MECHANISMS OF TRANSCRIPTIONAL REGULATION: GENE
REPRESSION BY KRAB ZINC FINGER PROTEINS AND GENE INDUCTION
BY ESTROGEN RECEPTOR beta
by
SMITHA P. SRIPATHY
Submitted in partial fulfillment of the requirements for a degree in Doctor of
Philosophy.
Dissertation advisors: David C. Schultz and Monica M. Montano
Department of Pharmacology
CASE WESTERN RESERVE UNIVERSITY
January 2009
CASE WESTERN RESERVE UNIVERSITY
SCHOOL OF GRADUATE STUDIES
We hereby approve the thesis/dissertation of Smitha P. Sripathy
for the Pharmacology degree *.
(signed)
Ruth A. Keri
(chair of the committee)
John J. Mieyal
David Samols
Monica M. Montano
(date) 07/15/08
*We also certify that written approval has been obtained for any
proprietary material contained therein.
i
Table of Contents
List of figures and tables iii
Acknowledgement viii
List of abbreviations 1
Abstract 3
Chapter I Transcriptional regulation: An overview 6
Gene repression by KRAB zinc finger proteins
Chapter II Introduction, review of literature and 17
statement of purpose
Chapter III Mechanism of transcriptional repression of 35
KRAB zfps
Chapter IV Summary and future directions 117
Gene induction by estrogen receptor beta
Chapter V Introduction, review of literature and 134
statement of purpose
Chapter VI Mechanism of transcriptional induction of 159
EpRE-genes by ERβ
Chapter VII Summary and future directions 201
Bibliography 211
ii
List of Tables and Figures
Figure I-1 Hox gene clusters regulate head-tail morphology 9
in humans
Table I-1 A minimal classification of transcriptional 11
coregulators
Figure II-1 Representation of a complex between DNA and 19
the ZIF268 protein, containing 3 zinc finger
motifs.
Figure II-2 Sequence alignment of the KRAB domain from 22
different KRAB–ZFPs
Figure II-3 Schematic representation of the TIF1 protein 26
family
Figure III-1 KAP1 is required for KRAB-mediated repression 75
Figure III-2 Stable depletion of KAP1 in HEK293 cells. 77
Figure III-3 Direct tethering of KAP1 to DNA is sufficient to 79
represses transcription
Figure III-4 The corepressor activity of KAP1 depends on its 81
interaction with HP1 and a functional PHD finger
and bromodomain
Figure III-5 Establishment of cell lines with a hormone- 83
iii
regulatable reporter transgene
Figure III-6 Characterization of cell lines with a chromatinized 85
5XGAL4-TK-luciferase transgene.
Figure III-7 Hormone dependent repression of chromatinized 87
luciferase transgenes by ERHBDTM-GAL4 fused
repressor proteins
Figure III-8 Hormone-dependent repression of the chromatin 89
template by ERHBD-GAL4-KRAB requires KAP1
Figure III-9 Hormone-dependent repression Kinetics of 91
ERHBD-GLA4-KRAB wildtype and mutant cells
Figure III-10 Hormone-dependent KRAB repression is 93
dependent on KAP1.
Figure III-11 KRAB mediated repression of a chromatin 95
template is independent of TIF1α and TIF1γ.
Figure III-12 Hormone-dependent repression of chromatin 97
templates by ERHBD-GAL4-KAP1
Figure III-13 Kinetics of hormone-dependent repression of 99
chromatin templates by ERHBD-GAL4-KAP1
wildtype and mutants
Figure III-14 Expression of the KAP1 repression machinery 101
Figure III-15 Chromatin immunoprecipitation analysis of a 103
iv
luciferase transgene repressed by ERHBD-GAL4-
KAP1.
Figure III-16 Hormone dependent repression mediated by 105
KAP1 involves hypoacetylation of histone H3 K9
of the transgenic reporter
Figure III-17 Disruption of the KAP1-HP1 interaction fails to 107
induce hormone-dependent changes in RNA
polymerase II recruitment, histone occupancy,
and histone modifications
Figure III-18 Chromatin immnoprecipitation analysis of 109
hormone-dependent changes in mediated by
ERHBDTM-GAL4-KAP1, mutated in either the PHD
finger or bromodomain, respectively
Figure III-19 Hormone-dependent KRAB repression of a 111
chromatin template requires KAP1, HP1, and
SETDB1.
Figure III-20 Transient knockdown of HP1β expression and its 113
effect on KRAB mediated repression.
Figure IV-1 Model of KRAB/KAP1 mediated repression 119
v
mechanism
Figure V-1 Metabolic activation and deactivation of estradiol 140
Figure V-2 Representation of cDNA variants that encode ERβ 151
isoforms
Figure VI-1 hPMC2 interacts directly with ERβ and is involved 182
in mediating Tamoxifen-dependent decrease in
ODD levels
Figure VI-2 Tamoxifen-dependent recruitment of coactivators 184
to the EpRE sequence of NQO1
Figure VI-3 Both ERβ and hPMC2 are required for tamoxifen- 186
mediated increase in antioxidative enzyme
expression and protection against ODD
Figure VI-4 Tamoxifen-dependent recruitment of coactivators 188
to the EpRE sequence of NQO1 requires both
ERβ and hPMC2.
Figure VI-5 Tamoxifen treatment induces increased 190
transcription of antioxidative enzyme levels.
Figure VI-6 Knockdown of PARP-1 attenuates tamoxifen- 192
dependent increase in the expression of
vi
antioxidative enzymes
Figure VI-7 PARP-1 down regulation does not significantly 194
affect basal expression levels of antioxidative
enzymes.
Figure VI-8 Ligand dependent recruitment of ERβ and hPMC2 196
to the ERE region
Table VI-1 Sequence of DNA oligos cloned into pSuper- 198
QRshRNA, pSuper-hPMC2shRNA and PCDNA-
hPMC2 569miR plasmids respectively.
Table VI-2 Primer sequences used in the various PCR 199
reactions described in Chapter VI
Table VI-3 Antibodies used in Chapter VI 200
Figure VII-1 Induction of antioxidative genes by tamoxifen 202
liganded ERβ
vii
Acknowledgements
First and foremost I would like to thank my parents who instilled in me a
sense of curiosity and wonder, encouraged me to explore the world around and
shared in my enthusiasm. These qualities led me to pursue science as a career
and were a major factor in my decision to do a Ph.D. They have been a source of constant support and encouragement all through the making of this dissertation.
I extend my sincere thanks to my first thesis advisor Dr. David Schultz for starting me out on the path of scientific thinking. I learnt to critically analyze experiments, both others and mine, and think about more than just the most obvious interpretation of a result. He had an invaluable teaching tool that was very effective in getting the message across to me, he simply taught by example.
His dedication to order, method, and attention to detail while planning and conducting experiments has made me realize that the “devil is indeed in the details”. I thank him for providing me with the foundations necessary for any student of science.
It is a pleasure to thank my current thesis advisor, Dr. Monica Montano whose support and encouragement was key to the completion of this dissertation. I thank her generosity in letting me join her lab and the independence she gave me in conducting my project. She has always been ready
viii to discuss any concerns regarding not just my dissertation work, but just about anything. Her open and hands off approach has helped me gain confidence in my ability to successfully generate and test hypotheses, a critical quality in any scientist. It has been a lot of fun being in Monica’s lab, her humor and ready sense of wit has made for many a lively conversation in the lab.
In more than 5 years that it took to complete this dissertation, I have been fortunate to meet some wonderful people who have become great friends.
I would especially like to acknowledge Bonnie Gorzelle, Ndiya Ogba and Laura O’
Donnell. Bonnie has been a very patient listener, her calmness has helped me many a times in not getting hyper over my negative results (though she is completely unaware of it). I am lucky to have a person like Ndiya for a friend.
She has been a great support for me in the lab, a very good source to discuss not only trouble shooting strategies or just about anything. I thank her for being who she is and also for the umpteen number of times she has driven me all over
Cleveland at very short notices. Her timely help has enabled me to actually focus on my experiments. Last but not the least, it is with joy that I thank Laura. She is just a great person to be around with and a good friend. I thank her for helping me seek reagents around the lab for the nth time without the slightest trace of irritation. I thank her for her constant bright smile and helping nature that endears her to everyone. Any acknowledgement of Laura is not complete
ix without the mention of her amazing cakes and cookies which were frequently to be found outside the lab.
x
List of Abbreviations
4-OHT 4-Hydroxytamoxifen
8-OHdG 8-Hydroxydeoxyguanine
APL Acute Promyelocytic Leukemia
ATRA All trans retinoic acid
CE Catechol estrogen
CE-SQ Catechol estrogen-semiquinone
CE-Q Catechol estrogen quinone
ChIP chromatin immunoprecipitation
DBD DNA binding domain
DBS DNA binding site
E2 17β-estradiol
EpRE electrophile response element
ERE Estrogen response element
ERHBD Estrogen receptor hormone binding domain
GCSh Gamma-glutamylcysteine Synthase, heavy subunit
GSTpi Glutathione-S-transferase
H3K9 Histone H3 Lysine 9
HDAC Histone deacetylase
1
HP1 Heterochromatin protein 1 hPMC2 Human homolog of Xenopus laevis prevention of mitotic
catastrophe 2
IP Immunoprecipitation
KAP1 KRAB associated protein 1
KRAB Kruppel associated box
ODD Oxidative DNA damage
PHD plant homeodomain
QR Quinone reductase
RBCC RING B-box Coiled-Coil
RING Really Interesting New Gene
SETDB1 SET domain bifurcated 1
TOT trans-Hydroxytamoxifen
TIF1 Transcription intermediary factor 1
Zfp zinc finger protein
2
Mechanisms Of Transcriptional Regulation: Gene Repression By KRAB
Zinc Finger Proteins And Gene Induction By Estrogen Receptor beta
Abstract
by
SMITHA P. SRIPATHY
DNA binding transcription factors are key players in the dynamic regulation of
distinct gene expression programs, essential for maintenance of cell physiology
and differentiation. Transcription factors either activate or repress transcription
by sequence-selective binding to cis DNA elements and subsequent recruitment
of distinct coregulator complexes.
The KRAB zinc finger proteins (zfp) are DNA binding-dependent
transcriptional repressors that are characterized by a conserved KRAB repressor
domain at their N-terminus. Here we demonstrate that the multidomain protein
KAP1 is an obligate corepressor of KRAB zfps. Mutational analysis of various
KAP1 domains coupled with transient knockdown of the KAP1 interacting
proteins, HP1 and SETDB1 reveal that, KRAB/KAP1-mediated repression is
dependent on interaction of KAP1 with HP1, a constituent of heterochromatin.
Effective repression also requires participation of the H3K9 methyltransferase
3
SETDB1 and the histone deacetylase Mi2α. ChIP analysis of promoter sequences targeted by KAP1 revealed localized recruitment of HP1, decreased histone acetylation and increased H3K9 methylation which are hallmarks of transcriptionally repressive chromatin. We conclude that KAP1 acts as a scaffolding protein to recruit histone modifying activities to KRAB zfp-targeted loci resulting in the formation of localized heterochromatin-like domains and subsequent transcriptional repression.
Previous studies from our lab demonstrated tamoxifen-dependent induction of EpRE (electrophile response element)-regulated promoters by the ligand-dependent transcription factor, ERβ. Here we demonstrate that transcriptional induction of the antioxidative genes QR, GSTpi and GCSh by TOT-
ERβ, is correlated with decrease in the levels of: carcinogenic estrogen metabolites and E2-induced oxidative DNA damage (ODD). These effects of
TOT-ERβ were dependent on the presence of a novel ERβ-interacting protein
hPMC2. ChIP analysis of the QR EpRE region revealed TOT-dependent
recruitment of a coactivator complex that included ERβ, hPMC2, PARP-1 and
topoisomerase IIβ. Recruitment of the coactivator complex and subsequent gene
induction was independent of ERα, but dependent on the presence of both ERβ
and hPMC2. Effective gene induction by TOT-ERβ was attenuated upon down
regulation of PARP-1 expression. We conclude that in the presence of tamoxifen,
ERβ and hPMC2 function together to recruit a coactivator complex to the EpRE,
4 resulting in gene induction and subsequent inhibition of E2-induced ODD in
breast epithelial cells.
5
CHAPTER I
TRANSCRIPTIONAL REGULATION: AN OVERVIEW
It makes intuitive sense to suppose that increased morphological and behavioral complexity in animals is accompanied by an increase in the number of genes. The human genome contains ~30,000 coding sequences (1, 2). A morphologically much simpler organism C.elegans codes for ~20,000 genes (3).
However, D.melanogaster which is a more complex organism than C.elegans codes for just about 14,000 genes (4). These observations clearly suggest that the actual number of genes is not an indicator of the ability to coordinate multiple gene regulatory processes required to maintain the homeostasis of a complex multicellular organism. Over the past three decades, studies aimed at understanding gene expression programs have revealed a complex network of transcription factors and cofactors that function together to coordinate various expression programs inside a cell. The central role played by DNA binding transcription factors is evident in 1) the increase in the number of transcription factors encoded and 2) increased number of cis regulatory elements that are
recognized and bound by these factors, along with increase in the complexity of
an organism (5). For example the yeast genome codes for ~300 transcription
factors while humans express greater than 3000 transcription factors (1, 6).
Added to the increased number of transcription factors themselves is the
fact that these factors do not function in isolation but rather regulate genes
6 through a combinatorial mechanism involving complex multi-protein systems.
Such cross-talk enormously expands the ability of transcription factors to coordinate multiple gene regulatory networks. Hence it is conceivable that the evolution of complex multicellular organisms was accompanied and facilitated by increased complexity of gene regulatory mechanisms (7). Over the past few years a number of sequence-specific DNA binding regulatory factors have been discovered. This is paralleled by increasing numbers of coregulators that are being discovered. The complexity is further increased by the existence of different layers of transcriptional regulation apart from a simple binding of a transcription factor to a given sequence and subsequent recruitment of RNA polymerase II machinery. The primary role of DNA binding transcription factors as effectors of gene actions in response to environmental cues is evident. They are required not only for maintenance of homeostasis in an organism but also for proper development and differentiation of various cells and tissues from a single embryonic cell into a whole organism. Finally they are dominant players in mediating gene expression programs that ultimately may be responsible for the distinct phenotypes seen among different species of organisms.
A typical example of the transcriptional regulation of basic developmental processes can be observed by studying the Hox (homeobox) gene cluster that encodes a large family of closely related transcription factors with similar DNA
binding preferences. Hox genes compose a distinct branch of the homeobox
7 gene superfamily, and have powerful functions in diversifying morphology on the
head-tail axis of animal embryos (Figure I-1, page 9). In fact, when one or more
of the Hox genes are activated in inappropriate axial positions in developing
animals, dramatic duplications of head-tail axial body structures, called homeotic
transformations, are observed (8, 9). The different HOX transcription factors are
expressed in distinct, often overlapping, domains on the head-tail body axis of
animal embryos and assign different regional fates to these axial domains. As
development proceeds, "head" HOX proteins specify the cell arrangements and structures that result in (for example) chewing organs, "thoracic" HOX proteins
specify (for example) locomotory organs, and "abdominal" HOX proteins specify
(for example) genital and excretory organs (Figure I-1, page 9). Not surprisingly,
extreme homeotic transformations are lethal at early stages of development. Hox
genes are also of great interest because there is abundant correlative evidence
that changes in Hox expression patterns and protein functions contributed to a
variety of small and large morphological changes during animal evolution (9).
Genomic analysis has revealed that HOX proteins work as a loosely coordinated
system, often with overlapping patterns of expression and function. They are
colinearly arranged and their regulation is coordinated in many animals, thus
enabling them to contribute to morphological change during the evolution of
animals.
8
These and other observations highlight the need to study the molecular mechanisms of transcriptional regulation and identify the key players involved in
Figure I-1: Hox gene clusters regulate head-tail morphology in humans
Four groups of similar Hox Genes, shown in color, appear to control related regions of the human body. Each box represents a single Hox Gene (10).
9 different gene expression programs. The ideal way of undertaking such a study would be to analyze gene expression programs in the context of an entire organism. However, such an undertaking can be highly complicated because of the sheer complexity involved. A more practical approach is to study distinct gene expression programs at a cellular level which would allow us to pinpoint the kind of expression programs triggered due to changes in the environment such as the presence of a growth factor, xenobiotic compounds and hormones. Even though studying transcriptional mechanisms at a cellular level is a lot less complex, it is still compounded by the fact that eukaryotic genomes are tightly packaged into chromatin. Variable degrees of DNA sequence accessibility exist
within chromatin throughout the cell cycle to accommodate essential biological
processes such as DNA replication, gene expression, and cell division.
Transcription factors also regulate DNA accessibility by interacting with various
coregulatory proteins that are capable of modifying chromatin structure.
Depending on the contribution of a given set of coregulators in regulating gene expression, they can be broadly grouped into five classes as shown in Table I-1
(page 11) (11). Over the past decade, studies focusing on the role of coregulators both in the activation and repression of gene expression have revealed that transcription regulatory complexes assembled by different transcription factors share many common coregulators, indicating a high degree of cross-talk. This would also enable a particular transcription factor to regulate a variety of gene expression programs by employing a combination of coregulatory
10 factors depending on the cell-type and DNA context (12). The activities of transcription factors and their coregulators are further regulated by a variety of mechanisms that include post translational modifications (acetylation, phosphorylation etc) and degradation via the proteasome machinery.
Class General properties Examples
I activator and repressor targets inherent to TAFs, TFIIA, NC2,
the core machinery, promoter recognition, PC4
and enzymatic functions
II activator and repressor adapters, OCA-B/OBF-1,
modulate DNA binding, target other Groucho, Notch,
coregulators and the core machinery CtBP, HCF, E1A,
VP16
III multifunctional structurally related but yeast: Mediator, SRBs
highly divergent coregulators: some human a: CRSP, PC2
interact with RNA Pol II and/or multiple human b:
types of activators, some also appear to ARC/DRIP/TRAP human
have inherent enzymatic functions or c: NAT, SMCC,
chromatin-selective properties Srb/Mediator
11
IV chromatin-modifying activator and CBP/p300,GCN5,
repressor adapters, acetyltransferase or P/CAF, p160s SRC1,
deacetylase activities with multiple TIF2, p/CIP, etc.),
substrates: histones, histone-related HDAC-1 and HDAC-2
proteins, activators, other coregulators (rpd3), Sir2
and the core machinery
V ATP-dependent chromatin remodeling SNF2-ATPase
activities (SWI/SNF, RSC)
and ISWI-ATPase
(NURF, ACF, ChrAC,
RSF, etc.)
Table I-1. A minimal classification of transcriptional coregulators
Nuclear Receptor superfamily of transcription factors
The physiological state of a eukaryotic cell is determined by endogenous
and exogenous signals, and often the endpoint of the pathways that interprets
these signals is gene transcription. Transcription factors modify gene expression
programs in response to these stimuli and hence are crucial for cell survival. The
nuclear receptor super family of DNA-binding transcription factors is uniquely
positioned to mediate such dynamic gene regulatory programs. They can
12 selectively bind to a variety of small molecule ligands that can cause the nuclear receptor either to activate or to repress transcription by inducing distinct conformational changes (13). As nuclear receptors regulate key physiological processes and their activities can be pharmacologically regulated, they constitute an attractive group of drug targets. Hence understanding the mechanism of nuclear receptor-mediated gene regulation, in particular, will the design of effective targeted therapies with minimal side effects for a variety of disorders, including asthma, rheumatoid arthritis and cancer. Studying the mechanism of transcriptional regulation and the effector proteins involved in this process will help us understand fundamental biological processes, such as the contribution of regulatory events to speciation, embryo development and differentiation. It will also provide us with an invaluable tool that can be applied to understand and manage conditions such as cancer which are a result of dysregulated gene expression programs.
A case in point is the successful treatment of Acute Promyelocytic
Leukemia (APL) which is characterized by the expansion of malignant myeloid cells blocked at the promyelocytic stage of hematopoietic development. Patients suffering from this cancer invariably posses reciprocal translocation of retinoic acid receptor α (RARα) giving rise to a fusion protein RARα-PML1 (Promyelocytic
Leukemia 1) (14, 15). RARα is a DNA binding transcription factor essential for proper terminal differentiation of hematopoietic lineages. Studies on the
13 transcriptional mechanism of RARα revealed that, RARα-PML interacts with and recruits the transcriptional corepressor NCOR. NCOR in turn targets histone deacetylase (HDAC) activity to sites bound by RARα-PML, thus repressing the
RARα-dependent transcription of genes required for differentiation of promyelocytic cells. However, the knowledge of RARα transactivation mechanism has enabled targeted therapy of APL with all trans retinoic acid
(ATRA) coupled with histone deacetylase inhibitors. ATRA binding shifts the
conformation of RARα such that it is unable to interact with NCOR, while blocking
the activity of HDAC prevents repression of RAR-PML targeted genes (14-16).
Another success story is the use of the antiestrogen drug tamoxifen for
estrogen receptor positive breast cancers. Estrogen receptor α (ERα) regulates
many genes via both direct and indirect binding to 5’ cis regulatory DNA
sequences (17, 18) Activation of transcriptional programs by estrogen-liganded
ERα plays a key role in mammary gland development and maintenance of bone
homeostasis. However, abnormally high expression or increased activity of ERα
results in dysregulation of ERα-mediated transcription programs and leads to
rapid accumulation of mutations and increased cell proliferation, thus
contributing to the progression of breast cancer (19, 20). Tamoxifen competes
with estrogen for the binding of ERα, and upon ERα binding induces a
conformational change that favors the recruitment of transcriptional corepressors
instead of coactivators to the ERα responsive genes in breast epithelial cells.
14
Blocking ERα-mediated transcription has been highly successful in reducing the mortality of breast cancer patients.
The Hox proteins, RARα and ERα are only a small part of the huge variety of transcription factors that function to coordinate cellular processes. Yet, studying the transcription mechanism of each factor has taken us farther in our understanding of basic developmental processes and yielded highly effective therapeutics. The study of these and other transcriptional programs has also shown us that transcription factors function by coordinating a series of events at the targeted promoter and the mechanism of coordination depends on the cell type and promoter context. Although a huge amount of information exists about various transcription paradigms, the complexity of eukaryotic combinatorial transcription programs is such that the unknown far exceeds our current knowledge. Hence there is a need to study transcription mechanisms that will help us understand key physiological processes and identify valuable drug targets.
This dissertation presents observations gained by studying two distinct transcription factors: KRAB (Kruppel associated box) zinc finger protein (zfp) and estrogen receptor β (ERβ) and their mechanisms of transcriptional regulation.
Our studies reveal the ability of both KRAB zfps and ERβ to mediate distinct transcriptional programs by recruiting either corepressors or coactivators to the
15 targeted DNA. Our studies on KRAB zfp mechanism indicates that the KRAB domain can repress transcription independent of the site of DNA it is targeted to.
A similar observation is also true for KAP1 (KRAB associated protein 1). These observations contribute to our understanding of how certain transcriptional programs can have a global impact on cellular programs, simply by the ability of the transcription factor to mediate similar transcription programs at different locations along the genome.
Our studies on ERβ-mediated transcription unveil a different paradigm where ERβ can either induce or repress transcription depending on its ligand and the DNA context. From a clinical standpoint, our studies on tamoxifen liganded-ERβ transcription argue for the development of ERβ selective agonists for the management of breast cancer. Our studies also suggest that a comprehensive result is more likely if receptor ligands are screened for both agonistic and antagonistic properties with respect to transcription from canonical and non-canonical gene promoters.
16
Gene Repression by KRAB Zinc Finger Proteins
CHAPTER II
INTRODUCTION, REVIEW OF LITERATURE AND STATEMENT OF PURPOSE
Introduction
C2H2 zinc finger family
The KRAB (Kruppel associated box) zinc finger proteins (zfp) belong to the
C2H2 superfamily of zinc finger proteins. The C-terminal region of these proteins
resembles the zinc finger structure of the Xenopus transcription factor TIFIIIA.
The superfamily is characterized by tandem repeats of a 30 amino acid motif that
folds in the presence of zinc to form a finger-like projection (21, 22). This
domain consists of a zinc ion tetrahedrally coordinated by two cysteines and two
histidines to form a ββα structure (23, 24). Coordination of the zinc ion results in
cross-linking between the strands, thus providing stability to the structure (21).
The first mammalian C2H2 zfp was identified based on sequence homology to the
Kruppel family of proteins which are characterized by a TIFIIIA like zinc finger
architecture (25, 26), and hence the superfamily is also referred to as Kruppel-
like zinc finger proteins.
The C2H2 zfps are amongst the most abundantly represented proteins in
the eukaryotic genome and most of the members belonging to this family are
17 capable of binding to DNA in a sequence selective manner (27). Analysis of crystal structures of the three fingered protein zif268, complexed to its target
DNA have shown that the amino acid residues on the α helix make contact with the base pairs on the major groove of DNA, with each finger contacting a four base pair subsite (23, 28) (Figure II-1, page 19).
The spacing of the zinc fingers is determined by the intervening number of residues between each zinc finger. A majority of the C2H2 proteins have a highly
conserved 5 amino acid sequence TGEKP (Thr-Gly-Glu-Lys-Pro), called the linker
region (29). Mutational analysis reveals that this motif is important for high
affinity DNA binding (30, 31). The linker region has a flexible structure in the free
protein, but assumes a rigid conformation upon DNA binding (32, 33). This rigid
conformation is conserved across diverse proteins and caps the C-terminus of
the preceding finger’s helix, upon binding of the finger to its cognate DNA
sequence. It is hypothesized that the flexibility of the linker region in the free
protein allows the protein to scan DNA sequences. Upon encountering the
correct sequence, the linker assumes a restricted structure and locks the zinc
fingers to the major groove of DNA, thus stabilizing the protein-DNA interactions
(29) .
18
Figure II-1: Representation of a complex between DNA and the ZIF268 protein, containing 3 zinc finger motifs.
Based on the x-ray structure of PDB 1A1L. Color coding: ZIF268: blue; DNA: orange; Zinc ions: green. (Adapted from the the open source molecular visualization tool PyMol (http://www.pymol.org/) credit: Thomas Splettstoesser).
19
However, it has not been possible to establish a one to one correlation between a given base pair on the DNA and an amino acid residue on the zinc finger. In fact synthetically engineered variants of zif268 recognize diverse sequences that are AT rich, in comparison to the GC rich region bound by zif268
(34). The side chain-side chain interaction between the finger loops has been shown to play an important part in determining context specificity of the DNA- protein contacts.
The KRAB domain
The Kruppel-like protein superfamily is further divided into subfamilies
depending on the presence of additional conserved domains. About 1/3rd of the
Kruppel-like fingers contain the KRAB domain which is characterized by a 75 amino acid conserved sequence present at the N-terminus of the zinc fingers
(35). When tethered to a heterologous DNA binding domain (DBD), KRAB acts as a transcriptional repressor (36-38). KRAB represses basal and activated transcription from a variety of RNA polymerase I and II dependent promoters, but does not seem to have any effect on transcription by RNA polymerase I (39-
41). Further, the repression is independent of the position or orientation of the
KRAB domain, relative to that of the target gene promoter (39).
The classical KRAB motif termed KRAB A, encodes a 45 amino acid core present within the 75 amino acid conserved sequence responsible for the
20 repressor activity. Mutations in the conserved residues of this motif abolish KRAB mediated repression (36). Based on sequence alignments of the KRAB domain, the KRAB zfps identified so far can be categorized as those possessing: 1) only the classical KRAB A motif 2) both KRAB A and a 32 amino acid KRAB B box 3)
KRAB A and a highly divergent KRAB b motif, termed KRAB b and 4) KRAB A and a 21 amino acid KRAB C domain (42, 43) (Figure II-2, page 22). All KRAB domains interact with a multidomain RING finger protein KAP1 (KRAB associated protein 1) and this interaction is necessary for KRAB repressor activity (44-46).
In particular, the KRAB domain directly binds to the RBCC (RING B box Coiled
Coil) domain present on the N-terminus of KAP1 (47, 48). In solution, both KRAB
A alone and KRAB AB are flexible and disordered and they assume a defined structure upon binding to the RBCC domain of KAP1 (49, 50). However, KRAB-
RBCC interaction is mediated by the KRAB A motif when it is present together with either KRAB b or KRAB C, however in case of KRAB AB interaction depends on both A and B motifs, indicating that the presence of the B motif influences the flexibility and conformation of KRAB A (49).
21
Figure II-2: Sequence alignment of the KRAB domain from different
KRAB–ZFPs. Letters in red represent conserved residues in each case (Adapted from, Peng H et al. JMB 2007).
22
Potential role of KRAB Zfps in regulating species-specific differentiation programs
The KRAB zfps make their first appearance in species representing the base of tetrapod divergence (35, 51, 52), and are vertebrate specific. The number of KRAB zfps show rapid expansion in mammalian genomes with 290 genes in mice and 423 genes in the human genome (53). Even though KRAB zfp genes occur individually across the genome, they typically occur as clusters of genes that share a high sequence similarity. Sequence analysis of human, rodent and primate genomes indicate that KRAB gene expansion involves lineage-specific duplications. Sequence divergence between different species is due to variations in the zinc finger regions, suggesting that the zinc fingers have evolved to recognize distinct DNA targets (51, 53).
Also, paralogous genes in a cluster are not typically coregulated which suggests that individual genes in a given cluster have roles in different biological processes (53). A case in point is the mouse rsl1 (regulator of sex-limitation) and rsl2 genes which are present within a cluster of 20 KRAB zfps on chromosome 13. Rsl1 represses transcription of slp (sex limited protein) and cyp2d9 in mouse liver, while Rsl2 represses mup (mouse urinary protein) transcription (54). Many independent research groups have reported differential expression of individual KRAB zfps in developmental processes such as
23 haematopoietic differentiation (55-57), chondrocyte differentiation (58) and embryonic development of the nervous system (59). But very few studies demonstrate direct regulation of specific genes by a given KRAB zfp. As mentioned earlier, a major hurdle for studying biological functions of specific
KRAB proteins has been their sheer number and high sequence similarity which makes it very hard to distinguish among individual KRAB zfps.
However the Rsl proteins and ZBRK1 (Zinc finger and BRCA1-interacting
protein with a KRAB domain 1) are among the KRAB zfps whose roles have been
relatively well characterized. As mentioned above, Rsl proteins regulate expression levels of mouse liver genes in a sexually dimorphic manner (54).
ZBRK1 recognizes and binds to sequences similar to the intron 3 region of the
Gadd45a (Growth Arrest and DNA Damage-inducible) gene and binds directly to
BRCA-1 (Breast Cancer 1) through the zinc fingers 5-8 (60-62). GADD45 is repressed by ZBRK1 in a KAP1 and BRCA-1 dependent manner and this repression is relieved upon DNA damage (61, 62). Taken together, the data so far strongly suggest a fundamental role for KRAB zfps in regulating development and differentiation programs specific to vertebrates and within vertebrates, regulation of transcriptional programs that result in both species- and gender- specific differences.
24
KAP1 is a multidomain corepressor of KRAB zfps
The only known protein that directly interacts with the KRAB domain is
KAP1. KAP1/TIF1β is a member of the TIF1 (Transcription intermediary factor 1) superfamily of proteins along with TIF1α, TIF1γ and TIF1δ. All four member of the family are characterized by a similar domain architecture consisting of individual structurally organized modules made up of highly conserved residues
(63-66) (Figure II-3, page 26). At the N-terminus is the tripartite motif (TRIM)
RBCC which consists of a RING domain, B box1, B box2 and Coiled Coil motifs and the C-terminus consists of PHD (Plant HomeoDomain) and Bromo domains.
The RING domain binds to two zinc atoms in a unique cross-braced conformation. The first zinc atom is coordinated by cys residues at positions
1,2,5 and 6 while the second one by cys residues at positions 3, 4, 7 and 8 (67-
69). It is a characteristic signature of many E3 ubiquitin ligases (70). However none of the TIF1 proteins have been reported to have E3 ligase activity except for Ectodermin, the Xenopus ortholog of TIF1γ, which ubiquitinates and
degrades
Smad4 (71). The B-box domains are also zinc containing motifs where the second Zn coordinating residue is cysteine in B1 and histidine in B2 (72) and no specific functions have been associated with these domains so far. The coiled- coil motifs are mainly involved in homo-oligomerization and cell compartmentalization (73). The RBCC domain of KAP1 interacts directly with the
KRAB domain and mutations in the conserved residues of any of the sub-
25
Figure II-3: Schematic representation of the TIF1 protein family
Amino acid numbers indicate the size of each TIF1.
26
domains disrupts this interaction, suggesting that the RBCC domain functions as a single structural unit to mediate KRAB interaction (47, 49, 74).
Solution structure analysis of KAP1 PHD domain at the C-terminus reveals that the PHD motif is structurally identical to the RING finger. The major difference is the presence of an additional hydrophobic core, in the PHD domain core compared to that of RING (75). The bromodomain at the C-terminus is a bundle of four left handed alpha helices and found in many proteins involved in chromatin remodeling. With a few exceptions, bromodomain recognizes and binds specifically to acetylated lysines on histones H3 and H4. Recognition is mediated by conserved residues that form a core hydrophobic pocket (76-79).
In addition to these well defined structural domains, all proteins of the
TIF1 family (except for TIF1γ) interact with the heterochromatin protein 1 (HP1) through a conserved penta peptide motif PxVxL (64, 80-82). It is important to note that despite having highly similar domain architectures, only KAP1 interacts with the KRAB domain, and only TIF1α with nuclear receptors (NR). Irrespective of their ability to interact with either, KRAB domains, NRs, or HP1, all TIF1s can repress transcription of RNA pol II driven promoters.
27
Role of HP1 in regulating KRAB repression
The ability of KAP1 to repress transcription is correlated with its ability to interact with HP1. Mutations in the PxVxL domain that disrupt interaction with
HP1, also result in loss of KAP1 repressor activity (80, 81). The importance of
KAP1 in development programs is evident by the fact that KAP1 knockout mice only develop until the blastocyst stage (embryonic day 5.5), but are completely resorbed by E 8.5 (83). Also, studies so far have not identified any other transcription factor apart from KRAB zfps that regulates transcription via KAP1.
This observation that KRAB domain depends on KAP1 to mediate repression suggests that embryonic lethality may be due to dysregulation of genes normally repressed by KRAB zfps during the course of embryogenesis. Observations made by independent researchers indicate that interaction between HP1 proteins and
KAP1 is a key component of KRAB domain-mediated silencing. KAP1 shows diffuse staining in undifferentiated mouse embyronic carcinoma cells (F9), but
KAP1 redistributes to centromeric heterochromatin upon differentiation of these cells to primitive endoderm-like cells (PrE). Disruption of KAP1-HP1 interaction not only disrupts relocation of KAP1, but also prevents subsequent differentiation of PrE cells into parietal endoderms or visceral endoderms (84, 85). These observations stress the biological significance of KAP1-HP1 interaction and also suggest a mechanistic basis for the corepressor ability of KAP1. HP1 is a constituent of heterochromatin and it is possible that KAP1 mediates repression
28 by repositioning of genes from a transcriptionally permissive euchromatin environment to a transcriptionally repressive heterochromatin environment which is enriched in HP1 occupancy. This suggestion is further strengthened by observations that KRAB zfps colocalize along with KAP1 to transcriptionally restrictive centromeric chromatin, but mutations in the HP1 binding domain
(HP1BD) of KAP1 disrupts centromeric association of the KRAB proteins (86).
HP1 has a key role in the formation and maintenance of heterochromatin domains
HP1 is a major constituent of heterochromatin and plays a key role in the establishment and maintenance of heterochromatin domains. Three isoforms
HP1α, HP1β and HP1γ, have been identified in humans (87-89). All three proteins share high sequence and structural homology, but differ in their distribution across the nucleus. In general, HP1α is predominantly localized to pericentromeric heterochromatin; HP1β shows diffused localization, while HP1γ is preferentially localized to euchromatic regions. The structure of HP1 is ideally suited for its role as an adapter protein with a highly conserved chromodomain
(CD) (90) and chromoshadow domain (CSD) at the N- and C-terminal ends respectively, linked together by a variable loop region (91). While the CD is a monomer, the CSD exists as a dimer and the loop region has a flexible structure, thus allowing the CD and CSD domains greater freedom in their interactions with other proteins (92, 93). This is an important structural feature, as HP1 interacts
29 with a wide variety of proteins via its CSD domain (94). The CD domain specifically recognizes and binds to di- or tri-methylated lysine 9 residues of histone H3 and acts as an adapter to recruit its interacting proteins to heterochromatin domains (95, 96). Mutations in the conserved residues of the
CD domain that disrupt methyl lysine binding also result in loss of silencing activity by HP1, thus underlining the importance of this interaction (97). It is hypothesized that chromatin condensation is achieved via HP1 binding to the methylated histones created by the activity of histone methyl transferases such as SUV39H1 (Suppressor of variegation 3-9 homolog 1). While the CD domain of
HP1 binds to the histones, the CSD domain forms homomers and recruits chromatin modifying enzymes such as histone deacetylases (example: NuRD
(nucleosomal remodeling and histone deacetylase) complex) and chromatin remodeling complexes (example: CAF150 complex). Combined activities of these enzyme complexes result in the formation of a transcriptionally non-permissive compact chromatin structure (98).
However, not all loci targeted by HP1 are repressed even when present in a heterochromatin environment and in fact certain genes require HP1 recruitment in order to be transcribed (99, 100). These observations clearly indicate that Histone H3 Lysine 9 (H3K9) methylation and HP1 binding are not the only factors that determine the transcriptional status of a gene. In concordance with these observations, many studies demonstrate that the binding
30 of HP1 to chromatin is dynamic, and methylation of H3K9 is a necessary but not sufficient for effective targeting of HP1 to a given locus. It has been shown that localization of HP1 is influenced by proteins that interact through the CSD domain of HP1 (101-104).
KAP1 coordinates histone methyltransferase and histone deacetylase activities to mediate repression
It is interesting to note that KAP1 not only interacts with HP1, but also interacts with a histone lysine 9 methyltransferase SETDB1 (SET domain bifurcated 1)(105, 106). Further, the PHD and bromodoamin of KAP1 form a cooperative unit to associate with Mi-2α, a component of the NuRD histone deacetylase complex (107-111). Hypoacetylation, enriched methylation of H3K9 and HP1 deposition are considered hallmarks of repressive chromatin (91, 112).
Repression of chromatinized reporter genes by heterologous KRAB domain targeting correlates with enrichment of KAP1, H3K9 methylation, SETDB1 and
HP1 deposition at upstream DNA binding sites (105, 113). It has been reported that transient targeting of KRAB domain to a chromatinized reporter was sufficient to repress the transcription for over 40 cell divisions.
Further, analysis of the promoter sequences from stable silenced promoters revealed increased DNA methylation levels (113). Also, in transgenic mice expressing KRAB repressible GFP reporters, GFP expression is stably
31 repressed only when KRAB-mediated silencing is induced in early embryogenesis.
This stable silencing correlates with DNA methylation of the coding regions of the reporter (114). Together, these studies indicate that the KRAB-KAP1 system plays important roles during embryo development. KAP1 is required for silencing of the M-MLV (Moloney Murine Leukemia) promoter only in embryonic cells but not in differentiated cells. Further, the KAP1 mediated silencing is dependent on its interaction with HP1 (115-117). These observations while supporting a crucial role for KRAB-KAP1 in early development also highlight the fact that KRAB-KAP1 mediated repression is cell type and developmental stage specific.
A reason for this specificity can be the involvement of developmentally restricted factors apart from HP1, SETDB1 and Mi-2α in the modulation of KAP1 activity. Another possibility is the differential expression of KRAB Zfps across cell types and developmental stages. As mentioned earlier, numerous studies report cell type and differentiation stage specific expression of KRAB Zfps, indicating dynamic regulation of this family of transcription factors in response to changing environmental cues. Interestingly, microarray analyses of DNA immunoprecipitated for KAP1 reveal ~7000 KAP1 binding sites that are enriched for KRAB Zfps, also KAP1 binding correlated with trimethylated H3K9 (118). An independent study observed that HP1 proteins form large domains across clusters of genes encoding KRAB zfps (119). Taken together these observations suggest that the KRAB-KAP1 system may be subject to autoregulation.
32
In conclusion, the unique ability of KRAB zfps to undergo adaptive mutations coupled with their ability to recruit histone methyltransferases, deacetylases and HP1, poise them to be not only effective but dynamic transcriptional regulators.
STATEMENT OF PURPOSE
The purpose of this part of the dissertation was to study the transcriptional
mechanism employed by the KRAB zinc finger proteins (zfps) to induce
transcriptional repression. KRAB zfps are potent repressors and are differentially
expressed during developmental programs. The number of genes encoding
KRAB zfps increase rapidly through vertebrates and especially in mammalian
genomes indicating that they have key roles to play in the evolution of species
specific phenotypes. However, to date very few genes have been demonstrated to be directly regulated by a KRAB zfp. Hence to study the repression
mechanism of these proteins, we used cells containing a stably integrated
luciferase reporter under the control of a GAL4 DNA binding site. This was
coupled with constitutive exogenous expression of a heterologous KRAB domain
that could be induced to bind to the GAL4 DBS via treatment with 4-Hydroxy-
tamoxifen (4-OHT).
We hypothesized that KAP1 is an obligate corepressors of the KRAB
domain and tested this hypothesis by testing for KRAB mediated repression
under stable KAP1 knockdown conditions. We analyzed the role of KAP1
33 towards transcriptional repression by directly tethering an RBCC domain truncated version of KAP1 to a GAL4 DNA binding domain to target it to GAL4
DBS. Like our KRAB construct KAP1 targeting could be induced by OHT treatment. Analysis of the roles played by individual domains of KAP1 was achieved by studying the ability of different mutant versions of KAP1 to mediate transcriptional repression. These were point mutations known to disrupt the HP1 binding domain, the PHD domain and the bromo domains of KAP1, respectively.
The molecular basis of KRAB-KAP1 mediated repression was analyzed by conducting chromatin immunoprecipitations (ChIP) at the GAL4 DBS to study 4-
OHT induced recruitment of KRAB and KAP1. The contribution of KAP1 interacting proteins HP1, SETDB1 and Mi2α was analyzed by comparative ChIP analysis of the protein recruitment patterns effected by the wild type and mutant
KAP1 proteins. HP1 and SETDB1 were also transiently down regulated in order to determine the importance of their role in mediating KRAB domain repression.
34
CHAPTER III
The KAP1 corepressor functions as a molecular scaffold to establish microenvironments of HP1 heterochromatin required for KRAB zinc finger protein mediated transcriptional repression
(This work has been published in the journal Molecular and Cellular Biology,
Sripathy SP, Stevens J and Schultz DC, 2006. 26:8623-38)
Introduction
Genetic and epigenetic programs that control proper spatial and temporal patterns of gene expression are instrumental for pluripotent stem cells to determine cellular identity and maintain homeostasis of adult metazoans. Though historically viewed as a passive packaging unit, remodeling of chromatin structure has emerged as a key target for programming of gene expression during early embryogenesis and tissue-specific gene expression.
The dynamic regulation of chromatin organization appears to be accomplished in part by at least four families of proteins, including: 1) macromolecular protein complexes that utilize energy from ATP hydrolysis to disrupt DNA: protein interactions; 2) proteins with intrinsic enzymatic activity to postranslationally modify the core histones; 3) non-histone chromosomal proteins; and 4) histone variants. Increasing experimental evidence indicates that the combinatorial use of histone variants, histone modifications, including acetylation, phosphorylation,
35 ubiquitination, and methylation, and non-histone chromatin-associated proteins that recognize these signals, represent an epigenetic marking system responsible for setting and maintaining heritable programs of gene expression (36, 95,
120, 121). However, several key questions in understanding this indexing system include: 1) how are histone modifications and variants targeted to gene specific regulatory elements, 2) what are the patterns of modifications at transcriptionally silenced loci, 3) how do patterns of modifications temporally change during active transcriptional silencing of gene expression, 4) what non-histone chromosomal proteins interpret this code, and 5) how is recognition of this code mechanistically translated into a change in gene activity?
TFIIIA/C2H2 containing zinc finger proteins represent the most abundant
family of sequence specific DNA binding proteins in higher eukaryotes (1). These proteins are characterized by a repeating three dimensional structural motif of a b-hairpin, followed by an α-helix, which is stabilized by the coordination of one zinc ion (122). Concatamers of two or more zinc finger motifs facilitate selective, high affinity binding to DNA with each finger module making specific contacts with a 3-5 base pair sub site in the major groove of double stranded
DNA (23). This family of DNA binding proteins can be further subcategorized by the presence of additional amino acid motifs. The KRAB
(Kruppel Associated Box)-zinc finger protein family represents nearly 1/3 of all
TFIIIA/C2H2-type zinc finger proteins encoded by the human genome (>200 of
36
~750) (1, 123). In addition to tandem arrays of C2H2 zinc finger motifs, this
subfamily is defined by a set of highly conserved amino acids commonly
referred to as the KRAB domain (43). Comparative genome analyses
indicate that this gene family is vertebrate specific and has rapidly
expanded across the vertebrate lineage, as the repertoire of KRAB-zfps differs
significantly between species, suggesting that KRAB-zfps regulate programs of
gene expression that contribute to speciation (45, 124). The KRAB domain,
defined by approximately 75 amino acids, is a transferable module and possesses
DNA binding-dependent transcriptional repression activity. This activity is common to all KRAB domains of independent zinc finger proteins that have been tested and can be disrupted by mutations at highly conserved amino acids that define the minimal KRAB domain consensus sequence (36, 38, 113, 125).
These data emphasize that the transcriptional repression activity associated with the KRAB domain is a common biochemical property of this motif, and that repression of transcription is achieved via a common mechanism or through a universal cofactor. Moreover, the abundant representation of KRAB-zfps in vertebrates potentially makes KRAB-directed transcriptional regulation one of the most widespread sequence specific mechanisms to repress gene transcription in higher eukaryotes.
Mechanistically, transcriptional repression by the KRAB domain correlates with binding to the KAP1 protein, also referred to as TIF1β r KRIP1
37
(KRAB Interacting Protein 1) (1, 45, 66, 86). The role of KAP1 in KRAB domain repression is supported by several pieces of experimental data, including: 1) KAP1 binds to multiple KRAB repression domains both in vitro and
in vivo; 2) KRAB domain mutations that abolish repression decrease or eliminate
the interaction with KAP1; 3) Exogenous expression of KAP1 enhances KRAB-
mediated repression; and 4) KAP1 directly tethered to DNA is sufficient to
repress transcription (40, 48, 80, 86, 126, 127). Despite these observations, it is
unclear whether other cellular proteins exist that are either necessary and/or
sufficient to mediate the repression activity of the KRAB domain.
The primary amino acid sequence of KAP1 reveals the presence of several
well conserved consensus signature motifs, including a RING finger, B
boxes, a coiled-coil region, a PHD finger, and a bromodomain (1, 45, 66, 86).
This spatial arrangement of motifs is the proto-type for the family of
transcriptional regulators that includes TIF1α, TIF1γ, TIF1δ, and Bonus (82, 112,
128, 129). Biochemical analyses of the RING finger, B-boxes, and coiled-coil,
collectively referred to as the RBCC/Trim domain indicate that this tri-partite
motif is both necessary and sufficient for homo-oligomerization and direct
binding to the KRAB repression module. Furthermore, the RBCC region of KAP1
is the only member of the TIF1 family that directly binds to the KRAB domain
(47, 48, 66, 74, 130). KAP1 also displays several biochemical properties that
suggest it functions as a molecular scaffold to coordinate activities that
38 regulate chromatin structure, including: 1) Interaction with Mi-2α, a core component of the multisubunit NuRD histone deacetylase complex (107);
2) Interaction with the histone H3 lysine 9 selective methyltransferase, SETDB1
(105); and 3) direct interaction with the chromoshadow domain of the heterochromatin protein 1 (HP1) family via a core PxVxL motif (HP1BD) in vitro and in vivo (80-82). The biological significance of the KAP1-HP1 interaction is
highlighted by observations in F9 cells where KAP1 associates with
heterochromatin in a PxVxL dependent manner upon induction of cellular
differentiation (84). Furthermore, the KAP1-HP1 interaction is required for
differentiation of F9 cells into parietal endoderm like cells in vitro (85).
Moreover, transcriptional repression of a chromatinized reporter gene by a heterologous KRAB repressor protein correlates with localized enrichment of
KAP1, SETDB1, HP1, and methylation of histone H3 lysine 9 at promoter sequences of the transgene (105, 113). Based on these data we hypothesize that KRAB-zfps require KAP1 and the network of proteins that interact with KAP1 to establish localized microenvironments of heterochromatin at gene-specific loci to repress gene transcription.
Our current model of transcriptional repression by KRAB-zfps is largely defined based on a network of biochemical interactions between
KAP1 and proteins with previously described roles in chromatin metabolism.
Previous studies have shown that mutations in the HP1BD/PxVxL motif, PHD
39 finger, and bromodomain of KAP1 that disrupt protein-protein interactions with
HP1, Mi-2a, and SETDB1, respectively, correlate with attenuated KAP1 repression. These data are consistent with the hypothesis that KRAB-mediated repression is dependent upon KAP1 and the network of proteins that associate with KAP1. However, the interpretation of these data is limited by the fact these experiments were done exclusively in transient transfection based reporter assays. Furthermore, many of these conclusions were drawn from the use of minimal peptides of KAP1 that function as autonomous repression domains when tethered directly to DNA. However, these data do not address whether the network of proteins that interact with KAP1 function cooperatively during KRAB mediated transcriptional repression, especially in the context of a chromatin template. Here, we use hormone responsive repressor proteins to genetically interrogate the requirement of KAP1 in KRAB mediated repression of stably integrated reporter transgenes. Using a combination of mutant KAP1 proteins and transient depletion of putative effector proteins of KAP1 repression by siRNA, we investigated the functional roles of
KAP1, HP1, and SETDB1 in the de novo establishment of a localized domain of
heterochromatin at a chromatinized transgene repressed by the KRAB repression
module.
40
RESULTS
The KAP1 corepressor is required for KRAB-mediated repression
Although KRAB-mediated repression of transcription correlates with KAP1
binding, the requirement for KAP1 has not been demonstrated. To address this
question, we have stably suppressed the expression of cellular KAP1 by 80 to
90% in HEK293 cells by constitutively expressing a shRNA directed against the
coding sequence of the KAP1 mRNA (Figure III-1A, Figure III-2A, pages 75 and
77). Stable depletion of KAP1 appeared to have little effect on cellular
morphology, viability, or the steady-state levels of proteins previously described
to interact with KAP1 (Figure III-2B, page 77). To test the requirement of KAP1 in
KRAB-mediated transcriptional repression, we used these cells in a transient-
transfection reporter assay with a 5XGAL4-TK-luciferase reporter plasmid and a
plasmid that expresses a GAL4-KRAB repressor protein (Figure III-1B, page 75).
In parental HEK293 cells, we observed a dose-dependent relationship between
the amount of plasmid, expressing GAL4-KRAB, transfected and the absolute
level of transcriptional repression. In contrast, KRAB repression was significantly
reduced in two independently isolated KAP1 knockdown cells (Figure III-1C, page
75). This observation is consistent with the hypothesis that KAP1 is an essential
cellular factor for KRAB-mediated transcriptional repression.
41
Previous site-directed mutagenesis studies have identified key residues that are required for protein-protein interactions between KAP1 and HP1, Mi-2 ,
and SETDB1 (80-82, 105, 107). These mutations also reduce the repression
activity of a heterologous KAP1 protein (amino acids 293 to 835) when tethered
to DNA in transient-transfection based reporter assays (45, 107) (Figure III-3, page 79). To test the effect of these mutations on the corepressor activity of full-
length KAP1, we exogenously expressed alleles of KAP1 in knockdown cells that
are refractory to the shRNA and possess deleterious amino acid substitutions in
the HP1BD, PHD finger, and bromodomain (Figure III-4A, page 81). As illustrated in Figure III-4B (page 81), expression of wild-type KAP1 complemented the
repression defect in the KAP1 knockdown cells. This observation suggests that
the repression defect observed in the stable KAP1 knockdown cells is unlikely to
be the result of an off-target effect of the shRNA. Although expressed at near-
equal levels or higher, exogenous expression of a KAP1 protein unable to interact with HP1 (RV487, 488EE) was incapable of restoring wild-type levels of
repression (Figure III-4B and C, page 81). This observation implies that the
interaction between KAP1 and HP1 is essential for KAP1 corepressor activity.
Expression of proteins with mutations in either the PHD finger or bromodomain
partially complemented the repression defect of the stable KAP1 knockdown
cells. The combination of these observations suggests that repression of gene
transcription by KRAB-zfps depends on the network of proteins that directly
interact with KAP1, especially HP1.
42
Establishment of GAL4-TK-luciferase cell lines
Our data indicate that KAP1 is an essential factor required for
transcriptional repression by the KRAB-zfp superfamily. Therefore, it was
important to investigate structure-function relationships of the HP1BD, PHD
finger, and bromodomain of KAP1 in transcriptional repression of a chromatin
template. In the absence of well-characterized endogenous genes regulated by
KRAB-zfps, we developed a modification of our previous strategy to conditionally
regulate transcription of a chromatinized reporter transgene (105, 113). As
schematically represented in Figure III-5A (page 83), we employed a sequential
transfection strategy to establish cells that possess a stably integrated GAL4-TK-
luciferase transgene and a hormone-responsive repressor protein. The advantage
of this approach is that we can directly compare the consequences of mutations
in KAP1 on regulation of gene transcription at an isogenic locus.
We initially established cell lines possessing a stably integrated copy of the
5XGAL4-TK-luciferase reporter (Figure III-1B, page 75). Basal expression of this transgene is driven by a minimal herpes simplex virus thymidine kinase promoter,
which can be regulated by heterologous proteins that bind GAL4 DNA binding
sites positioned immediately 5' to the TK promoter sequences. DNA for the
p5XGAL4-TK-luciferase plasmid and a second plasmid conferring puromycin
resistance were cotransfected into HEK293 cells. We chose this two-plasmid
transfection approach so that we could select for cells whose luciferase
43 transcription was not influenced by the promoter activity driving puromycin
resistance during the selection process. Approximately 100 puromycin-resistant
cell colonies were isolated and expanded. Southern blot analysis of genomic DNA
isolated from a representative selection of clones indicated that the copy number
of the transgene ranged from 1 to 10 copies, depending on the cell line, and
integrated at multiple independent loci (Figure III-6B, page 85). The normalized
basal luciferase activities of the individual clones varied from 5 to 200 light
units/optical density unit of protein (see Figure III-6C, page 85). The variability
of basal luciferase expression in each cell clone could be a direct reflection of the transgene copy number incorporated into the host cell genome or its site of
integration. Regardless, the detection of stable luciferase activity in these clones
is indicative of the transgene integrating in a transcriptionally permissive
euchromatic environment.
Hormone-regulated repressor proteins
In order to conditionally regulate the transcription of the chromatinized
luciferase transgene, we engineered our GAL4-KAP1 repressor protein to be
hormone regulated by fusing a tamoxifen-sensitive derivative of the estrogen receptor hormone binding domain (ERHBD) to the N terminus of the GAL4 DNA binding domain (Figure III-5B, page 83). The ERHBD is 1,000-fold less
responsive to serum estrogens and contains no intrinsic transcriptional activation
potential (131). Unlike other conditional expression systems that are
44 transcriptionally controlled, our chimerical repressor proteins are constitutively
expressed. The transcriptional regulatory activity of these proteins is posttranslationally controlled by the addition of 4-OHT to the tissue culture
medium. Hormone-regulated GAL4-KAP1 fusion proteins were created for the wild-type KAP1 sequence (amino acids 293 to 835) and mutations in the HP1BD
(RV487, 488EE), PHD finger (W664A), and bromodomain (L720A, F761A). Each
of these mutations disrupts the tertiary structure of these modular domains, which significantly affects the ability of KAP1 to interact with HP1, Mi-2 , and
SETDB1, respectively, and attenuates KAP1-mediated transcriptional repression
(Figure III-5B, page 83). Thus, this set of KAP1 mutant repressor proteins
enabled us to comprehensively investigate the functional role of these different
domains in KAP1-mediated regulation of transcription and chromatin structure. As
a control, an ERHBD-GAL4-KRAB protein was engineered which contained the 90-
amino-acid KRAB domain of Kox1. This minimal domain is sufficient to bind KAP1
and is a very potent, DNA binding-dependent repressor in vivo (105, 113, 132).
Individual subclones of stable GAL4-TK-Luc cells were transfected with plasmid DNA encoding either the ERHBD-GAL4-KRAB or ERHBD-GAL4-KAP1
proteins (Figure III-5A, page 83). For each repressor protein introduced, we
isolated 5 to 10 independent clones that demonstrated 4-OHT-dependent
repression of the chromatinized reporter (as described in ‘Materials and Methods’
section, page 69). Repression of the transgene's expression by the wild-type
45
ERHBD-GAL4-KRAB and ERHBD-GAL4-KAP1 repressor proteins was dependent on
the concentration of 4-OHT, with maximal effects being reached between 125
and 250 nM (Figure III-7A, page 87). Reduced luciferase activity also correlated
with a reduction in steady-state levels of the transgene's mRNA. At 48 hours post
OHT treatment, the protein and mRNA levels in the ERHBD-GAL4-KRAB clones
were approximately 14 fold and 12 fold repressed respectively, while the protein
and mRNA expression in the ERHBD-GAL4-KAP1 clones were 4 fold and 3 fold
repressed respectively. (Figures III-7A and III-7B, page 87). Further, chromatin
immunoprecipitation experiments with anti-GAL4 immunoglobulin G
demonstrated that repression of the transgene was tightly associated with 4-
OHT-induced DNA binding of the repressor proteins to the GAL4 DNA binding
sites and the HSVTK promoter sequences (Figure III-15, page 103).
Hormone-dependent repression of chromatin templates by ERHBD-
GAL4-KRAB is dependent upon KAP1
To study temporal characteristics of transcriptional repression by the
KRAB-KAP1 complex, we incubated cells expressing ERHBD-GAL4-KRAB with growth medium containing either 0.1% ethanol or 500 nM 4-OHT for various
amounts of time. Approach to steady-state repression of the chromatinized
transgene was observed between 72 and 96 h of continuous 4-OHT treatment
(Figure III-8A, page 89). Cells expressing a mutant KRAB domain (DV18, 19AA)
which lacked the ability to bind KAP1 failed to repress the chromatinized
46 transgene at any time point (Figure III-9, page 91). To investigate the role of
KAP1 in KRAB-mediated repression of chromatin templates, we developed an
experimental strategy to transiently deplete KAP1 by siRNA transfection (Figure
III-8B, page 89). The rationale for this scheme was to ensure sufficient
knockdown in steady-state levels of KAP1 prior to treating with 4-OHT. Similar to
the transfection-based reporter assays illustrated in Figure III-4 (page 81),
transient depletion of KAP1 by siRNA significantly crippled hormone-dependent
KRAB-mediated repression of the chromatinized reporter (Figure III-8C and 8D,
Figure III-10, pages 89 and 93). In general, the extent to which KRAB-mediated
repression was inhibited correlated with the efficiency of KAP1 depletion by the siRNA transfection. Furthermore, transient depletion of the highly related TIF1 and TIF1 proteins by siRNA transfection had no effect on hormone-dependent
KRAB-mediated transcriptional repression of a chromatin template (Figures III-10
and III-11, pages 93 and 95).
Direct tethering of KAP1 to a chromatin template is sufficient to
repress transcription
Our data indicate that KAP1 is an essential cellular factor for KRAB-
mediated transcriptional repression. To investigate whether directly tethering
KAP1 to DNA was sufficient to repress transcription of a chromatin template, we
incubated cells expressing ERHBD-GAL4-KAP1 with growth medium containing
either 0.1% ethanol or 500 nM 4-OHT for various amounts of time. In contrast to
47 the ERHBD-GAL4-KRAB repressor protein, direct tethering of KAP1 to DNA demonstrated lower levels of absolute repression but more rapid kinetics of
transcriptional repression, approaching steady-state repression within 48 to 72 h following 4-OHT treatment (Figure III-12A page 97). Furthermore, we observed
discrete differences in the absolute level of repression and subtle variations in the
kinetic patterns of repression for the ERHBD-GAL4-KAP1 repressor protein in the
different clones analyzed. We attribute these variations to differences in the
expression levels of the repressor proteins and/or the transgene's integration
site. Regardless, the combination of these data validate this model system as a
tool to further investigate the mechanism by which KAP1 coordinates changes in
histone modifications (i.e., histone deacetylation, methylation, etc.) and
deposition of HP1 proteins to alter chromatin structure and repress transcription
of a highly transcribed gene.
The HP1 interaction with KAP1 is required for repression of
chromatinized reporter transgenes
An advantage of our experimental system is that we can analyze the consequences of biochemically well-defined amino acid substitutions in KAP1 on
the transcriptional repression of a chromatinized reporter transgene, positioned
at an isogenic locus. To investigate the requirements of the HP1BD, the PHD finger, and bromodomain of KAP1 and their associated activities in the
transcriptional repression of a chromatin template, clone 12 cells (Figure III-6,
48 page 85) were stably transfected with plasmid DNA encoding the ERHBD-GAL4-
KAP1 repressor containing the RV487,488EE, W664A, L720A, or F761A mutation.
We isolated between 5 and 10 independent antibiotic-resistant cell clones that expressed either the wild-type or each mutant KAP1 repressor protein (Figure III-
5C, page 83). Two representative cell clones for each mutant protein were grown
in growth medium containing either 0.1% ethanol or 500 nM 4-OHT for 96 h in
order to characterize functional consequences associated with these specific
mutations on KAP1-mediated transcriptional repression. As illustrated in Figure
III-12C (page 97), two independent cell clones expressing wild-type ERHBD-
GAL4-KAP1 displayed a steady state of 10- to 14-fold repression. In contrast, the
cell clones expressing mutant forms of KAP1 demonstrated significantly attenuated levels of transcriptional repression, despite the expression levels of these mutant proteins being equal to or greater than the wild-type protein
(Figure III-12C, page 97). Temporal analysis of hormone-dependent transcriptional repression in these cell clones indicated that the difference in
absolute repression between the wild-type repressor protein and the various
mutant proteins was not the result of delayed kinetics (Figures III-13A and III-
13B, page 99). Furthermore, decreased repression activity appears to be intrinsic
to the mutant KAP1 repressor proteins, as Western blot analyses revealed little
variation in HP1 , HP1ß, HP1 , Mi-2 , and SETDB1 expression levels between the
different clones (Figure III-14, page 101). Overall, these data demonstrate a
49 fundamental requirement for these domains and their associated activities in
KAP1-mediated transcriptional repression of a chromosomally integrated target.
Hormone-dependent repression by KAP1 correlates with reduced
recruitment of RNA polymerase II and dynamic changes in histone tail
modifications
To determine molecular events that correlate with 4-OHT-induced
transcriptional repression of the chromatinized transgene by ERHBD-GAL4-KAP1,
we did chromatin immunoprecipitation experiments. To define spatial
relationships between histone modifications and specific DNA sequences within the transgene, we analyzed four loci along the transgene (Figure III-15A, page
103). Following a 96-h incubation with 4-OHT, we observed a hormone-
dependent enrichment (four- to sixfold) of the ERHBD-GAL4-KAP1 protein at
sequences that overlapped with the GAL4 DNA binding sites, HSVTK promoter,
and the transcription start site (Figures III-15B and III-16, pages 103 and 105).
Binding of the repressor protein was coincident with a fourfold reduction in hypophosphorylated RNA polymerase II occupancy at proximal regulatory
sequences. Analysis of total histone H3 revealed a twofold hormone-dependent
increase in histone H3 occupancy throughout the reporter transgene, coupled with a concomitant decrease in acetylated H3 K9/K14. Analysis of site-specific
histone H3 methylation (i.e., K4, K9, K27, and K36) indicated that both dimethyl and trimethyl H3K4 were reduced by 2-fold at promoter sequences following 4-
50
OHT treatment. Moreover, we observed enrichment of dimethyl histone H3K9 (2-
fold) and trimethyl histone H3K9 (2.5- to 6-fold), H3K27 (2.5- to 6-fold), H3K36
(2- to 4-fold), and histone H4K20 (3-fold) (Figure III-15B, page 103). When we
analyzed sequences 1.5 kbp and 2.5 kbp distal to the transcription start site in
the 3' coding region of the luciferase mRNA and polyadenylation signal sequence,
respectively, we observed a progressive reduction in trimethyl H3K4, H3K36, and histone H4K20 associated with a transcriptionally repressed transgene. In
contrast to proximal promoter sequences, the extent of hormone-dependent
changes in di- and trimethyl H3K9 and trimethyl H3K27 levels was less dramatic
in nucleosomes positioned in this region of the reporter transgene.
Immunoprecipitations with antisera against HP1 , HP1ß, and HP1 revealed a
twofold hormone-dependent enrichment of HP1 and HP1ß at promoter
sequences of the transgene (Figures III-15B and III-16, pages 103 and 105). We also observed increased binding of the histone H3 lysine 9-selective
methyltransferase SETDB1 to promoter sequences following treatment with 4-
OHT, which was coincident with the elevated levels of trimethyl H3K9 we
detected in this region. These data indicate that direct tethering of the KAP1
corepressor protein to a chromatinized reporter transgene is sufficient to
coordinate dynamic changes in histone modifications that support the
recruitment and deposition of HP1 proteins to form a localized heterochromatin-
like environment that blocks the recruitment of RNA polymerase II.
51
Our data indicate that disruption of the interaction between KAP1
and HP1 cripples the corepressor activity of KAP1. Thus, to begin to understand
at a molecular level the consequences of this mutation on 4-OHT-induced
changes in the chromatin structure of the transgene, we did chromatin
immunoprecipitation experiments in cells that express ERHBD-GAL4-KAP1
(RV487, 488EE). As illustrated in Figure III-17 (page 107), we observed a
hormone-dependent increase in the amount of GAL4-KAP1 repressor protein
bound to promoter sequences of the transgene. Despite the recruitment of the mutant repressor protein to the transgene's promoter, we did not observe any
decrease in hypophosphorylated RNA polymerase recruitment to promoter
sequences. Similarly, we observed very little 4-OHT-dependent change in any of the histone modifications we examined. Most striking was the absence of
hormone-induced enrichment of trimethyl H3K9, H3K27, H3K36, or histone
H4K20 at promoter sequences. Furthermore, we did not observe hormone-
dependent binding of either SETDB1 or any of the HP1 isoforms to promoter
sequences, a result that is consistent with the lack of hormone-dependent
enrichment of trimethyl H3K9 (Figure III-17B, page 107). ChIP analyses in cells
expressing either a PHD finger mutant (W664A) or bromodomain mutant (F761A)
ERHBD-GAL4-KAP1 protein yielded nearly identical results as the HP1BD mutant
protein (Figure III-18, page 109). In summation, our ChIP data indicate that the binding of HP1 to KAP1 is necessary to induce changes in patterns of histone
modifications that correlate with KAP1-dependent repression of transcription.
52
Moreover, these data are consistent with a role for KAP1 in de novo assembly of
highly localized microenvironments of HP1-demarcated heterochromatin.
HP1 and SETDB1 are required for KRAB-mediated repression of
chromatin templates
Our data demonstrate that KRAB repression is dependent upon KAP1 and
the network of proteins that interact with the HP1BD, PHD finger, and
bromodomain of KAP1. To test the role of known KAP1-interacting proteins in
hormone-dependent KRAB repression of a chromatinized transgene, we
transiently depleted KAP1, HP1 , HP1ß, HP1 , and SETDB1 using a siRNA
approach (Figure III-8B, page 89). Western blot analysis of protein extracts from
cells transfected with siRNAs targeting the mRNAs of HP1 , HP1ß, HP1 , and
SETDB1 indicated that the expression of these proteins was depleted by 75%
(Figures III-19 and III-20, pages 111 and 113). Interestingly, we observed a
slight reduction in the expression of HP1 in KAP1 knockdown cells, too. In
contrast to the reduction of cellular levels of KAP1, depletion of each HP1 isoform
individually resulted in little effect on KRAB repression, suggesting that the HP1 proteins are redundant in terms of function with KAP1 (Figures III-19B, page
111). However, simultaneous depletion of all three HP1 isoforms resulted in a
greater-than-50% loss of KRAB repressor activity. We observed a similar effect on hormone-dependent KRAB-mediated repression in cells where SETDB1 was
transiently depleted. Collectively, these genetic data support our biochemical data
53 and further suggest that the HP1 proteins and SETDB1 have a fundamental role
in site-specific regulation of chromatin structure and transcriptional repression by
the KRAB-zfp-KAP1 repressor-corepressor complex.
DISCUSSION
The KAP1 protein fulfills several important criteria that define it as a corepressor protein for the KRAB-zfp superfamily of transcriptional repressors.
However, the abundance of endogenous KAP1 has hindered the ability to
investigate its dependence in KRAB-mediated repression. Here we have
demonstrated that KRAB-mediated repression of both transiently transfected and
chromatinized reporter transgenes was attenuated in cells where the endogenous
level of KAP1 was reduced between 50 and 90%. Independent siRNAs against
KAP1 inhibited KRAB repression to varying extents. The extent of repression
directly correlated with the levels of KAP1 knockdown achieved by the siRNAs.
Furthermore, reexpression of the wild-type KAP1 protein in stable knockdown
cells complemented the defect in KRAB repression. These data strongly argue
against the repression defect arising completely from an off-target effect of the
siRNAs. In addition, transient depletion of TIF1 and TIF1 did not affect KRAB- mediated repression. This observation is consistent with in vitro biochemical
experiments demonstrating selective interaction between the KRAB repression
module and KAP1/TIF1ß (48, 66, 74, 130). Although we cannot rule out that
depletion of KAP1 from cells does not directly or indirectly affect the levels of
54 known and unknown cellular proteins that cooperate with KAP1 to optimally
repress transcription, the combination of these data is consistent with the
conclusion that KAP1 is an essential cellular factor necessary to repress
transcription by KRAB-zfps.
To further study the role of KAP1 and KAP1-interacting proteins in
mediating transcriptional repression of a chromatin template, we investigated
regulation of a stably integrated GAL4-responsive TK-luciferase transgene by
hormone-responsive GAL4-KRAB and GAL4-KAP1 repressor proteins, respectively.
In contrast to previous studies that have utilized a similar experimental strategy
(113), we first created a series of cell lines that stably express luciferase from a randomly integrated transgene. Subsequently, we transfected these cells with
plasmids that lead to stable expression of either wild-type or mutant repressor
proteins. This particular approach enabled us to study the effects of site-directed
mutations in KAP1 on its function as a transcriptional repressor within the context
of an isogenic, chromosomal locus.
Although direct tethering of KAP1 to a chromatin template is sufficient to
rapidly repress transcription, the absolute level of steady-state repression is
substantially less compared to tethering a heterologous KRAB repressor protein.
Thus, the collection of our data indicates that KAP1 is necessary but may not be
sufficient for KRAB repression. We speculate that the reduced efficiency of the
55 heterologous KAP1 repressor protein may be a consequence of the fact that endogenous KAP1 is a trimer in solution (47) and that this native oligomerization
state is not maintained by the ERHBD-GAL4-KAP1 protein. Furthermore, it is possible that in addition to facilitating oligomerization and the direct interaction
between KAP1 and the KRAB repression module, the RBCC/TRIM domain may
bind to additional cellular factors that are required for optimal levels of KAP1-
mediated corepression of transcription. Future studies will be needed to
determine whether the RBCC/TRIM domain of KAP1 contributes to transcriptional
repression beyond simple recognition of the KRAB domain. Alternatively, our data
suggest that KRAB-mediated repression results from the additive nature of a very
rapid (Figure III-12A, page 97) KAP1-dependent mechanism and a slower KAP1-
independent mechanism.
Previous studies have defined several KAP1 polypeptides that have the ability to autonomously repress transcription when directly tethered to DNA via a
heterologous DNA binding domain. However, the importance of these repression
domains in the context of the full-length KAP1 protein, and also their role in
regulating transcription of a chromatin template, has not been studied. Data from
transient-transfection reporter assays suggest that the repression mechanisms of the PHD finger/bromodomain and the HP1BD may be additive. Alternatively, these domains may work independently of one another and the different
functions of these domains may be invoked depending on the nature of the
56 target or the cell type. Our data demonstrate an obligate role for the interaction
between KAP1 and HP1 in KRAB-KAP1 repression. In contrast, mutations in the
PHD finger and bromodomain, respectively, display quite different results
depending on the context of the assay. In transient-transfection reporter assays,
mutations in either the PHD finger or bromodomain mildly impair KAP1-
dependent repression relative to the wild-type protein but do not ablate its function like the HP1BD mutation. However, our data demonstrate an essential
role for these domains in KAP1-mediated transcriptional repression of chromatin
templates. In fact, mutations in these domains appear to be epistatic (PHD and
bromodomain mutations mask the phenotype of HP1BD mutation) with the
HP1BD mutation in KAP1 repression. These observations are not entirely surprising, given that these motifs are almost exclusively found in proteins that have a role in regulating chromatin structure and function (133, 134) and have
been shown to bind specific posttranslational modifications of the histone
proteins (76-78, 129, 135-137). Therefore, one might predict that mutations in
these domains may have a more pronounced effect on the transcriptional
regulation of a chromatin template. Further insights into the functions of these domains in KAP1-directed transcriptional repression will be dependent upon
defining the specificity of the potential interactions these domains have with
epitopes on histones, nucleosomes, or higher-order chromatin structure and the identification of native target genes regulated by KRAB-zfps.
57
Understanding of transcriptionally silent chromatin assembly has been
largely limited to studies of cytologically defined heterochromatin in
Saccharomyces cerevisiae, Drosophila melanogaster, and mammalian X-
chromosome inactivation (138-141). Thus, how heterochromatin domains are
formed and how they function to repress transcription in euchromatin loci remain
important questions. An advantage of our experimental system is that we can
induce transcriptional silencing of a well-defined, highly transcribed transgene
embedded in a chromatin environment. Therefore, our system has great utility to
address fundamental questions regarding targeted gene silencing in time and
space. Our ChIP data indicate a reduced steady-state level of
hypophosphorylated RNA polymerase II at promoter sequences of a repressed
transgene, suggesting that recruitment of RNA polymerase II has been altered.
In S. cerevisiae, the formation of heterochromatin does not exclude the binding
of preinitiation complex components to transcriptionally silenced genes but rather
appears to attenuate productive initiation and/or elongation of transcription by
RNA polymerase II (142). The disparity between these two observations may represent fundamental differences in heterochromatin assembly in budding yeast
and higher eukaryotes. In this regard, S. cerevisiae lacks methylation of histone
H3K9 and an HP1 orthologue. Alternatively, these differences may be attributed to unique characteristics of the genomic loci targeted for silencing. Thus, further
insights into the impact of heterochromatin on RNA polymerase II activity will require the identification and characterization of endogenous targets that become
58 transcriptionally silenced in association with formation of localized heterochromatin environments.
The increase in histone H3 occupancy throughout the transgene under
repressed conditions may represent an indirect measurement of increased
nucleosome ordering. We have previously shown that repression of a
chromatinized reporter transgene by a KRAB repressor protein reduces
accessibility of DNA sequences to restriction endonucleases in situ (113). In D.
melanogaster, HP1 has been shown to induce long-range ordering of
nucleosomes associated with transgenes embedded within heterochromatin
environments (143). Although our ChIP data indicate a bias in HP1 deposition at
sequences surrounding the promoter of the transgene, we did detect hormone- dependent increases in the levels of HP1 within the coding sequences of the
transgene. These data could be indicative of HP1 spreading, ultimately leading to
increased ordering of nucleosomes throughout the transcription unit.
Interestingly, recent data indicate that methylation of histone H1K26 can be recognized by the chromodomain of HP1 (144-146). Histone H1 is instrumental in
the organization of oligonucleosomes into higher-order structures and, therefore,
it would be intriguing to investigate the potential role of KAP1 in the recruitment,
methylation, and codeposition of methylated histone H1 with HP1. Thus, one possibility is that the KRAB-KAP1 repression complex directs the assembly of a highly organized chromatin environment that stearically interferes with the
59 binding of transcriptional activator proteins and the ultimate
recruitment/engagement of RNA polymerase II. One way of measuring
chromatin accessibility is to determine the ability a restriction endonuclease such as microccocal nuclease to digest the DNA wound into chromatin, the more restrictive the chromatin structure, lesser will be the DNA digestion achieved by
the endonuclease. Analysis of ERHBD-GAL4-KRAB and ERHBD-GAL4-KAP1 wild
type cells and mutants for their ability to inhibit restriction digestion of DNA
digestion in the presence or absence of OHT-mediated repression would provide a more comprehensive understanding of KRAB-KAP1 repression mechanism
While many studies have investigated the correlation between a specific
histone modification and a particular cytological domain or transcriptional state of
a gene, few studies have looked into the temporal and spatial patterns of
multiple modifications during gene silencing. In our study we looked at the
spatial distribution of general histone occupancy, histone acetylation, and site- specific histone methylation. Induction of transcriptional silencing by direct
tethering of the KAP1 corepressor to DNA is characterized by increased histone
occupancy and a concomitant decrease in histone H3 acetylation, H3K4 methylation, an increase in trimethylation of H3K9, H3K27, H3K36, and H4K20, and enrichment of the HP1 proteins at proximal regulatory elements of the
transgene. The enrichment of the H3K9 trimethyl epitope, HP1, and SETDB1 at promoter sequences is consistent with our previous data (105, 113).
60
Furthermore, hormone-dependent KRAB repression is attenuated in cells where
SETDB1 has been transiently depleted. The preference for trimethylated H3K9 is
consistent with the observation that the SETDB1/mAM enzyme complex
possesses processivity to trimethylate substrates (147). The enrichment of
trimethyl histone H4K20 is not entirely surprising, as this epitope cytologically
localizes to constitutive heterochromatin domains in a histone H3K9 methylation-
dependent manner (148). This dependency may possibly explain the absence of a
4-OHT-dependent increase at transgene sequences in cell lines that express the
mutant ERHBD-GAL-KAP1 proteins. Moreover, this observation may suggest that
the formation of a highly localized domain of heterochromatin mimics the
structure of constitutive heterochromatin domains. The enrichment of
trimethylated H3K27 is an intriguing observation; however, the patterns of
H3K27 methylation consistently mirror the H3K9 methylation patterns, suggesting
that this result may be due to cross-reactivity of this antibody with methylated
H3K9 or H1K26. Thus, the relevance of this observation relative to KAP1
repression cannot be fully defined by the current study.
We also observe an increase in histone H3K36 methylation associated with the DNA sequence in the proximal regulatory elements of the transgene.
Methylation of histone H3K4, -K36, and -K79 is commonly associated with
transcriptional competence (140). Indeed, we did observe a 4-OHT-dependent
decrease in H3K4 methylation throughout the transgene and H3K36 methylation
61 associated with DNA sequences in the downstream transcriptional unit, as would
be expected for a repressed transcript. However, the precise function of
increased H3K36 methylation at promoter sequences in transcriptional repression
is unclear at this time. Interestingly, our data suggest that H3K4 and H3K9
methylation may coexist within the same regions of a transcriptionally silenced
transgene. This result may be explained by the fact that SETDB1 can methylate
substrates that possess methylation on K4 (105). To determine whether these modifications coexist in the same nucleosome or even on the same histone,
reimmunoprecipitation experiments will need to be done in the future.
Regardless, these data are in contrast with locus-wide data from the
Schizosaccharomyces pombe mating type locus, which demonstrate an inverse
correlation between H3K4 and H3K9 mehylations (149). These data indicate that
H3K4 methylation does not need to be completely removed in order for the transcription of a gene to be repressed and, therefore, our data may represent a
fundamental difference between constitutive and localized heterochromatin
domains. Interestingly, removal of 4-OHT from the growth medium reactivates
luciferase gene expression with kinetics that are nearly identical to the time (48
to 72 h) it takes the ERHBD-GAL4-KAP1 protein to reach steady-state repression
of the transgene (Figure III-21, page 115). The presence of histone H3K4 and
H3K36 methylation may explain the rapid kinetics of the transgene's
transcriptional reactivation following withdrawal of 4-OHT. Although these data
are the first to define the repertoire of histone modification patterns associated
62 with a transcriptional unit repressed by KAP1 future studies are needed to
examine the temporal changes in the patterns of histone modifications as a gene
transitions from a transcribed state to a transcriptionally repressed state.
Another unique advantage of our experimental system is that we can evaluate the consequences of well-defined mutations in the various domains of
KAP1 on molecular changes in chromatin structure of a target gene. Cells that express the HP1 binding mutant KAP1 repressor protein fail to repress
transcription of the integrated target. Consistent with this result, the
simultaneous reduction of HP1 /ß/ reduced the efficiency of KRAB-mediated
repression. Interestingly, cells that only express a KAP1 protein possessing a mutation in the HP1 binding domain fail to undergo endodermal differentiation in
vitro (85). The combination of these data suggests that the HP1-binding-deficient
allele of KAP1 in these cells fails to repress transcription of endogenous KRAB-zfp
target genes required for the cell to differentiate.
The major question that remains is how HP1 mechanistically influences the
transcriptional state of a KRAB-zfp target gene. At a molecular level, our data are
consistent with a hypothesis that KAP1 and HP1 direct the assembly of a localized
microenvironment of heterochromatin at gene-specific loci. In our experiments,
the HP1-binding-deficient KAP1 mutant protein not only failed to recruit HP1 to
the target locus but also failed to induce methylation of H3K9. Furthermore, the
63 magnitude of other changes in histone modifications appeared to be less severe
when compared to the wild-type repressor protein. These data suggest that HP1
has additional functions in KAP1-mediated transcriptional repression beyond
simple recognition of methylated H3K9 or H1K26. The binding of HP1 to KAP1
may lead to a change in structural conformation of the corepressor required for
the functions of the PHD finger and bromodomain. Alternatively, the binding of
HP1 may trigger the translocation of target genes from eu- to heterochromatin in
order to silence gene expression, including the coordination of activities that
modulate changes in histone modifications. This latter mechanism has been
proposed for the transcription factor Ikaros, which regulates the expression of
genes involved in T-cell activation (150, 151). The potential role for HP1-directed
nuclear compartmentalization in KRAB-KAP1 regulation of gene expression is
underscored by several pieces of experimental data. First, KAP1 that is unable to
interact with HP1 fails to associate with cytologically defined heterochromatin following stimulation of cellular differentiation in vitro (84). Second,
transcriptional repression of an integrated transgene by a hormone-responsive
KRAB repressor protein correlated with an increased frequency of association
with cytologically defined heterochromatin (113). Finally, the KRAB-zfps KRAZ1
and KRAZ2 colocalize with KAP1 and HP1 proteins within 4',6'-diamidino-2-
phenylindole-stained heterochromatin in fibroblasts (86). Future experiments are
needed to identify genes that are direct targets of KRAB-KAP1 transcriptional
regulation and how the KAP1 interaction with HP1 regulates the transcription of
64 these genes during cellular differentiation, organismal development, and possibly
human disease.
MATERIALS AND METHODS
Plasmids
The p5XGAL4-TK-Luciferase and pM1-KRAB plasmids have been previously
described (36, 107).
To construct pSUPERretro-K928, nucleotides 928 to 946 (5'-
GCATGAACCCCTTGTGCTG-3') of MN_005762 were subcloned into the
BglII/HindIII sites of pSUPERretro as a short hairpin (152).
To create the FLAG-KAP1 mammalian expression vector, a 1.2-kbp
EcoRI/BamHI fragment from pFASTBAC-Flag-KAP1 (74) and a 1.4-kbp
BamHI/XbaI fragment from pM2-KAP1 (45) were subcloned into the EcoRI/XbaI
restriction sites of pcDNA3 (Invitrogen). The cDNA insert encompassed
nucleotides 346 to 2797 of MN_005762, which encodes amino acids 20 to 835 of
KAP1 fused to an NH3-terminal FLAG epitope tag. To create an allele of KAP1
refractory to the short hairpin RNA (shRNA), a double nucleotide substitution at
nucleotides 937 (C>A) and 940 (T>A) was introduced into the pC3-FLAG-KAP1
expression vector by QuikChange PCR mutagenesis. These nucleotide
substitutions are silent with regard to the coding of amino acids at codons 216
and 217. The incorporation of the corresponding nucleotide substitutions and
integrity of the surrounding KAP1 coding sequence were confirmed by DNA
65 sequence analysis. Nucleotide substitutions giving rise to the RV487, 488EE,
W664A, L720A, and F761A mutations have been previously defined (75, 80, 105,
107). DNA fragments containing these mutations were subcloned into the pC3-
FLAG-KAP1 construct, replacing the corresponding wild-type sequence.
The pC3-ERHBD-GAL4 plasmid was created by a series of sequential
subcloning steps. First, nucleotides 1023 to 1979 of NM_007956 encoding amino
acids 281 to 599 of the murine estrogen receptor hormone binding domain
containing the G525R mutation (131) were PCR amplified and subcloned into the
HindIII/BamHI restriction sites of pcDNA3 (Invitrogen). Subsequently,
nucleotides encoding the GAL4 DNA binding domain (amino acids 2 to 147) were
PCR amplified from pM1 (153) and subcloned into the BamHI/EcoRI restriction
sites of pC3-ERHBD, destroying the BamHI site as a result of a BamHI/BglII fusion. The fusion junctions and integrity of PCR-amplified DNA were confirmed
by DNA sequence analysis.
The pC3-ERHBD-GAL4-KAP1 plasmid was created by subcloning a 1.4-kb
EcoRI/XbaI fragment from pM2-KAP1(293-835) (45) into the EcoRI/XbaI sites of
pC3-ERHBD-GAL4. The pC3-ERHBD-GAL4-KAP1 (RV487, 488EE) plasmid was
created by subcloning a 1.4-kb EcoRI/XbaI fragment from pM2-KAP1
(RV487,488EE) (80) into pC3-ERHBD-GAL4. Sequence-confirmed nucleotide
changes in the coding region of KAP1 encoding the W664A, L720A, and F761A
66 mutations (75, 105, 107) were first subcloned from pM1-KAP1 (nucleotides 618-
835) into pM2-KAP1(293-835). Subsequently, each mutation was subcloned from
pM2-KAP1(293-835) into the EcoRI/XbaI sites of pC3-ERHBD-GAL4 as described
above for the wild-type coding sequence. The pC3-ERHBD-GAL4-KRAB and pC3-
ERHBD-GAL4-KRAB (DV) plasmids were created by subcloning an EcoRI/XbaI
restriction fragment from pM1-KRAB and pM1-KRAB (DV) (36), respectively, into
the EcoRI/XbaI restriction sites of pC3-ERHBD-GAL4.
pQE32-HP1 (nucleotides 70 to 642 of NM_012117, encoding amino acids
1 to 191) and pQE32-HP1ß (nucleotides 283 to 840 of NM_006807, encoding
amino acids 1 to 185) bacterial expression plasmids have been previously
described (82, 94). The HP1 bacterial expression vector (nucleotides 152 to 703
of NM_016587, encoding amino acids 21 to 173) was created by subcloning an
XmaI/XhoI fragment from pC3-FLAG-HP1 (80) into pQE32 (QIAGEN).
Nucleotides encoding the GAL4 DNA binding domain (amino acids 2 to 147) were
PCR amplified from pM1 (153) and subcloned into the BamHI/HindIII sites of pQE30 (QIAGEN). Proteins were expressed in Escherichia coli and purified as
previously described (1, 80). Purified proteins were used to generate custom polyclonal antiserum (Rockland Immunochemicals).
Transient-transfection reporter assays
67
Cells (5 x 104) were plated in 17-mm tissue culture dishes 24 h prior to
transfection. Cells were cotransfected with the indicated plasmid constructs and
500 ng of pC3-ß-gal reporter plasmid using Fugene 6 reagent (Roche) at a ratio
of 1.5 µl of Fugene per 1 µg of plasmid DNA. Forty-eight hours posttransfection,
cells were harvested in 1x reporter lysis buffer, and whole-cell lysates were used
to determine luciferase activity (Promega). Raw luciferase values were
normalized to ß-galactosidase activity. Fold repression was calculated as the ratio
of normalized luciferase activity of cells transfected in the absence of an effector plasmid to that of the cells transfected with an effector plasmid.
Generation of cell lines with a stable reduction in endogenous KAP1
HEK293 cells were transfected with pSUPERretro-K928. Twenty-four hours posttransfection cells were grown in growth medium (Dulbecco's modified Eagle's
medium plus 10% fetal bovine serum) supplemented with 10 µg/ml puromycin.
Individual antibiotic-resistant colonies of cells were expanded and maintained in
growth medium containing 10 µg/ml puromycin. The absolute level of KAP1 in
antibiotic-resistant cells was determined by Western blotting with two
independent antibodies to nonoverlapping antigens in KAP1 (107).
68
Generation of cell lines with stable integration of the 5XGAL4-TK-
luciferase transgene
HEK293 cells were cotransfected with p5XGAL4-TK-luciferase and pBabe-
Puro at a molar ratio of 10:1. Twenty-four hours posttransfection, cells were
grown in growth medium (Dulbecco's modified Eagle's medium plus 10% fetal
bovine serum) supplemented with 1 µg/ml puromycin. Individual colonies of cells
were expanded and maintained in growth medium containing 1 µg/ml puromycin.
Five micrograms of genomic DNA isolated from established clones was digested
with HindIII and subjected to Southern blot analysis to verify stable incorporation
of the luciferase plasmid (154). Basal expression of the chromatinized reporter
was determined by measurement of luciferase activity in whole-cell extracts. Raw
luciferase values were normalized to the total protein concentration. Wild-type or
mutant versions of pC3-ERHBD-GAL4-KRAB and pC3-ERHBD-GAL4-KAP1,
respectively, were transfected into 5XGAL4-TK-LUC cells to generate double
stable cell clones that expressed a hormone-responsive repressor and luciferase.
Twenty-four hours posttransfection, cells were grown in growth medium
containing 1 µg/ml puromycin and 500 µg/ml of G418. Approximately 50 well-
isolated colonies of cells for each repressor plasmid transfected were expanded
and maintained in growth medium containing 1 µg/ml puromycin and 500 µg/ml of G418. Doubly antibiotic-resistant cells were screened for 4-hydroxytamoxifen
(4-OHT; Sigma)-dependent repression of luciferase activity in whole-cell extracts.
69
Luciferase assays
Cells were plated in triplicate into 17-mm wells and grown in medium
containing either 0.1% ethanol or 500 nM 4-OHT for the indicated times. The
cells were harvested with 1x reporter lysis buffer (Promega), and lysates were
used to measure luciferase activities. Raw luciferase values were normalized to total protein concentrations. Fold repression was calculated as the ratio of normalized luciferase activity in ethanol-treated cells to normalized luciferase
activity in 4-OHT-treated samples.
Transient siRNA transfection
Cells (4 x 105) were plated into 35-mm wells and transiently transfected
with double-stranded RNA (dsRNA) oligonucleotides against KAP1 (M-005046;
K928 [5'-GCATGAACCCCTTGTGCTG-3'], K1 [5'-GACCAAACCTGTGCTTATGTT-3'],
K2 [5'-GATGATCCCTACTCAAGTGTT-3'], K3 [5'-GCGATCTGGTTATGTGCAATT-3'],
and K4 [5'-AGAATTATTTCATGCGTGATT-3']; Dharmacon SMART pool), HP1 (5'-
AAGGAGCACAATACTTGGGAA-3'), HP1ß (M-009716; Dharmacon SMART pool),
HP1 (M-010033; Dharmacon SMART pool), and SETDB1 (M-020070; Dharmacon
SMART pool). Two hundred picomoles of each oligonucleotide was diluted into
250 µl of OPTIMEM (Invitrogen). For transfections designed to simultaneously
knock down expression of HP1 , HP1ß, and HP1 , 100 picomoles of each oligonucleotide was diluted in 250 µl of OPTIMEM. One microliter of
Lipofectamine 2000 reagent (Invitrogen) per 50 picomoles of siRNA was diluted
70 in 250 µl of OPTIMEM. Diluted Lipofectamine 2000 was added to diluted siRNA
and allowed to incubate for 20 min at room temperature before being added to the cells growing in 1.5 ml of standard growth medium minus antibiotics. A
second transfection was done 48 h after the first transfection. Twenty-four hours
following the second transfection, cells were trypsinized and plated (7 x 104 per
17-mm well in triplicate) in growth medium containing either 0.1% ethanol or
500 nM 4-OHT for 48 h.
Western blot analysis
Whole-cell lysates were prepared in RIPA buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 1% Triton X-100, 0.5% deoxycholic acid, 0.1% sodium dodecyl sulfate
[SDS], 10% glycerol) supplemented with 20 mM NaF, 0.1 M phenylmethylsulfonyl
fluoride, 10 mM Na3VO4, 10 µg/ml leupeptin, 10 µg/ml aprotonin, 10 µg/ml
pepstatin, and 1 mM benzamidine. Equal amounts of protein (25 µg) were
resolved by SDS-polyacrylamide gel electrophoresis and blotted to polyvinylidene
difluoride (Millipore) (80). Antigen-antibody complexes were visualized by
enhanced chemiluminescence and exposure to X-ray film. Expression levels of
specific proteins (i.e., KAP1, HP1, SETDB1, etc.) were determined from
densitometric traces of X-ray films and normalized to the expression levels of a
loading control (i.e., ß-actin or Rbap48).
71
Chromatin Immunoprecipitation (ChIP)
Cells were plated into 100-mm dishes and grown in medium containing
either 0.1% ethanol or 500 nM 4-OHT for the indicated times. Cells were fixed
with 1% formaldehyde for 10 min at 37°C. Excess formaldehyde was quenched
by adding a 1/10 volume of 1.25 M glycine for 5 min at room temperature.
Approximately 2 x 106 cell equivalents were lysed in 100 µl of SDS-lysis buffer (50
mM Tris, pH 8.0, 10 mM EDTA, 1% SDS, 0.1 M phenylmethylsulfonyl fluoride, 10
µg/ml leupeptin, 10 µg/ml aprotonin, 10 µg/ml pepstatin, 1 mM benzamidine).
Lysed cells were sonicated using a Branson 450 sonicator with a 3-mm two-step
tapered microtip at power setting 2 and 70% duty for 12 pulses/cycle and nine
cycles ( 5-W output for 8 to 10 seconds). Clarified, sonicated chromatin was
diluted 20-fold in chromatin immunoprecipitation (ChIP) dilution buffer (16.7 mM
Tris, pH 8.0, 1.2 mM EDTA, 167 mM NaCl, 1.1% Triton X-100, 0.01% SDS),
bringing the final concentration of SDS to 0.5%. Antibodies used to
immunoprecipitate chromatin were RNA polymerase II (MMS-126R; Covance), histone H3 (ab1791; Abcam), acetyl-H3 (06-599; Upstate Biotechnology), histone
H3-AcK9 (ab4441; Abcam), histone H3-AcK14 (ab2381; Abcam), acetyl-H4 (06-
866; Upstate), histone H4 AcK16 (ab1762; Abcam), histone H3-2XmeK4 (07-030;
Upstate), histone H3-3XmeK4 (ab8580; Abcam), histone H3-2XmeK9 (07-441;
Upstate), histone H3-3XmeK9 (07-442; Upstate), histone H3-3XmeK27 (07-449;
Upstate), histone H3-3XmeK36 (ab9050; Abcam), histone H4-3X-meK20 (07-463;
Upstate), antigen-purified custom polyclonal GAL4 (DNA binding domain), HP1 ,
72
HP1ß, HP1 , and SETDB1 (105) immunoglobulin G. Antigen-DNA complexes were
eluted in 200 µl of elution buffer (50 mM NaHCO3, pH 9.0, 1% SDS), cross-links
were reversed for 5 h at 65°C, and the DNA was purified by using spin columns
(MoBio Laboratories). A 1/10 volume of purified DNA was amplified under the
following PCR conditions: 1 mM MgCl2, 1 µM primer, 200 µM deoxynucleoside
triphosphate, and 0.25 U Taq DNA polymerase. DNA was denatured for 4 min at
94°C, followed by 28 cycles of 15 seconds at 94°C, 15 seconds at 55°C, and 30 seconds at 72°C. Primer sequences used to amplify immunoprecipitated DNA
were as follows: (i) GAL4(DBS), 5'-CACACAGGAAACAGCTATGAC-3'(sense) and 5'-
GAATTCGCCAATGACAAGAC-3'(antisense); (ii) HSVTK promoter, 5'-
GGATCCGACTAGATCTGACTTC-3'(sense) and 5'-CCAGGAACCAGGGCGTATCTC-
3'(antisense); (iii) LUC3', 5'-TACTGGGACGAAGACGAACAC-3'(sense) and 5'-
TCGTCCACAAACACAACTCC-3'(antisense); (iv) poly(A), 5'-
CACACAGGCATAGAGTGTCTG-3'(sense) and 5'-GATACATTGATGAGTTTGGAC-
3'(antisense). PCR-amplified products were run on a 2% agarose gel and
visualized by ethidium bromide staining. The fluorescence was captured by an
eight-bit digital camera, and signal intensities were quantitated using GeneTools
software from Syngene (Frederick, MD). Signals from specific
immunoprecipitations were normalized to signals from input DNA (0.0625%).
Enrichment was calculated as the ratio of normalized signal of amplified DNA
from chromatin immunoprecipitated from 4-OHT-treated cells to normalized
73 signal of amplified DNA from chromatin immunoprecipitated from ethanol-treated
cells.
ACKNOWLEDGEMENT
We thank Ruth Keri, Peter Harte, and John Mieyal for helpful comments
during the course of this work. We acknowledge technical assistance provided by
Bonnie Gorzelle to various parts of this work. We thank Yael Ziv and Yosef Shiloh for pSUPER-KAP1(928) vectors.
This work was supported by Public Health Service grant CA-99093 (D.C.S.) from the National Cancer Institute and funds from the Mount Sinai Healthcare
Foundation and Case Comprehensive Cancer Center (D.C.S.).
74
Figure III-1: KAP1 is required for KRAB-mediated repression. (A)
Western blot analysis of endogenous KAP1 expression in two independent stable knockdown cell lines, using antibodies that recognize either the N terminus (anti-
RBCC) or C terminus (anti-PHD/bromo) of KAP1. Detection of Rbap48 (p48) was used as a loading control. (B) Schematic illustration of the 5XGAL4-TK-luciferase reporter and GAL-KRAB repressor protein. (C) Stable KAP1 knockdown cells (cl4 and cl10) were transiently transfected with the p5XGAL4-TK-luciferase reporter and the indicated amounts of plasmid that expresses the GAL4-KRAB repressor protein. Luciferase activity was measured 48 h posttransfection and normalized for transfection efficiency. Repression (n-fold) represents the ratio of normalized luciferase activity in the absence of any effector plasmid to the activity measured in the presence of the indicated amount of GAL4-KRAB plasmid transfected. The data represent the averages of two independent experiments done in triplicate.
Error bars represent the standard deviations of the means (Published in MCB,
Sripathy SP et al, 2006. 26:8623-38).
75
76
Figure III-2: Stable depletion of KAP1 in HEK293 cells. (A) Quantitation of KAP1 depletion. KAP1 and Rbap48 (p48) levels were obtained from densitometric traces of autoradiographs. KAP1 levels were normalized to Rbap48
(p48) expression and plotted as the % reduction of KAP1 expression in cells stably transfected with either the pSUPERretro or pSUPERretro-K928, as compared to parental HEK293 cells. (B) Western blot analysis of KAP1 and its interacting proteins in two independent KAP1 stable knockdown cell lines. β-Actin expression was used as a loading control (Published in MCB, Sripathy SP et al,
2006. 26:8623-38).
77
78
Figure III-3: Direct tethering of KAP1 to DNA is sufficient to represses transcription. (A) Schematic illustration of the 5XGAL4-TK-luciferase reporter and GAL4-KAP1 repressor proteins. (B) Transient transfection reporter assay with the indicated GAL4-KAP1 constructs. Fold repression was calculated as described in figure III-1(Published in MCB, Sripathy SP et al, 2006. 26:8623-38).
79
80
Figure III-4: The corepressor activity of KAP1 depends on its interaction with HP1 and a functional PHD finger and bromodomain. (A)
Schematic illustration of KAP1's domain structure and location of synthetically introduced mutations. (B) Stable KAP1 knockdown cells were transiently transfected with the p5XGAL4-TK-luciferase reporter, 100 ng of pM1-KRAB, and increasing amounts of a plasmid that expresses FLAG-tagged KAP1 (wild type
[WT], RV487, 488EE, W664A, and F761A). Luciferase activity was measured 48 h posttransfection and normalized for transfection efficiency. Repression was calculated as described for Figure III-1. Data are representative of two independent experiments done in triplicate. Error bars represent the standard deviations of the means. The apparent absence of error bars indicates a standard deviation too small to be physically illustrated. (C) Western blot analysis of transfected HEK293 cells, confirming stable exogenous expression of the
FLAG-KAP1 proteins (using anti-FLAG and anti-RBCC antibodies). ß-Actin represents a loading control (Published in MCB, Sripathy SP et al, 2006.
26:8623-38).
81
82
Figure III-5: Establishment of cell lines with a hormone-regulatable reporter transgene. (A) Strategy to generate cell lines that possess a stably integrated 5XGAL4-TK-luciferase reporter and constitutive expression of a hormone-responsive repressor protein. (B) Schematic illustration of heterologous hormone-responsive repressor proteins. The KRAB domain (Kox1 amino acids 1 to 90) or amino acids 293 to 835 of KAP1 (wild type, RV487, 488EE, W664A,
L720A, and F761A) were fused in frame to the C terminus of the ERHBD-GAL4
DNA binding domain fusion (Published in MCB, Sripathy SP et al, 2006. 26:8623-
38).
83
84
Figure III-6: Characterization of cell lines with a chromatinized
5XGAL4-TK-luciferase transgene. (A) Schematic illustration of key regulatory elements in the luciferase reporter. The bold line represents the position of the DNA probe used in Southern blot analysis to detect transgene copy number and integration. (B) Southern blot analysis of six clonally expanded puromycin resistant cells following stable transfection with p5XGAL4-TK- luciferase. (C) Analysis of basal luciferase activities in cell lines possessing stable integration of the luciferase reporter following growth in medium containing either 0.1% ethanol or 500 nM 4-OHT for 48 hours (Published in MCB, Sripathy
SP et al, 2006. 26:8623-38).
.
85
86
Figure III-7: Hormone dependent repression of chromatinized luciferase transgenes by ERHBDTM-GAL4 fused repressor proteins. (A)
Repression of luciferase activity is dependent on the concentration of 4-OHT.
The indicated cell lines were treated with increasing concentrations of 4-OHT for
48 hours and harvested to assay for luciferase activity. The data represents the average of two independent experiments done in triplicate. The error bars represent standard deviation of the mean. (B) Hormone dependent repression of luciferase activity correlates with reduced steady state levels of luciferase mRNA. RTPCR amplification of luciferase mRNA from cells treated with either
0.1% ethanol or 500 nM OHT for 48 hours. The cDNA template was amplified for the indicated number of PCR cycles. Amplified products were run on a 2% agarose gel, visualized by ethidium bromide staining, and quantified. The graphs represent the level of luciferase mRNA in ethanol (-OHT) and 4-OHT (+OHT) treated cells, respectively, normalized to GAPDH levels at the indicated cycle numbers. The data is an average of two independent experiments with the error bars representing the standard deviation of the mean (Published in MCB,
Sripathy SP et al, 2006. 26:8623-38).
.
87
88
Figure III-8: Hormone-dependent repression of the chromatin template by ERHBD-GAL4-KRAB requires KAP1. (A, upper panel) Kinetics of hormone-dependent repression by ERHBD-GAL4-KRAB. Cells were grown in medium containing either 0.1% ethanol or 500 nM 4-OHT for the indicated times. Repression (n-fold) was calculated as the ratio of normalized luciferase
activity in the absence of the hormone to normalized luciferase activity in the
presence of the hormone. The data represent the averages of three independent
experiments done in triplicate, and the error bars represent the standard
deviations of the means. (Lower panel) Expression of ERHBD-GAL4-KRAB (arrow)
in the indicated cell clones was detected using an antibody against the GAL4
DNA binding domain. Detection of Rbap48 (p48) was used as a loading control.
(B) Overview of the experimental scheme. 12.10Kr cells were subjected to two
rounds of transfection with 100 nM of independent dsRNA oligonucleotides
designed to reduce expression of KAP1 prior to treatment with either 0.1%
ethanol or 500 nM 4-OHT for 48 h. (C) Western blot analysis of whole-cell
extracts prepared on day 6 from transfected cells with the indicated siRNA. (D)
Transient depletion of KAP1 by independent siRNA molecules targeted to
different regions of the KAP1 mRNA results in attenuation of hormone-dependent
KRAB-mediated repression. Repression was calculated as described for panel A.
Data are representative of two independent experiments done in triplicate. The
error bars represent the standard deviations of the means. UT, untransfected
(Published in MCB, Sripathy SP et al, 2006. 26:8623-38).
89
.
90
Figure III-9: Hormone-dependent repression Kinetics of ERHBD-GLA4-
KRAB wildtype and mutant cells. Cells were grown in medium containing either 0.1% ethanol or 500 nM 4-OHT for the indicated times. Repression (n- fold) was calculated as the ratio of normalized luciferase activity in the absence of the hormone to normalized luciferase activity in the presence of the hormone.
The data represent the averages of three independent experiments done in triplicate, and the error bars represent the standard deviations of the means
(unpublished data).
91
92
Figure III-10: Hormone-dependent KRAB repression is dependent on
KAP1. 7.18Kr cells (Figure III-8A) were subjected to two rounds of transfection with 100 nM of the indicated siRNAs prior to growth in medium containing either
0.1% ethanol or 500 nM 4-OHT for 48 hours. (A) Western blot analysis of whole cell extracts prepared on day 6 from transfected cells. (B) Quantitation of KAP1 expression. KAP1 expression levels in extracts prepared on day 6 from cells transfected with the indicated siRNAs was determined by densitometry and normalized to the expression of β-Actin. The graph represents the % of KAP1 expression as compared to mock transfected cells. (C) Transient depletion of
KAP1 attenuates hormone dependent repression of a chromatin template. The data is representative of two independent experiments done in triplicate. Error bars represent the standard deviation of the mean (Published in MCB, Sripathy
SP et al, 2006. 26:8623-38).
.
93
94
Figure III-11: KRAB mediated repression of a chromatin template is independent of TIF1α and TIF1γ. (A) 12.10Kr cells were subjected to two rounds of transfection with 100 nM of dsRNA oligonucleotides designed to knockdown expression of TIF1α, KAP1(TIF1β), and TIF1γ, prior to treatment with either 0.1% ethanol or 500 nM 4-OHT for 48 hours. The mRNA isolated from cells transfected with the indicated siRNAs was used as a template for a
RT-PCR reaction in order to detect the expression of the indicated mRNAs. β-
Actin expression was used as an indicator of RT efficiency. (B) Western blot analysis of KAP1 (TIF1β) expression in whole cell lysates prepared from cells transfected with the indicated siRNAs. β-Actin was used as a loading control. (C)
Transient depletion of TIF1β (KAP1) attenuates hormone dependent KRAB repression. Fold repression was calculated as described in figure 4. The data is average of two independent experiments done in triplicate. Error bars represent the standard deviation of the mean (Published in MCB, Sripathy SP et al, 2006.
26:8623-38).
.
95
96
Figure III-12: Hormone-dependent repression of chromatin templates by ERHBD-GAL4-KAP1. (A) Kinetics of hormone-dependent transcriptional repression by ERHBD-GAL4-KAP1 in the indicated cell clones. Cells were grown in medium containing either 0.1% ethanol or 500 nM 4-OHT for the indicated times. Repression was calculated as described for Figure III-8A. The data represent the means of two independent experiments done in triplicate, and the error bars represent the standard deviations of the means. (B) Expression of
ERHBD-GAL4-KAP1 (arrow) in the indicated cell clones was detected using an antibody against the GAL4 DNA binding domain. ß-Actin expression was used as a loading control. (C) Western blot analysis of stable cell lines expressing the indicated ERHBD-GAL4-KAP1 proteins (wild type [WT], RV487, 488EE, W664A,
L720A, and F761A). Numbers at the bottom represent the expression level
(arbitrary units) of each ERHBD-GAL4-KAP1 protein, normalized to ß-actin expression. (D) The indicated cell lines were grown in medium containing either
0.1% ethanol or 500 nM 4-OHT for 96 h. Repression was calculated as described in the legend for Figure III-8. The data represent the averages of three independent experiments done in triplicate, and the error bars represent the standard deviations of the means (Published in MCB, Sripathy SP et al, 2006.
26:8623-38).
.
97
98
Figure III-13: Kinetics of hormone-dependent repression of chromatin templates by ERHBD-GAL4-KAP1 wildtype and mutants. (A) Kinetics of hormone-dependent transcriptional repression by ERHBD-GAL4-KAP1 in the indicated cell clones. Cells were grown in medium containing either 0.1% ethanol or 500 nM 4-OHT for the indicated times. Repression was calculated as described for Figure III-8A. The data represent the means of two independent experiments done in triplicate, and the error bars represent the standard deviations of the means. (B) Expression of ERHBD-GAL4-KAP1 (unpublished data).
99
100
Figure III-14: Expression of the KAP1 repression machinery. Whole cell extracts prepared from stable cell lines expressing the indicated ERHBDTM-GAL4-
KAP1 proteins (wild-type, RV487, 488EE, W664A, L720A and F761A), respectively, were analyzed by Western blot analysis for expression of Mi2α,
SETDB1, HP1α, HP1β and HP1γ. β-Actin expression was used as a loading control (Published in MCB, Sripathy SP et al, 2006. 26:8623-38).
101
102
Figure III-15: Chromatin immunoprecipitation analysis of a luciferase transgene repressed by ERHBD-GAL4-KAP1. (A) Schematic illustration of the 5XGAL4-TK-luciferase transgene. Bold lines represent four regions of the transgene amplified by PCR in DNA recovered from immunoprecipitations. (B)
Formaldehyde-cross-linked chromatin from 12.32KA cells grown in medium containing either 0.1% ethanol (Etoh) or 500 nM 4-OHT for 96 h was immunoprecipitated with antibodies against the indicated antigens. The immunoprecipitated DNA was PCR amplified at the indicated loci to detect hormone-dependent changes in hypophosphorylated RNA polymerase II recruitment, histone H3 occupancy, and histone modifications (left panel) and changes in HP1 and SETDB1 occupancy (right panel). Input DNA represents
0.25, 0.125, and 0.0625% of the total amount of DNA immunoprecipitated, respectively (Published in MCB, Sripathy SP et al, 2006. 26:8623-38).
.
103
104
Figure III-16: (A) Hormone dependent repression mediated by KAP1 involves hypoacetylation of histone H3 K9 of the transgenic reporter.
Formaldehyde cross-linked chromatin from 12.32KA cells, treated with either
0.1% ethanol or 500 nM 4-OHT for 96 hours was immunoprecipitated with antibodies against the indicated antigens. DNA recovered from the immunoprecipitations was PCR amplified at the indicated loci along the transgene to detect hormone dependent changes in site specific acetylation of histones H3 and H4. (B) Quantitation of PCR amplified products from chromatin immnuprecipitation analysis of 12.32KA cells (Figure III-15 and Figure
III-16A). The PCR amplified products were run on a 2% agarose gel and visualized by ethidium bromide staining. The fluorescence intensities were quantified and the signals from specific immunoprecipitations were normalized to signals from the input DNA (0.0625%). Fold enrichment (>1) was calculated as the ratio of normalized signal of the immunoprecipitated DNA from 4-OHT treated cells to normalized signal of the DNA immunoprecipitaed from ethanol treated cells. Fold decrease (<-1) was calculated as the ratio of normalized signal of the immunoprecipitated DNA from ethanol treated cells to normalized signal of the DNA immunoprecipitaed from 4-OHT treated cells. The graph represents the average of three independent experiments. The error bars represent the standard deviation of the mean (Published in MCB, Sripathy SP et
al, 2006. 26:8623-38).
.
105
106
Figure III-17: Disruption of the KAP1-HP1 interaction fails to induce hormone-dependent changes in RNA polymerase II recruitment, histone occupancy, and histone modifications (A) or HP1/SETDB1 recruitment to promoter sequences of the luciferase reporter (B) following treatment with 4-OHT. Formaldehyde-cross-linked chromatin from 12.11M2 cells grown in medium containing either 0.1% ethanol or 500 nM 4-OHT for 96 h was immunoprecipitated with antibodies against the indicated antigens. Promoter and
3' luciferase coding sequences recovered from the immunoprecipitations were
PCR amplified. Input DNA represents 0.25, 0.125, and 0.0625% of the total amount of DNA immunoprecipitated (Published in MCB, Sripathy SP et al, 2006.
26:8623-38).
.
107
108
Figure III-18. Chromatin immnoprecipitation analysis of hormone dependent changes in RNA polymerase II, histone occupancy, and histone modifications at promoter sequences of an integrated transgene bound by ERHBDTM-GAL4-KAP1 mutated in either the PHD
finger or bromodomain, respectively. Formaldehyde cross-linked chromatin
from either 12.29WA (PHD mutant) or 12.18FA (Bromodomain mutant) cells
grown in medium containing either 0.1% ethanol or 500 nM 4-OHT for 96 hours
was immunoprecipitated with antibodies against the indicated antigens. The
immunoprecipitated DNA was PCR amplified at the indicated loci. Input DNA
represents 0.25, 0.125, and 0.0625% of the total amount of DNA
immunoprecipitated, respectively (Published in MCB, Sripathy SP et al, 2006.
26:8623-38).
.
109
110
Figure III-19: Hormone-dependent KRAB repression of a chromatin template requires KAP1, HP1, and SETDB1. 12.10Kr cells were subjected to two rounds of transfection with dsRNA oligonucleotides to transiently reduce levels of the indicated proteins, as described in the legend for Figure III-8B. For the triple knockdown of HP1 , -ß, and - , cells were transfected with 50 nM siRNA to each target. (A) Western blot analysis of whole-cell extracts prepared on day 6 of post-transfection with the indicated siRNA. UT: untransfected. (B)
Transient depletion of KAP1, HP1, and SETDB1 attenuates hormone-dependent
KRAB repression of a chromatin template. The data represent the averages of two independent experiments done in triplicate. Error bars represent the standard deviations of the means (Published in MCB, Sripathy SP et al, 2006.
26:8623-38).
.
111
112
Figure III-20: Transient knockdown of HP1β expression and its effect on KRAB mediated repression. (A) Quantitation of protein expression following transfection of 12.10Kr cells with the indicated siRNAs to knockdown expression of KAP1, HP1α, HP1β, HP1γ, and SETDB1, as indicated in the Figure
III-17. Whole cell extracts from the transfected cells were subjected to Western blot analysis. Expression levels of the each protein was determined by densitometry and normalized to β-Actin expression. The graph represents the expression of the indicated proteins as a % of the untransfected cells. Data represents the average of two independent experiments and the error bars represent the standard deviation of the mean. (B) 12.10Kr cells were subjected to two rounds of transfection with 100 nM of siRNA designed to knockdown expression of KAP1 and HP1β, prior to treatment with either 0.1% ethanol or 500 nM 4-OHT for 48 hours, as described in Figure III-8B. The mRNA isolated from cells transfected with the indicated siRNA was used as a template for a RTPCR reaction in order to detect the expression of the indicated mRNAs. β-Actin expression was used as an indicator of RT efficiency. (C) Transient depletion of
KAP1 attenuates hormone dependent repression of a chromatin template. Fold repression was calculated as described in Figure III-8. Data represents the average of two independent experiments done in triplicate and the error bars represent the standard deviation of the mean (Published in MCB, Sripathy SP et al, 2006. 26:8623-38).
.
113
114
Figure III-21: Kinetics of onset and offset of hormone-dependent repression a chromatin template by ERHBD-GAL4-KRAB. 12.10Kr Cells were grown in medium containing either 0.1% ethanol or 500 nM 4-OHT for the indicated times. Cells treated with OHT for 120 hours were grown in fresh media without OHT (washout) and assayed for luciferase activity at the time points indicated. Repression was calculated as described for Figure III-8A. The data represent the means of two independent experiments done in triplicate, and the error bars represent the standard deviations of the means.
115
116
CHAPTER IV
SUMMARY AND FUTURE DIRECTIONS
SUMMARY
Various studies strongly suggest a key role for the KRAB zfp subfamily of
proteins in mediating differentiation programs and contributing to species specific
differences. KRAB zfps function via an obligate corepressor protein KAP1. To
characterize the role of KAP1 and KAP1-interacting proteins in transcriptional
repression, we investigated the regulation of stably integrated reporter
transgenes by hormone responsive KRAB and KAP1 repressor proteins
(Chapter III). We demonstrate that depletion of endogenous KAP1 levels by
small interfering RNA (siRNA) significantly inhibited KRAB-mediated
transcriptional repression of a chromatin template. Similarly, reduction in cellular
levels of the heterochromatin proteins HP1α/β/γ and the histone
methyltransferase SETDB1 by siRNA attenuated KRAB-KAP1 repression. We
also found that direct tethering of KAP1 to DNA was sufficient to repress
transcription of an integrated transgene. This activity was absolutely dependent
upon the interaction of KAP1 with HP1 and on the presence of an intact PHD finger and bromodomain of KAP1, suggesting that these domains function
cooperatively in transcriptional corepression. The achievement of the repressed
state by wild-type KAP1 involves decreased recruitment of RNA polymerase II,
reduced levels of histone H3K9 acetylation and H3K4 methylation, an increase in
histone occupancy, enrichment of trimethyl histone H3K9, H3K36, and histone
117
H4K20, and HP1 deposition at proximal regulatory sequences of the transgene.
A KAP1 protein containing a mutation of the HP1 binding domain failed to induce any change in the histone modifications associated with DNA sequences of the transgene, implying that HP1-directed nuclear compartmentalization is required for transcriptional repression by the KRAB/KAP1 repression complex. The combination of these data suggests that KAP1 functions to coordinate activities that dynamically regulate changes in histone modifications and deposition of HP1 to establish a de novo microenvironment of heterochromatin, which is required for repression of gene transcription by KRAB-zfps (Figure IV-1, page 119).
118
Figure VI-1: KAP1 acts as a scaffold to coordinate histone modifying activities resulting in the formation of localized heterochromatin-like domains. KRAB zfps bind to the DNA via tandem arrays of zinc fingers, the
KRAB domain at the C-terminus is attached to the zinc finger domain via a flexible hinge region. DBS: DNA binding site
119
FUTURE DIRECTIONS
Based on our studies we concluded that KAP1 is an obligate corepressor of KRAB zfps and that effective repression by KAP1 requires a functional HP1 binding domain, PHD and bromo domains. The logical next step would be to study the transcriptional programs targeted by the KRAB-KAP1 system in a biologically relevant context. To understand the biological processes regulated by a particular KRAB zfp, it is essential to find out the target genes regulated by that protein. An unbiased approach is to conduct a “ChIP-on-chip” experiment where we can immunoprecipitate the chromatin bound by a particular KRAB zfp and after purification of DNA, use it as a template for microarray hybridization using a microarray chip made up of promoter sequences. Such arrays are commercially available and standardized for ChIP-on-chip experiments. The
Human Promoter 1.0R Array from Affymetrix consists of probes tiled to interrogate more than 25,000 promoter regions and the Mouse Tiling 2.0R Array
Set interrogates more than 28,000 mouse promoter regions.
We can potentially identify genes regulated by a given KRAB zfp based on the sequence of the immunoprecipitated DNA nucleotides that successfully hybridized to the microarray probes. Validation of the putative genes regulated by the KRAB zfp would involve RT-PCR analysis of the genes in the presence and absence of the KRAB zfp. Even though in principle it is possible to study KRAB zfp function this way, technically it is highly challenging. As mentioned before,
120
KRAB zfps occur as clusters of tandem and inverted repeats, and genes within a single cluster vary by a few nucleotides. These genes code for KRAB zfps that have highly similar structures and can regulate overlapping or distinct processes.
As such, it is highly probable that the hits generated by the ChIP-on-chip microarray experiment would include a large number of false positive results.
This is primarily because of the lack of specificity of anti-KRAB zfp antibodies in binding to a given KRAB protein.
Identification of KAP1/KRAB zfp binding sites- An unbiased approach
KAP1 is an obligate corepressor of KRAB zfps and mediates KRAB repression by directly binding to and interacting with the KRAB domain. KAP1 knockout mice exhibit embryonic lethality (83), highlighting the necessity of
KAP1 for early embryo development. Also, KAP1 is ubiquitously expressed and its expression levels do not change significantly with cell differentiation. KRAB zfps on the other hand, are differentially expressed in cell types and differentiation processes. One interpretation of these observations is that KAP1 mediates distinct transcriptional programs via interaction with various KRAB zfps. On the other hand, KAP1 might mediate processes that do not involve participation of a
KRAB zfp. However, KAP1 by itself cannot bind directly to DNA and to date, no other KAP1 interacting, DNA binding factor apart from KRAB zfps have been identified. These observations indicate that identification of KRAB zfp target sites can be achieved by performing a ChIP-on-chip experiment using DNA
121 immnunoprecipitated by anti-KAP1 antibodies. In fact, such an experiment has already been described where a ChIP-on-chip assay using antibodies against
KAP1 identified ~7000 KAP1 binding sites and a significant percentage of these binding sites were located within 5 Kb upstream or downstream of the transcription start sites of known genes (118). However, these studies were conducted using HEK293 cells which do not provide a good differentiation model.
The cell system used for KAP1 based ChIP-on-chip experiment is critical to interpret the output of this assay. As mentioned earlier, KRAB zfps are differentially expressed across cell-differentiation programs. Assuming that KAP1 is targeted to DNA predominantly via interaction with KRAB zfps, we can expect that a major fraction of the changes in global KAP1 binding is KRAB zfp- dependent. This assumption is supported by the observation that, overexpression of the VP16 transactivation domain fused with a KAP1 mutant that binds to the
KRAB domain but not to HP1 led to dramatic redistribution of the KRAB zfp,
KRAZ1 from repressive centromeric foci. Further, KRAZ1-mediated silencing was converted into strong transcriptional activation (86).
Hence, assaying for changes in KAP1 binding between undifferentiated and differentiated phenotypes of a given cell line would have increased accuracy for identifying KAP1/KRAB zfp targeted genes than a similar assay conducted on either phenotype alone. Mouse embryonic carcinoma cells (F9) is a good
122 example for a cell differentiation model that can be utilized for the unbiased microarray approach described above, especially so in the case of KAP1. A series of studies indicate that KAP1 is essential for mediating terminal differentiation of
F9 cells (84, 85, 155). F9 cells differentiate into primitive endoderm-like (PrE) cells upon trans-retinoic acid (tRA) treatment. PrE cells further differentiate into parietal endoderm-like (PE) cells upon addition of cAMP, and finally into visceral endoderm-like (VE) cells after treatment of vesicles with RA. Disruption of KAP1-
HP1 interaction does not prevent differentiation of F9 cells into PrEs, but terminal differentiation into PE and VE cell types is inhibited.
However, further experiments indicate that KAP1–HP1 interaction is essential only during a short window of time within early differentiating PrE cells to establish a selective transmittable competence to terminally differentiate after a cAMP inducing signal (85). The differentiation of F9 cells into PrE cells after treatment with tRA coincides with a change in the KAP1 and HP1 staining patterns in the nucleus. This indicates a correlation between the differentiation process and HP1-dependent changes in chromatin structure. Mouse endodermal differentiation requires silencing of oct3/4 expression and F9 cells treated with tRA exhibit down regulation of oct3/4 after 24 hours of tRA treatment. However, cells expressing KAP1 mutated for interaction with HP1 are unable to silence oct3/4 expression. Together, these observations highlight the role of KAP1 during
123 the early differentiation phase of PrE cells. They also suggest that KAP1-HP1 interaction primarily contributes to the ability of KAP1 to repress transcription.
Since KRAB zfps are the only group of transcription factors known to bind to KAP1, it is likely that endodermal differentiation is also regulated by KRAB-
KAP1 systems. As such, F9 cells constitute an ideal system to identify KRAB-KAP1 target genes involved in endodermal differentiation. This cellular differentiation model offers other distinct advantages: 1) It is a simple model and hence can be manipulated with greater precision, 2) The role of KAP1 in inducing the terminal differentiation of cells in this model system has been relatively well characterized. 3) It is very amenable to the generation of genetic KAP1 knockouts.
However, it must be acknowledged that we cannot rule out the possibility that KAP1 is targeted to DNA via interactions with as yet unidentified factors.
One way of identifying KRAB-independent DNA binding of KAP1 would be to introduce synthetic peptides that inhibit KAP1-KRAB interaction. The RBCC domain of KAP1 is responsible for interaction with KRAB domain and overexpression of RBCC domain alone is known to inhibit KRAB domain-mediated transcription. As such, a peptide mimicking the RBCC domain would act as a dominant negative inhibitor of KRAB zfp interaction with full length KAP1. The
RING and B box motifs in particular contain conserved residues that are critical
124 for binding to the KRAB domain. Hence, designing short peptides that span the
RING and B box motifs of KAP1 would be a good starting point for identifying the optimal peptide inhibitor. The pool of peptides can be tested for their efficiency by introducing them into F9 cells that: 1) have a stably integrated luciferase reporter regulated by a 5X-GAL4 DNA binding site, 2) constitutively express a
KRAB domain fused to GAL4 DNA binding domain. The peptides can be evaluated based on their ability to inhibit KRAB-mediated repression of the reporter.
Commercially available delivery agents (ChariotTM by Active Motif) are available
that can be used to introduce peptides into cells.
Incorporation of such an inhibitor as a control will help us filter out KRAB
zfp-independent DNA binding by KAP1. These sites can be easily identified
following ChIP-on-chip assay in the presence and absence of the inhibitor and
subsequent comparison of KAP1 binding sites revealed by these two treatments.
Putative KRAB-KAP1 targets are those genes that are not bound by KAP1 in the
presence of the peptide inhibitor but display KAP1 binding in the absence of the inhibitor. This approach would require four sets of samples to be analyzed 1) undifferentiated (F9) cells without the inhibitor 2) undifferentiated (F9) cells treated with the inhibitor 3) differentiated (PrE) cells without the inhibitor and 4) differentiated (PrE) cells with the inhibitor. Once we have selected putative
KAP1/KRAB zfp genes within the undifferentiated and differentiated cell phenotypes, we can then compare the KAP1-binding profiles between these two
125 phenotypes. Likely targets of KRAB regulation would be those genes that are located close (~5 Kb upstream or downstream of the transcription start site) to
KAP1 binding sites only in the PrE cells.
Identification of KRAB-KAP1 target genes in endodermal differentiation
In order to identify genes that are potentially regulated by the KRAB-KAP1 system, we can adopt the same strategy outlined in the previous section with one exception. Cells need to be processed for RNA isolation instead of chromatin immunoprecipitation, the isolated mRNA can then be hybridized to probes on an expression microarray chip after fluorescent tagging. A variety of mouse gene chip arrays are available such as the Mouse Expression Set 430 from Affymetrix that represents >45,000 transcribed genes from the mouse genome.
Comparative analysis of genes that are differentially expressed between undifferentiated F9 cells and PrE cells would provide putative targets of the
KRAB/KAP1 repression system. Hits from the microarray experiment must be validated by RT-PCR analysis of individual genes in the presence and absence of tRA (trans retinoic acid) treatment.
Comparison of putative KRAB/KAP1 binding targets from the ChIP-on-chip
assay and putative KRAB/KAP1 regulated genes from expression microarray
assay would yield a subset of genes that are differentially downregulated after
126 tRA treatment and also exhibit KAP1 binding closer to their transcription start sites. Initial validation of the putative targets would involve ChIP analysis of the
5’ regulatory regions of these genes for binding of at least KAP1 and HP1. Such an experiment would provide us with a good starting point to analyze KRAB-
KAP1 biology.
Biological significance of the PHD and bromodomains
A majority of the existing studies on the role of KAP1 in mediating differentiation programs are focused on the ability of KAP1 to interact with HP1, supporting the necessity for KAP1-HP1 interaction. However, we observed attenuated repression by KAP1 when either the PHD or the bromodomain were mutated (Chapter III). Also down regulation of SETDB1 attenuated KAP1 mediated repression indicating that the PHD and bromodomains play important roles in KAP1 biology via recruitment of histone modifying activities to repress
KRAB targeted genes (Chapter III). In fact, recent studies on the KAP1 PHD and bromodomains report that the PHD domain acts as an E3 ligase to sumoylate 6 different lysine residues that occur in the PHD and bromodoamins (156, 157).
Sumoylation of the bromo domain in particular is a prerequisite for recruitment of both SETDB1 and Mi2α. Further, interaction of KAP1 with the KRAB domain positively regulates the auto-sumoylation (156). Finally, it has been found that decreased sumoylation of KAP1 attenuates the ability of KAP1 to repress p21
127
(158, 159). These observations indicate a key role for the PHD and
Bromodomains in mediating the biological processes of KAP1.
The availability of F9 KAP1 null cells provides an excellent system to analyze the contribution of the PHD and bromodomains towards mediating biological processes by KAP1. Complementation and rescue experiments with
KAP1 PHD and bromo muatants would reveal key residues involved in mediating the functions of KAP1. Such an analysis is important because, even though sumoylation of the PHD and bromo domains was shown to be necessary for
KAP1 mediated repression, mutation of residues K779 and K804 in the bromo domain coincided with both a lack of sumoylation and a lack of repression, while other residues exhibited modest or no effect on the repressive ability of KAP1 even though their sumoylation was abrogated (156). This observation suggests that sumoylation may contribute to KAP1 functions other than just repression. It is possible that a change in conformation of KAP1 upon sumoylation of certain residues allows for interaction with chromatin modifying enzymes that may not contribute to repression per se, but might be involved in relocation of KAP1
bound DNA to repressive environments.
It would be informative to analyze the potential for other KAP1 mutants to
rescue differentiation. For example the residues W664 and F761 are important
for KAP1-mediated repression but are not targets for sumoylation (chapter III).
128
It would be interesting to transfect F9 cells with various PHD and bromodomain mutants of KAP1 and to examine their effect on tRA-induced PrE differentiation.
The PrE cells can be further differentiated to PE cells by treating them first with tRA followed by cAMP treatment. Finally, PE cells can be terminally differentiated into VE cells by tRA treatment. It is possible that the mutants differentially rescue either both or one of the differentiation programs, thus providing us with more insight into KAP1 biology.
KAP1 and DNA damage response
Recently, it has been shown that KAP1 can mediate response to DNA damage in addition to its role as a transcriptional corepressor. KAP1 is phosphorylated at ser824 upon induction of double stranded DNA breaks and phosphorylation can be mediated by members of the PIKK family of kinase (160).
Phosphorylation was observed even in the absence of ATM and ATR suggesting a global role for KAP1 in mediating DNA damage response. KAP1 present at the sites of DNA damage is phosphorylated by ATM within 5 minutes of the initiating event. The phosphorylated KAP1 spreads rapidly throughout the nucleus within
15 mins of inducing DNA damage. Kinetics of spreading of phosphorylated KAP1 is closely followed by a wave of chromatin relaxation and this relaxation is dependent on phosphorylation of KAP1 (161). Chromatin relaxation in turn facilitates greater accessibility of the DNA damaged sites to protein complexes that mediate DNA repair. These observations indicate an involvement of KAP1
129 phosphorylation in the relaxation of global chromatin condensation. This is further supported by the observation that cells expressing constitutively phosphorylated KAP1 also exhibit increased chromatin relaxation.
Curiously, no significant changes were observed with regards to KAP1 and
HP1 co-localization upon induction of DNA damage. This is interesting given that
HP1 is a constituent of heterochromatin and is a well characterized mediator of chromatin condensation. Since studies of transcriptional regulation by KAP1 indicate that KAP1 is capable of inducing heterochromatin-like environments in an HP1-dependent manner, it is conceivable that phosphorylation inhibits the ability of KAP1 to mediate chromatin condensation without affecting its ability to interact with HP1. This observation also indicates that KAP1 potentially interacts with factors other than HP1 to mediate heterochromatin condensation. Together these studies suggest a role for KAP1 in DNA damage response but it is not clear whether there is cross-talk between the proteins involved in the transcriptional and DNA response activities of KAP1.
However, it has also been reported that ser824 phosphorylation of KAP1 results in decreased sumoylation of KAP1 leading to attenuation of KAP1 repressor activity. Phosphorylation of KAP1 also correlates with increased expression of p21, GADD45α and Bax in MCF7 breast cancer cells indicating that phosphorylation also affects the transcriptional ability of KAP1 (158). It is
130 interesting to note that p21 and GADD45 are among the few genes demonstrated to be directly regulated by a KRAB zfp. p21 is repressed by binding the KRAB zfp ZBRK1 along with KAP1 on the p21 promoter, and this repression is lost upon induction of DNA damage and subsequent phosphorylation of KAP1
(159). These observations suggest a cross-talk between the two processes regulated by KAP1 and possible involvement of DNA damage response factors in transcriptional repression and vice versa.
In order to analyze this possibility, it is important to identify KAP1 interacting proteins that selectively interact with phosphorylated KAP1. An ideal approach would be to express either a constitutively phosphorylated (S824D) or a constitutively non-phosphorylated (S824A) mutant of KAP1 (161) in a KAP1 null background, and immunoprecipitate (IP) KAP1 in these two cell lines. Mass spectrometric analysis of proteins that selectively interact with the phosphorylated KAP1 would help us identify players associated with the role of
KAP1 in DNA damage response. As mentioned earlier, F9 cells that are KAP1 null would provide a good system to conduct these experiments. However, the role of
KAP-1 in mediating DNA damage response has only been studied so far using human cell lines. In order for us to use the F9 cells, these studies would have to be recapitulated using mouse cell lines such as F9 or NIH3T3. An alternative would be to conduct the proposed experiments in human cells that have been stably down regulated for endogenous KAP1 expression. However, residual
131 expression of the wild type protein could still interfere with the interpretation of the results derived from such an approach. This is especially so in the case of IP experiments as they concentrate the protein being immunoprecipitated and by extension, also the interacting partners of the given protein. The validity of immunoprecipitation results can be improved by incorporating a KAP1 IP using stable KAP1 knockdown cells that do not express any mutated forms of KAP1, thus eliminating proteins that are commonly pulled down by all three cell lines, from further Mass spectrometric analysis.
Another set of studies to examine the possibility of a cross-talk between the transcriptional and DNA damage response properties of KAP1 would be to study the effect of KAP1 mutants that inhibit KAP1 repressor ability (described in chapter III), on response to DNA damage. A series of mutant KAP1 that share a common S824D activating mutation in combination with mutations in HP1 BD,
PHD or bromo domains would help in uncoupling critical factors that regulate the different functions of KAP1. Another interesting experiment would be to test the involvement of a potential KRAB domain interaction in the process of KAP1- mediated chromatin relaxation. This suggestion is based on the observation that phosphorylation of KAP1 down regulates its sumoylation activity and results in chromatin relaxation in a manner that most likely does not involve HP1 interaction (158, 161). KAP1 sumoylation levels are increased upon KAP1-KRAB
132 interaction and expression of KRAB zfps has been shown to be sufficient to induce changes in KAP1 localization (86, 159).
Given these data, it is possible that phosphorylation changes the conformation of KAP1 in a manner that is unfavorable for KAP1-KRAB interaction.
Dissociation of KAP1 from KRAB could explain the chromatin relaxation observed and KAP1 relocation. In this respect it will be interesting to observe the ability of
KAP1 RBCC mutants to mediate DNA damage response. The RBCC domain of
KAP1 is responsible for direct interaction with the KRAB domain and it has been shown that mutations in the RBCC domain result in loss of KAP1 repressor activity (45). However, KAP1 is still capable of mediating repression even in the absence of RBCC domain, when targeted to DNA via a heterologous DNA binding domain (46). This observation indicates that the RBCC domain plays a prominent role in KAP1 targeting to KRAB enriched regions but does not contribute significantly to the silencing activity of KAP1. A similar situation might be true for the chromatin relaxation ability of KAP1. Hence analyzing RBCC mutants of KAP1 for the ability to induce chromatin relaxation following induction of DNA damage could help in dissecting the dual roles of KAP1 in repression and chromatin decondensation.
133
Gene Induction by Estrogen Receptor beta
CHAPTER V
INTRODUCTION, REVIEW OF LITERATURE AND STATEMENT OF PURPPOSE
Breast cancer treatment represents a success story that highlights the
usefulness of targeted therapies to cure cancer. However, the incidence of breast
cancer occurrence seems to be on a steady rise globally with 45% of the newly
reported breast cancer incidences arising in low and middle income countries
(162). There seems to be a correlation between increased incidence of breast
cancer and an urbanized lifestyle across the globe. The reason for this observed
correlation may be due to a combination of factors are more frequent in an urbanized population, including early age at menarche, increased fat content in the diet, increased age at first child birth, and reduced breast feeding (163).
Regular menstrual cycling leads to a constant exposure of breast tissues to circulating estrogens and the earlier the age at menarche, longer is the cumulative exposure to estrogens. Increased adipose tissue especially in post
menopausal women leads to increased production of estrogen, hence the heavier
the woman is after menopause, the more likely she is to develop breast cancer.
Early pregnancy and breast feeding on the other hand are thought to reduce breast cancer incidence by inducing terminal differentiation of cells that might
otherwise become tumorigenic. Not surprisingly environmental factors and
134 lifestyle changes are risk factors of developing breast cancer in women, along with increasing age and a familial history of breast cancer. With a global trend towards increased urbanization accompanied with change in lifestyle, effective strategies for the prevention and early diagnosis of breast cancer are the need of the hour and would go a long way in reducing the incidence of breast cancer globally.
Breast cancer types
Diagnosis of breast cancer also involves characterizing the type of cancer.
This is usually done by immunostaining for the expression of ERα, PR and HER2.
However subtyping breast cancers based on gene expression profiles together with standard histopathological procedures has yielded at least 3 distinct subtypes which differ with respect to the cell types involved, prognosis
(propensity to metastasize) and response to therapies. Among them Luminal A
(ER+ and/or PR+, HER2-) has the most favourable prognosis and responds well to selective estrogen receptor modulators (SERMS). The basal-like cancers (ER-
/PR-/HER2-/cytokeratins 5+, 14+, 17+) have the poorest prognosis and are treated with aggressive chemotherapy and inhibitors of angiogenesis. The luminobasal tumors (HER2+, ER+/-, PR+/-) respond to a combination of chemotherapy and the HER2 blocker Trastuzumab (164, 165). It is true that the availability of new drugs and a better understanding of cancer types have helped a lot in the treatment of breast cancer; however we still face problems of
135 resistance to chemotherapy, secondary cancers due to metastasis and recurrence, all of which contribute to breast cancer mortality. One way to circumvent these issues is to focus on developing drugs that can reduce the incidence of breast cancer by acting as chemopreventive agents. In order for such a strategy to be successful, we need to have knowledge of early events in breast cancer initiation and progression and also how risk factors such as estrogen exposure affect these early events.
In general, cancer is initiated due to the transformation of single cells and progression is achieved by the accumulation of genetic changes coupled with clonal selection and expansion. Accumulating evidence over the past few years indicates the existence of mammary stem cells that are at least bipotential and can give rise to both ER+ luminal epithelium and ER- myoepithelial cells (166).
The most compelling evidence for the existence of mammary stem cells comes from studies that demonstrate the regeneration of an entire mammary gland from a single cell (167, 168). During normal mammary development these stem cells can potentially be important players in mammary gland morphogenesis during pregnancy, lactation and involution. A series of studies from independent investigators has led to the hypothesis of a stem-cell driven model of breast carcinogenesis that can explain the heterogeneity of cell types found in breast cancers. The studies suggest the existence of a stem cell hierarchy with stem cell-types forming a continuum of phenotypes in vivo (164, 165, 169).
136
We can speculate that owing to aberrant differentiation programs, basal type cancers are derived from the most primitive progenitor cells (ER-, PR-
HER2-), the basoluminal cancers (HER2+, ER+/-, PR +/-) are derived from cells midway in the continuum, and the luminal cancers are derived from ER+ cells that have undergone transformation. Unfortunately, it is technically very challenging to isolate and study pure populations of these cell-types along the differentiation continuum. Even if such an analysis were undertaken, we have to be careful not to draw generalized conclusions as particular cancer subtypes occur differentially across populations. For example, immunohistochemical profiling of 496 women with breast cancers revealed that 36% of premenopausal
African-American women developed basal-like breast cancers as compared to only 16% of non African-american population who were more likely to develop luminal type breast cancer (59%) (170).
However studies that have analyzed the gene expression profiles of cells isolated from normal and cancerous breast tissue by sorting for expression of the cell surface markers CD44 and CD24. These studies indicate a distinctive profile for both of them. Cancer cells expressing CD44 had a more stem-cell like profile, and were more invasive and proliferative compared to the CD24+ cells (171-
173). It is interesting to note that CD44+ cells are also detected in normal breast tissue, but their number decreases following pregnancy (174), correlating with the predicted decline in progenitor cells due to pregnancy-induced differentiation.
137
While characterizing breast cancers by gene expression profiling has definitely been of advantage in terms of developing better targeted therapies, it still is not informative in analyzing the factors that predispose a putative stem cell to become a cancer cell. Hence studying contributing risk factors for breast cancer initiation is also of critical importance in devising strategies for breast cancer prevention.
Estrogen exposure as a risk factor for breast cancer.
Even though the risk factors for developing breast cancer include, age, reproductive history and inherited mutations (especially in the BRCA1 and BRCA2 genes), the strongest and most consistently reproducible risk factor has been lifetime exposure to estrogen (175-178). The strongest evidence highlighting the role of estrogens in initiating breast cancer comes from large scale clinical trials that included women taking an estrogen/progestin combination as part of hormone replacement therapy (HRT) in order to prevent osteoporosis and coronary heart disease. Analysis of risk/benefit ratio of HRT participants in the
WHI (Womens Health Initiative) trials revealed that the incidence of breast cancer increased by 53% following 5 years of HRT (179). Following the publication of this result, there was a sharp decline of 66% in the number of HRT prescriptions in 2002 and this was followed by a sharp decrease in the incidence of ER+/PR+ breast cancer among postmenopausal women in 2003 (180, 181).
138
This is not surprising given that estrogen activates ERα nuclear signaling, leading to cell proliferation. Increased rounds of DNA replication is in turn thought to increase the risk for accumulating mutations, resulting in tumorigenesis (182-184). Studies using breast epithelial cells in culture, animal breast cancer models and comparative analyses of normal and breast cancer samples from humans support a strong link between breast cancer initiation and levels of estrogen metabolites (175, 178, 185-187). Moreover, exposure to estrogen leads to a nearly 100% tumor incidence in ERα -/-/wnt-1 mice indicating that estrogens can act as carcinogens via an alternate pathway that does not involve ERα signaling (188).
Metabolism of estrogens
Estrone (E1) and 17β-Estradiol (E2), the pharmacologically active
metabolites of estrogen are hydroxylated selectively at the 2 and 4 positions by the cytochrome p450 family of enzymes to yield catechol estrogens (CE) (Figure
V-1, page 140) (189-191). CYP1A1 and CYP3A4 catalyze E2 to 2-OHE2 (2-
Hydroxyestradiol) and CYP1B1 catalyze conversion of E2 to 4-OHE2 (4-
Hydroxyestradiol) (192). The CE metabolites are highly unstable and further oxidizes to CE-quinones (CE-Q) via a semiquinone intermediate. 2-CE is catalyzed to 2-CE-Q (E2-2,3-quinone) by CYP1A1 and CYP3A4 while 4-CE is catalyzed to 4-
CE-Q (E2-3,4-quinone) by CYP1B1, CYPIA1 and CYP3A4 (192).
139
Figure V-1: Metabolism of estradiol
Italicized names represent enzymes that mediate indicated metabolic conversions.
140
The interconversion of CE-semiquinone and CE-quinone forms is catalyzed by
NADPH dependent cyp450 reductase (186).
CE-semiquinones are a highly reactive chemical species and can either react with molecular oxygen to generate quinone and superoxide radicals, or be nonenzymatically coupled with copper ions resulting in redox cycling that generates oxygen radicals (186, 193) leading to oxidative DNA damage (ODD)
(185, 186). A direct measure of the ability of E2 to induce oxidative DNA damage
(ODD) is evident by increased levels of 8-hydroxyguanine (8-OHdG) generated
due to the reaction of hydroxy radicals with guanine bases of the DNA (194).
Further, CE-Qs react directly with DNA and predominantly form the depurinating
adducts; 4-hydroxyestradiol-1-N3-adenine (4-OHE2-1-N3-Ade) and 4-
hydroxyestradiol-1-N7-guanine (4-OHE2-1-N7-Gua) (195-197).
Several studies report increased incidence of carcinogenesis with an
increase in the levels of catechol estrogens (178). Furthermore, activating mutations in the cytochrome P450 enzymes CYP1A1 and CYP1B1 are correlated with increased risk of breast cancer incidence (198-200). A number of studies have reported oxidative DNA damage followed by carcinogenesis in both cell and animal models of breast cancer (201). Error prone base excision repair of the depurinated DNA has been shown to give rise to mutations (202-204) that could predispose the cells to become cancerous (188). Taken together these studies
141 support the hypothesis that cumulative exposure to E2 can result in the initiation
of carcinogenesis.
Antioxidative enzymes involved in detoxification of estrogen
metabolites
A number of enzymes act to counter unchecked production of toxic
estrogen metabolites. These include catechol-O-methyltransferase (COMT),
quinone reductase (QR), glutathione-S-transferase (GSTpi) and γ-
glutamylcysteine synthetase heavy subunit (GCSh). Of these QR, GSTpi and
GCSh are regulated by an EpRE (Electrophile response element) sequence that is
typically regulated by the transcription factor Nrf2 (NE-EF-Related factor 2). As
mentioned earlier, E2 is metabolized to the CEs 2-OHE2 and 4-OHE2 both of which
are acted upon by COMT. Both 2-CE and 4-CE are converted to non-toxic O-
methyl conjugates by COMT which uses S-adenosyl-L-methionine (SAM) as the
methyl donor (205). In addition to getting methoxy-conjugated, CEs can also be
inactivated by sulfation and glucuronidation. Both of these conjugations increase
the water solubility of the target compound rendering it less toxic and easily excreted from the body. Sulfate conjugation is catalyzed by cytosolic sulfotransferases (SULTs) that catalyze the transfer of a sulfonate group from the active sulfate, 3-phosphoadenosine 5-phosphosulfate (PAPS), to an acceptor substrate compound containing a hydroxyl or an amino group (206). Conjugation to glucoronic acid is catalyzed by UDP-glucuronosyltransferases (UGTs). UGTs
142 are membrane-bound enzymes that are present in the endoplasmic reticulum, which mediate transfer of the ubiquitous co-substrate glucuronic acid group of uridine diphospho-glucuronic acid to the functional group (e.g. hydroxyl, carboxyl, amino, sulfur) of a specific substrate (207).
In vitro studies using recombinant SULTS indicate that at least 6 isoforms
catalyze the conjugation of E1, E2, CEs and their methoxy derivatives (206) with
SULT1 E1 being the most efficient enzyme. Interestingly, methylation of CEs
predominantly preceeds their sulfation, and inhibition of the methylating enzyme
COMT also leads to lower levels of sulfated CE conjugates (206). These results
indicate a concerted action of COMT and SULTs in the metabolism of CEs. In the
case of glucuronidation, 6 different UGTs have been shown to catalyze the
conjugation of E1, E2, CE and their methoxy conjugates (208, 209). Unlike
sulfation, glucuronidation seems to be independent of the methylation of CEs.
Conjugation of 2-CE is selectively and efficiently catalyzed by UGT1A1 and
UGT1A8 while 4-CE is selectively catalyzed by UGT2B7 (209). UGT enzymes are
highly polymorphic and it is possible that polymorphisms in these enzymes can
affect CE levels inside breast tissue and can be a contributing factor to breast
cancer initiation. However, no significant correlation has been found so far
between various polymorphisms of UGT1A1 and breast cancer risk (210, 211).
143
Even though sulfation and glucuronidation constiute a major pathway of detoxyfiying xenobiotics, it seems unlikely that any particular enzyme (SULT or
UGT) by itself plays a major role in the metabolism of CEs. The observation that conversion of CEs can be catalyzed at different efficiencies by at least 5 or more isoforms of either SULTS or UGTs suggests that the levels of CE-sulfate and CE- glucuronide conjugates depend on the relative activities of a group of enzymes rather than a single one. Unlike the existence of multiple SULTS and UGTs that catalyze CEs, CE-methoxy conjugates are predominantly formed by the COMT enzyme. This scenario is different that of CE-methoxy conjugates which are predominantly catalyzed by COMT. However, CE-methoxy conjugates could not be detected in female mice models of breast cancer, but instead high levels of
CE-GSH conjugates were detected. Also analysis of CE-conjugates in breast tissues from women with and without breast cancer revealed significantly higher levels of CE-GSH conjugates in breast cancer tissues compared to normal tissues
(212). These observations clearly suggest that COMT activity is insufficient by itself to maintain a balance between toxic and non-toxic estrogen metabolites.
In this view, it is significant that TOT-ERβ (tamoxifen-liganded ERβ) can induce expression of the antioxidative enzymes QR, GSTpi and GCSh all of which have direct roles in the detoxification of E2 metabolites (213). QR is an NADP(H)
dependent oxidoreductase that catalyzes a two electron reduction of quinones to
hydroquinones (214). Such a reduction prevents quinone-semiquinone
144 interconversion that gives rise to reactive hydroxyl radicals. In humans QR activity is high in many extrahepatic tissues including breast epithelial cells where
QR activity could be a key factor in inhibiting the initiation of E2-induced carcinogenesis. In vitro studies of QR activity using soft ionization electrospray
ionization–mass spectrometry (ESI-MS) techniques demonstrate the binding of
QR to CE-Q followed by the reduction of CE-Q (215). The importance of QR in
breast cancer is highlighted by considering that polymorphisms in the NQO1
gene that encodes for QR are associated with significant risk for breast cancer
(216, 217). A recent study indicated that the presence of a common homozygous
missense variant of NQO1 (558C>T) that disables NQO1 strongly correlates with poor prognosis for women with breast cancer, thus implicating loss of QR activity
in breast cancer progression (218). At a cellular level, breast epithelial cells down
regulated for QR expression exhibit increased sensitivity to E2-induced ODD
(219). Long term E2 treatment coupled with down regulation of QR expression
leads to transformation of normal breast epithelial cells. This correlates with increased levels of CE upon loss of QR expression (220).
The above mentioned observations clearly indicate the key role played by
QR in maintaining the oxidative homeostasis in breast epithelial cells. Apart from
QR, the enzyme activities of GSTpi and GCSh also directly contribute to the detoxification of CE-Qs. The GSTpi family of enzymes catalyzes the conjugation
of a variety of electrophilic compounds such as quinones to Glutathione (GSH).
145
GSTpi activity outside hepatic tissues is mediated solely by GSTP1-1(221). GSH
(γ-glutamyl-cysteinyl-glycine) is a non protein thiol and the predominant cellular antioxidant present in the cells with concentrations ranging from 1-10mM. GSTpi catalyzes the conjugation of reactive intermediates to form S-substituted GSH- adducts through the nucleophilic cysteine sulfydryl group of GSH. GSH- conjugates can be further converted to N-acetylcysteins and eventually excreted out of the system (222). As such the activity of GSTpi plays a major role in maintaining the redox balance inside the cells. The importance of GSTP1-1 activity in maintaining the redox balance in breast epithelial cells is highlighted by comparative studies of normal and cancerous breast tissues. These studies indicate that GSTpi promoter is frequently hypermethylated resulting in transcriptional silencing of GSTpi expression. Further, the studies also indicate that silencing of GSTpi is an early event in breast cancer progression, possibly contributing to breast cancer initiation (223, 224). The heavy (catalytic) subunit of GCS catalyzes the rate limiting step in the de novo synthesis of the
cellular antioxidant γ-glutamyl-cysteine (GSH) and thus directly controls the
redox balance within cells (225, 226).
Antiestrogen treatment of breast cancer
Perhaps the most successful anti-breast cancer drug is the non-steroidal compound tamoxifen. Tamoxifen is a selective estrogen receptor modulator
(SERM). It acts as an antagonist of ERα in the breast and an agonist in the bone and endometrium. This property of tamoxifen is highly desirable as women can
146 be treated for a longer term with tamoxifen without increasing the risk for developing osteoporosis as estrogen signaling is important for bone development. Tamoxifen was first introduced as a treatment option for advanced breast cancer in 1971 (227) and quickly came to be used as a first line therapy for early breast cancer, especially in ER+ cancers. Large scale randomized clinical trials indicated an overall 47% reduction in breast cancer recurrence following 5 years of tamoxifen treatment and 26% reduction in mortality. This trend was observed irrespective of age and menopausal status. Also, post- treatment follow up studies indicated a beneficial influence on survival beyond
10 years in early breast cancers treated with tamoxifen (228).
Improvements in the concept of selective modulation of estrogen signaling led to the development of raloxifene which is comparable to tamoxifen in reducing breast cancer incidence in high-risk women. Also Raloxifene treatment resulted in a reduced incidence of endometrial cancers as compared to tamoxifen, thus yielding a better safety profile for raloxifene (229). However even though tamoxifen and raloxifene are comparable in reducing risk for invasive cancer, tamoxifen is more effective in reducing the risk for ductal carcinoma in situ (230, 231). More importantly, tamoxifen is the only drug so far that has been shown to have beneficial side effects in the long run, even more than 5 years after the treatment has been stopped. Also, the adverse effects of
147 tamoxifen decrease after the treatment has been stopped while the beneficial effects continue to be evident (231-233).
The ability of SERMs to act as anti-cancer agents is attributed predominantly to their ability to compete with estrogen for binding to ERα and block ERα-mediated transcriptional activation of cell proliferation. Another strategy of blocking estrogen action is estrogen deprivation. Clinical trials including women who have been on estrogen deprivation therapy indicate a beneficial effect greater than that of tamoxifen (234). Most commonly estrogen deprivation is achieved by blocking the enzyme cytochrome P450 aromatase which is responsible for the synthesis of estrogen. However, aromatase inhibitors can be used only sparingly in premenopausal women and are limited to breast cancer patients who are postmenopausal.
Tamoxifen and chemoprevention of breast cancer
Women treated with tamoxifen continue to derive benefits long after they have stopped taking the drug. This finding led to large scale clinical trials that analyzed the use of tamoxifen as a chemopreventive measure for women at high risk to develop breast cancer. At least four large scale clinical trials evaluating the ability of tamoxifen and raloxifene concluded that both drugs had similar effects in reducing the risk for breast cancer. Tamoxifen treatment reduced the risk of invasive cancer by 58% while the incidence of benign breast tumor was reduced
148 by 38% (229, 235). However, raloxifene was equally effective and also had a better safety profile. Unlike tamoxifene, raloxifene did not increase the incidence of endometrial cancer. It has to be noted that raloxifene even though effective in postmenopausal women, is not indicated for premenopausal women. Also the incidence of non invasive breast cancer was higher with Raloxifene usage.
Tamoxifen especially was observed to be more effective in preventing ductal carcinoma in situ than raloxifene (236). While it is true that these results are
biased towards ERα positive breast cancers, about 5% to 10% of the patients
diagnosed with ERα negative cancers also respond to tamoxifen therapy (237).
Unfortunately the status of ERβ expression has not been determined in these
patients. Other studies have found that a correlation exists between favourable
response to tamoxifen treatment and expression on ERβ. However, ERβ acts as a
positive prognostic factor only in the absence of ERα (237).
Role of ERβ in breast cancer
ERβ was first discovered in rat prostrate and ovary in 1996 and studies
using ERβ KO mice have shown that ERβ expression is essential for maintenance
of differentiated phenotype in mammary glands (238). The role of ERβ in breast
cancer is currently being investigated. Analysis of ERβ expression at different
stages of breast cancer progression indicates a steady decrease in ERβ
expression levels with increased tumor grade (239-241). This suggests that at
least certain subtypes of breast cancers develop due to downregulation of ERβ
149 expression. This suggestion becomes stronger when we consider the observation that ERβ expression in ERα negative cancers acts as a positive prognostic marker, and such cancers are also responsive to tamoxifen treatment in spite of the absence of ERα expression. These observations suggest that tamoxifen liganded-ERβ is capable of regulating a different set of genes from that of ERα.
Many investigations into the influence of ERβ on breast cancer progression have been hampered by the existence of at least 5 distinct splice variants of ERβ.
(Figure V-2, page 151) Also, it is not possible to directly compare the results of
ERβ expression studies as some of the studies use mRNA levels while others use protein expression as a measure of ERβ expression. So far, protein expression has been demonstrated only in the cases of ERβ1, the full length protein, and
ERβ2, which contains an alternative exon 8 sequence (242). Both ERβ1 and
ERβ2 can bind to DNA in both liganded and unliganded states. They can also interfere with ERα-mediated transcription from ERE-regulated genes. However only ERβ1 has transcriptional activity of its own (243).
150
Figure V-2: Representation of cDNA variants that encode ERβ isoforms.
Full length ERβ consists (top) of a transcriptional activation domain (A/B); DNA binding domain (C); a hinge region (D); ligand binding domain (E) and an additional transactivation domain (F). Isoforms of ERβ are either deletion splice variants or have an additional stretch of 18 amino acids in their ligand binding domains
151
Transcriptional regulation by ERα and ERβ
To appreciate transcriptional regulation by ERβ, we have to compare it with ERα-mediated regulation. Similar to ERα, ERβ dimerizes and binds to both classical ERE and non-ERE sequences in order to regulate transcription.
However, unlike ERα, ERβ seems to have a weaker corresponding AF-1 function and thus depends more on the ligand-dependent AF-2 for its transcriptional activation function (244). Classically, ligand-activated ERs bind specifically to
DNA at EREs through their DNA binding domains and bring coregulators to the transcription start site. The consensus ERE consists of two half-sites
(aGGTCAnnnTGACCt) separated by a three-nucleotide spacer. However, many natural EREs deviate substantially from the consensus sequences (18).
Liganded ERα and ERβ can also modulate gene expression via interacting with other transcription factors, such as activating protein-1 (AP-1) and stimulating protein 1 (Sp1) and thus regulate non-ERE genes (245, 246).
However, in the presence of estrogen, ERα induces AP-1-driven reporter activity, whereas ERβ has no effect (247). Raloxifene binding to ERβ induces transcriptional activity through an AP-1 site, whereas binding to ERα results in minimal activation examined under the same conditions. ERβ activated an RARα1 promoter-reporter construct presumably by the formation of an ER:Sp1 complex
(248). Antagonist binding to ERβ caused an increase in reporter gene expression.
This effect was blocked by estrogen, which resembles the effect of ERβ on an
152
AP-1 site. Moreover, ERα and ERβ also exhibit different transcriptional effects in regulation of the cyclin D1 promoter (249). ERα mediates the stimulatory effect of estrogen on cyclin D1 expression, whereas ERβ has a repressive effect.
However, both ERα and ERβ induce the expression of cyclin D1 in response to antiestrogens.
ERβ regulates gene expression programs distinct from that of ERα
It is evident from the studies described above that ERα and ERβ regulate distinct sets of genes in addition to genes that are regulated by both the isoforms. In view of the positive prognostic value of ERβ in breast cancers several studies have been performed in breast cancer cell lines stably expressing
ERβ. The results of some of these studies are summarized in table 1 and clearly indicate distinct transcription profiles in response to E2-liganded ERα and ERβ.
However, not much is known about regulation of tamoxifen-liganded
transcription by ERβ. Tamoxifen acts as a partial antagonist of ERα for
transcription from classical ERE-regulated genes, but acts as a complete
antagonist for ERβ (250, 251). However, tamoxifen liganded ERβ activates
transcription from AP-1 sites via Jun/Fos complexs (252).
Interestingly, tamoxifen-ERβ also activates transcription of phase II
antioxidative enzymes such as quinone reductase (QR). Further, this activation is
dependent on a 5’ regulatory element EpRE (electrophile response element). ERα
153 on the other hand does not seem to have any effect on EpRE-genes either when liganded to tamoxifen or estrogen. These and other studies have indicated that the transcriptional activity of ERβ is dependent on the promoter context and also on the ligand bound.
ERβ-mediated regulation of antioxidative genes via the EpRE sequence
Experiments employing differential display RNA methods revealed that the antioxidative gene QR is upregulated in MCF7 cells in response to antiestrogens, while downregulated in the presence of estrogen (253, 254). Further studies demonstrated the ability of the EpRE sequence to be preferentially activated by tamoxifen-liganded ERβ. Studies from several laboratories reveal that EpRE sequence is activated by the formation of hydrogen peroxide generated by redox cycling of antioxidants and also through electrophilic compounds (255). The consensus sequence of EpRE contains a sequence very similar to that of TRE
(12-O-tetradecanoylphorbol 13-acetate response element). Sequence comparisons of EpRE elements reveals the existence of AP1 binding sites as part of the EpRE sequence (256). Mutational analysis indicates that the core EpRE consensus sequence is most likely: GTGACAnnnGC (257). However, the neighboring TRE sequences have also shown to be involved in effective induction of transcription from EpRE-regulated promoters (258-260). In response to oxygen radicals or electrophiles, EpRE is activated by the protein Nrf2. In fact,
Nrf2 is shown to be essential for both basal and induced expression of EpRE-
154 genes (261-263). Nrf2 binds directly to the EpRE sequence via a basic leucine zipper domain as a heterodimer either with Jun or small Maf proteins (264-266).
The existence of AP-1 sites and the recruitment of Fos/Jun to the EpRE site complicate the study of EpRE regulation. However, at least in the case of NQO1, it has been shown that EpRE is induced independently of AP-1 sites in response to antioxidants (267).
In addition to Nrf2, transcription of certain EpRE regulated genes can be induced by TOT-liganded ERβ (213, 254). Mutational analysis of the EpRE regions present 5’ of the antioxidative genes NQO1, GSTpi and GCSh reveal that
TOT-ERβ induced transcription is dependent on the presence of intact EpRE sequences. Studies using dominant negative mutants of Nrf2 and Fos indicate that Nrf2 is necessary for transcriptional induction by ERβ while the role of Fos was found to be context dependent. However, cells lacking ERβ expression do not exhibit transcriptional induction by TOT even in the presence of Nrf2, indicating that Nrf2 is necessary but not sufficient to mediate this effect (268).
The ability of ERβ to induce EpRE genes leads to the speculation that ERβ might bind to the EpRE element. In vitro gel shift assays indicate that ERβ indirectly
binds to the EpRE sequences from the GCSh and GSTpi genes as a complex with
Nrf2. However, ERβ could also bind to the GSTpi EpRE via a Fos/Jun interaction
but not to the GCSh EpRE (213). These studies suggest that TOT-ERβ mediated
regulation is highly context dependent. Another key regulator of ERβ activity at
155 the EpRE was discovered when yeast 2 hybrid assays using ERβ as bait pulled down the protein hPMC2. In vitro hPMC2 selectively interacts with ERβ over ERα and results in increased induction of QR EpRE by TOT-ERβ (269).
Inhibition of E2-induced oxidative DNA damage by tamoxifen
TOT-ERβ increases expression levels of antioxidative enzymes involved in
detoxification of E2 metabolism. As a result, tamoxifen treatment resulted in
inhibition of E2 -induced ODD. Interestingly, this property of tamoxifen is evident
only in the presence of ERβ (268). The ability of tamoxifen to protect against
ODD is compromised upon down regulation of QR or GSTpi, indicating that the
effect is most likely due to TOT-ERβ mediated regulation of these enzymes. This
supposition is further strengthened by observations in rat models of breast
cancer where increased ODD is observed in the mammary epithelial cells upon
prolonged exposure to E2. Simultaneous treatment of the rats with tamoxifen
significantly reduced the levels of ODD. Moreover, reduction in the levels of ODD
correlated with increased expression of QR in the mammary epithelium (220).
Studies analyzing the expression levels of tumor suppressor enzymes in women
treated with tamoxifen for breast cancer revealed that the expression levels of
GSTpi were significantly higher in women who responded positively to tamoxifen
treatment (270). Also, transformation of normal breast epithelial cells due to long
term E2 exposure was significantly inhibited by co-treatment with tamoxifen
156
(220). These observations indicate that tamoxifen can be used as a chemopreventive agent to inhibit estrogen-induced carcinogenesis.
The ability of TOT-ERβ to upregulate antioxidative genes together with its ability to mediate protection against ODD reveal a new role for antiestrogens in the treatment of breast cancer, distinct from that of blocking ERα transcription.
However, the transcriptional mechanisms involved in ERβ-mediated regulation of genes in the context of breast cancer, is not well defined. Even less is known about the molecular events that result in ERβ-mediated induction of EpRE-genes and about the key players involved in mediating such regulation. Studying these mechanisms and defining transcription complexes involved in regulating this process is needed for devising better targeted therapies that have minimal side effects.
STATEMENT OF PURPOSE
Tamoxifen is widely used in adjuvant therapy for treatment of breast cancer. Selective ER modulators such as tamoxifen are thought to act as anticancer drugs mainly by blocking estrogen receptor α (ERα) mediated cell proliferation. However, based on a series of studies from our laboratory, we observed that trans-hydroxytamoxifen (TOT) treatment can protect breast
epithelial cells against estrogen-induced oxidative DNA damage (ODD) in the
157 absence of ERα, but not ERβ. This is interesting when viewed in the context of studies that report a strong correlation between exposure to ODD causing agents including estrogen (E2) and increased risk of breast cancer. In the presence of
ERβ, TOT can induce transcription of anitoxidative genes such as QR, GCSh and
GSTpi, all of which are regulated by the electrophile response element (EpRE).
ER negative breast epithelial cells with down regulated levels of QR undergo
transformation upon long term E2 treatment and the incidence of transformation
is significantly reduced upon restoring ERβ expression and simultaneous
treatment with tamoxifen. The observation that ERβ directly interacts with a
novel protein, hPMC2, and binds to the EpRE sequence in vitro, offered a clue to
the mechanism of antioxidative gene regulation by ERβ. Based on our observations so far, we hypothesized that: Tamoxifen inhibits E2-induced ODD in
breast epithelial cells by upregulation of antioxidative stress enzymes resulting in
increased detoxification of E2 metabolites via an ERβ/hPMC2 dependent pathway.
Hence the purpose of this part of the dissertation was to elucidate the molecular
mechanism and define the transcriptional regulators employed by TOT-ERβ in
order to mediate the induction of EpRE-genes.
158
CHAPTER VI
hPMC2 is Required for Recruiting an ERβ Co-Activator Complex to Mediate
Transcriptional Upregulation of NQO1 and Protection Against Oxidative DNA
Damage by Tamoxifen
(Sripathy SP et al., 2008. Oncogene. In press)
INTRODUCTION
Prolonged exposure to estrogen is strongly associated with increased risk
for developing breast cancer (178). Metabolism of estrogens in breast epithelial
cells generates highly reactive catechol estrogen quinones (CE-Q) that form
mutagenic DNA adducts. Further, the interconversion of quinone-semiquinone
forms generates reactive oxygen species (ROS), which induce oxidative stress
inside the cells (188). The sustained oxidative stress, together with the
mutagenic potential of CE-Qs contributes to the initiation and progression of
breast cancer (201). This hypothesis is strengthened by our observation that
down regulation of the antioxidative enzyme QR, coupled with exposure to 17β-
estradiol (E2) leads to increased levels of CE-Q and transformation of non-
tumorigenic breast epithelial cells (220). QR catalyzes the reduction of estradiol-
3,4-quinone, thus preventing generation of ROS by quinone-semiquinone
interconversion (215). Also, a strong correlation exists between breast cancer incidence and polymorphisms in several antioxidative genes including NQO1,
159 indicating an important role for antioxidative enzymes in breast cancer prevention and treatment (271-274).
We observed that QR expression is increased in mammary glands of rats treated with tamoxifen and this increased expression correlates with a decrease in E2-induced ODD levels (220). Also, treatment of human breast epithelial cells
with trans-hydroxytamoxifen (TOT) prevents E2-induced increase in CE-Q levels
(220). This ODD protective ability of tamoxifen makes it an attractive candidate
for chemoprevention of breast cancer. In fact, the recently concluded STAR and
IBIS breast cancer trials highlight the advantage of using tamoxifen and
raloxifene as chemopreventive agents (236, 275).
TOT treatment increases activity of reporter genes regulated by an EpRE
sequence (213). EpRE is a cis regulatory element present upstream of many
antioxidative gene promoters including GCSH, GSTP1 and NQO1 (257, 276-278).
GSTpi detoxifies CE-Qs by conjugation with the cellular antioxidant glutathione,
while GCSh catalyzes the rate limiting step in the de novo synthesis of
glutathione (201). EpRE-regulated reporter genes are activated by TOT-ERβ, but
not significantly by TOT-ERα, and the ODD protective ability of TOT is not
observed in the absence of ERβ (213, 268). These data indicate a key role for
TOT-ERβ in mediating oxidative stress response via regulation of EpRE
promoters.
160
The role of tamoxifen in regulating ERα-dependent transcription is well characterized, but mechanism of ERβ-mediated transcription in breast cancer cells is not well defined. Even less is known about the pathway responsible for
TOT-ERβ mediated upregulation of EpRE-regulated genes. Hence, our goal in the current study was to analyze TOT-mediated events at the EpRE locus and thereby define the transcription co-activator complex responsible for ERβ- dependent upregulation of EpRE genes.
RESULTS
The ERβ interacting protein, hPMC2 is important for tamoxifen- mediated protection against estrogen induced oxidative DNA damage
TOT-induced transcription of EpRE genes is further enhanced by the ERβ- interacting protein hPMC2 (Prevention of Mitotic Catastrophe) (279), suggesting a role for hPMC2 in TOT-mediated protection against ODD. Hence, we transiently down-regulated the expressions of both hPMC2 and QR (Figure VI-1A, page 181) by retroviral delivery of shRNA. Cells were immunostained for 8-OHdG to analyze the relative roles of hPMC2 and QR in mediating TOT-dependent protection against E2-induced ODD.
161
Cells expressing knockdown levels of either hPMC2 or QR displayed higher basal levels of ODD compared to control-infected cells, while E2 treatment
resulted in >2 fold increase in the ODD levels of all three cell lines (hPMC2
shRNA, QR shRNA, control shRNA) (Figure VI-1B, page 181). Treatment with
TOT alone revealed 8-OHdG levels lesser than or similar to that of the ethanol-
treated samples, indicating that TOT by itself does not contribute to ODD.
Knockdown of QR did not result in a complete loss of the protective ability of
TOT, in confirmation with our previous observations (268). This finding suggests
the contribution of other antioxidative enzymes (GSTpi, GCSh etc.) in mediating
the ODD protective ability of TOT. However, in hPMC2 knockdown cells 8-OHdG
levels were similar to that of the E2-treated samples even with the TOT
treatment, indicating a requirement for hPMC2 in TOT-ERβ mediated protection
against E2-induced ODD.
In vitro studies indicate selective interaction of hPMC2 with ERβ (280).
Immunoprecipitation of MCF7 lysates for unliganded, E2-liganded and TOT-
liganded ERβ resulted in co-immunoprecipitation of hPMC2 in all three cases.
Conversely, hPMC2 immunoprecipitation pulled down ERβ both in the presence
and absence of E2 or TOT, indicating constitutive interaction between ERβ and hPMC2 (Figure VI-1C, page 181). Taken together, our data indicate a novel role for hPMC2 in TOT-mediated protection against ODD by selective interaction with
ERβ to induce antioxidative gene expression.
162
Tamoxifen treatment results in the recruitment of ERβ, hPMC2 and transcriptional coactivators to the EpRE
ERβ and hPMC2 interact in vivo and hPMC2 together with ERβ binds to the
EpRE sequence in vitro (279). These data suggest potential recruitment of ERβ
and hPMC2 to the EpRE regions in vivo. To test the TOT-dependent recruitment
of these proteins, we used ChIP to analyze the EpRE region of the NQO1 gene
(Figure VI-2A, page 183). We treated the cells for 3 hours with the indicated ligands, as that was the earliest time point at which we observed transcriptional induction (data not shown). To identify other factors involved in ERβ-mediated induction of EpRE, we first studied recruitment of known ERα-associated transcriptional coactivators: SRC-1 (Steroid Receptor Coactivator-1), PARP-1
(poly (ADP-ribose) polymerase 1) and Topoisomerase IIβ (281, 282). We also tested for the recruitment of the EpRE-binding transcriptional activator Nrf2 (NF-
E2-related factor-2) (261).
We observed weak recruitment of PARP-1 and Nrf2 in control-treated samples but a strong recruitment of all the factors tested in the TOT-treated cells indicating a predominantly TOT-dependent recruitment of not only ERβ, hPMC2 and Nrf2, but also ERα-associated coactivators (Figures VI-2B and VI-2C, page
183). Analysis of TOT-treated samples revealed little or no recruitment of any of the proteins tested either at ~800bp upstream of the EpRE or at the NQO1
163 promoter region located ~400bp downstream to the EpRE (Figure VI-2D, page
183), suggesting a localized and selective recruitment to the EpRE region.
An initial ChIP of the TOT-treated samples with an antibody to hPMC2, followed by re-immunoprecipitation of the chromatin using antibodies to the indicated proteins confirmed mutual recruitment of ERβ, hPMC2, Nrf2, ERα and
ERα-associated coactivators to the EpRE (Figure VI-2E, page 183). Taken together, the data indicate that TOT-ERβ together with hPMC2, recruits an ERα- like activation complex localized to the EpRE region, resulting in transcriptional induction.
ERβ and hPMC2 are required for effective inhibition of estrogen- induced oxidative DNA damage by tamoxifen
To examine the ERα-independent role of ERβ and hPMC2 in TOT- mediated induction of EpRE and in protection against E2-induced ODD, we used
the ER negative, non-tumorigenic breast epithelial cell line, MCF10A (220). The
lack of ERα expression enabled us to examine the ability of ERβ and hPMC2 to
recruit ER-associated coactivators and for inducing antioxidative genes in the
absence of ERα. We established two MCF10A cell lines, both of which
constitutively express FLAG-ERβ (FL-ERβ), and one of them expresses stably
knocked down levels of hPMC2 (hPMC2mi) (Figure VI-3A, page 185).
164
No significant change was observed in expression levels of QR, GCSh and
GSTpi upon TOT treatment of MCF10A(P) cells (Figure VI-3B, page 185). In contrast, a 2- to 3-fold increase in the expression of all three enzymes in FL-ERβ stables was observed, indicating that the transcriptional induction by TOT requires ERβ. Also, no significant changes were observed in antioxidative enzyme expression levels upon TOT treatment of hPMC2mi cells, strongly suggesting a requirement for hPMC2 in this process (Figure VI-3B, page 185). Similar results were obtained when the cell lines were tested for the ability of TOT to protect against E2-induced ODD (Figure VI-3C, page 185). The protective ability of TOT
was severely attenuated in the absence of either ERβ or hPMC2.
We also measured the effect of E2 treatment on the levels of catechol estrogens and their quinone-conjugates in the presence and absence of TOT
(Figure VI-3D, page 185). We measured the levels of 4-Hydroxyestradiol-quinone conjugate (4-con) as this metabolite is predominantly responsible for the formation of mutagenic DNA adducts (195). Following E2-treatment, we observed
comparable levels of 4-con induction in MCF10A(P) and FL-ERβ cells, while cells
down-regulated for hPMC2 exhibited lower levels of 4-con. This could be either
due to a slower conversion of 4-OHE to 4-con or increased metabolism of 4-con
to form DNA adducts in the absence of hPMC2. The similar levels of E2-induced
ODD across all three cell lines (Figure VI-3C, page 185), suggest that the lower
165 levels of 4-con observed in E2-treated hPMC2mi cells may not be reflective of
lower levels of 4-conformation.
As a comparison, we also measured the levels of 2-Hydroxyestradiol
quinone-conjugates (2-con). The FL-ERβ cells revealed significantly higher levels
of 2-con metabolite in response to E2 treatment in comparison with the other two
cell lines tested. Since the levels of 2-OHE are comparable across the cell lines, the higher levels of 2-con more likely reflects slower catabolism of the 2-con metabolite as a result of ERβ expression.
More importantly, TOT treatment reduced both 4- and 2-Hydroxyestradiol- quinone conjugates levels by 4- to 6-fold in FL-ERβ cells, while very little or no reduction were observed in the absence of ERβ (MCF10A(P) cells), or hPMC2 (FL-
ERβ/hPMC2mi cells). This indicated that TOT-dependent decrease in the levels of E2-induced quinone-conjugates is dependent on the presence of ERβ and hPMC2 and not on the levels of the conjugates themselves. Further, 2-OHE2 and
4-OHE2 (2-OHE and 4-OHE) levels were similar both in the presence and absence
of TOT in all three cell lines, indicating that the reduction in the levels of 2-con and 4-con is not simply a result of lower levels of 2-OHE2 and 4-OHE2 in
response to TOT treatment.
166
Collectively, these results demonstrate a strong link between expression levels of antioxidative enzymes and levels of quinone-conjugates, and underscore the key roles played by ERβ and hPMC2 in mediating oxidative stress responses.
Tamoxifen-mediated recruitment of the coactivators PARP-1, topoisomerase IIβ and SRC-1 to the EpRE, is dependent on both ERβ and hPMC2
TOT treatment resulted in the recruitment of an ERα-like activation complex at the EpRE that included both ERα and ERβ (Figure VI-2, page 187).
However, our previous studies indicate that TOT-ERα alone does not significantly induce antioxidative gene expression but TOT-ERβ induces expression even in the absence of ERα (213). Hence, we analyzed the E2- and TOT-dependent
recruitment of the ERβ, hPMC2, Nrf2 and ERα-associated coactivators to the
EpRE region, in MCF10A FL-ERβ cells (Figures VI-4A and VI-4B, page 187). We
observed TOT-dependent recruitment of Nrf2, ERβ, hPMC2, PARP-1,
topoisomerase IIβ and SRC-1, indicating that absence of ERα does not
significantly affect the recruitment of coactivators. However, induction levels of
antioxidative genes by TOT in MCF7 and MCF10A FL-ERβ cells are comparable
(Figure VI-5 and Figure VI-3B, page 187), suggesting that ERα recruitment is
unlikely to be a major contributor to ERβ-dependent induction of EpRE
promoters. In the absence of either ERβ or hPMC2 we observed little or no
recruitment of the factors tested, except for Nrf2. This indicates a dependence
167 on both ERβ and hPMC2 for TOT-mediated recruitment of the coactivator complex at the EpRE.
The above results taken together confirms our observations in MCF7 cells
(Figure VI-2, page 187) that ERβ and hPMC2 are capable of assembling an E2- liganded ERα-like transcription activation complex at the EpRE in response to
TOT. More importantly, this result is not cell line specific and demonstrates that
ERβ can assemble a functional ERα-like transcriptional complex in the absence of
ERα.
PARP-1 is involved in Tamoxifen-mediated increase of antioxidative enzyme expression
PARP-1 and topoisomerase IIβ together, are required for E2-dependent activation by ERα and only PARP-1 is recruited by TOT-ERα, resulting in
repression (281). However, we observe a TOT-dependent co-recruitment of
PARP-1 and topoisomerase IIβ along with hPMC2 and ERβ (Figures VI-2 and VI-
4, pages 183 and 187). To investigate the functional relevance of PARP-1
recruitment in the induction of EpRE-regulated genes, we transiently down-
regulated PARP-1 expression by siRNA transfection in MCF10A FL-ERβ cells
(Figure VI-6A, page 191). Down regulation of PARP-1 resulted in >50%
decrease in the ability of TOT to induce the antioxidative enzymes GCSh, QR and
GSTpi (Figure VI-6B, page 191). Similar results were obtained by four different
168 siRNA oligos, targeted against different regions of PARP-1 mRNA, indicating that the result is unlikely to be an off target effect (data not shown). The basal expression levels of GCSh, QR and GSTpi were comparable in both siRNA transfected and control cells (Figure VI-7, page 193), indicating that the enzyme expression levels in TOT-treated samples were not the consequence of a general decrease in protein expression. This observation, along with TOT-dependent recruitment to the EpRE, suggests an important role for PARP-1 in ERβ-mediated induction of antioxidative genes.
DISCUSSION
Our studies demonstrate a TOT-dependent, localized co-recruitment of the transcription factors: ERβ, hPMC2, PARP-1, topoisomerase IIβ, SRC-1, ERα and
Nrf2, to the EpRE. Recruitment of the entire complement of coactivators is observed both in the presence and absence of ERα. More importantly, except for
Nrf2, recruitment of all other coactivators is dependent upon the presence of both ERβ and hPMC2. Together, our data indicate TOT-dependent targeting of
ERβ and hPMC2 to the EpRE, with subsequent recruitment of other transcriptional regulators to form a coactivator complex resulting in transcriptional induction. The key roles played by ERβ and hPMC2 are highlighted by our observations that, lack of either of these factors results in an almost complete loss of 1) TOT-induced antioxidative gene expression, 2) TOT- dependent decrease in the levels of catechol estrogen quinones and 3) protection against E2-induced ODD by TOT. Transcriptional induction by ERβ and
169 hPMC2 at least in part is dependent upon PARP-1 recruitment, as down regulation of PARP-1 expression results in attenuated induction of antioxidative genes by TOT. However, our observations do not support a significant role for
ERα and Nrf2 in TOT-ERβ mediated transcription at the EpRE.
We verified induced transcription of the EpRE- regulated genes NQO1,
GCSH and GSTP1 in MCF7 breast epithelial cells (Figure VI-5, page 189). This induction exhibited a strong correlation with the ability of TOT to protect against
E2-induced ODD, as evidenced by the fact that absence of either ERβ or hPMC2
results in loss of both TOT-dependent gene induction and also protection against
E2-induced ODD after TOT treatment (Figures VI-1B and VI-3, pages 181 and
185). The EpRE-transcription factor Nrf2 is required for basal transcription of
EpRE genes and induces EpRE-promoters in response to high levels of oxidative
stress (283). Accordingly, we observed constitutive localization of Nrf2 at the
EpRE (Figures VI-2A, VI-2B,VI-4A and VI-4B, pages 183 and 187). However, in
the absence of either ERβ or hPMC2, Nrf2 was insufficient by itself to induce
antioxidative genes, or protect against E2-induced ODD (Figure VI-1B and VI-3,
pages 181 and 185), indicating that it does not contribute significantly to TOT-
ERβ mediated regulation of EpRE. A similar observation is true for ERα, even
though ERα is recruited to the EpRE in a TOT-dependent manner. FL-ERβ cells
lacking ERα expression revealed effective TOT-dependent antioxidative enzyme
induction and were protected against E2-induced ODD by TOT (Figure VI-3, page
170
185). This observation is also supported by our previous findings where TOT-ERα did not significantly induce transcription of EpRE-regulated reporter genes (254).
TOT treatment alone was insufficient to prevent MCF10A cells from acquiring tumorigenicity upon exposure to E2 (220). Also, exogenous expression of ERβ, not ERα, inhibited the tumorigenic potential of E2 in the presence of TOT (220).
Analysis of the mechanistic basis of ERβ and hPMC2 functions yielded a
constitutive interaction between ERβ and hPMC2 (Figure VI-1C, page 181). We
observe simultaneous recruitment of both ERβ and hPMC2 to the EpRE when
both of them are expressed, but little or no recruitment of either protein in the
absence of the other (Figures VI-2E and VI-4C, pages 183 and 187). These
results suggest recruitment of an ERβ-hPMC2 complex to the EpRE. ERβ-hPMC2
interaction is constitutive (Figure VI-1C, page 181) while recruitment to the EpRE
is TOT-dependent (Figure VI-2A and VI-4A, pages 183 and 187). An explanation
is that, while the presence or absence of a ligand does not affect the interaction
between ERβ and hPMC2, it could affect the conformation of the ERβ-hPMC2
complex. This could in turn result in differential recruitment of ERβ and hPMC2 to
target sequences in response to different ligands. This explanation seems more
likely when we consider the recruitment of ERβ and hPMC2 to the ERE region of
the pS2 gene. ChIP analysis in MCF10A FL-ERβ cells revealed recruitment of ERβ
and hPMC2 to the ERE under both E2 and TOT treatments (Figure VI-8, page
171
195), in contrast to the predominantly TOT-dependent recruitment observed at
EpRE sequences.
Transcriptional induction at the EpRE by the TOT-ERβ-hPMC2 pathway involves a coactivator complex very similar to that of E2-ERα (Figure VI-2A and
VI-2B, page 183), but independent of ERα recruitment (Figure VI-4A and VI-4B,
page 187). An explanation is that even though both ERα and ERβ are recruited
to the EpRE, only TOT-ERβ recruitment results in transcriptional induction. In
fact, studies on ligand-dependent recruitment to both classical and non classical
ER response genes indicate that the ability of either ERα or ERβ to activate
transcription is not solely dependent on their recruitment to DNA, but also
depend on both the ligand and the promoter context (247, 284-288). TOT-
dependent co-recruitment of ERα and ERβ to the EpRE (Figure VI-2E, page 183),
suggests the possibility of an ERα-ERβ heterodimer. It has been shown that ERα
and ERβ form functional heterodimers and that the heterodimer form
predominates when both isoforms are expressed (289). Alternately, the presence
of ERα could be a consequence of protein-protein interactions with the
coactivators recruited to the EpRE.
We observe increased recruitment of Nrf2 in response to TOT (Figures VI-
2A and VI-4A, pages 183 and 187). This can be explained by considering both in
vitro (213) and in vivo (Figure VI-2E, page 183) data that indicate co-recruitment
172 of Nrf2 with ERβ and hPMC2. Such an interaction can potentially result in more stable binding of Nrf2-EpRE or indirect recruitment of Nrf2 by the ERβ- coactivator complex. Even though Nrf2 is required for transcription of EpRE- regulated genes, TOT treatment neither increases antioxidative enzyme levels nor inhibits E2-induced ODD in the absence of ERβ or hPMC2 (Figure VI-3, page
185). This indicates that the ODD protective effect of TOT is primarily mediated
by ERβ and hPMC2, as Nrf2 alone is insufficient to induce antioxidative gene
expression in response to TOT.
Finally, both PARP-1 and Topoisomerase IIβ are co-recruited to the EpRE
in response to TOT (Figures VI-2A, VI-2E and VI-4A, pages 183 and 187). It has
been proposed that E2-ERα recruitment switches promoter occupancy from
PARP-1 to PARP-1/topoisomerase IIβ complex, resulting in gene activation (290).
Such a scenario would explain the presence of PARP-1 at the ERE region in
response to TOT treatment (Figure VI-8, page 195). However, at the EpRE
region we do not observe PARP-1/topoisomeasre IIβ recruitment in the absence
of either ERβ or hPMC2 (Figure VI-4C, page 187). It is possible that PARP-1 is recruited subsequent to ERβ and hPMC2, but plays a role in recruiting other
coactivators by protein-protein interaction. Such a role for PARP-1 is shown to be
necessary for transcription from RARβ2 promoters (291). Even though PARP-1 is
localized to the promoters of many actively transcribing genes, knockdown of
PARP-1 does not affect the transcription all the genes (292), indicating that mere
173 recruitment does not indicate a functional role. In this regard, our observation that downregulation of PARP-1 attenuates TOT-dependent induction of EpRE promoters coupled with TOT-dependent PARP-1/topoisomerase IIβ recruitment to the EpRE, underscores the contribution of PARP-1 in mediating oxidative stress responses via the TOT-ERβ pathway.
In conclusion, our data provide a mechanistic basis for the protective effect of TOT against E2-induced DNA damage. We show that the protective
effect of TOT is primarily mediated by ERβ and hPMC2. This is significant given
that ERβ expression is increasingly being recognized as a positive prognostic
factor for tamoxifen treatment, especially in ERα negative breast cancers (237,
238). Our studies imply an important role for hPMC2 in mediating
chemoprevention against breast cancer. Not much is known about the biological
role of this protein apart from its ability to increase transcription from EpRE
promoters in response to tamoxifen. The C-terminus of hPMC2 encodes a
putative exonuclease domain (ExoIII), while in vitro data suggest that the
interaction with ERβ is mediated through the N-terminus (unpublished data). We can speculate that transcriptional regulation by ERβ and hPMC2 could potentially
involve double strand DNA breaks induced by the exonuclease activity and
subsequent recruitment of multifunctional proteins such as PARP-1. However,
further studies are required to identify the biological functions of this protein.
Nevertheless, our characterization of the mechanism of transcriptional regulation
174 by antiestrogen-ERβ and identification of key players in this pathway will help us devise new strategies to prevent breast cancer progression.
MATERIALS AND METHODS
Plasmids
The construction of pCMV-Flag-ERβ has been previously described (254).
pSuper-QRshRNA and pSuper-hPMC2shRNA were constructed by annealing to its
complement an oligonucleotide that specifies a 19 nucleotides sequence derived
from the QR and hPMC2 genes, respectively, and separated by a short spacer
from the reverse complement of the same 19 nucleotide sequence.
Oligonucleotides were ligated to the pSUPER vector (293). To make pcDNA-
hPMC2 569miR, the oligos encoding the miRNA sequences were annealed and
cloned into the pcDNA 6.2 GW/EmGFP vector (Invitrogen) according to the
manufacturer’s instruction. The miRNA sequences used to construct the plasmids are listed in Table VI-1 (page 196).
Tissue culture and retroviral transfection
Breast epithelial cells (MCF7 and MCF10A) and PA317 amphotropic
packaging cells were obtained from American Type Culture Collection (Manassas,
VA) and maintained according to their recommended protocols. Retroviruses
were made by transfecting PA317 cells with the pSuper plasmid, pSuper-
hPMC2shRNA or pSuper-QRshRNA. MCF7 cells were infected with retrovirus as
175 previously described (268). For all experiments, breast epithelial cells were depleted of estrogen by growth in Improved Minimal Essential Media minus phenol red containing 5% charcoal-dextran treated calf serum for 5 days prior to ligand treatment. Twenty four hours post infection cells were treated with the indicated ligands for 24 h, and processed either for Western blot analysis or immunostaining.
RT-PCR assays
Cells were washed with PBS and total mRNA extracted using Trizol®
reagent from Invitrogen (Carlsbad, CA) as per the manufacturer’s protocol. Three
micrograms of mRNA was reverse transcribed using the M-MLV Reverse
Transcriptase kit (Invitrogen) following the recommended protocol. One micro
liter of the cDNA was PCR-amplified for varying cycle numbers using primers
listed in Table VI-2 (page 198). The amplified products were run on a 2%
agarose gel and visualized by ethidium bromide staining. Fluorescence was
captured by an eight-bit digital camera, and signal intensities were quantitated using GeneTools software from Syngene (Frederick, MD). Signals in each case
were normalized to their respective GAPDH values to calculate the relative
expression levels.
Generation of stable MCF10A cell lines
176
MCF10A FL-ERβ cell line was generated by transfecting MCF10A human breast epithelial cells with pCMV-Flag-ERβ using FuGene HD Transfection
Reagent (Roche, Indianapolis, IN). Cells were selected on growth media containing 500 μg/ml G418 and resistant colonies were isolated and expanded.
Positive colonies were identified by western blot analysis of FL-ERβ expression.
To generate MCF10A Fl-ERβ/hPMC2mi cell line, MCF10A FL-ERβ cells were transfected with pcDNA-hPMC2 569miR, and stable transfectants cells were selected by growing cells in the presence of 5 μg/ml Blasticidin. The cells were allowed to reach 70% confluency and were then flow sorted for GFP expression.
Knockdown of hPMC2 expression was confirmed by western blot analysis.
8-OHdG immnunostaining and quantitation
Cells grown on coverslips were immunostained for 8-OHdG as described
previously (268). Briefly, methacarn fixed cells were treated with 3% H2O2 to remove peroxidase activity. Cells were blocked with normal goat serum and permeabilized with proteinase K treatment. Cells were immunostained with anti-
8-oxo-dG monoclonal antibody 1F7 (1:100; Trevigen, Gaithersburg, MD).
Immunostaining was developed by the peroxidase-antiperoxidase procedure and
staining intensity (OD) measured as before. The OD of randomly selected fields
of cells was measured and the background OD was subtracted from each. Each
experiment was performed four times, and results were measured under the
177 same optical and light conditions. An electronic shading correction was used to
compensate for any unevenness that might be present in the illumination.
Statistical analysis was performed using two tailed Student’s t test.
Endogenous Immunoprecipitation
Cells were lysed with the IP buffer (50 mM Tris–HCl pH 8, 150 mM NaCl,
0.5% NP-40 and mM EDTA) and sonicated using a Branson 450 sonicator with a
3-mm tapered micro tip at power setting 2 and 70% duty for 3 cycles of 15
seconds each. The supernatant was precleared with protein A sepharose beads
from Pierce (Rockford, IL) for 1 h. The precleared lysates were split into three fractions of equal volume and each fraction was incubated for 3 hours at 40 C with 5 μg of the indicated antibodies preadsorbed to protein A beads. Precleared lysate was used as the input. The beads were washed 4 times with the IP buffer and resuspended in 50 μl of SDS loading buffer (Biorad) for western blot analysis.
Western blotting and quantitation
Whole cell lysates were prepared using mammalian protein extraction
reagent from Pierce. Fifty micro grams of the total protein extract was separated
on a 12% SDS-polyacrylamide gel and electrophoretically transferred onto
nitrocellulose membrane (Pall Corporation, Pansacola, FL). Membranes were
blocked with 5% BSA and probed with the indicated primary antibodies
178 overnight. The membranes were probed with HRP-conjugated anti-rabbit IgG or anti-mouse IgG secondary antibodies (1:5000) for detection with Super Femto reagents (Pierce Chemicals). Signals were visualized by exposure to X-ray films and the chemiluminescence was quantified using Syngene software as before.
The signal intensities in each case were normalized to their respective GAPDH loading controls. Fold change was calculated as the ratio of normalized protein expression in TOT-treated samples to that of the ethanol-treated samples.
Chromatin Immunoprecipitation (ChIP)
Cells were grown in 100-mm dishes and processed for ChIP analyses as
previously described (294). Briefly, cells were fixed with 1% formaldehyde and lysed in SDS-lysis buffer with protease inhibitors. Lysed cells were sonicated using a Branson 450 sonicator with a 3-mm tapered microtip at power setting 3 and 70% duty for 10 pulses/cycle and nine cycles ( 5-W output for 8 to 10
seconds). Clarified, sonicated chromatin was diluted 10-fold in chromatin
immunoprecipitation (ChIP) dilution buffer. One milliliter of the diluted chromatin
was used for overnight immunoprecipitation with a given antibody. The
antibodies used to immunoprecipitate the various proteins described in the
“Results”, are listed in Table VI-3 (page 199).The antibody-chromatin complexes
were pulled down using protein A beads. The beads were subjected to a series of
washes as described and the antigen-DNA complexes eluted. The eluates were
reverse crosslinked overnight at 650 C and the DNA was purified by
179 phenol:chloroform extraction. Ethanol precipitated pellets were resuspended in
50 μl of water and 2 to 4 μl of the suspension was used as template for PCR analysis. The primer sequences used for the PCR analysis are listed in Table VI-2
(page 198). PCR-amplified products were run on a 2% agarose gel and visualized
by ethidium bromide staining. The fluorescence was captured by an eight-bit digital camera and the images optimized using Adobe Photoshop. IP signals were
normalized to their respective inputs. Relative recruitment levels of TOT-treated
samples was calculated as the ratio of normalized IP signals from TOT-treated
cells to that of the vehicle-treated cells.
Liquid Chromatography/Mass Spectrometry (LC-MS) analysis of E2
metabolites
Cells were treated with either 50 nM E2 alone or together with 10 nM TOT for 24 hours. The media was collected and subjected to LC-MS analysis to detect levels of the indicated metabolites as described in (220). The amount of each metabolite was normalized to the total number of cells and the resulting values were used to calculate fold change. Fold change in metabolite levels for each cell line was calculated as the ratio of the normalized value from ligand-treated sample to that of its corresponding ethanol-treated cell line.
Transient siRNA transfection
180
Cells were maintained on stripped serum media two days prior to transfection. Trypsinized cells were then independently transfected with four different double stranded siRNA oligos from Qiagen (Valencia, CA) targeted along the PARP-1 mRNA. The transfections were done using siPORT Amine Reagent from Ambion (Austin, TX) as per the reverse transfection protocol provided by the manufacturer. After 72 hours, both transfected and untransfected cells were treated with 0.01% ethanol (control) or 10 nM TOT for 3 hours. Cells were harvested in PBS and whole cell lysates were analyzed by western blotting to determine protein expression levels.
181
Figure VI-1: hPMC2 interacts directly with ERβ and is involved in mediating Tamoxifen-dependent decrease in ODD levels. (A) Western blotting analysis of QR and hPMC2 expression levels to confirm transient knockdown in MCF7 cells. Cytokeratin 18 and GAPDH were used as loading controls, Con (control shRNA). (B) MCF7 cells transiently infected with retroviral particles to knockdown QR and hPMC2 expression were treated for 24 hours with
0.01% ethanol, 50 nM E2 or 100 nM TOT or a combination of E2 and TOT as indicated. Cells were immunostained with an antibody against 8-OHdG and the intensity of the staining quantitated in each case. The bars represent the average of 3 independent experiments and the standard error is given by the error bars. * represents p<0.001 vs the E2-treated samples. (C) Serum starved
MCF7 cells were treated with either 0.01% ethanol (con), 10 nM E2 or 10 nM
TOT for 3 hours. Lysates were immunoprecipitated using antibodies against ERβ or hPMC2 and analyzed for co-immunoprecipitating proteins by Western blotting.
Normal rabbit immunoglobilin was used as a specificity control. Input lanes represent 10% of the total protein (Published in Sripathy SP et al. Oncogene,
2008).
182
183
Figure VI-2: Tamoxifen-dependent recruitment of coactivators to the
EpRE sequence of NQO1 (A) Representation of NQO1 gene locus with the
thick lines representing regions analyzed in subsequent PCR experiments. (B)
Serum starved MCF7 cells were treated with either 0.01% ethanol (Con), 10 nM
E2 or 10 nM TOT for 3 hours and processed for ChIP analysis.
Immunoprecipitated DNA was PCR analyzed to determine recruitment patterns of
the indicated proteins. Input (Inp) represents 2% of total DNA. (C) Average
recruitment levels of the indicated factors in TOT-treated samples relative to that
of the ethanol-treated samples. Error bars represent standard errors of 3 or more
independent experiments. (D) ChIP analysis of TOT-treated cells for recruitment
of the indicated factors using primers upstream and downstream, of the EpRE
region (E) MCF7 cells were treated with 10 nM TOT for 3 hours and processed
lysates were subjected to ChIP using an antibody against hPMC2 (rabbit IgG was
used as a specificity control). hPMC2-precipitated chromatin was diluted 1:20
and re-immunoprecipitated using the indicated antibodies. The precipitates were
then used to isolate DNA and subjected to PCR analysis at the EpRE locus. The
results in each case are representative two or more independent experiments
(Published in Sripathy SP et al. Oncogene, 2008).
184
185
Figure VI-3: Both ERβ and hPMC2 are required for tamoxifen-mediated increase in antioxidative enzyme expression and protection against
ODD. (A) Western blot analysis of MCF7, MCF10A parental cells (10A(P)),
MCF10A stably expressing FLAG-ERβ (FL-ERβ) and MCF10A FL-ERβ positive cells that express stably knocked down levels of hPMC2 (hPMC2mi). An anti-FLAG antibody was used to detect FL-ERβ and GAPDH is used as a loading control. (B)
Cells belonging to the indicated cell lines were treated with either 0.01% ethanol or 10 nM TOT for 3 hours. Lysates were analyzed for expression levels of GCSh,
QR and GSTpi by Western blotting. The signals were normalized to their respective GAPDH loading controls and the fold change relative to the ethanol- treated group was calculated as described in the “Materials and Methods”. The bars indicate the average of 3 independent experiments and the standard error is given by the error bars. (C and D) Cells from the indicated cell lines were treated with either 0.01% ethanol (Con), 50 nM E2 or, a combination of 50 nM E2
and 10 nM TOT for 24 hours and (C) immunostained for 8-OHdG as before. The
bars represent average of 4 independent experiments and the error bars
represent the standard error. (D) The culture media was subjected to LC-MS
analysis to detect the indicated metabolite levels. The bars represent the average
fold change in metabolite levels relative to the ethanol-treated group and error
bars represent standard error from 3 independent experiments (Published in
Sripathy SP et al. Oncogene, 2008).
186
187
Figure VI-4: Tamoxifen-dependent recruitment of coactivators to the
EpRE sequence of NQO1 requires both ERβ and hPMC2. (A) The indicated cell lines were treated with either 0.01% ethanol (Con), 10 nM E2, or TOT for 3
hours. Cells were processed for ChIP analysis and immunoprecipitated DNA was
analyzed by PCR to determine the recruitment of the indicated proteins at the
EpRE region. (B) Average recruitment levels of the indicated factors in TOT-
treated samples relative to that of the ethanol-treated cells. Error bars represent
standard error of two or more independent experiments. (C) ChIP analysis of
TOT treated-samples for recruitment of the indicated factors in the absence of
ERα, ERβ and hPMC2 (Published in Sripathy SP et al. Oncogene, 2008).
188
189
Figure VI-5: Tamoxifen treatment induces increased transcription of antioxidative enzyme levels. MCF7 cells were treated with either 0.01% ethanol (con) or TOT for 3 hours. (A, B and C) The change in mRNA levels of
GCSH, NQO1 and GSTP-1 were determined using semi quantitative RT-PCR analysis. The quantified PCR signals in each case were normalized to their respective GAPDH controls. Error bars indicate standard error from 2 independent experiments. (D) Cell lysates from MCF7 cells treated with 10 nM E2 or TOT for 3 hours, were subjected to Western blot analysis using the indicated antibodies. The quantified signals were normalized to their respective GAPDH controls and used to calculate the fold change in expression as described. Error bars indicate standard error of 3 independent experiments (Published in Sripathy
SP et al. Oncogene, 2008).
190
191
Figure VI-6: Knockdown of PARP-1 attenuates tamoxifen-dependent increase in the expression of antioxidative enzymes. FL-ER β cells were treated for 3 hours with 0.01% ethanol (C) or 10 nM TOT (T) after 72 hours post transfection. (A) Western blot analysis of knockdown in PARP-1 expression. si 1, si 2, si 3 and si 4 are siRNA oligos targeted against different regions of PARP-1 mRNA. Numbers at the bottom represent the % knockdown in siRNA transfected cells as compared to their corresponding lanes in the control-transfected samples. GAPDH was used as a loading control. (B) Expression levels of GCSh,
QR and GAPDH were analyzed and normalized to that of GAPDH in each case.
The bars represent fold change in expression levels of the indicated enzymes as compared to their corresponding ethanol-treated samples. Error bars indicate standard deviation of 4 independent experiments (Published in Sripathy SP et al.
Oncogene, 2008).
192
193
Figure VI-7: PARP-1 down regulation does not significantly affect basal expression levels of antioxidative enzymes. MCF10A FL-ERβ (control and
PARP-1 siRNA transfected) cells were treated with 0.01% ethanol for 3 hours after 24 hours post transfection. Whole cell lysates were analyzed for expression levels of QR, GCSh and GSTpi by western blotting. The data represent average of four independent experiments and error bars represent standard deviation
(Published in Sripathy SP et al. Oncogene, 2008).
194
195
Figure VI-8:. Ligand dependent recruitment of ERβ and hPMC2 to the
ERE region. MCF10A FL-ERβ cells were treated with either 0.01% ethanol
(Con), 10 nM E2, or TOT for 3 hours, and were processed for ChIP analysis.
Immunoprecipitated DNA was analyzed by PCR to determine the recruitment of
the indicated proteins at the ERE region of the pS2 gene (Published in Sripathy
SP et al. Oncogene, 2008).
196
197
Table VI-1: Sequence of DNA oligos cloned into pSuper-QRshRNA, pSuper- hPMC2shRNA and PCDNA-hPMC2 569miR plasmids respectively (Published in
Sripathy SP et al. Oncogene, 2008).
198
Table VI-2: Primer sequences used in the various PCR reactions described in Chapter VI. (Published in Sripathy SP et al. Oncogene, 2008).
199
Table VI-3: Antibodies used in Chapter VI. (Published in Sripathy SP et al. Oncogene, 2008).
200
CHAPTER VII
SUMMARY AND FUTURE DIRECTIONS
SUMMARY
Investigation of the molecular mechanism responsible for the ERβ- mediated protective effects revealed a constitutive interaction of ER with a novel protein hPMC2 (Chapter VI). Using a combination of breast epithelial cell lines that are either positive or negative for ERα, we demonstrate TOT-dependent recruitment of both ERβ and hPMC2 to the EpRE-regulated antioxidative gene
NQO1 that codes for the enzyme QR. We further demonstrate TOT-dependent co-recruitment of the coactivators Nrf2, PARP-1 and topoisomerase IIβ, both in the presence and absence of ERα. However, the absence of either ERβ or hPMC2 results in non-recruitment of PARP-1 and topoisomerase IIβ, loss of antioxidative enzyme induction and attenuated protection against ODD by TOT even in the presence of Nrf2 and ERα. These findings indicate a minor role for Nrf2 and ERα in TOT-dependent antioxidative gene regulation. However, down regulation of
PARP-1 attenuates TOT-dependent antioxidative gene induction. We conclude that ERβ and hPMC2 are required for TOT-dependent recruitment of coactivators such as PARP-1 to the EpRE resulting in the induction of antioxidative enzymes and subsequent protection against ODD (Figure VII-1, page 202).
201
Figure VII-1: Induction of the EpRE-regulated genes, QR, GCSh and GSTpi by tamoxifen (TOT) is dependent on the recruitment of a coactivator complex by
ERβ and hPMC2. Increased expression of antioxidative genes correlates to decreased levels of catechol estrogens (CE) and inhibition of E2-induced oxidative
DNA damage (ODD) by tamoxifen liganded ERβ.
202
FUTURE DIRECTIONS
Our studies indicate that tamoxifen treatment can successfully inhibit E2- induced ODD in the presence of ERβ and hPMC2. We have previously reported that treatment with tamoxifen inhibits the transformation of normal breast epithelial cells that have been exposed to long term estrogen treatment (220).
Also, ACI rats treated with tamoxifen exhibit decreased E2-induced ODD levels
and increased expression of QR in their mammary epithelial cells (218). These observations strongly suggest the potential for tamoxifen to be used as a
chemopreventive drug to inhibit E2-induced mammary carcinogenesis. An ideal way to test this hypothesis would be to analyze the effect of tamoxifen in mice or rat models of E2-induced breast cancer.
Aromatase transgenic mice and E2-dependent carcinogenesis
The most commonly used mouse model to study E2-dependent
carcinogenesis is the aromatase transgenic mice model. E2 levels in these mice
exhibit a two fold increase compared to their wild type littermates. They exhibit
mammary hyperplasia and preneoplastic lesions. Treatment of these mice with
suboptimal concentrations of the carcinogen DMBA results in increased tumor
incidence compared to their wild type litter mates, indicating that E2 exposure
renders the mammary epithelial cells more sensitive to the carcinogenic effects
203 of other chemicals (295, 296). We have developed a mouse strain that overexpresses both aromatase and QR (Montano et al. unpublished). Preliminary data indicate that aromatase transgenic mice display high levels of oxidative DNA damage in their breast tissue compared to their wild type litter mates, however
QR overexpression results in lower levels of E2-induced ODD which also
correlates with a decreased incidence of mammary hyperplasia. These
observations indicate the important role played by antioxidative enzymes in the
inhibition of E2-induced carcinogenesis.
Our studies demonstrate that increased expression of QR results in a
concomitant decrease in carcinogenic E2 metabolites (Chapter VI). These data
together with our preliminary observations from the aromatase/QR double
transgenic mice make it reasonable to suppose that tamoxifen treatment of
aromatase transgenic mice would lead to decreased ODD levels and decreased
incidence of hyperplasia. As mentioned earlier, these mice are sensitive to
tumorigenesis induced by external carcinogens. Hence it would be worthwhile to
study the effect of DMBA treatment on tumor incidence in the presence and
absence of tamoxifen treatment in these mice.
ACI female rat model of E2-induced breast cancer and chemoprevention by tamoxifen
204
Another attractive animal model is the female ACI rat, which spontaneously develops mammary tumors in response to long term exposure to estrogen making it a model of choice to study E2-dependent carcinogenesis
(297). It has been shown that tamoxifen treatment significantly reduces E2- induced carcinogenesis in these rats. Our observations in ACI rats clearly indicate that increased QR expression is coupled to decreased levels of E2- induced ODD (220) suggesting that ERβ plays a key role in mediating the effects of tamoxifen. However, these rats express both ERα and ERβ and it is not clear whether the anticancer ability of tamoxifen is primarily mediated through blocking of ERα signaling or via ERβ-mediated increase in the transcription of
EpRE genes involved in detoxification of E2 metabolites. While it would be ideal
to generate ERβKO and ERαKO rat models and study the effect of E2-induced carcinogenesis against these backgrounds, to date it is not straightforward to produce gene knockouts through homologous recombination in rats.
An alternative strategy is to generate conditional ERα knockdown rats is to reduce ERα expression in rat embryonic cells. Heritable down regulation of target genes in rats via introduction of lentiviral particles expressing shRNA has been reported (298). A similar strategy can be adopted to conditionally knock down
ERα expression in rat under a mammary specific promoter such as the rat β- casein promoter which is activated during pregnancy and lactation. Such a rat model would allow us to more conclusively demonstrate the ability of tamoxifen
205 and ERβ to inhibit E2-induced carcinogenesis independent of the ERα antagonistic
activity of tamoxifen.
Selective activation of ERβ for breast cancer chemoprevention
Recent studies analyzing the role of ERβ in E2-dependent breast cancers
indicate that ERβ expression is a positive prognostic factor in breast cancer. This
was found to be especially true in the absence of ERα expression. The recently
concluded large scale clinical trial, STAR (Study of Tamoxifen and Roloxefene)
assessing the chemopreventive ability of tamoxifen and raloxifene indicated that
a subset of women diagnosed with ERα negative breast cancer respond
successfully to tamoxifen therapy. However, the status of ERβ expression in
these women was not known. We observe ERβ dependent induction of
antioxidative genes by the antiestrogen tamoxifen which results in inhibition of
E2-induced ODD (Chapter VI). Also, tamoxifen treatment can inhibit the E2- induced transformation of normal breast epithelial cells only in the presence of
ERβ (220). These observations suggest that ERβ may play a role in mediating the anticancerous property of tamoxifen in the absence of ERα. This ability of
ERβ may be partly linked to the transcriptional induction of antioxidative genes by tamoxifen liganded ERβ. The observation that a classical antagonist of estrogen receptors in the breast can also act as an agonist depending on the receptor subtype and promoter context outlines the usefulness of developing ligands that are ER subtype selective.
206
An ideal ligand would be an ERα and ERβ antagonist with respect to inducing transcription of E2-responsive genes such as cyclin D1 and c-Myc, while
being a selective ERβ agonist for induction of EpRE-regulated genes. These conditions are satisfied by tamoxifen which we have used in our study, hence the need is to generate SERMs that are more potent inducers of ERβ-EpRE agonistic
activity while still being antagonistic to ERβ-ERE regulated transcription. In fact
many ER subtype selective ligands already exist, for example, the synthetic
compounds DPN (2,3-bis(4-hyroxyphenyl)-propionitrile) and 8βVE2 (8-vinylestra-
1,3,5(10)-triene-3,17β-diol) preferentially activate ERβ (299). Phytoestrogens
that activate ERβ include liquiritigenin which induces E2-responsive genes via ERβ
(300). The most promising candidate is resveratrol which acts as an ERβ agonist
for induction of EpRE-genes while it does not activate ERE-regulated genes
either via ERα or ERβ (301, 302).
These observations clearly suggest that the ability of any ligand to act as
either an agonist or an antagonist is promoter context dependent. Hence a
comprehensive drug screen would be to adopt an unbiased approach and screen
a library of compounds for their ability to induce EpRE-mediated expression. This
can be easily accomplished by utilizing breast epithelial cell lines that stably
express either ERα or ERβ and contain a convenient reporter such as GFP under
the regulation of either an EpRE or an ERE sequence. Promising candidates
207 would be those drugs that induce GFP expression from cells expressing EpRE- regulated GFP but not ERE-GFP. These lead compounds should be further validated by analyzing their effects on endogenous expression of both ERE- and
EpRE-genes.
Role of hPMC2 in ERβ-mediated transcriptional regulation
Our results demonstrate the critical role played by hPMC2 in the induction of EpRE-genes by tamoxifen-ERβ. We observed tamoxifen dependent recruitment of hPMC2 to the EpRE region. However the precise role of hPMC2 in mediating the transcriptional induction by ERβ is not clear. The sequence of hPMC2 encodes a putative exonuclease domain, ExoIII. Proteins that possess the
ExoIII domain include DNA and RNA exonucleases capable of 3’- 5’ exonuclease activity. Hence we can speculate that hPMC2 facilitates transcriptional induction by inducing DNA breaks. Recent reports from independent labs demonstrate the primary role played by transient DNA breaks in ligand-induced transactivation of
ERE-genes by ERα (281). Accumulating evidence points to the involvement of
DNA repair components such as PARP1 in regulating transcriptional processes
(303). In this regard it is significant that we observed tamoxifen-dependent
PARP1 recruitment to the EpRE regions (Chapter VI). Based on these observations we can hypothesize that hPMC2 is a functional exonuclease that facilitates ERβ mediated transcription of EpRE-genes.
208
A first step towards testing this hypothesis would be to analyze purified hPMC2 for exonuclease activity in vitro. We have already purified full-length hPMC2 protein and will be purifying versions of these proteins that harbour
mutations in the conserved residues of the ExoIII domain. Both the wild type
and mutants can be tested for their ability to act as exonucleases by incubating
them with a synthetic stretch of radiolabeled nucleotides and analyzing the
degradation of the nucleotide substrate.
In order to detect induction of DNA breaks in vivo, it is necessary to first
establish the kinetics of hPMC2 recruitment to the EpRE following tamoxifen
treatment. Once this has been established, cells treated with tamoxifen can be
fixed using a tissue fixative to prevent further DNA damage. Fixed cells can be
permeabilized and incorporated with biotin-11-dUTP to detect DNA breaks. The
labeled cells can then be cross linked with formaldehyde and subjected to ChIP
analysis and subsequent PCR of the EpRE region to detect site-specific biotin
incorporation. An ideal negative control is provided by cells in which hPMC2 has
been knockeddown and hence do not exhibit tamoxifen-dependent induction of
QR.
Finally it is possible that even though hPMC2 is an active exonuclease, this
property might not play a major role in ERβ-mediated transcriptional induction.
To test such a possibility, we are creating cell lines that are stably down
209 regulated for hPMC2 expression by targeting the 3’UTR region of the hPMC2 mRNA with shRNA. We have already shown that down regulation of hPMC2 results in an inability of tamoxifen to induce QR, GSTpi or GCSh (Chapter VI).
Complementation of the hPMC2 down regulated cells with wild type hPMC2 would rescue this effect, while complementation with ExoIII mutant versions would rescue this effect only if the exonuclease activity is not a major contributor to ERβ induced transcription.
Our studies also demonstrate an endogenous and constitutive interaction between ERβ and hPMC2. Also, neither of the proteins are recruited to the EpRE in the absence of the other suggesting that ERβ and hPMC2 may act as a complex (Chapter VI). In vitro studies indicate that hPMC2 interacts with ERβ via
its N –terminus (Montano et al. unpublished). The importance of ERβ and hPMC2
interaction can be analyzed by complementation of hPMC2 knock down cells with
various N-terminal mutants that disrupt direct interaction with ERβ.
210
BIBILIOGRAPHY
1. Lander ES, Linton LM, Birren B, et al. Initial sequencing and analysis of the human genome. Nature 2001;409(6822):860-921.
2. Baltimore D. Our genome unveiled. Nature 2001;409(6822):814-6.
3. Ruvkun G, Hobert O. The taxonomy of developmental control in
Caenorhabditis elegans. Science 1998;282(5396):2033-41.
4. Adams MD, Celniker SE, Holt RA, et al. The genome sequence of
Drosophila melanogaster. Science 2000;287(5461):2185-95.
5. Levine M, Tjian R. Transcription regulation and animal diversity. Nature
2003;424(6945):147-51.
6. Wyrick JJ, Young RA. Deciphering gene expression regulatory networks.
Curr Opin Genet Dev 2002;12(2):130-6.
7. Chen K, Rajewsky N. The evolution of gene regulation by transcription factors and microRNAs. Nat Rev Genet 2007;8(2):93-103.
8. Lewis EB. A gene complex controlling segmentation in Drosophila. Nature
1978;276(5688):565-70.
9. Lemons D, McGinnis W. Genomic evolution of Hox gene clusters. Science
2006;313(5795):1918-22.
10. Ariniello L, kibiyuk l. Hox genes and birth defects. Society for
Neuroscience, brain briefings 1994.
211
11. Lemon B, Tjian R. Orchestrated response: a symphony of transcription factors for gene control. Genes Dev 2000;14(20):2551-69.
12. Rosenfeld MG, Lunyak VV, Glass CK. Sensors and signals: a coactivator/corepressor/epigenetic code for integrating signal-dependent programs of transcriptional response. Genes Dev 2006;20(11):1405-28.
13. Novac N, Heinzel T. Nuclear receptors: overview and classification. Curr
Drug Targets Inflamm Allergy 2004;3(4):335-46.
14. Fenrick R, Hiebert SW. Role of histone deacetylases in acute leukemia. J
Cell Biochem Suppl 1998;30-31:194-202.
15. Grignani F, De Matteis S, Nervi C, et al. Fusion proteins of the retinoic acid
receptor-alpha recruit histone deacetylase in promyelocytic leukaemia. Nature
1998;391(6669):815-8.
16. He LZ, Guidez F, Tribioli C, et al. Distinct interactions of PML-RARalpha and PLZF-RARalpha with co-repressors determine differential responses to RA in
APL. Nat Genet 1998;18(2):126-35.
17. Klinge CM. Estrogen receptor interaction with estrogen response elements. Nucleic Acids Res 2001;29(14):2905-19.
18. O'Lone R, Frith MC, Karlsson EK, Hansen U. Genomic targets of nuclear estrogen receptors. Mol Endocrinol 2004;18(8):1859-75.
19. Frasor J, Danes JM, Komm B, Chang KC, Lyttle CR, Katzenellenbogen BS.
Profiling of estrogen up- and down-regulated gene expression in human breast
212 cancer cells: insights into gene networks and pathways underlying estrogenic control of proliferation and cell phenotype. Endocrinology 2003;144(10):4562-74.
20. Nishidate T, Katagiri T, Lin ML, et al. Genome-wide gene-expression profiles of breast-cancer cells purified with laser microbeam microdissection: identification of genes associated with progression and metastasis. Int J Oncol
2004;25(4):797-819.
21. Frankel AD, Berg JM, Pabo CO. Metal-dependent folding of a single zinc finger from transcription factor IIIA. Proc Natl Acad Sci U S A 1987;84(14):4841-
5.
22. Brown RS, Sander C, Argos P. The primary structure of transcription factor
TFIIIA has 12 consecutive repeats. FEBS Lett 1985;186(2):271-4.
23. Pavletich NP, Pabo CO. Zinc finger-DNA recognition: crystal structure of a
Zif268-DNA complex at 2.1 A. Science (New York, NY 1991;252(5007):809-17.
24. Lee MS, Gippert GP, Soman KV, Case DA, Wright PE. Three-dimensional solution structure of a single zinc finger DNA-binding domain. Science
1989;245(4918):635-7.
25. Schuh R, Aicher W, Gaul U, et al. A conserved family of nuclear proteins
containing structural elements of the finger protein encoded by Kruppel, a
Drosophila segmentation gene. Cell 1986;47(6):1025-32.
26. Miller IJ, Bieker JJ. A novel, erythroid cell-specific murine transcription factor that binds to the CACCC element and is related to the Kruppel family of nuclear proteins. Mol Cell Biol 1993;13(5):2776-86.
213
27. Iuchi S. Three classes of C2H2 zinc finger proteins. Cell Mol Life Sci
2001;58(4):625-35.
28. Elrod-Erickson M, Benson TE, Pabo CO. High-resolution structures of variant Zif268-DNA complexes: implications for understanding zinc finger-DNA recognition. Structure 1998;6(4):451-64.
29. Laity JH, Dyson HJ, Wright PE. DNA-induced alpha-helix capping in conserved linker sequences is a determinant of binding affinity in Cys(2)-His(2) zinc fingers. J Mol Biol 2000;295(4):719-27.
30. Laity JH, Dyson HJ, Wright PE. Molecular basis for modulation of biological function by alternate splicing of the Wilms' tumor suppressor protein. Proc Natl
Acad Sci U S A 2000;97(22):11932-5.
31. Barbaux S, Niaudet P, Gubler MC, et al. Donor splice-site mutations in
WT1 are responsible for Frasier syndrome. Nat Genet 1997;17(4):467-70.
32. Bowers PM, Schaufler LE, Klevit RE. A folding transition and novel zinc
finger accessory domain in the transcription factor ADR1. Nat Struct Biol
1999;6(5):478-85.
33. Wuttke DS, Foster MP, Case DA, Gottesfeld JM, Wright PE. Solution
structure of the first three zinc fingers of TFIIIA bound to the cognate DNA
sequence: determinants of affinity and sequence specificity. J Mol Biol
1997;273(1):183-206.
214
34. Wolfe SA, Grant RA, Elrod-Erickson M, Pabo CO. Beyond the "recognition code": structures of two Cys2His2 zinc finger/TATA box complexes. Structure
2001;9(8):717-23.
35. Bellefroid EJ, Poncelet DA, Lecocq PJ, Revelant O, Martial JA. The evolutionarily conserved Kruppel-associated box domain defines a subfamily of eukaryotic multifingered proteins. Proc Natl Acad Sci U S A 1991;88(9):3608-12.
36. Margolin JF, Friedman JR, Meyer WK, Vissing H, Thiesen HJ, Rauscher FJ,
3rd. Kruppel-associated boxes are potent transcriptional repression domains.
Proc Natl Acad Sci U S A 1994;91(10):4509-13.
37. Pengue G, Calabro V, Bartoli PC, Pagliuca A, Lania L. Repression of transcriptional activity at a distance by the evolutionarily conserved KRAB domain present in a subfamily of zinc finger proteins. Nucleic Acids Res
1994;22(15):2908-14.
38. Witzgall R, O'Leary E, Leaf A, Onaldi D, Bonventre JV. The Kruppel- associated box-A (KRAB-A) domain of zinc finger proteins mediates transcriptional repression. Proc Natl Acad Sci U S A 1994;91(10):4514-8.
39. Senatore B, Cafieri A, Di Marino I, Rosati M, Di Nocera PP, Grimaldi G. A variety of RNA polymerases II and III-dependent promoter classes is repressed by factors containing the Kruppel-associated/finger preceding box of zinc finger proteins. Gene 1999;234(2):381-94.
40. Moosmann P, Georgiev O, Thiesen HJ, Hagmann M, Schaffner W.
Silencing of RNA polymerases II and III-dependent transcription by the KRAB
215 protein domain of KOX1, a Kruppel-type zinc finger factor. Biol Chem
1997;378(7):669-77.
41. Pengue G, Lania L. Kruppel-associated box-mediated repression of RNA polymerase II promoters is influenced by the arrangement of basal promoter elements. Proc Natl Acad Sci U S A 1996;93(3):1015-20.
42. Looman C, Hellman L, Abrink M. A novel Kruppel-Associated Box identified in a panel of mammalian zinc finger proteins. Mamm Genome 2004;15(1):35-40.
43. Mark C, Abrink M, Hellman L. Comparative analysis of KRAB zinc finger proteins in rodents and man: evidence for several evolutionarily distinct subfamilies of KRAB zinc finger genes. DNA Cell Biol 1999;18(5):381-96.
44. Kim S CY. A novel member of the RING finger family, KRIP-1, associates with the KRAB-A transcriptional repressor domain of zinc finger proteins. Proc
Natl Acad Sci USA 1996 Vol. 93:15299-304.
45. Friedman JR, Fredericks WJ, Jensen DE, et al. KAP-1, a novel corepressor
for the highly conserved KRAB repression domain. Genes Dev 1996;10(16):2067-
78.
46. Moosmann P, Georgiev O, Le Douarin B, Bourquin JP, Schaffner W.
Transcriptional repression by RING finger protein TIF1 beta that interacts with
the KRAB repressor domain of KOX1. Nucleic Acids Res 1996;24(24):4859-67.
47. Peng H, Begg GE, Schultz DC, et al. Reconstitution of the KRAB-KAP-1
repressor complex: a model system for defining the molecular anatomy of RING-
216
B box-coiled-coil domain-mediated protein-protein interactions. J Mol Biol
2000;295(5):1139-62.
48. Peng H, Begg GE, Harper SL, Friedman JR, Speicher DW, Rauscher FJ,
3rd. Biochemical analysis of the Kruppel-associated box (KRAB) transcriptional repression domain. J Biol Chem 2000;275(24):18000-10.
49. Peng H, Gibson LC, Capili AD, et al. The structurally disordered KRAB
repression domain is incorporated into a protease resistant core upon binding to
KAP-1-RBCC domain. J Mol Biol 2007;370(2):269-89.
50. Mannini R, Rivieccio V, D'Auria S, et al. Structure/function of KRAB repression domains: structural properties of KRAB modules inferred from hydrodynamic, circular dichroism, and FTIR spectroscopic analyses. Proteins
2006;62(3):604-16.
51. Shannon M, Hamilton AT, Gordon L, Branscomb E, Stubbs L. Differential expansion of zinc-finger transcription factor loci in homologous human and mouse gene clusters. Genome Res 2003;13(6A):1097-110.
52. Looman C, Abrink M, Mark C, Hellman L. KRAB zinc finger proteins: an analysis of the molecular mechanisms governing their increase in numbers and complexity during evolution. Mol Biol Evol 2002;19(12):2118-30.
53. Huntley S, Baggott DM, Hamilton AT, et al. A comprehensive catalog of
human KRAB-associated zinc finger genes: insights into the evolutionary history
of a large family of transcriptional repressors. Genome Res 2006;16(5):669-77.
217
54. Krebs CJ, Larkins LK, Price R, Tullis KM, Miller RD, Robins DM. Regulator of sex-limitation (Rsl) encodes a pair of KRAB zinc-finger genes that control sexually dimorphic liver gene expression. Genes Dev 2003;17(21):2664-74.
55. Han ZG, Zhang QH, Ye M, et al. Molecular cloning of six novel Kruppel-like
zinc finger genes from hematopoietic cells and identification of a novel
transregulatory domain KRNB. J Biol Chem 1999;274(50):35741-8.
56. Abrink M, Aveskogh M, Hellman L. Isolation of cDNA clones for 42
different Kruppel-related zinc finger proteins expressed in the human monoblast
cell line U-937. DNA Cell Biol 1995;14(2):125-36.
57. Bellefroid EJ, Marine JC, Ried T, et al. Clustered organization of
homologous KRAB zinc-finger genes with enhanced expression in human T
lymphoid cells. Embo J 1993;12(4):1363-74.
58. Hering TM, Kazmi NH, Huynh TD, et al. Characterization and chondrocyte
differentiation stage-specific expression of KRAB zinc-finger protein gene
ZNF470. Exp Cell Res 2004;299(1):137-47.
59. Alonso MB, Zoidl G, Taveggia C, et al. Identification and characterization
of ZFP-57, a novel zinc finger transcription factor in the mammalian peripheral
nervous system. J Biol Chem 2004;279(24):25653-64.
60. Tan W, Zheng L, Lee WH, Boyer TG. Functional dissection of transcription
factor ZBRK1 reveals zinc fingers with dual roles in DNA-binding and BRCA1-
dependent transcriptional repression. J Biol Chem 2004;279(8):6576-87.
218
61. Yun J, Lee WH. Degradation of transcription repressor ZBRK1 through the ubiquitin-proteasome pathway relieves repression of Gadd45a upon DNA damage. Mol Cell Biol 2003;23(20):7305-14.
62. Zheng L, Pan H, Li S, et al. Sequence-specific transcriptional corepressor
function for BRCA1 through a novel zinc finger protein, ZBRK1. Mol Cell
2000;6(4):757-68.
63. Le Douarin B, Zechel C, Garnier JM, et al. The N-terminal part of TIF1, a putative mediator of the ligand-dependent activation function (AF-2) of nuclear receptors, is fused to B-raf in the oncogenic protein T18. EMBO J
1995;14(9):2020-33.
64. Le Douarin B, Nielsen AL, Garnier JM, et al. A possible involvement of TIF1
alpha and TIF1 beta in the epigenetic control of transcription by nuclear
receptors. Embo J 1996;15(23):6701-15.
65. Venturini L, You J, Stadler M, et al. TIF1gamma, a novel member of the
transcriptional intermediary factor 1 family. Oncogene 1999;18(5):1209-17.
66. Khetchoumian K, Teletin M, Mark M, et al. TIF1delta, a novel HP1-
interacting member of the transcriptional intermediary factor 1 (TIF1) family
expressed by elongating spermatids. J Biol Chem 2004;279(46):48329-41.
67. Freemont PS. The RING finger. A novel protein sequence motif related to
the zinc finger. Ann N Y Acad Sci 1993;684:174-92.
219
68. Barlow PN, Luisi B, Milner A, Elliott M, Everett R. Structure of the C3HC4 domain by 1H-nuclear magnetic resonance spectroscopy. A new structural class of zinc-finger. J Mol Biol 1994;237(2):201-11.
69. Borden KL, Boddy MN, Lally J, et al. The solution structure of the RING
finger domain from the acute promyelocytic leukaemia proto-oncoprotein PML.
EMBO J 1995;14(7):1532-41.
70. Joazeiro CA, Weissman AM. RING finger proteins: mediators of ubiquitin ligase activity. Cell 2000;102(5):549-52.
71. Dupont S, Zacchigna L, Cordenonsi M, et al. Germ-layer specification and
control of cell growth by Ectodermin, a Smad4 ubiquitin ligase. Cell
2005;121(1):87-99.
72. Meroni G, Diez-Roux G. TRIM/RBCC, a novel class of 'single protein RING
finger' E3 ubiquitin ligases. Bioessays 2005;27(11):1147-57.
73. Reymond A, Meroni G, Fantozzi A, et al. The tripartite motif family
identifies cell compartments. EMBO J 2001;20(9):2140-51.
74. Peng H, Feldman I, Rauscher FJ, 3rd. Hetero-oligomerization among the
TIF family of RBCC/TRIM domain-containing nuclear cofactors: a potential mechanism for regulating the switch between coactivation and corepression. J
Mol Biol 2002;320(3):629-44.
75. Capili AD, Schultz DC, Rauscher IF, Borden KL. Solution structure of the
PHD domain from the KAP-1 corepressor: structural determinants for PHD, RING and LIM zinc-binding domains. Embo J 2001;20(1-2):165-77.
220
76. Dhalluin C, Carlson JE, Zeng L, He C, Aggarwal AK, Zhou MM. Structure and ligand of a histone acetyltransferase bromodomain. Nature
1999;399(6735):491-6.
77. Jacobson RH, Ladurner AG, King DS, Tjian R. Structure and function of a human TAFII250 double bromodomain module. Science 2000;288(5470):1422-5.
78. Owen DJ, Vallis Y, Pearse BM, McMahon HT, Evans PR. The structure and function of the beta 2-adaptin appendage domain. EMBO J 2000;19(16):4216-
27.
79. Mujtaba S, Zeng L, Zhou MM. Structure and acetyl-lysine recognition of the bromodomain. Oncogene 2007;26(37):5521-7.
80. Ryan RF, Schultz DC, Ayyanathan K, et al. KAP-1 corepressor protein
interacts and colocalizes with heterochromatic and euchromatic HP1 proteins: a
potential role for Kruppel-associated box-zinc finger proteins in heterochromatin-
mediated gene silencing. Mol Cell Biol 1999;19(6):4366-78.
81. Nielsen AL, Ortiz JA, You J, et al. Interaction with members of the
heterochromatin protein 1 (HP1) family and histone deacetylation are
differentially involved in transcriptional silencing by members of the TIF1 family.
EMBO J 1999;18(22):6385-95.
82. Lechner MS, Begg GE, Speicher DW, Rauscher FJ, 3rd. Molecular
determinants for targeting heterochromatin protein 1-mediated gene silencing:
direct chromoshadow domain-KAP-1 corepressor interaction is essential. Mol Cell
Biol 2000;20(17):6449-65.
221
83. Cammas F, Mark M, Dolle P, Dierich A, Chambon P, Losson R. Mice lacking the transcriptional corepressor TIF1beta are defective in early postimplantation development. Development 2000;127(13):2955-63.
84. Cammas F, Oulad-Abdelghani M, Vonesch JL, Huss-Garcia Y, Chambon P,
Losson R. Cell differentiation induces TIF1beta association with centromeric heterochromatin via an HP1 interaction. J Cell Sci 2002;115(Pt 17):3439-48.
85. Cammas F, Herzog M, Lerouge T, Chambon P, Losson R. Association of the transcriptional corepressor TIF1beta with heterochromatin protein 1 (HP1): an essential role for progression through differentiation. Genes Dev
2004;18(17):2147-60.
86. Matsuda E, Agata Y, Sugai M, Katakai T, Gonda H, Shimizu A. Targeting of
Kruppel-associated box-containing zinc finger proteins to centromeric heterochromatin. Implication for the gene silencing mechanisms. J Biol Chem
2001;276(17):14222-9.
87. Singh PB, Miller JR, Pearce J, et al. A sequence motif found in a
Drosophila heterochromatin protein is conserved in animals and plants. Nucleic
Acids Res 1991;19(4):789-94.
88. Saunders WS, Chue C, Goebl M, et al. Molecular cloning of a human
homologue of Drosophila heterochromatin protein HP1 using anti-centromere
autoantibodies with anti-chromo specificity. J Cell Sci 1993;104 ( Pt 2):573-82.
222
89. Ye Q, Worman HJ. Interaction between an integral protein of the nuclear envelope inner membrane and human chromodomain proteins homologous to
Drosophila HP1. J Biol Chem 1996;271(25):14653-6.
90. Paro R, Hogness DS. The Polycomb protein shares a homologous domain with a heterochromatin-associated protein of Drosophila. Proc Natl Acad Sci U S
A 1991;88(1):263-7.
91. Maison C, Almouzni G. HP1 and the dynamics of heterochromatin maintenance. Nat Rev Mol Cell Biol 2004;5(4):296-304.
92. Ball LJ, Murzina NV, Broadhurst RW, et al. Structure of the chromatin
binding (chromo) domain from mouse modifier protein 1. EMBO J
1997;16(9):2473-81.
93. Cowieson NP, Partridge JF, Allshire RC, McLaughlin PJ. Dimerisation of a
chromo shadow domain and distinctions from the chromodomain as revealed by
structural analysis. Curr Biol 2000;10(9):517-25.
94. Lechner MS, Schultz DC, Negorev D, Maul GG, Rauscher FJ, 3rd. The mammalian heterochromatin protein 1 binds diverse nuclear proteins through a common motif that targets the chromoshadow domain. Biochem Biophys Res
Commun 2005;331(4):929-37.
95. Lachner M, O'Carroll D, Rea S, Mechtler K, Jenuwein T. Methylation of
histone H3 lysine 9 creates a binding site for HP1 proteins. Nature
2001;410(6824):116-20.
223
96. Bannister AJ, Zegerman P, Partridge JF, et al. Selective recognition of
methylated lysine 9 on histone H3 by the HP1 chromo domain. Nature
2001;410(6824):120-4.
97. Platero JS, Hartnett T, Eissenberg JC. Functional analysis of the chromo domain of HP1. EMBO J 1995;14(16):3977-86.
98. Grewal SI, Jia S. Heterochromatin revisited. Nat Rev Genet 2007;8(1):35-
46.
99. Cryderman DE, Grade SK, Li Y, Fanti L, Pimpinelli S, Wallrath LL. Role of
Drosophila HP1 in euchromatic gene expression. Dev Dyn 2005;232(3):767-74.
100. Yasuhara JC, Wakimoto BT. Oxymoron no more: the expanding world of
heterochromatic genes. Trends Genet 2006;22(6):330-8.
101. Cheutin T, McNairn AJ, Jenuwein T, Gilbert DM, Singh PB, Misteli T.
Maintenance of stable heterochromatin domains by dynamic HP1 binding.
Science 2003;299(5607):721-5.
102. Festenstein R, Pagakis SN, Hiragami K, et al. Modulation of
heterochromatin protein 1 dynamics in primary Mammalian cells. Science
2003;299(5607):719-21.
103. Stewart MD, Li J, Wong J. Relationship between histone H3 lysine 9 methylation, transcription repression, and heterochromatin protein 1 recruitment.
Mol Cell Biol 2005;25(7):2525-38.
104. Eskeland R, Eberharter A, Imhof A. HP1 binding to chromatin methylated at H3K9 is enhanced by auxiliary factors. Mol Cell Biol 2007;27(2):453-65.
224
105. Schultz DC, Ayyanathan K, Negorev D, Maul GG, Rauscher FJ, 3rd.
SETDB1: a novel KAP-1-associated histone H3, lysine 9-specific methyltransferase that contributes to HP1-mediated silencing of euchromatic genes by KRAB zinc-finger proteins. Genes Dev 2002;16(8):919-32.
106. Harte PJ, Wu W, Carrasquillo MM, Matera AG. Assignment of a novel bifurcated SET domain gene, SETDB1, to human chromosome band 1q21 by in situ hybridization and radiation hybrids. Cytogenet Cell Genet 1999;84(1-2):83-6.
107. Schultz DC, Friedman JR, Rauscher FJ, 3rd. Targeting histone deacetylase complexes via KRAB-zinc finger proteins: the PHD and bromodomains of KAP-1 form a cooperative unit that recruits a novel isoform of the Mi-2alpha subunit of
NuRD. Genes Dev 2001;15(4):428-43.
108. Wade PA, Jones PL, Vermaak D, Wolffe AP. A multiple subunit Mi-2 histone deacetylase from Xenopus laevis cofractionates with an associated Snf2 superfamily ATPase. Curr Biol 1998;8(14):843-6.
109. Zhang Y, LeRoy G, Seelig HP, Lane WS, Reinberg D. The dermatomyositis- specific autoantigen Mi2 is a component of a complex containing histone deacetylase and nucleosome remodeling activities. Cell 1998;95(2):279-89.
110. Tong JK, Hassig CA, Schnitzler GR, Kingston RE, Schreiber SL. Chromatin deacetylation by an ATP-dependent nucleosome remodelling complex. Nature
1998;395(6705):917-21.
225
111. Xue Y, Wong J, Moreno GT, Young MK, Cote J, Wang W. NURD, a novel complex with both ATP-dependent chromatin-remodeling and histone deacetylase activities. Mol Cell 1998;2(6):851-61.
112. Jenuwein T, Allis CD. Translating the histone code. Science
2001;293(5532):1074-80.
113. Ayyanathan K, Lechner MS, Bell P, et al. Regulated recruitment of HP1 to
a euchromatic gene induces mitotically heritable, epigenetic gene silencing: a
mammalian cell culture model of gene variegation. Genes Dev
2003;17(15):1855-69.
114. Wiznerowicz M, Jakobsson J, Szulc J, et al. The Kruppel-associated box
repressor domain can trigger de novo promoter methylation during mouse early
embryogenesis. J Biol Chem 2007;282(47):34535-41.
115. Wolf D, Goff SP. TRIM28 mediates primer binding site-targeted silencing
of murine leukemia virus in embryonic cells. Cell 2007;131(1):46-57.
116. Ellis J, Hotta A, Rastegar M. Retrovirus silencing by an epigenetic TRIM.
Cell 2007;131(1):13-4.
117. Wolf D, Cammas F, Losson R, Goff SP. Primer binding site-dependent restriction of murine leukemia virus requires HP1 binding by TRIM28. J Virol
2008;82(9):4675-9.
118. O'Geen H, Squazzo SL, Iyengar S, et al. Genome-wide analysis of KAP1
binding suggests autoregulation of KRAB-ZNFs. PLoS Genet 2007;3(6):e89.
226
119. Vogel MJ, Guelen L, de Wit E, et al. Human heterochromatin proteins form
large domains containing KRAB-ZNF genes. Genome Res 2006;16(12):1493-504.
120. Turner BM. Memorable transcription. Nat Cell Biol 2003;5(5):390-3.
121. Hake SB, Xiao A, Allis CD. Linking the epigenetic 'language' of covalent
histone modifications to cancer. Br J Cancer 2004;90(4):761-9.
122. Laity JH, Lee BM, Wright PE. Zinc finger proteins: new insights into
structural and functional diversity. Curr Opin Struct Biol 2001;11(1):39-46.
123. Rousseau-Merck MF, Koczan D, Legrand I, Moller S, Autran S, Thiesen HJ.
The KOX zinc finger genes: genome wide mapping of 368 ZNF PAC clones with zinc finger gene clusters predominantly in 23 chromosomal loci are confirmed by human sequences annotated in EnsEMBL. Cytogenet Genome Res 2002;98(2-
3):147-53.
124. Dehal P, Predki P, Olsen AS, et al. Human chromosome 19 and related
regions in mouse: conservative and lineage-specific evolution. Science
2001;293(5527):104-11.
125. Vissing H, Meyer WK, Aagaard L, Tommerup N, Thiesen HJ. Repression of
transcriptional activity by heterologous KRAB domains present in zinc finger
proteins. FEBS Lett 1995;369(2-3):153-7.
126. Agata Y, Matsuda E, Shimizu A. Two novel Kruppel-associated box-
containing zinc-finger proteins, KRAZ1 and KRAZ2, repress transcription through
functional interaction with the corepressor KAP-1 (TIF1beta/KRIP-1). J Biol Chem
1999;274(23):16412-22.
227
127. Kim SS, Chen YM, O'Leary E, Witzgall R, Vidal M, Bonventre JV. A novel member of the RING finger family, KRIP-1, associates with the KRAB-A transcriptional repressor domain of zinc finger proteins. Proc Natl Acad Sci U S A
1996;93(26):15299-304.
128. Beckstead R, Ortiz JA, Sanchez C, et al. Bonus, a Drosophila homolog of
TIF1 proteins, interacts with nuclear receptors and can inhibit betaFTZ-F1-
dependent transcription. Mol Cell 2001;7(4):753-65.
129. Shi X, Hong T, Walter KL, et al. ING2 PHD domain links histone H3 lysine
4 methylation to active gene repression. Nature 2006;442(7098):96-9.
130. Abrink M, Ortiz JA, Mark C, et al. Conserved interaction between distinct
Kruppel-associated box domains and the transcriptional intermediary factor 1
beta. Proc Natl Acad Sci U S A 2001;98(4):1422-6.
131. Littlewood TD, Hancock DC, Danielian PS, Parker MG, Evan GI. A modified
oestrogen receptor ligand-binding domain as an improved switch for the
regulation of heterologous proteins. Nucleic Acids Res 1995;23(10):1686-90.
132. Ayyanathan K, Fredericks WJ, Berking C, et al. Hormone-dependent tumor
regression in vivo by an inducible transcriptional repressor directed at the PAX3-
FKHR oncogene. Cancer Res 2000;60(20):5803-14.
133. Bienz M. The PHD finger, a nuclear protein-interaction domain. Trends
Biochem Sci 2006;31(1):35-40.
134. Jeanmougin F, Wurtz JM, Le Douarin B, Chambon P, Losson R. The
bromodomain revisited. Trends Biochem Sci 1997;22(5):151-3.
228
135. Li H, Ilin S, Wang W, et al. Molecular basis for site-specific read-out of
histone H3K4me3 by the BPTF PHD finger of NURF. Nature 2006;442(7098):91-
5.
136. Pena PV, Davrazou F, Shi X, et al. Molecular mechanism of histone
H3K4me3 recognition by plant homeodomain of ING2. Nature
2006;442(7098):100-3.
137. Wysocka J, Swigut T, Xiao H, et al. A PHD finger of NURF couples histone
H3 lysine 4 trimethylation with chromatin remodelling. Nature
2006;442(7098):86-90.
138. Bernstein E, Allis CD. RNA meets chromatin. Genes Dev
2005;19(14):1635-55.
139. Chang SC, Tucker T, Thorogood NP, Brown CJ. Mechanisms of X-
chromosome inactivation. Front Biosci 2006;11:852-66.
140. Gerber M, Shilatifard A. Transcriptional elongation by RNA polymerase II
and histone methylation. J Biol Chem 2003;278(29):26303-6.
141. Smale ST. The establishment and maintenance of lymphocyte identity
through gene silencing. Nat Immunol 2003;4(7):607-15.
142. Sekinger EA, Gross DS. Silenced chromatin is permissive to activator
binding and PIC recruitment. Cell 2001;105(3):403-14.
143. Sun FL, Cuaycong MH, Elgin SC. Long-range nucleosome ordering is
associated with gene silencing in Drosophila melanogaster pericentric
heterochromatin. Mol Cell Biol 2001;21(8):2867-79.
229
144. Daujat S, Zeissler U, Waldmann T, Happel N, Schneider R. HP1 binds specifically to Lys26-methylated histone H1.4, whereas simultaneous Ser27 phosphorylation blocks HP1 binding. J Biol Chem 2005;280(45):38090-5.
145. Garcia BA, Busby SA, Barber CM, Shabanowitz J, Allis CD, Hunt DF.
Characterization of phosphorylation sites on histone H1 isoforms by tandem mass spectrometry. J Proteome Res 2004;3(6):1219-27.
146. Kuzmichev A, Jenuwein T, Tempst P, Reinberg D. Different EZH2- containing complexes target methylation of histone H1 or nucleosomal histone
H3. Mol Cell 2004;14(2):183-93.
147. Wang H, An W, Cao R, et al. mAM facilitates conversion by ESET of
dimethyl to trimethyl lysine 9 of histone H3 to cause transcriptional repression.
Mol Cell 2003;12(2):475-87.
148. Schotta G, Lachner M, Sarma K, et al. A silencing pathway to induce H3-
K9 and H4-K20 trimethylation at constitutive heterochromatin. Genes Dev
2004;18(11):1251-62.
149. Noma K, Allis CD, Grewal SI. Transitions in distinct histone H3 methylation
patterns at the heterochromatin domain boundaries. Science
2001;293(5532):1150-5.
150. Brown KE, Baxter J, Graf D, Merkenschlager M, Fisher AG. Dynamic
repositioning of genes in the nucleus of lymphocytes preparing for cell division.
Mol Cell 1999;3(2):207-17.
230
151. Brown KE, Guest SS, Smale ST, Hahm K, Merkenschlager M, Fisher AG.
Association of transcriptionally silent genes with Ikaros complexes at centromeric heterochromatin. Cell 1997;91(6):845-54.
152. Brummelkamp TR, Bernards R, Agami R. Stable suppression of tumorigenicity by virus-mediated RNA interference. Cancer Cell 2002;2:243-7.
153. Sadowski I, Bell B, Broad P, Hollis M. GAL4 fusion vectors for expression in yeast or mammalian cells. Gene 1992;118(1):137-41.
154. Schultz DC, Vanderveer L, Buetow KH, et al. Characterization of chromosome 9 in human ovarian neoplasia identifies frequent genetic imbalance on 9q and rare alterations involving 9p, including CDKN2. Cancer Res
1995;55(10):2150-7.
155. Cammas F, Janoshazi A, Lerouge T, Losson R. Dynamic and selective interactions of the transcriptional corepressor TIF1 beta with the heterochromatin protein HP1 isotypes during cell differentiation. Differentiation
2007;75(7):627-37.
156. Ivanov AV, Peng H, Yurchenko V, et al. PHD domain-mediated E3 ligase
activity directs intramolecular sumoylation of an adjacent bromodomain required
for gene silencing. Mol Cell 2007;28(5):823-37.
157. Mascle XH, Germain-Desprez D, Huynh P, Estephan P, Aubry M.
Sumoylation of the transcriptional intermediary factor 1beta (TIF1beta), the Co-
repressor of the KRAB Multifinger proteins, is required for its transcriptional
231 activity and is modulated by the KRAB domain. J Biol Chem 2007;282(14):10190-
202.
158. Li X, Lee YK, Jeng JC, et al. Role for KAP1 serine 824 phosphorylation and
sumoylation/desumoylation switch in regulating KAP1-mediated transcriptional
repression. J Biol Chem 2007;282(50):36177-89.
159. Lee YK, Thomas SN, Yang AJ, Ann DK. Doxorubicin down-regulates
Kruppel-associated box domain-associated protein 1 sumoylation that relieves its
transcription repression on p21WAF1/CIP1 in breast cancer MCF-7 cells. J Biol
Chem 2007;282(3):1595-606.
160. White DE, Negorev D, Peng H, Ivanov AV, Maul GG, Rauscher FJ, 3rd.
KAP1, a novel substrate for PIKK family members, colocalizes with numerous
damage response factors at DNA lesions. Cancer Res 2006;66(24):11594-9.
161. Ziv Y, Bielopolski D, Galanty Y, et al. Chromatin relaxation in response to
DNA double-strand breaks is modulated by a novel ATM- and KAP-1 dependent pathway. Nat Cell Biol 2006;8(8):870-6.
162. Porter P. "Westernizing" women's risks? Breast cancer in lower-income countries. N Engl J Med 2008;358(3):213-6.
163. Parkin DM, Fernandez LM. Use of statistics to assess the global burden of breast cancer. Breast J 2006;12 Suppl 1:S70-80.
164. Sims AH, Howell A, Howell SJ, Clarke RB. Origins of breast cancer subtypes and therapeutic implications. Nat Clin Pract Oncol 2007;4(9):516-25.
232
165. Polyak K. Breast cancer: origins and evolution. J Clin Invest
2007;117(11):3155-63.
166. Dontu G, Al-Hajj M, Abdallah WM, Clarke MF, Wicha MS. Stem cells in normal breast development and breast cancer. Cell Prolif 2003;36 Suppl 1:59-72.
167. Kordon EC, Smith GH. An entire functional mammary gland may comprise the progeny from a single cell. Development 1998;125(10):1921-30.
168. Shackleton M, Vaillant F, Simpson KJ, et al. Generation of a functional mammary gland from a single stem cell. Nature 2006;439(7072):84-8.
169. Villadsen R. In search of a stem cell hierarchy in the human breast and its relevance to breast cancer evolution. APMIS 2005;113(11-12):903-21.
170. Carey LA, Perou CM, Livasy CA, et al. Race, breast cancer subtypes, and
survival in the Carolina Breast Cancer Study. JAMA 2006;295(21):2492-502.
171. Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke MF.
Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci U
S A 2003;100(7):3983-8.
172. Shipitsin M, Campbell LL, Argani P, et al. Molecular definition of breast
tumor heterogeneity. Cancer Cell 2007;11(3):259-73.
173. Liu R, Wang X, Chen GY, et al. The prognostic role of a gene signature
from tumorigenic breast-cancer cells. N Engl J Med 2007;356(3):217-26.
174. Merlo LM, Pepper JW, Reid BJ, Maley CC. Cancer as an evolutionary and
ecological process. Nat Rev Cancer 2006;6(12):924-35.
233
175. Yager JD, Liehr JG. Molecular mechanisms of estrogen carcinogenesis.
Annu Rev Pharmacol Toxicol 1996;36:203-32.
176. Yue W, Santen RJ, Wang JP, et al. Genotoxic metabolites of estradiol in
breast: potential mechanism of estradiol induced carcinogenesis. J Steroid
Biochem Mol Biol 2003;86(3-5):477-86.
177. Gadducci A, Biglia N, Sismondi P, Genazzani AR. Breast cancer and sex steroids: critical review of epidemiological, experimental and clinical
investigations on etiopathogenesis, chemoprevention and endocrine treatment of
breast cancer. Gynecol Endocrinol 2005;20(6):343-60.
178. Bolton JL, Thatcher GR. Potential mechanisms of estrogen quinone
carcinogenesis. Chemical research in toxicology 2008;21(1):93-101.
179. Rossouw JE, Anderson GL, Prentice RL, et al. Risks and benefits of estrogen plus progestin in healthy postmenopausal women: principal results
From the Women's Health Initiative randomized controlled trial. JAMA
2002;288(3):321-33.
180. Jemal A, Ward E, Thun MJ. Recent trends in breast cancer incidence rates by age and tumor characteristics among U.S. women. Breast Cancer Res
2007;9(3):R28.
181. Linet MS. Invited commentary: postmenopausal unopposed estrogen and breast cancer risk in the Women's Health Initiative--before and beyond. Am J
Epidemiol 2008;167(12):1416-20.
234
182. Nandi S, Guzman RC, Yang J. Hormones and mammary carcinogenesis in mice, rats, and humans: a unifying hypothesis. Proc Natl Acad Sci U S A
1995;92(9):3650-7.
183. Feigelson HS, Henderson BE. Estrogens and breast cancer. Carcinogenesis
1996;17(11):2279-84.
184. Henderson BE, Feigelson HS. Hormonal carcinogenesis. Carcinogenesis
2000;21(3):427-33.
185. Liehr JG. Genotoxic effects of estrogens. Mutat Res 1990;238(3):269-76.
186. Liehr JG, Roy D. Free radical generation by redox cycling of estrogens.
Free Radic Biol Med 1990;8(4):415-23.
187. Gaikwad NW, Yang L, Muti P, et al. The molecular etiology of breast
cancer: evidence from biomarkers of risk. Int J Cancer 2008;122(9):1949-57.
188. Cavalieri E, Chakravarti D, Guttenplan J, et al. Catechol estrogen quinones
as initiators of breast and other human cancers: Implications for biomarkers of
susceptibility and cancer prevention. Biochim Biophys Acta 2006;[Epub ahead of
print].
189. Spink DC, Spink BC, Cao JQ, et al. Induction of cytochrome P450 1B1 and
catechol estrogen metabolism in ACHN human renal adenocarcinoma cells. J
Steroid Biochem Mol Biol 1997;62(2-3):223-32.
190. Badawi AF, Cavalieri EL, Rogan EG. Effect of chlorinated hydrocarbons on
expression of cytochrome P450 1A1, 1A2 and 1B1 and 2- and 4-hydroxylation of
235
17beta-estradiol in female Sprague-Dawley rats. Carcinogenesis
2000;21(8):1593-9.
191. Badawi AF, Cavalieri EL, Rogan EG. Role of human cytochrome P450 1A1,
1A2, 1B1, and 3A4 in the 2-, 4-, and 16alpha-hydroxylation of 17beta-estradiol.
Metabolism 2001;50(9):1001-3.
192. Zhang Y, Gaikwad NW, Olson K, Zahid M, Cavalieri EL, Rogan EG.
Cytochrome P450 isoforms catalyze formation of catechol estrogen quinones that react with DNA. Metabolism 2007;56(7):887-94.
193. Roy D, Liehr JG. Temporary decrease in renal quinone reductase activity induced by chronic administration of estradiol to male Syrian hamsters.
Increased superoxide formation by redox cycling of estrogen. J Biol Chem
1998;263:3646-51.
194. Mobley JA, Bhat AS, Brueggemeier RW. Measurement of oxidative DNA damage by catechol estrogens and analogues in vitro. Chem Res Toxicol
1999;12(3):270-7.
195. Li KM, Todorovic R, Devanesan P, et al. Metabolism and DNA binding
studies of 4-hydroxyestradiol and estradiol-3,4-quinone in vitro and in female ACI
rat mammary gland in vivo. Carcinogenesis 2004;25:289-97.
196. Zahid M, Kohli E, Saeed M, Rogan E, Cavalieri E. The greater reactivity of
estradiol-3,4-quinone vs estradiol-2,3-quinone with DNA in the formation of
depurinating adducts: implications for tumor-initiating activity. Chem Res Toxicol
2006;19(1):164-72.
236
197. Saeed M, Rogan E, Fernandez SV, Sheriff F, Russo J, Cavalieri E.
Formation of depurinating N3Adenine and N7Guanine adducts by MCF-10F cells cultured in the presence of 4-hydroxyestradiol. Int J Cancer 2007;120(8):1821-4.
198. Zheng W, Xie DW, Jin F, et al. Genetic polymorphism of cytochrome P450-
1B1 and risk of breast cancer. Cancer Epidemiol Biomarkers Prev 2000;9(2):147-
50.
199. Kisselev P, Schunck WH, Roots I, Schwarz D. Association of CYP1A1
polymorphisms with differential metabolic activation of 17beta-estradiol and
estrone. Cancer Res 2005;65(7):2972-8.
200. Wen W, Ren Z, Shu XO, et al. Expression of cytochrome P450 1B1 and
catechol-O-methyltransferase in breast tissue and their associations with breast
cancer risk. Cancer Epidemiol Biomarkers Prev 2007;16(5):917-20.
201. Liehr JG. Is estradiol a genotoxic mutagenic carcinogen? Endocr Rev
2000;21:40-54.
202. Chakravarti D, Mailander PC, Li KM, et al. Evidence that a burst of DNA
depurination in SENCAR mouse skin induces error-prone repair and forms
mutations in the H-ras gene. Oncogene 2001;20(55):7945-53.
203. Mailander PC, Meza JL, Higginbotham S, Chakravarti D. Induction of A.T
to G.C mutations by erroneous repair of depurinated DNA following estrogen
treatment of the mammary gland of ACI rats. J Steroid Biochem Mol Biol
2006;101(4-5):204-15.
237
204. Fernandez SV, Russo IH, Russo J. Estradiol and its metabolites 4- hydroxyestradiol and 2-hydroxyestradiol induce mutations in human breast epithelial cells. Int J Cancer 2006;118(8):1862-8.
205. Zhu BT. Catechol-O-Methyltransferase (COMT)-mediated methylation metabolism of endogenous bioactive catechols and modulation by endobiotics and xenobiotics: importance in pathophysiology and pathogenesis. Curr Drug
Metab 2002;3(3):321-49.
206. Hui Y, Yasuda S, Liu MY, Wu YY, Liu MC. On the sulfation and methylation of catecholestrogens in human mammary epithelial cells and breast cancer cells.
Biol Pharm Bull 2008;31(4):769-73.
207. Fisher MB, Paine MF, Strelevitz TJ, Wrighton SA. The role of hepatic and extrahepatic UDP-glucuronosyltransferases in human drug metabolism. Drug
Metab Rev 2001;33(3-4):273-97.
208. Gall WE, Zawada G, Mojarrabi B, et al. Differential glucuronidation of bile
acids, androgens and estrogens by human UGT1A3 and 2B7. J Steroid Biochem
Mol Biol 1999;70(1-3):101-8.
209. Guillemette C, Belanger A, Lepine J. Metabolic inactivation of estrogens in
breast tissue by UDP-glucuronosyltransferase enzymes: an overview. Breast
Cancer Res 2004;6(6):246-54.
210. Guillemette C, De Vivo I, Hankinson SE, et al. Association of genetic
polymorphisms in UGT1A1 with breast cancer and plasma hormone levels.
Cancer Epidemiol Biomarkers Prev 2001;10(6):711-4.
238
211. Adegoke OJ, Shu XO, Gao YT, et al. Genetic polymorphisms in uridine
diphospho-glucuronosyltransferase 1A1 (UGT1A1) and risk of breast cancer.
Breast Cancer Res Treat 2004;85(3):239-45.
212. Rogan EG, Badawi AF, Devanesan PD, et al. Relative imbalances in
estrogen metabolism and conjugation in breast tissue of women with carcinoma:
potential biomarkers of susceptibility to cancer. Carcinogenesis 2003;24:697-702.
213. Montano MM, Deng H, Liu M, Sun X, Singal R. Transcriptional regulation
by the estrogen receptor of antioxidative stress enzymes and its functional
implications. Oncogene 2004;23:2442-53.
214. Ross D. Quinone Reductases Multitasking in the Metabolic World. Drug
Metabolism Reviews 2004;36:639-54.
215. Gaikwad NW, Rogan EG, Cavalieri EL. Evidence from ESI-MS for NQO1- catalyzed reduction of estrogen ortho-quinones. Free Rad Biol Med
2007;43:1289-98.
216. Menzel HJ, Sarmanova J, Soucek P, et al. Association of NQO1
polymorphism with spontaneous breast cancer in two independent populations.
Br J Cancer 2004;90(10):1989-94.
217. Sarmanova J, Susova S, Gut I, et al. Breast cancer: role of polymorphisms
in biotransformation enzymes. Eur J Hum Genet 2004;12(10):848-54.
218. Fagerholm R, Hofstetter B, Tommiska J, et al. NAD(P)H:quinone
oxidoreductase 1 NQO1(*)2 genotype (P187S) is a strong prognostic and
predictive factor in breast cancer. Nat Genet 2008.
239
219. Bianco NR, Perry G, Smith MA, Templeton DJ, Montano MM. Functional implications of antiestrogen induction of quinone reductase: inhibition of estrogen-induced deoxyribonucleic acid damage. Mol Endocrinol
2003;17(7):1344-55.
220. Montano MM, Chaplin L, Deng H, et al. Protective roles of quinone
reductase and antiestrogens in estrogen-induced mammary cell tumorigenesis.
Oncogene 2007;26:3587-90.
221. Strange RC, Spiteri MA, Ramachandran S, Fryer AA. Glutathione-S-
transferase family of enzymes. Mutat Res 2001;482(1-2):21-6.
222. Blair IA. Endogenous glutathione adducts. Curr Drug Metab
2006;7(8):853-72.
223. Pasquali L, Bedeir A, Ringquist S, Styche A, Bhargava R, Trucco G.
Quantification of CpG island methylation in progressive breast lesions from
normal to invasive carcinoma. Cancer Lett 2007;257(1):136-44.
224. Lee JS. GSTP1 promoter hypermethylation is an early event in breast
carcinogenesis. Virchows Arch 2007;450(6):637-42.
225. Griffith OW, Mulcahy RT. The enzymes of glutathione synthesis: gamma-
glutamylcysteine synthetase. Adv Enzymol Relat Areas Mol Biol 1999;73:209-67.
226. Wild AC, Moinova HR, Mulcahy RT. Regulation of gamma-glutamylcysteine
synthetase subunit gene expression by the transcription factor Nrf2. J Biol Chem
1999;274(47):33627-36.
240
227. Cole MP, Jones CT, Todd ID. A new anti-oestrogenic agent in late breast cancer. An early clinical appraisal of ICI46474. Br J Cancer 1971;25(2):270-5.
228. Tamoxifen for early breast cancer: an overview of the randomised trials.
Early Breast Cancer Trialists' Collaborative Group. Lancet 1998;351(9114):1451-
67.
229. Vogel VG, Costantino JP, Wickerham DL, et al. Effects of tamoxifen vs raloxifene on the risk of developing invasive breast cancer and other disease outcomes: the NSABP Study of Tamoxifen and Raloxifene (STAR) P-2 trial. JAMA
2006;295:2727-41.
230. Martino S, Cauley JA, Barrett-Connor E, et al. Continuing outcomes
relevant to Evista: breast cancer incidence in postmenopausal osteoporotic
women in a randomized trial of raloxifene. J Natl Cancer Inst 2004;96(23):1751-
61.
231. Cuzick J, Forbes JF, Sestak I, et al. International Breast Cancer
Intervention Study I Investigators. Long-term results of tamoxifen prophylaxis for breast cancer--96-month follow-up of the randomized IBIS-I trial. J Natl
Cancer Inst 2007;99:272-82.
232. Fisher B, Costantino JP, Wickerham DL, et al. Tamoxifen for the
prevention of breast cancer: current status of the National Surgical Adjuvant
Breast and Bowel Project P-1 study. J Natl Cancer Inst 2005;97:1652-62.
241
233. Powles TJ, Ashley S, Tidy A, Smith IE, Dowsett M. Twenty-year follow-up of the Royal Marsden randomized, double-blinded tamoxifen breast cancer prevention trial. J Natl Cancer Inst 2007;99:283-90.
234. Castrellon AB, Gluck S. Chemoprevention of breast cancer. Expert Rev
Anticancer Ther 2008;8(3):443-52.
235. Vogel VG. Recent results from clinical trials using SERMs to reduce the risk of breast cancer. Ann N Y Acad Sci 2006;1089:127-42.
236. Fabian C. Tamoxifen or raloxifene in postmenopausal women for prevention of breast cancer: a tale of two choices--counterpoint. Cancer
Epidemiol Biomarkers Prev 2007;16(11):2210-2.
237. Gruvberger-Saal SK, Bendahl PO, Saal LH, et al. Estrogen receptor beta
expression is associated with tamoxifen response in ERalpha-negative breast
carcinoma. Clin Cancer Res 2007;13(7):1987-94.
238. Speirs V, Walker RA. New perspectives into the biological and clinical
relevance of oestrogen receptors in the human breast. J Pathol 2007;211(5):499-
506.
239. Roger P, Sahla ME, Makela S, Gustafsson JA, Baldet P, Rochefort H.
Decreased expression of estrogen receptor beta protein in proliferative
preinvasive mammary tumors. Cancer Res 2001;61(6):2537-41.
240. Esslimani-Sahla M, Simony-Lafontaine J, Kramar A, et al. Estrogen
receptor beta (ER beta) level but not its ER beta cx variant helps to predict tamoxifen resistance in breast cancer. Clin Cancer Res 2004;10(17):5769-76.
242
241. Shaaban AM, Jarvis C, Moore F, West C, Dodson A, Foster CS. Prognostic significance of estrogen receptor Beta in epithelial hyperplasia of usual type with known outcome. Am J Surg Pathol 2005;29(12):1593-9.
242. Ogawa S, Inoue S, Watanabe T, et al. Molecular cloning and characterization of human estrogen receptor ßcx: a potential inhibitor of estrogen action in human. Nucleic Acids Res 1998;26:3505-12.
243. Peng B, Lu B, Leygue E, Murphy LC. Putative functional characteristics of human estrogen receptor-beta isoforms. J Mol Endocrinol 2003;30(1):13-29.
244. Delaunay F, Pettersson K, Tujague M, Gustafsson JA. Functional differences between the amino-terminal domains of estrogen receptors alpha and beta. Mol Pharmacol 2000;58(3):584-90.
245. Saville B, Wormke M, Wang F, et al. Ligand-, cell-, and estrogen receptor
subtype (alpha/beta)-dependent activation at GC-rich (Sp1) promoter elements.
J Biol Chem 2000;275(8):5379-87.
246. Bjornstrom L, Sjoberg M. Mechanisms of estrogen receptor signaling:
convergence of genomic and nongenomic actions on target genes. Mol
Endocrinol 2005;19(4):833-42.
247. Paech K, Webb P, Kuiper GG, et al. Differential ligand activation of
estrogen receptors ERalpha and ERbeta at AP1 sites. Science 1997;277:1508-10.
248. Zou A, Marschke KB, Arnold KE, et al. Estrogen receptor beta activates the
human retinoic acid receptor alpha-1 promoter in response to tamoxifen and
243 other estrogen receptor antagonists, but not in response to estrogen. Mol
Endocrinol 1999;13(3):418-30.
249. Liu MM, Albanese C, Anderson CM, et al. Opposing action of estrogen
receptors alpha and beta on cyclin D1 gene expression. J Biol Chem
2002;277(27):24353-60.
250. Barkhem T, Carlsson B, Nilsson Y, Enmark E, Gustafsson J, Nilsson S.
Differential response of estrogen receptor alpha and estrogen receptor beta to
partial estrogen agonists/antagonists. Mol Pharmacol 1998;54(1):105-12.
251. Katzenellenbogen BS, Choi I, Delage-Mourroux R, et al. Molecular
mechanisms of estrogen action: selective ligands and receptor pharmacology. J
Steroid Biochem Mol Biol 2000;74(5):279-85.
252. Kushner PJ, Agard DA, Greene GL, et al. Estrogen receptor pathways to
AP-1. J Steroid Biochem Mol Biol 2000;74:311-7.
253. Montano MM, Katzenellenbogen BS. The quinone reductase gene: a
unique estrogen receptor-regulated gene that is activated by antiestrogens. Proc
Natl Acad Sci 1997;94:2581-6.
254. Montano MM, Jaiswal AK, Katzenellenbogen BS. Transcriptional regulation
of the human quinone reductase gene by antiestrogen-liganded estrogen
receptor-alpha and estrogen receptor- beta. J Biol Chem 1998;273:25443-9.
255. Forman HJ, Ruden D. Introduction to serial reviews on EpRE and its signaling pathway. Free Radic Biol Med 2004;36(10):1197-8.
244
256. Jaiswal AK. Nrf2 signaling in coordinated activation of antioxidant gene expression. Free Radic Biol Med 2004;36(10):1199-207.
257. Rushmore TH, Morton MR, Pickett CB. The antioxidant responsive element. Activation by oxidative stress and identification of the DNA consensus sequence required for functional activity. The Journal of biological chemistry
1991;266(18):11632-9.
258. Li Y, Jaiswal AK. Regulation of human NAD(P)H:quinone oxidoreductase gene. Role of AP1 binding site contained within human antioxidant response element. J Biol Chem 1992;267(21):15097-104.
259. Prestera T, Holtzclaw WD, Zhang Y, Talalay P. Chemical and molecular regulation of enzymes that detoxify carcinogens. Proc Natl Acad Sci U S A
1993;90(7):2965-9.
260. Wasserman WW, Fahl WE. Functional antioxidant responsive elements.
Proc Natl Acad Sci U S A 1997;94(10):5361-6.
261. Venugopal R, Jaiswal AK. Nrf1 and Nrf2 positively and c-Fos and Fra1 negatively regulate the human antioxidant response element-mediated expression of NAD(P)H:quinone oxidoreductase1 gene. Proceedings of the
National Academy of Sciences of the United States of America
1996;93(25):14960-5.
262. Nguyen T, Huang HC, Pickett CB. Transcriptional regulation of the antioxidant response element. Activation by Nrf2 and repression by MafK. J Biol
Chem 2000;275(20):15466-73.
245
263. Itoh K, Chiba T, Takahashi S, et al. An Nrf2/small Maf heterodimer
mediates the induction of phase II detoxifying enzyme genes through antioxidant
response elements. Biochem Biophys Res Commun 1997;236(2):313-22.
264. Venugopal R, Jaiswal AK. Nrf2 and Nrf1 in association with Jun proteins
regulate antioxidant response element-mediated expression and coordinated
induction of genes encoding detoxifying enzymes. Oncogene 1998;17(24):3145-
56.
265. Fujiwara KT, Kataoka K, Nishizawa M. Two new members of the maf
oncogene family, mafK and mafF, encode nuclear b-Zip proteins lacking putative
trans-activator domain. Oncogene 1993;8(9):2371-80.
266. Kataoka K, Noda M, Nishizawa M. Maf nuclear oncoprotein recognizes sequences related to an AP-1 site and forms heterodimers with both Fos and
Jun. Mol Cell Biol 1994;14(1):700-12.
267. Xie T, Belinsky M, Xu Y, Jaiswal AK. ARE- and TRE-mediated regulation of gene expression. Response to xenobiotics and antioxidants. J Biol Chem
270:6894-900.
268. Bianco NR, Perry G, Smith MA, Templeton DJ, Montano MM. Functional implications of antiestrogen induction of quinone reductase: inhibition of estrogen-induced DNA damage. Molecular Endocrinology 2003;17:1344-55.
269. Wittmann BM, Wang N, Montano MM. Identification of a novel inhibitor of cell growth that is down-regulated by estrogens and decreased in breast tumors.
Cancer Res 2003;63:5151-8.
246
270. Soubeyran I, Quenel N, Mauriac L, Durand M, Bonichon F, Coindre JM.
Variation of hormonal receptor, pS2, c-erbB-2 and GSTpi contents in breast carcinomas under tamoxifen: a study of 74 cases. Br J Cancer 1996;73(6):735-
43.
271. Menzel HJ, Sarmanova J, Soucek P, et al. Association of NQO1
polymorphism with spontaneous breast cancer in two independent populations.
Br J Cancer 2004;90:1989-94.
272. Oestergaard MZ, Tyrer J, Cebrian A, et al. Interactions between genes involved in the antioxidant defence system and breast cancer risk. British journal of cancer 2006;95(4):525-31.
273. Udler M, Maia AT, Cebrian A, et al. Common germline genetic variation in
antioxidant defense genes and survival after diagnosis of breast cancer. J Clin
Oncol 2007;25(21):3015-23.
274. Sakoda LC, Blackston CR, Xue K, et al. Glutathione S-transferase M1 and
P1 polymorphisms and risk of breast cancer and fibrocystic breast conditions in
Chinese women. Breast Cancer Res Treat 2007.
275. Jordan VC. Tamoxifen or raloxifene for breast cancer chemoprevention: a
tale of two choices--point. Cancer Epidemiol Biomarkers Prev 2007;16(11):2207-
9.
276. Strange RC, Spiteri MA, Ramachandran S, Fryer AA. Glutathione-S- transferase family of enzymes. Mutat Res 2001;482:21-6.
247
277. Tew KD. Glutathione-associated enzymes in anticancer drug resistance.
Cancer Res 1994;54:4313-20.
278. Prestera T, Talalay P. Electrophile and antioxidant regulation of enzymes that detoxify carcinogens. Proceedings of the National Academy of Sciences of the United States of America 1995;92(19):8965-9.
279. Montano MM, Wittmann BM, Bianco NR. Identification and characterization of a novel factor that regulates quinone reductase gene transcriptional activity. J
Biol Chem 2000;275:34306-13.
280. Montano MM, Wittmann BM, Bianco NR. Identification and characterization of a novel factor that regulates quinone reductase gene transcriptional activity. J
Biol Chem 2000;275(44):34306-13.
281. Ju BG, Lunyak VV, Perissi V, et al. A topoisomerase IIbeta-mediated
dsDNA break required for regulated transcription. Science (New York, NY
2006;312(5781):1798-802.
282. Shang Y, Brown M. Molecular determinants for the tissue specificity of
SERMs. Science (New York, NY 2002;295(5564):2465-8.
283. Ramos-Gomez M, Kwak MK, Dolan PM, et al. Sensitivity to carcinogenesis
is increased and chemoprotective efficacy of enzyme inducers is lost in nrf2
transcription factor-deficient mice. Proceedings of the National Academy of
Sciences of the United States of America 2001;98(6):3410-5.
248
284. Routledge EJ, White R, Parker MG, Sumpter JP. Differential effects of xenoestrogens on coactivator recruitment by estrogen receptor (ER) alpha and
ERbeta. The Journal of biological chemistry 2000;275(46):35986-93.
285. Chang EC, Frasor J, Komm B, Katzenellenbogen BS. Impact of estrogen receptor beta on gene networks regulated by estrogen receptor alpha in breast cancer cells. Endocrinology 2006;147(10):4831-42.
286. Matthews J, Wihlen B, Tujague M, Wan J, Strom A, Gustafsson JA.
Estrogen receptor (ER) beta modulates ERalpha-mediated transcriptional activation by altering the recruitment of c-Fos and c-Jun to estrogen-responsive promoters. Molecular endocrinology (Baltimore, Md 2006;20(3):534-43.
287. Wong CW, Komm B, Cheskis BJ. Structure-function evaluation of ER alpha and beta interplay with SRC family coactivators. ER selective ligands.
Biochemistry 2001;40(23):6756-65.
288. Matthews J, Gustafsson JA. Estrogen signaling: a subtle balance between
ER alpha and ER beta. Molecular interventions 2003;3(5):281-92.
289. Pettersson K, Grandien K, Kuiper GG, Gustafsson JA. Mouse estrogen receptor beta forms estrogen response element-binding heterodimers with estrogen receptor alpha. Molecular endocrinology (Baltimore, Md
1997;11(10):1486-96.
290. Lis JT, Kraus WL. Promoter cleavage: a topoIIbeta and PARP-1 collaboration. Cell 2006;125(7):1225-7.
249
291. Pavri R, Lewis B, Kim TK, et al. PARP-1 determines specificity in a retinoid
signaling pathway via direct modulation of mediator. Molecular cell
2005;18(1):83-96.
292. Krishnakumar R, Gamble MJ, Frizzell KM, Berrocal JG, Kininis M, Kraus WL.
Reciprocal binding of PARP-1 and histone H1 at promoters specifies
transcriptional outcomes. Science (New York, NY 2008;319(5864):819-21.
293. Brummelkamp TR, Bernards R, Agami R. A system for stable expression of
short interfering RNAs in mammalian cells. Science 2002;296:550-3.
294. Sripathy SP, Stevens J, Schultz DC. The KAP1 corepressor functions to coordinate the assembly of de novo HP1-demarcated microenvironments of heterochromatin required for KRAB zinc finger protein-mediated transcriptional repression. Molecular and cellular biology 2006;26(22):8623-38.
295. Keshava N, Mandava U, Kirma N, Tekmal RR. Acceleration of mammary neoplasia in aromatase transgenic mice by 7,12-dimethylbenz[a]anthracene.
Cancer Lett 2001;167(2):125-33.
296. Luthra R, Kirma N, Jones J, Tekmal RR. Use of letrozole as a chemopreventive agent in aromatase overexpressing transgenic mice. J Steroid
Biochem Mol Biol 2003;86(3-5):461-7.
297. Li SA, Weroha SJ, Tawfik O, Li JJ. Prevention of solely estrogen-induced mammary tumors in female aci rats by tamoxifen: evidence for estrogen receptor mediation. J Endocrinol 2002;175(2):297-305.
250
298. Dann CT, Alvarado AL, Hammer RE, Garbers DL. Heritable and stable gene knockdown in rats. Proc Natl Acad Sci U S A 2006;103(30):11246-51.
299. Escande A, Pillon A, Servant N, et al. Evaluation of ligand selectivity using
reporter cell lines stably expressing estrogen receptor alpha or beta. Biochem
Pharmacol 2006;71(10):1459-69.
300. Mersereau JE, Levy N, Staub RE, et al. Liquiritigenin is a plant-derived
highly selective estrogen receptor beta agonist. Mol Cell Endocrinol 2008;283(1-
2):49-57.
301. Matsumura A, Ghosh A, Pope GS, Darbre PD. Comparative study of
oestrogenic properties of eight phytoestrogens in MCF7 human breast cancer
cells. J Steroid Biochem Mol Biol 2005;94(5):431-43.
302. Bianco NR, Chaplin L, Montano MM. Differential induction of quinone
reductase by phytoestrogens and protection against estrogen-induced DNA damage. Biochem J 2005;385:279-87.
303. Perillo B, Ombra MN, Bertoni A, et al. DNA oxidation as triggered by
H3K9me2 demethylation drives estrogen-induced gene expression. Science
2008;319(5860):202-6.
251