BIOLOGICAL FUNCTION OF E2F7 AND E2F8 IS ESSENTIAL FOR EMBRYO DEVELOPMENT
DISSERTATION
Presented in Partial Fulfillment of the Requirements for
the Degree Doctor of Philosophy in the Graduate
School of the Ohio State University
By
Jing Li, M.S.
*****
The Ohio State University
2009
Approved by Dissertation Committee:
Gustavo Leone, Ph.D., Advisor
Susan Cole, Ph.D. ______Paul Herman, Ph.D. Advisor
Tim Huang, Ph.D. Graduate Program in Molecular Genetics Copyright
Jing Li
2009
ABSTRACT
The novel E2F7 and E2F8 family members are thought to function as transcriptional repressors important for the control of cell proliferation in vitro.
However, as the most recently indentified and least studied E2F members, their biological functions in vivo remain unknown. Here we have analyzed the consequences of inactivating E2f7 and E2f8 in mice. While loss of either E2f7 or
E2f8 did not significantly affect mouse development, their combined ablation
resulted in massive apoptosis, dilated blood vessels and severe placental
defects, culminating in embryonic lethality by day 11.5. E2F7 and E2F8 formed
homo-dimers and hetero-dimers that could recruit various co-repressor
complexes to E2F binding sites of target promoters, including E2f1. Consistent
with their important role in transcriptional repression, mouse embryonic
fibroblasts (MEFs) deficient for E2f7 and E2f8 expressed abnormally high levels
of E2f1 and other E2F-target mRNAs. These double knockout MEFs proliferated
surprisingly well, but accumulated high levels of p53 protein and were
hypersensitive to DNA damage-induced cell death. Importantly, loss of either
E2f1 or p53 suppressed the massive apoptosis observed in double mutant
embryos but failed to rescue their embryonic lethality.
ii In order to identify the leading cause of fetal death, the critical tissues,
related cellular processes and molecular mechanisms of E2F7 and E2F8
function, we utilized conditional knockout strategies to show that extra-embryonic
function of E2F7 and E2F8 is both necessary and sufficient for embryo
development, and thus define the placental abnormalities as the leading cause
of embryonic lethality in E2f7-/-E2f8-/- embryos. Consistent with this genetic
finding, cellular examination of double mutant placentas revealed ectopic DNA
replication and inappropriate mitosis in certain cell lineages of placenta. We also
provided two distinct molecular mechanisms of E2F7 and E2F8 function to
explain these cellular phenotypes. On one hand, E2F7 and E2F8 could directly
repress a novel transcription network that is critical for controlling DNA
replication. On the other hand, E2F7 and E2F8 could indirectly regulate the expression of mitotic cyclins and therefore coordinate mitosis progression. We believe disruption of these pathways in E2f7-/-E2f8-/- placentas culminates in G1-
S and G2-M specific defects and, presumably, leads to the profound placental
abnormality and its associated fetal death.
In summary, this study clearly provides the first in vivo evidence for the
biological functions of E2F7 and E2F8. We demonstrate, as a unique repressive
arm of the E2F program, E2F7 and E2F8 are not only critical for the control of
apoptotic in the fetus, but also essential for the regulation of cell cycle
progression in the placenta. We conclude that the synergistic function of E2F7
and E2F8 is essential for embryo development.
iii DEDICATION
To my mother and father with deep respect and love, and to all those whose constant support has helped me go through my graduate study.
iv
ACKNOWLEDGMENTS
I would like to thank my advisor Dr. Gustavo Leone for his consistent support, the enthusiasm and knowledge he brought to me during my Ph.D. study.
This study would have not been accomplished without his guidance. I am also grateful for having Dr. Susan Cole, Dr. Paul Herman, and Dr. Tim Huang in my committee. Their continuous mentorship and support in all my scientific endeavors are highly appreciated.
The past and present members of the Leone lab are equally precious for me during my graduate study. I would like to give my thanks to Dr. Alain de
Bruin, Dr. Pamela Wenzel, Shantanu Singh, Madhu Ouseph, Dr. Anthony
Trimboli, Dr. Hui Wang, Dr. Leon Chong, Dr. Enrico Caserta, Dr. Soo-in Bae, and
Lisa Rawahneh, not only for their advices and assistance during my graduate research, but also for the valuable discussions which have helped me become a better scientist. I can not thank enough my undergraduate students, Ran Cong,
Grant Comstock and Branxton Forde. It was such a pleasant experience working with them, as a mentor, or more as a friend.
v I sincerely thank my friends Xin Li, Fu Li, Hongtao Jia, Ainan Bao, Sudu
Sharma, Judy Kuo and many others, who have always been there whenever I need advice or help.
Last but not least, I will forever treasure the support and encouragement provided by my parents. Every accomplishment that I have ever achieved belongs to them.
vi
VITA
1996 – 2000……………………………………...B.S. Bioengineering Shanghai University
2000 – 2003……………………………………...M.S. Human Genetics Shanghai University
2003 – 2004……………………………………...University Fellow The Ohio State University
2004 – 2005……………………………………...Graduate Teaching Assistant The Ohio State University
2005 – 2009……………………………………..Graduate Research Assistant The Ohio State University
PUBLICATIONS
1. Li, J., Chen, Y.G. and Kong, X.Y. 2001. New progress of serial analysis of gene expression. Sheng Wu Gong Cheng Xue Bao 17: 613-616.
2. Li, J., Hu, L.D., Wang, W.J., Chen, Y.G. and Kong, X.Y. 2003. Linkage analysis of the candidate genes of familial restless legs syndrome. Yi Chuan Xue Bao 30: 325-329.
3. Chang, Q., Pang, J.C., Li, J., Hu, L., Kong, X. and Ng, H.K. 2004. Molecular analysis of PinX1 in medulloblastomas. Int. J. Cancer 109: 309-314.
4. Maiti, B., Li, J., de Bruin, A., Gordon, F., Timmers, C., Opavsky, R., Patil, K., Tuttle, J., Cleghorn, W. and Leone, G. 2005. Cloning and characterization of mouse E2F8, a novel mammalian E2F family member capable of blocking cellular proliferation. J. Biol. Chem. 280: 18211-18220.
vii 5. Li, J., Ran, C., Li, E., Gordon, F., Comstock, G., Siddiqui, H., Cleghorn, W., Chen, H.Z., Kornacker, K., Liu, C.G., Pandit, S.K., Khanizadeh, M., Weinstein, M., Leone, G. and de Bruin, A. 2008. Synergistic function of E2F7 and E2F8 is essential for cell survival and embryonic development. Dev. Cell. 14: 62- 75.
6. Lammens, T.*, Li, J.*, Leone, G. and De Veylder, L. 2009. Atypical E2Fs: new players in the E2F transcription factor family. Trends in Cell Biology 19: 111-118. *Co-first author.
7. Li, J., Ouseph, M., Comstock, G., Thompson, J., Forde, B., Chong, J. and Leone, G. E2F7 and E2F8 control proliferation and govern placental function during development. Dev. Cell. (In preparation).
FIELDS OF STUDY
Major Field: Molecular Genetics
viii
TABLE OF CONTENTS
Page
Abstract……………………………………………………………………….. ii
Dedication………………………………………...... ………………… iv
Acknowledgements………………………………………………………….. v
Vita…………………………………………………………………………….. vii
List of Tables………………………………………………………………..... xv
List of Figures………………………………………………………………… xvi
List of Abbreviations…………………………………………………………. xix
Chapters:
1 Introduction…………………………………………………………..….. 1
1.1 Cell cycle regulation……………………………………………………. 1
1.2 The E2F family of transcription factors………………………………. 2
1.3 Atypical E2Fs: new players in the E2F family…………………….…. 4
1.3.1 Molecular features of atypical E2F proteins……………..….... 5
1.3.2 Transcriptional regulation of atypical E2f expression……….. 7
1.3.3 Atypical E2F proteins and cell size control…………………… 10
ix 1.3.4 Atypical E2F proteins and cell cycle control………………..… 10
1.3.5 Atypical E2F proteins and DNA damage response………..… 13
2 Synergistic function of E2F7 and E2F8 is essential for embryonic development and cell survival…………………………………………. 19
2.1 Introduction……………………………………………………………… 19
2.2 Results…………………………………………………………………… 22
2.2.1 E2F7 and E2F8 are essential for embryonic viability………... 22
2.2.2 Genetic ablation of E2f7 and E2f8 in vivo induces massive apoptosis in the fetus…………………………………………………… 23
2.2.3 E2F7 and E2F8 form homo-dimers and hetero-dimers……... 24
2.2.4 E2f1 is a direct target of E2F7 and E2F8……………………... 25
2.2.5 E2f1 expression is deregulated in E2f7-/-E2f8-/- MEFs………. 26
2.2.6 E2f7-/-E2f8-/- cells are hypersensitive to DNA damage……… 28
2.2.7 Induction of apoptosis in E2f7-/-E2f8-/- embryos is dependent on E2F1 and p53………………………………………………………... 29
2.3 Discussion……………………………………………………………….. 30
2.4 Materials and Methods…………………………………………………. 34
2.4.1 Generation of E2f7 and E2f8 knockout mice…………………. 34
2.4.2 Quantitive reverse transcriptase PCR (RT-PCR)……...... 35
2.4.3 Affimetrix microarray analysis………………………………….. 35
2.4.4 Co-immunoprecipitation (Co-IP) assay…..……………….…... 36
2.4.5 Chromatin immunoprecipitation (ChIP) and sequential ChIP assays……………………………………………………………………. 36
x 2.4.6 Western blot and antibodies……………………………………. 38
2.4.7 Cell culture and viability assay…………………………………. 38
2.4.8 FACS analysis…………………………………………………… 38
2.4.9 BrdU and TUNEL assays…………………………...………… 39
3 Extra-embryonic function of E2F7 and E2F8 is essential for fetal survival…………………………………………………………………… 60
3.1 Introduction…………………………………….………………………… 60
3.2 Results…………………………………………………………………… 62
3.2.1 Loss of E2f7 and E2f8 leads to profound placental defects… 62
3.2.2 Wild type placenta is sufficient to support mutant fetuses to birth……………………………………………………………………….. 64
3.2.3 E2f7 and E2f8 are essential in the placenta………………….. 65
3.2.4 Loss of E2f7 and E2f8 in the placenta dictates molecular events in the fetus……………………………...……………………….. 66
3.2.5 Fetal function and placental function of E2F7/E2F8 both contribute to the full apoptotic phenotype observed in E2f7-/-E2f8-/- fetuses…………………………………………………………………… 68
3.3 Discussion……………………………………………………………….. 69
3.4 Material and Methods…………………………………………………... 70
3.4.1 Mouse strains and genotyping…………………………………. 70
3.4.2 Quantitative RT-PCR……………………………………………. 71
3.4.3 In situ hybridization……………………………………………… 71
3.4.4 Histological analysis and immunostaining……………………. 72
3.4.5 X-gal staining…………………………………………………….. 72
xi 3.4.6 TUNEL assay……………………………………………….……. 73
3.4.7 Affymetrix microarray analysis…………………………………. 73
4 E2F7 and E2F8 play a critical role in cell cycle regulation…………. 90
4.1 Introduction…………….………………………………………………… 90
4.2 Results…………………………………………………………………… 94
4.2.1 Identification of direct targets of E2F7 and E2F8…………….. 94
4.2.2 Ablation of E2f7 and E2f8 results in ectopic S-phase entry and aberrant mitosis……………………………………………………. 96
4.2.3 Molecular mechanisms underlying the observed cell cycle defects………………….………………………………………………… 98
4.2.4 Loss of E2f3a extends the life span of E2f7-/-E2f8-/- embryos by rescuing the ectopic DNA replication but not the aberrant mitosis in the placenta………………………………………………….. 99
4.3 Discussion……………………………………………………………….. 101
4.4 Material and Methods…………………………………………………... 104
4.4.1 Affymetrix microarray analysis…………………………………. 104
4.4.2 Quantitative RT-PCR……………………………………………. 104
4.4.3 Chromatin immunoprecipitation (ChIP)……………………….. 105
4.4.4 BrdU and TUNEL assay…..……………………………………. 106
4.4.5 Histological analysis and immunostaining……………………. 106
4.4.6 Confocal microscopy and 3D reconstruction…………………. 107
4.4.7 Transmission electron microscopy (TEM)……………………. 108
4.4.8 Western blot and antibodies……………………………………. 108
xii 4.4.9 Quantification and statistical analysis…………………………. 108
5 E2F7/E2F8-associated co-repressor complexes……………………. 137
5.1 Introduction…………………….………………………………………… 137
5.2 Results…………………………………………………………………… 138
5.2.1 E2F7 can specifically associate with CtBP and Rb family proteins……..……………………………………………………………. 138
5.2.2 Both E2F7 and E2F8 associate with components of the histone modifying machinery…………………..………………………. 141
5.2.3 Further characterization of E2F7/E2F8-associated repressor complexes……………………………………………………………….. 142
5.3 Discussion……………………………………………………………….. 143
5.4 Material and Methods…………………………………………………... 145
5.4.1 Co-immunoprecipitation (Co-IP) and sequential Co-IP assays……………………………………………………………………. 145
5.4.2 Western blot and antibodies……………………………………. 146
5.4.3 Site-directed mutagenesis……………………………………... 147
5.4.4 HDAC activity assay…………………………………………….. 147
6 Discussion……………………………………………………………….. 163
6.1 Role of E2Fs in development…………………………………….. 163
6.2 Critical tissues and cell lineages of E2F7and E2F8 function..... 165
6.3 Cellular processes and molecular pathways regulated by E2F7 and E2F8……………………………………………….…….…… 167
6.4 Feedback and antagonism: crosstalks between repressors E2F7, E2F8 and activators E2F1, E2F3a…………………………..… 170
xiii 6.5 E2F7 and E2F8, possible tumor suppressors?...... 171
6.6 Concluding remarks……………………………………………….. 172
References……………………………………………………………………... 176
xiv
LIST OF TABLES
Table Page
2.1 Genotypic analysis of embryos derived from E2f7+/-E2f8+/- or E2f7+/-E2f8-/- intercrosses at the indicated stages of development. 42
2.2 Primers used in the experiments employed in chapter 2…………. 59
4.1 20 direct targets of E2F7 and E2F8……………...... 114
4.2 29 placenta-specific targets of E2F7 and E2F8……………………. 115
xv
LIST OF FIGURES
Figure Page
1.1 Schematic representation of the E2F transcription factor family… 15
1.2 Schematic representation of the E2F activities during the cell cycle…………………………………………………………………….. 16
1.3 Phylogenetic analysis of the mouse E2F transcription factors family.…………………...……………………………………………… 17
1.4 Computer modeling of the interaction between DBD1 and DBD2 DNA binding domains………………………………………………… 18
2.1 Generation of E2f7 and E2f8 knockout mice………………………. 40
2.2 Global deletion of E2f7 and E2f8 results in developmental delay, vascular defects and widespread apoptosis in vivo……………….. 43
2.3 Analyses of proliferation and apoptosis in E2f7-/-E2f8-/- embryos... 45
2.4 E2F7 and E2F8 homo- and hetero-dimerize………………………. 47
2.5 E2F7 and E2F8 bind to the E2f1 promoter………………………… 49
2.6 Deregulation of E2f1 and p53 expression in MEFs deficient in E2f7 and E2f8…………………………………………………………. 51
2.7 MEFs deficient in E2f7 and E2f8 proliferate well but are hypersensitive to DNA damage induced apoptosis……………….. 53
2.8 Loss of E2f1 or p53 suppresses apoptosis in E2f7-/-E2f8-/- embryos………………………………………………………………… 55
2.9 Microarray analysis of E10.5 embryos……………………………… 57
xvi 3.1 E2F7 and E2F8 are highly expressed in the placenta……………. 75
3.2 E2f7-/-E2f8-/- placentas exhibit profound structural abnormalities... 77
3.3 Severe differentiation defects in E2f7-/-E2f8-/- placentas………….. 79
3.4 Wild type placenta is sufficient to rescue E2f7/E2f8-mutant fetuses to birth…………………………………………………………. 81
3.5 Extra-embryonic function of E2f7 and E2f8 is necessary for fetal survival………………….……………………………………………… 83
3.6 Specific deletion of E2f7 and E2f8 in spongiotrophoblasts does not affect placental development and fetal survival....……...... 85
3.7 Loss of E2f7 and E2f8 in the placenta dictates molecular events in the fetus……………………………………………………………... 87
3.8 Cell autonomous and non-cell autonomous functions of E2F7 and E2F8 in the fetus…………………………………………………. 89
4.1 E2F7/E2F8-dependent gene expression in the placenta…………. 110
4.2 Identification of E2F7 and E2F8 direct targets by microarray analysis……………………………………………………………….... 112
4.3 Confirmation and functional annotation of the identified direct targets…...……………………………………………..………………. 116
4.4 Ectopic proliferation in E2f7 and E2f8 double mutant placenta, lung and liver tissues…………………………………………………. 118
4.5 Aberrant mitosis in E2f7-/-E2f8-/- trophoblast giant cells…………... 120
4.6 E2f7-/-E2f8-/- placentas contain binucleated giant cells..………….. 122
4.7 Deregulation of CycA2 expression in E2f7-/-E2f8-/- giant cells……. 124
4.8 Deregulation of CycB1 expression in E2f7-/-E2f8-/- giant cells……. 126
4.9 Mitotic defects in Mx1-cre liver cells………………………………… 128
4.10 Loss of E2f3a rescues the ectopic DNA replication in E2f7-/-E2f8- /- placentas and prolongs life span of the fetus….…………………. 130
xvii 4.11 Ablation of E2f3a does not rescue the mitotic defect observed in E2f7-/-E2f8-/- placenta…………………………………………………. 132
4.12 E2F3a and E2F7/E2F8 co-regulate same set of G1-S target genes…………………………………………………………………… 134
4.13 A working model of E2F7 and E2F8 function in cell cycle control.. 136
5.1 E2F7, but not E2F8, associate with CtBP co-repressors…………. 149
5.2 E2F7, but not E2F8, associate with Rb family of proteins……...… 151
5.3 Both E2F7 and E2F8 associate with HDAC1/HDAC3/HDAC5 and Sin3a/Sin3b co-repressors……………………………………… 153
5.4 E2F7 and E2F8 can form ternary complexes with HDAC1 or Sin3a……………………………………………………………………. 155
5.5 E2F7 can form a macromolecular repressor complex with HDAC1, CtBP and Sin3a…………………………………………….. 157
5.6 Interactions between E2F7 and HDAC1 is independent of CtBP and Rb proteins……………………………………………………….. 158
5.7 Interactions between E2F7, E2F8 and their associated co- repressors are DNA-independent………………...…………………. 160
5.8 Schematic representation depicting the results obtaining from the Co-IP experiments…………………………………………………….. 162
6.1 Current model of E2F7 and E2F8 function during embryo development………………………………………………………….... 174
6.2 Models for the interactions between E2F1 and E2F7/E2F8, and E2F3a and E2F7/E2F8……………………………………………….. 175
xviii
LIST OF ABBREVIATIONS
bp base pair
°C degrees Celsius
ChIP chromatin immunoprecipitation
Co-IP co-immunoprecipitation d day(s)
DKO double knockout
DMSO dimethylsulfoxide
DP dimerization partner
E embryonic day
FBS fetal bovine serum g gram(s) h hour(s)
H&E hematoxylin and eosin
HU hydroxyurea
IB immunoblot
IHC immunohistochemistry
IP Immunoprecipitate
xix L liter(s)
LT labyrinth trophoblast
M moles per liter
MEF(s) mouse embryonic fibroblast(s) min minute(s) mol mole(s)
RT-PCR reverse transcriptase polymerase chain reaction
SP spongiotrophoblast
TG trophoblast giant cell
TKO triple knockout
TS trophoblast stem cell
TUNEL TdT-mediated dUTP nick end-labeling
xx
CHAPTER 1
INTRODUCTION
1.1 Cell cycle regulation
A huge amount of effort has been invested to understand how cell cycle is
regulated. We started to appreciate that there is a complex network of
mechanisms that regulate the cell cycle progression to ensure proper responses
to external signaling, appropriate entry into and successful completion of the cell cycle.
Cells respond to external growth stimuli by activating signaling cascades
that include growth factors, cyclins, protein kinases, transcription factors, and
chromatin modifying/remodeling complexes. A critical step in these signaling
cascades involves the downregulation of cyclin-dependent kinase inhibitors, the
induction of cyclins, and the subsequent association of these cyclins with their
catalytic cyclin-dependent kinase (CDK) subunits (Murray, 2004; Sánchez and
Dynlacht, 2005). Upon mitogenic stimulation, G1-cyclin (Cyclin D) accumulates
within the nucleus, activates CDK4 and CDK6 cyclin-dependent kinases and
thereby phosphorylates retinoblastoma (Rb) and Rb-related pocket proteins
(p107 and p130). The consequence of this phosphorylation is the dissociation of
1 pocket protein-E2F repressor complexes, accumulation of E2F activity and
subsequent activation of genes (such as Cdc6, Mcms) required for the G1-S
transition and DNA replication (Beijersbergen and Bernards, 1996; Frolov and
Dyson, 2004). In early G1, pre-replication complex (ORC, Cdc6, Cdt1, and
Mcm2-7 complex) forms on the DNA, which is referred to as licensing, to ensure
that DNA replication occurs only once per cell cycle (Blow and Laskey. 1988;
Mendez and Stillman, 2003; Takahashi et al., 2005). In late G1, rising G1-S-
cyclin (Cyclin E and Cyclin A) levels lead to the formation of active CDK2-Cyclin
A and CDK2-Cyclin E complexes that trigger G1-S progression. Cdh1, a subunit
of the APC complex, is an important CDK2-Cyclin A substrate. Phosphorylation
of Cdh1 inhibits APC-mediated proteolysis of S- and M-cyclins, and therefore
allows cells to complete S phase and later enter mitosis (Lukas et al., 1999;
Kramer et al., 2000; Hsu et al., 2002). As M phase begins, M-cyclin (Cyclin B1)
translocates into the nucleus, activates Cdc2 kinase (CDK1), and stimulates entry into mitosis. Whereas activation of specific cyclin-CDK complexes drives cell progression through the Start, G1-S and G2-M checkpoints, progression through mitosis is triggered by APC- and SCF-mediated protein destruction, leading to the final stages of cell division (Nurse, 1990; Pines and Hunter, 1991).
1.2 The E2F family of transcription factors
The family of E2F transcription factors was first identified as the essential downstream factors of Rb function (Rowland and Bernards, 2006). E2F target genes were initially implicated in the regulation of the G1-S phase transition of
2 the cell cycle and in DNA replication. The E2F proteins are now recognized to regulate the expression of a large number of genes associated with DNA replication, DNA repair, mitosis, apoptosis, and differentiation (Bracken et al.,
2004).
E2F consists of a family of related proteins (Attwooll et al., 2004;
DeGregori and Johnson, 2006). It had been a long time that people believed that there were only six members in this family (E2F1-E2F6). Based on functional studies, these E2F family members could be divided into two subclasses, transcription activators and repressors (Figure 1.1). Members of the activator subclass, consisting of E2F1, E2F2 and E2F3, bind E2F targets and induce gene expression. The combined ablation of the three activators in mouse embryo fibroblasts (MEFs) results in a profound inhibition of proliferation that is accompanied by a decrease in E2F-target expression (Wu et al., 2001). E2F4,
E2F5 and E2F6 comprise the repressor subclass which suppresses expression of their targets (Cam and Dynlacht, 2003). MEFs lacking E2f4 and E2f5 fail to repress G0-specific genes and lose responsiveness to growth inhibitory signals
(Gaubatz, et al., 2000). Regardless of their positive or negative role in regulating gene expression, E2F1-E2F6 require hetero-dimerization with their dimerization partner DP1/DP2 proteins to exert their function (Trimarchi and Lees, 2002).
E2F repressors are constitutively expressed throughout the cell cycle, whereas the activator subclass of E2Fs is regulated in a cell-cycle dependent manner (Figure 1.2) During G0, repressor E2Fs, particularly E2F4 form complexes with histone deacetylases (HDACs), chromatin-remodeling proteins
3 such as BRM/BRG-1, and histone methyl transferases such as SUV39H1
(DeGregori and Johnson 2006). Upon cell-cycle re-entry, E2F repressor complexes are replaced on promoters by E2F1-E2F3 and its associated activator complexes, including histone acetyl transferases (HATs) such as p300/CBP,
P/CAF, and Tip60 (DeGregori and Johnson 2006). The loading of E2F1- E2F3 activator complexes promote histone acetylation, open chromatin structure, and increased E2F target gene expression (Takahashi et al. 2000; Frolov and Dyson
2004). Therefore, besides action on their specific target genes, E2F repressors and activators may function sequentially to coordinate the expression of same set of genes required for S phase entry and progression through mitosis by binding their promoters in a cell cycle-dependent manner.
1.3 Atypical E2Fs: new players in the E2F family
Recently, the complexity of the E2F family was further extended by the identification of two novel family members. Through a genome survey for genes containing motif(s) homologous to the E2F DNA-binding domain (DBD), a novel evolutionarily conserved arm of E2F network was discovered, first in Arabidopsis thaliana, named as DP-E2F-Like (DEL1 to DEL3) (Kosugi and Ohashi, 2002;
Mariconti et al., 2002; Vandepoele et al., 2002), and subsequently in mice and humans, named as E2F7 and E2F8 (de Bruin et al., 2003; Di Stefano et al.,
2003; Logan et al., 2004; Christensen et al., 2005; Logan et al., 2005; Maiti et al.,
2005). Due to their unique structural properties, these new E2Fs were entitled atypical E2Fs. Interestingly, atypical E2Fs were also discovered in other
4 organisms, such as Caenorhabditis elegans (nematode worm) and Oryza sativa
(rice) by searching the OrthoMCL database (www.orthomcl.org).
1.3.1 Molecular features of atypical E2F proteins
The rather low overall sequence similarity (~20% on amino acid level)
between atypical and typical E2F proteins implies an important structural
difference between these two classes (Figure 1.3). These novel E2F family
members indeed have several salient features that distinguish them from other
family members (Figure 1.1). The most notable feature is that they each possess
two DNA binding domains (vs. E2F1-E2F6 only have one) and lack of DP dimerization domain (DIM), indicating that they may function in a DP-independent
manner. Three-dimensional modeling of E2F7 and E2F8, based on the solved
E2F4-DP2 structure, revealed a compatible structure of DNA binding, with DBD1
and DBD2 at the positions of E2F4 and DP2, respectively (Figure 1.4).
Therefore, the duplicated DBD of atypical E2Fs could structurally mimic the DNA-
binding interface of typical E2F and DP hetero-dimers. Consistent with this
model, both DBDs are essential for atypical E2Fs to bind DNA, since mutation of
either of these domains completely abolished their DNA binding ability (Kosugi
and Ohashi, 2002; Di Stefano et al., 2003; Logan et al., 2004; Christensen et al.,
2005; Logan et al., 2005; Maiti et al., 2005).
Interestingly, instead of dimerizing with DP proteins, E2F7 and E2F8 can form homo-dimers to themselves or form hetero-dimers to each other. These interactions were demonstrated by co-immunoprecipation of epitope-tagged
5 versions of E2F7 and E2F8 produced in human cells (Di Stefano et al., 2003;
Maiti et al., 2005, Chapter 2 in this study and Li et al., 2008). More interestingly,
E2F7 and E2F8 have preferential dimerization states with E2F7 homo-dimers formed preferentially over E2F7/E2F8 hetero-dimers and E2F8 homo-dimers being the least preferred form (Chapter 2 in this study and Li et al., 2008). Detailed homo- and hetero-dimerization studies indicated that the dimerization of
E2F7/E2F8 depends solely on the integrity of DNVLE (191-194) and DNVL (340-
342) residues within DBD1 and DBD2, respectively (Zalmas et al., 2008). The dimerization issue has not been conclusively addressed in plants.
While atypical E2F proteins contain an extra DBD, they do not have other domains found in most typical E2Fs (Figure 1.1). For example, they lack any trans-activation domain, which are possessed by E2F1-E2F5, suggesting that atypical E2F proteins most likely function as transcription repressors. Consistent with this idea, knockout of atypical E2f genes in plants and mammals led to upregulation of defined subsets of E2F-regulated genes expression (Ramirez-
Parra et al., 2004; Vlieghe et al., 2005; this study and Li et al., 2008; Zalmas et al., 2008). E2F7/E2F8 and E2F/DELs can also counteract the activating E2Fs function in competition assays (Kosugi and Ohashi, 2002; Mariconti et al, 2002; de Bruin et al., 2003; Di Stefano et al., 2003; Logan et al., 2004; Christensen et al., 2005; Logan et al., 2005; Maiti et al., 2005).
Rb and its family members (p107 and p130) have been shown to bind E2F and silence gene expression by recruiting chromatin-modifying factors such as
HP1, histone deacetylases, methyltransferases to E2F targets (Luo et al., 1998; 6 Nielsen et al., 2001; Vandel et al., 2001). Like E2F6, atypical E2Fs lack the
classical Rb-binding domain and thus are assumed to contribute gene silencing
in a manner that is independent of pocket proteins (Figure 1.1). In addition, the
mechanism of their repressing action could also well be competition with
activating E2F proteins for same E2F binding sites.
Information regarding domains other than those mentioned above is rather
limited, except studies with GFP reporter constructs have demonstrated the
nuclear localization of E2F7/E2F8 and E2F/DEL proteins (Kosugi and Ohashi,
2002; de Bruin et al., 2003; Maiti et al., 2005). Consistent with this observation,
all atypical E2F proteins contain a putative nuclear localization signal (NLS). In
contrast to the single NLS in the N-terminal regions of typical E2F proteins,
typical E2Fs hold a bipartite NLS at their C-terminus (Mariconti et al, 2002; de
Bruin et al., 2003; Logan et al., 2004; Christensen et al., 2005; Dimova and
Dyson, 2005). However, the functional significance of this difference remains
unknown. In addition, in silico analysis of E2F7 and E2F8 proteins identified
motifs that are known to be responsible to ubiquitin-mediated degradation, which
might provide an attractive explanation for the observed instability of these
atypical E2Fs (Christensen et al., 2005).
1.3.2 Transcriptional regulation of atypical E2f expression
E2F activity is tightly regulated during the cell cycle. As we discussed above, the expression of E2F activators reaches its maximum level at the G1-S transition, whereas that of typical E2F repressors remains unchanged throughout 7 cell cycle (Mariconti et al., 2002; Attwooll et al., 2004). Interestingly, unlike that of typical E2F repressors, the activity of atypical E2Fs is cell cycle-regulated
(Figure 1.2). Transcription of E2f7 and E2f8 is induced at the G1-S transition and mRNA levels peak during S-G2 (de Bruin et al., 2003; Di Stefano et al., 2003;
Logan et al., 2004; Christensen et al., 2005; Logan et al., 2005; Maiti et al.,
2005). Consistent with this observation and the role of E2F7/E2F8 as repressors of gene expression, E2f1, a direct target of E2F7/E2F8, has a complementary expression profile that was disturbed upon double ablation of E2f7 and E2f8, particularly during S-G2 phase of the cell cycle (Chapter 2 in this study and Li et al., 2008; Zalmas et al., 2008). The expression profile of atypical E2f genes in plants is slightly different: they all reach the highest level at the G2-M boundary, but E2Fe/DEL1 and E2Ff/DEL3 display an additional peak at G1-S (Mariconti et al., 2002; Lammens et al., 2008). The transcripts of the E2Fe/DEL1 target,
CCS52A2, have a complementary profile during G2-M as anticipated from the
repressing role of E2Fe/DEL1 (Lammens et al., 2008).
Not much is known about the tissue-specific expression of these atypical
E2f genes. E2f7 and E2f8 share a very similar expression pattern. In adults,
they are highly expressed in skin and thymus, moderately in spleen, and less
abundantly in intestine and testis. No, or very low, expression occurs in brain,
muscle and stomach (de Bruin et al., 2003; Maiti et al., 2005). Thus, in
mammals, atypical E2f transcripts are particularly present in tissues with high
proliferating potential. In plants, E2F/DEL transcripts were detected at high
levels in young, growing tissue, such as young leaves and immature flower buds,
8 whereas they were low in mature tissues, such as adult leaves (Kosugi and
Ohashi, 2002). Therefore, similar to their mammalian counterparts, E2F/DEL
transcript levels are also positively correlated with the proliferation state of cells
and tissues. E2Ff/DEL3, however, appears to be an exception. Analysis of
pollen transcriptome has indicated high levels of E2Ff/DEL3 expression in
mature pollen (Pina et al., 2005). The vegetative pollen nucleus is thought to be arrested in G1, and thus high E2Ff/DEL3 level in pollen was assumed to repress transcription of S-phase genes and be, at least partially, responsible for G1 arrest.
But surprisingly, no clear pollen phenotype could be detected in E2Ff/DEL3 knock-down plants (Ramirez-Parra et al., 2004).
Atypical E2fs themselves are E2F targets. E2f7 was independently
identified in a screen for human E2F1-regulated genes (Di Stefano et al., 2003).
High expression levels of E2f7 and E2f8 were observed in HeLaS3 cells that
have a deregulated Rb-E2F pathway (Christensen et al., 2005). Similarly, plants
over-producing E2Fa-DPa accumulate high level of the E2Ff/DEL3 transcripts
(Vandepoele et al., 2005). Consistent with these observations, sequence
analysis of atypical E2f genes revealed consensus E2F-binding sites on their
promoters. Chromatin immunoprecipitation assays further confirmed a direct
association of E2F1/E2F3/E2F4/E2F7 proteins to the E2f7 and E2f8 promoters
(Di Stefano et al., 2003; Christensen et al., 2005). These findings reinforce the
notion that the E2F network exists as a complex map of cross-talking pathways
(Attwooll et al., 2004; van den Heuvel and Dyson, 2008).
9 1.3.3 Atypical E2F proteins and cell size control
The study of E2Ff/DEL3 in Arabidopsis provided the first biological
evidence for atypical E2F protein function: depletion or overexpression of
E2Ff/DEL3 resulted in a significantly increased or decreased root and hypocotyl
size, respectively (Ramirez-Parra et al., 2004). These differences were caused
by changes in cell size, but not in cellular proliferation, because the size of
meristems, the plant proliferating centers, was not affected. In plants, growth of
root and hypocotyl relies extensively on cell expansion, which demands the
remodeling of preexisting cell walls (Sugimoto-Shirasu and Roberts, 2003). One
important step involved in this process is the loosening of cell walls, a process
carried out by the expansin protein family. E2Ff/DEL3 was shown to bind directly
to the promoter and modulate the expression of several expansin members.
Together, these data indicate an important role of E2Ff/DEL3 in restricting cell
expansion through transcriptional repression of genes involved in cell wall biosynthesis. In mammals, no function in cell size control has been assigned to any atypical E2Fs, it is probably because growth in animals occurs commonly by division and not often by cell expansion (Oldham et al., 2000).
1.3.4 Atypical E2F proteins and cell cycle control
Similar to what was observed for E2Ff/DEL3, overexpression of
E2Fe/DEL1 also had a negative impact on plant size (Vlieghe et al., 2005;
Lammens et al., 2008), but the mechanism underlying this phenotype is
completely different. In E2Fe/DEL1 transgenic plants, cell size decreases not
10 because of a repression of cell expansion, but due to a reduction in DNA ploidy
level (Vlieghe et al., 2005). In dicotyledonous plants (flowering plants with two
embryonic seed leaves), exit from mitotic cell cycle is often accompanied by an
onset of an alternative cell cycle, named endocycle (endoreduplication), in which
DNA replication is followed by incomplete mitosis. Mitotic cell cycle progression
and endoreduplication are linked events. A premature or delayed onset of
endoreduplication often results in an increased or decreased DNA ploidy level
and larger or smaller cells, respectively (Yu et al., 2003; Boudolf et al., 2004;
Verkest et al., 2005). The altered cell size and associated changes in DNA
ploidy levels in E2Fe/DEL1 knockout plants have been proven to result from a
premature onset of the endoreduplication program. More specifically,
E2Fe/DEL1 was found to control the expression of Anaphase-Promoting
Complex/Cyclosome (APC/C) activator gene CCS52A2, which is homologous to
the mammalian Cdh1 (Lammens et al., 2008). CCS52A2 has been implicated in
the control of endocycle onset in Arabidopsis, probably through degradation of
mitotic cyclins (Tarayre et al. 2004; Fulop et al., 2005). Thus, by regulating the
CCS52A2 expression, E2Fe/DEL1 probably determines the time point at which
cells switch from mitosis to endoreduplication and, consequently, controls cell
size. The distinct mechanisms employed by E2Fe/DEL1 and E2Ff/DEL3 to
govern cell size nicely support the idea that in plants cell expansion is regulated
in a ploidy-dependent (via E2Fe/DEL1) and a ploidy-independent (via
E2Ff/DEL3) manner. It would be interesting to study the functional relationship
11 between E2Fe/DEL1 and E2Ff/DEL3 to obtain a deeper insight into the coordination and crosstalk between these different levels of size control.
In mammals, the best example of developmentally programmed endocycle is placental trophoblast giant cells that amplify their genomes and attain a >1000C polytene configuration through successive rounds of DNA replication in the absence of karyokinesis or cytokinesis (MacAuley et al., 1998; Cross, 2005).
Interestingly, like E2Fe/DEL1, E2F7 and E2F8 also play a critical role in the control of endocycle. Trophoblast giant cells lacking E2f7 and E2f8 strikingly segregate their chromosomes and complete karyokinesis in the absence of cytokinesis
(Chapter 4 in this study). It appears that one of the conserved functions of atypical
E2Fs is to regulate endocycle, with the plant E2Fe/DEL1 protein controlling the timing of endocycle onset and mammalian E2F7/E2F8 likely governing the maintenance of endocycle progression.
Atypical E2Fs are also expressed in cell types that do not endoreduplicate, indicating that they might have other functions beyond endocycle. Early studies have demonstrated that overexpression of E2f7/E2f8 could repress cell proliferation and decrease colony formation of HeLa cells (de Bruin et al., 2003;
Di Stefano et al., 2003; Logan et al., 2004; Christensen et al., 2005; Logan et al.,
2005; Maiti et al., 2005). Conversely, loss of E2f7 and E2f8 caused ectopic DNA replication in spongiotrophoblast cells as well as in giant cells in vivo (Chapter 4 in this study). Consistent with these cellular observations, global gene expression profiling and chromatin immunoprecipitation analyses revealed that a majority of
12 E2F7 and E2F8 direct targets are functionally involved in cell cycle control, particularly in DNA replication during the G1-S transition (Chapter 4 in this study).
1.3.5 Atypical E2F proteins and DNA damage response
Studies of knocking out or knocking down E2f7 and E2f8 have provided a deep insight into the role of atypical E2Fs in cell survival. Cells lacking E2f7 and
E2f8 were hyper-sensitive to DNA damage-induced apoptosis (Chapter 2 in this study and Li et al., 2008; Zalmas et al., 2008). E2F7 and E2F8 proteins were induced upon stress (esp. DNA damage), which was coincided with an augmented E2F7/E2F8 occupancy onto the E2f1 promoter (Zalmas et al., 2008).
E2f1 appeared to be a physiological relevant target of E2F7 and E2F8, because simultaneous depletion of it could rescue the E2f7/E2f8-loss induced apoptosis
(Chapter 2 in this study and Li et al., 2008). Thus, E2F7 and E2F8 seem to act as critical regulators of E2F1 activity that had been proposed previously to determine the outcome of DNA damage. Low levels of E2F1 has been suggested to be necessary to recruit DNA repair complexes, whereas high level of E2F1 has been linked to activate pro-apoptotic genes expression and induce apoptosis (Lin et al., 2001; Stevens and La Thangue, 2003). Taken together, these data suggest that E2F7 and E2F8 act as an important arm of the E2F network, which is responsible for fine-tuning E2F1 activity upon stress and, consequently, regulating cell viability.
In contrast to mammals, knockout of individual atypical E2f genes did not affect cell survival in Arabidopsis (Ramirez-Parra et al., 2004; Vlieghe et al., 2005;
13 Lammens et al., 2008). This is probably because plants do not have a sensitive apoptotic program. Plant cells are fixed within their rigid cell walls and, therefore, there is no risk of metastasis of damaged cells. Rather than being eliminated by cell death, damaged cells are often pushed into a differentiation program that precludes cells with mutated DNA from becoming part of gametophytic cells.
Moreover, due to their plasticity, plants can easily escape DNA stress by forming new meristems. Nevertheless, future study of apoptosis in doubly and triply mutated E2F/DEL cohorts should be considered to exclude the possibility of functional redundancy in these atypical E2fs.
In summary, the family of E2F transcription factors has been shown to play a key role in controlling proliferation and apoptosis. However, our understanding of this family has been hampered by the fact that it was only very recently that E2F7 and E2F8, the last two family members, were identified.
Clearly, the completely understanding of the E2F activity will not be possible until physiological functions of E2F7 and E2F8 are fully revealed.
14 Activation NTD DBD DIM Mark. Rb bind. E2F1 E2F2 Activators E2F3 E2F4 E2F5 Repressors E2F6 E2F7 E2F8
Figure 1.1 Schematic representation of the E2F transcription factor family. A conserved DNA binding domain (DBD) is the hallmark of the E2F family of transcription factors. Unlike E2F1-E2F6, E2F7 and E2F8 have two distinct DBDs. The N-terminal domain (NTD) of E2F1-E2F3 harbors the
Cyclin A-binding and Skp-binding domains and is distinct from the NTD of
E2F6- E2F8. The DP dimerization domain (DIM) is common to E2F1-E2F6, but is absent in E2F7 and E2F8. The Marked (Mark.) domain common to
E2F1-E2F6 is involved in mediating interactions with co-factors. The classical
Rb binding (Rb bind.) domain buried within the activation domain is common to E2F1-E2F5.
15 E2F1 E2F7 E2F2 E2F8 E2F3a
E2F3b DNA DNA E2F4 Replication Replication E2F5
Go G1 S G2 M G1 S G2
Figure 1.2 Schematic representation of the E2F activities during the cell cycle. Typical E2F repressors (E2F4, E2F5 and potential E2F3b) are constitutively expressed throughout the cell cycle, whereas the activator subclass of E2Fs (E2F1, E2F2 and E2F3a) and atypical E2Fs (E2F7 and
E2F8) are regulated in a cell-cycle dependent manner, with E2F1, E2F2,
E2F3a peaking at the G1-S transition, and E2F7, E2F8 peaking at the G2-M transition.
16 mE2F2 mE2F3 mE2F1 mE2F4 mE2F5 mE2F6 mE2F7 mE2F8 156.2 140 120 100 80 60 40 20 0 Nucleotide Substitutions x100
Figure 1.3 Phylogenetic analysis of the mouse E2F transcription factors family. Full-length mouse E2F proteins were analyzed for evolutionary relationship on the basis of their primary structure. As shown above, these
E2Fs segregate into two main branches that consist of E2F1-E2F6 and E2F7-
E2F8. The first branch further separates into two groups that are reflective of their functional characteristics: activators (E2F1-E2F3) and repressors (E2F4-
E2F6).
17
E2F4/DP E2F7 E2F8
DP2 DBD2 DBD2
E2F4 DBD1 DBD1
Figure 1.4 Computer modeling of the interaction between DBD1 and
DBD2 DNA binding domains based on the solved E2F4-DP2 structure.
DBD1 and DBD2 of E2F7 or E2F8 can adopt similar structure as E2F4 and
DP2, with DBD1 and DBD2 in the position of E2F4 and DP2, respectively.
18
CHAPTER 2
SYNERGISTIC FUNCTION OF E2F7 AND E2F8 IS ESSENTIAL FOR
EMBRYONIC DEVELOPMENT AND CELL SURVIVAL
2.1 Introduction
A large body of work suggests that E2Fs function to activate and repress
the transcription of many essential genes involved in cell proliferation, apoptosis
and differentiation (Dimova and Dyson, 2005). The effects of deregulated E2F
activity are pleiotropic and vary in different experimental settings. A decrease in
E2F activity is generally associated with a reduction in the proliferation capacity
of cells (Leone et al., 1998; Humbert et al., 2000; Wu et al., 2001), whereas an
increase in E2F activity is often associated with inappropriate cell proliferation
and/or apoptosis (DeGregori et al., 1997). The regulation of E2F activity during
the cell cycle is thought to be critical for cellular homeostasis. Thus, a significant
amount of genetic currency has been invested to finely control E2F activity in
cells, including by direct binding of the Rb tumor suppressor, by transcription, by
post-transcriptional mechanisms involving miRNAs, and by post-translational mechanisms involving protein degradation, phosphorylation and acetylation
(Muller et al., 2000; O’Donnell et al., 2005).
19 Further complexity in the regulation of E2F function has been afforded by
the evolution of numerous family members and isoforms. The mammalian E2F
proteins are encoded by eight distinct genes (E2F1-8) and specific roles for each
family member in controlling cell cycle transitions and apoptosis have been
reported (Bracken et al., 2004; DeGregori and Johnson, 2006). Based on
structure-function studies and sequence analysis, the E2F family can be
conveniently divided into two subclasses, transcription repressors and activators.
Members of the activator subclass, consisting of E2F1, E2F2 and E2F3, accumulate late in G1 and are transiently recruited to E2F-binding elements on target promoters and participate in their acute activation. Consistent with an important function for activator E2Fs in regulating gene expression during G1-S, their overexpression in quiescent cells can potently transactivate many E2F- responsive genes and drive cells to enter S phase (DeGregori et al., 1997).
Conversely, the combined loss of the three activators results in a decrease of
E2F-target gene expression and a severe block in cell proliferation (Wu et al.,
2001). The repressor subclass consists of E2F4, E2F5, E2F6, E2F7 and E2F8.
A subset of this group that includes E2F4 and E2F5, serves to recruit Rb-related pocket proteins and associated co-repressors to target promoters during G0 and to repress their expression (Attwooll et al., 2004). As cells are stimulated to enter the cell cycle, cyclin-dependent kinases (CDKs) phosphorylate pocket proteins, resulting in dissociation of E2F-pocket protein repressor complexes and derepression of E2F-target genes (Muller et al., 2000; Seville et al., 2005). The
E2F6 repressor is part of a multi-subunit complex that includes polycomb group
20 proteins as well as Mga and Max, and appears to act on a different subset of
target genes than E2F4 (Ogawa et al., 2002; Giangrande et al., 2004). The
mechanism of how E2F7 and E2F8 repress gene expression is less clear. While
the molecular basis for how E2F repressors and activators orchestrate the acute
induction of E2F targets as cells transit through G0-G1-S is fairly well understood, how E2F-target genes are subsequently downregulated as cells
proceed through S-G2 phases of the cell cycle is not known.
The E2F7 and E2F8 proteins are conserved in mice and humans, and
related E2F-like proteins exist in Arabidopsis (Kosugi et al., 2002; Mariconti et al.,
2002; de Bruin et al., 2003; Di Stefano et al., 2003; Logan et al., 2004;
Christensen et al., 2005; Logan et al., 2005; Maiti et al., 2005). These two novel
factors have several unique features that distinguish them from other members in
the E2F family. They lack the leucine-zipper domain required for dimerization
with partner proteins DP1/DP2 and instead possess two DNA binding domains
that are predicted to interact with each other and foster DP-independent DNA-
binding activity. The expression of E2F7 and E2F8 during the cell cycle is also
different from that of other E2Fs, with peak levels found later in the cell cycle
during S-G2. Moreover, their overexpression in fibroblasts can, unlike that of
other E2Fs, repress E2F-target gene expression and block cell proliferation (de
Bruin et al., 2003; Di Stefano et al., 2003; Maiti et al., 2005), suggesting a role for these two E2Fs in controlling cell cycle transitions. Here we utilized homologous recombination strategies to inactivate E2f7 and E2f8 in mice and rigorously
investigate their functions in vivo. From these studies, we conclude that E2f7
21 and E2f8 represent a unique repressive arm of the E2F transcriptional network
that is critical for embryonic development and cell survival.
2.2 Results
2.2.1 E2F7 and E2F8 are essential for embryonic viability
To investigate E2F7 and E2F8 function in vivo, we utilized homologous
recombination techniques and cre-loxp technology to disrupt E2f7 and E2f8
function in mice. Targeting the inactivation of E2f7 and E2f8 was achieved by
flanking exon 4 of E2f7 and exons 3 and 4 of E2f8 with loxp sites (Figure 2.1A).
Cre-mediated recombination of loxp-flanked sequences resulted in the ablation of
domains required for DNA binding and in a shift of the open reading frames, leading to premature termination of translation. Pups lacking the neo cassette
(conditional knockout allele: E2f7loxp or E2f8loxp; Figure 2.1A) or lacking both the
neo cassette and the loxp-flanked regions of E2f7 or E2f8 (conventional knockout
allele: E2f7- or E2f8-; Figure 2.1A) were identified by Southern blot and PCR genotyping analysis (Figure 2.1B and data not shown).
To investigate the requirement for these E2Fs in development, we interbred E2f7+/- or E2f8+/- animals and found that E2f7-/- or E2f8-/- pups were
born. Mutant pups developed normally through puberty and lived to old age
(data not shown). Quantitative RT-PCR analysis of gene expression revealed
that E2f7 and E2f8 mRNA levels were unperturbed in E2f8-/- and E2f7-/- embryos,
respectively (Figure 2.1C), indicating that the absence of abnormalities in single
22 knockout animals is not due to simple compensation at the expression level. We
then explored functional redundancy between E2F7 and E2F8 by examining the
biological consequence of ablating both simultaneously. To this end, we
intercrossed E2f7+/-E2f8+/- animals and analyzed the resulting offspring. Whereas
E2f7-/-E2f8+/- and E2f7+/-E2f8-/- pups were born at the expected Mendelian ratios,
no E2f7-/-E2f8-/- double knockout (DKO) pups were found in newborn litters (P0)
(Table 2.1). This demonstrates that at least one allele of E2f7 or E2f8 is required
for embryonic development and viability. The contribution of E2F7 and E2F8
towards postnatal development, however, does not appear to be equal. Young
and adult E2f7+/-E2f8-/- mice were developmentally normal, whereas most E2f7-/-
E2f8+/- animals appeared runted (Figure 2.2A) and died within the first three
months of life (Figure 2.2B).
2.2.2 Genetic ablation of E2f7 and E2f8 in vivo induces massive apoptosis in the fetus
Analysis at earlier stages of embryonic development showed that all DKO embryos had a beating heart at embryonic day 9.5 (E9.5) (Table 2.1). These embryos were noticeably smaller than wild type littermates (Figure 2.2C, top panels), but macroscopic inspection did not reveal other obvious defects. By
E10.5, only 46% of DKO embryos were alive and by E11.5 all DKO embryos were dead (Table 2.1). Live E10.5 embryos often had vascular defects in the yolk sac and in the embryo proper, which were characterized by large dilated blood vessels associated with multifocal hemorrhages (Figure 2.2C, bottom
23 panels). Other phenotypes were also manifested in tissues distinct from the
vasculature. Specifically, multiple areas within the head mesenchyme, branchial
arch, somites and neural tube of DKO fetuses contained cells with pyknotic
nuclei surrounded by bright eosinophilic cytoplasm (Figure 2.2D).
These latter observations prompted us to examine proliferation and
apoptosis in E2f7-/-E2f8-/- embryos more closely. Since a significant fraction of
DKO embryos died by E10.5, proliferation and apoptosis were measured in E9.5 embryos. When assessed by immunohistochemistry using BrdU-specific antibodies, we observed no detectable difference in the percentage of cells incorporating BrdU between wild type and DKO embryos (Figure 2.3A, 2.3B).
We did observe, however, a marked increase in cells labeled by TdT-mediated dUTP nick end-labeling (TUNEL) in areas of DKO embryos previously noted to contain abundant pyknotic nuclei, confirming widespread apoptosis in these tissues (Figure 2.3C, 2.3D). In summary, global deletion of E2f7 and E2f8 resulted in a spectrum of embryonic defects impacting the vasculature and cell survival.
2.2.3 E2F7 and E2F8 form homo-dimers and hetero-dimers
Previous studies suggested that E2F7 and E2F8 form homo-dimers (Di
Stefano et al., 2003; Logan et al., 2004; Maiti et al., 2005). Co- immunoprecipitation (Co-IP) assays using HEK 293 cells co-expressing flag- tagged and HA-tagged versions of E2F7 or E2F8 confirmed their ability to form homo-dimers (Figure 2.4A and data not shown). We also explored the possibility
24 that E2F7 may physically interact with E2F8. Because immunoprecipitation-
quality antibodies for E2F7 and E2F8 are not yet available, we assessed the
ability of flag-tagged E2F7 to complex with HA-tagged E2F8 and vice versa. As
shown in Figure 2.4B, these Co-IP assays revealed hetero-dimerization between
E2F7 and E2F8.
Given the redundant functions of E2F7 and E2F8 in development, we
evaluated their preferred dimerization state in cells. To this end, HEK 293 cells
were co-transfected with flag-E2F7, HA-E2F7 and myc-E2F8, or alternatively
with flag-E2F8, HA-E2F7 and myc-E2F8. We then measured the relative
amounts of homo- versus hetero-dimers by immunoprecipitation with flag-
antibodies followed by immunoblotting with either HA- or myc-antibodies; the amount of E2F7 and E2F8 on blots was normalized to 1% of the input material
(Figure 2.4C). Quantification of three independent experiments showed that
E2F7 had a greater binding affinity to itself than to E2F8. On the other hand,
E2F8 had a greater binding affinity for E2F7 than for itself (Figure 2.4D). From this analysis we conclude that, at least in cultured cells, the preferred state of
dimerization is E2F7:E2F7 > E2F7:E2F8 > E2F8:E2F8.
2.2.4 E2f1 is a direct target of E2F7 and E2F8
We then utilized chromatin immunoprecipitation (ChIP) assays to assess
the ability of E2F7 and E2F8 to bind known E2F target promoters. Quantitative
PCR assays showed that flagged-tagged versions of E2F7 and E2F8 were
recruited to E2F binding sites on the E2f1 promoter but not to irrelevant
25 sequences in this gene (exons 1) or the tubulin promoter (Figure 2.5A, 2.5B).
This recruitment was specific, since IgG failed to immunoprecipitate target promoter sequences. Moreover, anti-flag antibodies failed to immunoprecipitate
target promoters from cell lysates expressing mutant forms of E2F7 or E2F8 that
are incapable of binding DNA.
We then used sequential ChIP techniques to address whether E2F7 and
E2F8 could be recruited to the target promoters as homo- and hetero-dimers. To
this end, HEK 293 cells expressing various combinations of HA-tagged and flag-
tagged versions of E2F7 and E2F8 were first immunoprecipitated with anti-flag
antibodies. Immunoprecipitated DNA-protein complexes were then eluted with
excess flag peptide and re-immunoprecipitated with HA-specific antibodies. The
final immunoprecipitates were amplified with primers specific for the E2f1 or
tubulin promoters as described above (Figure 2.5C-2.5E). These sequential
ChIP assays showed that both homo- and hetero-dimers of E2F7 and E2F8
could occupy E2F binding sites on E2F-target promoters, including E2f1.
2.2.5 E2f1 expression is derepressed in E2f7-/-E2f8-/- MEFs
To determine whether the recruitment of E2F7 and E2F8 to the target
promoters had any functional consequence on its expression, we initially examined E2F1 protein and mRNA levels in mouse embryo fibroblasts (MEFs) deficient for both E2f7 and E2f8. Because E2f7-/-E2f8-/- embryos died early
during mouse development, we derived MEFs from E13.5 E2f7loxp/loxpE2f8loxp/loxp
embryos and then ablated E2f7 and E2f8 expression in vitro using a cre
26 recombinase-expressing retrovirus. PCR genotyping of genomic DNA and quantitative RT-PCR analysis of E2f7 and E2f8 expression confirmed their efficient deletion (Figure 2.6A). These experiments showed that DKO cells have higher E2F1 protein and mRNA levels than control treated MEFs (Figure 2.6B and 2.6C, respectively). Interestingly, there was also an increase of p53 protein in DKO MEFs, consistent with the ability of E2F1 to induce the accumulation of p53 (Pomerantz et al., 1998; Hsieh et al., 2002; Rogoff et al., 2002; Russell et al.,
2002). Together with the ChIP assays shown in Figure 2.5, these data suggest that the repression of E2f1 by E2F7 and E2F8 is direct.
To examine whether E2F7/E2F8-mediated repression might be cell-cycle dependent, we compared E2F-target expression in synchronized populations of wild type and DKO MEFs stimulated to progress through the cell cycle. MEFs were synchronized in G1-S by serum deprivation followed by re-stimulation with medium containing 15% serum and 1mM hydroxyurea (HU). Cells were then washed and incubated with medium lacking HU and harvested at various times following HU-release. Cell cycle progression was monitored by flow cytometry
(Figure 2.6D). As expected, the expression of E2f1 in wild type cells peaked at
G1-S and subsequently decreased as cells entered S phase and progressed through G2 (Figure 2.6E) Strikingly, expression of E2f1 in DKO cells continued to increase during S and G2, accumulating up to 12-fold higher levels than in wild type cells. These data suggest an important role for E2F7/E2F8-mediated repression during S and G2, coinciding with the time of the cell cycle when these two E2Fs accumulate maximally (de Bruin et al., 2003; Maiti et al., 2005).
27 2.2.6 E2f7-/-E2f8-/- cells are hypersensitive to DNA damage
Given the marked increase in the expression of multiple E2F-target genes
in cells deficient for E2f7 and E2f8, we analyzed proliferation rates in these
double mutant cells. These analyses failed to reveal any significant difference in
the proliferation of DKO and wild type cells (Figure 2.7A), even though DKO cells
appeared to transit through S phase 2-3 hours faster than control cells (Figure
2.6D). Presumably, faster progression through S phase was compensated by a
concomitant delay in other phases of the cell cycle as previously reported for
cells lacking Rb or overexpressing dE2f1 (Resnitzky et al., 1994; Reis and Edgar,
2004). Because overexpression of E2f1 has been shown to elicit apoptosis in cells treated with DNA-damage inducing agents (Stevens and La Thangue,
2004), we tested the sensitivity of DKO cells to camptothecin and cisplatin, two
DNA-damage-inducing drugs. To this end, asynchronously proliferating wild type and DKO MEFs were treated with camptothecin or cisplatin and cell viability was determined by trypan blue exclusion. These drugs induced a significant acceleration of cell death in DKO MEFs when compared to wild type MEFs
(Figure 2.7B and data not shown). Drug-treated DKO cells detached from tissue culture plates and exhibited the characteristic blebbing morphology of apoptotic cells (Figure 2.7C). Consistent with this observation, drug treatment preferentially triggered cleavage of caspase-3 in DKO MEFs (Figure 2.7D). We
also evaluated the levels of E2F1 and p53 in drug-treated DKO MEFs. As
expected, E2F1 and p53 protein levels accumulated to higher levels in camptothecin-treated DKO cells than in similarly treated wild type cells (Figure
28 2.7E). The increase in p53 protein corresponded with a marked increase in the
expression of p53-responsive genes, including gadd45, noxa, and pidd (Figure
2.7F). These results suggest that E2F7 and E2F8 may attenuate DNA-damage
induced apoptosis by preventing the aberrant accumulation of E2F1 and p53.
2.2.7 Induction of apoptosis in E2f7-/-E2f8-/- embryos is dependent on E2F1 and p53
Given the above observations, we hypothesized that the apoptosis in
E2f7-/-E2f8-/- embryos might be mediated through the induction of E2F1 and/or
p53. To test this possibility, E2f7+/-E2f8-/- animals were bred with either E2f1-/- or
p53-/- animals in order to generate cohorts of triple knockout embryos (TKO).
TUNEL assays performed on E9.5 TKO embryos showed that the loss of either
E2f1 or p53 suppressed the massive apoptosis caused by a deficiency in E2f7
and E2f8 (Figure 2.8A, 2.8B). From these results, we conclude that E2F7 and
E2F8 represent a critical regulatory arm of the E2F network that controls apoptosis through the E2F1-p53 axis.
Interestingly, both cohorts of TKO embryos harvested at E10.5 had dilated vessels and extensive hemorrhaging similar to DKO embryos (data not shown).
Importantly, no live TKO embryos could be observed at E12.5 (Figure 2.8C,
2.8D). These results suggest that the widespread apoptosis observed in DKO embryos, which was suppressed in TKO embryos, is independent of vascular defects and embryonic lethality. These results also indicate that misregulated
29 E2F7/E2F8-target genes beyond E2f1 are likely to be involved in the mortality of
DKO embryos.
Consistent with this view, global analysis of gene expression showed that among the ~ 39,000 transcripts analyzed, 88 were upregulated and 33 were downregulated at least 3-fold in E2f7-/-E2f8-/- embryos relative to wild type embryos (Figure 2.9A). Quantitive RT-PCR assays confirmed the altered expression of many of these genes (Figure 2.9B). Functional annotation of misregulated genes revealed a bias for gene products known to be activated in response to stresses, including hypoxia, nutrient deprivation and apoptosis
(Figure 2.9C). These analyses also confirmed that E2f1 and additional E2F targets were increased in DKO versus wild type embryos, albeit the increase in their expression was only ~2-fold (data not shown). Interestingly, a subset of the
122 transcripts (88 up- and 33 down-regulated) were partially misregulated in
E2f7-/-, E2f8-/-, and/or E2f7+/-E2f8+/- embryos, underscoring the synergistic role of these two E2F factors in the control of gene expression during embryonic development.
2.3 Discussion
The E2F7 and E2F8 transcription factors likely represent the final members of the E2F family to be identified in mammals. We show here that these two novel factors are strictly required for embryonic development and are critical direct regulators of the E2F1-p53 apoptotic axis.
30 While disruption of E2f7 or E2f8 had little impact on mouse development,
their combined ablation resulted in widespread apoptosis, vascular defects and
hemorrhaging, leading to embryonic death by E11.5. Provision of even one
functional allele from either locus was sufficient to carry fetuses through
development all the way to birth. The contribution of E2f7 and E2f8 in postnatal
development, however, does not appear to be equal. Young and adult E2f7+/-
E2f8-/- mice were developmentally normal, whereas most E2f7-/-E2f8+/- animals appeared runted and died within the first three months of life. A bias in homo- versus hetero-dimerization may explain the differential requirement for E2f7 and
E2f8 in postnatal development. We found that in tissue culture experiments, under conditions where expression levels can be compared (by epitope tagging) and experimentally equalized, the formation of E2F7:E2F7 homo-dimers was preferred over E2F7:E2F8 hetero-dimers, and E2F8:E2F8 homo-dimers appeared to be the least preferred state (E2F7:E2F7 > E2F7:E2F8 >E2F8:E2F8).
While the molecular basis for homo- versus hetero-dimerization is not yet clear, these data suggest that inefficient homo-dimerization may compromise the ability of E2F8 to compensate for the loss of E2F7, an effect that might be aggravated in circumstances of limiting amounts of E2F8 (i.e. E2f7-/-E2f8+/-). Thus, the
observed bias for E2F7 homo-dimerization may explain the stricter postnatal
requirement for this subunit. While this interpretation is likely an
oversimplification, these results do provide a molecular explanation for their
functional redundancy in development. It will be interesting to investigate how
31 dimerization might be impacted by the levels of E2F7 and E2F8 proteins in vivo
or by tissue-specific signals.
Too much or unrestrained E2F activity can also compromise cell
homeostasis. The importance of restraining E2F activity is highlighted by the fact
that at least four independent mechanisms have evolved to achieve this. One
such repressive mechanism involves the binding and inhibition of E2F proteins
by Rb tumor suppressor proteins. Indeed, disruption of Rb leads to unregulated
proliferation and ectopic apoptosis that is partly suppressed by the concomitant
loss of E2F activators (Ziebold et al., 2001; Saavedra et al., 2002). A second
mechanism involves binding of CyclinA/CDK2 to E2F1, E2F2 and E2F3 proteins
during S phase, leading to the phosphorylation of their dimerization partners
DP1/DP2 and the inhibition of E2F/DP DNA binding activity (Dynlacht et al.,
1994). A third mechanism involves the p45SKP2/F-box-dependent degradation
of the three E2F activators during S phase (Leone et al., 1998; Marti et al., 1999).
Here we described a fourth repressive mechanism to keep E2F activity in control
that involves direct transcriptional repression of E2f1 by E2F7 and E2F8. In
contrast to other repressor E2Fs (E2F4-E2F6), which play a predominant role in
silencing gene expression in G0-G1, these two novel E2Fs may have a
particularly important role in repressing E2F targets as cells transit through S
phase and into G2. Gene expression analysis in synchronized E2f7-/-E2f8-/-
MEFs revealed that E2f1 mRNA levels, as well as that of many other E2F-target genes, increased acutely during S and G2. Thus, by participating in the repression of E2F-target genes as cells transit through S-G2, E2F7 and E2F8
32 represent the first transcriptional mechanism described in mammals that
contributes directly to the downswing in the oscillating pattern of cell cycle-
specific gene expression.
In vitro and in vivo experiments described here provide clear-cut evidence
for a role of E2F7 and E2F8 in the control of apoptosis that involves, at least in
part, the regulation of E2f1 expression. Genetic inactivation of E2f7 and E2f8
resulted in the accumulation of E2f1 mRNA and a corresponding increase in its
protein product. Consistent with a direct role in repression, ChIP experiments
showed that E2F7 and E2F8 are recruited to the E2f1 promoter and that this
requires intact DNA binding ability. The increase of E2F1 protein levels in E2f7-/-
E2f8-/- cells coincided with the accumulation of p53 protein. Several mechanisms
of how E2F1 may lead to the accumulation and activation of p53 have been described (Pomerantz et al., 1998; Hsieh et al., 2002; Rogoff et al., 2002; Russell et al., 2002). The increase of E2F1 and p53 in E2f7-/-E2f8-/- cells is of
physiological significance since ablation of either E2f1 or p53 suppressed the
widespread apoptosis observed in DKO embryos. The fact that these TKO
embryos still died suggests that apoptosis in E2f7-/-E2f8-/- embryos is not simply
due to an indirect consequence associated with embryonic lethality but rather
due to a specific signal emanating from deregulated E2F1. These results also
indicate that additional targets repressed by E2F7 and E2F8 must be involved in
the many pathologies that likely underlie the lethality of DKO embryos. Indeed,
profiling of global gene expression in DKO embryos confirmed the deregulation
of many additional genes. Interestingly, a substantial portion of these included
33 gene products known to be involved in various responses to stress such as hypoxia, nutrient deprivation and apoptosis. We view these results to mean that physiological stress, whether induced by vascular or other defects in DKO embryos or by the addition of DNA damaging agents in DKO MEFs, exacerbates the activation of the E2F1-p53 axis and unleashes a massive apoptotic response.
In summary, we show that E2F7 and E2F8 are essential for embryonic development. Their synergistic function may be viewed as a distinct arm of the
E2F network involved in repression of transcription during S-G2, where E2f1 represents a particularly critical target that if not appropriately repressed can elicit widespread apoptosis in developing embryos.
2.4 Materials and Methods
2.4.1 Generation of E2f7 and E2f8 knockout mice
E2f7 and E2f8 specific probes were used to screen a 129Sv/Ev bacterial artificial chromosome library. A 7.0 kb XbaI fragment of E2f7 spanning exon 4 and 5 was isolated from RPCI 22 431 H17, and a 6.6 kb fragment of E2f8 containing exon 3 and 4 was isolated from RPCI 22 539 P23. Standard cloning techniques were used to generate E2f7 and E2f8 targeting vectors, which were confirmed by direct sequencing. Targeting vectors were linearized with NotI and electroporated into TC1 129Sv/Ev embryonic stem (ES) cells. ES cells were selected in medium containing G418 plus ganciclovir and correct recombination was confirmed by Southern blots using the indicated probes in Figure 2.1A.
34 Selected ES clones were injected into C57BL/6 blastocysts to generate chimeric
mice, which were bred with Black Swiss females to obtain germline transmission.
Appropriate offspring were then bred to EIIa-Cre mice (Lakso et al., 1996) to
produce mice with conventional and conditional null alleles as illustrated in Figure
2.1A. Genotypic analysis of offspring was performed by Southern blot analysis
and multiplex PCR using allele-specific primers described in Table 2.2.
2.4.2 Quantitive reverse transcriptase PCR (RT-PCR)
Total RNA was isolated using Qiagen RNA miniprep columns as
described by the manufacturer, which included the optional DNase treatment
before elution from columns. Reverse transcription of total RNA was performed
using Superscript III reverse transcriptase (Invitrogen) and RNAse Inhibitor
(Roche) according to the manufacturer’s protocol. Quantitive PCR was performed using a BioRad iCycler and reactions were performed in triplicate and
relative amounts of cDNA were normalized to GAPDH. The RT-PCR primers for
E2f7 and E2f8 are located within the deleted regions. Quantitive RT-PCR primer
sequences are included in Table 2.2.
2.4.3 Affymetrix microarray analysis
Total RNA was prepared from E10.5 embryos using the RNasy Kit
(Qiagen) and included the optional DNase treatment before elution from
purification columns. RNA was processed as described in
www.osuccc.osu.edu/microarray/ and hybridized to oligonucleotide microarrays
35 (Mouse Genome 430 2.0 array, Affymetrix, Santa Clara, CA). Scanned image
files were quantified with GENECHIP 3.2 (Affymetrix). Genes that increased or
decreased at least 3-fold (p<0.008) in DKO samples relative to wild type were
used to generate the heatmaps.
2.4.4 Co-immunoprecipitation (Co-IP) assay
Human embryonic kidney (HEK) 293 cells were cultured in DMEM
medium supplemented with 15% fetal bovine serum (FBS) and used for co-
immunoprecipitation assay. Transient transfections with the indicated constructs
were performed by standard calcium chloride techniques. Cells were harvested and washed in PBS at 4˚C and cell pellets were lysed in 10 volumes of lysis buffer (0.05 M sodium phosphate pH 7.3, 0.3 M NaCl, 0.1% NP40, 10% glycerol with protease inhibitor cocktail, Roche). Lysates were incubated with Protein G
Plus/protein A-agarose beads (Calbiochem) at 4°C for 1 h to preclear. The
precleared lysates were incubated with appropriate antibody overnight. Protein
G Plus/protein A-agarose beads were added and incubated for 1 h at 4°C.
Protein complexes binding to the beads were precipitated and resolved by SDS-
PAGE followed by immunoblotting.
2.4.5 Chromatin immunoprecipitation (ChIP) and sequential ChIP assays
For ChIP assays, the EZ CHIPTM assay kit (Upstate) was used as
described by the manufacturer. Briefly, harvested HEK 293 cells overexpressing flag-E2F7 and/or flag-E2F8 were crosslinked and chromatin was sonicated to an
36 average size of 200-1000 bp. Lysates were subsequently pre-cleared with
Salmon Sperm DNA/Protein G agarose slurry. Antibodies specific to flag, HA, or
normal mouse IgG (Oncogene) were then added to each sample and incubated
overnight at 4°C. Antibody-protein-DNA complexes were recovered by addition
of 30 µl of Salmon Sperm DNA/Protein G agarose slurry and incubation for 1 h at
4°C. Following extensive washing, the complexes were eluted and de- crosslinked at 65°C for 4 h. Finally, samples were treated with Proteinase K
(Roche) and Rnase A (Roche) and purified through Qiaquick columns (Qiagen).
Quantitive PCR of immunoprecipitated DNA was performed using the Biorad
iCycler machine with primers specific for the indicated promoter regions. All
primer sequences are listed in Table 2.2. Reactions were performed in triplicate and normalized to 1% of the total input.
Sequential ChIP-PCR was carried out similarly using the same EZ CHIPTM protocol, except two rounds of chromatin immunoprecipitation steps were performed. The first round of immunoprecipitation was performed by incubation of extracts with M2 anti-flag antibody and Salmon Sperm DNA/Protein G agarose slurry. Precipitated DNA-protein complexes were eluted with excess flag peptide
(200 µg/mL, Sigma) for 1 h at 4°C and then subjected to a second round of precipitation with primary antibodies as indicated in Figure 2.5C-E. After extensive washing, sequentially precipitated complexes were recovered by addition of 30 µl of Salmon Sperm DNA/Protein G agarose slurry and incubation for another 1 h at 4°C. Protein-DNA complexes collected from sequential ChIP
37 were treated and subjected to quantitive PCR analysis as described above for
ChIP experiments.
2.4.6 Western blot and antibodies
Immunoblot analyses were performed by standard procedures using ECL
reagents as described by the manufacturer (Amersham Biosciences). The
following commercial antibodies were used as indicated in the figures: flag (M2,
Sigma), HA (12C5A, Roche), Myc (9E10, Santa Cruz), E2F1 (C-20, Santa Cruz),
tubulin (T6199, Sigma), caspase-3 (9662, Cell Signaling), p53 (Ab-1, Oncogene).
2.4.7 Cell culture and viability assay
E2f7loxp/loxpE2f8loxp/loxp mouse embryo fibroblasts (MEFs) were isolated
from E13.5 embryos and maintained in DMEM medium containing 15% FBS.
Immortalized cell lines were generated from primary MEFs using the standard
3T9 protocol. Immortalized MEFs were treated with retrovirus expressing cre
recombinase using standard methods (Wu et al., 2001).
Apoptosis was measured in MEFs at the indicated times following
treatment with 10 μm cisplatin (Sigma) or 20 μm camptothecin (Sigma) for 18 h.
Cell viability was determined by trypan blue exclusion.
2.4.8 FACS analysis
Cre-treated wild type and E2f7loxp/loxpE2f8loxp/loxp MEFs were synchronized
by starvation in DMEM containing 0.2% FBS for 48 h and blocked at the G1-S
38 transition by the addition of DMEM containing 15% FBS with 1 mM hydroxyurea
(HU) for 18 h. Cells were then washed 3 times with PBS and incubated with
fresh medium containing 15% FBS. Cells were harvested at the indicated time
points and fixed in 70% ethanol, followed by incubation in propidium iodide and
analyzed by flow cytometry using standard methods.
2.4.9 BrdU and TUNEL assays
Pregnant females at 9.5 days postcoitum were injected intraperitoneally
with BrdU (100 ug/grams of body weight) 30 min prior to harvesting. Embryos
were fixed in formalin and 5 μm paraffin embedded-sections were used for
immunohistochemistry. After deparaffinization, anti-BrdU antibody (MO-0744,
DAKO), Vectastain Elite ABC reagents (Vector labs) and DAB peroxidase
substrate kit (Vector labs) were used to detect BrdU incorporation according to the manufacturer’s instructions. Apoptotic cells were detected using TUNEL
(S7101, Chemican) assays, performed according to the manufacturer’s protocol.
All slides were counterstained with hematoxylin.
39 Figure 2.1 Generation of E2f7 and E2f8 knockout mice. (A) Targeting strategies used to inactivate E2f7 (left) and E2f8 (right). Top panels: partial exon and intron structures of the E2f7 and E2f8 genes. The black bars indicate the position of probes used for Southern analysis. Middle panels
(targeting vectors): position of the TK and Neo cassettes, as well as loxP sites
(filled triangles) are indicated. Bottom two panels (conditional knockout alleles and conventional knockout alleles): illustrate the two desired alleles resulting from possible recombination events. (B) Top panels: Southern analysis of genomic DNA isolated from conventional knockout mice using AseI for E2f7
(left) and EcoRV for E2f8 (right), and hybridized using probes A and B, respectively. Bottom panels: genotyping of tail DNA was performed using allele-specific PCR primers. (C) Quantitive RT-PCR analysis of E2f7 or E2f8 expression in embryos with the indicated genotypes using primers described in Table 2.2.
40 A E2f7 genomic locus E2f8 genomic locus
Ase I Ase I EcoRV EcoRV EcoRV 1 2 3 4 5 6 7 1 2 3 4 5 6 7 ProbeA ProbeB
targeting TK 4 PGK-Neo 5 vectors TK 2 3 4 PGK-Neo 5 short floxed long short floxed long arm region arm arm region arm + EIIa-cre + EIIa-cre
conditional 1 2 3 4 5 6 7 knockout alleles 1 2 3 4 5 6 7 or or conventional 1 2 3 5 6 7 knockout alleles 1 2 5 6 7
BCE2f7 E2f8 E2f7 genotype E2f8 genotype 1.2 - +/+ +/+ +/- +/- -/- -/- +/+ +/+ +/- +/- -/- -/- 1.0 - 0.8 - 7kb- wt 14.5kb- ko 0.6 - 6kb- ko 10.1kb- wt 0.4 - 506 - ko 506 - ko 396 - 220 - 0.2 - 298 - wt 154 - wt
Relative gene gene expression Relative 0.0 - +/+ -/- +/+ -/- +/+ -/- +/+ -/- :E2f7 +/+ +/+ -/- -/- +/+ +/+ -/- -/- :E2f8
Figure 2.1
41
Table: Genotypic analysis of embryos during development: E2f7/E2f8 crosses
E2f7+/+ E2f7+/- E2f7-/- total E2f8+/+ E2f8+/- E2f8-/- E2f8+/+ E2f8+/- E2f8-/- E2f8+/+ E2f8+/- E2f8-/- E9.5 6 16 29 16 72(2) 53(1) 12 58(1) 33 299 expected 5 28 22 19 75 56 13 47 34 E10.5 9 21 13 7(1) 45 28(1) 4(1) 28(1) 6(7)a 172 expected 7 20 13 16 43 28 8 23 14 E11.5 - 6 4 2 15(2) 14(2) 2(1) 8(1) 0(5)b 62 expected - 5 5 3 16 13 3 10 8 E12.5 - - 3 - 4 17(2) - 3 0(9)b 38 expected - - 5 - 4 15 - 4 10 P0 7 22 16 24 45 17 5 18 0b 154 expected 8 18 10 18 39 21 10 21 11
a b () number of dead embryos recovered; Exact binomial test: significant (p=0.0015), highly significant (p<0.0007).
Table 2.1 Genotypic analysis of embryos derived from E2f7+/-E2f8+/- or
E2f7+/-E2f8-/- intercrosses at the indicated stages of development.
42 Figure 2.2 Global deletion of E2f7 and E2f8 results in developmental delay, vascular defects and widespread apoptosis in vivo. (A) Pictures of
E2f7+/-E2f8+/-, E2f7+/-E2f8-/- and E2f7-/-E2f8+/- mice at 21 days of age. (B)
Tabulated number of E2f7+/-E2f8-/- and E2f7-/-E2f8+/- mice that live until 30
(P30) and 90 (P90) days of age. (C) Gross pictures of E2f7 and E2f8 double knockout embryos at E9.5 (top panels) and E10.5 (bottom panels). The right panel is a higher magnification view of the vascular defects in E10.5 E2f7-/-
E2f8-/- embryos. (D) H&E staining of E9.5 embryo mesenchymal tissues. The right panel highlights the nuclear morphology of mesenchymal cells in E2f7-/-
E2f8-/- embryos; black arrows point to examples of pyknotic nuclei.
43 AB Viability of E2f7+/-E2f8-/- and E2f7-/-E2f8+/- mice E2f7-/-E2f8+/- E2f7+/-E2f8-/- E2f7-/-E2f8+/- E2f7+/-E2f8-/- P30 57/57 14/19a P90 57/57 9/19b +/- +/- E2f7 E2f8 Numerator represents viable mice and denominator represents total number of mice with the indicated genotypes. Fisher’s exact test: a significant (p=0.001), b highly significant (p<0.001).
C E2f7+/+E2f8+/+ E2f7-/-E2f8-/- E9.5 E9.5
E10.5 E10.5
D E2f7+/+E2f8+/+ E2f7-/-E2f8-/-
Figure 2.2
44
Figure 2.3 Analyses of proliferation and apoptosis in E2f7-/-E2f8-/- embryos. (A) BrdU staining of embryos with the indicated genotypes. Far left panels: low magnification pictures of whole embryos. Right three panels: high magnification images of representative areas demarcated by the box in the low magnification images. (B) Quantification of proliferation in different tissue areas of embryos is presented as the average ± SD percentage of cells that are BrdU-positive. Three sections per embryo and two different embryos were counted for each genotype group. (C) E9.5 embryos with the indicated genotypes were analyzed by TUNEL assays. Far left panels: low magnification pictures of whole embryos. Right three panels: high magnification images of representative areas demarcated by the box in the low magnification images. (D) Quantification of TUNEL-positive cells in the indicated tissue areas are presented as the average ± SD percentage of cells that are TUNEL-positive. Three sections per embryo and at least three different embryos for each genotype were analyzed. Pairwise comparisons were evaluated by two-tailed Student’s t-test (** p<0.02).
45 A head branchial arch somite +/- E2f8 +/- E2f7 -/- E2f8 -/-
E2f7 4x 40x40x 40x
B head branchial arch somite 100- 100- 100-
80- 80- 80-
60- 60- 60-
40- 40- 40-
20- 20- 20-
% BrdU positive n=2 n=2 n=2 n=2 n=2 n=2 0- 0- 0- E2f7+/- E2f7-/- E2f7+/- E2f7-/- E2f7+/- E2f7-/- E2f8+/- E2f8-/- E2f8+/- E2f8-/- E2f8+/- E2f8-/-
C headbranchial arch somite +/- E2f8 +/- E2f7 -/- E2f8 -/-
E2f7 4x 40x40x 40x
D head branchial arch somite 30- 60- 50- ** 25- ** 50- 40- ** 20- 40- 30- 15- 30- 20- 10- 20-
5- 10- 10-
% TUNEL positive n=4 n=4 n=3 n=3 n=3 n=3 0- 0- 0- E2f7+/- E2f7-/- E2f7+/- E2f7-/- E2f7+/- E2f7-/- E2f8+/- E2f8-/- E2f8+/- E2f8-/- E2f8+/- E2f8-/-
Figure 2.3 46 Figure 2.4 E2F7 and E2F8 homo- and hetero-dimerize. (A) Homo- dimerization of E2F7. Lysates from HEK 293 cells transfected with both flag- tagged E2F7 (flag-7) and HA-tagged E2F7 (HA-7) were immunoprecipitated
(IP) and immunoblotted (IB) with anti-flag and anti-HA antibodies as indicated.
Immunoprecipitation with normal mouse IgG was used as a negative control.
(B) Hetero-dimerization of E2F7 and E2F8. Lysates derived from cells overexpressing both flag-E2F7 and HA-E2F8 were immunoprecipitated (IP) with anti-flag antibodies and immunoblotted (IB) with anti-HA antibodies (top panel) or vice versa (bottom panel). Immunoprecipitation with normal mouse
IgG was used as a negative control. (C) Kinetic analysis of dimerization. HEK
293 cells overexpressing the indicated constructs were subjected to anti-flag immunoprecipitation (IP), followed by anti-HA or anti-myc immunoblotting (IB) as indicated. The amount of detected E2F7 and E2F8 was measured by densitometry and quantified relative to 1% of the input material. The relative levels indicated at the bottom of each lane are presented as the average ±SD of 3 independent experiments where input is always equal to 1.00. (D) Bar graph presentation of the kinetic analysis of dimerization shown in (C).
47
AB flag-7 + HA-7 flag-7 + HA-8 IP: α-flag IgG Input IP: α-flag IgG Input IB: α-HA IB: α-HA
IP: α-HA IgG Input IP: α-HA IgG Input IB:α-flag IB:α-flag
C flag-7 + HA-7 + myc-8 flag-8 + HA-7 + myc-8 IP: α-flag IgG Input IP: α-flag IgG Input IB: α-HA IB: α-HA 1.72± 0.33 1.00 1.05± 0.27 1.00
IB:α-myc IB:α-myc 0.90± 0.14 1.00 0.21± 0.16 1.00
D 2.5 1.72± 0.33 2.0
1.5 1.05± 0.27 0.90± 0.14 1.0
0.5 0.21± 0.16 Relative affinities Relative
0.0 E2F7/E2F7 E2F7/E2F8 E2F7/E2F8 E2F8/E2F8
Figure 2.4
48 Figure 2.5 E2F7 and E2F8 bind the E2f1 promoter. (A) E2F7 binds the
E2f1 promoter. Chromatin from cells overexpressing wild type flag-E2F7 (wt) or flag-E2F7DBD1,2 (mut) was immunoprecipitated (IP) with anti-flag or IgG control antibodies. Immunoprecipitated DNA was amplified using primers specific for the E2f1 promoter (E2f1), irrelevant sequences in exon 1 of E2f1 and the tubulin promoter (control and tubulin, respectively). (B) E2F8 binds the E2f1 promoter. Cells overexpressing wild type flag-E2F8 (wt) or flag-
E2F8DBD1,2 (mut) was immunoprecipitated (IP) with anti-flag or normal mouse IgG antibodies. Immunoprecipitated DNA was amplified using primers specific for the E2f1 promoter (E2f1), exon 1 of E2f1 (control) and the tubulin promoter (tubulin). (C-E) Homo-dimers and hetero-dimers of E2F7 and E2F8 occupy the E2f1 promoter. Cell extracts expressing ectopic HA-E2F7 and flag-E2F7 (C), HA-E2F8 and flag-E2F8 (D), HA-E2F7 and flag-E2F8 (E) were used for sequential ChIP assays as described in Materials and Methods.
Antibodies used for the first and second round of immunoprecipitation are indicated (1st IP and 2nd IP, respectively). Immunoprecipitated DNA collected after two rounds of ChIP was amplified using primers specific for the E2f1 promoter (E2f1) or for the tubulin promoter (tubulin). For all the ChIP experiments, quantitive PCR was performed in triplicate and cycle numbers were normalized to 1% of the input DNA.
49 AB E2f1 control tubulin E2f1 control tubulin 0.06 0.06- .06 0 0.06 -10 0
0.05 .05 0 0.05 - 0.05 -08 0
0.04 0.04- .04 0 0.04 - 0 06 0.03 0.03- .03 0 0.03 - 0 04 0.02 0.02- .02 0 0.02 - 0
02
% of total input 0.01 0.01 - .01 0 % of total input 0.01 - 0
0.00 -0 0 0 0.00 -00 0 flag IgG flag flag IgG flag flagFH7 α FLAG IgGFH7 α mIgG flagFH7D α FLAG : IP flag IgG flag flag IgG flag flag IgG flag : IP wt mutwt mut wt mut wt mutwt mut wt mut flag-7 flag-8
C D E E2f1 tubulin E2f1 tubulin E2f1 tubulin 0.004 0.008 0.04
0.003 0.006 0.03
0.002 0.004 0.02
0.001 0.002 0.01 % of total input % of total input % of total input
0.000 0.000 0.00 flag flag flag flag :1st IP flag flag flag flag :1st IP flag flag flag flag :1st IP HA IgG HA IgG :2nd IP HA IgG HA IgG :2nd IP HA IgG HA IgG :2nd IP HA-7 + flag-7 HA-8 + flag-8 HA-7 + flag-8
Figure 2.5
50 Figure 2.6 Deregulation of E2F1 and p53 expression in MEFs deficient in
E2f7 and E2f8. (A) Top panels: PCR genotyping analysis of DNA isolated from E2f7+/+E2f8+/+ and E2f7loxp/loxpE2f8loxp/loxp MEFs treated with cre- expressing retroviruses. The knockout (ko) and wild type (wt) alleles are indicated. Bottom graphs: quantitive RT-PCR analysis of E2f7 and E2f8 expression in cells. For convenience, cre-deleted loxp alleles were labeled as
(-/-) at bottom of graphs. (B) Western blot analysis of lysates described in (A) using antibodies specific for E2F1 and p53 as indicated. Tubulin-specific antibodies were used as an internal loading control. (C) Expression of E2f1 in
MEFs treated as in (A) was measured by quantitive RT-PCR. (D) FACS analysis of synchronized cre-treated E2f7+/+E2f8+/+ and E2f7loxp/loxpE2f8loxp/loxp
MEFs. For convenience, cre-deleted loxp alleles were labeled as (-/-). MEFs were synchronized by serum deprivation and hydroxyurea (HU) block as described in Materials and Methods. At the indicated time points, cells were harvested and stained by propidium iodide for flow cytometry. Q: quiescent cells kept in serum-deprived medium. (E) Total RNA isolated from same synchronized MEFs samples as in (D) was analyzed by quantitive RT-PCR assays specific for E2f1 expression.
51 ABC E2f7 E2f8 +/+ -/- :E2f7 E2f1 ko +/+ -/- :E2f8 8- wt
1.2 - E2f7 E2f8 E2F1 6-
0.8 - p53 4-
2-
Relative Relative 0.4 - gene expression tubulin induction inFold gene expression 0.0 - 0- +/+ -/- +/+ -/-:E2f7 +/+ -/- :E2f7 +/+ -/- +/+ -/-:E2f8 +/+ -/- :E2f8
D
E2f7+/+ E2f8+/+
E2f7-/- E2f8-/-
Q 0 2 4 6 8 10 12 Hours post-HU release
E E2f1 14 - +/+ +/+ 12 - E2f7 E2f8 E2f7-/-E2f8-/- 10 - 8 - 6 - 4 - Fold induction in gene expression 2 - 0 - Q024 681012 Hours post-HU release
Figure 2.6
52 Figure 2.7 MEFs deficient in E2f7 and E2f8 proliferate well but are hypersensitive to DNA damage induced apoptosis. (A) Growth curve of cre-treated E2f7+/+E2f8+/+ and E2f7loxp/loxpE2f8loxp/loxp MEFs. Cells were plated and viable cells counted daily in triplicate. For convenience, cre-deleted loxp alleles were labeled as (-/-). (B) Cre-treated E2f7+/+E2f8+/+ and
E2f7loxp/loxpE2f8loxp/loxp MEFs were mock treated with DMSO (-camptothecin) or with camptothecin (+camptothecin). At the indicated time intervals cells were harvested and counted in triplicate as described in Materials and Methods.
The percentages of viable cells at the indicated time points are presented. (C)
Light microscopy images of MEFs treated as in (B) at 72h. (D) Lysates derived from MEFs treated as in (B) were analyzed at indicated time points by
Western blotting using caspase-3 specific antibodies. (E) Lysates derived from MEFs treated as in (B) were analyzed by Western blotting using E2F1 and p53 specific antibodies. Tubulin-specific antibodies were used as an internal loading control. (F) Total RNA was extracted from cells treated for
18h as in (B) and expression of the indicated p53 target genes was measured by quantitive RT-PCR.
53 A B - camptothecin + camptothecin 10 - 100 -
) +/+ +/+ 5 8 - E2f7 E2f8 80 - E2f7-/-E2f8-/- 6 - 60 -
4 - 40 - E2f7+/+E2f8+/+ cells (x 10 (x cells -/- -/- % viable cells 20 - E2f7 E2f8
Total number of 2 -
0 - 0 - 0 1 2 3 4 5 6 7 0 24 48 72 96 120 0 24 48 72 96 120 Days after plating Hours after treatment Hours after treatment CD E2f7+/+E2f8+/+ E2f7-/-E2f8-/- E2f7+/+E2f8+/+ E2f7-/-E2f8-/- 0 12 18 24 0 12 18 24 : hrs -camp. -camp. uncleaved caspase-3
cleaved caspase-3 +camp.
E F E2f7+/+E2f8+/+ E2f7-/-E2f8-/- gadd45 noxa pidd 25- 30- 90- 0 18 24 0 18 24 : hrs 20- 25- 70- E2F1 20- 15- 50- 15- p53 10- 30- 10-
5- 5- 10- Fold induction of gene expression
tubulin 0- 0- 0- -camp + camp - camp + camp - camp + camp E2f7+/+E2f8+/+ E2f7-/-E2f8-/-
Figure 2.7
54 Figure 2.8 Loss of E2f1 or p53 suppresses apoptosis in E2f7-/-E2f8-/- embryos. (A) Micrographs of TUNEL stained E9.5 E2f7-/-E2f8-/-E2f1-/- (top panels) and E2f7-/-E2f8-/-p53-/- embryos (bottom panels). Far left panels: low magnification pictures of whole embryos. Right three panels: high magnification images of representative areas demarcated by the box in the low magnification images. (B) Quantification of TUNEL-positive cells in the indicated tissue areas are presented as the average ± SD percentage of cells that are TUNEL-positive. Three sections per embryo and at least three different embryos for each genotype were analyzed. For comparison purposes, data derived from E2f7+/-E2f8+/- and E2f7-/-E2f8-/- in Figure 2.3D was included. (C) Genotypic analysis of embryos derived from E2f7+/-E2f8-/-E2f1+/- intercrosses at the indicated stages of development. (D) Genotypic analysis of embryos derived from E2f7+/-E2f8-/-p53+/- intercrosses at the indicated stages of development.
55 A head branchial arch somite
E2f7-/- E2f8-/- E2f1-/-
E2f7-/- E2f8-/- p53-/- 4x 40x 40x 40x
B head branchial arch somite 60-60 30-30 50-50 ** ** 50-50 ** 25-25 40-40
40-40 20-20 30-30 30-30 15-15 20-20 20-20 10-10
10-10 10-10 5-5 % TUNEL positive % TUNEL n=4 n=4 n=3 n=3 n=3 n=3 n=3 n=3 n=3 n=3 n=3 n=3 0-0 0-0 0-0 E2f7+/- E2f7-/- E2f7-/- E2f7-/- E2f7+/- E2f7-/- E2f7-/- E2f7-/- E2f7+/- E2f7-/- E2f7-/- E2f7-/- E2f8+/- E2f8-/- E2f8-/- E2f8-/- E2f8+/- E2f8-/- E2f8-/- E2f8-/- E2f8+/- E2f8-/- E2f8-/- E2f8-/- E2f1-/- p53-/- E2f1-/- p53-/- E2f1-/- p53-/-
C Genotypic analysis of embryos during development: E2f7/E2f8/E2f1 E2f1+/+ E2f1+/- E2f1-/- total E2f7-/-E2f8-/- E2f7-/-E2f8-/- E2f7-/-E2f8-/- E9.5 - 2 5 46 expected - 5 6 E12.5 - - 0(4)a 10 expected - - 3
() number of dead embryos; Exact binomial test: a highly significant (p<0.0007)
D Genotypic analysis of embryos during development: E2f7/E2f8/p53 p53+/+ p53+/- p53-/- total E2f7-/-E2f8-/- E2f7-/-E2f8-/- E2f7-/-E2f8-/- E9.5 3 6 4 66 expected 4 8 4 E12.5 0 0(1) 0(4)a 12 expected 1 2 1
() number of dead embryos; Exact binomial test: a highly significant (p<0.0007).
Figure 2.8 56 Figure 2.9 Microarray analysis of E10.5 embryos. (A) Heat maps of probe sets in Affymetrix microarrays that showed at least a 3-fold induction (left and middle) or a 3-fold reduction (right) of expression in E2f7-/-E2f8-/- relative to wild type embryos. Data are presented as the medium of 4 embryos from wild type, E2f7-/-E2f8+/+ and E2f7+/+E2f8-/- genotype groups, and the medium of 2 embryos from E2f7+/-E2f8+/- and E2f7-/-E2f8-/- genotype groups. Red and green indicate induction and reduction respectively. Probe sets representing genes of interest are indicated. (B) The expression of 8 target genes (6 up- regulated and 2 down-regulated) was evaluated by quantitive RT-PCR assays.
(C) Pie chart diagrams illustrate the non-random distribution of the stress- related genes among the total 88 up-regulated genes in DKO embryos.
57 A +/+ +/- -/- +/+ -/- :E2f7 +/+ +/- -/- +/+ -/- :E2f7 +/+ +/- -/- +/+ -/- :E2f7 +/+ +/- +/+ -/- -/- :E2f8 +/+ +/- +/+ -/- -/- :E2f8 +/+ +/- +/+ -/- -/- :E2f8 Stc2 2310056P07Rik Ppp1r1c Klk1b22 /// Klk1b9 Dtprp Slc17a6 Ier3 Krt19 ------Ppp1r3g Slc1a5 2610528A11Rik 4933409K07Rik /// --- Pdk1 Ero1l Eomes Bhlhb2 Ankrd1 6330527O06Rik Trib3 Mt1 AI661323 Ddit4 Plekha2 Bcl11a Ndrg1 Sesn2 Lect1 Adm Rad51l1 Ttyh1 /// Taf1 Ppp1r3c A2m /// LOC677369 Gsc Ankrd37 --- A530088I07Rik Vldlr Dbp AI606473 Ndrl Mt2 B230114P17Rik Ndrg1 /// Ndrl Smyd3 Ptpro A330076H08Rik ------BC064011 5830408C22Rik Myog Pfkp Ndrg2 Hyou1 Aldoc Punc Tial1 Bnip3 Flt1 Cntn2 Rgs11 Slc2a1 4930527B16Rik Pkp2 P4ha1 Jakmip2 Maff Pfkl Mtap1b Stc1 C330008K14Rik Dmn 4930583H14Rik Hk2 --- 2900016B01Rik Tcfl5 Dlx1as 2700089E24Rik St3gal1 --- BC062258 P4ha2 Apc2 Eno2 Pvr --- 2210418O10Rik Cbln3 D7Ertd316e Pfkfb3 --- Golt1b Slc16a3 --- AI854517 Gm129 LOC676974 Nppb Tmem45a Egln3 B230112C05Rik Zc3h6 Vegfa Aox4 Loxl2 Uck2 --- -3.0 0.0 3.0 Slc2a3 Pck2 Scmh1 Upp1 --- Ddit3 Rora Tnfaip3 p < 0.008 Pgm2 Asah3l
B 16-0 6-0 50-0 40- 0 0 Ndrg1 0 Eco1L Eomes Lect1 40-0 12-0 30- 0 4-0 0 30-0 8-0 0 0 20- 0 20-0 0 2- 0 4-0 10- 0 0 10- 0 0 0-0 0-0 0-0 0- +/+ +/- -/- +/+ -/- +/+ +/- -/- +/+ -/- :E2f7 8-0 12-0 +/+ +/- +/+ -/- -/- +/+ +/- +/+ -/- -/- :E2f8 0 Pfkp0 Slc2a3 6-0 0 0 8-0 4-0 0 0 4-0 C 2-0 0 0 0 0-0 0-0 Unknown Stress 8-0 12-0 26.1% (hypoxia,
Relative gene expression gene Relative 0 Bnip30 Trib3 6- nutrient, 0 8-0 apoptosis 4-0 0 Others 0 4-0 2- 25.0% 48.9% 0 0
0-0 0-0 +/+ +/- -/- +/+ -/- +/+ +/- -/- +/+ -/- :E2f7 +/+ +/- +/+ -/- -/- +/+ +/- +/+ -/- -/- :E2f8
Figure 2.9
58
Gene Forward Primer Reverse Primer Genotyping E2f7 AGGCAGCACACTTGACACG ACTTTTGGGACAGAGGTAGGA CCAAGATGAAGGCCGAGATGCTAC E2f8 TAAAAAGCTTTGCGGTCGTT AAGCCAACCTCGATGAATTG CTCGCATCATCGTCTGCTAA Quantitive RT-PCR E2f7 GCCAAGCAGGAAACAGAAGA ACCGTGCCAACCATACTGAT E2f8 GAGAAATCCCAGCCGAGTC CATAAATCCGCCGACGTT E2f1 GCCCTTGACTATCACTTTGGTCTC CCTTCCCATTTTGGTCTGCTC Cdc6 AGTTCTGTGCCCGCAAAGTG AGCAGCAAAGAGCAAACCAGG Gadd45 ACGACATCAACATCCTGCGG CAAAGTCATCTCTGAGCCCTCG Noxa GATGAGGAGCCCAAGCCCAACC CCCAAACGACTGCCCCCATACAA Pidd GCACCGTGTGAATCTCATTGC CAGGAAGTGAACCCCGATAAAAG Ndrg1 TCTTTGAGGCAGAGGGAGAA CAATGAAATCACACCCACCA Eco1L CCGAAAAACTGATCGCAAAT CAGAAACAGGCACATTCCAA Pfkp GGTTACTTGGCCTTGGTGAG CGATTGCTCCTTCAGACACA Slc2a3 AACTGTCCCCTCCTCCACTT GCCCCTTTCCATAGCAATCT Bnip3 GGGTTTTCCCCAAAGGAATA GACCACCCAAGGTAATGGTG Trib3 ACTTGGCTGTGGGATTCAAG GACTGTGGGCCTGGGTACTA Eomes GTGACAGAGACGGTGTGGAGG AGAGGAGGCCGTTGGTCTGTGG Lect1 CACCAGCAGGAAGGAGAAAG GGATTTACACCATGCCCAAG Quantitive ChIP-PCR E2f1-promoter CTGCCTGCAAAGTCCCGGCCACTT AGGAACCGCCGCCGTTGTTCCCGT E2f1-exon1 CGCCCAGACGCCACTTCATC TTCATTCCCTCACTCATTCAACAA Tubulin-promoter ATGGAGGGATGAATGGTTATGC CTTTTTGGGTCTGGCTTCTTTCAC Cdc6-promoter AAAGGCTCTGTGACTACAGCCA GATCCTTCTCACGTCTCTCACA
Table 2.2 Primers used in the experiments described in this chapter.
59
CHAPTER 3
EXTRA-EMBRYONIC FUNCTION OF E2F7 AND E2F8 IS ESSENTIAL FOR
FETAL SURVIVAL
3.1 Introduction
The placenta is one of the organs that distinguish eutherian mammals from other classes in the animal kingdom. Often referred to as the extra- embryonic compartment, it is the first organ to develop during mammalian embryogenesis, and serves as an interface between mother and fetuses for them to exchange oxygen, nutrients and wastes. The placenta is a complex tissue, which is primarily comprised of three specialized trophoblast cell lineages: trophoblast giant cell (TG), spongiotrophoblast (SP) and labyrinth trophoblast (LT)
(Rossant and Cross, 2001). These extra-embryonic trophoblast cells not only provide structural components to bring fetal and maternal components into close contact, but also release a variety of hormones and cytokines important for placental development. It is believed that all subtypes of trophoblasts are differentiated from trophoblast stem cells (TS) residing in the chorionic plate.
In mice with targeted gene disruption, defective placentation is often associated with fetal growth retardation and mid-gestation lethality (Watson and
60 Cross, 2005). Many of these genes are involved in signal transduction, encoding growth factors, receptors and proteases in the FGF, Wnt, EGF, HGF/c-Met and
TGFb/BMP/Activin pathways, which have been shown to regulate various stages of placental morphogenesis (Rossant and Cross, 2001). There is also an ever- increasing list of genes that function in the regulation of cell cycle progression.
Previous studies suggested that p57kip2 and p27kip1, two principal cell regulators, have an important role in the control of proliferation and differentiation of extra- embryonic cell lineages (Zhang et al., 1998). Similarly, DP1 and cyclinE1/E2 have been shown to be essential for placental function, specifically for the endoreduplication in trophoblast giant cells (Geng et al., 2003; Kohn et al., 2003;
Parisi et al., 2003). Moreover, ablation of Rb leads to severe disruption of the normal labyrinth architecture in the placenta, which likely results from the defects in trophoblast stem cells (Wu et al., 2003; Wenzel et al., 2007). While the molecular mechanisms underlying these cellular defects and the specific cell lineages contributing to the placental dysfunction have not been well studied, these findings suggest the high sensitivity of placenta to subtle alterations in cell cycle control and the restrictive requirement of a fully developed placenta to maintain a successful pregnancy.
We showed that embryos lacking both E2f7 and E2f8 exhibit widespread apoptosis and die by E11.5 (Chapter 2 of this study and Li et al., 2008). While the concomitant ablation of E2f1 suppressed most of the ectopic apoptosis observed in E2f7-/-E2f8-/- embryos, it failed to prolong fetus life or rescue any other defects, which strongly suggests that the apoptosis is not the leading cause
61 of these phenotypes. We sough to determine the primary cause underlying the
E11.5 lethality of DKO embryos and reasoned that identification of the tissues
and cells in which E2F7 and E2F8 are most critical for embryo
development/viability may provide more valuable insight into their physiological
functions. Utilizing conditional strategies of gene ablation, here we demonstrate
that extra-embryonic functions of E2f7 and E2f8 are both necessary and
sufficient for embryo viability. Interestingly, while disruption of placental function
of E2f7 and E2f8 alone is enough to result in the gross abnormalities and E11.5
lethality, it fails to cause the full apoptotic phenotype observed in DKO (total
knockout) fetuses.
3.2 Results
3.2.1 Loss of E2f7 and E2f8 leads to profound placental defects
Given the critical role of placenta during embryo development, we
investigated the possible functions of E2F7 and E2F8 in the placenta to
determine whether placental dysfunction might underlie the mid-gestation (E11.5)
lethality of E2f7-/-E2f8-/- embryos. We first examined their expression in the extra-
embryonic compartment. Quantitive RT-PCR expression analysis demonstrated
that E2f7 and E2f8 mRNA levels were higher in placental than fetal tissues
(Figure 3.1A). Immunohistochemistry (IHC) assays also detected robust expression of both E2F7 and E2F8 proteins in the three main extra-embryonic lineages, labyrinth trophoblasts (LT), spongiotrophoblasts (SP) and trophoblast
62 giant cells (TG) (Figure 3.1B and data not shown). Interestingly, E2F7 and E2F8
proteins were highly expressed in some cells but were undetectable in others,
presumably reflecting their cell cycle-dependent expression pattern (de Bruin et
al., 2003; Di Stefano et al., 2003; Maiti et al., 2005). Histological examination of
hematoxylin and eosin (H&E)-stained E2f7+/+E2f8+/+ placenta sections showed
that the placental architecture was well organized by E10.5, with the vasculature
arranged as a network of maternal sinusoids juxtaposed to fetal-derived blood
vessels (Figure 3.2A, left panels). In contrast, placental architecture in all E2f7-/-
E2f8-/- embryos was severely compromised (Figure 3.2A, right panels). There
was an abnormal accumulation of large clusters of densely packed trophoblast
cells that failed to effectively invade into the maternal decidua. The vasculature
network was also poorly formed and maternal sinusoids were rarely found
adjacent to fetal blood vessels. A similar phenotype was also observed in E9.5
DKO placental tissues (Figure 3.2B). Quantitative RT-PCR analysis of gene
expression showed that double mutant placentas had normal expression of
trophoblast stem (TS) cell-specific markers (Eomes and Cdx2), but lower levels of SP- and TG-specific markers (Tpbp, Pdgf and Proliferin, Csf1r, Pl-1, Prp,
respectively; Figure 3.3A). In situ hybridization confirmed the lower expression of
Tpbp and Proliferin (Figure 3.3B and data not shown) and IHC techniques
showed lower PL-1 protein in double mutant placentas (Figure 3.3C). Therefore,
in addition to the defects previously characterized in double mutant fetuses
(Chapter 2 in this study and Li et al., 2008), these analyses show that double mutant placentas have extensive architectural and functional abnormalities.
63 3.2.2 Wild type placenta is sufficient to support mutant fetuses to birth
The profound placental defects in E2f7-/-E2f8-/- embryos encouraged us to
rigorously evaluate the role of E2f7 and E2f8 in extra-embryonic lineages. To this end, we used conditional E2f7loxp and E2f8loxp alleles which were generated
previously in our laboratory (Chapter 2 in this study and Li et al., 2008) and took
two complementary genetic approaches to specifically delete Ef7/E2f8 either in
the fetal or in the placental compartment.
In the first approach, Sox2-cre transgenic mice were used to examine the consequence of inactivating E2f7 and E2f8 solely in fetal lineages. Sox2-cre
mice express cre in the embryo proper as early as E6.5, but not in extra-
embryonic lineages (Hayashi et al., 2002). Intercrosses between Sox2- cre;E2f7+/-E2f8+/- and E2f7loxp/loxpE2f8loxp/loxpRosa26loxp/loxp mice showed that live
Sox2-cre;E2f7loxp/-E2f8loxp/-Rosa26loxp/+ (Sox2-cre) fetuses could remarkably be
recovered at every embryonic stage analyzed, including at birth, although most
newborn pups died within their first two days of life (Figure 3.4A). Sox2-cre
E10.5 fetuses appeared normal and lacked any of the developmental phenotypes
characteristic of similarly staged E2f7-/-E2f8-/- embryos (Figure 3.4C, bottom
panels). As expected, E10.5 placentas associated with these fetuses were
morphologically normal (Figure 3.4C, top panels). The Rosa26loxp allele was
used in these crosses to visually monitor cre expression in the appropriate
tissues (Soriano, 1999). Indeed, X-gal staining and PCR genotyping confirmed
the fetal-specific expression of cre and the complete ablation of E2f7 and E2f8 in
Sox2-cre E10.5 fetuses and all the examined organs from pups recovered at
64 birth (Figure 3.4B, 3.4D). From these observations, we conclude that extra-
embryonic functions of E2f7 and E2f8 are sufficient for embryonic development.
3.2.3 E2f7 and E2f8 are essential in the placenta
As the second approach, we utilized Cyp19-cre transgenic mice to
examine the consequence of inactivating E2f7 and E2f8 in the placenta. Cyp19-
cre mice express cre in all trophoblast cell lineages of the mouse placenta as
early as E6.5, but not in the embryo proper (Wenzel and Leone, 2007).
Intercrosses between Cyp19-cre;E2f7loxp/+E2f8loxp/+ and
E2f7loxp/loxpE2f8loxp/loxpRosa26loxp/loxp mice showed that 8 out of 22 Cyp19-
cre;E2f7loxp/loxpE2f8loxp/loxpRosa26loxp/+ (Cyp19-cre) embryos died at E10.5 and all were dead by E11.5 (Figure 3.5A). X-gal staining and PCR genotyping of
Cyp19-cre positive embryos confirmed the specific expression of cre and deletion
of E2f7/E2f8 in the placenta (Figure 3.5B and 3.5C, top panels). Gross and histological examinations of live Cyp19-cre E10.5 embryos showed a severe disruption of placental architecture that was associated with fetal growth retardation, blood vessel dilation and hemorrhaging (Figure 3.5C, middle and bottom panels).
We also used Tpbp-cre transgenic mice, which expressed cre specifically in spongiotrophoblasts to further investigate the contribution made by E2F7 and
E2F8 in spongiotrophoblast cell lineage (Figure 3.6A). By similar breeding strategies as described above, we obtained offspring from crosses between
Tpbp-cre;E2f7loxp/+E2f8loxp/+ and E2f7loxp/loxpE2f8loxp/loxpRosa26loxp/loxp .
65 Interestingly, the targeted inactivation of E2f7/E2f8 in spongiotrophoblasts had no
adverse effect on placental development (Figure 3.6B) and the associated
fetuses could be carried to term (Figure 3.6C).
From these genetic analyses, we conclude that placental function of E2f7
and E2f8 is both necessary and sufficient to carry fetuses to term. While E2f7
and E2f8 may still play a role in spongiotrophoblast cell lineage, their
spongiotrophoblast-specific function is clearly dispensable for placental
development and fetal survival.
3.2.4 Loss of E2f7 and E2f8 in the placenta dictates molecular events in the fetus
Since changes in gene expression could directly underlie many of the placental and fetal phenotypes in E2f7-/-E2f8-/- embryos, we began evaluating global gene expression by performing Affymetrix microarray analysis on fetal
tissues derived from E2f7+/+E2f8+/+ and E2f7-/-E2f8-/- E10.5 embryos. In order to
discriminate between direct and indirect placenta-induced changes, we also
included Sox2-cre;E2f7loxp/-E2f8loxp/- (Sox2-cre) and Cyp19-
cre;E2f7loxp/loxpE2f8loxp/loxp (Cyp19-cre) E10.5 fetuses in our assay.
Unsupervised clustering of the entire set of fetal expression data
segregated the four cohorts into two arms, with E2f7+/+E2f8+/+ and Sox2-cre
fetuses in one arm and E2f7-/-E2f8-/- and Cyp19-cre fetuses in the other (Figure
3.7A). This degree of separation is remarkable given that each arm actually
consisted of two cohorts of fetuses with opposite genotypes. In one arm,
66 E2f7+/+E2f8+/+ fetuses segregated apart from Sox2-cre fetuses, suggesting that
these two cohorts had sufficient molecular differences between them to allow
fetuses containing E2f7/E2f8 to be distinguished from those lacking E2f7/E2f8.
In the second arm, however, E2f7-/-E2f8-/- and Cyp19-cre fetuses failed to segregate apart, despite the fact that Cyp19-cre fetuses had E2F7/E2F8 proteins while E2f7-/-E2f8-/- fetuses were devoid of them. The common feature between
E2f7-/-E2f8-/- and Cyp19-cre fetuses is that they are both attached to placentas lacking E2f7 and E2f8. Therefore, we conclude that the predominant force guiding gene expression profiles in these four fetal cohorts is the genotype of their placentas.
We then focused on the analysis of differentially expressed genes in E2f7-
/-E2f8-/- fetuses. Two striking observations were made by this analysis. First, we
noticed that most of the genes dysregulated in E2f7-/-E2f8-/- fetuses (>2-fold and
p<0.05) were also dysregulated in Cyp19-cre fetuses, but not in Sox2-cre fetuses
(Figure 3.7B). This was surprising because both E2f7-/-E2f8-/- and Sox2-cre
fetuses lack E2F7/E2F8 proteins, whereas Cyp19-cre fetuses contain a normal complement of these two E2Fs. The second striking observation came from the ontological classification of upregulated genes in E2f7-/-E2f8-/- fetuses, which revealed an enormous over-representation of genes with nutrition-, hypoxia- and
DNA-damage induced stress-related functions (92 of 183 upregulated genes in this experiment, Li et al., 2008). Scatter plots and quantitive RT-PCR showed that these stress-related genes were upregulated in Cyp19-cre fetuses but not in
Sox2-cre fetuses (Figure 3.7C, 3.7D). These latter observations suggest that
67 faulty placentas in E2f7-/-E2f8-/- and Cyp19-cre embryos elicit a nutrition- and
hypoxia-related stress response in attached fetuses, providing a plausible
explanation for how loss of E2f7 and E2f8 in the placenta dictates the
transcriptional programs engaged in the fetus.
3.2.5 Fetal function and placental function of E2F7/E2F8 both contribute to
the full apoptotic phenotype observed in E2f7-/-E2f8-/- fetuses
We then sought to determine the contribution of a dysfunctional placenta
to the massive apoptosis observed in E2f7-/-E2f8-/- fetuses. To this end,
apoptosis was compared in live Sox2-cre;E2f7loxp/-E2f8loxp/- (Sox2-cre) and
Cyp19-cre;E2f7loxp/loxpE2f8loxp/loxp (Cyp19-cre) E10.5 fetuses to that in
E2f7+/+E2f8+/+ and E2f7-/-E2f8-/- fetuses. As previously described, there was massive ectopic apoptosis in areas of the head, branchial arch and somites of
E2f7-/-E2f8-/- fetuses (Chapter 2 in this study and Li et al., 2008). In contrast, we
found no ectopic apoptosis in head regions and only intermediate levels of
apoptosis in the branchial arch and somites of Sox2-cre and Cyp19-cre fetuses
(Figure 3.8), even though Cyp19-cre embryos were destined to die soon after.
These findings genetically uncouple the execution of apoptosis in the fetus from
placental-mediated lethality, and demonstrate that ablation of E2f7/E2f8 function
in the fetus and in the placenta both contributes to the full level of apoptosis seen
in E2f7-/-E2f8-/- fetuses.
68 3.3 Discussion
The E2f7 and E2f8 family members are essential for embryo survival
(Chapter 2 in this study and Li et al., 2008). However, the primary cause of
lethality in E2f7-/-E2f8-/- embryos and the critical tissue where E2F7 and E2F8
function was unknown. In this study, by utilizing a complementary conditional
knockout strategy, we show that these two E2F family members play an essential
role in extra-embryonic lineages of the placenta. Their importance in these
lineages was highlighted by conditional knockout experiments showing that their
function in the placenta was both necessary and sufficient for fetal survival.
E2f7-/-E2f8-/- fetuses suffer from pleiotropic defects including widespread apoptosis, hemorrhaging and placental dysfunction, culminating in death by embryonic day E11.5 (Chapter 2 in this study and Li et al., 2008). In the effort to dissect this complex phenotype, we conditionally knocked out E2f7 and E2f8 in
embryonic cell lineages of the fetus, all extra-embryonic cell lineages of the placenta or specific spongiotrophoblast cell lineage of the placenta, and
examined resulting fetal phenotypes at the physiological, cellular and molecular
levels. We show that the mid-gestation lethality, hemorrhaging and acute stress
responses observed in E2f7-/-E2f8-/- embryos are primarily due to loss of E2f7
and E2f8 in all extra-embryonic cell lineages of the placenta with little contribution
from the lineage of trophoblast spongiotrophoblast cells.
Interestingly, while a mutant faulty placenta is sufficient to result in the
embryonic lethality and the gross abnormalities, it is not enough to cause the full
apoptotic phenotype seen in E2f7-/-E2f8-/- embryos. Based on this finding and
69 the previous observation that loss of E2f1 rescues E2f7/E2f8-loss induced
apoptosis, we postulate that the massive apoptosis in E2f7-/-E2f8-/- fetal tissues is
an additive effect of two related and necessary events, the derepression of E2f1 expression in fetal cells (Chapter 2 in this study and Li et al., 2008) and the induction of stress by a faulty placenta.
Since the inception of gene ablation techniques in mice, an ever- increasing number of knockout mouse models have been found to be embryonic lethal and associated with a variety of phenotypes. However, the underlying cause(s) of fetal phenotypes and critical site(s) of gene function have not yet been carefully analyzed. As we described here, an elegant genetic strategy together with appropriate systematic analyses would be very useful for investigating interactions between different cell lineages in a complex developmental system or in other complicated diseases.
3.4 Materials and Methods
3.4.1 Mouse strains and genotyping
The conventional or conditional E2f7 and E2f8 knockout mice, Sox2-cre,
Cyp19-cre transgenic mice, and Rosa26 reporter mice were maintained on a
mixed background of 129SvEv, C57B/L6 and FVB. Allele-specific primers used
for E2f7 and E2f8 PCR genotyping were described in Table 2.2.
70 3.4.2 Quantitive RT-PCR
Total RNA was extracted using Qiagen RNA miniprep columns according
to the manufacturer’s protocol, which included the optional DNase treatment
before elution from the column. Reverse transcription of total RNA was
performed using Superscript III reverse transcriptase (Invitrogen) and RNAse
Inhibitor (Roche) as described by the manufacturer. Quantitive PCR was
performed using SYBR Green reaction mix (BioRad) and the BioRad iCycler
PCR machine. All reactions were performed in triplicate and relative amounts of
cDNA were normalized to Gapdh.
3.4.3 In situ hybridization
In situ hybridization was performed on placental tissues using a previously
reported protocol (Christensen et al., 2002) that was modified for paraffin-
embedded sections. These modifications were deparaffinization in xylene and
Proteinase K (1 µg/ml) digestion. Plasmids for Proliferin and Tpbp (gifts from Dr.
J. Rossant) were linearized with HindIII and XbaI, respectively, to generate templates for riboprobe synthesis. Radio-labeled probes were generated by in vitro transcription with either T7 (Proliferin) or T3 (Tpbp) RNA polymerase
(Roche) using both 35S-CTP and 35S-UTP. The probes were hydrolyzed at 60°C
for 20 min in bicarbonate buffer (80 mM NaHCO3, 120 mM Na2CO3).
Hybridizations were performed with 1x107 dpm/slide for each probe. Following
application of the autoradiography emulsion NTB (Kodak), the slides were
71 exposed for 1 day (Proliferin) or 3 days (Tpbp) before the emulsion was developed.
3.4.4 Histological analysis and immunostaining
Placenta samples were fixed in 10% neutral buffered formalin and embedded in paraffin. 5 μm-thick sections were prepared. Standard hematoxylin and eosin (H&E) staining was used for general histopathological analysis. For immunostaining, slides were probed with primary antibodies specifically against E2F7 (ab56022, Abcam), E2F8 (homemade polyclonal antibody raised against a peptide representing amino acids 576-595 of murine
E2F8). Vectastain Elite ABC reagent (Vector labs) and DAB peroxidase substrate kit (Vector labs) were used in combination to immunohistochemically detect E2F7 and E2F8 signals by following the manufacturer’s instructions.
Samples were counterstained with hematoxylin. Cells belonging to giant cell lineage were detected by immunofluorescence (IF) staining using PL-1 antibodies (a gift from Dr. F. Talamantes). Nuclear DNA was counterstained with
DAPI.
3.4.5 X-gal staining
Whole-mount samples or 10 µm frozen sections were fixed in 2% formaldehyde and 0.2% glutaraldehyde in PBS for 1.5 h, and then washed 3 times for 5 min each in PBS at room temperature. Staining was carried out in 1 mg/ml X-gal, 5 mM potassium ferrocyanide, and 5 mM potassium ferricyanide in
72 lacZ wash buffer (0.01% deoxycholate, 0.02% NP-40 and 2 mM MgCl2) at 37°C
overnight. Nuclear Fast Red counterstaining was used to visualize lacZ-negative cells from the frozen sections.
3.4.6 TUNEL assay
Apoptotic cells were detected using TUNEL (S7101, Chemican) assays, performed according to the manufacturer’s protocol. TUNEL-stained slides were counterstained with hematoxylin. Quantification of TUNEL-positive cells was achieved by calculating the percentage of positive cells in different tissues.
Counting data are reported as the average ± SD fold induction of percentage of
positive cells. Three sections per sample and at least three different samples for
each genotype were analyzed.
3.4.7 Affymetrix microarray analysis
Total RNA was isolated using Qiagen RNA miniprep columns according to the manufacturer’s protocol. Global gene expression analyses were performed on Affymetrix Mouse Genome 430 2.0 arrays at the Ohio State University
Comprehensive Cancer Center (www.osuccc.osu.edu/microarray/). Expression
values were adjusted by quantile normalization and log2 transformation with
RMAExpress and data was analyzed with BRB-ArrayTools 3.7.0 (http:
linus.nci.nih.gov BRB-ArrayTools.html). Class comparison was used to select
genes differentially expressed at a significance level of p<0.05. Probes with a
>2-fold misexpression in E2f7-/-E2f8-/- when compared to E2f7+/+E2f8+/+ were
73 used and the average relative expression level of each genetic group was loaded into the TIGR MultiExperiment Viewer (MEV, version 4.0) (TIGR, Rockville, MD) to generate heatmaps. Clustering and scatter plot analyses were performed by using Clustering and Scatterplot functions on BRB Array Tools.
74 Figure 3.1 E2F7 and E2F8 are highly expressed in the placenta. (A)
Quantitive RT-PCR analysis of E2f7 or E2f8 mRNA expression in E10.5 wild type placentas (gray bars) and fetuses (white bars). (B)
Immunohistochemistry using antibodies specific against E2F7 protein (top panels) and E2F8 protein (bottom panels) on E10.5 placenta sections with the indicated genotypes. Positive trophoblast giant cells are indicated by arrows.
Negative trophoblast giant cells are indicated by arrows heads. Scale bars,
100 µm.
75 A E2f7 3-.0
.5
2-.0
.5
1-.0
.5
0-.0 E2f8 9-0 0 0 6-0 0
Relative gene expression 0 3-0 0 0 0-0 Placentas Fetuses
B E2f7+/+E2f8+/+ E2f7-/-E2f8-/- E2F7
E2f7+/+E2f8+/+ E2f7+/+E2f8-/- E2F8
Figure 3.1
76 Figure 3.2 E2f7-/-E2f8-/- placentas exhibit profound structural abnormalities. (A) E10.5 placenta sections with the indicated genotypes were stained with Hematoxylin and Eosin (H&E). The bottom panels are high magnification images of representative areas demarcated by the boxes in the top panels. The yellow dotted-line marks the interface between decidual and placental tissues. De., Decidua; La., Labyrinth. Scale bars, 100 µm. (B) E9.5 placenta sections with the indicated genotypes were stained with H&E. The bottom panels are high magnification images of representative areas demarcated by the boxes in the top panels.
77 A E10.5 placentas E2f7+/+E2f8+/+ E2f7-/-E2f8-/-
La. La.
De. De. H&E
B E9.5 placenta E2f7+/+E2f8+/+ E2f7-/-E2f8-/- H&E
Figure 3.2
78 Figure 3.3 Severe differentiation defects in E2f7-/-E2f8-/- placentas. (A)
Quantitive RT-PCR analysis of trophoblast cell lineage markers in E10.5
E2f7+/+E2f8+/+ (7+/+8+/+, black) and E2f7-/-E2f8-/- (7-/-8-/-, red) placentas. TS, trophoblast stem cells; SP, spongiotrophoblast cells; TG, trophoblast giant cells. All data represent the average of samples analyzed in triplicate; error bars indicate average ± SD from triplicates. (B) RNA in situ hybridization analysis of Proliferin, a giant cell-specific marker, on E9.5 placenta sections having the indicated genotypes. (C) Immunofluorescence staining of E10.5 placenta sections with the indicated genotypes, using antibodies specific for
Placental Lactogen 1 (PL-1), a giant cell-specific marker. The boxed areas in top panels are shown at higher magnification in bottom panels. Scale bars,
100 µm.
79
A TS SP TG TG 8- 8- 6- 6-6 Eomes Tpbp Proliferin Pl-1 5 6- 6- 4- 4-4 4- 4- 3 2- 2-2 2- 2- 1 0- 0- 0- 0-0 8- 6- 8- 6- Cdx2 Pdgf Csf1r Prp 6- 6- 4- 4- 4- 4- 2- 2- Relative gene expression 2- 2-
0- 0- 0- 0- 7+/+8+/+ 7-/-8-/- 7+/+8+/+ 7-/-8-/- 7+/+8+/+ 7-/-8-/- 7+/+8+/+ 7-/-8-/-
B E9.5 placentas E2f7+/-E2f8+/- E2f7-/-E2f8-/- Proliferin
C E10.5 placentas E2f7+/-E2f8+/- E2f7-/-E2f8-/- PL-1
Figure 3.3
80 Figure 3.4 Wild type placenta is sufficient to rescue E2f7/E2f8-mutant fetuses to birth. (A) Genotypic analysis of embryos derived from Sox2-cre crosses at the indicated stages of development. (B) Representative gel pictures of E2f7 (top) and E2f8 (bottom) PCR genotyping results using genomic DNA isolated from different organs of P0 pups. Bla., bladder; Int., intestine; Sto., stomach; Hea., heart; Liv., liver; Kid., kidney; Lun., lung; Bra., brain; Mus., muscle; Pla., placenta. ko, conventional knockout allele; loxP, conditional knockout allele. (C) Top panels: gross view of E2f7loxp/-E2f8loxp/-
(control) and Sox2-cre;E2f7loxp/-E2f8loxp/- (Sox2-cre) embryos at E10.5. Bottom panels: H&E staining of control and Sox2-cre E10.5 placental tissues. (D) Top left panel: whole-mount X-gal staining of Sox2-cre E10.5 fetus. Top right and bottom panels: X-gal staining of representative tissue sections from Sox2-cre newborn (P0) pups.
81 A Genotypic analysis of embryos during development: Sox2-cre crosses E2f7loxp/+ E2f7loxp/- Sox2-cre;E2f7loxp/+ Sox2-cre;E2f7loxp/- E2f8loxp/+ E2f8loxp/- E2f8loxp/+ E2f8loxp/- E2f8loxp/+E2f8loxp/- E2f8loxp/+ E2f8loxp/- total E10.5 -- 4 -- 4 -- 6 -- 4 28 expected -- 4 -- 4 -- 4 -- 4
E13.5 -- 4 -- 2 -- 7 -- 8 21 expected -- 5 -- 5 -- 5 -- 5
P0-P1 5 19 4 18 3 19 6 11(8)a 93 expected 4 20 4 20 4 20 4 19 () number of dead embryos; Exact binomial test: a highly significant (p< 0.001).
B E2f7 genotype control Sox2-cre Tail Tail Bla. Int. Sto. Hea. Liv. Kid. Lun. Bra. Mus. Pla. ko loxp
E2f8 genotype control Sox2-cre Tail Tail Bla. Int. Sto. Hea. Liv. Kid. Lun. Bra. Mus. Pla. ko loxp
C D control Sox2-cre fetus liver ) (fetus) Gross view Gross
sox2-cre lung intestine X-gal ( H&E (placenta)
Figure 3.4
82 Figure 3.5 Extra-embryonic function of E2f7 and E2f8 is necessary for fetal survival. (A) Genotypic analysis of embryos derived from Cyp19-cre crosses at the indicated stages of development. (B) PCR genotyping analyses of E2f7 (top) and E2f8 (bottom) using genomic DNA isolated from E10.5 fetuses (Fet.) and placentas (Pla.). Note that the loxp allele was largely diminished and the ko allele could be readily observed in the Cyp19-cre placenta. Cyp19-cre, Cyp19-cre;E2f7loxp/loxpE2f8loxp/loxpRosa26loxp/loxp; Control,
E2f7loxp/loxpE2f8loxp/loxpRosa26loxp/loxp. Because the placenta is composed of extra-embryonic trophoblast cells as well as embryonic endothelial cells, the intact loxp alleles are likely from fetal endothelial cell contamination. (C) Top panels: E10.5 control and Cyp19-cre fetus and placenta sections were stained with X-gal and counterstained with nuclear fast red. Middle panels: gross appearance of fetuses and their associated placentas. The boxed areas are demonstrated in the high magnification images on the right. Note that the
Cyp19-cre fetus is smaller in size. Arrows point to the hemorrhaging in the
Cyp19-cre fetus. Bottom panels: H&E staining of E10.5 placentas with indicated genotypes. The boxed areas are shown in the high magnification images on the right.
83
ABE2f7 genotype Cyp19-cre control Genotypic analysis of embryos: Cyp19-cre crosses Fet. Pla. Fet. Pla. 7loxp/+ 7loxp/loxp cre;7loxp/+ cre;7loxp/loxp ko total loxp 8loxp/+ 8loxp/loxp 8loxp/+ 8loxp/loxp 8loxp/+ 8loxp/loxp 8loxp/+ 8loxp/loxp E10.5 4 2 7(1) 21 3 9 10 14(8) 79 E2f8 genotype expected 5 7 10 18 5 7 10 18 Cyp19-cre control E11.5 -- -- 4 10 -- -- 2 0(11)a 27 Fet. Pla. Fet. Pla. expected -- -- 3 11 -- -- 3 11 ko () number of dead embryos; a highly significant (p< 0.0001). loxp
C control Cyp19-cre X-gal whole mount H&E
Figure 3.5
84 Figure 3.6 Specific deletion of E2f7 and E2f8 in spongiotrophoblasts does not affect placental development and fetal survival. (A) X-gal staining of E2f7loxp/loxpE2f8loxp/loxp (control) and Tpbp-cre;E2f7loxp/loxpE2f8loxp/loxp
(Tpbp-cre) placentas at E13.5. (B) H&E staining of control and Tpbp-cre placentas at E10.5. High magnification images are shown on the right. (C)
Genotypic analysis of embryos derived from Tpbp-cre crosses at the indicated stages of development.
85 A control Tpbp-cre
B control
Tpbp-cre
C
Genotypic analysis of embryos: Tpbp-cre crosses
loxp/+ loxp/loxp loxp/+ loxp/loxp 7 7 cre;7 cre;7 total 8loxp/+ 8loxp/loxp 8loxp/+ 8loxp/loxp 8loxp/+ 8loxp/loxp 8loxp/+ 8loxp/loxp E12.5 7 4 5 7 5 3 3 2 36 expected 5 5 5 5 5 5 5 5 P0 8 18 9 9 10 11 5 15 85 expected 11 11 11 11 11 11 11 11
Figure 3.6
86
Figure 3.7 Loss of E2f7 and E2f8 in the placenta dictates molecular events in the fetus. (A) Dendogram for unsupervised clustering analysis using centered correlation and average linkage. E10.5 fetuses with the indicated genotypes were presented individually. Note the segregation of these fetuses based on the genotypes of their associated placentas. (B)
Heatmap representation of microarray probe sets that showed a >2-fold misexpression in E2f7-/-E2f8-/- relative to E2f7+/+E2f8+/+ fetuses (p<0.05). Data from four different genetic groups are presented as an average expression level of E2f7+/+E2f8+/+ (7+/+8+/+, n=2), E2f7-/-E2f8-/- (7-/-8-/-, n=4), Sox2- cre;E2f7loxp/-E2f8loxp/- (Sox2, n=4) and Cyp19-cre;E2f7loxp/loxp;E2f8loxp/loxp
(Cyp19, n=3). (C) Scatter plot analyses of stress-related gene expression between indicated genetic groups. A 2-fold cutoff is shown. (D) The indicated stress-related gene expression was evaluated by quantitive RT-PCR assays.
Data are presented as the average ± SD-fold induction in samples analyzed in triplicate.
87 Fetuses A Fetuses B -0.2- 7+/+8+/+ 7-/-8-/- Sox2 Cyp19
0.0-
0.2-
0.4-
Correlation 0.6-
0.8-
1.0- -/- -/- -/- -/- +/+ +/+ 8 8 8 8 8 8 -/- -/- -/- -/- 7 7 7 7 +/+ +/+ Sox2 Sox2 Sox2 Sox2 7 7 Cyp19 Cyp19 Cyp19
fetuses with fetuses with 7/8-positive placentas 7/8-negative placentas n=2 n=4 n=4 n=3 -2.0 0.0 2.0 p < 0.05
C Stress-related gene expression in the fetus 13 13 13 -/-
-/- 11 11 11 9
9 E2f8 9 E2f8 -/- -/- 7 7 7
5 Cyp19-cre 5 5 E2f7 E2f7 3 3 3 35791113 3 5 7 9 11 13 3 5 7 9 11 13 E2f7+/+ E2f8+/+ E2f7+/+ E2f8+/+ Cyp19-cre D Bnip3 Ndrg1 Pfkp Eco1l 8-6.0 8-8.0 7.06- 3.03-
5.0 7.0 6.0 6- 6-6.0 5.0 4- 2.02- 4.0 5.0 4.0 4-3.0 4-4.0 3.0 3.0 2.0 2- 1.01- 2- 2-2.0 2.0 1.0 1.0 1.0 0-0.0 0-0.0 0.00- 0.00- Relative gene expression 7+/+8+/+7-/-8-/- Sox2 Cyp19 7+/+8+/+ 7-/-8-/- Sox2 Cyp19 7+/+8+/+7-/-8-/- Sox2 Cyp19 7+/+8+/+ 7-/-8-/- Sox2 Cyp19
Figure 3.7
88
Branchial arch Head Somites 8- 4-4 5-5 ** ** 4-4 6- 3-3 ** 3-3 4- ** ** 2-2 2-2 ** 2- 1- 1 1-1 Relative TUNEL +ve TUNEL Relative 0- 0-0 0-0 n=3 n=4 n=4 n=3 n=4 n=4 n=3 n=4 n=4 control 7-/-8-/- Sox2 Cyp19
Figure 3.8 Cell autonomous and non-cell autonomous functions of E2F7 and E2F8 in the fetus. Quantification of TdT-mediated dUTP nick end- labeling (TUNEL)-positive cells in the indicated tissue areas of E10.5 fetuses.
Results are shown as a fold induction of percentage of TUNEL positive cells in each indicated genetic group relative to percentage of TUNEL positive cells in their individual control which was set to 1. Pairwise comparisons were evaluated by two-tailed Student’s t-tests (** p<0.02). n, number of embryos measured for each genetic group; error bars represent standard deviations.
For comparison purposes, previous data derived from E2f7-/-E2f8-/- and their control group were included (Chapter in this study and Li et al., 2008).
89
CHAPTER 4
E2F7 AND E2F8 PLAY A CRITICAL ROLE IN CELL CYCLE REGULATION
4.1 Introduction
The majority of cells respond to external growth stimuli by activating signaling cascades that carry them through the cell cycle and culminate in the generation of two identical daughter cells. A critical step in these stimulatory signaling cascades involves the downregulation of cyclin-dependent kinase
(CDK) inhibitors, the induction of G1-cyclins, and the subsequent association of these cyclins with the catalytic CDK subunits (Murray, 2004; Sánchez and
Dynlacht, 2005). These events result in the activation of CDKs and phosphorylation of retinoblastoma (Rb) and Rb-related pocket proteins. The consequence of this phosphorylation is the dissociation of pocket protein-E2F repressor complexes, accumulation of E2F activity and activation of E2F- responsive genes (Beijersbergen and Bernards, 1996; Frolov and Dyson, 2004).
The execution of E2F-dependent transcription late in G1 is believed to represent the final step in the CDK-mediated mitogenic signaling that commits cells to enter
S phase. This E2F-dependent transcriptional program is attenuated later in S-G2 by a second wave of E2F-mediated repression that allows cells to complete S
90 phase, transit through G2 and enter mitosis. Hence, sequential waves of E2F-
mediated G0-repression, G1-S-activation and G2-repression coordinate the
completion of phase-specific events and the successful passage of cells through
the cell cycle.
E2F consists of a family of related proteins (Attwooll et al., 2004;
DeGregori and Johnson, 2006). Based on sequence conservation and structure-
function studies, the E2F family has been conveniently divided into transcription
activators and repressors. E2F activators, consisting of E2F1, E2F2 and E2F3,
have strong transcription activation domains that, in association with co-activator
proteins, robustly induce RNA polymerase II-dependent gene expression. The
combined ablation of the three activators in mouse embryo fibroblasts (MEFs)
results in a profound inhibition of proliferation that is accompanied by a decrease
in E2F-target expression (Wu et al., 2001). E2F repressors include E2F4, E2F5,
E2F6, E2F7 and E2F8. E2F4 and E2F5 serve to recruit pocket proteins, histone
deacetylases (HDACs), and associated co-repressors to E2F-target promoters
and mediate their repression during G0 (Trimarchi and Lees, 2002). MEFs
lacking E2f4 and E2f5 fail to repress G0-specific genes and lose responsiveness
to growth inhibitory signals (Gaubatz, et al., 2000). E2F6, E2F7 and E2F8 lack
the classical Rb-binding domain and are thus thought to repress gene expression in an Rb and pocket protein-independent manner. MEFs deficient for E2f6 or doubly deficient for E2f7 and E2f8 fail to adequately repress E2F-dependent gene expression during S-G2 and, as a result, are more sensitive to apoptotic stimuli (Giangrande et al., 2004; Chapter 2 in this study and Li et al., 2008;
91 Zalmas et al., 2008). While E2F6 is part of a multi-subunit repressor complex that includes polycomb group proteins as well as Mga and Max (Ogawa et al.,
2002), the mechanism of E2F7 and E2F8-mediated repression is less well understood but appears to be distinct from other E2Fs. Unlike other family members, E2F7 and E2F8 each have two tandem DNA-binding domains and bind DNA independent of dimerizing with DP1/DP2 proteins (de Bruin et al.,
2003; Di Stefano et al., 2003; Logan et al., 2004; Christensen et al., 2005; Logan et al., 2005; Maiti et al., 2005). Thus, functional and structural properties separate the E2F family into activators and two distinct subgroups of repressors, with E2F1/E2F2/E2F3 activators functioning in G1-S, E2F4/E2F5 repressors in
G0-G1 and E2F6/E2F7/E2F8 repressors in S-G2. Together, the E2Fs are believed to control cell cycle dependent oscillations of gene expression.
Studies in plants, worms and flies have highlighted the existence of important deviations from the archetypal cell cycle described above. The endocycle is an exquisite example of a specialized cell cycle utilized by a number of important cell types in protists, plants, and animals. Endocycles generate >2N genomes by undergoing successive rounds of DNA replication in the absence of an intervening karyokinesis or cytokinesis. The resulting genomes adopt different chromatin configurations that range from polyploid to polytene arrangements (Edgar and Orr-Weaver, 2001). At one extreme, cells in the salivary gland of flies and in the placenta of mammals use endocycles to attain polytenic genomes >1000N (Varmuza et al., 1988; Lilly and Duronio, 2005). The molecular mechanisms involved in achieving this incredible feat are largely
92 unknown, but likely require unique adaptations of mechanisms used by mitotic
cells to control the firing of DNA replication, assembly of chromatin, segregation of chromosomes, karyokinesis and cytokinesis. Components of the CDK,
cyclins, and E2F families are believed to contribute to the control of mitotic and
endocycles in flies (Edgar and Orr-Weaver, 2001); however, recent studies using
knockout mice have challenged the universal requirement for these activities in
governing cell cycle progression in most mammalian cell types (Geng et al., 2003;
Parisi et al., 2003; Kozar et al., 2004; Malumbres et al., 2004). Clearly, cell
cycles are not as uniform and rigid as previously portrayed by in vitro studies, and a great deal of work is still needed to understand how cell cycles are controlled in vivo.
The death of E2f7 and E2f8 double knockout embryos in utero is in stark
contrast to the much less severe adult and neonatal phenotypes seen with the
loss of other E2f family members. The genetic studies performed in Chapter 3 clearly defined the loss of extra-embryonic function of E2f7 and E2f8 as the leading event of the embryonic lethality in E2f7-/-E2f8-/- embryos. In this chapter,
we focus mainly on the placenta, and utilize molecular biology and cellular
approaches to further investigate the cellular processes and molecular pathways
regulated by E2F7 and E2F8 in the extra-embryonic compartment. We
demonstrate that E2F7/E2F8-mediated repression plays a critical role in the
control of cell cycle progression in the placenta, as well as in other tissues.
93 4.2 Results
4.2.1 Identification of direct targets of E2F7 and E2F8
Having known that the placenta is the critical tissue for E2F7 and E2F8
function, we decided to focus our attention on the placenta to further investigate cellular processes and molecular pathways regulated by these two novel E2Fs.
In effort to identify direct targets of E2F7 and E2F8, we performed global gene expression analyses on E2f7+/+E2f8+/+, E2f7-/-E2f8-/-, Sox2-cre and Cyp19-cre
E10.5 placental tissues. Unsupervised clustering analysis separated the four
placental cohorts based on their genotypes, with E2f7+/+E2f8+/+ and Sox2-cre
cohorts clustering together in one arm, and E2f7-/-E2f8-/- and Cyp19-cre cohorts clustering in the other (Figure 4.1A, 4.1B).
E2F7 and E2F8 are believed to act as transcriptional repressors (de Bruin et al., 2003; Di Stefano et al., 2003; Logan et al., 2004; Christensen et al., 2005;
Logan et al., 2005; Maiti et al., 2005). Therefore, in order to identify direct targets repressed by E2F7 and E2F8, we focused on genes that were upregulated in tissues lacking E2f7 and E2f8. Venn diagrams illustrated 49 genes that were commonly upregulated in E2f7-/-E2f8-/- and Cyp19-cre placentas (>2-fold and
p<0.05) (Figure 4.1C). Interestingly, annotation of this 49 gene-set revealed an
enrichment of genes with functions related to metabolism (22%) and cell cycle
regulation (39%) (Figure 4.1D). Quantitive RT-PCR assays confirmed these
results in a subset of selected targets (Figure 4.1E).
94 We then assessed whether these 49 genes were also upregulated in
fetuses lacking E2f7 and E2f8. Two sources of E2f7/E2f8-deficient fetuses were
considered, E2f7-/-E2f8-/- and Sox2-cre fetuses. However, to avoid the potential
effects of placental-induced stress on E2f7-/-E2f8-/- fetal gene expression, only
upregulated genes identified in Sox2-cre fetuses (>1.6-fold and p<0.05) were
used for this analysis. The Venn diagrams shown in Figure 4.2A revealed that
29 genes of the 49-gene-set were placenta-specific, whereas the other 20 genes
were upregulated in both placentas and fetuses lacking E2f7 and E2f8. Promoter
analysis by TFsearch showed that 70% of the 20 upregulated genes (14/20) had
conserved E2F binding sites in their promoters (Figure 4.2A, Table 4.1),
suggesting an enrichment of E2F-target genes in the 20 placental/fetal gene-set.
In contrast, only ~6.9% (2/29) of the placenta-specific promoters had conserved
E2F binding sites (Figure 4.2A, Table 4.2), indicating that these targets may be regulated indirectly through other DNA binding factors or directly through atypical
E2F binding sites. Consistent with the possible regulation of these 20 genes by
E2F7/E2F8 in multiple cell lineages, these targets were both upregulated in E2f7-
/-E2f8-/- placentas as well as E2f7-/-E2f8-/- MEFs (Figure 4.2B, 4.2C).
Next, we performed chromatin immunoprecipitation (ChIP) assays to
determine whether the 20 gene-set represented direct targets of E2F7 and E2F8.
Because reliable mouse E2F7/E2F8-specific ChIP-grade antibodies are not
available, we overexpressed flag-tagged E2F7/E2F8 in HEK 293 and performed
ChIP assays with flag-specific antibodies. These experiments showed that anti-
flag antibodies, but not control IgG, efficiently co-immunoprecipitated all seven
95 promoter sequences tested (Figure 4.3A). Importantly, anti-flag antibodies failed
to co-immunoprecipitate irrelevant sequences in E2f1 (exon 1) or the tubulin
promoter. Moreover, parallel ChIP assays using mutant versions of E2F7 or
E2F8 lacking DNA-binding capacity failed to co-immunoprecipitate target
promoter sequences. From these results, we conclude that the 20 target genes
identified by expression profiling of placental and fetal tissues represent bona
fide targets directly repressed by E2F7 and E2F8. Interestingly, 75% of these
direct target genes (15/20) encoded proteins with functions in cell cycle
regulation, particularly in G1-S control; whereas the other 29 placenta-specific
targets encoded gene products involved in metabolism and other functions
(Figure 4.3B, 4.3C respectively).
4.2.2 Ablation of E2f7 and E2f8 results in ectopic S-phase entry and
aberrant mitosis
The above findings prompted us to examine E2f7-/-E2f8-/- placentas more
closely for defects in cell cycle regulation. We first measured BrdU incorporation in the three main extra-embryonic lineages: TG, SP and LT and found ectopic S phase entry in E2f7-/-E2f8-/- TG and SP (Figure 4.4A, 4.4B). Interestingly, BrdU
analysis of other tissues, such as lung tissue from P0 Sox2-cre pups and liver
tissue from Mx1-cre adult mice also revealed a significant induction of DNA
replication in the absence of E2f7 and E2f8 (Figure 4.4C-4.4F). Together, these
observations suggest that E2F7 and E2F8 play a broad role in the control of S-
phase entry.
96 In addition to the ectopic DNA replication, we also detected many E2f7-/-
E2f8-/- giant cells in mitosis, as measured by IHC with phospho(S10)-H3 antibodies (Figure 4.5A, 4.5B). Trophoblast giant cells are remarkable among all mammalian cell types, because they undergo endocycle, a derivative of the mitotic cycle where cells undergo successive rounds of DNA replication without an intervening mitosis (Varmuza et al., 1988). Therefore, it was very striking to observe E2f7-/-E2f8-/- giant cells at different stages of mitosis, including
metaphase, anaphase and telophase (Figure 4.5C). Confocal microscopy and
3D reconstruction of serial placental sections demonstrated a significant number
of E2f7-/-E2f8-/- giant cells that were undergoing or had completed karyokinesis
(Figure 4.6A, 4.6B). The fact that these giant cells had either one or two nuclei
suggested that ectopic karyokinesis had taken place. Consistent with this
interpretation, transmission electron microscopy confirmed that many double
mutant giant cells contained two closely apposed but distinct nuclei (Figure
4.6C). From these experiments, we conclude that E2f7-/-E2f8-/- giant cells ectopically segregated their chromosomes, and remarkably, completed karyokinesis in the absence of cytokinesis.
We also evaluated the functions of E2F7 and E2F8 in the hepatocyte, since it has been suggested to undergo endocycle as well (Edger and Orr-
Weaver, 2001). Consistent with the idea that E2F7 and E2F8 may play a critical role in the regulation of mitosis in general or at least in endocycling cells, P-H3 immunochemistry revealed a significant high level of mitotic index in double
mutant hepatocytes (Figure 4.9A, 4.9B).
97 4.2.3 Molecular mechanisms underlying the observed cell cycle defects
Closer examination of the cell cycle genes directly repressed by E2F7 and
E2F8 showed that most of them (11 out of 15) have functions known to promote
G1-S transition, such as E2f1, Mcm4, Mcm5, Mcm6, and Cdc6 (Table 4.1). Their increased expression in E2f7/E2f8-deficient placentas provided a molecular explanation of the ectopic DNA replication observed in E2f7-/-E2f8-/- SP and TG
cells.
However, no G2-M specific gene could be readily discovered from
microarray analysis, which is likely due to the fact that mitosis is mainly regulated
at the protein level through proteolysis. In order to better understand the
underlying mechanism of the aberrant mitosis, we examined Cyclin A2 (CycA2)
and Cyclin B1 (CycB1) in the placenta. CycA2 and CycB1 are mitotic cyclins,
which have been suggested to play an essential role in the progression of cells
into mitosis (Yam et al., 2002). A western blot, shown in Figure 4.7A and 4.8A, revealed a reproducible increase in CycA2 and CycB1 protein levels in E2f7-/-
E2f8-/- placentas. Interestingly, IHC analysis demonstrated that CycA2 protein
levels, on a per-cell-basis, were similar in control and double mutant giant cells;
however, there were many more CycA2-positive giant cells in E2f7-/-E2f8-/- placentas (Figure 4.7B, 4.7C). The fact that double mutant placentas had a significant increase in CycA2-positive/BrdU-negative giant cells, but had a similar number of CycA2/BrdU double-positive giant cells as control placentas, suggests that the expression of CycA2 was elevated during interphase of endocycle.
Deregulation of CycB1 expression was more remarkable. As expected, no
98 CycB1 protein was detected in the wild type giant cells; however, in sharp
contrast, CycB1-positive giant cells could be easily observed in E2f7-/-E2f8-/- counterparts (Figure 4.8B, 4.8C). Intriguingly, consistent with the observed high mitotic index, significant induction of CycA2- and CycB1-positive cells was also found in adult hepatocytes lacking E2f7 and E2f8 (Figure 4.9C-4.9F).
Taken together, these results provide two distinct molecular explanations of the observed cell cycle defects in G1-S and G2-M transitions. We conclude that E2F7 and E2F8 play a dual role in the control of cell cycle progression in the placenta as well as in other tissues.
4.2.4 Loss of E2f3a extends the life span of E2f7-/-E2f8-/- embryos by
rescuing the ectopic DNA replication but not the aberrant mitosis in the
placenta
Previous work has demonstrated that uncontrolled E2F activity, by losing
the Rb tumor suppressor, causes unscheduled DNA replication and is
detrimental to placental development (Wu et al., 2003). Interestingly, loss of the activator E2f3, particularly its isoform E2f3a, suppresses the over-proliferation observed in Rb-/- placentas and extends fetal viability (Ziebold et al., 2001;
Saavedra et al., 2002; Chong et al., 2009). Therefore, E2F3a was believed to be
a major activator that contributes to the ectopic proliferation and severe
disruption of placenta structure in Rb mutant embryos.
Given these findings and the observed cell cycle defects in E2f7-/-E2f8-/- placentas, we speculated that the unharnessed E2F3a activity may also
99 contribute to the ectopic proliferation in E2f7 and E2f8 double mutant placentas
To test this possibility, E2f7-/-E2f8-/- animals were bred with E2f3a-/- in order to generate cohorts of E2f7-/-E2f8-/-E2f3a-/- triple knockout embryos (TKO). BrdU incorporation assays performed on E10.5 TKO placental samples showed that loss of E2f3a, indeed, largely suppressed the ectopic DNA replication caused by a deficiency in E2f7 and E2f8 (Figure 4.10A). Remarkably, these TKO placentas appeared normal, with few clustered trophoblast cells and well-organized vascular structure (Figure 4.10B). Consistent with this observation, scatter plot and TUNEL analyses failed to detect nutrition- and hypoxia-related stress response and widespread apoptosis in TKO fetuses (Figure 4.10C, 4.10D).
Strikingly, live TKO embryos could be recovered at every embryonic stage analyzed, including at birth, even though the penetrance of rescue was not 100%
(Figure 4.10E). Interestingly, in sharp contrast to the rescue in the ectopic S- phase entry, examination of P-H3 and CycA2 expression in TKO placentas revealed the same mitotic defects as in DKO placentas, indicating that E2F3a is not involved in this E2F7/E2F8-mediated mitotic phenotype (Figure 4.11A-4.11D).
From these results, we conclude that E2F3a genetically interacts with
E2F7/E2F8 and its unrestrained activity contributes to the ectopic proliferation but not the aberrant mitosis in E2f7-/-E2f8-/- placentas.
We then sought to elucidate the molecular mechanism underlying the genetic interaction between E2f3a and E2f7/E2f8. Quantitive RT-PCR analysis revealed a similar level of E2f3a mRNA expression in E2f7-/-E2f8-/- placental tissues when compared to wild type, suggesting that unlike E2f1, E2f3a is not a
100 direct downstream target of E2F7 and E2F8 (Figure 4.12A). Since E2F7 and
E2F8 were believed to control S-phase entry by directly repressing a subset of
G1-S genes and loss of E2f3a suppressed the ectopic S-phase entry caused by
a deficiency in E2f7 and E2f8, we hypothesized that E2F3a may function actively
on the same set of target genes, and therefore loss of E2F3a-mediated activation
could compensate for supra-activation of these genes in E2f7-/-E2f8-/- placenta.
To test this possibility, we focused our analysis on the G1-S genes which were
previously shown as E2F7/E2F8 direct targets (Figure 4.3A). Quantitive RT-PCR
analysis revealed that loss of E2f3a partially suppressed the over-expression of
these genes in E2f7-/-E2f8-/- placentas (Figure 4.12B). ChIP-PCR assays
confirmed that E2F3a could indeed bind these gene promoters and presumably
activated their expression. Interestingly, the occupancy of E2F3a to these gene
promoters remained unchanged in the absence of E2f7 and E2f8 (Figure 4.12C).
From these results, we conclude that activator E2F3a and repressor E2F7/E2F8 likely play antagonistic roles in the control of S-phase entry by independently co- regulating a same set of G1-S targets.
4.3 Discussion
The oldest evolutionary arm of the E2f family, consisting of E2f7 and E2f8,
is the least well understood. In this study, we demonstrate that E2F7 and E2F8
play a dual role in cell cycle regulation. On one hand, E2F7 and E2F8 directly
repress a subset of G1-S genes that serve to control S-phase entry (Figure 4.13).
Upregulation of these genes was tolerated by most cell types in E2f7-/-E2f8-/-
101 embryos, but had catastrophic consequences in trophoblast cells of the placenta,
resulting in ectopic S phase entry in TG and SP cell lineages. Interestingly,
these E2f7/E2f8-responsive genes are also regulated by activator E2F3a. Even
though inactivation of E2f3a failed to completely rescue the overexpression of
these genes, it significantly repressed the ectopic S-phase entry and the severe
placental defects in E2f7-/-E2f8-/- placentas, and more strikingly, prolonged the life span of E2f7-/-E2f8-/- embryos. Therefore, we propose that the ectopic S-phase
entry which is largely mediated by E2F3a has a major contribution to the
placental dysfunction and the fetal lethality in E2f7-/-E2f8-/- embryos. However,
whether such ectopic S-phase entry is the sole reason culpable for placental
dysfunction, and whether G1-S transition is the only process regulated by E2F3a
remain to be determined. Nonetheless, these results suggest that E2F7/E2F8
and E2F3a cooperatively regulate S-phase entry in the extra-embryonic lineages
and highlight the importance of keeping total E2F activity well-balanced.
On the other hand, while no G2-M specific genes were identified at the
mRNA level from microarray analysis, ectopic expression of mitotic cyclins
(CycA2 and CycB1) was readily observed at the protein level in double mutant
giant cells, which very likely contribute to the aberrant mitosis in these cells.
Analysis of E2f7-/-E2f8-/- placentas showed that E2F7 and E2F8 were not
required for the initiation of endocycle per se, since their ablation did not preclude
trophoblasts from differentiating into giant cells. Rather, the evidence presented
here suggests that the two E2Fs are required to complete endocycle
appropriately. At a time when the vast majority of normal giant cells stop
102 reduplicating their genome, E2f7/E2f8-deficient giant cells entered an additional
round(s) of S phase and remarkably, segregated their chromosomes without
undergoing cytokinesis, resulting in binucleated giant cells. These findings suggest that E2F7 and E2F8 play a critical role in the timely cessation of endoreduplication in mammals. Interestingly, while inactivation of E2f3a
corrected ectopic G1-S transition, it had little effect on the mitotic defect in DKO
giant cells. Therefore, E2F7 and E2F8 appear to function in two distinct cell
cycle pathways, that is, an E2F3a-related pathway, regulating G1-S transition, and an E2F3a-unrelated pathway, controlling G2-M progression (Figure 4.13).
We also explored the possibility of a broader impact of E2F7 and E2F8 on
DNA replication and mitosis in other cell lineages. We reasoned that the absence of overt cell cycle defects in embryonic cell lineages is likely due to the high proliferation potential of these cells and, therefore, expected to observe similar cell cycle defects in other cell types with low proliferation potential.
Indeed, cellular analyses of newborn lung samples and adult liver samples revealed that cells in these tissues lacking E2f7 and E2f8 are also over- proliferating. Therefore, E2F7 and E2F8 appear to function as important repressors to suppress cellular proliferation in general. In this regard, it is very intriguing to evaluate their potential tumor suppressor function in the future.
103 4.4 Materials and Methods
4.4.1 Affymetrix microarray analysis
Total RNA was isolated using Qiagen RNA miniprep columns according to
the manufacturer’s protocol. Global gene expression analyses were performed
on Affymetrix Mouse Genome 430 2.0 arrays at the Ohio State University
Comprehensive Cancer Center. Expression values were adjusted by quantile
normalization and log2 transformation with RMAExpress. Data was analyzed
with BRB-ArrayTools 3.7.0 (http: linus.nci.nih.gov BRB-ArrayTools.html). Class
comparison was used to select genes differentially expressed at a significance
level of p<0.05. When compared to E2f7+/+E2f8+/+, probes with a >2-fold
misexpression in E2f7-/-E2f8-/- were used and the average relative expression
level of each genetic group was loaded into the TIGR MultiExperiment Viewer
(MEV, version 4.0) (TIGR, Rockville, MD) to generate the heatmap. Clustering
and scatter plot analyses were performed by using Clustering and Scatterplot
functions on BRB Array Tools. Promoter sequences of each gene (-1000bp ~
+300bp) were obtained from UCSC genome browser (http://genome.ucsc.edu/).
TFSearch (http://www.cbrc.jp/research/db/TFSEARCH.html) aided in the
identification of genes containing E2F consensus binding sites.
4.4.2 Quantitative RT-PCR
Total RNA was extracted using Qiagen RNA miniprep columns according
to the manufacturer’s protocol, which included the optional DNase treatment
104 before elution from the column. Reverse transcription (RT) of total RNA was performed using Superscript III reverse transcriptase (Invitrogen) and RNAse
Inhibitor (Roche) as described by the manufacturer. Quantitative PCR was performed using SYBR Green reaction mix (BioRad) and the BioRad iCycler
PCR machine. All reactions were performed in triplicate and relative amounts of cDNA were normalized to Gapdh.
4.4.3 Chromatin immunoprecipitation (ChIP)
The EZ CHIPTM assay kit (Upstate Biotech) was used as described by the manufacturer. Briefly, for E2F7 and E2F8 ChIP, HEK 293 cells overexpressing wild type E2F7 or E2F8 (flag-E2F7 or flag-E2F8) or DNA binding mutant forms of
E2F7 or E2F8 (flag-E2F7DBD1,2 or flag-E2F8DBD1,2) were crosslinked, lysed and sonicated followed by incubation with 2 μg of anti-flag (M2, Sigma) or normal mouse IgG (Oncogene) antibodies at 4°C overnight. For E2F3 ChIP experiments, wild type and E2f7/E2f8-null MEFs were crosslinked, lysed and sonicated followed by incubation with 2 μg of anti-E2F3a (N-20, Santa Cruz), anti-E2F3 (C-18, Santa Cruz) or normal rabbit IgG (Oncogene) antibodies at 4°C overnight. Antibody-protein-DNA complexes were recovered by addition of 30 µl of Salmon Sperm DNA/Protein G agarose slurry (Upstate Biotech).
Immunoprecipitated DNA fractions were de-crosslinked and purified through
Qiaquick columns (Qiagen). Quantitive PCR of ChIP samples was performed using the Biorad iCycler machine with primers specific for the indicated promoter
105 regions. Reactions were performed in triplicate and normalized using the threshold cycle number for the 1% of total input sample.
4.4.4 BrdU and TUNEL assays
Pregnant females at 10.5 days postcoitum or newborn pups were injected intraperitoneally with BrdU (100 µg/grams of body weight) 30 min prior to harvesting. Samples were fixed in formalin and 5 µm paraffin embedded- sections were used. After deparaffinization, anti-BrdU antibodies (MO-0744,
DAKO), Vectastain Elite ABC reagents (Vector labs) and DAB peroxidase substrate kit (Vector labs) were used to detect BrdU incorporation according to the manufacturer’s instructions. Apoptotic cells were detected using TUNEL
(S7101, Chemican) assays, performed according to the manufacturer’s protocol.
All slides were counterstained with hematoxylin
4.4.5 Histological analysis and immunostaining
Placental tissues were fixed in 10% neutral buffered formalin and embedded in paraffin. 5 μm-thick sections were prepared. Standard hematoxylin and eosin (H&E) staining was used for general histopathological analysis. For immunochemistry (IHC), slides were probed with primary antibodies specifically against Ki-67 (550609, BD pharmingen), Phospho-Histone
3 (P-H3, Ser10) (06-570, Millipore) and CycA2 (C-19, Santa Cruz), CycB1 (H-
433, Santa Cruz). Vectastain Elite ABC reagent (Vector labs) and DAB peroxidase substrate kit (Vector labs) were used in combination to detect signals
106 by following the manufacturer’s instructions. Samples were counterstained with
hematoxylin.
Cells under M phase of cell cycle or belonging to giant cell lineage were
detected by immunofluorescence (IF) staining using P-H3 and PL-1 antibodies (a
gift from Dr. F. Talamantes), resepctively. Giant cells under G-S transition of
endocycle were detected by double IF staining using BrdU (MO-0744, DAKO)
and PL-1 antibodies. Giant cells under G-S transition of endocycle and/or
expressing CycA2 were detected by BrdU and CycA2 double IF staining.
Nuclear DNA was counterstained with DAPI.
4.4.6 Confocal microscopy and 3D reconstruction
Placental samples were collected from E10.5 mutant and control embryos,
fixed with 4% paraformaldehyde at 4°C overnight, and embedded in OCT. 70
µm-thick frozen sections were cut, washed with 0.2% Triton-X in PBS for 1.5 h and stained with 1 mM Draq5 (Biostatus) at 4°C overnight. Draq5-stained samples were examined by confocal laser scanning microscopy (Zeiss LSM 510) with the use of a 63X objective and 0.7X scan zoom. Optical sections with between-plane-plane resolution of 0.40 μm and axial resolution of 0.42 μm were acquired. Four to five separate regions were imaged with a field of view of 207
μm x 207 μm and depth of 40 ± 2 μm, resulting in a stack of 100 images for each region. For 3D reconstruction, the image stacks were processed using Insight
Toolkit and ITK-SNAP (Yushkevich et al., 2006). Each cell nucleus was segmented by active contour segmentation using region competition as the
107 stopping criterion (Zhu and Yuille, 1996). Nuclei of binucleated and
multinucleated cells were constructed separately and assigned different colors.
4.4.7 Transmission electron microscopy (TEM)
E10.5 placentas were harvested and fixed with 3% glutaraldehyde in 0.1
M phosphate buffer overnight at 4°C. Following fixation, tissues were processed
according to TEM Fixation Protocol (http://cmif.osu.edu/419.cfm) and embedded in Eponate resin for TEM observations. Ultrathin cross sections of placenta (70 nm) were mounted on copper grids, double stained with uranyl acetate and lead citrate, and observed with a Technai G2 Spirit transmission electron microscope
(FEI).
4.4.8 Western blot and antibodies
Immunoblot analyses were performed by standard procedures using ECL reagents as described by the manufacturer (Amersham Biosciences). CycA2 (C-
19, Santa Cruz), CycB1 (H-433, Santa Cruz) and Tubulin (T6199, Sigma)
antibodies were used for detection.
4.4.9 Quantification and statistical analysis
Images of immunostained sections were captured using Eclipse 50i
(Nikon) and Axioskop 40 (Zeiss) microscopes and quantified using Metamorph
Imaging 6.1 software. Quantification of BrdU-postive, P-H3-positive, CycA2-
positive and CycB1-positive cells was achieved by calculating the percentage of
108 positive cells in the different regions or in the different cell lineages. Counting data are reported as the average ± SD fold induction of percentage of positive cells. Three sections per sample and at least three different samples for each genotype were analyzed. Binucleated cells were quantified using LSM image browser (Zeiss). Results were represented as the average ± SD percentage of binucleated cells from three different samples. At least 143 giant cells were counted for each genotype group. Pairwise comparisons were evaluated by two- tailed Student’s t-tests.
109 Figure 4.1 E2F7/E2F8-dependent gene expression in the placenta. (A)
Dendogram for unsupervised clustering analysis using centered correlation and average linkage. E10.5 placentas with the indicated genotypes were presented individually. Note that the segregation of these placentas is based on their genotypes. (B) Heatmap representation of differentially expressed probe sets generated from the comparison of E2f7+/+E2f8+/+ and E2f7-/-E2f8-/- placenta microarray data (>2-fold, p<0.05). Results are presented as an average of relative expression level. n, number of placentas analyzed for each genetic group. (C) A two-set Venn diagram representation of genes differentially upregulated >2-fold in E2f7-/-E2f8-/- and Cyp19-cre placentas
(p<0.05). Note that the overlapping area between the two data sets contains
49 genes. (D) Pie chart illustrating the major gene ontology categories that the 49 differentially expressed genes contribute to. (E) Expression of representative target genes from each category was evaluated by quantitive
RT-PCR analyses. Wild type samples (7+/+8+/+) with the lowest expression value were normalized to 1. Data are shown as the average ± SD-fold induction of gene expression in samples analyzed in triplicate.
110 A Placentas B Placentas -0.2- 7+/+8+/+ 7-/-8-/- Sox2 Cyp19
0.0-
0.2-
0.4- Correlation 0.6-
0.8-
1.0- -/- -/- -/- -/- +/+ +/+ 8 8 8 8 8 8 -/- -/- -/- -/- n=2 n=4 n=4 n=3 7 7 7 7 +/+ +/+ Sox2 Sox2 Sox2 Sox2 Sox2 7 7 Cyp19 Cyp19 Cyp19 p<0.05 -2.0 0.0 2.0
C D -/- -/- 7 8 39% (placenta) 22% Metabol. Others 113 25 49 Cell Cycle Cyp19 (placenta) 39%
E Mcm4 (Cell cycle) Chek1 (Cell cycle) Bard1 (Cell cycle) 4- 4- 5-
3- 3- 4- 3- 2- 2- 2- 1- 1- 1- 0- 0- 0- Ak3l1 (Metabolism) Pfpl (Others) Elf5 (Others) 5-5 12- 4-
4- 10- 4 3- 8- 3-3 6- 2-
Relative gene expression 2-2 4- 1- 1-1 2- 0-0 0- 0- 7+/+8+/+ 7-/-8-/- Sox2 Cyp19 7+/+8+/+ 7-/-8-/- Sox2 Cyp19 7+/+8+/+ 7-/-8-/- Sox2 Cyp19
Figure 4.1
111 Figure 4.2 Identification of E2F7 and E2F8 direct targets by microarray analysis. (A) Left: a three-set Venn representation of genes differentially upregulated >2-fold in E2f7-/-E2f8-/- and Cyp19-cre placentas (p<0.05) and
>1.6-fold in Sox2-cre fetuses (p<0.05). Note that the overlapping area between the three data sets contains 20 genes. Right: promoter analysis of the two groups of target genes: the 20-gene and 29-gene data sets. Mouse and human promoter regions (-1000 bp ~ +300 bp) of each gene were examined for consensus E2F binding sites. Results are represented as the fraction and the percentage of genes that have at least one E2F binding site conserved between mice and humans (Conserved). (B) Expression of representative potential target genes (from the 20-gene data set) was examined in placental samples by quantitive RT-PCR assays. (C) Expression of same set of genes as in (B) was examined in primary mouse embryonic fibroblasts (MEFs) by quantitive RT-PCR assays. Control samples (7+/+8+/+ and 7+/-8+/- for placentas and MEFs, respectively) with the lowest expression value were normalized to 1. Data are shown as the average ± SD-fold induction of gene expression in samples analyzed in triplicate.
112 A Cyp19 Conserved (placenta) -/- -/- 7 8 2/29 (placenta) 29 E2F 104 6.9% -1000 +300 29 19 20 9 6 14/20 20 E2F Sox2 60 70% -1000 +300 (fetus)
B Mcm5 5- Gins2 4- Mcm4 3-3 4- 3- 2-2 3- 2- 2- 1-1 1- 1-
0- 0- 0-0
5- Mcm6 4- Cdc6 3- Zfp367 4- 3- 2- 3- 2- 2-
Relative gene expression 1- 1- 1-
0- 0- 0- 7+/+8+/+ 7-/-8-/- Sox2 Cyp19 7+/+8+/+ 7-/-8-/- Sox2 Cyp19 7+/+8+/+ 7-/-8-/- Sox2 Cyp19
C Gins2 Mcm4 Mcm5 3- 3- 3-
2- 2- 2-
1- 1- 1-
0- 0- 0- Mcm6 Cdc6 Zfp367 4- 4- 3- 3- 3- 2- 2- 2- 1- 1- 1- Relative gene expression 0- 0- 0- 7+/-8+/- 7-/-8-/- 7+/-8+/- 7-/-8-/- 7+/-8+/- 7-/-8-/-
Figure 4.2
113 E2F sites E2F sites Conserved Gene symbol Function (mouse) (human) Sites
Cell cycle (G1-S)
Uhrf1 (ICBP90) DNA replication , DNA repair, regulation of transcription 4 4 1
Mcm4 DNA replication, transcription 4 3 1
2810417H13Rik (Paf) DNA replication, PCNA-associated factor 3 3 2
Mcm6 DNA replication, transcription 3 3 2
Cdc6 DNA replication 3 2 2
Mcm5 DNA replication, transcription 3 2 1
Gins2 DNA replication 3 1 1
E2f1 G1 to S progression, apoptosis 2 2 2
2600005O03Rik (Dscc1) DNA replication 2 2 1
Mcm10 DNA replication 1 3 1
Cdc45l DNA replication, cell division 0 1 0 DNA damage and cell death
Dtl (Cdt2) DNA damage 2 2 2
Bard1 DNA damage, DNA repair, apoptosis 2 1 0
Chek1 DNA damamge, DNA repair, meiotic recombination 1 4 1 Ung DNA base-excision repair 0 3 0 Metabolism
Fbxl20 protein ubiquitination cycle 2 2 1 Asb7 protein ubiquitination cycle 0 1 0 Others
Zfp367 regulation of transcription, DNA-dependent 3 2 2 Tacstd1 epithelial cell adhesion molecule, signaling cascade 1 2 0 Ccdc59 regulation of transcription, DNA-dependent 1 0 0
Table 4.1 20 direct targets of E2F7 and E2F8.
114 E2F sites E2F sites Conserved Gene symbol Function (mouse) (human) sites
Cell cycle (G1-S) Cdkn1a (p21) G1/S arrest upon a variety of stress stimuli 1 0 0 Cell cycle (G2-M) chromatin silencing, centric heterochromatin formation, Hells 2 3 2 mitosis, apoptosis DNA damage and cell death oxygen sensor, promotes cell death through a caspase- Egln3 1 3 0 dependent mechanism Fasl cell death 0 0 0 Metabolism Srd5a1 androgen biosynthesis 3 2 0 Fbxo2 ubiquitin cycle, proteolysis 3 1 0 nucleobase, nucleoside, nucleotide and nucleic acid Ak3l1 /// LOC100047616 2 2 0 metabolism Trpm5 ion transport 2 0 0 Slco5a1 transport activity 1 1 0 Pthlh cAMP metabolism, cell proliferation, development 1 0 0 Aplp1 extracellular matrix, biogenesis 0 1 0 Ecm1 extracellular matrix, transport 0 0 0 Klk1b22 /// Klk1b9 proteolysis 0 0 0 Others Phf19 regulation of transcription 4 3 1 Plxnb2 regulation of axonogenesis 1 2 0 Timp1 (x-chromosome) cell proliferation, erythrocyte maturation 1 1 0 Elf5 ectoderm development, transcription, EST family 1 0 0 Irx1 transcription, development 0 3 0 Ceacam1 neovascularization, cell adhesion 0 2 0 Dll4 Angiogenesis, activates Notch 0 2 0 Pfpl invasive trophoblast giant cells 0 n/a 0 Myh6 muscle contraction 0 1 0 Rpgrip1 eye photoreceptor cell development, response to stimulus 0 0 0 5830482F20Rik unknown 0 0 0 Ifi30 unknown 0 0 0 LOC100040854 /// LOC100040861 /// unknown 0 n/a 0 OTTMUSG00000010207 BC062258 could not find n/a n/a n/a D13Ertd608e could not find n/a n/a n/a EG627488 /// EG667692 /// n/a n/a n/a EG667695 /// LOC100043292 could not find
Table 4.2 29 placenta-specific targets of E2F7 and E2F8.
115 Figure 4.3 Confirmation and functional annotation of the identified direct targets. (A) E2F7 and E2F8 promoter occupancy in HEK293 cells overexpressing flag-tagged wild type (w) and DNA binding mutants (m) of
E2F7 or E2F8 was confirmation by using a standard ChIP protocol. Flag antibodies (F) were used to immunoprecipitate E2F7 or E2F8 respectively.
Normal mouse IgG (Ig) was used as a negative control. Immunoprecipitates were amplified with specific primers flanking the E2F binding sites on the indicated promoters, with primers of irrelevant sequences from E2f1 exon1
(E2f1ds) and the Tubulin promoter (Tub). Quantitive PCRs were performed in triplicate and cycle numbers were normalized to 1% of the input. Error bars represent standard deviations. (B) Functional annotation of 20 direct targets of E2F7 and E2F8. (C) Functional annotation of 29 placenta-specific targets of E2F7 and E2F8. Details for each gene function are indicated in Table 4.1,
4.2. Met. Metabolism.
116 A
Mcm4Mcm5 Chek1 Ung DtlZfp367 E2f1 E2f1ds Tub 0.06 -2 4 0 5 2 0.05 -0 0 0 8 0.04 - 5 8 0.03 -6 0 6 5 0.02 -4 4 0 2 0.01 - 2 5 % of total input 0.00 -0 0 0 flag-E2F7 0.06 -2 7 4 4
0.05 -0 6 3 5 3 0.04 -8 4 0.03 -6 2 2 3 4 0.02 - 2
0.01 -2 01 % of total input 0.00 -0 0 0 0 FIgF FIgFFIgF FIg F F IgF FIgF F Ig F F Ig F F Ig F w m w m w m w m w m w m w m w m w m flag-E2F8
BC10% 15% 73% Met. Others Others G1/S Met. Cell cycle Cell cycle DNA 31% 55% 75% damage 14% 27%
Figure 4.3
117 Figure 4.4 Ectopic proliferation in E2f7 and E2f8 double mutant placenta, lung and liver tissues. (A) E10.5 placentas with the indicated genotypes were analyzed by BrdU/PL-1 double immunofluorescence staining with BrdU in green, PL-1 in red and DAPI in blue. (B) Percentages of BrdU-positive cells were quantified for the indicated extra-embryonic lineages (TG, trophoblast giant cells; SP, spongiotrophoblasts; LT, labyrinth trophoblasts). (C) BrdU analysis of lung tissues from E2f7loxp/-E2f8loxp/- (control) and Sox2-cre;E2f7loxp/-
E2f8loxp/- (Sox2-cre) pups at P0. (D) Quantification of BrdU-positive cells in lung tissues with the indicated genotypes. (E) Ki-67 IHC staining of liver tissues from E2f7loxp/loxpE2f8loxp/loxp (control) and Mx1-cre;E2f7loxp/loxpE2f8loxp/loxp
(Mx1-cre) mice at 6 months old. (F) Quantification of Ki-67-positive cells in liver tissues with the indicated genotypes. All quantification data are presented as the average ± SD percentage of positive cells. n, number of samples analyzed for each genetic group. Pairwise comparisons were evaluated by two-tailed Student’s t-tests (** p<0.006).
118 ABTG SP LT E2f7+/+E2f8+/+ E2f7-/-E2f8-/- 40- 50- 20- 35- ** ** 40- E2f7+/+E2f8+/+ E2f7-/-E2f8-/- 30- 15-
DAPI 25- / 30- 20- 10- 20- 15- PL-1 / 10- 5- 10-
%+ve BrdU cells 5- n=5 n=5 0- n=5 n=5 0- n=5 n=5 BrdU 0- E2f7+/+E2f8+/+ E2f7-/-E2f8-/-
C D Lung Control Sox2-cre 9- **
6-
BrdU 3- n=3 n=6
% BrdU +ve % BrdU cells Control
0- Sox2-cre
Liver E Control Mx1-cre F 7- 6- ** 5- 4- 3- Ki-67 2- n=1 n=2 Control % +ve Ki-67 cells 1- 0- Mx1-cre
Figure 4.4
119 Figure 4.5 Aberrant mitosis in E2f7-/-E2f8-/- trophoblast giant cells. (A)
E10.5 placentas with the indicated genotypes were analyzed by P-H3 IHC.
Arrow points to an aberrant P-H3-positive giant cells in the E2f7-/-E2f8-/- placenta. (B) Percentages of P-H3-positive cells were quantified for the indicated extra-embryonic lineages (TG, trophoblast giant cells; SP, spongiotrophoblasts; LT, labyrinth trophoblasts). Quantification data are presented as the average ± SD percentage of positive cells. n, number of samples analyzed for each genetic group. Pairwise comparisons were evaluated by two-tailed Student’s t-tests (** p<0.006). (C) Representative inappropriate mitotic figures (left: anaphase; right: telophase) observed in
E2f7-/-E2f8-/- giant cells by IF P-H3 staining. Insert represents an adjacent spongiotrophoblast cell with a normal size nucleus. (D) The identity of mitotic giant cells was verified by staining with a giant cell-specific marker, PL-1.
Scale bars, 10 μm.
120 A E2f7+/-E2f8+/- E2f7-/-E2f8-/- P-H3
B TG SP LT 20-25 8- 29-
20 0 15- ** 6- 86- 15 10- 4- 6 10 43- 5- 2- 5
% P-H3 +ve % P-H3 cells 2 n=7 n=7 n=7 n=7 n=7 n=7 0-0 0- 00- E2f7+/-E2f8+/- E2f7-/-E2f8-/-
C E2f7-/-E2f8-/- (TG) DAPI / P-H3
D E2f7-/-E2f8-/- (TG) DAPI / PL-1
Figure 4.5
121 Figure 4.6 E2f7-/-E2f8-/- placentas contain binucleated giant cells. (A) Top panels: confocal laser scanning microscopy of E10.5 placentas. Nuclear DNA was stained with Draq5 (pseudocolored in green). Note the binucleated giant cells in E2f7-/-E2f8-/- placentas. Scale bars, 10 μm. Bottom panels: 3D confocal reconstruction of a representative binucleated giant cell. (B)
Binucleated giant cells identified by confocal microscopy were quantified for the indicated genotypes. Data are represented as the average ± SD percentage of binucleated cells. n, number of placenta samples analyzed for each genetic group Pairwise comparisons were evaluated by two-tailed
Student’s t-tests (** p<0.04). (C) Top panel: transmission electron micrograph
(TEM) of giant cells in E2f7-/-E2f8-/- E10.5 placentas. The yellow dotted-line marks the giant cell membrane border and the yellow arrows point to two nuclei within the same giant cell. Bottom panel: a high magnification picture of the boxed area in top panel. Red arrows point to the closely apposed nuclear envelops (N.E.) of two different nuclei. Note the absence of an intervening cell membrane between the adjacent nuclei. Scale bars, 2 μm.
122
A E2f7-/-E2f8-/- (TG) E2f7+/-E2f8+/- (TG) 3D reconstruct. Draq5
B C -/- -/- TG E2f7 E2f8 60-60 **
40-40
20-20 TEM % Binucleated cells N.E. n=3 n=3 0- 0 N.E. E2f7+/-E2f8+/- E2f7-/-E2f8-/-
Figure 4.6
123 Figure 4.7 Deregulation of CycA2 expression in E2f7-/-E2f8-/- giant cells.
(A) Western blot analysis of CycA2 expression in placentas with the indicated genotypes. Tubulin was used as an internal loading control. (B)
Quantification of CycA2-positive giant cells with the indicated genotypes. The percentages of BrdU-positive and BrdU-negative giant cells that are CycA2- positve are shown as solid and dot bars, respectively. Data are represented as the average ± SD percentage of CycA2-positive cells. n, number of samples analyzed for each genetic group Pairwise comparisons were evaluated by two-tailed Student’s t-tests (** p<0.04). (C) BrdU/CycA2 double
IF staining of E10.5 placentas with BrdU in red, CycA2 in green and DAPI in blue. Arrows indicate CycA2-positive/BrdU-negative giant cells in the double mutant placentas.
124 AB 70- TG 60- ** 7+/-8+/+ 7-/-8-/- 50- CycA2 40- 30- Tubulin 20-
10- BrdU -ve % CycA2 +ve% CycA2 cells n=5
n=5 BrdU+ve 0- 7+/-8+/- 7-/-8-/- C E2f7+/-E2f8+/- E2f7-/-E2f8-/- DAPI BrdU CycA2 Merged
Figure 4.7
125 Figure 4.8 Deregulation of CycB1 expression in E2f7-/-E2f8-/- giant cells.
(A) Western blot analysis of CycB1 expression in placentas with the indicated genotypes. Tubulin was used as an internal loading control. (B)
Quantification of CycB1-positive giant cells with the indicated genotypes. Data are represented as the average ± SD percentage of CycB1-positive cells. n, number of samples analyzed for each genetic group Pairwise comparisons were evaluated by two-tailed Student’s t-tests (** p<0.003). (C) CycB1 IHC staining of E10.5 placentas with the indicated genotypes. Arrows indicate
CycB1-positive giant cells in the double mutant placenta.
126
A B 45- TG ** 36- 7+/-8+/+ 7-/-8-/- 27- CycB1 18- Tubulin % CycB1 +ve cells 9- n=5 n=5 0- E2f7-/-E2f8-/- E2f7+/-E2f8+/- C E2f7+/-E2f8+/- E2f7-/-E2f8-/- CycB1
Figure 4.8
127 Figure 4.9 Mitotic defects in Mx1-cre liver cells. (A) P-H3 IHC staining of liver tissues from E2f7loxp/loxpE2f8loxp/loxp (control) and Mx1- cre;E2f7loxp/loxpE2f8loxp/loxp (Mx1-cre) mice at 6 months old. (B) Quantification of
P-H3-positive cells in liver tissues with the indicated genotypes. (C) 6-month old mouse liver samples with the indicated genotypes were analyzed by
CycA2 IHC staining. (D) Quantification of CycA2-positive cells in the liver tissues. (E) 6-month old mouse liver samples with the indicated genotypes were analyzed by CycB1 IHC staining. (F) Quantification of CycB1-positive cells in the liver tissues. All quantification data are presented as the average ±
SD percentage of positive cells. n, number of samples analyzed for each genetic group Pairwise comparisons were evaluated by two-tailed Student’s t- tests (** p<0.003).
128 Liver A Control Mx1-cre B 2.5-
2.0- **
1.5-
1.0- P-H3 n=1 n=2
0.5-
% P-H3 +ve % P-H3 cells Control 0- Mx1-cre
Liver C Control Mx1-cre D 3.5- ** 3.0- 2.5- 2.0- 1.5-
CycA2 1.0- n=1 n=2 0.5- Control % CycA2 +ve cells 0- Mx1-cre
Liver E Control Mx1-cre F 1.5- **
1.0-
CycB1 0.5- Control % CycB1 +ve cells n=1 0- n=2 Mx1-cre
Figure 4.9
129 Figure 4.10 Loss of E2f3a rescues the ectopic DNA replication in E2f7-/-
E2f8-/- placentas and prolongs life span of the fetus. (A) Quantification of
BrdU-positive cells in trophoblast giant cell (TG) and spongiotrophoblast cell
(SP) lineages with the indicated genotypes. (B) HE staining of E10.5 placenta tissue sections with the indicated genotypes. (C) Scatter plot analyses of stress-related gene expression in the fetus between indicated genetic groups.
A 2-fold cutoff is shown. (D) Quantification of TUNEL-positive cells in the indicated tissue areas of E10.5 fetuses with the indicated genotypes. (E)
Genotypic analysis of embryos derived from E2f7/E2f8/E2f3a crosses at the indicated stages of development. All quantification data are presented as the average ± SD percentage of positive cells. n, number of samples analyzed for each genetic group Pairwise comparisons were evaluated by two-tailed
Student’s t-tests (* p<0.06, ** p<0.003).
130
AB TG SP 40- 50- E10.5 placentas * ** 32- 40- E2f7+/+E2f8-/-E2f3a-/- E2f7-/-E2f8-/-E2f3a-/-
24- 30-
16- 20- H&E
% BrdU +ve% BrdU cells 8- 10- n=5 n=5 n=4 n=5 n=5 n=4 0- 0- E2f7+/+E2f8+/+ E2f7-/-E2f8-/- E2f7-/-E2f8-/-E2f3a-/- CD Head Branchial arch Somite Stress-related gene expression in the fetus 6- 35- 28- ** 13 -/- 28- ** ** -/- 21- 8 11 3a -/-
-/- 4-
7 21- 9 8 -/- 14- 7 7 14- n=3 2- n=3 n=3
Relative 7- n=3 5 n=3 n=3 n=3 7+/+8+/+ 7+/+8+/+ 7- n=3 n=3 3 % +ve TUNEL cells expression (log 2) (log expression 3 5 7 9 11 13 35791113 0- 0- E2f7+/+E2f8+/+ E2f7-/-E2f8-/- E2f7-/-E2f8-/-E2f3a-/- Relative expression (log 2) E
Genotypic analysis of embryos during development: E2f7/E2f8/E2f3a crosses E2F3a+/+ E2f3a+/- E2f3a-/- total E2f7-/-E2f8-/- E2f7-/-E2f8-/- E2f7-/-E2f8-/- E10.5 - 9 23(1) 118 expected - 10 20 E14.5 - - 6(2)a 57 expected - - 10 P0 - 0 2(1) 37 expected - 3 5 () number of dead embryos; Exact binomial test: a highly significant (p<0.0007)
Figure 4.10
131 Figure 4.11 Ablation of E2f3a does not rescue the mitotic defect observed in E2f7-/-E2f8-/- placentas. (A) P-H3 IHC staining of E10.5 placental tissues from E2f7-/-E2f8-/-E2f3a-/- and its littermate control. (B)
Quantification of P-H3-positive cells in trophoblast giant cells with the indicated genotypes. (C) E10.5 placental samples with the indicated genotypes were analyzed by CycA2 IHC staining. (D) Quantification of CycA2-positive cells in placental tissues. All quantification data are presented as the average ± SD percentage of positive cells. n, number of placenta samples analyzed for each genetic group.
132 TG A B 40- Control E2f7-/-E2f8-/-E2f3a-/- 30-
20- P-H3 10- Control n=3 n=3 n=3 E2f7-/-E2f8-/-E2f3a+/-
% P-H3 +ve cells -/- -/- -/- 0- E2f7 E2f8 E2f3a
TG C D 90- Control E2f7-/-E2f8-/-E2f3a-/-
60-
30- CycA2 Control -/- -/- +/- n=3 n=3 n=3 E2f7 E2f8 E2f3a
% CycA2 +ve CycA2 % cells 0- E2f7-/-E2f8-/-E2f3a-/-
Figure 4.11
133 Figure 4.12 E2F3a and E2F7/E2F8 co-regulate same set of G1-S target genes. (A) Quantitive RT-PCR analysis of E2f3a expression in placenta tissues with the indicated genotypes. (B) The indicated G1-S genes expression was evaluated by quantitive RT-PCR assays in the placental tissues with the indicated genotypes. (C) ChIP assays of cell lysates from proliferating wild type (black) and double knockout (red) MEFs using antibodies against E2F3a, total E2F3 and normal rabbit IgG.
Immunoprecipitated DNA was measured by quantitive PCR using primers flanking the E2F binding sites on the indicted promoters. All data are presented as the average ± SD-fold induction in samples analyzed in triplicate.
134 AB E2f3a Cdc6 Mcm4 Mcm5 8- 4- 4- 3-
6- 3- 3- 2-
4- 2- 2- 1- 2- 1- 1- Relative gene expression Relative gene expression 0- 0- 0- 0- 7+/+8+/+ 7-/-8-/- 7+/+8+/+ 7-/-8-/- 7-/-8-/-3a-/- 7+/+8+/+ 7-/-8-/- 7-/-8-/-3a-/- 7+/+8+/+ 7-/-8-/- 7-/-8-/-3a-/-
C 0.07 - Cdc6 0.07 - Mcm4 0.07 - Mcm5 0.07 - Gapdh
0.06 - 0.06 - 0.06 - 0.06 - 0.05 - 0.05 - 0.05 - 0.05 - 0.04 - 0.04 - 0.04 - 0.04 - 0.03 - 0.03 - 0.03 - 0.03 - 0.02 - 0.02 - 0.02 - 0.02 -
% of total input 0.01 - 0.01 - 0.01 - 0.01 - 0.00 - 0.00 - 0.00 - 0.00 - E2F3a E2F3 IgG E2F3a E2F3 IgG E2F3a E2F3 IgG E2F3a E2F3 IgG E2f7+/+E2f8+/+ E2f7-/-E2f8-/-
Figure 4.12
135
E2F3a? E2F3a E2F7/E2F8
G1 S PHASE G2 MITOSIS •BrdU •P-H3 • G1-S targets •CycA2, CycB1
Figure 4.13 A working model of E2F7 and E2F8 function in cell cycle regulation. E2F7 and E2F8 play dual roles in the control of cell cycle progression. On one hand, they directly repress a subset of G1-S target gene expression, counteract E2F3a-mediated activation, and, therefore, control
DNA replication (assessed by BrdU incorporation). On the other hand, E2F7 and E2F8 regulate mitosis entry (assessed by P-H3 staining) by indirectly controlling CycA2 and CycB1 protein expression. While the direct targets of
E2F7 and E2F8 involved in this pathway remain unknown, E2F3a is irrelevant to this process. Other potential functions of E2F3a, such as differentiation and migration, may also contribute to the phenotypical rescue observed in E2f7-/-
E2f8-/-E2f3a-/- triple knockout embryos.
136
CHAPTER 5
E2F7/E2F8-ASSOCIATED CO-REPRESSOR COMPLEXES
5.1 Introduction
Transcription factors do not function in isolation but rather act in concert
with associated co-factors. The ability of transcription factors to mediate
transcriptional activation versus repression can often be distinguished by their association with specific types of chromatin modifying or remodeling activities.
For example, whereas activator E2Fs have been proposed to associate
transcriptional co-activators, such as CBP/p300, GCN5, TRAPP, Tip60 and
ACTR/AIB1 to promote histone acetylation, open chromatin structure, and
therefore induce gene expression, members of repressor E2Fs, E2F4 and E2F5,
form active repressor complexes with pocket proteins (Rbs) and their recruited
co-repressors, including HDACs, BRG1, RBP1, DNMT1, CtIP, CtBP, HPC2,
mSin3b, SUV39H and PRMT5 to remove acetyl groups from histone, close chromatin structure, and thus inhibit gene expression (DeGregori and Johnsons,
2006). While E2F6-mediated repression is Rb-independent, it also involves chromatin modifying/remodeling activity. E2F6 associated co-repressor complexes contain co-repressors HMTase, HP1 and polycomb group proteins,
137 which are believed to function in long-term transcriptional silencing (La Thangue,
2002; Ogawa et al., 2002). Thus, even though E2F activators and E2F repressors have opposite roles in regulating gene expression, the molecular bases of their actions are quite similar, that is, they all function by recruiting chromatin-modifying/remodeling protein complexes to their target gene promoter, which, in turn, modify the chromatin environment and determine the transcription status.
We and other investigators have recently demonstrated that E2F7 and
E2F8 can repress gene expression and suppress cellular growth (de Bruin et al.,
2003; Di Stefano et al., 2003; Maiti et al., 2005). However, as the most recently identified and least studied E2Fs, little was known about the binding partner of
E2F7 and E2F8 and the mechanism of their repression. In order to elucidate the molecular mechanism by which E2F7 and E2F8 exert their function, we took a candidate strategy to identify co-repressors that may associate with these two
E2Fs and contribute to E2F7/E2F8-mediated repression.
5.2 Results
5.2.1 E2F7 can specifically associate with CtBP and Rb family proteins
We first examined the protein sequences of E2F7 and E2F8 for functional motifs which may indicate potential protein-protein interactions. By using ELM
(http://elm.eu.org/links.html), a computational biology resource for investigating candidate functional sites in eukaryotic proteins, we found that E2F7 protein
138 contains three conserved CtBP binding sites, PLDLS, PADLS and PLSLV, at
amino acid positions 51-55, 567-580, and 615-619, respectively, indicating that
E2F7 may directly interact with CtBP. CtBP is a family of transcriptional co-
repressors (CtBP1 and CtBP2), which have been shown to repress gene
expression in a histone deacetylases (HDAC)-dependent or -independent
manner (Chinnadurai, 2002). Co-immunoprecipitation and reciprocal co-
immunoprecipitation assays confirmed the interaction between E2F7 and CtBP
(both CtBP1 and CtBP2) (Figure 5.1A). Further characterization of this
interaction showed that the PLDLS binding motif at position 51-55 is essential for
E2F7 to bind CtBP, since substitution of PLDLS sequence to PLASS (myc-
E2F7m1) largely abolished the binding (Figure 5.1C). Unlike E2F7, E2F8 lacks a
CtBP binding site, and consistently, when overexpressed alone, no E2F8 protein could be detected in the CtBP-containing complex (Figure 5.1B, top panel).
Interestingly, however, E2F8 could be easily found associated with CtBP in the presence of E2F7 overexpression (Figure 5.1B, bottom panel). Given the finding that E2F7 and E2F8 can hetero-dimerize to each other (Chapter 2 in this study and Li et al., 2008), these results suggest that E2F8 likely associates with CtBP proteins indirectly through its interaction with E2F7.
Inspection of protein sequences also revealed that both E2F7 and E2F8 lack the classical Rb binding domain (LxCxE) that is common to E2F1-E2F5.
However, E2F7 (but not E2F8) contains an LxCxD motif at its C-terminus that could potentially function to bind Rb-related proteins (Rb, p107 and p130). To test this interesting possibility, we performed Co-IP assays to examine the
139 interactions between E2F7 and Rb family members and showed that Rb-related proteins can be co-immunoprecipitated with E2F7 (Figure 5.2A). Substitution of the LxCxD motif with AxPxY reduced the interactions of E2F7 with p107 and p130, indicating that the LxCxD sequence is required for the binding, but the adjoining sequences may also contribute to the affinity of Rbs (Figure 5.2C).
Consistent with this idea, lacking LxCxD motif and its adjacent sequences in exon 12, E2F7a, an isoform of human E2F7 protein, lost its ability to associate with p107 (Figure 5.2D). While E2F8 did not interact with Rb-related proteins
(Figure 5.2B), it remained possible that E2F8 could co-exist in the same protein complex as Rb via its hetero-dimerization with E2F7. To test this idea, HEK 293 cells were co-transfected with flag-E2F7, myc-E2F8 and HA-p107. Protein complexes associated with E2F7 were immunoprecipitated with flag antibodies and eluted from the agarose beads using flag peptides, and then subjected to the second round of immunoprecipitation of p107 with HA-specific antibodies. The presence of E2F8 was finally examined by western blotting with antibodies against myc-tag (Figure 5.2E). This analysis reveals that E2F7, E2F8 and p107 could indeed form a ternary complex.
Together, these data suggest that E2F7 specifically and, most likely, directly interacts with CtBP and pocket proteins via the defined binding motifs, and thus can recruit E2F8 to CtBP- or Rb-containing complexes.
140 5.2.2 Both E2F7 and E2F8 associate with components of the histone modifying machinery
We next extended our candidates from the identified CtBP and RB co- repressors to Sin3 and HDAC proteins. Sin3 and HDAC proteins are the core components of the histone modifying machinery, which can remove acetyl group from histones, and therefore condense chromatin conformation and repress gene transcription (Silverstein and Ekwall, 2005, Glozak and Seto, 2007). Since Sin3 and HDAC proteins have been indicated to exist in the same co-repressor complexes as CtBP and RB (Chinnadurai, 2002; DeGregori and Johnsons, 2006), we speculated that E2F7/E2F8-mediated repression might involve the recruitment of Sin3/HDAC complex. To test this hypothesis, HEK 293 cells were transfected with differentially tagged versions of E2F7 or E2F8, along with flag- tagged HDAC1-5 or myc-tagged Sin3a. Protein complexes were immunoprecipitated with antibodies against the flag, and subsequently immunoblotted with antibodies against the HA-tagged E2F7, myc-tagged E2F8, myc-tagged Sin3a, or endogenous Sin3b. The Co-IP results shown in Figure 5.3 demonstrate that both E2F7 and E2F8 can interact with HDAC1, HDAC3,
HDAC5 and Sin3a, Sin3b. HDAC2 and HDAC4 were not detected in any of the
IP complexes. In each case, where an association between E2F7 or E2F8 and a co-repressor was apparent, the interaction was confirmed by reciprocal Co-IP assays (data not shown). Sequential Co-IP experiments further revealed that
E2F7 and E2F8 can form ternary repressor complexes with HDAC1 or Sin3a, respectively (Figure 5.4A, 5.4B).
141 Based on the fact that E2F7 and E2F8 can associate with exogenously expressed HDAC proteins, we were encouraged to test whether E2F7- associated complexes possess endogenous HDAC activity. To this end, fluorometric HDAC assays were performed on flag-E2F7-assoicated protein complexes immunoprecipitated by flag antibodies. In parallel experiments, empty vector and flag-tagged HDAC1 were used as a negative and a positive control, respectively. As expected, flag-tagged HDAC1 was associated with potent HDAC activity that is sensitive to trichostatin A (TSA), an HDAC inhibitor.
A significant amount of HDAC activity was also detected for E2F7-containing complexes in the absence of TSA (Figure 5.4C). While it is not clear whether the interactions between E2F7/E2F8 and HDAC1/HDAC3/HDAC5 or Sin3a/Sin3b are direct, these experiments demonstrate that both E2F7 and E2F8 have the ability to associate with components of the histone modifying machinery, such as
HDAC and Sin3, and presumably recruit them to the chromatin of their target gene promoter and therefore repress its expression.
5.2.3 Further characterization of E2F7/E2F8-associated repressor complexes
While informative, the experiments described above failed to determine whether E2F7-associated co-repressors co-exist within the same macromolecular complex or exist as distinct protein sub-complexes. To begin to address this question, we performed sequential co-immunoprecipitation assays followed by immunoblot analysis of lysates derived from HEK 293 cells co-
142 expressing flag-HDAC1 (or an empty vector as a control) along with HA-E2F7 and myc-Sin3a. The results shown in Figure 5.5 demonstrate that E2F7 can form a macromolecular complex with HDAC1, Sin3a and potentially CtBP.
Moreover, Co-IP experiments performed in CtBP DKO or Rb TKO MEFs showed that the interaction between E2F7 and HDAC1 is, or to a large extent, independent of CtBP or Rb family proteins (Figure 5.6). Characterization of
E2F7/E2F8-associated complexes was also performed by using E2F7 and E2F8
DNA binding mutants (Figure 5.7). These Co-IP assays revealed that the interactions between E2F7/E2F8 and their identified co-repressor proteins are
DNA-independent. While it may be too early to depict the E2F7/E2F8-associated co-repressor complexes, these preliminary results begin to describe the potential complexity of how E2F7 and E2F8 may function to regulate gene expression.
5.3 Discussion
Our previous studies and studies from other laboratories have shown that
E2F7 and E2F8 are transcriptional repressors (de Bruin et al., 2003; Di Stefano et al., 2003; Maiti et al., 2005), but the mechanism of repression is far from resolved. The experiments described here reveal that E2F7/E2F8 can associate with co-repressor CtBP, pocket proteins, and the chromatin modifying factors
Sin3a/3b and HDACs in a DNA-independent manner (Figure 5.8). We show that the preferential binding of E2F7 to CtBP or Rb co-repressors is likely direct, which depends on, or at least partially depends on, the predicted binding motifs.
The domains that are responsible for the binding of Sin3a/3b and HDACs have
143 not been well characterized. Clearly, a thorough analysis of various mutation and deletion forms of E2F7 and E2F8 would be required in the future to elucidate the motifs which may be responsible for these interactions. More importantly, in order to fully characterize the biochemical basis of E2F7 and E2F8 function, an affinity purification strategy should be considered to extensively study the
E2F7/E2F8-associated co-repressor complexes or sub-complexes.
Rb, HDAC and CtBP, Sin3 co-repressors have no intrinsic DNA-binding activity, yet they must localize to a gene promoter to exert any function. The targeting of these co-repressors is accomplished by physical interacting with many sequence-specific transcription factors, including E2F7 and E2F8. In mice, ablation of E2f7 and E2f8 leads to embryonic lethality (Chapter 2 in this study and Li et al., 2008), which is highly reminiscent of the essential roles of Rbs,
HDAC, CtBP and Sin3 during mouse embryo development (Hildebrand and
Soriano, 2002; Lagger et al., 2002; Wu et al., 2003; Dannenberg et al., 2005;
Montgomery et al., 2008). Given the physical interactions between E2F7/E2F8 and these co-repressor proteins, it is intriguing to speculate that E2F7 and E2F8 might be the important downstream mediators of these co-repressors function.
Thus, biological consequences of the ablation of these co-repressors could be recapitulated by the loss of E2F7/E2F8-mediated recruitment. Future experiments to identify common targets of E2F7/E2F8 and these co-repressors might provide a mechanistic explanation for this phenotypic coincidence.
Other questions that would be of great interest include that whether occupancy of target gene promoter by these repressors is dependent on E2F7 or
144 dependent on E2F8, and whether such recruitment is specific to certain gene
promoters and cellular context. We believe that answers to these questions
would provide a better view of how E2F7 and E2F8 regulate gene-specific or cell
cycle-specific repression through their association with these repressor
complexes. Moreover, the association of E2F7 and E2F8 with chromatin
modifying factors suggests that they might be important for establishment of
localized closed DNA conformations and perhaps heterochromatic regions. In
this regard, a global study of E2F7/E2F8-mediated histone biology would be very
intriguing.
In summary, our efforts presented here further support the idea that E2F7
and E2F8 function as transcriptional repressors and have begun to provide key
mechanistic insights into their function. While these experiments are by no
means exhaustive, they do provide convincing evidence for the interactions
between E2F7 and CtBP, Rb co-repressor proteins, and the association of
E2F7/E2F8 with HDAC and Sin3, the key components of the histone modifying machinery.
5.4 Materials and Methods
5.4.1 Co-immunoprecipitation (Co-IP) and sequential Co-IP assays
Human embryonic kidney (HEK) 293 cells, CtBP1-/- CtBP2-/- or Rb-/-p107-/-
p130-/- MEFs were cultured in DMEM medium supplemented with 15% FBS and
used for co-immunoprecipitation assays. Transient transfections with the
145 indicated constructs were performed by standard calcium chloride techniques.
Cells were harvested and washed in PBS at 4˚C and cell pellets were lysed in 10
volumes of lysis buffer (0.05 M sodium phosphate pH 7.3, 0.3 M NaCl, 0.1%
NP40, 10% glycerol with protease inhibitor cocktail, Roche). Lysates were
incubated with Protein G Plus/protein A-agarose beads (Calbiochem) at 4°C for 1
h to preclear. The precleared lysates were incubated with appropriate antibody
overnight. Protein G Plus/protein A-agarose beads were added and incubated
for 1 h at 4°C. Protein complexes binding to the beads were precipitated and
resolved by SDS-PAGE followed by immunoblotting.
Sequential Co-IP assays were performed similarly using the same
protocol, except two rounds of immunoprecipitation were included. The first
round of immunoprecipitation was performed by incubation of the indicated cell
lysates with flag antibodies (M2, Sigma) and protein G/A beads. Precipitated
protein complexes were eluted with excess flag peptide (200 μg/mL, Sigma) for 1 h at 4°C and then subjected to a second round of precipitation with primary antibodies as indicated. After extensive washing, sequentially precipitated complexes were recovered by addition of 30 μl of protein G/A agarose slurry and incubation for another 1 h at 4°C. Protein samples were then analyzed by
Western blotting with the indicated antibodies.
5.4.2 Western blot and antibodies
Immunoblot analyses were performed by standard procedures using ECL reagents as described by the manufacturer (Amersham Biosciences). The
146 following commercial antibodies were used as indicated in the figures: flag (M2,
Sigma), HA (12C5A, Roche), myc (9E10, Santa Cruz), CtBP1 (C32020, BD
Transduction Laboratories), CtBP2 (C32820, BD Transduction Laboratories),
CtBP (E-12, Santa Cruz), p107 (C-18, Santa Cruz), p130 (C-20, Santa Cruz),
and Sin3b (AK-12, Santa Cruz).
5.4.3 Site-directed mutagenesis
Site-directed mutagenesis was carried out using the QuikChange site-
directed mutagenesis kit (200518, Stratagene) following the manufacturer's
protocol. The changes of CtBP binding motifs in E2F7 are as follows: change
PLDLS to PLASS at the position of 51-55 (m1), change PADLS to PAASS at the position of 576-580 (m2), and change PLSLV to PLAAV at the position of 615-
619 (m3). The Rb binding motif in E2F7 was mutated from LPCTD to APGTY at
the position of 803-807.
5.4.4 HDAC activity assay
HDAC activity assay was performed using the Upstate HDAC assay kit
(17-356, Upstate) following the manufacturer's protocol. Briefly,
immunoprecipitated complexes from the indicated cell lysates were incubated
with 100 µM HDAC assay substrate at 30°C for 30 min. Fluorescence was measured using a Spectra Max M5 fluorescent plate reader (Molecular Devices),
147 with excitation 360 nm and emission 460 nm. HDAC inhibitor TSA (1 µM) was added to serve as negative controls. All reactions were set up in triplicate and arbitrary fluorescent units were plotted for HDAC activity.
148 Figure 5.1 E2F7, but not E2F8, associates with CtBP co-repressors. (A)
E2F7 associate with CtBP1 and CtBP2. HEK 293 cells were transfected with flag-tagged E2F7 expressing plasmid. Cell lysates were then immunoprecipitated, subjected to SDS-PAGE and probed with the indicated antibodies. (B) E2F8 can be recruited to CtBP-containing complexes in the presence of E2F7. HEK 293 cells were transfected with flag-tagged E2F8, or co-transfected with myc-E2F7 and flag-E2F8 expressing plasmids. Cell lysates were then immunoprecipitated with CtBP-specific antibodies, subjected to SDS-PAGE and probed with antibodies against flag. (C) The CtBP binding motifs of E2F7 is essential for its binding to CtBP. HEK 293 cells were transfected with myc-E2F7 expressing plasmid or various myc-tagged E2F7
CtBP binding domain mutants. Cell lysates were then immunoprecipitated, subjected to SDS-PAGE and probed with the indicated antibodies. m1: CtBP binding domain 1 mutant; m2: CtBP binding domain 2 mutant; m3: CtBP binding domain 3 mutant; m1,2: CtBP binding domain 1,2 mutant; m1,3: CtBP binding domain 1,3 mutant; m1,2,3: CtBP binding domain 1,2,3 mutant.
149 AB
Transfection: flag-E2F7 Transfection: flag-E2F8 IP: α-flag IgG Input IP: α-CtBP IgG Input IB: α-flag IB: α-CtBP1
flag-E2F8 + myc-E2F7 IB: α-CtBP2 Transfection: IP: α-CtBP IgG Input IP: α-CtBP IgG Input IB: α-flag IB: α-flag
C myc-E2F7m1,2,3 myc-E2F7m1 myc-E2F7m2 myc-E2F7m1,3 myc-E2F7m3 myc-E2F7m1,2 Transfection: myc-E2F7 myc-E2F7 IP: α-IgG CtBP IB: α-myc
IB: α-CtBP Input
IB: α-myc
Figure 5.1
150 Figure 5.2 E2F7, but not E2F8, associates with Rb family of proteins. (A-
B) E2F7, but not E2F8, associate with Rb family of proteins. HEK 293 cells were transfected with flag-E2F7(A) or flag-E2F8 (B), and with myc-Rb, HA- p107 or HA-p130 expressing plasmids as indicated. Cell lysates were then immunoprecipitated, subjected to SDS-PAGE and probed with the indicated antibodies. (C) The LxCxD motif of E2F7 is not essential for the binding of Rb.
HEK 293 cells were transfected with wild type E2F7 (myc-E2F7w.t.) or E2F7
Rb binding domain mutant (myc-E2F7Rbmut.), and HA-p107 or HA-p130 expressing plasmids as indicated. Cell lysates were then immunoprecipitated, subjected to SDS-PAGE and probed with myc antibodies. HEK 293 cells with no transfection were used as a control. (D) Exon 12 of human E2F7 is required for the binding of Rb. HEK 293 cells were transfected with human
E2F7a (shorter isoform lacking Exon12) or human E2F7b (full length), and HA- p107 expressing plasmids as indicated. Cell lysates were then immunoprecipitated, subjected to SDS-PAGE and probed with HA antibodies.
(E) E2F8 can be recruited to p107-containing complexes in the presence of
E2F7. HEK 293 cells were co-transfected with flag-E2F7, myc-E2F8 and HA- p107 expressing plasmids as indicated. Cell lysates were initially immunoprecipitated (1st IP) with flag-specific antibodies. Immunoprecipitates were then eluted with flag peptides. Eluents were then either subjected to
SDS-PAGE and directly probed with the indicated antibodies, or reprecipitated
(2nd IP) with HA or IgG antibodies and then probed with myc antibodies.
151 A B Transfection: flag-E2F7 + myc-Rb flag-E2F8 + myc-Rb IP: α-flag IgG Input α-flag IgG Input IB: α-myc
Transfection: flag-E2F7 + HA-p107 flag-E2F8 + HA-p107 IP: α-flag IgG Input α-flag IgG Input
IB: α-HA
Transfection: flag-E2F7 + HA-p130 flag-E2F8 + HA-p130 IP: α-flag IgG Input α-flag IgG Input IB: α-p130
t. . t. u .t u .t. m w b b bm bw R R C 7 7 7R 7R F 7 F 7 F F 2 0 2 0 2 07 2 07 -E 1 -E 1 -E 1 -E 1 c -p c -p c -p c -p y A y A 3 y A y A 93 m H m H 9 m H m H 2 Transfection: + + 2 + + α-p107 α-p130 IP: IB: α-myc
D
Transfection: flag-hE2F7b + HA-p107 flag-hE2F7a + HA-p107 IP: α-flag IgG Input α-flag IgG Input IB: α-HA
E 2nd IP: α-HA IgG
1st IP: α-flag flag Input Input α-flag flag Input flag-E2F7 ++- - ++ + Transfection HA-p107 + + ++ +++ myc-E2F8 + + ++ +++
IB: α-myc
IB: α-HA
Figure 5.2
152 Figure 5.3 Both E2F7 and E2F8 associate with HDAC1/HDAC3/HDAC5 and Sin3a/Sin3b co-repressors. (A) HEK 293 cells were co-transfected with
HA-E2F7 or myc-E2F8 and with flag-tagged HDAC1-5 expressing plasmid as indicated. Cell lysates were then immunoprecipitated with flag antibodies, subjected to SDS-PAGE, and Western immunoblotted (IB) using the indicated antibodies. (B) HEK 293 cells were co-transfected with flag-E2F7 or flag-
E2F8, and/or myc-Sin3a expressing plasmids as indicated. Cell lysates were then immunoprecipitated with flag antibodies, subjected to SDS-PAGE and probed with the indicated antibodies.
153 A Transfection: HA-E2F7+flag-HDAC1 myc-E2F8+flag-HDAC1 IP: α-flag IgG Input α-flag IgG Input IB: α-HA α-myc
Transfection: HA-E2F7+flag-HDAC2 myc-E2F8+flag-HDAC2 IP: α-flag IgG Input α-flag IgG Input
IB: α-HA α-myc
Transfection: HA-E2F7+flag-HDAC3 myc-E2F8+flag-HDAC3 IP: α-flag IgG Input α-flag IgG Input IB: α-HA α-myc
Transfection: HA-E2F7+flag-HDAC4 myc-E2F8+flag-HDAC4 IP: α-flag IgG Input α-flag IgG Input
IB: α-HA α-myc
Transfection: HA-E2F7+flag-HDAC5 myc-E2F8+flag-HDAC5 IP: α-flag IgG Input α-flag IgG Input IB: α-HA α-myc
B Transfection: flag-E2F7+myc-Sin3a flag-E2F8+myc-Sin3a IP: α-flag IgG Input α-flag IgG Input IB: α-myc α-myc
Transfection: flag-E2F7 flag-E2F8 IP: α-flag IgG Input α-flag IgG Input
IB: α-Sin3b α-Sin3b
Figure 5.3 154 Figure 5.4 E2F7 and E2F8 can form ternary complexes with HDAC1 or
Sin3a. (A) HEK 293 cells were co-transfected with HA-E2F7, flag-HDAC1 and myc-E2F8 expressing plasmids as indicated. Cell lysates were initially immunoprecipitated (1st IP) with flag-specific antibodies. Immunoprecipitates were then eluted with flag peptides. Eluants were then either subjected to
SDS-PAGE and directly probed with the indicated antibodies, or reprecipitated
(2nd IP) with HA or IgG antibodies and then probed with myc antibodies. (B)
HEK 293 cells were co-transfected with HA-E2F7, flag-E2F8 and myc-Sin3a expressing plasmids as indicated. Cell lysates were initially immunoprecipitated (1st IP) with flag-specific antibodies. Immunoprecipitates were then eluted with flag peptides. Eluents were then either subjected to
SDS-PAGE and directly probed with the indicated antibodies, or reprecipitated
(2nd IP) with HA or IgG antibodies and then probed with myc antibodies. (C)
HDAC enzyme activity was detected in E2F7-associated complexes. HEK
293 cells overexpressing flag-HDAC1 or flag-E2F7 were immunoprecipitated with flag antibodies or IgG as a negative control. Immunoprecipitated complexes were assayed for HDAC activity as described in Materials and
Methods. Error bars represent the standard deviations for three independent measurements.
155 A 2nd IP: α-HA IgG
1st IP: α-flag flag Input Input α-flag flag Input flag-HDAC1 ++- - ++ + Transfection HA-E2F7 + + ++ +++ myc-E2F8 + + ++ +++
IB: α-myc
IB: α-HA
B 2nd IP: α-HA IgG
1st IP: α-flag flag Input Input α-flag flag Input flag-E2F8 ++- - ++ + Transfection HA-E2F7 + + ++ +++ myc-Sin3a + + ++ +++
IB: α-myc
IB: α-HA
C 1800018000 16000 14000 1200012000 10000 8000 60006000 HDAC activity HDAC 4000 2000 0 IP: α-flag1234567flag IgG flag flag IgG flag TSA: -+ - - + - - Transfection: flag-HDAC1 flag-E2F7 -
Figure 5.4 156 2nd IP: α-myc IgG
1st IP: α-flag flag Input Input α-flag flag Input flag-HDAC1 ++- - ++ + Transfection myc-Sin3a + + ++ +++ HA-E2F7 + + ++ +++
IB: α-HA
IB: α-myc
IB: α-CtBP
Figure 5.5 E2F7 can form a macromolecular repressor complex with
HDAC1, CtBP and Sin3a. HEK 293 cells were co-transfected with HA-E2F7, flag-HDAC1 and myc-Sin3a expressing plasmids as indicated. Cell lysates were initially immunoprecipitated (1st IP) with flag-specific antibodies.
Immunoprecipitates were then eluted with flag peptides. Eluents were then either subjected to SDS-PAGE and directly probed with the indicated antibodies, or reprecipitated (2nd IP) with myc or IgG antibodies and then probed with HA antibodies.
157 Figure 5.6 Interactions between E2F7 and HDAC1 is independent of
CtBP and Rb proteins. (A) CtBP-deficient MEFs were co-transfected with flag-HDAC1 and myc-E2F7 expressing plasmids. Cell lysates were then immunoprecipitated with flag antibodies, subjected to SDS-PAGE, and
Western immunoblotted (IB) using antibodies against myc. (B) MEFs lacking
Rb family members were co-transfected with flag-HDAC1 and myc-E2F7 expressing plasmids. Cell lysates were then immunoprecipitated with flag antibodies, subjected to SDS-PAGE, and Western immunoblotted (IB) using antibodies against myc.
158 A CtBP-/- MEFs
Transfection:flag-HDAC1+ myc-E2F7
IP: α-flag IgG Input
IB: α-myc
B Rb-/-p107-/-p130-/- MEFs
Transfection: flag-HDAC1+ myc-E2F7
IP: α-flag IgG Input
IB: α-myc
Figure 5.6
159 Figure 5.7 Interactions between E2F7, E2F8 and their associated co- repressors are DNA-independent. (A) Co-immunoprecipitation assays were repeated with lysates expressing E2F7 DNA-binding domain 1,2 mutant (myc-
E2F7DBD1,2), and lysates from myc-E2F7DBD1,2 and HA-p107 or flag-
HDAC1 expressed cells. Cells with wild type myc-E2F7 overexpressing were used as positive controls. (B) Co-immunoprecipitation assays were repeated with lysates expressing E2F8 DNA-binding domain 1,2 mutant (myc-
E2F8DBD1,2), and lysates from myc-E2F8DBD1,2 and myc-sin3a or flag-
HDAC1 expressed cells. Cells with wild type myc-E2F8 overexpressing were used as positive controls.
160 A Transfection: myc-E2F7 myc-E2F7DBD1,2 IP: α-CtBP IgG Input α-CtBP IgG Input IB: α-myc α-myc
Transfection: myc-E2F7+HA-p107 myc-E2FDBD1,2+HA-p107 IP: α-HA IgG Input α-HA IgG Input
IB: α-myc α-myc
Transfection: myc-E2F7+flag-HDAC1 myc-E2F7DBD1,2+flag-HDAC1 IP: α-flag IgG Input α-flag IgG Input IB: α-myc α-myc
B Transfection: myc-E2F8+flag-HDAC1 myc-E2F8DBD1,2+flag-HDAC1 IP: α-flag IgG Input α-flag IgG Input IB: α-myc α-myc
Transfection: flag-E2F8+myc-Sin3a flag-E2F8DBD1,2+myc-Sin3a IP: α-flag IgG Input α-flag IgG Input
IB: α-myc α-myc
Figure 5.7
161
HDACs
RBs Sin3
CtBP E2F7 E2F8
Figure 5.8 Schematic representation depicting the results obtaining from the Co-IP experiments. Interactions indicated as black arrows have been described before. Interactions indicated as red arrows are shown in this chapter.
162
CHAPTER 6
DISCUSSION
6.1 Role of E2Fs in development
Because of the intense interest in E2Fs as major regulators of the cell
cycle and apoptosis, individual E2F family members, including E2f1 through
E2f6, have been extensively studied in vivo by gene targeting approaches in
mice. Surprisingly, with the exception of the E2f3 knockout, defects in embryos
deficient for each of the known E2Fs are rather subtle (Field et al., 1996;
Yamasaki et al., 1996; Lindeman et al., 1998; Humbert et al., 2000; Rempel et
al., 2000; Murga et al., 2001; Storre et al., 2002). Even the disruption of E2f3,
when in a mixed genetic background, yields viable mice (Humbert et al., 2000;
Wu et al., 2001). The virtual absence of cell proliferation and apoptotic defects in
E2F deficient embryos has raised questions about the physiological importance
of these factors, leaving the impression that E2Fs must either not be critical for
the control of these processes in vivo, or that there is sufficient functional
redundancy among family members to accommodate for a deficiency in any single E2F. In fact, in vivo and in vitro studies have argued for the latter point of
view and provided convincing evidence for functional redundancy among E2F1-
163 E2F3 activators and E2F4-E2F6 repressors in the control of proliferation
(Gaubatz et al., 2000; Wu et al., 2001; Giangrande et al., 2004; Tsai et al., 2009).
It is obviously a huge challenge, if not impossible, to knockout all E2fs in combination. Fortunately, analysis of DP proteins (DP1 and DP2) provides a simpler system to study E2F activity. Since E2F1-E2F6 need to dimerize with their dimerization partner DP proteins to bind DNA, ablation of DP1 and DP2 could functionally eliminate the total activities of these typical E2Fs. Embryos deficient in Dp1 died during embryonic stage between E10.5-E12.5 due to severe placental abnormalities, including impaired proliferation and reduced endoreplication in giant cells (Kohn et al., 2003). Although loss of Dp1 is still likely to be functionally compensated by DP2, this observation does argue for the essential role of E2F activity during development.
In this study, we investigated the physiological functions of E2F7 and
E2F8, the most recently identified and least understood E2F members.
Consistent with the idea of functional redundancy, we found disruption of either
E2f7 or E2f8 also had little consequence on mouse development. Their combined ablation, however, resulted in widespread apoptosis, profound placental defects, leading to embryonic death by E11.5. By identifying the critical tissues of E2F7 and E2F8 function, we were able to demonstrate their critical roles in the control of apoptosis and cell cycle progression, and elucidate the molecular mechanisms underlying these functions. Base on these findings, a working model for E2F7 and E2F8 function during embryonic development has been proposed (Figure 6.1). In this chapter, we will present this model and
164 discuss its implications for developmental biology and the diagram of E2F function.
6.2 Critical tissues and cell lineages of E2F7 and E2F8 function
E2F7 and E2F8 are widely expressed in fetuses and placentas during embryo development, but more restrictedly expressed in adult tissues (de Bruin et al., 2003; Maiti et al., 2005; Li et al., 2008). While function of E2F7 and E2F8 in these tissues has not been thoroughly evaluated, our analyses reveal that they are essential for fetal and placental development and play a critical role in maintaining the homeostasis of lung and liver tissues.
In vitro and in vivo experiments described in this study provide clear-cut evidence in support of a role for E2F7 and E2F8 in the control of apoptosis in the fetus (Figure 6.1). Total deletion of E2f7 and E2f8 (both in placental and in fetal compartments) induces widespread cell death in the fetus. While the extrinsic stress caused by loss of E2F7 and E2F8 in the placenta is required for achieving this full apoptotic phenotype, ablation of E2f7 and E2f8 in the fetus alone is sufficient to elicit apoptotic response in some cell lineages (e.g. bronchial arch and somites), but not in others (e.g. head). We speculate that the distinct apoptotic responses of these cell lineages may due be to the different developmental context or physiological microenvironment that they belong to.
In addition to their function of governing apoptosis in embryonic lineages,
E2F7 and E2F8 also play a critical role in the extra-embryonic compartment
(Figure 6.1). Actually, from the aspect of fetal survival, the placental function of
165 E2F7 and E2F8 is more important. This importance has been highlighted by conditional knockout experiments showing that their function in the placenta was both necessary and sufficient for embryo development and survival. Thus, E11.5 lethality of E2f7 and E2f8 DKO embryos could reflect the high susceptibility of placenta to loss of E2f7/E2f8, and the restrictive requirement of a well-functioning placenta to maintain a successful pregnancy. Interestingly, a recent survey of knockout mice with embryonic lethality revealed a prominent association of placental phenotypes (Rossant and Cross, 2001; Watson and Cross, 2005).
While the leading reason(s) for the embryonic lethality seen in these cases is not well-defined, our data suggest that the primary cause of fetal death is likely to be disrupted placental function.
The cell lineage(s) responsible for the abnormalities of E2f7/E2f8-mutant placenta is still not entirely clear. There are mainly three cell lineages in the placenta: labyrinth trophoblast cells, spongiotrophoblast cells and giant cells, two of which (i.e. spongiotrophoblast cells and giant cells) exhibited cell cycle defects in the absence of E2f7 and E2f8. Our data have shown that loss of E2f7 and
E2f8 function specifically in spongiotrophoblast cell lineage was not sufficient to disrupt the normal function of placenta. We also found that combinational ablation of E2f3a and E2f7/E2f8 could restore normal morphology of the placenta and partially prolong fetal life span. Interestingly, aberrant mitosis could still be easily observed in these TKO placentas, indicating that defective giant cells may not be solely responsible for the dysfunction of E2f7/E2f8-mutant placentas either. Based on these results, we speculate that the detrimental consequence
166 seen in the E2f7-/-E2f8-/- placentas is a combination of the effects of partially crippled functions of both cell lineages (i.e. spongiotrophoblast cells and giant
cells) that, when occurring simultaneously in the placenta, greatly magnifies the lethality phenotype in the affected embryos. Nevertheless, the contribution of giant cells to placental function needs to be well evaluated. In this regard, it would be very useful to generate a giant cell-specific cre mouse to achieve the specific deletion of E2f7 and E2f8 in the giant cell lineage.
We also have some indication that ablation of E2f7 and E2f8 perturbs cell
cycle regulation in lung tissues of newborn pups and liver tissues of adult mice.
It is not clear why this phenotype could be only observed in these tissues at
postnatal or later stages of development, but not during embryonic development.
One possible explanation could be that at embryonic stages, these tissues are
highly proliferative, which may make the difference in proliferation hardly visible.
Given this over-proliferation phenotype, it would be very exciting to examine the
biological consequence of loss of E2f7 and E2f8 in these tissues over time during
development, and study a potential tumor suppressor function of these two E2Fs
in a liver or lung cancer setting.
6.3 Cellular processes and molecular pathways regulated by E2F7 and
E2F8
In order to understand the cellular processes and molecular pathways
controlled by E2F7 and E2F8, we focused our effort on the placenta and the
fetus, two of the critical sites identified above. Close examination revealed that
167 E2F7 and E2F8 regulate a similar set of genes in these tissues, but control different cellular processes in a tissue-dependent manner. E2F7 and E2F8 function as negative regulators for apoptosis in the fetus. The mechanism of their action involves the direct repression of E2f1 (Figure 6.1). E2F7 and E2F8 also repress E2f1 expression in the placenta. But interestingly, instead of inducing massive apoptosis as seen in the fetus, loss of E2f7 and E2f8 in the placenta adversely affects cell cycle progression (Figure 6.1).
While the underlying reason for the lack of E2F1-mediated apoptosis in
DKO placentas remains unknown, we demonstrated that the molecular mechanisms of the observed cell cycle defects (ectopic DNA replication in SP,
TG cells, and aberrant mitosis in TG cells) involve the derepression of molecular pathways necessary for controlling the G1-S and G2-M transitions (Figure 6.1).
Interestingly, simultaneous deletion of E2f3a rescues the ectopic DNA replication, but not the aberrant mitosis, which nicely uncouples these events and strongly suggests that E2F7 and E2F8 play a dual role in the control of cell cycle by regulating an E2F3a-dependednt G1-S pathway and an E2F3a-independednt
G2-M pathway (Figure 6.1).
We propose that E2F7 and E2F8 regulate G1-S transition through direct transcriptional derepression of genes required for DNA synthesis. Loss of
E2F7/E2F8-mediated repression and, potentially, retention of E2F3a-mediated activation, upregulate these gene expression and lead to impaired DNA replication in SP and TG cells. The inappropriate nuclear division observed in double mutant TG cell, on the other hand, could be explained by the
168 constitutively high level of CyclinA2 and CyclinB1 proteins. However, the molecular interactions between E2F7/E2F8 and CyclinA2, CyclinB1 are less clear. Loss of E2f7 and E2f8 has only a mild effect on their mRNA expression, suggesting that these mitotic cyclins are not, at least to less extent, transcriptionally regulated by E2F7 and E2F8. It is believed that mitotic cyclins are mainly regulated at the protein level by APC/C-mediated proteolysis
(Malumbres and Barbacid, 2009). In flies, APC2, a subunit of the APC/C complex, has been shown to be critical for appropriate expression of mitotic cyclins, and therefore for restraining endoreplicated nurse cells from re-entry mitosis (Reed, et al, 1997). Thus, E2F7 and E2F8 might function directly or indirectly in the APC/C pathway to inactivate mitosis-promoting activity.
Consistent with the idea, in plant Arabidopsis thaliana, E2Fe/DEL1 (an
E2f7/E2f8-like gene) has been indicated to control the expression of Cdh1, an activator of the APC/C complex (Lammens et al., 2008). While the details of how atypical E2Fs function in mammals and plants may differ, these results indicate that the control of G2-M transition by these atypical E2Fs is an evolutionarily conserved strategy, which likely involves the APC/C pathway.
Taken together, the dichotomy of E2F7 and E2F8 function in the fetal and placental tissues clearly underscores the importance of understanding gene function in the context of cell lineages. While E2F7 and E2F8 might also function in other cellular processes, such as differentiation and migration, our study highlights their essential roles in the control of apoptosis in the fetus, and the G1-
S and G2-M transitions in the placenta.
169 6.4 Feedback and antagonism: crosstalks between repressors E2F7, E2F8 and activators E2F1, E2F3a
Disruption of repressor E2f7 and E2f8 leads to massive apoptosis in the fetus and over-proliferation in the placenta. Interestingly, concomitant loss of E2f activators, specifically E2f1 and E2f3a, suppresses these phenotypes, suggesting that these newly indentified E2F repressors are required to keep the
E2F family in balance by counteracting E2f1 and E2f3a activators function.
E2F7 and E2F8 control E2F1 activity by direct transcriptional repression of its expression. Interestingly, E2f1 is not just the downstream target of E2F7 and E2F8. Previous studies have shown that it also functions as an upstream regulator of E2f7 and E2f8 transcription (Di Stefano et al., 2003; Christensen et al., 2005). Therefore, the interactions between E2F1 and E2F7/E2F8 involves an elegant negative feedback loop, in which E2F1 activates E2f7 and E2f8 expression at the G1-S transition, and E2F7 and E2F8, in turn, represses E2f1 expression as cells transit through S phase and into G2 phase of the cell cycle
(Figure 6.2; Moon and Dyson, 2008).
Unlike E2f1, E2f3a is not a direct target of E2F7 and E2F8. Given the observations that E2F3a and E2F7/E2F8 proteins can occupy a common set of cell cycle gene promoters and have opposite effects on gene expression, we propose E2F3a and E2F7/E2F8 function in a parallel and antagonistic manner
(Figure 6.2). Depletion of the competing E2F7/E2F8 leaves the target gene promoter sites free for activation by the remaining E2F3a, and therefore tilts the balance of E2F activity toward activation.
170 In summary, we demonstrate here the complexity of crosstalks among the typical and atypical E2F family members: E2F1, E2F7/E2F8, and E2F1,
E2F7/E2F8, and highlight the significant functions of E2F7 and E2F8 in keeping the E2F activity in balance through either repressing activator mRNA expression or antagonizing activator protein functions. Clearly, with the identification of atypical E2Fs, our classical view of E2F activity needs to be refined to account for these newly identified E2Fs.
6.5 E2F7 and E2F8, possible tumor suppressors?
Uncontrolled cell proliferation and cell death is the hallmark of cancer.
Given the importance of E2F7 and E2F8 in the control of proliferation and apoptosis, it is very tempting to speculate that these newly identified E2F factors may function as putative tumor suppressors in a tumor setting. Consistent with this idea, a clinical study has associated low level of E2f7 expression with worse survival and potential development of resistance to anti-cancer drugs in ovarian cancer (Reimer et al., 2007). Interestingly, E2f7 is located at chromosome
12q21, the deletion of which has been associated with poor prognosis for pancreatic cancer (Kimura et al., 1998). Further studies with large sets of patients will be essential to confirm the predictive values of E2f7. In addition, whether expression of E2f8 may also reverse correlate with any tumor phenotypes remains to be determined. Additional evidence which may support the potential tumor suppressor function of E2F7 and E2F8 comes from the co- repressor study. We have shown that E2F7 and E2F8 can associate with Rb
171 family members in the same complex. Rb is the first tumor suppressor identified in humans. The idea that E2F7 and E2F8 might act as downstream mediators of
Rb tumor suppressor function is very intriguing.
Together, these preliminary data may suggest a potential role of E2F7 and
E2F8 as tumor suppressors, but confirmation of this hypothesis will require a thorough evaluation of their repressor functions in vivo. In this regard, E2f7 and
E2f8 conditional knockout mice would be very useful for further studies. We believe that investigating E2F7 and E2F8 function in later stages of development will provide us with more physiologically relevant information about their function in specialized tissues. With this knowledge and appropriate systems, it would be of great interest to further examine the potential suppressor function of E2F7 and
E2F8 in cancer contexts.
6.6 Concluding remarks
Using a combination of mouse genetics, bioinformatics, biochemistry, molecular biology and cell biology, this study clearly provides the first in vivo evidence for the physiological functions of E2F7 and E2F8, and successfully defined the critical cell lineages, the relevant cellular processes, the molecular pathways and the underlying mechanisms of E2F7 and E2F8 function. We conclude that E2F7/E2F8-mediated repression is essential for the control of apoptosis in the fetus and the regulation of cell cycle progression in the placenta.
172 These findings not only contribute greatly to our understanding of the ‘total E2F’ activity, but also suggest a potential tumor suppressor function of E2F7 and
E2F8 for the future study.
173 Placenta E2F3a E2F7/E2F8
G1-S & G2-M Fetus targets St E2F7/E2F8 ress
Proliferation E2f1
Dysfunction Apoptosis
Figure 6.1 Current model of E2F7 and E2F8 function during embryo development. E2F7 and E2F8 directly repress E2f1 expression in the fetus and therefore control apoptosis in a cell-autonomous manner. In the placenta,
E2F7 and E2F8 directly suppress a set of G1-S target gene expression and indirectly control G2-M gene expression. E2F3a-mediated activation counteracts E2F7/E2F8-mediated repression on the G1-S genes, but not on
G2-M targets. Upregulation of these target genes leads to the ectopic proliferation observed in spongiotrophoblast and giant cell lineages of the placenta, and consequentially placental dysfunction. Dysfunctional placenta elicits acute stress to the fetus, which contributes to fetal cell death in a non- cell autonomous manner.
174 Feedback Model
E2F1
E2F7 E2F8
Antagonism Model
E2F7 G1-S targets E2F3a E2F8
Figure 6.2 Models for the interactions between E2F1 and E2F7/E2F8, and
E2F3a and E2F7/E2F8. A feedback model for the interaction between E2F1 and E2F7/E2F8 (top), and an antagonism model for the interaction between
E2F3a and E2F7/E2F8 (bottom).
175
REFERENCES
Attwooll, C., Lazzerini Denchi, E., and Helin, K. (2004). The E2F family: specific functions and overlapping interests. EMBO J 23, 4709-4716.
Beijersbergen, R.L., and Bernards, R. (1996). Cell cycle regulation by the retinoblastoma family of growth inhibitory proteins. Biochim Biophys Acta 1287, 103-120.
Blow, J.J., and Laskey, R.A. (1988). A role for the nuclear envelope in controlling DNA replication within the cell cycle. Nature 332, 546-548.
Boudolf, V., Vlieghe, K., Beemster, G.T., Magyar, Z., Torres Acosta, J.A., Maes, S., Van Der Schueren, E., Inze, D., and De Veylder, L. (2004). The plant-specific cyclin-dependent kinase CDKB1;1 and transcription factor E2Fa-DPa control the balance of mitotically dividing and endoreduplicating cells in Arabidopsis. Plant Cell 16, 2683-2692.
Bracken, A.P., Ciro, M., Cocito, A., and Helin, K. (2004). E2F target genes: unraveling the biology. Trends Biochem Sci 29, 409-417.
Cam, H., and Dynlacht, B.D. (2003). Emerging roles for E2F: beyond the G1/S transition and DNA replication. Cancer Cell 3, 311-316.
Chinnadurai, G. (2002). CtBP, an unconventional transcriptional corepressor in development and oncogenesis. Mol Cell 9, 213-224.
Chong, J.L., Tsai, S.Y., Sharma, N., Opavsky, R., Price, R., Wu, L., Fernandez, S.A., and Leone, G. (2009). E2f3a and E2f3b contribute to the control of cell proliferation and mouse development. Mol Cell Biol 29, 414-424.
Christensen, J., Cloos, P., Toftegaard, U., Klinkenberg, D., Bracken, A.P., Trinh, E., Heeran, M., Di Stefano, L., and Helin, K. (2005). Characterization of E2F8, a novel E2F-like cell-cycle regulated repressor of E2F-activated transcription. Nucleic Acids Res 33, 5458-5470.
176 Christensen, R.N., Weinstein, M., and Tassava, R.A. (2002). Expression of fibroblast growth factors 4, 8, and 10 in limbs, flanks, and blastemas of Ambystoma. Dev Dyn 223, 193-203.
Cross, J.C. (2005). How to make a placenta: mechanisms of trophoblast cell differentiation in mice--a review. Placenta 26 Suppl A, S3-9.
Dannenberg, J.H., David, G., Zhong, S., van der Torre, J., Wong, W.H., and Depinho, R.A. (2005). mSin3A corepressor regulates diverse transcriptional networks governing normal and neoplastic growth and survival. Genes Dev 19, 1581-1595. de Bruin, A., Maiti, B., Jakoi, L., Timmers, C., Buerki, R., and Leone, G. (2003). Identification and characterization of E2F7, a novel mammalian E2F family member capable of blocking cellular proliferation. J Biol Chem 278, 42041- 42049. de Bruin, A., Wu, L., Saavedra, H.I., Wilson, P., Yang, Y., Rosol, T.J., Weinstein, M., Robinson, M.L., and Leone, G. (2003). Rb function in extraembryonic lineages suppresses apoptosis in the CNS of Rb-deficient mice. Proc Natl Acad Sci U S A 100, 6546-6551.
DeGregori, J., and Johnson, D.G. (2006). Distinct and Overlapping Roles for E2F Family Members in Transcription, Proliferation and Apoptosis. Curr Mol Med 6, 739-748.
DeGregori, J., Leone, G., Miron, A., Jakoi, L., and Nevins, J.R. (1997). Distinct roles for E2F proteins in cell growth control and apoptosis. Proc Natl Acad Sci U S A 94, 7245-7250.
Di Stefano, L., Jensen, M.R., and Helin, K. (2003). E2F7, a novel E2F featuring DP-independent repression of a subset of E2F-regulated genes. EMBO J 22, 6289-6298.
Dimova, D.K., and Dyson, N.J. (2005). The E2F transcriptional network: old acquaintances with new faces. Oncogene 24, 2810-2826.
Dynlacht, B.D., Flores, O., Lees, J.A., and Harlow, E. (1994). Differential regulation of E2F transactivation by cyclin/cdk2 complexes. Genes Dev 8, 1772- 1786.
Edgar, B.A., and Orr-Weaver, T.L. (2001). Endoreplication cell cycles: more for less. Cell 105, 297-306.
177 Field, S.J., Tsai, F.Y., Kuo, F., Zubiaga, A.M., Kaelin, W.G., Jr., Livingston, D.M., Orkin, S.H., and Greenberg, M.E. (1996). E2F-1 functions in mice to promote apoptosis and suppress proliferation. Cell 85, 549-561.
Frolov, M.V., and Dyson, N.J. (2004). Molecular mechanisms of E2F-dependent activation and pRB-mediated repression. J Cell Sci 117, 2173-2181.
Fulop, K., Tarayre, S., Kelemen, Z., Horvath, G., Kevei, Z., Nikovics, K., Bako, L., Brown, S., Kondorosi, A., and Kondorosi, E. (2005). Arabidopsis anaphase- promoting complexes: multiple activators and wide range of substrates might keep APC perpetually busy. Cell Cycle 4, 1084-1092.
Gaubatz, S., Lindeman, G.J., Ishida, S., Jakoi, L., Nevins, J.R., Livingston, D.M., and Rempel, R.E. (2000). E2F4 and E2F5 play an essential role in pocket protein-mediated G1 control. Mol Cell 6, 729-735.
Geng, Y., Yu, Q., Sicinska, E., Das, M., Schneider, J.E., Bhattacharya, S., Rideout, W.M., Bronson, R.T., Gardner, H., and Sicinski, P. (2003). Cyclin E ablation in the mouse. Cell 114, 431-443.
Giangrande, P.H., Zhu, W., Schlisio, S., Sun, X., Mori, S., Gaubatz, S., and Nevins, J.R. (2004). A role for E2F6 in distinguishing G1/S- and G2/M-specific transcription. Genes Dev 18, 2941-2951.
Glozak, M.A., and Seto, E. (2007). Histone deacetylases and cancer. Oncogene 26, 5420-5432.
Hayashi, S., Lewis, P., Pevny, L., and McMahon, A.P. (2002). Efficient gene modulation in mouse epiblast using a Sox2Cre transgenic mouse strain. Mech Dev 119 Suppl 1, S97-S101.
Hildebrand, J.D., and Soriano, P. (2002). Overlapping and unique roles for C- terminal binding protein 1 (CtBP1) and CtBP2 during mouse development. Mol Cell Biol 22, 5296-5307.
Hsieh, J.K., Yap, D., O'Connor, D.J., Fogal, V., Fallis, L., Chan, F., Zhong, S., and Lu, X. (2002). Novel function of the cyclin A binding site of E2F in regulating p53-induced apoptosis in response to DNA damage. Mol Cell Biol 22, 78-93.
Hsu, J.Y., Reimann, J.D., Sorensen, C.S., Lukas, J., and Jackson, P.K. (2002). E2F-dependent accumulation of hEmi1 regulates S phase entry by inhibiting APC(Cdh1). Nat Cell Biol 4, 358-366.
178 Humbert, P.O., Rogers, C., Ganiatsas, S., Landsberg, R.L., Trimarchi, J.M., Dandapani, S., Brugnara, C., Erdman, S., Schrenzel, M., Bronson, R.T., et al. (2000). E2F4 is essential for normal erythrocyte maturation and neonatal viability. Mol Cell 6, 281-291.
Humbert, P.O., Verona, R., Trimarchi, J.M., Rogers, C., Dandapani, S., and Lees, J.A. (2000). E2f3 is critical for normal cellular proliferation. Genes Dev 14, 690-703.
Jackson, C.J., Dixon-Woods, M., Hsu, R., and Kurinczuk, J.J. (2005). A qualitative study of choosing and using an NHS Walk-in Centre. Fam Pract 22, 269-274.
Kimura, M., Furukawa, T., Abe, T., Yatsuoka, T., Youssef, E.M., Yokoyama, T., Ouyang, H., Ohnishi, Y., Sunamura, M., Kobari, M., et al. (1998). Identification of two common regions of allelic loss in chromosome arm 12q in human pancreatic cancer. Cancer Res 58, 2456-2460.
Kohn, M.J., Bronson, R.T., Harlow, E., Dyson, N.J., and Yamasaki, L. (2003). Dp1 is required for extra-embryonic development. Development 130, 1295-1305.
Kosugi, S., and Ohashi, Y. (2002). E2Ls, E2F-like repressors of Arabidopsis that bind to E2F sites in a monomeric form. J Biol Chem 277, 16553-16558.
Kozar, K., Ciemerych, M.A., Rebel, V.I., Shigematsu, H., Zagozdzon, A., Sicinska, E., Geng, Y., Yu, Q., Bhattacharya, S., Bronson, R.T., et al. (2004). Mouse development and cell proliferation in the absence of D-cyclins. Cell 118, 477-491.
Kramer, E.R., Scheuringer, N., Podtelejnikov, A.V., Mann, M., and Peters, J.M. (2000). Mitotic regulation of the APC activator proteins CDC20 and CDH1. Mol Biol Cell 11, 1555-1569.
La Thangue, N.B. (2002). Transcription. Chromatin control--a place for E2F and Myc to meet. Science 296, 1034-1035.
Lagger, G., O'Carroll, D., Rembold, M., Khier, H., Tischler, J., Weitzer, G., Schuettengruber, B., Hauser, C., Brunmeir, R., Jenuwein, T., et al. (2002). Essential function of histone deacetylase 1 in proliferation control and CDK inhibitor repression. EMBO J 21, 2672-2681.
Lakso, M., Pichel, J.G., Gorman, J.R., Sauer, B., Okamoto, Y., Lee, E., Alt, F.W., and Westphal, H. (1996). Efficient in vivo manipulation of mouse genomic sequences at the zygote stage. Proc Natl Acad Sci U S A 93, 5860-5865.
179 Lammens, T., Boudolf, V., Kheibarshekan, L., Zalmas, L.P., Gaamouche, T., Maes, S., Vanstraelen, M., Kondorosi, E., La Thangue, N.B., Govaerts, W., et al. (2008). Atypical E2F activity restrains APC/CCCS52A2 function obligatory for endocycle onset. Proc Natl Acad Sci U S A 105, 14721-14726.
Leone, G., DeGregori, J., Yan, Z., Jakoi, L., Ishida, S., Williams, R.S., and Nevins, J.R. (1998). E2F3 activity is regulated during the cell cycle and is required for the induction of S phase. Genes Dev 12, 2120-2130.
Li, J., Ran, C., Li, E., Gordon, F., Comstock, G., Siddiqui, H., Cleghorn, W., Chen, H.Z., Kornacker, K., Liu, C.G., et al. (2008). Synergistic function of E2F7 and E2F8 is essential for cell survival and embryonic development. Dev Cell 14, 62-75.
Lilly, M.A., and Duronio, R.J. (2005). New insights into cell cycle control from the Drosophila endocycle. Oncogene 24, 2765-2775.
Lin, W.C., Lin, F.T., and Nevins, J.R. (2001). Selective induction of E2F1 in response to DNA damage, mediated by ATM-dependent phosphorylation. Genes Dev 15, 1833-1844.
Lindeman, G.J., Dagnino, L., Gaubatz, S., Xu, Y., Bronson, R.T., Warren, H.B., and Livingston, D.M. (1998). A specific, nonproliferative role for E2F-5 in choroid plexus function revealed by gene targeting. Genes Dev 12, 1092-1098.
Logan, N., Delavaine, L., Graham, A., Reilly, C., Wilson, J., Brummelkamp, T.R., Hijmans, E.M., Bernards, R., and La Thangue, N.B. (2004). E2F-7: a distinctive E2F family member with an unusual organization of DNA-binding domains. Oncogene 23, 5138-5150.
Logan, N., Graham, A., Zhao, X., Fisher, R., Maiti, B., Leone, G., and La Thangue, N.B. (2005). E2F-8: an E2F family member with a similar organization of DNA-binding domains to E2F-7. Oncogene 24, 5000-5004.
Lukas, C., Sorensen, C.S., Kramer, E., Santoni-Rugiu, E., Lindeneg, C., Peters, J.M., Bartek, J., and Lukas, J. (1999). Accumulation of cyclin B1 requires E2F and cyclin-A-dependent rearrangement of the anaphase-promoting complex. Nature 401, 815-818.
Luo, R.X., Postigo, A.A., and Dean, D.C. (1998). Rb interacts with histone deacetylase to repress transcription. Cell 92, 463-473.
MacAuley, A., Cross, J.C., and Werb, Z. (1998). Reprogramming the cell cycle for endoreduplication in rodent trophoblast cells. Mol Biol Cell 9, 795-807.
180 Maiti, B., Li, J., de Bruin, A., Gordon, F., Timmers, C., Opavsky, R., Patil, K., Tuttle, J., Cleghorn, W., and Leone, G. (2005). Cloning and characterization of mouse E2F8, a novel mammalian E2F family member capable of blocking cellular proliferation. J Biol Chem 280, 18211-18220.
Malumbres, M., and Barbacid, M. (2009). Cell cycle, CDKs and cancer: a changing paradigm. Nat Rev Cancer 9, 153-166.
Malumbres, M., Sotillo, R., Santamaria, D., Galan, J., Cerezo, A., Ortega, S., Dubus, P., and Barbacid, M. (2004). Mammalian cells cycle without the D-type cyclin-dependent kinases Cdk4 and Cdk6. Cell 118, 493-504.
Mariconti, L., Pellegrini, B., Cantoni, R., Stevens, R., Bergounioux, C., Cella, R., and Albani, D. (2002). The E2F family of transcription factors from Arabidopsis thaliana. Novel and conserved components of the retinoblastoma/E2F pathway in plants. J Biol Chem 277, 9911-9919.
Marti, A., Wirbelauer, C., Scheffner, M., and Krek, W. (1999). Interaction between ubiquitin-protein ligase SCFSKP2 and E2F-1 underlies the regulation of E2F-1 degradation. Nat Cell Biol 1, 14-19.
Mendez, J., and Stillman, B. (2003). Perpetuating the double helix: molecular machines at eukaryotic DNA replication origins. Bioessays 25, 1158-1167.
Montgomery, R.L., Potthoff, M.J., Haberland, M., Qi, X., Matsuzaki, S., Humphries, K.M., Richardson, J.A., Bassel-Duby, R., and Olson, E.N. (2008). Maintenance of cardiac energy metabolism by histone deacetylase 3 in mice. J Clin Invest 118, 3588-3597.
Moon, N.S., and Dyson, N. (2008). E2F7 and E2F8 keep the E2F family in balance. Dev Cell 14, 1-3.
Muller, H., and Helin, K. (2000). The E2F transcription factors: key regulators of cell proliferation. Biochim Biophys Acta 1470, M1-12.
Murga, M., Fernandez-Capetillo, O., Field, S.J., Moreno, B., Borlado, L.R., Fujiwara, Y., Balomenos, D., Vicario, A., Carrera, A.C., Orkin, S.H., et al. (2001). Mutation of E2F2 in mice causes enhanced T lymphocyte proliferation, leading to the development of autoimmunity. Immunity 15, 959-970.
Murray, A.W. (2004). Recycling the cell cycle: cyclins revisited. Cell 116, 221- 234.
181 Nielsen, S.J., Schneider, R., Bauer, U.M., Bannister, A.J., Morrison, A., O'Carroll, D., Firestein, R., Cleary, M., Jenuwein, T., Herrera, R.E., et al. (2001). Rb targets histone H3 methylation and HP1 to promoters. Nature 412, 561-565.
Nurse, P. (1990). Universal control mechanism regulating onset of M-phase. Nature 344, 503-508.
O'Donnell, K.A., Wentzel, E.A., Zeller, K.I., Dang, C.V., and Mendell, J.T. (2005). c-Myc-regulated microRNAs modulate E2F1 expression. Nature 435, 839-843.
Ogawa, H., Ishiguro, K., Gaubatz, S., Livingston, D.M., and Nakatani, Y. (2002). A complex with chromatin modifiers that occupies E2F- and Myc-responsive genes in G0 cells. Science 296, 1132-1136.
Oldham, S., Bohni, R., Stocker, H., Brogiolo, W., and Hafen, E. (2000). Genetic control of size in Drosophila. Philos Trans R Soc Lond B Biol Sci 355, 945-952.
Parisi, T., Beck, A.R., Rougier, N., McNeil, T., Lucian, L., Werb, Z., and Amati, B. (2003). Cyclins E1 and E2 are required for endoreplication in placental trophoblast giant cells. EMBO J 22, 4794-4803.
Pina, C., Pinto, F., Feijo, J.A., and Becker, J.D. (2005). Gene family analysis of the Arabidopsis pollen transcriptome reveals biological implications for cell growth, division control, and gene expression regulation. Plant Physiol 138, 744- 756.
Pines, J., and Hunter, T. (1991). Human cell division: the involvement of cyclins A and B1, and multiple cdc2s. Cold Spring Harb Symp Quant Biol 56, 449-463.
Pomerantz, J., Schreiber-Agus, N., Liegeois, N.J., Silverman, A., Alland, L., Chin, L., Potes, J., Chen, K., Orlow, I., Lee, H.W., et al. (1998). The Ink4a tumor suppressor gene product, p19Arf, interacts with MDM2 and neutralizes MDM2's inhibition of p53. Cell 92, 713-723.
Ramirez-Parra, E., Lopez-Matas, M.A., Frundt, C., and Gutierrez, C. (2004). Role of an atypical E2F transcription factor in the control of Arabidopsis cell growth and differentiation. Plant Cell 16, 2350-2363.
Reed, B.H., and Orr-Weaver, T.L. (1997). The Drosophila gene morula inhibits mitotic functions in the endo cell cycle and the mitotic cell cycle. Development 124, 3543-3553.
182 Reimer, D., Sadr, S., Wiedemair, A., Stadlmann, S., Concin, N., Hofstetter, G., Muller-Holzner, E., Marth, C., and Zeimet, A.G. (2007). Clinical relevance of E2F family members in ovarian cancer--an evaluation in a training set of 77 patients. Clin Cancer Res 13, 144-151.
Reis, T., and Edgar, B.A. (2004). Negative regulation of dE2F1 by cyclin- dependent kinases controls cell cycle timing. Cell 117, 253-264.
Rempel, R.E., Saenz-Robles, M.T., Storms, R., Morham, S., Ishida, S., Engel, A., Jakoi, L., Melhem, M.F., Pipas, J.M., Smith, C., et al. (2000). Loss of E2F4 activity leads to abnormal development of multiple cellular lineages. Mol Cell 6, 293-306.
Resnitzky, D., Gossen, M., Bujard, H., and Reed, S.I. (1994). Acceleration of the G1/S phase transition by expression of cyclins D1 and E with an inducible system. Mol Cell Biol 14, 1669-1679.
Riddle, R.D., Johnson, R.L., Laufer, E., and Tabin, C. (1993). Sonic hedgehog mediates the polarizing activity of the ZPA. Cell 75, 1401-1416.
Rogoff, H.A., Pickering, M.T., Debatis, M.E., Jones, S., and Kowalik, T.F. (2002). E2F1 induces phosphorylation of p53 that is coincident with p53 accumulation and apoptosis. Mol Cell Biol 22, 5308-5318.
Rossant, J., and Cross, J.C. (2001). Placental development: lessons from mouse mutants. Nat Rev Genet 2, 538-548.
Rowland, B.D., and Bernards, R. (2006). Re-evaluating cell-cycle regulation by E2Fs. Cell 127, 871-874.
Russell, J.L., Powers, J.T., Rounbehler, R.J., Rogers, P.M., Conti, C.J., and Johnson, D.G. (2002). ARF differentially modulates apoptosis induced by E2F1 and Myc. Mol Cell Biol 22, 1360-1368.
Saavedra, H.I., Wu, L., de Bruin, A., Timmers, C., Rosol, T.J., Weinstein, M., Robinson, M.L., and Leone, G. (2002). Specificity of E2F1, E2F2, and E2F3 in mediating phenotypes induced by loss of Rb. Cell Growth Differ 13, 215-225.
Sanchez, I., and Dynlacht, B.D. (2005). New insights into cyclins, CDKs, and cell cycle control. Semin Cell Dev Biol 16, 311-321.
Seville, L.L., Shah, N., Westwell, A.D., and Chan, W.C. (2005). Modulation of pRB/E2F functions in the regulation of cell cycle and in cancer. Curr Cancer Drug Targets 5, 159-170.
183 Silverstein, R.A., and Ekwall, K. (2005). Sin3: a flexible regulator of global gene expression and genome stability. Curr Genet 47, 1-17.
Soriano, P. (1999). Generalized lacZ expression with the ROSA26 Cre reporter strain. Nat Genet 21, 70-71.
Stanelle, J., and Putzer, B.M. (2006). E2F1-induced apoptosis: turning killers into therapeutics. Trends Mol Med 12, 177-185.
Stevens, C., and La Thangue, N.B. (2003). E2F and cell cycle control: a double- edged sword. Arch Biochem Biophys 412, 157-169.
Stevens, C., and La Thangue, N.B. (2004). The emerging role of E2F-1 in the DNA damage response and checkpoint control. DNA Repair (Amst) 3, 1071- 1079.
Storre, J., Elsasser, H.P., Fuchs, M., Ullmann, D., Livingston, D.M., and Gaubatz, S. (2002). Homeotic transformations of the axial skeleton that accompany a targeted deletion of E2f6. EMBO Rep 3, 695-700.
Sugimoto-Shirasu, K., and Roberts, K. (2003). "Big it up": endoreduplication and cell-size control in plants. Curr Opin Plant Biol 6, 544-553.
Takahashi, T.S., Wigley, D.B., and Walter, J.C. (2005). Pumps, paradoxes and ploughshares: mechanism of the MCM2-7 DNA helicase. Trends Biochem Sci 30, 437-444.
Takahashi, Y., Rayman, J.B., and Dynlacht, B.D. (2000). Analysis of promoter binding by the E2F and pRB families in vivo: distinct E2F proteins mediate activation and repression. Genes Dev 14, 804-816.
Tarayre, S., Vinardell, J.M., Cebolla, A., Kondorosi, A., and Kondorosi, E. (2004). Two classes of the CDh1-type activators of the anaphase-promoting complex in plants: novel functional domains and distinct regulation. Plant Cell 16, 422-434.
Trimarchi, J.M., and Lees, J.A. (2002). Sibling rivalry in the E2F family. Nat Rev Mol Cell Biol 3, 11-20. van den Heuvel, S., and Dyson, N.J. (2008). Conserved functions of the pRB and E2F families. Nat Rev Mol Cell Biol 9, 713-724.
Vandel, L., Nicolas, E., Vaute, O., Ferreira, R., Ait-Si-Ali, S., and Trouche, D. (2001). Transcriptional repression by the retinoblastoma protein through the recruitment of a histone methyltransferase. Mol Cell Biol 21, 6484-6494.
184 Vandepoele, K., Raes, J., De Veylder, L., Rouze, P., Rombauts, S., and Inze, D. (2002). Genome-wide analysis of core cell cycle genes in Arabidopsis. Plant Cell 14, 903-916.
Vandepoele, K., Vlieghe, K., Florquin, K., Hennig, L., Beemster, G.T., Gruissem, W., Van de Peer, Y., Inze, D., and De Veylder, L. (2005). Genome-wide identification of potential plant E2F target genes. Plant Physiol 139, 316-328.
Varmuza, S., Prideaux, V., Kothary, R., and Rossant, J. (1988). Polytene chromosomes in mouse trophoblast giant cells. Development 102, 127-134.
Verkest, A., Manes, C.L., Vercruysse, S., Maes, S., Van Der Schueren, E., Beeckman, T., Genschik, P., Kuiper, M., Inze, D., and De Veylder, L. (2005). The cyclin-dependent kinase inhibitor KRP2 controls the onset of the endoreduplication cycle during Arabidopsis leaf development through inhibition of mitotic CDKA;1 kinase complexes. Plant Cell 17, 1723-1736.
Vlieghe, K., Boudolf, V., Beemster, G.T., Maes, S., Magyar, Z., Atanassova, A., de Almeida Engler, J., De Groodt, R., Inze, D., and De Veylder, L. (2005). The DP-E2F-like gene DEL1 controls the endocycle in Arabidopsis thaliana. Curr Biol 15, 59-63.
Watson, E.D., and Cross, J.C. (2005). Development of structures and transport functions in the mouse placenta. Physiology (Bethesda) 20, 180-193.
Wenzel, P.L., and Leone, G. (2007). Expression of Cre recombinase in early diploid trophoblast cells of the mouse placenta. Genesis 45, 129-134.
Wenzel, P.L., Wu, L., de Bruin, A., Chong, J.L., Chen, W.Y., Dureska, G., Sites, E., Pan, T., Sharma, A., Huang, K., et al. (2007). Rb is critical in a mammalian tissue stem cell population. Genes Dev 21, 85-97.
Wu, L., de Bruin, A., Saavedra, H.I., Starovic, M., Trimboli, A., Yang, Y., Opavska, J., Wilson, P., Thompson, J.C., Ostrowski, M.C., et al. (2003). Extra- embryonic function of Rb is essential for embryonic development and viability. Nature 421, 942-947.
Wu, L., Timmers, C., Maiti, B., Saavedra, H.I., Sang, L., Chong, G.T., Nuckolls, F., Giangrande, P., Wright, F.A., Field, S.J., et al. (2001). The E2F1-3 transcription factors are essential for cellular proliferation. Nature 414, 457-462.
Yam, C.H., Fung, T.K., and Poon, R.Y. (2002). Cyclin A in cell cycle control and cancer. Cell Mol Life Sci 59, 1317-1326.
185 Yamasaki, L., Jacks, T., Bronson, R., Goillot, E., Harlow, E., and Dyson, N.J. (1996). Tumor induction and tissue atrophy in mice lacking E2F-1. Cell 85, 537- 548.
Yu, Y., Steinmetz, A., Meyer, D., Brown, S., and Shen, W.H. (2003). The tobacco A-type cyclin, Nicta;CYCA3;2, at the nexus of cell division and differentiation. Plant Cell 15, 2763-2777.
Yushkevich, P.A., Piven, J., Hazlett, H.C., Smith, R.G., Ho, S., Gee, J.C., and Gerig, G. (2006). User-guided 3D active contour segmentation of anatomical structures: significantly improved efficiency and reliability. Neuroimage 31, 1116- 1128.
Zalmas, L.P., Zhao, X., Graham, A.L., Fisher, R., Reilly, C., Coutts, A.S., and La Thangue, N.B. (2008). DNA-damage response control of E2F7 and E2F8. EMBO Rep 9, 252-259.
Zhang, P., Wong, C., DePinho, R.A., Harper, J.W., and Elledge, S.J. (1998). Cooperation between the Cdk inhibitors p27(KIP1) and p57(KIP2) in the control of tissue growth and development. Genes Dev 12, 3162-3167.
Zhu, S.C., and Yuille, A. (1996). Region competition: unifying snakes, region growing, and bayes/mdl for multiband image segmentation. IEEE Trans Pattern Anal 18, 884-900.
Zhu, W., Giangrande, P.H., and Nevins, J.R. (2004). E2Fs link the control of G1/S and G2/M transcription. EMBO J 23, 4615-4626.
Ziebold, U., Reza, T., Caron, A., and Lees, J.A. (2001). E2F3 contributes both to the inappropriate proliferation and to the apoptosis arising in Rb mutant embryos. Genes Dev 15, 386-391.
186