HMEC, resulting in a
partial loss of CDYL function (Figure 2E). In contrast, REST and
histone H3 occupancy were not significantly diminished (Fig-
ure 2G). Together, these data support the notion that CDYL is
a REST corepressor that facilitates G9a recruitment and H3K9
dimethylation at REST target genes for repression.
CDYL also contains a conserved N-terminal chromodomain
that binds to methylated H3K9 (Fischle et al., 2008 and unpub-
lished data). We speculate that CDYL chromodomain binding
to methyl-H3K9 may stabilize CDYL recruitment to REST-bind-
ing sites and/or facilitates spreading of the CDYL corepressor
complexalong the chromatin fiber fromtheinitial point ofrecruit-
ment. Such a model may explain why extended methyl-H3K9
domains are observed around certain REST-binding sites (Roo-
pra et al., 2004).
RNAi Knockdown of CDYL and G9a Induces
Transformation of TLM-HMEC
A recent finding described a somewhat unexpected function of
REST as a tumor suppressor required for suppression of TLM-
HMEC transformation in vitro (Westbrook et al., 2005). However,
it remains unclear how REST mediates this transformation
suppression activity. REST activity depends on the recruitment
of corepressor complexes, such as the CoREST complex, which
modifies chromatin to repress transcription (Andres et al., 1999;
Grimes et al., 2000). As discussed above, our evidence suggests
that CDYL also functions as a REST corepressor by recruiting
H3K9 methyltransferases, including G9a. This prompted us to
ask whether CDYL may be involved in suppression of cellular
transformation in TLM-HMEC. Using the same assay that identi-
fied REST as a tumor suppressor (Westbrook et al., 2005), we
found that knockdown of CDYL, but not CoREST, is sufficient
to induce transformation in these cells (Figures 3A and 3B).
These findings suggest that CDYL, but not CoREST, is required
for suppression of TLM-HMEC transformation.
The results of the transformation assays complement our bio-
chemical observations that CDYL, but not CoREST, bridges the
interaction between REST and G9a in vitro (Figure 2C) and is im-
determine whether transcriptional repression by REST-CDYL via
the recruitment of histone methyltransferases may be important
in transformation suppression, we investigated whether G9a
also plays a role in suppressing anchorage-independent growth.
As shown in Figures 3C and 3D, two independent shRNA
plasmids that efficiently inhibited G9a expression induced trans-
formation in TLM-HMEC. This suggests that suppression of
TLM-HMEC transformation by REST may depend on its direct
interaction with CDYL, which, in turn, recruits G9a to repress
TrkC/NTRK3 Induces Cellular Transformation
and Is Repressed by REST, CDYL, and G9a
We considered the possibility that CDYL and G9a may sup-
press transformation in TLM-HMEC by repressing the transcrip-
tion of an oncogene. The proto-oncogene TrkC/NTRK3 is a re-
ceptor tyrosine kinase that contains an intronic RE1 site and is
derepressed by CDYL RNAi in TLM-HMEC (Figure 2F). TrkC
plays a critical role in neurogenesis, cancers of the neural line-
age (Nakagawara, 2001), and other types of cancers (Jin et al.,
2007; McGregor et al., 1999). We found that stable transgenic
expression of TrkC in TLM-HMEC induced cellular transforma-
tion, identifying TrkC as a candidate oncogene in these cells
(Figure 3E). Expression of a TrkC mutant lacking the extracellu-
lar and transmembrane domains failed to induce transformation
(Figure 3E). Importantly, TrkC is upregulated in TLM-HMEC
upon shRNA knockdown of REST, CDYL, and G9a, but not in
shCoREST or shFF2 TLM-HMEC (Figure 3F). We then investi-
gated whether REST knockdown in the shREST TLM-HMEC
in which transformation induction was previously reported
(Westbrook et al., 2005) affected the occupancy of CDYL,
G9a, and H3K9me2 at the TrkC REST-binding site. We found
that knockdown of REST in TLM-HMEC not only results in di-
minished REST occupancy but also diminished occupancy of
CDYL, G9a, and H3K9me2 (Figures 3G and 3H). These findings
identify TrkC as a direct target of the REST/CDYL/G9a repres-
sor complex and suggest that TrkC repression may play
a role in REST/CDYL/G9a-mediated suppression of cellular
transformation. Future studies will determine whether the
REST repressor complex also regulates additional oncogenes
and tumor suppressor genes to accomplish tumor suppression.
Squamous Epithelial Cells Expressing HPV16
Oncoproteins E6 and E7
Given the role of CDYL in suppressing cellular transformation
(Figure 3B), we next investigated a potential role of CDYL loss of
LOH of 6p25, which encompasses the CDYL locus, in cervical
carcinoma (CC) (42.6%–70.4% LOH of 6p25 in invasive CC)
(Arias-Pulido et al., 2004; Chatterjee et al., 2001; Krul et al.,
1999; Mazurenko et al., 2006). The major risk factor for CC is in-
fection by specific ‘‘high-risk’’ HPV types, which express the E6
and E7 oncoproteins that target tumor suppressor pathways
and contribute directly to genomic instability (Howley and
Lowy, 2007). However, only a small percentage of HPV-infected
women develop CC (Munoz et al., 2003), suggesting that,
whereas HPV infection might be necessary for CC, it is not suffi-
cient. Furthermore, specific cytogenetic abnormalities are com-
lular genes might contribute to carcinogenic progression.
Significantly, 6p25 LOH was observed in the earliest stages of
CC and is more prevalent in precancerous lesions that have
tral-risk precancerous lesions (Arias-Pulido et al., 2004; Chatter-
jee et al., 2001; Krul et al., 1999). This suggests that mutation of
CDYL-G9a Regulation of REST Activity
Molecular Cell 32, 718–726, December 5, 2008 ª2008 Elsevier Inc. 721
CDYL-G9a Regulation of REST Activity
722 Molecular Cell 32, 718–726, December 5, 2008 ª2008 Elsevier Inc.
candidate tumor suppressor genes in this region may play a crit-
ical role in CC progression. However, the identity of such tumor
suppressor genes has remained elusive.
operates with high-risk HPV16 E6 and E7 oncoproteins to trans-
passage primary human squamous epithelial cells (normal oral
keratinocytes [NOK]) immortalized by expression of exogenous
hTERT were stably transduced with HPV16 oncoproteins E6
and E7 (E6/E7-NOK) or an empty viral vector (Figure 4A). CDYL
was then stably knocked down in these cells by shRNA, with
shFF2 used as a negative control (Figure 4B). The cells were
plated in semisolid media and assayed for anchorage-indepen-
dent growth of colonies 4 weeks later. As has previously been
demonstrated, the expression ofHPV16 E6and E7 oncoproteins
alone was sufficient for keratinocyte transformation (Munger
et al., 1989). Significantly, knockdown of CDYL in E6/E7-NOK
by two independent shRNAs resulted in a consistent 2.5- to 3-
(Figure 4C). This indicates that knockdown of CDYL augmented
the transformation phenotype of E6/E7-expressing NOK. Impor-
tantly, NOK that were not transduced with HPV16 E6 and E7 did
Together, these data suggest that CDYL mutation, though insuf-
the transformation phenotype of an HPV E6/E7-expressing cell.
Taken together, our findings support a model whereby CDYL
bridges the interaction between REST and G9a to repress
formation in TLM-HMEC, including the proto-oncogene TrkC. As
REST-binding sites have been mapped to some 1946 sites in the
genome (Johnson et al., 2007; Otto et al., 2007), we speculate
that other target genes may also participate in suppression of
transformation by the REST/CDYL/G9a repressor complex. The
failure of RNAi knockdown of CoREST to transform TLM-HMEC
suggests that different REST corepressors mediate different
of transformation. The fact that CDYL facilitates the interaction
between REST and G9a and that G9a is also required for
suppression of transformation in TLM-HMEC suggests that the
ability of CDYL to bridge REST-G9a interaction is important for
its mechanism of transformation suppression. Importantly, our
findings also suggest a role for CDYL in suppression of cellular
transformation by HPV16 oncoproteins in squamous epithelial
cells (Figure 4), the normal host cells of HPV infection. Given
the LOH of the locus encoding CDYL in cervical cancer and its
correlation with poor prognosis, this suggests that CDYL may
sociated with HPV infection. By demonstrating that RNAi
dent cell culture models, TLM-HMEC and E6/E7-NOK, our
findings suggest that CDYL may play a general role in suppres-
sion of cellular transformation and, potentially, clinical cancer.
Taken together, these findings cast light on the mechanism of
mechanisms that link a particular histone-modifying complex
with a specific biological process.
Cell Culture, Viral Transduction, and RNAi
HeLa cells were maintained as per Shi et al., (2003), TLM-HMEC cells were
maintained as previously described (Zhao et al., 2003; Westbrook et al.,
viral vectors expressing Flag-HA-CDYL mRNA isoform 2 were constructed as
described previously (Shi et al., 2003). Retroviral and lentiviral shRNA vectors
were obtained from Open Biosystems (clone Id: shCDYL #1, V2HS_68448;
shCDYL #2, V2HS_68451; shCoREST #1, V2HS_87301; shCoREST #2,
V2HS_87304; shG9a #1, TRCN0000115667; shG9a #2, TRCN0000115668) or
as previously described (Westbrook et al., 2005). pMSCV-neo (Clontech) ex-
Smogorzewska. Preparation of viruses and cell transduction was as described
previously (Westbrook et al., 2005). HeLa cells were transfected with pBlue-
script-U6 or pBluescript-U6-shRNA plasmid constructed to target CDYL se-
quence AAGGTACATCTCCGTTCATGG, as previously described (Sui et al.,
2002). Cells were cotransfected with pBABE-puro and selected with 1 mg/ml
puromycin for 48 hr before preparation of nuclear extracts.
CDYL Tandem Affinity Purification and Mass Spectrometry
CDYL-associated proteins were purified from HeLa cell nuclear extracts and
identified by mass spectrometry as previously described (Shi et al., 2003).
Figure 2. CDYL Bridges the Interaction between REST and the Histone Methyltransferases G9a and EuHMT1 and Is a Corepressor of REST
(A) Immunoblot analysis showing that purified GST-CDYL,but not GST, interacts with purified recombinant Flag-REST. Coomassieblue-stained SDS-PAGE gels
(B) Schematic diagram of the CDYL bridging assay to determine whether REST interaction with G9a/EuHMT1 depends on CDYL.
(C) CDYL bridging assay. Purified recombinant proteins were used to show that CDYL is required for the interaction of REST and G9a in vitro. Anti-HA resin was
used to immunoprecipitate HA-REST, and coimmunoprecipitation of purified Flag-G9a was detected by immunoblot analysis. REST did not interact directly with
G9a in vitro (lane 5). However, addition of GST-CDYL induced association of REST and G9a (lane 6), but not GST alone (lane 7) or CoREST (lane 8).
(D) REST was immunoprecipitated from control (U6) or CDYL RNAi HeLa nuclear extracts and assayed for coimmunoprecipitation of G9a by immunoblotting.
Coimmunoprecipitation of G9a with REST was significantly diminished upon CDYL knockdown (compare lanes 2 and 5). Nonspecific IgG did not immunopre-
cipitate REST, G9a, or CDYL (lanes 3 and 6). The position of the CDYL band is indicated by an arrow, and the IgG heavy chain is indicated by an asterisk.
(E) Immunoblot showing knockdown of CDYL in TLM-HMEC by two independent shCDYL hairpins. RNAi directed against firefly luciferase, shFF2, was used as
(F) Knockdown of CDYL in TLM-HMEC derepressed of a number of REST target genes, as determined by quantitative RT-PCR. Samples were normalized
to GAPDH expression levels and expressed as fold increase relative to control shRNA-treated cells (shFF2). Shown is the mean ± SD of three independent
(G) Knockdown of CDYL in TLM-HMEC resulted in decreased occupancy of REST-binding sites at the TrkC and NPTXR genes by CDYL, G9a, and H3K9me2.
Levels of REST and H3 occupancy were not affected. Samples were normalized to input chromatin and expressed as percentage occupancy relative to control
shRNA-treated cells. Shown is the mean ± SD of three independent assays.
CDYL-G9a Regulation of REST Activity
Molecular Cell 32, 718–726, December 5, 2008 ª2008 Elsevier Inc. 723
Glycerol Gradient Sedimentation
Flag-HA-CDYL tandem-affinity-purified material was analyzed as previously
described (Shi et al., 2003) using 4 ml, 15%–40% glycerol gradients in Buffer A
(20 mM Tris-HCl [pH 7.9], 5 mM MgCl2, 10% glycerol, 1 mM phenylmethylsul-
phonyl fluoride [PMSF], 0.1% Nonidet P40, 10 mM 2-mercaptoethanol) con-
taining 100 mM KCl, and the material was centrifuged for 15 hr at 55,000 rpm.
Fractions were collected from the bottom of the gradient, resolved by
SDS-PAGE, and analyzed by silver staining or immunoblotting.
Flag-purified Flag-HA-CDYL-associated proteins or HeLa nuclear extracts
were diluted 3-fold in IP buffer containing 50 mM Tris (pH 7.3), 150 mM NaCl,
Figure 3. CDYL and G9a, but Not CoREST,
Suppress Transformation of TLM-HMEC
and Repress Transcription of Proto-Onco-
(A) TLM-HMEC were stably transduced with two
independent shCoREST constructs. Knockdown
was confirmed by immunoblotting. CDYL knock-
down is shown in Figure 2E.
(B) Transformation assays were performed by
plating stable shRNA cells in semisolid media
and quantifying anchorage-independent (A.I.) col-
ony formation after 3 weeks. shRNAs targeting
firefly luciferase (shFF2) and PTEN tumorsuppres-
sor (shPTEN) were used as negative and positive
controls for transformation, respectively. Shown
is mean ± SD of triplicate samples.
(C and D) Similarly, two independent shRNA con-
structs were used to stably knock down G9a ex-
pression in TLM-HMEC, and the resultant cells
were plated in semisolid media to assay for trans-
(E) Transgenic expression of TrkC wild-type (TrkC
WT), but not a mutant TrkC lacking the extracellu-
lar and transmembrane domains (TrkC Mut), in-
duced A.I. growth of TLM-HMEC relative to empty
vector (pRoles). Shown is the mean ± SD of tripli-
cate samples. *p = 0.004.
(F) Quantitative RT-PCR demonstrating upregu-
lated TrkC expression in TLM-HMEC upon shRNA
knockdown of REST, CDYL, and G9a, but not
shFF2 or shCoREST. Shown is the mean ± SD of
three independent experiments.
(G) Proposed regulation of TrkC by REST binding
to a conserved RE1 site between exons 3 and 4,
thereby recruiting CDYL and G9a. shREST is
expected to cause loss of REST occupancy and
failure to recruit CDYL and G9a.
(H) Quantitative ChIP PCR analysis of the intronic
TrkC RE1 site in TLM-HMEC. shREST induced di-
minished RE1 occupancy by REST, with a con-
comitant decrease in CDYL and G9a occupancy
and H3K9me2 levels. shFF2 TLM-HMEC were
used as a negative control. Shown is the mean ±
SD of three independent experiments.
0.1% NP-40, 1 mM EDTA (pH 8), 1 mM PMSF,
5 mM 2-Mercaptoethanol, and 10 mM MG132
(Sigma) and incubated overnight with REST IgG
(gift of Gail Mandel) or nonspecific IgG at 4?C. Pro-
4?C and then washed four times with IP buffer.
Beads were boiled in SDS-PAGE loading buffer, and eluates were analyzed
In Vitro Protein Interaction Assays
Recombinant proteins were prepared as described in the Supplemental
Experimental Procedures. For GST pull-down assays, GST-CDYL (2 mg) was
incubated with Flag-REST (200 ng) or Flag-G9a/Flag-EuHMT1 (1 ug) overnight
at 4?C in 400 ml Buffer A containing 150 mM NaCl and 0.1% BSA. Beads were
sample buffer to elute bound proteins. For CDYL bridging assays, HA-REST
(200 ng) and Flag-G9a/Flag-EHMT1 (500 ng) were incubated overnight at
4?C with either GST-CDYL (190 ng), GST alone (50 ng), or His6-CoREST
CDYL-G9a Regulation of REST Activity
724 Molecular Cell 32, 718–726, December 5, 2008 ª2008 Elsevier Inc.
(150ng) in200 mlBufferA containing100 mMNaCl and 0.1% BSA.Complexes
were purified using anti-HA resin (Santa Cruz), washed four times in Buffer A
supplemented with 100 mM NaCl, eluted by boiling in sample buffer, and
analyzed by immunoblotting.
RT-PCR and Chromatin Immunoprecipitation
RNA was extracted using Trizol reagent (Invitrogen) as per manufacturer’s
protocol. Quantitative RT-PCR was performed using the Lightcycler 480 kit
(Roche). RT-PCR primer pairs were as previously described (Bruce et al.,
2004; Shi et al., 2004). ChIP was performed according to the Upstate protocol,
as previously described (Shi et al., 2003). ChIP PCR primers are listed in
Supplemental Experimental Procedures.
TLM-HMEC transformation assays were performed as previously described
(Westbrook et al., 2005). NOK cells were described previously (Piboonniyom
tal Experimental Procedures. For NOK transformation assays, 105cells were
plated in triplicate on 6 cm plates. Cells were seeded in a final volume of
1.5 ml of 0.3% noble agar (Sigma) on 0.6% bottom agar plates. Each week,
0.5 ml of fresh complete medium was added to the plates. Colonies of
100 mm or greater were counted by microscopy after 4 weeks.
The Supplemental Data include Supplemental Experimental Procedures, two
figures, and one table and can be found with this article at http://www.cell.
We gratefully acknowledge the assistance of Benedikt Kessler and Eric Spoo-
analysis. We thank Gail Mandel for providing REST antibody and reagents,
Augments Transformation of Immortalized
Primary Human Squamous Epithelial Cells
by HPV16 Oncoproteins E6 and E7
(A) NOK transformation assay. hTERT-immortal-
ized primary human squamous epithelial cells
(NOK) were engineered to stably express HPV16
oncoproteins E6 and E7 using retroviral transduc-
tion (E6/E7-NOK). These cells were then stably
transduced with two independent shRNA vectors
targeting knockdown of CDYL. As a negative
control, cells were stably transduced with shRNA
vectors targeting firefly luciferase (shFF2). Cells
were plated in semisolid media to assay for A.I.
(B) Immunoblotting confirmed shRNA knockdown
of CDYL in E6/E7-NOK.
(C) Transformation assay showing that shCDYL
augments A.I. colony formation of E6/E7-NOK.
Shown is the mean of triplicate samples ± SD.
4. RNAiKnockdown of CDYL
Grace Gill and Jian Ouyang for His-CoREST,
Agata Smogorzewska for pMSCV-neo-HPV16-
E6/E7, Karl Munger for NOK cells, and Yoshihiro
Nakatani and Bijan Sobhian for G9a antibody
and expression plasmid. We also thank Grace
Gill, Keith Blackwell, Wade Harper, and their labo-
ratories for critical discussion of the project. This
work was supported by NIH grants GM071004
(to Y.S.) and PO1 CA050661 (to P.M.H.). S.J.E is an investigator with the Ho-
ward Hughes Medical Institute, and T.F.W. is funded by grant PDF0403175
from the Susan G. Komen Breast Cancer Foundation. Y.S. is a cofounder of
Constellation Pharmaceuticals and member of its scientific advisory board.
Received: July 25, 2008
Revised: September 26, 2008
Accepted: October 22, 2008
Published: December 4, 2008
Andres, M.E., Burger, C., Peral-Rubio, M.J., Battaglioli, E., Anderson, M.E.,
Grimes, J., Dallaman, J., Ballas, N., and Mandel, G. (1999). CoREST: a func-
tional corepressor required for regulation of neural-specific gene expression.
Proc. Natl. Acad. Sci. USA 96, 9873–9878.
Arias-Pulido, H., Joste, N., and Wheeler, C.M. (2004). Loss of heterozygosity
on chromosome 6 in HPV-16 positive cervical carcinomas carrying the
Genes Chromosomes Cancer 40,
Ballas, N., and Mandel, G. (2005). The many faces of REST oversee epigenetic
programming of neuronal genes. Curr. Opin. Neurobiol. 15, 500–506.
Bruce, A.W., Donaldson, I.J., Wood, I.C., Yerbury, S.A., Sadowski, M.I., Chap-
man, M., Gottgens, B., and Buckley, N.J. (2004). Genome-wide analysis of
repressor element 1 silencing transcription factor/neuron-restrictive silencing
factor (REST/NRSF) target genes. Proc. Natl. Acad. Sci. USA 101,
Caron, C., Pivot-Pajot, C., van Grunsven, L.A., Col, E., Lestrat, C., Rousseaux,
S., and Khochbin, S. (2003). Cdyl: a new transcriptional co-repressor. EMBO
Rep. 4, 877–882.
Chatterjee, A., Pulido, H.A., Koul, S., Beleno, N., Perilla, A., Posso, H., Manu-
sukhani, M., and Murty, V.V. (2001). Mapping the sites of putative tumor
CDYL-G9a Regulation of REST Activity
Molecular Cell 32, 718–726, December 5, 2008 ª2008 Elsevier Inc. 725
suppressor genes at 6p25 and 6p21.3 in cervical carcinoma: occurrence of
allelic deletions in precancerous lesions. Cancer Res. 61, 2119–2123.
Chong, J.A., Tapia-Ramirez, J., Kim, S., Toledo-Aral, J.J., Zheng, Y., Boutros,
M.C., Altshuller, Y.M., Frohman, M.A., Kraner, S.D., and Mandel, G. (1995).
REST: a mammalian silencer protein that restricts sodium channel gene
expression to neurons. Cell 80, 949–957.
Ding, Z., Gillespie, L.L., and Paterno, G.D. (2003). Human MI-ER1 alpha and
beta function as transcriptional repressors by recruitment of histone deacety-
lase 1 to their conserved ELM2 domain. Mol. Cell. Biol. 23, 250–258.
Fischle, W.,Franz, H., Jacobs, S.A., Allis,C.D., and Khorasanizadeh, S.(2008).
Specificity of the chromodomain Y chromosome family of chromodomains for
lysine-methylated ARK(S/T) motifs. J. Biol. Chem. 283, 19626–19635.
Grimes, J.A., Nielsen, S.J., Battaglioli, E., Miska, E.A., Speh, J.C., Berry, D.L.,
Atouf, F., Holdener, B.C., Mandel, G., and Kouzarides, T. (2000). The co-
repressor mSin3A is a functional component of the REST-CoREST repressor
complex. J. Biol. Chem. 275, 9461–9467.
Howley, P.M., and Lowy, D.R. (2007). Papillomaviruses. In Fields Virology,
D.M.K.a, P.M. Howley, ed. (Philadelphia, PA: Lippincott Williams and Wilkins),
Huang, Y., Myers, S.J., and Dingledine, R. (1999). Transcriptional repression
by REST: recruitment of Sin3A and histone deacetylase to neuronal genes.
Nat. Neurosci. 2, 867–872.
Jin, W., Yun, C., Kwak, M.K., Kim, T.A., and Kim, S.J. (2007). TrkC binds to the
type II TGF-beta receptor to suppress TGF-beta signaling. Oncogene 26,
Johnson, D.S., Mortazavi, A., Myers, R.M., and Wold, B. (2007). Genome-wide
mapping of in vivo protein-DNA interactions. Science 316, 1497–1502.
Krul, E.J., Kersemaekers, A.M., Zomerdijk-Nooyen, Y.A., Cornelisse, C.J.,
Peters, L.A., and Fleuren, G.J. (1999). Different profiles of allelic losses in cer-
vical carcinoma cases in Surinam and The Netherlands. Cancer 86, 997–1004.
Lahn, B.T., Tang, Z.L., Zhou, J., Barndt, R.J., Parvinen, M., Allis, C.D., and
Page, D.C. (2002). Previously uncharacterized histone acetyltransferases
implicated in mammalian spermatogenesis. Proc. Natl. Acad. Sci. USA 99,
Matsumoto, K., Ishii, N., Yoshida, S., Shiosaka, S., Wanaka, A., and Tohyama,
M. (1998). Molecular cloning and distinct developmental expression pattern of
spliced forms of a novel zinc finger gene wiz in the mouse cerebellum. Brain
Res. Mol. Brain Res. 61, 179–189.
Mazurenko, N.N., Bliev, A., Bidzhieva, B.A., Peskov, D., Snigur, N.V., Savi-
nova, E.B., and Kiselev, F.L. (2006). Mol. Biol. (Mosk.) 40, 436–447.
McGregor, L.M., McCune, B.K., Graff, J.R., McDowell, P.R., Romans, K.E.,
Yancopoulos, G.D., Ball, D.W., Baylin, S.B., and Nelkin, B.D. (1999). Roles of
trk family neurotrophin receptors in medullary thyroid carcinoma development
and progression. Proc. Natl. Acad. Sci. USA 96, 4540–4545.
E6 and E7 genes of human papillomavirus type 16 together are necessary and
Munoz, N., Bosch, F.X., de Sanjose, S., Herrero, R., Castellsague, X., Shah,
K.V., Snijders, P.J., and Meijer, C.J. (2003). Epidemiologic classification of
human papillomavirus types associated with cervical cancer. N. Engl. J.
Med. 348, 518–527.
Nakagawara, A. (2001). Trk receptor tyrosine kinases: a bridge between
cancer and neural development. Cancer Lett. 169, 107–114.
A complex with chromatin modifiers that occupies E2F- and Myc-responsive
genes in G0 cells. Science 296, 1132–1136.
Otto, S.J., McCorkle, S.R., Hover, J., Conaco, C., Han, J.J., Impey, S.,
Yochum,G.S.,Dunn,J.J.,Goodman, R.H.,andMandel, G.(2007).Anewbind-
ing motif for the transcriptional repressor REST uncovers large gene networks
devoted to neuronal functions. J. Neurosci. 27, 6729–6739.
Piboonniyom, S.O., Duensing, S., Swilling, N.W., Hasskarl, J., Hinds, P.W.,
and Munger, K. (2003). Abrogation of the retinoblastoma tumor suppressor
checkpoint during keratinocyte immortalization is not sufficient for induction
of centrosome-mediated genomic instability. Cancer Res. 63, 476–483.
Roopra, A.,Sharling, L.,Wood,I.C., Briggs, T.,Bachfischer,U., Paquette, A.J.,
and Buckley, N.J. (2000). Transcriptional repression by neuron-restrictive
silencer factor is mediated via the Sin3-histone deacetylase complex. Mol.
Cell. Biol. 20, 2147–2157.
Roopra, A., Qazi, R., Schoenike, B., Daley, T.J., and Morrison, J.F. (2004).
Localized domains of G9a-mediated histone methylation are required for
silencing of neuronal genes. Mol. Cell 14, 727–738.
Schoenherr, C.J., and Anderson, D.J. (1995). The neuron-restrictive silencer
factor (NRSF): a coordinate repressor of multiple neuron-specific genes.
Science 267, 1360–1363.
Shi, Y., Sawada, J., Sui, G., Affar, el B., Whetstine, J.R., Lan, F., Ogawa, H.,
Luke, M.P., and Nakatani, Y. (2003). Coordinated histone modifications medi-
ated by a CtBP co-repressor complex. Nature 422, 735–738.
Shi, Y., Lan, F., Matson, C., Mulligan, P., Whetstine, J.R., Cole, P.A., Casero,
R.A., and Shi, Y. (2004). Histone demethylation mediated by the nuclear amine
oxidase homolog LSD1. Cell 119, 941–953.
Singh, S.K., Kagalwala, M.N., Parker-Thornburg, J., Adams, H., and
Majumder, S. (2008). REST maintains self-renewal and pluripotency of embry-
onic stem cells. Nature 453, 223–227.
Sui, G., Soohoo, C., Affar, el B., Gay, F., Shi, Y., Forrester, W.C., and Shi, Y.
(2002). A DNA vector-based RNAi technology to suppress gene expression
in mammalian cells. Proc. Natl. Acad. Sci. USA 99, 5515–5520.
Taunton, J., Hassig, C.A., and Schreiber, S.L. (1996). A mammalian histone
deacetylase related to the yeast transcriptional regulator Rpd3p. Science
Ueda, J., Tachibana, M., Ikura, T., and Shinkai, Y. (2006). Zinc finger protein
Wiz links G9a/GLP histone methyltransferases to the co-repressor molecule
CtBP. J. Biol. Chem. 281, 20120–20128.
Wang, H., An, W., Cao, R., Xia, L., Erdjument-Bromage, H., Chatton, B.,
Tempst, P., Roeder, R.G., and Zhang, Y. (2003). mAM facilitates conversion
by ESET of dimethyl to trimethyl lysine 9 of histone H3 to cause transcriptional
repression. Mol. Cell 12, 475–487.
Zhao, J.J., Roberts, T.M., Mandel, G., Hannon, G.J., et al. (2005). A genetic
screen for candidate tumor suppressors identifies REST. Cell 121, 837–848.
Zhao, J.J., Gjoerup, O.V., Subramanian, R.R., Cheng, Y., Chen, W., Roberts,
T.M., and Hahn, W.C. (2003). Human mammary epithelial cell transformation
through the activation of phosphatidylinositol 3-kinase. Cancer Cell 3,
CDYL-G9a Regulation of REST Activity
726 Molecular Cell 32, 718–726, December 5, 2008 ª2008 Elsevier Inc.