Molecular Cues in Pathfinding of Axial Motoneurons in the Developing Zebrafish

Home , Guidepost cells, Pleiotrophin, Plexin, Plexin A1, SEMA5A, Sema domain

Molecular cues in pathfinding of axial motoneurons in the developing zebrafish

DISSERTATION

Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University

Jona Dela Cruz Hilario

Graduate Program in Molecular, Cellular and Developmental Biology

The Ohio State University

2010

Dissertation Committee:

Christine Beattie, Advisor

Susan Cole

James Jontes

Harald Vaessin

Jona Hilario

2010

Abstract

Establishing neuromuscular specificity is an important step during development to ensure proper motor function. Motor axons may travel over relatively long distances from their positions in the CNS to muscle targets in the periphery. The zebrafish model has been useful for investigating the process of motor axon pathfinding. The embryonic motor system of the zebrafish is relatively simple and has been well studied. Both forward and reverse genetics have been used to study axon pathfinding of zebrafish trunk motor axons. Our study of two motor axon mutants, stumpy and topped, has led to the identification of two molecules that guide trunk motor axons to their muscle targets.

stumpy mutants exhibit a phenotype wherein motor axons stall for prolonged periods at intermediate targets prior to them reaching their final target. Positional cloning mapped the mutation to the zebrafish homolog of the collagenXIXa1 (colXIX). colXIX is expressed at known intermediate targets during the time of axon outgrowth. Knocking down ColXIX using morpholinos (MO) phenocopies stumpy. Also stumpy mutants were rescued by knocking down ColXIX and adding back mouse colXIX RNA. This suggests that ColXIX functions to enable growth cones to navigate intermediate target during development.

Semaphorin 5A (Sema5A) was initially identified as a candidate gene for the topped

mutant but this does not appear to be a case. However, Sema5A appears to play a role in

motor axon pathfinding. Sema5A is expressed in the myotome during the time of axon

outgrowth. Knocking down Sema5A using MOs specifically affects the Caudal Primary

(CaP) motor axon, inducing a delay in its extension to its muscle target as well as axon

branching. This MO phenotype can be rescued by adding back rat sema5A RNA.

Sema5A has been previously shown to act as a bifunctional cue in the rat habenula. In zebrafish, we saw that adding back RNA encoding the sema domain alone rescued the branching phenotype in sema5A morphants. Conversely, adding back RNA encoding the thrombospondin repeat (TSR) domain alone into sema5A morphants exclusively rescued delay in ventral motor axon extension. These data show that Sema5A is a bifunctional axon guidance cue for vertebrate motor axons in vivo.

The addition of ColXIX and Sema5A to the list of molecules that are involved in this seemingly simple pathfinding process demonstrates that numerous factors and pathways may be involved in establishing precise neuromuscular connections. Both these molecules have been shown to function in other contexts in the nervous system thus, understanding the roles these molecules play in axonal pathfinding can reveal novel mechanisms involved in wiring the nervous system.

iii

To Paul, who not only helped me with stats and feeding, but also believed I could do this

when I couldn’t see that for myself. You have my love and gratitude.

To my Family, for their love and support that could be felt over oceans.

To our Li’l One, can’t wait to meet you!

Acknowledgments

I am very grateful to my adviser Dr. Christine Beattie for being a mentor in the truest sense of the word throughout my graduate career. Her enthusiasm for science helped me see past the routine disappointments bench science often throws your way and look for the bigger more important picture. I would also like to thank my committee Dr. Susan

Cole, Dr. Jamie Jontes and Dr. Harald Vaessin for your encouragement, input and help when my project ran up against a wall.

I would also like to thank Louise Rodino-Klapac and Chunping Wang who helped and trained me early on and whose projects I eventually took over. I would not have had anything to work on if not for both your hard work! I would also like to thank all Beattie lab members past and present for their friendship and camaraderie as well as for their generosity in sharing their knowledge, reagents and their lives with me. Thanks too to the rest of the zebrafish group at CMN (PDH and JDJ labs) for making this a fun environment to work in. Also thanks to the fishroom staff who allow all of us to do our work well.

I thank Dr. Catherina Becker, Dr. Sarah Childs and Dr. Michael Granato for graciously providing reagents. This work has been supported by the National Science Foundation.

Vita

March 1998 ...... Philippine Science High School-Diliman

April 2002 ...... B.S. Molecular Biology and Biotechnology,

University of the Philippines-Diliman

2004 to present ...... Graduate Research Associate, Department

of Neuroscience, The Ohio State University

Publications

Hilario, J.D., Rodino-Klapac, L.R., Wang, C., Beattie, C.E., 2009. Semaphorin 5A is a Bifunctional Axon Guidance Cue for Axial Motoneurons in Vivo. Dev. Biol. 1, 190-200.

Fields of Study

Major Field: Molecular, Cellular and Developmental Biology

Table of Contents

Abstract ...... ii

Acknowledgments...... v

Vita ...... vii

Publications ...... vii

Fields of Study ...... vii

Table of Contents ...... viii

List of Tables ...... xiv

List of Figures ...... xv

Abbreviated terms ...... xviii

Chapter 1. Introduction to Motor Axon guidance ...... 1

Overview ...... 1

Growth cones and axon pathfinding ...... 2

Axon guidance molecules ...... 5

Netrins and its receptors ...... 6

Eph receptors and Ephrins ...... 10 vii

Slits and Robo ...... 11

Semaphorins and its receptors ...... 12

The role of the extracellular matrix in axon pathfinding ...... 16

Proteoglycans...... 18

Fibrillar ECM Molecules ...... 20

Laminins ...... 20

Tenascins ...... 21

Collagens ...... 22

The zebrafish as a model for studying axon guidance ...... 25

Forward genetics reveals molecules responsible guiding primary motoneurons ...... 28

diwanka mutants ...... 29

unplugged mutants ...... 30

stumpy mutants ...... 32

topped mutants ...... 33

Reverse Genetics confirms roles of known molecular motor axon guidance cues in

zebrafish...... 33

Chapter 2: Collagen XIXa1 is critical for intermediate target navigation by primary motor axons ...... 43

Abstract ...... 43

viii

Introduction ...... 44

Materials and Methods ...... 49

Fish care and maintenance ...... 49

Genetic Mapping and Cloning ...... 49

RNA in situ hybridization...... 51

Morpholino Analysis ...... 51

cDNA Constructs and RNA synthesis………………………………………………52

Whole mount antibody labeling ...... 53

Results ...... 54

Stumpy maps to the zebrafish homolog of ColXIX ...... 54

colXIX is expressed at intermediate targets during pathfinding of primary motor

axons ...... 55

Morpholino knockdown of ColXIX phenocopies stumpy…………………………..57

Identifying mutations in colXIX in stumpy mutants ...... 58

Mutations found in styb393 does not rescue colXIX MO phenotype ...... 59

Mouse colXIX mRNA with styb393 mutations acts in a dominant negative manner to

induce stumpy phenotype ...... 60

Overexpression of mouse ColXIX induces CaP axon defects ...... 63

Mouse ColXIX full-length RNA rescues stumpy only with knockdown of zebrafish

ColXIX ...... 62 ix

Overexpression of mouse ColXIX modules separately induced CaP axon defects .. 63

Discussion ...... 66

The stumpy gene encodes the zebrafish CollagenXIX ...... 67

Collagen XIX during development ...... 69

Collagen XIX has multiple interaction domains ...... 71

ColXIX might function to anchor cues at intermediate targets ...... 74

Chapter 3: Semaphorin 5A is a Bifunctional Axon Guidance Cue for Axial Motoneurons in vivo...... 91

Abstract ...... 91

Introduction ...... 92

Materials and Methods ...... 97

Fish Strains and Maintenance ...... 97

Genomic Cloning and Sequencing ...... 97

Morpholino Analysis ...... 98

RNA Rescue and Sema5A overexpression ...... 99

Calculations and Statistical Analysis ...... 100

Whole Mount Antibody Labeling………………………………………………….100

Whole-mount In Situ Hybridization ...... 100

Results ...... 101

The Zebrafish sema5A gene ...... 101

Zebrafish sema5A is expressed in the ventral myotome during development ...... 102

Knockdown of zebrafish Sema5A causes delay in motor axon extension into the

ventral muscle and axonal branching defects...... 103

RNA rescue of morpholino phenotype ...... 105

Sema5A and known zebrafish Plexins ...... 106

The thrombospodin repeat domain of rat Sema5A rescues delay of CaP axon

extension in sema5A morphants ...... 109

The sema domain of rat Sema5A rescues axonal branching defects in sema5A

morphants ...... 110

Discussion ...... 111

Sema5A is a bifunctional axon guidance cue for CaP axons ...... 112

Plexin A3: A Sema5A receptor? ...... 116

Sema5A may be modulated by proteoglycans in the myotome or CaP axon growth

cones ...... 118

Chapter 4: Topped candidate gene search ...... 130

Introduction ...... 130

Methods ...... 133

Fish Strains and Maintenance ...... 133

Morpholino analysis ...... 133 xi

RNA rescue...... 134

BAC rescue ...... 134

Whole mount antibody labeling ...... 134

RNA in situ hybridization...... 135

Results ...... 135

The Sema5A gene ...... 135

Injection of sema5A RNA may rescue topped mutant phenotype ...... 136

Knockdown of Sema5A has a different phenotype than topped ...... 136

Knockdown of Sema5A in topped mutants results in an additive phenotype ...... 137

cntnap2 is expressed in developing somites ...... 138

BAC containing cntnap2 does not rescue topped mutant ...... 138

Discussion ...... 139

Is sema5A the topped gene?...... 139

Is cntnap2 topped? ...... 140

Conclusion ...... 141

Chapter 5: Conclusions ...... 151

stumpy mutant ...... 151

topped mutant ...... 152

Modularity of axon guidance cues ...... 153

xii

Combining and modulating axon guidance cues ...... 155

Appendix A: Supplementary Data ...... 157

References ...... 160

xiii

List of Tables

Table Page

Table 1.1 Genes implicated in axon guidance of motor neurons in the developing zebrafish………………………………………………………………….34

Table 2.1 Amino acid mutations found in stumpy393…………………………….....88

Table 2.2 Overexpression of wild-type and mutant colXIX mouse RNA………….89

Table 2.3 ColXIX domains affect CaP axons………………………………………90

Table 3.1 Summary of data from sema5A RNA rescue experiments……………..129

Table 4.1 Annotated predicted genes in topped critical region……………………147

Table 4.2 Comparison of phenotypes of topped458 homozygous mutant embryos and sema5A MO knockdown embryos……………………………………...150

xiv

List of Figures

Figure Page

Figure 1.1 Classification of axon guidance molecules ...……………………………37

Figure 1.2 Illustration of two axon pathways and their intermediate targets ...……..39

Figure 1.3 Semaphorin family of ligands……………………………………………40

Figure 2.1 Genomic cloning of the stumpy/ColXIX………………………………...77

Figure 2.2 ColXIX expression during period of axon outgrowth…………………...78

Figure 2.3 MO knockdown of ColXIX phenocopies the stumpy mutation…………79

Figure 2.4 Alignment of human, mouse and zebrafish (zf) ColXIX protein sequences………………………………………………………………...80

Figure 2.5 Mouse ColXIX RNA rescues ColXIX MO phenotype………………….83

Figure 2.6 Mutant ColXIX acts as a dominant negative…………………………….84

Figure 2.7 Rescue of stumpy393 mutants…………………………………………….85

Figure 2.8 Overexpression of full-length ColXIX and ColXIX domains…………...87

Figure 3.1 sema5A is expressed in the myotome at 18 and 24 hpf………………...120

Figure 3.2 sema5B is expressed in the notochord………………………………….121

Figure 3.3 Morpholino knockdown of zebrafish Sema5A results in delay in CaP axon extension and axonal branching and is rescued by full-length rat Sema5A mRNA………………………………………………………………….122

Figure 3.4 CaP axon branching…………………………………………………….123

Figure Page

Figure 3.5 plexin B1-A and plexin B1-B morpholino injection phenotypes and co- injections with sema5A…………………………………………………124

Figure 3.6 Additive and synergistic interactions of Sema5A and Plexin A3………126

Figure 3.7 Rescue of sema5a knockdown embryos with rat Sema5a thrombospondin domain and rat sema5a sema domain…………………………………..127

Figure 4.1 Injection of Sema5A rat RNA into topped458 embryos rescues delay of CaP axon extension……………………………………………………..142

Figure 4.2 topped and sema5A MO knockdown phenotypes are different at 26 hpf…………………………………………………………………...143

Figure 4.3 sema5A MO induces branching of CaP axons in 30 hpf topped mutants…………………………………………………………144

Figure 4.4 Branching axons observed in sema5A MO-injected 30 hpf topped mutants……….……………………….………………………...145

Figure 4.5 Expression of cntnap2 in 24 hpf embryos……………………………...146

xvi

Abbreviated terms

A-term amino - terminal

C-term carboxy - terminal

CaP Caudal Primary motoneuron cDNA complementary DNA

Col collagenous

ColXIX Collagen XIXa1

CSPG chondroitin sulphate proteoglycan

ECM extracellular matrix

FACIT fibril-associated collagens with interrupted triple helices hpf hours post fertilization

HSPG Heparan sulphate proteoglycan

LamG Laminin G domain

MiP Medial Primary motoneuron

MO morpholino

NC non-collagenous

Npn neuropilin

RoP Rostral Primary motoneuron

RTK receptor tyrosine kinase

RT-PCR reverse transcription polymerase chain reaction

Sema semaphorin

Sema5A semaphorin 5A

xvii

TM transmembrane

TSR thrombospondin

TSPN amino domain of thrombospondin

xviii

Chapter 1. Introduction to Motor Axon guidance

Overview

Motoneurons are a subset of neurons responsible for controlled movement in both

invertebrates and vertebrates. These neurons send out axons from their positions in the

CNS to muscle targets in the periphery. These neurons need to make connections with the

proper muscle in order for coordinated movement to be possible. It was observed early on

almost 500 years ago by Leonardo da Vinci that nerves follow stereotyped pathways to

their targets (Pevsner 2002). The question of how proper connections are made correctly

by a huge number of neurons has been the subject of inquiry since then.

With the development of staining techniques, tissue culture and in vivo models, scientists in the 19th century were able to learn more about axon pathfinding. The identification of

the growth cone by Cajal and subsequent observation of it in culture by others led the way to the realization that the ability of growth cones to navigate through their environment is crucial to axons connecting to their proper targets. The later realization that growth cones are able to respond to specific molecular cues led to efforts to identify these cues. In this chapter, I will be discussing two groups of molecules found to influence pathfinding by axons during development. The first group consist of the four 1 main classes of axon guidance cues that have been extensively studied in the past two decades. We owe our understanding of the basic mechanisms of axon guidance to the study of these four classes and by discussing how they function we can understand mechanisms at work in guiding axons even in entirely different contexts. There will be an emphasis on the semaphorin family of ligands since a following chapter deals with the axon guidance cue Semaphorin 5A, a member of the semaphorin family. The second group of molecules I will discuss are molecules of the extracellular matrix (ECM) that until recently were only thought of as important for structural integrity of tissues but have increasingly been shown to be actively involved in patterning the nervous system. There will be an emphasis on Collagens in this portion since Collagen XIX will be discussed in a following chapter as well. I will also be discussing the zebrafish as a model for studying motor axon guidance in vivo and go into some of we have learned about motor axon guidance from studying the developing zebrafish motor system.

Growth cones and axon pathfinding

Growth cones were first observed in Golgi stained spinal cord section preparations from chick embryos by the great neuroscientist and histologist Santiago Ramon y Cajal (Cajal

1890). He first compared these structures at the tips of axons to “a living battering ram. . . endowed with amoeboid movements” that had the ability to “push aside mechanically the obstacles it finds in its path until it reaches the region of its peripheral termination”. He later on revised his views on the actions of growth cones from merely pushing aside cells to get to their targets to responding to certain diffusible elements from the target tissue

2 after observing ectopically positioned neurons still being able to reach their intended targets (Cajal, 1990). Taking into consideration the fact that Cajal was observing a static picture of the growing axon, it is quite remarkable that he was able to picture how dynamic and responsive growth cones are. These observations led Cajal to first pose questions about how axons grow and the mechanisms that enable them to make connections with the correct synaptic partner during development (Cajal, 1990).

Our understanding of just how dynamic the growing axon further benefited from the development of tissue culture techniques by Ross G. Harisson in 1907 (Harrison, 1907).

He observed for the first time in live frog embryonic neural tissue explants how growth cones extend and retract ‘filamentous’ processes while moving in amoeboid-like fashion.

He also correctly concluded that there is a requirement for a substratum for the axon to extend i.e., axons could not grow in a fluid medium (Harrison, 1914). Harrison’s work was confirmed, built upon and advanced by many others which resulted in a continually growing body of knowledge about growth cones and axon outgrowth from the early

1900s to the 1950s. Discoveries early on include rates of axonal elongation which revealed that axons could pause and retract, not just advance (Burrows, 1911). It was later on observed that growth cones could retract when coming into contact with other growth cones or neurites which was the first time contact inhibition was observed

(Hughes, 1953).

The first in vivo observation of growth cones was done by Speidel in 1933. Speidel used the transparent tail fin of the frog tadpole to observe the growth cones of sensory neurons

in anesthesized animals and was able to look at them for hours and over period of weeks

to see the axons establishing connections in the skin (Speidel, 1933). He was able to

conclude that the appearance of growth cones in the live animal was highly similar to

growth cones observed in culture, thereby giving credence to the tissue culture work done

by others (Speidel, 1933).

From these pioneering observations of the growing tip of axons, several fundamental

facts about how exactly growth cones are directed to their precise targets in the

developing embryos were established. Early on it was thought that axons grow out

initially at random and that only functional connections are maintained. This theory was

displaced by the ‘chemoaffinity hypothesis’ due largely to landmark experiments done by

Roger Sperry (1943). Sperry took advantage of the regenerative ability of the optic nerves

of adult amphibians. The optic nerve is composed of axons from the retinal ganglion cells

which innervate the tectum and effectively creates a map of the visual world. The map in

the tectum is reversed with respect to the map on the retina. Sperry severed the optic

nerve which contains axons of the retinal ganglion cells and rotated the eye by 1800 and for control animals he left the eye untouched. He found that the newts with the rotated eye, after regeneration and re-innervation of the nerve, behaved as if the world was upside down and left and right were reversed. This suggests that the regenerated optic nerve managed to grow back and innervate their original location in the tectum not taking into account the rotated eye. Sperry suggested that each axon had ‘identification tags’ or markers that matched markers on their targets in the tectum (Sperry, 1943).

The identification of these unique ‘tags’ and ‘markers’ then became a main objective in

understanding how nervous system patterning happens during development. In short, the

question of how axons pathfind to their final targets now have to be addressed by the

study of molecules or cues that guide the growth cone. However, these cues first had to

be identified. Two simple criterion need to be met for a molecule to be considered an

axon guidance molecule 1) expression, 2) function (Chien, 2005). The first criterion is

met if a molecule is expressed at the right location in the environment or in the growth

cone itself, and at the right time to influence the pathfinding of an axon. It should also be

in an active form and biologically-relevant concentration. To meet the second criterion, the question of sufficiency and necessity is often asked. Sufficiency is most easily addressed in culture where a potential cue can be presented to cultured axons to see if it

can induce the expected response. Necessity is best addressed in vivo by knocking out or

knocking down the expression of the molecule in question and observing its effect on the

process being observed. The interpretation of whether a molecule is necessary or not is

often complicated by redundancy in function of multiple molecules so even if a cue does

not appear to be strictly necessary, its potential involvement in that process should not be

discounted (Chien, 2005).

Axon guidance molecules

Axon guidance molecules have been classified as involved in mechanisms of chemo

repulsion, chemoattraction,, contact-dependent repulsion and contact-dependent

attraction (Figure 1.1). Positive cues may be merely permissive substrates to actively

attractive cues. Negative cues may be repulsive cues or be able to function as a barrier to

the growth cone (Mueller, 1999). This classification scheme may be an over-

simplification but it is most useful in thinking about different cues involved in axon guidance. The reality has been shown to be more complicated with some cues being both attractive and repulsive to the same growth cone but at different conditions such as cyclic

AMP levels in the growth cone (Ming et al., 1997). This implies that cues act attractive and/or repulsive not by any intrinsic characteristic of the molecule but by the ability of the growth cone to respond a certain way such as by expression of a receptor at the

growth cone surface or even by the levels of intracellular second messengers (Mueller,

1999).

For the rest of this section, I will be discussing families of classical axon guidance

molecules. By studying these molecular cues, we can understand different underlying

mechanisms of molecular guidance of growth cones. This will also show the variety of

methods and techniques that were used to identify and characterize these guidance cues.

Netrins and its receptors

Growth cones may be required to navigate relatively long distances to get to their final

target. Growing axons are observed to take stereotyped pathways to their targets. These

pathways are simplified by fragmentation into shorter segments which terminate in

structures or groups of cells called intermediate targets (Figure 1.2). Intermediate targets

will be discussed in further detail later. For now, we will look at one of the best described

intermediate targets, the midline. It is important that the two halves of an organism with a

bilateral nervous system be able to communicate with each other. Commissural neurons

provide this communication by sending out axons across the midline. These commissural

axons need to cross the midline for proper connections to occur. This pathfinding by the

commissural axons through the midline poses two questions. First, how are axons that need to form contralateral projections (those that cross the midline) differentiated from axons that from ipsilateral projections (those that do not cross the midline)? Second, how are the ipsilateral projections induced to stay on the same side and not cross back through the midline? These questions were what drove researchers to isolate and characterize the first diffusible axon guidance cue, netrins. It was observed in in vitro studies that the spinal cord floor plate of vertebrates acts as a chemoattractant to commissural axons grown in culture (Placzek et al., 1990; Tessier-Lavigne et al., 1988). This attractive factor was eventually isolated in an epic biochemical undertaking using embryonic chick brains to be the factors Netrin-1 and Netrin-2. Sequencing of the protein led to the identification of the genes and gave the indication that these proteins were probably secreted (Serafini et al., 1994). cDNA transfection for both netrins were then shown to be able induce axon growth and re-orienting which further proves the ability of netrin to attract and direct commissural axon growth cones. Expression of netrin-1 and netrin-2 at the floor plate was then confirmed by in situ hybridization and immunohistochemistry at the time when commissural axons are growing towards the floor plate (Kennedy et al., 2004).

It was also found that netrin does not always act as an attractive cue. The trochlear motor

nerve extends away, not towards, the floor plate. consistent with being repelled by a cue

from the floor plate (Colomarino & Tessier-Lavigne 1995). This effect is mimicked in

vitro by co-culturing COS cells expressing netrin-1 and trochlear explants (Colomarino &

Tessier-Lavigne 1995). These data demonstrate that netrins are able to function as

bifunctional cues.

Cloning of the netrin-1 gene revealed its similarities to the unc-6 gene in C. elegans

which has previously been found to be required for circumferential migrations of cells

and axons both dorsally and ventrally along the body wall (Ishii et al., 1992). This

similarity also led to the idea that identified unc-6 receptors in C. elegans may be homologues of vertebrate netrin receptors. Unc-40 and Unc-5 were two other genes that were required in C. elegans for circumferential migration of cells in the dorsoventral axis.

Using homology screening, DCC (deleted in colorectal cancer) was identified as a vertebrate homologue of Unc-40 (Kieno-Masu et al., 1998). The DCC family includes

DCC, neogenin which are both in vertebrates and , Unc-40 in C. elegans and Frazzled in

D. melanogaster (Manitt & Kennedy 2002). DCC is a transmembrane protein that binds netrin with high affinity in vitro and is expressed by commissural axons (Manitt and

Kennedy 2002, Fazeli at al., 1997). Further antibodies against DCC were shown to block axon outgrowth of commissural axons from explants in vitro (Kieno-Masu et al 1996).

The discovery of the Drosophila homolog of unc-40 Frazzled demonstrates the

conservation of this system. These suggest that unc-40/Frazzle/DCC is necessary for

netrin-mediated attraction.

Unc-5 is the other netrin receptor candidate found in C. elegans. Unc-5 is a single-pass transmembrane protein with extracellular Ig-like domains and two thrombospondin type I

domains (Manitt and Kennedy 2002). The intracellular domain consists of a ZU5 domain

and a death domain. In unc-5 mutants, the defective axons are those that migrate away

from the sources of Unc-6/Netrin. This was the first indication that unc-6/netrin may be

functioning as a repellant through unc-5. Further, ectopic expression of unc-5 in

C.elegans and D. melanogaster neurons that do not normally respond to netrin caused caused these axons to be repelled by netrin in vivo (Manitt and Kennedy 2002).

Unc-40/DCC/Frazzled and Unc-5 can interact independently with netrin. However, the ability of netrin to act as both an attractive cue and a repulsive cue is not simply a result of binding to one or the other. Early evidence in unc-40 mutants indicated that axons that were supposed to be affected by unc-5 also had subtle defects in these mutants. There was also data that showed that homologs of unc-5 can interact directly with DCC, presumably acting as a netrin receptor complex (Hong et al., 1999). In studies in done in

Drosophila, it was demonstrated that unc-5 by itself is sufficient for repulsive response to high concentrations of netrin but both unc-5 and frazzled/DCC are required for long range repellant response (Keleman and Dickson 2001). This finding suggests that the

9 relative number of DCC and Unc-6 may be another level of regulation of neuronal response to netrin (Manitt and Kennedy 2002).

Eph receptors and Ephrins

Receptor tyrosine kinases (RTKs) are cell-surface receptors that play key roles in cell-to- cell communication. During development of the nervous system precise signaling between cells is key to achieving a properly patterned and connected nervous system.

Eph receptors are the largest known subfamily of RTKs known so far with 16 presently known vertebrate Eph receptors. Almost all known Eph receptors are expressed in both the developing and/or adult nervous system. Their ligands called Ephrins were later identified as the graded cue that guides retinal axons to their appropriate location in the optic tectum (Cheng et al., 1995).

In the visual system Ephrin-A ligands and their EphA receptors act as a repulsive mechanism to position retinal axons along the anterior-posterior axis (Wilkinson 2001).

Ephrin-A ligands are expressed in a gradient in the tectum while EphA receptors are complementarily expressed in a gradient in the retina. Axons with higher levels of receptors migrate to lower points along the ephrin-A gradient (Wilkinson 2001).

Ephrins function in axon guidance in other contexts as well. The important aspect of Eph signaling which allows it to be a workhorse for cell signaling is its abilities to signal in

either direction and to elicit both attractive and repulsive responses (Wilkinson 2001). A

good example is in forebrain commissural axons wherein reverse signaling by ephrin-B

repels these axons from EphB expression while attracting them to EphA4 expression

(Henkemeyer et al., 1996, Kullander et al., 2001). In addition, Eph proteins can also

communicate with a variety of other cell surface proteins which adds even more

complexity and usefulness to Eph signaling (reviewed in Murai and Pasquale 2003).

Slits and Robo

As was already discussed with netrin, the midline is an important intermediate target

during development and acts as both repulsive and attractive to different axons. Robo was

identified in a genetic screen done in Drosophila to identify genes that regulate midline

crossing (Seeger et al., 1993). Robo encodes a transmembrane receptor member of the

immunoglobulin superfamily. It is highly conserved from C.elegans to vertebrates. Robo

expression is correlated with an axons inability to cross the midline (Kidd et al., 1999). In

robo mutants, axons are seen crossing and re-crossing the midline. This suggests that

Robo is a receptor for a repellent that prevents midline crossing and re-crossing (Brose

and Tessier-Lavigne 2000).

Slit was first identified in Drosophila as a secreted protein and was first thought to

function in patterning the midline. Slit is expressed by glial cells in the midline (Brose and Tessier-Lavigne 2000). It was at first unclear that it was the ligand for robo since slit

mutants have a different phenotype than robo mutants. Slit mutants have the “collapsed

midline” phenotype wherein axons extend towards the midline and stay there (Rothberg

et al., 1990). This difference in phenotype was actually the result of slit being able to bind

to two Robo receptors hence slit mutants had a stronger phenotype than the robo mutant.

Several lines of evidence later on showed that Slit was indeed a Robo ligand. First, transheterezygous mutant flies with one copy of slit and one copy of robo have a more severe phenotype than either of the single heterozygotes which suggests that both act in the same pathway (Kidd et al., 1999). Second, overexpressing Slit in developing muscle repelled axons that normally innervate it (Kidd et al., 1999). And finally, Slit and Robo were shown to interact physically in a cell-based binding assay (Brose et al., 1999).

Therefore Slit was confirmed to be the repulsive cue found in the midline that prevents crossing of ipsilateral axons and prevents the re-crossing of contralateral axons (Brose

and Tessier-Lavigne 2000).

Semaphorins and its receptors

The semaphorin family of molecules is one of the largest family of guidance cues known

today. This family is characterized by the presence of the 500 amino acid semaphorin

(sema) domain. Semaphorins are expressed widely and are conserved across animal phyla. Different functions have been attributed to semaphorins the most well-studied

being its role in axon guidance. Semaphorin functions in vascular growth, immune cell

regulation and tumor progression have also been elucidated (reviewed in Kruger et al.,

2005).

The first semaphorin identified Sema3A/Collapsin was isolated biochemically from

membrane preparations of chick brain that were able to induce collapse of sensory

growth cones. Sema1A/Fasciclin IV which was identified in a screen for glycoproteins

expressed in axonal pathways in the grasshopper was found to have high homology with

Sema3A. These two were the first members of the Semaphorin family which was

grouped based on the presence of the approximately 500 amino acid semaphorin (sema)

domain. At present, there are eight main classes of semaphorins (Figure 1.3) all of whom

may differ considerably in sequence and structural characteristics while having the sema

domain in common (Yazdani and Terman, 2006). Semas may be secreted (class 2 and 3),

transmembrane (classes 1, 4, 5 and 6) or GPI-linked (Class 7). Each class has

distinguishing characteristics including the presence of other domains such as

Immunoglobulin-like (Ig-like) domains and thrombospondin (TSR) domains.

Semaphorins are widely expressed in the developing nervous system with most if not all

semaphorins being expressed by both neuronal and non-neuronal cells (reviewed in Fiore

and Puschel, 2003). Expression of semaphorins appears to be dynamically regulated in particular areas during development, with expression going down with maturity (Yazdani and Terman, 2006). That different semaphorins have diverse expression profiles indicate that they are involved in a variety of functions during development. Often, semaphorins have been found to be required for viability. Functional assays reveal semaphorin’s

involvement in directing tissue morphogenesis by controlling or regulating numerous

cellular processes including adhesion, migration, process formation and cytoskeletal

organization among others (Yazdani and Terman, 2006).

It is highly likely that semaphorins are mainly involved in local, contact-mediated events

(Raper, 2000). Only one sema classes 2 and 3 have no cell attachment points but these

molecules have highly basic carboxy terminal tails that probably binds to cell surfaces or

the extracellular matrix (Raper, 2000).

In early functional studies, semaphorins were initially thought of as exclusively inhibitory

axon guidance cue due to their ability to collapse growth cones in culture (Tamagnone and Comoglio, 2004). However, semas were later shown to also be able to promote axon

extension/outgrowth, with some molecules being able to be both inhibitory and attractive

at different contexts (for example Polleux et al., 2000, Kantor et al., 2004). This ability

may be due to semaphorins being able to signal through different receptors or even

differences in levels of second messengers in the neuron (Kruger et al., 2005).

The first semaphorin receptor identified in an expression cloning screen wherein Sema3A

was fused to alkaline-phosphatase and this chimeric protein was used to probe COS cells

transfected with plasmid libraries (He and Tessier-Lavigne, 1997). Neuropilin-1was

identified in this screen. Neuropilin-1 is a known transmembrane protein expressed along

axonal pathways suggesting that it might be involved in axon guidance (Raper, 2000).

Also Neuropilin-1 antibodies are able to block Sema3A response of cultured primary

neurons (Takagi et al., 1995). Expression of a truncated form of neuropilin-1 also disrupts the response to Sema3A suggesting that this truncated form is able to compete for

Sema3A binding. Moreover, sensory ganglia from neuropilin-1 knockout mice are unable to respond to Sema3A (Kitsukawa et al., 1997).

An important aspect of neuropilins is that they form dimers in vivo. It appears that neuropilin-1 and -2 can form homomers and heteromers and that each combination confers a different binding specificity to the complex (Raper, 2000). Also it appears that neuropilins probably require a co-receptor to transducer a biological signal due to their very short cytoplasmic tail which has no known signaling motifs (Raper, 2000). The potential neuropilin co-receptor was first identified as a binding partner to the viral semaphorin sema A39R. The identified binding partner was related to a large family of transmembrane proteins called plexins which have highly conserved intracellular domains (Raper, 2000). Plexins are also distantly related to semaphorins and were already observed to be expressed in select axons (Raper, 2000). There are four subfamilies of plexins, Plexin A1-4, Plexin B1-3, Plexin C1 and Plexin D1 (Negisihi et al., 2005). The role of plexins in mediating axon repulsion and/or attraction by semaphorins have been well-studied (Negishi et al., 2005). Plexins employ mechanisms involving the Rho pathway and the axonal response to semaphorin is regulated by cyclic nucleotides (Negishi et al., 2005). Semaphorins and plexins will also be discussed further in Chapter 3 of this manuscript.

The role of the extracellular matrix in axon pathfinding

In early axon outgrowth experiments by Ross Harrison, it became evident that growth

cones require a substratum to extend and are unable to do so in a liquid medium

(Harrison, 1914). In vivo, growth cones extend either on cell surfaces or over the ECM. It was originally thought that simple adhesivity of the surface is favorable to growth cone advance. However this was concluded from in vitro studies using synthetic materials such as highly charged poly-amino acid substrates that do not occur in nature (Rogers et al.,

1983, Gundersen and Barrett, 1984, Harris et al., 1985). This idea changed once experiments were performed with naturally-occuring substrates and attachment of growth cones was measured by trying to dislodge them mechanically from the substrate

(Gundersen, 1987). In this experiment, chick dorsal root ganglia was grown on laminin, a naturally-occuring substrate and polylysine. Growth cones were observed to extend more rapidly in laminin than in polylysine but were more difficult to dislodge from the polylysine substrate than from laminin (Gundersen, 1987). A similar experiment was done with two naturally-occurring substrates (laminin and immunoglobulin superfamily member L1) which yielded the same result (Lemmon et al., 1992). The consensus then became that growth cone advance over a substratum is not due to differential adhesion between the growth cone and the substratum but rather guidance molecules, including molecules of the ECM, that exert influence on the advancing growth cone through signal transduction mechanisms that lead to changes in the growth cone cytoskeleton (Gomez and Latourneau, 1994, Kiryushko et al., 2004).

The ECM is a network of macromolecules that form an intricate network filling the

extracellular space. It was originally thought that ECM was an inert scaffold whose

function was to stabilize the physical structure of tissues (Alberts et al., 2007). It has

since become apparent that the ECM plays a more active and complicated role in

influencing cellular processes and interactions. With this complicated function is a

corresponding complex molecular composition. Though our understanding of the

organization of the ECM is by no means complete, much progress has been achieved in

identifying its different molecular components. The components of the ECM can be

classified into two groups 1) proteoglycans which are proteins that are covalently bonded

to glycosaminoglycan (GAG) chains and 2) fibrous proteins which include collagens, elastin, laminins and fibronectins (Alberts et al., 2007). The proteoglycans serve as

“ground substance” due to its gel-like consistency which enables fibrous proteins to be embedded in it. This foundation is strong enough to maintain stability of the tissue while allowing diffusion of other materials such as nutrients and hormones between the circulatory system and other tissues.

The role of the ECM in axon pathfinding has been demonstrated in several ways. It was observed during development that ECM molecules were highly spatially and temporally regulated in developing neural tissue which suggests that they function during the stage

wherein axons are pathfinding (Westerfield, 1987). This is more obvious in the central

nervous system wherein expression of ECM molecules are highly complex during

development then becomes less complex into adulthood. This however is not true of the

periphery. Other evidence of the role of ECM molecules in pathfinding is their ability to

promote neurite elongation in vitro. In fact, ECM molecules such as laminin are routinely

used as substrates for growing dissociated neuronal cultures (Gomez and Latourneau,

1994). Third, when ECM molecules were perturbed, axon pathfinding was affected.

Since this review chapter is concerned with motor axon guidance, I will focus on ECM molecules in the periphery The composition of the neural ECM is significantly different than the ECM in other tissues, thus I will only be discussing ECM molecules that have been shown to influence axon pathfinding in the periphery.

Proteoglycans

The role of proteoglycans in axon guidance was first demonstrated in studies done in the cockroach limb bud wherein enzymatic digestion of heparan sulphate proteoglycans

(HSPGs) induced pathfinding defects in pioneering axons (Holt & Dickson, 2000). Later on, other in vivo studies also showed that chondroitin sulphate proteoglycans (CSPGs) may also function in axon guidance (for example Walz et al., 2002). Recent studies performed in different cellular contexts and in different organisms has further expanded our understanding of the diversity of roles these molecules play during development

(Holt and Dickson, 2005). The ability of these molecules to function in diverse pathways is mostly due to their innate structural diversity. Proteoglycans are composed of a core protein covalently bonded to heparan- or chondroitin sulphate side chains. The type of core protein itself is a source of diversity for proteoglycans. In addition the GAG portion,

which is a linear chain of repeating disaccharides, is a substrate for numerous

modifications which result in specific patterns of sulfation and epimerization (Holt and

Dickson, 2005).

Studies have shown that HS are expressed along the axon tracts in the embryonic mouse brain. To study the function of HSPGs on axon tracts of the brain, a conditional mouse knockout was generated wherein the Ext1 gene which encodes the major HS glycosyltransferase which is necessary for HS function is knocked out specifically in

CNS (Inatani et al., 2003). These mutant mice die on the first day of life. There was also a complete absence of major commissures in the forebrain, a phenotype that was similar to Slit mutants (Holt and Dickson, 2005) This suggests a possible interaction between heparin sulphates and Slits. Further genetic evidence of HS function was demonstrated in

zebrafish ext2 and extl3 mutants generated from a mutagenesis screen for axon

pathfinding mutants in developing zebrafish retina. These genes are glycosyltransferases

necessary for HS biosynthesis. In the dackel (dak) and boxer (box) mutants HS levels are

dramatically decreased (Lee et al., 2004). It was also shown in these mutants by

disaccharide profiling that HS synthesis was significantly disrupted. This disruption of

HS synthesis resulted in specific defects in sorting of retinal ganglion cells (RGC) axons

in the optic tract (Lee et al., 2004). Both of these studies as well as others done in flies

and worms together confirm that HSPGs are important for normal axon guidance in vivo.

CSPGs have been demonstrated to act as an inhibitory cue to chick motoneurons (Oakley

and Tosney, 1991) and to axons of the mouse somatosensory cortex (Crossin et al.,

1989). Also removal of CSPGs from sections of adult rat spinal cord enhanced outgrowth

of dissociated DRGs (Zuo et al., 1998). However there is also evidence that CSPGs may

act to promote axon outgrowth in vitro (Clement et al., 1998) and in vivo in the chick

optic pathway (McAdams and Mcloon, 1995) and in the rat hippocampus (Clement et al.,

1999). In zebrafish, removal of chondroitin sulfates by injection with chondroitinase

ABC induced defects in ventral motor nerve outgrowth (Bernhardt and Schachner, 2000).

Both HSPG and CSPG modulate function of the semaphorin Sema5A in the rat habenula

(Kantor et al., 2004). HSPGs are required in the axon for Sema5A to act as an attractive

cue whereas CSPGs in the environment converts this attractive response to a repulsive response (Kantor et al., 2004). Taken together, these studies show that proteoglycans play important roles in diverse functions that guide axons to their targets possibly by interacting with other axon guidance cues.

Fibrillar ECM Molecules

Laminins

Laminins are large glycoproteins composed of three covalently bonded polypeptide chains (α, β and γ) in the triple coiled region of the molecule which form a cruciform structure (reviewed in Luckenbill-Edds 1997). In vertebrates there are at least five α, three β and two γ which combine to form at least fifteen different laminins (Yu et al.,

2007). Laminins are major components of the basement membrane and are among the

earliest ECM molecules required for proper development (Luckenbill-Edds 1997, Pollard

et al., 2006). Genetic studies have revealed roles for laminins during development in

assembly of basement membrane and cell polarization which impact cell survival and tissue organization (Li et al., 2003). This function of laminin is dependent upon its ability

to form polymers that anchor cells to the basement membrane via interactions with cell

surface receptors such as β1-integrins and dystroglycans (Li et al., 2003).

Tenascins

Tenascins are a family of oligomeric glycoproteins with five known members Tenascin-c,

Tenascin-R, Tenascin-X, Tenascin-Y and Tenascin-W. Tenascins have both attractive

and inhibitory functions in neural development (Joester and Faissner, 2001). Tenascin-C,

Tenascin-R and Tenascin-Y have all been observed in the nervous system. Tenascin-C

and Tenascin-R in particular appear to be highly spatially and temporally regulated

during development suggesting specific functions for these two Tenascins. Surprisingly,

Tenascin-C knockout mice were viable, fertile and displayed no overtly abnormal

behavior. However a more systematic analysis of behavior revealed certain abnormalities

in motor coordination and exploratory behavior which suggests defects in fine-tuning

neuronal connections or even synaptogenesis (Joester and Faissner, 2001). In zebrafish

knocking down Tenascin-C induces abnormal branching of ventral motor nerves and

appears to be a substrate for muscle-specific kinase (MuSK) (Schweitzer et al., 2005).

Collagens

Collagens are a family of ECM molecules that play important structural roles as well as

other cellular functions. These molecules are best known for forming elongated fibrils that provide tensile strength in vertebrate tissues like the tendon, cartilage, bone and skin

(Kadler et al., 2007). Collagens are also known components of basal lamina. Collagens

are composed of three polypeptide α chains and are characterized by collagenous domains containing Gly-X-Y repeats in each of the α chains. Some collagens are heterotrimers, composed of different α chains and others are homomers. The chains are held together by interchain hydrogen bonds. Some collagens may have interruptions in the triple helix (Kadler et al., 2007). These interruptions are termed non-collagenous domains (NC) and have been shown to play important roles in collagen function.

30% of the human body constitutes of collagen proteins making it the most abundant proteins by mass (Myllyharju and Kivirikko, 2004). There are at least 28 different vertebrate collagens (numbered I-XXVIII) (Kadler et al., 2007). In vertebrates, collagens are classified by domain homology and function. Fibril-forming collagens are the source of tensile strength in animal tissues (Kadler et al., 2007). These collagens are generated as procollagens with N- and C- propeptides at each end. Cleavage of these peptides is required for fibril formation (Kadler et al., 2007).

Network-forming collagens form an interlaced network in basement membranes by head- to-head interaction of two trimeric NC1 domains (Kadler at al., 2007). The prototypical network forming collagen is Collagen IV.

Transmembrane collagens have long interrupted helices as extracellular domains and short N-terminal cytosolic domains. The ectodomains may be proteolytically cleaved.

These collagens and other collagen-like transmembrane proteins have important roles in neural function, neural tube dorsalization and eye development among others (Kadler et al., 2007).

Other types of collagens are endostatin-producing collagens, which undergo proteolysis to release endostatin, a known inhibitor of endothelial cell migration and axon guidance cue in C. elegans, anchoring fibrils which is mainly Collagen VII which anchors fibrils to the lamina and beaded filament forming collagens which assemble into microfibrils that form structural links with cells (Kadler et al., 2007)

Fibril-associated collagens with interrupted triple helices (FACITs) is the largest subgroup of non-fibril forming collagens. FACITs include collagens type IX, XII, XIV,

XVI, XIX, XX, XXI and XXII. Collagen IX is a heterotrimer with three different α chains while the rest are homotrimers (Boudko et al., 2008). FACITs are characterized by short collagenous (COL) domains interrupted by several noncollagenous (NC) domains with relatively shorter carboxyl-terminal NC domains than fibril-forming collagens

(Boudko et al., 2008). FACITs share high sequence homology at their COL1/NC1 junctions and studies suggest this region is involved in chain selection and assembly of

FACIT collagens (Boudko et al., 2008). Collagen IX is the prototypical FACIT is the best-studied. Collagen IX is a minor cartilage component and is always found in complex

with the fibrillar Collagen II. Genetic evidence indicates that Collagen IX is important for

normal functioning joints (Eyre et al., 2006).

The role of collagens in the nervous system was not appreciated until recently. Non-fibril

forming collagens and other collagen-like proteins appear to be the most likely collagens

to be involved in patterning the nervous system (Fox, 2008). These collagens and

collagen-like protein are widely expressed in both the central and peripheral nervous

system (Fox, 2008). Recent in vivo studies revealed specific roles for a number of

collagens and collagen-like proteins in wiring the vertebrate nervous system (Fox, 2008).

Collagen XVIII is an endostatin-releasing collagen. Its role in zebrafish axon outgrowth

was discovered after identification of the diwanka mutant. The diwanka gene (which will

be discussed more in the following section) was mapped to the multifunctional lh3

glycosyltransferase gene (Schneider and Granato, 2006). lh3/diwanka appears to function

in modifying components of the ECM along the motor axon pathway. In diwanka

mutants, zebrafish primary motor neurons either fail to exit the spinal cord or stall soon

after exiting before getting to the first intermediate target (Zeller and Granato 1999).

When ColXVIII was knocked down using a morpholino, the diwanka phenotype was

recapitulated. This suggests that ColXVIII regulate motor axon outgrowth in zebrafish.

Another collagen identified to have a role in axon pathfinding is Collagen IVa5 (Col4a5).

The role of Col4a5 in target selection in the zebrafish optic tecum was identified by

mapping the mutant dragnet (Xiao and Baier, 2007). Col4a5 is a non-fibrillar, network-

forming collagen (Fox, 2008). In the tectal neuropil of wild type zebrafish larvae, axons

of the retinal ganglion cells (RGCs) terminate in one of the four laminae of the tectum

and only arborize within that layer. In dragnet mutants, RGC axons are unable to select the correct target layer within the tectum and some invade other laminae (Xiao and Baier,

2007). This suggests that Col4a5 is important for correct target selection of RGC axons.

Further, it was found that knocking down expression of HSPGs in the tectum phenocopies dragnet (Xiao and Baier, 2007). This led researchers to speculate that

Col4a5 assembles at the basement membrane and anchors cues such as HSPGs to the correct location (Xiao and Baier, 2007).

The zebrafish as a model for studying axon guidance

The breadth of knowledge we currently have about axon guidance is owed to the availability of a number of different models in which questions about axon guidance could be addressed. By simply growing neurons in different configuration in culture, the mechanism of action of specific molecules or combination of molecules have been elucidated. The grasshopper was also particularly useful due to its relatively large size

and amenability for whole embryo culture. With the use of single cell labeling and ablation, cellular interactions involved in pathfinding decisions by axons were easily observed (Van Vactor and Lorenz, 1999). When genetics was applied to address questions of axon guidance, D. melanogaster and C. elegans became the favored organism due to their smaller size, higher numbers, and ability to perform genetic screens

(Van Vactor and Lorenz, 1999). Mutagenesis screens in these organisms uncovered many genes involved in axon guidance. Initial skepticism to the applicability of knowledge gained from these invertebrates to higher vertebrates was common but the high degree of molecular conservations between systems soon became apparent (Van Vactor and

Lorenz, 1999). For the goal of applying what we know about axon guidance to treatment of human disease and disorders, however, using a vertebrate model is ideal. Studies in rat and mouse have been useful in confirming the importance of numerous molecules to axon guidance in the vertebrate context. However, these organisms present numerous challenges in studying axon guidance such as inability to observe axons during development in vivo as well as time limitations for development, etc. A model that has combined the advantages of working with smaller organisms like Drosophila and the characteristics of vertebrate organisms is the zebrafish.

The zebrafish (Danio rerio) has been a very useful model for studying axon guidance.

The developing zebrafish embryo is transparent and has a relatively simple nervous system. Dye labeling of developing axons reveal that developing motoneurons navigate correctly to muscle segments appropriate for their adult functions. In zebrafish, the

earliest developing neurons are called primary neurons and are distinguished by their

early birth, relatively large soma size, and soma position. Primary motoneurons in

particular have been well studied due to their ease of access. There are three primary

motoneurons in each hemisegment of the developing zebrafish trunk. These motoneurons

have been designated according to their axonal trajectory and are namely the Caudal

Primary (CaP) motoneuron, the Rostral Primary (RoP) motoneuron and the Medial

Primary (MiP) motoneuron. Studies using fluorescent dyes to follow these motoneurons

as they develop have revealed important characteristics of each motoneuron. It was seen

that the soma of each motoneuron occupy distinct positions within each spinal cord and

each axon also follows a unique, stereotyped trajectory (Eisen et al., 1986) All three

primary motor axons first exit the spinal cord ventrally following what is referred to as

the common pathway which is the medial surface of the myotome and along the edge of the notochord (Figure 1.4). The common pathway terminates in the horizontal myoseptum which is the first intermediate target or choice point for all three motor axons. From here on each one navigates a different trajectory. The CaP proceeds ventrally along the ventromedial myotome. The MiP extends along the medial portion of the dorsal muscle while the RoP extends laterally (see Figure 1.4). A fourth primay

motoneuron, the Variable Primary (VaP) is also present though not in all hemisegments

and these neurons die majority of the time. VaPs are usually seen extending ventrally

without extending past the horizontal myoseptum. These motoneurons therefore innervate portions of the muscle in each segment in a mutually exclusive manner. These observations suggest the complexity inherent in laying down the network of neurons

27 during development. These three primary motoneurons have differentiated enough from one another that they are able to respond to molecular cues in the environment which should be necessary for laying out their specific pathway. This demands that not only do each neuron differentially express ‘sensors’ such as receptors that could allow the growth cone to respond to the correct cues, but also that these environmental cues be differentially expressed in their respective fields along the unique pathways.

Forward genetics reveals molecules responsible guiding primary motoneurons

A technique that has been used in zebrafish that has proven fruitful in identifying axon guidance cues is the mutagenesis screen. The zebrafish is amenable to such screens due to the facts that they develop externally and can produce large numbers of eggs. Briefly, zebrafish males are exposed to mutagens which would introduce DNA mutations into the sperm. This sperm is then used to fertilize wild-type females to obtain embryos heterozygous for the mutations generated. By incrossing the progeny you can get homozygous mutants from the F3 generation. The key to a successful screen is picking the ideal phenotype to observe for the pathway or aspect of development that is of interest. In particular, there were two different screens have been useful for identifying genes involved in motor axon guidance in zebrafish. The first screen utilized motility and behavioral phenotypes to identify genes affecting development (Haffter et al., 1996). To identify genes that specifically affected motor axon guidance, a select number of the motility mutants that did not exhibit visible muscle/morphological defects were screened using antibodies to visualize axonal projections (Granato et al., 1996). Four axon

guidance mutants were identified through this with two being motor axon guidance

mutants, diwanka and unplugged (Granato et al., 1996). Since it is likely that not all motoneuron defects may cause an observable motility defect early in development, a second more focused screen was designed to identify genes involved in specifically in primary motor axon guidance (Beattie et al., 1999). Another difference between this screen and the previously discusses screen is that early pressure diploids were used which effectively decreased the number of generations needed to obtain mutants (Beattie, 2000).

This screen yielded two primary motor axon mutants; stumpy and topped. The following section will be a discussion of what we have learned about primary motor axon guidance from studying these four genes identified using forward genetics. All four mutants have distinct phenotypes suggesting that they all function in different pathways. Therefore

studying these mutants could also uncover four different signaling pathways or

mechanisms involved in motor axon guidance.

diwanka mutants

diwanka mutants were identified from the motility screen discussed earlier as an

‘accordion class’ mutant which has defects in side-to-side trunk movement which can be

identified at 23 hpf (Granato et al., 1996). Three embryonic lethal alleles of this mutation

were identified (Zeller and Granato, 1999). Using immunohistochemistry, it was

observed that diwanka CaP axons were unable to properly navigate the common pathway

with some axons being unable to exit the spinal cord whereas other axons stalled shortly

after exiting the spinal cord (Zeller and Granato, 1999). diwanka is also required for

proper pathfinding by the other two primary motor neurons MiP and RoP to their muscle

targets. In chimeric embryos generated by blastula transplants, mutant motoneurons in

wild-type embryos grew normally whereas wild type motoneurons transplanted into

mutant embryos had defective axon outgrowth (Zeller and Granato, 1999). This indicates

that diwanka is required in a non-cell autonomous manner relative to the neuron. Further,

they determined that Diwanka function is required specifically in adaxial muscle cells.

Positional cloning of diwanka mapped the mutation to the multifunctional enzyme LH3

that has both lysyl hydroxylase and glycosyltransferase domains (Schneider and Granato,

2006). The glycosyl transferase domain of the LH3 enzyme appears to be important for

modifying guidance molecules in the ECM, particularly collagens. Type XVIII Collagen

was identified as a potential substrate for LH3 that affects primary motor axon guidance

since knocking Col XVIII down phenocopies the diwanka mutant (Schneider and

Granato, 2006). The study of diwanka revealed the importance of modifications of

molecules of the ECM in axon guidance.

unplugged mutants

The unplugged mutant was identified from the motility screen as a mutant having

reduced locomotion (Granato et al., 1996). Antibody labeling to visualize primary motor

axons revealed that in this mutant the growth cone is able to exit the spinal cord but makes different pathfinding mistakes once they get towards the region of the first choice point, the horizontal myoseptum (Zhang and Granato, 2000). The unplugged gene is essential for proper outgrowth by spinal motor axons specifically CaPs and RoPs but

does not appear to be required by other axon tracts in the developing nervous system

(Zhang and Granato, 2000). To determine where unplugged is required, chimeric

embryos were generated that were either unplugged hosts with wild-type CaP or RoP

axons or wild type hosts with unplugged CaP or RoP axons. This experiment revealed that unplugged acts in a non-cell autonomous manner with regards to the axon since in the latter situation wild-type axons developed mostly mutant trajectories while in the latter, unplugged axons extended normally (Zhang and Granato, 2000). It was further found that unplugged was required in a specific subset of muscle cells, the dorsal adaxial cells (Zhang and Granato, 2000). These cells line the motor axon common pathway at the time the CaP axon pioneers the common pathway (Zeller and Granato, 1999). At the time of RoP pathfinding, these cells have already migrated towards the lateral surface of the somite. This indicates that unplugged function in these cells may enable the ‘priming’ of the common path which directs the growing CaP and RoP axons to make the correct pathfinding decision. Positional cloning of unplugged identified the gene affected in the mutant is the zebrafish homolog of the enzyme Muscle-specific kinase (MuSK) (Zhang et al., 2004). MuSK has a known role in formation of neuromuscular junctions (NMJ) particularly in recruiting rapsyn and acetylcholine receptors (AChRs) (ref). In zebrafish two splice variants of MuSK were identified SV1 and SV2 (Zhang et al., 2004).

However, MuSK SV1 appears to be the isoform of MuSK that is involved in guidance of primary motor axons. First, MuSK SV1 mRNA expression peaks between 10 and 24 somites which cover the time wherein primary motor axon growth cones navigate towards the first intermediate target (Zhang et al., 2004). Second, knocking down the

MuSK SV1 specifically without affecting the full-length MusK phenocopies the

unplugged phenotype while knocking down full-length MuSK specifically does not result

in any CaP axon defects (Zhang et al., 2004). Since it has already been speculated that

unplugged is somehow altering the common path thereby allowing for CaP and RoP to

make the correct pathfinding decisions, the question of whether unplugged/MuSK is

somehow altering the ECM was addressed. Expression of different ECM molecules was

observed in unplugged mutants versus wild types during motor axon outgrowth. Only the

expression and localization of chondroitin sulphate proteoglycan (CSP) showed a marked difference between mutant and wild type (Zhang et al., 2004). This suggests that MuSK functions in a non-rapsyn mediated manner to alter the ECM thereby enabling axons to pathfind properly.

stumpy mutants stumpy was identified in the second screen that utilized antibodies to visualize defects in motor axons (Beattie et al., 1999). The CaP axons in stumpy mutants have a very distinct phenotype with znp-1 antibody labeling wherein at 26 hpf the growth cone is stalled right at the horizontal myoseptum and the axon trunk is thickened (Beattie et al., 2000).

Further, it appears that the CaP axon also stalls at subsequent intermediate targets. Single cell-labeling experiments demonstrated that the other two primary motoneurons MiP and

RoP also exhibit prolonged pausing at intermediate targets. This mutant will be discussed

in further detail in Chapter 2 of this manuscript.

topped mutants

topped was another CaP axon mutant isolated from the antibody screen in zebrafish

(Beattie et al., 1999). In these mutants, CaP axons are delayed in innervating the ventral

muscle and are instead stalled at or around the horizontal myoseptum at 26 hpf as seen by znp-1 antibody labeling (Rodino-Klapac and Beattie, 2004). This phenotype is observed

solely in CaP motor axons and in secondary axons and not in other primary motor axons

(Rodino-Klapac and Beattie 2004). By using chimeras, topped was shown to be required in a subset of ventromedial muscle cells (Rodino-Klapac and Beattie, 2004). This mutant

is also discussed in further detail in Chapter 4 of this manuscript.

Reverse Genetics confirms roles of known molecular motor axon guidance cues in zebrafish

With the development of effective gene knockdown technology for zebrafish, it became

possible to confirm conserved vertebrate processes in the zebrafish system. Morpholinos

(MO) are an effective way to transiently knockdown expression of a gene of interest.

MOs are modified oligonucleotides that are designed to recognize and bind to specific mRNA sequences. By binding to these sequences, MOs are able to knockdown expression of the protein translated from these mRNA. There are two ways an MO can be designed to knockdown expression. A translation blocking MO is designed to the 5’ region of an mRNA sequence including the start site. Binding of the MO at this region

inhibits translation from this mRNA. The knockdown can be confirmed by Western blot

using an antibody against the protein of interest. A splice-blocking MO is designed to

intron-exon boundaries of a pre-mRNA sequence. Binding of the MO to this region 33 prevents proper splicing of the mRNA which can result in either an exon excision intron inclusion. This can be confirmed using reverse-transcription polymerase chain reaction

(RT-PCR) using primers designed to the same region where the MO binds. The presence of a different-sized band than in MO injected versus wild type indicates that the splice- blocking MO was affective.

MOs have been used to investigate the roles of a number of genes in primary motoneuron pathfinding in zebrafish. Some of these genes are summarized in Table 1.1.

Table 1.1 Genes implicated in axon guidance of motor neurons in the developing zebrafish Gene Description Method of Motor axon Reference

Knockdown phenotype

Agrin HSPG localized at Splice –site Truncations, Kim et al., 2007

NMJs and CNS MO erratic

trajectories

Chondroitin Proteoglycans, can Antibody Branching, Bernhardt and sulphate be both repulsive truncations Schachner 2000

and attractive to

axons

Chondroitin 4-O- Chondroitin sulfate Translation- Mostly truncated Mizumoto et al., sulfotransferase- biosynthesis blocking MO 2009

Collagen XVII Endostatin- Splice-site Stalled shortly Schneider and

producing collagen MO after exiting Granato 2006

spinal cord

Neuropilin-1a Semaphorin co- Translation- Truncated, Feldner et al.,

receptor blocking MO branched, erratic 2005

trajectories

Plexin A3 Semaphorin co- Translation- Branched, Feldner et al.,

receptor blocking MO ectopic roots 2007

Plexin B1 Semaphorin co- Splice- Branched, Hilario et al.,

receptor blocking MO delayed 2009

extension

Robo3 Slit receptor Translation- Aberrant Challa et al.,

blocking MO trajectories, 2005

short and ectopic

roots

Scn1bb β1 subunit of Na+ Translation branched Fein et al., 2008

channel blocking MO

Semaphorin 3A1 Repulsive Translation- Delayed Sato-Maeda et

semaphorin blocking MO extension, al., 2006

branched

Semaphorin 5A Bifunctional Splice- Short and Hilario et al.,

semaphorin blocking MO branched 2009

Stat-3 Transcription factor MO Stalled Conway 2006

One way MOs have been utilized is in determining whether two genes interact with each other or are part of the same pathway affecting a specific process. This can be done by injecting sub-threshold doses of each MO by itself and then injecting the two at sub- threshold doses together to determine whether there is a synergistic effect in the read-out phenotype. This was done by Feldner et al. (2005) and they were able to determine that neuropilin-1a likely affects CaP axon outgrowth through different ligands namely VEGF,

Sema3A1 and Sema3A2.

By identifying the remaining axon guidance mutants, we can potentially uncover new mechanisms involved in zebrafish motor axon guidance. And with the use of transient knockdowns it is then possible for us to tease out the different molecules involved in such mechanisms. By using these two approaches we can discover how these molecules togehter facilitate the deceptively simple process of guiding a motor axon from the spinal cord to its muscle target.

Figure 1.1 Classification of axon guidance molecules. Axon guidance molecules can be classified as A) chemoattractive cues; B) chemorepulsive cues; C) contact-mediated attractive cues and D) contact-mediated repulsive cues.

Figure 1.2 Intermediate targets are utilized by growing axons to simplify relatively complex pathways to their targets. In this illustration, a and b are two neurons with unique pathways. Both use intermediate target 1 then diverge to progress to intermediate targets 2a and 2b respectively. Neuron a has already reached target a while growth cone of neuron b is nearing its target b.

Figure 1.3 Semaphorin family of ligands. (Based on Neufeld and Kessler 2008) Classes 1, 2 and 5 are found in invertebrates. Classes 3, 4, 5, 6 and 7 are vertebrate semaphorins. V class is found in certain DNA viruses. Semaphorins all have the large amino-terminal sema domain that is essential for signaling.

Figure 1.4 Schematic of primary motor axon pathways in zebrafish (based on Beattie et al., 2000) A) Lateral view of developing trunk of zebrafish showing hemisegments and developing myotomes. Spinal cord (sc) and notochord (nc) are indicated CaP (orange), MiP (blue) and RoP (green) cell bodies are shown in their stereotypical positions in the spinal cord. (1) These primary motor axons exit the spinal cord at the same ventral root and traverse the common pathway (a). All three primary motor axons pause and then diverge at the 1st intermediate target, the horizontal myoseptum (indicated by gray dashed line). (2) CaP extends ventrally to the ventral muscle (b), MiP branches then extends dorsally along the medial edge of the dorsal muscle (c). RoP extends mediolaterally through the horizontal myoseptum. (3) By 26 hpf MiP and CaP has turned laterally. (B) Cross-section view showing CaP and MiP stereotyped pathways on separate sides of the section for clarity. The common pathway (a), the CaP axon target the ventral muscle (b) and the MiP axon target the dorsal muscle are indicated.

1 2 3

Chapter 2: Collagen XIXa1 is critical for intermediate target navigation by primary motor axons 1 Abstract

During development, motor axons navigate from the spinal cord to their muscle targets in the periphery using stereotyped pathways. These pathways are broken down into shorter segments by intermediate targets where growth cones at the tips of axons are believed to

coordinate guidance cues. In stumpy mutants, the primary motor axons are observed to pause for a prolonged period at these intermediate targets, though the entire embryo appears to develop normally. The stumpy mutation was mapped to the zebrafish homolog of the atypical collagen CollagenXIXa1 (ColXIX) indicating that this gene is involved in axon guidance at intermediate targets. ColXIX expression was observed at primary motor axon intermediate targets at the time of axon outgrowth. Knocking down zebrafish

ColXIX with antisense morpholinos phenocopies the stumpy phenotype and this morpholino phenotype can be rescued by adding back mouse colXIX RNA. Injection of mouse ColXIX protein having the amino acid mutations present in stumpy acts as a dominant negative suggesting the importance of ColXIX complex formation in this axon guidance function of ColXIX. The stumpy phenotype was also partially rescued in mutants by first knocking down zebrafish ColXIX and adding back mouse ColXIX RNA.

1 All experiments in this chapter were completed by JD Hilario with the exception of positional cloning of stumpy which was performed by Chunping Wang. 43

Together, these demonstrate the role of ColXIX in guiding motor axons through intermediate targets in the zebrafish embryo.

Introduction

Proper neuromuscular connections are essential for establishing a functional motor

system in vertebrates. For these connections to be formed, axons must pathfind to

specific targets which may be relatively far away. It has been observed that axons

consistently take stereotypical pathways to their targets and rarely make mistakes. This is due to growth cones at the tip of the axons that are able to navigate using cues present in the environment. This pathfinding is made simpler by the fact that the pathways are actually broken down into shorter segments which are delineated by endpoints known as intermediate targets or choice points (reviewed in Cook et al., 1998). At intermediate targets, growth cones are observed to pause and make extensive contacts with other cells, presumably coordinating cues that convey information to the axon that direct its trajectory (O’Connor, 1999). The growth cone then leaves the intermediate target and then progresses to either the next intermediate target or to its final target. Studies of axon pathfinding in different systems have led to the conclusion that proper connections between a neuron and its target is the result of a series of correct segmental decisions made at intermediate targets. Therefore, correct navigation of each intermediate target by growth cones is key for the proper wiring of the nervous system.

One of the first characterized intermediate targets were the guidepost cells in the developing grasshopper limb bud (O’Connor, 1999). The tibial (Ti) pioneer projection 44

that is the first projection into the limb bud, follows a stereotyped pathway that involves

making steering decisions at exact regions where these guidepost cells are located

(O’Connor, 1999). Sema1A and Sema2A were found to function as attractive and

repulsive cues respectively expressed in the pathway of the Ti projection. Their pattern of

expression guides the growth cone as they navigate through the guidepost cells on their

way to the CNS (Legg and O’Connor, 2003). The best characterized intermediate target

in vertebrates is the midline specifically the floor plate cells. Commisural neurons that sit on the dorsal region of the spinal cord send their projections ventrally and then cross the midline at the floor plate then proceed longitudinally along the spinal cord. Studies in different organisms reveal numerous molecules that influence midline crossing of the commissural axons (summarized in Kaprielan et al., 2001). These molecules include ligand-receptor pairs such as Netrin/DCC and Slit/Robo, and even adhesion molecules such as Axonin-1 and NrCAM (Stoeckli and Landmesser, 1995, Lustig et al., 1999).

These all point to the involvement of numerous molecules with diverse functions in

directing growth cones through intermediate targets on their way to their intended targets.

The zebrafish has proven to be a useful model for addressing questions concerning axon

pathfinding in vertebrates (reviewed in Beattie, 2000). Zebrafish embryos are optically

transparent making it amenable to visualizing early developmental events. The embryonic neuromuscular system is relatively simple and has been extensively studied.

Several mutagenesis screens have been performed to uncover cues involved in early axon

guidance (see Beattie et al., 1999, Beattie, 2000, Hutson and Chien, 2002, Schneider and

Granato, 2003). The axial primary motor axons in particular have been useful in

identifying guidance cues due to their visibility and accessibility. In each trunk

hemisegment in the developing zebrafish embryo, there are three primary motoneurons

that innervate the axial muscle between 18 to 24 hours post fertilization (hpf). These

three motoneurons, the Caudal Primary (CaP) motoneuron, the Medial Primary (MiP)

motoneuron and the Rostral Primary (RoP) motoneuron are distinguishable by the

position of their cell bodies on the spinal cord and from the stereotyped pathways their

axons take to their corresponding muscle targets. All three primary motoneurons send out

axons ventrally out of the spinal cord at first following the ‘common path’ to the first

intermediate target, the horizontal myoseptum. Each motor axon then takes a distinct

pathway to their respective targets. Several axon guidance cues have been previously

identified through their role in guiding the ventrally innervating Caudal Primary

motoneuron or CaP. In diwanka mutant embryos, the primary motoneurons are able to

navigate through and out of the spinal cord but are unable to extend to their common path

to get to the first intermediate target (Zeller and Granato, 1999). The diwanka gene was

later identified as the lysyl hydroxylase 3 (LH3) zebrafish homolog and the subsequent

studies revealed a role for the LH3 enzyme in guiding the CaP axon and other early developing motoneurons by modifying elements of the ECM, particularly collagens, along the common path of the primary motoneurons (Zeller and Granato, 1999;

Schneider and Granato, 2006). The unplugged mutant uncovered a cue required for growth cones to make correct pathfinding decisions after reaching the first intermediate

target (Zhang and Granato, 2000). In these mutants, the growth cones of the CaP and RoP

motoneurons are able to navigate correctly to the first intermediate target but make

incorrect pathfinding decisions at the intermediate target (Zhang and Granato, 2000).

The unplugged gene was later identified as the zebrafish homolog of the Muscle Specific

Kinase (MuSK) which appears to function as a receptor for Wnt ligands and whose signaling works to restrict growth cones to specific regions of the muscle during axon

outgrowth (Jing et al., 2009). The topped gene is another CaP axon guidance cue

identified through a mutagenesis screen. It is yet to be identified but it appears to be a cue required for the CaP axon to extend to the ventral muscle (Rodino-Klapac and Beattie,

2004).

The stumpy mutant was isolated in a screen designed to identify cues that guide primary motor axons (Beattie et al., 2000). stumpy was identified through its CaP axon phenotype wherein CaP axons are stalled at the first intermediate target, the horizontal myoseptum, at 25 hours post fertilization as opposed to wild-type embryos where the axon has already reached the ventral muscle (Beattie et al., 2000). stumpy CaP axons also exhibit a distinct thickening of the axon particularly at the horizontal myoseptum as labeled by the synaptotagmin (znp-1) antibody (Beattie et al., 2000). The styb393 allele is viable in homozygous state and appears to be acting in a partial dominant manner (Beattie et al.,

2000). At later time-points it was observed that the CaP axon also pauses at subsequent intermediate targets. It was further shown that the other two primary motor axons RoP and MiP also exhibit similar defects. Single cell transplants of mutant CaP axon into

wild-type hosts and vice versa also reveal that the function of stumpy is required in both

the CaP axon and in the surrounding cells (Beattie et al., 2000). These all point to a role

of the stumpy gene in the ability of motor axons to navigate past intermediate targets in

vivo.

In this paper we report the positional cloning of stumpy and show that it encodes a

member of the Fibril-Associated Collagens with Interrupted Triple helix (FACIT) family

of collagens, CollagenXIXa1 (ColXIX). FACITs are non-fibril forming collagens

characterized by the presence of short collagenous triple-helical domains interrupted by

short noncollagenous sequences (Prockop & Kivirikko, 1995). Expression of ColXIX has

been observed in human tissue in basement membrane zones of different organ systems

(Myers et al., 1997). In mice expression has been seen transiently in developing muscle

(Sumiyoshi et al., 2001) and in a subset of hippocampal neurons (Su et al., 2009).

In this paper, we show that knocking down ColXIX in zebrafish phenocopies stumpy and

that this morphant phenotype can be rescued by adding back full-length mouse ColXIX

RNA. This morphant phenotype cannot be rescued by mouse ColXIX RNA encoding

mutations present in stumpy mutants. In addition, overexpressing the mutant mouse

ColXIX RNA in wild type embryos induces stumpy phenotype which supports that the

mutant is acting as a dominant negative. Finally, we were also able to rescue the stumpy

phenotype by knocking down mutant ColXIX using a morpholino against zebrafish

colXIX and then adding back mouse ColXIX RNA. We also looked at the function of two

ColXIX protein domains, the amino-terminus globular LamG/TSPN-like domain and the carboxyl-terminus NC2-Col1-NC1 domain in an effort to ascertain which region of the molecule is important for this ColXIX function in motor axon guidance in vivo. We found both regions probably play important roles in the ability of the ColXIX molecule to form complexes with itself and most likely, other ECM molecules as well. Taken together, our data reveals a novel role for ColXIX in guiding motor axons through intermediate target during development.

Materials and Methods

Fish care and maintenance

AB* embryos, ABLF embryos and LF embryos embryos were used for morpholino and

RNA injections and were maintained between 25.5 and 28.5°C. Embryos were staged by converting the number of somites to hours post fertilization (hours post fertilization;

Kimmel et al., 1995). Rescue experiments were performed with stumpy393 homozygous mutants. Embryos were generated by natural mating of homozygous adults.

Genetic Mapping and Cloning

A map cross was generated between sty mutants on the AB* background and wild-type

WIK. Heterozygous fish were identified by incrossing, and haploids were generated by in vitro fertilization of eggs from heterozygous females with UV inactivated sperm. The sty phenotype was determined using whole mount antibody labeling with the znp1 antibody and visualizing motor axons at 26 hpf (Beattie et al., 2000). Microsatellite Z-marker 49

primers (http://zfin.org/cgi-bin/mapper_select.cgi) were used to place styb393 on

Chromosome 13. Any new primer was first tested on an agarose gel. Those that did not show polymorphisms were then analyzed by single strand conformational polymorphisms

(SSCP) (Sentinelli et al., 2000). The two closest markers on either side of styb393 were z25253 (2.1 cM) and z44520 (19.3 cM). For fine mapping a panel of 2542 haploid embryos was generated and more z-markers screened until two more closely linked markers AI476945 (fb55g08) (0.27 cM) and Fc22c08 (0.08 cM) were identified that flanked sty. Using these markers, BAC pools were screened (Danio Key 735, RZPD).

AI476945 mapped to BX3241887 and Fc22c08 mapped to BX322620. By mapping BAC ends and using overlapping BACS, styb393 was placed between 8-90 kB on BX322620.

The 5’ end of the zebrafish mRNA was obtained by performing a tBLASTn search on all

available zebrafish sequences using the first 268 amino acids sequence of the amino-

terminus of mouse ColXIX. This generated several hits but we selected sequences that

BLASTed to genomic sequence upstream from the known ColXIX sequence in LG 13.

This narrowed down the sequences to two Ab initio predicted RNA sequences which

have high homology to the mouse ColXIX amino acid sequence hmm478284 and

hmm476284. Primers were designed from these sequences and RT-PCR was performed.

To determine if these sequences are indeed in one mRNA with our previous sequence, we

performed RT-PCR using a forward primer of hmm476284 and a reverse primer designed

to exon 26. We were able to obtain exons 1-26 in one piece using the Takara One-Step

RNA PCR Kit.

RNA in situ hybridization

Whole-mount RNA in situ hybridization was performed as described by Thisse et al.

(1993). The probe used covers exons 1 – 26 of the colXIX cDNA which was TA cloned into a PCRII plasmid (Invitrogen Topo TA cloning Kit with PCRII vector). An antisense digoxigenin zebrafish colXIX riboprobe was synthesized from a plasmid linearized with

BamHI and transcribed with T7 RNA polymerase (Roche). The control sense probe was synthesized from the same plasmid linearized with NotI and transcribed with SP6 RNA

Polymerase (Roche).

Morpholino Analysis

For antisense oligonucleotide morpholino mediated knockdown of ColXIX, a splice blocking morpholino was designed to the splice-site junction of what was previously believed to be exon 3 and exon 4 but with the completion of mRNA sequencing, was actually exons 17 and 18 (Gene tools, colXIX MO:

GGCAAACCCTGCAAGCCAAAGGAG). 9 ng and 4.5 ng of MO were injected into wild-type embryos and one to two cell stage. Wild-type embryos were allowed to develop to 26 hpf, and subsequently stained with znp1 or used to isolate total RNA for

RT-PCR analysis. A translation blocking MO was also designed (Gene tools, colXIX

ATG MO2: TGCGGAGAAAGTTTATTATTCCAGC) and injected at 4 ng doses.

cDNA Constructs and RNA synthesis

Mouse ColXIX cDNA clone was obtained from Imagenes (Acc. No.: BC118970). The cDNA was cut from PCR-Blunt II-TOPO using XbaI and PvuII and ligated into the PCS2 vector which was cut with EcoRI and blunted. The mutated RNA constructs were generated in this construct using Stratagene Site-Directed Mutagenesis kit. The following primers were used to generate the K791R mutation, KRsense:

5’TCCTGGTCCCACTGGAGCAAGAGGTGACAAGGGTAGTGAGGG 3’,

KRantisense: CCCTCACTACCCTTGTCACCTCTTGCTCCAGTGGGACCAGGA and the following primers were used to confirm the sequence, sdm1:CCCTCACTACCCTTGTCACCTCTTGCTCCAGTGGGACCAGGA; sdm2:

TCCTGGTCCCACTGGAGCAAGAGGTGACAAGGGTAGTGAGGG. The following primers were used to generate the R901stop mutation, Rstopsense:

ACCAGGGGAGCAGGGTGAATGAGGACCTATTGGAGATACAG, Rstopantisense:

CTGTATCTCCAATAGGTCCTCATTCACCCTGCTCCCCTGGT and the following primers were used to confirm the sequence, sdmstop1: AATCCAGGGAGGGGTGAAT, sdmstop2: GAGGGCCTCTGTCTCCTG.

To create the mouse ColXIX amino terminus and carboxy terminus constructs, the region encoding amino acids 1 to 268 and amino acids 1004 to 1136 respectively were amplified by PCR using Invitrogen Platinum Taq. The PCR product was then TA cloned into the

Invitrogen gateway vector PCR8GW (TA cloning kit with PCR8GW, Invitrogen). The

52 orientation of the insert was confirmed via restriction analysis and sequencing. LR cloning was then performed using correctly-oriented constructs to clone the insert into the

PCSMT-Dest destination vector which was graciously provided by the Lawson Lab.

Capped polyA mRNA was transcribed from all the constructs using the mMessage mMachine (Ambion) SP6 or T7 kit. RNA was injected into one to four cell stage wild- type or mutant embryos by itself at 300 pg doses for mutant RNA injections, with 4.5 ng colXIX splice blocking MO at 250 pg for morphant rescue injections or with 4 ng colXIX translation blocking MO at 1000 pg and 300 pg doses for mutant rescue injections.

Whole mount antibody labeling

Whole mount antibody labeling was performed as described in Eisen et al., (1989) and

Beattie et al., (2000). The znp1 monoclonal antibody that recognizes primary and secondary motor axons (Trevarrow et al., 1990; Melancon et al., 1997) was detected using the Sternberger Clonal-PAP system with diaminobenzidine (DAB) as a substrate

(Beattie and Eisen, 1997). Znp1 recognizes synaptotagmin II (Fox and Sanes 2007) and is a good antibody for visualizing early developing motor axons. Embryos were analyzed with a Zeiss axioplan microscope. CaP axons in segments 5-15 on both sides of the embryo were analyzed. The Heparan sulfate proteoglycan antibody (10E4 epitope) was obtained from US Biological and used at 1:100. Anti-Laminin 1 (L 9393) obtained from

Sigma was used at 1:100. Anti-Chondroitin sulphate was (CS56) was obtained from

Sigma and used at 1:100.

Results

Stumpy maps to the zebrafish homolog of ColXIX

To positionally clone the stumpy mutation we generated a map cross between stumpy-/- (

on the AB* line) and WIK. All mapping was done on haploid DNA obtained from sty heterozygous embryos from the map cross. A haploid panel of 2542 embryos was collected and DNA was extracted from each one. The stumpy phenotype was determined by znp1 antibody labeling as previously described (Beattie et al., 2000). All primers were first tested on agarose gels. Those that did not show a polymorphism between the two strains were then analyzed by single strand conformation polymorphism (SSCP) on acrylamide gels (Sentinelli et al., 2000). Using z-markers for each chromosome, sty was mapped to Chromosome 13. We determined that the two closest z markers flanking the mutation were z25253 (2.1 cM) and z44520(19.3 cM). Using a combination of the meiotic map from the Talbot lab (Woods et al., 2005) and the Vega site on the Sanger

Zebrafish sequencing project (http://vega.sanger.ac.uk/Danio_rerio/index.html),

numerous markers were tested until we found two more closely linked markers AI476945

(4/1466, 0.27 cM) and Fc22c08 (2/2542, 0.08 cM) that flanked the sty mutation (Fig. 1).

By screening BAC pools (Danio Key BAC library 735, RZPD) we mapped AI476945

(fb55g08) on BX3241887 and Fc22c08 on BX322620. By SSCP mapping of BAC ends

and using overlapping BACs, sty was placed between 8-90 kB on BAC322620.

To determine what genes were on this BAC (between 1-90 kB), we used GenScan

(http://genes.mit.edu/GENSCAN.html) which identified 6 genes in this BAC (supp

FigX). The sequence corresponding to Gene 2 is exon 14-49 of zebrafish ColXIX . RT-

PCR with 3’ RACE was done to obtain the rest of the exons until a stop site was

identified. The 5’end containing the start site was obtained using tBLASTn to find

sequences in zebrafish chromosome 13 that is similar to the amino terminus of mouse

ColXIX. Primer were then designed to obtain these regions from cDNA.

The entire zebrafish ColXIX gene spans over 300 Kb in chromosome 13. The intron-exon

structure is illustrated in Figure 2.1A. By RT-PCR, we were able to sequence a 3,431 bp

colXIX cDNA. Conceptual translation of this cDNA sequence yields a 1,102 amino acid

protein with a signal peptide sequence, a 185 amino acid amino terminus non-

collagenous domain with similarities to the Laminin G/Thrombospondin N domain and five collagenous (Col) domains of varying lengths separated by short non-collagenous

(NC) domains (Figure 2.1C). The amino acid sequence has 45% and 44% overall sequence identity with human and mouse ColXIX respectively (Supplementary Figure 1).

There is especially high identity in the NC2-Col1-NC1 region at the carboxyl terminus

end of the protein with 62% and 61% identity in this region with human and mouse

ColXIX respectively.

ColXIX is expressed at intermediate targets during pathfinding of primary motor axons

If ColXIX protein is required for CaP axon navigation to the first intermediate target the horizontal myoseptum and to other subsequent intermediate targets, we would expect expression of colXIX at this location during this timepoint. We generated two RNA in situ probes against colXIX transcripts, one encompassing exons 17-21 and the other covering exons 1 – 26. We also generated sense probes as control. The probe that covers the amino terminus gave a much clearer and specific signal and was used to analyze expression in wild type embryos. The embryos were then cut into cross-sections of 16 um. We focused on sections in the middle trunk area. Primary motor axons start to grow out of the spinal cord at around 18 hpf and most have reached their muscle targets by 24-

26 hpf (Beattie 2000). Secondary motoneurons start migrating to their muscle targets at

26 hpf (Beattie 2000). If a gene is important for axons migrating to their targets, it should be expressed at these timepoints. We looked at colXIX expression in 19 hpf embryo sections and saw robust colXIX expression in the ventro-medial portion of the myotome immediately adjacent to the notochord. This location corresponds to the horizontal myoseptum which is the first intermediate target for all three primary motor axons

(Figure 2.2A, 2.2A’). By 24 hpf colXIX expression at the first intermediate target is reduced or mostly gone and expression is now observable in cells at dorso-lateral and ventro-lateral portions of the muscle (Figure 2.2B, 2.2B’). The dorso-lateral portion of the dorsal muscle is where the MiP axon is observed to turn ventrally along the edge of the dorsal muscle and where MiP stalls in stumpy mutants. Expression then expands in lateral cells of the dorsal and ventral myotome towards the horizontal axis by 30 hpf

through 36 hpf (Figure 2.2C, 2.2C’, 2.2D, 2.2D’). These findings suggest that colXIX is

expressed at the right place and time to influence growth cones of primary motor axons

as they navigate intermediate targets on their way to their final targets.

Morpholino knockdown of ColXIX phenocopies stumpy

If the stumpy gene is indeed colXIX we would expect that disrupting ColXIX would have

the same phenotype as sty. Morpholinos are an effective way for knocking down

expression of genes in zebrafish (Nasevicius et al., 2000). We designed two morpholinos

against zebrafish colXIX, a splice-site morpholino against the splice site junction of exons

17 and 18 and a translation blocking morpholinos. Morphants were then fixed at 26 hpf

then labeled with znp-1 antibody to visualize the CaP axons. Both morpholinos were

tested using different doses. For the splice-site morpholino, we used 4.5 and 9 ng doses.

We observed CaP axon defects at both doses that were identical to those observed in

styb393 mutants (Figure 2.3D), with the axons stalled at the horizontal myoseptum as well

as the characteristic thickening of the axons (Figure 2.3C) compared to wild type

uninjected (Figure 2.3B). However, increased death and off-targeting defects were

observed at the 9 ng dose so the 4.5 ng dose was used for the splice-blocking MO. The

efficacy of the splice-site morpholino was confirmed by RT-PCR using exon-specific primers designed for exons 14 to 17 (Figure 2.3A). An incorrectly spliced band that

indicated an inclusion of the 109 bp intron was observed in morphants and not in wild-

type uninjected embryos (Fig 2.3A). The same exact Stumpy-like phenotype was

observed with the injection of translation-blocking (TB) MO (4 ng, 9 ng), however due to

the unavailability of a fish ColXIX antibody, the efficacy of the colXIX TB MO could not

be confirmed. We therefore used the splice site MO for the rest of the experiments unless otherwise specified. With the injection of 4.5 ng colXIX MO we observe Sty-like axon

defects in 84.8%±2% of axons observed (see Table 2.2).

To ensure that the phenotype observed was specifically caused by ColXIX knockdown,

we asked whether the colXIX RNA could rescue the morphant phenotype. For these

experiments we generated ColXIX RNA from full length mouse ColXIX cDNA. We co-

injected 4.5 ng colXIX MO with 250 pg full-length mouse ColXIX RNA at the 1-cell

stage. Only 41.8±2% of observed CaP axons had stumpy-like phenotype versus the

84.8±2% seen in morphants. This indicates that mouse ColXIX is functional in fish and is

able to significantly rescue the phenotype caused by ColXIX knockdown.

Identifying mutations in ColXIX in stumpy mutants

If colXIX is stumpy, then we would expect to find DNA changes in the colXIX gene in

stumpy mutants. RT-PCR was performed using total RNA extracted from sty393 mutant and wild-type 26 hpf embryos. Sequencing of the 3.4 kb colXIX cDNA revealed three mutations in sty393. This relatively high number of mutations was unexpected but could

be the result of the colXIX gene being large. The mutations identified were A2410G,

C2739T and T2826C (see Table 2.2). The stumpy393 mutation should result in a protein

without the last collagen domain Col1 and the last two non-collagenous domains NC1

and NC 2 (see Figure 2.1C) as well as a Lysine to Arginine mutation in the collagen

domain Col 3.

Mutations found in styb393 does not rescue ColXIX MO phenotype

After identifying the amino acid mutations in ColXIX found in stumpy mutants, we then

generated these mutations in mouse ColXIX cDNA. To confirm that the mutant ColXIX

protein is not functional we used mouse ColXIX mRNA containing the stumpy393 mutations to attempt rescue of the morphant phenotype. The two stumpy393 mutated residues K763 and R873 are conserved in mouse (see Figure 2.4) so we generated the mutations K792R and R901stop in the mouse cDNA. We generated these mutations, both separately and together in mouse ColXIX cDNA and asked whether they could rescue the

CaP axon defect in colXIX morphant. We then made RNA and co-injected 250 pg with

4.5 ng ColXIX MO, a dose that induces the stumpy phenotype in WT embryos. The

K792R mutant RNA co-injection with colXIX MO was significantly different than colXIX

MO injection alone (Figure 2.4, percent short axons observed colXIX MO = 84.8 ± 2; colXIX MO + K792R RNA = 80.1 ± 2; p-value = 0.00012) suggesting that this form of

ColXIX does not function as well as the wild-type full-length protein (Figure 2.5, percent short axons observed MO + WT RNA = 41.8 ± 2; K792R = 80.1 ± 2 p-value < 0.0001 ).

The R901stop mutant RNA however was able to rescue to a similar degree as the wild- type full-length protein (Figure 2.5, percent short axons observed MO + WT RNA = 41.8

± 2; colXIX MO + R901stop = 42 ± 3; p value =0.91 ). When both mutations are

generated, the double mutant RNA not only does not rescue but also to a significant degree results in a worse phenotype than MO alone observed (Figure 2.5, percent short axons observed MO = 84.8 ± 2; double mutant = 90.1 ± 2; p-value = 6.26 x 10-6). This

suggests that ColXIX protein containing the styb393 mutations has lost its ability to

function in guiding CaP axons growth cone past the intermediate target.

Mouse ColXIX mRNA with styb393 mutations acts in a dominant negative manner to induce stumpy phenotype

Styb393 has a partially dominant phenotype (Beattie et al., 2000). One way this could occur is for styb393 to be a dominant negative. To confirm this hypothesis we injected

mouse ColXIX mRNA with the styb393 mutations into wild type embryos and if the mutant

protein is indeed acting as a dominant negative we predicted that we would observe the

CaP axon phenotype in injected embryos. We tested the two styb393 mutations separately

and together. We injected 300 pg each of the K792R, R901stop and K792R:R901stop

mouse ColXIX RNA into wild type embryos. We were not able to observe any axon

defects in K792R RNA injected embryos (Table. In embryos injected with the R901stop

mouse RNA, we observed axons exhibiting stumpy-like CaP axon phenotype. We saw

this phenotype in 38% ± 8% (n=152 embryos) of embryos observed but in only 3% ± 1%

of axons (n=3060 axons). When we inject mouse ColXIX RNA that has both mutations,

we observe stumpy-like axons in a significantly higher percentage of fish at 46% ± 9%

(n = 123 embryos) but at a still small percent of axons observed at 4% ± 1% (n = 2460

axons). This suggests that the mutant ColXIX protein is able to function in inhibiting

extension of the growth cone past intermediate targets. This finding is consistent with the

fact that the styb393 mutation acts in a dominant negative manner.

Overexpression of mouse ColXIX induces CaP axon defects

To address whether overexpression of wild type ColXIX by itself would cause some

phenotype we injected wild type colXIX RNA (300 pg) into wild type embryos. We

observed short stumpy like axons in these wild type injected embryos with 2 ± 1% of axons observed being short and this phenotype being observed in 23 ± 8% of the embryos

observed (Figure 2.6, Table 2.2) . In addition we also observed branching in these overexpressed embryos with 8% of CaP axons observed being branched and and this

phenotype being observed in 50% ± 9% of embryos observed. Since we have seen colXIX

transcripts present specifically at intermediate targets this data suggests that in addition to

a higher dosage, the axon defects observed is probably also due to colXIX being

misexpressed. That we observed both axon branching and short axons suggests that

overexpressing ColXIX creates a confusion of cues in the environment that the growth

cone sees at times as permissive (in the case of branching) or repulsive (in the case of

short axons). Alternatively, the stalling of the CaP axons at the intermediate target may

be due to overexpression of ColXIX making the intermediate too adhesive to the growth

cone thus not enabling the growth cone to migrate past it.

Mouse ColXIXa1 full-length RNA rescues stumpy only with knockdown of zebrafish ColXIXa1

Since we have determined that mouse ColXIX is functional in zebrafish we wanted to see

if this could rescue the CaP axon phenotype in stumpy mutants. We injected different

doses (250 pg, 500 pg and 1000 pg) of mouse colXIX mRNA into styb393and were unable

to observe any rescue (data not shown). This was not surprising since we have previously

shown the the mutant protein by itself induces the CaP axon phenotype which suggests that simply adding back wild-type protein will not work. We therefore co-injected both

splice-blocking and translation blocking MOs against ColXIX (4.5 ng and 4 ng doses

respectively) with 1000 pg wild type mouse colXIX mRNA into sty393 embryos with the

goal of knocking down expression of the mutant ColXIX protein which should get rid of

the dominant negative effect. We did not observe any rescue with the splice-blocking MO

co-injection (data not shown). This could be due to the possible maternal loading of pre-

spliced colXIX mRNA which would be impervious to the effect of the splice-blocking

MO. We then did co-injection of 4.5 ng translation blocking MO against ColXIX with the

1000 pg wild type mouse colXIX mRNA into styb393 embryos and we were able to see

CaP axons not exhibiting the short and thick CaP axon phenotype observed in sty

mutants. We define rescued CaP axons as axons that progress past the horizontal

myoseptum at 26 hpf. We saw 12 ± 2% (n = 1100 axons) of CaP axons observed were

rescued in ColXIX translation blocking MO and wild type mouse colXIX RNA co-

injected embryos whereas there are no rescued axons observed in uninjected

styb393embryos (n=600 axons, Figure 2.7). We see rescued axons in 78 ± 11% (n = 55 embryos) of the embryos scored. The ability of mouse colXIX mRNA to rescue the sty 62

phenotype after MO knockdown of zebrafish ColXIX shows that 1) the sty gene is indeed

ColXIX and 2) the mutant zebrafish ColXIX protein is acting as a dominant negative.

Overexpression of mouse ColXIX modules separately induced CaP axon defects

To gain insight into how ColXIX may function at intermediate target, we looked into the

domains present in the ColXIX molecule. With the identification of mutations that are

found in the ColXIX gene in styb393 mutant, we can infer that the protein present in this

mutant presumably lacks the C-terminus end of the protein which consists of the NC2-

Col1-NC1 domain. This suggests that this C-terminus domain of the protein is essential

for the function of the molecule in navigation of motor axons growth cones through

intermediate targets. The two C-terminal NC domains of ColXIX has been previously

demonstrated to have the ability to form trimers either by themselves or with the

requirement of a triple-helix (Boudko et al., 2008). This suggests that this carboxy- terminus domain of ColXIX may be able to function independently. To investigate whether this ability to recognize and bind other ColXIX chains is true for this domain in vivo, we analyzed the last 133 amino acids of the mouse ColXIX. RNA was injected at doses between 25 pg and 300 pg. We predict that if this domain is able to function, it should disrupt proper function of ColXIX by disrupting formation of trimers with the read-out being defects in CaP axon guidance. We saw low survival rate when injecting this RNA at the roughly the same dose as the full-length mouse colXIX RNA which was injected at 250 pg doses. At the 25 pg dose injection of the C-term RNA causes a death rate of 50% in injected embryos before the first day and 23% of those that survived had

morphological defects. These embryos with aberrant morphology were not scored. In the

morphologically normal embryos, we observed different CaP axon phenotypes including

short, stumpy-like axons, branched axons, missing axons and ectopic axon roots along

the spinal cord (Figure 2.8). We also observed embryos with a sequence of hemisegments exhibiting these axon defects.. 28% ±12% of C-term RNA injected fish observed (n=52) had short axons. 19% ± 1% of embryos observed had >2 adjacent hemisegments with axonal defects (summarized in Table 2.3).

We found that overexpressing mouse colXIX RNA encoding the stumpy mutations induced Stumpy-like CaP axon phenotype (Figure 2.6, 2.8). This suggests that the mutant protein may retain some function that could potentially inhibit proper functioning of the wild type ColXIX present in the developing embryo. This function of the protein may be related to interactions that occur between ColXIX chains which are believed to form trimers, or even interaction of ColXIX complexes with other ECM molecules (Boudko et al., 2008, Myers et al., 2003). The mutant protein should still have the amino terminus intact which leads us to believe that this portion of the ColXIX molecule may function in interactions involving the ColXIX molecule which could be how the mutant ColXIX molecule acts as a dominant negative. Thus, we studied the amino domain, specifically the amino globular non-collagenous amino globular domain of ColXIXa1 independent of the rest of the molecule. The amino domain of ColXIX (amino acid 1-268) has similarities to the Laminin G/Thrombospondin N domains as predicted by Interproscan.

This protein domain is known to bind heparins, sulphatides and α-dystroglycan (Tisi et

al., 2000). As with the C-terminus RNA injection, high toxicity was observed with this

amino-terminus RNA at our normally used doses. With 25 pg RNA injection we still observed a 30% death rate and 55% of the injected embryos displayed morphological defects and could not be used for scoring. Similar defects were observed here as was seen

with C-terminus domain. However a small number of missing axons were also seen

(8/1280 axons observed). We observed stumpy like axons in 48 ± 12% of injected

embryos observed and in 4 ± 1% of axons observed (see Table 2.3). We scored the

number of embryos exhibiting axonal defects in >2 adjacent hemisegments and we saw

that 16% ± 1% of the embryos observed exhibit this phenotype.

This data implies that the ColXIX amino-terminus may be binding to other molecules

that could influence CaP axon pathfinding and that overexpression and/or missexpression

of this domain may be affecting localization of such cues/molecules which may be what

is causing the various phenotypes. To address this, we used the sty mutants to visualize

changes in expression of components of the ECM. We looked at expression of

Chondroitin Sulphate Proteoglycans (CSPG), Heparan Sulphate Proteoglycan (HSPG)

and Laminin. Compared to wild type embryos, we did not see any obvious differences in

expression or levels of expression using any of these antibodies. In another study,

Schweitzer et al (2005) also looked at expression of Tenascin-C in stumpy mutants and

found no difference in expression level and pattern with wild-type. This is by no means

an exhaustive list of known ECM components so it would be interesting to look at even

more of these molecules as well as other known axon guidance cues.

Taken together these data suggests that ColXIX has multiple interaction domains that may facilitate its function through interaction with other ColXIX triple helices or other

ECM molecules. These interactions appear to be important for guiding motor axons through intermediate targets as they pathfind to their muscle targets.

Discussion

The collagens are a family of extracellular matrix molecules whose role as structural proteins has been well studied. Collagens are characterized by Gly-X-Y repeats and form triple helices comprised of homo- or heterogeneous polypeptide alpha (α) chains. At least

27 types of Collagens have been identified in vertebrates with 42 distinct α chains

(Myllyharju and Kivirikko 2004). Collagens are classified by their function and domain homology. Fibril-forming collagens are known to mainly provide structural support that maintains tissue integrity (Myllyharju and Kivirikko 2004). Non-fibril-forming collagens, on the other hand, appear to have various non-structural functions. Their roles in the developing vertebrate nervous system have only recently been appreciated (reviewed in

Fox 2008). In studying the stumpy mutant, we have revealed a novel role for a FACIT collagen, ColXIX in the formation of the neuromuscular network during early development.

The stumpy gene encodes the zebrafish ColXIX

The stumpy mutant reveals the existence of a gene necessary for proper navigation of intermediate targets by primary motor axons during development. In this mutant the primary motoneurons are observed to stall for longer periods at their intermediate targets as they pathfind to their final muscle target. Using positional cloning we identified

ColXIX as the gene that is disrupted in stumpy mutants. RNA in situ hybridization shows that colXIX transcripts are present at intermediate targets during the time of axon outgrowth. This puts ColXIX at the right location and timing for it to be functioning in promoting or allowing extension of growth cones past intermediate targets. Further proof that the stumpy gene is ColXIX was that MO knockdown of ColXIX phenocopies stumpy and this phenotype can be rescued by adding back mouse colXIX RNA.This shows that the phenotype observed with the MO knockdown was specific to knockdown of ColXIX and also demonstrated that mouse ColXIX is functional in zebrafish embryos.

We then attempted rescue of the stumpy mutant with mouse colXIX mRNA. Injecting the mouse colXIX mRNA into stumpy mutants by itself was not able to rescue the phenotype.

This is consistent with the stumpy mutation behaving as a partial dominant, that is, the phenotype does not arise from the complete absence of the functional protein but rather, the mutant protein itself is causing the phenotype by potentially rendering the wild type protein present non-functional. Co-injecting the mRNA with splice-blocking MO against

ColXIX also did not work (data not shown) which suggested to us that maternally loaded, pre-spliced mRNA was probably present. It was only by co-injecting the mouse colXIX mRNA with a translation blocking MO against ColXIX that we observed rescue in

stumpy mutants. This further proves that disruption of the colXIX gene causes the sty

phenotype and that ColXIX plays a key role in correct navigation of intermediate targets

by motor axon growth cones in the zebrafish embryo.

To further confirm that the stumpy phenotype is caused by a mutation in the ColXIX

gene, we demonstrated that MO knockdown of ColXIX in zebrafish embryos had a

similar CaP axon phenotype as stumpy mutants. This MO phenotype can be rescued by adding back full-length mouse colXIX mRNA. Interestingly enough, this suggests that stumpy may be a null mutation which we believe to be untrue. The styb393 mutation behaves in a partially dominant manner wherein heterozygous embryos exhibit an intermediate phenotype which indicates that the mutant protein is not acting as a null

(Beattie et al., 2000). To test whether the mutant sty protein is indeed acting as a dominant negative, we generated mouse colXIX mRNA containing the identified

mutations from the 393 allele. mRNA encoding ColXIX protein with both the K792R and

R901stop mutations were not able to rescue the MO phenotype. Moreover, injecting the

mutant colXIX mRNA by itself was sufficient to induce stumpy-like phenotype. This

suggests that the mutant protein by itself has some function contributing to the axonal

defect observed which also implies that the mutant protein that does not have the last 133

amino acids of the protein is still being translated. If that is the case, this mutant protein is

lacking the two non-collagenous domains NC1 and NC2 which has been suggested to be

key for chain selection and initiation of triple helix formation of ColXIX chains (Boudko

et al., 2008). In fact, the NC1 and NC2 domains of ColXIX by themselves or with a triple

helix forming length of amino acids have been shown to be able to form trimers in vitro

(Boudko et al., 2008). This suggests that either the function of ColXIX in CaP axon guidance does not require trimerization/oligomerization or that other domains in the protein may also be involved in trimerization/oligomerization. The second option is more

likely since ColXIX molecules have also been observed to interact via their amino terminus domains (Myers et al., 2003).

Collagen XIX during development

ColXIX remains a poorly characterized collagen especially in terms of its function. It has limited homology to other FACITs which makes it difficult to expand what is known about other FACITs to ColXIX. Electron microscopy images of purified ColXIX

molecules show the molecule consists of a polymorphic, sharply kinked tail and a

globular amino domain through which the molecule appears to form higher-order

structures (Myers et al., 2003). This suggests that the amino globular domain may be

important for forming homomeric complexes. An in vitro study however indicates that

the two carboxy-terminus non-collagenous domains NC1 and NC2 may be important for chain selection and stabilization (Boudko et al., 2008). These findings indicate the importance of interactions in the function of ColXIX. However, specific interaction partners for ColXIX are yet to be identified. Expression studies during early development in mice show that colXIX transcripts are restricted in the developing muscle at embryonic day 11.5 and decreases by embryonic day 16.5 (Sumiyoshi et al 2001). Expression was also observed in smooth muscle cells in the stomach and around the jaw (Sumiyoshi et

al., 2001). This data indicates that ColXIX is developmentally regulated and its function

is important in early development. Data from ColXIX mutants generated in mice

confirms this assumption (Sumiyoshi et al., 2004). ColXIX null mice are normal at birth

but the majority (~95%) of pups die within the first 3 weeks presumably caused by

inability to feed (Sumiyoshi et al., 2004). Defects in the muscle development in the lower

esophageal sphincter that likely caused this phenotype demonstrated a role for ColXIX in the development of skeletal muscle transdifferentiation in the mouse esophagus

(Sumiyoshi et al 2004). A secondary phenotype has also been observed in hippocampal neurons in these null mutant mice (Su et al., 2009). It was observed that neuronal morphology appears to be normal however in null ColXIX mutants however certain subtypes of hippocampal synapses were found to be malformed. This data demonstrates a role for ColXIX in the nervous system specifically in the formation of proper synapses

(Su et al., 2009). These studies done in mice suggest that ColXIX may be playing

multiple functions during development.

Our study in zebrafish shows that ColXIX may in addition be playing a role in setting up the developing motor network specifically by guiding motor axons through intermediate targets. Expression data from RNA in situ hybrization show that ColXIX transcripts are expressed specifically at identified intermediate targets for primary motoneurons during the time of axon outgrowth. This strengthens the argument that ColXIX is indeed playing a role at intermediate targets. The expression of the ColXIX transcripts also suggests that

ColXIX is very highly spatially and temporally regulated during early development as we

observe the expression pattern changes between the timepoints of 19 – 36 hpf. Signal was detectable at two known intermediate targets for the CaP and MiP axons, the horizontal

myoseptum which is an intermediate target for all three primary motor axons and the

dorso-lateral edge of the dorsal muscle which is roughly where MiP axons are observed

to turn (Myers et al., 1986).

Two other collagens have been found to function in axon guidance in the developing

zebrafish. Knockdown of Collagen XVIII induces stalling of CaP axons soon after

exiting the spinal cord, a phenotype similar to the diwanka mutant which has a mutation

in the gene for the multifunctional enzyme LH3 (Schneider and Granato 2006). It

appears that proper glycosyltransferase modification of Col XVIII is required for primary

neuron growth cones to pioneer into the periphery (Schneider and Granato 2006). A type

IV Collagen Col4a5 was also identified to play a role in proper targeting of the axons of

the retinal ganglion cells (RGCs) to the correct lamina of the tectal neuropil (Xiao and

Baier, 2007). Col4a5 functions by ancoring Heparan sulphate proteoglycans (HSPGs) and possibly other secreted factors onto the basement membrane which then guide axons to their proper target (Xiao and Baier, 2007).

Collagen XIX has multiple interaction domains

Collagens are known mosaic proteins due to their being composed of several known protein domains (reviewed in Bork, 1992). These domains are able to fold autonomously and are classified as such due to their similarities to each other in sequence and structure

(Engel, 1996). The NC domains of collagens specifically, have features of mobile

modules or domains.

The C-terminus domain consisting of the NC2-Col1-NC1 portions of ColXIX is absent from the mutant ColXIX present in sty mutants. This implies that this domain is important for the function ColXIX plays at intermediate targets. The carboxy-terminal

NC domains of collagens have been shown to exhibit diverse functions. In the FACIT

Collagen IX, the NC1 domain is responsible for forming crosslinks with the fibrillar

collagen Collagen II (Eyre et al., 2004). The NC1 domains of Collagen XVl exhibit anti-

angiogenic properties (reviewed in Ortega and Werb, 2002). The NC1 domains of different chains of Collagen IV have been shown to both be anti-angiogenic (Colorado et

al., 2000, Kamphaus et al., 2000, Petitclerc et al., 2000, Maeshima et al., 2001) and have

the ability to promote axonal outgrowth (Lein et al., 1991). It was shown in vitro that

NC2 can form stable trimers by itself while the NC1 domain can form trimers only with a

triple helix attached to it (Boudko et al., 2008). This indicates that 1) this domain of the

protein by itself folds properly and exhibits proper function and 2) this domain may

function as a mobile module. As for what specific function this domain has, the in vitro

study by Boudko et al. suggests that this domain may be key for initiating or stabilizing

trimerization of the ColXIX complex. That this domain is missing in the ColXIX protein

in sty mutants tells us that proper trimerization is important for the function of ColXIX in

intermediate targets. Injecting this domain by itself into wild type embryos induces

different CaP axon phenotypes. It could be possible that this domain independently

72 disrupts proper trimerization by binding to wild type ColXIX chains and preventing trimerization. Though there is no evidence yet that ColXIX associates with other fibrillar collagens, it could be possible that like Collagen IX this domain is also important for such interaction. This interaction with fibrillar collagens could potentially be important for anchoring ColXIX at basement membranes at intermediate targets.

We know from studying the different alleles of sty that the sty mutation in the 393 allele is most likely a dominant negative (Beattie et al., 2000). Therefore we know that any mutation identified in this allele should potentially result in a protein that could disrupt proper functioning of wild type protein. As mentioned above, the mutant ColXIX in sty does not have the C-terminus domain but the amino terminus globular domain is still present. This indicates that whatever function the mutant ColXIX possesses may be attributable to this domain. The amino-terminus domain RNA was toxic at doses that were normally used. The dose used for this experiment was 25 pg and at this dose, we see different axonal defects including short, stumpy-like axons and axon branching. This demonstrates that this domain by itself is able to disrupt proper axon pathfinding. It could be that this domain interacts with different cues that either promote extension past the intermediate targets or cues that delimit the axons to their specific path. Other known

LamG/TSPN domain-containing proteins belong to a class of multi-domain molecules that act as molecular briges between cells and the ECM and are also able to facilitate cell- cell communication (Beckmann et al., 1998). Heparin binding is one common activity that LamG/TSPN containing proteins share. It is tempting to speculate that the heparin

binding ability of the LamG/TSPN domain of ColXIX is involved in binding to heparin

sulphate proteoglycans (HSPGs). Proper HSPG function was shown to be required for

proper sorting of retinal ganglion cells axon tracts (Lee et al., 2004). It has been

speculated that HSPGs may be important in the ECM for guiding axons by controlling

the movement of ligands in the ECM (Lee and Chien, 2004).

That the phenotypes observed when overexpressing individual domains is different than

just overexpressing the mutant may tell us that the rest of the molecule is important for

specificity of binding of these domains. There are data that suggests that glycosylation of

hydroxylated lysines in Collagen type IV and VI is important for secretion and assembly

of these molecules (Sipila et al., 2007). This suggests that other NC or Col domains in

ColXIX may also be contributing to proper trimerization and/or complex formation.

ColXIX might function to anchor cues at intermediate targets

We now know that ColXIX plays an important role at intermediate targets that enables growth cones of growing primary motor axons to properly navigate these intermediate

targets. However, the mechanism or pathway that it is involved in is still unclear. From

what we know of its domains, we could infer some of its characteristics that may be key

to its function. The NC2-Col1-NC1 domain appears to be important for trimerization of

ColXIX chains. The NC1 domain could also potentially be involved in interactions with

fibrillar collagens which may be important for anchoring ColXIX on the basement

membranes at intermediate targets. The amino globular domain has similarities to the

LamG/TSPN domain which is also found in laminin family members as well as other extracellular matrix molecules (Elzie and Murphy-Ulrich, 2004). Functions that have been attributed to this domain include cell adhesion, migration, proliferation, and survival

(Elzie and Ulrich, 2004). The LamG/TSPN domain has been shown to have strong heparin binding capabilities (Elzie and Ulrich, 2004). Heparan sulphate proteoglycans which have heparin domains are known axon guidance cues (Inatani et al., 2003; Lee and

Chien, 2004). An interesting concept would be that ColXIX being localized to the intermediate targets, is able to anchor HSPGs or other heparin-containing cues, to the intermediate target which would then signal the growth cone to proceed through the intermediate target.

Alternatively, the LamG/TSPN domain of ColXIX could be recognized directly by receptors expressed by the growth cone. Integrins are known to recognize both laminins and thrombospondin domains (De Freitas et al., 1995, Hughes, 2001). In fact, Laminin-

α1 has previously been shown to function in guidance of multiple axons in the zebrafish

(Paulus and Halloran 2006).

Conclusions

By positional cloning of the stumpy mutant, we identified a novel role for a FACIT collagen ColXIX in navigation of intermediate targets by motor axons in early development. How ColXIX functions specifically in intermediate targets is unclear but our data suggests it may involve physical interactions between ColXIX and cues that

75 guide the growth cone at intermediate targets. There is also the possibility that ColXIX is itself being recognized by receptors in the growth cone. To confirm either one of this, binding partners of ColXIX need to be identified. This study demonstrates further that

Collagens are not merely structural molecules that lend stability to tissues and structures as was once thought but that these molecules are also playing active roles in different cellular processes such as wiring the nervous system.

Figure 2.1 Genomic cloning of the stumpy/ColXIX. (A) The stumpy mutation was mapped to chromosome 13 and further mapped to BAC BX322620 (B) Intron-exon structure of the ColXIX gene. The gene consists of 50 exons. (C) ColXIX protein structure with sty393 mutations indicated. ColXIX is composed of an amino terminus head consisting of a signal peptide and a LamG/TSPN domain. The carboxy terminal tail has five collagenous domains (Col1-Col5) ranging from 72 amino acids to 186 amino acids in length interrupted by five non-collagenous domains (NC1-NC5) ranging from 18 amino acids to 158 amino acids in length.

Figure 2.2 ColXIX expression during period of axon outgrowth. RNA in situ hybridization using colXIX anti-sense (AS) and sense (S) riboprobes which encompasses exons 1-27 of the ColXIX cDNA. Scale bar indicates 20 μm. Spinal cord (sc) and notochord (nc) are indicated by white borders. Arrows show localization of ColXIX transcripts at specific cells of the myotome during timepoints of primary and secondary axon pathfinding.

Figure 2.3 MO knockdown of ColXIX phenocopies the stumpy mutation. (A) RT- PCR shows efficacy of splice-blocking ColXIX MO. MO induces inclusion of 109 bp fragment. CaP axon phenotypes as visualized using znp-1 antibody in (B) uninjected wild-type embryos, (C) wild type embryos injected with 9 ng ColXIX MO and (D) sty393 homozygous mutant embryos. Scale bar indicates 70 µm

Figure 2.4 Alignment of human, mouse and zebrafish (zf) ColXIX protein sequences. Conserved Lysine and Arginine residues that are mutated in stumpy mutants and conserved in mouse and human highlighted in yellow.

Align human vs mouse vs zf colxix protein CLUSTAL W (1.83) multiple sequence alignment

human ------MRLTGPWKLWLWMSIFLLPASTSVTVRDKTEESCPILRIEGHQLTYD 47 mouse ------MRHTGSWKLWTWVTTFLLPACTCLTVRDKPETTCPTLRTERYQ---D 44 zf MFSRGPFSCAKDDMIHYLRWTVFLWIVN-SIPFASGMAVNERIDHTCPPLKLEDKWHTNV 59 * *.:: *: :* .: ::*.:: : :** *: * human NINKLEVSGFDLGDSFSLRR-AFCESDKTCFKLGSALLIRDTIKIFPKGLPEEYSVAAMF 106 mouse DRNKSELSGFDLGESFALRH-AFCEGDKTCFKLGSVLLIRDTVKIFPKGLPEEYAIAVMF 103 zf NLHR-EFTGFDLAEKFLLRKGTVTDSDTPLFRLGSKPLFKPTESVFPNGLXHEYSIVATF 118 : :: *.:****.:.* **: :. :.*.. *:*** *:: * .:**:** .**::.. * human RVRRNAKKERWFLWQVLNQQNIPQISIVVDGGKKVVEFMFQATEGDVLNYIFRNRELRPL 166 mouse RVRRSTKKERWFLWKILNQQNMAQISVVIDGTKKVVEFMFRGAEGDLLNYVFKNRELRPL 163 zf RIRKTTKKDRWFVXQIFDKGGTSQVSLIVDGAKKSVEFLALGFLKNSXLYVFKNRDLHAL 178 *:*:.:**:***: ::::: . .*:*:::** ** ***: . : *:*:**:*:.* human FDRQWHKLGISIQSQVISLYMDCNLIARRQTDEKDTVDFHGRTVIATRASDGKPVDIELH 226 mouse FDRQWHKLGIGVQSRVLSLYMDCNLIASRHTEEKNSVDFQGRTIIAARASDGKPVDIELH 223 zf FDRQFHKLGVSVESNAVSIYLDCELIERQVTAERSGIDVSGRTFITTRLEDGKPVDVELQ 238 ****:****:.::*..:*:*:**:** : * *:. :*. ***.*::* .******:**: human QLKIYCSANLIAQETCCEISDTKCPEQDGFGNIASSWVTAHASKMSSYLPAKQELKDQCQ 286 mouse QLRIYCNANFLAEESCCNLSPTKCPEQDDFGSTTSSWGTSNTGKMSSYLPGKQELKDTCQ 283 zf EILVFCDSRIADLDRCCDSPGAMCEPTVTHNPTAIPLVTGYLQKMLSMP--AQLPTDRCH 296 :: ::*.:.: : **: . : * .. : . *. ** * * .* *:

human CIPNKGEAGLPGAPGSPGQKGHKGEPGENGLHGAPGFPGQKGEQGFEGSKGETGEKGEQG 346 mouse CIPNKEEAGLPGTLRSIGHKGDKGEPGEHGLDGTPGLPGQKGEQGLEGIKGEIGEKGEPG 343 zf CPALKGLKGDPGPQGLPGLKGDKGDPGP---PGPGTLSVEKQAQ-----KGDQGPSGTPG 348 * . * * **. * **.**:** *. :. :* * **: * .* * human EKGDPALAGLNGENGLKGDLGPHGPPGPKGEKGDTGPPGPPALPGSLGIQGPQGPPGKEG 406 mouse AKGDSGLDGLNGQDGLKGDSGPQGPPGPKGDKGDMGPPGPPALTGSIGIQGPQGPPGKEG 403 zf EKGDVGPPGQPGAPGKEGKRGRRGKTGEPGTPGLQGPPGTCDAETVKGMKGDQGVAGERG 408 *** . * * * :*. * :* .* * * ****. *::* ** .*:.*

human QRGRRGKTGPPGKPGPPGPPGPPGIQGIHQTLGGYYNKDNKGNDEHEAGGLKGDKGETGL 466 mouse QRGRRGKTGPPGNPGPPGPPGPPGLQGLQQPFGGYFN---KGTGEHGASGPKGEKGDTGL 460 zf QKGDRGDSGLPGLEG------AGGMKGQKGEEGP 436 *:* **.:* ** * *.* **:**: * human PGFPGSVGPKGQKGEPGEPFTKGEKGDRGEPGVIGSQGVKGEPGDPGPPGLIGSPGLKGQ 526 mouse PGFPGSVGPKGHKGEPGEPLTKGEKGDRGEPGLLGPQGIKGEPGDPGPPGLLGSPGLKGQ 520 zf RGPPGPVMTAHGLRLSGPEGEGSEKGQKGEKGDQGEKGAEGSQGLQGPAGLTGPPGLEGK 496 * **.* . .* .***::** * * :* :*. * **.** *.***:*: human QGSAGSMGPRGPPGDVGLPGEHGIPGKQGIKGEKGDPGGIIGPPGLPGPKGEAGPPGKSL 586 mouse QGPAGSMGPRGPPGDVGLPGEHGIPGKQGVKGEKGDPGGRLGPPGLPGLKGDAGPPGISL 580 zf VGPPGPVGPRGLQGDPGSAGLIGLPGKDGMKGEKGDTGGAPGPAGPPGLKGEPGEP-CSV 555 *..*.:**** ** * .* *:***:*:******.** **.* ** **:.* * *: human PGEPGLDGNPGAPGPRGPKGERGLPGVHGSPGDIGPQGIGIPGRTGAQGPAGEPGIQGPR 646 mouse PGKPGLDGNPGSPGPRGPKGERGLPGLHGSPGDTGPPGVGIPGRTGSQGPAGEPGIQGPR 640 zf GLCGGEEVASGLRGLRGPKGERGVPGVPGDAGEKGEPGFGITGPPGLTGPKGEPGSQGPL 615 * : .* * ********:**: *..*: * *.**.* .* ** **** ***

human GLPGLPGTPGTPGNDGVPGRDGKPGLPGPPGDPIALPLLGDIGALLKNFCGNCQASVPGL 706 mouse GLPGLPGTPGMPGNDGAPGKDGKPGLPGPPGDPIALPLLGDIGALLKNFCGNCQANVPGL 700 zf GPQ---GPIGVPGIEGPPGPQGRPGLPGPPGEPTALPMVGDMGTLMKNACSVCQTRVPGL 672 * *. * ** :* ** :*:********:* ***::**:*:*:** *. **: ****

human KSNKGEEGGAGEPGKYDSMARKGDIGPRGPPGIPGREGPKGSKGERGYPGIPGEKGDEGL 766 mouse KSIKGDDGSTGEPGKYDPAARKGDVGPRGPPGFPGREGPKGSKGERGYPGIHGEKGDEGL 760 zf PGQKGEKGSPGASGIPGMDGEKGDQGLRGPAGNPGKEGPKGVKGERGFPGPMGDKGDEGS 732 . **:.*..* .* . ..*** * ***.* **:***** *****:** *:*****

human QGIPGIPGAPGPTGPPGLMGRTGHPGPTGAKGEKGSDGPPGKPGPPGPPGIPFNERNGMS 826 mouse QGIPGLSGAPGPTGPPGLTGRTGHPGPTGAKGDKGSEGPPGKPGPPGPPGVPLNEGNGMS 820 zf PGVPGLPGSTGRTGSQGMNGRPGAVGPQGQKGDRGSEGPPGQPGPPGPPGAPYSEGNGMS 792 *:**:.*:.* **. *: **.* ** * **::**:****:******** * .* **** human SLYKIKGGVNVPSYPGPPGPPGPKGDPGPVGEPGAMGLPGLEGFPGVKGDRGPAGPPGIA 886 mouse SLYKIQGGVNVPGYPGPPGPPGPKGDPGPVGEPGAMGLPGLEGFPGVKGDRGPAGPPGIA 880 zf SIYKLQNGAANGGQPGPPGPPGPKGDEGRMGEPGLMGLPGLEGLTGAKGDPGLPGPPGLN 852 *:**::.*. . ************ * :**** ********:.*.*** * .****:

human GMSGKPGAPGPPGVPGEPGERGPVGDIGFPGPEGPSGKPGINGKDGIPGAQGIMGKPGDR 946 mouse GISGKPGAPGPPGVPGEQGERGPIGDTGFPGPEGPSGKPGINGKDGLPGAQGIMGKPGDR 940 zf GPVGKPGPRGETGIPGEPGERGPVGETGFPGPEGPPGAPGRPGKDGVPGYEGATGRPGDR 912 * ****. * .*:*** *****:*: ********.* ** ****:** :* *:**** human GPKGERGDQGIPGDRGSQGERGKPGLTGMKGAIGPMGPPGNKGSMGSPGHQGPPGSPGIP 1006 mouse GPKGERGDQGIPGDRGPQGERGKPGLTGMKGAIGPVGPAGSKGSTGPPGHQGPPGNPGIP 1000 zf GTKGERGDPGIPGERGVQGERGK---TGDKGTIGPQGPPGQKGEPGPPGSLTSPGS---V 966 *.****** ****:** ****** ** **:*** **.*.**. *.** .**. human GIPADAVSFEEIKKYINQEVLRIFEERMAVFLSQLK-LPAAMLAAQAY-GRPGPPGKDGL 1064 mouse GTPADAVSFEEIKHYINQEVLRIFEERMAVFLSQLK-LPAAMLSAQAH-GRPGPPGKDGL 1058 zf KLLSDTAALEEIKTFIRNEVLRVFEEKFSDSQTLLQKTPAAILAAQGRQGPPGPPGNDGS 1026 :*:.::**** :*.:****:***::: : *: ***:*:**. * *****:** human PGPPGDPGP---QGYRGQKGERGEPGIGLPGSPGLPGTSALGLPGSPGAPGPQGPPGPSG 1121 mouse PGPPGDPGP---QGYRGQKGERGEPGIGLPGSPGLPGSSAVGLPGSPGAPGPQGPPGPSG 1115 zf PGPPGEPGPPGSQGYRGQKGERGQMGLGLPGAPGPAGPQGPVGLGPQGPSGPPGPPGPHG 1086 *****:*** ***********: *:****:** .*... *. *..** ***** * human RCNPEDCLYPVSHAHQRTGGN 1142 mouse RCNPEDCLYPAPPPHQQAGGK 1136 zf RCNPSDCFHPYG---RRDG-- 1102 ****.**::* :: *

Figure 2.5 Mouse ColXIX RNA rescues ColXIX MO phenotype. 250 pg of wild type mouse ColXIX RNA and mouse ColXIX RNA with mutations were co-injected with 4.5 ng ColXIX MO. P-values were computed for % short axons observed versus MO- injected with significance indicated by a black asterisk (*), and against MO + WT RNA indicated by a red (*).

Figure 2.6 Mutant ColXIX acts as a dominant negative. Sampling of CaP axon defects observed in embryos injected with mutant ColXIX RNA. White dashed line indicates the horizontal myoseptum. Arrowhead indicated sty-like CaP axons. Scale bar indicates 70 μm.

Figure 2.7 Rescue of stumpy393 mutants. Co-injection of translation blocking (ATG) MO and full-length mouse ColXIX RNA. A: stumpy393 mutant embryo uninjected. B – E: selection of embryos exhibiting rescued axons (indicated by arrowheads) and other abnormal CaP axon phenotypes (indicated by arrows). White dashed line indicates the horizontal myoseptum. Scale bar indicates 70 µm.

Figure 2.8 Overexpression of full-length ColXIX and ColXIX domains. Phenotypes observed in overexpression experiments with ColXIX NC2-Col1-NC1 carboxy domain RNA, LamG/TSPN amino domain RNA, and full-length ColXIX RNA. Arrow heads short or missing CaP axons. Arrows indicate ectopic and branched axons. Scale bar is 70 μm.

Allele Amino acid Amino acid WT Mutant Region position position (zebrafish) (mouse)

393 763 792 Lys Arg Col3

393 873 901 Arg Stop Col2

393 903 943 Lys His Col2

Table 2.1. Amino acid mutations found in stumpy393. All three mutations are found in the 393 allele of stumpy and all are conserved between human, mouse and zebrafish (see Supplementary Fig. 1). Column 3 indicates the equivalent amino acid position in the mouse ColXIX sequence.

Injection % fish with Total fish % axons Total axons short axons affected

K792R RNA 0 0

R873stop RNA 38 ± 8 152 3 ± 1 3060

K792R; R873stop 46 ± 9 123 4 ± 1 2460 RNA

Wild-type RNA 23 ± 8 111 2 ± 1 2220

Table 2.2 Overexpression of wild-type and mutant ColXIX mouse RNA. 250 pg of mutant or double mutant RNA and 300 pg of wild-type RNA were injected into one to two cell stage embryos.

Injected % short % fish % fish with # of axons # of fish axons with aberrant short trunk * axons

25 pg N-term 4±1 48±12 16±1 1280 64

25 pg C-term 2±1 28±12 19±1 1080 54

300 pg FL 2± 1 23± 8 4±3 2220 111 RNA

WT 0 0 0 1640 89 uninjected

*fish with >2 adjacent hemisegments exhibiting axon defects

Table 2.3 ColXIX domains affect CaP axons. Injection of RNA encoding mouse ColXIXa1 amino-terminus domain, mouse ColXIXa1 carboxy-terminus domain and mouse full-length ColXIXa1 induce axon defects in wild-type embyos.

Chapter 3: Semaphorin 5A is a Bifunctional Axon Guidance Cue for Axial Motoneurons in vivo.2

Abstract

Semaphorins are a large class of proteins that function throughout the nervous system to guide axons. It had previously been shown that Semaphorin 5A (Sema5A) was a bifunctional axon guidance cue for mammalian midbrain neurons. We found that zebrafish sema5A was expressed in myotomes during the period of motor axon outgrowth. To determine whether Sema5A functioned in motor axon guidance, we knocked down Sema5A, which resulted in two phenotypes: a delay in motor axon extension into the ventral myotome and aberrant branching of these motor axons. Both phenotypes were rescued by injection of full-length rat Sema5A mRNA. However, adding back RNA encoding the sema domain alone significantly rescued the branching phenotype in sema5A morphants. Conversely, adding back RNA encoding the thrombospondin repeat (TSR) domain alone into sema5A morphants exclusively rescued delay in ventral motor axon extension. Together, these data show that Sema5A is a bifunctional axon guidance cue for vertebrate motor axons in vivo. The TSR domain promotes growth of developing motor axons into the ventral myotome whereas the sema

2 The citation for this chapter of the thesis is: Hilario, J.D. et al. (2009). “Semaphorin 5A is a Bifunctional Axon Guidance Cue for Axial Motoneurons in vivo.” Dev. Biol. 1, 190-200. Reproduced from Developmental Biology, 2009, Vol 1, pp 190-200 by copyright permission of Academic Press. All experiments were completed by JD Hilario. 91

domain mediates repulsion and keeps these motor axons from branching into surrounding

myotome regions.

Introduction

The establishment of neuromuscular specificity is an essential step in nervous system

development. Motor axon growth cones, like all growth cones, respond to cues in the

environment via receptors. Major classes of axon guidance molecules such as netrins,

semaphorins, ephrins, slits and their receptors have been identified by biochemistry,

genetic screens, and homology cloning. Our understanding of the specific molecules

guiding motor axons to their targets have been steadily increasing in recent years and

some processes, such as commissural axon guidance, have been very well characterized

at the molecular level. However, there are still some aspects, such as what guides motor

axon growth cones to their ultimate muscle target, that have yet to be fully understood.

To characterize motor axon guidance cues both forward and reverse genetics have been

taken using zebrafish as a model system (see Beattie et al., 2000; Beattie et al., 2002,

Hutson and Chien 2002, Schneider and Granato 2003). In this report we show that the bifunctional molecule Semaphorin 5A (Sema5A) is a cue that guides motor axons into the ventral myotome.

Due to its simplicity and visibility during development, the zebrafish motor axon network has been an excellent system for elucidating guidance mechanisms. The three early developing spinal cord primary motoneurons that innervate trunk axial muscle have been 92

useful for this purpose. In particular, Caudal Primary (CaP) motoneurons that innervate

the ventral axial muscle region have been well studied due to their ease of visualization

and manipulation. Using forward genetics, diwanka/lysyl hydroxlyase 3 was shown to be

necessary for guiding CaP, and the other early developing motoneurons, along the initial

part of their pathway by modifying the extracellular matrix (ECM) in the region (Zeller

and Granato 1999; Schneider and Granato 2006). Unplugged/MuSK also acts early to direct CaP to its correct path after reaching the first intermediate target presumably by

also modifying the ECM (Zhang and Granato 2000; Zhang et al., 2004). Stumpy is

needed for CaP axons and other motor axons to extend past their intermediate targets and

into distal target regions (Beattie et al., 2000). Sidetracked/PlexinA3 defines axonal exit

points along the spinal cord and keeps CaP axons along their defined pathway (Palaisa

and Granato 2007) whereas Topped functions in the ventromedial fast muscle and is

essential for motor axon outgrowth into the ventral myotome (Rodino-Klapac and

Beattie, 2004). However, there is still a paucity of cues identified that affect motor axon

outgrowth into specific myotome regions, Topped functions in this capacity but the gene

is not yet known (Rodino-Klapac and Beattie 2004).

Semaphorins (Semas) are a large family of conserved axon guidance ligands that are

present in both vertebrates and invertebrates. They consist of eight classes that are either

secreted or membrane bound. Class 1 and 2 are present in invertebrates, classes 3-7 are

present in vertebrates, and class 8 is found in DNA viruses (Pasterkamp and Kolodkin,

2003; Semaphorin Nomenclature Committee, 1999). The receptors for these molecules

are members of the Plexin and the Neuropilin families or complexes thereof (Nakamura

et al., 2000; He et al., 2002). The function of a number of Semaphorins in zebrafish CaP

axon guidance have already been elucidated using reverse genetics in zebrafish.

Sema3A1 was identified as a repulsive cue for CaP axons that delimits it to the initial common pathway (Sato-Maeda et al., 2006). It was found that CaP axons sensitivity to this repulsive cue is dynamically regulated through expression of its receptor Neuropilin

1A (Nrp1A) in the axon (Sato-Maeda et al., 2006, Feldner et al., 2005) and that Plexin

A3 is a possible co-receptor (Feldner et al., 2007). Sema3AB was also implicated as an important inhibitory cue in position fine-tuning of CaP cell bodies which is important in establishing proper exit points for CaP motor axons (Sato-Maeda et al., 2008).

Sema5A was shown to be expressed in developing somites and limb buds in mouse embryos and is widely expressed in mesodermal tissues (Adams et al., 1996). In chick,

Sema5A is expressed in the developing telencephalon and heart (Jin et al., 2006, Pineda et al., 2005). Interestingly, inactivation of Sema5A in mice results in embryonic lethality

due not to axon guidance defects, but rather to defects in blood vessel stability in the developing embryo (Fiore et al., 2005). Moreover, it was found that Sema5A plays a very region-specific role in patterning the vasculature, possibly by regulating the position and angle of vessels in the embryonic vascular system (Fiore et al., 2005). Whether the nature of its role in this case is as an attractive or a repulsive cue for endothelial cells could not be determined. Although more emphasis has been placed on the Semaphorins as inhibitory axon guidance molecules due to the presence of the inhibitory Sema

domain, they also have been shown to have attractive capabilities (Wong et al., 1999,

Artigiani et al., 2004; Kantor et al., 2004). Sema1A was shown to function as an attractive cue for developing peripheral neurons when added as a soluble factor in the grasshopper limp bud (Wong et al., 1999). Sema5A as well, was shown to act as not just an inhibitory cue but also an attractive cue both in culture and in the rat diencephalon as seen in brain slices (Kantor et al., 2004). What makes vertebrate class 5 Semaphorins, including Sema5A unique is that they contain two clusters of type-1 thrombospondin repeats (TSR) 3’ to the Sema domain (Adams et al., 1996; Kantor et al., 2004). TSR domains have been shown to promote axon outgrowth in in vitro studies (Neugenbauer, et al., 1991, O’Shea et al., 1991; Osterhout et al., 1992; Rauvala et al, 2000; Kruger et al.,

2004; Pasterkamp and Kolodkin, 2003; Kantor et al., 2004). The presence of an inhibitory Sema domain and a putatively attractive TSR domain in class 5 molecules lends itself to the idea of Sema5A as a bifunctional molecule (Kantor et al., 2004). Other bifunctional axon guidance molecules have been identified namely, Netrin-1 and Sema3b

(Serafini et al., 1996, Julien et al., 2005). In mouse, netrin-1 was found to play a role as an attractant in guiding spinal commissural axons to the floor plate where it is expressed

(Serafini et al., 1996). At the same time, netrin-1 acts as a repellant to trochlear motor axons, steering it away from the floor plate (Colamarino and Tessier-Lavigne 1995).

Whether response to Netrin-1 is attraction or repulsion of a growth cone is dictated by the composition of netrin receptor complexes at the growth cone that consists of DCC by itself or DCC in complex with UNC5 (reviewed by Livesy 1999, Manitt and Kennedy

2002). Sema3b was also found to be a bifunctional cue in the mouse brain wherein it acts

as an attractant to neurons of the anterior pars of the the anterior commisure while it

induces growth cone collapse in neurons of the posterior pars of the anterior commisure

(Julien et al., 2005). Both responses are mediated by a receptor complex consisting of

NrCAM and Nrp2 (Julien et al., 2005). The difference in response is attributed to different signaling cascades being activated by Sema3b in neurons of the anterior pars versus neurons of the posterior pars (Julien et al., 2005).

Being a transmembrane semaphorin it is likely that Sema5A could signal through Plexin alone, without the requirement of a neuropilin co-receptor unlike Sema3A1 (Negishi et al., 2005, Tamagnone et al., 2000). Plexin B3 was previously identified as a putative

Sema5A receptor in COS cells (Artigiani et al., 2004), but no Plexin B3 homologue in zebrafish has been identified. Plexin B3 was also not detected in mice at the time-point at which Sema5A functions to pattern the vasculature suggesting the presence of other

Sema5A functional receptors (Fiore et al., 2005). Plexin B1 belongs to the same subfamily as Plexin B3 and has been shown to mediate Sema4D induced growth cone collapse in hippocampal neurons (Oinuma et al., 2004). In Drosophila, Plexin A was identified as a neuronal receptor for the class 1 Semaphorins Sema1a and Sema1b which are transmembrane semaphorins (Winberg et al., 1998). As mentioned, Plexin A3 genetically interacts with Sema3A suggesting that it may be a component of a receptor complex along with Neuropilin 1a. This signaling is important for guidance of spinal primary motor axons as well as for trigeminal and facial nerve axons (Feldner et al 2007,

Palaisa et al., 2007, Tanaka et al., 2007).

Here we report that Sema5A acts as both a growth-promoting and repulsive axon guidance cue for ventrally extending zebrafish CaP motor axons and is a unique cue as it specifically guides motor axons into a specific myotome region. We demonstrate that this bifunctionality is mediated by the different functions of its two domains, the Sema domain and the TSR domain. We also show that Sema5A is a potential ligand for Plexin

A3 and that it contributes to specifying axonal exit points in the spinal cord.

Materials and Methods

Fish Strains and Maintenance

AB* embryos, ABLF embryos, Tg(gata2:gfp) embryos, and Tg(islet:gfp) embryos were used for morpholino and RNA injections and were maintained between 25.5 and 28.5°C.

Embryos were staged by converting the number of somites to hours post fertilization

(hpf; Kimmel et al., 1995).

Genomic Cloning and Sequencing

The zebrafish semaphorin5A mRNA was predicted from the Sanger Ensembl database

(www.ensembl.org/Danio_rerio). The sema5A gene was isolated by reverse transcription

PCR (Qiagen One-Step RT-PCR kit) using total RNA from 24 hpf pooled AB* embryos.

Gene specific primers were designed based on the predicted mRNA coding sequence. To

determine the intron/ exon boundaries, cDNA sequence was aligned with the zebrafish genomic database and proper boundaries were deduced by comparing cDNA and genomic sequence. RT-PCR products were cloned using the Invitrogen Topo TA Cloning kit. Resulting colonies derived from AB* were sequenced with either SP6 or T7 primers.

Morpholino Analysis

For antisense oligonucleotide morpholino mediated knockdown of Sema5A, two splice blocking morpholinos were designed to the splice donor site of exon 3 (sema5A MO1:

CTTCTTTACTTACACATTACTGGTG) and splice donor site of exon 2 (sema5A MO2

:CCTGAAGAGATGATTTCTAAAGGA). A translation blocking MO was also designed to an ATG-containing sequence upstream of exon 1 sema5A translation blocking

MO(sema5A atg MO: GGTCCACGGGACACTGCTCAGTAC).18 ng of MO were injected into wild-type embryos and Tg(gata2:GFP] and Tg(islet1:GFP) embryos at the one to four cell stage. Wild-type embryos were allowed to develop to 26 hpf, and subsequently stained with znp1 or used to isolate total RNA for RT-PCR analysis.

Tg(gata2:GFP) and Tg(islet1:GFP) injected embryos were allowed to develop to 48 and

72 hpf respectively and motor axons were analyzed with a Zeiss axioplan microscope. A

control MO that does not have any known target in zebrafish was also used (ctl MO:

CCTCTTACCTCAGTTACAATTTATA ). To test the efficiency of the splice blocking

MOs, RT-PCR was performed using the One-Step RT-PCR kit (Qiagen). sema5A

transcript specific primers flanking the targeted MO site were used to amplify an 1100 bp

fragment in uninjected controls and an 850 bp fragment in MO injected embryos. A

different set of transcript specific primers were used to confirm splicing with sema5A

MO2 to amplify a 250 bp fragment in uninjected controls and an 100 bp fragment in

MO2-injected embryos. plexin B1-A and plexin B1-B MOs, designed against the two

isoforms of plexin B1 were kindly provided by Sarah Childs (plexin B1-A MO:

TATTGGCTACTGTACCATGCAGGTC; plexin B1-B MO:

TTCCAGCGGTAAGTTTGAAAACTAG). Both morpholinos were injected at 6 ng

doses. For sub-threshold co-injections with sema5A MO, 3 ng of plexin B1-A or plexin

B1-B MOs and 3 ng of sema5A MO were used. plexin A3 MO (plexin A3 MO:

ATACCAGCAGCCACAAGGACCTCAT)

was kindly provided by Catherine Becker (Feldner et al., 2007). Sub-threshold amounts of 1 ng plexin A3 and 3ng sema5A MO were injected alone or together. Non-specific cell

death, specifically in the head, was observed in plexin A3 morphants. To resolve this we

co-injected 3ng p53 MO (GCGCCATTGCTTTGCAAGAATTG) (Robu et al., 2007)

with the plexin A3 morpholino and the cell death was eliminated. For all injections,

embryos were injected at the one to two cell stage with ~1 nl volume of solution as

calculated by measuring injection bolus size with a stage micrometer. All injected

embryos were then allowed to develop to 26 hours post fertilization (hpf) and stained

with znp1 antibody.

RNA Rescue and Sema5A overexpression

For heterologous RNA rescue, the full-length rat sema5A gene (Kantor et al., 2004) was

cloned into the PCS2 vector. Two versions of the sema5A gene the first with the

99 thrombospondin repeat domain attached to the transmembrane domain (now referred to as TSR-TM) and the second with the Sema and transmembrane domains attached to the cartilage oligomeric protein domain (now referred to as Semacomp) were also cloned into the PCS2 vector (Kantor et al., 2004). Capped polyA mRNA was transcribed using the mMessage mMachine (Ambion) SP6 kit and injected at the one to four cell stage into wild-type zebrafish embryos at 500 pg doses with 18 ng of sema5A morpholino. 500 pg of the generated sema5A RNA alone was also injected into one to four cell stage embryos for Sema5A overexpression.

Whole Mount Antibody Labeling

Whole mount antibody labeling was performed as described in Eisen et al., (1989) and

(Beattie and Eisen, 1997). Znp1 recognizes synaptotagmin II (Fox and Sanes 2007) and is a good antibody for visualizing motoneurons in early development. Embryos were analyzed with a Zeiss axioplan microscope. CaP axons in segments 5-15 on both sides of the embryo were analyzed.

Calculations and Statistical Analysis

Percentage of axons with observed phenotype (axon extension or branching) were calculated from data combined from three different experiments per treatment. Error bars

100

for percentages were calculated using confidence interval for proportions at 95%

confidence with sample size (n) being the total number of axons observed. Difference in

percentages was considered statistically significant at 95% confidence. For rescue

experiments one-way ANOVA was also performed to determine p-values between 18 ng

sema5A MO treatments and co-injections of sema5A MO with either full length Sema5A

rat RNA, TSR-TM RNA or Semacomp RNA. Z-test for proportions was performed to

determine p-values for percentage of moderate and severely branched axons and percent

hemisegment with ectopic axonal exit points between sema5A MO+ plexin A3 MO co-

injections and single low-dose MO injections of each.

Whole-mount In Situ Hybridization

Whole-mount RNA in situ hybridization was performed as described by Thisse et al.

(1993). An antisense digoxigenin zebrafish sema5A and sema5b riboprobes were

synthesized from a plasmid linearized with HindIII and transcribed with T7. For sections,

labeled embryos were soaked in sterile 30% sucrose and then embedded in OCT at -800C

overnight and 20 micron sections were obtained. Sections were analyzed with a Zeiss

axioplan microscope.

Results

The Zebrafish sema5A gene

The sema5A gene is annotated by Sanger Ensembl (www.ensembl.org/Danio_rerio) and is located on linkage group 24. The gene (Ensembl Gene ID: ENSDARG00000058821) 101 spans 21 exons that covers over 200 Kb. The predicted transcript (Ensembl Transcript ID:

ENSDART00000081759) is 3480 kb and is predicted to encode for a 1046 amino acid protein. The domain structure of the predicted protein is identical to that of other vertebrate species consisting of a Sema domain followed by seven type 1 and type 1-like thrombospondin repeats, a transmembrane domain and a short intracellular domain. An incomplete clone containing exons 1 to 21 has been isolated and sequenced and was used as basis for designing primers, riboprobes and morpholinos. ClustalW analysis of sema5A transcripts from different vertebrates show that zebrafish sema5A is more closely related to rat sema5A with 79% homology. A zebrafish homolog of the other vertebrate class 5 semaphorin Sema5B located on chromosome 9. The predicted mRNA is approximately

3000 bp in length and shares 46% identity with sema5A, with a higher degree of homology in the Sema domain (74%) and the thrombospondin domain (71%).

Zebrafish sema5A is expressed in the ventral myotome during development

Zebrafish primary motor axons begin to extend out of the spinal cord at 18 hours post fertilization (hpf), and complete their outgrowth along the medial pathway by ~24 hpf

(Eisen et al., 1986, Beattie 2000), thus we would predict guidance cues functioning in this process would be expressed during these time points. To examine the expression pattern of sema5A in zebrafish, we conducted RNA in situ hybridization at 18 and 24 hpf. We generated anti-sense probes from four different sema5A fragments from non-overlapping regions. The sema5A transcript was detected throughout the embryo at these time-points with slightly stronger expression in the ventral myotome for both time-point (Figure

102

3.1B, 3.1C, 3.1E, 3.1F). Upon sectioning expression was seen in both dorsal and ventral

muscle (Figure 3.1G – F). A more diffused expression throughout the embryo is observed

at later timepoints (36 hpf, 48 hpf, data not shown). Presence of transcript was also detected by RT-PCR from as early as 18 hpf and as late as 48 hpf (data not shown).

These data indicate that Sema5A could function in guiding axons as they pathfind to the ventral myotome. We also looked at expression of sema5b by RNA in situ hybridization

and found that its expression at these time-points does not overlap with sema5A

suggesting that they have divergent functions (Figure 3.2).

Knockdown of zebrafish Sema5A causes delay in motor axon extension into the ventral muscle and axonal branching defects.

If Sema5A were involved in motor axon guidance, we would predict that decreasing the levels of Sema5A would have an effect on this process. To test this hypothesis, we employed splice-blocking anti-sense morpholino oliogonucleotides (MO). We designed a

MO to the splice donor site of sema5A exon 3 to preferentially result in the excision of exon 4 causing a frameshift mutation in the Sema5A protein. We injected 18 ng of the splice-site MO at the one to four cell stage. To verify that the MO was knocking down

Sema5A, we conducted RT-PCR using total RNA obtained from both MO injected embryos, and uninjected wild-type siblings. Using gene specific primers, we amplified an

1100 bp sema5A fragment in wild type and an 850 bp sema5A fragment in MO injected embryos, consistent with a loss of exon 4 that was confirmed by sequencing (Appendix

103

A1). In our hands, we found the MO was successful in excising out exon 4 although we

can still observe some full-length sema5A (Figure. 3.3D).

To observe the effect of Sema5A knockdown on axon outgrowth, sema5A MO injected

embryos (hereafter referred to as sema5A morphants) were fixed at 26 hpf and whole-

mount antibody staining was performed using the znp1 antibody. We observed 10 axons

per side (20 axons per embryo) and scored for position of CaP axon growth cones as well

as for other defects. At 26 hpf CaP axons have a stereotyped morphology and project

down into the either the proximal or the distal portion of the ventral muscle (Figure

3.3A). There was an overall delay in the extension of the CaP axon into the ventral

muscle in sema5A morphants. 20% of hemisegments observed had growth cones stalled

at the horizontal myoseptum or near the ventral edge of the notochord compared to only

1% stalled in uninjected embryos and in control MO-injected embryos (Figure 3.3B,

3.3E). In addition aberrant axon branching was observed in 27% of the hemisegments

observed in morphants compared to only 10% observed in uninjected embryos and 9% in

control MO injected embryos (Figure 3.3C, 3.3F). Examples of branching defects scored

are shown in Figure 3.3. Branching was characterized into slight, moderate and severe

branching phenotypes. Axons were characterized as slightly branching with the

observance of one or more short, fine projections along the axon (Figure 3.4A).

Moderately branched axons may have one to two slightly longer, thicker projections

(Figure 3.4B). Severely branched axons were characterized as having several long

projections or exhibited aberrant patterns of extension such as multiple (>2) branches,

104 loops, early bifurcations, etc (Figure 3.4C). The same phenotypes were observed in 18 ng injections of the translation-blocking morpholino although phenotypes were less consistent and robust (data not shown) and knockdown of protein levels could not be confirmed due to the lack of an antibody, therefore we used splice-blocking morpholinos for subsequent experiments. To determine whether knock down of Sema5A affects later- developing secondary spinal motor neurons, we injected the same amount of sema5A MO into Tg(gata2:GFP) embryos and observed motor axon outgrowth at 48 hpf. We observed no difference in secondary motor axon morphology in sema5A morphants compared to control MO-injected and uninjected embryos. The dorsally projecting motor nerves were also largely normal at 72 hpf as observed in Tg(islet1:GFP) embryos. These results suggest that Sema5A function in motor axon guidance is important during axon outgrowth early in development.

RNA rescue of morpholino phenotype

To ensure that the sema5A knockdown was specific, we rescued the CaP axon phenotype with rat sema5A RNA (Kantor et al., 2004). We generated capped polyA sema5A RNA from cDNA clones and co-injected 500 pg of RNA with 18 ng sema5A MO at the single cell stage. It is important to note that the MO was designed to recognize zebrafish sema5A mRNA but not rat sema5A mRNA. At 26 hpf, we processed embryos for whole- mount antibody labeling. Ten ventrally extending CaP axons per side in the mid-trunk section (20 axons per embryo) were scored for growth cone position and other defects.

Overexpression of full-length rat sema5A RNA alone showed no aberrant phenotype

105

(data not shown). When sema5A MO was co-injected with full-length rat Sema5A RNA,

however, there was a significant decrease in the percentage of motor axons that were

delayed before reaching the ventral muscle with 5% delayed compared to 20% delayed in

embryos injected with MO alone at 26 hpf (Figure 3.3E, Table 3.1). This corresponds to a

four-fold increase in the percentage of axons that reached the ventral muscle in the MO

and RNA co-injected embryos compared to MO injected alone. There was also a

significant decrease in the percentage of branched axons with 11% of motor axons

branched in embryos co-injected with MO and rat mRNA compared with 27% in

embryos injected with MO alone (Figure 3.3F, Table 3.1). This suggests that full-length rat sema5A RNA is able to significantly rescue the aberrant phenotypes in sema5A

morphants.

Sema5A and known zebrafish Plexins

We next asked whether any of the known zebrafish Plexins were receptors for Sema5A.

There are two isoforms of plexin B1 identified in zebrafish by a Blast search and these

two have been arbitrarily named plexin B1-A and plexin B1-B (Sarah Childs, personal

communication). plexin B1-A is located on chromosome 23 and plexin B1-B is located on

chromosome 6. Reduction of Plexin B1-A by a splice blocking MO resulted in an

increase in slight to severe branching of CaP axons while reduction of Plexin B1-B

resulted in delay in CaP axon extension (Figure 3.5A, 3.5C). To determine whether any

of these Plexins may be a receptor for Sema5A we performed co-injections of MOs

against Sema5A and Plexin B1-A or Plexin B1-B at low doses as well as injections of

106 each MO at low dose by itself. There was no synergistic increase in either of the two phenotypes observed in both co-injections (Figure 3.5B, 3.5D) which suggests that though Plexin B1-A and Plexin B1-B may be acting on CaP axon guidance, Sema5A is not the specific ligand for these plexins.

Genetic mutation in the plexin A3 gene, as well as morpholino knockdown exhibits aberrations in motor axon outgrowth (Feldner et al 2007, Palaisa et al 2007). The two phenotypes observed are branching of CaP axons and axons exiting the spinal cord at ectopic exit points (Birely et al., 2005, Feldner et al., 2007, Palaisa et al., 2007). Since the branching defects caused by both Plexin A3 and Sema5A knockdown were similar, we asked whether Plexin A3 may be mediating Sema5A function in primary motor axon guidance. Therefore we performed sub-threshold co-injections of sema5A MO and plexin

A3 MO and individual low dose injections of each MO. Embryos were injected with 1 ng plexin A3 MO, which is 1/9 of the optimal dose used by Feldner et al. (2007). At this dose we observed that 9% of CaP axons were moderately to severely branched. The branching phenotypes observed are identical to what we observe with Sema5A knockdown (Figure 3.6A). Injection of sema5A MO at 3 ng, a much lower dose than our optimal dose which is 18 ng, caused moderate and severe branching in 6% of CaP axons observed. Injecting 3 ng sema5A MO and 1 ng plexin A3 MO together resulted in 16% moderately and severely branching axons which is roughly the sum of severe and moderately branched axons in injections of each MO by itself at 3 ng, indicating that the effect of decreasing Sema5A and Plexin A3 is roughly additive. In addition to the

107

branching phenotype we observed axons coming out of ectopic exit points along the

spinal cord as was observed by Feldner et al., (2007) and Tanaka et al., (2007) in plexin

A3 morphants and by Birely et al (2005) in sidetracked/Plexin A3 mutants (Figure 3.6A).

This phenotype is not observed in uninjected wild type embryos and control MO injected

embryos. We counted the number of axons exiting from ectopic exit points within the ten

trunk hemisegments per side that we use for quantification and saw 5% of hemisegments

had ectopic axons (Figure 3.6C) in embryos injected with 1ng Plexin A3 MO. In sema5A

MO and plexin A3 MO co-injected embryos we observed 17% of hemisegments had ectopic axons. This value is comparable with what was observed in sidetracked/Plexin

A3 genetic mutants where ectopic axons were observed in 5-23% of total hemisegments

(Birely et al., 2005). Since ectopic exit points are not seen in the optimal dose (18 ng) nor

in the low dose (3 ng) injections of sema5A MO we repeated this co-injection experiment

using a second splice-site MO against Sema5A (sema5a MO2) designed to splice out

exon 3. Splicing was confirmed by RT-PCR (data not shown). An 18 ng dose of sema5A

MO2 induces ectopic axonal exit points in 6/2140 hemisegments observed which is about

0.3% incidence. Ectopic axonal exits were not observed with 3 ng dose of sema5A MO2.

Co-injection of 3 ng sema5A MO2 with 1 ng plexinA3 MO induces ectopic axon exit

points in 14% hemisegments observed (data not shown). This is consistent with what we

observe with the first sema5A MO when co-injected with the plexinA3 MO and confirms

Sema5A’s role in this process. Taken together, these data suggest that although Sema5A

and Plexin A3 do not work together to guide motor axons into the ventral myotome, they

do appear to interact to specify spinal cord exit points. Whether this interaction occurs

108 directly, that is as a ligand-receptor pair or indirectly as in they function in the same or parallel pathways is yet to be determined.

The thrombospodin repeat domain of rat sema5A rescues delay of CaP axon extension in sema5A morphants

It was previously shown that Sema5A may act as a bifunctional axon guidance cue via the function of its two domains, the thrombospondin repeat domain and the Sema domain

(Kantor et al 2004). To determine whether it functions as such in motor axon guidance in vivo, we used constructs that contain the different functional domains of rat Sema5A.

The thrombospondin repeat (TSR) domain is believed to be an attractive cue in axon guidance. We generated capped poly A RNA encoding for Sema5A TSR domain attached to a transmembrane domain (TSR-TM) from cDNA clones (Kantor et al., 2004).

Injection of TSR-TM RNA by itself does not induce any axon defects (data not shown).

We then co-injected this RNA with sema5A MO at the one to four cell stage and subsequently fixed and stained the embryos at 26 hpf. As in the full-length RNA rescue experiment, there was an observed decrease in percentage of axons that are delayed before reaching the ventral muscle, with 6% of axons delayed in sema5A MO and TSR-

TM RNA co-injected embryos compared to 20% in embryos injected with MO alone

(Figure 3.7A, Table 3.1). However there was no significant difference in the percentage of axons exhibiting slight and moderate axonal branching in embryos co-injected with sema5A MO and TSR-TM RNA compared to embryos injected with MO alone although

109

there was a decrease in percentage of axons exhibiting severe axonal branching in the

TSR-TM co-injected embryos (Figure 3.7B). This suggests that the thrombospondin

repeat domain of Sema5A influences CaP axon growth cone migration towards the

ventral muscle, but that it does not appear to function in CaP axon pathfinding on the

medial pathway prior to reaching the ventral muscle.

The sema domain of rat Sema5A rescues axonal branching defects in sema5A morphants

The Sema domain has been shown to mediate the inhibitory effect of Semaphorins

(Klostermann et al., 1998, Kantor et al., 2004). To determine whether the Sema domain is

important for CaP axon guidance, we made capped poly A RNA from cDNA clones

encoding for the Sema domain attached to a transmembrane domain and a COMP domain

(Semacomp RNA). The COMP domain is key to oligomerization of Sema proteins, which

is important for its function (Kantor et al., 2004). We then co-injected the Semacomp

RNA with sema5A MO into one to four cell stage embryos followed by analysis of motor

axons at 26 hpf. There was a significant decrease in axons stalled at or before reaching

the ventral edge of the notochord from 20% in MO-injected embryos to 12% of axons in

embryos co-injected with sema5A MO and Semacomp RNA (Figure 3.7C, Table 3.1).

However, this number is still significantly higher than observed delayed axons in full length rat sema5A RNA rescued embryos (5%) or in TSR-TM RNA rescued embryos

(6%) suggesting that the Sema domain only partially rescues this phenotype. There was, however, a significant decrease in the percentage of axons exhibiting aberrant branching in embryos co-injected with sema5A MO and Semacomp RNA (6%) compared to sema5A

110

morphants (29%) (Figure 3.7D). This finding suggests that the Sema domain of Sema5A functions to delimit the axons to their pathway as they migrate towards the ventral muscle.Taken together these RNA add-back experiments reveal that the thrombospondin domain of Sema5A acts in a positive manner to guide axons into the ventral myotome whereas the Sema domain acts primarily in a negative manner to prevent axon branching and keeps axons extending along a defined pathway.

Discussion

The stereotyped outgrowth of ventral axial motor axons is dependent upon the presentation and proper integration of the appropriate guidance cues in the ventral myotome. Our understanding of the cues guiding zebrafish primary motor axons from the spinal cord towards its target has increased in recent years, particularly with cues that govern outgrowth from the spinal cord as well as cues that guide the axons in the common pathway and up to the first intermediate target. However, cues that specifically guide axons toward their final target, the ventral myotome have yet to be determined. By morpholino knockdown of Sema5A, we demonstrate that the Sema5A ligand is a critical cue that guides CaP motor axons into the ventral muscle by its ability to act as a bifunctional axon guidance cue. In addition to promoting axon outgrowth to the ventral muscle, Sema5A appears to also be acting as a repulsive cue in the environment, preventing growth cones from migrating away from the defined axonal pathway. By

RNA rescue experiments we show that this dual function occurs via the two different domains of Sema5A, the thrombospondin domain and the Sema domain. Our data 111

confirms the bifunctionality of Sema5A as a consequence of the different functions of its

two domains and demonstrates the role of Sema5A in the motor axon system in an in vivo

model. We also identify Sema5A as a potential ligand for Plexin A3 by showing that the two genetically interact to prevent aberrant exiting of axons at ectopic points along the

spinal cord.

Sema5A is a bifunctional axon guidance cue for CaP axons

Kantor et al (2004) showed that Sema5A acts as a bifunctional axon guidance cue in the diencephalic axon tract, the fasciculus retroflexus , which originates from the habenula nucleus and extends between prosomere 1 and 2 without crossing into either prosomere.

Sema3F has been implicated as the repulsive cue in prosomere 1 (Funato et al., 2000,

Sahay et al., 2003). Sema5A expression was detected in prosomere 2 as well as in the

habenular neuron cell body. Their findings suggest that Sema5A acts as both an attractive

cue in the axons themselves, causing the axons to remain fasciculated, and as a repellant in prosomere 2 that prevents axons from crossing into that region. In zebrafish, we did not see Sema5A expression in motoneurons which tells us that both attractive and

repulsive effects of Sema5A are being mediated in the environment. The two phenotypes

observed in sema5A morphants, namely increase in branching of the CaP axons and delay

in CaP axon extension into the ventral muscle supports this conclusion. Sema5A may be

acting as a repulsive cue in the environment surrounding the CaP axon pathway and by

knocking down Sema5A levels, the surrounding area becomes permissive for the

migrating growth cones. This phenotype may be comparable to the observation in

112

organotypic diencephalon explants treated with antibodies against Sema5A wherein axon

tracts extended into prosomeres 1 and 2 and were no longer restricted to the boundary

between the prosomeres; and often do not reach their targets (Kantor et al., 2004).

Sema5A was demonstrated previously to also have the ability to be an attractive cue to

habenular neurons in an in vitro stripe membrane assay (Kantor et al., 2004). We show

here in vivo that Sema5A is acting as a cue that promotes extension of motor axon growth

cones into the ventral muscle. Knocking down Sema5A causes a delay in extension of the

CaP axon growth cone into the ventral muscle, suggesting that Sema5A normally functions to guide the axons into the ventral muscle. That the delay is transient and not permanent in our morphants suggests that there could be other cues acting on growth cones or that since we do not have a complete absence of Sema5A, the growth cones are merely slowed down by the low levels of Sema5A. Sema5A inactivation in mice results in embryonic lethality due to vascular instability (Fiore et al., 2005). The embryonic nervous system was largely normal in the knockout mouse although the possibility that subtle defects in a small subset of axons was not dismissed by the authors (Fiore et al.,

2005). Moreover, the defects observed in the mouse embryonic vasculature, much like the motor axon defects we see in our morphants, could be characterized as subtle, namely the slight decrease in complexity of branching of large blood vessels in null mutants versus wild type (Fiore et al., 2005).

Although Semaphorins are largely considered inhibitory cues, another transmembrane semaphorin, Sema1A has also been shown to function as an attractive cue to developing

113 peripheral neurons in the grasshopper limp bud (Wong et al., 1999). This Semaphorin only has the Sema domain so it is believed that a specific downstream signaling receptor/s is responsible for this attractive function for this Semaphorin (Wong et al.,

1999). For Sema5A however, the presence of the type-1 thrombospondin repeat domain was shown to be responsible for its ability to be both an attractive and a repulsive cue to rat habenular neurons (Kantor et al., 2004). Thrombospondin repeats are known to interact with different sulfated proteoglycans in the extracellular matrix and it was demonstrated that the presence of chondroitin sulfate proteoglycan (CSPG) versus heparan sulfate proteoglycan (HSPG) switches Sema5A from an inhibitory cue to an attractive one (Kantor et al., 2004). Another axon guidance cue, F-spondin which is an attractive cue that guides commissural axons to the floor plate in mouse (Burstyn-Cohen et al., 1999) also has the TSR domain and it was shown that the TSR domain is required for this attractive function.

By analyzing the Sema and TSR domains separately, we show that the bifunctionality of

Sema5A corresponds to the different functions of these two domains. That the thrombospondin domain alone rescues the delay phenotype of CaP axon growth cones in sema5A morphants but not the branching phenotype suggests that this domain is important specifically for promoting extension of the CaP axon into the ventral muscle.

On the other hand, the Sema domain alone rescues the branching phenotype in sema5A morphants as well as the the delay phenotype, though the latter is rescued to a much lesser extent. This suggests that the Sema domain is acting as a repulsive guidance cue in

114

the environment and serves to delimit the CaP axon to the pathway. The slight rescue of

the delay phenotype by the Sema domain could be attributed to the ability of this cue to

guide the growth cone at earlier choice points possibly the horizontal myoseptum. It was

previously shown that in embryos where expression of the repulsive semaphorin

Sema3a1 was knocked down by morpholinos, there was an observed delay of the CaP axon growth cones as it extends through the horizontal myoseptum (Sato-Maeda et al.,

2005). This suggests that the adding back of a repulsive cue, the Sema domain in our case, could possibly prevent prolonged pausing at this intermediate target and promote extension into the ventral muscle. The outcomes we observed agrees with what was determined in the habenular neuron culture where stripes that expressed the TSR-TM protein where permissive for axon growth while the stripes that expressed the Semacomp protein were repulsive to the axons (Kantor et al, 2004).

Interestingly, no axon defects were observed with overexpression of full-length Sema5A,

TSR-TM RNA or Sema domain alone by RNA injection (data not shown). This could possibly be due to a strict regulation of receptors expressed in the motor axon, that is, overexpression of a ligand should not have an effect if there are limited amounts of receptors that could detect it. The tight regulation of receptors expressed by the CaP axons have been observed with Neuropilin1a (npn1a) wherein npn1a is expressed in the cell body of the CaP axon while the axon traverses the common pathway (before reaching the horizontal myoseptum) and is then downregulated once the axon has passed the horizontal myoseptum (Sato-Maeda et al., 2005). This overexpression data also suggests

115

that TSR-TM is not acting as a strict attractant because if it is, then overexpression should

induce axon branching. It can therefore be argued that for the TSR-TM domain to act as

an attractant, it requires interaction with other factors that are specifically present in the

ventral muscle. Analysis of the Topped mutant supports this hypothesis. In these mutants

it was shown that when adding back wild type cells to Topped mutants, axons are able to

extend normally only where wild type cells are in the medial most portion of the ventral

muscle but not when wild type cells are even one cell body more lateral (Rodino-Klapac and Beattie 2004). These data suggest that the medial most muscle cells may contain the complement of proteins that can support ventral motor axons. In addition, the absence in present literature of a CaP axon mutant phenotype where axons invade more lateral regions of the ventral muscle is indicative of the presence of factors in this region, probably repulsive, that keep axons from extending into the muscle during this timepoint in development. Moreover, the CaP axon still extends medially even if Sema5A is already normally diffusely expressed suggesting that its function in promoting axon outgrowth is specifically regulated in these medial fast muscle cells.

Plexin A3: A Sema5A receptor?

Transmembrane Semaphorins such as Sema5A are believed to have the ability to signal

through Plexin receptors by themselves, without the requirement of neuropilin co-

receptors. Therefore, investigating potential interactions of Sema5A with Plexins is a way

to begin to define the zebrafish Sema5A receptor. Plexin A3 and Plexin B1 have both

116

been identified as important receptors for mediating repulsive Semaphorin cues, specifically Sema3a1 and Sema3a2 for Plexin A3 and Sema4D for Plexin B1 (Feldner et al., 2007, Ito et al., 2006). We have shown that though knockdown of two isoforms of

Plexin B1 affect CaP axon outgrowth, Sema5A does not appear to be involved. It was

previously shown that Plexin A3 knockdown, both by mutation and by morpholino yields

ectopic exit points as well as CaP axon branching, the latter being similar to that observed in Sema5A morphants (Feldner et al., 2007, Palaisa and Granato, 2007). That subthreshold co-injections of plexinA3 MO and sema5A MO shows a robust synergistic

(with synergism being defined as greater than simply additive) effect on the ectopic exit point phenotype compared to each MO on its own suggests that Sema5A could be a repulsive axon guidance cue that acts through the Plexin A3 receptor. It was shown

through similar means that both Sema3A1 and Sema3A2 act synergistically with Plexin

A3 to induce ectopic exit points (Feldner et al., 2007). This suggests that multiple repulsive cues may be responsible for defining exit points along the spinal cord.

Sema3A1 and Sema3A2 both have specific expression patterns in each somite during development (Shoji et al., 2003, Sato-Maeda et al., 2006, Feldner et al., 2007) whereas

Sema5A though more strongly expressed in the ventral muscle is still seen to be diffusedly expressed all through the developing trunk including the myotome region near the spinal cord ventral roots. Since knocking down Sema5A by itself induced only a slight increase in ectopic axonal exit points (as seen with the use of Sema5A MO2,

Figure 3.5), it indicates that other repulsive cues, most likely these other inhibitory

Semaphorins, are able to compensate for the knockdown of Sema5A. It was not

117

determined though whether Sema5A and Plexin A3 binds directly. However, it has been

shown that Plexin A1 which belongs to the same Plexin subfamily as Plexin A3 directly

binds a transmembrane semaphorin Sema 6D without the requirement of a neuropilin co-

receptor (Tamagnone et al., 1999, Toyofuku et al., 2004). This indicates that Sema5A,

which is also transmembrane, could bind Plexin A3 directly. The role of Sema5A in CaP

axon branching and extension into the myotome does not appear to be mediated through

Plexin A3. This suggests the existence of a yet unidentified Sema5A receptor/s that is present in the CaP motoneurons. In the study mentioned above,Sema6D was also found to bind another plexin receptor, Plexin A4, although weakly (Toyofuku et al., 2004). This indicates that it is likely that Sema5A may also be able to bind other Plexin receptors that could be responsible for regulating CaP axon extension into the myotome.

Sema5A may be modulated by proteoglycans in the myotome or CaP axon growth cones

It was further demonstrated that in the presence of chondroitin sulfate proteoglycans

(CSPGs) in the culture, Sema5A acts as an inhibitory molecule (Kantor et al., 2004).

Conversely, Sema5A mediates attraction in the presence of heparan sulfate proteoglycans

(HSPGs) in culture. Therefore, interaction of Sema5A with these two proteoglycans may be a mechanism through which the function of Sema5A is dictated. This interaction is believed to occur between the glycosaminoglycan (GAG) portions of the sulfated proteoglycans and the thrombospondin repeat domain of Sema5A (Adams and Tucker

2000). CSPG expression has been detected along the ventral portion of the motor axon

118 pathway during the period of primary motor axon pathfinding (Zhang et al., 2004) placing it in the right time and location to possibly interact with Sema5A. However, knock down of Chondroitin synthase I, an enzyme involved in CSPG synthesis, resulted in a decrease in immunoreactivity of CSPs but also did not cause motor axon defects in zebrafish (Zhang et al., 2004)) nor did low dose double knock down of Sema5A and

CSPG result in axon defects (JDH and CEB unpublished data). Thus, it may be that the

CSPGs observed in this study do not play a particular role in motor axon guidance since the antibody used in the study only reacts with Chondroitin Sulfate type A and type C and not Type B (Avnur & Geiger, 1984). HSPGs are also expressed diffusedly throughout the developing zebrafish embryo (Chen et al., 2005) and mutations in key enzymes in HSPG synthesis result in defects in both muscle and vascular development (Bink et al., 2003,

Chen et al., 2005). In addition, knocking down another HSPG, Agrin results in both truncations and branching of primary motor axons in addition to other morphological phenotypes (Kim et al., 2007). Another component of the extracellular matrix, Tenascin-

C, was previously shown to play a role in CaP axon guidance (Schweitzer et al., 2005).

This demonstrates that the extracellular matrix may be a rich source of guidance cues for the migrating growth cone and that understanding of interactions of these ECM components with other guidance molecules in the axons themselves or in the environment is important to fully understand the mechanisms that underlie axon guidance.

119

Figure 3.1 sema5A is expressed in the myotome at 18 and 24 hpf. RNA in situ hybridization shows expression of sema5A. (A, D, G) Sense and (B, E, H) anti-sense riboprobes were used. (C,F) Higher magnification views of trunk region of B and E. (G, H) 20 m cross-sections of in situ labeled 24 hpf embryos. Spinal cord (sc) and notochord (nc) are indicated. The myotome is outlined by dashed white line.

120

Figure 3.2 sema5b is expressed in the notochord. In situ hybridization with antisense sema5b riboprobe at 24 hpf, lateral view, arrow indicates notochord expression.

121

Figure 3.3 Morpholino knockdown of zebrafish Sema5A results in delay in CaP axon extension and axonal branching and is rescued by full-length rat Sema5A mRNA. (A-C) Lateral views of whole mount antibody labeling with znp1 of sema5A MO injected 26 hpf embryo (B, C) and uninjected WT sibling (A). Dashed white line indicate first intermediate target. Arrowheads indicate delayed axons. Arrow indicate an aberrant branch (D) RT-PCR showing efficacy of sema5A splice site MO in inducing aberrant splicing of sema5A mRNA in MO-injected embryos compared to uninjected wild type embryos. (E) Axon position was scored using different landmarks along the axon pathway in a dorsal to ventral direction. The landmarks on the x-axis are the horizontal myoseptum (HM), ventral edge of the notochord (VNC), proximal portion of the ventral muscle (PVM) and distal portion of the ventral muscle (DVM). (F) The axonal branching phenotype was scored by classifying branched axons into slight, moderate or severe branching. Data were quantified in sema5A MO injected embryos (n= 2020 axons, 101 embryos), sema5A MO and rat full-length sema5A RNA co-injected embryos (n= 2580, 129 embryos), WT uninjected (n=1640 axons, 89 embryos) and control MO injected embryos (n= 2360 axons, 118 embryos) with embryos obtained from at least three separate experiments. Error bars represent confidence interval for proportions at 95% confidence. Asterisks indicate significant difference between 18 ng sema5A MO-injected embryos and sema5a MO + rat sema5a RNA injected embryos at p<0.001 using One- way ANOVA.

122

Figure 3.4 CaP axon branching. Examples of different extents of branching observed in embryos (A) arrowhead indicates an axon with a slight branch characterized by short, thin extensions from the main axon trunk. (B) Arrowhead indicates axon displaying moderate branching characterized by either a longer extension from the main axon path into the surrounding myotome or a slightly thicker extension (C) arrowhead indicates a severely branched axon which is characterized by numerous and disordered branches.

123

Figure 3.5 plexin B1-A and plexin B1-B morpholino injection phenotypes and co- injections with sema5A. (A) Injection of 6 ng plexin B1-A MO shows an increase in branching CaP axons compared to control MO injected embryos. (B) Co-injection of 3 ng plexin B1-A MO and sema5A MO has little or no difference with injection of each MO by itself at 3 ng. (C) Injection 6 ng plexin B1-B MO results in a delay in CaP axon extension into the ventral myotome compared control MO injected embryos. (D) Co-injection of 3 ng of plexin B1-B MO and sema5A MO displayed no delay of CaP axon extension phenotype. Data quantified from 3 ng sema5A MO injected embryos (n=2020 axons, 101 embryos), 6 ng plexin B1-A MO injected embryos (n=1040 axons, 52 embryos), 3 ng plexin B1-A MO injected embryos (n=940 axons, 47 embryos), 3 ng sema5A MO + 3 ng plexin B1-A MO injected embryos (n=2140 axons, 107 embryos), 6 ng plxnB1B MO injected embryos (n=1960 axons, 98 embryos), 3 ng plexin B1-B MO injected embryos (n=1080 axons, 54 embryos), 3 ng sema5A MO + 3 ng plexin B1-B MO injected embryos (n=1180 axons, 59 embryos). Error bars represent confidence interval for proportions at 95% confidence. Asterisks indicate significant difference between control MO injected and plexin B1-A or plexin B1-B MO injected with p<0.001 using One-way ANOVA.

124

125

Figure 3.6 Additive and synergistic interactions of Sema5A and Plexin A3. A) Axonal phenotypes observed in embryos injected with plexin A3 MO alone or plexin A3 MO and sema5A MO. Arrows indicate moderately and severely branched CaP axons, arrowheads indicate axons exiting the spinal cord at ectopic exit points. B) Percentage of axons with moderate and severe branching in embryos injected with 3 ng of sema5A MO alone, 1 ng of plexin A3 MO alone and a combination of 3 ng sema5A MO and 1 ng plexin A3 MO. C) Percentage of hemisegments observed with ectopic axonal exit points in embryos injected with 3 ng sema5A MO, 1 ng plexin A3 MO and a combination of 3 ng sema5A MO and 1 ng plexin A3 MO. Data quantitated from 3 ng sema5A MO injections (n= 860 axons, 43 embryos), 1 ng plexin A3 MO injection (n=2860 axons, 140 embryos) and 3 ng sema5A MO + 1ng plexin A3 MO injections (n=2760 axons, 128 embryos) Ectopic axonal exit points are not observed in both wild-type uninjected or control MO-injected embryos (not shown). Error bars represent confidence interval for proportions at 95% confidence. Asterisk indicate significant difference between Sema5A MO + PlexinA3 MO injected and both 3ng Sema5A MO injected and PlexinA3 MO injected at p<0.01 using a Z-test of proportions.

126

Figure 3.7 Rescue of sema5a knockdown embryos with rat sema5a thrombospondin domain and rat sema5a sema domain Growth cone position (A, C) and extent of branching (B,D) were scored in sema5a MO- injected embryos (MO-injected), sema5a MO and rat thrombospondin repeat domain RNA co-injected embryos (MO + TSP-TM; A,B), sema5a MO and rat sema domain RNA co-injected embryos (MO + semacomp; C,D) and uninjected embryos (WT uninjected).Data were quantitated in sema5a MO injected embryos (n= 2020 axons, 101 embryos), sema5a MO and TSR-TM RNA co-injected embryos (n= 2300 axons, 115 embryos), sema5a MO and semacomp RNA co-injected embryos (n=2140 axons, 107 embryos),WT uninjected (n=1640 axons, 89 embryos) and control morpholino-injected (n=2360, 118 embryos). Asterisks indicate significant rescue by co-injected RNA at 95% confidence

127

128

Genotype n % stalled % branched axons (# of axons) axons Sema5a MO- 2020 20% ± 1.7% 27% ± 1.9% injected Sema5a MO 2580 4% ± 0.8% 4% ± 0.8% +FL sema5a RNA Sema5a MO+ 2300 6% ± 1.0% 22% ± 1.7% TSP-TM RNA Sema5a MO + 2140 12% ± 1.4% 7% ± 1.1% semacomp RNA

Table 3.1. Summary of Data from Sema5a RNA rescue experiments 18 ng of sema5a MO and 500 pg of RNA were used in each experiment. n represents total number of axons observed over three experiments. Percent stalled axons refers to axons stalled at or before reaching the ventral edge of the notochord (VNC). Percent branched axons refers to all classifications of observed branched axons. Confidence intervals calculated at 95% confidence.

129

Chapter 4: Topped candidate gene search3

Introduction

During development motoneurons that are positioned in the central nervous system need to make connections to muscle targets in the periphery. The axons sent out by these neurons may travel over relatively long distances and are observed to consistently follow stereotyped pathways. Growth cones at the tips of these axons are responsible for steering the axon to its destination by its ability to respond to cues from the environment. That the proper connections be established during development between motor axons and muscle cells is key to an organism’s ability to control movement.

The zebrafish is an ideal organism for studying motor axon guidance. Zebrafish embryos develop outside the mother and are optically transparent. This provides easy access to the motoneurons as they develop. Zebrafish also generate a large number of progeny which makes them amenable to large screens. Also, the embryonic neuromuscular system is relatively simple and has been well-characterized (Eisen et al., 1986; Myers et al., 1986).

Several mutagenesis screens have been performed with the goal of identifying genes involved in axonal pathfinding (Haffter et al 1996; Kalstrom et al., 1997; Beattie et al.,

3 All experiments in this chapter were completed by JD Hilario except for mapping topped which was performed by L Rodino-Klapac. 130

1999; Panzer et al., 2005). These screens have been discussed in detail in the first

chapter. One of the mutants isolated from the antibody screen is the topped mutant

(Beattie et al., 1999). topped is an ENU-induced, autosomal recessive, partially

homozygous viable mutation. It was identified through its distinctive CaP axon

phenotype. In wild type embryos at 26 hpf, CaP axon have already extended into the

ventral muscle. In topped mutants, CaP axons are stalled at or near the first intermediate target, the horizontal myoseptum at 26 hpf (Rodino-Klapac et al., 2004). The CaP axons then recover and topped mutants are indistinguishable from wild types by ~ 36 hpf which suggests that the mutation causes a delay in CaP axon extension into the ventral muscle.

Interestingly, the other two primary motor axons MiP and RoP are not affected. By generating chimeras between wild-type and topped mutant embryos, it was determined that Topped functions in a non-cell autonomous manner relative to the motoneuron since transplanted wild-type muscle cells in topped embryos were able to rescue the delay of

CaP axon extension. Further, a critical subset of cells, specifically ventromedial muscle cells where Topped is required was identified. Together these findings suggest that

Topped may be a cue in the ventral muscle that acts as an attractant specifically to the

CaP axon growth cone.

To identify the topped gene, positional cloning was undertaken. The region was narrowed down to a 5.2 Mb region between zebrafish markers z9321 and z3399. This region contains a number of protein coding sequences. Thus, we undertook a candidate gene approach for identifying topped. The first gene we investigated was Semaphorin 5A

131

(Sema5A). Sema5a is a member of the semaphorin family of axon guidance cues which

are characterized by the presence of the sema domain. The semaphorin family is discussed in detail in Chapter 1 of this manuscript. Sema5A is a transmembrane semaphorin that in addition to the sema domain, has an extracellular thrombospondin

(TSR) domain. Sema5A has previously been shown to act as a bifunctional axon

guidance cue in the rat habenula (Kantor et al., 2004). This bifunctionality was attributed

to modulation of growth cone response to Sema5A by different proteoglycans.

The second candidate gene we investigated was the Contactin associated protein-like 2

(CNTNAP2) gene. CNTNAP2 is a member of the neurexin family which are known cell

adhesion molecules with known functions in the vertebrate nervous system. Like other neurexins, CNTNAP2 contains epidermal growth factor (EGF) repeats and laminin G domains. Domains unique to CNTNAP2 include an F5/8 type C domain, discoidin/neuropilin-fibrinogen-like domains, thrombospondin N-terminal-like domains and a PDZ binding site. Caspr2 is known to function in paranodes of myelinated axons to regulate localization of K+ channels (Peles and Selzer 2000). At the end of the study our

experiments indicated that although both of these genes are present in the geneneral region where topped maps, neither sema5A nor contactin are the topped gene.

132

Methods

Fish Strains and Maintenance

AB* embryos, ABLF embryos and topped458 embryos were used for morpholino and

RNA injections and were maintained between 25.5 and 28.5°C. Embryos were staged by converting the number of somites to hours post fertilization (hpf; Kimmel et al., 1995).

Rescue experiments were performed in topped458 homozygous embryos which were generated by natural breeding of topped458 heterozygous or homozygous fish.

Morpholino analysis

For antisense oligonucleotide morpholino mediated knockdown of Sema5A, two splice blocking morpholinos were designed to the splice donor site of exon 3 (sema5A MO1:

CTTCTTTACTTACACATTACTGGTG) and splice donor site of exon 2 (sema5A MO2

:CCTGAAGAGATGATTTCTAAAGGA). 18 ng of MO were injected into two to four cell-stage wild-type embryos. Wild-type embryos were allowed to develop to 26 hpf, and subsequently stained with znp1 or used to isolate total RNA for RT-PCR analysis. motor axons were analyzed with a Zeiss axioplan microscope. A control MO that does not have any known target in zebrafish was also used (ctl MO:

CCTCTTACCTCAGTTACAATTTATA ). To test the efficiency of the splice blocking

MOs, RT-PCR was performed using the One-Step RT-PCR kit (Qiagen). sema5A transcript specific primers flanking the targeted MO site were used to amplify an 1100 bp fragment in uninjected controls and an 850 bp fragment in MO injected embryos.

133

RNA rescue

For heterologous RNA rescue, the full-length rat sema5A gene (Kantor et al., 2004) was

cloned into the PCS2 vector. Capped polyA mRNA was transcribed using the mMessage

mMachine (Ambion) SP6 kit and injected at the one to four cell stage into topped458 zebrafish embryos at 500 pg doses.

BAC rescue

BAC CH211-245H5 was obtained from BacPac (CHORI). Cultures were grown in

Chloramphenicol LB plates (100 mg/ml). BAC was isolated from cultures using Qiagen

Midi Prep Kit. 100 pg of isolated BAC was injected into one to four cell stage wild type embryos. Embryos were fixed at 26 hpf and processed for znp-1 labeling.

Whole mount antibody labeling

Whole mount antibody labeling was performed as described in Eisen et al., (1989) and

Beattie et al., (2000). The znp1 monoclonal antibody that recognizes primary and secondary motor axons (Trevarrow et al., 1990; Melancon et al., 1997) was detected using the Sternberger (now Clonetech, right?) Clonal-PAP system with diaminobenzidine

(DAB) as a substrate (Beattie and Eisen, 1997). Znp1 recognizes synaptotagmin II (Fox and Sanes 2007) and is a good antibody for visualizing early developing motor axons.

Embryos were analyzed with a Zeiss axioplan microscope. CaP axons in segments 5-15 on both sides of the embryo were analyzed.

134

RNA in situ hybridization

Whole-mount RNA in situ hybridization was performed as described by Thisse et al.

(1993). Zebrafish cntnap2 was partially cloned from zebrafish by RT-PCR using predicted sequence from Ensembl (http://uswest.ensembl.org/Danio_rerio/). The region

between 136 -398 bp of the transcript containing unique sequences was cloned. This RT-

PCR product was then TOPO-TA cloned to PCRII-Topo (Invitrogen). An antisense

digoxigenin zebrafish cntnap2 riboprobe was synthesized from this plasmid by linearizin

with HindIII and transcribed with T7.

Results

The Sema5A gene

The critical region of topped was obtained by positional cloning (Rodino-Klapac, 2005).

Two close markers were identified, z9321 and z3399 in chromosome 24. A list of

predicted annotated genes are listed in Table 4.1. Sema5A maps to the most 5’ area of

this region. The cloning of the sema5A gene is discussed in-depth in Chapter 3 of this

manuscript. sema5A was identified as a candidate gene since it is found between the two

closest identified marker for topped, markers z9321 and z3399 in Chromosome 24

(Rodino-Klapac 2005). sema5A is annotated in Sanger Zebrafish ensemble. It spans 21

exons, covers 200 Kb of the genome and conceptual translation predicts a 1046 amino

acid protein.

135

Injection of Sema5A RNA may rescue topped mutant phenotype

One way to determine if a gene is disrupted in a mutant is to see if adding back the

protein encoded by the gene will rescue the mutant phenotype. We obtained the mouse

Sema5A complete cDNA for this reason. We were able to show that the mouse Sema5A

RNA is functional by showing that it rescued the Sema5A morphant phenotype (in

Chapter 3 of this manuscript and Hilario et al., 2009). We then injected 500 pg mouse

Sema5A RNA into topped458 homozygous embryos. There was a significant difference in

the percentage of axons stalled at the horizontal myoseptum between RNA injected and

uninjected topped458 mutants (Figure 4.1). 74 ± 3% (n = 980 axons) of CaP axons were

observed to be stalled at the horizontal myoseptum in topped458 mutants whereas 64 ± 3%

(n = 800 axons) were stalled in sema5A RNA injected topped458 mutants (p-value <

0.0001). Also 14% of topped458 mutant embryos had axons extending into the ventral

muscle compared to 30% of Sema5A RNA injected topped458 mutants (p-value < 0.0001).

This data suggests that Sema5A rat RNA induces extension of CaP axons in topped

458mutants. This could mean that Sema5A is able to rescue the mutation or it could also

be that Sema5a may be acting as an attractant to the CaP growth cone independent of

Topped.

Knockdown of Sema5A has a different phenotype than topped

If Sema5A is the gene that is mutated in topped458 mutants, we would also expect that

knocking it down would have a similar phenotype as topped458. The splice-site morpholino against Sema5A was described in Chapter 2 of this manuscript. Knocking

136

down Sema5A yields two phenotypes, short axons and branching axons (Figure 4.2).

This is markedly different than the observed topped phenotype which is just axons stalled

at the horizontal myoseptum. There was also a significant difference in the number of

axons exhibiting these phenotypes in topped458 embryos versus Sema5A MO-injected embryos (Table 4.2). This data shows that knocking down Sema5A yields a different phenotype than the topped mutation. Unlike the result from the RNA rescue, this result is not consistent with sema5A being topped.

Knockdown of Sema5A in topped mutants results in an additive phenotype

We next asked whether knocking down Sema5A further in topped458 mutants would

exacerbate the observed phenotype that is, further delay CaP axon extension. For these

experiments we injected 18 ng sema5A MO into topped458 mutants and observed CaP

axons at 30 hpf. We chose this timepoint since in topped mutants the CaP axon starts to

extend past the horizontal myoseptum at around 29 hpf (Rodino-Klapac and Beattie

2000) so at 30 hpf we could see if knocking down Sema5A would exacerbate this

pheonotype i.e., further delay extension. This timepoint also allows the axons to extend longer such that any potential branching can be more easily observed. We did not see a significant difference between MO-injected and uninjected topped embryos in percent of

CaP axons extending to the ventral muscle. In uninjected topped mutants 91 ± 3% (n=480 axons) of CaP axons have not yet reached the ventral muscle while in sema5A MO- injected mutants 90 ± 3% CaP axons have not yet reached the ventral muscle. However injecting sema5A MO increases the number of branching axons in mutants. We saw 30 ±

137

4% of axons exhibited slight to severe branching in MO-injected topped458 mutant embryos compared to 15 ± 3% of axons in uninjected topped458 mutant embryos. There is

a particularly significant increase in moderately branched axons with MO injection

(Figure 4.3). This data disagrees with the RNA rescue result and indicate that Sema5A and Topped may be acting in different pathways.

cntnap2 is expressed in developing somites

Since Sema5A does not appear to be Topped, we then moved on to investigating a

different gene that is also found in the region were the topped mutation maps. The full

cntnap2 gene was found in this region of interest whereas sema5A after later genome annotation(http://uswest.ensembl.org/Danio_rerio/Location/View?h=Z9321;r=24:151904

32-18390569), only had parts of the gene in this region. To determine whether cntnap2 is

a good candidate for topped, we first looked at its expression pattern at the timepoint

when the topped phenotype is observe by doing RNA in situ hybridization in 26 hpf embryos. We found that cntnap2 is strongly expressed in the developing somites of the

trunk (Figure 4.3). There does not appear to be a difference in level of expression

between the dorsal and ventral muscle.

BAC containing cntnap2 does not rescue topped mutant

To test whether cntnap2 is topped, we decided to use BAC CH211-245H5 (from CHORI)

that encompasses the region where CNTNAP2 maps. We injected 100 pg of the purified

BAC into topped458 homozygous mutant embryos. The injected embryos were then fixed

138 at 26 hpf and labeled with znp-1 antibody. Embryos that had CaP axons past the horizontal myoseptum were counted and labeled ‘rescue’. Uninjected embryos were used as controls. 24 ± 8 % of BAC injected embryos had ‘rescued’ axons (24/99 embryos) while 12 ± 8% of unjected embryos had ‘rescued’ axons (9/73 embryos). This difference is not statistically significant. It is however difficult to rule out that cntnap2 is not topped since BAC injections are known to rescue mutants but to a small degree (Yan et al.,

1998).

Discussion

The identification of specific molecules that guide motor axons to their final targets has been a main objective in the field of neuroscience. Here we document the effort to uncover the identity of the topped gene. Topped was identified as an important cue for

CaP motor axons to be able to extend intothe ventral myotome. Topped acts very specifically on CaP axons and is required in a specific set of ventromedial cells in the developing muscle (Rodino-Klapac et al., 2004). Here we investigate two topped candidate genes sema5A and cntnap2.

Is Sema5A the topped gene? sema5A was initially identified as a candidate gene for topped because it was found in the

300 MB region where topped maps. sema5A was an obvious choice since it is a member of the semaphorin family of axon guidance molecules. Injecting rat Sema5A mRNA into topped458mutants resulted in slight and statistically significant rescue of the topped 139

phenotype. This indicated that sema5A could be the topped gene. However, further

experiments did not support this data. For example, knocking down Sema5A using MOs

has a distinctly different phenotype than the mutant. In topped458 mutants, the vast majority, if not all CaP axons are stalled at or near the horizontal myoseptum at 26 hpf.

Knocking down Sema5A induced stalling in a much smaller percentage of CaP axons and also caused CaP axon branching; a defect not seen in topped mutants. A second set of experiments also did not support that sema5A was topped. When MOs against sema5A

were injected into topped mutants, we observed a significant increase in CaP axon

branching (Figure 4.2) but no increase in stalled axons. Lastly, we have also sequenced

sema5A in both topped and wild-type embryos and were unable to find any base pair

differences between the two. Taken together, these data suggest that while Sema5A may

function to guide CaP axons, it is likely not the topped gene.

Is cntnap2 topped?

Another gene that is found in the region where topped maps is cntnap2. cntnap2 (also

known as caspr2) is a member of the neurexin superfamily which are known to mediate cell-cell interactions in the nervous system. The known function of CNTNAP2 in the mammalian nervous system is in local differentiation of axons into subdomains and is observed to colocalize with potassium channels at the juxtaparanodes of myelinated axons (Poliak et al., 1999). CNTNAP2 is known to interact with contactin, a GPI- anchored adhesion molecule that is a member of the immunoglobulin superfamily (Oiso et al., 2009). Contactin has known roles in axon outgrowth and guidance in mammalian

140

systems (Haenisch et al., 2004). Contactin 1 (Cntn1) expression has also been observed in

trunk motoneurons in zebrafish between 18 – 24 hpf (Gimnopoulus et al., 2002). The in situ expression of cntnap2 we observed was promising since it appeared to be expressed widely in the myotome during the time of axon outgrowth (Figure 4.4). However, injection of a BAC that contained the gene was not able to significantly rescue the topped phenotype. This data maybe considered preliminary though since BAC injections are not very reliable in being able to rescue mutations. It would be ideal if we could use zebrafish cntnap2 cDNA and use RNA generated from that to do the mutant rescue to fully conclude whether cntnap2 is or is not the topped gene. By cloning the entire cntnap2 gene we could also detect if there are any base pair differences in this gene between topped mutants and wild types.

Conclusion

Studying the topped mutation could potentially reveal a cue that attracts or promotes

extension of the CaP axon towards its muscle target, the ventral muscle. We have

investigated two candidate genes for topped, sema5A and cntnap2. We believe we have

eliminated Sema5A as a candidate gene for topped. However, cntnap2 still needs to be studied more, particularly the whole gene needs to be cloned and tested for ability to rescue topped. There are also more than 40 annotated genes plus more unidentified protein coding sequences in the topped region that could be investigated. Finally, finer

mapping of topped can be done to narrow down the known region.

141

P<0.0001

P<0.001 *

Figure 4.1 Injection of Sema5A rat RNA into topped458 embryos rescues delay of CaP axon extension. Axon position was scored using different landmarks along the axon pathway in a dorsal to ventral direction. The landmarks on the x-axis are the horizontal myoseptum (HM), ventral edge of the notochord (VNC), proximal portion of the ventral muscle (PVM) and distal portion of the ventral muscle (DVM). Data was quantified from topped458 embryos (n=980 axons, 49 embryos) and topped458 + Sema5A MO (n = 800 axons, 40 embryos). Error bars represent confidence interval for proportions at 95% confidence. Asterisks indicate significant difference between topped458 uninjected and topped458 RNA injected with indicated p-values.

142

Figure 4.2 topped and sema5A MO knockdown phenotypes are different at 26 hpf. (A) Lateral views of whole mount antibody labeling with znp1 topped458 homozygous mutant 26 hpf embryo. (B-C) Wild-type embryos injected with 18 ng Sema5A and (D) Wild type uninjected embryo. Sema5A MO injected embryos exhibit delayed CaP axon extension (arrowhead) and axonal branching (arrow). Dashed white line indicate first intermediate target.

143

Figure 4.3 sema5A MO induces branching of CaP axons in 30 hpf topped mutants. The axonal branching phenotype was scored by classifying branched axons into slight, moderate or severe branching. Data was obtained from topped458 homozygous mutant uninjected embryos (n = 480, 24 embryos) and topped458 injected with 18 ng sema5A MO (n=480, 24 axons).

144

Figure 4.4 Branching axons observed in sema5A MO-injected 30 hpf topped mutants. Lateral views of whole mount antibody labeling of znp1 (A) topped458 homozygous mutant 30 hpf and (B-C) topped458 homozygous mutant injected with 18 ng sema5A MO. Dashed white line indicate first intermediate target. Arrow heads indicate observed branched axons observed in sema5A MO-injected mutants.

145

Figure 4.5 Expression of cntnap2 in 24 hpf embryos. (A) A riboprobe designed against the first 200 bp of the CNTNAP2 mRNA reveals expression in the developing muscle. (B) Detail of expression in the mid-trunk hemisegments.

146

Table 4.1 Annotated predicted genes in topped critical region. Annotated genes in the region between markers z9321 and z3399, the closest mapping markers to topped. Data from search conducted in Sanger Ensembl (http://uswest.ensembl.org/Danio_rerio/).

147

Ensembl genes Possible function/known features (if known) z9321 Marker Sema5A With semaphorin domain Kelch-like protein 15 eukaryotic translation initiation factor 2 prostaglandin H2 D-isomerase complement component 8, gamma polypeptide Methionine-R-sulfoxide reductase B2, myeloid/lymphoid or mixed-lineage leukemia (trithorax homolog, Drosophila) Armadillo repeat-containing protein 3 Expressed in many cancer tissues and (Beta-catenin-like protein)(KU-CT- cell lines. 1)(Cancer/testis antigen 81)(CT81) sperm associated antigen 6 COMM domain containing 3 Polycystic kidney disease protein 1-like 1 May have a role in the heart and in the (Polycystin-1L1)(PC1-like 1 protein) male reproductive system. Polycomb complex protein BMI-1-A (Polycomb group Ring finger protein 4-A) Histone-lysine N-methyltransferase EZH2 Catalytic subunit of the prc2/eed-ezh2 complex, which methylates 'Lys-9' and 'Lys-27' of histone H3, leading to transcriptional repression of the affected target gene cullin 1-like CNTNAP2 May play a role in the formation of functional distinct domains critical for saltatory conduction of nerve impulses in myelinated nerve fibers. Seems to demarcate the juxtaparanodal region of the axo-glial junction. thiamin pyrophosphokinase 1 Sec61 gamma subunit LisH domain and HEAT repeat- containing protein K1AA1468 homolog Centrosome and spindle pole-associated May play a role in cell-cycle-dependent protein 1 microtubule organization

148 leucine zipper transcription factor-like 1 Serine/threonine-protein kinase VRK1 Serine/threonine kinase that phosphorylates 'Thr-18' of p53/TP53 and may thereby prevent the interaction between p53/TP53 and MDM2 Metal transporter CNNM3 (Cyclin- Probable metal transporter M3)(Ancient conserved domain-containing protein 3 tubulin, alpha 7 like Carboxypeptidase A6 Precursor Release of a C-terminal amino acid, but little or no action with -Asp, -Glu, -Arg, -Lys or -Pro. sulfatase 1 oxidative-stress responsive 1b myeloid differentiation primary response gene 88 Cryptochrome DASH (Protein CRY- May have a photoreceptor function DASH)(zCRY-DASH) acetyl-Coenzyme A acyltransferase 1 [ phospholipase C, delta 1a small CTD phosphatase 3-like Nktr protein Fragment vasoactive intestinal peptide receptor 1 cystatin B Coiled-coil domain-containing protein 58 UPF0389 protein FAM162B dopamine receptor D3 z3399 Marker

149

Genotype % axons at % branched # of axons

HM axons

Topped458 74 ± 3% 5 ± 1% 980

sema5A MO-injected 2 ± 1% 29 ± 1% 2020

wild type

Table 4.2 Comparison of phenotypes of topped458 homozygous mutant embryos and sema5A MO knockdown embryos. Percent of axons observed exhibiting delayed axons phenotype (stalled at horizontal myoseptum,HM) and branching axons.

150

Chapter 5: Conclusions

Our understanding of axon pathfinding has increased exponentially in just the past three decades. Several factors have contributed to this including advances in biochemistry, molecular biology, cell culture and in vivo techniques. Studies using the zebrafish have contributed significantly to the field of axon guidance by both the identification of molecules involved in axon guidance as well as by facilitating the elucidation of the mechanisms of axon guidance. As was discussed here, the identification of two zebrafish mutants that specifically affect pathfinding of primary motor axons has directly (and indirectly) led to the identification of two molecules involved in primary motor axon guidance and has also given us insight into how these cues may be interacting with other known axon guidance molecules and what mechanisms may be involved.

stumpy mutant

The phenotype of the stumpy mutant revealed several novel aspects of zebrafish motoneuron pathfinding. It identified previously unobserved intermediate targets along the pathway of the motor axon and it also demonstrated that one specific molecule i.e., the stumpy gene product is present at each intermediate target and is important for the growth cone to progress past each intermediate target (Beattie et al., 2000). It also showed that all three primary motoneurons are responsive to the effect of this specific

151 molecule. The subsequent identification of the stumpy gene as ColXIX was at first unexpected since collagens are more viewed as inert ECM molecules than molecules actively involved in cellular processes such as axon guidance. However ColXIX appears to possess unique characteristics that give it functions other than contributing to tissue stability. It is becoming more apparent that collagens, especially non-fibrillar collagens, function as more than just structural molecules. The identification of ColXIX as a molecule important for motor axon guidance is one more indication of this fact. What precisely ColXIX is doing at intermediate targets we can only still speculate about. We do know that its localization and time of expression is important. We can also surmise that the ability of its LamG/TSPN amino domain to bind other molecules, such as those containing heparin domains, may be important for its role here. HSPGs which bear heparin domains have been shown to affect axon guidance in zebrafish which makes it a likely interacting partner for ColXIX at intermediate targets (Kim et al., 2007). However we still cannot discount that its collagenous domains may also be recognized by other collagen binding proteins like integrins. Injection of integrin antagonists have also been shown to affect outgrowth of ventral motor axons (Becker et al., 2003). Either way, in this case ColXIX is directly affecting a very specific step in motor axon guidance specifically navigation of intermediate targets by growth cones.

topped mutant

By studying the topped mutant, we were also able to identify Sema5A as having an important role in guiding CaP axons. Serendipitously, even if Sema5A turned out not to 152 be the topped gene, it still appeared to be involved in guiding CaP axons. The knockdown phenotype of Sema5A was similar to the topped phenotype but not identical, with

Sema5A knockdown inducing a delay in CaP axon extension like in topped mutants but also causing axon branching. By overexpressing the TSR domain and the sema domain separately, we were able to tease out the separate functions of each in promoting axon extension to the ventral muscle and preventing axon branching respectively. This finding is consistent with Sema5A being a bifunctional axon guidance cue which has already been observed in the rat habenula (Kantor et al., 2004). We also showed that Sema5A may be acting through the PlexinA3 receptor which has previously been shown to be expressed in CaP motoneurons (Feldner et al., 2007).

Modularity of axon guidance cues

An interesting characteristic that most axon guidance molecules share is modularity (for example Seeger and Beattie 1999). More often than not axon guidance molecules are composed of discrete domains or modules that fold independently and presumably perform specific functions. This characteristic would be advantageous for axon guidance molecules whose specific function is promoting interactions between the growth cone and a cell surface or the ECM since such a molecule would essentially be recognizing both the growth cone and the surface of attachment. This modularity also enables these molecules to be multifunctional as is the case with Sema5A.

153

Collagens are known modular proteins (Bork 1999). Their triple-helical Col domains are usually surrounded by NC domains that have properties of modular domains. ColXIX has a very distinct amino-terminus NC domain that is similar to the LamG/TSPN domain. In this study, we show that this domain is probably important for the function of ColXIX in axon guidance presumably by its ability to bind heparins and heparin-containing molecules. This may have implications in understanding other known functions of

ColXIX in muscle development (Sumiyoshi et al. 2004) and hippocampal synaptogenesis in mice (Su et al., 2010).

Semaphorins are modular proteins which have the semaphorin domain in common. Each class of semaphorin is characterized by the presence of different domains in addition to the sema domain. Class 5 semaphorins of which Sema5A is a part of are characterized by the presence of the TSR domain. The combination of the sema and the TSR domain in one molecule is interesting since the sema domain is a known repulsive cue while the

TSR domain is present in a number of attractive axon guidance cues. Studies in culture and in the rat brain (Kantor et al., 2004) previously indicated that these domains separately have opposing effects on axons in vitro. Our work demonstrated that this function of Sema5A is relevant in vivo and that it also operates in pathfinding of motor axons in the periphery.

154

Combining and modulating axon guidance cues

Sema5A is the third semaphorin shown to influence CaP axon pathfinding along with

Sema3a1 and Sema3a2 (Roos et al., 1999, Sato-Maeda et al., 2005). This apparent redundancy is probably important for fine-tuning pathfinding by the CaP axon. Since axons follow a precisely-defined path to their targets there is an advantage to having multiple cues ‘label’ the specific pathway. Also semaphorins would be ideal for such redundant expression since different semaphorins are known to bind to the same receptor

(Feldner et al., 2005). Hence there is no need for a different receptor per ligand presented.

However there are two semaphorin receptors that have been shown to be expressed in the

CaP motoneuron, Plexin A3 and Neuropilin 1A (Feldner et al., 2005, Feldner et al., 2007,

Palaisa et al., 2007). Class 3 semaphorins which are secreted are known to bind both plexins and neuropilins, possibly as a receptor complex. Transmembrane semaphorins however are known to bind plexins directly (Negishi et al., 2005). Since both sema receptors are expressed by CaP motoneurons, CaP growth cones could potentially respond to all three semaphorins present in the environment which may be how CaP axon pathfinding is fine-tuned.

The ability of Sema5A to be influenced by proteoglycans adds another level of modulation that may contribute to the precision of the pathfinding of the axon. The TSR domain has the ability to bind proteoglycans. In the rat habenula, it was found that via

TSR interaction with either CSPGs or HSPGs, Sema5a can become a repulsive or an attractive cue respectively (Kantor et al., 2005). This ability to switch from being an

155

attractive to a repulsive cue depending on the presence of an ECM molecule is probably a common mechanism for TSR-containing molecules.

It is interesting that ColXIX also contains a thrombospondin domain in its amino- terminus domain. This indicates that ColXIX may function in regulating the localization of proteoglycans in the ECM. This may explain why the phenotype of Sema5A knockdown and ColXIX knockdown have similiarities though they are not identical. The

LamG/TSPN domain of ColXIX binds specifically to heparins so we can speculate that

ColXIX influences HSPGs specifically. The localization of specific proteoglycans by basement membrane molecules may be another way to modulate cues in axon guidance.

Understanding the molecules involved and mechanisms utilized in guiding an axon to its target has implications in understanding neurodevelopmental disorders. In a linkage and association genome-wide study, Sema5A was implicated as a susceptibility loci for autism and was seen to be reduced in the brains of autism patients (Weiss et al., 2009).

We would expect other genes that are involved in fine-tuning the wiring of the nervous system to also be involved in similar developmental diseases. Therefore understanding these molecules and the pathways they are involved in may potentially contribute to development of therapies and treatments for these disorders.

156

Appendix A: Supplementary Figures

Figure A1 Sema5A MO excises out exon 4. Alignment of sema5a cDNA generated from total RNA from wild-type uninjected embryos and sema5a MO-injected embryo showing loss of exon 4 in morphants.

157

WTsema5a GAGGCAGAATGGCATTGTGATGAGTTTACCAAAGGAGCCTGTTTCAGTCGTGGGAAATCA 60

Sema5aMO GAGGCAGAATGGCATTGTGATGAGTTTACCAAAGGAGCCTGTTTCAGTCGTGGGAAATCA 60

************************************************************

WTsema5a GAGGAGGAGTGCCAGAATTACATCCGTGTCCTCCTCGTCAATGGTGACAGGCTGTTTACC 120

Sema5aMO GAGGAGGAGTGCCAGAATTACATCCGTGTCCTCCTCGTCAATGGTGACAGGCTGTTTACC 120

************************************************************

WTsema5a TGCGGAACAAATGCATTCACTCCTATTTGCACCAATCGCACGCTGACTAACCTGACTGAG 180

Sema5aMO TGCGGAACAAATGCATTCACTCCTATTTGCACCAATCGCACGCTGACTAACCTGACTGAG 180

************************************************************

WTsema5a GTCCATGATCAAATCAGTGGGATGGCACGGTGCCCCTATAACCCTCTGCATAATTCCACC 240

Sema5aMO GTCCATGATCAAATCAGTGGGATGGCACGGTGCCCCTATAACCCTCTGCATAATTCCACC 240

************************************************************

WTsema5a GCCCTCATCACTTCCAGTGGAGAACTGTATGCTGCAACTGCAATGGACTTTTCAGGCAGA 300

Sema5aMO GCCCTCATCACTTCCAGTGGAGAACTGTATGCTGCAACTGCAATGGACTTTTCAGGCAGA 300

************************************************************

WTsema5a GACCCAGCCATCTACCGCAGCTTGGGAGGGCTTCCACCTCTGCGTACTGCTCAGTACAAC 360

Sema5aMO GACCCAGCCATCTACCGCAGCTTGGGAGGGCTTCCACCTCTGCGTACTGCTCAGTACAAC 360

************************************************************

WTsema5a TCCAAATGGCTCAATGAGCCCAACTTCGTCTCCTCCTATGACATCGGCAACTTCACGTAC 420

Sema5aMO TCCAAATGGCTCAATG------376

****************

WTsema5a TTCTTCTTCCGTGAAAATGCTGTGGAGCACGACTGCGGCAGGACTGTTTTCTCTCGGGCT 480

Sema5aMO ------

WTsema5a GCCCGCGTCTGCAAGAATGATATCGGGGGCCGTTTCCTTCTGGAGGACACCTGGACTACC 540

Sema5aMO ------

WTsema5a TTTATGAAAGCCCGGCTCAACTGCTCACGGCCTGGCGAGATCCCATTCAACTACAATGAG 600

Sema5aMO ------

WTsema5a TTGCAGGGAACCTTCTTTCTGCCTGAGCTCGAGCTCCTCTATGGGATTTTCACCACTAAT 660

Sema5aMO ------

158

WTsema5a GTTAACAGTATTGCGGCCTCAGCAGTGTGTGCCTTCAATCTGAGCGCTATTACCCAAGTG 720

Sema5aMO --TAACAGTATTGCGGCCTCAGCAGTGTGTGCCTTCAATCTGAGCGCTATTACCCAAGTG 434

**********************************************************

WTsema5a TTCAGCGGCCCTTTCAAGTACCAAGAGAACTCGCGCTCTGCTTGGCTTCCTTACCCCAAT 780

Sema5aMO TTCAGCGGCCCTTTCAAGTACCAAGAGAACTCGCGCTCTGCTTGGCTTCCTTACCCCAAT 494

************************************************************

WTsema5a CCTAACCCCGACTTCCAGTGTGGTACTATAGATTTTGGCTCGTATGTGAACTTAACGGAG 840

Sema5aMO CCTAACCCCGACTTCCAGTGTGGTACTATAGATTTTGGCTCGTATGTGAACTTAACGGAG 554

************************************************************

WTsema5a AGGAATCTGCAGGATGCTCAGAAGTTCATCCTGATGCATGAGGTGGTGCAGCCTGTGGTT 900

Sema5aMO AGGAATCTGCAGGATGCTCAGAAGTTCATCCTGATGCATGAGGTGGTGCAGCCTGTGGTT 614

************************************************************

WTsema5a CCTGTGCCGTATTTCATGGAGGACAATGTGCGCTTCTCTCATGTGGCTGTGGACGTGGTG 960

Sema5aMO CCTGTGCCGTATTTCATGGAGGACAATGTGCGCTTCTCTCATGTGGCTGTGGACGTGGTG 674

************************************************************

WTsema5a CAGGGCAAAGACATGCTTTACCACATCATTTATCTGGCAACAGATTACGGCACCATTAAG 1020

Sema5aMO CAGGGCAAAGACATGCTTTACCACATCATTTATCTGGCAACAGATTACGGCACCATTAAG 734

************************************************************

WTsema5a AAGGTGCTCTCCCCTCTCAACCAGACCACGGGCAGCTGCTTGCTGGACGAGATTGAGCTT 1080

Sema5aMO AAGGTGCTCTCCCCTCTCAACCAGACCACGGGCAGCTGCTTGCTGGACGAGATTGAGCTT 794

************************************************************

WTsema5a TTCCCCCTGAAGAAGAGGC 1099

Sema5aMO TTCCCCCTGAAGAAGAGGC 813

*******************

159

References

Adams, J.C., Tucker, R.P., 2000. The Thrombospondin Type 1 Repeat (TSR) Superfamily: Diverse Proteins with Related Roles in Neuronal Development. Dev. Dyn. 2, 280-299.

Adams, R.H., Betz, H., Puschel, A.W., 1996. A Novel Class of Murine Semaphorins with Homology to Thrombospondin is Differentially Expressed during Early Embryogenesis. Mech. Dev. 1, 33-45.

Alberts, B., Johnson, A., Lewis, J., Raff, M., Roberts, K., Walter, P., 2007. Molecular Biology of the Cell - Fifth Edition.

Artigiani, S., Conrotto, P., Fazzari, P., Gilestro, G.F., Barberis, D., Giordano, S., Comoglio, P.M., Tamagnone, L., 2004. Plexin-B3 is a Functional Receptor for Semaphorin 5A. EMBO Rep. 7, 710-714.

Beattie, C.E., Granato, M., Kuwada, J.Y., 2002. Cellular, Genetic and Molecular Mechanisms of Axonal Guidance in the Zebrafish. Results Probl. Cell Differ. 252- 269.

Beattie, C.E., 2000. Control of Motor Axon Guidance in the Zebrafish Embryo. Brain Res. Bull. 5, 489-500.

Beattie, C.E., Melancon, E., Eisen, J.S., 2000. Mutations in the Stumpy Gene Reveal Intermediate Targets for Zebrafish Motor Axons. Development 12, 2653-2662.

Beattie, C.E., Raible, D.W., Henion, P.D., Eisen, J.S., 1999. Early Pressure Screens. Methods Cell Biol. 71-86.

Beattie, C.E., Eisen, J.S., 1997. Notochord Alters the Permissiveness of Myotome for Pathfinding by an Identified Motoneuron in Embryonic Zebrafish. Development 3, 713-720.

Becker, T., McLane, M.A., Becker, C.G., 2003. Integrin Antagonists Affect Growth and Pathfinding of Ventral Motor Nerves in the Trunk of Embryonic Zebrafish. Mol. Cell. Neurosci. 1, 54-68.

160

Beckmann, G., Hanke, J., Bork, P., Reich, J.G., 1998. Merging Extracellular Domains: Fold Prediction for Laminin G-Like and Amino-Terminal Thrombospondin-Like Modules Based on Homology to Pentraxins. J. Mol. Biol. 5, 725-730.

Bernhardt, R.R., Schachner, M., 2000. Chondroitin Sulfates Affect the Formation of the Segmental Motor Nerves in Zebrafish Embryos. Dev. Biol. 1, 206-219.

Bink, R.J., Habuchi, H., Lele, Z., Dolk, E., Joore, J., Rauch, G.J., Geisler, R., Wilson, S.W., den Hertog, J., Kimata, K., Zivkovic, D., 2003. Heparan Sulfate 6-o- Sulfotransferase is Essential for Muscle Development in Zebrafish. J. Biol. Chem. 33, 31118-31127.

Bork, P., 1992. The Modular Architecture of Vertebrate Collagens. FEBS Lett. 1, 49-54.

Boudko, S.P., Engel, J., Bachinger, H.P., 2008. Trimerization and Triple Helix Stabilization of the Collagen XIX NC2 Domain. J. Biol. Chem. 49, 34345-34351.

Brose, K., Tessier-Lavigne, M., 2000. Slit Proteins: Key Regulators of Axon Guidance, Axonal Branching, and Cell Migration. Curr. Opin. Neurobiol. 1, 95-102.

Brose, K., Bland, K.S., Wang, K.H., Arnott, D., Henzel, W., Goodman, C.S., Tessier- Lavigne, M., Kidd, T., 1999. Slit Proteins Bind Robo Receptors and have an Evolutionarily Conserved Role in Repulsive Axon Guidance. Cell 6, 795-806.

Burrows, M.T., 1911. The Growth of Tissues of the Chick Embryo Outside the Animal Body, with Special Reference to the Nervous System. Journal of Experimental Zoology 63-84.

Burstyn-Cohen, T., Tzarfaty, V., Frumkin, A., Feinstein, Y., Stoeckli, E., Klar, A., 1999. F-Spondin is Required for Accurate Pathfinding of Commissural Axons at the Floor Plate. Neuron 2, 233-246.

Challa, A.K., McWhorter, M.L., Wang, C., Seeger, M.A., Beattie, C.E., 2005. Robo3 Isoforms have Distinct Roles during Zebrafish Development. Mech. Dev. 10, 1073- 1086.

Chen, E., Stringer, S.E., Rusch, M.A., Selleck, S.B., Ekker, S.C., 2005. A Unique Role for 6-O Sulfation Modification in Zebrafish Vascular Development. Dev. Biol. 2, 364-376.

Cheng, H.J., Nakamoto, M., Bergemann, A.D., Flanagan, J.G., 1995. Complementary Gradients in Expression and Binding of ELF-1 and Mek4 in Development of the Topographic Retinotectal Projection Map. Cell 3, 371-381.

161

Chien CB, Guidance of Axons and Dendrites. In: Developmental Neurobiology, Plenum Publishers, New York, 2005, .

Clement, A.M., Sugahara, K., Faissner, A., 1999. Chondroitin Sulfate E Promotes Neurite Outgrowth of Rat Embryonic Day 18 Hippocampal Neurons. Neurosci. Lett. 3, 125-128.

Clement, A.M., Nadanaka, S., Masayama, K., Mandl, C., Sugahara, K., Faissner, A., 1998. The DSD-1 Carbohydrate Epitope Depends on Sulfation, Correlates with Chondroitin Sulfate D Motifs, and is Sufficient to Promote Neurite Outgrowth. J. Biol. Chem. 43, 28444-28453.

Colamarino, S.A., Tessier-Lavigne, M., 1995. The Axonal Chemoattractant Netrin-1 is also a Chemorepellent for Trochlear Motor Axons. Cell 4, 621-629.

Colorado, P.C., Torre, A., Kamphaus, G., Maeshima, Y., Hopfer, H., Takahashi, K., Volk, R., Zamborsky, E.D., Herman, S., Sarkar, P.K., Ericksen, M.B., Dhanabal, M., Simons, M., Post, M., Kufe, D.W., Weichselbaum, R.R., Sukhatme, V.P., Kalluri, R., 2000. Anti-Angiogenic Cues from Vascular Basement Membrane Collagen. Cancer Res. 9, 2520-2526.

Conway, G., 2006. STAT3-Dependent Pathfinding and Control of Axonal Branching and Target Selection. Dev. Biol. 1, 119-136.

Cook, G., Tannahill, D., Keynes, R., 1998. Axon Guidance to and from Choice Points. Curr. Opin. Neurobiol. 1, 64-72.

Crossin, K.L., Hoffman, S., Tan, S.S., Edelman, G.M., 1989. Cytotactin and its Proteoglycan Ligand Mark Structural and Functional Boundaries in Somatosensory Cortex of the Early Postnatal Mouse. Dev. Biol. 2, 381-392.

DeFreitas, M.F., Yoshida, C.K., Frazier, W.A., Mendrick, D.L., Kypta, R.M., Reichardt, L.F., 1995. Identification of Integrin Alpha 3 Beta 1 as a Neuronal Thrombospondin Receptor Mediating Neurite Outgrowth. Neuron 2, 333-343.

Dickson, B.J., 2002. Molecular Mechanisms of Axon Guidance. Science 5600, 1959- 1964.

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., Belak, J., Boggs, C., 1996. A Genetic Screen for Mutations Affecting Embryogenesis in Zebrafish. Development 37-46.

162

Eisen, J.S., Pike, S.H., Debu, B., 1989. The Growth Cones of Identified Motoneurons in Embryonic Zebrafish Select Appropriate Pathways in the Absence of Specific Cellular Interactions. Neuron 1, 1097-1104.

Eisen, J.S., Myers, P.Z., Westerfield, M., 1986. Pathway Selection by Growth Cones of Identified Motoneurones in Live Zebra Fish Embryos. Nature 6059, 269-271.

Elzie, C.A., Murphy-Ullrich, J.E., 2004. The N-Terminus of Thrombospondin: The Domain Stands Apart. Int. J. Biochem. Cell Biol. 6, 1090-1101.

Engel, J., 1996. Domain Organizations of Modular Extracellular Matrix Proteins and their Evolution. Matrix Biol. 5, 295-299.

Eyre, D.R., Weis, M.A., Wu, J.J., 2006. Articular Cartilage Collagen: An Irreplaceable Framework? Eur. Cell. Mater. 57-63.

Fazeli, A., Dickinson, S.L., Hermiston, M.L., Tighe, R.V., Steen, R.G., Small, C.G., Stoeckli, E.T., Keino-Masu, K., Masu, M., Rayburn, H., Simons, J., Bronson, R.T., Gordon, J.I., Tessier-Lavigne, M., Weinberg, R.A., 1997. Phenotype of Mice Lacking Functional Deleted in Colorectal Cancer (Dcc) Gene. Nature 6627, 796- 804.

Fein, A.J., Wright, M.A., Slat, E.A., Ribera, A.B., Isom, L.L., 2008. Scn1bb, a Zebrafish Ortholog of SCN1B Expressed in Excitable and Nonexcitable Cells, Affects Motor Neuron Axon Morphology and Touch Sensitivity. J. Neurosci. 47, 12510-12522.

Feldner, J., Reimer, M.M., Schweitzer, J., Wendik, B., Meyer, D., Becker, T., Becker, C.G., 2007. PlexinA3 Restricts Spinal Exit Points and Branching of Trunk Motor Nerves in Embryonic Zebrafish. J. Neurosci. 18, 4978-4983.

Feldner, J., Becker, T., Goishi, K., Schweitzer, J., Lee, P., Schachner, M., Klagsbrun, M., Becker, C.G., 2005. Neuropilin-1a is Involved in Trunk Motor Axon Outgrowth in Embryonic Zebrafish. Dev. Dyn. 3, 535-549.

Fiore, R., Puschel, A.W., 2003. The Function of Semaphorins during Nervous System Development. Front. Biosci. s484-99.

163

Fox, M.A., Sanes, J.R., 2007. Synaptotagmin I and II are Present in Distinct Subsets of Central Synapses. J. Comp. Neurol. 2, 280-296.

Fox, M.A., 2008. Novel Roles for Collagens in Wiring the Vertebrate Nervous System. Curr. Opin. Cell Biol. 5, 508-513.

Funato, H., Saito-Nakazato, Y., Takahashi, H., 2000. Axonal Growth from the Habenular Nucleus Along the Neuromere Boundary Region of the Diencephalon is Regulated by Semaphorin 3F and Netrin-1. Mol. Cell. Neurosci. 3, 206-220.

Gimnopoulos, D., Becker, C.G., Ostendorff, H.P., Bach, I., Schachner, M., Becker, T., 2002. Expression of the Zebrafish Recognition Molecule F3/F11/contactin in a Subset of Differentiating Neurons is Regulated by Cofactors Associated with LIM Domains. Mech. Dev. S135-41.

Goodman, C.S., 1996. Mechanisms and Molecules that Control Growth Cone Guidance. Annu. Rev. Neurosci. 341-377.

Gordon-Weeks PR, 2000. Neuronal Growth Cones. Cambridge University Press, Cambridge, United Kingdom.

Granato, M., van Eeden, F.J., Schach, U., Trowe, T., Brand, M., Furutani-Seiki, M., Haffter, P., Hammerschmidt, M., Heisenberg, C.P., Jiang, Y.J., Kane, D.A., Kelsh, R.N., Mullins, M.C., Odenthal, J., Nusslein-Volhard, C., 1996. Genes Controlling and Mediating Locomotion Behavior of the Zebrafish Embryo and Larva. Development 399-413.

Gundersen, R.W., 1987. Response of Sensory Neurites and Growth Cones to Patterned Substrata of Laminin and Fibronectin in Vitro. Dev. Biol. 2, 423-431.

Haenisch, C., Diekmann, H., Klinger, M., Gennarini, G., Kuwada, J.Y., Stuermer, C.A., 2005. The Neuronal Growth and Regeneration Associated Cntn1 (F3/F11/Contactin) Gene is Duplicated in Fish: Expression during Development and Retinal Axon Regeneration. Mol. Cell. Neurosci. 2, 361-374.

Haffter, P., Granato, M., Brand, M., Mullins, M.C., Hammerschmidt, M., Kane, D.A., Odenthal, J., van Eeden, F.J., Jiang, Y.J., Heisenberg, C.P., Kelsh, R.N., Furutani- Seiki, M., Vogelsang, E., Beuchle, D., Schach, U., Fabian, C., Nusslein-Volhard, C., 1996. The Identification of Genes with Unique and Essential Functions in the Development of the Zebrafish, Danio Rerio. Development 1-36.

Harrison RG, 1914. The Reaction of Embryonic Cells to Solid Structures. Journal of Experimental Zoology 787-848.

164

Harrison, R.G., 1907. Observations on the Living Developing Nerve Fiber. Anatomical Record 116-118.

He, Z., Wang, K.C., Koprivica, V., Ming, G., Song, H.J., 2002. Knowing how to Navigate: Mechanisms of Semaphorin Signaling in the Nervous System. Sci. STKE 119, RE1.

He, Z., Tessier-Lavigne, M., 1997. Neuropilin is a Receptor for the Axonal Chemorepellent Semaphorin III. Cell 4, 739-751.

Heino, J., Huhtala, M., Käpylä, J., Johnson, M.S., 2009. Evolution of Collagen-Based Adhesion Systems. International Journal of Biochemistry and Cell Biology 2, 341- 348.

Henkemeyer, M., Orioli, D., Henderson, J.T., Saxton, T.M., Roder, J., Pawson, T., Klein, R., 1996. Nuk Controls Pathfinding of Commissural Axons in the Mammalian Central Nervous System. Cell 1, 35-46.

Hilario, J.D., Rodino-Klapac, L.R., Wang, C., Beattie, C.E., 2009. Semaphorin 5A is a Bifunctional Axon Guidance Cue for Axial Motoneurons in Vivo. Dev. Biol. 1, 190- 200.

Holt, C.E., Dickson, B.J., 2005. Sugar Codes for Axons? Neuron 2, 169-172.

Hong, K., Hinck, L., Nishiyama, M., Poo, M.M., Tessier-Lavigne, M., Stein, E., 1999. A Ligand-Gated Association between Cytoplasmic Domains of UNC5 and DCC Family Receptors Converts Netrin-Induced Growth Cone Attraction to Repulsion. Cell 7, 927-941.

Hughes, A., 1953. The Growth of Embryonic Neurites. A Study on Cultures of Chick Neural Tissues. Journal of Anatomy 150-162.

Hughes, A.L., 2001. Evolution of the Integrin α and β Protein Families. J. Mol. Evol. 1, 0063-0072.

Hutson, L.D., Chien, C.B., 2002. Wiring the Zebrafish: Axon Guidance and Synaptogenesis. Curr. Opin. Neurobiol. 1, 87-92.

Inatani, M., Irie, F., Plump, A.S., Tessier-Lavigne, M., Yamaguchi, Y., 2003. Mammalian Brain Morphogenesis and Midline Axon Guidance Require Heparan Sulfate. Science 5647, 1044-1046.

165

Ishii, N., Wadsworth, W.G., Stern, B.D., Culotti, J.G., Hedgecock, E.M., 1992. UNC-6, a Laminin-Related Protein, Guides Cell and Pioneer Axon Migrations in C. Elegans. Neuron 5, 873-881.

Ito, Y., Oinuma, I., Katoh, H., Kaibuchi, K., Negishi, M., 2006. Sema4D/plexin-B1 Activates GSK-3beta through R-Ras GAP Activity, Inducing Growth Cone Collapse. EMBO Rep. 7, 704-709.

Jing, L., Lefebvre, J.L., Gordon, L.R., Granato, M., 2009. Wnt Signals Organize Synaptic Prepattern and Axon Guidance through the Zebrafish unplugged/MuSK Receptor. Neuron 5, 721-733.

Joester, A., Faissner, A., 2001. The Structure and Function of Tenascins in the Nervous System. Matrix Biol. 1, 13-22.

Joester, A., Faissner, A., 2001. The Structure and Function of Tenascins in the Nervous System. Matrix Biology 1, 13-22.

Kadler, K.E., Baldock, C., Bella, J., Boot-Handford, R.P., 2007. Collagens at a Glance. J. Cell. Sci. Pt 12, 1955-1958.

Kamphaus, G.D., Colorado, P.C., Panka, D.J., Hopfer, H., Ramchandran, R., Torre, A., Maeshima, Y., Mier, J.W., Sukhatme, V.P., Kalluri, R., 2000. Canstatin, a Novel Matrix-Derived Inhibitor of Angiogenesis and Tumor Growth. J. Biol. Chem. 2, 1209-1215.

Kantor, D.B., Chivatakarn, O., Peer, K.L., Oster, S.F., Inatani, M., Hansen, M.J., Flanagan, J.G., Yamaguchi, Y., Sretavan, D.W., Giger, R.J., Kolodkin, A.L., 2004. Semaphorin 5A is a Bifunctional Axon Guidance Cue Regulated by Heparan and Chondroitin Sulfate Proteoglycans. Neuron 6, 961-975.

Kaprielian, Z., Runko, E., Imondi, R., 2001. Axon Guidance at the Midline Choice Point. Dev. Dyn. 2, 154-181.

Karlstrom, R.O., Trowe, T., Bonhoeffer, F., 1997. Genetic Analysis of Axon Guidance and Mapping in the Zebrafish. Trends Neurosci. 1, 3-8.

166

Keino-Masu, K., Masu, M., Hinck, L., Leonardo, E.D., Chan, S.S., Culotti, J.G., Tessier- Lavigne, M., 1996. Deleted in Colorectal Cancer (DCC) Encodes a Netrin Receptor. Cell 2, 175-185.

Keleman, K., Dickson, B.J., 2001. Short- and Long-Range Repulsion by the Drosophila Unc5 Netrin Receptor. Neuron 4, 605-617.

Kennedy, T.E., Serafini, T., de la Torre, J.R., Tessier-Lavigne, M., 1994. Netrins are Diffusible Chemotropic Factors for Commissural Axons in the Embryonic Spinal Cord. Cell 3, 425-435.

Khaleduzzaman, M., Sumiyoshi, H., Ueki, Y., Inoguchi, K., Ninomiya, Y., Yoshioka, H., 1997. Structure of the Human Type XIX Collagen (COL19A1) Gene, which Suggests it has Arisen from an Ancestor Gene of the FACIT Family. Genomics 2, 304-312.

Kidd, T., Bland, K.S., Goodman, C.S., 1999. Slit is the Midline Repellent for the Robo Receptor in Drosophila. Cell 6, 785-794.

Kim, M.J., Liu, I.H., Song, Y., Lee, J.A., Halfter, W., Balice-Gordon, R.J., Linney, E., Cole, G.J., 2007. Agrin is Required for Posterior Development and Motor Axon Outgrowth and Branching in Embryonic Zebrafish. Glycobiology 2, 231-247.

Kimmel, C.B., Ballard, W.W., Kimmel, S.R., Ullmann, B., Schilling, T.F., 1995. Stages of Embryonic Development of the Zebrafish. Dev. Dyn. 3, 253-310.

Kitsukawa, T., Shimizu, M., Sanbo, M., Hirata, T., Taniguchi, M., Bekku, Y., Yagi, T., Fujisawa, H., 1997. Neuropilin-Semaphorin III/D-Mediated Chemorepulsive Signals Play a Crucial Role in Peripheral Nerve Projection in Mice. Neuron 5, 995-1005.

Kruger, R.P., Aurandt, J., Guan, K.L., 2005. Semaphorins Command Cells to Move. Nat. Rev. Mol. Cell Biol. 10, 789-800.

Kruger, R.P., Lee, J., Li, W., Guan, K.L., 2004. Mapping Netrin Receptor Binding Reveals Domains of Unc5 Regulating its Tyrosine Phosphorylation. J. Neurosci. 48, 10826-10834.

Kullander, K., Croll, S.D., Zimmer, M., Pan, L., McClain, J., Hughes, V., Zabski, S., DeChiara, T.M., Klein, R., Yancopoulos, G.D., Gale, N.W., 2001. Ephrin-B3 is the Midline Barrier that Prevents Corticospinal Tract Axons from Recrossing, Allowing for Unilateral Motor Control. Genes Dev. 7, 877-888.

167

Kwok, J.C., Afshari, F., Garcia-Alias, G., Fawcett, J.W., 2008. Proteoglycans in the Central Nervous System: Plasticity, Regeneration and their Stimulation with Chondroitinase ABC. Restor. Neurol. Neurosci. 2-3, 131-145.

Lee, J.S., Chien, C.B., 2004. When Sugars Guide Axons: Insights from Heparan Sulphate Proteoglycan Mutants. Nat. Rev. Genet. 12, 923-935.

Lee, J.S., von der Hardt, S., Rusch, M.A., Stringer, S.E., Stickney, H.L., Talbot, W.S., Geisler, R., Nusslein-Volhard, C., Selleck, S.B., Chien, C.B., Roehl, H., 2004. Axon Sorting in the Optic Tract Requires HSPG Synthesis by ext2 (Dackel) and extl3 (Boxer). Neuron 6, 947-960.

Legg, A.T., O'Connor, T.P., 2003. Gradients and Growth Cone Guidance of Grasshopper Neurons. J. Histochem. Cytochem. 4, 445-454.

Lein, P.J., Higgins, D., Turner, D.C., Flier, L.A., Terranova, V.P., 1991. The NC1 Domain of Type IV Collagen Promotes Axonal Growth in Sympathetic Neurons through Interaction with the Alpha 1 Beta 1 Integrin. J. Cell Biol. 2, 417-428.

Lemmon, V., Burden, S.M., Payne, H.R., Elmslie, G.J., Hlavin, M.L., 1992. Neurite Growth on Different Substrates: Permissive Versus Instructive Influences and the Role of Adhesive Strength. J. Neurosci. 3, 818-826.

Li, S., Edgar, D., Fassler, R., Wadsworth, W., Yurchenco, P.D., 2003. The Role of Laminin in Embryonic Cell Polarization and Tissue Organization. Dev. Cell. 5, 613- 624.

Livesey, F.J., 1999. Netrins and Netrin Receptors. Cell Mol. Life Sci. 1-2, 62-68.

Luckenbill-Edds, L., 1997. Laminin and the Mechanism of Neuronal Outgrowth. Brain Res. Brain Res. Rev. 1-2, 1-27.

Lustig, M., Sakurai, T., Grumet, M., 1999. Nr-CAM Promotes Neurite Outgrowth from Peripheral Ganglia by a Mechanism Involving Axonin-1 as a Neuronal Receptor. Dev. Biol. 2, 340-351.

168

Maeshima, Y., Manfredi, M., Reimer, C., Holthaus, K.A., Hopfer, H., Chandamuri, B.R., Kharbanda, S., Kalluri, R., 2001. Identification of the Anti-Angiogenic Site within Vascular Basement Membrane-Derived Tumstatin. J. Biol. Chem. 18, 15240-15248.

Manitt, C., Kennedy, T.E., 2002. Where the Rubber Meets the Road: Netrin Expression and Function in Developing and Adult Nervous Systems. Prog. Brain Res. 425-442.

McAdams, B.D., McLoon, S.C., 1995. Expression of Chondroitin Sulfate and Keratan Sulfate Proteoglycans in the Path of Growing Retinal Axons in the Developing Chick. J. Comp. Neurol. 4, 594-606.

Melancon, E., Liu, D.W., Westerfield, M., Eisen, J.S., 1997. Pathfinding by Identified Zebrafish Motoneurons in the Absence of Muscle Pioneers. J. Neurosci. 20, 7796- 7804.

Mizumoto, S., Mikami, T., Yasunaga, D., Kobayashi, N., Yamauchi, H., Miyake, A., Itoh, N., Kitagawa, H., Sugahara, K., 2009. Chondroitin 4-O-Sulfotransferase-1 is Required for Somitic Muscle Development and Motor Axon Guidance in Zebrafish. Biochem. J. 2, 387-399.

Mueller, B.K., 1999. Growth Cone Guidance: First Steps Towards a Deeper Understanding. Annu. Rev. Neurosci. 351-388.

Murai, K.K., Pasquale, E.B., 2003. 'Eph'Ective Signaling: Forward, Reverse and Crosstalk. J. Cell. Sci. Pt 14, 2823-2832.

Myers, J.C., Li, D., Amenta, P.S., Clark, C.C., Nagaswami, C., Weisel, J.W., 2003. Type XIX Collagen Purified from Human Umbilical Cord is Characterized by Multiple Sharp Kinks Delineating Collagenous Subdomains and by Intermolecular Aggregates Via Globular, Disulfide-Linked, and Heparin-Binding Amino Termini. J. Biol. Chem. 34, 32047-32057.

Myers, J.C., Li, D., Bageris, A., Abraham, V., Dion, A.S., Amenta, P.S., 1997. Biochemical and Immunohistochemical Characterization of Human Type XIX Defines a Novel Class of Basement Membrane Zone Collagens. Am. J. Pathol. 6, 1729-1740.

Myers, P.Z., Eisen, J.S., Westerfield, M., 1986. Development and Axonal Outgrowth of Identified Motoneurons in the Zebrafish. J. Neurosci. 8, 2278-2289.

Myllyharju, J., Kivirikko, K.I., 2004. Collagens, Modifying Enzymes and their Mutations in Humans, Flies and Worms. Trends Genet. 1, 33-43.

169

Nakabayashi, K., Scherer, S.W., 2001. The Human Contactin-Associated Protein-Like 2 Gene (CNTNAP2) Spans Over 2 Mb of DNA at Chromosome 7q35. Genomics 1, 108-112.

Nakai, J., Kawasaki, Y., 1959. Studies on the Mechanism Determining the Course of Nerve Fibers in Tissue Culture. I. the Reaction of the Growth Cone to various Obstructions. Z. Zellforsch. Mikrosk. Anat. 108-122.

Nakamura, F., Kalb, R.G., Strittmatter, S.M., 2000. Molecular Basis of Semaphorin- Mediated Axon Guidance. J. Neurobiol. 2, 219-229.

Nasevicius, A., Ekker, S.C., 2000. Effective Targeted Gene 'Knockdown' in Zebrafish. Nat. Genet. 2, 216-220.

Nasevicius, A., Larson, J., Ekker, S.C., 2000. Distinct Requirements for Zebrafish Angiogenesis Revealed by a VEGF-A Morphant. Yeast 4, 294-301.

Negishi, M., Oinuma, I., Katoh, H., 2005. Plexins: Axon Guidance and Signal Transduction. Cellular and Molecular Life Sciences 12, 1363-1371.

Neufeld, ,Gera, Kessler, ,Ofra, The Semaphorins: Versatile Regulators of Tumour Progression and Tumour Angiogenesis.

Oakley, R.A., Tosney, K.W., 1991. Peanut Agglutinin and Chondroitin-6-Sulfate are Molecular Markers for Tissues that Act as Barriers to Axon Advance in the Avian Embryo. Dev. Biol. 1, 187-206.

O'Connor TP, 1999. Intermediate Targets and Segmental Pathfinding. CMLS 1358-1364.

Oinuma, I., Ishikawa, Y., Katoh, H., Negishi, M., 2004. The Semaphorin 4D Receptor Plexin-B1 is a GTPase Activating Protein for R-Ras. Science 5685, 862-865.

Oiso, S., Takeda, Y., Futagawa, T., Miura, T., Kuchiiwa, S., Nishida, K., Ikeda, R., Kariyazono, H., Watanabe, K., Yamada, K., 2009. Contactin-Associated Protein (Caspr) 2 Interacts with Carboxypeptidase E in the CNS. J. Neurochem. 1, 158-167.

Ortega, N., Werb, Z., 2002. New Functional Roles for Non-Collagenous Domains of Basement Membrane Collagens. J. Cell. Sci. Pt 22, 4201-4214.

O'Shea, K.S., Liu, L.H., Dixit, V.M., 1991. Thrombospondin and a 140 Kd Fragment Promote Adhesion and Neurite Outgrowth from Embryonic Central and Peripheral Neurons and from PC12 Cells. Neuron 2, 231-237.

170

Osterhout, D.J., Frazier, W.A., Higgins, D., 1992. Thrombospondin Promotes Process Outgrowth in Neurons from the Peripheral and Central Nervous Systems. Dev. Biol. 2, 256-265.

Palaisa, K.A., Granato, M., 2007. Analysis of Zebrafish Sidetracked Mutants Reveals a Novel Role for Plexin A3 in Intraspinal Motor Axon Guidance. Development 18, 3251-3257.

Panzer, J.A., Gibbs, S.M., Dosch, R., Wagner, D., Mullins, M.C., Granato, M., Balice- Gordon, R.J., 2005. Neuromuscular Synaptogenesis in Wild-Type and Mutant Zebrafish. Dev. Biol. 2, 340-357.

Pasterkamp, R.J., Kolodkin, A.L., 2003. Semaphorin Junction: Making Tracks Toward Neural Connectivity. Curr. Opin. Neurobiol. 1, 79-89.

Paulus, J.D., Halloran, M.C., 2006. Zebrafish bashful/laminin-Alpha 1 Mutants Exhibit Multiple Axon Guidance Defects. Dev. Dyn. 1, 213-224.

Petitclerc, E., Boutaud, A., Prestayko, A., Xu, J., Sado, Y., Ninomiya, Y., Sarras, M.P.,Jr, Hudson, B.G., Brooks, P.C., 2000. New Functions for Non-Collagenous Domains of Human Collagen Type IV. Novel Integrin Ligands Inhibiting Angiogenesis and Tumor Growth in Vivo. J. Biol. Chem. 11, 8051-8061.

Pevsner, J., 2002. Leonardo Da Vinci's Contributions to Neuroscience. Trends Neurosci. 4, 217-220.

Placzek, M., Tessier-Lavigne, M., Yamada, T., Dodd, J., Jessell, T.M., 1990. Guidance of Developing Axons by Diffusible Chemoattractants. Cold Spring Harb. Symp. Quant. Biol. 279-289.

Poliak, S., Gollan, L., Martinez, R., Custer, A., Einheber, S., Salzer, J.L., Trimmer, J.S., Shrager, P., Peles, E., 1999. Caspr2, a New Member of the Neurexin Superfamily, is Localized at the Juxtaparanodes of Myelinated Axons and Associates with K+ Channels. Neuron 4, 1037-1047.

Pollard, S.M., Parsons, M.J., Kamei, M., Kettleborough, R.N., Thomas, K.A., Pham, V.N., Bae, M.K., Scott, A., Weinstein, B.M., Stemple, D.L., 2006. Essential and Overlapping Roles for Laminin Alpha Chains in Notochord and Blood Vessel Formation. Dev. Biol. 1, 64-76.

Polleux, F., Morrow, T., Ghosh, A., 2000. Semaphorin 3A is a Chemoattractant for Cortical Apical Dendrites. Nature 6778, 567-573.

171

Prockop, D.J., Kivirikko, K.I., 1995. Collagens: Molecular Biology, Diseases, and Potentials for Therapy. Annu. Rev. Biochem. 403-434.

Ramon y Cajal, S., A Quelle Epoque Apparaissent Les Expansions Des Cellules Nerveuses De La Moelle Epiniere Du Poulet? Anatomischer Anzeiger 609-613-631- 639.

Raper, J.A., 2000. Semaphorins and their Receptors in Vertebrates and Invertebrates. Curr. Opin. Neurobiol. 1, 88-94.

Rauvala, H., Huttunen, H.J., Fages, C., Kaksonen, M., Kinnunen, T., Imai, S., Raulo, E., Kilpeläinen, I., 2000. Heparin-Binding Proteins HB-GAM (Pleiotrophin) and Amphoterin in the Regulation of Cell Motility. Matrix Biology 5, 377-387.

Robu, M.E., Larson, J.D., Nasevicius, A., Beiraghi, S., Brenner, C., Farber, S.A., Ekker, S.C., 2007. P53 Activation by Knockdown Technologies. PLoS Genet. 5, e78.

Rodino-Klapac, L.R., Beattie, C.E., 2004. Zebrafish Topped is Required for Ventral Motor Axon Guidance. Dev. Biol. 2, 308-320.

Roos, M., Schachner, M., Bernhardt, R.R., 1999. Zebrafish Semaphorin Z1b Inhibits Growing Motor Axons in Vivo. Mech. Dev. 1-2, 103-117.

Rothberg, J.M., Jacobs, J.R., Goodman, C.S., Artavanis-Tsakonas, S., 1990. Slit: An Extracellular Protein Necessary for Development of Midline Glia and Commissural Axon Pathways Contains both EGF and LRR Domains. Genes Dev. 12A, 2169- 2187.

Sahay, A., Molliver, M.E., Ginty, D.D., Kolodkin, A.L., 2003. Semaphorin 3F is Critical for Development of Limbic System Circuitry and is Required in Neurons for Selective CNS Axon Guidance Events. J. Neurosci. 17, 6671-6680.

Sato-Maeda, M., Obinata, M., Shoji, W., 2008. Position Fine-Tuning of Caudal Primary Motoneurons in the Zebrafish Spinal Cord. Development 2, 323-332.

Sato-Maeda, M., Tawarayama, H., Obinata, M., Kuwada, J.Y., Shoji, W., 2006. Sema3a1 Guides Spinal Motor Axons in a Cell- and Stage-Specific Manner in Zebrafish. Development 5, 937-947.

Schneider, V.A., Granato, M., 2006. The Myotomal Diwanka (lh3) Glycosyltransferase and Type XVIII Collagen are Critical for Motor Growth Cone Migration. Neuron 5, 683-695.

172

Schneider, V.A., Granato, M., 2003. Motor Axon Migration: A Long Way to Go. Dev. Biol. 1, 1-11.

Schweitzer, J., Becker, T., Lefebvre, J., Granato, M., Schachner, M., Becker, C.G., 2005. Tenascin-C is Involved in Motor Axon Outgrowth in the Trunk of Developing Zebrafish. Dev. Dyn. 3, 550-566.

Seeger, M., Tear, G., Ferres-Marco, D., Goodman, C.S., 1993. Mutations Affecting Growth Cone Guidance in Drosophila: Genes Necessary for Guidance Toward Or Away from the Midline. Neuron 3, 409-426.

Seeger, M.A., Beattie, C.E., 1999. Attraction Versus Repulsion: Modular Receptors make the Difference in Axon Guidance. Cell 7, 821-824.

Semaphorin Nomenclature Committee, 1999. Unified Nomenclature for the semaphorins/collapsins. Cell 5, 551-552.

Sentinelli, F., Lovari, S., Vitale, M., Giorgi, G., Di Mario, U., Baroni, M.G., 2000. A Simple Method for Non-Radioactive PCR-SSCP using MDE Gel Solution and a Midi Gel Format: Application for the Detection of Variants in the GLUT1 and CTLA-4 Genes. J. Biotechnol. 2, 201-204.

Serafini, T., Colamarino, S.A., Leonardo, E.D., Wang, H., Beddington, R., Skarnes, W.C., Tessier-Lavigne, M., 1996. Netrin-1 is Required for Commissural Axon Guidance in the Developing Vertebrate Nervous System. Cell 6, 1001-1014.

Shoji, W., Isogai, S., Sato-Maeda, M., Obinata, M., Kuwada, J., 2003. Semaphorin3a1 Regulates Angioblast Migration and Vascular Development in Zebrafish Embryos. Development 14, 3227-3236.

Sipila, L., Ruotsalainen, H., Sormunen, R., Baker, N.L., Lamande, S.R., Vapola, M., Wang, C., Sado, Y., Aszodi, A., Myllyla, R., 2007. Secretion and Assembly of Type IV and VI Collagens Depend on Glycosylation of Hydroxylysines. J. Biol. Chem. 46, 33381-33388.

Speidel, C.C., 1933. Studies of Living Nerves II. Activities of Ameboid Growth Cones Sheat Cells and Myelin Sheaths, as Revealed by Prolonged Observation of Individual Nerve Fibers in Frog Tadpoles. American Journal of Anatomy 1-79.

Sperry, R.W., 1943. Visuomotor Coordination in the Newt (Triturus Viridescens) After Regeneration of the Optic Nerve. Journal of Comparative Neurology 33-35.

173

Stoeckli, E.T., Landmesser, L.T., 1995. Axonin-1, Nr-CAM, and Ng-CAM Play Different Roles in the in Vivo Guidance of Chick Commissural Neurons. Neuron 6, 1165-1179.

Su, J., Gorse, K., Ramirez, F., Fox, M.A., 2010. Collagen XIX is Expressed by Interneurons and Contributes to the Formation of Hippocampal Synapses. J. Comp. Neurol. 2, 229-253.

Sumiyoshi, H., Mor, N., Lee, S.Y., Doty, S., Henderson, S., Tanaka, S., Yoshioka, H., Rattan, S., Ramirez, F., 2004. Esophageal Muscle Physiology and Morphogenesis Require Assembly of a Collagen XIX-Rich Basement Membrane Zone. J. Cell Biol. 4, 591-600.

Sumiyoshi, H., Laub, F., Yoshioka, H., Ramirez, F., 2001. Embryonic Expression of Type XIX Collagen is Transient and Confined to Muscle Cells. Dev. Dyn. 2, 155- 162.

Takagi, S., Kasuya, Y., Shimizu, M., Matsuura, T., Tsuboi, M., Kawakami, A., Fujisawa, H., 1995. Expression of a Cell Adhesion Molecule, Neuropilin, in the Developing Chick Nervous System. Dev. Biol. 1, 207-222.

Tamagnone, L., Comoglio, P.M., 2004. To Move Or Not to Move? Semaphorin Signalling in Cell Migration. EMBO Rep. 4, 356-361.

Tamagnone, L., Artigiani, S., Chen, H., He, Z., Ming, G.I., Song, H., Chedotal, A., Winberg, M.L., Goodman, C.S., Poo, M., Tessier-Lavigne, M., Comoglio, P.M., 1999. Plexins are a Large Family of Receptors for Transmembrane, Secreted, and GPI-Anchored Semaphorins in Vertebrates. Cell 1, 71-80.

Tanaka, H., Maeda, R., Shoji, W., Wada, H., Masai, I., Shiraki, T., Kobayashi, M., Nakayama, R., Okamoto, H., 2007. Novel Mutations Affecting Axon Guidance in Zebrafish and a Role for Plexin Signalling in the Guidance of Trigeminal and Facial Nerve Axons. Development 18, 3259-3269.

Tessier-Lavigne, M., Placzek, M., Lumsden, A.G., Dodd, J., Jessell, T.M., 1988. Chemotropic Guidance of Developing Axons in the Mammalian Central Nervous System. Nature 6201, 775-778.

Thisse, C., Thisse, B., Schilling, T.F., Postlethwait, J.H., 1993. Structure of the Zebrafish snail1 Gene and its Expression in Wild-Type, Spadetail and no Tail Mutant Embryos. Development 4, 1203-1215.

Trevarrow, B., Marks, D.L., Kimmel, C.B., 1990. Organization of Hindbrain Segments in the Zebrafish Embryo. Neuron 5, 669-679. 174

Van Vactor, D., Lorenz, L.J., 1999. Introduction: Invertebrate Axons Find their Way. Cell Mol. Life Sci. 11, 1355-1357.

Walz, A., Anderson, R.B., Irie, A., Chien, C.B., Holt, C.E., 2002. Chondroitin Sulfate Disrupts Axon Pathfinding in the Optic Tract and Alters Growth Cone Dynamics. J. Neurobiol. 3, 330-342.

Weiss, L.A., Arking, D.E., Gene Discovery Project of Johns Hopkins & the Autism Consortium, Daly, M.J., Chakravarti, A., 2009. A Genome-Wide Linkage and Association Scan Reveals Novel Loci for Autism. Nature 7265, 802-808.

Wilkinson, D.,G., Multiple Roles of Eph Receptors and Ephrins in Neural Development.

Wilkinson, D.G., 2001. Multiple Roles of EPH Receptors and Ephrins in Neural Development. Nat. Rev. Neurosci. 3, 155-164.

Winberg, M.L., Noordermeer, J.N., Tamagnone, L., Comoglio, P.M., Spriggs, M.K., Tessier-Lavigne, M., Goodman, C.S., 1998. Plexin A is a Neuronal Semaphorin Receptor that Controls Axon Guidance. Cell, 7, 903-916.

Wong, J.T., Wong, S.T., O'Connor, T.P., 1999. Ectopic Semaphorin-1a Functions as an Attractive Guidance Cue for Developing Peripheral Neurons. Nat. Neurosci. 9, 798- 803.

Woods, I.G., Wilson, C., Friedlander, B., Chang, P., Reyes, D.K., Nix, R., Kelly, P.D., Chu, F., Postlethwait, J.H., Talbot, W.S., 2005. The Zebrafish Gene Map Defines Ancestral Vertebrate Chromosomes. Genome Res. 9, 1307-1314.

Xiao, T., Baier, H., 2007. Lamina-Specific Axonal Projections in the Zebrafish Tectum Require the Type IV Collagen Dragnet. Nat. Neurosci. 12, 1529-1537.

Yan, Y.L., Talbot, W.S., Egan, E.S., Postlethwait, J.H., 1998. Mutant Rescue by BAC Clone Injection in Zebrafish. Genomics 2, 287-289.

Yazdani, U., Terman, J.R., 2006. The Semaphorins. Genome Biol. 3, 211.

Zeller, J., Granato, M., 1999. The Zebrafish Diwanka Gene Controls an Early Step of Motor Growth Cone Migration. Development 15, 3461-3472.

Zhang, J., Lefebvre, J.L., Zhao, S., Granato, M., 2004. Zebrafish Unplugged Reveals a Role for Muscle-Specific Kinase Homologs in Axonal Pathway Choice. Nat. Neurosci. 12, 1303-1309.

175

Zhang, J., Granato, M., 2000. The Zebrafish Unplugged Gene Controls Motor Axon Pathway Selection. Development 10, 2099-2111.

Zhou, Y., Gunput, R.A., Pasterkamp, R.J., 2008. Semaphorin Signaling: Progress made and Promises Ahead. Trends Biochem. Sci. 4, 161-170.

Zuo, J., Neubauer, D., Dyess, K., Ferguson, T.A., Muir, D., 1998. Degradation of Chondroitin Sulfate Proteoglycan Enhances the Neurite-Promoting Potential of Spinal Cord Tissue. Exp. Neurol. 2, 654-662.

176