Download 14-‐3-‐3 Proteins regulate axonal growth cone

Document related concepts
no text concepts found
Transcript
14-­‐3-­‐3 P roteins r egulate a xonal growth c one t urning r esponses Christopher B. Kent A thesis submitted to McGill University in partial fulfillment of the requirements of the degree of Doctor of Philosophy in Neurological Sciences Integrated Program in Neuroscience Department of Neurology and Neurosurgery -­‐ McGill University 3801 University Street -­‐ Montreal, QC, H3A 2B4 CANADA August 2012 © Christopher Kent 2012 All rights reserved Acknowledgments I would like to thank my supervisor Dr. Alyson Fournier for always having an open door. Her patience and willingness to listen has provided much encouragement as I have stumbled along through the awkward first stages of becoming a scientist. I must also thank Dr. Wayne Sossin, Dr. Edward Ruthazer and Dr. Keith Murai, the members of my thesis committee for their invaluable advice and continued guidance through the development of this work. I would also like thank the many collaborators I have had the privilege to work with namely Dr. Fred Charron, Dr. Tim Kennedy, Dr. Paul Wiseman, Dr. Santiago Constantino, Dr. Peter McPherson, Dr. Phil Barker and particularly Dr. Patricia Yam. They have all been the source of much insight and wisdom that has been a major contribution to my training experience. I have relied heavily on the experience, hard work and good nature of many of my past and present colleagues in the Fournier Lab and the MNI. I would especially like to thank Dr. Tadayuki Shimada, and Dr. Madeline Pool. As well as Isabel Rambaldi, Dr. Gino Ferraro, Dr. Yazan Alabed, Dr. Stefan Ong-­‐Tone, Dr. Francois Beaubien, Dr. Sathy Rajasekharan, Dr. Simon Moore, Dr. Nicolas Stifani, Sonia Rodrigues, Karen Lai, Benoit Vaillancourt, Rhalena Thomas and James Correia I would like to thank all the great friends, new and old that have been there for me over the last few years, bugging me about my cells and reminding me that there is more to life than the next experiment. I would also like to thank my siblings for their enthusiastic support and my parents for their good natured acceptance of my allergy for regular employment and their continued efforts to understand what I have been doing over the last few years. Lastly I would like to thank my daughter Violette, for bringing so much happiness into the world and my wife Jan who, despite her hatred for Saturdays at the lab, continues to share every joy and every anxiety of this journey and will always be the first ear for every idea, no matter what is to come. Contributions of Authors First Manuscript: “14-­‐3-­‐3 Proteins Regulate Protein Kinase A Activity to Modulate Growth Cone Turning Responses” -­‐Christopher B. Kent: Performed experiments, analysis and revisions pertaining to Figures 1-­‐5, Figure 7 and Supplementary Figures 3-­‐5. Design of project, analysis of data and writing and editing of manuscript. -­‐Tadayuki Shimada: Performed experiments and revision for Figures 6, 7 -­‐Gino B. Ferraro: Designed and performed experiments for Supplementary Figure 1. Analysis of proteomics data in Supplementary Figure 2 -­‐Brigitte Ritter: Design of microRNA delivery system used to perform experiments in Figure 4 -­‐Patricia T. Yam: Design of Dunn Chamber Turning Assay used to perform experiments pertaining to Figures 2-­‐5. -­‐Peter S. McPherson: Design of microRNA delivery system used to perform experiments in Figure 4. Edited manuscript. -­‐Frédéric Charron: Design of Dunn Chamber Turning Assay used to perform experiments pertaining to Figures 2-­‐5. Edited manuscript. -­‐Timothy E. Kennedy: Design of project and edited manuscript -­‐Alyson E. Fournier: Design of project, analysis of data, writing and editing of the manuscript. Second Manuscript: “A cell-­‐intrinsic switch from Sonic hedgehog-­‐mediated attraction to repulsion of commissural axons after midline crossing requires 14-­‐3-­‐
3 proteins” -­‐Patricia T. Yam: Performed experiments, analysis and revisions pertaining to Figures 3, 4, 5, and 7 and Supplementary Figures 4. Design of project, analysis of data and writing and editing of manuscript. -­‐Christopher B. Kent: Performed experiments, analysis and revisions pertaining to Figures 4-­‐6 and Supplementary Figure 2. Design of project, analysis of data and writing and editing of manuscript. -­‐ W. Todd Farmer: Performed experiments, analysis pertaining to Figure 1 and Supplementary Figure 1. -­‐Steves Morin: Performed Western Blotting for Figure 5, and chick embryo electroporation in Figure 7. -­‐Ricardo Alcini: Performed experiments for Figure 6 -­‐Léa Lepelletier: Performed Immunohistochemistry experiments in Fig. 4 -­‐David R. Colman: Experiments were performed in the lab of Dr. Colman -­‐Marc Tessier-­‐Lavigne: Initial project design -­‐Alyson E. Fournier: Design of project, analysis of data, editing of the manuscript. -­‐Frédéric Charron: Design of project, analysis of data, writing and editing of the manuscript. Third Manuscript: “14-­‐3-­‐3 proteins are required for DCC mediated netrin-­‐1 signalling” -­‐Christopher B. Kent: Performed experiments, analysis and revisions pertaining to Figures 1-­‐5. Design of project, analysis of data and writing and editing of manuscript. -­‐Ricardo Alcini: Performed immunopreciptation experiments for Figure 4 -­‐Timothy E. Kennedy: Initial design of the project. -­‐Alyson E. Fournier: Design of project, analysis of data, editing of the manuscript. Abstract The extension of axons to their appropriate targets during development or regeneration relies on the growth cone, a specialized structure that integrates extracellular signals to guide axon growth. Achieving the many complex connections formed by the nervous system while relying on a limited number of cues, requires the growth cone to modulate its response to signals in both space and time. The mechanisms by which the growth cone is able to spatially and temporally regulate its underlying cytoskeletal response to guidance cues are only beginning to become understood. Proteomic analysis of isolated growth cones identified the 14-­‐3-­‐3 adaptor proteins as major constituents of the growth cone. Given their role in regulating spatial and temporal activity of multiple cell signalling events, 14-­‐3-­‐3s are excellent candidates to play a similar role in the growth cone. This thesis establishes the critical role of 14-­‐3-­‐3 proteins in the spatial and temporal regulation of growth cone response to axon guidance cues. In chapter 2, we describe how loss of 14-­‐3-­‐3 function in dorsal root ganglion (DRG) neurons, converts their response to gradients of nerve growth factor (NGF) and myelin associated glycoprotein (MAG) from repellent to attractive. This switch can be blocked by antagonizing cAMP-­‐dependent kinase (PKA), implicating 14-­‐3-­‐3 in the regulation of PKA activity. Consistently, we find that 14-­‐3-­‐3s interact with PKA and that disrupting this interaction results in the release of active catalytic subunits of PKA. In Chapter 3, we describe the role of 14-­‐3-­‐3 proteins and PKA in establishing a cell intrinsic, time dependent switch in the turning response of commissural interneurons to gradients of Sonic Hedgehog (SHH). The change from an attractive to a repellent response to SHH at the midline is critical in the proper formation of the axonal projections of dorsal commissural neurons in the developing spinal cord, and interfering with this switch in vivo, either by prematurely overexpressing 14-­‐3-­‐3s or by blocking 14-­‐3-­‐3 function results in aberrant axon growth. In Chapter 4, we identify a role for 14-­‐3-­‐3 proteins in mediating the signalling of another guidance cue used by commissural neurons, Netrin-­‐1. Together these studies demonstrate the 14-­‐3-­‐3s are a multifunctional family of adaptor proteins that provide a critical layer of regulation to axon guidance signals, allowing the growth cone to respond to cues appropriately during development and provide a potentially interesting target for modulating growth cone responses after axon injury. Résumé La projection des axones vers leurs cibles appropriées au cours du développement ou de la régénération s'appuie sur le cône de croissance, une structure spécialisée qui intègre les signaux extracellulaires pour guider la croissance axonale. La mise en place des connexions multiples et complexes formées par le système nerveux, tout en s'appuyant sur un nombre limité de signaux, requière que le cône de croissance module sa réponse à ces signaux tant dans l'espace que dans le temps. Les mécanismes par lesquels le cône de croissance est en mesure réguler la réponse spacio-­‐temporelle de son cytosquelette à des signaux de guidage commencent seulement à être compris. L'analyse protéomique de cônes de croissance isolés a identifié les protéines adaptatrices 14-­‐3-­‐3 en tant que constituants principaux du cône de croissance. Compte tenu de leur rôle dans la régulation de l'activité spatiale et temporelle de multiples événements de signalisation cellulaires, les protéines 14-­‐3-­‐3 sont d'excellents candidats pour jouer un rôle similaire dans le cône de croissance. Cette thèse établit le rôle essentiel des protéines 14-­‐3-­‐3 dans la régulation spatiale et temporelle de la réponse du cône de croissance à des signaux de guidage axonal. Dans le chapitre 2, nous décrivons comment la perte de fonction des protéines 14-­‐3-­‐3 dans les neurones du ganglion de la racine dorsale (DRG), convertit leur réponse à des gradients de facteur de croissance nerveuse (NGF) et glycoprotéine associée à la myéline (MAG) de la répulsion à l’attraction. Ce commutateur peut être bloqué par l’inactivation de la kinase dépendante à AMPc (PKA), ce qui implique 14-­‐3-­‐3 dans la régulation de l'activité de PKA. Dans ce sens, nous constatons que les protéines 14-­‐3-­‐3 interagissent avec PKA et que perturber cette interaction résulte en la libération de sous-­‐unités catalytiques actives de la PKA. Dans le chapitre 3, nous décrivons le rôle des protéines 14-­‐3-­‐3 et de PKA dans l'établissement d'un commutateur intrinsèque à la cellule, et dépendant du temps lors de l’orientation des axones des interneurones commissuraux soumis à un gradient de Sonic Hedgehog (Shh). Le changement d'une réponse attractive à une réponse répulsive à SHH au niveau de la ligne médiane est essentiel à la formation des projections axonales appropriées des neurones commissuraux dorsaux de la moelle épinière lors du développement, et interférer avec ce commutateur in vivo, soit en surexprimant prématurément les 14-­‐3 – 3, soit en bloquant la fonction des 14-­‐3-­‐3, entraîne une croissance axonale aberrante. Dans le chapitre 4, nous identifions un rôle pour les protéines 14-­‐3-­‐3 dans la médiation d'un autre signal de guidage utilisé par les neurones commissuraux, la nétrine-­‐1. Ensemble, ces études montrent que les protéines 14-­‐3-­‐3 sont une famille multifonctionnelle de protéines adaptatrices qui fournissent un niveau supplémentaire de régulation du guidage axonal, permettant au cône de croissance de répondre à des signaux de façon appropriée au cours du développement et de fournir une cible potentiellement intéressante pour la modulation de cette croissance après blessure axonale. List o f a bbreviations
ADAM
ADP
ADF
AKAP
ANOVA
AP
Arg-1
BDNF
BHK-21
BMPs
Boc
BP
BSA
Ca2+
[Ca2+]I
CA1
CA3
CaM
CaMK
CaMKII
CaMKK
cAMP
CAMs
CaN
cGMP
CICR
ClCNG
CNS
Comm
CoREST
Cre
CREB
CSPGs
DCC
DISC1
DIV
DMEM
DRG
DSCAM
DV
ECD
ECM
a disintegrin and metalloproteinase
adenosine diphosphate
actin depolymerizing factor
protein kinase A anchoring protein
analysis of variance
anterior-posterior
arginase1
brain derived neurotrophic factor
baby hamster kidney cells-21
bone morphogenic protein
Brother of CDO
blocking peptide
bovine serum albumin
calcium ion
intracellular calcium concentration
Cornu Ammonis area 1
Cornu Ammonis area 3
Calmodulin
Calmodulin dependent kinase
Calmodulin dependent kinase II
Calmodulin dependent kinase kinase
cyclic adenosine monophosphate
cell adhesion molecules
Calcineurin
cyclic guanosine monophosphate
calcium induced calcium release
chloride ion
cyclic nucleotide gated channels
central nervous system
Commisureless
corepressor of RE1 silencing transcription factor
causes recombination
cAMP response element-binding
chondroitin sulphate proteoglycans
deleted in colrectal cancer
deleted in schizophrenia 1
days in vitro
Dulbecco modified Eagle's minimal essential medium
dorsal root ganglion
Down syndrome cell adhesion molecule
dorsal-ventral
extracellular domain
extracellular matrix
EDTA
EGFP
Epac
ERK1/2
ERM
FBS
FKBP52
FLIP
FP
Fra
FRET
GAP
GAPDH
GDF
GEF
GST
GTP
HDAC2
HEK
HH
Hhip
HSV
I-1
ICD
IgSFCAM
IL-6
IP3
IP3R
JNK
Kuz
LC-Q-ToF
LF
LIM
LIM-HD
LIMK
LIS1
lLMC
LMC
L-VDCC
MAG
MAIs
MAP2
MAPK
MATH1
MCF-7
MDS
miR
mLMC
mRNA
ethylenediaminetetraacetic acid
enhanced green fluorescent protein
exchange protein activated by cAMP
extracellular-signal-regulated kinase 1/2
Ezrin/Radixin/Moesin
fetal bovine serum
FK506-binding protein 52
focal laser induced photolysis
floor plate
Frazzled
flouresence resonance energy transfer
GTPase activating protein
glyceraldehyde 3-phosphate dehydrogenase
growth and differentiation factor
guanine exchange factor
glutathione S transferase
guanosine triphosphate
histone deacetylase 2
human embryonic kidney
Hamburger-Hamilton
Hedgehog interacting protein
herpes simplex virus
inhibitory protein 1
intracellular domain
IgG superfamily of cell adhesion molecules
interleukin 6
inositol 1,4,5-trisphosphate
IP3 Receptor
c-Jun N-terminal kinase
Kuzbanian
liquid chromatography quantitative time of flight
lateral funniculus
Lin11/Isl-1/Mec-3 domain
LIM homeodomain
LIM kinase
Lissencephaly protein 1
lateral cells of the lateral motor column
lateral motor column
L-type voltage dependent calcium channel
myelin associated glycoprotein
myelin associated inhibitors
microtubule associated protein 2
mitogen activated protein kinase
mouse Atonal homologue 1
Michigan Cancer Foundation cells-7
Miller-Dieker syndrome
microRNA
medial cells of the lateral motor column
messenger ribonucleic acid
MS
MSP
Na+
NCAM
NF
NGF
NOS
Npn
NrCAM
NUDEL
OC
OCT
ONH
p75NTR
PAK1
PBS
PC-12
PCR
PFA
PI(3,4,5)P3
PICK1
PKA
PKAcat
PKC
PKD
PKG
PKI
PLL
PP1
PVDF
RIIα
α
β
RIIβ
RGC
RhoGDI
RIPA
RNAi
ROBO
ROS
Rp-cAMPS
RyR
sAC
SCF
SDF-1
SDS-PAGE
SEM
Sema
Ser
Shh
shRNAmir
mass spectrometry
macrophage stimulating protein
sodium ion
neuron cell adhesion molecule
Neurofilament
nerve growth factor
nitric oxide synthase
Neuropilin
neuronal cell adhesion molecule
nuclear distribution protein nudE-like 1
optic chiasm
optimal cutting temperature
optic nerve head
p75 neurotrophin receptor
p21-activated kinase 1
phosphate buffered saline
pheochromocytoma cells-12
polymerase chain reaction
paraformaldehyde
phosphatidylinositol (3,4,5)-triphosphate
protein interacting with PRKCA 1
cAMP dependent protein kinase
protein kinase A catalytic subunit
protein kinase C
protein kinase D
cGMP –dependent protein kinase
protein kinase A inhibitory peptide
poly-L-lysine
protein phosphatase 1
polyvinylidene fluoride membrane
type II protein kinase A regulatory subunit α
type II protein kinase A regulatory subunit β
retinal ganglion cell
Rho-GDP dissociation inhibitor radioimmunoprecipitation assay
ribonucleic acid interference
Roundabout
reactive oxygen species
adenosine- 3', 5'- cyclic monophosphorothioate, Rp- isomer
ryanodine receptor
soluble adenyl cyclase
stem cell factor
stromal cell-derived factor 1
sodium dodecyl sulfate-polyacrylamide gel electrophoresis
standard error of the mean
Semaphorin
serine
Sonic Hedgehog
short hairpin RNA with a microRNA stem
Smo
Sp-cAMPS
SSH1L
TAT
TESK
TGN
Thr
TRPC
UNC5H1
VAMP2
VEGF
VF
VT
WAVE1
YFP
Smoothened
adenosine- 3', 5'- cyclic monophosphorothioate, Sp- isomer
Slingshot phosphatase long
trans-acting activator of transcription
Testin kinase
trans Golgi network
threonine
transient receptor potential cation channel
Unc5 homologue 1
vesicle associated membrane protein 2
vascular endothelial growth factor
ventral funniculus
ventro-temporal
WASP family Verprolin-homologous protein 1
yellow fluorescent protein
Table o f c ontents Introduction 1. Spatial and temporal switches in axon guidance during development _________________________________________ 5 1.1 Commissural axon guidance 1.2 Retinal ganglion cell axon guidance 5 10 2. Regulated growth cone responses to guidance cues _____ 14 2.1 Receptor expression, localization and processing 2.1.1 Transcriptional control of receptor expression 2.1.2 Translational control of receptor expression 2.1.3 Receptor processing at the cell surface. 1 14 2.2 Calcium ion signalling 2.3 Cyclic nucleotides and crosstalk 23 29 3.1 14-­‐3-­‐3 structure and function 3.2 14-­‐3-­‐3s role in nervous system development 3.3 14-­‐3-­‐3 proteins regulate signalling responses to external cues 37 41 15 19 20 3. 14-­‐3-­‐3 proteins as spatial and temporal regulators of cell signalling _______________________________________________ 37 43 4. Thesis rationale and objectives ___________________________ 50 Results 1. 14-­‐3-­‐3 proteins regulate protein kinase A activity to modulate growth cone turning responses ________________ 53 1.1 1.2 1.3 1.4 PREFACE ABSTRACT INTRODUCTION MATERIALS AND METHODS 1.4.1 1.4.2 1.4.3 1.4.4 1.4.5 1.4.6 1.4.7 Reagents Dorsal root ganglion cultures Proteomics Immunofluorescence 14-­‐3-­‐3 knockdown and virus preparation Dunn chamber turning assay Immunoprecipitation 1.5 RESULTS 1.5.1 14-­‐3-­‐3 proteins are expressed in neuronal growth cones i 54 55 56 58 63 51 58 59 59 60 60 61 62 63 1.5.2 A 14-­‐3-­‐3 antagonist converts NGF-­‐dependent repulsion to attraction 1.5.3 A 14-­‐3-­‐3 antagonist blocks MAG-­‐dependent repulsion 1.5.4 Knockdown of specific 14-­‐3-­‐3 proteins converts NGF-­‐dependent repulsion to attraction 1.5.5 14-­‐3-­‐3 proteins switch NGF-­‐dependent growth cone turning responses through PKA 1.5.6 14-­‐3-­‐3 proteins bind and regulate PKA 64 65 68 1.6 DISSCUSSION 74 1.7 SUPPLEMENTARY INFORMATION 80 2.1 2.2 2.3 2.4 85 86 87 90 1.6.1 14-­‐3-­‐3 expression in growth cones 1.6.2 14-­‐3-­‐3 proteins regulate PKA 1.6.3 PKA and growth cone turning 2. A cell-­‐intrinsic switch from Sonic hedgehog-­‐mediated attraction to repulsion of commissural axons after midline crossing requires 14-­‐3-­‐3 proteins _______________ 84 PREFACE ABSTRACT INTRODUCTION MATERIALS AND METHODS 2.4.1 Animals 2.4.2 Reagents 2.4.3 DiI axon tracing 2.4.4 Shh protein gradient detection 2.4.5 Dissociated commissural neuron culture 2.4.6 Dunn Chamber axon guidance assay 2.4.7 Rat open-­‐book cultures 2.4.8 Chick electroporation 2.4.9 Immunostaining 2.4.10 Western blotting 2.5 RESULTS 94 2.5.1 Guidance of post-­‐crossing commissural axons in vivo along the AP axis requires Smoothened 2.5.2 Shh protein is present in a longitudinal gradient along the AP axis of the spinal cord 2.5.3 Commissural neurons switch their response to Shh from attraction to repulsion over time in vitro 2.5.4 14-­‐3-­‐3 protein expression in commissural neurons is also time-­‐
dependent 2.5.5 PKA activity decreases in post-­‐crossing commissural neurons and changes in PKA activity alter turning responses to Shh 2.5.6 Inhibition of 14-­‐3-­‐3 proteins converts Shh repulsion to attraction through PKA activity 2.5.7 Altering 14-­‐3-­‐3 proteins in vivo perturbs AP guidance of commissural axons 2.6 DISCUSSION 2.6.1 Shh has multiple roles in commissural axon guidance. ii 116 69 71 75 76 78 90 90 91 91 92 92 93 93 93 94 94 97 99 102 106 108 111 118 2.6.2 A novel role for 14-­‐3-­‐3 proteins in switching the polarity of the turning response to Shh 2.6.3 Molecular mechanisms involved in modulating responses to guidance cues 2.7 ACKNOWLEDGEMENTS 2.8 SUPPLEMENTARY INFORMATION 124 125 3.1 3.2 3.3 3.4 130 131 132 134 120 121 3. 14-­‐3-­‐3 proteins regulate commissural neuron responses to Netrin-­‐1 ____________________________________ 129 PREFACE ABSTRACT INTRODUCTION MATERIALS AND METHODS 3.4.1 3.4.2 3.4.3 3.4.4 3.4.5 3.4.6 3.4.7 3.4.8 Reagents Virus and recombinant proteins Dissociated neuron culture Dunn Chamber turning assays Collagen explant assays Cell surface biotinylation Immunoprecipitation Western blotting 3.5 RESULTS 134 134 135 135 135 136 136 137 137 3.5.1 14-­‐3-­‐3 protein function is required for growth cone expansion in response to acute Netrin-­‐1 treatment. 3.5.2 14-­‐3-­‐3 activity is required for attractive turning responses of pre-­‐crossing commissural neurons to Netrin 3.5.3 14-­‐3-­‐3 protein activity is required for Netrin-­‐1 induced neurite outgrowth from dorsal explants of embryonic spinal cord. 3.5.4 14-­‐3-­‐3s are required for Netrin-­‐1 induced activation of MAPK signalling and interact with DCC but do not impact Netrin-­‐1 induced changes to DCC distribution . 3.6 DISCUSSION 3.6.1 Spatial regulation of DCC in Netrin-­‐1 signalling 3.6.2 14-­‐3-­‐3s as regulators of downstream signalling proteins Discussion and Conclusion 137 139 139 141 144 149 1. Summary __________________________________________________ 151 2. 14-­‐3-­‐3 proteins as negative regulators of PKA _________ 152 3. Potential alternative mechanisms for 14-­‐3-­‐3 regulation of growth cone turning responses __________ 155 4. Roles for 14-­‐3-­‐3 regulation of axon guidance in other neuronal cell types _______________________________________ 157 5. Implications of 14-­‐3-­‐3 regulation of PKA beyond axon guidance ___________________________________________________ 159 iii 145 146 6. Importance of PKA regulation in axonal regeneration 161 7. Conclusion ________________________________________________ 162 Bibliography 163 iv List o f f igures Introduction
Figure 1: Spatial regulation of vertebrate commissural axon guidance at the
midline.
Figure 2: Growth cone turning responses regulated at the receptor level
Figure 3: Calcium signalling regulates attractive and repellent growth cone
turning responses.
Figure 4: Cyclic nucleotide signals regulate both attractive and repellent
growth cone turning responses
Figure 5: 14-3-3 binding regulates target proteins in different ways
Chapter 1
Figure 1: 14-3-3 proteins are present in growth cones
Figure 2: The 14-3-3 antagonist R18 converts NGF-dependent repulsion to
attraction in E13 chick DRG neurons
Figure 3: The 14-3-3 antagonist R18 blocks MAG-dependent repulsion of
E13 chick DRG neurons
Figure 4: Loss of 14-3-3ε, β or γ converts NGF-dependent repulsion to
attraction in P5 rat DRG neurons
Figure 5: 14-3-3 proteins regulate growth cone turning responses through
PKA.
Figure 6: 14-3-3 proteins bind PKA
Figure 7: 14-3-3 proteins regulate the stability and activity of PKA
holoenzyme.
Figure 8: Model for 14-3-3 regulation of PKA in growth cone turning
response.
Figure 1S: Growth cone purification strategy.
Figure 2S: Growth cone proteomics analysis
Figure 3S: Validation of 14-3-3 antibodies.
Figure 4S: 14-3-3 isoform expression in chick DRG lysates.
Figure 5S: Validation of the R18 peptide.
Chapter 2
Figure 1: Smoothened is required for proper post-crossing commissural axon
guidance along the AP axis.
Figure 2: Shh protein accumulates in an AP gradient in the neural tube.
Figure 3: Commissural axons are repelled by Shh at 3-4 DIV, but are
attracted by Shh at 2 DIV.
Figure 4: 14-3-3 protein expression in commissural neurons is timedependent.
Figure 5: PKA activity in commissural neurons is time-dependent and
determines turning response to Shh
v 8
15 26 35 40 63
66
67
69
71
73
74
77
80
81
82
83
83
95
98
100
104
107
Figure 6: Inhibition of 14-3-3 protein function converts Shh repulsion to
attraction.
Figure 7: Inhibition of 14-3-3 protein function in vivo perturbs AP guidance of
post-crossing commissural axons.
Figure 8: Overexpression of 14-3-3 γ or β in pre-crossing commissural axons
induces Shh repulsion and perturbs AP guidance in vivo.
Figure 9: Differential pre- and post-crossing commissural axon responses to
Sonic hedgehog are regulated by a cell-intrinsic 14-3-3 protein
dependent switch.
Figure 1S: Math1+ neurons are commissural neurons.
Figure 2S: Inhibition of 14-3-3 function with Tat-R18-YFP has no effect on
axon growth.
Figure 3S: Hhip1 does not play a major role in commissural axon guidance.
Figure 4S: cAMP levels are similar in commissural neuron growth cones at 2
and 4 DIV
Chapter 3
Figure 1: Inhibition of 14-3-3 function blocks Netrin-1 induced growth cone
expansion
Figure 2: Inhibition of 14-3-3 function blocks attractive turning to Netrin-1
gradients in commissural neurons
Figure 3: Loss of 14-3-3 function blocks Netrin-1 induced outgrowth from
dorsal spinal cord explants
Figure 4: 14-3-3s mediate Netrin-1 induced MAPK signalling and interact
with Netrin-1 receptor DCC but do not effect its distribution at cell
surface
Figure 5: Model for 14-3-3 regulation of MAP Kinase signalling in response
to Netrin-1 to DCC
vi 109
112
115
117
125
126
127
128
138
140
142
144
146
Introduction Introduction One of the most fundamental characteristics of the nervous system, is the multitude and complexity of connections between neurons, formed with amazing precision, often over very large distances. These connections are established in part by the targeted extension of axons to their appropriate destinations during development. Understanding this process is critical to ensure that it is properly re-­‐established during the regeneration of axons after injury. As with so many fundamental questions in neurobiology, the earliest findings came from Santiago Ramon y Cajal who speculated that the ability of the axon to arrive at its target relies on the growth cone, a specialized structure at the distal tip of the growing axon which, “may be regarded as a sort of club or battering ram, endowed with exquisite chemical sensitivity, with rapid ameboid movements, and with certain impulsive force, thanks to which it is able to proceed forward and overcome obstacles met in its way, forcing cellular interstices until it arrives at its destination.” (Ramon y Cajal 1909). Ramon y Cajal’s proposal that the growth cone acts as an exquisite chemotropic machine would have to wait nearly a century for the advent of molecular biology to establish that indeed the navigation of the growth cone through the complex environment of the developing nervous system is achieved through directed growth responses of growth cones, both attractive and repulsive, to a number of evolutionarily conserved extracellular signalling molecules. Over the last two decades, the powerful combination of forward genetic screens in model organisms such as flies, worms and mice and biochemical techniques have provided tremendous insight into the highly evolutionarily conserved families of signalling molecules used by organisms to achieve proper nervous system development. While new signalling modalities are still being discovered, the field has established four canonical groups of axon guidance ligand proteins: Netrins, Slits, Ephrins and Semaphorins (Sema). As well as their respective receptors: Deleted in Colrectal Cancer (DCC)/Unc5, -­‐ 3 -­‐ Introduction Roundabout (Robo), Eph receptors and Plexins/Neuropilins (Npn) (Huber, Kolodkin et al. 2003). Other families of proteins such as secreted morphogens, neurotrophins and cell adhesion molecules have also been found to have axon guidance roles. Together, these molecules can act as short-­‐range, contact–mediated cues, or as long-­‐range diffusible cues and are attractive or repellent in nature. They are often bifunctional, alternatively attracting or repelling the growth cone in different contexts. They have many unique characteristics, employing various regulatory and downstream signalling mechanisms. How the growth cone, responding to a limited number of cues, reliably integrates complex molecular and physical environmental signals to make navigation decisions is only now becoming understood (Kolodkin and Tessier-­‐Lavigne 2011). In response to the binding of guidance cues to cell surface receptors (Bashaw and Klein 2010), the growth cone asymmetrically regulates the molecular dynamics of its underlying cytoskeleton (Dent, Gupton et al. 2011), adhesive properties and membrane turnover(Winckler and Mellman 2010). Complex intracellular signalling cascades modulate its response to signals in both space and time, allowing the growth cone to make critical navigational choices at intermediate targets during development. This includes responding to one cue while ignoring others, or changing from an attractive response to a repulsive response to the same cue. The signalling mechanisms by which the growth cone is able to spatially and temporally regulate its response to guidance cues are only partially understood (Vitriol and Zheng 2012), but they are an important outstanding area of research as they can provide insight into the fundamental cellular biology of the growth cone and an opportunity to make a positive impact in neuronal repair after injury (Hur, Saijilafu et al. 2012). In this thesis, we identify a novel regulator of growth cone responses, the 14-­‐3-­‐3 family of adaptor proteins, and establish their role in the spatial and temporal regulation of axon guidance. -­‐ 4 -­‐ Introduction 1. Spatial and temporal switches in axon guidance during d evelopment Molecular mechanisms that regulate the growth cone’s ability to respond to external cues also provide the extending axon with the ability to switch between attractive and repellent responses at the appropriate place and time during the development of the nervous system. The growth cone integrates multiple signals to make a series of navigational decisions at intermediate points and ultimately ensure that the axon reaches its synaptic target. Studies of the projections of interneurons in the spinal cord have provided much insight as to how this process occurs during development. First we review this common model of axon guidance, highlighting the importance of spatio-­‐
temporal regulation of signalling in the formation of neural circuits. 1.1 Commissural a xon g uidance The dorsal region of the vertebrate spinal cord contains the cell bodies of interneurons that project axons, some ipsilaterally and some contralaterally towards more rostral targets, providing an important pathway for the different sides of the body to communicate to each other. In order for these projections to form correctly, the axons are confronted with a series of guidance challenges. They must first grow ventrally, reach and decide whether to cross the midline, then position themselves laterally in one of three distinct bundles and turn to grow along the longitudinal axis without re-­‐crossing the midline (Fig. 1). The highly stereotypic nature of the commissural projections and the straight forward nature of the navigational choices faced, has made this a powerful model for studying and understanding how spatial and temporal regulation allow the growth cone to address these challenges. Interestingly, the commissures formed by the dorsal interneurons of the developing Drosophila central nervous system (CNS) show much the same pattern and have been a useful genetic tool for -­‐ 5 -­‐ Introduction uncovering many of the highly evolutionarily conserved mechanisms of axon guidance (Dickson and Zou 2010; Evans and Bashaw 2010). Initially, commissural axons project towards the midline, growing ventrally towards a collection of ependymal cells at the midline known as the floor plate. They are guided to the floor plate through a combination of attraction and repulsion (Fig. 1). The dorsal equivalent of the floor plate, known as the roof plate, generates gradients of repellents including bone morphogen proteins, (BMPs), BMP7 and growth and differentiation factor 7 (GDF7), down which commissural axons are repelled (Augsburger, Schuchardt et al. 1999; Butler and Dodd 2003; Yamauchi, Phan et al. 2008). A recent report also identified another molecule from the roof plate, Draxin that repels commissural axons in vitro (Islam, Shinmyo et al. 2009), although in vivo experiments suggest that Draxin plays only a minor role in the initial repellent growth away from the roof plate. The floor plate also secretes a number of factors to generate attractive gradients along which the axons grow, including Netrin-­‐1(Serafini, Colamarino et al. 1996) and the morphogen, Sonic hedgehog (Shh) (Charron, Stein et al. 2003). Netrin-­‐1 attracts the axons of dorsal commissural interneurons to the midline, in part, through binding to its receptor, DCC (Serafini, Kennedy et al. 1994; Keino-­‐
Masu, Masu et al. 1996). Genetic studies knocking out the DCC homologue, Frazzled (Fra), indicated that Netrin commissural axon guidance to the midline most likely involves multiple receptors (Garbe and Bashaw 2007; Garbe, O'Donnell et al. 2007). A series of recent studies have identified down syndrome cell adhesion molecule (DSCAM) as a Netrin-­‐1 binding protein that mediates attractive turning in commissural axons in parallel to DCC (Andrews, Tanglao et al. 2008; Ly, Nikolaev et al. 2008; Liu, Li et al. 2009). These studies also indicate that DSCAM may mediate attraction to the midline through an additional -­‐ 6 -­‐ Introduction unidentified ligand as well, suggesting that more floor plate guidance cues are yet to be discovered. Interestingly DSCAM normally functions through homophilic binding to mediate repellent effects in synaptic elaborations (Hattori, 2008). How the activity of DSCAM is regulated in the context of heterophilic Netrin interactions in the spinal cord to mediate attractive turning is an important outstanding question. The floor plate also generates repellent cues to prevent inappropriate midline crossing by ipsilateral projections as well as stalling and re-­‐crossing by commissural axons that have already entered the midline (Fig. 1). Slits, Sema3B and Sema3F (Fig. 1)(Kidd, Bland et al. 1999; Zou, Stoeckli et al. 2000; Nawabi, Briancon-­‐Marjollet et al. 2010; Parra and Zou 2010) and Ephrin-­‐
B3(Kullander, Butt et al. 2003) form a repellent midline barrier at the floor plate. Thus raising the question of how contralateral projections manage to cross the floor plate at all. The spatial and temporal regulation of commissural axon sensitivity to repellent cues is critical for proper extension to and crossing of the midline. Members of the receptor family, Robo, mediate sensitivity to Slit and differential expression of Robo receptors in commissural neurons establish the precise spatial and temporal regulation of Slit signalling, providing sensitivity to this repellent cue only when the growth cone reaches and crosses the midline (Sabatier, Plump et al. 2004). Midline exit is also promoted by the repellent effects of Sema3B and Sema3F. How the responsiveness to these cues is induced is not clear but two mechanisms have been proposed. Local down-­‐regulation of the protease calpain allows accumulation of the receptor Plexin A1 (Nawabi, Briancon-­‐
Marjollet et al. 2010) and Shh signals at the floor plate are required for Sem3B signalling, likely by reducing the level of cAMP dependent protein kinase (PKA) activity (Fig. 1) (Parra and Zou 2010), which has been shown to silence Plexin signalling in Drosophila motor neurons (Yang and Terman 2012). To provide an extra boost to the axons exiting the midline, they also -­‐ 7 -­‐ Introduction gain sensitivity to Stem Cell Factor (SCF) present at the floorplate. Acting through the receptor Kit, SCF enhances the outgrowth of the axons propelling them out of the midline (Gore, Wong et al. 2008). Figure 1: Spatial regulation of vertebrate commissural axon guidance at the midline. (Left) Axons from commissural interneurons initially project from the dorsal (D) half of the developing spinal cord to the ventral (V) floorplate in response to gradients of both repellent (orange, -­‐-­‐-­‐) and attractive (blue, +++) guidance cues. (Right) Prior to reaching floor plate, the growth cone is insensitive to the midline repellent Slit, due to expression of the Robo splice isoform 3.1. Upon entering the midline, Shh induces a drop in PKA levels and Robo3.1 expression is switched for Robo 3.2. These changes, allow the growth cone to become responsive to midline repellents Slits and Sema3B. Also, activated Robo blocks DCC mediated Netrin attraction. Finally, the axons turn longitudinally in the anterior (A) direction in response to gradients of both attractive and repellent cues In addition to gaining sensitivity to repellents, commissural axons lose their attractive responses to Netrin and Shh (Lyuksyutova, Lu et al. 2003). How this is achieved is not particularly clear. Some insight as to how this may occur, was provided by the finding that there is a hierarchical relationship between guidance cue receptors (Stein and Tessier-­‐Lavigne 2001). Ectopic -­‐ 8 -­‐ Introduction expression of Robo in Xenopus spinal neurons, which normally are attracted to Netrin-­‐1, was sufficient to confer a repellent response to Slit in the presence of Netrin-­‐1. The cytoplasmic tail of the active Robo receptor binds and silences DCC signalling (Fig. 1, 2) and is thus dominant, preventing confusion at the midline. Expression of Robo is also essential to silence DCC mediated attraction to Netrin-­‐1 at the midline in dorsal motor neurons as they grow longitudinally (Bai, Chivatakarn et al. 2011). Also, chick commissural neurons that have crossed the midline have been shown to increase their expression of Hedgehog interacting protein (Hhip), a negative regulator of Shh signalling (Bourikas, Pekarik et al. 2005). After crossing the midline and leaving the floor plate, commissural axons make a longitudinal turn and extend along the rostral–caudal axis. The longitudinal extension of commissural axons is marked by a turn towards the anterior of the spinal cord. The signals that establish this turn are generated by the interplay of attraction and repulsion along gradients of the secreted morphogens, Wnt4 and Shh(Fig. 1)(Lyuksyutova, Lu et al. 2003; Bourikas, Pekarik et al. 2005). Post-­‐crossing commissural axons gain sensitivity to Wnt signals upon arrival at the contralateral border of the floorplate, by unidentified mechanisms (Lyuksyutova, Lu et al. 2003). A high rostral to low caudal gradient of Wnt4 exists along the mammalian spinal cord (Lyuksyutova, Lu et al. 2003) and Wnt4 acting through a non-­‐canonical transcription-­‐independent signalling pathway induces attractive turning of commissural axons (Wolf, Lyuksyutova et al. 2008). In chick spinal cords Wnts are expressed but do not exist in a gradient (Domanitskaya, Wacker et al. 2010). However, Shh does exist in a high caudal to low rostral gradient and has been shown to induce an anterior turn in post-­‐crossing commissural axons (Bourikas, Pekarik et al. 2005), in part by inducing expression of the Wnt antagonist Secreted frizzled-­‐related protein 1 (Domanitskaya, Wacker et al. 2010) but also by acting as a repellent cue. The regulatory mechanisms -­‐ 9 -­‐ Introduction responsible for generating the spatio-­‐temporal switch in turning response to Shh is not clear, but are explored in Ch.3 of this thesis. 1.2 Retinal g anglion c ell a xon g uidance Another area where many species share a set of highly stereotypic projections through relatively simple anatomy is the developing visual system. As the optic nerve forms, the axons from the retinal ganglion cells (RGCs) also face the challenge of reliably reaching a number of intermediate targets, midline sorting and integrating multiple cues to reach their synaptic targets in the tectum or superior colliculus. Despite some species differences relating to the degree of binocular vision, many of the signalling mechanisms responsible for axon guidance of RGCs are highly evolutionarily conserved and much insight into growth cone function and the underlying molecular pathways has been derived from studies focused on RGC projections. Many excellent reviews can be found that describe in detail these pathways (Erskine and Herrera 2007; Petros, Rebsam et al. 2008). Here we briefly discuss a few examples highlighting how different spatio-­‐temporal responses to different cues ensure proper navigational decisions at key choice points. The initial projections of RGC axons extend along the optic fibre layer along the inner surface of the retina in a radial pattern towards the exit point of the eye, a structure known as the optic nerve head (ONH). This first intermediate target is the source of attractive cues, drawing the axons directly to the beginning of the optic nerve and then ensuring proper exiting. A key signal that directs RGC axon growth to the ONH, appears to be Shh. This cue is present in the retina in a radial gradient, with high concentrations in the center, and has transcription independent growth promoting effect on young RGCs in in vitro assays. Further, blocking Shh signalling prevents ONH formation in vivo (Kolpak, Zhang et al. 2005), suggesting that Shh acts as an attractant. The glial cells surrounding the ONH are also a strong source of Netrin-­‐1 and in vitro, cultured RGCs from this stage of projection show -­‐ 10 -­‐ Introduction attractive turning responses to Netrin-­‐1, making it another strong candidate for acting as an attractive cue at the ONH (Deiner, Kennedy et al. 1997; Shewan, Dwivedy et al. 2002). However, DCC and Netrin-­‐1 mutant mice show unperturbed RGC axon growth to the ONH, suggesting a redundancy of attractive cues (Deiner, Kennedy et al. 1997). Interestingly, DCC and Netrin-­‐1 mutants showed a dramatic failure of RGC axons in exiting the eye (Deiner, Kennedy et al. 1997). This suggests that Netrin-­‐1 at the ONH could act as a short-­‐range cue, instructing axons to grow through the developing ONH. Netrin-­‐mediated exit at the ONH, could be a mix of attraction and repulsion. The retinal surface near the ONH strongly expresses the extra-­‐cellular matrix (ECM) protein laminin, which in vitro was found to convert the Netrin-­‐1 attractive response of RGCs to repulsion(Hopker, Shewan et al. 1999). This tight spatial regulation of Netrin-­‐1 response promotes the growth of axons into and through the ONH, where only Netrin-­‐1 is found. Subsequently, RGCs cultured from an immediately post-­‐ONH stage show no response in vitro to Netrin-­‐1 and as they continue to age and extend through the optic nerve, slowly switch their response to repulsion (Shewan, Dwivedy et al. 2002). The switch to repellent turning becomes important to restrain axon outgrowth upon reaching distal targets in the tectum (Shewan, Dwivedy et al. 2002). In species with binocular vision, RGC axons approach the ventral midline at a precise point and segregate into contralateral and ipsilateral projections, forming a characteristic x-­‐like structure, known as the optic chiasm (OC). While most axons from the retina project to the contralateral tectum, the optic chiasm also serves as a segregation point, with axons from the ventro-­‐
temporal (VT) area of the retina, projecting ipsilaterally. The VT axons are segregated, in part, by the precise temporal expression of the receptor EphB1 and ephrin B2 (Williams, Mann et al. 2003) to induce repulsion of these -­‐ 11 -­‐ Introduction axons from the contralateral tract. This guidance effect occurs with spatial and temporal precision, avoiding premature repulsion. Shh-­‐mediated repulsion has been demonstrated to also play a role in segregation of retinal axons at the OC (Trousse, Marti et al. 2001; Sanchez-­‐Camacho and Bovolenta 2008; Fabre, Shimogori et al. 2010). The role of Shh in regulating the formation of this commissure has been controversial. Some findings, suggest that, acting as a repellent, Shh serves to constrain and guide contralateral RGC axons through the developing OC (Trousse, Marti et al. 2001; Sanchez-­‐
Camacho and Bovolenta 2008). Another study showed that growth cones of ipsilateral axons express distinct members of the Shh signalling pathway such as the receptor, Brother of CDO (Boc), and indicate that Shh at the OC serves to repel these axons from the midline and segregate them to the ipsilateral pathway (Fabre, Shimogori et al. 2010). Beyond the manner in which ipsilateral and contralateral projections are differentiated, the mechanism by which this previously attractive cue is spatially regulated to become repellent is unknown, but identifying this could provide important insight as to the function of Shh signalling in the optic pathway as well as deepen our understanding of growth cone signalling. As highlighted by the previous examples, the shift in responsiveness of axons to external cues from attractive turning and growth promotion to repellent turning and inhibition of growth during development is a common feature of axon growth. Increasing evidence suggests that the failure of adult CNS axons to regrow after injury is due in part to the growth inhibitory environment of the CNS. A number of molecules present in the mature CNS have been identified that contribute to the inhibitory environment, including the canonical myelin assosicated inhibitors (MAIs), and chondroitin sulphate proteoglycans (CSPGs)(Giger, Hollis et al. 2010). Interestingly, the presence of repellent axon guidance cues in CNS myelin such as Semaphorins and Ephrins (Giger, Hollis et al. 2010) as well as repellent growth cone responses -­‐ 12 -­‐ Introduction to remaining bi-­‐functional cues such as myelin associated glycoprotein (MAG) (Filbin 2003) and Netrin-­‐1(Low, Culbertson et al. 2008) also play a major role in preventing the re-­‐growth of axons. Thus, understanding the molecular mechanisms that underlie the spatial and temporal switch from attraction to repulsion could provide critical insight necessary for targeted therapeutic intervention after CNS injury. -­‐ 13 -­‐ Introduction 2. Regulated g rowth c one r esponses t o guidance c ues As mentioned above, growth cones respond to various extracellular guidance cues in a bi-­‐functional manner. The ability to respond to cues in either a repellent or an attractive manner, and often switch between the two, is fundamental to the spatial and temporal regulation that underlies accurate axon guidance events through development. Whether the growth cone is attracted or repelled by a particular cue at a given point in space or time is determined by a variety of mechanisms. We will review here the mechanisms by which this occurs and their impacts on different axon guidance signalling events. 2.1 Receptor e xpression, l ocalization a nd p rocessing The ability of the growth cone to direct axon extension in response to an extracellular cue depends fundamentally on the binding of the ligand molecule to a receptor protein or complex of proteins expressed at the cell surface of the growth cone. Unsurprisingly then, one of the best described ways of regulating growth cone signalling concerns the expression, processing and localization of guidance cue receptors within the plasma membrane of the growth cone. Here we review the state of knowledge in understanding how the spatial and temporal activity of receptors is modulated and how this is used to confer the sensitivity and accuracy required by the growing axon. We will focus on three main mechanisms: the transcriptional control of receptor expression; local translation controlling receptor expression and localized regulatory events at the growth cone such as proteolytic cleavage and endocytosis. -­‐ 14 -­‐ Introduction Figure 2: Growth cone turning responses regulated at the receptor level Various regulatory mechanisms are employed to regulate the sensitivity and polarity of response to guidance cues. Attractive turning (left) at the midline to gradients of Netrin-­‐1 (gray) is mediated by DCC/Frazzled receptors (green). PKA activity up-­‐regulates the sensitivity of growth cones to Netrin-­‐1 by increasing the levels of DCC at the cell surface. PKC phosphorylation of Pick1 and its association with the repellent Netrin-­‐1 receptor, Unc-­‐5, down-­‐regulates its expression through endocytosis. Cleavage of DCC/Frazzled by γ-­‐secretase, generates an intracellular domain (ICD), which is thought to regulate the transcription of Comm. Comm expression (purple) down-­‐regulates the repellent Slit receptor, Robo (red) by re-­‐directing it to the late endosome. Repellent turning (right), is mediated by Slit binding to Robo, which in turn supresses DCC activity. Also Unc5/DCC complexes signal repellent turning in response to Netrin and DCC sensitivity is down-­‐regulated by shedding of its extra-­‐cellular domain (ECD) by the ADAM family metalloproteinase, Kuzbanian (Kuz). 2.1.1 Transcriptional control of receptor expression Axon guidance decisions are often predicated on the distinct expression pattern of the receptor proteins at the cell surface, and the combination of different receptors can play important roles in regulating the signalling outcomes. For example, Netrin-­‐1 attracts the axons of dorsal commissural interneurons to the midline through binding to its receptor, DCC (Serafini, -­‐ 15 -­‐ Introduction Kennedy et al. 1994; Keino-­‐Masu, Masu et al. 1996), while in other developmental contexts, Netrin-­‐1 acts as a repellent cue and this is mediated through binding to another receptor, Unc-­‐5 (Leonardo, Hinck et al. 1997). Genetic evidence, demonstrating that mutations in the DCC homologue, Unc 40 suppressed Unc5 dependent axon guidance events (Colavita and Culotti 1998), suggests that repellent axon guidance by Netrin-­‐1 may also require DCC. Further findings (Hong, Hinck et al. 1999) have demonstrated that expression of Unc5 in a DCC-­‐expressing cell is sufficient to convert attraction to repulsion and that the repulsion was dependent on DCC function. Netrin binding to DCC in the presence of Unc5, induces the formation of a DCC/Unc5 complex which confers a repellent response (Fig. 2). Short range Netrin -­‐1 repulsion can be mediated by Unc-­‐5 alone, whereas repulsion by the DCC/Unc-­‐5 complex occurs developmentally over longer distances (Keleman and Dickson 2001), presumably due to the increased binding sensitivity of DCC. The expression of different guidance cue receptors and the subsequent spatial segregation and temporal ordering of guidance cue responses has been established by several studies to be a result of regulated gene expression (Butler and Tear 2007) in various axonal populations. Transcriptional regulation of Eph receptor expression is used to establish temporal and spatial segregation of axon guidance responses. In the retina, segregation of ipsilateral RGC axons at the OC depends on the precise temporal expression of EphB1 and ephrin B2 (Williams, Wu et al. 2003), achieved by the regulation of transcription factors, including Zic2 (Herrera, Brown et al. 2003) and FoxD1 (Herrera, Marcus et al. 2004).This can also be seen in the projections of motor neuron axons to the developing limb bud in vertebrates. Lateral motor column (LMC) motor neurons form a simple binary topographic map, with medial cells (mLMC) projecting axons to the ventral limb bud and lateral cells (lLMC) projecting axons dorsally. The two -­‐ 16 -­‐ Introduction distinct sets of neurons are established by a cross-­‐repression feedback loop between two LIM homeodomain(LIM-­‐HD) transcription factors, Lim1, expressed in the lLMC and Isl1 expressed in the mLMC (Kania and Jessell 2003). The segregation of axons is then achieved by Lim1 driving expression of the EphA4 receptor (Kania and Jessell 2003) and Isl1 driving expression of the EphB1 receptor (Luria, Krawchuk et al. 2008) allowing the axons to respond to different Ephrin ligands present in the limb bud. LIM HD transcription factors Lhx2/9 and Lhx 1/5 have also been found to delineate subpopulations of dorsal interneurons (Wilson, Shafer et al. 2008; Avraham, Hadas et al. 2009) and influence the determination of ipsilateral or contralateral projections, presumably through Lhx2/9 driven expression of the Robo, Rig1 (Wilson, Shafer et al. 2008). In addition to segregating responsive populations of axons, transcriptional control has also been shown to be involved in establishing a critical switch in growth cone responses at the midline. As previously mentioned, the projections of commissural axons across the midline rely on the spatial and temporal coordination of attraction to Netrin and subsequent repulsion by Slit. Insights as to how commissural axons regulate the spatio-­‐temporal sensitivity to Slit were provided by a genetic screen in Drosophila, which revealed the requirement of the trans-­‐membrane protein, Commissureless (Comm) for the formation of commissural axon tracts (Seeger, Tear et al. 1993; Tear, Harris et al. 1996). Further work established that expression of Comm in pre-­‐crossing axons, generated a Comm-­‐Robo complex that reroutes Robo-­‐containing vesicles from the trans Golgi network (TGN) to the lysosome, preventing the expression of Robo at the cell surface (Fig. 2). (Keleman, Rajagopalan et al. 2002; Keleman, Ribeiro et al. 2005). Recent work has also demonstrated that the DCC homologue, Frazzled plays two critical roles in commissure formation. Aside from mediating the Netrin-­‐1 attractive signal, the cytoplasmic domain of Fra is required for the regulated -­‐ 17 -­‐ Introduction expression of comm messenger ribonucleic acid (mRNA) in pre-­‐crossing axons (Fig. 2)(Yang, Garbe et al. 2009). Surprisingly, this function is cell autonomous and independent of netrin signalling, suggesting either a direct transcriptional regulatory role for Fra or the binding of some other unidentified ligand. Finally, at the midline by mechanisms that are still unclear, Comm mRNA levels drop and expression is down-­‐regulated (Georgiou and Tear 2002; Keleman, Rajagopalan et al. 2002). Consequently, Robo expression is elevated allowing for Robo-­‐mediated Slit repulsion. Efforts to identify the mammalian homologue of Comm have to date been unsuccessful, however the ability to regulate Robo activity in commissural axons relies on another transcriptional mechanism. There are three unique Robo genes, Robo1, 2 and 3. A splice variant of Robo3, (Robo3.1 or Rig-­‐1) exists which is required for crossing of the midline (Sabatier, Plump et al. 2004) (Chen, Gore et al. 2008). Pre-­‐crossing axons express both Robo3.1 and low levels of Robo1. Blocking Robo3.1 expression increases Slit responsiveness, however it has no effect on the level of Robo1 expression in pre-­‐crossing axons (Sabatier, Plump et al. 2004). This indicates that Robo3.1 serves to mediate growth towards the midline by down-­‐regulating Robo1 mediated Slit repulsion, but in a manner distinct from Comm. Interestingly, after reaching and crossing the midline, commissural axons show a precise spatial change in expression of Robo3 isoforms. The Robo3.1 isoform is no longer expressed and axons then express Robo3.2 along with Robo1 and 2, which is found to be cooperative in mediating Slit repulsion (Fig. 1) (Chen, Gore et al. 2008). Thus, the precise spatial control of expression of multiple receptor proteins is critical for accurately conveying guidance information to these axons. In addition to playing a critical role in establishing commissural axon pathways, the Robo receptors are also involved in maintaining the -­‐ 18 -­‐ Introduction longitudinal tracts of ipsilateral neurons in the developing Drosophila CNS. The “Robo code” (Simpson, Kidd et al. 2000) in which an orthogonal array of three distinct longitudinal tracts is organized by the combinatorial effect of expressing either Robo1 alone, Robo1 and Robo3, or all three Robos (Rajagopalan, Nicolas et al. 2000; Rajagopalan, Vivancos et al. 2000), was assumed to be generated by the interaction of the unique structural or biochemical properties of the different receptors. However, more recent findings, in which the coding regions of the robo genes were swapped for each other demonstrated that the expression of a different Robo receptor in place of the “coded” one had no effect on the positioning of the axon. This suggests that the code is one of transcriptional control (Spitzweck, Brankatschk et al. 2010). Differences in gene expression, not in receptor domain structure account for spatial differences in responses to Slit. 2.1.2 Translational control of receptor expression Although most characterized upstream regulation of guidance cue receptor gene expression has been focused on the transcriptional level, work that has demonstrated that local mRNA translation in the growth cone is required for proper axon guidance responses (Jung, Yoon et al. 2012) suggests the possibility that factors regulating local translation of receptor mRNA could be a powerful mechanism for providing tight spatio-­‐temporal control of growth cone signalling. This suggestion is supported by findings that Xenopus spinal neurons rely at least partially on local protein synthesis to recover cell surface receptor levels and sensitivity to Netrin-­‐1 and Sema3A, after exposures to high concentration of ligands (Piper, Salih et al. 2005). Also, commissural axons crossing the midline show increased expression and translation of EphA2 mRNA in the distal tip of the axon after crossing (Brittis, Lu et al. 2002). Thus up-­‐regulating the repellent response to ephrins located at the midline. -­‐ 19 -­‐ Introduction Furthermore, recent work supports the idea that local translational mechanisms may play an important role in altering turning responses with time. Using laser-­‐capture microdissection and microarray techniques, the mRNA profiles of early pathfinding RGC growth cones were compared to older ones. This study revealed the specialized localization of different mRNAs to growth cones of different ages (Zivraj, Tung et al. 2010). Another study, looking at the mechanisms underlying the late gain of responsiveness to Sema3A signals in RGCs, found that the temporal regulation of this signal was mediated by the expression of a microRNA (miR) (Baudet, Zivraj et al. 2012). miR-­‐124 regulates the expression of a Neuropilin1 repressor, corepressor of RE1 silencing transcription factor (CoREST). It targets and down-­‐regulates the mRNA of CoREST, reducing local translation, and releases the repression of Sema3A signalling when the growth cone reaches the appropriate age. One recent study has identified a potential regulatory factor, showing that RNA binding protein, Musashi1 is critical for driving protein translation of Robo3/Rig1 in precerebellar neurons and is required for midline crossing (Kuwako, Kakumoto et al. 2010). While examples of post-­‐
transcriptional regulation of receptor mRNA are few, this is an area for many potentially important findings. 2.1.3 Receptor processing at the cell surface Regulated expression of different combinations of guidance cue receptors provide important spatial segregation and temporal ordering of sensitivity to extracellular ligands in populations of neurons, but the level of precision required in regulating these signalling pathways often involves mechanisms that can act locally at the growth cone in very short time frames. Receptor proteins at the growth cone are subject to regulation by mechanisms such as proteolytic cleavage, induced endocytosis, targeted delivery to the cell surface and segregation into lipid rafts. All of these can establish exquisite control of guidance signalling in individual growth cones. -­‐ 20 -­‐ Introduction The targeted enzymatic cleavage of receptor proteins by proteases is a powerful regulatory mechanism that can have a wide variety of effects. Receptors can be silenced and cleared from the cell surface, or cleaved to generate unique extra-­‐ or intracellular signalling fragments that serve as regulators or downstream effectors. A number of examples of proteolytic cleavage involved in axon guidance exist. This possibility was first suggested by axon guidance defects in Dropsophila mutants with loss of the kuzbanian allele (Fambrough, Pan et al. 1996). Kuzbanian (Kuz) or ADAM10 is an a disintegrin and metalloproteinase (ADAM) family metalloproteinase, which was subsequently found to associate with ephrins (Hattori, Osterfield et al. 2000) or Eph receptors (Janes, Saha et al. 2005). This suggested a model to resolve an important issue. How could heterophillic ligand–receptor complexes that bind in trans with high efficiency mediate a cell contact repellent effect? Upon EphA3 binding to ephrin2 (Hattori, Osterfield et al. 2000), changes in confirmation occur (Janes, Saha et al. 2005) which expose target motifs for the associated ADAM10. Cleavage of the ephrin ligand occurs, enabling withdrawal of the growth one and the initial adhesion becomes repulsion (Hattori, Osterfield et al. 2000). Metalloproteinase activity has also been implicated in regulating the sensitivity of commissural axons to Netrin (Galko and Tessier-­‐Lavigne 2000). An inhibitor of metalloproteinases blocks the shedding of an extracellular domain of DCC and promotes the accumulation of DCC on the cell surface as well as increasing the axon-­‐growth promoting effects of netrin (Fig. 2). Extracellular domain cleavage of DCC must be combined with the proper regulation of Presenilin, a component of γ-­‐secretase, to ensure the proper patterning of axonal projections in the spinal cord. Presenilin activity is critical in motor neurons to prevent aberrant attraction to netrin and proper Slit repulsion (Bai, Chivatakarn et al. 2011). Extracellular domain (ECD) shedding of DCC creates intracellular stubs at the plasma membrane that are -­‐ 21 -­‐ Introduction able to bind full length DCC and prevent its association with and silencing by Robo. Cleavage of the intracellular stub of DCC by γ-­‐secretase, prevents this from occurring. Cleavage of receptors before they reach the plasma membrane can also play an important regulatory role. In vertebrate commissural neurons, post midline crossing axons are responsive to the repellent Sema3B. Prior to crossing, Sema3B signalling is suppressed by calpain cleavage of the co-­‐receptor Plexin A1(Nawabi, Briancon-­‐Marjollet et al. 2010). Calpain activity is down regulated at the midline, thus allowing Plexin A1 accumulation at the growth cone and inducing the appropriate spatio-­‐temporal sensitivity to Sema3B. Semaphorin signalling can also be modified by targeting the process of endocytosis and receptor internalization. Repellent growth cone collapse response to Sema3A induces the endocytosis of receptors Plexin and Neuropilin1 (Fournier, Nakamura et al. 2000). The internalization of the receptor complexes is mediated in part by L1, a member of the IgG superfamily of cell adhesion molecules (IgSFCAM)(Castellani, Falk et al. 2004). The regulation of L1 by other IgSFCAMs in various axon guidance contexts can lead to rapid desensitization or conversion to attractive responses to Sema3A by respectively, enhancing or blocking the endocytosis response to Sema3A (Bechara, Falk et al. 2007). Converting repellent responses to attractive ones through regulated endocytosis can also be seen in the context of Unc-­‐5/netrin signalling. Activation of protein interacting with PRKCA 1(PICK-­‐1) and subsequently protein kinase C alpha (PKCα) is sufficient to induce the formation of PICK-­‐1/ PKCα/Unc-­‐5 Homolouge 1(UNC5H1) complex in mammalian cells (Williams, Wu et al. 2003). This leads to the endocytosis of the complex and removal of UNC5H1 from the cell surface (Fig. 2), which in postnatal cerebellar neurons induces a switch from repellent to attractive responses to Netrin-­‐1 (Bartoe, McKenna et al. 2006). -­‐ 22 -­‐ Introduction In contrast to the PKCα mediated removal of Unc-­‐5 receptors from the cell surface, other findings have suggested that increased PKA (Bouchard, Moore et al. 2004) or decreased RhoA activity (Moore, Correia et al. 2008) can increase the insertion of DCC at the cell surface from an intracellular pool, in response to Netrin signalling, enhancing the sensitivity of commissural neurons to Netrin-­‐1(Fig. 2). Directed sub-­‐cellular localization of DCC is also supported by genetic experiments in C.elegans showing the active form of the RhoGTPase, MIG-­‐2 and upstream regulators can alter the sub-­‐cellular localization of Unc-­‐40 (DCC) and the Slit receptor, Sax-­‐3 (Levy-­‐Strumpf and Culotti 2007; Watari-­‐Goshima, Ogura et al. 2007). In addition to the directed trafficking of receptors to the cell surface, increasing evidence is emerging to indicate that receptor activity is dependent upon association with particular lipid microdomains. Localization of receptors to cholesterol and glycosphingolipid enriched “lipid rafts” occurs upon binding of Netrin-­‐1, brain derived neurotrophic factor (BDNF) or Sema3A, and lipid raft integrity is required for growth cone turning responses (Guirland, Suzuki et al. 2004). Additionally, DCC localization to lipid rafts has been shown to alter its association with downstream signalling effectors (Herincs, Corset et al. 2005; Petrie, Zhao et al. 2009). The impact of lipid raft associations on axon guidance events in vivo has yet to be identified but it could provide yet another mechanism for providing the exquisite spatial and temporal control over signalling required for accurate axonal pathfinding. 2.2 Calcium I on S ignalling While guidance cue receptors are very common targets for regulating growth cone response, the direction turned by the growth cone upon ligand binding also depends, in large part, on the state of diverse intracellular signalling mechanisms. Changes in intracellular calcium ion (Ca2+) concentration, or -­‐ 23 -­‐ Introduction [Ca2+]I, occur in many cellular contexts and provide a powerful way to convey very fast, frequency based signals globally or in highly localized micro-­‐ or nanodomains. [Ca2+]I changes impact a wide variety of downstream effectors and can be generated in a number of different ways, making it an ideal way to provide spatial and temporal regulation to axon guidance effects. Early efforts to understand how molecular cues were used to direct outgrowth of the growth cone focused on [Ca2+]I signals, as [Ca2+]I levels and temporal patterns had been shown to have important effects on the rate of axon outgrowth and growth cone motility (Gomez and Spitzer 1999). Ratiometric imaging of Ca2+ sensitive fluorophores expressed in growth cones of Xenopus spinal neurons revealed that application of a Netrin-­‐1 gradient induces attractive turning and a corresponding [Ca2+]I gradient, with high [Ca2+]I seen on the side of the growth cone exposed to Netrin-­‐1 (Hong, Nishiyama et al. 2000). The [Ca2+]I gradient is necessary for the Netrin turning response as it was abolished by completely blocking membrane Ca2+ channels or depleting intracellular Ca2+ stores. Further studies demonstrated that a [Ca2+]I gradient was also sufficient to induce growth cone turning. Focal laser induced photolysis (FLIP) was used to asymmetrically uncage Ca2+ in a growth cone causing a local and transient increase in [Ca2+]I, and this drove turning in the direction of the increased [Ca2+]I (Zheng 2000). Beyond establishing changes in [Ca2+]I as a key signalling event in growth cone turning responses, these experiments also revealed that altering the conditions of [Ca2+]I signalling could lead to bi-­‐directional turning responses. Greatly reducing the amount of Ca2+ available, either by blocking L-­‐type voltage dependent calcium channels (L-­‐VDCC) or by trapping closed the ryanodine receptors (RyR), responsible for Ca2+ induced Ca2+ release (CICR), switches the turning response to Netrin-­‐1 from attractive to repellent (Hong, Nishiyama et al. 2000). Furthermore, reducing resting levels of [Ca2+]I converted the FLIP-­‐induced attractive turning to repulsion (Zheng 2000) and -­‐ 24 -­‐ Introduction high frequency [Ca2+]I transients in growth cone filopodia also induced repellent turning (Gomez, 2001). These findings suggested a model whereby binding of a guidance cue induced a localized change in [Ca2+]I towards the side of ligand binding, but that the direction of the resulting turn was determined by the amplitude of the [Ca2+]I signal (Fig. 3). Low amplitude [Ca2+]I changes resulted in repulsion and high amplitude [Ca2+]I changes induced attraction. This model is supported by experiments examining the bi-­‐directional turning response to gradients of MAG. Normally repellent responses to MAG induce a [Ca2+]I gradient through CICR from intracellular stores. But depolarizing the cell, which drives up the amount of CICR and the levels of [Ca2+]I converts MAG turning to attractive (Henley, Huang et al. 2004). Furthermore, the same study found direct evidence for amplitude controlling direction. Asymmetric [Ca2+]I , induced by a gradient of ionomycin, which makes the cell permeable to Ca2+ ions, is sufficient to induce growth cone turning . The direction of turning can be controlled by increasing or decreasing the concentration of Ca2+ in the media to achieve attraction or repulsion respectively (Henley, Huang et al. 2004). The mechanisms by which the growth cone differentiates between a high amplitude and low amplitude [Ca2+]I signal to either turn towards or away from an external cue remain poorly understood, but a few different downstream effectors have been characterized. Activation of the protease calpain by [Ca2+]I transients in filopodia has been shown to mediate growth cone repulsion through regulation of phosphotyrosine kinases and proteases (Robles, Huttenlocher et al. 2003). In general, [Ca2+]I signals are tightly buffered and remain localized in part by the high binding affinity of the Ca2+ sensor, calmodulin (CaM) (Faas, Raghavachari et al. 2011). Ca2+ bound CaM activates a number of different downstream signalling pathways including various kinases and phosphatases. One study examining effectors of [Ca2+]I -­‐ 25 -­‐ Introduction Figure 3: Calcium signalling regulates attractive and repellent growth cone turning responses. (Left) Attractive turning is mediated by high amplitude [Ca2+]I signals that activate downstream kianses. [Ca2+]I signals can be driven through entry of extracellular Ca2+ either by depolarizing membrane potentials activating voltage dependent calcium channels (VDCC, red) or activation of TRP channels (TRPC, blue). Also release of Ca2+ from intracellular stores, through activation of Ryanodine receptors or IP3 receptors (green), can contribute to the [Ca2+]I signals. (Right) Low amplitude [Ca2+]I signals mediate repellent turning by activating downstream phosphatases. Hyperpolarizing membrane potentials inhibit VDCCs and cyclic nucleotide gated (CNG, blue) channels are activated instead. signalling in growth cones has proposed a model for bidirectional response based on the differential CaM activation of the kinase, Calmodulin dependent kinase II (CaMKII) and phosphatase, calcineurin, CaN (Wen, Guirland et al. 2004)(Fig. 3). The use of pharmacological inhibitors in FLIP induced Ca2+ release turning assays demonstrate that CaMKII activity is required for attractive turning and CaN activity, and subsequently protein phosphatase 1 (PP1) phosphatase activity, is required for repulsion (Wen, Guirland et al. 2004). Given that CaN is known to require much lower [Ca2+]I signals to -­‐ 26 -­‐ Introduction become activated than CaMKII, this led to the proposal that a low amplitude [Ca2+]I signal activates a CaN/PP1 pathway which also inhibits CaMKII, and leads to repulsion, but that high amplitude [Ca2+]I signals preferentially activate CaMKII and lead to attractive turning. Involvement of this pathway in response to a guidance cue was demonstrated by converting Netrin-­‐1 repulsion on a laminin substrate to attraction by blocking CaN and PP1 activity (Wen, Guirland et al. 2004). More recent studies have focused on the impact of direct manipulation of [Ca2+]I and second messengers on growth cone membrane organization through the regulation of vesicle trafficking and processes such as vesicle associated membrane protein 2 (VAMP2) exocytosis and clathrin mediated endocytosis (Tojima, Hines et al. 2011). While direct links to axon guidance signalling events in vivo remain to be established, the proposal of [Ca2+]I signalling providing an organizational axis determining bi-­‐directional turning responses by controlling the balance of exo-­‐ and endocytosis is potentially a powerful mechanistic model for these complex signalling events. The importance of [Ca2+]I signalling to growth cone turning outcomes is well established but emerging evidence is beginning to describe how [Ca2+]I signalling itself is regulated in the growth cone. Localized increases in [Ca2+]I can be achieved primarily in two ways. The opening of ion channels can allow for entry of extracellular Ca2+, which in turn can induce the release of Ca2+ from intracellular stores in the endoplasmic reticulum, also known as CICR (Fig. 3). Both these processes are mediated by a number of receptors and transmembrane proteins. As previously mentioned, early studies establishing the importance of [Ca2+]I signals in growth cone turning demonstrated the requirement of both the entry of extracellular Ca2+ and CICR for regulated turning responses (Hong, Nishiyama et al. 2000; Henley, Huang et al. 2004). Changes in [Ca2+]I achieved through release from intracellular Ca2+ stores, occur through ryanodine receptor (RyR) mediated CICR or IP3R (IP3 -­‐ 27 -­‐ Introduction receptor) detection of increases in Inositol 1,4,5-­‐triphosphste (IP3) (Fig. 3). Both the efficiency of CICR through RyRs (Ooashi, Futatsugi et al. 2005; Tojima, Itofusa et al. 2009) and the activation of IP3Rs (Akiyama, Matsu-­‐ura et al. 2009) are thought to be regulated in response to differing cell adhesion molecules and their effects on second messenger pathways such as cyclic nucleotides and nitric oxide synthase (NOS). Entry of extracellular Ca2+ through L-­‐type voltage gated Ca2+ channels (Hong, Nishiyama et al. 2000) have been shown to be critical for turning responses. The activity of these selective ion channels can be regulated directly through cyclic nucleotides (Nishiyama, Hoshino et al. 2003) or controlled by membrane potential. Thus, electrical and chemical treatments that depolarize the cell can have profound effects on growth cone turning, through increases in [Ca2+]I (Ming, Henley et al. 2001; Henley, Huang et al. 2004) (Fig. 3). Depolarizing potentials alter [Ca2+]I signals and convert repellent signals to attractive ones. Furthermore, binding of guidance cues has also been found to shift the membrane potential of cells through sodium ion (Na+) or chloride ion (Cl-­‐ ) channels, and thus regulate [Ca2+]I signals. Repellent cues such as Slit and Sema3A have a hyperpolarizing effect and attractive cues such as Netrin and BDNF depolarize the cell (Fig. 3) (Nishiyama, von Schimmelmann et al. 2008). Besides the requirement for voltage dependent calcium channel (VDCC) activity, several studies have demonstrated that turning responses to guidance cues require the activation of members of the transient receptor potential channel (TRPC) family to sustain the influx of extracellular calcium (Fig. 3) (Li, Jia et al. 2005; Shim, Goh et al. 2005; Wang and Poo 2005). This occurs through the asymmetric activation of Akt and accumulation of phosphatidylinositol (3,4,5)-­‐triphosphate (PI(3,4,5)P3) (Henle, Wang et al. 2011). TRPC channel activation occurs in response to and is required for both attractive turning to Netrin-­‐1 and BDNF as well as repellent turning to MAG, making it a general mechanism for generating [Ca2+]I signals. -­‐ 28 -­‐ Introduction Interestingly, one study suggests that regulation of TRP channels may play an important role in commissural axon guidance as TRPC1 activation by netrin-­‐
1/DCC signalling is regulated through isomerization in the presence of the immunophilin, FK506-­‐binding protein 52 (FKBP52) (Shim, Yuan et al. 2009). How the regulation of [Ca2+]I signalling impacts axon guidance in vivo remains elusive as until recently the ability to accurately image [Ca2+]I signals in living animals has been quite limited. The introduction of new optical techniques such as optogenetics and two-­‐photon microscopy should open the door to new findings linking our understanding of [Ca2+]I signalling to models of the temporal and spatial regulation of axon guidance decisions that occur in vivo. The regulation of [Ca2+]I signalling, both the generation of changes in [Ca2+]I and its impacts on downstream effectors, has provided powerful insights as to the ways in which growth cones convert the detection of extracellular cues into the activation of secondary messengers and downstream intracellular signalling cascades. The complex regulation of secondary messengers such as cyclic nucleotides and crosstalk with other downstream effectors, discussed in the next section, provides another critical layer of spatial and temporal control over growth cone signalling. 2.3 Cyclic n ucleotides a nd c rosstalk One of the first indications that the turning response of a growth cone to a gradient of an extracellular cue is subject to regulation by the state of intracellular signalling mechanisms, came from efforts to understand downstream signalling cascades that differentiate attractive turning responses from repellent ones. Studies with gradients of soluble cues known to be attractive such as brain derived neurotrophic factor (BDNF) and Netrin-­‐
1, (Ming, Song et al. 1997; Song, Ming et al. 1997) as well as those known to be repellent such as MAG and Sema III (Song, Ming et al. 1998), applied to Xenopus spinal neurons, revealed that the polarity of turning response is not -­‐ 29 -­‐ Introduction a fixed characteristic of the cue applied, with distinct downstream mechanisms. Attractive cues can be converted to repellent ones by blocking the activity of cyclic nucleotide signalling and conversely repellent cues can induce attractive responses if cyclic nucleotide signalling activity is elevated. Guidance cues can signal to either attractive or repellent pathways depending on the cellular context. These early studies quickly established a simple grouping of guidance cues. Type I cues are those subject to regulation by the activity of cyclic adenosine monophosphate (cAMP) and its effectors, such as cAMP-­‐dependent protein kinase (PKA), and type II are those regulated by cyclic guanine monophosphate (cGMP) and its effectors such as cGMP –dependent protein kinase (PKG)(Song, Ming et al. 1998; Ming, Henley et al. 2001; Dontchev and Letourneau 2002). It has now been well established that cyclic nucleotide regulation of guidance cue response provides a critical mechanism for axon guidance decisions in vivo. Retinal ganglion cells (RGCs) are initially attracted to Netrin at the optic nerve head, however as they continue to project along the optic nerve, this initial attraction switches to repulsion, mediated by a drop in cAMP levels which is both intrinsic (Shewan, Dwivedy et al. 2002) and driven by increases in the presence of laminin-­‐1 (Hopker, Shewan et al. 1999). Other external factors, such as Shh have also been shown to mediate spatial regulation of guidance cue responses through down-­‐regulating cAMP levels (Trousse, Marti et al. 2001; Parra and Zou 2010). The spatial regulation of cAMP levels in RGCs is further refined by the presence of the chemokine stromal cell-­‐derived factor 1 (SDF-­‐1). This extracellular cue has no intrinsic guidance effects (Chalasani, Sabelko et al. 2003), but down-­‐regulates RGC responsiveness to the midline-­‐repellent, Slit, through activating an adenylate cyclase and maintaining elevated cAMP levels (Chalasani, Sabelko et al. 2003; Xu, Leinwand et al. 2010), ensuring proper midline crossing at the optic chiasm. -­‐ 30 -­‐ Introduction The importance of understanding the role of cyclic nucleotide signalling in growth cones is highlighted by studies demonstrating that the elevated levels of cAMP in younger neurons underlie their ability to grow on normally inhibitory substrates such as MAG or myelin. A drop in neuronal cAMP levels during development contributes to the inhibitory growth responses to myelin and the inability of axons to re-­‐grow after injury (Cai, Qiu et al. 2001). Treatments to re-­‐elevate cAMP in older neurons, and activate PKA, result in increased axon growth on inhibitory substrates and some regeneration of axons in the spinal cord after injury (Neumann, Bradke et al. 2002; Qiu, Cai et al. 2002) as well as increased functional recovery (Pearse, Pereira et al. 2004). Further work on this critical regulatory pathway can build on the promise of these early findings. Because of the central role played by cyclic nucleotides and their effectors in regulating growth cone function, much work has been done to try to understand the mechanism by which cAMP and cGMP levels control directed axon extension and also establish bi-­‐functional turning. While this remains a very active area of research, and several findings are controversial, a few models of cyclic nucleotide signalling in growth cone turning are beginning to emerge. One important question that has given rise to mixed results is whether guidance cues themselves induce cyclic nucleotide signals. Netrin-­‐1 effects on axons from dorsal spinal cord were shown to require binding to the adenosine A2B receptor and in turn induce cAMP production (Corset, Nguyen-­‐Ba-­‐Charvet et al. 2000). Furthermore, Netrin-­‐1 induced cAMP production is mediated by soluble adenyl cyclase (sAC) (Wu, Zippin et al. 2006). Similarly, Sema1A signalling in Drosophila requires the activity of a receptor type guanalyl cyclase and Sema binding induces cGMP production (Ayoob, Yu et al. 2004). However, other groups have reported results that contradict some of these findings. The A2B receptor is not required for -­‐ 31 -­‐ Introduction Netrin-­‐1 signalling in rat commissural (Stein and Tessier-­‐Lavigne 2001) or Xenopus spinal neurons and furthermore Netrin-­‐1 treatment does not alter cAMP concentrations or PKA activity and does not require sAC (Moore, Correia et al. 2008). An alternative role is supported by the finding that cAMP driven PKA activity increases the expression of DCC at the cell surface in response to Netrin-­‐1 treatment, providing a feedback loop and increasing the sensitivity of commissural neurons to Netrin-­‐1 (Bouchard, Moore et al. 2004). Apart from acting as a direct mediator of guidance cue signals, much of the work surrounding the manner in which cyclic nucleotides can induce a switch in polarity of turning response has focused on the crosstalk between cyclic nucleotides and other intracellular signals such as [Ca2+]I and the Rho GTPases. The formation of [Ca2+]I gradients in the growth cone through CICR, is dependent on the activation of RyRs and down-­‐regulating PKA activity is sufficient to block this [Ca2+]I signal (Hong, Nishiyama et al. 2000). This finding led to the suggestion that cAMP switching could be mediated through altering CICR. Further pharmacological studies in Xenopus spinal neurons revealed that both cAMP and cGMP pathways could regulate [Ca2+]I signals (Nishiyama, Hoshino et al. 2003; Henley, Huang et al. 2004)through a variety mechanisms including direct modulation of L type VDCCs (Nishiyama, Hoshino et al. 2003), cAMP positive regulation (Ooashi, Futatsugi et al. 2005) or cGMP negative regulation (Tojima, Itofusa et al. 2009) of RyRs, and effecting membrane potential through gating ion channels (Nishiyama, von Schimmelmann et al. 2008; Togashi, von Schimmelmann et al. 2008). These results indicate that rather than there being two separate types of guidance cues, one being responsive to cAMP and the other to cGMP, both cyclic nucleotide pathways are involved in turning, and the critical factor for determining the polarity of response is the cAMP/cGMP ratio present (Fig. 4). A high ratio appears to amplify [Ca2+]I signals and lead to attraction and a low -­‐ 32 -­‐ Introduction ratio dampens [Ca2+]I signals and favours repellent turning (Fig. 4) (Nishiyama, Hoshino et al. 2003; Tojima, Hines et al. 2011). Another class of signalling molecules that are well established as critical mediators of axon guidance signals to the underlying cytoskeleton is the Rho family of GTPases (Hall and Lalli 2010). These molecular switches cycle from a GTP-­‐bound “on” state to a GDP-­‐bound “off” state and act on a wide array of effector proteins to coordinate changes to the dynamic arrangements of actin filaments and microtubules that drive growth cone turning. Generally speaking, attractive responses to proteins such as Netrin-­‐1, activate Rac and CDC-­‐42, while down regulating RhoA (Shekarabi, Moore et al. 2005; Moore, Correia et al. 2008) and repellent turning and growth cone collapse require the coordinated activation of RhoA and to some extent Rac1 (Jin and Strittmatter 1997; Kuhn, Brown et al. 1999; Gallo 2006), along with down-­‐
regulation of CDC-­‐42 (Myers, Robles et al. 2012). Thus, the switch from attractive to repellent responses can also be driven by altering the balance between Rac/CDC-­‐42 activity and RhoA activity. Recent findings from correlative image analysis of the leading edge of migrating cells have established that PKA is responsible for generating a pacemaker mechanism that governs the cycles of protrusion and retraction of the motile structures that drive cell migration. This is done by PKA phosphorylation of RhoA and the promotion of its association with the RhoA inhibitor, Rho-­‐GDP dissociation inhibitor (Rho-­‐GDI)(Tkachenko, Sabouri-­‐Ghomi et al. 2011). The ability of PKA to directly control the balance of RhoGTPase activation presents another likely mechanism by which it can alter growth cone turning responses. Outside of roles as upstream regulators of [Ca2+]I and Rho signals, cAMP and PKA have also been proposed to act as regulators of the kinase/phosphatase signals that act downstream of [Ca2+]I signals (Wen, Guirland et al. 2004; Han, -­‐ 33 -­‐ Introduction Han et al. 2007). Active PKA phosphorylates the phosphatase inhibitor, inhibitory protein-­‐1(I-­‐1), (Han, Han et al. 2007) which in turn directly antagonizes the CaN-­‐PP1 signals which low amplitude [Ca2+]I signals activate to mediate repulsion (Fig. 3) (Wen, Guirland et al. 2004). The attractive turning that results from this phosphatase inhibitory effect has been shown to rely on the proper localization of PKA to sub-­‐domains of the growth cone mediated by a class of proteins known as protein kinase A anchoring proteins (AKAPs) (Fig. 4) (Han, Han et al. 2007). AKAPs are adaptor proteins known to play a critical role in many cell-­‐signalling events by controlling the localization of PKA and promoting the phosphorylation of specific associated targets. The AKAP, Nervy, is associated with Plexin receptors in Drosophila (Terman and Kolodkin 2004) and mammals (Fiedler, Schillace et al. 2010) and is critical in mediating the targeted phosphorylation of PlexinA by PKA, antagonizing Plexin A signalling in Drosophila motor neurons and promoting a switch from Sema-­‐mediated repulsion to integrin adhesion (Yang and Terman 2012). The crucial involvement of AKAPs in growth cone signalling as well as the common existence of cAMP signalling microdomains (Zaccolo and Pozzan 2002) suggest that the complexity of cyclic nucleotide signalling in setting the polarity of turning responses, and the contradictory findings, may result in part from the interaction of cylic nucleotides with different targets and effectors, such as PKA and exchange protein activated by cAMP (Epac) (Murray, Tucker et al. 2009) under the control of very precise spatial regulation. -­‐ 34 -­‐ Introduction Figure 4: Cyclic nucleotide signals regulate both attractive and repellent growth cone turning responses (Left) Attractive growth cone turning is mediated by a high cAMP /cGMP ratio, where counter gradients are formed across the growth cone and localized cAMP signals lead to PKA activation, spatially regulated by AKAPs (blue), and phosphorylation of downstream targets, such as I-­‐1. Also, cAMP activates Ryanadine Receptors (RyR, green) to promote release of intracellular Ca2+stores. (Right) Repellent turning occurs with a low cAMP/cGMP ratio, with a more extensive cGMP gradient, which dampens Ca2+ signals by blocking RyRs and Voltage gated calcium channels (VDCC, red) and activating Cyclic nucleotide gated channels (CNGs, blue). More recent work has been focused on mechanisms that can establish this sub-­‐cellular spatial control. Interactions between cAMP and cGMP signals in neurons (Shelly, Lim et al. 2010) and the antagonistic relationship they appear to have in growth cone signalling (Nishiyama, Hoshino et al. 2003) has led to the proposal that cAMP and cGMP gradients may form and cross-­‐
repress each other across the growth cone, to amplify the initial asymmetric binding of a guidance cue, and define a protruding and retracting domain of the growth cone (Fig. 4) (Tojima, Hines et al. 2011). Another study, using optogenetic manipulations of [Ca2+]I signals and cAMP signals found that -­‐ 35 -­‐ Introduction cAMP signals in response to Netrin-­‐1 are transient and precede changes in [Ca2+]I signals. Interestingly, while cAMP transients occur throughout the growth cone, only those confined to the filopodia of the growth cone periphery are able to induce attractive turning (Nicol, Hong et al. 2011). As we continue to refine our understanding of the interplay between the many intracellular regulators involved in growth cone turning, models are emerging allowing us to predict turning outcomes based on the delicate balance between multiple inputs (Forbes, Thompson et al. 2012), highlighting the many ways in which axons are able to control the signalling mechanisms triggered by extracellular cues to address the navigational challenges they face. Spatial and temporal regulation of growth cone signalling play a central role in the directed growth of axons and the development of neural circuits and thus the goal of this thesis project was to further understanding of this process by identifying and characterizing proteins that mediate this regulation. The identification of 14-­‐3-­‐3 proteins as major constituents of the growth cone (Ch.2);(Nozumi, Togano et al. 2009), along with their known roles in cell signalling (see below), made them an important candidate for further investigation. -­‐ 36 -­‐ Introduction 3. 14-­‐3-­‐3 proteins as spatial and temporal regulators of c ell s ignalling Subsequent to their original identification by ion exchange chromatography in a systematic classification of proteins enriched in bovine brain (Moore 1967), 14-­‐3-­‐3 proteins, a name referring to their column fraction and gel migration in the original screen, have been shown to be a highly conserved family of adaptor proteins expressed in all eukaryotic organisms studied to date. They bind primarily to targets phosphorylated at serine or threonine residues and are involved in such fundamental cellular processes as metabolism, cell cycle control, signal transduction, apoptosis, trafficking and cytoskeletal organization (Mackintosh 2004). Their involvement in such a wide array of cellular processes has led to extensive study of these proteins and their numerous interactions with target proteins. Here we review their structure, proposed modes of function and genetic evidence of their role in neuronal development. As well, we examine a number of previously characterized, specific interactions with target proteins that highlight the manner in which 14-­‐3-­‐3 proteins impact the motile response of cells to external cues. A number of these targets are known to be involved in regulating growth cone signalling, and along with the enrichment of 14-­‐3-­‐3 proteins in the growth cone, the manner in which 14-­‐3-­‐3 protein function provides spatial and temporal regulation to key signalling pathways make these proteins excellent candidates for critical regulators of growth cone responses to axon guidance cues. 3.1 1 4-­‐3-­‐3 s tructure a nd f unction In mammals and other vertebrates, the 14-­‐3-­‐3 family consists of seven isoforms, (α/β, δ/ζ, ε, γ, η, τ/θ, and σ) encoded by unique genes (Rosenquist 2000), with the α and δ being identified as the phosphorylated form of β and ζ respectively (Aitken 2006). In other eukaryotic organisms, 14-­‐3-­‐3 -­‐ 37 -­‐ Introduction family members can consist of up to fifteen isoforms, such as in plants, and as few as two, in yeast, Drosophila and C. Elegans (Wang 1996). Extensive X-­‐ray diffraction crystallography data has now been compiled on all seven human isoforms in both their native and ligand-­‐bound states (Yang, Lee et al. 2006). Structural studies have shown that 14-­‐3-­‐3s readily form homo-­‐ or heterodimers, with each monomeric unit consisting of a bundle of nine α-­‐
helices (Liu, Bienkowska et al. 1995; Xiao, Smerdon et al. 1995). These helices are organized to form a highly conserved concave, amphipathic ligand binding groove, as well as a surface region critical for dimerization. Different features in this region are responsible for the different dimerization preferences of different isoforms, with ε preferentially forming only heterodimers, σ forming only homodimers and the other isoforms showing equal homo-­‐ and heterodimerization (Yang, Lee et al. 2006). Analysis of the crystal structure of the ligand bound forms of 14-­‐3-­‐3 proteins has revealed that binding of 14-­‐3-­‐3s to their targets is conferred by both a primary and secondary interaction. The primary interaction consists of the binding of the target phosphopeptide in a tightly regulated orientation to three conserved positively charged residues within the amphipathic groove (Yang, Lee et al. 2006). This binding occurs with a wide array of targets with the emerging consensus target sequence of RXXpS (Ballif, Cao et al. 2006). It should be pointed out that not all 14-­‐3-­‐3 binding is dependant on target phosphorylation. In vitro studies involving the bacterial adenosine diphosphate (ADP)-­‐ ribosyltransferase, exoenzyme S have shown that 14-­‐3-­‐
3s are able to bind to non-­‐phosphorylated targets in the amphipathic groove (Zhang, Wang et al. 1997). The initial high affinity interaction is followed by secondary interactions with more variable surface regions which could convey a level of target specificity (Yang, Lee et al. 2006). Indeed while a number of 14-­‐3-­‐3 target interactions have been shown to occur with equal affinity amongst several isoforms -­‐ 38 -­‐ Introduction (Rittinger, Budman et al. 1999), there is increasing evidence, both biochemically and in vivo, that isoform specificity is also functionally important, allowing for the integration of different targets into signalling complexes (Aitken, Baxter et al. 2002). Analysis of the solved crystal structures of all seven human isoforms bound to substrates reveals that structural differences between isoforms outside of the binding pocket may underlie the binding and dimerization preferences of the different isoforms, as well as differential post-­‐translational modifications (Gardino, Smerdon et al. 2006). The binding of 14-­‐3-­‐3 dimers to phosphorylated substrates can have a wide range of functional effects on the binding targets. The ability of 14-­‐3-­‐3 proteins to form dimers with two binding grooves as well as their ability to interact with target proteins at multiple sites has led to the “molecular anvil” hypothesis (Yaffe 2002). This hypothesis essentially states that initial binding of 14-­‐3-­‐3s is followed either by bipartite binding of a 14-­‐3-­‐3 dimer to its target through the ligand binding groove or secondary interactions at multiple sites along the target. This multiple binding results in 14-­‐3-­‐3 altering the conformation of its target and thereby regulating its enzymatic activity through masking or revealing functional motifs which are responsible for localization, stability or activity (Fig. 5). This model of 14-­‐3-­‐3 activity is unlikely to be the only valid one, as a number of other functional characterizations of 14-­‐3-­‐3 binding interactions have shown that 14-­‐3-­‐3s can act as adaptors, bringing together enzymes with their substrates or holoreceptor complexes such as the Ron receptor tyrosine kinase and an integrin dimer (Santoro, Gaudino et al. 2003), as well as sequestering molecules and preventing their association with binding partners such as the phosphatase Slingshot (SSH1L) and its target cofilin (Nagata-­‐Ohashi, Ohta et al. 2004) (Fig. 5). -­‐ 39 -­‐ Introduction A.
P
SSH
Ron
F-­Actin
B.
C.
P
P
filin
Co
LIMK
Integrin
D.
P
P
ilin
f
o
C
nnnnnn
Cofilin
Figure 5: 14-­‐3-­‐3 binding regulates target proteins in different ways A. Binding of a 14-­‐3-­‐3 dimer to the phosphatase Slingshot can regulate its activity by controlling its sub-­‐cellular localization and sequstering it from filamentous actin which acts as a co-­‐activator. B. 14-­‐3-­‐3s can also regulate the enzymatic activity of targets, such as the kinase LIMK, by facilitating the interaction with its substrate, cofilin, through bipartite binding. C. 14-­‐3-­‐3s can act as critical adaptor proteins in holoreceptors, such as Ron RTK and alpha6beta4 integrin, regulating their sub-­‐cellular location. D. The binding of a 14-­‐3-­‐3 dimer to multiple sites on the same protein can induce a conformational change or serve to stabilize a particular conformation induced by post-­‐
translational modifications such as the phosphorylation of cofilin. While 14-­‐3-­‐3 proteins can modulate enzymatic activity, they in turn can also be the targets of regulation. 14-­‐3-­‐3 activity can be regulated through the direct phosphorylation of 14-­‐3-­‐3 isoforms (Aitken 2006). While a number of -­‐ 40 -­‐ Introduction different kinases have been shown to target different 14-­‐3-­‐3 isoforms at different sites, broadly 14-­‐3-­‐3 phosphorylation can regulate its activity in two ways. The first is the disruption of 14-­‐3-­‐3 binding to a target by phosphorylation of residues near the ligand-­‐binding groove. This is exemplified by the finding that c-­‐Jun N-­‐terminal kinase (JNK) mediated phosphorylation of 14-­‐3-­‐3 ζ at serine (Ser) 184 results in the dissociation of the proapoptotic molecule Bad in COS cells and that this mediates stress induced apoptosis (Sunayama, Tsuruta et al. 2005). The second mode of regulation is by phosphorylating residues that are critical for 14-­‐3-­‐3 dimer formation. For example, the phosphorylation of 14-­‐3-­‐3ζ at Ser58 by PKA was shown to down-­‐regulate its ability to bind the p53 tumour suppressor by interfering with binding necessary for dimerization (Gu, Jin et al. 2006). This finding also indicates that in some contexts 14-­‐3-­‐3s are downstream of PKA, the effector of cAMP, previously discussed as a key regulator of growth cone turning responses. 3.2 14-­‐3-­‐3s r ole i n n ervous s ystem d evelopment Genetic analysis of 14-­‐3-­‐3s has been of limited use, presumably because of a large degree of functional redundancy within the protein family. However, a number of studies do implicate 14-­‐3-­‐3s in having some important isoform specific roles in early embryonic and neuronal development. One study (Lau 2006) aimed to take advantage of the ease of protein knockdown in the Xenopus antisense morpholino system. Having characterized expression differences between six different isoforms during early embryonic development, the group underwent a systematic analysis by injecting morpholinos against each of the isoforms at the four-­‐cell stage. Despite robust expression, knockdown of the β isoform showed no phenotypic effects, nor did ζ isoform knockdown. Interestingly, knockdown of the τ isoform, and to a lesser extent ε showed severe gastrulation defects and γ knockdown resulted in a failure in eye development suggesting a nervous -­‐ 41 -­‐ Introduction system specific role. Work in Drosophila, which lacks a γ isoform, has also implicated 14-­‐3-­‐3s in nervous system development. With only two 14-­‐3-­‐3 genes, leonardo (ζ ortholouge) and D14-­‐3-­‐3ε, genetic analysis in Drosophila is more straight forward. While homozygous leonardo null alleles are embryonic lethal, hypomorhic alleles have been shown to interfere with presynaptic function of mushroom body neurons and result in olfactory learning defects (Broadie, Rushton et al. 1997). Loss of function mutations in D14-­‐3-­‐3ε can act to suppress a rough-­‐eye phenotype generated by ectopic RAS-­‐1 signaling (Chang and Rubin 1997), indicating a positive regulatory role on Ras GTPase activty. Importantly, a very recent study has directly implicated D14-­‐3-­‐3ε in axon guidance as loss of D14-­‐3-­‐3ε results in highly penetrant misrouting and defasiculation phenotypes in motor neurons of the developing larvae (Yang and Terman 2012). The findings of this key study will be discussed later in this section. Mammalian studies have failed to confirm isoform specific roles in eye development, with the γ knockout mouse showing no obvious anatomical or behavioural defects (Steinacker, Schwarz et al. 2005), however closer characterization of axonal tracts are often required to detect guidance defects. More interestingly, mice that are homozygous null for 14-­‐3-­‐3 ε show severe defects in cortical layer formation, closely resembling the pathology seen in the lissencephaly, or “smooth-­‐brain” disease state, and the phenotype of knockouts of another lissencephaly-­‐linked protein, Lissencephaly protein 1 (LIS1). As well, defects in hippocampal formation were reported (Toyo-­‐oka, Shionoya et al. 2003). Further examination shows a dose dependant defect in neuronal migration, also seen in double heterozygotes of LIS1 and 14-­‐3-­‐3 ε. The link between 14-­‐3-­‐3 ε and cortical development is also supported in genetic studies of humans suffering from Miller-­‐Dieker Syndrome (MDS), a severe form of lissencephaly, which critically depends on deletions in chromosome 17p13.3, a region encoding 14-­‐3-­‐3 ε (Yingling 2003). Further -­‐ 42 -­‐ Introduction characterization of the role of 14-­‐3-­‐3 ε in MDS revealed direct binding to phosphorylated nuclear distribution protein nudE-­‐like 1 (NUDEL) and the formation of a protein complex consisting of NUDEL/LIS1/14-­‐3-­‐3ε whose proper localization to the growth cone of developing axons is critically dependant on 14-­‐3-­‐3ε expression (Toyo-­‐oka, Shionoya et al. 2003). The co-­‐
localization of the complex with end-­‐binding protein 1 at the plus end of microtubules is an indication that this complex may have a role in regulating the cytoskeletal dynamics of the growth cone. This is further supported by a study characterizing the role of the proteins, deleted in schizophrenia 1 (DISC1) and Kinesin-­‐1 in delivering the complex to the growth cones of hippocampal neurons (Taya, Shinoda et al. 2007). Preventing the accumulation of the complex at the distal tip by knocking down DISC1 significantly attenuated axon outgrowth, as did the knockdown of complex elements LIS1 and NUDEL. A further study looked at the effect of 14-­‐3-­‐
3ε knockdown on hippocampal pyramidal cell migration and found 14-­‐3-­‐
3ε is required for CA1 neurons to reach the CA1 layer (Hippenmeyer, Youn et al. 2010), suggesting a role for 14-­‐3-­‐3 regulation of the localization of new neurons. The importance of these findings was emphasized by results indicating that the gene encoding 14-­‐3-­‐3ε, is a susceptibility locus for schizophrenia (Ikeda, Hikita et al. 2008). 3.3 14-­‐3-­‐3 p roteins r egulate s ignalling r esponses t o e xternal cues The diversity of 14-­‐3-­‐3 homo-­‐ and heterodimers and the myriad ways in which their binding can be regulated allows this protein family to play important roles in many cellular events. Efforts to characterize 14-­‐3-­‐3 signalling at a more systemic level, gave rise to a number of attempts to utilize a large-­‐scale proteomics approach to identify the breadth of 14-­‐3-­‐3 interacting proteins (Jin, Smith et al. 2004; Meek, Lane et al. 2004; Pozuelo Rubio, Geraghty et al. 2004; Ballif, Cao et al. 2006). These studies used -­‐ 43 -­‐ Introduction different experimental paradigms including affinity chromatography of asynchronous and mitotic HeLa cell extracts (Meek, Lane et al. 2004), and embryonic murine brain extract (Ballif, Cao et al. 2006); in addition to immunoaffinity pulldowns from purified epitope tagged 14-­‐3-­‐3 isoforms from stable human embryonic kidney (HEK) 293 cell lines (Jin, Smith et al. 2004). These studies all used different isoforms as bait, but they each uncovered hundreds of binding partners with significant overlap between the data sets from each study. When compiled, proteomic studies have shown over 300 different binding partners with 14-­‐3-­‐3 proteins which suggest 14-­‐3-­‐
3s function in metabolism, cell cycle and transcriptional control, nuclear localization, membrane trafficking and other sub-­‐cellular localization, ubiquitination and proteosome regulation, protein synthesis and folding, apoptosis, cell signalling and cytoskeletal organization (Pozuelo Rubio, Geraghty et al. 2004). Given the vast number of binding targets, and the involvement in a wide array of cellular processes, examples of 14-­‐3-­‐3 function are too numerous to describe exhaustively here. Instead, this discussion will focus on two examples of signalling pathways, in which 14-­‐3-­‐
3 regulation plays a critical role in determining the motile response of cells to an external cue. Interestingly, many of the 14-­‐3-­‐3 targets, upstream regulators and downstream effectors described are also known to play key roles in growth cone signalling. One common effector protein which regulates the dynamic response of the cytoskeleton in response to guidance cues is actin depolymerizing factor (ADF)/cofilin (Aizawa, Wakatsuki et al. 2001; Gehler, Shaw et al. 2004; Hsieh, Ferraro et al. 2006; Wen, Han et al. 2007; Takemura, Mishima et al. 2009) The active form of cofilin is responsible for depolymerizing actin filaments as well as creating new barbed ends through a severing function. The activity o f cofilin in motile cells can be very precisely controlled through regulation of phosphorylation by a number of kinases, including LIM kinase and testin -­‐ 44 -­‐ Introduction kinase (TESK), which target serine residues to down-­‐regulate cofilin, and conversely, the phosphatase Slingshot-­‐1 (SSH1L) which activates cofilin by dephosphorylating the serine residues (Nishita, Tomizawa et al. 2005). The balance between the kinase/phosphatase regulatory axis has been proposed to underlie the bifunctional turning response to at least one guidance cue (Wen, Han et al. 2007). A yeast two hybrid screen for proteins interacting with 14-­‐3-­‐3ζ identified both LIM kinase (LIMK) and cofilin as binding targets of 14-­‐3-­‐3 proteins. (Birkenfeld, Betz et al. 2003). Cofilin, when phosphorylated, is bound and stabilized by 14-­‐3-­‐3ζ, protecting it from phosphatase activity and leading to a potentiation of a LIMK induced actin aggregation phenotype by 14-­‐3-­‐3ζ overexpression in baby hamster kidney (BHK-­‐21) cells (Gohla and Bokoch 2002). As 14-­‐3-­‐3 binding to LIMK was not seen to affect its kinase activity(Birkenfeld, Betz et al. 2003), this suggests that 14-­‐3-­‐3s may act to bring together LIMK and cofilin. These findings, in conjunction with the fact that SSH1L has a 14-­‐3-­‐3 consensus target site, led to a closer examination of the functional nature of these interactions. One group has shown that the regulation of cofilin by LIMK and SSH1L takes place within a multimeric protein complex that involves direct interactions between LIMK, SSH1L, F-­‐actin and 14-­‐3-­‐
3ζ (Soosairajah, Maiti et al. 2005). They were able to demonstrate in vitro that 14-­‐3-­‐3ζ binding to phosphorylated SSH1L occurs and serves to block its association with F-­‐actin which is necessary for its dephosphorylating and activating cofilin, suggesting yet another manner in which 14-­‐3-­‐3s could serve to increase levels of phosphorylated cofilin and regulate actin dynamics. The regulatory activity of this model is supported by another study(Nagata-­‐
Ohashi, Ohta et al. 2004) in which Neuregulin-­‐1β, a known trigger of lamellipodial formation and cell migration, was used to stimulate MCF-­‐7 -­‐ 45 -­‐ Introduction breast carcinoma cells. This was shown to result in the dephosphorylation of SSH1L at a key 14-­‐3-­‐3 binding site and its translocation to actin fibers and subsequent dephosphorylation of cofilin. This antagonizes the function of LIMK in MCF-­‐7 cells, which down-­‐regulates actin turnover in lamellipodia (Ohashi, Fujiwara et al. 2011), thus promoting motility. Overexpression of 14-­‐3-­‐3γ was shown to block the Neuregulin induced translocation and accumulation of SSH to the lammelipodia. Further work has established that the 14-­‐3-­‐3 binding site of SSH is phosphorylated by protein kinase D (PKD) and that PKD activity acting through 14-­‐3-­‐3, can effectively block the directed cell migration of these cancer cells (Eiseler, Doppler et al. 2009; Peterburs, Heering et al. 2009). In another cellular context, the promotion of cell motility can be achieved through angiotensin II induced production of reactive oxgen species (ROS) which target 14-­‐3-­‐3 for oxidation, and in turn release SSH (Kim, Huang et al. 2009). Thus 14-­‐3-­‐3 regulation of cofilin, through altering the balance of SSH/LIMK activity allows for controlled movements of various cell types. The direct relevance of 14-­‐3-­‐3 regulation of cofilin to growth cone signalling was demonstrated in a study which was looking at the effects of BDNF treatment on the filopodial dynamics of RGC growth cones (Gehler, Shaw et al. 2004). BDNF treatment induced an increase in the length and number of filopodia present on growth cones, which was mediated by a drop in phosphorylated cofilin. The pre-­‐loading of growth cones with recombinant glutathione-­‐S-­‐transferase (GST)-­‐14-­‐3-­‐3ζ, using the Chariot reagent, was able to block both the BDNF induced change in filopodia as well as the reduction in phosphorylated cofilin. While this would indicate that 14-­‐3-­‐3ζ acts a negative regulator of actin dynamics in RGCs, another recent study looking at the involvement 14-­‐3-­‐3s in RGC extension in vivo, found that interfering with 14-­‐3-­‐3 function or knockdown 14-­‐3-­‐3 protein levels corresponded with an increase in active cofilin, while diminishing the rate of axon extension and -­‐ 46 -­‐ Introduction sensitizing the growth cone to the collapse response to the repellent Slit (Yoon, Zivraj et al. 2011). Thus 14-­‐3-­‐3 regulation of cofilin both limits filipodial responses to BDNF and promotes proper axon extension in RGCs. In addition to directing cell movement through diffusible extracellular proteins, the regulation of adhesive properties through integrins and cell adhesion molecules (CAMs) can have important impacts on the motility of cells, and are known regulators of growth cone activity. 14-­‐3-­‐3s have also been found to play important roles in this regulatory mechanism. The growth factor, macrophage-­‐stimulating protein (MSP) acts through a receptor tyrosine kinase, Ron to induce activation of a wound healing pathway that alters the adhesive properties of keratinocytes and promotes migration. One study found that this is achieved in part by phosphorylation of both Ron and alpha6/beta4 integrin receptors and binding of 14-­‐3-­‐3, which promotes the formation of a receptor complex between the two (Santoro, Gaudino et al. 2003) Interestingly, 14-­‐3-­‐3 binding also induces the relocation of the complex from adhesive structures known as hemidesmosomes to the newly forming lammelipodia, where 14-­‐3-­‐3 continues to play a key signalling role, in regulating SSH and cofilin activity (Kligys, Claiborne et al. 2007; Kligys, Yao et al. 2009). Thus 14-­‐3-­‐3 regulation provides spatial regulation at the receptor level as well. 14-­‐3-­‐3 binding also plays an important role in regulating the interactions between integrins and their effector proteins. Phosphorylation of alpha4 integrins promotes cell motility by inducing 14-­‐3-­‐3 binding and promoting the localized interaction with paxillin, and the subsequent focal activation of the Rho family GTPase, Cdc42 (Deakin, Bass et al. 2009). Also, 14-­‐3-­‐3 binding to Beta1 integrin receptors at the leading edge of migrating cells has been shown to promote cell spreading and motility through the promotion of interactions with Tiam1. This guanine exchange factor (GEF) in turn locally -­‐ 47 -­‐ Introduction activates the Rho family GTPase, Rac1, promoting lamellipodial formation (Kobayashi, Ogura et al. 2011; O'Toole, Bialkowska et al. 2011). By regulating the precise spatial activation of these protein complexes, 14-­‐3-­‐3s play a central role in the intricate mechanisms involved in motile cell responses. Cell adhesion molecules, such as L1 and neural cell adhesion molecule (NCAM) play important roles regulating neurite outgrowth through homophilic interactions. Recent work has demonstrated that both L1 and NCAM mediated outgrowth of hippocampal neurons in culture are regulated by 14-­‐3-­‐3 binding interactions (Ramser, Buck et al. 2010; Ramser, Wolters et al. 2010). NCAM associates with a family of proteins, known as spectrins that coordinate changes in both the membrane and cytoskeleton. Phosphorylation of αII spectrin induces binding of 14-­‐3-­‐3β, in complex with NCAM. Using a version of αII spectrin in which the phosphorylated serine is mutated, thus blocking 14-­‐3-­‐3 binding, enhances NCAM mediated outgrowth (Ramser, Buck et al. 2010). Similarly, blocking interactions between 14-­‐3-­‐3 and L1 also promotes neurite outgrowth (Ramser, Wolters et al. 2010), suggesting that 14-­‐3-­‐3 acts as a negative regulator of cell adhesion molecule induced outgrowth. The importance of 14-­‐3-­‐3 regulation of receptor activation in providing spatial specificity to growth cone signals was highlighted by a very recent study, previously mentioned, in which 14-­‐3-­‐3ε loss of function mutants in Drosophila showed profound defects in motor neuron axon guidance (Yang and Terman 2012). These defects were very similar to mutants in which Sema1A/Plexin A signalling had been enhanced, and genetic interaction experiments suggested that 14-­‐3-­‐3ε plays a negative regulatory role in the Sema1A/Plexin A pathway. Subsequently, 14-­‐3-­‐3ε was shown to interact with Plexin A upon phosphorylation of Plexin A by PKA. This interaction silenced the repellent effect of Sema1A by blocking Plexin A interactions with -­‐ 48 -­‐ Introduction Ras GTPase and interfering with the GTPase activating protein (GAP) activity of Plexin A. Thus restoring Ras GTPase signalling and allowing for integrin promoted adhesion and outgrowth. The diversity of 14-­‐3-­‐3 regulatory interactions, their abundance and importance in nervous system development, and involvement in controlling the motile responses of cells, make them excellent candidates for study, as we try to extend our understanding of the spatial and temporal regulation of growth cone turning responses. -­‐ 49 -­‐ Introduction 4. Thesis R ationale a nd O bjectives Achieving accurate axon guidance through integrating its response to extracellular cues requires the growth cone to modulate its response to signals in both space and time. The mechanisms by which the growth cone is able to spatially and temporally regulate its underlying cytoskeletal response to guidance cues are only beginning to become understood. Given their role in regulating spatial and temporal activity of multiple cell signalling events, and their abundance in the nervous system, 14-­‐3-­‐3s are excellent candidates to play a similar role in the growth cone. This thesis establishes the critical role of 14-­‐3-­‐3 proteins in the spatial and temporal regulation of growth cone response to axon guidance cues. In chapter 2, proteomic analysis of isolated growth cones identifies the 14-­‐3-­‐3 adaptor proteins as major constituents of the growth cone. We describe how loss of 14-­‐3-­‐3 function converts growth cone turning responses to gradients of nerve growth factor (NGF) and MAG from repellent to attractive. We implicate 14-­‐3-­‐3s in the regulation of PKA activity in growth cones, finding that 14-­‐3-­‐3s interact with PKA and that disrupting this interaction results in the release of active catalytic subunits of PKA. In Chapter 3, we describe the role of 14-­‐3-­‐3 proteins and PKA in establishing a cell intrinsic, time dependent switch in the turning response of commissural interneurons to gradients of Shh, from an attractive to a repellent response. We show 14-­‐3-­‐3 function is critical in the proper formation of the axonal projections of dorsal commissural neurons in the developing spinal cord. In Chapter 4, we identify a role for 14-­‐3-­‐3 proteins in mediating the signalling of another guidance cue used by commissural neurons, Netrin-­‐1. -­‐ 50 -­‐ Results -­‐ 51 -­‐ Results 1. 14-­‐3-­‐3 proteins regulate protein kinase A activity to modulate g rowth c one t urning r esponses Christopher B. Kent1, Tadayuki Shimada1, Gino B. Ferraro1, Brigitte Ritter1, Patricia T. Yam2, Peter S. McPherson1, Frédéric Charron2, Timothy E. Kennedy1, Alyson E. Fournier1 Published in The Journal of Neuroscience, 2010, 30(42): 14059-14067
1
Department of Neurology and Neurosurgery, Montreal Neurological Institute,
3801 Rue University, Montreal, Quebec, H3A 2B4
2
Molecular Biology of Neural Development, Institut de Recherches Cliniques de
Montréal (IRCM), Montreal, QC, Canada
Corresponding Author: Alyson Fournier PhD -­‐ 53 -­‐ Results 1.1 PREFACE The extension of axons to their appropriate targets during development or regeneration relies on the growth cone, a specialized structure that integrates extracellular signals to guide axon growth. How growth cone signalling is regulated spatially and temporally is still poorly understood. The aim of this project was to use a proteomic analysis of growth cone constituent proteins to identify candidate molecules that could play essential roles in regulating the signalling pathways that allow growth cones to appropriately respond to guidance cues. We identified 14-­‐3-­‐3 proteins as candidates and characterized their role in regulating PKA signalling in the growth cone. -­‐ 54 -­‐ Results 1.2 ABSTRACT Growth cones regulate the speed and direction of neuronal outgrowth during development and regeneration. How the growth cone spatially and temporally regulates signals from guidance cues is poorly understood. Through a proteomic analysis of purified growth cones we identified isoforms of the 14-­‐3-­‐3 family of adaptor proteins as major constituents of the growth cone. 14-­‐3-­‐3 disruption via the R18 antagonist or knockdown of individual 14-­‐3-­‐3 isoforms switches nerve growth factor-­‐ and myelin-­‐
associated glycoprotein-­‐dependent repulsion to attraction in E13 chick and P5 rat dorsal root ganglion (DRG) neurons. These effects are reminiscent of switching responses observed in response to elevated cAMP. Intriguingly, R18-­‐dependent switching is blocked by inhibitors of PKA suggesting that 14-­‐
3-­‐3 proteins regulate PKA. Consistently, 14-­‐3-­‐3 proteins interact with PKA and R18 activates PKA by dissociating its regulatory and catalytic subunits. Thus, 14-­‐3-­‐3 heterodimers regulate the PKA holoenzyme and this activity plays a critical role in modulating neuronal responses to repellent cues. -­‐ 55 -­‐ Results 1.3 INTRODUCTION Axons accurately project to specific targets during development by interpreting a complex environment containing attractive and repulsive cues (Huber, Kolodkin et al. 2003). Binding of guidance cues to receptors on the surface of the growth cone stimulates intracellular signalling cascades that converge on cytoskeletal elements to control growth cone morphology (Huber, Kolodkin et al. 2003). How the activation state of cytoskeletal regulatory proteins is spatially and temporally regulated is an important question which has only been partially elucidated. Local translation, insertion and removal of receptors at the plasma membrane as well as receptor silencing all affect growth cone responses (Stein and Tessier-­‐
Lavigne 2001; Bouchard, Moore et al. 2004; Kim, Lee et al. 2005; Lin and Holt 2007). Further, regulating intracellular concentrations of cyclic nucleotides affects whether neuronal processes are attracted or repelled by guidance cues in vitro and in vivo (Ming, Song et al. 1997; Song, Ming et al. 1998; Song and Poo 1999; Polleux, Morrow et al. 2000; Shewan, Dwivedy et al. 2002; Nishiyama, Hoshino et al. 2003). Antagonizing cAMP in embryonic neurons can convert growth cone attraction to repulsion while elevating cAMP in post-­‐natal neurons can convert repulsion to attraction and can promote CNS regeneration following injury (Neumann, Bradke et al. 2002; Qiu, Cai et al. 2002). Both type II cAMP dependent protein kinase (PKA) and the exchange protein activated by cAMP (Epac) have been implicated as cAMP effectors affecting neurite outgrowth and growth cone turning (Cai, Qiu et al. 2001; Han, Han et al. 2007; Murray and Shewan 2008) but the mechanisms regulating these effectors are unclear. In a proteomic analysis of a retinal ganglion cell (RGC) growth cone preparation we identified several members of the 14-­‐3-­‐3 protein family (Tyrosine 3-­‐monooxygenase/tryptophan 5-­‐monooxygenase activation -­‐ 56 -­‐ Results proteins); excellent candidates to regulate the spatial and temporal activity of cAMP and its effectors. 14-­‐3-­‐3 proteins are adapter proteins that interact with specific phospho-­‐serine and phospho-­‐threonine motifs within a number of binding proteins and consequently regulate diverse cellular processes including cell survival, metabolism, proliferation and protein trafficking (Garbe and Bashaw 2004; Jin, Smith et al. 2004; van Heusden 2005). 14-­‐3-­‐3 proteins function as homodimers and heterodimers to control the spatial and temporal activity of substrate proteins by regulating subcellular localization (Muslin and Xing 2000; Nagata-­‐Ohashi, Ohta et al. 2004), binding partner proximity (Jones, Ley et al. 1995; Berruti 2000; Van Der Hoeven, Van Der Wal et al. 2000) and by inducing conformational changes that can affect interactions between binding proteins and their substrates (Roy, McPherson et al. 1998). In mammals, there are seven 14-­‐3-­‐3 isoforms designated beta or alpha (β,α), epsilon (ε), eta (η), gamma (γ), tau or theta (τ,θ), zeta or delta (ζ,δ) and sigma/stratifin (σ) (Bridges and Moorhead 2005). Multiple binding partners have been described for each 14-­‐3-­‐3 isoform, however the functional relevance of these interactions has only been elucidated for a small subset of proteins. The ability of 14-­‐3-­‐3 proteins to regulate the activity and localization of Serine/Threonine (Ser/Thr) phosphorylated substrates and our identification of 14-­‐3-­‐3 proteins as major constituents of the growth cone led us to investigate the role of 14-­‐3-­‐3 proteins in regulating neuronal growth cone responses to extracellular cues. We find that 14-­‐3-­‐3 proteins regulate PKA and loss of 14-­‐3-­‐3 function converts nerve growth factor-­‐ (NGF) and myelin-­‐associated glycoprotein-­‐ (MAG) dependent repulsion to attraction in E13 chick and P5 rat DRG neurons. Thus, we have identified a novel molecular switch that may be harnessed to attenuate the inhibitory CNS environment following injury. -­‐ 57 -­‐ Results 1.4 MATERIALS A ND M ETHODS 1.4.1 Reagents To construct expression vectors for 14-­‐3-­‐3 isoforms, cDNA was amplified by polymerase chain reaction (PCR) from vectors provided by Dr. Tony Pawson, (University of Toronto; pcDNA3.1FLAG 14-­‐3-­‐3γ, pSport6 14-­‐3-­‐3ε, pOTB7 14-­‐
3-­‐3ζ, pDNR-­‐Lib 14-­‐3-­‐3β) or Open Biosource Systems (pOTB7 14-­‐3-­‐3τ/θ) and ligated into pcDNA3.1A 2xmyc. R18 and WLRL expression vectors were constructed by synthesizing oligos encoding the peptide sequence (R18: ccccactgtgtcccccgagatctttcgtggttagatttagaagcaaatatgtgtttaccc WLRL: ccccactgtgtcccccgagatctttcgtggttaaggttagaagcaaatatgtgtttaccc), ligating into pCS2+ enhanced green fluorescent protein (EGFP) and then subcloning R18 EGFP and WLRL EGFP into pHSVPrPUC. To construct expression vectors for PKA RIIα and PKA RIIβ, cDNA was amplified by PCR from pET-­‐RIIα (mouse) and pET-­‐RIIβ (rat) provided by Dr. Susan Taylor (University of California, San Diego) and PCR products were ligated into pcDNA 3.1 V5 His. Histone deacetylase 2 (HDAC2) antibody was purchased from Zymed Laboratories Inc. (San Francisco, CA) and used at a 1:2500 dilution for Western blotting. Neurofilament-­‐200 antibody was purchased from Sigma-­‐
Aldrich (St.Louis, MO) and used at a dilution of 1:2000 for Western blotting. Anti-­‐pan-­‐ β-­‐, ζ-­‐, ε-­‐, γ-­‐, τ-­‐ and η-­‐14-­‐3-­‐3 primary antibodies were purchased from Santa Cruz Biotechnologies (Santa Cruz, CA) and used at a 1:1000 (ζ, ε, γ, τ) or 1:3000 (β) dilution for Western blotting and 1:50 for immunofluorescence. Anti-­‐PKA catalytic subunit and regulatory subunit II Beta primary antibodies were purchased from BD Biosciences (San Jose, CA) and used at 1:1000 for Western blotting. Anti-­‐phospho-­‐PKA catalytic (Thr197) antibody was purchased from Cell Signaling (Beverly, MA) and used at 1:1000 for Western blotting . Anti-­‐phospho I-­‐1/DARPP-­‐32 (Thr34) antibody was purchased from Novus Biologicals (Littleton, CO) and used at 1:100 for immunofluorescence. Sp-­‐cAMPS, Rp-­‐cAMPS and KT-­‐5720 were -­‐ 58 -­‐ Results purchased from Sigma-­‐Aldrich. Myrisolated protein kinase A inhibitory peptide (PKI) was purchased from Invitrogen (Carlsbad,CA). 1.4.2 Dorsal root ganglion c ultures DRGs were dissected from E8 chicks, E13 chicks, or P5 rats in L-­‐15 media and plated on poly-­‐L-­‐lysine and laminin coated substrates as explants. For dissociated cultures, explants were treated with 0.25% Trypsin-­‐EDTA and triturated in media to dissociate cells. DRGs were grown in F-­‐12 media or Neurobasal media with B-­‐27 (Invitrogen) for chick and rat DRGs, respectively. Media was supplemented with 1% penicillin/streptomycin, and 10ng/ml NGF (Calbiochem, La Jolla, CA). 1.4.3 Proteomics E6 Retinae were dissected from E6 chicks, cut into strips, placed on nitrocellulose membranes and grown on poly-­‐L-­‐lysine and laminin in a 1:1 Dulbecco modified Eagle's minimal essential medium (DMEM)/F-­‐12 (Invitrogen) mixture with 10% fetal bovine serum (FBS), 1% penicillin/streptomycin, 30ng/ml of BDNF and 1% N-­‐2 nutritional supplement (Invitrogen). Axons with growth cones were separated from cell bodies using a sharp borosilicate glass capillary. Cell bodies were removed and remaining neurites and growth cones were incubated at 37°C for 30 minutes for recovery. Growth cones were collected in radioimmunoprecipitation assay (RIPA) buffer (20mM HEPES pH 7.5, 150mM NaCl, 0.5% Sodium Deoxycholate, 0.1% SDS, 1% Triton X-­‐100) and lysates separated by sodium dodecyl sulfate-­‐polyacrylamide gel electrophoresis (SDS-­‐PAGE). Bands were processed by trypsin digestion and the resulting peptide mixtures were analyzed by nanoscale liquid chromatography quadrupole time-­‐of-­‐flight MS/ MS. MS profiles were analyzed using MASCOT software. -­‐ 59 -­‐ Results 1.4.4 Immunofluorescence For analysis of 14-­‐3-­‐3 isoform expression, primary neuronal cultures were fixed with modified Davidson’s Fix (30% Ethanol, 10% Acetic Acid, 20% Formaldehyde, 20% sucrose). Antigen blockade was performed with blocking peptide (Santa Cruz), by incubating the antibody with a 5:1 molar excess of the blocking peptide for 30 mins at 37°C immediately prior to staining. To validate developmental differences in expression, E6 retinal explants and E8 DRG explants were co-­‐cultured in the same well of a 4-­‐well chamber slide, fixed as described above and immunostained. Staining intensities were directly compared during image acquisition. For phosphorylated I-­‐1 levels, P5 rat DRG explants were fixed with 4% paraformaldehyde (PFA) with 20% sucrose.
1.4.5 14-­‐3-­‐3 knockdown and virus preparation shRNA with a microRNA stem (shRNAmir) was designed against target sequences (see below) for each 14-­‐3-­‐3 isoform using Invitrogen’s RNAi Designer. Oligonucleotides encoding the sequences were ligated into a pcDNA6.2-­‐GW/EmGFP-­‐miR vector and lentivirus was produced as described previous (Thomas, Ritter et al. 2009). Sequences of oligos used for knockdown of each 14-­‐3-­‐3 isoform: 14-­‐3-­‐3β; TGCTGATAACGCTCAGCTTGCTCAGCGTTTTGGCCACTGACTGACGCTGAGCACTGAGCGTTAT 14-­‐3-­‐3ζ; TGCTGATAAGCAACAGAGAGAAGGTTGTTTTGGCCACTGACTGACAACCTTCTCTGTTGCTTAT 14-­‐3-­‐3γ; TGCTGATCAGAGTGGAGTCCTTGTAGGTTTTGGCCACTGACTGACCTACAAGGTCCACTCTGAT 14-­‐3-­‐3τ; TGCTGTGAGGGTGCTGTCTTTGTAGGGTTTTGGCCACTGACTGACCCTACAAACAGCACCCTCA 14-­‐3-­‐3ε; TGCTGAAACCTTGGACTCGCCAGTGTGTTTTGGCCACTGACTGACACACTGGCGTCCAAGGTTT Control; TGCTGTTTGCAAAGAGCACTTTCTTTTCCACTACTGACGAAAGCCTTCTCTTGCAAA Dissociated P5 rat DRGs were transduced with recombinant lentivirus particles at the time of plating for 12 hours and then cultured for 5 days in vitro (DIV). Cells were detached with 5mM EDTA and re-­‐plated for use in -­‐ 60 -­‐ Results Dunn Chamber turning assays or lysed for Western blotting. Knockdown of targeted isoforms and specificity was examined by Western blot. Densitometry was performed on three independent experiments. Band density of each isoform was normalized to glyceraldehyde 3-­‐phosphate dehydrogenase (GAPDH) levels on the same blot and expressed as a percentage of expression in control lentivirus-­‐transduced cells, or the mean expression of the isoform in lysates transduced for knockdown of the off-­‐
target isoforms. Recombinant Herpes Simplex Virus was produced by transfecting pHSVPrPUC plasmids into 2-­‐2 Vero cells that were superinfected with 5dl 1.2 herpes simplex virus (HSV) helper virus 1 d later. Recombinant virus was amplified through three passages and stored at –80°C as described previously (Neve, Howe et al. 1997). 1.4.6 Dunn chamber turning assay Dissociated E8/E13 chick or P5 rat DRG neurons were cultured with 1ng/ml (chick) or 5ng/ml (rat) NGF overnight on #3D glass coverslips coated with poly-­‐L-­‐lysine and laminin (10µg/ml). Coverslips were used for Dunn Chamber assembly as described previously (Yam, Langlois et al. 2009) to establish an NGF or MAG gradient (Outer well concentrations of 200ng/ml MAG-­‐Fc or 50ng/ml NGF for chick DRGs and 100ng/ml NGF for rat DRGs). Cell images were acquired every 3-­‐4 min for 100-­‐120 minutes on a temperature controlled stage. Neurites of at least 15 microns in length were tracked and the final position of the growth cone was used to determine the trajectory relative to the initial 15 microns. The angle turned over two hours, relative to the gradient position, was recorded. Measurements are presented in rose histograms in bins of 10 degrees with the length of each segment representing the frequency of measurements in percent. Only transduced neurons were quantified as assessed by GFP staining. Where indicated DRGs -­‐ 61 -­‐ Results were treated with Sp-­‐cAMPS, Rp-­‐cAMPS, KT-­‐5720 or PKI for 1 hour prior to imaging in Dunn Chamber. 1.4.7 Immunoprecipitation HEK293T cells were transfected overnight with Lipofectamine 2000 and harvested in NP-­‐40 buffer (150 mM NaCl, 1% Nonidet P-­‐40, 10 mM Tris-­‐Cl, pH 8.0, 10% (w/v) glycerol) with complete EDTA free protease inhibitor cocktail (Roche) and phosphatase inhibitors (5mM NaF and 1mM Na3VO4). Following sonication, lysates were pre-­‐cleared with Protein A/G agarose beads (Santa Cruz) and then incubated with anti-­‐V5-­‐ or anti-­‐myc-­‐ conjugated agarose (Sigma). Precipitates were washed with NP-­‐40 buffer and eluted in 2× SDS sample buffer. For immunoprecipitations from DRGs, P5 rat DRG explants were collected and harvested in NP-­‐40 buffer. Cleared lysates were incubated with anti-­‐ regulatory subunit beta2 antibody (BD bioscience) and protein A/G agarose beads. Precipitates were washed with NP-­‐40 buffer and eluted in 2× SDS sample buffer. Pheochromocytoma (PC-­‐12) cells were transduced with recombinant HSV-­‐WLRL-­‐GFP or HSV-­‐R18-­‐GFP and plated overnight in RPMI media (Invitrogen) with 10% Horse Serum, 5% Fetal Bovine Serum and 1% Pen/Strep. Lysates were collected as above and incubated with anti-­‐ regulatory subunit beta2 antibody (BD bioscience) and protein A/G agarose beads. Precipitates were washed with NP-­‐40 buffer and eluted in 2× SDS sample buffer -­‐ 62 -­‐ Results 1.5 RESULTS 1.5.1 14-­‐3-­‐3 proteins are expressed in neuronal growth cones To identify constituents of neuronal growth cones, proteins from mechanically purified growth cone preparations from E6 chick RGCs were analyzed by tandem mass spectrometry (Fig. S1). Of the total number of tandem mass spectra assigned to peptides, 4.1% are assigned to members of the 14-­‐3-­‐3 family and 14-­‐3-­‐3 proteins represent 6.5% of total proteins suggesting that they are abundant components of growth cones (Fig. S2). To verify these results we identified antibodies specific for the β, ζ, ε, γ, and τ isoforms that recognize a protein of the correct molecular weight in brain lysates (Fig. S3). By immunofluorescence we find that 14-­‐3-­‐3 γ, ε, ζ and τ are expressed in E6 chick RGC growth cones confirming the results of our 14-­3-­3 14-­3-­3 14-­3-­3 14-­3-­3 14-­3-­3 Chick
E6 RGC
E6 RGC
+ BP
E8 DRG
Rat
E13 DRG
P5 DRG
Figure 1: 14-­‐3-­‐3 proteins are present in growth cones Immunofluorescence staining of growth cones with isoform-­‐specific 14-­‐3-­‐3 antibodies. For each antibody the immunofluorescence signal is lost by pre-­‐
adsorbing the antibody with isoform-­‐specific blocking peptide (BP). RGC, retinal ganglion cell; DRG, dorsal root ganglion. Scale bar, 5 µm. -­‐ 63 -­‐ Results proteomic analysis, and we identified 14-­‐3-­‐3β as an additional isoform present in these growth cones (Fig. 1). To test if 14-­‐3-­‐3 expression extends to other types of growth cones we analyzed 14-­‐3-­‐3 expression in chick dorsal root ganglia (DRG) growth cones. Isoforms β, γ, ε and τ but not ζ are present in E8 chick DRG growth cones (Fig. 1). To determine if 14-­‐3-­‐3 expression in growth cones is developmentally regulated we compared 14-­‐3-­‐3 expression in E8 and E13 chick DRGs. E12 represents a critical transitional time point in growth cone development, cAMP levels in growth cones are high during early development and these drop sharply at E12 in chick (equivalent to postnatal day 4 in rat). The change in cyclic nucleotide concentration mediates differential phenotypic responses to guidance cues such as netrins, neurotrophins and myelin-­‐associated inhibitors (Song, Ming et al. 1997; Song, Ming et al. 1998; Song and Poo 1999; Shewan, Dwivedy et al. 2002; Nishiyama, Hoshino et al. 2003). We find that expression of 14-­‐3-­‐3 β, γ, and τ are upregulated in E13 chick (Fig. 1). The altered expression of individual 14-­‐3-­‐3 isoforms suggests that they may developmentally-­‐regulate responses to individual guidance cues. By Western blotting we find that all 14-­‐3-­‐3 isoforms are upregulated in E13 chick DRG lysates (Supplementary Fig. 4) suggesting that 14-­‐3-­‐3 expression can be locally regulated within the growth cone independent of the cell body. 1.5.2 A 14-­‐3-­‐3 antagonist converts NGF-­‐dependent repulsion to attraction To determine if 14-­‐3-­‐3 proteins regulate ligand-­‐dependent growth cone responses we investigated the impact of a 14-­‐3-­‐3 antagonist on DRG growth cone turning in response to NGF using a Dunn Chamber turning assay (Yam, Langlois et al. 2009). R18 (PHCVPRDLSWLDLEANMCLP) is a potent 14-­‐3-­‐3 peptide antagonist, which disrupts binding of Ser/Thr phosphorylated proteins to 14-­‐3-­‐3 (Wang, Yang et al. 1999). We generated a recombinant R18-­‐green fluorescent protein expressing herpes simplex virus (HSV-­‐R18-­‐
GFP) and a control HSV-­‐WLRL-­‐GFP virus that does not bind to 14-­‐3-­‐3 -­‐ 64 -­‐ Results because of a substitution in the core-­‐binding motif (WLDL to WLRL) (Wang, Yang et al. 1999). R18-­‐GFP binds to 14-­‐3-­‐3 and antagonizes a previously characterized 14-­‐3-­‐3ζ-­‐Slingshot1 interaction (Fig. S5) (Nagata-­‐Ohashi, Ohta et al. 2004). We confirmed that application of an NGF gradient results in an attractive mean turning angle when applied to E8 chick DRG growth cones (Gundersen and Barrett 1980; Gehler, Gallo et al. 2004) and we find that this response is unaffected by HSV-­‐R18-­‐GFP (12.6° +/-­‐ 5.8, Fig. 2a, d). However, E13 chick DRG neurons are repelled by NGF (-­‐14.3° +/-­‐ 2.5) and intriguingly this response is converted to an attractive response in HSV-­‐R18-­‐GFP but not HSV-­‐WLRL-­‐GFP-­‐transduced growth cones (R18, 13.4° +/-­‐3.4; WLRL, -­‐9.6° +/-­‐
3.1; Fig. 2b,c,e) demonstrating that 14-­‐3-­‐3 proteins are important for conferring repulsive responses to NGF in postnatal neurons. 1.5.3 A 14-­‐3-­‐3 antagonist blocks MAG-­‐dependent repulsion To test if 14-­‐3-­‐3 proteins regulate responses to other inhibitory cues we assessed the impact of HSV-­‐R18-­‐GFP on repulsion in response to the myelin-­‐
associated inhibitor MAG. Non-­‐infected or control (HSV-­‐WLRL-­‐GFP)-­‐
transduced E13 chick DRGs are repelled by MAG (-­‐11.3° +/-­‐ 4.2 and -­‐12.1° +/-­‐ 4.5 respectively; Fig 3a, d). HSV-­‐R18-­‐GFP-­‐transduced neurons are not repelled by MAG (4.0° +/-­‐ 2.9) but are not significantly attracted to MAG suggesting that R18 attenuates rather than switches the growth cone response to MAG (Fig. 3b, d). However, we noted that in the control WLRL-­‐
transduced DRG neurons a notable number of DRG neurons (43%) are not repelled by MAG (Fig. 3a) raising the possibility that the E13 chick DRG population may be heterogeneous in its responsiveness to MAG. When -­‐ 65 -­‐ Results a E8 DRG
-­30
0
-­60
-­90
-­30
30
16
12
8
60
repulsion attraction
WLRL + NGF
0
-­60
20
16
12
90 -­90
-­30
30
60
-­60
30
30
25
20
90 -­90
R18 + NGF
0
0
-­30
60
-­60
20
16
12
90 -­90
NI + NGF
30
No Gradient
60
90
b E13 DRG
c
R18 + NGF
NGF
WLRL 0’ 60’ 120’
R18
0’ 60’ 120’
Mean Turning Angle (+ SEM)
d E8
-­60
90 -­90
No Gradient
NGF
*
*
*
20
15
10
5
0
-­5
49
30
47 61 56
-­10
-­30
60
20
16
12
NI + NGF
e E13
20
15
-­60
90-­90
30
40
30
20
No Gradient
60
90
No Gradient
NGF
*#
10
5
0
0
54
-­5
84 71
59
R18
90 -­90
60
20
16
12
0
NI
WLRL
-­60
-­30
NI
WLRL + NGF
60
30
Mean Turning Angle (+ SEM)
20
16
12
0
WLRL
R18
-­90
-­30
30
NI
-­60
0
NI
-­30
-­10
-­15
-­20
*
*
Figure 2: The 14-­‐3-­‐3 antagonist R18 converts NGF-­‐dependent repulsion to attraction in E13 chick DRG neurons a, b. Rose histograms illustrating turning responses of E8 (a) or E13 (b) chick DRG neurons in response to an NGF gradient in a Dunn Chamber turning assay (50ng/ml NGF in outer chamber). Responses of individual neurons are clustered in 10 degree bins and the percentage of total neurons per bin is represented by the radius of each segment. Non-­‐transduced neurons (NI) or neurons transduced with HSV-­‐WLRL-­‐GFP or HSV-­‐R18-­‐GFP for 12-­‐16 hours prior to the turning assay were analyzed. c. Phase images of representative E13 chick DRG growth cones transduced with WLRL or R18 and exposed to an NGF gradient in the Dunn chamber turning assay. Scale bar = 20 microns d,e. Mean turning angles of E8 or E13 chick DRG neurons in the absence of a cue (No Gradient) or in response to 50 ng/ml NGF. Numbers within the bars indicate the number of growth cones measured over at least three independent experiments. Statistics were performed by one-­‐way ANOVA with Tukey’s post-­‐test. * indicates p<0.05 compared to no gradient control. # indicates p<0.05 compared to the NGF-­‐
dependent repellent response in WLRL-­‐transduced neurons. -­‐ 66 -­‐ Results Figure 3: The 14-­‐3-­‐3 antagonist R18 blocks MAG-­‐dependent repulsion of E13 chick DRG neurons a-­‐b. Rose histograms illustrate turning responses of E13 chick DRG neurons in response to a MAG-­‐Fc gradient (200 ng/ml in outer chamber) in a Dunn Chamber turning assay. Neurons were transduced with HSV-­‐WLRL (a) or HSV-­‐
R18 (b) for 12-­‐16 hours prior to the turning assay. c. An overlay of the rose histograms presented in (a) and (b) illustrating a large population of R18-­‐
positive neurons with an attractive turning response to a MAG gradient. d. Mean turning angles of non-­‐infected, WLRL-­‐ or R18-­‐transduced E13 chick DRG neurons in the absence of a gradient or in response to MAG. Numbers within the bars indicate the number of growth cones measured over at least three independent experiments. Statistics were performed by one-­‐way ANOVA with Tukey’s post-­‐test. * indicates p<0.05 compared to no gradient control. # indicates p<0.05 compared to the MAG-­‐dependent repellent response in WLRL-­‐
transduced neurons. overlaying the turning data from WLRL-­‐ and R18-­‐transduced neurons, a proportion of DRG neurons do become attracted to MAG when transduced with R18 (Fig. 3c). When neurons are grouped by turning angle into positive or negative bins (greater or less than 0.5 standard deviations of no gradient controls, respectively), 56.9% of WLRL-­‐transduced neurons show negative turning, compared to 39% of R18-­‐transduced neurons. Positive turning angles are observed for 49% of R18-­‐transduced neurons, compared to 26.4% of WLRL neurons. We conclude that 14-­‐3-­‐3 proteins are important for conferring repulsive responses to both NGF and MAG in E13 chick DRG neurons and that loss of 14-­‐3-­‐3 can convert a repellent response to an attractive response. -­‐ 67 -­‐ Results 1.5.4 Knockdown of specific 14-­‐3-­‐3 proteins converts NGF-­‐dependent repulsion to attraction R18 is a powerful reagent that simultaneously inhibits binding of all 14-­‐3-­‐3 isoforms to Ser/Thr phosphorylated targets allowing one to assess the global role of 14-­‐3-­‐3 proteins in a given phenotype. To verify that 14-­‐3-­‐3 proteins play a role in growth cone turning, we transduced neurons with lentiviruses expressing shRNA target sequences with a microRNA stem (shRNAmir) for knockdown of individual 14-­‐3-­‐3 isoforms. These experiments were performed in P4-­‐P7 rat DRGs, a stage equivalent to ~E13 in chick when cAMP levels have been developmentally downregulated. P5 rat DRG growth cones express all 14-­‐3-­‐3 isoforms examined (β, ζ, ε, γ, and τ; Fig. 1). Similar to E13 chick, the repellent response to NGF in P5 rat DRG neurons is switched to an attractive response with transduction of HSV-­‐R18-­‐GFP, but not HSV-­‐WLRL-­‐
GFP (R18, 9.45 degrees +/-­‐4.75; WLRL, -­‐15.86 degrees +/-­‐5.05). Transduction of P4 rat DRG neurons with lentivirus for 14-­‐3-­‐3 isoform-­‐
specific knockdown effectively and specifically reduces expression levels of the targeted isoform by 60-­‐90% (Figs. 4a, b). In the Dunn chamber turning assay we find that knockdown of 14-­‐3-­‐3ε, β or γ but not ζ or τ converts NGF-­‐
dependent repulsion to attraction (Fig. 4c; ε, 10.6° +/-­‐ 3.5; β, 11.2° +/-­‐ 3.9; γ, 9.9° +/-­‐ 3.3; ζ, -­‐13.9° +/-­‐ 4.7; τ. -­‐9.5° +/-­‐ 4.8). This data provides an independent confirmation that 14-­‐3-­‐3 proteins are important for NGF-­‐
dependent repulsion and demonstrates that 14-­‐3-­‐3ε, β and γ mediate this response. -­‐ 68 -­‐ Results 14-­3-­3
14-­3-­3
14-­3-­3
14-­3-­3
14-­3-­3
GAPDH
Isoform Target
*
*
*
Isoform Target
*
20
15
NGF
No Gradient # * #
*
5
0
-­5
-­10
-­15
-­20
-­25
127
10
93 90 84 99
*
*
*#
111 103 106
14-­3-­3 NI
14-­3-­3 Control
14-­3-­3 c
NI
14-­3-­3 180
160
140
120
100
80
60
40
20 *
0 Mean Turning Angle (+ SEM)
b
14-­3-­3
Expression (% Control + SEM)
a
*
*
Figure 4: Loss of 14-­‐3-­‐3εε , β or γ converts NGF-­‐dependent repulsion to attraction in P5 rat DRG neurons a. Lysates from P5 rat DRG neurons transduced with lentiviruses for knockdown of individual 14-­‐3-­‐3 isoforms and analyzed by Western blot with anti-­‐GAPDH or anti-­‐14-­‐3-­‐3 isoform-­‐specific antibodies. The GAPDH blot shown is from the 14-­‐3-­‐3β gel and is representative of the equal loading achieved in these lysates. b. Expression of individual isoforms quantified by densitometry from Western blots of P5 rat DRG lysates transduced with lentiviruses for knockdown of individual isoforms. Densities were normalized to GAPDH levels for each blot. The mean expression of each isoform is expressed as a percentage of control from 3 independent experiments (+ SEM). * indicates p<0.05 compared to expression in control miRNA transduced neurons. c. Mean turning angles of P5 rat DRG neurons in the absence of a cue (No Gradient) or in response to 100 ng/ml NGF. Neurons were non-­‐transduced (NI) or were transduced with lentiviruses for expression of a non-­‐targeting control sequence or sequences for knockdown of individual 14-­‐3-­‐3 isoforms. Numbers within the bars indicate the number of growth cones measured over at least three independent experiments. Statistics were performed by one-­‐way ANOVA with Tukey’s post-­‐test. * indicates p<0.05 compared to no gradient control. # indicates p<0.05 compared to the NGF-­‐dependent repellent response in control shRNAmir-­‐transduced neurons. 1.5.5 14-­‐3-­‐3 proteins switch NGF-­‐dependent growth cone turning responses through PKA The switch in growth cone turning response is reminiscent of the switching responses observed in response to regulating levels of cyclic nucleotides (Song, Ming et al. 1997; Song, Ming et al. 1998). Neurotrophin-­‐dependent attraction can be converted to repulsion in the presence of the cAMP -­‐ 69 -­‐ Results antagonist Rp-­‐cAMPS and MAG-­‐dependent repulsion can be converted to attraction with the cAMP agonist Sp-­‐cAMPS (Song, Ming et al. 1997; Song, Ming et al. 1998). We find that Sp-­‐cAMPS converts NGF-­‐dependent repulsion to attraction in E13 chick DRG neurons (12.5° +/-­‐ 4.4) similar to the effect of R18 (Fig. 5). Combinatorial treatment with Sp-­‐cAMPS and R18 does not have an additive effect (16.5° +/-­‐ 2.2) suggesting that cAMP and 14-­‐3-­‐3 proteins function in a convergent pathway (Fig. 5a). Further, attractive turning in response to an Sp-­‐cAMPS gradient (13.1° +/-­‐ 3.5) is blocked by the introduction of R18 (3.1° +/-­‐ 3.7) suggesting that 14-­‐3-­‐3 proteins function downstream of cAMP to affect growth cone turning (Fig. 5b). To further investigate the relationship between 14-­‐3-­‐3 proteins and cAMP-­‐dependent signaling we asked if R18-­‐ dependent switching in an NGF gradient would be affected by applying the cAMP antagonist Rp-­‐cAMPs or the protein kinase A antagonists KT-­‐5720 or PKI (Kase, Iwahashi et al. 1987; Dalton and Dewey 2006). Both the small molecule inhibitor KT-­‐5720 and the myristoylated peptide PKI act by targeting the catalytic subunit of the PKA holoenzyme I and together these inhibitors selectively address the function of PKA (Murray 2008). As expected, inhibition of cAMP or PKA did not impact NGF-­‐ dependent repulsion in WLRL-­‐transduced E13 chick DRGs; a stage with low levels of cAMP (Fig. 5c, d; RpCAMPS, -­‐9.3° +/-­‐ 3.1; KT-­‐5720, -­‐8.0° +/-­‐ 2.9). Intriguingly, we find that while Rp-­‐cAMPS has no significant effect on NGF-­‐
dependent attraction of R18-­‐transduced growth cones (6.2° +/-­‐ 3.1), R18-­‐
transduced neurons are no longer attracted to NGF when treated with KT-­‐
5720 or PKI (KT-­‐5720, -­‐4.8° +/-­‐ 3.8; PKI, -­‐3.4° +/-­‐ 4.0, Fig. 5c). Similarly, repellent responses to MAG are restored in R18-­‐transduced DRGs treated with KT-­‐5720 (-­‐6.0 +/-­‐ 4.1; Fig. 5d). Together these findings suggest that antagonizing 14-­‐3-­‐3 proteins leads to activation of PKA consequently converting NGF-­‐ and MAG-­‐dependent repulsion to attraction. -­‐ 70 -­‐ Results 71 78 97
-­5
-­10
59 84
94 62
79 72
-­5
*
68
-­10
-­15
KT-­5720
0
KT-­5720
5
MAG
Control
10
Control
PKI
0
Mean Turning Angle (+ SEM)
0 59 60
5
*
KT-­5720
5
10
Rp-­cAMPS
*
15
KT-­5720
10
WLRL
R18
*
Rp-­cAMPS
WLRL
R18
15
20
Control
20
d
NGF
Control
-­20
Sp-­cAMPS
Mean Turning Angle (+ SEM)
-­10
71
49 59 115
Mean Turning Angle (+ SEM)
0
R18 + Sp-­cAMPS
10
Sp-­cAMPS
R18
20
c
b
NGF
Control
Mean Turning Angle (+ SEM)
a
49
WLRL
R18
-­15
Figure 5: 14-­‐3-­‐3 proteins regulate growth cone turning responses through PKA. a-­‐d. Mean turning angles of E13 chick DRG neurons in response to gradients of NGF (50 ng/ml), Sp-­‐cAMPS (20 µM) or MAG-­‐Fc (200 ng/ml). Neurons were transduced with HSV-­‐WLRL-­‐GFP or HSV-­‐R18-­‐GFP. Where indicated, neurons were treated with a bath application of 20 µM Sp-­‐cAMPS, 20µM Rp-­‐cAMPS, 200 nM KT-­‐5720 or 20 µM myristolated PKI for 60 minutes prior to the turning analysis. Numbers within the bars indicate the number of growth cones measured over at least three independent experiments. Statistics for a,c and d were performed by two-­‐way ANOVA with Bonferroni post-­‐tests. Satistics for b and PKI treatment were performed with unpaired Student’s t-­‐test. * indicates p<0.05 compared to the WLRL-­‐transduced control for each condition. 1.5.6 14-­‐3-­‐3 proteins bind and regulate PKA The PKA holoenzyme is a tetramer, with two catalytic subunits bound by two regulatory subunits that prevent the catalytic subunits from becoming active. cAMP binding to the regulatory subunits causes the release of the catalytic subunits and subsequent activation of PKA. We asked if 14-­‐3-­‐3 isoforms could interact with PKA regulatory subunits α or β (RII α and RII β) that are known to be present in growth cones (Han, Han et al. 2007). First, we preformed co-­‐immunoprecipitation experiments in HEK 293T co-­‐transfected with myc-­‐tagged 14-­‐3-­‐3 isoforms and V5-­‐tagged RIIα or RIIβ. Immunoprecipitation of RIIα or RIIβ robustly co-­‐immunoprecipitated 14-­‐3-­‐
3γ and the strength of the co-­‐immunoprecipitation was diminished by introducing R18 but not WLRL (Fig. 6a). RΙΙα weakly co-­‐immunoprecipitated -­‐ 71 -­‐ Results 14-­‐3-­‐3ε and both regulatory subunits failed to co-­‐immunoprecipitate 14-­‐3-­‐
3β, τ or ζ. PKA binding to 14-­‐3-­‐3-­‐γ and ε isoforms compliments our turning data as knockdown of these isoforms resulted in switching from NGF-­‐
dependent repulsion to attraction Immunoprecipitation of 14-­‐3-­‐3γ also co-­‐immunoprecipitated RIIα and RIIβ (Fig. 6b). In this direction, the strength of the co-­‐immunoprecipitation is weaker indicating that only a small percentage of 14-­‐3-­‐3 may interact with PKA. We also detect an interaction between endogenous 14-­‐3-­‐3γ and RIIβ in P5 rat DRG lysates (Fig. 6c) supporting the physiological relevance of the interaction To test if 14-­‐3-­‐3 binding could regulate PKA, we tested the impact of R18 on the interaction between the regulatory and catalytic subunits of PKA. PC12 cells were transduced with HSV-­‐R18-­‐EGFP or HSV-­‐WLRL-­‐EGFP. Endogenous PKA regulatory subunit (RIIβ) was immunoprecipitated from cell lysates and the amount of co-­‐immunoprecipitating catalytic subunit (PKAcat) was assessed by Western blot. Introduction of R18 resulted in a 33% decrease in the amount of PKAcat that co-­‐immunoprecipitated with RIIβ when compared to WLRL control lysates (Fig.7a; p <0.05 by paired student’s t-­‐test) indicating that 14-­‐3-­‐3 proteins stabilize binding between the regulatory and catalytic subunits of PKA. To determine if the dissociation between regulatory and catalytic subunits correlated with an increase in PKA activity we monitored active PKA using an antibody that recognizes an activated form of the catalytic subunit of PKA (Parra and Zou 2010). Lysates from DRGs expressing R18 showed a 37% increase in the levels of phospho-­‐PKA as compared to those expressing WLRL (Fig 7b; p <0.05 by paired student’s t-­‐
test, n=3). We further examined HSV-­‐R18-­‐EGFP-­‐ or HSV-­‐WLRL-­‐EGFP-­‐ transduced growth cones for indications of PKA activity. PKA regulates growth cone turning in part through targeting the PKA substrate, -­‐ 72 -­‐ Results a V5
IP
-­V5
-­myc
input
RII-­V5 -­
-­myc
b IP
RII-­V5
14-­3-­3-­myc
input
IP
-­myc
-­
-­
-­
-­
IP
-­
-­
WLRL R18
c myc
WLRL R18
-­V5
Input
RIIb
IgG
IP
14-­3-­3-­myc
-­RII
-­14-­3-­3
-­V5
Figure 6: 14-­‐3-­‐3 proteins bind PKA a, b. 293T cells were co-­‐transfected with V5-­‐tagged RIIα or RIIβ and myc-­‐tagged 14-­‐3-­‐3 constructs and GFP-­‐WLRL or GFP-­‐R18 and then subjected to immunoprecipitation with anti-­‐V5 (a) or anti-­‐myc (b) antibody. Cell lysates and immunoprecipitates were separated by SDS-­‐PAGE and analyzed by Western blotting with anti-­‐V5 and anti-­‐myc antibodies. c. Cell lysates from P5 rat DRG neurons were subjected to immunoprecipitation with control IgG or anti-­‐RIIβ antibody and cell lysates and immunoprecipitates were analyzed by Western blot with anti-­‐14-­‐3-­‐3γ and anti-­‐RIIβ antibodies. PP-­‐1 inhibitory protein I-­‐1 (Han, Han et al. 2007). We find that an antibody that recognizes the phosphorylated form of I-­‐1 robustly stains a subset of neurites in P5 rat DRG explants. We observe that phospho-­‐I-­‐1 staining extends beyond the neurite into the central domain of the growth cone in R18-­‐transduced explants while the staining is largely restricted to the neurite in WLRL-­‐transduced explants (Fig. 7c). The growth cone to neurite ratio of phosphorylated I-­‐1 shows a 1.73 fold increase in R18 transduced explants as compared to the WLRL control (p <0.05 by student’s t-­‐test, n=29 and 31 for -­‐ 73 -­‐ Results WLRL-­‐ and R18-­‐transduced growth cones, respectively) demonstrating that 14-­‐3-­‐3s regulate PKA activity in neuronal growth cones. 1.6 DISSCUSSION Intracellular signalling within the growth cone is critical for interpreting extracellular cues and for regulating how growth cones navigate during development and following injury. In this study we have demonstrated that 14-­‐3-­‐3 proteins are major constituent proteins of growth cones. Our findings suggest that 14-­‐3-­‐3 proteins stabilize the PKA holoenzyme. Antagonizing 14-­‐ IP
GFP
-­RII
-­cat
NI
WLRL R18
-­cat
-­GFP
b
-p-­PKA
-PKA
c
RII
NI WLRL R18
WLRL
R18
phall./p-­I-­1 p-­I-­1 phalloidin
input
IP
a Figure 7: 14-­‐3-­‐3 proteins regulate the stability and activity of PKA holoenzyme. a. PC12 cells were transduced with HSV-­‐WLRL or HSV-­‐R18. Following transduction, cell lysates were subjected to immunoprecipitation with anti-­‐RIIβ antibody. Cell lysates and immunoprecipitates were analyzed by Western blot with anti-­‐RIIβ, anti-­‐PKA catalytic subunit (cat) and anti-­‐GFP antibodies. b. P5 rat DRGs were transduced with HSV-­‐WLRL or HSV-­‐R18. Cell lysates were analyzed by Western blot for levels of phosphorylated PKA catalytic subunit and total levels of PKA catalytic subunit. c. Immunofluorescence of P5 rat DRG growth cones transduced with HSV-­‐WLRL or HSV-­‐R18 and stained with an anti-­‐
phospho-­‐I-­‐1 antibody and rhodamine-­‐phalloidin. Scale bar = 10 microns -­‐ 74 -­‐ Results 3-­‐3 proteins activates PKA and converts NGF-­‐ and MAG-­‐dependent repulsion to attraction through the regulation of PKA. These findings provide novel insights into the intracellular mechanisms that control axon repulsion and suggest a novel molecular mechanism to convert repellent responses to growth promoting responses, a process which could positively impact neuronal repair following CNS injury. 1.6.1 14-­‐3-­‐3 expression in growth cones Here, we identify a number of growth cone constituent proteins from a mechanically isolated growth cone preparation. Of the proteins that we identified as growth cone constituents by tandem mass spectrometry, 85% overlap with those identified in a recent study of biochemically isolated growth cone particle fraction from rat forebrain (Nozumi, Togano et al. 2009). Both approaches identified 14-­‐3-­‐3ε, γ, ζ and τ as growth cone constituents while Nozumi and colleagues additionally identified 14-­‐3-­‐3β and η. As expected, the biochemical isolation is a valuable high yield approach, while the mechanical purification may more accurately predict growth cone constituent proteins. 14-­‐3-­‐3β, γ and τ are upregulated between E8 and E13 in chick DRG growth cones while 14-­‐3-­‐3ε and ζ are not. By Western blotting, a general increase in expression of each 14-­‐3-­‐3 isoform can be detected in E13 DRG cell lysates (Fig. S4) suggesting that 14-­‐3-­‐3 expression is regulated locally through translation of 14-­‐3-­‐3 mRNA as has been seen for other proteins (Lin and Holt 2007). The weak 14-­‐3-­‐3 expression in E8 chick DRG growth cones may explain why responses to NGF are unaltered by antagonizing 14-­‐3-­‐3 function with R18 at this stage (Fig. 2). Alternatively, cAMP levels and associated PKA activity are elevated at E8 and thus R18-­‐dependent enhancement of PKA activity would not necessarily be expected to affect attractive turning responses. For example, bath application of Sp-­‐cAMPS to Xenopus spinal neurons does not significantly impact netrin-­‐dependent attraction (Ming, -­‐ 75 -­‐ Results Song et al. 1997). 14-­‐3-­‐3 proteins are however strongly expressed in E6 chick RGCs (Fig. 1). At this stage RGCs are undergoing a number of well characterized guidance decisions raising the possibility that 14-­‐3-­‐3 proteins may also play additional roles in regulating axon guidance during development through PKA-­‐dependent or –independent mechanisms. 1.6.2 14-­‐3-­‐3 proteins regulate PKA Our functional and biochemical data suggests a model whereby 14-­‐3-­‐3 proteins regulate PKA activity to modify neuronal responses to NGF and MAG (Fig. 8). We postulate that 14-­‐3-­‐3 proteins regulate PKA because R18 converts NGF-­‐ and MAG-­‐dependent repulsion to attraction in the presence of the cAMP antagonist Rp-­‐cAMPS but not the PKA antagonists KT5720 or PKI. Introduction of R18 dissociates the PKA catalytic and regulatory subunits suggesting that 14-­‐3-­‐3 proteins stabilize the PKA holoenzyme. However it is also well established that the spatial localization and regulation of PKA activity is essential for proper growth cone turning in Xenopus spinal neurons (Han, Han et al. 2007) and it is conceivable that 14-­‐3-­‐3 proteins may have a parallel function cooperating with A-­‐Kinase anchoring proteins (AKAPs) which anchor PKA to phosphorylation targets consequently regulating local PKA activity. Upon local stimulation of DRG growth cones with an NGF gradient, we do not find any evidence of asymmetric localization of 14-­‐3-­‐3γ by immunofluorescence (data not shown). However, this does not exclude the possibility that localized changes in the phosphorylation state of 14-­‐3-­‐3 proteins or binding targets enable spatial regulation of PKA activity Knockdown of 14-­‐3-­‐3γ, β or ε expression mimics the ability of R18 to convert NGF-­‐dependent repulsion to attraction. Biochemically we observe a robust interaction between PKA and 14-­‐3-­‐3γ but weak or no interaction with 14-­‐3-­‐
3β or ε, respectively. One possible explanation for our findings is that 14-­‐3-­‐
3γ/β and 14-­‐3-­‐3γ/ε heterodimers bind to PKA and that both complexes are necessary for regulating PKA, perhaps through favouring specific regulatory -­‐ 76 -­‐ Results or catalytic PKA subunits. Our biochemical data suggests a model whereby 14-­‐3-­‐3γ may mediate binding to PKA and that 14-­‐3-­‐3β and 14-­‐3-­‐3ε may complex with PKA through binding to 14-­‐3-­‐3γ in an unstable heterodimer that is difficult to retain in our co-­‐immunoprecipitation analysis. An alternative possibility is that a 14-­‐3-­‐3γ/ε heterodimer may stabilize PKA while 14-­‐3-­‐3β may regulate growth cone responses through a PKA-­‐
independent pathway. Treatment with the PKA inhibitors, KT-­‐5720 and PKI did not completely restore repellent responses to MAG and NGF in R18-­‐ Figure 8: Model for 14-­‐3-­‐3 regulation of PKA in growth cone turning response. In P5 rat DRG growth cones, 14-­‐3-­‐3γ/ε heterodimers (shown in green) bind to PKA directly or through an adaptor protein (AP) and downregulate the activity of the PKA holoenzyme, resulting in a repellent response to a gradient of NGF (left). In the presence of the R18 peptide (red square), 14-­‐3-­‐3 binding is disrupted, leading to a dissociation of the active catalytic subunits of PKA (orange) from the regulatory subunits (blue) and an increase in PKA phosphorylation of downstream targets. This switches the response of the growth cone from repulsion to attraction (right).. -­‐ 77 -­‐ Results infected neurons (Fig. 5). While this could be ascribed to incomplete inactivation of PKA it does allow for the possibility that 14-­‐3-­‐3 proteins may affect growth cone responses through additional mechanisms. In addition to directly regulating PKA activity it is possible that 14-­‐3-­‐3s may also act as a tether to affect the localization of PKA. Further, 14-­‐3-­‐3 proteins may regulate growth cone responses by affecting the activity of cytoskeletal modulators including ADF/cofilin (Soosairajah, Maiti et al. 2005), members of the Calmodulin dependent kinase (CaMK) pathway (Davare, Saneyoshi et al. 2004), myosin light chain phosphatase (Koga and Ikebe 2008), and Rho GEFs (Meiri, Greeve et al. 2009). It is also important to acknowledge that viral-­‐
mediated transduction of R18-­‐GFP and shRNAmir affects 14-­‐3-­‐3 function throughout the cell body as well as the growth cone. 14-­‐3-­‐3 proteins are expressed in a number of cellular compartments and given the wide range of binding partners previously reported (Jin, Smith et al. 2004), it is possible that 14-­‐3-­‐3s also alter growth cone responses through additional mechanisms such as transcriptional regulation and vesicle trafficking. 1.6.3 PKA and growth cone turning How 14-­‐3-­‐3-­‐dependent PKA activation may convert NGF-­‐ and MAG-­‐
dependent repulsion to attraction remains an open question. PKA can effect the insertion of guidance cue receptors into the plasma membrane (Bouchard, Moore et al. 2004). Both NGF and MAG can engage p75 neurotrophin receptor (p75NTR) as part of their receptor complex in certain contexts raising the possibility that receptor dynamics may be altered by PKA activation at the site of ligand binding. PKA also targets IP-­‐1, an inhibitor of PP1. PP1 is an important downstream effector of Ca2+ signalling, another known modulator of growth cone response (Hong, Nishiyama et al. 2000) and these findings suggest a high degree of cross-­‐talk between cAMP and Ca2+ signalling. This idea is further supported by studies suggesting cAMP levels can alter Ca2+ signals by modifying voltage dependent Ca2+ channels (Henley, Huang et al. 2004) and through activation of ryanodine receptors, -­‐ 78 -­‐ Results modify Ca2+-­‐dependent release of intracellular Ca2+ (Ooashi, Futatsugi et al. 2005). 14-­‐3-­‐3 proteins may play a critical role in spatially and temporally regulating the interaction between PKA and such target substrates to regulate growth cone turning responses. -­‐ 79 -­‐ Results c
300 mm
50 mm
10 mm
d
Total Retina
b
Hoescht Phalloidin Hoescht
+ Phalloidin a
Retinal Growth Cone
1.7 SUPPLEMENTARY I NFORMATION -­NF
-­HDAC
50 mm
Figure 1S: Growth cone purification strategy. a. Schematic of the growth cone purification procedure. E6 chicken retinae are flat mounted on nitrocellulose filters, cut into 400 µm tissue strips and grown on PLL/laminin-­‐coated culture dishes for 48 hours. Neurites were cut, explants removed from the culture dish and growth cones collected in lysis buffer. b. Three magnifications of rhodamine-­‐phalloidin-­‐stained retinal cultures cut with a glass electrode. Growth cones are clearly visible at the highest magnification. c. Hoescht and phalloidin stain of a retinal explant demonstrating that Hoescht-­‐
positive cell bodies do not migrate out of the explant and into the region of neurite/growth cone outgrowth. d. Western blot of a total retinal lysate compared to an aliquot of the neurite/growth cone preparation which was analyzed by LC-­‐Q-­‐Tof-­‐MS/MS. As expected the neurite/growth cone preparation is highly enriched for neurofilament (NF) and devoid of the nuclear histone deacetylase marker (HDAC). -­‐ 80 -­‐ Results a
Accession #
gi|82197919
gi|63018
gi|38322686 gi|57529758
gi|63689
gi|50759528
gi|119331154
gi|135393
gi|135423
gi|135474
gi|135468
gi|417855
gi|86467 gi|135446
gi|123903408
Alpha-­enolase
Aspartate aminotransferase, mitochondrial precursor
ATP synthase subunit alpha
ATP synthase subunit beta, mitochondrial precursor
Carbonic anhydrase 2
Creatine kinase B-­type
Fatty acid synthase
Glyceraldehyde-­3-­phosphate dehydrogenase
L-­lactate dehydrogenase A chain
L-­lactate dehydrogenase B chain
3-­phosphoglycerate dehydrogenase
Phosphoglycerate kinase
Phosphoglycerate mutase 1
Pyruvate kinase muscle isozyme
Protein name
Actin, cytoplasmic 2
Beta-­actin
Capping protein (actin filament) muscle Z-­line, alpha 2 Dynein, cytoplasmic 1, intermediate chain 2 Neurofilament, middle polypeptide
similar to neurofilament-­L isoform 2
Profilin-­2
Tubulin alpha-­1 chain
Tubulin alpha-­5 chain
Tubulin beta-­7 chain
Tubulin beta-­4 chain
Tubulin beta-­2 chain
Tubulin beta-­3 chain
Tubulin beta-­1 chain
Vimentin
1
1
2
2
4
6
6
5
4
Cytoskeleton
and regulators
gi|1706653 gi|112981
gi|45383566
gi|50804057
gi|833606
gi|45384340
gi|45384052
gi|211797
gi|65925 gi|126046
gi|50751002
gi|45384486
gi|82082619
gi|45382651
60S ribosomal protein L30
Chromodomain-­helicase-­DNA-­binding protein 7
Eukaryotic initiation factor 4A-­II
Elongation factor 1-­alpha
Eukaryotic translation elongation factor 1
Histone H2B 1/2/3/4/6
Histone H4-­VIII
Histone H3.2
Histone H2A
2
3
3
2
4
1
1
Cell signaling and membrane receptor
Accession #
gi|164429959
gi|462065
gi|122378
gi|50751796 gi|82115183
60 kDa heat shock protein, mitochondrial precursor
Heat shock cognate 70 kDa protein
Heat shock protein HSP 90-­alpha
T-­complex protein 1 delta subunit
Ubiquitin
Ubiquitin carboxyl-­terminal hydrolase isozyme L1
Ubiquitin-­ribosomal protein fusion protein
Protein name
Cellular retinoic acid binding protein I
Fatty acid-­binding protein, brain
Hemoglobin subunit alpha-­A
similar to retinol dehydrogenase 8 (all-­trans)
Voltage-­dependent anion channel
1
2
3
3
1
2
1
2
4
2
1
1
Catabolism, 22.6%
peptide #
Proteasomal degradation and stress response
gi|50750210 gi|2996407
gi|47604960
gi|8670805
gi|51703335
gi|121490095
gi|82069516
2
1
2
1
1
Other
Vesicular transport
Coatomer protein complex, subunit beta Rab-­GDP dissociation inhibitor
RAB5A, member RAS oncogene family similar to Vesicle amine transport protein 1 homolog
Vesicle-­associated membrane protein 2
Vesicular transport, 8.1%
Proteasomal degradation and stress response, 11.3%
Other, 8.1%
Cell signaling , 11.3%
Cytoskeleton, 24.2%
gi|57530020
gi|45384364
gi|57530730
gi|50760787
gi|82130924
b
14-­ 3-­3, 6.5%
Transcription and translation, 14.5%
-­‐ 81 -­‐ peptide #
11
2
2
1
6
3
1
31
1
25
16
5
3
3
6
Catabolism
gi|54039169
gi|117606190
gi|45383077
gi|2119922
gi|44969758 gi|119331174
gi|1493813
gi|55977060
gi|82207461
14-­3-­3 protein epsilon
14-­3-­3 protein gamma
14-­3-­3 protein zeta
14-­3-­3 protein tau
Collapsin response mediator protein-­2B
similar to guanine nucleotide binding protein alpha oB
Low-­density lipoprotein receptor precursor
12
3
2
3
1
2
2
7
3
2
1
2
4
5
Transcription and translation
gi|82197924
gi|82194891
gi|82197807
gi|53127458 gi|33340025
gi|50753611 gi|31455368
Results Figure 2S: Growth cone proteomics analysis a. Proteins identified in the growth cone proteomics divided by function/localization. Peptide counts for each identified protein are shown. b. Pie chart representing the distribution of proteins identified. The 14-­‐3-­‐3 proteins represent 6.5% of total hits. a
14-­3-­3 14-­3-­3 -pan 14-­3-­3 *
-14-­3-­3 *
*
*
*
-14-­3-­3 27 kDa
25 kDa
14-­3-­3 14-­3-­3 14-­3-­3 14-­3-­3 pcDNA
-14-­3-­3 -14-­3-­3 -14-­3-­3 -14-­3-­3 -14-­3-­3 b
-14-­3-­3 -14-­3-­3 -14-­3-­3 -14-­3-­3 *
-myc Figure 3S: Validation of 14-­‐3-­‐3 antibodies. a. P5 Rat DRG lysates Western blotted with individual isoform-­‐specific 14-­‐3-­‐
3 antibodies. b. Isoform specific antibodies were assessed for specificity by transfecting HEK293T cells with plasmids encoding 2X myc tagged versions of each isoform to confer a molecular weight shift. Transfected lysates were analyzed by Western blot with isoform-­‐specific 14-­‐3-­‐3 antibodies. All antibodies expect for the 14-­‐3-­‐3η antibody specifically detect the correct 14-­‐
3-­‐3 isoform. * indicates endogenous protein and arrows indicate transfected protein. -­‐ 82 -­‐ 14
-­3
-­3
14
-­3
-­3
-­3
-­3
14
14
14
-­3
-­3
-­3
-­3
Fold Change (E13/E8)
Results Figure 4S: 14-­‐3-­‐3 isoform expression in chick DRG lysates. Quantification of 14-­‐3-­‐3 isoform expression in E8 and E13 chick DRG lysates as assessed by densitometry of Western blots. Fold change is presented as the ratio of the E13 to E8 signal for each isoform normalized to GAPDH to control for equal protein loading. a
IP
G
Ig
b
-­3
-­3
14
myc SSH1 +
V5 14-­3-­3 +
R18 EGFP WLRL EGFP -­myc
-­V5
-­GFP
+
+
+
+
+
+
-
IP
-­GFP
Input
Input
IP
R18 EGFP + +
-­14-­3-­3
myc IP IgG -­V5
Figure 5S: Validation of the R18 peptide. a. PC12 cells were transduced with HSV-­‐R18-­‐EGFP. Cells were lysed 20 hours after transduction and subjected to immunoprecipitation with a 14-­‐3-­‐3γ antibody or control IgG. R18 co-­‐immunoprecipitates with 14-­‐3-­‐3γ. b. HEK 293T cells were transfected with plasmids encoding V5-­‐14-­‐3-­‐3ζ, the known 14-­‐3-­‐3 binding target myc-­‐SSH1, and WLRL-­‐EGFP or R18-­‐EGFP. R18-­‐EGFP but not WLRL-­‐EGFP disrupts the co-­‐immunoprecipitation of 14-­‐3-­‐3ζ with SSH1. -­‐ 83 -­‐ Results 2. A cell-­‐intrinsic switch from Sonic hedgehog-­‐
mediated attraction to repulsion of commissural axons a fter m idline c rossing r equires 1 4-­‐3-­‐3 p roteins Patricia T. Yam‡,1,2,5, Christopher B. Kent‡2, W. Todd Farmer,1,Steves Morin1, Ricardo Alchini 2,5 , Léa Lepelletier1, David R. Colman2,5, Marc Tessier-­‐
Lavigne6, Alyson Fournier2,5, * and Frédéric Charron1,3,4,5,* (Submitted) 1Molecular Biology of Neural Development, Institut de Recherches Cliniques de Montréal (IRCM), 110 Pine Ave West, Montreal, Quebec, Canada, H2W 1R7 2Department of Neurology and Neurosurgery, Montreal Neurological Institute and Hospital, McGill University, Montreal, Quebec H3A 2B4, Canada 3Department of Medicine, University of Montreal, Montreal, Quebec, Canada 4Department of Anatomy and Cell Biology, Department of Biology, Division of Experimental Medicine, McGill University, Quebec, Canada 5Program in Neuroengineering, McGill University, Montreal, Quebec, Canada 6Genentech, Inc., 1 DNA Way, South San Francisco, California 94080, USA; Present address: The Rockefeller University, New York, NY 10065, USA *Co-­‐Corresponding authors Alyson Fournier, PhD and Frédéric Charron, Ph.D. ‡ PTY and CBK contributed equally to this work -­‐ 84 -­‐ Results 2.1 PREFACE We have identified 14-­‐3-­‐3 proteins as major constituents of the axonal growth cone and established a novel function of 14-­‐3-­‐3 isoforms in modulating directional growth cone responses to external cues by regulating the activity of PKA. Disrupting the 14-­‐3-­‐3/PKA interaction in cultured DRG neurons results in a release of active PKA catalytic subunit and the reversal of a repellent turning response to a number of external cues. This finding raises the possibility that 14-­‐3-­‐3 proteins are required for the proper directional response of the growth cone and are critical to establishing correct axonal pathfinding during the developing nervous system. The aim of the present study was to investigate whether 14-­‐3-­‐3 regulation of PKA is required for the proper extension of axons of dorsal commissural interneurons across the midline in the developing mammalian spinal cord. -­‐ 85 -­‐ Results 2.2 ABSTRACT During development of the nervous system, axons rely on a limited set of external cues to navigate through complex environments. This process relies on the growth cone’s ability to switch its directional response to gradients of external ligands. Sonic Hedgehog (Shh) attracts spinal cord commissural axons ventrally towards the floorplate, and after crossing the floorplate, commissural axons switch their response to Shh from attraction to repulsion. This prevents re-­‐crossing of the midline, as well as establishes anterior projections along the longitudinal axis by repelling axons down a posterior-­‐
high/anterior-­‐low Shh gradient. We show this switch is recapitulated in vitro with dissociated commissural neurons as they age, indicating that the switch is intrinsic and time-­‐dependent. In dissociated neurons, the switch correlated with increased expression of 14-­‐3-­‐3 proteins, which are also enriched in post-­‐crossing commissural axons in vivo. Inhibition of 14-­‐3-­‐3 function converted Shh-­‐mediated repulsion of aged dissociated neurons to attraction, similar to activation of PKA. As well, loss of 14-­‐3-­‐3 function prevented the correct anterior turn of post-­‐crossing commissural axons in vivo. Conversely, premature overexpression of 14-­‐3-­‐3 isoforms or inhibiting PKA function is sufficient to induce the repellent response to Shh in vitro and aberrant pre-­‐
crossing axonal projections in vivo. Therefore we identify a novel 14-­‐3-­‐3 protein-­‐dependent mechanism for a cell-­‐intrinsic temporal switch in the polarity of axon turning responses. -­‐ 86 -­‐ Results 2.3 INTRODUCTION Commissural interneurons in the developing spinal cord have stereotypic axonal projections that result from the coordinated activity of numerous guidance cues. Originating in the dorsal section of the neural tube, the axons project along a dorsal-­‐ventral (D-­‐V) axis, initially repelled by BMPs secreted from the roofplate, as well as attracted by Netrin-­‐1(Kennedy, Serafini et al. 1994; Serafini, Kennedy et al. 1994), Shh (Charron, Stein et al. 2003) and VEGF (Ruiz de Almodovar, Fabre et al. 2011), which derive from the floorplate at the ventral midline. Upon reaching and crossing the midline, the axons alter their responsiveness , becoming repelled by cues derived by the floorplate, such as Slits and Semaphorins (Zou, Stoeckli et al. 2000; Long, Sabatier et al. 2004). This ensures proper exit from the floorplate and prevents re-­‐crossing of the midline. After the floorplate, axons are induced to turn their projections along the anterior-­‐posterior (A-­‐P) axis, continuing towards targets in the anterior direction. The mechanisms by which commissural axons alter their responsiveness to guidance cues at the floorplate and make this anterior turn have yet to be fully understood. The secreted morphogen Wnt4 has been shown to be present at the mRNA level in the developing spinal cord of mammals in a high anterior-­‐low posterior gradient along the AP axis and acts as an attractant cue to post-­‐
crossing commissural neurons, (Lyuksyutova, Lu et al. 2003) indicating a role for Wnt4 in establishing the anterior projection, but findings in chick also suggest another morphogen, Shh, is present in a high posterior-­‐low anterior gradient and is responsible for inducing a repellent turn down the A-­‐P axis (Bourikas, Pekarik et al. 2005). A previous study of Shh signalling affecting post-­‐crossing commissural neurons in mammals, advances the idea that Shh is responsible for inducing -­‐ 87 -­‐ Results responsiveness to Semaphorin repulsion after midline crossing (Parra and Zou 2010). Interfering with Shh signalling in the rat open-­‐book preparation results in axon stalling, knotting and re-­‐crossing in the midline, reminiscent of the phenotypes induced by knocking out Sema receptor Neuropilin 2 (Zou, Stoeckli et al. 2000). Notably, Shh disruption also showed guidance defects in post-­‐crossing axons along the A-­‐P axis, with significant numbers of axons projecting in the posterior direction instead of anterior (Parra and Zou 2010). We confirm the rat open-­‐book findings by examining the AP projections of commissural neurons in mutant mice lacking either a floorplate (Gli2-­‐/-­‐) or conditionally knocked-­‐out for a key positive regulator of Shh signaling, smoothend (Smo). (Wnt1-­‐cre, Smo n/c; Math1-­‐Cre, Smo n/c). These mutants also showed randomization of axonal projections in the A-­‐P axis. In addition, whole mount staining of open book preparations of developing spinal cords with anti Shh revealed the presence of a high posterior-­‐low anterior gradient along the A-­‐P axis in mammals, similar to that found in chick (Bourikas, Pekarik et al. 2005). Together, these findings support the possibility that Shh plays an additional role in mammalian post-­‐
crossing commissural neurons establishing a proper anterior projection. Beyond inducing responsiveness to Semaphorins, we find that Shh acts directly as a repellent cue on post-­‐crossing commissural neurons and hypothesize that this repellent signal is responsible for establishing a proper anterior turn along the A-­‐P axis. Intriguingly, pre-­‐crossing commissural neurons are attracted along the D-­‐V axis by Shh and thus must switch their responsiveness to be repelled by the same cue after crossing the midline. How this post-­‐crossing switch occurs is unknown. It has been well established in vitro that many mechanisms can switch axon guidance responses from attraction to repulsion, including altering levels of cyclic nucleotides, calcium signalling and growth substrates (Ming, Song et al. 1997; Song, Ming et al. 1997; Song, Ming et al. 1998). Much -­‐ 88 -­‐ Results fewer examples of guidance switches occurring in vivo during development have been described, with the most well established being the change from attraction to Netrin-­‐1 to repulsion in the optic nerve projections of retinal ganglion cells due to changes in the level of cAMP (Hopker, Shewan et al. 1999; Shewan, Dwivedy et al. 2002). Here we show that the in vitro repellent response of commissural neurons to gradients of Shh occurs after an extended period in culture, and does not require the presence of the floorplate. This indicates a cell-­‐intrinsic time dependent mechanism that can recapitulate the behaviour of pre-­‐ and post-­‐
crossing axons in vivo. We previously identified 14-­‐3-­‐3 proteins as modulators of axon guidance responses. Through negatively regulating Protein Kinase A (PKA) activity, 14-­‐3-­‐3 proteins are important for establishing the repulsive effects of MAG and NGF in older dorsal root ganglion neurons (Kent, Shimada et al. 2010). We find 14-­‐3-­‐3s are enriched in post-­‐crossing commissural axons and also increase in a time-­‐dependent manner in vitro, coinciding with a decrease in levels of active PKA. Inhibition of 14-­‐3-­‐3 function in older neurons switches the response from Shh repulsion to attraction in vitro, similar to selective pharmacological activation of PKA. Conversely, overexpression of 14-­‐3-­‐3 proteins in younger neurons is sufficient to induce a repellent turning response to Shh in vitro, similar to blocking PKA activity. Loss of 14-­‐3-­‐3 function in vivo randomizes the AP turning of post-­‐crossing commissural axons, and overexpression of 14-­‐3-­‐3s induces pre-­‐crossing commissural axons to turn in the anterior direction prematurely. Hence we identify a novel 14-­‐3-­‐3 protein-­‐dependent mechanism for a cell-­‐
intrinsic time-­‐dependent switch in the polarity of axon turning responses. This allows commissural axons, which are first attracted towards the ventral -­‐ 89 -­‐ Results midline of the spinal cord by Shh, to switch their response to Shh so that they become repelled by Shh after crossing the floorplate and migrate anteriorly along the longitudinal axis. 2.4 MATERIALS A ND M ETHODS 2.4.1 Animals All animal work was performed in accordance with the Canadian Council on Animal Care Guidelines and approved by the IRCM Animal Care Committee. Mice were maintained in the IRCM specific pathogen-­‐free animal facility. All mice lines have been previously described: Gli2 (Mo, Freer et al. 1997), Wnt1-­‐
Cre (Danielian, Muccino et al. 1998), Math1-­‐Cre (Matei, Pauley et al. 2005), Smon null allele (Zhang, Ramalho-­‐Santos et al. 2001), Smoc conditional allele (Long, Zhang et al. 2001), Hhip1 (Chuang, Kawcak et al. 2003), Math1-­‐LacZ (Ben-­‐Arie, Hassan et al. 2000), and Math1-­‐GFP (Lumpkin, Collisson et al. 2003). Wnt1-­‐Cre;Smon/+ or Math1-­‐Cre;Smon/+ mice were crossed with Smoc/c mice to generate Wnt1-­‐Cre;Smon/c and Math1-­‐Cre;Smon/c embryos. Non-­‐Cre+ and non-­‐Smon/c embryos were used as controls. Embryonic day 0 (E0) was defined as midnight of the night before a plug was found. 2.4.2 Reagents 14-­‐3-­‐3 isoform specific antibodies (β, γ, τ, ε and ζ) were from Santa Cruz Biotechnology and have been shown to be specific (Kent et al., 2010). Rat anti-­‐L1 antibody (clone 324) was purchased from Millipore, rabbit (monoclonal) anti-­‐PKA C (αβγ) T197 phosphospecific antibody from Cell Signaling (#5661), mouse anti-­‐PKA C from BD Transduction Laboratories (#610980), and anti-­‐phospho I-­‐1/DARPP-­‐32(Thr34) from Novus Biologicals. Mouse anti-­‐Shh antibody 5E1 and mouse anti-­‐Tag1 IgM monoclonal antibody 4D7 were obtained from the Developmental Studies Hybridoma Bank. Tat-­‐
R18-­‐YFP and Tat-­‐WLKL-­‐YFP vectors (Dong, Kang et al. 2008) were provided by Dr. Jing Chen (Emory University). The 14-­‐3-­‐3 isoform, R18-­‐EGFP and WLRL-­‐EGFP expression vectors have been previously described (Kent, -­‐ 90 -­‐ Results Shimada et al. 2010). Expression vectors with shRNA target sequences with a microRNA stem (shRNAmir) for knockdown of individual 14-­‐3-­‐3 isoforms were generated by subcloning previously identified target sequences (Kent, Shimada et al. 2010) into pCAGGS/ES vector. 6-­‐BNZ-­‐cAMP and KT-­‐5720 were purchased from Sigma-­‐Aldrich. Myrisolated PKI was purchased from Invitrogen (Carlsbad,CA) 2.4.3 DiI axon tracing Embryos were fixed at least overnight at 4°C in 4% paraformaldehyde in phosphate buffered saline (PBS). After fixation, the neural tubes were dissected from the embryo, pinned open and small 1,1’-­‐dioctadecyl-­‐3,3,3’,3’-­‐
tetramethylindocarbocyanine perchlorate (DiI, Molecular Probes, Eugene, OR) crystals were inserted to the medial neural tube dorsal of the motor column to label several individual cohorts per embryo (around 5-­‐9) at multiple levels along the AP axis (Farmer, Altick et al. 2008). The DiI was allowed to diffuse for 1 or 2 days at room temperature. After diffusion of the dye, the neural tubes were mounted open-­‐book and imaged on a Yokogawa spinning disk confocal system (Perkin-­‐Elmer, Wellesley, MA) connected to a Nikon (Tokyo, Japan) Eclipse TE2000. Z-­‐stacks were collected using MetaMorph imaging software (Molecular Devices, Palo Alto, CA). Analysis of axon guidance and image processing was performed on the resulting Z-­‐stacks in ImageJ (NIH). All images shown are maximum projections of the Z-­‐stacks. 2.4.4 Shh protein gradient detection Spinal cords dissected from rat E13.5 embryos were fixed for 1 hr in 4% PFA at 4°C. After washing in PBS, spinal cords were blocked in PBS + 10% heat-­‐
inactivated goat serum + 0.1% Triton (PHT), incubated overnight at 4°C with the 5E1 anti-­‐Shh antibody diluted 1/50 in PHT, washed extensively in PHT, incubated with a goat anti-­‐mouse-­‐HRP antibody overnight, and washed extensively in PHT. Shh immunoreactivity was detected using DAB. Rat spinal -­‐ 91 -­‐ Results cord tissue was used instead of mouse tissue to avoid immunoreactivity with the anti-­‐mouse secondary antibody. 2.4.5 Dissociated commissural neuron culture Staged pregnant Sprague Dawley rats were obtained from Charles River (St. Constant, Canada). Dissociated commissural neuron cultures were prepared from the dorsal fifth of E13 rat neural tubes as previously described (Yam, Langlois et al. 2009; Langlois, Morin et al. 2010). They were plated in Neurobasal supplemented with 10% heat-­‐inactivated FBS and 2 mM L-­‐
glutamine. After ~21 h, the medium was changed to Neurobasal supplemented with 2% B27 and 2 mM L-­‐glutamine. Dissociated commissural neuron cultures were used for experiments at 2 DIV (50-­‐55 h after plating) and 3-­‐4 DIV (76-­‐102 h after plating). For low-­‐density cultures, commissural neurons were plated at 50 000 cells/well for a 12-­‐well plate and 90-­‐120 000 cells/well for a 6-­‐well plate. 2.4.6 Dunn Chamber axon guidance assay The Dunn chamber axon guidance assay, imaging and analysis was performed as previously detailed (Yam, Langlois et al. 2009). Briefly, commissural neurons were grown on poly-­‐L-­‐lysine (100 μg/ml) coated square #3D coverslips (Fischer Scientific) at low density, and then assembled in a Dunn chamber after specified times in culture. Gradients were generated in the Dunn chamber with 0.1 μg/ml recombinant human Shh (C24II, R&D Systems), or buffer containing bovine serum albumin (BSA) (the vehicle for Shh) as the control in the outer well. After Dunn chamber assembly, time-­‐
lapse phase contrast images were acquired for a minimum of 2 h. The angle turned was defined as the angle between the original direction of the axon and a straight line connecting the base of the growth cone from the first to the last time point of the assay period (Figure 3D). -­‐ 92 -­‐ Results 2.4.7 Rat open-­‐book cultures Open-­‐book preparations of rat E13 spinal cords were isolated and cultured as previously described (Lyuksyutova, Lu et al. 2003). After 1 hr in culture, Tat-­‐YFP-­‐R18 or the control Tat-­‐YFP-­‐ WLKL was added to the culture media to a final concentration of either 100 or 150 μg/ml and cultured for 24 hr. Open-­‐book explants were fixed at room temperature with 4% PFA, washed with PBS and injected with DiI. 2.4.8 Chick electroporation Chick spinal cord electroporation was performed at HH st. 18/19 as described (Luria, Krawchuk et al. 2008). Briefly, a 5-­‐10 μg/μl solution of plasmid DNA was injected into the lumbar neural tube. The chick embryos were then electroporated using platinum/iridium electrodes (FHC) with an ECM 830 Electro Square Porator (BTX; Harvard Apparatus; 30 V, 5 pulses, 50 ms, at 1 s interval). Shells were sealed with Parafilm (Pechiney Plastic Packaging Company) and incubated at 38°C until harvesting at HH st. 28/29. 2.4.9 Immunostaining For immunostaining of spinal cord cross-­‐sections, mouse embryos were collected at E10.5 or E11.5, fixed in 4% PFA for 2-­‐3 h at room temperature, transferred to 30% sucrose at 4°C overnight, then and embedded in OCT and frozen. 12 μm sections were cut, and standard immunostaining techniques used. For immunostaining of dissociated commissural neuron cultures, cells were fixed for 15 min by adding an equal volume of 8% PFA, 30% sucrose in PBS to the cultures. The cells were washed in PBS and standard immunostaining techniques used. For cAMP immunostaining, dissociated commissural neurons were fixed with 5.5% acrolein in acetate buffer pH 4.75 for 30 min, quenched in 1% glycine in acetate buffer for 30 min, and reduced with 0.1% NaBH4 in TBS (Tris buffered saline) pH 7.4. Neurons were then blocked with TBS + 10% goat serum + 0.1% Triton X-­‐100 (TGT 10%) for 1 hr, incubated overnight at 4˚C with a rabbit anti-­‐cAMP antibody (Chemicon) diluted 1/1000 in TGT 1% serum, and the rabbit anti-­‐cAMP antibody -­‐ 93 -­‐ Results detected with a goat anti-­‐rabbit IgG-­‐Alexa Fluor 488 (Invitrogen) for 1hr. cAMP immunofluorescence intensity was measured using Volocity on images acquired under identical conditions. 2.4.10 Western blotting Cells were lysed in RIPA buffer (50 mM HEPES pH 7.4, 150 mM NaCl, 10% glycerol, 1.5 mM MgCl2, 1% Triton, 1% SDS, 1 mM EDTA) and boiled in SDS sample buffer for 3 min. Protein samples were separated by SDS-­‐PAGE and transferred to polyvinylidene fluoride membrane (PVDF). Secondary antibodies were conjugated to horseradish peroxidase and western blots were visualized with chemiluminescence. 2.5 RESULTS 2.5.1 Guidance of post-­‐crossing commissural axons in vivo along the AP axis requires Smoothened To evaluate the role of floorplate-­‐derived cues in the AP projection of post-­‐
crossing commissural axons, we analyzed open book preparations from Gli2-­‐
/-­‐ mouse embryos (Figure 1A), which lack a floorplate. In these mutants, commissural axons still project to the midline, due to expression of Netrin-­‐1 in the ventral ventricular zone (Matise, Lustig et al. 1999). In Gli2-­‐/-­‐ neural tubes, DiI labeled axons still switched from a DV to an AP axis of migration at the midline, but they became severely disorganized in AP axis. Their AP directionality appeared random, when compared to control Gli2+/-­‐ embryos, with no apparent preference for anterior versus posterior turns (Figure 1B), consistent with previous studies (Matise, Lustig et al. 1999). ~50% of the total fluorescence of the axons was distributed anteriorly, indicating complete randomization of the AP guidance of axons at the midline (Figure 1C). Thus, while the floorplate is not required for axons to switch from a DV to an AP axis of migration, it is required for the axons to correctly turn anteriorly after midline crossing. This suggested that a floorplate-­‐derived cue is important for correct anterior turning of post-­‐crossing commissural axons. -­‐ 94 -­‐ Results Figure 1: Smoothened is required for proper post-­‐crossing commissural axon guidance along the AP axis. -­‐ 95 -­‐ Results (A) Schematic of the stereotypical commissural axon trajectory in an intact spinal cord (left) and open-­‐book preparation (right). Axons travel ventrally towards the FP (gray) where they cross the midline. After crossing, the axons turn anteriorly and travel along the FP. (B) Top, DiI labeling of post-­‐crossing commissural axons in open-­‐book preparations of Gli2+/-­‐ and Gli2-­‐/-­‐ embryos at E11.5. Middle, zoom of the boxed region. Bottom, schematic of the axon guidance phenotype. Control Gli2+/-­‐ axons make a stereotypical anterior turn after exiting the FP (3 embryos, 13 cohorts). Gli2-­‐/-­‐ axons migrate ventrally to the midline, but become severely disorganized at the midline (3 embryos, 12 cohorts). Although axons still switch from a DV to an AP axis of migration, their AP directionality appeared random. (C) Relative fluorescence (+/-­‐SEM) of the anterior directed axons versus total fluorescence of anterior and posterior directed axons. In Gli2+/-­‐ cohorts (n=11), most of the fluorescence is on the anterior side of the label, while in the Gli2-­‐/-­‐ cohorts (n=11) the anterior fluorescence is equivalent to the posterior fluorescence. (D) Top, DiI labeling of post-­‐crossing commissural axons in open-­‐book preparations of Wnt1-­‐
Cre;Smon/c and Math1-­‐Cre;Smon/c embryos at E11.5. Middle, zoom of the boxed region. Bottom, schematic of the axon guidance phenotype. Axons from control littermates turn anteriorly after exiting the FP and do not re-­‐enter the FP (7 embryos, 33 cohorts). Axons of Wnt1-­‐Cre;Smon/c cohorts are extremely disorganized with a similar number of axons turning posteriorly and anteriorly (17/21 cohorts, 4 embryos). Several axons stall or re-­‐enter and re-­‐cross the FP (arrowheads). Post-­‐crossing commissural axons of Math1-­‐Cre;Smon/c cohorts displayed similar guidance defects to those observed with the axons of Wnt1-­‐
Cre;Smon/c cohorts, such as posterior-­‐turning axons (arrow) and axons which recross the FP (arrowheads), but with alower frequency and severity. (E) Relative fluorescence of the anterior directed axons versus total fluorescence of anterior and posterior axons in Wnt1-­‐Cre;Smon/c and Math1-­‐Cre;Smon/c cohorts. Wnt1-­‐Cre;Smon/c and Math1-­‐Cre;Smon/c have more posterior directed axons. (F) Quantification of recrossing axons as percentage of cohorts (+/-­‐SEM) per embryo. Unpaired t-­‐test was used for all statistical comparisons. D: dorsal; V: ventral; A: anterior; P: posterior; FP: floorplate; bracket: floorplate; dashed line: midline. Scale bars: (B,D Top) 50μm. (B,D Middle) 20μm. Shh is a floorplate-­‐derived molecule that acts as a guidance cue, attracting pre-­‐crossing commissural axons ventrally to the floorplate in mammals (Charron, Stein et al. 2003), and has been implicated in chick as a repellent for post-­‐crossing commissural axons (Bourikas, Pekarik et al. 2005). To test the involvement of Shh signaling in AP guidance, we analyzed conditional knockout mice for Smoothened (Smo), a key positive regulator of the Shh -­‐ 96 -­‐ Results signaling pathway (Charron, Stein et al. 2003; Dessaud, McMahon et al. 2008; Yam, Langlois et al. 2009). To generate commissural neurons null for Smo, we used Wnt1-­‐Cre to express Cre recombinase in the dorsal spinal cord (Charron, Stein et al. 2003), in Smon/c embryos which have one null allele (Smon) and one floxed allele (Smoc) at the Smo locus. In Wnt1-­‐Cre;Smon/c embryos, in which all dorsal commissural neurons are null for Smo, cohorts showed severe axon guidance defects as compared to littermate controls (Figure 1D). Most (17/21) had no clear anterior bias, with a ratio of anterior versus posterior axons of almost 50% (Figure 1E), indicating almost complete randomization of AP turning, similar to that of the Gli2-­‐/-­‐ mutants. To delete Smo specifically in commissural neurons, we used the Math1-­‐Cre driver. Math1-­‐Cre;Smon/c embryos displayed clear guidance defects along the AP axis (Figure 1D, right). The proportion of anteriorly directed axons was significantly lower than the control (Figure 1E), and many cohorts had axons that recrossed the floorplate (9/18 for mutants versus 1/33 for controls) (Figure 1F). Post-­‐crossing commissural axons of Math1-­‐Cre;Smon/c embryos were not as disorganized as those from Wnt1-­‐Cre;Smon/c embryos. However, this is expected since Math1 is expressed only in a subset of commissural neurons (Figure S1),(Gowan, Helms et al. 2001), and not all commissural neurons in these mutants are null for Smo. Together, these results show that Smo is required cell-­‐autonomously in commissural neurons for their post-­‐crossing axons to project in the proper direction along the AP axis, indicating that Shh signaling through Smo plays a role in guidance of post-­‐crossing commissural axons along the longitudinal axis. 2.5.2 Shh protein is present in a longitudinal gradient along the AP axis of the spinal cord The failure of Smo mutant axons to turn correctly along the AP axis might be due to an inability of Smo mutant post-­‐crossing axons to respond to Shh. To visualize the distribution of Shh protein in the neural tube we performed anti-­‐Shh whole-­‐mount staining on open-­‐book preparations (Figure 2A,B). A -­‐ 97 -­‐ Results Shh protein gradient was readily detectable and showed an approximately linear, posterior high, anterior low distribution (Figure 2C,D; correlation coefficient R2=0.88). To our knowledge, this is the first demonstration at the protein level of a diffusible guidance cue accumulating in a gradient along the AP axis. This finding is consistent with results from the developing chick spinal cord, where Shh mRNA was found to be expressed in a posterior high, anterior low gradient (Bourikas, Pekarik et al. 2005). The presence of Shh in an AP gradient along the floorplate, together with our results showing that Smo is required cell-­‐autonomously for post-­‐crossing commissural axons to turn anteriorly along the AP axis supports a model where a Shh gradient directs AP guidance of post-­‐crossing commissural axons in mammals. Shh
B
C
A
A
Relative Intensity
(a.u.)
A
Posterior
Anterior
Position on Neural Tube P
P
D
Relative Gradient Slope
Sample 1: -­1.15 ± 0.10
Sample 2: -­1.31 ± 0.29
Sample 3: -­1.97 ± 0.24
Figure 2: Shh protein accumulates in an AP gradient in the neural tube. (A) Open-­‐book-­‐view and (B) side-­‐view of wholemount Shh immunostaining of a rat E13.5 neural tube. A gradient of Shh protein is present along the neural tube with high Shh levels posterior (black arrowheads) and low Shh levels anterior (white arrowheads) (C) Plot of relative Shh staining intensity versus relative position along the neural tube. (D) Each independent neural tube measured (n=3) had a negative slope for the gradient. A: anterior; P: posterior -­‐ 98 -­‐ Results 2.5.3 Commissural neurons switch their response to Shh from attraction to repulsion over time in vitro The directionality of the Shh gradient, decreasing anteriorly, implies that Shh acts as a repellent on post-­‐crossing commissural axons. Although Shh has been proposed to function as a guidance cue for post-­‐crossing commissural axons in the chick (Bourikas, Pekarik et al. 2005), Shh gradients have not been shown to directly repel commissural axons. Explant-­‐based assays in which an explant is cultured a short distance from a source of the guidance cue, such as those performed by Bourikas et al. (Bourikas, Pekarik et al. 2005), cannot distinguish between biased outgrowth of axons and actual turning. To test whether Shh gradients can directly cause commissural axons to turn away, we used an in vitro assay for axon guidance based on the Dunn chamber (Yam, Langlois et al. 2009). Dissociated commissural neurons were grown in culture and then exposed to a gradient of Shh in the Dunn chamber after a specified numbers of days in culture. With this assay, the turning of axons can be imaged and measured in response to a defined gradient of a chemical cue over a short time period. Since the response of commissural neurons to guidance cues such as Slit changes with the age of the neurons (Stein and Tessier-­‐Lavigne 2001), we assayed neurons cultured from 2 to 4 DIV (days in vitro). We found that for commissural neurons at 3-­‐4 DIV, axons dramatically changed their direction of growth on exposure to a Shh gradient, turning away from higher concentrations of Shh (Figure 3A-­‐C). The Shh gradient rapidly stimulated the repulsion of axons with turning commencing within 1 h of application of the gradient, indicating that Shh has a direct effect on the guidance of commissural axons. Quantification of the angle turned (Figure 3D) indicated that axons in a control gradient have no net turning; the angle turned varied around 0° (-­‐0.82±4.3°, mean±SEM; Figure 3E,F). In a Shh gradient, however, axons from commissural neurons at 3-­‐4 DIV had a significant bias towards negative angles turned (-­‐14.8±5.0°, p<0.05, one way ANOVA; Figure 3E,F), indicating repulsion by Shh. The -­‐ 99 -­‐ Results 0 h
1 h
B
2 h
2 h
0 h
1 h
2 h
Shh
0 h
1 h
2 h
40
m
Control
90°
40
m
Shh
30
30
20
20
10
10
0
0
0°
attraction
C
1 h
D 20m
repulsion
Control
0 h
Shh
Control
angle turned
A
-­90°
final
10
initial
-­40
-­30
E
-­20
-­10
0
10
20
30
-­40
40
m
-­30
-­20
-­10
-­10
-­10
-­20
-­20
3-­4 DIV: Control gradient
-­30°
60°
repulsion
attraction
-­60°
60°
19.5
8
0
-­10
-­20
Control Shh
Shh
3-­4 DIV
2 DIV
p = 0.4025
20
15
10
5
0
Netrin
2 DIV
60°
H
p = 0.8287
25
20
15
-­150
50
25
-­100
-­50
0
50
100
Angle turned ()
10
J
5
0
75
Net axon growth (m)
10
25
20
90°
Control Shh
3-­4 DIV
3-­4 DIV
240
Axon length (m)
G
Net axon growth (m)
*
Mean angle turned Mean angle turned ()
20
*
12
16
90° -­90°
-­90°
I
F
n=49
30°
26
90°
20
0°
-­30°
-­60°
13
10
10
= initial angle
= angle turned
> 0° for turns up gradient
< 0° for turns down gradient
2 DIV: 0.1 g/ml Shh gradient
n=76
30°
20
-­90°
40
m
30
initial axon position
growth over 2 h
-­30°
30
20
0°
30°
-­60°
10
3-­4 DIV: 0.1 g/ml Shh gradient
n=46
0°
0
0
-­150
180
120
60
-­100
-­50
0
50
Angle turned ()
Figure 3: Commissural axons are repelled by Shh at 3-­‐4 DIV, but are attracted by Shh at 2 DIV. -­‐ 100 -­‐ 100
Results (A) Axons of commissural neurons at 3-­‐4 DIV do not change their direction of growth in a control gradient. (B) In a 0.1 μg/ml Shh gradient, axons of commissural neurons at 3-­‐4 DIV are repelled by high concentrations of Shh and turn toward low concentrations of Shh. For all images the concentration gradient increases along the y-­‐axis. (C) Trajectory plots of a sample of 10 axons in a control (left) or 0.1 μg/ml Shh (right) gradient. The initial axon position is black and the axon growth over 2 h colored according to the angle turned. Axons in the Shh gradient tend to turn down the gradient. (D) Definition of the initial angle, a, the angle between the initial axon position and the gradient, and angle turned, b, the angle between the vectors representing the initial and final position of the axon. (E) Rose histograms of the distribution of turned angles of commissural neurons at 3-­‐4 DIV or 2 DIV in a 0.1 μg/ml Shh gradient. Reponses of individual neurons were clustered in 15° bins and the percentage of total neurons per bin is represented by the radius of each segment. (F) The mean angle turned (+/-­‐ SEM) for axons in a control and 0.1 μg/ml Shh gradient for commissural neurons at 2 and 3-­‐4 DIV. p=0.0007, one way ANOVA, * = p<0.05. (G) The mean angle turned (+/-­‐ SEM) for axons in a 0.2µg/ml Netrin-­‐1 gradient for commissural neurons at 2 and 3-­‐4DIV. There is no significant change in the turning response to Netrin with time (p=0.40255, unpaired t-­‐test). (H) Net axon growth (+/-­‐ SEM) of commissural neurons at 3-­‐4 DIV. Axons in a 0.1 mg/ml Shh gradient have no significant difference in net growth compared to those in a control gradient (p=0.8287, unpaired t-­‐test). (I) Scatter plot of the net axon growth versus angle turned for axons of commissural neurons at 3-­‐4 DIV in a 0.1 mg/ml Shh gradient. The angle turned is independent of the net growth of the axon. (J) Scatter plot of the axon length versus the angle turned for axons of commissural neurons at 3-­‐4 DIV in a 0.1 mg/ml Shh gradient. The angle turned is independent of the initial axon length. Scale bar (A,B): 20 μm degree of repulsion by Shh was even more dramatic when those axons oriented towards increasing Shh concentrations, i.e. with initial angles between 0 and 90°, were considered. In this case, the mean angle turned was -­‐23.9°. Shh appeared to only affect turning, not growth, of these axons as the Shh gradient did not significantly change the growth rate of the axons compared to the control (p=0.8287) (Figure 3H). Furthermore the net axon growth in a Shh gradient showed no correlation with the angle turned (Figure 3I). -­‐ 101 -­‐ Results As previously shown (Yam, Langlois et al. 2009), commissural axons at 2 DIV were attracted up a Shh gradient, with a mean angle turned of 11.1±4.6° (Figure 3E,F). This contrasts sharply with the repulsion by Shh that we observed at 3-­‐4 DIV and suggests that the response of commissural neurons in vitro to Shh gradients changes over time. This is not a general change in the guidance response to all cues as commissural neurons cultured 3-­‐4DIV retain their attractive turning response to gradients of Netrin-­‐1, with a mean angle turned of 12.8±3.6°(Figure3G). Moreover, the length of the axon had no bearing on the degree of repulsion by Shh (Figure 3J), suggesting that the switch from attraction to repulsion by Shh is independent of axon length. The change in response to Shh over time is reminiscent of the changes in the response of commissural neurons in vivo to Shh gradients during development, with younger pre-­‐crossing axons attracted to Shh along the DV axis and older post-­‐crossing axons repelled by Shh along the AP axis. That isolated commissural neurons in culture maintain the ability to switch their response to Shh gradients suggests that the switch is cell intrinsic and temporally regulated. 2.5.4 14-­‐3-­‐3 protein expression in commissural neurons is also time-­‐
dependent We next looked for endogenous proteins that are expressed in a time-­‐
dependent manner and which could mediate the switch in Shh response. 14-­‐
3-­‐3 proteins were good candidates because they are major constituent proteins of growth cones and are important for the repulsive response to MAG and NGF in DRG neurons (Kent, Shimada et al. 2010). 14-­‐3-­‐3 proteins are adapter proteins that interact with phospho-­‐serine and phospho-­‐
threonine motifs in their binding partners. They control the spatial and temporal activity of their binding partners through regulating their subcellular localization, conformation or accessibility (Muslin and Xing 2000; Bridges and Moorhead 2005). -­‐ 102 -­‐ Results We used immunostaining to examine the expression of five 14-­‐3-­‐3 isoforms in the developing mouse spinal cord at E10.5 and E11.5, when commissural axons are crossing the floorplate. To visualize commissural axon tracts, we stained for Tag1, a marker of pre-­‐crossing commissural axons, and for L1, a marker of post-­‐crossing commissural axons. As illustrated by the Tag1 staining at E10.5 and E11.5, pre-­‐crossing commissural axons have a stereotyped DV trajectory towards the floorplate (Figure 4A, arrows). L1 staining, which reflects the path of post-­‐crossing commissural axons, changes between E10.5 and E11.5. At E10.5, commissural axons have just crossed the floorplate, and L1 expression is present predominantly at the floorplate and ventral funiculus. At E11.5, L1 expression extends up the lateral funiculi and widens in the ventral funiculus, illustrating the progression of the post-­‐
crossing axons (Figure 4A, arrowheads). The different 14-­‐3-­‐3 isoforms are all expressed in neural tissue (Figure 4B). Strikingly, both 14-­‐3-­‐3β and γ have an expression pattern in the neural tube that correlates with that of L1. Although 14-­‐3-­‐3β is expressed faintly in pre-­‐
crossing commissural axons, at E10.5 both 14-­‐3-­‐3β and γ are enriched at the floorplate and ventral funiculi, and at E11.5, their expression expands along the lateral funiculi and widens in the ventral funiculi. These changes in the distribution of 14-­‐3-­‐3β and γ mimics the changes in the pattern of L1 expression, and indicate that 14-­‐3-­‐3β and γ are enriched in post-­‐crossing commissural axons. 14-­‐3-­‐3τ is also present in post-­‐crossing commissural axons, being present at the floorplate and ventral funiculi at E10.5, and at the lateral funiculi at E11.5. However, in contrast to 14-­‐3-­‐3β and γ, it is expressed at significant levels in pre-­‐crossing commissural axons, with staining along the DV axonal tracts. 14-­‐3-­‐3ε and ζ are also present in neural tissue. However, they are expressed predominantly in cell bodies, rather than axonal processes. Hence, isoforms β, γ and τ are those expressed in post-­‐
crossing commissural axons. If 14-­‐3-­‐3 proteins are involved in the switch in -­‐ 103 -­‐ Results A
E10.5
B
E11.5
E10.5
E11.5
Tag1
14-­3-­3
D
V
L1
14-­3-­3
FP
VF
LF
VF
FP
C
2 DIV
FP
3 DIV
14-­3-­3
-­14-­3-­3
-­GAPDH
-­14-­3-­3
-­GAPDH
-­14-­3-­3
-­GAPDH
14-­3-­3
-­14-­3-­3
-­GAPDH
-­14-­3-­3
-­GAPDH
2.5
2.0
14-­3-­3
Fold change (3 DIV / 2 DIV)
D
1.5
1.0
0.5
0.0
E
14-­3-­3 isoform
-­‐ 104 -­‐ 14-­3-­3
isoform
Enriched post-­
crossing
in vivo
Change in
expression
levels
in vitro
+ +
+ +
+/-­
-­
-­
-­
Results Figure 4: 14-­‐3-­‐3 protein expression in commissural neurons is time-­‐
dependent. (A) Mouse E10.5 and 11.5 spinal cord cross-­‐sections immunostained for Tag1 (top), a marker of pre-­‐crossing commissural axons (arrows), and L1 (bottom), a marker for post-­‐crossing commissural axons (filled arrowheads). (B) Mouse E10.5 and 11.5 spinal cord cross-­‐sections immunostained for 14-­‐3-­‐3 isoforms. 14-­‐3-­‐3β, γ and τ are enriched in post-­‐crossing commissural axons. Arrows: pre-­‐
crossing commissural axon trajectories; filled arrowheads: post-­‐crossing commissural axon trajectories; open arrowheads: floorplate. (C) Dissociated commissural neurons were cultured for 2 or 3 DIV. Cells were lysed and levels of various 14-­‐3-­‐3 isoforms were detected by Western blot. (D) Densitometry was performed on Western blots stained for each 14-­‐3-­‐3 isoform and GAPDH. Protein levels were normalized to GAPDH before calculating the ratio of expression at 3 DIV to 2 DIV. Bar represents the mean of 3 independent experiments. (E) Summary of the 14-­‐3-­‐3 isoform expression patterns in vivo in spinal cord cross-­‐sections and in vitro in dissociated commissural neuron culture. Both 14-­‐3-­‐3β and γ are enriched in post-­‐crossing commissural neurons in vivo and increase in expression levels over time in vitro. Scale bar (A,B): 100 μm. FP: floorplate; VF: ventral funiculus; LF: lateral funiculus. Shh responsiveness, their expression should also change in vitro over time. We cultured dissociated commissural neurons for 2 or 3 DIV, and analyzed the levels of 14-­‐3-­‐3 isoforms in cell lysates by western blotting. 14-­‐3-­‐3β, γ and τ, all of which are expressed in post-­‐crossing commissural axons, all have higher expression at 3 DIV compared to 2 DIV (Figure 4C,D). Of the two isoforms that are predominantly expressed in the cell bodies, 14-­‐3-­‐3ε also has higher expression at 3 DIV compared to 2 DIV, but 14-­‐3-­‐3ζ does not (Figure 4C,D). Given that 14-­‐3-­‐3β and γ are strongly expressed in post-­‐
crossing commissural axons compared to pre-­‐crossing commissural axons, and that their expression increases in vitro with age in culture, they are good candidates for mediating the switch in Shh response from attraction to repulsion (Figure 4E). -­‐ 105 -­‐ Results 2.5.5 PKA activity decreases in post-­‐crossing commissural neurons and changes in PKA activity alter turning responses to SHH 14-­‐3-­‐3 proteins have been postulated to modulate growth cone turning through PKA by binding to the regulatory subunits of PKA and stabilizing the interaction between the regulatory and catalytic subunits of PKA, thereby reducing PKA activity (Kent, Shimada et al. 2010). Therefore an increase in 14-­‐3-­‐3 activity should lead to decrease in PKA activity. To assess the levels of active PKA, we used an antibody that recognizes the activated form of the catalytic subunit of PKA – phospho-­‐PKA. Western blotting of lysates from dissociated commissural neurons showed that the levels of phospho-­‐PKA at 3-­‐4 DIV decreased by about one-­‐third compared to the levels at 2 DIV (Figure 5A). PKA phosphorylates the PP-­‐1 inhibitory protein I-­‐1 in growth cones; thus, phospho-­‐I-­‐1 staining is another indicator of PKA activity (Han, Han et al. 2007). Consistent with the decrease in phospho-­‐PKA observed by Western blotting, phospho-­‐I-­‐1 staining in commissural neuron growth cones was also significantly lower at 3 DIV compared to 2 DIV (p=0.0158) (Figure 5B). Hence, the increase in 14-­‐3-­‐3 protein expression at 3 DIV correlated with a decrease in PKA activity. This is consistent with high levels of 14-­‐3-­‐3 proteins inactivating PKA, as has been shown for DRG neurons (Kent, Shimada et al. 2010). To determine if changing levels of PKA activity are sufficient to alter the turning response of commissural neurons to SHH, we pre-­‐treated 2DIV commissural neurons with inhibitors of PKA activity and then examined their response to gradients of SHH. Both the small molecule inhibitor KT-­‐5720 (-­‐
11.1±3.9°) and the inhibitory peptide PKI (-­‐19.6±5.1°) were sufficient to induce a switch to repellent turning responses compared to vehicle (12.7±4.3°) (Figure 5C). Conversely, when we selectively activated PKA in 3-­‐
4DIV commissural neurons, by pre-­‐treating with 6-­‐BNZ-­‐cAMP, we found a switch back to an attractive turning response to SHH gradients (9.7±4.8°) when compared to vehicle control (-­‐14.9±5.6°) (Figure 5C). Thus PKA activity levels are sufficient to determine the directional turning response of commissural neurons to SHH. -­‐ 106 -­‐ Results A
2 DIV
3-­4 DIV
2 DIV
3-­4 DIV
B
2 DIV
3 DIV
phospho-­I-­1
-­phospho
-­PKA
-­PKA
0.8
0.6
0.4
0.2
Mean angle turned ()
20
1.2
***
***
10
0
relative phospho-­I-­1 fluorescence
0.0
C
phalloidin
relative phospho-­PKA levels
1.0
1.0
0.8
0.6
0.4
0.2
0.0
-­10
p = 0.0158*
2 DIV
3 DIV
n=87
n=80
-­20
-­30
DMSO KT-­5720 PKI
2 DIV
PBS 6-­BNZ-­cAMP
Shh
3-­4 DIV
Figure 5: PKA activity in commissural neurons is time-­‐dependent and determines turning response to Shh (A) Commissural neurons were cultured for 2 or 3-­‐4 DIV. Cells were lysed and levels of phosphorylated PKA catalytic subunit and total PKA were detected by Western blot. Phospho-­‐PKA levels were normalized to total PKA and quantified from 2 experiments. Phospho-­‐PKA levels are lower at 3-­‐4 DIV compared to 2 DIV. (B) Commissural neurons were cultured for 2 or 3 DIV, then fixed and immunostained for phospho-­‐I-­‐1, a target of phospho-­‐PKA. Fluorescent phalloidin was used to label F-­‐actin. The average phospho-­‐I-­‐1 fluorescence signal was measured for each growth cone, and normalized to the mean growth cone fluorescence signal at 2 DIV. At least 25 growth cones were measured for each condition in 3 independent experiments. There is a significant decrease in phospho-­‐I-­‐1 fluorescence intensity at 3 DIV compared to 2 DIV (unpaired t-­‐test, p = 0.0158). (C) Mean turning angles for axons of commissural neurons cultured for 2 or 3-­‐4DIV in a 0.1µg/ml gradient of Shh. Neurons were pre-­‐treated for 1 hour before Dunn Chamber assembly with 200 nM KT-­‐5720, 20 µM myristolated PKI or 30µM 6-­‐Bnz-­‐cAMP. Altering PKA activity induces switches in turning responses to Shh (*** =p< 0.001, One way ANOVA). Scale bar (B): 10 μm. -­‐ 107 -­‐ Results 2.5.6 Inhibition of 14-­‐3-­‐3 proteins converts Shh repulsion to attraction through PKA activity We hypothesized that the increase in 14-­‐3-­‐3 protein levels may mediate the switch in Shh response from attraction to repulsion. To test this hypothesis, we inhibited 14-­‐3-­‐3 activity with R18 (PHCVPRDLSWLDLEANMCLP), a peptide antagonist which inhibits binding of all 14-­‐3-­‐3 isoforms to their Ser/Thr phosphorylated targets. The control WLKL peptide (WLDL to WLKL) does not bind to 14-­‐3-­‐3 because of an amino acid substitution in the core-­‐
binding motif. Both the R18 peptide and WLKL control peptide were fused to YFP and to Tat to allow entry into cells (Dong, Kang et al. 2008). Commissural neurons, which are normally repelled by Shh at 3 DIV (Figure 3A-­‐F), continue to do so in the presence of the control Tat-­‐WLKL-­‐YFP, with a mean angle turned of -­‐9.5±3.8° (Figure 6). Remarkably, in the presence of the inhibitory Tat-­‐R18-­‐YFP, 3 DIV commissural neurons were attracted by a Shh gradient, with a mean angle turned of 7.6±4.1° (Figure 6). There was a dramatic shift in the distribution of the angles turned from mostly negative in the presence of WLKL, to mostly positive when 14-­‐3-­‐3 proteins were inhibited by R18 (Figure 6A). In contrast, R18 had no effect on the net axon growth under the same conditions (Figure S2). In addition to interfering with 14-­‐3-­‐3 binding activity, we also examined the effect of knocking down levels of the specific 14-­‐3-­‐3 protein isoforms which show increased expression in post-­‐crossing commissurals, 14-­‐3-­‐3γ and β. Commissural neurons were electroporated at the time of plating with vectors expressing shRNA target sequences with a microRNA stem (shRNAmir) that have previously been shown to knockdown individual 14-­‐3-­‐3 isoforms (Kent, et al. 2010) or a scrambled control sequence. After 3-­‐4DIV, control sequence expressing cells exhibited a robust repellent turning response to Shh (-­‐19.8±6.1°), but neurons in which either 14-­‐3-­‐3 γ (14.2±4.5°) or 14-­‐3-­‐3 β (9.0±5.1°) had been knocked down were attracted to gradients of Shh (Figure 6C,D). Hence, 14-­‐3-­‐
3 activity is important for conferring the repulsive response to Shh at 3 DIV, -­‐ 108 -­‐ Results 3 DIV: Control WLKL
0.1 mg/ml Shh gradient
n=104
0°
-­30°
20
16
4
-­90°
8
30°
-­60°
60°
60°
24
18
12
12
6
90° -­90°
repulsion
n=86
0°
-­30°
30°
-­60°
B
3 DIV: R18
0.1 mg/ml Shh gradient
p = 0.0028**
15
Mean angle turned (°)
A
10
5
0
-­5
-­10
90°
attraction
-­15
WLKL
R18
0.1 mg/ml Shh gradient
C
3 DIV: Control scrambled shRNAmir
0.1 mg/ml Shh gradient
3 DIV: 14-­3-­3β shRNAmir
0.1 mg/ml Shh gradient
n=52
0°
-­60°
20
10
-­90°
repulsion
D
E
20
Mean angle turned (°)
Mean angle turned (°)
***
10
0
-­10
-­20
-­30
Ctl
-­90°
12
30°
-­30°
60°
20
-­60°
8
90°
12
16
60°
20
-­90°
90°
attraction
***
20
8
90°
16
n=61
0°
30°
-­30°
60°
30
-­60°
n=74
0°
30°
-­30°
3 DIV: 14-­3-­3γ shRNAmir
0.1 mg/ml Shh gradient
β
γ
14-­3-­3 shRNAmir
knockdown
n.s.
10
0
-­10
-­20
-­30
WLKL
R18
0.1 mg/ml Shh gradient
+KT-­5720
Figure 6: Inhibition o f 1 4-­‐3-­‐3 p rotein f unction c onverts S hh r epulsion t o attraction t hrough P KA a ctivity. (A) Rose histograms of the distribution of turned angles of commissural neurons at 3 DIV in a 0.1 μg/ml Shh gradient. Reponses of individual neurons were clustered in 15° bins and the percentage of total neurons per bin is represented by the radius of each segment. Blue segments represent repulsion and yellow represent attraction. Neurons were treated with 100 ng/ml of either Tat-­‐WLKL-­‐ YFP (control) or Tat-­‐R18-­‐YFP for 6 hours prior to placing the neurons in the Dunn chamber for the turning assay. Inhibition of 14-­‐3-­‐3 proteins with Tat-­‐R18-­‐
YFP switches the response to Shh from repulsion to attraction. (B) Mean angle turned (+/-­‐ SEM); p=0.0028, unpaired t test. (C) Rose histograms of the distribution of turned angles of commissural neurons at 3 DIV in a 0.1 μg/ml Shhgradient. Neurons were electroporated at the time of plating with plasmids -­‐ 109 -­‐ Results expressing shRNAmir targeting 14-­‐3-­‐3 γ,β or control. Knockdown of individual 14-­‐3-­‐3 isoforms switches the response to Shh from repulsion to attraction. (D) Mean angle turned (+/-­‐ SEM); ***=p<0.001, One way ANOVA. (E) Mean angle turned (+/-­‐ SEM) of neurons treated with 100 ng/ml of either Tat-­‐WLKL-­‐ YFP (control) or Tat-­‐R18-­‐YFP and the PKA inhibitor, KT-­‐5720, prior to placing the neurons in the Dunn chamber for the turning assay. Inhibition of PKA activity is sufficient to rescue the Tat-­‐R18-­‐YFP switch in the turning response to Shh. n.s.= not significant, One way ANOVA. and inhibition of 14-­‐3-­‐3 activity switches the response to Shh from repulsion to attraction. The attractive response when 14-­‐3-­‐3 levels are inhibited appears to be a similar state to commissural neurons at 2 DIV, when 14-­‐3-­‐3 levels are also lower and they are also attracted by Shh (Figure 9 ). As the level of PKA activity in commissural neuron growth cones is sufficient to determine the direction of turning response to gradients of Shh (Figure 5), we asked if loss of 14-­‐3-­‐3 function in 3-­‐4DIV neurons acts through PKA to switch the turning direction of growth cones. In order to determine if the switch in turning that results from blocking 14-­‐3-­‐3 binding is dependent on increasing PKA activity, we treated 3-­‐4DIV commissural neurons with either TAT-­‐R18-­‐YFP or TAT-­‐WLKL-­‐YFP, as well as KT-­‐5720, prior to subjecting them to turning assays. The positive turning angles that cells treated with TAT-­‐R18-­‐YFP show (Figure 6A,B) are completely blocked by the presence of the PKA inhibitor, KT-­‐5720. (Figure 6E). With KT-­‐5720, there is no significant difference between the mean turning angles of cells treated with TAT-­‐WLKL-­‐
YFP (-­‐11.6±5.5°), and TAT-­‐R18-­‐YFP (-­‐17.1±6.0°). Thus indicating that blocking 14-­‐3-­‐3 binding in 3-­‐4DIV commissural neurons depends on increased PKA activity to cause positive turning responses to gradients of Shh. -­‐ 110 -­‐ Results 2.5.7 Altering 14-­‐3-­‐3 proteins in vivo perturbs AP guidance of commissural axons Our in vitro experiments implicate the increase in 14-­‐3-­‐3 protein expression in the switch from attraction to repulsion of commissural neurons by Shh. To test whether 14-­‐3-­‐3 proteins are important in vivo for the repulsion of post-­‐
crossing commissural axons anteriorly along the longitudinal axis, we inhibited 14-­‐3-­‐3 activity in embryonic rat neural tubes. Neural tubes were dissected, cultured as open-­‐books, and Tat-­‐R18-­‐YFP or the control Tat-­‐
WLKL-­‐YFP added to the cultures. One day later, the cultures were fixed and the trajectories of post-­‐crossing commissural neurons visualized with DiI anterograde labeling. Post-­‐crossing axons in the presence of WLKL exhibited a stereotyped commissural axon trajectory, turning anteriorly after crossing the floorplate (Figure 7A). However, in the presence of R18, many axons failed to turn anteriorly upon floorplate exit. For each cohort, we scored whether they had randomization of AP guidance along the longitudinal axis, stalled axons in the floorplate, and recrossing axons. R18 clearly increased the degree of AP randomization (p=0.0041), but there was no significant effect on the amount of stalling or recrossing axons (Figure 7B). When we quantified the degree of AP randomization by measuring the fraction of anterior fluorescence, there was a significant decrease in the amount of anterior axons in the presence of R18 (Figure 7C). The decrease was in between that observed in the Wnt1-­‐Cre;Smon/c and Math1-­‐Cre;Smon/c mice. -­‐ 111 -­‐ A
Control WLKL
R18
B
A
40
p=0.0041**
R18
ns
ns
stalled
recrossing
20
10
0
C
random AP
1.00
D
100
p=0.002**
0.75
0.50
0.25
0.00
percentage of axons per embryo
WLKL
30
fraction anterior fluorescence
ant/(ant + post)
P
cohorts with phenotype (%)
Results WLKL
R18
***
anterior turn
***
posterior turn
80
stalled in FP
60
stalled after FP exit
40
20
0
WLRL-­EGFP
R18-­EGFP
Figure 7: Inhibition o f 1 4-­‐3-­‐3 p rotein f unction in v ivo p erturbs A P g uidance o f post-­‐crossing c ommissural a xons. (A) Top, DiI labeling of post-­‐crossing commissural axons in rat open-­‐book cultures treated with either 100 or 150 μg/ml Tat-­‐WLKL-­‐YFP or Tat-­‐R18-­‐YFP. Middle, zoom of the boxed region. Bottom, schematic of the axon guidance phenotype. Control Tat-­‐WLKL-­‐YFP-­‐treated axons make a stereotypical anterior turn after exiting the floorplate (12 open-­‐books, 64 cohorts). Tat-­‐R18-­‐YFP-­‐
treated axons show errors in turning after exiting the floorplate, with many axons turning posteriorly (arrows) (13 open-­‐books, 75 cohorts). -­‐ 112 -­‐ Results (B)Quantification of random AP turning, stalled axons in the floorplate, and recrossing axons, as percentage of cohorts (+/-­‐SEM) per open-­‐book (unpaired t-­‐
test). (C) Relative fluorescence (+/-­‐SEM) of the anterior directed axons versus total fluorescence of anterior and posterior directed axons. In WLKL-­‐treated axons (6 open-­‐books, 37 cohorts), most of the fluorescence is on the anterior side of the label, while in R18-­‐treated axons (7 open-­‐books, 43 cohorts), there is a significant decrease in the amount of anterior fluorescence. (p=0.002, unpaired t-­‐test). (D) Chick neural tubes were electroporated with either WLRL-­‐GFP or R18-­‐GFP plasmids and allowed to develop for 2 days in vivo. The behavior of individual axons was quantified. Expression of R18-­‐EGFP significantly decreased the number of axons which turn anteriorly and increased the number of axons which turn posteriorly after crossing the floorplate. n= 6 embryos per condition, minimum 199 axons per condition (two-­‐way ANOVA, ***=p<0.001). Scale bars: (A, Top) 50μm. (A, Middle) 20μm. Bracket (in (A)): floorplate. To inhibit 14-­‐3-­‐3 activity specifically in neurons, we turned to the developing chick embryo. Plasmids encoding R18 or a control WLRL peptide fused to EGFP (Wang, Yang et al. 1999; Kent, Shimada et al. 2010) were injected into the chick neural tube and electroporated unilaterally. Two days later, the embryos were dissected and the commissural axon trajectories analyzed in the open-­‐book format. Electroporated neurons were identified by EGFP expression. In neurons expressing control WLRL-­‐EGFP, the vast majority (92%) of axons correctly made an anterior turn after crossing the floorplate, with very few axons making a posterior turn. In contrast, neurons expressing R18-­‐EGFP had a significantly different distribution of phenotypes, with a significantly lower percentage of axons (58%) making the correct anterior turn (p<0.001), and a significantly higher percentage of axons (35%) turning posteriorly (p<0.001) (Figure 7D). We next asked if increasing the 14-­‐3-­‐3 isoforms (γ and β) which normally increase in post-­‐crossing commissural axons, prematurely in pre-­‐crossing commissural neurons would be sufficient to induce guidance defects in the AP axis. Our in vitro experiments, in which vectors expressing either 14-­‐3-­‐3 γ or β were transduced by HSV into 2DIV commissural neurons before -­‐ 113 -­‐ Results subjecting them to Shh gradients, showed that overexpression of 14-­‐3-­‐3 γ or β at an early stage was sufficient to induce a repellent response (-­‐11.8±4.9°) as compared to the attractive turning (11.3±4.2°) seen with a control vector (Figure 8A,B). When we electroporated the same expression vectors in developing chick embryos along with JX2-­‐GFP to label Math1+ neurons, we saw that the embryos expressing the control vector or 14-­‐3-­‐3ζ, exhibited the stereotypical projections to the midline, with 96% of axons turning anteriorly after floorplate crossing. However those embryos in which 14-­‐3-­‐3 γ or β were overexpressed, showed a significant percentage of axons(>21%) projecting anteriorly before crossing the midline at the floorplate indicating a premature repellent guidance response to the Shh gradient along the AP axis (Figure 8C,D). Thus increasing 14-­‐3-­‐3 γ or β in vivo is sufficient to induce a repellent turning response to Shh. Notably in both loss and gain of 14-­‐3-­‐3 function chick embryo experiments, as with the Tat-­‐R18-­‐YFP treated rat open-­‐book cultures (Figure 7A,B), there was no significant difference in the percentage of axons stalled in the floorplate or after floorplate exit, indicating that proper 14-­‐3-­‐3 activity is required for AP guidance, but not floorplate crossing and exit. This is different to what we observed in the Wnt1-­‐Cre;Smon/c and Math1-­‐
Cre;Smon/c mice, which have defects in both AP guidance and floorplate crossing and exit. In conclusion, inhibition of 14-­‐3-­‐3 activity in neurons disrupts the turning of post-­‐crossing commissural axons and increases in 14-­‐
3-­‐3 can induce improper AP turning in pre-­‐crossing commissural axons, with -­‐ 114 -­‐ Results A
2 DIV: Control
0.1 mg/ml Shh gradient
2 DIV: 14-­3-­3β overexpression
0.1 mg/ml Shh gradient
n=43
0°
-­30°
n=47
0°
-­30°
30°
-­60°
30
35
23
B
Mean angle turned (°)
25
C
8
12
16
90° -­90°
20
60°
90°
attraction
*
***
20
15
10
5
0
-­5
-­10
-­15
-­60°
60°
10
90° -­90°
γ
β
Ctl
14-­3-­3 overexpression
Control
D
percentage of axons per embryo
repulsion
30°
20
12
-­90°
n=70
0°
-­30°
30°
-­60°
60°
2 DIV: 14-­3-­3γ overexpression
0.1 mg/ml Shh gradient
100
***
80
anterior turn post-­crossing
***
anterior turn pre-­crossing
stalled in floorplate
60
40
20
0
Control
***
***
β
γ
ζ
14-­3-­3 overexpression
14-­3-­3β overexpression
14-­3-­3γ overexpression
14-­3-­3ζ overexpression
A
P
Figure 8: O verexpression o f 1 4-­‐3-­‐3 γ o r β in p re-­‐crossing c ommissural a xons induces S hh r epulsion a nd p erturbs A P g uidance i n v ivo. (A) Rose histograms of the distribution of turned angles of commissural neurons at 2 DIV in a 0.1 μg/ml Shh gradient. Neurons were transduced with HSV expressing either 14-­‐3-­‐3 γ,β or control. Overexpression of these 14-­‐3-­‐3 isoforms induces a repellent response to Shh. (B) Mean angle turned (+/-­‐ SEM); *=p<0.05,***=p<0.001, One way ANOVA. (C,D) Chick neural tubes were electroporated with either control, 14-­‐3-­‐3β,γ,or ζ plasmids and allowed to develop for 2 days in vivo. The behavior of individual axons was quantified. -­‐ 115 -­‐ Results Expression of 14-­‐3-­‐3γ,or β significantly increased the number of axons which turn anteriorly before crossing the floorplate. (arrows) n=2 embryos per condition per experiment, 3 independent experiments, minimum 136 axons per condition (two-­‐way ANOVA, ***=p<0.001). no effect on floorplate crossing and exit. Together with our data that demonstrates that 14-­‐3-­‐3 activity is necessary and sufficient for the switch in Shh response from attraction to repulsion, this implicates 14-­‐3-­‐3 proteins in mediating the repulsive response to Shh in post-­‐crossing commissural axons. 2.6 DISCUSSION Our data supports a model where Shh acts as a bifunctional guidance cue, attracting commissural axons towards the floorplate and then repelling them anteriorly along the AP axis (Figure 9A). Furthermore, dissociated commissural neurons in vitro, in the absence of other cell types, can switch from being attracted to being repelled by Shh (Figure 9B). This recapitulates the change in response to Shh between pre-­‐ and post-­‐crossing commissural axons in vivo, suggesting that the switch in response is intrinsic, cell autonomous and time-­‐dependent. We found that 14-­‐3-­‐3 proteins are required for this intrinsic switch. In vivo, 14-­‐3-­‐3 proteins are enriched in post-­‐crossing commissural axons, and in vitro their expression levels increase with time in culture. Antagonizing or knocking down 14-­‐3-­‐3 proteins in vitro switched the response to Shh from repulsion to attraction, and inhibition of 14-­‐3-­‐3 protein function in vivo disrupts post-­‐crossing commissural axon guidance. Conversely, overexpressing individual 14-­‐3-­‐3 -­‐ 116 -­‐ Results Figure 9: D ifferential p re-­‐ a nd p ost-­‐crossing c ommissural a xon r esponses t o Sonic h edgehog a re r egulated b y a c ell-­‐intrinsic 1 4-­‐3-­‐3 p rotein d ependent switch. (A) Shh is a bifunctional guidance cue for commissural axons. (left) During pre-­‐
crossing commissural axon guidance, commissural axons are attracted by Shh. The dorsal-­‐high repellent gradient of BMPs (red) and the ventral-­‐high attractant gradient of Netrin-­‐1, Shh and VEGF (blue) act together to guide pre-­‐crossing commissural axons ventrally towards the floorplate. (right) After reaching and crossing the floorplate, axon behavior switches and post-­‐crossing commissural axons become repelled by Shh. Axons are guided anteriorly by the posterior-­‐high repellent gradient of Shh (blue) and an anterior-­‐high attractant gradient of Wnt4 (yellow). This switch in the polarity of the Shh response depends on 14-­‐3-­‐3 levels, which are low in pre-­‐crossing commissural axons and high in post-­‐
crossing commissural axons. (B) The switch in Shh response is recapitulated in vitro with dissociated commissural neuron cultures. At 2 DIV, commissural neurons are attracted towards high concentrations of Shh. At 3-­‐4 DIV, commissural neurons are repelled by high concentrations of Shh. This switch also correlates with an increase in 14-­‐3-­‐3 protein levels. -­‐ 117 -­‐ Results family members in vitro induced an early switch to a repellent response to Shh. In vivo, 14-­‐3-­‐3 overexpression resulted in axons turning prematurely along the anterior axis of the developing spinal cord. This suggests that 14-­‐3-­‐
3 proteins are required for repulsion of post-­‐crossing commissural neurons by Shh, and that they play an important role in establishing the switch in response of commissural neurons to Shh. 2.6.1 Shh has multiple roles in commissural axon guidance. Shh secreted from the floorplate has multiple roles in nervous system development, from cell fate specification to axon guidance. The DV gradient of Shh, together with BMPs from the roofplate, initially specifies the identity of various neuron types in the spinal cord (Dessaud, McMahon et al. 2008). Subsequently, the DV Shh gradient, together with Netrin-­‐1 and VEGF, is re-­‐
used to guide pre-­‐crossing commissural axons to the floorplate. Upon reaching the floorplate, Shh induces the response to Semaphorins, which allows for the correct exit of commissural axons from the floorplate (Parra and Zou 2010). We now show that Shh also has a direct effect on post-­‐
crossing commissural axons in mammals, guiding them along the AP axis, consistent with evidence from chick (Bourikas, Pekarik et al. 2005). We further show that the switch to Shh repulsion along the AP axis depends on 14-­‐3-­‐3 proteins and a cell autonomous timer. The multiple roles of Shh at the floorplate are reflected in the phenotypes observed when Shh function or signaling is disrupted. In both chick and mouse, perturbation of Shh function results in defects in floorplate crossing and exit and defects in turning along the AP axis (Figure 1), (Bourikas, Pekarik et al. 2005; Parra and Zou 2010). Thus it is reasonable to propose that these phenotypes reflect the dual role of Shh at the floorplate for crossing commissural axons – both i) to induce Semaphorin repulsion of commissural axons at the floorplate for correct floorplate exit and ii) to guide post-­‐crossing commissural axons anteriorly along the longitudinal axis. -­‐ 118 -­‐ Results Intriguingly, inhibition of 14-­‐3-­‐3 function in rat and chick affected only AP guidance of post-­‐crossing commissural neurons, but not floorplate crossing and exit (Figure 7). This implies that 14-­‐3-­‐3 proteins are specifically involved in AP guidance of post-­‐crossing commissural neurons by Shh, but not in the induction of Semaphorin repulsion by Shh. This highlights that these two functions of Shh at the floorplate are distinguishable and act through different mechanisms. Our conditional mutants show that Smo is required for the AP guidance of post-­‐crossing commissural axons. This is consistent with independent experiments showing that downregulation of Smo in rat open-­‐book explant cultures also leads to AP guidance defects (Parra and Zou 2010). Importantly, our genetic experiments selectively inactivate Smo in commissural neurons (Figure 1), convincingly demonstrating a cell autonomous requirement for Smo in the AP guidance of commissural neurons. Both our results and those of Parra and Zou (Parra and Zou 2010) contrast with those obtained in chick with in ovo RNAi, which suggest that the guidance of post-­‐crossing axons by Shh is independent of Smo (Bourikas, Pekarik et al. 2005). Additionally, in chick, the receptor mediating the effect of Shh on post-­‐crossing axons has been proposed to be Hhip1 (Bourikas, Pekarik et al. 2005). We failed to find a role for Hhip1 in the AP guidance of commissural axons through genetic analysis in the mouse, with post-­‐crossing commissural axons turning anteriorly in Hhip1-­‐/-­‐ mice (Figure S3). Although it is possible that Shh-­‐
mediated AP axon guidance in the chick and mammals use different molecular mechanisms, this would be somewhat surprising given that all of the other guidance effects described so far for Shh are mediated in a Smo-­‐
dependent manner (Charron, Stein et al. 2003; Sanchez-­‐Camacho and Bovolenta 2008; Yam, Langlois et al. 2009; Fabre, Shimogori et al. 2010). -­‐ 119 -­‐ Results 2.6.2 A novel role for 14-­‐3-­‐3 proteins in switching the polarity of the turning response to Shh The ability of axons to change responsiveness to guidance cues is critical for axons to properly navigate through complex environments. We show that the switch in Shh response from attraction to repulsion depends on 14-­‐3-­‐3 proteins. 14-­‐3-­‐3 proteins are important for conferring repulsive responses to NGF in postnatal rat DRGs, and antagonism of 14-­‐3-­‐3 proteins converts this NGF-­‐mediated repulsion to attraction (Kent, Shimada et al. 2010). Work in Drosophila has shown 14-­‐3-­‐3ε to be critical for proper axon guidance in vivo. 14-­‐3-­‐3 ε acts to mediate highly localized signalling events, allowing for rapid silencing of Sema1A repulsion and the re-­‐establishment of integrin mediated adhesion (Yang and Terman 2012). We now demonstrate a novel role for 14-­‐
3-­‐3 proteins in establishing a temporal developmental switch in the polarity of turning response to a single guidance cue. 14-­‐3-­‐3 proteins function as homodimers and heterodimers to control the spatial and temporal activity of substrate proteins (Muslin and Xing 2000; Bridges and Moorhead 2005). One way through which 14-­‐3-­‐3 proteins act to modulate growth cone turning is by inhibiting PKA activity. 14-­‐3-­‐3 proteins bind the regulatory subunits of PKA and stabilize the PKA holoenzyme (Kent, Shimada et al. 2010). Consistent with 14-­‐3-­‐3 proteins modulating growth cone turning through PKA, we also observed that the increase in 14-­‐3-­‐3 levels in 3-­‐4 DIV commissural neurons was accompanied by a decrease in active PKA levels. If this decrease is antagonized by activating PKA in 3-­‐4DIV commissural neurons, the switch to repulsion fails to occur. Furthermore, mimicking this decrease with pharmacological inhibition of PKA is also able to switch turning responses to Shh from attraction to repulsion. In addition, pharmacological inhibition of PKA is sufficient to block the switch to positive turning angles induced by blocking 14-­‐3-­‐3 function. Our data indicate that 14-­‐3-­‐3s act to determine the direction of growth cone turning mainly through regulating PKA activity but do not exclude the possibility that 14-­‐3-­‐3 -­‐ 120 -­‐ Results proteins may also act in a PKA-­‐independent manner to mediate the switch in response to Shh. 14-­‐3-­‐3 proteins have a large number of binding partners, many of which regulate the cytoskeleton. For example, 14-­‐3-­‐3 proteins also affect the activity of myosin light chain phosphatase (Koga and Ikebe 2008), Rho GEFs (Meiri, Greeve et al. 2009), and ADF/cofilin (Gohla and Bokoch 2002; Yoon, Zivraj et al. 2011). In rat DRGs, knockdown of 14-­‐3-­‐3 β, γ or ε converts NGF-­‐dependent repulsion to attraction (Kent, Shimada et al. 2010), demonstrating that these isoforms are important for mediating the repulsive response. Interestingly, our data implicates an overlapping set of isoforms, β and γ, in the switch in response to Shh, as they are enriched in post-­‐crossing commissural axons and their expression increases over time in vitro. Overexpression of these isoforms, but not 14-­‐3-­‐3ζ, in vivo, induces premature anterior turning and knockdown of either of these isoforms alone was sufficient to reverse the switch in turning response to Shh. Furthermore, γ, and to a lesser extent, β, have been shown to interact with the regulatory subunit of PKA (Kent, Shimada et al. 2010). Given that 14-­‐3-­‐3 isoforms can form heterodimers, it is possible that more than one isoform, in different heterodimeric combinations, may be involved in mediating the switch in response to Shh. 2.6.3 Molecular mechanisms involved in modulating responses to guidance cues In vitro, changes in the relative levels of other intracellular molecules, such as cyclic nucleotides and Ca2+ can switch responses to guidance cues (Ming, Song et al. 1997; Song, Ming et al. 1997; Song, Ming et al. 1998; Wen, Guirland et al. 2004). In vivo, cAMP is involved in the switch from Netrin1-­‐mediated attraction to repulsion of retinal ganglion cell axons (Shewan, Dwivedy et al. 2002), as well as silencing Sema1A repulsion in Drospohila motor axons through 14-­‐3-­‐3ε (Yang and Terman 2012). However, we have not detected -­‐ 121 -­‐ Results any differences in cAMP levels between commissural neuron growth cones at 2 and 4 DIV (Figure S4). Alternatively, differential guidance responses can also result from differential expression patterns of receptors, as has been demonstrated for Sema3E, Netrin and Slit (Hong, Hinck et al. 1999; Shewan, Dwivedy et al. 2002; Chauvet, Cohen et al. 2007; Chen, Gore et al. 2008). For example, in the forebrain, neurons expressing PlexinD1 are repelled by Sema3E whilst neurons expressing PlexinD1 and Neuropilin1 are attracted (Chauvet, Cohen et al. 2007). In Xenopus axons in vitro, expression of Unc5 converts Netrin-­‐
mediated, DCC-­‐dependent attraction to repulsion (Hong, Hinck et al. 1999). Axon turning responses can also be modified by extrinsic signals such as extracellular matrix components or other guidance cues. Laminin can switch Netrin attraction to repulsion in retinal ganglion cell axons (Hopker, Shewan et al. 1999). The presence of heparin sulfate proteoglycans confers attraction to Sema5A for axons in the fasciculus retroflexus, whereas in the presence of chondroitin sulfate proteoglycans Sema5A is repulsive (Kantor, Chivatakarn et al. 2004). Moreover, at the floorplate, activation of Robo by Slit silences the attractive effect of Netrin-­‐1 on commissural axons (Stein and Tessier-­‐Lavigne 2001) whilst neuronal cell adhesion molecule (NrCAM) triggers a gain of response to Sema3B (Nawabi, Briancon-­‐Marjollet et al. 2010). Commissural neurons also gain responsiveness to class 3 Semaphorins through exposure to Shh (Parra and Zou 2010). However, in our case, the switch in response to Shh is intrinsic and occurs in the absence of extrinsic cues. The switch we observed can be recapitulated in vitro with dissociated commissural neuron cultures in the absence of Shh and other cell types. This temporal switch between attraction and repulsion is reminiscent of the switch in responsiveness to Netrin-­‐1 that has been observed in retinal explant cultures (Shewan, Dwivedy et al. 2002), the only -­‐ 122 -­‐ Results prior example of a time-­‐dependent switch in the polarity of a response to a guidance cue. In the retinal explant cultures (Shewan, Dwivedy et al. 2002), it is possible that extrinsic factors may be involved since there is contact with neighboring cells and different cell types present in the explants. As our commissural neuron cultures are almost pure (90-­‐98%), (Yam, Langlois et al. 2009) and cultured at very low density, it is unlikely that the switch we observe is triggered by other cell types, as has been documented for the eventual switch of retinal ganglion cells to a non-­‐growth state (Goldberg, Klassen et al. 2002). Furthermore, we isolate commissural neurons from the dorsal fifth of the spinal cord, at an age where axons that have extended from these neurons have not yet reached the floorplate. Hence the neurons we use for our in vitro experiments are floorplate-­‐naïve. Our work reinforces the idea proposed by Holt and colleagues that time-­‐dependent switches can regulate responses to axon guidance cues during development and in addition illustrates that these time-­‐dependent switches can be intrinsic. A cell intrinsic switch that changes the response of neurons to midline cues adds another level of regulation to the switching of cellular responses at the midline. Crossing the floorplate is not required for commissural axons to become sensitive to the longitudinal gradients of Wnt4 and Shh, since commissural axons in Robo3 mutants make the correct anterior turn without crossing the floorplate (Chen, Gore et al. 2008). Although the floorplate derived cues Wnt4 and Shh are required to guide the anterior projection of axons, an intrinsic temporal switch may be involved in sensitizing the axons to these cues. Extrinsic factors in the spinal cord would add an additional level of regulation to modulate and fine-­‐tune the guidance program. In vivo, extrinsic spatial and intrinsic temporal regulation might act together to confer highly localized control to signalling mechanisms, allowing for the switch in commissural axon trajectory from DV to AP at the floorplate, and ensuring high fidelity in axon turning at this intermediate target. -­‐ 123 -­‐ Results 2.7 ACKNOWLEDGEMENTS We thank E. Ruthazer for critical reading of the manuscript. We are grateful to K.K. Murai for access to his spinning-­‐disc confocal microscope. We thank J. Barthe, J. Cardin, A. Daigneault, S. D. Langlois and T. Shimada for expert assistance. We thank D. Rowitch for Math1-­‐Cre mice and P.T. Chuang for Hhip1 mice. The 4D7 and 5E1 antibodies were obtained from the Developmental Studies Hybridoma Bank developed under the auspices of the NICHD and maintained by The University of Iowa. This work was supported by grants from the Canadian Institutes of Health Research (CIHR), the Peter Lougheed Medical Research Foundation, the McGill Program in Neuroengineering, the Fonds de Recherche en Santé du Québec (FRSQ), and the Canada Foundation for Innovation (CFI). FC is a FRSQ Chercheur-­‐
Boursier. -­‐ 124 -­‐ Results 2.8 SUPPLEMENTARY I NFORMATION Figure 1S: Math1+ neurons are commissural neurons. (A) Position of Math1+ neuron cell bodies as visualized by b-­‐galactosidase staining in Math1-­‐LacZ spinal cord sections. (B-­‐D) Retrograde DiI labeling in Math1-­‐GFP spinal cords. (B) Overview of the location of Math1+ nuclei and the contralateral DiI site (arrow). (C) Close-­‐up of box in (B) showing retrograde labeling to the Math1+ region of the spinal cord. (D) Close up of box in (C). DiI from the contralateral DiI site retrogradely labels the cell bodies containing GFP+ nuclei (arrows) showing that Math1+ neurons are commissural. -­‐ 125 -­‐ Results Figure 2S: Inhibition o f 1 4-­‐3-­‐3 f unction w ith T at-­‐R18-­‐YFP h as n o e ffect o n axon g rowth. Commissural neurons at 3 DIV were treated with 100 μg/ml of either Tat-­‐WLKL-­‐
YFP (control) or Tat-­‐R18-­‐YFP for 6 hours prior to placing the neurons in the Dunn chamber for the turning assay. N et axon growth (+/-­‐ SEM) of commissural neurons in a 0.1 mg/ml Shh gradient was the same in the presence of either Tat-­‐
WLKL-­‐YFP or Tat-­‐R18-­‐YFP (p=0.2290, unpaired t-­‐test). -­‐ 126 -­‐ Results Figure 3S: H hip1 d oes n ot p lay a m ajor r ole in c ommissural a xon g uidance. (A, B) Open-­‐book DiI labeling of commissural axons in E11.5 WT and Hhip1-­‐/-­‐ neural tubes. (A) Control axons make a stereotypical anterior turn after exiting the floor plate. (B) Post-­‐crossing commissural axons in Hhip1-­‐/-­‐ embryos project anteriorly in the same manner as axons in WT embryos, indicating that Hhip1 is not required for AP post-­‐crossing axon guidance (5 embryos, 23/23 cohorts indistinguishable from WT). (C) Tag1 immunostaining of pre-­‐crossing commissural axons in a WT E11.5 spinal cord cross-­‐section showing the normal commissural axon trajectory projecting towards the floorplate. (D) Tag1 immunostaining of Hhip1-­‐/-­‐ E11.5 spinal cord cross-­‐sections shows that the commissural axon trajectories in Hhip1-­‐/-­‐ embryos resemble the trajectories in WT embryos, indicating that Hhip1 is not a receptor for pre-­‐crossing commissural axon guidance in the mouse. Brackets: width of floorplate in (A) and (B). Arrowhead: floorplate in (C) and (D). -­‐ 127 -­‐ Results Figure 4S: c AMP levels a re s imilar in c ommissural n euron g rowth c ones a t 2 and 4 D IV Commissural neurons were cultured for 2 and 4 DIV and growth cone cAMP levels were measured by immunofluorescence, then normalized to the mean fluorescence intensity at 2 DIV. Similar immunofluorescence intensity was detected in the growth cones of commissural neurons at 2 DIV (n=37) and 4 DIV (n=33) (p=0.1158, unpaired t-­‐test). DIV, days in vitro. Error bars are standard deviation. -­‐ 128 -­‐ Results 3. 14-­‐3-­‐3 proteins regulate commissural neuron responses t o N etrin-­‐1 Christopher B. Kent, Ricardo Alchini, Timothy Kennedy and Alyson E. Fournier* Department of Neurology and Neurosurgery, Montreal Neurological Institute and Hospital, McGill University, Montreal, Quebec H3A 2B4, Canada *Corresponding Author: Alyson E. Fournier -­‐ 129 -­‐ Results 3.1 PREFACE The regulation of PKA activity in axonal growth cones by 14-­‐3-­‐3 proteins is required for establishing switches in the directional response to gradients of external ligands. This process is critical for establishing the proper axonal trajectories of commissural interneurons in the developing spinal cord and is mediated by select 14-­‐3-­‐3 isoforms in a non-­‐redundant manner. The high expression of other 14-­‐3-­‐3 isoforms in commissural neurons, prior to their crossing the midline, raises the possibility that other roles exist for 14-­‐3-­‐3 proteins in regulating the signalling pathways that guide the direction of axonal growth. The present study establishes the requirement of 14-­‐3-­‐3 activity for Netrin-­‐1 effects on pre-­‐crossing commissural neurons and examines the interaction of 14-­‐3-­‐3 proteins with the Netrin-­‐1 receptor DCC and its downstream effector proteins. -­‐ 130 -­‐ Results 3.2 ABSTRACT The ability of neurons to properly extend their axons during development or regeneration to appropriate targets depends on the precise spatial and temporal regulation of multiple signalling pathways. 14-­‐3-­‐3 adaptor proteins play a critical role in regulating axon guidance signalling pathways. We previously demonstrated that 14-­‐3-­‐3 antagonism switches repellent growth cone responses to attraction in mature DRG neurons through regulation of protein kinase A. Also we found that 14-­‐3-­‐3 regulation of PKA is required to establish the proper axonal trajectories of commissural interneurons after they cross the midline of the developing spinal cord. Here, we report that 14-­‐
3-­‐3s co-­‐immunoprecipitate with the Netrin receptor Deleted in Colorectal Cancer (DCC) and are required for attractive responses to Netrin-­‐1. We find that pre-­‐crossing embryonic commissural neurons express several 14-­‐3-­‐3 isoforms and that loss of 14-­‐3-­‐3 activity blocks Netrin-­‐1 induced changes to the growth cone structure, attractive turning responses and enhanced neurite outgrowth. Cell surface biotinylation analysis reveals that DCC trafficking to the cell surface is normal in commissural neurons when 14-­‐3-­‐3 activity is blocked. Intriguingly, 14-­‐3-­‐3 function is required for the proper netrin-­‐dependent phosphorylation and activation of downstream signalling effectors such as extracellular-­‐signal-­‐regulated kinase 1/2 (ERK1/2). Thus we identify a novel role for 14-­‐3-­‐3 proteins in mediating DCC signalling and proper attractive axon guidance in pre-­‐crossing commissural neurons. -­‐ 131 -­‐ Results 3.3 INTRODUCTION In order for neurons to extend their axons during development or regeneration to the appropriate targets they must navigate complex environments with multiple extracellular cues. Many cues function both as attractive and repellent signals and the appropriate response by the growth cone depends on the precise spatial and temporal regulation of multiple signalling pathways (Vitriol and Zheng 2012). The family of 14-­‐3-­‐3 adaptor proteins bind pSer/pThr motifs in a diverse array of targets, conferring temporal and spatial regulation to multiple cellular signalling events (Muslin and Xing 2000; Bridges and Moorhead 2005; van Heusden 2005). They play a number of critical roles in regulating axon guidance signalling pathways. 14-­‐
3-­‐3 isoforms interact with important effector proteins that modulate cytoskeletal changes involved in axon outgrowth (Yoon, Zivraj et al. 2011). They can spatially regulate axon guidance signalling by restricting interactions between guidance cue receptors and downstream signalling proteins such as the small GTPase, Ras (Yang and Terman 2012). As well, temporal regulation of PKA activity in growth cones by 14-­‐3-­‐3 proteins is required for establishing switches from attractive to repellent turning responses to different guidance cues (Kent, Shimada et al. 2010), (Chapter 2). 14-­‐3-­‐3 regulation of axon guidance through regulation of PKA, is critical for establishing the proper axonal trajectories of commissural interneurons in the developing spinal cord (Ch. 2). Commissural interneurons in the dorsal part of the spinal cord initially extend their axons ventrally towards the midline, responding in an attractive manner to gradients of the floorplate derived cues: Netrin-­‐1(Serafini, Kennedy et al. 1994), Shh (Charron, Stein et al. 2003) and vascular endothelial growth factor (VEGF) (Ruiz de Almodovar, Fabre et al. 2011). As axons reach the midline, an intrinsic, cell autonomous switch occurs in the response to gradients of Shh. The initial attractive response becomes repellent ensuring a proper turn down the AP axis of the -­‐ 132 -­‐ Results developing spinal cord. This switch is mediated by 14-­‐3-­‐3 β and γ isoforms in a non-­‐redundant manner (Ch. 2). The high expression of other 14-­‐3-­‐3 isoforms in commissural neurons, prior to their crossing the midline, raises the possibility that other roles exist for 14-­‐3-­‐3 proteins in regulating the signalling pathways that guide the direction of axonal growth. The attractive response of pre-­‐crossing commissural neurons to Netrin-­‐1 is mediated by the receptor, DCC (Keino-­‐Masu, Masu et al. 1996). Netrin-­‐1 signalling through DCC has been shown to involve the regulation of multiple downstream effectors, including kinases (Forcet, Stein et al. 2002; Li, Lee et al. 2004; Shekarabi, Moore et al. 2005), GEFs (Briancon-­‐Marjollet, Ghogha et al. 2008) and Rho family GTPases (Gitai, Yu et al. 2003; Shekarabi, Moore et al. 2005; Moore, Correia et al. 2008). Further, trafficking of DCC to the cell surface and to appropriate microdomains within the plasmia membrane is regulated and important for regulating the response to Netrin-­‐1 (Bouchard, Moore et al. 2004; Guirland, Suzuki et al. 2004; Petrie, Zhao et al. 2009; Bai, Chivatakarn et al. 2011). The manner in which this occurs is not completely understood. Here we examine the role of 14-­‐3-­‐3 proteins in regulating Netrin-­‐1 signalling through DCC. We examined the requirement of 14-­‐3-­‐3 activity for Netrin-­‐1 signalling in pre-­‐crossing commissural neurons in vitro through loss of function studies and found that loss of 14-­‐3-­‐3 function blocks Netrin-­‐1 dependent growth cone expansion. Further, 14-­‐3-­‐3 function is required for the attractive turning response of pre-­‐crossing commissural neurons to gradients of Netrin-­‐1 and loss of 14-­‐3-­‐3 function blocks Netrin-­‐1-­‐dependent neurite outgrowth out of embryonic dorsal spinal cord explants into a collagen matrix. We find that the DCC co-­‐immunoprecipitates with 14-­‐3-­‐3 proteins, and that 14-­‐3-­‐3 binding is required for the proper phosphorylation and activation of the downstream signalling effector, ERK1/2. Thus we identify a -­‐ 133 -­‐ Results novel role for 14-­‐3-­‐3 proteins in mediating DCC signalling and proper attractive axon guidance in pre-­‐crossing commissural neurons. 3.4 MATERIALS A ND M ETHODS 3.4.1 Reagents Commercial antibodies were acquired for full length DCC from BD Biosciences (San Jose, CA), total and phosphorylated ERK 1/2 from Cell Signalling (Beverly, MA), and pan-­‐14-­‐3-­‐3 from Santa Cruz Biotechnologies (Santa Cruz, CA). Tat-­‐R18-­‐YFP and Tat-­‐WLKL-­‐YFP vectors (Dong, Kang et al. 2008) were generously provided by Dr. Jing Chen (Emory University). The pCEP4 plasmid expressing full-­‐length rat DCC and pCS2 and pHSVPrPUC plasmids expressing GFP-­‐R18 and GFP-­‐WLRL were previously described (Shekarabi and Kennedy 2002; Kent, Shimada et al. 2010). 3.4.2 Virus and recombinant proteins Recombinant Herpes Simplex Virus was produced by transfecting pHSVPrPUC plasmids into 2-­‐2 Vero cells that were superinfected with 5dl 1.2 herpes simplex virus (HSV) helper virus 1 d later. Recombinant virus was amplified through three passages and stored at –80°C as described previously(Neve, Howe et al. 1997). Soluble Recombinant netrin-­‐1 proteins were purified from a HEK293T cell lines secreting netrin-­‐1 as previously described (Serafini, Kennedy et al. 1994; Shirasaki, Mirzayan et al. 1996). TAT fusion peptides were purified as previously described (Dong, Kang et al. 2008). In brief, peptides were purified by sonication of high expressing BL21(DE3)pLysS cells obtained from culture with IPTG-­‐induction. Cellular lysates were resolved by centrifugation and loaded onto a ProBond Ni-­‐
column (Invitrogen, Carlsbad, Ca.) in 20 mm imidazole. After washing, the protein was eluted with 250 mm imidazole. TAT fusion proteins were desalted on a PD-­‐10 column (GE Healthcare, Buckinghamshire, UK). -­‐ 134 -­‐ Results 3.4.3 Dissociated neuron culture Staged pregnant Sprague Dawley rats were obtained from Charles River (St. Constant, Canada). Dissociated commissural neuron cultures were prepared from the dorsal fifth of E13 rat neural tubes as previously described (Yam, Langlois et al. 2009). They were plated in Neurobasal (Invitrogen, Carlsbad, Ca.) supplemented with 10% heat-­‐inactivated FBS and 2 mM L-­‐glutamine. After ~21 h, the medium was changed to Neurobasal supplemented with 2% B27 and 2 mM L-­‐glutamine. HSV infections were performed ~24hrs after plating Dissociated commissural neuron cultures were used for experiments at 2 DIV (50-­‐55 h after plating). For low-­‐density cultures, commissural neurons were plated at 90-­‐120 000 cells/well for a 6-­‐well plate. 3.4.4 Dunn chamber turning a ssays The Dunn chamber axon guidance assay was performed as previously detailed (Yam, Langlois et al. 2009). Briefly, commissural neurons were dissociated and infected with HSV-­‐GFP WLRL or HSV-­‐GFP R18, grown on poly-­‐L-­‐lysine (20 μg/ml) coated square #3D coverslips (Fischer Scientific, Waltham, MA) at low density, and then assembled in a Dunn chamber after specified times in culture. Gradients were generated in the Dunn chamber with 0.2 μg/ml recombinant VI,V fragment of Netrin-­‐1 in the outer well. After Dunn chamber assembly, time-­‐lapse phase contrast images were acquired for a minimum of 2 h. The angle turned was defined as the angle between the original direction of the axon and a straight line connecting the base of the growth cone from the first to the last time point of the assay period. 3.4.5 Collagen explant a ssays E13 rat spinal cords were dissected and the dorsal fifth was mechanically separated to produce approximately 0.1mm2 explants. Explants were embedded in a Type I rat tail collagen (BD Biosciences) matrix and cultured for 19hrs in Neurobasal medium (Invitrogen) supplemented with 10% FBS, Pen/Strep and Glutamine. Culture media also included 200 ng/ml of Netrin-­‐1 or vehicle control and 100µg/ml of purified Tat R18 or WLRL peptides, -­‐ 135 -­‐ Results which were present throughout the experiment. DIC photomicrographs were taken using a 10x objective on an Axiovert microscope (Zeiss, Toronto,ON) and analysed using the NeuronJ plugin for ImageJ. Total axon length in the focal plane was measured per explant and was then normalized by the area of the explant. Statistical comparison was done with Two way ANOVA, with Bonferroni post-­‐test. 3.4.6 Cell surface biotinylation E13 rat dorsal spinal cords were dissociated, and commissural neurons were plated and cultured for 2DIV on poly-­‐L-­‐lysine (PLL) coated 60mm plates. Cells were infected overnight with HSV-­‐GFP WLRL or HSV-­‐GFP R18 and the next treated with 200ng/ml netrin-­‐1 or vehicle for 10 min. Cells were then washed with ice-­‐cold PBS containing 0.1 mM calcium chloride and 1 mM magnesium chloride, pH 7.4. Surface biotinylation was performed by adding EZ-­‐Link Sulfo-­‐NHS-­‐LC-­‐biotin (Pierce, Waltham, MA), 2 ml per plate at 0.5 mg/ml in PBS at 4°C for 30 min, removed, and the reaction was quenched by the addition of 10mM ice-­‐cold glycine in PBS at 4°C for two 10 min periods. Subsequently, cells were washed with ice-­‐cold PBS and lysed with RIPA buffer. Biotinylated proteins were precipitated with streptavidin–agarose (Pierce) and analyzed by Western blot. 3.4.7 Immunoprecipitation HEK 293T Cells were transfected with plasmid expressing full-­‐length DCC for 5 hours with Lipofectamine 2000 and then serum starved for 16 hrs. Cells were treated with 200ng/ml Netrin-­‐1 for 10 min and then were lysed in NP-­‐
40 buffer (150 mM NaCl, 1% Nonidet P-­‐40, 10 mM Tris-­‐Cl, pH 8.0, 10% (w/v) glycerol) with complete EDTA free protease inhibitor cocktail (Roche) and phosphatase inhibitors (5mM NaF and 1mM Na3VO4). Following sonication, lysates were pre-­‐cleared with Protein A/G agarose beads (SantaCruz) and then incubated with anti-­‐pan 14-­‐3-­‐3 antibody (Santa Cruz). Precipitates were washed with NP-­‐40 buffer and eluted in 2× SDS sample buffer. -­‐ 136 -­‐ Results 3.4.8 Western blotting HEK 293T Cells were transfected for 5 hours with Lipofectamine 2000 and then serum starved for 16 hrs. Cells were treated with 200ng/ml Netrin-­‐1 for 10 min and then were lysed in RIPA buffer (50 mM HEPES pH 7.4, 150 mM NaCl, 10% glycerol, 1.5 mM MgCl2, 1% Triton, 1% SDS, 1 mM EDTA) and boiled in SDS sample buffer for 5 min. Protein samples were separated by SDS-­‐PAGE and transferred to PVDF. Secondary antibodies were conjugated to horseradish peroxidase and western blots were visualized with chemiluminescence 3.5 RESULTS 3.5.1 14-­‐3-­‐3 protein function is required for growth cone expansion in response to acute Netrin-­‐1 treatment. The axon guidance effects of Netrin-­‐1 are in part mediated by the response of multiple proteins downstream of DCC that act rapidly to alter the cytoskeletal dynamics of commissural neuron growth cones (Shekarabi, Moore et al. 2005). 14-­‐3-­‐3 proteins regulate numerous signalling proteins that effect cytoskeletal dynamics. In order to examine if 14-­‐3-­‐3 proteins were required for acute responses to Netrin-­‐1 in commissural neurons, we performed loss of function experiments with the 14-­‐3-­‐3 binding inhibitory peptide R18 (PHCVPRDLSWLDLEANMCLP) and an inactive control peptide. R18 is a potent 14-­‐3-­‐3 peptide antagonist, which disrupts binding of Ser/Thr phosphorylated proteins to 14-­‐3-­‐3(Wang, Yang et al. 1999). We generated a recombinant R18-­‐green fluorescent protein expressing herpes simplex virus (HSV-­‐R18-­‐GFP) and a control HSV-­‐WLRL-­‐GFP virus that does not bind to 14-­‐
3-­‐3 because of a substitution in the core-­‐binding motif (WLDL to WLRL/WLKL) (Wang, Yang et al. 1999). Dissociated commissural neurons were transduced with HSV GFP-­‐R18 or HSV GFP-­‐WLRL, cultured for 2DIV and then treated with Netrin-­‐1 for 10 min. The cells were fixed and the underlying actin cytoskeleton was visualized -­‐ 137 -­‐ Results A
Untreated
B
Netrin-­1
p = 0.0193*
Phalloidin
GFP
R18
WLRL
Growth Cone Area
(% of untreated +/-­ SEM)
160% 140% 120% WLRL R18 100% 80% 60% 40% 20% Figure 1: Inhibition of 14-­‐3-­‐3 function blocks Netrin-­‐1 induced growth cone expansion (A,B) Dissociated commissural neurons were transduced with HSV-­‐GFP WLRL or HSV-­‐GFP R18 and cultured for 2DIV. Prior to fixation, neurons were treated for 10 min with 200ng/ml Netrin-­‐1 or vehicle control (PBS). The actin cytoskeleton of individual neurons was visualized with rhodamine conjugated phalloidin (Red), and the area of GFP (Green) expressing growth cones was measured. Netrin-­‐1 treatment induces a 45% increase in growth cone area in WLRL expressing cells, but not in R18 expressing cells. (n=4 independent experiments, p<0.05, unpaired t-­‐test)Scale bar: 5µm. with a fluorescently conjugated phalloidin. Previous reports show Netrin-­‐1 treatment results in an expansion of the growth cone (Shekarabi, Moore et al. 2005), and consistently Netrin-­‐1 treatment of cells expressing the WLRL control results in an expansion of growth cone area to 145% (+/-­‐ 7%) of the growth cones of untreated cells. R18 expression in untreated commissural neurons had no significant effect on growth cone area compared to WLRL expression (95.5% +/-­‐7.7%). Furthermore, the growth cone area of Netrin-­‐1 treated cells expressing R18 were 79% (+/-­‐12%) of those in untreated cells (Fig. 1), indicating that Netrin-­‐1 had no significant effect on growth cone expansion in the absence of 14-­‐3-­‐3 activity. Thus 14-­‐3-­‐3s are required for Netrin-­‐1 signalling to alter the cytoskeleton of commissural neurons. -­‐ 138 -­‐ Results 3.5.2 14-­‐3-­‐3 activity is required for attractive turning responses of pre-­‐crossing commissural neurons to Netrin. To directly test whether 14-­‐3-­‐3 binding is required for the axons of commissural neurons to be attracted along gradients of Netrin-­‐1, we used the Dunn Chamber turning assay in conjunction with HSV transduction of R18 or WLRL in commissural neurons. Cells were cultured for 2DIV and then live imaged while being subjected to a gradient of the DCC binding fragment of Netrin-­‐1. Neurons transduced with the control HSV WLRL GFP showed an attractive turning response, with a mean turning angle of 12.6±4.6°, similar to previously reported data (Ch. 2). Expression of the R18 resulted in abolishing any significant turning response to the Netrin-­‐1 gradient. Neurons showed a mean turning angle of -­‐0.9±4.0° which suggests a randomization of turning responses and is not significantly different (p=0.3376) than neurons imaged in the absence of any gradient (Fig. 2 A,B). Importantly, when the amount of axon extension occurring during the imaging period was measured, no significant difference was seen between the neurons expressing R18 (22.5 +/-­‐ 0.75µm) and those expressing WLRL (20.21+/-­‐
0.59µm) (Fig.2C). This indicates that the lack of Netrin-­‐1 turning response in R18 expressing commissural neurons is not attributable to any impairment in axon outgrowth. Therefore, we have shown 14-­‐3-­‐3 protein activity is required for Netrin-­‐1 attractive turning responses in commissural neurons. 3.5.3 14-­‐3-­‐3 protein activity is required for Netrin-­‐1 induced neurite outgrowth from Dorsal explants of embryonic spinal cord. In addition to inducing attractive turning responses along two-­‐dimensional gradients, long-­‐term Netrin-­‐1 treatment has been shown to direct axon outgrowth in a number of neuronal cell types (Serafini, Kennedy et al. 1994; Li, Saint-­‐Cyr-­‐Proulx et al. 2002; Liu, Beggs et al. 2004), In dorsal spinal cord explants embedded in a three dimensional -­‐ 139 -­‐ Results A
n=82
0°
-­30°
R18
30°
-­60°
60°
8
-­90°
repulsion
Mean Turning Angle (degrees +/-­ SEM)
B
12
0°
-­30°
30°
n=106
-­60°
60°
16
90°
8
-­90°
12
16
90°
attraction
C
25 20 *
15 10 5 0 -­5 WLRL R18 Mean Displacement (m +/-­ SEM) WLRL
30 25 20 15 10 5 0 WLRL R18 Figure 2: Inhibition of 14-­‐3-­‐3 function blocks attractive turning to Netrin-­‐1 gradients in commissural neurons (A) Rose histograms of the distribution of turned angles of commissural neurons at 2 DIV in a 0.2 μg/ml gradient of the VI,V fragment of Netrin-­‐1. Reponses of individual neurons were clustered in 10° bins and the percentage of total neurons per bin is represented by the radius of each segment. Neurons transduced with HSV-­‐ GFP WLRL or HSV-­‐GFP R18. Inhibition of 14-­‐3-­‐3 proteins with R18 randomizes turning responses in response to the Netrin-­‐1 gradient. (B) Mean angle turned (+/-­‐ SEM); *p<0.05, One way ANOVA (C) Mean displacement of axons during the 2 hour imaging. period (+/-­‐ SEM). R18 does not impair the outgrowth of axons, compared to the WLRL control. collagen matrix, Netrin-­‐1 directed outgrowth results in the increased emergence of neurites from the explant. In order to test if 14-­‐3-­‐3 proteins play a role in the long-­‐term directed outgrowth of axons, in addition to the turning response, we microdissected E13 rat spinal cords and made explants from the most dorsal fifth, which contain the cell bodies of commissural -­‐ 140 -­‐ Results interneurons. The explants were embedded in a collagen matrix and cultured in the presence or absence of Netrin-­‐1 along with purified WLKL or R18 peptides. Both peptides were tagged with the HIV derived Tat sequence in order to make them cell permeable. Consistent with previous reports (Moore, Correia et al. 2008), dorsal explants cultured with no peptides showed multiple bundles of axons extending into the matrix after ~20hrs in culture when Netrin-­‐1 was included in the media, but showed very little outgrowth outside of the explant without Netrin-­‐1 (Fig. 3A). This pattern was seen with Tat-­‐WLKL treated explants, with explants cultured without Netrin-­‐1 showing only a mean of 1.14 +/-­‐0.32mm of total axon outgrowth per mm2 of explant and those cultured in the presence of Netrin-­‐1 showing a mean of 6.96 +/-­‐
1.57mm of total axon outgrowth per mm2 of explant. Tat-­‐R18 treated explants had similar outgrowth outside of the explant when cultured without Netrin-­‐1 (1.17+/-­‐0.52 mm/mm2 of explant area), and notably failed to show any significant increase in visible axon outgrowth in the presence of Netrin-­‐1, with only 2.18+/-­‐0.7 mm of total axon outgrowth per mm2 of explant area (Fig.3B). Hence, 14-­‐3-­‐3 activity is required for Netrin-­‐1 to direct axon outgrowth from commissural neurons into the collagen matrix. 3.5.4 14-­‐3-­‐3s are required for Netrin-­‐1 induced activation of MAPK signalling and interact with DCC but do not impact Netrin-­‐1 induced changes to DCC distribution. A recent report has shown that 14-­‐3-­‐3 proteins can also play a critical role in axon guidance through binding to a guidance cue receptor and regulating interactions with downstream effector proteins (Yang and Terman 2012). This led us to ask whether the role of 14-­‐3-­‐3 function in mediating Netrin-­‐1 signalling may be through interactions between the receptor DCC and downstream effector proteins. DCC has been shown to interact in a complex with proteins involved in multiple signalling pathways (Forcet, Stein et al. 2002; Li, Lee et al. 2004; Shekarabi, Moore et al. 2005) -­‐ 141 -­‐ Results Untreated
TAT-­WLKL
TAT-­R18
Netrin-­1
0 ng/ml
200 ng/ml
B
Total Outgrowth/Explant Area (microns/ mm² +/-­‐SEM)
A
*
5000
0 ng/ml Netrin-­1
*
200 ng/ml Netrin-­1
4000
3000
n.s.
2000
1000
0
Untreated
WLKL
R18
Figure 3: Loss of 14-­‐3-­‐3 function blocks Netrin-­‐1 directed outgrowth from dorsal spinal cord explants (A,B)Embryonic spinal cords were dissected and explants were isolated from the most dorsal fifth and embedded in a collagen matrix. Explants were then cultured for 19hrs with or without 200ng/ml Netrin-­‐1 and 100µg/ml of either Tat tagged WLKL or R18 and then imaged. Total axon outgrowth length was measured for each explant and the normalized by the area of the explant. Untreated and control explants show significant increases in the amount of visible axon outgrowth. R18 treated explants fail to show a significant increase with Netrin-­‐1. (n= 5 independent experiments, *p>0.05, Two way ANOVA, with Bonferroni post-­‐test). Scale Bar:100µm and upon Netrin-­‐1 binding, trigger activation of a number of different kinases shown to be required for growth cone changes and axon turning including, Src Family Kinases (Li, Lee et al. 2004; Meriane, Tcherkezian et al. 2004), p21-­‐activated kinase 1 (PAK1) (Shekarabi, Moore et al. 2005), PKA (Bouchard, Moore et al. 2004; Lebrand, Dent et al. 2004) and mitogen activated protein kinase (MAPK) (Forcet, Stein et al. 2002). 14-­‐3-­‐3 proteins have also been shown to play an important regulatory role in the MAPK signalling cascade (Roy, McPherson et al. 1998). -­‐ 142 -­‐ Results To ask if 14-­‐3-­‐3 activity is required for Netrin-­‐1 activation of MAPK signalling we co-­‐transfected HEK 293T cells with plasmids expressing DCC and either WLRL or R18 peptides. Cells were treated with Netrin-­‐1 and lysates were examined by Western blotting for the active phosphorylated form of ERK1/2, a downstream kinase in the MAPK signalling cascade (Fig 5 A,B). Netrin-­‐1 treatment of cells expressing DCC and WLRL resulted in an increase in levels of phospho-­‐ERK1/2 to 151.3% of untreated, whereas those cells expressing R18 and DCC failed to elevate levels of phospho-­‐ERK1/2 remaining at 92.3% of untreated (n=5, p>0.05 unpaired t-­‐test). Thus, 14-­‐3-­‐3 activity is required for Netrin-­‐1 signalling and downstream activation of MAPK signalling . To look for an interaction between DCC and the 14-­‐3-­‐3 proteins, we used an antibody that recognizes all isoforms of the 14-­‐3-­‐3 family, to immunoprecipitate the 14-­‐3-­‐3 proteins from lysates of HEK 293T cells that were transfected with DCC and treated with either Netrin-­‐1 or vehicle control. We found that DCC co-­‐immunoprecipitates with 14-­‐3-­‐3 proteins and that the interaction is not regulated by Netrin-­‐1 stimulation (Fig.4C)(n=3 independent experiments). We asked if the lack of response to Netrin-­‐1 in the presence of R18 could be explained by insufficient cell surface expression of DCC. Dissociated commissural neurons transduced with either WLRL or R18 were cultured for 2DIV and then treated with either Netrin-­‐1 or control. Prior to lysis, the cells were treated with biotin to label all cell surface proteins, which were then isolated from cell lysates by streptavidin pulldown. Western blotting of the eluates from the streptavidin columns shows that Netrin-­‐1 treatment of WLRL expressing cells results in a 35% increase in the total levels of DCC at the cell surface (Fig.4A) (n=6, p<0.05, unpaired t-­‐test). Blocking 14-­‐3-­‐3 function through R18 expression had no impact on the levels of DCC at the cell surface in untreated neurons and the neurons show a similar increase upon Netrin-­‐1 treatment. -­‐ 143 -­‐ Results WLRL
R18
-­ + + -­ + + DCC
-­ -­ + -­ -­ + Netrin-­1
IB: DCC
IB: Total ERK1/2
IB: phospho ERK1/2
C
IP
IgG
14-­3-­3
-­ + Netrin
IB:DCC
D
pERK Intensity (fold change over control +/-­SEM) B
A
WLRL
2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 R18
-­ + -­ + IP
*
WLRL R18 Netrin
DCC biotin
Input
DCC input
Figure 4: 14-­‐3-­‐3s mediate Netrin-­‐1 induced MAPK signalling and interact with Netrin-­‐1 receptor DCC but do not effect its distribution at cell surface (A,B) HEK 293T cells were transfected with plasmids expressing DCC and WLRL or R18, serum starved and then treated with 200ng/ml Netrin-­‐1 for 10 minutes prior to lysing. Western blots using antibodies against total and phosphorylated ERK1/2 show WLRL expressing cells undergo a significant increase in levels of phosphorylated ERK1/2 with Netrin-­‐1 treatment. R18 expressing cells do not show an increase. (n= 6 independent experiments, *p>0.05, unpaired t-­‐test) (C) HEK 293T cells were transfected with a plasmid to express full-­‐length DCC. Lysates from untreated or Netrin-­‐1 treated cells were used for pan 14-­‐3-­‐3 or IgG immunoprecipitation. DCC co-­‐immunoprecipitates with 14-­‐3-­‐3s in both untreated and Netrin-­‐1 treated cells. (D) Cell surface biotinylation of dissociated commissural neurons expressing either WLRL or R18. Cells were treated with 200ng/ml of Netrin-­‐1 prior to lysis and streptavidin pulldown. Both WLRL and R18 expressing cells show an increase in DCC at the cell surface upon Netrin-­‐1 treatment. 3.6 DISCUSSION Together our data shows that 14-­‐3-­‐3 protein function plays a critical role in mediating signalling by Netrin-­‐1 through DCC and the resulting attractive axon guidance effects on pre-­‐crossing commissural neurons in the -­‐ 144 -­‐ Results developing spinal cord. In the absence of 14-­‐3-­‐3 function, Netrin-­‐1 loses its ability to alter growth cone morphology, induce attractive axon turning and to induce axon outgrowth out of a dorsal spinal cord explant. 14-­‐3-­‐3 activity is not required for the insertion of DCC at the cell surface but 14-­‐3-­‐3 proteins are found in complex with DCC in the presence or absence of Netrin. Further, 14-­‐3-­‐3 function is required for the activation of MAPK signalling downstream of DCC. Thus, in addition to their role in regulating repellent turning responses, we show that 14-­‐3-­‐3 proteins are critical for attractive axon guidance in vitro. 3.6.1 Spatial regulation of DCC in Netrin-­‐1 signalling We find that, although 14-­‐3-­‐3 proteins do interact with DCC and are critical for Netrin-­‐1 signalling at the growth cone, blocking 14-­‐3-­‐3 function does not have any impact on the cell surface expression of DCC. R18 is known to enhance the activity of PKA in neurons (See chapter 1 and chapter 2) and Netrin-­‐1 enhances cell surface levels of DCC in a PKA-­‐dependent manner (Moore and Kennedy 2006). Together, this suggests that the 14-­‐3-­‐3 isoforms expressed in pre-­‐crossing commissural neurons are not expressed at sufficient levels to suppress the activity of PKA. There is increasing evidence that the spatial distribution of DCC to localized signalling microdomains is also critical for Netrin-­‐1 based axon guidance. Netrin-­‐1 signalling through DCC has been shown to illicit distinct patterns of cAMP and Ca2+ second messenger responses in different compartments of the growth cone (Nicol, Hong et al. 2011). The localization of DCC to detergent insoluble lipid rafts is required for the growth cone turning effects of Netrin-­‐1 (Guirland, Suzuki et al. 2004). Raft associated DCC interacts with different downstream effectors than non-­‐raft DCC (Petrie, Zhao et al. 2009) and early DCC signalling events such as MAPK activation is raft association dependent (Guirland, Suzuki et al. 2004; Herincs, Corset et al. 2005). As one of their many functions, 14-­‐3-­‐3s have been shown to bind and regulate the localization of receptor complexes to signalling microdomains, localizing integrin receptor complexes to the -­‐ 145 -­‐ Results lammelapodia of migrating keratinocytes (Santoro, Gaudino et al. 2003). Given our findings that 14-­‐3-­‐3s are critical for mediating MAPK activation, it will be interesting to explore whether disrupting 14-­‐3-­‐3/DCC interactions has effects on the spatial distribution of DCC to lipid raft microdomains, which would not have been detected in our cell surface biotinylation experiments. Netrin-­1
DCC
DCC
MEK
DCC
MEK
ERK1/2
14-­3-­3
B-­RAF
14-­3-­3
Figure 5: Model for 14-­‐3-­‐3 regulation of MAP Kinase signalling in response to Netrin-­‐1 binding to DCC. The requirement of 14-­‐3-­‐3 protein binding to activate ERK1/2 upon Netrin-­‐1 treatment and prior reports indicating 14-­‐3-­‐3 can promote B-­‐Raf activity leads us to propose the following model. 14-­‐3-­‐3 dimers are in complex with DCC and MEK (left). Upon Netrin-­‐1 binding (right), formation of DCC signalling complexes includes 14-­‐3-­‐3 mediated activation of B-­‐Raf, which in turn activates the associated MEK and subsequently the phosphorylation of ERK1/2. 3.6.2 14-­‐3-­‐3s as regulators of downstream signalling proteins In addition to altering receptor distribution, the requirement of 14-­‐3-­‐3 protein function to mediate Netrin-­‐1 signalling through DCC could also be explained by the ability of 14-­‐3-­‐3 proteins to regulate the interactions -­‐ 146 -­‐ Results between different proteins involved in downstream signalling pathways such as MAPK. 14-­‐3-­‐3s are known to be involved in the activation of B-­‐Raf (Qiu, Zhuang et al. 2000), a kinase upstream of MEK-­‐1/2, which in turn has been shown to bind DCC independent of Netrin-­‐1 and is a known activator of ERK1/2 (Forcet, Stein et al. 2002). Furthermore, the activity of the B-­‐Raf activating GTPase Ras has been shown to be enhanced by 14-­‐3-­‐3 suppression of RasGAPs (Benzing, Yaffe et al. 2000; Feng, Yunoue et al. 2004). Particularly relevant for axon guidance, 14-­‐3-­‐3ε has been shown in Drospohila to bind the Sema1A receptor, Plexin A and block its RasGAP function, providing spatial regulation to Sema1A repulsion and allowing Ras to re-­‐establish integrin mediated adhesion (Yang and Terman 2012). Our finding that 14-­‐3-­‐3 interacts with the Netrin-­‐1 receptor DCC raises the possibility that it could provide similar spatial regulation by determining the interactions between DCC and other downstream effectors. One possible model, (Fig.5) would be that 14-­‐3-­‐3 dimers exist in complex with unbound DCC and MEK. Upon Netrin-­‐1 binding, formation of DCC signalling complexes promotes 14-­‐3-­‐3 mediated activation of B-­‐Raf, perhaps by recruiting it to the complex and limiting the access of a RasGAP. This in turn activates the associated MEK and subsequently the phosphorylation of ERK1/2. Given these findings, we will perform further experiments to examine the impacts of loss of 14-­‐3-­‐3 binding to DCC on its interactions with and ability to activate signalling components such as MEK1/2 (Forcet, Stein et al. 2002), the Src Family kinases (Li, Lee et al. 2004; Meriane, Tcherkezian et al. 2004), the ERM (Ezrin/Radixin/Moesin) proteins (Antoine-­‐Bertrand, Ghogha et al. 2011)and GEFs such as Trio (Briancon-­‐Marjollet, Ghogha et al. 2008). -­‐ 147 -­‐ Discussion a nd C onclusion Discussion and Conclusion 1. Summary The major aim of this thesis was to identify and characterize constituent proteins of the axonal growth cone that act as critical spatial and temporal regulators of the signalling pathways that direct axonal outgrowth towards targets during development and after injury. The 14-­‐3-­‐3 family of adaptor proteins were initially identified as candidates through a proteomic screen of isolated growth cones described in Chapter 2 and the results presented herein establish the importance of these molecules in regulating guidance decisions both in vitro and in vivo. Chapter 2 of this thesis describes how expression of 14-­‐3-­‐3 isoforms is cell specific and developmentally regulated. In older neurons, when 14-­‐3-­‐3 expression is higher and levels of cAMP decrease, growth cone turning responses to NGF and MAG are repellent. These repellent responses are reversed back to attractive turning in vitro, with the general loss of 14-­‐3-­‐3 function, and the knockdown of specific 14-­‐3-­‐
3 isoforms. Thus indicating 14-­‐3-­‐3 regulation as an important contributor to establishing axonal repulsion. Our results establish that the switch back to attractive turning is mediated in part by an increase in active PKA and that 14-­‐3-­‐3 interactions with the PKA holoenzyme serve to stabilize the complex and down-­‐regulate PKA catalytic activity in the growth cone. In Chapter 3 of this thesis, we extend these findings to an in vivo model of axon guidance. We describe the developmental regulation of 14-­‐3-­‐3 isoform expression in the axonal tracts of commissural interneurons of the spinal cord, showing a cell-­‐intrinsic time dependent increase in the presence of certain 14-­‐3-­‐3 isoforms. This coincides with the crossing of the midline by these cells as well as a cell-­‐intrinsic temporal switch in the responsiveness of the growth cone to gradients of Shh from attraction to repulsion. We demonstrate that 14-­‐3-­‐3 expression is necessary and sufficient to induce the -­‐ 151 -­‐ Discussion and Conclusion switch in responsiveness and that this is mediated and can be mimicked by changes in PKA activity. Importantly we demonstrate that either pre-­‐crossing overexpression of 14-­‐3-­‐3s or loss of 14-­‐3-­‐3 function at the midline results in axons being misdirected along the anterior-­‐posterior axis of the spinal cord. Thus indicating that 14-­‐3-­‐3s are critical in vivo to establish a switch in guidance response to Shh. This occurs most likely through a temporal down-­‐
regulation of PKA activity at the midline, in agreement with previous findings (Parra and Zou 2010). Beyond their role in regulating switches to repellent responses, in Chapter 4 of this thesis, we establish 14-­‐3-­‐3s as important mediators of an attractive axon guidance signal. Our results demonstrate that loss of 14-­‐3-­‐3 function in pre-­‐crossing commissural neurons blocks the growth promoting and attractive turning responses that these cells show in the presence of Netrin-­‐1. We describe an interaction between 14-­‐3-­‐3s and the Netrin-­‐1 receptor DCC and show how loss of 14-­‐3-­‐3 binding does not effect the insertion of DCC at the plasma membrane but does alter its ability to trigger downstream signalling cascades. This suggests that 14-­‐3-­‐3 regulation of receptor protein interactions could provide a key pathway for spatially regulating axon guidance signalling. 2. 14-­‐3-­‐3 p roteins a s n egative r egulators o f P KA We find that 14-­‐3-­‐3 ε and γ interact with the type II regulatory subunits of PKA. In the presence of inhibitors of 14-­‐3-­‐3 binding, less catalytic subunit is found to be bound to the regulatory subunits and an increase in the autophosphorylation of PKA and other substrates in the growth cone is observed. Together with growth cone turning phenotypes that depend on increases in PKA activity, these findings indicate that 14-­‐3-­‐3s act as negative regulators of PKA activity, potentially by binding and stabilizing the holoenzyme complex. -­‐ 152 -­‐ Discussion and Conclusion PKA activity has been shown to be under very precise subcellular spatial control in a number of different cellular contexts (Gervasi, Tchenio et al. 2010; Lignitto, Carlucci et al. 2011; Choi, Berrera et al. 2012; Sample, DiPilato et al. 2012). Spatial and temporal patterning of PKA activity is critical for regulating the activity of Rho family GTPases which in turn coordinate dynamic changes in the cytoskeleton at the leading edge of migrating cells (Tkachenko, Sabouri-­‐Ghomi et al. 2011). The precise spatial control of PKA activity relies on the coordination of changes in cAMP levels with the targeting of PKA to signalling microdomains. It has been well established that this is achieved by a diverse family of proteins known as protein kinase A anchoring proteins (AKAPs) (Logue and Scott 2010). In general, AKAPs bind the type II regulatory (RII) subunits of the PKA holoenzyme, although some type I isoform specific AKAPs have been found (Gold, Lygren et al. 2006). In addition to targeting the holoenzyme to microdomains, they promote the formation of signalling complexes with relevant downstream targets (Tasken and Aandahl 2004). For example, in the case of migrating cells, the cytoskeletal associated proteins, WASP family Verprolin-­‐homologous protein 1, (WAVE1) (Westphal, Soderling et al. 2000), microtubule associated protein 2 (MAP2) (Theurkauf and Vallee 1982) and ezrin (Dransfield, Bradford et al. 1997) have all been found to be AKAPs in addition to their roles in regulating actin and microtubule dynamics. AKAP-­‐lbc has been found to also have a guanine exchange factor (GEF) activity and coordinates PKA signalling with activation of Rho GTPases (Jin, Smith et al. 2004). Interestingly its’ GEF activity has also been found to be regulated by 14-­‐3-­‐3 binding. In growth cones, the ability of PKA activity to alter the turning response to at least one guidance cue has been shown to depend on AKAP mediated localization of RII subunits to the base of filopodia (Han, Han et al. 2007). However, one report showed that disrupting AKAP binding in growth cones does not entirely displace PKA (Rivard, Birger et al. 2009), raising the possibility that the -­‐ 153 -­‐ Discussion and Conclusion existence of PKA signalling microdomains within the growth cone is also mediated in part by 14-­‐3-­‐3 binding. One interesting feature of the dynamics of PKA activity within microdomains that has begun to emerge in recent studies is that signalling is often mediated by oscillatory patterns of kinase activity (Ni, Ganesan et al. 2011; Tkachenko, Sabouri-­‐Ghomi et al. 2011). In growth cones this is supported by recent findings that showed the ability of cAMP to drive attractive turning depends on localized cycling of changes in cAMP concentration (Nicol, Hong et al. 2011). In order to generate dynamic oscillatory signals, one requirement of the signalling system is to incorporate a feedback inhibition mechanism. In the case of cAMP-­‐PKA signalling this feedback inhibition has often been found to involve PKA phosphorylation of AKAP bound adenyl cyclases or phosphodiesterases, which are responsible for controlling cAMP levels (Vandamme, Castermans et al. 2012). However PKA activity within microdomains can be regulated in the absence of changes to cAMP (Barzi, Berenguer et al. 2010). Active PKA autophosphorylates and targets RII subunits as well, generating 14-­‐3-­‐3 binding motifs and regulating AKAP interactions (Budillon, Cereseto et al. 1995; Manni, Mauban et al. 2008). Our data indicates that 14-­‐3-­‐3 is a negative regulator of PKA and raises the possibility that 14-­‐3-­‐3 regulation could work in conjunction with AKAPs to provide feedback inhibition of PKA in microdomains, independent of cAMP changes, providing the spatial and temporal specificity required in axon guidance signalling. Intriguingly one recent report has shown that Shh signalling in the developing cerebellum requires AKAP mediated localization of RII subunits of PKA to microdomains at the base of cilia in cerebellar granule cell precursors (Barzi, Berenguer et al. 2010). Activation of Shh signalling leads to a decrease in PKA activity within the microdomain, similar to reported changes in commissural neurons at the midline (Parra and Zou 2010), raising the possibility that a similar mechanism could be used later in -­‐ 154 -­‐ Discussion and Conclusion development to regulate switches in axon guidance responses. Further experiments to address these possibilities could be performed making use of a new generation of fluorescence resonance energy transfer (FRET)-­‐ based reporters of PKA activity (Allen and Zhang 2006) in growth cones to examine the spatio-­‐temporal dynamics of PKA microdomains when 14-­‐3-­‐3 activity is perturbed. 3. Potential alternative mechanisms for 14-­‐3-­‐3 regulation o f g rowth c one t urning r esponses While the results reported in this thesis demonstrate that 14-­‐3-­‐3 regulation of PKA activity is critical for the spatial and temporal control of growth cone turning responses, we cannot rule out the possibility that 14-­‐3-­‐3s may act through other mechanisms to alter axon guidance signalling. Calcium signalling plays a number of critical roles in regulating the spatial and temporal outcomes of growth cone signalling (see Introduction 1.2). Amongst the many binding targets reported for 14-­‐3-­‐3 proteins, a number of findings indicate 14-­‐3-­‐3s can regulate intracellular calcium signals through a few different mechanisms. The influx of Ca2+ into the cell through voltage dependent calcium channels is a major source of Ca2+ signalling in neurons, and the ability of these channels to be inactivated is regulated by 14-­‐3-­‐3s. Binding of 14-­‐3-­‐3s to CaV2.2 channels slows their inactivation, leading to sustained Ca2+ influx and reduced short–term depression in hippocampal cells (Li, Wu et al. 2006). Also intracellular Ca2+ levels can be sustained by 14-­‐
3-­‐3 mediated inhibition of Na+/Ca2+ exchangers (Pulina, Rizzuto et al. 2006). The direct modulation of plasma membrane channel properties by 14-­‐3-­‐3s could be a potential mechanism for altering growth cone turning. 14-­‐3-­‐3s also impact Ca2+ signalling through targeting downstream effectors. The Ca2+ dependent kinase, PKC has been shown to promote attractive growth cone turning (Williams, Wu et al. 2003; Wolf, Lyuksyutova et al. 2008) and to be -­‐ 155 -­‐ Discussion and Conclusion inhibited by 14-­‐3-­‐3 binding (Aitken, Howell et al. 1995). Also Calmodulin dependent kinase kinase, (CaMKK) activates calmodulin dependent kinases, such as the attractive turning regulator CaMKII, and is inhibited by 14-­‐3-­‐3 binding (Davare, Saneyoshi et al. 2004). Further downstream, 14-­‐3-­‐3s are known to play key regulatory roles through interactions with proteins that directly control the coordinated changes in the cytoskeleton that occur in response to extracellular cues. The activity of small GTPases, such as the Rho family, is a key mediator of such signals. Their activation is promoted by guanine exchange factors (GEFs) and downregulated by GTPase activating proteins (GAPs). In non-­‐neuronal cells, 14-­‐3-­‐3s have been found to interact directly with and inhibit a number of Rho GEFs such as Lbc (Jin, Smith et al. 2004), Lfc (Meiri, Greeve et al. 2009) and GEF-­‐H1 (Zenke, Krendel et al. 2004), indicating another potential mechanism for altering growth cone turning responses. Furthermore, 14-­‐3-­‐3 inhibition of the GAP activity of the Sema receptor, Plexin, has already been shown to be critical for mediating axon guidance effects in the motor neurons of the Drosophila CNS (Yang and Terman 2012). The targets of Rho signalling, include proteins such as the actin severing protein ADF/cofilin and the motor protein myosin II, which have direct impacts on the dynamics of actin filaments within the growth cone. 14-­‐3-­‐3s have been shown to inhibit cofilin activity in neurons (Yoon, Zivraj et al. 2011), and in non-­‐neuronal contexts inhibit myosin light chain phosphatase (Koga and Ikebe 2008) and promote myosin II activity (Zhou, Kee et al. 2010). While it is not clear exactly how 14-­‐3-­‐3 regulation of these effector proteins would result in switches between attractive and repellent turning responses of growth cones, these reports indicate that 14-­‐3-­‐3s may play a number of complex signalling roles in axon guidance. -­‐ 156 -­‐ Discussion and Conclusion 4. Roles for 14-­‐3-­‐3 regulation of axon guidance in other n euronal c ell t ypes Our results clearly demonstrate that 14-­‐3-­‐3s are critical for regulating the turning response of growth cones during the development of dorsal root ganglion neurons and commissural interneurons. Both these neuronal cell types are powerful models of axon guidance, providing many insights into the mechanisms underlying the signalling complexity present in the growth cone. However, many other neurons are faced with complex guidance decisions during development and our findings together with other recent reports, suggest that 14-­‐3-­‐3s may play key roles in providing spatial and temporal specificity in other axon guidance contexts. The initial screen which identified 14-­‐3-­‐3 proteins as significant constituent molecules in the growth cone was performed with growth cones isolated from early stage embryonic chick retinal ganglion cells. The strong expression of 14-­‐3-­‐3s in RGCs was collaborated by another report showing the presence of 14-­‐3-­‐3 mRNAs in Xenopus RGC growth cones (Zivraj, Tung et al. 2010) along with protein expression (Yoon, Zivraj et al. 2011). Interestingly in Xenopus, both mRNA and protein levels of 14-­‐3-­‐3s are developmentally regulated. RGCs isolated from very early stage (stage 24) embryos show little 14-­‐3-­‐3 transcript before the axons are actively projecting along the optic nerve. A dramatic increase in 14-­‐3-­‐3 mRNA and protein expression is reported in stage 28 to stage 33 growth cones, corresponding to the period of active axon extension towards the tectum. Finally, RGC growth cones that have already reached their destinations in the tectum (stage 40), are reported to have a significant decrease in 14-­‐3-­‐3 protein levels (Yoon, Zivraj et al. 2011). This pattern of regulated 14-­‐3-­‐3 expression is similar to the one we demonstrate in commissural neurons, in that as axons project through developmental pathways involving -­‐ 157 -­‐ Discussion and Conclusion intermediate targets, expression is increased. A possible role for 14-­‐3-­‐3s regulating RGC axon guidance is suggested by misrouted axonal projections found when expressing the R18 inhibitor in developing RGCs (Yoon, Zivraj et al. 2011). Interestingly, the projections of RGCs in mammals, have been shown to rely on both attractive Shh signalling at very early stages (Kolpak, Zhang et al. 2005) and later repellent Shh signalling at the optic chiasm (Trousse, Marti et al. 2001; Sanchez-­‐Camacho and Bovolenta 2008; Fabre, Shimogori et al. 2010). While the polarity of RGC turning response to Shh was shown, in vitro, to be concentration dependent (Kolpak, Zhang et al. 2005), it is unknown how RGCs switch their responsiveness in vivo. The increase in 14-­‐3-­‐3 expression, along with the established decrease in cAMP activity (Shewan, Dwivedy et al. 2002), raises the possibility that a mechanism similar to the one we demonstrate in commissural neurons exists in RGCs to switch the responsiveness to Shh signalling through PKA regulation. One striking report that indicates 14-­‐3-­‐3s may play a role in regulating axon guidance in another neuronal context, is the phenotype of 14-­‐3-­‐3ε knockout mice. Mostly characterized for defects in the radial migration of neurons in the cortical layers and their relationship with lissencephaly alleles, these mice also show disordered hippocampal structures, with loosely packed and heterotopic pyramidal and granule cell layers (Toyo-­‐oka, Shionoya et al. 2003). While the axonal projections of these structures were not characterized, it is well established that the proper formation of the hippocampus relies on both the radial and tangential migration of neurons and is regulated by a number classical axon guidance cues (Marin and Rubenstein 2003; Stanco, Szekeres et al. 2009). Furthermore, these cues also act to organize both afferent and efferent projections and proper formation of hippocampal layers (Skutella and Nitsch 2001). For example, mice lacking Netrin-­‐1 show altered patterns of layer specific terminations of the CA1 to CA3 projections (Barallobre, Del Rio et al. 2000), as well as aberrant -­‐ 158 -­‐ Discussion and Conclusion positioning of interneurons (Stanco, Szekeres et al. 2009). Our findings that 14-­‐3-­‐3s are an important mediator of Netrin-­‐1 signalling through DCC and the existence of disordered hippocampal layers in both 14-­‐3-­‐3 and Netrin-­‐1 mutants raises the possibility that 14-­‐3-­‐3 regulation of axon guidance signalling provides important spatial regulation to guidance decisions in the developing hippocampus. 5. Implications of 14-­‐3-­‐3 regulation of PKA beyond axon g uidance Both the 14-­‐3-­‐3 family of adaptor proteins and PKA activity are widely expressed in eukaryotic cells, and have been shown to play critical roles in signalling events in many important systems. Subsequently, our findings that show 14-­‐3-­‐3 isoforms act as a negative regulator of PKA activity and provide temporal and spatial specificity to PKA signalling during axon guidance, raises the possibility that a similar mechanism exists in other cellular contexts where PKA signalling needs precise regulation. For example, cardiac hypertrophy in response to hormones, relies on subdomains of PKA activity (McConnachie, Langeberg et al. 2006) mediated by AKAP-­‐lbc, (Carnegie, Soughayer et al. 2008) which interacts with both PKA and 14-­‐3-­‐3s (Diviani, Abuin et al. 2004). Also, insulin effects on cells relies both on signals mediated by anchored PKA (Zhang, Hupfeld et al. 2005) and increased 14-­‐3-­‐3 binding to targets (Chen, Synowsky et al. 2011). Perturbations to 14-­‐3-­‐3 function have been linked to a number of cancers, and many years of studies have elucidated multiple connections to cancer progression (Freeman and Morrison 2011; Zhao, Meyerkord et al. 2011). Of particular relevance to our findings, is the role 14-­‐3-­‐3s play in modifying the ability of breast tumor cells to migrate in a directed manner. Two separate groups have previously identified 14-­‐3-­‐3 regulation of Slingshot (Eiseler, -­‐ 159 -­‐ Discussion and Conclusion Doppler et al. 2009; Peterburs, Heering et al. 2009), as well as regulation of the actin nucleating factor, cortactin (Eiseler, Hausser et al. 2010), via PKD phosphorylation, as down-­‐regulating tumor cell migration. Interestingly, the directed movement of cancer cells has also been shown to involve AKAP –
localized PKA microdomains at the leading edges of the cell (Paulucci-­‐
Holthauzen, Vergara et al. 2009; McKenzie, Campbell et al. 2011), similar to those reported in growth cones (Han, Han et al. 2007). If a mechanism similar to what we find existing in growth cones, is present in cancer cells, it is possible that loss of 14-­‐3-­‐3 function would promote tumor migration through both de-­‐regulated actin modifiers such as cortactin and Slingshot, and overactive PKA signalling. Within neurons, 14-­‐3-­‐3 regulation of PKA could potentially play an interesting role in the ability of neurons to modulate their connections after synaptogenesis, in processes that underlie memory formation. Some evidence for this exists in genetic studies done in Drosophila. Leonardo is the Drosophila homologue of 14-­‐3-­‐3ζ and is expressed primarily in the mushroom bodies of the CNS, which are structures responsible for olfactory memory formation. Mutations affecting the adenyl cyclase, rutabaga, the phosphodiesterase, dunce, or the catalytic subunit of PKA, strongly implicate PKA signalling in this process (Skoulakis and Grammenoudi 2006). Similarly, alleles knocking out leonardo expression greatly diminish performance in an olfactory learned avoidance task (Skoulakis and Davis 1996). It was shown that the memory defects were not due to disrupted development of the neurons, but that the requirement of leonardo for memory formation is acute (Philip, Acevedo et al. 2001). Interestingly, a more recent study using two-­‐
photon microscopy to image subcellular domains of PKA activity in the mushroom bodies of living flies was able to demonstrate that proper learning and memory activities results in distinct subcellular domains of PKA activity and that these are perturbed in the rutabaga and dunce mutants (Gervasi, -­‐ 160 -­‐ Discussion and Conclusion Tchenio et al. 2010). It would be very interesting to extend these studies to the leonardo mutants as well. 6. Importance of PKA regulation in axonal regeneration Our understanding of how PKA activity is regulated in growth cones is critical to making targeted interventions after axonal injury. This was highlighted by studies demonstrating that the elevated levels of cAMP in younger neurons underlie their ability to grow on normally inhibitory substrates such as MAG or myelin. A drop in neuronal cAMP levels during development contributes to the inhibitory growth responses to myelin and the inability of axons to re-­‐grow after injury (Cai, Qiu et al. 2001). Treatments to re-­‐elevate cAMP in older neurons, such as the systemic addition of rolipram, result in increased axon growth on inhibitory substrates and some regeneration of axons in the spinal cord after injury (Neumann, Bradke et al. 2002; Qiu, Cai et al. 2002) as well as increased functional recovery (Pearse, Pereira et al. 2004). This has been found to rely initially on the activation of PKA and then subsequently on transcriptional changes initiated by phosphorylated cAMP response element-­‐binding protein (CREB) (Gao, Deng et al. 2004). Studies following up on this strategy have been focused on the regulated transcription products such as Interleukin-­‐6 (IL-­‐6) (Cao, Gao et al. 2006) and Arginase-­‐1 (Arg-­‐1) (Ma, Campana et al. 2010), but the emergence of specialized PKA signalling microdomains also indicates that more targeted ways of activating PKA could prove promising and further work on the critical regulatory mechanism played by 14-­‐3-­‐3s could build on our findings, providing a novel target for therapeutic intervention. -­‐ 161 -­‐ Discussion and Conclusion 7. Conclusion Understanding the molecular pathways involved in regulating the behaviour of the growth cone is critical to our broader understanding of the complexities of the nervous system and to making targeted and effective interventions to promote recovery after injury. Our work in this thesis establishes 14-­‐3-­‐3s as novel regulators of growth cone signalling and demonstrates the important role they play in regulating PKA activity. We demonstrate how 14-­‐3-­‐3 activity provides spatial and temporal control to mechanisms that are responsible for targeted axonal outgrowth during development. Our work provides new insights into the mechanisms that underlie the exquisite precision of axon guidance signalling events and indicate new pathways of investigation for potential therapeutic interventions after injuries to the nervous system. -­‐ 162 -­‐ Bibliography Bibliography Aitken, A. (2006). "14-­‐3-­‐3 proteins: a historic overview." Seminars in Cancer Biology 16(3): 162-­‐172. Aitken, A., H. Baxter, et al. (2002). "Specificity of 14-­‐3-­‐3 isoform dimer interactions and phosphorylation." Biochem Soc Trans 30(4): 351-­‐360. Aitken, A., S. Howell, et al. (1995). "Post-­‐translationally modified 14-­‐3-­‐3 isoforms and inhibition of protein kinase C." Mol Cell Biochem 149-­‐150: 41-­‐49. Aizawa, H., S. Wakatsuki, et al. (2001). "Phosphorylation of cofilin by LIM-­‐kinase is necessary for semaphorin 3A-­‐induced growth cone collapse." Nat Neurosci 4(4): 367-­‐373. Akiyama, H., T. Matsu-­‐ura, et al. (2009). "Control of neuronal growth cone navigation by asymmetric inositol 1,4,5-­‐trisphosphate signals." Sci Signal 2(79): ra34. Allen, M. D. and J. Zhang (2006). "Subcellular dynamics of protein kinase A activity visualized by FRET-­‐based reporters." Biochem Biophys Res Commun 348(2): 716-­‐721. Andrews, G. L., S. Tanglao, et al. (2008). "Dscam guides embryonic axons by Netrin-­‐
dependent and -­‐independent functions." Development 135(23): 3839-­‐3848. Antoine-­‐Bertrand, J., A. Ghogha, et al. (2011). "The activation of ezrin-­‐radixin-­‐moesin proteins is regulated by netrin-­‐1 through Src kinase and RhoA/Rho kinase activities and mediates netrin-­‐1-­‐induced axon outgrowth." Mol Biol Cell 22(19): 3734-­‐3746. Augsburger, A., A. Schuchardt, et al. (1999). "BMPs as mediators of roof plate repulsion of commissural neurons." Neuron 24(1): 127-­‐141. Avraham, O., Y. Hadas, et al. (2009). "Transcriptional control of axonal guidance and sorting in dorsal interneurons by the Lim-­‐HD proteins Lhx9 and Lhx1." Neural Dev 4: 21. Ayoob, J. C., H. H. Yu, et al. (2004). "The Drosophila receptor guanylyl cyclase Gyc76C is required for semaphorin-­‐1a-­‐plexin A-­‐mediated axonal repulsion." J Neurosci 24(30): 6639-­‐6649. Bai, G., O. Chivatakarn, et al. (2011). "Presenilin-­‐dependent receptor processing is required for axon guidance." Cell 144(1): 106-­‐118. Ballif, B. A., Z. Cao, et al. (2006). "Identification of 14-­‐3-­‐3epsilon substrates from embryonic murine brain." Journal of Proteome Research 5(9): 2372-­‐2379. Barallobre, M. J., J. A. Del Rio, et al. (2000). "Aberrant development of hippocampal circuits and altered neural activity in netrin 1-­‐deficient mice." Development 127(22): 4797-­‐
4810. Bartoe, J. L., W. L. McKenna, et al. (2006). "Protein interacting with C -­‐kinase 1/protein kinase Calpha-­‐mediated endocytosis converts netrin-­‐1-­‐mediated repulsion to attraction." J Neurosci 26(12): 3192-­‐3205. Barzi, M., J. Berenguer, et al. (2010). "Sonic-­‐hedgehog-­‐mediated proliferation requires the localization of PKA to the cilium base." J Cell Sci 123(Pt 1): 62-­‐69. Bashaw, G. J. and R. Klein (2010). "Signaling from axon guidance receptors." Cold Spring Harb Perspect Biol 2(5): a001941. Baudet, M. L., K. H. Zivraj, et al. (2012). "miR-­‐124 acts through CoREST to control onset of Sema3A sensitivity in navigating retinal growth cones." Nat Neurosci 15(1): 29-­‐38. Bechara, A., J. Falk, et al. (2007). "Modulation of semaphorin signaling by Ig superfamily cell adhesion molecules." Adv Exp Med Biol 600: 61-­‐72. Ben-­‐Arie, N., B. A. Hassan, et al. (2000). "Functional conservation of atonal and Math1 in the CNS and PNS." Development 127(5): 1039-­‐1048. Benzing, T., M. B. Yaffe, et al. (2000). "14-­‐3-­‐3 interacts with regulator of G protein signaling proteins and modulates their activity." J Biol Chem 275(36): 28167-­‐28172. Berruti, G. (2000). "A Novel Rap1/B-­‐Raf/14-­‐3-­‐3 θ Protein Complex Is Formed in Vivo during the Morphogenetic Differentiation of Postmeiotic Male Germ Cells." Experimental Cell Research 257(1): 172-­‐179. Birkenfeld, J., H. Betz, et al. (2003). "Identification of cofilin and LIM-­‐domain-­‐containing protein kinase 1 as novel interaction partners of 14-­‐3-­‐3 zeta." Biochemical Journal 369(Pt 1): 45-­‐54. i Bibliography Bouchard, J.-­‐F., S. W. Moore, et al. (2004). "Protein Kinase A Activation Promotes Plasma Membrane Insertion of DCC from an Intracellular Pool: A Novel Mechanism Regulating Commissural Axon Extension." The Journal of Neuroscience 24(12): 3040-­‐3050. Bourikas, D., V. Pekarik, et al. (2005). "Sonic hedgehog guides commissural axons along the longitudinal axis of the spinal cord." Nat Neurosci 8(3): 297-­‐304. Briancon-­‐Marjollet, A., A. Ghogha, et al. (2008). "Trio mediates netrin-­‐1-­‐induced Rac1 activation in axon outgrowth and guidance." Mol Cell Biol 28(7): 2314-­‐2323. Bridges, D. and G. B. G. Moorhead (2005). "14-­‐3-­‐3 Proteins: A Number of Functions for a Numbered Protein." Sci. STKE 2005(296): re10-­‐. Brittis, P. A., Q. Lu, et al. (2002). "Axonal protein synthesis provides a mechanism for localized regulation at an intermediate target." Cell 110(2): 223-­‐235. Broadie, K., E. Rushton, et al. (1997). "Leonardo, a Drosophila 14-­‐3-­‐3 Protein Involved in Learning, Regulates Presynaptic Function." Neuron 19(2): 391-­‐402. Budillon, A., A. Cereseto, et al. (1995). "Point mutation of the autophosphorylation site or in the nuclear location signal causes protein kinase A RII beta regulatory subunit to lose its ability to revert transformed fibroblasts." Proc Natl Acad Sci U S A 92(23): 10634-­‐10638. Butler, S. J. and J. Dodd (2003). "A role for BMP heterodimers in roof plate-­‐mediated repulsion of commissural axons." Neuron 38(3): 389-­‐401. Butler, S. J. and G. Tear (2007). "Getting axons onto the right path: the role of transcription factors in axon guidance." Development 134(3): 439-­‐448. Cai, D., J. Qiu, et al. (2001). "Neuronal Cyclic AMP Controls the Developmental Loss in Ability of Axons to Regenerate." The Journal of Neuroscience 21(13): 4731-­‐4739. Cao, Z., Y. Gao, et al. (2006). "The cytokine interleukin-­‐6 is sufficient but not necessary to mimic the peripheral conditioning lesion effect on axonal growth." J Neurosci 26(20): 5565-­‐5573. Carnegie, G. K., J. Soughayer, et al. (2008). "AKAP-­‐Lbc mobilizes a cardiac hypertrophy signaling pathway." Mol Cell 32(2): 169-­‐179. Castellani, V., J. Falk, et al. (2004). "Semaphorin3A-­‐induced receptor endocytosis during axon guidance responses is mediated by L1 CAM." Mol Cell Neurosci 26(1): 89-­‐100. Chalasani, S. H., K. A. Sabelko, et al. (2003). "A chemokine, SDF-­‐1, reduces the effectiveness of multiple axonal repellents and is required for normal axon pathfinding." J Neurosci 23(4): 1360-­‐1371. Chang, H. C. and G. M. Rubin (1997). "14-­‐3-­‐3 epsilon positively regulates Ras-­‐mediated signaling in Drosophila." Genes Dev. 11(9): 1132-­‐1139. Charron, F., E. Stein, et al. (2003). "The morphogen sonic hedgehog is an axonal chemoattractant that collaborates with netrin-­‐1 in midline axon guidance." Cell 113(1): 11-­‐23. Chauvet, S., S. Cohen, et al. (2007). "Gating of Sema3E/PlexinD1 signaling by neuropilin-­‐1 switches axonal repulsion to attraction during brain development." Neuron 56(5): 807-­‐822. Chen, S., S. Synowsky, et al. (2011). "The capture of phosphoproteins by 14-­‐3-­‐3 proteins mediates actions of insulin." Trends Endocrinol Metab 22(11): 429-­‐436. Chen, Z., B. B. Gore, et al. (2008). "Alternative splicing of the Robo3 axon guidance receptor governs the midline switch from attraction to repulsion." Neuron 58(3): 325-­‐332. Choi, K. R., M. Berrera, et al. (2012). "Rapsyn mediates subsynaptic anchoring of PKA type I and stabilisation of acetylcholine receptor in vivo." J Cell Sci 125(Pt 3): 714-­‐723. Chuang, P. T., T. Kawcak, et al. (2003). "Feedback control of mammalian Hedgehog signaling by the Hedgehog-­‐binding protein, Hip1, modulates Fgf signaling during branching morphogenesis of the lung." Genes Dev 17(3): 342-­‐347. Colavita, A. and J. G. Culotti (1998). "Suppressors of ectopic UNC-­‐5 growth cone steering identify eight genes involved in axon guidance in Caenorhabditis elegans." Dev Biol 194(1): 72-­‐85. ii Bibliography Corset, V., K. T. Nguyen-­‐Ba-­‐Charvet, et al. (2000). "Netrin-­‐1-­‐mediated axon outgrowth and cAMP production requires interaction with adenosine A2b receptor." Nature 407(6805): 747-­‐750. Dalton, G. D. and W. L. Dewey (2006). "Protein kinase inhibitor peptide (PKI): A family of endogenous neuropeptides that modulate neuronal cAMP-­‐dependent protein kinase function." Neuropeptides 40(1): 23-­‐34. Danielian, P. S., D. Muccino, et al. (1998). "Modification of gene activity in mouse embryos in utero by a tamoxifen-­‐inducible form of Cre recombinase." Curr Biol 8(24): 1323-­‐
1326. Davare, M. A., T. Saneyoshi, et al. (2004). "Inhibition of Calcium/Calmodulin-­‐dependent Protein Kinase Kinase by Protein 14-­‐3-­‐3." Journal of Biological Chemistry 279(50): 52191-­‐52199. Deakin, N. O., M. D. Bass, et al. (2009). "An integrin-­‐alpha4-­‐14-­‐3-­‐3zeta-­‐paxillin ternary complex mediates localised Cdc42 activity and accelerates cell migration." J Cell Sci 122(Pt 10): 1654-­‐1664. Deiner, M. S., T. E. Kennedy, et al. (1997). "Netrin-­‐1 and DCC mediate axon guidance locally at the optic disc: loss of function leads to optic nerve hypoplasia." Neuron 19(3): 575-­‐
589. Dent, E. W., S. L. Gupton, et al. (2011). "The growth cone cytoskeleton in axon outgrowth and guidance." Cold Spring Harb Perspect Biol 3(3). Dessaud, E., A. P. McMahon, et al. (2008). "Pattern formation in the vertebrate neural tube: a sonic hedgehog morphogen-­‐regulated transcriptional network." Development 135(15): 2489-­‐2503. Dickson, B. J. and Y. Zou (2010). "Navigating intermediate targets: the nervous system midline." Cold Spring Harb Perspect Biol 2(8): a002055. Diviani, D., L. Abuin, et al. (2004). "Anchoring of both PKA and 14-­‐3-­‐3 inhibits the Rho-­‐GEF activity of the AKAP-­‐Lbc signaling complex." EMBO J 23(14): 2811-­‐2820. Domanitskaya, E., A. Wacker, et al. (2010). "Sonic hedgehog guides post-­‐crossing commissural axons both directly and indirectly by regulating Wnt activity." J Neurosci 30(33): 11167-­‐11176. Dong, S., S. Kang, et al. (2008). "Targeting 14-­‐3-­‐3 sensitizes native and mutant BCR-­‐ABL to inhibition with U0126, rapamycin and Bcl-­‐2 inhibitor GX15-­‐070." Leukemia 22(3): 572-­‐577. Dontchev, V. D. and P. C. Letourneau (2002). "Nerve growth factor and semaphorin 3A signaling pathways interact in regulating sensory neuronal growth cone motility." J Neurosci 22(15): 6659-­‐6669. Dransfield, D. T., A. J. Bradford, et al. (1997). "Ezrin is a cyclic AMP-­‐dependent protein kinase anchoring protein." EMBO J 16(1): 35-­‐43. Eiseler, T., H. Doppler, et al. (2009). "Protein kinase D1 regulates cofilin-­‐mediated F-­‐actin reorganization and cell motility through slingshot." Nat Cell Biol 11(5): 545-­‐556. Eiseler, T., A. Hausser, et al. (2010). "Protein kinase D controls actin polymerization and cell motility through phosphorylation of cortactin." J Biol Chem 285(24): 18672-­‐18683. Erskine, L. and E. Herrera (2007). "The retinal ganglion cell axon's journey: insights into molecular mechanisms of axon guidance." Dev Biol 308(1): 1-­‐14. Evans, T. A. and G. J. Bashaw (2010). "Axon guidance at the midline: of mice and flies." Curr Opin Neurobiol 20(1): 79-­‐85. Faas, G. C., S. Raghavachari, et al. (2011). "Calmodulin as a direct detector of Ca2+ signals." Nat Neurosci 14(3): 301-­‐304. Fabre, P. J., T. Shimogori, et al. (2010). "Segregation of ipsilateral retinal ganglion cell axons at the optic chiasm requires the Shh receptor Boc." J Neurosci 30(1): 266-­‐275. Fambrough, D., D. Pan, et al. (1996). "The cell surface metalloprotease/disintegrin Kuzbanian is required for axonal extension in Drosophila." Proc Natl Acad Sci U S A 93(23): 13233-­‐13238. iii Bibliography Farmer, W. T., A. L. Altick, et al. (2008). "Pioneer longitudinal a xons navigate using floor plate and Slit/Robo signals." Development 135(22): 3643-­‐3653. Feng, L., S. Yunoue, et al. (2004). "PKA phosphorylation and 14-­‐3-­‐3 interaction regulate the function of neurofibromatosis type I tumor suppressor, neurofibromin." FEBS Lett 557(1-­‐3): 275-­‐282. Fiedler, S. E., R. V. Schillace, et al. (2010). "Myeloid translocation gene 16b is a dual A -­‐kinase anchoring protein that interacts selectively with plexins in a phospho-­‐regulated manner." FEBS Lett 584(5): 873-­‐877. Filbin, M. T. (2003). "Myelin-­‐associated inhibitors of axonal regeneration in the adult mammalian CNS." Nat Rev Neurosci 4(9): 703-­‐713. Forbes, E. M., A. W. Thompson, et al. (2012). "Calcium and cAMP Levels Interact to Determine Attraction versus Repulsion in Axon Guidance." Neuron 74(3): 490-­‐503. Forcet, C., E. Stein, et al. (2002). "Netrin-­‐1-­‐mediated axon outgrowth requires deleted in colorectal cancer-­‐dependent MAPK activation." Nature 417(6887): 443-­‐447. Fournier, A. E., F. Nakamura, et al. (2000). "Semaphorin3A enhances endocytosis at sites of receptor-­‐F-­‐actin colocalization during growth cone collapse." J Cell Biol 149(2): 411-­‐422. Freeman, A. K. and D. K. Morrison (2011). "14-­‐3-­‐3 Proteins: diverse functions in cell proliferation and cancer progression." Semin Cell Dev Biol 22(7): 681-­‐687. Galko, M. J. and M. Tessier-­‐Lavigne (2000). "Function of an axonal chemoattractant modulated by metalloprotease activity." Science 289(5483): 1365-­‐1367. Gallo, G. (2006). "RhoA-­‐kinase coordinates F-­‐actin organization and myosin II activity during semaphorin-­‐3A-­‐induced axon retraction." J Cell Sci 119(Pt 16): 3413-­‐3423. Gao, Y., K. Deng, et al. (2004). "Activated CREB is sufficient to overcome inhibitors in myelin and promote spinal axon regeneration in vivo." Neuron 44(4): 609-­‐621. Garbe, D. S. and G. J. Bashaw (2004). "Axon Guidance at the Midline: From Mutants to Mechanisms." Critical Reviews in Biochemistry & Molecular Biology 39(5/6): 319-­‐
341. Garbe, D. S. and G. J. Bashaw (2007). "Independent functions of Slit-­‐Robo repulsion and Netrin-­‐Frazzled attraction regulate axon crossing at the midline in Drosophila." J Neurosci 27(13): 3584-­‐3592. Garbe, D. S., M. O'Donnell, et al. (2007). "Cytoplasmic domain requirements for Frazzled-­‐
mediated attractive axon turning at the Drosophila midline." Development 134(24): 4325-­‐4334. Gardino, A. K., S. J. Smerdon, et al. (2006). "Structural determinants of 14-­‐3-­‐3 binding specificities and regulation of subcellular localization of 14-­‐3-­‐3-­‐ligand complexes: a comparison of the X-­‐ray crystal structures of all human 14-­‐3-­‐3 isoforms." Semin Cancer Biol 16(3): 173-­‐182. Gehler, S., G. Gallo, et al. (2004). "p75 Neurotrophin Receptor Signaling Regulates Growth Cone Filopodial Dynamics through Modulating RhoA Activity." The Journal of Neuroscience 24(18): 4363-­‐4372. Gehler, S., A. E. Shaw, et al. (2004). "Brain-­‐Derived Neurotrophic Factor Regulation of Retinal Growth Cone Filopodial Dynamics Is Mediated through Actin Depolymerizing Factor/Cofilin." J. Neurosci. 24(47): 10741-­‐10749. Georgiou, M. and G. Tear (2002). "Commissureless is required both in commissural neurones and midline cells for axon guidance across the midline." Development 129(12): 2947-­‐2956. Gervasi, N., P. Tchenio, et al. (2010). "PKA dynamics in a Drosophila learning center: coincidence detection by rutabaga adenylyl cyclase and spatial regulation by dunce phosphodiesterase." Neuron 65(4): 516-­‐529. Giger, R. J., E. R. Hollis, 2nd, et al. (2010). "Guidance molecules in axon regeneration." Cold Spring Harb Perspect Biol 2(7): a001867. iv Bibliography Gitai, Z., T. W. Yu, et al. (2003). "The netrin receptor UNC-­‐40/DCC stimulates axon attraction and outgrowth through enabled and, in parallel, Rac and UNC-­‐115/AbLIM." Neuron 37(1): 53-­‐65. Gohla, A. and G. M. Bokoch (2002). "14-­‐3-­‐3 regulates actin dynamics by stabilizing phosphorylated cofilin." Curr Biol 12(19): 1704-­‐1710. Gold, M. G., B. Lygren, et al. (2006). "Molecular basis of AKAP specificity for PKA regulatory subunits." Mol Cell 24(3): 383-­‐395. Goldberg, J. L., M. P. Klassen, et al. (2002). "Amacrine-­‐signaled loss of intrinsic axon growth ability by retinal ganglion cells." Science 296(5574): 1860-­‐1864. Gomez, T. M. and N. C. Spitzer (1999). "In vivo regulation of axon extension and pathfinding by growth-­‐cone calcium transients." Nature 397(6717): 350-­‐355. Gore, B. B., K. G. Wong, et al. (2008). "Stem cell factor functions as an outgrowth-­‐promoting factor to enable axon exit from the midline intermediate target." Neuron 57(4): 501-­‐
510. Gowan, K., A. W. Helms, et al. (2001). "Crossinhibitory activities of Ngn1 and Math1 allow specification of distinct dorsal interneurons." Neuron 31(2): 219-­‐232. Gu, Y. M., Y. H. Jin, et al. (2006). "Protein kinase A phosphorylates and regulates dimerization of 14-­‐3-­‐3 zeta." FEBS letters 580(1): 305-­‐310. Guirland, C., S. Suzuki, et al. (2004). "Lipid rafts mediate chemotropic guidance of nerve growth cones." Neuron 42(1): 51-­‐62. Gundersen, R. W. and J. N. Barrett (1980). "Characterization of the turning response of dorsal root neurites toward nerve growth factor." The Journal of Cell Biology 87(3): 546-­‐
554. Hall, A. and G. Lalli (2010). "Rho and Ras GTPases in axon growth, guidance, and branching." Cold Spring Harb Perspect Biol 2(2): a001818. Han, J., L. Han, et al. (2007). "Spatial targeting of type II protein kinase A to filopodia mediates the regulation of growth cone guidance by cAMP." The Journal of Cell Biology 176(1): 101-­‐111. Hattori, M., M. Osterfield, et al. (2000). "Regulated cleavage of a contact-­‐mediated axon repellent." Science 289(5483): 1360-­‐1365. Henle, S. J., G. Wang, et al. (2011). "Asymmetric PI(3,4,5)P3 and Akt signaling mediates chemotaxis of axonal growth cones." J Neurosci 31(19): 7016-­‐7027. Henley, J. R., K.-­‐h. Huang, et al. (2004). "Calcium Mediates Bidirectional Growth Cone Turning Induced by Myelin-­‐Associated Glycoprotein." Neuron 44(6): 909-­‐916. Herincs, Z., V. Corset, et al. (2005). "DCC association with lipid rafts is required for netrin-­‐1-­‐
mediated axon guidance." J Cell Sci 118(Pt 8): 1687-­‐1692. Herrera, E., L. Brown, et al. (2003). "Zic2 patterns binocular vision by specifying the uncrossed retinal projection." Cell 114(5): 545-­‐557. Herrera, E., R. Marcus, et al. (2004). "Foxd1 is required for proper formation of the optic chiasm." Development 131(22): 5727-­‐5739. Hippenmeyer, S., Y. H. Youn, et al. (2010). "Genetic mosaic dissection of Lis1 and Ndel1 in neuronal migration." Neuron 68(4): 695-­‐709. Hong, K., L. Hinck, et al. (1999). "A ligand-­‐gated association between cytoplasmic domains of UNC5 and DCC family receptors converts netrin-­‐induced growth cone attraction to repulsion." Cell 97(7): 927-­‐941. Hong, K., M. Nishiyama, et al. (2000). "Calcium signalling in the guidance of nerve growth by netrin-­‐1." Nature 403(6765): 93-­‐98. Hopker, V. H., D. Shewan, et al. (1999). "Growth-­‐cone attraction to netrin-­‐1 is converted to repulsion by laminin-­‐1." Nature 401(6748): 69-­‐73. Hsieh, S. H. K., G. B. Ferraro, et al. (2006). "Myelin-­‐Associated Inhibitors Regulate Cofilin Phosphorylation and Neuronal Inhibition through LIM Kinase and Slingshot Phosphatase." J. Neurosci. 26(3): 1006-­‐1015. v Bibliography Huber, A. B., A. L. Kolodkin, et al. (2003). "SIGNALING AT THE GROWTH CONE: Ligand-­‐
Receptor Complexes and the Control of Axon Growth and Guidance." Annu Rev Neurosci 26(1): 509-­‐563. Hur, E. M., Saijilafu, et al. (2012). "Growing the growth cone: remodeling the cytoskeleton to promote axon regeneration." Trends Neurosci 35(3): 164-­‐174. Ikeda, M., T. Hikita, et al. (2008). "Identification of YWHAE, a gene encoding 14-­‐3-­‐3epsilon, as a possible susceptibility gene for schizophrenia." Hum Mol Genet 17(20): 3212-­‐
3222. Islam, S. M., Y. Shinmyo, et al. (2009). "Draxin, a repulsive guidance protein for spinal cord and forebrain commissures." Science 323(5912): 388-­‐393. Janes, P. W., N. Saha, et al. (2005). "Adam meets Eph: an ADAM substrate recognition module acts as a molecular switch for ephrin cleavage in trans." Cell 123(2): 291-­‐304. Jin, J., F. D. Smith, et al. (2004). "Proteomic, Functional, and Domain-­‐Based Analysis of In Vivo 14-­‐3-­‐3 Binding Proteins Involved in Cytoskeletal Regulation and Cellular Organization." Current biology : CB 14(16): 1436-­‐1450. Jin, Z. and S. M. Strittmatter (1997). "Rac1 mediates collapsin-­‐1-­‐induced growth cone collapse." J Neurosci 17(16): 6256-­‐6263. Jones, D. H., S. Ley, et al. (1995). "Isoforms of 14-­‐3-­‐3 protein can form homo-­‐ and heterodimers in vivo and in vitro: implications for function as adapter proteins." FEBS Letters 368(1): 55-­‐58. Jung, H., B. C. Yoon, et al. (2012). "Axonal mRNA localization and local protein synthesis in nervous system assembly, maintenance and repair." Nat Rev Neurosci 13(5): 308-­‐
324. Kania, A. and T. M. Jessell (2003). "Topographic motor projections in the limb imposed by LIM homeodomain protein regulation of ephrin-­‐A:EphA interactions." Neuron 38(4): 581-­‐596. Kantor, D. B., O. Chivatakarn, et al. (2004). "Semaphorin 5A is a bifunctional axon guidance cue regulated by heparan and chondroitin sulfate proteoglycans." Neuron 44(6): 961-­‐975. Kase, H., K. Iwahashi, et al. (1987). "K-­‐252 compounds, novel and potent inhibitors of protein kinase C and cyclic nucleotide-­‐dependent protein kinases." Biochemical and Biophysical Research Communications 142(2): 436-­‐440. Keino-­‐Masu, K., M. Masu, et al. (1996). "Deleted in Colorectal Cancer (DCC) encodes a netrin receptor." Cell 87(2): 175-­‐185. Keleman, K. and B. J. Dickson (2001). "Short-­‐ and long-­‐range repulsion by the Drosophila Unc5 netrin receptor." Neuron 32(4): 605-­‐617. Keleman, K., S. Rajagopalan, et al. (2002). "Comm sorts robo to control axon guidance at the Drosophila midline." Cell 110(4): 415-­‐427. Keleman, K., C. Ribeiro, et al. (2005). "Comm function in commissural axon guidance: cell-­‐
autonomous sorting of Robo in vivo." Nat Neurosci 8(2): 156-­‐163. Kennedy, T. E., T. Serafini, et al. (1994). "Netrins are diffusible chemotropic factors for commissural axons in the embryonic spinal cord." Cell 78(3): 425-­‐435. Kent, C. B., T. Shimada, et al. (2010). "14-­‐3-­‐3 proteins regulate protein kinase a activity to modulate growth cone turning responses." J Neurosci 30(42): 14059-­‐14067. Kidd, T., K. S. Bland, et al. (1999). "Slit is the midline repellent for the robo receptor in Drosophila." Cell 96(6): 785-­‐794. Kim, J. S., T. Y. Huang, et al. (2009). "Reactive oxygen species regulate a slingshot-­‐cofilin activation pathway." Mol Biol Cell 20(11): 2650-­‐2660. Kim, T.-­‐H., H. K. Lee, et al. (2005). "Netrin induces down-­‐regulation of its receptor, Deleted in Colorectal Cancer, through the ubiquitin–proteasome pathway in the embryonic cortical neuron." Journal of Neurochemistry 95(1): 1-­‐8. Kligys, K., J. N. Claiborne, et al. (2007). "The slingshot family of phosphatases mediates Rac1 regulation of cofilin phosphorylation, laminin-­‐332 organization and motility behavior of keratinocytes." J. Biol. Chem.: M707041200. vi Bibliography Kligys, K., J. Yao, et al. (2009). "14-­‐3-­‐3zeta/tau heterodimers regulate Slingshot activity in migrating keratinocytes." Biochem Biophys Res Commun 383(4): 450-­‐454. Kobayashi, H., Y. Ogura, et al. (2011). "Involvement of 14-­‐3-­‐3 proteins in the second epidermal growth factor-­‐induced wave of Rac1 activation in the process of cell migration." J Biol Chem 286(45): 39259-­‐39268. Koga, Y. and M. Ikebe (2008). "A Novel Regulatory Mechanism of Myosin Light Chain Phosphorylation via Binding of 14-­‐3-­‐3 to Myosin Phosphatase." Molecular Biology of the Cell 19(3): 1062-­‐1071. Kolodkin, A. L. and M. Tessier-­‐Lavigne (2011). "Mechanisms and molecules of neuronal wiring: a primer." Cold Spring Harb Perspect Biol 3(6). Kolpak, A., J. Zhang, et al. (2005). "Sonic hedgehog has a dual effect on the growth of retinal ganglion axons depending on its concentration." J Neurosci 25(13): 3432-­‐3441. Kuhn, T. B., M. D. Brown, et al. (1999). "Myelin and collapsin-­‐1 induce motor neuron growth cone collapse through different pathways: inhibition of collapse by opposing mutants of rac1." J Neurosci 19(6): 1965-­‐1975. Kullander, K., S. J. Butt, et al. (2003). "Role of EphA4 and EphrinB3 in local neuronal circuits that control walking." Science 299(5614): 1889-­‐1892. Kuwako, K., K. Kakumoto, et al. (2010). "Neural RNA-­‐binding protein Musashi1 controls midline crossing of precerebellar neurons through posttranscriptional regulation of Robo3/Rig-­‐1 expression." Neuron 67(3): 407-­‐421. Langlois, S. D., S. Morin, et al. (2010). "Dissection and culture of commissural neurons from embryonic spinal cord." J Vis Exp(39). Lau, J. M., Wu, C., Muslin, A.M. (2006). "Differential role of 14-­‐3-­‐3 family members in Xenopus development." Developmental Dynamics 235(7): 1761-­‐1776. Lebrand, C., E. W. Dent, et al. (2004). "Critical role of Ena/VASP proteins for filopodia formation in neurons and in function downstream of netrin-­‐1." Neuron 42(1): 37-­‐
49. Leonardo, E. D., L. Hinck, et al. (1997). "Vertebrate homologues of C. elegans UNC-­‐5 are candidate netrin receptors." Nature 386(6627): 833-­‐838. Levy-­‐Strumpf, N. and J. G. Culotti (2007). "VAB-­‐8, UNC-­‐73 and MIG-­‐2 regulate axon polarity and cell migration functions of UNC-­‐40 in C. elegans." Nat Neurosci 10(2): 161-­‐168. Li, W., J. Lee, et al. (2004). "Activation of FAK and Src are receptor-­‐proximal events required for netrin signaling." Nat Neurosci 7(11): 1213-­‐1221. Li, X., E. Saint-­‐Cyr-­‐Proulx, et al. (2002). "Rac1 and Cdc42 but not RhoA or Rho kinase activities are required for neurite outgrowth induced by the Netrin-­‐1 receptor DCC (deleted in colorectal cancer) in N1E-­‐115 neuroblastoma cells." J Biol Chem 277(17): 15207-­‐15214. Li, Y., Y. C. Jia, et al. (2005). "Essential role of TRPC channels in the guidance of nerve growth cones by brain-­‐derived neurotrophic factor." Nature 434(7035): 894-­‐898. Li, Y., Y. Wu, et al. (2006). "Modulation of inactivation properties of CaV2.2 channels by 14-­‐3-­‐
3 proteins." Neuron 51(6): 755-­‐771. Lignitto, L., A. Carlucci, et al. (2011). "Control of PKA stability and signalling by the RING ligase praja2." Nat Cell Biol 13(4): 412-­‐422. Lin, A. C. and C. E. Holt (2007). "Local translation and directional s teering in axons." EMBO J 26(16): 3729-­‐3736. Liu, D., J. Bienkowska, et al. (1995). "Crystal structure of the zeta isoform of the 14-­‐3-­‐3 protein." Nature 376(6536): 191-­‐194. Liu, G., H. Beggs, et al. (2004). "Netrin requires focal adhesion kinase and Src family kinases for axon outgrowth and attraction." Nat Neurosci 7(11): 1222-­‐1232. Liu, G., W. Li, et al. (2009). "DSCAM functions as a netrin receptor in commissural axon pathfinding." Proc Natl Acad Sci U S A 106(8): 2951-­‐2956. Logue, J. S. and J. D. Scott (2010). "Organizing signal transduction through A -­‐kinase anchoring proteins (AKAPs)." FEBS J 277(21): 4370-­‐4375. vii Bibliography Long, F., X. M. Zhang, et al. (2001). "Genetic manipulation of hedgehog signaling in the endochondral skeleton reveals a direct role in the regulation of chondrocyte proliferation." Development 128(24): 5099-­‐5108. Long, H., C. Sabatier, et al. (2004). "Conserved roles for Slit and Robo proteins in midline commissural axon guidance." Neuron 42(2): 213-­‐223. Low, K., M. Culbertson, et al. (2008). "Netrin-­‐1 is a novel myelin-­‐associated inhibitor to axon growth." J Neurosci 28(5): 1099-­‐1108. Lumpkin, E. A., T. Collisson, et al. (2003). "Math1-­‐driven GFP expression in the developing nervous system of transgenic mice." Gene Expr Patterns 3(4): 389-­‐395. Luria, V., D. Krawchuk, et al. (2008). "Specification of motor axon trajectory by ephrin-­‐
B:EphB signaling: symmetrical control of axonal patterning in the developing limb." Neuron 60(6): 1039-­‐1053. Ly, A., A. Nikolaev, et al. (2008). "DSCAM is a netrin receptor that collaborates with DCC in mediating turning responses to netrin-­‐1." Cell 133(7): 1241-­‐1254. Lyuksyutova, A. I., C. C. Lu, et al. (2003). "Anterior-­‐posterior guidance of commissural axons by Wnt-­‐frizzled signaling." Science 302(5652): 1984-­‐1988. Ma, T. C., A. Campana, et al. (2010). "A large-­‐scale chemical screen for regulators of the arginase 1 promoter identifies the soy isoflavone daidzeinas a clinically approved small molecule that can promote neuronal protection or regeneration via a cAMP-­‐
independent pathway." J Neurosci 30(2): 739-­‐748. Mackintosh, C. (2004). "Dynamic interactions between 14-­‐3-­‐3 proteins and phosphoproteins regulate diverse cellular processes." Biochemical Journal 381(Pt 2): 329-­‐342. Manni, S., J. H. Mauban, et al. (2008). "Phosphorylation of the cAMP-­‐dependent protein kinase (PKA) regulatory subunit modulates PKA-­‐AKAP interaction, substrate phosphorylation, and calcium signaling in cardiac cells." J Biol Chem 283(35): 24145-­‐24154. Marin, O. and J. L. Rubenstein (2003). "Cell migration in the forebrain." Annu Rev Neurosci 26: 441-­‐483. Matei, V., S. Pauley, et al. (2005). "Smaller inner ear sensory epithelia in Neurog 1 null mice are related to earlier hair cell cycle exit." Dev Dyn 234(3): 633-­‐650. Matise, M. P., M. Lustig, et al. (1999). "Ventral midline cells are required for the local control of commissural axon guidance in the mouse spinal cord." Development 126(16): 3649-­‐3659. McConnachie, G., L. K. Langeberg, et al. (2006). "AKAP signaling complexes: getting to the heart of the matter." Trends Mol Med 12(7): 317-­‐323. McKenzie, A. J., S. L. Campbell, et al. (2011). "Protein kinase A activity and anchoring are required for ovarian cancer cell migration and invasion." PLoS One 6(10): e26552. Meek, S. E., W. S. Lane, et al. (2004). "Comprehensive proteomic analysis of interphase and mitotic 14-­‐3-­‐3-­‐binding proteins." Journal of Biological Chemistry 279(31): 32046-­‐
32054. Meiri, D., M. A. Greeve, et al. (2009). "Modulation of Rho Guanine Exchange Factor Lfc Activity by Protein Kinase A-­‐Mediated Phosphorylation." Molecular and Cellular Biology 29(21): 5963-­‐5973. Meriane, M., J. Tcherkezian, et al. (2004). "Phosphorylation of DCC by Fyn mediates Netrin-­‐1 signaling in growth cone guidance." J Cell Biol 167(4): 687-­‐698. Ming, G., J. Henley, et al. (2001). "Electrical activity modulates growth cone guidance by diffusible factors." Neuron 29(2): 441-­‐452. Ming, G.-­‐l., H.-­‐j. Song, et al. (1997). "cAMP-­‐Dependent Growth Cone Guidance by Netrin-­‐1." Neuron 19(6): 1225-­‐1235. Mo, R., A. M. Freer, et al. (1997). "Specific and redundant functions of Gli2 and Gli3 zinc finger genes in skeletal patterning and development." Development 124(1): 113-­‐123. Moore, B. W. a. P., V. J. (1967). Specific Acid Proteins in the Nervous System. Englewood Cliffs, New Jersey, Prentice-­‐Hall. viii Bibliography Moore, S. W., J. P. Correia, et al. (2008). "Rho inhibition recruits DCC to the neuronal plasma membrane and enhances axon chemoattraction to netrin 1." Development 135(17): 2855-­‐2864. Moore, S. W. and T. E. Kennedy (2006). "Protein kinase A regulates the sensitivity of spinal commissural axon turning to netrin-­‐1 but does not switch between chemoattraction and chemorepulsion." J Neurosci 26(9): 2419-­‐2423. Murray, A. J. (2008). "Pharmacological PKA Inhibition: All May Not Be What It Seems." Sci. Signal. 1(22): re4-­‐. Murray, A. J. and D. A. Shewan (2008). "Epac mediates cyclic AMP-­‐dependent axon growth, guidance and regeneration." Molecular and Cellular Neuroscience 38(4): 578-­‐588. Murray, A. J., S. J. Tucker, et al. (2009). "cAMP-­‐dependent axon guidance is distinctly regulated by Epac and protein kinase A." J Neurosci 29(49): 15434-­‐15444. Muslin, A. J. and H. Xing (2000). "14-­‐3-­‐3 proteins: regulation of subcellular localization by molecular interference." Cellular Signalling 12(11–12): 703-­‐709. Myers, J. P., E. Robles, et al. (2012). "Focal adhesion kinase modulates Cdc42 activity downstream of positive and negative axon guidance cues." J Cell Sci 125(Pt 12): 2918-­‐2929. Nagata-­‐Ohashi, K., Y. Ohta, et al. (2004). "A pathway of neuregulin-­‐induced activation of cofilin-­‐phosphatase Slingshot and cofilin in lamellipodia." The Journal of Cell Biology 165(4): 465-­‐471. Nagata-­‐Ohashi, K., Y. Ohta, et al. (2004). "A pathway of neuregulin-­‐induced activation of cofilin-­‐phosphatase Slingshot and cofilin in lamellipodia." Journal of Cell Biology 165(4): 465-­‐471. Nawabi, H., A. Briancon-­‐Marjollet, et al. (2010). "A midline switch of receptor processing regulates commissural axon guidance in vertebrates." Genes Dev 24(4): 396-­‐410. Neumann, S., F. Bradke, et al. (2002). "Regeneration of Sensory Axons within the Injured Spinal Cord Induced by Intraganglionic cAMP Elevation." Neuron 34(6): 885-­‐893. Neve, R. L., J. R. Howe, et al. (1997). "Introduction of the glutamate receptor subunit 1 into motor neurons in vitro and in vivo using a recombinant herpes simplex virus." Neuroscience 79(2): 435-­‐447. Ni, Q., A. Ganesan, et al. (2011). "Signaling diversity of PKA achieved via a Ca2+-­‐cAMP-­‐PKA oscillatory circuit." Nat Chem Biol 7(1): 34-­‐40. Nicol, X., K. P. Hong, et al. (2011). "Spatial and temporal second messenger codes for growth cone turning." Proc Natl Acad Sci U S A 108(33): 13776-­‐13781. Nishita, M., C. Tomizawa, et al. (2005). "Spatial and temporal regulation of cofilin activity by LIM kinase and Slingshot is critical for directional cell migration." J Cell Biol 171(2): 349-­‐359. Nishiyama, M., A. Hoshino, et al. (2003). "Cyclic AMP/GMP-­‐dependent modulation of Ca2+ channels sets the polarity of nerve growth-­‐cone turning." Nature 423(6943): 990-­‐
995. Nishiyama, M., M. J. von Schimmelmann, et al. (2008). "Membrane potential shifts caused by diffusible guidance signals direct growth-­‐cone turning." Nat Neurosci 11(7): 762-­‐
771. Nozumi, M., T. Togano, et al. (2009). "Identification of functional marker proteins in the mammalian growth cone." Proc Natl Acad Sci U S A 106(40): 17211-­‐17216. O'Toole, T. E., K. Bialkowska, et al. (2011). "Tiam1 is recruited to beta1-­‐integrin complexes by 14-­‐3-­‐3zeta where it mediates integrin-­‐induced Rac1 activation and motility." J Cell Physiol 226(11): 2965-­‐2978. Ohashi, K., S. Fujiwara, et al. (2011). "LIM kinase has a dual role in regulating lamellipodium extension by decelerating the rate of actin retrograde flow and the rate of actin polymerization." J Biol Chem 286(42): 36340-­‐36351. Ooashi, N., A. Futatsugi, et al. (2005). "Cell adhesion molecules regulate Ca2+-­‐mediated steering of growth cones via cyclic AMP and ryanodine receptor type 3." J Cell Biol 170(7): 1159-­‐1167. ix Bibliography Parra, L. M. and Y. Zou (2010). "Sonic hedgehog induces response of commissural axons to Semaphorin repulsion during midline crossing." Nat Neurosci 13(1): 29-­‐35. Paulucci-­‐Holthauzen, A. A., L. A. Vergara, et al. (2009). "Spatial distribution of protein kinase A activity during cell migration is mediated by A-­‐kinase anchoring protein AKAP Lbc." J Biol Chem 284(9): 5956-­‐5967. Pearse, D. D., F. C. Pereira, et al. (2004). "cAMP and Schwann cells promote axonal growth and functional recovery after spinal cord injury." Nat Med 10(6): 610-­‐616. Peterburs, P., J. Heering, et al. (2009). "Protein kinase D regulates cell migration by direct phosphorylation of the cofilin phosphatase slingshot 1 like." Cancer Res 69(14): 5634-­‐5638. Petrie, R. J., B. Zhao, et al. (2009). "Compartmentalized DCC signalling is distinct from DCC localized to lipid rafts." Biol Cell 101(2): 77-­‐90. Petros, T. J., A. Rebsam, et al. (2008). "Retinal axon growth at the optic chiasm: to cross or not to cross." Annu Rev Neurosci 31: 295-­‐315. Philip, N., S. F. Acevedo, et al. (2001). "Conditional rescue of olfactory learning and memory defects in mutants of the 14-­‐3-­‐3zeta gene leonardo." J Neurosci 21(21): 8417-­‐8425. Piper, M., S. Salih, et al. (2005). "Endocytosis-­‐dependent desensitization and protein synthesis-­‐dependent resensitization in retinal growth cone adaptation." Nat Neurosci 8(2): 179-­‐186. Polleux, F., T. Morrow, et al. (2000). "Semaphorin 3A is a chemoattractant for cortical apical dendrites." Nature 404(6778): 567-­‐573. Pozuelo Rubio, M., K. M. Geraghty, et al. (2004). "14-­‐3-­‐3-­‐affinity purification of over 200 human phosphoproteins reveals new links to regulation of cellular metabolism, proliferation and trafficking.[see comment]." Biochemical Journal 379(Pt 2): 395-­‐
408. Pulina, M. V., R. Rizzuto, et al. (2006). "Inhibitory interaction of the plasma membrane Na+/Ca2+ exchangers with the 14-­‐3-­‐3 proteins." J Biol Chem 281(28): 19645-­‐
19654. Qiu, J., D. Cai, et al. (2002). "Spinal axon regeneration induced by elevation of cyclic AMP." Neuron 34(6): 895-­‐903. Qiu, W., S. Zhuang, et al. (2000). "Cell type-­‐specific regulation of B-­‐Raf kinase by cAMP and 14-­‐3-­‐3 proteins." J Biol Chem 275(41): 31921-­‐31929. Rajagopalan, S., E. Nicolas, et al. (2000). "Crossing the midline: roles and regulation of Robo receptors." Neuron 28(3): 767-­‐777. Rajagopalan, S., V. Vivancos, et al. (2000). "Selecting a longitudinal pathway: Robo receptors specify the lateral position of axons in the Drosophila CNS." Cell 103(7): 1033-­‐1045. Ramon y Cajal, S. (1909). Histologie du System Nerveux de l’Homme et des Vertebres. Paris, Malonine. Ramser, E. M., F. Buck, et al. (2010). "Binding of alphaII spectrin to 14-­‐3-­‐3beta is involved in NCAM-­‐dependent neurite outgrowth." Mol Cell Neurosci 45(1): 66-­‐74. Ramser, E. M., G. Wolters, et al. (2010). "The 14-­‐3-­‐3zeta protein binds to the cell adhesion molecule L1, promotes L1 phosphorylation by CKII and influences L1-­‐dependent neurite outgrowth." PLoS One 5(10): e13462. Rittinger, K., J. Budman, et al. (1999). "Structural Analysis of 14-­‐3-­‐3 Phosphopeptide Complexes Identifies a Dual Role for the Nuclear Export Signal of 14-­‐3-­‐3 in Ligand Binding." Molecular Cell 4(2): 153-­‐166. Rivard, R. L., M. Birger, et al. (2009). "AKAP-­‐independent localization of type-­‐II protein kinase A to dynamic actin microspikes." Cell Motil Cytoskeleton 66(9): 693-­‐709. Robles, E., A. Huttenlocher, et al. (2003). "Filopodial calcium transients regulate growth cone motility and guidance through local activation of calpain." Neuron 38(4): 597-­‐609. Rosenquist, M., Sehnke, P., Ferl, R.J., Sommarin, M., Larsson, C. (2000). "Evolution of the 14-­‐3-­‐
3 Protein Family: Does the Large Number of Isoforms in Multicellular Organisms Reflect Functional Specificity?" Journal of Molecular Evolution 51: 446-­‐458. x Bibliography Roy, S., R. A. McPherson, et al. (1998). "14-­‐3-­‐3 facilitates Ras-­‐dependent Raf-­‐1 activation in vitro and in vivo." Mol Cell Biol 18(7): 3947-­‐3955. Ruiz de Almodovar, C., P. J. Fabre, et al. (2011). "VEGF mediates commissural axon chemoattraction through its receptor Flk1." Neuron 70(5): 966-­‐978. Sabatier, C., A. S. Plump, et al. (2004). "The divergent Robo family protein rig-­‐1/Robo3 is a negative regulator of slit responsiveness required for midline crossing by commissural axons." Cell 117(2): 157-­‐169. Sample, V., L. M. DiPilato, et al. (2012). "Regulation of nuclear PKA revealed by spatiotemporal manipulation of cyclic AMP." Nat Chem Biol 8(4): 375-­‐382. Sanchez-­‐Camacho, C. and P. Bovolenta (2008). "Autonomous and non-­‐autonomous Shh signalling mediate the in vivo growth and guidance of mouse retinal ganglion cell axons." Development 135(21): 3531-­‐3541. Santoro, M. M., G. Gaudino, et al. (2003). "The MSP receptor regulates alpha6beta4 and alpha3beta1 integrins via 14-­‐3-­‐3 proteins in keratinocyte migration." Dev Cell 5(2): 257-­‐271. Santoro, M. M., G. Gaudino, et al. (2003). "The MSP receptor regulates alpha6beta4 and alpha3beta1 integrins via 14-­‐3-­‐3 proteins in keratinocyte migration." Developmental Cell 5(2): 257-­‐271. Seeger, M., G. Tear, et al. (1993). "Mutations affecting growth cone guidance in Drosophila: genes necessary for guidance toward or away from the midline." Neuron 10(3): 409-­‐
426. Serafini, T., S. A. Colamarino, et al. (1996). "Netrin-­‐1 is required for commissural axon guidance in the developing vertebrate nervous system." Cell 87(6): 1001-­‐1014. Serafini, T., T. E. Kennedy, et al. (1994). "The netrins define a family of axon outgrowth-­‐
promoting proteins homologous to C. elegans UNC-­‐6." Cell 78(3): 409-­‐424. Shekarabi, M. and T. E. Kennedy (2002). "The netrin-­‐1 receptor DCC promotes filopodia formation and cell spreading by activating Cdc42 and Rac1." Mol Cell Neurosci 19(1): 1-­‐17. Shekarabi, M., S. W. Moore, et al. (2005). "Deleted in colorectal cancer binding netrin-­‐1 mediates cell substrate adhesion and recruits Cdc42, Rac1, Pak1, and N-­‐WASP into an intracellular signaling complex that promotes growth cone expansion." J Neurosci 25(12): 3132-­‐3141. Shelly, M., B. K. Lim, et al. (2010). "Local and long-­‐range reciprocal regulation of cAMP and cGMP in axon/dendrite formation." Science 327(5965): 547-­‐552. Shewan, D., A. Dwivedy, et al. (2002). "Age-­‐related changes underlie switch in netrin-­‐1 responsiveness as growth cones advance along visual pathway." Nat Neurosci 5(10): 955-­‐962. Shim, S., E. L. Goh, et al. (2005). "XTRPC1-­‐dependent chemotropic guidance of neuronal growth cones." Nat Neurosci 8(6): 730-­‐735. Shim, S., J. P. Yuan, et al. (2009). "Peptidyl-­‐prolyl isomerase FKBP52 controls chemotropic guidance of neuronal growth cones via regulation of TRPC1 channel opening." Neuron 64(4): 471-­‐483. Shirasaki, R., C. Mirzayan, et al. (1996). "Guidance of circumferentially growing axons by netrin-­‐dependent and -­‐independent floor plate chemotropism in the vertebrate brain." Neuron 17(6): 1079-­‐1088. Simpson, J. H., T. Kidd, et al. (2000). "Short-­‐range and long-­‐range guidance by slit and its Robo receptors. Robo and Robo2 play distinct roles in midline guidance." Neuron 28(3): 753-­‐766. Skoulakis, E. M. and R. L. Davis (1996). "Olfactory learning deficits in mutants for leonardo, a Drosophila gene encoding a 14-­‐3-­‐3 protein." Neuron 17(5): 931-­‐944. Skoulakis, E. M. and S. Grammenoudi (2006). "Dunces and da Vincis: the genetics of learning and memory in Drosophila." Cell Mol Life Sci 63(9): 975-­‐988. Skutella, T. and R. Nitsch (2001). "New molecules for hippocampal development." Trends Neurosci 24(2): 107-­‐113. xi Bibliography Song, H., G. Ming, et al. (1998). "Conversion of neuronal growth cone responses from repulsion to attraction by cyclic nucleotides." Science 281(5382): 1515-­‐1518. Song, H. J., G. L. Ming, et al. (1997). "cAMP-­‐induced switching in turning direction of nerve growth cones." Nature 388(6639): 275-­‐279. Song, H. J. and M. M. Poo (1999). "Signal transduction underlying growth cone guidance by diffusible factors." Curr Opin Neurobiol 9(3): 355-­‐363. Soosairajah, J., S. Maiti, et al. (2005). "Interplay between components of a novel LIM kinase-­‐
slingshot phosphatase complex regulates cofilin." EMBO J 24(3): 473-­‐486. Spitzweck, B., M. Brankatschk, et al. (2010). "Distinct protein domains and expression patterns confer divergent axon guidance functions for Drosophila Robo receptors." Cell 140(3): 409-­‐420. Stanco, A., C. Szekeres, et al. (2009). "Netrin-­‐1-­‐alpha3beta1 integrin interactions regulate the migration of interneurons through the cortical marginal zone." Proc Natl Acad Sci U S A 106(18): 7595-­‐7600. Stein, E. and M. Tessier-­‐Lavigne (2001). "Hierarchical organization of guidance receptors: silencing of netrin attraction by slit through a Robo/DCC receptor complex." Science 291(5510): 1928-­‐1938. Steinacker, P., P. Schwarz, et al. (2005). "Unchanged Survival Rates of 14-­‐3-­‐3{gamma} Knockout Mice after Inoculation with Pathological Prion Protein." Mol. Cell. Biol. 25(4): 1339-­‐1346. Sunayama, J., F. Tsuruta, et al. (2005). "JNK antagonizes Akt-­‐mediated survival signals by phosphorylating 14-­‐3-­‐3." Journal of Cell Biology 170(2): 295-­‐304. Takemura, M., T. Mishima, et al. (2009). "Ca2+/calmodulin-­‐dependent protein kinase IV-­‐
mediated LIM kinase activation is critical for calcium signal-­‐induced neurite outgrowth." J Biol Chem 284(42): 28554-­‐28562. Tasken, K. and E. M. Aandahl (2004). "Localized effects of cAMP mediated by distinct routes of protein kinase A." Physiol Rev 84(1): 137-­‐167. Taya, S., T. Shinoda, et al. (2007). "DISC1 Regulates the Transport of t he NUDEL/LIS1/14-­‐3-­‐
3{varepsilon} Complex through Kinesin-­‐1." J. Neurosci. 27(1): 15-­‐26. Tear, G., R. Harris, et al. (1996). "commissureless controls growth cone guidance across the CNS midline in Drosophila and encodes a novel membrane protein." Neuron 16(3): 501-­‐514. Terman, J. R. and A. L. Kolodkin (2004). "Nervy links protein kinase a to plexin-­‐mediated semaphorin repulsion." Science 303(5661): 1204-­‐1207. Theurkauf, W. E. and R. B. Vallee (1982). "Molecular characterization of the cAMP-­‐dependent protein kinase bound to microtubule-­‐associated protein 2." J Biol Chem 257(6): 3284-­‐3290. Thomas, S., B. Ritter, et al. (2009). "Intersectin regulates dendritic spine development and somatodendritic endocytosis but not synaptic vesicle recycling in hippocampal neurons." J Biol Chem 284(18): 12410-­‐12419. Tkachenko, E., M. Sabouri-­‐Ghomi, et al. (2011). "Protein kinase A governs a RhoA-­‐RhoGDI protrusion-­‐retraction pacemaker in migrating cells." Nat Cell Biol 13(6): 660-­‐667. Togashi, K., M. J. von Schimmelmann, et al. (2008). "Cyclic GMP-­‐gated CNG channels function in Sema3A-­‐induced growth cone repulsion." Neuron 58(5): 694-­‐707. Tojima, T., J. H. Hines, et al. (2011). "Second messengers and membrane trafficking direct and organize growth cone steering." Nat Rev Neurosci 12(4): 191-­‐203. Tojima, T., R. Itofusa, et al. (2009). "The nitric oxide-­‐cGMP pathway controls the directional polarity of growth cone guidance via modulating cytosolic Ca2+ signals." J Neurosci 29(24): 7886-­‐7897. Toyo-­‐oka, K., A. Shionoya, et al. (2003). "14-­‐3-­‐3[epsi] is important for neuronal migration by binding to NUDEL: a molecular explanation for Miller-­‐Dieker syndrome." Nat Genet 34(3): 274-­‐285. Trousse, F., E. Marti, et al. (2001). "Control of retinal ganglion cell axon growth: a new role for Sonic hedgehog." Development 128(20): 3927-­‐3936. xii Bibliography Van Der Hoeven, P. C., J. C. Van Der Wal, et al. (2000). "14-­‐3-­‐3 isotypes facilitate coupling of protein kinase C-­‐zeta to Raf-­‐1: negative regulation by 14-­‐3-­‐3 phosphorylation." Biochem J 345 Pt 2: 297-­‐306. van Heusden, G. P. (2005). "14-­‐3-­‐3 proteins: regulators of numerous eukaryotic proteins." IUBMB Life 57(9): 623-­‐629. Vandamme, J., D. Castermans, et al. (2012). "Molecular mechanisms of feedback inhibition of protein kinase A on intracellular cAMP accumulation." Cell Signal 24(8): 1610-­‐1618. Vitriol, Eric A. and James Q. Zheng (2012). "Growth Cone Travel in Space and Time: the Cellular Ensemble of Cytoskeleton, Adhesion, and Membrane." Neuron 73(6): 1068-­‐
1081. Vitriol, E. A. and J. Q. Zheng (2012). "Growth cone travel in space and time: the cellular ensemble of cytoskeleton, adhesion, and membrane." Neuron 73(6): 1068-­‐1081. Wang, B., H. Yang, et al. (1999). "Isolation of high-­‐affinity peptide antagonists of 14-­‐3-­‐3 proteins by phage display." Biochemistry 38(38): 12499-­‐12504. Wang, G. X. and M. M. Poo (2005). "Requirement of TRPC channels in netrin-­‐1-­‐induced chemotropic turning of nerve growth cones." Nature 434(7035): 898-­‐904. Wang, W. a. S., D. C. (1996). "Molecular evolution of the 14-­‐3-­‐3 protein family." Journal of Molecular Evolution 43: 384-­‐398. Watari-­‐Goshima, N., K. Ogura, et al. (2007). "C. elegans VAB-­‐8 and UNC-­‐73 regulate the SAX-­‐3 receptor to direct cell and growth-­‐cone migrations." Nat Neurosci 10(2): 169-­‐176. Wen, Z., C. Guirland, et al. (2004). "A CaMKII/calcineurin switch controls the direction of Ca(2+)-­‐dependent growth cone guidance." Neuron 43(6): 835-­‐846. Wen, Z., L. Han, et al. (2007). "BMP gradients steer nerve growth cones by a balancing act of LIM kinase and Slingshot phosphatase on ADF/cofilin." J Cell Biol 178(1): 107-­‐119. Westphal, R. S., S. H. Soderling, et al. (2000). "Scar/WAVE-­‐1, a Wiskott-­‐Aldrich syndrome protein, assembles an actin-­‐associated multi-­‐kinase scaffold." EMBO J 19(17): 4589-­‐
4600. Williams, M. E., S. C. Wu, et al. (2003). "Surface expression of the netrin receptor UNC5H1 is regulated through a protein kinase C-­‐interacting protein/protein kinase-­‐dependent mechanism." J Neurosci 23(36): 11279-­‐11288. Williams, S. E., F. Mann, et al. (2003). "Ephrin-­‐B2 and EphB1 mediate retinal axon divergence at the optic chiasm." Neuron 39(6): 919-­‐935. Wilson, S. I., B. Shafer, et al. (2008). "A molecular program for contralateral trajectory: Rig-­‐1 control by LIM homeodomain transcription factors." Neuron 59(3): 413-­‐424. Winckler, B. and I. Mellman (2010). "Trafficking guidance receptors." Cold Spring Harb Perspect Biol 2(7): a001826. Wolf, A. M., A. I. Lyuksyutova, et al. (2008). "Phosphatidylinositol-­‐3-­‐kinase-­‐atypical protein kinase C signaling is required for Wnt attraction and anterior-­‐posterior axon guidance." J Neurosci 28(13): 3456-­‐3467. Wu, K. Y., J. H. Zippin, et al. (2006). "Soluble adenylyl cyclase is required for netrin-­‐1 signaling in nerve growth cones." Nat Neurosci 9(10): 1257-­‐1264. Xiao, B., S. J. Smerdon, et al. (1995). "Structure of a 14-­‐3-­‐3 protein and implications for coordination of multiple signalling pathways." Nature 376(6536): 188-­‐191. Xu, H., S. G. Leinwand, et al. (2010). "The calmodulin-­‐stimulated adenylate cyclase ADCY8 sets the sensitivity of zebrafish retinal axons to midline repellents and is required for normal midline crossing." J Neurosci 30(21): 7423-­‐7433. Yaffe, M. B. (2002). "How do 14-­‐3-­‐3 proteins work? -­‐ Gatekeeper phosphorylation and the molecular anvil hypothesis." FEBS letters 513(1): 53-­‐57. Yam, P. T., S. D. Langlois, et al. (2009). "Sonic hedgehog guides axons through a noncanonical, Src-­‐family-­‐kinase-­‐dependent signaling pathway." Neuron 62(3): 349-­‐362. Yamauchi, K., K. D. Phan, et al. (2008). "BMP type I receptor complexes have distinct activities mediating cell fate and axon guidance decisions." Development 135(6): 1119-­‐1128. xiii Bibliography Yang, L., D. S. Garbe, et al. (2009). "A frazzled/DCC-­‐dependent transcriptional switch regulates midline axon guidance." Science 324(5929): 944-­‐947. Yang, T. and J. R. Terman (2012). "14-­‐3-­‐3epsilon Couples Protein Kinase A to Semaphorin Signaling and Silences Plexin RasGAP-­‐Mediated Axonal Repulsion." Neuron 74(1): 108-­‐121. Yang, X., W. H. Lee, et al. (2006). "Structural basis for protein-­‐protein interactions in the 14-­‐
3-­‐3 protein family." Proceedings of the National Academy of Sciences of the United States of America 103(46): 17237-­‐17242. Yingling, J., Toyo-­‐oka, K., Wynshaw-­‐Boris, A. (2003). "Miller-­‐Dieker Syndrome: Analysis of a Human Contiguous Gene Syndrome in the Mouse." American Journal of Human Genetics 73: 475-­‐488. Yoon, B. C., K. H. Zivraj, et al. (2011). "14-­‐3-­‐3 proteins regulate retinal axon growth by modulating ADF/cofilin activity." Dev Neurobiol. Zaccolo, M. and T. Pozzan (2002). "Discrete microdomains with high concentration of cAMP in stimulated rat neonatal cardiac myocytes." Science 295(5560): 1711-­‐1715. Zenke, F. T., M. Krendel, et al. (2004). "p21-­‐activated kinase 1 phosphorylates and regulates 14-­‐3-­‐3 binding to GEF-­‐H1, a microtubule-­‐localized Rho exchange factor." J Biol Chem 279(18): 18392-­‐18400. Zhang, J., C. J. Hupfeld, et al. (2005). "Insulin disrupts beta-­‐adrenergic signalling to protein kinase A in adipocytes." Nature 437(7058): 569-­‐573. Zhang, L., H. Wang, et al. (1997). "Raf-­‐1 Kinase and Exoenzyme S Interact with 14-­‐3-­‐3zeta through a Common Site Involving Lysine 49." J. Biol. Chem. 272(21): 13717-­‐13724. Zhang, X. M., M. Ramalho-­‐Santos, et al. (2001). "Smoothened mutants reveal redundant roles for Shh and Ihh signaling including regulation of L/R symmetry by the mouse node." Cell 106(2): 781-­‐792. Zhao, J., C. L. Meyerkord, et al. (2011). "14-­‐3-­‐3 proteins as potential therapeutic targets." Semin Cell Dev Biol 22(7): 705-­‐712. Zheng, J. Q. (2000). "Turning of nerve growth cones induced by localized increases in intracellular calcium ions." Nature 403(6765): 89-­‐93. Zhou, Q., Y. S. Kee, et al. (2010). "14-­‐3-­‐3 coordinates microtubules, Rac, and myosin II to control cell mechanics and cytokinesis." Curr Biol 20(21): 1881-­‐1889. Zivraj, K. H., Y. C. Tung, et al. (2010). "Subcellular profiling reveals distinct and developmentally regulated repertoire of growth cone mRNAs." J Neurosci 30(46): 15464-­‐15478. Zou, Y., E. Stoeckli, et al. (2000). "Squeezing axons out of the gray matter: a role for slit and semaphorin proteins from midline and ventral spinal cord." Cell 102(3): 363-­‐375. xiv