University of Southern California Dissertations and Theses

The mechanisms of somatic cell reprogramming

(USC Thesis Other)

The mechanisms of somatic cell reprogramming

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
THE MOLECULAR MECHANISMS OF SOMATIC CELL REPROGRAMMING

by

Zong Wei

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(GENETIC, MOLECULAR AND CELLULAR BIOLOGY)

August 2012

Copyright 2012 Zong Wei

ii

Acknowledgements
First I would like to thank my mentor Dr. Wange Lu, for his continuous support and
encouragement. It would not be possible for me to start from scratch in a new and
competitive field without the guidance from him. His philosophy about scientific research
and enthusiasm in exploring unknown fields has greatly inspired and influenced me.
Besides experimental techniques, I am also lucky to learn the art of designing, executing
and managing projects in daily lab life. Particularly, I am grateful for his generosity in
offering me tremendous freedom for conducting my research, a challenge but also a
privilege.
I would also like to acknowledge many colleagues for their contribution and help. Rose
Andrianakos and Yang Yang contributed to the molecular experiments regarding Klf4’s
interactions. Fan Gao provided essential data analysis for 4C sequencing. I have learned a
lot from senior colleagues in the Lu Lab as well as our collaborators, including Peilin
Zhang, Jungmook Lyu, Kouichi Hasegawa, and SiHo Choi. Xi Chen, Chunming Liu,
Sewoon Kim and Chenchen Yang generously provided me critical reagents. I have also
benefited hugely from daily discussion with people in the lab, especially with Jingyang
Zhong, Wen-Hsuan Chang and Hongzhen Yang.
I am particularly lucky to have great guidance committees which provided critical
comments and suggestions. I would like to thank the members in the committees
including Dr. Michael Stallcup, Dr. Gerald Coetzee, Dr. Neil Segil, Dr. Martin Pera, and
Dr. Gregor Adams.
iii

Lastly and most importantly, I would like to offer my most sincere thanks to my parents,
Guangqun Wei and Ling Xu, and the love of my life, Haoyi Wang. Numerous times their
unconditional love and support lead me out of darkness. It is my greatest luck and fortune
to always have their blessings with me.

iv

Table of Contents
Acknowledgements ii
List of Tables v
List of Figures vi
Abstract viii
Chapter 1 Induced pluripotency and nuclear architecture dynamics 1
1.1 Early studies of nuclear reprogramming 1
1.2 Discovery of embryonic stem cells and its application 3
1.3 Discovery of induced pluripotent stem cells (iPSCs) 7
1.4 Application of iPSCs in disease modeling and regenerative
medicine 13
1.5 Molecular mechanisms of reprogramming from somatic cells into
iPSCs 20
1.6 Discovery of long-range interactions via chromosome conformation
capture and fluorescence in situ hybridization (FISH) 26
1.7 Higher order chromatin structure and transcription factory 31
1.8 Current understanding of the regulatory mechanisms of nuclear
organization 34
Chapter 2 Klf4 directly interacts with Oct4 and Sox2 to promote
reprogramming 39
2.1 Abstract 39
2.2 Introduction 39
2.3 Material and methods 41
2.4 Results 46
2.5 Discussion 64

Chapter 3 Interchromosomal interactions modulate Oct4 expression
in reprogramming and pluripotency 67
3.1 Abstract 67
3.2 Introduction 68
3.3 Material and methods 70
3.4 Results and discussion 77

Chapter 4 Future perspectives on induced pluripotency and nuclear
architecture 102

Bibliography 106
v

List of Tables
Table 1 Summary of different cell types for reprogramming 9
Table 2 Summary of patient specific iPSCs and phenotype observed 15
in previous studies.
Table 3 List of BACs used in this study 100
Table 4 List of primers used in this study 101

vi

List of Figures
Figure 1 Summary of 3C based methods 26

Figure 2 Establishment of inducible reprogramming system using 48
defined factors

Figure 3 Characterization of first generation iPS cells 50

Figure 4 Generation of a homogeneous population of NSCs and
2
nd
generation iPS cells 52

Figure 5 Klf4 interacts with Oct4 and Sox2 54

Figure 6 The Klf4 C-terminus directly interacts with Oct4 and Sox2 56

Figure 7 Dominant negative Klf4 mutants compete with wild-type Klf4
for binding with Oct4 and Sox2, resulting in disruption of
reprogramming 58

Figure 8 Transduction of neural stem cells with lentiviruses
expressing Klf4 mutants 61

Figure 9 Disrupting endogenous O/S/K complex by introducing
Klf4 dominant negative mutant results in failure of reprogramming
and differentiation in wild type ES cells 63

Figure 10 Schematic show of 4C assay and bait region 78

Figure 11 Long-range interactions between Oct4 and various loci reveal unique
higher-order chromatin structure of the Oct4 locus in PSCs 79

Figure 12 NSCs from different origins are similar in co-localization
efficiency 81

Figure 13 Sizes of nuclei measured by DAPI staining in NSCs, iPSCs, ESCs,
and pre-iPSCs 81

Figure 14 Characterization of the secondary reprogramming system 82

Figure 15 PSC-specific interchromosomal interactions of Oct4 loci are
established early in reprogramming 84

vii

Figure 16 Interchromosomal interactions are correlated with endogenous Oct4
transcription in PSCs 86

Figure 17 Detection of Oct4 primary transcript by RNA FISH in ESCs 88

Figure 18 Enrichment of RNAPII-S5P in Oct4-474J5 and Oct4-280I15
interacting sites 88

Figure 19 Enrichment in pluripotency factor binding sites in 4C positive
LID regions 89

Figure 20 Characterization of the potential Klf4 target gene 89

Figure 21 Characterization of the role of Llgl2, a gene located in 4C
enriched region, in PSCs self-renewal 90

Figure 22 Characterization of the role of Grb7, a gene located in 4C
enriched region, in PSCs self-renewa 90

Figure 23 Characterization of role of Unc84a in PSCs differentiation
potential 91

Figure 24 Immuno-staining indicates that Klf4 protein (green) form
punctuated structure which overlay with RNAPII-S5P (red)
inside nucleus 92

Figure 25 Klf4 is essential for organizing the interchromosomal interactions
of the Oct4 locus 94

Figure 26 Klf4 protein decrease significantly after 2 days of knockdown 96

Figure 27 Characterizing the effect of inducible knockdown of Klf4 to PSCs
self-renewal 97

Figure 28 Inhibition of RNA polymerase II transcription by α-amanitin
treatment does not change interchromosomal interactions
frequencies 98

Figure 29 Generation of inducible Klf4 overexpression stable cell lines
in mouse ES cells 99

viii

Abstract

The discovery of induced pluripotent stem cells (iPSCs) has transformed the research of
stem cells and provided infinite possibilities in regenerative medicine. In classical
Yamanaka protocol, somatic cells from various sources can be reprogrammed to iPSCs
with forced expression of Oct4, Sox2, Klf4, and cMyc. Numerous other combinations of
factors and various delivery methods have also been developed to optimize the efficiency
and accustom to different applications. These iPSCs are useful for disease modeling,
toxicology studies and cell therapy. However, the molecular mechanisms of this
transformation remain largely unclear. The transition from somatic cells to iPSCs
involved comprehensive changes on epigenetic level of the cells induced by
reprogramming factors. Reprogramming factor Klf4 can physically interact with Oct4
and Sox2. These three transcription factors co-occupy promoters of many pluripotency
related genes, such as endogenous Nanog and Oct4. The physical interactions depend on
the C2H2 zinc fingers in Klf4. Abrogation of these interactions will lead to failure of
reprogramming due to the inability of defective complexes in activating key downstream
genes. These results suggest that direct interactions between reprogramming factors are
essential for initiating key downstream genes. During reprogramming, nuclear
architecture of the cells also experience dramatic changes. In pluripotent stem cells
(PSCs), endogenous Oct4 loci interact with distant regions in cis and in trans. Many of
these long range interactions are specific to PSCs. PSC-specific interchromosomal
interactions are established prior to transcriptional activation of endogenous Oct4 during
ix

reprogramming. In PSCs, Oct4-colocalized domains are enriched in active genes and
pluripotency factor binding. Transcription of Oct4 is facilitated when the Oct4 locus is
co-localized with its interchromosomal partners. Finally, depletion or overexpression of
Klf4 causes changes in interchromosomal interactions prior to loss of Oct4 transcription
and PSC differentiation, suggesting that Klf4 regulates interchromosomal interactions
independent of its role as a transcription factor. Together these results reveal two novel
essential factors in facilitating reprogramming: physical interactions between
reprogramming factors and nuclear architecture dynamics.

1

Chapter 1 Induced pluripotency and nuclear architecture dynamics

1.1 Early studies of nuclear reprogramming
The development of organisms is usually described as a unidirectional process. In mammals and
other vertebrates, the zygote represents the totipotent state, which is defined as the ability to give
rise into essentially every lineage of the cells of the organism, including embryonic and
extraembryonic tissues (Kelly, 1977). During the development, cells gradually lose their
differentiation potential. In mammals, the next stage following totipotency is usually referred as
pluripotency. The cells in the inner cell mass (ICM) are considered pluripotent, as they have the
potential to differentiate into every linage except extraembryonic tissues. Later on in
development, cells lose their pluripotency and derive into multipotent progenitors, which are able
to give rise to various cell types, but usually restricted within one lineage. The well-characterized
examples of multipotent progenitors in mammals include hematopoietic stem cells, neural
stem/progenitor cells and skin stem cells. These multipotent progenitors further become
unipotent cells, which are only able to commit to one specific type of cells.
In 1950s, the technique of ‘cloning’ developed to isolate nuclei from late-stage embryos of
tadpoles and transplant them into enucleated oocytes. This leads to the groundbreaking discovery
by Gurdon in 1960s and 1970s that the nuclei from differentiated cells have the potential to
generate cloned frogs (Gurdon, 1962), which proved that the genetic information in somatic cells
is sufficient to generate the whole animal. The experiment supported the ‘epigenetic landscape’
hypothesis from Waddington (Waddington, 1957), in which he use a marble rolling down a hill
and ‘trapped’ in a valley to illustrate the increasing epigenetic restriction during the development.
In 1997, Wilmut and colleagues generate the first cloned mammal, Dolly the sheep (Wilmut et
2

al., 1997). Following experiments using terminally differentiated cells such as lymphoid cells
(Hochedlinger and Jaenisch, 2002; Inoue et al., 2005) and postmitotic neurons (Eggan et al.,
2004; Li et al., 2004) demonstrated that the nuclei of terminally differentiated cells still have the
‘totipotent potential’. However, generation of cloned animal from terminal differentiated cells is
inefficient. This indicates that the ‘totipotent potential’ of nuclei is decreased in differentiated
cells. One example is that direct cloning from skin stem cells are more efficient than that from
transient amplifying keratinocytes derived from them (Li et al., 2007). However, systematic
studies using donor cells from different differentiation stages for direct cloning is lacking, and
controversial results have been presented in hematopoietic stem cell lineages (Sung et al., 2006),
where evidence suggested that terminal differentiated cells may be more efficient donors than
progenitor cells in nuclear transfer.
Besides producing whole animals, nuclear transfer can also be used to generate embryonic stem
cells (ESCs). In fact, to more efficiently generate whole animals, usually ESCs are first
generated and then injected into the blastocysts to produce whole animals (Jaenisch and Young,
2008). Thus the concept of ‘therapeutic reprogramming’ has been proposed. Patient specific
tissue can be used to generate customized ESCs and used for disease studies. However, early
works suggested that only oocytes but not fertilized eggs can be used to generate cloned animals
(McGrath and Solter, 1984). The difficulty in obtaining human oocytes without compensation
had become a significant hurdle for therapeutic application (Egli et al., 2011). Although the idea
of therapeutic reprogramming has been proposed for more than 10 years, only recently has it
been shown to work in human (Noggle et al., 2011). It is worth noting that even in this seminal
experiment, ESCs can only be obtained when the oocyte genome is not removed (Noggle et al.,
2011). This interesting finding indicates that removal of oocyte genome cause developmental
3

failure after genome exchange. The triploid ESCs produced in this study, though useful for
mechanism studies, will not be suitable for clinical application.
Besides nuclear transfer, there are other ways of reprogramming nuclei to pluripotency state. The
best characterized way is to fuse the embryonic cell with somatic cells and produce hybrids that
the donor nuclei will be reprogrammed to a pluripotent state. This has been shown using
embryonic carcinoma cells, embryonic germ cells and ESCs in mouse (Solter, 2006; Zwaka and
Thomson, 2005) and human (Cowan et al., 2005; Yu et al., 2006). The pluripotent marker such
as Oct4 as well as inactive X chromosomes can be reactivated after fusion (Yu et al., 2006).
However, this produces tetraploid cells and again will not be suitable for clinical application.
There has been consistent attempts to fuse the somatic cells only with nuclear extract by
incubating permeabilized cells with ESCs extracts. Although it has been reported that genome-
wide reprogramming can be observed using this method (Taranger et al., 2005), convincing
confirmation of this phenomenon has never been reported so far.
1.2 Discovery of embryonic stem cells and its application
At E3.5, the blastocysts of mice can be divided into 3 different linages. The outer layer is
trophectoderm, which encloses the cavity and inner cell mass (ICM). The ICM can be further
divided into primitive ectorderm, or epiblast, and primitive endoderm, or hypoblast.
Trophectoderm and primitive endoderm will eventually give rise to extraembryonic tissues
including placenta and yolk sacs. The primitive ectorderm will give rise to the entire fetus
(Rossant, 2008). Embryonic stem cells (ESCs) were first isolated by Evans and Kaufman (Evans
and Kaufman, 1981) isolating the inner cell mass of the embryos and explanting them in the petri
dishes plated with mitomycin C-inhibited STO cells. These ESCs are equivalent to mouse inner
4

cell mass. When injected into blastocysts, ESCs can contribute to all three germ layers including
the germ line of the chimeric animals.
The embryonic stem cells exhibit characteristics similar to the ICM cells. They both express
pluripotency genes, such as Oct4, Nanog, and Rex1. Both of the X chromosomes in female ESCs
remain activated, which is also similar to pre-implanted blastocysts. The major discrepancy
between in vivo cells and in vitro culturing ESCs is that ESCs can be stably passaged for
unlimited times, indicating that the pluripotency state in vitro is stable. Whereas in vivo, the
epiblast cells only transiently maintain their pluripotency signature. The cells in the ICM are not
self-renewing (Santos et al., 2002), and have a hypomethylated genome when compared to ESCs
(Meissner et al., 2008).
The pluripotent state of ESCs is well characterized. ESCs are capable of embryonic body
formation in proper conditions (Hanna et al., 2010b). When injected into immuno-deficient mice,
ESCs are able to form terotoma, a benign encapsulated tumor with tissue or organ components
resembling normal derivatives of all three germ layers (Hanna et al., 2010b). By injecting into
blastocysts, ESCs are able to contribute to all three germ layers including primordial germ cell
specification in chimaeras (Hanna et al., 2010b). On the molecular signaling level, early studies
have identified LIF as the key factor in activating Stat3 and maintaining pluripotency (Ying et al.,
2008). Typical mouse ESCs thus can be maintained with LIF and serum without feeder support.
Further analysis has identified BMP4 as the key factor in the serum (Ying et al., 2003). Recently,
a ‘2i’ condition was defined to use only LIF and 2 small-molecule inhibitors, which inhibit
ERK1/2 and GSK3β (Ying et al., 2008). Together these indicate the significance of MEK/ERK,
JAK-STAT, and WNT pathways in maintaining pluripotenty of ESCs.
5

Mouse ESCs are also featured by the transcriptional circuitry controlled by master genes such as
Oct4, Nanog, Sox2, Klf4 and Tcf3 (Boyer et al., 2005; Jiang et al., 2008; Loh et al., 2006;
Marson et al., 2008). Oct4, Sox2 and Nanog co-occupy a subset of important genes in
pluripotency and differentiation (Jaenisch and Young, 2008). The core regulatory module
activates the pluripotency genes and represses the differentiation genes.
Human ESCs are also isolated from explanted blastocysts (Thomson et al., 1998). Unlike mouse
ESCs, human ESCs cannot be maintained with LIF. On the contrary, human ESCs depend on
FGF2/Activin signaling (Thomson et al., 1998). Other characteristics also differentiate human
ESCs from their mouse counterparts. Human ESCs colonies have flat morphology, while mouse
ESCs have domed morphology. These human ESCs survive poorly when isolated in single cell.
For female human ESCs, one of the X chromosomes is partially inactivated. Some of the
pluripotency markers, such as transcription factors Rex1 and Stella, are also absent in human
ESCs (Hanna et al., 2010b). All these suggest that human ESCs are representing a more primed,
or a more differentiated state comparing to mouse ESCs. However, since starting materials, the
pre-implantation blastocysts from mouse and human, should represent similar developmental
stages, it is hard to explain why human ESCs are more primed to differentiation than mouse
ESCs.
The hypothesis that human ESCs are actually not representing ICM but a later stage developing
epiblast is further supported by the discovery of mouse epiblast stem cells (EpiSCs). These cells
are established from embryos after implantation (E5.5-E7.5). EpiSCs are cultured in supplement
of FGF2 and Activin (Brons et al., 2007; Tesar et al., 2007). The morphology of EpiSCs is
similar to that of human ESCs. They also cannot be maintained after trypsinization. Moreover,
these cells are inactive for one of the X chromosomes, again resembling the human ESCs.
6

EpiSCs are positive for many pluripotency markers and are able to form embryonic bodies and
teratoma. However, the contribution from EpiSCs to the germline after injection into blastocysts
is negative, suggesting developmentally EpiSCs are representing a different population
compared with ESCs. This is further supported by the fact that several key pluripotency factors
such as Rex1, Nanog, and Klf2/4/5 are less expressed and differentiation marker such as FGF5
and MHC class I are highly expressed in EpiSCs comparing to ESCs. To better understand the
different potential of ESCs and EpiSCs, Smith. et.al proposed a model in which ESCs represent
‘naïve’ or ‘ground’ state of pluripotency, whereas EpiSCs represent ‘primed’ state of
pluripotency (Nichols and Smith, 2009). EpiSCs are ‘primed’ to differentiation, but still
considered as pluripotent. Interestingly, mouse EpiSCs can be reverted to mouse ESCs by
overexpressing Klf4 or Nanog, suggesting the primed pluripotent state is plastic and are possible
to be reverted into ‘naïve’ pluripotent state (Guo et al., 2009). More importantly, Surani et. al.
demonstrated that even with the embryos from day 5.5-7.5, a special cell line, termed rESCs,
which is equivalent to ESCs in almost every aspect, can still be generated by rigorous selection
but not overexpression of transcription factors (Bao et al., 2009). This seems to suggest that
using an optimized condition, human blastocysts should also be able to generate ‘naïve’ ESCs. In
fact, overexpression of Klf4/Oct4 or addition of Forskolin, an activator of Klf4, together with
LIF/2i medium, are able to convert human ESCs to a naïve state which are highly efficient in
single cell colony formation and re-activation of the inactive X chromosome (Hanna et al.,
2010a). Moreover, culturing human ESCs under 5% oxygen condition will also lead to pre-X-
inactivation state, suggesting that applying physiological oxygen concentrations may help the
cells maintain naïve pluripotent state (Lengner et al., 2010).

7

1.3 Discovery of induced pluripotent stem cells (iPSCs)
The groundbreaking discovery of iPSCs from Yamanaka and Takahashi (Takahashi and
Yamanaka, 2006) in 2006 started with reprogramming mouse embryonic fibroblasts (MEFs).
The authors applied a candidate approach, selecting 24 factors which are essential for
pluripotency and transduced MEFs with a pool of retrovirus expressing all 24 of them. The
MEFs are drug resistant (βgeo) under the control of Fbx15 promoter, which is only activated in
pluripotent state. The combination of 24 factors, together with the selection, was able to
transform the MEFs into cell which are similar to ESCs. By ruling out factors which are not
essential for generation iPSCs, the authors eventually narrowed down the list to 4 transcription
factors: Oct4, Sox2, Klf4 and cMyc. Using only 4 factors, they were able to reprogram the cells
into iPSCs which are positive for pluripotency-related surface marker SSEA1and nuclear marker
Nanog. Genome-wide expression analysis showed these cells are highly similar to ESCs but
completely different from fibroblasts. These iPSCs were able to form teratoma and differentiate
into all three germ layer tissues, thus meeting the basic criteria of pluripotency. However, some
of the pluripotency marker such as Rex1, were absent in the iPSCs. More importantly, the iPSCs
in this study seemed unable to contribute to the germline of chimeric mice, suggesting that these
original iPSCs may not be fully reprogrammed. The following studies from several independent
labs soon figured out a better way to generate high quality iPSCs (Maherali et al., 2007; Okita et
al., 2007; Wernig et al., 2007). Generally, these studies used selection under ectopic or
endogenous promoter of Nanog or Oct4, instead of Fbx15. This subtle change led to isolation of
iPSCs which are able to contribute to germline of chimera, which meet the stringent criteria of
pluripotency. To further rule out the possibility that iPSCs may still be ‘influenced’ by in vivo
environment and then contribute to all lineage differentiation in chimeric embryos, ‘all-iPSC’
8

chimeric mice were generated when iPSCs were injected into tetraploid blasotcysts (Zhao et al.,
2009). This is so far the most stringent test for pluripotency, since in this case all the tissues in
the F1 mice are from iPSCs. Together, functionally iPSCs has been shown to be equivalent to
ESCs in every functional assay.
Similar to mouse iPSCs, human iPSCs were generated using similar method (Lowry et al., 2008;
Park et al., 2008b; Takahashi et al., 2007b; Yu et al., 2007). The human ESCs culture condition
was used to select iPSCs which resemble the morphology of human ESCs. This method can
actually be extended to other species, such as rat (Chang et al., 2010; Li et al., 2009; Liao et al.,
2009), rhesus monkeys (Liu et al., 2008), pigs (Esteban et al., 2009), and marmoset (Wu et al.,
2010). Interestingly, for some of the species such as rat, the identification of iPSCs precede the
discovery of ESCs (Liao et al., 2009), suggesting that iPSCs can be applied to many other
species for which ESCs are not available due to technical difficulties.
iPSCs can also be derived from various origins. Embryonic fibroblasts are still the most
commonly used cells, since they are easy to transduce by retrovirus and lentivirus. However,
numerous other types of cells are also able to be reprogrammed into iPSCs (Table 1).
Reprogramming of terminally differentiated cells such as B cells (Hanna et al., 2008) suggest
that with proper conditions, induced pluripotency can be achieved in most, if not all, types of
cells.

9

Species Cell type Reprogramming
Factors
Efficiency
(%)
Reference
Mouse Fibroblasts OSKM or OSK or
OS
0.002-
0.02
(Nakagawa et al., 2008;
Takahashi and Yamanaka,
2006; Wernig et al., 2008)
Neural stem cells OK or O 0.01-0.1 (Kim et al., 2008b)
Dermal papilla OKM or OK 0.02-1.4 (Tsai et al., 2010)
Adipose-derived stem cells OSKM 0.2 (Sugii et al., 2010)
B cells, T cells OSKM 0.02-3 (Eminli et al., 2009; Hanna
et al., 2008)
Myeloid progenitors OSKM 25 (Eminli et al., 2009)
Hematopoietic stem cells OSKM 13 (Eminli et al., 2009)
Hepatic endoderm OSK ND (Aoi et al., 2008)
Melanocytes OKM 0.2 (Utikal et al., 2009)
Human Fibroblast OSKM or OSLN 0.002-
0.02
(Takahashi et al., 2007b;
Yu et al., 2007)
Neural stem cells O 0.004 (Kim et al., 2009a)
Hepatocytes OSKM 0.1 (Liu et al.)
Keratinocytes OSKM 0.01 (Aasen et al., 2008)
Mobilized peripheral blood OSKM 0.01 (Loh et al., 2009)
Cord blood stem cells OSKM or OS 0.01 (Eminli et al., 2009)
Adipose derived stem cells OSKM or OSK 0.1-0.5 (Sugii et al., 2010)
Amiotic cells OSKM or OSN 0.05-1.5 (Zhao et al., 2010)
Rat Fibroblasts OSKM or OSK 0.01-0.05 (Liao et al., 2009)
Liver progenitor cells OSK ND (Li et al., 2009)
Neural progenitor cells OSK 0.01 (Chang et al., 2010)
Table.1 Summary of different cell types for reprogramming
(O) Oct4; (S) Sox2; (K) Klf4; (M) cMyc or nMyc; (L) Lin28; (N) Nanog; (ND) Not determined.
The initial attempts to produce iPSCs rely mostly on retroviral transduction. Although highly
efficient and convenient, retroviral vectors insert into multiple sites of the host genome
(Takahashi et al., 2007b). These random insertions may activate or inactive genes that brings
selective advantages to certain iPSC lines, bringing a potential threat for therapeutic application.
Also, retroviral vectors are usually silenced in pluripotent cells, which means the transgenes
usually are not expressed in iPSCs (Mikkelsen et al., 2008). However, the silencing is usually
incomplete and partial reprogrammed iPSCs are observed with some of the reprogramming
factors still expressing (Mikkelsen et al., 2008; Sridharan et al., 2009; Takahashi et al., 2006).
10

These cells are problematic because 1) some lines may rely on overexpression of exogenous
transgenes without fully activating endogenous pluripotency genes; 2) constitutive activation of
cMyc in iPSCs could be potentially oncogenic. In fact, the observation that some of the chimeras
from iPSCs seem more frequently to generate tumors further confirm these speculations (Okita et
al., 2007).
Lentiviral vectors are also popular in delivering genes to target cells. Unlike retroviral vectors,
which can only infect dividing cells, lentiviral vectors can transduce both dividing and non-
dividing cells (Amado and Chen, 1999). However, lentiviral vectors are more likely to escape
silencing in pluripotent stem cells, which may cause expression of reprogramming factors even
in differentiation conditions and lead to skewed differentiation (Brambrink et al., 2008; Stadtfeld
et al., 2008a). To overcome the problem, inducible lentiviral vectors were used to express the
reprogramming factors only during reprogramming (Brambrink et al., 2008; Wei et al., 2009).
Cells only express reprogramming factors with addition of doxycycline, which allows
researchers to select iPSCs which are independent of transgene expression. The other advantage
of inducible system is that after iPSCs are further differentiated into different lineages such as
neural progenitors or fibroblasts, addition of doxycycline will generate second generation of
iPSCs (Wei et al., 2009). The secondary system only requires small molecular doxycycline and
no viral transduction is required. In this case, reprogramming efficiency is elevated (Brambrink
et al., 2008; Wei et al., 2009). Also, primary reprogramming suffers from inconsistent efficiency
due to transduction efficiency variance in each experiment. Secondary systems, however, are
more consistent in terms of reprogramming efficiency. Lastly, genetically homogenous somatic
cells can be obtained via in vitro differentiation or from embryonic tissues of chimeric mice
(Hanna et al., 2008; Wei et al., 2009). This is particularly important for mechanism study, since
11

reprogramming using genetically homogenous cells rules out the influence of random integration
(Hanna et al., 2008; Hanna et al., 2009b). A more advanced and convenient version of secondary
reprogramming system, ‘reprogrammable mice’, took advantages of inducible polycistronic
trangene which contains all 4 reprogramming factors and is inserted at a constitutive active locus
(Carey et al., 2010; Stadtfeld et al., 2010b). In this case, essentially all the tissues from the
‘reprogrammable mice’ are able to generate secondary iPSCs. A similar approach was also
applied to generate mice carrying only 3 reprogramming factors (Markoulaki et al., 2009). Cells
derived from these mice can be used to screen for small molecules that can substitute the one
missing reprogramming factor, as fibroblasts from these mice will not be reprogrammed or only
be reprogrammed at very low efficiency with only 3 factors overexpression.
From a clinical application prospective, insertion by retroviral or lentiviral vector, even at a low
copy number, can still be potential threat for therapy. This is particularly troublesome when
generating massive parallel iPSC lines from individual patients. With different insertion in each
line, identifying various insertions and evaluating the safety and quality of every line will be
laborious and expensive. To avoid using insertion-based gene delivery, multiple methods has
been adopted to generate integration free iPSCs. Generally three types of strategy were applied:
excisable vectors; integration-free vectors; and direct delivery of RNA or protein.
Initial design of excisable vectors was using lentiviral vectors with loxP sites on both ends of
insertion. With transient expression of Cre recombinase after generation of iPSC lines, insertion
will be excised from the host genome (Kaji et al., 2009; Soldner et al., 2009). However, with the
leftover sequence of loxP, it still could potentially disrupt gene expression when inserting into
coding region. An alternative method is to use piggyBac transposons, which are integrated into
genome but can be removed when the transposase is transiently expressed (Woltjen et al., 2009;
12

Yusa et al., 2009). Theoretically a seamless excision can be achieved using this method.
However, excision may not be achieved in each single cell. Thus confirming the excision after
transposase expression will be required and exact insertion sites will also need to be
characterized first. Moreover, human iPSCs survive poorly after single cell isolation, which
brings practical difficulties in generating subclone lines. Together it is still technically
challenging to apply piggyBac method in clinical research.
Integration free vectors can also be applied to deliver reprogramming genes. One of the first
proof of principle experiments used adenoviral vectors to reprogram adult mouse hepatocytes
and MEFs (Stadtfeld et al., 2008c). Similar experiment were performed using human fibroblasts
with various types of vectors, including adenoviral vectors (Zhou and Freed, 2009), sendai virus
(Fusaki et al., 2009), polycistronic mionicircle vectors (Jia et al., 2010) and self replicating
selectable episomes (Yu et al., 2009). Nevertheless, expression of the reprogramming factors
using these methods usually cannot be maintained at high level for a period of time long enough
for reprogramming. Thus the reprogramming efficiency using these non-integrating methods
(0.0001%) are much lower comparing with the conventional methods (0.1-1%).
Another type of method is directly delivering proteins or mRNAs into target cells. This is
achieved by using purified recombinant proteins of Oct4, Sox2, Klf4 and cMyc, each fused with
‘11R’ short peptide which facilitates protein transduction (Zhou et al., 2009). However, this
approach requires purification of all four individual proteins, most of which are insoluble when
expressed in E.coli and need to be solubilized and refolded. The transfected proteins are only
stable for 48 hours, which means multiple rounds of transfection are needed before iPSCs appear.
Therefore the efficiency is extremely low and required addition of small molecule inhibitors such
as valproic acid. More recently, modified mRNA of Oct4, Sox2, Klf4 and cMyc has also been
13

shown to be capable of reprogramming mouse and human fibroblast (Warren et al., 2010). This
relatively straight forward method utilized synthetic mRNA modified to avoid innate antiviral
responses and be relatively stable in the cells. Although multiple rounds of transfection are still
required, the efficiency seems to surpass the traditional methods, reaching 0.6-4.4%. Although
producing large amount of mRNA by in vitro transcription can be laborious, mRNA transfection
still is the most promising non-integrating technology currently.
Besides ectopic expression of transcription factors, multiple small molecules have also been
discovered to replace certain reprogramming factors or boost reprogramming efficiency. For
example, using valproic acid, an inhibitor of HDAC, fibroblasts can be reprogrammed using only
Oct4 and Sox2 (Huangfu et al., 2008a; Huangfu et al., 2008b). DNMT1 inhibitor 5-azacytidine
and histone methyltransferase G9a inhibitor BIX-01294 have also shown to improve efficiency
and transition from partial iPSCs to iPSCs (Mikkelsen et al., 2008; Shi et al., 2008). Tgf-β
signaling inhibitor RepSox, was identified in a screening to replace Sox2 in reprogramming
(Ichida et al., 2009). Although novel molecules have been constantly identified in recent studies,
it is worth noting that 1) so far a small molecule recipe that can replace all 4 reprogramming
factors has not been identified, largely due to difficulty in finding a substitute for Oct4; 2) many
of these inhibitors are potent modulators of DNA methyltransferase or histone modifiers, which
may result in genetic instability and mutant iPSCs.
1.4 Application of iPSCs in disease modeling and regenerative medicine
The ability to generate patient specific iPSCs provided infinite possibilities in modeling disease
in vitro. In fact, the increasing collection of differentiation protocols of iPSCs provide amazing
potential for generation of desired tissues for specific diseases and capturing the disease
14

phenotype from these patient-specific cells. A list of patient specific iPSCs and the diseases
phenotypes can be found in the following table.

15

Table 2. Summary of patient specific iPSCs and phenotype observed in studies. (NA) not available.
Disease category Disease name Genetic cause Phenotype observed in studies Reference
Neurological Parkinson’s
Disease
Polygenic NA (Hargus et al.;
Soldner et al.,
2009)
Polygenic (LRRK2
mutation)
Increased caspase-3 activation
and DA neuron death (in
stressed condition)
(Nguyen et al.,
2011)
Polygenic (with
known PINK1
mutation)
Impairment of mitochondrial
translocation of Parkin in DA
neurons (in stressed condition)
(Seibler et al.,
2011)
Huntinton’s
disease
Monogenic NA (Park et al.,
2008a)
Amyotrophic
lateral sclerosis
Polygenic (with
known SOD1
mutation)
NA (Boulting et al.,
2011; Dimos et
al., 2008)
Spinal muscular
atrophy
Monogenic Reduced number of motor
neurons, loss of SMN gene
expression
(Ebert et al.,
2009)
Friedreich ataxia Monogenic NA (Ku et al.,
2010; Liu et al.,
2011)
Familial
dysautonmia
Monogenic Loss of neural crest cells (Lee and
Studer, 2011)
Duchenne
muscular
dystrophy
Monogenic NA (Park et al.,
2008a)
Rett Syndrome Monogenic Defect in neuronal morphology
and synapse function
(Marchetto et
al., 2010)
Gaucher disease Monogenic NA (Park et al.,
2008a)
Lysch-nyhan
syndrome
Monogenic NA (Park et al.,
2008a)
Down syndrome Chromosome 21
trisomy
NA (Park et al.,
2008a)
Fragile-X
syndrome
Monogenic NA (Urbach et al.,
2010)
Prader-Willi
syndrome
Monogenic NA (Chamberlain
et al., 2010)
Angelman
syndrome
Monogenic NA (Chamberlain
et al., 2010)
Schizophrenia Polygenic Reduced synaptic connectivity (Brennand et
al.)
Cardiovascular Long QT1
syndrome
Monogenic Increased cardiomyocyte
depolarization
(Moretti et al.,
2010)
Long QT2
syndrome
Monogenic Increased cardiomyocyte
depolarization
(Itzhaki et al.)
LEOPARD
syndrome
Monogenic Increased cardiomyocyte size
and decreased MAPK
signaling
(Carvajal-
Vergara et al.,
2010)
Timothy syndrome Monogenic Increased cardiomyocyte
depolarization
(Yazawa et al.)
Hutchinson
Gilford progeria
Monogenic Smooth muscle cell aging
phenotype
(Zhang et al.,
2011)
16

Table 2: continued
Cardiovascular Duchene muscular
dystrophy
Monogenic NA (Park et al.,
2008a)
Blood Fanconi anaemia Monogenic Corrected FANCA iPS cells
can differentiate normally into
myeloid and erythroid linages
(Raya et al.,
2009)
Pancreatic Type 1 diabetes Polygenic NA (Maehr et al.,
2009)
Hepatic A1-antitrypsin
deficiency
Monogenic Loss if A1-antitrypsin
expression
(Rashid et al.)

In most of these studies, patient specific iPSCs were generated from biopsy samples of
fibroblasts and then further differentiated into disease relevant tissues. Since the differentiation
protocols into neural lineages are best characterized, neurological diseases are the best explored
disease category in iPSCs disease modeling.
The best examples of iPSCs disease modeling usually come from those diseases with a clear
genetic cause. For example, spinal muscular atrophy (SMA) is an autosomal recessive disease
caused by homozygous mutation in Survival of Motor Neuron-1 (SMN1) gene. Although SMN1
is a general housekeeping gene functioning in biogenesis of small nuclear ribonucleoproteins for
pre-mRNA splicing, SMN1 may play specific roles in RNA transport in neurons (Burghes and
Beattie, 2009). Ebert et. al. obtained fibroblasts from a SMA type 1 patient and his healthy
mother and reprogrammed them with lentiviral vectors expressing OCT4, SOX2, NANOG and
LIN28 (Ebert et al., 2009). The SMA iPSCs lack SMN1 expression, similar to fibroblasts. When
differentiated into spinal motor neurons for 4 and 6 weeks, the disease line and control line
showed little differences in total neuron numbers based on neuronal marker TUJ1 staining.
However, when motor neuron specific marker ChAT was coupled with TUJ1, the percentage of
motor neurons defined by ChAT and TUJ1 double staining dramatically decreased in patient
samples compared to wild type cells. Also, diseased motor neurons soma size was smaller and
synapse formation was affected. These phenotypes, although evaluated based on rough assays,
17

may represent true defects in disease neurons (Ebert et al., 2009). More importantly, although
SMN1 is disrupted in patient specific samples, SMN2, a homolog gene usually expressed at low
level and only contributing to 10% of total SMN level, can be upregulated modestly in patient
iPSCs derived neurons via addition of tobramycin or valproic acid (Ebert et al., 2009). This proof
of principle experiment demonstrates the potential of using these neurons for therapeutic
screening.
Another extensively studied iPSCs disease model is Rett Syndrome, a neurodevelopmental
disorder caused by X-linked mutation in methyl-CpG binding protein (MeCP2). Fibroblasts
obtained from patients were reprogrammed with OCT4, SOX2, KLF4 and CMYC (Marchetto et
al., 2010). Neurons derived from these samples survive normally, which may not be able to
explain the decreased brain size in Rett syndrome patients. However, VGLUT1 staining revealed
decreased density of puncta in disease glutamatergic neurons, suggesting that glutamatergic
synapse number could be reduced. Reduced numbers of neuritic spines and smaller soma size
were also observed in patient samples. Lastly, irregular function and circuitry of the disease
neurons was demonstrated by a decrease in intracellular calcium oscillations and decreased
frequency and amplitude of spontaneous postsynaptic currents compared to controls. Similar to
the SMA experiment, part of the phenotype can be rescued by adding IGF-1, which will increase
glutamatergic synapse number, and aminoglycoside gentamicine, which will read through the
nonsense mutation of MeCP2 (Marchetto et al., 2010). However, the value of this study may be
compromised by the insufficient characterization of differentiating cell populations. The protocol
utilized retinoic acid (RA) which may induce posteriorization of neuroectoderm, thus producing
cells most likely representing dorsal spinal cord interneurons. Since the phenotypes in vivo are
mostly described in neocortex and hippocampus, the relevance of the iPSC-derived neuron
18

phenotype to the real disease pathology may be questionable (Hansen et al., 2011). These
arguments emphasize the importance of characterization of differentiation cell population, which
is indeed the problem for many diseases that require differentiating into specific lineages which
optimized protocol is still absent currently.
Although numerous disease specific iPSCs have been generated and the ‘bank’ to storage and
distribute these lines may soon be available, there are still lots of hurdles ahead if we are going to
apply these lines into disease studies or phenotypic screening. One of the most puzzling
questions is how we can mimic the late-onset diseases using progenitors or differentiated cells
which usually resemble the embryonic stage of development. Neurodegenerative diseases, such
as PD, AD, and ALS, usually occur in patients over 50 years old. Whether it is valid to use
neurons cultured only 1-2 months in petri dishes to capture the disease has always been
controversial. Indeed, most of the studies focusing on PD so far have not been able to observe
severe disease phenotypes (see Table 2). In fact transplantation assay suggested that
dopaminergic neurons from PD patients can survive and partially rescue amphetamine-induced
rotational asymmetry, but do not exhibit pathological hallmark: inclusion of α-synuclein. This
raised the concern that using standardized differentiation protocol it may be difficult to obtain the
real ‘aged’ cells in vitro. Nevertheless, stressing the cells using oxidative stress, starvation or
neurotoxins, may ‘accelerate’ the aging of cells. In one study, iPS cell lines were generated from
patient with mutation in LRRK2, which led to increased vulnerability when exposed to hydrogen
peroxide, proteosome inhibition dn 6-OHDA. With these stressing conditions, higher level of α-
synuclein was observed in disease lines compared to control lines from healthy patient (Nguyen
et al., 2011). Additionally, for some diseases such as PD, cells may start the process of
degeneration earlier than the onset characterized by pathology. In PD, patients started to lose
19

significant number of dopaminergic neuron before the onset of the motor component of PD (Han
et al., 2011b). Thus even the ‘new’ neurons produced from patient specific iPSCs may still
capture some disease phenotype.
Another challenge of disease modeling is to establish in vitro systems to capture non-cell-
autonomous effects of diseases. In ALS, glial cells are believed to be toxic to SOD-1 mutation
and further lead to the death of motor neurons (Di Giorgio et al., 2008; Marchetto et al., 2008).
Therefore, a co-culture system of glial and motor neurons will ideally identify the specific
contribution from neurons and glias to the disease phenotype. A microfluidic or 3D culture
system will also help building an in vitro disease unit that mimic the in vivo environment.
Whether polygenic diseases can also be effectively studied using patient iPSCs is another
challenge. Complex neurological disorders, such as schizophrenia and autism, usually do not
have clear defined genetic causes. However, cells generated from these patients may still exhibit
phenotypes related to diseases. Schizophrenia patient iPSCs derived neurons demonstrated
decrease synaptic connectivity and can be rescued by antipsychotic loxapine (Brennand et al.).
Expression profiling also suggests that 25% of the differentially expressed genes are related to
schizophrenia. These proof-of-principle experiments are promising for future studies which may
involve a large set of patients with various genetic causes.
With extensive follow-up studies in future, is it possible to perform drug screening and
toxicology studies using patient specific iPSCs? Since iPSCs can be generated from a large
group of patients, we expect to observe diverse responses for drug candidates when treating the
drugs to a library of iPSCs derived cell lines. These responses are supposed to represent the
spectrum of drug responses in real patients. This is attractive because that high failure rate of
20

clinical trial is partly resulted from imperfect animal disease models in preclinical development.
Using human cells mimicking the ‘real’ response should be a better screening strategy. Moreover,
there is a constant debate about advantages in the ‘target oriented’ screening vs. ‘phenotypic
screening’. Patient specific iPSCs derived tissues may eventually be perfect resources for
phenotypic screening, in which only the changes in disease parameters are screened. This may
especially useful for diseases with unknown mechanisms. For example, a large percentage of
ALS and AD patients are sporadic and may be caused by a spectrum of mutations. Thus
screening a library of patient specific iPSCs-derived tissues may be able to discover molecules
which can cover a wider range of patients.

1.5 Molecular mechanisms of reprogramming from somatic cells into iPSCs
Understanding the molecular mechanisms of reprogramming is essential for designing safer and
more efficient reprogramming strategy. Firstly, reprogramming takes more than 2 weeks and the
efficiency is usually low. I will discuss the possible explanation of why only a small percentage
of cells can be reprogrammed. Secondly, reprogramming factors and enhancing elements are
mostly discovered by screening a panel of candidate factors. The specific function of these
factors to initiate or enhance reprogramming will be then explored.
The ‘barrier’ of reprogramming
The ‘Yamanaka recipe’ of reprogramming is inefficient (0.001-0.1%) for most somatic cells.
Using progenitor cells such as neural stem cells and hematopoietic stem cells as starting
population will elevate the efficiency up ~10-20% (see Table 1). Secondary systems will also
elevate the efficiency 10-100 folds (Wei et al., 2009). However, in most cases, only a minor
21

portion of the cells can be reprogrammed, even when every cell is expressing the reprogramming
factors at a similar level (in secondary systems). Therefore, two models, ‘elite’ model and
‘stochastic’ model have been proposed by Yamanaka to explain the phenomenon (Yamanaka,
2009).
In the ‘elite’ model, the low efficiency of reprogramming can be easily attributed to the
heterogeneity of fibroblast cells. The primary fibroblast could contain ‘progenitor’ cells which
are the only cells to be reprogrammed. Since the very first experiment of reprogramming was
using embryonic or adult fibroblast primary culture, the ‘elite’ model was raised and whether
direct reprogramming can be achieved in ‘real’ somatic cells is once unclear. However, the
report that a defined terminal differentiated population such as B and T lymphocytes or β cells
can also be reprogrammed (Eminli et al., 2009; Hanna et al., 2008; Stadtfeld et al., 2008a)
challenged this ‘elite’ model. The most convincing study came from an elegant design by Hanna
et. al. (Hanna et al., 2009b). Monitoring the clonal population of early B cells which are
expressing same level of reprogramming factors for over 8 weeks revealed that almost all clones
can eventually give rise to iPSCs. This suggests that 1) most dividing cells are capable of being
reprogrammed; 2) the efficiency of reprogramming is actually referring to the number of
divisions which can give rise to the first iPS daughter cell. Compared to progenitor cells,
terminally differentiated cells need more divisions to produce the first iPS daughter cell. In other
words, progenitors can be reprogrammed faster than terminally differentiated cells. This can be
better explained by the ‘stochastic’ model. In this model, each cell needs to overcome a number
of ‘roadblocks’ to finally reach pluripotency stage. The expression of reprogramming factors is
not sufficient to overcome these ‘roadblocks’. Thus the stochastic epigenetic events will only
occur in a random small portion of cells. These events led to activation of pluripotency related
22

genes, silencing of integrated reprogramming factors, and reactivation of X chromosome and
telomerase. Since stem/progenitor cells may express some of the pluripotency genes (such as
Sox2 in neural stem cells), a progenitor cell needs to overcome fewer stochastic events to acquire
pluripotency. Thus on the population level, progenitors are reprogrammed at a higher efficiency
and in a shorter period of time.
What is the molecular basis of reprogramming barriers? The transcriptome and epigenome of
somatic cells are very different from pluripotent stem cells. The first responses in somatic cells in
reprogramming include down-regulation of somatic cell markers and morphological changes
similar to a mesenchymal-to-epithelial transition (MET) (Li et al., 2010; Sridharan et al., 2009;
Stadtfeld et al., 2008b). Depletion of genes involved in MET such as E-cadherin and BMP
signaling causes a decrease of reprogramming efficiency (Li et al., 2010; Samavarchi-Tehrani et
al.). Surface markers SSEA1 and alkaline phosphatase were also activated early, followed by
activation of endogenous Oct4, Nanog and other pluripotency circuitry genes (Brambrink et al.,
2008; Wei et al., 2009). The activation of endogenous pluripotency circuitry is usually
considered to be essential for the cells to become independent of ectopic reprogramming factor
expression (Brambrink et al., 2008). Originally it was considered successful reprogramming after
the cells reach independence of ectopic factors. However, further studies suggest that the cells
may still be premature in early passages, since the telomere length and DNA methylation
patterns are different between early and late passage of iPSCs (Chin et al., 2009; Marion et al.,
2009b; Polo et al., 2010). On epigenome level, the histone modification and DNA methylation
are also gradually remodeled to pluirpotency state (Maherali et al., 2007; Mikkelsen et al., 2008).
However, most of these studies are only reflecting changes on population level, while iPSCs are
23

relatively rare and asynchronous. Thus single cell based assay will be essential to determine the
temporal sequence of these events.
Activation of endogenous pluripotency circuitry is usually considered as the most important
event in reprogramming. Although Oct4, Sox2 and Klf4 are key masters in regulating
pluripotency circuitry, activation of endogenous pluripotency circuitry does not occur until late
stage of reprogramming. These seemingly contradictory observations could mainly be caused by
the inaccessibility of the promoters of pluripotency factors (Nanog and Oct4, for example). In
somatic cells the pluripotent genes are highly methylated and not accessable for transcription
factor binding (Sridharan et al., 2009). In transition state represented by partial iPSCs (or pre-
iPSCs), Oct4, Sox2, Klf4 and cMyc can gradually access some of the pluripotency genes.
However, many genes which are occupied by Oct4, Sox2, and Klf4 together in PSCs are only
occupied by only one or two factors in partial iPSCs and thus not activated (Sridharan et al.,
2009). Given the fact that physical interaction between Klf4 and Oct4/Sox2 are crucial for
reprogramming (Wei et al., 2009), we propose a cooperative model, in which Oct4/Sox2/Klf4
complex occupy and activate key endogenous pluripotency genes such as Nanog (Wei et al.,
2009).
It is still unclear how the promoters of endogenous Nanog and Oct4 are demethylated and thus
being accessible for ectopic reprogramming factors. Passive and active demethylation may both
be involved in this process. Inhibition of DNMT1 has been shown to increase reprogramming
efficiency and facilitate the transition from partial iPSCs to mature iPSCs (Mikkelsen et al.,
2008). DNMT1 is required for inheritance of DNA methylation during replication. Therefore
inhibition of passive demethylation may result in partial demethylation of the newly synthesized
genome, especially in CpG islands of pluripotency gene promoters. This is followed by
24

occupation of reprogramming factors and activation of pluripotency circuitry. Consistent with
this hypothesis, ascorbic acid, which has been shown to partially demethylate human ESCs
genome (Chung et al.), has also been shown to significantly increase reprogramming efficiency
(Esteban et al., 2010).
Whether active demethylation is directly involved in reprogramming is still controversial. In
nuclear transfer experiments, reprogramming of donor nuclei does not require replication, which
indicates that some extent of active demethylation must be involved. Direct reprogramming by
transcription factors, on the contrary, requires replication of cells. Also, the potential function of
active demethylation players, such as AID and TET proteins (Bhutani et al.; Ito et al., 2010), has
not yet been confirmed in direct reprogramming.
Another roadblock for somatic cells to overcome during reprogramming is the senescence and
apoptosis. In fibroblasts, inhibiting p53 or ink4a/arf leads to significant increase of
reprogramming efficiency and speed (Banito et al., 2009). On the other hand, post-mitotic cells
such as neurons cannot be reprogrammed unless p53 is suppressed (Kim et al., 2011). This
suggests that immortalization is required for reprogramming. p53 may also be responsible for
inhibiting cell death induced by DNA damage and overexpression of cMyc (Marion et al.,
2009a). Besides, suppression of p53 may also accelerate the cell cycle and increase the efficiency
of reprogramming (Hanna et al., 2009b).
Divergent roles of transcription factors in initiating and enhancing reprogramming
What is the role of reprogramming factors in overcoming these hurdles and promoting epigenetic
remodeling? Oct4, Sox2, and Klf4 are pluripotency genes and part of the pluripotency circuitry
(Jiang et al., 2008). The pluripotency circuitry consist of master genes such as Oct4, Sox2,
25

Nanog and Klf4. These players form complex forward feedback loops and also activate a panel
of downstream ESCs specific genes (Jaenisch and Young, 2008). At the same time, these factors
also work together with repressive chromatin modifying complexes, such as Polycomb complex
PRC2, to repress lineage specific genes (Boyer et al., 2006; Lee et al., 2006). In nuclear fusion
assays, PRC1 and PRC2 have been demonstrated to be required for the nuclear fusion between
ESCs and lymphocytes (Pereira et al., 2010). However, similar experiments have not been
reported in direct reprogramming. The activation and repression of target genes seems to have
different requirement of reprogramming factors. Occupancy of single factor may cause silence of
somatic genes, while co-binding of multiple factors is required for activation of pluripotency
genes (Kim et al., 2008a; Sridharan et al., 2009). This may be able to explain the interesting
observation that silencing of somatic marker occurs immediately after reprogramming start,
whereas activation of endogenous pluripotency circuitry is in the last stage of reprogramming.
Unlike Oct4, Sox2 and Klf4, cMyc seems to have different roles. The binding sites of cMyc do
not significantly overlap with Oct4, Sox2, Nanog and Klf4 (Chen et al., 2008). Analysis from
partial iPSCs also revealed that cMyc activates cell proliferation and energy metabolism early in
reprogramming (Sridharan et al., 2009). Ectopic expression of cMyc was quickly found to be
dispensable for reprogramming in fibroblasts (Nakagawa et al., 2008). However, since cMyc is
lowly expressed in fibroblast and omitting cMyc leads to significant decrease of reprogramming
efficiency, the endogenous cMyc may also play significant role in facilitating reprogramming.

26

1.6 Discovery of long-range interactions via chromosome conformation capture and
fluorescence in situ hybridization (FISH)
The nucleus is a three-dimensional entity. However, most of the epigenetic studies only consider
DNA as one-dimensional linear structure. The higher order chromatin structure has been
suggested to play significant roles in regulating multiple biological functions. However, early
studies are mostly based on microscopy observations and only the general organization such as
chromosome position can be revealed. Recently, various strategies have been developed based
on chromosome conformation capture (3C) (Dekker et al., 2002). The discoveries of 3C can also
be validated by high resolution confocal imaging. I will discuss the various approaches to study
nuclear architecture and their potential and limitation (see summary in Figure 1).

Figure 1. Summary of 3C based methods
27

The groundbreaking 3C technology was introduced by Dekker et. al. to quantitatively study the
interactions between distant DNA regions (Dekker et al., 2002). The cells are fixed with
formaldehyde. The fixed chromatin is then digested by restriction enzymes such as ‘six cutters’-
HindIII, BalmHI and EcoRI, recognizing 6 bp sequence, or ‘four cutters’- DpnII, MboI, and
Csp6I, recognizing 4bp sequence. Then, the fragments containing protein complex and sticky
end DNA are diluted and ligated by T4 ligase. The DNA/protein complexes are so diluted that
most of the ligation is considered to occur between DNA fragments that are in the same complex.
By this way, chimeric DNA fragments which are ligated between two different DNA fragments
are generated. When the amount of certain types of ligation is quantified by PCR using primers
targeted towards the ends of restriction enzyme cutting sites, the frequency of two DNA
fragment located in the same DNA/protein complex can be indirectly measured. Usually various
combinations of primers targeting every cutting site in a certain region are used and then a matrix
of ligation efficiency can be used to describe the interaction frequencies within a certain region.
The original study applied 3C in studying yeast chromosome III. Follow-up studies have adapted
this method in mammalian system and applied quantitative PCR to replace semiquantitative PCR
(Splinter et al., 2006; Tolhuis et al., 2002).
3C has successfully modeled several chromatin loops with distinct biological functions. One of
the best illustrated examples was the contact between upstream LCR and active β-globin gene,
which loops out the 30-50kb sequence between these two sites. The long range interaction is
dynamic and accompanies the changes of transcription of β-globin during development (Palstra
et al., 2003). The mediators of this long range interaction include multiple transcription factors
(Drissen et al., 2004; Splinter et al., 2006; Vakoc et al., 2005). Other examples of long range
interactions discovered by 3C include H19-Igf2 locus, interleukin T
H
2 locus, α-globin locus and
28

CFTR locus (Gheldof et al., 2010; Murrell et al., 2004; Spilianakis et al., 2005; Vernimmen et al.,
2007).
3C aims to discover ‘one to one’ long range interactions. However, in many cases there is no
prior knowledge indicating where the potential interacting sites are. Thus a screening strategy
based on 3C was developed (Simonis et al., 2006; Zhao et al., 2006), named circular
chromosome conformation capture or chromosome conformation capture-on-chip (4C). There
are various protocols using slightly different strategies of 4C. Generally fixation and digestion
are similar to 3C and circular DNA loops were formed using ligation. Inverse PCR primers were
designed to amplify all the unknown sequence that ligated to the bait region. The mixture of
amplified DNA can be analyzed by microarray or next generation sequencing.
To further discover unbiased ‘many to many’ or ‘all to all’ long range interactions, more
comprehensive methods such as 5C and HiC are developed (Dostie et al., 2006). For 5C, oligo
pairs are designed to match every ligation site of interacting 3C fragments in a genomic region.
After amplification, the readout of the junction was performed using microarray or next
generation sequencing. Since the readout includes frequencies of interactions between any two
sites, a 3D conformation of a large genomic region can be constructed. However, the pitfall of
this technology is that a huge amount of oligos need to be synthesized if we are going to study
the conformation of a whole chromosome or the whole genome. The cost of synthesizing large
amount of primers prevents the application of this technology to genome-wide studies.
Compared to 5C, HiC does not rely on designing specific pairs of oligos (Lieberman-Aiden et al.,
2009). Instead, after digestion of 3C templates, restriction ends are filled with biotin-labeled
nucleotides. After blunt end ligation, biotin-labeled fragments are pulled down to perform high
29

throughput sequencing. Only the sequences with restriction enzyme cutting sites and two
sequences from different loci are considered as interaction. The resolution of the published HiC
experiment can reach ~1Mb resolution based on 10 million pair end reads. Nevertheless,
increasing the resolution of HiC seems to be challenging: since it reveals ‘all to all’ interactions,
10-fold increase of resolution will require 100-fold increase in sequence depth. In yeast, due to
the smaller size of genome, resolution can reach kb level (Duan et al., 2010).
Since the resolution of HiC is on Mb level, correlation of HiC data to specific genes or
epigenetic marks will be unrealistic. However, it is still powerful in revealing the chromosome
territories and compartmentalization of genomes. To specifically screen ‘point-to-point’
interactions, a new strategy ChIA-PET combining chromatin immunoprecipitaion (ChIP) with
3C was developed to specifically discover DNA loops which bound by a certain protein
(Fullwood et al., 2009). In the initial study, Fullwood et. al. discovered that the ERα-binding
sites consisted of thousands of intrachromosomal DNA loops. However, ChIA-PET cannot
address whether these loops depends on ERα, since depletion of ERα will also result in failure of
chromatin immunoprecipitation.
The 3C based biochemical approaches are powerful in discovering and confirming long range
interactions. However, long range interactions, especially in trans interactions, are rare events in
general. 4C and HiC rely on statistics to calculate the contact probability instead of raw
sequencing counts, in order to ‘smooth’ the data. The low signal-to-noise ratio is not only caused
by limited sequencing coverage, but also reflects the heterogeneity of population. This is
confirmed by the fluorescence in situ hybridization (FISH). Although limited by resolution of
microscopy, FISH allows researchers to identify proximity of two regions in single cell level.
Careful examination in various studies has indentified that the frequencies of in trans long range
30

interactions occur only in 5-15% of cells (Ling et al., 2006; Lomvardas et al., 2006; Sandhu et al.,
2009; Simonis et al., 2006; Spilianakis et al., 2005; Zhao et al., 2006). Unfortunately, since FISH
require fixation of the cells, there are at least two explanations of the data. One explanation is
that the interactions are transient, thus at a given time, we are only able to observe a small
percentage of interactions. The other explanation, though, can be that certain long range
interactions are stable, but only present in a subset of cells. Only with powerful time-lapse
experiments to trace the movement of two loci in realtime can we have a definite answer.
Nevertheless, both theories may explain why it is difficult to obtain high quality interacting map
using biochemical approaches which often use 10
7
cells at a time.
The fact that long range interactions are only happening in a small portion of cells at a given time
emphasizes the significance of validating the long range interactions using FISH based methods.
In fact, FISH based methods have been used to study the chromosome territory for a long time
(Cremer et al., 1982). FISH experiments targeting active genes such as HoxB have revealed that
active genes tend to loop out of their chromosome territory (Chambeyron and Bickmore, 2004;
Ferrai et al., 2010). Moreover, FISH also demonstrated that genes located far away from each
other can colocalize when activated (Osborne et al., 2004). In the next session I will discuss
further using FISH to investigate the role of long range interactions in transcription.
Although FISH is suitable for single cell analysis, it is low throughput and time consuming. Also,
the resolution of FISH is low, usually only representing regions spanning MBs. Thus it is not
suitable for studying looping between sites within hundreds of kilobases away, limiting its
application in studying many distal enhancer promoter looping.

31

1.7 Higher order chromatin structure and transcription factory
Although in the text book transcription is usually described as one-dimensional model in which
RNA polymerase slides through the transcript region, the transcription in vivo may be very
different. Early immunolabeling experiment of nascent RNA transcripts are not evenly
distributed in the nuclei. Rather, discrete foci were discovered and these sites are sensitive to
transcription inhibition (Wansink et al., 1993). Even after reaching saturation, the number of foci
observed in the labeling studies is only around hundreds to several thousand per nucleus
(Jackson et al., 1993; Osborne et al., 2004; Pombo et al., 1999). The number of sites is several
times lower than the number of active transcription units (Jackson et al., 1993; Jackson et al.,
1998). Thus, several transcription events may share the same foci. The concept of ‘transcription
factory’ was used to describe the gathering of several transcription units (Iborra et al., 1996b),
which is similar to replication factories.
Immuno-gold and immunofluorescent detection of RNA polymerase II (RNAPII) coupled with
nascent RNA revealed that many of RNAPII foci and nascent RNA foci are overlapped (Iborra et
al., 1996a). However, not all RNAPII sites are positive for nascent RNA labeling, indicating that
not all the RNAPII labeled factories are equally active. In fact, recent studies found out that with
inhibition of transcription initiation, the RNAPII-Ser5 (labeled active form of RNAPII) remains
at the foci 30 min after expressed genes move away from the foci (Mitchell and Fraser, 2008).
These studies suggest that transcription factories are not self-assembled RNAPII proteins which
are only brought together as the result of transcription. Rather, these are relatively stable
subnuclear compartments. There are other evidences supporting this hypothesis. For example,
one intriguing question is whether it is the DNA or the transcriptional machinery that actually
moves. Early in vitro studies showed that anchored polymerase can rotate the double helix in a
32

clockwise manner and one strand will slide through the protein during transcription (Kabata et al.,
1993). A recent in vivo study also indicates that it is the DNA loci that dynamically move
towards transcription sites (Papantonis et al., 2010). Together these evidences suggest stable and
organized RNAPII factories are the main sites for transcription.
One of the key features of the transcription factory model is that several transcription units will
share the same factory. This is originally referred to as the aggregation of several neighboring
active genes (Cook, 2002). However, following studies revealed that active genes up to 40 MBs
away can share the same factories at a higher-than-expected frequency (Osborne et al., 2004).
This is confirmed by both RNA-FISH/RNAPII-Ser5 colocalization and 3C assay. The
colocalizations of active loci are apparently not limited to genes in the same chromosome, but
can expand to whole genome, with reduced frequency for in trans colocalizations (Schoenfelder
et al., 2010). In these two examples, α-globin (Hba) and β-globin (Hbb), which are constantly
active in erythroid cells, are considered as ‘super genes’. These ‘super genes’ are constantly been
transcribed judged by nascent RNA FISH (Osborne et al., 2004). Most other active genes, on the
contrary, can only be detected by being transcribed in a small portion of cells at a given time
point, indicating ‘burst’ instead of continuous transcription. Thus one appealing model from
these studies is that ‘super genes’ such as Hba and Hbb constantly occupy transcription factories,
whereas other active genes transiently move in the factory when transcription is needed.
Supporting this model, study of immediate early genes such as Fos and Myc suggest that
induction for only 5 min is enough for Fos to relocate from outside of transcription factories into
constitutively transcribed locus Igh. More interestingly, although Myc is located on a different
chromosome, 25% of the Myc alleles also colocalize with Igh locus. This highly preferential
colocalization, together with the fact that MYC and IGH are the most common translocation
33

partners in Burkitt’s Lymphoma, also suggest that one of the unexpected consequence of sharing
the same transcription factory for in trans partners is the change of translocation. However, it
should be noticed that whether transcription is the direct and instant force that drives the loci to
transcription factories is still under debate. Inhibition of general transcription for 5 hr does not
disturb co-localization (Palstra et al., 2008). Similar results can also be found in Chapter 3. These
conflicting results seem to suggest that co-localizations, though definitely correlated with
transcription, are not dynamically synchronized to the transcriptional burst and pulse.
If the gathering of active alleles is the pure result of sharing basic transcriptional machinery, one
would expect the chance of any pairs of active alleles coming to the same spot will be similar.
This is apparently not the case. Thus a specialized transcription factory which prefers the non-
random organization of specialized chromosomal regions seems more reasonable. Indeed, when
mini-chromosome containing active genes was introduced into the cells, immuno-RNA FISH
showed that they only localized together with endogenous genes in a subset of transcription
facotories, indicating a non-random preferential localization of these active alleles (Xu and Cook,
2008). Schoenfelder et al. provide more direct evidence of in vivo clustering of subsets of active
genes (Schoenfelder et al., 2010). When they introduce a human allele of HBB trangene into
mouse genome, the integrated HBB preferably localizes with the endogenous mouse Hbb locus
compared to mouse Hba locus. When an enhanced 4C assay was applied to screen loci that are
occupied by active RNAPII and interacting with Hbb or Hba, they discover hundreds of
overlapping loci that contain both in cis and in trans transcription partners.
In summary, emerging data suggest a critical part of functionally organized nucleus are the
specialized transcriptional factories. Higher order chromatin structure, including intra- and inter-
chromosomal interactions, ensure the compartmentalization of non-random organization of
34

transcription. These structures can be tissue specific and dynamically regulated in many
biological events. The most challenging question, though, is how these structures are constructed
and maintained, which will be discussed in the following section.
1.8 Current understanding of the regulatory mechanisms of nuclear organization
The nucleus is highly organized with specific conformation. In the past decades many powerful
tools including 3C based methods and FISH have enabled us to investigate the detail of
organization. However, for most of these methods, the readout is DNA sequence. Therefore so
far the description of nuclear architecture is mainly restricted to the conformation of
chromosomes, whereas the organizers of these higher order chromatin conformations remains
elusive. In this section I will discuss a few documented examples of protein and RNA factors
that contribute to the organization of nuclear architecture.
Transcription factors EKLF, GATA-1 and FOG1 in controlling the β-globin locus
The β-globin loci are one of the most extensively studied examples of long range interaction. The
locus control region (LCR) is responsible for regulation of a set of genes encoding variants of the
β-chain of hemoglobin. LCR is located 25 kb from the closest gene and can regulate the gene as
far as 80 kb away. In fetal stages, LCR regulates γ-globin, whereas in adult, LCR controls the
expression of β-globin. 3C analysis confirmed the hypothesis that LCR can loop out genes and
interact with γ-globin in fetal stage cells and with β-globin in adult cells (Palstra et al., 2003).
What are the factors that control the looping? The LCR contains binding sites of EKLF and
GATA-1, which are essential transcription factors regulating expression of β-globin
(Stamatoyannopoulos et al., 1995). When EKLF is knocked out, β-globin expression is severely
reduced. At the same time, long range interactions between LCR and β-globin are also
35

diminished (Drissen et al., 2004). The role of GATA-1 is proved in a reverse manner (Vakoc et
al., 2005). GATA-1 is fused with ER and inducibly expressed in a GATA1 -/- background. After
GATA-1 starts to express, both long range interactions and occupancy of GATA-1 at LCR/β-
globin loci are observed. As a GATA-1 interacting co-factor, FOG-1 is also essential in
mediating the looping. More importantly, the function of FOG-1in mediating long range
interactions depends on its interaction with GATA-1 (Vakoc et al., 2005).
In a more recent study, when genome-wide long range interactions involving β-globin loci were
discovered by an enhanced 4C assay (Schoenfelder et al., 2010), many of these loci are co-
localized with EKLF staining and contain EKLF regulated genes. This again validates the
EKLF’s role in organizing specialized transcription including β-globin and its in cis and in trans
interacting partners. The co-regulation of EKLF, GATA-1 and FOG-1 also serve as an example
that long range interactions are usually regulated by protein complexes and multiple players act
together to establish these interactions.
CTCF regulates the genome organization in pluripotent stem cells
As the only characterized insulator in vertebrate, CTCF is unique for its capability of blocking
enhancers and demarcating the boundaries between euchromatin and heterochromatin (Phillips
and Corces, 2009). However, genome-wide occupancy profiling indicated that CTCF binds to a
huge number of places with unclear function (Chen et al., 2008). CTCF has been demonstrated in
correlating with long range interactions (Donohoe et al., 2009; Ling et al., 2006; Splinter et al.,
2006). However, direct evidences of CTCF regulating genome-wide long range interactions were
not available until Handoko et. al. presented elegant work of characterizing CTCF-mediated
functional chromatin interactomes recently (Handoko et al., 2011).
36

Using ChIA-PET technology, Handoko et. al. pulled down CTCF in 2 biological replicates and 2
technical replicates and applied deep sequencing to recover long range interactions in cis and in
trans. Together the authors identified 1480 intrachromosomal interactions and 336
interchromosomal interactions. These interactions involve 3306 CTCF binding sites. Parts of
interactions can be confirmed by 4C and FISH assay. Using 3C and FISH to compare control and
CTCF knockdown cells, the authors also demonstrated the dependence of CTCF in these
interactions. Most importantly, studying the epigenetic features of genes inside and outside the
loop allowed the authors to define five different categories of looping and further implicate the
biological function of looping. The first category features active epigenetic marks inside the loop
with repressive marks outside. The second category is the opposite of first category, with
repressive marks inside and active genes outside. The third category represents enhancer
promoter loops. The fourth category seems to be the barrier between active and repressive
regions, with these two types of epigenetic marks on opposite sides of the loop. The last category
does not have a common epigenetic feature. Defining these categories of looping revealed the
multifunction of CTCF in regulating higher order chromatin structure and the link between
looping structure and local transcription.
However, several important questions remain to be answered for CTCF’s function as organizers.
Firstly, CTCF has more than 68,000 binding sites, but only 3000 of them are involved in looping.
These three thousands sites also are not distinguished from other CTCF sites. This suggests other
cofactors may work together with CTCF. Another explanation could be these binding
interactions are again transient and only occurs in a subpopulation of cells, thus the sequencing
depth right now for ChIA-PET is far from enough. Secondly, CTCF is ubiquitously expressed.
37

So far there are no comprehensive data showing whether transient depletion of CTCF has
profound influence in transcription. Thus the biological meaning of these looping is still elusive.
Xist, Tsix, and other long non-coding RNAs in X chromosome inactivation
Perhaps the best example of long non-coding RNAs (lncRNAs) in mediating higher order
chromatin structure is the X chromosome inactivation (XCI). Xist is the first lncRNA identified
to be essential for XCI, followed by discovery of other lncRNA players such as RepA, Tsix and
Jpx (Lee, 2010).
Before differentiation, Tsix is expressed biallelically and prevents the binding of RepA-PRC2
complex to the Xist promoter. During the initiation of differentiation, Tsix expression is limited
to only one allele (randomly selected). This leads to the loading of RepA-PRC2 complex to Xist
loci in X chromosome to be inactivated (Xi), and paradoxically activates Xist expression. Xist
transcripts then recruit PRC2 in a co-transcriptional manner. At the same time, YY1 binds to the
‘nucleation center’ of Xi but not the activated X chromosome (Xa). Finally, the PRC2-Xist
complex is loaded to YY1 and starts to spread from the nucleation center to the neighboring
regions. The spreading eventually results in the ‘coating’ of PRC2 and Xist on the surface of the
entire inactivated X chromosome (Lee, 2011).
The nuclear architecture of Xi is dynamic during differentiation. Repeating elements such as
LINEs and SINEs are aggregated into the core of the Xi territory, whereas active gene-rich
regions are located in the outer surface. 4C analysis of gene-rich regions in Xi and Xa also
showed distinct interacting partners (Splinter et al., 2011). Interestingly, although Xist is clearly
involved in almost every step of XCI, depletion of Xist after XCI seems to have little effect on
the chromosomal confirmation (Splinter et al., 2011).
38

The central role of lncRNAs in XCI may not be a special case, but rather indicate the universal
role of lncRNAs as mediators of higher order chromatin structure. lncRNAs are naturally
suitable for mediators of long range interactions. They can tether to its transcription units,
therefore serve as the allele specific tag. The binding site of lncRNA can also be more selective
compared to DNA binding proteins. Moreover, lncRNA can serve as adaptors and interact with
multiple proteins. Considering the fact that deep sequencing indentified lncRNA are transcribed
from 60-90% of the genome, one may imagine that many of these lncRNAs may serve as
essential regulators of nuclear architecture.
In general, our knowledge of regulation of higher order chromatin structure is still limited. Case
studies of transcription factors, insulators and lncRNAs are discussed above. Nevertheless,
composition of the organizing complexes remain largely unknown. How these proteins and
RNAs regulate the dynamic nuclear architecture is also an important but unanswered question.
Developing novel tools to systematically and quantitatively study the composition of nuclear
organization will be the key to the future research.

39

Chapter 2 Klf4 Directly Interacts with Oct4 and Sox2 to Promote Reprogramming

2.1 Abstract
Somatic cells can be reprogrammed to induced pluripotent stem (iPS) cells by ectopic expression
of specific sets of transcription factors. Oct4, Sox2, and Klf4, factors that share many target
genes in embryonic stem (ES) cells, are critical components in various reprogramming protocols.
Nevertheless, it remains unclear whether these factors function together or separately in
reprogramming. Here we show that Klf4 interacts directly with Oct4 and Sox2 when expressed
at levels sufficient to induce iPS cells. Endogenous Klf4 also interacts with Oct4 and Sox2 in iPS
cells and in mouse ES cells. The Klf4 C-terminus, which contains three tandem zinc fingers, is
critical for this interaction and is required for activation of its target gene Nanog. In addition,
Klf4 and Oct4 co-occupy the Nanog promoter. Specific dominant negative Klf4 mutant can
compete with wild-type Klf4 to form defective Oct4/Sox2/Klf4 complexes and strongly inhibit
reprogramming. In the absence of Klf4 overexpression, endogenous Klf4 and its interaction with
Oct4/Sox2 are also required for reprogramming. This study supports the idea that direct
interactions between Klf4, Oct4 and Sox2 are critical in somatic cell reprogramming.
2.2 Introduction
The recent development of methods to reprogram somatic cells to induced pluripotent stem (iPS)
cells using retroviral or lentiviral transduction of four genes (Oct3/4, Sox2, c-Myc and Klf4)
represents a major breakthrough in stem cell research (Hochedlinger and Plath, 2009; Jaenisch
and Young, 2008; Takahashi and Yamanaka, 2006). Further analysis shows that three genes,
Oct4, Sox2, and Klf4, are critical to the process and that c-Myc functions as an enhancer of
40

reprogramming efficiency (Feng et al., 2009b; Nakagawa et al., 2008; Wernig et al., 2008).
Additionally, inducible systems have been developed to better control transgene expression
(Brambrink et al., 2008). Novel methods requiring no viral integration have also been developed
(Kaji et al., 2009; Okita et al., 2008; Stadtfeld et al., 2008c; Woltjen et al., 2009; Zhou et al.,
2009), as have strategies using small molecules to promote reprogramming efficiency (Huangfu
et al., 2008b; Marson et al., 2008; Shi et al., 2008).
As a key factor in reprogramming, Kruppel-like factor 4 (Klf4/GKLF/EZF) functions as both a
transcriptional activator and repressor to regulate proliferation and differentiation of different
cell types (Evans et al., 2007). RNAi experiments confirm that Klf4 is redundant with the other
two family members, Klf2 and Klf5, in regulating expression of pluripotency-related genes
(Jiang et al., 2008). In ES cells, Klf4 has been shown to be important for activating Lefty1
together with Oct4 and Sox2 (Nakatake et al., 2006). Genome-wide ChIP-Chip demonstrates that
the DNA binding profile of Klf4 overlaps with that of Oct4 and Sox2 on promoters of genes
specifically underlying establishment of iPS cells, suggesting transcriptional synergy among
these factors (Sridharan et al., 2009). Furthermore, studies also suggest that Klf4 may function in
establishing an “authentic” and “metastable” pluripotent state in various pluripotent cell types
(Guo et al., 2009; Hanna et al., 2009a).
The fact that other Klf family members can substitute for Klf4 in reprogramming (Nakagawa et
al., 2008) suggests that motifs common to this family are important for reprogramming activity.
The only structural similarities common to Klf family proteins are C-terminal tandem zinc finger
motifs (Philipsen and Suske, 1999). Interactions between Klf1 (EKLF) and GATA-1 suggest that
Klf family C2H2 zinc fingers can also bind other transcriptional partners (Merika and Orkin,
1995), which may be functionally conserved among Klf family members (Wolfe et al., 2000).
41

However, little is known about potential interaction partners of Klf4 and the significance of these
interactions in reprogramming.
Here we show that Klf4 interacts directly with Oct4 and Sox2 in iPS cells and ES cells. The Klf4
C-terminal three zinc fingers mediate both interactions and are required for transcriptional
activation of the target gene Nanog. We found that specific Klf4 mutants can compete with
ectopic or endogenous wild-type (WT) Klf4 to form transcriptionally defective complexes with
Oct4 and Sox2, inhibiting reprogramming efficiency. These results indicate that Oct4, Sox2 and
Klf4 function via direct interaction to regulate downstream targets and facilitate reprogramming.
2.3 Materials and Methods
Plasmid construction
To generate the doxycycline-inducible viral expression vector, the digested rtTA2S-M2 fragment
(Urlinger et al., 2000) was inserted into the vector FUIPW, containing an internal ribosomal
entry site (IRES) followed by the puromycin resistance gene. The ubiquitin promoter of FUW
(Lois et al., 2002) was replaced with a tetracycline-responsive element (TRE) containing a CMV
minimal promoter to construct FTRE. cDNAs encoding Oct4, Sox2, Klf4, and Klf4 mutants
were subsequently cloned into FTRE and FUIPW. Klf4 and its mutants were also cloned into
pCS2 as described (Evans et al., 2007).
Cell culture, lentivirus preparation, iPS cell generation and in vitro differentiation
293T cells were maintained in DMEM (Cellgro) containing 10% FBS. FTRE-based lentiviruses
were generated in 293T cells as described previously (Lyu et al., 2008). Virus-containing
42

medium was collected at 48 hours after transfection and virus was concentrated by
ultracentrifugation at 28,000rpm for 2 hrs. Concentrated viruses were reconstituted in PBS.
Reprogramming of primary MEFs was performed as described previously (Takahashi et al.,
2007a). Briefly, primary mouse embryonic fibroblasts (MEF) were generated from E13.5
embryos carrying a transgene GFP under the control of Oct4 promoter (Szabo et al., 2002).
6×105 MEFs were seeded in 100mm dishes and transduced two times with a cocktail of 5
lentivirus, including 4 of the reprogramming factors and one containing rtTA. Mouse ES
medium (GMEM with 15% FBS, 2 mM glutamine, 0.1 mM β-mercaptoethanol, 1% non-
essential amino acid, 1% sodium pyruvate, LIF 1,000 u/ml) containing 0.1μg/ml of doxycycline
were used after 2 days and changed everyday afterwards. 3 weeks later mature iPS colonies were
isolated by manual cutting and individual lines were maintained and characterized.
For in vitro differentiation, iPS cells were maintained on feeder layers of irradiated mouse
embryo fibroblasts (MEFs) in mES growth medium. To obtain neural stem cells (NSCs), iPS
cells were detached from the MEF layer with 1 mg/ml collagenase for 15 min at 37°C and
clumps were transferred to 9 cm bacterial dishes as suspension cultures in mouse ES medium
lacking LIF. After 4 days, differentiating clusters resembling embryoid bodies (EBs) were
transferred to tissue culture dishes in 2% B27 (Invitrogen) defined medium with 20 ng/ml of
bFGF for 14 days differentiation to form NS with medium change every 2 days. Neurospheres
were collected and treated with 500 µl of 0.05% trypsin at 37°C for 10 min, then triturated and
neutralized with 1mg/ml trypsin inhibitor (Sigma). NSCs were subsequently cultured as
monolayer on poly-L-lysine and fibronectin coated dishes.

43

RT-PCR analysis
Total RNA was isolated from iPS cells using an RNAeasy mini kit (Qiagen). 2μg of RNA was
subjected to the RT reaction using Superscript II (Invitrogen). Semi-quantitative PCR was
performed to evaluate total gene expression, using primers previously described (Takahashi and
Yamanaka, 2006).
Immunostaining and AP staining
iPS cells grown on feeders were fixed at 2% paraformaldehyde for 10 min. Cells were incubated
with SSEA1 primary antibody (Hybridoma bank, 1:400) for 1hr, washed with PBS, and
incubated with secondary antibody (goat anti mouse IgM, Jackson immune, 1:1000) for 30 min.
Alkaline phosphatase staining was done using the manufacturer’s protocol (Vector).
Teratoma formation and histological analysis
iPS cells were suspended at 1×107 cells per ml in PBS. 100μl of cell suspension were injected
subcutaneously into the dorsal flank of SCID mice (Charles River). Six weeks later, samples
were fixed in Bouin’s fixation buffer. Sections were stained with hematoxylin and eosin.
Coimmunoprecipitation and Western blotting
Co-immuniprecipitation and Western blotting were performed as previously described (Lyu et al.,
2008). Antibodies used were: anti-Flag (Sigma), anti-HA (Santa Cruz), anti-Klf4 (a gift from Dr.
Ng), anti-Sox2 (R&D), anti-Oct4 (Santa Cruz), and anti-myc (Santa Cruz). The relative quantity
of western blotting bands were quantified by ImageJ.

44

Purification of recombinant proteins and in vitro binding
GST-Klf4 (300-483) was constructed in pGEX41T, expressed in BL-21 E. coli and purified by
affinity chromatography using glutathione-sepharose (GE Healthcare) as described (Harper and
Speicher, 2001). HA-Oct4 and HA-Sox2 were cloned in to pGEX41T and further purified by the
same method. GST-tagged Oct4 and Sox2 were further cleaved by Thrombin (GE healthcare) at
room temperature for 2hr. After further pre-clearing with glutathione-sepharose, the supernatant
was checked by Western blotting using anti-HA antibody to confirm that Oct4 and Sox2
remaining in the supernatant were completely cleaved. In vitro binding was assayed by mixing
the glutathione-sepharose bound GST-Klf4 (300-483) or GST with Oct4 (HA) and Sox2 (HA) at
4°C overnight. Sepharose beads were washed, boiled, and ready for Western blotting.
Luciferase reporter assays
Luciferase reporter assays were performed as described (Lu et al., 2004).
Electrophoretic mobility shift assays (EMSA)
293T cells were transfected with vectors expressing wild-type or mutant Klf4. Cells transfected
with GFP vector served as a control. Nuclear extracts were prepared as described (Lyu et al.,
2008) with modifications.
For EMSA, Klf-binding DNA (Jiang et al., 2008) oligonucleotides were synthesized and labeled
with IRD800 dye (IDT). They were then annealed at 5mM. For DNA binding reactions, 1μl of
DNA was added to a 10-μl reaction containing 1μl of nuclear extract, 2μg of salmon sperm DNA
(Gibco), and 1μl of 10× binding buffer (100mM Tris, 10mM EDTA, 1M KCl, 1mM DTT, 50%
45

glycerol). After 30 minutes incubation, the mixture was resolved on pre-run 10% native PAGE
gels in 1×TBE. Gels were imaged directly on glass plates using Li-Cor Odyssey imager.
Chromatin immunoprecipitation (CHIP) assay
CHIP assays with mouse ES cells (or iPS cells) were carried out as described (Wells and
Farnham, 2002). For all ChIP experiments, relative occupancy values were calculated by
determining the apparent immunoprecipitate efficiency (ratios of the amount of
immunoprecipitated DNA over that of the input sample) and normalized to the level observed at
a control region (primer 1), which was defined as 1.0.
Reprogramming efficiency assay
To perform secondary generation iPS cells efficiency assay, 3×105 NSCs from iPS cell in vitro
differentiation or second generation MEFs derived from chimeric mice were plated on 10cm
plates. Equal amount of viruses encoding mutant forms of Klf4 or control virus were added and
the medium was changed to fresh medium after 24 hours. Forty-eight hours later, cells were
trypsinized and re-plated into 6-well plates at 5×104 cells per well. NSCs were then seeded on
irradiated MEF coated plates, while MEFs were seeded on 0.1% gelatin. Duplicate or triplicate
wells were usually prepared for each sample. The medium was again changed into mouse ES cell
medium and doxycycline was added to 1ug/ml. Medium was changed every day and
supplemented with doxycycline throughout the induction period. FACS analysis was performed
after 1 to 2 weeks. GFP-positive colonies were counted at various time points.
To evaluate the quantity of AP positive colonies in primary MEFs infected with Oct4, Sox2 and
Klf4 mutants without exogenous wild-type Klf4 expression, 2×105 or 1×105 primary MEFs
isolated from Oct4-GFP mice were seeded in 60mm dishes or one well of six-well-plates, and
46

transduced with various combination of lentiviruses as indicated in the text. Alkaline
phosphatase staining were performed 2 weeks after transduction.
Flow cytometry
Cells were trypsinized, washed once in PBS, and resuspended in PBS. Cells were stained with
anti-SSEA1 antibody (Hybridoma bank) diluted 1:250 for 15 min. After one washing with PBS,
a secondary antibody (Cy3-conjugated goat anti-mouse IgM, Jackson Immune) diluted 1:800
was added for 15min. Cells were then washed twice in PBS and resuspended in PBS for analysis.
Lentivirus transduction on mouse ES (mES) cells
FUIPW based Klf4 mutants or control vectors were used for transduction. 8μg/ml polybrene
were used to help transduction. 48 hours after transduction, 0.5μg of puromycin were added and
medium were changed everyday afterwards. After 6 days selection, the remaining cells were
stained for SSEA1 and counted for the percentage of SSEA1 positive cells.
2.4 Results
Generation of secondary iPS cells using genetically homogeneous cells
Previous studies showed that the efficiency of direct reprogramming is generally low and
variable (Feng et al., 2009b). To investigate the detailed function of Klf4 and other factors in
reprogramming events, it was necessary to establish a high efficiency system with consistent
readout. Therefore, we set up a secondary iPS cell induction system using inducible expression
of reprogramming factors (Fig. 2A). In this system, mouse embryonic fibroblast (MEF) cells
were transduced by lentivirus expressing Klf4, Oct4, and Sox2 whose expression is controlled by
doxycycline induction. These cells were also transduced with lentivirus expressing doxycycline
47

co-activator rtTA. After weeks of induction in ES cell medium in the presence of doxycycline,
doxycycline was then removed and many colonies emerged. Stable first generation iPS cell lines
were isolated, characterized and then differentiated in vitro to generate neural stem cells. After
several passages under differentiation conditions, this neural stem cell population became
homogeneous, and second generation iPS cells can be induced by addition of doxycycline. In
addition, first generation iPS cells were injected into mouse blastocysts to generate chimeric
mice. MEFs derived from first generation iPS cells chimeric mouse can be isolated. These MEFs
can also be reprogrammed to generate second generation iPS cells by addition of doxycycline.

48

Figure 2. Establishment of inducible reprogramming system using defined factors.
(A) Schematic representation of the experimental design showing first generation iPS cell derivation and secondary
iPS cells generated from NSCs differentiated in vitro from iPS cells and MEFs from chimeric mice. iPS cells
are generated by doxycycline-induced expression of Oct4, Sox2, c-Myc and Klf4 in mouse embryonic
fibroblast (MEF) cells using lentiviral transduction. MEFs are derived from E13.5 mouse embryos containing
an Oct4-driven GFP transgene. The construct encoding rtTA for doxycycline induction expresses the
puromycin resistance gene. Stable iPS cell lines are isolated, characterized and then differentiated in vitro to
generate neural stem cells. After several passages under differentiation condition, this neural stem cell
population becomes homogeneous and second generation iPS cells can be induced by addition of doxycycline.
First generation iPS cells are also injected into mouse blastocysts to generate chimeric mice. iPS cell-derived
MEFs are isolated from E13.5 chimeric embryos using puromycin selection. These cells can form second
generation iPS cells after addition of doxycycline.
(B) Pluripotency markers are expressed in (Oct4-GFP) iPS cell lines. iPS cell lines were positive for endogenous
Oct4 shown by GFP expression, alkaline phosphatase (AP) and SSEA1 (Bars, 5 µm).
(C) Teratomas derived from iPS cells show differentiation of iPS cells into cartilage (mesoderm), muscle
(mesoderm), gut-like structures (endoderm), and neural epithelium (ectoderm).
(D) A second generation of homogeneous NSCs can be reprogrammed at high efficiency following doxycycline
(Dox) treatment with or without valproic acid (VPA) (n=3; error bars indicate s.d.). NSCs derived from in vitro
differentiation were seeded 5×10
4
cells/well. Doxycycline was added at the concentration of 0.5μg/ml or
1μg/ml. VPA was added at the final concentration of 2μM to some cultures for 14 days. GFP expression is
driven by the Oct4 promoter. The number of GFP-positive colonies was determined after three weeks of
induction.

49

To establish first generation iPS cell lines, lentiviral vectors expressing Oct4, Sox2, Klf4 and c-
Myc under doxycycline control (TRE promoter) (Fig. 3A), together with a constitutively active
lentivirus expressing rtTA with puromycin selection (Fig. 3A), were transduced into mouse
fibroblasts (MEFs) isolated from E13.5 transgenic mice embryos expressing green fluorescent
protein (GFP) driven by the pluripotent cell-specific Oct4 promoter (Oct4-GFP) (Szabo et al.,
2002). After three weeks of induction (Fig. 3B), ES cell-like colonies with GFP signals emerged
(Fig. 3C). Further characterization suggested that, in addition to Oct4 (as indicated by GFP
expression), these iPS cell lines were also positive for other pluripotency markers such as
alkaline phosphatase (AP) and SSEA1 (Fig. 2B). RT-PCR analysis confirmed the expression of
pluripotency-related genes in these cell lines (Fig. 3E). All lines were positive for SSEA1 and
AP, although some did not express endogenous Oct4 or Nanog as indicated by RT-PCR,
indicating these cells may be partially reprogrammed iPS cells. We chose lines most similar to
ES cells in pluripotency gene expression for further experiments. To further examine
pluripotency, we injected different iPS cell lines into SCID mice and evaluated teratoma
formation. After 7 to 9 weeks, multiple lines developed into teratomas, and further analysis
indicated that they contained derivatives of all three germ layers (Fig. 2C). The same cell lines
were injected into mouse blastocysts to generate chimeric mice. These cell lines contained
transgenes of these reprogramming factors as shown by PCR using primers specific to the
transgenes (Fig. 3F).
50

Figure 3. Characterization of first generation iPS cells
(A) Inducible lentiviral constructs contain cDNA encoding Oct4, Sox2, c-Myc and Klf4. A separate lentiviral
construct expresses rtTA and a puromycin resistance gene under ubiquitin promoter control.
(B) A schematic outline showing experimental time-line in producing first generation iPS cells from MEFs. Two
transductions were made at Day 0 and Day 5 to guarantee high efficiency of transduction. Cells were cultured
in the presence of puromycin in mES medium and 0.1μg/ml doxycycline. iPS colonies were manually isolated
after three weeks.
(C) Bright field and fluorescent images of a GFP-positive iPS colony, three weeks after transduction with
lentiviruses.
(D) RT-PCR analysis of pluripotency-related gene expression. N-acetyltransferase 1 (NAT1) is used as an internal
loading control. MEF OG: MEF cells containing the Oct4 GFP transgene which are initially used for
reprogramming.
(E) Transgene integration was detected by PCR in individual iPS lines 54 and 76, and in NSCs generated from in
vitro differentiation of iPS 76. Mouse ES cells serves as control. Genomic DNA is extracted from cells using a
DNeasy kit (Qiagen). Forward primers are: Oct4-FP, CTATGGAAGCCCCCACTTC; Sox2-FP,
CGCCCAGTAGACTGCACAT; Klf4-FP, CACTACCGCAAACACACAGG; and cMyc-FP,
TCAAGAGGCGAACACACAAC. The reverse primer, complementary to the lentivirus WRE region, is WRE-
RP, ACTGTGTTTGCTGACGCAAC.
(F) Genotyping of chimeric embryos with transgene integration using primers specific for each transgene. A wild-
type embryo (lane 1) serves as a negative control.

To derive genetically homogeneous neural stem cells (NSCs) directly from first generation iPS
cells, embryoid bodies were generated from iPS cells (Fig. 4A), and cells were then further
51

differentiated into neural progenitors. Oct4-dependent GFP expression disappeared in later
stages of embryoid bodies and no Oct4-driven GFP-positive cells were detectable after passages
in monolayer NSC culture conditions (Fig. 4B), indicating that no pluripotent cells remained.
To derive second-generation iPS cells, NSCs derived from in vitro differentiation were plated in
the presence of doxycycline at the same density in multiple wells with mES medium (Fig. 4C).
Colonies emerged within a week, and mature Oct4-GFP-positive colonies appeared in 2 weeks
(Fig. 4D). We calculated reprogramming efficiency by counting the number of GFP-positive
colonies after 3 weeks (Fig. 2D). To optimize conditions favoring iPS cell induction, NSCs were
treated with 0.5 µg/ml or 1 µg/ml doxycycline and with or without the HDAC inhibitor valproic
acid (VPA) (Huangfu et al., 2008b). iPS cell induction efficiency increased with both increased
doxycycline concentration and VPA treatment (Fig. 2D). By counting GFP-positive colonies
after 3 weeks, we estimated reprogramming efficiency of NSCs to be approximately 0.04%
(without VPA) to 0.2% (with VPA), a frequency higher than that obtained by directly generating
iPS cells from transduction of neural progenitors (Eminli et al., 2008). These iPS colonies also
exhibited SSEA1 and AP expression (Fig. 4E).
52

Figure 4. Generation of a homogeneous population of NSCs and 2
nd
generation iPS cells
(A) Bright field and fluorescent images of embryoid bodies from iPS cells at day 3 after induced differentiation.
Note that GFP (driven by the Oct4 promoter) is still expressed at this stage.
(B) Bright field and fluorescent images of NSCs at passage 2. Cells have become homogeneous in morphology and
are GFP-negative.
(C) Schematic representation of the 2nd generation iPS cell induction using homogeneous NSCs from in vitro
differentiation. GFP-positive colonies are visible after 1 week of induction with doxycycline. SSEA1-positive
cells appear earlier than GFP-positive cells.
(D) Bright field and fluorescent images (insets) of second generation iPS cells show cell morphology and Oct4-GFP
expression of homogeneous NSCs treated with doxycycline for 6, 12, and 17 days. (Bars, 2.5μm)
(E) SSEA1 and AP staining in iPS colonies from secondary generation of iPS cells, indicating that they are true iPS
colonies. GFP expression driven by the Oct4 promoter is also detectable.

To obtain MEFs specifically derived from first generation iPS cells, we used chimeric mice
generated by blastocyst injection with iPS cells. MEFs from E13.5 chimeric embryos were
isolated and wild-type cells were eliminated by puromycin selection. After doxycycline
treatment to induce iPS cells, Oct4-GFP positive colonies started to appear within two weeks.
These iPS cells are also SSEA1 and AP positive (data not shown). The secondary iPS cell
53

induction from NSCs and MEFs will be used as assays for reprogramming efficiency in the
following study.
Klf4 directly interacts with Oct4 and Sox2
Klf4, Sox2 and Oct4 are critical for reprogramming, suggesting that these factors may function
as a complex. To determine whether they physically interact with each other, we first examined
their potential interactions in 293T cells. In 293T cells over-expressing Flag-tagged Klf4 and
untagged Oct4, Oct4 was co-immunoprecipitated by an anti-Flag antibody (Fig. 5A). Alternative
experiments in which antibodies were reversed showed that Klf4 was co-immunoprecipitated by
an anti-Oct4 antibody (Fig. 5A), indicating that Klf4 binds to Oct4. Similar results were obtained
analyzing Klf4 interaction with Sox2 (Fig. 5B). To confirm interaction of endogenous Oct4,
Sox2 and Klf4 in iPS cells, we performed co-immunoprecipitations from lysates of iPS cells not
treated with doxycycline. When Klf4 was immunoprecipitated, Oct4 was detected by Western
blotting (Fig. 5C). Oct4 and Sox2 interaction was also demonstrated by immunoprecipitating
Sox2 and blotting for Oct4 (Fig. 5D). Interaction between endogenous Klf4 and Sox2 in iPS cells
was undetectable (data not shown), possibly due to either weak interactions or low sensitivity of
the Sox2 antibody. To determine whether these interactions were unique to iPS cells, we
performed the same experiments on mouse ES cells and detected similar interactions (Fig. 5E
and F). These data indicate that Klf4 interacts with Oct4 and Sox2 in iPS as well as ES cells.
54

Figure 5. Klf4 interacts with Oct4 and Sox2.
(A-B) Klf4 interacts with Oct4 (A) and Sox2 (B) when over-expressed in 293T cells. Constructs encoding Flag-
tagged Klf4 or untagged Oct4 were transfected into 293T cells alone or together. Cell lysates were
immunoprecipitated using anti-Flag antibody followed by Western analysis with anti-Oct4 antibody. Klf4 and Oct4
expression in whole cell lysates was determined by Western blot.
(C-F) Endogenous Klf4 interacts with endogenous Oct4 in iPS (C) and ES (E) cells, and endogenous Oct4 and Sox2
interact with each other in iPS (D) and ES (F) cells. iPS cells were maintained and passaged without doxycycline.

The Klf4 C-terminus contains three consecutive, highly conserved C2H2 zinc fingers. These
domains are generally thought to mediate Klf4’s DNA binding function (Mahatan et al., 1999;
Xie et al., 2008). To investigate which domains are required for Oct4 and Sox2 interaction, we
generated Klf4 deletion mutants including those deleted in one or more zinc fingers (Fig. 6A).
Interactions of these mutants with Oct4 and Sox2 were determined by co-immunoprecipitation.
Comparison of mutants having truncations of the last one (Klf4ΔZF3), last two (Klf4ΔZF2-3), or
all three (Klf4ΔZF1-3) zinc fingers indicated that deletion of all three zinc finger motifs reduced
55

interaction between Klf4 and Oct4 (Fig. 6B), while deleting all three motifs completely
abolished Klf4 interaction with Sox2 (Fig. 6C). Deletion of the C-terminal two zinc fingers
(Klf4ΔZF2-3) resulted in an intermediate reduction in interaction with Sox2 (Fig. 6C). Other
mutants, such as Klf4ΔM, which lacks the middle of the protein but retains all three zinc fingers,
did not alter interaction of Klf4 with Oct4 or Sox2 (Fig. 6B and C). These results indicate that
the Klf4 C-terminus is important for binding to Oct4/Sox2, and also that Oct4 and Sox2 may
bind to different regions of Klf4. We also found that neither the Klf4 middle region nor the N-
terminus is sufficient for binding Oct4 and Sox2 in 293T cells (Fig. 6D and E), while Klf4’s C-
terminus is sufficient for interacting with both Oct4 and Sox2. Also, a recombinant GST fusion
of the Klf4 C-terminus directly bound to bacterially purified recombinant HA-tagged Oct4 and
Sox2 in in vitro pull down assays (Fig. 6F and G). Competitive binding assays performed to
determine whether Oct4 and Sox2 bind to the same domain in Klf4 showed that increasing
amounts of Sox2 protein did not alter Klf4 and Oct4 interaction, supporting the idea that Oct4
and Sox2 do not compete for the same Klf4 binding site (Fig. 6H).
56

Figure 6. The Klf4 C-terminus directly interacts with Oct4 and Sox2.
(A) Design of Klf4 deletions used in the experiments.
(B) Interaction between Klf4 and Oct4 requires zinc finger motifs at the Klf4 C-terminus. Deletion of all three zinc
finger motifs (Klf4ΔZF1-3) significantly decreased the Klf4/Oct4 interaction. Deletion of the C-terminal or two last
C-terminal zinc fingers (Klf4ΔZF3 and Klf4ΔZF2-3, respectively), as well as deletion of the middle region
(Klf4ΔM), did not affect the interaction. The relative quantity of each band was measured and listed below each blot.
(C) Klf4 and Sox2 interaction requires zinc finger motifs at the Klf4 C-terminus. Deletion of all three Klf4 zinc
fingers abolishes Klf4/Sox2 interaction, whereas deleting last two zinc fingers significantly reduces the interaction.
The relative quantity of each band was measured and listed below each blot.
(D and E) The Klf4 C-terminus interacts with Sox2 (D) and Oct4 (E) when over-expressed in 293T cells, while the
N-terminus (Klf4-N) and Middle region (Klf4-M) of Klf4 alone do not.
(F and G) Recombinant GST-Klf4 (300-483) interacts with bacterially-purified recombinant Oct4 (F) and Sox2 (G)
in vitro. Asterisks indicate intact GST-Klf4 (300-483).
(H) Oct4 and Sox2 do not compete for interaction with Klf4. Klf4/Oct4 interaction was not disrupted by increasing
Sox2 expression.
(I) Klf4 and Oct4 co-occupy the Nanog promoter, shown schematically above, in iPS cells. Cross-linked chromatin
was first immunoprecipitated with Oct4 antibody and then with a control IgG or anti-Klf4 antibody. The precipitated
DNA was amplified by PCR, normalized by control IgG, and then normalized by the first pair of primer. Results
indicate that Oct4 and Klf4 co-occupy the Nanog proximal promoter.

We next asked whether a Klf4, Oct4, and Sox2 complex co-occupies a promoter. Sequential
ChIP using Oct4 antibody followed by Klf4 antibody plus PCR analysis using six pairs of
57

primers spanning approximately 1.5kb of the Nanog proximal promoter (Rodda et al., 2005) was
performed on iPS cells (Fig. 6I). The results suggested that Klf4 and Oct4 co-occupy the same
region of the Nanog promoter.
Klf4 mutants compete with wild-type Klf4 to interact with Oct4 and Sox2 and significantly
reduce reprogramming efficiency.
We hypothesize that Klf4 recruits Oct4 and Sox2 through direct interaction and activates
downstream targets required for reprogramming. Since different zinc finger deletions of the Klf4
C-terminus alter its interaction affinities for Oct4 and Sox2, mutants interacting with Oct4 and
Sox2 but lacking transcriptional activation capacity should be able to compete with wild-type
Klf4 when introduced into an inducible reprogramming system and thus serve as dominant
negative constructs. To evaluate transactivation potential of Klf4 mutants, we utilized the Nanog
proximal promoter in a reporter assay and found that Klf4 activates the Nanog promoter (Fig.
7A). None of the C-terminal zinc finger deletion mutants, including Klf4ΔZF3, Klf4ΔZF2-3, and
Klf4ΔZF1-3, activated the promoter, indicating that zinc fingers are critical for target gene
activation (Fig. 7B). Furthermore, electrophoretic mobility shift assay (EMSA) showed that all
zinc finger deletion mutants lacked DNA binding ability (Fig. 7C), confirming that the zinc
fingers are critical for DNA binding.
58

Figure 7. Dominant negative Klf4 mutants compete with wild-type Klf4 for binding with Oct4 and Sox2, resulting
in disruption of reprogramming.

(A) Klf4 can activate the Nanog luciferase reporter in a dose-dependent manner.
(B) Deletion of C-terminal zinc finger motifs abolishes Klf4’s transcriptional activation capacity. Compared to
wild-type Klf4, deleting the last (Klf4ΔZF3), the last two (Klf4ΔZF2-3), or all three zinc finger motifs (Klf4ΔZF1-3)
significantly decreased Nanog-luciferase activity.
(C) EMSA assay shows that only wild-type Klf4 binds DNA, while deletion of the last, the last two, or all three zinc
fingers abolishes DNA binding capability.
(D, E) The dominant negative mutant (Klf4ΔZF3) inhibits binding between wild-type Klf4 with Oct4 (D) or Sox2
(E) by competing with wild-type Klf4. However, Klf4ΔZF1-3, which exhibits low binding affinity for Oct4 or Sox2,
does not compete with wild-type Klf4 to bind Oct4 or Sox2. Cell lysates were immunoprecipitated using anti-HA
antibody followed by immunoblotting with anti- HA, Myc or Flag antibodies. The relative quantity of each band
was measured and listed below each blot.
(F) Dominant negative Klf4 mutants show significantly reduced reprogramming capacity (n=3 ; error bars indicate
s.d. **P<0.01) in genetically homogeneous secondary NSCs. In vitro differentiated NSCs were transduced with
doxycycline induced lentivirus expressing Klf4 mutants. GFP-positive colonies were counted after three weeks of
doxycycline induction. Klf4ΔZF3 and Klf4ΔZF2-3 strongly inhibited reprogramming, whereas Klf4ΔZF1-3 did not
show significant inhibition.
(G) FACS analysis of the SSEA1-positive (Cy-3 labeled) in experiments described in (F). Samples were collected
at day 21. Numbers show percentage of SSEA1 positive cells.
(H) Klf4 mutants inhibit reprogramming in secondary MEFs in a manner similar to NSCs. 5×10
4
secondary MEFs
from chimeric mice were plated per well. Puromycin was used at 0.5μg/ml for 3 days before transduction of new
virus. The total number of GFP-positive colonies in triplicate wells was determined after 12 days of doxycycline
induction (n=3, **P<0.01). Similarly, Klf4ΔZF3 and Klf4ΔZF2-3 strongly inhibited reprogramming, whereas
Klf4ΔZF1-3 did not show significant inhibition.
(I) FACS analysis of the SSEA1-positive population in sample from (H).

59

To further investigate competition between wild-type and mutant forms of Klf4 in interacting
with Oct4 and Sox2, we performed a competition assay between wild-type Klf4 and mutants
lacking only the last or all three zinc fingers, Klf4ΔZF3 and Klf4ΔZF1-3, respectively. As shown
in Fig.7 D and E, Klf4ΔZF3 bound to Oct4 and Sox2 similarly to wild-type Klf4. On the other
hand, Klf4ΔZF1-3 could not interact with Sox2 and only weakly interacted with Oct4, due to
deletion of all three zinc finger motifs. Constructs encoding either of these Flag-tagged Klf4
mutants were co-transfected with constructs encoding Myc-tagged wild-type Klf4 and HA-
tagged Oct4, and the amount of Flag-tagged mutant Klf4 and Myc-tagged wild-type Klf4
associated with HA-tagged Oct4 was determined by immunoprecipitation of Oct4, followed by
immunoblotting with anti-Flag and anti-Myc antibodies. In cells expressing Klf4ΔZF3, Flag-
tagged Klf4ΔZF3 directly bound to Oct4, while the interaction between Myc-tagged wild-type
Klf4 and Oct4 was significantly reduced (Fig. 7D), suggesting that Klf4ΔZF3 significantly
competes with wild-type Klf4. As a control, Klf4ΔZF1-3 could not compete with wild-type Klf4
in binding to Oct4 (Fig. 7D). Similar results were obtained in analyses of Klf4 and Sox2
interactions (Fig. 7E). These results suggest that KlfΔZF3 functions as a dominant negative
mutant of Klf4 in competing with wild-type Klf4 to form complexes with Oct4 and Sox2.
Next, with a dominant negative form of Klf4 (Klf4ΔZF3) in hand, we determined whether or not
the interactions among Klf4, Sox2 and Oct4 were important for iPS cell induction. The same
inducible lentiviral system was used to introduce Flag-tagged mutant or wild-type forms of Klf4
into iPS cell-derived second generation homogeneous NSCs or MEFs described above. High titer
viruses were generated to ensure transduction efficiency of all NSCs or MEFs, and equivalent
transduction efficiency in different samples was confirmed by Flag immunostaining (Fig. 8).
Thus, difference in reprogramming efficiency observed among various samples was assumed to
60

be due primarily to different activities of over-expressed products. Cells from the same passage
of NSCs or MEFs were split and plated at the same density. Klf4ΔZF3- and Klf4ΔZF2-3-
expressing cells showed significantly inhibited reprogramming efficiency: the numbers of Oct4-
GFP-positive colonies and percentage of SSEA1-positive cells in these samples were
significantly reduced compared with cells over-expressing wild-type Klf4 (Fig. 7F and G). To
exclude the possibility that this effect was limited to NSCs, MEFs derived from chimeric mice
were similarly analyzed. Although reprogramming efficiency in MEFs was significantly lower
than in NSCs, we observed a similar inhibition pattern in MEFs as in NSCs (Fig. 7H and I). As
Klf4ΔZF3 or Klf4ΔZF2-3 interact with Oct4 and Sox2 in a manner similar to wild-type Klf4, we
conclude that these mutants compete with wild-type Klf4 to form complexes that cannot activate
transcription, possibly due to the mutant Klf4’s inability to bind DNA. Klf4ΔZF1-3, however,
did not inhibit reprogramming, as GFP-positive colonies and the proportion of SSEA1-positive
cells were similar to those observed in controls (Figure 7F-I). This finding is consistent with the
observation that Klf4ΔZF1-3 cannot interact with Sox2 and only poorly interacts with Oct4.
Thus, there was no competition between Klf4ΔZF1-3 and wild-type Klf4 in forming an
Oct4/Sox2/Klf4 complex.
61

Figure 8. Transduction of neural stem cells with lentiviruses expressing Klf4 mutants.

Immunocytochemistry using anti-Flag antibody (middle panels) and DAPI staining (left panels) of NSCs transduced
with different viruses. Klf4ΔZF2-3, Klf4ΔZF3, and Klf4ΔZF1-3 are Flag-tagged and detected using the Flag
antibody (Sigma). Empty vector FTRE serves as a negative control for Flag staining. Bars, 5 µm. Transduction
efficiency is quantified by counting the Flag-positive cells among DAPI-positive cells (right panel). The
transduction efficiencies are similar among all samples.

Endogenous Klf4 in MEF is required for reprogramming.
Although Klf4 has been shown to be important for reprogramming, protocols were also
developed to induce iPS cells without over-expression of Klf4. These protocols uses expression
of Oct4 and Sox2 with addition of chemicals such as VPA (Feng et al., 2009a; Huangfu et al.,
2008b; Kim et al., 2009b; Lyssiotis et al., 2009). It has been shown that endogenous Klf4 is
expressed in MEF (Lyssiotis et al., 2009; Rowland et al., 2005). It will be very important to
determine if endogenous Klf4 is involved in the Oct4/Sox2/Klf4 complex during reprogramming
in these protocols. To address this, we first tested whether endogenous Klf4 is required for iPS
cell induction from wild type MEFs using only Oct4 and Sox2 as described by Huangfu et al.
62

(Huangfu et al., 2008b). Knock-down of endogenous Klf4 using Klf4 shRNA significantly
decreased the amount of AP-positive colonies compared to controls (Figure 9A and B). This
result indicates that the endogenous Klf4 is required for reprogramming.
To further determine whether endogenous Klf4 facilitates reprogramming by forming complex
with Oct4 and Sox2, we expressed dominant negative mutants of Klf4 together with Oct4 and
Sox2 described above to primary MEFs using the same protocol described by Huangfu et al.
(Huangfu et al., 2008b). When dominant negative mutant Klf4ΔZF3 was over-expressed, AP-
positive colonies was significantly reduced compared to those samples with empty vector control
or over-expression of Klf4ΔZF1-3 (Fig. 9 C, D). These results further confirm that the
Oct4/Sox2/Klf4 complex formation is important for iPS cell induction, in spite of the levels of
Klf4 expression.

Oct4/Sox2/Klf4 complex is required for self-renewal in wild-type mouse ES cells.
The significance of Klf4 in maintaining ES cell self-renewal has been extensively studied
(Jiang et al., 2008). The Klf4/Oct4/Sox2 complex formation is required for somatic cell
reprogramming. It is therefore also important to know if such a complex formation is also
required for self-renewal of ES cells. To address this, we transduced lentiviral constructs
expressing the dominant negative mutant Klf4 and a puromycin resistance gene into mouse ES
cells. After 6 days of puromycin selection, we immunostained the remaining cells with SSEA1
antibody and calculated the percentage of SSEA1 positive cells. Over-expression of Klf4ΔZF3
resulted in almost complete loss of SSEA1 positive cells, while only small part of Klf4ΔZF1-3
63

over-expressed cells differentiated, suggesting that disrupting wild-type Oct4/Sox2/Klf4
complex interferes normal self-renewal of ES cells and causes differentiation (Figure 9E, F).

Figure 9. Disrupting endogenous O/S/K complex by introducing Klf4 dominant negative mutant results in failure of
reprogramming and differentiation in wild type ES cells.
(A) Knock-down of endogenous Klf4 results in a decrease of AP positive colony number compared to control.
1×10
5
primary MEFs were seeded in one well of six-well plates and transduced with lentivirus expressing Oct4,
Sox2, Klf4shRNA or a control vector. VPA was added at the final concentration of 0.5μM. AP staining were
performed 2 weeks after induction of transgenes expression.
(B) Quantification of AP positive colonies in (A) (n=2 independent experiments, **P<0.01).
(C) Primary MEFs over-expressing Klf4ΔZF3 fail to produce AP positive cells, while over-expression of
Klf4ΔZF1-3 has no significant effect compared to control. 2×10
5
MEFs were seeded in each 60mm dish. MEFs
were transduced by lentivirus expressing Oct4, Sox2, and Klf4 mutants. Control samples were transduced with an
empty vector instead of Klf4 mutants. VPA were added at the final concentration of 0.5μM. AP staining were
performed 2 weeks after induction of transgenes expression.
(D) Quantification of AP positive colonies in (C) (n=2 independent experiments, **P<0.01).
(E) Over-expression of Klf4ΔZF3 lead to almost complete differentiation of wild-type mouse ES cells, while a
substantial proportion of cells over-expressing Klf4ΔZF1-3 remain SSEA1 positive. Wild-type mouse ES cells were
transduced with lentiviral vectors containing Klf4 mutant genes followed by IRES and a puromycin resistance gene.
Empty vector served as control. Cells were selected by puromycin for 6 days and subsequently stained for SSEA1.
Bars, 10µm
(F) Quantitative analysis of the percentage of SSEA1 positive cells in (E) revealed a statistically significant
decrease (n=3 independent experiments, *P<0.05, **P<0.01) of pluripotent cells in Klf4ΔZF3 over-expressed cells
when comparing to both control and over-expression of Klf4ΔZF1-3.
64

Overall, deletion of the Klf4 third zinc finger (Klf4ΔZF3) or of the second and third zinc fingers
(Klf4ΔZF2-3), which allows formation of a complex with Oct4 and Sox2 but lacks
transcriptional activity, significantly inhibits Klf4’s activity in iPS cell induction, as well as ES
cell self-renewal. However, Klf4ΔZF1-3, which lacks all three zinc fingers, fails to suppress
reprogramming or inhibit self-renewal. These striking differences in activity of various Klf4
mutants indicate that the Oct4/ Sox2’s binding to Klf4 is critical for reprogramming.
2.5 Discussion
Klf4 has been suggested to play an important role in ES cell self-renewal (Jiang et al., 2008). Its
target genes overlap with those of Oct4, Sox2 and Nanog. Previous research indicated that Oct4
and Sox2 co-localize at target gene promoters (Rodda et al., 2005; Wang et al., 2006). Here, we
found that Klf4 directly interacts with Oct4 and Sox2 through its C-terminus in vitro and in vivo,
which contains three zinc finger motifs. Oct4 and Sox2 do not compete with each other in
binding to Klf4, indicating that each binds to a different site. This result was confirmed by the
fact that loss of two zinc finger motifs significantly decreased Klf4 binding affinity for Sox2,
while the same mutant did not differ from the wild-type protein in terms of interaction with Oct4.
We hypothesize that a complex containing Klf4, Oct4, and Sox2 activates downstream targets
required for reprogramming. Interestingly, Klf4 can activate the Nanog promoter alone in 293T
cells in a transfection assay, while Oct4 or Sox2 alone cannot (data not shown), suggesting that
Klf4 may function as the transactivator in the complex.
To analyze Klf4’s function in reprogramming, we analyzed the ability of mutant forms of Klf4 to
induce reprogramming. We found that Klf4ΔZF3 and KlfΔZF2-3, which interact with Oct4 and
Sox2 but lack DNA binding activity, significantly inhibit normal reprogramming. However,
65

Klf4ΔZF1-3, which lacks all three zinc fingers, fails to suppress reprogramming. The striking
differences between Klf4ΔZF1-3 and Klf4ΔZF3/Klf4ΔZF2-3 activity indicate that Klf4 binding
with Oct4 and Sox2, which is disrupted by deletion of the first zinc finger, is critical for
reprogramming.
Only recently has the mechanism underlying direct reprogramming started to be revealed
(Maherali et al., 2007; Sridharan et al., 2009). Profiling of Oct4, Sox2 and Klf4’s DNA binding
in the whole genome in both ES cells and iPS cells suggested that these 3 factors co-localize in
promoters of essential pluripotency genes (Jiang et al., 2008; Sridharan et al., 2009). However,
whether this co-localization is simply a synergistic effect or involves active recruitment and
assembly of a certain functional protein complex remains elusive. In our model, a complex of
Klf4, Oct4 and Sox2, which are assembled via direct interaction, is likely required to activate
transcription of critical pluripotency genes, such as Nanog. The requirement for complex
formation may also play a regulatory role: not only the quantity but also the stoichiometry of
these factors may determine whether certain amount of complexes sufficient for reprogramming
is available in a single cell. The recent finding that reprogramming is more efficient when factors
are expressed on a polycistronic vector (Okita et al., 2008; Yu et al., 2009) supports this idea.
Such design may guarantee more efficient co-localization of these factors immediately after
translation. It may also ensure equivalent amounts of each factor get produced. These conditions
could facilitate complex formation and lead to successful reprogramming.
The ChIP-on-Chip studies indicate that common targets of Klf4, Oct4 and Sox2 are most
differentially bound by Klf4, Oct4 and Sox2 between fully reprogrammed and partially
reprogrammed iPS cells (Sridharan et al., 2009). This interesting phenomenon suggests that
binding of target gene promoters by Oct4, Sox2 and Klf4 may not be through individual binding
66

but may instead require assembly of a functional complex containing all three factors. Since
these three proteins are abundant even in partially reprogrammed cells (by forced expression),
the formation of a functional complex clearly requires other interacting partners, especially ones
that can help erase the repressive chromatin structures in the target genes. Identifying potential
partners will be interesting and critical for further investigation.
Altogether, our results demonstrate that direct interaction among reprogramming factors Klf4,
Oct4 and Sox2 is required for the success of reprogramming. Klf4 interacts directly with Oct4
and Sox2 and they co-occupy the Nanog promoter. When formation of the normal complex is
disrupted by introducing specific mutants of Klf4, the efficiency of reprogramming is
significantly reduced. Further studies of this interaction should facilitate our understanding of
reprogramming mechanisms.

67

Chapter 3 Interchromosomal interactions modulate Oct4 expression in reprogramming
and pluripotency
3.1 Abstract
The mechanisms of somatic cell reprogramming have been extensively studied (Takahashi and
Yamanaka, 2006; Yu et al., 2007). Whereas most studies have focused on epigenetic changes,
including DNA methylation and histone modifications (Hanna et al., 2009b; Kim et al., 2010;
Lister et al.; Mikkelsen et al., 2008; Silva et al., 2009), very little is known about the nuclear
architecture in pluripotent stem cells (PSCs) and its function in reprogramming and pluripotency.
Here we show that interchromosomal interactions modulate the expression of Oct4 and correlate
with the activation and maintenance of the pluripotency circuitry in reprogramming and self-
renewal. Using circular chromosome conformation capture (4C) and fluorescence in situ
hybridization (FISH), we identified chromosomal domains in trans that co-localize with the Oct4
locus in PSCs. In addition, we discovered that PSC-specific interchromosomal interactions are
established prior to transcriptional activation of endogenous Oct4 during reprogramming. In
PSCs, Oct4-colocalized domains are enriched in active genes and pluripotency factor binding.
Transcription of Oct4 is facilitated when the Oct4 locus is co-localized with its
interchromosomal partners. Finally, we demonstrated that depletion or overexpression of Klf4
causes changes in interchromosomal interactions prior to loss of Oct4 transcription and PSC
differentiation, suggesting that Klf4 regulates interchromosomal interactions independent of its
role as a transcription factor. Taken together, our results delineate a basic level of nuclear
organization at the Oct4 locus in PSCs and elucidate a role for higher-order chromatin structure
in regulating the expression of Oct4 in reprogramming and pluripotency.
68

3.2 Introduction
The discovery of direct reprogramming by overexpression of the transcription factors Oct4, Sox2,
cMyc and Klf4 has transformed the research of pluripotent stem cells (PSCs) and inspired
immense interest in translational research (Takahashi and Yamanaka, 2006; Yu et al., 2007),
with the intention of applying iPSCs for various therapeutic purposes (Wu and Hochedlinger,
2011). Understanding the molecular mechanisms of direct reprogramming is critical in
generating safer and better iPSCs (Hanna et al., 2010b; Orkin and Hochedlinger, 2011).
Numerous studies were undertaken to compare iPSCs with embryonic stem cells (ESCs),
including analyzing genomic and epigenetic changes, and dissecting the dynamics of
reprogramming (Hanna et al., 2009b; Kim et al., 2010; Lister et al.; Mikkelsen et al., 2008; Silva
et al., 2009). Nevertheless, several key questions remain unanswered. One interesting
phenomenon is that, even when using genetically identical cells that equally express
reprogramming factors, only a small fraction of cells or a fraction of the daughter cells from a
single cell can be successfully reprogrammed (Hanna et al., 2009b). A ‘stochastic model’ has
been proposed in which cells need to express reprogramming factors and undergo specific
epigenetic changes to overcome an epigenetic ‘bump’ and reach the pluripotent state (Plath and
Lowry, 2011; Yamanaka, 2009). However, the molecular foundation of this ‘bump’ remains
largely unknown.
A key observation from nuclear reprogramming in pioneer nuclear transfer experiments is that
nuclear enlargement and chromatin decondensation precede alterations in gene expression
(Gurdon and Melton, 2008). Given the fact that the nuclear architecture of PSCs and somatic
cells are divergent (Jaenisch and Young, 2008; Meshorer et al., 2006), somatic cells must
undergo enormous changes in higher-order chromatin structure during reprogramming. Since
69

higher-order chromatin structure has been suggested to be involved in genome-wide
transcriptional activity (Chakalova and Fraser, 2010), such changes in nuclear architecture
during reprogramming may act as a significant driving force to transform the epigenome and
transcriptome and eventually lead to the establishment of induced pluripotency. The recent
development of chromosome conformation capture (3C)-based methods has enabled detection of
long-range interactions and discovery of novel interactions both locally and in a genome-wide
fashion (Dekker et al., 2002; Fullwood et al., 2009; Lieberman-Aiden et al., 2009; Simonis et al.,
2006; Zhao et al., 2006). Long-range interactions that recruit distant enhancers in cis and in trans
have various biological functions (Williams et al., 2010). Long-range intrachromosomal
interactions in PSCs have been described for the Nanog and the Gata4 loci, which are important
for self-renewal and differentiation, respectively (Levasseur et al., 2008; Tiwari et al., 2008).
Nonetheless, the biology of higher-order chromatin structure in reprogramming is mostly unclear.
In the present study, we first apply circular chromosome conformation capture (4C) and
fluorescence in situ hybridization (FISH) to identify a group of chromosomal regions both in cis
and in trans that co-localizes with Oct4 in PSCs specifically. Furthermore, analysis of these co-
localizations during reprogramming reveals that establishment of interchromosomal interactions
at Oct4 locus arise specifically in a subset of cells and before activation of endogenous Oct4.
These long-range interactions facilitate Oct4 transcription in PSCs. Interestingly, the mRNA of
several genes in close proximity to Oct4 locus are enriched in PSCs and involved in the
maintenance of PSC self-renewal and pluripotency. More importantly, Oct4 and its interacting
partners are regulated by reprogramming factor Klf4. Klf4 is also co-localized with the long-
range interacting sites and its depletion leads to dissociation of these interchromosomal
interactions. Taken together, we have discovered a set of novel Klf4-mediated interchromosomal
70

interactions involving the Oct4 enhancer, and further present evidence of critical roles for these
interactions in regulating Oct4 expression in PSCs by activating Oct4 during reprogramming.
3.3 Methods and Materials
Cell culture
E14 ESCs and iPSCs were cultured in traditional ES medium (GMEM with 15% FBS, L-
glutamine, penicillin-streptomycin, non-essential amino acids, β-mercaptoethanol and 1,000
U/ml LIF) on gelatin-coated dishes or on irradiated MEF cells. NSCs were cultured as
monolayers in NSC medium (DMEM/F12 with 2% B27, 20 ng/ml β-FGF and 10 ng/ml EGF) on
poly-L-lysine and fibronectin-coated dishes. In vitro differentiation of NSCs from iPSCs was
performed as described previously (Wei et al., 2009). Reprogramming of secondary NSCs was
performed as described previously (Wei et al., 2009). Briefly, NSCs were seeded in NSC
medium on coverslips and doxycycline was added at a concentration of 1 μg/ml for 2 days. The
medium was then replaced by ES medium with daily addition of doxycycline. pre-iPSCs were
isolated using a protocol similar to a previously described protocol (Mikkelsen et al., 2008), and
were cultured in ES medium with doxycycline. Reprogramming from pre-iPSCs to iPSCs was
performed as described previously (Mikkelsen et al., 2008).
4C assay
4C was performed using a modified protocol adapted from a previous 3C-qPCR protocol. Briefly,
10
7
ESCs were trypsinized to single cells and resuspended in 0.5 ml GMEM/10% FBS. 9.5 ml of
2% paraformaldehyde/10% FBS was then added and the cells were incubated for 10 min. After
quenching by glycine, the cells were lysed in 5 ml lysis buffer (10 mM Tris-HCl, pH7.5; 10 mM
NaCl; 5 mM MgCl
2
; 0.1 mM EGTA; 1X protease inhibitor) for 10 min on ice, then centrifuged
71

to remove the supernatant. Nuclei were then incubated in 0.3% SDS for 0.5 hr, 2% triton for 1 hr,
and 400U Csp6I restriction enzyme overnight, all at 37ºC. After inactivation by 1.6% SDS at
65ºC for 20 min, samples were diluted in 6.125 ml of 1.15x ligation buffer and 100U T4 ligase
and incubated at 16ºC for 3 days with addition of ATP every day. The ligated chromatin was
digested by proteinase K, purified by phenol-chloroform extraction, and the DNA precipitated by
ethanol precipitation. 100 ng of DNA was used for nested PCR amplification with the primers in
Table 4. PCR products were run on a 1.5% agarose gel and DNA from 100-500 bp was extracted
and purified and ligated into PCR2.1 vectors. Individual clones were then digested to check the
insertion size. Insertions with different sizes were sent for sequencing. Only sequences with
intact bait sequences were considered and mapped to genomic locations.
4C-seq data analysis
The 91bp paired-end sequencing data from Illumina HiSeq platform was processed as follows.
The first 24bp long tags from the raw fastq files of paired-ends were trimmed and aligned to
mm9 assembly separately using Burrows-Wheeler Aligne. Paired tags uniquely mapped to the
mouse genome were processed to select paired tags with one tag aligned to the forward nested
primer used for 4C library construction, and the other tag aligned to different chromosomes or
regions on chr17 that are distal from the bait primer (20kb away). One additional filter was used
to remove redundant tags aligned to the same location. Finally the tags that are within 200kb
genomic distance to each other were merged, and the count of tags per 1MB genomic length in
each large domain formed is treated as interaction frequency of individual domain with the bait.
When the signal density of a LID is higher than twice the genome average signal density, it is
considered as a positive LID. mRNA-seq and ChIP-seq data in ESCs are from a previous study
72

(Brookes et al., 2012). Here, RNAPII-S5P is referred to as S5P (TSS), and RNAPII-S2P is
referred to as S2P (TES) according to the data set from Brookes et al. (Brookes et al., 2012)
Genes are defined as positive/bivalent/inactive based on previous definitions using available
RNAPII, PRC mapping and RNA-seq data (Brookes et al., 2012). To simplify the definition,
genes defined as ‘active’ or ‘PRC active’ in the previous study are considered here as ‘active’.
‘PRC intermediate’, ‘PRC repressed’ and ‘PRC only’ genes are considered as ‘bivalent’.
‘Inactive’ genes in Brookes et al. carry the same designation here. Genes without available ChIP-
seq data (labeled as ‘NA’ in the previous study) are excluded.
Immunofluorescence
For GFP, Oct4 and RNAPII-S5P staining, cells were crosslinked with 4% paraformaldehyde for
15 min, and washed in 0.2% PBS. Cells were then permeabilized and blocked by incubation in
0.2% Triton X-100, 10% normal goat serum and 3% BSA in PBS for 1 hr, followed by
incubation with the 1
st
antibody in 0.2% Triton and 3% normal goat serum in PBS overnight at 4
ºC. After washing with PBS three times, the cells were incubated with the 2
nd
antibody for 1 hr,
washed and stained with DAPI. SSEA1 staining was performed using similar methods except
without addition of Triton X-100.
For Klf4 and RNAPII-S5P double staining, cells were first incubated with anti-Klf4 antibody
using blocking buffer with normal donkey serum, followed with donkey anti-goat 2
nd
antibody
conjugated with Cy2. Cells were then incubated with anti-RNAPII-S5P antibody using blocking
buffer with normal goat serum, and finally with goat anti-mouse 2
nd
antibody conjugated with
Cy3.
73

Antibodies used in this study and their dilutions were: anti-GFP (Aves Labs, GFP-1020, 1:300),
anti-Oct4 (Santa Cruz, SC-5279, 1:50), monoclonal anti-RNAPIIS5P CTD4H8 (Millipore, 05-
623, 1:1000), anti-Klf4 (R&D, AF3158, 1:20), anti-SSEA1 (Hybridoma bank, 1:400), Cy3 goat
anti-mouse IgG (Jackson ImmunoResearch, 1:400), fluorescein goat anti-chicken IgY (Aves
Labs, 1:200), Cy2 donkey anti-goat IgG (Jackson ImmunoResearch, 1:400), Alexa488 goat anti-
mouse IgM (Invitrogen, A-21042, 1:500), Alexa 350 goat anti-mouse IgG (Invitrogen, A-21049,
1:400).
shRNA knockdown
shRNA constructs are: Klf4 shRNA (TRIPZ-based vectors, miRNA sequences from
V3LMM_469914, V3LMM_524009 were moved from GIPZ vectors to TRIPZ vectors, Open
Biosystems), Llgl2 shRNA (GIPZ-based vectors, V2LMM_25726, V2LMM_31673, Open
Biosystems), Grb7 shRNA (pLKO.1-based vectors, TRCN0000097206, TRCN0000097208,
Open Biosystems), Unc84a shRNA (pLKO.1-based vectors, TRCN0000126979,
TRCN0000126981, Open Biosystems). For GIPZ-based vectors, 6-10×10
7
ES cells were
transfected with 8µg DNA using a Nucleofactor kit for mES (Lonza) on a Nucleofactor I device
(Lonza). Cells were fixed 48 hr after transfection or FACS sorted after 48 hr for GFP positive
cells. For pLKO.1-based vectors, ESCs were transfected, incubated overnight and treated with
1µg/ml puromycin for 48 hr. Controls are non-target shRNA construct from Sigma (for pLKO
constructs) and from Openbiosystems (for pGIPZ constructs).
Generation of stable ES cell lines
For inducible Klf4 knockdown, shRNA was expressed from a pTRIPZ vector (Open Biosystems).
ESCs were transduced with lentivirus and plated as single cells in 96-well plates. Puromycin was
74

added at a concentration of 1µg/ml to select the positive clones. Doxycycline was added at a
concentration of 5 µg/ml for induction, and the medium were changed daily. For inducible Klf4
overexpression, the turboRFP-shRNAmir of TRIPZ was replaced by Klf4-IRES-GFP. The
transduction and selection procedures were similar to those used for the shRNA lines.
Doxycycline was added at a concentration of 2 µg/ml.
Quantitative PCR
Total RNA was isolated from cells using an RNAeasy mini kit with in-column DNase I digestion
(Qiagen). 0.1-1μg of RNA was subjected to cDNA synthesis using an M-MuLV first-strand
cDNA synthesis kit (NEB). Real-time quantitative PCRs were set up in duplicate using iQ SYBR
Green Supermix (Bio-Rad) and run on a CFX96 real-time PCR detection system (Bio-Rad).
Primers are listed in Table 4.
Chromatin immunoprecipitation
10
7
ESCs or iPSCs were fixed in 2% formaldehyde for 10 min and then lysed in lysis buffer (1%
SDS, 10 mM EDTA, 50 µM Tris-HCl, pH8.1, proteinase inhibitor) for 30 min on ice. The lysate
was then split into ~100 µg per sample and sonicated 4 times for 2 min, 15s on and 15s off. The
supernatant was pre-cleared for 1 hr with agarose beads in 1ml dilution buffer (0.01% SDS, 1.1%
Triton X-100, 1.2 mM EDTA, 16.7 mM Tris-HCl, pH8.1, 150 µM NaCl). 2 µg of antibody and
10 µl of agarose beads were then added and incubated overnight. Washing and elution was
performed as described previously. For ChIP using cells after FACS sorting, ~10
5
SSEA1
positive or negative cells were lysed initially and ~10 µg of chromatin was used for each sample.

75

DNA fluorescence in situ hybridization (FISH)
In-house probes representing various regions were produced from BAC probes listed in Table 3.
Briefly, BACs were amplified by using a GenomiPhi DNA Amplification Kit (GE Healthcare)
and labeled with in-house-conjugated Alexa488-dUTP, Cy3-dUTP, or Cy5-dUTP by using a
Nick Translation Mix (Roche).
DNA FISH was performed by using a previously described protocol (Chaumeil et al., 2008).
Briefly, cells were seeded on coverslips, directly fixed with 4% formaldehyde and permeabilized
with 0.5% Triton X-100 in PBS on ice for 7 min. Samples were then denatured in 50%
formamide/2× SSC (pH 7.2) for 30 min at 80 ºC. 10 µl of hybridization mix containing 0.1 µg of
probes was then added and hybridized overnight at 42 ºC. After washing with 50%
formamide/2×SSC and 2×SSC, three times each, the samples were stained with DAPI. Images
were collected using an LSM510 with 3D z-stacks at 0.5µm intervals. The two probes were
considered co-localized when the two FISH signals overlapped or were juxtaposed (distance
between the two centers of the signals ≤450 nm). For co-localization efficiencies in two-color
FISH, 150-250 nuclei were analyzed.
Triple-label immuno-FISH
SSEA1 and RNAPII-S5P staining was performed first as indicated in the immunofluorescence
section. The samples were then post-fixed with 3% paraformaldehyde for 10 min and washed
twice with 2×SSC. After permeabilization in 0.1M HCl/0.7% Triton X-100 for 10 min on ice, the
samples were washed twice with 2×SSC and then denatured and hybridized as described for
DNA FISH. To measure the RNAPII-S5P intensity, the images were processed by ImageJ and
76

the normalized intensity of a specified locus was determined by the median intensity of the locus
divided by median staining intensity of the same nucleus.
RNA FISH and RNA/DNA FISH
RNA FISH was performed by following previous protocols. Briefly, an intron sequence of Oct4
was amplified and cloned into pBluescript II-KS (see Table 4 for primer sequences). Linearized
vectors were in vitro-transcribed and RNA was further reverse-transcribed with DIG-labeled
dUTP (Roche). After RNase H digestion, the single-strand DIG-cDNA probe was precipitated,
resuspended in formamide, denatured for 7 min at 75 ºC and mixed with 2×hybridization mix
(4×SSC, 20% dextran sulfate, 2 mg/ml BSA, and 40 mM Vanadyl Ribonucleoside Complex) .
Cells were permeabilized in CSK buffer (200 mM NaCl, 300 mM sucrose, 3 mM MgCl
2
, and 10
mM PIPES, pH 6.8) /0.5% Triton X-100 containing 2 mM Vanadyl Ribonucleoside Complex.
After the samples were fixed in 4% paraformaldehyde for 10 min and dehydrated with 70%, 80%,
95% and 100% ethanol, the denatured probes were deposited on the sample and hybridized at 37
ºC overnight. After washing with 50% formamide/2×SSC and 2×SSC, three times each, the
samples were quenched with 3% H
2
O
2
for 1 hr, and then blocked by 0.5% blocking powder
(Perkin Elmer) in TST (0.1M Tris-HCl, pH7.5, 0.15 M NaCl, 0.05% Tween-20) for 3 hr at 25 ºC.
Anti-DIG-POD antibody (Roche, 1:200) was applied to the samples and incubated at 4 ºC
overnight. Samples were then washed with TST buffer, and TSA-fluorescein (Perkin Elmer,
1:100 in amplification diluents) was applied to the sample for 10 min at 25 ºC. After washing in
TST three times, 10 min each, the samples were stained with DAPI and mounted.
77

For RNA/DNA FISH, RNA FISH was performed first. After post-fixation with 4%
paraformaldehyde, the cells were denatured in 70% formamide /2×SSC for 4 min at 75 ºC. DNA
FISH was performed afterwards.
Statistical analyses
One-way ANOVA with a nonparametric test was used to compare sizes of nuclei. Paired two-
tailed t-tests were applied in all other statistical analyses. The Wilcoxon single-rank test was
used in Figure 3a. The chi-square test was used in Fig.3c. Two-tailed t-tests were used for Fig.
2c,d and Fig. 5f. All other statistical analyses used Fisher’s exact test.
3.4 Results and discussion
The nuclear architectures of PSCs and somatic cells are divergent (Jaenisch and Young, 2008;
Meshorer et al., 2006), suggesting that somatic cells must undergo enormous changes in higher-
order chromatin structure during reprogramming. Oct4 is an essential player in the pluripotency
circuitry and in direct reprogramming (Kim et al., 2009b). Either by direct ectopic expression or
indirect activation, Oct4 has been shown to be indispensible during late stages of reprogramming
(Anokye-Danso et al., 2011; Heng et al., 2010; Kim et al., 2009b; Plath and Lowry, 2011). Thus
we focused on the dynamics of higher-order chromatin structure of the Oct4 locus. To identify
potential long-range interactions involving the Oct4 locus, we performed a modified circular
chromosome conformation capture (4C) assay in embryonic stem cells (ESCs) (Fig. 10a). We
focused on the distal enhancer (DE) region of Oct4 (Fig. 10b), located ~2 kb upstream of the
transcription start site and responsible for Oct4 activation in ESCs (Pan et al., 2002). Previous
ChIP-seq data revealed binding of Oct4, Sox2, and Klf4 in this region (Chen et al., 2008). Thus
we applied the 4C assay in ESCs using the DE region as bait and amplified a group of unknown
78

sequences. We first sequenced more than 200 individual clones by Sanger sequencing, which
revealed a wide range of sequences located both in cis and in trans. To verify that these potential
interactions exist in vivo, we selected 17 loci and performed DNA FISH on ESCs using bacterial
artificial chromosome (BAC)-derived probes containing these loci together with a BAC probe
containing the endogenous Oct4 locus (Fig. 11a, b. The full names of the BAC probes and their
locations are listed in Table 3). The majority of the candidate regions (9 of 17) co-localized with
Oct4 in about 5-12% of the cells (Fig. 11b, black bars), whereas 3 control probes (representing
active, bivalent and inactive loci) were found to co-localize with Oct4 in less than 2% of cells
(Fig. 11b, white bars).

Figure 10. Schematic show of 4C assay and bait region.
a. Schematic show of experimental procedures of 4C
b. Schematic show of Oct4 DE region. Modified from UCSC genome browser.
79

Figure 11. Long-range interactions between Oct4 and various loci reveal unique higher-order chromatin structure of
the Oct4 locus in PSCs.
a, DNA FISH of Oct4 and various loci in ESCs. b, Summary of percentage of interchromosomal co-localization of
Oct4 locus and candidate loci identified by 4C in ESCs. Controls were performed using a BAC of an unrelated
region instead of Oct4. p=0.0004. The p value represents differences in co-localization frequencies between Oct4-
candidates and control-candidates. c, Images of co-localization of the Oct4 locus and its interacting partners, Llgl2
and Grb7, in ESCs, iPSCs, pre-iPSCs and NSCs. d, e, f, Co-localization percentages of Oct4-Llgl2 (d), Oct4-Grb7
(e) and Oct4-Unc84a (f) in ESCs, iPSCs, pre-iPSCs and NSCs. g, Sizes of nuclei measured by DAPI staining in
NSCs, iPSCs, ESCs, and pre-iPSCs. Error bars represent s.e.m. (***p<0.0001).
We then repeated the 4C experiment and generated a library for high-throughput DNA
sequencing (4C-seq). Nearby unique interacting sites were merged to produce 2150 large
80

interacting domains (LIDs). We found that the positive regions verified by FISH are mostly
located in LIDs with higher signal densities (represented by red dots in Fig. 11c). In contrast, 3
‘negative’ probes are either located in LIDs with low signal density (315L23 and 102A1,
represented by black dots in Fig. 11c) or do not overlap with any LIDs (304A22). This
demonstrates that by combining 4C-seq and FISH we were able to identify regions that co-
localized with Oct4 loci.
To investigate whether the long-range interactions observed at the Oct4 locus are specific to
PSCs, we examined the co-localization frequency in iPSCs, ESCs, pre-iPSCs (or partial iPSCs)
and neural stem/progenitor cells (NSCs) derived from in vitro differentiation or from embryonic
brain tissue. Several of the long-range interactions were found to be specific to iPSCs and ESCs
only, but not to pre-iPSCs or NSCs (Fig. 11d-g, Fig. 12), indicating that these long-range
interactions are specific to PSCs. Nuclear staining by 4',6-diamidino-2-phenylindole (DAPI) in
various cell types excludes the possibility that the higher co-localization frequency in PSCs is a
consequence of these cells having a smaller size nucleus (Fig.13).
81

Figure
12. NSCs from different origins are similar in co-localization efficiency
a. Images of DNA FISH in NSCs isolated from E14.5 brain tissue
b. Co-localization efficiency between Oct4 loci and 143F14 probe in NSCs from in vitro differentiation and
primary culture.

Figure 13. Sizes of nuclei measured by DAPI staining in NSCs, iPSCs, ESCs, and pre-iPSCs. Error bars represent
s.e.m. (***p<0.0001).

82

Next we investigated how cells establish PSC-specific, long-range interactions during
reprogramming. Previously we established a doxycycline-inducible reprogramming system to
study the mechanisms of direct reprogramming (Wei et al., 2009). Upon doxycycline induction,
iPSC-derived NSCs can be reprogrammed back into 2
nd
-generation iPSCs within 2 weeks (Wei
et al., 2009), with a gradual increase in the percentage of cells carrying the pluripotency surface
marker SSEA1 (Fig. 14a, b). GFP, which is driven by the Oct4 promoter and distal enhancer,
begins to appear after day 13, indicating that endogenous Oct4 is activated after approximately 2
weeks (Fig. 14c). This is confirmed by quantitative PCR (qPCR) showing that the level of
endogenous Oct4 remains low at day 8 (Fig. 14d).

Figure 14. Characterization of the secondary reprogramming system.
a. Images of SSEA1+ cells (green) during different stages of reprogramming. Scale bar, 50µm.
b. Quantitation of percentage of SSEA1+ cells during reprogramming.
c. Bright field and GFP images of reprogramming NSCs at day1, day 5, day 8 and day 13 after addition of
doxycycline. GFP cannot be observed in day 1, day 5, and day 8, and only start to be expressed weakly in a
small proportion of cells at day 13.
d. Expression of endogenous Oct4 during reprogramming and in PSCs.

83

To decipher when NSCs begin to form pluripotent-specific long-range interactions, we collected
samples at day 1, 5, 8, and 13 and applied immuno-FISH to examine the co-localization
frequency between Oct4 and 2 other loci, 143F14 and 474J5, in SSEA1+ cells and SSEA1- cells.
Interestingly, in contrast to endogenous Oct4 expression, which remained low at day 8 (Fig. 14d),
co-localization levels began to rise as early as day 5 (Fig. 15a-d). This observation indicates that
long-range interchromosomal interactions involving Oct4 begin to form in the early stages of
reprogramming and, more importantly, the co-localizations take place before actual activation of
the endogenous Oct4 gene occurs. Moreover, the co-localizations happened mostly in SSEA1+
cells (Fig. 2c,d). Since mature iPSCs arise from the SSEA1+ population, this indicates that long-
range interactions occur specifically in the progenitors of mature iPSCs.
To determine whether the dynamics during the transition from pre-iPSCs to iPSCs are similar to
those during normal reprogramming, we examined the dynamics of co-localization during
reprogramming from pre-iPSCs to mature iPSCs (Fig. 2e,f). After applying the DNMT1 inhibitor
AZA, it took about 5 days to reprogram all pre-iPSCs into mature iPSCs, as shown by GFP
expression (Fig. 15e,f). However, the extent of interchromosomal co-localization right after
addition of AZA at day 1 rose to a level similar to that in PSCs (Fig. 2f). The co-localization can
also be observed in GFP-negative cells (Fig. 15e). Again, in this transition, initiation of long-
range interchromosomal interactions precedes activation of endogenous Oct4.
84

Figure 15. PSC-specific interchromosomal interactions of Oct4 loci are established early in reprogramming.
a-d, Representative images (a, b) and quantification (c, d) showing co-localization between Oct4 and its interacting
partners 143F14 (a, c) and 474J5 (b,d), with enrichment only in the SSEA1+ population. Scale bar, 5 µm. e, DNA
FISH images of co-localization during AZA treatment of pre-iPSCs. Insets show an enlarged single nucleus. Scale
bar, 50 µm. f, Statistical summary of the co-localization frequency during reprogramming. Error bars represent
standard deviation with n=3 biological replicates in c,d and f. *p<0.05, **p<0.01.

The function of higher-order chromatin structure varies, from organizing transcription factories
to maintaining condensed heterochromatin (Chakalova and Fraser, 2010; Probst and Almouzni,
2011). By combining our 4C-seq and published mRNA-seq results (Brookes et al., 2012), we
found that the genes located in the positive Oct4-interacting LIDs are expressed at a higher level
than genes located outside positive LIDs in ESCs (Fig. 16a). Similarly, a higher than average
percentage of genes were found to be positive for active gene markers such as RNA polymerase
II (RNAPII) S5P, RNAPII S2P and H3K4Me3, while the percentage of Suz12 positive genes
remain unchanged (Fig. 16b). When we classified all the genes into three categories— active,
85

bivalent and inactive —based on the published genome-wide profiling data (Brookes et al.,
2012), 60% of the genes in positive LIDs are classified as ‘active’, while only 45% of genes in
the whole genome are considered ‘active’ (Fig. 16c).

86

Figure 16. Interchromosomal interactions are correlated with endogenous Oct4 transcription in PSCs.
a, Expression of all Refseq genes and genes located in the positive 4C-LIDs in ESCs. After filtering out technical
noise which FPKM<1, Log
10
(FPKM) of 4788 genes in the 4C positive LIDs and 5512 genes not located in LIDs are
represented as boxplots, with the original data from Brookes et al.
19
b, percentage of genes with RNAPII S5P,
RNAPIIS2P, H3K4Me3 and H3K27Me3 modification in ESCs. c, The proportion of active, bivalent and inactive
state of all genes and 4C-LID genes (p<0.0001). d-f, Images of triple-labeled RNA/DNA FISH of Oct4 RNA
(green), Oct4 DNA (yellow), and DNA probe 143F14 (red) (d), 474J5 (red) (e), or 264I8 (red) (f) in ESCs. Side
panels show enlarged images of the co-localized signals (top to bottom: yellow, red, green, yellow and green, red
and yellow, triple label). g-i, Percentage of Oct4 loci which are positive for RNA in triple-labeled RNA/DNA FISH.
The percentage of 143F14-interacting Oct4 loci (g), 474J5-interacting Oct4 loci (h), or 264I8-interacting Oct4 loci (i)
are compared to the percentage of the non-interacting Oct4 allele in the same cell or to the average percentage of all
cells. n indicates the total number of nuclei analyzed in 2 biological samples. (**p<0.01).

87

To further test this hypothesis at the single-cell level, we performed both RNA/DNA FISH and
triple-label immuno-DNA FISH in ESCs. For RNA/DNA FISH, a probe targeting the first intron
of Oct4 served as a marker of Oct4 transcription. With this probe we were able to detect Oct4
primary transcript in ~40-50% of ESCs (Fig. 17), a proportion similar to that of a previous study
of Rex1 (also known as Zfp42) expression in ESCs (Hiratani et al., 2010). When we combined
RNA FISH of Oct4 with DNA FISH of Oct4, 143F14, 474J5 and 264I8, we found that 66-80%
of the 143F14-, 474J5- and 264I8-interacting Oct4 alleles were positive for Oct4 RNA FISH (Fig.
16d-i). This is in sharp contrast to the other Oct4 allele in the same cell that is not co-localized
with 143F14, 474J5 or 264I8, of which only ~40% were positive for Oct4 RNA FISH.
Considering that Oct4 is very likely to have multiple long-range interacting partners, the 40%
transcriptionally active Oct4 alleles in the controls may represent Oct4 loci that are involved in
other long-range interactions. In addition, we performed triple-label immuno-DNA FISH using
an antibody that recognizes RNAPII-S5P. RNAPII-S5P staining intensity at most Oct4 loci was
found to be above the median intensity of the same nucleus. Notably, the Oct4 loci participating
in long-range interactions exhibited particularly high RNAPII-S5P intensity when compared with
“non-interacting” Oct4 loci (Fig. 18). Together these results suggest that the Oct4 loci that are
involved in long-range interactions are much more likely to be in an active transcriptional state.

88

Figure 17. Detection of Oct4 primary transcript by RNA FISH in ESCs.
a. An overview of ESCs with FISH signals (red) and DAPI.
b- d. Individual examples of single nucleus with mono-allelic (b), bi-allelic (c), and non-detectable (d) RNA
FISH signals.

e. Percentages of nuclei with bi-allelic, mono-allelic, or non-allelic transcription of Oct4. **p<0.01.

Figure 18. Enrichment of RNAPII-S5P in Oct4-474J5 and Oct4-280I15 interacting sites.
a. Images of Triple label immuno-DNA FISH, including DNA FISH of Oct4 (red) and 474J5 (green), as well
as RNAPII-staining (blue). Side panels are enlarged interacting sites (top to bottom: green, red, blue, triple
label).
b. Normalized median values of RNAPII-S5P intensity at various Oct4 loci are measured. Control groups are
Oct4 loci which do not have interactions with 474J5. **p<0.01.
c. Images of Triple label immuno-DNA FISH, including DNA FISH of Oct4 (red) and 280I15 (green), as well
as RNAPII-staining (blue). Side panels are enlarged interacting sites (top to bottom: green, red, blue, triple
label).
d. Normalized median values of RNAPII-S5P intensity at various Oct4 loci are measured. Control groups are
Oct4 loci which do not have interactions with 280I15. *p<0.05.

89

The bait used in our 4C assay, the DE region of Oct4, contains Oct4, Sox2 and Klf4 binding sites
(Chen et al., 2008) (Fig. 10b). Interestingly, 4C positive LIDs are enriched in Klf4 binding sites,
followed by Oct4, Sox2 and Nanog sites (Fig. 19). This raises an interesting point, since another
Sp/Klf family member, EKlf/Klf1, has been studied extensively for its role in regulating long-
range interactions at β-globin loci (Drissen et al., 2004; Schoenfelder et al., 2010). Moreover,
Klf1 is able to replace Klf4 in reprogramming, though at a lower efficiency (Nakagawa et al.,
2008), suggesting at least partial functional redundancy between Klf4 and Klf1.

Figure 19. Enrichment in pluripotency factor binding sites in 4C positive LID regions.
The enrichment represent ratio between binding sites density in 4C positive LIDs and that in 10,000 randomly
generated regions with 1MB length.

Figure 20. Characterization of the potential Klf4 target gene.
a. Chromatin immunoprecipitation of Klf4 in ESCs and iPSCs shows occupancy by Klf4 in the promoters of
Oct4, Llgl2, Grb7 and Unc84a.
b. Depletion of Klf4 by shRNA leads to an immediate reduction of expression of Llgl2, Grb7 and Unc84a.
90

Figure 21. Characterization of the role of Llgl2, a gene located in 4C enriched region, in PSCs self-renewal.
a. qPCR of the expression of Llgl2 in ESCs, iPSCs, NSCs and MEFs.
b. qPCR showing changes in expression of several pluripotency-related genes after depletion of Llgl2.
c. Alkaline phosphatase (AP) staining of ESCs with knockdown of Llgl2.
d. Statistics of single-cell survival by number of AP positive colonies after Llgl2 knock down by shRNA. All
error bars indicate standard deviations with n=2 in a,b and n=3 in d.

Figure 22. Characterization of the role of Grb7, a gene located in 4C enriched region, in PSCs self-renewal.
a. qPCR of the expression of Grb7 in ESCs, iPSCs, NSCs and MEFs.
b. qPCR showing changes in expression of several pluripotency-related genes after depletion of Grb7.
c. Alkaline phosphatase (AP) staining of ESCs with knockdown of Grb7.
d. Statistics of single-cell survival by number of AP positive colonies after Grb7 knock down by shRNA. All
error bars indicate standard deviations with n=2 in a,b and n=3 in d.
91

Figure 23. Characterization of role of Unc84a in PSCs differentiation potential.
a. qPCR of Unc84a expression after shRNA depletion.
b. qPCR showing changes in expression of self-renewal genes after Unc84a depletion
c-h. qPCR of expression of lineage specific markers during embryonic body differentiation day0, day3, day6,
day10 in non-specific shRNA lines and Unc84a shRNA lines. Lineage specific genes include ectoderm marker
Fgf5 (c) and Nestin (d), mesoderm marker T (e) and Pax3 (f), and endoderm marker Gata4 (g) and Gata6 (h).
All error bar indicate standard deviation with n=3.

92

Figure 24. Immuno-staining indicates that Klf4 protein (green) form punctuated structure which overlay with
RNAPII-S5P (red) inside nucleus.
To determine Klf4’s function in mediating long-range interactions, we first identified three
representative Klf4 targets, Llgl2 (located in 143F14), Grb7 (located in 474J5), and Unc84a
(located in 264I8), in the Oct4-interacting regions (Fig. 20), all of which are highly expressed in
PSCs (Han et al., 2011a; Heng et al., 2010; Mikkelsen et al., 2008; Sridharan et al., 2009;
Takahashi and Yamanaka, 2006). Further characterization demonstrates that Llgl2 and Grb7 are
required for PSC self-renewal (Figs. 21 and 22), whereas Unc84a is essential for differentiation
(Fig. 23).
It has been demonstrated that Klf1 is located in discrete sites and overlaps with RNAPII-S5P foci,
which function as specialized transcription factories. Interestingly, immunostaining of Klf4
shows a similar discrete localization pattern (Fig. 24). Most of these Klf4 foci also overlap with
RNAPII-S5P foci (Fig.24). To confirm that the Klf4-occupied genes are indeed located inside
these Klf4 foci, we performed immuno-FISH to detect simultaneously Klf4 protein and Oct4,
143F14, 474J5 and 264I8 in PSCs. All these loci were found to be associated with Klf4 foci at a
much higher frequency than “background association” (Fig. 25a, b). More importantly, when the
Oct4 locus was co-localized with its partner-loci 143F14, 474J5, and 264I8, its frequency of
association with Klf4 foci was further increased (Fig. 25 c, d). Together, these results strongly
93

suggest that the interactions between Oct4 loci and its interacting partners occur inside Klf4-
enriched granules, in which Klf4 and RNAPII-Ser5 constitute specialized transcription factories.

94

Figure 25. Klf4 is essential for organizing the interchromosomal interactions of the Oct4 locus.
a, Immuno-FISH showing co-localization of Klf4 protein (green) and genomic loci, including the control locus
102A1 (red) and 143F14, Oct4, 474J5 and 264I8 (red). Side panels show enlarged images of co-localization (top to
bottom: red, green, double labeling). b, Graph shows percentages of various loci co-localized with Klf4 protein
(black) or not co-localized with Klf4 protein (white). The red line shows ‘background’ co-localization percentage. c,
Triple-labeled immuno-FISH of Klf4 protein (green) and co-localized Oct4-264I8 or Oct4-474J5 (yellow and red, as
indicated). Side panels show enlarged co-localization signals (top to bottom: red, yellow, red and yellow, green,
triple label). d, Graph shows the interacting pairs (143F14-Oct4, 474J5-Oct4, and 264I8-Oct4) co-localized (black)
or not co-localized (white) with Klf4 foci. e, schematic illustration of the experimental procedure in (f). f, Ratio of
Klf4 enrichment in SSEA1+ cells to that in SSEA1- cells at Oct4, Llgl2 (located in 143F14), Grb7 (located in 474J5)
and Unc84a (located in 264I8) loci at reprogramming day 5 (white bar) and day 10 (black bar). g-h, qPCR of
expression of Klf4 (g) and Oct4 (h) upon depletion of Klf4 by shRNA. i, Percentage of Oct4 alleles with detectable
nascent RNA signal in control and Klf4 knockdown samples. All 3 lines were treated with doxycycline 5 µg/ml for
48 hr. j-l, Frequencies of co-localization of Oct4-143F14 (j), Oct4-474J5 (k), and Oct4-264I8 (l) in control and Klf4
knockdown samples. m-o, Co-localization frequencies of Oct4-143F14 (m), Oct4-474J5 (n), and Oct4-264I8 (o) for
the Oct4 transcriptionally active allele (RNA positive) and the Oct4 transcriptionally inactive allele (RNA negative).
p,q, Co-localization frequencies of Oct4-143F14 (p) and Oct4-474J5 (q) in control and Klf4-overexpressed cells. In
j-q, n indicates the total number of nuclei analyzed in 2 biological samples. *p<0.05, **p<0.01.
95

To determine Klf4’s role in organizing Oct4-locus long-range interactions during reprogramming,
we investigated whether Klf4’s occupancy at these genes is different between SSEA1+ and
SSEA1- cells (Fig. 25e). Whereas the control region showed a ratio of ~ 1 for SSEA1+ versus
SSEA1- cells, enrichment of Klf4 at Oct4, Llgl2, and Grb7 was much higher in SSEA1+ than
that in SSEA1- cells at day 5 and day 10 (Fig. 25f). This suggests that, although Klf4 expression
is induced right after initiation of reprogramming in all cells, Klf4’s occupancy of Oct4 loci and
its interacting partners (Llgl2, Grb7, and Unc84a) only occurs in SSEA1+ cells, which is
consistent with the previous observation that the interchromosomal interactions occur in
SSEA1+ cells. This differential recruitment of Klf4 further supports the hypothesis that Klf4 is
one of the core players in organizing these long-range interactions.
We then generated stable ESC lines with inducible Klf4 shRNA to test whether Klf4 is required
for the aforementioned long-range interactions. Induction of shRNA by doxycycline for 2 days
led to a decrease of Klf4 mRNA and protein (Fig. 25g and Fig. 26), while prolonged depletion
for more than 1 passage led to differentiation (Fig. 27a). On knockdown day 2, whereas Oct4
mRNA and protein levels were still maintained (Fig. 25h and Fig. 27b), interchromosomal
interactions of Oct4 had already begun to dissociate significantly (Fig. 25i-k). This dissociation
is due neither to cell differentiation (Fig.27a, b) nor to a decrease in Oct4 transcription (Fig. 25l
and Fig. 28). More importantly, when we separate Oct4 alleles in the Klf4 knockdown day-2
samples into two subgroups, transcriptionally active and transcriptionally inactive based on RNA
FISH, the reduction in co-localization occurs in both subgroups (Fig. 25m-o). All these data
demonstrate that Klf4 depletion leads to dissociation of Oct4-interchromosomal interactions
independent of transcription. Finally, when Klf4 is induced to be overexpressed in ESCs (Fig.
29), we observe a significant increase in co-localization frequencies (Fig. 25p and q). Together
96

these results suggest that Klf4 is essential for the organization of specific long-range
interchromosomal interactions at Oct4 loci.
Collectively, we demonstrated that PSC-specific long-range interactions are quickly established
during reprogramming. Klf4 participates in organizing these long-range interactions by
occupying Oct4 and its interacting regions. These interactions are correlated with the activation
of endogenous Oct4 during reprogramming and facilitate Oct4 transcription in PSCs. Our results
present a possible explanation of why only a small fraction of cells is able to be transformed into
iPSCs despite the ubiquitous overexpression of reprogramming factors. The transition of higher-
order chromatin structures, represented here by long-range interactions at endogenous Oct4 loci,
occurs only in a sub-population of cells and is highly correlated with transcription of Oct4.
Furthermore, the fact that Klf4 acts as an organizer of interchromosomal interactions in
reprogramming and pluripotency suggests novel roles of reprogramming factors in sculpting the
unique nuclear architecture of PSCs.

Figure 26. Klf4 protein decrease significantly after 2 days of knockdown.
Immunoflurorescece imaging shows that in two different inducible shRNA ES lines, Klf4 staining intensities
decrease significantly after two days of induction of shRNA. While in non-target lines shows no difference in
staining intensities while treated with doxycycline.
97

Figure 27. Characterizing the effect of inducible knockdown of Klf4 to PSCs self-renewal.
a. Images of stable TRIPZ-Klf4 ES cell lines after induction of knockdown. While no apparent morphology
changes were observed in day 1 and day3, part of the cells start to show differentiation in passage 2 and 3.
Doxycycline was added at 5ug/ml and culture medium were changed daily.
b. Immunoflurosence images show the Oct4 staining intensity was not changed at day 2, but was decreased on
day5 (passage 2), indicating that at cells has not committed to differentiation at day 2.
98

Figure 28. Inhibition of RNA polymerase II transcription by α-amanitin treatment does not change
interchromosomal interactions frequencies.
a. Representative images of Oct4 RNA FISH in mock (vehicle only) and α-amanitin treated mouse ES cells.
α-amanitin was added to a final concentration f 100ug/ul and incubated for 5h.
b. Most cells lost nascent Oct4 mRNA signal, indicating the transcription is inhibited.
c. Representative images of co-localization between Oct4 (green) and 143F14 (red) or 474J5 (red) in control
and α-amanitin treated samples.
d,e. Statistics of co-localization frequencies of Oct4-143F14 (d) and Oct4-474J5 (e). The frequencies in control
and α-amanitin samples show no statistical difference.
99

Figure 29. Generation of inducible Klf4 overexpression stable cell lines in mouse ES cells.
a. Schematic design of the lentiviral construct carrying Klf4 cDNA and IRES-GFP under the control of
doxycline inducible promoter.
b. Images of the clonal cell line expressing GFP when treated with doxycycline for 2 days.
c. qPCR of gene expression changes when Klf4 is overexpressed. The cell line is treated with doxycline at the
concentration of 2ug/ml for 2 days. While Klf4 level is elevated, Oct4 and other genes remain largely
unchanged.

100

BAC name Genomic region (mm9)

Representative genes
located in the region
and characterized in
this study
Probes tested in FISH
RP23-143F14
Chr11: 115559980-
115767566 Llgl2

RP24-352D10
Chr5: 125866932-
126002087

RP24-264I8
Chr5: 139542928-
139753935 Unc84a

RP23-37G12
Chr7: 38974113-
39139666

RP23-94F8
Chr14: 122866457-
123033331

RP23-474J5
Chr11: 98210673-
98380153 Grb7

RP23-280I15
Chr18: 69141918-
69295055

RP24-362E16
Chr4: 138738950-
138897728

RP23-366P12
Chr11: 116596256-
116699855

RP24-132L24
Chr16: 35694112-
35842722

RP23-422N13
Chr9: 115391763-
115588122

RP23-458J22
Chr7: 13397752-
13598389

RP24-104G18
Chr1: 133314862-
133479649

RP24-245F10
Chr11: 51130933-
51283557

RP24-315L23
Chr3: 68417244-
68574456

RP24-102A1
Chr11: 25934737-
26115468

RP23-304A22
ChrX: 90082895-
90258791

RP24-248K18
Chr17: 35537970-
35677662 Oct4

Controls:
RP23-449D9
Chr8: 44215086-
44394683 Rex1

RP23-129L1
Chr18: 10974418-
11178480 Gata6

RP23-394O14
Chr3: 140743144-
140908628 Phda2
Table 3. List of BACs used in this study
101

Table 4 List of primers used in this study.

102

Chapter 4 Future perspectives on induced pluripotency and nuclear architecture

4.1 Remaining controversial questions for induced pluripotency
Since the discovery of iPSCs, the identity, safety and quality has always been under debate. Most
of the concerns are related to the potential clinical use of iPSCs. Compared to 6 years ago, our
understanding of generating and maintaining iPSCs has significantly increased. However, some
of the fundamental questions remain unanswered. For example, one of the key questions is: how
good are iPSCs? When we talk about the quality of iPSCs, one should always be reminded that
unlike ESCs, iPSCs are derived from individual labs with various methods. Therefore there are
actually two questions here. 1) Is it possible to derive iPSCs which are as good as standard ESCs?
2) Is there any way to compare and ‘score’ the quality of iPSCs? In the following section I will
discuss these two questions separately.
Are there any systematic discrepancies in iPSCs compared to ESCs?
Since most of reprogramming methods involve transgenes integration, individual lines are
different in genetic background. Although iPSCs meet the most stringent functional test in mouse
and human, the question remains that whether iPSCs are equivalent to ESCs on transcriptional
and epigenetic level. Although most would agree that iPSCs show a high similarity to ESCs,
early studies indicate that a small portion of mRNA and miRNA are differentially expressed in
iPSCs (Chin et al., 2009; Wilson et al., 2009). DNA methylation in iPSCs is also shown to be
slightly different from that in ESCs (Deng et al., 2009; Doi et al., 2009; Pick et al., 2009).
However, many of these studies are done with human ESCs and iPSCs from different individuals.
Therefore the genetic background may also contribute to these discrepancies. In fact, careful
examination of mouse ESCs and iPSCs from same genetic background found no consistent gene
103

expression differences between ESCs and iPSCs, with the exception of aberrant expression of the
imprinting Dlk1-Dio3 gene cluster in some iPSC lines (Stadtfeld et al., 2010a). The imprinting
status of Dlk-Dio3 has been suggested to be indicator of ‘good’ iPSCs. However, recent
discoveries suggested that that it is the stoichiometry of reprogramming factors rather than the
imprinting of Dlk-Dio3 that determine the quality of iPSCs (Carey et al., 2011). From these
conflicting data, the conclusion so far can be summarized as 1) with identical genetic
background, iPSCs and ESCs can be mostly identical in terms of gene expression. Systematical
aberrantly expressed genes are rare. 2) There are variances between iPSCs lines. These variances
could be originated from stochastic events early in reprogramming even with the same
integration and expression level of reprogramming factors. Whether these variance are
functionally important remains unclear. In general, there are no systematic difference in gene
expression between mouse iPSCs and mouse ESCs. 3) Erasure of somatic methylation pattern
can be incomplete in iPSCs. These minor differences could influence the in vitro differentiation
efficiency. Extended culture can attenuate the aberrant DNA methylation. 4) Chimera
experiments are unavailable for human iPSCs. Therefore the functional assay in human so far is
not enough to provide definitive answer for human iPSCs quality. Since no comparison of
human ESCs and iPSCs from same genetic background has been described, it is also difficult to
judge whether human iPSCs are systematically different from human ESCs.
How can we identify the ‘good’ iPSCs?
Current data suggest that there are no common defects for iPSCs. However, iPSC lines are very
different in terms of quality and certain gene expression (Carey et al., 2011). In this case, how
are we going to pick the ‘best’ iPSC lines? Unfortunately, common pluripotency markers are
similar for most iPSC lines. Sensitive functional assays, such as ‘all-iPS’ mice assay, are
104

expensive, time consuming and unavailable for human cells. Currently the best criteria in daily
practices are still based on morphology and experience. Thus it is of great interest to identify an
unbiased and sensitive method to ‘score’ the iPSCs.
One of the promising methods comes from Bock et. al., who performed systematic assays to
chart the gene expression and DNA methylation on genome-wide (Bock et al., 2011). With 20
human ESC lines, the authors characterized a ‘normal range’ of ESCs. These ‘scores’ can
correlate with in vitro differentiation quality of ESCs. When the authors tested 12 iPSC lines
from various origins, some but not all iPSC lines fall into the normal range of ESCs variation.
The genomic ‘scoreboard’ provide a predictive and straightforward way to characterize the
quality of iPSC lines. In future, other epigenetic modification can also be included in the
scoreboard system and further strengthen its potential in identifying good quality iPSCs.
4.2 Building better models for higher order chromatin structure
Although numerous derivatives have made 3C suitable for various applications, how to build a
reliable model from high throughput sequencing results remains challenging. Here I will discuss
the emerging methods and concepts that will enhance our understanding of higher order
chromatin structure in future research.
One of the characteristics of 3C based method, especially for HiC, is to use a large quantity of
cells, usually ~10
7
-10
8
cells. Although huge amount of interactions can be discovered by these
methods, the frequency of interactions actually represents the average frequency in all cells.
Unfortunately, FISH experiments indicate that a certain long range interaction can only be found
in a small percentage of cells. This indicates that cells are very dynamic in terms of nuclear
architecture. For 3C and 4C experiments, the aim is usually only to reveal as many interacting
105

partners as possible. Therefore using the average frequency will lower the signal-to-noise ratio,
but usually will not affect the presentation of final model of long range interactions. However,
for 5C, HiC and other ‘all-to-all’ methods, the presentation is usually a general pattern which in
theory fits all the point to point interactions discovered. This average but unique pattern
apparently does not fit our understanding of nuclear dynamics. To better address this issue,
Kalhor et. al. recently proposed a new method of modeling (Kalhor et al., 2012). This population
based modeling is not aimed at finding a ‘one-fits-all’ model. Instead, a set of various models are
first identified. These models are selected by an algorithm that considers the statistical variability
in the original data. These models of organization are non-random, but representing the parallel
nuclear organizations in different cells.
Developing novel modeling strategies and experimental techniques are essential for studying
nuclear architecture. Besides probabilistic model of organization, long range interactions can
also be related to co-transcribed hubs and trans-regulation (Williams et al., 2010). In these
scenarios, it will be critical to address all the possible structures or follow the dynamic of
transition between different structures, as these dynamics are part of the regulating process. To
this end, applying deep sequencing to obtain higher resolution HiC data, and establishing time-
lapse experiments following the movement of particular loci in real-time, will be essential.

106

Bibliography

Aasen, T., Raya, A., Barrero, M.J., Garreta, E., Consiglio, A., Gonzalez, F., Vassena, R., Bilic, J.,
Pekarik, V., Tiscornia, G., et al. (2008). Efficient and rapid generation of induced pluripotent
stem cells from human keratinocytes. Nature biotechnology 26, 1276-1284.
Amado, R.G., and Chen, I.S. (1999). Lentiviral vectors--the promise of gene therapy within
reach? Science (New York, NY 285, 674-676.
Anokye-Danso, F., Trivedi, C.M., Juhr, D., Gupta, M., Cui, Z., Tian, Y., Zhang, Y., Yang, W.,
Gruber, P.J., Epstein, J.A., et al. (2011). Highly efficient miRNA-mediated reprogramming of
mouse and human somatic cells to pluripotency. Cell stem cell 8, 376-388.
Aoi, T., Yae, K., Nakagawa, M., Ichisaka, T., Okita, K., Takahashi, K., Chiba, T., and Yamanaka,
S. (2008). Generation of pluripotent stem cells from adult mouse liver and stomach cells. Science
(New York, NY 321, 699-702.
Banito, A., Rashid, S.T., Acosta, J.C., Li, S., Pereira, C.F., Geti, I., Pinho, S., Silva, J.C., Azuara,
V., Walsh, M., et al. (2009). Senescence impairs successful reprogramming to pluripotent stem
cells. Genes & development 23, 2134-2139.
Bao, S., Tang, F., Li, X., Hayashi, K., Gillich, A., Lao, K., and Surani, M.A. (2009). Epigenetic
reversion of post-implantation epiblast to pluripotent embryonic stem cells. Nature 461, 1292-
1295.
Bhutani, N., Brady, J.J., Damian, M., Sacco, A., Corbel, S.Y., and Blau, H.M. Reprogramming
towards pluripotency requires AID-dependent DNA demethylation. Nature 463, 1042-1047.
Bock, C., Kiskinis, E., Verstappen, G., Gu, H., Boulting, G., Smith, Z.D., Ziller, M., Croft, G.F.,
Amoroso, M.W., Oakley, D.H., et al. (2011). Reference Maps of human ES and iPS cell
variation enable high-throughput characterization of pluripotent cell lines. Cell 144, 439-452.
Boulting, G.L., Kiskinis, E., Croft, G.F., Amoroso, M.W., Oakley, D.H., Wainger, B.J., Williams,
D.J., Kahler, D.J., Yamaki, M., Davidow, L., et al. (2011). A functionally characterized test set
of human induced pluripotent stem cells. Nature biotechnology 29, 279-286.
Boyer, L.A., Lee, T.I., Cole, M.F., Johnstone, S.E., Levine, S.S., Zucker, J.P., Guenther, M.G.,
Kumar, R.M., Murray, H.L., Jenner, R.G., et al. (2005). Core transcriptional regulatory circuitry
in human embryonic stem cells. Cell 122, 947-956.
Boyer, L.A., Plath, K., Zeitlinger, J., Brambrink, T., Medeiros, L.A., Lee, T.I., Levine, S.S.,
Wernig, M., Tajonar, A., Ray, M.K., et al. (2006). Polycomb complexes repress developmental
regulators in murine embryonic stem cells. Nature 441, 349-353.
Brambrink, T., Foreman, R., Welstead, G.G., Lengner, C.J., Wernig, M., Suh, H., and Jaenisch,
R. (2008). Sequential expression of pluripotency markers during direct reprogramming of mouse
somatic cells. Cell stem cell 2, 151-159.
107

Brennand, K.J., Simone, A., Jou, J., Gelboin-Burkhart, C., Tran, N., Sangar, S., Li, Y., Mu, Y.,
Chen, G., Yu, D., et al. Modelling schizophrenia using human induced pluripotent stem cells.
Nature 473, 221-225.
Brons, I.G., Smithers, L.E., Trotter, M.W., Rugg-Gunn, P., Sun, B., Chuva de Sousa Lopes, S.M.,
Howlett, S.K., Clarkson, A., Ahrlund-Richter, L., Pedersen, R.A., et al. (2007). Derivation of
pluripotent epiblast stem cells from mammalian embryos. Nature 448, 191-195.
Brookes, E., de Santiago, I., Hebenstreit, D., Morris, K.J., Carroll, T., Xie, S.Q., Stock, J.K.,
Heidemann, M., Eick, D., Nozaki, N., et al. (2012). Polycomb associates genome-wide with a
specific RNA polymerase II variant, and regulates metabolic genes in ESCs. Cell stem cell 10,
157-170.
Burghes, A.H., and Beattie, C.E. (2009). Spinal muscular atrophy: why do low levels of survival
motor neuron protein make motor neurons sick? Nature reviews Neuroscience 10, 597-609.
Carey, B.W., Markoulaki, S., Beard, C., Hanna, J., and Jaenisch, R. (2010). Single-gene
transgenic mouse strains for reprogramming adult somatic cells. Nature methods 7, 56-59.
Carey, B.W., Markoulaki, S., Hanna, J.H., Faddah, D.A., Buganim, Y., Kim, J., Ganz, K., Steine,
E.J., Cassady, J.P., Creyghton, M.P., et al. (2011). Reprogramming factor stoichiometry
influences the epigenetic state and biological properties of induced pluripotent stem cells. Cell
stem cell 9, 588-598.
Carvajal-Vergara, X., Sevilla, A., D'Souza, S.L., Ang, Y.S., Schaniel, C., Lee, D.F., Yang, L.,
Kaplan, A.D., Adler, E.D., Rozov, R., et al. (2010). Patient-specific induced pluripotent stem-
cell-derived models of LEOPARD syndrome. Nature 465, 808-812.
Chakalova, L., and Fraser, P. (2010). Organization of transcription. Cold Spring Harbor
perspectives in biology 2, a000729.
Chamberlain, S.J., Chen, P.F., Ng, K.Y., Bourgois-Rocha, F., Lemtiri-Chlieh, F., Levine, E.S.,
and Lalande, M. (2010). Induced pluripotent stem cell models of the genomic imprinting
disorders Angelman and Prader-Willi syndromes. Proceedings of the National Academy of
Sciences of the United States of America 107, 17668-17673.
Chambeyron, S., and Bickmore, W.A. (2004). Chromatin decondensation and nuclear
reorganization of the HoxB locus upon induction of transcription. Genes & development 18,
1119-1130.
Chang, M.Y., Kim, D., Kim, C.H., Kang, H.C., Yang, E., Moon, J.I., Ko, S., Park, J., Park, K.S.,
Lee, K.A., et al. (2010). Direct reprogramming of rat neural precursor cells and fibroblasts into
pluripotent stem cells. PloS one 5, e9838.
Chaumeil, J., Augui, S., Chow, J.C., and Heard, E. (2008). Combined immunofluorescence,
RNA fluorescent in situ hybridization, and DNA fluorescent in situ hybridization to study
chromatin changes, transcriptional activity, nuclear organization, and X-chromosome
inactivation. Methods in molecular biology (Clifton, NJ 463, 297-308.
108

Chen, X., Xu, H., Yuan, P., Fang, F., Huss, M., Vega, V.B., Wong, E., Orlov, Y.L., Zhang, W.,
Jiang, J., et al. (2008). Integration of external signaling pathways with the core transcriptional
network in embryonic stem cells. Cell 133, 1106-1117.
Chin, M.H., Mason, M.J., Xie, W., Volinia, S., Singer, M., Peterson, C., Ambartsumyan, G.,
Aimiuwu, O., Richter, L., Zhang, J., et al. (2009). Induced pluripotent stem cells and embryonic
stem cells are distinguished by gene expression signatures. Cell stem cell 5, 111-123.
Chung, T.L., Brena, R.M., Kolle, G., Grimmond, S.M., Berman, B.P., Laird, P.W., Pera, M.F.,
and Wolvetang, E.J. Vitamin C promotes widespread yet specific DNA demethylation of the
epigenome in human embryonic stem cells. Stem cells (Dayton, Ohio) 28, 1848-1855.
Cook, P.R. (2002). Predicting three-dimensional genome structure from transcriptional activity.
Nature genetics 32, 347-352.
Cowan, C.A., Atienza, J., Melton, D.A., and Eggan, K. (2005). Nuclear reprogramming of
somatic cells after fusion with human embryonic stem cells. Science (New York, NY 309, 1369-
1373.
Cremer, C., Cremer, T., and Gray, J.W. (1982). Induction of chromosome damage by ultraviolet
light and caffeine: correlation of cytogenetic evaluation and flow karyotype. Cytometry 2, 287-
290.
Dekker, J., Rippe, K., Dekker, M., and Kleckner, N. (2002). Capturing chromosome
conformation. Science (New York, NY 295, 1306-1311.
Deng, J., Shoemaker, R., Xie, B., Gore, A., LeProust, E.M., Antosiewicz-Bourget, J., Egli, D.,
Maherali, N., Park, I.H., Yu, J., et al. (2009). Targeted bisulfite sequencing reveals changes in
DNA methylation associated with nuclear reprogramming. Nature biotechnology 27, 353-360.
Di Giorgio, F.P., Boulting, G.L., Bobrowicz, S., and Eggan, K.C. (2008). Human embryonic
stem cell-derived motor neurons are sensitive to the toxic effect of glial cells carrying an ALS-
causing mutation. Cell stem cell 3, 637-648.
Dimos, J.T., Rodolfa, K.T., Niakan, K.K., Weisenthal, L.M., Mitsumoto, H., Chung, W., Croft,
G.F., Saphier, G., Leibel, R., Goland, R., et al. (2008). Induced pluripotent stem cells generated
from patients with ALS can be differentiated into motor neurons. Science (New York, NY 321,
1218-1221.
Doi, A., Park, I.H., Wen, B., Murakami, P., Aryee, M.J., Irizarry, R., Herb, B., Ladd-Acosta, C.,
Rho, J., Loewer, S., et al. (2009). Differential methylation of tissue- and cancer-specific CpG
island shores distinguishes human induced pluripotent stem cells, embryonic stem cells and
fibroblasts. Nature genetics 41, 1350-1353.
Donohoe, M.E., Silva, S.S., Pinter, S.F., Xu, N., and Lee, J.T. (2009). The pluripotency factor
Oct4 interacts with Ctcf and also controls X-chromosome pairing and counting. Nature 460, 128-
132.
Dostie, J., Richmond, T.A., Arnaout, R.A., Selzer, R.R., Lee, W.L., Honan, T.A., Rubio, E.D.,
Krumm, A., Lamb, J., Nusbaum, C., et al. (2006). Chromosome Conformation Capture Carbon
109

Copy (5C): a massively parallel solution for mapping interactions between genomic elements.
Genome research 16, 1299-1309.
Drissen, R., Palstra, R.J., Gillemans, N., Splinter, E., Grosveld, F., Philipsen, S., and de Laat, W.
(2004). The active spatial organization of the beta-globin locus requires the transcription factor
EKLF. Genes & development 18, 2485-2490.
Duan, Z., Andronescu, M., Schutz, K., McIlwain, S., Kim, Y.J., Lee, C., Shendure, J., Fields, S.,
Blau, C.A., and Noble, W.S. (2010). A three-dimensional model of the yeast genome. Nature
465, 363-367.
Ebert, A.D., Yu, J., Rose, F.F., Jr., Mattis, V.B., Lorson, C.L., Thomson, J.A., and Svendsen,
C.N. (2009). Induced pluripotent stem cells from a spinal muscular atrophy patient. Nature 457,
277-280.
Eggan, K., Baldwin, K., Tackett, M., Osborne, J., Gogos, J., Chess, A., Axel, R., and Jaenisch, R.
(2004). Mice cloned from olfactory sensory neurons. Nature 428, 44-49.
Egli, D., Chen, A.E., Saphier, G., Powers, D., Alper, M., Katz, K., Berger, B., Goland, R., Leibel,
R.L., Melton, D.A., et al. (2011). Impracticality of egg donor recruitment in the absence of
compensation. Cell stem cell 9, 293-294.
Eminli, S., Foudi, A., Stadtfeld, M., Maherali, N., Ahfeldt, T., Mostoslavsky, G., Hock, H., and
Hochedlinger, K. (2009). Differentiation stage determines potential of hematopoietic cells for
reprogramming into induced pluripotent stem cells. Nature genetics 41, 968-976.
Eminli, S., Utikal, J.S., Arnold, K., Jaenisch, R., and Hochedlinger, K. (2008). Reprogramming
of Neural Progenitor Cells into iPS Cells in the Absence of Exogenous Sox2 Expression. Stem
cells (Dayton, Ohio).
Esteban, M.A., Wang, T., Qin, B., Yang, J., Qin, D., Cai, J., Li, W., Weng, Z., Chen, J., Ni, S., et
al. (2010). Vitamin C enhances the generation of mouse and human induced pluripotent stem
cells. Cell stem cell 6, 71-79.
Esteban, M.A., Xu, J., Yang, J., Peng, M., Qin, D., Li, W., Jiang, Z., Chen, J., Deng, K., Zhong,
M., et al. (2009). Generation of induced pluripotent stem cell lines from Tibetan miniature pig.
The Journal of biological chemistry 284, 17634-17640.
Evans, M.J., and Kaufman, M.H. (1981). Establishment in culture of pluripotential cells from
mouse embryos. Nature 292, 154-156.
Evans, P.M., Zhang, W., Chen, X., Yang, J., Bhakat, K.K., and Liu, C. (2007). Kruppel-like
factor 4 is acetylated by p300 and regulates gene transcription via modulation of histone
acetylation. The Journal of biological chemistry 282, 33994-34002.
Feng, B., Jiang, J., Kraus, P., Ng, J.H., Heng, J.C., Chan, Y.S., Yaw, L.P., Zhang, W., Loh, Y.H.,
Han, J., et al. (2009a). Reprogramming of fibroblasts into induced pluripotent stem cells with
orphan nuclear receptor Esrrb. Nature cell biology 11, 197-203.
110

Feng, B., Ng, J.H., Heng, J.C., and Ng, H.H. (2009b). Molecules that promote or enhance
reprogramming of somatic cells to induced pluripotent stem cells. Cell stem cell 4, 301-312.
Ferrai, C., Xie, S.Q., Luraghi, P., Munari, D., Ramirez, F., Branco, M.R., Pombo, A., and Crippa,
M.P. (2010). Poised transcription factories prime silent uPA gene prior to activation. PLoS
biology 8, e1000270.
Fullwood, M.J., Liu, M.H., Pan, Y.F., Liu, J., Xu, H., Mohamed, Y.B., Orlov, Y.L., Velkov, S.,
Ho, A., Mei, P.H., et al. (2009). An oestrogen-receptor-alpha-bound human chromatin
interactome. Nature 462, 58-64.
Fusaki, N., Ban, H., Nishiyama, A., Saeki, K., and Hasegawa, M. (2009). Efficient induction of
transgene-free human pluripotent stem cells using a vector based on Sendai virus, an RNA virus
that does not integrate into the host genome. Proceedings of the Japan Academy 85, 348-362.
Gheldof, N., Smith, E.M., Tabuchi, T.M., Koch, C.M., Dunham, I., Stamatoyannopoulos, J.A.,
and Dekker, J. (2010). Cell-type-specific long-range looping interactions identify distant
regulatory elements of the CFTR gene. Nucleic acids research 38, 4325-4336.
Guo, G., Yang, J., Nichols, J., Hall, J.S., Eyres, I., Mansfield, W., and Smith, A. (2009). Klf4
reverts developmentally programmed restriction of ground state pluripotency. Development
(Cambridge, England) 136, 1063-1069.
Gurdon, J.B. (1962). Adult frogs derived from the nuclei of single somatic cells. Developmental
biology 4, 256-273.
Gurdon, J.B., and Melton, D.A. (2008). Nuclear reprogramming in cells. Science (New York,
NY 322, 1811-1815.
Han, D.W., Greber, B., Wu, G., Tapia, N., Arauzo-Bravo, M.J., Ko, K., Bernemann, C., Stehling,
M., and Scholer, H.R. (2011a). Direct reprogramming of fibroblasts into epiblast stem cells.
Nature cell biology 13, 66-71.
Han, S.S., Williams, L.A., and Eggan, K.C. (2011b). Constructing and deconstructing stem cell
models of neurological disease. Neuron 70, 626-644.
Handoko, L., Xu, H., Li, G., Ngan, C.Y., Chew, E., Schnapp, M., Lee, C.W., Ye, C., Ping, J.L.,
Mulawadi, F., et al. (2011). CTCF-mediated functional chromatin interactome in pluripotent
cells. Nature genetics 43, 630-638.
Hanna, J., Cheng, A.W., Saha, K., Kim, J., Lengner, C.J., Soldner, F., Cassady, J.P., Muffat, J.,
Carey, B.W., and Jaenisch, R. (2010a). Human embryonic stem cells with biological and
epigenetic characteristics similar to those of mouse ESCs. Proceedings of the National Academy
of Sciences of the United States of America 107, 9222-9227.
Hanna, J., Markoulaki, S., Mitalipov, S.M., Cheng, A.W., Cassady, J.P., Staerk, J., Carey, B.W.,
Lengner, C.J., Foreman, R., Love, J., et al. (2009a). Metastable Pluripotent States in NOD-
Mouse-Derived ESCs. Cell stem cell.
111

Hanna, J., Markoulaki, S., Schorderet, P., Carey, B.W., Beard, C., Wernig, M., Creyghton, M.P.,
Steine, E.J., Cassady, J.P., Foreman, R., et al. (2008). Direct reprogramming of terminally
differentiated mature B lymphocytes to pluripotency. Cell 133, 250-264.
Hanna, J., Saha, K., Pando, B., van Zon, J., Lengner, C.J., Creyghton, M.P., van Oudenaarden,
A., and Jaenisch, R. (2009b). Direct cell reprogramming is a stochastic process amenable to
acceleration. Nature 462, 595-601.
Hanna, J.H., Saha, K., and Jaenisch, R. (2010b). Pluripotency and cellular reprogramming: facts,
hypotheses, unresolved issues. Cell 143, 508-525.
Hansen, D.V., Rubenstein, J.L., and Kriegstein, A.R. (2011). Deriving excitatory neurons of the
neocortex from pluripotent stem cells. Neuron 70, 645-660.
Hargus, G., Cooper, O., Deleidi, M., Levy, A., Lee, K., Marlow, E., Yow, A., Soldner, F.,
Hockemeyer, D., Hallett, P.J., et al. Differentiated Parkinson patient-derived induced pluripotent
stem cells grow in the adult rodent brain and reduce motor asymmetry in Parkinsonian rats.
Proceedings of the National Academy of Sciences of the United States of America 107, 15921-
15926.
Harper, S., and Speicher, D.W. (2001). Expression and purification of GST fusion proteins. Curr
Protoc Protein Sci Chapter 6, Unit 6 6.
Heng, J.C., Feng, B., Han, J., Jiang, J., Kraus, P., Ng, J.H., Orlov, Y.L., Huss, M., Yang, L.,
Lufkin, T., et al. (2010). The nuclear receptor Nr5a2 can replace Oct4 in the reprogramming of
murine somatic cells to pluripotent cells. Cell stem cell 6, 167-174.
Hiratani, I., Ryba, T., Itoh, M., Rathjen, J., Kulik, M., Papp, B., Fussner, E., Bazett-Jones, D.P.,
Plath, K., Dalton, S., et al. (2010). Genome-wide dynamics of replication timing revealed by in
vitro models of mouse embryogenesis. Genome research 20, 155-169.
Hochedlinger, K., and Jaenisch, R. (2002). Monoclonal mice generated by nuclear transfer from
mature B and T donor cells. Nature 415, 1035-1038.
Hochedlinger, K., and Plath, K. (2009). Epigenetic reprogramming and induced pluripotency.
Development (Cambridge, England) 136, 509-523.
Huangfu, D., Maehr, R., Guo, W., Eijkelenboom, A., Snitow, M., Chen, A.E., and Melton, D.A.
(2008a). Induction of pluripotent stem cells by defined factors is greatly improved by small-
molecule compounds. Nature biotechnology 26, 795-797.
Huangfu, D., Osafune, K., Maehr, R., Guo, W., Eijkelenboom, A., Chen, S., Muhlestein, W., and
Melton, D.A. (2008b). Induction of pluripotent stem cells from primary human fibroblasts with
only Oct4 and Sox2. Nature biotechnology.
Iborra, F.J., Pombo, A., Jackson, D.A., and Cook, P.R. (1996a). Active RNA polymerases are
localized within discrete transcription "factories' in human nuclei. Journal of cell science 109 ( Pt
6), 1427-1436.
112

Iborra, F.J., Pombo, A., McManus, J., Jackson, D.A., and Cook, P.R. (1996b). The topology of
transcription by immobilized polymerases. Experimental cell research 229, 167-173.
Ichida, J.K., Blanchard, J., Lam, K., Son, E.Y., Chung, J.E., Egli, D., Loh, K.M., Carter, A.C., Di
Giorgio, F.P., Koszka, K., et al. (2009). A small-molecule inhibitor of tgf-Beta signaling replaces
sox2 in reprogramming by inducing nanog. Cell stem cell 5, 491-503.
Inoue, K., Wakao, H., Ogonuki, N., Miki, H., Seino, K., Nambu-Wakao, R., Noda, S., Miyoshi,
H., Koseki, H., Taniguchi, M., et al. (2005). Generation of cloned mice by direct nuclear transfer
from natural killer T cells. Curr Biol 15, 1114-1118.
Ito, S., D'Alessio, A.C., Taranova, O.V., Hong, K., Sowers, L.C., and Zhang, Y. (2010). Role of
Tet proteins in 5mC to 5hmC conversion, ES-cell self-renewal and inner cell mass specification.
Nature 466, 1129-1133.
Itzhaki, I., Maizels, L., Huber, I., Zwi-Dantsis, L., Caspi, O., Winterstern, A., Feldman, O.,
Gepstein, A., Arbel, G., Hammerman, H., et al. Modelling the long QT syndrome with induced
pluripotent stem cells. Nature 471, 225-229.
Jackson, D.A., Hassan, A.B., Errington, R.J., and Cook, P.R. (1993). Visualization of focal sites
of transcription within human nuclei. The EMBO journal 12, 1059-1065.
Jackson, D.A., Iborra, F.J., Manders, E.M., and Cook, P.R. (1998). Numbers and organization of
RNA polymerases, nascent transcripts, and transcription units in HeLa nuclei. Molecular biology
of the cell 9, 1523-1536.
Jaenisch, R., and Young, R. (2008). Stem cells, the molecular circuitry of pluripotency and
nuclear reprogramming. Cell 132, 567-582.
Jia, F., Wilson, K.D., Sun, N., Gupta, D.M., Huang, M., Li, Z., Panetta, N.J., Chen, Z.Y.,
Robbins, R.C., Kay, M.A., et al. (2010). A nonviral minicircle vector for deriving human iPS
cells. Nature methods 7, 197-199.
Jiang, J., Chan, Y.S., Loh, Y.H., Cai, J., Tong, G.Q., Lim, C.A., Robson, P., Zhong, S., and Ng,
H.H. (2008). A core Klf circuitry regulates self-renewal of embryonic stem cells. Nature cell
biology 10, 353-360.
Kabata, H., Kurosawa, O., Arai, I., Washizu, M., Margarson, S.A., Glass, R.E., and Shimamoto,
N. (1993). Visualization of single molecules of RNA polymerase sliding along DNA. Science
(New York, NY 262, 1561-1563.
Kaji, K., Norrby, K., Paca, A., Mileikovsky, M., Mohseni, P., and Woltjen, K. (2009). Virus-free
induction of pluripotency and subsequent excision of reprogramming factors. Nature 458, 771-
775.
Kalhor, R., Tjong, H., Jayathilaka, N., Alber, F., and Chen, L. (2012). Genome architectures
revealed by tethered chromosome conformation capture and population-based modeling. Nature
biotechnology 30, 90-98.
113

Kelly, S.J. (1977). Studies of the developmental potential of 4- and 8-cell stage mouse
blastomeres. The Journal of experimental zoology 200, 365-376.
Kim, J., Chu, J., Shen, X., Wang, J., and Orkin, S.H. (2008a). An extended transcriptional
network for pluripotency of embryonic stem cells. Cell 132, 1049-1061.
Kim, J., Lengner, C.J., Kirak, O., Hanna, J., Cassady, J.P., Lodato, M.A., Wu, S., Faddah, D.A.,
Steine, E.J., Gao, Q., et al. (2011). Reprogramming of postnatal neurons into induced pluripotent
stem cells by defined factors. Stem cells (Dayton, Ohio) 29, 992-1000.
Kim, J.B., Greber, B., Arauzo-Bravo, M.J., Meyer, J., Park, K.I., Zaehres, H., and Scholer, H.R.
(2009a). Direct reprogramming of human neural stem cells by OCT4. Nature 461, 649-643.
Kim, J.B., Sebastiano, V., Wu, G., Arauzo-Bravo, M.J., Sasse, P., Gentile, L., Ko, K., Ruau, D.,
Ehrich, M., van den Boom, D., et al. (2009b). Oct4-induced pluripotency in adult neural stem
cells. Cell 136, 411-419.
Kim, J.B., Zaehres, H., Wu, G., Gentile, L., Ko, K., Sebastiano, V., Arauzo-Bravo, M.J., Ruau,
D., Han, D.W., Zenke, M., et al. (2008b). Pluripotent stem cells induced from adult neural stem
cells by reprogramming with two factors. Nature 454, 646-650.
Kim, K., Doi, A., Wen, B., Ng, K., Zhao, R., Cahan, P., Kim, J., Aryee, M.J., Ji, H., Ehrlich, L.I.,
et al. (2010). Epigenetic memory in induced pluripotent stem cells. Nature 467, 285-290.
Ku, S., Soragni, E., Campau, E., Thomas, E.A., Altun, G., Laurent, L.C., Loring, J.F., Napierala,
M., and Gottesfeld, J.M. (2010). Friedreich's ataxia induced pluripotent stem cells model
intergenerational GAATTC triplet repeat instability. Cell stem cell 7, 631-637.
Lee, G., and Studer, L. (2011). Modelling familial dysautonomia in human induced pluripotent
stem cells. Philosophical transactions of the Royal Society of London 366, 2286-2296.
Lee, J.T. (2010). The X as model for RNA's niche in epigenomic regulation. Cold Spring Harbor
perspectives in biology 2, a003749.
Lee, J.T. (2011). Gracefully ageing at 50, X-chromosome inactivation becomes a paradigm for
RNA and chromatin control. Nature reviews 12, 815-826.
Lee, T.I., Jenner, R.G., Boyer, L.A., Guenther, M.G., Levine, S.S., Kumar, R.M., Chevalier, B.,
Johnstone, S.E., Cole, M.F., Isono, K., et al. (2006). Control of developmental regulators by
Polycomb in human embryonic stem cells. Cell 125, 301-313.
Lengner, C.J., Gimelbrant, A.A., Erwin, J.A., Cheng, A.W., Guenther, M.G., Welstead, G.G.,
Alagappan, R., Frampton, G.M., Xu, P., Muffat, J., et al. (2010). Derivation of pre-X inactivation
human embryonic stem cells under physiological oxygen concentrations. Cell 141, 872-883.
Levasseur, D.N., Wang, J., Dorschner, M.O., Stamatoyannopoulos, J.A., and Orkin, S.H. (2008).
Oct4 dependence of chromatin structure within the extended Nanog locus in ES cells. Genes &
development 22, 575-580.
114

Li, J., Greco, V., Guasch, G., Fuchs, E., and Mombaerts, P. (2007). Mice cloned from skin cells.
Proceedings of the National Academy of Sciences of the United States of America 104, 2738-
2743.
Li, J., Ishii, T., Feinstein, P., and Mombaerts, P. (2004). Odorant receptor gene choice is reset by
nuclear transfer from mouse olfactory sensory neurons. Nature 428, 393-399.
Li, R., Liang, J., Ni, S., Zhou, T., Qing, X., Li, H., He, W., Chen, J., Li, F., Zhuang, Q., et al.
(2010). A mesenchymal-to-epithelial transition initiates and is required for the nuclear
reprogramming of mouse fibroblasts. Cell stem cell 7, 51-63.
Li, W., Wei, W., Zhu, S., Zhu, J., Shi, Y., Lin, T., Hao, E., Hayek, A., Deng, H., and Ding, S.
(2009). Generation of rat and human induced pluripotent stem cells by combining genetic
reprogramming and chemical inhibitors. Cell stem cell 4, 16-19.
Liao, J., Cui, C., Chen, S., Ren, J., Chen, J., Gao, Y., Li, H., Jia, N., Cheng, L., Xiao, H., et al.
(2009). Generation of induced pluripotent stem cell lines from adult rat cells. Cell stem cell 4,
11-15.
Lieberman-Aiden, E., van Berkum, N.L., Williams, L., Imakaev, M., Ragoczy, T., Telling, A.,
Amit, I., Lajoie, B.R., Sabo, P.J., Dorschner, M.O., et al. (2009). Comprehensive mapping of
long-range interactions reveals folding principles of the human genome. Science (New York, NY
326, 289-293.
Ling, J.Q., Li, T., Hu, J.F., Vu, T.H., Chen, H.L., Qiu, X.W., Cherry, A.M., and Hoffman, A.R.
(2006). CTCF mediates interchromosomal colocalization between Igf2/H19 and Wsb1/Nf1.
Science (New York, NY 312, 269-272.
Lister, R., Pelizzola, M., Kida, Y.S., Hawkins, R.D., Nery, J.R., Hon, G., Antosiewicz-Bourget,
J., O'Malley, R., Castanon, R., Klugman, S., et al. Hotspots of aberrant epigenomic
reprogramming in human induced pluripotent stem cells. Nature 471, 68-73.
Liu, H., Ye, Z., Kim, Y., Sharkis, S., and Jang, Y.Y. Generation of endoderm-derived human
induced pluripotent stem cells from primary hepatocytes. Hepatology (Baltimore, Md 51, 1810-
1819.
Liu, H., Zhu, F., Yong, J., Zhang, P., Hou, P., Li, H., Jiang, W., Cai, J., Liu, M., Cui, K., et al.
(2008). Generation of induced pluripotent stem cells from adult rhesus monkey fibroblasts. Cell
stem cell 3, 587-590.
Liu, J., Verma, P.J., Evans-Galea, M.V., Delatycki, M.B., Michalska, A., Leung, J., Crombie, D.,
Sarsero, J.P., Williamson, R., Dottori, M., et al. (2011). Generation of induced pluripotent stem
cell lines from Friedreich ataxia patients. Stem cell reviews 7, 703-713.
Loh, Y.H., Agarwal, S., Park, I.H., Urbach, A., Huo, H., Heffner, G.C., Kim, K., Miller, J.D., Ng,
K., and Daley, G.Q. (2009). Generation of induced pluripotent stem cells from human blood.
Blood 113, 5476-5479.
115

Loh, Y.H., Wu, Q., Chew, J.L., Vega, V.B., Zhang, W., Chen, X., Bourque, G., George, J.,
Leong, B., Liu, J., et al. (2006). The Oct4 and Nanog transcription network regulates
pluripotency in mouse embryonic stem cells. Nature genetics 38, 431-440.
Lois, C., Hong, E.J., Pease, S., Brown, E.J., and Baltimore, D. (2002). Germline transmission
and tissue-specific expression of transgenes delivered by lentiviral vectors. Science (New York,
NY 295, 868-872.
Lomvardas, S., Barnea, G., Pisapia, D.J., Mendelsohn, M., Kirkland, J., and Axel, R. (2006).
Interchromosomal interactions and olfactory receptor choice. Cell 126, 403-413.
Lowry, W.E., Richter, L., Yachechko, R., Pyle, A.D., Tchieu, J., Sridharan, R., Clark, A.T., and
Plath, K. (2008). Generation of human induced pluripotent stem cells from dermal fibroblasts.
Proceedings of the National Academy of Sciences of the United States of America 105, 2883-
2888.
Lu, W., Yamamoto, V., Ortega, B., and Baltimore, D. (2004). Mammalian Ryk is a Wnt
coreceptor required for stimulation of neurite outgrowth. Cell 119, 97-108.
Lyssiotis, C.A., Foreman, R.K., Staerk, J., Garcia, M., Mathur, D., Markoulaki, S., Hanna, J.,
Lairson, L.L., Charette, B.D., Bouchez, L.C., et al. (2009). Reprogramming of murine fibroblasts
to induced pluripotent stem cells with chemical complementation of Klf4. Proceedings of the
National Academy of Sciences of the United States of America 106, 8912-8917.
Lyu, J., Yamamoto, V., and Lu, W. (2008). Cleavage of the Wnt receptor Ryk regulates neuronal
differentiation during cortical neurogenesis. Developmental cell 15, 773-780.
Maehr, R., Chen, S., Snitow, M., Ludwig, T., Yagasaki, L., Goland, R., Leibel, R.L., and Melton,
D.A. (2009). Generation of pluripotent stem cells from patients with type 1 diabetes. Proceedings
of the National Academy of Sciences of the United States of America 106, 15768-15773.
Mahatan, C.S., Kaestner, K.H., Geiman, D.E., and Yang, V.W. (1999). Characterization of the
structure and regulation of the murine gene encoding gut-enriched Kruppel-like factor (Kruppel-
like factor 4). Nucleic acids research 27, 4562-4569.
Maherali, N., Sridharan, R., Xie, W., Utikal, J., Eminli, S., Arnold, K., Stadtfeld, M., Yachechko,
R., Tchieu, J., Jaenisch, R., et al. (2007). Directly reprogrammed fibroblasts show global
epigenetic remodeling and widespread tissue contribution. Cell stem cell 1, 55-70.
Marchetto, M.C., Carromeu, C., Acab, A., Yu, D., Yeo, G.W., Mu, Y., Chen, G., Gage, F.H., and
Muotri, A.R. (2010). A model for neural development and treatment of Rett syndrome using
human induced pluripotent stem cells. Cell 143, 527-539.
Marchetto, M.C., Muotri, A.R., Mu, Y., Smith, A.M., Cezar, G.G., and Gage, F.H. (2008). Non-
cell-autonomous effect of human SOD1 G37R astrocytes on motor neurons derived from human
embryonic stem cells. Cell stem cell 3, 649-657.
Marion, R.M., Strati, K., Li, H., Murga, M., Blanco, R., Ortega, S., Fernandez-Capetillo, O.,
Serrano, M., and Blasco, M.A. (2009a). A p53-mediated DNA damage response limits
reprogramming to ensure iPS cell genomic integrity. Nature 460, 1149-1153.
116

Marion, R.M., Strati, K., Li, H., Tejera, A., Schoeftner, S., Ortega, S., Serrano, M., and Blasco,
M.A. (2009b). Telomeres acquire embryonic stem cell characteristics in induced pluripotent stem
cells. Cell stem cell 4, 141-154.
Markoulaki, S., Hanna, J., Beard, C., Carey, B.W., Cheng, A.W., Lengner, C.J., Dausman, J.A.,
Fu, D., Gao, Q., Wu, S., et al. (2009). Transgenic mice with defined combinations of drug-
inducible reprogramming factors. Nature biotechnology 27, 169-171.
Marson, A., Foreman, R., Chevalier, B., Bilodeau, S., Kahn, M., Young, R.A., and Jaenisch, R.
(2008). Wnt signaling promotes reprogramming of somatic cells to pluripotency. Cell stem cell 3,
132-135.
McGrath, J., and Solter, D. (1984). Inability of mouse blastomere nuclei transferred to
enucleated zygotes to support development in vitro. Science (New York, NY 226, 1317-1319.
Meissner, A., Mikkelsen, T.S., Gu, H., Wernig, M., Hanna, J., Sivachenko, A., Zhang, X.,
Bernstein, B.E., Nusbaum, C., Jaffe, D.B., et al. (2008). Genome-scale DNA methylation maps
of pluripotent and differentiated cells. Nature 454, 766-770.
Merika, M., and Orkin, S.H. (1995). Functional synergy and physical interactions of the
erythroid transcription factor GATA-1 with the Kruppel family proteins Sp1 and EKLF.
Molecular and cellular biology 15, 2437-2447.
Meshorer, E., Yellajoshula, D., George, E., Scambler, P.J., Brown, D.T., and Misteli, T. (2006).
Hyperdynamic plasticity of chromatin proteins in pluripotent embryonic stem cells.
Developmental cell 10, 105-116.
Mikkelsen, T.S., Hanna, J., Zhang, X., Ku, M., Wernig, M., Schorderet, P., Bernstein, B.E.,
Jaenisch, R., Lander, E.S., and Meissner, A. (2008). Dissecting direct reprogramming through
integrative genomic analysis. Nature 454, 49-55.
Mitchell, J.A., and Fraser, P. (2008). Transcription factories are nuclear subcompartments that
remain in the absence of transcription. Genes & development 22, 20-25.
Moretti, A., Bellin, M., Welling, A., Jung, C.B., Lam, J.T., Bott-Flugel, L., Dorn, T., Goedel, A.,
Hohnke, C., Hofmann, F., et al. (2010). Patient-specific induced pluripotent stem-cell models for
long-QT syndrome. The New England journal of medicine 363, 1397-1409.
Murrell, A., Heeson, S., and Reik, W. (2004). Interaction between differentially methylated
regions partitions the imprinted genes Igf2 and H19 into parent-specific chromatin loops. Nature
genetics 36, 889-893.
Nakagawa, M., Koyanagi, M., Tanabe, K., Takahashi, K., Ichisaka, T., Aoi, T., Okita, K.,
Mochiduki, Y., Takizawa, N., and Yamanaka, S. (2008). Generation of induced pluripotent stem
cells without Myc from mouse and human fibroblasts. Nature biotechnology 26, 101-106.
Nakatake, Y., Fukui, N., Iwamatsu, Y., Masui, S., Takahashi, K., Yagi, R., Yagi, K., Miyazaki, J.,
Matoba, R., Ko, M.S., et al. (2006). Klf4 cooperates with Oct3/4 and Sox2 to activate the Lefty1
core promoter in embryonic stem cells. Molecular and cellular biology 26, 7772-7782.
117

Nguyen, H.N., Byers, B., Cord, B., Shcheglovitov, A., Byrne, J., Gujar, P., Kee, K., Schule, B.,
Dolmetsch, R.E., Langston, W., et al. (2011). LRRK2 mutant iPSC-derived DA neurons
demonstrate increased susceptibility to oxidative stress. Cell stem cell 8, 267-280.
Nichols, J., and Smith, A. (2009). Naive and primed pluripotent states. Cell stem cell 4, 487-492.
Noggle, S., Fung, H.L., Gore, A., Martinez, H., Satriani, K.C., Prosser, R., Oum, K., Paull, D.,
Druckenmiller, S., Freeby, M., et al. (2011). Human oocytes reprogram somatic cells to a
pluripotent state. Nature 478, 70-75.
Okita, K., Ichisaka, T., and Yamanaka, S. (2007). Generation of germline-competent induced
pluripotent stem cells. Nature 448, 313-317.
Okita, K., Nakagawa, M., Hyenjong, H., Ichisaka, T., and Yamanaka, S. (2008). Generation of
Mouse Induced Pluripotent Stem Cells Without Viral Vectors. Science (New York, NY.
Orkin, S.H., and Hochedlinger, K. (2011). Chromatin connections to pluripotency and cellular
reprogramming. Cell 145, 835-850.
Osborne, C.S., Chakalova, L., Brown, K.E., Carter, D., Horton, A., Debrand, E., Goyenechea, B.,
Mitchell, J.A., Lopes, S., Reik, W., et al. (2004). Active genes dynamically colocalize to shared
sites of ongoing transcription. Nature genetics 36, 1065-1071.
Palstra, R.J., Simonis, M., Klous, P., Brasset, E., Eijkelkamp, B., and de Laat, W. (2008).
Maintenance of long-range DNA interactions after inhibition of ongoing RNA polymerase II
transcription. PloS one 3, e1661.
Palstra, R.J., Tolhuis, B., Splinter, E., Nijmeijer, R., Grosveld, F., and de Laat, W. (2003). The
beta-globin nuclear compartment in development and erythroid differentiation. Nature genetics
35, 190-194.
Pan, G.J., Chang, Z.Y., Scholer, H.R., and Pei, D. (2002). Stem cell pluripotency and
transcription factor Oct4. Cell research 12, 321-329.
Papantonis, A., Larkin, J.D., Wada, Y., Ohta, Y., Ihara, S., Kodama, T., and Cook, P.R. (2010).
Active RNA polymerases: mobile or immobile molecular machines? PLoS biology 8, e1000419.
Park, I.H., Arora, N., Huo, H., Maherali, N., Ahfeldt, T., Shimamura, A., Lensch, M.W., Cowan,
C., Hochedlinger, K., and Daley, G.Q. (2008a). Disease-specific induced pluripotent stem cells.
Cell 134, 877-886.
Park, I.H., Lerou, P.H., Zhao, R., Huo, H., and Daley, G.Q. (2008b). Generation of human-
induced pluripotent stem cells. Nature protocols 3, 1180-1186.
Pereira, C.F., Piccolo, F.M., Tsubouchi, T., Sauer, S., Ryan, N.K., Bruno, L., Landeira, D.,
Santos, J., Banito, A., Gil, J., et al. (2010). ESCs require PRC2 to direct the successful
reprogramming of differentiated cells toward pluripotency. Cell stem cell 6, 547-556.
Philipsen, S., and Suske, G. (1999). A tale of three fingers: the family of mammalian Sp/XKLF
transcription factors. Nucleic acids research 27, 2991-3000.
118

Phillips, J.E., and Corces, V.G. (2009). CTCF: master weaver of the genome. Cell 137, 1194-
1211.
Pick, M., Stelzer, Y., Bar-Nur, O., Mayshar, Y., Eden, A., and Benvenisty, N. (2009). Clone- and
gene-specific aberrations of parental imprinting in human induced pluripotent stem cells. Stem
cells (Dayton, Ohio) 27, 2686-2690.
Plath, K., and Lowry, W.E. (2011). Progress in understanding reprogramming to the induced
pluripotent state. Nature reviews 12, 253-265.
Polo, J.M., Liu, S., Figueroa, M.E., Kulalert, W., Eminli, S., Tan, K.Y., Apostolou, E., Stadtfeld,
M., Li, Y., Shioda, T., et al. (2010). Cell type of origin influences the molecular and functional
properties of mouse induced pluripotent stem cells. Nature biotechnology 28, 848-855.
Pombo, A., Jackson, D.A., Hollinshead, M., Wang, Z., Roeder, R.G., and Cook, P.R. (1999).
Regional specialization in human nuclei: visualization of discrete sites of transcription by RNA
polymerase III. The EMBO journal 18, 2241-2253.
Probst, A.V., and Almouzni, G. (2011). Heterochromatin establishment in the context of
genome-wide epigenetic reprogramming. Trends Genet 27, 177-185.
Rashid, S.T., Corbineau, S., Hannan, N., Marciniak, S.J., Miranda, E., Alexander, G., Huang-
Doran, I., Griffin, J., Ahrlund-Richter, L., Skepper, J., et al. Modeling inherited metabolic
disorders of the liver using human induced pluripotent stem cells. The Journal of clinical
investigation 120, 3127-3136.
Raya, A., Rodriguez-Piza, I., Guenechea, G., Vassena, R., Navarro, S., Barrero, M.J., Consiglio,
A., Castella, M., Rio, P., Sleep, E., et al. (2009). Disease-corrected haematopoietic progenitors
from Fanconi anaemia induced pluripotent stem cells. Nature 460, 53-59.
Rodda, D.J., Chew, J.L., Lim, L.H., Loh, Y.H., Wang, B., Ng, H.H., and Robson, P. (2005).
Transcriptional regulation of nanog by OCT4 and SOX2. The Journal of biological chemistry
280, 24731-24737.
Rossant, J. (2008). Stem cells and early lineage development. Cell 132, 527-531.
Rowland, B.D., Bernards, R., and Peeper, D.S. (2005). The KLF4 tumour suppressor is a
transcriptional repressor of p53 that acts as a context-dependent oncogene. Nature cell biology 7,
1074-1082.
Samavarchi-Tehrani, P., Golipour, A., David, L., Sung, H.K., Beyer, T.A., Datti, A., Woltjen, K.,
Nagy, A., and Wrana, J.L. Functional genomics reveals a BMP-driven mesenchymal-to-
epithelial transition in the initiation of somatic cell reprogramming. Cell stem cell 7, 64-77.
Sandhu, K.S., Shi, C., Sjolinder, M., Zhao, Z., Gondor, A., Liu, L., Tiwari, V.K., Guibert, S.,
Emilsson, L., Imreh, M.P., et al. (2009). Nonallelic transvection of multiple imprinted loci is
organized by the H19 imprinting control region during germline development. Genes &
development 23, 2598-2603.
119

Santos, F., Hendrich, B., Reik, W., and Dean, W. (2002). Dynamic reprogramming of DNA
methylation in the early mouse embryo. Developmental biology 241, 172-182.
Schoenfelder, S., Sexton, T., Chakalova, L., Cope, N.F., Horton, A., Andrews, S., Kurukuti, S.,
Mitchell, J.A., Umlauf, D., Dimitrova, D.S., et al. (2010). Preferential associations between co-
regulated genes reveal a transcriptional interactome in erythroid cells. Nature genetics 42, 53-61.
Seibler, P., Graziotto, J., Jeong, H., Simunovic, F., Klein, C., and Krainc, D. (2011).
Mitochondrial Parkin recruitment is impaired in neurons derived from mutant PINK1 induced
pluripotent stem cells. J Neurosci 31, 5970-5976.
Shi, Y., Desponts, C., Do, J.T., Hahm, H.S., Scholer, H.R., and Ding, S. (2008). Induction of
pluripotent stem cells from mouse embryonic fibroblasts by Oct4 and Klf4 with small-molecule
compounds. Cell stem cell 3, 568-574.
Silva, J., Nichols, J., Theunissen, T.W., Guo, G., van Oosten, A.L., Barrandon, O., Wray, J.,
Yamanaka, S., Chambers, I., and Smith, A. (2009). Nanog is the gateway to the pluripotent
ground state. Cell 138, 722-737.
Simonis, M., Klous, P., Splinter, E., Moshkin, Y., Willemsen, R., de Wit, E., van Steensel, B.,
and de Laat, W. (2006). Nuclear organization of active and inactive chromatin domains
uncovered by chromosome conformation capture-on-chip (4C). Nature genetics 38, 1348-1354.
Soldner, F., Hockemeyer, D., Beard, C., Gao, Q., Bell, G.W., Cook, E.G., Hargus, G., Blak, A.,
Cooper, O., Mitalipova, M., et al. (2009). Parkinson's disease patient-derived induced pluripotent
stem cells free of viral reprogramming factors. Cell 136, 964-977.
Solter, D. (2006). From teratocarcinomas to embryonic stem cells and beyond: a history of
embryonic stem cell research. Nature reviews 7, 319-327.
Spilianakis, C.G., Lalioti, M.D., Town, T., Lee, G.R., and Flavell, R.A. (2005).
Interchromosomal associations between alternatively expressed loci. Nature 435, 637-645.
Splinter, E., de Wit, E., Nora, E.P., Klous, P., van de Werken, H.J., Zhu, Y., Kaaij, L.J., van
Ijcken, W., Gribnau, J., Heard, E., et al. (2011). The inactive X chromosome adopts a unique
three-dimensional conformation that is dependent on Xist RNA. Genes & development 25, 1371-
1383.
Splinter, E., Heath, H., Kooren, J., Palstra, R.J., Klous, P., Grosveld, F., Galjart, N., and de Laat,
W. (2006). CTCF mediates long-range chromatin looping and local histone modification in the
beta-globin locus. Genes & development 20, 2349-2354.
Sridharan, R., Tchieu, J., Mason, M.J., Yachechko, R., Kuoy, E., Horvath, S., Zhou, Q., and
Plath, K. (2009). Role of the murine reprogramming factors in the induction of pluripotency. Cell
136, 364-377.
Stadtfeld, M., Apostolou, E., Akutsu, H., Fukuda, A., Follett, P., Natesan, S., Kono, T., Shioda,
T., and Hochedlinger, K. (2010a). Aberrant silencing of imprinted genes on chromosome 12qF1
in mouse induced pluripotent stem cells. Nature 465, 175-181.
120

Stadtfeld, M., Brennand, K., and Hochedlinger, K. (2008a). Reprogramming of pancreatic beta
cells into induced pluripotent stem cells. Curr Biol 18, 890-894.
Stadtfeld, M., Maherali, N., Borkent, M., and Hochedlinger, K. (2010b). A reprogrammable
mouse strain from gene-targeted embryonic stem cells. Nature methods 7, 53-55.
Stadtfeld, M., Maherali, N., Breault, D.T., and Hochedlinger, K. (2008b). Defining molecular
cornerstones during fibroblast to iPS cell reprogramming in mouse. Cell stem cell 2, 230-240.
Stadtfeld, M., Nagaya, M., Utikal, J., Weir, G., and Hochedlinger, K. (2008c). Induced
Pluripotent Stem Cells Generated Without Viral Integration. Science (New York, NY.
Stamatoyannopoulos, J.A., Goodwin, A., Joyce, T., and Lowrey, C.H. (1995). NF-E2 and GATA
binding motifs are required for the formation of DNase I hypersensitive site 4 of the human beta-
globin locus control region. The EMBO journal 14, 106-116.
Sugii, S., Kida, Y., Kawamura, T., Suzuki, J., Vassena, R., Yin, Y.Q., Lutz, M.K., Berggren,
W.T., Izpisua Belmonte, J.C., and Evans, R.M. (2010). Human and mouse adipose-derived cells
support feeder-independent induction of pluripotent stem cells. Proceedings of the National
Academy of Sciences of the United States of America 107, 3558-3563.
Sung, L.Y., Gao, S., Shen, H., Yu, H., Song, Y., Smith, S.L., Chang, C.C., Inoue, K., Kuo, L.,
Lian, J., et al. (2006). Differentiated cells are more efficient than adult stem cells for cloning by
somatic cell nuclear transfer. Nature genetics 38, 1323-1328.
Szabo, P.E., Hubner, K., Scholer, H., and Mann, J.R. (2002). Allele-specific expression of
imprinted genes in mouse migratory primordial germ cells. Mechanisms of development 115,
157-160.
Takahashi, K., Ichisaka, T., and Yamanaka, S. (2006). Identification of genes involved in tumor-
like properties of embryonic stem cells. Methods in molecular biology (Clifton, NJ 329, 449-458.
Takahashi, K., Okita, K., Nakagawa, M., and Yamanaka, S. (2007a). Induction of pluripotent
stem cells from fibroblast cultures. Nature protocols 2, 3081-3089.
Takahashi, K., Tanabe, K., Ohnuki, M., Narita, M., Ichisaka, T., Tomoda, K., and Yamanaka, S.
(2007b). Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell
131, 861-872.
Takahashi, K., and Yamanaka, S. (2006). Induction of pluripotent stem cells from mouse
embryonic and adult fibroblast cultures by defined factors. Cell 126, 663-676.
Taranger, C.K., Noer, A., Sorensen, A.L., Hakelien, A.M., Boquest, A.C., and Collas, P. (2005).
Induction of dedifferentiation, genomewide transcriptional programming, and epigenetic
reprogramming by extracts of carcinoma and embryonic stem cells. Molecular biology of the cell
16, 5719-5735.
Tesar, P.J., Chenoweth, J.G., Brook, F.A., Davies, T.J., Evans, E.P., Mack, D.L., Gardner, R.L.,
and McKay, R.D. (2007). New cell lines from mouse epiblast share defining features with human
embryonic stem cells. Nature 448, 196-199.
121

Thomson, J.A., Itskovitz-Eldor, J., Shapiro, S.S., Waknitz, M.A., Swiergiel, J.J., Marshall, V.S.,
and Jones, J.M. (1998). Embryonic stem cell lines derived from human blastocysts. Science
(New York, NY 282, 1145-1147.
Tiwari, V.K., McGarvey, K.M., Licchesi, J.D., Ohm, J.E., Herman, J.G., Schubeler, D., and
Baylin, S.B. (2008). PcG proteins, DNA methylation, and gene repression by chromatin looping.
PLoS biology 6, 2911-2927.
Tolhuis, B., Palstra, R.J., Splinter, E., Grosveld, F., and de Laat, W. (2002). Looping and
interaction between hypersensitive sites in the active beta-globin locus. Molecular cell 10, 1453-
1465.
Tsai, S.Y., Clavel, C., Kim, S., Ang, Y.S., Grisanti, L., Lee, D.F., Kelley, K., and Rendl, M.
(2010). Oct4 and klf4 reprogram dermal papilla cells into induced pluripotent stem cells. Stem
cells (Dayton, Ohio) 28, 221-228.
Urbach, A., Bar-Nur, O., Daley, G.Q., and Benvenisty, N. (2010). Differential modeling of
fragile X syndrome by human embryonic stem cells and induced pluripotent stem cells. Cell
stem cell 6, 407-411.
Urlinger, S., Baron, U., Thellmann, M., Hasan, M.T., Bujard, H., and Hillen, W. (2000).
Exploring the sequence space for tetracycline-dependent transcriptional activators: novel
mutations yield expanded range and sensitivity. Proceedings of the National Academy of
Sciences of the United States of America 97, 7963-7968.
Utikal, J., Polo, J.M., Stadtfeld, M., Maherali, N., Kulalert, W., Walsh, R.M., Khalil, A.,
Rheinwald, J.G., and Hochedlinger, K. (2009). Immortalization eliminates a roadblock during
cellular reprogramming into iPS cells. Nature 460, 1145-1148.
Vakoc, C.R., Letting, D.L., Gheldof, N., Sawado, T., Bender, M.A., Groudine, M., Weiss, M.J.,
Dekker, J., and Blobel, G.A. (2005). Proximity among distant regulatory elements at the beta-
globin locus requires GATA-1 and FOG-1. Molecular cell 17, 453-462.
Vernimmen, D., De Gobbi, M., Sloane-Stanley, J.A., Wood, W.G., and Higgs, D.R. (2007).
Long-range chromosomal interactions regulate the timing of the transition between poised and
active gene expression. The EMBO journal 26, 2041-2051.
Waddington, C.H. (1957). The strategy of the genes; a discussion of some aspects of theroetical
biology (London,: Allen & Unwin).
Wang, J., Rao, S., Chu, J., Shen, X., Levasseur, D.N., Theunissen, T.W., and Orkin, S.H. (2006).
A protein interaction network for pluripotency of embryonic stem cells. Nature 444, 364-368.
Wansink, D.G., Schul, W., van der Kraan, I., van Steensel, B., van Driel, R., and de Jong, L.
(1993). Fluorescent labeling of nascent RNA reveals transcription by RNA polymerase II in
domains scattered throughout the nucleus. The Journal of cell biology 122, 283-293.
Warren, L., Manos, P.D., Ahfeldt, T., Loh, Y.H., Li, H., Lau, F., Ebina, W., Mandal, P.K., Smith,
Z.D., Meissner, A., et al. (2010). Highly efficient reprogramming to pluripotency and directed
differentiation of human cells with synthetic modified mRNA. Cell stem cell 7, 618-630.
122

Wei, Z., Yang, Y., Zhang, P., Andrianakos, R., Hasegawa, K., Lyu, J., Chen, X., Bai, G., Liu, C.,
Pera, M., et al. (2009). Klf4 interacts directly with Oct4 and Sox2 to promote reprogramming.
Stem cells (Dayton, Ohio) 27, 2969-2978.
Wells, J., and Farnham, P.J. (2002). Characterizing transcription factor binding sites using
formaldehyde crosslinking and immunoprecipitation. Methods (San Diego, Calif 26, 48-56.
Wernig, M., Lengner, C.J., Hanna, J., Lodato, M.A., Steine, E., Foreman, R., Staerk, J.,
Markoulaki, S., and Jaenisch, R. (2008). A drug-inducible transgenic system for direct
reprogramming of multiple somatic cell types. Nature biotechnology 26, 916-924.
Wernig, M., Meissner, A., Foreman, R., Brambrink, T., Ku, M., Hochedlinger, K., Bernstein,
B.E., and Jaenisch, R. (2007). In vitro reprogramming of fibroblasts into a pluripotent ES-cell-
like state. Nature 448, 318-324.
Williams, A., Spilianakis, C.G., and Flavell, R.A. (2010). Interchromosomal association and
gene regulation in trans. Trends Genet 26, 188-197.
Wilmut, I., Schnieke, A.E., McWhir, J., Kind, A.J., and Campbell, K.H. (1997). Viable offspring
derived from fetal and adult mammalian cells. Nature 385, 810-813.
Wilson, K.D., Venkatasubrahmanyam, S., Jia, F., Sun, N., Butte, A.J., and Wu, J.C. (2009).
MicroRNA profiling of human-induced pluripotent stem cells. Stem cells and development 18,
749-758.
Wolfe, S.A., Nekludova, L., and Pabo, C.O. (2000). DNA recognition by Cys2His2 zinc finger
proteins. Annu Rev Biophys Biomol Struct 29, 183-212.
Woltjen, K., Michael, I.P., Mohseni, P., Desai, R., Mileikovsky, M., Hamalainen, R., Cowling,
R., Wang, W., Liu, P., Gertsenstein, M., et al. (2009). piggyBac transposition reprograms
fibroblasts to induced pluripotent stem cells. Nature 458, 766-770.
Wu, S.M., and Hochedlinger, K. (2011). Harnessing the potential of induced pluripotent stem
cells for regenerative medicine. Nature cell biology 13, 497-505.
Wu, Y., Zhang, Y., Mishra, A., Tardif, S.D., and Hornsby, P.J. (2010). Generation of induced
pluripotent stem cells from newborn marmoset skin fibroblasts. Stem cell research 4, 180-188.
Xie, D., Cai, J., Chia, N.Y., Ng, H.H., and Zhong, S. (2008). Cross-species de novo identification
of cis-regulatory modules with GibbsModule: application to gene regulation in embryonic stem
cells. Genome research 18, 1325-1335.
Xu, M., and Cook, P.R. (2008). Similar active genes cluster in specialized transcription factories.
The Journal of cell biology 181, 615-623.
Yamanaka, S. (2009). Elite and stochastic models for induced pluripotent stem cell generation.
Nature 460, 49-52.
123

Yazawa, M., Hsueh, B., Jia, X., Pasca, A.M., Bernstein, J.A., Hallmayer, J., and Dolmetsch, R.E.
Using induced pluripotent stem cells to investigate cardiac phenotypes in Timothy syndrome.
Nature 471, 230-234.
Ying, Q.L., Nichols, J., Chambers, I., and Smith, A. (2003). BMP induction of Id proteins
suppresses differentiation and sustains embryonic stem cell self-renewal in collaboration with
STAT3. Cell 115, 281-292.
Ying, Q.L., Wray, J., Nichols, J., Batlle-Morera, L., Doble, B., Woodgett, J., Cohen, P., and
Smith, A. (2008). The ground state of embryonic stem cell self-renewal. Nature 453, 519-523.
Yu, J., Hu, K., Smuga-Otto, K., Tian, S., Stewart, R., Slukvin, II, and Thomson, J.A. (2009).
Human Induced Pluripotent Stem Cells Free of Vector and Transgene Sequences. Science (New
York, NY.
Yu, J., Vodyanik, M.A., He, P., Slukvin, II, and Thomson, J.A. (2006). Human embryonic stem
cells reprogram myeloid precursors following cell-cell fusion. Stem cells (Dayton, Ohio) 24,
168-176.
Yu, J., Vodyanik, M.A., Smuga-Otto, K., Antosiewicz-Bourget, J., Frane, J.L., Tian, S., Nie, J.,
Jonsdottir, G.A., Ruotti, V., Stewart, R., et al. (2007). Induced pluripotent stem cell lines derived
from human somatic cells. Science (New York, NY 318, 1917-1920.
Yusa, K., Rad, R., Takeda, J., and Bradley, A. (2009). Generation of transgene-free induced
pluripotent mouse stem cells by the piggyBac transposon. Nature methods 6, 363-369.
Zhang, J., Lian, Q., Zhu, G., Zhou, F., Sui, L., Tan, C., Mutalif, R.A., Navasankari, R., Zhang, Y.,
Tse, H.F., et al. (2011). A human iPSC model of Hutchinson Gilford Progeria reveals vascular
smooth muscle and mesenchymal stem cell defects. Cell stem cell 8, 31-45.
Zhao, H.X., Li, Y., Jin, H.F., Xie, L., Liu, C., Jiang, F., Luo, Y.N., Yin, G.W., Wang, J., Li, L.S.,
et al. (2010). Rapid and efficient reprogramming of human amnion-derived cells into
pluripotency by three factors OCT4/SOX2/NANOG. Differentiation; research in biological
diversity 80, 123-129.
Zhao, X.Y., Li, W., Lv, Z., Liu, L., Tong, M., Hai, T., Hao, J., Guo, C.L., Ma, Q.W., Wang, L.,
et al. (2009). iPS cells produce viable mice through tetraploid complementation. Nature 461, 86-
90.
Zhao, Z., Tavoosidana, G., Sjolinder, M., Gondor, A., Mariano, P., Wang, S., Kanduri, C.,
Lezcano, M., Sandhu, K.S., Singh, U., et al. (2006). Circular chromosome conformation capture
(4C) uncovers extensive networks of epigenetically regulated intra- and interchromosomal
interactions. Nature genetics 38, 1341-1347.
Zhou, H., Wu, S., Joo, J.Y., Zhu, S., Han, D.W., Lin, T., Trauger, S., Bien, G., Yao, S., Zhu, Y.,
et al. (2009). Generation of Induced Pluripotent Stem Cells Using Recombinant Proteins. Cell
stem cell.
Zhou, W., and Freed, C.R. (2009). Adenoviral gene delivery can reprogram human fibroblasts to
induced pluripotent stem cells. Stem cells (Dayton, Ohio) 27, 2667-2674.
124

Zwaka, T.P., and Thomson, J.A. (2005). A germ cell origin of embryonic stem cells?
Development (Cambridge, England) 132, 227-233.

Abstract (if available)

Abstract The discovery of induced pluripotent stem cells (iPSCs) has transformed the research of stem cells and provided infinite possibilities in regenerative medicine. In classical Yamanaka protocol, somatic cells from various sources can be reprogrammed to iPSCs with forced expression of Oct4, Sox2, Klf4, and cMyc. Numerous other combinations of factors and various delivery methods have also been developed to optimize the efficiency and accustom to different applications. These iPSCs are useful for disease modeling, toxicology studies and cell therapy. However, the molecular mechanisms of this transformation remain largely unclear. The transition from somatic cells to iPSCs involved comprehensive changes on epigenetic level of the cells induced by reprogramming factors. Reprogramming factor Klf4 can physically interact with Oct4 and Sox2. These three transcription factors co-occupy promoters of many pluripotency related genes, such as endogenous Nanog and Oct4. The physical interactions depend on the C2H2 zinc fingers in Klf4. Abrogation of these interactions will lead to failure of reprogramming due to the inability of defective complexes in activating key downstream genes. These results suggest that direct interactions between reprogramming factors are essential for initiating key downstream genes. During reprogramming, nuclear architecture of the cells also experience dramatic changes. In pluripotent stem cells (PSCs), endogenous Oct4 loci interact with distant regions in cis and in trans. Many of these long range interactions are specific to PSCs. PSC-specific interchromosomal interactions are established prior to transcriptional activation of endogenous Oct4 during reprogramming. In PSCs, Oct4-colocalized domains are enriched in active genes and pluripotency factor binding. Transcription of Oct4 is facilitated when the Oct4 locus is co-localized with its interchromosomal partners. Finally, depletion or overexpression of Klf4 causes changes in interchromosomal interactions prior to loss of Oct4 transcription and PSC differentiation, suggesting that Klf4 regulates interchromosomal interactions independent of its role as a transcription factor. Together these results reveal two novel essential factors in facilitating reprogramming: physical interactions between reprogramming factors and nuclear architecture dynamics.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Exploring stem cell pluripotency through long range chromosome interactions

PDF

Derivation and characterization of human embryonic stem (hES) cells and human induced pluripotent stem (hiPS) cells in clinical grade conditions

PDF

IPS and CNS cell models of peroxisomal disorders

PDF

Roles of Klf4 in embryonic stem cells

PDF

Novel roles for Maf1 in embryonic stem cell differentiation and adipogenesis

PDF

Epigenetic plasticity of cultured female human embryonic stem cells and regulation of gene expression and chromatin by PR-SET7 mediated H4K20me1

PDF

Investigating molecular roadblocks to enhance direct cellular reprogramming

PDF

CpG poor promoter SULT1C2 regulated by DNA methylation and is induced by cigarette smoke condensate in lung cell lines

PDF

Functional DNA methylation changes in normal and cancer cells

PDF

Deletion of Krüppel-like factor 4 (KLF4) affects NPCs to embark on differentiative progression and leads to delayed neurogenesis in mouse embryonic cortex

PDF

Investigation of the molecular mechanisms of ototoxicity

PDF

Differential role of two coactivators, CCAR1 and CARM1, for dysregulated beta-catenin activity in colorectal cancer cell growth and gene expression

PDF

Role of STAT3 phosphorylation in mouse embryonic stem cell self-renewal and differentiation

PDF

Pleotropic potential of Stat3 in determining self-renewal, apoptosis, and differentiation in mouse embryonic stem cells

PDF

Breast epithelial cell type specific enhancers and functional annotation of breast cancer risk loci

PDF

The molecular mechanism underlying the autoimmune-associated PTPN22 R620W variation and the quest for therapeutics

PDF

Identification of novel epigenetic biomarkers and microRNAs for cancer therapeutics

PDF

The role of HES/HEY transcriptional repressors in specification and maintenance of cell fate in the mouse organ of Corti

PDF

Transcriptional regulation in nephron progenitor cells

PDF

Characterization of human embryonic stem cell derived retinal pigment epithelial cells for age-related macular degeneration

Asset Metadata

Creator Wei, Zong (author)

Core Title The mechanisms of somatic cell reprogramming

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Genetic, Molecular and Cellular Biology

Publication Date 07/30/2012

Defense Date 06/05/2012

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag mechanism,molecular,OAI-PMH Harvest,reprogramming

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Stallcup, Michael R. (committee chair), Coetzee, Gerhard (Gerry) A. (committee member), Lu, Wange (committee member), Segil, Neil (committee member)

Creator Email zongwei@gmail.com,zongwei@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-77661

Unique identifier UC11289121

Identifier usctheses-c3-77661 (legacy record id)

Legacy Identifier etd-WeiZong-1070.pdf

Dmrecord 77661

Document Type Dissertation

Rights Wei, Zong

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA