@prefix vivo: . @prefix edm: . @prefix ns0: . @prefix dcterms: . @prefix dc: . @prefix skos: . vivo:departmentOrSchool "Medicine, Faculty of"@en ; edm:dataProvider "DSpace"@en ; ns0:degreeCampus "UBCV"@en ; dcterms:creator "Wong, June Tan-Whea"@en ; dcterms:issued "2009-07-20T19:47:42Z"@en, "2000"@en ; vivo:relatedDegree "Doctor of Philosophy - PhD"@en ; ns0:degreeGrantor "University of British Columbia"@en ; dcterms:description """During development, elongating axons must make important steering decisions in order to establish the intricate and precise web of connections in a functional nervous system. Developing axons are currently believed to be guided by cues present in their local environment which act through four general mechanisms: chemoattraction/chemorepulsion (secreted molecules), and contact mediated attraction/contact mediated repulsion (cell surface molecules) which effectively steer the axon growth cone along a directed path to its correct target. The semaphorin family of secreted and cell surface glycoproteins were first described in 1993 and are thought to play an integral role in axon pathfinding by acting as chemorepulsive guidance molecules. However, although chemorepulsion had been demonstrated for several secreted semaphorin family members, little was known about the function of the majority of the semaphorin members, including the transmembrane forms, at the time that this thesis project began in 1995. The aim of this thesis was to investigate the role of Semaphorin-1a, the founding member of the semaphorin family, using the highly accessible peripheral nervous system of the developing grasshopper limb bud with the aspiration of obtaining a better understanding of the functional role of a transmembrane semaphorin. The studies outlined in this thesis investigate the effects of Sema-1a on the pathfinding events of several early arising neurons. The data presented herein describe the first evidence of an attractive guidance activity for a semaphorin family member and highlights the complexity and utility of the semaphorins in axon guidance."""@en ; edm:aggregatedCHO "https://circle.library.ubc.ca/rest/handle/2429/10999?expand=metadata"@en ; dcterms:extent "13716142 bytes"@en ; dc:format "application/pdf"@en ; skos:note "Functional Analysis of Semaphorin-la in the Developing PNS by J U N E T A N - W H E A W O N G B.Sc. (Biochemistry), University of British Columbia, 1994 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L M E N T OF T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R OF P H I L O S O P H Y in T H E F A C U L T Y OF G R A D U A T E STUDIES (Department of Anatomy) We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y OF B R I T I S H C O L U M B I A Apr i l 2000 © June Tan-Whea Wong, 2000 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada DE-6 (2/88) A B S T R A C T During development, elongating axons must make important steering decisions in order to establish the intricate and precise web of connections in a functional nervous system. Developing axons are currently believed to be guided by cues present in their local environment which act through four general mechanisms: chemoattraction/chemorepulsion (secreted molecules), and contact mediated attraction/contact mediated repulsion (cell surface molecules) which effectively steer the axon growth cone along a directed path to its correct target. The semaphorin family of secreted and cell surface glycoproteins were first described in 1993 and are thought to play an integral role in axon pathfinding by acting as chemorepulsive guidance molecules. However, although chemorepulsion had been demonstrated for several secreted semaphorin family members, little was known about the function of the majority of the semaphorin members, including the transmembrane forms, at the time that this thesis project began in 1995. The aim of this thesis was to investigate the role of Semaphorin-la, the founding member of the semaphorin family, using the highly accessible peripheral nervous system of the developing grasshopper limb bud with the aspiration of obtaining a better understanding of the functional role o f a transmembrane semaphorin. The studies outlined in this thesis investigate the effects of Sema-1 a on the pathfinding events of several early arising neurons. The data presented herein describe the first evidence of an attractive guidance activity for a semaphorin family member and highlights the complexity and utility of the semaphorins in axon guidance. T A B L E O F C O N T E N T S Page Abstract 1 1 Table of Contents n i List of Figures v n Acknowledgements x Chapter 1 Introduction: 1 The semaphorins 4 Discovery of the semaphorins 4 Semaphorin structure 7 Functional studies 8 Class 1 semaphorins: function unclear 8 Class 2 semaphorins: chemorepellents in the invertebrate nervous system 9 Class 3 semaphorins: chemorepulsive guidance cues in the vertebrate system 12 Receptor 16 Transmembrane semaphorins 22 Class 4 semaphorins 25 Class 5 semaphorins 26 Class 6 semaphorins 27 Class 7 semaphorins 27 The model system: The developing grasshopper nervous system 29 Why study the invertebrate system? 29 The grasshopper 33 The grasshopper nervous system 37 The grasshopper limb bud 38 Early events in PNS development 42 The role of Sema-1 a in peripheral axon pathfinding 44 Objectives 46 Chapter 2 Differential expression of Grasshopper Semaphorin-la in the developing limb bud epithelium and relationship to developing peripheral axon tracts during neuroembryogenesis 47 ui Introduction 47 Materials and Methods 49 Staging of grasshopper embryos 49 Preparation of 6F8 antibodies 50 Immunocytochemistry 50 Antibody double labeling 52 Results 52 Developmental time course of Sema-la expression in limb bud 52 Summary of Sema-la expression 61 Relationship of identified neurons to Sema-la 64 Relationship of scolopidial organs to Sema-la 72 Discussion 74 Expression of Sema-1 a is developmentally regulated 74 Expression of SemaO 1 a correlates with role in axon guidance 75 Chapter 3 Sema-la mediates S G O axon stalling after removal of guidepost cells but is not an inhibitory or collapsing cue for the S G O growth cones 77 Introduction 77 Materials and Methods 81 Immunocytochemistry 81 Heat shock 81 Limb fillet 82 D i l labeling o f neurons 82 T i 1 neuron removal 82 Results 83 The S G O growth cones are arrested within a band o f Sema-la 83 The S G O projection is pioneered by the most proximal neurons 93 Proximal extension of SGO arrested in limb fillet 93 Discussion 95 Sema-la is an attractive cue for the S G O 97 Chapter 4 Sema-la promotes S G O axon outgrowth 100 Introduction 100 Materials and Methods 101 Preparation of 6F8 blocking antibodies 101 Antibody blocking 102 Visualization of neurons 102 Results 103 6F8 blocking of Sema-la inhibits S G O outgrowth and extension 103 6F8 blocking of Sema-la later in development prevents migration 106 Discussion 107 A role for Sema-la in the patterning of the grasshopper PNS 107 The semaphorins are a growing family with multiple functions 110 Chapter 5 The activity of Sema-la is mediated by the semaphorin domain 113 iv Introduction 113 Materials and Methods 115 Recombinant Sema-1 a protein production 115 Anti-cmyc immunoaffinity column preparation 115 Embryo culture 116 Results 117 Discussion 132 Chapter 6 Transmembrane Sema-la requires clustering for its full functional activity: Sema-la clustering as a cellular mechanism to control activity 135 Introduction 135 Materials and Methods 136 Recombinant Sema-la construct 136 Embryo culture 137 Antibody clustering . . . . 137 rSema-la localization 138 Immunofluorescence microscopy of S2 and epithelial cells 138 Visualization of neurons 138 Results 139 Discussion 149 Dimerization of semaphorins 149 Role of oligomerization in regulating Sema-la activity 150 Chapter 7 Localized ectopic Sema-la actively attracts T i l growth cones in vivo 153 Introduction 153 Materials and Methods 155 Construction of C O S cell expression vector 155 S2 cell transfection 155 Production of S2 cell balls 156 C O S cell transfection 156 C O S cell aggregates 156 Limb fillet assay system 156 Visualization of neurons 157 R N A isolation and northern analysis 157 Results 158 Discussion 162 Summary 168 Appendix A : The role of heterotrimeric G proteins in semaphorin signaling reassessed... . 171 Introduction 171 Heterotrimeric G proteins, pertussis toxin, and mastoparan 172 Heterotrimeric G proteins in neuronal guidance 173 Heterotrimeric G proteins mediate Sema5A induced growth cone collapse 175 Do heterotrimeric G proteins mediate Sema-la growth cone attraction? . . . . 176 Materials and Methods 177 v T i l assay 177 S G O assay 177 Visualization of neurons 178 Results 178 Pertussis toxin does not disrupt T i l pathfinding 178 Pertussis toxin does not disrupt SGO pathfinding or outgrowth 181 Discussion 183 Sema-la attractive activity is not mediated by a G protein coupled receptor 183 The semaphorin receptor is not a G protein coupled receptor 185 Recent controversy over Semaphorin signaling 186 New molecules implicated in semaphorin signaling 188 Chapter 9 Towards the future of Sema-la function analysis: Cloning and sequencing of C S E M A 5B, a vertebrate semaphorin with high homology to Sema-la 190 Introduction 190 Materials and Methods 192 Clustal alignment of semaphorins 192 Design degenerate P C R primers 192 R N A extraction from E10 chick brain 192 Reverse Transcription 193 L o w stringency P C R cloning 193 Cloning of P C R products into pCR2.1 plasmid 193 Library screening 193 Clone analysis 194 Sequencing 194 Results 194 Identification of conserved sequences unique to transmembrane semaphorins 194 Identification of Chick S E M A 5 B 195 Structure of C S E M A 5 B 202 Discussion 202 Chick S E M A 5 B has highest homology to grasshopper Sema-la 202 Thrombospondins mediate neurite outgrowth and attachment 203 Appendix C : Development of a grasshopper culture medium 206 References 207 VI Lis t of Figures Page Figure 1.1 Schematic summary of the four types of mechanisms which contribute to guidance of growth cones 2 Figure 1.2 Schematic diagram of the effects of 6F8 antibody on T i l pathfinding in the developing grasshopper limb bud 5 Figure 1.3 Schematic diagram illustrating the classic experiments which indicate that secreted class 2 and class 3 semaphorins are inhibitory guidance molecules 10 Figure 1.4 Structure of the class 3 semaphorin and its receptor, neuropilin 19 Figure 1.5 Structural schematic of the Semaphorin family 23 Figure 1.6 Schematic representation of the grasshopper limb bud 31 Figure 1.7 Schematic diagram summarizing the limb fillet procedure 35 Figure 1.8 Schematic diagram of the initial axon paths in the developing grasshopper peripheral nervous system 39 Figure 2.1 Embryonic expression of Sema-1 a in the developing grasshopper limb bud 54 Figure 2.2 Schematic summary of Sema-la expression in the developing grasshopper limb bud epithelium 56 Figure 2.3 Embryonic expression o f Sema-la in the developing grasshopper l imb bud and its relationship to developing axon tracts 65 Figure 2.4 Schematic summary of the developmental expression pattern of Sema-la in the grasshopper limb bud epithelium and its relationship to developing axon tracts 67 Figure 3.1 Proximal migration of the T i l growth cones is arrested in the absence of the T i l cell bodies 79 Figure 3.2 Elimination of the T i l pioneer pathway by heat shock prevents the proximal extension of the S G O growth cones 85 Figure 3.3 The growth cones of the SGO are arrested within a band of Sema-la 87 VI Figure 3.4 Number and location of SGO neurons with axon projections in the T i l pathway 89 Figure 3.5 Removal of T i l cell bodies in the limb fillet preparation inhibits proximal extension of the S G O growth cones 91 Figure 4.1 m A b 6F8 blocking of Sema-1 a prevents axon outgrowth from the S G O 104 Figure 4.2 Schematic summary of heat shock and limb fillet experiments 108 Figure 5.1 Recombinant Sema-la construct 118 Figure 5.2 Recombinant Sema-la construct - Sequence 120 Figure 5.3 Penetrance and distribution of soluble rSema-la in the limb bud during embryo culture 122 Figure 5.4 Ectopic rSema-la induces defects in T i l - axon guidance - Histogram . . . . 124 Figure 5.5 Phenotype of defects induced by ectopic rSema-la 126 Figure 5.6 Schematic summary of rSema-la induced T i l axon pathfinding defects. . . . 128 Figure 5.7 Dose-response curve for soluble rSema-la 130 Figure 6.1 Penetrance and distribution of clustered rSema-la in the limb bud during embryo culture 140 Figure 6.2 Dose-response curve comparing the activity of soluble, dimerized or clustered rSema-la 142 Figure 6.3 T i l axon defects induced by clustered rSema-la 144 Figure 6.4 Full-length Sema-la forms microscopically visible aggregates on the cell surface 146 Figure 7.1 T i l growth cones turn towards ectopic S2 and C O S cells expressing Sema-la 159 Figure 7.2 Structure of the class 1 semaphorin and its receptor, Plexin A 165 Figure 8.1 Effects of P T X on T i l pathfinding 179 Figure 9.1 Clustal alignment of the Chick Sema V sequence with related semaphorins 196 Figure 9.2 Hydrophobicity plot for the predicted Chick SemalV protein 198 vi i i Figure 9.3 Phylogenetic tree for semaphorin domains 200 IX Acknowledgements I have enjoyed my experience as graduate student in the Department of Anatomy at the University of British Columbia. I would like to thank my graduate supervisor, Dr. T im O'Connor for his support and positive encouragement over this period of time. T im has provided an open and supportive atmosphere, allowing for individual creative thinking and resulting in a truly rewarding learning environment. M y research and studies were also facilitated greatly by the members of my supervisory committee, and I would like to thank Dr. Vanessa Au ld , Dr. Cal Roskelly and Dr. John Church for time, guidance, and direction throughout the course of my program. The Department of Anatomy as a whole, the students, faculty, and staff, all contributed to a constructive working atmosphere. Finally, but most importantly, I am very grateful to my family for their never-ending encouragement and support. Chapter 1: Introduction The formation of a functional nervous system during development is a highly complex 12 15 process. For instance, to generate a functioning human brain, 10 neurons make over 10 connections with target cells. During development, axonal growth cones must navigate over long distances along highly specific pathways in order to establish the complex patterns of neuronal connections in the adult nervous system. The enormous complexity of the nervous system is intriguing, and for over a century neurobiologists have struggled to obtain an understanding of some of the fundamental processes of how axons are guided during development. In 1892, a renowned Spanish neurobiologist, Ramon y Cajal, proposed that axonal growth cones are directed to their targets by chemotaxis; extending along an attractive gradient of diffusible factors emanating from the target (Ramon y Cajal, 1893). In the 1960's Roger Sperry proposed the basic idea of the \"chemoaffinity hypothesis\" (Sperry, 1963) that individual neurons acquire distinct molecular markers early in development, and that establishment of appropriate connections between two neurons would depend on the correct matching of molecules present on the pre- and postsynaptic neuron. We now believe that growth cones are guided by at least four general mechanisms (see Tessier-Lavigne and Goodman, 1996): chemoattraction, chemorepulsion, contact mediated attraction, and contact mediated repulsion (Fig. 1.1). In addition, a number of molecules which are candidate guidance cues have been identified over the past decade, and include members of the netrin family, semaphorin family and E P H family (see Goodman, 1996, Tessier-Lavigne and Goodman, 1996). The semaphorins are the largest family identified to date, with a number of potential roles in neurodevelopment. 1 CONTACT ATTRACTION CONTACT REPULSION m CHEMOATTRACTION + + + + + + + o + + c CHEMOREPULSION - - J3 J O Figure 1.1. Schematic summary of the four types of mechanisms which contribute to guidance of growth cones. A x o n guidance mechanisms could be categorized into four groups: contact attraction, contact repulsion, chemoattraction, and chemorepulsion. The term attraction refers to a range o f permissive, outgrowth promoting, adhesive and attractive effects. The term repulsion refers to a range o f inhibitory, collapsing, and repulsive effects. 3 The semaphorins Discovery of the semaphorins The semaphorins are a large family of transmembrane and secreted glycoproteins which share a highly conserved domain of approximately 500 amino acids known as the semaphorin domain (Kolodkin et al., 1992, 1993; Luo et al., 1993, 1995; Puschel et al., 1995). The first member of the semaphorin family was originally isolated in a monoclonal antibody screen for differentially expressed surface glycoproteins in the developing grasshopper central nervous system (CNS) in an attempt to identify molecules involved in selective axonal fasciculation (Kolodkin et al., 1992). Unlike most of the molecules isolated from the screen, one of the molecules isolated, Fasciclin IV (now Sema-la), was not a homophilic cell adhesion molecule. Most interestingly however, was the finding that Sema-la was not only expressed on subsets of axons in the C N S , but also on bands of epithelial cells in the developing grasshopper limb bud where it was implicated to potentially play an important role in the guidance of axons in the developing grasshopper PNS (Kolodkin et al., 1992). In particular, the growth cones of two sensory neurons, the T i l pioneer neuron pair, were observed to make a sharp ventral turn upon encountering a band of Sema-la expressing epithelial cells located at the trochanter limb segment (Fig. 1.2). Antibody perturbation of Fasciclin IV function resulted in aberrant pathfinding and defasciculation of the T i l axons at the region of the band, suggesting that Fasciclin IV is required for the proper guidance of the T i l growth cones in the limb bud. Subsequent cloning of homologous sequences from Tribolium confusum, Drosophila melanogaster, and human led to the identification of a new gene family, the semaphorins (Kolodkin et al., 1993). Due to its potential role as an axonal guidance molecule, Fasciclin IV 4 6\" Figure 1.2. Schematic diagram of the effects of 6F8 antibody on T i l pathfinding in the developing grasshopper limb bud. Wild-Type. The T i l axons initially extend proximally towards the C N S but reorient once they encounter a band of Sema-la expressing cells (gray) at the T r -Cx limb segment boundary. After contacting the Sema-la band, the T i l growth cones turn sharply to migrate within the Sema-la band and reorient a second time after reaching the C x i neurons to migrate proximally into the C N S . Antibody block of Sema-la. In limb buds which were cultured in the presence o f antibodies against Sema-la, the T i l growth cones extend aberrantly across the Sema-la stripe. Arrows indicate the location of limb segment boundaries. 6 was renamed semaphorin I (Sema I, now known as Sema-la), to signify the first member of a family of molecules which provide axon guidance information . Within the first few years after the cloning of the initial semaphorin, other members of the family were rapidly identified in many animal species (Kolodkin et al., 1993, 1995; Luo et al., 1993, 1995; Puschel et al., 1995), allowing the semaphorins to be grouped into subclasses based on sequence and structural similarity (Kolodkin et al., 1993; Messersmith et al., 1995; Puschel et al., 1995; Giger et al., 1996). While the founding semaphorin family member and other homologous invertebrate transmembrane semaphorins subsequently identified were classified as class 1 semaphorins, secreted semaphorins were also identified in invertebrates and were classified as class 2 semaphorins (Kolodkin et al., 1993). However, the vast majority of the semaphorins identified in the initial flood of cloning papers reported the identification of secreted vertebrate semaphorins which are classified as class 3 semaphorins (Kolodkin et al., 1993, 1995; Luo et al., 1993, 1995; Puschel et al., 1995). Semaphorin structure A l l semaphorin family members have in common a 500 amino acid semaphorin domain, conserved cysteine residues and potential N-glycosylation sites (semaphorin nomenclature committee, 1999). However, unlike the transmembrane class 1 semaphorins which contain no obvious motifs outside of the semaphorin domain, class 2 and 3 semaphorins are secreted glycoproteins that possess a C2 type immunoglobulin domain in addition to the semaphorin domain. Class 3 semaphorins also have a short highly basic positively charged domain at the carboxyl terminus (Kolodkin et al., 1993; Luo et a l , 1993; Messersmith et al., 1995; Puschel et * semaphorin (abbreviated SEMA) is derived from the word \"semaphore\", meaning to convey information by a al., 1995). The function of secreted semaphorins in growth cone guidance has been addressed most extensively and have been demonstrated to function as chemorepulsive axon guidance cues. Funct ional studies Class 1 semaphorins: function unclear Although the first semaphorin discovered, Sema-la, was implicated to play a role in the guidance of the T i l pioneer neuron' pair in the grasshopper peripheral nervous system (Kolodkin et al., 1992), it was unclear from the studies whether this transmembrane molecule was acting as an attractive or repulsive cue for the T i l neurons (Fig. 1.2). The aberrant pathfinding o f the T i l neurons in the region o f the Sema-la band induced by antibody perturbation of endogenous Sema-la could be due to a number of possibilities: 1. It is possible that Sema-la normally acts as an inhibitory cue which prevents the proximal extension of the T i l neurons, resulting in the sharp ventral turn of the T i l growth cones upon encountering the Sema-la band. This is supported by the observation that the T i l neurons are able to extend proximally past the band when Sema-la is functionally blocked by 6F8 monoclonal anti-Sema-la antibody. 2. Alternatively, Sema-la may be functioning as an attractive cue for the T i l growth cones. This hypothesis is supported by the observation that upon contacting the Sema-la epithelial band, the T i l growth cones turn and migrate strictly within the Sema-la band (Fig. signaling system often used in rail and maritime transportation * Neurons in the embryonic grasshopper limb bud are named according to the limb segment in which they instance the T i l neurons are the first neurons to arise in the tibia limb segment. * Pioneer neurons are the first neurons to arise and establish the first or \"pioneering\" projections upon whi axons of later arising neurons often follow (Bate, 1976). 8 1.2). Additionally, when Sema-la is functionally blocked, the T i l growth cones are able to extend outside o f the Sema-la band after its ventral turn. Other than the studies presented for transmembrane Sema-la, little was known about the function o f any other transmembrane semaphorins at the time that this thesis project began in 1995, and the function of transmembrane class 1 semaphorins remained unclear. A role for secreted semaphorins in chemorepulsion The first connection of the semaphorin family to chemorepulsion was made when Luo et al. (1993) independently identified collapsin 1 (a secreted class 3 semaphorin which has now been renamed Sema3A) as the first molecularly characterized growth cone collapsing factor (Fig. 1.3 A ) . Sema3 A was initially purified and cloned from chick brain on the basis of a sensory growth cone collapsing activity which was present in chick brain membranes. Sema3A is a potent inducer o f sensory growth cone collapse which is able to induce the collapse o f dorsal root ganglia growth cones at picomolar concentrations in vitro. Fol lowing this initial demonstration of a potential repulsive activity for Sema3A, a flurry of investigations ensued which clearly demonstrated that the class 2 and 3 semaphorins are chemorepellent guidance molecules with important roles in nervous system development (see Marke t al., 1997) (Fig. 1.3). Class 2 semaphorins: cliemorepellents in the invertebrate nervous system In Drosophila, a secreted form of semaphorin, Sema-2a (previously Sema II) is expressed during embryonic development by a subset of neurons in the C N S (Kolodkin et al., 1993). Loss of function mutations in the Sema-2a gene resulted in mutant flies with increased lethality and 9 DRG GROWTH CONE GROWTH CONE COLLAPSE B Sema3A COATCD BEAD DRG GROWTH CONE Figure 1.3. Schematic diagram illustrating the classic experiments which indicate that secreted class 2 and class 3 semaphorins are inhibitory guidance molecules. A . Seina3A/Collapsin 1 induces the collapse of DRG growth cones in vitro at concentrations as low as 10 picomolars. B . DRG growth cones turn away from beads coated with Sem3A/Collapsin 1 in vitro after filopodiai contact, demonstrating that inhibitory guidance cues could steer growth cones by localized inhibition of lamellipodial protrusion. C. Sema3A is expressed at high levels in the ventral half of the spinal cord (gray shading). Axons of NT3 responsive group la muscle sensory afferents (M) project to the ventral region of the spinal cord. Axons from the NGF-responsive cutaneous sensory afferents (C) are excluded from the ventral half of the spinal cord where Sema3A is expressed. In a collagen gel assay, COS cells expressing Sema3A as well as explants from the ventral half of the spinal cord repel the NGF-responsive cutaneous sensory afferent growth cones while having no effect on the axons of the NT3 responsive la muscle sensory afferents. D. Diagram of the Drosophila embryo ventral muscles (oval) and motor nerves viewed in cross section. The RP3 motor neuron normally establishes synapses (black dot) with the target muscles 6 and 7 in the Drosophila embryo. Ectopic expression of Sema-2a (gray shading) on the target muscles stalls the RP3 growth cones and prevents them from forming a synapse on muscles 6 and 7. 11 behavioral abnormalities, which may reflect defects in the underlying neural circuits (Kolodkin et al., 1993) and suggests a role for Sema-2a in nervous system development. Evidence for an inhibitory activity for Sema-2a came when Matthes et al. (1995) tested the function of Sema-2a by producing transgenic Drosophila which express Sema-2a in muscles which normally do not express it (Fig. 1.3D). They showed that ectopic expression of Sema-2a by a muscle effectively prevents the contact and synaptogenesis of the motorneuron growth cones which normally innervate it, suggesting that Sema-2a can function in vivo as a selective target derived signal that inhibits the formation of specific terminal arbors (Matthes et al., 1995). Class 3 semaphorins: chemorepulsive guidance cues in the vertebrate nervous system Much of our current knowledge about the function of the semaphorins stem from functional analysis of members of the secreted class 3 vertebrate semaphorins, in particular Sema3A. The demonstration of chemorepulsion for several class 3 semaphorins has led to the speculation that the semaphorins represent a family of chemorepulsive guidance molecules. Sema3A can steer sensory growth cones without inducing full collapse Although acute application of Sema3A induces rapid growth cone collapse and arrested axon migration (Luo et al., 1993), Fan and Raper (1995) subsequently demonstrated that Sema3A can repel axons without inducing full growth cone collapse when presented from a point source (Fig. 1.3B). Brain-derived membrane extracts enriched for Sema3A and immobilized to beads effectively provided sensory neurons in culture with a localized repulsive cue capable of steering growth cones away from the beads (Fan and Raper, 1995). Importantly, this finding suggests that one mechanism by which Sema3A may act to direct growth cone 1 2 steering is by causing local collapse of specific regions of the growth cone which have come into contact with the source of Sema3A, effectively directing growth cone extension in another direction. Sema3A functions in vivo to pattern axon projections in the vertebrate spinal cord Sema3A has been implicated to play an important role as a selective diffusible chemorepellent in the patterning of the vertebrate C N S during development (Messersmith et al., 1995; Luo et al., 1995, Wright et al., 1995; Puschel et al., 1996; Behar et al., 1996; Giger et al., 1996). In the mammalian nervous system, the projection of neurons in the spinal cord is highly stereotyped. For instance, sensory axons from the dorsal root ganglia enter the dorsal horn and extend ventrally to different termination sites within the spinal cord which are characteristic for different classes of neurons (Fig. 1.3Ci). The N G F responsive nociceptive neurons which are involved in thermoreception and nociception, normally enter the dorsal horn of the spinal cord and terminate in the dorsal-most laminae. However, the NT3 responsive l a muscle spindle afferents terminate in the ventral spinal cord. Previous studies by Fitzgerald et al. (1993) demonstrated using a collagen gel assay that the ventral spinal cord from rat E14-E15 embryos secretes a long range diffusible factor that inhibits the growth of the N G F responsive axons from the dorsal root ganglia while having no effect on the NT3 responsive axons. Messersmith et al. (1995) implicated Sema3A in mediating this effect by demonstrating that Sema3A is expressed at high levels in the ventral part of the spinal cord during the time of D R G outgrowth (Fig. 1.3Ci). Additionally, C O S cells secreting Sema3A in vitro were shown to mimic the chemorepulsive effect of the ventral spinal cord explant on the D R G sensory neurons (Fig. 1.3Cii), suggesting that Sema3A is a selective chemorepellent which is important in the 13 patterning of the spinal cord during development (Fitzgerald et al., 1993; Messersmith et al., 1995; Puschel etal . , 1996). This hypothesis is supported by findings by Behar et al. (1996) who showed that in mice with a targeted deletion in the Sema3A gene, some sensory axons projected into inappropriate regions of the spinal cord where Sema3A is normally expressed. Defects were also observed in other areas of the nervous system as well as other organs, including the bones and the heart, and most mice died a few days postnatally. Additional investigations by Taniguchi et al., (1997) indicated that deletion of the SemD gene in mice also results in abnormal and excess branching in the periphery, consistent with the protein exerting a chemorepulsive guidance role in vivo. Sema3A is not just a repellent for developing sensory axons Recent investigations involving gene gun transfection of Sema3A into corneal epithelial cells of adult rabbits demonstrated that Sema3A repels established sensory axons (Tanelian et al., 1997). This finding indicates that small-diameter sensory neurons retain the ability to respond to Sema3A into adulthood. Collagen gel coculture assays further demonstrate that in addition to developing and adult sensory neurons, Sema3A can act as a selective chemorepellent for several populations of spinal motor neurons and a subset o f cranial motor axons (Varela-Echavaria et al., 1997). Coupled with extensive analysis of Sema3A expression (Wright et al., 1995; Giger et al., 1996; Shepherd et a l , 1996), all of these data suggest that a number of populations of embryonic and adult neurons require Sema3A for establishment and possibly maintenance, of their appropriate patterns of connections. 14 Sema3A is important in the patterning of cortical projections in the vertebrate CNS Recently, slice overlay experiments have indicated that Sema3A is important in the establishment of cortical projections during development (Polleux et al., 1998). When cortical neurons are plated over slices of cortex, Sema3A expressed in the marginal zone of the cortical slice directs the extension of axons from plated neurons toward the ventricular surface. Additionally, ectopic Sema3A repels cortical axons, suggesting that Sema3A is a repulsive guidance molecule that is important in regulating the growth of cortical axons by directing cortical growth cones towards the ventrical. This finding is further supported by the observation that in Sema3 A null mice, cortical projections are abnormal with aberrantly oriented apical dendrites and axons (Behar et al., 1996; Polleux et al., 1998). Sema3A is important in the development of olfactory neurons A number of recent studies have also suggested an important role for Sema3A in the development of olfactory neurons. Kobayashi et al. (1997) demonstrated that Sema3A induces collapse of embryonic olfactory receptor neuron growth cones in culture. Additionally, the expression pattern of Sema3A is consistent with a role as a chemorepulsive molecule in the patterning and maintenance of the primary olfactory pathway (Giger et al., 1996; Sheperd et al., 1996; Williams-Hogarth et al., 1997; Yoshida et al., 1997). Differential expression of Sema3A after axotomy and bulbectomy suggests that Sema3A may also have a role in preventing the regeneration o f primary olfactory axons after bulbectomy (Pasterkamp et al., 1998). 1 5 Other class 3 semaphorins also have chemorepellent activity Using the in vitro collagen gel assay, a repulsive activity was also indicated for several other members of the class 3 semaphorins (Puschel et al., 1995; Varela-Echavarria et al., 1997). For instance, the related molecules, Sema3B, Sema3C and Sema3F (previously SemA, SemE, and SemalV) have been shown to act in vivo as repulsive cues for the growth cones of sympathetic neurons, but have no effect on sensory axons (Adams et al., 1997; Feiner et al., 1997; Koppel et al., 1997). Receptor Neuropilins are the receptors for class 3 semaphorins Recently, neuropilin, an axonal glycoprotein previously characterized as a homophilic cell adhesion molecule in the vertebrate nervous system, was identified as the receptor for several members of the secreted class 3 semaphorins (see Kolodkin and Ginty, 1997). Neuropilin was originally isolated by Takagi et al. in 1991 as a 140kDa type 1 membrane protein known as A 5 which is expressed on subsets of neurons in the developing Xenopus nervous system and is highly conserved among vertebrates including Xenopus, chicken, and mouse. Previous work suggested that neuropilin functions in axonal growth and guidance due to its dynamic expression by a variety of different classes of axons in Xenopus, mouse and chick, and its ability to promote neurite outgrowth in vitro (Takagi et al., 1991, 1995; Satoda et al., 1995; Kawakami et al., 1996). Cel l aggregation assays showed that neuropilin can mediate cell adhesion through a heterophilic molecular interaction (Takagi et al., 1991, Takagi et al., 1995). In Xenopus, neuropilin is expressed in the olfactory axon subclasses and had been suggested to 16 play a role in selective fasciculation of the axons (Satoda et al., 1995). Additionally, Kitsukawa et al. (1995) found that transgenic mice with forced ectopic expression of neuropilin had defects in axonal fasciculation and aberrant sprouting of PNS fibers. Thus, initial observations had suggested that neuropilin likely functions as a cell adhesion molecule which mediates axon-axon interactions, with an integral role in axon fasciculation and guidance during pathfinding. The identification of neuropilin as a semaphorin receptor was accomplished by two different groups (He and Tessier-Lavigne, 1997; Kolodkin et al., 1997) using C O S cell expression cloning. Both groups identified a clone from E l 8 rat D R G which binds with high affinity to alkaline phosphatase conjugated Sema3A. This clone was identified as Rat neuropilin-1 and found to bind with high affinity to all members of the Sema3A subclass (Feiner et al., 1997). Additional evidence suggesting that neuropilin-1 is a semaphorin receptor came from antibody perturbation studies (He et al., 1997; Kolodkin et al, 1997) which showed that antibodies directed against neuropilin-1 prevent the collapse o f D R G growth cones in the presence of Sema3A. Also, anti-neuropilin-1 antibodies prevented the repulsive action of Sema3A on D R G axons in collagen gel assays (He and Tessier-Lavigne, 1997; Kolodkin et al., 1997). Importantly, D R G growth cones from mice lacking an intact neuropilin-1 gene do not collapse upon application of Sema3A, strongly suggesting that neuropilin-1 is a functional receptor for the repulsive activity of Sema3A (Kitsukawa et al., 1997). A n interaction between Sema3A and neuropilin-1 is also implicated from in vivo studies demonstrating that both Sema3A and neuropilin-1 mutations result in identical axonal projection defects (Kitsukawa et al., 1997; Taniguchi et al., 1997). Together, these results demonstrated that neuropilins are the receptor or at least an essential component of the receptor for class 3 semaphorins. 17 Neuropilin-1 likely belongs to a family of receptors since a second neuropilin, neuropilin-2 was also identified in rat (Kolodkin et al., 1997; Chen et al., 1997). While neuroplin-1 appears to be a receptor for Sema3A, recent reports indicate that neuropilin-2 is a receptor for other class 3 semaphorins. Neuropilin-2 is expressed in a population of neurons that respond to Sema3F (previously Sema IV) and expression of neuropilin-2 is both necessary and sufficient to produce a collapse response to Sema3F (Giger et al., 1998). Neuroplin-2 has also recently been reported to function as a receptor for Sema3B and Sema3C (previously SemaA and SemaE respectively) (Takahashi et al., 1998). Thus, it is likely that different neuropilins mediate the activities of different class 3 semaphorins. Neuropilin structure Neuropilins are type one membrane proteins (Fig. 1.4). The extracellular portion of neuropilins contain three domains referred to as al/a2, b l /b2 , and c that are postulated to mediate protein/protein interactions, followed by a transmembrane domain and a short cytoplasmic domain. The al/a2 complement binding ( C U B ) domains have similarity to motifs found in the complement components C l r and C l s . The bl /b2 coagulation factor domains are similar to domains found in coagulation factors V and VIII. The c M A M domain is similar to M A M domains found in a diverse group of proteins that includes four receptor tyrosine phosphatases and the meprin metalloendopeptidases (Beckmann and Bork, 1993). In contrast to the size and complexity of the extracellular domain, the cytoplasmic domain of neuropilin-1 and neuroplin-2 is small, approximately 40 amino acids in length. Because the cytoplasmic domain has no obvious conserved structural motifs or similarity to other known proteins, the mode of semaphorin signaling is not obvious through examination o f the primary structure of the 18 C L A S S 3 S E M A P H O R I N S N E U R O P I L I N S SIGNAL SEQUENCE a l a 2 CUB SEMAPHORIN DOMAIN b l COAGULATION FACTOR BASIC DOMAIN b 2 lg DOMAIN M A M TRANSMEMBRANE DOMAIN Figure 1.4. Structure of the class 3 semaphorin and its receptor, neuropilin. Class 3 semaphorins contain a 500 amino acid semaphorin domain followed by an lg domain and a basic tail. The extracellular part of neuropilins contain three unique domains referred to as al/a2, bl/b2, and c. The al/a2 domains are similar to complement CIr/s homology domains (CUB), the bl/b2 domains are homologous to coagulation factors V and VII (FV/VIII), and the c domain is also known as the M A M domain. 20 neuropilins, and the possibility exists that neuropilins represent only a component of a receptor complex. Recently, the receptor for Sema3A has been implicated to consist of neuropilin-1/plexin complexes (Takahashi et al., 1999; Tamagnone et al., 1999). Additionally, a protein that interacts with the cytoplasmic domain of neuropilin-1, neuropilin-1-interacting protein (NIP), has been identified (Cai and Reed, 1999). NIP contains a single P D Z domain and a C-terminal acyl carrier protein domain. Thus, it is possible that NIP may potentially participate in the regulation of neuropilin-1 signaling. A number of initial studies were aimed at determining both the physical and functional specificities of Sema3/neuropilin interactions. Binding studies indicate that while both the N as well as the C terminus of Sema3A bind to neuropilin expressed on C O S cells, the carboxy tail binds neuropilin with higher affinity than the semaphorin domain (Feiner et al., 1997). Additional binding studies with neuropilin deletion constructs demonstrated that the C U B domain is necessary for binding of the semaphorin domain and confers semaphorin specificity while the coagulation factor domains bind the lg domain of all class 3 semaphorins (Giger et al., 1998; Nakamura et al., 1998). The M A M domain has been implicated to contribute to the oligomerization of neuropilin (Nakamura et al., 1998). Chimeric constructs indicate that the transmembrane and cytoplasmic domains of neuropilin are not necessary for neuropilin mediated growth cone collapse, suggesting that neuropilin likely interacts with a second signal-transducing transmembrane molecule (Nakamura et al., 1998). Other potential receptor components for class 3 semaphorins have not yet been identified. 21 Transmembrane semaphorins B y 1995, only a few years after the initial discovery of the semaphorins, a handful of secreted vertebrate class 3 semaphorins had been isolated and functionally analyzed. However, while the biological functions of the secreted semaphorins were rapidly unfolding, nothing was known about the function o f the transmembrane members. In fact, until recently, transmembrane forms of semaphorin were represented only by the class 1 semaphorins found in invertebrates (Kolodkin et al., 1992, 1993). However, in a second wave of cloning papers presented over the past few years, a large number of transmembrane forms of semaphorin were identified in mouse, rat, human, and chick (Luo et al., 1995; Inagaki et al., 1995; Adams et al., 1996; Hal l et al., 1996; Furuyama et al., 1996; Zhou et al., 1997; X u et al., 1998; Sato et al., 1998). The identification of new and unique transmembrane semaphorins in vertebrates resulted in the formation of four new subclasses, resulting in a total of seven semaphorin subclasses to date (Fig. 1.5)*. Structurally, all members of the semaphorin family have a highly conserved 500 amino acid extracellular semaphorin domain which is highly conserved between invertebrates and vertebrates. There is more sequence diversity outside of the semaphorin domain. The class-specific C terminus of the semaphorins is diverse and often contains additional sequence motifs. While secreted semaphorins contain an Ig domain that is C-terminal to the semaphorin domain, transmembrane semaphorins can have an Ig domain, type 1 thrombospondin repeat, or no obvious domain motif N-terminal to their transmembrane domain. * With the rapid numbers of semaphorins pouring into the databases, the semaphorins are now grouped into seven subclasses, and recently, a unified nomenclature for the semaphorins has been developed to standardize the naming of the increasing numbers of semaphorins (Semaphorin nomenclature committee, 1999). 2 2 Figure 1.5. Structural schematic of the Semaphorin family. The Semaphorins are a family of secreted, transmembrane and GPI linked glycoproteins. A l l semaphorins contain a -500 amino acid semaphorin (sema) domain but have a class-specific C terminus that may contain additional sequence motifs, allowing the semaphorins to be classified into seven subclasses. 2 4 Recent identification of vertebrate transmembrane semaphorins Class 4 semaphorins The first transmembrane semaphorin identified in vertebrates was mouse Sema4C (previously Sema F) (Inagaki et al., 1995). Structurally, Sema4C is similar to class 1 semaphorins because it contains a transmembrane and cytoplasmic domain in addition to the semaphorin domain. However, it differs from the class 1 semaphorins in that it contains an additional lg domain adjacent to the semaphorin domain which is not present in the invertebrate class 1 semaphorins, resulting in a new subclass, class 4. Hal l et al. (1996) subsequently identified an additional class 4 semaphorin member when they cloned and sequenced human C D 100. C D 100 was originally isolated in a monoclonal antibody screen for molecules expressed on the surface of activated T cells (Bougeret et al., 1992). Interestingly, C D 100 is found as a homodimer on the surface of T cells and the addition of anti-CD 100 antibodies triggers T cell proliferation, suggesting that C D 100 is a cell surface receptor. The finding that C D 100 is a class 4 semaphorin (now renamed Sema4D) led to the first indication that members of the semaphorin family may also have a role in the immune system. Sema4D has also been independently isolated from mouse (Furuyama et al., 1996) and is found to be expressed in the developing nervous system as well as organs of the immune system. Unlike other semaphorin family members which have high sequence conservation only in the semaphorin domain, class 4 semaphorins share approximately 94% identity in the cytoplasmic domain, suggesting that the intracellular domain of these transmembrane semaphorins may be of significant importance. Interestingly, a number of unique features of transmembrane Sema4D indicate that it may function both as a ligand as well as a receptor. Sema4D expressing T cells as well as cells transfected with Sema4D are able to induce B cell 25 aggregation, suggesting that Sema4D acts as a ligand in the immune system. However, antibody cross-linking o f Sema4D on the surface of resting T cells induces T cell proliferation, suggesting that Sema4D may also act as a receptor. Additional studies revealed several consensus sites for serine and tyrosine phosphorylation within the cytoplasmic domain. Recent studies have also shown that a kinase activity can be immunoprecipitated with Sema4D (Elhabazi et al., 1997). These findings have led to the speculation that at least some transmembrane semaphorin members may be capable of bidirectional signaling. Whether bidirectional signaling by semaphorin family members is functionally significant in the developing nervous system is currently unknown. Class 5 semaphorins Class 5 semaphorins, consisting of Sema5A and Sema5B (previously SemF and SemG respectively), were identified in mouse and contain a N-terminus semaphorin domain, no Ig domain and most notably, a novel domain with seven thrombospondin repeats followed by a short 40aa cytoplasmic domain. The thrombospondin repeats are type 1 thrombospondins, which have been shown to promote neurite adhesion and outgrowth (Neugebauer et al., 1991; O'Shea et al., 1991; Osterhout et al., 1992). This suggests that class 5 transmembrane semaphorins may potentially act as chemoattractive guidance cues. Sema5A and Sema5B are expressed abundantly in both the C N S and the periphery of the mouse embryo (Adams et al., 1996). Although the function of class 5 semaphorins is currently unknown, recent studies have demonstrated that the human S E M A 5 A gene is deleted in patients with Cri-du-chat syndrome 26 suggesting that haploinsufficiency for S E M A 5 A may result in abnormal brain development * leading to the features of this disorder . Class 6 semaphorins Another novel transmembrane semaphorin recently identified is mouse Sema6A (previously Sema V i a , Zhou et al., 1997), a semaphorin with structural similarity to Sema-la in invertebrates. Like the invertebrate class 1 semaphorins, class 6 semaphorins are transmembrane molecules with an N-terminal 500aa semaphorin domain and no lg domain, but unlike the insect Sema-la, it has a larger intracellular domain. The cytoplasmic domain of class 6 semaphorins contains several proline-rich potential SH3 domain-binding sites. In vitro binding assays have demonstrated that Sema6B binds specifically to the SH3 domain of c-src. Additionally, C O S cells which have been transfected with Sema6B coimmunoprecipitate with c-src, suggesting that like the class 4 semaphorins, class 6 semaphorins may potentially function as receptors (Eckhardt et al., 1997). However, the function of Sema6B either as ligands or as receptors is currently still unknown. Class 7 semaphorins Recently a novel semaphorin, Sema7A (previously Sema K l ) was identified which is associated with cell surfaces via a glycosylphosphatidylinositol (GPI) linkage (Xu et al., 1998; Sato et al., 1998). Interestingly, Sema7A is highly homologous to viral semaphorins and binding studies indicate that Sema7A binds specifically to several immune cells lines, suggesting a role * Cri-du-chat syndrome is a human contiguous gene deletion syndrome resulting from hemizygous deletions of the short arm of chromosome 5 (Lejeune et al., 1963; Niebuhr, 1978a; Overhauser et al., 1986). A diagnostic feature of 27 in the immune system. However, its abundant expression in the brain and spinal cord suggests that Sema7A is also likely to play an important role in the nervous system. Function of transmembrane semaphorins: unknown To date, the semaphorin family includes at least 19 different vertebrate members and at least 3 different invertebrate members (Semaphorin Nomenclature committee, 1999). With the recent identification of transmembrane semaphorins in vertebrates (Inagaki et al., 1995; Adams et al., 1996; Hal l et al., 1996; Furuyama et al., 1996; Zhou et al., 1997; X u et al., 1998; Sato et al., 1998), approximately one half of all mammalian semaphorins isolated are transmembrane proteins and sequence and structural data suggest that the transmembrane forms may potentially have diverse and interesting functions. Most of the functional studies to date have focused on the secreted class 3 semaphorins, in particular Sema3A, and demonstrated that these semaphorins act as chemorepulsive guidance cues (see Mark et al., 1997), leading to the generalization that the semaphorins represent a family of secreted and transmembrane repulsive guidance molecules. A t the time that this thesis project began in 1995, none of the transmembrane semaphorins had been functionally characterized. However, their novel sequence, structure, and expression indicated that the transmembrane semaphorins may have a unique and important role in the developing nervous system. Thus, it was important to investigate the function o f transmembrane semaphorins in axon guidance. Additionally, most o f the information acquired from functional studies of axon guidance molecules had been through the investigation o f events which occur in vitro, and in vivo observations in general were lacking. It was also important to this disorder is a high-pitched monochromatic cry that resembles the mewing o f a cat with developmental delay and 28 investigate the role of a guidance molecule in an in vivo environment. Many of the investigations in this thesis take advantage of the many benefits of the invertebrate nervous system to address fundamental questions regarding the function of a transmembrane semaphorin, Sema-la in the developing grasshopper nervous system. The model system: The Developing Grasshopper Nervous System Why study the invertebrate nervous system? In the field of axon guidance, it is important to be able to study and experimentally manipulate guidance events in an in vivo context. Although popular and powerful vertebrate genetic organisms such as the mouse are invaluable for molecular and genetic research, the sheer size and complexity of the vertebrate nervous system makes it difficult to impossible to attempt to dissect the precise and intricate individual events that occur during nervous system development. Thus, the less complex nervous system of invertebrates are invaluable model systems for answering fundamental questions about nervous system development. Additionally, neurogenesis in invertebrates is readily accessible to molecular, genetic, and physical manipulations and investigations. Economically, invertebrates are usually much less expensive to breed and maintain with the added advantage of conveniently short developmental time periods, allowing large numbers of invertebrates such as insects to be easily studied per experiment. Importantly, from an ethical viewpoint, invertebrates possess little or no pain receptors (Boyan and Bal l , 1993), making the use of invertebrate systems for research much more humane than the use of vertebrate systems such as the mouse. Thus, the use of severe mental retardation (Niebuhr, 1978b). 29 invertebrates as a tool for research has a large number of benefits and invertebrates are becoming the systems on which rapid progress is being made in the field of developmental neurobiology. Fortunately, the genes which are found in vertebrates are often highly conserved in sequence as well as function in invertebrates. Thus, findings made in invertebrate systems are usually relevant to the more complex vertebrate systems as well , and effectively, many of the questions about nervous system development can be best answered by studying invertebrate systems (see Boyan and B a l l , 1993). The grasshopper, an orthopteran of the genus Schistocerca, is now becoming recognized as a very useful model organism for developmental neurobiologists (see Sanchez et al., 1995). The grasshopper provides many of the benefits of an invertebrate system including the accessibility and simplicity of its nervous system, with the added advantage of the large size and ease of artificial culture of the embryo. However, the greatest advantage of the grasshopper embryo as a model system for nervous system development is the large size and accessibility of its easily identifiable neurons which makes them simple to locate and experimentally manipulate in vivo. Thus, the grasshopper model system allows for in vivo and in situ manipulations of the live embryo during embryonic development, providing a very powerful in vivo system with the ease and simplicity of an in vitro system. Additionally, because of its increasing popularity for neurobiological research, a relatively detailed background knowledge of neuronal development, staging, structure and function have been established to which researchers can easily refer (Bentley et al., 1979; Goodman, 1982; Keshishian and Bentley, 1982a,b). Studies of growth cone guidance in the grasshopper have focused not only on the C N S , but also on the P N S , most notably the sensory neurons in the developing grasshopper limb bud (Fig. 1.6). 30 3 | Figure 1.6. Schematic representation of the grasshopper limb bud. The developing grasshopper limb bud is a hollow structure consisting of several layers including an external layer of ectodermal epithelium, a basal lamina, and an inner mesoderm layer. The axon eel bodies arise from the epithelium and lie between the epithelium and the basal lamina. The grasshopper Grasshopper eggs are laid by the female into sand in clutches of 20 to 100 embryos, each enclosed in a tough protective eggshell. Grasshopper embryonic development occurs entirely within the eggshell, and upon hatching, the insect resembles a miniature form of the adult. Post-embryonically, the grasshopper undergoes several nymphal stages (also referred to as instars) before reaching adulthood. A l l o f the embryos in one clutch are at approximately the same developmental time point. This is useful for researchers because different eggs from the same clutch can be used with the knowledge that they are all at similar stages of development. The rate of embryonic development is highly dependent upon temperature. The optimal temperature for development is 33°C, with development significantly retarded at lower temperatures. The length of embryonic development, corresponding to the time period between the time that the eggs are laid and the time that the eggs hatch is approximately 20 days at 33°C. Thus, the time period for embryonic development can be conveniently divided into percentage of development. B y definition, embryos on the day eggs are laid are at 0% development while embryos on the day of hatching are at 100% of development. This allows the embryos to be staged as a percentage of total embryonic development corresponding to 5% o f development occurring every 24 hours. The precise age o f an embryo can be easily confirmed by direct examination o f the external morphology of the embryo using a very detailed and effective staging system described by Bentley et al. (1979). Thus, live unstained embryos can be rapidly and accurately staged over the entire period of embryogenesis. This is useful for researchers because it allows embryos to be examined at precise developmental time points. Another very useful property of the embryonic grasshopper is that although development of the embryo normally occurs within the amnion and eggshell, it is possible to remove the entire 33 embryo from the eggshell and amnion for in vitro culture in a specially developed grasshopper culture medium (see appendix C) . In culture, embryonic development occurs normally, and the embryos could be cultured for at least 48 hours. This corresponds to 10% of embryonic development. Thus embryos can be experimentally manipulated using in vitro conditions for at least a period of 10% of embryonic development. One very useful manipulation of the embryo which can be performed when the embryo is in culture is called the embryonic limb fillet (Lefcort and Bentley, 1987) (Fig. 1.7). The limb fillet involves securing the limb of the embryo onto an adhesive poly-l-lysine coated coverslip. Once secured, the limb can be dissected open to reveal the peripheral neurons which are developing within the limb. This procedure effectively makes the individual neurons of the nervous system accessible for manipulation. Neurons and cell bodies can be physically manipulated once exposed by the limb fillet procedure. Additionally, cells and beads expressing recombinant proteins can be artificially introduced into ectopic regions o f the limb to examine the effects of potential guidance molecules on axon development. Most of the important pathfinding events in the formation of the grasshopper nervous system occur during embryonic development (Goodman, 1982; Keshishian and Bentley, 1982a,b). Fortunately, the neurons and axons of the embryonic grasshopper can be easily visualized during development using several different methods. The availability of excellent neuron specific monoclonal antibodies which recognize insect neurons allows researchers to examine the axon pathways at every stage of embryonic development using standard immunohistochemistry. A n antibody which is commonly used to visualize insect neurons is anti-horseradish peroxidase antibody (Jan and Jan, 1982). Antiserum against horseradish peroxidase recognizes a neural specific carbohydrate moiety of glycoproteins found on the surface of insect 34 3b Figure 1.7. Schematic diagram summarizing the limb fillet procedure. The limb fillet procedure is an operation of the limb bud which makes the neuronal cell bodies and axons accessible for experimental manipulations. The embryo is severed at the thorax and the anterior half of the embryo is removed. The isolated limb bud is secured onto a poly-L-lysine coated coverslip and the posterior side of the limb is cut open with a glass needle. The limb bud is opened and pinned down flat on the coverslip. A suction pipette is used to remove the mesodermal cells overlying the neurons to expose the neurons. 36 neurons (Snow et al., 1987) and effectively labels the entire surface of the neuron, including the somata, the axons, as well as the growth cones and filopodia. A n t i - H R P antibody can be used to label the surface o f all axon pathways in the central and peripheral nervous of insect embryos and has proven to be a powerful marker in studies o f neuronal development in insects (Jan and Jan, 1982). In addition, immunohistochemical analysis of the grasshopper embryo is very simple as the C N S and PNS pathways are readily identifiable in whole mount preparations. Thus, whole intact embryos can be fixed, stained, and visualized using microscopy. Also , live unstained neurons and their axons can be identified in vivo with interference contrast optics allowing for relatively easy labeling, injection or physical manipulation. The Grasshopper nervous system The nervous system of the grasshopper is relatively simple and most of nervous system development occurs during embryogenesis. The central nervous system of the grasshopper contains approximately 800,000 neurons, consisting of a brain at the anterior end and a segmentally repeated chain of ganglia located ventrally along the longitudinal body axis of the animal (Goodman, 1982). Each ganglion contains only 500 to 3000 neurons, and their axons form a ladder-like chain along the length of the embryo. Many of the neurons are easily identifiable because of their stereotyped position and pattern of neuronal projections. The head contains the optic lobes, a supraoesophageal ganglion which forms the brain, followed by the suboesophageal ganglion. The neurons of the C N S first delaminate from the epithelium at 20% of embryonic development and eventually establish the C N S . Although embryogenesis of the grasshopper C N S has been documented in great detail, it is the peripheral nervous system of the grasshopper which is of greatest potential interest for 37 axon guidance studies (Bate, 1976; Bentley and Keshishian, 1982; Keshishian and Bentley, 1983a,b). The grasshopper peripheral nervous system consists of a number of proprioreceptors and surface mechanoreceptors and chemoreceptors which are served by a number of highly stereotyped nerve trunks and branches composed of 10 to 100 motor axons and thousands of afferent axons. Most of the peripheral axon tracts in the limb are generated during embryogenesis. The highly stereotyped nerve tracks are generated by the initial directed projections of pioneer neurons (Fig. 1.8A) which establish the first axon pathways in the limb bud during development (Bate, 1976; Keshishian, 1980); as development proceeds, the nerves are eventually formed by the axonal projections of later arising neurons which fasciculate with the primary axon tracts (Fig. 1.8B). The strict pathfinding events which occur in the developing grasshopper limb bud early in neurogenesis has been the focus of numerous studies aimed at elucidating some o f the fundamental mechanisms in pathfinding. The grasshopper limb bud: The peripheral nervous system The peripheral nervous system of the developing grasshopper limb bud has become a popular tool for axon guidance studies because the early axon tracts which form in the limb are highly stereotyped, easily identifiable and accessible, and can be manipulated in situ using the limb fillet procedure (Lefcort and Bentley, 1987). The limb bud initially evaginates from the embryonic body at approximately 20 to 25% of embryonic development. The limb bud is a hollow structure which consists of an outer ectodermally derived columnar epithelium lined by a thick basal lamina with an inner layer of mesodermal cells which later form loose mesenchyme, the precursors for muscles and other limb structures (Bentley and Keshishian, 1982) (Fig. 1.6). 38 Figure 1.8. Schematic diagram of the initial axon paths in the developing grasshopper peripheral nervous system. A . The T i l pioneer neuron pathway at 35% of embryonic development. The T i l pioneer neuron pair arise at the distal tip of the limb bud. Their growth cones migrate proximally to the central nervous system along a highly stereotyped route. Other neurons which arise early in development along the T i l pathway include the Fe l , T r l , and C x i neurons. B. By 45% of embryonic development, the initial axon tract in the limb bud contains six neurons and two scolopidial organs (SGO and FCO); the stereotyped axonal trajectories and projections between these neurons give rise to the 5B and 5B1 nerve branches (circled by dashed lines) of the limb. Limb segment boundaries are indicated by arrows. 40 During development, the limb bud elongates and segmentally differentiates, eventually forming five leg segments, the coxa, trochanter, femur, tibia, and tarsus (Bentley et al., 1979) . At 30% of development, the first neurons differentiate from the epithelial layer to lie between the epithelium and the basal lamina, forming the pioneering axon tracts to the C N S (Keshishian, 1980). Axonogenesis and extension of the pioneering axons occurs directly on the epithelial layer, suggesting that the pioneering axons in the limb likely respond to molecular guidance cues which may be expressed by the epithelial cells. Initial axon pathfinding in the limb occurs early during embryogenesis and later arising axons fasciculate with the pioneering axon tracts (Keshishian and Bentley, 1983a,b; Klose and Bentley, 1989). Later in development, glial cells enwrap the nerves, and the axon bundles eventually lose their association with the epithelium and relocate to the interior of the limb. In the adult nervous system, only 10% of the axon length is established as a result o f axon pathfinding as the majority o f axon growth during limb elongation occurs through passive elongation of the axon due to subsequent leg growth after the axon tract has already been established (Keshishian and Bentley, 1983a). The majority o f axon pathfinding events in the limb occur early in development when the distances that must be traversed by the growth cone are relatively short. For example, the T i 1 pathway which forms early in development is initially less than 250 pm in length (Keshishian, 1980). 41 Early events in PNS development Although PNS development occurs throughout embryogenesis, the most important pathfinding events in the developing grasshopper limb bud PNS occur early in neurogenesis (corresponding to 30 to 45% of embryonic development), when the first or \"pioneering\" axon tracts are established in the limb. The axons of pioneer neurons which arise early in neurogenesis play an important role in the formation of nerve tracts by providing distinct pathways for the axons of later developing neurons to follow when the embryonic environment is much more complex (Klose and Bentley, 1989). The follower neurons make specific contacts with the pioneer axons and form bundles based on the process of selective fasciculation. The first neurons to arise in the limb bud peripheral nervous system are the pair of T i 1 neurons which differentiate from a single mother cell in the limb tip epithelium at 28% of embryonic development and begin to stain positively with antibodies against the H R P epitope by 30%> o f embryonic development (Fig. 1.8A). A s the limb elongates and segmentally differentiates, the T i 1 neurons are progressively displaced from the apical end of the limb bud and are eventually located in the tibia at the tibia-femur limb segment boundary. The T i l neurons undergo axonogenesis at 30%) of development and establish a highly stereotyped axon tract across the limb into the C N S by 35%o of embryonic development. This initial pathway formed by the T i l neurons is called the pioneering pathway because it is the first pathway which is established in the PNS and it is also the pathway upon which the axons of many of the neurons which arise later in development fasciculate with in order to reach the C N S . * Peripheral neurons which arise in the limb are named according to the segment in which their somata are located followed by a numeral indicating the order in which it appears (e.g. Ti l = tibial 1, the first neurons to arise in the tibia). 42 The T i l neurons establish a highly stereotyped pioneering process in the limb. The T i l growth cones initially extend proximally across the femur, but upon reaching the femur-trochanter limb segment boundary, the T i l growth cones make a sharp 90° turn and continue to migrate in the ventral direction within the trochanter limb segment. The Ti 1 axons then make a second 90° turn at a ventral region of the limb and continue to migrate in a proximal direction directly into the C N S (Fig. 1.8). On their way to the C N S , the T i l neurons contact other sensory neurons called guidepost cells which arise at different positions in the limb bud along the T i l pathway. These neurons represent strong adhesive substrates for the pioneer growth cones, and the T i l growth cones often enwrap and form dye-coupled contacts with the guidepost cells (see Palka et al., 1992). Guidepost cells along the T i l pathway include the C x i , T r l , and F e l neurons. Concomitant with T i l axonogenesis, the C x i neuron pair differentiate individually in the posterior and anterior limb bud epithelium and migrate ventrally to the posterior-ventral ectoderm of the coxa-trochanter region of the limb bud, corresponding to the location of the second turn o f the T i l axon trajectory. Previous studies have indicated that the presence o f the C x i cells is essential for the second turn of the T i l axons, and in their absence, the T i l axons are unable to continue their projection into the C N S . The C x i neurons project axons directly into the C N S along the T i l pathway (Fig. 1.8). Soon after the differentiation of the T i l and C x i neurons, the T r l neuron arises at the femur-trochanter limb segment boundary corresponding to the location where the T i l axons make their first ventral turn. Although the T i l growth cones contact the T r l cells before turning in the ventral direction, cell ablation studies have indicated that the proximal migration and ventral turn of the T i l growth cones can occur even in the absence of the T r l cell. The T r l cell 43 initiates axonogenesis soon after contact by the T i l neurons and its axon fasciculates with the T i l pathway to grow into the C N S , developing into a multipolar sensory neuron (Fig. 1.8). Another neuron which appears early in neurogenesis is the F e l neuron which arises in the midanterior femur adjacent to the T i l axons between the T i l and T r l cell bodies. The F e l afferent neuron also extends an axon along the T i l pathway into the C N S . B y 38% of development, a nerve composed of six axons arises from the fasciculated axonal projections of the T i l and C x i cell pairs and the follower afferent neurons F e l and T r l (Fig. 1.8). Specialized junctions which pass Lucifer yellow, a 457-MW dye, are formed between the cells in this chain. Another notable event which occurs early in the developing limb bud PNS is the development of two scolopidial organs: the femoral chordotonal organ (FCO) and the subgenual organ (SGO) which also jo in the T i l pioneering pathway (Fig. 1.8B). The F C O arises dorsally in the femur at 35%> of development and its axons project ventrally, contacting the F e l cells before entering the T i l pathway to extend into the C N S . A t 38%> of embryonic development, the S G O arises distally in the limb bud epithelium and its axons enter the C N S by fasciculating with the T i l pathway at the region of the T i l cell bodies. In the adult animal, the F C O and S G O function as stretch transducing structures. Together, the axon tract formed by the neurons which arise in the first 10%> of neurogenesis (which include six neurons and two scolopidial organs) eventually give rise to a major nerve trunk in the limb. The role of Sema-la in axon pathfinding Sema-la, the founding member of the semaphorin family, is a transmembrane glycoprotein which has been implicated to play an essential role in the formation of the grasshopper T i l pioneer pathway (Fig. 1.2). However, while antibody blocking experiments 44 suggest that Sema-la expression in the epithelium is necessary for the characteristic turn of the T i l axons, the studies provide very little insight into the potential mechanism of action of Sema-l a in axon pathfinding. The aberrant phenotypes induced by blocking Sema-la may be interpreted in a number of ways and whether it is functioning through repulsion like the secreted semaphorins, or attraction cannot be discerned. Other than our limited knowledge o f Sema-la function, nothing was known about the function of any of the other transmembrane semaphorins in both the vertebrate and invertebrate nervous system at the time that this thesis project began in 1995. Additionally, chemorepulsion was only demonstrated for a subset of secreted semaphorins, thus it was important to determine whether all members of the semaphorin family, including the transmembrane forms, possess a similar guidance activity. 45 Objectives The formation of a functional nervous system is a complex and intriguing process. With the discoveries within the past decade of novel gene families with a potential role in axon guidance, we are now just beginning to dissect the molecular mechanisms underlying the complex process of axon guidance. The semaphorins are a recently identified gene family which have been implicated to function as axon guidance molecules. However, while new members of this family in both vertebrates and invertebrates have been rapidly identified, there is very little understanding of the function of most of the members of this large family. While the semaphorins were initially identified as a family of repulsive guidance molecules, evidence for chemorepulsion was limited to only a few secreted semaphorin members and the function of the rest of the semaphorins, including the transmembrane members, was largely unknown at the time that this thesis project began in 1995. The aim of this thesis is to examine the role of transmembrane Sema-la, the founding member of the semaphorin family, in establishing the highly stereotyped axon tracts in the peripheral nervous system of the grasshopper embryo with the aspiration that this w i l l provide some insight into the function of a transmembrane member of the semaphorin family. We hope that by understanding the cellular and molecular mechanisms of potential axon guidance molecules such as the semaphorins, in the development o f the relatively simple and highly accessible nervous system of the grasshopper, we wi l l contribute some answers to some o f the fundamental questions of how the intriguing and complex nervous system is generated. 46 Chapter 2: Differential expression of Grasshopper semaphorin-la in the developing limb bud epithelium and relationship to developing peripheral axon tracts during neurogenesis In this section I describe the developmental expression pattern of Sema-la in the embryonic grasshopper limb bud. Additionally, the correlation of Sema-1 a expression with developing axons in the PNS is established in an attempt to elucidate the functional significance of Sema-la in the development of axon tracts. The data from this section are crucial in establishing the studies outlined in the ensuing chapters. Introduction: One important mechanism by which axons are directed to their targets during neurogenesis is axon based guidance: the selective fasciculation of extending axons with preexisting axon tracts which were established early in development (Raper et al., 1984; Bastiani et al., 1984; Goodman et al., 1984). However, a look one step back leads to the question of how the initial axon tracts are established. In the developing grasshopper P N S , the growth cones o f neurons which arise early in neurogenesis migrate upon a relatively axon free epithelium. Previous studies have indicated that the mesoderm and the basal lamina o f the embryonic l imb bud do not provide the necessary guidance information to direct the stereotyped projections o f the initial axonal trajectories (Lefcort and Bentley, 1987; Koefoed, 1987; Condic and Bentley, 1989), suggesting that the most likely candidate for guidance information comes from molecular cues which may be expressed on the surface of the epithelial cells upon which the neurites migrate (see Goodman, 1996). Preliminary studies have indicated that the potential axon guidance molecule, Sema-la is expressed by distinct bands of epithelial cells in the limb and 47 may be an important guidance molecule for at least one of the early axon tracts (Kolodkin et al., 1992). The first nerve tracts which are established in the peripheral nervous system of the grasshopper limb bud are the 5B and 5B1 nerves (Keshishian and Bentley, 1983a) (Fig. 1.8). The 5B nerve provides the primary innervation of the leg, and its anterior branch, the 5B1 nerve, serves as a primary sensory nerve to the limb. Both nerves 5B and 5B1 are initially established by the pioneering projection of the T i l neuron pair as they project from the distal tip of the limb bud into the C N S between 30 to 35% of embryonic development, establishing the first axon tract in the limb (Bate, 1976; Keshishian, 1980). During the ensuing 10% of embryonic development, a total of six neurons and two scolopidial organs develop in the peripheral nervous system, and the stereotyped projections between these neurons eventually give rise to the 5B and 5B1 nerve branches of the limb, all established between 30 and 45% of development (Keshishian and Bentley, 1983a). Kolodkin et al. (1992) had previously reported that at 34.5% of development, Sema-la is expressed in three circumferential stripes o f limb epithelium in the tibia, femur, trochanter, and coxa. Further investigations demonstrated that one of the three bands of Sema-la expressing epithelial cells provides guidance information for the T i l neurons and is essential for its highly stereotyped pathfinding (Kolodkin et al., 1992). The importance of the expression pattern of Sema-la at 34.5% of development in the pathfinding of the T i l neurons leads to the speculation that Sema-la expression at other developmental time points may also play a key role in the guidance of the other early peripheral axon trajectories which are normally established between 30 and 45% of development (Keshishian and Bentley, 1983a). However, little is known about the spatiotemporal expression of Sema-la in the limb during this critical period of axonogenesis. 48 Thus, an essential initial step is to establish the developmental expression pattern of Sema-la in the limb bud epithelium and its relationship to identified peripheral axon tracts in the limb. This chapter examines the spatial and temporal expression of Sema-la in the developing limb bud epithelium using immunohistochemistry with the monoclonal anti-Sema-la antibody 6F8 (Kolodkin et al., 1992). Additionally, antibody double labeling was used to determine the relationship of Sema-la expressing epithelial cells in the limb with developing axon tracts throughout the period of Sema-la expression. I show that Sema-la is dynamically expressed in bands of epithelial cells during the time period that a number of initial axon tracts are established. Additionally, the expression pattern of Sema-la in relation to the axon tracts suggests that a number of axon paths may normally be guided by Sema-la, indicating an important role for Sema-la in axon guidance. Mater ia ls and methods: Staging of grasshopper embryos Grasshopper embryos were obtained from a colony maintained at U B C . Eggs are laid into sand in glass jars. The eggs are collected from the sand and incubated on moist Kimwipes in a sealed container at 30°C. A t various stages of embryonic development, individual eggs are separated from the pod for staging to determine the developmental time point of the embryos. Eggs are immersed in saline and the dorsal tip of the egg is snipped off. The embryo is then pushed out of the eggshell into the saline. The limb morphology, antennae and pleuropodia are examined to determine the stage of development of the embryo as described by Bentley et al. (1979). The age 49 of an embryo from a single pod is usually representative of the age of the rest of the embryos in the same pod. Preparation of 6F8 antibodies Hybridoma cells which produce 6F8 antibodies were a gift of the Goodman lab, University of California, Berkeley. The hybridomas were cultured in R P M I (Gibco) supplemented with 10% fetal bovine serum (Hyclone) and I X antimycotic (Gibco) at 37°C with 5% CO2. Cells were passaged every 3 days. Supernatant was collected from the cells and tested for antibody production by SDS P A G E . Immunocytochemistry Shistocerca gregaria embryos were obtained from a colony maintained at the University of British Columbia. Embryos were staged by percentage of total embryonic development (Bentley et al., 1979). To visualize neurons and Sema-la expressing epithelium, embryos were dissected in saline, and fixed in P E M - F A (0.1 M PIPES, 2 m M E G T A , 1.0 m M MgSC-4, 3.7% formaldehyde) for 1 hour, washed for 45 minutes with 6 changes of P B T ( l x P B S , 0.1% B S A , 0.1%) Triton X-100), and incubated in primary antibody overnight at 4°C with gentle rocking. The following primary antibodies were diluted in PBT: mAb 6F8 (anti-Sema-la antibody from a hybridoma cell line kindly provided by Dr. C. Goodman; 1:1) and rabbit anti-HRP (Jackson Immunoresearch Laboratories; 1:1250). After primary incubation, the embryos were washed for 45 minutes with 6 changes of P B T and incubated for 2 hours in secondary antibody. The following secondary antibodies were diluted in P B T : HRP-conjugated goat anti-mouse IgG, HRP-conjugated goat anti-rabbit IgG, FITC-conjugated goat anti-rabbit IgG, and Cy3-50 conjugated goat anti-mouse IgG (Jackson Immunoresearch Laboratories, all 1:250). Following secondary incubation, embryos were washed for 45 minutes with 6 changes of PBT. Fluorescently labeled embryos were mounted in Slowfade antifade reagent (Molecular Probes) and viewed under fluorescence microscopy. The HRP labeled embryos were reacted in 0.5% diaminobenzidine in P B T , and the reaction was stopped by 5 serial washes in P B T . The embryos were cleared overnight with 70% glycerol in PBS and mounted in 100% glycerol and viewed using a light microscope. For double labeling, embryos were incubated simultaneously in the two primary antibodies overnight, washed, and the 6F8 antibody labeling was visualized first by incubating in secondary antibody H R P conjugated goat anti-mouse for two hours, then washed and reacted in diaminobenzidine solution containing 0.06%) nickel chloride. Embryos were then incubated in the H R P conjugated goat anti-rabbit secondary antibody for 2 hours and the second antibody was visualized with a diaminobenzidine reaction as described above. For fluorescent secondary antibodies, embryos were incubated in the first secondary antibody for two hours, washed and then incubated in the fluorescence conjugated second secondary antibody for two hours. The embryos were washed and mounted as above for fluorescence labeled embryos. Visualization of neurons For indirect immunofluorescence studies, neurons were labeled with antibodies against horseradish peroxidase (anti-HRP), which selectively bind to insect neurons (Jan and Jan, 1982: Snow et al., 1987). Embryos were incubated with rabbit anti-HRP antibodies (Jackson Immunoresearch Laboratories) diluted 1:1000 in P B T for 24 hours at 4°C, then washed in P B T containing 1%> B S A for 4 hours. Fluorescein-conjugated goat anti-rabbit IgG (Jackson) was 51 applied at a dilution of 1:200 in P B T + 1% B S A overnight at 4°C. Embryos were then rinsed in several changes o f P B T + 1 % B S A and mounted in 90% glycerol in PBS . Slides were viewed and photographed with a Zeiss epifluorescence microscope. Antibody double labeling of Sema-] a expressing cells and developing axon tracts The period of dynamic Sema-la expression corresponds to the time period that many important axon tracts are established. The morphology of the adult nerve has been indicated to be determined by the stereotyped positioning of neurons in the differentiating limb bud and by the resultant axon paths which are established during the first 10% of peripheral neurogenesis. To visualize both the Sema-la expressing cells as well as the neurons, embryos were incubated simultaneously in both primary antibodies. Following primary antibody incubation, embryos were washed and incubated in each secondary antibody separately. For embryos labeled with fluorescent antibodies, rhodamine conjugated antibodies were used to visualize Sema-la, and fluoresein conjugated secondary antibodies were used to visualize neurons. For embryos that are visualized with H R P staining, Sema-la expression was visualized as a black staining product (developed in the presence of N i C l ) , neurons were visualized with brown staining (developed in the absence of N i C l ) . Results: Developmental time course of Sema-la expression in the developing grasshopper limb bud To better understand the possible involvement of Sema-la in the pathfinding of peripheral neurons, the embryonic expression pattern of Sema-la in the developing grasshopper 52 limb bud of over 300 embryos was examined using immunohistochemistry from 30% to 45% of embryonic development, the time period that the major peripheral nerves 5B and 5B1 are established. Monoclonal anti-Sema-la antibody 6F8 was generated from hybridoma cells and used for immunohistochemistry to examine the distribution of Sema-la in the developing grasshopper limb bud epithelium. Consistent with the hypothesis that Sema-la can act as a guidance cue, I find that Sema-l a is dynamically expressed in defined regions of the limb bud epithelium between 30% to 45% of embryonic development, corresponding to the time period that initial axonal pathways are forming. Figures 2.1 and 2.2 give a summary of the time course of Sema-la expression in the limb bud. Sema-la epithelial expression was first detected at low levels in the limb bud epithelium at 30%) of embryonic development but its first expression in detectable bands occurs at 32% of embryonic development. A t 32% of development, a faint continuous band of Sema-la expressing epithelial cells approximately 2 epithelial cells wide is detected in the trochanter. A n additional faint band is apparent at the distal tip of the limb bud epithelium. B y 33%> of embryonic development, three distinct bands of Sema-la expressing epithelial cells are detectable in the limb bud. The most intensely stained band is the band in the trochanter which is still only 2 to 3 epithelial cells in width but now expresses high levels of Sema-la. The second band, located at the mid-femur is discontinuous. The third band of Sema-la expressing epithelial cells just distal to the Ti-Fe limb segment boundary becomes more apparent. At 34%) of development, the band in the trochanter and the discontinuous Sema-la band at the mid-femur are still present and stain intensely with the anti-Sema-la antibody. The third band just distal to the Ti-Fe limb segment boundary now stains intensely and is approximately 3 53 5 - f Figure 2.1. Embryonic expression of Sema-la in the developing grasshopper limb bud. Sema-la is expressed in bands of epithelial cells as well as by the femoral chordotonal organ (FCO) in the developing grasshopper limb bud. The pattern of Sema-la expression in the limb bud at three different stages of development as visualized with HRP immunohistochemistry using mAb 6F8. Proximal is to the left, and dorsal is up. A . Limb bud at 38% of development showing bands of Sema-la expression. Multiple bands of Sema-la expressing epithelial cells are present on the limb. Most notably, a prominent band of Sema-la expressing epithelial cells (arrow) is apparent at this stage just distal to the approximate location of the tibia-femur segment boundary (Ti-Fe; arrowheads). B. At 40% of development, the Sema-la band at the Ti-Fe region is broader and staining is more intense (arrows). Also, the FCO staining is more apparent (arrowhead). C. At 42% of development, the number of Sema-la bands in the limb continue to increase. The band of Sema-la at the Ti-Fe region is still present and has become very broad at the dorsal part of the band (arrows). Scale bar, 200 pm (A and C); 175 um (B). 55 DORSAL • DISTAL 38% x 54 Figure 2.2. Schematic summary of Sema-la expression in the developing grasshopper limb bud epithelium. Sema-la is dynamically expressed on the epithelium of the developing grasshopper limb bud between 30% and 45% of development. At 30%, Sema-la is distributed homogeneously within the limb bud epithelium. By 32%, Sema-la expression is limited to defined epithelial bands. At 34% of development, a prominent Sema-la band is present at the Tr-Cx limb segment boundary. Additional bands are located at the femur (discontinuous band), Ti-Fe boundary, and Ta-Ti boundary. At 38%, the Sema-la band at the Ti-Fe boundary becomes the most prominent band in the limb. As development proceeds, the number of bands increases through the addition of new bands at the distal tip of the limb bud and existing bands widen and eventually split into two bands. 57 epithelial cells in diameter. Also at this stage, a fourth band appears at the distal tip of the elongating limb bud. At 35% of embryonic development, the band in the trochanter still expresses high levels of Sema-la and does not differ much from its phenotype at 34% of development. However, the band at the mid-femur begins to widen, now extending at least 5 epithelial cells in diameter at the dorsal end of the band. Ventrally, the band also widens but is only approximately 3 epithelial cells in diameter. The Sema-la epithelial band just distal to the Ti-Fe limb segment boundary is also more intense at this stage and the newest band which appeared at the distal tip of the limb is now very apparent. This pattern of expression is similar to that previously reported by Singer et al. (1995) for a 35% embryo. B y 36%> of embryonic development, the band in the trochanter remains unchanged, but the band at the mid-femur continues to widen. The band just distal to the Ti-Fe limb segment boundary continues to widen becoming 4 epithelial cells in diameter and staining becomes more intense. The newest band which had originated from the distal tip o f the limb bud is now located just distal to the Ta-Ti limb segment boundary and stains more intensely. At 37%o of development, the band in the trochanter is still present and remains unchanged. The band at the mid femur continues to express high levels of Sema-la. The most notable band in the limb at this stage of development is the band just distal to the Ti-Fe limb segment boundary which has now increased in width to at least 5 or 6 epithelial cells. Much like the band at the mid-femur, widening of the band seems to increase dorsally, resulting in a \" V \" shaped band. Thus, although the dorsal regions of the band are at least 5 to 6 cells in width, ventral regions of the band are only approximately 3 epithelial cells in width. The band just 58 distal to the Ta-Ti limb segment boundary also becomes wider and stains more intensely and becomes wider at ventral regions of the band resulting in an \" A \" shaped band. A t 38% of embryonic development (Fig. 2.1 A ) , the band in the trochanter still remains unchanged. The band at the mid-femur continues to widen and split. The band just distal to the Ti-Fe limb segment boundary remains the most intense band in the limb, and Sema-la expression has now increased to at least 7 epithelial cells in diameter. The band just distal to the Ta-Ti limb segment boundary continues to widen and is beginning to split from the ventral region of the band. At 39% of development, the Fe band continues to widen. The dorsal part of the Fe band which had split into two thinner bands are moving further apart as the limb elongates. The ventral portion of the Fe band continues to widen but has not yet split into two bands. The band in the trochanter is still the most prominent band in the limb and is at least 8 epithelial cells in width in the dorsal regions of the band. More ventrally, the band is at least 5 epithelial cells in width. Sema-1 a expression by epithelial cells in this band remain very high. The band distal to the Ta-Ti limb segment boundary continues to split, and at this stage, the band has split from the ventral side of the limb halfway up the band. A t 40% o f development, the band in the trochanter is still present (Fig. 2. IB) . The Fe band continues to widen with the two bands at the dorsal region of the limb growing further apart as the more ventral region of the band continues to widen and split. The band just distal to the Ti-Fe segment boundary still remains the most prominent band in the l imb and continues to widen. The band just distal to the Ta-Ti limb segment boundary has split into two but is still joined at the dorsal part of the band. 59 At 41% of development, the band in the trochanter is still present. The Fe band continues to widen, with the two bands at the dorsal region widening and separating even further from each other. The band is still connected ventrally at this point but is continuing to split into two. The band just distal to the Ti-Fe segment boundary still expresses high levels o f Sema-la and continues to widen. The band just distal to the Ta-Ti segment boundary has completely separated into two distinct bands at this stage which we name Ta band 1 (proximal band) and Ta band 2 (distal band). B y 42%) of development, the band in the trochanter is still present (Fig. 2.1C). The Fe band continues to widen as the limb elongates. The Sema-la expressing epithelial band just distal to the Ti-Fe limb segment boundary is still intense, and this remains the most intensely stained band in the limb at this time point. The band just distal to the Ta-Ti segment boundary which had spilt into two separate bands continue to widen and express high levels of Sema-la. Ta band 1 is the wider of the two bands in the Ta-Ti region and is beginning to split into two, starting at the ventral part of the band. A t 43% o f development, the band in the trochanter is still present but Sema-la expression seems to be decreasing. The band just distal to the Ti-Fe limb segment boundary continues to widen and is beginning to split into two bands. Ta band 1 also continues to split into two bands. Ta band 2 is still present but does not widen and shows no sign of splitting. A t 44% o f development, the band in the trochanter is still present but the level o f Sema-l a expression has decreased, resulting in a fainter band. The two Fe bands continue to separate further from each other but staining is also becoming fainter. The band just distal to the Ti-Fe segment boundary continues to widen, forming a distinct \" V \" shaped band that extends at least 60 10 epithelial cells a the dorsal most part of the band. Ta band 1 has now split into two bands which are still connected dorsally. Ta band 2 remains unchanged. A t 45% of development, the overall expression of Sema-la in the limb is decreasing with the band in the trochanter and the two Fe bands becoming more faint. The band just distal to the Ti-Fe segment boundary is still present but has split into two bands which are connected ventrally. Ta band 1 has split into two separate unjoined bands. Ta band 2 is still present. After 45% of development, the banding pattern of Sema-la in the limb rapidly diminishes, becoming localized to small regions of the epithelium at 50% of development. Additionally, later in development, expression of Sema-la is beginning to be detected in the developing muscle cells in the limb. Summary of Sema-la expression To summarize, Sema-la is dynamically expressed by bands of epithelial cells in stripes at intervals along the length of the l imb bud from 30 to 45% of development (Fig. 2.2). H igh levels of Sema-la are also expressed in one group of neurons known as the F C O neurons. Sema-la is not expressed by any o f the other early arising neurons and is not expressed on the T i 1 axon tract or on any of the guidepost cells along the T i l pathway. Expression in the epithelium starts at 30% o f development and quickly becomes restricted to distinct regions o f the limb, forming discreet bands by 32% of development. The limb bud elongates rapidly during the course of embryogenesis. L imb growth is accomplished by apical elongation of the limb bud accompanied by differentiation of new segments as well as elongation within each delineated segment. The expression o f Sema-la in the limb is highly dynamic: new bands continuously form distally in the limb bud, and as the 61 limb bud elongates during development, the Sema-la expressing epithelial band integrates into more proximal regions of the limb bud. Many of the bands which are formed in the limb widen as the limb elongates, forming a \" V \" shaped band and eventually becomes two separate bands. After 45% of embryonic development, epithelial staining rapidly decreases. After 48% of embryonic development, expression becomes progressively restricted to specific sets of muscles in the limb bud. Bands of staining diminish in the epithelium, but could still be found in small patches scattered in the dorsal region of the limb at 48% of development. Immunohistochemical characterization of Sema-la expression indicates that it is dynamically expressed in the limb bud epithelium between 30 and 45% of embryonic development. The timing o f Sema-la epithelial expression corresponds to the time that many initial neurons are generated in the peripheral nervous system of the limb bud, indicating that Sema-la may play an important role in the guidance of a number of these neurons within this time period. Thus, it is important to further investigate the location of Sema-la expression in relation to the location and phenotypes of identified neurons and axon tracts which are formed within this time period. Double staining for Sema-la and growing nerve fiber tracts in the grasshopper peripheral nervous system: Localization of Sema-la epithelial bands in relation to peripheral neurons in the limb The highly dynamic regionalized expression of Sema-la on the limb bud epithelium suggests that Sema-la might regulate the development of peripheral axon tracts of a number o f peripheral axons at early stages of neurogenesis. During PNS development, the formation of 62 important pioneering axon tracks occurs between 30% and 45% of embryonic development. This is the time period that axons pathfind along an axon free environment, and likely respond to molecular guidance cues in the limb bud epithelium to establish the initial axon pathways. This time period also corresponds exactly to the time period that Sema-la is dynamically expressed, suggesting that Sema-la may play an important role in the establishment of the pathways which form during this time period. As a first step in identifying the axon tracks which could potentially be influenced by this potential guidance molecule, it is important to examine the correlation between Sema-la expressing cells to identified neurons and axon paths which develop during this time period. In the second part of this chapter, I examine the initial morphogenesis of the major peripheral nerves 5B and 5B1 at the level of identified cells in comparison to Sema-la expression in the limb to define the potential role of Sema-1 a in establishing the early nerve tracts. To determine the relationship between the cellular distribution of Sema-la and the development of nerve fiber tracts, I performed antibody double staining to visualize Sema-la protein (using monoclonal anti-Sema-la antibody) and neurons (using an antibody directed against the H R P epitope which recognizes carbohydrates found on neuronal glycoproteins) in over 300 embryos. I examined the morphogenesis of each o f the 6 neurons and 2 scolopidial organs that arise between 30 and 45% of development, forming the 5B and 5B1 nerves, and examined their relationship to Sema-la expressing epithelial cells to determine whether Sema-la may play a role in the pathfinding of these neurons. 63 Relationship of six identified neurons to Sema-la expressing epithelial cells Kolodkin et. al. (1992) had previously reported that from 34.5 to 35.5% of development, • bands o f Sema-la are localized in circumferential stripes o f limb epithelium in the tibia, femur, trochanter, and coxa. One of these bands has been implicated to play an important role in the pathfinding of the T i l axons between 33% and 34% of development. I now extend these studies by examining the correlation between neurons in the T i l pathway and Sema-la expression between 30% and 45% of development (Figs. 2.3 and 2.4). At 30% of development, the T i l neuron pair arises in the distal tip of the limb bud. At this time, Sema-la is expressed at low levels homogeneously throughout the limb bud epithelium. B y 31% of development, the T i l neuron pair initiates axon outgrowth onto epithelium which expresses low levels of Sema-la. A t 32% of development, Sema-la expression has become restricted to bands of epithelial cells in the limb bud. At this time, the axons of the T i l neurons have extended proximally within the limb and reached the mid-femur. The T r l cells arise at this time point and are situated within a band of Sema-la expressing cells in the trochanter. B y 33%) of development, the growth cones of the T i l neurons have contacted the T r l cell bodies. Previous studies have indicated that at this time point, the T i l growth cones are actively exploring its environment by extending filopodia both dorsally and ventrally. Because the T r l cells are situated within a band of Sema-la expressing cells, this would indicate that the T i l filopodia are exploring Sema-la expressing epithelium at the trochanter. A t 34% of embryonic development, the T i l axons have turned sharply in a ventral direction and migrates within the band of Sema-la expressing epithelial cells before contacting the pair of C x i cells which are located exactly at the proximal edge of the same band of Sema-la expressing cells. 64 6s Figure 2.3. Embryonic expression of Sema-la in the developing grasshopper limb bud and its relationship to developing axon tracts. Embryos were double stained with mAb 6F8 and anti-HRP antibody to visualize the position of the epithelial bands of Sema-la expression in relation to neuronal cell bodies and processes. Proximal is to the left and dorsal is up. A and B. Limb bud at 38% of development showing the location of the SGO (arrowhead), which is beginning to differentiate at the distal border of a band of Sema-la expressing epithelial cells. The T i l cell bodies are located on the proximal edge of the same band (double arrowheads). C and D. At 40% of development, the SGO (arrowhead in C) has begun to send a proximal projection into the band of Sema-la (arrows). The band of Sema-la is still located between the Ti 1 cell bodies (double arrowheads) and the cells of the SGO, and the growth cones of the SGO have extended into the band (arrowhead in D). E. By 42% of development, the SGO (arrowhead) axons have contacted the Ti 1 cell bodies (double arrowheads) and have fasciculated with the pioneer pathway, exiting the band of Sema-la (arrows) at the Ti-Fe boundary. F. High magnification of a limb at 42% of development single labeled with anti-HRP shows that at this stage, the SGO axons (arrowhead) have contacted the T i l cell bodies, and fasciculated with the T i l pathway. Scale bar, 200 pm (A, E); 175 urn (C); 100 um (B, D, F). 66 DORSAL -HOMOGENOUS LOW LEVEL Sema - l a EXPRESSION -Til NEURONS ARISE AT DISTAL LIMB EPITHELIUM -HOMOGENOUS LOW LEVEL Sema-1 a EXPRESSION -Trl NEURON ARISES -Sema- la EXPRESSION LIMITED TO TWO BANDS -Til INITIATES AXONOGENESIS -Til GROWTH CONES REACH Trl NEURON -Fel CELL ARISES -DISCONTINUOUS Sema-1 a BAND APPEARS AT FEMUR -Cxi CELLS REACH VENTRAL Tr-Cx REGION -Til GROWTH CONES REACH Cx i CELLS -Fel CELL ARISES IN MID-FEMUR ALONG Til PATHWAY -Sema-1 a BAND APPEARS AT DISTAL TIP OF LIMB Cl • DISTAL CONES MIGRATE TOWARDS CNS -FCO UNDERGOES AXONOGENESIS -FCO GROWTH CONES MIGRATE VENTRALLY 38% -FCO GROWTH CONES REACH Fel AND ENTER Til PATH -Senna-1 a BAND AT Tr-Cx BEGINS TO SPLIT -SGO NEURONS DIFFERENTIATE DORSAL -SGO GROWTH CONES ENTER Til PATHWAY Figure 2.4. Schematic summary of the developmental expression pattern of Sema-la in the grasshopper limb bud epithelium and its relationship to developing axon tracts. Schematic diagrams o f the embryonic limb bud from 30% to 45% of development showing the location of epithelial bands of Sema-la expressing cells (gray stripes) in relation to the locations o f identified neurons. L imb segment boundaries are indicated by arrows. 71 B y 35% of development, the T i l growth cones have exited the band of Sema-la expressing cells, growing directly into the C N S and forming the first projection from the limb into the C N S . Relationship of the scolopidial organs FCO and SGO to Sema-la expressing cells The FCO expresses cell surface Sema-la The femoral chordotonal organ (FCO) neurons arise at 35% of embryonic development in the embryonic grasshopper limb and eventually differentiate into bipolar sensory neurons in the femur (Slifer, 1935; Debaisieux, 1938; Usherwood et al., 1968; Burns, 1974). The F C O arises at 35%o of embryonic development in the dorsal mid-femur in a location corresponding to the dorsal half of an incomplete band of Sema-la expressing epithelial cells. Unlike the other early neurons in the limb, the F C O cells express high levels of Sema-la. The number of cells in the F C O cluster increases as development progresses, and at 37% of development, the F C O initiates axonogenesis ventrally onto epithelium which does not express Sema-la. B y 38% of development, the F C O axons contact the F e l cells and fasciculate with the T i l pathway to extend into the C N S . A s the limb matures, the number of F C O cells increases to at least 30 cells, all of which continue to express high levels of Sema-la. Axons of the SGO extend into a band of Sema-la expression by 40% of embryonic development The subgenual organ neurons are bipolar sensory neurons which arise just distal to the T i -Fe limb segment boundary at approximately 38%> of embryonic development (Debaiseaux, 1938; Jagers-Rohr, 1968; Heitler and Burrows, 1977a,b). The S G O neurons eventually form the subgenual organ, a stretch-transducing structure that projects its axons to the C N S from nerve 5B1 (Burns, 1974; Campbell, 1961). To address the possibility that Sema-la plays a role in the 72 outgrowth of axons from the SGO, I first examined whether Sema-la was expressed in the developing limb bud in a manner consistent with a role in influencing the S G O projections. Since the S G O does not arise until approximately 38% of development, I examined the expression pattern of Sema-la in relation to neurons at later stages of development. At 38% of development, corresponding to the time that the S G O first arises, the limb epithelium stains intensely for multiple bands of Sema-1 a located in the coxa, trochanter, femur, tarsus and a very distinct band just distal to the approximate location of the tibia-femur (Ti-Fe) segment boundary (Figs 2.3A,B). The S G O is first identifiable at approximately 38%) of development, typically along the distal edge of the band of epithelial cells expressing Sema-la at the Ti-Fe limb segment boundary (Fig. 2.3B). The T i l cell bodies are located on the opposite side along the proximal edge of the same band of Sema-la. A t approximately 40% of development, the S G O neurons initiate axon outgrowth directly onto Sema-la expressing epithelial cells (Figs. 2.3C,D). The axons of the SGO extend into the Sema-la band at 41% of development, but the growth cones have not yet reached the T i l cell bodies. A t this stage the S G O growth cones are within filopodial contact of the T i l pioneer neuron cell bodies. B y 42% of development, the axons of the S G O have traversed the band of Sema-la expressing epithelial cells and have contacted the T i l cell bodies, fasciculating with the T i l pioneer pathway (Figs. 2.3E,F). During this stage, the band of Sema-la just distal to the Ti-Fe boundary is approximately 8-10 cells wide. The T i l cell bodies are still located on the proximal edge of the band of Sema-la, but the S G O is now partially within the band. The spatial and temporal expression of Sema-la is consistent with Sema-la playing a role in the guidance of the S G O growth cones. 73 Discussion: Expression of Sema-la is developmentally regulated Growth cone guidance is proposed to be mediated by a number of different mechanisms, including contact-mediated attraction and repulsion as well as chemoattraction and chemorepulsion at long range (Tessier-Lavigne and Goodman, 1996). Sema-la has been identified as a primary candidate for contact-mediated guidance in the developing grasshopper peripheral nervous system (Kolodkin et al., 1992). Sema-la is a transmembrane glycoprotein which is expressed during embryonic development by bands o f epithelial cells in the l imb bud where it has been implicated in guiding neurons in the peripheral nervous system (Kolodkin et al., 1992). A s a first step in correlating nerve track formation and Sema-la expression in vivo, I have analyzed the temporal and spatial expression patterns of Sema-la during grasshopper neurogenesis to determine the relationship between the cellular distribution of Sema-la in the limb bud epithelium and the development of early axon tracts. Immunohistochemistry in the embryonic grasshopper reveals a widely distributed and dynamic expression pattern of Sema-la protein in distinct epithelial bands in the limb bud (Fig. 2.4). Embryonic expression of Sema-la in the developing limb bud epithelium extends over a protracted period of time. The first expression begins at 30% of embryonic development and continues past 45% of embryonic development in the limb. Whereas epithelial expression is first evident at 30% of embryonic development as a homogeneously distributed protein, the expression pattern is highly dynamic, forming distinct epithelial bands by 33% of embryonic development. Staining becomes more intense and banding patterns constantly change during development with maximum expression around 40% of embryonic development. A s 74 development proceeds and the limb elongates, an increase in the number of bands of Sema-la expressing epithelial cells in the limb bud epithelium is also apparent. After 45% of development, expression rapidly declines in the epithelium but becomes increasingly more apparent in the muscles. Sema-la epithelial expression correlates with a vole in the guidance of early arising neurons in the developing grasshopper limb bud The dynamic expression of Sema-la in distinct bands throughout the limb bud indicates that Sema-la may contribute to the development and patterning of the initial T i l neurons as well as neurons which arise at later stages of development. To determine neurons which may potentially be guided by Sema-la during this period of dynamic Sema-la expression, the relationship o f Sema-la expressing cells to axon paths that form during development was also examined. Double staining for Sema-la and peripheral nerve tracks revealed remarkable complementarity. The period of development in which Sema-la is dynamically expressed in the limb epithelium encompasses the critical time of establishment of important nerve tracts in the limb bud. Sema-la is already present in the limb when the first neurons, the T i l pioneering neurons, arise in the limb. However, after the T i l neurons initiate axon outgrowth at 31% of embryonic development, Sema-la expression becomes limited to precisely defined bands of epithelial cells in the limb bud epithelium. In agreement with previous studies, a stripe of Sema-l a at the femur-coxa limb segment boundary corresponds precisely to a sharp turn in the T i l axon tract. Detailed examination of the relationship of the T i l axon with this band of Sema-la expressing epithelial cells indicates that the axons of the T i l neurons extend directly within the Sema-la band which is only 3 epithelial cells in diameter. Although Sema-la has been 75 implicated to play a crucial role in mediating this turn, the preference of the T i l growth cones to extend directly upon the narrow stripe of epithelial cells is not consistent with a repulsive guidance function for this molecule. The role of Sema-la in T i l guidance wi l l be examined and discussed in chapters 5, 6 and 7. The Sema-la expression pattern suggests that another group of neurons which may be guided by Sema-la expressing epithelial cells are the S G O neurons (Debaiseaux, 1938; Jagers-Rohr, 1968; Heitler and Burrows, 1977a,b). A t 38% of embryonic development, the S G O neurons arise adjacent to a band of Sema-la expressing epithelial cells and initiate axonogenesis directly onto Sema-la expressing epithelial cells by 40% of development. The S G O growth cones normally cross the band of Sema-la expressing cells in order to contact the T i l cell bodies which are situated on the proximal edge of the Sema-la band, suggesting that Sema-la may potentially be important in guiding the S G O axons. This study provides the first complete mapping of the spatial and temporal expression of Sema-la in the embryonic grasshopper limb bud epithelium and documents the expression o f Sema-la in relation to developing peripheral axon tracts. The complex and highly dynamic expression patterns of Sema-la presented in this study are consistent with the hypothesis that Sema-la can pattern the peripheral axons in the limb bud peripheral nervous system. The temporal and spatial expression pattern of Sema-la and its relationship to emerging axon tracts suggests that Sema-la may be important in guiding the axons of the T i l and subgenual organ (SGO) neurons. The focus o f the following chapter is to examine whether Sema-la is a guidance cue for the axons of the SGO. Investigating the effects of Sema-la on the S G O axons may provide valuable insight into the function of this molecule. 76 Chapter 3 : Sema-la mediates SGO axon stalling after removal of guidepost cells but is not an inhibitory or collapsing cue for the SGO growth cones Previous studies had indicated that the proximal extension of the SGO growth cones is abruptly arrested at the Ti-Fe limb segment boundary in the absence of the Til neurons. In this section I examine whether the arrest of the SGO growth cones is due to an inhibitory/repulsive guidance activity of endogenous Sema-la. Surprisingly, Sema-la did not appear to function as a repulsive guidance molecule and the results are instead consistent with an attractive role for Sema-la. Introduction: The spatiotemporal expression pattern of Sema-la in the developing grasshopper limb bud established in the previous chapter suggests that one of the groups of neurons which may normally be guided by Sema-la are the S G O neurons. Shortly after the T i l pioneer neurons have established the first axonal projection into the C N S , the S G O neurons arise in the ventral ectoderm of the limb bud just distal to the T i l cell bodies at 38% of embryonic development. The S G O neurons undergo axonogenesis at 40%> of development and their axons project directly to the T i l somata to join the T i l pathway. Thus, like a number of other neurons which arise later in embryogenesis, the growth cones of the S G O reach the C N S by fasciculating with the pathway pioneered by the T i l neurons (Keshishian and Bentley, 1983). The S G O neurons eventually form stretch transducing structures that project their axons to the C N S via the 5B1 nerve branch in the mature leg (Debaiseauz, 1938; Jagers-Rohr, 1968; Heitler and Burrows, 1977a,b). Interestingly, Klose and Bentley (1989) previously demonstrated that elimination of the T i 1 pathway by heat shock treatment early in development prevents the proximal extension of 77 the S G O growth cones into the C N S (Fig. 3.1). They showed that heatshock treatment of embryos at 27% of development (the T i l neurons start to differentiate at approximately 28% of development) results in the failure of the T i l neurons to differentiate in a fraction of the embryos, but the differentiation of other cells which arise later in the limb is not affected. Elimination of the T i l neurons by heatshock resulted in the inability of the S G O growth cones to migrate past the approximate region of the Ti-Fe boundary, indicating that the presence of the T i l pathway is necessary for the development of the SGO projection (Klose and Bentley, 1989) (Fig. 3.1). The growth cones of the S G O were typically arrested in the approximate location of the Ti-Fe limb segment boundary. Although some circumferential spreading of the S G O processes along the Ti-Fe boundary was observed, none of the SGO growth cones extended past the region of the Ti-Fe segment boundary in the T i l free limb buds even by 55% of development (Klose and Bentley, 1989). Klose and Bentley had proposed that the presence of pioneer neurons is necessary for the \"successful migration of distal neurons over regions of peripheral tissue in which it is difficult to travel successfully, due possibly to chemical or mechanical properties of the tissue\". However, at the time, a candidate molecule which may be expressed in this region of the limb bud epithelium to mediate this phenotype was unknown. In the previous chapter, I have shown that during the time of S G O axonogenesis, Sema-l a is expressed intensely by a band of epithelial cells situated directly in the path of the SGO growth cones. The S G O growth cones extend across this band in their normal path to the C N S . Interestingly, the band of Sema-la expressing epithelial cells is situated just distal to the Ti-Fe limb segment boundary, a location which corresponds approximately to the region of the limb that the S G O growth cones are arrested in the absence of the T i l cell bodies. Since semaphorins have been implicated to have growth cone inhibitory properties, I examined the possibility that 78 DORSAL • DISTAL WILD TYPE Til REMOVAL 7? Figure 3.1. Proximal migration of the T i l growth cones is arrested in the absence of the T i l cell bodies. Wild Type. The growth cones of the subgenual organ (SGO; blue) extend proximally to fasciculate with the T i l cell bodies and extend along the T i l pathway into the CNS. T i l Removal. Klose and Bentley (1989) demonstrated that in the absence of the Ti 1 cell bodies, the growth cones of the SGO were arrested at the approximate location of the Ti-Fe limb segment boundary and prevented from migrating into the CNS. 80 the arrest of the S G O axons is mediated by endogenous Sema-la in the limb bud epithelium. I demonstrate that the S G O axons extend into a band of Sema-la expressing epithelial cells located just distal to the T i l cell bodies at approximately 40% of embryonic development. In the absence of the T i l pathway, the S G O growth cones extended into the band of Sema-la and typically stopped growing within the proximal edge of the Sema-la band. These results were surprising since they suggest that the Sema-la band may have a permissive/attractive function. Previous studies have indicated that class 3 semaphorins are repulsive guidance molecules (see Mark et al., 1997), and this study provides the first indication that a semaphorin may act as an attractive or permissive guidance cue in vivo. Mater ia ls and Methods: Imm unocytochem is try See chapter 2 Heat Shock In order to prevent the development of the T i l neurons, clutches of grasshopper embryos at 27% of development were positioned in 1.5 ml eppendorf tubes and subjected to heat shock treatment by immersion in a hot water bath at 45-47°C for 30 minutes as described previously (Klose and Bentley 1989). Heat shocked eggs and control eggs (no heat shock) were then incubated at 30°C until 45% development (approximately 4 days), fixed, and the neurons and semaphorin expressing epithelium were labeled as above. This method for removal of T i l neurons was 81 identical to the method originally used by Klose and Bentley and was highly inefficient, resulting in missing T i l neurons in only 3% of the heat-shocked limbs. Limb Fillet Grasshopper embryos at 38-43% of development were anchored ventral side down on a poly-L-lysine coated glass coverslip. The posterior side of the T3 limb was cut longitudinally with a glass needle and spread apart on the coverslip to flatten the epithelium. A suction pipette was used to remove the mesodermal cells overlying the neurons. Dil labeling of neurons For D i l labeling experiments, fillets were lightly fixed for 10 minutes and a small crystal of D i l (Molecular Probes) was placed on the S G O projection at its point of contact with the T i l neurons. After labeling, the fillets were fixed for an additional 30 minutes and then processed for anti-HRP fluorescence immunocytochemistry as described above with the exception that the solutions did not contain Triton X-100. Develop alternative method for removal of Til neurons Due to the inefficiency of the original heat shock method for removal of T i l neurons, a more efficient method for assaying the effect of removal of the T i l neurons on S G O pathfinding had to be developed. The improved method that I established involved physically removing the T i l cell bodies from a limb fillet preparation. The pioneer neuron cell bodies which are located on the basal, upward-facing surface of the epithelium were located with Nomarski optics. To remove neurons, a glass needle was used to remove only the T i l cell bodies, or both the T i l and F e l cell 82 bodies without disturbing other cells in the limb. The limbs were then cultured in freshly made R P M I culture media (supplemented with 5% heat inactivated fetal bovine serum) for 30 hours until 47% o f development. A t the end o f the culture period, the embryos were fixed and processed as above. Results: The SGO growth cones are arrested within a band of Sema-la after elimination of the Til neurons by heat shock The previous chapter had indicated that Sema-la is expressed in a band o f epithelial cells in the region o f the Ti-Fe segment boundary at the time o f axonal outgrowth from the S G O . Based on these findings, we further examined the possibility that this band o f Sema-la expression may play a role in arresting the proximal extension of the S G O growth cones by determining the location o f arrest o f the S G O growth cones in relation to the Sema-la band. We first repeated the heat shock experiments o f Klose and Bentley (1989). Similar to the previous report, we found that the T i l neurons failed to differentiate in a fraction o f the heat shocked embryos*. Approximately 18% of the embryos subjected to heat shock at 27% of development had at least one limb bud lacking T i l pioneer neurons. The limb morphology and differentiation of later arising neurons were not affected. In the absence o f the T i l cells, the F e l and T r l guidepost neurons established the initial projection toward the C N S and the later arising F C O extended along this projection (Fig. 3.2). However, as reported previously by Klose and * Heatshock treatment of embryos was performed by directed studies student Wilfred Yu . 83 Bentley, elimination of the T i l neurons prevented the proximal extension of the S G O growth cones in all of the limbs examined, both at 45% of development (Fig. 3.2A,B), as well as at 50% of development (Fig. 3.2C,D). The growth cones did not appear collapsed in any of these limbs, even as late as 50% of development (Fig. 3.2D), indicating that the arrest o f the S G O growth cones was not due to growth cone collapse. To obtain a better understanding of the possible effects of Sema-la on the growth cones of the S G O , we used fluorescence double labeling to determine the location of S G O growth cone arrest in relation to the band of Sema-la. We heat shocked 361 embryos and examined 22 limbs that did not contain T i l neurons and were successfully double labeled. In the absence of the T i l neurons, the growth cones of the SGO extended into the Sema-la band and typically reached the proximal edge of the band by 42%> of development (Fig. 3 .3A,B) and remained at this location throughout later stages examined (Fig. 3.3D,E). In each case, the growth cones of the S G O extended well into the band of Sema-la expression, usually to the proximal edge o f the band, but were never observed to leave the band. The growth cones o f the S G O did not appear collapsed while within the band of Sema-la (Fig. 3.3C). The ability of the S G O growth cones to migrate within the band of Sema-la during normal development and in the absence of the T i 1 cells suggests that it is unlikely that Sema-la may be preventing the proximal extension of the S G O growth cones by acting as an inhibitory or repulsive guidance molecule. However, the S G O growth cones are able to extend within the Sema-la band until they reach the edge of the band, but are not able to leave the band, suggesting that Sema-la may contribute to the arrest of the S G O growth cones by acting as an attractive or adhesive guidance cue. In order to test this, an alternative approach was used to remove the T i l cell bodies from the developing limb bud. 84 Em, 1 c * WKF Figure 3.2. Elimination of the Ti l pioneer pathway by heat shock prevents the proximal extension of the SGO growth cones. Heat shock treatment of embryos at 27% of development prevented the differentiation of the Ti 1 cell bodies as visualized by anti-HRP immunocytochemistry of heat shocked embryos at 45% and 50% of development. A and B . Heat shocked embryos at 45% of development. In the absence of the Ti 1 neurons, the axons of the SGO are arrested from proximal migration in the approximate location of the Fe-Ti limb segment boundary. Note that the growth cones of the SGO are not collapsed (arrowhead). C. By 50% of development, the axons of the SGO (arrowhead) are still arrested at the location of the Fe-Ti segment boundary. D. Higher magnification of C, the growth cones of the SGO (arrowhead) do not appear collapsed even at 50% of development. Scale bar, 100 um (A, B , D); 200 urn (C). 8 6 Figure 3.3. The growth cones of the S G O are arrested within a band of Sema-la expression. Heat shocked embryos were double labeled to reveal neuronal projections and the bands of Sema-la expressing epithelial cells. A and B. The growth cones of the SGO (arrowhead) are arrested at the proximal edge of a band of Sema-1 a (arrows) at 45% of development in the absence of the Ti 1 neurons. C. Higher magnification of A shows that the growth cones of the SGO (arrowhead) do not appear collapsed. Numerous filopodia and branches extend from the growth cones. D and E. A different limb showing the arrest of the axons of the SGO (arrowhead) within the proximal edge of a band of Sema-la (arrows) at 42% of development. Scale bar, 200 urn (A, B); 150 um (D, E); 100 urn (C). 88 A • ^9 c D • * Figure 3.4. Number and location of S G O neurons with axon projections in the T i l pathway. The lipophillic dye, Di l , was applied to the entire bundle of SGO axons at the point of contact with the T i l cell bodies, allowing visualization of the individual SGO neurons which have extended growth cones to the T i l neurons. A . At 42% of development, the growth cones which had initially migrated across the Sema-la band to fasciculate with the T i l cell bodies consists of the growth cones of one to two SGO neurons (arrowhead). B . HRP labeling of the same limb immediately following Di l labeling indicates that the first SGO neurons to extend axons to the T i l neurons are located at the proximal tip (arrowhead) of a large group of SGO neurons, consisting of at least 15 cells. C. Di l labeling of the entire SGO axon bundle at approximately 43% of development indicates that at least three SGO neurons (arrowhead) have extended axons to the T i l neurons. D. HRP labeling of the neurons in the limb shown in C indicates that the SGO consists of numerous cells, and the neurons which have extended axons to the T i l neurons are located at the tip of the group of SGO neurons (arrowhead). Scale bar, 100 um. 90 Figure 3.5. Removal of T i l cell bodies in the limb fillet preparation inhibits proximal extension of the S G O growth cones. Limb fillets at 38% of development were prepared, allowing the Til neuronal cell bodies to be selectively extracted. The limb fillets were cultured until 43% of development and neurons were subsequently visualized using standard immunohistochemistry. A. At 38% of development, the cell bodies of the SGO and FCO are beginning to differentiate but do not yet have any axonal processes. B . A control limb fillet following 30 hours in culture, showing that the SGO axons have contacted the Ti l cell bodies and have fasciculated with the Ti 1 pioneer pathway. Note also that the FCO has fasciculated with the Fel cell and has extended along the Ti 1 pathway. C . Removal of the Ti 1 neurons at 38% of development prevented the proximal extension of the SGO growth cones resulting in a dorsal turn within the Sema-la domain (arrowhead) when visualized 30 hours after culturing. The FCO axons have fasciculated with the Fel neuron (arrow) and have extended along its axon. D. Both the Fel and Til neurons were removed from a 38%> limb fillet before culturing. After 30 hours in culture, the SGO axons have arrested their proximal extension (arrowhead) while the FCO axons have extended ventrally past their normal turning point at the region previously occupied by the Fel cell (arrow). The FCO axons extended to approximately the ventral midline of the limb before turning proximally toward the CNS. E and F . Higher magnification of the arrested SGO growth cones in limbs where the Til neurons were removed indicates that although the SGO axons have arrested their proximal migration, their growth cones do not appear collapsed, and they have extended multiple branches both dorsally and ventrally (arrowheads). Scale bar, 100 pm. 92 The SGO projection is pioneered by the most proximal 1-3 neurons While the S G O matures into a large collection of sensory neurons (> 20 neurons), its initial projection to the T i l cell bodies appears to be pioneered by only one or a few growth cones (see Fig. 3.4). In order to determine the number of growth cones that establish the projection from the S G O to the T i l neurons, we used the limb fillet preparation to label the first axons that contact the T i l neurons. The lipophilic tracer D i l indicated that the initial projection from the S G O to the T i l neurons is pioneered by the most proximal 1-3 neurons (Fig. 3.4A,B), and as more neurons differentiate in the S G O they extend along the established projection (Fig. 3.4C, D). Proximal extension of the SGO growth cones is arrested when the Til cell bodies are removed from a limb fillet preparation A s an alternate approach to the heat shock experiments, the T i l cell bodies were physically extracted from the developing limb bud prior to the extension o f the S G O axons. Using the limb fillet preparation (Lefcort and Bentley, 1987), the T i l cell bodies were removed using a sharp glass needle. The T i l cell bodies were selectively extracted from limbs at 38% o f development, a stage at which the S G O is first beginning to differentiate, but does not yet have any processes (Fig. 3.5A). Similar to the heatshock experiments, the S G O axons extended proximally for a short distance and then ceased to grow in the absence of the T i l neurons (Fig. 3.5C, D , E , F). N o proximal extension of the S G O growth cones past the region that was previously occupied by the T i l cell bodies was observed, but the growth cones often exhibited complex morphologies, typically extending dorsal and ventral branches circumferentially (Fig. 3.5C, E , F). This indicates that the S G O growth cones are able to migrate within the band of 93 Sema-la, but are not able to grow out of the band in the absence of the T i l cell bodies as observed in the heatshock experiments. The axons of the S G O of the contralateral control limb fillets with intact T i l cell bodies fasciculated normally with the pioneer pathway and extended proximally toward the C N S (Fig. 3.5B). Removal of the T i l neurons had no effect on the extension of the F C O neurons, and the F C O extended its normal projection ventrally and fasciculated with the F e l guidepost cell. In some cases, as a consequence of the removal of the T i l neurons, the F e l guidepost neurons were removed. This resulted in the removal of the entire pioneer pathway through the femur allowing us to test whether removal of the pathway had any effects on the extension of axons from the F C O . In limbs where both the T i l and F e l cells were removed, the S G O growth cones were arrested; however, in contrast to the S G O , removal of the T i l and F e l cell bodies did not arrest the ventral extension of the F C O growth cones (Fig. 3.5D). During normal development the axons of the F C O extend ventrally, contact the F e l neuron and then extend along the pioneer pathway into the C N S (Keshishian and Bentley 1983). However, in the absence of the T i l and F e l cells, the axons of the F C O extended past the region normally occupied by the missing F e l cell body, finally making a sharp proximal turn at the ventral midline region of the limb. These results indicate that the removal of neurons from the limb does not result in cessation of neurite outgrowth in other cells in the limb, suggesting that the cessation of growth of the S G O axons is likely due to the expression of Sema-la. Both heat shock removal and physical removal of the T i l cell bodies early in development resulted in the arrested proximal migration of the SGO growth cones. Elimination of the T i l neurons by both methods resulted in the inability of the S G O growth cones to migrate past the region of the Ti-Fe limb segment boundary. The growth cones did not appear collapsed 94 in any of the limbs, even as late as 50% of development, indicating that the arrest of the S G O growth cones was not due to growth cone collapse. However, although none of the numerous S G O growth cones extended past the region of the Ti-Fe segment boundary in the T i l free limb buds, circumferential spreading of the S G O processes along the Ti-Fe boundary was observed. Antibody double labeling revealed that Sema-la is expressed by a band o f epithelial cells at the Ti-Fe segment boundary. The S G O is first identifiable at approximately 38% o f development and is located immediately adjacent to the distal edge o f the band o f epithelial cells expressing Sema-la at the Ti-Fe limb segment boundary. The T i l cell bodies are located along the proximal edge of the same band of Sema-la. Unlike the results that would be expected for an inhibitory guidance cue, the S G O growth cones normally initiate axon outgrowth onto Sema-l a expressing epithelium at 39% of embryonic development, and extend normally within the Sema-la expressing epithelium. The growth cones of the S G O did not appear collapsed while within the band of Sema-la. In the absence of the T i l neurons, the growth cones of the SGO extended into the Sema-la band and typically reached the proximal edge of the band by 42% o f development and remained at this location throughout the later stages examined. Upon reaching the edge of the Sema-la expressing band of epithelial cells, the S G O growth cones often branch and continue to migrate circumferentially within the band but do not exit it. The phenotype of the S G O growth cones in relation to Sema-la expression are consistent with an attractive/permissive role for Sema-la. Discussion: Intermediate targets are essential for the patterning of the developing nervous system in a variety of systems (Klose and Bentley, 1989; Ghosh et al., 1990; Marcus et al., 1995; Tessier 95 Lavigne and Goodman, 1996). In the developing grasshopper limb bud, the first projection towards the C N S is established by the T i l pioneer neurons between 30-34% of development (Keshishian and Bentley, 1983a). Later in development, neurons arising in distal regions of the limb bud fasciculate with and extend along the T i l pathway into the C N S . One such group of neurons, the mechanoreceptor neurons of the subgenual organ (SGO), arises at approximately 38% of development, distal to the T i l cell bodies. Klose and Bentley (1989) have previously shown that the T i l neurons are necessary to mediate the extension of axons from the S G O into the C N S . Removal of the T i l cell bodies by heat shock prevented proximal extension of the S G O axons past the approximate region of the Ti-Fe segment boundary of the limb, suggesting that unique mechanical or chemical properties of the limb epithelium in the region between the T i l neurons and the S G O were preventing the proximal migration of the S G O growth cones in the absence of the T i l neurons. Using a monoclonal antibody directed against Sema-la, I found that Sema-la is expressed in a band o f limb epithelium just distal to the Ti-Fe limb segment boundary directly between the T i l cell bodies and the S G O . During development, the axons o f the developing S G O extend through the band of Sema-la and fasciculate with the T i l pioneer pathway to enter the C N S . In limb buds where the T i l neurons were not present, the growth cones of the S G O were prevented from extending beyond the band of Sema-la and were arrested within the proximal edge of the band, consistent with a role for Sema-la in mediating this effect. However, unlike the functions of other semaphorins, this effect was not mediated by collapse or growth cone repulsion. Sema3A (previously chick collapsin 1) has previously been shown to prevent the extension o f chick sensory growth cones in vitro by inducing growth cone collapse (Luo et al., 1993). This involves filopodial and lamellipodial retraction as the actin network disassembles in 96 affected growth cones (Fan et al., 1993). In contrast, the S G O growth cones do not appear collapsed while contacting the band of Sema-la and readily extend on the Sema-la substrate. Presently, it is unclear what role growth cone collapse may play during development as there have been few observations of collapse in vivo. This is not unexpected as it would be unusual for the entire growth cone to come into contact with an inhibitory environment. Growth cones are likely deflected by high concentrations of collapsing activity due to localized collapse o f a small region o f the growth cone which first contacts the source o f inhibitory cues, effectively steering the growth cone extension away from the source of inhibitory cues before complete growth cone collapse occurs. Consistent with this, Fan and Raper (1995) have shown that Sema3A immobilized onto beads induces the turning of growth cones away from the bead following contact, rather than resulting in the complete collapse of the growth cone. Alternatively, a few filopodial contacts can result in the collapse of sympathetic neuron growth cones when they contact retinal neurites in vitro (Kapfhammer et al., 1987), suggesting that under the right conditions, a small number of contacts is sufficient for growth cone collapse. Unlike the in vitro results with chick collapsin, the ability o f the S G O growth cones to migrate within the band o f Sema-la suggests that Sema-la is not an inhibitory cue for the S G O axons, and the absence o f growth cone collapse, even after the arrest of the S G O growth cones within the band, indicates that the arrest of the proximal extension of the S G O growth cones is not mediated by growth cone collapse. Sema-la is an attractive/permissive cue for the growth cones of the SGO I have shown that the proximal extension of the S G O growth cones is arrested within a band of Sema-la expression at the Ti-Fe segment boundary in the absence of the T i l neurons. 97 The ability of the S G O growth cones to leave the band of Sema-la in the presence of the T i l neurons is likely due to a preference o f the S G O growth cones for the T i l neurons. Similarly, a band of Sema-la in the trochanter limb segment has previously been demonstrated to prevent the proximal extension of the T i l growth cones. The T i l growth cones make a sharp ventral turn when they reach the T r l guidepost cell in a band of Sema-la expressing epithelium and grow ventrally within the Sema-la band. The growth cones do not leave the band until they contact the C x i guidepost cells located at the proximal edge of the band. Laser ablation of the C x i cells prevents the T i l growth cones from making this proximal turn (Bentley and Caudy, 1983). The S G O neurons seem to respond to the T i l neurons in a similar manner. Thus, the T i l cells likely play an analogous role to the C x i cells in allowing growth cones to exit regions o f Sema-la expression. Members of the semaphorin family have previously been shown to function as inhibitory guidance cues. However, our results are more consistent with Sema-la being an attractive/adhesive cue for the S G O growth cones for a number of reasons: 1. The S G O growth cones are able to extend well into the band of Sema-la expression both during normal development and in the absence of the T i l cell bodies. 2. In the absence of the T i l cell bodies the S G O growth cones typically extend to the proximal edge of the Sema-la band but do not extend further into the femur intrasegmental epithelium. It is therefore likely that the epithelium outside the Sema-la band is not as permissive for axonal growth, thereby preventing the extension of the S G O growth cones out of the band of Sema-la. The inability of the S G O growth cones to leave the Sema-la band until contacting the T i l cells ensures that the S G O growth cones establish a projection into the C N S along the T i l pathway. 98 3. The growth cones of the S G O are complex and highly branched while within the band of Sema-la, typical of growth cones extending along an adhesive substrate (Payne et al., 1992; Burden-Gulley, 1995). Similarly, observations of T i l pioneer growth cone morphology (Caudy and Bentley, 1986a, b) and adhesive interactions (Condic and Bentley, 1989) have suggested that growth cones in the vicinity of segment boundaries, corresponding to the location of Sema-la expression, exhibit a high affinity for the epithelium. Caudy and Bentley (1986a) have shown that T i l growth cones extend more lamellae, filopodia and branches at these locations. The ability o f the S G O growth cones to initiate axon outgrowth onto Sema-la expressing cells and extend normally upon them without any signs of growth cone collapse suggests that unlike the potent inhibitory and collapsing activity of the secreted forms of semaphorin, Sema-la seems to have a different function. The phenotype of the S G O growth cones upon reaching the edge of the semaphorin band also indicate that Sema-la may have a function opposite to that of the secreted forms, possibly acting as an attractive guidance molecule. Thus this particular set of experiments leads to the hypothesis that Sema-la may have a novel attractive function. These results are consistent with an alternative function for a transmembrane form of semaphorin and may explain the previously reported arrest of the proximal extension of the subgenual organ growth cones in the absence of the T i l pioneer pathway. This hypothesis is tested by a number of different approaches in the following chapters. 99 Chapter 4: Sema- la promotes S G O axon outgrowth In this section I investigate the role of Sema-la in SGO development. By blocking endogenous Sema-la function using monoclonal Sema-la antibodies, I found that Sema-la is requiredfor axon outgrowth from the SGO as well as axon extension, further supporting an attractive function for Sema-la. Introduction: Axonal growth cones are guided to their targets during development by a wide variety of guidance cues. Generally, guidance molecules act as permissive/attractive cues, or as repulsive/inhibitory cues to effectively steer growth cones towards or away from a specific region (reviewed by Tessier-Lavigne, 1994; Goodman, 1996; Tessier-Lavigne and Goodman, 1996). While some of these cues are diffusible and mediate their effects from a distance, other cues are membrane bound and have a short range contact-mediated effect. Both cell surface and secreted forms of semaphorin have now been identified in several species and it was originally thought that both forms play an important role in axonal pathfinding by functioning as inhibitory guidance cues (Kolodkin et al., 1993; Luo et al., 1993, 1995; Messersmith e al., 1995; Puschel et al, 1995; Messersmith et a l , 19995; Matthes et al., 1995; Kolodkin et al., 1992; Luo et al., 1993; Adams et al., 1996). Transmembrane Sema-la had previously been shown to play an important role in the development of the grasshopper limb bud PNS, but its function as a repulsive axon guidance molecule in the limb was only speculative. In the previous chapter, I illustrated a set of experiments which suggested a role for Sema-la in S G O development which is not consistent with axon repulsion. Findings from the previous chapter suggest that transmembrane Sema-la 100 which is expressed by bands of epithelial cells in the developing grasshopper limb bud may function as an attractive/permissive cue for the growth cones of the S G O . This chapter further investigates the role of Sema-la in axon guidance by examining the functional importance o f endogenous Sema-la in S G O axon outgrowth and extension. The role of Sema-la in S G O guidance is analyzed by blocking endogenous Sema-la at various stages of SGO development. To perturb Sema-la function, I used the anti-Sema-la monoclonal antibody 6F8 which had previously been demonstrated by Kolodkin et al. (1992) to cause aberrant T i l pathway formation. Antibodies to Sema-la disrupted normal SGO development, effectively preventing axon outgrowth from the S G O and preventing proximal extension of the S G O growth cones onto the Sema-la expressing band. These results support the hypothesis established in chapter 3 that Sema-la is an attractive guidance molecule in the developing grasshopper P N S . Materials and Methods: Preparation of limbs and removal of Til neurons were as described in the previous chapter. Preparation of 6F8 blocking antibodies For blocking experiments, the m A b reagents were prepared as follows. Hybridoma cell lines secreting 6F8 monoclonal antibody or anti-cmyc monoclonal antibody were cultured in serum free medium. Ammonium sulphate saturated H2O was added to the tissue culture supernatant bringing it to a 25% solution, incubated on ice for one hour and spun at 150, 000 R P M at 4°C for one hour. The supernatant was then brought to 55% with H20-saturated (NH4)2S04, incubated overnight at 4°C and spun as above. The pellet was resuspended in P B S using approximately 101 1/40 volume of the original hybridoma supernatant, dialyzed against I X PBS overnight with three changes, filtered through a 0.2 pm filter, and further concentrated using centricon filters (Amicon) to a final concentration of 600 p M . The concentrated antibody was then dialyzed overnight in R P M I with three changes. For blocking experiments, each antibody reagent was diluted with three portions of R P M I supplemented with 5% heat inactivated fetal bovine serum resulting in an antibody concentration of 200 u M . Limb fillets were cultured in either m A b 6F8 or m A b anti-cmyc reagent overnight and then fixed and stained as above. Antibody blocking In order to examine the role of Sema-la in both S G O axon initiation as well as extension, embryos were collected at 38% o f development and also at additional 0.5%> developmental intervals until 41% o f embryonic development. Embryos were immediately processed using the limb fillet preparation and the T i l cell bodies were located using Nomarski optics and extracted as described in Chapter 3. Embryos were then cultured for 30 hours in the presence of blocking antibody. Visualization of neurons Following fixation, anti-HRP antibodies were used to visualize the neurons using standard immunohistochemistry. See chapter 2. 102 Results: 6F8 mAb blocking of Sema-la inhibits SGO axon outgrowth and extension The monoclonal antibody 6F8 had previously been shown to cause aberrant T i 1 pathway formation by blocking the function of Sema-la in the trochanter limb segment (Kolodkin et al. 1992). To examine whether Sema-la plays a role in S G O outgrowth, limb fillets at 38% of development were cultured in the presence of mAb 6F8. Prior to culturing, the T i l neurons were removed from contralateral limb fillets and limbs were cultured for 30 hours until approximately 43% of development. Control cultures were bathed in a monoclonal antibody to the cmyc protein. A s indicated in chapters 2 and 3, the S G O has differentiated by 38% but does not yet have any axons, allowing us to examine the role of Sema-1 a in axon outgrowth. Outgrowth from the S G O neurons was completely prevented in the majority of the limbs which were cultured in medium that contained m A b 6F8, both in the absence (Fig. 4.1 A , n=27/32), and presence (Fig. 4. I B , n= 17/21) o f the T i l cell bodies, indicating that Sema-la is necessary for initial axon outgrowth from the S G O . A l l other neurons in the limb differentiated and extended normally, and the F C O axons fasciculated normally with the F e l cells (Fig. 4.1). In the presence of the monoclonal antibody against the cmyc protein, the axons of the S G O extended normally, fasciculating with the T i l neurons in control limbs (Fig. 4 . ID, n=30/30), and in the absence of the T i l cell bodies, the growth cones were arrested (Fig. 4.1C, n=37/37). 103 Figure 4.1. mAb 6F8 blocking of Sema-la prevents axon outgrowth from the SGO. Limb fillets at 38% of development were incubated until 43% of development in the presence of either 6F8 mAb or anti-cmyc mAb and then stained using standard immunocytochemistry. A . Following incubation in mAb 6F8, the SGO axons were completely absent (arrowhead) in the limb fillets where the T i l cell bodies were removed prior to culture. Note that the FCO has differentiated and extended axons and fasciculated with the Fel cell. B. In the presence of mAb 6F8, no axon outgrowth from the SGO was apparent (arrowhead) in limb fillets with intact T i l neurons (double arrowheads), while the FCO projection extended normally. C. Incubation of limb fillets without the T i l cell bodies in the presence of anti-cmyc mAb did not prevent axon outgrowth from the SGO, but the axon was arrested from proximal migration (arrowhead) as seen in the absence of antibody. D. In the presence of anti-cmyc mAb, and with the T i l neurons intact (arrowheads), the axons of the SGO were able to fasciculate normally with the T i l pathway. Scale bar, 100 um. 105 When Sema-la was blocked later in SGO development shortly after axon outgrowth, proximal migration of the SGO growth cones was prevented. To examine the role of Sema-la in axon extension after outgrowth from the S G O has occurred, the same experiment was performed on embryos at a later stage, at approximately 40-41% of development. The precise extent of outgrowth from an individual S G O prior to antibody addition was difficult to accurately ascertain since variations in developmental age between animals from the same pod is typically +/- 1%. During this period of development there is extensive variation in the extent of S G O outgrowth, ranging from little to no axon at 40.5%) of development, to crossing of the Sema-la band and fasciculation with the T i l pathway by 42% of development. L i m b fillets at approximately 40-41%) of development with axon lengths ranging from less than 50pm to more than 100pm were cultured as above in the presence o f m A b 6F8 or anti cmyc antibody. After 30 hours in culture (44% of development) in the presence of m A b 6F8, approximately 33% (20/62) o f the limbs that were missing T i l neurons still had axons less than 50pm in length, while 94% (47/50) o f the limbs with intact T i l pathways had axons which extended further than 50pm, the majority o f which had fasciculated with the T i l pathway. In the presence o f anti cmyc antibody, 91% (31/34) o f the axons in limbs missing T i l neurons extended further than 50pm, stopping at the Sema-la boundary, and 100% (30/30) o f the limbs with an intact T i l pathway had axons which extended greater than 50pm, and had fasciculated normally with the T i 1 pathway. Although there is great variability in the lengths of the axons in this set of experiments, the key finding is that in the absence of the T i l neurons, many of the S G O axons were still less than 50pm in length after culturing in the presence of mAb 6F8, indicating that in addition to 106 stimulating axon outgrowth, Sema-la may also be needed for the proximal extension of the S G O axons on the limb epithelium after outgrowth has already occurred. Although Sema-la seems to be absolutely required for S G O axon initiation regardless of the presence or absence of the T i l neurons, the presence of the T i l neurons in the limb after outgrowth has occurred allowed the proximal extension of the majority of the SGO axons onto the T i l pathway in the presence of mAb 6F8. The ability of the T i l neurons to allow proximal extension of the SGO axons onto the T i l pathway is likely due the establishment of filopodial contact of the S G O growth cones with the T i l neurons prior to, or soon after culturing in the presence of mAb 6F8. This would allow the growth cones to extend to the T i l neurons. The results of these experiments and the heatshock experiments suggest that Sema-la may act as a high affinity substrate for the SGO growth cones (Fig. 4.2). These results also suggest that the S G O growth cones have a higher affinity for the T i l neurons and that the T i l neurons are necessary for continuous proximal extension out of the Sema-la band. This scheme of affinities is consistent with previous reports on the hierarchy of affinities in the developing grasshopper limb bud (Caudy and Bentley, 1986b). Discussion: A role for Sema-la in the patterning of the grasshopper PNS The results presented in this chapter indicate that Sema-la is a guidance molecule which plays an important role in the proper development of the grasshopper P N S . I found that Sema-la is needed for initial axon outgrowth and extension from the S G O neurons. In the presence of anti-Sema-la monoclonal antibody 6F8, axonal outgrowth was completely prevented in the 1 0 7 DORSAL • DISTAL WILD TYPE ANTI-Sema-la Figure 4.2. Schematic summary of heat shock and limb fillet experiments. Wi ld Type. The SGO axon normally fasciculates with the Ti 1 pathway to grow towards the CNS. T i l Ablation. When the T i l pathway is eliminated by heat shock, the SGO axon is prevented from extending proximally to the CNS, and is arrested within a band of Sema-la expressing epithelial cells. The growth cones are not collapsed and typically extend to the proximal edge of the Sema-la band. Anti-Sema-la. The SGO axon is missing in experimental limb fillets cultured in media containing mAb 6F8 in the absence (or presence) of the T i l neurons. 109 majority of subgenual organs both in the presence and in the absence of the T i l cell bodies. In addition, in the absence of the T i l cell bodies, growth was also arrested in a significant number of subgenual organs that had initiated an axon at the time of 6F8 antibody addition. However, while Sema-la is needed for outgrowth from the S G O , the absence of Sema-la expressing epithelium adjacent to other neurons in the limb indicates axon initiation from other neurons, such as the T i l pioneers does not require Sema-la. Additionally, although the proximal migration of the T i l axons is prevented by a band of Sema-la at the Fe-Tr segment boundary of the limb, addition of 6F8 m A b to block the function of Sema-la did not prevent the extension of the T i l growth cones onto the band of Sema-la expression, but instead allowed the growth cones to extend across the band (Kolodkin et al., 1992), indicating the requirement for Sema-la for axon extension differs between different neurons. These observations suggest that Sema-la expressed on bands of limb epithelium functions to pattern afferent projections in the limb by stalling or stopping the proximal projection of growth cones, possibly to allow other guidance cues to direct further growth. Resumption of proximal migration of growth cones out of a Sema-l a band requires contact and fasciculation of the growth cones with other neurons at the proximal edge of the band. The semaphorins are a large family of axon guidance molecules with multiple functions during development Previous antibody blocking experiments in the developing grasshopper limb bud indicated that Sema-la expressing cells at the Tr-Cx limb segment boundary are important in mediating a characteristic 90° turn in the T i l axon tract (Kolodkin et al., 1992). Although alternative interpretations are consistent this finding (Kolodkin et al., 1993), a 110 permissive/attractive role for Sema-la is supported by observations that T i l growth cones exhibit a preference for epithelium in the region of the Tr-Cx limb segment boundary (Caudy and Bentley, 1986a, b; Condic and Bentley, 1989). The results presented in this chapter also support an attractive role for transmembrane Sema-la. The finding that a transmembrane semaphorin may act as an attractive cue is interesting since it raises several intriguing possibilities. It is possible that transmembrane forms o f semaphorin may have a function which is different from that of the secreted forms. However, in light of the homology to the secreted forms of semaphorin, an alternative explanation is that Sema-la molecules may act as bifunctional guidance cues, functioning both as attractive as well as repulsive guidance cues. This has been demonstrated for members of the netrin family. Netrin 1 and unc-6 (a nematode homologue of netrin) are bifunctional molecules, acting as chemoattractants and chemorepellents (Kennedy et al., 1994; Colamarino and Tessier-Lavigne, 1995; Hedgecock et al., 1990). U N C 6, a secreted protein which is believed to form a gradient in the extracellular environment, plays a role in circumferential guidance in the nematode. N u l l mutations disrupt both dorsal and ventral migrations while partial loss-of-function mutations disrupt either ventral or dorsal migrations, suggesting that different domains of the protein direct dorsal or ventral growth cone guidance (Hedgecock et al., 1990). In addition, axons extend either dorsally or ventrally along the presumptive U N C 6 gradient depending on the particular U N C - 6 receptor expressed (Hamelin et al., 1993, Chan et al., 1996). In chick, netrin 1 secreted by the floor plate o f the neural tube has been shown to attract the growth cones o f commissural neurons whose axons extend towards the floor plate, and repel trochlear motor axons which originate near the floor plate and extend dorsally (Kennedy et al., 1994; Colamarino and Tessier-Lavigne, 1995). Importantly, these experiments provided the first evidence that guidance molecules may not simply be acting only as repulsive or attractive cues, but could act as bifunctional guidance cues, acting either to attract or repel different subsets of neurons. The semaphorin family is large, structurally diverse, and has been implicated in multiple functions during development. Initial studies have indicated that members of the semaphorin family act as repulsive guidance cues and are essential in the patterning of the developing nervous system. This chapter indicates an alternative function for a transmembrane member of the semaphorin family, Sema-la, in acting as a cell surface attractive or permissive guidance cue in neuronal development. Additionally, recent evidence has also suggested that members of the semaphorin family may have various functions in development other than its known function as a neuronal guidance molecule. Mouse knockout experiments have revealed that secreted Sema3A is essential in the normal development of heart, bones, and muscle (Behar et. al., 1996). The finding that two semaphorins are encoded in viral genomes (Kolodkin et al., 1993), as well as its expression on B lymphocytes (Hall et al., 1996) indicates that semaphorins may play a role in the immune system. Although semaphorins were initially implicated as chemorepellent cues, it is very likely that members of this family have diverse functions during development. Based on expression studies and antibody blocking phenotypes, the previous chapters have provided evidence which suggests that Sema-la is not an inhibitory or repulsive guidance molecule, but instead functions as an attractive or permissive guidance molecule. To test this hypothesis, the next few chapters are aimed at determining whether an ectopic source of Sema-l a protein can act directly as an attractive guidance cue in vivo. 112 Chapter 5: The activity of Sema-la is mediated by the semaphorin domain In this section I directly examine the function of Sema-la protein and determine that the extracellular semaphorin domain is sufficient to perturb the pathfinding of the Til axons. Ectopic soluble Sema-la allowed the Til axons to migrate into normally non-permissive epithelium, further indicating an attractive guidance function for Sema-la. Introduction: During development, elongating axons are directed to their correct targets by a combination of attractive and repulsive guidance molecules expressed in their local environment (Tessier-Lavigne and Goodman, 1996). The semaphorins are a large family of structurally diverse secreted and transmembrane proteins which are typically classified as a family of inhibitory neuronal guidance molecules (Kolodkin et al., 1993; Luo et al., 1995; Kolodkin, 1996). While in vitro studies have indicated that a number of secreted semaphorin family members are able to repel axons (Messersmith et al., 1995; Fan and Raper, 1995; Puschel et al., 1995) and cause rapid collapse of neuronal growth cones in culture (Luo et al., 1993), little is known about the function of the numerous transmembrane semaphorins which have been identified to date. However, recent findings are beginning to suggest that the function of the semaphorins may not be limited to a role as a repulsive guidance molecule (Wong et al., 1997; Bagnard et al., 1998; Adams et al., 1996; Hal l et al., 1996). Chapters 3 and 4 examined the effects of pioneer axon pathway disruption and antibody blocking of endogenous transmembrane Sema-la on S G O axon pathfinding and indicated that Sema-la may possibly have axon outgrowth promoting properties (Wong et al., 1997). Independently, a secreted semaphorin has 113 been implicated in growth cone attraction in vitro (Bagnard et al., 1998). Additionally, semaphorins have been found which contain thrombospondin repeats (a motif known to promote neurite outgrowth) in the developing mouse (Adams et al., 1996) and chick (see appendix B) nervous system. Although these findings suggest that members of the semaphorin family may have a function other than repulsion, direct evidence for semaphorin mediated attractive guidance activity in vivo is lacking. In this chapter, I attempt to directly address the role of Sema-la as an axon guidance molecule by examining the effects of a homogenous ectopic source of Sema-la protein on the highly stereotyped pathfinding of the T i l neurons. As mentioned in chapters 1 and 2, the expression pattern of Sema-la in relation to the T i l axon tract is consistent with a role for Sema-l a in mediating the highly stereotyped pioneering axon projection of the T i l neurons as they extend from the distal tip of the limb bud into the C N S . Additionally, previous studies have shown that antibody blocking of Sema-la in the limb disrupts T i l pathfinding in the region of the Sema-la band, suggesting that Sema-la is important for T i l pathfinding (Kolodkin et al., 1992). The T i l neurons first arise in the limb bud epithelium at 29% of embryonic development and initiate axon outgrowth at 30%> of development. Although the axons initially migrate proximally, upon encountering a band of epithelial cells which express Sema-la, the axons turn sharply and continue to extend ventrally within the Sema-la band (Kolodkin et al., 1992). B y 35%o of embryonic development, the T i l axon pathway is fully established and the T i l axons have traversed the limb epithelium and reached the C N S . To determine whether Sema-la can be directly responsible for axon guidance, I examined whether an ectopic source of Sema-la protein can perturb T i l pathfinding. I present in vivo data which indicate that the extracellular 114 semaphorin domain o f Sema-la actively perturbs peripheral axon pathfinding in vivo when added ectopically during development as a homogenous recombinant freely soluble factor. Mater ia ls and Methods: Recombinant Sema-la protein production R E C O M B I N A N T Sema-la C O N S T R U C T . The recombinant Sema-1 a construct is truncated at amino acid 597 and contains the entire semaphorin domain of Sema-la, missing the entire transmembrane and cytoplasmic domains and is tagged at the truncated C-terminus by a cmyc tag ( E Q K L I S E E D L L R K R R E Q L K H K L E ) . Recombinant Sema-la protein was generated using a Baculovirus expression system (Invitrogen), and the soluble protein was harvested from the viral supernatant. Western analysis using both anti-Sema-la antibody (6F8) and anti-cmyc antibody demonstrated the presence of secreted recombinant Sema-la in the viral supernatant. rSema-la was purified and concentrated using an anti-cmyc immunoaffinity column and further concentrated using Centricon filters (Amicon) to a final concentration of 1.1 mg/ml and dialyzed overnight at 4°C with three changes of R P M I . Anti-cmyc immunoaffinity column preparation Anti-cmyc antibodies were produced using hybridoma cells. Large batches of hybridoma cell culture supernatant were produced and concentrated by ammonium persulfate precipitation followed by further concentration using Centricon filters and dialyzed. The column was produced by covalent association of the anti-cmyc antibody to the protein A beads of the column. 115 Embryo culture For culture experiments, embryos at 30% of embryonic development were sterilized and dissected (Chang et al., 1992). The entire amnion and dorsal membrane were removed from the embryo to ensure access of the reagents during culturing. Grasshopper embryos at 30%> of development were incubated in varying concentrations of rSema-la protein diluted in grasshopper culture medium. For control experiments, embryos were incubated with heat inactivated rSema-la or anti-cmyc antibody. Both experimental and control embryos were further cultured at 37°C for 24 hours, fixed, and stained using standard immunohistochemistry with anti-HRP primary antibody and FITC-conjugated secondary antibody to visualize the neurons. Visualization Following fixation, neurons were visualized with anti-HRP antibodies using standard immunohistochemistry. See chapter 2. For each experiment, the scoring of the number of abnormal T i 1 pathways in each limb was confirmed independently by a second observer. A n y variations between the two observers were not statistically significant. The T i l pathway in experimental limbs were compared to the wild-type T i l pathway. The T i l pathway was scored as abnormal i f any major deviation from the normal Ti2 pathway was observed. The T i l pathway was scored as abnormal i f any of the following characteristics were observed: axon branching into dorsal limb epithelium, axon branches that migrated distally, abnormal outgrowth of the T i l axons distally and dorsally. For each concentration of rSema-la tested, error bars represent any uncertainties in distinguishing between abnormal versus normal T i 1 pathways. 116 Results: A recombinant Sema-la (rSema-la) construct was generated (Fig 5.1a, b) and added as a soluble protein to embryos in culture in order to examine the effect of a homogenous ectopic source of rSema-la on T i l pathfinding and outgrowth. The recombinant soluble molecule contains the entire semaphorin domain but is missing the entire transmembrane and cytoplasmic domains (Fig. 5.2). Using a grasshopper in vivo assay system (Chang et al., 1992) soluble rSema-la was added to embryos at 30% of development, a stage in which the T i l axon pathway has not yet been established. The limbs were cultured in the presence of the peptide until 35% of development. Immunohistochemical examination of the experimental limbs to determine the accessibility of the recombinant protein indicate that rSema-la successfully enters and distributes homogeneously within the limb bud (Fig. 5.3). At concentrations of rSema-la of 13.5 ug/ml and above (Fig. 5.4), the T i l pathway fails to form correctly in a significant number of embryos and lacks the sharp characteristic turn at the region of the endogenous band of Sema-la (Fig. 5.5B to H), suggesting that rSema-la overwhelms the Sema-la receptors on the T i l growth cone, and is capable of preventing them from recognizing endogenous Sema-la as a target. However, unlike the phenotypes expected for a repulsive or collapsing cue, rSema-la did not prevent axon outgrowth or extension, instead, the presence of rSema-la at concentrations of 13.5 ug/ml and above induced a number of unique defects in the T i l axon pathway (Fig. 5.4). T i l axon defective phenotypes were characterized by axon branching and aberrant growth into distal and dorsal regions of the limb bud epithelium (Fig. 5.5B to H), and could generally be grouped into 4 different classes of defects (Fig. 5.6). In the wild-type embryo, the T i l axons normally do not migrate into distal and dorsal regions of the limb epithelium (Fig. 5.5A), due possibly to the presence of inhibitory guidance 117 S E M A P H O R I N D O M A I N [7\\^jJ+1 SECRETED Sema3A S E M A P H O R I N D O M A I N M C Y U TRANSMEMBRANE S e m a - l a S E M A P H O R I N D O M A I N jpmyc) RECOMBINANT Sema-1 A 95 — • • m * <• —I— —I— 6F8 ant i -cmyc ANTIBODY ANTIBODY Figure 5.1. Recombinant Sema-la construct. A . Structural schematic comparing Sema3A, Sema-la, and soluble recombinant Sema-la (rSema-la) construct. rSema-la is truncated at amino acid 597 (arrow) and contains a cmyc epitope tag at the truncated C terminus. B . Immunoblot analysis of Baculovirus supernatant using both anti-cmyc antibody and anti-Sema-la (6F8) reveal a predominant 95 kDa rSema-la band (additional bands may be due to differences in glycosylation). 119 SEQUENCE OF rSema-1 a CONSTRUCT 1 mraalvavaa l l w v a l h a a a wvndvspkmy v q f g e e r v q r 61 121 181 241 301 361 421 481 541 flgneshkdh f k l l e k d h n s l l v g a r n i v y n i s l r d l t e f teqriewhss g a h r e l c y l k gkseddcqny i r v l a k i d d d r v l i c g t n a y k p l c r h y a l k dgdyvvekey e g r g l c p f d p d h n s t a i y s e g q l y s a t v a d f s g t d p l i y r g p l r t e r s d l kqlnapnfvn t m e y n d f i f f f f r e t a v e y i n c g k a i y s r v arvckhdkgg phqfgdrwts f l k s r l n c s v pgdypfyfne i q s t s d i i e g n y g g q v e k l i y g v f t t p v n s i g g s a v c a f s m k s i l e s f d g pfkeqetmns nwlavpslkv peprpgqcvn d s r t l p d v s v nfvkshtlmd e a v p a f f t r p i l i r i s l q y r f t k i a v d q q v rtpdgkaydv l f i g t d d g k v i k a l n s a s f d s s d t v d s v v i e e l q v l p p g v pvknlyvvrm dgddsklvvv s d d e i l a i k l h r c g s d k i t n c r e c v s l q d p ycawdnvelk ctavgspdws a g k r r f i q n i slgehkaccjg r p q t e i v a s p v p t q p t t k s s gdpvhsihqa e f e p e i d n e i v i g v d d s n v i 601 p n t l a e i n h a g s k l p s s q e k l p i y t a e t l t i a i v t s c l g a l v v g f i s g f l f s l r r c r g e d y 661 tdmpfpdqrh q l n r l t e a g l nadspylppc a n n k a a i n l v lnvppknang knanssaenk 721 p i q k v k k t y i SEQUENCE OF c m y c TAG EQKLISEEDLLRKRREQLKHKLE SIGNAL SEQUENCE SEMAPHORIN DOMAIN TRUNCATION _ ^ TRANSMEMBRANE DOMAIN INTRACELLULAR DOMAIN Figure 5.2. Recombinant Sema-la construct - Sequence. Amino acid sequence of rSema-la construct. rSema-la contains the entire semaphorin domain but is missing the transmembrane and cytoplasmic domains. rSema-1 a contains a 23 amino acid cmyc tag at the truncated C-terminus. Arrows indicate point of truncation. 121 Figure 5.3. Penetrance and distribution of soluble rSema-la in the limb bud during embryo culture. Distribution of rSema-la in the limb bud is detected using anti-Sema-la antibody (6F8). A . Endogenous Sema-la expression in the absence of rSema-la. B. Distribution of rSema-la in the limb after culture in the presence of 13.5 ug/ml rSema-la. 123 Figure 5.4. Ectopic rSema-la induces defects in Til-axon guidance - Histogram. Histogram showing effect of rSema-la on T i l pathfinding. Concentrations of soluble rSema-la below 9.0 ug per ml, control heat inactivated rSema-la (50 ug per ml) or anti-cmyc antibody (200 uM) alone did not induce T i l pathfinding defects compared with control embryos cultured in plain media, which normally show a baseline of approximately 7.5% defective pathways. Pathfinding defects are induced by rSema-la at concentrations of 13.5 ug per ml and above. Data shown are the averages and standard error of the mean for four to six independent experiments. The numbers above the bars represent the total number of limbs analyzed. 125 Figure 5.5. Phenofype of defects induced by ectopic rSema-la. A . In the absence of rSema-la, the T i l pathway forms normally after 24 hours in culture. B to H . Embryos cultured in the presence of 13.5 ug per ml rSema-la exhibit pathfinding defects which are typically characterized by aberrant migration of the Ti 1 growth cones into distal and dorsal limb bud epithelium and defasciculation. Pathfinding defects are noticed at all stages of the pathway, even at early stages soon after outgrowth (B, F) as well as defects in axon initiation (G, H) characterized by axon initiation from the T i l cell bodies in the wrong direction into distal limb epithelium. Fe l , T r l , and C x i are guidepost cells along the T i l pathway. Scale bar, 60 um. 127 A Wild Type B I3£ Figure 5.6. Schematic summary of rSema-la induced T i l axon pathfinding defects. Abnormal phenotypes induced by rSema-la can generally be classified into four classes of T i l pathfinding defects. A . The wild-type T i l pathway. B. Class 1 defects are characterized by incorrect turning of the T i l growth cones into dorsal epithelium at the femur/trochanter limb-segment boundary, often followed by an abrupt 180° loop in the dorsal region of the limb to continue migration in the correct ventral direction. Class 2 defects are characterized by initial axon migration in the correct proximal direction followed by an abrupt turn to continue migration in the incorrect distal direction. Class 3 defects are characterized by abnormal T i l axon outgrowth into distal limb epithelium. Class 4 defects are characterized by aberrant axon-branch formation in multiple directions in the limb. 129 100 yj so 60 40 20 i 04 and then fixed with 4% paraformaldehyde in P E M for 1 hour at room temperature. Fixed cells were washed with P B T and then processed using standard immunohistochemistry with anti-Sema-la (6F8) primary antibody, and Cy3 conjugated secondary antibody. To visualize Sema-la expressing epithelial cells on the grasshopper embryo limb bud epithelium, embryos at 35% of development were dissected out of the eggshell, the amnion and dorsal membrane are removed. The embryos are fixed with 4%o paraformaldehyde in P E M for 1 hour at room temperature. The embryos are washed with P B T and then processed using standard immunohistochemistry as above. Visualization of neurons After fixation, neurons were visualized with anti-HRP antibodies using standard immunohistochemistry. 138 Results: To determine whether Sema-la functions more efficiently in the clustered form, antibodies directed against the C-terminal cmyc tag of rSema-la were used to dimerize and cluster the ligand (Davis et al., 1994). Immunohistochemical examination of the experimental limbs indicated that the penetrance and distribution of rSema-la in the limb bud is not affected by antibody mediated dimerization or clustering (Fig. 6.1). Dimerization of rSema-la at its C-terminus by anti-cmyc monoclonal antibodies greatly potentiated the activity of the soluble recombinant protein (Fig. 6.2). Pathfinding defects induced by the dimerized molecule closely resembled the defects induced by the soluble molecule but characteristically showed increased axon branching and combinations of multiple classes of defects were observed within one pathway. Antibody mediated clustering of the dimerized molecule resulted in perturbation of T i 1 pathfinding in nearly 100% of the experimental limbs. The majority (90%) of the defects observed were characterized by massive axon branching in all directions into all regions of the limb (Fig. 6.3). Dose-response studies demonstrated that dimerized rSema-la was 3 fold more active than the non-dimerized ligand and clustered ligands were at least 10 times more active than unclustered ligands (Fig. 6.2). Additionally, responses were saturable with low concentrations of clustered ligand, perturbing T i l pathfinding in close to 100% of the limbs at concentrations as low as 2.25 ug/ml. The increased activity o f soluble rSema-la after ligand dimerization and clustering suggests that Sema-la may normally be clustered on the cell surface by the transmembrane and cytoplasmic domains and interacts with its receptor most efficiently as a ligand cluster. A n 1 EMBRYO CULTURE SUPERNATANT I SOLUBLE CLUSTERED Figure 6.1. Penetrance and distribution of clustered rSema-la in the limb bud during embryo culture. i . Embryo culture medium collected at the end of the culture period and analyzed by western blot confirm that a source of intact rSema-la, dimerized rSema-la (not shown) and clustered rSema-la are present in the culture medium throughout the embryo culture period, i i . Distribution of rSema-la in the limb bud is detected using anti-Sema-la antibody (6F8). A. Endogenous Sema-la expression in the absence o f rSema-la (same picture as Fig . 53A). B. Distribution of clustered rSema-la in the limb after culture. 1 4 1 10 20 30 40 50 60 70 r S e m a - l a (jug/ml) CLUSTERED rSema- la DIMERIZED rSema-la SOLUBLE rSema- la Figure 6.2. Dose-response curve comparing the activity of soluble, dimerized or clustered rSema-la . Grasshopper embryos at 30% of development were cultured in the presence of various concentrations of either soluble, dimerized or clustered rSema-la for 24 hours and then processed using standard immunohistochemistry. Antibody-mediated dimerization/clustering of rSema-la greatly potentiates its activity. The data are the mean and standard error of the mean for six to ten independent experiments. 14 Dorsal Distal J m Figure 6.3. T i l axon defects induced by clustered rSema-la. Abnormal phenotypes induced by clustered rSema-la. A. Defects induced by clustered rSema-la are typically characterized by aberrant pathfinding and excess branching, particularly early in the pathway. Increased axon branching into all regions of the limb. B. Schematic representation of Ti l axon defect induced by rSema-la. Scale bar, 60 pm. A B - *if • • Figure 6.4. Full-length Sema-la forms microscopically visible aggregates on the cell surface. Transfected and control non-transfected S2 cells were processed for immunofluorescent localization of Sema-la using monoclonal anti-Sema-la antibody 6F8 and Cy3-conjugated secondary antibodies. A. Nontransfected S2 cells do not express Sema-la. B. S2 cells transfected with full length Sema-la express Sema-la as microscopically visible clusters on the cell surface. C. Immunofluorescence analysis of endogenously expressed Sema-la on embryonic grasshopper epithelial cells at 35% of development indicate that Sema-la is not distributed evenly on the cell surface and is found as visible aggregates, particularly at cell borders. 147 examination of the sequence of the cytoplasmic tail of Sema-la reveals a potential P D Z binding site (T /S-X(D/E/A)-V/ I ) (Cohen et al., 1996) at the C terminal amino acids 728 to 730. The P D Z binding consensus sequence has been shown to mediate the clustering of a number of transmembrane proteins (Ponting et al., 1997). Thus, it is possible that the cytoplasmic domain facilitates ligand dimerization or clustering, possibly through interactions with molecules in the cytoplasm (such as P D Z domain containing molecules). Previous studies have indicated that some P D Z proteins with P D Z binding sites are capable o f forming clusters which are microscopically visible in heterologous cells ( K i m et al., 1995). The distribution of Sema-la on S2 cells which have been transfected with full-length Sema-la was examined at the light microscopic level using anti-Sema-la antibody 6F8. The distribution of Sema-la protein on the cell surface is not uniform (Fig. 6.4B). Instead, the protein is localized in distinct punctate regions of the cell surface. To further define the endogenous distribution of Sema-la, I examined the distribution of Sema-la on embryonic grasshopper epithelial tissue which are known to produce bands of Sema-la expressing cells. In agreement with previous studies by Kolodkin et al. (1992), Sema-la immunoreactivity detected on the epithelial cells also appeared as intensely stained aggregates (Fig. 6.4c), similar to those observed in transfected S2 cells. This indicates that Sema-la distribution both on endogenous Sema-la expressing limb bud epithelial cells as well as on S2 cells which have been transfected with full length Sema-la is not uniform. 148 Discussion: The findings in chapter 5 had indicated that although the semaphorin domain of Sema-la can actively perturb axon pathfinding, its activity is lower than expected, raising the question of whether the transmembrane and cytoplasmic domains are important for Sema-la activity. Here I provide evidence indicating that the transmembrane and cytoplasmic domains of Sema-la may be important for regulating the oligomerization state of Sema-la, and that the oligomerization state of Sema-1 a is important for its activity. Antibody clustering studies from this chapter indicate that Sema-la functions most efficiently in the clustered form. Furthermore, full-length Sema-la appears as distinct aggregates on the cell surface of transfected cells as well as endogenously on epithelial cells in the grasshopper limb bud. The additional observation that Sema-la contains a putative P D Z binding site within its cytoplasmic tail suggests that membrane attachment may facilitate ligand oligomerization through interactions with P D Z proteins in the cytoplasm. Dimerization of semaphorins The hypothesis that semaphorins do not signal in isolation is also supported by a number of recent studies which indicate that several semaphorin family members form dimers. Klostermann et al. (1998) recently demonstrated that dimerization of the secreted semaphorin, Sema3A, by a cysteine residue at amino acid 723 is essential for its chemorepulsive activity. Similarly, Koppel and Raper (1998) have shown recently that the sema domain of Sema3A alone is sufficient for chemorepulsive activity once dimerized. They demonstrate that truncation of the carboxy terminus of Sema3 A results in a considerable decrease in activity. Surprisingly, the activity of Sema3 A can be restored by addition of the carboxy terminus o f another semaphorin 149 family member or the addition o f the lg domain o f IgG, suggesting that dimerization mediated by the carboxy terminus domain is important for potentiating Sema3A activity. Moreover, experiments in which portions o f the semaphorin domain are swapped between family members demonstrate that the semaphorin domain alone controls the specificity of the biological response (Koppel et al., 1997). This echoes the findings from chapters 5 and 6 which indicate that Sema-1 a activity is also specified by the semaphorin domain, and suggests that the semaphorin domain alone is sufficient to specify the activity of semaphorin family members, and that the C terminus domains are important for regulating its level of activity. The transmembrane semaphorins Sema4A (previously SemB) and Sema4D (previously C D 100) have also been demonstrated to form homodimers. However, because the function of vertebrate transmembrane semaphorins is currently unknown, it is uncertain whether dimerization is important for the function of these molecules (Herold et al., 1995; Hal l , et al., 1996). Given the findings in this chapter that oligomerization of an invertebrate transmembrane semaphorin, possibly by P D Z interactions, is important for regulation of its activity, it would be important to determine whether any of the numerous vertebrate transmembrane semaphorins are also clustered on the cell surface by P D Z interactions. Perhaps one method for regulating the activity of transmembrane semaphorins is through cytoplasmic proteins which may regulate the oligomerization state of the semaphorin. The role of oligomerization in regulating the activity of Sema-la Oligomerization can increase the number of ligand types: increased versatility The ability of the extracellular domain of Sema-la to actively perturb T i l pathfinding suggests that the activity o f Sema-la is likely mediated by the semaphorin domain. However, 150 the activity o f truncated rSema-la is greatly increased by antibody mediated clustering. This indicates that aggregation of the Sema-la ligand, which may normally be mediated by the transmembrane and cytoplasmic domains, may be necessary for the molecule to bind to and efficiently activate its receptor. A number of semaphorin receptor candidates have now been identified and include members of the neuropilin (He and Tessier-Lavigne, 1997; Kolodkin et al., 1997) and plexin (Winberg et al., 1998) family, both of which are implicated to likely form homo or heteromeric receptor complexes. It is thus tempting to speculate that the Sema-la receptor is a receptor complex, and that binding to multiple semaphorins may serve to cluster and activate the receptor. Since semaphorin family members homodimerize and as we show in this chapter, functions more efficiently in the clustered form, the interesting possibility arises that different semaphorin family members may form heteromeric complexes that have distinct biological activities. Heteromeric interactions would add a significant degree of versatility to the potential signaling role of the semaphorin family. The semaphorin family is large, currently with at least 20 different vertebrate members. Thus, the formation of both homo and hetero oligomer combinations could potentially generate a large number of different ligand complexes with potentially different biological activities. It would be important to determine whether the semaphorins could form heterodimers in vivo. Koppel and Raper (1998) discussed this phenomenon using two members of the transforming growth factor (3 signaling family, the inhibin and activin. While inhibin proper is generated by the heterodimerization of one inhibin a and one inhibin P subunit, activin is formed from the homodimerization o f two inhibin P subunits. Both inhibin and activin are functionally distinct. Inhibin functions in suppressing the secretion of follicle-stimulating hormone while activin enhances the secretion o f follicle-151 stimulating hormone (Ying 1989, Kingsley, 1994; Knight, 1996). Thus, it is also possible that different semaphorin members may form functionally distinct dimers or oligomers. Role of Oligomerization in Regulating Activity Given our finding that monomeric Sema-la has limited activity, while dimerized Sema-l a has greater activity and clustered Sema-la has greatest activity, it is tempting to speculate that oligomerization can serve to regulate the function of some transmembrane semaphorins. The identification of a P D Z binding domain in the cytoplasmic domain of Sema-la suggests that oligomerization of some transmembrane semaphorins is regulated by a cytoplasmic protein. A recent study now indicates that Sema 5A, a transmembrane class 5 semaphorin is also localized in microscopically visible clusters on the cell surface. Interestingly, a P D Z binding motif is also present in the cytoplasmic domain of Sema 5A. The P D Z containing neural cytoplasmic proteins S E M C A P - 1 and S E M C A P - 2 (SemF cytoplasmic domain-associated protein) were identified and implicated to cluster and alter the distribution of Sema 5 A in the brain (Wang et al., 1999). Although the functional importance of clustering on the function of Sema 5 A is currently unknown, it is tempting to speculate that clustering may be important in regulating the function of Grasshopper Sema-la as well as other transmembrane semaphorins. 152 Chapter 7: Ectopic Sema- la actively attracts T i l growth cones in vivo All of the studies presented in the previous chapters strongly indicate an attractive guidance activity for Sema-la. In order to show that Sema-la is an attractive guidance molecule which can directly steer axons, I presented Sema-la expressing cells in localized regions of the limb epithelium during embryonic development of the grasshopper PNS. In this final chapter, I demonstrate that Sema-la can directly guide the Til axons by acting as an attractive cue and indicate that the phenotype of the Til pathway can be artificially controlled using ectopic Sema-la. Introduction: The previous chapters have addressed the function of Sema-la using a number of different approaches. The results obtained strongly suggest an attractive guidance function for Sema-la and also provided insight into the importance of several structural domains of the molecule as well as possible methods to regulate its activity. However, in this final chapter, I would like to return to the key objective of the thesis which is to determine the function of Sema-l a in axon guidance. Although data presented in the previous chapters suggest that Sema-la may have an attractive function, it is important to investigate whether Sema-la can indeed act directly as an attractive guidance molecule which can steer growth cones. The phenotype seen after T i 1 pathway disruption in chapter 3 gave the first indication that Sema-1 a is able to guide growth cones in a manner other than repulsion and growth cone collapse, as was previously observed for the secreted semaphorin members. Chapter 4 further addressed the function of Sema-la and indicated that antibody perturbation of Sema-la prevents axon outgrowth, 153 suggesting that Sema-la may be an attractive guidance molecule. Further investigations into the function of Sema-la in chapters 5 and 6 examined the effects of a soluble recombinant Sema-la protein on T i l pathfinding. Preincubation of the limb with soluble truncated recombinant Sema-1 a allowed the T i 1 axons to extend aberrantly into normally non-permissive regions of the limb bud epithelium, strongly suggesting an attractive function for ectopic Sema-la. However, it is also possible that the soluble ectopic recombinant Sema-la is binding to and perturbing the activity of repulsive molecules which are present in the limb, such as Sema-2a. This possibility has recently been presented by Takahashi et al (1998) who demonstrated that SemA and SemaE could antagonize SemD binding to its receptor and effectively attenuate the repulsive activity of SemD. Thus, it is possible that the excess soluble rSema-la may be acting as an antagonist to an endogenous repulsive activity o f an inhibitory semaphorin family member. In order to investigate whether Sema-la can act directly as a contact mediated guidance molecule, this chapter examines whether a localized non-soluble source of Sema-la can steer neurons once differentially expressed ectopically on restricted regions of the limb bud epithelium in vivo. Drosophila S2 cells transfected with transmembrane Sema-la were used to provide an ectopic source of Sema-la in the embryonic grasshopper limb bud epithelium early in development to examine its effects on T i l axon pathfinding. When an ectopic source o f Sema-l a is localized onto specific epithelial regions, I find that its effect on axon growth is mediated through a strong attractive activity which can actively guide neurons when differentially expressed. Taken together with the previous chapters, these data provides extensive evidence for an attractive guidance activity for Sema-la and this chapter demonstrates that Sema-la can function directly as a guidance molecule. 154 Materials and Methods: Construction of COS Cell Expression Vector Ful l length grasshopper Sema-la c D N A in the A K B 72 plasmid was a gift o f Dr. A lex Kolodkin. Sema-la was directionally subcloned into the mammalian expression plasmid p B K C M V (Stratagene) in P s t l - X h o l . S2 CELL TRANSFECTION The. Sema-la expression plasmid A K 7 4 (Kolodkin et al., 1992) was used to express Sema-la in S2 cells (Schneider, 1972). A K 7 4 contains the full-length Sema-la c D N A cloned into pRmHa-3 (Bunch et al., 1998) to allow the expression of Sema-la under the control of the inducible metallothionein promoter. A K 7 4 was cotransfected into Drosophila S2 cells along with the plasmid pPC4 (Jokerst et al., 1989) which confers oc-amanitin resistance. S2 cells were transfected using Lipofectamine ( G I B C O B R L ) and stable transfectants were selected using oc-amanitin. Sema-la expression was induced by adding CuS04 to 0.7 m M and incubating for at least 72 hours. Sema-la expression was confirmed both by Northern blot analysis as well as immunohistochemical staining for cell surface Sema-la protein using monoclonal antibody 6F8 (Kolodkin et al., 1992) which is specific for grasshopper Sema-la. Single clones were obtained by serial dilution into soft agar and individual clones were subsequently grown up. Expression of semaphorin protein was detected using 6F8 antibody, and the highest expressing clone was used for experiments. Production of S2 cell balls Sema-la expression was induced with CUSO4 and transfected S2 clusters were generated in soft agar. Transfected S2 cells were grown in 0.3% agar, 0.7 m M CuSCu to generate S2 cell clusters. Individual clusters were removed and used for experiments. COS CELL TRANSFECTION Ful l length Sema-la was cloned into the expression plasmid p B K - C M V (Stratagene). C O S cells were transfected with p B K - C M V plasmid containing full length Sema-la using Lipofectamine ( G I B C O B R L ) . Stable transfectants were selected using Geneticin ( G I B C O B R L ) . Sema-la expression was confirmed by both Northern blot analysis as well as immunohistochemical staining using 6F8 antibody (Kolodkin et al., 1992). COS CELL AGGREGATES C O S cell aggregates were generated using the hanging drop method (Kennedy et al., 1994) and used immediately for functional studies. LIMB FILLET ASS A Y SYSTEM The T i l pathway was made accessible by the limb fillet preparation (Lefcort and Bentley, 1987). Grasshopper embryos at 30%> of embryonic development were anchored ventral side down on a poly-L-lysine coated glass coverslip. The posterior side of the T3 limb was cut longitudinally with a glass needle and spread apart on the coverslip to flatten the epithelium. A suction pipette was used to remove the mesodermal cells overlying the limb epithelium to expose the epithelium on which the T i 1 growth cones migrate. S2 cell clusters were strategically seeded 156 onto the limb using a sharp glass needle. In control experiments, non-transfected S2 cell clusters were used. Both experimental and control limbs were cultured at 30°C for another 5 % of development ( 2 4 hours) in grasshopper culture medium (see appendix) supplemented with 0.7 m M C U S C M , fixed and stained using standard immunohistochemistry. Visualization of neurons Embryos were fixed and stained as described in chapter 2. RNA Isolation and Northern Analysis Cellular R N A was extracted from S 2 and C O S cells using Trizol reagent ( G I B C O B R L ) . Twenty micrograms of total R N A were separated on 1 % agarose-formaldehyde gel and transferred to HybondN+ nylon membrane (Amersham Pharmacia Biotech) by capillary elution with 20X SSC. Sema-la D N A probes were prepared by random priming with the Klenow fragment of D N A polymerase in the presence of [a-32P]dCTP. Unincorporated nucleotides were removed from the probes using a D N A clean up column (Qiagen). Hybridization was performed at 68C in hybridization buffer overnight. The nylon membrane was then washed 4 X 5 minutes with 2 X SSC, 0.1% SDS at room temperature, followed by 2 X 1 hour with I X S S C , 0.1% SDS at 68C. Finally, the membrane was washed for 2 X 15 minutes with 0.1X SSC, 0.1% SDS at 68C. Washed nylon membranes were exposed to Kodak Biomax film for 36 hours with a Kodak B ioMax intensifying screen at 80°C. 157 Results: During axon pathfinding, the T i l growth cones have long filopodial processes which can extend 75 to 100pm (Bentley and Keshishian, 1982; Keshishian and Bentley, 1983a; O'Connor et al., 1990) effectively sampling large areas of the limb bud epithelium. The growth cones and axons are deflected from areas which are inhibitory but are attracted to and prefer to grow on regions which are more permissive (contact-mediated attraction). Drosophila S2 cells were transfected with transmembrane Sema-la and used to provide a localized ectopic source of Sema-la on the limb epithelium (Fig. 7.1 A ) . Antibody labeling without detergent indicate that the transmembrane protein is located on the cell surface. Western blot analysis did not detect any semaphorin secreted into the tissue culture supernatant, suggesting that the transmembrane molecule is not secreted or cleaved but is found tightly associated with the cell surface. To examine the ability of localized ectopic expression of Sema-la to direct pathfinding, the transfected cells or control non-transfected cells were implanted onto specific regions of the limb bud epithelium at 30% o f development both within and surrounding the path o f T i l growth cone migration. Embryos were cultured for 24 hours until 35%> of embryonic development, the normal time period required for the establishment of the T i l pathway. The localized ectopic source of Sema-la had a profound effect on normal T i l pathfinding in 100% (147/147) of the experimental limbs. In all cases, when transfected cells were placed within filopodial contact (Bentley and Keshishian, 1982; Keshishian and Bentley, 1983a; O'Connor et al., 1990) of the normal T i l pathway in regions where the T i l axons normally do not extend, T i l axon branches consistently contacted the transfected cells (Fig. 7.1F,G). When the cells were placed directly within the pathway, Sema-la did not deflect the T i l axons, induce growth cone collapse or arrest 158 Figure 7.1. T i l growth cones turn towards ectopic S2 and COS cells expressing Sema-la. Neurons are visualized using rhodamine-conjugated secondary antibodies (red, B-K, left panels). Sema-la expressing S2 and COS cells are visualized using floweriest-conjugated secondary antibodies (green, B-K, right panels). The brightly labeled cells (E, G, K) could also be seen at the rhodamine wavelength (D, F, J) due to \"bleed-through\" with rhodamine filters. A. Northern-blot analysis of Sema-la expression in S2 and COS cells detected a 3.0 kb Sema-la transcript only in transfected cells. \"S\" denotes Sema-la transfected S2 and COS cells. \"C\" denotes control, nontransfected cells. RNA loading control by ethidium bromide staining. B, H. Wild-type Til pathway. B, C. Control nontransfected S2 cells do not express Sema-la and typically have little effect on the pathfinding of the Til growth cones. D, E. Sema-la expressing S2 cells (arrows) placed within the normal Til pathway disrupt Til pathfinding. F, G. Sema-la expressing S2 cells placed in limb epithelium outside of the normal Til pathway attracts the Ti l growth cones away from their normal pathway to contact the Sema-la expressing cells (arrows). H , I. Control nontransfected COS cells have no effect on Til pathfinding. J , K. Transfected COS cells expressing full-length Sema-la actively attract Til growth cones (arrows). Scale bar, 60 pm. 160 further T i l axon extension as would be expected for a repulsive guidance molecule. Instead, the axons consistently made direct contact with the transfected cells in their pathway, extending in a dot-to-dot fashion (Fig. 7.1 D,E). After contacting a Sema-la expressing cell, the T i l axons continued to elongate along other transfected cells that were within filopodial reach (Fig. 7 .ID to G) . Thus, the phenotype of the developing T i l axon pathway is strongly dictated by the pattern specified by the location of ectopic Sema-la. In this manner, the phenotype of the resulting T i l axon pathway can be controlled by the location in which ectopic sources of Sema-la are placed. Surprisingly, after contacting a Sema-la expressing S2 cell, the T i l axons only continued to extend along other Sema-la expressing cells and were rarely observed to leave Sema-la expressing cells to grow aimlessly into limb epithelium. This behavior is reminiscent of previous observations where the growth cones of both the T i 1 and S G O neurons would not exit a band of epithelial cells expressing Sema-la unless they contacted a high affinity substrate such as another neuron (Wong et al., 1997; Klose and Bentley, 1989; Bentley and Caudy, 1983). Lastly, as long as other Sema-la cells were within growth cone reach, the lengths of the T i l axons were comparable to that of the wild-type limb at the same developmental time point (approximately 330 urn, Fig. 7.1) indicating that Sema-la does not inhibit axon elongation. In some cases where growth cones could not contact additional Sema-la expressing cells, the growth cone remained associated with the last contacted Sema-la cell (Fig. 7.1J,K). While transfected S2 cells disrupted the T i l axon pathway in 100% of the limbs examined, control limbs show slightly higher rates of abnormalities than limbs grown in the absence of S2 cells (15.7%, 14/89). In order to confirm that the guidance activity of the transfected cells is mediated by cell surface Sema-la, the experiment was repeated with transfected C O S cells. Whereas non-transfected COS cells had no effect on axon pathfinding in 161 any of 216 limbs examined (Fig. 7.1H,I), C O S cells expressing cell surface Sema-la actively perturbed T i l pathfinding, leading to branching of axons onto the transfected cells (Fig. 7.1J,K). Again, the Sema-la expressing cells did not deflect or inhibit the migration of the T i l growth cones as would be expected for a repulsive guidance molecule. The ability to dictate the migration pattern of the T i l axons by the regions in which the ectopic Sema-la was placed supports a direct attractive role for Sema-la as a contact-mediated attractive guidance molecule which can instruct axons to grow onto regions where it is expressed. Discussion: During embryonic development, Sema-la is dynamically expressed by bands of epithelial cells in the grasshopper limb bud epithelium (Kolodkin et al., 1992; Wong et al. , 1997, Chapter 2). The ability of Sema-la to guide neurons when it is differentially expressed suggests that the precise patterning of Sema-la expression in the epithelium may provide pre-established permissive pathways which are important for guiding axons in the peripheral nervous system. This would explain the previous observations that the T i l growth cones exhibit a preference for epithelium in the region of limb-segment boundaries (Caudy and Bentley, 1986a; Caudy and Bentley, 1986b; Condic and Bentley, 1989), regions which coincide with bands of epithelial cells which express Sema-la. This would also explain the characteristic sharp turn o f the T i l axons upon encountering a band of Sema-la expressing epithelial cells, and its directed migration within the narrow band of Sema-la expressing cells (Kolodkin et al., 1992; Wong et al., 1997, Chapter 2). Once the growth cones reach and enter the Sema-la band, they change their course of migration, staying strictly within the Sema-la path until they reach another attractive substrate 1 6 2 such as a guidepost cell or another neuron. The T i l growth cones are observed to exit the Sema-l a band when they encounter the C x i guidepost cells (Caudy and Bentley, 1983). Additionally, previous studies indicated that monoclonal antibody perturbation of Sema-la allowed the aberrant migration of the T i l growth cones out of the Sema-la band (Kolodkin et al., 1992). If Sema-la forms a preferred substrate for the T i l growth cones, disruption of the permissive substrate could explain the growth of axons out of the Sema-la band. The ability of Sema-la to provide a permissive substrate to promote axon outgrowth and branching and the ability of an artificially introduced ectopic source of Sema-la to successfully direct axon extension demonstrates that Sema-la actually functions as a contact-mediated attractive guidance molecule with an essential role in axon guidance in the developing grasshopper peripheral nervous system. Plexins are receptors for Class 1 semaphorins While neuropilins have been recently recognized as the receptors for class 3 semaphorins (Chen et al., 1997; He and Tessier-Lavigne, 1997: Kolodkin et a l , 1997), Plexin A has recently been reported as the receptor for class 1 semaphorins (Winberg et al., 1998). Like the neuropilins, the plexins were initially identified through a monoclonal antibody screen for antigens expressed in the Xenopus optic tectum (Takagi et al., 1987) and was also subsequently isolated from the mouse (Kameyama et al., 1996). Plexin was originally characterized as a homophilic cell adhesion molecule (Ohta et al., 1995) with a potential role in selective fasciculation (Satoda et al., 1995; Fujisawa et al., 1997). The first report to indicate a potential role for the plexins in semaphorin signaling came with the cloning of the receptor for a virus encoded semaphorin (Comeau et al., 1998). A39R, a semaphorin encoded by vaccinia virus was used to affinity purify an A 3 9 R receptor from a 163 human B cell line. Sequence analysis of the virus encoded semaphorin receptor ( V E S P R ) indicated that it is a novel member of the Plexin family (Comeau et al., 1998). Subsequently, two Drosophila Plexins were identified, PlexinA and PlexinB (Winberg et al., 1998). Plexin A is expressed in the developing Drosophila nervous system, and genetic analysis implicates that Plexin A is a receptor for class 1 semaphorins. Additionally, binding assays demonstrated that Sema-la and Sema-lb alkaline phosphatase fusion proteins bind C O S cells expressing PlexinA (Winberg et al., 1998). Plexin A is a large 220 kDa type one transmembrane glycoprotein (Fig. 7.2). Surprisingly, the extracellular domain of the Plexins contain a complete 500 amino acid semaphorin domain near the N terminus. The Plexins are also related to the MET-related tyrosine kinase receptors (including M E T , R O N , and SEA) . However, phylogenetic analysis of the semaphorin domains of the semaphorin family, the Plexin family and the MET-related tyrosine kinase receptors indicate that the three groups cluster separately, with the Plexins appearing as the ancestral molecules (Winberg et al., 1998). Due to the divergence in the semaphorin domain of the Plexins, the semaphorin naming nomenclature and family grouping do not include the plexins (semaphorin Nomenclature Committee, 1999). L ike the neuropilins, the cytoplasmic tail o f the Plexins contain no obvious motifs that suggest a mechanism by which this receptor may transmit signals and it is likely that the Plexins are a component of a receptor complex. Recently, plexin has been demonstrated to form an important part of a receptor complex with neuropilin-1 to mediate the repulsive activity of Sema3A. Recently, another study has indicated that Sema-1 a in Drosophila may function as a repulsive cue through its receptor Plexin A (Winberg et al., 1998, Tamagnone et al., 1999) and is 164 P L E X I N A / B C L A S S 1 S E M A P H O R I N S SIGNAL SEQUENCE SEMAPHORIN DOMAIN SEMAPHORIN DOMAIN MET-RELATED SEQUENCE CYSTEINE-RICH REPEATS TRANSMEMBRANE DOMAIN • GLYCINE-PROLINE-RICH REPEATS Figure 7.2. Structure of the class 1 semaphorin and its receptor, Plexin A . Class 1 semaphorins contain a 500 amino acid semaphorin domain and a transmembrane region. Plexins (Plexin A and Plexin B) are transmembrane glycoproteins. The extracellular domains include a full semaphorin domain, a motif of three cysteine-rich repeats (called MET-related sequences), and three glycine and proline rich repeats spaced ~50 amino acids apart. The intracellular region is divided into two blocks of strong homology separated by a variable linker. 166 important for motor axon extension and fasciculation (Yu et al., 1998). One simple explanation for this apparent discrepancy is that the semaphorin domain may simply be bifunctional, capable of mediating both adhesion as well as repulsion. This leads to the speculation that semaphorin mediated attraction and repulsion is specified by and dependent upon either the components of a receptor complex, events downstream of the receptor, or other molecules which may be associated with the ligand. A recent study now indicates that the chemorepulsive activity of the secreted semaphorins can be converted to attraction by pharmacological manipulation of cytosolic cyclic nucleotide (cGMP) levels (Song et al., 1998), suggesting that attractive and repulsive responses may be mechanistically related, potentially allowing an extending axon to respond differently to the same guidance molecule at different points along its journey to its final target. 167 Summary The past decade has seen the identification of a number of potential axon guidance molecules. Both cell surface and secreted guidance molecules provide guidance information in the form of attraction or repulsion of the growth cone, a sensory and motor apparatus at the tip of the extending axon, ultimately steering the axon to its correct target. The semaphorins are a large family of secreted and transmembrane molecules first identified in 1993, and recognized as the largest family of repulsive guidance molecules; functioning as a repulsive cue in both the vertebrate and invertebrate nervous system. This thesis outlines a series of studies which had provided the first demonstration that semaphorins can also have the opposite function, acting as an attractive guidance molecule in the developing grasshopper peripheral nervous system. Using the developing grasshopper nervous system, this thesis investigated the role o f Sema-la, the founding member of the Semaphorin family, in the development of the Grasshopper PNS and led to a number of observations: 1. Examination of the spatial-temporal expression pattern of Sema-1 a revealed that Sema-la is dynamically expressed by bands of epithelial cells during the time that a number of peripheral axon tracks are established in the developing limb bud. 2. Antibody double labeling studies indicated that Sema-la may play an important role in the pathfinding of the T i l and S G O axons. 3. Further investigation of the role of Sema-la in S G O pathfinding suggests that the arrested phenotype of the S G O growth cones in the absence of the T i l neurons is likely due to an attractive/permissive function of a band of epithelial cells expressing high levels of Sema-1 a.. 168 4. Antibody blocking of Sema-la prevents S G O outgrowth and extension, further supporting an attractive/permissive function for Sema-la. 5. A recombinant Sema-la construct (rSema-la) which contains only the Semaphorin domain is able to perturb T i l pathfinding. Ectopic rSema-la allows the T i l growth cones to migrate into normally unpermissive limb epithelium, suggesting that Sema-1 a is either an attractive guidance molecule, or is antagonizing the activity of a repulsive molecule. 6. Dimerization and clustering of rSema-la potentiates its activity, indicating that the activity of Sema-la can be regulated by its oligomerization state. Furthermore, Sema-la seems to be clustered when expressed heterologously in a cell line and when visualized in vivo. 7. A direct attractive guidance function for Sema-la is established by confronting the T i l axons with a localized ectopic source of Sema-la. Cells expressing full-length membrane bound Sema-la placed on developing limb buds attracted the T i l growth cones away from their normal pathway to contact the transfected cells. Once in contact with the transfected cells, the axons were unable to leave and grew only along cells which expressed ectopic Sema-la, demonstrating that Sema-la functions as a contact-mediated attractive guidance cue. The results presented in this thesis indicate that the function of the semaphorins is not limited to axon repulsion and further suggests that even the subcelluar organization of a membrane bound ligand may have a profound consequence for its biological activity. Recent studies are now indicating an attractive activity for other semaphorin members which were 169 previously known to only have a role in chemorepulsion. In vitro studies by Bagnard et al., 1998 indicate that secreted Sema E can attract cortical axons when the growth cones are exposed to increasing concentrations or a patterned distribution of Sema E . Independently, de Casro et al. (1999) found that Sema A , which was previously demonstrated to repel sympathetic axons (Adams et al., 1997; Takahashi et al., 1998) surprisingly acts as an attractant for olfactory bulb axons in collagen gel assays. Thus, there is now strong evidence that the semaphorins may be bifunctional guidance molecules with a wide range o f activities. The recent finding that the chemorepulsive activity of a secreted semaphorin, Sema 3a, can be converted to attraction by merely altering the cytosolic c G M P levels illuminates the complexity of the signaling mechanisms of the semaphorins. A s we are beginning to understand the complexity of the function and regulation of semaphorins in axon guidance, our efforts w i l l hopefully help us to answer some of the fundamental questions of how the intriguing and complex nervous system is generated. 170 Appendix A: The role of heterotrimeric G proteins in semaphorin signaling reassessed - Previous studies had implicated a role for a pertussis toxin sensitive G protein coupled receptor in mediating the collapsing activity of Sema3A. This appendix investigates the possible role ofpertussis toxin sensitive heterotrimeric G proteins in Sema-la signaling. I found that unlike previous reports, pertussis toxin sensitive G protein coupled receptors do not appear to play a role in Sema-la signaling during axon guidance. Introduction: Since the discovery of Fasciclin IV (Sema-la), the founding member of the semaphorin family, in 1992 (Kolodkin et al., 1992), molecules which contain the semaphorin domain have rapidly poured into the databases. The semaphorins now comprise a large gene family of structurally and functionally diverse glycoproteins with at least 20 members from different species, which could be subclassified into at least 7 subclasses (Semaphorin Nomenclature Committee, 1999). The semaphorin family of glycoproteins has been recognized as an important regulator of axon pathfinding. Although the semaphorins were at first believed to be a large family of chemorepulsive guidance molecules, recent findings indicate that the semaphorins may also function as attractive cues (see Puschel, 1999). A s the function of the semaphorins is beginning to unfold, the question of the intracellular events which occur during semaphorin signaling arises. What are the intracellular events which occur during semaphorin signaling, and how do they differ between different members of the semaphorin family and the different functions that they mediate? The mechanism by which semaphorins mediate their effects on the growth cone is currently unclear. 171 However, the high sensitivity of the growth cone to nanomolar concentrations of Sema3A in vitro (Luo et al., 1993) and the ability of a single ectopic Sema-la expressing cell to rearrange axonal morphology in vivo (Wong et al., 1999) suggests a powerful signal amplification system. Recently, the contribution of heterotrimeric G proteins to the regulation of semaphorin signaling has come under investigation (Igarashi et al., 1993; Goshima et al., 1995; Jin and Strittmatter, 1997), with evidence suggesting a role for G protein coupled receptors in mediating the collapsing activity of class 3 semaphorins. In this appendix, I focus on reexamining the role of G proteins in semaphorin signaling. Heterotrimeric G proteins, pertussis toxin, and mastoparan G protein coupled receptors contain a single polypeptide chain with seven membrane spanning domains and functionally couple to heterotrimeric G proteins. Heterotrimeric G -proteins are composed of a, (3, and y subunits. The activity of heterotrimeric G proteins is dependent on the guanine nucleotide bound to the a subunit (Gilman, 1987). In the basal inactive state, G proteins are bound to G D P , and, upon interaction with a receptor-ligand complex, G D P is released, allowing G T P to bind. Association with G T P shifts the a subunit to its active conformation. The GTP-bound G protein then activates various second messenger systems. The cascade is terminated by the intrinsic GTPase activity of the a subunit, which switches the G protein back to its GDP-bound inactive state. Indications of a role for G proteins in regulating growth cone motility could be detected by directly modifying the guanine nucleotide binding state through the use of various pharmacological agents. The wasp venom peptide mastoparan is a G o / G i stimulator and stimulates G proteins directly by a receptor-like mechanism (Higashijima at al., 1988; 172 Higashijima et al., 1990). Stimulation of Go and G i by either receptor activation or mastoparan is blocked by pretreatment with pertussis toxin (Gilman, 1987; Higashijima et al., 1990). Pertussis toxin A D P ribosylates the a subunit of heterotrimeric G-proteins of the Go or G i class and blocks their activation by receptors (Gilman, 1987; Tamura et al., 1982). Thus, both mastoparan and pertussis toxin are often used to detect the involvement of receptors which are coupled to heterotrimeric G proteins of the Go and G i classes. Heterotrimeric G proteins in neuronal guidance A x o n pathfinding results from the navigation of growth cones in response to specific guidance cues. Guidance cues provided by cell surface and secreted molecules have a number of effects on growth cones, including: stimulation of neurite outgrowth regulation of axonal fasciculation chemoattraction/contact-mediated attraction chemorepulsion/contact-mediated repulsion - induction o f growth cone paralysis or collapse (in vitro) There are a number of indications that heterotrimeric G-proteins may play an important role in growth cone guidance. The growth cone membrane has been found to contain very high levels of the heterotrimeric G T P binding protein G i and Go (Strittmatter et al., 1990; Strittmatter et al., 1991; Strittmatter et al., 1992), with Go in greater abundance, suggesting that the Go activation state may regulate neurite extension (Strittmatter and Fishman, 1991; Strittmatter, 1992). In fact, Go and G i are the most prominent noncytoskeletal components of the growth cone membrane (Strittmatter et al., 1990). Other than its enrichment in the growth cone, a role for heterotrimeric 173 G-proteins in regulating neurite growth is also compatible with a number of observations which indicate that the activation state of G proteins can drastically affect growth cone morphology. For example: 1. N C A M and N-cadherin enhance growth cone and neurite extension from P C 12 cells. However, N C A M and N-cadherin induced neurite outgrowth from PC 12 cells can be blocked by pertussis toxin (Doherty et al., 1991). Additionally, antibodies to N C A M and L I decrease phosphoinositide hydrolysis and increase intracellular calcium levels in PC12 cells by a pertussis toxin-sensitive mechanism (Schuch et al., 1989). 2. Activation of pertussis toxin sensitive G proteins by mastoparan increases neurite outgrowth from neuroblastoma cells. 3. Expression of activated mutants of the a subunit of Go promotes neurite outgrowth (Strittmatter et al., 1994). However, in some other neuronal cell types, G-protein activation has been associated with decreased neurite outgrowth. For example: 1. 5-HT (Haydon et al., 1984), dopamine (Lankford et al., 1988; Rodrigues and Dowling, 1990), and thrombin (Suidan et al., 1992) interact with G protein coupled receptors, causing growth cone collapse and the cessation of neurite extension. 2. Ligands for G-protein-coupled receptors such as serotonin and dopamine induce growth cone collapse (Haydon et al., 1984; Lankford et al., 1988; Rodrigues and Dowling, 1990; Suidan et al., 1992). 3. Mastoparan which stimulates G proteins (Higashimima et al., 1990) and C N S myelin (Caroni and Schwab, 1988; Bandtlow et al., 1990) induce collapse of dorsal root 174 ganglion growth cones in a pertussis toxin sensitive fashion (Kaziro et al., 1991; Igarashi et al., 1993; Strittmatter et al., 1994b). 4. G-protein stimulation of embryonic chick sympathetic neurons also decreases neurite extension (Strittmatter et al., 1992b). 5. The growth cone collapsing activity of embryonic brain membranes is blocked by pertussis toxin (Igarashi et al., 1993). 6. Expression of activated mutants of the a subunit of Go increases neurite outgrowth from neuroblastoma and pheochromocytoma cells (Strittmatter et al., 1994). Taken together, these studies indicate that the G-protein activation state can alter growth cone motility in different ways. Heterotrimeric G proteins mediate Sema3A induced growth cone collapse Previous studies have indicated that the growth cone collapsing effect of the secreted semaphorins (collapsing) are also mediated by a pertussis toxin sensitive G protein coupled receptor (Igarashi et al., 1993; Goshima et al., 1995; Jin et al., 1997). Initial investigations indicate that both mastoparan (which activates heterotrimeric G proteins of the Go and G i subtypes) (Higashijima at al., 1988; Higashijima et al., 1990) as well as chick brain membrane extract ( B M E ) (Raper and Kapfhammer, 1990) which contains Sema3A (Luo et al., 1993) stimulate growth cone collapse when added to chick D R G and retinal-neuron growth cones in vitro. However, growth cone collapse induced by Sema3A from B M E can be prevented by pretreatment with lOOng/ml pertussis toxin, implicating a G protein signaling pathway (Igarashi et al., 1993). The findings suggest that Sema3A induced growth cone collapse is mediated by a pertussis toxin sensitive G protein coupled receptor. 175 Do heterotrimeric G proteins mediate Sema-J a growth cone attraction? The semaphorins are a large family of molecules with important functions in nervous system development, acting both as chemorepellents as well as attractants. Based on previous studies which have indicated that the collapsing activity of Sema3 A is mediated by a pertussis toxin sensitive G-protein coupled receptor (Igarashi et al., 1993), it is possible that a pertussis toxin sensitive receptor is also responsible for mediating semaphorin attractive activity. Whether contact-dependent signaling via membrane-bound Sema-la (Wong et al., 1997; Wong et al., 1999, Chapters 3-7) is also mediated by a pertussis toxin sensitive G-protein coupled receptor is unknown. Given that pertussis toxin sensitive G protein coupled receptors are capable of mediating both collapsing as well as neurite promoting responses in the growth cone as discussed previously, I investigated whether the attractive outgrowth promoting activity of Sema-la is also mediated by a pertussis toxin sensitive G protein coupled receptor. This appendix investigates the role of a pertussis toxin sensitive G protein coupled receptor in the formation of two peripheral pathways which are critically dependent upon Sema-la signaling for guidance during development. The correct outgrowth and navigation of the T i l (Kolodkin et al., 1992; Wong et al., 1999, Chapters 5-7) and S G O (Wong et al., 1997, Chapters 2 and 3) axons is critically dependent upon guidance information from Sema-la. Previous studies have demonstrated that perturbation of Sema-la leads to drastic abnormalities in both the T i l and S G O axon tracts (Kolodkin et al., 1992; Wong et al., 1997; Wong et a l , 1999, Chapters 3-7). To see whether G protein coupled receptors play a role in mediating semaphorin attractive activity, the role of pertussis toxin sensitive G protein receptors on T i l and S G O 176 pathfinding was examined. Pertussis toxin was added to peripheral neurons extending in vivo to examine its ability to affect normal Sema-la responses in the limb bud. The aim was to understand whether different members of the semaphorin family are all dependent upon pertussis toxin sensitive G protein couple receptors, and more specifically, whether semaphorin induced growth cone collapse and attraction are mediated by similar signaling mechanisms. Mater ia ls and Methods Til assay In order to examine the effects of pertussis toxin on T i 1 pathfinding, embryos were collected at 29%, 30%, 31%, 32.5% and 33% of development. The entire amnion was removed and the embryos were immediately cultured for 30 hours in the presence of various concentrations of pertussis toxin. SGO assay In order to examine the effects of pertussis toxin on SGO initiation as well as extension, embryos were collected at 38%, 39% and 40% of development. Embryos were immediately processed using the limb fillet preparation and in some embryos, the T i l cell bodies were located and extracted. In some of the embryos, the T i l neurons were removed as described in chapter 3. Embryos were then cultured for 30 hours in the presence of various concentrations of pertussis toxin. 177 Visualization of neurons Following fixation, anti-HRP antibodies were used to visualize the neurons using standard immunohistochemistry. Results: Pertussis Toxin does not disrupt Til pathfinding As a first step to assess the possible interaction of Sema-la with a pertussis toxin sensitive G protein coupled receptor, grasshopper embryos at 29%, 30%, 31%, 32.5% and 33% of development were cultured for 24 hours in the presence of lOOng/ml pertussis toxin. In control embryos incubated in the absence of pertussis toxin, the T i 1 pathway developed normally, forming the complete T i l pathway within the 24 hour culture period. Embryos incubated in the presence of 100 ng/ml pertussis toxin also developed normally. Increases in defective T i l pathfinding phenotypes in experimental limbs were not detected (Fig. 8.1). The T i l pathway was normal and the rate of extension was unaffected. When embryos were cultured in the presence of up to 2000ng/ml pertussis, T i l pathfinding was still unaffected (Fig. 8.1). In particular, T i l pathfinding in the region of endogenous Sema-la expressing epithelium was normal. The T i l growth cones migrated normally along the Sema-la band and increased branching outside of the band was not observed. Previous studies had indicated that antibody blocking of Sema-la resulted in increased branching in the region of the Sema-la band. Pertussis toxin treatment of the embryos to block G protein coupled receptors in the T i l growth cone did not prevent the T i l growth cones from responding to endogenous Sema-la, suggesting that Sema-la signaling in the T i l growth cones is not mediated by a pertussis toxin 178 o CO CO o *3 m <3 •O O C O -o C O 00 o CO •SI-L O o CN L O DL-DLU9SJ |uu/6n 5'zz |UJ/6u 0002 |UJ/6U 001 ^0 c> 00 CO |W/6u 0002 |uu/6u OQI ^0 c> LO CN CO |UJ/6U 00 L |uu/6u 0002 |uu/6u 0091 c> |uu/6u o g / CO |uu/6u 001 |W/6u 0002 |UJ/6U 0091 |W/6u 000 L ^0 o> O CO |uu/6u 09/ |uu/6u 00 L |w/6u 0002 |uu/6u 00S L |W/6u 000 L s@ c> 0 s CN |iu/6u 09/ |UJ/6u 00 L S3LLIlWMcJONaV% Figure 8.1. Effects of P T X on T i l pathfinding. Embryos at 29%, 30%, 31%, 32.5% and 33% of embryonic development were cultured for 24 hours the presence of various concentrations of PTX and then fixed and processed using standard immunohistochemistry. Concentrations of PTX up to 2000 ng/ml had no significant effect on T i l pathfinding. sensitive G protein coupled receptor. Although concentrations of 750 ng/ml P T X and above did not significantly affect T i l pathfinding, defects in embryogenesis in other regions of the limb were observed, including defects in neuronal induction, and epithelial cell death. Pertussis toxin does not disrupt SGO pathfinding or outgrowth The S G O growth cones normally initiate axon outgrowth onto Sema-la expressing epithelium.and Sema-la is necessary for S G O axon outgrowth. Additionally, Sema-la acts as an attractive cue for the S G O growth cones. This attraction becomes very apparent when the pioneering axon tract in the limb is artificially disrupted. Artificial removal of the T i l cell bodies which are guidepost cells for the S G O growth cones results in arrested S G O growth cone migration within a band o f Sema-la which is likely due to Sema-la attractive activity. When Sema-la is blocked in the limb at 40% of development using anti-Sema-la antibodies, S G O axonogenesis is prevented. Furthermore, when Sema-la is blocked in the limb after axonogenesis has already occurred, the S G O growth cones are prevented from continuing to migrate across the Sema-la expressing epithelium, likely because the S G O growth cones require Sema-la signaling for axonogenesis and extension. Because of the sensitivity and dependence of the S G O axons to Sema-la signaling, I further tested the possibility that the S G O growth cones respond to Sema-la through a pertussis toxin sensitive G protein coupled receptor. Embryos at 38% of development were cultured in the presence of 100 ng/ml, 750 ng/ml, 1000 ng/ml, 1500 ng/ml and 2000 ng/ml pertussis toxin for 24 hours in the presence of anti-Sema-la monoclonal antibody 6F8. Pertussis toxin blocking of G protein coupled receptors had no effect on S G O axonogenesis; outgrowth from the S G O neurons was comparable to that of control limbs in 100% of experimental limbs (43/43). 181 Sema-la is also important in acting as a permissive cue for the SGO axons after outgrowth has already occurred, allowing the SGO growth cones to extend only along Sema-la expressing cells in the absence of the T i l cell bodies. Chapter 3 had indicated that artificial removal of the T i l axons in the limb at 38% of development allows us to examine the subtle effects of Sema-1 a on S G O axon extension. Normally, in the presence of the T i l neurons, the S G O axons are able to extend across a band of Sema-la to contact the T i l cell bodies even when Sema-la is functionally blocked. However, when the T i l cell bodies are physically removed, the proximal extension of the S G O growth cones is highly dependent upon the Sema-la expressing epithelial cells. Antibody blocking of Sema-la in the absence of the T i l cell bodies prevents S G O growth cone extension (Wong et al., 1997). In order to investigate whether a pertussis toxin sensitive G protein coupled receptor is responsible for S G O growth cone extension in the absence of the T i l cell bodies, the T i l neurons were removed form embryos at 38%), 39% and 40% of development and embryos were cultured in the presence of either anti-Sema-la antibody or pertussis toxin for 24 hours. A s noted previously, antibody blocking of Sema-la prevented the proximal migration of the SGO growth cones in the absence of the T i l cell bodies. However, pertussis toxin perturbation of pertussis toxin sensitive G protein coupled receptors had no effect on S G O extension in 100% of experimental limbs (39/39). In all limbs examined, the S G O growth cones extended completely across the band of Sema-1 a expressing cells and reached the proximal side of the Sema-la band. Pertussis toxin had no detectable effect on S G O outgrowth or pathfinding at concentrations up to 2000 ng/ml. 182 Discussion: Members of the semaphorin family are able to steer growth cones by two different mechanisms: attraction, and repulsion (see Puschel, 1999). It is unknown whether the attractive and repulsive actions of the semaphorins are mediated by common receptor types and similar intracellular mechanisms. One of the first insights into the intracellular events which occur during semaphorin signaling came with the finding that heterotrimeric G proteins may play an important role in Sema3A induced growth cone collapse (Igarashi et al., 1993; Goshima et al., 1995; Jin et al., 1997). Growth cone collapse induced by chick brain membrane extract which contains Sema3A is prevented by pretreatment with pertussis toxin (Igarashi et al., 1993; Jin et al., 1997) which inactivates several types of heterotrimeric G proteins, including Go and G i (Gilman, 1987; Tamura et al., 1982). The finding that the repulsive activity of Sema3A is mediated by a pertussis toxin sensitive G protein coupled receptor lead us to question whether the attractive activity of Sema-la is also mediated by pertussis toxin sensitive heterotrimeric G proteins. Sema-la attractive activity is not likely mediated by a pertussis toxin sensitive G protein coupled receptor Sema-la is an important guidance molecule in the developing grasshopper P N S and its precise spatial-temporal expression pattern is important for the formation of the highly stereotyped projections of the T i l and SGO neurons (Kolodkin et al., 1992; Wong et al., 1997; Wong et al., 1999). Previous findings have indicated that antibody perturbation of Sema-la during development leads to drastic pathfinding abnormalities by these neurons (Kolodkin et al., 1992; Wong et al., 1997). Thus, the T i l and SGO pathways of the developing grasshopper PNS 183 provided a good assay system for examining the potential role of pertussis toxin sensitive G proteins in Sema-1 a signaling. Correct pathfinding by the T i l growth cones is critically dependent upon guidance information from a band of Sema-la expressing epithelial cells at the approximate location of the Tr-Fe limb segment boundary (Kolodkin et al., 1992; Wong et al., 1999). Surprisingly, pertussis toxin inactivation of Go and G i had little effect on T i l and S G O axon outgrowth and pathfinding. Even at high concentrations of pertussis toxin, the T i l growth cones responded normally to Sema-la cues in the limb bud epithelium. The S G O neurons are also acutely dependent upon Sema-la signaling during development (Wong et al., 1997). Sema-la had previously been shown to be essential for S G O outgrowth and extension. Antibody blocking of endogenous Sema-la in the limb bud during development arrested S G O extension and prevented SGO axon outgrowth. However, pertussis toxin inactivation of Go and G i did not prevent the S G O growth cones from responding normally to Sema-la in the limb. Both S G O axon extension as well as axon outgrowth were not affected. While previous studies had indicated that 100 ng/ml pertussis toxin was able to prevent growth cones from responding to the collapsing activity of Sema3A (Igarashi et al., 1993), I found that peripheral growth cones were able to respond normally to Sema-la even in the presence of 2000 ng/ml pertussis toxin. The presence of pertussis at levels which are toxic to the developing limb bud did not prevent the T i l and S G O growth cones from responding normally to Sema-la, suggesting that it is unlikely that Go or G i are responsible for mediating the attractive guidance activity of Sema-la. However, in the absence of a positive control, it could be argued that the lack of effect of pertussis toxin on Sema-la signaling could be contributed to 184 the possibility that G protein coupled receptors in the grasshopper are not sensitive to pertussis toxin. The semaphorin receptor is not a G protein coupled receptor Although G protein coupled receptors have been implicated to mediate the effects o f Sema3A inhibitory guidance information (Igarashi et al., 1993; Goshima et al. , 1995; Jin et al., 1997), our studies indicate that the response of peripheral growth cones to the attractive guidance activity of Sema-la is likely independent of a pertussis toxin sensitive G protein coupled receptor. The inability of pertussis toxin to affect the pathfinding of axons which are highly dependent on semaphorin signaling leads to several speculations. It is possible that while semaphorin inhibitory signaling is mediated by a G protein coupled receptor, attractive activity is mediated by a different receptor type. It is also possible that secreted semaphorins bind to G -protein coupled receptors while the transmembrane Sema-la receptor is not G-protein coupled. The recent identification of the receptors for both the secreted Sema3A as well as the transmembrane class 1 semaphorins is beginning to shed some light into this issue. In agreement with our findings that the activity of Sema-la is not mediated by a G protein coupled receptor, the receptor for Sema-la, PlexinA, has only one potential transmembrane domain and does not resemble a G protein coupled receptor (Comeau et al., 1998; Winberg et al., 1998). Additionally, neuropilins, which are the functional receptors for the repulsive class 3 semaphorins also are not G protein coupled receptors (see Kolodkin and Ginty, 1997). The recent identification of potential semaphorin receptors leads to a number of speculations regarding the role of pertussis toxin sensitive G protein coupled receptors in Semaphorin signaling: 185 1. The method of neuropilin signaling is currently unknown. It is possible that neuropilin is a component of a receptor complex and is functionally associated with a G protein coupled receptor. First, the cytoplasmic domain of neuropilin is short and contains no motifs which are known to participate in signal transduction (see Kolodkin and Ginty, 1997). Second, semaphorins with equal affinity for neuropilin 1 exhibit different binding patterns in situ (Chen et al., 1997; Feiner et al., 1997). Third, Neuropilin deletion and chimera studies demonstrate that the transmembrane/cytoplasmic regions are not necessary for neuropilin-1 signaling (Nakamura et al., 1998). Fourth, neuropilin-1 has been demonstrated to function as a nonsignaling coreceptor with a receptor tyrosine kinase for the vascular endothelial growth factor V E G F (Soker et al., 1998), leading to the speculation that a receptor tyrosine kinase may couple with neuropilins in the growth cone to mediate semaphorin signaling. The answer to this question w i l l become apparent with the cloning of molecules which interact with the receptor subunits. 2. Class 3 semaphorins may be the ligand for more than one type o f receptor. A G protein coupled receptor for class 3 semaphorins has not yet been identified. 3. The inhibition of Sema 3A growth cone collapse by P T X is not mediated by A D P ribosylation of a G protein and thus pertussis toxin sensitive G protein coupled receptors do not play a role in Sema3A signaling. Recent controversy over Sema3A signaling Chick E10 brain membrane extracts ( B M E ) containing Sema3A cause dorsal root ganglion growth cone collapse in a pertussis toxin sensitive fashion (Igarashi et al., 1993), 186 suggesting that Sema3A may act through a G protein-mediated cascade. However, some effects of pertussis toxin do not involve A D P ribosylation of G proteins and can be mimicked by a form of the toxin, the P oligomer, that lacks the SI subunit (Tamura et al., 1983; Gray et al., 1989; for review, see Kaslow and Burns, 1992) . Recent data have questioned the involvement of A D P ribosylation of G proteins in the blocking action of pertussis toxin (Kindt and Lander, 1995). Kindt and Lander (1995) reported that the P oligomer of P T X containing the cellular binding subunits but not the catalytic SI subunit is able to block growth cone collapse induced by Sema3A containing chick E10 brain membrane extract. This indicates that the inhibition of growth cone collapse by P T X is not mediated by A D P ribosylation of G protein. However, Goshima et al. (1997) and Jin et al. (1997) subsequently claimed that pertussis toxin and not just its P oligomer was needed to block the growth cone collapsing activity of purified Sema 3 A . They indicated that the potent effects o f the p oligomer observed by Kindt and Lander (1995) might be due to the blocking of growth cone collapsing activities other than Sema 3 A which may present in the B M E preparation (Luo et al., 1993). Nevertheless, the findings reported by Kindt and Lander, the findings reported in this chapter, as well as the recent * O f the four polypeptides that comprise the P oligomer, two bind oligosaccharides, especially oligosaccharides with terminal sialic acid residues (Armstrong et a l , 1988; Witvliet et al., 1989). Pertussis toxin may act by binding cell surface carbohydrates. Neurites lacking complex type N-linked oligosaccharides are insensitive to the effects of the toxin (Kindt and Lander, 1995). 187 identification of semaphorin receptors, has greatly weakened the idea that G protein coupled receptors play a role in Semaphorin signaling. New molecules implicated in semaphorin signaling The mechanism by which semaphorin receptors signal to the interior of the growth cone remains unknown. The intracellular domain o f neuropilin is short and contains no motifs with obvious catalytic function nor any domains that offer clues regarding the mechanism of Sema3 A signal transduction. However, because the intracellular domains of neuropilin-1 and neuropilin-2 are similar with respect to both primary sequence and length, it is likely that they share a common signaling mechanism (Kolodkin et al., 1997). Recently, neuropilin-1-interacting protein (NIP) was identified as a P D Z domain containing cytoplasmic protein which may participate in the regulation of neuroplin-1 mediated signaling. Further insight into semaphorin signaling came with the identification of the cytosolic protein termed collapsin response mediator proteins ( C R M P family). The C R M P s were identified through expression cloning in Xenopus laevis oocytes in an attempt to identify molecules that might be involved in Sema3A action (Goshima et al., 1995). Injection o f C R M P -62 protein renders oocytes Sema3A responsive within minutes, indicating that the protein functions directly in semaphorin response pathway. Cells treated with ant i -CRMP antibodies do not respond to Sema3A, indicating that C R M P - 6 2 is required for the Sema3A signaling process. H o w C R M P - 6 2 is involved in Sema3A signaling is, at present, not understood. It has been proposed that C R M P - 6 2 is an intracellular component of the multimeric receptor complex that couples semaphorin binding transmembrane receptors to the signaling machinery and functions upstream of a heterotrimeric G protein (Goshima et al., 1995; Wang and Strittmatter, 1996). 188 Other than heterotrimeric G proteins, recent evidence indicates that r ac l , a small monomeric G protein of the rho subclass may mediate the inhibitory effects of collapsin-1 on neurite outgrowth (Jin and Strittmatter, 1997). Evidence that racl mediates collapsin-1 action in D R G neurons comes from studies where the trituration of dominant negative racl abolishes growth cone collapse by Sema3A and greatly reduces neurite outgrowth inhibition by Sema3A. Constitutively active racl weakly mimics the action of Sema3A. Additionally, dominant negative racl does not prevent myelin-induced growth cone collapse, suggesting that racl specifically acts to transduce the collapsing signal of Sema3A. Although the activity of semaphorins were originally thought to be mediated by G protein coupled receptors, in light of new data, the role of G proteins in semaphorin signaling is uncertain. Answers to this question wi l l become more apparent as more receptor components of the semaphorins are identified. 189 Appendix B: Towards the future of Sema-la function analysis: Cloning and sequencing of C S E M A 5 B , a vertebrate Semaphorin with high homology to Grasshopper Sema-la thesis investigated the function of Sema-la and provided extensive evidence that transmembrane Sema-la functions as an attractive axon guidance molecule in the developing invertebrate grasshopper nervous system. Whether a transmembrane molecule with a similar function exists in the vertebrate nervous system is currently unknown. In this appendix, I describe the cloning of a transmembrane semaphorin with high sequence homology to grasshopper Sema-la in the developing chick nervous system and postulate that this molecule likely functions as an attractive guidance molecule in the vertebrate nervous system. Introduction: During development, axonal growth cones navigate to their correct target by responding to a number of guidance molecules in their local environment. Molecular guidance cues generally guide growth cones through four basic mechanisms, including chemoattraction and chemorepulsion mediated by soluble guidance molecules and contact attraction and repulsion mediated by cell surface non-soluble guidance molecules (Tessier-Lavigne and Goodman, 1996). The critical role of the semaphorin family of guidance molecules in nervous system development has now been demonstrated in a number of systems. Since their discovery in 1992, a number of functional assays have demonstrated a chemorepulsive role for secreted semaphorins in both invertebrate as well as vertebrate systems (see Mark et al., 1997), leading researchers to initially believe that the semaphorins are a large family of chemorepellent guidance molecules. 190 This thesis was aimed at addressing the functional role of the first member of the semaphorin family identified, Sema-la, a transmembrane semaphorin of the invertebrate nervous system (Kolodkin et al., 1993). Whereas most secreted semaphorins identified to date have been shown to function as chemorepellents, this thesis provides evidence that transmembrane Sema-la mediates the opposite function, acting as a potent contact mediated attractant which is essential for peripheral nervous system development (see Puschel, 1999). Whether a transmembrane semaphorin has a similar function in the vertebrate nervous system is unknown. Given the finding that transmembrane Sema-la is an important chemoattractive guidance molecule in the developing invertebrate nervous system, I wondered whether a transmembrane semaphorin with a similar function exists in the developing vertebrate nervous system. In order to search for additional semaphorin molecules which might regulate the axon projection of vertebrate neurons in a manner similar to that of invertebrate Sema-la, a polymerase chain reaction-based strategy was used to clone a vertebrate transmembrane semaphorin from the developing chick brain with high homology to grasshopper Sema-la. This appendix describes the identification and cloning of Chick SemaV (now renamed C S E M A 5 B ) , a chick semaphorin with close sequence homology to grasshopper Sema-la within the semaphorin domain, suggesting that this molecule may potentially function as an attractive guidance cue in the vertebrate nervous system. 191 Materials and Methods: Clustal alignment of semaphorins The sequences of the transmembrane and secreted semaphorin family members were obtained using the Entrez nucleotide search (NCBI) . The sequences in Fasta format were aligned using the Clustal W program which is available online from the Bayor College of Medicine. This included the sequences of the secreted semaphorins, Sema D , Sema E, Sema-2a and the transmembrane semaphorins Sema3F, C D 100, and Grasshopper Sema-la. The aligned sequences were examined in detail to note regions of interest which are unique to the semaphorin domain of the transmembrane semaphorins. Design degenerate PCR primers In an attempt to isolate a transmembrane Sema-la-like molecule from the embryonic chick brain, R T - P C R was performed using degenerate primer pairs for conserved motifs o f the semaphorin domain. Based on the sequences of the highly conserved regions which are found in only the transmembrane semaphorins, a number of degenerate primer pairs were generated which encompass conserved regions within the semaphorin domain. The primers were generated by Gibco Laboratories. RNA extraction from E10 chick brain Fertilized white leghorn chicken eggs were obtained from Coastline Chicks in Abbotsford, B C . The eggs are kept up to one week at 8°C. For experiments, eggs were placed in a 30°C incubator. The state of embryonic development when first placed in the incubator is E0. At E10, the chick 192 is decapitated and the entire brain is removed. Total R N A is immediately extracted from the E10 brain using Trizol (Gibco B R L ) following the manufacturer's instructions. Reverse Transcription Chick E10 c D N A was generated from chick E10 total R N A using reverse transcriptase (Gibco) and oligo dT primers. Low stringency PCR cloning Each of the degenerate primer pairs were used in an attempt to amplify from the c D N A . Adjustment of both M g concentration and P C R cycling conditions were necessary in order to obtain successfully amplified products (visible amplified D N A bands). Clone PCR products into pCR2.1 plasmid The P C R reaction for each of the primer pairs was separated by agarose gel electrophoresis. The D N A was then extracted from the gel and purified (Qiagen). The P C R products were cloned into the plasmid pCR2.1 (Invitrogen) using T-tail cloning and positive clones were identified using blue/white screening. A total of 128 positive clones were sequenced (USB Sequenase Version 2.3) and analyzed before a semaphorin clone was successfully identified. Library screening The semaphorin clones isolated were used to generate 3 2 P labeled D N A probes. The D N A probes were used to screen a chicken E10 c D N A library (Clonetech). The plaques were transferred onto nitrocellulose filters and screened at low stringency at 68°C. Over 10 mil l ion 193 clones were screened in the primary screen and 25 positives were obtained. Four more series of screens were performed, resulting in ten positive clones. Over 5 mil l ion clones were screened in the primary screen using a different probe and over 35 positive clones were obtained. Four more series of screens were performed, resulting in 17 positive clones. Analysis of positive clones Each of the positive clones obtained from the final screen were isolated and grown up in order to isolate lambda D N A . Lambda D N A was prepared using a Lambda prep kit (Qiagen). Due to the difficulty in isolating lambda D N A , the D N A from each of the 42 clones were individually subcloned into either pBluescript or pLitmus for sequencing. Sequencing of largest clones The D N A from the ten largest clones obtained was isolated and sequenced using the Perkin Elmer A B I 377 Sequencer. One full length clone (3.5 Kb) was obtained from this screen. Internal primers were designed for the entire length of the clone and were used to sequence the entire clone. Results Identification of conserved sequences which are unique to transmembrane semaphorins In an effort to identify a vertebrate homologue of transmembrane grasshopper Sema-la, the nucleotide sequences of known secreted and transmembrane Semaphorins were aligned using ClustalW. The most obvious differences noted between transmembrane and secreted 194 semaphorins are the regions found outside of the semaphorin domain. However, it is not known whether the transmembrane semaphorins share common sequences within the semaphorin domain which are absent in the secreted forms. Detailed analysis of ClustalW alignments of secreted and transmembrane semaphorins indicated that within the semaphorin domain, several small groups of nucleotides are more highly conserved between transmembrane forms o f semaphorins than in the secreted semaphorin members. This suggests that other than structural differences, the transmembrane and secreted forms also demonstrate distinct differences in their semaphorin domains. Cloning ofC SEMA 5B P C R primers designed based on the regions o f conservation within the Semaphorin domain of transmembrane Semaphorins were used to amplify chick Semaphorin sequences from the E l 0 chick brain. P C R products were used as probes to screen a chick E10 c D N A library. This resulted in the isolation of numerous c D N A s , including one 3.0Kb full length clone encoding a candidate Semaphorin gene. The c D N A predicts a 1090 amino acid novel transmembrane semaphorin, Chick SemaV (now renamed C S E M A 5 B ) with high sequence homology to grasshopper Sema-la within the semaphorin domain (Fig. 9.1). Examination o f the amino acid sequence deduced from the isolated c D N A showed that the predicted protein is a transmembrane molecule with a hydrophobic membrane spanning domain followed by a short intracellular domain at the C terminus (Fig. 9.2). A n N-terminal stretch of 15-25 amino acids with the characteristics of a signal sequence is followed by a semaphorin domain. Adjacent to the semaphorin domain is a chain of thrombospondin repeats. A consensus N-glycosylation site 195 C S e m a V M W S R T . K A T B T . S T . P S T . F T . T . A F H T . S A S O N V A ILS--EHQQCVRKEH PTIAFEDLKPV