University of Southern California Dissertations and Theses

Epigenetic checks and balances: PHF6 activity restricts neural crest migration

(USC Thesis Other)

Epigenetic checks and balances: PHF6 activity restricts neural crest migration

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

EPIGENETIC CHECKS AND BALANCES: PHF6 ACTIVITY RESTRICTS NEURAL
CREST MIGRATION

by

Erin Catherine Moran

A dissertation submitted in partial fulfillment of the requirements of the degree of
Doctor of Philosophy, Genetics Molecular and Cellular Biology
Faculty of the USC Gradute School
University of Southern California
December, 2017

2
Table of Contents
Abstract……………………………………………………………………………………………………3
Acknowledgements………………………………………………………………………………………..4
Chapter 1: Introduction……………………………………………………………………………………5
Section 1.1 Epigenetic regulation is crucial to development………………………………………..…5
Section 1.2 Neural crest is particularly susceptible to changes in epigenetic regulators…………….17
Section 1.3 PHF6 is an adapter protein known to cause developmental defects…………………….26
Section 1.4 PHF6 and epigenetic regulation of neural crest: thesis outline………………………….31
Chapter 2: Detailed Materials and Methods………………………………………………………………32
Section 2.1 Xenopus laevis embryology……………………………………………...………………32
Section 2.2 Human neural crest differentiation……………………………………..………………..46
Section 2.3 Bacterial protein preparation………………………………………….…………………58
Chapter 3: Spatio-temporal expression of Phf6…………………………………………………………..60
Section 3.1 Introduction……………………………………………………………..……………….60
Section 3.2 Results……………………………………………………………………………..…….62
Section 3.3 Discussion………………………………………………………………………..……...75
Chapter 4: PHF6 negatively regulates neural crest cell migration in Xenopus laevis.……………………85
Section 4.1 Introduction……………………………………………………………..…………….....85
Section 4.2 Results…………………………………………………………………..…………….....87
Section 4.3 Discussion………………………………………………………………..…………….106
Chapter 5: Mechanistic evaluation of PHF6 interaction with chromatin……………………………….110
Section 5.1 Introduction……………………………………………………………..……………...110
Section 5.2 Results………………………………………………………………………..………...112
Section 5.3 Discussion………………………………………………………………………..…….126
Chapter 6: A role in mesoderm formation, and other PHF6 details………..............…………….…….137
Section 6.1 Introduction………………………………………………………………………..…...137
Section 6.2 Results……………………………………………………………………………..…...138
Section 6.3 Discussion……………………………………………………………………….……..145
Chapter 7: The role of PHF6 in chromatin biology and development: perspectives and conclusions…..150
Appendix I. References……………………………………………………..………………………..…158

3
Abstract
Börjeson-Forssman-Lehmann Syndrome (BFLS) is an X-linked intellectual disability
syndrome that is characterized by unique facial features that involve broad, course faces, heavy-
set jaw bones, and thickened calvaria among other symptoms. This syndrome is characterized by
mutation in a dual PHD protein, PHF6. This protein has been recently characterized as being
associated with ribosomal DNA transcription, but very few studies have shown any role for PHF6
in development or a cellular phenotype in a developmental setting despite its role in BFLS. It is
the subject of this thesis to define the role of PHF6 in neural crest development, in order to
determine how this characteristic craniofacial phenotype arises, and to determine a cellular
mechanism of PHF6 activity within neural crest cells. I used an approach that combined the
strengths of Xenopus laevis embryology to define the role of Phf6 in neural crest in vivo, and
human embryonic stem cells differentiated to neural crest cells to define the cellular and
biochemical characteristics of PHF6 in vitro. I found that PHF6 protein and Phf6 mRNA is highly
expressed in the neural plate border and in pre-migratory neural crest cells, but that there is reduced
to absent expression of PHF6 in migratory neural crest cells. I found that Phf6 depletion leads to
a loss of temporal control of neural crest migration, while ectopic expression prevents migration
in a tissue- and cell-autonomous manner. Furthermore, I found that PHF6 tracks to poised but
inactive enhancer regions in pre-migratory neural crest, and physically associates with the
chromatin remodeler CHD7 in this context without the presence of other neural crest active
chromatin remodeling factors. This led me to develop the model that PHF6 binds to poised
enhancer regions to prevent precocious activation of these regions through the activity of
chromatin remodelers like CHD7, and it maintains the poised state until it is down-regulated at the
point of neural crest epithelial-to-mesenchymal transition.
4
Acknowledgements

I would like to acknowledge the help of my mentor, Dr. Ruchi Bajpai, for providing
encouragement, advice, and support.
I would like to acknowledge the guidance and support of my current committee members,
Dr. Michael Stallcup, Dr. Robert Maxson, and Dr. Jian Xu, for guidance and support, and for
going above and beyond for me when called. I would also like to acknowledge the guidance and
support of previous committee members, Dr. Peggy Farnham, Dr. Judd Rice, Dr. Peter Jones, Dr.
Gage Crump, and Dr. Pedro Sanchez.
For instruction, mentorship, guidance, and advice, I would like to thank Dr. Raymond
Keller. For working with me day in and day out on my experiments, I would like to thank Dr.
Katherine Pfister. For indispensable assistance with ChIP assays, I would like to acknowledge
the present and past Farnham lab members, especially Dr. Esther Tak, Dr. Malaina Gaddis, and
Dr. Adam Blattler. For allowing me use of lab space, reagents, and for advice, I would like to
thank Dr. Marianne Bronner.
For assistance within the lab and on various experiments, I would like to thank Dr.
Richard Pelikan, Dr. Soma Samantha, Dr. Daniela Schmid, Susan Smith, Mallory Holland,
Jennifer Oki, William Ciodza, Philbert Mach, Maria Nava, Annie Lynch, Kaivalya Shevade,
Casey Griffin, and many other summer students, STAR students, and dental students within the
Bajpai lab. For discussions of data, projects, and advice, I would like to acknowledge Dr. Amy
Merril, Dr. Cynthia Neben, Joanna Salva, Ryan Roberts, and Dr. Creighton Tuzon from the
Merril Lab.

5
Chapter 1: Introduction

Section 1.1. Epigenetic regulation is crucial to development
1.1.1 Epigenetics is the study of the structure of Chromatin
The term epigenetics has come to describe the study of that which “stands upon” the genetic
code, or modifies the way in which the genetic code is structured or read out within the nucleus.
The DNA strand combined with the proteins that encapsulate and shape it are collectively known
as chromatin, and the basic unit of chromatin is the nucleosome. The nucleosome consists of a
histone octamer, where 2 of each histone H2A, H2B, H3, and H4 form a wheel-like structure with
protruding positively charged tails. Approximately 144 bases of DNA are wrapped around the
histone octamer, its negatively charged backbone interacting electrostatically with positively
charged residues at 14 contact points
1
(Fig. 1.1). This basic unit is repeated millions of times across
the genome, with varying lengths of unoccupied DNA interspersing the nucleosome-occupied
DNA regions.
The study of nucleosome-unoccupied DNA, or nucleosome-free regions, has been one of
high focus in recent years, as RNA polymerases and many transcription factors cannot bind
directly to a piece of DNA that is nucleosome-occupied. The average length of nucleosome-free
regions is highly dependent upon its location in active chromatin regions, or euchromatin, or in
repressed chromatin regions, or heterochromatin. Within euchromatin, it is common to have
stretches of DNA 200 base pairs or more in length
2
that are nucleosome-free in order to allow
transcription factor and polymerase binding
2–5
, whereas in heterochromatin, the DNA is highly
compacted and coated with nucleosomes, heterochromatic proteins such as HP1, and scaffolding
6
Figure 1.1. Schematic representation of the histone octamer embedded in DNA to form a
nucleosome. Lysine residues within the tail regions of the proteins are represented.

H3 H4
H2A H2B
K4
K9
R8
K27
K5
K16
K5
K9
K9
K12
K36
7

proteins in order to condense the DNA into small pockets within the nucleus
6,7
. Indeed, recent
research has identified distinct pockets of different chromatin conformations within the nucleus,
with heterochromatin largely localized along the nuclear periphery and often in contact with the
nuclear laminin, and active chromatin often localized towards the center of the nucleus
8
. These
distinct pockets of chromatin are thought to allow the cell to sequester chromatin into active
regions of the nucleus to increase efficiency of transcription, and are characterized by differential
post-translational modifications on the histone tails
8
.
Histone post-translational modifications (PTMs) have long been associated with distinct
roles of certain areas of chromatin, and correlated with accessibility of regions of chromatin. These
roles can be found in detailed view in Table 1.1. Examples of histone PTMs associated with
heterochromatin are the di- or tri-methylation of lysine 9 of Histone 3 (H3K9me2/3), which have
been shown to bind directly to HP1 in order to condense the chromatin
9
. Conversely, methylation
of lysine 4 of Histone 3 (H3K4me1/2/3) is well associated with euchromatin, with H3K4me1/2
primarily associated with transcriptional enhancer regions
10,11
and H3K4me3 primarily associated
with active transcriptional start sites
12
. While the association of specific K4 methylation marks with
enhancers or promoters has been questioned recently as being imperfectly correlated, it still stands
that active enhancer regions are more likely to be marked with H3K4me1/2 than not
13–15
.
The study of histone PTMs largely focus on the deposition and removal of these
modifications, or writing and erasing, and the binding of these modifications, or reading, by
various proteins. Writers and erasers set the stage for the formation of protein complexes that affect
change in the chromatin structure, and readers, which recognize the deposited PTMs, form the core
of these protein complexes. Most chromatin remodeling complexes involve at least one reader
protein
16–20
, which allows specific targeting of chromatin remodeling complexes to the regions of
8
Table 1.1. Histone modifications and their functions.
Histone Modification Activation/Repression Function
H2A K5 Acetylation Activation Transcriptional Activation
H2A R3 Symmetric
dimethylation
Repression Transcriptional Repression
H2A R3 Asymmetric
dimethylation
Activation Transcriptional Activation
H2B K5, K12, K15, K20
Acetylation
Activation Transcriptional Activation
H3 K9, K27, K36
Acetylation
Activation Transcriptional Activation
H3 K14, K18, K23
Acetylation
Activation Transcriptional Activation, DNA
repair
H3 R2 methylation Repression Transcriptional Repression
H3 R8 asymmetric
dimethylation
Repression Transcriptional Repression, blocking
of H3K4methyl binding
H3 R17, R26, R42
methylation
Activation Transcriptional Activation
H3 K4 methylation Activation Me1/Me2, active enhancer regions;
Me3, active promoter regions
H3 K9 methylation Repression Heterochromatin, silencing genomic
imprinting
H3 K27 methylation Repression Transcriptional repression,
bivalency, X-inactivation
H3 K36 methylation Activation Transcriptional elongation
H3 K79 Activation Transcriptional elongation,
Checkpoint response
H4 K5, K8, K12, K16
acetylation
Activation Transcriptional Activation
H4 R3 Symmetric
dimethylation
Activation Transcriptional Activation
H4 R3 Asymmetric
dimethylation
Repression Transcriptional Repression
H4 K20 methylation Repression Transcriptional silencing,
Heterochromatin

9
the genome where remodeling is necessary. The exact mechanisms behind these activities has been
an area of intense research, and this thesis involves one mechanism by which reader proteins can
not only affect remodeling at specific regions, but can also prevent premature remodeling at these
same regions.

1.1.2 Many epigenetic factors are mutated in complex developmental disorders
Complex developmental disorders are those that involve birth defects that affect more than
one system; here, I will discuss those that have been linked to a genetic cause involving an
epigenetic factor. These factors can be a writer, reader, eraser, or remodeler. Developmental
disorders highlight the necessity for the genes involved within developmental processes as
mutations of these genes cause defects that are not able to be rescued by the embryo via a redundant
mechanism. By studying how a gene causes certain developmental disorders, we can therefore
gain some insight into which developmental processes for which a gene is necessary. We also gain
a medicinal utility of study of the gene, in that by learning about the function of a gene we can
understand the precise mechanism of developmental failure within the system with the aim of
ultimately finding means of correcting this developmental failure and ameliorating the lives of
patients.

i. CHARGE syndrome (OMIM 214800)
CHARGE syndrome is a disorder that affects multiple systems, most notably involving
craniofacial abnormalities, coloboma of the eye, heart defects, choanal atresia, growth retardation,
genital abnormalities, and ear abnormalities
21–23
. The large majority of CHARGE syndrome patients
have a de novo mutation in the CHD7 gene
24–26
. CHD7 is an ATP-dependent chromatin remodeler,
10
which is known to bind methylated lysine 4 of H3 through its chromodomains, and has been found
to track with H3K4me1 within the genomes of stem cells
27,28
. This protein is not only a reader, but
has the enzymatic ability to shift and remove nucleosomes in order to expose or mask regions of
DNA. CHD7 is required for the acquisition of multiple cellular phenotypes, most notably that of
migratory and multipotent neural crest
29,30
.

ii. Coffin-Siris Syndrome (OMIM 135900)
Coffin-Siris Syndrome is a complex disorder that is caused by genetic mutations in a
number of epigenetic factors. The majority of the cases have been linked to a mutation in the
ARID1B gene
31
, but other mutations have been noted, such as ARID1A, SMARCA4, SMARCB1,
SMARCE1, and SOX11
32
. All mutations have been shown to be inherited in an autosomal-dominant
fashion, and many of the identified proband patients have de novo mutations
32
. Patients with this
syndrome exhibit developmental or cognitive delay, hypoplasia of the 5
th
and additional digits,
hypotonia, malformation of cardiac, gastrointestinal, genitourinary, and central nervous systems
with varying penetrance of each phenotype
33
. All of the genes found mutated in this syndrome with
the exception of SOX11 are known to be subunits or associated proteins of the SWI/SNF chromatin
remodeling complex in its various forms. This complex is one of the crucial reading/remodeling
complexes involved in a large number of chromatin functions. ARID1B and ARID1A are
mutually-exclusive proteins in this complex
34
, and both forms are known to play a role in chromatin
remodeling in neural progenitors and neural remodeling complexes (npBAF and nBAF)
35–37
.
SMARCA4, otherwise known as BRG1, is a highly-noted protein component of all BAF
complexes and also of the CREST-BRG1 calcium-responsive chromatin remodeling complex
38
.
11
SMARCB1 is a core component of the BAF complex, while SMARCE1 is an interchangeable
component of the BAF complex
38
.

iii. Nicolaides-Baraitser Syndrome (OMIM 601358)
Nicolaides-Baraitser syndrome (NCBRS) is defined by mutations in the SMARCA2 gene
in an autosomal dominant fashion
39
. NCBRS is defined by severe to moderate intellectual disability,
characteristic coarse facial features, microcephaly, seizures, and decreased subcutaneous fat
39
. In
all the reported cases, SMARCA2 was found to contain a de novo mutation
39
. SMARCA2 is a
component of the BAF complex, and is a homologue of the Drosophila Brahma protein which has
both helicase and ATPase activities and is thought to perform chromatin remodeling
40
.

iv. Cornelia de Lange Syndrome (OMIM 122470)
Cornelia de Lange Syndrome (CdLS) is separated into 5 different subsets based on the
genetic mutation causing the disorder, namely NIPBL, RAD21, SMC3, HDAC8, or SMC1A. CdLS
is defined by characteristic craniofacial features, microcephaly, growth retardation, limb reduction
defects, intellectual disability, autistic tendencies, cardiac septal defects, genital defects, hearing
loss, gastrointestinal dysfunction, and myopia
41
. Mutations in NIPBL, RAD21, and SMC3 are
inherited in an autosomal dominant fashion
42
, while HDAC8 and SMC1A mutations are inherited
in an X-linked fashion
43
. NIPBL, RAD21, SMC3, and SMC1A are all thought to be involved in
overall chromatin structure through association with centromeres or scaffolding
44,45
, while HDAC8
removes acetyl groups on acetylated lysines, thereby being characterized as a chromatin eraser
46
.

v. Rubinstein-Taybi Syndrome (OMIM 180849)
12
Rubinstein-Taybi Syndrome (RTS) is characterized by mutations in either p300 or CBP,
two related histone acetyltransferases. Patients with this disorder have characteristic palate
malformations as well as several other craniofacial malformations, brachydactyly, short stature,
intellectual disability, joint hypermobility, all with close to complete penetrance
47,48
. These two
proteins together make up a large proportion of the histone acetyltransferase activity within the
cell, and are both involved in gene activation in response to b-catenin nuclear localization.

vi. X-linked Syndromic Mental Retardation, Claes-Jensen Type (OMIM 300534)
This disorder, hereby known as MRXSJ, is the group of X-linked intellectual disability
patients with known mutations in JARID1C. These mutations have been relatively recently
identified as a cause of X-linked intellectual disability
49
, and since then characteristic phenotypes
have emerged. These include severe intellectual disability, seizures, hypertonia, short stature,
palate abnormalities, head size abnormalities, abnormalities of the testes, and other phenotypes
49
.
JARID1C recognizes H3K9me3 through its PHD domain and demethylates H3K4 specifically,
making this protein a reader as well as an eraser
50
.

vii. Sifrim-Hitz-Weiss Syndrome (OMIM 617159)
Two recent studies
51,52
have identified that the ATP-dependent chromatin remodeler CHD4
also has a developmental role; mutations in this gene cause an intellectual disability disorder that
has characteristic phenotypes. Patients exhibit intellectual disability, short stature, macrocephaly,
facial dysmorphism, palatal anomalies, cardiac abnormalities, hearing loss, hypogonadism, and
other phenotypes with varying penetrance
51,52
. The CHD4 protein is an ATP-dependent chromatin
13
remodeler associated with the nucleosome remodeling and histone deacetylation complex
(NuRD), a chromatin remodeling and erasing complex associated with gene repression
52
.

viii. Autism Spectrum Disorders
In addition to intellectual disability disorders, a number of genetic causes of autism
spectrum disorders (ASD) have been linked to mutations in chromatin-associated and chromatin-
remodeling proteins. O’Roak et al.
53
performed genome analysis of exome sequencing samples of
autism spectrum disorders, finding that approximately 1% of these patients have linked mutations
in a set of 6 genes. This gene list included the chromatin remodeler CHD8, and chromatin
remodeling complex member BAF180 was found to contain mutations at a lower rate. Subsequent
analysis for de novo mutations causing ASD identified additional epigenetic factors DNMT3B and
CHD7
54
. Additional analyses have continued to confirm the link between autism spectrum disorders
and epigenetic regulators, such as BAF155, BAF170
55
.

ix. Additional disorders
The list of currently known developmental disorders caused by chromatin regulators and
epigenetic factors can be found in Table 1.2, adapted from Ronan, Wu, and Crabtree, 2013
56
to
include more recent discoveries.
Each of these disorders qualifies as a rare disease, but combined, they offer a more
complete idea of the importance of epigenetics in development. Also notable is the fact that each
disorder involves multiple systems, indicating that these proteins are indispensable for the fate of
multiple types of cells. This provides a picture whereby epigenetics is indispensable for the
formation of almost all cell types within the developed organism.
14
Table 2.2. Mutations in epigenetic factors cause diverse developmental defects. Adapted from
Ronan, Wu, and Crabtree 2013
56
.
15
OMIM Disorder Gene Phenotype Craniofacial
135900 Coffin-Siris
Syndrome
ARID1A, ARID1B,
SMARCA2, BRG1,
BAF47
Course facial features, mental retardation, hypertrichosis,
sparse scalp hair, hypoplastic or absent fifth fingernails,
congenital heart defects.
Course facial
features
209850 Autism ARID1B, BAF155,
BAF170, BAF180,
CHD7, CHD8,
MBD5, EHMT1
Limited or absent verbal communication, lack of reciprocal
social interaction, restricted and ritualized patterns of interests
and behavior.
Facial structure
dysmorphism
214800 CHARGE
syndrome
CHD7 Coloboma of the eye, heart anomaly, choanal atresia, mental
retardation, ear abnormalities, facial palsy, cleft palate
Dysmorphic facial
structures, cleft
palate
301040 X-linked ⍺-
thalassaemia and
mental syndrome
ATRX Mental retardation, alpha-thalassemia, microcephaly, flat face,
hemoglobin-H disease, genital abnormalities
Microcephaly, flat
face
180849 Rubinstein-Taybi
Syndrome
P300, CBP Mental retardation, postnatal growth deficiency, microcephaly,
broad thumbs, dysmorphic facial features
Microcephaly,
dysmorphic facial
features
603736 Say-Barber-
Biesecker-
Young-Simpson
Syndrome
KAT6B Distinctive facial appearance, severe blepharophimosis,
hypotonia, joint laxity, long thumbs, cardiac defects, thyroid
abnormalities
Distinctive facial
appearance
600430 Brachydactyly
Mental
Retardation
Syndrome
HDAC4 Brachydactyly affecting the metacarpals and metatarsals,
moderate intellectual disability, behavioral abnormalities,
dysmorphic facial features
Dysmorphic facial
features
277590 Weaver’s
Syndrome
EZH2 Pre- and postnatal overgrowth, accelerated osseous maturation,
broad forehead and face, micrognathia, developmental delay
Broad forehead and
face, micrognathia
610253 Kleefstra’s
Syndrome
EHMT1, BAF47,
MLL3, MBD5
Mental retardation without speech development, hypotonia,
microcephaly, brachycephaly, hypertelorism, midface
hypoplasia
Microcephaly,
brachycephaly,
hypertelorism,
midface hypoplasia
605130 Wiedemann-
Steiner
Syndrome
MLL Short stature, hypertrichosis cubiti, thick or arched eyebrows
with a lateral flare, downslanting and vertically narrow
palpebral fissures, intellectual disability
Characteristic
facial features
147920 Kabuki’s
Syndrome
MLL2 Postnatal dwarfism, long palebral fissures, broad and depressed
nasal tip, cleft or high arched palate, large ear lobes, scoliosis
Cleft or high
arched palate,
characteristic facial
features
300534 Claes-Jensen
Type X-linked
mental
retardation
(XLMR)
JARID1C Severe mental retardation, facial hypotonia, maxilliary
hypoplasia, aggressive behavior, short stature, epilepsy
Maxilliary
hypoplasia, facial
hypotonia
16

300263 Siderius XLMR PHF8 Mild mental retardation, long faces, broad nasal tip, cleft lip
and palate
Long faces, cleft lip
and palate
300706 XLMR, Turner
type
HUWE1 Mental retardation, macrocephaly, holoprosencephaly, tapering
fingers
Macrocephaly
312750 Rett’s Syndrome MECP2 Arrested development between 6 and 18 months, loss of
speech, microcephaly, mental retardation
Microcephaly
309520 Lujan-Fryns
Syndrome
MED12 Marfanoid habitus, long, narrow face, small mandible, high-
arched palate
Narrow face, small
mandible, high-
arched palate
614249 Non-syndromic
Intellectual
Disability
MED23 Mild to moderate mental retardation
601224 Potocki-Shaffer
Syndrome
PHF21A Developmental delay, craniofacial abnormalities, intellectual
disability, biparietal foramina
Craniofacial
abnormalities
617159 Sifrim-Hitz-
Weiss Syndrome
CHD4 Intellectual disability, cardiac abnormalities, hearing loss, non-
specific facial dysmorphism
Facial
dysmorphism
17

Section 1.2. Neural crest cells are particularly susceptible to changes in epigenetic regulators
1.2.1 Most developmental disorders caused by epigenetic regulators involve craniofacial
malformations
As we observe from Table 1.2, almost all of the developmental disorders linked to
epigenetic factors that have been identified at this point involve a craniofacial component. In fact,
it can be said that two hallmarks of germline inheritance of epigenetic mutations are cognitive
defects and craniofacial malformations. This is significant because neural tissue and neural crest
tissue, the development of which can lead to cognitive defects and craniofacial defects
respectively, have a very similar developmental origin. Both tissues form from dorsal ectoderm
collectively known as the neural ectoderm. This region then differentiates into the neural plate and
neural plate border during late gastrulation to early neurulation
57
. The neural plate then closes to
form the neural tube, bringing the neural plate border regions in close proximity as the neural fold
regions
57
. After fusion of the neural tube, these specified neural crest cells within the neural fold
undergo an epithelial-to-mesenchymal transition to form migratory neural crest
57
. The neural tube
then goes on to form the central nervous system and its derivatives, whereas the crest goes on to
form the craniofacial structures, melanocytes, neurons and glia of the peripheral nervous system,
and numerous other cell types
57
.

1.2.2 Craniofacial disorders are caused by defects in neural crest during development
The developmental origin of the craniofacial structures can be traced to the development
of a particular cell type, the neural crest. The neural crest is a cell type that becomes specified
during early neurulation along the neural plate border, and undergoes a transition to become
18
migratory and multipotent during late neurulation. The multi-potentiality of these cells is unusual,
as the vast majority of developmental cell types become increasingly lineage-restricted during the
developmental timeline; the neural crest is one of the few that is able to reverse the trend toward
developmental restriction and gain the ability to have mesenchymal fates as well as ectodermal
cell fates
57
. This switch in cell fates is associated with a large transcriptional network that is
necessary to derive the neural crest cell phenotype. The neural plate border is initially specified
during gastrulation, where signaling from the forming mesoderm and the surrounding ectoderm
induces the neural ectoderm using WNT, BMP and FGF signals
58
(Fig. 1.2). The interaction of
WNT and BMP signaling from the lateral ectoderm and WNT and BMP inhibitors secreted from
the neural ectoderm creates a region of intermediate signaling of these two factors, leading to the
specification of the neural plate border in this region
59
. FGF signaling from the underlying
mesoderm and NOTCH signaling help further define the neural plate border region
60–62
. The
culmination of these signaling events creates a cascade of transcriptional activation, activating
transcription factors ZIC1, MSX1, GBX2, PAX3, PAX7, TFAP2a, DLX5, and DLX6
63–65
. These
then feed back onto themselves
66–69
, reinforcing the expression of these genes across the neural plate
border region, and creating a broad swath of expression in the area that is competent to form neural
crest. By the end of gastrulation and the transition to neurulation, the neural plate border is
specified. From this region, neural crest is then specified
70
as the neural plate begins to close,
resulting in the activation of neural crest specific transcription factors FOXD3, SNAIL1, SNAIL2,
and ETS1 early
65
, followed by SOX9, nMYC, cMYB, ID, and TWIST
71
, along with continued
activation of TFAP2a, PAX3, and PAX7
72,73
. The neural crest specifier genes are expressed more
medially within the neural plate border. The early neural crest markers begin to turn on as the
neural plate border becomes mature during early neurulation, while the later neural crest markers
19
Figure 1.2. Neural crest specification and epithelial to mesenchymal transition. Left, schematic in
cross section view of neural plate (blue), neural plate border (green), surrounding non-neural
ectoderm (gray), and the notochord (red), over the course of neural crest specification and initial
migration. Right, gene regulatory modules that govern each stage of this development, and the
feed forward and feedback loops that promote transcriptional control of neural crest. Adapted from
Simões-Costa and Bronner, 2015
70
.

Figure 1.2. Neural crest morphogenesis and the gene regulatory network that derives it. Simões-Costa and Bronner, 2015
Signaling Module
WNT BMP
Notch FGF
Neural Plate Border Module
Zic1 Msx1 Gbx2
Pax3 Pax7 Tfap2
Dlx5 Dlx6
Neural Crest Specification Module
FoxD3 Snail1 Tfap2
Twist1 Sox5 Pax3
Pax7 Sox9 Ets1
Myc Snail2 Id
Sox10 Myb
EMT
Neural Crest Migration Module
Sox10 Sox5 Ebf1
Pax3 FoxD3 Pax7
Tfap2 RxrG Myc
Sox9 Id Snail1
Snail2 Myb
Neural Crest Differentiation
20

turn on as the neural crest becomes specified and increase in expression as neural crest begin to
delaminate from their original position along the neural plate border during late neurulation.
During neural crest migration, the expression of neural crest transcription factors that turn on
during specification are largely maintained through forward feedback control loops
71,74–76
, while
migration-specific transcription factors, namely SOX5, SOX10, EFB1, and RXRG, are also highly
expressed. It is also at the point of neural crest delamination that cells undergo a cadherin switch
in most model organisms from type 1 cadherins to type 2 cadherins
77–81
; this defines those crest cells
that have delaminated versus those that have yet to delaminate. After delamination, the neural crest
cell becomes increasingly mesenchymal and migratory in phenotype
82
, with multiple cytoskeletal
elements becoming involved in determining the path of migration via chemoattraction and
chemorepulsion
82–84
. The pathway that the cells migrate down is highly determined by the time point
of delamination as well as the position along the anterior/posterior axis; in turn, the pathway that
the cells migrate down is informative in terms of the type of cell into which the neural crest cell
will differentiate
85–87
. The most anterior portion of neural crest, collectively known as the cranial
neural crest, form specific streams known as the pharyngeal arches (PA) and the frontonasal
process (FNP); these streams will eventually converge to form the bone, cartilage, and connective
tissue of the face (Fig. 1.3). Extensive work using model organisms has shown that defects in some
part of neural crest development ultimately leads to malformation of the craniofacial structures
88–91
.
We can therefore link the epigenetic regulators that, when mutated, have a craniofacial
malformation associated with a defect in the correct development of neural crest.

21
Figure 1.3. Neural crest development and the craniofacial structures that arise from the cranial
neural crest. Left, cranial neural crest migration in Xenopus laevis leads to the formation of the
brachial cartilage structures at tadpole stages, with defined contributions from each pharyngeal
arch. Right, cranial neural crest migration in human embryos leads to the formation of the facial
bones with defined contributions from each pharyngeal arch.

Figure 1.3. Neural crest in the context of craniofacial development. Baltzinger et al, 2005, Santagati and Rijli, 2003
Xenopus laevis Homo sapiens
Neural crest migration
Craniofacial Structures
Frontal bone
Zygomatic bone
Maxilliary bone
Mandible
Hyoid
Nasal bone
Malleus
Incus
Stapes
Meckel’s Cartilage
Palatoquadrate
Cartilage
Ceratohyal
Cartilage
Gill
Cartilage
22

1.2.3 Epigenetics plays a significant role in neural crest delamination.
Neural crest delamination has several defining characteristics (Reviewed in Simões-Costa
and Bronner, 2015
70
). First, the cell undergoes a large phenotypic change, starting from an
epithelial, columnar type cell that is bound tightly to its neighbors and becoming migratory, with
a different cytoskeletal organization, and distinct lamellipodial projections, secreting extracellular
matrix components and binding to matrix components while not making connections with nearby
neural crest cells. Second, the cell activates a large number of new transcription factors that are
responsible for creating the phenotypic changes that are observed. Third, the cells at this point
become less lineage restricted, as discussed above. Fourth, only a subset of the cells that express
neural plate border genes then go on to delaminate and form bona fide neural crest cells. These
points are important to note in that they highlight the importance of epigenetic regulation in the
specification and delamination of neural crest cells.
Activation of a unique set of transcriptional regulators that then lead to downstream
phenotypic changes in the cell indicate that there is a distinct shift in the transcriptional profile of
these cells before and after delamination. This shift in the transcriptional profile must be
accomplished with the help of epigenetic regulators, as the access of transcription factors to
specific sites within the DNA is limited by the epigenetic context of these sites. For example, an
important site of neural crest specific transcription promotion or enhancement might be blocked
by a nucleosome bearing repressive chromatin marks before neural crest delamination, but become
nucleosome depleted and thereby revealed with the help of chromatin writers, erasers, and
remodelers during the delamination process, allowing the transcription of the target (Fig. 1.4). An
important point to note is that not only is there a shift of promoter activation in cells before versus
during the delamination process, but there is also a unique set of enhancers that become activated
23
Figure 1.4. Epigenetic regulation of neural crest transcription by activation of neural crest
enhancers.

Figure 1.4. Model of epigenetic regulation of neural crest specific transcription
Neural Crest Enhancer Neural Crest Promoter
Bivalent Inactive
Active Remodeling
Remodeling Complex
Looping interaction
TF TF
HAT
RNA Pol
II
Active Gene Complex
H3K4me1
H3K4me3
H3K27me3
H3K27ac
Histone PTM legend
24
during this process
92
. This indicates that there is a large chromatin remodeling shift during this
time, changing the epigenetic landscape and allowing a unique group of enhancers to activate
specific neural crest targets.
The third point, namely that cells become less lineage restricted during neural crest
delamination, is important in that it indicates why so many epigenetic factors seem to be involved
in this process. The Waddington epigenetic landscape hypothesis
93
(Fig. 1.5), depicts a cell in an
embryo as a rolling ball, which has several lineage decisions to make. These are illustrated as
valleys within the hill landscape, and the ball rolls down the diagram making several cell fate
choices as it goes. This is a striking visual representation, in that it not only eloquently illustrates
the point of lineage restriction, but also allows that there is an ability to change this lineage
restriction; however, significant energy input is required to make these changes. In this sense, the
transition from neural plate border ectoderm to neural crest represents a dangerous time in the cell
fate decision process whereby many different epigenetic factors are required to overcome the
“energy hill”, in order to change the lineage restriction of the cell. The observation that depleting
either CHD7 or BRG1 or both together prevents the migration of neural crest cells
29
is one measure
of proof of the requirement for chromatin remodeling to change this lineage restriction.
The fourth point, that only a subset of cells that are specified to be neural plate border cells
end up becoming migratory neural crest cells, further illustrates this need for epigenetic regulation.
While most of these cells receive many of the same signals from surrounding tissues, they express
very similar levels of transcription factors and express a broad measure of the same markers.
However, only those that are able to set up the correct chromatin environment to allow the shift in
the epigenetic landscape and form the correct downstream accessibility of transcription factors to
the correct targets are the cells that are successfully able to become migratory and
25
Figure 1.5. Waddington’s Epigenetic landscape, adapted to demonstrate the epigenetic barriers
between neural differentiation and neural crest cell fate.

Human embryonic stem cell
Xenopus blastomere
Neural epithelial
cell
Neural crest cell
26

multipotent neural crest cells. This, however, also emphasizes that, much in the same way that we
have transcriptional enhancers and repressors, we must also have epigenetic factors that promote
a certain cell fate switching as well as epigenetic factors that repress a certain cell fate. On the
epigenetic level, cells must be responsive to negative feedback signals in order to not allow cells
to undergo cell fate switching too early, or too often. In this work, I will expand on this idea to
provide one mechanism by which neural crest cell fate is restricted.

Section 1.3. PHF6 is a chromatin adapter protein known to cause developmental defects
1.3.1 Börjeson-Forssman-Lehmann Syndrome is caused by PHF6 mutation in patients
Börjeson-Forssman-Lehmann Syndrome (BFLS) is a multi-symptom birth defect
syndrome initially characterized in 1962 by Drs. Börjeson, Forssman, and Lehmann as a novel X-
linked recessive mental deficiency syndrome that is accompanied by epilepsy and endocrine
disorder
94
. Further studies over the next 40 years, accompanied by identification of additional
patients
95–99
, led to progressive narrowing of the chromosomal window over which the causative
mutation was located
100–104
, finally leading to the discovery of mutations within the PHF6 gene that
were identified as causative of the disease
105
. This disorder primarily affects male patients, causing
characteristic facial features, intellectual disability, epilepsy, obesity with gynocomastia, enlarged
and elongated ear lobes, hypogonadism, and developmental delay
94,105
. Female carriers of the disease
either are completely asymptomatic but with a pattern of skewed X-inactivation
105
, or show a set
of more subtle phenotypes
106,107
. More recent work has discovered a set of female patients with de
novo mutations in PHF6 and a set of similar, though not identical, clinical manifestations
108,109
; these
mutations are likely to be more severe in nature as many of the described patients have large
portions of the PHF6 protein deleted, and these patients show haploinsufficiency that is not
27
observed in hereditary forms of BFLS
109
. Specifically, these patients seem to have their own
characteristic facial phenotypes, dental abnormalities, linear hyperpigmentation, and digit
abnormalities
109
. Continued research has solidified the causative relationship between the PHF6
gene and this developmental disorder, outlining the need to determine the function of this protein
within the various affected cell types.

1.3.2 PHF6 is a dual PHD finger protein with no identified enzymatic activity
Prior to the identification of PHF6 as the causative gene for BFLS in 2002, there were no
published reports of PHF6 gene function. Indeed, this paper identified this gene not only as the
causative gene for BFLS, but also outlined the two plant homeodomains (PHD) of this protein as
well as 4 nuclear localization signals and a prospective nucleolar localization signal
105
(Fig. 1.6).
Further structural studies have not identified new domains within this protein, but have defined
the second PHD domain as having a “zinc knuckle” domain N-terminal to the standard PHD
110
,
meaning that there is an additional zinc coordination motif that causes folding of the protein in
such a way that it can block the standard protein binding pocket of the PHD. EMSA studies on the
second PHD determined that this portion of the protein binds non-specifically to DNA and does
not bind histone tails
110
. A recent study has identified portions of the secondary structure of the first
PHD, but has not identified the binding partners of this domain
111
. The lack of enzymatic activity
of this protein suggests that PHF6 likely acts as an adapter protein, forming a complex between
the chromatin and other binding partners
112
. As such, the need to identify the binding partners of
PHF6 as well as the precise chromatin signature that this protein recognizes is key to unlocking
the function of the protein.

28
Figure 1.6. Schematic representation of the PHF6 protein and the mutations associated in male
BFLS (top) and female BFLS (bottom).

PHDF1 PHDF2
M1T, M1V
K8X
C45Y
C99F
R342X
K234E
M46Dfs
H229R
R257G
I314V
D333del
PHF6
NLS1 NLS2 3 4
Male
BFLS
G10
fsX21
G226E
fsX279
C305F
R319X
Female
BFLS
Missense Frameshift Nonsense Deletion Large Deletion Duplication
G89Y
29
1.3.3 PHF6 has been observed within the nucleolus of cancer cell lines and is thought to regulate
the cell cycle through regulating rDNA transcription
Several different studies have focused on the sub-cellular localization of the PHF6 protein
in multiple cell lines. High resolution microscopy of GFP-tagged PHF6 constructs expressed in
HeLa cells with progressive deletion of the C-terminus of the protein showed that the full-length
protein is localized primarily to the nucleoli of these cells, with a more dispersed nuclear staining
105
.
As the NLS sequences are progressively deleted, the staining becomes more nuclear and
cytoplasmic
105
. Further studies have verified this nucleolar staining in HEK293T cells
113
and in HeLa
cells
114
through co-immunofluorescence with Nucleolin and Fibrillarin, respectively. Additionally,
deletion studies show that the first and second NLS sequences along with the first PHD are
necessary and sufficient for the nucleolar localization in HeLa cells
114
. There is some variability
among cell types, however, because endogenous and exogenous expression within HEK293T cells
is much more dispersed within the nucleus than endogenous or exogenous expression within HeLa
cells
113,114
. To date, no study has shown restricted nucleolar expression of PHF6 within cells of a
developing embryo. In fact, staining within developing brain tissue of mouse embryos shows Phf6
subcellular localization to be primarily nuclear with some cytoplasmic staining, and the nuclear
staining seems to be more closely linked to DAPI low regions, most likely euchromatin
115
. This
may signify a divergent role for PHF6 in developing tissues and in adult tissues.
Immunoprecipitation of exogenous, tandem affinity purified PHF6 in HEK293T cells
followed by mass spectrometric analysis of PHF6 binding partners showed that PHF6 has an
affinity for ribosomal biogenesis factors as well as chromatin remodeling factors
113
. It has been
speculated that PHF6 within the nucleoplasm has a primary interaction with chromatin remodeling
factors and operates primarily as a chromatin adaptor protein
113
while, in the nucleosome, PHF6
30
primarily acts to control ribosome biogenesis and splicing
113
. In support of this hypothesis, cell
fractionation showed that the PHF6 complex with 4 members of the Nucleosome Remodeling and
Deacetylation (NuRD) complex exists only in the nucleoplasm, and not the nucleolus
113
.
Additionally, studies showing the effect of PHF6 on ribosome biogenesis were primarily
performed in cell types where PHF6 has been shown to be almost exclusively nucleolar in nature,
namely, HeLa cells
114
. We therefore extend this hypothesis further and postulate that in a
developmental setting, where PHF6 is primarily expressed in the nucleoplasm and not in the
nucleolus, PHF6 likely acts to regulate chromatin as its primary function.

1.3.4. PHF6 has a defined effect on neuronal pathfinding in the developing mouse brain.
In an effort to determine the developmental origin of the intellectual disability phenotype
observed in patients, Zhang et al.
116
performed shRNA depletion of PHF6 in the cerebral cortexes
of mice at E14, and observed a reduced migration phenotype from the ventricular zone to the upper
cortical plate. They also observed an increase in the number of cells that showed a bipolar/
multipolar phenotype and abnormal branching. This suggests that PHF6 has a role in correct
migration of developing neurons. The multipolar phenotype and abnormal branching suggest that
cells depleted for PHF6 have an intrinsic defect in their pathfinding capabilities, which then may
result in the reduced ability for cells to reach the cortical plate. PHF6 therefore has a defined role
in neuronal migration, and may have a role in the migratory potential of other cell types.

Section 1.4. PHF6 and epigenetic regulation of neural crest: Thesis outline
31
As I have outlined here, there is a need to understand the mechanism by which neural crest
delamination can be controlled on an epigenetic level, to prevent the precocious formation of
neural crest which would lead to the hyperplasia of craniofacial structures. We believe that PHF6
fits this role, and here we will set out the evidence for this role as well as a prospective mechanism
for precisely how PHF6 performs this activity. In chapter 2, I will outline the methodology that I
use to determine the biochemical and developmental role of PHF6, including a detailed evaluation
of the model systems used in this work. In chapter 3, I will outline the specific expression pattern
of Phf6 in Xenopus laevis embryos and PHF6 in neural crest differentiated from human embryonic
stem cells. I will also address how the expression pattern of PHF6 could allow for a role in pre-
migratory neural crest in preventing precocious migration. In chapter 4, I will give evidence for a
cell-autonomous restrictive role that Phf6 plays in the development of neural crest in Xenopus
laevis. In chapter 5, I will outline the evidence for a biochemical role of PHF6 at neural crest-
specific enhancers to prevent precocious activation of these enhancers. In chapter 6, I will give
evidence for protein-protein interactions between PHF6 and additional complex members and give
evidence of a role for Phf6 in mesoderm development in Xenopus laevis. And finally, in chapter
7, I will evaluate what this work contributes to the field of developmental epigenetics as a whole.

32
Chapter 2: Detailed Materials and Methods.
Section 2.1 Xenopus laevis embryology
2.1.1 Xenopus laevis as a model system and the advantages it brings to the study of epigenetics
in craniofacial development
The Xenopus laevis embryological system is highly efficient at addressing a number of
developmental problems. This system has been used to high effect to determine questions of
signaling in development, the arrangement of cytoskeletal elements in collective cellular
movement
117–119
, the generation of forces within an embryo
120,121
, and even to develop ideas involving
histone proteins and their biochemical properties
122,123
. This is largely due to two main
methodological advantages of the system for experimental manipulations: microinjection and
explantation/transplantation. Microinjection of Xenopus laevis oocytes and early stage embryos
has long been used to introduce mRNA, which induces ectopic expression, and morpholinos,
antisense oligonucleotides stabilized to deplete protein levels within the embryo over a long period
of time. Microinjection has allowed scientists to target either the whole embryo or create tissue-
targeted injections, due to the fact that this model organism has a particularly well defined fate
map for each blastomere, or cell, up to the 64-cell stage
124–126
, and that this organism has particularly
large blastomeres, so that each one is differentiable by a trained eye through the early blastula
stages. Additionally, the left-right symmetry of the organism is set up at the first cell division, so
one can target specific tissues on only one side of the organism, and leave the opposing side as an
internal control. While microinjection is common amongst model organisms used for
developmental biology, Xenopus is rare in that it simultaneously allows for specific targeting of
cell types with a relatively high degree of accuracy
127
, and provides a model for vertebrate
development where most of the early developmental events are highly conserved between
33
mammals and Xenopus. In addition, this organism creates a large number of embryos in a single
day, which allows for a level of statistical power for a large number of downstream quantitative
assays that is usually only seen in in vitro cell culture assays. The other main methodological
strength of the Xenopus system lies in the ability to remove selected tissues and either culture them
externally, explantation, or replace them within a different embryological setting, transplantation.
Transplantation of selected targeted or untargeted tissues allows scientists to establish the
contributions of selected tissues on surrounding tissues, or the contributions of surrounding tissues
on selected tissues. Seminal experiments using this methodology has allowed the developmental
field to determine the effects of signaling on many developmental events
128–131
, and these experiments
are still considered the gold standard for determining competence, determination, and specification
of a tissue at points of time or due to experimental manipulation.
The specific suitability of this system to epigenetics and particularly the question of how
epigenetic regulators affect craniofacial development is apparent when we examine the details of
the Xenopus system. The large cells within the early stages and the defined fate map means that
we can introduce humanized mRNAs into specific tissues with microinjection. This allows us to
perform proteomics to study protein-protein interaction on specific tissues within an in vivo setting,
even in circumstances where no Xenopus specific antibody exists. Furthermore, the curation of the
Xenopus laevis genome over recent years
132
has allowed the probing of protein-DNA and protein-
RNA interactions, as well as downstream effects on the transcriptome of specific tissues. When
we combine these technologies with the ability to extract specific tissues from Xenopus embryos
using classic explant techniques, we have a situation where we can perform in vivo based
epigenome experiments to determine how epigenetic regulators function specifically within cell
types and tissues of interest. As these techniques become more developed in the coming years, we
34
can also begin to probe the effects that signaling has on epigenomic changes directly by using the
well-defined signaling environments of Xenopus embryos to transplant tissues between signaling
environments, thus probing how signaling events are able to contribute to specific changes within
the epigenome.
This model organism is also specifically useful for ascertaining the role of epigenetics in
craniofacial development due to the well-defined progression of cranial neural crest development
within Xenopus embryos and the ability to screen many different effects on this system. The
progression from early competence to terminal differentiation of neural crest cells has been well
established within the Xenopus system
133,134
, and manipulations that affect the neural crest
transcriptional network has been intensively studied in this system in recent years. Readouts for
manipulations of the epigenome are therefore fully available, and have been demonstrated as to
their efficacy in a number of publications modeling deficiency of CHD7 in Xenopus
29
. Furthermore,
a number of recent publications
135,136
have performed transcriptome analysis on Xenopus neural crest
models, therefore allowing a standard of comparison for epigenetic manipulation effects on the
neural crest transcriptome. The high statistical power of this model organism also allows for
screening of a large number of different constructs for the specific changes to the epigenome and
transcriptome for each patient mutation within the number of different diseases involving
epigenetic regulators and craniofacial disorders. This model system therefore not only has the
ability to probe the effects of epigenetic regulation on neural crest and craniofacial development,
but has the ability to determine the specific effects of each of the mutations in a relatively short
time frame.

35
2.1.2 Xenopus laevis embryology
Female adult Xenopus laevis organisms are kept in a temperature-regulated room in water
treated to remove chlorine and other harmful additives. Females are pre-primed up to 2 weeks
before priming with 50 units Pregnant Mare Serum Gonadotropin (PMSG) by subcutaneous
injection to the ovary region of the frog. Females are then induced to ovulate using 500 units
human Chorionic Gonadotropin (hCG) 12-18 hours before oocytes are needed. Females are gently
massaged to promote oocyte laying into a dish between 5 and 9 times during oocyte production,
and allowed to rest at least an hour and a half between laying. After a day of oocyte production,
females are allowed to rest 3 months before a new cycle of priming begins.
Oocytes within a dish are incubated with 1-1.5 mm of male testes, dounced loosely 2 times
in 0.3x Marc’s Modified Ringer (MMR, at 1x concentration 0.1M NaCl, 2 mM KCl, 1 mM MgSO 4,
2 mM CaCl 2, 5 mM HEPES, pH 7.8, 0.1 mM EDTA), for 30-45 minutes. Fertilized embryos are
then de-jellied using a 2% L-cysteine solution in 0.3x MMR, pH normalized to 7.5 for 5 minutes,
on gentle rotation, then rinsed thoroughly using 0.3x MMR and placed into a new dish. Embryos
are then allowed to develop at a temperature of 16
o
C-25
o
C, depending on staging and time frame
needed, according to temperature dependent development charts.

2.1.3 Transcription of mRNA for microinjection
Transcription is performed with the assistance of the mMessage mMachine in vitro
transcription reactions from Ambion. The appropriate mRNAs are cloned into the pCS2+
backbone, which has several RNA-polymerase specific promoters which can be utilized to
transcribe the RNA. These plasmids are then prepared to high concentration (>1 µg/µl) and high
purity to ensure the transcription reaction is effective. Plasmids are then linearized using the
36
appropriate restriction enzyme with a 5’ overhang that is downstream of the intended transcript.
Some evidence suggests that a 3’ overhang may interfere with processivity of the RNA
polymerase. The linearized plasmid is then run on an agarose gel in order to size-select only for
the linearized plasmid and verify complete cutting of the plasmid; the linearized band is cut out
and purified. This is then used as a template for the transcription reaction. The transcription
reaction is assembled according to manufacturer specifications, and run 2 hours at 37
o
C for T7 or
T3 polymerase, or 4 hours for SP6 polymerase. The resulting reaction is then precipitated using
LiCl, washed with RNase-free isopropanol and 80% ethanol, and resuspended in Ultrapure water,
10-20 µl depending on the size of the pellet after precipitation. The product is then aliquoted in 2
µl aliquots and frozen at -80
o
C. One aliquot is held back and checked for concentration on a
nanodrop machine and for RNA integrity using a denaturing agarose gel.
Briefly, 1.5 g of Agarose is melted into 72 ml of autoclaved water. This is taken to a fume
hood where 10 ml of 10x MOPS buffer (200 mM MOPS, pH 7.0, 10 mM EDTA, 50 mM NaOAc)
and 18 ml of 37% formaldehyde is added, and the gel is immediately poured into a tray and
covered. This is left for 30 minutes to harden. Meanwhile, RNA samples are denatured alongside
of an RNA ladder (Sigma) by diluting sample to 5 µl using ultrapure water, adding 5 µl Ambion
Formaldehyde load dye, and Ethidium Bromide to 10 µg/ml. These samples are then heated at
70
o
C for 15 minutes. The gel is submerged into 1x MOPS buffer, the samples and ladder is loaded,
and samples are run at 50 V until the bromophenol blue has migrated 2/3 of the way down the
gel.

37
2.1.4 Microinjection of Xenopus embryos
Dejellied embryos are placed in a 4% Ficoll in 0.3x MMR solution 10-15 minutes ahead
of injection. RNase-free microcapillary tubes are pulled using a micropipette pulling machine from
Sutter Instruments using 1 x 0.75 capillary glass with a pull value of 30, velocity of 120, time of
200, pressure of 250. Tips are then broken off to 3-5 µm diameter tip. These are then calibrated to
estimate the time amount in milliseconds (ms) to get 1 nl injection by taking up 1 µl of liquid and
dispensing it in 100 ms increments; the number of increments, n, into 1000 yields number of nl
per 100 ms increment, 100 ms divided by number of nl per increment yields ms/1nl. See table 2.1
for dilutions into the various blastomeres per injected volume. Embryos are then allowed to rest in
4% Ficoll for 30 minutes after injection, and transferred to 0.3x MMR.

2.1.5 Neural crest transplantation
Danilchick’s media modified for Amy (DFA, 53 mM NaCl, 5 mM Na 2CO 3, 4.5 mM
Potassium Gluconate, 32 mM Sodium Gluconate, 1 mM CaCl 2, 1 mM MgSO 4, pH 8.3, 0.1% BSA)
is prepared, autoclaved, and stored at -80
o
C. This media is thawed immediately before use and
brought to room temperature, where antibiotic/antimicotic is added at 1:1000 dilution. Embryos
are placed in a clay dish in DFA at NF stage 14 and allowed to develop to NF 15-17. WT embryos
are stage matched and paired with their microinjected counterparts, ideally from the identical
fertilization clutch. Using a hair loop and an eyebrow knife, the apex of the neural ridge
immediately posterior to the eye bulge and just lateral to that point is lifted using a shallow cut that
does not disturb the underlying, lighter colored mesoderm. Two cuts are made, one behind the eye
bulge, and one posterior to this but parallel, at the apex of the curvature of the embryo (Fig. 2.1).
A lateral cut is then made, connecting these two cuts, and one at the apex of the neural ridge. This
38
Table 2.1 Dilution factors associated with volume injections into blastomere at each cell division.
1 cell 1/2 cell 1/4 cell 1/8 cell 1/16 cell
1 nl 1:1000 1:500 1:250 1:125 1:62.5
2 nl 1:500 1:250 1:125 1:62.5 -
5 nl 1:200 1:100 1:50 - -
6 nl 1:167 1:83 1:42 - -
8 nl 1:125 1:62.5 - - -
10 nl 1:100 1:50 - - -

39
Figure 2.1. Schematic representation of the transplanted or explanted region of the embryo,
outlined in the black box, on a lateral view Xenopus laevis embryo at NF16. It should be noted that
the neural crest is only several cell thickness deep, and care should be taken to remove the neural
crest away from the underlying mesoderm.

NF16
A P
40
tissue is then set aside, with care taken to note the anterior region versus the posterior, as the
corresponding region is lifted out of the stage-matched partner embryo. This tissue is then placed
in the corresponding region hole, and a small piece of cover glass is gently placed over the embryo
to provide pressure and aid in healing. The embryo is allowed to heal in this position for 30
minutes, and then transferred to 0.3x MMR.

2.1.6 Neural crest explant and culture
Cover glass is coated with 100 nM fibronectin in phosphate buffered saline (PBS)
overnight (Alfandari et al., 2003). The PBS from these coverslips are aspirated, imaging chambers
are placed over the coverslips and filled with DFA medium. Cranial neural crest cells are dissected
out of the embryo, pipetted up carefully, and allowed to settle onto the coverslip, positioning so
that the tissue that was adjacent to the mesoderm is now adjacent to the fibronectin. These explants
are allowed to heal for 20 minutes, and are then either filmed continuously using phase contrast or
fluorescent microscopy, or allowed to develop for 15-24 hours, fixed, and stained using phalloidin.

2.1.7 Tissue fixation from Xenopus embryos
Embryos are allowed to develop to the appropriate stage. They are then fixed by
submerging into MEMFA solution (100 mM MOPS, pH 7.4, 2 mM EGTA, 1 mM MgSO 4, 3.7%
V/V formaldehyde) for 1-2 hours at room temperature or overnight at 4
o
C. They are then rinsed 5
times in 100% ethanol and stored at -20
o
C.

41
2.1.8 In situ hybridization probe preparation and testing
Probes are prepared first by linearization of the probe plasmid followed by in vitro
transcription using the appropriate polymerase, T7 or SP6. Probe linearization enzymes and
polymerases can be found in Table 2.1. Probes are transcribed using the Ambien Megascript kit of
the appropriate polymerase, using the Roche dig-UTP mix as the nucleotide source. Probes are
transcribed for 2-4 hours, the resultant reaction mix is measured for nucleotide concentration, and
the reaction mix is combined 1:3 with hybridization buffer to eliminate RNase activity.
Probe reactivity is tested using a modified Northern blot protocol. Probe plasmids, one
specific and one non-specific, are dotted on a piece of N+-hybond paper starting with 10 ng in 10-
fold serial dilutions, and allowed to dry completely. The DNA is then denatured by placing the
membrane onto Watman filter paper soaked in a 0.5M NaOH, 1.5 M NaCl solution for 5 minutes.
The membrane is then blotted on dry filter paper and placed onto filter paper saturated with 2x
saline-sodium citrate (SSC, 300 mM NaCl, 30 mM Na 3Citrate, pH 7) buffer and incubated for 5
minutes. This is repeated, and the membrane is placed into hybridization buffer (50% formamide,
5x SSC, 1 mg/ml Torula RNA, 100 µg/ml Heparin, 0.1% Tween-20, 0.1% CHAPS, 10 mM EDTA,
2% BMB) to block for 1 hour. The membrane is then incubated in hybridization buffer + 0.5 µg/ml
probe overnight at 65
o
C. The membrane is then washed to remove non-specific interactions of the
probe with the following solutions: 5 minutes in 2x SSC, 5 minutes in 2x SSC, 5 minutes in 2x
SSC + 0.1% SDS, 5 minutes in 2x SSC + 0.5% SDS, 5 minutes in 2x SSC + 1% SDS, 5 minutes
in 1x SSC + 1% SDS, 5 minutes in 0.2x SSC + 1% SDS, 5 minutes in 0.2x SSC + 1% SDS. The
membrane is then incubated in 2% w/v Boehringer-Manneheim Block (BMB) in Maleic Acid
Buffer (MAB, 100 mM Maleic Acid, 150 mM NaCl) for 30 minutes, then 2% BMB in MAB +
1:3000 Fab anti-digoxygenin antibody for 30 minutes. The membrane is then washed in Tris-base
42
saline with Tween (TBST), 3 times for 10 minutes each. The membrane is washed with Alkaline
phosphatase buffer (100 mM Tris, pH 9.5, 50 mM MgCl 2, 100 mM NaCl, 0.1% Tween-20, 2.5 mM
Levasimol) for 10 minutes, and incubated with BM purple substrate until color develops.
Specificity is determined by analyzing the concentration at which the specific plasmid is detected
by the color reaction versus the concentration at which the non-specific plasmid is detected by the
color reaction. Probe preparations that are not at least 100x more sensitive to the specific plasmid
are discarded.

2.1.9 Whole mount in situ hybridization
RNase-free conditions are crucial for the first day of in situ hybridization, so care is taken
to make all first day reagents fresh with RNase-free water (DEPC treated or Ultrapure water).
Following probe incubation, reagents can be used from previous experiments or made fresh using
deionized or double-distilled water; RNase-free conditions are not necessary.
Fixed embryos are rehydrated stepwise by washes in 100% ethanol, 75%
ethanol/25%water, 50% ethanol/50% water, 25% ethanol/75% PBS + 1% Tween-20 (PTw),
followed by 3 washes in 100% PTw. The embryo is then permeabilized by a 12-minute incubation
with PTw + 10 µg/ml Proteinase K. In order to denature secondary structure of mRNAs within the
embryo, the embryos are washed twice with 0.1 M triethanolamine (TEA), followed by two washes
of 0.1 M TEA + 2.5 µl acetic anhydride / ml TEA. Excess acetic anhydride is washed away with
two washes in PTw, and the embryo is re-fixed with a 5-minute wash of 4% PFA in PTw, followed
by quenching using 0.1 M glycine in PTw for 5 minutes. Excess glycine and PFA are washed away
with 5 washes in PTw, and then embryos are equilibrated in hybridization buffer. Xenopus laevis
embryos, which are large and have a high yolk content, then require at least 6 hours in
43
hybridization buffer at 60
o
C to ensure complete penetration of the buffer and blocking. The
embryos are then incubated overnight at 60
o
C in hybridization buffer with 1 µg/ml of the
appropriate probe, at least 12 hours.
The next day, excess probe is washed away with 2 washes in hybridization buffer, followed
by 5 stringency washes in 2x SSC at 60
o
C for 20 minutes. Excess non-hybridized probe is then
digested away using RNase A and T1 in 2x SSC at 37
o
C for 30 minutes, followed by 1 wash in 2x
SSC to remove excess RNase. Stringency washes are then continued with 0.2x SSC for 30 minutes
at 60
o
C, followed by 3 washes in Maleic Acid Buffer (MAB) at room temperature for 10 minutes.
Embryos are then blocked against excessive antibody binding using 2% Boehringer-Manneheim
Block (BMB, Roche) in MAB for 2 hours at room temperature, followed by an overnight
incubation of 2% BMB in MAB with 1/3000 dilution of Fab anti-digoxygenin antibody (Roche)
at 4
o
C.
The following day, embryos are washed 5 times, 1 hour each, in MAB at room temperature.
They are then equilibrated to high pH to facilitate the color reaction using an Alkaline Phosphatase
buffer (made fresh, see above), two washes for 5 minutes each. Embryos are then incubated with
the color reaction substrate, BM purple (Roche), at room temperature or 4
o
C until color develops
to the appropriate level, depending on the probe used. Table 2.2 indicates the approximate times
for color to develop using these exact conditions and probe preparations; expect variability if
conditions or probe preparations are varied.
To stop the color reaction, embryos are washed twice in MAB. In order to fix the color, the
embryos are then incubated in Bouin’s fixative (70 ml saturated Picric acid, 5 ml glacial acetic
acid, 25 ml 37% formaldehyde), with care taken to prevent spilling this fixative, for 2-4 hours.
Embryos are then washed in 70% ethanol buffered with 0.2x SSC in continuous washes until
44

Table 2.2 Approximate times for color to develop for the probes used in this thesis.
Probe Time at room temperature Time at 4
o
C
Phf6a/Pan-Phf6 5-6 hours 2 days
Twist 1-2 hours Overnight
Sox2 2-4 hours 1 day
Slug 2-3 hours 1 day
Sox9 1-2 hours Overnight

45
embryos return to white from the yellow Bouin’s fixative. Embryos are then progressively
rehydrated using 75% buffered ethanol/25% 1x SSC, 50% buffered ethanol/50% 1x SSC, 25%
buffered ethanol/75% 1x SSC, and 2 washes of 1x SSC. In order to bleach the pigmentation of the
embryos to image the staining (can be omitted if albino embryos are used), embryos are incubated
with Bleaching solution (1.25 ml 20x SSC, 2.5 ml formamide, 2 ml 30% hydrogen peroxide, 45
ml distilled water, made fresh) for 1-2 hours under a lamp until pigmentation is removed. Embryos
are then washed twice in 1x SSC, then imaged in 1x SSC using an 1% agarose in SSC dish.

2.1.10 Phalloidin staining
In this work, phalloidin staining was used following explantation or rough sectioning of
xenopus embryos. For rough sectioning, following fixation in MEMFA, embryos were dehydrated
using 100% ethanol, followed by rehydration into PBS. Embryos were then sectioned using a razor
blade using microscopy to guide the sectioning, making dorsal/ventral cuts along the most anterior
portion of the embryo to expose the migration pattern of the cranial neural crest. Phalloidin was
then used to distinguish morphology within the section, in order to differentiate tissues under
confocal microscopy. Phalloidin staining was used on explants to distinguish neural crest actin
morphologies after depletion or ectopic expression of PHF6 within embryos; explants were fixed
using MEMFA and gently washed several times using PBS.
Phalloidin stain is rehydrated in TBST, 10 µl phalloidin in 500 µl of TBST. Sufficient
phalloidin solution is added to cover the embryo sections or explants after removal of the PBS,
and this solution is left for 1 hour at room temperature. This solution is then removed, and the
sections or explants are either stained with a 5 µg/ml solution of Hoescht in PBS for 5 minutes, or
simply covered in PBS and imaged using confocal microscopy.
46
Section 2.2. Human neural crest cell differentiation
2.2.1 Neural crest derived from human embryonic stem cells and the advantages this system brings
for modeling biochemical processes in a developmental context
The use of human embryonic stem cells in cell culture systems has opened up numerous
advantages for researchers over the years. Human embryonic stem cells (hESC) allow modeling
of human developmental processes, drug discovery and testing of drug efficacy in a relevant
human system, and discovery of biochemical processes in a developmental setting, among the vast
applications of this technology.
We have developed a methodology for differentiating human embryonic stem cells to
neural crest cells in a way that mimics the normal developmental process
29,92
. Cells are differentiated
from the human embryonic stem cell state to a self-adherent neural ectodermal sphere
intermediate; these cell spheres then settle onto the culture dish through contacts formed by early
delaminating cells, and neural crest cells migrate out. The phenotypic differences between the self-
adherent neural ectoderm and the dish-adherent, extracellular matrix secreting neural crest cells
make these two populations easy to separate physically for separate for downstream analyses. Both
the neural crest cells and the neural ectodermal cells formed from this system mimic the expression
of the appropriate markers. Neural ectodermal cells express Sox2, Nestin, N-cadherin, and other
neural ectodermal cell markers, while neural crest cells express Sox10, Sox9, AP2a, Twist, and
other neural crest markers (data not shown, performed by Ruchi Bajpai).
The advantages of this technology are that we can perform biochemical analyses, test the
effects of genetic perturbations to the system in a way that we can assay the direct functionality of
these changes, and that we can ultimately perform drug screening on a highly relevant human
neural crest system. The ability to culture these cells on a large scale makes feasible analyses that
47
require a large amount of starting material, such as ChIP-seq, Mass spectrometry, Chromatin
confirmation capture and the associated high throughput versions, as well as protein-RNA
interaction cataloging with technologies like RIP-seq. This culture technique also allows for a
reproducible way to perform smaller scale biochemical analyses, such as assaying direct protein
interactions using co-immunoprecipitation, native protein complex analysis, or protein purification
and structural analysis. We have also had success in developing assays to test protein function.
These cells have been used to determine CHD7 binding to certain human neural crest enhancers
92
,
and we can then test the efficiency of mutant CHD7 binding to these loci through the creation of
patient induced pluripotent cell lines and use in the same assays. Similar assays can be used to
determine the functionality of patient mutations seen in Chapter 1, and therefore can be a highly
efficient tool for assaying the direct effect of patient mutations on biochemical activity.
Furthermore, the ability to use these assays on induced pluripotent cell lines means that we can
perform eventual drug screening to determine what small molecules can help reverse the effects
of these genetic mutations in order to directly help patients.

2.2.2 Human Embryonic Stem Cell Culture
Human embryonic stem cells must be grown at sub-confluency with consistent daily media
changes in order to prevent starvation. Stem cells used here are grown on Geltrex extracellular
matrix in feeder-free conditions and using mTeSR1 complete media without serum. High levels of
care must be taken to ensure proper Geltrex coating prior to cell plating; Geltrex is diluted 1:90
using serum-free, additive free DMEM and coated on 6 well plates overnight at 4
o
C. Plates are
removed and allowed to equilibrate to room temperature 1-2 hours prior to plating, after which
DMEM is aspirated and replaced with mTeSR. Cells are passaged using Accutase enzymatic
48
passage; cells are rinsed twice with PBS, and enough Accutase is added to cover cells. Accutase
is monitored carefully over the next 2-5 minutes, and when cells along the edges have extensive
separation between groups of cells, Accutase is removed and cells are rinsed briefly twice to
remove any single cells using PBS. mTeSR is added to cover cells, and cells are then scraped using
the tip of a 2 ml pipette. Cells are then distributed with a 1:6 to 1:12 split into new wells with
Geltrex and mTeSR. Passaging must be performed every 5-6 days to maintain sub-confluency.

2.2.3 Neural Crest Culture
Neural crest cells are differentiated using 7-14 days of treatment using a neural crest
differentiation medium, depending on the quantity and maturity of the neural crest cells desired.
The neural crest differentiation media consists of 50% DMEM-F12, 50% Neurobasal media,
supplemented with Gem21, N2, bFGF, EGF, and Insulin. Human embryonic stem cells are placed
into a 2 mg/ml Collagenase IV solution in DMEM (filter sterilized) and incubated at 37
o
C for 30
minutes to 1 hour to allow for most of the cells within a cell cluster to lift off the plate. Cells are
then washed 3-4 times using excess PBS to remove any trace of Collagenase, and clusters are
broken up either by pipetting up and down for completely released cells, or by scraping using a 2
ml pipette for cells still on the plate. Cells are then placed into neural crest media and incubated at
37
o
C overnight. Media is replaced the next day to remove any debris from the passaging process
by collecting media and allowing cells to settle for 15-20 minutes. Cells are rinsed once in PBS,
allowed to settle again, then resuspended in neural crest differentiation media and added to a new
plate. Approximately 3-6 wells of human embryonic stem cells are added to one 15 cm plate during
the differentiation process. Equivalent density should be aimed for between differentiation trials;
cell clusters should make up about 15% of the total dish surface when settled. At this density, cell
49
clusters should become adherent on day 6 of differentiation, and migratory cells should appear at
day 7. If cell clusters are small, <50 cells per cluster, all neural ectodermal cells should become
neural crest by day 10; larger cell clusters, >200 cells per cluster, will never entirely dissolve into
neural crest as neural ectodermal cells will continue to proliferate throughout the course of the
differentiation and thus they will continuously self-replace if the clusters are too large. Neural crest
cells will begin migrating as compact cells with multiple lamellipodial projections, and will
radially migrate from the central cell cluster to become more elongated and fibroblast-like in
appearance. Mature neural crest cells can be harvested using trypsin re-plated onto fibronectin and
continue to proliferate. We define these as cranial mesenchymal stem cells; these cells can be
further differentiated into osteoblasts, chondrocytes, or neurons using defined differentiation
conditions.

2.2.4 Cellular Fractionation
In order to perform immunoprecipitation on endogenous proteins, we perform a series of
extraction procedures to specifically separate those proteins that are innately localized in various
cellular compartments. Single cells are carefully isolated; Neural ectodermal clusters are incubated
with Accutase for 10 minutes prior to gentle centrifugation (10 minutes, 500 RPM) in order to gain
single cells. In lieu of counting cells to estimate numbers, packed cell volume (PCV) is estimated
and used as a standard to add the appropriate amount of buffers downstream. 5x PCV of Dignam
A buffer (10 mM Hepes-KOH, pH 7.9, 1.5 mM MgCl 2, 10 mM KCl, 1 mM DTT, 1 mM PMSF,
1x Complete protease inhibitor cocktail) is added to cells, and cells are incubated on ice for 10
minutes. Cells are centrifuged at 2000 RPM in a large tissue culture centrifuge at 4
o
C for 5 minutes,
and the supernatant is discarded. 2.5x PCV of Dignam A buffer is added to the cell pellet and cells
50
are resuspended, followed by tight douncing in a cold dounce 20 times. This solution is transferred
to a cold 15 ml conical tube and centrifuged at 2500 RPM at 4
o
C for 10 minutes. The supernatant
is preserved as the cytoplasmic extract. 1x PCV of a high salt extraction buffer (100 mM Tris-HCl,
pH 8, 300 mM NaCl, 0.1% NP-40, 10% Glycerol, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 1 mM
PMSF, 1x Complete Protease Inhibitor cocktail) is added to the nuclei and resuspended gently;
nuclei are then added to a 1.5 ml Eppendorf tube, and incubated on a rotator at 4
o
C for 30 minutes.
This solution is then centrifuged at maximum speed on a desktop centrifuge (>12,000xg) for 15
minutes at 4
o
C. Supernatant is removed, and salts are normalized to immunoprecipitation level
using no salt extraction buffer (100 mM Tris-HCl, pH 8, 0.1% NP-40, 10% Glycerol, 1 mM EDTA,
1 mM EGTA, 1 mM DTT, 1 mM PMSF, 1x Complete Protease Inhibitor cocktail), usually
equivalent volume to extracted nuclear extract supernatant to yield 150 mM NaCl. This salt level
is highly dependent on the antibody used to precipitate, and should be determined individually.

2.2.5 Immunoprecipitation
Immunoprecipitations involved in this work were all performed using a nuclear extract of
the cells in question. Care must be taken to measure the protein level in the nuclear extract using
a BSA assay in order to accurately set up immunoprecipitation experiments. The necessary
quantity of protein per µg of antibody used can vary greatly depending on the antibody, as can the
salt level of the interaction reaction and washes, the time of incubation, the amount of magnetic
beads required to precipitate the antibody, and the number and stringency of washes. Each of these
values was determined and noted for each immunoprecipitation antibody, and a non-specific rabbit
or mouse IgG antibody was used as a negative control for each immunoprecipitation to account
for non-specific binding in each condition. The general protocol is that a specific amount of protein
51
is incubated with the antibody for a designated time, on a rotation machine at 4
o
C. After this period,
magnetic Protein A and Protein G beads are added to the immunoprecipitation, and the reaction is
returned to the rotation machine at 4
o
C for 2 hours. The reaction is then placed on a magnet, and
the supernatant is removed and kept as a depleted fraction; the level of depletion within this
fraction contributes to the efficacy of the assay. The reaction is then washed a set number of times
either with agitation or not for a set period of time per wash using a solution of high salt extraction
buffer (see Section 2.2.4) mixed with no salt extraction buffer to gain a specific concentration of
salt. The reaction is placed on the magnet and the wash buffer is removed after each wash. The
immunoprecipitation reaction is then eluted using a base elution buffer (50 mM NH 4, 50 mM NaCl)
on ice for 15 minutes, resuspending the beads every 3 minutes. The elution is removed from the
beads after pull down with the magnet and equilibrated to neutral pH with equivalent amount of
neutralization buffer (100 mM Tris-HCl, pH 7, 150 mM NaCl). The remaining beads are incubated
with 1x Laemmli loading buffer at 70
o
C for complete elution of remaining protein.

2.2.6 Chromatin Immunoprecipitation
Chromatin Immunoprecipitation is performed to assay genomic binding of chromatin
modifiers, transcription factors, and the deposition of histone post-translational modifications.
Here, we used this technology to assay PHF6 binding and to track the histone modifications
associated with PHF6 binding.
Cells are carefully prepared to be in single-cell suspensions in the case of neural ectodermal
cells by treatment with Accutase for 10 minutes followed by gentle centrifugation (500 RPM, 10
minutes) and resuspension in neural crest differentiation media. In the case of neural crest cells,
neural ectodermal cell clusters are gently rinsed off the plate, and cells are covered with neural
52
crest differentiation media. A buffered formaldehyde solution (50 mM Hepes-KOH, pH 7.5, 100
mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 11% formaldehyde) is then added at 1:10 dilution, and
cells are gently agitated for 13 minutes. A 2.5 M glycine solution is then added to the crosslinking
mix at a 1:20 dilution and cells are left to gently agitate for an additional 5 minutes. In the case of
neural ectodermal cells (NEC), cells are then centrifuged at 1350xg, 4
o
C, 5 minutes, and
resuspended in progressive cold PBS washes until cell pellet is completely white in color. For
neural crest cells (NCC), the crosslinking mix is removed, and the cells are washed twice gently
in cold PBS. Minimal PBS supplemented with 1 mM PMSF and 0.1% BSA is then added to the
plate, and cells are scraped and collected. Cells are then centrifuged as above until the cell pellet
is completely white in color.
Cell volume is important at this point, as ~10
8
NEC or NCC cells amount to approximately
1 ml packed cell volume (PCV). Cells are resuspended in 5x PCV of Lysis Buffer 1 (LB1, 50 mM
Hepes-KOH, pH 7.5, 140 mM NaCl, 1 mM EDTA, 10% glycerol, 0.5% NP-40, 0.25% Triton X-
100) and gently agitated on a rotator at 4
o
C for 10 minutes. Cells are then centrifuged at 1350xg,
4
o
C, 5 minutes, and the supernatant is removed. Pellets are then resuspended in 5x PCV in Lysis
Buffer 2 (10 mM Tris-HCl, pH 8.0, 200 mM NaCl, 1 mM EDTA, 0.5 mM EGTA) and agitated
gently on a rotator at room temperature for 10 minutes. Nuclei are pelleted at 2500xg for 5 minutes
at 4
o
, and the supernatant is removed. Nuclei are then resuspended in 3x PCV Lysis Buffer 3 (10
mM Tris-HCl, pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 0.1% Na-Deoxycholate,
0.5% N-lauroylsarcosine) and aliquoted in polystyrene 15 ml conical tubes, 1.5 ml of nuclei per
tube, and flash frozen, storing at -80
o
C.
Chromatin sonication is then performed on these samples; this is one of the more crucial
steps in this method, and care is needed to prepare chromatin effectively. Samples should be
53
thawed on ice. Here, I used the Bioruptor machinery to sonicate samples; if other machinery is
used then sonication should be optimized for this machinery. Sonication probes are inserted into
the polystyrene tubes and tubes are placed in the wheel holder and placed in the machine. Cold
water should be added to completely submerge the chromatin sample. Chromatin is then sonicated
for 30 seconds on, 60 seconds off on the medium setting for approximately 45 minutes, with 10 µl
samples taken every 10-15 minutes to ensure correct sonication. This time will vary from sample
to sample, so each chromatin preparation should be tested individually. To test the sonication, the
sample should be taken aside, and 40 µl ChIP elution buffer added (50 mM Tris-HCl, pH 8.0, 10
mM EDTA, 1% SDS). Debris is then centrifuged out using a benchtop centrifuge at maximum
speed for 10 minutes. 1 µl of a 10 mg/ml RNase A solution is then added to the separated
supernatant, and the sample is left at room temperature for 10 minutes. 12 µl of 5M NaCl is then
added to assist the reverse crosslink, and the sample is boiled for 15 minutes. DNA is then isolated
using a standard PCR clean up. This DNA is then run on a 1.5% agarose gel. The DNA should
appear to be a smear of variable lengths, but two important points should be noted: first, optimal
chromatin should not have any DNA that is over 2 kilobases in length, as this will increase the
background of any chromatin immunoprecipitation, and second, the majority of the DNA should
be approximately 200-300 bases in length, though oversonicated chromatin (majority 150-200
bases) will be sufficient to assay histone post-translational modifications as this corresponds to
mono-nucleosomal DNA.
Once chromatin is sonicated appropriately, all chromatin from a sample is pooled and 10%
Triton X-100 is added to a final concentration of 1%. The sample is then spun at 20,000xg to pellet
debris from the sonication process. Immunoprecipitation is then set up; 250-1000 µl of chromatin
is used per immunoprecipitation, 3-5 µg of antibody. PHF6: 500 µl chromatin to 3 µg of antibody,
54
CHD7: 1000 µl of chromatin to 5 µg of antibody, histone post-translational modifications: 250 µl
of chromatin to 3 µg of antibody. This is then incubated at 4
o
C overnight on a rotatator. The next
morning, Protein A and Protein G beads are used to precipitate the antibody. For 3 µg of antibody,
25 µl of each is used, for 5 µg of antibody, 50 µl of each. Beads are placed into a new Eppendorf
tube and washed twice with PBS supplemented with 0.5% w/v BSA for 5 minutes each. The
immunoprecipitation reaction is then transferred into the tube with the beads, and the reaction is
incubated at 4
o
C for 4 hours, rotating. The reaction is then placed on a magnetic stand, and the
supernatant is removed and kept as a depleted fraction. This fraction can be used to test the
efficiency of the immunoprecipitation using western blot. The beads are then resuspended in a
RIPA wash buffer (50 mM Hepes-KOH, pH 7.5, 500 mM LiCl, 1 mM EDTA, 1% NP-40, 0.7%
Na-deoxycholate) and transferred to a new, clean tube to eliminate background based on DNA
bound to the eppendorf tube. The beads are then washed a number of times, depending on the
observed background from each antibody. Histone post-translational modification ChIPs are
usually washed 7 times briefly, PHF6 ChIP is washed 10 times, 5 minutes each at 4
o
C, and CHD7
ChIP is washed 4 times briefly. The reaction is then washed once briefly with a modified TE buffer
(10 mM Tris-HCl, pH 8, 1 mM EDTA, 50 mM NaCl), which is aspirated twice to ensure complete
removal before elution. 210 µl of elution buffer (50 mM Tris-HCl, pH 8.0, 10 mM EDTA, 1%
SDS) is added to the beads at room temperature, and the mixture is incubated on a 65
o
C shaking
platform for 15 minutes. The elution is then gently centrifuged, and pulled down using a magnetic
stand. The elution supernatant is removed and transferred to a new tube, 8 µl 5M NaCl is added
and the eluate is incubated at 65
o
C overnight to reverse the crosslink. At this point, a 10% input is
prepared by taking 10% of the original IP material, diluting to 200 µl with elution buffer, adding
8 µl 5M NaCl, and incubated with the immunoprecipitations at 65
o
C overnight to reverse crosslink.
55
The next day, 200 µl of TE buffer is added to each IP to dilute the SDS in the solution. 8
µl of 10 mg/ml RNase A is then added to remove RNA, and the solutions are incubated at 37
o
C for
2 hours. 4 µl of 20 mg/ml Proteinase K is then added to remove protein from the solution, and
solutions are incubated at 55
o
C for 2 hours. DNA is then precipitated using phenol chloroform
precipitation, and resuspended in 10 mM Tris-HCl, pH 8. The resultant DNA is then used as a
template for quantitative PCR for genomic loci of interest. Primers should be designed to amplify
a target of 75-150 bases in length, have only one target, and should have an annealing temperature
of 60
o
C to allow multiple PCR reactions to be run on the same plate. Primers used for this work
are listed in Appendix 1.

2.2.7 Histone preparation
Here, histone extraction is adapted from Shechter et al., Nature Protocols, 2007
137
. I used
both acid extraction as well as high salt extraction followed by low molecular weight cut off
dialysis to normalize salt levels as described in this paper. These isolated histones were then used
downstream to determine the affinity of PHF6, recombinant and endogenous, for various histone
post-translational modifications using immunoprecipitation or GST precipitation.

2.2.8 Lentiviral preparation and infection
Lentiviruses are used to transduce a short hairpin RNA or coding cDNA into embryonic
stem cells in order to influence the directed differentiation of neural crest. The TRIPZ system is
used to create inducible depletion via shRNA or inducible overexpression of specific genes. The
appropriate shRNA or cDNA is first cloned into this vector, with care taken that the overall cloned
transcript is not longer than 2 kilobases in length. This vector is then prepared to high concentration
56
and purity in endonuclease-free conditions, along with lentiviral packaging plasmids pCMV-
dR8.74, which contains a number of necessary enzymes for packaging, and VSVG plasmid, which
encodes the capsid protein. These plasmids are then transfected into the human cell line HEK293T
that has been passaged rapidly at 1:3 or 1:4 splits for 2 passages to ensure rapidly growing, non-
contact-inhibited cells. Plasmid solution is created using a ratio of 5 µg viral vector, 1 µg pCMV-
dR8.74, 1 µg VSVG in 250-1000 µl serum-free DMEM, depending on the size of culture dish. 10
µg of viral vector is used for a 10 cm plate, and 30 µg of viral vector is used for a 15 cm plate. PEI
(1.5 µg/µl) is added to the equivalent volume of DMEM in a separate tube at 2 µl per µg DNA to
be transvected. These mixes are allowed to sit separately for 15 minutes, then combined and
vortexed for 1 minute and incubated for 30 minutes. During this time, 293T cells are trypsinized
and passaged 1:2 or 1:3 into a new plate, and the lentiviral mix is dropped onto the cells
immediately after plating, swirling while adding. Cells are incubated at 37
o
C overnight, 24 hours
total. After 24 hours, DMEM is replaced with serum-free Ultraculture media supplemented with
2x Glutamax. 30 hours later, this media is removed and replaced with additional Ultraculture
media, and the removed media, which now contains viral particles, is centrifuged at 2500 RPM for
10 minutes to remove debris, then supernatant is centrifuged in an ultracentrifuge at 12,500 RPM,
at 4
o
for 24 hours. This procedure is repeated 24 hours later, with the depleted supernatant from
the ultracentrifuge bleached and discarded. 24 hours later, remaining ultraculture media on the
cells is removed, cells are bleached and discarded, and this media is added to the ultracentrifuge
and centrifuged at 19,000 RPM, 4
o
C, 8 hours. After 8 hours, all remaining media in the
ultracentrifuge tubes is discarded by dumping, and any remaining media in the tubes is sealed in
the casings and placed on ice with gentle agitation overnight to resuspend the virus in small
volume. The next morning, human embryonic stem cells are split without scraping by long
57
exposure to accutase followed by rinsing with PBS to release the cells. The cells are then split 6
ways and incubated with the virus in polystyrene FACS tubes with 300 µl mTESR and 10 µg/ml
Polybrene at 37
o
C for 1 hour. Cells are then plated onto a Geltrex coated plate.

2.2.9 FACS sorting
Although the TRIPZ system incorporates a puromycin selection cassette, human
embryonic stem cells are known to progressively silence transgenes in culture
138
, and we therefore
use FACS sorting to enrich high expressing populations of human embryonic stem cells. Cells are
treated with doxycycline to induce expression of the transgene, which is usually cloned with a RFP
or GFP reporter. These cells are then passaged to single cell suspension with Accutase for 10-15
minutes, followed by scraping cells into PBS. Cells are then added to a polystyrene FACS tube
with several drops of mTESR 5x supplement to maintain the cells. Cells are then sorted by GFP
or RFP expression as appropriate, and positive and negative populations are sorted into 0.5 ml
mTESR 5x supplement. These cells are then centrifuged for 3 minutes at 800 RPM, and the
supernatant is removed and replaced with 100 µl geltrex and 100 µl mTESR media, allowed to sit
at room temperature for 30 minutes to reform clusters, and added to a 96 well plate. Cells are then
grown up and passaged.

2.2.10 Immunofluorescence
Immunofluorescence is used to determine the protein expression and cellular localization
of selected proteins of interest during the process of neural crest differentiation. In order to perform
immunofluorescence, neural ectodermal cells of day 6 or later are plated into wells with fibronectin
covered chamber wells or cover glass. These cells are then allowed to progress to the appropriate
58
stage, whereby media is removed and cells are gently washed and fixed with 4% paraformaldehyde
in PBS on ice for 20-30 minutes. Cells are then washed once with PBS, and blocked with a 2.5%
BSA solution in PBS with 0.1% TritonX-100 for 1 hour at room temperature. The protein-specific
antibody is then diluted in the blocking solution at 1:100 to 1:500 dilution, depending on the
antibody, and incubated overnight at 4
o
C. Most primary antibodies used in this study were used at
a 1:100 dilution. The antibody is then removed, and cells are washed 3 times without agitation
using blocking solution. The species-specific antibody is then diluted 1:500 in blocking solution,
and incubated for 45 minutes at room temperature. The cells are then washed 3 times without
agitation in PBS, followed by 1 µg/ml Hoechst 33342 solution in PBS for 15 minutes at room
temperature. Cells are then washed once more and checked for fluorescence, followed by confocal
imaging.

Section 2.3. Bacterial protein preparation and large scale purification for the characterization
of innate protein characteristics
Bacterial protein expression has long been used as the gold standard for determining innate
protein characteristics for the simple fact that bacteria often will produce proteins that are folded
in the same way that they would be in human cells but we can separate these proteins away from
any other complex members and ascertain the structural and binding properties of these proteins
in absence of any other complex members. Given that relatively little was known about PHF6 and
its structural or binding properties, I therefore endeavored to create a bacterial protein expression
system that would allow us to address basic questions about the structure of this protein. I first
cloned the two major isoforms of this protein into the GST fusion vector pGOOD6P, as well as a
protein fragment containing only the second PHD domain. Once purified, these vectors were
59
transformed into Arctic Express bacteria, a modified version of BL21 that contains cold-shock
chaperone proteins. Single colonies are picked, and grown in an overnight culture with ampicillin
(to maintain the pGOOD6P) and gentamycin (to maintain cold-shock chaperones). These cultures
were diluted 1:100 in 100 ml LB with ampicillin and grown to an optical density of approximately
0.6, and placed in a shaker at 4
o
C. The cultures were allowed to equilibrate for 10 minutes, then
induced with IPTG to 1 mM and grown for 36 hours at 4
o
C. Bacteria were then pelleted, flash
frozen, and placed at -80
o
C overnight. Pellets were resuspended in 4 ml of a lysis buffer (16.7 mM
Tris, pH 7.0, 100 mM NaCl, 2.7 mM KCl, 10 µM ZnCl 2) and lysozyme was added to 2 mg/ml.
Bacteria were allowed to lyse on ice for 2-3 hours until a gelatinous mixture appears. Sonication
can be used to help lyse bacteria if lysozyme is insufficient to produce a gelatinous lysate. 10 ml
of 1% Tween-20 and 1 mM DTT was added to the lysate, and this was incubated for 1 hour. The
lysate was then spun down at 3000xg for 30 minutes. A crude isolate and pellet sample was then
taken, and remaining supernatant was incubated with 2 ml of a 50% Glutathione sepharose slurry
in lysis buffer overnight. Beads were then spun down gently at 500xg for 10 minutes, and washed
twice with the lysis buffer. A sample is taken at each step to verify the purification process. Elution
of protein is then performed by 2 rounds of incubation with 2 ml elution buffer (10 mM reduced
glutathione, 50 mM Tris-HCl, pH 8, 10 µM ZnCl 2). Purification is verified using SDS-PAGE gel
followed by coomassie stain. These purified proteins are then used for binding assays involving
arrays of histone peptides, individual biotinylated peptides, neural crest nuclear extract to verify
specific interactions, and purified histones to determine the range of histone modifications with
which this protein could interact.

60
Chapter 3: Spatio-temporal expression of Phf6

Section 3.1. Introduction
Birth defects affecting the central nervous system and craniofacial structures are amongst
the most prevalent birth malformations
139,140
, affecting greater than 1 in 1000 live births. Birth defect
syndromes that affect multiple systems, while less common, pose a significant health problem as
patients require continual health monitoring by a team of specialists throughout the course of their
lives
141
. Börjeson-Forssman-Lehmann Syndrome (BFLS) is one such birth defect syndrome that
affects males, and in rare cases females
94,98,106,107,109
and is caused by mutations in the gene PHF6
142
. PHF6
is thought to be a chromatin adapter protein
112
but its role in a developmental setting is poorly
understood. Clinically, BFLS is defined by a discrete set of features that can have varying
penetrance, including intellectual disability, obesity with gynocomastia, microcephaly,
characteristic craniofacial features, hypogonadism, seizures, developmental delay, and
occasionally digit elongation
94,106,109
. BFLS shows significant similarities to Coffin-Siris syndrome,
which is characterized by mutation in the ARID1B gene. Indeed, exome sequencing of several
clinically diagnosed Coffin-Siris syndrome patients revealed PHF6 mutations
143,144
. Thus,
understanding the molecular mechanisms of PHF6 effects in early development has broad
implications for complex birth defect syndromes as well as for development as a whole.
Xenopus laevis has long been used as a model organism of early development and has
unique features that allow in-depth study of proteomics and structure-function in a highly-
controlled manner that is especially suited for understanding PHF6 function and modeling BFLS.
Since BFLS is a birth defect that likely manifests in very early stages of development, given that
cellular derivatives of multiple germ layers are affected in this disease, externally developing
61
Xenopus embryos offer an added advantage of studying PHF6 function in early development in
real time. To understand the effect of the large number of distinct mutations that can result in
BFLS, Xenopus is ideal as it produces large clutches of synchronously developing embryos,
allowing robust statistical analysis of multiple control and experimental groups within the same
clutch
145
. This provides an ideal model to perform screening of structure-function analysis of many
different BFLS patient mutations. Further details of the Xenopus laevis system and its utility in
studying complex developmental disorders can be found in Chapter 2.
Here, we defined the expression pattern of Phf6 in this model organism and compared the
expression pattern to that which has been previously determined within mouse embryos. Previous
work has shown the expression of Phf6 in mouse embryos from E9.5 onward
115
. This study showed
expression of Phf6 mRNA within the neural crest and neural tube of these embryos, as well as
within developing brain sections at E12.5 and E15.5
115
. They showed the mRNA expression pattern
generally in earlier embryos via northern blot, and protein expression in the developing brain,
frontonasal process, developing limb bud, and testis on frozen sections of embryos at E12.5 and
later
115
. They further analyzed the postnatal and adult expression patterns of Phf6 mRNA and
protein
115
. This provides a good base for determining the early developmental expression patterns,
and yet matching later, late tailbud stage expression patters with the expression patterns that have
already been established. This comparison will assist in determining the viability of study of the
role of Phf6 in Xenopus.
In this chapter, we determined that Xenopus laevis can be used as a model for study of
Phf6 in early development. Examination of the phylogeny of PHF6 protein structure established
that Phf6 is sufficiently conserved between Xenopus laevis and humans to imply highly similar
structure and function in the two species. We examined the expression patterns of Phf6 mRNA
62
using in situ hybridization on whole mount embryos throughout early development. We also
determined that an alternate isoform, observed in human cells, is expressed in Xenopus laevis, and
performed expression analysis of both this novel isoform as well as the established isoform of Phf6
using RNA-seq. We conclude that Xenopus laevis serves as a valuable model system to understand
the effect of Phf6 on multiple tissue types in early development, and for the role of Phf6 in BFLS
and other developmental disorders. By drawing attention to the specific pattern of expression of
Phf6 in early development, we show the tissues where Phf6 mutation are the most relevant and
where further study would link Phf6 activity to the specific defects observed in patients.

Section 3.2. Results
3.2.1 Phf6 gene structure is highly conserved in Xenopus laevis
We performed a phylogenic analysis of the Xenopus laevis Phf6 protein structure compared
with that of the Homo sapiens PHF6 to determine the validity of using Xenopus laevis as a model
to study the early developmental effects of Phf6 perturbation. Using ClustaLW based alignment
algorithms and allowing the protein structure of Phf6 to cluster undirected using the bioinformatics
tool Geneious version R10.0.6
146
using a Jukes-Cantor amino acid substitution model to calculate
evolutionary distance
147
, we found that Xenopus laevis Phf6 showed over 75% amino acid identity
with human PHF6 protein, and clustered closest to human PHF6 among model organisms that
develop ex ovo and ex utero (Fig. 3.1A). Genetic distance, shown on the right of Fig. 3.1A, is
calculated using a Patristic Distance Matrix, equivalent to the length of tree branches between two
of the PHF6 proteins on this tree
148
, as related to human PHF6 protein sequence. Our analysis also

Figure 3.1. A. Phylogenic analysis of PHF6 protein sequence between Homo sapiens, Canis
lupis familiaris, Rattus norvegicus, Mus musculus, Gallus gallus, Xenopus laevis, Xenopus
63
tropicalis, and Danio rerio. Class Mammalia in green, class Aves in yellow, class Amphibia in
orange, and class Actinopterygii in black. Phylogenic distance listed to the right. B and C.
ClustaLW pairwise alignment of Xenopus laevis Phf6.L (A.) and Phf6.S (B.) protein sequences
against hPHF6. Green bars above show identical residues, PHD domains are outlined by grey
bars. Distribution of hydrophobic residues are shown by the blue and red bars below each protein
sequence. Black residues in protein sequence show complete identity, grey residues show similar
residue type.

Distance:
Homo sapiens 0.000
Canislupus familiaris 0.003
Rattusnorvegicus 0.024
Musmusculus 0.026
Gallus gallus 0.112
Xenopuslaevis 0.278
Xenopustropicalis 0.260
Daniorerio 0.355
A.
B.
C.
0.03
ePHD1
ePHD1
ePHD1
ePHD1
ePHD2
ePHD2
ePHD2
ePHD2
Identity
Identity
Homo sapiens
Homo sapiens
Xenopuslaevis
Xenopuslaevis
Hydrophobicity
Hydrophobicity
Hydrophobicity
Hydrophobicity
Figure 3.1
64
identified an evolutionary node between Zebrafish and all other model organisms that shows a
remarkable amount of protein sequence divergence, even between Xenopus and Zebrafish. Taken
together Xenopus embryos can serve as a powerful system to model PHF6 protein function in early
vertebrate development.
We further examined the protein structure of Phf6 by comparing the distribution of polar
and hydrophobic residues within the human and Xenopus laevis proteins as an indication of
whether the two proteins are likely to have similar folding mechanisms. Xenopus laevis is a
pseudo-tetraploid organism, and there are 2 different loci that have Phf6 alleles, Phf6.L on
chromosome 8L and Phf6.S on chromosome 8S; these two genes have slightly diverging mRNA
sequences and intron-exon structures (Fig. 3.2). Given this gene duplication, we compared the
protein structure of both Xenopus Phf6.L and Phf6.S against human PHF6 (Fig. 3.1B and 3.1C).
Both proteins show a high degree of conservation of amino acid sequence to human PHF6 (>75%,
identical amino acids are shown in the top identity bars), broadly distributed along the length of
the protein, but more remarkable is that these two proteins show almost identical distributions of
hydrophobic residues (hydrophobicity lines) with human Phf6. This suggests that the identical
regions of the proteins are likely to be folded towards the inside of the 3D protein structure, a
correlation that has been well established within the literature
149,150
. Because the folded protein
structure is highly linked to protein function, we therefore conclude that protein function is likely
to be conserved between human and Xenopus laevis. Additional evidence for this concept comes
from the fact that the known protein motifs of the 2 PHD domains have the highest degree of
sequence conservation (>90% identity), which suggests that these functional domains are under
evolutionary pressure for maintenance from Xenopus laevis to human proteins. Given the nature
of evolutionary conservation between these two organisms in Phf6 protein sequence and structure,
65
Figure 3.2. A. Phf6 genomic locus on Xenopus laevis Chromosome 8L. B. Schematic of Phf6.L
pre-mRNA structure. White: UTR, Red: undefined coding regions, Blue: ePHD finger 1 coding
regions, Purple: ePHD finger 2 coding regions. C. Phf6 genomic locus on Xenopus laevis
Chromosome 8S. D. Schematic of Phf6.S pre-mRNA structure. E. Pairwise comparison of amino
acid structure resultant from Phf6 from the L locus (top) and the S locus (bottom). Top green bars:
sequence identity at each position, Grey: positions of the ePHD domains, Blue and Red bars: score
of each amino acid in the sequence on hydrophobicity scale. Black bars: similarity of polarity of
each comparitive residue.

Chr8L
Phf6.L
Chr8S
Phf6.S
PHD1 PHD2
Phf6.S genomic locus
Phf6.L genomic locus
PHD1 PHD2
A.
B.
C.
D.
E.
ePHD1
ePHD1
ePHD2
ePHD2
Identity
Phf6.L
Phf6.S
Hydrophobicity
Hydrophobicity
Figure 3.2
66

Xenopus laevis is likely to be a useful model organism for the study of the early developmental
defects involved in BFLS.

3.2.2. In situ hybridization of Phf6 shows a highly dynamic pattern of expression in Xenopus.
Based on the evidence that the function of Phf6 in humans may be reflected in the behavior
of Phf6 in Xenopus, we sought to determine the distribution of Phf6 in Xenopus during early
development. In order to investigate the overall pattern of expression of Phf6 mRNA in Xenopus
laevis embryos, we created two RNA probes against the phf6.L allele: one in the 3’UTR of the
longest version of the mRNA, and one in the coding region of the mRNA (Fig. 3.3). These probes
were then used for whole mount in situ hybridizations on embryos ranging from early gastrulation
to late tailbud stage (NF11-30). An antisense probe targeting Twist1 was used to mark pre-
migratory and migratory neural crest, while a probe targeting Sox2 was used to mark neural and
pre-placodal tissues. These two markers were chosen based on previous observations of neuronal
Phf6 expression and the presence of both mental disability and distinct craniofacial phenotypes in
patients, and these two probes are classically defined markers of neural crest and neural tissues,
respectively.
In gastrulation stage Xenopus embryos, Phf6 is widely expressed in the prospective
mesoderm and ectoderm, with highest expression around the blastopore (Fig. 3.4 A, B). The broad
expression of Phf6 suggests there may be a role in mesodermal and ectodermal lineage
determination during gastrulation. The pattern of expression of Twist contrasted with this, as
expected; Twist was not expressed in the gastrula stage embryo.
Neurula stage embryos showed a much more restricted and dynamic expression pattern of Phf6.
In early neurula embryos, we observe highest expression of Phf6 in the region where the lateral
67
Figure 3.3. A. Splicing from Phf6.L pre-mRNA leading to Phf6a.L mature mRNA, and the portion
of the transcript used to create the antisense probe against Phf6a transcripts. B. Splicing from
Phf6.L pre-mRNA leading to Phf6b.L mature mRNA, and the portion of this transcript used to
create the antisense probe highlighting the expression of all Phf6 transcripts.

Phf6b
1 2 3 4 5 6 7 8 9 10
pre-mRNA
PHD Finger1 PHD Finger2
mRNA
Pan-Phf6 probe
1 2 3 4 5 6 7 8 9 10 pre-mRNA
Phf6a
PHD Finger1
PHD Finger2
mRNA
Phf6aprobe
A.
B.
Figure 3.3
Pan-Phf6 probe
Phf6.L
Phf6.L
68
Figure 3.4. In situ hybridization of Xenopus laevis embryos. i. Model of embryos, tissue types
shown. Yellow: Endoderm. Red: Mesoderm. Blue: Ectoderm. Orange: Neural Tube. Green: Neural
Crest. Pink: Placode. Dark Red: Somites. ii. Twist1 in situ hybridization. iii. Sox2 in situ
hybridization. iv. Phf6 in situ hybridization, using a probe designed from the 3’UTR of the longest
isoform made from Phf6.L locus. v. Phf6 in situ hybridization, using a probe designed from the
coding region of the Phf6.L locus. A. embryos at NF11, mid gastrula stage. B. embryos at NF12.5,
late gastrula stage. C. embryos at NF16, early neurula stage. D. embryos at NF18, mid neurula
stage. E. embryos at NF23, early tailbud stage. F. embryos at NF26, mid tailbud stage. G. embryos
at NF29, late tailbud stage. Filled arrow points out present neural crest staining, hollow arrow
points out absent neural crest staining. All are representative images of 6 in situ hybridization
experiments, >5 embryos per stage per experiment.

Twist1 Sox2 Phf63’UTR Phf6 coding
NF11
Mid-
Gastrula
NF12.5
Late-
Gastrula
A. i.
B. i.
ii.
ii.
iii.
iii.
iv.
iv.
v.
v.
NF16
Early Neurula
NF18
Mid Neurula
C. i.
D. i.
ii.
ii.
iii.
iii.
iv.
iv.
v.
v
.
NF23
Early Tailbud
NF26
Mid Tailbud
NF29
Late Tailbud
E. i.
F. i.
G. i.
ii.
ii.
iii.
iii.
iv.
iv.
v.
v.
ii.
iii.
iv.
v.
Figure 3.4
Ectoderm Endoderm Mesoderm Neural Neural Crest Cranial Placode Somite
H. i. H. ii.
69
cranial neural crest and anteriormost pre-placodal regions are specified, with low expression across
the neural plate (Fig. 3.4C). In a later neurula stage, the expression of Phf6 appears to be
upregulated in the entire neural plate, resembling the Sox2 expression domain, but Phf6 continues
to be expressed in pre-migratory neural crest (Fig. 3.4D). During both early and later neurula
stages, expression of Phf6 overlaps with Twist expression, with perhaps a larger zone of expression
of Phf6 at the earlier time point (Fig. 3.4C, D). Initially, Phf6 expression overlapped well with the
pre-placodal Sox2 expression and, later in development, overlaps with the distribution of Sox2 in
the developing neural tube. Both probes against Phf6 transcripts reflect highly similar expression
patterns during neurulation, indicating that this is likely to reflect the pattern of expression for the
intact mRNA (Fig. 3.4C, D).
In situ hybridization in tailbud stage embryos showed that Phf6 is highly expressed within
the developing neural tube as well as in the optic and otic vesicle region, but not within the
migrating neural crest (Figs. 3.4E, F). The spatial distribution of Phf6 overlapped well with Sox2
expression, but not with Twist expression, which marks the migrating neural crest during these
stages (Fig. 3.4E, F). Phf6 exhibited a broader expression pattern than Sox2, indicating expression
within the axial mesoderm. In mid- to late-tailbud stage embryos, there is a punctuated Phf6 somite
expression pattern in the trunk region (Fig. 3.4F, G, H), indicating an overall expression within
paraxial mesoderm (see the arrow in Fig. 3.4G panel iv.). This is also evident in higher
magnification as well as in a transverse section, where staining appears in the oval-shaped somite
regions flanking the outlined notochord (Fig. 3.4 H. panels i. and ii.). Embryos stained at late
tailbud stages also seem to regain expression of Phf6 within the distinctive, migrating neural crest
arches; this Phf6 expression closely resembles Twist expression at this late stage. This is consistent
with the previously published expression patterns, where Phf6 is expressed in the most ventrally
70
localized crest within the first arch and the frontnasal process in mice at embryonic day 15
115
.
Overall, the expression patterns demonstrate a role for Phf6 in early development beginning in
gastrulation, and extending through neural crest formation to neural and somite development and
maturation.

3.2.3. Human neural crest differentiated from embryonic stem cells show a similar pattern of
PHF6 expression.
In order to verify this expression pattern, we performed differentiation of neural ectoderm
and neural crest from human embryonic stem cells. We then performed immunofluorescence using
an N-terminal antibody against PHF6, and paired these with immunofluorescence against neural
crest marker AP2a or HNK1. AP2a (Fig. 3.5B panel ii’) shows early expression in the nuclei of
migratory neural crest cells, while HNK1 (Fig. 3.5C) is generally expressed on the cell surface of
approximately 50% of migrating and mature neural crest cells. We found that PHF6 is highly
expressed in the neural ectodermal cells, but is rapidly downregulated upon neural crest
delamination (Fig. 3.5A, B). This closely matches what we see on the RNA level within Xenopus
laevis embryos, in that migrating neural crest cells have no appreciable level of Phf6 transcripts
(Fig. 3.4E, F). However, within this differentiation system, we then see PHF6 protein in neural
crest cells that have migrated to some distance away from the neural ectodermal cluster, and have
begun to mature (Fig. 3.5C). We observe that a higher percentage of cells express PHF6 protein
as we move away from the neural ectodermal cell cluster (Fig. 3.5C panel i. and 3.5C panel ii.).
This re-activation of expression matches what we see in Xenopus laevis embryos, in that late
tailbud stage embryos regain pharyngeal arch Phf6 expression at NF28-30 (Fig. 3.4G), as well as
previously published expression patterns
115
. The expression patterns of PHF6 within the neural
71
Figure 3.5. Immunofluorescence analysis of PHF6 in human neural crest cells. A., B. PHF6
immunofluorescence in pre-migratory neural crest cells (A.) and early migratory neural crest cells
(B.), green; nuclear fluorescence using Hoescht, red. First inset, I, outlined in B’; second inset, ii,
imaged at 20x in another region of the dish; AP2a fluorescence in purple. C. PHF6
immunofluorescence in mature neural crest cells, green, nuclear fluorescence using using Hoescht,
red; HNK1 immunofluorescence, blue. First inset, i, and second inset, ii, correspond to the outlined
regions in C’. Single experiment.

Premigratory !
Day 8!
Early Migratory!
Day 10!
*"
*"
*"
*
*"
*"
*" *"
*"
*"
*"
*"
Late Migratory!
Day 14!
*"
PHF6!
PHF6!
DAPI!
DAPI! Merge!
AP2α!
PHF6! HNK1! DAPI! Merge!
*
*
*
* *
*
*
*
*
*
*
*
*
*
*
PHF6
PHF6
PHF6
PHF6
PHF6
PHF6
PHF6
Premigratory NCC Migratory NCC Mature NCC
A A’
B B’
i
Bi Bi’
Bii Bii’
C
C’ i
ii
Ci
Cii
Ci’ Ci’’ Ci’’’
Cii’
Cii’’ Cii’’’
Figure 3.5
Hoescht PHF6
Hoescht
Hoescht
PHF6
PHF6
Hoescht
Hoescht
Hoescht
AP2⍺
HNK1
HNK1
HNK1
72
crest differentiated from human embryonic stem cells seems to mimic embryonic expression
patterns.

3.2.4. Xenopus laevis expresses an alternate isoform of Phf6 that lacks the second PHD domain.
Alternate isoforms of PHF6 have been described in humans, including an isoform that
lacks the second PHD domain, here referred to as PHF6b. In order to determine whether Xenopus
laevis also expresses an alternate isoform of similar structure, we took a data mining approach to
perform de novo transcriptome assembly. We used published RNA-seq data sets containing paired-
end sequencing reads of cultured Xenopus laevis animal caps that had been untreated, treated with
Noggin mRNA in order to create neural precursor-like cells, or treated with Noggin and Wnt
mRNA to create neural-crest like cells
151,152
. The use of cultured animal caps to model cell types of
interest is well characterized
153
, and has been used to perform proteomics and genomics on neural
precursor like and neural crest like cells in many studies
154,155
. We performed de novo transcriptome
assembly using two different methods on a set of paired end sequencing reads described in Hatch
et al., 2015, to determine if there is an isoform of Phf6 in Xenopus laevis that has a similar protein
structure to that of PHF6b. In the first of these approaches, we performed genome-independent
assembly through the Trinity RNA-seq protocol
156
followed by BLAST alignment
157
; in the second,
we used cufflinks genome-dependent assembly
158
and cuffdiff tracking of the established
transcripts
159
. In both methods, we identified an alternate isoform from the Phf6.L locus and the
Phf6.S locus. To independently confirm the expression of the shorter transcript, we cloned the
Phf6b.L coding sequence from cDNA from whole Xenopus laevis embryos, NF 20. This sequence
is shown in Table 3.1. We then compiled data from wild type animal caps (vehicle), animal caps
injected with Noggin mRNA, and animal caps injected with Noggin and Wnt mRNA from Hatch
73
et al. and Thelie et al. Using HTSeq
160
, we then generated raw counts of the various loci and used
edgeR
161
to normalize and calculate expression values. We performed further relative normalization
to the expression of beta-actin, which was very consistently expressed across the samples.
Significance level was calculated using a two-tailed student’s t-test with Bonferroni multiple
comparison correction.
The resultant data shows the expected expression patterns of neural markers, Pax6.L,
Pax6.S, and Sox2, where all three of these mRNAs are significantly upregulated in Noggin treated
caps (Fig. 3.6A and 3.6B). Additionally, analysis of several neural crest markers (Snail, Twist and
Sox9), as well as several neural plate border markers (Msx1 and Zic1) showed significant
upregulation in animal caps treated with Noggin and Wnt (Fig. 3.6C-G). The expression of these
markers suggests that, as expected, the treatment of caps with Noggin mRNA provides a good
model for neural precursors, while the treatment of caps with Noggin and Wnt provides a good
model for pre-migratory neural crest. We went on to analyze expression levels of both the
established isoforms of Phf6, which will be referred to here as Phf6a.L and Phf6a.S, as well as the
newly identified Phf6 isoforms that resemble the shorter isoform in humans, here referred to as
Phf6b.L and Phf6b.S, for the L and S loci respectively (Fig. 3.6H). Phf6a.L is significantly
upregulated in both Noggin and Noggin+Wnt treated animal caps, which is consistent with our in
situ hybridization data suggesting that Phf6 is highly expressed in neural tissue and pre-migratory
neural crest. Interestingly, though Phf6a.S shows a trend towards an increase in expression in both
Noggin and Noggin+Wnt animal caps, there was no significant difference in expression,
suggesting that this locus may maintain a more stable, but lower, level of expression (Fig. 3.6H,
Phf6a.S). However, expression from the S locus may not be represented within the existing in situ
hybridization data, since both probes were designed based upon the sequence of Phf6.L (Fig. 3.3).
74
Figure 3.6. RNA-seq analysis of specific transcripts in Xenopus laevis animal caps treated with
vehicle (grey), Noggin mRNA (blue), and Noggin and Wnt mRNA (green). A. Differential
expression of Pax6 mRNA, B. Sox2 mRNA, C. Snai1 and Snai2 mRNA, D. Msx1 mRNA, E. Zic1
mRNA, F. Twist1 mRNA, G. Sox9 mRNA, and H. Phf6 mRNA. * p<0.05 ** p<0.01 *** p<0.001,
vs vehicle, calculated using a two-tailed student’s t-test with Bonferroni multiple comparison
correction. Compiled transcripts, n=4 paired-end RNA-seq tracks for vehicle, noggin; n=2 paired-
end RNA-seq tracks for Noggin + Wnt. Error bars represent SEM.

*
*
0
0.002
0.004
0.006
0.008
0.01
0.012
phf6a.L phf6a.S phf6b.S phf6b.L
Vehicle Noggin Noggin+Wnt
H.
*
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
pax6.L pax6.S
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
sox2.S
**
***
*
***
0
0.01
0.02
0.03
0.04
0.05
0.06
snai1.S snai2.L
**
***
*
A. B. C.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0.16
0.18
msx1.L msx1.S
***
***
***
**
0
0.01
0.02
0.03
0.04
0.05
0.06
zic1.L zic1.S
**
***
D. E.
0
0.0005
0.001
0.0015
0.002
0.0025
0.003
twist1.L twist1.S
0
0.005
0.01
0.015
0.02
0.025
0.03
0.035
0.04
sox9.L sox9.S
*
*
**
*
F.
G.
Figure 3.6
75
Furthermore, both Phf6b versions showed consistent expression across the 3 samples, with Phf6b.L
showing a significant increase in Noggin treated animal caps. This data provides higher resolution
of the expression of Phf6 in neural crest and neural precursor tissues as compared to animal cap
expression, showing the distinct upregulation observed in our in situ hybridization studies in neural
plate and pre-migratory neural crest is consistent with overall mRNA expression. Additionally,
these results demonstrate the alternate isoform, Phf6b, to be broadly expressed but at a low level.
The presence of this alternatively spliced mRNA confirms that Phf6 is not only conserved in terms
of protein structure, but also in terms of splicing, indicating that the role of the two isoforms can
be studied in this organism and extrapolated to human cells.

Section 3.3 Discussion
3.3.1. Conservation of protein domains
The results shown here demonstrate that the protein sequence of Phf6 in Xenopus laevis is
highly conserved with the protein sequence of PHF6 in human cells, both in amino acid identity
as well as amino acid hydrophobicity and polarity. This suggests that the structure and function of
these two proteins are likely to be highly conserved. In particular, the defined PHD domains seem
to contain the highest sequence similarity between these two organisms. PHD domains are
common protein motifs that have been shown to bind a number of different protein modifications
and small molecules
162–164
but, in the context of nuclear expression
142
, PHD domains are classically
known to bind modified histone tails
164,165
. The PHD domains of PHF6 have been partially
characterized, as the crystal structure of the second PHD domain has been resolved, and others
have shown that this domain can bind non-specifically to DNA
110
. A recent study has resolved
portions of the secondary structure of the first PHD domain
111
, but has not identified the specific
76
binding partners of this domain. Molecular features of epigenetic control are highly conserved
between Xenopus laevis and humans, and many elements of histone structure and positioning were
originally defined using this system
166–168
. The high degree of conservation of PHF6 combined with
the need to further resolve the structure and function of this protein and its defined domains
underscores the need for a model system like Xenopus laevis, where the function of the domains
of the protein can not only be defined but then can be related back to the overall development of
the organism. Thus, use of this organism to study the PHD domains of Phf6 can provide a true
structure-function analysis within a developmental setting.

3.3.2. Conservation of splicing and function of alternative splicing
Our results demonstrate that an alternate isoform of Phf6 is produced in Xenopus laevis,
which produces a protein lacking the second PHD domain. This isoform has been previously
observed in human cells, demonstrating that the splicing of Phf6 is also conserved between human
and Xenopus. The specific function of this isoform is currently not known, but evidence from
patient mutations and protein biology suggest that this alternate isoform may act in a dominant
negative fashion.
Given the X-linked nature of this disorder, patients with PHF6 mutations that are
symptomatic for BFLS are primarily male, with the majority of carrier females being unaffected.
However, a few heterozygous female patients have been diagnosed with BFLS
109
. In spite of one
normal copy of the gene, the higher degree of penetrance of the disease in the affected females
suggests that the mutations they carry may have a stronger effect than the mutations originally
described in male patients. Interestingly, male patients mostly have missense mutations (Fig. 3.7A,
top), while the symptomatic female patients are more likely to have large or full deletions of the
77
Figure 3.7. A. Schematic of BFLS patient mutations. White Circles: missense mutations, red
circles: nonsense mutations, blue circles: frameshift mutations, green circle: amino acid deletion,
green bars: deletion over the area outlined by the bar, blue bar: duplication over the area outlined
by the bar. Male patient mutations are outlined above the protein schematic, with the precise
residue affected next to each circle, while female patient mutations are outlined below the protein
schematic, with point mutations having the precise residue affected next to each circle.
105,106,108,109,144,169–174

B. Conservation analysis of distinct amino acids in the two PHD domains, the maintained
structural amino acids in the two PHD domains, and an unstructured region of PHF6. NS: no
significance, p>0.05; *** p<0.001 using a two-tailed student’s t-test. Error bars represent SEM.
C. Mapping of divergent amino acids between the two PHD domains and the structural amino
acids in the two PHD domains. Blue: structural amino acids; Orange: divergent amino acids. D.
Conservation of divergent and structural amino acids, PHD1. E. Conservation of divergent and
structural amino acids, PHD2. Identity bars indicate conservation score for each amino acid, which
are averaged in B.

78
protein or mutations affecting the second PHD domain
109
. Furthermore, reports of deletion of the
second PHD domain
109
in female patients, which has the potential to generate a shorter transcript,
similar to the alternate isoform PHF6b, leads us to speculate that the patients inherited a dominant-
negative form of the protein which increases the likelihood of exhibiting symptoms of BFLS.
Corollary to this, the wide spectrum of PHF6 mutations in male BFLS patients whose carrier
mothers are unaffected are likely to represent a partial loss of function. Given that some of the
female mutations are suggestive of dominant negative effects and bear a striking resemblance to
the alternate isoform of PHF6, we wonder if the alternate isoform may be an intrinsic mechanism
to negatively regulate PHF6 function and precisely control the activity of full-length Phf6.
The two PHD fingers of PHF6 are very similar to each other and likely arose via tandem
duplication. It remains to be seen if they have retained their function or diverged to acquire
different roles. These two protein domains are highly similar in the distribution of similar
characteristic amino acids (H. sapiens: 48.3% amino acid identity, 66.9% Similarity via Blosum45
with threshold 1, X. laevis: 44.9% amino acid identity, 67.8% Similarity via Blosum45 with
threshold 1; Fig. 4B). We find that the small numbers of amino acids that differ between the first
and second PHD fingers are as conserved across multiple vertebrate species as the PHD domain
overall and more highly conserved than other regions of the protein. This suggests that there is
evolutionary pressure to maintain these residues and their positions and functions (Fig. 3.7 B-D).
This observation, along with the differences in the three dimensional structure of the two PHD
fingers
111,175
suggest that these two terminal PHD fingers may bind different targets, functioning as
a bridge. The shorter transcript then would lack the adaptor function and could function as a natural
dominant negative.
79
Further efforts to characterize this alternate isoform and the various BFLS mutations
biochemically and in the context of a developing embryo are required to test the hypothesis that
this alternate isoform acts as a dominant negative to the canonical isoform. These studies could be
easily performed in the context of Xenopus laevis embryology.

3.3.3. High-resolution map of expression in early development
In situ hybridization studies presented here show that Phf6 is expressed highly in
developing neural plate border, neural tube, developing axial and paraxial mesoderm, and around
the otic and optic placodes. This study of how Phf6 expression changes over early development
provides clues for the study of the individual defects observed in patients with BFLS. The
observation that Phf6 is expressed within the neural plate border region but not in the migratory
neural crest indicates that the craniofacial abnormalities observed in patients may involve a defect
in very early stages of neural crest migration, that of neural crest epithelial-to-mesenchymal
transition, and that transient, but near complete, downregulation in the early migrating cranial
neural crest suggests a rapid switch in Phf6 status that may be necessary for the normal
development of neural crest cells. That Phf6 is continually expressed in the developing neural
tissue from mid-neurula stage throughout the timecourse of this study suggests that there is a
continuing role for Phf6 in the formation of an intact central nervous system. This is corroborated
by the fact that a hallmark of BFLS is intellectual disability and failure to reach mental
developmental timepoints; these characteristics could be caused by the failure of the central
nervous system to develop correctly. Furthermore, the expression of Phf6 in paraxial mesoderm
shows that there is a role for this protein in the correct formation of mesodermal cell types, and
this could be linked to the phenotypes of obesity with gynocomastia and hypogonadism through
80
the improper formation of adipose tissues as well as gonad tissues. We therefore present a roadmap
of tissues to study throughout embryological development in order to address the effects of Phf6
and Phf6 mutation and their role in the pathogenesis of BFLS.

Section 3.4 Experimental Procedures
3.4.1. Molecular Cloning
Probes and Phf6b alternate isoform were cloned into the pCS2+ backbone. Primers were as
follows: Phf6 3’UTR forward: 5’-AAAAGGATCCACTTTATTGTAAAAATCA CAGTGG-3’
and reverse: 5’-AAAACTCGAGATTTCTAGACTTTGGAATTTATTTTCCC-3’; Phf6 Coding
forward: 5’-AAAAGGATCCATGTCTAGCTCCACTGGAC-3’ and reverse: 5’-
AAAACTCGAGTCACAAAATACTACACATCAATTC-3’; Phf6b coding forward: 5’-AAAA
GGATCCATGTCTAGCTCCACTGGAC-3’ and reverse: 5’-AAAACTCGAGTCACAAAATA
CTACACATCAATTC-3’. Amplicons were created from whole Xenopus laevis embryo cDNA
template using the following PCR conditions: 95
o
C 1 minute, 95
o
C 30s, 60
o
C 30s, 72
o
C 1 minute,
30 rounds. Amplicons were purified and digested using BamHI and XhoI (NEB) double digestion
for 1 hour, purified, ligated to digested pCS2+ backbone, and transformed into DH5α bacteria.
Colonies were grown, prepped, and sequence verified.

3.4.2. Probe preparation
Probes were created using Ambion T7 megascript and Roche Dig-UTP mix. Briefly, probe
plasmids were linearized as follows: Phf6 3’UTR and Phf6 coding: BamHI; probes for Twist1 and
Sox2 have been previously described
176,177
. Linearized plasmids were verified using gel
electrophoresis and purified. 1 μg of linearized plasmid was incubated in a standard T7 megascript
81
reaction mix at 37
o
C for 2 hours, and probe concentration was estimated. Probes were stored at -
80
o
C in a 1:3 dilution with hybridization buffer to prevent RNase activity.

3.4.3. Phylogenic Analysis
Protein sequences of PHF6 were downloaded from the NCBI database with accession numbers as
follows: H. sapiens: NP_001015877, M. musculus: NP_081918, R. norvegicus: XP_008771882,
C. lupis familiaris: XP_005641890, G. gallus: XP_004940719, X. tropicalis: XP_012824040, X.
laevis: XP_018084592 (corresponds to the L homologue), and D. rerio: XP_005173256. Using
these sequences, the Geneious tree building tool arranged the phylogenic distances and relation
based on pairwise alignment via global alignment with free end gaps and a Blosum62 cost matrix
146
. The Jukes-Cantor
147
method of approximating genetic distance was then used to assemble the
tree. Distance calculations were performed using a Patristic Distance Matrix, which is equal to the
sum of the lengths of the branches that link two proteins on a tree
148
, and thus represents the amount
of evolutionary distance between two PHF6 proteins. Individual pairwise comparisons were
performed first by creating individual protein translations from the Phf6 loci on Chromosome 8L
(Phf6.L, extracted from Xenbase J-Strain 9.1 genome, www.xenbase.org) and on Chromosome 8S
(Phf6.S, extracted from Xenbase J-strain 9.1 genome)
178,179
. Pairwise alignment was performed using
ClustalW Alignment and a BLOSUM cost matrix.

3.4.4. De-novo Transcriptome analysis
Raw paired-end read files were downloaded from the ENA database
(www.ebi.ac.uk/ena/data/view/SRX1834645-SRX1834650) for Hatch et al. data and
(www.ebi.ac.uk/ena/data/view/SRX825218-SRX825227) for Thelie et al. data. Read files were
82
uploaded to Galaxy web-based bioinformatics server, both to usegalaxy.org for genome-directed
alignment, henceforth called Tuxedo Pipeline, and galaxy.ncgas-trinity.indiana.edu for genome-
independent transcriptome generation, henceforth called Trinity Pipeline. Briefly, the Tuxedo
Pipeline proceeded as follows: TopHat gapped read alignment
180
using paired-end parameters,
using J-Strain 9.1 genome with repeats masked as the reference genome, intron length allowed
between 70-500000 bases, and mean inner distance between mate pairs as 300 bases; BAM files
for all TopHat outputs were merged to create overall alignment for all reads; Cufflinks transcript
assembly
158
using 300000 Max intron length, 0.01 minimum isoform fraction, 0.1 pre-mRNA
Fraction, no reference annotation, Bias correction using the reference genome, and Cufflinks
effective Length Correction; Transcripts were tracked back to the reference annotation using
Cuffcompare
158
and J-strain version 9.1 all transcripts gff3 file from Xenbase; single exon
transcripts were ignored. Trinity Pipeline proceeded as follows: fastq files were concatenated and
trimmed so that all left reads were in one file and had corresponding right reads, transcripts were
created using the three-step process of the Trinity suite
181
. Transcripts were then cross-referenced
between the two protocols, common transcripts were compiled using python parsing, and
transcripts were filtered based on Phf6 and reference loci. Count tables were created using HTSeq
160
, and differential expression was tested using EdgeR
161
.

83
Section 3.5. Tables

Table 3.1. Coding sequence of Phf6 alternate isoform, cloned

Sequence
Phf6b.
L
5’-
ATGTCTAGCTCCACTGGACAGAGAAAGGGCTCTTCACATCAGCAAC
ACA
AATGTGGATTCTGCAGATCTAACCGAGAAAAGGAATGTGGACATCT
ACTAATATCAAGCAACCAAAAGGTGGCAGCACATCACAGATGTAT
GCTGTTTTCATCAGCACTGGTATCATCACAGTCTGACAGTGAGAAT
CTCGGGGGCTTTTCAATAGAAGATATCCAAAAGGAGCTTAAGAGA
GGAAAAAAGCTGATGTGCTCACTTTGTCACTGCCCAGGAGCCACAA
TTGGGTGTGATGTTAAATCCTGCCATAGAAGCTACCATTACCACTG
TGCCCTGCGTGATAAAGCTCATATTCAGGAAAATCCATCACAGGGA
ATGTATACGATTTTTTGCAGAAAGCATAAGGAACAGGTGCGGAATT
CAGACGATGAATTAGAAGGTAGTTTTACACGTAGAGGTTTGACTCC
ATCTCCTCAACGTGGCCGAGGAAGAGGGTCTAGAGGAAAATCACG
GACATCAAATTCCAGAGGGCAGTCTGAGGAGAGCAGACTTTCATCC
TCACATTGTACAGAAGAAACCGAAAGCAGCTCTAGTAGAGATCGTT
CTCCACACAGAAGCAGTCCCAGTGACACTAGACCTAAATGTGGTTT
CTGTCACGCAGGAGATGAAGAAAATGAGACACGAGGCAAGCTCCA
TGTATTTAATGCAAAAAAGGCCGCGGCTCATTACAAATGCATGCTT
TTCTCCTCTGGCACAGTACAGCTCACCACAACATCAAGAGCAGAAT
TTGGAGACTTTGATATCAAAACTGTGATCCAAGAGATGAAACGTGG
GAAGAGAATGGTATTTTTCCAGGTTTTGTTTTTTGAATTGATGTGTA
GTATTTTGTGA-3’

84
Table 3.2. Probe sequences. Both probes are anti-sense to the respective targets, cloned into
pCS2+, linearized using BamHI and transcribed using T7 megascript polymerase.
Probe Sequence
Phf6
3’UTR
5’-atttctagactttggaatttattttccctaaaactagaacattaaagtttgcaatcagccaaacgct
ttgcaatgatacaccatgtattcatattgtgtaataaaggtaactgtacattaccgcagaacaaaagacacata
ttgacgtgtatgaaggtcattagtaacagcccacattgtatacatttacaaactgagctgcgtgggggccaa
acagactttttacgcaccaaagtctttcctcgcaatattttgcagcatgttgctgggcagtttttaaataaagaa
agtaaatagctatcttgtgctatgtcaacagtaagcaactggcagcacagccatagtttaaacagcagtgtg
acagcctgtctaaccctttccattggctttaaactgcttaaaacgtgtgaaatgcataaagcataaaaagtaat
aaccaatgggtacaaaccaagacatatgtaggataacaaccgctgccctttttgtgattatcacattttcacta
actggtgaaaaaaatacatatttcccatttccccgcagtccttttcacaatcctagataaatatatgacttttatct
aggaaattgaaactgtgtctgcattcccacaactggcagactgggcactggcttatatgacagctttgatcct
tagttcccattgagctgctgcacagtgctctgattgtgccctactgttgcattgctcttgcttctgatttcacgttc
ctcgtcctcttcatctctttcatcgtttccactgtgatttttacaataaagt-3’
Phf6
Coding
5’-
TCACAAAATACTACACATCAATTCAAAAAACAAAACCTGGAAA
AAT
ACCATTCTCTTCCCACGTTTCATCTCTTGGATCACAGTTTTGATA
TCAAAGTCTCCAAATTCTGCTCTTGATGTTGTGGTGAGCTGTACT
GTGCCAGAGGAGAAAAGCATGCATTTGTAATGAGCCGCGGCCT
TTTTTGCATTAAATACATGGAGCTTGCCTCGTGTCTCATTTTCTT
CATCTCCTGCGTGACAGAAACCACATTTAGGTCTAGTGTCACTG
GGACTGCTTCTGTGTGGAGAACGATCTCTACTAGAGCTGCTTTC
GGTTTCTTCTGTACAATGTGAGGATGAAAGTCTGCTCTCCTCAG
ACTGCCCTCTGGAATTTGATGTCCGTGATTTTCCTCTAGACCCTC
TTCCTCGGCCACGTTGAGGAGATGGAGTCAAACCTCTACGTGTA
AAACTACCTTCTAATTCATCGTCTGAATTCCGCACCTGTTCCTTA
TGCTTTCTGCAAAAAATCGTATACATTCCCTGTGATGGATTTTCC
TGAATATGAGCTTTATCACGCAGGGCACAGTGGTAATGGTAGCT
TCTATGGCAGGATTTAACATCACACCCAATTGTGGCTCCTGGGC
AGTGACAAAGTGAGCACATCAGCTTTTTTCCTCTCTTAAGCTCC
TTTTGGATATCTTCTATTGAAAAGCCCCCGAGATTCTCACTGTCA
GACTGTGATGATACCAGTGCTGATGAAAACAGCATACATCTGTG
ATGTGCTGCCACCTTTTGGTTGCTTGATATTAGTAGATGTCCACA
TTCCTTTTCTCGGTTAGATCTGCAGAATCCACATTTGTGTTGCTG
ATGTGAAGAGCCCTTTCTCTGTCCAGTGGAGCTAGACAT-3’

85
Chapter 4: PHF6 negatively regulates neural crest cell migration in Xenopus laevis.

Section 4.1. Introduction
Börjeson-Forssman-Lehmann Syndrome, or BFLS, is a relatively rare X-linked disorder
arising from mutation of the protein PHF6. BFLS affects multiple systems within patients,
including the neural system, causing intellectual disability, the neural crest, causing characteristic
facial phenotypes, metabolic and developmental regulation of adipose tissue causing obesity, as
well as correct hormonal development, causing gynocomastia and hypogonadism
182
. Classic BFLS
affects male patients, and mutations are largely conservative point mutations spread throughout
the coding region of the gene (Fig. 1.6). It is therefore a challenge to find a model to assess the
etiology of developmental defects observed, both in very early development in the case of neuronal
and neural crest defects, as well as later development in the case of metabolic and hormonal
development, and in a way where the structure-function of the protein can be assessed given the
breadth of mutations that are observed in patients.
PHF6 is a dual plant homeodomain (PHD) protein, and its structure involves two atypical
PHD motifs with additional “zinc knuckle” regions N-terminal to PHD region, creating an
extended PHD (ePHD)
183
. Multiple groups have found roles for this protein in cancerous cell lines,
but only one group has found a role for PHF6 in a developmental system, that of a defect in
neuronal pathfinding
116
. Due to the presence of defects in neuronal and neural crest cell types, it is
necessary to delve further into the early role of PHF6 in development in order to determine whether
PHF6 has a role in the initial specification and formation of neural cell types and neural crest cell
types. We assessed in Chapter 3 the viability of using Xenopus laevis as a model for the early
developmental defects within BFLS and found that this system not only recapitulates established
86
expression patterns, but likely has conserved the molecular function of PHF6 based on protein
structure similarities. Therefore, this system is appropriate for use to address the effect of PHF6
on cells like the neural crest.
Neural crest development in Xenopus laevis is a multi-step process that involves initial
specification, delamination, migration, pathfinding, and terminal differentiation
133
. In the cranial
neural crest region, neural crest progenitor cells are initially formed at the neural plate border and
migrate ventrally within well-defined streams to eventually converge and form the brachial
cartilage
133
. Modeling patient disorders in Xenopus has allowed a more in-depth understanding of
the basic processes surrounding neural crest development. In particular, modeling in this system
has yielded insights into craniofacial disorders such as CHARGE syndrome, in which many
patients carry de novo mutations in CHD7
29
, several groups have used the Xenopus system to model
Fetal Alcohol Syndrome and the craniofacial disorders associated with it
184
, and work in Xenopus
has helped uncover the role of Tbx1 in the craniofacial defects involved in DiGeorge Syndrome
185
.
The Xenopus system therefore has the potential to aid in the understanding of the role of PHF6 in
neural crest development.
I report here that Phf6 in Xenopus laevis is involved as a tissue-autonomous negative
regulator of the neural crest cell migratory phenotype, and that perturbations in the levels of PHF6
in the progenitor population change the cellular morphology and the migratory ability of neural
crest cells. These observations allow me to propose that PHF6 loss in BFLS patients leads directly
to precocious migration of neural crest cells and thus hyperplasia of the neural crest associated
structures within the craniofacial region. This then causes the observed characteristic facial
phenotype observed in BFLS patients.

87

Section 4.2. Results
4.2.1. Targeted Phf6 knockdown in neural ectoderm of embryos causes premature migration of
neural crest cells.
In order to determine the effect of endogenous Phf6 in the regions where it is most highly
expressed in early development, I developed two splice-blocking antisense morpholinos for
depletion of Phf6 within embryos. The first splice-blocking morpholino, henceforth MO1, showed
exon-skipping of exon 8 of Phf6, resulting in loss of 150 bases of RNA in the spliced product (Fig.
4.1B, lower band) and loss of the zinc-knuckle region of the second ePHD motif. This region is
thought to be indispensable for the interaction between this domain and DNA
183
and thus we expect
a loss of functionality in the second ePHD motif. Approximately 2/3 of the total mRNA was
differentially spliced at 4 µM of MO1. The second splice-blocking morpholino, henceforth MO2,
is predicted to cause exon skipping of exon 3 of Phf6, which would result in the loss of the first
PHD motif and a pre-mature stop codon. This exon skipping has yet to be successfully tested,
though it is clear that this transcript is not degraded by nonsense-mediated decay to any significant
extent (Fig. 4.1B, MO2). Both morpholinos were then injected into one of the dorsal, animal
blastomeres at NF4 along with a Ruby Dextran to fluorescently trace the targeting of the
morpholino. These embryos were then selected for correct targeting of the neural plate border and
some involvement of the neural plate, but minimal involvement of the more lateral ectoderm
(pattern demonstrated in Fig. 4.1C). Embryos with such an expression pattern were allowed to
develop to NF20, and the migration of neural crest was tracked via the fluorescent dextran until
NF27, whereby embryos with significant expression of the fluorescent dextran within the
88
migrating neural crest cells were imaged for fluorescence. Embryos targeted with a morpholino
against GFP, which acts as a non-targeting control, showed a predictable pattern of pharyngeal
Figure 4.1. Phf6 depletion causes loss of temporal control of neural crest migration. A.
Schematic representation of splice-blocking morpholino targeting. B. Phf6 MO1 causes exon
skipping at exon 8. RT-PCR of whole cDNA isolated from whole-embryo injections of 4 µM of
morpholino, results in loss of 150 base pairs of Phf6 transcript. Representative image of two
experiments. C. Schematic representation of targeted injections for neural crest tracking. D. Neural
crest tracking experiments with GFP, or control, morpholino, Phf6 MO1 or MO2, or rescue of
Phf6 MO1 using PHF6A mRNA. MOs injected at 4 µM, mRNA at 300 pg. E. In situ hybridization
of Twist1 on morphant embryos, MOs injected at 4 µM at NF4. Area stained by Twist1 quantified
and normalized to the control, contralateral side of the embryo in F. *, p<0.05; **, p<0.01 using a
two-tailed student’s T-test. Error bars indicate SEM. G and H. In situ hybridization of NF27
embryos with neural crest markers Twist1 and Sox9. Black arrows indicate increased neural crest
migration caused by Phf6 MO1 at 4 µM. N values given for each individual experiment.

A
GFP MO Phf6 MO1 Phf6 MO2
Phf6 MO1
PHF6A mRNA
N=9/11 N=5/8 N=7/7
MO: 4 μM
iv
D
Phf6 pre-mRNA MO1 MO2
UTR
Coding Exon
Intron
C
i
iii
NF4 NF13
Phf6 MO1
Target Control
GFP MO Phf6 MO1
Twist1, NF 21
N=30
N=21 N=31
Target Control
Phf6 MO2
Target Control
E
COMO
SB1
SB2
SB3
0.0
0.5
1.0
1.5
2.0
Twist
Ratio of Stained area
Morpholino, 4uM
F
B Control MO Phf6 MO1 Phf6 MO2
- 350 bp
- 200 bp
Twist1
Ratio of Stained Area
Twist1, NF27
Target Control
GFP MO Phf6 MO1
Control Target
G
Sox9, NF27
GFP MO Phf6 MO1
Target Control Target Control
H
* **
i ii
iii
iv
N=15/21
N=12/14
NF27
89
arch neural crest migration, where crest migrating from the frontonasal process or the first arch
have migrated completely to the ventral side of the embryo and the second arch lags behind and
has not yet reached the ventral side, and the third and 4
th
arch lag behind the second (Fig. 4.1D,
panel i). When we observe the migration of Phf6 morphant embryos in either MO1 or MO2, the
majority of the embryos show complete loss of stagger of the more posterior pharyngeal arches,
whereby even the 3
rd
and 4
th
arches have reached the ventral side of the embryo in MO1 (Fig. 4.1D,
panel ii, white arrow) or a turning towards the posterior of the third arch in MO2 (Fig. 4.1D, panel
iii, white arrow). The posterior arch stagger is completely restored when MO1 is co-injected with
humanized PHF6 mRNA (Fig. 4.1D, panel iv), indicating that normalization of functional PHF6
protein can rescue this defect.
I further analyzed this defect by examining the effects of Phf6 depletion on the expression
of neural crest markers Twist1 and Sox9 using in situ hybridization. Analysis of Twist1 espression
at NF21 showed a slight, but significant increase in the area marked by Twist1 at this early time
point (Fig. 4.1E, quantitated in F, p<0.05 and p<0.01 via two-tailed student’s T-test, respectively).
Close examination of the staining patterns revealed a slight blurring of the boundaries between
stained arch regions, particularly in the more posterior arches (data not shown), indicating the
possibility of a slight disorganization of migration at this time point. In situ hybridization of Twist1
and Sox9 showed a re-capitulation of the phenotype observed in Fig. 4.1D, where embryos targeted
with Phf6 MO1 showed a distinct loss of stagger in the posterior arches (Fig. 4.1G and H, black
arrows) when compared to GFP morphant embryos or the pattern on the contralateral, non-targeted
side. This loss of posterior arch stagger may indicate a loss of timing control at the point of neural
crest migration, particularly in light of the fact that Phf6 is not expressed in migratory neural crest
cells (Fig. 3.4E and F).
90
4.2.2. Ectopic expression of hPHF6 leads to reduction of neural crest migration.
To complement the depletion studies, we determined how overexpression of humanized
PHF6 mRNA targeted to Xenopus laevis neural plate border would affect neural crest and neuronal
tissue development. To this end, I microinjected 300 pg of PHF6 full-length mRNA along with
mCherry mRNA as a fluorescent marker using a similar strategy to that described in section 4.2.1.
A higher dose of mCherry mRNA was used as a control for excess mRNA presence, which can be
toxic to the cells. In this experiment, I chose to use mRNA to mark targeted cells due to the
theoretical discrepancy between dispersion rates of small molecular weight dextran and large
molecular weight mRNA in the yolk-filled Xenopus blastomeres. mCherry injected embryos
showed the same characteristic migration pattern at NF27, with the pharyngeal arches showing
staggered migration anterior to posterior (Fig. 4.2B, panel i). Embryos overexpressing PHF6
mRNA, on the other hand, showed a marked reduction of fluorescence within the pharyngeal
arches, particularly the more posterior arches (Fig. 4.2B, panel ii, hollow arrow). Interestingly, we
observed an increase in fluorescent signal within and nearby the neural tube region, and we
wondered if this was due to increase in fluorescent targeting of the neural tube region or if more
targeted neural plate border cells were failing to migrate and remaining within the neural tube
region. I hypothesized that overexpression of PHF6 caused failure of neural crest cells to migrate,
I would see an increase of Twist1 positive cells that did not migrate ventrally, or if overexpression
caused a shift towards neural precursors and away from specification of neural crest cells, that
there would be an increase in Sox2 positive cells that remained within the neural tube. In order to
differentiate between these possibilities, I performed in situ hybridization of Twist1 and Sox2 in
mid and late neurula embryos targeted as in Fig. 4.2A. In NF17 stage embryos, we observed a
slight decrease in Twist1 stained region with PHF6 overexpression (Fig. 4.2C,
91
Figure 4.2. Overexpression of PHF6 causes loss of neural crest migration. A. Schematic
representation of neural crest targeting for neural crest tracking experiments. B. Neural crest
tracking of embryos injected with mCherry control mRNA, or PHF6A mRNA, 300 pg. C. and D.
In situ hybridization of Twist1 at mid-neurula stage, NF17, and late neurula stage, NF19. Hollow
arrows indicate reduced Twist1 staining due to PHF6A overexpression. Stained area is quantitated
and normalized to stained area on the contralateral side in E. **, p<0.01 using a two-tailed
student’s T-test. F and G. Sox2 in situ hybridization on NF17 and NF19 embryos, * indicates
targeted side. Black arrows indicates increased Sox2 staining, which is quantified in H. *, p<0.05
using a two-tailed student’s T-test. Error bars indicate SEM.

PHF6A mCherry
mCherry PHF6A
Target Target Control Control
*
*
*
*
mCherry PHF6A
mCherry PHF6A
0.0
0.5
1.0
1.5
mRNA Injected
twist1
Ratio of distance from midline
Cherry PHF6A
0.0
0.5
1.0
1.5
2.0
sox2
Ratio of distance from midline
mRNA Injected
A
i
iii
NF4 NF13
B
i ii
Twist1, NF17
N=7
C
Target Target Control Control
mCherry PHF6A
N=5
Twist1, NF19
D
E
mCherry PHF6A
**
Twist1
Ratio of Stained Area
Sox2, NF17
F
mCherry PHF6A
Sox2, NF19
mCherry PHF6A
Sox2
Ratio of Distance from Midline
H G
*
NF27
92
hollow arrow), as compared to control embryos and the contralateral side of the embryo, that
correlated to a very slight increase in the Sox2 stained region in similarly targeted embryos (Fig.
4.2F, black arrow). This indicated that there may be some shift towards neural precursors over
neural crest cells upon PHF6 overexpression. In NF19 stage embryos, we observed a complete
lack of migration of the PHF6 overexpressing Twist1 positive cells (Fig. 4.2D, hollow arrow) as
compared with control embryos and the contralateral side of the embryo. The reduction of Twist1
positive area is statistically significant (Fig. 4.2E, p<0.01 in two-tailed student’s T-test), as is the
continued subtle increase in the Sox2 stained region upon PHF6 overexpression (Fig. 4.2G,
quantified in H, p<0.05 in two-tailed student’s T-test). This signifies that there is some contribution
of both facets of our hypothesis, that there is a slight shift towards neural precursors in PHF6
overexpression, but that non-migrating specified neural crest cells also persist around the neural
tube region.

4.2.3. Overexpression of an alternative isoform of PHF6 phenocopies knockdown of endogenous
Xenopus laevis Phf6.
In Chapter 3, we discussed the presence of a shorter isoform of PHF6 that is made by
human cells as well as Xenopus laevis cells. In order to determine the function of this transcript,
we created mRNA of the human form of this shorter isoform (Fig. 4.3A) for overexpression in the
same manner as in sections 4.2.1 and 4.2.2. Interestingly, when we overexpress this PHF6B
isoform, we observe a similar loss of stagger as with Phf6 depletion with our MO1 and MO2 (Fig.
4.1D, white arrows, Fig. 4.3B panel ii, white arrow). This observation is consistent with this
isoform acting as a dominant negative, since we did not perform any depletion of the endogenous

93
Figure 4.3. Overexpression of PHF6B isoform phenocopies depletion of endogenous Phf6. A.
Schematic representation of the alternate PHF6B isoform in human cells. B. Neural crest tracking
experiment, as in Figure 4.2. White arrow indicates increased neural crest migration caused by
overexpression of PHF6B at 300 pg. C. Twist1 In situ hybridization on embryos at NF17. Black
arrow indicates increased stained area, quantitated and normalized to the contralateral side in D.
*, p<0.05 using a two-tailed student’s T-test. F. In situ hybridization of Sox2 at NF19,
demonstrating neural tube closure defect. * indicates targeted side. Distance from midline
quantitated in E. Error bars indicate SEM.

mCherry PHF6B
0.0
0.5
1.0
1.5
2.0
2.5
mRNA Injected
twist1
Ratio of distance from midline
PHF6B (ΔF2)
*
*
Cherry PHF6B
0.0
0.5
1.0
1.5
2.0
sox2
Ratio of distance from midline
mRNA Injected
mCherry
PHF6b
1 2 3 4 5 6 7 8 9 10
pre-mRNA
PHD Finger1 PHD Finger2
mRNA
A B
i
ii
Target Control Target Control
mCherry PHF6B
Twist1, NF17
Twist1
Ratio of Stained Area
mCherry PHF6B mCherry PHF6B
Sox2
Ratio of Distance from Midline
mCherry PHF6B
Sox2, NF19
*
C
D
F
E
Figure 4.3
N=5/8
N=2/7
NF27
94
transcript. I therefore wanted to examine if this transcript promoted precocious formation or
migration of neural crest cells, perhaps at the expense of neural precursor cells, opposing the
phenotypes observed by overexpression of the PHF6A isoform. I found that there was a
statistically significant increase in the Twist1 stained region in mid and late neurula stage embryos
(Fig. 4.3C, quantitated in Fig. 4.3D, p<0.05 in a two-tailed student’s T-test). I also found that there
was a slight trend towards reduction of Sox2 stained region on the targeted side, and that there was
a small incidence of neural tube closure defects at late neurula stage (Fig. 4.3F, quantitated in Fig.
4.3E, p=0.08, n=2/7 for neural tube closure defects). This incidence of neural tube closure defects
may indicate that although the reduction in neural precursors is small, it may be sufficient to
perturb the integrity of the dorsal neural tube. This should be further investigated by following
these embryos to later stages to assay the formation of the central nervous system.

4.2.4. PHF6A overexpression cell autonomously represses formation of migrating neural crest
cells.
In order to resolve whether the effects observed by PHF6A overexpression within neural
crest are caused by migration defect within the crest themselves or caused by aberrant signaling
from nearby tissues, I performed neural crest transplantation experiments to determine the ability
of overexpressing neural crest cells to migrate within a wild type signaling environment. These
experiments were performed by E.C. Moran in the laboratory of Dr. Raymond Keller in
collaboration with Dr. Katherine Pfister at UVA. In these experiments, targeted embryos with
expression of the desired construct were isolated and stage-matched with wild-type, same clutch
embryos. The prospective cranial neural crest region of the targeted embryo was removed at NF15-
17 and transplanted into the corresponding location and orientation within the wild-type embryo
95
Figure 4.4. A. Schematic of neural crest transplantation. Morpholinos injected at 4 µM, mRNA
injected at 300 pg. B. Neural crest transplantation, typical embryological phenotype displayed. N
values below each panel state the number of embryos matching the depicted phenotype. C. Dorsal
to ventral migration quantified by the proportion of the embryo crossed by migratory neural crest
cells. *, p<0.05 by two-tailed student’s T-test. Error bars indicate SEM.

Figure 4.4
mCherry PHF6A Phf6 MO1 + PHF6A
GFP MO + PHF6A
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Cherry PHF6A GFP MO +
PHF6A
PHF6 MO +
PHF6A
Dorsal to Ventral Migration
*
*
NS
Donor WT Host
NF4 NF15-18 NF26 Stage Match
MO: 4 μM
mRNA: 300 pg
A
B
i ii iii iv
C
N=9/9
NF27
N=5/6 N=6/8
96
(Fig. 4.4A). The transplanted embryos were then allowed to heal and the migration of the neural
crest from the dorsal side of the embryo to the ventral side was followed through tailbud stages
and imaged at NF27-28. As expected, wild-type neural crest marked with mCherry mRNA formed
the distinct pharyngeal arch staggered pattern of migration after transplantation (Fig. 4.4B, panel
i), and the second arch migrated approximately 70% across the embryo by this stage (Fig. 4.4C,
first bar). PHF6A overexpressing embryos, however, showed almost no migration from the
transplantation site to the ventral side of the embryo (Fig. 4.4B, panel ii), which resulted in a
statistically significant reduction in the dorsal to ventral migration (Fig. 4.4C, second bar, p<0.05
by two-tailed student’s T-test). Similarly, when this overexpression was paired with a non-
targeting GFP morpholino, there was the same severe reduction in dorsal to ventral migration of
transplanted neural crest (Fig. 4.4B, panel iii, and Fig. 4.4C, third bar, p<0.05 by two-tailed
student’s T-test). The wild-type pattern of migration was restored, however, when the
overexpression of PHF6A was paired with the depletion of endogenous Phf6 through use of MO1
at 4 µM (Fig. 4.4B, panel iv, and Fig. 4.4C, fourth bar, p=0.57 in a two-tailed student’s T-test
compared to mCherry).
In order to resolve whether the signaling from the surrounding tissues is deficient in PHF6
overexpression embryos, I performed transplantation experiments of wild-type prospective cranial
neural crest marked with GFP mRNA into embryos where either mCherry or PHF6A was
overexpressed on one side of the embryo (Schematic in Fig. 4.5A). I then followed the migration
of the neural crest through the embryo to NF27, whereby we imaged and determined the proportion
of the embryo traveled by the migrating neural crest. I found that wild-type neural crest
transplanted into PHF6A overexpressing embryos showed no migration defect, and were able to
migrate statistically indistinguishably from wild-type neural crest cells in an mCherry
97
Figure 4.5. A. Schematic representation of converse transplantation experiment. mRNAs injected
at 300 pg. B. Typical migration patterns of wild-type neural crest migrating within embryos of
wild-type, or PHF6A overexpression. N values below show number of successful transplants, all
had a similar phenotype. C. Dorsal-to-ventral migration of neural crest cells. No significant
differences. Error bars indicate SEM.

Figure 4.5
0
0.2
0.4
0.6
0.8
1
Cherry PHF6A Phf6 MO
Dorsal to Ventral Migration
NS
NF2 NF15-18 NF26
Donor Host
Stage Match
mCherry
PHF6A GFP
GFP
N=5
A
B
C
i ii
N=3
NF27
98
overexpressing background (Fig. 4.5B and C, p=0.82 by two-tailed student’s T-test). I therefore
concluded that the observed defect in neural crest migration caused by PHF6A overexpression
occurs in a tissue-autonomous manner.

4.2.5. Loss of timing control caused by Phf6 depletion occurs independent of signaling
background.
In order to determine whether the loss of timing control observed in Phf6 morphant
embryos occurred in a tissue-autonomous fashion, I performed transplantation experiments similar
to those found in Figure 4.4 (Schematic in Fig. 4.6A). I observed a consistent neural crest migration
pattern with a typical staggering of the pharyngeal arches when neural crest injected with non-
targeting GFP morpholino were transplanted to wild type embryos (Fig. 4.6B, panel i). Consistent
with the phenotypes we observed in the neural crest tracking experiments, I found that neural crest
targeted with Phf6 MO1 lost the posterior-arch staggered migration pattern (Fig. 4.6 C, panel ii).
I quantified this observation by measuring the distance migrated by the second arch and the
distance migrated by the fourth arch, and took the difference between these two measurements to
create a stagger index, expressed in a fraction of the embryo (Fig. 4.6C). There was a statistically
significant difference between the stagger index on Phf6 morphant embryos as compared with
GFP morphant embryos (Fig. 4.6C, first two bars, p=0.008 by two-tailed student’s T-test).
Embryos targeted with Phf6 MO1 and GFP mRNA together showed the same loss of stagger
phenotype (Fig. 4.6B panel iii), however, no statistics could be performed here as the N value was
insufficient. Full rescue was achieved by co-injection of Phf6 MO1 and PHF6A mRNA, as in Fig.
4.4B panel iv (Fig. 4.6B panel iv). Due to the consistency of this loss of timing control

99
Figure 4.6. A. Schematic of neural crest transplantation experiment. B. Characteristic phenotype
of neural crest in each condition. N values indicate number of embryos transplanted. C. Stagger
quantified between the second and fourth pharyngeal arches, expressed as a percentage of the
embryo. **, p<0.01 by two-tailed student’s T-test. Error bars indicate SEM.

Figure 4.6
GFP MO Phf6 MO Phf6 MO + GFP Phf6 MO + PHF6A
0
0.05
0.1
0.15
0.2
0.25
0.3
GFP MO Phf6 MO Phf6 MO + GFP Phf6 MO +
PHF6A
Fraction of embryo
Stagger between Second and Fourth Arch
**
Donor WT Host
NF4 NF15-18 NF26 Stage Match
MO: 4 μM
mRNA: 300 pg
A
B
C
i ii iii iv
NF27
N=1 N=6/8 N=5
100
phenotype within multiple experimental conditions, we concluded that this phenotype is
independent of signaling environment.
In order to determine complete tissue-autonomous nature, the converse experiment where
expression of Phf6 morpholino in half of the embryo and transplant of wild-type neural crest into
this background would need to be performed. However, when performing expression of Phf6
morpholino into half of the embryo we observed a significant increase in gastrulation defects as
well as an overall loss of neural crest migration; this is further detailed in Chapter 6. Therefore, at
this time we cannot conclude that this phenotype is completely cell-autonomous.
In order to observe the migration of transplanted neural crest through the embryo, we took
a course transverse section of the transplanted embryo and stained these sections using Phalloidin-
Alexa 488. The transplanted neural crest was dually marked with a small amount of mCherry
mRNA and fix-safe Ruby Dextran in order to facilitate the visualization of transplanted neural
crest cells after fixation. Phalloidin was used to facilitate identification of specific structures within
the embryo, as cells attached to basement membranes express a high level of actin along the surface
of these membranes. We can therefore visually orient via definition of the neural tube (Fig. 4.7,
long-short dashes), the notochord (Fig. 4.7, short dashes), and the archenteron roof (Fig. 4.7, long
dashes). In these sections, we can observe that both the control embryos and the Phf6 MO rescued
with PHF6A mRNA show a fully intact, mostly symmetrical neural tube, and a stream of red neural
crest migrating through the tissue layers and stopping at the roof of the archenteron (Fig. 4.7A and
C). Phf6 morphant transplant embryos, on the other hand, show a thinning of the neural tube on
the transplanted side (Fig. 4.7B, indicated by the arrow), and the neural crest appear to have
migrated through the somite region of the embryo (Fig. 4.7B, bright green region, right side) and
101
Figure 4.7. Sectioned and Phalloidin stained neural crest transplant embryos (A-C). Neural tube
is outlined with short-long dashes, notochord is outlined with short dashes, archenteron roof
outlined with long dashes. Arrow indicates origin of transplant.

Control Phf6 MO Phf6 MO
+ PHF6A
Figure 4.7
A
B
C
Notochord
Archenteron roof
Neural Tube
Transplant origin
102
the stream of red neural crest cells stretch down along side of the archenteron (Fig. 4.7B, lower
right side). This may be indicative of increased invasive behavior of these cells.

4.2.6. Phf6 expression affects the cellular morphology and migratory potential of individual neural
crest cells.
Since Phf6-dependent phenotypes were determined to be cell-autonomous, we wanted to
resolve whether there were specific differences in the cellular morphologies of these cells that
could be central to the migratory differences that we observed. Xenopus embryos provide a
mechanism for investigating individual neural crest cell dynamics, as the prospective neural crest
can be explanted onto fibronectin-coated coverslips and the neural crest cells will migrate out from
the explant as they would in vivo. Embryos were injected at NF4 into one of the dorsal animal
blastomeres, and explanted at NF15-17 (Fig. 4.8G). These explants were allowed to migrate for
24 hours, and were fixed and stained using Phalloidin to highlight the actin structures within the
cells. Imaging at 20x using confocal microscopy, we observed that explants targeted with Phf6
MO1 showed significantly elongated neural crest cells when compared with those targeted with
control morpholino. Control neural crest cells averaged between 5-10 µm in length after exiting
the region of the explant, and gradually elongated to 15-20 µm in length with inherent
directionality and lammelipodial projections extending away from the central mass of the explant
(Fig. 4.8A, scale bar 30 µm). Phf6 MO1 cells, on the other hand, showed an elongated pattern
immediately upon exiting the central mass of the explant, with the average cell length of >30 µm
(Fig. 4.8B, scale bar 30 µm). This is quantified in terms of pixels at 20x magnification (Fig. 4.8C,
p<0.001 by two-tailed student’s T-test). Additionally, the Phf6 morphant neural crest cells show

103
Figure 4.8. Confocal images of neural crest explants stained with Phalloidin. A-C. Images at 20x
magnification of control (A.) or Phf6 depleted (B.) explants, scale bar 30 µm, cell length quantified
in C. ***, p<0.001 by two-tailed student’s T-test. D-F. Images at 63x magnification of control
(D.) or PHF6 overexpression (E.), scale bar 10 µm, distance migrated quantified in F. ***,
p<0.001 by two-tailed student’s T-test. Schematic of explantation, G. Error bars indicate SEM.

Figure 4.8
Fibronectin coverslip
NF4
NF16
Determining the role of PHF6 in epigenetic regulation of
Neural Crest Development
Erin Moran, Mallory Holland, Maria Nava, Ruchi Bajpai
Figure 1. Neural Crest formation and the tissues
and cell types that form from these cells.
6, 7
2.8 Mb
duplication
CHARGE like
ears
6.5 Mb
duplication
CHARGE BFLS
Dysmorphic,
reduced face
Broad,
Square Face
CHD7 PHF6
Syndrome
Facial
phenotype
External
ear (lobe)
Subcutaneous fat
WT
CHD7 +/-
Loss-of-function phenotypes
Symptoms
Figure 2. Contrasting phenotypes of CHD7 deficient
CHARGE patients and PHF6 deficient BFLS patients.
Thursday, October 18, 2012
Introduction
Aim 2. Determine Chromatin Associating Characteristics of
PHF6
Aim 1. Identify PHF6 Expression Patterns in In vitro Neural
Crest Differentiation System
Aim 3. Investigate the Cellular and Developmental Phenotypes of
PHF6 deficiency
PHF6 is expressed in NCC and NPC populations
in our hESC culture system
Rosettes w/o
NCC
Rosettes
generating NCC
Expression of PHF6 in differentiating NCC
Figure 7. RNA-seq for PHF6 in differentiated Neural Rosettes and Neural Rosettes actively
generating Neural Crest Cells.
Thursday, October 18, 2012
Figure 5.
PHF6
DAPI
Figure 6. Human Embryonic Stem Cells Differentiated into Neural Rosettes with migrating Neural Crest Cells.
Immunofluorescence was performed with anti-PHF6 antibody and nuclei were stained using DAPI. The border of the
Neural Rosette is shown with the dotted line, and the direction of migrating Neural Crest is indicated by the arrow to the
right.
Embryonic Stem Cell Culture
-Differentiate Neural Crest Cells (NCC)
and Neural Precursor Cells (NPC) from
Human Embryonic Stem Cells using
specific balances of morphogenic factors
(See Figure 3)
-Perform RNA-seq on NCC and NPC
to characterize expression levels
-Perform Immunofluorescence on fixed
cells to determine which subset of
cells highly express PHF6
Embryonic Stem Cell Culture
-Differentiate NCC and NPC
-Perform Immunoprecipitation with
Mass Spectrometry Analysis and
Western Blot Analysis to study
binding partners of PHF6
-Perform ChIP-Seq to determine
Global Chromatin binding locations
In vitro Binding Assays
-Clone different isoforms of PHF6 into a
bacterial expression vector containing
GST and 6xHis Tags, pGOOD6P
-Overexpress PHF6 isoforms in
bacteria, purify, and perform
binding arrays to determine which
chromatin marks to which PHF6
can bind
PHF6 forms a complex with CHD7 in these NCC
and NPC cells
PHF6 and CHD7 IP
Western Blot
PHF6
CHD7
IP: Input IgG PHF6
Cells: NP M NP M NP M
Figure 9. Cells grown in conditions to create neural precursor (NP)
or mixed neural precursor and neural crest (M) cell populations were
immunoprecipitated using IgG and PHF6, then western blotted for
PHF6 and CHD7. 1% of the input was run in the Input lanes.
PHF6 isoforms can be overexpressed in Arctic
Express bacteria

PHF6a PHF6b PHF6 PHDF2 GST-6xHis only
Time (hours) 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3
70 kDa
Figure 7. Coomassie stained SDS PAGE gels of PHF6 overexpression samples from Arctic Express
bacteria. PHF6 isoforms were cloned into the pGOOD6P GST-6xHis construct and used to
transform Arctic Express Bacteria, which were then grown and induced using 1 mM IPTG over a
timecourse of 3 hours at 37 degrees. Arrows indicate overexpression band.
Figure 6.
b.
c.
Figure 6. a. Strategy for design of PHF6 morpholinos. b. Schematic of
strategy for delivering morpholinos (MO) and mRNA to the perspective
neural plate and neural crest of the Xenopus embryo
1
. c. Results of CHD7
mRNA injection in tadpole stage embryos
1
. d. Results of CHD7 MO
knockdown in neurula stage embryos
1
.
d.
a.
Figure 8.
Xenopus laevis Embryology
-Determine the PHF6 knockdown
phenotype in this system using
Morpholino technology (see
Figure 4)
-Inject mRNA of the various PHF6
isoforms to investigate the
overexpression phenotype of
this protein
-The neural crest phenotype will
be quantified by performing in
situ hybridization using RNA
probes for neural crest
markers
Human Embryonic Stem Cell Culture
-Knock down PHF6 using pTRIPZ inducible shRNA against PHF6
-Study phenotype by differentiating knockdown hESC and observing NCC
migration, formation level
Future Directions
Determine Chromatin Binding Characteristics of PHF6
Study Overexpression and Knockdown Phenotypes of PHF6
References
1. Bajpai, R., Chen, D.A., Rada-Iglesias, A., Zhang, J., Xiong, Y ., Helms, J., Chang, C., Zhao, Y ., Swigut, T., and Wysocka, J., 2010. CHD7
cooperates with PBAF to control multipotent neural crest formation. Nature 463, 958-96
2. Etchevers, H. C., Amiel, J., Lyonnet, S., 2006. Molecular Basis of Human Neurocristopathies. Neural Crest Induction and Differentiation. Ed.
Jean-Pierre Saint-Jeannet. 1-22
3. V oss, A.K., Gamble, R., Collin, C., Shoubridge, C., Corbett, M., Gécz, J., Thomas, T. 2007. Protein and gene expression analysis of PHF6, the
gene mutated in the Börjeson-Forssman-Lehmann Syndrome of intellectual disability and obesity. Gene Expression Patterns 7 (2007) 858-871
4. Lower, K.M., et al. 2002. Mutations in PHF6 are associated with Börjeson-Forssman-Lehmann Syndrome. Nature Genetics 32, 661-665
5. Amouroux, C., et al. 2012. Duplication 8q12: confirmation of a novel recognizable phenotype with duane retraction syndrome and
developmental delay. European Journal of Human Genetics 20, 580-583
6. Laura S. Gammill & Marianne Bronner-Fraser. Neural Crest Specification: migrating into genomics. Nature Reviews Neuroscience 4,
795-805. October 2003
7. McKeown S, Lee V, bronner-Frasser M., Newgreen D., Farlie P. Sox10 Overexpression Induces Neural Crest-Like Cells from All
Dorsoventral Levels of the Neural Tube but Inhibits Differentiation. Dev Dyn. 2005 June; 233 (2): 430-44. 2005 Wiley-Liss Inc
Acknowledgements
Bajpai Lab
Dr. Ruchi Bajpai
Dr. Soma Samanta
Dr. Daniela Schmid
Dr. Richard Pelikan
Maria Nava
NIH T32 Cellular
Biochemical and Molecular
Biology
Investigate other proteins associated with the PHF6-CHD7
complex in Neural Crest and Neural Precursors
Mallory Holland
William Ciodza
Dr. Ankita Das
Yanke Lu
The Neural Crest is a population of cells that become specified along the border region between the neural plate and the surrounding ectoderm
2
. Neural Crest cells undergo an epithelial to mesenchymal transition, where

they extrapolate themselves from the epithelium where they once resided and migrate to distant locations (see Figure 1). These cells are especially intriguing in that they are able to regain multipotency at a stage in embryogenesis
where cell lineages are becoming progressively restricted.

These cells then go on to form the facial bones and much of the connective tissue in the face, as well as the peripheral nervous system, part of the outflow tract of the heart,
and integrate into almost all other organs
2
. Neurocristopathies, and particularly craniofacial abnormalities, are among the most prevalent birth defects, with craniofacial abnormalities affecting approximately 1 in 1000 children
born
2
. Neural Crest cell differentiation from neural ectoderm also exists in a fine balance with neural precursor cell differentiation from the same progenitors, and as such defects in neural development and defects in neural crest
development overlap
2
. Specification of cells towards a Neural Crest cell fate is tightly controlled transcriptionally through factors like Sox9 and Snail
2
.

CHD7, an ATP dependent chromatin remodeller, has been found to be essential to the formation of migratory neural crest cell populations in the developing embryo, a function which is carried out through neural crest specific
gene expression activation, such as activation of Sox9 and Snail, in a complex with PBAF
1
. In humans, null mutations in CHD7 are known to cause CHARGE syndrome. This syndrome has many characteristic symptoms, such as
dysmorphic, reduced facial bones and small external ear lobes.

Interestingly, these patients show no generalized intellectual disability.

PHF6 is a nuclear dual PHD finger protein (isoform a) with unknown function, but has been hypothesized to have a role in transcription due to the historical association of PHD finger proteins with transcription regulation. It
has been found to be highly expressed in the neural tube and the migrating neural crest of developing embryos
3
. Inactivating mutations of PHF6 in human patients are known to cause Börjeson-Forssman-Lehmann Syndrome
(BFLS), a syndrome which is characterized by broad, square faces, intellectual disability, enlarged, elongated external ear lobes, and obesity
4
. BFLS patients have remarkably contrasting phenotypes to CHARGE patients (see
Figure 2), and there have been reported cases where CHD7 duplication causes BFLS-like phenotypes
5
. Interestingly, PHF6 was identified in an immunoprecipitation with mass spectrometric analysis to physically

associate with CHD7 in a complex distinct from the CHD7-PBAF complex. Here, I present the data verifying the expression patterns of PHF6 in our cell culture system for producing Neural Crest Cells, as well as the verification
of the physical association of PHF6 with CHD7.

I

also will demonstrate the methods by which we will ascertain the precise role of PHF6 in neural crest development, as well as the function of the PHF6-

CHD7 complex.
neural crest cell
Neuroectoderm
How is cellular diversity
achieved ?
neural precursor cells
Embryonic Stem Cell
Neural Differentiation
Medium
Noggin Low levels of BMP
Figure 4. Strategy for differentiation of neural crest and neural
precursor cells.
Thursday, October 18, 2012
Figure 3.
Figure 4. RNA-seq data for the SOX9 and PHF6 locus in Neural Rosettes and Neural Rosettes
generating NCC cells differentiated from Human Embryonic Stem Cells. SOX9 marks Neural
Crest lineage, showing that PHF6 is downregulated in early NCC. R Bajpai, unpublished data.
Figure 5.
Epigenetic checks and balances promote control of neural crest migration.
Erin Moran
1,2
, Katherine Pfister
3
, Richard Pelikan
4
, Ruchi Bajpai
1,2
1
Center for Craniofacial Molecular Biology, Ostrow School of Dentistry, USC;
2
Department of Biochemistry, Keck School of Medicine, USC;

3
Department of Cell Biology, UVA;
4
Arthritis and Clinical Immunology Research Program, Oklahoma Medical Research Foundation

Abstract
40 kDa
10%
Input PHF6 IgG CHD7
IP
α-CHD7
α-PHF6
*NS-Heavy Chain
50 kDa
250 kDa
Western blot
5%
Input

IgG
Western blot
α-BAF180
α-BAF170
α-BRG1

PHF6

CHD7
Co-IP
α-CHD7
PHF6
α-PHF6

IgG
5%
Input
IgG
5%
Input CHD7
control IP
40 kDa
200 kDa
PHF6 and CHD7 form a mutually exclusive complex to CHD7-PBAF
α-H3K4Me1
α-H3K9Me3
α-PHF6
Western blot
IgG PHF6
10%
Input
IP
K4me1
40 kDa
15 kDa
15 kDa
*NS-Heavy Chain
α-PHF6
α-Biotin
10%
Input

Western Blot
H3(1-27) Peptide
10 kDa
40 kDa
PHF6 can physically associate with histones marked with K4me1
30x
Day7( NEC) Day9-14( NCC)
H3K4me1 P300 H3K4me1 P300
P300
H3K27Ac
H3K4Me1
No migratory NCC NEC with migratory NCC
-2 0 2 -2 0 2 -2 0 2 -2 0 2
The majority of NCC enhancers are
premarked with K4me1 in premigratory NCC
CHD7
-Can not remodel
nucleosomes due to
inaccessibility and histone
“tethering” by PHF6
-Area remains
inaccessible to
transcripVon
factors
NCC specific
TF
No NCC specific transcrip@on
PHF6 downregulaVon
PHF6 physically associates with NCC enhancers in premigratory NCC
NC
E

-1.5Kb
FLI1 locus
NC
E

-356Kb
TWIST1 locus
0
0.5
1
1.5
2
2.5
3
3.5
FLI1E
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
TWIST1E
PHF6 ChIP
0
2
4
6
8
10
12
SOX9E SOX9P
Percent Input
0
10
20
30
40
50
60
70
80
FLI1E
0
2
4
6
8
10
12
TWIST1E
H3K4me1 ChIP
NC
E

-251Kb
SOX9 locus
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
SOX9E SOX9P
Percent Input
Neuroectoderm
Pre-migratory
NCC
Mature NCC
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
SOX9E SOX9P
CHD7 ChIP
H3K4me1 ChIP
PHF6 ChIP
Phf6 Twist Sox2
NF17
Early Neurula
Pre-migratory
NCC
NF19
Mid Neurula
Pre-migratory and
migratory NCC
NF24
Early Tailbud
Migrating NCC
n=32
n=45
n=31
Neural Tube Cranial Neural Crest Cranial Placode
NF2 NF15-18
NF26
Donor Host
Stage Match
mCherry
PHF6
GFP GFP
mCherry PHF6 Phf6 MO + PHF6 GFP MO + PHF6
0
0.2
0.4
0.6
0.8
Cherry PHF6A GFP MO + PHF6A PHF6 MO +
PHF6A
Maximum Dorsal to Ventral
Migra@on
*
*
NS
N>5 per sample
GFP MO
Phf6 MO Phf6 MO + GFP Phf6 MO + PHF6
0
0.05
0.1
0.15
0.2
0.25
0.3
GFP MO Phf6 MO Phf6 MO + GFP Phf6 MO + PHF6A
Frac@on of embryo
Stagger between First and Third Arch
**
NS
Donor
WT Host
NF15-18
NF26
Stage Match
Figure 1. A. Co-immunoprecipitaVon of CHD7 with PHF6, performed in Neuroectodermal cells. B. Mutual
Co-immunoprecipitaVon of CHD7 and PHF6. C. Co-immunoprecipitaVon of PBAF complex members
BRG1, BAF180, and BAF170 with CHD7, but not PHF6.
A.
A.
B.
B.
C.
Figure 2. A. Co-immunoprecipitaVon of PHF6 with histones marked with H3K4me1, but not marked with
heterochromaVn modificaVon H3K9me3. B. Binding assay of N-terminal H3 pepVdes marked with
K4me3, K4me1, and unmodified with endogenous PHF6.
Figure 3. A.i. SOX9 locus diagram. ii. PHF6 ChIP at the SOX9 locus. iii. H3K4me1 ChIP at the SOX9 locus. iv.
CHD7 ChIP at the SOX9 locus, data from Bajpai et al., 2010. B. i. Diagram of the FLI1 locus. ii. PHF6 ChIP at
the FLI1 locus. iii. H3K4me1 ChIP at the FLI1 locus. C. i. Diagram of the TWIST1 locus. ii. PHF6 ChIP at the
TWIST1 locus. iii. H3K4me1 ChIP at the TWIST1 locus.
A. i.
ii.
iii.
iv.
PHF6 ChIP
H3K4me1 ChIP
B. i.
ii.
iii.
iii.
ii.
C. i.
Figure 4. Heatmap of H3K4me1 and p300 at the 3454 Neural Crest
specific enhancers at Day 7 of differenVaVon, represenVng
neurectodermal and premigratory neural crest cells, and day 9-14 of
differenVaVon, represenVng premigratory and migratory neural
crest cells.
H2B
H2A
H3
H4
CHD7
NCC specific
TF
Distal NCC transcripVon
PHF6 is highly expressed in pre-migratory neural crest but not
migraVng neural crest
NEC
NCC
Figure 5. A. In situ hybridizaVon at NF17 of Sox2 (ii.) Phf6 (iii.) and Twist (iv.) diagram of neural tube,
cranial neural crest, and cranial placode at the same stage (i.). B. In situ hybridizaVon of the same 3
mRNAs at NF19. C. In situ hybridizaVon of the same 3 mRNAs at NF24.
Figure 6. A. GFP Morpholino (non-targeVng) control, injected at 1/8 cell into one of the dorsal animal
blastomeres then followed to NF26 to view the migraVon paierns. B. and C. Two separate Phf6
morpholinos, injected the same way as control. Note the loss of stagger of the posterior pharyngeal
arches. D. Rescue of Phf6 MO1 using human PHF6 mRNA, longest isoform. E. mCherry mRNA, injected as
before (mRNA overexpression control). F. PHF6 mRNA overexpression and G. PHF6-ΔF2 mRNA, lacking
the second PHD domain, overexpression. Note the opposing phenotypes using these two mRNAs,
suggesVng the potenVal dominant negaVve funcVon of PHF6-ΔF2.
Figure 8. A. Diagram of how the transplantaVon experiment is performed. B-D. TransplantaVon of neural crest overexpressing
mCherry mRNA (B.), PHF6 mRNA (C.), PHF6 mRNA with a GFP (non-targeVng) morpholino (mock rescue, D.) or PHF6 mRNA with Phf6
morpholino 1 ( E.). F. QuanVtaVon of the maximum dorsal to ventral migraVon of the explant, quanVtated by ImageJ. G-H.
TransplantaVon of neural crest targeted with GFP (non-targeVng) morpholino ( G.), Phf6 morpholino 1 ( H.), Phf6 morpholino 1
overexpressing GFP mRNA (mock rescue, I.) or Phf6 morpholino 1 overexpressing human PHF6 mRNA (J.). K. QuanVtaVon of the
stagger between the first and third arch, performed by quanVtaVon of the distance migrated by the first arch subtracVng the
distance migrated by the third arch. Phf6 knockdown transplants show a significant loss of stagger of the pharyngeal arches.
Figure 9. A. Diagram of how the transplantaVon of wild-type neural
crest to mutant background is performed. This experiment is designed
to control for changes in signaling to the neural crest due to
overexpression of PHF6. B. Wild-type neural crest cells marked by GFP
are transplanted to embryos overexpressing mCherry. C. Wild-type
neural crest cells, marked by GFP, are transplanted to embryos
overexpressing PHF6. Note the normal migraVon paierns of these
neural crest cells.
Figure 11. A. Diagram of how the experiment is performed. B. Embryos are targeted by Chd7 Morpholino, as published in
Bajpai et al., 2010. C. Embryos are targeted by Chd7 morpholino and overexpressing PHF6 mRNA. D. Embryos are
targeted by Chd7 morpholino and overexpressing PHF6-ΔF2 mRNA. Note the rescue of neural crest migraVng through the
pharyngeal arches. E. Embryos are targeted by Chd7 morpholino and overexpressing PHF6-ΔF1 mRNA. F. QuanVtaVon of
the percent of embryos with no Chd7 MO associated defect, some Chd7 MO associated defect, and complete Chd7 MO
associated defect.
A. i. ii. iii. iv.
iv.
iv.
iii.
iii.
ii.
ii.
B. i.
C. i.
Part I. Biochemical analysis of PHF6 and its interac@on with CHD7
Part II. Expression PaWern of Phf6
Acknowledgements and Funding
Funding: NIH T32 Training Grant (PI: Stallcup) Cellular, Biochemical and Molecular Biology, NIH NIDCR F31
DE024688
Collaborators:
Ray Keller, Katherine Pfister (UVA)
Marianne Bronner, Shuyi Nie (Cal Tech)
MaW Lee, Susan Smith (USC)
Peggy Farnham

Bajpai Lab:
Ruchi Bajpai
Jennifer Oki Leslie Dominguez
Bajpai Lab Former Members
Richard Pelikan
Philbert Mach Maria Nava
USC Stem Cell Core, USC Ins@tute for Gene@c Medicine, USC Center for Craniofacial Molecular Biology, USC
Department of Biochemistry and Molecular Biology, Gene@c Molecular Cellular Biology Graduate Program
Part III. In vivo effect of Phf6 on neural crest migra@on
Effect of PHF6 overexpression on Twist expression paiern
mCherry PHF6
Injected Injected Control Control
NF17
NF19
Twist
mCherry PHF6A
0.0
0.5
1.0
1.5
mRNA Injected
twist1
Ratio of distance from midline
**
Injected/Control
Twist
Ra@o of stained area
mCherry PHF6-ΔF2
Injected
Injected
Control
Control
NF17
NF19
Twist
mCherry PHF6B
0.0
0.5
1.0
1.5
2.0
2.5
mRNA Injected
twist1
Ratio of distance from midline
*
Twist
Ra@o of stained area
Injected/Control
mCherry
PHF6-ΔF2
mRNA injected
mCherry PHF6
Effect of Phf6 on neural crest migraVon in situ
GFP MO Phf6 MO1 Phf6 MO2
Phf6 MO1
PHF6 mRNA
mCherry mRNA PHF6 mRNA
PHF6-ΔF2 mRNA
NF4
NF26
A. B.
C.
D.
E.
F.
G.
A. B.
C. D.
Figure 7. A. In situ hybridizaVon of Twist in embryos overexpressing mCherry mRNA on the injected side,
or PHF6 mRNA on the injected side. B. quanVtaVon of the raVo of stained areas on the injected side vs.
control side. Areas were picked up un-aided by ImageJ sooware. C. In situ hybridizaVon of Twist in
embryos overexpressing mCherry mRNA on the injected side, or PHF6-ΔF2 mRNA on the injected side. D.
QuanVtaVon of the raVo of stained areas, as in B.

Neural crest transplantaVon experiments show that Phf6 affects are cell-autonomous
A.
B. C. D. E.
F.
G. H. I.
J.
K.
A.
B.
C.
Chd7 MO+PHF6-ΔF2
Chd7 MO
Chd7 MO+PHF6-ΔF1
Defect
Rescue
Chd7 MO+PHF6
**
*
Percent embryos
100%
80%
60%
40%
20%
0%
A.
B. C.
D.
E.
F.
PHF6-ΔF2 mRNA rescues Chd7 morpholino phenotype
0
50
100
150
Control MO Phf6 MO
Pixels at 20x
Average NC Cell Length
***
GFP MO Phf6 MO
Top scale bar: 30 um; Boiom scale bar: 10 um
mCherry PHF6
0
50
100
150
200
250
mCherry PHF6
Pixels at 63x
Average Distance Migrated
***
0
5
10
15
20
25
30
Control Phf6 MO PHF6A
Pixels/second
Average Velocity
***
**
A. B.
D. E.
C.
F.
G.
Figure 10. A-C. 20x maximum intensity projecVon of neural crest explant on fibronecVn, injected
with GFP MO (A.), Phf6 MO (B.), quanVtaVon of cell size (C.). D-F. 63x maximum intensity
projecVon of neural crest explants on fibronecVn, injected with mCherry (D.) and PHF6 (E.),
quanVtaVon of distance migrated by NCC (F.). G. Average velocity of cells in explant, via 30 minute
confocal movie, data not shown.
FN coverslip
Neural crest explants show effect of Phf6 on individual cell dynamics
Determining the role of PHF6 in epigenetic regulation of
Neural Crest Development
Erin Moran, Mallory Holland, Maria Nava, Ruchi Bajpai
Figure 1. Neural Crest formation and the tissues
and cell types that form from these cells.
6, 7
2.8 Mb
duplication
CHARGE like
ears
6.5 Mb
duplication
CHARGE BFLS
Dysmorphic,
reduced face
Broad,
Square Face
CHD7 PHF6
Syndrome
Facial
phenotype
External
ear (lobe)
Subcutaneous fat
WT
CHD7 +/-
Loss-of-function phenotypes
Symptoms
Figure 2. Contrasting phenotypes of CHD7 deficient
CHARGE patients and PHF6 deficient BFLS patients.
Thursday, October 18, 2012
Introduction
Aim 2. Determine Chromatin Associating Characteristics of
PHF6
Aim 1. Identify PHF6 Expression Patterns in In vitro Neural
Crest Differentiation System
Aim 3. Investigate the Cellular and Developmental Phenotypes of
PHF6 deficiency
PHF6 is expressed in NCC and NPC populations
in our hESC culture system
Rosettes w/o
NCC
Rosettes
generating NCC
Expression of PHF6 in differentiating NCC
Figure 7. RNA-seq for PHF6 in differentiated Neural Rosettes and Neural Rosettes actively
generating Neural Crest Cells.
Thursday, October 18, 2012
Figure 5.
PHF6
DAPI
Figure 6. Human Embryonic Stem Cells Differentiated into Neural Rosettes with migrating Neural Crest Cells.
Immunofluorescence was performed with anti-PHF6 antibody and nuclei were stained using DAPI. The border of the
Neural Rosette is shown with the dotted line, and the direction of migrating Neural Crest is indicated by the arrow to the
right.
Embryonic Stem Cell Culture
-Differentiate Neural Crest Cells (NCC)
and Neural Precursor Cells (NPC) from
Human Embryonic Stem Cells using
specific balances of morphogenic factors
(See Figure 3)
-Perform RNA-seq on NCC and NPC
to characterize expression levels
-Perform Immunofluorescence on fixed
cells to determine which subset of
cells highly express PHF6
Embryonic Stem Cell Culture
-Differentiate NCC and NPC
-Perform Immunoprecipitation with
Mass Spectrometry Analysis and
Western Blot Analysis to study
binding partners of PHF6
-Perform ChIP-Seq to determine
Global Chromatin binding locations
In vitro Binding Assays
-Clone different isoforms of PHF6 into a
bacterial expression vector containing
GST and 6xHis Tags, pGOOD6P
-Overexpress PHF6 isoforms in
bacteria, purify, and perform
binding arrays to determine which
chromatin marks to which PHF6
can bind
PHF6 forms a complex with CHD7 in these NCC
and NPC cells
PHF6 and CHD7 IP
Western Blot
PHF6
CHD7
IP: Input IgG PHF6
Cells: NP M NP M NP M
Figure 9. Cells grown in conditions to create neural precursor (NP)
or mixed neural precursor and neural crest (M) cell populations were
immunoprecipitated using IgG and PHF6, then western blotted for
PHF6 and CHD7. 1% of the input was run in the Input lanes.
PHF6 isoforms can be overexpressed in Arctic
Express bacteria

PHF6a PHF6b PHF6 PHDF2 GST-6xHis only
Time (hours) 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3
70 kDa
Figure 7. Coomassie stained SDS PAGE gels of PHF6 overexpression samples from Arctic Express
bacteria. PHF6 isoforms were cloned into the pGOOD6P GST-6xHis construct and used to
transform Arctic Express Bacteria, which were then grown and induced using 1 mM IPTG over a
timecourse of 3 hours at 37 degrees. Arrows indicate overexpression band.
Figure 6.
b.
c.
Figure 6. a. Strategy for design of PHF6 morpholinos. b. Schematic of
strategy for delivering morpholinos (MO) and mRNA to the perspective
neural plate and neural crest of the Xenopus embryo
1
. c. Results of CHD7
mRNA injection in tadpole stage embryos
1
. d. Results of CHD7 MO
knockdown in neurula stage embryos
1
.
d.
a.
Figure 8.
Xenopus laevis Embryology
-Determine the PHF6 knockdown
phenotype in this system using
Morpholino technology (see
Figure 4)
-Inject mRNA of the various PHF6
isoforms to investigate the
overexpression phenotype of
this protein
-The neural crest phenotype will
be quantified by performing in
situ hybridization using RNA
probes for neural crest
markers
Human Embryonic Stem Cell Culture
-Knock down PHF6 using pTRIPZ inducible shRNA against PHF6
-Study phenotype by differentiating knockdown hESC and observing NCC
migration, formation level
Future Directions
Determine Chromatin Binding Characteristics of PHF6
Study Overexpression and Knockdown Phenotypes of PHF6
References
1. Bajpai, R., Chen, D.A., Rada-Iglesias, A., Zhang, J., Xiong, Y ., Helms, J., Chang, C., Zhao, Y ., Swigut, T., and Wysocka, J., 2010. CHD7
cooperates with PBAF to control multipotent neural crest formation. Nature 463, 958-96
2. Etchevers, H. C., Amiel, J., Lyonnet, S., 2006. Molecular Basis of Human Neurocristopathies. Neural Crest Induction and Differentiation. Ed.
Jean-Pierre Saint-Jeannet. 1-22
3. V oss, A.K., Gamble, R., Collin, C., Shoubridge, C., Corbett, M., Gécz, J., Thomas, T. 2007. Protein and gene expression analysis of PHF6, the
gene mutated in the Börjeson-Forssman-Lehmann Syndrome of intellectual disability and obesity. Gene Expression Patterns 7 (2007) 858-871
4. Lower, K.M., et al. 2002. Mutations in PHF6 are associated with Börjeson-Forssman-Lehmann Syndrome. Nature Genetics 32, 661-665
5. Amouroux, C., et al. 2012. Duplication 8q12: confirmation of a novel recognizable phenotype with duane retraction syndrome and
developmental delay. European Journal of Human Genetics 20, 580-583
6. Laura S. Gammill & Marianne Bronner-Fraser. Neural Crest Specification: migrating into genomics. Nature Reviews Neuroscience 4,
795-805. October 2003
7. McKeown S, Lee V, bronner-Frasser M., Newgreen D., Farlie P. Sox10 Overexpression Induces Neural Crest-Like Cells from All
Dorsoventral Levels of the Neural Tube but Inhibits Differentiation. Dev Dyn. 2005 June; 233 (2): 430-44. 2005 Wiley-Liss Inc
Acknowledgements
Bajpai Lab
Dr. Ruchi Bajpai
Dr. Soma Samanta
Dr. Daniela Schmid
Dr. Richard Pelikan
Maria Nava
NIH T32 Cellular
Biochemical and Molecular
Biology
Investigate other proteins associated with the PHF6-CHD7
complex in Neural Crest and Neural Precursors
Mallory Holland
William Ciodza
Dr. Ankita Das
Yanke Lu
The Neural Crest is a population of cells that become specified along the border region between the neural plate and the surrounding ectoderm
2
. Neural Crest cells undergo an epithelial to mesenchymal transition, where

they extrapolate themselves from the epithelium where they once resided and migrate to distant locations (see Figure 1). These cells are especially intriguing in that they are able to regain multipotency at a stage in embryogenesis
where cell lineages are becoming progressively restricted.

These cells then go on to form the facial bones and much of the connective tissue in the face, as well as the peripheral nervous system, part of the outflow tract of the heart,
and integrate into almost all other organs
2
. Neurocristopathies, and particularly craniofacial abnormalities, are among the most prevalent birth defects, with craniofacial abnormalities affecting approximately 1 in 1000 children
born
2
. Neural Crest cell differentiation from neural ectoderm also exists in a fine balance with neural precursor cell differentiation from the same progenitors, and as such defects in neural development and defects in neural crest
development overlap
2
. Specification of cells towards a Neural Crest cell fate is tightly controlled transcriptionally through factors like Sox9 and Snail
2
.

CHD7, an ATP dependent chromatin remodeller, has been found to be essential to the formation of migratory neural crest cell populations in the developing embryo, a function which is carried out through neural crest specific
gene expression activation, such as activation of Sox9 and Snail, in a complex with PBAF
1
. In humans, null mutations in CHD7 are known to cause CHARGE syndrome. This syndrome has many characteristic symptoms, such as
dysmorphic, reduced facial bones and small external ear lobes.

Interestingly, these patients show no generalized intellectual disability.

PHF6 is a nuclear dual PHD finger protein (isoform a) with unknown function, but has been hypothesized to have a role in transcription due to the historical association of PHD finger proteins with transcription regulation. It
has been found to be highly expressed in the neural tube and the migrating neural crest of developing embryos
3
. Inactivating mutations of PHF6 in human patients are known to cause Börjeson-Forssman-Lehmann Syndrome
(BFLS), a syndrome which is characterized by broad, square faces, intellectual disability, enlarged, elongated external ear lobes, and obesity
4
. BFLS patients have remarkably contrasting phenotypes to CHARGE patients (see
Figure 2), and there have been reported cases where CHD7 duplication causes BFLS-like phenotypes
5
. Interestingly, PHF6 was identified in an immunoprecipitation with mass spectrometric analysis to physically

associate with CHD7 in a complex distinct from the CHD7-PBAF complex. Here, I present the data verifying the expression patterns of PHF6 in our cell culture system for producing Neural Crest Cells, as well as the verification
of the physical association of PHF6 with CHD7.

I

also will demonstrate the methods by which we will ascertain the precise role of PHF6 in neural crest development, as well as the function of the PHF6-

CHD7 complex.
neural crest cell
Neuroectoderm
How is cellular diversity
achieved ?
neural precursor cells
Embryonic Stem Cell
Neural Differentiation
Medium
Noggin Low levels of BMP
Figure 4. Strategy for differentiation of neural crest and neural
precursor cells.
Thursday, October 18, 2012
Figure 3.
Figure 4. RNA-seq data for the SOX9 and PHF6 locus in Neural Rosettes and Neural Rosettes
generating NCC cells differentiated from Human Embryonic Stem Cells. SOX9 marks Neural
Crest lineage, showing that PHF6 is downregulated in early NCC. R Bajpai, unpublished data.
Figure 5.
Epigenetic checks and balances promote control of neural crest migration.
Erin Moran
1,2
, Katherine Pfister
3
, Richard Pelikan
4
, Ruchi Bajpai
1,2
1
Center for Craniofacial Molecular Biology, Ostrow School of Dentistry, USC;
2
Department of Biochemistry, Keck School of Medicine, USC;

3
Department of Cell Biology, UVA;
4
Arthritis and Clinical Immunology Research Program, Oklahoma Medical Research Foundation

Abstract
40 kDa
10%
Input PHF6 IgG CHD7
IP
α-CHD7
α-PHF6
*NS-Heavy Chain
50 kDa
250 kDa
Western blot
5%
Input

IgG
Western blot
α-BAF180
α-BAF170
α-BRG1

PHF6

CHD7
Co-IP
α-CHD7
PHF6
α-PHF6

IgG
5%
Input
IgG
5%
Input CHD7
control IP
40 kDa
200 kDa
PHF6 and CHD7 form a mutually exclusive complex to CHD7-PBAF
α-H3K4Me1
α-H3K9Me3
α-PHF6
Western blot
IgG PHF6
10%
Input
IP
K4me1
40 kDa
15 kDa
15 kDa
*NS-Heavy Chain
α-PHF6
α-Biotin
10%
Input

Western Blot
H3(1-27) Peptide
10 kDa
40 kDa
PHF6 can physically associate with histones marked with K4me1
30x
Day7( NEC) Day9-14( NCC)
H3K4me1 P300 H3K4me1 P300
P300
H3K27Ac
H3K4Me1
No migratory NCC NEC with migratory NCC
-2 0 2 -2 0 2 -2 0 2 -2 0 2
The majority of NCC enhancers are
premarked with K4me1 in premigratory NCC
CHD7
-Can not remodel
nucleosomes due to
inaccessibility and histone
“tethering” by PHF6
-Area remains
inaccessible to
transcripVon
factors
NCC specific
TF
No NCC specific transcrip@on
PHF6 downregulaVon
PHF6 physically associates with NCC enhancers in premigratory NCC
NC
E

-1.5Kb
FLI1 locus
NC
E

-356Kb
TWIST1 locus
0
0.5
1
1.5
2
2.5
3
3.5
FLI1E
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
TWIST1E
PHF6 ChIP
0
2
4
6
8
10
12
SOX9E SOX9P
Percent Input
0
10
20
30
40
50
60
70
80
FLI1E
0
2
4
6
8
10
12
TWIST1E
H3K4me1 ChIP
NC
E

-251Kb
SOX9 locus
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
SOX9E SOX9P
Percent Input
Neuroectoderm
Pre-migratory
NCC
Mature NCC
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
SOX9E SOX9P
CHD7 ChIP
H3K4me1 ChIP
PHF6 ChIP
Phf6 Twist Sox2
NF17
Early Neurula
Pre-migratory
NCC
NF19
Mid Neurula
Pre-migratory and
migratory NCC
NF24
Early Tailbud
Migrating NCC
n=32
n=45
n=31
Neural Tube Cranial Neural Crest Cranial Placode
NF2 NF15-18
NF26
Donor Host
Stage Match
mCherry
PHF6
GFP GFP
mCherry PHF6 Phf6 MO + PHF6 GFP MO + PHF6
0
0.2
0.4
0.6
0.8
Cherry PHF6A GFP MO + PHF6A PHF6 MO +
PHF6A
Maximum Dorsal to Ventral
Migra@on
*
*
NS
N>5 per sample
GFP MO
Phf6 MO Phf6 MO + GFP Phf6 MO + PHF6
0
0.05
0.1
0.15
0.2
0.25
0.3
GFP MO Phf6 MO Phf6 MO + GFP Phf6 MO + PHF6A
Frac@on of embryo
Stagger between First and Third Arch
**
NS
Donor
WT Host
NF15-18
NF26
Stage Match
Figure 1. A. Co-immunoprecipitaVon of CHD7 with PHF6, performed in Neuroectodermal cells. B. Mutual
Co-immunoprecipitaVon of CHD7 and PHF6. C. Co-immunoprecipitaVon of PBAF complex members
BRG1, BAF180, and BAF170 with CHD7, but not PHF6.
A.
A.
B.
B.
C.
Figure 2. A. Co-immunoprecipitaVon of PHF6 with histones marked with H3K4me1, but not marked with
heterochromaVn modificaVon H3K9me3. B. Binding assay of N-terminal H3 pepVdes marked with
K4me3, K4me1, and unmodified with endogenous PHF6.
Figure 3. A.i. SOX9 locus diagram. ii. PHF6 ChIP at the SOX9 locus. iii. H3K4me1 ChIP at the SOX9 locus. iv.
CHD7 ChIP at the SOX9 locus, data from Bajpai et al., 2010. B. i. Diagram of the FLI1 locus. ii. PHF6 ChIP at
the FLI1 locus. iii. H3K4me1 ChIP at the FLI1 locus. C. i. Diagram of the TWIST1 locus. ii. PHF6 ChIP at the
TWIST1 locus. iii. H3K4me1 ChIP at the TWIST1 locus.
A. i.
ii.
iii.
iv.
PHF6 ChIP
H3K4me1 ChIP
B. i.
ii.
iii.
iii.
ii.
C. i.
Figure 4. Heatmap of H3K4me1 and p300 at the 3454 Neural Crest
specific enhancers at Day 7 of differenVaVon, represenVng
neurectodermal and premigratory neural crest cells, and day 9-14 of
differenVaVon, represenVng premigratory and migratory neural
crest cells.
H2B
H2A
H3
H4
CHD7
NCC specific
TF
Distal NCC transcripVon
PHF6 is highly expressed in pre-migratory neural crest but not
migraVng neural crest
NEC
NCC
Figure 5. A. In situ hybridizaVon at NF17 of Sox2 (ii.) Phf6 (iii.) and Twist (iv.) diagram of neural tube,
cranial neural crest, and cranial placode at the same stage (i.). B. In situ hybridizaVon of the same 3
mRNAs at NF19. C. In situ hybridizaVon of the same 3 mRNAs at NF24.
Figure 6. A. GFP Morpholino (non-targeVng) control, injected at 1/8 cell into one of the dorsal animal
blastomeres then followed to NF26 to view the migraVon paierns. B. and C. Two separate Phf6
morpholinos, injected the same way as control. Note the loss of stagger of the posterior pharyngeal
arches. D. Rescue of Phf6 MO1 using human PHF6 mRNA, longest isoform. E. mCherry mRNA, injected as
before (mRNA overexpression control). F. PHF6 mRNA overexpression and G. PHF6-ΔF2 mRNA, lacking
the second PHD domain, overexpression. Note the opposing phenotypes using these two mRNAs,
suggesVng the potenVal dominant negaVve funcVon of PHF6-ΔF2.
Figure 8. A. Diagram of how the transplantaVon experiment is performed. B-D. TransplantaVon of neural crest overexpressing
mCherry mRNA (B.), PHF6 mRNA (C.), PHF6 mRNA with a GFP (non-targeVng) morpholino (mock rescue, D.) or PHF6 mRNA with Phf6
morpholino 1 ( E.). F. QuanVtaVon of the maximum dorsal to ventral migraVon of the explant, quanVtated by ImageJ. G-H.
TransplantaVon of neural crest targeted with GFP (non-targeVng) morpholino ( G.), Phf6 morpholino 1 ( H.), Phf6 morpholino 1
overexpressing GFP mRNA (mock rescue, I.) or Phf6 morpholino 1 overexpressing human PHF6 mRNA (J.). K. QuanVtaVon of the
stagger between the first and third arch, performed by quanVtaVon of the distance migrated by the first arch subtracVng the
distance migrated by the third arch. Phf6 knockdown transplants show a significant loss of stagger of the pharyngeal arches.
Figure 9. A. Diagram of how the transplantaVon of wild-type neural
crest to mutant background is performed. This experiment is designed
to control for changes in signaling to the neural crest due to
overexpression of PHF6. B. Wild-type neural crest cells marked by GFP
are transplanted to embryos overexpressing mCherry. C. Wild-type
neural crest cells, marked by GFP, are transplanted to embryos
overexpressing PHF6. Note the normal migraVon paierns of these
neural crest cells.
Figure 11. A. Diagram of how the experiment is performed. B. Embryos are targeted by Chd7 Morpholino, as published in
Bajpai et al., 2010. C. Embryos are targeted by Chd7 morpholino and overexpressing PHF6 mRNA. D. Embryos are
targeted by Chd7 morpholino and overexpressing PHF6-ΔF2 mRNA. Note the rescue of neural crest migraVng through the
pharyngeal arches. E. Embryos are targeted by Chd7 morpholino and overexpressing PHF6-ΔF1 mRNA. F. QuanVtaVon of
the percent of embryos with no Chd7 MO associated defect, some Chd7 MO associated defect, and complete Chd7 MO
associated defect.
A. i. ii. iii. iv.
iv.
iv.
iii.
iii.
ii.
ii.
B. i.
C. i.
Part I. Biochemical analysis of PHF6 and its interac@on with CHD7
Part II. Expression PaWern of Phf6
Acknowledgements and Funding
Funding: NIH T32 Training Grant (PI: Stallcup) Cellular, Biochemical and Molecular Biology, NIH NIDCR F31
DE024688
Collaborators:
Ray Keller, Katherine Pfister (UVA)
Marianne Bronner, Shuyi Nie (Cal Tech)
MaW Lee, Susan Smith (USC)
Peggy Farnham

Bajpai Lab:
Ruchi Bajpai
Jennifer Oki Leslie Dominguez
Bajpai Lab Former Members
Richard Pelikan
Philbert Mach Maria Nava
USC Stem Cell Core, USC Ins@tute for Gene@c Medicine, USC Center for Craniofacial Molecular Biology, USC
Department of Biochemistry and Molecular Biology, Gene@c Molecular Cellular Biology Graduate Program
Part III. In vivo effect of Phf6 on neural crest migra@on
Effect of PHF6 overexpression on Twist expression paiern
mCherry
PHF6
Injected Injected Control Control
NF17
NF19
Twist
mCherry PHF6A
0.0
0.5
1.0
1.5
mRNA Injected
twist1
Ratio of distance from midline
**
Injected/Control
Twist
Ra@o of stained area
mCherry PHF6-ΔF2
Injected
Injected
Control
Control
NF17
NF19
Twist
mCherry PHF6B
0.0
0.5
1.0
1.5
2.0
2.5
mRNA Injected
twist1
Ratio of distance from midline
*
Twist
Ra@o of stained area
Injected/Control
mCherry
PHF6-ΔF2
mRNA injected
mCherry PHF6
Effect of Phf6 on neural crest migraVon in situ
GFP MO Phf6 MO1 Phf6 MO2
Phf6 MO1
PHF6 mRNA
mCherry mRNA PHF6 mRNA
PHF6-ΔF2 mRNA
NF4
NF26
A. B.
C.
D.
E.
F.
G.
A. B.
C. D.
Figure 7. A. In situ hybridizaVon of Twist in embryos overexpressing mCherry mRNA on the injected side,
or PHF6 mRNA on the injected side. B. quanVtaVon of the raVo of stained areas on the injected side vs.
control side. Areas were picked up un-aided by ImageJ sooware. C. In situ hybridizaVon of Twist in
embryos overexpressing mCherry mRNA on the injected side, or PHF6-ΔF2 mRNA on the injected side. D.
QuanVtaVon of the raVo of stained areas, as in B.

Neural crest transplantaVon experiments show that Phf6 affects are cell-autonomous
A.
B. C. D. E.
F.
G. H. I.
J.
K.
A.
B.
C.
Chd7 MO+PHF6-ΔF2
Chd7 MO
Chd7 MO+PHF6-ΔF1
Defect
Rescue
Chd7 MO+PHF6
**
*
Percent embryos
100%
80%
60%
40%
20%
0%
A.
B. C.
D.
E.
F.
PHF6-ΔF2 mRNA rescues Chd7 morpholino phenotype
0
50
100
150
Control MO Phf6 MO
Pixels at 20x
Average NC Cell Length
***
GFP MO Phf6 MO
Top scale bar: 30 um; Boiom scale bar: 10 um
mCherry
PHF6
0
50
100
150
200
250
mCherry PHF6
Pixels at 63x
Average Distance Migrated
***
0
5
10
15
20
25
30
Control Phf6 MO PHF6A
Pixels/second
Average Velocity
***
**
A. B.
D. E.
C.
F.
G.
Figure 10. A-C. 20x maximum intensity projecVon of neural crest explant on fibronecVn, injected
with GFP MO (A.), Phf6 MO (B.), quanVtaVon of cell size (C.). D-F. 63x maximum intensity
projecVon of neural crest explants on fibronecVn, injected with mCherry (D.) and PHF6 (E.),
quanVtaVon of distance migrated by NCC (F.). G. Average velocity of cells in explant, via 30 minute
confocal movie, data not shown.
FN coverslip
Neural crest explants show effect of Phf6 on individual cell dynamics
Determining the role of PHF6 in epigenetic regulation of
Neural Crest Development
Erin Moran, Mallory Holland, Maria Nava, Ruchi Bajpai
Figure 1. Neural Crest formation and the tissues
and cell types that form from these cells.
6, 7
2.8 Mb
duplication
CHARGE like
ears
6.5 Mb
duplication
CHARGE BFLS
Dysmorphic,
reduced face
Broad,
Square Face
CHD7 PHF6
Syndrome
Facial
phenotype
External
ear (lobe)
Subcutaneous fat
WT
CHD7 +/-
Loss-of-function phenotypes
Symptoms
Figure 2. Contrasting phenotypes of CHD7 deficient
CHARGE patients and PHF6 deficient BFLS patients.
Thursday, October 18, 2012
Introduction
Aim 2. Determine Chromatin Associating Characteristics of
PHF6
Aim 1. Identify PHF6 Expression Patterns in In vitro Neural
Crest Differentiation System
Aim 3. Investigate the Cellular and Developmental Phenotypes of
PHF6 deficiency
PHF6 is expressed in NCC and NPC populations
in our hESC culture system
Rosettes w/o
NCC
Rosettes
generating NCC
Expression of PHF6 in differentiating NCC
Figure 7. RNA-seq for PHF6 in differentiated Neural Rosettes and Neural Rosettes actively
generating Neural Crest Cells.
Thursday, October 18, 2012
Figure 5.
PHF6
DAPI
Figure 6. Human Embryonic Stem Cells Differentiated into Neural Rosettes with migrating Neural Crest Cells.
Immunofluorescence was performed with anti-PHF6 antibody and nuclei were stained using DAPI. The border of the
Neural Rosette is shown with the dotted line, and the direction of migrating Neural Crest is indicated by the arrow to the
right.
Embryonic Stem Cell Culture
-Differentiate Neural Crest Cells (NCC)
and Neural Precursor Cells (NPC) from
Human Embryonic Stem Cells using
specific balances of morphogenic factors
(See Figure 3)
-Perform RNA-seq on NCC and NPC
to characterize expression levels
-Perform Immunofluorescence on fixed
cells to determine which subset of
cells highly express PHF6
Embryonic Stem Cell Culture
-Differentiate NCC and NPC
-Perform Immunoprecipitation with
Mass Spectrometry Analysis and
Western Blot Analysis to study
binding partners of PHF6
-Perform ChIP-Seq to determine
Global Chromatin binding locations
In vitro Binding Assays
-Clone different isoforms of PHF6 into a
bacterial expression vector containing
GST and 6xHis Tags, pGOOD6P
-Overexpress PHF6 isoforms in
bacteria, purify, and perform
binding arrays to determine which
chromatin marks to which PHF6
can bind
PHF6 forms a complex with CHD7 in these NCC
and NPC cells
PHF6 and CHD7 IP
Western Blot
PHF6
CHD7
IP: Input IgG PHF6
Cells: NP M NP M NP M
Figure 9. Cells grown in conditions to create neural precursor (NP)
or mixed neural precursor and neural crest (M) cell populations were
immunoprecipitated using IgG and PHF6, then western blotted for
PHF6 and CHD7. 1% of the input was run in the Input lanes.
PHF6 isoforms can be overexpressed in Arctic
Express bacteria

PHF6a PHF6b PHF6 PHDF2 GST-6xHis only
Time (hours) 0 1 2 3 0 1 2 3 0 1 2 3 0 1 2 3
70 kDa
Figure 7. Coomassie stained SDS PAGE gels of PHF6 overexpression samples from Arctic Express
bacteria. PHF6 isoforms were cloned into the pGOOD6P GST-6xHis construct and used to
transform Arctic Express Bacteria, which were then grown and induced using 1 mM IPTG over a
timecourse of 3 hours at 37 degrees. Arrows indicate overexpression band.
Figure 6.
b.
c.
Figure 6. a. Strategy for design of PHF6 morpholinos. b. Schematic of
strategy for delivering morpholinos (MO) and mRNA to the perspective
neural plate and neural crest of the Xenopus embryo
1
. c. Results of CHD7
mRNA injection in tadpole stage embryos
1
. d. Results of CHD7 MO
knockdown in neurula stage embryos
1
.
d.
a.
Figure 8.
Xenopus laevis Embryology
-Determine the PHF6 knockdown
phenotype in this system using
Morpholino technology (see
Figure 4)
-Inject mRNA of the various PHF6
isoforms to investigate the
overexpression phenotype of
this protein
-The neural crest phenotype will
be quantified by performing in
situ hybridization using RNA
probes for neural crest
markers
Human Embryonic Stem Cell Culture
-Knock down PHF6 using pTRIPZ inducible shRNA against PHF6
-Study phenotype by differentiating knockdown hESC and observing NCC
migration, formation level
Future Directions
Determine Chromatin Binding Characteristics of PHF6
Study Overexpression and Knockdown Phenotypes of PHF6
References
1. Bajpai, R., Chen, D.A., Rada-Iglesias, A., Zhang, J., Xiong, Y ., Helms, J., Chang, C., Zhao, Y ., Swigut, T., and Wysocka, J., 2010. CHD7
cooperates with PBAF to control multipotent neural crest formation. Nature 463, 958-96
2. Etchevers, H. C., Amiel, J., Lyonnet, S., 2006. Molecular Basis of Human Neurocristopathies. Neural Crest Induction and Differentiation. Ed.
Jean-Pierre Saint-Jeannet. 1-22
3. V oss, A.K., Gamble, R., Collin, C., Shoubridge, C., Corbett, M., Gécz, J., Thomas, T. 2007. Protein and gene expression analysis of PHF6, the
gene mutated in the Börjeson-Forssman-Lehmann Syndrome of intellectual disability and obesity. Gene Expression Patterns 7 (2007) 858-871
4. Lower, K.M., et al. 2002. Mutations in PHF6 are associated with Börjeson-Forssman-Lehmann Syndrome. Nature Genetics 32, 661-665
5. Amouroux, C., et al. 2012. Duplication 8q12: confirmation of a novel recognizable phenotype with duane retraction syndrome and
developmental delay. European Journal of Human Genetics 20, 580-583
6. Laura S. Gammill & Marianne Bronner-Fraser. Neural Crest Specification: migrating into genomics. Nature Reviews Neuroscience 4,
795-805. October 2003
7. McKeown S, Lee V, bronner-Frasser M., Newgreen D., Farlie P. Sox10 Overexpression Induces Neural Crest-Like Cells from All
Dorsoventral Levels of the Neural Tube but Inhibits Differentiation. Dev Dyn. 2005 June; 233 (2): 430-44. 2005 Wiley-Liss Inc
Acknowledgements
Bajpai Lab
Dr. Ruchi Bajpai
Dr. Soma Samanta
Dr. Daniela Schmid
Dr. Richard Pelikan
Maria Nava
NIH T32 Cellular
Biochemical and Molecular
Biology
Investigate other proteins associated with the PHF6-CHD7
complex in Neural Crest and Neural Precursors
Mallory Holland
William Ciodza
Dr. Ankita Das
Yanke Lu
The Neural Crest is a population of cells that become specified along the border region between the neural plate and the surrounding ectoderm
2
. Neural Crest cells undergo an epithelial to mesenchymal transition, where

they extrapolate themselves from the epithelium where they once resided and migrate to distant locations (see Figure 1). These cells are especially intriguing in that they are able to regain multipotency at a stage in embryogenesis
where cell lineages are becoming progressively restricted.

These cells then go on to form the facial bones and much of the connective tissue in the face, as well as the peripheral nervous system, part of the outflow tract of the heart,
and integrate into almost all other organs
2
. Neurocristopathies, and particularly craniofacial abnormalities, are among the most prevalent birth defects, with craniofacial abnormalities affecting approximately 1 in 1000 children
born
2
. Neural Crest cell differentiation from neural ectoderm also exists in a fine balance with neural precursor cell differentiation from the same progenitors, and as such defects in neural development and defects in neural crest
development overlap
2
. Specification of cells towards a Neural Crest cell fate is tightly controlled transcriptionally through factors like Sox9 and Snail
2
.

CHD7, an ATP dependent chromatin remodeller, has been found to be essential to the formation of migratory neural crest cell populations in the developing embryo, a function which is carried out through neural crest specific
gene expression activation, such as activation of Sox9 and Snail, in a complex with PBAF
1
. In humans, null mutations in CHD7 are known to cause CHARGE syndrome. This syndrome has many characteristic symptoms, such as
dysmorphic, reduced facial bones and small external ear lobes.

Interestingly, these patients show no generalized intellectual disability.

PHF6 is a nuclear dual PHD finger protein (isoform a) with unknown function, but has been hypothesized to have a role in transcription due to the historical association of PHD finger proteins with transcription regulation. It
has been found to be highly expressed in the neural tube and the migrating neural crest of developing embryos
3
. Inactivating mutations of PHF6 in human patients are known to cause Börjeson-Forssman-Lehmann Syndrome
(BFLS), a syndrome which is characterized by broad, square faces, intellectual disability, enlarged, elongated external ear lobes, and obesity
4
. BFLS patients have remarkably contrasting phenotypes to CHARGE patients (see
Figure 2), and there have been reported cases where CHD7 duplication causes BFLS-like phenotypes
5
. Interestingly, PHF6 was identified in an immunoprecipitation with mass spectrometric analysis to physically

associate with CHD7 in a complex distinct from the CHD7-PBAF complex. Here, I present the data verifying the expression patterns of PHF6 in our cell culture system for producing Neural Crest Cells, as well as the verification
of the physical association of PHF6 with CHD7.

I

also will demonstrate the methods by which we will ascertain the precise role of PHF6 in neural crest development, as well as the function of the PHF6-

CHD7 complex.
neural crest cell
Neuroectoderm
How is cellular diversity
achieved ?
neural precursor cells
Embryonic Stem Cell
Neural Differentiation
Medium
Noggin Low levels of BMP
Figure 4. Strategy for differentiation of neural crest and neural
precursor cells.
Thursday, October 18, 2012
Figure 3.
Figure 4. RNA-seq data for the SOX9 and PHF6 locus in Neural Rosettes and Neural Rosettes
generating NCC cells differentiated from Human Embryonic Stem Cells. SOX9 marks Neural
Crest lineage, showing that PHF6 is downregulated in early NCC. R Bajpai, unpublished data.
Figure 5.
Epigenetic checks and balances promote control of neural crest migration.
Erin Moran
1,2
, Katherine Pfister
3
, Richard Pelikan
4
, Ruchi Bajpai
1,2
1
Center for Craniofacial Molecular Biology, Ostrow School of Dentistry, USC;
2
Department of Biochemistry, Keck School of Medicine, USC;

3
Department of Cell Biology, UVA;
4
Arthritis and Clinical Immunology Research Program, Oklahoma Medical Research Foundation

Abstract
40 kDa
10%
Input PHF6 IgG CHD7
IP
α-CHD7
α-PHF6
*NS-Heavy Chain
50 kDa
250 kDa
Western blot
5%
Input

IgG
Western blot
α-BAF180
α-BAF170
α-BRG1

PHF6

CHD7
Co-IP
α-CHD7
PHF6
α-PHF6

IgG
5%
Input
IgG
5%
Input CHD7
control IP
40 kDa
200 kDa
PHF6 and CHD7 form a mutually exclusive complex to CHD7-PBAF
α-H3K4Me1
α-H3K9Me3
α-PHF6
Western blot
IgG PHF6
10%
Input
IP
K4me1
40 kDa
15 kDa
15 kDa
*NS-Heavy Chain
α-PHF6
α-Biotin
10%
Input

Western Blot
H3(1-27) Peptide
10 kDa
40 kDa
PHF6 can physically associate with histones marked with K4me1
30x
Day7( NEC) Day9-14( NCC)
H3K4me1 P300 H3K4me1 P300
P300
H3K27Ac
H3K4Me1
No migratory NCC NEC with migratory NCC
-2 0 2 -2 0 2 -2 0 2 -2 0 2
The majority of NCC enhancers are
premarked with K4me1 in premigratory NCC
CHD7
-Can not remodel
nucleosomes due to
inaccessibility and histone
“tethering” by PHF6
-Area remains
inaccessible to
transcripVon
factors
NCC specific
TF
No NCC specific transcrip@on
PHF6 downregulaVon
PHF6 physically associates with NCC enhancers in premigratory NCC
NC
E

-1.5Kb
FLI1 locus
NC
E

-356Kb
TWIST1 locus
0
0.5
1
1.5
2
2.5
3
3.5
FLI1E
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
TWIST1E
PHF6 ChIP
0
2
4
6
8
10
12
SOX9E SOX9P
Percent Input
0
10
20
30
40
50
60
70
80
FLI1E
0
2
4
6
8
10
12
TWIST1E
H3K4me1 ChIP
NC
E

-251Kb
SOX9 locus
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
SOX9E SOX9P
Percent Input
Neuroectoderm
Pre-migratory
NCC
Mature NCC
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
SOX9E SOX9P
CHD7 ChIP
H3K4me1 ChIP
PHF6 ChIP
Phf6 Twist
Sox2
NF17
Early Neurula
Pre-migratory
NCC
NF19
Mid Neurula
Pre-migratory and
migratory NCC
NF24
Early Tailbud
Migrating NCC
n=32
n=45
n=31
Neural Tube Cranial Neural Crest Cranial Placode
NF2 NF15-18
NF26
Donor Host
Stage Match
mCherry
PHF6
GFP GFP
mCherry PHF6 Phf6 MO + PHF6 GFP MO + PHF6
0
0.2
0.4
0.6
0.8
Cherry PHF6A GFP MO + PHF6A PHF6 MO +
PHF6A
Maximum Dorsal to Ventral
Migra@on
*
*
NS
N>5 per sample
GFP MO
Phf6 MO
Phf6 MO + GFP
Phf6 MO + PHF6
0
0.05
0.1
0.15
0.2
0.25
0.3
GFP MO Phf6 MO Phf6 MO + GFP Phf6 MO + PHF6A
Frac@on of embryo
Stagger between First and Third Arch
**
NS
Donor
WT Host
NF15-18
NF26
Stage Match
Figure 1. A. Co-immunoprecipitaVon of CHD7 with PHF6, performed in Neuroectodermal cells. B. Mutual
Co-immunoprecipitaVon of CHD7 and PHF6. C. Co-immunoprecipitaVon of PBAF complex members
BRG1, BAF180, and BAF170 with CHD7, but not PHF6.
A.
A.
B.
B.
C.
Figure 2. A. Co-immunoprecipitaVon of PHF6 with histones marked with H3K4me1, but not marked with
heterochromaVn modificaVon H3K9me3. B. Binding assay of N-terminal H3 pepVdes marked with
K4me3, K4me1, and unmodified with endogenous PHF6.
Figure 3. A.i. SOX9 locus diagram. ii. PHF6 ChIP at the SOX9 locus. iii. H3K4me1 ChIP at the SOX9 locus. iv.
CHD7 ChIP at the SOX9 locus, data from Bajpai et al., 2010. B. i. Diagram of the FLI1 locus. ii. PHF6 ChIP at
the FLI1 locus. iii. H3K4me1 ChIP at the FLI1 locus. C. i. Diagram of the TWIST1 locus. ii. PHF6 ChIP at the
TWIST1 locus. iii. H3K4me1 ChIP at the TWIST1 locus.
A. i.
ii.
iii.
iv.
PHF6 ChIP
H3K4me1 ChIP
B. i.
ii.
iii.
iii.
ii.
C. i.
Figure 4. Heatmap of H3K4me1 and p300 at the 3454 Neural Crest
specific enhancers at Day 7 of differenVaVon, represenVng
neurectodermal and premigratory neural crest cells, and day 9-14 of
differenVaVon, represenVng premigratory and migratory neural
crest cells.
H2B
H2A
H3
H4
CHD7
NCC specific
TF
Distal NCC transcripVon
PHF6 is highly expressed in pre-migratory neural crest but not
migraVng neural crest
NEC
NCC
Figure 5. A. In situ hybridizaVon at NF17 of Sox2 (ii.) Phf6 (iii.) and Twist (iv.) diagram of neural tube,
cranial neural crest, and cranial placode at the same stage (i.). B. In situ hybridizaVon of the same 3
mRNAs at NF19. C. In situ hybridizaVon of the same 3 mRNAs at NF24.
Figure 6. A. GFP Morpholino (non-targeVng) control, injected at 1/8 cell into one of the dorsal animal
blastomeres then followed to NF26 to view the migraVon paierns. B. and C. Two separate Phf6
morpholinos, injected the same way as control. Note the loss of stagger of the posterior pharyngeal
arches. D. Rescue of Phf6 MO1 using human PHF6 mRNA, longest isoform. E. mCherry mRNA, injected as
before (mRNA overexpression control). F. PHF6 mRNA overexpression and G. PHF6-ΔF2 mRNA, lacking
the second PHD domain, overexpression. Note the opposing phenotypes using these two mRNAs,
suggesVng the potenVal dominant negaVve funcVon of PHF6-ΔF2.
Figure 8. A. Diagram of how the transplantaVon experiment is performed. B-D. TransplantaVon of neural crest overexpressing
mCherry mRNA (B.), PHF6 mRNA (C.), PHF6 mRNA with a GFP (non-targeVng) morpholino (mock rescue, D.) or PHF6 mRNA with Phf6
morpholino 1 ( E.). F. QuanVtaVon of the maximum dorsal to ventral migraVon of the explant, quanVtated by ImageJ. G-H.
TransplantaVon of neural crest targeted with GFP (non-targeVng) morpholino ( G.), Phf6 morpholino 1 ( H.), Phf6 morpholino 1
overexpressing GFP mRNA (mock rescue, I.) or Phf6 morpholino 1 overexpressing human PHF6 mRNA (J.). K. QuanVtaVon of the
stagger between the first and third arch, performed by quanVtaVon of the distance migrated by the first arch subtracVng the
distance migrated by the third arch. Phf6 knockdown transplants show a significant loss of stagger of the pharyngeal arches.
Figure 9. A. Diagram of how the transplantaVon of wild-type neural
crest to mutant background is performed. This experiment is designed
to control for changes in signaling to the neural crest due to
overexpression of PHF6. B. Wild-type neural crest cells marked by GFP
are transplanted to embryos overexpressing mCherry. C. Wild-type
neural crest cells, marked by GFP, are transplanted to embryos
overexpressing PHF6. Note the normal migraVon paierns of these
neural crest cells.
Figure 11. A. Diagram of how the experiment is performed. B. Embryos are targeted by Chd7 Morpholino, as published in
Bajpai et al., 2010. C. Embryos are targeted by Chd7 morpholino and overexpressing PHF6 mRNA. D. Embryos are
targeted by Chd7 morpholino and overexpressing PHF6-ΔF2 mRNA. Note the rescue of neural crest migraVng through the
pharyngeal arches. E. Embryos are targeted by Chd7 morpholino and overexpressing PHF6-ΔF1 mRNA. F. QuanVtaVon of
the percent of embryos with no Chd7 MO associated defect, some Chd7 MO associated defect, and complete Chd7 MO
associated defect.
A. i. ii. iii.
iv.
iv.
iv.
iii.
iii.
ii.
ii.
B. i.
C. i.
Part I. Biochemical analysis of PHF6 and its interac@on with CHD7
Part II. Expression PaWern of Phf6
Acknowledgements and Funding
Funding: NIH T32 Training Grant (PI: Stallcup) Cellular, Biochemical and Molecular Biology, NIH NIDCR F31
DE024688
Collaborators:
Ray Keller, Katherine Pfister (UVA)
Marianne Bronner, Shuyi Nie (Cal Tech)
MaW Lee, Susan Smith (USC)
Peggy Farnham

Bajpai Lab:
Ruchi Bajpai
Jennifer Oki Leslie Dominguez
Bajpai Lab Former Members
Richard Pelikan
Philbert Mach Maria Nava
USC Stem Cell Core, USC Ins@tute for Gene@c Medicine, USC Center for Craniofacial Molecular Biology, USC
Department of Biochemistry and Molecular Biology, Gene@c Molecular Cellular Biology Graduate Program
Part III. In vivo effect of Phf6 on neural crest migra@on
Effect of PHF6 overexpression on Twist expression paiern
mCherry
PHF6
Injected Injected Control Control
NF17
NF19
Twist
mCherry PHF6A
0.0
0.5
1.0
1.5
mRNA Injected
twist1
Ratio of distance from midline
**
Injected/Control
Twist
Ra@o of stained area
mCherry PHF6-ΔF2
Injected
Injected
Control
Control
NF17
NF19
Twist
mCherry PHF6B
0.0
0.5
1.0
1.5
2.0
2.5
mRNA Injected
twist1
Ratio of distance from midline
*
Twist
Ra@o of stained area
Injected/Control
mCherry
PHF6-ΔF2
mRNA injected
mCherry PHF6
Effect of Phf6 on neural crest migraVon in situ
GFP MO Phf6 MO1 Phf6 MO2
Phf6 MO1
PHF6 mRNA
mCherry mRNA PHF6 mRNA
PHF6-ΔF2 mRNA
NF4
NF26
A. B.
C.
D.
E.
F.
G.
A. B.
C. D.
Figure 7. A. In situ hybridizaVon of Twist in embryos overexpressing mCherry mRNA on the injected side,
or PHF6 mRNA on the injected side. B. quanVtaVon of the raVo of stained areas on the injected side vs.
control side. Areas were picked up un-aided by ImageJ sooware. C. In situ hybridizaVon of Twist in
embryos overexpressing mCherry mRNA on the injected side, or PHF6-ΔF2 mRNA on the injected side. D.
QuanVtaVon of the raVo of stained areas, as in B.

Neural crest transplantaVon experiments show that Phf6 affects are cell-autonomous
A.
B. C. D. E.
F.
G. H. I.
J.
K.
A.
B.
C.
Chd7 MO+PHF6-ΔF2
Chd7 MO
Chd7 MO+PHF6-ΔF1
Defect
Rescue
Chd7 MO+PHF6
**
*
Percent embryos
100%
80%
60%
40%
20%
0%
A.
B. C.
D.
E.
F.
PHF6-ΔF2 mRNA rescues Chd7 morpholino phenotype
0
50
100
150
Control MO Phf6 MO
Pixels at 20x
Average NC Cell Length
***
GFP MO Phf6 MO
Top scale bar: 30 um; Boiom scale bar: 10 um
mCherry
PHF6
0
50
100
150
200
250
mCherry PHF6
Pixels at 63x
Average Distance Migrated
***
0
5
10
15
20
25
30
Control Phf6 MO PHF6A
Pixels/second
Average Velocity
***
**
A. B.
D.
E.
C.
F.
G.
Figure 10. A-C. 20x maximum intensity projecVon of neural crest explant on fibronecVn, injected
with GFP MO (A.), Phf6 MO (B.), quanVtaVon of cell size (C.). D-F. 63x maximum intensity
projecVon of neural crest explants on fibronecVn, injected with mCherry (D.) and PHF6 (E.),
quanVtaVon of distance migrated by NCC (F.). G. Average velocity of cells in explant, via 30 minute
confocal movie, data not shown.
FN coverslip
Neural crest explants show effect of Phf6 on individual cell dynamics
G.
104
no inherent directionality, and send out multiple lammelipodial projections per cell (Fig. 4.8B),
which may account for why these cells appear elongated.
As anticipated, neural crest explants overexpressing PHF6A show very little migration. I
explanted these cells as with morphant embryos, and imaged these cells at 63x magnification using
confocal microscopy. Interestingly, I observed that the overexpressing neural crest cells seemed
to extend lammelipodial projections (Fig. 4.8E, scale bar 10 µm), but did not separate from the
main mass of the explant. Control mCherry overexpressing neural crest cells, on the other hand,
were able to separate from the main mass of the explant and migrate away (Fig. 4.8D, scale bar 10
µm). This migration is quantified in terms of pixels migrated (Fig. 4.8F, p<0.001 by two-tailed
student’s T-test).
We then observed the migration of these cells in real time. In order to accomplish this, we
performed injection of Myosin-GFP mRNA alone or accompanied by Phf6 MO1 or PHF6A
mRNA in one of the neural crest precursor blastomeres at NF6, located in the second tier,
blastomere b2 or b3 (Moody 1987b). This injection strategy was used to highlight less than 50%
of the cells with fluorescence, and follow these cells specifically and not lose them within a field
of cells. We then screened for partial neural crest targeting with the GFP label, and explanted
neural crest onto fibronectin and followed the targeted cells after a 30-minute rest period for 2
hours, taking an image in a single z-plane every 5 minutes.
Using ImageJ single-cell tracking plugin
186
, I then tracked each fluorescent cell in the case
of Phf6 morphant cells, or 10 cells in the case of control and PHF6A overexpressing cells, to
determine the velocity and displacement of these cells within the scope of imaging. I found that
Phf6 morphant cells showed a significant increase in velocity over control cells (Fig. 4.9B, p<0.01
by two-tailed student’s T-test), but that there was no significant increase in net displacement
105
Figure 4.9. Dot-and-line confocal movies, tracking neural crest cells through a field upon Phf6
depletion or PHF6A overexpression. A. Final frame of the movies, with line trajectory marked. B.
Mean velocity of tracked cells, **, p<0.01, ***, p<0.001 by two-tailed student’s T-test. C. Total
displacement of each cell, **, p<0.01 by two-tailed student’s T-test. Error bars indicate SEM.
Control and PHF6A, n=10. Phf6 MO, n=4.

Figure 4.9
Control Phf6 MO PHF6A
0
5
10
15
20
25
30
Control Phf6 MO PHF6A
Pixels/second
Velocity
0
20
40
60
80
100
120
Control Phf6 MO PHF6A
Pixels
Displacement
**
**
***
A
B C
i ii iii
106
because these cells did not routinely migrate in one direction (Fig. 4.9A, panel ii, migration lines,
and Fig. 4.9C). PHF6A overexpressing cells, on the other hand, did put out projections away from
the central mass of the explant, but never actively migrated in that direction, causing a much lower
average velocity (Fig. 4.9B, p<0.001 by two-tailed student’s T-test). In fact, the only movement
by these cells appeared to be the passive spreading of the explant (Fig. 4.9A, panel iii). This caused
a significant decrease in net displacement (Fig. 4.9C).
Based on the observations gleaned in these two sets of experiments, I concluded that there
appears to be inherent differences in the migratory potential of the neural crest cells based
on the expression levels of Phf6. This likely operates through a pathway downstream of Phf6
activity. Further investigation into these phenotypes is required to fully comprehend these
dynamics.

Section 4.3. Discussion
Craniofacial disorders are a highly prevalent and rapidly growing group of disorders;
understanding the etiology of these disorders is critical to determining the cellular and tissue
biology from which these structures arise. During gastrulation in Xenopus laevis embryos, neural
crest cells become specified to the neural crest cell fate along the prospective neural plate border
region. However, these cells do not become migratory until much later in neurulation, despite
progressive upregulation of multiple transcription factors permissive for migration, pathfinding,
and terminal differentiation of these cells. Additionally, neighboring cells within this region,
which receive the same external signals, have the developmental potential to either remain within
the neural plate region and become neural progenitors, or undergo an epithelial to mesenchymal
transition and become migratory neural crest cells. Our hypothesis within this Chapter revolved
107
around whether the presence of PHF6 at the neural plate border and later throughout the neural
plate could provide a restrictive mechanism on neural crest formation to protect the integrity of
the neural plate, and subsequently the neural tube, by prohibiting precocious migration of these
specified, transcriptionally competent neural crest progenitors. Our data supports a neural crest
restrictive role for PHF6, since knockdown of this factor either by morpholino or by
overexpression of a dominant negative isoform causes precocious neural crest cell migration in
the posterior arch regions. Furthermore, the structure of the neural tube is crucially changed as
seen by the frequency of neural tube closure delays seen in these embryos (Fig. 4.3). Ectopic
expression of humanized PHF6, which is not subjected to endogenous Xenopus laevis
mechanisms of PHF6 expression control, restricted the formation of neural crest and resulted in
a broadening of the neural tube region (Fig. 4.2), which supports the idea that PHF6 restricts NCC
migration. This effect proved to be tissue autonomous, suggesting that PHF6 is acting through a
mechanism that is key to the functionality of neural crest cells either collectively or individually.
Thus, we propose that prospective mechanisms of PHF6 activity are limited to those involving
neural plate border cells and their transcriptional and translational activities.
While PHF6 is a protein with relatively unknown function, there are several prospective
mechanisms of activity that could be at play here. Tandem Affinity Purification followed by mass
spectrometry (TAP-MS/MS) data was produced from isolation of tagged PHF6 from HEK293
cells
187
, and the predominant categories of proteins identified to bind to PHF6 were chromatin
modifiers and RNA processing/translation proteins. This suggests a potential dual role for PHF6
in transcription control as well as post-transcriptional modification and/or translational control.
The evidence available for a transcriptional control mechanism of PHF6 activity involved in this
developmental process is fairly high. Todd et al.
187
identified the physical association of PHF6
108
with the NuRD complex in HEK293 cells, and validated its interaction with multiple members of
the complex within these cells, including several members of the CHD family of proteins.
Additionally, the second PHD domain of this protein has been determined through electrophoretic
mobility shift assays to bind directly, though non-specifically, to DNA
183
. ChIP and ChIP-seq
assays in HeLa and Jurkat cells
188,189
, respectively, and found that PHF6 is bound to chromatin at
the core rDNA promoter as well as at many active promoters throughout the genome, suggesting
that there is a specific mechanism for targeting PHF6 towards these regions, perhaps through the
first plant homeodomain. This first PHD has not been assayed for binding due to its insolubility
and therefore intractability for binding assays, but recent efforts have identified some elements of
its structure
190
. Despite this, PHD proteins have been historically known to bind modified histone
tails
191
, and this first PHD domain could target the protein towards modifications that determine
its specific genomic binding pattern. Combined with this data, it is therefore likely that in neural
ectodermal cells PHF6 has a specific pattern of genome binding that would either be inhibitory
to the neural crest transcriptional network or permissive to the neural progenitor transcriptional
network. This is supported by the fact that PHF6 was shown to bind CHD family member
proteins
187
, and work in Xenopus laevis has shown that CHD7 is highly important for promotion
of the neural crest cell state
29
.
Alternately, PHF6 could be involved in RNA processing and control of translation through
control of ribosomal biogenesis. Besides the NuRD complex, TAP-MS/MS
187
identified a large
number of splicing factors bound to PHF6, suggesting a role in post-transcriptional modification
of spliced mRNA transcripts. There are many transcripts that are differentially spliced within
different differentiation lineages, and perhaps PHF6 forms a complex to promote neuronal
precursor-type splicing
192–195
. PHF6 has also been identified by several groups to strongly localize
109
to the nucleolus in HeLa cells and HEK293T cells, and interact with nucleolar proteins
187
and
has been found at the core rDNA promoter
196
, all suggesting a role for PHF6 in ribosome
biogenesis in rapidly growing cells. Thus, PHF6 might impact translation control through either
promotion or suppression of ribosome biogenesis. Neural crest cells are rapidly growing, highly
migratory cells which require a large number of proteins to maintain this functionality, so if PHF6
is involved in suppression of ribosome biogenesis then relief of this suppression would lead to
rapid recruitment of ribosomes and increased translation rates.
The presence of multipolar morphologies of individual cells is highly reminiscent of the
phenotype found in Zhang et al.
116
. In both neurons and neural crest cells, it is apparent that
pathfinding is deficient. The ability to track the migration within this system in real time, however,
shows us that these individual neural crest cells display an increased velocity (Fig. 4.9) in Phf6
depleted cells. This may also be the case in individual neurons, as real-time studies have not been
done. This increase in velocity paired with the decrease in directionality could allow for the
subtlety of the phenotype observed in late tailbud stage embryos (Fig. 4.2).
We demonstrate here that PHF6 is involved in the cell fate decision to become neural crest
through changes in morphology. However, our understanding of the mechanistic interactions
formed by PHF6 in these cell types is still limited. Understanding of the protein-protein and
protein-DNA interactions formed within these cell types will be crucial to the understanding of
these phenomena, and to the understanding of the etiology of patient phenotypes in BFLS. Further
delving into the mechanism by which these cellular phenotypes are altered will also provide us
with a more complete understanding of this cell fate decision on a very crucial level, and will
advance our knowledge of craniofacial disorders as a group.

110
Chapter 5: Mechanistic evaluation of PHF6 interaction with chromatin
Section 5.1 Introduction
The neural crest is a population of cells that is formed during neurulation from precursor
cells in the neural plate border region. These cells undergo a transition during which they gain
migratory potential and leave the organized epithelium. Pivotal work has been performed by a
number of different groups to understand the transcriptional regulatory network surrounding this
cell fate decision, and this work has revealed a number of transcription factors that are necessary
at each level of the specification and migration of these cells
63,197,198
. For example, neural plate border
transcription factors are required to activate transcription factors that are involved in the epithelial
to mesenchymal transition of these cells
63
. This work has not only illuminated the cascade of
transcription required for the neural crest cellular phenotype to arise, but has also made evident
that these steps are tightly controlled both temporally and spatially
63
.
Despite the relatively high understanding of the gene regulatory network of neural crest
cells, we still have relatively little understanding on the precise mechanisms of transcriptional
activation at the gene loci in these cells. The transcription factors that are required for neural crest
cell fate have to access the DNA at thousands of loci across the genome in order to promote or
repress transcription, which highlights the need for a permissive chromatin state at these loci. This
permissive chromatin state can either be facilitated or blocked by the activity of multiple epigenetic
regulators, including chromatin modifiers and chromatin remodelers, which simultaneously
facilitates specific transcription as well as allowing for additional checkpoint measures to be in
place to prevent erroneous or ectopic expression of certain transcripts. Epigenetic reprogramming
is known to be highly necessary during cell fate decisions for this reason; multiple studies have
shown that the epigenetic state of the cell, including enhancer position, repressed chromatin
111
regions, and other hallmarks, are highly cell-type specific
199
. Indeed, the kinetics of transcription
during the development of higher eukaryotes is thought to be linked to the kinetics of creating the
appropriate epigenetic state. These twin ideas of kinetic control and restrictive measures on
inappropriate transcription allow for the level of specific transcriptional regulation of the neural
crest program that we observe in vivo, in particular the level of specificity both spatially and
temporally.
Our current knowledge as to the epigenetics of creating the neural crest permissive
chromatin state involves the ATP-dependent chromatin remodeler CHD7 and the remodeling
pBAF complex. Studies performed in a number of different model organisms have shown that
CHD7 interacts with the pBAF complex
29
, and that together this complex promotes the
transcription of certain neural crest specific transcription factors, downstream of neural plate
border genes
29
. Additionally, we know that these proteins are physically localized to several neural
crest specific enhancers
29
. However, no protein elements of these complex show the potential
ability to act as a restrictive measure for inappropriate neural crest specific transcription.
We found in Chapter 4 that depletion of Phf6 in Xenopus laevis neural crest causes a loss
of temporal control of neural crest migration, leading to premature neural crest formation at the
more posterior cranial neural crest regions. We therefore hypothesized that this protein could act
as a restrictive measure at the chromatin level that prevents precocious activation of neural crest
transcription. In this chapter, I will discuss the evidence that I have found for this model, including
the ability of this protein to bind to CHD7 in the absence of pBAF complex members, bind to
enhancer associated histone PTM H3K4me1, and indeed bind to enhancers in neural crest cells
prior to migration. I will then discuss the ways in which this model can be further evaluated with
continuing studies.
112

Section 5.2. Results
5.2.1 CHD7 and PHF6 form a complex that is mutually exclusive to the CHD7/pBAF complex
ATP-dependent chromatin remodeler CHD7 has a well-established role in the regulation
of nucleosome-free regions, and has been found to track with H3K4me1 in human embryonic stem
cells
27
. Native immunoprecipitation reactions of CHD7 protein in a mixed population of human
neural ectodermal cells mixed with neural crest cells followed by bulk mass spectrometry analysis
identified a low number of PHF6 peptides (data not shown, performed by Ruchi Bajpai), which
signifies a potentially transient interaction between these proteins. Co-immunoprecipitation of
endogenous PHF6 performed in human embryonic stem cells differentiated into human neural
ectodermal cells and a mixed population of neural ectodermal cells and neural crest cells verified
this interaction as well as demonstrating the relative scarcity of the interaction (Fig. 5.1A,
representative of 4 such immunoprecipitations). In both cell populations, nearly complete
depletion of PHF6 by immunoprecipitation led to approximately 10% of the input CHD7 in the
eluted fraction, indicating that the majority of PHF6 in these cells is unlikely to be involved solely
in this complex. Furthermore, PHF6 is not involved in the complex between CHD7 and BRG1 and
other members of the pBAF complex. Co-immunoprecipitation experiments performed on neural
ectodermal cell nuclear extracts show that PHF6 immunopreciptation reactions are able to co-
immunoprecipitate CHD7, but unable to co-immunoprecipitate BRG1, BAF170, or BAF180, three
members of the pBAF complex (Fig. 5.1B panel iii and iv, Fig. 5.1C panel ii, and Fig. 5.1F).
Conversely, immunoprecipitation of BRG1 did not co-immunoprecipitate PHF6, but did co-
immunoprecipitate CHD7 (Fig. 5.1C, D, and E). This immunoprecipitation reaction did not enrich
BRG1 to a high level, but it nevertheless shows that immunoprecipitation reaction of BRG1 causes
113
Figure 5.1. Co-immunoprecipitation of PHF6 with chromatin remodeling proteins. A. Co-
immunoprecipitation of CHD7 (panel i) with PHF6 (panel ii), in reference to non-specific
immunoprecipitation with rabbit IgG and 1% of the total immunoprecipitation input.
Representative image of 3 experiments. B. Co-immunoprecipitation of CHD7 (panel ii), BRG1
(panel iii), BAF180 and BAF170 (panel iv) with PHF6 (panel i) and CHD7 in reference to non-
specific immunoprecipitation with rabbit IgG and 5% of the total immunoprecipitation input.
Representative image of 3 experiments. C. Co-immunoprecipitation of CHD7 (panel i), BRG1
(panel ii), and PHF6 (panel iii) with PHF6, CHD7, and BRG1 in reference to non-specific
immunoprecipitation with rabbit IgG and 10% of the total immunoprecipitation input.
Representative image of 3 experiments. D-F. Composite enrichment across all
immunoprecipitation experiments of D. CHD7, E. PHF6, and F. BRG1. Error bars indicate SEM.

PHF6 IgG CHD7 BRG1
Immunoprecipitation from NEC
α-CHD7
α-BRG1
α-PHF6
*NS-Heavy Chain
α-BAF180
α-BAF170
α-BRG1
α-CHD7
PHF6
α-PHF6
IgG
5%
Input
IgG
5%
Input CHD7
-5
0
5
10
15
20
25
30
10%
Input
IP IgG IP
PHF6
IP
CHD7
IP
BRG1
Relative Density
CHD7 enrichment
-5
0
5
10
15
20
25
30
35
40
45
10%
Input
IP IgG IP PHF6 IP
CHD7
IP
BRG1
Relative Density
PHF6 Enrichment
-5
0
5
10
15
20
25
Input
5%
IP IgG IP
PHF6
IP
CHD7
IP
BRG1
Relative Density
BRG1 Enrichment
Western blot
Western blot
Western blot
250 ⎯
150 ⎯
50 ⎯
50 ⎯
250 ⎯
250 ⎯
150 ⎯
200 ⎯
40 ⎯
A. B.
i
ii
i
ii
iii
iv
i
ii
iii
C.
D. E. F.
114
co-immunoprecipitation of CHD7, a result that is consistent with published data
29
. Given that PHF6
immunoprecipitation was unable to co-immunoprecipitate any of the pBAF complex members, it
is unlikely that a replicate of this BRG1 immunoprecipitation reaction would co-
immunoprecipitate PHF6. Therefore, we can reasonably conclude that while both the pBAF
complex and PHF6 form a complex with CHD7, the two complexes are mutually exclusive and
form independently of each other. While the CHD7/pBAF complex has been shown to have a
synergistic positive effect on neural crest transcription and migration
29
, the role of the CHD7/PHF6
complex is currently unknown, and given that PHF6 seems to have a restrictive role on neural crest
migration, this complex may be involved in sequestering CHD7 at the regions of chromatin where
it is required, but preventing its eminent activity until the appropriate signals are received.

5.2.2. PHF6 binds modified histone tails through recognition of H3K4me1
In order to test one part of the hypothesis that PHF6 associates with CHD7 and sequesters
it, we wanted to determine if PHF6 can associate with chromatin. To test this, I created
recombinant PHF6 constructs which could be induced in BL21 bacterial derivatives and purified
using a Glutathione-S-Transferase (GST) tag. Three PHF6 constructs were cloned, the full-length
protein, as well as two constructs each lacking one of the PHD motifs in order to test the reactivity
of each PHD motif separately. These constructs were cloned, along with the GST tag alone as a
negative control for background histone binding, into a BL21 derivative designed to aid protein
folding of proteins that are unstable in recombinant assays, Arctic Express (Aligent). Initial
experiments performed by us and other groups (data not shown) suggested that the N-terminal
region of this protein would be highly unstable in standard recombinant purification assays. The
use of this system allowed for the purification of three PHF6 recombinant proteins: full length
115
(FL), deletion of PHD motif 2 (DF2), and deletion of PHD motif 1 (DF1), with a small amount of
some degradation products of PHF6-FL and PHF6-DF2 (Fig. 5.2A). These recombinant proteins
were then used in order to determine the range of histone modifications that could interact with
PHF6 by performing GST pulldown assays whereby the recombinant proteins were incubated with
purified cellular histone extracts, and using the GST tag the recombinant protein is precipitated
and assayed for co-precipitated histones by western blot. I used a highly-diluted histone extract
(1:50 dilution) to perform this assay, as well as extensive washing of the precipitated proteins to
remove non-specific interactions. These precautions were sufficient to prevent non-specific
interactions between GST and the histone proteins, as the GST only control showed no non-
specific interaction with H3 (Fig. 5.2B, far right). Additionally, all recombinant PHF6 proteins
showed an interaction with histone H3 (Fig. 5.2B, panel I, representative of 2 such reactions),
signifying that PHF6 can interact specifically with histones. Furthermore, when assaying some
common histone post-translational modifications, we found that only PHF6 constructs containing
the first PHD motif can interact with H3 bearing the modification of Lysine 4 mono-methylation
(Fig. 5.2B, panel ii, representative of two such reactions), and that only the PHF6-DF2 can interact
with H3 bearing the modification of Lysine 27 acetylation (Fig. 5.2B, panel iii). These two
modifications commonly co-exist as these are two modifications associated with activated
enhancers, so at least one of these modifications is likely passively carried on the actively bound
histone protein. PHD proteins are commonly bound to methylated lysines on histone tails
191
, so we
considered H3K4me1 a possible binding partner of PHF6.
In order to determine the ability of endogenous PHF6 to bind to modified histone tails, we
performed co-immunoprecipitation experiments, whereby we immunoprecipitated PHF6 and
H3K4me1 out of protein extracts containing nuclear and histone extracts. Immunoprecipitated
116
Figure 5.2. Interaction between PHF6 and modified histone tails. A. Coomassie stain of
recombinant PHF6 proteins, PHF6-FL (diagram, right side), PHF6-DF2 (diagram, right side),
PHF6-DF1 (diagram, right side), and GST only (diagram, right side). Diagram aligns with the
corresponding band in the coomassie stained SDS-PAGE gel. B. GST-pulldown analysis of
binding between recombinant PHF6 proteins and dilute modified histone extracts. Precipitated
extracts are probed using western blotting for H3 (panel i), H3K4me1 (panel ii), and H3K27ac
(panel iii). Representative image of 2 experiments. C. Co-immunoprecipitation of PHF6 (panel
i), H3K4me1 (panel ii), and H3K9me3 (panel iii) with PHF6 and H3K4me1, with reference to
non-specific immunoprecipitation binding with rabbit IgG and 10% of total immunoprecipitation
input. Representative image of 4 experiments. D. Streptavidin-pulldown analysis of binding
between biotinylated H3 peptides, amino acids 1-27 bearing the modification of K4me3, K4me1,
or unmodified K4, and endogenous PHF6 (panel i). PHF6 enrichment relative to biotin quantity
precipitated in each reaction (panel ii) is quantitated in the inset, iii. Single Experiment. E. Peptide
competition assay, whereby co-immunoprecipitation experiments of K4me1 with precipitating
antibody against PHF6 is competed with equimolar (+) or 5-fold excess (++) of each peptide, H3
peptides amino acids 1-27 with K4me3, K4me1, or unmodified K4. Single Experiment. F. GST
pulldown analysis of endogenous PHF6 (panel ii, arrow) with recombinant PHF6 (recognized by
GST antibody, panel i, and PHF6 antibody, panel ii). Single Experiment.

Figure 5.2.
α-H3
α-H3K4me1
α-H3K27ac
PHF6-FL
PHF6-ΔF2
PHF6-ΔF1
GST - - - +
- - + -
- + - -
+ - - -
15 ⎯
15 ⎯
15 ⎯
Input 10%
K4me1 K4
α-PHF6
α-Biotin
No
Peptide Input K4me3
Western Blot
H3(1-27) Peptide
α-H3K4me1
Western Blot
IgG PHF6 IP:
Input
5%
10
40
IP
IgG PHF6 Input K4me1
40 ⎯
15 ⎯
15 ⎯
α-H3K4Me1
α-H3K9Me3
α-PHF6
A.
B.
C. D.
E.
F.
Competing Peptide:
K4me1 + ++
K4me3 + ++
K4 + ++
Western Blot
Western Blot
40 ⎯
0
0.1
0.2
0.3
0.4
K4me3 K4me1 K4
Relative Density, Biotin
Normalized
PHF6 Enrichment
50 ⎯
50 ⎯
α-PHF6
Western Blot
α-GST
i
ii
iii
i
ii
iii
i
ii
i
ii
117

PHF6 was readily able to co-immunoprecipitate histone proteins bearing the modification of
H3K4me1 (Fig. 5.2C panel ii, representative of four such reactions), but not repressive histone
post-translational modification H3K9me3 (Fig. 5.2 panel iii, representative of four such reactions).
Converse immunoprecipitation experiment of H3K4me1 was able to only co-immunoprecipitate a
small amount of PHF6 over background (Fig. 5.2 panel i, representative of four such reactions),
indicating that either binding of PHF6 interfered with the precipitation reaction, as this antibody
recognizes a relatively small epitope, or that only a small amount of the H3K4me1 contained
within the cell binds PHF6.
To assay the binding of PHF6 to singly-modified histone tails, we performed peptide
pulldown experiments whereby biotinylated singly modified histone peptides were incubated with
nuclear extracts, and peptides were then precipitated using streptavidin. These precipitates were
then assayed for bound PHF6 using western blot. H3 amino acids 1-27 were used, in unmodified
form, with K4 mono-methylated, and with K4 tri-methylated. Many PHD motifs that are able to
bind methylated lysines show a preference for one methylation state over another
191
, and I wanted
to determine for which methylation state PHF6 had preference. Methylated K4 bound more PHF6
than unmodified K4, with a slight preference for H3K4me1 over H3K4me3 (Fig. 5.2D, single
experiment). This becomes even more clear when normalized for biotin levels (Fig. 5.2 D inset).
In order to test the sufficiency of the interaction of PHF6 and H3K4me1 for the PHF6/H3
interaction, I performed a peptide competition assay in which the co-immunoprecipitation
experiment was spiked with equimolar amounts (+) or a 5-fold excess (++) of H3 amino acid 1-27
in unmodified form, with K3K4me1, or H3K4me3. The rationale behind this experiment is that if
the interaction between PHF6 and this histone post-translational modification is sufficient for the
interaction between PHF6 and histone protein, then having excess of that peptide in the reaction
118
would cause the immunoprecipitation of PHF6 to not have bound H3 or H3K4me1. This was not
the case, as none of these peptides were able to successfully compete the interaction (Fig. 5.2E,
single experiment). This indicates that there is likely another interaction necessary for the
specificity of the PHF6/Histone protein interaction.

5.2.3. PHF6 forms a homodimer
In order to verify the binding capabilities of recombinant, bacterially purified PHF6 to
bind to neural crest nuclear complexes in a similar manner to endogenous PHF6, I performed a
GST pulldown experiment where recombinant PHF6 was incubated with neural ectodermal
nuclear extracts and recombinant PHF6 was precipitated using Glutathione sepharose. The binding
of these recombinant proteins was then assayed for protein binding by western blotting. In the
process, precipitation was verified by western blot of GST and PHF6 N-terminal and C-terminal
antibodies. Endogenous PHF6 was observed to be precipitated by the GST pulldown reaction in
both the PHF6-FL and PHF6-DF2 samples, but not the PHF6-DF1 sample (Fig. 5.2F, arrow; single
experiment). This band is distinct from the recombinant proteins in that the GST-tag on the
recombinant protein causes a 25 kD and 10 kD shift in the PHF6-FL and PHF6-DF2 proteins,
respectively, allowing us to observe the endogenous protein (Fig. 5.2F). In chapter 4, we found
that a shorter isoform of PHF6 lacking the second PHD motif acts in a dominant-negative fashion;
this data suggests a potential mechanism for this activity in that overexpression of this shorter
isoform may have the ability to create inactive dimers with functional protein, thereby titrating
away active protein.

5.2.4. PHF6 binds to neural crest specific enhancers prior to complete transcriptional activation.
119
Given that PHF6 has an interaction with H3K4me1, which is generally thought of as a
histone post translational modification associated with pre-activated and activated enhancer
regions, I wanted to assay whether PHF6 binds to these regions in neural ectodermal cells as these
cells become neural crest. Work in our lab and in the lab of Joanna Wysocka at Stanford has
identified a set of active enhancers in our human neural crest differentiation system, and we
identified a set of enhancers that are distal to several neural crest markers and that have been
demonstrated to be activated in the neural crest differentiation system (data not shown, work of
Mallory Holland, Ruchi Bajpai). I therefore performed Chromatin Immunoprecipitation
experiments over the timecourse of our differentiation system, separating out 3 different cell
populations: Neural ectodermal cells that had not yet attached to the dish (NEC), Neural
ectodermal cells that had attached to the dish but had not yet started neural crest migration
(Premigratory NCC), and a pure population of migrated neural crest cells (Mature NCC). From
these different populations of cells, I then precipitated chromatin associated with PHF6,
H3K4me1, H3K27ac, and H3K4me3, and assayed the selected enhancers and several associated
promoters for deposition of the histone PTMs and PHF6 binding.
Figure 5.3 shows the composite data for 4 such experiments. PHF6 shows similar
enrichment levels to other transcription factors and PHD motif proteins
200,201
(Fig. 5.3A). PHF6
shows significant enrichment over background at SOX9 enhancer and TWIST1 enhancer (Fig. 5.3A
SOX9E and TWIST1E) over the control heterochromatic region, p=0.005 and p=0.02 respectively
by two-tailed student’s T-test with Bonferroni correction. Specifically, SOX9 enhancer, TWIST1
enhancer, and FLI1 enhancer show significant enrichment at the pre-migratory neural crest time
point over background, p=0.004, p=0.01, and p=0.03 respectively by two-tailed student’s T-test

120
Figure 5.3. ChIP analysis of PHF6 bound to neural crest enhancers and promoters. Top
diagrams, illustrations of the loci assayed and the primers designed for each region. A-D.
Composite enrichment at each locus of A. PHF6, B. H3K4me1, C. H3K27ac, D. H3K4me3. *,
p<0.05, **, p<0.01, by two-tailed student’s T-test. Error bars indicate SEM. PHF6 ChIP, K4me1
ChIP, all loci, n=4. K27ac ChIP, all loci but SOX2P, n=2. K4me3, mature NCC, all loci but
SOX2P, n=2. All bars without error, n=1.

Figure 3.
NC
E
-251Kb
SOX9 locus
E
NC
E
-1.5Kb
FLI1 locus
E
NC
E
-356Kb
TWIST1 locus
E P
P
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Control SOX9E SOX9P TWIST1E TWIST1P FLI1E BMP7E LMX1AE SOX2P
PHF6 Enrichment, % Input
NEC
Premigratory NCC
Mature NCC
0
5
10
15
20
25
30
35
Control SOX9E SOX9P TWIST1E TWIST1P FLI1E BMP7E LMX1AE SOX2P
K4me1 Enrichment, % Input
NEC
Premigratory NCC
Mature NCC
0
1
2
3
4
5
6
7
8
9
10
Control SOX9E SOX9P TWIST1E TWIST1P FLI1E BMP7E LMX1AE SOX2P
K42ac Enrichment, % Input
NEC
Premigratory NCC
Mature NCC
K27ac ChIP
0
0.5
1
1.5
2
2.5
3
3.5
Control SOX9E SOX9P TWIST1E TWIST1P FLI1E BMP7E LMX1AE SOX2P
K4me3 Enrichment, % Input
NEC
Premigratory NCC
Mature NCC
K4me1 ChIP
K4me3 ChIP
PHF6 ChIP
A.
B.
C.
D.
*
**
**
**
*
121
(Fig. 5.3A SOX9E, TWIST1E, and FLI1E). There is a 10-fold composite enrichment of SOX9
enhancer occupation over background, and a 30-fold composite enrichment of FLI1 enhancer
occupation over background at the pre-migratory neural crest time point. These statistics together
indicate that at these enhancers, PHF6 is bound at the pre-migratory neural crest time point.
In order to assay enhancer activation, we look to two histone PTMs, H3K4me1 and
H3K27ac. These two histone PTMs are the same that were used to classify the set of enhancers
within this system
92
. H3K4me1 is highly deposited at the same three enhancers that we observed
PHF6 binding, SOX9 enhancer, TWIST1 enhancer, and FLI1 enhancer (Fig 5.3B, p=2.3x10
-6
,
p=3.9x10
-4
, p=2.2x10
-4
). Surprisingly, two other enhancers, BMP7 enhancer and LMX1A enhancer
showed limited activation across the time course of neural crest differentiation, with limited
deposition of H3K4me1 and H3K27ac (Fig. 5.3B, C). It is unknown in this particular set of
experiments if this resulted in a lack of transcription of these two genes, as RNA analysis was not
performed and promoter ChIP analysis for active promoter chromatin PTM H3K4me3 was not
assayed. However, SOX9 promoter and TWIST1 promoter showed significant enrichment of
H3K4me3 at the mature neural crest time point, indicating activation of these promoters (Fig. 5.3
D, p=0.03 and p=0.03 respectively). In these experiments, the highest deposition of H3K4me1
corresponded to the highest binding of PHF6, at the point of pre-migratory neural crest. H3K27ac,
while more variable in enrichment across experiments, tended to become more enriched over the
time course of differentiation. This is particularly clear in the case of TWIST1 enhancer (Fig. 5.3
C). Overall, however, this histone modification seemed to have the highest level of variability, and
ChIP experiments in general using this antibody can be highly batch dependent.
Despite the variability of the H3K27ac ChIP experiments, we can draw the conclusion that
PHF6 is most highly bound to a subset of enhancers where H3K4me1 is highly enriched, at the
122
time point before the nearby promoters gain the activation PTM of H3K4me3. Therefore, we can
conclude that PHF6 is present at these enhancers prior to the activation of associated promoter
transcription.

5.2.5. PHF6 enrichment follows enrichment of H3K4me1.
In order to further analyze the correlation between PHF6 enrichment and H3K4me1
enrichment, I compiled all ChIP data for all loci and plotted corresponding values of PHF6 and
H3K4me1 enrichment in heatmap form as well as in a single scatter plot where the values were
transformed over log 10 to gain a normal distribution for statistical testing (Fig. 5.4 A and B). As a
control, I performed the same analysis using H3K4me3 enrichment and PHF6 enrichment (Fig.
5.4 C and D). I performed standard R
2
analysis of the resultant line of best fit for each scatter plot
(inset, Fig. 5.4 B and C), which resulted in an R
2
value of 0.62 for PHF6 and H3K4me1 versus
0.061 for PHF6 and H3K4me3. This R
2
value is highly indicative of a strong level of correlation
between PHF6 binding at a locus and H3K4me1 deposition at the same locus, across all loci and
all experiments. Conversely, the R
2
statistic between PHF6 and H3K4me3 indicates that the null
hypothesis is maintained in this case, that there is no correlation between PHF6 binding at a locus
and H3K4me3 deposition. This is a significant point, as data in Fig. 5.2 showed a preference for
PHF6 binding to H3K4me1 over H3K4me3, but there was still some affinity for the H3K4me3
histone PTM. Furthermore, ChIP-seq experiments performed in Jurkat cells indicated a preference
for PHF6 binding to promoter regions in this cell type
189
. There may be some cell type specificity
for the two apparent differential activities of this protein, for this analysis suggests that PHF6 has
a clear preference for enhancer loci binding within the context of neural crest differentiation.

123
Figure 5.4. Correlation analysis between H3K4 methylation and PHF6 enrichment. A, D.
Heatmap analysis of H3K4me1 and PHF6 (A.) and H3K4me3 and PHF6 (D.) across all loci and
in all experiments. B, C. Scatter plot analysis used for calculation of R
2
values, inset into the plot,
of PHF6 enrichment versus H3K4me1 deposition, transformed over log 10 to gain a normal
distribution (B.) and PHF6 enrichment versus H3K4me3 deposition transformed over log 10 to gain
a normal distribution (C.).

Figure 5.4.
K4me1
K4me1 enrichment
PHF6
PHF6 enrichment
K4me1 enrichment
10
8
6
4
2
PHF6 enrichment
0.4
0.3
0.2
0.1
0
K4me3
K4me3 enrichment
PHF6
PHF6 enrichment
K4me3 enrichment
3
2
1
0
PHF6 enrichment
0.08
0.06
0.04
0.02
0
K4me1
K4me1 enrichment
PHF6
PHF6 enrichment
K4me1 enrichment
10
8
6
4
2
PHF6 enrichment
0.4
0.3
0.2
0.1
0
K4me3
K4me3 enrichment
PHF6
PHF6 enrichment
K4me3 enrichment
3
2
1
0
PHF6 enrichment
0.08
0.06
0.04
0.02
0
A. B.
C.
D.
-2
-1
0
1
2
-2.0 -1.5 -1.0 -0.5 0.0 0.5
log10(PHF6)
log10(K4me1)
R
2
=0.62
Log
10
PHF6 enrichment
Log
10
H3K4me1 enrichment
-3
-2
-1
0
-3 -2 -1
Log_PHF6
Log_K4me3
Log
10
PHF6 enrichment
Log
10
H3K4me3 enrichment
R
2
=0.091
Correlation of enrichment,
H3K4me1 and PHF6
Correlation of enrichment,
H3K4me3 and PHF6
124

5.2.6. Depletion of PHF6 in human neural crest differentiation causes precocious neural crest
migration.
We have seen in chapter 4 that loss of neural crest timing control is seen when Phf6 is
depleted in vivo in Xenopus embryos, yet we have not, up to this point, addressed the phenotype
of PHF6 depletion in the human neural crest differentiation system. This is an important point, as
linking the biochemical activity of PHF6 that has been discussed in this chapter with the in vivo
phenotype of Phf6 relies on the maintenance of a neural crest phenotype in this differentiation
system. To address this question, we established human embryonic cell lines infected with TRIPZ
shRNA of either a non-targeting control (shControl) or against PHF6 (shPHF6). This system is
doxycycline inducible and induction is associated with expression of TurboRFP alongside of the
shRNA. These cell lines were selected and progressively cell sorted, resulting in 30-40% RFP
positive cells in the shPHF6 line and ~50% RFP positive cells in the shControl line after three days
of induction. These cells were then placed into differentiation in the presence of doxycycline and
imaged progressively through the course of differentiation. Interestingly, shPHF6 cells showed an
interesting proclivity towards early neural crest migration, irrespective of the RFP status of the
earliest migrating cells. That is to say, the earliest migrating cells in the shPHF6 sample were not
necessarily RFP positive (data not shown). There are two possible causes for this phenotype, either
RFP negative cells had some expression of the shRNA, or early migration of initial shRNA
positive cells release signaling molecules, causing the early migration of nearby cells. Despite this
lack of discrepancy in early stage migration of neural crest cells, progressive migration favors
preference for the migration of RFP positive cells and the depletion of RFP positive cells from the
central neural ectodermal cluster (Fig. 5.5 C and D, neural ectodermal cell cluster outlined with a
dotted line) in comparison to the equitable distribution of RFP positive cells in the shControl cell
125
Figure 5.5. Precocious migration of neural crest cells in response to PHF6 depletion. A-D.
Fluorescent imaging of live cells expressing RFP associated small hairpin RNA molecules against
a non-targeted, random sequence (shControl, panels A. and B.) or against PHF6 (shPHF6, panels
C. and D.) at two different time points in the differentiation process, early migration (day 7, panels
A. and C.) and later migration (day 9, panels B. and D.). Edges of the neural ectodermal cluster
are outlined in white dashes. Representative images of >30 clusters per trial. E. Western blot
analysis of the effective level of shRNA knockdown of PHF6 in a 100% RFP positive HEK293T
cell population. Doxycycline induction of the PHF6 shRNA leads to ~90% protein level reduction
in PHF6 (panel i), while levels of loading control Tubulin (panel ii) remain consistent.
Representative image of 2 experiments. F.-I. Phase images corresponding to the fluorescent
images A-D., indicating the presence of a neural ectodermal cluster in the outlined region.

shControl shControl
shPHF6 shPHF6
Early migratory NCC (day 7) Migratory NCC (day 9)
PHF6
a-tubulin
shControl + + + + - - - -
shPHF6 - - - - + + + +
Dox - - + + - - + +
Western Blot
293T
A.
B.
C. D.
E.
F.
G.
H.
I.
126
line between migrating neural crest cells and neural ectodermal cluster (Fig. 5.5 A and B). Western
blot assay of infected 293T cells with the same set of shControl and shPHF6 viruses show that
doxycycline induction of shPHF6 for 36 hours leads to ~90% reduction in PHF6 protein in a
strongly RFP positive cell population (Fig. 5.5 E, representative of 3 such experiments).
Corresponding phase images to the RFP images are seen in Fig. 5.5 F-I, which verifies the presence
of neural ectodermal cells in Fig. 5.5 D despite complete lack of RFP signal from within the dotted
line region. This data clearly indicates a preference for PHF6 depleted cells to migrate early from
the neural ectodermal precursor population. Further investigation is required to determine whether
the PHF6 depleted cells experience precocious activation of the migratory neural crest
transcriptional program, with activation of SOX9 and TWIST1. Given the binding of PHF6 to the
enhancers for these genes and the precocious migration of neural crest cells upon PHF6 depletion
within this system, we hypothesize that depletion of PHF6 within this system would cause
precocious activation of these transcription factors and others within the migratory neural crest
transcriptional program.

Section 5.3. Discussion
5.3.1 Assessing the role of PHF6 in chromatin remodeling
The biochemical activity of PHF6 has been the subject of many recent publications. While
some groups have focused on the structure of the protein
183,190,202
, others have focused on the
complexes that it forms
187,196
, and yet others have focused on the binding of the protein to
chromatin
188,189
. Despite this flurry of information, only one cohesive study has been published in
recent years on the biochemical activity of this protein within the developmental setting
116
. Indeed,
perhaps due to the relative instability of this protein and the relative insolubility of the N-terminal
127
region of the protein, many groups have chosen to study this protein through introducetion of
exogenous constructs into cancer cell lines
189,196
, or to study endogenous protein some of the longest
cultured cells known to the scientific community
187,189
. While this does not detract from the quality
of the data gained from these studies, it does leave un-answered questions as to the changes in
protein functionality in these settings. Study of endogenous functionality, while more difficult to
achieve, allows for more direct conclusions to be drawn. This is the true strength of human
embryonic differentiation systems. It allows us to draw conclusions about endogenous activity,
and assay endogenous interactions, without encountering the hurdles associated with deriving the
material for such assays from embryos. However, these systems are not without their drawbacks.
The relative youth of stem cell culture systems and the difficulty of gaining and establishing cell
lines leaves many points of their culture still debated. Indeed, the maintenance of true “stem-ness”
is the subject of intense research, and hundreds of journal articles have been written about this
topic to date. These drawbacks certainly affect the neural crest cell differentiation system, and as
discussed in chapter 2 extreme care has to be taken to ensure identical culture conditions between
experiments in order to produce reproducible results.
The observation of PHF6 bound to CHD7 is an extension of existing literature, in that
immunoprecipitation experiments in other studies found an association between PHF6 and another
CHD family member, CHD4
187
. In this context, this association is extremely important as CHD7
mutation leads to CHARGE syndrome, a birth defect syndrome that involves a marked reduction
and dysmorphia of the craniofacial structures, among other defects
23–26
. In fact, CHARGE syndrome
and Börjesson-Forssman-Lehmann syndrome (BFLS) have remarkably contrasting phenotypes
across a number of different tissues, including craniofacial defects, external ear phenotypes,
subcutaneous fat, and digit defects (Fig. 5.6 A). These striking contrasts in patient
128
Figure 5.6. Relevance of PHF6 in the scope of neural crest epigenetics. A. Contrasting
phenotypes between PHF6 mutation in BFLS and CHD7 mutation in CHARGE in craniofacial
abnormalities, external ear phenotypes, obesity phenotypes, pigment defects, and digit defects. B.
Model of PHF6 activity in pre-migratory neural crest (left side) and neural crest (right side). C.
Heatmap enrichment of histone modifications H3K4me1 and H3K27ac in pre-migratory neural
crest (left side) and migrating neural crest (right side).

129
phenotypes indicate that the antagonistic complex described in this chapter may be a common
mechanism across several different tissues in development. Further work in the other tissues of
origin will determine if this is the case.
Mutual exclusivity of PHF6 and pBAF complex members in a complex with CHD7 may
indicate a complex that is prevented from performing active chromatin remodeling. Current
understanding of PHF6 activity indicates that the second PHD motif of this protein binds non-
specifically to DNA
183
, an observation which does not come into significant contestation by the data
presented in this chapter as I only observed a weak interaction between my recombinant PHF6-
DF1 and H3 protein (Fig. 5.2B panel i), which could easily be explained by a passive interaction.
If we combine the understanding that the second PHD motif binds non-specifically to DNA with
the conclusion that the first PHD motif binds to H3K4me1 and that this protein may dimerize, the
resultant structure of PHF6 bound to the nucleosome resembles a clamped wheel structure (Fig.
5.6B, left side). If we then add to this the interaction of CHD7 at these loci, but without the
chromatin remodeling pBAF complex, it becomes unlikely that CHD7 would be capable of
performing active chromatin remodeling in this context due to the inaccessibility of the
nucleosome for remodeling or due to the loss of assistance in remodeling by the pBAF complex
(Figure 5.6B left side). It is currently not well understood whether the pBAF complex or CHD7
directly facilitates chromatin remodeling at these enhancer loci or nearby promoter regions to
promote transcription, only that the two act synergistically to promote active neural crest
transcription
29
. ChIP-seq analysis of the enhancer loci across the genome suggests that active
chromatin remodeling must take place, as approximately 75% of these loci transition from a
nucleosome occupied and occluded region in the pre-migratory neural crest time point to gaining
a central nucleosome depleted region that is positive for transcription factor p300 at the migratory
130
neural crest time point (Fig. 5.6 C, analysis performed by Richard Pelikan, data from Rada-Iglesias
et al., 2012
92
).
The observation that PHF6 protein likely binds to H3K4me1 histone tails and that this
protein tracks with high levels of H3K4me1 throughout independent ChIP experiments support
one another. Fig. 5.2 gives strong evidence for an interaction between H3K4me1 and PHF6, but
at this point there is no conclusive evidence for a direct interaction between the two proteins. In
fact, Fig. 5.2E indicates that there is an additional confounding modification necessary for the
interaction between PHF6 and histone tails with the PTM H3K4me1. However, the observation
that PHF6 tracks with H3K4me1 across the genome in experiments signifies that while this
additional modification may provide an additional level of specificity to the targeting of PHF6,
H3K4me1 is an important modification involved in the mechanism of activity of PHF6.
The combination of the biochemical results from this chapter, along with expression data
found in chapter 3 and developmental role of PHF6 in chapter 4 leads to an overall model of the
role of PHF6 in epigenetic regulation of neural crest migration. Within neural crest cell precursors
in the neural plate border, PHF6 binds to neural crest enhancers and nearby DNA and prevents
enhancers from becoming remodeled and activating nearby promoters by forming a clamped-
wheel structure around the nucleosome, and yet sequestering CHD7 to these regions (Fig. 5.6B,
left side). Downregulation of PHF6 upon neural crest migration both at the RNA and protein level
through a currently unknown mechanism then frees CHD7 to perform active chromatin remodeling
with the pBAF complex, allowing for active neural crest transcription to occur (Fig. 5.6B, right
side). The purpose of this mechanism is to prevent precocious migration of neural crest cells,
particular in the more posterior pharyngeal arch regions of the embryo thereby allowing pathway
and positional identity to be conveyed.
131
5.3.2 Critical evaluation of the PHF6 model
This model of PHF6 activity represents the most likely scenario given the data uncovered
in this chapter. However, this model could represent an over-simplification of the real biological
picture. Particularly, in light that PHF6 tracks to primarily active promoters in Jurkat cells
189
, there
may be additional complexity necessary for targeting PHF6 in the genome. PHF6 is not mutated
in the Jurkat cell line
203
which eliminates the possibility of direct binding pocket changes. However,
PHF6 is not only targeted to active promoters, but also repressed and bivalent promoters as well.
Since the distribution of H3K4me1 was not assayed in this cell in reference to the PHF6 ChIP, we
cannot know the distribution of this PTM under the PHF6 ChIP conditions, yet it is unlikely that
it is distributed to silenced promoter regions. Therefore, we must conclude that there is either an
additional measure of complexity, either a protein-protein interaction or alternate chromatin
modification that PHF6 binds, to the targeting of PHF6 within the genome, or that an alternate
mechanism of PHF6 activity reigns within Jurkat cells. PHF6 has been found to be a target of
several kinases
204–206
, and perhaps assaying the phosphorylation status of PHF6 in these two cell types
could shed light on how a differential mechanism of PHF6 activity is achieved.
Furthermore, it is altogether possible that PHF6 does not bind directly to H3K4me1, but
rather a protein-protein interaction between PHF6 and another protein causes targeting to loci with
this modification. Consistent problems with PHF6-FL and PHF6-DF2 solubility in protein
purification assays prevented a direct assay between recombinant PHF6 and H3K4me1 modified
biotinylated peptide, which would provide evidence for a direct interaction. A tertiary interaction
is therefore possible, but improbable. The consistency of the co-immunoprecipitation between
PHF6 and H3K4me1, and the strength of the interaction despite multiple different manipulations
(Fig. 5.2 C and E, and co-immunoprecipitation with high salt extracted histones, data not shown)
132
suggests that this interaction is extremely difficult to titrate away despite salt and interfering
peptide concentrations within the immunoprecipitation reaction. In and of itself, this kinetics
suggest a strong, primary interaction between PHF6 and H3K4me1.
The interaction between PHF6 and CHD7 and therefore the idea that CHD7 would be
present at the enhancer regions at the point of necessary remodeling is also highly probable. The
interaction between PHF6 and CHD7 has been observed in multiple co-immunoprecipitation
experiments presented here, as well as in immunoprecipitation with mass spectrometry analysis
experiments performed by Dr. Bajpai in the lab of Joanna Wysocka. Furthermore, CHD7 ChIP
experiments performed in Bajpai et al.
29
identified CHD7 enrichment at the same SOX9 enhancer
and TWIST1 enhancer that was assayed in experiments performed within this chapter. CHD7
enrichment was specifically observed at the point of neural crest migration, and not assayed in the
pre-migratory neural crest population, yet this observation is consistent with the model presented.
The specific mechanism of PHF6 downregulation at both protein and mRNA levels is an
important part of the overall mechanism presented here, but has not been addressed at this point.
Presence of several FOXD3 binding sites in the promoter region of PHF6 (data not shown),
predicted by FOXD3 matrix binding
207
, suggests the possibility of transcriptional repression of
PHF6 by FOXD3. This could be addressed through ChIP experiments involving FOXD3
immunoprecipitation followed by qPCR assessment of enrichment of the PHF6 promoter.
Additionally, the 3’UTR of PHF6 contains binding sites for multiple micro RNAs, most notably
miR128-3p, which has been identified as an oncomiR due to the targeting of PHF6 in T-ALL
samples
208
. Targeting of PHF6 mRNA for translation blocking and mRNA degradation through a
micro-RNA dependent pathway could easily account for a steep drop-off in level of PHF6 mRNA
and/or protein in migrating neural crest cells. Furthermore, the identification of multiple
133
phosphorylation sites in PHF6
204–206
could provide a mechanism for phosphorylation-dependent
proteasome degradation for existing protein. One or more of these mechanisms could be at work
at the junction between pre-migratory neural crest and migratory neural crest and could provide
for the downregulation of mRNA and protein in this region.

Section 5.4. Methods
5.4.1 Immunoprecipitation
Immunoprecipitation reactions followed the general outline from Chapter 2. Each antibody was
tested specifically as to the optimal salt concentration and input protein level, listed in Table 5.1
below.

134
Table 5.1 Immunoprecipitation antibody information.
Antibody Product Information IP salt concentration
Input protein per µg antibody
PHF6 Bethyl A301-451A 150 mM
300 µg
CHD7 Bethyl A301-223A 250 mM
500 µg
BRG1 Crabtree lab 150 mM
500 µg
IgG Sigma 12-370 150 mM
300 µg
H3K4me1 Abcam ab8895 200 mM
250 µg

135
5.4.2 Western Blot
Western blotting was performed using standard 4-15% single concentration or gradient SDS-
PAGE gels, depending on the size resolution required for the particular protein, followed by
Turbo Transfer (BioRad) to nitrocellulose membrane, blocking using Roche blocking reagent,
overnight incubation with antibody at the notated dilution. This was followed by TBST washes,
incubation with a secondary HRP conjugated secondary antibody at 1:10,000 dilution, and
exposure with Roche ECL reagent.

5.4.3 Molecular Cloning
PHF6 constructs were cloned into pGOOD6P vector after PCR amplification. These amplicons
were then digested using BamHI and XhoI double digest conditions, and cloned into digested
pGOOD6P vector. Each clone was sequence verified before transformation into Arctic Express
(Aligent).

5.4.3. Chromatin Immunoprecipitation
Chromatin Immunoprecipitation was performed using the methodology described in Chapter 2.
Antibody information, including product information, input chromatin amounts, and ChIP
antibody quantity used can be found in Table 5.3 below.

136
Table 5.2. Chromatin Immunoprecipitation antibody information.
Antibody Product information Input Chromatin Antibody Quantity
PHF6 Bethyl A301-451A
250 µl 3 µg
H3K4me1 Abcam ab8895
125 µl 1.5 µg
H3K27ac Abcam ab4729
125 µl 1.5 µg
H3K4me3 Active Motif 39159
125 µl 1.5 µg
IgG Sigma 12-370
250 µl 3 µg

137
Chapter 6: A role in mesoderm formation, and other PHF6 details.

Section 6.1. Introduction
Patients with Börjeson-Forssman-Lehmann Syndrome, on top of a distinct craniofacial
phenotype, have many other defects including intellectual disability, obsesity with gynocomastia,
hypogonadism, large external ears, and other developmental disorders
182,209,210
. The complexity of this
disorder demonstrates that the PHF6 role has diverse roles in development in other cell types than
the neural crest and neural precursors. The developmental expression pattern of PHF6 allows for
a potential roadmap of the cellular roles of PHF6 but not the potential mechanisms that are
involved in these cell types. This mechanism may be partially or completely maintained in other
cell types, or protein interactions may be highly diverse depending on the cell type. What is clear
is that our understanding of PHF6 is far from complete.
The overall interactions surrounding this protein, which thus far has shown itself to be a
chromatin adapter protein
211
, are likely to be highly complex. Immunoprecipitation experiments
followed by mass spectrometry analysis of protein interactions in HEK293 cells revealed several
kinds of interactions: those with chromatin remodelers, those with spliceosome factors and RNA
processing factors, those with histone proteins, and other miscellaneous factors
187
. Characterization
of each of these interactions is necessary to get a complete picture of the activity of this protein,
and analysis of these complexes and the extent to which they are formed in each cell type where
PHF6 is expressed is necessary to fully comprehend the role of this protein in development.
In this chapter, I will offer some evidence that I have gained in my work on some additional
protein-protein interactions that PHF6 forms within the context of neural crest. I will also offer
some additional observations that we have come across during the course of study of Phf6 in the
138
Xenopus laevis developmental system. Although these observations are preliminary, they can
provide a framework for additional study of this protein within the developmental context, and can
provide a start for further systems-based analysis of the role of PHF6 in development.

Section 6.2. Results
6.2.1. PHF6 protein interacts with multiple pluripotency-associated factors
PHF6 immunoprecipitation followed by mass spectrometry analysis of co-
immunoprecipitated proteins (performed by Ruchi Bajpai, analyzed by Minjia Tan in the lab of
Yingming Zhao at University of Chicago) in neural ectodermal cells found a number of
pluripotency-related factors as potential binding partners of PHF6. These included DPPA4,
IGF2BP2, and Lin28. In an effort to validate these interactions, I performed a GST pulldown assay
whereby I used the recombinant PHF6 proteins described in Chapter 5 to precipitate PHF6 binding
proteins in neural ectodermal nuclear extract, and assayed PHF6 binding by western blot using
antibodies against DPPA4 and IGF2BP2. The recombinant PHF6 proteins all precipitated some
IGF2BP2 specifically (Fig. 6.1 panel i), with PHF6-FL and PHF6-DF2 enriching this protein to
the highest level of the recombinant proteins despite much lower overall PHF6 protein (Fig. 6.1
panel iii). Interestingly, in the DPPA4 western blot, I observed a slight shift in the position at which
the protein DPPA4 ran, which may indicate a slightly different pH in the precipitated samples
versus the input sample, or may indicate post-translational modification of the DPPA4 protein
(Fig. 6.1 panel ii). Additionally, there was a non-specific interaction of a protein recognized by the
antibody (Fig. 6.1 panel ii, demarcated by the *, top band). However, of the specific interactions,
DPPA4 40 KD was precipitated with all of the PHF6 recombinant proteins, and enriched to the
highest extent by PHF6-DF2 (Fig. 6.1 panel ii) despite much lower protein level (Fig. 6.1 panel
139
Figure 6.1. GST-pulldown analysis of recombinant proteins with IGF2BP2 (panel i) and DPPA4
(panel ii), verified by GST expression (panel iii). *, non-specific band. Single experiment.

⍺-IGF2BP2
⍺-DPPA4
⍺-GST
70 ⎯
40 ⎯
40 ⎯
Figure 6.1
* Non-specific
i
ii
iii
140
iii). PHF6-FL and PHF6-DF2 precipitation reactions both contained lower protein levels due to
protein instability, detailed further in Chapter 2 and 5. These results should be followed up with
co-immunoprecipitation experiments, but overall indicate a predilection of PHF6 to interact with
DPPA4 as well as IGF2BP2.

6.2.2. Phf6 depletion in embryos leads to a high rate of gastrulation defects
Results observed in Chapter 4 were performed using a 1 in 8 cell microinjection strategy
in order to target the neural and neural crest derivatives specifically and to limit morpholino
presence in mesoderm and endoderm derivatives. In initial experiments, I observed that broader
depletion of Phf6 led to increased embryo mortality. I observed a statistically significant increase
in the incidence of gastrulation failure (Fig. 6.2A, p=0.04 by one-tailed student’s T-test, n=4) in
Phf6 morphant embryos over multiple clutches when compared to control morphant embryos when
targeted microinjection was performed at NF2 (Fig.6.2C, injection strategy). In order to further
investigate the observed gastrulation defect, I filmed gastrulating embryos over the course of 6
hours, taking images every 2 minutes over this time course. I then calculated the average time until
the blastopore reached its minimal diameter in number of frames (Fig. 6.2B, n=8) and noticed that
there was a statistically significant delay in blastopore closure in the Phf6 morphant embryos
(p=0.049 by one-tailed student’s T-test). Blastopore closure delay was calculated per clutch by
taking the frame where the earliest embryo had completed gastrulation and subtracting this value
from all other points of gastrulation completion, thereby indicating how many frames after the
earliest each embryo completed gastrulation.

141
Figure 6.2. Gastrulation analysis upon Phf6 depletion. A. Incidence of gastrulation failure across
4 separate experiments where Control morpholino and Phf6 morpholino were microinjected at 1
in 2 cell stage at 4 µM final concentration. *, p=0.04 in one-tailed student’s t-test against control
morphant embryos. B. Quantitation of blastopore closure delay observed during imaging of
gastrulation stage embryos, n=8 for each. *, p=0.049 in one-tailed student’s t-test against control
morphant embryos. C. Schematic of injection strategy used in analysis of gastrulation phenotype.
D. Timecourse of gastrulating embryos in control morpholino injected embryos, Phf6 morpholino
injected embryos, and Phf6 morpholino plus PHF6 mRNA injected embryos at the indicated
concentrations. Time is indicated after the start of imaging in minutes. White arrow indicates the
point of bottle cell constriction failure. Error bars indicate SEM.

Time, min 80 120 160 200 240 280 320 360
Control MO,
4 μM
Phf6 MO, 4
μM
Phf6 MO, 4
μM
PHF6
mRNA, 500
pg
Figure 6.2
0
10
20
30
40
50
60
70
80
Control PHF6 mRNA Phf6 MO Phf6 MO +
PHF6 mRNA
Blastopore Closure Delay
*
0
10
20
30
40
50
60
Control Phf6
Percent of total injected embryos
Morpholino, 4 μM
Gastrulation Failure
*
D
A B
NF2
NF19
C
142

Analysis of the dynamics of blastopore closure across the various groups also indicated a
characteristic phenotype. Figure 6.2D shows a representative embryo across control morphant,
Phf6 morphant, and Phf6 rescue groups. Interestingly, at 160 minutes, we can observe that the
bottle cells which had previously formed in the Phf6 morphant (see 120 minutes) fail to constrict
(Fig. 6.2D, white arrow) along the dorsal side. Despite this loss, the remaining sides of the
blastopore continue to constrict, and we can observe what appears to be the posterior neural plate
forming despite this loss of internalization of the mesoderm and endoderm at 320 minutes. This
phenotype is not present in either the control morphant group nor the rescue group, suggesting that
this phenotype is likely specific for loss of Phf6. This loss of bottle cell constriction was initially
observed by Dr. Katherine Pfister, an expert in convergence and extension during gastrulation
from Dr. Raymond Keller’s lab at University of Virginia. This loss of bottle cell constriction during
gastrulation would likely lead to incomplete formation of mesodermal structures in later embryos,
and ultimately high mortality of embryos. Considering the fact that previously described
phenotypes involving Phf6 depletion caused the dysregulation of temporal control of neural crest
migration, a similar effect may occur in this case. This constriction of bottle cells is extremely
highly conserved in developing embryos across species
212,213
, and slight changes in the expression of
downstream effectors including those involved in WNT signaling in this region
214,215
would have
devastating effects on bottle cell constriction and the collective cell migration resulting from this
initial event.

143
6.2.3. Phf6 depletion in half of the embryo results in failure of neural crest migration, potentially
through a mesoderm-mediated mechanism.
Interestingly, even in those embryos that proceed to undergo gastrulation seemingly
normally, there appears to be continuing developmental process failure. A set of embryos were
microinjected at NF2 and selected for those that completed gastrulation and expressed the intended
construct in half of the embryo. These were then fixed at NF 19, and in situ hybridization was
performed for neural crest marker Twist, much like those seen in Chapter 4 (Fig. 6.3A, schematic).
Interestingly, these embryos showed the opposite phenotype from that observed in embryos with
targeted neural crest injection. Here, embryos depleted for Phf6 using either morpholino against
Phf6 splice junctions (Fig. 6.3B, panels ii and iii) showed a trend towards a collapse in neural crest
migration on the targeted side. This tended to be rescued by co-injection with PHF6 mRNA (Fig
6.3B, panel v). Overall, representative embryos depleted for Phf6 and co-injected with GFP
mRNA did not rescue the phenotype (Fig 6.3B, panel iv). Quantitation of the ratio of staining on
the affected side versus the unaffected side for a control GFP targeting morpholino, the first Phf6
morpholino, the second Phf6 morpholino, the first Phf6 morpholino with GFP mRNA and the first
Phf6 morpholino with PHF6 mRNA showed an average ratio of 0.85, 0.61, 0.67, 0.87, and 1.03
respectively (Fig. 6.3C), though due to variations in this phenotype I observed no statistically
significant differences in neural crest migration. Embryos which have lost migration of neural crest
cells express Twist in the neural plate border region, and therefore this region is likely competent
to form migratory neural crest; however, this should be verified with additional neural crest in situ
hybridization experiments for neural plate border markers like Zic1 and Msx2. Given the
prevalence of gastrulation defects in embryos similarly depleted for Phf6 described above, this
failure of neural crest migration is likely to be a downstream effect of partial failure of gastrulation
144
Figure 6.3. Neural crest cell migration failure upon systemic Phf6 depletion. A. Injection strategy
associated with systemic expression in half of the embryo. B. Characteristic Twist expression
pattern at NF19 after in situ hybridization. * indicates the affected side. C. Quantification of Twist
staining in each embryo with means depicted as a bar, and the standard error depicted with error
bars. D. Quantitation of the variability of the phenotypes, expressed as percentages of the total
number of embryos with reduced, equivalent, or increased Twist staining on the targeted side per
condition. Single experiment, n>5 per condition. Error bars indicate SEM.

GFP MO Phf6 MO1 Phf6 MO2
Phf6 MO1
PHF6A mRNA
Phf6 MO1
GFP mRNA
Twist
NF19
* * * * * *
MO: 4uM
NF2
NF19
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
GFP MO Phf6 MO1 Phf6 MO2 Phf6
MO1+GFP
Phf6 MO1+
PHF6A
Twist staining on targeted side
Less
Even
More
GFP MO
Phf6 MO1
Phf6 MO2
Phf6 MO1+ GFP
Phf6 MO1 + PHF6A
Cherry
0.0
0.5
1.0
1.5
2.0
2.5
Twist Ratio of Stained Areas
GFP MO
Phf6 MO1
Phf6 MO2
Phf6 MO1+ GFP
Phf6 MO1 + PHF6A
Figure 6.3
A
B
C D
i ii
iii
iv v
145
despite apparent progression through this developmental phase and therefore incorrect
specification of the mesodermal cell types. The manifestation of the incorrect specification of
mesodermal cell types would involve a partial or complete lack of signaling from the mesodermal
structures to specify normal neural crest and migration of neural crest
70
. This hypothesis could be
further addressed with in situ hybridization for axial and paraxial mesodermal structures, and
transplant of wild-type neural plate border regions to this background.

Section 6.3. Discussion
6.3.1. Additional PHF6 Biochemical interactions point to formation of an RNA processing
complex.
Despite the relative simplicity of this protein, given the two PHD motifs and no defined
enzymatic activity, previously published data as well as data produced within our group indicates
that PHF6 is likely to have multiple functions both within neural crest cells and without. Our data
and the published IP-MS/MS data suggests that multiple binding partners of PHF6 also bind RNA
either specifically or as part of a general splicing mechanism. One such protein is IGF2 mRNA
binding protein 2 (IGF2BP2 or IMP2), which binds IGF2 mRNA transcripts and sequesters them
to allow for timed translation of the mRNA
216,217
. This interaction was identified as part of the IP-
MS/MS experiment performed in our group, and is verified here. This suggests that PHF6 may be
involved in an RNA processing complex, along with other mRNA binding proteins and processing
proteins. Further verification of other RNA binding proteins identified within the IP-MS/MS
experiments and performing RNA-immunoprecipitation experiments with PHF6 would help
identify the extent to which PHF6 is involved in RNA processing and identify the RNAs that PHF6
binds in complex. The complex between PHF6 and IGF2BP2 is interesting within the context of
146
neural crest in that IGF2BP2 activation has been correlated with increased cellular invasiveness,
metastasis, and poor prognosis in multiple cancers
218,219
. Considering the increased cellular
invasiveness phenotype observed with Phf6 depletion in Xenopus laevis explants (Fig. 4.8 and
4.9), the hypothesis that PHF6 forms an inhibitory complex with IGF2BP2 is consistent with PHF6
loss leading to increased IGF2BP2 activation and increased cellular invasiveness.
The identification of binding to developmental pluripotency associated protein 4 (DPPA4)
is less clear in terms of the type of complex this interaction forms. This protein is a downstream
target of the Sox2/Oct4 pathway
220,221
and relatively little is known about its overall protein function
other than it has an interaction with polycomb repressor complex 1 component PGCF1
222
. Based on
this interaction, it is assumed that this protein may be involved in epigenetic maintenance of
pluripotency, and it is possible that this protein also operates within the PHF6/CHD7 complex or
a completely separate epigenetic complex.

6.3.2. Failure of bottle cell constriction suggests to a role for PHF6 in the blastopore lip, early
mesoderm formation.
In this chapter, we show evidence for failure to gastrulate due to Phf6 depletion. When we
observe embryos depleted for Phf6, we can observe a lack of bottle cell constriction during
gastrulation, ultimately resulting in the gastrulation failure. This phenotype is also rescuable by
co-injection of humanized PHF6 mRNA, suggesting the specificity of this phenotype. In chapter
3, we can observe Phf6 mRNA expression broadly throughout the prospective ectoderm and
mesoderm during gastrulation, and throughout the blastopore lip region. This combined data
suggest that Phf6 has a role in gastrulation in the dorsal blastopore lip and in the correct
constriction of the apical membrane of bottle cells during gastrulation. This loss of bottle cell
147
constriction and invagination would result in the failure to establish the three germ layers in a
correct manner, and would likely result in a partial loss of mesoderm formation. This conclusion
is further supported by the observation that Phf6 depletion on one side of the embryo, even when
gastrulation is completed, leads to loss of migratory neural crest, which is contrary to the
observations within Chapter 4. This loss of migratory neural crest despite correct location of the
neural plate border suggests that perhaps the underlying mesoderm is not correctly formed, and
fails to provide the appropriate signals for correct neural crest formation. Somite and notochord in
situ hybridization or immunofluorescent staining in these embryos could provide evidence as to
the verity of this hypothesis.
The mechanism proposed here that PHF6 protein regulates the temporal activation of
neural crest enhancers and thereby regulates the timing of neural crest specific transcription
through this epigenetic regulation could also be involved in the temporal regulation of gastrulation
and the expression of transcripts crucial to this collective cell migration. While the epithelial-to-
mesenchymal transition of neural crest involves a different type of collective migration, we
nevertheless have a signaling input that results in the creation of a migratory phenotype resulting
in collective cell migration in both neural crest epithelial-to-mesenchymal transition and in bottle
cell constriction during gastrulation. It is therefore within the scope of this work to postulate that
PHF6 may act in an epigenetic manner to restrict collective cell migration overall. This is also
consistent with the observation that PHF6 has been identified as an oncogene and is one of the
highest-frequency mutated proteins across cancer types
223
.

148
6.3.3 Severe loss of PHF6 causing gastrulation defects points to additional insights into the
genetics of BFLS.
Depletion of Phf6 throughout one side of the embryo causes severe gastrulation
phenotypes in Xenopus laevis embryos, yet human male patients are able to survive to birth and
beyond with one mutated copy of the PHF6 gene. This indicates that these patients might not have
a complete loss of the functionality of this protein. If we examine the mutations observed within
male patients (Fig. 3.4), we see that the majority of these mutations are point mutations, often
conservative, or nonsense or frameshift mutations resulting in early stop codons within the protein.
Further analysis of the human mature mRNA has indicated multiple in-frame start codons within
the first 3 coding exons of the protein, indicating early nonsense or frameshift mutations may not
cause complete loss of function of the protein. Overall, these mutations may result in partial loss
of function of the protein in male patients rather than complete loss of function, which would allow
patients to develop much milder defects than would otherwise be the case with complete loss of
protein function.

6.4. Methods
6.4.1. Microinjection and imaging
Xenopus laevis embryos are prepared for microinjection as described in Chapter 2.
Embryos are microinjected with the appropriate concentrations of control, Phf6 morpholino and/or
PHF6 mRNA with Ruby Dextran, 2.5 nM, at the end of NF 2 to prevent contamination of the other
blastomere. Embryos were then allowed to rest for 1 hour in Ficoll, and placed into a 1/3x MBS
solution to develop to mid-blastula stage. Embryos were then checked for expression of the dextran
on one half of the embryo, and select embryos had their vitelline envelopes removed and were
149
positioned in a clay dish to develop under a standard inverted microscope with a black and white
camera attachment for 6 hours, with one frame taken every 2 minutes.

150
Chapter 7: The role of PHF6 in chromatin biology and development: perspectives and
conclusions.
Section 7.1. PHF6 and chromatin remodeling.
PHF6 is an interesting protein to be involved in chromatin remodeling, in that it has no
discernable enzymatic activity. The structure of the protein, the topic of several recent studies
183,190,202
,
contains two plant homeodomain regions, with an additional “zinc knuckle”, or additional zinc
coordination region, N-terminal to the second PHD and proposed to be N-terminal to the first
creating extended PHD motifs (ePHD). Other than these domains, this protein contains very little
defined structure. The structure of the protein alone suggests that it would operate as a chromatin
adapter protein, as PHD regions are typically associated with modified histone tail binding. Despite
this, initial publications from other laboratories suggested that this protein would act as a
transcription factor
224
. The findings presented within this work suggest that this protein would
indeed act as a chromatin adapter protein, and while it may indeed act to affect transcription, it
does not do so directly.
At the end of Chapter 5, I proposed a mechanism of PHF6 activity within neural crest cells.
This mechanism proposes that PHF6 binds to designated enhancer regions at a stage when they
are marked for activation by H3K4me1 but are not yet fully activated. Current understanding of
enhancer activation suggests that H3K4me2 replaces H3K4me1 to a certain extent, and that the
region gains H3K27ac as well as a central nucleosome depleted region that binds transcription
factors. The extent to which active enhancers are marked by H3K4me1 versus H3K4me2 is
currently under intense debate
14,225–228
. Empirical evidence using ChIP-seq and FAIRE-seq within
human embryonic stem cell system differentiated to neural crest cells used here (Fig. 5.6) suggests
that the three enhancers that are observed to be enriched in PHF6 chromatin immunoprecipitation,
151
however, do not gain p300 binding or a central nucleosome depleted region until active neural
crest migration occurs. Once neural crest migration occurs, PHF6 has been observed to be
downregulated at the protein and mRNA levels, and PHF6 binding decreases at these enhancers.
We hypothesize that this reduction in PHF6 presence at these enhancers allows for active
chromatin remodeling to occur, promoting neural crest specific transcripts to be made.
There are several ideas inherent to this model. First, I will address the idea that chromatin
remodeling proteins like CHD7 have to be checked in their activity. This idea arises from the fact
that we still have very little understanding of how a chromatin remodeler can act to promote one
cell type over another. We observe that CHD7 promotes the neural crest cell type in a specific
manner
29
, yet we do not completely understand how this would occur. ATP-dependent chromatin
remodelers could easily promote differentiation of neuronal cell types over neural crest, or
maintenance of the current cell type, simply by removing nucleosomes in neuronal enhancers
instead of neural crest enhancers, or masking neural crest specific transcriptional start sites by
moving nucleosomes into these regions. The activity of ATP-dependent chromatin remodelers
therefore has to be regulated extrinsically, through interactions with other proteins and attraction
to certain regions within the genome by these proteins, or intrinsically through protein post-
translational modifications on the protein itself, which changes an element of its activity. Some
combination of these two is likely. The mechanism that we are proposing here involves a measure
of extrinsic regulation, whereby CHD7 is attracted to the regions where it will become active
through protein-protein interactions (Fig. 5.1 and 5.3), but physically prevented from performing
its activity through a steric hindrance method, i.e. physically blocking the interaction between the
enzyme and its substrate (Illustrated in Fig. 5.6). Downregulation of PHF6 would therefore
alleviate this steric hindrance and lift this inhibition.
152
This idea implies that PHF6 has a higher level of specificity for certain regions in the
genome than could be accomplished by CHD7 alone. Preference for H3K4me1, while offering a
certain measure of specificity, is unlikely to be the only modification bound by PHF6 for several
reasons. First, we observe that the binding of PHF6 to H3K4me1 is not sufficient for the overall
interaction between PHF6 and modified H3 (Fig. 5.2). Secondly, H3K4me1 is thought to be a
highly promiscuous post-translational modification, at times being deposited in regions with no
known functionality (data based observations). Additional, combinatorial specificity for dually
modified histone tails would therefore provide a higher order of specificity. The additional
modification is likely to involve one of the immediately surrounding H3 residues, perhaps one of
the nearby arginines, as the PHD pocket is relatively small. RAG2 is one example of a PHD-
containing protein that binds dually modified histone tails, and binds modified H3K4me3 and
H3R2 that has been symmetrically dimethylated
229
. Specificity for a lack of modification on another
nearby residue of H3 could also lend an additional mechanism for specificity of PHF6 to specific
genomic regions. Multiple PHD proteins are prevented from binding histone tails by certain
modifications
191
. Therefore, PHF6 could provide the appropriate level of specificity to target CHD7
while preventing immediate activity of this enzyme.

Section 7.2. Enhancer activity and timing of neural crest migration.
Inherent within the model presented in this dissertation is the idea that the set of neural
crest enhancers have activity that is indispensable for the correct timing of neural crest migration.
The set of neural crest specific enhancers are highly cell type specific, with approximately 80% of
these enhancers becoming specifically activated in neural crest cells and not in the precursor neural
ectodermal cell type nor in human embryonic stem cells
92
. The specificity of these enhancers is on
153
the same order as other cell-type specific enhancers
230–232
. In fact, increasing understanding of
transcription in recent years has highlighted that with only 25,000 total transcripts, there are
relatively few that are truly cell-type specific. What is more cell-type specific is the mechanisms
by which the level of transcription from each promoter is controlled to allow the correct amount
of a transcript to be created at the correct time in the cell of interest. One such way in which
transcriptional level and time is coordinated is through transcriptional enhancement, whereby a
region of DNA has a certain affinity for a set of transcription factors, which recruit polymerases
with a specific efficiency and perform a DNA looping activity with a certain set of kinetics. This
provides an elegant model by which transcriptional level can become highly cell type specific and
regulated. We observe that when enhancers do not become activated, as is the case with Chd7
depletion
29
or PHF6 overexpression (Fig. 4.6), the embryos lose the ability to form the cellular
phenotype of migratory neural crest cells. This suggests the necessity of enhancer activation for
the ability to create the cellular readout of migration of neural crest. RNA-seq experiments using
PHF6 overexpression and depletion in Xenopus laevis neural crest would allow a more direct
analysis of the effects of PHF6 on overall gene expression, and thereby an indirect measure of the
extent to which the enhancers affected by PHF6 effect gene expression. The utility of neural crest
enhancers in promoting the correct timing of neural crest transcription is also possible to assess
within the neural crest system by creating a fusion of PHF6 with the Glucocorticoid Receptor
(GR). With this, it may be possible to ablate specific waves of neural crest migration through
pulses of nuclear translocation of PHF6 with dexamethasone treatment. This would allow PHF6
to be overexpressed at specific times in the region where it is active, and sequestered away from
the nucleus in the absence of dexamethasone.

154
Section 7.3. PHF6 and restriction of neural crest formation.
An additional idea behind this mechanism is that there is an inherent cell type decision that
has to be made among the cells of the neural plate border and the adjacent neural ectoderm. Gene
regulatory networks established by multiple groups working on neural crest transcriptional
regulation point to a cascade of transcriptional activation initially promoted by intermediate levels
of BMP signaling and bFGF signaling as the primary signals, and reinforced by continuous
feedback loops, resulting in regions of neural crest competency along the neural plate border
70
.
While this gene regulatory network provides an eloquent framework for activation of neural crest
specific transcription, the zone of expression of each of the neural plate border specifying genes is
broader than the region that ultimately forms the neural crest
70
. This indicates that there must be
some cell fate decision that occurs to promote certain cells to become neural crest, while other
immediately adjacent cells can remain and ultimately become part of the dorsal neural tube. This
balance between the two cell types can be observed within the literature; in multiple cases,
restriction of the neural crest cell type causes expansion of the neural tube region
76,233
. Here, we can
observe to some extent the opposite, where expression of a dominant-negative PHF6 mRNA leads
to a reduction in the Sox2 expressing region, as well as a small rate of neural tube closure defects
(Fig. 4.5C). This data supports the idea that there is a cell fate choice at hand, and that specific
activation of a permissive epigenetic state may be inherent to this cell fate decision.
PHF6 is unlikely to be the only epigenetic regulator involved in restricting the permissive
epigenetic state of neural crest migration. As outlined in Chapter 1, there are multiple other
epigenetic factors that, when mutated, cause craniofacial phenotypes which may be described as
hyperplasia of the craniofacial structures. Most notable is the ARID1B protein, which causes most
of the cases of Coffin-Siris Syndrome. Interestingly, this syndrome has such a high rate of
155
phenotypic crossover with BFLS that patients with BFLS have been mistakenly diagnosed as
Coffin-Siris Syndrome patients
234
. ARID1B may also be involved in restriction of the permissive
epigenetic state of neural crest transcription; however, it is unlikely that it functions in the same
complexes described within this work as none of the immunoprecipitation-mass spectrometry
experiments identified this protein as a binding partner of either CHD7 or PHF6. Further research
into ARID1B and other proteins with similar patient phenotypes may unveil new ways in which
this permissive epigenetic state is promoted and restricted.

Section 7.4. Epigenetic regulation of cell fate decisions
The idea that the chromatin environment must allow and support differentiation of different
cell types is not a new idea. This was the basis of John Gurdon’s Nobel Prize winning discovery
that transplantation of an adult cell nucleus into an ablated embryo can yield an entire organism.
The concept was originally postulated in Dr. Waddington’s 1962 epigenetic landscape (Fig. 1.5),
and it is re-emphasized every time we find a new epigenetic factor that increases the efficiency of
induced pluripotent stem cell reprogramming
235,236
. However, the mechanism behind epigenetic
restriction of a cell fate decision is a relatively young idea. The idea of maintenance of a cell fate
by forestalling lineage commitment is the very idea behind embryonic stem cell culture, and the
mechanism by which certain promoters and enhancers are deposited with histone modifications
depicting both positive modifications, “go” signals, and negative modifications, or “stop” signals,
has been termed bivalency. While the mechanism presented here is not identical to these bivalent
promoters and enhancers, it does bear some similarity. For example, neural crest enhancers that
are coated with H3K4me1 and bound to PHF6 are poised for activation via the active enhancer
modification H3K4me1, but prevented from complete activation by PHF6. Resolution of this state
156
is possible either by downregulation of PHF6 and progression into the neural crest cell state or
maintenance of PHF6 expression and continued progression towards the neuronal pathway.
However, this mechanism also differs from the bivalent mechanism in one key way: PHF6 does
not provide a strict stop signal, particularly given the fact that it complexes with CHD7. Rather,
PHF6 provides an activation barrier for cell fate change, much like a hill in Waddington’s
epigenetic landscape.
This restriction of neural crest cell fate through PHF6 binding of neural crest enhancers
may represent a mechanism operating in other cell fate decisions, particularly since PHF6 and
CHD7 show opposing defects in other cell types. Most notable is the development of adipose
tissue. CHD7 loss results in a reduced subcutaneous fat phenotype in heterozygous knockout
mice
237,238
, while PHF6 causes an obesity phenotype
182
. This may represent a similar cell fate decision
mechanism at the point of adipose differentiation. Moreover, this may not be the only example of
a chromatin adapter protein that prevents precocious activation of enhancer, repressor, or
transcriptional start site elements. Given that there are hundreds of cell fate decisions over the
course of human development, it is highly likely that a similar mechanism occurs in other cell
types. It is therefore worthwhile to fully comprehend how this particular mechanism arises and the
complexity within the system.
This mechanism is also important because understanding the complexity and identifying
similar mechanisms within other cell fate choices allows for possibilities to develop therapeutics
for developmental disorders like CHARGE and BFLS. While it is unlikely that therapeutics can
be effectively delivered to affect the cell fate decision of neural crest migration, as this occurs
within the first days of conception, other later cell fate choices may be influenced by novel
157
therapeutics that can significantly alleviate the suffering of these patients and improve their
livelihoods.
Furthermore, understanding of basic mechanisms of activity of these proteins has
implications for cancer research. PHF6 has been identified as one of the most commonly mutated
proteins across all cancers
223
, and is particularly highly affected in T-ALL and other leukemias.
Despite the fact that the mechanism of oncogenesis due to PHF6 mutation is still unknown, this
protein has been identified as a tumor suppressor. The mechanism outlined in this work represents
a potential mechanism by which PHF6 is able to prevent invasive, proliferative cell phenotypes,
as neural crest cells are highly migratory and proliferative. Further research into activity of PHF6
in the cell types that become transformed would determine if a similar mechanism occurs in these
cases.

158
Appendix I. References
1. Luger, K., Rechsteiner, T. J., Flaus, A. J., Waye, M. M. & Richmond, T. J. Characterization of
nucleosome core particles containing histone proteins made in bacteria. J. Mol. Biol. 272,
301–311 (1997).
2. Yuan, G.-C. et al. Genome-scale identification of nucleosome positions in S. cerevisiae.
Science 309, 626–630 (2005).
3. Bernstein, B. E., Liu, C. L., Humphrey, E. L., Perlstein, E. O. & Schreiber, S. L. Global
nucleosome occupancy in yeast. Genome Biol. 5, R62 (2004).
4. Lee, C.-K., Shibata, Y., Rao, B., Strahl, B. D. & Lieb, J. D. Evidence for nucleosome depletion
at active regulatory regions genome-wide. Nat. Genet. 36, 900–905 (2004).
5. Sekinger, E. A., Moqtaderi, Z. & Struhl, K. Intrinsic histone-DNA interactions and low
nucleosome density are important for preferential accessibility of promoter regions in
yeast. Mol. Cell 18, 735–748 (2005).
6. Snyder, M. J. et al. Anchoring of Heterochromatin to the Nuclear Lamina Reinforces
Dosage Compensation-Mediated Gene Repression. PLoS Genet. 12, e1006341 (2016).
7. Towbin, B. D., Gonzalez-Sandoval, A. & Gasser, S. M. Mechanisms of heterochromatin
subnuclear localization. Trends Biochem. Sci. 38, 356–363 (2013).
8. Giorgetti, L. et al. Predictive polymer modeling reveals coupled fluctuations in
chromosome conformation and transcription. Cell 157, 950–963 (2014).
9. Jacobs, S. A. et al. The active site of the SET domain is constructed on a knot. Nat. Struct.
Biol. 9, 833–838 (2002).
159
10. Ernst, J. & Kellis, M. Discovery and characterization of chromatin states for systematic
annotation of the human genome. Nat. Biotechnol. 28, 817–825 (2010).
11. Tee, W.-W. & Reinberg, D. Chromatin features and the epigenetic regulation of
pluripotency states in ESCs. Dev. Camb. Engl. 141, 2376–2390 (2014).
12. Guenther, M. G., Levine, S. S., Boyer, L. A., Jaenisch, R. & Young, R. A. A chromatin
landmark and transcription initiation at most promoters in human cells. Cell 130, 77–88
(2007).
13. Barski, A. et al. High-resolution profiling of histone methylations in the human genome.
Cell 129, 823–837 (2007).
14. Creyghton, M. P. et al. Histone H3K27ac separates active from poised enhancers and
predicts developmental state. Proc. Natl. Acad. Sci. U. S. A. 107, 21931–21936 (2010).
15. Maston, G. A., Landt, S. G., Snyder, M. & Green, M. R. Characterization of enhancer
function from genome-wide analyses. Annu. Rev. Genomics Hum. Genet. 13, 29–57 (2012).
16. Wysocka, J. et al. A PHD finger of NURF couples histone H3 lysine 4 trimethylation with
chromatin remodelling. Nature 442, 86–90 (2006).
17. Wysocka, J. et al. WDR5 associates with histone H3 methylated at K4 and is essential for
H3 K4 methylation and vertebrate development. Cell 121, 859–872 (2005).
18. Shi, X. et al. ING2 PHD domain links histone H3 lysine 4 methylation to active gene
repression. Nature 442, 96–99 (2006).
19. Pray-Grant, M. G., Daniel, J. A., Schieltz, D., Yates, J. R. & Grant, P. A. Chd1 chromodomain
links histone H3 methylation with SAGA- and SLIK-dependent acetylation. Nature 433,
434–438 (2005).
160
20. Sims, R. J. et al. Human but not yeast CHD1 binds directly and selectively to histone H3
methylated at lysine 4 via its tandem chromodomains. J. Biol. Chem. 280, 41789–41792
(2005).
21. Hall, B. D. Choanal atresia and associated multiple anomalies. J. Pediatr. 95, 395–398
(1979).
22. Hittner, H. M., Hirsch, N. J., Kreh, G. M. & Rudolph, A. J. Colobomatous microphthalmia,
heart disease, hearing loss, and mental retardation--a syndrome. J. Pediatr. Ophthalmol.
Strabismus 16, 122–128 (1979).
23. Pagon, R. A., Graham, J. M., Zonana, J. & Yong, S. L. Coloboma, congenital heart disease,
and choanal atresia with multiple anomalies: CHARGE association. J. Pediatr. 99, 223–227
(1981).
24. Davenport, S. L., Hefner, M. A. & Mitchell, J. A. The spectrum of clinical features in
CHARGE syndrome. Clin. Genet. 29, 298–310 (1986).
25. Oley, C. A., Baraitser, M. & Grant, D. B. A reappraisal of the CHARGE association. J. Med.
Genet. 25, 147–156 (1988).
26. Tellier, A. L. et al. CHARGE syndrome: report of 47 cases and review. Am. J. Med. Genet.
76, 402–409 (1998).
27. Schnetz, M. P. et al. Genomic distribution of CHD7 on chromatin tracks H3K4 methylation
patterns. Genome Res. 19, 590–601 (2009).
28. Schnetz, M. P. et al. CHD7 targets active gene enhancer elements to modulate ES cell-
specific gene expression. PLoS Genet. 6, e1001023 (2010).
161
29. Bajpai, R. et al. CHD7 cooperates with PBAF to control multipotent neural crest formation.
Nature 463, 958–962 (2010).
30. Sperry, E. D. et al. The chromatin remodeling protein CHD7, mutated in CHARGE
syndrome, is necessary for proper craniofacial and tracheal development. Dev. Dyn. Off.
Publ. Am. Assoc. Anat. 243, 1055–1066 (2014).
31. Wieczorek, D. et al. A comprehensive molecular study on Coffin-Siris and Nicolaides-
Baraitser syndromes identifies a broad molecular and clinical spectrum converging on
altered chromatin remodeling. Hum. Mol. Genet. 22, 5121–5135 (2013).
32. Schrier Vergano, S. et al. Coffin-Siris Syndrome. in GeneReviews(®) (eds. Pagon, R. A. et al.)
(University of Washington, Seattle, 1993).
33. Vergano, S. S. & Deardorff, M. A. Clinical features, diagnostic criteria, and management of
Coffin-Siris syndrome. Am. J. Med. Genet. C Semin. Med. Genet. 166C, 252–256 (2014).
34. Hurlstone, A. F. L., Olave, I. A., Barker, N., van Noort, M. & Clevers, H. Cloning and
characterization of hELD/OSA1, a novel BRG1 interacting protein. Biochem. J. 364, 255–
264 (2002).
35. Lessard, J. et al. An essential switch in subunit composition of a chromatin remodeling
complex during neural development. Neuron 55, 201–215 (2007).
36. Wu, J. I. et al. Regulation of dendritic development by neuron-specific chromatin
remodeling complexes. Neuron 56, 94–108 (2007).
37. Wu, J. I., Lessard, J. & Crabtree, G. R. Understanding the words of chromatin regulation.
Cell 136, 200–206 (2009).
162
38. Hargreaves, D. C. & Crabtree, G. R. ATP-dependent chromatin remodeling: genetics,
genomics and mechanisms. Cell Res. 21, 396–420 (2011).
39. Abdul-Rahman, O. Nicolaides-Baraitser Syndrome. in GeneReviews(®) (eds. Pagon, R. A. et
al.) (University of Washington, Seattle, 1993).
40. Muchardt, C. & Yaniv, M. A human homologue of Saccharomyces cerevisiae SNF2/SWI2
and Drosophila brm genes potentiates transcriptional activation by the glucocorticoid
receptor. EMBO J. 12, 4279–4290 (1993).
41. Rohatgi, S. et al. Facial diagnosis of mild and variant CdLS: Insights from a dysmorphologist
survey. Am. J. Med. Genet. A. 152A, 1641–1653 (2010).
42. Kumar, D., Blank, C. E. & Griffiths, B. L. Cornelia de Lange syndrome in several members of
the same family. J. Med. Genet. 22, 296–300 (1985).
43. Hoppman-Chaney, N., Jang, J. S., Jen, J., Babovic-Vuksanovic, D. & Hodge, J. C. In-frame
multi-exon deletion of SMC1A in a severely affected female with Cornelia de Lange
Syndrome. Am. J. Med. Genet. A. 158A, 193–198 (2012).
44. Sumara, I., Vorlaufer, E., Gieffers, C., Peters, B. H. & Peters, J. M. Characterization of
vertebrate cohesin complexes and their regulation in prophase. J. Cell Biol. 151, 749–762
(2000).
45. Watrin, E. et al. Human Scc4 is required for cohesin binding to chromatin, sister-chromatid
cohesion, and mitotic progression. Curr. Biol. CB 16, 863–874 (2006).
46. Van den Wyngaert, I. et al. Cloning and characterization of human histone deacetylase 8.
FEBS Lett. 478, 77–83 (2000).
163
47. Rubinstein, J. H. & Taybi, H. Broad thumbs and toes and facial abnormalities. A possible
mental retardation syndrome. Am. J. Dis. Child. 1960 105, 588–608 (1963).
48. Hennekam, R. C. M. Rubinstein-Taybi syndrome. Eur. J. Hum. Genet. EJHG 14, 981–985
(2006).
49. Jensen, L. R. et al. Mutations in the JARID1C gene, which is involved in transcriptional
regulation and chromatin remodeling, cause X-linked mental retardation. Am. J. Hum.
Genet. 76, 227–236 (2005).
50. Tahiliani, M. et al. The histone H3K4 demethylase SMCX links REST target genes to X-linked
mental retardation. Nature 447, 601–605 (2007).
51. Sifrim, A. et al. Distinct genetic architectures for syndromic and nonsyndromic congenital
heart defects identified by exome sequencing. Nat. Genet. 48, 1060–1065 (2016).
52. Weiss, K. et al. De Novo Mutations in CHD4, an ATP-Dependent Chromatin Remodeler
Gene, Cause an Intellectual Disability Syndrome with Distinctive Dysmorphisms. Am. J.
Hum. Genet. 99, 934–941 (2016).
53. O’Roak, B. J. et al. Multiplex targeted sequencing identifies recurrently mutated genes in
autism spectrum disorders. Science 338, 1619–1622 (2012).
54. Jiang, Y. et al. Detection of clinically relevant genetic variants in autism spectrum disorder
by whole-genome sequencing. Am. J. Hum. Genet. 93, 249–263 (2013).
55. Neale, B. M. et al. Patterns and rates of exonic de novo mutations in autism spectrum
disorders. Nature 485, 242–245 (2012).
56. Ronan, J. L., Wu, W. & Crabtree, G. R. From neural development to cognition: unexpected
roles for chromatin. Nat. Rev. Genet. 14, 347–359 (2013).
164
57. Le Douarin, N. & Kalcheim, C. The neural crest. (Cambridge University Press, 1999).
58. Streit, A., Berliner, A. J., Papanayotou, C., Sirulnik, A. & Stern, C. D. Initiation of neural
induction by FGF signalling before gastrulation. Nature 406, 74–78 (2000).
59. Groves, A. K. & LaBonne, C. Setting appropriate boundaries: fate, patterning and
competence at the neural plate border. Dev. Biol. 389, 2–12 (2014).
60. Endo, Y., Osumi, N. & Wakamatsu, Y. Bimodal functions of Notch-mediated signaling are
involved in neural crest formation during avian ectoderm development. Dev. Camb. Engl.
129, 863–873 (2002).
61. Monsoro-Burq, A.-H., Fletcher, R. B. & Harland, R. M. Neural crest induction by paraxial
mesoderm in Xenopus embryos requires FGF signals. Dev. Camb. Engl. 130, 3111–3124
(2003).
62. Yardley, N. & García-Castro, M. I. FGF signaling transforms non-neural ectoderm into
neural crest. Dev. Biol. 372, 166–177 (2012).
63. Meulemans, D. & Bronner-Fraser, M. Gene-regulatory interactions in neural crest
evolution and development. Dev. Cell 7, 291–299 (2004).
64. Nichane, M., Ren, X., Souopgui, J. & Bellefroid, E. J. Hairy2 functions through both DNA-
binding and non DNA-binding mechanisms at the neural plate border in Xenopus. Dev.
Biol. 322, 368–380 (2008).
65. Khudyakov, J. & Bronner-Fraser, M. Comprehensive spatiotemporal analysis of early chick
neural crest network genes. Dev. Dyn. Off. Publ. Am. Assoc. Anat. 238, 716–723 (2009).
66. Bhat, N., Kwon, H.-J. & Riley, B. B. A gene network that coordinates preplacodal
competence and neural crest specification in zebrafish. Dev. Biol. 373, 107–117 (2013).
165
67. McLarren, K. W., Litsiou, A. & Streit, A. DLX5 positions the neural crest and preplacode
region at the border of the neural plate. Dev. Biol. 259, 34–47 (2003).
68. Matsuo-Takasaki, M., Matsumura, M. & Sasai, Y. An essential role of Xenopus Foxi1a for
ventral specification of the cephalic ectoderm during gastrulation. Dev. Camb. Engl. 132,
3885–3894 (2005).
69. Kwon, H.-J., Bhat, N., Sweet, E. M., Cornell, R. A. & Riley, B. B. Identification of early
requirements for preplacodal ectoderm and sensory organ development. PLoS Genet. 6,
e1001133 (2010).
70. Simões-Costa, M. & Bronner, M. E. Establishing neural crest identity: a gene regulatory
recipe. Dev. Camb. Engl. 142, 242–257 (2015).
71. Aybar, M. J., Nieto, M. A. & Mayor, R. Snail precedes slug in the genetic cascade required
for the specification and migration of the Xenopus neural crest. Dev. Camb. Engl. 130,
483–494 (2003).
72. Milet, C., Maczkowiak, F., Roche, D. D. & Monsoro-Burq, A. H. Pax3 and Zic1 drive
induction and differentiation of multipotent, migratory, and functional neural crest in
Xenopus embryos. Proc. Natl. Acad. Sci. U. S. A. 110, 5528–5533 (2013).
73. de Crozé, N., Maczkowiak, F. & Monsoro-Burq, A. H. Reiterative AP2a activity controls
sequential steps in the neural crest gene regulatory network. Proc. Natl. Acad. Sci. U. S. A.
108, 155–160 (2011).
74. Nikitina, N., Sauka-Spengler, T. & Bronner-Fraser, M. Dissecting early regulatory
relationships in the lamprey neural crest gene network. Proc. Natl. Acad. Sci. U. S. A. 105,
20083–20088 (2008).
166
75. Simões-Costa, M. S., McKeown, S. J., Tan-Cabugao, J., Sauka-Spengler, T. & Bronner, M. E.
Dynamic and differential regulation of stem cell factor FoxD3 in the neural crest is
Encrypted in the genome. PLoS Genet. 8, e1003142 (2012).
76. Honoré, S. M., Aybar, M. J. & Mayor, R. Sox10 is required for the early development of the
prospective neural crest in Xenopus embryos. Dev. Biol. 260, 79–96 (2003).
77. Vallin, J., Girault, J. M., Thiery, J. P. & Broders, F. Xenopus cadherin-11 is expressed in
different populations of migrating neural crest cells. Mech. Dev. 75, 171–174 (1998).
78. Kashef, J. et al. Cadherin-11 regulates protrusive activity in Xenopus cranial neural crest
cells upstream of Trio and the small GTPases. Genes Dev. 23, 1393–1398 (2009).
79. Cheung, M. et al. The transcriptional control of trunk neural crest induction, survival, and
delamination. Dev. Cell 8, 179–192 (2005).
80. Dottori, M., Gross, M. K., Labosky, P. & Goulding, M. The winged-helix transcription factor
Foxd3 suppresses interneuron differentiation and promotes neural crest cell fate. Dev.
Camb. Engl. 128, 4127–4138 (2001).
81. Fairchild, C. L. & Gammill, L. S. Tetraspanin18 is a FoxD3-responsive antagonist of cranial
neural crest epithelial-to-mesenchymal transition that maintains cadherin-6B protein. J.
Cell Sci. 126, 1464–1476 (2013).
82. Theveneau, E. et al. Collective chemotaxis requires contact-dependent cell polarity. Dev.
Cell 19, 39–53 (2010).
83. Theveneau, E. et al. Chase-and-run between adjacent cell populations promotes
directional collective migration. Nat. Cell Biol. 15, 763–772 (2013).
167
84. Barriga, E. H., Maxwell, P. H., Reyes, A. E. & Mayor, R. The hypoxia factor Hif-1α controls
neural crest chemotaxis and epithelial to mesenchymal transition. J. Cell Biol. 201, 759–
776 (2013).
85. Takahashi, Y., Sipp, D. & Enomoto, H. Tissue interactions in neural crest cell development
and disease. Science 341, 860–863 (2013).
86. Minoux, M. et al. Gene bivalency at Polycomb domains regulates cranial neural crest
positional identity. Science 355, (2017).
87. Attanasio, C. et al. Fine tuning of craniofacial morphology by distant-acting enhancers.
Science 342, 1241006 (2013).
88. Han, J. et al. Concerted action of Msx1 and Msx2 in regulating cranial neural crest cell
differentiation during frontal bone development. Mech. Dev. 124, 729–745 (2007).
89. Sasaki, T. et al. TGFbeta-mediated FGF signaling is crucial for regulating cranial neural crest
cell proliferation during frontal bone development. Dev. Camb. Engl. 133, 371–381 (2006).
90. Tabler, J. M., Rice, C. P., Liu, K. J. & Wallingford, J. B. A novel ciliopathic skull defect arising
from excess neural crest. Dev. Biol. 417, 4–10 (2016).
91. Li, L. et al. Augmented BMPRIA-mediated BMP signaling in cranial neural crest lineage
leads to cleft palate formation and delayed tooth differentiation. PloS One 8, e66107
(2013).
92. Rada-Iglesias, A. et al. Epigenomic annotation of enhancers predicts transcriptional
regulators of human neural crest. Cell Stem Cell 11, 633–648 (2012).
93. The Strategy of the Genes. (Routledge, 2014).
168
94. Borjeson, M., Forssman, H. & Lehmann, O. An X-linked, recessively inherited syndrome
characterized by grave mental deficiency, epilepsy, and endocrine disorder. Acta Med.
Scand. 171, 13–21 (1962).
95. Veall, R. M., Brett, E. M., Rivinus, T. M. & Stephens, R. The Börjeson-Forssman-Lehmann
syndrome: a new case. J. Ment. Defic. Res. 23, 231–242 (1979).
96. Weber, F. T., Frias, J. L., Julius, R. L. & Felman, A. H. Primary hypogonadism in the
Borjeson-Forssman-Lehmann syndrome. J. Med. Genet. 15, 63–66 (1978).
97. Robinson, L. K. et al. The Börjeson-Forssman-Lehmann syndrome. Am. J. Med. Genet. 15,
457–468 (1983).
98. Ardinger, H. H., Hanson, J. W. & Zellweger, H. U. Börjeson-Forssman-Lehmann syndrome:
further delineation in five cases. Am. J. Med. Genet. 19, 653–664 (1984).
99. Dereymaeker, A. M. et al. The Borjeson-Forssman-Lehmann syndrome. A family study.
Clin. Genet. 29, 317–320 (1986).
100. Turner, G. et al. Börjeson-Forssman-Lehmann syndrome: clinical manifestations and gene
localization to Xq26-27. Am. J. Med. Genet. 34, 463–469 (1989).
101. Mathews, K. D., Buetow, K., Turner, G. & Mulley, J. Börjeson-Forssman-Lehmann
syndrome localization. Am. J. Med. Genet. 34, 475 (1989).
102. Mathews, K. D. et al. Linkage localization of Börjeson-Forssman-Lehmann syndrome. Am.
J. Med. Genet. 34, 470–474 (1989).
103. Chudley, A. E. Gene localization in Börjeson-Forssman-Lehmann syndrome and other X-
linked mental retardation syndromes. Am. J. Med. Genet. 38, 248 (1991).
169
104. Gedeon, A. K. et al. Refinement of the background genetic map of Xq26-q27 and gene
localisation for Börjeson-Forssman-Lehmann Syndrome. Am. J. Med. Genet. 64, 63–68
(1996).
105. Lower, K. M. et al. Mutations in PHF6 are associated with Börjeson-Forssman-Lehmann
syndrome. Nat. Genet. 32, 661–665 (2002).
106. Turner, G. et al. The clinical picture of the Börjeson-Forssman-Lehmann syndrome in males
and heterozygous females with PHF6 mutations. Clin. Genet. 65, 226–232 (2004).
107. Gécz, J., Turner, G., Nelson, J. & Partington, M. The Börjeson-Forssman-Lehman syndrome
(BFLS, MIM #301900). Eur. J. Hum. Genet. EJHG 14, 1233–1237 (2006).
108. Crawford, J. et al. Mutation screening in Borjeson-Forssman-Lehmann syndrome:
identification of a novel de novo PHF6 mutation in a female patient. J. Med. Genet. 43,
238–243 (2006).
109. Zweier, C. et al. A new face of Borjeson-Forssman-Lehmann syndrome? De novo mutations
in PHF6 in seven females with a distinct phenotype. J. Med. Genet. 50, 838–847 (2013).
110. Liu, Z. et al. Structural and functional insights into the human Börjeson-Forssman-Lehmann
syndrome-associated protein PHF6. J. Biol. Chem. 289, 10069–10083 (2014).
111. Bao, Y., Liu, Z., Zhang, J., Wu, J. & Shi, Y. 1H, 13C and 15N resonance assignments and
secondary structure of the human PHF6-ePHD1 domain. Biomol. NMR Assign. 10, 1–4
(2016).
112. Todd, M. A. M., Ivanochko, D. & Picketts, D. J. PHF6 Degrees of Separation: The
Multifaceted Roles of a Chromatin Adaptor Protein. Genes 6, 325–352 (2015).
170
113. Todd, M. A. M. & Picketts, D. J. PHF6 interacts with the nucleosome remodeling and
deacetylation (NuRD) complex. J. Proteome Res. 11, 4326–4337 (2012).
114. Wang, J. et al. PHF6 regulates cell cycle progression by suppressing ribosomal RNA
synthesis. J. Biol. Chem. 288, 3174–3183 (2013).
115. Voss, A. K. et al. Protein and gene expression analysis of Phf6, the gene mutated in the
Börjeson-Forssman-Lehmann Syndrome of intellectual disability and obesity. Gene Expr.
Patterns GEP 7, 858–871 (2007).
116. Zhang, C. et al. The X-linked intellectual disability protein PHF6 associates with the PAF1
complex and regulates neuronal migration in the mammalian brain. Neuron 78, 986–993
(2013).
117. Keller, R. E. The cellular basis of epiboly: an SEM study of deep-cell rearrangement during
gastrulation in Xenopus laevis. J. Embryol. Exp. Morphol. 60, 201–234 (1980).
118. Keller, R. & Danilchik, M. Regional expression, pattern and timing of convergence and
extension during gastrulation of Xenopus laevis. Dev. Camb. Engl. 103, 193–209 (1988).
119. Keller, R. E., Danilchik, M., Gimlich, R. & Shih, J. The function and mechanism of
convergent extension during gastrulation of Xenopus laevis. J. Embryol. Exp. Morphol. 89
Suppl, 185–209 (1985).
120. Wilson, P. A., Oster, G. & Keller, R. Cell rearrangement and segmentation in Xenopus:
direct observation of cultured explants. Dev. Camb. Engl. 105, 155–166 (1989).
121. Poznanski, A., Minsuk, S., Stathopoulos, D. & Keller, R. Epithelial cell wedging and neural
trough formation are induced planarly in Xenopus, without persistent vertical interactions
with mesoderm. Dev. Biol. 189, 256–269 (1997).
171
122. Berlowitz, L. & Birnstiel, M. L. Histones in the wild-type and the anucleolate mutant of
Xenopus laevis. Science 156, 78–80 (1967).
123. Nguyen-Kim-Do, null & Csaba, G. Histochemical examination of histones in early phases of
the development of amphibia. Z. Mikrosk. Anat. Forsch. 87, 57–64 (1973).
124. Moody, S. A. Fates of the blastomeres of the 16-cell stage Xenopus embryo. Dev. Biol. 119,
560–578 (1987).
125. Moody, S. A. Fates of the blastomeres of the 32-cell-stage Xenopus embryo. Dev. Biol. 122,
300–319 (1987).
126. Bauer, D. V., Huang, S. & Moody, S. A. The cleavage stage origin of Spemann’s Organizer:
analysis of the movements of blastomere clones before and during gastrulation in
Xenopus. Dev. Camb. Engl. 120, 1179–1189 (1994).
127. Sive, H. L., Grainger, R. M. & Harland, R. M. Early development of Xenopus laevis: a
laboratory manual. (Cold Spring Harbor Laboratory Press, 2000).
128. Balls, M. & Ruben, L. N. VARIATION IN THE RESPONSE OF XENOPUS LAEVIS TO NORMAL
TISSUE HOMOGRAFTS. Dev. Biol. 10, 92–104 (1964).
129. Hunt, P. M. Effects of rotating neural tissue and underlying mesoderm in Xenopus laevis
embryos. Acta Embryol. Exp. (Palermo) 2, 211–229 (1969).
130. Hunt, R. K. & Jacobson, M. Specification of positional information in retinal ganglion cells
of Xenopus: stability of the specified state. Proc. Natl. Acad. Sci. U. S. A. 69, 2860–2864
(1972).
131. Tarin, D. Histological features of neural induction in Xenopus laevis. J. Embryol. Exp.
Morphol. 26, 543–570 (1971).
172
132. Session, A. M. et al. Genome evolution in the allotetraploid frog Xenopus laevis. Nature
538, 336–343 (2016).
133. Pegoraro, C. & Monsoro-Burq, A. H. Signaling and transcriptional regulation in neural crest
specification and migration: lessons from xenopus embryos. Wiley Interdiscip. Rev. Dev.
Biol. 2, 247–259 (2013).
134. Dubey, A. & Saint-Jeannet, J.-P. Modeling human craniofacial disorders in Xenopus. Curr.
Pathobiol. Rep. 5, 79–92 (2017).
135. Hatch, V. L. et al. The positive transcriptional elongation factor (P-TEFb) is required for
neural crest specification. Dev. Biol. 416, 361–372 (2016).
136. Thélie, A. et al. Prdm12 specifies V1 interneurons through cross-repressive interactions
with Dbx1 and Nkx6 genes in Xenopus. Dev. Camb. Engl. 142, 3416–3428 (2015).
137. Shechter, D., Dormann, H. L., Allis, C. D. & Hake, S. B. Extraction, purification and analysis
of histones. Nat. Protoc. 2, 1445–1457 (2007).
138. Xia, X., Zhang, Y., Zieth, C. R. & Zhang, S.-C. Transgenes delivered by lentiviral vector are
suppressed in human embryonic stem cells in a promoter-dependent manner. Stem Cells
Dev. 16, 167–176 (2007).
139. Hoyert, D. L., Mathews, T. J., Menacker, F., Strobino, D. M. & Guyer, B. Annual summary of
vital statistics: 2004. Pediatrics 117, 168–183 (2006).
140. Yoon, P. W. et al. Contribution of birth defects and genetic diseases to pediatric
hospitalizations. A population-based study. Arch. Pediatr. Adolesc. Med. 151, 1096–1103
(1997).
173
141. Centers for Disease Control and Prevention (CDC). Hospital stays, hospital charges, and in-
hospital deaths among infants with selected birth defects--United States, 2003. MMWR
Morb. Mortal. Wkly. Rep. 56, 25–29 (2007).
142. Lower, K. M. et al. Mutations in PHF6 are associated with Börjeson-Forssman-Lehmann
syndrome. Nat. Genet. 32, 661–665 (2002).
143. Miyake, N., Tsurusaki, Y. & Matsumoto, N. Numerous BAF complex genes are mutated in
Coffin-Siris syndrome. Am. J. Med. Genet. C Semin. Med. Genet. 166C, 257–261 (2014).
144. Zweier, C. et al. Females with de novo aberrations in PHF6: clinical overlap of Borjeson-
Forssman-Lehmann with Coffin-Siris syndrome. Am. J. Med. Genet. C Semin. Med. Genet.
166C, 290–301 (2014).
145. Sive, H. L., Grainger, R. & Harland, R. M. Early development of Xenopus laevis: a laboratory
manual. (Cold Spring Harbor Laboratory Press., 2010).
146. Kearse, M. et al. Geneious Basic: an integrated and extendable desktop software platform
for the organization and analysis of sequence data. Bioinforma. Oxf. Engl. 28, 1647–1649
(2012).
147. Jukes, T. & Cantor, C. Evolution of Protein Molecules. in Mammalian Protein Metabolism
(ed. Munro, H.) 21–132 (Academic Press, 1969).
148. Fourment, M. & Gibbs, M. J. PATRISTIC: a program for calculating patristic distances and
graphically comparing the components of genetic change. BMC Evol. Biol. 6, 1 (2006).
149. Sarkar, A. & Kellogg, G. E. Hydrophobicity--shake flasks, protein folding and drug discovery.
Curr. Top. Med. Chem. 10, 67–83 (2010).
174
150. Vlassi, M., Cesareni, G. & Kokkinidis, M. A correlation between the loss of hydrophobic
core packing interactions and protein stability. J. Mol. Biol. 285, 817–827 (1999).
151. Hatch, V. L. et al. The positive transcriptional elongation factor (P-TEFb) is required for
neural crest specification. Dev. Biol. 416, 361–372 (2016).
152. Thélie, A. et al. Prdm12 specifies V1 interneurons through cross-repressive interactions
with Dbx1 and Nkx6 genes in Xenopus. Dev. Camb. Engl. 142, 3416–3428 (2015).
153. Green, J. The animal cap assay. Methods Mol. Biol. Clifton NJ 127, 1–13 (1999).
154. Blower, M. D., Jambhekar, A., Schwarz, D. S. & Toombs, J. A. Combining different mRNA
capture methods to analyze the transcriptome: analysis of the Xenopus laevis
transcriptome. PloS One 8, e77700 (2013).
155. Popov, I. K. et al. Identification of new regulators of embryonic patterning and
morphogenesis in Xenopus gastrulae by RNA sequencing. Dev. Biol. (2016).
doi:10.1016/j.ydbio.2016.05.014
156. Grabherr, M. G. et al. Full-length transcriptome assembly from RNA-Seq data without a
reference genome. Nat. Biotechnol. 29, 644–652 (2011).
157. Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J. Basic local alignment
search tool. J. Mol. Biol. 215, 403–410 (1990).
158. Trapnell, C. et al. Transcript assembly and quantification by RNA-Seq reveals unannotated
transcripts and isoform switching during cell differentiation. Nat. Biotechnol. 28, 511–515
(2010).
159. Trapnell, C. et al. Differential analysis of gene regulation at transcript resolution with RNA-
seq. Nat. Biotechnol. 31, 46–53 (2013).
175
160. Anders, S., Pyl, P. T. & Huber, W. HTSeq--a Python framework to work with high-
throughput sequencing data. Bioinforma. Oxf. Engl. 31, 166–169 (2015).
161. Robinson, M. D., McCarthy, D. J. & Smyth, G. K. edgeR: a Bioconductor package for
differential expression analysis of digital gene expression data. Bioinforma. Oxf. Engl. 26,
139–140 (2010).
162. Ali, M., Hom, R. A., Blakeslee, W., Ikenouye, L. & Kutateladze, T. G. Diverse functions of
PHD fingers of the MLL/KMT2 subfamily. Biochim. Biophys. Acta 1843, 366–371 (2014).
163. DiNitto, J. P., Cronin, T. C. & Lambright, D. G. Membrane recognition and targeting by lipid-
binding domains. Sci. STKE Signal Transduct. Knowl. Environ. 2003, re16 (2003).
164. Matthews, J. M. et al. It takes two to tango: the structure and function of LIM, RING, PHD
and MYND domains. Curr. Pharm. Des. 15, 3681–3696 (2009).
165. Sanchez, R. & Zhou, M.-M. The PHD finger: a versatile epigenome reader. Trends Biochem.
Sci. 36, 364–372 (2011).
166. Adamson, E. D. & Woodland, H. R. Histone synthesis in early amphibian development:
histone and DNA syntheses are not co-ordinated. J. Mol. Biol. 88, 263–285 (1974).
167. Byrd, E. W. & Kasinsky, H. E. Histone synthesis during early embryogenesis in Xenopus
laevis (South African clawed toad). Biochemistry (Mosc.) 12, 246–253 (1973).
168. Claycomb, W. C. & Villee, C. A. DNA-binding proteins of Xenopus laevis: synthesis during
oogenesis-ovulation and early embryogenesis. Exp. Cell Res. 83, 191–199 (1974).
169. Berland, S., Alme, K., Brendehaug, A., Houge, G. & Hovland, R. PHF6 Deletions May Cause
Borjeson-Forssman-Lehmann Syndrome in Females. Mol. Syndromol. 1, 294–300 (2011).
176
170. Chao, M. M. et al. T-cell acute lymphoblastic leukemia in association with Börjeson-
Forssman-Lehmann syndrome due to a mutation in PHF6. Pediatr. Blood Cancer 55, 722–
724 (2010).
171. Di Donato, N. et al. Distinct phenotype of PHF6 deletions in females. Eur. J. Med. Genet.
57, 85–89 (2014).
172. Ernst, A. et al. The PHF6 Mutation c.1A>G; pM1V Causes Börjeson-Forsman-Lehmann
Syndrome in a Family with Four Affected Young Boys. Mol. Syndromol. 6, 181–186 (2015).
173. Jahani-Asl, A., Cheng, C., Zhang, C. & Bonni, A. Pathogenesis of Börjeson-Forssman-
Lehmann syndrome: Insights from PHF6 function. Neurobiol. Dis. 96, 227–235 (2016).
174. Visootsak, J. et al. Clinical and behavioral features of patients with Borjeson-Forssman-
Lehmann syndrome with mutations in PHF6. J. Pediatr. 145, 819–825 (2004).
175. Liu, Z. et al. Structural and functional insights into the human Börjeson-Forssman-Lehmann
syndrome-associated protein PHF6. J. Biol. Chem. 289, 10069–10083 (2014).
176. Hopwood, N. D., Pluck, A. & Gurdon, J. B. A Xenopus mRNA related to Drosophila twist is
expressed in response to induction in the mesoderm and the neural crest. Cell 59, 893–
903 (1989).
177. Mizuseki, K., Kishi, M., Matsui, M., Nakanishi, S. & Sasai, Y. Xenopus Zic-related-1 and Sox-
2, two factors induced by chordin, have distinct activities in the initiation of neural
induction. Dev. Camb. Engl. 125, 579–587 (1998).
178. James-Zorn, C. et al. Xenbase: Core features, data acquisition, and data processing. Genes.
N. Y. N 2000 53, 486–497 (2015).
177
179. Karpinka, J. B. et al. Xenbase, the Xenopus model organism database; new virtualized
system, data types and genomes. Nucleic Acids Res. 43, D756-763 (2015).
180. Kim, D. et al. TopHat2: accurate alignment of transcriptomes in the presence of insertions,
deletions and gene fusions. Genome Biol. 14, R36 (2013).
181. Haas, B. J. et al. De novo transcript sequence reconstruction from RNA-seq using the
Trinity platform for reference generation and analysis. Nat. Protoc. 8, 1494–1512 (2013).
182. Borjeson, M., Forssman, H. & Lehmann, O. An X-linked, recessively inherited syndrome
characterized by grave mental deficiency, epilepsy, and endocrine disorder. Acta Med.
Scand. 171, 13–21 (1962).
183. Liu, Z. et al. Structural and functional insights into the human Börjeson-Forssman-Lehmann
syndrome-associated protein PHF6. J. Biol. Chem. 289, 10069–10083 (2014).
184. Nakatsuji, N. Craniofacial malformation in Xenopus laevis tadpoles caused by the exposure
of early embryos to ethanol. Teratology 28, 299–305 (1983).
185. Ataliotis, P., Ivins, S., Mohun, T. J. & Scambler, P. J. XTbx1 is a transcriptional activator
involved in head and pharyngeal arch development in Xenopus laevis. Dev. Dyn. Off. Publ.
Am. Assoc. Anat. 232, 979–991 (2005).
186. Cordelieres, F. Manual Tracking. (2004).
187. Todd, M. A. M. & Picketts, D. J. PHF6 interacts with the nucleosome remodeling and
deacetylation (NuRD) complex. J. Proteome Res. 11, 4326–4337 (2012).
188. Wang, J. et al. PHF6 regulates cell cycle progression by suppressing ribosomal RNA
synthesis. J. Biol. Chem. 288, 3174–3183 (2013).
178
189. Meacham, C. E. et al. A genome-scale in vivo loss-of-function screen identifies Phf6 as a
lineage-specific regulator of leukemia cell growth. Genes Dev. 29, 483–488 (2015).
190. Bao, Y., Liu, Z., Zhang, J., Wu, J. & Shi, Y. 1H, 13C and 15N resonance assignments and
secondary structure of the human PHF6-ePHD1 domain. Biomol. NMR Assign. 10, 1–4
(2016).
191. Sanchez, R. & Zhou, M.-M. The PHD finger: a versatile epigenome reader. Trends Biochem.
Sci. 36, 364–372 (2011).
192. Ashiya, M., Zhang, L. & Grabowski, P. J. Regulated splicing of gamma2 pre-messenger RNA
in neuronal cells. Nucleic Acids Symp. Ser. 215–216 (1995).
193. Murray, B. A. et al. Cell surface modulation of the neural cell adhesion molecule resulting
from alternative mRNA splicing in a tissue-specific developmental sequence. J. Cell Biol.
103, 1431–1439 (1986).
194. Stone, D. M. & Nikolics, K. Tissue- and age-specific expression patterns of alternatively
spliced agrin mRNA transcripts in embryonic rat suggest novel developmental roles. J.
Neurosci. Off. J. Soc. Neurosci. 15, 6767–6778 (1995).
195. Stamm, S., Casper, D., Lees-Miller, J. P. & Helfman, D. M. Brain-specific tropomyosins
TMBr-1 and TMBr-3 have distinct patterns of expression during development and in adult
brain. Proc. Natl. Acad. Sci. U. S. A. 90, 9857–9861 (1993).
196. Todd, M. A. M., Huh, M. S. & Picketts, D. J. The sub-nucleolar localization of PHF6 defines
its role in rDNA transcription and early processing events. Eur. J. Hum. Genet. EJHG 24,
1453–1459 (2016).
179
197. Southard-Smith, E. M., Kos, L. & Pavan, W. J. Sox10 mutation disrupts neural crest
development in Dom Hirschsprung mouse model. Nat. Genet. 18, 60–64 (1998).
198. Labosky, P. A. & Kaestner, K. H. The winged helix transcription factor Hfh2 is expressed in
neural crest and spinal cord during mouse development. Mech. Dev. 76, 185–190 (1998).
199. Ko, J. Y., Oh, S. & Yoo, K. H. Functional Enhancers As Master Regulators of Tissue-Specific
Gene Regulation and Cancer Development. Mol. Cells 40, 169–177 (2017).
200. Chung, H.-R. et al. PHF13 is a molecular reader and transcriptional co-regulator of
H3K4me2/3. eLife 5, (2016).
201. Qiu, Y. et al. Combinatorial readout of unmodified H3R2 and acetylated H3K14 by the
tandem PHD finger of MOZ reveals a regulatory mechanism for HOXA9 transcription.
Genes Dev. 26, 1376–1391 (2012).
202. Liu, Z. et al. Structural basis of plant homeodomain finger 6 (PHF6) recognition by the
retinoblastoma binding protein 4 (RBBP4) component of the nucleosome remodeling and
deacetylase (NuRD) complex. J. Biol. Chem. 290, 6630–6638 (2015).
203. Kalender Atak, Z. et al. High accuracy mutation detection in leukemia on a selected panel
of cancer genes. PloS One 7, e38463 (2012).
204. Dephoure, N. et al. A quantitative atlas of mitotic phosphorylation. Proc. Natl. Acad. Sci. U.
S. A. 105, 10762–10767 (2008).
205. Mayya, V. et al. Quantitative phosphoproteomic analysis of T cell receptor signaling
reveals system-wide modulation of protein-protein interactions. Sci. Signal. 2, ra46 (2009).
206. Zhou, H. et al. Toward a comprehensive characterization of a human cancer cell
phosphoproteome. J. Proteome Res. 12, 260–271 (2013).
180
207. Kearse, M. et al. Geneious Basic: an integrated and extendable desktop software platform
for the organization and analysis of sequence data. Bioinforma. Oxf. Engl. 28, 1647–1649
(2012).
208. Mets, E. et al. MicroRNA-128-3p is a novel oncomiR targeting PHF6 in T-cell acute
lymphoblastic leukemia. Haematologica 99, 1326–1333 (2014).
209. Brun, A., Börjeson, M. & Forssman, H. An inherited syndrome with mental deficiency and
endocrine disorder. A patho-anatomical study. J. Ment. Defic. Res. 18, 317–325 (1974).
210. Turner, G. et al. The clinical picture of the Börjeson-Forssman-Lehmann syndrome in males
and heterozygous females with PHF6 mutations. Clin. Genet. 65, 226–232 (2004).
211. Todd, M. A. M., Ivanochko, D. & Picketts, D. J. PHF6 Degrees of Separation: The
Multifaceted Roles of a Chromatin Adaptor Protein. Genes 6, 325–352 (2015).
212. Lee, J.-Y. Uncorking gastrulation: the morphogenetic movement of bottle cells. Wiley
Interdiscip. Rev. Dev. Biol. 1, 286–293 (2012).
213. Tamulonis, C. et al. A cell-based model of Nematostella vectensis gastrulation including
bottle cell formation, invagination and zippering. Dev. Biol. 351, 217–228 (2011).
214. Ossipova, O., Chuykin, I., Chu, C.-W. & Sokol, S. Y. Vangl2 cooperates with Rab11 and
Myosin V to regulate apical constriction during vertebrate gastrulation. Dev. Camb. Engl.
142, 99–107 (2015).
215. Choi, S.-C. & Sokol, S. Y. The involvement of lethal giant larvae and Wnt signaling in bottle
cell formation in Xenopus embryos. Dev. Biol. 336, 68–75 (2009).
216. Mu, Q. et al. Imp2 regulates GBM progression by activating IGF2/PI3K/Akt pathway.
Cancer Biol. Ther. 16, 623–633 (2015).
181
217. Christiansen, J., Kolte, A. M., Hansen, T. v O. & Nielsen, F. C. IGF2 mRNA-binding protein 2:
biological function and putative role in type 2 diabetes. J. Mol. Endocrinol. 43, 187–195
(2009).
218. Barghash, A., Golob-Schwarzl, N., Helms, V., Haybaeck, J. & Kessler, S. M. Elevated
expression of the IGF2 mRNA binding protein 2 (IGF2BP2/IMP2) is linked to short survival
and metastasis in esophageal adenocarcinoma. Oncotarget 7, 49743–49750 (2016).
219. Degrauwe, N., Suvà, M.-L., Janiszewska, M., Riggi, N. & Stamenkovic, I. IMPs: an RNA-
binding protein family that provides a link between stem cell maintenance in normal
development and cancer. Genes Dev. 30, 2459–2474 (2016).
220. Chakravarthy, H. et al. Identification of DPPA4 and other genes as putative Sox2:Oct-3/4
target genes using a combination of in silico analysis and transcription-based assays. J.
Cell. Physiol. 216, 651–662 (2008).
221. Masaki, H., Nishida, T., Kitajima, S., Asahina, K. & Teraoka, H. Developmental pluripotency-
associated 4 (DPPA4) localized in active chromatin inhibits mouse embryonic stem cell
differentiation into a primitive ectoderm lineage. J. Biol. Chem. 282, 33034–33042 (2007).
222. Oliviero, G. et al. The variant Polycomb Repressor Complex 1 component PCGF1 interacts
with a pluripotency sub-network that includes DPPA4, a regulator of embryogenesis. Sci.
Rep. 5, 18388 (2015).
223. Kandoth, C. et al. Mutational landscape and significance across 12 major cancer types.
Nature 502, 333–339 (2013).
224. Lower, K. M. et al. Mutations in PHF6 are associated with Börjeson-Forssman-Lehmann
syndrome. Nat. Genet. 32, 661–665 (2002).
182
225. Kim, S. G., Harwani, M., Grama, A. & Chaterji, S. EP-DNN: A Deep Neural Network-Based
Global Enhancer Prediction Algorithm. Sci. Rep. 6, 38433 (2016).
226. Kim, S. G. et al. Opening up the blackbox: an interpretable deep neural network-based
classifier for cell-type specific enhancer predictions. BMC Syst. Biol. 10 Suppl 2, 54 (2016).
227. He, Y. et al. Improved regulatory element prediction based on tissue-specific local
epigenomic signatures. Proc. Natl. Acad. Sci. U. S. A. 114, E1633–E1640 (2017).
228. Erwin, G. D. et al. Integrating diverse datasets improves developmental enhancer
prediction. PLoS Comput. Biol. 10, e1003677 (2014).
229. Yuan, C.-C. et al. Histone H3R2 symmetric dimethylation and histone H3K4 trimethylation
are tightly correlated in eukaryotic genomes. Cell Rep. 1, 83–90 (2012).
230. Maatouk, D. M. et al. Genome-wide identification of regulatory elements in Sertoli cells.
Dev. Camb. Engl. 144, 720–730 (2017).
231. Rhie, S. K. et al. Identification of activated enhancers and linked transcription factors in
breast, prostate, and kidney tumors by tracing enhancer networks using epigenetic traits.
Epigenetics Chromatin 9, 50 (2016).
232. Tu, W. J. et al. Priming of transcriptional memory responses via the chromatin accessibility
landscape in T cells. Sci. Rep. 7, 44825 (2017).
233. Pohl, B. S. & Knöchel, W. Overexpression of the transcriptional repressor FoxD3 prevents
neural crest formation in Xenopus embryos. Mech. Dev. 103, 93–106 (2001).
234. Zweier, C. et al. Females with de novo aberrations in PHF6: clinical overlap of Borjeson-
Forssman-Lehmann with Coffin-Siris syndrome. Am. J. Med. Genet. C Semin. Med. Genet.
166C, 290–301 (2014).
183
235. Park, J. et al. Identification of the early and late responder genes during the generation of
induced pluripotent stem cells from mouse fibroblasts. PloS One 12, e0171300 (2017).
236. Pfaff, N. et al. Inhibition of miRNA-212/132 improves the reprogramming of fibroblasts
into induced pluripotent stem cells by de-repressing important epigenetic remodelling
factors. Stem Cell Res. 20, 70–75 (2017).
237. Ogier, J. M. et al. CHD7 deficiency in ‘Looper’, a new mouse model of CHARGE syndrome,
results in ossicle malformation, otosclerosis and hearing impairment. PloS One 9, e97559
(2014).
238. Layman, W. S., Hurd, E. A. & Martin, D. M. Reproductive dysfunction and decreased GnRH
neurogenesis in a mouse model of CHARGE syndrome. Hum. Mol. Genet. 20, 3138–3150
(2011).

Abstract (if available)

Abstract Börjeson-Forssman-Lehmann Syndrome (BFLS) is an X-linked intellectual disability syndrome that is characterized by unique facial features that involve broad, course faces, heavy- set jaw bones, and thickened calvaria among other symptoms. This syndrome is characterized by mutation in a dual PHD protein, PHF6. This protein has been recently characterized as being associated with ribosomal DNA transcription, but very few studies have shown any role for PHF6 in development or a cellular phenotype in a developmental setting despite its role in BFLS. It is the subject of this thesis to define the role of PHF6 in neural crest development, in order to determine how this characteristic craniofacial phenotype arises, and to determine a cellular mechanism of PHF6 activity within neural crest cells. I used an approach that combined the strengths of Xenopus laevis embryology to define the role of Phf6 in neural crest in vivo, and human embryonic stem cells differentiated to neural crest cells to define the cellular and biochemical characteristics of PHF6 in vitro. I found that PHF6 protein and Phf6 mRNA is highly expressed in the neural plate border and in pre-migratory neural crest cells, but that there is reduced to absent expression of PHF6 in migratory neural crest cells. I found that Phf6 depletion leads to a loss of temporal control of neural crest migration, while ectopic expression prevents migration in a tissue- and cell-autonomous manner. Furthermore, I found that PHF6 tracks to poised but inactive enhancer regions in pre-migratory neural crest, and physically associates with the chromatin remodeler CHD7 in this context without the presence of other neural crest active chromatin remodeling factors. This led me to develop the model that PHF6 binds to poised enhancer regions to prevent precocious activation of these regions through the activity of chromatin remodelers like CHD7, and it maintains the poised state until it is down-regulated at the point of neural crest epithelial-to-mesenchymal transition.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Opposing function of chd7 and phf6 in zebrafish craniofacial development

PDF

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

PDF

Chromatin remodeling factors Chd7 and Phf6 in craniofacial and heart development

PDF

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

PDF

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

PDF

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

PDF

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

PDF

Genetics and the environment: evaluating the role of noncoding RNA in autism spectrum disorder

PDF

A novel role for post-migratory neural crest cells in cardiac outflow tract alignment: insights into the molecular basis of concomitant congenital cardiac and cleft malformations

PDF

Mixed lineage leukemia proteins (MLLs), their effect as coregulators on target gene expression and global histone methylation

PDF

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

PDF

The histone H2A deubiquitinase MYSM1 regulates CCR9 expression on CD4⁺ T cells and thymocytes

PDF

Induction of hypersignaling as a therapeutic approach for treatment of BCR-ABL1 positive Acute Lymphoblastic Leukemia (ALL) cells

PDF

Identification of novel epigenetic biomarkers and microRNAs for cancer therapeutics

PDF

Molecular characterization of the HIV-1 Vpu protein and its role in antagonizing the cellular restriction factor BST-2/tetherin both in vitro and in vivo

PDF

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

PDF

Characterization of a fragment in secreted Hsp90α with potential therapeutic benefits in wound healing and cancer

Asset Metadata

Creator Moran, Erin Catherine (author)

Core Title Epigenetic checks and balances: PHF6 activity restricts neural crest migration

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Genetic, Molecular and Cellular Biology

Publication Date 12/13/2017

Defense Date 04/25/2017

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

Tag developmental disorders,enhancer,epigenetics,histone tail reader,neural crest,OAI-PMH Harvest

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Stallcup, Michael (committee chair), Bajpai, Ruchi (committee member), Maxson, Robert (committee member), Xu, Jian (committee member)

Creator Email decembercat17@gmail.com,erinmora@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-462454

Unique identifier UC11266915

Identifier etd-MoranErinC-5714.pdf (filename),usctheses-c40-462454 (legacy record id)

Legacy Identifier etd-MoranErinC-5714.pdf

Dmrecord 462454

Document Type Dissertation

Rights Moran, Erin Catherine

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