THE CHROMATIN REMODELING FACTOR CHD1L IN THE
PREIMPLANTATION EMBRYO AND IN ES CELLS
A DISSERTATION
SUBMITTED TO THE DEPARTMENT OF GENETICS
AND THE COMMITTEE ON GRADUATE STUDIES
OF STANFORD UNIVERSITY
FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
Alyssa Christine Snider
August 2010
http://creativecommons.org/licenses/by-nc/3.0/us/
This dissertation is online at: http://purl.stanford.edu/tk349zq2724
© 2010 by Alyssa Christine Snider. All Rights Reserved.
Re-distributed by Stanford University under license with the author.
This work is licensed under a Creative Commons Attribution-Noncommercial 3.0 United States License.
ii
I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.
Matthew Scott, Primary Adviser
I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.
Joanna Wysocka, Co-Adviser
I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.
Gerald Crabtree
I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.
Margaret Fuller
I certify that I have read this dissertation and that, in my opinion, it is fully adequatein scope and quality as a dissertation for the degree of Doctor of Philosophy.
Joseph Lipsick
Approved for the Stanford University Committee on Graduate Studies.
Patricia J. Gumport, Vice Provost Graduate Education
This signature page was generated electronically upon submission of this dissertation in electronic format. An original signed hard copy of the signature page is on file inUniversity Archives.
iii
iv
ABSTRACT
Early embryonic cell types such as the zygote, blastomeres of the preimplantation
embryo, and embryonic stem (ES) cells have powerful chromatin remodeling activities
that facilitate DNA-dependent processes such as transcription and DNA repair. These
chromatin-regulated processes are crucial for enacting complex gene expression
programs and ensuring genomic integrity for the developing embryo. Improving our
basic knowledge of chromatin remodeling in the preimplantation embryo and in
embryonic stem cells has implications for addressing human infertility and regenerative
medicine.
Chd1l encodes a chromatin remodeling factor and was highlighted as a candidate
developmental regulator from a screen in which factors were identified whose transcripts
are more highly expressed in the isolated inner cell mass (ICM) compared to the whole
blastocyst. Chd1l expression is developmentally regulated during a time course of
preimplantation development, peaking at the late morula stage, just prior to the formation
of the blastocyst. In addition, Chd1l is expressed in ES cells. Prior to this dissertation
research, the role of Chd1l had not been addressed, and its intriguing expression patterns
suggested Chd1l could be a novel regulator of DNA-dependent processes in early
developmental cell types. This dissertation describes research undertaken to address the
role of Chd1l in chromatin remodeling in the preimplantation embryo and in ES cells.
Four questions were addressed: 1) Is Chd1l essential in ES cells? 2) Does Chd1l
regulate gene expression in ES cells? 3) Is Chd1l essential in the preimplantation
embryo? 4) Does Chd1l contribute to the DNA damage response in ES cells or in the
preimplantation embryo?
To address the first question, Chd1l was knocked-down in mouse ES cells using a
shRNA targeting the Chd1l transcript. Reducing Chd1l protein to nearly undetectable
levels reveals that Chd1l is dispensable for ES cell viability, proliferation, and pluripotent
morphology. The second question was addressed by subjecting ES cells in which Chd1l
had been knock-down to genome-wide expression analysis. This study demonstrated that
global gene expression patterns were unaltered by Chd1l knock-down, confirming that
Chd1l is dispensable for transcription in general and, in particular, for maintaining
v
pluripotent transcriptional network. Chd1l is also dispensable for gene expression
programs associated with the formation of the primary germ layers, as differentiating
embryoid bodies demonstrate temporally appropriate repression of pluripotency markers
and activation of germ layer lineage markers.
To address whether Chd1l is essential in the preimplantation embryo, mouse
embryos were micro-injected at the single-cell stage (zygote stage) with antisense
morpholino (MO) oligos targeting the Chd1l transcript. Development was observed in
vitro for four days, during which time control embryos progressed to the blastocyst stage.
Embryos injected with Chd1l-MO arrested prior to the multi-cell stage, indicating that
Chd1l plays a crucial role during preimplantation embryogenesis. Knock-down by the
MO was confirmed at the transcript levels by microfluidic qPCR, and the arrest
phenotype was confirmed to be due to Chd1l deficiency by partial rescue upon co-
injection of Chd1l mRNA and Chd1l-MO.
During the course of this study, evidence from independent research groups
identified a role for Chd1l in the DNA damage response pathway in somatic cultured
cells. ES cells and cells of the early embryo are known to have stringent and unique
pathways to repair DNA damage to prevent mutation and genomic instability from
arising in the organism. To address whether Chd1l participates in the DNA damage
response in ES cells, ES cells in which Chd1l had been knocked-down were treated with
DNA damaging agents and assayed for survival rates. In direct contrast to the published
literature, in which reduction of Chd1l in somatic cells induced hypersensitivity to
induced DNA damage, reduction of Chd1l in ES cells conferred resistance to induced
DNA damage. The most likely explanation for this is that Chd1l participates in a
damage-induced apoptotic response in ES cells. This is supported by data showing that
Chd1l over-expression is sufficient to induce apoptosis in ES cells.
Interestingly, apoptosis induced by over-expression of Chd1l occurs specifically
in ES cells, as ES cells differentiated for as little as two days by removal of the cytokine
LIF no longer undergo Chd1l-induced apoptosis. This switch in the effect of Chd1l over-
expression during differentiation suggests that Chd1l responds to DNA damage very
differently in ES cells than in differentiated somatic cells. Indeed, ES cells are
significantly more proficient in initiating apoptosis to eliminate DNA damage than their
vi
differentiated counterparts. A model in which Chd1l responds to DNA damage by
initiating apoptosis in ES cell but not in differentiated cells could explain the conflicting
results presented in this dissertation and in other published studies that used differentiated
cells.
A likely role for Chd1l in DNA repair in the preimplantation embryo could
explain the Chd1l arrest phenotype. DNA repair is particularly critical in the zygote to
repair the paternal genome, and inefficient DNA repair leads to decreased fertility in
humans. The zygote relies heavily on non-homologous end joining (NHEJ) to repair
double-stranded breaks (DSBs). Reliance on NHEJ is in common with differentiated
cells but not ES cells, which primarily use homologous recombination (HR) to repair
DSBs. Biochemical evidence supports the involvement of Chd1l in NHEJ as it interacts
with NHEJ-specific proteins but not HR-specific proteins. Modest increases in staining
for a marker of DSBs can be seen in embryos injected with Chd1l-MO, indicating that
deficient DNA repair could underlie the Chd1l arrest phenotype.
In summary, this dissertation describes an essential role for Chd1l in the
preimplantation embryo that could be due to defects in DNA repair. In contrast, Chd1l is
dispensable for ES cell gene expression, pluripotency, and differentiation. Chd1l likely
contributes to an ES cell-specific apoptotic response to DNA damage. It is proposed here
that Chd1l functions through the NHEJ pathway, a pathway critical in the zygote and in
differentiated cells, but not in ES cells. Therefore, Chd1l functions minimally, or not at
all, in regulating gene expression and contributes to the DNA damage response in a
developmental stage-specific and/or cell type-specific manner.
vii
PREFACE
The first decade of the twenty-first century during which I began graduate school
was an extraordinary time of revolutionary advancements in genetics, epigenetics, and
biotechnology. The Human Genome Project reported the first draft of the sequenced
human genome in 2001 [1,2] and its completion in 2007, sparking a burst in activity in
the field of genetics. Advancements in high throughput technology led to the so-called
“genomics era” and have culminated in the advent of “personal genomics” where
individuals can have their genomes sequenced or affordably SNP-genotyped by
companies such as 23andMe and Navigenics. Dolly the sheep, the first mammal cloned
through the reprogramming of an adult somatic cell, was born in 1996, undermining the
dogma that differentiation is an irreversible phenomenon and sparking interest in the field
of epigenetics [3]. Epigenetic research has transitioned in parallel with genetic research
to large-scale, genome-wide studies that have lead to great advancements in our
knowledge of the “epigenome” and its surprising plasticity. The public watched in
dismay as Dolly passed away in 2003, showing signs of premature aging and raising
questions about the molecular basis of “youth” and pluripotency. In 2007, when bans on
federal funding of human ES cells were still in effect under the Bush administration, two
groups transformed adult human cells into “induced pluripotent stem” (iPS) cells by the
introduction of only a few factors [4,5]. The ability to avoid the controversial use of
embryos to generate pluripotent stem cells, and the ability to use adult cell types instead,
revolutionized the fields of personal and regenerative medicine.
It was in the midst of this climate of excitement, undiminished by the clear skies
and temperate weather in Palo Alto, California, that I began my studies in the department
of Genetics at Stanford University. I was enticed to join Dr. Matthew Scott’s laboratory
by his attitude of scientific open-mindedness and by a chromatin project that was
underway. The laboratory was located in the Clark Center, built in 2003 for the purpose
of housing the multi-disciplinary Bio-X Department. With developmental biologists,
statisticians, bioengineers, computer scientists, and neurobiologists as neighbors, and a
Pete’s Coffee shop that was strategically situated in the Clark Center to promote
scientific discussions, I pursued the most difficult academic challenge of my life.
viii
Dr. Tian Wang, now a professor at the University of Chicago, Illinois, was a post-
doc in Dr. Scott’s laboratory who was interested in the property of pluripotency and how
it was established and maintained in embryonic stem (ES) cells. She conducted a screen
to identify candidate pluripotency regulatory factors by asking which transcripts are
enriched in the inner cell mass (ICM), the part of the embryo that is pluripotent and from
which ES cells are derived. We became interested in the subset of these ICM-enriched
factors that were involved in regulating chromatin dynamics. Among the genes whose
transcripts were enriched in the ICM was Chd1l (Chromodomain, ATPase/Helicase, DNA
binding 1-like), a chromatin remodeling factor that is part of a very important protein
family with diverse developmental roles. At the inception of the project, no
developmental or molecular studies had been reported on Chd1l, making the gene an
intriguing subject of study to me.
This dissertation is an account of research undertaken to determine the role of
Chd1l in the preimplantation embryo and in embryonic stem cells. The first introductory
chapter will focus on chromatin remodeling in these systems. Because Chd1l is a
member of the SNF2 family, members of which have well established roles in gene
expression, and because Chd1l has more recently been shown to have a role in the
response to DNA damage, the introduction will describe chromatin remodeling activities
that pertain in particular to gene regulation and DNA repair. The role of Chd1l was
studied in ES cells and, in collaboration with the laboratory of Dr. Mylene Yao, in the
preimplantation embryo. Chapter 2 documents the results of these studies, revealing that
Chd1l is essential in the preimplantation embryo but dispensable in ES cells. Along with
Dr. Wang, my Yao laboratory collaborators, my advisor and co-advisor, I will submit
Chapter 2 as a manuscript for publication to PlosOne. Chapters 3 and 4 describe my
investigations into the molecular function of Chd1l that could reveal hidden phenotypes
in ES cells or explain the arrest phenotype in preimplantation embryos. Chapter 5
discusses the implications of these findings and the directions in which I see the field
progressing.
ix
ACKNOWLEDGEMENTS
This thesis project is the result of the efforts of many people at Stanford,
beginning with my advisor Dr. Matthew Scott, who always encouraged living a fulfilling,
balanced life. In particular, I would like to express my regards to my two “lab moms,”
our laboratory administrator Diane Bush, and our research associate Kaye Suyama,
without whom I may never have graduated. My experience would have been very
different without the humor, intelligence, and kindness of the graduate students beside
whom I worked, Monique Barakat, Manuel Lopez, and Tyler Hillman, and three special
colleagues, Dariya Glaser, Fraser Tan, and Dr. Timothy Reddy. It was a pleasure to work
with my collaborators in Dr. Mylene Yao’s laboratory, especially my micro-injectionist
Denise Leong. I am endlessly grateful to my co-advisor Dr. Joanna Wysocka and her
husband and collaborator Dr. Tomek Swigut for lending their genius toward my project.
I am also grateful to the rest of my committee members, Dr. Jerry Crabtree, Dr. Margaret
Fuller, Dr. Julie Baker (on my committee until the final three months), and Dr. Joseph
Lipsick. I am honored to have known these remarkable people.
I would like to thank my family for their support during the challenging six years
of graduate school: my father, whose passion for academia first inspired me to pursue
higher education; my mother, whose unconditional love carried me through difficult
times of self-perceived failure; my older sister, who encouraged me to trail-blaze my own
path; my younger sister, who reminded me to stay true to myself; and my husband, who
gave me a new reason for living and excelling.
x
DEDICATION
This dissertation is dedicated to my father, Dr. Mervin G. Wright
xi
CONTENTS
1. Chromatin Remodeling in Preimplantation Embryos and
Embryonic Stem Cells ............................................................................... 1
Organization of the Mammalian Genome ...................................................................... 1
Chromatin Regulatory Proteins ...................................................................................... 2
DNA methyltransferases .............................................................................................. 2
Post-translational modifications .................................................................................. 3
ATP-dependent chromatin remodeling factors ............................................................ 6
Epigenetic Pluripotency in ES Cells ............................................................................... 8
DNA Repair in ES cells .................................................................................................. 9
Chromatin Remodeling in the Zygote and Cleavage-Stage Embryo ............................ 13
Chromatin states of the gametes ................................................................................ 13
Chromatin remodeling in the zygote ......................................................................... 14
Maternal/zygotic transition ........................................................................................ 15
Chromatin Remodeling in the Blastocyst ..................................................................... 16
Differentiation of the ICM ......................................................................................... 16
Epigenetic differences between the TE and ICM ...................................................... 18
X chromosome inactivation ....................................................................................... 19
Derivation of ES cells from the ICM ......................................................................... 20
DNA Repair in the Zygote ............................................................................................ 22
Chd1l as an Oncogene and DNA Damage Response Protein ....................................... 24
Summary ....................................................................................................................... 26
2. The Chromatin Remodeling Factor Chd1l Is Required in
the Preimplantation Embryo ..................................................................33
Abstract ......................................................................................................................... 34
Introduction ................................................................................................................... 34
Results ........................................................................................................................... 37
xii
Chromatin factors are compartmentalized in the blastocyst ...................................... 37
Chd1l expression patterns suggest a developmental role .......................................... 40
Chd1l is dispensable for ES cell pluripotency and proliferation ............................... 40
Chd1l does not regulate gene expression in ES cells ................................................ 41
Chd1l is not required for differentiation of ES cells .................................................. 42
Chd1l transcripts are abrogated in MO-injected embryos ......................................... 42
Embryos injected with Chd1l-targeting MOs arrest prior to blastocyst
stage ...................................................................................................................... 43
Chd1l phenotype is partially rescued by co-injection of Chd1l mRNA .................... 44
Discussion ..................................................................................................................... 44
Methods ........................................................................................................................ 46
3. Finding Direct Transcriptional Targets of Chd1l in ES
Cells ...........................................................................................................57
Introduction ................................................................................................................... 57
Results ........................................................................................................................... 59
Chromatin Immunoprecipitation ............................................................................... 59
Sequencing and Analysis ........................................................................................... 60
Re-analysis using deeper sequencing ........................................................................ 61
HA-tagged Chd1l targeting vector ............................................................................. 62
Discussion ..................................................................................................................... 62
Methods ........................................................................................................................ 64
Contributing Collaborators ........................................................................................... 66
4. Molecular Functions of Chd1l in Early Development .........................74
Introduction ................................................................................................................... 74
Results ........................................................................................................................... 78
Reduction of Chd1l increases DNA damage tolerance in ES cells ........................... 78
Over-expression of Chd1l kills ES cells but not differentiated cells ......................... 80
xiii
Chd1l forms a ~500 kD complex in ES cells ............................................................ 83
γ-H2AX marks uninjected embryos and embryos injected with Chd1l-
MO ........................................................................................................................ 83
PAR marks uninjected embryos and embryos injected with Chd1l-
MO ........................................................................................................................ 85
Discussion ..................................................................................................................... 86
Methods ........................................................................................................................ 90
5. General Discussion ................................................................................105
Chd1l is Essential in the Preimplantation Embryo ..................................................... 105
Chd1l is Non-Essential in ES cells ............................................................................. 108
Role of Chd1l in DNA Repair in ES cells .................................................................. 110
Oncogenic Potential of Chd1l ..................................................................................... 112
Paradoxes in Chd1l Function in ES Cells ................................................................... 114
REFERENCES ................................................................................................117
xiv
LIST OF FIGURES
Figure 1.1 PAR modifies nuclear proteins............................................................... 28
Figure 1.2 ATP-dependent chromatin remodeling ................................................. 29
Figure 1.3 SNF2 family of chromatin remodeling factors ..................................... 30
Figure 1.4 Chromatin remodeling in the preimplantation embryo ...................... 31
Figure 1.5 Chd1l and the PARP-dependent DNA damage response .................... 32
Figure 2.1 Chromatin remodeling factors are enriched in the ICM ..................... 50
Figure 2.2. Chd1l is a candidate developmental regulator ...................................... 51
Figure 2.3. Chd1l is non-essential in ES cells ............................................................ 53
Figure 2.4. Chd1l MO knock-down .......................................................................... 54
Figure 2.5. Chd1l embryonic arrest phenotype ........................................................ 56
Figure 3.1 Preliminary validations for ChIP ........................................................... 67
Figure 3.2 Distribution of reads around TSS .......................................................... 68
Figure 3.3 Validation of Chd1l ChIP peaks ............................................................ 71
Figure 3.4. Re-analysis of two combined Chd1l ChIP sequencing runs ............... 72
Figure 3.5. Chd1l-HA targeting vector .................................................................... 73
Figure 4.1. Sensitivity to DNA damage in ES cells expressing Chd1l-shRNA ..... 93
Figure 4.2. Effects of over-expression of K71R or WT Chd1l. .............................. 95
Figure 4.3. Chd1l over-expression specifically kills undifferentiated ES cells ..... 96
Figure 4.4. Size Fractionation of a Chd1l-containing protein complex ................ 98
Figure 4.5. γH2AX staining in embryos injected with Chd1l-MO. ..................... 101
Figure 4.6. PAR staining in embryos injected with Chd1l-MO. .......................... 104
Figure 5.1 Chd1l Phenotypes .................................................................................. 116
LIST OF TABLES
Table 3.1 Number of reads obtained for ChIP libraries ........................................ 68
Table 3.2. Number of peaks called for ChIP samples ............................................. 69
Table 3.3. Number of peaks called for the combined analysis ............................... 69
1
1. Chromatin Remodeling in Preimplantation Embryos and
Embryonic Stem Cells
Organization of the Mammalian Genome
Within the diploid human cell nucleus is approximately 6 billion base pairs of
DNA divided into 46 macromolecules called chromosomes [1]. Between them, they
contain a little over 20,000 genes that encode the proteins necessary for life. Some of
these proteins, in turn, handle DNA-related processes such as transcription, DNA
replication, and DNA repair. Elaborate and diverse transcriptional programs establish the
identity and functional specificity of tens of trillions of cells that comprise the adult body.
Housekeeping genes are constitutively expressed in all cell types, whereas genes
encoding proteins with highly specific functions must be repressed in one cell type but
active in another. Cell type specific gene expression patterns are maintained
epigenetically by modification of DNA, histones, and other proteins that interact with
DNA. During cell division, DNA replicates and condenses into discrete mitotic
chromosome structures, and sister chromatids are physically tethered and distributed to
daughter cells. Chromosomes are huge molecules, the smallest having a molecular
weight of about 15 billion g/mol and the largest 80 billion g/mol, compared to a molecule
of water whose molecular weight is 18 g/ml. Each nucleic acid is comprised of many
chemical bonds that are subject to spontaneous hydrolysis, oxidation, and alkylation, and
it is estimated that a single cell experiences tens of thousands of lesions due to
endogenous sources every day [6]. Because of this, chromosomes must be constantly
monitored and repaired by specialized DNA repair machinery.
If stretched end to end, the size of the genome is 2 meters long, about 340,000
times the diameter an average nucleus (6 µm). DNA must therefore be organized by
several degrees of complexity to fit into the nucleus. Histone octamers (two of each
H2A, H2B, H3 and H4) function like spools, wrapping DNA around themselves like
thread and forming nucleosomal arrays, the primary structure of chromatin [7].
Micrococcal nuclease digestion reveals that 146 base pairs of DNA is wrapped around
each histone octamer, and each core nucleosome is separated by ~100 base pairs of linker
2
DNA [8,9]. The presence of the histone H1 within the linker sequence facilitates
formation of secondary chromatin structure, the 30 nm fiber [10]. The crystal structure of
the core nucleosome was solved in 1997 and remains, in this author’s biased opinion, one
of the most beautiful crystal structures to date [11].
But such highly ordered packaging of the genome presents a significant challenge
for proteins that need access to DNA [12]. If DNA is rendered inaccessible, processes
necessary for viability including replication, transcription, and DNA surveillance and
repair would be fatally impaired [12,13,14]. Fortunately, nucleosomes and higher-
ordered chromatin are dynamic structures [15], and a host of chromatin regulatory
proteins are dedicated to remodeling chromatin structure to facilitate chromatin-
dependent processes. The focus of this introductory chapter is on chromatin remodeling
in the pluripotent cells of preimplantation embryo and in embryonic stem (ES) cells, cells
that contain the some of the most potent chromatin remodeling activities, except perhaps
for primordial germ cells [16]. Here chromatin remodeling is broadly defined as any
alteration of chromatin structure for the purpose of regulating gene expression or other
chromatin-dependent processes, such as DNA repair.
Chromatin Regulatory Proteins
Chromatin regulatory proteins can be categorized into three broad classes: (1)
enzymes that regulate methylation on DNA directly, (2) enzymes that add post-
translational modifications onto histone tails, and (3) enzymes that harness the power of
ATP to force torsional changes in chromatin, catalyze nucleosome sliding, and facilitate
histone eviction or deposition [17,18].
DNA methyltransferases
The enzymes that methylate CpG islands in the genome are called DNA
methyltransferases, or DNMTs, and their activity usually opposes transcription [19,20].
DNA methylation is required for allele-specific expression of imprinted genes [21], X
chromosome inactivation in females [22,23], the silencing of transposable elements
[24,25], and is associated with differentiation in somatic cells [20]. However, the
simplistic view that DNA methylation always leads to chromatin compaction and gene
3
silencing has more recently been under scrutiny, and a more complex scenario probably
functions in vivo [26]. The essential DNMT1 maintains CpG methylation through cell
division by methylating the 5th
carbon on cytosine nucleotides in the newly synthesized
strand of hemi-methylated DNA. The DNMTs 3a and 3b are responsible for de novo
DNA methylation, particularly during development [27]. A mechanism of active DNA
demethylation has been proposed, but the enzyme responsible for this activity has not
been identified in mammals.
Post-translational modifications
Histones contain flexible C-terminal tails that are subject to a variety of post-
translational modifications (PTMs) including methylation, acetylation, PARylation,
phosphorylation, and ubiquitinilation. Because PTMs can modulate differential gene
expression patterns, even in genetically identical cells, a “histone code” has been
proposed [28]. It is becoming increasingly clear that PTMs function in combination with
each other, and there is not always a direct relationship between one PTM and gene
expression [29]. Although PTMs are typically included when discussing “epigenetics,” it
is unresolved how PTMs are inherited through cell divisions. This is in contrast with
DNA methylation, which can be propagated during replication by the activity of the
maintenance DNMT1 that recognizes hemimethylated DNA.
A variety of enzymes exist that post-translationally modify specific residues in
histone tails. The enzymes that catalyze histone acetylation are called histone
acetyltransferases, or HATs, and the enzymes that remove acetyl groups are the histone
deacetylases, or HDACs. Histone acetylation on the promoters of genes is associated
with transcriptional activity. Histone methyltransferases (KMTs) are the enzymes that
add methyl groups to histone tails. This PTM can be associated with either
transcriptional activation or repression, depending on which lysine residue is modified.
For example, methylation of lysine 4 on histone 3 (H3K4me) is associated with gene
expression, and trimethylation of H3K27me3 is strongly associated with gene repression.
It was long expected that histone methylation would be a reversible modification similar
to histone acetylation, but the first histone demethylase (KDM), LSD1, now called
4
KDM1a, was only recently discovered [30]. Since then, a large number of jumonji-
domain containing proteins have been shown to have histone KDM activity [31].
Poly(ADP-ribosyl)ation, or PARylation, is a PTM that covalently attaches
negatively charged polymers of ADP-ribose onto histones and other nuclear proteins that
interact with DNA (Fig. 1) [32]. The PARP family of polymerases contains 18 members,
and most, if not all, can synthesize long linear or branching chains of up to 200 ADP-
ribose moieties using NAD+
as a substrate [32,33]. The primary acceptor protein is
PARP-1 itself, which, when activated, PARylates itself in an automodification reaction
[32]. The presence of PAR is extremely transient in nature, with a half-life of less than 1
minute due to the abundance of a single glycohydrolase enzyme, PARG, which rapidly
cleaves PAR polymers endo- and exo-glycositically, releasing free ADP-ribose units
[34,35]. The role of PAR in the cell is diverse and depends on the PARP family member
involved. PARP-1 and PARP-2 are the most studied and best understood PARPs, having
traditional roles in DNA damage response and repair [34,36,37,38,39]. One PARP,
PARP3, localizes to centrosomes and its over-expression interferes with progression
through the G1/S phase transition and may participate in DNA damage surveillance [40].
Another PARP, Tankyrase1 (Tank1), associates with TRF1 and other telomeric proteins
and is involved in telomere homeostasis [41]. Long term over-expression of Tank1
induces telomere elongation [42]. On the other hand, over-expression of another
tankyrase, Tank2, that also associates with TRF1 induces rapid and widespread PARP-
dependent cell death [43].
It has long been known that PARylation can contribute to chromatin relaxation
[44,45,46]. PARylation of the major histone acceptor H1 by PARP-1 in biochemical
assays induces relaxation of the 30 nm fiber, in a manner highly similar to chromatin
relaxation in the absence of histone H1, and is proposed to increase DNA accessibility
[46]. It was later shown that PARP-1 activity could contribute to either chromatin
relaxation or compaction, depending on the context and NAD +
supplies [47]. More
recently, the involvement of PARP-1 in regulating transcription has become apparent.
Like histone H1, PARP-1 binds to nucleosomes. In in vitro assays without NAD+,
PARP-1 incorporates into chromatin, decreases micrococcal nuclease sensitivity, and
represses reporter gene expression. When NAD+ is introduced, chromatin relaxes,
5
PARP-1 is displaced from nucleosomes, and reporter gene transcription is re-activated
[48]. In vivo ChIP-chip studies have shown that the presence of PARP-1 and absence of
histone H1 at promoters of genes is associated with gene activation [49].
Automodification of PARP-1 is the dominant mode of PARylation and, because it can
regulate gene regulation, it may also be a part of the “histone code” [45,50].
The longest known role of PARP-1 is its ability to respond to DNA damage and
facilitate repair [34]. The specific activity of PARP-1 is amplified 500-fold by the
presence of single strand breaks or double strand breaks in chromatin [32]. Its activation
leads to PARP-1 dimerization, automodification, and DNA binding, and to chromatin
relaxation. PARylated PARP-1 then recruits components of the base excision repair
(BER) pathway, including XRCC1 [51]. Parp-1 knock-out mice show hypersensitivity to
γ-irradiation and treatment with DNA damaging agents MMS, MNU, and MNNG at the
whole-animal level [38,52,53] and defects in DNA repair at the cellular level [54]. Even
in these mice, some residual PAR activity was observed, and this was found to be due to
a highly similar PARP-2 protein [55,56]. Parp-2 knock-out mice also show
hypersensitivity to γ-irradiation and DNA damage induced by alkylating agents [57].
PARP-1 and PARP-2 have partially redundant roles, as double knock-out mice of Parp-1
and Parp-2 display embryonic lethality at E7.5 [57]. However, specific roles for PARP-2
are becoming elucidated, such as binding to a telomeric protein TRF2 [58], and
preferential heteromodification of histone H2B [59], as opposed to preferential
heteromodification of histone H1 by PARP-1 [46].
The macro domain is a module that binds PAR. Macro domain-containing
proteins can be found in eukaryotes, bacteria, and archaeans, but include only a small
mammalian family [60,61]. Several PARPs, PARP-14, PARP-15, and PARP-9 (bal, the
lymphoma risk factor protein), contain a macro domain, and although it is tempting to
propose feedback signaling loops, these PARPs have not been confirmed to bind PAR
and may not because of divergent sequence [33]. The H2A histone variant macroH2A is
a component of chromatin in the inactive X chromosome [62] and its ability to bind PAR
has been shown [50]. Upon PARP-1 activation, macroH2A facilitates chromatin
compaction and hinders the recruitment of DNA repair proteins. These studies
underscore the conclusion that PAR has diverse roles in the cell that can facilitate
6
heterochromatin or euchromatin formation, depending on the context in which it is
synthesized and recognized.
In addition to its cytoprotective role, PARP-1 can mediate an alternative response
to DNA damage—apoptosis. A cell has two choices when confronted by DNA damage:
to repair the damage or to initiate programmed cell death. Surprisingly, Parp-1 knock-
out mice also show increased resistance to ischemia-reperfusion injury, inflammation-
related injury [63], diabetes [64], and hyperoxic damage. Under certain conditions,
PARP-1 appears to initiate apoptosis through the caspase-independent AIF-mediated
pathway [65]. AIF is normally sequestered in the mitochondrion, but upon PARP-1
activation, AIF translocates to the nucleus where it induces widespread chromatin
condensation and fragmentation. Translocation of AIF to the nucleus is suppressed by
PARP inhibitors, and PARP-dependent cell death can be suppressed by microinjection of
inhibitory AIF antibodies, demonstrating co-dependence. The mechanism by which
PARP activation in the nucleus triggers AIF release from the mitochondrion remains
mysterious [65].
ATP-dependent chromatin remodeling factors
The third class of chromatin regulatory enzymes is the ATP-dependent chromatin
remodeling factors (CRFs) that assemble into large, multisubunit complexes. These
CRFs utilize ATP to drive conformational changes in chromatin, and the outcome of their
activity depends on what other factors associate with them (Fig. 2) [66,67]. A single
CRF can have diverse functions due to the combinatorial nature of complex assembly; a
single subunit switch can alter or reverse its function [68]. This can happen between cell
types where a subunit is expressed in one cell type but repressed in another, or even in the
same cell where assembly is context dependent [69,70,71,72].
The best-known chromatin remodeling complex is the BAF complex, which
contains the prototypical mammalian Brm homologs, Brg1 and Brm. These chromatin
remodeling factors define the SNF2 family of chromatin remodeling factors by the
presence of the split DNA-dependent helicase domain. Rather than unwinding DNA like
other helicases, SNF2-like ATPase/helicase proteins translocate along DNA and induce
superhelical torsion and/or slide nucleosomes [73,74,75,76,77,78,79,80]. Chromatin
7
remodeling factors contain, in addition to an N-terminal ATPase/helicase domain, other
accessory domains which provide specificity of function [81]. SNF2-like CRFs achieve
the ability to bind DNA sequence-specifically by incorporating sequence-specific
transcription factors into the complex [82].
Classically, there are four subclasses of the SNF2 superfamily: (1) SWI/SNF,
including Brg1 and Brm, both of which have C-terminal bromodomains that recognize
specific acetylated lysines in histones (2) the CHD class, members of which are
characterized by the presence of chromodomains that recognize methylated lysines on
histone tails, (3) the INO80 class, which is characterized by an ATPase/helicase domain
with an extended central split, and (4) the ISWI class, which contains two mammalian
homologs, Snf2l and Snf2h, having SANT and SLIDE domains responsible for
mobilizing nucleosomes [17]. There remain a number of proteins that contain the SNF2-
like DNA-dependent ATPase/helicase domain but that do not fit into these four major
categories, including Chd1l, which is the subject of this dissertation (Fig. 3) [81].
The SNF2 class of chromatin remodeling factors has traditional roles in regulating
gene expression. The founding member, Snf2p was identified in yeast through genetic
screens that identified mutants defective in mating type switching (SWItch) and sucrose
fermentation (Sucrose-Non-Fermenting), and is required for expression of genes that
regulate those processes [83,84]. The mammalian homologs of Snf2p, Brg1 and Brm,
associate with the BAF complex, and as previously mentioned, can regulate gene
activation or repression, depending on the context [71,72]. Like Brg1-containing
complexes, the NURF complex, which contains one of two mammalian ISWI homologs,
Snf2l, also regulates gene expression [85,86,87]. Mice with non-functional NURF are
embryonic lethal with major deficiencies in expression of BMP signaling genes [85].
Members of the CHD class can also regulate transcription, as evidenced by CHD7, the
chromatin remodeling factor whose dysfunction in humans is the cause of CHARGE
syndrome [88]. CHD7 co-localizes with the permissive H3K4me3 mark and interacts
with the SMAD family of transcription factors to control gene activation of cell-type
specific genes [89]. The INO80-containing complexes in yeast, Drosophila, and
mammals also control gene expression [90]. In addition, INO80 complexes have
multifunctional roles in histone deposition of the histone variant H2AZ [91] and in
8
regulating DNA repair by interacting with the phosphorylated H2AX that is immediately
localized to sites of DNA damage [92,93]. Thus, the SNF2 family of CRFs have roles in
transcriptional regulation, but can also have roles in other chromatin-dependent processes
[94].
Epigenetic Pluripotency in ES Cells
ES cells are a pluripotent cell type that can give rise to the three primary germ
layers, endoderm, mesoderm, and ectoderm [95,96]. They are derived from the
pluripotent inner cell mass (ICM) of the blastocyst, but unlike the ICM, ES cells self-
renew in culture and maintain pluripotency indefinitely [95]. The pluripotent state is
orchestrated by a network of transcription factors including POU domain transcription
factor Oct4, Sox2, and Nanog [97,98].
Chromatin in pluripotent ES cells is generally more relaxed than differentiated
counterparts, a state thought to preserve the plasticity of future gene expression programs.
Histone modifications permissive for transcription, including trimethylation of H3K4 and
acetylation of H4, are more abundant in ES cells than in differentiated cells [99]. In
addition, modifications that repress gene expression are less frequent in ES cells. ES cell
chromatin is hyperdynamic with respect to histones and architectural chromatin proteins
[100]. While core histones bind stably to chromatin in differentiated cells, they are
loosely bound in ES cells. The dynamic exchange of the linker histone H1 is required for
pluripotency, as its immobilization causes ES cell differentiation [100].
A unique chromatin landscape can be found in ES cells. Trimethylation of H3K4
is a transcriptionally permissive mark, while trimethylation of H3K27 is a repressive
mark. In ES cells and in other multipotent cells, the appearance of “bivalent” domains
that contains overlapping H3K4me3 and H3K27me3 is intriguing [101]. In ES cells,
genes that are marked with bivalent domains are either not expressed or expressed at very
low levels and are strongly correlated with developmental roles [101]. Upon
differentiation, however, only one of the marks is retained and the gene either becomes
activated or is continued to be repressed. Which mark is retained depends on which
lineage the cell has adopted. In this way, ES cells can repress genes associated with
differentiation without shutting them off permanently.
9
It has been proposed that pluripotency is the default state [102]. Transcriptionally
permissive, open chromatin allows the expression of many genes, but those of the
pluripotency transcriptional network dominantly repress differentiation genes and
establish positive feedback loops to ensure the continued expression of pluripotency
genes. The pluripotent state is further locked in by epigenetic changes in chromatin such
as histone acetylation and methylation. One of these epigenetic changes is the
demethylation of H3K9 by the histone demethylase enzymes KDM3a (previously known
at Jmjd1a) and KDM4c (Jmj2c), both of which are encoded by Oct4 target genes [103].
In line with this hypothesis, it has recently been shown that somatic cells can be
induced to become pluripotent by the introduction of just a few transcription factors,
namely Oct4, Sox2, c-myc, and Klf4 [4]. It has been nearly sixty years since the first
report demonstrating that live tadpoles could be obtained with high efficiency from
nuclear transplantation of blastula cells into enucleated amphibian oocytes [104] and
more than ten years since the birth of “Dolly the Sheep,” the first mammal cloned from
the nucleus of a fully differentiated somatic cell [3]. These studies have clearly
demonstrated that cellular differentiation is a reversible process that can be mediated by
factors present in the oocyte. The identity of these reprogramming factors remain
elusive, but studies on induced pluripotent stem cells (iPS) show that forced expression of
only a few transcription factors is sufficient to induce complete reprogramming [4]. iPS
cells are indistinguishable from ES cells in terms of their DNA methylation patterns and
ability to generate all the cell types of an adult organism [5,105,106]. Despite the
successful derivation of iPS, the frequency of obtaining truly pluripotent cells remains
low, and a large number of stochastic factors are likely involved [107].
DNA Repair in ES cells
Like the cells of the early embryo, ES cells have strict requirements to ensure that
mutations are prevented from being passed down to daughter cells and subsequent
lineages. Rapidly proliferating cell types such as ES cells must be particularly efficient at
repairing damage because unrepaired DNA lesions can lead to stalling and collapse of the
replication fork, resulting in highly deleterious double-stranded breaks (DSBs).
Emerging evidence indicates that ES cells are somewhat unusual in the way they repair
10
DNA, and this could be due to several factors including the strict need for genetic
fidelity, rapid proliferation, culture adaptations, and unique “open chromatin” structure
[108,109,110]. The “open chromatin” of ES could render them initially more vulnerable
to acquiring DNA damage, but also facilitate repair since repair proteins would have
easier access to sites of damage.
The mutation frequency of ES cells is considerably lower than in MEFs. To
measure spontaneous mutation frequencies, the Aprt locus is often used a reporter locus
because APRT deficiency can be positively selected for in culture [111]. Cells
heterozygous for APRT activity are allowed to accumulate spontaneous mutations, and
then cells in which the second copy of APRT is inactivated through spontaneous
mutation are selected for. The most frequent mutation is loss of heterozygosity (LOH,
80%), either through non-disjunction or mitotic recombination, and point mutations also
occur (20%). Experiments using this Aprt reporter system reveal that the spontaneous
mutation frequency of ES cells is 100 times lower than isogenic MEFs (10-6
in ES cells
versus 10-4
in MEFs) [112]. Interestingly, whereas MEFs sustained LOH mainly through
mitotic recombination, ES cells sustained LOH primarily through non-disjunction. When
the X-linked Hprt locus is used as a reporter, the mutation frequency in ES cells is 1000
times lower than in isogenic MEFs (<10-8
in ES cells vs. 10-5
in MEFs) [112,113]. The
difference in mutation frequencies between the two loci is probably due to the location of
Hprt on the single-copy X chromosome, making it immune to LOH events. When
treated with DNA damaging agents, accumulation of mutations at both the Aprt and Hprt
loci is dose-dependent in ES cells as well as MEFs, eliminating the argument that lower
mutation frequencies occur solely because ES cells are resistant to DNA damage
[114,115]. These data show that ES cells are indeed proficient in removing DNA
damage.
A cell can respond to DNA damage in one of three ways: (1) it can ignore the
damage and risk propagating a deleterious mutation, (2) it can repair the damage through
a variety of pathways, and (3) it can terminate itself through programmed cell death. The
latter response is chosen more readily in ES cells than in other cell types, and this can be
seen in their hypersensitivity to DNA damaging agents [116,117,118]. The propensity to
initiate apoptosis is not directly mediated by the tumor suppressor p53 because it remains
11
inactivated in the cytoplasm of ES cells, even upon insult by DNA damaging agents
[119]. Instead, the lack of a G1 arrest allows cells to slip into S phase where lesions can
be exacerbated by conversion to DSBs upon stalling and collapse of the replication fork,
and apoptosis may result as a consequence. Interestingly, ES cells are deficient in the
two pathways that regulate a functional G1/S phase checkpoint, the p53/p21-mediated
pathway and the ATM-Chk2-Cdc25A-mediated pathway [120]. In addition to non-
functional cytoplasmic p53, the Cdk inhibitors, p21 and p27 are expressed at undetectable
levels in ES cells [120]. The second pathway is rendered ineffective because of Chk2
localization to centrosomes [108]. Ectopic Chk2 expression in ES cells restores the G1/S
phase checkpoint and reduces sensitivity to DNA damaging agents. The authors suggest
that the lack of a normal G1 arrest is beneficial for ES cells and encourages damaged
cells to be removed from the population by apoptosis.
In light of the lack of G1/S phase checkpoint in ES cells, it is perhaps not
surprising that ES cells stain strongly for phosphorylation of the histone variant H2AX, a
marker of DSBs [109]. Despite the high prevalence of DSBs, ES cells are more efficient
in repairing DSBs induced by ionizing radiation than either their differentiating embryoid
body or MEF counterparts [121]. DSBs can be repaired through one of two pathways
[122,123]. One is non-homologous end joining (NHEJ) that ligates together two double-
stranded ends of DNA. This method is error-prone since any two broken ends can be
used. The second pathway is homologous recombination repair (HRR) that utilizes
homologous recombination and is error-free. Because HRR utilizes the sister chromatid
as a template, this pathway is active during late S phase [123,124,125]. A second
advantage of skipping G1/S phase checkpoint and allowing damaged cells to enter S
phase, therefore, is that error-free HRR pathway can be utilized in place of NHEJ [109].
The protracted S phase in ES cells supports this notion [126]. Therefore, the high
prevalence of γH2AX staining may not reflect a deficiency in repairing DSBs, but instead
a choice to send cells with damaged nucleotides into S phase where DSBs can be
preferentially repaired by HRR.
In one study, DSBs were introduced into ES cell lines by transfecting a plasmid
encoding the rare cutting endonuclease I-SceI, which cut two I-SceI cleavage sites
engineered into two different chromosomes [127]. Of the clones that successfully
12
repaired the DSB, 79% were repaired through conservative HRR, and 21% were
characterized by chromosomal translocation characteristic of NHEJ. The data suggest
that DSBs are preferentially repaired by HRR when only two DSBs are present. The
same authors showed that NHEJ was never used when only one DSB was introduced. A
separate group reported that two DSBs located in cis in ES cells are repaired by NHEJ,
but highly inefficiently [128]. The studies highlight a potentially compounding factor,
that the number of DSBs and where they are located could have a major impact on which
DSB pathway is used [129]. In other words, the more broken ends are available, the
more opportunities may be found for NHEJ and the more difficult it may be to align
chromosomes for homologous recombination. Several other studies support a prominent
role for HRR in ES cells [130,131,132]. While it has long been supposed that ES cells
are unique in their utilization of HRR and exclusion of NHEJ, it was only recently that
Tichy et. al. demonstrated that HRR is the preferential repair pathway in ES cells, and
upon ES cell differentiation by retinoic acid, the primary repair pathway switches to
NHEJ [133]. Similar results were simultaneously reported by Serrano et. al., but this
study did not compare ES cells with cells directly differentiated from ES cells [134].
Instead, Serrano et. al. showed that ES cells can utilize HRR prior to DNA replication,
resolving the discrepancy between inefficient mitotic recombination yet high prevalence
of HRR in ES cells [134].
Other repair pathways are active in ES cells, including mismatch repair (MMR)
and nucleotide excision repair (NER). Mutations in either of two MMR genes, Msh2 and
Mlh1 in humans are primarily responsible for the development of hereditary non-
polyposis colorectal cancer (HNPCC) [135,136]. ES cells lacking Msh2 have
spontaneous mutation frequencies 30-fold higher than wild-type ES cells [137], and Msh2
mutant ES cells are also more resistant to apoptosis upon treatment with DNA damaging
agents [138]. A similar effect was seen in ES cells null for Mlh1 [139]. Photolesions
induced by UV-C radiation are repaired by the NER pathway. The mutation rate in ES
cells treated with low doses of UV-C were comparable to that of MEFs, but at higher
doses, ES cells acquired disproportionately more mutations, indicating that the NER
pathway has a saturation level in ES cells [117]. ES cells lacking ERCC1, a major NER
component, also exhibit 10-fold higher mutation frequencies than their wild-type
13
counterparts, and at high doses of UV-C underwent massive apoptosis [117]. The
literature is strangely silent regarding the BER pathway in ES cells, but null mutations in
at least one BER member, Aag results in hypersensitivity to DNA damaging agents [140].
Thus, ES cells have lower mutation frequencies than their somatic counterparts,
undergo apoptosis more readily in response to damage, and preferentially use error-free
HRR during S phase to repair damage. Other repair pathways are also active, and these
may function similarly in ES cells than in other cell types.
Chromatin Remodeling in the Zygote and Cleavage-Stage Embryo
Chromatin states of the gametes
Primordial germ cells are set aside early during mammalian development [141].
In females, primordial germ cells enter meiosis and replicate DNA, becoming 4N in
chromosome number. These primary oocytes remain paused in meiosis I until puberty.
In response to hormonal signals, the primary oocyte completes meiosis I, extruding the
first polar body and becoming 2N in number, and is ovulated. In the oviducts, the
secondary oocyte initiates meiosis II, but remains arrested in metaphase II until
fertilization. In contrast, the sperm has completed meiosis I and II and enter the oocyte as
a haploid cell.
Methylation of imprinted genes is already established through reprogramming
during development of the germ cell lineage [142]. Chromatin in the sperm is even more
compacted than normal chromatin in somatic cells, owing to the ~10-fold reduction in
size of the sperm nucleus [143,144]. To achieve this, the sperm adopts a “frozen” state
by exchanging positively-charged protamines in place of histones and eliminating the
need for active transcription by decoupling translation from transcription [143,145,146].
Chromatin compaction through the incorporation of protamines is essential for sperm
function. Human infertility is positively correlated with low abundance of protamines,
and mice with defective protamine production have severe defects in fertility
[147,148,149,150]. Chromatin of both the sperm and the oocyte has high levels of DNA
methylation and is transcriptionally inactive at the time of fertilization [151,152].
14
Chromatin remodeling in the zygote
Upon fertilization, the oocyte completes meiosis II and extrudes the second polar
body. The cortical reaction, accompanied by a sharp increase in intracellular Ca++
levels,
induces changes in the zona pellucid that prevents polyspermy [153]. Chromatin of the
maternal and paternal genomes decondenses and becomes encapsulated in nuclear
envelopes, forming two pronuclei. The pronuclei migrate through the nucleoplasm and
ultimately fuse in a process called synkaryogamy. The zygote then undergoes its first
round of DNA replication and development progresses. Prior to pronuclear fusion,
however, large scale changes in chromatin structure occur, especially in the paternal
genome that must recover from the extreme chromatin compaction and general dearth of
proteins in the sperm (Fig. 4).
The cytoplasm of the oocyte has powerful reprogramming potential, as evidenced
by reproductive cloning through somatic cell nuclear transfer [154]. Within the first
several hours of fertilization, the sperm chromatin is massively remodeled by factors
present in the oocyte [155]. Even prior to the first round of replication, DNA methylation
rapidly decreases and protamines are evicted from chromatin and exchanged for histones
[156]. Interestingly, the incorporation of specific histones into paternal chromatin
establishes an asymmetry between the paternal and maternal genomes. For example,
paternal chromatin preferentially incorporates acetylated histones, which aids in sperm
decondensation upon fertilization [157,158]. In addition, the replication independent
histone H3 variant, H3.3 is utilized, and these are generally replaced by canonical H3
histones during the first cell cycle [159,160,161]. Monomethylation of H4K2, a
modification important for DNA repair, is asymmetrically localized to paternal chromatin
[159].
The paternal genome is rapidly demethylated within 8 hours of fertilization, even
in the presence of the replication inhibitor aphidicolin [162,163]. In contrast, the maternal
genome retains DNA methylation in the zygote and, due to the absence of Dnmt1, is
passively demethylated during subsequent cell divisions [164,165]. In plants
demethylation of 5-meC is performed directly by the glycosylases DEMETER, ROS1,
DML1 and DML2 and unmethylated cytosines are inserted into abasic sites through the
base excision repair (BER) pathway [166,167]. Despite DNA demethylation being well
15
established in plants, the identification of a parallel mechanism in mammals has been
elusive. Many groups have reported the identification of demethylase activities
[168,169] or enzymes, including the glycosylase TDG [170], the methyl-binding protein
MBD2 [171], a nuclear protein called GADD45a [172], and more recently, the de novo
methyltransferases DNMT3a and 3b [173,174], but so far, biochemical studies have been
irreproducible, refuted, or controversial in terms of mutant phenotypes [175]. The
discovery of bona fide demethylase enzymes has lead to controversy over whether they
exist in mammals at all, particularly because demethylation can occur passively during
replication in any dividing cell type, including primordial germ cells in which 5-meC on
imprinted genes is known to be removed [175]. While the extent of 5-meC
demethylation in the developing mammalian embryo and in the adult may be uncertain,
rapid, replication-independent demethylation is clear in the paternal genome, supporting
the existence of an active demethylase enzyme [163]. A recent publication by Hajkova
et. al. demonstrates that demethylation in the zygote is partially impaired in the presence
of PARP inhibitor and is co-localized with members of the BER pathway [176],
suggesting that the mechanism of demethylation may indeed be similar to that in plants.
Further studies focusing on the zygote should be rewarding. How the demethylase
targets the parental genome while allowing 5-meC to remain on the maternal genome is
also an interesting open question.
Intriguingly, the methylation of imprinted genes and of some repeat sequences is
somehow protected from the rapid demethylation of paternal chromatin and the passive
demethylation of maternal chromatin [177]. One explanation is that methylation at
imprinted genes is quickly re-established by factors in the oocyte after active
demethylation. An alternative and more likely explanation is that a counteracting
enzyme may function to protect imprinted genes, or chromatin at imprinted genes may
have additional marks that fend off the demethylase.
Maternal/zygotic transition
While the oocyte is supplied with maternally contributed transcripts and proteins
[178], these cannot sustain the embryo for long, and transcription must be reinitiated in
the zygote through a process called zygotic genome activation (ZGA, Fig. 4). Massive
16
degradation of maternal RNAs that might interfere with cleavage-stage development
immediately precedes ZGA [179]. This is a highly ordered process and regulated by the
Smaug or EDEN-BP proteins and by the microRNA mir-430 that induce deadenylation of
the 3’ UTRs of maternal transcripts and trigger nuclease degradation [180,181].
In mice ZGA occurs at the two-cell stage; in humans it occurs between the four-
cell and the eight-cell stage, and results in a whole new profile of gene expression
[152,182,183]. Little is known about how the parental genomes are silenced in the germ
cells, nor about what triggers the activation of transcription from the zygotic (ZGA).
Interestingly, some basal level of transcription can be detected from the paternal
pronucleus, suggesting that paternally derived transcripts could participate in the
initiation of ZGA [184]. It is likely that ZGA involves extensive and rapid remodeling of
chromatin. Brg1, the enzymatic subunit of the BAF complex, is required for
reprogramming of permeabilized human somatic cells using Xenopus laves oocyte
extract, because antibody depletion of Brg1 prevented reprogramming [185]. More
recently, Brg1 was shown to be required specifically for ZGA in mice. Brg1 knock-out
mice, which retain maternal Brg1 expression, exhibit peri-implantation lethality [186].
When Brg1 is specifically depleted in the oocyte, fertilized embryos arrest at the 2- to 4-
cell stage with a 30% reduction in genes involved in transcription, RNA processing, and
cell cycle regulation [187]. The authors note a reduction in the levels of dimethylation of
H3K4 (a mark associated with transcriptional activity) and propose a model in which
Brg1 affects chromatin structure to promote ZGA [187].
Chromatin Remodeling in the Blastocyst
Differentiation of the ICM
After several rounds of cell division without concomitant growth, the embryo
undergoes compaction and achieves a unique organization: there are now cells on the
inside of the embryo surrounded by cells on the outside [152]. It has long been suggested
that the inside/outside organization of cells leads to the first differentiation choice to
become either inner cell mass (ICM) or outer trophectoderm (TE) cells (Fig. 4) [188].
The outer cells are polarized, by virtue of having cellular connections to one side but not
the other, and become epithelialized [189]. In contrast, the inner cells are surrounded on
17
all sides by neighboring cells and lack polarity [190]. TE cells are required for
implantation and do not contribute to the embryo proper. Instead, they will give rise to
the extra-embryonic tissues including the placenta. All the cells of the embryo proper are
exclusively derived from the ICM, although a derivative of the ICM, the primitive
endoderm, will also contribute to extra-embryonic endoderm [191].
How the ICM and TE lineages are determined is still under investigation, and the
role of epigenetics is unclear. Initially, asymmetric cell divisions along the apical/basal
axis during early cleavage stages allocate one daughter cell to the inside of the morula
and one to the outer surface of the morula; symmetric cell divisions result in two
daughter cells with an outer-facing surface [188,192]. There has been considerable
debate regarding whether early asymmetric cell divisions in the 4- to 8-cell stage embryo
contribute to an intrinsic bias leading toward differentiation into either the ICM or TE,
but the emerging consensus seems to be that apical/basal polarity in the late morula
overrides earlier biases [193,194,195,196]. An elegant blastomere dissociation study
showed that the TE is determined prior to the ICM by the 32 cell stage [197]. Labeled
and dissociated cells of the 16-cell morula were fully competent to become either ICM or
TE, regardless of whether they had been labeled as “outer” or “inner.” In striking
contrast, aggregates of “outer” cells of a 32-cell morula formed TE but not ICM. At the
same time, “inner” cell aggregates were able to form an ICM as well as a TE layer
competent to implant into surrogate mothers, showing that inner cells of the 32-cell
morula are not yet determined to become ICM.
Transcriptional changes are known to participate in the differentiation of the TE
and ICM. The TEA-domain containing transcription factor Tead4 initiates the
differentiation of the TE, and Tead4-/-
embryos cannot form TE and arrest prior to
blastocyst formation [198]. Because of this severe phenotype, it is proposed to be the
most upstream transcription factor in the specification of the TE, but it should also be
noted that many early knock-out phenotypes are masked by maternally contributed
transcripts and proteins. Tead4 regulates the expression of the caudal-type homeobox
transcription factor Cdx2 [199]. Cdx2 promotes symmetric cell divisions and thus
regulates the allocation of cells [200]. Cdx2 and the Pou-domain transcription factor
Oct4 are co-expressed in blastomeres but mutually oppose each other as the blastocyst
18
forms [192,201,202]. Upon formation of the blastocyst, Cdx2 becomes excluded from
the ICM, and Oct4 is excluded from the TE [203,204,205,206]. How cell polarity signals
the activation of specific transcription factors remains a mystery.
Epigenetic differences between the TE and ICM
In additional to transcriptional changes, there are also epigenetic differences
between the ICM and TE. Whether epigenetic modifications participate in the
specification of the two tissues or whether they lock in differentiation choices and set up
future lineage choices is not clear. Interestingly, despite delayed determination of ICM
cells compared to TE cells [197], the ICM is characterized by much higher levels of DNA
methylation [158]. Chromatin is passively demethylated during the cleavage stage
embryo due to inactivity of Dnmt1, hitting a trough at the late morula stage; but as the
blastocyst forms, de novo methylation mediated by Dnmt3a and 3b begins, and this
occurs more prominently in the ICM than TE lineages [165]. Appropriate methylation
patterns are essential to the developing embryo, because Dnmt3a/3b null mice show
severe abnormalities and lethality at E8.5-E9.5 [27]. The transcription factor Elf5 plays a
role in the determination of TE lineage. In the epiblast and in ES cells, loss of DNA
methylation at the promoter of Elf5 alleviates its repression and induces aberrant TE
formation, suggesting that the enrichment of DNA methylation in the ICM may
participate in repression of TE lineage determination genes [198].
In addition to differential DNA methylation, the ICM also differs from TE in
trimethylation of H3K27, a transcriptionally repressive mark that is catalyzed by the
Polycomb Repressive Complex 2, PRC2 [207,208]. Genome-wide µChIP studies show
that signaling and developmentally-regulated genes in the ICM were preferentially
methylated at H3K27, suggesting that PCR2 is more active in the embryonic lineage
[208]. No differences between ICM and TE were found in methylation of H3K4, a mark
associated with gene activation [208]. Taken together with evidence of increased DNA
methylation, the theme arises that the cells of the ICM contain more transcriptional
repressive chromatin than the TE [209]. One possibility is that TE differentiation
programs are intrinsically dominant, and wide-spread transcriptional repression is
necessary to maintain ICM fate. An alternative explanation is that the generally
19
repressive chromatin marks are necessary to ensure accurate gene expression programs
during subsequent development.
X chromosome inactivation
X chromosome inactivation (XCI) is a critical process that achieves dosage
compensation in mammals [210]. XCI occurs early during development and involves the
expression of non-coding RNA Xist [211,212,213]. Upon XCI initiation, the two X
chromosomes transiently pair at the X inactivation control center and at the X pairing
region, and Xist is expressed mono-allelically from the future inactive X chromosome
(Xi) [214]. Xist binds along the length of the Xi and triggers a stream of epigenetic
events that shut off Xi, including widespread DNA methylation , methylation of H3K9
and H3K27, incorporation of the macroH2A histone variant, and hypoacetylation of
histones H2A, H2B, H3 and H4 [23,207,215,216,217,218,219]. Interestingly, the
incorporation of macroH2A has been shown to interfere with the mobilization of
nucleosomes by SWI/SNF [220]. In addition, Xi becomes localized to the nuclear
periphery, where heterochromatin is frequently relegated [216].
The traditionally held view that X chromosome inactivation is random is not so
straight forward. In fact, random X inactivation occurs specifically in the ICM, and the
paternally-inherited X (Xp) is non-randomly inactivated in trophectoderm derived tissues
[221]. During fertilization, Xp is delivered in the inactivated form but becomes active
upon sperm chromatin remodeling [222]. At the four cell stage, Xp is again inactivated
and remains so until the differentiation of the ICM [223]. In the trophectoderm,
inactivation of Xp is maintained; but in the ICM, Xp undergoes a second round of
reactivation [223]. Only after implantation does random inactivation of ICM-derived
cells occur.
How the X chromosome is reactivated is not well understood. The first and
second waves of Xi reactivation differ in at least three critical points: (1) the first round
reactivates a Xp that was inactivated during spermatogenesis by a process mechanistically
distinct from XCI called meiotic sex chromosome inactivation (MSCI), which occurs in
an Xist-independent manner [224,225]; (2) the second event is followed by random XCI,
but the first reactivation event does not remove parent-of-origin marks because it is
20
followed by imprinted XCI of Xp [226]; and (3) the second wave occurs ICM-
specifically and may depend on pluripotency factors [223].
In somatic cells, Xist expression is not required to maintain XCI, but loss of Xist
expression is accompanied by loss of macroH2A localization to the Xi [227,228]. In the
ICM and in ES cells, however, loss Xist reactivates the Xi, potentially through loss of
macroH2A localization to Xi [223]. In the zygote, MacroH2A transcripts are degraded
prior to pronuclear formation, raising the possibility that loss of macroH2A participates
in both the first and second X reactivation events [229].
Two mechanisms have emerged to explain why Xi is reactivated in the ICM
alone, the first being repression of Xist in the ICM, and the second being PcG action in
the TE. Intriguingly, Xist expression is repressed in early ICM and ES cells, and unlike
in somatic cells, Xist is required to maintain XCI in both these pluripotent cell types
[223]. Reactivation of Xi in the TE is inhibited by the PcG members Eed and Enx1.
H3K27me3, catalyzed by PcG complexes, while generally low in the TE, shows distinct
Xi localization in TE cells [207]. Eed and Enx1 proteins localize to the Xi in TE [230],
and Eed mutant mice do not retain XCI in the TE [231]. One hypothesis to explain these
elaborate mechanisms to allow random XCI in the ICM but imprinted XCI in the TE is
that repression of paternally inherited genetic material in the placenta could be beneficial
to the mother, but random expression of paternally or maternally derived genetic material
could be beneficial to the offspring.
Derivation of ES cells from the ICM
Mouse ES cells are derived from cells of the ICM and, because they can be
differentiated and cultured to large numbers, are frequently used as a surrogate system to
study the biochemistry of pluripotency and early differentiation [95,96]. Several
different procedures are used to derive ES cells that involve the propagation of ICM cells
on a gelatin-coated culture dish in defined media [232]. Some of these procedures
include whole-embryo culture, immunosurgery whereby the TE cells are selectively lysed
by the complement cascade, and physical microdissection of ICM cells. The techniques
are invariably inefficient and involve drastic and stressful changes of environment for the
21
cells of the ICM [95]. It is therefore not surprising that heated debate on the similarity
between ES cells and ICM has arisen.
In females, both X chromosomes are active at the developmental time period at
which ES cells are derived [223,233]. Consequently, both chromosomes remain active in
ES cells. In addition, parent-of-origin epigenetic marks appear to be lost, because XCI is
random in TE lineages of embryos derived from ES cells [226]. This is in contrast with
normal embryos and embryos derived from SCNT of somatic cells in which XCI in TE
lineages is non-random. The Xi of the somatic cell donor becomes reactivated normally
during ICM differentiation and in the TE is faithfully and non-randomly inactivated again
[226]. With regard to X chromosome inactivation, ES cells resemble the ICM of the pre-
implantation embryo.
Expression of transcription factors Oct4, Sox2, Nanog, and other pluripotency
factors is preserved between ICM and ES cells [234]. One of the most striking
differences between ES cells and the ICM is that the cells of the ICM are characterized
by relatively repressed chromatin while ES cells have hyperdynamic chromatin structure
and an “open chromatin” conformation [100,102,207,209]. An early study using a target-
specific µChIP approach showed that the repression of some key differentiation genes
was lessened in ES cells compared to ICM [235]. Dahl et. al. used a µChIP-chip
approach to analyze H3K4me3 and H3K27me3 distribution in ICM and in ES cells
directly derived from ICMs of the same strain [208]. Methylation patterns on both the
repressive chromatin mark H3K27 and the permissive mark H3K4 were drastically
altered between the ICM and ES cells. Only 30% of promoters with H3K4me3 and 20%
of promoters with H3K27me3 in the ICM retained those marks in ES cells. This change
in methylation patterns is not due to deficient methylation because a large number of
promoters in ES cells also gained H3K4 (40%) or H3K27 (80%) methylation.
Approximately 500 promoters in the ICM contained the so called “bivalent” domains
containing both H3K4me3 and H3K27me3 marks. In ES cells, this number nearly
doubles, but less than 10% of promoters with bivalent domains in ES cells were also
H3K27/H3K4me3-marked in the ICM. Furthermore, 50% of H3K27/H3K4me3-marked
genes were expressed in ICM, in sharp contrast with bivalent domains being mostly
22
associated with repressed genes in ES cells. These studies show that there are major
epigenetic differences between the two cell types.
Dahl et. al. console the stem cell field by showing that the epigenetic differences
between the ICM and ES cells were not as great as those between the ICM and the TE.
They suggest that during the generation of chimeric mice, ES cells are again
reprogrammed and that epigenetic changes in ES cells could explain why some ES cells
are competent to form chimeric mice and others are not. Further investigation into ES
cell epigenetics and reprogramming is clearly needed to realize the therapeutic potential
of ES cells.
DNA Repair in the Zygote
Even in the mammalian embryo that is relatively protected from exogenous
sources of DNA damage such as UV radiation, a typical cell experiences greater than ten
thousand lesions from exogenous sources [6]. These sources can include oxidation,
alkylation, and spontaneous hydrolysis that affect the chemical bonds of DNA,
threatening the accuracy and processivity of replication machinery. If mutations arise in
the early embryo and are allowed to fall into the germ line, that individual’s offspring
may inherit highly deleterious phenotypes. Even if mutations do not enter the germ line,
mutations acquired early during development can be propagated through other cellular
lineages and can result in embryonic lethality and disease.
Mutations in DNA repair machinery can result in a wide range of human postnatal
diseases including Xeroderma Pigmentosa, Hereditary Non-Polyposis Colorectal Cancer,
breast and ovarian cancer, and Ataxia-Telangiectasia [236]. Because of the late onset of
these conditions, their development is probably the result of multiple compounding
factors, including environmental variables, and is not covered under the scope of this
introductory chapter on preimplantation development.
Unexpectedly, most de novo germ line mutations in humans can be linked to
paternally-derived chromosomal anomalies and are therefore believed to have arisen from
failure to repair the parental genome [237,238,239]. DNA repair is fully functional in the
zygote, but cannot occur after the completion of spermiogenesis due to the extreme
compaction of chromatin [144,240,241,242,243]. Perhaps one of the most important
23
early reprogramming events, therefore, is the repair of the parental genome before the
onslaught of the replication machinery. In addition to remodeling DNA methylation and
the histone composition of paternal chromatin following fertilization, the zygote must
also repair the paternal genome before the onslaught of the replication machinery.
Upon nuclear condensation during spermatogenesis, DSBs are introduced into
chromatin in a histone hypoacteylation-, topoisomerase II-dependent manner
[244,245,246]. The intentional “nicking” of DNA is thought to be necessary to relieve
tension during histone withdrawal, and the majority of nicks are repaired during the final
stages of transition protein and protamine incorporation [148,247,248,249]. Repair of
double-stranded breaks is a delicate process, and defects in protamine levels are tightly
correlated with chromosomal aberrations and infertility in men [250,251]. Spermatids
retain very little ability to repair DNA after nuclear condensation is completed [252];
thus, remaining DSBs and lesions accumulated during the variable delay between
spermatogenesis and fertilization are reliant upon DNA repair in the zygotic environment
[240,241,242].
Because many mutations in components involved in the repair of DSBs result in
early lethality [253], much effort has been focused on determining the relative
contribution of error-free homologous recombination repair (HRR) versus rapid but error-
prone non-homologous end joining (NHEJ) in the embryo [123]. The duration of the cell
cycle in embryonic cells can be as rapid as 2-3 hours, resulting in the shortening of G1
and G2 gap phases [254]. The first cellular division of the zygote, however, does not
occur until 18-20 hours after fertilization, drastically extending the duration of the G1
phase and the time in which NHEJ predominates [122,123,254]. The zygote is a very
expensive and valuable cell in terms of its rarity (there is only one per embryo in contrast
to the many cells of the gastrulation-stage embryo) and its purpose (to give rise to all
cellular lineages). Unlike plentiful, rapidly dividing cells that are able to abort instead of
risk propagating mutational errors, the zygote would be very unwise to initiate apoptosis.
Therefore, rapid repair of DNA through NHEJ followed by cellular division may be more
beneficial than forcing cells to S phase where repair failures can cause cell death. To
determine which DSB repair pathways are involved in zygotic DNA repair, Derijck et. al.
compared the chromosomal abnormalities in mouse zygotes deficient in NHEJ (the skid
24
mouse, with null mutations in DNA-PKCs) or in HRR (Rad54/Rad54B double knock-out)
after irradiation. The authors found that DSBs in sperm-derived chromatin is primarily
repaired by NHEJ during G1 of the zygote, and that the majority of chromosomal
deficiencies in skid zygotes were paternal. Interestingly, markers of DNA damage
indicate that spontaneous damage occurs more frequently on paternal chromosomes. In
addition, they demonstrate that both HRR and NHEJ function during S/G2 phase.
Chd1l as an Oncogene and DNA Damage Response Protein
At the time this dissertation project began, very little was known about the
chromatin remodeling factor Chd1l, or Chromodomain, Helicase/DNA Binding Protein 1,
except what could be gleaned from genetic conservation, protein structure, and
expression studies. Even its nomenclature was misleading, as it does not contain the
chromodomain that characterizes the CHD subfamily of chromatin remodeling factors.
Chd1l is also called ALC1, or Amplified in Liver Carcinoma 1, because 1q21, a large
genomic region encompassing 12.4 million base pairs on chromosome 1, is frequently
amplified in human patients with hepatocellular carcinoma [255,256]. However, 320
other genes are found in 1q21, and it was therefore unclear whether Chd1l deserved the
synonym.
Chd1l homologs can be found in plants, mammals and most other vertebrates, but
not in the Xenopus genus. No homologs can be found in the Saccharomyces genus,
Neurospora crassa, or multicellular invertebrates including C. elegans and D.
melanogaster [81,257]. Data from the Scott laboratory published in 2004 show that
Chd1l is expressed in the preimplantation embryo, with peak expression at the morula
stage, just prior to the formation of the blastocyst [258]. Expression studies showing
genes whose transcripts are enriched in the inner cell mass (ICM) (presented in Chapter
2) identify Chd1l transcripts as showing ICM enrichment.
Chd1l contains an N-terminal, SNF2-like DNA-dependent ATPase/helicase and a
C-terminal macrodomain. Other proteins that contain the ATPase/helicase domain are
members of the SNF2 family of chromatin remodeling factors discussed previously.
These CRFs have broad roles in DNA processing, most notably transcriptional activation
and/or repression and DNA repair. Other macrodomain containing proteins bind to PAR
25
and carry out PARP processes, including transcriptional regulation, DNA repair, and
apoptosis. The presence of these domains raised the possibility that Chd1l participates in
regulating gene expression or DNA repair. This laboratory was interested in exploring
pluripotency and differentiation choices in early development, thus a role for Chd1l in
development and in gene expression was pursued in this dissertation.
Two other groups investigated the role of Chd1l in DNA repair and published
their results in 2009. Both groups confirmed that Chd1l binds to PAR through the macro
domain and has PARP-dependent, NAD+-dependent ATPase and nucleosome sliding
activities [259,260]. Mass spectrometry analysis in HEK293 cells revealed that Chd1l
associates with histones H2A and H2B and base excision repair (BER) pathway and
double-stranded break (DSB) repair pathway members, including PARP-1, XRCC1, and
APLF [259]. In addition, when treated with oxidative damage-inducing H2O2, Chd1l
becomes associated with DNA-PKcs. Association with the DNA repair proteins can be
blocked by treatment with a PARP inhibitor. Transient transfection of YFP fused to
Chd1l and laser irradiation of cells demonstrated that Chd1l rapidly localizes to sites of
DNA damage, and its localization depends on both functional macro and ATPase
domains (Fig. 5) [259,260]. Knock-down of Chd1l significantly increases the sensitivity
of U2OS cells to the DNA damaging agents H2O2 and phleomycin, suggesting that these
cells may be deficient in DNA repair. Chd1l therefore participates in the DNA damage
response, but despite these elegant experiments, a direct role in the repair of damage has
yet to be shown.
Eight years after Guan et. al. published their findings that a minimal genomic
region containing 1q21 was amplified in over 50% of patients suffering hepatocellular
carcinoma (HCC) [255], Guan’s laboratory identified ALC1, or CHD1L as the oncogene
responsible for carcinogenesis [261]. cDNA synthesized from a primary HCC tumor was
hybridized to microdissected 1q21 DNA, isolating CHD1L-encoded transcripts. Over-
expression of CHD1L increased colony formation of HCC and liver cells in soft agar
assays, and xenografts of CHD1L over-expressing cells increased teratoma formation in
nude mice. The tumor suppressor protein p53 was reduced, as was p21, a negative
regulator of CyclinE/Ckd2-mediated G1/S phase transition. Transgenic over-expression
of CHD1L in mice resulted in spontaneous tumor formation in nearly a quarter of the
26
mice, and MEFs cultures from transgenic mice displayed reduced levels of p53 [262].
Furthermore, knock-down of CHD1L using siRNA in HCC cells decreased colony
formation in soft agar assays, increased apoptosis, and decreased levels of apoptotic
proteins Caspase 3 and Bax [261]. These studies provide compelling evidence for the
role of Chd1l as an oncogene, although the causality of Chd1l over-expression in altering
cell cycle and apoptotic proteins has yet to be determined.
How can a DNA repair protein act as an oncogene? Ahel et. al. addressed this
question by over-expressing Chd1l in HEK293 cells and assaying for the γH2AX, a
marker of DSBs [259]. Over-expression of CHD1L alone did not change the intensity of
γH2AX, but when CHD1L over-expressing cells were treated with phleomycin, γH2AX
intensity increased nearly two-fold. The authors propose that over-expression of CHD1L
induces chromatin relaxation, rendering DNA susceptible to DNA damage. Interestingly,
the increase in γH2AX intensity was not seen with phleomycin treatment of cells over-
expressing a mutant CHD1L deficient in binding ATP. Although the authors did not
show chromatin relaxation, there is precedence for over-activation of oncogenes inducing
DNA damage [263]. Activation of oncogenes ras, myc, cyclin E, mos, cdc25A, and E2F1
has been observed to induce DSBs [263]. In the model proposed by Halazonetis et. al.,
precancerous tissues are associated with increases in DSBs and apoptotic indices, which
are then followed by loss of p53 function and sharp increases in proliferation rates as
cancer develops. The loss of p53 tumor suppression is frequently found in cancer and
can be due to mutations in Trp53 itself or in other genes that activate p53 [263]. Thus
cells over-expressing Chd1l accumulate DSBs and may activate p53-mediated apoptosis
followed by selection of cells with impaired p53 function. If this model is valid,
reduction of p53 and other tumor suppressor proteins and activation of cell cycle proteins
would an indirect effect of Chd1l over-expression in cells and transgenic mice.
Summary
Chromatin organization in the eukayotic cell is complex, and DNA-dependent
nuclear processes require the activity of many different chromatin remodeling factors.
The zygote, embryonic cells, and ES cells have particularly active remodeling to
reprogram chromatin to achieve the transition from two parental gametes to a single
27
embryo, to establish and/or maintain a pluripotent state, or to begin carrying out complex
differentiation as the embryo develops. Increasing our understanding of the chromatin
remodeling factors that participate in these early cell types will have broad implications
for nuclear reprogramming, DNA repair, assisted reproduction techniques (ART), and
developmental biology as a whole. The study of Chd1l, in particular, could shed light on
how transcription and/or DNA repair are controlled in the developing embryo, or could
yield novel discoveries on preimplantation reprogramming processes that are not well
understood.
28
Figure 1.1 PAR modifies nuclear proteins
Parp-1 synthesizes polymers of ADP-ribose (PAR) onto nuclear proteins. A. PAR
modifies histones and is proposed to contribute to the “histone code” in a similar way as
histone acetylation (Ac) and methylation (Me), which are recognized by bromo and
chromo domains, respectively. B. The dominant recipient of PAR is Parp-1 itself, which
catalyzes PARylation in an automodification reaction.
PARylation of histones
A B
Auto-PARylation of Parp-1
29
Figure 1.2 ATP-dependent chromatin remodeling
Chromatin remodeling factors such as Brg1 and Chd1l (orange) can remodel nucleosome
structure through the hydrolysis of ATP. Chromatin remodeling factors are often found
in association with other subunits that lend functional specificity to a complex (grey).
Chromatin remodeling factors can catalyze a number of nucleosome alterations including
nucleosome sliding and nucleosome displacement (shown). Their activity facilitates
DNA accessibility to enzymes required for transcription, replication, and DNA repair.
30
Figure 1.3 SNF2 family of chromatin remodeling factors
The four predominate SNF2 subclasses of chromatin remodeling factors, Swi/Snf, CHD,
Ino80, and ISWI, contain the core DNA-dependent ATPase domain that defines the
SNF2 family. The Swi/Snf family is characterized by bromo domain-containing proteins,
and the CHD family is characterized by PHD and chromo domain-containing proteins.
INO80 members have a unique extended split within the ATPase domain. ISWI
members contain SANT and SLIDE domains. Chd1l lacks the domains that characterize
the four major subclasses, but contains instead a macro domain not found in the other
subclasses.
31
Figure 1.4 Chromatin remodeling in the preimplantation embryo
Extensive reprogramming occurs in the zygote as the sperm decondenses. In the paternal
genome, DNA is actively demethylated, protamines are exchanged for histones, and
DNA that was damaged during the “frozen” state of the sperm is repaired. In mice, the
maternal-embryonic transition occurs at the two-cell stage, when maternal transcripts are
degraded, and transcription is initiated from the zygotic genome. Upon compaction,
blastomeres become polarized, and “outer” cells differentiate into the trophectoderm
(TE), while “inner” cells remain part of the pluripotent inner cell mass (ICM).
Differentiation into TE or ICM is accompanied by dissimilarity in the epigenetic
landscapes. For example, the ICM is marked by higher levels of repressive H3K27me3
than the TE.
32
Figure 1.5 Chd1l and the PARP-dependent DNA damage response
The Parp-1 enzyme binds to DNA strand breaks with high affinity and PARylates itself in
an automodification reaction. Chd1l and other DNA repair enzymes are recruited to sites
of DNA damage in a PARP-dependent manner. Chd1l is the only DNA damage response
protein to contain a macro domain known to bind PAR. It is believed that the DNA-
dependent and NAD+-dependent nucleosome remodeling activity of Chd1l facilitates
DNA accessibility for DNA repair enzymes.
33
2. The Chromatin Remodeling Factor Chd1l Is Required
in the Preimplantation Embryo
Alyssa C. Wright1, Denise Leong
2, Q. Tian Wang
1, 4,
Joanna Wysocka1, 3
, Mylene W. M. Yao2, and Matthew P. Scott
1, 5*
1Departments of Developmental Biology, Genetics, and Bioengineering
2Department of Obstetrics and Gynecology
3Department of Chemical and Systems Biology
4University of Illinois, Chicago
5Howard Hughes Medical Institute
Clark Center West W252, 318 Campus Drive
Stanford University School of Medicine
Stanford, California, U.S.A. 94305-5439
*Corresponding author:
650-725-7680
650-725-2952 (fax)
Keywords:
Chd1l, ALC1, preimplantation development, ICM, ES cells, chromatin remodeling
34
Abstract
Preimplantation development is marked by a multitude of chromatin changes that
allow the zygote to reprogram epigenetic marks of the parent genomes, establish
totipotency, and enact the earliest differentiation choices. To identify novel
developmental regulators, we screened for genes that are preferentially transcribed in the
pluripotent inner cell mass (ICM) of the mouse blastocyst. Genes that encode chromatin
remodeling factors were prominently represented in the ICM, including Chd1l, a member
of the Snf2 gene family. Chd1l is developmentally regulated and expressed in embryonic
stem (ES) cells, but its role in development has not been investigated. Here we show that
reducing Chd1l protein by microinjection of antisense morpholinos causes arrest prior to
the blastocyst stage. Despite this important function in vivo, Chd1l is non-essential for
ES cell survival, pluripotency, or differentiation, suggesting that Chd1l is vital for events
in embryos that are distinct from events in ES cells. Our data reveal a novel role for the
chromatin remodeling factor Chd1l in the earliest cell divisions of mammalian
development.
Introduction
The first differentiation decision in the mammalian embryo is made prior to the
blastocyst stage, when blastomeres must commit to becoming either part of the
trophectoderm (TE) or the inner cell mass (ICM). Cells of the ICM possess the property
of pluripotency and will contribute to the embryo proper, whereas the TE will give rise to
extra-embryonic material. We reasoned that factors compartmentalized in the pluripotent
ICM could be novel developmental regulators of pluripotency or early differentiation. To
identify candidate preimplantation regulators, we performed an expression analysis on
ICM separated by immunosurgery and whole blastocysts and identified genes enriched in
the ICM. Gene ontology clustering revealed a large group of chromatin regulatory
enzymes.
The high degree of chromatin organization within the nucleus is oppressive to
transcription and other processes that require DNA accessibility [12,13]. Chromatin
modifying enzymes can be divided into two broad classes, those that covalently modify
DNA and histone tails and those that utilize energy to alter nucleosome positioning. The
35
latter class is made up of chromatin remodeling factors (CRFs) that contain a core SNF2-
like ATPase/helicase domain responsible for nucleosome remodeling [81,264]. CRFs
participate in key chromatin-dependent processes including transcriptional activation and
repression, histone exchange, cell cycling, DNA repair, and many others [68,92,94,264].
CRFs assemble into multi-subunit complexes, and their functions depend in part on the
composition of the complex [18,66,265].
Profound chromatin changes take place in the zygote and preimplantation embryo
that allow the parental genomes to achieve a state of totipotency and that are necessary
for normal development. Despite successful reprogramming of somatic cells through
somatic cell nuclear transfer (SCNT) and more recently, through viral introduction of
only nuclear factors Oct4, Sox2, Klf4, and c-myc, [4,5,105,106,154], reprogramming in
the embryo remains largely enigmatic. Reprogramming is a multi-step process that
involves extensive changes in chromatin structure beginning immediately upon
fertilization. First, DNA methylation and histone modifications associated with
differentiation must be removed [209]. Next, the embryo must gain independence from
maternally provided proteins and transcripts by activating transcription from the zygotic
genome at the two-cell stage in the mouse [182,266,267]. Lastly, epigenetic
modifications erased during reprogramming must be reestablished on imprinted genes,
the inactive X chromosome (in females), and genes associated with differentiation [209].
As cell division proceeds, DNA must replicate, condense on the metaphase plate, divide,
and decondense again, all the while maintaining genomic integrity.
Relatively few factors involved in preimplantation development have been
identified because phenotypes of traditional genetic manipulations are often masked by
maternally provided transcript and proteins. Homozygous mutation of Brg1, the Snf2-
like catalytic core of the multi-subunit BAF complex, in mice causes peri-implantation
lethality [186]. The phenotype is even more severe when the maternal contribution of
Brg1 is eliminated by an oocyte-specific deletion. In this case, embryos fail to initiate
zygotic genome activation and arrest at the 2-cell stage [187]. It is likely that many more
chromatin modifiers are essential in the preimplantation embryo, so techniques aimed at
early development will likely be fruitful.
36
Among the genes identified in our screen was Chd1l, a largely unexplored CRF
that is a member of the Snf2-like family. Its compartmentalization in the ICM,
expression in ES cells, and temporal regulation prior to the blastocyst stage [258] led us
to hypothesize that Chd1l is critical for early development. The protein has a Snf2-like
ATPase domain but does not fall into any of the four major subclasses because it contains
a C-terminal “macro” domain not present in other Snf2 members and lacks signature
domains of other classes [17,257]. The macro domain binds poly(ADP-ribose), or PAR,
a post-translational modification added to nuclear acceptor proteins by the PARP family
of ADP-ribose polymerases (PARPs) [61,257,259,260]. PARPs, and by deduction the
PAR modification, have well established roles in DNA repair and transcription, among
others [33]. The nucleosome remodeling activity of Chd1l is dependent on PAR
synthesis, indicating that the PAR-binding macro domain is central to its function as a
chromatin remodeler [259,260].
Consistent with its ability to bind PAR, Chd1l is involved in the DNA damage
response. The kinetics of Chd1l localization to, and dissociation from, sites of induced
DNA damage is dependent on the ATPase and macro domains, respectively [259,260].
Recent studies have also implicated Chd1l as an oncogene. The majority of
hepatocellular carcinomas in humans are associated with genomic amplification of a
region that includes Chd1l, and its over-expression in liver cell lines and mouse models is
tumorigenic [261,262]. While evidence is accumulating for a role for Chd1l as an
oncogene and in DNA repair, its importance during development has not been addressed.
We find that Chd1l is expressed in embryonic stem cells (ES cells), which are
derived from the ICM and share the ability to differentiate into the three major germ
layers. Our data show that Chd1l is not required for ES cell viability, pluripotency or
differentiation. To determine whether Chd1l is required during earlier stages in
development, we knocked-down Chd1l in zygote-stage embryos with morpholinos (MOs)
and discovered that embryos arrest prior to the blastocyst stage.
37
Results
Chromatin factors are compartmentalized in the blastocyst
The decision to become inner cell mass (ICM) or trophectoderm (TE) is the first
lineage commitment a totipotent blastomere must make. The inner cell mass (ICM)
retains pluripotency, the ability to give rise to the three primary germ layers, whereas the
TE will give rise to extra-embryonic tissue. We reasoned that mRNAs enriched in the
ICM would encode proteins that contribute to the development of the blastocyst and/or
the establishment of pluripotency. To screen for ICM-enriched mRNAs, we purified the
ICM by immunosurgery [268], taking advantage of the structural organization of the
blastocyst (Fig. 1A). Outer TE cells of the blastocyst were labeled with IgG by
incubation with rabbit anti-mouse serum and specifically lysed by the complement
cascade, leaving behind purified ICMs. RNA extracted from the ICM was compared
with total blastocyst RNA using genome-wide expression analysis.
Transcripts encoding Oct4 and Nanog, factors known to be critical for
pluripotency, were enriched in the ICM 1.9- and 2.4-fold, respectively, providing proof
of sound methodology (Fig. 1B). In addition, mRNAs encoding Cdx2 and Eomes,
markers of extra-embryonic material, were repressed 4.5-fold and 2.4-fold, respectively,
in ICM compared to the whole blastocyst (Fig. 1B). Gene clustering revealed three major
classes of genes whose transcripts are enriched in ICM: cell signaling molecules,
transcription factors, and chromatin-modifying enzymes. Some of the chromatin factors
have known enzymatic activity and/or developmental roles (Fig. 1C). A discussion of
these notable factors follows.
Compared to the TE, chromatin of ICM cells is characterized by modifications
that are indicative of a more “transcriptionally repressive” state [209], displaying among
other repressive marks, higher levels of DNA methylation [158,198]. We find
enrichment of the mRNAs encoding the de novo DNA methyltransferases, Dnmt3a (3.7-
fold) and Dnmt3l (2.8-fold), and the maintenance DNA methyltransferase Dnmt1 (2.5-
fold) in the ICM (Fig. 1C). The higher levels of these enzymes in the ICM could explain
in part the higher DNA methylation observed in the ICM.
Chromatin of ICM cells also contains higher global levels of H3K27
trimethylation (H3K27me3), as shown by both immunostaining [207] and promoter tiling
38
arrays [208]. The repressive modification H3K27me3 is catalyzed and read by Polycomb
Repressive Complexes PRC2 and PRC1, respectively [269]. PRC1 and PRC2 are
complexes of the Polycomb Group (PcG) that function during development to establish
the body plan [270]. We observed ICM enrichment of mRNA encoding the PRC2
subunit Eed (2.9-fold, Fig. 2C), but only moderate enrichment of mRNA encoding the
lysine methyltransferase (KMT) Ezh2 (KMT6, 1.3-fold, data not shown). In ES cells, the
noncanonical KMT Ezh1can partially compensate for loss of Ezh2 [271]. Interestingly,
Ezh1 was even more enriched in the ICM (10.2-fold, Fig. 2C) than Ezh2, raising the
possibility that Ezh1 contributes to preferential H3K27me3 in the ICM.
We observed ICM enrichment for mRNAs encoding members of the Trithorax
Group (TrxG) of factors, including Mll5 (2.16-fold) and Ash1l (2.7-fold, Fig. 2C), that
modify chromatin by methylating H3K4 to increase transcriptional potential and
antagonize PRC2 function on target genes [272,273,274]. The presence of high levels of
both PcG and TrxG proteins, whose functions are opposing, may seem paradoxical; but
TrxG proteins and PcG proteins have crucial functions in pluripotent ES cells in
establishing “bivalent domains” containing both repressive H3K27me3 and permissive
H3K4me3 [101]. Upon differentiation, only one of the marks is retained, allowing for
either epigenetic repression or activation of transcription. Therefore, ICM enrichment of
both PcG and TrxG proteins suggests that the interplay between the two opposing groups
may be more prominent in the pluripotent ICM that will give rise to the developing
embryo proper than in extra-embryonic tissues.
The histone methyltransferases (KMTs) and demethylases (KDMs) are another
class of chromatin regulatory proteins whose transcripts are enriched in the ICM of the
blastocyst. We observed ICM enrichment of transcripts encoding KDMs Jarid1b
(KDM5b, 2.7-fold) and Jarid1c (KDM5c, 2.7-fold), enzymes that demethylate H3K4me
[31] (Fig. 2C). Reduction of trimethylated H3K4, a transcriptionally permissive mark
[275], is likely to contribute to the unique repressive ICM chromatin architecture.
Predictions on the methylation pattern of H3K9 are not as straight forward. ICM
enrichment of transcripts encoding two KMTs of H3K9, Suv39h1 (KMT1a, 3.5-fold) and
Suv39h2 (KMT1b, 4.0-fold), and ICM repression of two KDMs of H3K9me, KDM3a
and KDM3b, indicate that H3K9me3 would be enriched in ICM. However, transcripts
39
encoding two H3K9 demethylases, KDM4c (2.3-fold), whose expression is activated by
Oct4 in ES cells, and LSD1 (KDM1a, 1.9-fold) are enriched in ICM, provide an opposing
prediction about the status of H3K9me in the ICM.
A methyltransferase of H4K20me1, Set7, is required for development to the eight
cell stage in preimplantation embryos, with null mutants showing massive DNA damage
[276]. Transcripts encoding Set7 are enriched in ICM (7.3-fold, Fig. 2C)
The HDAC/mSin3a complex deacetylates histones and is part of the pluripotency
gene network in ES cells [277]. We observed ICM enrichment of mRNAs encoding
components of the Sin3a-HDAC1 histone deacetylase complex (Sin3a, 3.3-fold; HDAC1,
2.4-fold; Sap30 2.0-fold), as well as histone deacetylase HDAC6 (1.9-fold), but not of
HDAC2 (Fig. 2C). This finding corroborates a recent study showing that HDAC1, but
not HDAC2 controls differentiation of ES cells [278]. Enrichment of HDAC complexes
may contribute to more repressive chromatin in the ICM.
Among the mRNAs enriched in ICM were those encoding members of the Snf2
family of ATP-dependent nucleosome remodelers. These include some components of
the BAF complex, Brm (5.1-fold), Baf155 (2.1-fold) and Baf53a (5.7-fold). The BAF
complex exhibits combinatorial assembly of subunits; variations of the BAF complex are
essential at different developmental stages [69,186,187]. Transcripts encoding Snf2l, a
member of NURF chromatin remodeling complex, were enriched in the ICM (12.1-fold)
[279]. Null mutants of BPTF, the largest subunit of the NURF complex, are embryonic
lethal at E8.5, with failure to form visceral endoderm [85]. Analysis in ES cells shows
that NURF is required for proper formation of all three primary germ layers.
Chromatin remodeling factors are often found in large, multisubunit complexes
[265]. Subtle changes in the composition of a complex can have dramatic effects on its
function, and on the differentiation status of a cell [69,280,281]. Enrichment (or
repression) of one or more subunits of a complex is one way in which the composition of
a complex can be regulated [282].
In general, our data support a model in which, compared to the trophectoderm, the
ICM is characterized by a chromatin state with tight transcriptional control and an
abundance of chromatin proteins that mediate transcription and differentiation.
40
Chd1l expression patterns suggest a developmental role
Among the Snf2 family of chromatin enzymes whose mRNAs were enriched in the
ICM was the CRF called Chd1l. Its enrichment score of 4.28-fold was higher than that of
the “master regulator” of pluripotency, Oct4 (1.8-fold, Fig. 2A). The Snf2 family of
chromatin remodeling factors has powerful and diverse roles in development and
transcriptional regulation [280,283], and Chd1l is a member of this family by virtue of
the split DNA-dependent ATPase/helicase domain [81]. Chd1l is the only member of the
Snf2 family that contains a poly(ADP-ribosyl)ation binding macro domain (Fig. 2B)
[257]. As expected, Chd1l expression is observed in human and mouse ES cells (data not
shown).
Our lab previously reported genome-wide gene expression profiles over a time course
of preimplantation development from the zygote through the blastocyst stage [258]. In
these studies, Chd1l expression was found to increase through the first several cell
divisions of development, peaking at the late morula stage (Fig. 2C). Upon formation of
the blastocyst, total Chd1l expression is seen to decrease slightly; our ICM data indicate
it then becomes preferentially expressed in the ICM. Compartmentalization in the ICM,
expression in ES cells, and developmental regulation support a potential role for Chd1l in
pluripotency and during early embryogenesis. In addition to these selection criteria, we
were interested in studying a novel CRF of the powerful Snf2 superfamily (Fig. 2D).
Chd1l is dispensable for ES cell pluripotency and proliferation
Mouse ES cells are derived from the ICM of blastocyst stage embryos and
maintain the property of pluripotency indefinitely. Because Chd1l mRNA is enriched in
the ICM and abundant in ES cells, we asked whether Chd1l is essential for ES cell
survival and pluripotency. To knock-down Chd1l in ES cells, we introduced shRNA-
encoding sequence into the EBRTcH3 ES cell line [284] that allows for stable, Cre-
mediated integration and inducible transgene expression under the control of a CMV
promoter (Tet-Off, Fig. 3A). First, we created a control ES cell line, NS-shRNA
EBRTcH3, by integrating cDNA encoding shRNA that does not target any transcript in
the mouse genome (“Non-Silencing”). Transcription of the shRNA from the CMV
promoter was confirmed by observing robust Venus reporter gene expression 24 hours
41
after inducing expression by Tetracycline withdrawal (“Tet-Off” induction). We created
the Chd1l-shRNA EBRTcH3 ES cell line by integration of sequencing encoding shRNA
that targets the Chd1l transcript. Chd1l protein was reduced to nearly undetectable levels
in Chd1l-shRNA EBRTcH3 cells 48 hours after tetracycline withdrawal (Fig. 3B). In
contrast, NS-shRNA ES cells induced to express NS-shRNA had normal levels of Chd1l.
Chd1l-shRNA ES cells with reduced Chd1l had normal levels of Oct4 expression (Fig.
3B), no obvious abnormalities in ES cell morphology or colony formation (Fig. 3C), and
normal proliferation over a period of eight days, or approximately 10 doublings (Fig.
3D). Our results are consistent with a recent RNAi screen performed in ES cells in which
Chd1l was included a set of chromatin factors screened, but was not found necessary for
ES cell proliferation or for expression of a pluripotency reporter gene [285].
Chd1l does not regulate gene expression in ES cells
A primary function of the SNF2 family of DNA-dependent ATPases is
transcriptional regulation [75,286]. The four major subfamilies, SWI/SNF, CHD, ISWI,
and INO80 all regulate gene expression during development [280]. Chd1l contains a
seven -motif, DNA-dependent ATPase module that defines the SNF2 family of chromatin
remodeling factors as well as a macro domain that recognizes PAR-modified nuclear
proteins, including PAR-modified histones. We hypothesized that Chd1l might also
regulate transcription and that ES cells lacking Chd1l could have transcription changes
even in the absence of obvious morphological changes. We took a whole-genome
approach and obtained the expression profiles of induced (-Tet) EBRTcH3 ES cells
expressing Chd1l-shRNA or NS-shRNA and uninduced (+Tet) ES cells that did not
express shRNA.
Expression indices show ~70% reduction of Chd1l at the transcript level in ES
cells expressing Chd1l-shRNA (Fig. 3E). However, we found only a small number of
transcripts that changed more than 1.4-fold between induced ES cells expressing Chd1l-
shRNA and uninduced ES cells (~30), and these transcripts were also differentially
expressed between induced ES cells expressing NS-shRNA and uninduced ES cells,
indicating the expression changes were a byproduct of inducing shRNA expression. We
found no statistically significant changes in expression of pluripotency markers,
42
differentiation markers, or cell cycling genes (Fig. 3E). Although a very small amount of
remaining Chd1l could in principle be sufficient to maintain normal gene expression
programs, our data suggest that Chd1l does not regulate transcription in ES cells.
Chd1l is not required for differentiation of ES cells
Like the ICM, ES cells are capable of differentiating into the three germ layers.
While ES cells maintain this property indefinitely in vitro, the ICM is only transiently
pluripotent as cells rapidly differentiate during embryogenesis. In the absence of the
pluripotency cytokine LIF, ES cells can be grown into embryoid bodies (EBs),
differentiating cellular aggregates that mimic in vivo post-implantation development. We
reasoned that Chd1l may not regulate gene expression in pluripotent ES cells, but may do
so in differentiating cells, when new gene expression patterns are being established. To
ask whether Chd1l is required for the formation of the germ layers, we knocked down
Chd1l in Chd1l-shRNA ES cells then differentiated them into EBs and observed the
expression of a panel of lineage markers by q-rtPCR over multiple time points. For
comparison, we measured gene expression in EBs made from induced and uninduced
NS-shRNA ES cells. Quantitative rt-PCR confirmed ~70% knock-down of Chd1l
mRNA in induced Chd1l-shRNA EBs but not in induced NS-shRNA EBs over nine days
of differentiation (Fig. 3F). The panel of lineage markers included genes associated with
the establishment of endoderm (Sox17, AFP, Gata4), mesoderm (Lhx1), and ectoderm
(Fgf5, Otx2), as well as pluripotency (Oct4) and extra-embryonic (Eomes) tissues. EBs
expressing Chd1l-shRNA reduced Oct4 expression in a manner similar to EBs expressing
NT-shRNA (Fig. 3F). Expression of markers for all three germ layers was induced in a
temporally appropriate manner (Fig. 3F). Our results indicate Chd1l does not control
gene expression in pluripotent ES cells or in differentiating embryoid bodies under
normal culture conditions.
Chd1l transcripts are abrogated in MO-injected embryos
Next, we addressed the question of whether Chd1l plays a role in development
prior to differentiation of the ICM. The preimplantation embryo can be cultured in vitro
through the blastocyst and hatching stages. We took a rapid knock-down approach,
43
utilizing the synthetic antisense oligos called morpholinos (MOs) that inhibit translational
and splicing machinery. MOs have been used to effectively reduce production of specific
proteins in the preimplantation embryo, with minimal toxicity or off-target effects [287].
Splice-blocking MOs were designed to target Chd1l pre-mRNA. The predicted
splice mutants produce truncated proteins due to stop codons within the intron (Fig.4A).
Chd1l MO-1 was microinjected into the cytoplasm of one-cell stage mouse embryos
collected from superovulated and mated females. To confirm that MO-1 was functioning
as predicted, we used microfluidic q-rtPCR on RNA collected from single MO-injected
and control embryos. We used a TaqMan primer-probe assay that targeted the junction
between exons 2 and 3. This junction would be present in the wild-type Chd1l transcript
but absent if the MO blocks its targeted splicing event. As expected, the amplification of
the splice junction was nearly absent in injected embryos, showing that the transcript is
abrogated (Fig. 4B and C). A second splice-blocking MO targeting Chd1l (MO-2)
impaired a separate splicing; qPCR amplification of the splice junction showed reduction
greater than 99% compared to uninjected embryos (Fig. 4C). Using either MO, the
altered splicing would lead to the introduction of a stop codon within the intron, and
consequently the only products would be mutant proteins truncated near the N-terminus
prior to production of any functional domains.
Embryos injected with Chd1l-targeting MOs arrest prior to blastocyst stage
To ask whether Chd1l is required during early development, we microinjected the
zygote-stage embryo with MO-1 targeting Chd1l and observed embryos for a period of
four days. MO-injected embryos did not reach the blastocyst stage and instead arrested at
the compaction stage (Fig. 5A and B). An arrest prior to blastocyst formation is
consistent with the peak in Chd1l expression at the late morula stage and enrichment in
the ICM. In contrast, the majority of embryos microinjected with MO that targets
another Snf2-like chromatin remodeling factor, Snf2l, reached the blastocyst stage (Fig.
5A and B). This result demonstrates that embryonic arrest upon microinjection of a CRF
MO is not a general effect.
To further test our finding, we microinjected a second Chd1l MO that targets a
different splice junction (MO-2). These embryos also arrested prior to the blastocyst
44
stage (Fig. 5A and B). The precise timing of the arrest varied between the MOs, perhaps
due to different binding affinities of the MO sequences.
Chd1l phenotype is partially rescued by co-injection of Chd1l mRNA
To confirm that the embryonic arrest phenotype is a result of disrupting Chd1l
protein production, mRNA encoding Chd1l was co-injected along with Chd1l MO. We
reasoned that embryos arrested at an earlier stage would be more able to progress to later
developmental stages with addition of mRNA than embryos arrested at later stages, so we
used MO-2 for co-injection. MO-2 targets a splicing junction and therefore will not
target the mRNA synthesized from cDNA lacking intronic sequence. Figure 5C shows
the average result from three independent rescue experiments. Embryos injected with
MO-2 alone did not progress to the multi-cell stage, nor did embryos co-injected with
MO-2 plus GFP mRNA. As many as 50% of the embryos co-injected with MO plus
Chd1l mRNA progressed to the multi-cell stage or further. Mitigation of the
developmental arrest phenotype by Chd1l mRNA confirms that loss of Chd1l is
responsible for the embryonic arrest.
Discussion
A previous study attempted to identify pluripotency factors by comparing
expression profiles of cultured ES cells and trophoblast stem (TS) cells, derived from the
ICM and TE, respectively [288]. Our approach has the significant advantage of using
true embryonic cells that retain in vivo gene expression programs. Many genetic studies
in the preimplantation embryo focus on genes known a priori to be essential for ES cell
viability or pluripotency. We used gene expression data from early embryos to select
candidate regulators.
Chromatin remodeling activities are abundant in preimplantation embryos and in
ES cells, and many of these activities are geared toward initiating pluripotent
transcriptional competence and ensuring differentiation programs are locked in
epigenetically [209,234]. Although Chd1l is part of the Snf2 family of DNA-dependent
ATPases [81], many of which are potent transcriptional regulators, Chd1l itself does not
seem to regulate gene expression, at least in ES cells.
45
Chd1l is distinct from the other members of its family because of the presence of
a C-terminal macro module that binds poly ADP-ribose, or PAR [61,259,260]. PAR is a
posttranslational modification synthesized by the PARP family of PAR polymerases and
has important roles in diverse chromatin-dependent processes. Evidence is accumulating
for the importance of PAR regulation in the embryo. Double knock-out of PAR
polymerases Parp-1 and the partially redundant Parp-2 in mice is embryonic lethal at the
onset of gastrulation [57], whereas knock-out of the PAR depolymerase PARG is lethal at
E3.5 [289]. These data suggest that PAR levels are tightly regulated in the embryo, and
that fluctuations are highly deleterious. Chd1l contains a module responsible for binding
PAR. It is interesting to speculate that Chd1l contributes to PAR regulation and that the
Chd1l embryonic arrest phenotype is due to aberrant PAR levels or PAR signaling.
One of the most prominent trigger of PAR modification is DNA damage. Parp-1
is activated by DNA damage and synthesizes PAR onto itself in an automodification
reaction [290,291]. Blocking Parp-1 activity with specific inhibitors or through null
mutations results in cellular hypersensitivity to DNA damaging agents and defects in
DNA repair [292]. Two independent groups recently demonstrated the ability of Chd1l
to respond to DNA damage through its association with PAR [259,260]. The early
embryo has unique and stringent requirements to repair any errors to maintain genomic
integrity for the future organism. Damage to DNA occurs as a result of normal cellular
metabolism and during DNA replication. Genes involved in all of the major DNA repair
pathways are expressed in the preimplantation embryo [253], and a large number of
double-stranded break repair proteins are embryonic lethal when deleted [293].
In particular, DNA repair within the zygote is especially crucial for the paternal
genome that must recover from the “frozen” state of the sperm [240,294]. Defects in
repairing paternal DNA are thought to be a major cause of chromosomal aberrations and
human infertility. Double-stranded breaks are the most toxic form of DNA damage and
can be repaired through either non-homologous end-joining (NHEJ) or homologous
recombination (HR). NHEJ can function throughout the cell cycle, whereas HR is
restricted to S/G2 phase [295]. The zygote spends ~20 hours in G1 prior to the first cell
division, and much of the paternal DNA is repaired through NHEJ [296,297,298,299].
46
The ability of Chd1l to function in NHEJ is suggested by its PARP-dependent association
with a major NHEJ component, DNA-PKcs, upon induced DNA damage [259].
An intriguing question is why is Chd1l essential in the earliest stages of
embryogenesis but not in ES cells? In contrast to the zygote, ES cells have rapid cell
cycles with abbreviated G1 and G2 gap phases and rely heavily on HR to repair lesions
during S phase [126]. Therefore, one explanation for why reduced Chd1l causes
preimplantation arrest but is dispensable for ES cells is that Chd1l plays a role in NHEJ,
and NHEJ is essential during early embryogenesis but not in ES cells.
PAR modifications are involved in the regulation of transcription as well as DNA
repair and other processes [300]. While we cannot rule out a role for Chd1l in
transcriptional regulation in the embryo through its interaction with PAR, given our data
in ES cells showing that Chd1l is not involved in regulation of gene expression, we
suggest that the embryonic arrest due to reduction of Chd1l is more likely to be due to a
defect in DNA repair.
In summary, chromatin remodeling enzymes are active in the ICM of the
blastocyst that will differentiate into all the cellular lineages of the adult organism. One
candidate regulator of pluripotency, Chd1l, turned out to be required even earlier than the
formation of the ICM. Despite its essential role in the preimplantation embryo and
expression in ES cells, Chd1l is dispensable for ES cell viability, pluripotency, or
differentiation. Chd1l contains a macro domain that binds to PAR, and the Chd1l arrest
phenotype could be due to impaired PAR signaling. Recent studies have demonstrated a
role for Chd1l as a DNA damage response protein that interacts with members of the
NHEJ pathway in a PARP-dependent manner. A role for Chd1l in NHEJ could explain
why Chd1l deficiencies in the preimplantation embryo result in an early arrest phenotype
whereas deficiencies in ES cells result in no detectable abnormalities.
Methods
Immunosurgery and expression profiling
E3.5 blastocysts were collected from timed pregnant mothers and washed in M2
medium. The zona pellucida was removed by incubation in Acid Tyrode solution for 3
minutes. Outer TE cells were labeled with IgG’s by incubation with 10% rabbit α-
47
mouse serum for 60 minutes. Embryos were washed three times in M2 medium, and then
TE cells were lysed through the complement cascade by incubation with 30% guinea pig
complement for 15-30 minutes, or until lysis was visible. Remaining ICMs were washed
three times in M2 medium with a fine pipette to remove residual TE cells. Total RNA
was extracted from purified ICMs and whole blastocysts with Trizol reagent. Purified
RNA was amplified, labeled, and hybridized to Affymetrix mouse 430 2.0 Expression
Arrays.
ES cell lines
The EBRTcH3 cell line contains a cassette acceptor utilizing loxP and loxPV sites at
the Rosa locus to allow efficient and directional integration of a transgene by Cre-
mediated recombination. ShRNA-mir cDNAs were subcloned from pGIPZ vectors
(OpenBiosystems, Chd1l shRNA Oligo ID: V2LMM_18041 and “non-silencing”
shRNA-mir) into the pPthC exchange vector for recombination into the EBRTcH3 ES
cell line. The parental EBRTcH3 ES cells and the pPthC exchange vector were gifts
from the lab of Dr. Hitoshi Niwa of Japan.
The exchange vector containing the shRNA-mir sequence was cotransfected along
with a Cre expression plasmid by lipofectamine. Transfected cells were plated single-cell
density and cultured in the presence of Puromycin (1.5 µg/ml) to select for successful
integrants and of Tetracycline (1.0 µg/ml) to repress transgene expression. Clones were
confirmed by PCR genotyping of the 5’ and 3’ recombination sites. To induce shRNA
expression, the derived ES cell lines were cultured in the absence of Tetracycline and
high Puromycin (7.5 µg/ml). Control, uninduced ES cells were cultured in high
Tetracycline (1.5 µg/ml) and high Puromycin (7.5 µg/ml).
ES cell expression profiling
Total RNA was extracted from Chd1l-shRNA and NS-shRNA ES cells three days
after inducing the expression of Chd1l-shRNA or NS-shRNA by Tetracycline removal
(7.5 µg/ml Puromycin), and from uninduced Chd1l-shRNA and NS-shRNA ES cells that
do not express shRNA (7.5 µg/ml Puromycin, 1.5 µg/ml Tetracycline). Three different
Chd1l-shRNA EBRTcH3 clones and one NS-shRNA EBRTcH3 clone were used. RNA
48
was amplified from the eight samples and labeled using Affymetrix kit and hybridized to
mouse 430 2.0 Expression Arrays. Fold changes in expression indices were calculated
for shRNA-induced ES cells vs. shRNA-uninduced ES cells. Statistical significance of
fold-changes between Chd1-shRNA induced and uninduced samples was carried out
using a paired t-test, a minimum fold change of 1.4, and a delta value of 1.9 (SAM
Analysis [301]).
Differentiation of embryoid bodies
Expression of shRNA was induced by Tetracycline withdrawal in Chd1l-shRNA
and NS-shRNA EBRTcH3 ES cells for three days prior to differentiation into embryoid
bodies (EBs) to ensure complete Chd1l knock-down. RNA was collected at “Day 0” of
differentiation from induced and uninduced Chd1l-shRNA and NS-shRNA EBRTcH3 ES
cells. ES cells were suspended at a density of 2x104 cells/ml of media without LIF, and
EBs were made using hanging droplets of 500 cells in 25 µl. After two days, embryoid
bodies were collected into 10-cm Ultralow Attachment plates (Corning) and cultured for
an additional seven days in the absence of LIF. RNA was collected every three days after
LIF removal. cDNA was synthesized from each sample and subjected to qPCR. Relative
quantities for each cell line were calculated using Gapdh as the internal control and
shRNA-uninduced, “Day 0” samples as references.
Embryo Culture and microinjection
Three to five week old wild-type F1 (C57BL6xDBA/2) females (Charles Rivers)
were superovulated by intraperitonial injections of 5 IU of pregnant mare’s serum
gonadotropin (Sigma) followed by 5 IU of human chorion gonadotropin (Sigma) 48
hours later and mated with wild-type males. Mice were sacrificed by cervical dislocation
17 hours after hCG injection, and 1-cell embryos were dissected and released from
oviducts. Cumulus cells were removed hyaluronidase digestion, and embryos at the two
pronuclei stage were recovered and immediately micro-injected cytoplasmicly with 5-10
pL of 0.6 mM antisense morpholino. Prior to injection, the MO was heated at 65° for 15
minutes to remove any secondary structure.
49
Preimplantation embryos were cultured in vitro in 20 µl droplets of Quinn’s
Advantage Cleavage Medium (Sage) supplemented with 10% SPS serum and covered
with mineral oil. Dishes were placed in a dessicator filled with mixed gas (90% nitrogen,
5% oxygen, 5% carbon dioxide) in a 37° incubator. Embryos were observed every 24
hours for a period of four days.
Morpholinos were obtained from GeneTools. Chd1l MO-1:
tcattccacagcagatacCTGGCAG (in2-EX2). Chd1l MO-2: ttggagagaagcagagggctaCCTC
(in4-EX4). Snf2l: tgctgtttaccaccttacCAAGGGC (in2-EX2).
Microfluidic qPCR
Single embryos were collected 48 hours after injection and lysed by one freeze
thaw cycle. cDNA was synthesized using the CellsDirect One-Step rtPCR kit
(Invitrogen) and subjected to 18 rounds of gene-specific amplification using TaqMan
primer/probe assays (Applied Biosystems). TaqMan primer/probe assays and cDNA
from single embryos were loaded onto a Fluidigm 48.48 microfluidic array for qPCR
analysis using the Biomark thermalcycler.
α-Chd1l antibody
A hydrophilic sequence of 122 aa corresponding to amino acids #557-678 of the
Chd1l protein was selected for the antigenic region. The antigen was produced as a TrpE
fusion protein from the pATH11 vector in BL21 E. coli, solubilized, and subjected to
SDS-PAGE. The gel slice was excised and submitted to Josman, LLC for injection into
rabbits. The antiserum was affinity purified over a GST-Chd1l-bound Sepharose column
and eluted with low pH buffer.
50
Figure 2.1 Chromatin remodeling factors are enriched in the ICM
A. Schematic of immunosurgery followed by whole-genome expression analysis. B.
Compartmentalization of known ICM and trophectoderm factors.
C. Enrichment of selected classes of chromatin factors in the ICM. TF: transcription
factor. DNMT: DNA methyltransferase. PcG: Polycomb group. ETP: Enhancers of
Trithorax and Polycomb. TrxG: Trithorax group. KMT: lysine (histone)
methyltransferase. KDM: lysine (histone) methyltransferase. HDAC: histone
deacetylase.
A B
C
51
Figure 2.2. Chd1l is a candidate developmental regulator
A. Chd1l is enriched in the ICM compared to the whole blastocyst. B. Chd1l is a
SNF2 chromatin remodeling enzyme containing a split ATPase/helicase and a macro
domain. C. Chd1l expression during pre-implantation development [258]. Chd1l
expression peaks at the late morula stage before it becomes compartmentalized in the
inner cell mass of the blastocyst. D. Decision tree for choosing Chd1l.
A B
C
ICM Enrichment
Chromatin Remodeling
Expressed in ESCs
Dynamic PreimplantationExpression
Unknown Developmental Role
Chd1l
D
900 aa
52
0
50
100
150
200
250
300
350
Day 0 Day 3 Day 6 Day 9
Lhx1
0
10
20
30
40
50
Day 0 Day 3 Day 6 Day 9
Gata4
0
5
10
15
20
Day 0 Day 3 Day 6 Day 9
Otx2
0
20
40
60
80
100
Day 0 Day 3 Day 6 Day 9
Sox17
0
1
2
3
4
5
6
Day 0 Day 3 Day 6 Day 9
Eomes
0
500
1000
1500
2000
2500
Day 0 Day 3 Day 6 Day 9
Fgf5
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Day 0 Day 3 Day 6 Day 9
Oct4
0
500
1000
1500
2000
2500
Day 0 Day 3 Day 6 Day 9
AFP
0
0.5
1
1.5
Day 0 Day 3 Day 6 Day 9
Chd1l
A
E
F
C
B
D
-Tet
+Tet
53
Figure 2.3. Chd1l is non-essential in ES cells
A. Strategy for knocking-down Chd1l in EBRTcH3 ES cells. The tetracycline
transactivator (tTA) is expressed from the endogenous Rosa26 locus and is bound in the
inactive form in the presence of tetracycline. In the absence of tetracycline, the tTA
activates the CMV promoter and induces expression of the shRNA-IRES-Venus
transcript. B. Efficiency of knocking down Chd1l in uninduced (+Tet) and induced (-
Tet) Chd1l-shRNA EBRTcH3 cells. Chd1l protein levels are nearly undetectable in
shRNA-expressing cells 48 hours after Tetracycline withdrawal. Oct4 levels do not
change upon knock-down of Chd1l. C. Colony morphology of Chd1l-shRNA EBRTcH3
cells with induced (+Tet) or uninduced (+Tet) shRNA expression. Expression of shRNA
was induced 24 hours prior to plating cells at clonal density and allowing colonies to
grow for six days. D. Proliferation curve of Chd1l-shRNA ES cells expressing Chd1l-
shRNA (-Tet) or uninduced (+Tet). E. Expression changes of selected genes from
global expression analysis from ES cells expressing Chd1l-shRNA or NS-shRNA. F.
Expression of lineage markers over embryoid body (EB) differentiation. All values are
normalized to uninduced “Day 0” samples from each respective cell line. Similar to EBs
expressing NS-shRNA, EBs expressing Chd1l-shRNA are able to differentiate and form
the three germ layers as evidenced by markers for pluripotency (Oct4), extra-embryonic
(Eomes), endoderm (Gata4, AFP, Sox17), mesoderm (Lhx1), and ectoderm (Fgf5, Otx2).
54
Figure 2.4. Chd1l MO knock-down
A. Mechanism of splice-blocking morpholinos (MOs). Two splice-blocking MOs were
designed targeting exon-intron boundaries. Both are predicted to produce mutant
proteins truncated prior to the ATPase domain, thus lacking any functional activity. B.,
C. Validation of Chd1l MO activity. Heat map (A) and quantitation (B) of microfluidic
qPCR of Chd1l transcripts. Ct values for each PCR reaction were subtracted from an
arbitrary value of 40 to reflect a positive correlation with expression levels. The “Ex20-
21” probe targets the 3’ end of the Chd1l transcript and will amplify all Chd1l transcripts,
regardless of splicing aberrations. The “Ex2-3” and “Ex4-5” probes target exon-exon
junctions and will only amplify if that splicing event occurs. MO-1 disrupts exon2-exon3
splicing. MO-2 disrupts exon4-exon5 splicing.
C
B
. . .
Predicted Mutant Protein Pre-mRNA
A
55
B C
A
A
56
Figure 2.5. Chd1l embryonic arrest phenotype
A. Uninjected embryos and embryos injected with indicated MOs at 4 days after
microinjection. B. Quantification of development to blastocyst stage in uninjected
embryos and embryos injected with different MOs. C. Partial rescue of Chd1l arrest
phenotype with co-injection of Chd1l mRNA.
57
3. Finding Direct Transcriptional Targets of Chd1l in ES Cells
Introduction
Specification of a totipotent or multipotent cell into more differentiated daughter
cells is associated by differences in gene expression and epigenetic alterations in
chromatin. At the inception of this thesis project, I was intrigued by the earliest
differentiation choice a cell makes: to become trophectoderm (extra-embryonic) or inner
cell mass (embryonic); and later, to depart from the pluripotent state and become one of
the three primary germ layers: endoderm, mesoderm, or ectoderm. There is debate within
the field regarding the primary event that drives differentiation, whether it is the genetic,
transcriptional activity of transcription factors that activates downstream genes including
histone modifying proteins, or the epigenetic activity of enzymes that post-translationally
modify histones and subsequently activate or repress the expression of genes, including
transcription factors [102]. Regardless of the victor in the circular “chicken-or-the-egg”
conundrum, it is clear that feedback loops exist and that both types of regulation are
crucial in establishing and maintaining gene networks.
A large body of literature has established that control of gene expression through
modification of nucleosome dynamics is a powerful way in which chromatin remodeling
enzymes mediate differentiation. The first chromatin remodeling factor shown to
mediate gene expression was the yeast Swi/Snf that was identified in genetic screens for
mutants of mating-type switching (SWI) and sucrose fermentation (Sucrose-Non-
Fermenting) [83,84]. Since then, a number of proteins containing a conserved SNF2-like
DNA-dependent helicase/ATPase have been discovered in mammals that also have
powerful functions in regulating gene expression during development [66,280].
Although the SNF2-like domains show similarity to the prototypical helicase domain,
they do not actually possess helicase activity [75,76]. Rather, this class of proteins
functions by interacting with the minor groove of the DNA double helix and inducing
torsional forces that cause strand distortion and interfere with the DNA-histone interface
58
[94]. SNF2-like chromatin factors harnesses the energy of ATP hydrolysis to force
movement of histone cores along DNA, exposing regions of DNA to a variety of nuclear
factors [76]. Thus, ATP-dependent chromatin remodeling factors counteract compacted
chromatin and can drive changes in gene expression by providing either transcription
factors or histone modifying enzymes (or both) access to chromatin.
Chd1l is developmentally regulated during in the preimplantation embryo, is
preferentially expressed in the inner cell mass of the blastocyst, and is expressed in ES
cells. As such, it became the candidate developmental regulator studied in this
dissertation. Because of homology to the SNF2-like family of chromatin remodeling
enzymes, I hypothesized that Chd1l is also a chromatin remodeling enzyme that regulates
gene expression. The goal of this project was to find direct targets of Chd1l through
chromatin immunoprecipitation followed by unbiased, whole-genome sequencing (ChIP-
seq). To determine functional relevance, I endeavored to find Chd1l-bound genes whose
expression depends on Chd1l. In parallel with ChIP-seq analysis, I performed whole-
genome expression analysis on ES cells that had normal or reduced levels of Chd1l
protein (see Chapter 2). My approach was to use the intersection of ChIP-seq data and
expression analysis to identify genomically bound and transcriptionally responsive gene
targets of Chd1l.
The purpose of this project was two-fold. The first was to identify a role for
Chd1l in ES cells, despite the lack of an obvious morphological phenotype in culture.
Gene ontology clustering of transcriptional targets can reveal roles in various cellular
processes. Enrichment of target genes involved in one process would implicate a role for
Chd1l in that process. The second purpose was to use the list of target genes in ES cells
to identify a role for Chd1l in the preimplantation embryo, where Chd1l is essential.
Because of the requirement for large amounts of starting material for ChIP-seq
experiments, procuring for de novo targets on a genome-wide scale from preimplantation
embryos that contain at most several hundred cells was not possible. The list of Chd1l-
bound genes in ES cells (regardless of expression changes in ES cells upon Chd1l knock-
down) would be scanned for genes having crucial roles in the preimplantation embryo.
These genes would then serve as candidates to be tested in early embryos using standard
PCR-based ChIP experiments that require much less input material.
59
A major hurdle of this project was the absence of previously identified Chd1l
target genes to confirm that my in-house generated Chd1l antibody is capable of
immunoprecipitating cross-linked chromatin. Indeed, the results of this experiment
yielded very few target genes, and none of the candidates selected could be validated by
q-PCR. It is difficult to differentiate between the explanations that the antibody did not
work for ChIP or that Chd1l truly does not bind distinct loci in ES cells. This appendix
documents the experimental procedures and analysis undertaken in the attempt to find
Chd1l gene targets.
Results
Chromatin Immunoprecipitation
To confirm that the lab-generated α-Chd1l antibody had specific recognition for
Chd1l, ES cell extracts with and without knock-down of Chd1l were subjected to SDS-
PAGE and blotted with α-Chd1l (Fig. 1A). A strong band at the predicted size of ~100
kD was apparent in ES cells with Chd1l, and that band became nearly undetectable in ES
cells expressing Chd1l-shRNA. Production of other genes including β-tubulin and Oct4
did not change. The ability of the α-Chd1l antibody to efficiently immunodeplete Chd1l
from supernatants was confirmed in ES cell extracts (Fig. 1B). Association of Chd1l
with chromatin was confirmed by cell fractionation assays (Fig. 1C).
Two ChIP experiments were conducted in parallel: one using α-Chd1l antibody
and a second using α-RNA PolII antibody. The RNA PolII ChIP was used for quality
control, since this antibody has been extensively used in ChIP experiments and PolII has
many known targets. A negative control ChIP was avoided because sequencing uses
total, genomic chromatin as a reference, and a negative control antibody would pull down
and amplify a very small fraction of the genome.
ES cells were cultured to subconfluence, crosslinked, and lysed. Chromatin was
sheared to small fragment sizes that are ideal for sequencing (between 100 and 500 base
pairs) and then immunoprecipitated by either α-Chd1l or α-PolII. The
immunoprecipitated DNA was purified and quantified. The amount of purified DNA for
PolII ChIP was approximately 4 times greater than for Chd1l ChIP (1.14 µg vs 0.25 µg),
indicating that in general, Chd1l binds fewer targets or binds targets with lower affinity
60
than PolII. Given that PolII is necessary for general transcription, this was not overly
concerning.
The immunoprecipitated and total chromatin DNA libraries were prepared for
sequencing by adapter ligation and amplification. DNA size, purity, and concentration
were acceptable for all samples. Prior to submitting samples for sequencing, the PolII
library was validated by PCR amplification of known targets (Fig. 1D). Confirmation of
pull down of PolII targets shows that protocol procedures were enacted accurately. The
Chd1l library could not be validated in the absence of previously identified targets.
Sequencing and Analysis
DNA libraries for PolII ChIP, Chd1l ChIP, and total genomic chromatin were
submitted to the laboratory of Dr. Arend Sidow for Solexa sequencing. Raw reads were
passed through a quality filter that excludes short reads or reads with ambiguous base
calling and mapped to the mouse genome. After filtering out repetitive reads,
approximately 4 million quality reads were obtained for each sample (Table 1). For
Solexa sequencing, which generates reads of about 25 base pairs, this corresponds to
roughly 3% coverage of the mouse genome, assuming uniform distribution. The
distribution of reads around the transcription start sites (TSSs) was measured (Fig. 2).
This analysis further validates PolII ChIP library, as PolII is known to bind at TSSs. The
Chd1l ChIP reads show a modest enrichment around TSSs, but only slightly above that
seen for total chromatin reads. In general, chromatin shears more easily at TSSs,
therefore some accumulation of reads there is expected.
Non-repetitive reads mappable to the mouse genome were delivered to the
laboratory of Dr. Wing Wong who used CisGenome for peak calling software and
genome browser [302]. The software analyzes forward and reverse reads separately to
form bi-horned normal distributions around factor binding sites. A peak, or “hit,” is
called between the modes of each distribution of forward and reverse reads (e.g. Fig. 3A).
Total genomic chromatin was used for background correction.
The number of peaks called for PolII ChIP and Chd1l ChIP for three different
false-discovery rates (FDRs) is shown in Table 2. At a FDR of .05, ~200 peaks were
called for Chd1l compared to almost 18,000 peaks for PolII. The robust number of peaks
61
for PolII and the location of peaks at actively transcribed genes show that the procedures
followed for ChIP, sequencing, and analysis were sound. The number of Chd1l peaks
was surprisingly low, and resembled ChIP-seq data for antibodies that do not ChIP well
(personal communication). Approximately 30-40% of Chd1l peaks were associated with
nearby PolII peaks (Fig. 4D). This association can reveal potential interaction of a
transcription factor or chromatin remodeling factor with PolII, an interaction that
suggests a role in gene expression. The caveat is that regions of active transcription and
PolII association are also more easily sheared, and sheared ends will result in an
accumulation of sequencing reads and peak artifacts.
Twenty-seven hits were manually selected for validation by quantitative PCR.
Many of the highest-ranking hits were excluded because of excessively high signal in the
background (e.g. Fig. 4B). Only those hits with minimal background were selected (e. g.
Fig. 4C). While peaks for PolII ChIP had maximum values around 20, maximum peak
values for Chd1l peaks tended to be small, on the order of 4 or 5. A subset of the 27
regions were associated with nearby PolII peaks, and qPCR of these regions from PolII
ChIP samples showed enrichment as expected (Fig. 4D). Regions where no PolII peaks
were called did not show PolII enrichment by qPCR. In contrast, none of the Chd1l peak
regions selected showed enrichment in Chd1l ChIP samples. These results demonstrate
that gene targets for PolII, but not for Chd1l, were successfully identified by ChIP-seq.
Re-analysis using deeper sequencing
In order to get deeper sequencing and peaks of higher confidence, the Chd1l ChIP
library was re-sequenced. A similar number of reads was obtained for the second run as
for the first. The second set of reads was combined with the first set of reads for analysis
by CisGenome. The combined analysis yielded about 200 peaks at a FDR of 0.01, a
number greater than that obtained either individual analysis (Table 3). Surprisingly, no
peaks called were common between the first and the second set, and the peaks called for
the combined analysis were largely distinct from either individual set (Fig. 4A). In
addition, the appearance of the peaks called for the combined analysis was similar to the
peaks called for the first analysis, with a peak height of about 4 or 5 (Fig. 4B). This
pattern is representative of reads randomly distributed, or enriched at easily sheared loci
62
such as TSSs, or enriched by amplification bias. It was decided at this point to
discontinue Chd1l ChIP-seq experiments.
HA-tagged Chd1l targeting vector
At the inception of Chd1l ChIP-seq experiments, it was anticipated that a second
set of data generated from the use of a separate antibody would be required to increase
the confidence of Chd1l target genes. To address this, a targeting vector was created that
would introduce an HA tag onto the 3’ end of Chd1l (Fig. 5). An antibody against HA
could then be used for ChIP-seq in ES cells that expressed tagged Chd1l at endogenous
levels. The data generated from α-Chd1l ChIP and α-HA ChIP would then be compared,
and the gene targets common to both sets would be selected for further analysis. The
generation of the targeting vector was done in parallel with the α-Chd1l ChIP-seq and
expression analysis studies; recombination into ES cells was not pursued when it became
apparent that there were no changes in gene expression upon Chd1l knock-down and no
high-confidence gene targets from α-Chd1l ChIP-seq.
Discussion
ChIP with an α-Chd1l antibody followed by whole-genome Solexa sequencing
and analysis yielded a small list of 200 peaks with a FDR of 0.10. Quantitative PCR
amplification of 27 genomic regions failed to validate any candidate peak. Two
explanations cannot be resolved: a) the α-Chd1l antibody does not work in ChIP
experiments, and b) Chd1l does not bind to distinct chromatin targets in ES cells.
However, given that there were no significant changes in gene expression upon reduction
of Chd1l levels in ES cells, the latter explanation is likely. For this reason, ChIP-seq
from stably transfected, HA-tagged Chd1l from ES cells using an α-HA antibody was no
longer pursued.
After completion of this project, Ahel et. al. reported that treatment of 293T cells
with hydrogen peroxide, a DNA damaging agent, caused mobilization of Chd1l to
chromatin [259]. In addition, the composition of the Chd1l-containing complex is altered
upon hydrogen peroxide treatment [259,260], indicating recruitment of Chd1l to
chromatin is different between treated and untreated cells. Cell fractionation assays
63
conducted in this study show Chd1l in both nucleosol and in the chromatin-bound
fraction. It is plausible that treatment with hydrogen peroxide treatment of ES cells
would induce Chd1l to become more tightly bound to chromatin and that direct targets
could then be found by ChIP-seq. This avenue of study was not pursued because of time
constraints and the lack of evidence that Chd1l contributes to DNA repair in ES cells
(See Chapter 3).
Recently, Chen et. al. reported the identification of putative CHD1L target genes
in a human hepatocellular carcinoma (HCC) cell line, one of which was associated with
gene expression changes upon modulation of CHDL levels in HCC cells. Their approach
differed from the one described here in several regards. GFP-tagged CHD1L was
transfected and over-expressed in HCC cells and CHD1L was immunoprecipitated with a
GFP antibody. DNA immunoprecipitates were cloned into a vector and identified by
sequencing. The finding that CHD1L binds target genes in this study is unconvincing for
several reasons. The over-expression of tagged CHD1L is likely to promote non-specific
binding to chromatin. The cloning-based strategy employed by the authors does not
account for sticky, non-specific binding, either between over-expressed CHD1L and
chromatin, or between GFP antibody chromatin DNA. PCR validation of targets was not
quantitative or semi-quantitative, and the functional relevance of CHD1L binding to
targets (e.g. through reporter assays) was not attempted. Therefore, although the authors
report the identification of CHD1L gene targets, severe experimental flaws bring this
finding into doubt.
The inability to find direct targets in ES cells does not preclude a transcriptional
role for Chd1l in the preimplantation embryo, where a strong arrest phenotype results
from injection with antisense morpholinos targeting Chd1l. No proliferation,
morphological, or gene expression aberrations were seen in ES cells with reduced Chd1l.
Chromatin remodeling factors are known to bind to distinct sets of targets in different cell
types, and it remains possible that Chd1l binds distinct targets in early embryos but not in
ES cells. However, until genome-wide, micro-scale ChIP experiments become more
practical, discovering targets de novo in the early embryo will remain challenging.
It is also likely that Chd1l exerts its function in the preimplantation embryo not
through transcriptional regulation, but through the modulation of other nuclear processes.
64
ATP-dependent chromatin remodeling enzymes are utilized in a broad spectrum of
nuclear processes including homologous recombination, sister chromatin separation
during cell division, and DNA repair [283]. A role for Chd1l in the DNA damage
response has been confirmed by two independent groups, although a direct role in DNA
repair has yet to be confirmed. If the composition and function of a Chd1l-containing
complex is preserved throughout development, then the Chd1l phenotype of
preimplantation embryos will not be answered at the transcriptional level.
Methods
Antibodies
A polyclonal rabbit Anti-Chd1l antibody produced in our lab was used for
immunoprecipitation and ChIP experiments. A polyclonal rat α-Chd1l antibody, also
raised in-house, was used for immunoblotting of Rb α-Chd1l immunoprecipitates. The
RNA PolII antibody (clone 8WG16) is a mouse monoclonal antibody from Covance.
Chd1l Immunoprecipitation
For each IP, 120 µg of ES cell extract was used with 50 µl of Protein G magnetic
beads (Dynal) pre-conjugated to varying amounts of α-Chd1l (5 µl, 25 µl).
Immunoprecipitation was carried out over night at 4° in RIPA buffer. Beads were
washed the following day and eluted with 50 µl Laemeli buffer. 10% of the supernatant
and 10% of the eluate were loaded onto SDS-PAGE gel for analysis.
Cell Fractionation
ES cell nuclei were separated from a cytoplasmic fraction by lysing mild
detergent (0.1% Triton) and brief, gentle centrifugation. Nuclei were ruptured by
incubation in hypotonic buffer. The nucleosol fraction and chromatin pellet were
obtained by centrifugation.
Chromatin Immunoprecipitation and library preparation
ChIP experiments were performed essentially as described [303]. Feeder-free,
parental EBRTcH3 ES cells [284] were grown to subconfluence on 15-cm plates. Each
65
ChIP library was derived from four pooled ChIP reactions of 2x107 cells each, and the
total chromatin library contained DNA from 2x107 cells. Cell viability was checked
(>95% viability) by Trypan blue staining prior to crosslinking with 1% formaldehyde for
20 minutes. Cells were collected by scraping and lysed at a density of 4x107 cells/ml.
Chromatin was sheared to an average size of 250 base pairs by sonicating with a 3mm
stepped microtip for 1 pulse at 50% power and 6 pulses at 60% power. For each IP,
lysate containing the equivalent of 2x107 cells was incubated with 50 µl of Protein G
magnetic beads (Dynal) pre-conjugated with primary antibody (5 µg of α-PolII and 50 µl
of α-Chd1l). Beads were washed, crosslinks were reversed, and DNA was purified by
phenol-chloroform extraction. DNA libraries from the three samples were all of
acceptable purity (A-260/A280 ratios between 1.8 and 1.9).
Adapter oligos (Illumina) were then ligated to the ends of DNA
immunoprecipitates and sheared total, genomic DNA and subjected to 20 cycles of
amplification using primers to the adapter oligos. Fragments of 100 to 300 base pairs
were size selected by gel electrophoresis, subjected to a second round of 18 amplification
cycles, and purified again by phenol-chloroform extraction.
Recombineering of Chd1l-HA Targeting Vector
Approximately 9 Kb of genomic DNA containing the last three exons of the
Chd1l gene and downstream sequence was “captured” from a BAC clone (RP24-273N16,
BAC PAC Resources) by recombineering [304] into a pBluescript vector in SW105 cells
(NCI-Frederick). Intermediate constructs containing two short homology arms (~500
base pairs) flanking a NeoR cassette, GalK cassette, or HA tag were linearized and used
for sequential recombineering. Electrocompetent bacterial SW105 cells (NCI-Frederick)
were prepared by desalting, heat shocked at 42° to induce the expression of λ prophage
recombination proteins, and electroporated with linearized construct. First, the neomycin
resistance cassette was inserted into a region displaying low conservation between exons
22 and 23. The cassette containing the PGK promoter, neomycin resistance, and bGH
polyA tail was subcloned from the pL452 vector (NCI-Frederick). Next, HA was
inserted at the end of the Chd1l coding region. HA was inserted by sequential selection
first for the presence of GalK by plating clones on galactose agar medium, and then for
66
the loss of GalK (and replacement by HA) by plating on MacKonkey agar medium. The
GalK cassette was subcloned from pGalK (NCI-Frederick). Lastly, the DTA cassette was
inserted at the end of the 3’ Chd1l homology arm. The cassette containing the MCI
promoter, DT-A, and PGK polyA was subcloned from the pMC1-DTpA vector.
Contributing Collaborators
I would like to thank Ziming Weng1,4
and for performing Solexa sequencing, Dr.
Phil Lacroute1 for processing sequencing reads, their advisor Dr. Arend Sidow
1,4, Wenxiu
Ma5,6
for doing CisGenome analysis, her advisor Dr. Wing Wong6, and my advisor Dr.
Matthew P. Scott1,2,3
.
1Department of Genetics
2Department of Developmental Biology
3Department of Bioengineering
4Department of Pathology
5Department of Computer Science
6Department of Biostatistics
67
Figure 3.1 Preliminary validations for ChIP
A. Validation of α-Chd1l specificity. ES cell extracts induced (-Tet) or uninduced
(+Tet) to express a Chd1l-shRNA were immunoblotted with α-Chd1l, α-Oct4, and α-β-
Tubulin. α-Chd1l detects a band ~100 KD in uninduced ES cell extracts, but not in
extracts in which Chd1l had been knocked down. B. α-Chd1l efficiently
immunodepletes Chd1l from ES cell extracts in a dose dependent manner, but does not
pull down Oct4, another nuclear protein. C. Cell fractionation shows that, like Snf2l,
Chd1l fractionates with chromatin, while cytoplasmic β-tubulin does not. D. Validation
of PolII ChIP. PolII ChIP library was validated by qPCR amplification of selected
genomic loci. CNAP is an intergenic region where PolII does not bind, and ChIP does no
enrichment of PolII at this locus. Conversely, promoters of Gapdh and β-Actin are
enriched for PolII binding.
A C B
D
68
# Raw Reads # Post Filter % # Mappable % # Non-Repetitive %
Chd1l 12,540,570 6,415,343 51.20% 4,498,942 70.10% 3,588,770 79.80%
PolII 8,446,088 5,452,102 64.60% 3,895,115 71.40% 3,668,476 94.20%
Total C’tin 12,631,182 6,759,394 53.50% 3,652,432 54.00% 3,588,770 98.30%
Table 3.1 Number of reads obtained for ChIP libraries
The reads obtained from sequencing of Chd1l and PolII ChIP and total chromatin (c’tin) libraries were passed through a quality filter
(no arbitrary nucleotides for 25 base pairs) and then mapped to the mouse genome. Repetitive reads were filtered out to avoid
skewing peak calling software. Post-filter, mappable, non-repetitive reads were used for peak calling.
Figure 3.2 Distribution of reads around TSS
Reads from PolII and Chd1l ChIP libraries and total chromatin library were mapped 1000 base pairs upstream and 1000 base pairs
downstream of transcription start sites (TSSs). A small peak can be observed in the “total chromatin” sample because chromatin
typically shears more easily at the TSSs, and these ends of chromatin get preferentially sequenced. The peak from the PolII sample is
much more significant, confirming that PolII is frequently located at TSSs. The Chd1l sample shows a modest peak at the TSS.
69
ChIP Control
# of Peaks
FDR<0.01 FDR<0.05 FDR<0.10
Chd1l Total C'tin 98 206 792
PolII Total C'tin 15304 17582 17599
Table 3.2. Number of peaks called for ChIP samples
The number of peaks for Chd1l ChIP and PolII ChIP sequencing experiment a varying
false discovery rates (FDR) is shown. At a FDR of <0.05, ~200 peaks were called for
Chd1l ChIP. This is in sharp contrast with the nearly 18,000 peaks called for PolII ChIP.
ChIP Control
# of Peaks
FDR<0.01 FDR<0.05 FDR<0.10
Run #1 Chd1l Total C’tin 98 206 792
Run #2 Chd1l Total C’tin 8 178 699
Combined Chd1l Total C’tin 201 2836 9160
Table 3.3. Number of peaks called for the combined analysis
Combined analysis of two sequencing runs of the same Chd1l ChIP library produced a
greater number of Chd1l peaks than either individual sequencing run alone.
70
A
B
Conservation
RefSeq
PolII Peak
PolII Signal
Total C’tin
Chd1l Peak
Chd1l Signal
C
Conservation
RefSeq
PolII Peak
PolII Signal
Total C’tin
RefSeq
Chd1l Signal
Conservation
PolII Peak
PolII Signal
Total C’tin
Chd1l Peak
Fig. 8 cont. next pg.
71
Figure 3.3 Validation of Chd1l ChIP peaks
A-C. Each track is labeled at the right. Forward reads are coded in red; reverse reads are
coded in blue. Black bars above reads represent a peak called by the software. RefSeq
tracks show presence of genes. Arrowed regions are intergenic and show the direction of
transcription. A. PolII peak calling at the Gapdh locus. Bi-horned peaks of forward
and reverse reads flank the PolII peak called at the promoter of the Gapdh locus.
Background signal from total chromatin (c’tin) is minimal. B. Example of an excluded,
top-ranking peak called for Chd1l. Peaks were manually excluded from analysis if the
background showed excessively high signal. C. Example of a Chd1l ChIP peak
manually selected for further validation. Chd1l peaks were selected if background
was minimal and if flanked by forward and reverse reads. D. Overlap between Chd1l
and PolII peaks at FDR<0.05. The Venn diagram shows that a significant portion
of Chd1l peaks overlap with PolII. E. Validation of Chd1l peaks by qPCR. Primers
were designed around the peaks and used in qPCR analysis of Chd1l and PolII ChIP
libraries. A subset of the Chd1l peaks selected was associated with nearby PolII peaks.
qPCR of PolII ChIP samples at these loci show enrichment of PolII binding as expected.
However, no enrichment of Chd1l binding was found for any candidate locus in Chd1l
ChIP samples.
17495
87/88
118 PolII Chd1l
E
D
72
Figure 3.4. Re-analysis of two combined Chd1l ChIP sequencing runs
A. Overlap of the reads called for each individual sequencing run and the combined
analysis of both sequencing runs. No peaks called for sequencing run #1 were common
to peaks called for sequencing run #2. Peaks called for the combined analysis represent a
largely distinct set from either run #1 or run #2. B. Example of a top-ranked peak
called for a combined analysis of two Chd1l ChIP sequencing runs. The appearance
of peaks called for the combined analysis was not noticeably different than for peaks
called for the first analysis. Peak height remained between 4 and 5.
A
196
1 97
4 4
Run #1 Combined
Run #2
RefSeq
Chd1l Signal
Conservation
PolII Peak
PolII Signal
Total C’tin
Chd1l Peak
Comb. Chd1l Signal
Comb. Chd1l
Peak
Comb. Total C’tin
B
73
Figure 3.5. Chd1l-HA targeting vector
The Chd1l-HA targeting vector was built to incorporate an HA tag onto the 3’ end of the
Chd1l gene. Recombineering technology was used to produce a vector containing two 4
and 5 Kb arms of sequence homologous to the Chd1l locus (red), a neomycin resistance
(NeoR) cassette for positive selection (green), and a DTA cassette for negative selection
(blue). The HA tag (bright yellow bar) is incorporated prior to the stop codon at the end
of the last exon in the Chd1l gene. The pBlueScript sk(+) backbone is digested away
upon linearization prior to recombination in ES cells (yellow). After selection by
neomycin, the neomycin cassette is floxed out by Cre-mediated recombination at the
LoxP sites.
74
4. Molecular Functions of Chd1l in Early Development
Introduction
Multiple studies published within the last year have identified a role for Chd1l in
the DNA damage response pathway and as an oncogene that promotes spontaneous tumor
formation in mice [259,260,262]. Thus far, experiments conducted in this thesis were
directed at elucidating developmental and transcriptional roles for Chd1l. Having
identified an essential role for Chd1l in the preimplantation embryo, it remains unknown
whether defects in gene expression underlie the arrest phenotype (Chapter 2). However,
no changes in transcriptional regulation could be identified in ES cells deficient in Chd1l,
providing evidence that its role may not be at the level of gene regulation. Other ATP-
dependent chromatin remodeling factors are involved in a broad spectrum of nuclear
processes other than transcription [94,280]. The INO80 complex, for example, responds
to early indicators of DNA damage and repairs double-stranded breaks (DSBs) [92]. The
first half of this chapter explores a potential role for Chd1l in promoting apoptosis in ES
cells as a response to DNA damage.
Given that Chd1l is essential in the preimplantation embryo, the question remains:
what is the molecular mechanism of Chd1l in early development? Perhaps more
intriguing is the puzzle of why Chd1l is essential in the preimplantation embryo but not
in ES cells? One possibility is that Chd1l is involved in a process that occurs in the early
embryo but that does not occur or is no longer required in ES cells. In the second half of
this chapter, I propose two hypotheses regarding what this process could be, based on
knowledge about the protein structure of Chd1l, which includes a macro domain that
binds polymers of ADP-ribose (PAR), and from recent reports demonstrating that Chd1l
is a key player in the DNA repair response. The first hypothesis is that the arrest
phenotype of embryos injected with Chd1l MO is due to a defect in DNA repair; the
second hypothesis is that the arrest phenotype is due to disregulation in PAR levels,
which may also have implications in DNA repair.
Everyday, tens of thousands of nucleotides are damaged in a typical cell due to
endogenous sources of DNA damage as well as exogenous sources [6]. Such endogenous
75
sources can arise from byproducts of normal metabolism called reactive oxygen species
that oxidize nucleotides such that they become toxic, mispairing, or miscoding [305].
The sugar bonds of nucleotides are susceptible to spontaneous hydrolysis, resulting in
miscoding bases or abasic sites. Mismatch errors can arise from basepairing inaccuracies
during replication. In addition, S-adenosylmethionine can transfer methyl groups to
nucleotides, changing their composition and basepairing properties. Stalling of the
replication fork due to various types of lesions can generate double-stranded breaks
(DSBs), a highly deleterious mutation. If lesions are not repaired, programmed cell death
may ensue or mutations can be passed down to future cell lineages.
Fortunately, repair mechanisms exist to protect the genome from endogenous and
exogenous DNA damage. Repair mechanisms that function on nucleotides of at least a
partially intact strand of DNA include base excision repair (BER), in which a single
damaged nucleotide is removed and replace with another; nucleotide excision repair
(NER), in which a lesion causing distortion of the DNA strand is removed along with a
number of nucleotides flanking either side of the lesion [306]. Mismatches introduced
during replication are repaired by mismatch repair (MMR) [305]. Repair mechanisms
that respond to DSBs include error-free homologous recombination-mediated repair
(HRR) or error-prone non-homologous end joining (NHEJ), in which any two ends of a
broken strand of DNA can be ligated together [305].
Defects in repair mechanisms can lead to a broad spectrum of human diseases,
including many types of cancer [236]. Mutations in the BRCA1 or BRCA2 genes, for
example, lead to disruptions in DSB repair and predispose individuals to breast and
ovarian cancer. Mutations in ATM render cells deficient in the recognition of DSBs and
lead to a disease called Ataxia-telangiectasia that is characterized by childhood leukemias
and lymphomas. Disruptions in a number of genes involved in MMR lead to a condition
called hereditary non-polyposis colon cancer (HNPPC) whereas disruptions in genes that
participate in BER can lead to Xeroderma pigmentosum in which individuals suffer high
prevalence of skin cancer.
Because of the link with cancer, DNA repair is intensely studied, but repair
mechanisms during early embryogenesis are only recently becoming elucidated
[253,307,308]. Understanding how DNA repair functions during preimplantation
76
development can have major impacts on assisted reproductive technology (ART), since
infertility can be correlated with chromosomal abnormalities and inefficient DNA repair
[237,309]. Because the phase of the cell cycle influences which DNA repair pathways
are used, it is expected that the manner in which DNA is repaired in the preimplantation
embryo, in which cells have very short G1 and G2 phases and no apparent G1/S phase
check point, will be distinct [310,311]. These cells also have increased need to maintain
genetic fidelity. Several studies have shown that transcripts encoding proteins involved
in all the major DNA repair pathways are present in the zygote and in preimplantation
embryos [253,308]. However, not only is there strong translational regulation in embryos
[312], but control of DNA repair mechanisms occurs largely through modifications at the
protein level, leaving many unknowns about DNA repair during early embryogenesis.
Genetic studies have demonstrated that components of the BER pathway such as Polβ
[313], Apex [314], Lig1 [315], and Fen1 [316] are necessary for development. The
majority of mice null for components of NER are viable but display sensitivity to DNA
damage [317]. The components of NER necessary for embryonic survival are Xpd [318],
Rad23A/Rad23B (double knock-out) [319], and Xab2 [320]. There are few, if any,
components of MMR that result in embryonic lethality when deleted [317], but mutations
in members of this pathway lead to deleterious diseases and cancers in post-natal
individuals [307]. In contrast, proteins that are involved in the repair of DSBs that are
essential for development are many, suggesting high dependence on the HRR and/or
NHEJ pathways in the embryos [317].
ES cells are pluripotent cells that are derived from the inner cell mass of the
blastocyst and may also have unique ways of coping with DNA damage. ES cells rapidly
proliferate in culture while maintaining self-renewal indefinitely. This ability requires
stringent DNA repair mechanisms to ensure that damage is repaired quickly, or that cells
with damaged DNA are removed from the population to prevent mutations from being
passed down to daughter cells. ES cells are highly sensitive to DNA damage, being more
likely to undergo apoptosis as a response to DNA damage than other cells types
[108,116,117]. This is not to say that ES cells are less proficient in repairing damage
than other cell types. On the contrary, the spontaneous mutation frequency in ES cells is
at least 100 times lower than in isogenic MEFs, depending on which reporter gene is
77
assayed [109,110,112], indicating robust machinery efficiently repairs DNA damage in
ES cells. Similar to early development, understanding of DNA repair pathways in ES
cells lags behind knowledge of repair in somatic cells and differentiated cultured cells
[109].
PAR, or poly(ADP-ribose) is a post-translational modification that is immediately
added to nuclear receptor proteins in response to DNA damage. Because ADP-ribose is a
positively charged moiety, addition onto chromatin promotes chromatin relaxation and is
proposed to allow access to DNA repair machinery. The PAR modification is
recognized by the macro domain of proteins such as Chd1l [61,259,260]. Long and
branching chains of PAR are synthesized by the PAR polymerase (PARP) family of
enzymes using the respiratory metabolite NAD+ as a substrate [32]. PAR levels are
tightly regulated by polymerization by a large number of Parps and by degradation by a
single enzyme called poly(ADP-ribose) glycohydrolase, or PARG. Disregulation of PAR
levels during development leads to lethality; Double knock-outs of two major PARPs,
PARP-1 and PARP-1, are embryonic lethal at about E5.5 [57], and PARG knock-out
mice die even earlier, at E3.5 [289]. As the PAR modification has diverse roles in DNA
repair, transcription , telomere length homeostasis, and others [321], it is unclear what the
primary cause of lethality is when PAR levels are reduced or amplified in these mouse
mutants.
The role of PAR in the cell is best described in terms of the function of the
various PARP themselves. PARP-1 and PARP-2, the best described PAR polymerases,
have partially redundant functions [57]. Their activity is strongly activated by DNA
damage, either single-stranded or double stranded breaks, and mouse knock-outs of
PARP-1 or PARP-2 are hypersensitive to ionizing radiation and alkylation [52,57].
PARP-3 localizes to the centrosome, and its over-expression interferes with G1/S phase
progression [40]. Parp-5a and Parp-5b, also known as Tankyrase 1 and Tankyrase 2,
respectively, associate with telomeric proteins, including TRF1 [41,43]. Tankyrase 1
levels and perhaps Tankyrase 2 levels influence telomere length and segregation during
cell division [42].
In addition to safeguarding the genome, there is also a “dark side” to PARP-1
function [33]. PARP-1 promotes apoptosis through a caspase-independent AIF pathway,
78
an activity that is halted by treatment of PARP-1 inhibitors [65]. In the AIF-mediated
apoptosis, AIF, which is normally sequestered in mitochondria, is translocated to the
nucleus where it triggers rapid condensation and fragmentation of chromatin [322,323].
PARP-1 is required for AIF translocation to the nucleus, but how it accomplishes this
remains unknown [65]. Interestingly, Tankyrase 2 when over-expressed also induces
caspase-independent apoptosis, an effect diminished by treatment of a PARP inhibitor
[43].
Because Chd1l binds to PAR through the macro domain, Chd1l may contribute to
DNA repair through PAR recognition. Alternatively, Chd1l may recognize PARylated
proteins and carry out other nuclear functions, including the activation of caspase-
independent apoptosis. Like other chromatin remodeling factors, Chd1l may function as
the ATP-dependent motor of a complex that has multiple and diverse functions
throughout development and in different cell types. This chapter provides data that
indicate that Chd1l indeed recognizes DNA damage in ES cells, but may propagate a
response that is very different than that reported in the literature for other cell types.
Results
Reduction of Chd1l increases DNA damage tolerance in ES cells
In cultured cells, Chd1l participates in the response to several types of DNA
damage. Chd1l localizes to double-stranded breaks induced by laser microirradiation
[259,260], and reduction of Chd1l protein causes hypersensitivity to phleomycin- and
H2O2-induced damage [259]. Studies described in this dissertation demonstrate that
Chd1l is not essential for ES cells under normal culture conditions. However, these
experiments have not addressed whether Chd1l participates in DNA repair in ES cells.
Given that Chd1l participates in the DNA damage response in the U20S
osteosarcoma cell line [259] and in the HeLa cervical cancer cell line [260], it is
reasonable to hypothesize that Chd1l also participates in DNA repair in ES cells. To test
this, sensitivity to H2O2-induced oxidative DNA damage was assayed in ES cells
expressing Chd1l-shRNA or NS-shRNA. Expression of Chd1l-shRNA or NS-shRNA
was induced by removing Tetracycline 48 hours prior to seeding. Knock-down of Chd1l
in Chd1l-shRNA expressing ES cells was confirmed by SDS-PAGE (Fig. 1A). ES cells
79
were treated with H2O2 concentrations varying between 5 µM and 500 µM 24 hours after
seeding. Surprisingly, two days after treatment, a slight increase in survival was seen for
ES cells with reduced Chd1l at the lower concentrations of H2O2 (Fig. 1B). Although
subtle, the effect with H2O2 treatment is reproducible; multiple experiments showed
greater survival for ES cells expressing Chd1l-shRNA than those expressing NS-shRNA.
To test whether ES cells with reduced Chd1l are more sensitive to strand breaks,
ES expressing Chd1l-shRNA or NS-shRNA were treated with varying concentrations of
the radiomimetic drug phleomycin and assayed for percent survival relative to untreated
cells (Fig. 1C). In contrast to H2O2 treatment, treatment with phleomycin rapidly killed
control ES cells even with very low phleomycin concentration (5 µM). Therefore,
phleomycin treatment is reported after only 24 hours. ES cells expressing Chd1l-shRNA
showed higher percent survival than ES cells expressing NS-shRNA at all concentrations
of phleomycin treatment, a finding reproduced over multiple experiments. Intriguingly,
the increased resistance to DNA damage-induced killing seen in ES cells with Chd1l
knocked-down is more pronounced when phleomycin is used than when H2O2 is used,
suggesting that Chd1l may play a more prominent role in the recognition of double-
stranded breaks than lesions induced by oxidative damage.
To test whether knock-down of Chd1l provided long-term resistance to killing by
drug-induced DNA damage, survival rates over a three day period were measured. At 1
day after treatment, Chd1l-shRNA expressing ES cells survive slightly better than NS-
shRNA expressing cells in the lower H2O2 concentrations (<250 µM) (Fig. 1D). This
difference was reduced by 2 days (Fig. 1E) and not seen at 3 days (Fig. 1E). The data
suggest that increased resistance is an acute affect and may not confer long-term
survivability to ES cells. The data show that ES cells with reduced Chd1l levels do not
show hypersensitivity when treated with H2O2 or phleomycin, as has been reported in
other cell types. Instead, reducing Chd1l levels in ES cells confers acute resistance to
DNA damage-induced killing.
One explanation for this surprising result is that in ES cells, Chd1l promotes
apoptosis as a response to DNA damage. Cells that lack Chd1l could be impaired in
initiating apoptosis. Such a finding is reminiscent of a phenomenon termed “damage
tolerance” seen when ES cells deficient in a mismatch repair protein, Msh2, are treated
80
with DNA certain damaging agents [138,324,325,326]. Damage tolerance is observed in
cancer cells that are able to circumvent apoptosis as a response to DNA damage.
Over-expression of Chd1l kills ES cells but not differentiated cells
The result that ES cells with reduced Chd1l are more resistant to killing brought
to attention data obtained previously in this thesis showing that over-expression of Chd1l
causes widespread death in ES cells. EBRTcH3 ES cells [284] were used for stable
integration of Chd1l transgene and inducible expression by removal of Tetracycline from
culture medium. Two forms of Chd1l were over-expressed: a wild-type form (WT), and
a mutant predicted to lack ATPase activity due to a point mutation in the ATP binding
domain (K71R). The lysine in ATP binding motifs is critical to its function; mutating
this residue abolishes ATP hydrolysis in many different proteins, including SNF2
proteins in yeast and human [76,327,328,329,330,331]. Ahel et. al. showed that an
analogous mutation in the human CHD1L protein abolished ATP hydrolysis [259],
strongly suggesting that this mutation in mouse Chd1l is deficient in ATP hydrolysis.
Tetracycline was removed from culturing media to induce the expression of either
WT or K71R Chd1l. After six days of culturing without Tetracycline, ES cell
populations over-expressing either form of Chd1l had ~90% fewer cells than control ES
cell populations cultured in the presence of Tetracycline to maintain transgene repression
(Fig. 2A).
To address the question of whether over-expression triggers ES cells to
differentiate, Chd1l over-expression (either K71R or WT) was induced by removal of
Tetracycline for 24 hours and then ES cells were seeded at clonal density and cultured for
six days. Colonies were stained with Leishman’s Stain, a chromatin stain that labels
pluripotent cells. Colony morphology is indicative of the pluripotent state of ES cells; ES
cell colonies that are round, small in diameter, and multilayered indicate pluripotency,
whereas colonies that are monolayered and diffuse are indicative of differentiation (Fig.).
Some spontaneous differentiation (~5%) is expected in ES cell cultures. Chd1l over-
expressing colonies (either K71R or WT) were smaller in diameter than control ES cell
colonies (Fig. 2C). The number of colonies obtained for ES cells over-expressing Chd1l
was reduced by more than 60%, indicating that the loss in cell number was not only due
81
to lack of proliferation (Fig. 2B). Colonies were scored as either “differentiated” or
“undifferentiated.” Chd1l over-expressing colonies were 95% (K71R) or 94% (WT)
undifferentiated (pluripotent) compared to 95% undifferentiated in control colonies (Fig.
2D). Therefore, the effect of Chd1l over-expression on differentiation in ES cells is
minimal.
The drastic loss in cell number upon Chd1l over-expression could be due to
changes in proliferation or cell death. To address whether Chd1l over-expression causes
cell cycling, ES cells were cultured without Tetracycline for three days to induce the
expression of either WT or K71R forms of Chd1l, permeabilized, and stained with
propidium iodide (PI) to label DNA. DNA content was determined by flow cytometry
and percentages of cells in each phase of the cell cycle were calculated using FlowJo
software. Consistent with reports that Chd1l may contribute to G1/S phase transition
[262], 4.5% fewer of K71R and 3% fewer of WT Chd1l over-expressing cells were found
in G1 phase compared to their uninduced (+Tet) counterparts, small but significant
differences (χ2 test, p<0.001) (Fig. 2E). However, no gross arrest in any phase was
observed that could explain ~90% reductions in cell numbers.
To ask if ES cells over-expressing Chd1l were undergo apoptosis, ES cells that
were induced for four days to express either WT or K71R forms of Chd1l were collected
and co-stained with PI and AnnexinV. The population of ES staining positive for
AnnexinV but negative for PI was counted using flow cytometry and gating was done
using FlowJo software. There were ~3.5 times (K71R) or ~2.5 times (WT) as many
apoptotic Chd1l-over-expressing ES cells as control ES cells (χ2 test, p<0.001) (Fig. 2F),
showing that over-expression of Chd1l initiates apoptosis in ES cells.
To test whether over-expression of Chd1l is toxic to other cell types, Chd1l was
over-expressed in 3T3 fibroblasts. 3T3 cells were transfected with a GFP-containing
vector only or WT or K71R forms of Chd1l. No readily apparent differences in cell
death were seen between cells transfected with vector only and cells transfected with
Chd1l-containing vector. As cell division ensues, the transfected vector will be diluted
out. If Chd1l is toxic in 3T3s, cells that lose the Chd1l-containing vector will have a
selective advantage over those that retain it. To test this, cells expressing GFP were
selected two days after transfection using FACS sorting and cultured for an additional
82
two days. A second-round of flow cytometry revealed that 68% or 70% of cells
transfected with the K71R or WT Chd1l-containing vector retained GFP expression,
whereas only 58% of cells transfected with the empty vectors retained GFP expression
(Fig. 3A), showing that Chd1l is not toxic to 3T3s and suggesting instead that over-
expression of Chd1l endows a slight (~10%) selective advantage over those transfected
with vector only.
The cell death caused by over-expression of Chd1l may be a phenotype specific to
ES cells. Mouse models over-expressing Chd1l are viable and experience ~24% increase
in spontaneous tumor formation in adult mice [262]. To test whether differentiating ES
cells can tolerate the over-expression of Chd1l, LIF was removed from culturing medium
and expression of Chd1l transgene was induced by tetracycline removal. ES cells were
cultured in the absence of LIF for 0, 2, 4, or 6 days prior to tetracycline removal, and LIF
was withheld throughout an additional six days of culturing with Chd1l over-expression
(Fig. 3B). Survival of ES cells over-expressing Chd1l was impaired in undifferentiated
ES cells as well as cells in which Chd1l over-expression was induced simultaneously
with LIF removal (D0-LR) (Fig. 3C). However, Chd1l over-expressing ES cells
differentiated for 2 or 4 days prior to Chd1l induction (D2-LR or D4-LR) survived as
well as (and better than) control ES cells (Fig. 3C), indicating that at some critical
moment during differentiation, ES cells become resistant to Chd1l over-expression.
Unexpectedly, the results were highly similar between over-expression of WT and
K71R forms of Chd1l for all experiments described above. It was originally thought that
over-expression was a loss of function phenotype because over-representation of a
subunit of a complex can disrupt protein stoichiometry and yield partially formed,
nonfunctional complexes. However, reducing Chd1l levels by shRNA (a more likely
loss-of-function manipulation) produces a very different effect (Chapter 2); therefore, the
apoptosis-inducing phenotype of Chd1l over-expression in ES cells is probably a
hypermorphic or neomorphic function that does not depend on the ATPase domain. It
would be interesting to address whether the macro domain plays a role in the apoptotic
phenotype by over-expressing a macro domain mutant in ES cells.
83
Chd1l forms a ~500 kD complex in ES cells
ATP-dependent chromatin remodeling factors are found in multi-subunit
complexes than can change in composition and function throughout development and in
different cell types [69,70,71]. To address the question of whether Chd1l forms a
complex in ES cells, nuclear extracts were prepared and size fractionated on a 10-30%
glycerol gradient. Large proteins and complexes will filter farther through the gradient
than smaller proteins and complexes. Fractions were then run on SDS-PAGE and blotted
for Chd1l, pluripotency proteins Oct4 and Sox2, the DNA repair protein PARP-1, and the
chromatin remodeling factor Brg1.
Chd1l was found most abundant in fractions 4-6, cofractionating with PARP-1
(Fig. 4). Consistent with previous reports, Brg1 was most abundant in fractions 12-14
[69], correlating with a complex size of ~2 mDa. Oct4 and Sox2 form a complex of
about 440 kDa [332]. Therefore, fractions 4-6 in which Chd1l and PARP-1 are found are
estimated to contain complexes about 500 kDa. Chd1l and PARP-1 associate with each
other in HEK293 and in 293T cells [259,260]. Co-fractionation with Chd1l and PARP-1
suggest that these proteins also interact in ES cells, although co-immunoprecipitation
experiments are necessary to confirm this.
Because Chd1l is itself 100 kDa, a sum of ~400 kDa of subunits remain to be
identified. PARP-1 is 113 kDa, potentially accounting for some of the mass.
Identification of complex subunits could yield valuable insight into the function of Chd1l,
as could characterization of a Chd1l-containing complex in ES cells that have been
treated with DNA damaging agents.
γ-H2AX marks uninjected embryos and embryos injected with Chd1l-MO
Given that Chd1l participates in the DNA damage response in cultured cells, it
could also govern a DNA damage response in the preimplantation embryo. Although
DNA repair is still being elucidated in very early development, if any cells of an
organism should need strict repair mechanisms, it should be the zygote and blastomeres
as these are the progenitor cells that anchor all developmental lineages. Indeed, a large
number of expression studies reveal that transcripts encoding DNA repair proteins for
mismatch repair (MMR), base excision repair (BER), nucleotide excision repair (NER),
84
homologous recombination- mediated repair (HRR), and non-homologous end joining
(NHEJ) are present in the zygote and throughout preimplantation development [253].
Many DNA repair proteins are essential for development as their knock-out phenotypes
results in early embryonic lethality, especially those contributing to HRR of double
strand breaks [317]. Therefore, a defect in DNA repair could potentially explain the
arrest phenotype seen in embryos injected with MO targeting Chd1l.
One way in which to test for DNA damage is by assaying the levels of
phosphorylation of the histone variant H2AX. Phosphorylated H2AX associates with
sites of DNA damage and is essential for the repair of DSBs [333]. Increased levels of
γH2AX have been demonstrated to occur in the embryo upon treatment with DNA
damaging agents [334,335]. Quantification at the protein level is inherently difficult in
the preimplantation embryo due to the low abundance of starting material that is further
limited by the number of embryos that can be µ-injected in a single experiment (30 to
50). Multiple attempts at obtaining reliable results from SDS-PAGE or dot-blotting using
as many as 100 embryos were unsuccessful. Although less quantitative, general protein
levels can be observed using immunocytochemistry (ICC) staining in embryos.
Therefore, to address whether embryos deficient in Chd1l experience an
accumulation of DNA damage, embryos injected with Chd1l MO were subjected to ICC
staining using the α-All uninjected embryos developed normally to the 16- to 32- cell
stage. Embryos injected with MO-1 developed to the 8- to 16-cell stage; previous
experiments (Chapter 2) show that embryos injected with MO-1 arrest at this stage and
do not develop further. Embryos injected with MO-2 were arrested at either the 2-cell
stage or the 4-cell stage. At least fifteen embryos were collected for each uninjected,
MO-1 injected, and MO-2 injected embryos and stained with α-γH2AX antibody. .
Phosphorylated H2AX also marks nuclei of embryos injected with Chd1l MO-1, and
staining in some nuclei is brighter than nuclei of uninjected embryos (Fig. 5A). Nuclei of
embryos injected with Chd1l MO-2 also showed strong γH2AX staining that appear
somewhat brighter than control embryos (Fig. 5B). The data is consistent with a role for
Chd1l in the early embryo, suggesting that a loss of Chd1l may increase DNA damage,
particularly damage that ends in DSBs. The differences in staining signals are difficult to
quantify because ICC staining is only semi-quantitative, and further experiments are
85
necessary to confirm that accumulation of DNA damage is the underlying cause of arrest
in preimplantation embryo injected with Chd1l MO.
PAR marks uninjected embryos and embryos injected with Chd1l-MO
PAR levels are carefully regulated in the preimplantation embryo by the activity
of the PARP family of enzymes that add PAR onto nuclear receptor proteins, and by the
activity of PARG, which rapidly degrades PAR modifications. Mouse knockouts of
PARG are unable to catabolize PAR [289], and double knockouts of partially redundant
polymerases PARP-1 and PARP-2 are deficient in PAR synthesis [57]. Knock-out of
either PARG or both PARP-1 and PARP-2 results in early embryonic lethality at E3.5 and
7.5, respectively.
Because Chd1l binds to PAR through its C-terminal macro domain, and because
embryos injected with Chd1l-MO arrest prior to the blastocyst stage, it was intriguing to
hypothesize that the arrest phenotype was due to aberrant PAR levels in the embryo.
Given that Chd1l binds to PAR in vivo, Chd1l could be involved in the propagation of a
PAR recognition and degradation process. Loss of Chd1l could disrupt PAR recognition
and subsequent degradation by PARG, leading to toxic PAR levels. Alternatively,
through binding to PAR, Chd1l could stabilize its presence long enough for it to be
recognized by other proteins. Loss of Chd1l could lead to PAR destabilization, rapid
degradation, and untenably low levels in the cell. If either of these scenarios were true,
PAR levels would be altered in embryos injected with Chd1l MOs.
To test whether altered PAR levels could explain the Chd1l arrest phenotype,
embryos injected with Chd1l-MO were subjected to ICC and stained with α-PAR.
Zygote-stage embryos were µ-injected with Chd1l MO-1 or MO-2 and fixed three days
later (E3.5). Uninjected embryos appeared to develop normally to this stage, and
embryos injected with MO-1 developed to and were likely arrested at o the 8- to 16-cell
stage. Embryos injected with MO-2 were arrested at either the 2-cell stage or the 4-cell
stage. At least fifteen embryos were collected for each uninjected, MO-1 injected, and
MO-2 injected embryos and stained with α-PAR antibody. Uninjected embryos show
low levels of PAR staining in the nucleus, consistent with reports that PAR modifies
nuclear proteins (Fig. 6). Some globular staining can be seen outside of the nucleus, but
86
this is likely non-specific sticking to membranes or the zona pellucida. One uninjected
embryo arrested at the 2-cell stage, and this embryo is shown to contrast Chd1l MO-2
embryos that consistently arrest at this stage (Fig. 6C).
PAR staining in Chd1l MO-1 embryos also shows nuclear localization of PAR
(Fig. 6A), demonstrating that the arrest phenotype is not due to global reduction of PAR.
Some blastomeres of MO-1 injected embryos show PAR staining that is slightly greater
than those in uninjected embryos, suggesting PAR levels could be higher in arrested
embryos. As ICC is semi-quantitative at best, it is difficult to make conclusions between
samples with similar signals. Even if PAR levels are higher in embryos injected with
MO-1, the difference may not be great enough to explain the early arrest phenotype.
PAR staining in Chd1l MO-2 embryos show nuclear localization of PAR, but the signal
in some cells is reduced compared to cells of uninjected embryos (Fig. 6B, note
uninjected embryo arrested at the 2-cell stage), suggesting that PAR levels are actually
reduced. The difference is subtle and inconsistent between embryos, and the discrepancy
between embryos injected with MO-1 and MO-2 is difficult to rationalize. One
possibility is that the transient nature of PAR modification causes it to degrade in arrested
embryos. Thus PAR in MO-2 injected embryos that had been arrested for a longer
duration than MO-1 injected embryos may show reduced PAR levels.
The data show that PAR levels are not consistently or dramatically altered in
embryos injected with Chd1l MO, indicating that the arrest phenotype is not due to large-
scale changes in PAR levels. More quantitative techniques are necessary to determine
whether PAR levels change subtly in MO-injected embryos versus uninjected embryos;
however due to the discrepancy between MO-1 and MO-2 injected embryos, identifying
such subtle changes remains unlikely to explain the arrest phenotype seen in embryos
injected with Chd1l MO.
Discussion
Upon experiencing DNA damage, a cell has two choices: 1) repair the damage, or
2) initiate programmed cell death or apoptosis to remove the damaged cell from the
population. ES cells lack a G1 checkpoint in which some DNA repair takes place and are
instead more prone to initiate apoptosis in response to DNA damage than other cell types
87
[108,116,117,119,338]. Two independent groups have demonstrated that Chd1l is part of
the DNA damage response [259,260]. In ES cells, Chd1l may respond to DNA damage
by promoting apoptosis; therefore, reduction of Chd1l in ES cells may increase their
tolerance to drug-induced DNA damage. Further experiments will be necessary to test
whether DNA damage accumulates and whether the programmed cell death pathway is
impaired in ES cells deficient in Chd1l.
Consistent with the speculation that Chd1l pushes ES cells toward apoptosis in
response to DNA damage, over-expression of Chd1l in ES cells causes widespread
killing and increased staining of the apoptotic marker AnnexinV. This effect is
reminiscent Tankyrase 2- (PARP-5b-) induced cell death upon transient transfection
[43]. PARP-1, when over active can also trigger apoptosis in a caspase-independent
manner. Cell death induced by both Tank2 over-expression and Parp-1 over-activation
is blocked by addition of PARP inhibitors, suggesting that the presence of additional
PAR molecules may contribute to the initiation of apoptosis [339]. An alternative
explanation is that over-active PARP depletes cellular NAD+ and ATP levels, and
necrosis ensues [37]. Interestingly, Parp1 deficiency in mouse disease models is
cytoprotective to ischemia-reperfusion injury, inflammation-related injury [63], diabetes
[64], and hyperoxic damage. However, several studies indicate that energy depletion
alone may not account for Parp-1-induced cell death [340]. Cell death induced by over-
activation of PARP-1 requires the translocation of AIF to the nucleus, and inhibition of
AIF by microinjection of AIF antibodies blocks PARP-1-dependent cell death [65]. The
release of AIF from the mitochondrion is PARP-1 dependent, but how PARP-1 trigger
AIF release is not known [65]. It is interesting to speculate that Chd1l may play a role in
PAR recognition and signaling leading to AIF release. While it is puzzling to think how
a chromatin remodeling factor might participate in this, cell death caused by Chd1l over-
expression seems to be independent of nucleosome remodeling activity, since over-
expression of the K71R mutant Chd1l, predicted to be deficient in ATP binding, and
over-expression of wild-type Chd1l phenocopy each other.
The increased survival of ES cells with reduced Chd1l after DNA damage was
more pronounced when that damage was induced by the radiomimetic drug Phleomycin
than with oxidative damage-inducing H2O2. Converging lines of evidence suggest that
88
ES cells are less sensitive to oxidative damage as they can be cultured in hyperoxic
conditions with fewer deleterious effects than other cell types, and are instead more
dependent on strand break repair pathways, particularly error-free recombination-
mediated repair (HRR) [109]. ES cells show high basal levels of γH2AX staining, an
established marker of double stranded breaks (DSB) [109]. Because of increased
dependence on HRR to repair DSBs, loss of Chd1l could show a stronger phenotype in
ES cells treated with drugs that induce strand breaks than drugs that introduce oxidative
damage.
The data shown here present several striking paradoxes with the published
literature. Knock-down of Chd1l in ES cells and subsequent treatment with DNA
damaging agents produces cells that are more resistant to killing than control ES cells.
This is a direct contradiction with literature showing that knock-down of Chd1l in U2OS
cells causes hypersensitivity to treatment with the same DNA damaging agents [259].
That over-expression in ES cells induces apoptosis also conflicts with studies showing
that Chd1l can act as an oncogene, because over-expression in cell culture increases
colony formation in soft agar assays and transgenic over-expression in mouse models
induces spontaneous tumors [261,262]. Lastly, it is surprising that Chd1l is essential for
development to the blastocyst state, but not essential in ES cells that are derived from the
inner cell mass of the blastocyst (Chapter 2).
The data raise the possibility that Chd1l has a very different (or opposite) function
in ES cells that are adapted for culturing and maintenance of pluripotency. Supporting
this theory, apoptosis induced by Chd1l over-expression is specific to ES cells. The same
ES cells that are killed by Chd1l over-expression are no longer sensitive with as little as
two days of differentiation by LIF removal prior to introduction of Chd1l transgene.
DNA repair studies focusing on mechanisms in ES cells are beginning to show that ES
cell DNA repair mechanisms may be exceptional. On a conceptual level, ES cells should
have robust repair mechanisms to account for the rapid proliferation of a pluripotent cell
type. Spontaneous lesions, mismatches, and double strand breaks accumulate with every
cell cycle and can be highly detrimental for daughter cells and developmental lineages
[6]. During the derivation of ES cells from blastocysts, many epigenetic changes take
place, including those that induce the expression of some DNA repair genes [208].
89
Evidence that ES cells have extraordinary repair mechanisms arises from low
spontaneous mutation frequencies (10-6
) found at the Aprt reporter locus, compared to 10
-
4 in isogenic MEFs [111,112]. One way in which ES cells achieve this is by quickly
removing damaged cells from the population through apoptosis [116,117].
ES cells are also unique in that their rapid cell cycling is driven by extremely high
levels of Cdk2 [120], and they lack a functional G1/S phase checkpoint [108,119,338].
This is in part due to very low levels of the Cdk inhibitors p21 and p27 and sequestering
of p53 in the cytoplasm that renders it partially nonfunctional [120]. Restoration of G1/S
phase checkpoint by ectopic expression of the serine/threonine kinase Chk2 impairs the
ability of ES cells to undergo apoptosis in response to DNA damage [108].
If Chd1l participates in the G1/S phase transition as has been proposed here and in
Chen, et. al. [262], its function in ES cells that lack a G1/S phase checkpoint could be
drastically different. Instead of responding to DNA damage by facilitating repair, Chd1l
may orchestrate an apoptotic response instead. It should be noted that while Chd1l has
been shown to respond to DNA damage, it has not yet been shown to actually repair
DNA [259,260]. A chromatin remodeling complex can achieve highly diverse (and
opposite) functions through context-dependent and cell type-specific assembly of
subunits [70,280]. Chd1l forms a ~500 kD complex in ES cells. Investigation of
complex members in ES cells should yield interesting insights.
Therefore, Chd1l may play a very specific role in ES cells in the DNA damage
response pathway, potentially pushing cells towards apoptosis through the PARP-1/AIF-
dependent, caspase-independent pathway. This theory is supported by the striking switch
in the ability of Chd1l over-expression to kill cells when ES cells are differentiated by
two days of LIF removal. What happens to the DNA repair and response machinery at
the molecular during these early stages of differentiation is unclear, but would be
intriguing to investigate.
Whether defects in DNA repair can explain the early arrest phenotype of embryos
injected with Chd1l MO remains an open question. Evidence presented here showing
increased presence of γH2AX in Chd1l deficient embryos is suggestive rather than
conclusive due to the challenges of comparing semi-quantitative ICC signal staining
between samples. Staining with other indicators of DNA damage could be informative.
90
Other approaches include FISH analysis to detect large scale chromosomal anomalies and
mutation analysis at reporter loci such as the Arpt locus [111].
Alternatively, defects in DNA repair might not explain the Chd1l arrest phenotype
at all. The effects of PARylation on the embryo are still being elucidated, and although
DNA damage activates PARP-1 to PARylate itself and other nuclear receptors, activated
PARP-1 has multiple roles, many of which are independent on DNA damage. In
addition, Chd1l may recognize PAR modifications catalyzed by another member of the
PARP family. PAR levels are not consistently altered in Chd1l deficient arrested
embryos; however, the phenotype could be due to processes downstream of PAR. PAR
binding by Chd1l through its macro domain then would not affect PAR synthesis or
degradation, and the consequence of the interaction between Chd1l and PAR would be
strictly the execution of the PAR signal.
Methods
DNA Sensitivity Assay in ES cells
The expression of Chd1l-shRNA or NS-shRNA was induced in EBRTcH3 ES
cells for two days prior to seeding. “Induced” ES cells were cultured in 7.5 µg/ml
Puromycin; “Uninduced” ES cells that do not express shRNA were cultured in 7.5 µg/ml
Puromycin and 1.5 µg/ml Tetracycline. ES cells were seeded onto 24 well plates to allow
subconfluent growth for 1, 2, or 3 days for each induced and uninduced Chd1l-shRNA
and NT-shRNA EBRThH3 cell lines. H2O2 or phleomycin was added 24 hours after
seeding with concentrations ranging from 0 to 500 µM (H2O2) or 0 to 80 µM
(phleomycin). After 24 hours, 48 hours, and 72 hours, surviving ES cells were counted
using an automated cell counter that incorporates a trypan blue cell viability
measurement. Percent survival was recorded for each induced ES cell line compared to
the corresponding uninduced cell line.
ES cell lines
The EBRTcH3 cell line contains a cassette acceptor utilizing loxP and loxPV sites
at the Rosa locus to allow efficient and directional integration of a transgene by Cre-
mediated recombination. The derivation of Chd1l-shRNA and NS-shRNA cell lines is
91
described in Chapter 2 of this dissertation. To create an ATPase-deficient mutant of
Chd1l, a point mutation was created using site-directed mutagenesis (Stratagene)
encoding a single amino acid change in the ATP-binding domain (K71R). cDNA
encoding wild-type for K71R forms of Chd1l was subcloned into the pPthC exchange
vector for recombination into the EBRTcH3 ES cell line.
The exchange vector containing the shRNA-mir sequence was cotransfected
along with a Cre expression plasmid by lipofectamine. Transfected cells were plated
single-cell density and cultured in the presence of Puromycin (1.5 µg/ml) to select for
successful integrants and of Tetracycline (1.0 µg/ml) to repress transgene expression.
Clones were confirmed by PCR genotyping of the 5’ and 3’ recombination sites. To
induce shRNA expression, the derived ES cell lines were cultured in the absence of
Tetracycline and high Puromycin (7.5 µg/ml). Control, uninduced ES cells were cultured
in high Tetracycline (1.5 µg/ml) and high Puromycin (7.5 µg/ml).
Size fractionation
Untreated parental EBRTcH3 cells (without the introduction of transgene) were
cultured in 100 µg/ml Hygromycin and grown to subconfluence on 15-cm plates.
Nuclear extract was dialyzed in 5% glycerol buffer. To make the gradient, a 30%
glycerol buffer was layered below a 10% glycerol buffer and rotated at an angle of 81.5°
for 2” at a speed of 14. Nuclear extract (0.5 ml, 10 mg/ml) was layered on top of the
gradient and ultra-centrifuged at a speed of 32,000 rpm at 4° for 20 hours. Twenty six
fractions of ~0.5 ml were collected and run on SDS-PAGE gel.
FACS analysis
For analysis of cell cycle, ES cells were grown in the absence of tetracycline for
three days to induce the expression of K71R or wild-type Chd1l. Cells were trypsinized
and resuspended at a density of 106/ml, fixed for 1hour at 4° in 100% EtOH, stained with
PI, permeabilized in 1% Triton for 10 minutes, then subjected to flow cytometry. For
analysis of apoptosis, ES cells were cultured in the absence of tetracycline for four days
to induce the expression of K71R or wild-type Chd1l. Cells were trypsinized and
resuspended at a density of 106/ml, fixed in 100% EtOH, stained with PI and AnnexinV
92
without permeabilization, then subjected to flow cytometry. Data was analyzed using
FlowJo software. Statistics were calculated using χ2.
Embryo immunocytochemistry
Unless otherwise specified, all incubations and washes were carried out at room
temperature in 20-50 µl droplets of PBS-PVP solutions covered in mineral oil. Embryos
were fixed in 4% PFA for 20” at room temperature and washed for 10 minutes in PBS-
PVP. Embryos were permeabilized in a large volume of 0.3% Triton X-100 for 10
minutes without mineral oil and then washed for 5” in PBS-PVP droplets under mineral
oil. Embryos were incubated in Image iT signal enhancer solution (Invitrogen) for 30”,
washed in PVP-PBS for 15 minutes, blocked in 10% normal goal serum for 30”, and then
incubated overnight in primary antibody at 4°. Embryos were washed for 30” in PBS-
PVP, incubated in secondary antibody for 2 hours and then Hoechst for 10”. After final
washing in PBS-PVP for 30”, embryos were mounted in fibrinogen/thrombin clots. Clots
were made of equal volumes of 25 mg/ml fibrinogen (Sigma) in Ringer’s Solution and
100U/ml thrombin (Sigma) in PBS. Embryos were imaged using confocal microscopy
under 63x oil immersion. Five to ten embryos were placed into each clot.
Antibodies
Mouse α-PAR antibody was obtained from Trevigen (4335-AMC-050). Mouse
α-γH2AX was obtained from Abcam (ab18311).
93
Figure 4.1. Sensitivity to DNA damage in ES cells expressing Chd1l-shRNA
Expression of Chd1l-shRNA or NS-shRNA was induced by removal of Tetracycline 48
hours prior to treatment with varying concentrations of Phleomycin or H2O2. A. Chd1l
knock-down. Knock-down of Chd1l is observed in Chd1l-shRNA expressing ES cells
(-Tet), but not in NS-shRNA expressing ES cells (-Tet) or in uninduced ES cells (+Tet).
B-E. Sensitivity assays. Percent survival of ES cells expressing shRNA was measured to
determine sensitivity to H2O2-induced oxidative damage (B-D) or Phleomycin-induced
strand breaks (E) relative to untreated ES cells.
A
B C
D E
Figure 1. DNA repair in Chd1l-shRNA ES cells
F
94
E
C
A
-Tet +Tet
Undiffe
rentiate
d
Diffe
rentiate
d
Cell Number Colony Number
Colony Morphology D
F
B
Cell Cycling Apoptosis
Figure 2. Effect of Chd1l Over-Expression in ES cells
95
Figure 4.2. Effects of over-expression of K71R or WT Chd1l.
A. Cell number. After six days of culturing in the absence of Tetracycline, there is
~90% reduction in the total number of ES cells over-expressing K71R or WT Chd1l. B.
Colony Survival. Tetracycline was removed and cells were seeded at clonal density and
then cultured for six days. The number of colonies obtained for ES cells expressing
K71R or WT Chd1l was reduced by ~60%. C. Colony Morphology. Representative
images of pluripotent or differentiated ES cell colonies expressing Chd1l-K71R or held in
the presence of Tetracycline to repress transgene expression. D. Quantitation of Colony
Morphology. About 95% of colonies expressing either K71R or WT forms of Chd1l
appear undifferentiated, similar to control colonies held in the presence of Tetracycline.
E. Cell Cycling. The percentage of cells in G1, S, or G2/M phase in each population of
ES cells expressing K71R or WT forms of Chd1l is compared to control ES cells cultured
in the presence of Tetracycline. F. Apoptosis. Fixed, non-permeabilized cells were
labeled with PI or AnnexinV-FITC and analyzed by flow cytometry. The percentage of
cells that labeled AnnexinV+/PI- are shown for ES cells expressing either K71R or WT
versions of Chd1l or control cells cultured in the presence of Tetracycline.
96
Figure 4.3. Chd1l over-expression specifically kills undifferentiated ES cells
A. Competition in 3T3 cells transiently transfected with Chd1l. Transfected 3T3 cells
were selected using FACS sorting on GFP-expressing cells and then cultured for an
additional two days. The percentage of cell that retain the empty vector or vector
containing K71R or WT Chd1l is shown (GFP+). The cells that have lost the vector (i.e.
through dilution upon cell division) are GFP-. Cells were stained with PI to label dead
cells. B. Experimental Design of ES cell differentiation assay. Tetracycline was
γH2AX
γH2AX
γH2AX+Hoechst γH2AX+Hoechst
γH2AX+Hoechst
Competition in 3T3 Cells
A B
Figure 3. Chd1l over-expression kills ES cells only
C Cell Number in Differentiating ES Cells
Experimental Design
97
removed from ES cells differentiated for 0, 2, or 4 days by removal of LIF (D0 LR, D2
LR, or D4 LR). Differentiating ES cells were cultured for an additional six days in the
absence of Tetracycline and in the absence of LIF. C. Results of ES cell differentiation
assay. Cell number was assayed after 0, 2, or 4 days of LIF removal and an additional 6
days of LIF removal and over-expression of K71R or WT Chd1l. Chd1l over-expression
kills ES cells cultured in the presence of LIF and “D0 LR” differentiating cells. Neither
“D2 LR” or “D4 LR” differentiating cells are susceptible to over-expression of Chd1.
98
Figure 4.4. Size Fractionation of a Chd1l-containing protein complex
ES cell nuclear extracts were submitted to size fractionation of a glycerol gradient
ranging from 10% 5o 30%. Fractions were run on SDS-PAGE and blotted for Chd1l. A
Chd1l-containing complex is smaller than the Brg1-containing BAF complex (~2 mDa)
and larger than the Oct4/Sox2 complex (<440 kDa), and is estimated at ~500 kDa. Chd1l
cofractionates with the DNA repair protein PARP-1.
Figure 4. Size Fractionation of Chd1l complex
γH2AX γH2AX
γH2AX
99
Figure 5. α-γH2AX Staining
A. Chd1l MO-1 Injected
γH2AX γH2AX
γH2AX
γH2AX+Hoechst γH2AX+Hoechst
γH2AX+Hoechst
C. Uninjected
γH2AX γH2AX
γH2AX
γH2AX+Hoechst γH2AX+Hoechst
γH2AX+Hoechst
100
Figure 5 cont.: α-γH2AX Staining
B. Chd1l MO-2 Injected
γH2AX γH2AX
γH2AX
γH2AX+Hoechst γH2AX+Hoechst
γH2AX+Hoechst
C. Uninjected
γH2AX γH2AX
γH2AX
γH2AX+Hoechst γH2AX+Hoechst
γH2AX+Hoechst
101
Figure 4.5. γH2AX staining in embryos injected with Chd1l-MO.
Embryos injected with Chd1l MO-1 (A) or Chd1l MO-2 (B) were fixed 2 days after
injection (E3.5) and stained with α-γH2AX . Embryos injected with MO-1 arrest at the
8- to 16- cell stage. Embryos injected with MO-2 arrest at the 2- to 4- cell stage.
Uninjected embryos at the morula stage (E3.5) were fixed and stained in parallel with
MO-injected embryos (C, shown twice for comparison). Images are taken from 63x oil
immersion confocal microscopy.
102
Figure 6. α-PAR Staining
C. Uninjected
PAR PAR PAR
PAR+Hoechst PAR+Hoechst PAR+Hoechst
A. Chd1l MO-1 Injected
PAR PAR PAR
PAR+Hoechst PAR+Hoechst PAR+Hoechst
103
Figure 6 cont.: α-PAR Staining
C. Uninjected
PAR PAR PAR
PAR+Hoechst PAR+Hoechst PAR+Hoechst
B. Chd1l MO-2 Injected
PAR PAR PAR
PAR+Hoechst PAR+Hoechst PAR+Hoechst
104
Figure 4.6. PAR staining in embryos injected with Chd1l-MO.
Embryos injected with Chd1l MO-1 (A) or Chd1l MO-2 (B) were fixed 2 days after
injection (E3.5) and stained with α-PAR. Embryos injected with MO-1 arrest at the 8- to
16- cell stage. Embryos injected with MO-2 arrest at the 2- to 4- cell stage. Uninjected
embryos at the morula stage (E3.5) were fixed and stained in parallel with MO-injected
embryos (C, shown twice for comparison). Images are taken from 63x oil immersion
confocal microscopy.
105
5. General Discussion
Chd1l is Essential in the Preimplantation Embryo
Chd1l has stage and cell-type specific expression patters during embryonic
development [258]. There are many nuclear reprogramming and chromatin remodeling
activities that occur during embryogenesis, and many of them are not well understood.
To address the question of whether Chd1l could be a novel player during chromatin
remodeling during early embryogenesis, Chd1l was knocked-down using synthetic,
antisense oligos called morpholinos (MOs). The morpholino system has the advantage of
being able to do rapid reverse genetic studies in a cell type where maternal effect proteins
usually mask the phenotypes of typical knock-out mice.
Microinjection of MO targeting Chd1l in pronuclear zygote stage embryos
resulted in arrest of the embryo. Several MOs were used, all of which caused arrest prior
to the formation of the blastocyst, some as early as the two cell stage. Knock-down was
confirmed at the transcript level by measuring the abundance of Chd1l transcripts that
were abrogated at the splice junction targeted by the MO. The early arrest phenotype was
partially rescued by co-injection of MO with mRNA encoding Chd1l. Because cDNA
sequences lack introns, the mRNA was not targeted by the MO. Embryos injected with
MO that result in a very early arrest between the two and four cell stage progressed
through several more cell divisions with co-expression of Chd1l mRNA, some of them
obtaining blastocyst formation. These results show that Chd1l is required for an early
chromatin remodeling even during preimplantation development.
The reason behind the early arrest remains unknown. Knock-down of Chd1l in
ES cells followed by genome-wide expression profiling revealed no statistically
significant gene expression changes, and therefore could not provide any clue as to a
Chd1l function in the embryo. Gene expression profiling in preimplantation embryos in
which Chd1l has been knocked-down has not been attempted but may reveal changes in
expression of genes involved in key regulatory processes. Alternatively, the function of
Chd1l may not be at the transcriptional level. If Chd1l participates in the DNA damage
response in embryos as it does in cultured cell types, its mode of action could be entirely
106
post-translational, dealing exclusively with altering chromatin accessibility for DNA
repair factors.
One intriguing possibility is that Chd1l regulates parental genome
reprogramming. Upon fertilization, the sperm chromatin decondenses and undergoes
extensive and rapid chromatin-dependent reprogramming [143]. DNA methylation is
actively and globally reduced except at imprinted loci [163], and protamines are removed
and replaced with replication-independent histone variant H3.3, acetylated histone H4,
and histone H4 monomethylated at lysine 20 [157]. This process is completed within
eight hours after fertilization, just prior to the formation of pronuclear membranes. In
addition, the paternal genome must recover from its “frozen state” and repair endogenous
DNA damage accrued prior to fertilization and/or upon decondensation in the zygotic
environment.
DNA demethylation has been studied for decades, but despite the molecular
pathway being well characterized in plants, it remains ill-defined and nebulous in animals
[166,167]. It was recently shown that 5-meC demethylation occurs through the base
excision repair (BER) pathway in preimplantation embryos [176]. Because Chd1l is
known to interact with BER and DSB repair proteins and to respond to DNA damage by
localizing to lesions [259,260], its role in facilitating DNA demethylation represents an
attractive hypothesis. However, given the timing of embryo microinjection at
approximately 10 hpf, the Chd1l arrest phenotype is probably not consistent with a role
for Chd1l in active DNA demethylation, which is already complete by 8hpf [162,163].
For Chd1l to have a role in active DNA demethylation, a significant delay would have to
occur between excision of 5-meC and repair of the abasic site.
X-inactivation and reactivation events are also important features of embryonic
reprogramming. The parental X chromosome is delivered to the zygote in the inactive
form and is re-activated upon initial sperm remodeling, supposedly as part of the global
demethylation process although this has not been determined [23,222]. The Xp
undergoes imprinted XCI at the four cell stage and remains inactive in all cells of the
cleavage-stage embryo and the trophectoderm upon blastocyst formation [221,223,233].
The early ICM, however, specifically reactivates Xp and initiates random XCI at the peri-
implantation stage [223]. The timing of the Chd1l arrest phenotype (as early as the two
107
cell stage and as late as morula stage) would not be consistent with any XCI (four cell
stage or peri-implantation) or X reactivation (pre-pronuclear formation or early ICM
formation) event.
In contrast, a role of Chd1l in the repair of paternal DNA damage is a temporally
relevant hypothesis. Repair of damaged paternal DNA (and maternal DNA) occurs
during the extended ~20-24 hours of the first cell cycle and during sequent cleavages
[240,336]. Failure to repair paternal chromatin in the zygote is a leading cause of human
infertility [237,251,310,341]. Irradiation of zygote stage embryos results in chromosomal
abnormalities and arrest prior to the blastocyst stage, demonstrating that the presence of
DNA damage in the zygote is deleterious to the embryo, but that development may
progress through several cell cycles before arrest occurs [241,242]. This phenotype,
although non-specific, resembles that seen in embryos microinjected with Chd1l-MO.
The zygote is distinctive from other embryonic cell types in that it has an extended G1
phase in which NHEJ predominates [336]. The potential of Chd1l to contribute to NHEJ
is supported by the observation that upon induced DNA damage in HEK293 cells, Chd1l
associates with DNA-PKcs, a major component of NHEJ repair [259]. Chd1l may also
contribute to other modes of DNA repair, including HRR, which functions during S/G2
phases, and BER, although the evidence for these pathways is not as strong as for NHEJ.
If a Chd1l arrest phenotype is due to impaired DNA repair, particularly repair of
DSBs, it would be expected that arrested embryos would display an accumulation of
DSBs. Immunocytochemistry staining of control embryos and embryos injected with
Chd1l-MO using γH2AX as a marker for DSBs revealed that some, but not all, cells of
arrested embryos had a higher frequency of DSBs. Further analysis is necessary to
analyze the presence of DSBs in Chd1l arrested embryos, although this is not an easy
task. Single-cell gel electrophoresis, or the comet assay, is often used in cultured cells to
detect double- and single-stranded breaks, but even with ample material, error values are
extremely high. Novel in vitro approaches may be necessary to test repair of lesions in
control embryos and Chd1l-arrested embryos.
Another temporally viable hypothesis is that Chd1l contributes to the maternal-
zygotic transition through chromatin remodeling leading to zygotic genome activation.
The primary events that trigger ZGA are still under investigation, but initiation of
108
expression of at least three prominent classes of genes is known to require chromatin
remodeling by Brg1 [186]. The foundation for this second hypothesis is tenuous because
there is no current evidence to suggest that Chd1l regulates gene expression through
chromatin remodeling. Expression analysis in control embryos and embryos injected
with Chd1l-MO should reveal whether Chd1l-mediated gene regulation is a feasible
hypothesis.
Due to the inherent difficulties associated with embryo microinjection, namely,
physical insult to the embryo and low statistical power, mutant mouse models should be
created to corroborate the findings presented here. Because maternal contribution is
likely to be a factor, a strategy that allows conditional knock-out in the zygote should be
employed. With the use of Chd1l-null zygotes, the critical window of time in which
Chd1l can be more accurately defined and levels of PAR or γH2AX can be more reliably
assessed.
Chd1l is Non-Essential in ES cells
ICM enrichment and expression in ES cells suggest that an essential role for
Chd1l could be found in ES cell pluripotency or subsequent differentiation. To address
this question, inducible ES lines were created that expresses a shRNA when tetracycline
is removed from culturing medium. cDNA encoding a shRNA targeting Chd1l was
stably integrated into the Rosa locus to create the Chd1l-shRNA EBRTcH3 ES cell line.
As a control, cDNA encoding a “non-silencing” shRNA was also integrated to create the
NS-shRNA eBRTcH3 ES cell line.
Chd1l-shRNA ES cells cultured in the absence of tetracycline for two days show
efficient knock-down of Chd1l. There were no appreciable differences in growth rates
compared to induced NS-shRNA ES cells over a period of eight days, and no
morphological differences in the formation of colonies grown for six days. To address
whether gene expression patterns were altered even in the absence of a readily apparent
phenotype, the expression profiles of ES cells expressing either Chd1l-shRNA or NS-
shRNA were examined. Surprisingly, there were no significant changes in gene
expression due to knock-down of Chd1l. While Chd1l may not function to regulate gene
expression in pluripotent ES cells, it may be critical for regulating gene expression
109
patterns during differentiation. To test this, ES cells were used to form into embryoid
bodies (EBs), differentiating aggregates of cells that mimic embryonic development.
EBs expressing Chdl1-shRNA or NS-shRNA were analyzed for the formation of the
three primary germ layers, endoderm, mesoderm, and ectoderm. Quantitative PCR using
a panel of lineage markers revealed that all three germ layers formed in EBs expressing
Chd1l-shRNA, similar to EBs expressing NS-shRNA.
Protein analysis of ES cells expressing Chd1l-shRNA reveals that knock-down
was very efficient. Although it remains possible that a small amount of Chd1l remaining
is sufficient to carry out the normal Chd1l functions in ES cells, the data indicate that
Chd1l is non-essential for ES cell viability, proliferation, pluripotency, and
differentiation. Furthermore, genome-wide chromatin immunoprecipitation (ChIP) failed
to identify genomic loci bound by Chd1l. The finding that the α-Chd1l antibody did not
pull down any direct Chd1l targets has two explanations. The first is that Chd1l does not
bind discrete genomic loci, and the second is that the α-Chd1l antibody did not
immunoprecipitate cross-linked Chd1l. Because no direct Chd1l targets have been
identified in any cell type have, validation of the α-Chd1l antibody was impossible.
Given that gene expression does not change upon Chd1l knock-down in ES cells, it is
likely that Chd1l also does not bind discrete loci.
After completion of these studies, two independent laboratories reported their
findings that Chd1l is involved in the DNA damage response. If Chd1l also functions in
the DNA damage response ES cells, the inability to identify Chd1l target genes in ES
cells could be explained by semi-random binding to sites of endogenous DNA damage
rather than discrete gene regulatory loci. Prior knowledge of the role of Chd1l in DNA
damage response would not have precluded investigation into an additional role in gene
expression, because other chromatin remodeling factors, such as INO80, have dual roles
in DNA repair and regulating gene expression. However, given the data presented here
and in the literature, Chd1l appears to have a role in DNA repair but not in transcriptional
regulation.
110
Role of Chd1l in DNA Repair in ES cells
Studies conducted early in this dissertation project demonstrate that over-
expression of Chd1l in ES cells induces widespread cellular death through apoptosis.
Intriguingly, Chd1l-induced cell death was a phenotype specific to undifferentiated ES
cells. ES cells that had been differentiated for two days or more by removal of LIF prior
to Chd1l over-expression were resistant to cell death, and even had increases in cell
number compared to control ES cells. Differentiation mediated by LIF removal is a
challenging task for ES cells, so it is not clear whether the increase in cell number reflects
greater proliferation or increased survival in differentiating cells over-expressing Chd1l.
These data suggest that at some unspecified key point during differentiation, the function
of Chd1l, or the cellular response to Chd1l, changes.
It is worth mentioning that over-expression of wild-type Chd1l or Chd1l
containing a single amino acid mutation predicted to eliminate ATPase activity (K71R)
resulted in nearly indistinguishable phenotypes for all parameters tested, including cell
death, differentiation, proliferation, and apoptosis. Therefore, ES cell death caused by
Chd1l over-expression does not depend on a functional ATPase domain, but may instead
result from the effects of over-expressing the macro domain. This line of research was
not pursued, but it would be interesting to know whether expression of the Chd1l macro
domain alone also triggered apoptosis in ES cells. Over-expressing either form of Chd1l
was predicted to result in a loss of function phenotype due to disrupted stoichiometry of a
Chd1l-containing chromatin remodeling complex. Hypermorphic or neomorphic
phenotypes could result due to over-expression of a macro domain-containing protein, a
possibility made more likely given that Chd1l-induced cell death was independent of the
ATPase domain.
A more reliable loss of function manipulation was pursued by expressing a
shRNA directed against the Chd1l transcript. Unexpectedly, the phenotype of Chd1l
knock-down in ES cells was very different than Chd1l over-expression; that is, there was
no readily apparent phenotype. However, when ES cells expressing Chd1l-shRNA or
NS-shRNA were treated with DNA damaging agents, they displayed damage resistance.
This effect was readily apparent when the radiomimetic drug phleomycin was used, and a
more subtle but reproducible affect was seen when H2O2 treatment was used. These
111
results suggest that Chd1l participates in the decision to induce apoptosis in response to
DNA damage, and that ES cells are more sensitive to radiomimetic damage than
oxidative damage, as has been proposed previously.
The data presented here are consistent with studies demonstrating that Chd1l
participates in the DNA damage response, but how Chd1l responds to DNA damage
appears to be very different in ES cells that readily undergo apoptosis in response to
DNA insults than in other cultured cell types. In human U2OS cells, loss of CHD1L
contributes to damage sensitivity, presumably because cells are deficient in repairing
DNA. In mouse ES cells, loss of Chd1l contributes to damage resistance, presumably
because cells are deficient in initiating apoptosis. The apparent paradox between Chd1l
function in ES cells and in other cell types brought to attention the previous over-
expression studies demonstrating that too much Chd1l induced apoptosis, but only in ES
cells, and not from their differentiating counterparts. Consistent with this, Chd1l over-
expression in 3T3 fibroblasts also does not cause decreased cellular survival. Therefore,
the data show that Chd1l functions very differently in ES cells than in other cell types.
There are many remaining questions regarding the role of Chd1l in DNA repair.
In U2OS cells, CHD1L responds to DNA damage by rapidly localizing to sites of DNA
damage. This localization was shown to depend on PARP1 activation and a functional
macro domain. Chd1l dissociation from damaged sites was assumed to reflect successful
DNA repair and was shown to depend on a functional ATPase domain. However, a
direct role for Chd1l in the repair process was not shown; thus the molecular consequence
of Chdl1 localizing to sites of DNA damage remains unknown.
Because early over-expression studies were not pursued, it remains unknown
whether Chd1l-induced apoptosis in ES cells occurs through a Parp-1- or AIF-dependent
manner. This is a fascinating question because Parp-1 has also been implicated in the
apoptotic damage response through the AIF pathway, but this phenomenon is not well
understood. Mice deficient in Parp-1 through mull mutations show increased damage
resistance after ischemia and reperfusion. This role in mediating apoptosis is unique to
certain cell types and is in contrast the typical role of Parp-1 in responding to DNA
damage by facilitating DNA repair. Why Parp-1 triggers apoptosis in certain cell types
but not in others is unclear but probably has to do with cellular NAD+ levels and specific
112
tolerance to damage. In particular, it is not known how activation of Parp-1 in the
nucleus triggers AIF release from the mitochondrion. That Chd1l participates in the
decision to initiate apoptosis in ES cells but not in other cell types is reminiscent of Parp-
1 function and raises the intriguing possibility that Chd1l participates in the decision to
initiate apoptosis by responding Parp-1-synthesized PAR modifications through the
macro domain. Chd1l may be a piece of the puzzle, but because Chd1l is a nuclear
protein, there would still be a missing link to explain how the Parp-1 signal is transmitted
to the mitochondrion.
In U2OS cells, Chd1l becomes tightly associated with chromatin upon DNA
damage. Localization to lesions depends on the macro domain but not the ATPase
domain. In ES cells, over-expression of either WT or K71R versions of Chd1l, both of
which have a functional macro domain, may have the hypermorphic affect of localizing
to chromatin even in the absence of induced DNA damage. Because cell death induced
by over-expression of Chd1l occurred independently of the ATPase domain, the
mechanisms of Chd1l in initiating apoptosis is independent of nucleosome remodeling
but may still depend on a functional macro domain and localization to chromatin.
Consistent with this hypothesis, ES cells expressing Chd1l-shRNA and treated with DNA
damaging agents may be deficient in initiating apoptosis because Chd1l is not present to
localize to lesions. Thus a model is proposed in which Chd1l responds to DNA damage
in all cell types, but the signal that is propagated by its response is different in ES cells
than in other cell types. Upon DNA damage, Chd1l localizes to lesions. In ES cells that
have high propensity to choose apoptosis, this signal is interpreted as a trigger for cell
death; whereas in other cell types, the signal facilitates DNA repair, most likely through
chromatin relaxation and nucleosome mobilization.
Oncogenic Potential of Chd1l
Given that over-expression of Chd1l induces cell death in ES cells, reports that
Chd1l functions as an oncogene were surprising. The evidence supporting the oncogenic
properties of CHD1L is compelling. CHD1L was identified as a candidate oncogene
because its genomic amplification and over-expression is associated with over 50% of
human hepatocellular carcinoma patients [255]. Transient over-expression of Chd1l in
113
U2OS cells and in primary liver cells increases colony formation in soft agar assays, and
xenografts of CHD1L over-expressing cells form multiple teratomas [261]. Transgenic
mice over-expressing CHD1L develop spontaneous tumors with high frequency (25%),
and MEFs generated from transgenic mice showed decreased levels of the tumor
suppressor proteins p53 and Rb and increased levels of cell cycling proteins Cdk2 and
Cyclin A [262].
Somewhat paradoxically, over-expression of Chd1l in ES cells induces
widespread apoptosis. U2OS cells, when over-expressing CHD1L and upon phleomycin
treatment, accumulate of double-stranded breaks, as measured by γH2AX flow cytometry
and single-cell gel electrophoresis (comet) assays [259]. This result provides a
mechanism by which CHD1L could promote tumorigeneis. Over-expression of some
well characterized oncogenes such as ras, myc, cyclin E, mos, cdc25A, and E2F1 has
been observed to induce DSBs because cells are forced through the cell cycle before
lesions can be repaired [263]. The accumulation of DSBs results in genomic instability,
and when mutations in tumor suppressor genes arise, cancer can develop. Chd1l over-
expression could result in the accumulation of DSBs in ES cells as it does in U2OS cells,
and this could result in genomic instability. Because ES cells have a higher propensity to
undergo apoptosis in response to DNA damage, this genomic instability could trigger
apoptosis in ES cells but lead to oncogenesis in other cell types [116,117,118].
Over-expression of DNA repair genes sometimes confers protection to
cytotoxicity and mutagenic agents; but many times it does not [305,342]. PARP-1 is
among the genes whose over-expression does not. Instead, an increase in sensitivity to γ-
irradiation was observed [343]. Deleterious effects including increased rates of
spontaneous mutations and sensitivity to various DNA damaging agents is observed with
over-expression of the NER repair protein ERCC-1 [344], the BER repair proteins DNA
polβ [345,346] and ANPG (a glycosylase) [347], the HRR repair protein Rad51 [348].
The mechanisms by which over-expression of each of these proteins disrupt repair are
unique and depend on their specific activity during DNA repair. For example, over-
expression of the HRR repair protein Rad51 lead to unusual recombination events
resulting in translocations and genomic instability [348], whereas over-expression of
ANGP is thought to create an excess of AP sites and gapped DNA [347].
114
Therefore, the mechanisms by which over-expression of DNA repair proteins can
lead to increased DNA damage and genomic instability are many, and without further
studies it is difficult to predict a mechanism by which over-expression of Chd1l could
induce genomic instability in ES cells. Chd1l-induced apoptosis does not depend on a
functional ATPase domain, but may be a consequence of specifically over-expressing the
macro domain. Because the macro domain is responsible for localization to chromatin
upon DNA damage it is possible that over-expression of Chd1l has a hypermorphic
effect, causing aberrant localization to chromatin, and that this interferes with DNA
repair or signals to the ES cell apoptosis pathway. Further studies are necessary to test
this hypothesis.
Paradoxes in Chd1l Function in ES Cells
Knock-down of Chd1l in preimplantation embryos results in lethality prior to the
blastocyst stage, demonstrating that Chd1l is essential during early embryogenesis.
Given this result, it was surprising that ES cells in which Chd1l had been efficiently
knocked-down were viable, proliferated normally, did not differentiate, and did not have
altered gene expression patterns. In contrast, over-expression of Chd1l in ES cells
induces apoptosis. Paradoxically, Chd1l over-expression in transgenic mouse models
leads to spontaneous tumor formation in combination with increased levels of some cell
cycling proteins and decreased levels of tumor suppressor genes. Transgenic Chd1l mice
grew normally to term without changes in body size [262].
ES cells expressing Chd1l-shRNA display decreased sensitivity to treatment with
DNA damaging agents H2O2 and especially phleomycin. This result is in sharp contrast
with published studies demonstrating that other cultured cells are increasingly sensitive to
induced DNA damage in cells with reduced Chd1l [259]. It is rather the over-production
of Chd1l in ES cells that experience widespread cell death through apoptosis. These
phenotypes are summarized in Figure 1.
How can these conflicting results be resolved? The common theme from the data
is that results in ES cells conflicts with results in other systems. That early embryos with
reduced Chd1l arrest prior to the blastocyst stage is not in conflict with the role of Chd1l
as an oncogene or as a DNA damage response protein. Therefore, the explanation might
115
be that Chd1l has a very different function in ES cells than it does in any other cell type.
The unique cell cycle, DNA repair strategies, and apoptotic response to DNA damage in
ES cells support this theory. The underlying mechanism could be that Chd1l functions
primarily through NHEJ repair pathway, a pathway that is predominantly used in the
zygote-stage embryo and in differentiated cells. In contrast, pluripotent ES cell primarily
use HR to repair DNA, and are easily triggered to apoptose in response to DNA damage.
The involvement of Chd1l in NHEJ is suggested by its PARP-dependent, DNA damage-
dependent association with a major NHEJ component DNA-PKcs. This represents an
interesting hypothesis that should be tested.
116
Figure 5.1 Chd1l Phenotypes
Reduction of Chd1l in the zygote through morpholino microinjection results in
embryonic arrest prior to the blastocyst stage. What the effects of loss of function are in
the developing post-implantation animal or in the adult remains unknown. Chd1l over-
expression in transgenic mice results in no reported developmental abnormalities, but
adult mice have higher spontaneous tumor formation [262]. In tissue cultured cells, over-
expression of Chd1l and treatment of the radiomimetic drug phleomycin leads to an
accumulation of double-stranded breaks as measured by γH2AX flow cytometry and
comet assays [259]. Over-expression of Chd1l in ES cells leads to widespread cell death
in the absence of induced DNA damage. Apoptosis in response to Chd1l over-expression
is specific to ES cells and does not occur after ES cell differentiation by LIF removal.
Knock-down of Chd1l in ES cells does not result in impaired pluripotency, proliferation,
or differentiation, but confers damage resistance when ES cells are treated with DNA
damaging agents. In contrast, knock-down of Chd1l in differentiated tissue cultured cells
leads to increased hypersensitivity to treatment with DNA damaging agents.
117
REFERENCES
1. Venter JC, Adams MD, Myers EW, Li PW, Mural RJ, et al. (2001) The sequence of
the human genome. Science 291: 1304-1351.
2. Lander ES, Linton LM, Birren B, Nusbaum C, Zody MC, et al. (2001) Initial
sequencing and analysis of the human genome. Nature 409: 860-921.
3. Wilmut I, Schnieke AE, McWhir J, Kind AJ, Campbell KH (1997) Viable offspring
derived from fetal and adult mammalian cells. Nature 385: 810-813.
4. Takahashi K, Yamanaka S (2006) Induction of pluripotent stem cells from mouse
embryonic and adult fibroblast cultures by defined factors. Cell 126: 663-676.
5. Yu J, Vodyanik MA, Smuga-Otto K, Antosiewicz-Bourget J, Frane JL, et al. (2007)
Induced pluripotent stem cell lines derived from human somatic cells. Science
318: 1917-1920.
6. Lindahl T, Barnes DE (2000) Repair of endogenous DNA damage. Cold Spring Harb
Symp Quant Biol 65: 127-133.
7. Kornberg RD (1977) Structure of chromatin. Annu Rev Biochem 46: 931-954.
8. Widom J (1989) Toward a unified model of chromatin folding. Annu Rev Biophys
Biophys Chem 18: 365-395.
9. McGhee JD, Felsenfeld G (1980) Nucleosome structure. Annu Rev Biochem 49: 1115-
1156.
10. Bednar J, Horowitz RA, Grigoryev SA, Carruthers LM, Hansen JC, et al. (1998)
Nucleosomes, linker DNA, and linker histone form a unique structural motif that
directs the higher-order folding and compaction of chromatin. Proc Natl Acad Sci
U S A 95: 14173-14178.
11. Luger K, Mader AW, Richmond RK, Sargent DF, Richmond TJ (1997) Crystal
structure of the nucleosome core particle at 2.8 A resolution. Nature 389: 251-
260.
12. Lorch Y, LaPointe JW, Kornberg RD (1987) Nucleosomes inhibit the initiation of
transcription but allow chain elongation with the displacement of histones. Cell
49: 203-210.
13. Knezetic JA, Luse DS (1986) The presence of nucleosomes on a DNA template
prevents initiation by RNA polymerase II in vitro. Cell 45: 95-104.
14. Iizuka M, Smith MM (2003) Functional consequences of histone modifications. Curr
Opin Genet Dev 13: 154-160.
15. Flaus A, Luger K, Tan S, Richmond TJ (1996) Mapping nucleosome position at
single base-pair resolution by using site-directed hydroxyl radicals. Proc Natl
Acad Sci U S A 93: 1370-1375.
16. Reik W, Dean W, Walter J (2001) Epigenetic reprogramming in mammalian
development. Science 293: 1089-1093.
17. Mohrmann L, Verrijzer CP (2005) Composition and functional specificity of
SWI2/SNF2 class chromatin remodeling complexes. Biochim Biophys Acta 1681:
59-73.
18. Flaus A, Owen-Hughes T (2004) Mechanisms for ATP-dependent chromatin
remodelling: farewell to the tuna-can octamer? Curr Opin Genet Dev 14: 165-173.
19. Bird A (2002) DNA methylation patterns and epigenetic memory. Genes Dev 16: 6-
21.
118
20. Meehan R, Lewis J, Cross S, Nan X, Jeppesen P, et al. (1992) Transcriptional
repression by methylation of CpG. J Cell Sci Suppl 16: 9-14.
21. Li E, Beard C, Jaenisch R (1993) Role for DNA methylation in genomic imprinting.
Nature 366: 362-365.
22. Liskay RM, Evans RJ (1980) Inactive X chromosome DNA does not function in
DNA-mediated cell transformation for the hypoxanthine
phosphoribosyltransferase gene. Proc Natl Acad Sci U S A 77: 4895-4898.
23. Mohandas T, Sparkes RS, Shapiro LJ (1981) Reactivation of an inactive human X
chromosome: evidence for X inactivation by DNA methylation. Science 211:
393-396.
24. Bestor TH (1998) The host defence function of genomic methylation patterns.
Novartis Found Symp 214: 187-195; discussion 195-189, 228-132.
25. Yoder JA, Soman NS, Verdine GL, Bestor TH (1997) DNA (cytosine-5)-
methyltransferases in mouse cells and tissues. Studies with a mechanism-based
probe. J Mol Biol 270: 385-395.
26. Suzuki MM, Bird A (2008) DNA methylation landscapes: provocative insights from
epigenomics. Nat Rev Genet 9: 465-476.
27. Okano M, Bell DW, Haber DA, Li E (1999) DNA methyltransferases Dnmt3a and
Dnmt3b are essential for de novo methylation and mammalian development. Cell
99: 247-257.
28. Strahl BD, Allis CD (2000) The language of covalent histone modifications. Nature
403: 41-45.
29. Berger SL (2007) The complex language of chromatin regulation during
transcription. Nature 447: 407-412.
30. Shi Y, Lan F, Matson C, Mulligan P, Whetstine JR, et al. (2004) Histone
demethylation mediated by the nuclear amine oxidase homolog LSD1. Cell 119:
941-953.
31. Lan F, Nottke AC, Shi Y (2008) Mechanisms involved in the regulation of histone
lysine demethylases. Curr Opin Cell Biol 20: 316-325.
32. D'Amours D, Desnoyers S, D'Silva I, Poirier GG (1999) Poly(ADP-ribosyl)ation
reactions in the regulation of nuclear functions. Biochem J 342 ( Pt 2): 249-268.
33. Ame JC, Spenlehauer C, de Murcia G (2004) The PARP superfamily. Bioessays 26:
882-893.
34. de Murcia G, Menissier de Murcia J (1994) Poly(ADP-ribose) polymerase: a
molecular nick-sensor. Trends Biochem Sci 19: 172-176.
35. Miwa M, Sugimura T (1971) Splitting of the ribose-ribose linkage of poly(adenosine
diphosphate-robose) by a calf thymus extract. J Biol Chem 246: 6362-6364.
36. Juarez-Salinas H, Sims JL, Jacobson MK (1979) Poly(ADP-ribose) levels in
carcinogen-treated cells. Nature 282: 740-741.
37. Berger NA, Sims JL, Catino DM, Berger SJ (1983) Poly(ADP-ribose) polymerase
mediates the suicide response to massive DNA damage: studies in normal and
DNA-repair defective cells. Princess Takamatsu Symp 13: 219-226.
38. Shall S, de Murcia G (2000) Poly(ADP-ribose) polymerase-1: what have we learned
from the deficient mouse model? Mutat Res 460: 1-15.
39. Smith S (2001) The world according to PARP. Trends Biochem Sci 26: 174-179.
119
40. Augustin A, Spenlehauer C, Dumond H, Menissier-De Murcia J, Piel M, et al. (2003)
PARP-3 localizes preferentially to the daughter centriole and interferes with the
G1/S cell cycle progression. J Cell Sci 116: 1551-1562.
41. Smith S, Giriat I, Schmitt A, de Lange T (1998) Tankyrase, a poly(ADP-ribose)
polymerase at human telomeres. Science 282: 1484-1487.
42. Smith S, de Lange T (2000) Tankyrase promotes telomere elongation in human cells.
Curr Biol 10: 1299-1302.
43. Kaminker PG, Kim SH, Taylor RD, Zebarjadian Y, Funk WD, et al. (2001) TANK2,
a new TRF1-associated poly(ADP-ribose) polymerase, causes rapid induction of
cell death upon overexpression. J Biol Chem 276: 35891-35899.
44. Huletsky A, de Murcia G, Muller S, Hengartner M, Menard L, et al. (1989) The effect
of poly(ADP-ribosyl)ation on native and H1-depleted chromatin. A role of
poly(ADP-ribosyl)ation on core nucleosome structure. J Biol Chem 264: 8878-
8886.
45. Rouleau M, Aubin RA, Poirier GG (2004) Poly(ADP-ribosyl)ated chromatin
domains: access granted. J Cell Sci 117: 815-825.
46. Poirier GG, de Murcia G, Jongstra-Bilen J, Niedergang C, Mandel P (1982)
Poly(ADP-ribosyl)ation of polynucleosomes causes relaxation of chromatin
structure. Proc Natl Acad Sci U S A 79: 3423-3427.
47. Quenet D, El Ramy R, Schreiber V, Dantzer F (2009) The role of poly(ADP-
ribosyl)ation in epigenetic events. Int J Biochem Cell Biol 41: 60-65.
48. Kim MY, Mauro S, Gevry N, Lis JT, Kraus WL (2004) NAD+-dependent modulation
of chromatin structure and transcription by nucleosome binding properties of
PARP-1. Cell 119: 803-814.
49. Krishnakumar R, Gamble MJ, Frizzell KM, Berrocal JG, Kininis M, et al. (2008)
Reciprocal binding of PARP-1 and histone H1 at promoters specifies
transcriptional outcomes. Science 319: 819-821.
50. Timinszky G, Till S, Hassa PO, Hothorn M, Kustatscher G, et al. (2009) A
macrodomain-containing histone rearranges chromatin upon sensing PARP1
activation. Nat Struct Mol Biol 16: 923-929.
51. El-Khamisy SF, Masutani M, Suzuki H, Caldecott KW (2003) A requirement for
PARP-1 for the assembly or stability of XRCC1 nuclear foci at sites of oxidative
DNA damage. Nucleic Acids Res 31: 5526-5533.
52. de Murcia JM, Niedergang C, Trucco C, Ricoul M, Dutrillaux B, et al. (1997)
Requirement of poly(ADP-ribose) polymerase in recovery from DNA damage in
mice and in cells. Proc Natl Acad Sci U S A 94: 7303-7307.
53. Wang ZQ, Stingl L, Morrison C, Jantsch M, Los M, et al. (1997) PARP is important
for genomic stability but dispensable in apoptosis. Genes Dev 11: 2347-2358.
54. Wang ZQ, Auer B, Stingl L, Berghammer H, Haidacher D, et al. (1995) Mice lacking
ADPRT and poly(ADP-ribosyl)ation develop normally but are susceptible to skin
disease. Genes Dev 9: 509-520.
55. Shieh WM, Ame JC, Wilson MV, Wang ZQ, Koh DW, et al. (1998) Poly(ADP-
ribose) polymerase null mouse cells synthesize ADP-ribose polymers. J Biol
Chem 273: 30069-30072.
120
56. Ame JC, Rolli V, Schreiber V, Niedergang C, Apiou F, et al. (1999) PARP-2, A
novel mammalian DNA damage-dependent poly(ADP-ribose) polymerase. J Biol
Chem 274: 17860-17868.
57. Menissier de Murcia J, Ricoul M, Tartier L, Niedergang C, Huber A, et al. (2003)
Functional interaction between PARP-1 and PARP-2 in chromosome stability and
embryonic development in mouse. Embo J 22: 2255-2263.
58. Dantzer F, Giraud-Panis MJ, Jaco I, Ame JC, Schultz I, et al. (2004) Functional
interaction between poly(ADP-Ribose) polymerase 2 (PARP-2) and TRF2: PARP
activity negatively regulates TRF2. Mol Cell Biol 24: 1595-1607.
59. Huber A, Bai P, de Murcia JM, de Murcia G (2004) PARP-1, PARP-2 and ATM in
the DNA damage response: functional synergy in mouse development. DNA
Repair (Amst) 3: 1103-1108.
60. Allen MD, Buckle AM, Cordell SC, Lowe J, Bycroft M (2003) The crystal structure
of AF1521 a protein from Archaeoglobus fulgidus with homology to the non-
histone domain of macroH2A. J Mol Biol 330: 503-511.
61. Karras GI, Kustatscher G, Buhecha HR, Allen MD, Pugieux C, et al. (2005) The
macro domain is an ADP-ribose binding module. Embo J 24: 1911-1920.
62. Costanzi C, Pehrson JR (1998) Histone macroH2A1 is concentrated in the inactive X
chromosome of female mammals. Nature 393: 599-601.
63. Szabo C, Dawson VL (1998) Role of poly(ADP-ribose) synthetase in inflammation
and ischaemia-reperfusion. Trends Pharmacol Sci 19: 287-298.
64. Masutani M, Suzuki H, Kamada N, Watanabe M, Ueda O, et al. (1999) Poly(ADP-
ribose) polymerase gene disruption conferred mice resistant to streptozotocin-
induced diabetes. Proc Natl Acad Sci U S A 96: 2301-2304.
65. Yu SW, Wang H, Poitras MF, Coombs C, Bowers WJ, et al. (2002) Mediation of
poly(ADP-ribose) polymerase-1-dependent cell death by apoptosis-inducing
factor. Science 297: 259-263.
66. Wang W, Xue Y, Zhou S, Kuo A, Cairns BR, et al. (1996) Diversity and
specialization of mammalian SWI/SNF complexes. Genes Dev 10: 2117-2130.
67. Wu JI, Lessard J, Crabtree GR (2009) Understanding the words of chromatin
regulation. Cell 136: 200-206.
68. Cairns BR (2005) Chromatin remodeling complexes: strength in diversity, precision
through specialization. Curr Opin Genet Dev 15: 185-190.
69. Lessard J, Wu JI, Ranish JA, Wan M, Winslow MM, et al. (2007) An essential switch
in subunit composition of a chromatin remodeling complex during neural
development. Neuron 55: 201-215.
70. Zhang B, Chambers KJ, Faller DV, Wang S (2007) Reprogramming of the SWI/SNF
complex for co-activation or co-repression in prohibitin-mediated estrogen
receptor regulation. Oncogene 26: 7153-7157.
71. Chi TH, Wan M, Lee PP, Akashi K, Metzger D, et al. (2003) Sequential roles of Brg,
the ATPase subunit of BAF chromatin remodeling complexes, in thymocyte
development. Immunity 19: 169-182.
72. Trotter KW, Archer TK (2008) The BRG1 transcriptional coregulator. Nucl Recept
Signal 6: e004.
73. Saha A, Wittmeyer J, Cairns BR (2002) Chromatin remodeling by RSC involves
ATP-dependent DNA translocation. Genes Dev 16: 2120-2134.
121
74. Whitehouse I, Stockdale C, Flaus A, Szczelkun MD, Owen-Hughes T (2003)
Evidence for DNA translocation by the ISWI chromatin-remodeling enzyme. Mol
Cell Biol 23: 1935-1945.
75. Flaus A, Owen-Hughes T (2001) Mechanisms for ATP-dependent chromatin
remodelling. Curr Opin Genet Dev 11: 148-154.
76. Cote J, Quinn J, Workman JL, Peterson CL (1994) Stimulation of GAL4 derivative
binding to nucleosomal DNA by the yeast SWI/SNF complex. Science 265: 53-
60.
77. Havas K, Flaus A, Phelan M, Kingston R, Wade PA, et al. (2000) Generation of
superhelical torsion by ATP-dependent chromatin remodeling activities. Cell 103:
1133-1142.
78. Lia G, Praly E, Ferreira H, Stockdale C, Tse-Dinh YC, et al. (2006) Direct
observation of DNA distortion by the RSC complex. Mol Cell 21: 417-425.
79. Van Komen S, Petukhova G, Sigurdsson S, Stratton S, Sung P (2000) Superhelicity-
driven homologous DNA pairing by yeast recombination factors Rad51 and
Rad54. Mol Cell 6: 563-572.
80. Yu S, Owen-Hughes T, Friedberg EC, Waters R, Reed SH (2004) The yeast
Rad7/Rad16/Abf1 complex generates superhelical torsion in DNA that is required
for nucleotide excision repair. DNA Repair (Amst) 3: 277-287.
81. Flaus A, Martin DM, Barton GJ, Owen-Hughes T (2006) Identification of multiple
distinct Snf2 subfamilies with conserved structural motifs. Nucleic Acids Res 34:
2887-2905.
82. Peterson CL, Workman JL (2000) Promoter targeting and chromatin remodeling by
the SWI/SNF complex. Curr Opin Genet Dev 10: 187-192.
83. Neigeborn L, Carlson M (1984) Genes affecting the regulation of SUC2 gene
expression by glucose repression in Saccharomyces cerevisiae. Genetics 108:
845-858.
84. Abrams E, Neigeborn L, Carlson M (1986) Molecular analysis of SNF2 and SNF5,
genes required for expression of glucose-repressible genes in Saccharomyces
cerevisiae. Mol Cell Biol 6: 3643-3651.
85. Landry J, Sharov AA, Piao Y, Sharova LV, Xiao H, et al. (2008) Essential role of
chromatin remodeling protein Bptf in early mouse embryos and embryonic stem
cells. PLoS Genet 4: e1000241.
86. Lazzaro MA, Picketts DJ (2001) Cloning and characterization of the murine Imitation
Switch (ISWI) genes: differential expression patterns suggest distinct
developmental roles for Snf2h and Snf2l. J Neurochem 77: 1145-1156.
87. Banting GS, Barak O, Ames TM, Burnham AC, Kardel MD, et al. (2005) CECR2, a
protein involved in neurulation, forms a novel chromatin remodeling complex
with SNF2L. Hum Mol Genet 14: 513-524.
88. Vissers LE, van Ravenswaaij CM, Admiraal R, Hurst JA, de Vries BB, et al. (2004)
Mutations in a new member of the chromodomain gene family cause CHARGE
syndrome. Nat Genet 36: 955-957.
89. Schnetz MP, Bartels CF, Shastri K, Balasubramanian D, Zentner GE, et al. (2009)
Genomic distribution of CHD7 on chromatin tracks H3K4 methylation patterns.
Genome Res 19: 590-601.
122
90. Conaway RC, Conaway JW (2009) The INO80 chromatin remodeling complex in
transcription, replication and repair. Trends Biochem Sci 34: 71-77.
91. Mizuguchi G, Shen X, Landry J, Wu WH, Sen S, et al. (2004) ATP-driven exchange
of histone H2AZ variant catalyzed by SWR1 chromatin remodeling complex.
Science 303: 343-348.
92. Morrison AJ, Highland J, Krogan NJ, Arbel-Eden A, Greenblatt JF, et al. (2004)
INO80 and gamma-H2AX interaction links ATP-dependent chromatin
remodeling to DNA damage repair. Cell 119: 767-775.
93. van Attikum H, Fritsch O, Hohn B, Gasser SM (2004) Recruitment of the INO80
complex by H2A phosphorylation links ATP-dependent chromatin remodeling
with DNA double-strand break repair. Cell 119: 777-788.
94. Fyodorov DV, Kadonaga JT (2001) The many faces of chromatin remodeling:
SWItching beyond transcription. Cell 106: 523-525.
95. Brook FA, Gardner RL (1997) The origin and efficient derivation of embryonic stem
cells in the mouse. Proc Natl Acad Sci U S A 94: 5709-5712.
96. Keller G (2005) Embryonic stem cell differentiation: emergence of a new era in
biology and medicine. Genes Dev 19: 1129-1155.
97. Boyer LA, Lee TI, Cole MF, Johnstone SE, Levine SS, et al. (2005) Core
transcriptional regulatory circuitry in human embryonic stem cells. Cell 122: 947-
956.
98. Loh YH, Wu Q, Chew JL, Vega VB, Zhang W, et al. (2006) The Oct4 and Nanog
transcription network regulates pluripotency in mouse embryonic stem cells. Nat
Genet 38: 431-440.
99. Azuara V, Perry P, Sauer S, Spivakov M, Jorgensen HF, et al. (2006) Chromatin
signatures of pluripotent cell lines. Nat Cell Biol 8: 532-538.
100. Meshorer E, Yellajoshula D, George E, Scambler PJ, Brown DT, et al. (2006)
Hyperdynamic plasticity of chromatin proteins in pluripotent embryonic stem
cells. Dev Cell 10: 105-116.
101. Bernstein BE, Mikkelsen TS, Xie X, Kamal M, Huebert DJ, et al. (2006) A bivalent
chromatin structure marks key developmental genes in embryonic stem cells. Cell
125: 315-326.
102. Niwa H (2007) Open conformation chromatin and pluripotency. Genes Dev 21:
2671-2676.
103. Loh YH, Zhang W, Chen X, George J, Ng HH (2007) Jmjd1a and Jmjd2c histone
H3 Lys 9 demethylases regulate self-renewal in embryonic stem cells. Genes Dev
21: 2545-2557.
104. Briggs R, King TJ (1952) Transplantation of Living Nuclei From Blastula Cells into
Enucleated Frogs' Eggs. Proc Natl Acad Sci U S A 38: 455-463.
105. Wernig M, Meissner A, Foreman R, Brambrink T, Ku M, et al. (2007) In vitro
reprogramming of fibroblasts into a pluripotent ES-cell-like state. Nature 448:
318-324.
106. Okita K, Ichisaka T, Yamanaka S (2007) Generation of germline-competent induced
pluripotent stem cells. Nature 448: 313-317.
107. Yamanaka S, Blau HM (2010) Nuclear reprogramming to a pluripotent state by
three approaches. Nature 465: 704-712.
123
108. Hong Y, Stambrook PJ (2004) Restoration of an absent G1 arrest and protection
from apoptosis in embryonic stem cells after ionizing radiation. Proc Natl Acad
Sci U S A 101: 14443-14448.
109. Tichy ED, Stambrook PJ (2008) DNA repair in murine embryonic stem cells and
differentiated cells. Exp Cell Res 314: 1929-1936.
110. Hong Y, Cervantes RB, Tichy E, Tischfield JA, Stambrook PJ (2007) Protecting
genomic integrity in somatic cells and embryonic stem cells. Mutat Res 614: 48-
55.
111. Stambrook PJ, Shao C, Stockelman M, Boivin G, Engle SJ, et al. (1996) APRT: a
versatile in vivo resident reporter of local mutation and loss of heterozygosity.
Environ Mol Mutagen 28: 471-482.
112. Cervantes RB, Stringer JR, Shao C, Tischfield JA, Stambrook PJ (2002) Embryonic
stem cells and somatic cells differ in mutation frequency and type. Proc Natl Acad
Sci U S A 99: 3586-3590.
113. Larson JS, Yin M, Fischer JM, Stringer SL, Stringer JR (2006) Expression and loss
of alleles in cultured mouse embryonic fibroblasts and stem cells carrying allelic
fluorescent protein genes. BMC Mol Biol 7: 36.
114. Chen Y, Yee D, Dains K, Chatterjee A, Cavalcoli J, et al. (2000) Genotype-based
screen for ENU-induced mutations in mouse embryonic stem cells. Nat Genet 24:
314-317.
115. Munroe RJ, Bergstrom RA, Zheng QY, Libby B, Smith R, et al. (2000) Mouse
mutants from chemically mutagenized embryonic stem cells. Nat Genet 24: 318-
321.
116. de Waard H, de Wit J, Gorgels TG, van den Aardweg G, Andressoo JO, et al. (2003)
Cell type-specific hypersensitivity to oxidative damage in CSB and XPA mice.
DNA Repair (Amst) 2: 13-25.
117. Van Sloun PP, Jansen JG, Weeda G, Mullenders LH, van Zeeland AA, et al. (1999)
The role of nucleotide excision repair in protecting embryonic stem cells from
genotoxic effects of UV-induced DNA damage. Nucleic Acids Res 27: 3276-
3282.
118. Roos WP, Christmann M, Fraser ST, Kaina B (2007) Mouse embryonic stem cells
are hypersensitive to apoptosis triggered by the DNA damage O(6)-
methylguanine due to high E2F1 regulated mismatch repair. Cell Death Differ 14:
1422-1432.
119. Aladjem MI, Spike BT, Rodewald LW, Hope TJ, Klemm M, et al. (1998) ES cells
do not activate p53-dependent stress responses and undergo p53-independent
apoptosis in response to DNA damage. Curr Biol 8: 145-155.
120. Stead E, White J, Faast R, Conn S, Goldstone S, et al. (2002) Pluripotent cell
division cycles are driven by ectopic Cdk2, cyclin A/E and E2F activities.
Oncogene 21: 8320-8333.
121. Saretzki G, Armstrong L, Leake A, Lako M, von Zglinicki T (2004) Stress defense
in murine embryonic stem cells is superior to that of various differentiated murine
cells. Stem Cells 22: 962-971.
122. Kanaar R, Hoeijmakers JH, van Gent DC (1998) Molecular mechanisms of DNA
double strand break repair. Trends Cell Biol 8: 483-489.
124
123. Takata M, Sasaki MS, Sonoda E, Morrison C, Hashimoto M, et al. (1998)
Homologous recombination and non-homologous end-joining pathways of DNA
double-strand break repair have overlapping roles in the maintenance of
chromosomal integrity in vertebrate cells. EMBO J 17: 5497-5508.
124. Haber JE (2000) Partners and pathwaysrepairing a double-strand break. Trends
Genet 16: 259-264.
125. Morrison C, Sonoda E, Takao N, Shinohara A, Yamamoto K, et al. (2000) The
controlling role of ATM in homologous recombinational repair of DNA damage.
EMBO J 19: 463-471.
126. Savatier P, Lapillonne H, Jirmanova L, Vitelli L, Samarut J (2002) Analysis of the
cell cycle in mouse embryonic stem cells. Methods Mol Biol 185: 27-33.
127. Richardson C, Jasin M (2000) Frequent chromosomal translocations induced by
DNA double-strand breaks. Nature 405: 697-700.
128. Boubakour-Azzouz I, Ricchetti M (2008) Low joining efficiency and non-
conservative repair of two distant double-strand breaks in mouse embryonic stem
cells. DNA Repair (Amst) 7: 149-161.
129. Allen C, Miller CA, Nickoloff JA (2003) The mutagenic potential of a single DNA
double-strand break in a mammalian chromosome is not influenced by
transcription. DNA Repair (Amst) 2: 1147-1156.
130. Griffin C, Waard H, Deans B, Thacker J (2005) The involvement of key DNA repair
pathways in the formation of chromosome rearrangements in embryonic stem
cells. DNA Repair (Amst) 4: 1019-1027.
131. Orii KE, Lee Y, Kondo N, McKinnon PJ (2006) Selective utilization of
nonhomologous end-joining and homologous recombination DNA repair
pathways during nervous system development. Proc Natl Acad Sci U S A 103:
10017-10022.
132. Pierce AJ, Jasin M (2005) Measuring recombination proficiency in mouse
embryonic stem cells. Methods Mol Biol 291: 373-384.
133. Tichy ED, Pillai R, Deng L, Liang L, Tischfield JA, et al. (2010) Mouse embryonic
stem cells but not somatic cells predominantly use homologous recombination to
repair double strand DNA breaks. Stem Cells Dev.
134. Serrano L, Liang L, Chang Y, Deng L, Maulion C, et al. (2010) Homologous
Recombination conserves DNA sequence integrity throughout the cell cycle in
embryonic stem cells. Stem Cells Dev.
135. Marti TM, Kunz C, Fleck O (2002) DNA mismatch repair and mutation avoidance
pathways. J Cell Physiol 191: 28-41.
136. Peltomaki P (2001) DNA mismatch repair and cancer. Mutat Res 488: 77-85.
137. Borgdorff V, Pauw B, van Hees-Stuivenberg S, de Wind N (2006) DNA mismatch
repair mediates protection from mutagenesis induced by short-wave ultraviolet
light. DNA Repair (Amst) 5: 1364-1372.
138. Meyers M, Wagner MW, Mazurek A, Schmutte C, Fishel R, et al. (2005) DNA
mismatch repair-dependent response to fluoropyrimidine-generated damage. J
Biol Chem 280: 5516-5526.
139. Meyers M, Wagner MW, Hwang HS, Kinsella TJ, Boothman DA (2001) Role of the
hMLH1 DNA mismatch repair protein in fluoropyrimidine-mediated cell death
and cell cycle responses. Cancer Res 61: 5193-5201.
125
140. Engelward BP, Dreslin A, Christensen J, Huszar D, Kurahara C, et al. (1996)
Repair-deficient 3-methyladenine DNA glycosylase homozygous mutant mouse
cells have increased sensitivity to alkylation-induced chromosome damage and
cell killing. EMBO J 15: 945-952.
141. Hayashi K, de Sousa Lopes SM, Surani MA (2007) Germ cell specification in mice.
Science 316: 394-396.
142. Lucifero D, Mertineit C, Clarke HJ, Bestor TH, Trasler JM (2002) Methylation
dynamics of imprinted genes in mouse germ cells. Genomics 79: 530-538.
143. Miller D, Brinkworth M, Iles D (2010) Paternal DNA packaging in spermatozoa:
more than the sum of its parts? DNA, histones, protamines and epigenetics.
Reproduction 139: 287-301.
144. Ward WS, Coffey DS (1991) DNA packaging and organization in mammalian
spermatozoa: comparison with somatic cells. Biol Reprod 44: 569-574.
145. Balhorn R, Gledhill BL, Wyrobek AJ (1977) Mouse sperm chromatin proteins:
quantitative isolation and partial characterization. Biochemistry 16: 4074-4080.
146. Braun RE (2001) Packaging paternal chromosomes with protamine. Nat Genet 28:
10-12.
147. de Boer P, Redi CA, Garagna S, Winking H (1990) Protamine amount and cross
linking in mouse teratospermatozoa and aneuploid spermatozoa. Mol Reprod Dev
25: 297-301.
148. Cho C, Willis WD, Goulding EH, Jung-Ha H, Choi YC, et al. (2001)
Haploinsufficiency of protamine-1 or -2 causes infertility in mice. Nat Genet 28:
82-86.
149. Oliva R (2006) Protamines and male infertility. Hum Reprod Update 12: 417-435.
150. Aoki VW, Emery BR, Liu L, Carrell DT (2006) Protamine levels vary between
individual sperm cells of infertile human males and correlate with viability and
DNA integrity. J Androl 27: 890-898.
151. Moore GP (1975) The RNA polymerase activity of the preimplantation mouse
embryo. J Embryol Exp Morphol 34: 291-298.
152. Schultz GA (1986) Molecular biology of the early mouse embryo. Biological
Bulletin 171: 291-309.
153. Whitaker M (2006) Calcium at fertilization and in early development. Physiol Rev
86: 25-88.
154. Niemann H, Tian XC, King WA, Lee RS (2008) Epigenetic reprogramming in
embryonic and foetal development upon somatic cell nuclear transfer cloning.
Reproduction 135: 151-163.
155. Palermo GD, Colombero LT, Rosenwaks Z (1997) The human sperm centrosome is
responsible for normal syngamy and early embryonic development. Rev Reprod
2: 19-27.
156. Nonchev S, Tsanev R (1990) Protamine-histone replacement and DNA replication
in the male mouse pronucleus. Mol Reprod Dev 25: 72-76.
157. Adenot PG, Mercier Y, Renard JP, Thompson EM (1997) Differential H4
acetylation of paternal and maternal chromatin precedes DNA replication and
differential transcriptional activity in pronuclei of 1-cell mouse embryos.
Development 124: 4615-4625.
126
158. Santos F, Hendrich B, Reik W, Dean W (2002) Dynamic reprogramming of DNA
methylation in the early mouse embryo. Dev Biol 241: 172-182.
159. van der Heijden GW, Dieker JW, Derijck AA, Muller S, Berden JH, et al. (2005)
Asymmetry in histone H3 variants and lysine methylation between paternal and
maternal chromatin of the early mouse zygote. Mech Dev 122: 1008-1022.
160. Torres-Padilla ME, Bannister AJ, Hurd PJ, Kouzarides T, Zernicka-Goetz M (2006)
Dynamic distribution of the replacement histone variant H3.3 in the mouse oocyte
and preimplantation embryos. Int J Dev Biol 50: 455-461.
161. Orsi GA, Couble P, Loppin B (2009) Epigenetic and replacement roles of histone
variant H3.3 in reproduction and development. Int J Dev Biol 53: 231-243.
162. Oswald J, Engemann S, Lane N, Mayer W, Olek A, et al. (2000) Active
demethylation of the paternal genome in the mouse zygote. Curr Biol 10: 475-
478.
163. Mayer W, Niveleau A, Walter J, Fundele R, Haaf T (2000) Demethylation of the
zygotic paternal genome. Nature 403: 501-502.
164. Carlson LL, Page AW, Bestor TH (1992) Properties and localization of DNA
methyltransferase in preimplantation mouse embryos: implications for genomic
imprinting. Genes Dev 6: 2536-2541.
165. Monk M, Adams RL, Rinaldi A (1991) Decrease in DNA methylase activity during
preimplantation development in the mouse. Development 112: 189-192.
166. Gehring M, Huh JH, Hsieh TF, Penterman J, Choi Y, et al. (2006) DEMETER DNA
glycosylase establishes MEDEA polycomb gene self-imprinting by allele-specific
demethylation. Cell 124: 495-506.
167. Penterman J, Zilberman D, Huh JH, Ballinger T, Henikoff S, et al. (2007) DNA
demethylation in the Arabidopsis genome. Proc Natl Acad Sci U S A 104: 6752-
6757.
168. Weiss A, Keshet I, Razin A, Cedar H (1996) DNA demethylation in vitro:
involvement of RNA. Cell 86: 709-718.
169. Gjerset RA, Martin DW, Jr. (1982) Presence of a DNA demethylating activity in the
nucleus of murine erythroleukemic cells. J Biol Chem 257: 8581-8583.
170. Jost JP (1993) Nuclear extracts of chicken embryos promote an active demethylation
of DNA by excision repair of 5-methyldeoxycytidine. Proc Natl Acad Sci U S A
90: 4684-4688.
171. Bhattacharya SK, Ramchandani S, Cervoni N, Szyf M (1999) A mammalian protein
with specific demethylase activity for mCpG DNA. Nature 397: 579-583.
172. Barreto G, Schafer A, Marhold J, Stach D, Swaminathan SK, et al. (2007) Gadd45a
promotes epigenetic gene activation by repair-mediated DNA demethylation.
Nature 445: 671-675.
173. Kangaspeska S, Stride B, Metivier R, Polycarpou-Schwarz M, Ibberson D, et al.
(2008) Transient cyclical methylation of promoter DNA. Nature 452: 112-115.
174. Metivier R, Gallais R, Tiffoche C, Le Peron C, Jurkowska RZ, et al. (2008) Cyclical
DNA methylation of a transcriptionally active promoter. Nature 452: 45-50.
175. Ooi SK, Bestor TH (2008) The colorful history of active DNA demethylation. Cell
133: 1145-1148.
127
176. Hajkova P, Jeffries SJ, Lee C, Miller N, Jackson SP, et al. (2010) Genome-wide
reprogramming in the mouse germ line entails the base excision repair pathway.
Science 329: 78-82.
177. Shibata H, Yoda Y, Kato R, Ueda T, Kamiya M, et al. (1998) A methylation imprint
mark in the mouse imprinted gene Grf1/Cdc25Mm locus shares a common feature
with the U2afbp-rs gene: an association with a short tandem repeat and a
hypermethylated region. Genomics 49: 30-37.
178. Bachvarova R (1985) Gene expression during oogenesis and oocyte development in
mammals. Dev Biol (N Y 1985) 1: 453-524.
179. Schier AF (2007) The maternal-zygotic transition: death and birth of RNAs. Science
316: 406-407.
180. Tadros W, Lipshitz HD (2005) Setting the stage for development: mRNA translation
and stability during oocyte maturation and egg activation in Drosophila. Dev Dyn
232: 593-608.
181. Giraldez AJ, Mishima Y, Rihel J, Grocock RJ, Van Dongen S, et al. (2006)
Zebrafish MiR-430 promotes deadenylation and clearance of maternal mRNAs.
Science 312: 75-79.
182. Flach G, Johnson MH, Braude PR, Taylor RA, Bolton VN (1982) The transition
from maternal to embryonic control in the 2-cell mouse embryo. EMBO J 1: 681-
686.
183. Telford NA, Watson AJ, Schultz GA (1990) Transition from maternal to embryonic
control in early mammalian development: a comparison of several species. Mol
Reprod Dev 26: 90-100.
184. Ram PT, Schultz RM (1993) Reporter gene expression in G2 of the 1-cell mouse
embryo. Dev Biol 156: 552-556.
185. Hansis C, Barreto G, Maltry N, Niehrs C (2004) Nuclear reprogramming of human
somatic cells by xenopus egg extract requires BRG1. Curr Biol 14: 1475-1480.
186. Bultman S, Gebuhr T, Yee D, La Mantia C, Nicholson J, et al. (2000) A Brg1 null
mutation in the mouse reveals functional differences among mammalian
SWI/SNF complexes. Mol Cell 6: 1287-1295.
187. Bultman SJ, Gebuhr TC, Pan H, Svoboda P, Schultz RM, et al. (2006) Maternal
BRG1 regulates zygotic genome activation in the mouse. Genes Dev 20: 1744-
1754.
188. Tarkowski AK, Wroblewska J (1967) Development of blastomeres of mouse eggs
isolated at the 4- and 8-cell stage. J Embryol Exp Morphol 18: 155-180.
189. Marikawa Y, Alarcon VB (2009) Establishment of trophectoderm and inner cell
mass lineages in the mouse embryo. Mol Reprod Dev 76: 1019-1032.
190. Fleming TP (1987) A quantitative analysis of cell allocation to trophectoderm and
inner cell mass in the mouse blastocyst. Dev Biol 119: 520-531.
191. Pedersen RA, Wu K, Balakier H (1986) Origin of the inner cell mass in mouse
embryos: cell lineage analysis by microinjection. Dev Biol 117: 581-595.
192. Ralston A, Rossant J (2008) Cdx2 acts downstream of cell polarization to cell-
autonomously promote trophectoderm fate in the early mouse embryo. Dev Biol
313: 614-629.
193. Rossant J, Tam PP (2009) Blastocyst lineage formation, early embryonic
asymmetries and axis patterning in the mouse. Development 136: 701-713.
128
194. Reeve WJ, Ziomek CA (1981) Distribution of microvilli on dissociated blastomeres
from mouse embryos: evidence for surface polarization at compaction. J Embryol
Exp Morphol 62: 339-350.
195. Reeve WJ (1981) Cytoplasmic polarity develops at compaction in rat and mouse
embryos. J Embryol Exp Morphol 62: 351-367.
196. Johnson MH, Ziomek CA (1981) The foundation of two distinct cell lineages within
the mouse morula. Cell 24: 71-80.
197. Suwinska A, Czolowska R, Ozdzenski W, Tarkowski AK (2008) Blastomeres of the
mouse embryo lose totipotency after the fifth cleavage division: expression of
Cdx2 and Oct4 and developmental potential of inner and outer blastomeres of 16-
and 32-cell embryos. Dev Biol 322: 133-144.
198. Ng RK, Dean W, Dawson C, Lucifero D, Madeja Z, et al. (2008) Epigenetic
restriction of embryonic cell lineage fate by methylation of Elf5. Nat Cell Biol 10:
1280-1290.
199. Nishioka N, Yamamoto S, Kiyonari H, Sato H, Sawada A, et al. (2008) Tead4 is
required for specification of trophectoderm in pre-implantation mouse embryos.
Mech Dev 125: 270-283.
200. Jedrusik A, Parfitt DE, Guo G, Skamagki M, Grabarek JB, et al. (2008) Role of
Cdx2 and cell polarity in cell allocation and specification of trophectoderm and
inner cell mass in the mouse embryo. Genes Dev 22: 2692-2706.
201. Niwa H, Toyooka Y, Shimosato D, Strumpf D, Takahashi K, et al. (2005)
Interaction between Oct3/4 and Cdx2 determines trophectoderm differentiation.
Cell 123: 917-929.
202. Dietrich JE, Hiiragi T (2007) Stochastic patterning in the mouse pre-implantation
embryo. Development 134: 4219-4231.
203. Beck F, Erler T, Russell A, James R (1995) Expression of Cdx-2 in the mouse
embryo and placenta: possible role in patterning of the extra-embryonic
membranes. Dev Dyn 204: 219-227.
204. Scholer HR, Ruppert S, Suzuki N, Chowdhury K, Gruss P (1990) New type of POU
domain in germ line-specific protein Oct-4. Nature 344: 435-439.
205. Rosner MH, Vigano MA, Ozato K, Timmons PM, Poirier F, et al. (1990) A POU-
domain transcription factor in early stem cells and germ cells of the mammalian
embryo. Nature 345: 686-692.
206. Okamoto K, Okazawa H, Okuda A, Sakai M, Muramatsu M, et al. (1990) A novel
octamer binding transcription factor is differentially expressed in mouse
embryonic cells. Cell 60: 461-472.
207. Erhardt S, Su IH, Schneider R, Barton S, Bannister AJ, et al. (2003) Consequences
of the depletion of zygotic and embryonic enhancer of zeste 2 during
preimplantation mouse development. Development 130: 4235-4248.
208. Dahl JA, Reiner AH, Klungland A, Wakayama T, Collas P (2010) Histone H3 lysine
27 methylation asymmetry on developmentally-regulated promoters distinguish
the first two lineages in mouse preimplantation embryos. PLoS One 5: e9150.
209. Corry GN, Tanasijevic B, Barry ER, Krueger W, Rasmussen TP (2009) Epigenetic
regulatory mechanisms during preimplantation development. Birth Defects Res C
Embryo Today 87: 297-313.
129
210. Lyon MF (1961) Gene action in the X-chromosome of the mouse (Mus musculus
L.). Nature 190: 372-373.
211. Brockdorff N, Ashworth A, Kay GF, McCabe VM, Norris DP, et al. (1992) The
product of the mouse Xist gene is a 15 kb inactive X-specific transcript containing
no conserved ORF and located in the nucleus. Cell 71: 515-526.
212. Brown CJ, Hendrich BD, Rupert JL, Lafreniere RG, Xing Y, et al. (1992) The
human XIST gene: analysis of a 17 kb inactive X-specific RNA that contains
conserved repeats and is highly localized within the nucleus. Cell 71: 527-542.
213. Penny GD, Kay GF, Sheardown SA, Rastan S, Brockdorff N (1996) Requirement
for Xist in X chromosome inactivation. Nature 379: 131-137.
214. Augui S, Filion GJ, Huart S, Nora E, Guggiari M, et al. (2007) Sensing X
chromosome pairs before X inactivation via a novel X-pairing region of the Xic.
Science 318: 1632-1636.
215. Clemson CM, McNeil JA, Willard HF, Lawrence JB (1996) XIST RNA paints the
inactive X chromosome at interphase: evidence for a novel RNA involved in
nuclear/chromosome structure. J Cell Biol 132: 259-275.
216. Lucchesi JC, Kelly WG, Panning B (2005) Chromatin remodeling in dosage
compensation. Annu Rev Genet 39: 615-651.
217. Jeppesen P, Turner BM (1993) The inactive X chromosome in female mammals is
distinguished by a lack of histone H4 acetylation, a cytogenetic marker for gene
expression. Cell 74: 281-289.
218. Peters AH, Mermoud JE, O'Carroll D, Pagani M, Schweizer D, et al. (2002) Histone
H3 lysine 9 methylation is an epigenetic imprint of facultative heterochromatin.
Nat Genet 30: 77-80.
219. Boggs BA, Cheung P, Heard E, Spector DL, Chinault AC, et al. (2002)
Differentially methylated forms of histone H3 show unique association patterns
with inactive human X chromosomes. Nat Genet 30: 73-76.
220. Angelov D, Molla A, Perche PY, Hans F, Cote J, et al. (2003) The histone variant
macroH2A interferes with transcription factor binding and SWI/SNF nucleosome
remodeling. Mol Cell 11: 1033-1041.
221. Takagi N, Sasaki M (1975) Preferential inactivation of the paternally derived X
chromosome in the extraembryonic membranes of the mouse. Nature 256: 640-
642.
222. Huynh KD, Lee JT (2003) Inheritance of a pre-inactivated paternal X chromosome
in early mouse embryos. Nature 426: 857-862.
223. Mak W, Nesterova TB, de Napoles M, Appanah R, Yamanaka S, et al. (2004)
Reactivation of the paternal X chromosome in early mouse embryos. Science 303:
666-669.
224. Salido EC, Yen PH, Mohandas TK, Shapiro LJ (1992) Expression of the X-
inactivation-associated gene XIST during spermatogenesis. Nat Genet 2: 196-199.
225. Turner JM, Mahadevaiah SK, Elliott DJ, Garchon HJ, Pehrson JR, et al. (2002)
Meiotic sex chromosome inactivation in male mice with targeted disruptions of
Xist. J Cell Sci 115: 4097-4105.
226. Eggan K, Akutsu H, Hochedlinger K, Rideout W, 3rd, Yanagimachi R, et al. (2000)
X-Chromosome inactivation in cloned mouse embryos. Science 290: 1578-1581.
130
227. Brown CJ, Willard HF (1994) The human X-inactivation centre is not required for
maintenance of X-chromosome inactivation. Nature 368: 154-156.
228. Csankovszki G, Panning B, Bates B, Pehrson JR, Jaenisch R (1999) Conditional
deletion of Xist disrupts histone macroH2A localization but not maintenance of X
inactivation. Nat Genet 22: 323-324.
229. Chang CC, Ma Y, Jacobs S, Tian XC, Yang X, et al. (2005) A maternal store of
macroH2A is removed from pronuclei prior to onset of somatic macroH2A
expression in preimplantation embryos. Dev Biol 278: 367-380.
230. Mak W, Baxter J, Silva J, Newall AE, Otte AP, et al. (2002) Mitotically stable
association of polycomb group proteins eed and enx1 with the inactive x
chromosome in trophoblast stem cells. Curr Biol 12: 1016-1020.
231. Wang J, Mager J, Chen Y, Schneider E, Cross JC, et al. (2001) Imprinted X
inactivation maintained by a mouse Polycomb group gene. Nat Genet 28: 371-
375.
232. Buehr M, Smith A (2003) Genesis of embryonic stem cells. Philos Trans R Soc
Lond B Biol Sci 358: 1397-1402; discussion 1402.
233. Okamoto I, Otte AP, Allis CD, Reinberg D, Heard E (2004) Epigenetic dynamics of
imprinted X inactivation during early mouse development. Science 303: 644-649.
234. Albert M, Peters AH (2009) Genetic and epigenetic control of early mouse
development. Curr Opin Genet Dev 19: 113-121.
235. O'Neill LP, VerMilyea MD, Turner BM (2006) Epigenetic characterization of the
early embryo with a chromatin immunoprecipitation protocol applicable to small
cell populations. Nat Genet 38: 835-841.
236. Kastan MB (2008) DNA damage responses: mechanisms and roles in human
disease: 2007 G.H.A. Clowes Memorial Award Lecture. Mol Cancer Res 6: 517-
524.
237. Zini A, Libman J (2006) Sperm DNA damage: importance in the era of assisted
reproduction. Curr Opin Urol 16: 428-434.
238. Zini A, Meriano J, Kader K, Jarvi K, Laskin CA, et al. (2005) Potential adverse
effect of sperm DNA damage on embryo quality after ICSI. Hum Reprod 20:
3476-3480.
239. Bonduelle M, Van Assche E, Joris H, Keymolen K, Devroey P, et al. (2002)
Prenatal testing in ICSI pregnancies: incidence of chromosomal anomalies in
1586 karyotypes and relation to sperm parameters. Hum Reprod 17: 2600-2614.
240. Generoso WM, Cain KT, Krishna M, Huff SW (1979) Genetic lesions induced by
chemicals in spermatozoa and spermatids of mice are repaired in the egg. Proc
Natl Acad Sci U S A 76: 435-437.
241. Matsuda Y, Seki N, Utsugi-Takeuchi T, Tobari I (1989) Changes in X-ray
sensitivity of mouse eggs from fertilization to the early pronuclear stage, and their
repair capacity. Int J Radiat Biol 55: 233-256.
242. Matsuda Y, Tobari I, Yamada T (1985) Studies on chromosome aberrations in the
eggs of mice fertilized in vitro after irradiation. II. Chromosome aberrations
induced in mature oocytes and fertilized eggs at the pronuclear stage following X-
irradiation. Mutat Res 151: 275-280.
243. Sega GA, Sotomayor RE, Owens JG (1978) A study of unscheduled DNA synthesis
induced by X-rays in the germ cells of male mice. Mutat Res 49: 239-257.
131
244. McPherson SM, Longo FJ (1993) Nicking of rat spermatid and spermatozoa DNA:
possible involvement of DNA topoisomerase II. Dev Biol 158: 122-130.
245. Laberge RM, Boissonneault G (2005) On the nature and origin of DNA strand
breaks in elongating spermatids. Biol Reprod 73: 289-296.
246. McPherson S, Longo FJ (1993) Chromatin structure-function alterations during
mammalian spermatogenesis: DNA nicking and repair in elongating spermatids.
Eur J Histochem 37: 109-128.
247. Cho C, Jung-Ha H, Willis WD, Goulding EH, Stein P, et al. (2003) Protamine 2
deficiency leads to sperm DNA damage and embryo death in mice. Biol Reprod
69: 211-217.
248. Shirley CR, Hayashi S, Mounsey S, Yanagimachi R, Meistrich ML (2004)
Abnormalities and reduced reproductive potential of sperm from Tnp1- and Tnp2-
null double mutant mice. Biol Reprod 71: 1220-1229.
249. Zhao M, Shirley CR, Mounsey S, Meistrich ML (2004) Nucleoprotein transitions
during spermiogenesis in mice with transition nuclear protein Tnp1 and Tnp2
mutations. Biol Reprod 71: 1016-1025.
250. Aoki VW, Moskovtsev SI, Willis J, Liu L, Mullen JB, et al. (2005) DNA integrity is
compromised in protamine-deficient human sperm. J Androl 26: 741-748.
251. Chevaillier P, Mauro N, Feneux D, Jouannet P, David G (1987) Anomalous protein
complement of sperm nuclei in some infertile men. Lancet 2: 806-807.
252. van Loon AA, Sonneveld E, Hoogerbrugge J, van der Schans GP, Grootegoed JA, et
al. (1993) Induction and repair of DNA single-strand breaks and DNA base
damage at different cellular stages of spermatogenesis of the hamster upon in
vitro exposure to ionizing radiation. Mutat Res 294: 139-148.
253. Jaroudi S, SenGupta S (2007) DNA repair in mammalian embryos. Mutat Res 635:
53-77.
254. Mac Auley A, Werb Z, Mirkes PE (1993) Characterization of the unusually rapid
cell cycles during rat gastrulation. Development 117: 873-883.
255. Guan XY, Fang Y, Sham JS, Kwong DL, Zhang Y, et al. (2000) Recurrent
chromosome alterations in hepatocellular carcinoma detected by comparative
genomic hybridization. Genes Chromosomes Cancer 29: 110-116.
256. Wong N, Lai P, Lee SW, Fan S, Pang E, et al. (1999) Assessment of genetic changes
in hepatocellular carcinoma by comparative genomic hybridization analysis:
relationship to disease stage, tumor size, and cirrhosis. Am J Pathol 154: 37-43.
257. Yan L, Echenique V, Busso C, SanMiguel P, Ramakrishna W, et al. (2002) Cereal
genes similar to Snf2 define a new subfamily that includes human and mouse
genes. Mol Genet Genomics 268: 488-499.
258. Wang QT, Piotrowska K, Ciemerych MA, Milenkovic L, Scott MP, et al. (2004) A
genome-wide study of gene activity reveals developmental signaling pathways in
the preimplantation mouse embryo. Dev Cell 6: 133-144.
259. Ahel D, Horejsi Z, Wiechens N, Polo SE, Garcia-Wilson E, et al. (2009) Poly(ADP-
ribose)-dependent regulation of DNA repair by the chromatin remodeling enzyme
ALC1. Science 325: 1240-1243.
260. Gottschalk AJ, Timinszky G, Kong SE, Jin J, Cai Y, et al. (2009) Poly(ADP-
ribosyl)ation directs recruitment and activation of an ATP-dependent chromatin
remodeler. Proc Natl Acad Sci U S A 106: 13770-13774.
132
261. Ma NF, Hu L, Fung JM, Xie D, Zheng BJ, et al. (2008) Isolation and
characterization of a novel oncogene, amplified in liver cancer 1, within a
commonly amplified region at 1q21 in hepatocellular carcinoma. Hepatology 47:
503-510.
262. Chen M, Huang JD, Hu L, Zheng BJ, Chen L, et al. (2009) Transgenic CHD1L
expression in mouse induces spontaneous tumors. PLoS One 4: e6727.
263. Halazonetis TD, Gorgoulis VG, Bartek J (2008) An oncogene-induced DNA
damage model for cancer development. Science 319: 1352-1355.
264. Hirschhorn JN, Brown SA, Clark CD, Winston F (1992) Evidence that SNF2/SWI2
and SNF5 activate transcription in yeast by altering chromatin structure. Genes
Dev 6: 2288-2298.
265. Wang W, Cote J, Xue Y, Zhou S, Khavari PA, et al. (1996) Purification and
biochemical heterogeneity of the mammalian SWI-SNF complex. EMBO J 15:
5370-5382.
266. Braude P, Bolton V, Moore S (1988) Human gene expression first occurs between
the four- and eight-cell stages of preimplantation development. Nature 332: 459-
461.
267. Aoki F, Worrad DM, Schultz RM (1997) Regulation of transcriptional activity
during the first and second cell cycles in the preimplantation mouse embryo. Dev
Biol 181: 296-307.
268. Solter D, Knowles BB (1975) Immunosurgery of mouse blastocyst. Proc Natl Acad
Sci U S A 72: 5099-5102.
269. Min J, Zhang Y, Xu RM (2003) Structural basis for specific binding of Polycomb
chromodomain to histone H3 methylated at Lys 27. Genes Dev 17: 1823-1828.
270. Ringrose L, Paro R (2004) Epigenetic regulation of cellular memory by the
Polycomb and Trithorax group proteins. Annu Rev Genet 38: 413-443.
271. Shen X, Liu Y, Hsu YJ, Fujiwara Y, Kim J, et al. (2008) EZH1 mediates
methylation on histone H3 lysine 27 and complements EZH2 in maintaining stem
cell identity and executing pluripotency. Mol Cell 32: 491-502.
272. Sebastian S, Sreenivas P, Sambasivan R, Cheedipudi S, Kandalla P, et al. (2009)
MLL5, a trithorax homolog, indirectly regulates H3K4 methylation, represses
cyclin A2 expression, and promotes myogenic differentiation. Proc Natl Acad Sci
U S A 106: 4719-4724.
273. Tripoulas N, LaJeunesse D, Gildea J, Shearn A (1996) The Drosophila ash1 gene
product, which is localized at specific sites on polytene chromosomes, contains a
SET domain and a PHD finger. Genetics 143: 913-928.
274. Gregory GD, Vakoc CR, Rozovskaia T, Zheng X, Patel S, et al. (2007) Mammalian
ASH1L is a histone methyltransferase that occupies the transcribed region of
active genes. Mol Cell Biol 27: 8466-8479.
275. Santos-Rosa H, Schneider R, Bannister AJ, Sherriff J, Bernstein BE, et al. (2002)
Active genes are tri-methylated at K4 of histone H3. Nature 419: 407-411.
276. Oda H, Okamoto I, Murphy N, Chu J, Price SM, et al. (2009) Monomethylation of
histone H4-lysine 20 is involved in chromosome structure and stability and is
essential for mouse development. Mol Cell Biol 29: 2278-2295.
133
277. Baltus GA, Kowalski MP, Tutter AV, Kadam S (2009) A positive regulatory role for
the mSin3A-HDAC complex in pluripotency through Nanog and Sox2. J Biol
Chem 284: 6998-7006.
278. Dovey OM, Foster CT, Cowley SM (2010) Histone deacetylase 1 (HDAC1), but not
HDAC2, controls embryonic stem cell differentiation. Proc Natl Acad Sci U S A
107: 8242-8247.
279. Barak O, Lazzaro MA, Lane WS, Speicher DW, Picketts DJ, et al. (2003) Isolation
of human NURF: a regulator of Engrailed gene expression. EMBO J 22: 6089-
6100.
280. Ho L, Crabtree GR (2010) Chromatin remodelling during development. Nature 463:
474-484.
281. Ho L, Ronan JL, Wu J, Staahl BT, Chen L, et al. (2009) An embryonic stem cell
chromatin remodeling complex, esBAF, is essential for embryonic stem cell self-
renewal and pluripotency. Proc Natl Acad Sci U S A 106: 5181-5186.
282. Peng JC, Valouev A, Swigut T, Zhang J, Zhao Y, et al. (2009) Jarid2/Jumonji
coordinates control of PRC2 enzymatic activity and target gene occupancy in
pluripotent cells. Cell 139: 1290-1302.
283. Eisen JA, Sweder KS, Hanawalt PC (1995) Evolution of the SNF2 family of
proteins: subfamilies with distinct sequences and functions. Nucleic Acids Res
23: 2715-2723.
284. Masui S, Shimosato D, Toyooka Y, Yagi R, Takahashi K, et al. (2005) An efficient
system to establish multiple embryonic stem cell lines carrying an inducible
expression unit. Nucleic Acids Res 33: e43.
285. Fazzio TG, Huff JT, Panning B (2008) An RNAi screen of chromatin proteins
identifies Tip60-p400 as a regulator of embryonic stem cell identity. Cell 134:
162-174.
286. Fry CJ, Peterson CL (2001) Chromatin remodeling enzymes: who's on first? Curr
Biol 11: R185-197.
287. Foygel K, Choi B, Jun S, Leong DE, Lee A, et al. (2008) A novel and critical role
for Oct4 as a regulator of the maternal-embryonic transition. PLoS One 3: e4109.
288. Tanaka TS, Kunath T, Kimber WL, Jaradat SA, Stagg CA, et al. (2002) Gene
expression profiling of embryo-derived stem cells reveals candidate genes
associated with pluripotency and lineage specificity. Genome Res 12: 1921-1928.
289. Koh DW, Lawler AM, Poitras MF, Sasaki M, Wattler S, et al. (2004) Failure to
degrade poly(ADP-ribose) causes increased sensitivity to cytotoxicity and early
embryonic lethality. Proc Natl Acad Sci U S A 101: 17699-17704.
290. Berger NA (1985) Poly(ADP-ribose) in the cellular response to DNA damage.
Radiat Res 101: 4-15.
291. Benjamin RC, Gill DM (1980) ADP-ribosylation in mammalian cell ghosts.
Dependence of poly(ADP-ribose) synthesis on strand breakage in DNA. J Biol
Chem 255: 10493-10501.
292. Herceg Z, Wang ZQ (2001) Functions of poly(ADP-ribose) polymerase (PARP) in
DNA repair, genomic integrity and cell death. Mutat Res 477: 97-110.
293. Hakem R (2008) DNA-damage repair; the good, the bad, and the ugly. Embo J 27:
589-605.
134
294. Matsuda Y, Yamada T, Tobari I (1985) Studies on chromosome aberrations in the
eggs of mice fertilized in vitro after irradiation. I. Chromosome aberrations
induced in sperm after X-irradiation. Mutat Res 148: 113-117.
295. Rothkamm K, Kruger I, Thompson LH, Lobrich M (2003) Pathways of DNA
double-strand break repair during the mammalian cell cycle. Mol Cell Biol 23:
5706-5715.
296. Fiorenza MT, Bevilacqua A, Bevilacqua S, Mangia F (2001) Growing dictyate
oocytes, but not early preimplantation embryos, of the mouse display high levels
of DNA homologous recombination by single-strand annealing and lack DNA
nonhomologous end joining. Dev Biol 233: 214-224.
297. Hagmann M, Adlkofer K, Pfeiffer P, Bruggmann R, Georgiev O, et al. (1996)
Dramatic changes in the ratio of homologous recombination to nonhomologous
DNA-end joining in oocytes and early embryos of Xenopus laevis. Biol Chem
Hoppe Seyler 377: 239-250.
298. Hagmann M, Bruggmann R, Xue L, Georgiev O, Schaffner W, et al. (1998)
Homologous recombination and DNA-end joining reactions in zygotes and early
embryos of zebrafish (Danio rerio) and Drosophila melanogaster. Biol Chem 379:
673-681.
299. Lee SE, Mitchell RA, Cheng A, Hendrickson EA (1997) Evidence for DNA-PK-
dependent and -independent DNA double-strand break repair pathways in
mammalian cells as a function of the cell cycle. Mol Cell Biol 17: 1425-1433.
300. Schreiber V, Dantzer F, Ame JC, de Murcia G (2006) Poly(ADP-ribose): novel
functions for an old molecule. Nat Rev Mol Cell Biol 7: 517-528.
301. Tusher VG, Tibshirani R, Chu G (2001) Significance analysis of microarrays
applied to the ionizing radiation response. Proc Natl Acad Sci U S A 98: 5116-
5121.
302. Ji H, Jiang H, Ma W, Johnson DS, Myers RM, et al. (2008) An integrated software
system for analyzing ChIP-chip and ChIP-seq data. Nat Biotechnol 26: 1293-
1300.
303. Johnson DS, Mortazavi A, Myers RM, Wold B (2007) Genome-wide mapping of in
vivo protein-DNA interactions. Science 316: 1497-1502.
304. Copeland NG, Jenkins NA, Court DL (2001) Recombineering: a powerful new tool
for mouse functional genomics. Nat Rev Genet 2: 769-779.
305. Frosina G (2000) Overexpression of enzymes that repair endogenous damage to
DNA. Eur J Biochem 267: 2135-2149.
306. Lindahl T, Karran P, Wood RD (1997) DNA excision repair pathways. Curr Opin
Genet Dev 7: 158-169.
307. Jackson SP, Bartek J (2009) The DNA-damage response in human biology and
disease. Nature 461: 1071-1078.
308. Vinson RK, Hales BF (2002) DNA repair during organogenesis. Mutat Res 509: 79-
91.
309. Bielanska M, Jin S, Bernier M, Tan SL, Ao A (2005) Diploid-aneuploid mosaicism
in human embryos cultured to the blastocyst stage. Fertil Steril 84: 336-342.
310. Delhanty JD, Handyside AH (1995) The origin of genetic defects in the human and
their detection in the preimplantation embryo. Hum Reprod Update 1: 201-215.
135
311. Harrison RH, Kuo HC, Scriven PN, Handyside AH, Ogilvie CM (2000) Lack of cell
cycle checkpoints in human cleavage stage embryos revealed by a clonal pattern
of chromosomal mosaicism analysed by sequential multicolour FISH. Zygote 8:
217-224.
312. van Blerkom J, Brockway GO (1975) Qualitative patterns of protein synthesis in the
preimplantation mouse embryo. I. Normal pregnancy. Dev Biol 44: 148-157.
313. Gu H, Marth JD, Orban PC, Mossmann H, Rajewsky K (1994) Deletion of a DNA
polymerase beta gene segment in T cells using cell type-specific gene targeting.
Science 265: 103-106.
314. Xanthoudakis S, Smeyne RJ, Wallace JD, Curran T (1996) The redox/DNA repair
protein, Ref-1, is essential for early embryonic development in mice. Proc Natl
Acad Sci U S A 93: 8919-8923.
315. Bentley D, Selfridge J, Millar JK, Samuel K, Hole N, et al. (1996) DNA ligase I is
required for fetal liver erythropoiesis but is not essential for mammalian cell
viability. Nat Genet 13: 489-491.
316. Larsen E, Gran C, Saether BE, Seeberg E, Klungland A (2003) Proliferation failure
and gamma radiation sensitivity of Fen1 null mutant mice at the blastocyst stage.
Mol Cell Biol 23: 5346-5353.
317. Friedberg EC, Meira LB (2006) Database of mouse strains carrying targeted
mutations in genes affecting biological responses to DNA damage Version 7.
DNA Repair (Amst) 5: 189-209.
318. de Boer J, Donker I, de Wit J, Hoeijmakers JH, Weeda G (1998) Disruption of the
mouse xeroderma pigmentosum group D DNA repair/basal transcription gene
results in preimplantation lethality. Cancer Res 58: 89-94.
319. Ng JM, Vrieling H, Sugasawa K, Ooms MP, Grootegoed JA, et al. (2002)
Developmental defects and male sterility in mice lacking the ubiquitin-like DNA
repair gene mHR23B. Mol Cell Biol 22: 1233-1245.
320. Yonemasu R, Minami M, Nakatsu Y, Takeuchi M, Kuraoka I, et al. (2005)
Disruption of mouse XAB2 gene involved in pre-mRNA splicing, transcription
and transcription-coupled DNA repair results in preimplantation lethality. DNA
Repair (Amst) 4: 479-491.
321. Hassa PO, Hottiger MO (2008) The diverse biological roles of mammalian PARPS,
a small but powerful family of poly-ADP-ribose polymerases. Front Biosci 13:
3046-3082.
322. Penninger JM, Kroemer G (2003) Mitochondria, AIF and caspases--rivaling for cell
death execution. Nat Cell Biol 5: 97-99.
323. Ye H, Cande C, Stephanou NC, Jiang S, Gurbuxani S, et al. (2002) DNA binding is
required for the apoptogenic action of apoptosis inducing factor. Nat Struct Biol
9: 680-684.
324. Koi M, Umar A, Chauhan DP, Cherian SP, Carethers JM, et al. (1994) Human
chromosome 3 corrects mismatch repair deficiency and microsatellite instability
and reduces N-methyl-N'-nitro-N-nitrosoguanidine tolerance in colon tumor cells
with homozygous hMLH1 mutation. Cancer Res 54: 4308-4312.
325. Umar A, Koi M, Risinger JI, Glaab WE, Tindall KR, et al. (1997) Correction of
hypermutability, N-methyl-N'-nitro-N-nitrosoguanidine resistance, and defective
136
DNA mismatch repair by introducing chromosome 2 into human tumor cells with
mutations in MSH2 and MSH6. Cancer Res 57: 3949-3955.
326. Branch P, Hampson R, Karran P (1995) DNA mismatch binding defects, DNA
damage tolerance, and mutator phenotypes in human colorectal carcinoma cell
lines. Cancer Res 55: 2304-2309.
327. Moolenaar WH, Bierman AJ, Tilly BC, Verlaan I, Defize LH, et al. (1988) A point
mutation at the ATP-binding site of the EGF-receptor abolishes signal
transduction. EMBO J 7: 707-710.
328. Fischer H, Glockshuber R (1994) A point mutation within the ATP-binding site
inactivates both catalytic functions of the ATP-dependent protease La (Lon) from
Escherichia coli. FEBS Lett 356: 101-103.
329. Laurent BC, Treich I, Carlson M (1993) The yeast SNF2/SWI2 protein has DNA-
stimulated ATPase activity required for transcriptional activation. Genes Dev 7:
583-591.
330. Khavari PA, Peterson CL, Tamkun JW, Mendel DB, Crabtree GR (1993) BRG1
contains a conserved domain of the SWI2/SNF2 family necessary for normal
mitotic growth and transcription. Nature 366: 170-174.
331. Tsukiyama T, Palmer J, Landel CC, Shiloach J, Wu C (1999) Characterization of the
imitation switch subfamily of ATP-dependent chromatin-remodeling factors in
Saccharomyces cerevisiae. Genes Dev 13: 686-697.
332. Wang J, Rao S, Chu J, Shen X, Levasseur DN, et al. (2006) A protein interaction
network for pluripotency of embryonic stem cells. Nature 444: 364-368.
333. Stiff T, O'Driscoll M, Rief N, Iwabuchi K, Lobrich M, et al. (2004) ATM and DNA-
PK function redundantly to phosphorylate H2AX after exposure to ionizing
radiation. Cancer Res 64: 2390-2396.
334. Barton TS, Robaire B, Hales BF (2007) DNA damage recognition in the rat zygote
following chronic paternal cyclophosphamide exposure. Toxicol Sci 100: 495-
503.
335. Grenier L, Robaire B, Hales BF (2010) Paternal exposure to cyclophosphamide
affects the progression of sperm chromatin decondensation and activates a DNA
damage response in the prepronuclear rat zygote. Biol Reprod 83: 195-204.
336. Derijck A, van der Heijden G, Giele M, Philippens M, de Boer P (2008) DNA
double-strand break repair in parental chromatin of mouse zygotes, the first cell
cycle as an origin of de novo mutation. Hum Mol Genet 17: 1922-1937.
337. Yukawa M, Oda S, Mitani H, Nagata M, Aoki F (2007) Deficiency in the response
to DNA double-strand breaks in mouse early preimplantation embryos. Biochem
Biophys Res Commun 358: 578-584.
338. Hirao A, Kong YY, Matsuoka S, Wakeham A, Ruland J, et al. (2000) DNA damage-
induced activation of p53 by the checkpoint kinase Chk2. Science 287: 1824-
1827.
339. Hong SJ, Dawson TM, Dawson VL (2004) Nuclear and mitochondrial conversations
in cell death: PARP-1 and AIF signaling. Trends Pharmacol Sci 25: 259-264.
340. Chiarugi A (2002) Poly(ADP-ribose) polymerase: killer or conspirator? The 'suicide
hypothesis' revisited. Trends Pharmacol Sci 23: 122-129.
341. McFadden DE, Friedman JM (1997) Chromosome abnormalities in human beings.
Mutat Res 396: 129-140.
137
342. Kaina B, Fritz G, Ochs K, Haas S, Grombacher T, et al. (1998) Transgenic systems
in studies on genotoxicity of alkylating agents: critical lesions, thresholds and
defense mechanisms. Mutat Res 405: 179-191.
343. Fritz G, Auer B, Kaina B (1994) Effect of transfection of human poly(ADP-
ribose)polymerase in Chinese hamster cells on mutagen resistance. Mutat Res
308: 127-133.
344. Bramson J, Panasci LC (1993) Effect of ERCC-1 overexpression on sensitivity of
Chinese hamster ovary cells to DNA damaging agents. Cancer Res 53: 3237-
3240.
345. Canitrot Y, Frechet M, Servant L, Cazaux C, Hoffmann JS (1999) Overexpression
of DNA polymerase beta: a genomic instability enhancer process. FASEB J 13:
1107-1111.
346. Clairmont CA, Sweasy JB (1996) Dominant negative rat DNA polymerase beta
mutants interfere with base excision repair in Saccharomyces cerevisiae. J
Bacteriol 178: 656-661.
347. Kolodner RD (1995) Mismatch repair: mechanisms and relationship to cancer
susceptibility. Trends Biochem Sci 20: 397-401.
348. Richardson C, Stark JM, Ommundsen M, Jasin M (2004) Rad51 overexpression
promotes alternative double-strand break repair pathways and genome instability.
Oncogene 23: 546-553.