the role of diencephalon/mesencephalon homeobox 1 …...iv acknowledgments it is undeniable that i...
TRANSCRIPT
The role of Diencephalon/Mesencephalon Homeobox 1
(Dmbx1) in Zebrafish Visual System Development
by
Loksum Wong
A thesis submitted in conformity with the requirements
for the degree of Doctor of Philosophy
Department of Cell and Systems Biology
University of Toronto
© Copyright by Loksum Wong 2013
ii
The role of diencephalon/mesencephalon homeobox 1 (Dmbx1) in
zebrafish visual system development
Loksum Wong
Doctor of Philosophy
Department of Cell and Systems Biology
University of Toronto
2013
Abstract
Visual system development is highly conserved across all vertebrates. The retina receives visual
inputs and transfers the sensory information to the brain for processing. In zebrafish, the largest
synaptic target of the retinal projections is the optic tectum, which acts as a relay center that
organizes visuomotor activities. My research has revealed that the transcriptional regulator
Diencephalon/mesencephalon homeobox 1 (dmbx1), of which there are two paralogs
represented in the zebrafish genome, plays a pivotal role in coordinating neurogenesis in these
functionally integrated regions. During the course of my doctoral research, I have characterized
the functions of dmbx1 paralogs in zebrafish retinal and tectal development and showed that
they are required for sustaining growth in the retina and the optic tectum. Functional deficiency
of the dmbx1 genes caused a severe cell cycle defect in progenitor cells, resulting in reduced
tissue growth and delayed differentiation within those structures. Further analysis revealed that
progenitor cells require dmbx1 to reduce the level of cyclinD1 expression, in order to promote
cell cycle exit. In addition, I discovered that misexpression of both dmbx1 genes could duplicate
wnt3a expression and lead to enhanced proliferation at the tectal midline. This suggested that
the two transcription factors could regulate different cellular mechanisms with the presence of
canonical Wnt signaling. Overall, this research demonstrates that dmbx1 is involved in the
iii
timing of retinal and tectal neurogenesis, and provides novel insight into the transcriptional
regulation of visual system development in vertebrates.
iv
Acknowledgments
It is undeniable that I have learnt so much and mature a lot during my years in grad school.
However, I know none of this would have happened if it were not for the following people, so I
would like to express my deepest gratitude to them.
My Ph.D. supervisor ˗ Vince, thank you for believing in me all these years. I was never one of
those students that have glorious marks or exceptional research background. But you are willing
to take a chance on me and I hope I have not disappointed you. I feel like we are growing as a
team since I was a new grad student while you were a new PI back then. I think we both had
"evolve" a lot along the way. Thank you so much for guiding me and showing me how to
become an independent scientist.
To my committee member, Ashley, even though you are not my supervisor, I always think of
you as my mentor. Thank you for teaching me the techniques I need to complete my
experiments and everything I now know about zebrafish. Your devotion to your lab and how
much you take care for everyone is something so rare and so precious. I understand it is
frustrating at times but I do believe that the quality of work that comes out of your lab will be
much better because you care.
Mei, my committee member and ex-supervisor, thank you for letting me carried out my first
official project in your lab. I have never been in a lab that has such a great dynamics, and the
way you manage your lab is something I wish I can do one day. Moreover, I appreciate the
advice you have for my project and your guidance as well. Your suggestions are always very
helpful.
v
I would also like to thank the following people in the department. Tamar, Ian, Nalini – without
you guys I will never be able to complete my degree as smooth as it is. You make all the
administrative works look so easy. Jim (Canada Best Handyman) – thank you for fixing all these
equipment in our lab so we can carry out our experiment. Terry – it never fail to amaze me how
you remember all these things that are in the shop. Henry and Audrey – thank you for teaching
me everything that I need to know about cryosection and confocal. You guys keep me company
for all those countless hours in the imaging facility. Olivia, Erich, Jason, Hannah, Nicole, Jaffer
– thank you for all those fruitful conversations we had and all the lab stories that we shared.
Mike - thank you for making our lab meeting actually worth going. Bruno - thank you for
sharing all those crazy stories and pranks to keep us entertained. Namita and Sue – I will miss
you girls, you two are really fun to work with. Monica – thank you for listening to us and
someone that holds the lab together. Lan – thank you for feeding us and taking care of us. You
are a really great friend. Steph – I am glad that you are the person whom I get to share my grad
school experience with. You are one of those very few people which I hope will be a PI one day
since you are a great scientist. Last but not least, thanks all past and present members of the
Tropepe Lab, as well as all those 5th and 6
th floor people who I get to meet during my years in
grad school. It is my pleasure to be able to work with you all.
I am also very thankful to have a group of wonderful friends surrounding me over the years as I
know it is very easy for friends to drift apart growing up, especially our jobs lead us to different
paths in life. I am very fortunate to have friends like you guys. Thank you “Haigers”, especially
Joni, Alex, and Vivian. Even though you have no ideas what molecular biology is, you guys still
listen and support every work-related things I said over the years. You three are lifelong friends
that we can share our problems with no matter where, when, or what.
vi
I would also like to thank my in-laws – especially Esa and Kenneth. Ever since I have known
you all, you have been treating me like your daughter. I know I could not have been luckier to
be able to join you all as part of this great family.
I am forever in debt to my family. To my little brother, Lokman, I miss our coffee session at the
Starbucks in AC. It was nice to be able to talk to someone in the family who understand what it
is like to work in a lab. To Mom and Dad, thank you for letting me pursue my science dream
even though it may not lead to anywhere. I know how frustrating you folks are since it takes me
forever to finish. I may not be the best daughter in the world, but I truly hope that one day I will
make you proud. Thank you for loving me unconditionally and you should know that I love you
guys very much as well.
Lastly, thank you Eric, my "better" half, who keeps me sane (or you are just as crazy as I am) all
these years. I honestly believe my marriage will fail if I had not married a scientist who
understands what kind of life a graduate student lives. No normal person will accept our
irregular work hours and minimum wages. Without your constant courage and support, I would
never be able to achieve this. Thank you for being my partner, my colleague, and my companion
on this journey (plus many more in the future). It would be unimaginable to be with someone
who does not know what PCR is....
vii
Table of Contents
Acknowledgments ...................................................................................................................... iv
Table of Contents ......................................................................................................................vii
List of Tables ........................................................................................................................... xiv
List of Figures ........................................................................................................................... xv
List of Abbreviations................................................................................................................ xix
Chapter 1 .................................................................................................................................... 1
General Introduction ............................................................................................................... 2
1.1 Diversity of visual systems ............................................................................................ 2
1.2 Conserved building blocks and molecular mechanisms in regulating the formation
of different types of eyes ............................................................................................... 3
1.3 Visual image processing in the nervous system .............................................................. 4
1.4 Visual system development in vertebrates...................................................................... 6
1.4.1 Architecture of the vertebrate visual system ......................................................... 6
1.4.1.1 Anatomy of the eye ................................................................................ 6
1.4.1.2 Laminar organization of the retina .......................................................... 6
1.4.1.3 Information processing centers in the brain and synaptogenesis of
the visual system .................................................................................... 9
1.4.2 Molecular mechanisms involved in establishing the visual system ..................... 13
1.4.2.1 Intrinsic factors that control patterning of the visual system.................. 13
1.4.2.2 Transcription factors that regulate neurogenesis in the visual
structures .............................................................................................. 14
1.5 Zebrafish visual system development .......................................................................... 16
1.5.1 Anatomy of zebrafish visual system ................................................................... 17
1.5.2 Neurogenesis in zebrafish retina and optic tectum .............................................. 18
1.5.3 Transcription factors are key players in zebrafish visual system development .... 20
1.6 Diencephalon/Mesencephalon Homeobox 1(Dmbx1) .................................................. 21
viii
1.6.1 A novel member of the paired-type homeodomain gene family .......................... 21
1.6.2 Role of Dmbx1 in visual system development .................................................... 21
1.7 Aims and Objectives .................................................................................................... 22
Chapter 2 .................................................................................................................................. 24
Characterization of Diencephalon/Mesencephalon Homeobox 1 (dmbx1) Paralogs in
Zebrafish .............................................................................................................................. 25
2.1 Introduction ................................................................................................................. 25
2.1.1 Duplicated gene pairs with diverged functions ................................................... 25
2.1.1.1 Neofunctionalization ............................................................................ 26
2.1.1.2 Subfunctionalization ............................................................................ 27
2.1.2 Phylogenetic analyses of Dmbx1 orthologs ........................................................ 27
2.1.3 Duplication of dmbx1 occurred independently in Cnidarian and Teleost ............. 29
2.1.4 Expression patterns of dmbx1 across different species ........................................ 30
2.2 Results......................................................................................................................... 33
2.2.1 Spatiotemporal expression profiling of dmbx1a and dmbx1b in zebrafish ........... 33
2.2.2 Efficacy and dosage requirements for dmbx1a and dmbx1b morpholinos............ 41
2.2.3 Midbrain phenotype in dmbx1a- and dmbx1b-morphant embryos ....................... 46
2.2.4 Hindbrain abnormality in dmbx1a and dmbx1b morphant embryos ..................... 48
2.2.5 Retinal defects in dmbx1a and dmbx1b knock down embryos ............................. 50
2.2.6 Functional divergence of dmbx1 paralogs in brain and retina .............................. 52
2.2.7 Rescue of dmbx1a and dmbx1b morphant phenotypes with zebrafish and
mouse dmbx1 mRNA ......................................................................................... 55
2.3 Discussion ................................................................................................................... 60
2.3.1 Spatiotemporal expression of zebrafish dmbx1 paralogs is highly conserved
with other vertebrates ......................................................................................... 60
2.3.2 Dmbx1a and Dmbx1b have partially overlapping expression patterns and
functions in the central nervous system .............................................................. 61
2.3.3 A limitation in morphant rescue analyses due to the dorsalization phenotype. .... 62
ix
2.3.4 Conservation of autoregulation in mouse and zebrafish dmbx1 genes ................. 63
2.3.5 Evolution of dmbx1 protein coding sequences .................................................... 64
2.3.6 Potential retention mechanism of duplicated dmbx1 gene pair ............................ 65
2.3.7 Functional divergence of the teleostean dmbx1 paralogs compared to other
vertebrates.......................................................................................................... 66
Chapter 3 .................................................................................................................................. 68
Dmbx1 Promotes Cell Cycle Exit in Retinal Progenitor Cells ............................................... 69
3.1 Introduction ................................................................................................................. 69
3.1.1 Retinogenesis in the zebrafish retina .................................................................. 69
3.1.2 Birth order of retinal neurons and glia in the zebrafish eye ................................. 70
3.1.3 Molecular components involved in cell cycle regulation..................................... 71
3.1.4 Retinal defects when cell cycle components are disrupted .................................. 73
3.1.5 Transcription factors involved in cell cycle regulation in retinal progenitor
cells ................................................................................................................... 74
3.1.6 The role of dmbx1 in retinal development........................................................... 75
3.2 Results......................................................................................................................... 76
3.2.1 Retinal differentiation is delayed in dmbx1 double morphants ............................ 76
3.2.2 Reduced eye size and retinal differentiation is not caused by pervasive cell
death .................................................................................................................. 78
3.2.3 Cell cycle defects in dmbx1 double morphants ................................................... 80
3.2.4 Retinal progenitor cells in dmbx1 morphants undergo complete mitosis ............. 83
3.2.5 Dmbx1 paralogs regulate cell cycle kinetics in retinal progenitor cells ................ 85
3.2.6 Delayed retinal differentiation in dmbx1 morphants............................................ 90
3.2.7 Cell autonomy of dmbx1 .................................................................................... 90
3.2.8 Identify potential cell cycle components controlled by Dmbx1 ........................... 96
3.2.9 Dmbx1 paralogs regulate cell cycle progression through cyclin D1 .................... 98
3.2.10 Cyclin D1 knockdown can rescue the differentiation defects in dmbx1 double
morphants ........................................................................................................ 100
x
3.2.11 Dmbx1 over-expression resulted in premature cell cycle exit ........................... 102
3.2.12 Misexpression of dmbx1 enhances the production of earlier-born neurons in
the retina .......................................................................................................... 104
3.3 Discussion ................................................................................................................. 106
3.3.1 Dmbx1 regulates retinal neurogenesis .............................................................. 106
3.3.2 The transcriptional relationship between dmbx1 and ccnd1 .............................. 107
3.3.3 Dual functions of Dmbx1 during retinal development in pre- and post-larval
stages ............................................................................................................... 110
3.3.4 Working model for Dmbx1 regulation of cell cycle exit during retinogenesis ... 110
3.3.5 Dmbx1 may interact with Notch and Wnt signaling pathways in the retina
during neurogenesis ......................................................................................... 113
Chapter 4 ................................................................................................................................ 115
Role of Dmbx1 in Midbrain Formation during Embryonic Development in Zebrafish ........ 116
4.1 Introduction ............................................................................................................... 116
4.1.1 Regionalization of midbrain in the neural tube ................................................. 116
4.1.2 Neurogenesis in the zebrafish brain .................................................................. 118
4.1.3 Specification of the optic tectum ...................................................................... 120
4.1.4 The canonical Wnt signaling pathway and its importance in midbrain
development..................................................................................................... 121
4.1.5 The role of Dmbx1 in regulating optic tectum development ............................. 123
4.2 Results....................................................................................................................... 124
4.2.1 Dmbx1 is required for specification of the optic tectum in zebrafish ................. 124
4.2.2 MHB is unaffected in dmbx1 morphants........................................................... 126
4.2.3 Neuronal differentiation in the hindbrain is affected in dmbx1 double
morphant .......................................................................................................... 129
4.2.4 Forebrain and ventral markers are unaffected in dmbx1 morphant embryos ...... 129
4.2.5 The loss of tectal tissue in dmbx1 double morphant is independent of cell
death ................................................................................................................ 132
xi
4.2.6 Delayed cell cycle exit in tectal progenitor cells in dmbx1 knock down embryos136
4.2.7 Dmbx1 is necessary and sufficient for cell cycle exit in tectal progenitor cells .. 139
4.2.8 Dmbx1 represses cyclinD1 in tectal progenitor cells ......................................... 139
4.2.9 Dmbx1 interacts with the Wnt canonical signaling pathway.............................. 144
4.2.10 Dmbx1 is positioned downstream of the Wnt ligand ......................................... 147
4.2.11 Dmbx1 induces ectopic midbrain structure ....................................................... 148
4.2.12 Hyperactivated Wnt signaling may induce the formation of ectopic tectal
structure ........................................................................................................... 151
4.3 Discussion ................................................................................................................. 157
4.3.1 Dmbx1 is dispensable for global brain patterning ............................................. 157
4.3.2 Dmbx1 is required for tectal development and neuronal specification in the
hindbrain .......................................................................................................... 158
4.3.3 Fewer apoptotic cells in the optic tectum of dmbx1 morphant embryos............. 159
4.3.4 Dmbx1 acts as a transcriptional repressor to regulate cyclinD1 level in tectal
progenitor cells ................................................................................................ 160
4.3.5 Dmbx1 is synergized with the canonical Wnt pathway ..................................... 161
4.3.6 Low penetrance of ectopic tectal structure in dmbx1-overexpressing embryos .. 165
4.3.7 Testing the functionality of the ectopic tectal structure ..................................... 165
Chapter 5 ................................................................................................................................ 168
Materials and Methods ....................................................................................................... 169
5.1 Zebrafish husbandry .................................................................................................. 169
5.2 Morpholino injections................................................................................................ 169
5.3 GFP fusion proteins ................................................................................................... 170
5.4 Standard and quantitative real time RT-PCR .............................................................. 170
5.5 Whole-mount in situ hybridization............................................................................. 171
5.6 Histology ................................................................................................................... 172
5.7 Cross-section area measurement ................................................................................ 172
xii
5.8 Retinotectal projections ............................................................................................. 172
5.9 Ectopic gene expression ............................................................................................ 173
5.10 Wholemount antibody staining .................................................................................. 173
5.11 Immunohistochemistry .............................................................................................. 173
5.12 Cyclopamine treatment .............................................................................................. 174
5.13 Cell death analyses .................................................................................................... 174
5.14 BrdU labeling ............................................................................................................ 175
5.15 Flow cytometry ......................................................................................................... 175
5.16 Transplantation .......................................................................................................... 176
5.17 Molecular evolutionary analyses ................................................................................ 176
5.18 Yeast two-hybrid ....................................................................................................... 176
5.19 Western Blot ............................................................................................................. 177
5.20 Statistical Analyses .................................................................................................... 177
5.21 Cloning for myc-tagged proteins................................................................................ 177
Chapter 6 ................................................................................................................................ 178
General Discussion ............................................................................................................. 179
6.1 Discrepancy between previously reported dsmbx1a phenotypes and the double
dmbx1 morphants examined in this study ................................................................... 179
6.2 Mechanism underlying the retention of dmbx1 genes during evolution ...................... 181
6.2.1 Neofunctionalization is likely the retention mechanism of dmbx1 paralogs in
zebrafish .......................................................................................................... 183
6.3 A potentially novel function for zebrafish dmbx1 paralogs in visual system
development .............................................................................................................. 185
6.3.1 Dmbx1 is the first paired-type homeodomain protein identified that
negatively regulates the cell cycle .................................................................... 185
6.3.2 Dmbx1-interacting partners during retinal and tectal development ................... 185
6.3.2.1 Dmbx1 may antagonize Vsx2 in a cis-regulatory network .................. 185
6.3.2.2 Protein-protein interaction of Dmbx1 for cell cycle regulation ............ 186
xiii
6.4 Future Experiments ................................................................................................... 187
6.4.1 Verify potential downstream targets of Dmbx1 in vitro .................................... 188
6.4.1.1 Whole transcriptome shotgun sequencing with dmbx1 morphants....... 188
6.4.1.2 Verifying potential direct targets of Dmbx1a/Dmbx1b ....................... 189
6.4.2 Examine the importance of dmbx1 in prey capture behaviour ........................... 191
6.4.2.1 Generation of dmbx1a and dmbx1b mutant stains ............................... 191
6.4.2.2 Prey capture study with dmbx1 mutant fish......................................... 192
6.5 Conclusion ................................................................................................................ 193
References or Bibliography ..................................................................................................... 196
Appendices ............................................................................................................................. 239
Appendix 1: Transcriptional auto-regulation of dmbx1 paralogs .............................................. 240
Appendix 2: The search for Dmbx1a protein-protein interactions ............................................ 246
Appendix 3: Dmbx1a and Dmbx1b antibodies ........................................................................ 250
Appendix 4: Other Publications ............................................................................................... 253
Copyright Acknowledgements ................................................................................................ 254
List of Tables
Table 2-1 A summary of dmbx1 expression patterns observed in different species. ..................31
Table 2-2 Concentrations of Dmbx1 mRNAs tested for counteracting the zebrafish knockdown
phenotype. ...............................................................................................................................56
Table S1: Summary of potential Dmbx1a interacting genes obtained from a yeast two-hybrid
screen ..................................................................................................................................... 248
xv
List of Figures
Figure 1.1 Schematic diagram of a vertebrate eye ..................................................................... 7
Figure 1.2 Architecture of non-mammalian midbrain...............................................................10
Figure 1.3 Retinal axon trajectories in non-mammals and mammals. .......................................12
Figure 2.1 Evolutionary models to address the retention of duplicated genes ...........................28
Figure 2.2 Semi-quantitative and quantitative RT-PCR analysis of dmbx1a and dmbx1b
expressions...............................................................................................................................34
Figure 2.3 Whole-mount in situ hybridization patterns of dmbx1a. ..........................................36
Figure 2.4 Whole-mount in situ hybridization patterns of dmbx1b. ..........................................39
Figure 2.5 Specificity of morpholino induced dmbx1a and dmbx1b knockdown using fusion
protein constructs Dmbx1a-GFP and Dmbx1b-GFP. ................................................................43
Figure 2.6 Dose-dependent changes in foxb1.2 and rhodopsin gene expression in dmbx1
morphants. ...............................................................................................................................45
Figure 2.7 Hypoplasia of the optic tectum in dmbx1 morphants at 48 hpf. ...............................47
Figure 2.8 Morphological defects in medulla oblongata observed in dmbx1 knockdown
morphants. ...............................................................................................................................49
Figure 2.9 Retinal hypoplasia in dmbx1 morphants at 72 hpf. ..................................................51
Figure 2.10 Distinct patterns of hindbrain foxb1.2 expression and retinotectal projections in
dmbx1a and dmbx1b morphant embryos. ..................................................................................54
Figure 2.11 Dmbx1 morphant rescued with zebrafish or mouse dmbx1 genes. .........................59
Figure 3.1 Several major retinal cell types are absent in dmbx1-deficiency embryos. ...............77
xvi
Figure 3.2 Retinal progenitor genes and optic stalk genes are expanded in dmbx1 knocked
down embryos ..........................................................................................................................79
Figure 3.3 No increase in cell death in dmbx1 morphants. .......................................................81
Figure 3.4 Cell proliferation assays in the retina of 72hpf embryos. .........................................82
Figure 3.5 Using propidium iodide analysis to examine cell cycle progression. .......................84
Figure 3.6 Cumulative BrdU analysis was used to examine whether the cell cycle length has
increased in the retinal progenitor cells of the morphants. .........................................................86
Figure 3.7 Using cumulative BrdU labelling analysis to determine cell cycle kinetics in control
and morphant embryos at 72hpf ...............................................................................................88
Figure 3.8 Differentiated retinal markers are observed in 4dpf and 5dpf dmbx1 morphant
embryos. ..................................................................................................................................91
Figure 3.9 Cell transplantation experiment showed that dmbx1 paralogs act cell autonomously
in the RGC. ..............................................................................................................................93
Figure 3.10 Transplantation experiments demonstrated that dmbx1 functions cell autonomously
in the developing retina. ...........................................................................................................94
Figure 3.11 Cryosection on wholemount in situ hybridization embryos with various cell cycle
markers at 2-3 dpf. ...................................................................................................................97
Figure 3.12 Wholemount in situ hybridization of cyclinD1 (ccnd1) expression changes when
dmbx1 is perturbed ...................................................................................................................99
Figure 3.13 Repression of cyclinD1 partially rescued dmbx1 morphants phenotype. .............. 101
Figure 3.14 Retinal progenitor cells of those dmbx1-overexpressing embryos undergo
premature cell cycle exit. ........................................................................................................ 103
Figure 3.15 Early cell cycle exit in dmbx1-overexpressing embryos lead to bias in earlier-born
retinal neuron at 72hpf. .......................................................................................................... 105
xvii
Figure 3.16 Potential Dmbx1 binding sites on ccnd1 promoter. ............................................. 109
Figure 3.17 Model of dmbx1 regulation of cell cycle exit in retinal progenitor cells. .............. 112
Figure 4.1 Schematic diagram of canonical Wnt pathway. ..................................................... 122
Figure 4.2 Whole-mount in situ hybridization of tectal markers in dmbx1 double knockdown
embryos at 48hpf embryos. .................................................................................................... 125
Figure 4.3 Whole mount in situ hybridization of various midbrain-hindbrain boundary (MHB)
markers. ................................................................................................................................. 127
Figure 4.4 Whole mount in situ hybridization of dmbx1a and dmbx1b in acerebellar (ace-/-
)
mutants. ................................................................................................................................. 128
Figure 4.5 Whole mount in situ hybridization of various hindbrain markers. ......................... 130
Figure 4.6 Whole mount in situ hybridization of various forebrain and ventral markers. ........ 131
Figure 4.7 Whole mount in situ hybridization of dmbx1a and dmbx1b in the absence of Shh
signaling. ............................................................................................................................... 133
Figure 4.8 Cell death assays on dmbx1 morphants between 48-72hpf. ................................... 134
Figure 4.9 Delay in cell cycle exit leads to reduced tectal growth in dmbx1 morphants. ......... 137
Figure 4.10 Overexpressing dmbx1 genes promote premature cell cycle exit in the tectum. ... 140
Figure 4.11 Dmbx1 can regulate ccnd1 expression in the optic tectum. ................................. 142
Figure 4.12 Repression of cyclinD1 rescues dmbx1 morphants phenotype. ............................ 143
Figure 4.13 Expression of Wnt signaling components in dmbx1 double morphants embryos. 145
Figure 4.14 Misexpression of dmbx1a and dmbx1b mRNA at 1-cell stage leads to an ectopic
structure at the dorsal midbrain in a small percentage of injected embryos at 48hpf. ............... 149
Figure 4.15 Coronal sections of loss- and gain-of-function tectal phenotypes at 48hpf and
72hpf...................................................................................................................................... 150
xviii
Figure 4.16 Cells in the ectopic midbrain structure process tectal identity.............................. 152
Figure 4.17 Expression patterns of Wnt signaling components in dmbx1-overexpressing
embryos at 48hpf. ................................................................................................................... 153
Figure 4.18 Cells in the ectopic midbrain structure have acquired tectal cell fate but they are
still proliferative. .................................................................................................................... 156
Figure 4.19 Schematic diagram of how dmbx1 regulates the development of the optic tectum.
.............................................................................................................................................. 164
Figure S1 Dmbx1a and dmbx1b expression patterns in dmbx1 knocked down embryos. ......... 241
Figure S2 Dmbx1a expression in single and double dmbx1 morphant embryos at 72hpf. ........ 243
Figure S3 Dmbx1b expression in single and double dmbx1 morphant embryos at 72hpf. ........ 245
Figure S4 Generating Dmbx1a-specific peptide fragment with GST tag at the N-term. ........... 251
Figure S5 Generating Dmbx1b-specific peptide fragment with GST tag at the N-term. .......... 252
xix
List of Abbreviations
Alx Aristaless-like homeobox
Apc Adenomatosis polyposis coli
Ascl Achaete-scute homolog
Ath5 Atonal homolog 7
BAC Bacterial artificial chromosome
BCIP 5-bromo-4-chloro-3'-indolyphosphate
bp Base pair
C.elegans Caenorhabditis elegans
Celsr2 Cadherin, EGF LAG seven-pass G-type receptor 2
c-fos Cellular oncogene fos
chx10 Ceh-10 homeodomain containing homolog
c-myc Cellular myelocytomatosis oncogene
Csx Cardiac-specific homeobox
C- Carboxyl
DAB 3,3'-Diaminobenzidine
DAPI 4',6-diamidino-2-phenylindole
dGFP Destabilized green fluorescent protein
DIG Digoxigenin
Dlx Distal-less homeobox gene
GRASP Immunoglobulin-like restricted axonal surface protein
DNA Deoxyribonucleic acid
dNTPs Deoxynucleotide triphosphates
E Embryonic day (for mouse)
Egr2b Early growth response
Elav Embryonic lethal, abnormal vision
Eng Engrailed
Etv Ets variant
FGF Fibroblast growth factor
FITC Fluorescein isothiocyanate
Fox Foxhead box
GABA Gamma aminobutyric acid
xx
Gbx Gastrulation brain homeobox
GFP Green fluorescent protein
Gsk-3 Glycogen synthase kinase 3
GST Glutathione S-transferase
Hesr Hairy/Enhancer of split [E(spl)]
HH Hamburger-Hamilton (embryonic stages of chick embryo)
Id Inhibitor of DNA binding
IPTG Isopropylthio-β-galactoside
Irx Iroquois homeobox protein
Isl Islet
K50 Lysine at amino acid position 50
lacZ Gene that encodes β-galactosidase
Lef Lymphocyte enhancer binding factor
LIM Lin11-Isl1-Mec-3
Meis Myeloid ecotropic viral integration site
MeLc Caudal medial lateral cell
MeLr Rostral medial lateral cell
MgCl2 Magnesium chloride
mM Millimolar
mMO Mismatch morpholino
MO Morpholino
MYC Myelocytomatosis oncogene
NBT Nitro blue tetrazolium chloride
Nkx2 NK2 transcription factor related
n-myc Neuroblastoma-Derived V-Myc Avian Myelocytomatosis Viral Related
Oncogene
N- Amino
OE Over-expression
Otx2 Orthodenticle homolog
P Post-embryonic day (for mouse)
Pax Paired box gene
PBT Phosphate-buffered saline
Pbx Pre-B-cell leukemia homeobox factor
PCR Polymerase chain reaction
xxi
pg picogram
Pit-1 POU domain, class1, transcription factor 1
PKC Protein kinase C
Pknox Pbx/knotted 1 homeobox
POU Pit-Oct-Unc
Prox1 Prospero-related homeobox gene
Q50 Glutamine at amino acid position 50
RNA Ribonucleic acid
RT-PCR Reverse transcription polymerase chain reaction
Rx Retinal homeobox gene
Shh Sonic hedgehog
Six Six homeobox 1
TALE Transcription activator-like effectors
TOP Tcf optimal promoter
TUNEL Terminal deoxynucleotidyl transferase dUTP nick end labeling
UTR Untranslated region
Vsx Visual system homeobox
Wnt Wingless-type MMTV integration site family, member
WT Wildtype
Zic Zinc finger protein family member
μm micrometer
1
Chapter 1
General Introduction
2
General Introduction
Human vision has been extensively studied over the past century and a substantial number of
discoveries have helped us understand how we see. However, it is still not completely clear how
neural development affects the assembly of neural circuits in the visual system. There is an
immense knowledge gap in the regulatory mechanism of the development of the visual
structures. Thus, studying the molecular components involved in formation of the visual
processing organs will provide important insights into how the visual system works.
1.1 Diversity of visual systems
The ability of living organisms to sense environments is essential for their reproduction, survival
and overall fitness. Visual, auditory, somatosensory, gustatory and olfactory systems are all part
of the sensory system that has been extensively studied over the years. Throughout evolution,
organisms have developed various specialized cells or organelles to sense their surroundings. In
general, more sophisticated sensory systems are found in vertebrates compared to invertebrates,
and this is likely associated with the complexity of their nervous systems (Farris, 2008; Kaas,
1989).
One of the sensory processing units in the metazoans lineage is the visual system. Cnidarians
and many basal bilaterians do not have the typical eyes as their visual organs, but these
organisms do possess photoreceptor cells (and in most cases the pigmented cells as well) that
carry out phototransduction similar to the vertebrate eye (Piatigorsky and Kozmik, 2004; Ward
et al., 2008). These photosensory cells are usually found across the surface of the animals’
bodies. For example, planarians (freshwater flatworms) have two ocelli (eyespots) to sense light
intensity (Arees, 1986); whereas the sea urchins have photoreceptors at the distal tip of each
tube foot, allowing them detect light radially like a giant compound eye (Lesser et al., 2011;
3
Ullrich-Lüter et al., 2011). Some invertebrates such as arthropods have two compound eyes that
are made up of highly organized ommatidial array (Osorio and Bacon, 1994). Furthermore,
some vertebrates have simple camera-type eyes, which receive light through the lens and project
an inverted image on the retina. Many terrestrial and aquatic vertebrates have this type of eye
since it provides higher resolution and visual acuity distance as these animals rely on distance
vision to secure food sources and avoid predators (Land, 1997). The diversity of visual system
sensory organs in these organisms is usually adaptive to their living environments. For instance,
fish in the deep sea have distinct visual pigments in their lens that allow them to detect
bioluminescence (Douglas et al., 1998; Warrant, 2000).
1.2 Conserved building blocks and molecular mechanisms in regulating the formation of different types of eyes
Despite the variations in visual system, it appears that the many of the structural components
involved in establishing a visual system are conserved across different phylogenetic lineages.
Photoreceptor cells, pigment cells and crystallins proteins are some of the basic modules that
can be found in most simple or compound eyes (Nilsson, 2009; Vopalensky and Kozmik, 2009).
Moreover, given the variability in the eye morphologies, it is intriguing to observe that a
common molecular pathway that is crucial in regulating eye development persisted throughout
evolution. In jellyfish (Tripedalia cystophora), c-opsins, Mitf, and J1-crystallin were expressed
in the photoreceptors, retinal pigment cells, and lens respectively (Kozmik et al., 2008). The
orthologs of these set of genes can also be found in the same cell types within the human eye,
although they have diverged into multi-member gene families and have become more
specialized in their functions (Shibahara et al., 2000; Terakita, 2005; Wistow, 2012). In
addition, regulation of eye development appears to be evolutionarily conserved at the gene
regulatory level. The most prominent example is the role of pax genes during eye evolution. A
4
single paxB gene was detected in the eyes of the Cnidarians (Kozmik et al., 2003). Structural
analysis suggested that this paxB gene was likely an ancestral copy of the vertebrate pax2 and
pax6 genes, as this single copy pax gene has a paired domain homologous to the one in pax2 as
well as a homeodomain found only in pax6 (Kozmik et al., 2003; Piatigorsky and Kozmik,
2004). In Drosophila, pax6 had been demonstrated to be both necessary and sufficient to induce
eye development (Halder et al., 1995; Quiring et al., 1994), whereas pax2 was shown to have a
role in lens formation (Dziedzic et al., 2009). Moreover, pax2 and pax6 are also involved in eye
development in vertebrates. Studies in zebrafish (Macdonald et al., 1995; Nornes et al., 1998),
chicken (Li et al., 1994a; Thanos et al., 2004), mouse (Grindley et al., 1995; Otteson et al.,
1998) and human (Glaser et al., 1992; Sanyanusin et al., 1995) found that pax2 was required for
the proper formation of the optic stalk, while pax6 is responsible for lens development and
correct eye size. Although aspects of the eye development are still regulated by the paxB/2/6
families, functional analyses on the evolution of these pax genes revealed that the ancestral
paxB gene from jellyfish (cnidarians) could only substitute the regulatory function of Pax6 in
Drosophila but not mouse (Kozmik et al., 2003; Ruzickova et al., 2009). This suggested that
there might be other novel regulators that arose through a more complex gene network
developed by the vertebrates during their visual system formation.
1.3 Visual image processing in the nervous system
When light enters the retina, photon particles are converted into electrical signals in a reaction
called phototransduction. During this process, opsins with specific absorbance spectra that are
present in rod and cone photoreceptors will change their conformations when stimulated by light
at particular wavelengths (Palczewski, 2012). These opsins are G-protein couple receptors
which consist of 11-cis retinal. Upon photoisomerization, the 11-cis retinal transforms into 11-
trans retinal and activates transducin, which is a regulatory unit of the G protein (Kefalov, 2012;
5
Palczewski, 2012). Transducin is active when GTP instead of GDP is bound to its alpha-subunit,
allowing this GTP-bound alpha-subunit to dissociate from the complex and subsequently
activates cyclic guanosine monophospate (cGMP) phosphodiesterase (Collery and Kennedy,
2010; Tesmer, 2008). In the dark, substantial amount of cGMP that are present in the cytoplasm
will keep those cGMP-gated sodium channels open and causes cells to depolarize (Baylor, 1996;
Stryer, 1986). However, the level of cGMP drops when phototransduction occurs due to the
activation of phosphodiesterase (Baylor, 1996; Stryer, 1986). This causes the closure of sodium
channels and hyperpolarization of the cell, which eventually leads to a decrease in
neurotransmitters level released from the photoreceptor cells to their downstream neurons
(Baylor, 1996). Thus, phototransduction cascade allows light particles to convert into synaptic
signals during visual inputs.
In addition to the sensory organ that receives visual cues, a nervous system also needs to be in
place in to convey the information into appropriate motion outputs. There are noticeable
differences in the neural structure at the anatomical and physiological level among these
organisms. In cnidarians, neurons are dispersed as a nerve net or nerve ring and they form a
network (Koizumi et al., 1992; Marlow et al., 2009; Satterlie, 2011). Thus no visual image will
be formed in those organisms within the phylum. In most insects, these visual inputs are sent to
the optic lobes within their brain, which is then processed for motion detection (Borst, 2009;
Fischbach and Hiesinger, 2008; Strausfeld, 2009). Vertebrates have a much more complex
neuronal organization called the central nervous system (CNS) that can be divided into many
small groups of neurons that specialize in integrating stimuli received from the environment and
providing appropriate commands to an organism’s behavior. In vertebrates, visual signals are
processed in the optic tectum in bony fish, amphibians, reptiles and avian (Nakamura, 2001;
Nevin et al., 2010; Udin, 2007). However, this relay center partially shifted from the superior
6
colliculus to the cerebral cortex in mammals (Espinosa and Stryker, 2012; Lee, 2011; Leopold,
2012).
1.4 Visual system development in vertebrates
1.4.1 Architecture of the vertebrate visual system
1.4.1.1 Anatomy of the eye
Most vertebrates have a pair of camera-type eyes as their visual sensory organs. The architecture
of the eye is very similar between mammals and non-mammals. The overall shape of the eyeball
is spherical (Figure 1.1A). The lens is situated in the front and the optic nerve exits from the
back of the eye to connect to the brain. The lens is made up of crystallins that help to refract
light and its flexibility provides better focus for the retina. The iris and the ciliary body together
hold the lens in the proper position, suspending it in the middle of the eye. The multilayered
transparent cornea covers both the lens and the iris in the front and enhances the optical power
by strengthening the focus of light in terrestrial animals. The posterior half of the eye consists of
three closely arranged layers. The outermost one is called sclera which offers structural support
and protection for the eye, the inner layer is the retina where photo transduction and signals
processing takes place. The middle layer is known as choroid and it consists of connective tissue
and blood vessels. The center of the eyeball is filled with a gel-like matter called the vitreous
humour.
1.4.1.2 Laminar organization of the retina
The structural organization of retina is highly conserved among vertebrates (Figure 1.1B). The
neural retina is covered by an epithelial cell layer called the retinal pigmented epithelial (RPE),
which protects photoreceptor cells against photo-oxidation and supplies nutrients to the retina
(Strauss, 2005). The retina is comprised of three neuronal layers which are separated by two
7
Figure 1.1 Schematic diagram of a vertebrate eye
(A) Cross-section of a vertebrate eye. The front of the eye is facing towards the left and the back
to the right. (B) A schematic for different cell types found in a laminated retina. Basal side is on
top, whereas the outer layer is at the bottom.
8
synaptic layers. The innermost cell layer is called the ganglion cell layer, which houses the
retinal ganglion cells (RGCs) that arborize to various regions in the brain, as well as displaced
amacrine cells in many species. The tier furthest back from the eye is the outer nuclear layer that
contains rod and cone photoreceptors. Rod cells are much more sensitive to light than cone
cells, thus rod photoreceptors are responsible for vision in dim light (or scotopic vision) (Okawa
and Sampath, 2007). By contrast, there are three major spectral classes of cones under direct
light (or photopic vision) – long-wavelength (maximum absorbance with red light), medium-
wavelength (maximum absorbance with green light) and short-wavelength (maximum
absorbance with blue light) (Bowmaker, 1998; Dacey, 1996). In non-mammals, there is an
additional type of cone that is sensitive to violet/ultraviolet light (Bowmaker, 1998; Hawryshyn,
2010; Witkovsky, 2000). These photoreceptors are activated by light and hyperpolarized during
the phototransduction process. Furthermore, the middle stratum is identified as the inner nuclear
layer that mainly consists of interneurons. The three major groups of neurons found within the
tier, arranged from basal to apical, are the amacrine cells, bipolar cells, and horizontal cells.
There are many subpopulations of amacrine (MacNeil and Masland, 1998) and bipolar cells
(Sanes and Zipursky, 2010) that connect with RGCs and their dendrites and synapses can be
found in the inner plexiform layer (Yamagata and Sanes, 2008), which are situated between the
two basal nuclear layers. Bipolar cells connect both the photoreceptors and the RGCs to transmit
information received from the rods and cones to the ganglion cells (Bloomfield and Dacheux,
2001; Hensley et al., 1993; Okawa and Sampath, 2007). Horizontal cells are closest to the outer
nuclear layer and they connect with photoreceptor cells in order to receive light transduction
signals (Thoreson and Mangel, 2012). Dendrites from both bipolar and horizontal cells can be
found connecting to the photoreceptor cells in the outer plexiform layer (Raviola and Gilula,
1975).
9
In addition to those six subtypes of neurons, glial cells are also found in the vertebrate retina.
Müller glia are major supporting cells that stretch across the three nuclear layers with their cell
bodies inserted into the inner nuclear layer. It has been shown that these type of glial cells have
regenerate capacity after injury in the retina, particularly in non-mammals (Fischer and Bongini,
2010; Reh and Levine, 1998). Another type of retinal glia found in plexiform layers and
ganglion cell layer is called microglia, which serves as an innate immune cell during infection or
is involved in phagocytosis to clean up any debris from dying retinal cells (Chen et al., 2002a;
Langmann, 2007). All these neuronal and glial cells can be found radially across the retina.
1.4.1.3 Information processing centers in the brain and synaptogenesis of the
visual system
In vertebrates, particularly non-mammals, the midbrain (or mesencephalon) is the major vision
recipient region in the brain. The midbrain is a neural structure that can be roughly partitioned
into two compartments. The dorsal half of the mesencephalon forms three major structures,
torus longitudinalis (dorsomedial), the optic tectum (dorsolateral) and the torus semicircularis
(ventrolateral). In contrast, the tegmentum is situated in the ventral midbrain. Tectum and
tegmentum are two morphologically distinct compartments. The architecture of the tectum is
arranged in stratified neuropils, and there are five major laminae in the tectum. From
dorsolateral to ventromedial, they are stratum opticum (SO), stratum fibrosum et griseum
superficiale (SFGS), stratum griseum centrale (SGC), stratum album centrale (SAC) and stratum
periventriculare (SPV) (Corbo et al., 2012; Nevin et al., 2008; Nevin et al., 2010). It has been
well-established that the optic nerve projects to the midline ventrally and crosses at the midline
optic chiasm. Since these organisms (such as zebrafish) do not have binocular vision, all RGC
axon fascicles continue to project dorsally to the contralateral side of the brain along the optic
tract (Figure 1.3A). When the optic fibers arrive at the rostral tectum, these axons begin to target
10
Figure 1.2 Architecture of non-mammalian midbrain.
On the left is a schematic diagram showing the location of the midbrain along the neural tube.
Top right is a cartoon of a coronal section of the midbrain, which is divided dorsoventrally. The
top section of the cartoon is the dorsal midbrain, which consists of three refined structures and
the bottom is the ventral tegmentum. One of the regions in the dorsal midbrain is the optic
tectum. It is a stratified structure and five main layers are indicated in the bottom right of the
diagram.
11
specific tectal layers in order to establish a functional visual circuitry (Del Bene et al., 2010;
Nikolaou et al., 2012). The topography of these retinotectal projections mapped from
dorsoventral and nasal-temporal axes of the retinal neurons to the ventrodorsal and
anteroposterior axes.
In mammals, there are two main areas in the brain that connect with the RGCs, one of them is
the superior colliculus (Figure 1.3B). This layered structure is very similar to the optic tectum in
other non-mammal vertebrates. However, unlike the optic tectum, the superior colliculus does
not reconstitute the visual field but instead regulates reflex orienting responses. It can control
quick movement of the eyes towards any stimuli, called saccadic eye movement or gaze shift
(Freedman and Sparks, 1997; Lünenburger et al., 2001). The actual visual processing structure
in mammals is the visual cortex (Figure 1.3B). It is part of the cerebral cortex and is located in
the posterior part of the brain. The visual cortex is subdivided into more refined regions and the
main recipient of visual signals is the primary visual cortex. This stratified domain receives
afferent inputs from a region in the thalamus called lateral geniculate nucleus (LGN), which is
the other major area targeted by the RGCs (Nassi and Callaway, 2009). The optic nerves
segregate at the optic chiasm with some of the neurons projecting to the contralateral LGN
while the rest of the axon fascicles continue to arborize in the ipsilateral LGN (Alonso et al.,
2006). The LGN in both sides of the brain acts as a relay center in which it collects and
processes the visual information from the RGCs, then sends those signals to the visual cortex
within the same hemisphere (Alonso et al., 2006; Sanes and Zipursky, 2010). Cells from
different layers of the LGN innervate various designated subregions of the visual cortex (Nassi
and Callaway, 2009).
12
Figure 1.3 Retinal axon trajectories in non-mammals and mammals.
(A) In non-mammals, RGCs project to the contralateral optic tectum. Topology of the
retinotectal projections show that ventrodorsal axons synapse with mediolateral neurons in the
tectum. The nasal and temporal RGCs send their axons to the caudal and rostral tectal region
respectively. The optic nerves from the left and right eyes cross each other at the optic chiasm.
(B) In mammals, RGCs send their axons to the lateral geniculate nucleus on the contralateral
and ipsilateral sides. These connections then extend towards the visual cortex for visual
processing. The RGCs also target the contralateral superior colliculus which controls quick eye
movement.
13
1.4.2 Molecular mechanisms involved in establishing the visual system
Transcription factors (TFs) and signaling molecules (such as Wnt, Shh, and Fgf) are key players
that regulate formation of the visual system. They have important roles in tissue growth, neural
patterning, neuronal identity specification, and neural circuitry formation. Expression of
different TFs and morphogens along the dorsal-ventral, anterior-posterior, and left-right axes
dictate the positions of subdivided domains in the CNS. These factors are also important for
regulating cell cycle in progenitor cells and they help specify neuronal identities within a
particular neural structure. Many conserved pathways are shared among vertebrates (and
invertebrates as well) to establish and maintain a proper visual system.
1.4.2.1 Intrinsic factors that control patterning of the visual system
As mentioned at the beginning, one of the fundamentally conserved genes in retinal
development are Pax genes. In Drosophila, mutation or inhibition of the Pax6 gene leads to an
eyeless phenotype (Quiring et al., 1994). In mouse and frog, mutation in the Pax6 gene resulted
in small eye with no lens (Grindley et al., 1995; Hirsch and Harris, 1997a; Quinn et al., 1996;
Thompson et al., 2007; Zygar et al., 1998), and in human it usually results in Anirida and
(absence of the iris) and Peters' anomaly (defects in cornea) (Dansault et al., 2007; Quiring et
al., 1994). It has also been shown in other vertebrates that Pax6 is important for photoreceptor
regeneration and RPE formation (Bharti et al., 2012; Thummel et al., 2010).
Six3 is another homeobox gene that is crucial for visual system development. In flies, mutation
at the sine oculis locus leads to increased cell death and severe disruption in eye morphogenesis
and optic lobes formation (Cheyette et al., 1994). In teleosts, Six3 has been demonstrated to be
necessary and sufficient for forebrain and retina formation (Carl et al., 2002; Loosli et al., 1999).
In humans, Six3 is associated with holoprosencephaly (Wallis et al., 1999). Rx, another
14
transcription factor, is a pivotal patterning gene for eye development since mutations in the Rx
gene in human can lead to anophthalmia (absence of one or both eyes) and coloboma (failure of
closure of the optic fissure) (Bailey et al., 2004; London et al., 2009; Loosli et al., 2003;
Mathers et al., 1997).
The homeobox gene Otx2 is a crucial factor for both retinal and tectal development. Null
mutation of Otd (the Drosophila homolog of Otx) resulted in brain malformation and absence of
ocelli (Finkelstein et al., 1990). Conditional mutation in the retina revealed that Otd has a role to
in inducing photoreceptors (Terrell et al., 2012). Similar to the findings in Drosophila, Otx1-/-
;Otx2-/-
mutant mice have prominent ocular defects. Conditional knockout of Otx2 demonstrated
that this gene is necessary for the production of photoreceptors (Martinez-Morales et al., 2001;
Omori et al., 2011). Otx2 was also found to be a key regulator of RPE in vertebrates (Bovolenta
et al., 1997; Lane and Lister, 2012; Reinisalo et al., 2012), together with another well-conserved
transcription factor Mitf (Adijanto et al., 2012; Bora et al., 1999; Tsukiji et al., 2009).
1.4.2.2 Transcription factors that regulate neurogenesis in the visual structures
In addition to retina formation, many transcription factors are also required for neurogenesis in
the retina. Proneural genes (such as atonal) promote the production of photoreceptors in
Drosophila (Jarman et al., 1994). The homolog of Atonal in vertebrates, Math5 (or ath5), has
also been reported to be essential for the specification of retinal ganglion cells (Kay et al., 2001;
Mu et al., 2005). POU-domain transcription factor Brn3 (or Pou4f) is expressed in RGCs and
dorsal midbrain from fish to mouse (Brombin et al., 2011; DeCarvalho et al., 2004; Fedtsova
and Turner, 2001; Hirsch and Harris, 1997b; Liu et al., 2000; Pan et al., 2005). This gene
appears to be important for RGC specification and is possibly involved in axon pathfinding
during retinotectal projection. Another transcription factor, Vsx2 is a positive regulator for
15
retinal progenitors during early retinogenesis and specifies bipolar neuron cell fates during
differentiation. In fish, mouse and human, disrupting the function of Vsx2 early on causes
microphthalmia due to reduced proliferation and the absence of bipolar cells (Burmeister et al.,
1996; Passini et al., 1997; Reis et al., 2011). NeuroD is important for photoreceptor
specification in chick, zebrafish and rodent. Overexpressing this basic helix-loop-helix
transcription factor can increase the number of photoreceptors at the expense of Müller glia
formation in the retina (Morrow et al., 1999; Ochocinska and Hitchcock, 2008; Yan and Wang,
1998).
Many transcription factors identified over the years have been demonstrated to be conserved in
retinal patterning or promoting a particular subtype of retinal neuron. However, there is no
known transcription factor that can control the production or specification of Müller glia across
the vertebrates. Hesr2 has been shown to promote gliogenesis and repress rod genesis in mouse,
contrary to NeuroD (Satow et al., 2001). Transcription factor Prox1 appears to help maintain
cell survival in Müller glia cells in mouse (Cid et al., 2010). It is known that these Muller glia
serve as quiescent stem cells in the retina and are activated upon injury for regeneration
purposes in many non-mammalian species (Bernardos et al., 2007). Thus it is possible that the
molecular mechanisms controlling Müller glia specification and maintenance are unique among
vertebrates. Last but not least, transcription factors can also antagonize retinogenesis to maintain
progenitor cell identity. Members of the Id gene family are helix-loop-helix transcription factor
that are present in the retina, and they have the ability to prevent differentiation by regulating
the cell cycle kinetics in mitotic retinal cells (Du and Yip, 2011; Liu and Harland, 2003; Uribe
and Gross, 2010; Uribe et al., 2012). Numerous studies have illustrated that different
combinations of intrinsic factors are required to promote patterning, regulate cell growth,
specify neuronal differentiation, and establish proper neural circuitry. These molecular
16
mechanisms, together with other structural architecture and cellular mechanisms discussed
above highlighted the degree of conservation among vertebrates in visual system development.
1.5 Zebrafish visual system development
Zebrafish (Danio rerio) is an attractive model organism to study the vertebrate nervous system
given its rapid development, optical transparency, and the availability of numerous molecular
and genetic tools for manipulating gene functions. Embryos are fertilized externally in large
clutch sizes, thus facilitating a more substantial sample size. Recent advances in molecular and
cellular techniques provide new ways to study the visual system in zebrafish more efficiently.
Over the years, a great number of mutants, which have defects in vision, have been generated
through several large-scale ENU-screens. Many of these isolated mutant lines have specific
defects in neuroanatomy, neurogenesis, neural circuitry, physiology and behaviours that are all
related to the visual system (Brockerhoff et al., 1998; Jiang et al., 1996; Karlstrom et al., 1997;
Muto et al., 2005). In addition, antisense morpholino oligonucleotides have been extensively
used for knocking down specific target genes during embryonic development in zebrafish that
may be responsible for proper visual functions (Nasevicius and Ekker, 2000).
In addition to gene manipulations, zebrafish is also a great model to evaluate alternations in
visual behaviours given that there are well-established standard vision tests for optokinetic
reflex (OKR) and optomotor response (OMR) (Portugues and Engert, 2009). Recently, a new
technique has been developed using a calcium indicator GCaMP to monitor individual tectal
neuron firing in vivo upon stimulation received from the retina. This powerful tool allowed
neuroscientists to image neuronal activity at the cellular level (Muto et al., 2013). Overall, the
prominent similarities (both structurally and functionally) between the zebrafish visual system
17
and that of other vertebrates make it an excellent model organism for studying visual
development.
1.5.1 Anatomy of zebrafish visual system
Formation of the nervous system in zebrafish begins at tailbud stage, which is around 10 hours
post-fertilization (hpf). However, cells with neural fates begin to be specified much earlier
during gastrulation (5-10hpf) on the dorsal side of the embryos. These precursor cells then
migrate towards the dorsal midline to form the future neural plate. Establishment of the neural
tube in zebrafish initiates by thickening of the neuroepithelium (Strähle and Blader, 1994).
Neural cells then fold inwards and close at the midline to form the neural keel, with the medial
cells along the neural plate becoming the ventral part of the neural tube (Strähle and Blader,
1994). The neural keel subsequently detaches from the epithelium and a lumen cavitates at the
midline (Lowery and Sive, 2004). After a series of dynamic cellular movements that turn a flat
neural epithelial sheet into a three-dimensional tube, the central nervous system continues to
develop by undergoing neurogenesis and establishing complex sensory-motor circuitries during
late embryonic and early larval stages. One of such neuronal networks is the visual system,
which involves two major neuroanatomical structures – the retina and the optic tectum. The
retina receives sensory inputs and sends visual information to multiple processing sites within
the central nervous system. One of the largest recipients these RGCs synapses is the optic
tectum, which regionalizes within the dorsal midbrain (Burrill and Easter, 1994).
In zebrafish, patterning of the retina and the neural tube begin concurrently. Presumptive
forebrain, midbrain and hindbrain are marked along the neural tube rostrocaudally by patterning
genes such as Pax6, Otx2 and Gbx2 respectively (Kikuta et al., 2003; Kurokawa et al., 2006;
Matsunaga et al., 2000; Mercier et al., 1995; Rhinn and Brand, 2001). The bilateral eye field
18
also appears on both sides of the ventral forebrain during early somite stages and Rx is one of
the genes that is required for eye field specification (Stigloher et al., 2006). Once these neural
structures are compartmentized, progenitor cells within each region continue to proliferate until
the substantial size is reached and differentiation begins. Retinogenesis initiates earlier than
tectal neurogenesis in zebrafish (see below for details). By 36hpf, some of these differentiated
retinal neurons extend their axons as optic nerves and exit the retina (Laessing and Stuermer,
1996; Stuermer, 1988). The axons from both eyes extend towards the ventral midline and cross
each other to form the optic chiasm. The retinal nerve fibers continue to travel dorsally along the
optic tracts until they reach their target, which is the contralateral tectal lobe, and begin to
innervate that region. RGCs originated from the dorsal and nasal sides of the retina will arborize
in the ventral and temporal tectum, forming a topographic map of the retinotectal projections in
zebrafish by 72hpf (Burrill and Easter, 1994; Stuermer, 1988).
1.5.2 Neurogenesis in zebrafish retina and optic tectum
In the retina, the first cohort of retinal neurons begin to differentiate at 28hpf and retinogenesis
is completed around 72hpf when RGC axons reach various visual targets in the brain.
Neurogenesis of the retina produces six different subtypes of neurons and one type of glial cell
arranged across three distinct layers similar to other vertebrates as described above (Hu and
Easter, 1999; Stenkamp, 2007). These differentiated retinal neurons are situated in the central
retina, whereas retinal stem cells remain at the peripheral ciliary marginal zone. The continual
presence of actively proliferating cells in the central nervous system and the ability of these cells
to replace any injured or damaged neurons make zebrafish an attractive model for regeneration
studies, since mammals also have neural stem cells but they appear to be quiescent without any
regenerative potential. Retinal neurons connect with one another across the laminated retina in
order to establish proper transduction of light from the photoreceptors (Pujic and Malicki,
19
2004). Visual information received by the retinal ganglion cells is transmitted to various regions
in the brain, with optic tectum being the largest recipient of these axon inputs (Burrill and
Easter, 1994).
Since the tectum is a visuomotor processing center, it receives visual afferents from the retina
and sends motion efferents to various regions in the hindbrain. Neurogenesis in the tectum
begins around 2dpf and proliferating cells around the ventricle progressively migrate towards
the neruopil area close to the alar plate. It has been shown that tectum in teleost fish has the
ability to regenerate as there are stem cell niches at the medial, lateral, and posterior regions of
the tectum, called the germinal zone (Raymond and Easter, 1983). In the tectum, there are three
major types of neurons: superficial interneurons, periventricular projection neurons, and
periventricular interneurons. With the exception of superficial interneurons, which are located at
the most superficial stratum opticum (SO) and arborize ventrally into stratum fibrosum et
griseum superficiale (SFGS), the cell bodies of the other two tectal neurons are situated in the
stratum periventriculare layer (SPV) and project their dendrites dorsally to SFGS, stratum album
centrale (SAC), and stratum griseum centrale (SGC) (Corbo et al., 2012; Nevin et al., 2010). In
addition, these tectal neurons are a mixture of glutamatergic, GABAergic, or cholinergic cells,
depending on the kind of activities their carry out in the tectum (Robles et al., 2011).
Retinotectal projections begin to innervate the multi-laminated tectum around 60hpf, and each
retinal ganglion cell can only arborize in a single retinal-recipient layer (Xiao and Baier, 2007).
These retinal axons synapse with different tectal interneurons in the neuropil. It appears that
neurons receiving visual information are typically glutamatergic neurons, whereas those
connecting to the reticuspinal neurons in the spinal cord are usually GABAergic (Robles et al.,
2011; Sato et al., 2007). However, intratectal neurons can be either glutamatergic (excitatory) or
20
GABAergic (inhibitory) depending on the types of information received in the superficial
neuropils and how it transfers to the deep layers after processing (Del Bene et al., 2010; Nevin
et al., 2010).
1.5.3 Transcription factors are key players in zebrafish visual system
development
Transcription factors are known to be highly involved in the visual system development in
vertebrates. Proper development of the retina and its primary target, the midbrain optic tectum,
are crucial for the formation of a normal visual system in zebrafish (Picker et al., 1999). The
developmental interdependency of these two major anatomical structures is apparent at earlier
stages of embryogenesis when regional territories of the brain are being patterned. It has been
observed that many TFs temporally and/or spatially regulate the initial patterning,
differentiation, or wiring of neurons developed in both regions. Evidence from various studies
on TFs demonstrated different crucial roles for visual system development.
For instance, Pbx2/4 are TALE-class homeodomain transcription factors that regulate both
retinal and tectal patterning, as well as setting up the polarity in the tectum in order to guide
those retinotectal projections (French et al., 2007). Mab21l2 is another transcription factor that
is expressed in the presumptive eye field and midbrain primordium, and it is required for cell
survival for both retinal and tectal progenitor cells (Kennedy et al., 2004). Moreover, Meis1 is a
TALE-class homeodomain TF that is crucial for providing axial identity for both the retina and
the tectum during early embryonic development (Erickson et al., 2007). Furthermore, basic
helix-loop-helix Her4.2 plays an important role in maintaining retinal and tectal progenitor cells
identity via the Notch-signaling pathway (Jung et al., 2012; Yamaguchi et al., 2005). In
addition, Brn-3b and Islet2b are POU- and LIM-domain proteins that are expressed in retinal
ganglion cells and optic tectum. Both transcription factors are required to regulate proper
21
development of the two regions, as they also affect the topology of retinotectal projections
(DeCarvalho et al., 2004; Kikuchi et al., 1997). Overall, transcription factors have been
demonstrated to have diverse roles in regulating multiple aspects of the visual system during
zebrafish embryogenesis.
1.6 Diencephalon/Mesencephalon Homeobox 1(Dmbx1)
1.6.1 A novel member of the paired-type homeodomain gene family
Diencephalon/mesencephalon homeobox 1 (Dmbx1) is the most recent addition to the paired-
type homeodomain transcription factor family. The paired-typed protein family members all
share a well-conserved homeodomain, but what distinguishes each sub-family of the paired-type
homeodomain is the amino acid at position 50 within the motif, as it determines the specific
DNA-binding sequences among each subclass of paired-type protein (Chaney et al., 2005;
Galliot et al., 1999; Pellizzari et al., 1997). One type of homeodomain is the K50 whose
members include Otx, Gsc, and Dmbx1 (Takahashi et al., 2002). Other paired-type gene family
such as Pax and Arx belong to S50 and Q50 subclass respectively (Schneitz et al., 1993). Another
feature of the paired-type homeodomain transcription factors is the OAR (otp, arx, rx) peptide at
the carboxyl-terminal. It has been shown in some studies that this OAR domain may bind to
other transcription factors and allow its own homeodomain to activate transcription (Amendt et
al., 1999; Brouwer et al., 2003).
1.6.2 Role of Dmbx1 in visual system development
Embryonic expression of Dmbx1 in mouse and chicken revealed strong expression in the
presumptive midbrain (Broccoli et al., 2002; Gogoi et al., 2002; Kimura et al., 2005; Miyamoto
et al., 2002; Ohtoshi et al., 2002). In many teleosts, there are two copies of the dmbx1 genes due
to an additional round of whole genome duplication in the lineage of the ray-finned fish around
22
350 million years ago (Meyer and Van de Peer, 2005). Both dmbx1a and dmbx1b are expressed
in the optic tectum, although their expression domains do not overlap completely. Compared to
other basal chordates (such as Ciona) which express dmbx1 only in the posterior neural tube
(Stolfi and Levine, 2011; Takahashi and Holland, 2004), the abundance of dmbx1 in the
midbrain of the neural tube suggested that this gene might have taken on a novel role in
regulating midbrain development when the vertebrates diverged from the invertebrates .
In zebrafish, strong dmbx1a and weak dmbx1b expression can also be found in the retina in
addition to their tectal domains, and functional analysis showed that dmbx1a knock-down
embryos have a small eye phenotype and reduced tectal size (Kawahara et al., 2002). These two
pieces of evidence suggest a strong possibility that dmbx1a, together with dmbx1b, may have
acquired new function during evolution in coordinating the development of visual organs in
zebrafish.
1.7 Aims and Objectives
Taken from other reported paralogs and the types of evolutionary model that help explain
retention of duplicated gene pair, I hypothesize that dmbx1a and dmbx1b have diversified their
expression domains and possibly their functions in regulating visual system development in
zebrafish. Comparing the functions between zebrafish dmbx1 paralogs and their mouse ortholog
may help us understand how the gene pair evolved and took on a more crucial role during
evolution in establishing the visual function.
In this thesis, my research focuses mainly on how dmbx1 paralogs control the development of
the visual system. My specific aims are as follows:
23
1) To understand why the dmbx1 gene duplicates have been retained and to test whether the
zebrafish paralogs preserve the same function as the mouse Dmbx1;
2) To investigate the function of dmbx1 genes in retinal and tectal development using gain and
loss of function approaches;
3) To identify potential downstream targets of dmbx1 paralogs and investigate the molecular
mechanism through which these two transcription factors act.
24
Chapter 2
Characterization of Diencephalon/mesencephalon Homeobox 1
(dmbx1) Paralogs in Zebrafish
Part of this chapter is adapted from:
Chang, L., Khoo, B., Wong, L. and Tropepe, V. (2006). Genomic sequence and
spatiotemporal expression comparison of zebrafish mbx1 and its paralog, mbx2. Dev Genes
Evol. 216, 647–654.
My contribution to this paper includes the determination of dmbx1a and dmbx1b spatiotemporal
expression using RT-PCR and quantitative RT-PCR.
Wong, L., Weadick, C. J., Kuo, C., Chang, B. S. and Tropepe, V. (2010). Duplicate dmbx1
genes regulate progenitor cell cycle and differentiation during zebrafish midbrain and retinal
development. BMC Dev Biol. 10, 100.
I performed in situ hybridization to profile the spatiotemporal expression of dmbx1a and
dmbx1b. Also, I used specific morpholinos to knock down both dmbx1 genes and characterized
the morphological phenotypes in the single and double morphant embryos. I also performed
rescue experiments on the dmbx1 single morphants with zebrafish- or mouse-specific mRNA to
determine their conservation during evolution.
25
Characterization of Diencephalon/Mesencephalon Homeobox 1
(dmbx1) Paralogs in Zebrafish
2.1 Introduction
Paralogous genes that arise from whole genome duplication events are initially redundant and
they can perform the same biochemical functions (Ohno, 1999; Roth et al., 2007). This is typical
of more recent gene duplicates, as they have not had the opportunity to accumulate non-
synonymous changes in their sequences. The ability to buffer against any new mutation in one
duplicate ensures that the essential gene function is being carried out normally. This type of
“safety-net” strategy reduces lethality caused by spontaneous changes in the genetic material,
and helps increase an organism’s fitness or survival.
During evolution, these gene pairs often accumulate degenerative mutations that typically
transform one of the copies into a pseudogene, while the other one maintains its original
function. However, in some cases, both duplicated genes are retained in the genome. On
average, around 3-4% of gene duplicates are retained in the teleost genome, with the exception
of zebrafish, which is predicted to maintain both copies of about 20% of their duplicated genes
(Kassahn et al., 2009). While these paralogs usually have sequences with high similarity, the
majority of the gene pairs documented does not behave redundantly. In fact, the functions of
most of these paralogs appear to have modestly diverged. Their spatiotemporal expression
profiles and their biochemical functions may only partially overlap or even be completely
distinct from the ancestral gene.
2.1.1 Duplicated gene pairs with diverged functions
Retention of duplicated gene pairs usually is the result of diverged gene expression or altered
gene function between the two paralogs. These changes typically develop when selection is
26
more relaxed for one of the redundant copies allowing it to accumulate mutations that can
sometimes lead to different functions, without compromising their original role (Pegueroles et
al., 2013). Mutations can accumulate anywhere in the regulatory sequences (gene promoter,
introns, and 5’- or 3’- untranslated regions) or coding regions of the gene duplicates as long as
they are not both deleterious (Li et al., 2005). Changes in the regulatory regions are likely to
result in different spatiotemporal expression patterns between the two paralogs; however, their
protein function may remain the same. Engrailed 1/2 and Hoxa1/b1 in mouse are examples
where the protein coding regions of the paralogs have been shown to preserve the same
biochemical function even though their regulatory regions have altered (Hanks et al., 1995;
Tvrdik and Capecchi, 2006). However, the biochemical functions of the paralogs can diverge
from one another when degenerative mutations accumulate in the protein coding regions of the
gene pair. For instance, the Kit ligand and its receptors have both duplicated in zebrafish and
have developed more critical roles compared to other orthologs in regulating folliculogenesis
during ovarian development (Yao and Ge, 2010). The mechanism by which these paralogs are
modified at the genomic level can often be attributed to one of the established evolutionary
models — neofunctionalization or subfunctionalization.
2.1.1.1 Neofunctionalization
Neofunctionalization helps explain the retention of duplicate genes due to the acquisition of a
new biochemical function (Conant and Wolfe, 2008; Hahn, 2009; Han et al., 2009; Otto and
Yong, 2002). This model suggests that while one copy of the paralog retains its original
function, the constraint on the other is more relaxed and adaptive mutations are allowed to fix in
a population and acquire new function without any major consequences (Figure 2.1A).
Examples of this evolutionary model are rare due to the fact that experimental evidence for
neofunctionalization in vertebrates is rather difficult to analyze given the lack of ancestral genes
27
available (Tirosh and Barkai, 2007). However, several studies have reported this phenomenon in
mouse and zebrafish (Bertrand et al., 2007; Han et al., 2009; Huminiecki and Wolfe, 2004; Tello
et al., 2008). It is often unclear whether a new function observed in duplicated genes is due to
neofunctionalization, or instead the result of a parallel loss of the ancestral function in multiple
lineages. Hence, careful in depth analysis with the proper basal outgroup is required to confirm
that the duplicated pairs are retained in the genome due to neofunctionalization.
2.1.1.2 Subfunctionalization
Another evolutionary model has also been proposed to account for the retention of duplicate
genes during evolution. Subfunctionalization proposes that both gene duplicates together have
expression and/or functions complimentary to their pre-duplicated ancestral gene (Lynch and
Force, 2000; MacCarthy and Bergman, 2007; Massingham et al., 2001). In this case, gene
paralogs are no longer redundant to one another nor did they inherit a new function. Instead,
these duplicated genes each lost some aspects of their original functions and together they
perform the same task as the ancestral gene (Figure 2.1B). In most cases, studies have concluded
that duplicated gene pairs undergo subfunctionalization based on their expression domains (de
Souza et al., 2005; Rissone et al., 2010) rather than their functionalities (Hurley et al., 2007)
when compared to other single orthologous genes.
2.1.2 Phylogenetic analyses of Dmbx1 orthologs
Dmbx1 was initially thought to be a vertebrate specific gene which arose along with the
invention of the tripartite brain in the vertebrate central nervous system (Holland and Takahashi,
2005). Dmbx1 is an early midbrain specific marker in mouse, chicken, and zebrafish, although
its expression expands into the hindbrain and retina during later embryonic stages (Kawahara et
al., 2002; Zhang et al., 2002). However, closer examination across other phyla revealed that
28
Figure 2.1 Evolutionary models to address the retention of duplicated genes
Duplicated gene pairs initially have identical function. Subsequently the paralogs accumulate
mutations either in the protein coding or regulatory regions of these sister genes that change
either their biochemical functions or expression patterns or both. Once these mutations are fixed
in the population, the gene pairs are preserved in the genome due to function divergence. In
neofunctionalization model (A), one of the copies retains the ancestral gene function while the
other one takes on new function. In subfunctionalization model (B), each paralog become
specialized in different roles and together they complement the ancestral function.
29
this homeodomain transcription factor was also found in Ciona and Amphioxus, suggesting that
this gene could have originated in the chordate lineage (Takahashi and Holland, 2004). Due to
the gradual advance in DNA sequencing, genome annotation of other basal metazoans became
available. This greatly assisted in the search for additional dmbx1 orthologs in different species.
In other basal metazoans genomes, dmbx1 was identified in sea anemone (Nematostella
vectensis) (Ryan et al., 2006) and coral (Acropora millepora) (Hislop et al., 2005), but no
orthologs have been reported in other metazoan groups. Moreover, dmbx1 gene was never
reported in Drosophila or C. elegans, suggesting that this gene family might have been lost only
in protostomes during Bilaterian evolution (Ryan et al., 2006; Takahashi et al., 2002). However,
a paired-class homeodomain protein named Pph13 in Drosophila has been shown to resemble
other paired-type homeodomain transcription factors (such as Aristaless and Goosecoid) in
vertebrates (Goriely et al., 1999). In fact, this gene is required for morphogenesis and
phototransduction function in photoreceptor cells in fruit flies (Mishra et al., 2010; Zelhof,
2003), suggesting a role in visual system development. Overall, phylogenetic analyses indicated
that Dmbx1 is an ancient metazoans gene that might have been lost.
2.1.3 Duplication of dmbx1 occurred independently in Cnidarian and Teleost
All chordates have only a single copy of the dmbx gene, except for the teleost which underwent
an additional round of genome duplication and that resulted in two copies of the dmbx1 gene ‒
termed dmbx1a and dmbx1b. Interestingly, an independent duplication of the dmbx gene has
also been found in Acropora millepora but not Nematostella vectensis, which are both members
of the Cnidarian phylum (Hislop et al., 2005; Ryan et al., 2006). It is unclear whether the
duplication event in Acropora is an isolated incident, or whether some other Cnidarians may
also possess two copies of the dmbx genes. However, it appears that one of the Acropora
paralogs, Dmbx2-Am is the ortholog of the chordate Dmbx1, whereas Dmbx1-Am more closely
30
resembles other Cnidarian dmbx genes (Hislop et al., 2005). Phylogenetic comparison of dmbx1
paralogs in teleosts and coral could yield further insights into the evolutionary mechanism of
how these two gene pairs arose and become fixed in their respective genome, and whether these
two duplication events are truly independent from each other or not.
2.1.4 Expression patterns of dmbx1 across different species
As mentioned above, dmbx1 is evolutionarily conserved in metazoans and the expression pattern
of this gene have been reported in amphioxus, ascidians, teleosts, avians and mammals
(Summarized in Table 2-1). AmphiDmbx is expressed at the end of gastrulation in the anterior
mesendoderm, and extends rostrally into Hatschek`s diverticula; however, AmphiDmbx is not
detectable in the neural tube during embryogenesis or early larval stages (Takahashi and
Holland, 2004). On the other hand, CiDmbx is initially present in the Ciona neural tube at mid-
tailbud stage, and its expression is gradually restricted to a pair of visceral ganglion neurons
called A12.239 (Stolfi and Levine, 2011; Takahashi and Holland, 2004). It appears that CiDmbx
is co-expressed with Ci-Fgf8/17/18 and CiHox3 but positions posterior to Ci-pax2/5/8-A,
suggesting that CiDmbx is localized at the trunk-tail junction (Takahashi and Holland, 2004).
Expression of Dmbx1 in mouse and chick defines a discrete neural region that overlaps rostrally
with the pretectum and caudally with the posterior midbrain at the beginning of neurulation
(Broccoli et al., 2002; Ferran et al., 2007; Gogoi et al., 2002; Miyamoto et al., 2002; Ohtoshi et
al., 2002; Takahashi et al., 2002; Zhang et al., 2002). Expression of dmbx1 in the chick embryo
becomes apparent at HH stage 4 and continues to mark the midbrain and posterior diencephalon
at HH17 (Gogoi et al., 2002). Later during development, dmbx1begins to be expressed in a
bilateral stripe of cells rostrocaudally adjacent to the rhombic lip in the hindbrain (Gogoi et al.,
2002).
31
Table 2-1 A summary of dmbx1 expression patterns observed in different
species.
Spatial expressions of dmbx1 in different species are summarized in the table below. “+” and
“-” indicated the presence and absence of dmbx1 expression in the region listed at the top. “-/+”
indicates that the observation in that region is inconsistent among studies.
Eye Forebrain Midbrain Hindbrain Non-neural
Tissue
Amphioxus - - - - +
Ciona - - - + -
Chick - + + + +
Mouse -/+ + + + -
Zebrafish + + + + -
32
In mouse, expression of Dmbx1 first appears at E7.5 in the anterior neural tube and its
expression sharpens by E9.0 to localize only to the presumptive midbrain (Broccoli et al., 2002;
Miyamoto et al., 2002; Ohtoshi et al., 2002; Takahashi et al., 2002; Zhang et al., 2002). Around
E9.5, Dmbx1 begins to expand rostrally to the first prosomere, which will later develop into the
posterior commissure (Broccoli et al., 2002; Ohtoshi et al., 2002; Takahashi et al., 2002).
Dmbx1 becomes more restricted to the alar plate at E14.5 but it is never detected in the roof
plate (Broccoli et al., 2002; Ohtoshi et al., 2002). Other regions such as the dorsal optic cup are
also labeled by the Dmbx1 probe at E10-11 (Takahashi et al., 2002). In hindbrain, expression of
Dmbx1 is observed at E10.5 in the rhombomeres at the lateral edge of the rhombic lip (Broccoli
et al., 2002; Takahashi et al., 2002).
In zebrafish, dmbx1a and dmbx1b are the orthologs that derived from the single ancestral dmbx1
gene (Chang et al., 2006). Early expression of dmbx1a in the zebrafish embryo is found in the
presumptive mesencephalic and retinal territories based on previous vertebrate fate mapping
studies (Kawahara et al., 2002). Loss-of-function experiments have demonstrated that reduced
dmbx1a expression caused growth defects in both retina and optic tectum (Kawahara et al.,
2002). The paralog of this gene, dmbx1b, first arose from a fish-specific gene duplication event.
Both copies are remarkably similar to each other, with 72% overall identity. The paired-type
homeodomain, the N-terminus and the C-terminal otp-aristaless-rax (OAR) domains of both
genes are 100% identical.
In this chapter, I investigate whether dmbx1a and dmbx1b in zebrafish have diverged expression
patterns which may have led to paralog-specific function in regulating visual system
development. Detail analysis comparing the similarities and differences between dmbx1a and
dmbx1b showed that they have partial overlapping expression profiles and protein functions
33
during zebrafish embryonic development. In addition, rescue experiments between zebrafish
dmbx1 paralogs and mouse Dmbx1 ortholog yield further insights on the degree of functional
conservation within this gene family. Together with the expression and functional data,
phylogenetic analysis would help deduce the potential evolutionary model that dmbx1 gene
family may utilize to retain in the genome.
2.2 Results
2.2.1 Spatiotemporal expression profiling of dmbx1a and dmbx1b in zebrafish
To analyze the temporal expression of dmbx1a and dmbx1b, RT-PCR and quantitative real-time
RT-PCR were performed. No dmbx1a or dmbx1b mRNA was detected between 3-6 hpf,
suggesting that neither gene was likely to be maternally expressed (Figure 2.2A). At 9 hpf,
dmbx1a expression was first detected and its expression was maintained until 48 hpf (Figure
2.2A). In contrast, very low levels of dmbx1b mRNA were detected between 9 and 48 hpf
(Figure 2.2A). Quantitative analyses of dmbx1a and dmbx1b expression at 9 hpf and 24 hpf
indicated that levels of dmbx1a mRNA were 140-fold greater than dmbx1b at 9-10 hpf (tailbud),
but that only a 2-fold increase in dmbx1a expression compared to dmbx1b was observed by 24
hpf (Figure 2.2C). These data indicated that dmbx1a has a robust, early onset of expression at
tailbud stage compared to dmbx1b, and that in general dmbx1b expression was quantitatively
less than dmbx1a throughout the first 2 days of development. To determine if the dmbx1 genes
are expressed in adult zebrafish, I performed RT-PCR analysis on tissues from 6-month-old
animals. Strong expression of dmbx1a was observed in the forebrain, midbrain and hindbrain
tissue samples, whereas strong expression of dmbx1b was only observed in the midbrain (Figure
2.2B). The dmbx1a result was consistent with previous work reporting the expression pattern in
zebrafish by in situ hybridization (Kawahara et al., 2002).
34
Figure 2.2 Semi-quantitative and quantitative RT-PCR analysis of dmbx1a
and dmbx1b expressions.
(A) Semi-quantitative RT-PCR analysis of dmbx1a and dmbx1b with various embryonic stages
(listed on top). Both transcripts were first detected at 9hpf and continued to be expressed at
48hpf, and dmbx1a was expressed at higher levels than dmbx1b at all stages examined. Actin
served as a control. (B) Expression of dmbx1 paralogs was assessed in the zebrafish adult brain
by RT-PCR. RNA was isolated from forebrain, midbrain, and hindbrain tissue and RT-PCR
showed that dmbx1a was expressed more broadly than dmbx1b, which mostly localized in the
midbrain only. Actin serves as a control. Template from 24 hpf cDNA was used as positive
control based on (A), and dH2O was the negative control. (C) Quantitative analysis of dmbx1
transcript level at tailbud (~9hpf) and 24hpf. Dmbx1a (blue) is expressing 140 times more
compared to dmbx1b (yellow) at tailbud but their discrepancies closed in by 24hpf to only two-
fold.
35
To determine the spatial localization of dmbx1a transcripts, I performed whole-mount in situ
hybridization on 9-10 hpf (tailbud), and 1-6 days post-fertilization (dpf) zebrafish samples using
antisense dmbx1a and dmbx1b RNA probes (Figure 2.3 and 2.4). Dmbx1a expression patterns
from tailbud to 3 dpf were identical to previously published data (Kawahara et al., 2002). At 10
hpf, dmbx1a expression was evident in the presumptive mesencephalic, diencephalic and retinal
regions in a characteristic annulus shape (Figure 2.3A). At 1 dpf, dmbx1a was highly expressed
in the mesencephalon, the dorsal diencephalon and small bilateral areas in the
rhombencephalon, but it was never detected in the midbrain-hindbrain boundary or the retina
(Figure 2.3B). Dorsal views revealed that dmbx1a was present throughout the mesencephalic
periventricular zone (Figure 2.3C).
Transverse sections of whole-mount stained embryos were used to determine the
neuroanatomical distribution of the dmbx1a gene expression, which was detected primarily in
the dorsal domain of the mesencephalon (Figure 2.3D). By 2 dpf, dmbx1a expression was
apparent along the midline of the tectal region, as well as in the cells along the edge of the
hindbrain (Figure 2.3E-F, 2.3H). Coronal cryosections of the embryos revealed a puncta-like
pattern of dmbx1a-positive cells within the dorsal tecta (Figure 2.3G), suggesting that dmbx1a
was expressed in a subset of tectal neurons. From 3 to 6 dpf, expression of dmbx1a from the late
embryonic to early larval stages remained more or less the same. Dmbx1a was still detected in
the optic tectum and predominantly localized to the medial and lateral compartments of the
tectal region (Figure 2.3I-BB). Scattered cells of the retinal inner nuclear layer also expressed
dmbx1a, but not where the optic nerves exit (Figure 2.3K, P, U, Z). It is likely dmbx1a has a role
in regulating cells in the inner nuclear layer of the retina, possibly bipolar neurons, however
further analysis is needed to confirm their identities. Dmbx1a was also expressed in discrete
36
Figure 2.3 Whole-mount in situ hybridization patterns of dmbx1a.
The expression of dmbx1a was examined from embryonic tailbud stage (10hpf) to 6 dpf larva.
Stages are indicated at bottom right. Dorsal views, anterior is to the top (A, C, F, J, O, T, Y);
Lateral views, anterior to the left (B, E, I, N, S, X). Coronal sections taken where the magenta
lines are indicated (D, G, H, K, L, M, P, Q, R, U, V, W, Z, AA, BB). Dmbx1a transcripts were
mainly found in the posterior forebrain, midbrain (MB) [particularly in the optic tectum (OT)],
and the lateral sides of the hindbrain (HB). dmbx1a is also strongly expressed in the inner
nuclear layer of the retina. For embryos examined at 10hpf, n=50; 1dpf, n=81; 2dpf, n=30; 3dpf,
n=15; 4dpf, n=42; 5dpf, n=22, 6dpf, n=15. Scale bar = 50μm.
37
FIGURE 2.3
38
hindbrain cell populations that appear to demarcate the cerebellar eminentia granularis (EG)
anteriorly and the medulla oblongata posteriorly (Figure 2.3H, M, R, W, BB).
Instead of the annulus expression pattern that dmbx1a had at tailbud stage, dmbx1b was only
weakly detected at tailbud stage (Figure 2.4). However, expression of the dmbx1b gene became
stronger in the midbrain territory around mid-somitogenesis (~13 hpf; data not shown). At 1dpf,
dmbx1b expression was confined to the mesencephalon where dmbx1a was also expressed, but it
never expanded rostrally into the diencephalon (Figure 2.4B). Dorsal and coronal views
revealed dmbx1b expression in both dorsal tectum and ventral tegmentum (Figure 2.4C-D). As
with dmbx1a, there was little or no expression of dmbx1b in the midbrain-hindbrain boundary.
By 2 dpf, dmbx1b expression was localized to the midline as well as to the caudal midbrain, at
the boundary between optic tecta and tegmentum. Intense expression was also detected in the
ventricular zone (Figure 2.4E-G). Similar to dmbx1a in the hindbrain, dmbx1b expression was
primarily in the lateral regions of the rhombic lip (Figure 2.4H). From 3-6 dpf, dmbx1b was
expressed throughout the midbrain, but was only weakly expressed in the retina and hindbrain
when compared to dmbx1a expression. In the optic tectum, dmbx1b paralog localized around the
margin of both tecta and its expression was stronger near the dorsal midline (Figure 2.4I-BB). It
is also worth noting that dmbx1b-expressing cells in the tecta did not seem to overlap with those
that are dmbx1a-positive. This suggested that the dmbx1 paralogs were marking different
subgroups of tectal neurons. It is known that there are various types of neurons in each tectal
neuropil (Nevin et al., 2010). Thus, it will be of interest to identify the particular cell types that
express these two dmbx1 genes, and whether the absence of these paralogs at later stages could
disrupt the development of specific mature tectal neurons. Furthermore, dmbx1b is present
bilaterally near the medial region of the medulla oblongata, which consists of many GABAergic
neurons (Higashijima et al., 2004a; Higashijima et al., 2004b); however, further analysis is
39
Figure 2.4 Whole-mount in situ hybridization patterns of dmbx1b.
The expression of dmbx1b was examined from tailbud stage (10hpf) to 6 days post-fertilization.
Stages indicated at bottom right. Dorsal views, anterior to the top (A, C, F, J, O, T, Y); Lateral
views, anterior to the left (B, E, I, N, S, X). Coronal sections taken where the magenta lines are
indicated (D, G, H, K, L, M, P, Q, R, U, V, W, Z, AA, BB). dmbx1b transcripts were localized
in the midbrain (MB) [particularly in the optic tectum (OT)], and in two small bilateral strips in
the medial region of the hindbrain (HB). Expression of dmbx1b was faintly detected in the inner
nuclear layer of the retina (K, P). For embryos examined at 10hpf, n=59; 1dpf, n=81; 2dpf,
n=42; 3dpf, n=30; 4dpf, n=42; 5dpf, n=17, 6dpf, n=12. Scale bar = 50μm.
40
FIGURE 2.4
41
required to confirm if dmbx1b-expressing cells are GABAergic (Figure 2.4H, M, R, W, Y, BB).
An apparent difference between the expression of dmbx1a and dmbx1b was also found in the
retina. Expression of dmbx1b was only observed in the retina from 3 to 4dpf, and it appeared to
be localized to the amacrine cells in the inner nuclear layer (Figure 2.4K, P). Beyond 5dpf,
expression of dmbx1b was no longer detected in the retina (Figure 2.4U, Z).
Taking into account the relative temporal, spatial and quantitative expression patterns of the
dmbx1 paralogs in the first 6 days of life, my data suggested that there were several notable
differences in the spatial expression patterns of the two dmbx1 genes in zebrafish. Both dmbx1a
and dmbx1b are expressed throughout the optic tecta and medulla oblongata, but the expression
of dmbx1b is consistently lower in the hindbrain. Moreover, both dmbx1a and dmbx1b are
expressed throughout the retinal inner nuclear layer, but dmbx1a appears to be highly expressed
and localized to the apical region of the inner nuclear layer, whereas dmbx1b expression is
relatively weaker throughout the basal region. Overall the in situ hybridization data is consistent
with the RT-PCR results, and the strong tectal expression of dmbx1 paralogs is similar to that of
mouse and chicken dmbx1 homologs. Based on the expression patterns of the dmbx1 genes and
dmbx1 loss-of-function phenotypes from other studies, it was predicted that the two dmbx1
paralogs in zebrafish would predominantly function in midbrain formation, plus play a role in
retinal and hindbrain development during later differentiation. Therefore, I wanted to compare
the functional requirements of dmbx1a and dmbx1b during zebrafish embryogenesis.
2.2.2 Efficacy and dosage requirements for dmbx1a and dmbx1b morpholinos
To investigate the function of the dmbx1 paralogs, specific antisense morpholino
oligonucleotides were used to carry out loss-of-function assays. Given the high degree of
sequence similarity between the dmbx1 paralogs, it was necessary to confirm that the gene MOs
42
could specifically block the translation of the targeted gene. In the absence of commercially
available antibodies to verify the efficacy of the knockdown experiments, an alternative method
was used. In this approach, an in-frame Dmbx1-GFP fusion construct, which included the MO
target sequence, was generated. This construct was used to estimate the level of protein
knockdown for each paralog. Each MOs was co-injected with their corresponding in vitro
transcribed fusion mRNAs (Figure 2.5A and 2.5B). When either dmbx1a-GFP (Figure 2.5C, n =
62) or dmbx1b-GFP (Figure 2.5G, n = 73) fusion mRNA was injected, ~75% of the embryos
had bright ubiquitous GFP protein expression after 24 hpf. GFP expression was completely
suppressed in embryos that were coinjected with MO1a + dmbx1a-GFP (Figure 2.5D, n = 74) or
MO1b + dmbx1b-GFP (Figure 2.5H, n = 59), indicating that the MOs result in very efficient
translation inhibition. In order to control for MO sequence specificity, I also co-injected the GFP
fusion mRNA constructs with a 5-bp mismatched MO (mMO1a or mMO1b) and quantified the
percentage of injected embryos that were GFP positive after 24 hpf. Of the embryos that were
co-injected with mMO1a + dmbx1a-GFP (Figure 2.5E, n = 53) or mMO1b+ dmbx1b-GFP
(Figure 2.5I, n = 64), ~75 - 80% of the embryos demonstrated ubiquitous GFP expression,
which was similar to the percentage of fusion construct injected embryos expressing GFP
without MO co-injection. These results indicate that the knockdown of either GFP fusion
construct depends precisely on the complementary MO sequences. Further examination on the
paralog specific knockdown confirmed that there were no cross-target effects. Co-injection of
MO1b + dmbx1a-GFP (Figure 2.5F, n = 60) or MO1a + dmbx1b-GFP (Figure 2.5J, n = 62)
resulted in ~75 - 80% of the embryos with ubiquitous GFP expression in injected embryos after
24 hpf. Again, the numbers of GFP positive embryos within the test pool were comparable to
the results obtained from injecting the GFP fusion mRNA alone, indicating that it is unlikely
that cross targeting of MOs occurred. These data suggest that the MO knockdown of the
43
Figure 2.5 Specificity of morpholino induced dmbx1a and dmbx1b
knockdown using fusion protein constructs Dmbx1a-GFP and Dmbx1b-GFP.
(A, B) Schematic diagrams showing sequence of fusion constructs: 5' UTR (black horizontal
line); 5' end of the CDS for either dmbx1a or dmbx1b (white box), the full-length coding
sequence for GFP (green box); site of polyadenylation (AAAA); relative position binding sites
(red horizontal line) where the MOs block translation of the fusion proteins. (C-J) dmbx1a- and
dmbx1b-GFP mRNA was either injected alone or co-injected with MOs as indicated at the
bottom left of each panel, and embryos were examined for the presence of GFP fluorescence at
24 hpf (lateral view with anterior to the left). The percentage of GFP positive embryos is shown
on the top right. UTR, untranslated region; CDS, coding sequence; MO, morpholino; mMO,
mismatch morpholino; GFP, green fluorescent protein.
44
Dmbx1-GFP fusion proteins, and by inference the endogenous Dmbx1 proteins, is both efficient
and paralog specific.
A dose-response analysis for morpholino efficacy was performed using foxb1.2 gene expression
as a reliable correlated readout for brain development defects observed at 48 hpf. When 5 ng of
either MO1a or MO1b was used, there was no discernible change in foxb1.2 expression in the
midbrain (or hindbrain) compared to un-injected controls (Figure 2.6A, 2.6B and 2.6D).
However, when 10 ng of either MO1a or MO1b was injected separately (Figure 2.6C and 2.6E),
a reduction in foxb1.2 expression was observed. Combined injection of 5 ng MO1a and 5 ng
MO1b caused a similar reduction in foxb1.2 expression (Figure 2.6F). The same dose response
analysis was also performed, using rhodopsin (rho) gene expression as a reliable correlated
readout for any retinal defects observed at 72 hpf. In contrast to the observations made for
foxb1.2 expression in the midbrain and hindbrain, 5 ng of MO1a (Figure 2.6H) resulted in a
substantial reduction (~50%) in the extent of rho expression in the retina, whereas 5 ng of
MO1b (Figure 2.6J) resulted in a negligible difference in rho expression compared to un-
injected controls (Figure 2.6G). However, when 5 ng each of MO1a and MO1b was combined
for the injection (Figure 2.6L), reduction in rho expression was observed. When 10 ng of MO1a
was injected individually (Figure 2.6I), a similar reduction in rho expression was observed when
compared to the double morphants derived from injections of 5 ng of each morpholino
combined (Figure 2.6L). When 10 ng of MO1b (Figure 2.6K) was injected individually, a
relatively mild reduction in rho expression was observed compared to controls (Figure 2.6G).
These data suggest that dmbx1a has a predominant functional role in retinal differentiation and
that dmbx1b may only have a minor, additive role.
45
Figure 2.6 Dose-dependent changes in foxb1.2 and rhodopsin gene expression
in dmbx1 morphants.
Analysis of gene expression in un-injected (A, G), MO1a-injected (B, C, H, I), MO1b-injected
(D, E, J, K) or MO1a+MO1b-injected (F, L) embryos using the MO concentrations listed. For
foxb1.2 expression at 48hpf, embryos are shown in lateral view with anterior to the left (A-F).
For rhodopsin expression at 72hpf, embryos are shown in ventral view, anterior to the left (G-
L). Control embryos injected with mismatch MOs at similar concentrations showed no change
in expression and are not shown.
46
2.2.3 Midbrain phenotype in dmbx1a- and dmbx1b-morphant embryos
The early onset and sustained expression of dmbx1a and dmbx1b within the midbrain suggested
that the genes may play an important role in the development of this neuroanatomical structure.
Morphological analyses were carried out to compare the MO-injected embryos (10ng) to the
mMO-injected (10ng/embryo each) and un-injected control embryos. After 48 hpf, prominent
differences in the size of the midbrain were detected. The size of the dorsal tectum of the MO1a-
injected embryos was reduced (Figure 2.7C-D) compared to un-injected (Figure 2.7A-B) and
mMO1a-injected embryos (Figure 2.7E-F), consistent with previous results using the same
morpholino (MO1a) (Kawahara et al., 2002). The overall cross-sectional area of the tectal
hemispheres was diminished in the MO1a morphant, but the shape of the tectal hemispheres and
the extent of the cerebellar plate (CeP) remained relatively unaffected. To quantify these
differences, transverse sections were analyzed by measuring the average cross-sectional area of
the tectal wall unilaterally [from the lateral sulcus separating the tectum dorsally from the region
of the torus semicircularis (TS) ventrally]. A reduction of ~ 50% was observed in the MO1a
morphants compared to controls at the same position along the anteroposterior axis (Figure
2.7O). Interestingly, in MO1b morphant embryos, the tectal morphology was also affected
compared to mMO1b control injected embryos (Figure 2.7G-J). Cross-sectional area
measurements revealed a ~ 35% reduction in size in MO1b morphants (Figure 2.7O).
Knockdown of both dmbx1 genes simultaneously (MO1a + MO1b at 5ng each) resulted in an
obvious change in the overall morphology of the tectum (Figure 2.7K-L) compared to the
double control injected (mMO1a + mMO1b) embryos (Figure 2.7M-N). In contrast to the single
morphant phenotype, the shape of the tectal hemispheres in the double morphant often was
abnormal, which can be observed from a dorsal perspective (compare red dotted line in Figure
47
Figure 2.7 Hypoplasia of the optic tectum in dmbx1 morphants at 48 hpf.
Gross morphology of the optic tecta are compared between morphants (C, G, K) and both the
un-injected (A) and mMO-injected (E, I, M) embryos. Dorsal views, anterior to the left. Contour
of the medial-posterior ridge of the optic tectum is demarcated by red dotted lines. The MOs
used in each group are shown on the bottom right of the panel. Arrows in A, C, E, G, I, K, M
represent the relative position where ~ 1 μm plastic sections are obtained as depicted in B, D, F,
H, J, L, N, respectively. Measurements of the cross-sectional tectal area are summarized in the
graph (O). Asterisk indicates significant difference (p < 0.05) between the morphant samples
and the controls. TeO, optic tectum; CeP, cerebellar plate
48
2.7K with red dotted lines in Figure 2.7A, C, G). Despite the more extensive morphological
alterations in the double morphants, the overall cross-sectional area of the tectal hemispheres
were reduced to a similar degree (~ 60%), when compared to the differences observed in the
single morphant analyses (Figure 2.O). Overall the average tectal cross-sectional areas in a
transverse section of the un-injected and mMO1 injected embryos ranged between 5800 - 7600
μm2, whereas the morphant embryos had an average area that ranged between 2800-3900 μm
2,
which were statistically significantly less than their cognate controls, but not significantly
different from each other. The overall growth of the morphant embryos was not significantly
impaired by 48 hpf [body length of WT = 2986 ± 51 μm; MO1a+b = 2910 ± 210 μm;
mMO1a+b = 3018 ± 65 μm (n = 5 per group)], and after normalizing the tectal cross-sectional
area measurements to embryo length, the area in the double morphants was significantly
reduced (p<0.05) compared to un-injected and mMO1a + mMO1b injected embryos, whereas
the two control groups were not significantly different.
2.2.4 Hindbrain abnormality in dmbx1a and dmbx1b morphant embryos
Coronal sections of 48hpf embryos showed that un-injected control (Figure 2.8A) and mMO1a
injected embryos (Figure 2.8C) had thick layer of cells packed in the medulla oblongata,
whereas the single MO1a injected morphants appeared to have fewer cells in this region (Figure
2.8B). A similar phenomenon was observed in MO1b- and mMO1b-injected embryos (Figure
2.8D-E). When both dmbx1a and dmbx1b were knocked down, the tissue thickness of the
medulla oblongata (the granular layer) was significantly decreased when compared to the
controls (Figure 2.8F-G). Moreover, it appeared that there was a gap at the midline between the
two lobes (red arrowhead), which is likely a result of growth reduction in the medulla oblongata.
Size comparison of the medulla oblongata between morphants and control groups was done by
taking sections at the level of the otic vesicles are (with the otolith visible) and determining the
49
Figure 2.8 Morphological defects in medulla oblongata observed in dmbx1
knockdown morphants.
Morpholino-injected embryos (B, D, G) are compared with un-injected embryos (A) and
control-injected embryos (C, E, G) at 48hpf. The morpholinos used in each group are shown on
the bottom left. Arrowhead indicates the furrow that situated above the floor plate and between
the two medulla oblongatas in each of the single knockdown morphant embryos. The drift
between the medulla might be an outcome of inadequate growth in the medulla and the double
morphants have a more severe phenotype. Measurements of the tectal area are summarized in
the graph (H). Asterisk indicates significant difference (p < 0.05) between the double morphants
and the un-injected groups.
50
cross-sectional area of the medulla oblongata. Measurements obtained from un-injected and
mMO groups ranged from 6933-8629 μm2 (average 7666±871 μm
2) and 7492-8320 μm
2
(average 7837±431 μm2) respectively, while the area obtained from the three morphants groups
were 4920 μm2 (for MO1a injected), 5646 μm
2 (for MO1b injected) and 5138 μm
2 (for
MO1a+1b injected) respectively. The quantifications revealed that there were no significant size
differences between single dmbx1 morphants and their respective un-injected or mMO-injected
controls, given the morphological distortion at the midline. However, the double morphants did
show a significant size reduction in the medulla oblongata. This indicates that dmbx1a and
dmbx1b function redundantly in the hindbrain. Since neurogenesis across the rhombomeres and
spinal cord occurs earlier than the expression of dmbx1 paralogs in that region, it is possible that
dmbx1 is not responsible for patterning and growth of the hindbrain. Instead, these transcription
factors may help specify or maintain particular subtypes of neurons. Further analysis is required
to pinpoint which types of cells Dmbx1a and Dmbx1b are regulating in the hindbrain.
2.2.5 Retinal defects in dmbx1a and dmbx1b knock down embryos
In order to determine the relative requirement for Dmbx1a and Dmbx1b in the development of
the retina, transverse sections of the retina at 72 hpf were examined. Single MO1a injected
embryos had severe defects in lamination (Figure 2.9B) compared to the un-injected (Figure
2.9A) or mMO1a injected (Figure 2.9C) embryos. Furthermore, there was a significant decrease
in the average overall area of a midtransverse section of the retina in the dmbx1a morphants
compared to controls (Figure 2.9H). In contrast, the MO1b injected embryos displayed a
relatively mild retinal phenotype (most prominently in the dorsal regions) (figure 2.9D)
compared to the un-injected (Figure 2.9A) and mMO1b injected (Figure 2.9E) controls.
However, similar to the dmbx1a morphants, the dmbx1b morphants exhibited an overall
reduction in the average midtransverse area of the retina (Figure 2.9H). The combined MO1a +
51
Figure 2.9 Retinal hypoplasia in dmbx1 morphants at 72 hpf.
Coronal plastic sections (~ 1 μm) of 72hpf retina were taken from the un-injected (A), MO-
injected (B, D, F), and mMO-injected (C, E, G) embryos. The morpholinos used in each group
are shown on the bottom left. At 72hpf, retina can be clearly distinguished into three layers (A):
retinal ganglion cell layer (RGC), inner nuclear layer (INL), and photoreceptor layer (PR).
Retinae from MO-injected embryos were defective in lens development and failed to
differentiate into proper lamination. Measurements of the retinal cross-sectional area
summarized in the graph (H). Asterisk indicates significant difference (p < 0.05) between the
morphant samples and the controls.
52
MO1b injected embryos resembled the MO1a injected embryos in that the double morphants
had severe defects in lamination (Figure 2.9F) compared to controls (Figure 2.9A, G).
Interestingly, the reduction in the average mid-transverse area in the double morphants was not
significantly different from that observed in the single morphant embryos, ranging from 40 -
60% less than controls (Figure 2.9H).
These observations indicated that retinal growth or retinal cell survival was primarily dependent
on the dmbx1a paralog, which could partially compensate for the lack of dmbx1b, but that
dmbx1b is also required for the proper continued growth or survival of the retina (from
approximately 48 - 72 hpf). The overall growth by 72 hpf [body length of WT = 3302 ± 54 μm;
MO1a+b = 2996 ± 159 μm; mMO1a+b = 3278 ± 65 μm (n = 5 per group)] of the double
morphant embryos was significantly less (p<0.05) than un-injected and mMO1a + mMO1b
injected embryos. However, despite this difference, the retinal sectional area measurements in
the double morphants after normalizing to the embryo length was still significantly reduced
(p<0.05) compared to un-injected as well as the mMO1a + mMO1b injected embryos, whereas
the two control groups were not significantly different. Furthermore, the lens was also
underdeveloped in all morphant groups. However, further experiments are required to
characterize this lens defect in more detail.
2.2.6 Functional divergence of dmbx1 paralogs in brain and retina
The specific loss of foxb1.2 expression in midbrain and hindbrain regions in the double
morphants suggested that the development of specific sub-regions with dmbx1 gene expression
were compromised as a result of gene knock down. However, given that dmbx1a and dmbx1b
have partially non-overlapping expression domains, the expression of foxb1.2 was investigated
in single morphant embryos. Knockdown of either Dmbx1a or Dmbx1b caused a significant
53
reduction in foxb1.2 expression in the midbrain, which was enhanced further in the double
morphants (Figure 2.10A-D, red arrowhead in D). However, in the hindbrain, foxb1.2
expression in the anterolateral domain (presumptive eminentia granularis) was eliminated only
in the presence of MO1a but not MO1b (red arrow in Figure 2.10B, D). Although the trigeminal
ganglion is also present in this anterior-lateral domain, it did not appear to be affected in the
MO1a morphant as indicated by the relatively normal expression of GFP in the isl2b:GFP
transgenic embryos (Figure 2.10 E-H). On the other hand, the ventral diencephalic/hypothalamic
and MHB expression of foxb1.2 was relatively less affected in the morphants. Thus, dmbx1a and
dmbx1b appeared to be independently required for region-specific midbrain and hindbrain
development perhaps through a synergistic mechanism.
Previous analyses demonstrated that knockdown of dmbx1a alone caused defects in retinotectal
projections and reduced terminal fields within the optic tectum (Kawahara et al., 2002). This
observation was confirmed in MO1a injected embryos using a Tg(isl2b:GFP)zc7
transgenic
zebrafish line that robustly marks the RGCs in the retina, their axonal trajectories along the
retinotectal pathway, and the contralateral terminal fields in the tectum (Pittman et al., 2008).
Approximately 75% of MO1a injected embryos (Figure 2.10F, n = 25) showed a defasciculated
optic nerve prior to the chiasm as well as reduced terminal fields in the optic tectum when
compared to control (Figure 2.10E, n=30). In contrast, the retinotectal projection in MO1b
injected embryos appeared normal (Figure 2.10G, n=25), which could be due to the presence of
normal levels of dmbx1a. If so, then a functional role of dmbx1b in RGC development may be
absent. Thus, the retinotectal projection in the double morphants would resemble that of the
MO1a injected embryos. Again in approximately 75% of embryos, a significant defect in the
retinotectal projections when both paralogs are knocked down (n = 35), with significantly
reduced terminal fields in the optic tectum and optic nerve defasciculation (Figure 2.10H).
54
Figure 2.10 Distinct patterns of hindbrain foxb1.2 expression and retinotectal
projections in dmbx1a and dmbx1b morphant embryos.
(A-D) Expression patterns of hindbrain foxb1.2 expression. Lateral view with anterior to the left
of un-injected (A), MO1a- (B), MO1b- (C) or MO1a+MO1b-injected (D) embryos at 48hpf.
The mismatch control injected morpholinos resembled the un-injected controls and are not
shown. Red arrow indicates the missing expression domains of foxb1.2 in the rostrolateral
hindbrain in both MO1a- and MO1a+MO1b animals. (E-H) Retinotectal projections observed in
Tg(Islet2b:GFP) line. Dorsal view anterior to the top of un-injected (E), MO1a injected (F),
MO1b injected (G), and MO1a + MO1b injected (H) embryos. Islet2b:GFP transgenics label
retinal ganglion cells (RGC) and axonal fibers of the optic nerve, trigeminal ganglion, and
Rohon-Beard neurons in green fluorescence. In (H), red arrowheads indicate the terminal field
of the retinal ganglion cells in the optic tectum is missing in the MO1a+MO1b-injected groups.
The terminal field is smaller in MO1a-injected embryos, but overall axon projections seem to be
targeting correctly in all groups. The mismatch control injected morpholinos resembled the un-
injected controls and are not shown.
55
These data suggested that dmbx1a has a predominant role in the development of the retinotectal
projection.
2.2.7 Rescue of dmbx1a and dmbx1b morphant phenotypes with zebrafish and
mouse dmbx1 mRNA
In order to confirm the specificity of the knockdown phenotype, either MO1a or MO1b
morpholino and the corresponding zebrafish full-length mRNAs (lacking the morpholino
binding sequence) were co-injected. Microinjection of either dmbx1a mRNA or dmbx1b mRNA
resulted in a significant dose-dependent dorsalization phenotype that was evident in embryos by
24 hpf. Thus, titration of the appropriate amount of mRNA to use and evaluation of the
percentage of embryos co-injected with optimal amounts of mRNA and morpholino that
demonstrated a rescued morphant phenotype were necessary. The highest concentration of either
dmbx1a or dmbx1b mRNA (250 pg) caused mild to severe dorsalization in ~70 - 75% of the
embryos injected and ≥50% of these were in the severe category (Table 2-2). Because some of
these embryos showed signs of necrosis at 24 hpf, in particular in the tail region, a lower
concentration of mRNA (150 pg for dmbx1a and 188 pg for dmbx1b, yielding similar
phenotypic results) was opted for in order to test whether the paralog specific mRNA could
rescue the morphant phenotype. Using this lower concentration, ~ 50-70% fewer embryos were
severely dorsalized (Table 2-2), so proper phenotypic analyses could be carried out with these
animals.
In a separate set of experiments I wanted to confirm whether zebrafish dmbx1a mRNA could
rescue the zebrafish MO1a morphant phenotype by analyzing foxb1.2 gene expression. The
overall number of embryos examined is summarized in Figure 2.11F, and non-parametric exact
binomial tests were used to evaluate if there were significant differences between the
56
Table 2-2 Concentrations of Dmbx1 mRNAs tested for counteracting the
zebrafish knockdown phenotype.
Embryos were injected with dmbx1 zebrafish mRNA (z1a or z1b), or mouse mRNA (m1).
Amount of mRNA injected to each group is indicated on the left. Co-injection of either 10ng of
dmbx1 morpholinos (MO1a or MO1b) reduced severely dorsalized phenotype caused by medial
doses of zebrafish mRNA (at 150pg z1a and 188pg z1b) but not mouse mRNA (400pg m1).
Results from all groups were tabulated from 2-5 separate experiments.
57
58
proportions of normal and disrupted foxb1.2 staining between any two groups. The majority of
embryos injected with MO1a resulted in a reduction of foxb1.2 expression in the midbrain and
the rostrolateral hindbrain (Figure 2.11B, and 2.11F, p=2.45×10-21
) compared to the un-injected
controls (Figure 2.11A). By co-injecting zebrafish dmbx1a mRNA with MO1a, the expression
of foxb1.2 in the midbrain and rostrolateral hindbrain was rescued in ~55% of the injected
embryos (Figure 2.11C and 2.11F, p=2.56×10-8
). Similar experiments were carried out with
MO1b (Figure 2.11B’ and 2.11F, p=5.07×10-12
) and midbrain expression of foxb1.2 was rescued
in over ~95% of the embryos (Figure 2.11C’ and 2.11F, p=2.1×10-15
). I also tested if the
dmbx1b gene could rescue the dmbx1a morphant phenotype by examining foxb1.2 expression in
the brain, and vice versa. As shown above, foxb1.2 is expressed in the dorsal midbrain and in the
anterolateral hindbrain and this pattern is similar to dmbx1a. In contrast, dmbx1b is not
expressed in this anterolateral hindbrain domain, although it is expressed in the dorsal midbrain.
Knockdown of dmbx1a, but not dmbx1b, results in the loss of foxb1.2 expression in the
anterolateral hindbrain (Figure 2.11B-B’). Therefore, I reasoned that if dmbx1b was capable of
rescuing the dmbx1a morphant phenotype, then the expression of foxb1.2 in the anterolateral
hindbrain would be restored. The results showed that while dmbx1a mRNA could partially
rescue the midbrain of those embryos co-injected with MO1b (Figure 2.11D’ and 2.11F,
p=6.98×10-21
), dmbx1b mRNA co-injected with MO1a was ineffective at restoring foxb1.2
expression in the anterolateral hindbrain (Figure 2.11D), although midbrain expression was
restored to control levels in ~50% of the embryos (Figure 2.11D and 2.11F, p=4.34×10-9
).
Hence, the functionality of the dmbx1 paralogs, in this experimental context, is not completely
interchangeable.
Given the coding-sequence conservation between the mammalian Dmbx1 genes and the teleost
dmbx1 genes, I also examined whether the full-length mouse Dmbx1 mRNA (lacking the
59
Figure 2.11 Dmbx1 morphant rescued with zebrafish or mouse dmbx1 genes.
Foxb1.2 images represent dorsal views, anterior to the top. The expression domain of foxb1.2 in
un-injected 48 hpf embryos (A) is reduced in the MO1a (B) and MO1b injected embryos (B’).
Co-injection of zebrafish dmbx1a (zf1a) and dmbx1b (zf1b) mRNA with their respective
morpholino oligonucleotides rescues these phenotypes (C and C’), whereas co-injection of
mouse Dmbx1 mRNA (mDmbx1) does not (E and E’). Zebrafish dmbx1b can only partially
rescue MO1a morphants’ phenotypes (D), while dmbx1a appeared to demonstrate stronger
rescue effects on MO1b injected embryos (D’). Percent of embryos examined in each group that
process normal (green colour) or defective (magenta colour) foxb1.2 expression pattern is
summarized in graph (F). Exact binomial tests (p<0.05) were performed to see if the extend of
rescue was significant (*) between two groups.
60
morpholino binding site) could rescue the MO1a phenotype. Injection of 400ng mouse Dmbx1
mRNA, which is 2-3 fold more concentrated than the optimized zebrafish dmbx1a and dmbx1b
injection experiment, yielded a comparable dorsalization phenotype (Table 2-2). However, the
mouse Dmbx1 mRNA was unable to rescue the altered foxb1.2 expression in the midbrain and
anterolateral hindbrain in dmbx1a morphant (Figure 2.11E and 2.11F, p=0.303), but did appear
to partially rescue the MO1b embryos (Figure 2.11E’ and 2.11F, p=0.02).
2.3 Discussion
2.3.1 Spatiotemporal expression of zebrafish dmbx1 paralogs is highly
conserved with other vertebrates
Given the DNA sequence similarity in the duplicate Dmbx1 genes and their conservation with
other vertebrate Dmbx1 proteins, it is not surprise that the zebrafish dmbx1 paralogs would have
very similar expression profiles. Indeed, both copies of the dmbx1 gene in zebrafish are
expressed in the caudal forebrain, midbrain, bilateral regions of the hindbrain and inner nuclear
layer of the retina. The spatiotemporal expression profile of dmbx1a and dmbx1b suggested that
they may mediate aspects of early brain development, and in particular the formation of the
mesencephalon. Consistent with the studies in mouse and chick, dmbx1 appears to be a unique
vertebrate invention for specifying midbrain in the CNS (Holland and Takahashi, 2005;
Takahashi, 2005). Recently, it has been reported that dmbx1 expression is also detected in
human midbrain at Carnegie Stage 15 and 19 (Ramos et al., 2007). Together, these data
highlighted the importance of this gene during midbrain development.
In the hindbrain, dmbx1a is expressed at the bilateral sides of the rhombic lip, whereas dmbx1b
is expressed in the medial region close to the midline. Both domains, when combined,
recapitulate the expression profiles of dmbx1 in mouse’s and chicken’s hindbrain. In addition,
zebrafish dmbx1 genes are detected in the retinal regions, which could be a teleost-specific
61
novel expression domain as only one mouse study (without any in depth analysis) has
mentioned a similar observation (Zhang et al., 2002). Furthermore, both dmbx1 genes initiate
expression at the beginning of neurulation in zebrafish, which coincides temporally with other
vertebrates. Overall, the expression domains of dmbx1 paralogs in zebrafish appear to
recapitulate those reported in other vertebrates.
2.3.2 Dmbx1a and Dmbx1b have partially overlapping expression patterns and
functions in the central nervous system
Through the loss-of-function analyses, it is obvious that the dmbx1 paralogs have partially
overlapping as well as distinct functions in the brain. dmbx1a has stronger expression in the
retina and hindbrain when compared to dmbx1b, which may suggest a more important role in
those regions. The functional differences between the two genes in zebrafish may be an outcome
of spatiotemporal divergence in expression patterns between dmbx1a and dmbx1b. Changes
accumulated in the regulatory regions of the paralogs throughout evolution resulted in the
alternations of dmbx1a and dmbx1b expression patterns. However, one cannot eliminate the
possibility that degenerative mutations can also accumulate in the coding region of these two
genes, resulting in distinct biochemical function of each paralog. It will be of interest to
distinguish whether the functional roles of dmbx1 paralogs have diverged due to changes in the
regulatory region or the coding region or both.
Based on the spatiotemporal pattern of dmbx1 gene expression, it is not surprising that
perturbation of dmbx1 results in brain malformation. The effects of gene knock down revealed
that dmbx1a and dmbx1b are independently responsible for tectal and retinal growth. However,
their roles in regulating hindbrain are not as prominent, as size reduction in medulla oblongata
was only observed when both dmbx1 genes were knocked down. It was noticeable that the
double morphant embryos had more severe phenotypes when compared to single morphants. As
62
well, loss of retinal lamination and foxb1.2 positive eminentia granularis neurons were specific
to dmbx1a knock down only. All of these data suggest that the dmbx1 paralogs are only partially
redundant. They have also diverged during the course of evolution to take on separate functional
roles, especially in the retina. Indeed, when dmbx1a was knocked down, the morphants resulted
in more severe phenotypes in those zones compared to dmbx1b deficient embryos. Moreover, a
synergistic effect in the optic tecta was observed when both genes were knocked down, which
revealed that dmbx1a and dmbx1b may regulate different subtypes of tectal neurons. Thus, it is
possible that different expression profiles of the two dmbx1 paralogs may contribute to their
functional divergence.
2.3.3 A limitation in morphant rescue analyses due to the dorsalization
phenotype.
The concentration of injected dmbx1 mRNA at 1-cell stage appears to be proportional to the
number of dorsalized embryos at 24hpf. The relative high level of dmbx1 expression in the
neural tissue compared to the posterior region could also help explain the dorsalized phenotype
at high concentrations of mRNA. Further experiments are required to better understand the
underlying mechanisms that might account for the dorsalization defects. The mRNA-induced
dorsalization phenotype was rescued to a certain extent, when morpholinos were co-injected at
1-cell stage. Co-injection of morpholino (MO1a/MO1b) and zebrafish mRNA was sufficient to
mitigate the dorsalization defects. In contrast, the co-injection of morpholino and mouse mRNA
could not antagonize these early abnormalities to the same extend as zebrafish mRNA,
suggesting that the downstream effects of ectopic mouse mRNAs may be different. In other
words, the dorsalization phenotypes from zebrafish and mouse mRNA injections could be
caused by distinct molecular mechanisms. Overall these data suggested that mouse dmbx1 gene
63
cannot be fully substituted for the zebrafish homologs, but it appears to have a more conserved
role with dmbx1b compared to dmbx1a.
There is a potential possibility that the mouse Dmbx1 proteins is less stable in the zebrafish
embryo, which could account for the reduced potency of mouse Dmbx1 to induce a dorsalized
phenotype. Therefore, to address whether the mouse Dmbx1 protein was less stable, I generated
c-myc epitope-tagged mouse Dmbx1 and zebrafish Dmbx1a constructs and monitored protein
levels by immunohistochemistry using an anti-c-myc antibody at tailbud stage after injecting
into 1-2 cell stage embryos. Compared to un-injected controls, embryos injected with either
mouse myc-Dmbx1 or zebrafish myc-dmbx1a resulted in similar protein expression (data not
shown). These results suggested that there was no overt difference in the mRNA stability of
zebrafish and mouse mRNA that would account for the inability of the mouse Dmbx1 to rescue
midbrain and retinal morphant gene expression phenotypes in dmbx1a morphant embryos.
2.3.4 Conservation of autoregulation in mouse and zebrafish dmbx1 genes
In mouse, mutant Dmbx1lacZ/-
embryos showed neonatal lethality and poor growth rate (Ohtoshi
and Behringer, 2004). In contrast to loss of function observations in zebrafish, mouse Dmbx1 is
not required for early mesencephalic specification or morphogenesis, nor does it have an
obvious role in retinal development (Ohtoshi and Behringer, 2004). It is possible that there are
subtle defects in the mouse null mutant embryos that are reminiscent of the embryonic zebrafish
morphants, but it is evident that the loss of dmbx1 in both species is dispensable for survival
during embryogenesis and gross neural patterning. Using a lacZ cassette to generate an
insertional mutation at the Dmbx1 locus, these Dmbx1lacZ/-
animals have fewer Dmbx1-
expressing cells in cerebellum, medulla oblongata (Ohtoshi and Behringer, 2004; Ohtoshi et al.,
2006). It is possible that this transcription factor is only required for specific subtypes of
64
neurons in those regions (Ohtoshi and Behringer, 2004). Dmbx1lacZ/-
embryos also have
increased expression of Dmbx1 in the inferior colliculus (midbrain), which suggests a potential
autoregulation of Dmbx1 expression (Ohtoshi and Behringer, 2004). Similar to what was
reported in Dmbx1lacZ/-
mutant mice, I also observed an upregulation of dmbx1a and dmbx1b
expressions in the optic tectum and medulla oblongata in the respective dmbx1-deficiency
zebrafish embryos (see appendix). However, this phenomenon was never detected in the retina.
Thus, identifying enhancer elements that are present in both dmbx1a and dmbx1b promoters
may reveal unique midbrain/hindbrain regulatory sequences that are targeted by Dmbx1
transcription factors.
2.3.5 Evolution of dmbx1 protein coding sequences
The morphant rescue results indicated that the putative functional differences between the
zebrafish and mouse Dmbx1 genes may correlate with changes in the protein coding sequences
among the vertebrate Dmbx1 genes family. Some teleosts have undergone an additional round
of genome duplication, which occurred after its divergence from other vertebrates (Woods et al.,
2005). The N-termini and the homeodomains of the vertebrate dmbx1 homologs share high
sequence similarity with one another, and the OAR domain is commonly observed in this class
of homeobox proteins (Brouwer et al., 2003). The apparent structural difference among
members of the Dmbx1 protein family is the region between the homeodomain and the OAR
domain, and this variable area may accumulate degenerate mutations that can lead to changes in
protein functions or interactions.
To evaluate the types of selection pressure exerted on the dmbx1 paralogs, the amino acid
substitution rates within the coding sequences of both Dmbx1a and Dmbx1b in zebrafish were
analyzed (in collaboration with Belinda Chang’s lab). Determining the proportion of non-
65
synonymous to synonymous (dN/dS) changes in the coding sequences of Dmbx1 genes can
estimate the rate of amino acid evolution as well as allow inferences about any changes in the
selective constraints during the evolution of this gene family in vertebrates to be made (Yang
and Nielsen, 1998). A dN/dS value of one implies that there is no selection pressure (or neutral)
on the protein sequence. Anything above one suggests that the protein is under positive
selection, whereas below one indicates stabilizing selection. The results of our molecular
evolutionary analyses of Dmbx1 genes suggest that although evolution conservation can be
observed in this gene family, with overall dN/dS of only 0.036, there was a dramatic change in
selective constraint after the duplication event which gave rise to the dmbx1a and dmbx1b
families in fish. The elevation in dN/dS after this gene duplication in the dmbx1a lineage is
significant (dN/dS = 46.88), suggestive of positive selection; whereas the dmbx1b lineage has a
much lower value (dN/dS = 0.37), which would be consistent with weak stabilizing selection.
However, these preliminary results are based on analyses of a fairly small data set, which is
particularly deficient in basal fish Dmbx1 genes. The inclusion of additional sequence data will
improve the ability of these statistical methods to detect changes in the form and strength of
selection across the Dmbx1 gene family. Although the molecular evolutionary analyses show
evidence of positive selection in the coding region of the duplicated fish sequences, further
studies are required to resolve whether changes at the amino acid level have direct consequences
for new functions.
2.3.6 Potential retention mechanism of duplicated dmbx1 gene pair
Data from the molecular evolution analyses suggest that there was positive selection on both
paralogs after the fish-specific genome duplication event over 300 million years ago, and the
selection was slightly stronger on dmbx1a than dmbx1b. In contrast with mouse Dmbx1,
zebrafish Dmbx1a and Dmbx1b appear to have a critical role during retinal development.
66
Together with the cross-species rescue data which demonstrated that the protein function of
Dmbx1 in mouse and zebrafish is not interchangeable, the data seems to indicate that the
possible retention model of dmbx1 paralogs in zebrafish is due to neofunctionalization.
Although it is tempting to conclude that the dmbx1 genes were retained through the
neofunctionalization model, examining the functional role of this ortholog further across
different species is crucial to support that.
As mentioned above, Dmbx1z/-
mutant mouse embryos do not have any morphological defects in
the brain, but they show fewer tectal neurons (Ohtoshi and Behringer, 2004). These observations
are quite distinct from those obtained in my study. It is possible that mouse Dmbx1 alone has
progressively lost some of its functions, although more evidence is needed from cross-species
examination of other terrestrial organisms such as chicken and frog to further deduce that. In
addition, analyzing the ancestral function of dmbx1 in other non-teleost fish groups, such as
dogfish and bichir, may also provide insights into the possible retention mechanism of dmbx1
paralogs in zebrafish. Furthermore, it will be useful to also look at other teleosts and investigate
whether dmbx1 can perform similar functions in the nervous system.
2.3.7 Functional divergence of the teleostean dmbx1 paralogs compared to
other vertebrates
In Amphioxus and in Ciona, expression of dmbx1 begins at mid-tail bud stage in a pair of motor
ganglion interneurons in the visceral ganglion (which may be equivalent to the hindbrain in
vertebrates) (Stolfi and Levine, 2011; Takahashi and Holland, 2004). These examples from the
basal chordate lineage highlight the fact that the ancestral function of dmbx1 is in the hindbrain,
and the presence of dmbx1 in the midbrain only occurred after the vertebrate diverged. As
mentioned above, the phenotypic defects in Dmbx1-/-
mutant mice are most prominent in the
hindbrain (Ohtoshi and Behringer, 2004). On the other hand, dmbx1 paralogs in the teleost
67
appear to have a pivotal role in maintaining growth in the optic tectum and the retina, but they
are less important for hindbrain development. With an extra copy of dmbx1 that arose in the
teleost lineage, it is possible that the restraint on the original dmbx1 became much more relaxed.
Thus, allowing the expression of dmbx1a in zebrafish to diverge from other vertebrates. The
changes in dmbx1a expression domain may provide an opportunity for this gene to take on a
new function in the retina, and it may have helped retain both genes in the teleost lineage during
evolution.
In zebrafish, the unique expansion of the dmbx1a expression domain to the retina may
contribute to its potential role in coordinating the development of anatomical structures that are
responsible for prey-capture activity. In zebrafish, the retina receives visual cues and sends
information to the optic tectum for information processing (Burrill and Easter, 1994). It has been
reported that this brain region is required for visually mediated prey capture activity in fish
(Gahtan et al., 2005). The prey capture neural circuitry also requires the reticulospinal neurons
to execute quick bending for startle response (Gahtan et al., 2005), particularly the Mauthner
cells in the hindbrain which are known to be involved in fast evoke response (Eaton et al., 2001;
Gahtan and Baier, 2004). Based on the expression domains identified in this chapter, it appears
that Dmbx1a is present in all those brain regions and may be a potential regulator to control the
development of those visuomotor structures for proper visual-based behaviours. Hence, it is
possible that the diverged expression of dmbx1a in the zebrafish retina may lead to this novel
function that has only been described in teleost fish.
68
Chapter 3
Dmbx1 Promotes Cell Cycle Exit in Retinal Progenitor Cells
Part of this chapter is adapted from:
Wong, L., Weadick, C. J., Kuo, C., Chang, B. S. and Tropepe, V. (2010). Duplicate dmbx1
genes regulate progenitor cell cycle and differentiation during zebrafish midbrain and retinal
development. BMC Dev Biol. 10, 100.
In this paper, I examined the loss-of-function retinal phenotypes in dmbx1 morphant embryos.
My work showed that differentiation is compromised in retinal progenitor cells due to cell cycle
defects in dmbx1-deficiency embryos. I performed immunostaining, propidium iodide analysis,
and cumulative BrdU labeling to provide evidence that support this finding.
69
Dmbx1 Promotes Cell Cycle Exit in Retinal Progenitor Cells
3.1 Introduction
Neural progenitor cells proliferate until they reach the appropriate size and subsequently become
post-mitotic and develop into specific neurons or glia. In the retina, both intrinsic factors and
cell-cell signals are required to perform their tasks in a strict and timely manner in order to
ensure proper formation and function of the eye. It has been suggested that the timing of cell
cycle exit plays a critical role in cell type specification, and it is likely that terminal cell cycle
and differentiation may coordinate with one another. However, we still lack a complete
understanding of the regulators that are involved in the transition between progenitor cells and
mature retinal neurons. In this chapter, I want to understand how retinal progenitor cells
determine the timing of their terminal exit from the cell cycle and undergo differentiation by
studying the function of a paired-type transcription factor, Diencephalon/mesencephalon
homeobox1 (Dmbx1).
3.1.1 Retinogenesis in the zebrafish retina
Formation of a functional retina requires precise coordination between cell proliferation and cell
differentiation. The timing of these events will dictate the fates of the cells during retinogenesis
(Stenkamp, 2007). Progenitor cells in the neural retina gradually become post-mitotic and start
to differentiate into various subtypes of neurons (Hu and Easter, 1999; Mueller and Wullimann,
2003). After these differentiated retinal neurons are specified, they begin to arborize to their
corresponding targets to develop a functional neural circuitry (Culverwell and Karlstrom, 2002).
In a mature zebrafish retina, there are six major groups of neurons and one type of glia that are
organized into three separate strata (Fadool and Dowling, 2008) (see Figure 1.1). The most basal
tier is the ganglion cell layer (GCL), containing the retinal ganglion cells (RGCs) and displaced
70
amacrine cells (Becker and Becker, 2007). The middle section is the inner nuclear layer (INL),
which is comprised of amacrine cells, bipolar cells, horizontal cells, and Müller glia
(Connaughton et al., 2004; Pujic and Malicki, 2004). The photoreceptor layer (PL) is at the
apical region of the retina, right adjacent to the RPE. There are two main classes of
photoreceptors. They are the rod and cone photoreceptors, which have rather distinct
phototransduction properties between one another (Raymond et al., 1995; Tsujikawa and
Malicki, 2004). The majority of the retinal cells are differentiated by around 60 hpf, with the
exception of those in the ciliary (or circumferential) marginal zone (CMZ), which is located
around the dorsoventral axis of the lens within the neural retina. CMZ is a neurogenic niche that
continues to produce new retinal cells throughout adulthood (Mueller and Wullimann, 2003;
Stenkamp, 2007), with the exception of the rod photoreceptors which are derived from rod
precursor cells within the INL (Otteson and Hitchcock, 2003; Stenkamp, 2011). In addition to
the retinal stem cells in the CMZ, a subset of Müller glia within the INL are also characterized
as stem cell since they have the ability to generate new neurons in the retina when triggered by
injury (Otteson and Hitchcock, 2003; Raymond et al., 2006; Stenkamp, 2011).
3.1.2 Birth order of retinal neurons and glia in the zebrafish eye
It is known that temporal regulation of cell cycle exit together with the presence of different
intrinsic/extrinsic factors can determine which types of retinal neurons the progenitor cells will
become (Stenkamp, 2007). In mouse, the birth order of retinal neurons is rather distinct and
neurogenesis spans from E11 all the way to P12. The first group of differentiated neurons born
in the retina are the RGCs, followed by horizontal cells, cone photoreceptors, amacrine cells,
rod photoreceptors, bipolar cells, and Müller glia (Donovan and Dyer, 2005; Marquardt, 2003).
Teleost retinae have a similar but less distinct birthdating order. Early post-mitotic cells in the
retina are destined to become RGCs, while the Müller glia are the last cell type to be specified
71
(Fadool and Dowling, 2008). These differentiation events, similar to the neurogenic waves,
progress from ventronasally to dorsotemporally across the neural retina (Schmitt and Dowling,
1996; Schmitt and Dowling, 1999; Stenkamp, 2007). All cells in the retina are still mitotic at
27hpf, but within an hour a small group of cells will exit the cell cycle and become RGCs (Hu
and Easter, 1999; Li et al., 2000). Some later-born retinal neurons such as bipolar cells appear
around 50-60hpf (Passini et al., 1997; Schmitt and Dowling, 1999). The last cell type to arise in
the retina are the Muller glia, which are undetectable until 60hpf (Peterson et al., 2001).
3.1.3 Molecular components involved in cell cycle regulation
Neurogenesis is a tightly regulated process that largely depends on proper cell cycle progression
(Dyer and Cepko, 2001b). The rate of cell division has been well-documented during early
zebrafish development. During early cleavage stages (3 hpf), synchronized divisions of each
blastomere is less than 15 minutes, but the cell division time lengthens during blastula and
gastrula stages (Kane and Kimmel, 1993). By 9 hpf (before neurulation), the cell cycle length is
about 4 hours (Kimmel et al., 1994). Due to the heterogeneous population of neurons across the
central nervous system, it becomes difficult to access the cycling time. Generally the cell cycle
time increases with age during development. Zebrafish retina, on the other hand, has been
reported to have dynamic cell cycle length during neurogenesis. Between 16-24hpf, progenitor
cells in the zebrafish neural retina have low mitotic index (number of mitotic cells/total number
of cells). However, after 24hpf it changes significantly to almost four times higher. This
translates to the fact that the cell cycle time in retinal progenitor cells is around 32-49 hours
before 24hpf and then shortens dramatically to 8-10 hours afterwards (Li et al., 2000). Thus, it is
possible that the rate of neurogenesis may vary according to specific cell type and
developmental stages.
72
A cell cycle consists of four major phases: Gap 1 (G1), synthesis (S), Gap 2 (G2), and mitosis
(M). During S-phase, new genetic material is synthesized by DNA replication and sister
chromatids are then separated at mitosis during cell division. Between these two phases are G1
and G2, which are crucial checkpoints for S- and M-phase correspondingly. During the gap
phases, cells repair any DNA damage or terminate cell cycle progression if necessary. There is
also an additional gap phase called G0. This third gap phase allows cells to remain at a quiescent
state, which means that they are no longer proliferative but are yet to differentiate.
How does a uniform group of retinal progenitor cells differentially regulate the timing of their
terminal cell divisions? There appears to be a fine balance between the levels of cell cycle
activators and inhibitors. Extensive studies on the cell cycle revealed that many different
regulators monitor each step of the cell cycle and control the progressions of these phases. When
the level of activators such as cyclins and cyclin-dependent kinases (cdks) are high, cells tend to
remain in the cell cycle. However, when cell cycle repressors are dominant, then cells become
post-mitotic. When progenitor cells remain proliferative at G1, two members of the cyclin
family – cyclin D and cyclin E, together with their associated cyclin-dependent kinases (CDKs)
will phosphorylate the retinoblastoma protein (Rb) (Dyer and Cepko, 2001a). When Rb is
phosphorylated, it releases the activated E2F-DP complex to drive S-phase until it becomes
inactivated by the phosphorylation of cyclin A-CDK2 (Dyer and Cepko, 2001b; Harbour and
Dean, 2000). Followed by a short G2 driven by cyclin B, cells will progress through mitosis/M-
phase and undergo cell division (John et al., 2001). If the newly divided cells exit the cell cycle
after returning to G1, then the cyclin-CDK phosphorylation activities need to be blocked so that
these progenitor cells can become post-mitotic and differentiate. One of the most intensely
studied negative regulators of the cell cycle is the cyclin-dependent kinase inhibitors (CKIs).
They consist of the INK (such as p15, p16, p18, p19) and the Cip/Kip (such as p21cip1
, p27kip1
,
73
and p57kip2
) family (Bilitou and Ohnuma, 2010). These proteins antagonize the cyclin-CDK
activities; however the molecular mechanisms by which they interfere with the cyclin-CDK
complexes are distinct from one another (Cheng, 2004).
3.1.4 Retinal defects when cell cycle components are disrupted
There are many mechanisms which control whether a cell exits the cell cycle. The presence of
high level cell cycle activators or lack of inhibitors will lengthen cell cycle time or progress into
cell cycle arrest, while reduced amount of cyclin proteins or excess of CKIs will shorten cell
division time or force mitotic cells out of cell cycle prematurely. In Ccnd1-/-
null mutant mice,
retinal progenitor cells have longer cell cycle length but earlier cell cycle withdrawal than
wildtype cells (Das et al., 2009). This results in more RGCs and photoreceptors being produced
at the expense of horizontal and amacrine cells (Das et al., 2009). Although there was a shift in
the proportion of retinal cell types in Ccnd1 mutant embryos, the onset of neurogenesis was
unaltered (Das et al., 2009). In zebrafish, targeted knockdown of ccnd1 hindered retinal and
tectal development but did not affect differentiation, suggesting that ccnd1 only affects tissue
growth but not cell cycle exit (Duffy et al., 2005).
Using a specific morpholino to eliminate p57kip2
(or cdkn1c) in the zebrafish embryos resulted in
reduced differentiation and a smaller retina (Shkumatava and Neumann, 2005). In the cdkn1c
morphants, the number of proliferating cells increased in the retina, but there was also a
significant increase in cell death in the same domain. However, when p21cip1
(cdkn1a) was
downregulated, there was an increase in the number of mitotic events and a decrease in
apoptosis (Liu et al., 2009). It has been shown that p21cip1
, but not p57kip2
, is downstream of the
p53 pathway that facilitates cell cycle arrest at the checkpoint (el-Deiry et al., 1993).
74
Another cell cycle component, Rb, also disrupts cell cycle progression when it is absent. Since
Rb is required for G1-to-S progression during cell division, mutation at the retinoblastoma 1
(rb1) locus in zebrafish causes delayed differentiation of pioneer RGCs, which in turn affected
the cell cycle exit timing of other later born RGCs (Gyda et al., 2012). However, rb1 mutant
embryos (also known as space cadet) do not have any obvious morphological retinal phenotype,
except for optic nerve hypoplasia. Instead the retinotectal axons in the mutant embryos are mis-
projected when innervating the optic tecta, and consequently leads to locomotion and visual
behaviours deficiencies (Gyda et al., 2012; Lorent et al., 2001).
3.1.5 Transcription factors involved in cell cycle regulation in retinal
progenitor cells
Extrinsic, as well as intrinsic factors are known to affect many of these cell cycle regulators. It is
well-established that transcription factors play an important role in coupling the progression of
cell cycle to cell specification in order to direct the multipotent retinal progenitor cell fates.
Several transcription factors have been shown to regulate cell cycle in zebrafish retina. NeuroD
is a basic helix-loop-helix protein that is expressed in amacrine cell and photoreceptor
progenitors (Ochocinska and Hitchcock, 2007). When neuroD is mis-expressed in the retina,
cells exited the cell cycle early due to significant upregulation of cyclin inhibitors
(p27kip1
/cdkn1b and cdkn1c) and mild decrease in cyclins expressions (cyclinD1, cyclinB, and
cyclinE) (Ochocinska and Hitchcock, 2008). On the other hand, morpholino knock down of
neuroD had no effect on cyclin inhibitors but cyclinD1 expression expanded in the CMZ and
remained in those cells at the photoreceptor layer despite the fact that all photoreceptors are
supposed to be post-mitotic by 3dpf (Ochocinska and Hitchcock, 2008). It appeared that this
differentiation defect was specific to photoreceptors and the lamination in the rest of the eye
looked normal (Ochocinska and Hitchcock, 2008).
75
Id2a is another helix-loop-helix protein that controls the transition between S- to M-phase in
retinal progenitor cells. Loss of function of id2a resulted in longer S-phase (longer cell cycle
time) that led to smaller eye size and lack of retinal differentiation (Uribe and Gross, 2010). On
the other hand, id2a caused macrophthalmia by shortening cell cycle to increase number of
mitotic cells in the gain of function assay (Uribe and Gross, 2010). This study demonstrated that
a transcription factor could regulate cell cycle kinetics at a specific phase, although the
molecular mechanism of how id2a interacted with particular cell cycle components requires
further investigation.
3.1.6 The role of dmbx1 in retinal development
As shown in chapter 2, both dmbx1 paralogs displayed a small eye phenotype when knocked
down using morpholinos. It was observed that retinal lamination was severely affected in the
morphants, especially in single dmbx1a and dmbx1a+dmbx1b double knocked down embryos.
Moreover, defects in retinal morphology also disrupted retinotectal projections in these animals.
Based on their spatiotemporal expression, I hypothesize that Dmbx1a and Dmbx1b can affect
development of retinal progenitor cells in zebrafish. The prediction is that these genes can
promote the differentiation of retinal progenitor cells.
In this chapter, I provide evidence that Dmbx1 expression in the retina represses cyclinD1
expression, and allowing retinal progenitor cells to become post-mitotic and undergo
differentiation at the proper time. Without the dmbx1 paralogs, these cells linger in the cell cycle
much longer than they should due to high levels of cyclinD1. Knocking down these two genes
resulted in microphthalmia (possibly due to fewer cell divisions), as well as delayed retinal
differentiation. Dmbx1 is the first paired-type homeodomain transcription factor reported that
can negatively regulate cell cycle progression during retinogenesis in zebrafish.
76
3.2 Results
3.2.1 Retinal differentiation is delayed in dmbx1 double morphants
As reported in the previous chapter, dmbx1a and dmbx1b are required for the development of
the retina. Phenotypic analysis revealed that in the absence of single or both dmbx1 paralogs,
retinal growth was reduced and the typical lamination observed in the retina was missing
(Figure 2.9). Since the combined MO1a + MO1b injected embryos had more severe defects, all
of the retinal studies were carried out using the double morphants. To test whether retinogenesis
was compromised in morphant embryos, several cell-type specific retinal markers were
analyzed using immunohistochemistry and confocal microscopy. Using an isl2b:gfp transgene
to mark retinal ganglion cells (Pittman et al., 2008), I observed a significant reduction, but not a
complete loss, of retinal ganglion cells (GFP-positive cells) in the double morphant embryos
compared to controls (Figure 3.1A, 3.1A’). The expression of Pax6 (Macdonald et al., 1995),
which marks most amacrine cells and a subpopulation of ganglion cells, and PKC (Yazulla and
Studholme, 2001), which is expressed in bipolar neurons, were both significantly reduced
(Figure 3.1B- C, 3.1B’-C’). Consistent with these observations, expression of general markers
for cone photoreceptors (Zpr1) (Larison and Bremiller, 1990) , rod photoreceptors (Zpr3)
(Schmitt and Dowling, 1996), and Müller glia (glutamine synthetase, GS) (Peterson et al., 2001)
were also almost absent in the dmbx1 double morphant embryos (Figure 3.1 D-F, 3.1D’-F’).
Thus, with the exception of a small group of differentiated ganglion cells and amacrine cells, the
majority of the cells in the dmbx1 double morphant retina were not differentiated at 72 hpf. If
the cells in the neural retina were undifferentiated in the morphants, then it is possible that they
remained as progenitors. Since the lack of stratification in the retina of the double morphants
resembled the control at an earlier stage, I examined progenitor cell markers in the retina. Using
in situ hybridization, a relative increase in progenitor cell marker expression at 72hpf was
77
Figure 3.1 Several major retinal cell types are absent in dmbx1-deficiency
embryos.
Differential retinal markers were examined through immunostaining on coronal section of 72hpf
embryos (n=6 for each group). Antibodies used are listed at the bottom right corner. Expression
pattern of these retinal markers were compared among the controls (A-F) and MO1a+1b-
injected (A’-F’) embryos. It appears that the staining of Isl2b (A’) and Pax6 (B’) are reduced in
dmbx1 double morphants (A’ B’) compared to controls (A-B). Others markers for bipolar cells
(PKC), photoreceptors (Zpr1, Zpr3), and Muller glia (GS) are undetectable in dmbx1 double
morphants (C’-F’).
78
detected in animals with reduced levels of the dmbx1 genes. In control embryos, otx2 was
mostly absent from the RGC layer, ONL and the CMZ, but was expressed in the central region
of the INL at 72 hpf (Figure 3.2A) (Shen and Raymond, 2004). In the double morphant
embryos, expression of otx2 appeared expanded and relatively uniform throughout the central
retina, except in the CMZ (Figure 3.2A’). In addition, a significant expansion of vsx2-expressing
stem and progenitor cells in the CMZ was observed (Vitorino et al., 2009) compared to controls
(Figure 3.2B-B”). There was also an increase in neurod-expressing cells (Figure 3.2C-C’),
which is a marker of photoreceptor progenitor cells (Ochocinska and Hitchcock, 2007).
Interestingly, expression of pax2a (Macdonald et al., 1995) and fgf8 (Walshe and Mason, 2003)
within the optic stalk region was expanded (Figure 3.2D-E, 3.2D’-E’) in morphant embryos. It is
known that there is a defined interface between the retina and the optic stalk (Macdonald et al.,
1995). In the absence of dmbx1a and dmbx1b, the optic stalk expanded as the retina became
smaller. Together these data revealed that reduced levels of Dmbx1 resulted in a persistent
progenitor identity in cells throughout the retina, which correlates with the histological analyses
in the previous data chapter.
3.2.2 Reduced eye size and retinal differentiation is not caused by pervasive
cell death
Data from the dmbx1 double morphants demonstrated that the size of the retina is reduced and
that neural differentiation is significantly impaired. One possible mechanism to account for this
phenotype is increased cell death in retinal progenitor cells, resulting in diminished growth and
lack of differentiation potential in the eye. To test this hypothesis, the level of cell death was
examined using various cell death markers. At 72 hpf, there is substantial remodeling occurring
in the teleostean retinotectal pathway, which results in apoptotic cell death when an erroneous
connection is made (Candal et al., 2005). Thus, a basal level of cell death was expected in
79
Figure 3.2 Retinal progenitor genes and optic stalk genes are expanded in
dmbx1 knocked down embryos
(A-C) Coronal sections of 72hpf retina are examined with various markers, which are shown on
bottom left corners. Number of embryos examined is on the bottom right. Expression patterns of
these markers are compared among the control-injected (A-C) and MO-injected (A’-C’)
embryos through in situ hybridization. Retinal progenitor markers (otx2, vsx2, neuronD) had all
increase their expressions spatially in the central retina. (D-E) Whole mount in situ
hybridization of optic stalk markers. Genes express in the optic stalk has increased level of
expression in the MO-injected (D’, E’) embryos compared to the control-injected (D, E) group.
Red arrow points to the expanded domain of each marker.
80
control-injected embryos. Cell death was measured by three different assays. First, using the
chromatin binding fluorescent marker acridine orange (AO) to label dying cells on live embryos,
it was found that combined MO1a + MO1b injected embryos (Figure 3.3A’) showed less AO
labeling in the retina compared to controls (Figure 3.3A). The same result was observed by
performing TUNEL labeling and Caspase 3 antibody staining at 72 hpf to detect apoptotic
nuclei. The number of apoptotic cells in a transverse section of the retina appeared to be slightly
more in the controls (Figure 3.3B-C) than in the double morphants (Figure 3.3B’-C’). The
number of Caspase 3-positive cells per section was not significantly different between control-
injected (~4 cells/section) and MO-injected (~3 cells/section) embryos (Figure 3.3D, p<0.05).
However, other types of cell death such as necrosis and autophagy currently cannot be ruled out.
Overall, these data suggest that pervasive cell death likely does not account for the defects in
size and differentiation of the retina in dmbx1 morphants.
3.2.3 Cell cycle defects in dmbx1 double morphants
An alternative mechanism that could contribute to the reduced size and attenuated
differentiation of the retina may involve changes in the proliferative capacity of progenitor cells.
To address this, proliferating cell nuclear antigen (PCNA) protein expression by
immunohistochemistry was performed to label cells that are actively in cell cycle (Wullimann
and Knipp, 2000). By 72 hpf, cell proliferation normally becomes substantially restricted to the
CMZ (Marcus et al., 1999). Thus, a defect in cell cycle regulation would be most evident by 72
hpf when neurogenic compartments are relatively small and very well circumscribed under
normal conditions. PCNA labeling at 72 hpf revealed that cell proliferation in the double
morphants was greatly expanded from the CMZ towards the central retina (Figure 3.4A’),
compared to the relatively few PCNA-positive cells in the CMZ of control retinas (Figure 3.4A).
This observation was confirmed with another cell cycle marker, 5-bromo-2’-deoxyuridine
81
Figure 3.3 No increase in cell death in dmbx1 morphants.
Embryos at 72hpf were examined with different cell death assays. Acridine orange staining (A,
n=10 in each group) and TUNEL-labelled (B, n=3 in each group) and Caspase 3
immunohistochemistry (C, n=12 in each group) all mark apoptotic cells. Wholemount embryos
in lateral views, anterior are to the left (A). Coronal sections are stained with TUNEL (B) and
Caspase 3 (C), as indicated at the bottom right. There is a base level of cell death present in the
retina from the control embryos. Yellow arrowhead indicates cells undergoing apoptosis. The
amount of cell death between control- and MO-injected embryos is not significantly different
(D).
82
Figure 3.4 Cell proliferation assays in the retina of 72hpf embryos.
Injection group is indicated on the left. Markers are shown at the bottom right. Coronal section
of retina (A-C) were examined with PCNA (A-A’, n=9 in each group), BrdU (B-B’, n=6 in each
group) and PHH3 (C-C’, n=9 in each group). Proliferating cells can only be found in the ciliary
marginal zone (white arrowhead) in control-injected embryos (A-C), but morphant embryos
(A’-C’) have many more that are presented in the central retina (white parenthesis). Number of
PHH3-positive cells is significantly higher in the dmbx1 morphant group (p<0.05). The presence
of S- and M-phase markers in the morphants suggests that these cells can progress through
mitosis and they are not arrested at a particular phase in the cell cycle.
83
(BrdU), a thymidine analog that is incorporated into newly synthesized DNA during S phase of
the cell cycle. A significant increase in BrdU labeled cells was observed in double morphants
(Figure 3.4B’) compared to controls (Figure 3.4B). Given that the retina is smaller in the
morphants, an increase in PCNA- and BrdU-positive cells suggests that cells might be delayed
or stalled in the G1/S transition of the cell cycle. To address this, the expression of phospho-
histone H3 (PHH3), which labels cells in M-phase of the cell cycle, was examined in transverse
sections (Figure 3.4C, C’). Phospho-histone expression was shown to be increased by 3-fold in
MO-injected embryos when compared to controls (Figure 3.4D). In summary, the progenitor
cells are able to progress beyond S-phase of the cell cycle in the dmbx1 double morphants.
3.2.4 Retinal progenitor cells in dmbx1 morphants undergo complete mitosis
Propidium iodide assay was carried out to determine whether retinal progenitor cells were
undergoing endoreduplication of their DNA. The average DNA content of a cell can be
determined after staining it with propidium iodide. Cells at G1 (2n) are detected at a wavelength
that is half of what G2/M phase (4n) cells would be. Using flow cytometry, a large population of
retinal cells can be systematically categorized into each phase of the cell cycle for more detailed
analyses.
Embryos at 72 hpf were compared between MO1a + MO1b- injected and un-injected control
embryos of equivalent age. Dissected retinal tissues from 120 embryos in both groups were
pooled for this analysis. In a pool of un-injected retinal cells, 89% of them were in G1 of the cell
cycle, whereas only 6% were in S phase (Figure 3.5A). In contrast, 61% of the morpholino
injected cells were in G1 phase of the cell cycle and 27% were in S phase (Figure 3.5B). The
proportion of cells in G2-M of the cell cycle in both groups was <1% and this is due to the fact
that the overall fraction of the cell population captured in these short phases of the cell cycle is
84
Figure 3.5 Using propidium iodide analysis to examine cell cycle progression.
There is a significant increase of cells remained in S-phase (from 5.7% to 27%) between un-
injected (A) and MO1a+MO1b-injected (B) embryos. The graphs show the number of G1 cells
(green peak) and S-phase cells (yellow small peak, indicated with blue arrow. The propidium
iodide analysis showed that MO-injected embryos (B) have fewer G1 cells (green peak) and
more S-phase cells (yellow small peak, indicated with blue arrow) compared to control embryos
(A) but these cells still progress through mitosis as there was higher percentage of the G2-M
cells in the morphant groups.
85
rather low using this method. Nonetheless, there was a ~7 fold increase in the proportion of cells
in G2-M in the morphant retinas. Importantly, there was no evidence of polyploidy in either
control or morphant samples. Therefore, these data indicate that retinal progenitor cells in
morphant embryos are not stalled in any particular phase of the cell cycle and that they complete
mitosis. Taken together, the smaller retinal size and increased proportion of cells that remain in
cycle at 72 hpf suggests that progenitor cell cycle length is significantly increased.
3.2.5 Dmbx1 paralogs regulate cell cycle kinetics in retinal progenitor cells
To quantify potential changes in cell cycle length in the dmbx1 double morphants at 72 hpf, a
BrdU cumulative labeling experiment focusing on the retina was performed. A 5 mM pulse of
BrdU was given by intracerebroventricular injection into embryos at 72hpf (n=12 at each time
point) with chase times ranging from 0.5 hours to 10.5 hours at 2-hour intervals. For example,
embryos in the 0.5 hour group received a single BrdU injection and were processed for
immunolabeling after 30 min. In contrast, embryos in the 10.5 hour group received a total of 6
separate injections (2 hours apart) and were processed for immunolabeling 30 min after the last
injection (Figure 3.6A). The central assumption in this analysis is that an asynchronous
population of cells exhibits single population kinetics (i.e. all cells in the population have the
same cell cycle time) (Morshead and van der Kooy, 1992). Representative confocal images of
BrdU labeled cells in control retinas at 0.5 hr, 6.5 hr, and 10.5 hr are shown in Figure 3.6 B-D
where proliferating cells are exclusively confined to the CMZ. In contrast, in dmbx1 double
morphant retinas, BrdU-positive cells appear scattered throughout the peripheral and central
retina (Figure 3.6B’-D’).
Proliferating cells entering S-phase over time will incorporate BrdU and become labeled until
they re-enter S-phase, at which point they can still incorporate BrdU, but they will not be
86
Figure 3.6 Cumulative BrdU analysis was used to examine whether the cell
cycle length has increased in the retinal progenitor cells of the morphants.
For cumulative cell cycle analysis using BrdU incorporation, embryos (n=12) were injected
every two hours up to 10 hours followed by 4% paraformaldehyde fixation 30 minutes post-
BrdU injection (A). Proliferating cells (red circles) entering S-phase over time will eventually
incorporate BrdU and become labelled (solid red circles). During the 10-hours cumulative BrdU
assay, proliferating cells (BrdU-positive) are found only in the CMZ region of the controls (B-
D) at all time points. In the morphants (B’-D’), there are many more BrdU positive cells and
they spread across the centre of the retina in the inner and outer nuclear layer. The time points at
which these samples are taken from are listed at the bottom left.
87
marked as newly positive cells. In controls, the data during this interval (~0 – 5.5 hr) were fit to
a linear regression model (R2 = 0.93815), which allowed us to estimate when the maximal
number of BrdU-positive cells in the population that were labeled (the first time point when the
plateau is reached) (Figure 3.7A). Thus, by ~ 5.5 hours of cumulative BrdU labeling, all of the
cells that are cycling in the population (the growth fraction) are labeled and further
incorporation of BrdU at later time points does not increase this value. This allowed us to
estimate the growth fraction in the controls to be ~11%. Using these values obtained from the
plotted data (Figure 3.7A), the progenitor cell cycle was estimated to be ~ 10.5 hours in the
control retina (see Material and Methods for calculation).
The same analysis for dmbx1 double morphant retinas resulted in a significantly different cell
cycle estimate. First, the fraction of cells incorporating BrdU over time continued to increase
over the entire labeling interval (R2 = 0.99055; Figure 3.7B). Therefore, I was unable to
accurately determine the growth fraction for the 72 hpf morphant retinae, which would have
required continuing the cumulative BrdU labeling well beyond 10 hours. However, I reasoned
that the last time point assayed (10.5 hours) could be used as a minimum estimate for the time at
which the growth fraction (i.e. ~35%) is reached (Figure 3.7B). Therefore, a minimal estimate
for the cell cycle in these morphant progenitor cells is ~50.6 hours, which is approximately 5-
fold longer than in the control retina. This increase in cell cycle length could account for the fact
that the size of the retina at 72hpf is significantly smaller since on average progenitor cells in the
morphant retinas would not have completed a cell division between 48hpf and 72hpf. Our data
indicate that a reduction in Dmbx1 proteins causes an increase in the cell cycle time of
progenitor cells in the retina resulting in fewer differentiated cells.
88
Figure 3.7 Using cumulative BrdU labelling analysis to determine cell cycle
kinetics in control and morphant embryos at 72hpf
Over a 10-hour time course, three sections were analyzed at each time point from the control
and morphant group respectively and the numbers of BrdU positive cells from each section were
counted. Cell density (number of DAPI positive nuclei/area of section) was used to estimate the
number of total nuclei from each section and calculated the labelling index (BrdU positive
cells/total nuclei) at all six time points. Cell cycle kinetics in control and morphant embryos was
determined by plotting hours of BrdU injection (T) against the labelling index (LI). Growth
fraction (maximum LI on the y-axis, LIm) can be determined from where the curve plateaus off.
The time when the maximum amount of BrdU positive cells was labelled is equal to total cell
cycle time (Tc) minus S-phase time (Ts). By extrapolating the curve back to time = 0, the
labelling index at Ts (LI0) can be determined. With those information, the total cell cycle time
can be estimated using the equation LI0/LIm = Ts/Tc. Total cell cycle time in the control
embryos was calculated to be 10.5 hours with a maximum of 11% growth fraction (A). Total
cell cycle time and growth fraction cannot be accurately determined in the morphants because
not all S-phase cells were captured with the 10 hours time frame (B). With the assumption that
the graph plateau off at 10.5 hour, the minimum total cell cycle time and growth fraction were
estimated to be 50.6 hours and 35% respectively. This analysis reveals that the cell cycle time of
retinal progenitors is 5 times longer due to the deficiency of dmbx1a and dmbx1b.
89
FIGURE 3.7
90
3.2.6 Delayed retinal differentiation in dmbx1 morphants
To determine if dmbx1 paralogs are responsible for retinal differentiation in addition to
regulating the correct timing of cell cycle exit, double morphant embryos at 4dpf and 5dpf were
examined. Using Pax6 and Zpr1 antibodies (which label RGCs, amacrine cells and
photoreceptors), differentiated retinal neurons were detected in the central retina of both control-
injected (Figure 3.8A-B) and MO-injected (Figure 3.8A’-B’) embryos. This supported the
hypothesis that progenitor cells were progressing through cell cycle much more slowly without
dmbx1 paralogs, and hence, differentiation was severely delayed. However, it was also possible
that the knocked down effect of dmbx1 morpholinos had started to subside. Supposing that
dmbx1 morpholinos remained effective at 4-5dpf, this data demonstrated that dmbx1a and
dmbx1b were not involved in retinal differentiation, at least not for those three retinal cell types
tested. Although more in depth analysis with other retinal makers, such as PKC for bipolar
neurons, is required before definitive conclusions can be reached. Since dmbx1a is strongly
expressed in the INL, where most of the bipolar cells are located, it is possible that dmbx1a has
a role in specifying or maintaining certain subtypes of bipolar neurons. More data is needed to
determine whether the functional role of dmbx1 paralogs is solely to regulate cell cycle or if they
may be able to couple cell cycle progression to neuronal specification.
3.2.7 Cell autonomy of dmbx1
Given that dmbx1 genes encode transcription factors, it is likely that they behave cell
autonomously. However, recent studies reveal that some known transcription factors can
function non-cell-autonomously, such as Irx1a (Cheng et al., 2006), Pax6 (Lesaffre et al., 2007)
and Engrailed (Fuchs et al., 2012). Therefore, assessing whether dmbx1 functions cell-
autonomously will help us to clarify its cellular function.
91
Figure 3.8 Differentiated retinal markers are observed in 4dpf and 5dpf
dmbx1 morphant embryos.
Pax6 (RGCs and amacrine cells) and Zpr1 (cone photoreceptor) are both examined in un-
injected (A, B) and dmbx1-deficiency embryos (A’, B’) at 4dpf (A-A’) and 5dpf (B-B’)
respectively. Comparing the coronal sections of both groups (n=3 in each group) shows that the
size of the retina appears to remain smaller than the control-injected animals but mature retinal
makers are detectable in dmbx1 morphants.
92
To address whether dmbx1 behaves cell autonomously, cell transplantation experiments were
carried out at the blastula stage to study how the wildtype and the morphant clones developed
under different circumstances. Labeled donor cells were placed at the animal pole of the host
embryo as fate-mapping studies have shown that this region will give rise to the eyes, in
addition to other anterior brain structures (Kimmel et al., 1990). The first set of genetic mosaic
experiments were performed by injecting the dmbx1a and dmbx1b morpholinos plus rhodamine
dextran dye into Tg(HuC:Kaede) embryos at the 1-cell stage, then transplanting some of the
donor cells into host embryos at mid-blastula stage (Figure 3.9A). HuC is a pan-neuronal
marker, and the HuC:Kaede transgenic line labels neurons with fluorescent protein Kaede (Sato
et al., 2006). If donor cells from the trangenics were able to differentiate into neurons, they
would fluoresce green. Our results showed that most donor cells from MO-injected embryos
remained undifferentiated in the MO-injected group (Figure 3.9E), except for a few ganglion
cells that managed to mature (similar to the Isl2b expression) (Figure 3.9E’). In the case with
un-injected wildtype hosts, MO1a+MO1b donor cells were unable to differentiate, suggesting
that these morphant cells were behaving cell autonomously in a wildtype environment (Figure
3.9D). On the other hand, transplanted control-injected cells were able to differentiate normally
in un-injected or MO1a+1b knockdown hosts based on the presence of GFP+ neurons (Figure
3.9B, C). Thus, these experimental outcomes again confirmed that dmbx1 genes act cell-
autonomously in the retina.
Due to the relatively limited expression of HuC in the retina (only in the RGCs and amacrine
cells) (Sato et al, 2006), a second transgenic line was used as a source for donor cells. Donor
cells from Tg(β-actin:mGFP) transgenic embryos are easily distinguished from the host cells
since they express membrane-bound GFP. After transplantation at the blastula stage, embryos
93
Figure 3.9 Cell transplantation experiment showed that dmbx1 paralogs act
cell autonomously in the RGC.
Control-injected or MO-injected donor Tg(HuC:Kaede) embryos are labeled with rhodamine at
one cell stage, and donor cells are transplanted to un-injected or MO-injected AB hosts between
3-4hpf (A). Only hosts that have rhodamine positive cells in the retina were examined.
Transplanted cells from control-injected donor are able to differentiate into RGCs normally in
both un-injected (B) and MO-injected (C) hosts (white arrows), whereas MO-injected donor
cells cannot differentiate properly in either the un-injected, or MO-injected hosts (D, E). Some
of the donor cells become neurons in the morphant hosts (E’), and they are potentially the
pioneer retinal ganglion cells.
94
Figure 3.10 Transplantation experiments demonstrated that dmbx1 functions
cell autonomously in the developing retina.
Coronal sections of 72hpf host embryos transplanted with β-actin:GFP positive donor cells at
blastula stage. Cells from un-injected (UN) or morpholino-injected (MO) Tg(β-actin:GFP)
donor embryos were transplanted to either UN- or MO host embryos. Animals with mosaic
retinae were selected for further analysis using immunohistochemistry: (A) Pax6 (RGCs and
amacrine cells) and Zpr1 (cone photoreceptor); (B) PKC (bipolar cells) and Zpr3 (rod
photoreceptor); and (C) Bromodeoxyuridine [BrdU] (S-phase) and phospho-Histone H3 [PHH3]
(M-phase). Images were captured using confocal microscope without any maximum projection.
Numbers of PHH3 positive cells within those transplanted cells in the retina of the host embryos
are summarized in graph (D).
95
FIGURE 3.10
96
were allowed to develop until 72hpf for analysis. Immunohistochemistry with various retinal
differentiated cell type markers (Pax6, PKC, Zpr1, and Zpr3) were performed on the
transplanted embryos, and the data suggested that dmbx1 acted cell-autonomously (Figure 3.10).
When un-injected donor cells were transplanted to a morphant host, the cells expressed all these
tested markers normally (Figure 3.10A-B). On the other hand, when MO-injected donor cells
were transplanted to a wildtype background, they failed to express this set of retinal markers
(Figure 3.10A-B). I showed previously that the expression of differentiated retinal markers were
compromised in the morphant embryos due to the fact that these retinal progenitor cells were
remained in the cell cycle, and thus were labeled for PHH3 (M-phase markers) . To further
illustrate that Dmbx1 acted cell-autonomously, I looked for these cell cycle markers in the
transplanted embryos. As expected, only cells from the morphant donors were found to be
PHH3 positive in the central retina where neurons are post-mitotic by 72hpf in the wildtype
(Figure 3.10C-D). The results from all three sets of transplantation experiments were consistent
with a cell autonomous role for Dmbx1 in the retina.
3.2.8 Identify potential cell cycle components controlled by Dmbx1
I previously showed that retinal progenitor cells in MO1a+MO1b injected embryos had
lengthened cell cycles so only a small portion of neurons differentiated into ganglion and
amacrine cells at 72hpf. To understand how dmbx1 paralogs can regulate cell cycle exit, several
major components of the cell cycle machinery were examined to see if they were affected in
dmbx1 morphant embryos. Two cyclins (cyclin D1 and cyclin E2) and two CKIs (p27kip1
and
p57kip2
) were chosen for further analysis based on their spatiotemporal expressions in the retina.
Wholemount in situ hybridization experiments were performed with these four probes on
control and morphant embryos, and there were obvious differences in the expression of cyclin
D1 (ccnd1) and p57kip2
(cdkn1c) between the two groups (Figure 3.11). At 2dpf and 3dpf, ccnd1
97
Figure 3.11 Cryosection on wholemount in situ hybridization embryos with
various cell cycle markers at 2-3 dpf.
All sections are in coronal view, development stages and probe used in each case is listed on the
left and on the top respectively. Number of embryos examined is listed at the bottom right.
Coronal retinal sections of ccnd1 (A-B and A’-B’), ccne2 (C-D and C’-D’), cdkn1b (E-F, and
E’-F’) and cdkn1c (G-H, G’-H’) are all taken from wholemount in situ embryos. Only
expression of ccnd1 has expanded its expression in the dmbx1 morphants (A’-B’) and cdkn1c
expression came up slower in dmbx1-deficiency embryos (G’-H’) when compare to un-injected
embryos (A-B, G-H)
98
expression is normally found near the CMZ region where most of the cycling cells are (Figure
3.11A-B). However, in dmbx1 MO-injected embryos, ccnd1 expression had expanded from the
CMZ all the way to the central retina and the level seemed to be more intense (Figure 3.11A’-
B’). On the other hand, cdkn1c expression was temporally delayed in the morphants (Figure
3.11G’-H’) when compared to the controls (Figure 3.11G-H) without any noticeable spatial
distortion. For the other two cell cycle markers, there were no obvious changes in their
expression in double morphants (Figure 3.11C-F, 3.11C’-F’). These findings suggest that there
was not an alternation in pan-cell cycle gene expression, but instead Dmbx1a and Dmbx1b had a
specific effect on the transcriptional regulation of ccnd1 and cdkn1c. It appeared that ccnd1
expression was not only upregulated in retinal progenitor cells but also in those progenitors that
were in the medial and lateral proliferative regions of the optic tecta (see chapter 4).
3.2.9 Dmbx1 paralogs regulate cell cycle progression through cyclin D1
To verify whether Dmbx1 can control ccnd1 transcription in the retina, my first approach was to
overexpress dmbx1a and dmbx1b by mRNA injection at 1-cell stage to look for any changes in
ccnd1 expression by in situ hybridization. My hypothesis was that if Dmbx1 transcription
factors negatively regulate the expression of ccnd1, then misexpression of Dmbx1 paralogs
should lead to repression of ccnd1. Similar to the phenotype reported above, knocking down
dmbx1a and dmbx1b led to significant increase in ccnd1 expression in the central retina (Figure
3.12B, 3.12E) compared to un-injected controls where ccnd1 can only be found at the CMZ
(Figure 3.12A, 3.12D). On the contrary, the level of ccnd1 was reduced at the CMZ in embryos
over-expressing Dmbx1 paralogs (Figure 3.12C, 3.12F). However, it was also observed that in
MO-ccnd1 embryos, there was no change in dmbx1a expression at 3dpf (Figure 3.12G-H),
suggesting that Dmbx1 is upstream of ccnd1. Since it is known that the functions of ccnd1 is to
99
Figure 3.12 Wholemount in situ hybridization of cyclinD1 (ccnd1) expression
changes when dmbx1 is perturbed
Between 48 and 72hpf, ccnd1 expression was detected in the CMZ region in control embryos
(A, D) but in dmbx1 MO-injected embryos, expression had expanded from the CMZ to the
central retina (B, E). When dmbx1 was ectopically expressed, there was a strong reduction of
ccnd1 expression around the CMZ region (C, F). Knock down of ccnd1 did not affect the
expression of dmbx1a in the retina (G, H).
100
promote mitotic cells to re-enter the cell cycle (Coqueret, 2002), the potential role of these two
dmbx1 genes is likely to downregulate the expression of ccnd1 in the retinal progenitor cells so
they can undergo cell cycle exit and differentiation at the appropriate time.
3.2.10 Cyclin D1 knockdown can rescue the differentiation defects in dmbx1
double morphants
Another approach to investigate how ccnd1 works with Dmbx1 in a molecular pathway is to see
if knocking down the increased level of ccnd1 with morpholino (MO-ccnd1) in the dmbx1
morphant embryos (MO1a+1b) can allow the progenitor cells to exit cell cycle and begin to
differentiate. In other words: could knock down of ccnd1 in dmbx1 morphant embryos rescue
the cell cycle defect? Differentiated retinal markers such as Zpr1 and Calbindin were chosen to
examine mature cone photoreceptors, RGCs and amacrine cells. When MO-ccnd1 alone was
injected (Figure 3.13B, 3.13F), the embryos had small eyes but overall cell specification was
unchanged, similar to what was previous reported (Duffy et al., 2005). On the other hand,
differentiation of retinal neurons was compromised in MO1a+MO1b embryos (Figure 3.13C,
3.13G). When the level of ccnd1 was brought down using MO-ccnd1 in the MO1a+1b
morphants (Figure 3.13D, 3.13H), some of these retinal progenitor cells were able to withdraw
from the cell cycle in a timely manner and become specified into various mature retinal neurons.
However, only 4 out of the 6 triple morphant embryos examined had differentiated retinal
markers present. The number of Calbindin-positive cells in the amacrine cell layer recovered in
the triple morphants (Figure 3.13I) suggested that knocking down ccnd1 cannot completely
rescue the dmbx1 morphants. Immunostaining with Zpr1 and Calbindin on these single, double,
and triple morphants confirmed that Dmbx1 is required to downregulate the expression of
ccnd1, which allows the retinal progenitor cells to exit their cell cycles. However, the results
also showed that the cell cycle defects in those triple morphants were not fully rescued, which
101
Figure 3.13 Repression of cyclinD1 partially rescued dmbx1 morphants
phenotype.
Coronal cryosections of 72hpf retina (A-H). Retinal differentiation in the central retina can by
seen using immunohistochemistry with Zpr1 (labels photoreceptors; n=6 in each group) and
Calbindin (labels retinal ganglion + amacrine cells; n=3 in each group) markers. Both un-
injected (A, E) and MOccnd1 (B, F) embryos express differentiation retinal markers. No
differentiation markers are shown in dmbx1 morphants (MO1a+MO1b) (C, G). When dmbx1
morphants are knocked down with cyclin D1 morpholino (MO1a+MO1b+MOccnd1) (D, H),
cells in the central retina are able to differentiate into specific cell types normally. Arrows point
to the differentiation markers that are present in the retina. Numbers of Calbindin-positive cells
observed in the amacrine cell layer per retinal section from each injection groups are
summarized in graph (I), asterisks represent significant difference in the number of Calbindin-
positive cells between two injection groups (p<0.05)
102
indicated that Dmbx1 may interact with other cell cycle components in addition to ccnd1 to
regulate the process of cell cycle exit in progenitor cells.
3.2.11 Dmbx1 over-expression resulted in premature cell cycle exit
To further confirm that Dmbx1a and Dmbx1b play an important role in cell cycle regulation,
dmbx1a and dmbx1b mRNAs were ectopically expressed at 1-cell stage. About 30% of these
injected embryos (n=425) were dorsalized, and therefore, could not be analyzed. However, 50%
of the injected embryos appeared normal, suggesting that the effect of RNA-injection was
variable. Only some of the remaining dmbx1-overexpressing embryos (~20% of total injected
embryos) had various degrees of cell cycle related phenotypes.
Since overexpression of Dmbx1 could lead to downregulation of ccnd1, it is possible that retinal
progenitor cells may exit the cell cycle early and thus the overall proliferation is reduced. I
decided to investigate if these retinal progenitor cells exit their cell cycles prematurely in
Dmbx1a+1b over-expression embryos. Using BrdU and PHH3 as cell cycle markers, it was
observed that the staining of both markers in these animals at 24hpf when there were excess of
Dmbx1 (Figure 3.14B, D) was reduced when compared to un-injected embryos (Figure 3.14A,
C). Since retinal cells in a Dmbx1-overexpressing animal did not have proper differentiation and
lamination, and yet they were no longer mitotic, there was a possibility that these post-mitotic
cells might be undergoing cell death without proper differentiation. To verify whether these
cells are undergoing apoptosis, I performed Caspase 3 immunohistochemistry on these embryos.
There was no noticeable increase in the number of dying cells in the retina at 48hpf when
Dmbx1 was ectopically expressed (Figure 3.14E-F). This indicated that the lack of retinal
differentiation in these progenitor cells that had exited their cell cycle prematurely was not due
to increased cell death.
103
Figure 3.14 Retinal progenitor cells of those dmbx1-overexpressing embryos
undergo premature cell cycle exit.
Dmbx1a and dmbx1b mRNAs were co-injected in the embryos at the 1-cell stage to test whether
Dmbx1 can sufficiently promote cell cycle exit in retinal progenitor cells. Indeed, when dmbx1
genes are misexpressed, these embryos (B, D) have reduced amount of proliferating cells in the
retina when compared to un-injected embryos (A, C) as indicated by BrdU (A-B, n=3 in each
group) and PHH3 (C-D, n=3 in each group) staining at 24hpf. It seems that these progenitor
cells are forced to become post-mitotic prematurely. Moreover, these retinal neurons that
prematurely exited the cell cycle do not appear to induce cell death. White arrows point to
positive-staining cells.
104
3.2.12 Misexpression of dmbx1 enhances the production of earlier-born
neurons in the retina
When dmbx1 was misexpressed by injecting mRNA at the 1-cell stage, the retinal laminations in
these embryos were also lost. It is known that the type of retinal neurons these progenitor cells
become is partially dependent on the timing of cell cycle exit during retinogenesis (Dyer and
Cepko, 2001a; Dyer and Cepko, 2001b; Ohnuma et al., 2002), thus investigating the type of
retinal neurons these cells are differentiated into will be informative. If overexpression of
dmbx1 paralogs can lead to early cell cycle exit of these progenitor cells, then earlier-born
retinal neurons such as the RGCs may be more abundant in those embryos.
In a wildtype animal, the first group of post-mitotic RGCs appears at the ventronasal retina
between 27hpf and 28hpf (Hu and Easter, 1999). These pioneer RGCs express Alcam-a, which
is a cell adhesion molecule that is transiently up-regulated in new RGCs as well as in their axons
(Laessing and Stuermer, 1996). Expression of Alcam-a can be detected through
immunohistochemistry across ganglion cell layer in the central retina (Laessing and Stuermer,
1996). By 3dpf when these RGCs have reached the optic tectum, the level of Alcam-a also drops
significantly (Laessing and Stuermer, 1996).
Using Zn-5 to detect Alcam-a immunoreactivity at 72hpf, I found that the expressions of this
marker was reduced in dmbx1 morphants but enhanced in dmbx1-overexpressing embryos. It is
possible that that reduced level of dmbx1 genes led to longer cell cycle in these retinal
progenitor cells, and hence only a few of them become post-mitotic and differentiated into
RGCs (Figure 3.15C-D). On the other hand, ectopic expression of Dmbx1 paralogs forced those
progenitor cells to undergo terminal division much earlier so that they take up the early-born
RGC cell fate (Figure 3.15E-F). The current view on neurogenesis is that cell cycle exit and cell
specification are tightly linked and are well-coordinated in timing so that when cells undergo
105
Figure 3.15 Early cell cycle exit in dmbx1-overexpressing embryos lead to
bias in earlier-born retinal neuron at 72hpf.
Cryosections from 72hpf embryos, coronal view (n=3 in each group). Using Zn5 to label RGCs,
knock down of dmbx1 (C-D) causes a reduction in Zn5 expression compared to un-injected
controls (A-B). During late retinogenesis, a significant increase of Zn5 positive cells is observed
in 72hpf embryos that overexpressed dmbx1 at the 1-cell stage, suggesting that these dmbx1-
overexpressing cells may have exited the cell cycle early on and therefore, majority of them had
taken up the RGC (earlier-born retinal neuron) cell fate instead. White rectangular box
represents the magnified region in (B, D, F).
106
their last round of cell division, they are almost immediately targeted for differentiation, so
different subtypes of retinal neurons are generated depending on the timing of their cell cycle
exit (Bilitou and Ohnuma, 2010; Dyer and Cepko, 2001a; Dyer and Cepko, 2001b). Thus, too
much dmbx1 can push these retinal cells to exit cell cycle early and take on earlier-born cell fate
during retinogenesis.
3.3 Discussion
3.3.1 Dmbx1 regulates retinal neurogenesis
The present results expand upon dmbx1a and dmbx1b knockdown experiments discussed in
chapter 2 and demonstrated that the zebrafish dmbx1 genes have a fundamental role in
retinogenesis through cell cycle regulation of retinal progenitor cells. My data revealed that
transcription factors Dmbx1a and Dmbx1b can downregulate the level of ccnd1 in retinal
progenitor cells during the transition between progenitors to mature neurons. However, these
retinal progenitor cells in the double morphants did not arrest in G1, possibly because other
known ccnd1 regulators could still regulate this gene in the retina during development. For
instance, both p27kip1
/cdkn1b and p57kip2
/cdkn1c are well-studied CKIs that antagonize ccnd1 to
promote cell cycle exit in the retina (Dyer and Cepko, 2001a). In dmbx1 double morphants, the
expression of cdkn1b and cdkn1c were still maintained to downregulate other cyclins. This helps
to explain why these retinal progenitor cells with high level of ccnd1 transcripts did not
completely halted at the G1 phase, but instead took five times longer to progress through the cell
cycle.
How does high level of ccnd1 perturb the whole cell cycle instead of just the G1 phase? It has
been documented in synchronized cell culture studies that Ccnd1 is induced around G2 and
peaks at G1 to promote another round of cell division; however, the level of ccnd1 needs to drop
107
before a cell can enter S-phase (Stacey, 2003). Overexpression of ccnd1 in cell lines and
fibroblasts showed that DNA synthesis is inhibited since excess amount of Ccnd1 binds and
inactivates PCNA, which is part of the DNA replication machinery (Fukami-Kobayashi and
Mitsui, 1999; Pagano et al., 1994). Thus, the increased level of ccnd1 found in dmbx1 double
morphants may accelerate the progression of G1 but prolong G1 to S transition and S-phase due
to insufficient amount of PCNA to carry out proper DNA synthesis. Therefore the overall length
of the cell cycle time in retinal progenitor cells is longer in double morphant embryos. My
results indicate that cells with lower level of dmbx1 were able to reach mitosis, suggesting that
the cells did not just stall at S-phase as a result of depleted PCNA. It will be of interest to assess
the duration of each phase in a cell cycle between control-injected and dmbx1 morpholino-
injected embryos in order to confirm this hypothesis.
It was also observed that retinal differentiation in dmbx1 double morphants had caught up to the
controls by 5dpf even though the overall growth defect was permanent. This illustrates that the
number of cell cycles each progenitor cell progresses through is not intrinsically programmed;
otherwise, these retinal progenitor cells would continue to proliferate until the appropriate eye
size is reached before any differentiation events take place. Instead, these progenitor cells switch
to terminal differentiation when affected by extrinsic cues in their environment. Thus the size of
the eye remains small in the dmbx1 morphants but differentiated neurons are apparent in the
neural retina.
3.3.2 The transcriptional relationship between dmbx1 and ccnd1
All the evidence so far supports the idea that cyclin D1 functions in the same pathway as the
dmbx1 paralogs, possibly downstream. The effects on ccnd1 transcript levels in MO1a+MO1b
morphants and dmbx1-overexpressing embryos suggested the possibility of direct transcriptional
108
regulation by Dmbx1 on cyclinD1. Not much is known about the binding specificity of zebrafish
Dmbx1 transcription factors; however, Kimura et al. showed that in vitro, mouse Dmbx1 can
bind to the consensus sequences TAATCCGATTA and TAATCC(N2-4)TAATCC (Kimura et al.,
2005). Since the homeodomain regions of these two orthologs are 100% conserved, I predict
that the zebrafish Dmbx1 proteins can potentially bind the same DNA sequences. Preliminary
examination of ccnd1 promoter did not yield any potential binding regions within 30kb
upstream of the ccnd1 transcription start site. Thus, I lowered the stringency and looked for the
common K50 paired-type homeodomain binding sites ‒ TAATCCG or TAATCC (Baird-Titus et
al., 2006; Chaney et al., 2005; Zhang et al., 2002). From the modified search, I found three
TAATCC sites within 5kb upstream of the cyclinD1 transcription start site (Figure 3.16).
There are several ways to verify whether Dmbx1 can bind those sites. One approach is to
perform chromatin immunoprecipitation (ChIP) to test if Dmbx1 can bind to ccnd1 promoter in
vitro. I am currently making specific antibodies for both Dmbx1a and Dmbx1b in order to
perform the ChIP experiments (see Appendix 3 for details). In addition, validating the direct
transcriptional regulation between Dmbx1 and ccnd1 in vivo will be more compelling. I plan to
subclone a 5kb endogenous fragment of the ccnd1 promoter upstream of a luciferase reporter
plasmid. I will then co-inject this construct together with either dmbx1 morpholinos
(MO1a+MO1b) or dmbx1a+dmbx1b mRNAs into the zebrafish embryos and assay for changes
in luciferase activity. My prediction is that Dmbx1 will transcriptionally repress ccnd1
expression, so the luciferase activity should increase in the absence of Dmbx1 proteins, and
decrease when dmbx1 genes are ectopically expressed. If warranted, this analysis could be
extended to include subcloning the same 5kb ccnd1 promoter fragment without all the putative
Dmbx1 binding sites into the same luciferase reporter plasmid to check if the transcriptional
regulation is abolished.
109
Figure 3.16 Potential Dmbx1 binding sites on ccnd1 promoter.
The examined region is taken from chromosome 7 of the zebrafish genome at position
54550500 – 54577321. Ccnd1 gene structure is shown on the bottom right with blue boxes
representing exons and horizontal lines as introns. Blue arrowheads indicate the direction of
transcription. Three TAATCC sites are found within 5kb of the ccnd1 transcription start site.
110
3.3.3 Dual functions of Dmbx1 during retinal development in pre- and post-
larval stages
There are several examples in the literature where the transcription factors have different
functional roles in the retina during embryonic and post-embryonic development. Pax6 is known
to be an early retinal marker that helps pattern the eye field and promote proliferation in retinal
stem cells (Marquardt et al., 2001; Nornes et al., 1998; Xu et al., 2007). However, it has also
been shown to regulate the production of specific ganglion and amacrine cells (Marquardt et al.,
2001; Zaghloul and Moody, 2007a; Zaghloul and Moody, 2007b). Moreover, Chx10/vsx2 plays
a role in expanding the pool of retinal progenitor cells during embryogenesis (Barabino et al.,
1997; Burmeister et al., 1996; Dhomen et al., 2006; Green et al., 2003; Passini et al., 1997;
Vitorino et al., 2009), but it has also been shown to be required for maintaining subtypes of
bipolar neurons in the post-embryonic retina (Burmeister et al., 1996; Hatakeyama et al., 2001;
Kokkinopoulos et al., 2008; Livne-Bar et al., 2006; Passini et al., 1997). In post-embryonic
stages, strong expression of dmbx1a is maintained in the apical INL and dmbx1b is expressed
transiently in the GCL and basal INL. The functional roles of both genes in the retina of a
zebrafish larva will require further examination. Based on the spatiotemporal expression of
dmbx1 genes from 4 – 6 dpf (see chapter 2), it is possible that they may be required for
specification or maintenance of other neuronal subtypes in the GCL and INL. Performing
double in situ hybridization with the dmbx1 genes and other known retinal markers expressed in
GCL and INL will help characterize the neuronal identities of dmbx1a- and dmbx1b-positive
cells.
3.3.4 Working model for Dmbx1 regulation of cell cycle exit during
retinogenesis
In this chapter, I have demonstrated that the Dmbx1 can regulate cell cycle exit in retinal
progenitor cells. Reducing the level of dmbx1 paralogs can cause a remarkable increase in the
111
amount of ccnd1 mRNA in the retina. Moreover, knocking down the elevated level of ccnd1
using antisense morpholino in dmbx1 morphants allowed partial rescue of the cell cycle defects.
Upon more detailed examination of dmbx1a expression in the retina, it appears that this gene
begins to turn on in the INL and PL of the retina at 2dpf (~55hpf) before its expression is
restricted to a more refined domain in the INL at 3dpf (Figure 3.17A-B). The initiation of
dmbx1a expression in the retina coincides with the period in which mitotic progenitors are fated
to become later-born retinal neurons are born. The spatiotemporal expression of dmbx1a
suggests that this paralog plays a pivotal role in regulating the cell cycle of some of these retinal
progenitor cells during their transitions to become matured differentiated neurons or glia. In
addition, the spatial expression of ccnd1 and dmbx1 paralogs also fits into this model quite well.
Both genes are detected in the same tissue (retina and tectum), but they are non-overlapping and
appeared to be mutually exclusive — ccnd1 is presents in mitotic cells, while dmbx1 genes are
maintained at terminally differentiated neurons. Thus Dmbx1a and Dmbx1b are needed for
some of these progenitor cells to exit the cell cycle.
A working model is proposed in Figure 3.17C to explain how Dmbx1a and Dmbx1b might
function during retinogenesis. In early retinal development, cells within the neural retina are
proliferative and ccnd1 can be detected in these progenitor cells. Transcription factors Dmbx1a
and Dmbx1b begin to initiate their expressions in the INL and gradually downregulate the level
of ccnd1 in progenitor cells in the INL and PL of the central retina, to allow these retinal cells to
become post-mitotic and undergo proper differentiation. Some mature neurons in the INL will
continue to express dmbx1a and dmbx1b in the post-embryonic retina. In this model, the
assumption is that Dmbx1a/b only regulates ccnd1, but it was discovered from the triple
morphants (MO1a+MO1b+MO-ccnd1) that Dmbx1a/b may also be regulating additional
112
Figure 3.17 Model of dmbx1 regulation of cell cycle exit in retinal progenitor
cells.
(A-B) Dmbx1a begins to be expressed in the INL of the retina around 55hpf and its expression
is more concentrated by 3dpf. (C) Diagram of Dmbx1 (particularly Dmbx1a) regulation of
retinal progenitor cells during zebrafish embryonic development. Dmbx1a expression initiates
around 2dpf, when most progenitor cells are gradually become post-mitotic. This transcription
factor helps downregulate the level of cyclin D1 in these retinal progenitors so they can undergo
terminal differentiation. Expression domains of dmbx1a and ccnd1 are mutually exclusive at
3dpf. Once these retinal cells are differentiated, dmbx1a expression is retained in a subset of
cells (likely bipolar neurons) in the INL.
113
components in the cell cycle (such as cdkn1c or Rb). Examining addition regulators at various
phases of the cell cycle will further reveal the effect of dmbx1 paralogs on cell cycle kinetics.
3.3.5 Dmbx1 may interact with Notch and Wnt signaling pathways in the
retina during neurogenesis
Components from major signaling pathways have also been proven to interact with Dmbx1
during retinal and tectal development. Studies in zebrafish and Ciona have both demonstrated
that dmbx1 works downstream of Delta-Notch signaling. In Ciona, Notch signaling is turned off
in visceral ganglion cells that are expressing Dmbx1 (Stolfi and Levine, 2011). Progenitor cells
that are fated to express Dmbx1 will have proneural gene such as Neurogenin (Ngn) to
upregulate Delta, which then inhibit neighbouring cells to become neurons by activating Notch
signaling (Stolfi and Levine, 2011). In zebrafish, the mind bomb mutant (mibhi904
), which fails to
trigger Notch signaling due to a dysfunctional E3 ubiquitin ligase that helps internalize Delta
and Notch in the signaling cells (Chen and Casey Corliss, 2004; Itoh et al., 2003), expression of
dmbx1a decreased by two-fold and loss of dmbx1a expression was observed in midbrain,
hindbrain, and retina at 72hpf (Hortopan and Baraban, 2011). Several papers have characterized
that the mind bomb mutants have excess primary neurons but diminished number of neural
progenitor cells and later-born neuronal subtypes (Bingham et al., 2003). It is possible that the
excess neurons that are born prematurely can no longer express differentiated markers in tectal
and retinal regions, resulting in a reduction of dmbx1a expression at 72hpf in those mutants.
Another key signaling pathway involved in retinal development is the Wnt/Frizzled signaling
cascade. Numerous Wnt molecules and Frizzled receptors are present in the vertebrate retina
during embryonic development (Van Raay and Vetter, 2004). It has been shown in mouse and
chicken that both canonical and non-canonical Wnt signaling are involved in retinal
development (Fuhrmann, 2008; Van Raay and Vetter, 2004). Several mutant strains with
114
disrupted canonical Wnt signaling pathway have been identified in zebrafish. In apc mutant,
scaffolding protein that targets β-Catenin degradation is mutated, resulting in constitutively
active Wnt pathway (Stephens et al., 2010). Another mutant strain masterblind (mbl) has a point
mutation in axin1, which leads to inhibition of GSK3β and causes an increase in stabilized β-
Catenin and the hyperactivation of canonical Wnt signaling (van de Water et al., 2001). Both
apc and mbl mutants have retinal defects due to over-proliferation of progenitor cells in the
CMZ, resulting in the absence of proper differentiation in the neural retina (Meyers et al., 2012;
Stephens et al., 2010). Retinal phenotype in dmbx1 double morphant and mutants with up-
regulated Wnt signaling highly resemble one another. Expression domains of retinal progenitor
genes expand from the CMZ and mitotic cells (PCNA and BrdU positive) still remain in the
central retina beyond 50hpf when they should undergo retinal differentiation (Meyers et al.,
2012; Stephens et al., 2010). It is possible that dmbx1 may be antagonizing Wnt signals in the
retina in order to facilitate cell cycle exit in those retinal progenitor cells. Thus, it would be
interesting to see whether constitutively active Wnt mutants (such as apc and mbl) can be
rescued by overexpressing dmbx1 genes.
115
Chapter 4
Role of Dmbx1 in Midbrain Formation during Embryonic
Development in Zebrafish
Part of this chapter is adapted from:
Wong, L., Weadick, C. J., Kuo, C., Chang, B. S. and Tropepe, V. (2010). Duplicate dmbx1
genes regulate progenitor cell cycle and differentiation during zebrafish midbrain and retinal
development. BMC Dev Biol. 10, 100.
I analyzed the function of dmbx1a and dmbx1b in the optic tectum, and demonstrated that
morpholino knocked down of both dmbx1 genes resulted in the loss of tectal-specific markers
using in situ hybridization. I illustrated through immunostaining experiments that the loss of
tectal tissue was due to the inability of these tectal progenitor cells to proliferate properly and
exit their cell cycles in a promptly manner.
116
Role of Dmbx1 in Midbrain Formation during Embryonic
Development in Zebrafish
4.1 Introduction
Midbrain, or mesencephalon, is part of the brainstem associated with the visual and auditory
system (Aubie et al., 2012; Deeg et al., 2009; Nevin et al., 2010; Shen et al., 2011). It receives
sensory information and acts as a relay center that regulates an organism’s motor output. The
midbrain is partitioned dorsoventrally and each region has a distinct anatomical organization
and function. In mammals, the dorsal midbrain consists of superior and inferior colliculi, which
are equivalent to the optic tectum and torus semicircularis in other vertebrates. Studies across
species have illustrated that the tectum plays a much more important role in non-mammalian
vertebrates compared to mammals, since many of the information processing tasks of the
superior/inferior colliculi have been reassigned to the cerebral cortex in mammals (Huberman
and Niell, 2011; Nevin et al., 2010; Shimizu et al., 2010). The ventral tegmentum consists of a
group of nuclei (dopamine neurons) in the substantia nigra that is crucial for motor functions
(Smidt and Burbach, 2007). Degeneration of dopaminergic neurons is the cause of Parkinson’s
disease (Sulzer, 2007). In zebrafish, medial longitudinal fasciculus are efferent nerve fibers that
control movements such as swimming and prey capture (Gahtan et al., 2005; Sankrithi and
O’Malley, 2010), and the nucleus of this axon bundle can be found in the ventral tegmentum.
Cell bodies of some of the cranial and trigeminal nerves are also located in the ventral midbrain
(Higashijima et al., 2000; Higashijima et al., 2004a), indicating that tegmentum is linked to
multiple complex motor responses.
4.1.1 Regionalization of midbrain in the neural tube
Patterning begins along the rostrocaudal axis of the brain rudiment to subdivide the neural tube
into more refined neuromeres after neurulation is completed (Kimmel, 1993). Partitioning of
117
the neural tube initiates around 16 hpf and begins with two indents on the dorsal side that mark
the margins between prospective neuromeres (Ross et al., 1992). These boundary lines
eventually become signaling centers that help pattern the neighbouring territories. The rostral
line is called the diencephalic-mesencephalic boundary (DMB), while the caudal one is the
midbrain-hindbrain boundary (MHB). These two organizing centers segregate the neural tube
into three major primordia: forebrain, midbrain, and hindbrain.
Regionalization of the zebrafish midbrain during embryonic development is a coordinated
process. Otx2 and Gbx2 are two key transcription factors that set up the rostral-caudal axis
along the neural tube. Otx2 is necessary for the formation of forebrain and midbrain (Li and
Joyner, 2001; Li et al., 1994b; Scholpp and Brand, 2003; Scholpp et al., 2007), whereas gbx2 is
required for cerebellum and hindbrain development (Kikuta et al., 2003; Rhinn et al., 2003;
Rhinn et al., 2009). Analysis of the otxH morphant in zebrafish (simultaneous knocked down of
both otx2 and otx1-like genes) illustrated that caudal midbrain was transformed into
rhombomere 1 (Foucher et al., 2006). Otx2 and Gbx2 co-repress each another along the neural
tube and the boundary between the two defines the position of an important signaling center in
the brain, called midbrain-hindbrain boundary (MHB) (Hidalgo-Sánchez et al., 2005; Li and
Joyner, 2001; Raible and Brand, 2004). Three other factors — wnt1, fgf8, and pax2a, are
subsequently expressed in this region where they activate other genes that initiate midbrain and
hindbrain patterning. Wnt1 is a morphogen that promotes proliferation in the midbrain and Fgf8
is responsible for restricting hindbrain cell fate. In the fgf8 mutant acerebellar (ace -/-
), the
midbrain domain is enlarged and extends into the cerebellar region (Jászai et al., 2003). Pax2a is
a transcription factor expressed at the MHB, where it controls the formation of both MHB and
the midbrain via eng2a/2b (Lun and Brand, 1998; Scholpp and Brand, 2001). The MHB will fail
to form when any one of these factors is compromised (Belting et al., 2001; Buckles et al., 2004;
118
Lekven et al., 2003; Lun and Brand, 1998). When the balance among these regulators is lost, the
midbrain domain may expand into hindbrain or vice versa, where hindbrain cells transform into
cells with midbrain identity (Jászai et al., 2003; Rhinn et al., 2009). Thus, the MHB is a key
regulator for the regionalization of the posterior midbrain and anterior hindbrain.
The midbrain and forebrain are separated by the diencephalic-mesencephalic boundary (DMB),
which is also formed at the interface between two transcription factors — pax6a and eng2/3.
Pax6 is a key forebrain marker that can repress midbrain genes such as pax2 and eng2/3
(Matsunaga et al., 2000; Scholpp et al., 2003), whereas eng2/3 can promote midbrain identity
and inhibit forebrain cell fate by repressing pax6 (Araki and Nakamura, 1999; Scholpp and
Brand, 2003; Scholpp et al., 2003). This mutual repression between pax6 and eng2/3 refines the
DMB and forms the anterior boundary of midbrain. Other factors such as Fgf8 and Pbx have
also been shown to help set up the DMB by cooperating with Eng2/3 to repress Pax6 during
regionalization of the midbrain (Erickson et al., 2007; Scholpp et al., 2003).
4.1.2 Neurogenesis in the zebrafish brain
In zebrafish embryos, neurogenesis in the central nervous system begins around 16-18hpf. This
first wave of neurogenesis generates several clusters of post-mitotic cells across the brain. Using
anti-acetylcholinestarase (AChE) antibody to mark differentiated neurons, AChE staining was
first observed bilaterally in dorsal forebrain, ventral midbrain and ventral hindbrain between 16-
24 hpf (Ross et al., 1992). By contrast, some regions, like the optic tectum, are AChE-negative
at the end of the primary neurogenesis. Using proliferating cell nuclear antigen (PCNA) as a cell
cycle marker, it was confirmed that cells in those regions remain mitotic at 24hpf (Wullimann
and Knipp, 2000).
119
Secondary neurogenesis occurs from around 48hpf until 5dpf, but most cells become post-
mitotic from 48 to 72 hpf (Mueller and Wullimann, 2003). As proliferating cells across the
central nervous system gradually decrease, the remaining mitotic cells become restricted to
certain areas in the brain called stem cell niches. Studies using antibodies against PCNA and
HuC/D to label mitotic and early differentiated cells respectively demonstrated that the spatial
expression patterns of these markers are virtually complementary to each other (Mueller and
Wullimann, 2003). It was revealed that proliferative zones are typically situated close to the
ventricular space, while differentiated mature neurons migrate towards the periphery. In the
optic tectum, there are two identified proliferation zones. They are located at the dorsomedial
(close to the midline beneath the roof plate) and ventrolateral regions (dorsal to where the torus
semicirularis is) (Wullimann and Knipp, 2000). In the adult optic tectum, proliferating cells can
be found in the periventricular gray zone, which is situated at the caudal region of the optic
tectum (Ito et al., 2010; Kizil et al., 2012).
Another mechanism that potentially regulates neurogenesis is programmed cell death. A detailed
study across different regions of the nervous system at various developmental stages during
zebrafish embryogenesis revealed apoptosis to be in sync with axon-pruning activities (Cole and
Ross, 2001). Using the TUNEL method to label dying cells, was shown that apoptotic events are
required to eliminate damaged neurons and also to facilitate morphogenic movement by
providing adequate space for further growth (Cole and Ross, 2001). During neurulation, the
majority of cell death is concentrated at the dorsal midline which serves to eliminate defective
cells in the neural keel (Cole and Ross, 2001). Once the neural tube is patterned, apoptotic cells
are scattered along the neural tube and the number of dying cells peaks at particular areas during
regional synaptogenesis. Apoptosis is most frequently observed in the retina around 36hpf when
retinal cells begin to send fibers across the optic chiasm at the ventral midline that arborize on
120
the contralateral dorsal brain hemisphere (Cole and Ross, 2001). Cell death helps eliminate any
aberrant axonal outgrowths and provides room for the optic nerve to exit the retina. The highest
number of apoptotic cells in the optic tectum is detected at 60hpf (Cole and Ross, 2001). This
coincides with the timing of axon pruning and the elimination of target cells that lack neuronal
connections.
4.1.3 Specification of the optic tectum
Dorsoventral patterning factors such as Wnt1/3a, Zic2/5, and Pax3/7 are all required for the
proper formation of the optic tectum. Wnt1/3a is an important signaling molecule that is
expressed at the midline of the dorsal roof plate between two tectal lobes (Buckles et al., 2004;
Clements et al., 2009; Krauss et al., 1991; Molven et al., 1991). In the absence of Wnt signaling,
increased apoptosis via p53 was observed in the midbrain as well in both signaling organizers,
the DMB and MHB (Buckles et al., 2004; Mattes et al., 2012). In addition, Wnt molecules are
also known for their mitogenic role during neurogenesis. It had been shown in chicken and frog
that ectopic Wnt signaling induced additional brain tissues, as well as resulting in axial
bifurcation (Lee et al., 2000; McMahon and Moon, 1989; Megason and McMahon, 2002; Wolda
et al., 1993). Zinc-finger transcription factors Zic2a/5 are also responsible for tectal
neurogenesis in zebrafish. In the study by Nyholm et al., the authors demonstrated that Wnt
signaling could activate the transcription of zic2/5 (Nyholm et al., 2007). When the zic genes
were knocked down, the tectal domain was reduced, suggesting that zic2a/5 were required for
growth of the optic tectum possibly by regulating the cell cycle (Nyholm et al., 2007).
Moreover, it has been well-established in chicken and mouse that Pax3/7 are key transcription
factors that specify the optic tectum. In one study, it was shown that misexpression of Pax3/7 in
the diencephalon induced ectopic tectal cells in chicken (Matsunaga et al., 2001). Pax7 is
required to maintain a subpopulation of tectal cells, and it also regulates the polarity in the optic
121
tectum to ensure proper tectal topography during development (Thomas et al., 2006; Thompson
et al., 2007; Thompson et al., 2008). Moreover, Shh is expressed at the midline of the ventral
midbrain (tegmentum), where it specifies tegmental cell fate by antagonizing the expression
domains of dorsal tectal markers (Fedtsova and Turner, 2001; Watanabe and Nakamura, 2000).
However, a recent study showed that Shh could also have a positive effect on neurogenesis in
the zebrafish optic tectum, suggesting a diverse role of Shh during midbrain development
(Feijóo et al., 2011).
4.1.4 The canonical Wnt signaling pathway and its importance in midbrain
development
The mechanism by which the canonical Wnt pathway promotes cell proliferation has been
studied extensively. Experiments have shown that Wnt1 and Wnt3/3a play crucial roles in
midbrain and rostral hindbrain development in mouse, chicken, frog and zebrafish; highlighting
the fact that this evolutionarily conserved pathway is indispensable for brain patterning (Bally-
Cuif and Wassef, 1994; Buckles et al., 2004; Clements et al., 2009; Saint-Jeannet et al., 1997;
Takada et al., 1994; Thomas and Capecchi, 1990). These Wnt molecules bind and activate
Frizzled (Fz) — the seven-transmembrane receptor (Bhanot et al., 1996; Yang-Snyder et al.,
1996). There are 15 Fz genes in zebrafish, and Wnt1/3a in zebrafish may interact with multiple
Fz receptors based on their overlapping spatiotemporal expression patterns (Nikaido et al., 2013;
Sisson and Topczewski, 2009). Similar to the mouse Fz homologs, Wnt receptors in zebrafish
may be functionally redundant in canonical Wnt signaling (Fischer et al., 2007). Recently,
studies in mouse have shown that Fzd3 and Fzd6 double mutant embryos had severe midbrain
morphogenesis defects (Stuebner et al., 2010), which will help us narrow our focus on those Fz
homologs in zebrafish. Extensive studies have been performed to determine the components of
the canonical Wnt transduction pathway (Figure 4.1). When the cell is in the “off-state”, the
122
Figure 4.1 Schematic diagram of canonical Wnt pathway.
(A) Without Wnt, Axin/GSK-3/APC complexes degrade β-cat before it can translocate into the
nucleus. Transcription factor Lef1acts as a repressor in the absence of β-cat. (B) When the Wnt
ligand is present, it interacts with the Fzd receptor and prevents the Axin/GSK-3/APC
complexes from degrading β-cat. The β-cat proteins can enter the nucleus and binds with Lef1,
which then becomes a transcriptional activator.
123
cytoplasmic Axin/GSK-3/APC complex binds to Beta-catenin (β-cat), a key component of this
cascade, and GSK-3 kinase phosphorylates resulting in proteolytic degradation of the cytosolic
β-cat (Farr et al., 2000; Hart et al., 1998). In the presence of Wnt molecules, the Fz receptor
disassembles the Axin/GSK-3/APC destruction complexes and allows β-cat to accumulate
(Alves dos Santos and Smidt, 2011; Li et al., 2012). β-cat then enters the nucleus to bind with
the T cell factor/lymphoid enhancer factor (Tcf/Lef) repressor complexes, and transforms them
into transcriptional activators, which then turn on Wnt target genes (Hsu et al., 1998; Porfiri et
al., 1997). It is known that Wnt pathway target genes include ccndD1 and n-myc, which are
crucial for proliferation during development (Kuwahara et al., 2010; Shtutman et al., 1999).
4.1.5 The role of Dmbx1 in regulating optic tectum development
In zebrafish, it has been shown that when dmbx1a and dmbx1b are knocked down, the growth of
the optic tecta was significantly reduced (Kawahara et al., 2002, see chapter 2). Studies in
mouse have demonstrated that a null allele of Dmbx1 results in neurogenesis defects in
homozygous mutant embryos (Ohtoshi and Behringer, 2004). In addition, a recent study showed
that duplication of a copy number variant region, which includes the Dmbx1 gene, is linked to
brain malformation in human fetuses (Serra-Juhé et al., 2012). However, the molecular
mechanism by which Dmbx1 regulates the development of the tectum remains poorly
understood. Hence, I wanted to investigate the potential role of dmbx1 in regulating the
development of the optic tecta during zebrafish embryogenesis. Evidence presented in the
previous chapter suggested that dmbx1 paralogs could regulate the level of cyclinD1 (ccnd1) in
retinal progenitor cells. Thus I wanted to examine whether these paralogous genes could
perform the same function in tectal progenitor cells. I hypothesized that Dmbx1a and Dmbx1b
are necessary to reduce the level of ccnd1 present within tectal progenitor cells. Failure to lower
the level of Cyclins might lead to slower cell cycles and delay in cell cycle exit, which would
124
result in a proliferation deficiency and lack of differentiation in the optic tectum. Furthermore, I
proposed an alternative mechanism by which Dmbx1a/1b could regulate tectal development. I
have evidence to support the hypothesis that Dmbx1a and Dmbx1b interact with the canonical
Wnt pathway, downstream of the Wnt ligand, to induce tectal identity in cells that are competent
to become tectum. This is the first study to show that Dmbx1 transcription factors are specific
and pivotal tectal determinants that work synergistically with the canonical Wnt signaling in the
midbrain.
4.2 Results
4.2.1 Dmbx1 is required for specification of the optic tectum in zebrafish
Data in chapter 2 showed that when dmbx1a and dmbx1b were knocked down using
morpholinos, the size of the optic tectum was reduced by half and the medulla oblongata was
smaller when compared to control-injected embryos (ctrl-injected). In addition to growth
retardation, loss of dmbx1 paralogs also affected the specification of the optic tectum. To
investigate how dmbx1 affected tectal development, I performed wholemount in situ
hybridization on ctrl- and MO-injected embryos at various stages to assay whether expression of
different brain markers was affected, and if there were any changes in tectal cell fate in the
morphants. First, I examined genes known to be expressed in the optic tecta. foxb1.2, lef1, and
otx2 are previously characterized tectal markers (Bonkowsky et al., 2008; Dorsky et al., 1999;
Mercier et al., 1995; Thisse et al., 2001). The results indicated that expression of these markers
was eliminated or strongly reduced in the tectal region of dmbx1 morpholino-injected embryos
at 48hpf (Figure 4.2). This finding suggested that dmbx1 paralogs were required for tectal
neurogenesis.
125
Figure 4.2 Whole-mount in situ hybridization of tectal markers in dmbx1
double knockdown embryos at 48hpf embryos.
All panels are lateral views, anterior is to the left (A-F). The marker used in each group is shown
on the left and the number of embryos examined is indicated at the bottom right. Expression of
dorsal midbrain markers had reduced in the morphant embryos (D-F) compared to the control-
injected embryos (A-C). Arrow marks the reduced expression of these tectal markers in the
dmbx1 morphant embryos.
126
4.2.2 MHB is unaffected in dmbx1 morphants
To further investigate whether brain patterning is affected, I examined the expression of genes
expressed in the MHB since it serves as an organizer that provides important signaling to
regulate the formation of midbrain (Jászai et al., 2003; Raible and Brand, 2004). In dmbx1
double morphants, no change was detected in the expression of several MHB markers, including
eng2b (previously known as eng3) (Ekker et al., 1992), etv5b (previously known as erm)
(Münchberg et al., 1999), fgf8 (Reifers et al., 1998), or pax2a (Krauss et al., 1991) (Figure 4.3).
These results suggested that MHB signals were upstream of dmbx1 paralogs during midbrain
patterning, and that the loss of tectal structure did not affect regionalization of the midbrain.
To follow up on whether the dmbx1 paralogs were downstream of MHB genes, I looked for
changes in dmbx1a and/or dmbx1b expression that might occur in the absence of the MHB. To
do this, I analyzed the expression of dmbx1 genes in the acerebellar (ace-/-
) mutants, which
carry a mutation in the fgf8 gene. Homozygous mutant embryos lack the MHB organizer and
cerebellum (Reifers et al., 1998). Wholemount in situ hybridization of dmbx1a revealed a caudal
expansion similar to previous studies (Jászai et al., 2003; Kawahara et al., 2002) (Figure 4.4B).
However, further examination at a later stage indicated that dmbx1a expression was restored
(Figure 4.4D). This suggested that even though other MHB genes had taken longer to turn on in
the absence of Fgf8, they were able to prevent dmbx1a from expanding into the hindbrain. On
the other hand, dmbx1b expression remained the same in the ace-/-
mutants (Figure 4.4F-G).
Hence, both dmbx1 genes were blocked by the MHB from being expressed beyond the posterior
midbrain, and they were likely involved in specification rather than regionalization of the
midbrain.
127
Figure 4.3 Whole mount in situ hybridization of various midbrain-hindbrain
boundary (MHB) markers.
All panels are lateral views, anterior to the left with embryos at 24hpf (A-B, E-F) and 48hpf (C-
D, G-H). The marker used in each group is shown on the left and the number of embryos
examined is indicated at the bottom right. All four MHB markers had no expression change in
the dmbx1 double morphant embryos (E-H) compare to control-injected embryos (A-D). Arrow
points to the unaltered expression of these MHB markers in the dmbx1 knock down embryos.
128
Figure 4.4 Whole mount in situ hybridization of dmbx1a and dmbx1b in
acerebellar (ace-/-
) mutants.
All panels are lateral views, anterior to the left at 24hpf (A-B, E-F) and 48hpf (C-D, G-H). The
marker used in each group is shown on top and the number of embryos examined is indicated at
the bottom right. Expressions of dmbx1a and dmbx1b are compared between ace-/-
mutants (B,
D, F, H) and their siblings (A, C, E, G). Dmbx1a expands posteriorly at 24hpf in the ace-/-
mutants (B) beyond where MHB normally is but the expression domain is restored by 48hpf
(D). Expression of dmbx1b remains the same from 24hpf to 48hpf between mutants (F, H) and
their sibling (E, G). Arrow points to where the MHB is.
129
4.2.3 Neuronal differentiation in the hindbrain is affected in dmbx1 double
morphant
Strong expression of dmbx1a and dmbx1b was initiated in the hindbrain at 24hpf and continued
until at least early larval stage (6 dpf) as reported in chapter 2. However, when dmbx1 genes
were knocked down, defects in the hindbrain were rather subtle. No significant growth reduction
in the medulla oblongata was observed in the morphants and early hindbrain patterning genes
such as egr2b (Woo and Fraser, 1998) were normal in the double morphant (Figure 4.5A, 4.5F).
Instead, various differentiated neuronal markers in the hindbrain were found to have altered
expression patterns. Dlx2a labels cranial neural crest in the pharyngeal arches (Sperber et al.,
2008), while lim1 and pax2a mark interneurons (Batista and Lewis, 2008; Toyama and Dawid,
1997). In situ hybridization with these markers indicated that they were reduced in the double
morphants, indicative of fewer differentiated hindbrain neurons (Figure 4.5G-I). In the case of
foxb1.2, specific hindbrain neurons were completely eliminated. It is possible that dmbx1 may
have little effect on the growth and patterning of the hindbrain; but is required for the
establishment of specific subtypes of neurons during late neurogenesis.
4.2.4 Forebrain and ventral markers are unaffected in dmbx1 morphant
embryos
Despite the significant decrease in expression of tectal and hindbrain markers, there was
relatively little change observed in the forebrain after dmbx1 knockdown. Dlx2a is present in the
telencephalon and diencephalon (Akimenko et al., 1994), and it was expressed normally when
dmbx1 genes were knocked down (Figure 4.6D). This suggested that the loss of tectal tissue was
not due to change of cell fate to forebrain identity during development. Since the optic tectum is
situated dorsally in the midbrain, it was important to confirm whether the loss of tectal tissue in
dmbx1 double morphants was due to a shift from dorsal to ventral mesencephalic cell fate.
130
Figure 4.5 Whole mount in situ hybridization of various hindbrain markers.
All panels are lateral views, anterior to the left with embryos at 24hpf (A-B, F-G) and 48hpf (C-
E, H-J). The marker used in each group is shown on the left and the number of embryos
examined is indicated at the bottom right. Expression of all hindbrain neuronal markers have
decreased in dmbx1 double morphants (G-J) compare to control-injected embryos (B-E), but
hindbrain patterning was not affected (A-B). Arrow points to the reduced expression of
hindbrain neuronal markers.
131
Figure 4.6 Whole mount in situ hybridization of various forebrain and
ventral markers.
All panels are lateral views, anterior to the left with embryos at 24hpf. The marker used in each
group is shown on the left and the number of embryos examined is indicated at the bottom right.
Expression of both forebrain and ventral markers remain the same between dmbx1 double
morphants (D-F) compare to control-injected embryos (A-C). Arrow points to dlx2a expression
in the forebrain.
132
Using wholemount in situ hybridization, I examined two ventral markers ‒ foxa2 (previously
known as axial) and shh (Odenthal et al., 2000; Strähle et al., 1993). No change in their
expression was observed in the double morphants (Figure 4.6E-F). To further investigate
whether perturbing ventral patterning through Shh signaling pathway would affect dmbx1a or
dmbx1b expression in the midbrain, I inhibited Shh signaling by soaking embryos in
cyclopamine (which blocks Smoothened activity) (Chen et al., 2002b) and then examining
whether dmbx1a or dmbx1b expression was expanded ventrally. The results showed very minor
ventral expansion of the dmbx1a expression domain and no change in dmbx1b expression in the
absence of Shh activity (Figure 4.7C-D). Forebrain expression of dmbx1a was also missing
(Figure 4.7C) and this was likely due to the effect that loss of Shh signaling had on the
development of zona limitans intrathalamica, which is an important signaling center in the
forebrain.
4.2.5 The loss of tectal tissue in dmbx1 double morphant is independent of
cell death
One possible explanation for the decrease in tectal size in dmbx1 double morphant is elevated
level of cell death in the dorsal midbrain. To test this, acridine orange staining was used to
check for cell death in embryos. Acridine orange is a fluorescent dye that permeates into dying
cells and binds to their DNA. Results from the acridine orange staining did not show any
obvious increase of apoptotic cells at 48hpf and 72hpf in dmbx1 double morphants (Figure
4.8A-D). Acridine orange staining was performed on live embryos, immunostaining on sections
would allow more in depth analyses in tissue deep inside the brain. The results of TUNEL and
anti-Caspase 3 immunohistochemistry similarly confirmed that there was no increase of cell
death in the optic tecta in morphant embryos (Figure 4.8E-G). In fact, the number of dying cells
was reduced in morphants compared to controls. Thus, the loss of tectal tissue in dmbx1 double
133
Figure 4.7 Whole mount in situ hybridization of dmbx1a and dmbx1b in the
absence of Shh signaling.
All panels are lateral views, anterior to the left with embryos at 24hpf. The marker used in each
group is shown on the left and the number of embryos examined is indicated at the bottom right.
Expression of both dmbx1a and dmbx1b remain the same between un-injected (A-B) and
cyclopamine treated (C-D) embryos.
134
Figure 4.8 Cell death assays on dmbx1 morphants between 48-72hpf.
Acridine orange was used to stain dying cells in wholemount embryos (A-D). Panels are dorsal
view, anterior to the left. No staining is observed at the tectum in both control-injected (A) and
MO-injected embryos (B) at 48hpf (n=10 in each group). At 72hpf, there are many acridine
orange positive cells in the optic tectum during pruning (C), but the dmbx1 morphants have very
few dying cells. To further validate that, TUNEL (n=3 in each group) and Caspase 3 (n=12 in
each group) immunostaining on cryosections of 72hpf embryos were performed. Coronal
sections of morpholino-injected embryos (F, H) do not have evaluated cell death compared to
controls (E, G). Types of staining methods are indicated on the left. Numbers of Caspase 3
positive cells between control-injected (n=6) and morpholino-injected embryos (n=3) were
compared in graph (I), p<0.05.
135
FIGURE 4.8
136
morphant was not a result of tectal progenitor cells dying; instead it was likely that these tectal
cells did not proliferate properly.
4.2.6 Delayed cell cycle exit in tectal progenitor cells in dmbx1 knock down
embryos
The morphological defect in the optic tectum of dmbx1 morphants could reflect a failure to
proliferate. To test this hypothesis, I used PCNA to label cells in the cell cycle. At 72hpf, most
mitotic cells in control-injected embryos were restricted to the dorsomedial and ventrolateral
proliferation zones of the optic tectum (Figure 4.9A-C). In MO-injected embryos, PCNA
positive cells were observed across the whole optic tectum (Figure 4.9D). Since the optic tectum
was smaller when both dmbx1 genes were knocked down, the increased number of cycling
progenitor cells in the tectum led to the hypothesis that these tectal progenitor cells might be
unable to exit the cell cycle for differentiation. To investigate this, bromodeoxyuridine (BrdU)
was injected into the embryos to label cells that are in S-phase and phosphor-histone H3 (PHH3)
antibody was used to perform immunohistochemistry in order to detect cells in mitosis. The
rationale was that if these progenitor cells were arrested in S-phase of the cell cycle, then only
the number of BrdU-positive cells would increase. However, if these progenitors were stalled at
M-phase, then the number of PHH3-positive cells would be higher than expected. I found that at
72hpf, there were significantly more BrdU-positive and PHH3-positive cells in the double
morphants (Figure 4.9E-F) compared to the controls (Figure 4.9B-C). This result showed that
the tectal cells in the morphants were not arrested in the S- or M-phase of the cell cycle. It was
likely that the tectal cells were behaving similarly to the retinal cells in the dmbx1 double
morphant as in chapter 3. The analysis of retinal cells in dmbx1 morphants showed that
progenitor cells were cycling at least 5 times slower when dmbx1 paralogs were knocked down.
137
Figure 4.9 Delay in cell cycle exit leads to reduced tectal growth in dmbx1
morphants.
Immunohistochemistry was performed on coronal cryosections of 72hpf animals. Mitotic cells
can be found in the medial dorsal tectal region (white arrows) in control-injected embryos (A-
C), but morphant embryos had many more cells that are still cycling (PCNA positive, n=9 in
each group), and they spread across both tecta at 48hpf (D). However, those tectal cells in the
morphants are not proliferative since tectal size is severely decreased. Nonetheless, the presence
of both S- (BrdU positive, n=6 in each group) and M-phase (PHH3 positive, n=9 in each group)
markers in the morphants suggested that these tectal cells could progress through mitosis and
they were not arrested at a particular phase in the cell cycle (E, F). Size reduction and the lack of
differentiation in the dmbx1 morphants tectum is a result of delay cell cycle exit in those tectal
progenitor cells. Numbers of mitotic cells in control-injected and morpholino-injected embryos
are summarized in graph (G).
138
FIGURE 4.9
139
Thus the tectal cells might be behaving in a similar manner. The progenitor cells in the optic
tectum had cell cycle defects in the dmbx1 double morphants, resulting in a smaller optic tectum
with very fewer post-mitotic differentiated tectal cells.
4.2.7 Dmbx1 is necessary and sufficient for cell cycle exit in tectal progenitor
cells
It was shown in the previous section that dmbx1 double morphants had cell cycle exit delay in
the tectal progenitor cells. To examine whether the dmbx1 paralogs were sufficient to force
tectal progenitor cells to become post-mitotic prematurely, the S-phase marker BrdU was
examined in embryos overexpressing dmbx1a and dmbx1b mRNAs (Dmbx1a+1b over-
expression). At 24hpf, proliferation was at its peak in the brain, especially in the optic tectum.
By over-expressing dmbx1a and dmbx1b at 1-cell stage, it was expected that many of tectal
progenitor cells would no longer be mitotic at 24hpf. Indeed, the result showed that after a one
hour BrdU pulse-chase, there were fewer BrdU positive cells in the brain of Dmbx1a+1b
overexpression embryos (Figure 4.10). This indicated that excess amounts of Dmbx1 could lead
to early cell cycle exit of tectal progenitor cells; however whether these prematurely exited
tectal cells could undergo differentiation remained unknown.
4.2.8 Dmbx1 represses cyclinD1 in tectal progenitor cells
The results from Chapter 3 demonstrated that dmbx1 genes regulate the cell cycle by interacting
with cyclinD1 (ccnd1) in the retina. It was possible that these two transcription factors could
also be using the same molecular mechanism in the optic tectum to regulate cell cycle. At 72hpf,
ccnd1 was expressed at the medial and lateral proliferative regions of the optic tectum (Figure
4.11A-A’), which are well-characterized proliferation zones in the brain (Mueller and
Wullimann, 2005). In dmbx1 double morphant embryos, expression of ccnd1 was expanded in
140
Figure 4.10 Overexpressing dmbx1 genes promote premature cell cycle exit
in the tectum.
Embryos overexpressed with dmbx1a+1b mRNA from the 1-cell stage were injected with BrdU
together with controls. These animals were sacrificed an hours post-BrdU injections and process
for BrdU staining. Immunohistochemistry was performed on coronal cryosections of 24hpf
animals (n=3 in each group). S-phase proliferating cells can be found throughout the tectum in
control-injected embryos (A), whereas very few BrdU positive cells are observed when dmbx1
genes are misexpressed.
141
both dorsal and lateral domains (Figure 4.11B-B’). On the other hand, injecting dmbx1a and
dmbx1b mRNA into 1-cell stage embryos, resulted in reduced expression of ccnd1 (Figure
4.11C-C’). These data provided evidence that Dmbx1 regulates ccnd1 transcriptions in the optic
tectum, which is similar to the molecular mechanism that Dmbx1 employed in the retina. My
analysis revealed that Dmbx1 represses ccnd1 in tectal progenitor cells to trigger cells to exit the
cell cycle and undergo differentiation.
To verify that ccnd1 is downstream of Dmbx1, the elevated ccnd1 level observed in dmbx1
morphant embryos (MO1a+1b) was attenuated by co-injecting a ccnd1 morpholino (MO-ccnd1)
into 1-cell stage embryos. Animals injected with MO-ccnd1 alone developed a smaller but well-
differentiated optic tectum, as the expression of tectal marker such as lef1 was properly
expressed in the tectal tissue (Figure 4.12B-B’). As mentioned before, there was minimal
amount of lef1 expression in the optic tectum of MO1a+1b-injected embryos (Figure 4.12C-C’).
The reason was that when dmbx1 genes were knocked down, tectal cells progressed through cell
cycle slowly which led to reduced growth and lack of differentiated post-mitotic cells that
expressed lef1. By lowering ccnd1 level in the triple morphants (dmbx1a+dmbx1b+ccnd1
morpholinos injected simultaneously), I was able to demonstrate that it was possible to partially
rescue the cell cycle delay in the optic tectum. The recovery of lef1 expression in the tectal lobes
of triple morphant embryos (Figure 4.12D-D’) illustrated that by reverting the up-regulated
amount of ccnd1 in dmbx1 double morphant embryos, these tectal cells could undergo
neurogenesis properly. The triple morpholinos rescue experiments, together with the dmbx1
loss- and gain-of-function assays which showed an inverse correlation of ccnd1 expression
level, all suggest that cyclinD1 is the downstream component through which Dmbx1a and
Dmbx1b regulates cell cycle in the tectum.
142
Figure 4.11 Dmbx1 can regulate ccnd1 expression in the optic tectum.
Wholemount in situ hybridization of ccnd1 expression at 72hpf shown in lateral view, anterior
to the left, and the number of embryos examined is indicated at the bottom right (A-C). Yellow
lines indicate the position of where the coronal retina sections at the bottom are taken from (A’-
C’). Increase ccnd1 expression can be seen at the dorsomedial neurogenic zone in dmbx1
morphants, whereas the same region in the dmbx1-overexpressing embryos has much lower
ccnd1 expression when compared to un-injected control. Ccnd1 expression is present in lateral
and dorsomedial regions as indicated by yellow arrows and arrowhead respectively. These
results suggest that Dmbx1 can regulate ccnd1 mRNA level in the tectal progenitor cells during
their termination cycle so that these cells will undergo cell cycle exit and subsequently
differentiate into proper tectal neurons.
143
Figure 4.12 Repression of cyclinD1 rescues dmbx1 morphants phenotype.
Wholemount in situ hybridization of lef1 expression at 48hpf shown in dorsal view, anterior at
the top, the number of embryos examined is indicated at the bottom right (A-D). Yellow lines
indicate the position of where the coronal retina sections at the bottom are taken from (A’-D’).
Lef1 is expressed throughout the optic tectum in un-injected embryos (A-A’), whereas dmbx1
morphants have almost complete absent of lef1 (C-C’). Ccnd1 morphants have smaller optic
tectum but the expression of lef1 is still clearly visible (B-B’). Knocking down cyclin D1 with
morpholino (MOccnd1) in dmbx1 morphants (MO1a+1b) allowed partial growth recovery in the
midbrain as shown with lef1. More tectal progenitor cells in dmbx1 morphants were able to
progress through cell cycle when the level of ccnd1 was down, and thus rescue the tectal
phenotype in triple morphants.
144
4.2.9 Dmbx1 interacts with the Wnt canonical signaling pathway
One major signaling pathway that is known to be important for the optic tectum development is
the canonical Wnt pathway. It is possible that dmbx1a and dmbx1b are part of the Wnt signaling
cascade during neurogenesis. Another compelling reason to look at dmbx1 and Wnt signaling in
the midbrain is that both of them can alter the expression of ccnd1 in the optic tectum. Studies
have shown that ccnd1 is regulated by Wnt signaling, possibly via β–catenin/Lef1
transactivation (Ota et al., 2012; Shtutman et al., 1999). Hence, I begin to look more closely at
the relationship between dmbx1 paralogs and Wnt signaling pathway.
To test if the Wnt signaling pathway was compromised in dmbx1 morphants, the
Tg(TOP:dGFP) transgenic line was used as an in vivo readout for Wnt activity. The
Tg(TOP:dGFP) transgene has four consecutive Tcf/Lef binding sites under the c-fos promoter
that drives the expression of destabilized GFP as a reporter for activated Wnt signaling (Dorsky
et al., 2002). Since lef1 expression was strongly reduced in dmbx1 double morphant embryos, I
predicted that the expression of Wnt-reporter might also be reduced in dmbx1 double morphants.
At 24hpf, Wnt activity was concentrated in the medial region of the optic tectum in un-injected
embryos (Figure 4.13A). However, expression of GFP was absent in double dmbx1 morphants
(Figure 4.13B). Similar outcomes were consistently observed at 48hpf, as no GFP expression
was detected in dmbx1 morphants (Figure 4.13D). Since expression of the transgene relied on
the expression of Lef1, this verified that dmbx1 genes were upstream of lef1 in the canonical
Wnt signaling cascade. Another study also confirmed the relationship between dmbx1 and lef1
by showing that there was no change in dmbx1a expression when lef1 was knocked down by
morpholino (Bonkowsky et al., 2008), suggesting that dmbx1 were upstream regulators of lef1.
145
Figure 4.13 Expression of Wnt signaling components in dmbx1 double
morphants embryos.
Coronal sections of Tg(TOP:dGFP) embryos at 24hpf (A-B, n=3 in each group) and 48hpf (C-
D, n=3 in each group). GFP positive cells suggested that cells are activated by Wnt signaling. In
un-injected embryos these cells are localized in the dorsal medial region at 24hpf (A) and
scattered within the optic tectum at 48hpf (C). Knock down of dmbx1 genes lead to severe loss
of activated Wnt signaling in the tectum at 24hpf (B) and 48hpf (D). However, wnt3a
expression remained the same between un-injected and MO1a+1b-injected embryos (E-H).
Lateral views, anterior to the left at 24hpf (E-F). Coronal sections are shown for 48hpf embryos
(G, H), and the number of embryos examined is indicated at the bottom left. β-cat, a
downstream component of Wnt in the signaling cascade, expressed within the tectal lobes in un-
injected embryos (I, n=6) but it was lost in the dmbx1 morphants embryos (J, n=6). The Frizzled
receptor, fzd3a, did not show any change in expression in the optic tectum between the two
groups (K-L). Coronal sections are taken from wholemount in situ embryos, and the number of
embryos examined is indicated at the bottom left. Red arrow indicates where the optic tectum is.
146
FIGURE 4.13
147
4.2.10 Dmbx1 is positioned downstream of the Wnt ligand
To further investigate the relationship between dmbx1 and the Wnt signaling pathway, several
other components in the Wnt signaling cascade were examined. Given that Wnt was the source
of this signaling cascade, it would be informative to check for any changes in wnt1/3a
expression when dmbx1 was knocked down. Wholemount in situ hybridization with wnt3a
riboprobe revealed no obvious change in expression at the midline of the dorsal alar plate in
double dmbx1 morphants (Figure 4.13F, 4.13H) when compared to the control (Figure 4.13E,
4.13G). Since the expression of wnt3a was unaffected, it placed dmbx1 downstream of this Wnt
ligand in the canonical pathway. However, other canonical Wnt signaling molecules such as
Wnt1 may interact with Dmbx1 to induce this ectopic midbrain phenotype.
If dmbx1 was downstream of Wnt ligands but upstream of lef1, then dmbx1 was likely to
interact with β-cat. I performed immunostaining with β-Cat antibody on dmbx1 morphants and
found that expression of this protein was absent (Figure 4.13J). At 48hpf, most of β-Cat
expression was present in the central tectal lobe in un-injected embryos (Figure 4.13I) as
expected given that Wnt morphogens released from the medial neurogenic region diffuse across
the tectal lobes to stabilize β-cat in order to turn the Lef1 transcriptional complex from a
repressor into an activator (Porfiri et al., 1997). This result suggested that dmbx1 disrupted the
Wnt pathway before the Wnt ligands could stabilize β-cat in the nucleus, possibly at the receptor
level. One potential receptor for the Wnt3a based on wholemount in situ expression pattern is
Frizzled3a (Fzd3a), which is expressed in the midbrain during early development (Nikaido et
al., 2013). However, detail examination of fzd3a expression at 48hpf revealed that the gene
localized more strongly in the ventral midbrain instead of the dorsal tectum, but there was no
apparent changes between un-injected and MO1a+1b-injected embryos (Figure 4.13K-L).
Further study on which Frizzled is the receptor for Wnt1/3a may help to address how dmbx1
148
interacts with Wnt signaling pathway during tectal neurogenesis as most of the major
downstream components are missing.
4.2.11 Dmbx1 induces ectopic midbrain structure
To investigate whether misexpression of dmbx1a and dmbx1b together might have any
overexpression phenotype in the tectum, I injected 150pg dmbx1a and 188pg dmbx1b mRNA
(optimized dosage determined in chapter 2) into 1-cell stage embryos. At 48hpf, about 70% of
the injected embryos looked similar to un-injected controls, whereas 30% of the dmbx1
overexpressing (OE1a+1b) embryos were dorsalized. However, I noticed that in about 10-15%
of those embryos with relatively normal tectum, they had an ectopic structure grown from the
midline within the tectal region (Figure 4.14B). Based on where this extra tissue was located, it
could potentially be an ectopic tectal lobe. To investigate this further, wholemount in situ
hybridization for lef1 was performed on embryos overexpressing dmbx1a and dmbx1b. The
positive staining of lef1 in the ectopic structure suggested that it could be an additional tectal
lobe (Figure 4.14D).
To examine this phenotype more closely, OE1a+1b embryos were sectioned and it appeared that
the additional tissue extended ventrally from the dorsomedial region into the ventricle in 48hpf
embryos (Figure 4.15C). This symmetrical tissue was situated perfectly at the midline and
persisted at least until 72hpf (Figure 4.15F). This gain-of-function phenotype suggested that
dmbx1 might be able to induce cells around the midline to take on a tectal fate. To further
confirm that the ectopic structure had tectal but not isthmic characteristics, I performed in situ
hybridization with known tectal (otx2) and MHB (pax2a) markers on dmbx1 overexpressing
embryos. Coronal sections of stained embryos showed that the extra tissue at the midline was
149
Figure 4.14 Misexpression of dmbx1a and dmbx1b mRNA at 1-cell stage leads
to an ectopic structure at the dorsal midbrain in a small percentage of
injected embryos at 48hpf.
All panels are dorsal views, anterior is at the top. Light view of un-injected (A) and dmbx1
overexpressed embryos (B) at 48hpf. Ectopic structure is situated at the midline within the
midbrain region, as indicated with red arrow (B). This structure is also expressing lef1 (D),
which is a tectal marker. This suggested that overexpressing dmbx1 genes can lead to an ectopic
structure that may be a potential tectal lobe.
150
Figure 4.15 Coronal sections of loss- and gain-of-function tectal phenotypes
at 48hpf and 72hpf.
Cryosections are stained with DAPI. Tectum in un-injected embryos at 48hpf and 72hpf (A, D)
were symmetrical and the two lobes were right beside one another, and the minimal space
between them is the ventricle. In dmbx1 morphants, tectum size was reduced (B, E). Embryos
overexpressing with dmbx1a and dmbx1b occasionally processed an ectopic structure at the
midline of the tectal region (C, F). White dotted line marks the margin of a tectal lobe. White
arrow indicates where the ectopic structure is.
151
positive for otx2 but not for pax2a (Figure 4.16B, 4.16D). Overall, the results indicated that cells
within the ectopic structure had tectal identity.
4.2.12 Hyperactivated Wnt signaling may induce the formation of ectopic
tectal structure
Given the mitogenic property of Wnt in the mesencephalon and the expression of Wnt1/3a at the
midline of the dorsal midbrain, I hypothesized that the ectopic structure induced by dmbx1
overexpression might be caused by hyperactivated Wnt signaling at the center of the midbrain
alar plate. Since this extra tectal tissue came from the midline where dmbx1a, dmbx1b, and wnt
are normally expressed at 48hpf, I examined if this ectopic structure was responsive to Wnt
signaling. Indeed, in OE1a+1b-injected Tg(TOP:dGFP) embryos, this ectopic structure was
GFP-positive (Figure 4.17B), suggesting that this tissue was responsive to active Wnt-signaling.
However, GFP expression was not present in the nucleus like those in the original two tectal
lobes. Instead, the expression resembled the two small patches of GFP-positive cells located
bilaterally at the roof plate in the control embryos (Figure 4.17A). This led us to further
investigate whether wnt expression might be altered at the dorsal midline. The initial hypothesis
was that the ectopic tectal structure originated from the midline, possibly another miniature
neural tube developed within the fourth ventricle. However, the wnt3a in situ hybridization
showed that the expression of wnt was split and flanked the ectopic structure (Figure 4.17D).
The result ruled out the previous hypothesis as to how the ectopic structure was formed. A
modified hypothesis suggested that the two Wnt signaling centers in the dorsal midbrain region
led to a mirror duplicated tecta along the midline, and the tectal tissue near the midline formed a
continual structure as the middle lobes. Closer examination of the GFP expression in OE1a+1b-
injected Tg(TOP:dGFP) embryos showed that there were two patches of GFP-negative zones
152
Figure 4.16 Cells in the ectopic midbrain structure process tectal identity.
In situ hybridization was performed on un-injected (A, C) and dmbx1 overexpressed embryos
(B, D) at 48hpf using otx2 and pax2a probes. Coronal sections are taken from wholemount in
situ hybridization embryos, and the number of embryos examined is indicated at the bottom
right. The tissue (white arrow) situated at the midline was otx2-positive but pax2a-negative,
suggesting that it has tectal identity and not part of the MHB.
153
Figure 4.17 Expression patterns of Wnt signaling components in dmbx1-
overexpressing embryos at 48hpf.
Coronal sections of Tg(TOP:dGFP) embryos at 48hpf where GFP positive cells are present in
the ectopic structure (B), suggesting that those cells are activated by Wnt signaling (n=6 in each
group). White arrow indicates the cytosolic GFP expression close to the midline, which is
different from the nucleus GFP expression within the optic tectum. Wnt3a expression split into
two in dmbx1 overexpressing embryos (D). Coronal sections are shown and the number of
embryos examined is indicated at the bottom right. Wholemount embryos are shown at the
upper left in dorsal view, anterior at the top. β-catenin expresses in the middle of the ectopic
structure, as well as on both sides where the new duplicated midlines are and within the lateral
tectal lobes (F) (n=6 in both un-injected and dmbx1-overexpressing groups respectively). A
downstream component of Wnt signaling cascade, lef1, expresses within the tectal lobes in the
un-injected embryos (G), and it is also expresses in the ectopic structure in dmbx1-
overexpressing embryos (H). Frizzled receptors, fzd3a, appears to express between the tectal
lobes in un-injected embryos (I) and its expression remains in between those tectal lobes on both
sides and the ectopic structure at the midline in dmbx1-overexpressing embryos (J). Coronal
sections are taken from wholemount in situ embryos in G-J, and the number of embryos
examined is indicated at the bottom right. Yellow arrowhead indicates where the new midline is.
154
FIGURE 4.17
155
between the ectopic tissue at the midline and the tectal lobe on each side, and those two GFP
negative regions probably coincided with the two spots where wnt3a were expressed.
To verify the new hypothesis further, several downstream components of the canonical Wnt
pathway were examined. Both β-cat and lef1 were expressed in the ectopic structure in addition
to their normally domains in the tectal lobes (Figure 4.17F, 4.17H). Expression of fzd3a was
localized between the ectopic tectal tissue and the other tectal lobes in dmbx1-overexpressing
embryos (Figure 4.17J). All these results suggested that the tissue sandwiched between the two
lateral tectal lobes was responding to Wnt signal normally.
Since Wnt is a known mitogen factor (McMahon and Moon, 1989; Megason and McMahon,
2002), it was possible that this ectopic structure was a consequence of too much Wnt signaling.
Both BrdU and PHH3 immunostaining showed that the majority of cells in the ectopic tectum
were still cycling (Figure 4.18B, 4.18D). Moreover, excess Wnt signaling has been shown to
lead to upregulation of cyclinD1, which in turn may cause overproliferation (Megason and
McMahon, 2002). I examined the expression of ccnd1 in OE1a+1b embryos and found that at
72hpf, ccnd1 expression was restricted to the medial region of the dorsal midbrain in control
embryos but its expression remained high in the ectopic tectum of dmbx1-overexpressing
embryos (Figure 4.18E-F). Using HuC to label post-mitotic neurons also confirmed that cells in
the ectopic tectal structure were still mitotic at 48hpf (Figure 4.18H). It is possible that these
cells in the central tectum were developing more slowly than the two normal tectal lobes on both
sides, and they would become post-mitotic and differentiate into tectal cells later during
development.
156
Figure 4.18 Cells in the ectopic midbrain structure have acquired tectal cell
fate but they are still proliferative.
Coronal cryosections stained with various mitotic and differentiated neuronal markers as stated
on the left. It appears that the ectopic structure is developing slower than the lateral tectal lobes
in dmbx1-overexpressing embryos. At 48hpf, cells within the ectopic structure are mitotic as
indicated by the positive staining of PHH3 (B) and BrdU (D) (n=9 in each un-injected and
dmbx1-overexpressing group). By 72hpf, ccnd1 expression still remains high in the ectopic
structure (F), suggesting that those cells are still undifferentiated compared to the rest of the
tectum. Coronal sections are taken from wholemount in situ embryos (E-F), and the number of
embryos examined is indicated at the bottom right. To confirm the lack of differentiated neurons
in the ectopic tectal structure, HuC staining on cryosection was performed. Result suggests that
central region may differentiate first but overall, cells within the extra tissue are not post-mitotic
at 48hpf (n=3 in each group). White arrow indicates where the ectopic structure is.
157
4.3 Discussion
4.3.1 Dmbx1 is dispensable for global brain patterning
In this chapter, the role of dmbx1a and dmbx1b in tectal development was examined more in
depth, with a focus on the molecular mechanism by which they regulated tectal growth and cell
specification. My data demonstrate that the loss of optic tectum in the dmbx1 morphants did not
affect the overall patterning of the brain, and very minor changes were observed in the
expression of the dmbx1 genes when patterning was disrupted in the animals. Many studies have
shown that the development of midbrain and hindbrain are regulated by signals originating from
the MHB during embryogenesis (Cavodeassi and Houart, 2012; Imai et al., 2009; Langenberg
and Brand, 2005), thus one would predict that genes expressed within the MHB might be
targeted by the dmbx1 paralogs to regulate the development of both mid- and hindbrain
simultaneously. However, my data revealed that MHB markers were normal regardless of the
level of dmbx1 paralogs present in the brain. On the other hand, in the absence of MHB (as in
the ace-/-
mutants), expression of the dmbx1 genes were initiated properly. These data implied
that the initiation and maintenance of dmbx1 genes and MHB genes are independent of one
other. Instead, the role of the MHB appears to be to restrict the tectal domain from extending
towards the hindbrain, as the experiment from ace-/-
mutants revealed the caudal expansion of
dmbx1a expression at 24hpf. With the MHB acting as a barrier, patterning of the hindbrain
region was normal in the dmbx1 double morphants as shown from egr2b expression.
Dmbx1 paralogs also appeared to have no patterning effect in the forebrain. Expression of the
forebrain marker, dlx2a, was relatively normal. This indicates that the size reduction of optic
tectum in the midbrain was not due to the concomitant expansion of the forebrain. In addition,
ventral markers remained unchanged in MO-injected embryos. Moreover, when the major
ventral patterning signaling cascade was disrupted by blocking Shh activity with a drug called
158
cyclopamine, there was no change in dmbx1b expression and only a minor increase in dmbx1a
expression at 24hpf. Taken together these data indicate that dmbx1a and dmbx1b are not major
patterning genes for the midbrain. Overall, the data indicated that the dmbx1a was more
sensitive to any changes along the dorsoventral or rostrocaudal axes during initial brain
patterning compared to dmbx1b. However, both dmbx1 genes can initiate and maintain their
expression in the tectal domain even though patterning within the midbrain is compromised.
This suggests that dmbx1 is solely required for tectal neurogenesis during midbrain
development.
4.3.2 Dmbx1 is required for tectal development and neuronal specification in
the hindbrain
Wholemount in situ hybridization assays provided strong evidence that only tectal and hindbrain
genes are disturbed in the dmbx1 double morphants, consistent with the expression domains of
dmbx1. In addition to the significant growth reduction in the optic tectum at 48hpf when dmbx1a
and dmbx1b were knocked down, specific loss of tectal markers and the disorganized expression
of hindbrain neuronal genes both suggested that the role of dmbx1 in the brain was to regulate
tectal development and specify differentiation of certain subtypes of neurons in the hindbrain.
Data from chapter 2 showed that tectal size in dmbx1 double morphants had decreased by half
when compared to the controls. If dmbx1 genes solely affected tectal growth, one would expect
the remaining tectal cells in the morphants to still undergo differentiation normally. However,
my results indicated that the MO-injected embryos had apparent losses of several tectal markers,
suggesting that dmbx1 genes might have an additional role in promoting tectal differentiation
besides just regulating cell growth.
Analyses with different hindbrain markers also revealed that dmbx1 paralogs were required for
neuronal differentiation in the hindbrain. Initial expression of dmbx1a and dmbx1b coincided
159
with the onset of neurogenesis in the hindbrain. Several types of hindbrain neurons were
affected in the dmbx1 morphants, including interneurons and neural crest cells. However, the
mechanistic details of how dmbx1 paralogs regulated this region had not been well-explored. In
Ciona, it was shown that Dmbx (ortholog of Dmbx1) labels a specific subtype of interneuron
(A12.239) in the visceral ganglion, which are cholinergic neurons that innervate the tail muscle
to facilitate swimming (Stolfi and Levine, 2011). This A12.239 neuronal pair had been
characterized to be the equivalent of the Mauthner neurons in teleosts (Stolfi and Levine, 2011),
suggesting that this gene might be important for escape behaviours (Koyama et al., 2011). More
experiments are needed to confirm the types of hindbrain neurons that dmbx1a and dmbx1b are
expressed in, and it will be informative to characterize their roles in hindbrain neurogenesis.
Furthermore, it will be interesting to know how the neural circuitry is affected in dmbx1
morphants since there are many axon projections between the midbrain and the hindbrain.
Malformation of the optic tectum may also affect the wiring between tectal neurons and many
other neurons in different regions of the central nervous system, which in turn may alter any
visual or motor behaviors. Thus, it is important in the future to examine whether the neural
circuit is different between morphants and controls.
4.3.3 Fewer apoptotic cells in the optic tectum of dmbx1 morphant embryos
Loss of tectal tissue in the dmbx1 double morphants was not due to increased cell death. When
axons of the retinal ganglion cells arborized to their target sites in the optic tectum, some of
these neurons may overshoot and need to undergo axonal pruning, a process by which incorrect
connections will be terminated through programmed cell death in order to refine the topographic
map (Ichijo, 1999; Low and Cheng, 2006; Nakamura and O’Leary, 1989). In zebrafish, it was
reported that the peak of cell death in the optic tectum was around 60hpf, which coincided with
the timing in which these retinotectal projections arrived at their tectal targets (Cole and Ross,
160
2001). Data from chapter 3 showed that retinotectal projections were compromised in MO-
injected embryos and the axons were unable to reach to the optic tectum, thus, it is reasonable
that very few cell death events were detected in the morphants’ tecta.
4.3.4 Dmbx1 acts as a transcriptional repressor to regulate cyclinD1 level in
tectal progenitor cells
Results from studying various cell cycle markers suggested that knocking down dmbx1 genes
led to cell cycle defects in tectal progenitor cells. Although detailed cell cycle analysis with flow
cytometry and cumulative BrdU labeling has not been performed on tectal cells from dmbx1
morphant embryos, the fact that these progenitor cells within both retinal and tectal regions had
very similar cell cycle defects suggested that tectal cells might also progress through cell cycle
more slowly compared to controls. Moreover, overexpressing dmbx1a and dmbx1b could lead to
premature cell cycle exit in tectal progenitors, strengthening support for a role of dmbx1 in cell
cycle regulation.
There are many examples in the literature of transcription factors that are cell cycle regulators,
especially homeobox genes (Klein and Assoian, 2008). Rx1 is a paired-typed homeodomain
transcription factor that has been shown to positively control ccnd1 transcription level (Del
Bene and Wittbrodt, 2005). Six3 can regulate ccnd1 and p27 in the anterior neural plate (Gestri
et al., 2005). Dmbx1, on the other hand, is the first transcription factor from the paired-type
homeodomain family that has been shown to be able to repress cyclinD1. My data provides
evidence that Dmbx1 is a novel negative regulator of the cell cycle, as expression of ccnd1 was
de-repressed in the dmbx1 double morphants. Previous work from Megason and McMahon
demonstrated that ectopic expression of ccnd1 was not sufficient to cause an overgrowth in the
neural tube of a chicken embryo (Megason and McMahon, 2002), instead the accumulation of
ccnd1 expression leads to prolonged G1 to S transition (Das et al., 2009). On the other hand,
161
overexpressing dmbx1 reduces ccnd1 expression, which resulted in shorten cell cycle length and
premature cell cycle exit. Another strong piece of evidence that ccnd1 is downstream of Dmbx1
came from the triple morphants data which illustrated how lowering cyclinD1 transcript levels
can facilitate tectal progenitor cells to become post-mitotic and differentiate properly. Without
the Dmbx1 paralogs, ccnd1 is de-repressed and this leads to prolonged cell cycle in the tectal
progenitor cells, resulting in fewer rounds of cell division and lack of differentiation.
4.3.5 Dmbx1 is synergized with the canonical Wnt pathway
Canonical Wnt signaling has been implicated in midbrain development in mouse, chicken, frog
and zebrafish (Bally-Cuif and Wassef, 1994; Buckles et al., 2004; Kunz et al., 2004; Thomas
and Capecchi, 1990; Wolda et al., 1993). Loss- and gain-of-function studies on several major
components of the pathway have revealed the consequences of perturbing the pathway. In
zebrafish, zygotic mutant embryos apc-/-
have hyperactivated Wnt signaling which resulted in
increased cell death in the tectum (Paridaen et al., 2009). Knock down of Nemo-like kinase
(which phosphorylates Lef1 to mediate transcriptional activation in tectal progenitor cells) or
lef1 itself in zebrafish embryos also resulted in a smaller tectum similar to wnt1 morphants (Ota
et al., 2012). The loss- and gain-of-function assays of dmbx1 genes revealed that several major
Wnt signaling components were affected. It will be informative to identify where dmbx1
paralogs are positioned in relation to the overall canonical Wnt signaling pathway. A previous
study in mouse has reported a similar tectal phenotype in Fzd3-/-
; Fzd6-/-
double mutant (3 out of
7 embryos), in which case there was a massive tectal tissue outgrowth in the third ventricle
around E12.5 (Stuebner et al., 2010). It appeared that cells in the protruding structure expressed
Ccnd1 but were devoid of any post-mitotic markers (Stuebner et al., 2010), which suggested that
they were likely progenitor cells, similar to the dmbx1-overexpressing embryos. The tectal
phenotype observed in Fzd3-/-
; Fzd6-/-
mutant mice strongly resembles the ectopic midbrain
162
phenotype observed in dmbx1-overexpressing zebrafish embryos. Thus, the possibility that the
Dmbx1 proteins may be interacting with Frizzled near the midline in the optic tectum is very
appealing. Several frizzled genes (fzd3a, fzd6 and fzd10) are potential Wnt3a receptors based on
their expression patterns in the brain (Nikaido et al., 2013; Sisson and Topczewski, 2009);
however, more careful examination of fzd3a expression pattern revealed that this gene is not
expressed in the optic tectum. Thus, Fzd10 becomes the top candidate receptor for Wnt3a since
it is expressed bilaterally along the midline in the tectum (Nikaido et al., 2013). Cell culture
studies and in vivo experiments on Xenopus embryos have demonstrated that overexpression of
mouse Frizzled-1 can antagonize Wnt3a induced Wnt/β-cat canonical signaling by reducing
transcriptional activity and blocking translocation of β-cat to the nucleus (Roman-Roman et al.,
2004). dmbx1 morphants embryos can also block Wnt transduction events, since expression of
β-catenin and lef1 were both reduced, yielding similar phenotypes as animals misexpressing
frizzled. Hence, it is possible that Dmbx1 negatively regulates frizzled when canonical Wnt
signaling needs to be down-regulated. Thus it will be of interest to determine if dmbx1 can
influence the expression of fzd10 in the optic tectum.
As the embryo develops, wnt3a expression becomes more restricted to the midline and hence
some Wnt target genes (such as cyclin D1) are expressed close to the midline only. This also
allows proper neuronal differentiation in the rest of the tectal lobes further from the source of
Wnts. The wnt3a in situ hybridization data revealed that Dmbx1 transcription factors act
downstream of the Wnt ligand. But with dmbx1a and dmbx1b overexpressed at 1-cell stage,
embryos could be found occasionally with the extra structure at the tectal region given the
correct spatiotemporal amount of Dmbx1 proteins which triggered the threshold, and these
embryos also have ectopically induced expression of wnt3a at the dorsal midbrain. Although the
mechanism of how wnt3a split into two signaling centers remains unresolved, the revelation of
163
duplicated midlines may help explain this ectopic tectum phenotype. This shift in midline
position from medial to bilateral alters the way tectal progenitor cells responded to the canonical
Wnt signaling pathway. In the study on Fzd3-/-
; Fzd6-/-
double mutant mice, the authors did not
look at the expression of Wnt3a in those embryos. Thus, it would be interesting to confirm if the
comparable midbrain phenotype between dmbx1-overexpressing animals and Fzd3-/-
; Fzd6-/-
double mutant is due to bifurcation of the Wnt expression in the tectal region.
Since Wnts are mitogenic, the duplicated Wnt signaling centers increased the number of cells in
the optic tectum that are likely to have activated canonical Wnt signaling. The overgrown tissue
around the dorsal midline appears to be a mirror image of two regions that are close to the
midline (see Figure 4.19). Because the tissue develops within a confined space, it folds inward
(as a teardrop shape) into the ventricle as it proliferates. My hypothesis is that the structure in
the middle of the tectum actually begins to be established between 24-48 hpf, when cells within
the ectopic structure receive ample amount of Wnt signaling and they are highly mitogenic.
However, since they develop a day later than the original two tectal lobes, cells within the extra
tissues are still undergoing mitosis (with many BrdU-, PHH3- and ccnd1-positive cells) and lack
any differentiated neuronal maker such as HuC. Further examination of the relationship between
dmbx1 paralogs and canonical Wnt signaling components will allow us to hypothesize whether
dmbx1 helps define the position of the tectal midline(s). Moreover, identifying the potential
molecular mechanism by which dmbx1 paralogs utilize to regulate the canonical Wnt signaling
pathway in the optic tectum.
164
Figure 4.19 Schematic diagram of how dmbx1 regulates the development of
the optic tectum.
Diagram represents the coronal view of the optic tectum from a 48hpf embryo. In un-injected
embryos, the dorsoventral region near the midline (blue colour) is where the mitotic progenitor
cells (red colour) are. Some of these mitotic progenitors undergo terminal differentiation and
they become tectal precursor cells (green colour). When the appropriate signals for differentiate
are present, these cells will develop into differentiated tectal neurons (yellow colour) in the
tectal lobe. In dmbx1 double morphants (MO1a+MO1b), many of these progenitor cells have
cell cycle defects that lead to decrease in proliferation and failure to exit cell cycle on time. On
the other hand, a small portion of the dmbx1 overexpressing embryos give rise to an ectopic
structure in the midbrain due to the duplicated midlines at the tectal region. Cells in the extra
tissue are mixture of progenitors and precursors cells.
165
4.3.6 Low penetrance of ectopic tectal structure in dmbx1-overexpressing
embryos
In a given batch of injected embryos overexpressed with dmbx1a and dmbx1b, the number of
animals with the ectopic tectal structure was consistently low. Although the penetrance of this
phenotype was low, it was highly replicable. In addition to the rarity of this phenotype, the
expressivity of this trait was also variable which can be detected through the irregular size of
this midbrain structure found in a portion of the dmbx1-overexpressing embryos. However, the
dorsomedial position of the alar plate where this extra tissue is originated remained the same
across embryos that carry the ectopic tectal structure. Furthermore, this usual anatomic structure
is virtually symmetrical at the midline of the neural tube.
It is not entirely clear why the penetrance of this trait is so low. My hypothesis is that in order to
generate this ectopic structure, the excess amount of Dmbx1 has to be over the threshold level at
the dorsal anterior neural tube from 24 – 48 hpf. However, the setup of my misexpression
experiment makes it difficult to spatially control where the dmbx1 mRNA injected at 1-cell
stage will localize in the embryo at 24hpf. In addition, the amount of mRNA that can be injected
initially is limited by the fact that too much Dmbx1 leads to severe dorsalization of the embryo.
Thus, the level of Dmbx1 remains in the midbrain region by 24hpf may not be sufficient to elicit
the bifurcation of the Wnt signaling center to produce an ectopic tectal structure.
4.3.7 Testing the functionality of the ectopic tectal structure
So far my data indicates that by 2-3dpf, the majority of the cells within the ectopic structure are
still undergoing mitosis and have not exited the cell cycle. It would be interesting to examine
whether these cells will eventually become post-mitotic and differentiate into mature tectal
neuron in dmbx1-overexpressing embryos at 4-5 dpf. A quick and easy way to distinguish tectal
progenitor cells versus mature tectal neurons is to compare the expression pattern of GFP-
166
positive cells in the Tg(TOP:dGFP) transgenic line. In un-injected embryos, it is obvious that
progenitor cells close to the midline have membrane-like GFP expression, whereas
differentiated neurons at the center of the tectal lobes have puncta-like GFP expression pattern.
By overexpressing dmbx1 paralogs in Tg(TOP:dGFP) embryos and monitoring how the GFP
expression pattern changes within the ectopic structure will help deduce the differentiation
status of these cells. It is not entirely clear why there are two sets of GFP positive cells in the
Tg(TOP:dGFP) fish and how their patterns change depending on the level of commitment a cell
is at within the neuronal lineage inside the tectum. It has been shown before that the TOP:dGFP
construct in other organisms has behaved inconsistently, and that the expression of GFP does
not truly reflect the activity of Wnt as there are many negative or false positive GFP expression
patterns reported from different transgenic animals made from this construct over the years
(Barolo, 2006; Shimizu et al., 2012). In Tg(TOP:dGFP) fish, this construct is driven under
minimal c-fos promoter with four Tcf/Lef-binding sites (Dorsky et al., 2002). Depending on
where this transgene is integrated in the genome, it is possible that the chromatin status of
progenitor cells and mature neurons may affect how the GFP expression changes within a cell.
Nonetheless, this assay provides a quick and preliminary assessment of how these cells
transform within the ectopic structure and displays a dynamics observation for the system.
Additional immunohistochemistry performed with different neuronal markers on these dmbx1-
overexpressing embryos at 4-5 dpf will help verify the identities of these cells within the ectopic
structure.
It would also be interesting to know in the future if these cells within the ectopic structure do
become mature tectal neurons or not, and whether this extra tissue will affect the overall
retinotectal topography and visual ability of those animals overexpressed with dmbx1 paralogs.
Retinal ganglion axons send out their projections by 36hpf and will not reach the rostral tip of
167
the optic tectum until 45hpf (Stuermer, 1988). Since cells in the ectopic structure develop later
than those within the lateral tectal lobes on both sides, the timing of arborization and any
changes in the topographic map of these retinotectal projections will provide insights into the
potential function of this ectopic structure. To investigate this, one can inject dmbx1a and
dmbx1b mRNA into Tg(isl2b:GFP) embryos, which expresses GFP in the retinal ganglion cells.
These GFP positive cells will then send out axons to their appropriate target regions in the brain.
By examining the retinotectal projections over time in dmbx1-overexpressing transgenic
embryos, one can make observations on whether the RGC axons will branch off and connect to
the ectopic structure at the center, or they will avoid connecting with the ectopic structure. Last
but not least, it will be worthwhile to test if these embryos ectopically expressing with dmbx1
can still see normally with an ectopic structure in the tectum. Fish with ablated optic tectum
have lost their prey capture ability (Gahtan and Baier, 2004; Gahtan et al., 2005), and thus it will
be fascinating to see if any changes in prey capture activity is observed in these dmbx1-
overexpressing animals and how this extra tectal tissue helps contribute to those behaviours
during development.
168
Chapter 5
Materials and Methods
The contents of this chapter were published as:
Chang, L., Khoo, B., Wong, L. and Tropepe, V. (2006). Genomic sequence and spatiotemporal
expression comparison of zebrafish mbx1 and its paralog, mbx2. Development Genes and
Evolution 216, 647–654.
Wong, L., Weadick, C. J., Kuo, C., Chang, B. S. and Tropepe, V. (2010). Duplicate dmbx1
genes regulate progenitor cell cycle and differentiation during zebrafish midbrain and retinal
development. BMC Developmental Biology 10, 100.
169
Materials and Methods
5.1 Zebrafish husbandry
Adult zebrafish (Danio rerio) used in this study were maintained at 28oC on a 14-hour light/10-
hour dark cycle and housed in an automated re-circulating system (Aquaneering). Embryos were
staged as described in Kimmel et al. (Kimmel et al., 1995) and reared according to standard
procedures (Westerfield, 2000). The wildtype strain used was AB (Zebrafish International
Resource Center). Tg(HuC:Kaede) [also known as Tg(elavl3:Kaede)rw0130a
] and
Tg(isl2b:GFP)zc7
transgenic strains were kind gifts from Dr. Hitoshi Okamoto and Dr. Chi-Bin
Chien respectively. Tg(TOP:dGFP)w25
and Tg (β-actin-mGFP) [also known as Tg(Ola-
Actb:Hsa.HRAS-EGFP)vu119
] fish were provided by Dr. Ashley Bruce (generated in the Moon
Lab and the Solnica-Krezel Lab respectively). The acerebellar (ace-/-
) mutant strain (genotype
fgf8ati282a/ti282a
) was generated in the Nüsslein-Volhard Lab and was obtained through Dr. Ian
Scott.
5.2 Morpholino injections
Antisense morpholinos (MOs) were obtained from Gene Tools, LLC. dmbx1a-MO and ccnd1-
MO were described previously and the suggested dosages were used (Duffy et al., 2005;
Kawahara et al., 2002).The dmbx1b-MO was designed to target sequences upstream of the
dmbx1b start codon in the 5-UTR, and the working concentration was determined by a series of
dosage-response experiments. The sequences of the MOs are as follows, dmbx1a-MO (MO1a):
5-ACTCCGTAGTGCTGCATGATTCACA-3, dmbx1b-MO (MO1b): 5-
TCGAGCTTCTCTCTGGGAAGTTTTG-3, and cyclinD1 MO (MOccnd1): 5-
ACTGGTGCTCCATATCTTCA-3. 3-5 nucleotides mismatched morpholinos were also
synthesized for dmbx1a (mMO1a): 5-ACTgCGTAcTGCTcCATcATTgACA-3, dmbx1b
(mMO1b): 5-TCcAGCTTgTCTgTGcGAAcTTTTG-3, and ccnd1 (mMOccnd1): 5-
ACTGGTaCTCtATATaTTCA-3 as controls. Unless otherwise noted, embryos were injected
with 10 ng of a single MO1a/MO1b and 20 ng of MOccnd1 alone, or 5 ng each of the combined
MO1a and MO1b and 2.5 ng of MOccnd1 into the yolk at 1- to 2-cell stages.
170
5.3 GFP fusion proteins
Primers were designed to flank part of the 5-UTR and N-terminal domains of dmbx1a and
dmbx1b that are complementary to the morpholino sequences. The previously described
dmbx1a-MO (MO1a) was complementary to the sequences surrounding the ATG start codon
and the corresponding fusion protein was generated by cloning 70 bp of the 5-UTR and the first
16 amino acids of Dmbx1a in frame with the EGFP gene (Dmbx1a-FP primers: F:5-
CGAGCTAGAAGCAAGAAAATATCA-3 and R:5- GAGTTCATGGCGTGGAGAGAGTA-
3. The dmbx1b-MO (MO1b) targeted the 5-UTR sequences just upstream of the start codon.
The fusion protein consisted of the 99 bp of the 5-UTR plus amino acids 1-16 of Dmbx1b,
followed by the EGFP gene sequences (Dmbx1b-FP primers: F:5-
TGGGAAAAATCACTCGTGTTC-3 and R: 5- GAGTTCATGGCGTGCAAA-3). The PCR
fragments of Dmbx1a-FP and Dmbx1b-FP were cloned upstream and in frame with EGFP in
pCS2+. Plasmids were linearized with BstX1 and in vitro transcribed with the SP6 messenger kit
(Ambion). For each fusion construct, 500 pg of mRNA was injected at the 1-cell stage embryo
in the presence or absence of morpholinos.
5.4 Standard and quantitative real time RT-PCR
For standard RT-PCR, total RNA was extracted from embryos (n=20-40, at 6-, 9-, 11-, 24, 48-
hpf) or adult brain tissue (n=5-10) with Trizol Reagent (Invitrogen). First Strand cDNAs were
reverse transcribed from oligo(dT)12-18 primed total RNA (DNase treated) using SuperScript
III (Invitrogen). Each 25μL reaction consisted of 1× PCR buffer, 1.5mM MgCl2, 0.2mM
dNTPs, 0.4μM each of forward and reverse primers, 0.5U Platinum Taq DNA Polymerase, and
diluted cDNA template (1:100). PCR conditions were as follows: 95oC for 5 minutes, then 35-
40 cycles of 95oC for 30 seconds, 52-58
oC for 30 seconds, and 72
oC for 30 seconds. Annealing
temperatures and cycle number for each primer pair were determined using gradient PCR.
Primer sequences are as follow:
actin-F 5´-AAGCAGGAGTACGATGAGTC-3´ actin-R 5´-TGGAGTCCTCAGATGCATTG-3´
dmbx1a-F 5´-GACAGATGGAGCCCTAGCAG-3´ dmbx1a-R 5´-CTCCTCTTCCTTGTCGGTTG-3´
dmbx1b-F 5´-CCGTCCTCCTTACCTTACCTG-3 ́ dmbx1b-R 5´-ATGGCTCCCTGTTGGTTC-3´
171
Products were resolved on a 1% agarose gel. For real-time RT-PCR I used the Rotor-Gene 3000
Thermal Cycler (Corbett Research). First Strand cDNAs were reverse transcribed from
oligo(dT)12-18 primed total RNA (DNase treated) using SuperScript III (Invitrogen). Each
20μL reaction consisted of 1× PCR buffer, 3mM MgCl2, 0.2mM dNTPs, 0.2μM each of
forward and reverse primers, 0.5U Platinum Taq DNA Polymerase, 0.5× SYBR Green I
(Invitrogen), and diluted cDNA template (1:50). Test samples were carried out in duplicate
including a control reaction with template omitted: 95oC for 5 minutes, 40 cycles of 95
oC for 15
seconds, 60oC for 20 seconds, and 72
oC for 30 seconds. Post-PCR amplification for melt curve
analyses was performed by ramping from 72oC to 99
oC (at 0.2
oC/second) in order to check the
specificity of the amplicons. PCR products were also verified on a 1% agarose gel to ensure
correct amplification sizes were obtained. Quantification of relative gene expression was
calculated with Rotor-Gene Software, version 6.0 (Corbett Research).
5.5 Whole-mount in situ hybridization
Embryos treated with 0.003% of 1-phenyl-2-thiourea (Sigma) were fixed in 4%
paraformaldehyde and kept in methanol before performing in situ hybridization. Samples were
treated with proteinase K according to their ages and hybridized with ~100ng antisense
DIG/FITC-RNA probes overnight at 65oC. Excess probes were washed off the next day and
embryos were incubated with anti-DIG/FITC antibodies (1:4000, Roche) at 4oC overnight.
Colour reactions were performed by mixing in NBT+BCIP (Roche) as substrates to the samples.
Embryos were then post-fixed and left in glycerol until being processed (see histology section
below for embryos subjected to sectioning). Images of those embryos were captured with a
Leica MZ16F dissecting microscope (whole mounted samples) or a Leica DM4500B compound
microscope (flat-mounted and cryosectioned samples) with a QIMAGING digital camera and
OpenLab software. The following antisense RNA probes were used: foxa2/axial (Strähle et al.,
1993), dlx2a (Akimenko et al., 1994), eng2b/eng3 (Ekker et al., 1992), fgf8 (Reifers et al.,
1998), foxb1.2/mariposa (Moens et al., 1996), egr2b/krox20 (Oxtoby and Jowett, 1993), islet1
(Inoue et al., 1994), lef1 (Dorsky et al., 1999), lim1 (Toyama and Dawid, 1997), pax2a (Krauss
et al., 1991), otx2 (Mercier et al., 1995), shh (Krauss et al., 1993), and wnt3a (Krauss et al.,
1992) (all the probes mentioned above were kind gifts from Dr. Ashley Bruce); erm
(Münchberg et al., 1999) (kind gift from Dr. Herbert Steinbeisser); dmbx1a and dmbx1b;
172
neurod, pax6a, rho, and vsx2 (bought from Open Biosystems); and ccnd1(cb161), ccne2(cb165),
p27kip1(cb611), p57kip2(cb961), fzd3a (ordered from ZIRC).
5.6 Histology
Embryos were fixed in 4% paraformaldehyde and rinsed in phosphate-buffered saline solution
after. For semi-thin sectioning, embryos were first dehydrated using increasing concentrations
of ethanol, followed by embedding with increasing concentrations of spurr’s resin in ethanol.
Embryos were then left to polymerization at 65°C in 100% spurr’s resin. Semithin coronal
sections (approximately 1µm thick) were cut with a glass knife using an ultramicrotome and
dried onto glass slides. This procedure was followed by counterstaining with toluidine blue to
visualize zebrafish morphology. Whole-mount in situ hybridization embryos in 100% glycerol
were washed with PBT and followed by the same embedding and sectioning steps as above.
Sections were 1.5 micrometers thick without counterstaining to maximize visualization for the
NBT+BCIP precipitate. For cryosectioning, 1-6dpf embryos from each group were fixed with
4% paraformaldehyde overnight at 4oC and washed in sucrose series (from 5% to 30% sucrose
in PBS) for cryoprotection. Except for PCNA labeling embryos, which were fixed in 37%
formaldehyde:95% ethanol (3:7 ratio) solution. Samples were left in 30% sucrose:OCT (2:1
ratio) at -20oC before cutting into 14µm/20µm sections with a cryostat.
5.7 Cross-section area measurement
To measure the area of the retina and optic tecta, five plastic sections with similar focal plane
were chosen to represent each embryo, and images were taken on a Leica DM4500B compound
microscope with a QIMAGING digital camera and OpenLab software. The areas of interest
were outlined and measured using the program ImageJ (http://rsb.info.nih.gov/ij/) (Abràmoff et
al., 2004). Results represent the average obtained from at least 5 embryos from each group.
Statistical analyses between injected and un-injected groups were performed using student’s t-
test. Differences were regarded as significant for p<0.05.
5.8 Retinotectal projections
Tg(isl2b:GFP)zc7
larvae treated with 0.003% of 1-phenyl-2-thiourea (PTU) at 4dpf were
immobilized on long coverslip and sandwiched with another smaller elevated coverslip in order
173
to observe their retinotectal projections dorsally. Images were collected using Zeiss LSM 510
inverted confocal microscope.
5.9 Ectopic gene expression
Dmbx1aCDS primers (F:5´-ATGCAGCACTACGGAGTGAA-3´ and R:5´-
TCAGTTGGGCAGTGTGTCC-3´) and Dmbx1bCDS primers (F:5´-
ATGCAGCACTACGGGGTGA-3´ and R: 5´-TTAGTTTGGTAGCGTGTCCAGG-3´)
amplified the full coding sequences of dmbx1a and dmbx1b, respectively, and both mRNAs lack
the 5´-UTR to avoid binding of the corresponding morpholinos. Both PCR fragments were
cloned into pCS2+ and linearized with SacII for in vitro RNA transcriptions using the
mMESSAGE mMACHINE SP6 kit (Ambion). Mouse Dmbx1 mRNA was synthesized from
pCMV6-Kan/Neo plasmid containing the cloned full length mouse Dmbx1 cDNA (OriGene).
The template was linearized with SacII and transcribed using the mMESSAGE mMACHINE T7
kit (Ambion). RNA or RNA+MO were injected into the yolk of 1-cell stage embryos at the
concentrations indicated in text.
5.10 Wholemount antibody staining
10 embryos were fixed in 4% paraformaldehyde at 4oC overnight and washed with PBS.
Samples were incubated with block solution (PBS+1%BSA+1%DMSO+0.8% TritonX-100) for
an hour at room temperature and overnight in c-myc (9E10) antibody (1:20) at 4oC. Embryos
were then washed with PBS+Triton and incubated in goat anti-mouse HRP antibody (1:500)
overnight. Before DAB staining, samples were washed with PBS+Trition followed by PBS
only. Peroxidase was detected with DAB and 3% hydrogen peroxide in the dark. Images of the
stained embryos were taken from Leica MZ16F dissecting microscope and Leica DM4500B
compound microscope.
5.11 Immunohistochemistry
Cryosections were re-hydrated with 1xPBS and blocked for 2 hours in 0.2% Triton X-100 + 2%
goat serum in PBS at room temperature. Primary antibody in block solution was applied on
sections overnight at 4oC. Slides were washed with PBS + 0.1% Tween-20 and incubated with
secondary antibody for 2 hours at 4oC. Nuclei were counterstained with DAPI before mounting
the slides. No staining was performed after cryosectioning Tg(isl2b:GFP)zc7
and Tg(β-
174
actin:mGFP) embryos, but Tg(TOP:dGFP)w25
transgenic line required antibody staining in order
to detect the GFP expression. Anti-GFP antibody (rabbit, 1:100, life technologies™) was
conjugated to Alexa Fluor® 488 so no secondary antibody was required. The following primary
antibodies were used: mouse anti-β-catenin (1:250, BD Transduction Laboratories™), rat anti-
BrdU (1:100, Cedarlane Laboratories, Ltd), rabbit calbindin D-28K (1:200, EMD Millipore),
rabbit cleaved Caspase 3 (Asp175) (1:250, Cell Signaling Technology®), mouse HuC (1:100,
Life Technologies-Novex), mouse anti-PCNA (1:100, ZYMED Laboratories), rabbit anti-
Phospho-histone H3 (Ser10) (1:250, Cell Signaling Technology®), rabbit anti-Pax6 (1:100,
Covance), rabbit anti-PKC (1:100, Santa Cruz Biotechnology, Inc.), mouse anit-GS (1:500,
Chemicon), mouse Zn5 (1:100, ZIRC), mouse anti-Zpr1 (1:200, ZIRC), and mouse anti-Zpr3
(1:200, ZIRC). Secondary antibodies used for detection were all from Jackson ImmunoResearch
Laboratories, Inc.: mouse, rat and rabbit Cy3 (1:500); mouse and rabbit Cy5 (1:200). Images
were taken from Leica TCS SP5 II Confocal Microscope and analyzed with Leica LAS AF
software. For cleaved Caspase 3 and phosphor-histone H3 positive cell count, I obtained three
sections from different embryos of each group and took the average number for comparison.
Statistical analyses between MO-injected and control-injected groups were performed using
student’s t-test. Differences were regarded as significant for p<0.05.
5.12 Cyclopamine treatment
Embryos were soaked in 0.04mM cyclopamine (dissolve in ethanol) at 10 hpf until 24hpf. Drug
was removed by washing embryos with embryo medium (3×5mins washes). Both treated and
untreated embryos were fixed with 4% PFA immediately afterwards for wholemount in situ
hybridization assays.
5.13 Cell death analyses
Embryos from 24 – 72 hpf (n=10) were bathed in embryo media that contained 5µg/mL acridine
orange (Sigma) for 15 min at room temperature, and immediately followed by 410 min washes
using regular media. AO-positive cells were imaged using a Leica DM4500B compound
microscope with a QIMAGING digital camera and OpenLab software. For TUNEL labeling,
embryos at 72 hpf (n=10) were fixed in 4% paraformaldehyde at 4oC overnight followed by
cryoprotection with sucrose and cryosectioning (same as immunohistochemistry). Sections were
rehydrated and TUNEL assay was carried out according to manufacturer's instruction (Apo-
175
Direct TUNEL Assay Kit, Millipore) and images were captured using Leica TCS SP5 II
Confocal Microscope and analyzed with Leica LAS AF software.
5.14 BrdU labeling
To label cells that were in S-phase, 5 mM of 5´-bromo-2´-deoxyuridine (BrdU) was injected
into the tectal brain ventricle of 24 or 72 hpf embryos and fixed the animals (n=5-10) with 4%
paraformaldehyde ½-1 hour later. For cumulative cell cycle analysis using BrdU incorporation,
embryos (n=12) were injected with 5 mM BrdU every two hours up to 10 hours followed by 4%
paraformaldehyde fixation 30 minutes post-BrdU injection. Cryosectioning procedures were
performed as mentioned above. For BrdU-immunostaining, slides were treated with 20U/mL
DNase I at room temperature for 30 minutes followed by extensive washes with PBS+1%
DMSO+0.1% Tween-20 (PBDT). Sections were blocked for two hours and incubated in rat anti-
BrdU primary antibody overnight at 4oC, which was then detected with Cy3 secondary
antibody. Images were obtained from mounted slides using Leica TCS SP5 II Confocal
Microscope and analyzed with Leica LAS AF software. For the cumulative BrdU assay, I
counted the number of BrdU positive cells per section (averaged over at least 3 separate retinas)
and cell density (number of DAPI positive nuclei/area of section) were used to estimate the
number of total nuclei from each section and calculated the labeling index (BrdU positive
cells/total nuclei) at all six time points. Cell cycle kinetics in control and morphant embryos was
determined as described in (Kim and Shen, 2008) assuming this was a single population model.
Briefly, hours of BrdU injection (T) was plotted against the labeling index (LI). Growth fraction
(maximum LI on the y-axis, LIm) can be determined from where the curve begins to plateau.
The time when the maximum amount of BrdU positive cells was labeled is equal to total cell
cycle time (Tc) minus S-phase time (Ts). By extrapolating the curve back to time = 0, the
labeling index at Ts (LI0) can also be determined. With this information, the total cell cycle time
can be estimated using the equation LI0/LIm = Ts/Tc.
5.15 Flow cytometry
Retinas from 60-80 embryos (~2 x 106 cells/mL) were dissected out from sample and control
groups and left in Hank's Buffered Salt Solution (HBSS) with trypsin for 30 minutes. Pellet cells
and discarded supernatant but re-suspended cells in 50μL of HBSS with 2% FBS. Added 1mL
of ice-cold 80% ethanol and kept samples in -20oC for at least 30 minutes. Cells were collected
176
by centrifugation and washed twice with HBSS + 2% FBS. Removed supernatant and added
500μL of 0.1 mg/mL propidium iodide (Sigma) in HBSS with 0.6% NP-40, together with
500μL of 2mg/mL RNaseA and incubated in the dark for 30 minutes. Samples were filtered
through 85μm Nitex mesh filter and analyzed with BD FACSAriaTM
cell sorter. Flow cytometry
data were analyzed using FlowJo software.
5.16 Transplantation
Donor Tg(elavl3:Kaede)rw0130a
embryos injected with dmbx1a and dmbx1b morpholinos or their
corresponding mismatch morpholinos (controls) were labeled with rhodamine at one cell stage,
and donor cells were transplanted to un-injected or dmbx1a+dmbx1b morpholino-injected AB
hosts between 3-4hpf. Only hosts that had rhodamine positive cells in the retina were examined.
Another set of blastula transplantation experiment was carried out between the control/MO-
injected Tg(β-actin:mGFP) donor embryos and un-/MO-injected AB hosts. Host animals that
had GFP positive cells in the retina were cryosectioned for further analysis with various retinal
and cell cycle markers using immunohistochemistry.
5.17 Molecular evolutionary analyses
Maximum likelihood phylogenetic methods were used to estimate the ratio of non-synonymous
to synonymous rates (dN/dS) along lineages (Yang, 1998; Yang and Nielsen, 1998) in a pruned
Dmbx1 phylogeny consisting of a subset of the sequences used for the phylogenetic analysis.
dN/dS ratios can be used to estimate the form and strength of selection operating in the Dmbx1
gene family. Assuming no selection pressure, dN/dS value is equal to on (Kimura, 1983; Yang
and Bielawski, 2000). Positive selection is indicated by dN/dS values greater than one, while
negative selection is indicated by dN/dS values near zero. Codon models that allow for variation
in dN/dS along branches were implemented in the PAML package v4.2a (Yang, 2007).
Likelihood ratio tests were used to determine which among nested models provided a
statistically significantly better fit to the data at hand (Navidi et al., 1991; Yang, 1994).
5.18 Yeast two-hybrid
pBluescript-dmbx1a (full length cDNA) plasmid was sent to Hybrigenics Services (France) for
yeast-two-hybrid analysis on 18-20 hpf zebrafish embryos library. Thirty-two out of sixty-nine
177
clones were pulled out as potential baits for Dmbx1a and the outcomes were summarized based
on the confidence level of their interactions (see appendix).
5.19 Western Blot
Fragments of Dmbx1a (amino acids 225-293) and Dmbx1b (amino acids 170-267) coding
regions were subcloned in-frame downstream of the GST-tag into a pGEX plasmid. Bacteria
(BL21) containing plasmid pGEX-dmbx1a and pGEX-dmbx1b are induced with IPTG and
protein lysates were collected using B-PER® Bacterial Protein Extraction Reagents (Thermo
scientific). Soluble and insoluble fractions were separated and those GST-fusion proteins are
then subjected for purification from other bacterial proteins using commercially available
Glutathione Sepharose® 4B beads (GE Healthcare). Purified proteins were analyzed on 10%
SDS-PAGE gel following standard Western blot protocol.
5.20 Statistical Analyses
Two-tailed non-parametric exact binomial and two-tailed unpaired student t-tests (p<0.05) were
performed to compare whether two samples were significantly different or not.
5.21 Cloning for myc-tagged proteins
Myc-tag at the N-terminal of mouse Dmbx1 was subcloned using EcoRI and XbaI from
pCMB6Kan/Neo-dmbx1 into pCS2+-MT plasmid, whereas Myc-Dmbx1a (zebrafish) was
generated by cloning klenow-treated BamHI+XhoI pCS2+-Dmbx1aCDS into CIAP-treated
(Invitrogen) XbaI digested pCS2+-MT plasmid. RNA sequences of Myc-Dmbx1 and Myc-
Dmbx1a were synthesized by linearizing both plasmids with NotI, and then transcribed them
with the mMESSAGE mMACHINE SP6 kit (Ambion).
178
Chapter 6
General Discussion
179
General Discussion
6.1 Discrepancy between previously reported dsmbx1a phenotypes
and the double dmbx1 morphants examined in this study
Dmbx1a (previously known as Mbx) was first described in zebrafish in Kawahara et al and the
authors had reported in this paper, the expression pattern of dmbx1a during embryogenesis and
the loss of function phenotypes in dmbx1a knocked down embryos (Kawahara et al., 2002). My
work here has confirmed some of the findings from this paper and also enhanced our knowledge
about the dmbx1 gene family when I characterized and compared the paralogs ‒ dmbx1a and
dmbx1b. My expression profile analyses complemented with detail examination using
cryosections from 1-6dpf offers a more comprehensive study of both genes, revealing any minor
differences in their dynamic expressions, as it will help address any divergence between the
paralogs during evolution. Moreover, I have characterized the role of dmbx1 genes during tectal
development in much greater details.
In addition to identify the functional role of dmbx1b in zebrafish, I have also repeated some of
the analyses performed in Kawahara et al on dmbx1a and obtained similar outcomes (Kawahara
et al., 2002). My results have supported the general notion that knocking down dmbx1a at 1-cell
stage leads to smaller retina and tectum in the morphant embryos, but the tissue size in the
hindbrain remains unaffected. Since dmbx1a has a much stronger role in retinal development, I
noticed that some retinal defects were comparable between dmbx1a morphants reported in
Kawahara et al and the dmbx1 double morphants. In both studies, retinal progenitor markers
(rx1, vsx2) persisted in 48-72hpf morphant embryos, but differentiated markers (Zn5, islet1)
were reduced (Kawahara et al., 2002). Furthermore, the retinotectal topology defects were seen
in both morphants, with thinner optic nerves and smaller arborization field in the tectum
(Kawahara et al., 2002).
180
Despite some of the similarities mentioned above, there are several noticeable differences
between my work and the dmbx1a phenotypes reported previously (Kawahara et al., 2002).
First, the expression of optic stalk makers (pax2a and fgf8) had increased in dmbx1 double
morphants (Figure 3.2 ) but it remained the same in Kawahara et al. (Kawahara et al., 2002).
Second, the authors observed an elevated level of TUNEL labeling throughout the anterior
region of the dmbx1a morphants at 15-somite stage (Kawahara et al., 2002). However, my cell
death analyses on dmbx1 double morphants using TUNEL, acridine orange, and activated
Caspase 3 all indicated that there was no increase in apoptotic cells within both retinal and tectal
regions (Figure 3.3 and 4.8). Third, there was no change in the level of PHH3 staining in
dmbx1a morphants at 15-somite stage as shown in Kawahara et al. (Kawahara et al., 2002). But
I detected a strong increase in PCNA, BrdU, and PHH3 staining when both dmbx1a and dmbx1b
were knocked down (Figure 3.4 and 4.9).
One of the possible explanations for this discrepancy may be due to the developmental stages
we had chosen to perform our analyses. In Kawahara et al., the authors performed their cell
death and cell proliferation experiments on much younger embryos (before 24hpf). However, all
my assays were carried out on 48-72hpf embryos. The reason for looking at a later time point
was due to the fact that expression and functional differences between dmbx1a and dmbx1b
were more prominent during those stages. Hence, I continued to characterize the dmbxl double
morphants between 48-72hpf and focused on identifying their functional roles during
neurogenesis. When I examined PCNA staining at 24hpf, I could not detect any differences
between control-injected and dmbx1 double morphant embryos (data not shown). This suggested
that the dmbx1 genes are not involved in primary neurogenesis (completed before 24hpf);
instead they are required for secondary neurogenesis in the brain. Furthermore, the variation in
level of expression obtained from those optic stalk markers can also be explained through age
181
differences. It is known that genes expressed in retina (pax6) and optic stalk (pax2) can mutually
repress one another (Macdonald et al., 1995; Schwarz et al., 2000). Thus, the optic stalk might
have expanded in the dmbx1 double morphants during 24-48hpf since the retina has become
relatively smaller compared to the control embryos overtime.
It has been reported that p53-dependent cell death activation during early development can be a
morpholino off-targeting effect (Robu et al., 2007). Thus, it is possible that increase in cell death
detected in dmbx1a morphants during somite stages in Kawahara et al. might be a nontarget-
related phenotype (Kawahara et al., 2002). Since dmbx1a was not ubiquitously expressed in the
neural tissue but localized to the midbrain and the retina, the reported finding on TUNEL
staining spread across the neural tube in Kawahara et al. suggested that this might be mediated
by morpholino-dependent p53 activation. We cannot rule out the fact that this non-specific cell
death might contribute to part of the tissue size reduction in those visual structures observed in
dmbx1a and double dmbx1 morphants at 48 and 72hpf; however, the lack of pervasive cell death
detected in those morphant embryos during late embryogenesis in my study suggested that the
significant decrease in tectal and retinal tissue size could not be merely attributed from the early
apoptotic events. In the future, this morpholino off-targeting effect should be controlled for by
co-injecting dmbx1 morpholinos with p53 morpholino in order to prevent this early non-specific
cell death, and allowing us to assay those target-specific defects.
6.2 Mechanism underlying the retention of dmbx1 genes during evolution
Gene duplication has been demonstrated to be a driving force for new genes to arise during
evolution (Kassahn et al., 2009; Mazet and Shimeld, 2002; Ohno, 1999; Otto and Yong, 2002).
A small number of duplicated genes are retained in the genome during natural selection usually
as the result of their functions diverging (Kassahn et al., 2009). Throughout evolution, adaptive
182
mutations were accumulated in preserved gene duplicates that led to diversification in their
expression patterns or functions. There are two main types of divergence in the genomic
sequences that can lead to either expression or functional differences between paralogs. One of
these is regulatory-motif divergence where degenerative mutations accumulate in the promoter
region of the gene pair, and leads to expression pattern differences. The other type of divergence
is in the coding region, which results in distinct protein functions between the two paralogs.
From analyses of gene duplication, two different but not mutually exclusive evolutionary
models have been proposed to address the retention mechanisms utilized by duplicated gene
pairs. In the neofunctionalization model, one copy of the gene maintains the ancestral
expression pattern or function, while the other paralog diverges and acquires new functions
(Massingham et al., 2001; Mazet and Shimeld, 2002). The other model, subfunctionalization,
proposes that random distinct mutations are accumulated in the two gene copies which results in
diverged functions of the paralogs (MacCarthy and Bergman, 2007; Massingham et al., 2001;
Mazet and Shimeld, 2002; Rastogi and Liberles, 2005; Roth et al., 2007). Only when they
complement each other will they be able to maintain their ancestral function, and that forces
both duplicates to be retained in the genome.
Comparison between vertebrate and non-vertebrate chordates (such as amphioxus) revealed that
many gene families that are duplicated in vertebrates are crucial for the development of the
central nervous system (Mazet and Shimeld, 2002). The relationship between gene duplication
and the innovation of specialized structures in the brain is a fascinating and largely unexplored
mechanism of vertebrate brain evolution. Some duplicated gene pairs such as the hox clusters
have a deep evolutionary history and their gene family has undergone a significant expansion
(Amores et al., 1998). However, in order to understand the process of how gene duplicates are
183
first fixed in the population, one will need to examine gene paralogs that have only recently
duplicated. Zebrafish offers an excellent opportunity to study these gene duplicates due to the
most recent whole genome duplication event that occurred in the teleosts after they separated
from the lobe-finned fishes (Meyer and Van de Peer, 2005).
6.2.1 Neofunctionalization is likely the retention mechanism of dmbx1
paralogs in zebrafish
Studies in zebrafish are increasingly playing an important role in deciphering the functional
consequences of gene duplication. Significant insight into the fundamental changes in brain
development that are due to the gene duplication in zebrafish comes from studies of the Dlx
(Ghanem et al., 2003; Ghanem et al., 2007), Pax (Kleinjan et al., 2008), and Zic gene families
(Elsen et al., 2008; Nyholm et al., 2007). In many cases, subfunctionalization is reported to play
an important role in the retention of duplicate genes (Bruce et al., 2001; de Souza et al., 2005;
Kleinjan et al., 2008; Lynch and Force, 2000; MacCarthy and Bergman, 2007; Tvrdik and
Capecchi, 2006). However, characterization of the dmbx1 paralogs revealed that these recent
duplicates were preserved by acquiring new functions in visual system development.
My work shows that Dmbx1a and Dmbx1b function cooperatively in a similar molecular
pathway that is essential to facilitate proper cell cycle progression in retinal and tectal progenitor
cells. However, detailed analysis of their expression revealed partially overlapping yet distinct
patterns. This might be the result of degenerative mutations in their regulatory sequences. It
seems that the protein functions of these duplicates remains relatively conserved, but their non-
coding elements have diverged into unique spatiotemporal expression patterns of both genes
during zebrafish embryogenesis. In addition, Dmbx1 in the teleost lineage appears to have a
more crucial role in regulating neurogenesis compared to other vertebrates that harbour a single
ortholog. It is because the zebrafish morphant embryos have more severe morphological defects
184
compared to the homozygous mutant mice. Furthermore, functional divergence was observed in
the zebrafish morpholino rescue experiments. These experiments revealed that the mouse
Dmbx1 gene does not have the same functional capacities as the zebrafish dmbx1 paralogs.
Moreover, analyses of the amino acid substitution rate of dmbx1 coding sequences suggest that
both zebrafish dmbx1a and dmbx1b have evolved under positive selection, which means the
proteins may have acquired new functions. Overall, my studies appear to support
neofunctionalization model as the mechanism for Dmbx1 retention.
There are two conventional models by which duplicated genes gain new function. They either
accumulated mutations in the protein coding region which changes the function of the protein,
or the regulatory elements are modified resulting in a change in expression pattern. The new
expression or function becomes fixed in the population during evolution. One of the new
functions of zebrafish Dmbx1 might be to downregulate the expression of ccnd1 in retinal and
tectal progenitor cells. In order to address whether the transcriptional repression of ccnd1 is a
teleost-specific invention, it would be useful to examine whether depleting Dmbx1 in other
vertebrates (such as mouse and chick) can also lead to an increase of ccnd1 level in their neural
tissues. Another possible scenario for Dmbx1 to show positive selection is that the zebrafish
Dmbx1 has expanded its expression domain into the eye, which amplified its role during visual
system development. If the relationship between the transcriptional repressor Dmbx1 and its
target ccnd1 is indeed conserved among vertebrates, then we would be able to detect this
regulatory phenomenon in other non-retinal domains with samples from other vertebrates where
Dmbx1 is expressed. This suggests that the new role of Dmbx1 is incurred through changes in
the gene regulatory region, not the protein coding region.
185
6.3 A potentially novel function for zebrafish dmbx1 paralogs in visual system development
6.3.1 Dmbx1 is the first paired-type homeodomain protein identified that
negatively regulates the cell cycle
Dmbx1 is a member of the K50 homeodomain group (lysine at position 50 of the homeodomain),
together with Otx2, Gsc, and Pitx2. Members of this family typically function as transcriptional
repressors (Bai et al., 2012; Berry et al., 2006; Kimura et al., 2005; Yao and Kessler, 2001).
Thus, the finding that Dmbx1 can downregulate the expression of cyclinD1 is consistent with a
role as a transcriptional repressor. Many paired-type homeodomain proteins, such as Vsx2 and
Pitx2, can interact with different cell cycle components (Baek et al., 2003; Fung et al., 2012;
Green et al., 2003). In fact, they are generally positive regulators of the cell cycle (Baek et al.,
2003; Budhram-Mahadeo et al., 2008; Bunt et al., 2012; Fung et al., 2012; Green et al., 2003).
However, Dmbx1a and Dmbx1b are the first two K50 paired-type homeoproteins that
demonstrate a negative regulation of the cell cycle by repressing ccnd1.
6.3.2 Dmbx1-interacting partners during retinal and tectal development
6.3.2.1 Dmbx1 may antagonize Vsx2 in a cis-regulatory network
To further characterize where the dmbx1 genes are positioned in the transcriptional network that
regulates the transition from proliferating progenitor cells to post-mitotic retinal neurons or glial
cells, I examined other paired-type homeodomain proteins that may interact with Dmbx1 during
retinogenesis in order to help regulate aspect of the cell cycle, particularly ccnd1. A potential
candidate that might interact with Dmbx1 is Vsx2 (Chx10 homolog), which is a Q50 paired-like
homeobox gene that is known to be important for retinal development. In mouse, it has been
demonstrated that Chx10 upregulates ccnd1 expression in order to prevent the accumulation of
p27kip1
proteins in the retina (Green et al., 2003). Studies in zebrafish have shown that vsx2
186
morphants have microphthalmia but retinal differentiation is apparently normal, similar to the
ccnd1 morphant embryos (Das et al., 2009; Duffy et al., 2005; Vitorino et al., 2009). This
evidence suggested that Dmbx1 may be antagonizing Vsx2 during retinogenesis. Support for
this hypothesis comes from work in Ciona which demonstrated that ectopic expression of Vsx2
resulted in reduced dmbx1 expression and vice versa, indicating that these two genes strongly
repress one another in the spinal cord (Stolfi and Levine, 2011).
Our lab has recently begun to examine the relationship between Vsx2 and Dmbx1a in the retina.
Knocking down dmbx1 paralogs resulted in expanded vsx2 expression in the retina. When vsx2
was misexpressed using a heat-activated transgene in zebrafish embryos, the opposite outcome
was observed where dmbx1a expression was reduced (Namita Power, personal communication).
These data suggest that vsx2 and dmbx1a may be mutually repressing one another during
retinogenesis. It is known that Vsx2 functions to maintain proliferation of retinal progenitor
cells (Vitorino et al., 2009), while Dmbx1a promotes cell cycle exit in those retinal progenitor
cells. There is a possibility that these two transcription factors are directly repressing one
another or they could be affecting each other by competing for the control of ccnd1 at the
transcription level.
6.3.2.2 Protein-protein interaction of Dmbx1 for cell cycle regulation
The current working model is that Dmbx1 paralogs negatively regulate ccnd1 expression level
to control the cell cycle in retinal and tectal progenitor cells. However, it is also possible that
Dmbx1 may require other binding partners in order to control ccnd1 expression in a spatially
and temporally restricted manner. To identify novel potential interacting partners of Dmbx1, a
yeast-two-hybrid (Y2H) screen was performed using Dmbx1a as bait (see Appendix 2 for
details). From a list of proteins pulled out from the Y2H screen, I predict that the following four
187
proteins - Rassf1, Meis1, Raraa, and Rarab, are good candidate partners for Dmbx1a based on
their expression patterns and functions.
Studies have shown that Ras-Association Domain Family 1 (Rassf1) is a tumour suppressor
gene that can negatively control Cyclin D1 accumulation, which in turn inhibits cell cycle
progression (Shivakumar et al., 2002). Another candidate, Meis1 has been shown to upregulate
cyclinD1 expression in the retinal primordia to promote eye growth during early development,
and together with ccnd1 both genes are downregulated when neurogenesis begins (Bessa et al.,
2008). Thus, it is possible that Meis1 binds to the cyclinD1 promoter to initiate its transcription
during early proliferation within the retina. Once retinogenesis starts and Dmbx1a begins to
interact with Meis1 on the cyclinD1 promoter, this heterodimer complex could become a
transcriptional repressor of cyclinD1. In addition, retinoic acid receptors alpha a and b (Raraa &
Rarab respectively) are both potential binding partners of Dmbx1a. The transition between G1
and S requires rapid breakdown of CyclinD1 in order for proliferating cells to undergo DNA
synthesis. Degradation of CyclinD1 proteins is mainly carried out by 26S proteasomes and it is
known that retinoid acid receptors play a major role in controlling CyclinD1 stability through
ubiquitin-dependent proteolysis (Alao, 2007; Langenfeld et al., 1997; MJ et al., 1999). Hence, it
is possible that Dmbx1 may interact with retinoic acid receptors to induce cyclinD1 degradation
in retinal and tectal progenitor cells.
6.4 Future Experiments
From my Ph.D. work, I have characterized the cellular mechanisms that Dmbx1a and Dmbx1b
participated in during zebrafish embryonic development. In particular, my research has
contributed to the understanding of how transcription factors orchestrate the transition between
progenitor cells to mature neurons during neurogenesis through cell cycle regulation. However,
188
the molecular mechanisms which Dmbx1 utilizes to control the cell cycle in retinal and tectal
progenitor cells still remain largely unknown. Future work should focus on unveiling the gene
regulatory networks that Dmbx1a and Dmbx1b are engaged in, both collectively and
individually. Moreover, emphasis should be given to identifying the direct downstream targets
of each Dmbx1 transcription factor, in order to understand how cell cycle kinetics is controlled.
Lastly, it will be important to analyze the effects that dmbx1 paralogs might have on visual
processing behaviours. This will help to address how neural development and neuronal
networks are linked.
6.4.1 Verify potential downstream targets of Dmbx1 in vitro
To discover additional pathways that Dmbx1a and Dmbx1b might be involved in, or to identify
downstream targets, a biochemical approach is needed to demonstrate their physical
interactions. It will be useful to take an unbiased approach and perform a large-scale
transcriptome analysis to identify new candidate genes/pathways that Dmbx1a and Dmbx1b
regulate. An experiment such as whole transcriptome shotgun sequencing, which allows
differentially expressed RNA sequences to be compared between control and dmbx1-deficiency
embryos in a high-throughput manner, will be a suitable assay to look for new targets.
6.4.1.1 Whole transcriptome shotgun sequencing with dmbx1 morphants
Massive parallel transcriptome sequencing (RNA-seq) can be performed on different platforms,
which vary according to the coverage and the size of each read. For example, several published
papers have used the paired end sequencing from Illumina (Collins et al., 2012; Gomez et al.,
2012; Palstra et al., 2013; Yang et al., 2012) which provides on average about a total amount of
54-60 Gb reads from many short 100bp sequences obtained from both protein coding and non-
coding regions (as listed on the Illumina website). This type of experiment can provide
189
comprehensive information on the alternatively expressed transcripts between samples.
Moreover, it can also detect details such as proportion of alternatively-spliced transcripts of each
gene or even different gene allele.
To increase the likelihood of identifying alternatively expressed transcripts in our assays that are
meaningful, we need to enrich our samples and increase the signal-to-noise ratio. Since strong
expression of dmbx1a and dmbx1b can be found in the head region, it will be rational to use
total RNA samples taken from the heads of the control- and dmbx1-morpholino injected
embryos for this comparative analysis. Given the tremendous amount of sequencing data that
will be collected from the analyses, another good way to help decipher the data and pull out
some strong candidate genes is to cross-reference the RNA-seq data obtained from single and
double-morpholino injected embryos and select only those that are both differentially expressed
in the two sets of data compared to controls for subsequent analyses. The list of potential
downstream target genes between Dmbx1a and Dmbx1b may be fairly similar due to their
conserved N-terminal region and homeodomain. However, differences in their expression
domains may allow them to interact with other genes that are unique targets of each protein.
One major weakness with this experiment is the inability to distinguish direct or indirect targets
of Dmbx1a or Dmbx1b, which required further validations as discuss below.
6.4.1.2 Verifying potential direct targets of Dmbx1a/Dmbx1b
There are several ways to confirm our findings from the whole transcriptome shotgun
sequencing experiment. Transcripts that are linked to cell cycle (ccnd1 can be our internal
control) or related to visual system development (such as vsx2) will be our top candidates for
further investigations. Gene candidates from Dmbx1a- or Dmbx1b-target list will be selected
from the RNA-seq data for detail analyses. To first validate the changes detected at the
190
transcription level of those target genes, quantitative RT-PCR will be performed to evaluate the
actual fold changes in those genes that may interact with Dmbx1a or Dmbx1b. We will also use
wholemount in situ hybridization to examine how the expression patterns of these potential
target genes changed in dmx1a or dmbx1b morphant embryos. Specific changes in the spatial
expression of those target genes will also refine our search for direct Dmbx1a or Dmbx1b
downstream target genes, as we expect those modifications to localize at the
retina/tectum/bilateral hindbrain regions close to where Dmbx1a and Dmbx1b are expressed.
Once the transcriptional changes of these potential Dmbx1-target genes have been confirmed in
dmbx1a and dmbx1b single morphants, we will need to perform biochemical assays to examine
if this transcriptional regulation is direct or not. Regulatory sequences from these direct target
genes can be subjected to the Electrophoretic Mobility Shift Assay (EMSA) to see whether
Dmbx1a and/or Dmbx1b transcription factors can bind to them. However, the major challenge
in verifying these protein-DNA interactions is the lack of specific antibodies for Dmbx1a and
Dmbx1b in zebrafish. To overcome this issue, GST-tagged full length Dmbx1a and Dmbx1b
proteins have been generated and GST antibody can be used in order to detect the recombinant
Dmbx1a- and Dmbx1b-DNA bound complexes in a gel shift assay. Successful binding between
Dmbx1 transcription factors and the promoter sequences of these target genes will indicate that
they are likely the direct downstream targets of Dmbx1a and/or Dmbx1b in a transcriptional
network. By subcloning the promoter region of these target genes into a reporter construct (such
as luciferase), we can directly assay the changes in transcriptional activities of these target genes
in the presence and absence of Dmbx1.
191
6.4.2 Examine the importance of dmbx1 in prey capture behaviour
In dmbx1 double morphant embryos, retinotectal projections were defective at 4dpf. It appeared
that the RGCs were able to cross the midline and formed a proper chiasm even though the axon
bundle appeared thinner than normal. However, upon reaching the tectum, arborization was
barely observed. Since the optic tectum is a relay center that receives visual cues from the eyes
and processes the information into motion outputs, this led us to suspect that morphant embryos
may not have normal vision. However, it is well-documented that the effectiveness of antisense
morpholino oligonucleotides injected at 1-cell stage will gradually decrease as the concentration
becomes diluted with cell division as the embryo develops (Nasevicius and Ekker, 2000). Thus,
Dmbx1 mutants will be required in order to study visually guided behaviours at later stages.
6.4.2.1 Generation of dmbx1a and dmbx1b mutant stains
There are several strategies that can be used to generate mutants in zebrafish. Although forward
genetic screening such as traditional ENU-mutagenesis has been used for years to create
mutants, this method offers low probability of targeting the gene of interest, as certain regions in
the genome are more or less susceptible to those mutagens (Driever et al., 1996). Hence, more
gene specific targeting techniques to generate mutants have been developed over the years.
The Wellcome Trust Sanger Institute has produced a dmbx1a knockout mutant using the
Targeting Induced Local Lesions in Genomes (TILLING) method with a point mutation that
removes the splice site between exon 2 and 3, resulting in a premature stop codon. This
truncated protein lacks the homeodomain and thus is potentially a null mutation. Unfortunately
the Dmbx1b mutant is not available at present, but our lab is currently testing out the Clustered
Regularly Interspaced Short Palindromic Repeats (CRISPR) system to generate site specific
mutant lines.
192
The CRISPR-Cas method can be used to generate insertions or deletions at targeted sites in the
dmbx1b coding region. Non-coding RNA (single-guided RNA) complementary to the targeted
Dmbx1b genomic region will recruit the CRISPR-associated (Cas) protein to create a double
strand break where the RNA-DNA duplex is formed (Hwang et al., 2013; Makarova et al.,
2011). The breakpoint on the DNA strand by this nuclease undergoes a non-homologous end-
joining repair mechanism that introduces insertions or deletions at the cleavage site (Hwang et
al., 2013; Makarova et al., 2011). This targeted error prone repair system at the genomic level
will likely disrupt the protein coding sequence of Dmbx1b, resulting in a defective transcription
factors. By injecting a RNA sequence that harbors the dmbx1b target site and the Cas mRNA
together into 1-cell stage zebrafish embryos, different allelic mutants can be raised and
genotyped for further analysis. The possibility of generating multiple alleles simultaneously
allows us to study the functional roles of different domains in the Dmbx1b genes. Ultimately, it
would be informative to examine these dmbx1 double mutants and check whether those embryos
have a more severe phenotype compared to dmbx1a or dmbx1b single mutant fish. Furthermore,
it will also be helpful to utilize these dmbx1 single and double mutant fish to discover more
visual processing behaviours that the Dmbx1 transcription factors are regulating within the
visual system.
6.4.2.2 Prey capture study with dmbx1 mutant fish
The tecta is responsible for prey capture (Gahtan et al., 2005), but not the optokinetic reflex and
optomotor response (Portugues and Engert, 2009; Roeser and Baier, 2003). In order to learn
whether prey capture behaviours are compromised in the absence of dmbx1 during development,
the dmbx1 mutant fish will be ideal for this study. The first sign of food hunting in zebrafish
larvae begins around 5dpf with two evident behaviours. Eye convergence response is the first
sign that prey capturing has started, followed by the J-turn tail movement (Bianco et al., 2011).
193
Using the “prey capture” assay developed by Bianco et al, the authors showed that hunting
responses can be observed with the same magnitude in larvae that were either restrained by low
melt agarose or free swimming in a petri dish (Bianco et al., 2011). By measuring the ocular and
tail movement induced by visual stimuli in dmbx1 single/double mutant larvae and wildtype
larvae under the retrained method, we can compare those values and determine whether prey
capture behaviour is abolished in the absence of dmbx1 genes. This will allow us to examine
whether neural development during embryogenesis can have a permanent effect on the
establishment of neuronal circuitries within the visual system.
6.5 Conclusion
My work has demonstrated the importance of Dmbx1 in regulating visual system development.
In addition, the work helped establish Dmbx1a and Dmbx1b as key cell cycle regulators during
retinal and tectal neurogenesis. I examined in great detail the dynamic spatiotemporal
expression patterns of dmbx1a and dmbx1b from embryonic to larval stages, as well as
characterizing the phenotypic defects in single and double dmbx1 knock down embryos using
morpholino antisense oligonucleotides. Loss of function analyses showed that morphant
embryos had severe defects in retinal and tectal growth, and differentiation was delayed in areas
where the dmbx1 paralogs are expressed. Further investigation of the causation of these
developmental defects should clarify the cellular mechanism by which Dmbx1 regulates cell
cycle progression in retinal and tectal progenitor cells.
A strong negative correlation between levels of Dmbx1 and the cell cycle component ccnd1
indicates that progenitor cells in the retina and optic tectum require Dmbx1 to downregulate
ccnd1 during the G1 to S phase transition so that these cells can undergo mitosis in a timely
manner or exit the cell cycle to differentiate. Moreover, I provided evidence that Dmbx1a and
194
Dmbx1b are both necessary and sufficient to regulate the cell cycle in retinal and tectal
progenitor cells. It appears that their roles are more prominent in the tectal region, likely due to
their involvement with the Wnt-signaling pathway that is activated at the dorsal mesencephalic
midline. Overall, I have demonstrated that Dmbx1a and Dmbx1b are indispensable during
neurogenesis and they are crucial for maintaining specific sets of neurons in the retina, tectum,
and rhombomeres during post-embryonic stages. However, it is not clear what subtypes of
neurons they specify and hence, a systematic analysis to identify those cell types will provide a
good framework for subsequent studies to look for any downstream targets genes that are
specific to Dmbx1a or Dmbx1b.
Gene duplication is known to be a mechanism for acquiring new gene functions during the
course of evolution. In particular, the zebrafish and other teleost species have undergone an
additional round of whole genome duplication compared to other non-teleost vertebrates,
resulting in more recent duplicated gene pairs that are retained in the genome. A paper by Lu et
al has reported that there are 3,991 duplicated gene sets amongst 26,842 genes in the zebrafish
genome (approximately 15% of the genes have paralogs) (Lu et al., 2012), but only a few papers
in the literature have examined the function of these paralogs. Many evolutionary models have
been proposed to explain why duplicated genes are retained; however, without any meaningful
functional data from these gene pairs to verify them it is difficult to argue that their presence has
help promote gene diversification.
My research on the functions of dmbx1 paralogs in zebrafish has contributed to the verification
of function divergence observed in duplicated genes. I showed that both dmbx1a and dmbx1b
are independently required to regulate neurogenesis in the midbrain and they are clearly
expressed in different subpopulation of tectal neurons during later development. Moreover,
195
dmbx1a has a predominant role in regulating neurogenesis in the retina and anterior hindbrain,
and is therefore functionally diverged from dmbx1b. Furthermore, I provide evidence for post-
duplication positive selection in teleost dmbx1 genes and show that the mouse Dmbx1 gene is
not sufficient to functionally compensate for the reduced levels of endogenous zebrafish dmbx1a
or dmbx1b. Therefore, compared to the single mouse Dmbx1 gene, zebrafish dmbx1 duplicate
genes appear to be functionally diverged and have an important role in controlling neurogenesis
during development. Evolution of the Dmbx1 paralogs provides an example of how duplicated
genes utilize the neofunctionalization mechanism to preserve both gene copies in the genome.
196
References or Bibliography
Abràmoff, M. D., Magalhães, P. J. and Ram, S. J. (2004). Image Processing with ImageJ.
Biophotonics International 11, 36–42.
Adijanto, J., Castorino, J. J., Wang, Z.-X., Maminishkis, A., Grunwald, G. B. and Philp,
N. J. (2012). Microphthalmia-associated transcription factor (MITF) promotes
differentiation of human retinal pigment epithelium (RPE) by regulating microRNAs-
204/211 expression. J Biol Chem. 287, 20491–503.
Akimenko, M. A., Ekker, M., Wegner, J., Lin, W. and Westerfield, M. (1994).
Combinatorial expression of three zebrafish genes related to distal-less: part of a homeobox
gene code for the head. J Neurosci. 14, 3475–86.
Alao, J. P. (2007). The regulation of cyclin D1 degradation: roles in cancer development and
the potential for therapeutic invention. Mol Cancer. 6, 24.
Alonso, J.-M., Yeh, C.-I., Weng, C. and Stoelzel, C. (2006). Retinogeniculate connections: A
balancing act between connection specificity and receptive field diversity. Prog Brain Res.
154, 3–13.
Alves dos Santos, M. T. M. and Smidt, M. P. (2011). En1 and Wnt signaling in midbrain
dopaminergic neuronal development. Neural Dev. 6, 23.
Amendt, B. A., Sutherland, L. B. and Russo, A. F. (1999). Multifunctional role of the Pitx2
homeodomain protein C-terminal tail. Mol Cell Biol. 19, 7001–10.
Amores, A., Force, A., Yan, Y. L., Joly, L., Amemiya, C., Fritz, A., Ho, R. K., Langeland,
J., Prince, V., Wang, Y. L., et al. (1998). Zebrafish hox clusters and vertebrate genome
evolution. Science 282, 1711–4.
Araki, I. and Nakamura, H. (1999). Engrailed defines the position of dorsal di-mesencephalic
boundary by repressing diencephalic fate. Development. 126, 5127–35.
Arees, E. A. (1986). Absence of light response in eyeless planaria. Physiol Behav. 36, 445–449.
197
Aubie, B., Sayegh, R. and Faure, P. A. (2012). Duration tuning across vertebrates. J Neurosci.
32, 6373–90.
Baek, S. H., Kioussi, C., Briata, P., Wang, D., Nguyen, H. D., Ohgi, K. A., Glass, C. K.,
Wynshaw-Boris, A., Rose, D. W. and Rosenfeld, M. G. (2003). Regulated subset of G1
growth-control genes in response to derepression by the Wnt pathway. Proc Natl Acad Sci
U S A. 100, 3245–50.
Bai, R.-Y., Staedtke, V., Lidov, H. G., Eberhart, C. G. and Riggins, G. J. (2012). OTX2
represses myogenic and neuronal differentiation in medulloblastoma cells. Cancer Res. 72,
5988–6001.
Bailey, T. J., El-Hodiri, H., Zhang, L., Shah, R., Mathers, P. H. and Jamrich, M. (2004).
Regulation of vertebrate eye development by Rx genes. Int J Dev Biol. 48, 761–70.
Baird-Titus, J. M., Clark-Baldwin, K., Dave, V., Caperelli, C. A., Ma, J. and Rance, M.
(2006). The solution structure of the native K50 Bicoid homeodomain bound to the
consensus TAATCC DNA-binding site. J Mol Biol. 356, 1137–51.
Bally-Cuif, L. and Wassef, M. (1994). Ectopic induction and reorganization of Wnt-1
expression in quail/chick chimeras. Development. 120, 3379–94.
Barabino, S. M., Spada, F., Cotelli, F. and Boncinelli, E. (1997). Inactivation of the zebrafish
homologue of Chx10 by antisense oligonucleotides causes eye malformations similar to the
ocular retardation phenotype. Mech Dev. 63, 133–43.
Barolo, S. (2006). Transgenic Wnt/TCF pathway reporters: all you need is Lef? Oncogene. 25,
7505–11.
Batista, M. F. and Lewis, K. E. (2008). Pax2/8 act redundantly to specify glycinergic and
GABAergic fates of multiple spinal interneurons. Dev Biol. 323, 88–97.
Baylor, D. (1996). How photons start vision. Proc Natl Acad Sci U S A. 93, 560–5.
Becker, C. G. and Becker, T. (2007). Growth and pathfinding of regenerating axons in the
optic projection of adult fish. J Neurosci Res. 85, 2793–9.
198
Belting, H. G., Hauptmann, G., Meyer, D., Abdelilah-Seyfried, S., Chitnis, A., Eschbach,
C., Söll, I., Thisse, C., Thisse, B., Artinger, K. B., et al. (2001). spiel ohne grenzen/pou2
is required during establishment of the zebrafish midbrain-hindbrain boundary organizer.
Development. 128, 4165–76.
Bernardos, R. L., Barthel, L. K., Meyers, J. R. and Raymond, P. A. (2007). Late-stage
neuronal progenitors in the retina are radial Müller glia that function as retinal stem cells. J
Neurosci. 27, 7028–40.
Berry, F. B., Lines, M. A., Oas, J. M., Footz, T., Underhill, D. A., Gage, P. J. and Walter,
M. A. (2006). Functional interactions between FOXC1 and PITX2 underlie the sensitivity
to FOXC1 gene dose in Axenfeld-Rieger syndrome and anterior segment dysgenesis. Hum
Mol Genet. 15, 905–19.
Bertrand, S., Thisse, B., Tavares, R., Sachs, L., Chaumot, A., Bardet, P.-L., Escrivà, H.,
Duffraisse, M., Marchand, O., Safi, R., et al. (2007). Unexpected novel relational links
uncovered by extensive developmental profiling of nuclear receptor expression. PLoS
Genet. 3, e188.
Bessa, J., Tavares, M. J., Santos, J., Kikuta, H., Laplante, M., Becker, T. S., Gómez-
Skarmeta, J. L. and Casares, F. (2008). meis1 regulates cyclin D1 and c-myc expression,
and controls the proliferation of the multipotent cells in the early developing zebrafish eye.
Development. 135, 799–803.
Bhanot, P., Brink, M., Samos, C. H., Hsieh, J. C., Wang, Y., Macke, J. P., Andrew, D.,
Nathans, J. and Nusse, R. (1996). A new member of the frizzled family from Drosophila
functions as a Wingless receptor. Nature. 382, 225–30.
Bharti, K., Gasper, M., Ou, J., Brucato, M., Clore-Gronenborn, K., Pickel, J. and
Arnheiter, H. (2012). A regulatory loop involving PAX6, MITF, and WNT signaling
controls retinal pigment epithelium development. PLoS Genet. 8, e1002757.
Bianco, I. H., Kampff, A. R. and Engert, F. (2011). Prey capture behavior evoked by simple
visual stimuli in larval zebrafish. Front Syst Neurosci. 5, 101.
199
Bilitou, A. and Ohnuma, S. (2010). The role of cell cycle in retinal development: cyclin-
dependent kinase inhibitors co-ordinate cell-cycle inhibition, cell-fate determination and
differentiation in the developing retina. Dev Dyn. 239, 727–36.
Bingham, S., Chaudhari, S., Vanderlaan, G., Itoh, M., Chitnis, A. and Chandrasekhar, A.
(2003). Neurogenic phenotype of mind bomb mutants leads to severe patterning defects in
the zebrafish hindbrain. Dev Dyn. 228, 451–63.
Bloomfield, S. A. and Dacheux, R. F. (2001). Rod vision: pathways and processing in the
mammalian retina. Prog Retin Eye Res. 20, 351–84.
Bonkowsky, J. L., Wang, X., Fujimoto, E., Lee, J. E., Chien, C.-B. and Dorsky, R. I.
(2008). Domain-specific regulation of foxP2 CNS expression by lef1. BMC Dev Biol. 8,
103.
Bora, N., Defoe, D. and Smith, S. B. (1999). Evidence of decreased adhesion between the
neural retina and retinal pigmented epithelium of the Mitfvit (vitiligo) mutant mouse. Cell
Tissue Res. 295, 65–75.
Borst, A. (2009). Drosophila’s view on insect vision. Curr Biol. 19, R36–47.
Bovolenta, P., Mallamaci, A., Briata, P., Corte, G. and Boncinelli, E. (1997). Implication of
OTX2 in pigment epithelium determination and neural retina differentiation. J Neurosci.
17, 4243–52.
Bowmaker, J. K. (1998). Evolution of colour vision in vertebrates. Eye. 12 ( Pt 3b, 541–7.
Broccoli, V., Colombo, E. and Cossu, G. (2002). Dmbx1 is a paired-box containing gene
specifically expressed in the caudal most brain structures. Mech Dev. 114, 219–23.
Brockerhoff, S. E., Dowling, J. E. and Hurley, J. B. (1998). Zebrafish retinal mutants. Vision
Res. 38, 1335–9.
Brombin, A., Grossier, J.-P., Heuzé, A., Radev, Z., Bourrat, F., Joly, J.-S. and Jamen, F.
(2011). Genome-wide analysis of the POU genes in medaka, focusing on expression in the
optic tectum. Dev Dyn. 240, 2354–63.
200
Brouwer, A., ten Berge, D., Wiegerinck, R. and Meijlink, F. (2003). The OAR/aristaless
domain of the homeodomain protein Cart1 has an attenuating role in vivo. Mech Dev. 120,
241–52.
Bruce, A. E., Oates, A. C., Prince, V. E. and Ho, R. K. (2001). Additional hox clusters in the
zebrafish: divergent expression patterns belie equivalent activities of duplicate hoxB5
genes. Evol Dev. 3, 127–44.
Buckles, G. R., Thorpe, C. J., Ramel, M.-C. and Lekven, A. C. (2004). Combinatorial Wnt
control of zebrafish midbrain-hindbrain boundary formation. Mech Dev. 121, 437–47.
Budhram-Mahadeo, V. S., Irshad, S., Bowen, S., Lee, S. A., Samady, L., Tonini, G. P. and
Latchman, D. S. (2008). Proliferation-associated Brn-3b transcription factor can activate
cyclin D1 expression in neuroblastoma and breast cancer cells. Oncogene 27, 145–54.
Bunt, J., Hasselt, N. E., Zwijnenburg, D. A., Hamdi, M., Koster, J., Versteeg, R. and Kool,
M. (2012). OTX2 directly activates cell cycle genes and inhibits differentiation in
medulloblastoma cells. Int J Cancer. 131, E21–32.
Burmeister, M., Novak, J., Liang, M. Y., Basu, S., Ploder, L., Hawes, N. L., Vidgen, D.,
Hoover, F., Goldman, D., Kalnins, V. I., et al. (1996). Ocular retardation mouse caused
by Chx10 homeobox null allele: impaired retinal progenitor proliferation and bipolar cell
differentiation. Nat Genet. 12, 376–84.
Burrill, J. D. and Easter, S. S. (1994). Development of the retinofugal projections in the
embryonic and larval zebrafish (Brachydanio rerio). J Comp Neurol. 346, 583–600.
Candal, E., Anadón, R., DeGrip, W. J. and Rodríguez-Moldes, I. (2005). Patterns of cell
proliferation and cell death in the developing retina and optic tectum of the brown trout.
Brain Res Dev Brain Res. 154, 101–19.
Carl, M., Loosli, F. and Wittbrodt, J. (2002). Six3 inactivation reveals its essential role for the
formation and patterning of the vertebrate eye. Development. 129, 4057–63.
Cavodeassi, F. and Houart, C. (2012). Brain regionalization: of signaling centers and
boundaries. Dev Neurobiol. 72, 218–33.
201
Chaney, B. A., Clark-Baldwin, K., Dave, V., Ma, J. and Rance, M. (2005). Solution structure
of the K50 class homeodomain PITX2 bound to DNA and implications for mutations that
cause Rieger syndrome. Biochemistry. 44, 7497–511.
Chang, L., Khoo, B., Wong, L. and Tropepe, V. (2006). Genomic sequence and
spatiotemporal expression comparison of zebrafish mbx1 and its paralog, mbx2. Dev Genes
Evol. 216, 647–654.
Chen, W. and Casey Corliss, D. (2004). Three modules of zebrafish Mind bomb work
cooperatively to promote Delta ubiquitination and endocytosis. Dev Biol. 267, 361–73.
Chen, L., Yang, P. and Kijlstra, A. (2002a). Distribution, markers, and functions of retinal
microglia. Ocul Immunol Inflamm. 10, 27–39.
Chen, J. K., Taipale, J., Cooper, M. K. and Beachy, P. A. (2002b). Inhibition of Hedgehog
signaling by direct binding of cyclopamine to Smoothened. Genes Dev. 16, 2743–8.
Cheng, T. (2004). Cell cycle inhibitors in normal and tumor stem cells. Oncogene. 23, 7256–66.
Cheng, C. W., Yan, C. H. M., Hui, C., Strähle, U. and Cheng, S. H. (2006). The homeobox
gene irx1a is required for the propagation of the neurogenic waves in the zebrafish retina.
Mech Dev. 123, 252–63.
Cheyette, B. N., Green, P. J., Martin, K., Garren, H., Hartenstein, V. and Zipursky, S. L.
(1994). The Drosophila sine oculis locus encodes a homeodomain-containing protein
required for the development of the entire visual system. Neuron. 12, 977–96.
Cid, E., Santos-Ledo, A., Parrilla-Monge, M., Lillo, C., Arévalo, R., Lara, J. M., Aijón, J.
and Velasco, A. (2010). Prox1 expression in rod precursors and Müller cells. Exp Eye Res.
90, 267–76.
Clements, W. K., Ong, K. G. and Traver, D. (2009). Zebrafish wnt3 is expressed in
developing neural tissue. Dev Dyn. 238, 1788–95.
Cole, L. K. and Ross, L. S. (2001). Apoptosis in the developing zebrafish embryo. Dev Biol.
240, 123–42.
202
Collery, R. F. and Kennedy, B. N. (2010). Photoreceptor guanylate cyclases and cGMP
phosphodiesterases in zebrafish. Adv Exp Med Biol. 664, 55–61.
Collins, J. E., White, S., Searle, S. M. J. and Stemple, D. L. (2012). Incorporating RNA-seq
data into the zebrafish Ensembl genebuild. Genome Res. 22, 2067–78.
Conant, G. C. and Wolfe, K. H. (2008). Turning a hobby into a job: how duplicated genes find
new functions. Nat Rev Genet. 9, 938–50.
Connaughton, V. P., Graham, D. and Nelson, R. (2004). Identification and morphological
classification of horizontal, bipolar, and amacrine cells within the zebrafish retina. J Comp
Neurol. 477, 371–85.
Coqueret, O. (2002). Linking cyclins to transcriptional control. Gene. 299, 35–55.
Corbo, C. P., Othman, N. a, Gutkin, M. C., Alonso, A. D. C. and Fulop, Z. L. (2012). Use of
different morphological techniques to analyze the cellular composition of the adult
zebrafish optic tectum. Microsc Res Tech. 75, 325–33.
Culverwell, J. and Karlstrom, R. O. (2002). Making the connection: retinal axon guidance in
the zebrafish. Semin Cell Dev Biol. 13, 497–506.
Dacey, D. M. (1996). Circuitry for color coding in the primate retina. Proc Natl Acad Sci U S A.
93, 582–8.
Dansault, A., David, G., Schwartz, C., Jaliffa, C., Vieira, V., de la Houssaye, G., Bigot, K.,
Catin, F., Tattu, L., Chopin, C., et al. (2007). Three new PAX6 mutations including one
causing an unusual ophthalmic phenotype associated with neurodevelopmental
abnormalities. Mol Vis. 13, 511–23.
Das, G., Choi, Y., Sicinski, P. and Levine, E. M. (2009). Cyclin D1 fine-tunes the neurogenic
output of embryonic retinal progenitor cells. Neural Dev. 4, 15.
De Souza, F. S. J., Bumaschny, V. F., Low, M. J. and Rubinstein, M. (2005).
Subfunctionalization of expression and peptide domains following the ancient duplication
of the proopiomelanocortin gene in teleost fishes. Mol Biol Evol. 22, 2417–27.
203
DeCarvalho, A. C., Cappendijk, S. L. T. and Fadool, J. M. (2004). Developmental
expression of the POU domain transcription factor Brn-3b (Pou4f2) in the lateral line and
visual system of zebrafish. Dev Dyn. 229, 869–76.
Deeg, K. E., Sears, I. B. and Aizenman, C. D. (2009). Development of multisensory
convergence in the Xenopus optic tectum. J Neurophysiol. 102, 3392–404.
Del Bene, F. and Wittbrodt, J. (2005). Cell cycle control by homeobox genes in development
and disease. Semin Cell Dev Biol. 16, 449–60.
Del Bene, F., Wyart, C., Robles, E., Tran, A., Looger, L., Scott, E. K., Isacoff, E. Y. and
Baier, H. (2010). Filtering of visual information in the tectum by an identified neural
circuit. Science. 330, 669–73.
Dhomen, N. S., Balaggan, K. S., Pearson, R. A., Bainbridge, J. W., Levine, E. M., Ali, R. R.
and Sowden, J. C. (2006). Absence of chx10 causes neural progenitors to persist in the
adult retina. Invest Ophthalmol Vis Sci. 47, 386–96.
Donovan, S. L. and Dyer, M. a (2005). Regulation of proliferation during central nervous
system development. Semin Cell Dev Biol. 16, 407–21.
Dorsky, R. I., Snyder, A., Cretekos, C. J., Grunwald, D. J., Geisler, R., Haffter, P., Moon,
R. T. and Raible, D. W. (1999). Maternal and embryonic expression of zebrafish lef1.
Mech Dev. 86, 147–50.
Dorsky, R. I., Sheldahl, L. C. and Moon, R. T. (2002). A transgenic Lef1/beta-catenin-
dependent reporter is expressed in spatially restricted domains throughout zebrafish
development. Dev Biol. 241, 229–37.
Douglas, R. H., Partridge, J. C. and Marshall, N. J. (1998). The eyes of deep-sea fish. I: Lens
pigmentation, tapeta and visual pigments. Prog Retin Eye Res. 17, 597–636.
Driever, W., Solnica-Krezel, L., Schier, a F., Neuhauss, S. C., Malicki, J., Stemple, D. L.,
Stainier, D. Y., Zwartkruis, F., Abdelilah, S., Rangini, Z., et al. (1996). A genetic
screen for mutations affecting embryogenesis in zebrafish. Development 123, 37–46.
204
Du, Y. and Yip, H. K. (2011). The expression and roles of inhibitor of DNA binding helix-
loop-helix proteins in the developing and adult mouse retina. Neuroscience. 175, 367–79.
Duffy, K. T., McAleer, M. F., Davidson, W. R., Kari, L., Kari, C., Liu, C.-G., Farber, S. a,
Cheng, K. C., Mest, J. R., Wickstrom, E., et al. (2005). Coordinate control of cell cycle
regulatory genes in zebrafish development tested by cyclin D1 knockdown with
morpholino phosphorodiamidates and hydroxyprolyl-phosphono peptide nucleic acids.
Nucleic Acids Res. 33, 4914–21.
Dyer, M. A. and Cepko, C. L. (2001a). p27 Kip1 and p57 Kip2 Regulate Proliferation in
Distinct Retinal Progenitor Cell Populations. J Neurosci. 21, 4259–4271.
Dyer, M. a and Cepko, C. L. (2001b). Regulating proliferation during retinal development. Nat
Rev Neurosci. 2, 333–42.
Dziedzic, K., Heaphy, J., Prescott, H. and Kavaler, J. (2009). The transcription factor D-Pax2
regulates crystallin production during eye development in Drosophila melanogaster. Dev
Dyn. 238, 2530–9.
Eaton, R. C., Lee, R. K. and Foreman, M. B. (2001). The Mauthner cell and other identified
neurons of the brainstem escape network of fish. Prog Neurobiol. 63, 467–85.
Ekker, M., Wegner, J., Akimenko, M. A. and Westerfield, M. (1992). Coordinate embryonic
expression of three zebrafish engrailed genes. Development. 116, 1001–10.
el-Deiry, W. S., Tokino, T., Velculescu, V. E., Levy, D. B., Parsons, R., Trent, J. M., Lin,
D., Mercer, W. E., Kinzler, K. W. and Vogelstein, B. (1993). WAF1, a potential
mediator of p53 tumor suppression. Cell. 75, 817–25.
Elsen, G. E., Choi, L. Y., Millen, K. J., Grinblat, Y. and Prince, V. E. (2008). Zic1 and Zic4
regulate zebrafish roof plate specification and hindbrain ventricle morphogenesis. Dev Biol.
314, 376–92.
Erickson, T., Scholpp, S., Brand, M., Moens, C. B. and Waskiewicz, A. J. (2007). Pbx
proteins cooperate with Engrailed to pattern the midbrain-hindbrain and diencephalic-
mesencephalic boundaries. Dev Biol. 301, 504–17.
205
Espinosa, J. S. and Stryker, M. P. (2012). Development and plasticity of the primary visual
cortex. Neuron. 75, 230–49.
Fadool, J. M. and Dowling, J. E. (2008). Zebrafish: a model system for the study of eye
genetics. Prog Retin Eye Res. 27, 89–110.
Farr, G. H., Ferkey, D. M., Yost, C., Pierce, S. B., Weaver, C. and Kimelman, D. (2000).
Interaction among GSK-3, GBP, axin, and APC in Xenopus axis specification. J Cell Biol.
148, 691–702.
Farris, S. M. (2008). Evolutionary convergence of higher brain centers spanning the
protostome-deuterostome boundary. Brain Behav Evol. 72, 106–22.
Fedtsova, N. and Turner, E. E. (2001). Signals from the ventral midline and isthmus regulate
the development of Brn3.0-expressing neurons in the midbrain. Mech Dev. 105, 129–44.
Feijóo, C. G., Oñate, M. G., Milla, L. a and Palma, V. a (2011). Sonic hedgehog (Shh)-Gli
signaling controls neural progenitor cell division in the developing tectum in zebrafish. Eur
J Neurosci. 33, 589–98.
Ferran, J. L., Sánchez-Arrones, L., Sandoval, J. E. and Puelles, L. (2007). A model of early
molecular regionalization in the chicken embryonic pretectum. J Comp Neurol. 505, 379–
403.
Finkelstein, R., Smouse, D., Capaci, T. M., Spradling, A. C. and Perrimon, N. (1990). The
orthodenticle gene encodes a novel homeo domain protein involved in the development of
the Drosophila nervous system and ocellar visual structures. Genes Dev. 4, 1516–27.
Fischbach, K.-F. and Hiesinger, P. R. (2008). Optic lobe development. Adv Exp Med Biol.
628, 115–36.
Fischer, A. J. and Bongini, R. (2010). Turning Müller glia into neural progenitors in the retina.
Mol Neurobiol. 42, 199–209.
206
Fischer, T., Guimera, J., Wurst, W. and Prakash, N. (2007). Distinct but redundant
expression of the Frizzled Wnt receptor genes at signaling centers of the developing mouse
brain. Neuroscience. 147, 693–711.
Foucher, I., Mione, M., Simeone, A., Acampora, D., Bally-Cuif, L. and Houart, C. (2006).
Differentiation of cerebellar cell identities in absence of Fgf signalling in zebrafish Otx
morphants. Development. 133, 1891–900.
Freedman, E. G. and Sparks, D. L. (1997). Activity of Cells in the Deeper Layers of the
Superior Colliculus of the Rhesus Monkey: Evidence for a Gaze Displacement Command.
J Neurophysiol. 78, 1669–1690.
French, C. R., Erickson, T., Callander, D., Berry, K. M., Koss, R., Hagey, D. W., Stout, J.,
Wuennenberg-Stapleton, K., Ngai, J., Moens, C. B., et al. (2007). Pbx homeodomain
proteins pattern both the zebrafish retina and tectum. BMC Dev Biol. 7, 85.
Fuchs, J., Stettler, O., Alvarez-Fischer, D., Prochiantz, A., Moya, K. L. and Joshi, R. L.
(2012). Engrailed signaling in axon guidance and neuron survival. Eur J Neurosci. 35,
1837–45.
Fuhrmann, S. (2008). Wnt signaling in eye organogenesis. Organogenesis 4, 60–7.
Fukami-Kobayashi, J. and Mitsui, Y. (1999). Cyclin D1 inhibits cell proliferation through
binding to PCNA and cdk2. Exp Cell Res. 246, 338–47.
Fung, F. K. C., Chan, D. W., Liu, V. W. S., Leung, T. H. Y., Cheung, A. N. Y. and Ngan,
H. Y. S. (2012). Increased expression of PITX2 transcription factor contributes to ovarian
cancer progression. PLoS One. 7, e37076.
Funnell, A. P. W. and Crossley, M. (2012). Homo- and heterodimerization in transcriptional
regulation. Adv Exp Med Biol. 747, 105–21.
Gahtan, E. and Baier, H. (2004). Of lasers, mutants, and see-through brains: functional
neuroanatomy in zebrafish. J Neurobiol. 59, 147–61.
207
Gahtan, E., Tanger, P. and Baier, H. (2005). Visual prey capture in larval zebrafish is
controlled by identified reticulospinal neurons downstream of the tectum. J Neurosci. 25,
9294–303.
Galliot, B., de Vargas, C. and Miller, D. (1999). Evolution of homeobox genes: Q50 Paired-
like genes founded the Paired class. Dev Genes Evol. 209, 186–97.
Garnaas, M. K., Cutting, C. C., Meyers, A., Kelsey, P. B., Harris, J. M., North, T. E. and
Goessling, W. (2012). Rargb regulates organ laterality in a zebrafish model of right atrial
isomerism. Dev Biol. 372, 178–89.
Gestri, G., Carl, M., Appolloni, I., Wilson, S. W., Barsacchi, G. and Andreazzoli, M.
(2005). Six3 functions in anterior neural plate specification by promoting cell proliferation
and inhibiting Bmp4 expression. Development. 132, 2401–13.
Ghanem, N., Jarinova, O., Amores, A., Long, Q., Hatch, G., Park, B. K., Rubenstein, J. L.
R. and Ekker, M. (2003). Regulatory roles of conserved intergenic domains in vertebrate
Dlx bigene clusters. Genome Res. 13, 533–43.
Ghanem, N., Yu, M., Long, J., Hatch, G., Rubenstein, J. L. R. and Ekker, M. (2007).
Distinct cis-regulatory elements from the Dlx1/Dlx2 locus mark different progenitor cell
populations in the ganglionic eminences and different subtypes of adult cortical
interneurons. J Neurosci. 27, 5012–22.
Glaser, T., Walton, D. S. and Maas, R. L. (1992). Genomic structure, evolutionary
conservation and aniridia mutations in the human PAX6 gene. Nat Genet. 2, 232–9.
Gogoi, R. N., Schubert, F. R., Martinez-Barbera, J.-P., Acampora, D., Simeone, A. and
Lumsden, A. (2002). The paired-type homeobox gene Dmbx1 marks the midbrain and
pretectum. Mech Dev. 114, 213–7.
Gomez, G., Lee, J.-H., Veldman, M. B., Lu, J., Xiao, X. and Lin, S. (2012). Identification of
vascular and hematopoietic genes downstream of etsrp by deep sequencing in zebrafish.
PLoS One. 7, e31658.
208
Goriely, a, Mollereau, B., Coffinier, C. and Desplan, C. (1999). Munster, a novel paired-class
homeobox gene specifically expressed in the Drosophila larval eye. Mech Dev. 88, 107–10.
Green, E. S., Stubbs, J. L. and Levine, E. M. (2003). Genetic rescue of cell number in a
mouse model of microphthalmia: interactions between Chx10 and G1-phase cell cycle
regulators. Development. 130, 539–52.
Grindley, J. C., Davidson, D. R. and Hill, R. E. (1995). The role of Pax-6 in eye and nasal
development. Development. 121, 1433–42.
Gyda, M., Wolman, M., Lorent, K. and Granato, M. (2012). The tumor suppressor gene
retinoblastoma-1 is required for retinotectal development and visual function in zebrafish.
PLoS Genet. 8, e1003106.
Hahn, M. W. (2009). Distinguishing among evolutionary models for the maintenance of gene
duplicates. J Hered. 100, 605–17.
Halder, G., Callaerts, P. and Gehring, W. J. (1995). Induction of ectopic eyes by targeted
expression of the eyeless gene in Drosophila. Science. 267, 1788–92.
Han, M. V, Demuth, J. P., McGrath, C. L., Casola, C. and Hahn, M. W. (2009). Adaptive
evolution of young gene duplicates in mammals. Genome Res. 19, 859–67.
Hanks, M., Wurst, W., Anson-Cartwright, L., Auerbach, A. B. and Joyner, A. L. (1995).
Rescue of the En-1 mutant phenotype by replacement of En-1 with En-2. Science. 269,
679–82.
Harbour, J. W. and Dean, D. C. (2000). The Rb/E2F pathway: expanding roles and emerging
paradigms. Genes Dev. 14, 2393–409.
Hart, M. J., de los Santos, R., Albert, I. N., Rubinfeld, B. and Polakis, P. (1998).
Downregulation of beta-catenin by human Axin and its association with the APC tumor
suppressor, beta-catenin and GSK3 beta. Curr Biol. 8, 573–81.
Hatakeyama, J., Tomita, K., Inoue, T. and Kageyama, R. (2001). Roles of homeobox and
bHLH genes in specification of a retinal cell type. Development. 128, 1313–22.
209
Hawryshyn, C. W. (2010). Ultraviolet polarization vision and visually guided behavior in
fishes. Brain Behav Evol. 75, 186–94.
Hensley, S. H., Yang, X. L. and Wu, S. M. (1993). Relative contribution of rod and cone
inputs to bipolar cells and ganglion cells in the tiger salamander retina. J Neurophysiol. 69,
2086–98.
Hidalgo-Sánchez, M., Millet, S., Bloch-Gallego, E. and Alvarado-Mallart, R.-M. (2005).
Specification of the meso-isthmo-cerebellar region: the Otx2/Gbx2 boundary. Brain Res
Brain Res Rev. 49, 134–49.
Higashijima, S., Hotta, Y. and Okamoto, H. (2000). Visualization of cranial motor neurons in
live transgenic zebrafish expressing green fluorescent protein under the control of the islet-
1 promoter/enhancer. J Neurosci. 20, 206–18.
Higashijima, S.-I., Mandel, G. and Fetcho, J. R. (2004a). Distribution of prospective
glutamatergic, glycinergic, and GABAergic neurons in embryonic and larval zebrafish. J
Comp Neurol. 480, 1–18.
Higashijima, S.-I., Schaefer, M. and Fetcho, J. R. (2004b). Neurotransmitter properties of
spinal interneurons in embryonic and larval zebrafish. J Comp Neurol. 480, 19–37.
Hirsch, N. and Harris, W. A. (1997a). Xenopus Pax-6 and retinal development. J Neurobiol.
32, 45–61.
Hirsch, N. and Harris, W. A. (1997b). Xenopus Brn-3.0, a POU-domain gene expressed in the
developing retina and tectum. Not regulated by innervation. Invest Ophthalmol Vis Sci. 38,
960–9.
Hislop, N. R., de Jong, D., Hayward, D. C., Ball, E. E. and Miller, D. J. (2005). Tandem
organization of independently duplicated homeobox genes in the basal cnidarian Acropora
millepora. Dev Genes Evol. 215, 268–73.
Holland, P. W. H. and Takahashi, T. (2005). The evolution of homeobox genes: Implications
for the study of brain development. Brain Res Bull. 66, 484–90.
210
Hortopan, G. a and Baraban, S. C. (2011). Aberrant expression of genes necessary for
neuronal development and Notch signaling in an epileptic mind bomb zebrafish. Dev Dyn.
240, 1964–76.
Hsu, S. C., Galceran, J. and Grosschedl, R. (1998). Modulation of transcriptional regulation
by LEF-1 in response to Wnt-1 signaling and association with beta-catenin. Mol Cell Biol.
18, 4807–18.
Hu, M. and Easter, S. S. (1999). Retinal neurogenesis: the formation of the initial central patch
of postmitotic cells. Dev Biol. 207, 309–21.
Huberman, A. D. and Niell, C. M. (2011). What can mice tell us about how vision works?
Trends Neurosci. 34, 464–73.
Huminiecki, L. and Wolfe, K. H. (2004). Divergence of spatial gene expression profiles
following species-specific gene duplications in human and mouse. Genome Res. 14, 1870–
9.
Hurley, I. A., Scemama, J.-L. and Prince, V. E. (2007). Consequences of hoxb1 duplication
in teleost fish. Evol Dev. 9, 540–54.
Hwang, W. Y., Fu, Y., Reyon, D., Maeder, M. L., Tsai, S. Q., Sander, J. D., Peterson, R. T.,
Yeh, J.-R. J. and Joung, J. K. (2013). Efficient genome editing in zebrafish using a
CRISPR-Cas system. Nat Biotechnol. 31, 1–3.
Ichijo, H. (1999). Differentiation of the chick retinotectal topographic map by remodeling in
specificity and refinement in accuracy. Brain Res Dev Brain Res. 117, 199–211.
Imai, K. S., Stolfi, A., Levine, M. and Satou, Y. (2009). Gene regulatory networks underlying
the compartmentalization of the Ciona central nervous system. Development. 136, 285–93.
Inoue, A., Takahashi, M., Hatta, K., Hotta, Y. and Okamoto, H. (1994). Developmental
regulation of islet-1 mRNA expression during neuronal differentiation in embryonic
zebrafish. Dev Dyn. 199, 1–11.
211
Ito, Y., Tanaka, H., Okamoto, H. and Ohshima, T. (2010). Characterization of neural stem
cells and their progeny in the adult zebrafish optic tectum. Dev Biol. 342, 26–38.
Itoh, M., Kim, C.-H., Palardy, G., Oda, T., Jiang, Y.-J., Maust, D., Yeo, S.-Y., Lorick, K.,
Wright, G. J., Ariza-McNaughton, L., et al. (2003). Mind bomb is a ubiquitin ligase that
is essential for efficient activation of Notch signaling by Delta. Dev Cell. 4, 67–82.
Jacobson, E. M., Li, P., Leon-del-Rio, A., Rosenfeld, M. G. and Aggarwal, A. K. (1997).
Structure of Pit-1 POU domain bound to DNA as a dimer: unexpected arrangement and
flexibility. Genes Dev. 11, 198–212.
Jarman, A. P., Grell, E. H., Ackerman, L., Jan, L. Y. and Jan, Y. N. (1994). Atonal is the
proneural gene for Drosophila photoreceptors. Nature. 369, 398–400.
Jászai, J., Reifers, F., Picker, A., Langenberg, T. and Brand, M. (2003). Isthmus-to-
midbrain transformation in the absence of midbrain-hindbrain organizer activity.
Development. 130, 6611–23.
Jiang, Y. J., Brand, M., Heisenberg, C. P., Beuchle, D., Furutani-Seiki, M., Kelsh, R. N.,
Warga, R. M., Granato, M., Haffter, P., Hammerschmidt, M., et al. (1996). Mutations
affecting neurogenesis and brain morphology in the zebrafish, Danio rerio. Development.
123, 205–16.
John, P. C., Mews, M. and Moore, R. (2001). Cyclin/Cdk complexes: their involvement in cell
cycle progression and mitotic division. Protoplasma. 216, 119–42.
Jung, S.-H., Kim, H.-S., Ryu, J.-H., Gwak, J.-W., Bae, Y.-K., Kim, C.-H. and Yeo, S.-Y.
(2012). Her4-positive population in the tectum opticum is proliferating neural precursors in
the adult zebrafish brain. Mol Cells. 33, 627–32.
Kaas, J. H. (1989). The evolution of complex sensory systems in mammals. J Exp Biol. 146,
165–76.
Kane, D. A. and Kimmel, C. B. (1993). The zebrafish midblastula transition. Development.
119, 447–56.
212
Karlstrom, R. O., Trowe, T. and Bonhoeffer, F. (1997). Genetic analysis of axon guidance
and mapping in the zebrafish. Trends Neurosci. 20, 3–8.
Kassahn, K. S., Dang, V. T., Wilkins, S. J., Perkins, A. C. and Ragan, M. A. (2009).
Evolution of gene function and regulatory control after whole-genome duplication:
comparative analyses in vertebrates. Genome Res. 19, 1404–18.
Kawahara, A., Chien, C.-B. and Dawid, I. B. (2002). The Homeobox Gene mbx Is Involved
in Eye and Tectum Development. Dev Biol. 248, 107–117.
Kay, J. N., Finger-Baier, K. C., Roeser, T., Staub, W. and Baier, H. (2001). Retinal ganglion
cell genesis requires lakritz, a Zebrafish atonal Homolog. Neuron. 30, 725–36.
Kefalov, V. J. (2012). Rod and cone visual pigments and phototransduction through
pharmacological, genetic, and physiological approaches. J Biol Chem. 287, 1635–41.
Kennedy, B. N., Stearns, G. W., Smyth, V. a, Ramamurthy, V., van Eeden, F.,
Ankoudinova, I., Raible, D., Hurley, J. B. and Brockerhoff, S. E. (2004). Zebrafish rx3
and mab21l2 are required during eye morphogenesis. Dev Biol. 270, 336–49.
Kikuchi, Y., Segawa, H., Tokumoto, M., Tsubokawa, T., Hotta, Y., Uyemura, K. and
Okamoto, H. (1997). Ocular and cerebellar defects in zebrafish induced by overexpression
of the LIM domains of the islet-3 LIM/homeodomain protein. Neuron. 18, 369–82.
Kikuta, H., Kanai, M., Ito, Y. and Yamasu, K. (2003). gbx2 Homeobox gene is required for
the maintenance of the isthmic region in the zebrafish embryonic brain. Dev Dyn. 228,
433–50.
Kim, W.-Y. and Shen, J. (2008). Presenilins are required for maintenance of neural stem cells
in the developing brain. Mol Neurodegener. 3, 2.
Kimmel, C. (1993). Patterning the brain of the zebrafish embryo. Annu Rev Neurosci. 16, 707–
32.
Kimmel, C. B., Warga, R. M. and Schilling, T. F. (1990). Origin and organization of the
zebrafish fate map. Development. 108, 581–94.
213
Kimmel, C. B., Warga, R. M. and Kane, D. A. (1994). Cell cycles and clonal strings during
formation of the zebrafish central nervous system. Development. 120, 265–76.
Kimmel, C. B., Ballard, W. W., Kimmel, S. R., Ullmann, B. and Schilling, T. F. (1995).
Stages of embryonic development of the zebrafish. Dev Dyn. 203, 253–310.
Kimura, M. (1983). Rare variant alleles in the light of the neutral theory. Mol Biol Evol. 1, 84–
93.
Kimura, K., Miki, T., Shibasaki, T., Zhang, Y., Ogawa, M., Saisho, H., Okuno, M.,
Iwanaga, T. and Seino, S. (2005). Functional analysis of transcriptional repressor
Otx3/Dmbx1. FEBS Lett. 579, 2926–32.
Kizil, C., Kaslin, J., Kroehne, V. and Brand, M. (2012). Adult neurogenesis and brain
regeneration in zebrafish. Dev Neurobiol. 72, 429–61.
Klein, E. a and Assoian, R. K. (2008). Transcriptional regulation of the cyclin D1 gene at a
glance. J Cell Sci. 121, 3853–7.
Kleinjan, D. A., Bancewicz, R. M., Gautier, P., Dahm, R., Schonthaler, H. B., Damante, G.,
Seawright, A., Hever, A. M., Yeyati, P. L., van Heyningen, V., et al. (2008).
Subfunctionalization of duplicated zebrafish pax6 genes by cis-regulatory divergence.
PLoS Genet. 4, e29.
Knoepfler, P. S., Bergstrom, D. A., Uetsuki, T., Dac-Korytko, I., Sun, Y. H., Wright, W. E.,
Tapscott, S. J. and Kamps, M. P. (1999). A conserved motif N-terminal to the DNA-
binding domains of myogenic bHLH transcription factors mediates cooperative DNA
binding with pbx-Meis1/Prep1. Nucleic Acids Res. 27, 3752–61.
Koizumi, O., Itazawa, M., Mizumoto, H., Minobe, S., Javois, L. C., Grimmelikhuijzen, C.
J. and Bode, H. R. (1992). Nerve ring of the hypostome in hydra. I. Its structure,
development, and maintenance. J Comp Neurol. 326, 7–21.
Kokkinopoulos, I., Pearson, R. A., Macneil, A., Dhomen, N. S., MacLaren, R. E., Ali, R. R.
and Sowden, J. C. (2008). Isolation and characterisation of neural progenitor cells from
the adult Chx10(orJ/orJ) central neural retina. Mol Cell Neurosci. 38, 359–73.
214
Koyama, M., Kinkhabwala, A., Satou, C., Higashijima, S. and Fetcho, J. (2011). Mapping a
sensory-motor network onto a structural and functional ground plan in the hindbrain. Proc
Natl Acad Sci U S A. 108, 1170–5.
Kozmik, Z., Daube, M., Frei, E., Norman, B., Kos, L., Dishaw, L. J., Noll, M. and
Piatigorsky, J. (2003). Role of Pax genes in eye evolution: a cnidarian PaxB gene uniting
Pax2 and Pax6 functions. Dev Cell. 5, 773–85.
Kozmik, Z., Ruzickova, J., Jonasova, K., Matsumoto, Y., Vopalensky, P., Kozmikova, I.,
Strnad, H., Kawamura, S., Piatigorsky, J., Paces, V., et al. (2008). Assembly of the
cnidarian camera-type eye from vertebrate-like components. Proc Natl Acad Sci U S A.
105, 8989–8993.
Krauss, S., Johansen, T., Korzh, V., Moens, U., Ericson, J. U. and Fjose, A. (1991).
Zebrafish pax[zf-a]: a paired box-containing gene expressed in the neural tube. EMBO J.
10, 3609–3619.
Krauss, S., Korzh, V., Fjose, a and Johansen, T. (1992). Expression of four zebrafish wnt-
related genes during embryogenesis. Development. 116, 249–59.
Krauss, S., Concordet, J. P. and Ingham, P. W. (1993). A functionally conserved homolog of
the Drosophila segment polarity gene hh is expressed in tissues with polarizing activity in
zebrafish embryos. Cell. 75, 1431–44.
Kunz, M., Herrmann, M., Wedlich, D. and Gradl, D. (2004). Autoregulation of canonical
Wnt signaling controls midbrain development. Dev Biol. 273, 390–401.
Kurokawa, D., Sakurai, Y., Inoue, A., Nakayama, R., Takasaki, N., Suda, Y., Miyake, T.,
Amemiya, C. T. and Aizawa, S. (2006). Evolutionary constraint on Otx2 neuroectoderm
enhancers-deep conservation from skate to mouse and unique divergence in teleost. Proc
Natl Acad Sci U S A. 103, 19350–5.
Kuwahara, A., Hirabayashi, Y., Knoepfler, P. S., Taketo, M. M., Sakai, J., Kodama, T.
and Gotoh, Y. (2010). Wnt signaling and its downstream target N-myc regulate basal
progenitors in the developing neocortex. Development. 137, 1035–44.
215
Laessing, U. and Stuermer, C. a (1996). Spatiotemporal pattern of retinal ganglion cell
differentiation revealed by the expression of neurolin in embryonic zebrafish. J Neurobiol.
29, 65–74.
Land, M. F. (1997). Visual acuity in insects. Annu Rev Entomol. 42, 147–77.
Lane, B. M. and Lister, J. a (2012). Otx but not Mitf transcription factors are required for
zebrafish retinal pigment epithelium development. PLoS One. 7, e49357.
Langenberg, T. and Brand, M. (2005). Lineage restriction maintains a stable organizer cell
population at the zebrafish midbrain-hindbrain boundary. Development. 132, 3209–16.
Langenfeld, J., Kiyokawa, H., Sekula, D., Boyle, J. and Dmitrovsky, E. (1997).
Posttranslational regulation of cyclin D1 by retinoic acid: a chemoprevention mechanism.
Proc Natl Acad Sci U S A. 94, 12070–4.
Langmann, T. (2007). Microglia activation in retinal degeneration. J Leukoc Biol. 81, 1345–51.
Larison, K. D. and Bremiller, R. (1990). Early onset of phenotype and cell patterning in the
embryonic zebrafish retina. Development. 109, 567–76.
Lee, B. B. (2011). Visual pathways and psychophysical channels in the primate. J Physiol. 589,
41–7.
Lee, S. M., Tole, S., Grove, E. and McMahon, A. P. (2000). A local Wnt-3a signal is required
for development of the mammalian hippocampus. Development. 127, 457–67.
Lekven, A. C., Buckles, G. R., Kostakis, N. and Moon, R. T. (2003). Wnt1 and wnt10b
function redundantly at the zebrafish midbrain-hindbrain boundary. Dev Biol. 254, 172–87.
Leopold, D. A. (2012). Primary visual cortex: awareness and blindsight. Annu Rev Neurosci.
35, 91–109.
Lesaffre, B., Joliot, A., Prochiantz, A. and Volovitch, M. (2007). Direct non-cell autonomous
Pax6 activity regulates eye development in the zebrafish. Neural Dev. 2, 2.
216
Lesser, M. P., Carleton, K. L., Böttger, S. a, Barry, T. M. and Walker, C. W. (2011). Sea
urchin tube feet are photosensory organs that express a rhabdomeric-like opsin and PAX6.
Proc Biol Sci. 278, 3371–9.
Li, J. Y. and Joyner, a L. (2001). Otx2 and Gbx2 are required for refinement and not induction
of mid-hindbrain gene expression. Development. 128, 4979–91.
Li, H. S., Yang, J. M., Jacobson, R. D., Pasko, D. and Sundin, O. (1994a). Pax-6 is first
expressed in a region of ectoderm anterior to the early neural plate: implications for
stepwise determination of the lens. Dev Biol. 162, 181–94.
Li, Y., Allende, M. L., Finkelstein, R. and Weinberg, E. S. (1994b). Expression of two
zebrafish orthodenticle-related genes in the embryonic brain. Mech Dev. 48, 229–44.
Li, Z., Hu, M., Ochocinska, M. J., Joseph, N. M. and Easter, S. S. (2000). Modulation of cell
proliferation in the embryonic retina of zebrafish (Danio rerio). Dev Dyn. 219, 391–401.
Li, W.-H., Yang, J. and Gu, X. (2005). Expression divergence between duplicate genes.
Trends Genet. 21, 602–7.
Li, V. S. W., Ng, S. S., Boersema, P. J., Low, T. Y., Karthaus, W. R., Gerlach, J. P.,
Mohammed, S., Heck, A. J. R., Maurice, M. M., Mahmoudi, T., et al. (2012). Wnt
signaling through inhibition of β-catenin degradation in an intact Axin1 complex. Cell.
149, 1245–56.
Liu, K. J. and Harland, R. M. (2003). Cloning and characterization of Xenopus Id4 reveals
differing roles for Id genes. Dev Biol. 264, 339–51.
Liu, W., Khare, S. L., Liang, X., Peters, M. A., Liu, X., Cepko, C. L. and Xiang, M. (2000).
All Brn3 genes can promote retinal ganglion cell differentiation in the chick. Development.
127, 3237–47.
Liu, X., Huang, S., Ma, J., Li, C., Zhang, Y. and Luo, L. (2009). NF-kappaB and Snail1a
coordinate the cell cycle with gastrulation. J Cell Biol. 184, 805–15.
217
Livne-Bar, I., Pacal, M., Cheung, M. C., Hankin, M., Trogadis, J., Chen, D., Dorval, K. M.
and Bremner, R. (2006). Chx10 is required to block photoreceptor differentiation but is
dispensable for progenitor proliferation in the postnatal retina. Proc Natl Acad Sci U S A.
103, 4988–93.
London, N. J. S., Kessler, P., Williams, B., Pauer, G. J., Hagstrom, S. A. and Traboulsi, E.
I. (2009). Sequence alterations in RX in patients with microphthalmia, anophthalmia, and
coloboma. Mol Vis. 15, 162–7.
Loosli, F., Winkler, S. and Wittbrodt, J. (1999). Six3 overexpression initiates the formation
of ectopic retina. Genes Dev. 13, 649–54.
Loosli, F., Staub, W., Finger-Baier, K. C., Ober, E. a, Verkade, H., Wittbrodt, J. and
Baier, H. (2003). Loss of eyes in zebrafish caused by mutation of chokh/rx3. EMBO
reports. 4, 894–9.
Lorent, K., Liu, K. S., Fetcho, J. R. and Granato, M. (2001). The zebrafish space cadet gene
controls axonal pathfinding of neurons that modulate fast turning movements.
Development. 128, 2131–42.
Low, L. K. and Cheng, H.-J. (2006). Axon pruning: an essential step underlying the
developmental plasticity of neuronal connections. Philos Trans R Soc Lond B Biol Sci. 361,
1531–44.
Lowery, L. A. and Sive, H. (2004). Strategies of vertebrate neurulation and a re-evaluation of
teleost neural tube formation. Mech Dev. 121, 1189–97.
Lu, J., Peatman, E., Tang, H., Lewis, J. and Liu, Z. (2012). Profiling of gene duplication
patterns of sequenced teleost genomes: evidence for rapid lineage-specific genome
expansion mediated by recent tandem duplications. BMC Genomics. 13, 246.
Lun, K. and Brand, M. (1998). A series of no isthmus (noi) alleles of the zebrafish pax2.1
gene reveals multiple signaling events in development of the midbrain-hindbrain boundary.
Development. 125, 3049–62.
218
Lünenburger, L., Kleiser, R., Stuphorn, V., Miller, L. E. and Hoffmann, K. P. (2001). A
possible role of the superior colliculus in eye-hand coordination. Prog Brain Res. 134,
109–25.
Lynch, M. and Force, A. (2000). The probability of duplicate gene preservation by
subfunctionalization. Genetics. 154, 459–73.
MacCarthy, T. and Bergman, A. (2007). The limits of subfunctionalization. BMC Evol Biol.
7, 213.
Macdonald, R., Barth, K. A., Xu, Q., Holder, N., Mikkola, I. and Wilson, S. W. (1995).
Midline signalling is required for Pax gene regulation and patterning of the eyes.
Development. 121, 3267–78.
MacNeil, M. A. and Masland, R. H. (1998). Extreme diversity among amacrine cells:
implications for function. Neuron. 20, 971–82.
Makarova, K. S., Haft, D. H., Barrangou, R., Brouns, S. J. J., Charpentier, E., Horvath, P.,
Moineau, S., Mojica, F. J. M., Wolf, Y. I., Yakunin, A. F., et al. (2011). Evolution and
classification of the CRISPR-Cas systems. Nat Rev Microbiol. 9, 467–77.
Marcus, R. C., Delaney, C. L. and Easter, S. S. (1999). Neurogenesis in the visual system of
embryonic and adult zebrafish (Danio rerio). off. Vis Neurosci. 16, 417–24.
Marlow, H. Q., Srivastava, M., Matus, D. Q., Rokhsar, D. and Martindale, M. Q. (2009).
Anatomy and development of the nervous system of Nematostella vectensis, an anthozoan
cnidarian. Dev Neurobiol. 69, 235–54.
Marquardt, T. (2003). Transcriptional control of neuronal diversification in the retina. Prog
Retin Eye Res. 22, 567–77.
Marquardt, T., Ashery-Padan, R., Andrejewski, N., Scardigli, R., Guillemot, F. and Gruss,
P. (2001). Pax6 is required for the multipotent state of retinal progenitor cells. Cell. 105,
43–55.
219
Martinez-Morales, J. R., Signore, M., Acampora, D., Simeone, A. and Bovolenta, P. (2001).
Otx genes are required for tissue specification in the developing eye. Development. 128,
2019–30.
Massingham, T., Davies, L. J. and Liò, P. (2001). Analysing gene function after duplication.
BioEssays. 23, 873–6.
Mathers, P. H., Grinberg, A., Mahon, K. A. and Jamrich, M. (1997). The Rx homeobox
gene is essential for vertebrate eye development. Nature. 387, 603–7.
Matsunaga, E., Araki, I. and Nakamura, H. (2000). Pax6 defines the di-mesencephalic
boundary by repressing En1 and Pax2. Development. 127, 2357–65.
Matsunaga, E., Araki, I. and Nakamura, H. (2001). Role of Pax3/7 in the tectum
regionalization. Development. 128, 4069–77.
Mattes, B., Weber, S., Peres, J., Chen, Q., Davidson, G., Houart, C. and Scholpp, S. (2012).
Wnt3 and Wnt3a are required for induction of the mid-diencephalic organizer in the caudal
forebrain. Neural Dev. 7, 12.
Mazet, F. and Shimeld, S. M. (2002). Gene duplication and divergence in the early evolution
of vertebrates. Curr Opin Genet Dev. 12, 393–6.
McMahon, A. P. and Moon, R. T. (1989). Ectopic expression of the proto-oncogene int-1 in
Xenopus embryos leads to duplication of the embryonic axis. Cell. 58, 1075–84.
Megason, S. and McMahon, A. (2002). A mitogen gradient of dorsal midline Wnts organizes
growth in the CNS. Development. 2098, 2087–2098.
Mercier, P., Simeone, a, Cotelli, F. and Boncinelli, E. (1995). Expression pattern of two otx
genes suggests a role in specifying anterior body structures in zebrafish. Int J Dev Biol. 39,
559–73.
Meyer, A. and Van de Peer, Y. (2005). From 2R to 3R: evidence for a fish-specific genome
duplication (FSGD). BioEssays. 27, 937–45.
220
Meyers, J. R., Hu, L., Moses, A., Kaboli, K., Papandrea, A. and Raymond, P. A. (2012). β-
catenin/Wnt signaling controls progenitor fate in the developing and regenerating zebrafish
retina. Neural Dev. 7, 30.
Mishra, M., Oke, A., Lebel, C., McDonald, E. C., Plummer, Z., Cook, T. A. and Zelhof, A.
C. (2010). Pph13 and orthodenticle define a dual regulatory pathway for photoreceptor cell
morphogenesis and function. Development. 137, 2895–904.
Miyamoto, T., Kawahara, A., Teufel, A., Mukhopadhyay, M., Zhao, Y., Dawid, I. B. and
Westphal, H. (2002). Mbx, a novel mouse homeobox gene. Dev Genes Evol. 212, 104–6.
MJ, S., SJ, F., D, S., JH, C., AJ, C. and E, D. (1999). Retinoic Acid Promotes Ubiquitination
and Proteolysis of Cyclin D1 during Induced Tumor Cell Differentiation. J Biol Chem. 274,
22013–22018.
Moens, C. B., Yan, Y. L., Appel, B., Force, A. G. and Kimmel, C. B. (1996). valentino: a
zebrafish gene required for normal hindbrain segmentation. Development. 122, 3981–90.
Molven, A., Rasmus, P. and Fjose, A. (1991). Genomic structure and restricted neural
expression of. EMBO J. 10, 799–807.
Morrow, E. M., Furukawa, T., Lee, J. E. and Cepko, C. L. (1999). NeuroD regulates
multiple functions in the developing neural retina in rodent. Development. 126, 23–36.
Morshead, C. M. and van der Kooy, D. (1992). Postmitotic death is the fate of constitutively
proliferating cells in the subependymal layer of the adult mouse brain. J Neurosci. 12, 249–
56.
Mu, X., Fu, X., Sun, H., Beremand, P. D., Thomas, T. L. and Klein, W. H. (2005). A gene
network downstream of transcription factor Math5 regulates retinal progenitor cell
competence and ganglion cell fate. Dev Biol. 280, 467–81.
Mueller, T. and Wullimann, M. F. (2003). Anatomy of neurogenesis in the early zebrafish
brain. Brain Res Dev Brain Res. 140, 137–55.
221
Mueller, T. and Wullimann, M. (2005). Atlas of Early Zebrafish Brain Development: A Tool
for Molecular Neurogenetics. 1st ed. Amsterdam: Elsevier Science.
Münchberg, S. R., Ober, E. A. and Steinbeisser, H. (1999). Expression of the Ets
transcription factors erm and pea3 in early zebrafish development. Mech Dev. 88, 233–6.
Muto, A., Orger, M. B., Wehman, A. M., Smear, M. C., Kay, J. N., Page-McCaw, P. S.,
Gahtan, E., Xiao, T., Nevin, L. M., Gosse, N. J., et al. (2005). Forward genetic analysis
of visual behavior in zebrafish. PLoS Genet. 1, e66.
Muto, A., Ohkura, M., Abe, G., Nakai, J. and Kawakami, K. (2013). Real-Time
Visualization of Neuronal Activity during Perception. Curr Biol. 23, 307–11.
Nakamura, H. (2001). Regionalisation and acquisition of polarity in the optic tectum. Prog
Neurobiol. 65, 473–88.
Nakamura, H. and O’Leary, D. D. (1989). Inaccuracies in initial growth and arborization of
chick retinotectal axons followed by course corrections and axon remodeling to develop
topographic order. J Neurosci. 9, 3776–95.
Nasevicius, a and Ekker, S. C. (2000). Effective targeted gene “knockdown” in zebrafish. Nat
Genet. 26, 216–20.
Nassi, J. J. and Callaway, E. M. (2009). Parallel processing strategies of the primate visual
system. Nat Rev Neurosci. 10, 360–72.
Navidi, W. C., Churchill, G. A. and von Haeseler, A. (1991). Methods for inferring
phylogenies from nucleic acid sequence data by using maximum likelihood and linear
invariants. Mol Biol Evol. 8, 128–43.
Nevin, L. M., Taylor, M. R. and Baier, H. (2008). Hardwiring of fine synaptic layers in the
zebrafish visual pathway. Neural Dev. 3, 36.
Nevin, L. M., Robles, E., Baier, H. and Scott, E. K. (2010). Focusing on optic tectum circuitry
through the lens of genetics. BMC Biol. 8, 126.
222
Nikaido, M., Law, E. W. P. and Kelsh, R. N. (2013). A systematic survey of expression and
function of zebrafish frizzled genes. PLoS One. 8, e54833.
Nikolaou, N., Lowe, A. S., Walker, A. S., Abbas, F., Hunter, P. R., Thompson, I. D. and
Meyer, M. P. (2012). Parametric functional maps of visual inputs to the tectum. Neuron.
76, 317–24.
Nilsson, D.-E. (2009). The evolution of eyes and visually guided behaviour. Philos Trans R Soc
Lond B Biol Sci. 364, 2833–47.
Nornes, S., Clarkson, M., Mikkola, I., Pedersen, M., Bardsley, a, Martinez, J. P., Krauss,
S. and Johansen, T. (1998). Zebrafish contains two pax6 genes involved in eye
development. Mech Dev. 77, 185–96.
Nyholm, M. K., Wu, S.-F., Dorsky, R. I. and Grinblat, Y. (2007). The zebrafish zic2a-zic5
gene pair acts downstream of canonical Wnt signaling to control cell proliferation in the
developing tectum. Development. 134, 735–46.
Ochocinska, M. J. and Hitchcock, P. F. (2007). Dynamic Expression of the Basic Helix-
Loop-Helix Transcription Factor NeuroD in the Rod and Cone Photoreceptor Lineages in
the Retina of the Embryonic and Larval Zebrafish. J Comp Neurol. 501, 1–12.
Ochocinska, M. J. and Hitchcock, P. F. (2008). NeuroD regulates proliferation of
photoreceptor progenitors in the retina of the zebrafish. Mech Dev. 126, 128–41.
Odenthal, J., van Eeden, F. J., Haffter, P., Ingham, P. W. and Nüsslein-Volhard, C. (2000).
Two distinct cell populations in the floor plate of the zebrafish are induced by different
pathways. Dev Biol. 219, 350–63.
Ogata, K., Sato, K., Tahirov, T. H. and Tahirov, T. (2003). Eukaryotic transcriptional
regulatory complexes: cooperativity from near and afar. Curr Opin Struct Biol. 13, 40–8.
Ohno, S. (1999). Gene duplication and the uniqueness of vertebrate genomes circa 1970-1999.
Semin Cell Dev Biol. 10, 517–22.
223
Ohnuma, S., Hopper, S., Wang, K. C., Philpott, A. and Harris, W. a (2002). Co-ordinating
retinal histogenesis: early cell cycle exit enhances early cell fate determination in the
Xenopus retina. Development. 129, 2435–46.
Ohtoshi, A. and Behringer, R. (2004). Neonatal lethality, dwarfism, and abnormal brain
development in Dmbx1 mutant mice. Mol Cell Biol. 24, 7548–7558.
Ohtoshi, A., Nishijima, I., Justice, M. J. and Behringer, R. R. (2002). Dmbx1, a novel
evolutionarily conserved paired-like homeobox gene expressed in the brain of mouse
embryos. Mech Dev. 110, 241–4.
Ohtoshi, A., Bradley, A., Behringer, R. R. and Nishijima, I. (2006). Generation and
maintenance of Dmbx1 gene-targeted mutant alleles. Mamm Genome. 17, 744–50.
Okawa, H. and Sampath, A. P. (2007). Optimization of single-photon response transmission at
the rod-to-rod bipolar synapse. Physiology. 22, 279–86.
Omori, Y., Katoh, K., Sato, S., Muranishi, Y., Chaya, T., Onishi, A., Minami, T., Fujikado,
T. and Furukawa, T. (2011). Analysis of transcriptional regulatory pathways of
photoreceptor genes by expression profiling of the Otx2-deficient retina. PLoS One. 6,
e19685.
Osorio, D. and Bacon, J. P. (1994). A good eye for arthropod evolution. BioEssays. 16, 419–
24.
Ota, S., Ishitani, S., Shimizu, N., Matsumoto, K., Itoh, M. and Ishitani, T. (2012). NLK
positively regulates Wnt/β-catenin signalling by phosphorylating LEF1 in neural progenitor
cells. EMBO J. 31, 1904–15.
Otteson, D. C. and Hitchcock, P. F. (2003). Stem cells in the teleost retina: persistent
neurogenesis and injury-induced regeneration. Vision Res. 43, 927–36.
Otteson, D. C., Shelden, E., Jones, J. M., Kameoka, J. and Hitchcock, P. F. (1998). Pax2
expression and retinal morphogenesis in the normal and Krd mouse. Dev Biol. 193, 209–
24.
224
Otto, S. P. and Yong, P. (2002). The evolution of gene duplicates. Adv Genet. 46, 451–83.
Oxtoby, E. and Jowett, T. (1993). Cloning of the zebrafish krox-20 gene (krx-20) and its
expression during hindbrain development. Nucleic Acids Res. 21, 1087–95.
Pagano, M., Theodoras, A. M., Tam, S. W. and Draetta, G. F. (1994). Cyclin D1-mediated
inhibition of repair and replicative DNA synthesis in human fibroblasts. Genes Dev. 8,
1627–39.
Palczewski, K. (2012). Chemistry and biology of vision. J Biol Chem. 287, 1612–9.
Palstra, A. P., Beltran, S., Burgerhout, E., Brittijn, S. A., Magnoni, L. J., Henkel, C. V,
Jansen, H. J., van den Thillart, G. E. E. J. M., Spaink, H. P. and Planas, J. V (2013).
Deep RNA sequencing of the skeletal muscle transcriptome in swimming fish. PLoS One.
8, e53171.
Pan, L., Yang, Z., Feng, L. and Gan, L. (2005). Functional equivalence of Brn3 POU-domain
transcription factors in mouse retinal neurogenesis. Development. 132, 703–12.
Paridaen, J. T. M. L., Danesin, C., Elas, A. T., van de Water, S., Houart, C. and Zivkovic,
D. (2009). Apc1 is required for maintenance of local brain organizers and dorsal midbrain
survival. Dev Biol. 331, 101–12.
Passini, M. a, Levine, E. M., Canger, a K., Raymond, P. a and Schechter, N. (1997). Vsx-1
and Vsx-2: differential expression of two paired-like homeobox genes during zebrafish and
goldfish retinogenesis. J Comp Neurol. 388, 495–505.
Pegueroles, C., Laurie, S. and Albà, M. M. (2013). Accelerated evolution after gene
duplication: a time-dependent process affecting just one copy. Mol Biol Evol. first
published online April 26, 2013 doi:10.1093/.
Pellizzari, L., Tell, G., Fabbro, D., Pucillo, C. and Damante, G. (1997). Functional
interference between contacting amino acids of homeodomains. FEBS Lett. 407, 320–4.
225
Peterson, R. E., Fadool, J. M., McClintock, J. and Linser, P. J. (2001). Müller cell
differentiation in the zebrafish neural retina: evidence of distinct early and late stages in
cell maturation. J Comp Neurol. 429, 530–40.
Piatigorsky, J. and Kozmik, Z. (2004). Cubozoan jellyfish: an Evo/Devo model for eyes and
other sensory systems. Int J Dev Biol. 48, 719–29.
Picker, a, Brennan, C., Reifers, F., Clarke, J. D., Holder, N. and Brand, M. (1999).
Requirement for the zebrafish mid-hindbrain boundary in midbrain polarisation, mapping
and confinement of the retinotectal projection. Development. 126, 2967–78.
Pittman, A. J., Law, M.-Y. and Chien, C.-B. (2008). Pathfinding in a large vertebrate axon
tract: isotypic interactions guide retinotectal axons at multiple choice points. Development.
135, 2865–71.
Porfiri, E., Rubinfeld, B., Albert, I., Hovanes, K., Waterman, M. and Polakis, P. (1997).
Induction of a beta-catenin-LEF-1 complex by wnt-1 and transforming mutants of beta-
catenin. Oncogene. 15, 2833–9.
Portugues, R. and Engert, F. (2009). The neural basis of visual behaviors in the larval
zebrafish. Curr Opin Neurobiol. 19, 644–7.
Pujic, Z. and Malicki, J. (2004). Retinal pattern and the genetic basis of its formation in
zebrafish. Semin Cell Dev Biol. 15, 105–14.
Quinn, J. C., West, J. D. and Hill, R. E. (1996). Multiple functions for Pax6 in mouse eye and
nasal development. Genes Dev. 10, 435–46.
Quiring, R., Walldorf, U., Kloter, U. and Gehring, W. J. (1994). Homology of the eyeless
gene of Drosophila to the Small eye gene in mice and Aniridia in humans. Science. 265,
785–9.
Raible, F. and Brand, M. (2004). Divide et Impera--the midbrain-hindbrain boundary and its
organizer. Trends Neurosci. 27, 727–34.
226
Ramos, C., Vouyiouklis, D., Scott, M., Andras, A. and Lindsay, S. (2007). Embryonic
expression of OTX2 and DMBX1 in human brain. In European Society of Human Genetics
conference, Nice, France.
Rastogi, S. and Liberles, D. a (2005). Subfunctionalization of duplicated genes as a transition
state to neofunctionalization. BMC Evol Biol. 5, 28.
Raviola, E. and Gilula, N. B. (1975). Intramembrane organization of specialized contacts in the
outer plexiform layer of the retina. A freeze-fracture study in monkeys and rabbits. J Cell
Biol. 65, 192–222.
Raymond, P. A. and Easter, S. S. (1983). Postembryonic growth of the optic tectum in
goldfish. I. Location of germinal cells and numbers of neurons produced. J Neurosci. 3,
1077–91.
Raymond, P. A., Barthel, L. K. and Curran, G. A. (1995). Developmental patterning of rod
and cone photoreceptors in embryonic zebrafish. J Comp Neurol. 359, 537–50.
Raymond, P. a, Barthel, L. K., Bernardos, R. L. and Perkowski, J. J. (2006). Molecular
characterization of retinal stem cells and their niches in adult zebrafish. BMC Dev Biol. 6,
36.
Reh, T. A. and Levine, E. M. (1998). Multipotential stem cells and progenitors in the
vertebrate retina. J Neurobiol. 36, 206–20.
Reifers, F., Böhli, H., Walsh, E. C., Crossley, P. H., Stainier, D. Y. and Brand, M. (1998).
Fgf8 is mutated in zebrafish acerebellar (ace) mutants and is required for maintenance of
midbrain-hindbrain boundary development and somitogenesis. Development. 125, 2381–
95.
Reinisalo, M., Putula, J., Mannermaa, E., Urtti, A. and Honkakoski, P. (2012). Regulation
of the human tyrosinase gene in retinal pigment epithelium cells: the significance of
transcription factor orthodenticle homeobox 2 and its polymorphic binding site. Mol Vis.
18, 38–54.
227
Reis, L. M., Khan, A., Kariminejad, A., Ebadi, F., Tyler, R. C. and Semina, E. V (2011).
VSX2 mutations in autosomal recessive microphthalmia. Mol Vis. 17, 2527–32.
Rhinn, M. and Brand, M. (2001). The midbrain--hindbrain boundary organizer. Curr Opin
Neurobiol. 11, 34–42.
Rhinn, M., Lun, K., Amores, A., Yan, Y.-L., Postlethwait, J. H. and Brand, M. (2003).
Cloning, expression and relationship of zebrafish gbx1 and gbx2 genes to Fgf signaling.
Mech Dev. 120, 919–936.
Rhinn, M., Lun, K., Ahrendt, R., Geffarth, M. and Brand, M. (2009). Zebrafish gbx1 refines
the midbrain-hindbrain boundary border and mediates the Wnt8 posteriorization signal.
Neural Dev. 4, 12.
Rissone, A., Sangiorgio, L., Monopoli, M., Beltrame, M., Zucchi, I., Bussolino, F., Arese,
M. and Cotelli, F. (2010). Characterization of the neuroligin gene family expression and
evolution in zebrafish. Dev Dyn. 239, 688–702.
Robles, E., Smith, S. J. and Baier, H. (2011). Characterization of genetically targeted neuron
types in the zebrafish optic tectum. Front Neural Circuits. 5, 1.
Robu, M. E., Larson, J. D., Nasevicius, A., Beiraghi, S., Brenner, C., Farber, S. a and
Ekker, S. C. (2007). P53 Activation By Knockdown Technologies. PLoS Genet. 3, e78.
Roeser, T. and Baier, H. (2003). Visuomotor behaviors in larval zebrafish after GFP-guided
laser ablation of the optic tectum. J Neurosci. 23, 3726–34.
Roman-Roman, S., Shi, D.-L., Stiot, V., Haÿ, E., Vayssière, B., Garcia, T., Baron, R. and
Rawadi, G. (2004). Murine Frizzled-1 behaves as an antagonist of the canonical Wnt/beta-
catenin signaling. J Biol Chem. 279, 5725–33.
Ross, L. S., Parrett, T. and Easter, S. S. (1992). Axonogenesis and morphogenesis in the
embryonic zebrafish brain. J Neurosci. 12, 467–82.
228
Roth, C., Rastogi, S., Arvestad, L., Dittmar, K., Light, S., Ekman, D. and Liberles, D. A.
(2007). Evolution After Gene Duplication : Models , Mechanisms , Sequences , Systems ,
and Organisms. J Exp Zool B Mol Dev Evol. 308, 58–73.
Ruzickova, J., Piatigorsky, J. and Kozmik, Z. (2009). Eye-specific expression of an ancestral
jellyfish PaxB gene interferes with Pax6 function despite its conserved Pax6/Pax2
characteristics. Int J Dev Biol. 53, 469–82.
Ryan, J. F., Burton, P. M., Mazza, M. E., Kwong, G. K., Mullikin, J. C. and Finnerty, J. R.
(2006). The cnidarian-bilaterian ancestor possessed at least 56 homeoboxes: evidence from
the starlet sea anemone, Nematostella vectensis. Genome Biol. 7, R64.
Saint-Jeannet, J. P., He, X., Varmus, H. E. and Dawid, I. B. (1997). Regulation of dorsal fate
in the neuraxis by Wnt-1 and Wnt-3a. Proc Natl Acad Sci U S A. 94, 13713–8.
Sanes, J. R. and Zipursky, S. L. (2010). Design principles of insect and vertebrate visual
systems. Neuron. 66, 15–36.
Sankrithi, N. S. and O’Malley, D. M. (2010). Activation of a multisensory, multifunctional
nucleus in the zebrafish midbrain during diverse locomotor behaviors. Neuroscience. 166,
970–93.
Sanyanusin, P., Schimmenti, L. A., McNoe, L. A., Ward, T. A., Pierpont, M. E., Sullivan,
M. J., Dobyns, W. B. and Eccles, M. R. (1995). Mutation of the PAX2 gene in a family
with optic nerve colobomas, renal anomalies and vesicoureteral reflux. Nat Genet. 9, 358–
64.
Sarno, J. L., Kliman, H. J. and Taylor, H. S. (2005). HOXA10, Pbx2, and Meis1 protein
expression in the human endometrium: formation of multimeric complexes on HOXA10
target genes. J Clin Endocrinol Metab. 90, 522–8.
Sato, T., Takahoko, M. and Okamoto, H. (2006). HuC:Kaede, a useful tool to label neural
morphologies in networks in vivo. Genesis. 44, 136–42.
229
Sato, T., Hamaoka, T., Aizawa, H., Hosoya, T. and Okamoto, H. (2007). Genetic single-cell
mosaic analysis implicates ephrinB2 reverse signaling in projections from the posterior
tectum to the hindbrain in zebrafish. J Neurosci. 27, 5271–9.
Satow, T., Bae, S. K., Inoue, T., Inoue, C., Miyoshi, G., Tomita, K., Bessho, Y., Hashimoto,
N. and Kageyama, R. (2001). The basic helix-loop-helix gene hesr2 promotes gliogenesis
in mouse retina. J Neurosci. 21, 1265–73.
Satterlie, R. A. (2011). Do jellyfish have central nervous systems? J Exp Biol. 214, 1215–23.
Schmitt, E. A. and Dowling, J. E. (1996). Comparison of topographical patterns of ganglion
and photoreceptor cell differentiation in the retina of the zebrafish, Danio rerio. J Comp
Neurol. 371, 222–34.
Schmitt, E. A. and Dowling, J. E. (1999). Early retinal development in the zebrafish, Danio
rerio: light and electron microscopic analyses. J Comp Neurol. 404, 515–36.
Schneitz, K., Spielmann, P. and Noll, M. (1993). Molecular genetics of aristaless, a prd-type
homeo box gene involved in the morphogenesis of proximal and distal pattern elements in
a subset of appendages in Drosophila. Genes Dev. 7, 114–29.
Scholpp, S. and Brand, M. (2001). Morpholino-induced knockdown of zebrafish engrailed
genes eng2 and eng3 reveals redundant and unique functions in midbrain--hindbrain
boundary development. Genesis. 30, 129–33.
Scholpp, S. and Brand, M. (2003). Integrity of the midbrain region is required to maintain the
diencephalic-mesencephalic boundary in zebrafish no isthmus/pax2.1 mutants. Dev Dyn.
228, 313–22.
Scholpp, S., Lohs, C. and Brand, M. (2003). Engrailed and Fgf8 act synergistically to
maintain the boundary between diencephalon and mesencephalon. Development. 130,
4881–93.
Scholpp, S., Foucher, I., Staudt, N., Peukert, D., Lumsden, A. and Houart, C. (2007).
Otx1l, Otx2 and Irx1b establish and position the ZLI in the diencephalon. Development.
134, 3167–76.
230
Schonemann, M. D., Ryan, A. K., Erkman, L., McEvilly, R. J., Bermingham, J. and
Rosenfeld, M. G. (1998). POU domain factors in neural development. Adv Exp Med Biol.
449, 39–53.
Schwarz, M., Cecconi, F., Bernier, G., Andrejewski, N., Kammandel, B., Wagner, M. and
Gruss, P. (2000). Spatial specification of mammalian eye territories by reciprocal
transcriptional repression of Pax2 and Pax6. Development. 127, 4325–34.
Serra-Juhé, C., Rodríguez-Santiago, B., Cuscó, I., Vendrell, T., Camats, N., Torán, N. and
Pérez-Jurado, L. a (2012). Contribution of rare copy number variants to isolated human
malformations. PLoS One. 7, e45530.
Shen, Y. and Raymond, P. a (2004). Zebrafish cone-rod (crx) homeobox gene promotes
retinogenesis. Dev Biol. 269, 237–51.
Shen, K., Valero, J., Day, G. S. and Paré, M. (2011). Investigating the role of the superior
colliculus in active vision with the visual search paradigm. Eur J Neurosci. 33, 2003–16.
Shibahara, S., Yasumoto, K., Amae, S., Udono, T., Watanabe, K., Saito, H. and Takeda, K.
(2000). Regulation of pigment cell-specific gene expression by MITF. Pigment Cell Res.
13 Suppl 8, 98–102.
Shimizu, T., Patton, T. B. and Husband, S. A. (2010). Avian visual behavior and the
organization of the telencephalon. Brain Behav Evol. 75, 204–17.
Shimizu, N., Kawakami, K. and Ishitani, T. (2012). Visualization and exploration of Tcf/Lef
function using a highly responsive Wnt/β-catenin signaling-reporter transgenic zebrafish.
Dev Biol. 370, 71–85.
Shivakumar, L., Minna, J., Sakamaki, T., Pestell, R. and White, M. A. (2002). The
RASSF1A Tumor Suppressor Blocks Cell Cycle Progression and Inhibits Cyclin D1
Accumulation. Mol Cell Biol. 22, 4309–4318.
Shkumatava, A. and Neumann, C. J. (2005). Shh directs cell-cycle exit by activating p57Kip2
in the zebrafish retina. EMBO reports. 6, 563–9.
231
Shtutman, M., Zhurinsky, J., Simcha, I., Albanese, C., D’Amico, M., Pestell, R. and Ben-
Ze’ev, a (1999). The cyclin D1 gene is a target of the beta-catenin/LEF-1 pathway. Proc
Natl Acad Sci U S A. 96, 5522–7.
Sisson, B. E. and Topczewski, J. (2009). Expression of five frizzleds during zebrafish
craniofacial development. Gene Expr Patterns. 9, 520–7.
Smidt, M. P. and Burbach, J. P. H. (2007). How to make a mesodiencephalic dopaminergic
neuron. Nat Rev Neurosci. 8, 21–32.
Sperber, S. M., Saxena, V., Hatch, G. and Ekker, M. (2008). Zebrafish dlx2a contributes to
hindbrain neural crest survival, is necessary for differentiation of sensory ganglia and
functions with dlx1a in maturation of the arch cartilage elements. Dev Biol. 314, 59–70.
Stacey, D. W. (2003). Cyclin D1 serves as a cell cycle regulatory switch in actively
proliferating cells. Current opinion in cell biology 15, 158–63.
Stenkamp, D. L. (2007). Neurogenesis in the fish retina. Int Rev Cytol. 259, 173–224.
Stenkamp, D. L. (2011). The rod photoreceptor lineage of teleost fish. Prog Retin Eye Res. 30,
395–404.
Stephens, W. Z., Senecal, M., Nguyen, M. and Piotrowski, T. (2010). Loss of adenomatous
polyposis coli (apc) results in an expanded ciliary marginal zone in the zebrafish eye. Dev
Dyn. 239, 2066–77.
Stigloher, C., Ninkovic, J., Laplante, M., Geling, A., Tannhäuser, B., Topp, S., Kikuta, H.,
Becker, T. S., Houart, C. and Bally-Cuif, L. (2006). Segregation of telencephalic and
eye-field identities inside the zebrafish forebrain territory is controlled by Rx3.
Development. 133, 2925–35.
Stolfi, A. and Levine, M. (2011). Neuronal subtype specification in the spinal cord of a
protovertebrate. Development. 138, 995–1004.
Strähle, U. and Blader, P. (1994). Early neurogenesis in the zebrafish embryo. FASEB J. 8,
692–8.
232
Strähle, U., Blader, P., Henrique, D. and Ingham, P. W. (1993). Axial, a zebrafish gene
expressed along the developing body axis, shows altered expression in cyclops mutant
embryos. Genes Dev. 7, 1436–46.
Strausfeld, N. J. (2009). Brain organization and the origin of insects: an assessment. Proc Biol
Sci. 276, 1929–37.
Strauss, O. (2005). The retinal pigment epithelium in visual function. Physiol Rev. 85, 845–81.
Stryer, L. (1986). Cyclic GMP cascade of vision. Annu Rev Neurosci. 9, 87–119.
Stuebner, S., Faus-Kessler, T., Fischer, T., Wurst, W. and Prakash, N. (2010). Fzd3 and
Fzd6 deficiency results in a severe midbrain morphogenesis defect. Dev Dyn. 239, 246–60.
Stuermer, C. A. (1988). Retinotopic Organization the Zebrafish Embryo of the Developing
Retinotectal Projection in. J Neurosci. 8, 4513–4530.
Sulzer, D. (2007). Multiple hit hypotheses for dopamine neuron loss in Parkinson’s disease.
Trends Neurosci. 30, 244–50.
Takada, S., Stark, K. L., Shea, M. J., Vassileva, G., McMahon, J. A. and McMahon, A. P.
(1994). Wnt-3a regulates somite and tailbud formation in the mouse embryo. Genes Dev. 8,
174–89.
Takahashi, T. (2005). The evolutionary origins of vertebrate midbrain and MHB: insights from
mouse, amphioxus and ascidian Dmbx homeobox genes. Brain Res Bull. 66, 510–7.
Takahashi, T. and Holland, P. W. H. (2004). Amphioxus and ascidian Dmbx homeobox genes
give clues to the vertebrate origins of midbrain development. Development. 131, 3285–94.
Takahashi, T., Holland, P. W. H., Cohn, M. J., Shimizu, K. and Kurokawa, M. (2002). An
orphan PRD class homeobox gene expressed in mouse brain and limb development. Dev
Genes Evol. 212, 293–297.
Tello, J. A., Wu, S., Rivier, J. E. and Sherwood, N. M. (2008). Four functional GnRH
receptors in zebrafish: analysis of structure, signaling, synteny and phylogeny. Integr Comp
Biol. 48, 570–87.
233
Terakita, A. (2005). The opsins. Genome Biol. 6, 213.
Terrell, D., Xie, B., Workman, M., Mahato, S., Zelhof, A., Gebelein, B. and Cook, T.
(2012). OTX2 and CRX rescue overlapping and photoreceptor-specific functions in the
Drosophila eye. Dev Dyn. 241, 215–28.
Tesmer, J. J. G. (2008). Guanylyl cyclase sees the light. J Biol. 7, 31.
Thanos, S., Püttmann, S., Naskar, R., Rose, K., Langkamp-Flock, M. and Paulus, W.
(2004). Potential role of Pax-2 in retinal axon navigation through the chick optic nerve
stalk and optic chiasm. J Neurobiol. 59, 8–23.
Thisse, B., Pflumio, S., Fürthauer, M., Loppin, B., Heyer, V., Degrave, A., Woehl, R., Lux,
A., Steffan, T., Charbonnier, X.Q. and Thisse, C. (2001). Expression of the zebrafish
genome during embryogenesis. ZFIN Direct Data Submission (http://zfin.org).
Thomas, K. R. and Capecchi, M. R. (1990). Targeted disruption of the murine int-1 proto-
oncogene resulting in severe abnormalities in midbrain and cerebellar development.
Nature. 346, 847–50.
Thomas, M., Beazley, L. and Ziman, M. (2006). A multiphasic role for Pax7 in tectal
development. Exp Brain Res. 169, 266–71.
Thompson, J. a, Lovicu, F. J. and Ziman, M. (2007). Pax7 and superior collicular polarity:
insights from Pax6 (Sey) mutant mice. Exp Brain Res. 178, 316–25.
Thompson, J. A., Zembrzycki, A., Mansouri, A. and Ziman, M. (2008). Pax7 is requisite for
maintenance of a subpopulation of superior collicular neurons and shows a diverging
expression pattern to Pax3 during superior collicular development. BMC Dev Biol. 8, 62.
Thoreson, W. B. and Mangel, S. C. (2012). Lateral interactions in the outer retina. Prog Retin
Eye Res. 31, 407–41.
Thummel, R., Enright, J. M., Kassen, S. C., Montgomery, J. E., Bailey, T. J. and Hyde, D.
R. (2010). Pax6a and Pax6b are required at different points in neuronal progenitor cell
proliferation during zebrafish photoreceptor regeneration. Exp Eye Res. 90, 572–82.
234
Tirosh, I. and Barkai, N. (2007). Comparative analysis indicates regulatory
neofunctionalization of yeast duplicates. Genome Biol. 8, R50.
Toyama, R. and Dawid, I. B. (1997). lim6, a novel LIM homeobox gene in the zebrafish:
comparison of its expression pattern with lim1. Dev Dyn. 209, 406–17.
Tsujikawa, M. and Malicki, J. (2004). Genetics of photoreceptor development and function in
zebrafish. Int J Dev Biol. 48, 925–34.
Tsukiji, N., Nishihara, D., Yajima, I., Takeda, K., Shibahara, S. and Yamamoto, H. (2009).
Mitf functions as an in ovo regulator for cell differentiation and proliferation during
development of the chick RPE. Dev Biol. 326, 335–46.
Tvrdik, P. and Capecchi, M. R. (2006). Reversal of Hox1 gene subfunctionalization in the
mouse. Dev Cell. 11, 239–50.
Udin, S. B. (2007). The instructive role of binocular vision in the Xenopus tectum. Biol Cybern.
97, 493–503.
Ullrich-Lüter, E. M., Dupont, S., Arboleda, E., Hausen, H. and Arnone, M. I. (2011).
Unique system of photoreceptors in sea urchin tube feet. Proc Natl Acad Sci U S A. 108,
8367–72.
Uribe, R. a and Gross, J. M. (2010). Id2a influences neuron and glia formation in the zebrafish
retina by modulating retinoblast cell cycle kinetics. Development. 137, 3763–74.
Uribe, R. a, Kwon, T., Marcotte, E. M. and Gross, J. M. (2012). Id2a functions to limit
Notch pathway activity and thereby influence the transition from proliferation to
differentiation of retinoblasts during zebrafish retinogenesis. Dev Biol. 371, 280–92.
Van de Water, S., van de Wetering, M., Joore, J., Esseling, J., Bink, R., Clevers, H. and
Zivkovic, D. (2001). Ectopic Wnt signal determines the eyeless phenotype of zebrafish
masterblind mutant. Development. 128, 3877–88.
Van Raay, T. J. and Vetter, M. L. (2004). Wnt/frizzled signaling during vertebrate retinal
development. Dev Neurosci. 26, 352–8.
235
Vitorino, M., Jusuf, P. R., Maurus, D., Kimura, Y., Higashijima, S.-I. and Harris, W. a
(2009). Vsx2 in the zebrafish retina: restricted lineages through derepression. Neural Dev.
4, 14.
Vopalensky, P. and Kozmik, Z. (2009). Eye evolution: common use and independent
recruitment of genetic components. Philos Trans R Soc Lond B Biol Sci. 364, 2819–32.
Wallis, D. E., Roessler, E., Hehr, U., Nanni, L., Wiltshire, T., Richieri-Costa, A., Gillessen-
Kaesbach, G., Zackai, E. H., Rommens, J. and Muenke, M. (1999). Mutations in the
homeodomain of the human SIX3 gene cause holoprosencephaly. Nat Genet. 22, 196–8.
Walshe, J. and Mason, I. (2003). Unique and combinatorial functions of Fgf3 and Fgf8 during
zebrafish forebrain development. Development. 130, 4337–49.
Ward, A., Liu, J., Feng, Z. and Xu, X. Z. S. (2008). Light-sensitive neurons and channels
mediate phototaxis in C. elegans. Nature neuroscience 11, 916–22.
Warrant, E. (2000). The eyes of deep-sea fishes and the changing nature of visual scenes with
depth. Philos Trans R Soc Lond B Biol Sci. 355, 1155–9.
Watanabe, Y. and Nakamura, H. (2000). Control of chick tectum territory along dorsoventral
axis by Sonic hedgehog. Development. 127, 1131–40.
Westerfield, M. (2000). The zebrafish book. A guide for the laboratory use of zebrafish (Danio
rerio). 4th ed. Eugene: University of Oregon Press.
Wistow, G. (2012). The human crystallin gene families. Hum Genomics. 6, 26.
Witkovsky, P. (2000). Photoreceptor classes and transmission at the photoreceptor synapse in
the retina of the clawed frog, Xenopus laevis. Microsc Res Tech. 50, 338–46.
Wolda, S. L., Moody, C. J. and Moon, R. T. (1993). Overlapping expression of Xwnt-3A and
Xwnt-1 in neural tissue of Xenopus laevis embryos. Dev Biol. 155, 46–57.
Woo, K. and Fraser, S. E. (1998). Specification of the hindbrain fate in the zebrafish. Dev Biol.
197, 283–96.
236
Woods, I. G. I., Wilson, C., Friedlander, B., Chang, P., Reyes, D. K., Nix, R., Kelly, P. D.,
Chu, F., Postlethwait, J. H. and Talbot, W. S. (2005). The zebrafish gene map defines
ancestral vertebrate chromosomes. Genome Res. 15, 1307–1314.
Wullimann, M. F. and Knipp, S. (2000). Proliferation pattern changes in the zebrafish brain
from embryonic through early postembryonic stages. Anat Embryol. 202, 385–400.
Xiao, T. and Baier, H. (2007). Lamina-specific axonal projections in the zebrafish tectum
require the type IV collagen Dragnet. Nature neuroscience 10, 1529–37.
Xu, S., Sunderland, M. E., Coles, B. L. K., Kam, A., Holowacz, T., Ashery-Padan, R.,
Marquardt, T., McInnes, R. R. and van der Kooy, D. (2007). The proliferation and
expansion of retinal stem cells require functional Pax6. Dev Biol. 304, 713–21.
Yamagata, M. and Sanes, J. R. (2008). Dscam and Sidekick proteins direct lamina-specific
synaptic connections in vertebrate retina. Nature. 451, 465–9.
Yamaguchi, M., Tonou-Fujimori, N., Komori, A., Maeda, R., Nojima, Y., Li, H., Okamoto,
H. and Masai, I. (2005). Histone deacetylase 1 regulates retinal neurogenesis in zebrafish
by suppressing Wnt and Notch signaling pathways. Development. 132, 3027–43.
Yan, R. T. and Wang, S. Z. (1998). neuroD induces photoreceptor cell overproduction in vivo
and de novo generation in vitro. J Neurobiol. 36, 485–96.
Yang, Z. (1994). Maximum likelihood phylogenetic estimation from DNA sequences with
variable rates over sites: approximate methods. J Mol Evol. 39, 306–14.
Yang, Z. (1998). Likelihood ratio tests for detecting positive selection and application to
primate lysozyme evolution. Mol Biol Evol. 15, 568–73.
Yang, Z. (2007). PAML 4: phylogenetic analysis by maximum likelihood. Mol Biol Evol. 24,
1586–91.
Yang, Z. and Bielawski, J. (2000). Statistical methods for detecting molecular adaptation.
Trends Ecol Evol. 15, 496–503.
237
Yang, Z. and Nielsen, R. (1998). Synonymous and nonsynonymous rate variation in nuclear
genes of mammals. J Mol Evol. 46, 409–18.
Yang, D., Liu, Q., Yang, M., Wu, H., Wang, Q., Xiao, J. and Zhang, Y. (2012). RNA-seq
liver transcriptome analysis reveals an activated MHC-I pathway and an inhibited MHC-II
pathway at the early stage of vaccine immunization in zebrafish. BMC Genomics. 13, 319.
Yang-Snyder, J., Miller, J. R., Brown, J. D., Lai, C. J. and Moon, R. T. (1996). A frizzled
homolog functions in a vertebrate Wnt signaling pathway. Curr Biol. 6, 1302–6.
Yao, K. and Ge, W. (2010). Kit system in the zebrafish ovary: evidence for functional
divergence of two isoforms of kit (kita and kitb) and kit ligand (kitlga and kitlgb) during
folliculogenesis. Biol Reprod. 82, 1216–26.
Yao, J. and Kessler, D. S. (2001). Goosecoid promotes head organizer activity by direct
repression of Xwnt8 in Spemann’s organizer. Development. 128, 2975–87.
Yazulla, S. and Studholme, K. M. (2001). Neurochemical anatomy of the zebrafish retina as
determined by immunocytochemistry. J Neurocytol. 30, 551–92.
Zaghloul, N. A. and Moody, S. A. (2007a). Changes in Rx1 and Pax6 activity at eye field
stages differentially alter the production of amacrine neurotransmitter subtypes in Xenopus.
Mol Vis. 13, 86–95.
Zaghloul, N. A. and Moody, S. A. (2007b). Alterations of rx1 and pax6 expression levels at
neural plate stages differentially affect the production of retinal cell types and maintenance
of retinal stem cell qualities. Dev Biol. 306, 222–40.
Zelhof, A. C. (2003). Mutation of the photoreceptor specific homeodomain gene Pph13 results
in defects in phototransduction and rhabdomere morphogenesis. Development. 130, 4383–
4392.
Zhang, Y., Miki, T., Iwanaga, T., Koseki, Y., Okuno, M., Sunaga, Y., Ozaki, N., Yano, H.,
Koseki, H. and Seino, S. (2002). Identification, tissue expression, and functional
characterization of Otx3, a novel member of the Otx family. J Biol Chem. 277, 28065–9.
238
Zygar, C. A., Cook, T. L. and Grainger, R. M. (1998). Gene activation during early stages of
lens induction in Xenopus. Development. 125, 3509–19.
239
Appendices
240
Appendix 1: Transcriptional auto-regulation of dmbx1 paralogs
Previous analyses of the mouse knock-in Dmbx1-lacZ transgenic showed that expression of the
lacZ reporter was altered in mutants with one null allele and one knock-in lacZ allele (Ohtoshi
and Behringer, 2004). For example, up-regulated transcription from the endogenous Dmbx1
locus occurred within the normal Dmbx1 expression domains in the embryo (e.g. midbrain),
down regulation of transgene expression occurred in the postnatal medulla oblongata, and
ectopic transgene expression was observed in the postnatal inferior colliculus, where Dmbx1 is
not normally expressed (Ohtoshi and Behringer, 2004). These observations indicate that in the
absence of functional Dmbx1 protein, the transcription of the Dmbx1 gene is deregulated
leading to enhanced, reduced or ectopic gene expression in a region-specific manner.
To test whether the zebrafish dmbx1 genes depend on Dmbx1 function, I analyzed dmbx1a and
dmbx1b gene expression by in situ hybridization in morpholino knockdown embryos. The
analysis assumed that the extent of gene expression was correlated with transcriptional activity,
although the possibility that mRNA stability may change through post-transcriptional
modification and affect transcription cannot be ruled out. At 24hpf, expression of dmbx1a or
dmbx1b appeared to remain the same in the respective single morphants (Figure S1). However,
at 72hpf, the intensity of dmbx1a expression within the optic tectum and the hindbrain was
enhanced in MO1a injected embryos (Figure S2B, S2B’) compared to mMO1a injected embryos
(Figure S2C, S2C’). Strikingly, the normally robust retinal expression of dmbx1a at this stage
was almost completely eliminated in MO1a injected embryos (Figure S2B, B’). Because
variation in the intensity of the hybridization signal may be unrelated to the specific changes in
RNA levels, further investigation on dmbx1a expression using semi-quantitative RT-PCR from
dissected 72 hpf brain tissue (not including retinal tissue) as well as isolated retinal tissue
241
Figure S1 Dmbx1a and dmbx1b expression patterns in dmbx1 knocked down
embryos.
Lateral views, anterior to the left (A-L). At 24hpf, expression of dmbx1a in dmbx1a morphants
has extended ventrally (B), whereas dmbx1b expanded ventrally and posteriorly (K). Other
morphant groups (E, H) and control-injected embryos (C, F, I, L) have similar dmbx1a and
dmbx1b expressions as un-injected controls (A, D, G, J).
242
provided more quantitative analysis (Figure S2J, S2K). The changes in dmbx1a expression in
the brain and retina matched the in situ hybridization data. On the other hand, the pattern of
dmbx1a expression in MO1b injected embryos was less obvious. By in situ hybridization, the
overall pattern of dmbx1a expression in the brain was similar in MO1b (Figure S2E, S2E’) and
mMO1b injected (Figure S2F, S2F’) embryos; however the RT-PCR analyses revealed that the
overall level of dmbx1a expression was reduced (Figure S2J, S2K).
The expression of dmbx1a in the combined MO1a+MO1b injected embryos at 72 hpf resembled
the phenotype of single MO1a injected embryos. Consistent with the single MO1a morphant
data, dmbx1a expression in the retina was significantly reduced in the double morphant embryos
(Figure S2H, S2H’) compared to control (Figure S2G, S2G’, S2I, S2I’), suggesting that
continual dmbx1a expression in the retina is dependent on Dmbx1a function, and verified by
semi-quantitative RT-PCR (Figure S2J, S2K). These data also demonstrate that the concomitant
loss of Dmbx1a and Dmbx1b function may partially suppress the up-regulation of dmbx1a gene
expression in the brain relative to the MO1a phenotype, indicating that Dmbx1b may have an
opposite role in regulating dmbx1a expression.
The same experiment was performed to analyze dmbx1b expression in the brain and retina of 72
hpf morphant embryos. Unlike the results obtained with dmbx1a expression, there was little
evidence for a change in dmbx1b expression in the hindbrain in any injected embryos (Figure
S3). Strong upregulation of dmbx1b could be seen at the periphery of the optic tecta in MO1b
injected and double morphant embryos (Figure S3E, S3E’, S3H, S3H’), but the effect was minor
in the single MO1a injected embryos (Figure S3B, S3B’). However, this change in brain
expression in MO-injected embryos was unable to detect by semi-quantitative RT-PCR (Figure
S3J). Thus, dmbx1a and dmbx1b gene expression in the brain was likely to be differentially
243
Figure S2 Dmbx1a expression in single and double dmbx1 morphant embryos
at 72hpf.
Lateral views (A-I) and dorsal views (A’-I’) of wholemount in situ hybridization embryos with
dmbx1a probe. Number of embryos examined is shown on the bottom right. Un-injected
embryos (A, D, G) and control-injected (mMO) embryos (C, F, I) have similar dmbx1a
expression, and so are dmbx1b (MO1b) embryos (E). Expression of dmbx1a strengthens at the
medial and caudal tectal regions in dmbx1a (MO1a) (B and B’) or both dmbx1a and dmbx1b
(MO1a+MO1b) knocked down embryos. Dmbx1a expression is abolished in double dmbx1
morphants (H and H’) at the tecta but gained significant level in the hindbrain. In both dmbx1a
and double morphants, expression of dmbx1a is not detected in the retina. Semi-quantitative RT-
PCR was performed to evaluate the changes in dmbx1a expression in brain (J) and retina (K)
separately, while β-actin serves as controls. PCR results correlates with those in situ
hybridization patterns observed.
244
regulated by Dmbx1a and Dmbx1b proteins. Because of the low levels of dmbx1b expression in
the retina at this stage, no dmbx1b expression was detected in the retina of embryos from any of
the treatment groups RT-PCR (Figure S3K). As mention before, the possibility that morpholinos
might stabilize transcripts could not be ruled out, and this might cause an increase in staining
independent of transcriptional regulation. However, the enhanced transcriptional activity in the
brain but reduced transcriptional level in the retina of dmbx1a expression within the same
embryos indicated that these effects were unlikely to be caused by non-specific transcript
stabilization. Together, these data suggested that Dmbx1 proteins might repress their respective
transcripts in the optic tectum and the hindbrain, although the molecular mechanism of how
these proteins carried out their regulatory functions remained unclear. Biochemistry analysis
revealed that the homeodomain of mouse Dmbx1 functioned as a transcriptional repressor
(Kimura et al., 2005), which complemented with the results observed from the auto-regulation
study here.
245
Figure S3 Dmbx1b expression in single and double dmbx1 morphant embryos
at 72hpf.
Lateral views (A-I) and dorsal views (A’-I’) of wholemount in situ hybridization embryos with
dmbx1b probe. Number of embryos examined is shown on the bottom right. Un-injected
embryos (A, D, G) and control-injected (mMO) embryos (C, F, I) have similar dmbx1b
expression. Expression of dmbx1b in MO1a embryos remains the same (B), but it intensifies at
the ventricle region outlining the tectal lobes in both dmbx1b (MO1b) single morphants and
double morphants (MO1a+MO1b) (E, E’ and H, H’ respectively). Level of dmbx1b in hindbrain
remains the same in all three groups of knocked down embryos. Semi-quantitative RT-PCR was
performed to evaluate the changes in dmbx1b expression in brain (J) and retina (K) separately,
while β-actin serves as controls. Level of dmbx1b was too low to be detected in the retina.
Changes in the amount of dmbx1b transcript were subtle among all three groups.
246
Appendix 2: The search for Dmbx1a protein-protein interactions
Many transcription factors are known to form homo- or heterodimers, or multimeric complexes
to regulate transcriptions (Funnell and Crossley, 2012; Ogata et al., 2003). There are many
examples in which a transcription factor works with various protein complexes and bind to
different DNA sequences to carry out specific functions. For examples, Pit-1 is a POU domain
transcription factor that homodimerizes when bound to DNA (Jacobson et al., 1997;
Schonemann et al., 1998). On the other hand, Pbx can heterodimerize with Meis or Hox
homeodomain proteins, or together they can form trimeric complexes (Knoepfler et al., 1999;
Sarno et al., 2005). To investigate proteins that may interact with Dmbx1a, a yeast-two-hybrid
screen with the Hybrigenics Services was performed. Using full length Dmbx1a protein as bait
to screen though a protein library built from 18-20hpf zebrafish embryos, 69 potential clones
with prey sequences obtained from different gene fragments were selected for further validation
(summarized in Table S1). Only 32 out of 69 clones showed potential interaction with Dmbx1a.
Of those 32 prey clones, only one interaction fell into the category with good confidence level,
while the rest of the 31 clones only showed moderate confidence in the interaction. The prey
fragment which interacted relatively strongly with Dmbx1a encoded a protein called Retinoic
acid receptor, gamma b (Rargb). In zebrafish, Rargb can regulate lateral plate mesoderm
migration and help position the liver properly (Garnaas et al., 2012). Based on the expression
patterns of rargb and dmbx1a, it was unlikely that these two proteins could interact with each
other since rargb was expressed in the caudal spinal cord and the branchial arches at 18-20hpf.
However, the possibility that these two proteins may be binding to one another cannot be ruled
out until a co-immunoprecipitation experiment is performed to validate this interaction. As for
the rest of the prey fragments that were pulled out from the screen, several of these proteins
were promising candidates that might bind with Dmbx1a based on their overlapping expression
247
patterns and their functions in the nervous system (see Chapter 6 for details). However, clones
that fell into this category with only moderate confidence level of the interaction are sometimes
difficult to validate because their interactions with Dmbx1a are usually rather weak, thus
resulted in those ambiguous outcomes from the yeast-two-hybrid screen. In some cases, they
may also be false-positive. Hence, more sensitive and more powerful assays to discover protein-
protein interactions, such as mass spectrometry, may be a better way to identify potential
Dmbx1a binding partners compared to the yeast-two-hybrid study.
248
Table S1: Summary of potential Dmbx1a interacting genes obtained from a
yeast two-hybrid screen
249
250
Appendix 3: Dmbx1a and Dmbx1b antibodies
To generate specific dmbx1a and dmbx1b antibodies, I chose to target the diverged region
between the two paralogs. The N-terminal and the homeodomain region were avoided due to the
high level of sequence similarity between the two dmbx1 paralogs. Using endogenous restriction
sites, I subcloned a short fragment of Dmbx1a into pGEX vector. The insert includes mostly the
variable regions, starts from the coiled coil domain and continues to the C-terminal (red dotted
line above the dmbx1a sequence in Figure S4A). Similar to dmbx1a, I also cloned a dmbx1b
sequence into pGEX (blue dotted line underneath the dmbx1b sequence in Figure S5A). pGEX
plasmid with or without inserts are induced with IPTG in bacteria and lysates were collected to
examine if recombinant proteins were the right size. Since I was unsure about the solubility of
these proteins, both soluble and insoluble lysate fractions separate were kept for the analyses.
The predicted size for GST alone was 26kDa, while the recombinant GST-Dmbx1a and GST-
Dmbx1b proteins weight 33kDa and 36kDa respectively. From the blot in Figure S4B, it
showed that GST was soluble and the size was correct (Figure S4B). Dmbx1a with GST tag was
slightly lighter (~31kDa) than expected (33kDa) and soluble (Figure S4B). However, Dmbx1b
protein was found to be much heavier (thickest band was at ~52kDa) than expected (36kDa) and
appeared to be quite non-specific since there were multiple bands detected in the lane (Figure
S5B). This result was rather surprising for GST-Dmbx1b proteins because the peptide chains at
least 10kDa heavier than expected; however, it was possible that Dmbx1b was modified by
glycosylation which could cause this discrepancy. The subsequent goal is to induce GST tagged
Dmbx1a and Dmbx1b proteins in bacteria, then cleave off the GST tag and inject the purified
Dmbx1a and Dmbx1b proteins into rabbit to generate antibodies. However, due to time
constraint, the productions of these two antibodies are not completed yet.
251
Figure S4 Generating Dmbx1a-specific peptide fragment with GST tag at the
N-term.
(A) Alignments between Dmbx1a and Dmbx1b protein sequences. Red dotted line above the
sequences indicates the region of Dmbx1a that was cloned in-frame and downstream of the
GST-tag. (B) pGEX and pGEX-Dmbx1a (pGEX-1a) in BL21 bacteria cells were induced by
IPTG, followed by an analysis with the GST antibody on the Western Blot to identify the
solubility of Dmbx1a peptide. Lysates from un-induced bacteria, soluble fraction of IPTG-
induced bacteria, and insoluble fraction of IPTG-induced bacteria were respectively loaded in
left, middle, and right lanes from each group. Protein ladder is shown on the left, and expected
protein size from each group is listed on top. Red asterisk indicates that the induced proteins are
soluble.
252
Figure S5 Generating Dmbx1b-specific peptide fragment with GST tag at the
N-term.
(A) Alignments between Dmbx1a and Dmbx1b protein sequences. Blue dotted line beneath the
sequences indicates the region of Dmbx1b that was cloned in-frame and downstream of the
GST-tag. (B) pGEX and pGEX-Dmbx1b (pGEX-1b) in BL21 bacteria cells were induced by
IPTG, followed by an analysis with the GST antibody on the Western Blot to identify the
solubility of Dmbx1b peptide. Lysates from un-induced bacteria, soluble fraction of IPTG-
induced bacteria, and insoluble fraction of IPTG-induced bacteria were respectively loaded in
left, middle, and right lanes from each group. Protein ladder is shown on the left, and expected
protein size from each group is listed on top. Red asterisk indicates that the induced proteins are
soluble.
253
Appendix 4: Other Publications
Published Articles:
1) Beharry AA, Wong L, Tropepe V, Woolley GA. (2011). Fluorescence imaging of
azobenzene photoswitching in vivo. Angew Chem Int Ed Engl. 50(6):1325-7
Using zebrafish to demonstrate that the photochemical switching ability of azobenzene-based
peptide is possible in living organisms, this fluorescent reporter allows precise
spatiotemporal controls on when switching occurs in an in vivo system.
2) Dang LT, Wong L, Tropepe V. (2012). Zfhx1b induces a definitive neural stem cell fate in
mouse embryonic stem cells. Stem Cells Dev. 21(15):2838-51.
To investigate the role of transcription factor Zfhx1b in regulating neural cell fate in vitro,
Zfhx1b overexpression and siRNA knocked down studies were performed in embryonic stem
cells. Results from the analyses demonstrated that Zfhx1b was necessary and sufficient for
cells to maintain as definitive neural stem cells, and that this gene was a downstream target of
the intercellular FGF signaling pathway.
Manuscript in Preparation:
1) Olsen JB, Wong L, Marcon E, Guo H, Ni Z, Zhong G, Guo X, Li Y, Phanse S, Lsserlin
R, Fong V, Smiley S, Pogoutse O, Moffat J, Zhang Z, Greenblatt JF, Tropepe V, Emili
A. (2013). G9a and ZNF644 physically associate to suppress progenitor cell identity and
cell cycle progression in the retina. (co-first author, manuscript in preparation for
Developmental Cell)
ZNF644 is a transcription factor that was identified as a novel interacting protein with G9a,
which forms a histone methylation complex to regulate transcriptions. Mouse and zebrafish
znf644 ortholog genes were both examined in order to study whether its functions was
conserved between mammals and non-mammals. Results obtained from the zebrafish
indicated that the znf644 paralogs were both interacting with G9a, although znf644a was
required for controlling cell survival in differentiated retinal neurons while znf644b was
necessary for promoting retinal differentiation during early stages.
254
Copyright Acknowledgements
1) Chang, L., Khoo, B., Wong, L. and Tropepe, V. (2006). Genomic sequence and
spatiotemporal expression comparison of zebrafish mbx1 and its paralog, mbx2.
Development Genes and Evolution 216, 647–654.
2) Wong, L., Weadick, C. J., Kuo, C., Chang, B. S. and Tropepe, V. (2010). Duplicate dmbx1
genes regulate progenitor cell cycle and differentiation during zebrafish midbrain and retinal
development. BMC Developmental Biology 10, 100.
Licensee BioMed Central Ltd. BMC Developmental Biology is an open access article that
allows unrestricted use.