University of Southern California Dissertations and Theses

Tissue interactions & molecular pathways in specification of the ectomesenchyme from cranial neural crest

(USC Thesis Other)

Tissue interactions & molecular pathways in specification of the ectomesenchyme from cranial neural crest

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
TISSUE INTERACTIONS & MOLECULAR PATHWAYS IN SPECIFICATION OF THE
ECTOMESENCHYME FROM CRANIAL NEURAL CREST

by

Ankita Das

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)

May 2012
Copyright 2012 Ankita Das

ii

ACKNOWLEDGEMENTS
I am extremely grateful to Dr. Gage D Crump for providing me with the opportunity to
pursue my graduate studies in his laboratory and also for the constant guidance and
support in my endeavors.
I would like to thank Drs Rob Maxson, Henry Sucov, Wange Lu, and Ellen Lien for willing
to serve on my thesis committee and for guidance and support in developing my thesis
project and with career development advice.
I would also like to acknowledge Dr. Samuel G Cox for mentoring me when I initially
started in the lab.
I would like to thank Megan Matsutani for her immense support not just as a manager
of the lab, but also for her guidance and mentoring during my initial years in the lab and
for her helpful scientific discussions as a colleague , and for help with experiments as
and when necessary.
I would like to thank Elizabeth Zuniga, Dr. Bartek Balczerski, Dr. Sandeep Paul and
Simone Schindler for sharing ideas and reagents for successful execution of experiments
and supporting me in my efforts.

iii

TABLE OF CONTENTS
Acknowledgements ii

List of Figures v

Abstract viii

Chapter1: Introduction
The Ectomesenchyme 1
Origin and Importance of the Ectomesenchyme 1
Cranial Versus Trunk Neural Crest 3
Mechanisms of Specification in the Trunk Neural Crest 5
Temporal Versus Spatial Models of Lineage Segregation 5
Tissue Interactions in Regulation of Lineage 6
Cell Intrinsic Molecules in Regulating Fate Switch 7
Mechanisms of Specification of the Cranial Neural Crest 9
Temporal Versus Spatial Bases of Lineage Segregation of The Cranial Neural Crest 10
Cell Extrinsic Factors in Specification of The Lineages Of the Cranial Neural Crest 10
Cell Intrinsic Molecules in Regulating Fate Switch 11
Role of Migration in Specification of The
Different Lineages of the Cranial Neural Crest 11
Zebrafish in Neural Crest Development 12
Cranial Neural Crest Development in Zebrafish 13

Chapter 2: Role of Cell Extrinsic Factors in Specification of the
Ectomesenchyme Lineage of Cranial Neural Crest: Introduction 17
Materials and Methods 20
Results
The Ectoderm is a Likely Source of Bmps
in Ectomesenchyme Specification
Change in Bmp Activity in the Migratory Neural Crest 43
Bmp4 Misexpression Inhibits Ectomesenchyme Specification 46
Loss of Bmp Signaling in the Neural Crest Results in Delayed
Migration of The Neural Crest and Loss of Non-Ectomesenchyme Markers 48
Id2a is Regulated by Bmps and Excluded from The Ectomesenchyme 50
Misexpression of Id2a Results in Similar Defects in Specification of
Ectomesenchyme as Misexpression of Bmp4 52
Id2a is Required for Bmps to Inhibit Ectomesenchyme Formation 54
Twist1 Transcription Factor is A Candidate for Mediating Bmp function 56
iv

Twist1a and Twist1b Have Redundant and Synergistic Roles in Specification of The
Ectomesenchyme Lineage 57
Loss of Twist1 Results in Defects in Specification of Ectomesenchyme 58
Possible Mechanisms of Id2a Mediated Twist1 Function in Specification
of Ectomesenchyme 61
Twist1 Function is Required in Migratory CNCCs for Ectomesenchyme Specification 62
Twist1 function Non-autonomously to cause migration defects in the CNCCs 63
Fgf Signaling Regulates Part of Ectomesenchyme Specification
Defining the Temporal Basis of Specification of the Ectomesenchyme 66
Discussion 69

Chapter 3: Defining the Molecular Bases for Twist1 Function in Specification of the
Ectomesenchyme Lineage during Cranial Neural Crest Development
Introduction 75
Materials and Methods 76
Results 86
Loss of Twist1 Does Not Maintain Neural Crest Cells in an Undifferentiated State 87
Twist1 Functions in the CNCC to Repress
Expression of the Non-ectomesenchyme Genes
Identification of Novel Ectomesenchyme Specific Genes Downstream of Twist1 89
Twist1 regulates Fli1a Likely Directly 97
Loss of Fli1a Results in Defects in Ectomesenchyme Formation 98
Introduction of Fli1a mRNA into Fli1a Loss of Function Results
In More Severe Phenotype 99
Upregulation of Sox10 Results in Defects in Ectomesenchyme Formation 100
Reduction of Sox10 in Twist1a/1b-MO Partially Rescues Skeletal Defects 101
Twist1 Regulation of Sox10 via Regulation of Fgf Signaling 103
Discussion 105

Chapter 4: Conclusions and future directions 110

References 117

v

LIST OF FIGURES
Figure 1.1: Schematic Representation of Cranial Neural Crest
Development into Ectomesenchyme in Zebrafish 14

Figure 2.1: Schematic Representation of the Possible Models for
Specification of the Ectomesenchyme Lineage 17

Figure 2.2: Mesoderm and Endoderm are Not Essential for Ectomesenchyme 42

Figure2.3: Bmp Activity is Selectively Down-regulated in
Ectomesenchyme Precursors 44

Figure2.4: Time-course of sox10:Gal4VP16-dependent Trans-gene expression 45

Figure 2.5: Misexpression of Bmp4 in Migrating CNCCs
Inhibits Ectomesenchyme Formation 46

Figure 2.6: Induction and Early Migration of CNCCs
is Unaffected by Bmp4 Misexpression 47

Figure 2.7: Cell Death Analysis in sox10:Gal4VP16; UAS: Bmp4 embryos. 48

Figure 2.8: Effect on the Non-ectomesenchyme Upon
Treatment with Bmp Inhibitor LDN-193189 48

Figure 2.9: Effect of Dorsomorphin Treatment on Embryos 49

Figure 2.10: Id2a is Regulated by Bmps and Excluded from the Ectomesenchyme. 51

Figure 2.11: Exclusion of Id2a Precedes Exclusion
of sox10 from the Ectomesenchyme 52

Figure2.12: Forced Expression of Id2a in CNCCs
Inhibits Ectomesenchyme Specification 53

vi

Figure 2.13: Knockdown of Id2a in the Background of Bmp4 Misexpression
Rescues Gene Expression Defects in Ectomesenchyme Specification 55

Figure 2.14: Twist1a and Twist1b Function Redundantly to Specify Ectomesenchyme 57

Figure 2.15: Twist1 Genes are Required for Ectomesenchyme Specification
in Zebrafish 60

Figure 2.16: Cell Death in Twist1a/1b MO Injected Embryos 61

Figure 2.17: Crest Autonomous and Non-autonomous Functions of Twist1 64

Figure 2.18: Fgf Signaling Regulates a Subset of Ectomesenchyme Development 66

Figure 2.20: Birthdating of Ectomesenchyme Versus Non-ectomesenchyme
by sox10:KikGR Photoconversion
at 4 Somite Stage Embryos 67

Figure 2.21: Photo-conversion of Neural Crest at Different Stages During
Development Shows Difference in Contribution of CNCC
to Ectomesenchyme versus Non- ectomesenchyme 68

Figure 2.22: Model for Specification of the Ectomesenchyme Lineage from CNCC 69

Figure3.1: Fate of the Ectomesenchyme Crest in Twist1a/1b-MO 86

Figure 3.2: Gene Expression Profiling in Twist1a/1b-MO Embryos 88

Figure 3.3: Validation of Novel Downstream Targets of Twist1 89

Figure 3.4: Expression of Genes from the Microarray and in the Non-canonical
Wnt Signaling Pathway 90

Figure 3.5: Gene-ontology Analyses of Micro-array Data Using Arraystar 92

Figure 3.6: Validation of Microarray Data From Table 3.1 95

Figure 3.7: Twist1 Directly Regulate Fli1a Expression in Ectomesenchyme 96

Figure 3.8: Loss of Fli1a Results in Defects in Ectomesenchyme Formation 99

vii

Figure3. 9: Misexpression of Hsox10 Results in Defects in Ectomesenchyme 101

Figure 3.10: Twist1 Regulates Fgf Activity 104

Figure 3.11: Model for Function of Twist1 in the Cranial neural crest 105

viii

ABSTRACT
Vertebrate cranial neural crest cells (CNCCs) contribute not only to ectodermal lineages
like neurons, glia and pigment but also to “ectomesenchymal” lineages like cartilage and
bone. Whereas studies have established that in zebrafish the CNCCs are lineage
restricted at the neural tube, the molecular bases for regulation of the cell lineage
remains unknown. In this thesis work, I will discuss my studies on the role of Bmp
signaling from the ectoderm in restricting the ectomesenchyme potential of the CNCC. I
will provide evidence for functions of Id2a and Twist1 proteins in specification of the
ectomesenchyme. We show that although twist1 genes are expressed in the CNCC
starting at pre-migratory stages, presence of Id2a in these cells prevents Twist1 from
functioning, and that a loss of Id2a in the migratory CNCCs over time in development
facilitates specification of the ectomesenchyme lineage. Furthermore, I will discuss a
detailed characterization of the roles of Twist1 in specification of the ectomesenchyme
lineage in zebrafish. We propose that Twist1 functions as a master-regulator in
specification of the ectomesenchyme lineage via regulating induction of early
ectomesenchyme genes and repressing the non-ectomesenchyme.
1

CHAPTER 1: INTRODUCTION
The ectomesenchyme:
Origin and importance of the ectomesenchyme
The neural crest is a transient, migratory cell population that arises at the boundary
between the neural and non-neural ectoderm (Le Douarin, 1999) . As a developing
vertebrate embryo undergoes neurulation, neural crest is induced at the border of
neural and non-neural ectoderm via response to inductive signals from juxtaposed
tissues: the ectoderm and the mesoderm. As the embryo develops, these cells undergo
epithelial to mesenchymal transition to migrate out of the neural tube to their final
destinations in the embryo and then later in development, they differentiate into their
bonafide cell types. Although both cranial and trunk neural crest cells differentiate into
non-ectomesenchyme derivatives, such as neurons, glia and pigment cells, CNCCs also
generate ectomesenchyme derivatives, in particular many of the cartilage-, bone-, and
teeth-forming cells of the head (Baroffio et al., 1991). Soon after the cranial neural crest
cells start migrating in their bonafide streams, the cells that have migrated most
ventrally from the neural tube contribute to the precursors of the ectomesenchyme.
Later in development, these ectomesenchymal precursor cells will condense and
differentiate into craniofacial cartilage. In contrast cells that migrate out later from the
neural tube and reside in the dorsal aspects of the embryo, express a different set of
2

markers from the ectomesenchyme precursors, and subsequently differentiate into
neurons, glia and pigment.
Besides the unique developmental origin of the ectomesenchyme that has interested
developmental biologists, this cell type has also been of interest to evolutionary
biologists about the question of evolution of vertebrates. In 1983, Gans & Northcutt
proposed “The new head hypothesis’’, which emphasizes that the neural crest was an
important embryonic feature in facilitating the evolution of vertebrates by allowing
them to evolve predatory behaviors as well as better feeding mechanisms (Northcutt
and Gans, 1983).
In addition, recently, given the multi-potent nature of the cell population, and its
promises as an ideal source for multi-potent adult stem cells, stem cell biologists have
also turned their attention to understand the regenerative potential of the
ectomesenchyme (Teng and Labosky, 2006) .
Although the ectomesenchyme is of central importance to biologists, the molecular and
cellular bases for specification of the ectomesenchyme from the multi-potent cranial
neural crest remains debated (Weston et al., 2004).

3

Cranial versus trunk neural crest
Not only are cranial and trunk neural crest cells specified at different times during
embryonic development, they contact very different sets of neighboring tissues, have
distinct migratory behaviors, and do so along very different pathways. Whereas the
cranial neural crest cells are the first ones to migrate out in a rostro-caudal manner, the
trunk neural crest cells follow. The cranial neural crest cells migrates out of the
ectoderm and over the underlying the head mesoderm and then contact the endoderm
(Knight and Schilling, 2006). For the most part, the cranial neural crest cells migrate in
close contact with the surface ectoderm. The trunk neural crest cells instead migrate
much deeper into the embryo and through deeper tissue layers like in between the
somites (Le Douarin, 1999; Theveneau and Mayor, 2011).
These neighboring tissues could provide signals that induce or inhibit specific lineages of
the neural crest. Also these tissues could provide guidance cues to the migratory crest
which allows them to migrate along or avoid certain paths (Theveneau and Mayor,
2011).
The cranial neural crest migrates along two discrete pathways: cells emigrating from
midbrain level migrate to populate the fronto-nasal prominences and the first
pharyngeal arch, and cells migrating from the hindbrain levels migrate ventrally to
populate the pharyngeal arches (Knight and Schilling, 2006). Interestingly whereas
cranial neural crest cells migrate as a collected group of cells, and subsequently split into
4

their bonafide streams as they migrate, the trunk neural crest cells leave the neural tube
as chains of cells at a later stage in embryonic development. Given the distinct cell
biology of the cranial versus the trunk neural crest, it might be important to carefully
study how these differences arise and how they contribute to the final lineage outcomes
of the neural crest.
Whereas they express a common set of bonafide neural crest markers (Meulemans and
Bronner-Fraser, 2004), the cranial neural crest cells subsequently express a distinct set
of transcription factors; for example Twist1 in some species of vertebrates like zebrafish
and mice (Germanguz et al., 2007; Hopwood et al., 1989) . In vitro clonal analysis studies
have shown that both the head and the trunk neural crest are multipotent in time and
well as space (Baroffio et al., 1991), lineage tracing studies both in the head and the
trunk have indicated that cells are fate restricted in vivo (Dorsky et al., 1998; Schilling
and Kimmel, 1994). Therefore, although the neural crest cells themselves might have
the potential to form all lineages, the environment in the context of the embryo might
restrict the potential to certain lineages versus others (Dupin et al., 2004).
Since several studies have highlighted the importance of signals from neighboring
tissues in specification of the lineages of the neural crest, I will review the literature
briefly to highlight the commonalities and differences between the cranial and trunk
neural crest cells.
5

In order to understand how during development, the different factors within the
embryo and within the neural crest work together to determine the lineage outcomes of
the multi-potent neural crest, one needs to first have a clear idea about the lineage
state of the neural crest during their different stages along its development. It is
possible that these cells are lineage restricted at the neural tube or they might undergo
progressive lineage restriction as they migrate out of the neural tube to their final
destinations. It is also possible that they are lineage restricted to begin with, and over
time they change their intrinsic properties to migrate along certain routes where they
can then differentiate(Harris and Erickson, 2007).
Mechanisms of specification in the trunk neural crest:
Temporal versus spatial models of lineage segregation:
In the trunk, neural crest cells migrate out in two waves. Previous studies have
established that depending on the timing of birth, cells are destined to contribute to
neuro-glial versus the pigment lineages (Henion and Weston, 1997). Signaling molecules
at the neural tube, signals along the migratory routes, and cell intrinsic factors regulate
the final lineage outcomes of the trunk neural crest (Theveneau and Mayor, 2011).
Whereas at the origin of the neural crest in the neural tube, there is an inbuilt temporal
mechanism for attribution of an initial fate restriction to these cells, as the cells leave
the neural tube, there exists a spatial difference attributed by signaling molecules along
6

the migratory routes which determine the final differentiation outcomes of the trunk
neural crest lineages (Theveneau and Mayor, 2011).
In the following paragraph, I discuss the molecular bases for the temporal and spatial
differences that contribute to the final lineage outcomes of the trunk neural crest.
Tissue interactions in regulation of lineage choice:
In the trunk, studies have shown that Wnt3a at the neural tube promotes
melanogenesis and Bmp4 signaling regulates neuro-glial specification (Jin et al., 2001).
Subsequently, these neural crest cells emigrating in the first wave of migrating trunk
crest take the ventral route since they are inhibited from taking the dorso-lateral route
by the cell-extrinsic factors expressed in the surrounding environment (Santiago and
Erickson, 2002). Subsequently, cells that take the ventral route need to be sorted for
differential development into neurons versus glia. For differential development of the
ventral population of the neural crest, a role of cell-cell communication is highlighted
since change in the cell-cell communications disrupt the patterning of the glia derived
from the trunk crest (Santiago and Erickson, 2002) .
Furthermore, these migrating out in the first wave are intrinsically programmed at the
neural tube to express molecules that allow them to migrate along the ventral route as
a response to repulsive cues in the path of dorso-lateral migratory pathway; the
7

alternate route of migration. Taken together, these studies highlight the importance of
signals from the surrounding tissues in controlling the migratory routes of the cells.
Cell intrinsic molecules in regulating fate switch:
Cells migrating out in the first wave are intrinsically programmed to contribute to the
neuro-glial lineages versus pigment lineages (Henion and Weston, 1997). This is
regulated by the expression of the transcription factor foxd3 specifically in the first wave
of cells, which in turn represses expression of mitf (important for specification of the
pigment lineages). Therefore, loss of foxd3 expression in the second wave of neural
crest emigration regulates the fate switch between the neural crest cells migrating in
the first and the second wave of migration (Thomas and Erickson, 2009) . Subsequently,
expression of chemokine receptors defines the migratory pathways of the trunk crest. In
addition, there are several transcription factors that further specify and promote
differentiation of the non-ectomesenchyme lineages like neurons, glia and pigment
(Thomas and Erickson, 2009).
Taken together, during trunk neural crest development, evidence from several studies
shows that signaling between the neural crest and the neighboring tissues, as well as
communication between the cells determine the migratory routes of the neural crest.
Migration along the different routes then determines the final lineage outcomes of the
neural crest (Theveneau and Mayor, 2011).
8

Mechanisms of specification of the cranial neural crest
Whereas studies have suggested that cranial neural crest is lineage restricted, the
precise timing of when ectomesenchyme might be specified remains unknown. This is
likely due to the fact that unlike the trunk crest where change in the expression of
transcription factor foxd3 determines when the fate switch occurs in the migratory
trunk neural crest (Thomas and Erickson, 2009), no such transcription factor has been
identified in the cranial crest that would regulate the fate switch between the
ectomesenchyme and the non-ectomesenchyme precursors. In the head, genes like
Twist1 are expressed in migratory neural crest along all streams (Germanguz et al.,
2007) but a function in regulating specification of the ectomesenchyme lineage at the
expense of the non-ectomesenchyme has not been shown. Further, whereas genes like
dlx2a and fli1a are uniquely expressed in the ectomesenchyme lineage as the cranial
neural crest cells progress in development, dlx2a appears to be dispensable for
ectomesenchyme formation (Sperber et al., 2008) and the function of fli1a in
ectomesenchyme development remains unknown. Although the lack of specification
defects in the dlx2a loss of function can be attributed to functional redundancy due to
expression of dlx1a in the same population of cells, a lack of definitive markers of the
ectomesenchyme lineage has hindered the progress of understanding the molecular
basis for specification of the ectomesenchyme. In addition, based on the expression of
the known markers of the ectomesenchyme, CNCC specification into ectomesenchyme
9

appears to be a process with progressive lineage restriction where they progressively
express different sets of markers as they migrate (shown in Figure 1.1 for zebrafish).
Further, the cranial crest cells migrate out in bonafide paths from the midbrain levels to
the fronto-nasal prominence and from the hindbrain levels to the pharyngeal arches. In
addition although the cranial neural crest cells that migrate along bonafide pathways,
the molecular bases for how the cells collectively migrate along these routes remains
unknown. Taken together both the molecular and cellular basis for specification of the
ectomesenchyme remains unknown.
Temporal versus spatial mechanisms of lineage specification of the cranial neural
crest:
Whereas in the trunk, several studies have provided a comprehensive understanding of
the molecular and cellular bases of specification of the neural crest lineages, in the head
specification of the cell- lineages of the neural crest still remains debated(Weston et al.,
2004). Although in the head lineage tracing studies have indicated lineage restriction
(Dorsky et al., 1998; Schilling and Kimmel, 1994), and subsequent studies have provided
evidence for signals in the neural tube that regulate neuro-glial versus pigment lineages,
there exists no evidence for molecular regulation of the ectomesenchyme versus the
non-ectomesenchyme lineages (Dorsky et al., 1998). One possible reason is that
although in the trunk, there is evidence for the temperal lineage switch in the neural
crest precursors guided by Wnt signaling and Bmp signaling pathways; which then
10

determines specification of the neuro-glial versus pigment lineages (Jin et al., 2001),
and in the head although a spatial mechanism for specification of the neuro-glial versus
pigment lineages, the ectomesenchyme precursors do not segregate into any of the pre-
specified domains described in the lineage tracing studies (Schilling and Kimmel, 1994).
Cell extrinsic factors in specification of the lineages of the cranial neural crest
Dorsky and colleagues have shown that neural crest cells in the vicinity versus at a
distance from the medially expressed Wnt3a signaling are specified to form pigment
versus neuro-glial lineages (Dorsky et al., 1998). It was proposed by Graham and
colleagues that CNCCs might acquire ectomesenchyme identity upon arrival in the
pharyngeal arches, potentially as a result of endoderm-secreted Fgfs (Blentic et al.,
2008). Previous studies have suggested roles for Fgf signaling, in particular Fgf20b and
Fgfr1, in ectomesenchyme specification in avians and zebrafish (Blentic et al., 2008;
Yamauchi et al., 2011). Although there exists substantial evidence for a role of Fgf
signaling in specification of the ectomesenchyme, the timing of specification as well as
the tissues of importance in this process still remains debatable and will require further
investigation.

11

Cell intrinsic factors
As mentioned previously, a handful of genes like Twist1, Dlx2 and Fli1a (Germanguz et
al., 2007; Liu et al., 2008; Sperber et al., 2008) have been shown to be expressed in the
ectomesenchyme precursor population, starting at different times during development.
Although these genes are expressed in the ectomesenchyme precursor population in
multiple species, their roles in instructing the neural crest to form ectomesenchyme has
not been defined. In mice, Twist loss of function results in loss of ectomesenchyme
(Bildsoe et al., 2009; Soo et al., 2002). But the function of Twist1 in regulating the
lineage switch between ectomesenchyme and non-ectomesenchyme precursors
remains to be established.
Role of migration in specification of the different lineages of the CNCC
Signals that define the migratory routes are known. Although how CNCCs migrate in a
collective fashion has recently been characterized (Theveneau et al., 2010), how specific
subsets of the CNCC generate differential collective migratory behaviors remains to be
explored. In addition, roles for cell-cell communication and cell-extracellular matrix
communications in determining the differential migration and patterning of the
ectomesenchyme versus non-ectomesenchyme precursors populations remains to be
determined. Identification of novel candidate molecules will be necessary to completely
understand how cranial neural crest cells communicate with each other and regulate
12

differential migration and sorting to segregate into distinct domains of
ectomesenchyme and non-ectomesenchyme.
Having underlined the various themes during development that could determine the
final lineage outcomes of the CNCC into ectomesenchyme versus non-ectomesenchyme,
I will discuss in this thesis our studies on understanding the roles of all of the above
mentioned factors, and try to propose a model for comprehensive understanding of the
development of the ectomesenchyme development from the cranial neural crest.
Zebrafish in neural crest development
In the next few paragraphs, I introduce the model system that I used to understand the
process of specification of the ectomesenchyme lineage from CNCC. Zebrafish is an
important organism of choice for these studies for several reasons: Not only is the
genetics of the organism is well established, there are several transgenic tools available
for studying genetic pathways. Furthermore, given the transparent nature of the
embryos, they are amenable to live imaging studies, which makes it suitable to study in
detail cellular events that facilitate morphogenesis. In addition, the zebrafish genome
had undergone duplication, which has resulted in sub-functionalization of a lot of the
homologous proteins in other species of vertebrates. Since sub-functionalization results
in similar proteins sharing the functions that a conserved protein would normally have
in other species, it allows often surpassing embryonic lethality caused by an early
13

developmental gene (accounted by genetic redundancy) and studying its later functions
in developmental processes.
Cranial neural crest development in zebrafish
In my graduate studies, I utilized the zebrafish larval head skeleton to understand
molecular and cellular mechanisms underlying specification of the ectomesenchyme
lineage. The entire process of CNCC development into the craniofacial cartilage occurs in
4 days. Summarized in the figure to follow, are the steps and the molecular markers
used to define the transition in the different steps of development of the CNCC into
ectomesenchyme derived head-skeleton. In zebrafish, CNCCs are first apparent within
the anterior neural plate border at 10.5 hours-post-fertilization (hpf), when they begin
to express sox10, foxd3, sox9b, and tfap2a. Within the next few hours, three streams of
CNCCs can be seen migrating away from the neural tube to more ventral positions.
Starting around 15.5 hpf, ectomesenchyme precursors begin to down regulate early
CNCC genes such as sox10, foxd3, sox9b, and tfap2a (Kelsh et al., 2000; Meulemans and
Bronner-Fraser, 2004) and upregulate ectomesenchyme-specific genes such as dlx2a
(Sperber et al., 2008) and fli1a (Brown et al., 2000). These ectomesenchyme cells then
go on to populate a series of pharyngeal arches from which develops the support
skeleton of the jaw and gills in zebrafish, and the jaw, middle ear, and larynx in
mammals(Wada et al., 2005).

14

Cranial neural crest development in zebrafish:

Figure 1.1: Schematic representation of cranial neural crest development into
ectomesenchyme in zebrafish: Lateral view of the zebrafish embryo shows migrating cranial
neural crest cells. On the right, confocal images of fluorescent in situ hybridization shows
expression of sox10 in the pre-migratory neural crest at 11hpf. As the embryo develops, and
cells migrate out of the neural tube, they continue to express sox10. (e) Indicates ear staining of
sox10. Subsequently, cells that have migrated out ventrally express dlx2a (shown in green) and
downregulate expression of sox10 in the ectomesenchyme domain. Cells that reside more
dorsally, shown in red, continue to express sox10 and will later on differentiate into the non-
ectomesenchyme lineage. Schematic on the left shows that cells that express dlx2a will
condense to form transient epithelial structures called the pharyngeal arches (shown in green).
These pharyngeal arch crest cells then start to express fli1a shown on the right. Cells express
dlx2a and fli1a will then differentiate into the craniofacial cartilage and bone starting at 4dpf.

15

In chapter 2, I will discuss the role of Bmp signaling from the ectoderm and role of Fgf
signaling in the migratory crest in specification of the ectomesenchyme. Also, I will
discuss a role of Bmp signaling from the ectoderm in regulation of the precise temporal
basis for specification of the ectomesenchyme via expression of id2a in the neural crest.
I will then discuss our results on genetic interactions between Id2a and transcription
factor Twist1 in specification of the ectomesenchyme lineage. Since we show
molecularly, that there is a Bmp dependent fate switch in the CNCC to form
ectomesenchyme versus non-ectomesenchyme, I will discuss our lineage tracing to
determine when the lineage switch occurs. Taken together, based on our molecular and
cellular analyses, I will propose a temporal model for specification of the
ectomesenchyme.
In Chapter 3, I discuss what we find about the function of Twist1 in regulation of the
ectomesenchyme lineage. I will present evidence that Twist1 regulates the lineage
switch between the ectomesenchyme and the non-ectomesenchyme lineage by
repressing expression of non-ectomesenchyme specific genes in the ectomesenchyme
lineage and in the non-ectomesenchyme precursors. Furthermore, I also discuss how
molecules downstream of Twist1 like fli1a, sox10 and other newly identified candidates
function in specification of the ectomesenchyme. Taken together, from this chapter I
propose that Twist1 is a master-regulator for specification of the ectomesenchyme
lineage of CNCC. I will conclude with some preliminary evidence showing that Twist1
16

might function to regulate expression of several classes of molecules important for
regulation of cell-cell communication, cell adhesion re-organization of the cell-
extracellular matrix to facilitate sorting of cells into different populations.

17

Chapter 2: ROLE OF CELL EXTRINSIC FACTORS IN SPECIFICATION
OF THE ECTOMESENCHYME LINEAGE OF CRANIAL NEURAL CREST
Introduction:

Figure 2.1: Schematic representation of the possible models for specification of the
ectomesenchyme lineage: A. Shows lateral views of the head of a zebrafish embryo with
migrating neural crest cells (in green). (B) Cross-section through the neural tube shows that pre-
migratory neural crest (shown in green) receives signals from the neural and the non-neural to
be specified, which introduces the idea of signal from the origin. (C) Cross-section through the
neural tube at a later stage in development shows that neural crest cells can also receive signals
from the pharyngeal endodermal tissue that contacts the neural crest as they contribute to the
pharyngeal arches.

Although it is known that CNCCs can receive signals either at their origin at the neural
tube or at their destination in the pharyngeal arches (Figure2.1), there is no definitive
evidence for either. Not only do the tissues of interest remain to be characterized, the
molecular signals from those tissues directing the cells down a certain lineage remain to
be discovered. In this chapter, I will first discuss our studies to identify the neighboring
tissues of the neural crest in providing signals for specification of the ectomesenchyme
by utilizing a previously established genetic tissue ablation studies in zebrafish. I will
18

then discuss our findings on the roles of signaling pathways like Bmp and Fgf in the
regulation of the ectomesenchyme lineage. Further, I will then discuss our studies on
how these signaling pathways mediate their function in the specification of the
ectomesenchyme lineage. Also, I will discuss our findings on what molecules in the
migratory CNCC mediate the function of these signaling pathways.
Bmps such as bmp2b in zebrafish are prominently expressed in the non-neural
ectoderm, where they influence neural crest induction, migration, survival, and
differentiation (Liem et al., 1995; Nguyen et al., 1998; Steventon et al., 2009). However,
a role for Bmp signaling in ectomesenchyme lineage decisions has not been previously
described, likely because of the earlier essential roles of Bmps in neural crest induction
(Nguyen et al., 1998). Using transgenic tools, we characterize a precise temporal
requirement of Bmp signaling in regulating the fate switch between the
ectomesenchyme and the non-ectomesenchyme.
Id genes are widely expressed in the early neural crest, and Id2 has been shown to
promote neural crest at the expense of epidermis in avians (Martinsen and Bronner-
Fraser, 1998). In zebrafish, Id2a has been shown to regulate neuron and glia formation
in the retina, albeit non-cell-autonomously, yet its role in CNCC development has not
been explored (Uribe and Gross, 2010). Here we show a novel role of Id2a in CNCC
lineage decisions, with down-regulation of id2a in migrating CNCCs being essential for
ectomesenchyme specification.
19

One factor critical for ectomesenchyme specification reported previously in mouse is
the basic-helix-loop-helix (bHLH) transcription factor Twist1. Both the conventional
Twist1 knockout and a conditional Twist1 neural-crest-specific (Wnt1-CRE, Twist-flox-
flox) knockout displayed defective ectomesenchyme development, including abnormal
perdurance of Sox10 and loss of expression of many arch ectomesenchyme genes
(Bildsoe et al., 2009; Soo et al., 2002). Furthermore, the neural-crest-specific knockout
showed severe reductions of the CNCC-derived craniofacial skeleton, although the lower
jaw was less affected. Zebrafish have two Twist1 orthologs, with twist1b being
expressed in early CNCCs and twist1a restricted to ectomesenchyme precursors from 16
hpf onwards (Germanguz et al., 2007). Given the known expression pattern of twist1
genes and the reported craniofacial abnormalities in mouse, we characterize in detail
the function of Twist1 in specification of the ectomesenchyme lineage.
Taken together, in this chapter from our detailed genetic analyses of the roles of Bmp
signaling, Id2a and Twist1, we propose a pathway that defines the molecular bases for
specification of the ectomesenchyme lineage from CNCC and also provide insights into
the temporal and spatial bases for specification of the ectomesenchyme lineage from
the cranial neural crest.

20

Materials and methods:
Zebrafish lines and transgenic constructs
Zebrafish were staged as described (Kimmel et al., 1995). Previously reported lines
include Tg(hsp70I:Gal4)
kca4
(Scheer and Campos-Ortega, 1999), Tg(~3.4her5:EGFP)
ne1911

(Tallafuss and Bally-Cuif, 2003), Tg(sox10:kikGR)
el2
(Balczerski et al., 2012) and
Tg(UAS:Bmp4; cmlc2:GFP)
el49
(Zuniga et al., 2011). Tg(sox10:Gal4VP16)
el159
,
Tg(UAS:mKR; cmlc2:GFP)
el15
, Tg(UAS:kikGR; -crystallin:Cerulean)
el377
,

Tg(UAS:Id2a; -
crystallin:Cerulean)
el405
, Tg(UAS:dnFgfr1a; cmlc2:GFP)
el28
, Tg(UAS:dnTwist1b;
UAS:mcherryCAAX; cmlc2:GFP)
el179
, Tg(sox10:dsRED)
el10
, lines were generated using
Gateway cloning (Invitrogen) and the Tol2 kit as described (Kwan et al., 2007).
sox10:Gal4VP16 was created by combining p5E-sox10, pME-Gal4VP16, p3E-IRES-GFP-
pA, and pDestTol2pA2. The IRES-GFP was later found to be non-functional in stable
transgenics. UAS: mKR, UAS:kikGR, and UAS:dnFgfr1a were generated by combining
p5E-UAS; pME-mKR, pME-kikGR, or pME-dnFgfr1a; p3E-polyA; and pDestTol2CG2. UAS:
dnTwist1b was made by combining p5E-UAS, pME-dnTwist1b, p3E-UAS:
mCherryCAAXpA, and pDestTol2CG2. The dnTwist1b clone was made by first amplifying
the full-length zebrafish twist1b cDNA with primers rTwist1bikL: 5’-
GGGGACAAGTTTGTACAAAAAAGCAGGCTCGGCCACCATGCCCGAAGAGCCCGCGGAGA-3’
and Twrk: 5’-
GGGGACCACTTTGTACAAGAAAGCTGGGTACTTAGATGCAGACATGGACCAAGCGCC-3’. PCR
21

based mutagenesis was then performed with primers twdnEK1L: 5’-
GCGCTTTCGCAGAGACTTA-3’ and twdnEK1R 5’-CTGTCGCTTGCGCACGTT-3’ as described
(Heckman and Pease, 2007) to change nucleotide G250 of the twist1b cDNA to A, which
results in conversion of amino acid 84E to K, with the PCR product cloned into
pDNOR221 to create pME-dnTwist1b. UAS: twist1bAlanine and UAS: twist1b-glutamic
acid transgenic lines were constructed by combining p5E-UAS, pME-twist1b-alanine,
p3E-pA, and pDestTol2CG2 and p5E-UAS, pME-twist1b-glutamic acid, p3E-pA, and
pDestTol2CG2 respectively. The twist1b-alanine and glutamic acid clones were made by
first amplifying the full-length zebrafish twist1b cDNA with primers rTwist1bikL: 5’-
GGGGACAAGTTTGTACAAAAAAGCAGGCTCGGCCACCATGCCCGAAGAGCCCGCGGAGA-3’
and Twrk: 5’-
GGGGACCACTTTGTACAAGAAAGCTGGGTACTTAGATGCAGACATGGACCAAGCGCC-3’. PCR
based mutagenesis was then performed with primers AlanineB: 5’- TTC GTT GAG AGC
CTG AGC CCT CTG TCG CTC GCG CA GCGCTTTCGCAGAGACTTA-3’ and Alanine C 5’- TGC
GCG AGC GAC AGA GGG CTC AGG CTC TCA ACG AA CTGTCGCTTGCGCACGTT-3’ as
described (Heckman and Pease, 2007) to change nucleotide G250 of the twist1b cDNA
to A and the PCR product was then cloned into pDNOR221 to create pME-Twist1b
alanine. To construct pME-dnFgfr1a, we amplified a truncated form of zebrafish fgfr1a
based on the previously reported dominant-negative Fgf receptor in Xenopus laevis
(Amaya et al., 1991) and cloned the PCR product into pDONR221; primers used were 5’-
GGGGACAAGTTTGTACAAAAAAGCAGGCTCCACCATGATAATGAAGACCACGCTG-3’ and
22

5’GGGGACCACTTTGTACAAGAAAGCTGGGTCTAAGAGCTGTGCATTTTGGC-3’. Full-length
zebrafish id2a was amplified with primers id2a-F: 5’-
GGGGACAAGTTTGTACAAAAAAGCAGGCTCGGCCACCATGAAGGCAATAAG-3’ and id2a-R:
5’-GGGGACCACTTTGTACAAGAAAGCTGGGTGTTAACGGTAAAGTGTCCT-3’ and then
cloned into pDONR221 to generate pME-Id2a. UAS: Id2a was generated by combining
p5E-UAS, pME-Id2a, p3E-polyA, and pDestTol2AB2. sox10: dsRED was generated by
combining p5E-sox10, p3E-polyA, and pDestTol2PA2 with pME-dsRed. UAS:mKR was
made by combining p5E-UAS, pME-mKR, p3E-polyA, and pDestTol2CG2. pME-mKR was
made by PCR amplification of membraneKillerRed (mKR) with primers kilredL: 5’-
GGGGACAAGTTTGTACAAAAAAGCAGGCTCGGTCGCCACCATGCTGTG-3’ and kilredR: 5’-
GGGGACCACTTTGTACAAGAAAGCTGGGTTTAATCCTCGTCGCTACCGA-3’ and insertion into
pDONR221. UAS:kikGR was made by combining p5E-UAS, pME-kikGR , p3E-polyA, and
pDestTol2AB2.
To construct stable transgenics, one-cell-stage embryos were injected with vectors and
transposase RNA and multiple stable lines were identified for each transgene:
Tg(sox10:Gal4VP16)

(1), Tg(UAS:Id2a; -crystallin:Cerulean) (5),
Tg(UAS:dnTwist1b:UAS:mcherryCAAX; cmlc2:GFP) (7), Tg(UAS:dnFgfr1a; cmlc2:GFP) (6),
Tg(UAS:mKR; cmlc2:GFP) (2), Tg(UAS:kikGR; cmlc2:GFP) (3), Tg(sox10:dsRED)

(1),
Tg(sox10:LOX-GFP-LOX-hDLX3) (2), and Tg(fli1-F-hsp70I:GFP)

(2). For
sox10:Gal4VP16/UAS:transgene experiments, genotyping for Gal4 was performed using
23

G4L: CTCCCAAAACCAAAAGGTCTCC and G4R: TGAAGCCAATCTATCTGTGACGG and cmlc2
:cmcl2-L 5_-TGGTGCAGATGA -ACTTCAGG-3 and cmcl2-R 5’-
TGCTGGAATCTGAGCACTTG-3’- confirmed the observed phenotypes, except for
sox10:Gal4VP16/UAS:Id2a experiments where genotyping for Id2a was performed with
primers id2aF: 5’-AGAACACCCCTGACAACACT-3’ and id2aR: 5’-
GCTAATACGACTCACTATAGGTACCGGCAGTCCAATTTC-3’.
Skeletal analysis: Alcian Blue staining of cartilage and Alizarin Red staining of bone at 5
dpf was performed according to the double staining protocol. In situ hybridizations were
performed according to the protocol from Kimmel lab with modifications.
Fluorescent in situ hybridization: Fluorescent in situ hybridization was performed by
following the protocol kindly shared by Jared Talbot from Charles Kimmel lab in
University of Oregon, Eugene. For making the probes, the protocol is listed below. I used
half reaction for making DIG probes and full reaction for DNP probes.
Things to prepare for probe synthesis: For synthesis of DNP labeled probe,
Need to make up:
1. 20XNTP mix: Mix: 10µl each of ATP, GTP, CTP (100mM NTP’s obtained from
Amersham Bioscience (cat# 27202501). with 6.5µl UTP in 13.5µl nuclease free H2O
resulting in: 20mM each ATP, GTP, CTP, 13mM UTP stock.
24

2. 20X DNP-11-UTP stock: Obtain a 250nmol/25µl stock of DNP-11-UTP from Perkin-
Elmer (cat# NEL555001EA). Add 10.7µl nuclease free H2O for a 10mM stock
Probe synthesis:
The DIG probe synthesis materials can be obtained from Roche in a kit (Cat#
11175025910), and the DNP probe synthesis reagents can also be obtained in the same
kit except for the ones listed above.
Protocol for in-situ hybridizations:
Use multistage cDNA to generate temperaturelate for probe synthesis
Protocol for cDNA synthesis:
1. Use 20 embryos per stage. Add 200ul of Lysis/Binding buffer
2. Disrupt tissue by a small plastic pestle
3. Add 200ul of 64 % ethanol to the lysate. Mix gently by pipetting several times
4. Centrifuge for one minute until the lysate/ ethanol mixture has gone through the
filter
5. Discard flow through
6. Pre-heat elution solution at 75 C ( 50ul per cartridge)
7. Apply 700ul of wash solution #1 to the filter cartridge, and centrifuge for one
minute until the wash solution has gone through the filter
8. Wash 2 times with 500ul of wash solution 2/3 and centrifuge for one minute and
discard flow through
9. Put the filter cartridge into a fresh collection tube
25

10. Pipette pre-heated solution to the center of the filter. Close the cap of the tube
11. Elute RNA with 40ul pre-heated solution and centrifuge for 30 seconds.
12. In the same tube, elute RNA with 10ul of pre-heated elution solution and re-
centrifuge for 30 seconds
13. Perform DnaseI in-activation using the following steps:
14. Add 5ul of 10X DnaseI buffer and add 1ul of DnaseI
15. Mix gently and incubate at 37 C for 30 minutes
16. Add 5.6ul of DnaseI inactivation reagent ( vortex prior to use)
17. Mix gently and incubate at room temperature for 2 minutes
18. Centrifuge for one minute, RNA will be in solution and the DnaseI inactivation
reagent will settle to the bottom as a pellet.
19. Proceed to cDNA synthesis using Retroscript kit from Ambion:
20. Add 2ug of total RNA
21. Add 2ul of random decamers
22. Mix and heat at 85 C for 3 minutes
23. Add the remaining components:
2ul of 10X RT buffer
4ul of dNTP mix
1ul of RNase inhibitor
1ul of MMLV-RT reverse transcriptase
24. Mix gently
26

25. Incubate at 44C for one hour
26. Incubate at 92C for 10 minutes
27. Store at -20c
1. Mix together:
DNP Probes DIG probes
1-2µg linear plasmid 1-2µG linear plasmid/ 8-9ul of PCR product
1µl 20X NTP 2µl DIG NTP mix
1µl 20X DNP-11-UTP
2µl 10X TXN buffer 2µl 10X TXN buffer
1µl RNAse inhibitor 1µl RNAse inhibitor
2µl T7/SP6/T3 2µl T7/SP6/T3
Nuclease free H2O to 20µl Nuclease free H2O to 20µl
2. Incubate 2 hours @ 37˚C
3. Mix in 2µl DNAse I
4. Incubate 15 minutes at 37˚C
5. Add 0.8µl EDTA
27

6. Add 2µl 5M LiCl
7. Add 75µl Prechilled 100% Etoh
8. Place @ -80˚C for one hour to overnight
9. Centrifuge 10 minutes
10. Remove liquid
11. Rinse with 200µl Prechilled 80% Etoh
12. Centrifuge 5 minutes
13. Remove as much liquid as possible
14. Air dry 5-10 minutes at room temperature
15. Resuspend in 30-40µl nuclease free H2O
16. Mix in 1µl RNAse inhibitor
16. Store at –80˚C ASAP
Diagnostic gel:
1. Mix 2µl probe with 5µl formamide, 3µl nuclease free H2O.
2. Heat this aliquot 3 minutes @ 68-70˚C
3. Run products 10-15 minutes on a 1.5% gel @ 130V
28

Anti-DNP-POD: Perkin-Elmer Cat#NEL747A001KT.
Anti-DIG-POD: Roche cat# 1207733 or individually. This kit can also be used for DNP
synthesis, for the reagents listed above.
In situ bybridization:
Embryo preparation:
1. Fix embryos overnight at 4˚c in 4% PFA/1X PBS or at room temperature for 2-3 hours.
2. Wash twice in PBST (1X PBS, 0.25% tween-20)
3. Dechorionate using a pair of watchmakers forceps
4. Dehydrate with a series of methanol/PBST solutions (25%, 50%, 75% methanol mixed
with PBST), then twice with 100% methanol. Shake 3-5 minutes in each solution.
5. Store the embryos in 100% methanol at –20˚C from 24 hours to several months

Day 1:
1. Rehydrate the embryos through a methanol/PBST series (75%, 50%, 25% methanol
mixed with PBST) 3-5 minutes per wash.
2. Wash five minutes in PBST, four times.
29

3. Treat the embryos with 1µg/ml proteinase K in PBST to increase the permeability of
the membrane. Time of ProK treatments are as follows:
Embryonic stage Length of ProK
24 hours 10 min
30 hours 20 min
36 hours 30 min
48 hours 45 min
55 hours 1 hour
72 hours 45 min 10ug/ml ProK
4. Wash the fish twice, quickly in PBST to remove ProK
5. Fix the embryos in 4% PFA/1XPBS for 20 minutes to ensure the ProK has stopped.
6. Wash 5 X 5 minutes in PBST
7. Add 500µl Pre-Hyb solution (50% formamide, 5X SSC, 100µg/ml yeast RNA, 50µg/ml
Heparin, 0.125% tween-20, Citric acid to pH 6.0)
8. Incubate at 68-70˚C for at least 2 hours. I’ve always done this at least four hours.
9. Replace prehyb with 200µl fresh Pre-Hyb.
30

10. Add no more than 200ng of each probe to this hyb solution.
11. Incubate overnight @ 68-70˚C
Antibody preabsorption:

(Done the day antibodies will be used)
1. Prepare100 embryos of mixed stages with in situ protocol steps 1-7, ending in pre-
hybridization solution.
2. Wash three times in TNT to get rid of the pre-hybridization solution.
3. Wash once in 400µl TBSTB (TNT with 0.5% Perkin-Elmer blocking powder). To get
blocking powder into solution, I shake the tubes one hour at 68-70˚C. Aliquots may be
stored at –20˚C, but once they have been thawed, never refreeze them.
4. Add TBSTB to the tubes, (volume=400µlX[number of samples+1])
5. Add anti-DIG-peroxidase (@ 1:500) or anti-DNP peroxidase (@ 1:200) to the tubes.
6. Shake several hours at room temperature.
7. After use, store the embryos in pre-hybridization solution. They can be re-used
several times. Be sure to label what antibody was used on those embryos, and only
reuse them for that antibody. To reuse, remove fish from freezer and enter pre-
absorption protocol on step 2. Typically, I use the pre-absorbed embryos 3-4 times.

31

Day 2:
12. Remove probes, and save them for reuse
13. Wash 5 minutes in “5X” (5X SSC, 50% formamide, 0.25% tween-20) @ 68-70˚C
14. Wash 5 minutes in 3:1 5X: 2X @ 68-70˚C
15. Wash 5 minutes in 1:1 5X: 2X @ 68-70˚C
16. Wash 5 minutes in 1:3 5X:2X @ 68-70˚C
17. Wash 5 minutes in “2X” (2x SSC, 0.25% tween-20) @ 68-70˚C
18. Wash 3 X 20 minutes in 0.2X SSC, 0.25% tween-20 @ 68-70˚C
19. Wash 2 X10 minutes in PBST @ room temperature
20. Replace with 500µl 2% Hydrogen-peroxide in PBST
21. Shake 60 minutes @ room temperature
22. Wash 4 X 5 minutes in TNT (0.1 M Tris-Hcl pH 7.5; 0.15 M NaCl; 0.5% Tween20)
23. Block at least two hours (I have typically done four hours) in 400µl TBSTB
24. Replace with 400µl pre-absorbed anti-DIG-POD 1:500.
25. Rock overnight @ 4˚C

32

Day 3:
26. Wash 8 X in TNT at room temperature over the course of 1-2 hours
27. Wash five minutes in Perkin Elmer Amplification Diluent
28. Prepare 1:50 Tyr-Fluorescein in Amplification Diluent for 50µlX(n+1) samples.
29. Replace amplification diluent with 50µl 1:50 Tyr-Fluorescein
30. Shake one hour in the dark, tubes upright. Do not exceed one hour. All steps from
here on out are done in the dark.
31. Wash 2 X 5 minutes in TNT
32. Wash one hour in 2%H2O2/TNT
33. Wash 4 X 5 minutes in TNT
34. Block with 400µl TBSTB 1-4 hours
35. Replace with preabsorbed 1:200 anti-DNP-peroxidase in TBSTB to the tubes.
36. Rock overnight at 4˚C.
Day 4:
26. Wash 8 X 10 minutes in TNT at room temperature over the course of 1-2 hours
27. Wash five minutes in Perkin Elmer Amplification Diluent
28. Prepare 1:50 Tyr-Cy3 in Amplification Diluent for 50µlX(n+1) samples.
33

29. Replace amplification diluent with 50µl 1:50 Tyr-Cy3
30. Shake one hour, tubes upright.
31. Wash 2 X 5 minutes in TNT
32. Wash one hour in 2%H2O2/TNT (optional. I find this improves the second stain’s
signal: noise somewhat)
33. Wash several times in TNT
33. If desired, leave in room temperature TNT washes overnight
Modifications to the protocol:
1. The order of developing the probes:
For double in-situ protocol, one needs to keep in mind that the stronger probe
signal is sometimes harder to quench, and can bleed through into the
development of the second probe, and therefore, it is better to develop the
weaker probe better. Also, I always label the stronger probe with DIG and the
weaker probe with DNP.
2. For the antibody pre-absorption, I have used about 50 embryos and have
successfully executed the protocol. I have noticed that embryos re-used for 3-4
times start to give background, so I collect fresh embryos for the pre-absorption
step after that.
34

Reported probes include sox9b (Yan et al., 2005), tfap2a (Furthauer et al., 1997), and
fli1a (Brown et al., 2000). id2a, sox10, fli1a, twist1a, ntla, dct, xdh, pea3, kikGR, mkr, and
foxd3 probes were synthesized with T7 RNA polymerase from PCR products amplified
from multistage zebrafish cDNA, with the exception of id2a that was amplified from an
id2a cDNA plasmid (Zebrafish International Resource Center). We used the following
primer pairs: id2a: id2aF and id2aR, twist1a: 5'-CAGAGTCTCCGGTGGACAGT-3' and 5'-
GCTAATACGACTCACTATAGGGTCTTTTCCTGCAGCGAGTC-3', ntla: 5’-
GACCACAAGGAAGTCCCAGA-3’ and 5’-
GCTAATACGACTCACTATAGGCATTGAGGAGGGAGAGGACA-3’, dct: 5’-
CGTACTGGAACTTTGCGACA-3’ and 5’-
GCTAATACGACTCACTATAGGACCAACACGATCAACAGCAG-3’, xdh: 5’-
TGAACACTCTGACGCACCTC-3’ and 5’-
GCTAATACGACTCACTATAGGTGTTGAAGCTCCAGCAACAC-3’, foxd3: 5’-
CGGCATTGGGAATCCATA-3’ and 5’-
GCTAATACGACTCACTATAGGCAACGAAATGAAATAGAAAGAAGGA-3’, pea3: 5’-
CCCATATGATGGTCAAACAGG-3’ and 5’-
GCTAATACGACTCACTATAGGATTGTCGGGAAAAGCCAAG-3’, kikGR: 5’-
GAAGATCGAGCTGAGGATGG-3’ and 5’-
GCTAATACGACTCACTATAGGGCTCGTACAGCTTCACCTTGT-3’, mkr: 5’-
GGGCAGAAGTTCACCATCG-3’ and 5’-
GCTAATACGACTCACTATAGGGTGGTGAAGCCGATGAAGG-3’, sox10: 5’-
35

TGCATTACAAGAGCCTGCAC-3’ and 5’-
GCTAATACGACTCACTATAGGAGGAGAAGGCGGAGTAGAGG-3’.
Immunostaining:
1. Fix embryos in 4% para-formaldehyde/PBS overnight at 4 C or at room
temperature for 3 hours. From experience it appears that fixing at room
temperature does not make any difference
2. Wash embryos 3x’s for 5 minutes with PBT
3. Rinse in PBS/1%DMSO/0.1%TritonX 3x’s for 5 minutes
4. Rinse in ddH
2
O for 5 minutes
5. Wash in cold acetone (should be kept at -20 C) for 10 minutes
6. Rinse in ddH
2
O for 5 minutes
7. Wash in PBT for 5 minutes
8. Rinse in PBS/1%DMSO/0.1%TritonX 3x’s for 5 minutes
9. Block for 3 hours in PBSTx/1%DMSO/10% Normal Goat Serum
10. Incubate with primary antibody rabbit anti p-SMAD1/5/8 (Cell Signaling, #9511S,
1:1000), overnight at 4 C
36

11. Wash 8x’s in PBS/1%DMSO/0.1%TritonX for 10 minutes each
12. Incubate with secondary antibody ( Alexa 568 1:1000) ( Invitrogen) overnight at
4 C
13. Rinse 4x’s in PBT for 30 minutes each
Fluorescent in situ & Immuno-staining combined protocol:
For combined fluorescent in situ hybridization and fluorescent immuno-staining
protocol process the embryos first for in situ, wash the embryos in TNT buffer and then
equilibriate the embryos in PBT and follow the immune-staining protocol listed above.
For anti-GFP immuno-staining, we used 1:1000 rabbit polyclonal anti-GFP primary
antibody (Torrey Pines Biolabs, East Orange, NJ, USA) and 1:300 AlexaFluor488 goat
anti-rabbit secondary antibody (Invitrogen).
Cell death analysis with Lysotracker Staining:
Day1:
1. Dechorionate live embryos of appropriate stage and transfer to a small dish
2. Add 25ul of Lysotracker red ( 1mM solution) to 5ml of embryo medium
3. Replace embryo medium from dish using small bore Pasteur pipette and replace
with Lysotracker Red ( Invitrogen)
37

4. Cover from light and incubate for 30- 45 minutes. I left it for 45 minutes
5. Rinse 4-5 times gently with embryo medium
6. Fix in 4% PFA/PBS overnight at 4 C with rocking
Day2:
7. Rinse in PBT
8. Dehydrate in methanol to reduce background by serial washes in ( 25, 50 and
75% methanol/PBT respectively)
9. Store in methanol at -20 overnight
10. Rehydrate into PBT
Morpholino and mRNA injections
One-cell-stage embryos were injected with 3 nl of the following translation-blocking
MOs (GeneTools, Philomath, OR, USA): twist1a-MO 5’-
ACCTCTGGAAAAGCTCAGATTGCGG-3’ (400 M), twist1b–MO 5’-
TTAAGTCTCTGCTGAAAGCGCGTG-3’ (400 M), twist1a/1b-MO (200/200 M), and id2a-
MO 5’-GCCTTCATGTTGACAGCAGGATTTC-3’ (800 M) (Uribe and Gross, 2010). For
morpholino validation, we performed rescue experiments with in vitro transcribed
mRNA generated using the Ambion Message machine kit (Applied Biosystems) from a
CMV/SP6-Twist1b template generated by combining p5E-CMV/SP6, pME-Twist1b, p3E-
38

polyA, and pDestTol2pA2. Capped mRNA was transcribed as described previously, and
then injections were performed with and without MO to examine GFP expression. In the
control embryos, no GFP expression was observed. In addition, we attempted to test
whether the morpholinos for twist1a and twist1b were blocking translation efficiently.
To test for that, we designed GFP fusion proteins for twist1a and twist1b to utilize the
change in GFP read-out upon morpholino injection as a means to detect efficient
translation. The experiment involves co-injection of GFP fusion version of RNA of
twist1a and twist1b with the morpholino and injection of the RNA encoding the RNA
alone and assessing the proteins of twist1a and twist1b with eGFP were generated by
fusion PCR method described previously, and cloned into middle entry vector Pdonr 221
as described previously. For generating twist1a-eGFP and twist1b-eGFP the following
primers were used: twist1begfpa : 5’- GGG GAC AAG TTT GTA CAA AAA AGC AGG CTT
TAT CTC CCT CCT CTC- 3’, and twist1begfpb- 5’- GCC CTT GCT CAC CAT CGC GGG AGA
CAC and twist1begfpc : GTG TCT CCC GCG ATG GTG AGC AAG GGC, twist1a-eGFP 5’-
GGG GAC AAG TTT GTA CAA AAA AGC AGG CTC GGC CAC CAT GTT TGA and GGG GAC
CAC TTT GTA CAA GAA AGC TGG GTT TAG TGA GAT GTT G. Pme-twist1b-eGFP was then
combined with p5E-CMV/SP6 and p3E into Pdesttol2ap2 destination vector to generate
CMV/SP6-twist1a-Egfp-polyA and CMV/SP6-twist1b-Egfp- polya respectively. Antivin
mRNA was synthesized and injected as described (Ragland and Raible, 2004). The
Antivin plasmid was obtained from David Raible Lab and linearized using their
instructions and mRNA was generated using the method described previously.
39

Imaging
Larval skeletons and colorimetric in situ hybridization embryos were imaged using a
Leica MZ16F dissecting scope using Camera Window software for a Canon60 Shot 80S
digital camera. Fluorescent images were captured using a Zeiss LSM5 confocal
microscope with Zen software. Identical gain settings were used to acquire fluorescent
in situ images across data sets. Images were processed using Photoshop CS4 with care
taken to apply identical adjustments throughout data sets.
KikGR photoconversion and imaging:
Embryos were raised in the dark at 28.5 C until 10.5hpf. At 10.5hpf, embryos were
carefully examined under Leica Dissecting scope MZ16F for expression of green
fluorescence of KikGR protein. Embryos are dechorionated with watch-makers ’ forceps
with minimum exposure to white light under the dissecting scope. These embryos were
then carefully mounted on a depression slide with a drop of embryo medium. After
mounting the embryos, they were exposed to ultraviolet radiation through the ultraviolet
filter on the dissecting scope for 30 seconds to a minute. After this process, we observed
the embryos under the red channel to detect photo-conversion from green to red. Once
all the green cells were photo-converted into red, the embryos were carefully recovered
into 10cm petri-dishes for observing and imaging later. Of the entire sample of embryos
were photo-converted, a representative sample was mounted using agarose onto cover
slip slides for confocal imaging using the procedure described previously.
40

Photoconversion of cells expressing sox10: kikGR were performed at stages 11.5hpf,
13hpf and 14hpf respectively.
Drug treatments:
Dorsomorphin Treatment: To treat embryos with dorsomorphin drug 71260 AMPK
Inhibitor, Compound C from EMD Chemicals: dissolve the drug in appropriate amount
of DMSO to make up a 10mM stock solution. Then for the experiment, make serial
dilutions in Embryo Medium to get a final concentration of 10um. Pipette the drug and
Embryo Medium mixture into 6 well plates with 10 embryos each, during 6-10hpf, 8-
12hpf and 10-14hpf. As a control, add equivalent amounts of DMSO only to Embryo
Medium.
LDN-193189 Treatment: LDN-193189 was purchased from Stemgent (catalog # 04-
0074). A stock solution of 10uM was made in DMSO and serial dilution was prepared in
Embryo Medium. The drug was added to embryos at 1uM and 10uM concentrations at
12hpf and 14hpf. As a vehicle control, equal amount of DMSO was added to the embryo
medium.
Heat shock treatments of dominant negative BmpR: For heat shocking
hsp70I:dnBmpr1a-GFP the clutch of embryos was divided into batches of 10 each, and
then aliquoted into 0.5ml eppendorf tubes with 30ul of embryo medium in each, and
then heat shocked at 40C in the thermo-cycler for 30 minutes. The embryos were then
transferred gently into 10cm petri-dishes and allowed to develop. For further
41

phenotypic analysis, they were sorted into GFP positive and negative populations. Heat
shocks were performed starting at 10.5hpf, 11hpf and 11.5hpf.
Statistical analysis: Statistics was performed with JMP7 software. One-way analysis of
variance was calculated with a Tukey-Kramer HSD test (alpha= 0.05) and standard errors
of the mean were plotted.
Results:
The ectoderm is a likely source of Bmps for ectomesenchyme specification
A previous study had reported that ablation of mesoderm and endoderm, by injection of
mRNA encoding the Nodal antagonist Antivin, resulted in severe reductions of dlx2a-
positive ectomesenchyme at 24 hpf, yet it was unclear whether the loss of dlx2a
reflected defects in the specification or later endoderm-mediated survival of
ectomesenchyme precursors (Ahlgren and Bronner-Fraser, 1999; Crump et al., 2004;
David et al., 2002; Kikuchi et al., 2001). We decided to revisit these experiments to
determine whether the mesoderm and the endoderm could be sources of signals for
induction of the ectomesenchyme lineage markers and downregulation of the early
neural crest/ non-ectomesenchyme markers. As reported previously, we found that
injection of Antivin mRNA eliminated mesoderm (as assayed by ntl and myod expression
at 12 hpf) and endoderm (as shown by her5: GFP pharyngeal endoderm staining at 24
hpf) (Fig. 2.2A-F).
42

Figure 2.2: -Mesoderm and endoderm are not essential for ectomesenchyme formation. (A-D)
In situs at 12 hpf show loss of axial ntl and mesodermal myod expression in Antivin-mRNA-
injected embryos compared to un-injected controls.(E and F) Confocal projections of her5:GFP
expression at 19 hpf show loss of pouch endoderm (arrows) but not brain (arrowheads) in
Antivin-mRNA-injected embryos compared to un-injected controls. (G and H) In situs at 15.5 hpf
show normal ectomesenchyme induction of dlx2a in Antivin-mRNA-injected embryos (n=12) and
un-injected controls (n=13). Arches are numbered.(I and J) Double fluorescent in situs at 15.5
hpf show normal induction of dlx2a (green) in sox10-positive CNCCs (red, yellow indicates co-
localization) in Antivin-mRNA-injected embryos (n=15) and un-injected controls (n=9).(K and L)
Double fluorescent in situs at 19 hpf show complementary expression of fli1a (green) in
ectomesenchyme and sox10 (red) in non-ectomesenchyme in Antivin-mRNA-injected embryos
(n=5) and un-injected controls (n=8). Scale bar = 50 m.

However, in contrast to the loss of dlx2a-expressing cells at 24 hpf, we observed three
roughly equivalent streams of dlx2a-positive CNCCs earlier at 15.5 hpf (Fig. 2.2 G, H),
with co-expression of dlx2a and sox10 confirming the CNCC identity of these cells (Fig.
2.2I, J). Furthermore, fli1a expression was largely unaffected in CNCCs of 19 hpf
Antivin-injected embryos (Fig.2.2 K, L). We therefore conclude that signals from the
mesoderm and endoderm are not required for the initial specification of ectomesenchyme,
43

although these tissues may play roles in the later maintenance, survival, and further
differentiation of ectomesenchyme derivatives. Hence, rather than signals deriving from
the arch endoderm and/or mesoderm (Blentic et al., 2008), our data suggests that
signals from the non-neural ectoderm, such as Bmps, might influence specification of
the ectomesenchyme fates.
Change in Bmp activity in the migratory neural crest
Based on our previous findings, we decided to examine possible signaling pathways
from the ectoderm that might be important for specification of the ectomesenchyme.
Since previous studies have shown that high Bmp signaling inhibits neural crest
specification (LaBonne and Bronner-Fraser, 1998) and loss of Bmps leads to severe
developmental defects (Nguyen et al., 1998), there have been no reports on the roles of
Bmp signaling in specification of distinct lineages of the CNCC. Given the genetic tools
we had developed in the laboratory that would allow us to manipulate Bmp signaling
after the initial window of specification, we decided to examine whether Bmp signaling
might play a role in specification of the ectomesenchyme. In order to visualize Bmp
activity in CNCCs, we performed immune-staining for the phosphorylated form of
Smad1/5/8 (pSmad1/5/8), which is thought to broadly indicate canonical Bmp activity ,
in embryos expressing sox10:GFP in the neural crest (Carney et al., 2006). Consistent
with the known role of Bmps in neural crest induction (Tribulo et al., 2003), all CNCCs
displayed high levels of pSmad1/5/8 at pre-migratory stages (12 hpf) (Fig. 2.3A).
44

However, by 16.5 hpf ectomesenchyme precursors had down-regulated pSmad1/5/8,
while non-ectomesenchyme precursors located more dorsally continued to exhibit high
levels of pSmad1/5/8 (Fig. 2.3B, C). Cross-sectional views demonstrated that ventral
CNCCs with low pSmad1/5/8 were located deep to the ectoderm whereas dorsal CNCCs
with high pSmad1/5/8 had yet to delaminate and remained in the neural plate ectoderm
(Fig. 2.3D).

Figure2.3: Bmp activity is selectively down-regulated in ectomesenchyme precursors. (A)
Confocal projection of pSMAD1/5/8 immunostaining at 12 hpf shows that all sox10: GFP-positive
CNCCs display high pSMAD1/5/8. (B-D) At 16.5 hpf, a projection (B) and a high-magnification
section (C, boxed area in B) show that pSMAD1/5/8 remains high in the dorsal non-
ectomesenchyme but is lower in the more ventral ectomesenchyme. An orthogonal section (D,
taken at the level of the line in C) shows that high pSMAD1/5/8 CNCCs remain in the neural
epithelium whereas low pSMAD1/5/8 CNCCs are positioned medial and ventral. The eye also
displays high pSMAD1/5/8. Abbreviations: D, dorsal; V, ventral; A, anterior; P, posterior; M,
medial; L, lateral; m, mandibular arch ectomesenchyme; h, hyoid arch ectomesenchyme; n, non-
ectomesenchyme; e, ectomesenchyme; ve, ventral epithelium. Scale bar = 50 m.

Since we observed a correlation between changing Bmp activity and specification of the
ectomesenchyme, we formulated the hypothesis that down-regulation of Bmp activity
in the migratory neural crest is necessary for specification of the ectomesenchyme. To
45

test our hypothesis, we decided to prolong Bmp activity in the migratory neural crest by
crossing sox10:gal4VP16 into UAS: Bmp4 zebrafish and obtaining embryos carrying both
transgenics. We first determined the time course for when Bmp activity by crossing the
sox10:gal4VP16 with a reporter line UAS: kikGR and performing in –situ hybridization for
kikGR transcript. We find that although at 11hpf, the pre-migratory neural crest had no
expression of kikGR, at 13hpf; kikGR was expressed in the migratory neural crest
(Figure2.4). Since we found that by crossing sox10:gal4VP16 into UAS: Bmp4
transgenics, we could prolong Bmp signaling in the neural crest after they migrate out of
the neural tube, we then carried out experiments to test our hypothesis.

Bmp4 misexpression inhibits ectomesenchyme specification
We next examined whether the observed down-regulation of Bmp activity in
ectomesenchyme precursors was necessary for their formation. Strikingly, Bmp4
misexpression in migratory CNCCs of sox10:Gal4VP16; UAS: Bmp4 embryos resulted in
Figure2. 4: Time-course of sox10:Gal4VP16-
dependent transgene expression. (A-D) In
situs for kikGR mRNA in sox10:Gal4VP16;
UAS:kikGR embryos show transgene
expression in migratory CNCCs at 13 hpf but
not in pre-migratory CNCCs at 11 hpf.
sox10:Gal4VP16 control embryos show no
expression.

46

profound defects in ectomesenchyme specification, including persistent sox10
expression and reduced expression of fli1a and dlx2a (Figure 2.5A-H).

Figure 2.5: Misexpression of Bmp4 in migrating CNCCs inhibits ectomesenchyme formation.
(A-D) Whole mount in situs at 18 hpf show ectopic expression of sox10 in the arches (numbered)
of sox10:Gal4VP16; UAS:mKR; UAS:Bmp4 embryos (n=8/8) compared to sox10:Gal4VP16;
UAS:mKR controls (n=0/8) and reductions of dlx2a in sox10:Gal4VP16; UAS:mKR; UAS:Bmp4
embryos (n=4/4) compared to sox10:Gal4VP16; UAS:mKR controls (n=0/4). Arrows indicate the
second arch and white arrowheads the developing ear. (E-H) Double fluorescent in situs for mKR
(red) and fli1a or dlx2a (green) at 24 hpf show reduction of fli1a arch expression in
sox10:Gal4VP16; UAS:mKR; UAS:Bmp4 embryos (n=5/5) compared to sox10:Gal4VP16; UAS:mKR
controls (n=0/9) and reduction of dlx2a arch expression in sox10:Gal4VP16; UAS:mKR;
UAS:Bmp4 embryos (n=4/4) compared to sox10:Gal4VP16; UAS:mKR controls (n=0/3). (I and J)
Skeletal staining at 5 dpf shows severe loss of craniofacial skeleton in sox10:Gal4VP16;
UAS:mKR; UAS:Bmp4 embryos (n=7/7) compared to sox10:Gal4VP16; UAS:mKR controls (n=0/9).
(K and L) In situs for foxd3 at 48 hpf reveal largely normal patterns of glia in sox10:Gal4VP16;
UAS:mKR; UAS:Bmp4 embryos (n=3) and sox10:Gal4VP16; UAS:mKR controls (n=5). (M-P) In
situs at 28 hpf show normal dct-positive melanophore precursors in sox10:Gal4VP16; UAS: mKR;
UAS:Bmp4 embryos (n=3) and sox10:Gal4VP16; UAS:mKR controls (n=8) and normal xdh-positive
xanthophore precursors in sox10:Gal4VP16; UAS:mKR; UAS:Bmp4 embryos (n=4) and
sox10:Gal4VP16; UAS:mKR controls (n=2). Scale bars = 50 m.

47

Importantly, the reduction of fli1a and dlx2a arch expression was not simply due to a
lack of CNCC formation, migration, or survival. The examination of sox10-positive CNCCs
in 15 hpf Bmp4-misexpression embryos revealed no major defects in CNCC induction or
migration (Figure 2.6), and Lysotracker staining also revealed no major increases in
CNCC apoptosis (Figure 2.7). Moreover, although fewer CNCCs were found in the arches
of Bmp4-misexpression embryos at later stages (as shown by lineage tracing using the
mKR red fluorescent protein at 24 hpf), those CNCCs found in the arches also displayed
reduced fli1a and dlx2a expression (Figure 2.5 E-H).

Figure 2.7: Cell death analysis in sox10:Gal4VP16; UAS: Bmp4 embryos. Compared to
sox10:Gal4VP16 only controls (n=3), Lysotracker staining reveals no major increase in cell death
at 24 hpf in sox10:Gal4VP16; UAS: Bmp4 embryos (n=3). Scale bar = 50 M

Figure 2.6: Induction and early migration of CNCCs is unaffected by Bmp4 misexpression. In
situs for sox10 at 15 hpf show no difference in early migrating CNCCs between sox10:Gal4VP16;
UAS: Bmp4 embryos (n=5) and sox10:Gal4VP16 only controls (n=12). Scale bar = 50 m.

48

Loss of Bmp signaling in the neural crest results in delayed migration of the neural
crest in the pharyngeal arches and loss of non-ectomesenchyme markers
Since we determined that loss of Bmp activity was essential to promote the
ectomesenchyme formation, we then asked whether loss of Bmp activity in the non-
ectomesenchyme would result in expansion of the ectomesenchyme marker fli1a in the
non-ectomesenchyme and loss of markers of the non-ectomesenchyme dct and xdh.

Figure 2.8: Effect on the non-ectomesenchyme upon treatment with Bmp inhibitor LDN-
193189: Confocal projections of double fluorescent in situ hybridization to detect expression of
dct and immunostaining for GFP to detect expression sox10: GFP shows that compared to
controls (A-A’’), embryos with loss of Bmp signaling (B-B’’) have loss of dct expression in the
neural crest derived pigment in the head and also a reduction in the pharyngeal arches.
Compared to controls (C-C’’), embryos with loss of Bmp signaling (D-D’’) have a slight change in
the pattern of expression of xdh, but no loss of expression, and there is defective migration or
delay in development of the neural crest into the pharyngeal arches.
49

Figure 2.9: Effect of Dorsomorphin treatment of embryos at 10-14hpf: In- situ hybridizations
to examine expression of sox10 and dlx2a shows that compared to controls (A,C), embryos
treated with Dorsomorphin have increased expression of sox10 in the migratory CNCC (B),
fusion of the streams of the neural crest, and no change in expression of dlx2a in the
ectomesenchyme (D).

In loss of Bmp function, there is increased cell death of the migratory neural crest.
When we examined expression of melanophore marker dct, we find that there is a loss
of expression in the head region of the embryo. In addition, there is a defect in
migration of the neural crest cells and loss of pharyngeal arch crest. There appears to be
increased cell death in the dorsal aspect of the embryo, which could result from a defect
in migration of the cells. When embryos were treated with another Bmp inhibitor
dorsomorphin, we observed that there was defect in migration of the neural crest, and a
fusion of the neural crest streams, which resulted in increased number of sox10 positive
cells in the dorsal aspect of the embryo. Furthermore, we did not observe any expansion
of the ectomesenchyme marker dlx2a into the non-ectomesenchyme domain (Figure
2.9). In addition, when also decided to block Bmp signaling by heat-shocking embryos
50

expressing heat-shock inducible form of dominant negative Bmp receptor. Consistent
with what we found with our drug treatment studies, there was a defect in migration of
the neural crest cells, but no expansion of the ectomesenchyme lineage into the non-
ectomesenchyme domain
id2a is regulated by Bmps and excluded from the ectomesenchyme.
In order to understand the molecular basis for function of Bmp activity in the migratory
neural crest, we decided to examine the expression of id genes in the neural crest over
the course of neural crest development into ectomesenchyme. Intriguingly, we found
that zebrafish id2a was broadly co-expressed with sox10 in early CNCCs at 12and 14hpf
(Figure 2.10 A-E’’) yet by 16.5 hpf was restricted to sox10-positive non-ectomesenchyme
and excluded from ectomesenchyme expressing dlx2a and twist1a (Figure 2.10 F-H’’).
Furthermore, upon misexpression of Bmp4 in the migratory neural crest, compared to
controls (Figure 2.10 F-F’’), we observed ectopic expression of id2a in the
ectomesenchyme precursor crest (Figure 2.10 I-I’’). Taken together, as the down-
regulation of id2a expression tightly correlates with ectomesenchyme specification, as
well as exclusion of id2a precedes the exclusion of sox10 from the ectomesenchyme
precursor crest, and we showed that id2a is a transcriptional target of Bmp4 in the
migratory neural crest, we hypothesized that down-regulation of id2a might be
important for specification of the ectomesenchyme lineage and might mediate the
function of Bmp signaling.
51

Figure 2.10:id2a is regulated by Bmps and is excluded from the ectomesenchyme. (A-C)
Colorimetric in situs show weak expression of id2a in pre-migratory CNCCs at 12 hpf (dorsal
views with anterior up, arrows indicate bilateral CNCC fields) and increasing expression in non-
ectomesenchyme precursors at 17.5 and 19 hpf (lateral views). (D-I) Confocal projections of
fluorescent in situs show co-localization of id2a with sox10 in CNCCs at 12 hpf (D) and 14hpf (E),
as well as co-localization in the non-ectomesenchyme at 16.5 hpf (F). At 16.5 hpf, id2a is
excluded from ectomesenchyme CNCCs marked by dlx2a (G) and twist1a (H), with twist1a also
being expressed in head mesoderm. In sox10:Gal4VP16; UAS: Bmp4 embryos, id2a and sox10
are expressed ectopically in the ectomesenchyme (white arrows). Arrowheads indicate otic
expression of sox10. Scale bars = 50 m
52

Exclusion of id2a from the ectomesenchyme precedes exclusion of sox10 and
induction of dlx2a

Figure 2.11: Exclusion of id2a precedes exclusion of sox10 from the ectomesenchyme:
Confocal projections of fluorescent in situs of sox10/ id2a at 15hpf shows that whereas id2a
expression is restricted to the non-ectomesenchyme crest and is expressed in the surface
ectoderm ( indicated by the white arrow) , sox10 is expressed both in the ectomesenchyme
precursors and the non-ectomesenchyme.

Misexpression of Id2a results in similar defects in specification of ectomesenchyme as
misexpression of Bmp4
To test whether exclusion of id2a is necessary for specification of the ectomesenchyme
lineage, we decided to misexpress id2a throughout the migratory neural crest using the
UAS: Gal4 system. Indeed, when id2a was mis-expressed throughout migratory CNCCs in
sox10:Gal4VP16; UAS: Id2a embryos, we observed persistent expression of sox10 and
reduced expression of fli1a in arch CNCCs, as well as mild reductions of dlx2a expression
(Figure 2.12 A-F). The effect of CNCC misexpression of id2a was specific to the
ectomesenchyme, as we observed severe reductions of the craniofacial skeleton but no
defects in the non-ectomesenchyme-derived melanophore and xanthophore precursors
53

Figure2.12: Forced expression of Id2a in CNCCs inhibits ectomesenchyme specification. (A-F)
Whole mount in situs show ectopic arch expression of sox10 at 18 hpf in sox10:Gal4VP16;
UAS:Id2a embryos (n=7/7) but not sox10:Gal4VP16 only controls (n=0/9), loss of fli1a arch
expression at 24 hpf in sox10:Gal4VP16; UAS:Id2a embryos (n=4/4) but not sox10:Gal4VP16 only
controls (n=0/17), and mild reductions of dlx2a arch expression at 18 hpf in sox10:Gal4VP16;
UAS:Id2a embryos (n=5/5) but not sox10:Gal4VP16 only controls (n=0/8). Arches are numbered.
Arrows denote the second arch, white arrowheads the developing ear and red arrowheads the
vasculature. (G and H) Skeletal staining at 5 dpf shows severe reduction of the craniofacial
skeleton in sox10:Gal4VP16; UAS: Id2a embryos (n=17/17) compared to sox10:Gal4VP16 only
controls (n=0/22). J) and (I In situs for dct expression at 28 hpf show that
melanophoreprecursors are unaffected in sox10:Gal4VP16; UAS: Id2a embryos (n=8) and
sox10:Gal4VP16 only controls (n=8). (K and L) In situs for xdh expression at 28 hpf show that
xanthophore precursors are unaffected in sox10:Gal4VP16; UAS: Id2a (n=4) and sox10:Gal4VP16
only controls (n=4). Scale bars = 50 m.

54

(Figure 2.12 G-L). However, since sox10 promoter also drives later expression in
chondrocytes, we cannot rule out that the skeletal defects of sox10:Gal4VP16; UAS: Id2a
animals are additionally or alternatively due to this later phase of Id2a misexpression. In
any event, the ectomesenchyme specificity of gene expression and differentiation
defects suggests that Id2a does not generally inhibit CNCC differentiation but instead
specifically restricts ectomesenchyme fates.
Id2a is required for Bmps to inhibit ectomesenchyme formation
As Id2a and Bmp4 misexpression resulted in similar ectomesenchyme defects, we next
examined whether Id2a might mediate the negative effects of Bmps on
ectomesenchyme development. To do so, we injected a previously validated translation-
blocking id2a-MO (Uribe and Gross, 2010 ) into one-cell-stage embryos to deplete Id2a
protein embryos and then examined development of the ectomesenchyme. Compared
to controls, whereas injection of id2a-MO caused no ectomesenchyme defects on its
own (2.15C, G-K), injection into sox10:Gal4VP16; UAS: Bmp4 embryos fully rescued
persistent sox10 and reduced dlx2a arch expression, as well as partially rescuing
expression of fli1a. (Uribe and Gross, 2010). We therefore conclude that Id2a functions
downstream of Bmp signaling to regulate its functions in specification of the
ectomesenchyme.

55

Figure 2.13: Knockdown of id2a in the background of Bmp4 misexpression rescues gene
expression defects in ectomesenchyme specification: (A-L) Colorimetric in situs show ectopic
sox10 arch expression at 18 hpf and reductions of dlx2a and fli1a arch expression at 24 hpf in
sox10:Gal4VP16; UAS: Bmp4 embryos but not sox10:Gal4VP16 control or id2a-MO-injected
embryos. sox10:Gal4VP16; UAS: Bmp4 embryos injected with id2a-MO showed complete rescue
of sox10 and dlx2a expression and partial rescue of fli1a expression. Arrows denote the second
arch, white arrowheads the developing ear and red arrowheads the vasculature. (M-O)
Quantification of gene expression defects: The mutant index is based on the following: 0 =
normal, 1 = partially defective, 2 = fully defective. The rescue of Bmp4 misexpression defects
with id2a-MO injection was statistically significant for all genes based on a Tukey-Kramer HSD
test ( =0.05). Standard errors of the mean are shown. Scale bar = 50 m.

Twist1 transcription factor is a candidate for mediating Bmp function
Although we found that Id2a could mediate the function of Bmp signaling in the neural
crest to regulate expression of the markers of the ectomesenchyme, Id2a is not a
transcription factor, but is involved in modulating the activity of transcription factors of
56

the bHLH family (Connerney et al., 2006). Similar to what we showed in prolonged
expression of Bmp4 as well as of Id2a in the migratory neural crest, previous studies in
mice have shown that loss of Twist1 results in similar defects in prolonged expression of
sox10 in the ectomesenchyme of the pharyngeal arches (Soo et al., 2002). In order to
test whether Twist1 genes are required for ectomesenchyme development in zebrafish,
we designed translation-blocking MOs to deplete both zebrafish Twist1 orthologs -
twist1a and twist1b. Although there are four twist genes in zebrafish (Germanguz et al.,
2007), since twist1a and twist1b are the two twist genes expressed in the head region
during early neural crest development, we decided to examine the functions of twist1a
and twist1b in formation of the ectomesenchyme lineages. To study the functions, we
designed translation blocking morpholinos to inject into single cell stage Although
injections of either twist1a-MO or twist1b-MO alone resulted in only very subtle
changes in sox10, embryos injected with both MOs (twist1a/1b-MO) displayed very
severe abnormal persistence of sox10 at 18hpf (Figure 2.14) , prior to contribution to
the pharyngeal arches by these cells. Since misexpression of Id2a and loss of Twist1
resulted in similar defects in specification of the ectomesenchyme lineage, this data
would suggest that the Twist1 might be an important transcription factor in regulating
specification of the ectomesenchyme.

57

twist1a and twist1b have redundant and synergistic roles in specification of the
ectomesenchyme lineage
Since injection of either twist1a and twist1b results in defects in sox10 expression, we
decided to trace the development of the ectomesenchyme, by injecting morpholinos
into fli1a: GFP, sox10: dsRED transgenics,(thus allowing us to follow the
ectomesenchyme precursors as well as the development of the neural crest )

Figure 2.14: twist1a and twist1b function redundantly to specify ectomesenchyme. (A-C) In
situs at 18 hpf show sox10 expression in un-injected, twist1a-MO, and twist1b-MO embryos. A
few ectopic sox10-positive cells were seen in the second arches (arrows) of twist1a-MO and
twist1b-MO embryos. White arrowheads denote the developing ear. (D-G) Confocal projections
of fli1a: GFP; sox10: dsRed doubly transgenic embryos at 28 hpf show normal fli1a: GFP
expression in un-injected control, twist1a-MO, and twist1b-MO embryos and loss of fli1a: GFP
arch expression in twist1a/b-MO embryos. Arrows indicate fli1a: GFP vascular expression that is
unaffected in twist1a/b-MO embryos. (H-L) Skeletal staining shows malformed mandibular and
hyoid skeletons in twist1a-MO and twist1b-MO embryos compared to un-injected controls. In
addition, co-injection of a Twist1b mRNA not targeted by the MOs partially rescued the head
skeleton of twist1a/b-MO embryos (n=24/24), whereas co-injection of a control kikGR mRNA
never rescued (n=0/11). Scale bars = 50 m.

58

We find that injection of Twist1a-MO alone caused no changes in fli1a: GFP transgene
expression, but injection of Twist1b-MO caused reductions in the size of the pharyngeal
arches, without changing the expression of ectomesenchyme marker fli1a. Analysis of
doubly transgenic sox10: dsRed; fli1a: GFP embryos injected with twist1a/b-MO
confirmed that CNCCs still formed arches, although reduced in the absence of Twist1
function yet failed to initiate fli1a: GFP expression (Figure 2.14). These results indicate
that twist1a and twist1b function synergistically in the pathway of specification of the
ectomesenchyme. Also, loss of twist1a and twist1b resulted in mild defects in the
ventral mandibular and hyoid cartilages, these defects are distinct, suggesting that
twist1a and twist1b might have independent functions in development of the
craniofacial skeleton.
Loss of Twist1 results in defects in specification of ectomesenchyme:
We further characterized the function of Twist1 in specification of the ectomesenchyme
lineage by examining expression of fli1a, and dlx2a by in situ hybridizations. We find
that compared to controls, in embryos injected with twist1a/1b- MO (Figure 2.15 E, I),
there was a complete loss of fli1a in the ectomesenchyme of the pharyngeal arches (F),
although the blood vessel expression of fli1a was un-affected (indicated by the red
arrow), and no change in the expression of dlx2a. We further analyzed the fate of the
ectomesenchyme precursor neural crest cells by performing alcian blue staining. We
found that compared to controls, embryos with loss of twist1 had near complete loss of
59

the neural crest derived head skeleton (2.15M, N). Importantly, injection of a twist1b
mRNA not targeted by the MOs, but not a control kikGR mRNA, partially rescued
skeletal defects in 24/24 twist1a/1b-MO embryos, showing specificity of MO-generated
defects for Twist1 (Figure 2. 14 I, K, L). To determine the cellular bases for the skeletal
phenotype in twist1a/1b –MO, we decided to perform Lysotracker staining for apoptosis
in sox10: GFP transgenic embryos. We found that at 36hpf, there was a statistically
significant increase in the number of apoptotic cells in the pharyngeal arch neural crest,
as well as in the non-ectomesenchyme (Figure 2.16). Therefore, we conclude that
ectomesenchyme precursor crest cells in the absence of twist1 might undergo cell
death.
Furthermore, upon examination of the differentiation of the other lineages of the
cranial neural crest such as melanophores, xanthophores and the glia of the cranial
ganglia by in situ hybridization for dct, xdh and foxd3 respectively, we found that there
was no change in expression of these markers in twist1a/1b-MO compared to controls,
although expression of foxd3 was slightly increased in cranial ganglia in twist1a/1b-MO
injected embryos. This might be attributed to the fact that the in-situ probe might have
preferentially penetrated the twist1a/1b-MO injected embryos due to increased cell
death that that we observed upon Lysotracker staining in the twist1a/1b -MO injected
embryos at 36hpf compared to controls (Figure 2.16).

60

Figure 2.15: Twist1 genes are required for ectomesenchyme specification in zebrafish.
(A-D) Whole mount in situ hybridizations of sox10 expression at 18 hpf show ectopic expression
in the arches (numbered and second arch indicated by black arrow) of twist1a/b-MO (n=16/16)
and sox10:Gal4VP16; UAS: dnTwist1b (n=4/4) embryos compared to un-injected (n=0/14) and
sox10:Gal4VP16 only (n=0/9) controls. White arrowheads indicate otic expression. (E-H) Whole
mount in situs at 24 hpf show reduction of fli1a expression in the arch ectomesenchyme
(numbered) of twist1a/b-MO (n=12/12) and sox10:Gal4VP16; UAS: dnTwist1b (n=5/5) compared
to un-injected (n=0/13) and sox10:Gal4VP16 only (n=0/8) controls. Vascular expression of fli1a
(red arrowheads) is unaffected. (I-L) Whole mount in situs at 18 hpf show a moderate reduction
of dlx2a in sox10:Gal4VP16; UAS: dnTwist1b embryos (n=4/4) but not un-injected (n=0/6),
twist1a/b-MO (n=0/8), and sox10:Gal4VP16 only (n=0/6) embryos. (M-P) Skeletal staining at 5
dpf shows severe loss of CNCC-derived head skeleton in twist1a/1b-MO embryos (n=21/21) and
primarily jaw reductions in sox10:Gal4VP16; UAS: dnTwist1b embryos (n=9/9) compared to no
defects in un-injected (n=0/24) and sox10:Gal4VP16 only (n=0/16) controls. (Q-T) In situs for dct
expression at 28 hpf show normal melanophore precursors in un-injected (n=14), twist1a/b-MO
(n=12), sox10:Gal4VP16 only (n=3), and sox10:Gal4VP16; UAS: dnTwist1b (n=3) embryos. (U-X)
In situs for xdh expression at 28 hpf show normal xanthophore precursors in un-injected (n=7),
twist1a/b-MO (n=5), sox10:Gal4VP16 only (n=3), and sox10:Gal4VP16; UAS: dnTwist1b (n=3)
embryos.

61

Figure 2.16: Cell death in twist1a/1b MO injected embryos (A-D) Confocal projections of
Lysotracker Red staining in 36 hpf sox10:GFP transgenic embryos show increased cell death in
the pharyngeal arches (shown in high-magnification views in C and D from the boxes in A and B)
in twist1a/b-MO-injected embryos (n=6) compared to un-injected controls (n=6). Arrowhead
shows increased cell death in non-ectomesenchyme CNCCs as well. Scale bar = 50 m. (E)
Quantification of Lysotracker-positive cells per arch area. Mandibular and hyoid arches were
used for the analysis. Asterisk indicates statistical significance using a Tukey-Kramer HSD test
( =0.05).

Possible mechanisms of id2a mediated twist1 function:
Based on our genetic analyses, I suggest that Id2a and Twist1 function in the same
pathway to specify the ectomesenchyme lineage ectomesenchyme. To understand
completely how Id2a and Twist1 function together to regulate specification of the
ectomesenchyme lineage, immune-precipitation experiments with Id2a antibody would
be necessary to identify potential interacting partners of Twist1. Furthermore,
chromatin immune-precipitation experiments would be necessary to determine
whether Twist1 binds to its downstream target genes at the promoter and regulates
their expression and how Id2a might be modulating its function.
62

Another possible mechanism for precise transcriptional control of bHLH proteins like
Twist1 is via phosphorylation at the serine and threonine residues in the basic helix1. To
test whether phosphorylation of Twist1 was important for regulation of Twist1 function
in the ectomesenchyme specification process, we decided to generate transgenic lines
that would result in loss of phosphorylation as well as mimic the phosphorylation.
Although, the transgenic constructs were sequence verified for having the desired
mutations, upon introduction of these constructs into the migratory neural crest cells,
we never observed any changes in the specification of the ectomesenchyme lineage.
One possible explanation is that, the levels of expression of the transgenes were not
sufficient to tip the balance of the protein. Although not conclusive, based on these
experiments, I would conclude that changes in phosphorylation might not be important
for specification of the ectomesenchyme lineage.
Twist1 function is required in migratory CNCCs for ectomesenchyme specification
Since twist1 genes are expressed both in the cranial neural crest as well as in the
mesoderm, we decided to examine whether the defects observed in specification of
ectomesenchyme were crest autonomous or due to non-autonomous functions of
twist1. Furthermore, published expression patterns of twist1 shows that twist1 genes
are expressed early in the pre-migratory crest and in the migratory CNCCs later in
development, suggesting that Twist1 might have multiple roles in craniofacial skeleton.
To dissect more precisely when Twist1 is required for specification of the
63

ectomesenchyme lineage, we employed a transgenic strategy to misexpress a dominant-
negative Twist1b protein specifically in migratory CNCCs. Caenorhabditis elegans has a
single Twist gene, hlh8, which is required for mesodermal patterning and vulval muscle
development (Corsi et al., 2002), and mutation of Glutamine 29 to Lysine in the DNA-
binding domain has been shown to dominantly interfere with its function (Corsi et al.,
2002). As this residue is conserved in zebrafish Twist1 genes (Glutamine 84 in Twist1b),
we reasoned that the equivalent mutation might also result in a dominant-negative
protein. We therefore constructed a transgenic line in which Twist1b-E84K (referred to as
DN-Twist1b) was expressed under the Gal4VP16-sensitive UAS promoter. We then used
the sox10:Gal4VP16 transgenic to specifically block Twist1 function in the migratory
neural crest cells. Upon doing so, we observed persistent arch expression of sox10 and
severe reductions of fli1a but only minor reductions of dlx2a expression in
sox10:Gal4VP16; UAS:DN-Twist1b embryos, whereas sox10:Gal4VP16 controls were
unaffected (Fig. 2.15) Moreover, inhibition of Twist1 function in migratory CNCCs
resulted in partial losses of facial skeleton, in particular the jaw skeleton (Fig. 2.15 M-P).
Therefore, similar to studies in mice (Bildsoe et al., 2009; Soo et al., 2002), Twist1 genes
function in migratory CNCCs to regulate specification of the ectomesenchyme lineage.
Twist1 functions non-autonomously to cause migration defects in the CNCCs
Although the gene expression defects observed in the ectomesenchyme precursor
population were similar in several respects between the neural crest specific loss of
64

function of Twist1 and the loss of function mediated by morpholino knockdown, we
observed some interesting differences in another population of the cranial neural crest.

Figure 2.17: Crest autonomous and non-autonomous functions of Twist1: Confocal projections
of fluorescent in situ hybridizations at 24hpf show that in twist1a/1b-MO injected embryos,
there is ectopic expression of sox10 and foxd3 ( A’, B’) above the eye (indicated by the white
arrow ) in comparison to the sox10:Gal4VP16; UAS:dntwist1b transgenic embryos ( A’’, B’’) ,
which have ectopic expression of sox10 and foxd3 in the pharyngeal arches similar to embryos
with twist1a/1b –knockdown by morpholino injection .

When we essayed for changes in changes in expression of general neural crest / later
non-ectomesenchyme marker sox10, we noticed in embryos injected with twist1a/1b
MO but not in sox10:gal4VP16; UAS:DNtwist1b embryos, there was ectopic expression
of sox10 and foxd3 in neural crest around the eye. Although we did not follow-up on this
observation with precise lineage tracing to determine the origin of the cells, our data
suggests that twist1 function non-autonomously to regulate migration of the neural
crest. This is similar to a previously described role of Twist1 from the mesoderm in
regulation of migratory of the arch neural crest in mice (Bildsoe et al., 2009), although
65

the defects there related to difference in localization of the ectomesenchyme neural
crest to specific domains in the branchial arches.

Fgf signaling regulates specification part of ectomesenchyme specification:
Although previous studies have established a role for Fgf signaling in downregulation of
sox10 as well in induction of dlx2a (Yamauchi et al.), there have been no reports on how
it functions to regulate the expression of ectomesenchyme gene fli1a. Furthermore, how
Fgf signaling is integrated with the Bmp/Id2a/Twist1 pathway during ectomesenchyme
development. In addition, inhibition of Fgf signaling specifically in migratory CNCCs, by
transgenic misexpression of a dominant-negative version of Fgfr1 (DN-Fgfr1) in
sox10:Gal4VP16; UAS: DN-Fgfr1 embryos resulted in abnormal persistence of sox10
and reduction of dlx2a arch expression (Figure 2.18). However, in contrast to Twist1-
deficient and Bmp4-misexpression embryos, fli1a ectomesenchyme expression of fli1a
was largely unaffected by Fgf inhibition (Fig. 2.18E, F). Thus, Fgf signaling is required
downstream of Twist1 to repress sox10 and activate dlx2a expression, yet plays little or
no role in activation of fli1a.
66

Figure 2.18: Fgf signaling regulates a subset of ectomesenchyme development. E-J) Embryos
misexpressing sox10:Gal4VP16; UAS: dnFgfR1a show ectopic expression of sox10 at 18hpf (n= 8/
30), reduction of dlx2a (n= 3/12) and no change in fli1a (n=7/25) compared to controls at 24hpf.
Scale bar= 50um

Studies to define the temporal basis of specification of the ectomesenchyme:

In order to determine the precise timing when neural crest cells no longer form
ectomesenchyme , we decided to follow neural crest cells after birth-dating them using
KikGR photoconversion and determine their contribution to the neural crest derived
lineages. We photo-converted neural crest cells in embryos as early as 4 somites
(11hpf) stage embryos and 6 somites (12hpf), and 10 somites (14hpf) stages of embryos
subsequently. We find that whereas the earliest forming CNCCs preferentially
67

contributed to the ectomesenchyme of the pharyngeal arches, cells photo-converted at
10 somites stage, there is preferential contribution to the non-ectomesenchyme. This
data shows that neural crest cells photo-converted early preferentially contribute to the
ectomesenchyme, and at 10 somites stage, there is a fate switch in the contribution of
the CNCC from the ectomesenchyme to the non-ectomesenchyme.

Figure 2.19: Birthdating of ectomesenchyme versus non-ectomesenchyme by sox10: KikGR
photoconversion at 4 somite stage embryos: Confocal projections of sox10: KikGR embryos
shows that compared to controls (A-A’), in embryos with all KikGR positive cells photoconverted
at 4 somite stage, although faint, cells contributed to the ectomesenchyme at 17hpf (C, C’,C’’)
and also to the non-ectomesenchyme lineage.

68

Figure 2.20: Photo-conversion of neural crest at different stages during development shows
difference in contribution CNCC to ectomesenchyme versus non-ectomesenchyme: (A-A’’.
Confocal projections of sox10: kikGR embryos show that embryos photo-converted at 4 somite
stage neural crest cells contribute preferentially to pharyngeal arches shown in red. The faint
red in the non-ectomesenchyme domain indicates photo-conversion of the non-
ectomesenchyme in the process of imaging. (B) shows that neural crest cells at 6 somite stage in
development still contributes to the ectomesenchyme of the pharyngeal arches, but (C) shows a
preferential contribution of the neural crest to the non-ectomesenchyme domain as indicated
by brighter red in the non-ectomesenchyme domain in embryos photoconverted at 10 somite
stage compared to 4 or 6 somite stage.

69

Discussion:

Figure 2.21: Model for specification of the ectomesenchyme lineage from CNCC: Early in
development, CNCCs shown in red have High Bmp activity which leads to activation of Id2a and
block of Twist1-E protein complex required to promote ectomesenchyme and block non-
ectomesenchyme. At 16.5hpf in development, there are two distinct populations of CNCC with
differences in Bmp activity. Whereas neural crest cells close to the neural tube shown in red,
maintain High Bmp activity and no Twist1 activity, ectomesenchyme precursor cells shown in
green migrate out of the Bmp source. These cells have low Bmp activity, thereby loss of id2a and
activation of Twist1 function, resulting in activation of ectomesenchyme specific genes and loss
of non-ectomesenchyme markers. Together, these genetic interactions promote specification of
the ectomesenchyme lineage.

70

Bmp signaling from the ectoderm: Possible functions of the surrounding tissues in
specification of the ectomesenchyme lineage of CNCC
Here we present evidence for a new model of ectomesenchyme formation. Upon
delamination from the neural plate border, early migrating CNCCs would down-regulate
Bmp activity and id2a expression, thus allowing activation of Twist1. The down-
regulation of Bmp activity in ectomesenchyme precursors could reflect CNCCs distancing
themselves from a Bmp source in the non-neural ectoderm (e.g. Bmp2b) and/or active
inhibition of Bmp signaling during migration. Whereas the Bmp antagonist Noggin1 is
expressed in the paraxial mesoderm surrounding migratory CNCCs in zebrafish
(Furthauer et al., 1999), genetic ablation of endoderm and mesoderm did not disrupt
ectomesenchyme gene expression, although there are likely other sources of Bmp
inhibitors (Zuniga et al., 2011). Alternatively, the down-regulation of Bmp activity could
result from the delamination of CNCCs from the neural plate ectoderm, with the
basement membrane of the ectodermal epithelium, under which CNCCs migrate,
serving as a barrier to apical Bmp secretion and signaling (Eom et al., 2011).
Since loss of Bmp was not sufficient to ectopically induce the ectomesenchyme in the
non-ectomesenchyme population, one possibility is that the cells that are born later
have a slightly different potential. To understand these further, additional experiments
would be necessary to see whether the cells in an older embryo have a slightly different
potential. Infact, our studies on birth dating of CNCCs to then follow their lineage
71

contributions, show that over time the CNCCs have different fates. This evidence
supports a temporal model for specification of the ectomesenhcyme lineage. We also
find that high Bmp activity is maintained in the neural tube close to where the non-
ectomesenchyme would form. Upon loss of bmp signaling, we observed that there was
reduced migration of the neural crest, which is consistent with what is known in the
literature, but there was induction of the non-ectomesenchyme markers, but increased
death was observed. Therefore, Bmp activity might not be important for induction of
the non-ectomesenchyme, but maybe for maintaining the non-ectomesenchyme
precursor fate.
Id2a restricts Twist1 activity to the ectomesenchyme
Similar to the differential localization of pSMAD that we observed between the
ectomesenchyme and the non-ectomesenchyme precursors, upon careful
characterization of the expression patterns of Bmp target id2a, we observed that
differential expression of id2a in the ectomesenchyme versus non-ectomesenchyme.
Using transgenic tools, we find that id2a is critical in defining the temporal basis for
specification of the ectomesenchyme lineage. Id genes typically modulate
transcriptional complexes of bHLH group of transcription factors. Interestingly, the
perdurance of sox10 indicative of a specification defect, was reported in earlier studies
in loss of Twist1 mice. Upon careful characterization of the ectomesenchyme defects,
we find that all of the gene expression changes indicative of defects in transition of the
72

CNCC into ectomesenchyme were shared between the prolonged expression of id2a and
loss of Twist1. Previous work suggests that Id2a does not inhibit Twist1 function per se,
but instead would modulate what types of dimers Twist1 forms (Connerney et al., 2006;
Yokota, 2001). Since the expression of Twist1 genes is throughout early CNCCs in both
zebrafish (Germanguz et al., 2007) and mice (Gitelman, 1997; Hopwood et al., 1989),
but the activity of Twist1 in ectomesenchyme is not until later, therefore a post-
transcriptional or post-translational mechanism might be at play either through
enzymatic modifications or changes in dimerization partners. Our studies provide
genetic evidence for the latter, with the Bmp target gene Id2a functioning oppositely to
Twist1 for ectomesenchyme specification. In the future, identification of binding
partners for Twist1 in CNCCs will allow us to test what roles Twist1 heterodimers versus
homodimers play in ectomesenchyme specification. Furthermore, we note that Twist1
genes continue to be expressed in arch ectomesenchyme and likely have additional
roles in facial skeletal patterning, for example by interacting with other bHLH factors
such as Hand2 (Firulli et al., 2005). Hence, different Twist1 heterodimers may have
distinct roles during craniofacial development (i.e. ectomesenchyme formation versus
arch patterning), with the later disruption of these heterodimers in Twist1-deficient and
Id2a-misexpression embryos also contributing to some aspects of the observed facial
skeletal defects.
73

We also observed that there was high id2a expression in the non-ectomesenchyme
precursors, but the function of id2a in the non-ectomesenchyme remains un-explored.
Future experiments would be necessary to determine the precise mechanisms of id2a
function both in the ectomesenchyme precursors as well as in the non-
ectomesenchyme.
Implications of sub-functionalization of Twist1
Whereas knockdown of either twist1a or twist1b results in defects in sox10 gene
expression, combined expression of twist1a and twist1b is required to see gene
expression defects in fli1a. This would suggest that distinct transcriptional complexes
might be regulating transcriptional outcomes of sox10 and fli1a. Given the fact that
these functions of twist1a and twist1b in the zebrafish is attributed to sub-
functionalization, we suggest that the zebrafish is useful in understanding molecular
pathways that would be harder to study in other organisms. In the next chapter, I will
discuss further our studies on understanding how the different transcription factors
downstream of Twist1 function together or independently to regulate lineage outcomes
of the CNCC into ectomesenchyme.
Migration defects can cause defects in specification
In the Twist1-MO we observed an interesting ectopic expression of non-
ectomesenchyme genes sox10 and foxd3 over the eye. This could be explained by the
74

fact that there might be an aberrant migration of the ectomesenchyme precursors in
that area, which would then result in them being fated to form non- ectomesenchyme.
This is interesting in that previous studies in the trunk have discussed roles of migration
in actively determining the lineage choices of neural crest. Although, several molecules
have been studied in the head that define the migratory routes of the cranial neural
crest, a role of migration in determining the fate choices of the cranial neural crest has
never been discussed. Therefore, although preliminary, upon further characterization, I
believe that one might find active roles of differential migration of various sub-
populations of the cranial neural crest. In the following chapter, I also discuss
preliminary evidence from characterization of Twist1 downstream targets, how we
identified candidate molecules that might play roles in regulating the collective cell
behavior of the CNCC and in turn regulate specification of the ectomesenchyme.

75

CHAPTER 3: DEFINING THE MOLECULAR BASES FOR TWIST1
FUNCTION IN SPECIFICATION OF THE ECTOMESENCHYME
LINEAGE DURING CRANIAL NEURAL CREST DEVELOPMENT
Introduction:
We show previously that Twist1 is essential for specification of the ectomesenchyme by
positively regulating expression of the known markers of the ectomesenchyme lineage
like dlx2a and fli1a, as well as negatively regulating expression of early neural crest
marker sox10. Although sox10 labels the early neural crest, later in development, it
labels the non-ectomesenchyme precursors and has been shown to be important for
specification of the non-ectomesenchyme (Kelsh, 2006). Therefore, ectopic
maintenance of sox10 in the ectomesenchyme of the pharyngeal arches might indicate
that these cells are now more non-ectomesenchyme like. To examine the possibilities
further, we decided to take two approaches: One where we examined expression of all
known markers of the early neural crest at 24hpf, after the CNCC have contributed to
the pharyngeal arches to assay for ectopic expression of a majority of the markers in the
CNCC that should have contributed to the pharyngeal arches. In addition, we decided to
take a global approach to determine the transcriptional profile of the neural crest cells
upon knockdown of twist1a/1b function. By utilizing this approach, we would be able to
76

determine in an unbiased manner the altered transcriptional program allowing us to
determine whether the cells had been programmed into an alternate lineage.
Previous studies have shown that sox10 is important for regulating the expression of
non-ectomesenchyme(Kelsh, 2006). In addition, studies have shown that dlx2a is not
essential for specification of the ectomesenchyme (Sperber et al., 2008), but function of
fli1a in this process has not been examined, possibly due to the embryonic lethal nature
of the Fli1 mutants in mice (Spyropoulos et al., 2000). Since fli1a is one of the earliest
markers of the ectomesenchyme lineage, we decided to first investigate how Twist1
regulates fli1a expression in the ectomesenchyme and then we performed some loss of
function studies to investigate the functions of fli1a in the ectomesenchyme.
In addition to investigate the functions of known downstream targets of Twist1, we also
decided to utilize the microarray approach to identify novel players downstream of
Twist1 involved in specification of the ectomesenchyme and investigate how these
players would contribute to Twist1 mediated specification of the ectomesenchyme.
Taken together, these studies will provide us with a more picture of the cellular and
molecular bases for specification of the ectomesenchyme lineage of CNCC.
Materials and methods:
Transgenic lines: Tg (sox10:LOX-GFP-LOX-hDLX3)
el8
was generated by combining p5E-
sox10 pME-LOX-GFP-LOX-Dlx3. In the absence of CRE, Tg(sox10:LOX-GFP-LOX-hDLX3)
el8

77

(referred to as sox10:GFP in the Results) drives similar GFP expression as the previously
reported Tg(~4725sox10:GFP)
ba2
line but without neural crest toxicity (Carney et al.,
2006) The UAS: hsox10 transgenic line was constructed combining p5E-UAS, pME-
hsox10, p3E-polyA, and pDestTolAB2.. pME -hsox10 was generated by a Invitrogen BP
reaction to recombine hsox10 cDNA containing pOTB7 vector ( Open Biosystems #
MHS1011-62084 ) with pDonr 221. UAS: fli1a was generated by combining p5E-UAS,
pME-hsox10, p3E-polyA, and pDestTolAB2. Fli1a cDNA was generated by PCR from
zebrafish using primers attFli1aF: 5’-
GGGGACAAGTTTGTACAAAAAAGCAGGCTCGGCCACATGGACGGAACTATTAAG-3’ and
attFli1ar- 5’- GGGGACCACTTTGTACAAGAAAGCTGGGTGTTAGTAACTACCAAGG-3’. The
PCR product was then combined with pDonr 221 vector to generate pME- Fli1a. For the
UAS: hsox10 transgenic line, genotyping of embryos for UAS:hsox10 transgene was
performed using primers : hsox10genF: 5’ ACAAAGTTCCCCCGTGTGCAT-3’ and
hsox10genR: 5’- TCAGCAGCCTCCAGAGCTT-3’, and genotyping of gal4 was performed
using G4L and G4R primers described previously. For the UAS:fli1a transgenic line,
embryos were sorted for eye CFP , and genotyping was performed using the Gal
4primers.
Morpholino and mRNA injections: Injection of twist1a/1b-MO was performed as
described previously. Injection of fli1aMO (5' - TTTCCGCAATTTTCAGTGGAGCCCG - 3')
was injected at 400um concentration into single cell stage embryos. mRNA for fli1a was
78

transcribed from CMV/SP6-fli1a-polyA using the method described previously.
CMV/SP6-fli1a-polyA was generated by combining p5E-CM/SP6, pME- fli1a and p3E-
polyA and pDestTol2AB2. mRNA injections of fli1amRNA was performed at 200ng/ul
and 500ng/ul respectively.
Sample preparation for microarray analysis:
Embryo preparation for FACS: After injection of the twist1a/1b morpholinos as
described previously, the embryos were allowed to develop at 28.5 C until 4hpf. After
that the embryos were kept at room temperatureerature until they were at 10hfp. At
10hpf, the embryos were again placed at 28.5C to allow them to develop to 17hpf for
FACS next morning. By following this protocol, embryos were ready to sort at 9am.
FACS sorting was done according to the protocol from Ilya from James Chen lab at
Stanford University. This protocol is a modification from the FACS protocol the zebrafish
book.
RNA extraction using Trizol from FACS sorted cells:
Cells were sorted into 500ul of trizol reagent from Invitrogen. Tubes of sorted cells
should be placed on ice until ready to process further.
Cells can be frozen at -80c for months or cells could be used at that time for RNA
extraction
79

If using frozen cells, thaw cells and then pipette cells to lyse cells
Incubate homogenized samples for 5 minutes at room temperatureerature
Add 200ul of chloroform, tap tubes and shake vigorously fo 15 seconds
Incubate at room temperatureerature for 2-3 minutes
Centrifuge samples at no more than 12000g for 15 minutes at 2-8 C
Transfer the aqueous phase to a new tube
Add 500ul of Iso-propyl alcohol per mls of trizol.
Incubate samples at room temperatureerature for 10 minutes
Centrifuge samples at 12000g at 2-8 c for 10 minutes
Remove supernatant
Wash RNA pellet once with 75 % ethanol
Mix the sample by vortexing and centrifuge at no more than 7500g for 5 minutes at 2-8C
After RNA extraction, measure concentration using Nanodrop and use 250ng to make
cDNA library

80

Use WTA2 kit from Sigma to generate the cDNA library
The cDNA library was purified using a modified version of the recommended protocol on
cDNA synthesis in the Roche Nimblegen protocol for sample preparation for Microarray
analysis.
Purify the cDNA samples using precipitation with Ammonium acetate and other stuff:
Update this part of the protocol
Microarray was performed using the Nimblegen 385K arrays for gene expression
analysis,
Analysis of microarray data was performed using the Array star software
Steps for microarray data analysis:
We used Arraystar 4.0 (dnastar.com) and we also utilized help from Dr. Yibu Chen in the
Bioinformatics Department at the Norris Medical Library to analyze the data using
Partek Genomic Suite. We then combined both the analyses to generate the final list of
candidates to study. I will summarize the steps for analysis of the data using Arraystar
4.0. In addition to generating table of genes by applying various filtering criteria, I also
utilized Arraystar to do hierarchical clustering of the data as well as gene ontology
analysis. Summarized below are the steps for Analysis of data using Arraystar 4.0.
1. Import RMA files into new project on Arraystar.
81

2. Once the experiment list is generated, then use the Create Replicate set function
to group replicate experiments. This feature allows to average the values of
multiple experiments as well allow doing analysis of individual experiments one
at a time.
3. Select the folders to compare, in my case I compared to GFP+ vs GFP- in control
and GFP +vs GFP- in treated and then generated the scatter plot view to plot all
the data points. Subsequently, I performed a student’s t-test for statistical
significance.
4. After this, I used the filter all feature to select genes that were upregulated by 2
fold in GFP+ vs GFP negative and had 1.5 fold changes in expression in GFP + vs
GFP negative treated population.
5. I then went to graphs, and plotted the Venn diagram and gene ontology to
obtain the gene ontology table.
6. Used Arraystar 4 to first generate gene lists, set 1.5 fold upregulated and
downregulated to generate lists, and used p less than 0.05 to generate a list of
genes that were enriched in the GFP positive neural crest cells compared to
other populations, and then did the same for the treated samples and then
Using excel generated a ratio of the fold changes to then get to the working list
of genes
82

I then used the gene lists generated using Partek Genomics Suite to generate a similar
list. Also, for more detailed analyses of the genes in the upregulated category, I also
generated sub-lists using Microsoft excel of genes that were expressed moderately in
GFP positive population in controls and enriched significantly in the GFP positive
population from the treated samples. I also generated a working list of genes that were
expressed at moderate levels in the GFP positive population and upregulated slightly in
the GFP positive population in the treated samples.
Validation of microarray data:
For validation of microarray data, we performed in-situ hybridization experiments in
control and twist1a/1b-MO injected embryos at 17hpf and for some at 24hpf.
Additionally, we also performed double fluorescent in situ hybridization experiments, to
identify genes that were co-expressed in specific populations of the neural crest.
In situ hybridizations: In situ hybridizations were performed as described previously. The
probes used were synthesized using the protocol described previously. For all probes,
PCR products generated from multi-stage zebrafish cDNA was used. The primer
sequences are nr4a2b: 5’-CCAGGCTCAGTATGGGACAT-3’ and 5’-
GCTAATACGACTCACTATAGGTATGTGACGTCGCCAGGTAG-3’,
gch2: 5’- GGAATACCAAAAGGCAGCAG-3’ and 5’-
GCTAATACGACTCACTATAGGCTCTCTGAACACACCCAGCA-3’.
83

nkd2aF: 5’- CACATCAGAGGCCAGACTGA-3’, nkd2aR: 5’-
GCTAATACGACTCACTATAGGTGCTGTCTGCTCTGCTCTGT-3’, loxl2aF: 5’-
ACCCATATCATCCCACCTCA-3’, loxl2aR: 5’-
GCTAATACGACTCACTATAGGAACAGCTGCGTCCTCTTCAT-3’, rgs3F 5’-
GCCTGCGAGGATTACAAGAA-3’, rgs3R: 5’-
GCTAATACGACTCACTATAGGCGGATTTAGGACGAGGATGA-3’, rab3ipF: 5’-
CAGGAGCTGGAGGAACTCAC-3’, rab3ipR: 5’-
GCTAATACGACTCACTATAGGTTGAACTCGGTGAACAGCAG-3’, cart1F: 5’-
CAAAACAGACCTCGACGAAC-3’, cart1R:
GCTAATACGACTCACTATAGGCTGTGTGTTCTTTGGCCTTC-3’, atf3F: 5’-
GCTTCAGCACCCTGGTTTG-3’, atf3R: 5’-
GCTAATACGACTCACTATAGGCTGGTTGGTATGGCGT-3’, hist1h2BF: 5’-
ATGCCCGAACCTGCGAAGTC-3’, hist1h2BR- 5’-
GCTAATACGACTCACTATAGGCCAGCTCCAAGTG-3’.
Functional analysis of the microarray data:
For functional analysis of the data, we used multiple approaches, including in silico and
experimental. For the in-silico approach, as a starting point, I used the gene-ontology
feature of Arraystar to obtain an idea of the kinds of processes that these genes could
be involved. Furthermore, I also performed a detailed literature search using the gene
ontology features from zfin and other resources like pubmed to determine the possible
84

roles of some of the candidate genes. In addition, I manually annotated the cell-types
that were predominantly represented in the gene lists by utilizing published expression
data available via zfin (zfin.org) and other publications via pubmed. For the functional
assays, I have designed morpholinos for some of the genes that were represented in the
microarray and will test them for function.
Transfections into 293T cells: 293T cells cultured in the lab of Dr. Wange Lu’s lab at the
Broad CIRM Center at USC were utilized in this experiment. Cells were cultured in a
10cm dish (in DMEM/10% FBS/ 1% Penicillin/Streptomycin), and were split and plated
after counting using a Hema-cytometer onto 24 well plates to 60% confluency. The
cells were then allowed to attach to the plate and tranfected 12 hours later. For the
transfection, PEI transfection reagent from Sigma (408727-100ml) was used. Each
reaction was performed in triplicates. The PEI: DNA ratio was optimized to be 1ug:0.5ul.
Transfection was performed following these steps:
1. The DNA was prepared in a mastermix and DMEM medium was added to make up for
the volume.
2. In a separate tube, PEI was mixed with medium without antibiotic and vortexed to
mix.
3. Mix 1 and 2 are then mixed and allowed to sit at room temperatureerature for 15
minutes.
85

4. The DNA PEI mixture should then be pipetted onto the cells
5. Swirl to disperse the liquid evenly without allowing the cells to detach from the
bottom
6. The cells are then incubated at 37 C CO2 incubator and harvested for luciferase assay
after 48 hrs
Luciferase assay: The Twist1 expression vector was made by cloning the zebrafish
twist1b cDNA into pCDNA3. Luciferase reporter constructs were generated by cloning
wild-type and mutant versions of the fli1a-F enhancer into the Pgl3 basic vector
(Progema #E1715). We then transfected 293T cells with the Twist1b expression and
luciferase reporter plasmids, and used the Dual Luciferase Assay kit (Promega #E1910)
per manufacturer instructions to measure firefly luciferase activity relative to renilla
luciferase activity 48 hours after transfection. All experiments were performed in
triplicates.

86

Results:
Loss of Twist1 does not result in neural crest cells to be maintained in an un-
differentiated state
Figure3.1: Fate of the ectomesenchyme crest in twist1a/1b-MO : In wild types, the expression
of sox10, foxd3, sox9b, and tfap2a were downregulated in arch ectomesenchyme by 18 hpf, and
foxd3, sox9b, and tfap2a were similarly extinguished in the arch ectomesenchyme of twist1a/b-
MOs embryos (Supplemental Figure 2). Thus, Twist1 depletion does not result in the persistence
of early multipotent CNCCs. In contrast, by 24 hpf, a time when foxd3 and sox10 are co-
expressed in early glial precursors of wild types, we observed ectopic expression of of sox10 and
foxd3 in the ectomesenchyme of Twist1a/b-depleted arches.

In order to determine whether loss of Twist1 results in CNCCs being maintained in an
undifferentiated state, or they acquire a non-ectomesenchyme identity, we decided to
examine expression of several markers of early neural crest like sox10, foxd3, sox9b and
tfap2a. Interestingly, we found that compared to controls at 24hpf, in embryos injected
with twist1a/1b-MO, there was ectopic expression of sox10 and foxd3 but not other
markers like sox9b and tfap2a (Figure 3.1). Since both sox10 and foxd3 label the non-
ectomesenchyme lineages, and we did not observe an up-regulation of all markers of
the early neural crest lineage, we conclude that ectomesenchyme precursor crest cells
in the absence of Twist1 are being fated to form the non-ectomesenchyme.
87

Twist1 functions in the CNCC to repress expression of the non-ectomesenchyme
specific genes
To determine the lineage state of the twist1 knockdown cells, we performed a
microarray analysis of FACS sorted cells from control and twist1a/1b –MO knockdown
sox10: GFP transgenic embryos (shown in figure 3.2). Subsequently, using Array-star
software, we calculated the fold enrichment of genes in GFP positive versus negative
fractions, and then calculated the ratio of fold enrichment in treated GFP (+) / GFP (-)
fraction versus controls. Also, I performed functional annotation of the 25 most up
regulated and 25 most down regulated genes. As indicated in figure 3.2, most of the
genes in the upregulated category represented expression in of the non-
ectomesenchymal lineages like neurons, glia and pigment. Furthermore, for the down-
regulated list of genes, most of the genes with known function annotated as being
expressed in the ectomesenchyme. To validate the results, we performed in-situ
hybridizations at 18hpf. In-situ hybridizations of the top genes like gch2 and nr4a2b
showed an increased expression of these genes in the non-ectomesenchyme domain,
with no ectopic expression into the ectomesenchyme. Based on these findings, we
conclude that twist1a/1b function globally both in the ectomesenchyme and the non-
ectomesenchyme precursor population to repress expression of the non-
ectomesenchyme specific genes.

88

Figure 3.2: Gene expression profiling in twist1a/1b-MO embryos. (A and B) sox10: GFP-
positive and -negative cells were isolated from un-injected or twist1a/b-MO embryos at 18 hpf
by FACS. (C) Fold changes of the GFP+/GFP- ratios between twist1a/b-MO and un-injected
controls show the top 25 up-regulated (blue) and down-regulated (red) genes after Twist1
depletion. Color codes indicate where genes are expressed based on the published literature).

89

Figure 3.3: Validation of novel downstream targets of Twist1: Confocal projections of
fluorescent in situ hybridization shows that compared to controls (A, C, E, and G), expression of
genes gch2 and nr4a2b were upregulated in the non-ectomesenchyme domain (B, D), grem2
was reduced in the second stream of CNCC (F), and cd248a was reduced in streams 1&2 (H).

Identification of novel ectomesenchyme specific genes downstream of Twist1:
In addition to defining more precisely the lineage state of the CNCC in the absence of
Twist1, we also obtained a list of genes that were up and down-regulated in twist1a/1b-
MO knockdown and had never been reported before to be reported in the CNCC or
regulated by Twist1 in any given species of vertebrates (Soo et al., 2002) (Figure 3.2).
Upon annotation of the down-regulated set of genes for cell types being represented in
the group, we found that a subset of those genes for example grem2, fli1a and pcdh18a
(Aamar and Dawid, 2008, (Liu et al., 2008; Muller et al., 2006), had been shown to be
expressed in the ectomesenchyme in previous studies. In addition, in situ hybridizations
of some other candidates like cd248a, and loxla2a, nkd2a revealed expression in the
ectomesenchyme crest at 18hpf and 24hpf respectively (Figure 3.3 and 3.4). Of the
ones that are expressed in the ectomesenchyme, we decided to validate the top 2 by
fluorescent in situ hybridization. We find that compared to controls (Figure 3.3 E,G) , in
90

embryos with twist1a/1b reduction of function, there was a slight reduction in
expression of grem2 in the second stream of migratory CNCC (Figure 3.3 F) and a
reduction in the expression of cd248a in the first and second streams of migratory
CNCCs (Figure 3.3 H). Also, genes like fli1a were represented on the list as one of the
most down-regulated genes. Based on the validation of the microarray data by in situ
hybridizations, we conclude that our microarray experiment has worked successfully.
Furthermore, to add to the previously described role of Twist1 in promoting expression
of some of the markers of the ectomesenchyme lineage, we suggest that Twist1
functions as a master-regulator to promote expression of the several of the
ectomesenchyme genes, all of which are expressed in the migratory CNCC precursors.

Figure 3.4: Expression of genes from the microarray and in the non-canonical Wnt signaling
pathway: Images of in situ hybridizations for genes rgs3 and nkd2a show that these genes are
expressed in the migratory CNCC that will give rise to ectomesenchyme.

In addition to the annotation of the expression patterns by cell-types, I also performed
an analysis to determine the pathways and cellular processes that the candidate genes
might be involved in. Upon gene-ontology analysis using Arraystar, two of the
interesting categories were neural crest migration and cell recognition (Figure 3.5).
91

Furthermore, upon detailed literature search, we found that genes like rgs3, nkd2a that
have been reported to function in the non-canonical Wnt pathway in zebrafish and
other systems (Freisinger et al., ; Van Raay et al., 2007). Whereas rgs3 has been
described as a downstream effector of the nkd2a Wnt5b pathway, is a negative
regulator of the Wnt signaling pathway. We examined the expression patterns of these
two genes in embryos at 18hpf and 24hpf, and found they were expressed in the CNCC
(Figure3.4). Taken together, based on our microarray data, Twist1 might function
upstream of the Wnt5 Planar Cell Polarity Pathway. Furthermore, genes like pcdh18a
have been shown to be important for regulation of cell-cell adhesion (Aamar and Dawid,
2008), loxl2 and lama4 are known to be a shown to be important for remodeling the
extracellular matrix , and essential components of the basement membrane and phlda3
has been defined as being important in cell-polarity regulation. All of these processes
have been shown to be important for regulation of cell migration and differential cell
sorting. Taken together, based on these studies, not only have we have identified
several new classes of molecules that are expressed specifically in the cranial neural
crest fated to form ectomesenchyme, we have also found that several of these
molecules might function in pathways that regulate specific cellular processes like
establishment of specific cell-cell adhesion, and also polarity and also migration.

92

Figure 3.5: Gene-ontology analyses of micro-array data using Arraystar:
The list above shows output from gene ontology analyses generated using Arraystar 4.0 that
represents a prediction of the most common processes that these candidates might be involved.

93

Table 3.1a: List of candidate genes enriched slightly but upregulated: The table
Lists candidate genes obtained from the microarray by sorting using the criteria stated above.

94

Table 3.1a continued: This table is a continuation of the table from the previous page.
The table is being presented in this manner to maintain a high enough resolution for clear
visibility and representation of the data.

From the list of genes represented in the table above, we decided to validate some
candidates via in situ at 18hpf. We find that compared to controls, embryos with
95

knockdown of twist1a/1b by morpholinos. We find that compared to controls, in
embryos injected with twist1a/1b, there is upregulation of atf3, cart1, and gch1, but no
significant change in expression of hish2l (Figure 3.6). Since atf3, cart1 and gch1 appear
to be co-expressed in a specific subset of the CNCC; these results might suggest that
these genes are involved in a common pathway or common biological process.

Figure 3.6: Validation of microarray data from Table 3.1: In situ hybridizations of atf3, cart1,
gch1 and hist2hl shows that in controls (A, C, E,G ) atf3, cart1, gch1 are expressed in the anterior
non- ectomesenchyme precursor population and hist2hl is expressed both in the pre-migratory
and migratory CNCC. Compared to controls, embryos injected with twist1a/1b-MO have slight
increase in expression of atf3, increased expression of cart1, of gch1 and slightly increased
expression of hist2hl.

96

Twist1 regulates fli1a expression likely directly

Figure 3.7: Twist1 directly regulate fli1a expression in ectomesenchyme. (A) Schematic
shows the fli1a genomic locus with hatch marks at 1 kb intervals. Predicted enhancers (grey
boxes, “A-H”) and the first exon (red box) are shown. Below are the various enhancer constructs
analyzed, with the hsp70I core promoter in blue and GFP in green. Wild-type (black) and mutant
(red) versions of the putative Twist1 binding site are also shown, as is the percentage of
transient transgenic embryos showing GFP expression in the pharyngeal arches or vasculature.
(B) Alignment of the central portion of the F enhancer between five fish species and mouse. The
putative bHLH/Twist1 binding site is boxed in red. (C and D) Confocal projections of 32 hpf
sox10: dsRed embryos injected at the one-cell-stage with fli1a-F enhancer constructs. The wild-
type (C) but not mutant (D) enhancer drives arch expression. (E and F) Confocal sections of
merged GFP and DIC channels show that a stable Tg(fli1a-F-hsp70I:GFP) line displays arch GFP
expression in 32 hpf wild-type (E) but not twist1a/b-MO (F) embryos. (G) Luciferase activity
relative to renilla firefly activity in 293T cells transfected with wild-type or mutant fli1a-F
enhancer reporter constructs, with or without a Twist1 expression plasmid. Asterisks indicate
significant comparisons using a student’s t-test (p<0.05). Arbitrary units and standard error of
the mean are shown.

Since we identified fli1a has one of the earliest markers of ectomesenchyme being
regulated by Twist1, we decided to characterize further how Twist1 regulates the
expression of fli1a. To do so, we first used a comparative genomics and transient
97

transgenic approach to identify ectomesenchyme-specific regulatory regions of the fli1a
gene. A ~15 kb region centered around the first exon of the fli1a gene had previously
been shown to drive expression in both the ectomesenchyme and vasculature (Lawson
and Weinstein, 2002), and we used the mVISTA program (Frazer et al., 2004) to identify
seven short (<500 bp) sub-regions (A-H) that were conserved between zebrafish, puffer
fishes (Takifugu rubripes and Tetraodon nigroviridis), stickleback (Gasterosteus
aculeatus), and medaka (Oryzias latipes). By testing the ability of these sub-regions to
drive GFP expression in conjunction with the hsp70I core promoter, we identified two
vasculature-specific enhancers (G and H) and one ectomesenchyme-specific enhancer
(F) (Figure 3.7A). Element F, located ~1.5 kb upstream of the fli1a promoter, was
sufficient to drive GFP expression in the ectomesenchyme from 19 hpf onwards, as
confirmed by co-localization with a sox10: dsRed transgene (Figure 3.6C). We also
identified homologous sequence to enhancer F ~1.7 kb upstream of the mouse Fli1
gene, which contained a perfectly conserved element, CAGATG, that matches the bHLH
binding site, CANNTG (though Twist1 most strongly prefers CATATG) (Kophengnavong et
al., 2000) (Figure 3.7B). This putative Twist1 binding site was required for enhancer
function, as mutation to GTATAC abolished the ability of element F to direct
ectomesenchyme expression in zebrafish embryos and to potentiate Twist1-dependent
transgene expression in mammalian 293T cells (Figure 3.7D,G). Twist1 was also required
for enhancer F activity as co-injection of twist1a/1b-MO prevented ectomesenchyme
expression in a stable Tg(fli1a-F-hsp70I:GFP) transgenic line (Figure 3.7E,F). Together,
98

our results strongly suggest that Twist1 directly activates fli1a expression through
binding to a conserved enhancer element.
Loss of fli1a results in defects in ectomesenchyme formation.
To examine the role of fli1a in specification of the ectomesenchyme lineage, we decided
to knockdown function by injecting a previously characterized translation blocking
morpholino. Injection of the morpholino into sox10: GFP transgenic embryos showed
severely reduced contribution of the neural crest to the ectomesenchyme of the
pharyngeal arches. Furthermore, skeletal staining at 5dpf showed severely reduced
head skeleton, similar to what was observed in loss of Twist1 function. We further
decided to examine the embryos for increased cell death in the fli1a loss of function,
and found that compared to controls, in embryos injected with fli1a-MO, there was
increased cell death in the migratory neural crest cells, which might have resulted in the
reduced pharyngeal arches. Although the morpholinos were published to be effective
for blocking translation of fli1a, we decided to examine the specificity of the phenotype
by co-injecting in vitro transcribed RNA for fli1a into single cell stage embryos. We
observed a high rate of mortality in the embryos injected with fli1a RNA on its own,
possibly due to the known role of this gene in gastrulation, which made the analysis of
the experiment challenging. Also, of the embryos co-injected with fli1aMO and fli1a
mRNA, there was enhancement of phenotype. Taken together, our preliminary analyses
show that fli1a might be involved in migration and survival of the neural crest.
99

Figure 3.8: Loss of fli1a results in defects in ectomesenchyme formation: (A, B, C) Confocal
projections of sox10: GFP embryos reduced contribution of neural crest cells to the pharyngeal
arches in fli1a Mo injected embryos (as indicated by the yellow arrow). Furthermore, there is
increased numbers of sox10: GFP positive cells in the more dorsal regions of the embryos. (D, E,
F)Alcian staining of the head-skeleton at 5dpf shows loss of cartilage elements derived from the
maxillary and mandibular arches in fli1a MO injected embryos. (G, H, I) Lysotracker staining of
embryos at 24hpf shows increased numbers of lysotracker positive cells in the fli1aMO injected
embryos compared to controls.

Injection of fli1a mRNA results in more severe phenotypes:
Upon injection of fli1a mRNA, there is an increased penetrance of the morpholino
phenotype, suggesting a dominant negative effect of excessive RNA. Since fli1a has
multiple splice variants, maybe changing the balance between the proteins affects
embryonic development. Therefore, in order to characterize the function of the protein,
one needs to generate a mutant that affects splicing of multiple variants of the protein
and thereby knocks out function of multiple proteins. Alternatively, a splice blocking
morpholino can be utilized. Taken together, since we obtain cell death in the neural
crest which could result from a defect in specification of the ectomesenchyme and a
100

defect in the head skeleton that closely resembles other mutants and morphants with
specification defects, our data suggests that fli1a might play an active role in
specification of the ectomesenchyme lineage.
Upregulation of sox10 results in defects in ectomesenchyme formation:
Since we observed an upregulation of sox10 in the ectomesenchyme precursors and a
subsequent defect in the development of the neural crest derived head skeleton, we
decided to investigate further how loss of sox10 played into the Twist1 dependent
pathway for specification of the ectomesenchyme lineage. To do so, we generated a
transgenic line that would allow us to misexpress sox10 in the embryo and examine
defects in specification of the ectomesenchyme lineage. We also wanted to establish
whether sox10 regulated fli1a expression. We therefore misexpressed UAS: hsox10 by
doing heat shock treatments between 12-16hpf and between 20-28hpf. Whereas upon
heat shocking for 12-16hpf, we observed a range of craniofacial defects at 4dpf in the
embryos positive for hsp70:Gal4; UAS: hsox10 compared to the control embryos (Figure
3.8) , there were no defects in craniofacial cartilage development in embryos heat shock
treated between 20-28hpf.

101

Figure3. 9: Misexpression of hsox10 results in defects in ectomesenchyme: Alcian staining of
4dpf larval head skeleton shows that compared to control (A), embryos with hsp70:gal4; UAS:
hsox10 have reduction in Hs and Hm as indicated by the black arrow (B,C), truncation in pq
indicated by black arrow (D,E) and severe reductions of all elements of the head skeleton
accompanied by heart edema (F).

In situ hybridization of fli1a at 28hpf shows that compared to controls, embryos
expressing hsp70:gal4; UAS: hsox10 reduction in the arch size of pharyngeal arches 1 &2
with no changes in the posterior arches 3-7 and in the expression of fli1a transcript. For
the embryos heat shocked between 12-16hpf, predominantly there were defects in the
Hm and Hs elements with increased severity showing reduction of the entire CNCC head
skeleton. We further performed these experiments using hsp70:gal4; fli1a:GFP embryos
to observe effects on the development of the pharyngeal arches. We found that
compared to controls in embryos expressing hsp70:gal4; UAS: hsox10, there was a
reduction of the pharyngeal arch size although expression of fli1a transcript was
unaffected.
102

Reduction of sox10 in twist1a/1b-MO partially rescues skeletal defects:
To determine whether sox10 actively blocks induction of the ectomesenchyme lineage
by Twist1, we decided to inject Twist1a/1b-morpholino in the background of sox10
mutants called colourless (Kelsh and Eisen, 2000). Upon injections of twist1a/1b
morpholino into the colourless mutant, we observed that there was a reduced
penetrance of the severe phenotype of the twist1a/1b- MO as compared to the
complete penetrance of the twist1a/1b morpholino injected embryos. The results from
this experiment are summarized in Table 3.2. Since we did observe a reduced
penetrance of the twist1 morphants, this would indicate that sox10 participates in the
pathway of ectomesenchyme specification, but there might be processes downstream
of Twist1 but in parallel to sox10 which are also crucial for specification of the
ectomesenchyme.

103

Genotype

Phenotype
twist1a/1b-
MO
sox10+/-
sox10+/-,
twist1a/1b-MO
sox10-/-
sox10-/-,
twist1a/1b-MO
no defect 0 13/14 0 2/2 2/7
intermediate 0 0 5/14 0 1/7
severe 18 0 9/14 0 4/7
edema only 0 1/14 0 0 0

Table 3.2: Phenotypic scoring of craniofacial defects in colourless mutants and twist1a/1b
injected embryos: Summary of number of embryos with no phenotype, edema only,
intermediate and severe phenotypes in each of the genotypes stated. There is a shift in the
number of embryos in each class between twist1a/1b-MO only and sox10
mutant+twist1a/1bMO injected embryos.

104

Twist1 regulation of Fgf and Fgf regulating sox10:
Since others and we have shown that Fgf signaling was important for specification of the
ectomesenchyme and in addition found that Fgf signaling does not regulate expression
of twist1 target fli1a, we decided to investigate further Fgf signaling could be
downstream of Twist1. To do so, we decided to assay for Fgf activity in the twist1a/1b
morphants by examining expression of transcription factor pea3, which has been shown
to be a downstream target of Fgf signaling in zebrafish (Roehl and Nusslein-Volhard,
2001). We find that compared to controls, in embryos injected with twist1a/1b
morpholinos, there was a reduction of pea3 expression in the ectomesenchyme crest
(Figure 3.10).

Figure 3.10: Twist1 regulates Fgf activity: In situ hybridizations of pea3 expression at 18hpf in
(A) control and (B) twist1a/1b-MO injected embryos shows reduction of pea3 in migrating
CNCCs in loss of Twist1.

105

Discussion:

Figure 3.11: Model for function of Twist1 in the Cranial neural crest: The above presented
model shows that Twist1 functions as a master-regulator of fate switch between the early
neural crest and ectomesenchyme by promoting the expression of key players of the
ectomesenchyme as well as repressing the non-ectomesenchyme

Twist1 promotes ectomesenchyme at the expense of non-ectomesenchyme fates
Twist1 is required within CNCCs for ectomesenchyme development in mice (Bildsoe et
al., 2009), and here in the previous chapter we show that zebrafish twist1a and twist1b
are redundantly required at the earliest steps of ectomesenchyme specification. In
theory, Twist1 could promote the transition from a multipotent CNCC precursor to a
more lineage-restricted ectomesenchyme cell, or alternatively influence fate choices
between ectomesenchyme and non-ectomesenchyme lineages. However, the
observation that only a subset of early CNCC genes persisted in twist1a/1b-MO arches,
106

combined with the normal development of CNCC-derived cranial pigment and glial cells,
suggest that CNCCs do not remain as multipotent precursors in the absence of Twist1
function. Instead, the ectopic arch expression of sox10 and foxd3 in Twist1-depleted
embryos, at a stage when these genes are normally co-expressed in the non-
ectomesenchyme, suggests that presumptive ectomesenchyme precursors partially
adopt a non-ectomesenchyme expression profile in the absence of Twist1. This
conclusion was confirmed by our microarray analysis in which non-ectomesenchyme but
not early CNCC genes were the most up-regulated in Twist1-deficient embryos.
Unexpectedly, some non-ectomesenchyme genes, such as nr4a2b and gch2, were up-
regulated in the non-ectomesenchyme domain but not ectopically expressed in Twist1-
deficient arches, suggesting that Twist1 may have additional roles in restricting non-
ectomesenchyme gene expression within the early non-ectomesenchyme domain.
Consistent with only a subset of non-ectomesenchyme genes showing ectopic arch
expression, we also found no evidence for pigment cells and functional glia forming
ectopically in Twist1-deficient arches. Instead, rather than fully trans-fating into other
CNCC derivatives, the increase in cell death observed in Twist1-deficient arches suggests
that some of these mis-specified arch CNCCs may undergo cell death in the absence of
Twist1.

107

Twist1 promotes ectomesenchyme through both Fgf regulation and direct gene
activation of Fli1a
Our data also indicate that Twist1 promotes ectomesenchyme fates through two largely
parallel pathways. For one, the finding that Twist1 was required for expression of the
Fgf target gene pea3 in arch ectomesenchyme suggests a role for Twist1 in potentiating
Fgf signaling. Twist1 might do so by regulating expression of Fgf ligands and/or
receptors, as Twist1 has been shown to regulate the CNCC expression of both Fgf10 and
Fgfr1 in mice (Soo et al., 2002). Hence, Twist1 could function upstream of
autocrine/paracrine Fgf signaling within CNCCs that promotes ectomesenchyme
development. However, although Fgf signaling was required to repress sox10 and
activate dlx2a expression, we found that fli1a ectomesenchyme expression was Fgf-
independent. Instead, fli1a appears to be a potentially direct target of Twist1, and we
have identified a conserved ectomesenchyme enhancer element of fli1a that is Twist1-
dependent. However, chromatin immunoprecipitation experiments (currently not
feasible due to the lack of a high quality Twist1 antibody) will eventually be needed to
show enhancer occupancy by Twist1. Moreover, the function of Fli1a in
ectomesenchyme formation also remains to be investigated. Targeted disruption of the
Fli1 gene in mouse results in embryonic lethality by E12.5 due to severe hemorrhaging
thus precludes an analysis of its role in development of the ectomesenchyme-derived
108

craniofacial skeleton (Spyropoulos et al., 2000). In addition, since fli1a has multiple
splice variants, knockdown or loss of one of the isoforms or a mutation that blocks
splicing of a single isoform might not result in a complete loss of function. Thereby,
either a splice blocking morpholino that generates non-functional transcripts for all
transcripts or a mutant that blocks splicing of multiple isoforms will be necessary to
obtain a genetically null situation for the fli1a gene. In the future, the generation of
zebrafish fli1a mutants and/or conditional Fli1 mouse knockouts will allow us to
determine the extent to which Fli1 mediates the effects of Twist1 in craniofacial
development.
Stepwise model for specification of the ectomesenchyme:
Taken together, based on our studies, I would like to propose a step-wise model for
specification of the ectomesenchyme mediated by transcription factor Twist1, whereby
the migratory CNCC need to first shut off transcription of non-ectomesenchyme genes
and further turn on expression of genes that would further facilitate differentiation of
the ectomesenchyme lineage.

109

Sox10 and fli1a function in independent pathways to regulate specification of
ectomesenchyme
Based on our genetic analyses, I suggest that sox10 and fli1a are both important
molecules for specification of the ectomesenchyme lineage, yet they function in distinct
pathways involved in this process and play independent roles.
Identification of novel regulators of the ectomesenchyme lineage
Since we show previously that specification of the ectomesenchyme versus the non-
ectomesenchyme occurs prior to cells migrating to the pharyngeal arches,
characterization of the function of these genes might provide ectomesenchyme gene,
we studies, we conclude that our microarray had worked and we could utilize the data
to investigate the cellular bases of specification of the ectomesenchyme. Taken
together, several of the molecules identified downstream of Twist1 appear to play roles
in cell-cell and cell-matrix interactions. Thereby, detailed analyses of the functions of
these molecules might facilitate a complete understanding of how cell- cell and cell-
matrix interactions of the CNCC with their neighbors play a crucial role in specification of
the ectomesenchyme.

110

CHAPTER 4: CONCLUSIONS AND FUTURE DIRECTIONS
When I initiated my graduate studies, although ectomesenchyme specification had been
a topic of extensive interest and research, the precise timing of specification of the
lineage was unknown. Whereas some studies had suggested that there was lineage
restriction at the neural tube, it was undetermined how the fate switch occurred in the
pre-migratory CNCC such that it would give rise to one lineage versus the other. In
addition, whether the fate switch was dependent on signals at the origin, along the way
or at the destination was also unknown. Some studies had suggested Fgf signaling from
the destination as being important in regulating a fate choice of the CNCC to the
ectomesenchyme, yet some of the defining molecular events of the transition from a
more multipotent precursor to the ectomesenchyme precursor such as down-
regulation of early neural crest markers, and induction of ectomesenchyme specific
genes occurred prior to these cells reaching their final destination. This prompted me to
investigate the problem further of how ectomesenchyme is specified in a migratory
population like the CNCC as the embryo develops and the CNCCs contact multiple
tissues as they migrate.
In my initial studies, I decided to investigate whether the surrounding tissues of the
mesoderm and endoderm were necessary for specification of the ectomesenchyme, and
by genetic ablation studies utilizing the effects of blocking Nodal signaling on the
embryo, I uncovered that both the mesoderm and the endoderm were dispensable for
111

specification of the ectomesenchyme lineage, and then I turned my attention to search
for signals at the ectoderm. Since Bmp signaling had been shown to be important for
specification of the neural crest, and change in Bmp signaling resulted in early
embryonic defects, it has never been investigated whether change in Bmp activity could
be important for specification of ectomesenchyme lineage from the CNCC. Taking
advantage of a Gal4 driver in the laboratory that allowed for manipulation of Bmp
signaling after the cells had migrated out of the neural tube; I was able to investigate in
detail a role of Bmp signaling in specification of the ectomesenchyme.
Firstly, I observed that there was a change in Bmp activity in the migratory neural crest
that correlated with specification of the ectomesenchyme. Through our genetic
analyses of conditional misexpression of Bmp4 in the migratory neural crest specifically,
I found that down regulation of Bmp signaling is essential for specification of the
ectomesenchyme. From these studies, I would like to conclude that a change in Bmp
activity in the migratory CNCC acts as the temporal mechanism for a fate switch in the
cranial neural crest to form ectomesenchyme versus non-ectomesenchyme. Although
we show conclusively that differential Bmp activity in the migratory CNCC might be the
molecular bases for specification of the ectomesenchyme, it remains to be discovered
how Bmp activity is being regulated in a precise manner, which then facilitates the
crucial lineage transition. Bmp activity is known to be regulated by inhibitors like noggin
and gremlins in several contexts. Although noggin2 is expressed in the head in the
112

surrounding paraxial mesoderm, we show that loss of mesoderm was dispensable in
induction of the ectomesenchyme lineage. Another possible candidate is gremlin2
(grem2), which is expressed in the migratory CNCC and has been shown to regulate
Bmp4 activity in the neural crest. In order to determine the role of grem2 in regulation
of the Bmp activity, I would utilize the grem2 morpholino to first knockdown grem2
function and examine a change in the pSMAD activity. If the loss of the grem2 via
morpholino results in prolonged pSMAD1/5/8 localization in the CNCC, then I would
conclude that the grem2 is essential in exerting a precise temporal control of
specification of the ectomesenchyme lineage. Interestingly, microarray analysis of
neural crest cells from control versus twist1a/1b-MO treated embryos identified grem2
as one of the top hits in regulation by twist1a/1b. Upon validation of this result via in-
situ hybridization, I found that there was a partial reduction of grem2 expression in
twist1a/1b knockdown. This would suggest that twist1 regulates Bmp activity. In
addition, a recent study identified snw1 as a novel molecule that regulates the spatial
acitivity of Bmp signaling (Wu et al.). Published expression pattern suggests that the
gene is expressed in the head of the zebrafish embryo. Therefore, this molecule could
be an interesting candidate to investigate further for its possible role in regulation of
Bmp activity.
Whereas we show that Bmp activity at the origin restricts ectomesenchyme formation,
others and we have found that Fgf activity in the migratory CNCCs is essential for
113

induction of some of the essential features of the ectomesenchyme lineage (Blentic et
al., 2008; Yamauchi et al.). Although there is Fgf signaling from the brain, and the oral
ectoderm, as well as pharyngeal endoderm, signaling from which these tissues is
important remains unknown. Fgf activity is present in the neural crest cells, as soon as
start migrating out of the neural crest, yet the gene expression changes due to Fgf is not
manifested until later. This might be attributed to the fact that reporter of Fgf activity,
pea3 is regulated by Twist1 and we know from our studies that Twist1 does not function
until a certain time during CNCC development, although it is expressed starting in the
pre-migratory neural crest.
In addition to understanding the role of Bmp signaling in regulating the lineage decisions
of the CNCC into ectomesenchyme, I also investigated the role of cell intrinsic molecules
in this process. Previous studies in mice had established that Twist1 was crucial in the
process of ectomesenchyme development. What was unknown was whether this
molecule regulated an early fate switch of the CNCC to ectomesenchyme or it functions
later to maintain or promote differentiation of the ectomesenchyme precursor
population. After several loss of function studies, I would like to conclude that Twist1
plays a crucial role in the transition between the ectomesenchyme and the non-
ectomesenchyme lineage. Whereas twist1 genes are expressed starting early in
development, they do not function until later. I discovered that the precise regulation of
Twist1 activity is regulated by Id2a, which classically functions by altering transcriptional
114

complexes of bHLH proteins like Twist1. Since some of other studies might indicate that
Twist1 possibly forms different transcriptional complexes in different subpopulations of
the neural crest during development, and these differential complexes have distinct
transcriptional consequences (Connerney et al., 2006), it would make sense to think that
altering the transcriptional complex that Twist1 would normally form in the
ectomesenchyme precursor would alter the lineage outcomes. Our genetic analyses of
misexpression of Id2a and loss of Twist1 suggest that Id2a modulation of Twist1
complex formation could alter lineage outcomes of the CNCC into ectomesenchyme.
Although our genetic analyses show that id2a is critical for regulation of twist1 function,
how id2a regulates twist1 function remains to be discovered by biochemical methods.
One limitation in that regard is the ability to get enough starting material for
biochemical analysis. Recent studies (unpublished) have established several neural crest
cell lines that could be utilized to understand the molecular mechanisms for
specification of the various cell types of the neural crest. From my studies so far, since I
also found that exclusion of Id2a precedes induction of ectomesenchyme markers and
coincides with specification of the ectomesenchyme, I would like to conclude that Id2a
is a critical molecule in determining the fate switch from the multi-potent precursor to
the ectomesenchyme.
Although through my studies, I have determined that Twist1 functions in the early
migratory crest to specify the ectomesenchyme, the precise timing when Twist1 is
115

functioning remains to be discovered. To determine this, we need to first identify genes
that are expressed in different subsets of the pre-migratory and migratory neural crest
cells and then generate transgenic lines that would drive expression of the dominant
negative Twist1 protein in specific contexts. From our microarray studies, I have
identified several candidates that are expressed in specific subsets of the neural crest.
Upon further characterization by fluorescent two color in situ hybridization studies, I
plan to identify specific subsets that these genes label, then clone the regulatory
elements for expression of these genes, and then drive expression of dominant negative
Twist1 in those domains. These detailed analyses will facilitate a more complete
understanding function of Twist1 in the process of specification of the
ectomesenchyme.
An initial survey of the literature of the possible functions of the novel downstream
targets of Twist1, several of them could be playing active roles in remodeling of the cell-
extracellular matrix and establishing specific adhesion properties between groups of
cells. Since these processes have been shown to be important for establishment of
proper patterns or domains of precursor cells for later tissue differentiation, an
understanding of these molecules might lead us to understand how the different
precursor populations within the CNCC establish and maintain their patterns in the head
of the embryo, which then determines how they differentiate. In other words, our
future loss and gain of function genetic analyses of the newly identified molecules like
116

pchd18a, loxl2a, rgs3, nkd2a and others might explain the cellular bases for segregation
of the ectomesenchyme from the non-ectomesenchyme in the head.

117

REFERENCES
Aamar, E. and Dawid, I. B. (2008). Protocadherin-18a has a role in cell adhesion,
behavior and migration in zebrafish development. Dev Biol 318, 335-46.

Ahlgren, S. C. and Bronner-Fraser, M. (1999). Inhibition of sonic hedgehog signaling in
vivo results in craniofacial neural crest cell death. Curr Biol 9, 1304-14.

Amaya, E., Musci, T. J. and Kirschner, M. W. (1991). Expression of a dominant negative
mutant of the FGF receptor disrupts mesoderm formation in Xenopus embryos. Cell 66,
257-70.

Balczerski, B., Matsutani, M., Castillo, P., Osborne, N., Stainier, D. Y. and Crump, J. G.
(2012). Analysis of Sphingosine-1-phosphate signaling mutants reveals endodermal
requirements for the growth but not dorsoventral patterning of jaw skeletal precursors.
Dev Biol 362, 230-41.

Baroffio, A., Dupin, E. and Le Douarin, N. M. (1991). Common precursors for neural and
mesectodermal derivatives in the cephalic neural crest. Development 112, 301-5.

Bildsoe, H., Loebel, D. A., Jones, V. J., Chen, Y. T., Behringer, R. R. and Tam, P. P.
(2009). Requirement for Twist1 in frontonasal and skull vault development in the mouse
embryo. Dev Biol 331, 176-88.

Blentic, A., Tandon, P., Payton, S., Walshe, J., Carney, T., Kelsh, R. N., Mason, I. and
Graham, A. (2008). The emergence of ectomesenchyme. Dev Dyn 237, 592-601.

Brown, L. A., Rodaway, A. R., Schilling, T. F., Jowett, T., Ingham, P. W., Patient, R. K.
and Sharrocks, A. D. (2000). Insights into early vasculogenesis revealed by expression of
the ETS-domain transcription factor Fli-1 in wild-type and mutant zebrafish embryos.
Mech Dev 90, 237-52.

Carney, T. J., Dutton, K. A., Greenhill, E., Delfino-Machin, M., Dufourcq, P., Blader, P.
and Kelsh, R. N. (2006). A direct role for Sox10 in specification of neural crest-derived
sensory neurons. Development 133, 4619-30

Connerney, J., Andreeva, V., Leshem, Y., Muentener, C., Mercado, M. A. and Spicer, D.
B. (2006). Twist1 dimer selection regulates cranial suture patterning and fusion. Dev Dyn
235, 1345-57.
118

Corsi, A. K., Brodigan, T. M., Jorgensen, E. M. and Krause, M. (2002). Characterization
of a dominant negative C. elegans Twist mutant protein with implications for human
Saethre-Chotzen syndrome. Development 129, 2761-72.

Crump, J. G., Maves, L., Lawson, N. D., Weinstein, B. M. and Kimmel, C. B. (2004). An
essential role for Fgfs in endodermal pouch formation influences later craniofacial
skeletal patterning. Development 131, 5703-16.

David, N. B., Saint-Etienne, L., Tsang, M., Schilling, T. F. and Rosa, F. M. (2002).
Requirement for endoderm and FGF3 in ventral head skeleton formation. Development
129, 4457-68.

Dorsky, R. I., Moon, R. T. and Raible, D. W. (1998). Control of neural crest cell fate by
the Wnt signalling pathway. Nature 396, 370-3.

Dupin, E., Calloni, G. W. and Le Douarin, N. M. (2004). The cephalic neural crest of
amniote vertebrates is composed of a large majority of precursors endowed with
neural, melanocytic, chondrogenic and osteogenic potentialities. Cell Cycle 9, 238-49.

Eom, D. S., Amarnath, S., Fogel, J. L. and Agarwala, S. (2011). Bone morphogenetic
proteins regulate neural tube closure by interacting with the apicobasal polarity
pathway. Development 138, 3179-88.

Firulli, B. A., Krawchuk, D., Centonze, V. E., Vargesson, N., Virshup, D. M., Conway, S.
J., Cserjesi, P., Laufer, E. and Firulli, A. B. (2005). Altered Twist1 and Hand2 dimerization
is associated with Saethre-Chotzen syndrome and limb abnormalities. Nat Genet 37,
373-81.

Frazer, K. A., Pachter, L., Poliakov, A., Rubin, E. M. and Dubchak, I. (2004). VISTA:
computational tools for comparative genomics. Nucleic Acids Res 32, W273-9.

Freisinger, C. M., Fisher, R. A. and Slusarski, D. C. Regulator of g protein signaling 3
modulates wnt5b calcium dynamics and somite patterning. PLoS Genet 6, e1001020.

Furthauer, M., Thisse, B. and Thisse, C. (1999). Three different noggin genes antagonize
the activity of bone morphogenetic proteins in the zebrafish embryo. Dev Biol 214, 181-
96.

Furthauer, M., Thisse, C. and Thisse, B. (1997). A role for FGF-8 in the dorsoventral
patterning of the zebrafish gastrula. Development 124, 4253-64.
119

Germanguz, I., Lev, D., Waisman, T., Kim, C. H. and Gitelman, I. (2007). Four twist
genes in zebrafish, four expression patterns. Dev Dyn 236, 2615-26.

Gitelman, I. (1997). Twist protein in mouse embryogenesis. Dev Biol 189, 205-14.
Harris, M. L. and Erickson, C. A. (2007). Lineage specification in neural crest cell
pathfinding. Dev Dyn 236, 1-19.

Heckman, K. L. and Pease, L. R. (2007). Gene splicing and mutagenesis by PCR-driven
overlap extension. Nat Protoc 2, 924-32.

Henion, P. D. and Weston, J. A. (1997). Timing and pattern of cell fate restrictions in the
neural crest lineage. Development 124, 4351-9.

Hopwood, N. D., Pluck, A. and Gurdon, J. B. (1989). A Xenopus mRNA related to
Drosophila twist is expressed in response to induction in the mesoderm and the neural
crest. Cell 59, 893-903.

Jin, E. J., Erickson, C. A., Takada, S. and Burrus, L. W. (2001). Wnt and BMP signaling
govern lineage segregation of melanocytes in the avian embryo. Dev Biol 233, 22-37.

Kelsh, R. N. (2006). Sorting out Sox10 functions in neural crest development. Bioessays
28, 788-98.

Kelsh, R. N., Dutton, K., Medlin, J. and Eisen, J. S. (2000). Expression of zebrafish fkd6 in
neural crest-derived glia. Mech Dev 93, 161-4.

Kelsh, R. N. and Eisen, J. S. (2000). The zebrafish colourless gene regulates development
of non-ectomesenchymal neural crest derivatives. Development 127, 515-25.

Kikuchi, Y., Agathon, A., Alexander, J., Thisse, C., Waldron, S., Yelon, D., Thisse, B. and
Stainier, D. Y. (2001). casanova encodes a novel Sox-related protein necessary and
sufficient for early endoderm formation in zebrafish. Genes Dev 15, 1493-505.

Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B. and Schilling, T. F. (1995).
Stages of embryonic development of the zebrafish. Dev Dyn 203, 253-310.

Knight, R. D. and Schilling, T. F. (2006). Cranial neural crest and development of the
head skeleton. Adv Exp Med Biol 589, 120-33.

120

Kophengnavong, T., Michnowicz, J. E. and Blackwell, T. K. (2000). Establishment of
distinct MyoD, E2A, and twist DNA binding specificities by different basic region-DNA
conformations. Mol Cell Biol 20, 261-72.

Kwan, K. M., Fujimoto, E., Grabher, C., Mangum, B. D., Hardy, M. E., Campbell, D. S.,
Parant, J. M., Yost, H. J., Kanki, J. P. and Chien, C. B. (2007). The Tol2kit: a multisite
gateway-based construction kit for Tol2 transposon transgenesis constructs. Dev Dyn
236, 3088-99.

LaBonne, C. and Bronner-Fraser, M. (1998). Neural crest induction in Xenopus: evidence
for a two-signal model. Development 125, 2403-14.

Lawson, N. D. and Weinstein, B. M. (2002). In vivo imaging of embryonic vascular
development using transgenic zebrafish. Dev Biol 248, 307-18.

Le Douarin, N. M. (1999). The neural crest: Cambridge university press.

Liem, K. F., Jr., Tremml, G., Roelink, H. and Jessell, T. M. (1995). Dorsal differentiation
of neural plate cells induced by BMP-mediated signals from epidermal ectoderm. Cell
82, 969-79.

Liu, F., Walmsley, M., Rodaway, A. and Patient, R. (2008). Fli1 acts at the top of the
transcriptional network driving blood and endothelial development. Curr Biol 18, 1234-
40.

Martinsen, B. J. and Bronner-Fraser, M. (1998). Neural crest specification regulated by
the helix-loop-helix repressor Id2. Science 281, 988-91.

Meulemans, D. and Bronner-Fraser, M. (2004). Gene-regulatory interactions in neural
crest evolution and development. Dev Cell 7, 291-9.

Muller, II, Knapik, E. W. and Hatzopoulos, A. K. (2006). Expression of the protein
related to Dan and Cerberus gene--prdc--During eye, pharyngeal arch, somite, and swim
bladder development in zebrafish. Dev Dyn 235, 2881-8.

Nguyen, V. H., Schmid, B., Trout, J., Connors, S. A., Ekker, M. and Mullins, M. C. (1998).
Ventral and lateral regions of the zebrafish gastrula, including the neural crest
progenitors, are established by a bmp2b/swirl pathway of genes. Dev Biol 199, 93-110.

Northcutt, R. G. and Gans, C. (1983). The genesis of neural crest and epidermal
placodes: a reinterpretation of vertebrate origins. Q Rev Biol 58, 1-28.
121

Ragland, J. W. and Raible, D. W. (2004). Signals derived from the underlying mesoderm
are dispensable for zebrafish neural crest induction. Dev Biol 276, 16-30.

Roehl, H. and Nusslein-Volhard, C. (2001). Zebrafish pea3 and erm are general targets
of FGF8 signaling. Curr Biol 11, 503-7.

Santiago, A. and Erickson, C. A. (2002). Ephrin-B ligands play a dual role in the control of
neural crest cell migration. Development 129, 3621-32.

Scheer, N. and Campos-Ortega, J. A. (1999). Use of the Gal4-UAS technique for targeted
gene expression in the zebrafish. Mech Dev 80, 153-8.

Schilling, T. F. and Kimmel, C. B. (1994). Segment and cell type lineage restrictions
during pharyngeal arch development in the zebrafish embryo. Development 120, 483-
94.

Soo, K., O'Rourke, M. P., Khoo, P. L., Steiner, K. A., Wong, N., Behringer, R. R. and Tam,
P. P. (2002). Twist function is required for the morphogenesis of the cephalic neural
tube and the differentiation of the cranial neural crest cells in the mouse embryo. Dev
Biol 247, 251-70.

Sperber, S. M., Saxena, V., Hatch, G. and Ekker, M. (2008). Zebrafish dlx2a contributes
to hindbrain neural crest survival, is necessary for differentiation of sensory ganglia and
functions with dlx1a in maturation of the arch cartilage elements. Dev Biol 314, 59-70.

Spyropoulos, D. D., Pharr, P. N., Lavenburg, K. R., Jackers, P., Papas, T. S., Ogawa, M.
and Watson, D. K. (2000). Hemorrhage, impaired hematopoiesis, and lethality in mouse
embryos carrying a targeted disruption of the Fli1 transcription factor. Mol Cell Biol 20,
5643-52.

Steventon, B., Araya, C., Linker, C., Kuriyama, S. and Mayor, R. (2009). Differential
requirements of BMP and Wnt signalling during gastrulation and neurulation define two
steps in neural crest induction. Development 136, 771-9.

Tallafuss, A. and Bally-Cuif, L. (2003). Tracing of her5 progeny in zebrafish transgenics
reveals the dynamics of midbrain-hindbrain neurogenesis and maintenance.
Development 130, 4307-23.

Teng, L. and Labosky, P. A. (2006). Neural crest stem cells. Adv Exp Med Biol 589, 206-
12.
122

Theveneau, E., Marchant, L., Kuriyama, S., Gull, M., Moepps, B., Parsons, M. and
Mayor, R. (2010). Collective chemotaxis requires contact-dependent cell polarity. Dev
Cell 19, 39-53.

Theveneau, E. and Mayor, R. (2011). Collective cell migration of the cephalic neural
crest: the art of integrating information. Genesis 49, 164-76.

Thomas, A. J. and Erickson, C. A. (2009). FOXD3 regulates the lineage switch between
neural crest-derived glial cells and pigment cells by repressing MITF through a non-
canonical mechanism. Development 136, 1849-58.

Tribulo, C., Aybar, M. J., Nguyen, V. H., Mullins, M. C. and Mayor, R. (2003). Regulation
of Msx genes by a Bmp gradient is essential for neural crest specification. Development
130, 6441-52.

Uribe, R. A. and Gross, J. M. (2010). Id2a influences neuron and glia formation in the
zebrafish retina by modulating retinoblast cell cycle kinetics. Development 137, 3763-74.

Van Raay, T. J., Coffey, R. J. and Solnica-Krezel, L. (2007). Zebrafish Naked1 and Naked2
antagonize both canonical and non-canonical Wnt signaling. Dev Biol 309, 151-68.

Wada, N., Javidan, Y., Nelson, S., Carney, T. J., Kelsh, R. N. and Schilling, T. F. (2005).
Hedgehog signaling is required for cranial neural crest morphogenesis and
chondrogenesis at the midline in the zebrafish skull. Development 132, 3977-88.

Weston, J. A., Yoshida, H., Robinson, V., Nishikawa, S., Fraser, S. T. and Nishikawa, S.
(2004). Neural crest and the origin of ectomesenchyme: neural fold heterogeneity
suggests an alternative hypothesis. Dev Dyn 229, 118-30.

Wu, M. Y., Ramel, M. C., Howell, M. and Hill, C. S. SNW1 is a critical regulator of spatial
BMP activity, neural plate border formation, and neural crest specification in vertebrate
embryos. PLoS Biol 9, e1000593.

Yamauchi, H., Goto, M., Katayama, M., Miyake, A. and Itoh, N. Fgf20b is required for
the ectomesenchymal fate establishment of cranial neural crest cells in zebrafish.
Biochem Biophys Res Commun 409, 705-10.

Yamauchi, H., Goto, M., Katayama, M., Miyake, A. and Itoh, N. (2011). Fgf20b is
required for the ectomesenchymal fate establishment of cranial neural crest cells in
zebrafish. Biochem Biophys Res Commun 409, 705-10.
123

Yan, Y. L., Willoughby, J., Liu, D., Crump, J. G., Wilson, C., Miller, C. T., Singer, A.,
Kimmel, C., Westerfield, M. and Postlethwait, J. H. (2005). A pair of Sox: distinct and
overlapping functions of zebrafish sox9 co-orthologs in craniofacial and pectoral fin
development. Development 132, 1069-83.

Yokota, Y. (2001). Id and development. Oncogene 20, 8290-8.

Zuniga, E., Rippen, M., Alexander, C., Schilling, T. F. and Crump, J. G. (2011). Gremlin 2
regulates distinct roles of BMP and Endothelin 1 signaling in dorsoventral patterning of
the facial skeleton. Development 138, 5147-56.

Abstract (if available)

Abstract Vertebrate cranial neural crest cells (CNCCs) contribute not only to ectodermal lineages like neurons, glia and pigment but also to “ectomesenchymal” lineages like cartilage and bone. Whereas studies have established that in zebrafish the CNCCs are lineage restricted at the neural tube, the molecular bases for regulation of the cell lineage remains unknown. In this thesis work, I will discuss my studies on the role of Bmp signaling from the ectoderm in restricting the ectomesenchyme potential of the CNCC. I will provide evidence for functions of Id2a and Twist1 proteins in specification of the ectomesenchyme. We show that although twist1 genes are expressed in the CNCC starting at pre-migratory stages, presence of Id2a in these cells prevents Twist1 from functioning, and that a loss of Id2a in the migratory CNCCs over time in development facilitates specification of the ectomesenchyme lineage. Furthermore, I will discuss a detailed characterization of the roles of Twist1 in specification of the ectomesenchyme lineage in zebrafish. We propose that Twist1 functions as a master-regulator in specification of the ectomesenchyme lineage via regulating induction of early ectomesenchyme genes and repressing the non-ectomesenchyme.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Elucidating the role of neural crest specific Stat3 signaling in maintaining coronal suture patency during embryonic development

PDF

Epigenetic checks and balances: PHF6 activity restricts neural crest migration

PDF

Developmental biology of epicardial adipose tissue

PDF

TWIST1 functions in both mesoderm and neural crest derived cranial tissues to establish and maintain coronal suture patency

PDF

The role of α-catlulin during tumor progression and early mouse development

PDF

Lineage boundaries and cell migration in the patterning of the mammalian skull

PDF

Jagged1 functions downstream of Twist1 in the specification of the coronal suture and the formation of a boundary between osteogenic and non-osteogenic cells

PDF

Tissue-specific action of Msx genes in the regulation of skull vault development

PDF

The mechanisms of somatic cell reprogramming

PDF

Utilizing zebrafish and mouse models to uncover the underlying genetics of human craniofacial anomalies

PDF

Novel roles for Maf1 in embryonic stem cell differentiation and adipogenesis

PDF

The role of Prkci in stem cell maintenance and cell polarity using a 3-D culture system

PDF

Asymmetric cell division during neurogenesis, and the mechanisms behind GABAergic cortical interneuron development and specification

PDF

Multiple functions of the PR-Set7 histone methyltransferase: from transcription to the cell cycle

PDF

The function of BS69 in mouse embryogenesis and embryonic stem cell differentiation

PDF

Molecular basis of mouse epiblast stem cell and human embryonic stem cell self‐renewal

PDF

Identification and characterization of adult stem cells in the oral cavity

PDF

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

PDF

Elucidating the functional role of CHD7 associated nuclear PDH complex and other associated proteins on neural crest development

PDF

The calcium-sensing receptor in the specification of normal and malignant hematopoietic cell localization in the bone marrow microenvironment

Asset Metadata

Creator Das, Ankita (author)

Core Title Tissue interactions & molecular pathways in specification of the ectomesenchyme from cranial neural crest

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Genetic, Molecular and Cellular Biology

Publication Date 05/15/2012

Defense Date 01/27/2012

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

Tag ectomesenchyme,neural crest,OAI-PMH Harvest

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Maxson, Robert E., Jr. (committee chair), Lien, Ellen (committee member), Lu, Wange (committee member), Sucov, Henry (committee member)

Creator Email ankita23@gmail.com,ankitada@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-40333

Unique identifier UC11289304

Identifier usctheses-c3-40333 (legacy record id)

Legacy Identifier etd-DasAnkita-852.pdf

Dmrecord 40333

Document Type Dissertation

Rights Das, Ankita

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