Repulsive Axonal Pathfinding Requires the Ena/VASP family of Actin Regulatory
Proteins in Vertebrates
by
John Edward van Veen
B.S. Genetics, Mathematics
University of California, Davis, 2001
SUBMITTED TO THE DEPARTMENT OF BIOLOGY IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY IN BIOLOGY
AT THE
MASSACHUSETTS INSTITUTE IF TECHNOLOGY
MAY I
FEBRUARY 2012
@ 2012 Massachusetts Institute of Technology. All rights reserved.
ARCHIVES
-7
Signature of
Author:
(
Department of Biology
January 12, 2012
Certified by:
Frank B. Gertler
Professor of Biology
Thesis Supervisor
Accepted by:
Alan Grossman
Professor of Biology
Chairman, Committee of Graduate Students
2
Repulsive Axonal Pathfinding Requires the Ena/VASP Family of Actin Regulatory
Proteins in Vertebrates
by
John Edward van Veen
Submitted to the Department of Biology on January 12, 2012 in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy in Biology
Thesis Abstract
Vertebrate nervous system development requires the careful interpretation of
many attractive and repulsive guidance molecules. For the incredibly complicated wiring
diagram comprising the vertebrate nervous system to elaborate properly, the highly
motile "growth cone" at the tip of an axon must sense extracellular embryonic cues and
respond through a number of intracellular interactions leading ultimately to coordinated
changes in cytoskeletal morphology and modulation of the axonal path.
Here I describe axon pathfinding defects displayed by mice genetically deficient
for all three vertebrate Ena/VASP homologues: Mena, VASP, and EVL. As has been
reported previously in invertebrates, these defects share phenotypic overlap with those
seen in mice genetically deficient for the repulsive guidance molecules Slit and Robo. I
find that the pathfinding errors observed in Ena/VASP deficient mice are likely a result of
failure to respond to Slit/Robo. Furthermore, based on my findings, I propose a "four-step"
model of growth cone responses to repulsive cues.
Finally I find that the direct binding of Ena/VASP proteins to Robo seen in
invertebrates is conserved and expanded in vertebrates. These interactions appear to be
tunable by phosphorylation, suggesting a model by which context dictates the
Ena/VASP:Robo interaction, potentially leading to changes in growth cone
responsiveness to guidance cues.
Thesis Supervisor: Frank Gertler
Title: Professor of Biology
3
4
ACKNOWLEDGEMENTS
First and foremost I would like it to be known that mentoring me has not
generally been an easy task. While my love for problem solving is what led me to
become a scientist, it has at times, led me to become overly enamored with my own
hypotheses. While many advisors certainly would insist on the type of scientific rigor that
heartbreakingly reveals the truth behind a flawed idea, Frank Gertler somehow manages
to do so in a way that isn't discouraging. Thanks to Frank's guidance and friendship, my
time at MIT has been immensely rewarding and for this I will never be able to thank him
enough.
I need to thank the members of my committee for their continual guidance over
the years. Troy Littleton, Carlos Lois, Adam Martin, and Mike Hemann have all made
indelible marks on the direction of this work. Furthermore, while many people only meet
their external committee member on the day of their defense, I was fortunate enough to
have asked Elke Stein for her help. Elke has a genuine love for science like few others,
and has become an integral part of this project. I look forward to continued collaboration
with her and cannot wait to see what comes of the growing relationship between the
Gertler and Stein labs.
The list of lifelong friends that I have made at MIT cannot be completely
enumerated here due to space restraints, however without them I can confidently state
that I would never have been able to survive this experience. Gordon, Luke, Nate, Shomit,
and I fancied ourselves MIT's own rat pack and were partners in both scientific
edification and general raucous carousing. During graduate school, we attended each
other's weddings and more recently have begun to meet each other's children. In spite of
now being scattered across this earth, I look forward to many more years of amazing
friendship.
Leslie Mebane began as just a classmate with whom to commiserate about our
shared inexperience with the cold weather and lack of decent Mexican food in the city of
Boston, but she later became a labmate, and eventually thanks to her caring became one
of the best friends that I have ever had. In the preparation of this thesis, there were times
when I lost confidence in my ability to finish it, but Leslie was always there to help me
get my head back on straight.
As any scientist can attest, laboratory life can easily become all consuming. MIT
is unique in that it has an amazing place on campus for graduate students to either forget
about science entirely for a few hours or to connect with scientists from different
disciplines. My friends at the Muddy Charles Pub have always been there to provide
much needed stress relief and a way to really get out of the lab, if only for a short period
of time.
The help I have received extends well beyond the confines of Cambridge. Even
from across the country, my friends and family in California have sent continual support
not only mentally but physically, sending care packages and surprise pizza deliveries
when life got the most hectic.
5
Most of all, I need to thank my wife Stephanie, who has been remarkably patient
and understanding about my need to stay behind in Cambridge when she moved to San
Francisco so that I could feel comfortable about the completeness of the study presented
here. It is also really a wonderful thing being married to a scientist more talented than
you are, as you have a constant role model and resource for bouncing ideas off of.
Without her love and support I have no idea where I would be now.
6
Table of Contents
T itle page...............................................................................................1
A bstract.............................................................................................
3
Acknowledgements................................................................................5
Chapter 1 - Introduction: Precise Regulation of the Cytoskeleton Shapes the
Nervous System.....................................................................................
9
Chapter 2 - Mice genetically deficient in Mena, VASP, and EVL display axon
pathfinding defects suggestive of impaired sensitivity to repulsive guidance cues....59
Chapter 3 - Normal growth cone response to the chemorepellent, Slit requires
Ena/VASP induced reorganization of the cytoskeleton.................................84
Chapter 4 - EVH1 domains of vertebrate Ena/VASP proteins interact directly with
multiple sites in the cytoplasmic tail of Robo in a phospho dependent manner.....122
Chapter 5 - Conclusions and future directions.............................................142
7
8
Chapter 1
Introduction: Precise Regulation of the Cytoskeleton Shapes the Nervous
System.
9
1.1 Connections in the nervous system ..................................................................................
11
1.2 Axonal Pathfinding ...............................................................................................................
12
1.3 Guidance cues and their receptors....................................................................................
13
1.3.1 Slit/Robo ..........................................................................................................................
14
1.3.2 Netrin/DCC/Unc5/DsCAM ...........................................................................................
21
1.2.3 Sem aphorin/Plexin/Nrp................................................................................................
22
1.2.4 Ephrins/EphRs .................................................................................................................
22
1.4 Downstream effectors and the cytoskeleton .......................................................................
23
1.4.1 Rho GTPases....................................................................................................................
23
1.5 Regulation of the growth cone cytoskeleton ....................................................................
24
1.5.1 Ena/VASP proteins are critical for filopodia formation ...............................................
30
1.5.2 Ena/VASP proteins are regulated by localization.........................................................
31
1.5.3 Ena/VASP proteins are regulated by phosphorylation. ................................................
34
1.5.4 Ena/VASP proteins regulate neural development in vertebrates and invertebrates ........ 34
1.6 References..............................................................................................................................
10
43
1.1 Connections in the nervous system.
At the cellular level, the basic building blocks of the mammalian nervous system share
extensive similarities with those found in nearly every multicellular animal. Even the most
rudimentary of extant neural nets, as those of the Cnidarian jellyfish, contain the same three
fundamental neuronal types found in vertebrates: sensory neurons, which receive environmental
information and stimulate interneurons, which then stimulate specific populations of motor
neurons, which finally translate this information into changes in behavior mediated by
orchestrated muscular contraction. Comparative examination at the biochemical or ultrastructural
level reveals few differences in neuronal structure even between very distantly related
organisms. This then begs the question of how more complex animals have evolved such a
variety of behaviors without more significant changes in the most fundamental units of our
nervous systems.
The modem "neuron theory" stems ultimately from the work of Santiago Ram6n
Cajal
who used Golgi staining to
visualize the structure of the
cells that comprise the nervous
system. Ram6n y Cajal was
i~
able to identify, and draw with
i
high accuracy the
morphologies of individual
Figure 1 - Drawing of neurons in the chick cerebellum
1905.
Santiago Ramon y Cajal, neurons (Fig. 1). The quality of
11
Ram6n y Cajal's preparations allowed him to propose the idea that the nervous system is
contiguous rather than continuous; that is to say that each neuron is an autonomous unit of a
network in which connections are formed between the axon of one neuron and the dendrites, or
cell body of another. It has been long appreciated that the number and elaboration of these
connections correlates more closely with an organism's complexity of behavior than any other
known indicator. Understanding how these connections are made during development remains a
problem with many unsolved questions, a small fraction of which are addressed here.
1.2 Axonal Pathfinding
Connections in the nervous system are made with remarkable specificity. Invertebrates,
though orders of magnitude less complicated than vertebrates, are simultaneously fantastically
complicated, yet reliable in their development. The nematode worm Caenorhabditiselegans, for
example, has 282 somatic neurons which make approximately 6400 inter-neuronal synaptic
connections and 1400 neuromuscular junctions(Varshney et al., 2011). These connections are
made stereotypically; under normal developmental conditions, two equivalent neurons from two
different C. elegans worms of the same sex will make precisely the same inter-neuronal synapses
and neuromuscular junctions. The human nervous system is significantly more complex;
estimates for the number of neurons center around 1011, which are thought to make up to 1014
synaptic connections(Sporns et al., 2005). Given its complexity, mapping all of the connections
in the human nervous system at the single neuron level, as has been done in C. elegans, remains
unfeasible. While there is almost certainly more variability in exactly which neurons connect in
humans, it is clear that under normal conditions, the general pathways between different regions
of the nervous system are remarkably conserved.
12
For the correct synapse or neuromuscular junction to
eventually be made, during development the axon must be able
to recognize and respond to extracellular landmarks present in
an embryo. To achieve this, at the tip of every growing axon is
a highly specialized structure called a growth cone (Fig. 2).
Growth cones are thin, fan-shaped structures made up of
Figure 2 - A neuronal growth cone
(DIC)
lamellipodia, protrusive actin meshworks, and numerous
filopodia, actin rich finger-like projections(Yamada, 1971).
Growth cones are highly dynamic, continually scanning the local microenvironment for guidance
cues, responding with coordinated changes in the cytoskeleton that lead to modifications of
either speed or trajectory, which ultimately determine the final path of the axon(Bentley and
Toroian-Raymond, 1986)(Caudy and Bentley, 1986).
1.3 Guidance cues and their receptors
The axonal guidance cues interpreted by growth cones have a wide variety of molecular
origin and action. The "canonical" guidance cues were originally described as proteins that could
act as either chemoattractants or chemorepellents at long range when secreted and at short range
when membrane tethered(Kennedy et al., 1994)(Kidd et al., 1999)(Cheng et al., 1995)(Kolodkin
et al., 1993). Cell adhesion molecules such as cadherin and integrin have also been demonstrated
to participate independently in short range axon guidance (Kuhn et al., 1995)(Matsunaga et al.,
1988) and in cooperation with the canonical cues in long range axon guidance(Stevens, 2002).
Finally, more recent research has demonstrated that molecules classically described as
morphogens or growth factors have strong effects on axon guidance at a subset of decision
points(Charron et al., 2003)(Lyuksyutova et al., 2003). Adding to the complexity and tunability
13
of this system, cross talk between different members of these families can lead to context
specific silencing or conversion of guidance response polarity (Stein and Tessier-Lavigne,
2001)(Hong et al., 1999)(Holt et al., 1999)(Song, 1998). One of the best studied systems of
pathfinding is the guidance of axons originating from neurons in the dorsal portion of the spinal
cord. These neurons form commissures by extending their axons across the midline, synapsing
with neurons on the contralateral side. This context provides a unique challenge for a growth
cone in which guidance factor sensitivity must be context dependent. During this developmental
window, the floor plate of the spinal cord continually produces both attractive and repulsive
guidance cues. For a growth cone to cross the midline, it must first be attracted to this milieu so
that it may reach it. Then, it must switch its attraction to repulsion effecting midline exit and
preventing re-crossing. While much progress has been made in understanding how midline
guidance is coordinated (Chen et al., 2008)(Stein and Tessier-Lavigne, 2001)(Simpson et al.,
2000)(Holt et al., 1999)(Song et al., 1997)(Stein et al., 1998), context dependent sensitivity to
guidance cues remains one of the areas of axon guidance with many unsolved questions.
Following is a brief description of the four canonical guidance cue families, their receptors, and
their roles in axon guidance.
1.3.1 Slit/Robo
The founding member of the Robo family of transmembrane receptors, roundabout, was
discovered in a D. melanogasterscreen for genes involved in midline repulsion of axons(Seeger
et al., 1993). The D. melanogaster central nervous system normally contains non-commissural
axons that form longitudinal tracts and commissural axons that cross the midline before being
guided longitudinally. In Robo mutants, non-commissural axons aberrantly reach the midline.
Furthermore, rather than switching from midline attraction to repulsion, commissural axons in
14
Robo mutant animals repeatedly re-cross the midline. Together, these defects lead to a reduction
in longitudinal axon tract formation and the appearance of circular rings of axon fibers centered
around the midline. These functions of robo are conserved in vertebrates, which have four Robo
paralogs, as animals that express loss of function alleles of Robo 1, Robo2, and Robo3 show
stalling and re-crossing of commissural axon fibers in the spinal cord midline (Long and
Sabatier, 2004).
Mammals carrying loss of function mutations in Robo 1 and Robo2 also display defects in
axon guidance in the peripheral nervous system. Afferent axons from the dorsal root ganglia
(DRG) normally extend to the dorsal funiculus where they bifurcate, and send their daughter
branches orthogonally along the spinal cord. In the absence of Robol and Robo2, these axons
project aberrantly into the spinal cord, often reaching or extending beyond the midline. Further,
axons emanating from the trigeminal ganglia in Robol/Robo2 double knockout animals display
defects in axonal outgrowth and branching (Ma and Tessier-Lavigne, 2007).
Robo receptors belong to the immunoglobulin superfamily of cell adhesion molecules.
The canonical structure of Robo is one of an amino-terminal extracellular region with five Ig
domains and three fibronectin type III domains (Fig. 3a). Robos contain a single transmembrane
domain followed by a cytoplasmic tail with four short motifs (CCO-CC3) conserved in
vertebrates and invertebrates (Fig. 3c). No catalytic domains have been described in the
cytoplasmic tail of Robo. Instead, ligand dependent recruitment of specific binding partners to
the conserved cytoplasmic motifs leads to activation of downstream effectors. While the entire
signaling pathway downstream of Robo activation remains to be elucidated, it is clear that proper
response requires recruitment of signaling molecules that act upon the cytoskeleton (Fan et al.,
2003).
15
The canonical Robo organization is shared by the two D. melanogasterRobos and Robo 1
and Robo2 in vertebrates. Other vertebrate homologs share varying levels of conservation.
Vertebrate Robo3 has the same extracellular domains, but lacks the CC3 motif. Interestingly, the
cytoplasmic tail of Robo3 can be alternatively spliced, and heterodimerization of one of these
spliceforms, Robo3.1 with other Robo members silences their repulsive response, an activity
crucial in commissural axons for initial attraction to the midline (Chen et al., 2008). Vertebrate
Robo4 is the most divergent in the family, having fewer extracellular Ig and FnIII domains, and
lacking both CC1 and CC3. Robo4 is expressed largely in endothelial cells and so far it's
function has been primarily described in angiogenesis and vascular stability (Bedell et al.,
2005)(Jones et al., 2009).
Slit was initially identified in a drosophila hybridization screen for genes containing
EGF-like repeats (Rothberg et al., 1988), and later determined to be the primary ligand for Robo
(Kidd et al., 1999). Flies with loss of function mutations in Slit display potent defects in axon
guidance and development of midline glia (Rothberg et al., 1990). Slit mutants display a CNS
scaffold that is collapsed into a single longitudinal nerve bundle, a defect similar but more severe
than those seen in flies lacking Robo (Rothberg et al., 1990). As is the case with Robo, the
phenotypes of mutations affecting Slit in vertebrates seem to be largely conserved, and similar
though sometimes slightly more pronounced than the Robo knockout phenotypes (Ma and
Tessier-Lavigne, 2007)(Long and Sabatier, 2004).
Structurally, all known Slit proteins in both invertebrates and vertebrates share extensive
similarities. They are all large secreted glycoproteins of roughly 200kDa with four aminoterminal domains containing leucine rich repeats (LRRs). Immediately carboxy-terminal to these
LRRs in vertebrates are nine EGF-like repeats, or seven in the case of invertebrates. Finally, Slits
16
all have a cysteine knot at the carboxy-terminus (Fig. 3b). All Slits bind the canonical Robos
with similar affinity through an interaction between the second LRR and the first two Ig domains
of Robo (Brose et al., 1999). While the second LRR of Slit is necessary and sufficient for
binding and repulsive activity, the full length protein binds Robo with higher affinity, as a result
of increased dimerization mediated by a homophilic interaction in the fourth LRR that is
stabilized by heparan sulfate proteoglycans (Seiradake et al., 2009). While this homophilic
interaction suggests that ligand dependent receptor dimerization may be important for Robo
activation, this has yet to be demonstrated directly. Slit proteins have a conserved site for
proteolytic cleavage near the carboxy-terminus, which when acted upon by a yet unidentified
protease, produces a -180 kDa amino-terminal fragment which can independently act as a
repellent, and a 20kDa carboxy-terminal fragment of unknown function (Nguyen Ba-Charvet et
al., 2001).
Plasma membrane localization of Robo underlies its function as a transmembrane
receptor. In D. melanogaster,the transmembrane protein Commissureless (Comm) regulates
midline guidance by temporarily diverting Robo from a plasma membrane directed pathway to
an endosomal one when growth cones approach the midline (Keleman et al., 2002). A Comm
homologue has not been discovered in vertebrates, though in cancer cells, Slit stimulation causes
a relocation of Robo to the plasma membrane, a process which requires the deubiquitinating
enzyme, USP33 (Yuasa-Kawada et al., 2009a). Interestingly, in commissural neurons, the
opposite effect is observed, as Robo is lost from the plasma membrane upon Slit stimulation.
siRNA mediated depletion of USP33 in neurons potentiates this relocation in vitro and depletion
of USP33 in neurons in vivo causes a guidance defect in which axons cannot exit the midline
(Yuasa-Kawada et al., 2009b). These seemingly conflicting observations imply that while
17
Figure 3
Robo
Slit
B
Ig domains
Leucine rich
repeats
IFnIllI domains
L
TM domain
EGF repeats
r
Conserved
cytoplasmic
motifs
Cysteine knot
Cleavage site
CCI
Abl, Ena
CC2
CrGAPNilse, Dock/NCK, Ena
CC3
SrGAP
18
Figure 3 - Domain structure of the chemorepellent Slit and it's receptor, Robo
(A) Canonical structure of Robo showing five extracellular Ig domains followed by three
fibronectin type III domains. The cytoplasmic tail of Robo contains three conserved cytoplasmic
motifs. (B) Canonical structure of the ligand, Slit, encompassing four leucine rich repeats
followed by seven to nine EGF repeats and a cysteine knot. (C) Known interactions of the
cytoplasmic tail of Robo. Abl: Non-receptor tyrosine kinase. Ena: Actin regulatory protein.
CrGAP/Vilse: Dual function Rac/Cdc42 GAP. Dock/NCK: Rac GEF. SrGAP Dual function
RhoA/Cdc42 GAP.
19
ubiquitination status plays an important role in Robo function, other factors likely cooperate in
context dependent changes in localization of Robo upon Slit application.
While Slit and Robo were both originally described as effectors of commissural midline
pathfinding, Slit/Robo signaling has subsequently been shown to be required for proper guidance
in a wide variety of axonal projections. Mutations in Slit or Robo result in misprojection of
various types of sensory axons, including those in the optic chiasm (Plump et al., 2002),
projections from the olfactory bulb (Jhaveri et al., 2004), projections from the vomeronasal organ
(Cloutier et al., 2004), and projections from the trigeminal and dorsal root ganglia (Ma and
Tessier-Lavigne, 2007).
Beyond their roles in repulsive guidance, Slit proteins have been shown to promote
axonal elongation and arborization both in vitro (Wang et al., 1999) and in a subset of sensory
neurons in vivo (Ma and Tessier-Lavigne, 2007). The branch promoting activity of Slit has been
reported as being promoted uniquely by the amino-terminal fragment, and antagonized by the
full length protein (Nguyen Ba-Charvet et al., 2001), though the mechanism for this antagonism
remains unknown.
Recently it has become appreciated that Slit/Robo signaling plays many roles outside of
the nervous system. It has been reported that Slit/Robo is frequently lost in certain breast cancers
where it plays a role in silencing SDF-1/CXCL12 mediated invasiveness (Prasad et al., 2004). It
has also been reported that Slit may act as a chemoattractant in brain specific breast cancer
metastasis (Schmid et al., 2007). While these observations are seemingly in conflict, they may
represent another example of context specific changes in polarity of chemotactic responses to
guidance factors.
20
1.3.2 Netrin/DCC/Unc5/DsCAM
Of the other canonical guidance cues, the pathway with the strongest demonstrated link to
Slit/Robo signaling is the Netrin/DCC pathway. Like Slit, Netrin is a chemotropic factor secreted
by the midline shown to act in long range commissural axon guidance. Structurally, Netrin also
contains EGF repeats, but is much smaller, ranging in size from 70-80 kDa. As opposed to Slits,
Netrins were originally described as midline attractants (Ishii et al., 1992).
Characterized Netrin receptors fall in to four major families: DCC/Neogenin, Unc5,
DsCAM, and the NGL family. The fourth class of Netrin receptor, the GPI linked NGL receptors
bind to GPI linked Netrin G ligands but not to canonical Netrins. While there is one report of
Netrin-G/NGL proteins being important for axon outgrowth (Lin et al., 2003), these proteins'
main characterized function is in synaptogenesis, so will not be discussed here further.
The first characterized Netrin receptor, DCC, is the best-studied mediator of Netrin induced
chemoattraction. The extracellular portion of DCC has four Ig domains followed by six
fibronectin type III domains. Similar to Robo, DCC has no known cytoplasmic domains with
catalytic activity, but does have three evolutionarily conserved motifs (P1 -P3), which bind to
effector proteins upon ligand stimulation. Netrin/DCC mediated attraction can be switched to
repulsion by either substrate dependent modulation of growth cone cyclic nucleotide levels
(Nishiyama et al., 2003), or heterophilic interactions with members of the Unc5 family. The
Unc5 family of receptors can mediate a chemorepulsive activity of Netrin through heterophilic
interaction with DCC (Hong et al., 1999), or in certain contexts act as DCC independent
repulsive guidance receptors via homophilic interactions (Keleman, 2001) . The final described
class of Netrin receptor is the DsCAM family of cell adhesion molecules, which can mediate
attractive growth cone turning in response to Netrin both independently of, and in collaboration
21
with DCC (Ly et al., 2008). Netrin and Slit signaling converge through a context specific, ligand
dependent interaction between CCl of Robo and P3 or DCC. This interaction leads to a silencing
of growth cone attraction to Netrin and has been demonstrated to be the primary role of Robo in
certain developmental contexts (Stein and Tessier-Lavigne, 2001).
1.2.3 Semaphorin/Plexin/Nrp
Similar to Slit/Robo, Semaphorins were originally described as potent chemorepulsive
molecules involved in many steps of the normal development of the nervous system (Fan and
Raper, 1995)(Luo et al., 1993)(Adams et al., 1996)(Cheng et al., 2001). Semaphorins can be
secreted, transmembrane, or membrane-tethered, and as such have roles in both long and short
range guidance. Vertebrates have up to twenty members of the Semaphorin family of ligand,
which signal in different combinations through their nine Plexin receptors. Many Semaphorins
bind directly to Plexins, but others must bind to one of two co-receptors, Neuropilin 1 and 2. As
is the case with the other canonical guidance cues, Semaphorins also have demonstrated roles in
attractive guidance that depend on context. Very recently, cross-talk between the Semaphorin
and Slit pathways was demonstrated in that Robo potentiates the response of cortical
interneurons to Semaphorins, and that Robo binds directly to Neuropilin-1 (Hemindez-Miranda
et al., 2011).
1.2.4 Ephrins/EphRs
The final members of the canonical guidance cues belong to the Ephrin family. Differing
from the other guidance cues, Ephrins are exclusively plasma membrane associated, either being
integral or tethered via a GPI linkage, and as such have been described exclusively as short range
guidance cues. Further differentiating Ephrins from the other canonical cues, their receptors
(EphRs) have a kinase domain that becomes catalytically active upon receptor activation.
22
Interestingly, in contrast to other described receptor tyrosine kinases, dimerization is insufficient
to effect full cellular response, instead, Ephrins and their receptors must be clustered in tetramers
(Stein et al., 1998). Like the other canonical cues, Ephrins can be attractive or repulsive
depending on context (Marquardt et al., 2005).
1.4 Downstream effectors and the cytoskeleton
For a guidance cue to be effective, it's recognition must result in a modulation growth
cone migratory behavior. Any change in growth cone speed or trajectory ultimately requires
directed changes in growth cone morphology, orchestrated by spatial regulation of the actin
cytoskeleton. While guidance receptors have not been shown to act directly upon the
cytoskeleton, there is a clear link formed by interaction with downstream effectors.
1.4.1 Rho GTPases
The Rho family of small GTPases is a subfamily of the Ras superfamily long known to
be involved in regulation of cytoskeletal dynamics. Rho GTPases are active when bound to the
cyclic nucleotide GTP but possess catalytic activity that renders them inactive through hydrolysis
of GTP to GDP. Switching between the active and inactive form is accomplished primarily by
four known means: The intrinsic hydrolytic activity of the GTPase; The enhanced hydrolytic
activity of the GTPase when it is bound to a GTPase activating protein (GAP); The reactivation
of the GTPase by GDP-GTP exchange mediated by a guanine nucleotide exchange factor (GEF);
Or the prevention of this exchange by a guanine nucleotide dissociation inhibitor (GDI).
The three best studied members of the Rho GTPases, RhoA, Rac, and CDC42 have been
shown to have differing roles in cytoskeletal regulation and consequently different loss of
function phenotypes in axon guidance. Given the number of guidance receptors in vertebrates,
23
the value of describing all of their downstream effectors is quickly outweighed by space
limitations; Therefore, I here describe only what is known about downstream effectors of Robo
activation.
The GAP Vilse was originally discovered in a D. melanogasterscreen for genes affecting
the formation of tracheal airways, but was shown in the same study to potentiate the midline
axon guidance defects observed in flies heterozygous for mutations of either Slit or Robo. Vilse
binds directly to the CC2 motif of Robo, and is a GAP for Rac, and more weakly for CDC42
(Lundstr6m et al., 2004). Interestingly, overexpression of Vilse shares the same phenotype of
potentiation of Slit/Robo defects, indicating that Rac activation likely must be tightly controlled
during axon guidance (Hu et al., 2005). Further evidence for the involvement of Rac in effecting
changes downstream of Robo come from the phenotype of the Rae GEF, Son of Sevenless (Sos).
In D. melanogasterSos mutants, aberrant axonal guidance phenotypes similar to those in Robo
mutants are observed in lateral axon tracts. Upon Slit stimulation, Sos is recruited into a ternary
complex with Robo mediated by the adaptor protein, Dock/Nek (Yang and Bashaw, 2006).
Furthermore, loss of function mutations in Dock result in axon guidance defects (Garrity et al.,
1996) and Dock displays a genetic interaction with Robo(Fan et al., 2003). While one study
describes a link to RhoA activation upon Slit stimulation via formation a Robo - SrGAP1
complex in tissue culture cells (Wong et al., 2001), it is unclear whether this complex has any
role in axon guidance.
1.5 Regulation of the growth cone cytoskeleton
Integration of guidance receptor signaling is a phenomenally complicated process, likely
involving many yet to be discovered interactions. Whatever proteins make up this network,
24
however, it is clear that for their activities to be productive, their outputs must eventually
converge on spatially regulated changes in the growth cone cytoskeleton.
Growth cones can be functionally divided into two regions: the central region, which is
most proximal to the axon shaft more distal peripheral region (Letoumeau, 1983). The peripheral
region of the growth cone is comprised of lamellipodia, sheet-like F-actin meshworks, and
filopodia, finger-like projections containing many parallel, unbranched actin filaments. The
peripheral region of the growth cone is highly dynamic; F-actin is under a continual process of
tightly regulated turnover. Actin monomers (globular or G-actin) are added to the membraneoriented, fast growing "barbed ends" (Pollard and Mooseker, 1981) of actin microfilaments
thereby producing protrusive force on the plasma membrane, an activity crucial in inducing local
changes in growth cone morphology.
Free barbed ends represent a rate-limiting component of protrusion as they are efficiently
bound by ubiquitous proteins that inhibit polymerization. Actin monomer addition is regulated
negatively by the activity of capping proteins (CPs), which bind to filaments and block further
monomer addition (Isenberg et al., 1980). CP associates quickly with free barbed ends but
dissociates slowly: In cells, free barbed ends have a half-life of -. 2 sec whereas capped filaments
have a half-life of ~28 min (Schafer et al., 1996). Elaboration of a protrusive structure requires
coordination of actin polymerization at many barbed ends simultaneously and as such the first
step involves uncapping filaments or creating new barbed ends.
Polymerization competent barbed ends are created near the plasma membrane primarily
by two mechanisms: Nucleation of new branches from existing filaments and capped end
removal by filament severing proteins. Spatially restricted activation of the actin severing
protein, Cofilin upon release from the phosphoinositide PIP 2 severs the ends of capped filaments,
25
creating new barbed ends near the plasma membrane in response to growth factor stimulation
(Fig. 4a)(van Rheenen et al., 2007). The major mediator of barbed end formation via branch
nucleation is a highly conserved seven member complex (Arp2/3) including two actin related
proteins, Arp2 and Arp3. The Arp2/3 complex binds to the sides of nascent actin filaments
promoting nucleation of new filaments at a characteristic 700 angle (Mullins et al., 1997)(Mullins
et al., 1998). Interestingly, capping protein directly cooperates with Arp2/3 to nucleate new actin
filaments. This complex has the dual activity of barbed end capping and barbed end formation,
leading to the formation of a dense, highly branched actin network (Akin, 2008).
The Ena/VASP family of actin regulatory proteins act as filament elongation factors by
binding the barbed ends of growing actin filaments enhancing polymerization and blocking
capping activity (Fig. 4a) (Barzik et al., 2005). Expression of Ena/VASP proteins promotes an
actin geometry comprised of longer, less branched microfilaments. The requirement of barbed
end actin polymerization for protrusive activity leads one to assume that Ena/VASP proteins
would be positive regulators of cell speed. This idea is supported by experiments demonstrating
that Ena/VASP proteins potentiate the rate of motility of the intracellular pathogen Listeria
monocytogenes (Laurent, 1999). In an observation that seems antithetical to this model, loss of
Ena/VASP function increases fibroblast speed of translocation while overexpression has the
opposite effect. Closer examination reveals that Ena/VASP function does indeed correlate with
increased protrusive behavior, but that these protrusions are highly labile and do not form
adhesive contacts with the substratum required for subsequent cellular translocation (Bear et al.,
2000)(Bear et al., 2002). These data support a model in which dense highly branched actin
networks possess the rigidity required to produce stable lamellipodial protrusions, and a network
26
of aberrantly long, unbranched, unbundled actin fibers may not be able to resist the
countervailing force exerted by the membrane, leading to premature lamellipodial collapse.
Vertebrates have three Ena/VASP family members with many overlapping functions,
Mena, Vasodilator-stimulated phosphoprotein (VASP), and Ena-VASP Like (EVL). All
Ena/VASP family members share a conserved tripartite structure in which two Ena/VASP
homology domains, EVH 1 and EVH2, flank a central proline rich region. During filament
elongation, Profilin-Actin complexes bind to "recruiting" sites in the amino-terminal portion of
the proline rich region of Ena/VASP. Recruiting sites have the sequence GPPPPP: a sequence
able to bind multiple Profilin-Actin complexes simultaneously. Profilin-Actin is then passed to a
highly conserved "loading" site more proximal to the EVH2 domain, which binds to profilin
with higher affinity than the recruiting sites. Profilin-actin is then passed to the G-actin binding
site within EVH2, where structural insights imply that it will be in the appropriate orientation for
direct addition to the barbed end of a growing actin filament (Ferron et al., 2007). In vitro
experiments demonstrate that profilin, actin, and Ena/VASP can form a trimeric complex at the
barbed end of a microfilament, which potently inhibits binding of capping protein across a range
of salt concentrations. As Profilin-Actin complexes represent the major pool of polymerization
competent actin monomers found in cells (Kaiser et al., 1999), it is believed that this complex
represents the physiological state of a barbed end bound to Ena/VASP, and that as such
Ena/VASP proteins are highly processive anti-cappers (Fig. 4b)(Hansen and Mullins, 2010).
27
Figure 4
ly
Capping
protein
ADP/ATP
Actin
Profilin:
Actin
Cofilin
I
Localization
EnaVASP
Arp2/3
D
ProfilinActin
Profilin:Actin
CappingProtein
G-Actin
F-Actin
28
Figure 4 - Barbed end binding proteins regulate actin polymerization
(A) Polymerization barbed ends are capped preventing further polymerization. New barbed ends
are created by nucleation and severing. Ena/VASP proteins are obligate tetramers that promote
filament elongation at the barbed end (B) Ena/VASP proteins form a ternary complex with the
barbed ends of actin filaments, preventing the activity of capping protein. Profilin binding motifs
are found adjacent to G and F-actin binding sites promoting formation of this complex at the
barbed end.
29
1.5.1 Ena/VASP proteins are critical for filopodia formation
Filopodia are fingerlike projections cellular projections containing multiple unbranched
actin filaments oriented with barbed ends pointed outward (Fig. 1). Filopodia are stabilized by
the bundling activity of the crosslinking protein, Fascin (Vignjevic et al., 2006). Filopodia are
implicated in diverse processes including wound healing in mammalian cells in vitro (Conrad et
al., 1993)and D. melanogasterembryos in vivo(Wood et al., 2002), dorsal closure of epithelial
sheets during embryogenesis in D. melanogaster (Jacinto et al., 2000)and C. elegans (Raich et
al., 1999), endothelial cell motility (Sheldon et al., 2009) as well as during many stages of
neuronal development including neuritogenesis (Kwiatkowski et al., 2007)(Dent et al., 2007),
and both repulsive and attractive axon guidance (Gitai et al., 2003)(Bashaw, 2000). Ena/VASP
proteins are found at the tips of filopodia in many contexts, including neuronal growth cones. As
in fibroblasts, loss of Ena/VASP in growth cones dramatically shifts actin network geometry.
These growth cones also display an almost complete lack of filopodia.
The function of filopodia remains somewhat unclear. Many authors have proposed that
filopodia act as a sort of cellular antennae whose function is in receiving extracellular cues
(Chien et al., 1993). Others have proposed that filopodia may be playing a more mechanical role:
Filopodial stabilization in one particular direction may lead to microtubule capture and
subsequent changes in cellular trajectory (Buck and Zheng, 2002)(Dent and Kalil, 2001).
Filopodia also contain integrins enabling them to make adhesive contacts with permissive
substrates (Letourneau, 1983)(Galbraith et al., 2007), though it's unclear whether filopodial
attachment to substratum is important for adhesion per se or if filopodia are acting as sensors
directing cellular motility in the direction of a permissive environment. Adding to the
controversy regarding filopodial function, some authors have claimed that filopodia are
30
dispensable for growth cone guidance in certain contexts (Kim et al., 2002)(Dwivedy et al.,
2007).
1.5.2 Ena/VASP proteins are regulated by localization
The molecular organization of Ena/VASP family members provides insight into how they
link the growth cone microenvironment to actin remodeling. While the EVH1 domain is
dispensable for both the anti-capping and actin polymerization enhancing activities, it is required
for many of the direct protein-protein interactions known to involve Ena/VASP. The EVH1
domain has been demonstrated to bind directly to a large number of proteins through a wellcharacterized interaction with a proline rich motif comprised of the sequence
(D/E)(F/L)PXCDPX(D/E)(D/E), abbreviated FP4(Niebuhr, 1997). This sequence binds in a Vshaped groove found on the surface of EVH1(Prehoda et al., 1999). FP4 motifs of varying levels
of conservation are found in proteins with diverse functions, (Table 1) but show enrichment for
those involved in cellular motility. Another, more recently characterized interaction between
Mena and the focal adhesion protein, Tes (Coutts et al., 2003), reveals that the EVH1 domain can
also bind to non-FP4 ligands by interacting with certain LIM domains. This interaction is able to
compete with EVH1:FP4 interactions, and is thus believed to bind using the same pocket as FP4.
Furthermore, Tes does not appear to bind VASP or EVL, but is instead a specific Mena
interaction, though the biological significance of this paralog specificity is unclear (Bodda et al.,
2007).
Localization of Ena/VASP is integral to function. Expressing fusions of the Ena/VASP
consensus binding motifs found in the ActA protein of the intracellular pathogen Listeria
monocytogenes with a mitochondrial targeting sequence leads to a depletion of Ena/VASP from
the cell periphery and from focal adhesions, sequestering them on the mitochondrial surface
31
(Bear et al., 2000). This sequestration is sufficient to mimic the effects of lack of endogenous
Ena/VASP proteins on fibroblast motility as well as on the geometry of the actin
cytoskeleton(Bear et al., 2002). Furthermore, flies transgenic for a construct expressing FP4
phenocopy the observed axon guidance defects observed in Ena loss of function(Gates et al.,
2007). Interestingly, while these non-physiological EVHl:FP4 interactions are sufficient to
sequester Ena/VASP proteins to the mitochondria, EVH 1:FP4 interactions are insufficient for
full endogenous localization of Ena/VASP. Genetic fusions of EVH 1 with the green fluorescent
protein of A. victoria (EVH1:GFP) only show weak leading edge and focal adhesion localization.
Furthermore treatment of fibroblasts and neuronal growth cones with the actin polymerization
inhibitor Cytochalasin D leads to a loss of Ena/VASP localization at the leading edge of
lamellipodia and from filopodial tips (Bear et al., 2002). These experiments indicate that beyond
the EVHl:FP4 interaction, interaction of Ena/VASP with filamentous actin is crucial for proper
localization. While the interaction of EVH 1 with FP4 motifs is critical for the function of
Ena/VASP, regulation of this interaction remains an area that is unexplored.
32
EVH1 interactions
Partner
Vinculin
Zyxin
Lamellipodin
RIAM
EVH1 binding
Ligand
Ligand
Sequence
Localization
Function
Focal adhesions
Cell-matrix
(Brindle et
adhesion
al., 1996)
Cell-matrix
(Drees et al.,
adhesion
2000)
Lamellipodial
(Krause et
dynamics
al., 2004)
Integrin
(Lafuente et
activation,
al., 2004)
DFPPPP
DFPLP
DFPPPP, ALPPPP
(D/E/F)LPPPP
Focal adhesions
Plasma membrane
Focal adhesions
Reference
adhesion
Palladin
ActA
FatI
Tes
Sema6A
Robo
PFPPPP, DVFPLPP
(D/E)FPPPP(D/E)
DFPPPP
IIN(X) 4 2 M(X)10 M
DVPPKP
EFLPPPP
Focal Adhesions
Bacterial surface
Plasma membrane
Focal adhesions
Plasma membrane
Plasma membrane
Table 1 - EVH1 interactions
33
Cell-cell
(Boukhelifa
adhesion
et al., 2004)
Bacterial
(Niebuhr et
pathogenesis
al., 1997)
Cell polarity,
(Moeller et
motility
al., 2004)
Stress fiber
(Boeda et al.,
formation
2007)
Axon
(Klostermann
guidance
et al., 2000)
Axon
(Bashaw,
guidance
2000)
1.5.3 Ena/VASP proteins are regulated by phosphorylation.
Vertebrate Mena, VASP, and EVL share a conserved phosphorylation site immediately
amino-terminal to the central proline rich region, which is shown to be a substrate for the cyclic
nucleotide dependent Ser/Thr kinase PKA. Phosphorylation at this site (S236 in Mena, S157 in
VASP, and S 160 in EVL. Figure 3b) is important for VASP localization, but seems to have no
effect on the in vitro actin elongation/anti-capping activities of VASP (Barzik et al., 2005)(Benz
et al., 2009). Furthermore, phosphorylation at this site ablates interaction of VASP with the SH3
domain containing protein aI-Spectrin (Benz et al., 2008)and phosphorylation of either Mena or
EVL at their homologous sites ablates their interaction with both Abl and Src(Lambrechts et al.,
2000). A second site for phosphorylation found in Mena and VASP, but not in EVL, is
preferentially a target for the cyclic nucleotide dependent kinase, PKG. Phosphorylation of this
site has no demonstrated effect upon binding to non-cytoskeletal partner proteins but has a clear
dampening effect on the rate of both Mena and VASP mediated actin filament assembly(Benz et
al., 2009). Beyond Ena/VASP phosphorylation, the importance of cyclic nucleotide levels and
PKA/PKG activity in axon guidance has been studied extensively(Lohof et al., 1992)(Ming et
al., 1997)(Song et al., 1997)(Song et al., 1998)(Dontchev and Letourneau, 2002)(Moore and
Kennedy, 2006), raising the interesting but as of yet unproven idea that Ena/VASP localization
and activity may be coordinated in concert with other signaling effectors in response to guidance
cues by cyclic nucleotide signaling.
1.5.4 Ena/VASP proteins regulate neural development in vertebrates and invertebrates
D. melanogasterand C. elegans each have only one Ena/VASP homolog named Ena and
Unc-34 respectively. The founding member of the Ena/VASP family, D. melanogasterEna was
34
originally discovered in a screen for dominant suppressors phenotypes associated with loss of
Ableson non receptor tyrosine kinase homolog. Flies with loss of function mutations in Ena
show guidance defects affecting both non-commissural longitudinal axons and axons which
create both the anterior and posterior commissures. Intriguingly, Ena displays a genetic
interaction with Robo. Heterozygosity for a loss of function allele of Robo in flies causes no
phenotype alone, however when Robo dosage is decreased in the context of homozygous loss of
Ena function, the observed guidance defects are potentiated. In C. elegans, loss of Unc-34 leads
to defects in axon guidance that are similar but less sever than those caused by loss of their sole
Robo homolog (Sax-3). Furthermore, loss of Ena/VASP in worms can ameliorate the gain of
function defects that are the result of Slit overexpression. Together these results argue for a role
of Ena/VASP function in the Slit/Robo but do not distinguish whether Ena/VASP acts upstream
or downstream of Robo (Yu et al., 2002).
The three Ena/VASP vertebrate paralogs exhibit significant functional overlap in the
nervous system, as loss of any individual family member results in relatively restricted or
undetectable defects in major axonal structures in the brain and spinal cord. Mena appears to
have the most significant role in the nervous system, as loss of Mena causes a partially penetrant
defect in formation of the corpus callosum and hippocampal commissure(Lanier et al.,
1999)(Menzies et al., 2004). Deletion of VASP, EVL or both in animals that express Mena does
not cause obvious nervous system defects. Whether the apparent differential requirements for
Mena and EVL/VASP in the nervous system arise for differences in expression or a unique
function for Mena is unclear.
To observe the full requirement for Ena/VASP function in vertebrates, all three paralogs
must be inactivated. Deletion of Mena/VASP/EVL ("mmvvee") results in widespread, highly
35
penetrant defects in nervous system morphogenesis. Neural tube closure defects result in 86% of
mmvvee mice displaying exencephaly wherein the brain develops outside of the head, precluding
analysis of axonogenesis and subsequent axon guidance. Examination of the brains of rare nonexencephalic mmvvee mice reveals formation of ectopias in which cortical neurons are
aberrantly present beyond the meninges, and cobblestone lissencephaly, a defect in which rather
than being organized in layers, the cortex is disorganized (Kwiatkowski et al., 2007).Cobblestone
lissencephalies are a hallmark of the human conditions Walker-Warburg syndrome, Fukuyama
syndrome, and Muscle-Eye-Brain (MEB) disease(Verloes), though no involvement of
Ena/VASP proteins has as of yet been demonstrated in these human diseases. Interestingly, this
defect in cortical organization does not appear to be the result of a cell autonomous defect in
cortical migration. Chimeric mice with roughly fifty percent mmvvee cells and fifty percent wild
type cells show proper layering (Fig. 5a), and no ectopia formation. Furthermore, the mmvvee
neurons show proper positioning within the cortex (Fig. 5 b-e). It appears instead that the defects
in cortical layering arise from the inability of Ena/VASP deficient radial glia to seal the cortex
through contacts made by tufting of their endfeet (Kwiatkowski et al., 2007).
Cortices of mmvvee animals display an almost complete lack of axon fiber tract
formation (Fig. 6b)(Kwiatkowski et al., 2007). When cultured in vitro, cortical neurons normally
elaborate many filopodia that extend radially from the periphery as the neurons attach and
spread. Within 40 hours the majority of neurons have undergone neuritogenesis. One of these
neurites eventually undergoes specification and becomes an axon. Cortical neurons isolated from
mmvvee animals display a smooth round "fried egg" phenotype devoid of filopodia. While
dynamic microtubules still explore the periphery of these neurons, their dynamics are altered,
they cannot be stabilized and neurites do not form. Contrary to the cobblestone cortex phenotype,
36
this defect is cell-autonomous as chimeric embryos show axon formation from wild-type cortical
neurons, but no mutant axon fiber tracts are observed (Fig. 6 c-h) (Dent et al.,
2007)(Kwiatkowski et al., 2009). This defect can be rescued however, by restoring frequent
filopodia formation, which can be achieved by ectopic expression of a formin protein, such as
mDia, or by culturing these neurons on a non-physiological substrate such as laminin. Laminin
rescues filopodia formation by activating integrins which leads to a switch from VAMP2
mediated exocytosis to VAMP7 mediated exocytosis (Gupton and Gertler, 2010). While these
phenotypes are by themselves very interesting they preclude further analysis of the role of
Ena/VASP in cortical axon guidance.
Ena/VASP proteins clearly play important roles in axon guidance in vertebrates and
invertebrates as outlined above. Furthermore, the best described function of Ena/VASP in the
growth cone is in filopodia formation. As of yet however, no study has directly shown how or if
Ena/VASP induced filopodia are important directly in axon guidance. The following work
explores this issue in the context of repulsive guidance in response to the chemorepellent, Slit.
37
Figure 5
-uwv
EC"P
ES 00
ES 066
Ty" skal"Y"
500
E 2oo
i0
38
~
cEb
p a 0.0006
400
I
~
-1n0i-
Figure 5 - Chimeric embryo analysis reveals no cell autonomous defect in position of
mmvvee neurons
(A) mmvvee embryonic stem cells are infected with lentiviruses carrying GFP constructs and
mixed with unlabeled wild-type ES cells before blastocyst injection, leading to the production of
a chimeric embryo. (B) Tbr1 labels neurons in cortical layer VI. (C) Tbrl positive mmvvee
neurons do not migrate significantly farther than Tbri positive wild-type neurons. (D) FoxpI
labels neurons in cortical layers III-V. (E) Foxpl positive mmvvee neurons do not migrate
significantly farther than FoxpI positive wild-type neurons.
39
40
Figure 6
41
Figure 6 - mmvvee animals display defects in neuritogenesis due to a cell autonomous
defect in neuritogenesis
(A) Neurofilament labeling of IOum Coronal sections of el 6.5 control embryo cortex reveals
normal axonal architecture. (B) Similar preparations of mmvvee cortices display a lack of axon
fibers. (C-G) mmvvee neurons cultured in vitro fail to form neurites. (H) Neurons from chimeric
embryos cultured in vitro fail to form neurites indicating cell autonomy of the defect.
42
1.6 References
Adams, R. H., Betz, H., and Puschel, A. W. (1996). A novel class of murine semaphorins with
homology to thrombospondin is differentially expressed during early embryogenesis. Mech.
Dev. 57, 33-45.
Akin, 0. (2008). ScienceDirect - Cell: Capping Protein Increases the Rate of Actin-Based
Motility by Promoting Filament Nucleation by the Arp2/3 Complex. Cell.
Barzik, M., Kotova, T. I., Higgs, H. N., Hazelwood, L., Hanein, D., Gertler, F. B., and Schafer,
D. A. (2005). Ena/VASP proteins enhance actin polymerization in the presence of barbed
end capping proteins. J. Biol. Chem. 280, 28653-28662.
Bashaw, G. J. (2000). Repulsive Axon GuidanceAbelson and Enabled Play Opposing Roles
Downstream of the Roundabout Receptor. Cell 101, 703-715.
Bear, J. E., Svitkina, T. M., Krause, M., Schafer, D. A., Loureiro, J. J., Strasser, G. A., Maly, I.
V., Chaga, 0. Y., Cooper, J. A., Borisy, G. G., et al. (2002). Antagonism between
Ena/VASP proteins and actin filament capping regulates fibroblast motility. Cell 109, 509521.
Bear, J. E., Loureiro, J. J., Libova, I., Fassler, R., Wehland, J., and Gertler, F. B. (2000).
Negative regulation of fibroblast motility by Ena/VASP proteins. Cell 101, 717-728.
Bedell, V. M., Yeo, S.-Y., Park, K. W., Chung, J., Seth, P., Shivalingappa, V., Zhao, J., Obara,
T., Sukhatme, V. P., Drummond, I. A., et al. (2005). roundabout4 is essential for
angiogenesis in vivo. Proc Natl Acad Sci USA 102, 6373-6378.
43
Bentley, D., and Toroian-Raymond, A. (1986). Disoriented pathfinding by pioneer neurone
growth cones deprived of filopodia by cytochalasin treatment. Nature 323, 712-715.
Benz, P. M., Blume, C., Moebius, J., Oschatz, C., Schuh, K., Sickmann, A., Walter, U., Feller, S.
M., and Renne, T. (2008). Cytoskeleton assembly at endothelial cell-cell contacts is
regulated by alphall-spectrin-VASP complexes. J. Cell Biol. 180, 205-219.
Benz, P. M., Blume, C., Seifert, S., Wilhelm, S., Waschke, J., Schuh, K., Gertler, F., MtInzel, T.,
and Renne, T. (2009). Differential VASP phosphorylation controls remodeling of the actin
cytoskeleton. J. Cell. Sci. 122, 3954-3965.
Bodda, B., Briggs, D. C., Higgins, T., Garvalov, B. K., Fadden, A. J., McDonald, N.
Q., and
Way, M. (2007). Tes, a specific Mena interacting partner, breaks the rules for EVH1
binding. Mol. Cell 28, 1071-1082.
Boukhelifa, M., Parast, M. M., Bear, J. E., Gertler, F. B., and Otey, C. A. (2004). Palladin is a
novel binding partner for Ena/VASP family members. Cell Motil. Cytoskeleton 58, 17-29.
Brindle, N. P., Holt, M. R., Davies, J. E., Price, C. J., and Critchley, D. R. (1996). The focaladhesion vasodilator-stimulated phosphoprotein (VASP) binds to the proline-rich domain in
vinculin. Biochem. J. 318 (Pt 3), 753-757.
Brose, K., Bland, K. S., Wang, K. H., Arnott, D., Henzel, W., Goodman, C. S., Tessier-Lavigne,
M., and Kidd, T. (1999). Slit proteins bind Robo receptors and have an evolutionarily
conserved role in repulsive axon guidance. Cell 96, 795-806.
Buck, K. B., and Zheng, J. Q. (2002). Growth cone turning induced by direct local modification
44
of microtubule dynamics. J Neurosci 22, 9358-9367.
Caudy, M., and Bentley, D. (1986). Pioneer growth cone steering along a series of neuronal and
non-neuronal cues of different affinities. J Neurosci 6, 1781-1795.
Charron, F., Stein, E., Jeong, J., McMahon, A. P., and Tessier-Lavigne, M. (2003). The
morphogen sonic hedgehog is an axonal chemoattractant that collaborates with netrin- 1 in
midline axon guidance. Cell 113, 11-23.
Chen, Z., Gore, B. B., Long, H., Ma, L., and Tessier-Lavigne, M. (2008). Alternative splicing of
the Robo3 axon guidance receptor governs the midline switch from attraction to repulsion.
Neuron 58, 325-332.
Cheng, H. J., Nakamoto, M., Bergemann, A. D., and Flanagan, J. G. (1995). Complementary
gradients in expression and binding of ELF-I and Mek4 in development of the topographic
retinotectal projection map. Cell 82, 371-381.
Cheng, H. J., Bagri, A., Yaron, A., Stein, E., Pleasure, S. J., and Tessier-Lavigne, M. (2001).
Plexin-A3 mediates semaphorin signaling and regulates the development of hippocampal
axonal projections. Neuron 32, 249-263.
Chien, C. B., Rosenthal, D. E., Harris, W. A., and Holt, C. E. (1993). Navigational errors made
by growth cones without filopodia in the embryonic Xenopus brain. Neuron 11, 237-251.
Cloutier, J.-F., Sahay, A., Chang, E. C., Tessier-Lavigne, M., Dulac, C., Kolodkin, A. L., and
Ginty, D. D. (2004). Differential requirements for semaphorin 3F and Slit-i in axonal
targeting, fasciculation, and segregation of olfactory sensory neuron projections. J Neurosci
45
24, 9087-9096.
Conrad, P. A., Giuliano, K. A., Fisher, G., Collins, K., Matsudaira, P. T., and Taylor, D. L.
(1993). Relative distribution of actin, myosin I, and myosin II during the wound healing
response of fibroblasts. J. Cell Biol. 120, 1381-1391.
Coutts, A. S., MacKenzie, E., Griffith, E., and Black, D. M. (2003). TES is a novel focal
adhesion protein with a role in cell spreading. J. Cell. Sci. 116, 897-906.
Dent, E. W., Kwiatkowski, A. V., Mebane, L. M., Philippar, U., Barzik, M., Rubinson, D. A.,
Gupton, S., van Veen, J. E., Furman, C., Zhang, J., et al. (2007). Filopodia are required for
cortical neurite initiation. Nat. Cell Biol. 9, 1347-1359.
Dent, E. W., and Kalil, K. (2001). Axon branching requires interactions between dynamic
microtubules and actin filaments. J Neurosci 21, 9757-9769.
Dontchev, V. D., and Letourneau, P. C. (2002). Nerve growth factor and semaphorin 3A
signaling pathways interact in regulating sensory neuronal growth cone motility. J Neurosci
22, 6659-6669.
Drees, B., Friederich, E., Fradelizi, J., Louvard, D., Beckerle, M. C., and Golsteyn, R. M. (2000).
Characterization of the interaction between zyxin and members of the Ena/vasodilatorstimulated phosphoprotein family of proteins. J. Biol. Chem. 275, 22503-22511.
Dwivedy, A., Gertler, F. B., Miller, J., Holt, C. E., and Lebrand, C. (2007). Ena/VASP function
in retinal axons is required for terminal arborization but not pathway navigation.
Development 134, 2137-2146.
46
Fan, J., and Raper, J. A. (1995). Localized collapsing cues can steer growth cones without
inducing their full collapse. Neuron 14, 263-274.
Fan, X., Labrador, J. P., Hing, H., and Bashaw, G. J. (2003). Slit stimulation recruits Dock and
Pak to the roundabout receptor and increases Rac activity to regulate axon repulsion at the
CNS midline. Neuron 40, 113-127.
Ferron, F., Rebowski, G., Lee, S. H., and Dominguez, R. (2007). Structural basis for the
recruitment of profilin-actin complexes during filament elongation by Ena/VASP. EMBO J.
26, 4597-4606.
Galbraith, C. G., Yamada, K. M., and Galbraith, J. A. (2007). Polymerizing actin fibers position
integrins primed to probe for adhesion sites. Science 315, 992-995.
Garrity, P., Rao, Y., Salecker, I., McGlade, J., and Pawson, T. (1996). ScienceDirect - Cell:
Drosophila Photoreceptor Axon Guidance and Targeting Requires the Dreadlocks SH2/SH3
Adapter Protein. Cell.
Gates, J., Mahaffey, J. P., Rogers, S. L., Emerson, M., Rogers, E. M., Sottile, S. L., Van Vactor,
D., Gertler, F. B., and Peifer, M. (2007). Enabled plays key roles in embryonic epithelial
morphogenesis in Drosophila. Development 134, 2027-2039.
Gitai, Z., Yu, T., Lundquist, E., and Tessier-Lavigne, M. (2003). ScienceDirect - Neuron : The
Netrin Receptor UNC-40/DCC Stimulates Axon Attraction and Outgrowth through Enabled
and, in Parallel, Rac and UNC- 115/AbLIM. Neuron.
Gupton, S. L., and Gertler, F. B. (2010). Integrin signaling switches the cytoskeletal and exocytic
47
machinery that drives neuritogenesis. Dev. Cell 18, 725-736.
Hansen, S. D., and Mullins, R. D. (2010). VASP is a processive actin polymerase that requires
monomeric actin for barbed end association. J. Cell Biol. 191, 571-584.
Hemindez-Miranda, L. R., Cariboni, A., Faux, C., Ruhrberg, C., Cho, J. H., Cloutier, J.-F.,
Eickholt, B. J., Parnavelas, J. G., and Andrews, W. D. (2011). Robol regulates semaphorin
signaling to guide the migration of cortical interneurons through the ventral forebrain. J
Neurosci 31, 6174-6187.
Holt, C., Hbpker, V. H., Shewan, D., Tessier-Lavigne, M., and Poo, M.-M. (1999). Growth-cone
attraction to netrin- 1 is converted to repulsion by laminin- 1. Nature 401, 69-73.
Hong, K., Hinck, L., Nishiyama, M., Poo, M. M., Tessier-Lavigne, M., and Stein, E. (1999). A
ligand-gated association between cytoplasmic domains of UNC5 and DCC family receptors
converts netrin-induced growth cone attraction to repulsion. Cell 97, 927-941.
Hu, H., Li, M., Labrador, J. P., McEwen, J., Lai, E. C., Goodman, C. S., and Bashaw, G. J.
(2005). Cross GTPase-activating protein (CrossGAP)/Vilse links the Roundabout receptor to
Rac to regulate midline repulsion. Proc Natl Acad Sci USA 102, 4613-4618.
Isenberg, G., Aebi, U., and Pollard, T. D. (1980). An actin-binding protein from Acanthamoeba
regulates actin filament polymerization and interactions. Nature 288, 455-459.
Ishii, N., Wadsworth, W. G., Stem, B. D., Culotti, J. G., and Hedgecock, E. M. (1992). UNC-6, a
laminin-related protein, guides cell and pioneer axon migrations in C. elegans. Neuron 9,
873-881.
48
Jacinto, A., Wood, W., Balayo, T., Turmaine, M., Martinez-Arias, A., and Martin, P. (2000).
Dynamic actin-based epithelial adhesion and cell matching during Drosophila dorsal closure.
Curr. Biol. 10, 1420-1426.
Jhaveri, D., Saharan, S., Sen, A., and Rodrigues, V. (2004). Positioning sensory terminals in the
olfactory lobe of Drosophila by Robo signaling. Development 131, 1903-1912.
Jones, C. A., Nishiya, N., London, N. R., Zhu, W., Sorensen, L. K., Chan, A. C., Lim, C. J.,
Chen, H., Zhang,
Q.,
Schultz, P. G., et al. (2009). Slit2-Robo4 signalling promotes vascular
stability by blocking Arf6 activity. Nat. Cell Biol. 11, 1325-133 1.
Kaiser, D. A., Vinson, V. K., Murphy, D. B., and Pollard, T. D. (1999). Profilin is predominantly
associated with monomeric actin in Acanthamoeba. J. Cell. Sci. 112 (Pt 21), 3779-3790.
Keleman, K. (2001). ScienceDirect - Neuron: Short- and Long-Range Repulsion by the
Drosophila Unc5 Netrin Receptor. Neuron.
Keleman, K., Rajagopalan, S., Cleppien, D., Teis, D., Paiha, K., Huber, L. A., Technau, G. M.,
and Dickson, B. J. (2002). Comm Sorts Robo to Control Axon Guidance at the Drosophila
Midline. Cell 110, 415-427.
Kennedy, T. E., Serafini, T., la Torre, de, J. R., and Tessier-Lavigne, M. (1994). Netrins are
diffusible chemotropic factors for commissural axons in the embryonic spinal cord. Cell 78,
425-435.
Kidd, T., Bland, K. S., and Goodman, C. S. (1999). Slit is the midline repellent for the robo
receptor in Drosophila. Cell 96, 785-794.
49
Kim, M. D., Kolodziej, P., and Chiba, A. (2002). Growth cone pathfinding and filopodial
dynamics are mediated separately by Cdc42 activation. J Neurosci 22, 1794-1806.
Klostermann, A., Lutz, B., Gertler, F., and Behl, C. (2000). The orthologous human and murine
semaphorin 6A-1 proteins (SEMA6A-1/Sema6A-1) bind to the enabled/vasodilatorstimulated phosphoprotein-like protein (EVL) via a novel carboxyl-terminal zyxin-like
domain. J. Biol. Chem. 275, 39647-39653.
Kolodkin, A. L., Matthes, D. J., and Goodman, C. S. (1993). The semaphorin genes encode a
family of transmembrane and secreted growth cone guidance molecules. Cell 75, 13891399.
Krause, M., Leslie, J. D., Stewart, M., Lafuente, E. M., Valderrama, F., Jagannathan, R.,
Strasser, G. A., Rubinson, D. A., Liu, H., Way, M., et al. (2004). Lamellipodin, an
Ena/VASP ligand, is implicated in the regulation of lamellipodial dynamics. Dev. Cell 7,
571-583.
Kuhn, T. B., Schmidt, M. F., and Kater, S. B. (1995). Laminin and fibronectin guideposts signal
sustained but opposite effects to passing growth cones. Neuron 14, 275-285.
Kwiatkowski, A. V., Garner, C. C., Nelson, W. J., and Gertler, F. B. (2009). Cell autonomous
defects in cortical development revealed by two-color chimera analysis. Mol. Cell. Neurosci.
41, 44-50.
Kwiatkowski, A. V., Rubinson, D. A., Dent, E. W., Edward van Veen, J., Leslie, J. D., Zhang, J.,
Mebane, L. M., Philippar, U., Pinheiro, E. M., Burds, A. A., et al. (2007). Ena/VASP Is
Required for neuritogenesis in the developing cortex. Neuron 56, 441-455.
50
Lafuente, E. M., van Puijenbroek, A. A. F. L., Krause, M., Carman, C. V., Freeman, G. J.,
Berezovskaya, A., Constantine, E., Springer, T. A., Gertler, F. B., and Boussiotis, V. A.
(2004). RIAM, an Ena/VASP and Profilin ligand, interacts with Rap 1-GTP and mediates
Rap1-induced adhesion. Dev. Cell 7, 585-595.
Lambrechts, A., Kwiatkowski, A. V., Lanier, L. M., Bear, J. E., Vandekerckhove, J., Ampe, C.,
and Gertler, F. B. (2000). cAMP-dependent protein kinase phosphorylation of EVL, a
Mena/VASP relative, regulates its interaction with actin and SH3 domains. J. Biol. Chem.
275, 36143-36151.
Lanier, L. M., Gates, M. A., Witke, W., Menzies, A. S., Wehman, A. M., Macklis, J. D.,
Kwiatkowski, D., Soriano, P., and Gertler, F. B. (1999). Mena Is Required for Neurulation
and Commissure Formation. Neuron 22, 313-325.
Laurent, V. (1999). Role of Proteins of the Ena/VASP Family in Actin-based Motility of Listeria
monocytogenes. J. Cell Biol. 144, 1245-1258.
Letourneau, P. C. (1983). Differences in the organization of actin in the growth cones compared
with the neurites of cultured neurons from chick embryos. J. Cell Biol. 97, 963-973.
Lin, J. C., Ho, W.-H., Gurney, A., and Rosenthal, A. (2003). The netrin-GI ligand NGL-1
promotes the outgrowth of thalamocortical axons. Nat. Neurosci. 6, 1270-1276.
Lohof, A. M., Quillan, M., Dan, Y., and Poo, M. M. (1992). Asymmetric modulation of cytosolic
cAMP activity induces growth cone turning. J Neurosci 12, 1253-1261.
Long, H., and Sabatier, C. (2004). ScienceDirect - Neuron : Conserved Roles for Slit and Robo
51
Proteins in Midline Commissural Axon Guidance. Neuron.
Lundstr6m, A., Gallio, M., Englund, C., Steneberg, P., Hemphala, J., Aspenstr6m, P., Keleman,
K., Falileeva, L., Dickson, B. J., and Samakovlis, C. (2004). Vilse, a conserved Rac/Cdc42
GAP mediating Robo repulsion in tracheal cells and axons. Genes Dev. 18, 2161-2171.
Luo, Y., Raible, D., and Raper, J. A. (1993). Collapsin: a protein in brain that induces the
collapse and paralysis of neuronal growth cones. Cell 75, 217-227.
Ly, A., Nikolaev, A., Suresh, G., Zheng, Y., Tessier-Lavigne, M., and Stein, E. (2008). DSCAM
is a netrin receptor that collaborates with DCC in mediating turning responses to netrin- 1.
Cell 133, 1241-1254.
Lyuksyutova, A. I., Lu, C.-C., Milanesio, N., King, L. A., Guo, N., Wang, Y., Nathans, J.,
Tessier-Lavigne, M., and Zou, Y. (2003). Anterior-posterior guidance of commissural axons
by Wnt-frizzled signaling. Science 302, 1984-1988.
Ma, L., and Tessier-Lavigne, M. (2007). Dual branch-promoting and branch-repelling actions of
Slit/Robo signaling on peripheral and central branches of developing sensory axons. J
Neurosci 27, 6843-685 1.
Marquardt, T., Shirasaki, R., Ghosh, S., Andrews, S. E., Carter, N., Hunter, T., and Pfaff, S. L.
(2005). Coexpressed EphA receptors and ephrin-A ligands mediate opposing actions on
growth cone navigation from distinct membrane domains. Cell 121, 127-139.
Matsunaga, M., Hatta, K., Nagafuchi, A., and Takeichi, M. (1988). Guidance of optic nerve
fibres by N-cadherin adhesion molecules. Nature 334, 62-64.
52
Menzies, A. S., Aszodi, A., Williams, S. E., Pfeifer, A., Wehman, A. M., Goh, K. L., Mason, C.
A., Fassler, R., and Gertler, F. B. (2004). Mena and vasodilator-stimulated phosphoprotein
are required for multiple actin-dependent processes that shape the vertebrate nervous system.
J Neurosci 24, 8029-8038.
Ming, G. L., Song, H. J., Berninger, B., Holt, C. E., Tessier-Lavigne, M., and Poo, M. M. (1997).
cAMP-dependent growth cone guidance by netrin-1. Neuron 19, 1225-1235.
Moeller, M. J., Soofi, A., Braun, G. S., Li, X., Watzl, C., Kriz, W., and Holzman, L. B. (2004).
Protocadherin FATI binds Ena/VASP proteins and is necessary for actin dynamics and cell
polarization. EMBO J. 23, 3769-3779.
Moore, S. W., and Kennedy, T. E. (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, 2419-2423.
Mullins, R. D., Stafford, W. F., and Pollard, T. D. (1997). Structure, subunit topology, and actinbinding activity of the Arp2/3 complex from Acanthamoeba. J. Cell Biol. 136, 331-343.
Mullins, R. D., Heuser, J. A., and Pollard, T. D. (1998). The interaction of Arp2/3 complex with
actin: nucleation, high affinity pointed end capping, and formation of branching networks of
filaments. Proc Natl Acad Sci USA 95, 6181-6186.
Nguyen Ba-Charvet, K. T., Brose, K., Ma, L., Wang, K. H., Marillat, V., Sotelo, C., TessierLavigne, M., and Chedotal, A. (2001). Diversity and specificity of actions of Slit2
proteolytic fragments in axon guidance. J Neurosci 21, 4281-4289.
53
Niebuhr, K. (1997). A novel proline-rich motif present in ActA of Listeria monocytogenes and
cytoskeletal proteins is the ligand for the EVH 1 domain, a protein module present in the
Ena/VASP family. EMBO J. 16, 5433-5444.
Niebuhr, K., Ebel, F., Frank, R., Reinhard, M., Domann, E., Carl, U. D., Walter, U., Gertler, F.
B., Wehland, J., and Chakraborty, T. (1997). A novel proline-rich motif present in ActA of
Listeria monocytogenes and cytoskeletal proteins is the ligand for the EVHI domain, a
protein module present in the Ena/VASP family. EMBO J. 16, 5433-5444.
Nishiyama, M., Hoshino, A., Tsai, L., Henley, J. R., Goshima, Y., Tessier-Lavigne, M., Poo, M.M., and Hong, K. (2003). Cyclic AMP/GMP-dependent modulation of Ca2+ channels sets
the polarity of nerve growth-cone turning. Nature 423, 990-995.
Plump, A. S., Erskine, L., Sabatier, C., Brose, K., Epstein, C. J., Goodman, C. S., Mason, C. A.,
and Tessier-Lavigne, M. (2002). Slitl and Slit2 cooperate to prevent premature midline
crossing of retinal axons in the mouse visual system. Neuron 33, 219-232.
Pollard, T. D., and Mooseker, M. S. (1981). Direct measurement of actin polymerization rate
constants by electron microscopy of actin filaments nucleated by isolated microvillus cores.
J. Cell Biol. 88, 654-659.
Prasad, A., Fernandis, A. Z., Rao, Y., and Ganju, R. K. (2004). Slit protein-mediated inhibition
of CXCR4-induced chemotactic and chemoinvasive signaling pathways in breast cancer
cells. J. Biol. Chem. 279, 9115-9124.
Prehoda, K. E., Lee, D. J., and Lim, W. A. (1999). Structure of the enabled/VASP homology 1
domain-peptide complex: a key component in the spatial control of actin assembly. Cell 97,
54
471-480.
Raich, W. B., Agbunag, C., and Hardin, J. (1999). Rapid epithelial-sheet sealing in the
Caenorhabditis elegans embryo requires cadherin-dependent filopodial priming. Curr. Biol.
9, 1139-1146.
Rothberg, J. M., Hartley, D. A., Walther, Z., and Artavanis-Tsakonas, S. (1988). slit: an EGFhomologous locus of D. melanogaster involved in the development of the embryonic central
nervous system. Cell 55, 1047-1059.
Rothberg, J. M., Jacobs, J. R., Goodman, C. S., and Artavanis-Tsakonas, S. (1990). slit: an
extracellular protein necessary for development of midline glia and commissural axon
pathways contains both EGF and LRR domains. Genes Dev. 4, 2169-2187.
Schafer, D. A., Jennings, P. B., and Cooper, J. A. (1996). Dynamics of capping protein and actin
assembly in vitro: uncapping barbed ends by polyphosphoinositides. J. Cell Biol. 135, 169179.
Schmid, B. C., Rezniczek, G. A., Fabjani, G., Yoneda, T., Leodolter, S., and Zeillinger, R.
(2007). The neuronal guidance cue Slit2 induces targeted migration and may play a role in
brain metastasis of breast cancer cells. Breast Cancer Res. Treat. 106, 333-342.
Seeger, M., Tear, G., Ferres-Marco, D., and Goodman, C. S. (1993). Mutations affecting growth
cone guidance in Drosophila: genes necessary for guidance toward or away from the
midline. Neuron 10, 409-426.
Seiradake, E., Philipsborn, von, A. C., Henry, M., Fritz, M., Lortat-Jacob, H., Jamin, M.,
55
Hemrika, W., Bastmeyer, M., Cusack, S., and McCarthy, A. A. (2009). Structure and
functional relevance of the Slit2 homodimerization domain. EMBO Rep. 10, 736-741.
Sheldon, H., Andre, M., Legg, J. A., Heal, P., Herbert, J. M., Sainson, R., Sharma, A. S.,
Kitajewski, J. K., Heath, V. L., and Bicknell, R. (2009). Active involvement of Robol and
Robo4 in filopodia formation and endothelial cell motility mediated via WASP and other
actin nucleation-promoting factors. FASEB J. 23, 513-522.
Simpson, J. H., Bland, K. S., Fetter, R. D., and Goodman, C. S. (2000). Short-range and longrange guidance by Slit and its Robo receptors: a combinatorial code of Robo receptors
controls lateral position. Cell 103, 1019-1032.
Song, H. (1998). Conversion of Neuronal Growth Cone Responses from Repulsion to Attraction
by Cyclic Nucleotides. Science 281, 1515-1518.
Song, H., Ming, G., He, Z., Lehmann, M., McKerracher, L., Tessier-Lavigne, M., and Poo, M.
(1998). Conversion of neuronal growth cone responses from repulsion to attraction by cyclic
nucleotides. Science 281, 1515-1518.
Song, H. J., Ming, G. L., and Poo, M. M. (1997). cAMP-induced switching in turning direction
of nerve growth cones. Nature 388, 275-279.
Sporns, 0., Tononi, G., and K6tter, R. (2005). The human connectome: A structural description
of the human brain. PLoS Comput. Biol. 1, e42.
Stein, E., Lane, A. A., Cerretti, D. P., Schoecklmann, H. 0., Schroff, A. D., Van Etten, R. L., and
Daniel, T. 0. (1998). Eph receptors discriminate specific ligand oligomers to determine
56
alternative signaling complexes, attachment, and assembly responses. Genes Dev. 12, 667678.
Stein, E., and Tessier-Lavigne, M. (2001). Hierarchical organization of guidance receptors:
silencing of netrin attraction by slit through a Robo/DCC receptor complex. Science 291,
1928-1938.
Stevens, A. (2002). Integrins Regulate Responsiveness to Slit Repellent Signals. The Journal of
neuroscience.
van Rheenen, J., Song, X., van Roosmalen, W., Cammer, M., Chen, X., Desmarais, V., Yip, S.C., Backer, J. M., Eddy, R. J., and Condeelis, J. S. (2007). EGF-induced PIP2 hydrolysis
releases and activates cofilin locally in carcinoma cells. J. Cell Biol. 179, 1247-1259.
Varshney, L. R., Chen, B. L., Paniagua, E., Hall, D. H., and Chklovskii, D. B. (2011). Structural
properties of the Caenorhabditis elegans neuronal network. PLoS Comput. Biol. 7,
e1001066.
Verloes, A. Verloes: Lissencephaly, generic term. Orphanet encycloped... - Google Scholar.
Vignjevic, D., Kojima, S., and Aratyn, Y. (2006). Role of fascin in filopodial protrusion. The
Journal of cell ....
Wang, K., Brose, K., Arnott, D., Kidd, T., and Goodman, C. (1999). ScienceDirect - Cell:
Biochemical Purification of a Mammalian Slit Protein as a Positive Regulator of Sensory
Axon Elongation and Branching. Cell.
Wong, K., Ren, X. R., Huang, Y. Z., Xie, Y., Liu, G., Saito, H., Tang, H., Wen, L., Brady-
57
Kalnay, S. M., Mei, L., et al. (2001). Signal transduction in neuronal migration: roles of
GTPase activating proteins and the small GTPase Cdc42 in the Slit-Robo pathway. Cell 107,
209-221.
Wood, W., Jacinto, A., Grose, R., Woolner, S., Gale, J., Wilson, C., and Martin, P. (2002).
Wound healing recapitulates morphogenesis in Drosophila embryos. Nat. Cell Biol. 4, 907912.
Yamada, K. M. (1971). ULTRASTRUCTURE AND FUNCTION OF GROWTH CONES AND
AXONS OF CULTURED NERVE CELLS. J. Cell Biol. 49, 614-635.
Yang, L., and Bashaw, G. J. (2006). Son of sevenless directly links the Robo receptor to rac
activation to control axon repulsion at the midline. Neuron 52, 595-607.
Yu, T. W., Hao, J. C., Lim, W., Tessier-Lavigne, M., and Bargmann, C. I. (2002). Shared
receptors in axon guidance: SAX-3/Robo signals via UNC-34/Enabled and a Netrinindependent UNC-40/DCC function. Nat. Neurosci. 5, 1147-1154.
Yuasa-Kawada, J., Kinoshita-Kawada, M., Rao, Y., and Wu, J. Y. (2009a). Deubiquitinating
enzyme USP33/VDU1 is required for Slit signaling in inhibiting breast cancer cell
migration. Proc Natl Acad Sci USA 106, 14530-14535.
Yuasa-Kawada, J., Kinoshita-Kawada, M., Wu, G., Rao, Y., and Wu, J. Y. (2009b). Midline
crossing and Slit responsiveness of commissural axons require USP33. Nat. Neurosci. 12,
1087-1089.
58
Chapter 2
Mice genetically deficient in Mena, VASP, and EVL display axon pathfinding
defects suggestive of impaired sensitivity to repulsive guidance cues
59
2.1 A bstract..................................................................................................................................
61
2.2 Introduction...........................................................................................................................61
2.3 R esults ....................................................................................................................................
62
2.3.1 mm vvee m ice display axon pathfinding defects...........................................................
62
2.3.2 Organotypic spinal cord explants show normal architecture ........................................
68
2.3.3 mm vvee m ice display a disorganized dorsal funiculus................................................
71
2.3.4 mmvvee growth cones appear devoid of filopodia ex vivo
73
.....................
2.4 Discussin..y.....O............................................76
2.4 D iscussion .............................................................................................................................. 76
2.5 M ethods..................................................................................................................................
2.6 R eferences..............................................................................................................................81
60
80
2.1 Abstract
Defects in repulsive axon pathfinding have long been known to require Ena/VASP in
invertebrates. Mice deficient in Mena and VASP display defects in formation of the optic
chiasm, though these are relatively mild and partially resolve during development potentially due
to the expression of EVL. Here I report that in mice deficient for Mena, VASP, and EVL,
pathfinding defects are evident in afferent axons emanating from the peripheral nervous system
derived dorsal root ganglia. Pathfinding defects are also apparent in central nervous system
derived motor axons. The defects in PNS pathfinding show a striking similarity to those
observed in either Slit1/Slit2 double knockout mice or Robol/Robo2 double knockout mice,
implying a role for Ena/VASP in Slit/Robo mediated repulsive guidance.
2.2 Introduction
Defects in axonal pathfinding are associated with a loss or reduction of Ena/VASP
expression in D. melanogaster,C. elegans, and M musculus (Gertler et al., 1995)(Gitai et al.,
2003)(Menzies et al., 2004). Phenotypic overlaps with defects observed in Slit/Robo mutants are
present in many of these examples, suggesting a role for Ena/VASP in repulsive axon guidance.
Axonal pathfinding defects observed in invertebrate Ena/VASP mutants are potentiated by
genetic interactions with Slit or Robo, more directly indicating a role of Ena/VASP in Slit/Robo
signaling. This model is supported by in vitro studies demonstrating direct binding between
Ena/VASP proteins and the cytoplasmic tail of Robo. Notably, however, Ena/VASP loss of
function defects are generally less severe than those observed in Slit/Robo mutants suggesting
61
that other pathways may work in parallel, or that Ena/VASP may not be required for all aspects
of a repulsive guidance response in invertebrates.
2.3 Results
2.3.1 mmvvee mice display axon pathfinding defects
Antibody labeling of the neuronal marker neurofilament, of the spinal cords of elO.5
mmvvee mice reveals that these neurons are able to bypass the neuritogenesis defect observed in
other neuronal types (Fig. lb, arrowheads). Interestingly, while the midline of littermate control
animals is free of axons, there are multiple axons extending aberrantly toward the midline in
mmvvee animals (Fig. Ib, arrows). Although the origin of these fibers is difficult to comment on
unequivocally in this preparation, they appear to be DRG derived.
DRG axon pathfinding defects have recently been described in mice lacking either Slit or
Robo (Ma and Tessier-Lavigne, 2007). DRG neurons predominantly express two homologs of
Slit, Sliti and Slit2, and two homologs of Robo, Robol and Robo2. Loss of both Slit1 and Slit2
leads to an aberrant invasion of afferent DRG axons toward the spinal cord midline, easily
observed in transverse sections of mutant embryos. These phenotypes are mirrored closely by
phenotypes observed in e13.5 mmvvee animals as seen by antibody labeling with the axon
specific marker, Gap43. In e13.5 littermate controls, which are null for VASP an EVL, but wildtype or heterozygous for Mena (MMvvee and Mmvvee, respectively) DRGs appear well formed
(Fig. 2a, arrow) extending their afferent axons to the DREZ, where the dorsal funiculus appears
well organized. In contrast, mmvvee mice display profound guidance defects. While DRGs still
appear relatively well formed (Fig.2b, arrow), multiple axons invade the spinal cord midline
(Fig. 2b, box). Furthermore, these axons seem to extend directly from the dorsal funiculus
62
implying that they are DRG derived. Notably, these phenotypes are highly similar to what is seen
in either Slitl/Sit2 or Robol/Robo2 double knockout animals. Interestingly, a subset of
misguided axons have been observed reversing course (Fig. 2c, arrows).
Figure 1
IMMvveeI
Immvvee
4-a
E
0
L..
Figure 1- mmvvee animals display pathfinding defects.
(A) Dorsal view of whole mount immunostaining of e10.5 littermate embryos with the axonal
marker neurofilament reveals well organized axons extending from DRGs (B) mmvvee animals
in similar preparations display disorganized axon bundles emanating from DRGs (arrowhead)
and multiple axons fibers aberrantly invading the spinal cord midline (arrows)
63
Figure 2
MMvvee
mmvvee
Ai'
64
Figure 2- mmvvee animals display pathfinding defects
(A)Immunostaining with the axonal marker, Gap43 of Transverse 100um Vibratome sections
taken from e 13.5 littermate control embryos display normal guidance and a well-organized
dorsal funiculus. (B) mmvvee embryos in similar preparations display normal looking DRGs
(arrow), but have multiple axon fibers invading the spinal cord midline (box). Increased
magnification of region in box shows extent of axonal pathfinding errors. Some axons are seen
wandering or reversing course (arrows)
65
Failure to be repelled at the spinal cord midline may represent a neuronal defect in Robo
localization. To determine whether Robo expression is present during a time period when
invasion is occurring, embryos were examined one day earlier, at e12.5. Co-labeling with
antibodies reactive to Robol and Gap43, reveals that misguided axons display immunoreactivity
to Robol antibody (Figure 3e, arrow). An increased magnification of the same sections shows
that Robo 1 appears to be expressed in what appear to be growth cones, indicating that this
guidance defect may not be a result of Robo mis-localization. These preparations do not
determine whether Robo is localized on the growth cone surface or within a cytoplasmic pool.
Examination of these sections further reveals that while e12.5 littermate controls display an
organized dorsal funiculus and no invasion of axons toward the spinal cord midline (Figure 3ac), e12.5 mmvvee animals display an aberrant invasion of axons toward the spinal cord midline
(Figure 3d, arrow). The width of the congregation of afferent axons appears increased near the
dorsal funiculus in mmvvee mice indicating poor axonal association (Figure 3d, asterisk). This
may be a result of the process of axonal bundling called fasciculation which positively correlates
with the formation of intra-axonal filopodial contacts (Raper and Bastiani, 1983).
I also observed a severe defect in the guidance of what appear to be spinal motor axons
(Figure 3d, arrowhead). It has been previously demonstrated that explants of spinal motor axons
can be repelled by the presence of Slit indicating another potential context for the involvement of
Ena/VASP in Slit/Robo signaling (Brose et al., 1999). Interestingly in the characterization of the
Slit1/Slit2 and Robol/Robo2 knockout animals, this defect is not observed in spinal motor
axons, indicating a potential role of Ena/VASP in modulating response mediated by other
molecules.
66
Figure 3
IGap43
IROBO-1
IMerge
E
E
E
E
Figure 3 - mmvvee embryos display defects in both CNS and PNS pathfinding.
(A-C) Gap43 and Robol staining of 100um Vibratome sections of e12.5 littermate embryos
reveal normal spinal cord guidance. (B-F) mmvvee animals display pathfinding defects in DRG
guidance (box, arrows) and in spinal motor axons (arrowheads). (G-I) Increased magnification
reveals immunoreactivity misguided axons to anti-Robol antibody.
67
2.3.2 Organotypic spinal cord explants show normal architecture
In wild-type mice, afferent axons extend directly from the dorsal root ganglia (DRGs,
Fig. 4a) to the dorsal portion of the spinal cord, termed the dorsal funiculus (Fig. 4b, arrow).
Axons do not directly enter the spinal cord, but branch, sending one projection rostrally, toward
the head and the other caudally, toward the tail (Fig.4a). Later these projections sprout collateral
branches, which enter the spinal cord, synapsing with interneurons or motor neurons. To
visualize axons extending toward and branching at the dorsal funiculus, I infected organotypic
spinal cord explants with Herpes Simplex Viruses (HSVs) genetically engineered to express the
fluorescent protein, tdTomato. Within 16 hours of viral treatment, fluorescent protein expression
is readily visible in both the perikarya and the extended neuronal processes of the DRGs (Fig.
4b). In this preparation, axons appear to be extending toward the dorsal funiculus (Fig. 4b
arrowheads), as well as in the region of the dorsal funiculus where axons branch and extend
antiparallel rostrally and caudally (Fig. 4b, arrows) recapitulating physiologically normal
neuroanatomy observed in vivo.
68
Figure 4
Dorsal View
A
To Head
*.****
e****I
i0
@0
e****
I***
gO
0
.e.e
0*.00..
*e'e~e
-ee'*
---
T T
*To Tail
69
@0
=
ml
Figure 1- Organotypic spinal cord explants infected with Herpes Simplex Virus reveal
DRG derived axonal pathways.
(A) Cartoon depiction of guidance in the Dorsal Root Ganglia. (B) Axons extending toward the
spinal cord from DRGs (arrowheads) before reaching the dorsal funiculus of the spinal cord
(arrows) where they branch; the daughter branches turn sharply toward the head and tail.
70
2.3.3 mmvvee mice display a disorganized dorsal funiculus
To determine the role of Ena/VASP in the organization of the dorsal funiculus at higher
resolution, I once again examined organotypic spinal explants. Spinal cords of littermate controls
were cultured alongside the spinal cords of mmvvee animals in the presence of tdTomato
expressing HSV for 16 hours before fixation and visualization. High-resolution confocal
microscopy of the littermate dorsal funiculus reveals a well-organized structure. Upon reaching
the dorsal funiculus axons branch and these branches turn at sharp angle, extending in a direction
completely opposite each other and orthogonal to the axons emanating from the dorsal root
ganglia (Figure 5b). This is very similar to what has been reported in studies of labeling of
individual axons with the lipophilic dye DiI in fixed tissue(Ma and Tessier-Lavigne, 2007). In
contrast to this, the dorsal funiculus of mmvvee animals is highly disorganized (Figure 5c).
While many axons extend generally along an anterior-posterior axis, the width of the dorsal
funiculus is greatly increased (compare width between dashed lines in Fig. 5b vs 5c) and the
turning angle is often wrong. Again, these phenotypes resemble those seen the dorsal funiculus
of Slit1/Slit2 knockout animals, in which roughly fifty percent of axons turn generally anteriorposterior upon entering the dorsal funiculus, but often at an angle of less than the appropriate
90 0(Ma and Tessier-Lavigne, 2007).
71
Figure 5
Dorsal View
A
* To Head
er*r
C,
'-o.
0
WO
=
=
T
To Tail
IMMvveeI
I mmvveel
-
Am C1
I
Figure 5- Organotypic spinal cord explants infected with Herpes Simplex Virus reveal
disorganized dorsal funiculus of mmvvee spinal cords.
(A) Cartoon showing location of dorsal funiculus magnified(dashed box). (B-C) 40X
magnification of dorsal funiculus of mmvvee animals reveals profound disorganization.
72
2.3.4 mmvvee growth cones appear devoid of filopodia ex vivo.
Growth cones become more complex by increasing the total f-actin content and number
of filopodia present when they reach points along their pathway at which a guidance decision
must be made(Moon and Gomez, 2005). As the dorsal funiculus represents a point at which
DRG derived growth cones must turn, I next examined growth cone morphology in this region.
Using e 12.5 organotypic spinal cord cultures infected at a lower viral titer than previous
experiments, I was able to observe individual growth cone morphologies ex vivo. DRG derived
growth cones in littermate control animals show multiple filopodia (Fig. 6a, arrows) and fanshaped lamellipodia (Fig. 6a, arrowhead) upon reaching the dorsal funiculus, when a turn has
just been made. Contrasting this, growth cones in mmvvee derived dorsal funiculi appear to be
devoid of filopodia.
73
mmvvee
IMMvvee
'1
c~E
C
-I
CD
............
......
..
....
........
..
....
Figure 6- Organotypic spinal cord explants infected with Herpes Simplex Virus reveal
growth cone morphologies ex vivo.
(A) 60X magnification of littermate growth cones in the dorsal funiculus reveals filopodia
(arrows) and lamellipodia (arrwoheads). (B) mmvvee growth cones in similar preparations
appear to lack growth cone filopodia (arrowhead).
75
2.4 Discussion
In this chapter, I have described axon guidance defects associated with genetic loss of
Ena/VASP. It has long been appreciated that in invertebrates, loss of the one homologue of
Ena/VASP leads to axon guidance phenotypes. As of yet, the only described neuronal
phenotypes associated with complete genetic loss of Ena/VASP have been in neuritogenesis, a
phenotype not seen in invertebrates lacking Ena. The extent of the axonal pathfinding defects
observed in the peripheral and central nervous systems argues that the role of Ena/VASP is
conserved and potentially expanded in vertebrates.
Neurons in the central and peripheral nervous system of mmvvee mice display sufficient
axonogenesis to allow analysis of axon guidance. Neurite formation can be rescued in mmvvee
cortical neurons by culturing them on the non-physiological substrate, laminin. It has been long
appreciated that while the cortex is a laminin-poor environment, neurons of the Dorsal Root
Ganglia are intermixed with Schwann cells and boundary cap derived satellite glia, both cell
types that directly contact neuronal perikarya and produce laminin(Duband and Thiery,
1987)(Schiff and Rosenbluth, 1986). The lack of a defect in neuritogenesis may be the result of
rescue by this laminin rich environment, or it may reveal that cortical neuritogenesis is more
complicated than other systems in which axon formation has been studied, and therefore requires
the activity of Ena/VASP proteins.
Ena/VASP deficient mice display pathway finding defects in axons derived from both
peripheral and central nervous system neurons. In the peripheral nervous system, these defects
very closely phenocopy those seen in Slitl/Slit2 or Robol/Robo2 knockout mice. Similar to what
I see in mammals, in C. elegans loss of function mutations in the one Ena/VASP homolog, Unc76
34, result in multiple axon guidance defects. Ventral guidance of the AVM axon and midline
guidance of the PVQ axons is impaired in a very similar fashion in both Unc-34 and Sax-3
(Robo) mutants. Furthermore, these defects are not attenuated in Unc-34/Sax-3 double mutants,
implying that the role of Ena/VASP in guidance of these neurons is related uniquely to Robo
signaling (Yu et al., 2002). In drosophila on the other hand, loss of function mutations in the one
Ena/VASP homolog, Ena, result in axon guidance defects phenotypically distinct from those
seen in loss of function mutants for either Slit or Robo. Genetic evidence shows that Ena and
Slit/Robo are in the same pathway, but given the non-identical phenotypes seen, Ena and
Slit/Robo may also have non-overlapping roles in axon guidance in the drosophila CNS. The
high degree of phenotypic overlap observed argues for a conserved central role of Ena/VASP
proteins in Slit/Robo signaling in the PNS. Intriguingly, a small subset of DRG derived mmvvee
axons that initially extend aberrantly toward the spinal cord midline only to loop, reversing
trajectory. This may be the result of random wandering of mmvvee growth cones, or some
growth cones may reach a critical concentration of Slit sufficient to effect repulsion. The latter
hypothesis would argue for a role of Ena/VASP proteins in sensitivity to Slit, upstream of Robo.
Central nervous system derived motor neurons also display profound pathfinding defects.
Motor neurons send efferent axons out of the spinal cord that eventually synapse with muscles at
neuromuscular junctions. Rather than projecting out of the spinal cord, mmvvee derived motor
axons extend aberrantly toward the midline. Spinal motor neurons are repelled by the presence of
Slit proteins in vitro (Brose et al., 1999), indicating that midline repulsion may also be Slit/Robo
mediated in this context. Interestingly, there is not a similar phenotype observed in the Slitl/2 or
Robol/2 knockout animals. This could be the result of the fact that while Robol/2 are the
predominant family members expressed in the dorsal root ganglia and the spinal cord roof plate
77
only expresses Slit1/2, motor neurons express Robo3, and the spinal cord floorplate expresses
Slit3. The presence of this receptor/ligand combination may be enough to prevent aberrant
guidance in the spinal motor axons. This would be analogous to what has been observed in spinal
commissural axons. Slitl/Sit2 double mutant animals show no commissural axon midline
guidance defects(Plump et al., 2002), however animals genetically deficient for all three Slit
homologs show far more severe guidance defects including axonal stalling and re-crossing of the
midline(Long and Sabatier, 2004). While no study to date has examined guidance in spinal motor
axons in Slit 1/2/3 or Robol/2/3 knockout animals, if they were to have a motor guidance defect
similar to what I have observed in the mmvvee knockout mice, it would imply a role for
Ena/VASP in sensitivity or response mediated by all three neuronal Robos. Another equally
likely possibility is that Ena/VASP proteins also mediate response to other repulsive guidance
molecules. As Semaphorins and Ephrins both have demonstrated roles in motor axon
pathfinding(Luria et al., 2008)(Huber et al., 2005), a future area of investigation in the laboratory
will be the role of Ena/VASP in response to these guidance molecules.
It is interesting that while guidance is clearly impaired in mmvvee animals, it is clearly
not abolished altogether. This is best illustrated in the dorsal funiculus of organotypic spinal cord
explants. While axons in mmvvee animals are clearly disorganized in the dorsal funiculus, and
most do not adopt the proper turning angle, it's obvious that many of them still turn in a
generally anterior or posterior fashion. Interestingly, using a different technique of axonal
labeling with the lipophilic dye, DiI, it has been reported that Slit1/2 and Robol/2 knockout
animals have similar phenotypes in which roughly fifty percent of axons make the appropriate
turn, although many of those not at an angle completely orthogonal to the incoming axon.
Together, these phenotypes imply that there is a Slit/Robo independent guidance pathway acting
78
on DRG derived growth cones in the dorsal funiculus. It remains unclear whether this pathway is
comprised of a known diffusible repulsive cue that is mechanistically similar in action to Slit or
another ligand or receptor yet to be described. Alternatively, there may be substrate bound
attractive pathways along the dorsal funiculus mechanistically quite different than Slit/Robo
signaling.
Finally, I have observed ex vivo that mmvvee growth cones appear devoid of filopodia.
Similar Club-like growth cone morphologies have been reported in physiological contexts in
which guidance decisions likely do not require exquisite sensitivity. Growth cones devoid of
filopodia appear in axons which grow along tracts already made by "pioneer axons" (Kim et al.,
199 1)(Lopresti and Macagno, 1973)(Nordlander, 1987). During X laevis midline guidance,
commissural interneurons approach the midline in an undeviating pathway and display a clublike morphology with few long filopodia. Upon reaching the midline they turn longitudinally at
an angle orthogonal to the angle of incidence. Concomitant with this turning, F-actin rich growth
cone filopodia appear and the growth cones become repelled by the midline, a process mediated
by Slit/Robo(Moon and Gomez, 2005). This suggests the exciting possibility that Ena/VASP
proteins promote filopodia emergence or elongation in a context dependent fashion, crucial for
responsiveness to the repulsive activities of Slit/Robo. Further, my observation that mmvvee
growth cones in the dorsal funiculus appear to lack filopodia suggests that the defects observed
in mmvvee mice may be at least partially the result of a failure to induce filopodia in a setting
crucial for proper axon guidance. The notion that Ena/VASP induced filopodia formation may
only be required at certain points in an axons path, may also underlie an unsolved controversy
stemming from the observation that some axons are not misguided in the absence of Ena/VASP
(Dwivedy et al., 2007).
79
2.5 Methods
For organotypic spinal cord explants, embryos were dissected from timed-pregnancies of
Swiss-Webster female mice 12.5 days post-conception (e12.5). Care was taken during dissection
of embryonic spinal cords to not disconnect the dorsal root axons, those spinal cords in which
significant perturbation to the spinal cord occurred were not used in these analyses. Spinal cords
were taken with DRGs intact and floated onto hanging Millipore 35mm tissue culture inserts
with .45um pores. Spinal cords were oriented dorsal side up, and the medium was then removed
from the top of the membrane, leaving only the bottom exposed to medium. The spinal cords
were then exposed to roughly 2x1 04 infectious units of Herpes Simplex Virus (Rachael Neve,
McGovern institiute viral core facility) carrying a transgene for expression of the fluorescent
protein, Tandem-dimer Tomato, in a total volume of 1Oul. These cultures were incubated for 16
hours in standard 5% C02 tissue culture incubator before fixation in 4% paraformaldehyde in IX
Tris-buffered Saline. Spinal cords were then floated off of the membranes and placed dorsal side
down in glass bottom dishes in Fluoromount G. Images were taken on a Nikon AIR scanning
confocal using a 20X DICN2 objective. For growth cone imaging, everything was done
identically excepting the viral titer which was reduced 10-fold, and imaging was done using a
40X DICN2 water immersion objective.
For transverse section imaging, e12.5 and e13.5 mouse embryos were fixed overnight in
4% paraformaldehyde in phosphate buffered saline, then washed 3 times in PBS the following
morning. Embryos were embedded in blocks of 4% low-melting point agarose and 100um
sections taken on a Vibratome. Free floating sections from the brachial region of the embryo
were transferred to individual wells of a 12 well dish, where they were immunostained using
80
standard ABC protocols as previously described. Primary antibody incubations were done
overnight at 40 C in PBS + 1%normal donkey serum, +.025% triton-x1OO using the following
dilutions: Neurofilament, 1:500 (Mouse monoclonal, DSHB 2H3). Gap43, 1:1000 (Rabbit
polyclonal, Abcam ab16053). Robol, 1:500 (Goat polyclonal, Abcam ab1 11726).
2.6 References
Brose, K., Bland, K., Wang, K., Amott, D., and Henzel, W. (1999). ScienceDirect - Cell: Slit
Proteins Bind Robo Receptors and Have an Evolutionarily Conserved Role in Repulsive
Axon Guidance. Cell.
Duband, J. L., and Thiery, J. P. (1987). Distribution of laminin and collagens during avian neural
crest development. Development 101, 461-478.
Dwivedy, A., Gertler, F. B., Miller, J., Holt, C. E., and Lebrand, C. (2007). Ena/VASP function
in retinal axons is required for terminal arborization but not pathway navigation.
Development 134, 2137-2146.
Gertler, F. B., Comer, A. R., Juang, J. L., Ahern, S. M., Clark, M. J., Liebl, E. C., and Hoffmann,
F. M. (1995). enabled, a dosage-sensitive suppressor of mutations in the Drosophila Abl
tyrosine kinase, encodes an Abl substrate with SH3 domain-binding properties. Genes Dev.
9, 521-533.
Gitai, Z., Yu, T., Lundquist, E., and Tessier-Lavigne, M. (2003). ScienceDirect - Neuron : The
Netrin Receptor UNC-40/DCC Stimulates Axon Attraction and Outgrowth through Enabled
81
and, in Parallel, Rac and UNC- 115/AbLIM. Neuron.
Huber, A., Kania, A., Tran, T., and Gu, C. (2005). ScienceDirect - Neuron: Distinct Roles for
Secreted Semaphorin Signaling in Spinal Motor Axon Guidance. Neuron.
Kim, G. J., Shatz, C. J., and McConnell, S. K. (1991). Morphology of pioneer and follower
growth cones in the developing cerebral cortex. J. Neurobiol. 22, 629-642.
Kwiatkowski, A. V., Rubinson, D. A., Dent, E. W., Edward van Veen, J., Leslie, J. D., Zhang, J.,
Mebane, L. M., Philippar, U., Pinheiro, E. M., Burds, A. A., et al. (2007). Ena/VASP Is
Required for neuritogenesis in the developing cortex. Neuron 56, 441-455.
Long, H., and Sabatier, C. (2004). ScienceDirect - Neuron : Conserved Roles for Slit and Robo
Proteins in Midline Commissural Axon Guidance. Neuron.
Lopresti, V., and Macagno, E. (1973). Structure and Development of Neuronal Connections in
Isogenic Organisms: Cellular Interactions in the Development of the Optic Lamina of
Daphnia. In Proceedings of the ....
Luria, V., Krawchuk, D., Jessell, T., and Laufer, E. (2008). ScienceDirect - Neuron:
Specification of Motor Axon Trajectory by Ephrin-B:EphB Signaling: Symmetrical Control
of Axonal Patterning in the Developing Limb. Neuron.
Ma, L., and Tessier-Lavigne, M. (2007). Dual branch-promoting and branch-repelling actions of
Slit/Robo signaling on peripheral and central branches of developing sensory axons. J
Neurosci 27, 6843-6851.
Menzies, A. S., Aszodi, A., Williams, S. E., Pfeifer, A., Wehman, A. M., Goh, K. L., Mason, C.
82
A., Fassler, R., and Gertler, F. B. (2004). Mena and vasodilator-stimulated phosphoprotein
are required for multiple actin-dependent processes that shape the vertebrate nervous system.
J Neurosci 24, 8029-8038.
Moon, M.-S., and Gomez, T. M. (2005). Adjacent pioneer commissural interneuron growth
cones switch from contact avoidance to axon fasciculation after midline crossing. Dev. Biol.
288, 474-486.
Nordlander, R. H. (1987). Axonal growth cones in the developing amphibian spinal cord. J.
Comp. Neurol. 263, 485-496.
Plump, A. S., Erskine, L., Sabatier, C., Brose, K., Epstein, C. J., Goodman, C. S., Mason, C. A.,
and Tessier-Lavigne, M. (2002). Slitl and Slit2 cooperate to prevent premature midline
crossing of retinal axons in the mouse visual system. Neuron 33, 219-232.
Raper, J., and Bastiani, M. (1983). Guidance of Neuronal Growth Cones: Selective Fasciculation
in the Grasshopper Embryo. Cold Spring Harbor ....
Schiff, R., and Rosenbluth, J. (1986). Ultrastructural localization of laminin in rat sensory
ganglia. J. Histochem. Cytochem. 34, 1691-1699.
Yu, T. W., Hao, J. C., Lim, W., Tessier-Lavigne, M., and Bargmann, C. I. (2002). Shared
receptors in axon guidance: SAX-3/Robo signals via UNC-34/Enabled and a Netrinindependent UNC-40/DCC function. Nat. Neurosci. 5, 1147-1154.
83
Chapter 3
Normal growth cone response to the chemorepellent, Slit requires Ena/VASP
induced reorganization of the cytoskeleton
84
3.1 A bstract..................................................................................................................................
86
3.2 Introduction...........................................................................................................................
86
3.3 R esults ....................................................................................................................................
87
3.3.1 mmvvee DRG explants display normal rates of axon outgrowth.................................
87
3.3.2 mmvvee DRG growth cones are defective in filopodia formation...............................
87
3.3.3 Slit application prom otes filopodial elongation...........................................................
93
3.3.4 m mvvee grow th cones display reduced sensitivity to Slit...........................................
98
3.3.5 The Slit receptor, Robo localizes preferentially in filopodia.........................................
103
3.3.6 Asymmetric filopodia elongation and collapse appear to correlate with repulsive turning
.................................................................................................................................................
10 6
3.4 D iscussion ............................................................................................................................
112
3.5 M ethods................................................................................................................................
117
3.6 References............................................................................................................................
118
85
3.1 Abstract
Mice genetically deficient for all three Ena/VASP family members (mmvvee mice)
display defects in axonal pathfinding in both the central and peripheral nervous system.
Pathfinding errors observed in axons derived from the dorsal root ganglia bear phenotypic
similarity to defects observed in Robol/2 and Slitl/2 knockout mice. These observations and
known genetic interactions between Ena/VASP and Slit/Robo in invertebrates suggest a
conserved role of Ena/VASP in repulsive guidance in vertebrates. Here I report that mmvvee
growth cones display an almost complete lack filopodia, and respond with reduced sensitivity to
application of the chemorepellent Slit. Furthermore, upon Slit stimulation, filopodia extend
dramatically in wild-type but not mmvvee growth cones. Intriguingly, it appears that filopodia
elongate preferentially in the direction of the chemorepellent. Together these observations
strongly argue for a role of filopodia upstream of Robo in sensitizing the growth cone to the
chemorepellent, Slit2.
3.2 Introduction
Filopodia are fingerlike cellular protrusions implicated in axon guidance in both
vertebrates and invertebrates, though their mechanistic involvement in these processes remains
unclear. In certain contexts, upon reaching decision points or after being exposed to guidance
cues, growth cones become more complex, elaborating numerous filopodia (Lebrand et al.,
2004)(Moon and Gomez, 2005). While many authors propose that filopodia act as cellular
antennae(Chien et al., 1993), this idea remains controversial, as others have observed loss of
filopodia without loss of proper pathfinding(Kim et al., 2002)(Dwivedy et al., 2007).
86
Ena/VASP proteins have been demonstrated to regulate filopodial dynamics and both
attractive and repulsive axonal pathfinding in vertebrates and invertebrates, though the
requirement for Ena/VASP induced filopodia formation in guidance has yet to be demonstrated
directly(Gertler et al., 1995)(Gitai et al., 2003)(Bashaw, 2000)(Lebrand et al., 2004)(Dent et al.,
2007)(Kwiatkowski et al., 2007).
3.3 Results
3.3.1 mmvvee DRG explants display normal rates of axon outgrowth
Ena/VASP deficient mice display sufficient axonogenesis in vivo to allow the observation
of pathfinding defects. I cultured dorsal root ganglia in vitro to test whether this axonal extension
occurs at normal rates in the absence of Ena/VASP. 48h after dissection, staining with the
neuronal specific marker,
p111-tubulin reveals that littermate explants display radial axon
outgrowth in 360 degrees (Fig. I a). Proximal to the neuronal cell bodies, littermate axons are
bundled together in fascicles (Fig. la, arrowhead). Over the same time period, mmvvee and
littermate explants form axons that reach the same length of approximately 1500um (Fig. la-c).
3.3.2 mmvvee DRG growth cones are defective in filopodia formation
As proper axonal pathfinding is believed to require distinct growth cone morphologies, I
next examined the structure of mmvvee growth cones in higher resolution. DRG explants shown
were cultured 48h then stained with markers for neuronal identity and morphology. Littermate
growth cones display multiple actin rich filopodia extending from the plasma membrane (Fig. 2
b,c, arrowhead). Each littermate growth cone has around 8 filopodia averaging 4um in length.
The manner in which mmvvee embryos are generated results in littermates always being
deficient in both VASP and EVL. I find no significant difference in either filopodia number or
87
length in comparison of littermate and Swiss-Webster (wild-type) DRG growth cones, indicating
that Mena is able to fully compensate for complete loss of VASP and EVL in filopodia formation
(Fig. 2ij).
Examination of Gap43 staining demonstrates that mmvvee growth cones are almost
completely devoid of filopodia (Fig. 2f). Ena/VASP deficient DRG growth cones display an
average of less than one filopodium. The rare "escaper filopodia" observed are significantly
shorter than those seen in littermate or wild-type growth cones (p<.001). Normal filopodia
display bundled actin filaments within the filopodium that extend into the peripheral region of
the growth cone (Fig. 2c, arrowhead). Escaper filopodia display no obvious F-actin filaments
within the structure or in the proximal growth cone periphery, indicating that these structures are
likely fundamentally different than physiologically normal filopodia (Fig. 2g, arrowhead).
88
Figure 1
A
MMvvee
mmvvee
1800
1600
1400 -
E
-c
1200
1000
0
800
V.
600
400
200 -
0-
MMvvee
mmvvee
89
Figure 1- mmvvee DRG explants display normal rates of axon outgrowth.
(A) 48h in culture results in similar explant radius in littermate and mmvvee DRG explants. (B)
Littermates and mmvvee explants show no significant difference in distance from the center of
DRG explant to growth cone (r) .
90
Figure 2
IB3-tubl
lActin
IGap43
E,
E
E
6
ns
5
I-
4
3
CL 2 0
EL
0
mmvvee
MMvvee
Swiss-Webster
12
10
0
8
0
6
0
4
0I...
--
2
Swiss-Webster
MMvvee
ns--
mmvvee
91
Swiss-Webster + 100nm
CytoD
IMere
Figure 2- mmvvee growth cones are almost completely devoid of filopodia.
(A-D) Littermate growth cones display multiple actin rich filopodia (arrowhead) (E-H) mmvvee
growth cones display an almost complete absence of filopodia. (I) Few filopodia present in
mmvvee growth cones are significantly shorter than in either littermate or wild-type growth
cones. (J) Wild-type and littermate growth cones have similar numbers of filopodia, both of
which are significantly higher than those observed in mmvvee growth cones or wild-type growth
cones treated with the actin polymerization inhibitor, cytochalasin D.
92
3.3.3 Slit application promotes filopodial elongation
The similar defects in axonal pathfinding observed in mmvvee and Slit/Robo mutants
lead me to hypothesize that mmvvee growth cones would be unable to respond to Slit properly.
The reported effect of Slit application on DRG growth cones is a loss of lamellipodial veils
leading to growth cone collapse(Ma and Tessier-Lavigne, 2007). To capture the dynamics of
growth cone collapse, DRG explants were imaged by time-lapse microscopy as they were treated
with Slit. I confirmed in these preparations that bath application of either wild-type or mmvvee
littermate control cultures with Slit leads to growth cone collapse (Fig. 3f-h). Within 10 minutes
of Slit application, lamellipodial veils have disappeared and growth cones have halted their
forward translocation.
Concomitant with growth cone collapse, I observed a novel effect of Slit addition more
readily examined in higher temporal resolution. Upon stimulation, growth cone filopodia rapidly
elongated dramatically (Fig. 3 e,f arrowheads). Examination of images taken every minute
reveals the highly dynamic nature of Slit induced filopodial elongation (Fig. 4 a-h). Within 5
minutes of Slit addition average filopodial length has more than doubled (Fig. 4 i). These images
further reveal that during the initial stages of collapse, filopodia begin to extend (Fig. 4a) while
lamellipodia are still present (Fig. 4b), allowing for the possibility that this filopodial extension is
somehow related to collapse
93
After Slit addition!
Before Slit addition
*CD
-I
c~E
'1
Figure 3- Slit application promotes filopodial extension in DRG growth cones.
(A-D) Growth cones display highly dynamic changes in morphogenesis before slit addition.
Filopodia (arrowheads) and lamellipodia (arrows) are continually remodeled. (E-H) Application
of Slit (900ng/ml) causes transient dramatic filopodial extension (arrowheads). Images shown
were captured at three minute intervals.
95
After Slit addition
U.'
lCD
'1
(~.
Figure 4
w
-
-
20 T ~ ~~
-
-
~ -
--
-
~~ --- ~ ~~
- ~-
18
E
16
T
ca
14 t--
12
E
_0
il0
0
10
-
-
T
t
-
j
"LM
ammvvee
-
8
2
!I
--
A
~~
~
~
0 4--0
W
-
--- ~-~
6
-
IN
--
- --
11
1
-
-
~~
S
W
~--~
24
810
12
Minutes after Slit addition
Figure 4- Filopodia extension is highly labile.
(A-H) Filopodia begin extending almost immediately upon application of Slit. (I) Maximal
filopodial extension occurs an average of four minutes after stimulation. mmvvee growth cones
do not extend filopodia after stimulation. Images shown were captured at one minute intervals
97
3.3.4 mmvvee growth cones display reduced sensitivity to Slit
Imaging of mmvvee derived growth cones revealed that in the absence of Ena/VASP, growth
cones are highly dynamic, ruffling and continually reorganizing lamellipodia (Fig. 5a-d). Upon
Slit stimulation, mmvvee growth cones often fail to collapse, continuing their forward
translocation (Fig. 5e-h). mmvvee growth cones do appear to sense the presence of Slit, as some
edges of their lamellipodia adopt a more scalloped morphology than is seen prior to collapse
(Fig. 5 e,f arrowheads). While stimulation of mmvvee derived growth cones with levels of Slit
capable of causing modest increases in collapse in littermates (750ng/mi) failed to induce
collapse above background levels, increasing the concentration of Slit applied led to a
concentration dependent reappearance of growth cone collapse (Fig. 6). At each concentration of
Slit applied, collapse was significantly reduced with respect to littermates (p<.05 or .01). This
partial reappearance of collapse indicates that the role of Ena/VASP proteins in Slit response is
in the sensitivity to the ligand rather than in collapse per se. To further confirm that filopodia are
required for proper sensitivity to Slit, I applied low levels (1OOnM) of the actin polymerization
inhibitor, Cytochalasin D (Mousa et al., 1978) to Swiss-Webster DRGs. At these levels,
Cytochalasin D caps actin filaments without showing appreciable severing activity (Urbanik and
Ware, 1989). Within one hour of application of Cytochalasin D, nearly all filopodia disappear
(Fig. 7 a-g) (Fig. 2j). These drug treated growth cones show a decrease in sensitivity to
application of Slit quantitatively similar to mmvvee growth cones (Fig. 7i).
98
.-
I
.. ..
....
...
.... ........................
After Slit addition
Before Slit addition
.
(31
CD
U.'
'1
c~u
Figure 6
-
-~~
75.0 -r
70.0 t
-
-
- - -
~ -
-
65.0 t
(1)
60.0
CL)
55.0
0
~
-
7
me6Littermate
50.0 +-
ammvvee
45.0
40.0
-
35.0 f
30.0
2 5 .0
500
-.600
700
, - ~ - - - - , - --.....
800
900
1000
- -, 1100
--.....- - - - -,-1200
1300
Slit Concentration (ng/ml)
Figure 5 - mmvvee growth cones display reduced sensitivity to application of Slit.
(A-H) mmvvee growth cones often fail to collapse upon stimulation with 900ng/ml Slit.
Figure 6 - mmvvee growth cones collapse at higher concentrations of Slit
Increasing the amount of Slit applied to either littermate or mmvvee growth cones results in a
concentration dependent increase in collapse. Littermate growth cones collapse at significantly
higher rates than mmvvee growth cones at each concentration of Slit tested. Images shown were
captured at three minute intervals
100
Figure 7
j 7bUng/ml Slit stimulation
75
~.~.-~ .
- . ...-.
- -. ~..... ~ -
~...
65
.
-- ~
- ~ ~~-
70
60
-
-..... .......
-
-
55
50
L)
45
-
-
.. -~
-
~ --- -
-~
40
U
2
.
35
..-
-
~
-
~-
30~
Swiss-Webster
101
Swiss-Webster + 100nm CD
Figure 7 - Treatment of wild-type growth cones with Cytochalasin D results in a loss of
filopodia and loss of sensitivity to Slit application.
(A-H) At the time of application, wild type growth cones display multiple filopodia. Filopodia
are gradually lost over the course of 30 minutes. Images shown were captured at five minute
intervals. (I) Cytochalasin D treated growth cones show a significant reduction in response to
application of 750ng/ml Slit similar to what is observed in mmvvee growth cones.
102
3.3.5 The Slit receptor, Robo localizes preferentially in filopodia
Given that filopodia are required for proper sensitivity to Slit, I hypothesized that its
receptor, Robo would display preferential filopodial localization. Immunoreactivity of an antiRobol antibody confirms a high level of Robol within filopodia and at their tips (Fig. 8b,
arrows). To confirm that the receptor is surface exposed and therefore able to bind its ligand, I
treated growth cones with Fluorescently labeled Slit. Combination DIC/fluorescence live
microscopy reveals that Slit binds to filopodia and is subsequently trafficked toward the body of
the growth cone (Fig. 8d, arrows).
103
Figure 8
IF-Actin|
IROBO-1|I
t=1s
t=
t=1 5s
t=2
IMergel
Figure 8 - Filopodia are able to bind Slit.
(A-C) Wild-Type growth cone filopodia show immunoreactivity to a Robol antibody. (D)
Fluorescently conjugated Slit binds filopodia indicating Robo is surface exposed in growth
cones.
105
3.3.6 Asymmetric filopodia elongation and collapse appear to correlate with repulsive
turning
It is tempting to speculate that the filopodial elongation seen upon bath application of Slit
is a mechanism for increasing either the sensitivity of the growth cone to this chemorepellent, or
a way for the growth cone to sample a larger environment before turning. Answering this
question requires live observation of asymmetric collapse and growth cone turning. To observe
asymmetric collapse, I added Slit coated polystyrene beads to DRG explant preparations isolated
from Swiss-Webster e 12.5 embryos. Upon contact with Slit coated beads, growth cones appear
to change course in avoidance of the chemorepellent (Fig. 9 a-h). Interestingly, after the turn
occurs, the axon itself does not remain static. During and soon after the turning event, a kink in
the axon is clearly visible (Fig. 9 e,f arrowheads), but the axon is quickly smoothened and adopts
an arc shaped morphology by the end of the time course (Fig. 9h). Examination of this turning
event at higher temporal resolution (frames taken at 30 second intervals) reveals that upon
contact with Slit coated beads, the growth cone elongates its filopodia preferentially on the side
of the chemorepellent (Fig. 9c-f, arrowheads). Intriguingly, concomitant with this extension a
new lamellipodium appears, orthogonal to the angle of filopodial extension (Fig. lOc-h, arrows.
Fig. 11 a-h Morphology shown more clearly in tracings). This new lamellipodial protrusion
becomes the main body of the growth cone as it continues along it's new trajectory (Fig. 9 e-h)
106
Figure 9
107
Figure 9 - Growth cones appear to avoid Slit coated beads.
(A-B) Growth cone shown initially adopts a straight trajectory. (C-D) Upon contact with Slit
coated beads, growth cone begins to turn. (E-F) After turning event, clear kink is present in axon
(arrowhead) (G-H) After turn, kink is smoothened, and axon adopts an arc shape. Images shown
were captured at five minute intervals.
108
Figure 10
r A K7.Mw
109
Figure 11
U
3
3
I
110
Figure 10 - Filopodia appear to extend asymmetrically upon contact with Slit coated beads.
(A-B) Growth cones encounter Slit coated beads (C-F) Filopodium extends in the direction of
bead contact (arrowhead); simultaneously, lamellipodia extend in direction away from contact
(arrows). (G-H) Newly elaborated lamellipodium persists leading to turning (arrows)
Figure 11 - Filopodia appear to extend asymmetrically upon contact with Slit coated beads.
(A-H) Growth cone from figure ten traced to more clearly show changes in morphology upon
contact with Slit coated beads (red dots)
111
3.4 Discussion
The observation that DRG explants derived from mmvvee animals have numerous axons
that extend at a similar rate as their littermate controls appears at first to be in conflict with our
previous studies demonstrating a complete lack of cortical axon formation in mmvvee animals.
Our previous studies (Kwiatkowski et al., 2007)(Dent et al., 2007)(Gupton and Gertler,
2010)argue for a model in which the laminin rich environment of the dorsal root ganglia
promotes sufficient filopodia fonnation to bypass the defects observed in cortical neurons.
While numerous studies have found Ena/VASP family members to be regulators of
repulsive axon guidance responses in vivo, the mechanistic basis by which Ena/VASP is
affecting pathfinding has not previously been investigated (Bashaw, 2000)(Yu et al., 2002).
Vertebrate studies have examined the effect of Ena/VASP inactivation on growth cone
morphology and found defects in steady state and Netrin induced filopodia formation, but have
not demonstrated a failure of these growth cones to be guided properly in vivo(Lebrand et al.,
2004). Other vertebrate studies have demonstrated phenotypes in cortical and hippocampal
commissure formation, as well as pathfinding in the optic chiasm in animals deficient in Mena or
Mena and VASP but have not examined the mechanism underlying these defects (Menzies et al.,
2004)(Lanier et al., 1999).
It has long been unclear how a family of proteins that promote actin polymerization
might be mechanistically involved in the process of repulsive guidance. Growth cone collapse
has not previously been shown to require actin polymerization, and therefore I initially
hypothesized that Ena/VASP proteins may only be required for induction of actin polymerization
on the face of the growth cone opposite the repulsive cue. Alternatively, upon Slit stimulation,
112
Ena/VASP proteins might be polymerizing long unbranched actin filaments that cannot exert
sufficient force on the plasma membrane to cause stable lamellipodial protrusion. This would
lead to growth cone stalling in a mechanism analogous to how Ena/VASP proteins have been
observed to negatively regulate fibroblast motility(Bear et al., 2002).
Ena/VASP deficient growth cones display a profound defect in filopodia formation.
Interestingly, in the absence of Ena/VASP, application of Slit at a concentration that collapses
littermate growth cones has very little effect. These defects are recapitulated in wild-type
neurons in which filopodia formation is impaired by the application of low levels of
Cytochalasin D. While we don't observe any defects in embryonic progression, this observation
argues that there is no subtle delay in axonal maturation underlying our observed reduction in
Slit responsiveness.
The observation that application of increased concentrations of Slit leads to a
reappearance of collapse in mmvvee growth cones argues that the cellular machinery underlying
collapse is intact. This indicates a defect in sensitivity to Slit rather than a defect in collapse per
se, though I cannot yet formally rule out the possibility that multiple independent pathways are
involved in growth cone collapse and loss of Ena/VASP merely results in a loss of one of these
parallel pathways. Together, these data argue that contrary to my initial hypotheses, filopodia
formation mediated by Ena/VASP induced actin polymerization is required to prime the growth
cone for sensitivity to Slit. This model of sensitivity is further supported by my observation that
Robo localizes preferentially within filopodia, and that Slit protein can bind directly to filopodia.
During my examination of the collapse of wild-type growth cones, I discovered a striking
new effect of Slit addition. Upon stimulation, filopodia present in the growth cone extend
dramatically concomitant with collapse. This extension is highly labile, beginning almost
113
immediately upon Slit application, and often persisting for less than five minutes. Previous
studies of Slit induced collapse have not examined growth cone dynamics in higher resolution
than five-minute intervals, and thus it is likely that previous authors were not able to observe the
striking nature of this effect.
Filopodia appear to extend preferentially on the side of the growth cone that is exposed to
a chemorepellent. I have observed that upon contact with Slit coated polystyrene beads, there is a
filopodial elaboration in the direction of the bead, concomitant with a loss of lamellipodial veils
restricted to the area proximal to the extending filopodia. Asymmetric collapse has been
demonstrated in growth cones encountering Semaphorin coated beads, and is believed to be
involved in repulsive turning(Fan and Raper, 1995). I observed that this asymmetric collapse is
followed by lamellipodial protrusion on the side opposite this extension, leading to a turning
event. If filopodia indeed are involved in sensitivity of the growth cone to Slit, elaborating these
filopodia prior to making a stable turn may represent a way for the growth cone to either increase
sensitivity or sample a larger area. Opposed to surface bound growth cone turning techniques
such as stripe assays, this technique rules out the possibility that bound ligand obscures adhesive
substrate components. Therefore using this technique, we can state unequivocally that turning
events observed are not a result of simple inability of the growth cone to adhere on one side.
Together these data argue for a role of Ena/VASP induced filopodia formation in
sensitizing the growth cone to Slit upstream of Robo. They do not, however, argue against an
additional role of Ena/VASP proteins downstream of Robo in inducing lamellipodial and
filopodial extension on the face of the growth cone opposite a repulsive cue. Based on these data,
I propose the following "four-step" model of growth cone repulsion by Slit/Robo signaling (Fig.
12): Step 1 - Scanning. Loss of Ena/VASP induced filopodia formation results in a significant
114
reduction in growth cone sensitivity to the presence of the ligand Slit, indicating that as has been
previously hypothesized, filopodia are acting in a sensory fashion upstream of Robo. Step 2 Reinforcing. Upon ligand binding, Ena/VASP induced filopodia extend in the direction of the
repellent to increase sensitivity or sampling area, reinforcing the initial recognition. Step 3 Asymmetric collapse. Lamellipodial veils are lost on the side of the growth cone proximal to the
chemorepellent, likely in an Ena/VASP independent fashion. Step 4 - Protrusion. Lamellipodia
are induced on the face of the growth cone opposite the repulsive cue. It remains unclear whether
this step involves Ena/VASP induced actin polymerization. It will be interesting in the future to
see if Ena/VASP proteins or asymmetric filopodial extension are general features of repulsive
guidance, or are specific to Slit/Robo mediated repulsion.
115
Figure 12
Four Step Model of Repellent Avoidance
Step I - Scanning
EnaNASP mediated filopodia
scan the growth cone microenvironment
Step 2 - Reinforcing
EnaNASP mediated Filopodia
extend in the direction of the
signal, potentiating sensitivity.
Step 3 - Asymmetric collapse
Lamellipodial veils are lost proximal to the chemorepellent in an
EnaNASP independent fashion
Step 4 - Protrusion
Lamellipodia elaborate distal to
site of chemorepellent encounter. EnaNASP dependent?
116
3.5 Methods
Whole dorsal root ganglia were isolated from e12.5 Swiss-Webster, mmvvee, or
littermate mice using Dumont #5 forceps. DRG explants were cultured in Mattek glass bottom
dishes coated with 500ug/ml Poly-D-Lysine (Sigma) and 50ug/ml Laminin (Southern Biotech).
DRGs were cultured for 8-16hours in Neurobasal medium supplemented with B27, NGF, and
5% fetal bovine serum (Hyclone), then switched to identical medium without serum. For analysis
of outgrowth and fasciculation, explants were allowed to develop in culture for 48h before
fixation in 4% paraformaldehyde+.05% glutaraldehyde. Explants were immunostained using
standard ABC protocols. Primary antibody incubations were done 2hr at room temperature in
TBS + 1%BSA, + .025% triton-x100 using an antibody recognizing pIII-tubulin at a dilution of
1:5000 (Mouse monoclonal, Promega 5G8). Individual fluorescence images were taken using a
Nikon Ti wide-field microscope and 20X DICN2 objective, then stitched to form one cohesive
image of the entire ganglion using the NIS Elements imaging/analysis software. Explant
outgrowth was quantitated using unstained DRGs by taking four evenly spaced measurements of
maximal distance from ganglion center to axon tip.
Growth cone immunohistochemistry was done using similar fixation and staining
conditions but with the following dilutions: Robo1 1:500 (Rabbit polyclonal, Abcam ab7279),
p111-tubulin
1:5000 (Mouse monoclonal, Promega 5G8), Alexa-594 conjugated phalloidin
1:1 00(Molecular probes). Growth cone imaging was done using a deltavision deconvolution
microscope with a 1OOX DIC objective. For live imaging of growth cones, images were taken on
a Nikon Tslkdfjs equipped with PerfectFocus, using either a 1OOX DICN2 oil immersion or 40X
DICN2 water immersion objective.
117
Purification of Slit protein was done by transient transfection of a COS- 1 cell line with an
expression construct encoding the amino-terminal fragment of Slit-2 protein and subsequent 2448h incubation in opti-mem reduced serum medium. Protease inhibitors were added to combined
high salt washes and cultured media, then concentrated with Amicon-ultrao centrifugal filter
units. Slit was purified from concentrates by affinity chromatography to wheat germ agglutinin.
Growth cone collapse was quantified by fifteen minute bath applications of Slit
performed by removal of 1ml of dish volume and immediately replacing with iml serum free
medium + ligand or vehicle. Dishes were fixed, then imaged using phase microscopy. "Collapse"
was counted as a complete or nearly complete absence of lamellipodial veils. Filopodia number
and length were quantified using Nikon NIS image analysis software. Slit-2N was covalently
labeled with a fluorophore using DyLight Microscale o 594 Antibody Labeling kit.
lum Opti-bindo polystyrene beads were purchased from Thermo-fisher scientific. .5ul of
beads were added directly to 20ul of Slit-2N protein solution at a concentration of 50ng/ul. This
mixture was incubated for one hour at 37c in a tissue culture incubator, blocked for one hour
with 1mg/ml BSA, then diluted into 500ul of Neurobasal+B27 before direct addition of 2ul
diluted beads to DRG explant cultures.
3.6 References
Bashaw, G. J. (2000). Repulsive Axon GuidanceAbelson and Enabled Play Opposing Roles
Downstream of the Roundabout Receptor. Cell 101, 703-715.
Bear, J.E., Svitkina, T. M., Krause, M., Schafer, D. A., Loureiro,
118
J.J., Strasser, G.A.,
Maly, I.V.,
Chaga, 0. Y., Cooper, J.A., Borisy, G.G., et al. (2002). Antagonism between Ena/VASP
proteins and actin filament capping regulates fibroblast motility. Cell 109, 509-521.
Chien, C.B., Rosenthal, D. E., Harris, W. A., and Holt, C. E. (1993). Navigational errors made
by growth cones without filopodia in the embryonic Xenopus brain. Neuron 11, 237251.
Dent, E.W., Kwiatkowski, A. V., Mebane, L. M., Philippar, U., Barzik, M., Rubinson, D. A.,
Gupton, S., van Veen, J.E., Furman, C., Zhang,
J., et al. (2007).
Filopodia are required for
cortical neurite initiation. Nat. Cell Biol. 9, 1347-1359.
Dwivedy, A., Gertler, F. B., Miller, J., Holt, C.E., and Lebrand, C.(2007). Ena/VASP function in
retinal axons is required for terminal arborization but not pathway navigation.
Development 134, 2137-2146.
Fan, J., and Raper, J.A. (1995). Localized collapsing cues can steer growth cones without
inducing their full collapse. Neuron 14, 263-274.
Gertler, F. B., Comer, A. R., Juang, J.L., Ahern, S. M., Clark, M. J., Liebl, E. C., and Hoffmann, F.
M. (1995). enabled, a dosage-sensitive suppressor of mutations in the Drosophila Abl
tyrosine kinase, encodes an Abl substrate with SH3 domain-binding properties. Genes
Dev. 9, 521-533.
Gitai, Z., Yu, T., Lundquist, E., and Tessier-Lavigne, M. (2003). ScienceDirect - Neuron: The
Netrin Receptor UNC-40/DCC Stimulates Axon Attraction and Outgrowth through
Enabled and, in Parallel, Rac and UNC-115/AbLIM. Neuron.
119
Gupton, S. L., and Gertler, F. B. (2010). Integrin signaling switches the cytoskeletal and
exocytic machinery that drives neuritogenesis. Dev. Cell 18, 725-736.
Kim, M. D., Kolodziej, P., and Chiba, A. (2002). Growth cone pathfinding and filopodial
dynamics are mediated separately by Cdc42 activation. J Neurosci 22, 1794-1806.
Kwiatkowski, A. V., Rubinson, D.A., Dent, E. W., Edward van Veen, J., Leslie, J. D., Zhang, J.,
Mebane, L. M., Philippar, U., Pinheiro, E. M., Burds, A. A., et al. (2007). Ena/VASP Is
Required for neuritogenesis in the developing cortex. Neuron 56, 441-455.
Lanier, L. M., Gates, M. A., Witke, W., Menzies, A. S., Wehman, A. M., Macklis,
J.D.,
Kwiatkowski, D., Soriano, P., and Gertler, F. B. (1999). Mena Is Required for Neurulation
and Commissure Formation. Neuron 22, 313-325.
Lebrand, C., Dent, E.W., Strasser, G.A., Lanier, L. M., Krause, M., Svitkina, T. M., Borisy, G.G.,
and Gertler, F. B. (2004). Critical role of Ena/VASP proteins for filopodia formation in
neurons and in function downstream of netrin-1. Neuron 42, 37-49.
Ma, L., and Tessier-Lavigne, M. (2007). Dual branch-promoting and branch-repelling
actions of Slit/Robo signaling on peripheral and central branches of developing sensory
axons. J Neurosci 27, 6843-6851.
Menzies, A. S., Aszodi, A., Williams, S. E., Pfeifer, A., Wehman, A. M., Goh, K. L., Mason, C.A.,
Fassler, R., and Gertler, F. B. (2004). Mena and vasodilator-stimulated phosphoprotein
are required for multiple actin-dependent processes that shape the vertebrate nervous
system. J Neurosci 24, 8029-8038.
120
Moon, M.-S., and Gomez, T. M. (2005). Adjacent pioneer commissural interneuron growth
cones switch from contact avoidance to axon fasciculation after midline crossing. Dev.
Biol. 288,474-486.
Mousa, G.Y., Trevithick, J.R., Bechberger,
J., and
Blair, D. G. (1978). Cytochalasin D induces
the capping of both leukaemia viral proteins and actin in infected cells. Nature 274,
808-809.
Urbanik, E., and Ware, B. R. (1989). Actin filament capping and cleaving activity of
cytochalasins B, D, E, and H.Arch. Biochem. Biophys. 269, 181-187.
Yu, T. W., Hao, J. C., Lim, W., Tessier-Lavigne, M., and Bargmann, C.I. (2002). Shared
receptors in axon guidance: SAX-3/Robo signals via UNC-34/Enabled and a Netrinindependent UNC-40/DCC function. Nat. Neurosci. 5, 1147-1154.
121
Chapter 4
EVH1 domains of vertebrate Ena/VASP proteins interact directly with
multiple sites in the cytoplasmic tail of Robo in a phospho regulated manner
122
4.1 A bstract................................................................................................................................
124
4.2 Introduction.........................................................................................................................
124
4.3 Results ..................................................................................................................................
128
4.3.1 EV I binds CC2 and CC3 directly ...............................................................................
128
4.3.2 EVHI binds CCl ...........................................................................................................
131
4.3.3 EV H 1 binding of CC3 m ay be phospho-dependent ......................................................
134
4.4 Discussion ............................................................................................................................
135
4.5 M ethods................................................................................................................................
138
4.5 References............................................................................................................................
138
123
4.1 Abstract
In invertebrates it has been shown that Ena/VASP binds directly to the cytoplasmic tail of
Robo. The specificity and regulation of this interaction, however, have yet to be determined.
Here I describe the mapping of three Ena/VASP binding sites in the cytoplasmic tail of
vertebrate Robo. Two of these sites are canonical EVH1 ligands, and the third is a novel EVH1
binding motif. Phosphorylation has been shown to take place within or near two of the three
sites, and I have found that these phosphorylation events dramatically affect the strength of the
EVH1:Robo interaction. This may provide a mechanism by which context dictates the affinity of
the Ena/VASP:Robo interaction.
4.2 Introduction
Ena/VASP Homology Domain 1 (EVH 1) is the most amino terminal domain within
Ena/VASP and is conserved in all known Ena/VASP homologs. The best characterized role of
EVH 1 is in mediating protein:protein interactions through specific interactions with a
characterized consensus sequence, (D/E)(F/L)PXCDPX(D/E)(D/E), commonly abbreviated "FP4"
(Niebuhr et al., 1997). FP4 motifs are found in proteins with diverse functions, with an
enrichment for proteins involved in cellular motility (Brindle et al., 1996)(Drees et al.,
2000)(Krause et al., 2004)(Lafuente et al., 2004). Aside from binding to FP4 motifs, EVH1
domains found in Ena/VASP have been described to bind to only one other motif. The sequence
HN(X)4 2 M(X) 10 M found in the focal adhesion protein, Tes, has been demonstrated to bind
specifically to the EVH1 domain of Mena (Bodda et al., 2007).
The cytoplasmic tails of all studied Robo receptors are devoid of any demonstrated
domains with catalytic activity of their own. Instead, upon activation Robo binds to conserved
downstream effector proteins. The cytoplasmic tail of vertebrate Robo 1 contains four described
124
conserved cytoplasmic motifs (Figure 1a), three of which share a high level of conservation from
invertebrates to vertebrates (Figure lb-d). Conserved cytoplasmic motif #1 (CCl) is found in
vertebrate Robo 1 and Robo2 and D. melanogasterRobo 1 and Robo2, but absent in vertebrate
Robo3 and Robo4. Within CCl is a conserved tyrosine that is a demonstrated substrate for the
Ableson non-receptor tyrosine kinase in flies. Phosphorylation of CCl has been shown to be
important for axon guidance in flies; flies expressing a version of Robo with the tyrosine mutated
to the non-phosphorylatable tyrosine mimetic amino acid, phenylalanine, display a dominant
phenotype in which the midline is completely devoid of axon fibers(Bashaw, 2000). This
mutation has been interpreted as causing hyperactivity of Robo, which would indicate that
Ableson is a negative regulator of Robo in drosophila. CCl has also been shown to directly
associate with the attractive guidance receptor, DCC, and participate in silencing of
attraction(Stein and Tessier-Lavigne, 2001).
The second conserved cytoplasmic motif, CC2, is present in all four vertebrate Robo
homologs, as well as drosophila Robo 1. CC2 of vertebrate Robo 1/2/3 and D. melanogaster
Robo 1 contain the potential EVH 1 binding motif, LPPPP. CC2 has been shown to be crucial for
Robo activity, as the adaptor protein Dock/Nck binds to D. melanogasterRobo CC2 and recruits
the Rac GEF, Son of Sevenless (Sos)(Yang and Bashaw, 2006). CC2 has also been shown to
directly bind the Rac GAP Vilse(Lundstr6m et al., 2004), which is thought to play a role in
signaling downstream of Robo(Hu et al., 2005). Both Vilse and Nck are believed to interact with
CC2 through Src homology 3 (SH3) domains, which bind to proline rich sequences with the
consensus motif PxxP. The third conserved cytoplasmic motif of Robo, CC3, is present in
vertebrate Robo 1/2/3 and contains an Ena/VASP binding motif. Interestingly, while CC3 in the
vertebrate Robos contains the EVHI consensus binding motif, the D. melanogasterCC3
125
Figure 1
Ig domains
Fnlil domains
O
TM domain
'CC1I
M.musRobol
M.mus_Robo2
D.melRobol
D.melRobo2
DGRFVNPSGCPTPYATTQLIANLSNNMNN
TYNSSSQITCATPYATTQIL SNSIHELAV
FYNCRKSPD1PTPYATTMII TSSSETCTK
SEYGRHGNAPAPYATSSI+ PHQQQQQQQ
M.musRobol
M.mus_Robo2
M.mus_Robo3
M.musRobo4
D.melRobol
GGMNWDLLPPPPAHPP
GNFGRGDVLPPVPGQGD
PSLNW EALPPPPPSCEL
LQAPSDPLPAAPLSVLN
EPINW EFLPPPPEHPPP
M.musRobol
M.mus_Robo2
M.musRobo3
D.melRobol
EAYDDLPPPPVPPP IKSPTVQSKA
EGMDDLPPPPDPPPGQGLRQQIGLS
ENS FGDLPPPPLPPP EEASWALELR
ATPEPMQPPPPVPVP GWYQPVHPNS
126
SNSEEYN
ATMLSDG
CLEGPEE
SRPSSPQ
STYGYAQ
0
0
Figure 1 - Sequence conservation in the cytoplasmic tail of Robo.
(A) Canonical structure of Robo showing location of conserved cytoplasmic motifs (B)
Conserved cytoplasmic motif #1. Multiple alignment showing region of highest conservation
(box)(C) Multiple alignment of conserved cytoplasmic motif #2 demonstrates conservation of
EVH 1 consensus binding LPPPP motif (D) Multiple alignment of conserved cytoplasmic motif
#3 showing conservation of EVH 1 consensus binding LPPPP motifs in vertebrate but not
invertebrate Robo. Multiple alignments performed using ClustalW algorithm.
127
lacks the N-terminal hydrophobic (F/L) residue crucial for binding of EVH 1. CC3 has also been
demonstrated to bind the Slit-Robo Gap (SRGAP) family of Rho GTPase GAPs, which may
have roles in axon guidance downstream of Robo.
4.3 Results
4.3.1 EVH1 binds CC2 and CC3 directly
To determine whether vertebrate Ena/VASP proteins bind to the conserved motifs in
Robo, I overlaid peptide SPOTs arrays with purified GST:EVH1 fusion protein. Peptides 25
amino acids in length were arrayed such that each sequential spot overlapped 20 of the 25
residues in the previous spot (Window:25, Step:5). This technique was used to scan the
cytoplasmic tail of mouse Robol for interaction sites. GST-EVH1 overlay of these membranes
reveals that only peptides which contain the entire (D/E)(F/L)PXCDP consensus motif show
appreciable EVH1 binding (Fig. 2a, 3a). Indicating specificity of this interaction, loss of the
fourth proline, a residue shown to be crucial for EVH1 binding(Ball et al., 2000) confers a loss of
EVH1 binding(Fig. 2a, row 3). Similarly, loss of the N-terminal acidic residue also shown to be
important for EVHl binding affinity greatly reduces binding (Fig. 3a, row 7). Further confirming
the specificity of this interaction, the same sequences with alanine substitutions in the internal
hydrophobic residues in the case of CC2 or the acidic and hydrophobic residues in the case of
CC3 show no appreciable binding to purified EVH1 (Fig. 2b,3b) indicating that EVH1 is binding
to both CC2 and CC3 in a specific fashion.
128
Figure 2
ROBO CC2
RGSSTSGSQGHKKGARTPKAPKQGG
SGSQGHKKGARTPKAPKQGGMNWAD
HKKGARTPKAPKQGGMNWADLLPPP
RTPKAPKQGGMNWADLLPPPPAHPP
PKQGGMNWADLLPPPPAHPPPHSNS
MNWADLLPPPPAHPPPHSNSEEYNM
LLPPPPAHPPPHSNSEEYNMSVDES
PAHPPPHSNSEEYNMSVDESYDQEM
ROBO CC2
RGSSTSGSQGHKKGARTPKAPKQGG
SGSQGHKKGARTPKAPKQGGMNWAD
HKKGARTPKAPKQGGMNWADAAPPP
RTPKAPKQGGMNWADAAPPPPAHPP
PKQGGMNWADAAPPPPAHPPPHSNS
MWADAAPPPPAHPPPHSNSEEYNM
AAPPPPAHPPPHSNSEEYNMSVDES
PAHPPPHSNSEEYNMSVDESYDQEM
Figure 2 - Specific binding of EVH1 to CC2.
(A) Peptide sequences walking through CC2 of mouse Robol shown on sequential lines.
Consensus EVH1 binding motif residues displayed in bold are required for specific binding. (B)
Core hydrophobic residues mutated to alanine shown in red ablate binding.
129
Figure 3
ROBO CC3
SAVVIQKARPAKKQKHQPGHLRREA
QKARPAKKQKHQPGHLRREAYADDL
AKKQKHQPGHLRREAYADDLPPPPV
HQPGHLRREAYADDLPPPPVPPPAI
LRREAYADDLPPPPVPPPAIKSPTV
YADDLPPPPVPPPAIKSPTVQSKAQ
PPPPVPPPAIKSPTVQSKAQLEVRP
PPPAIKSPTVQSKAQLEVRPVMVPK
ROBO CC3
SAVVIQKARPAKKQKHQPGHLRREA
QKARPAKKQKHQPGHLRREAYADDL
KKQKHQPGHLRREAYADAAPPPPV
HQPGHLRREAYADAAPPPPVPPPAI
LRREAYADAAPPPPVPPPAIKSPTV
YADAAPPPPVPPPAIKSPTVQSKAQ
PPPPVPPPAIKSPTVQSKAQLEVRP
PPPAIKSPTVQSKAQLEVRPVMVPK
Figure 3 - Specific binding of EVH1 to CC3.
(A) Peptide sequences walking through CC3 of mouse Robol shown on sequential lines.
Consensus EVH1 binding motif residues displayed in bold are required for specific binding. (B)
Core hydrophobic and acidic residues mutated to alanine shown in red ablate binding.
130
4.3.2 EVH1 binds CC1
While CCl contains no canonical EVH 1 binding motifs, based on the report that deletion
of CC 1 confers loss of Ena binding in D. melanogaster,I next used a peptide array to probe the
specificity of this reported interaction. EVH 1 binds to a stretch of amino acids spanning the most
highly conserved region of CC 1 (Fig. 4, rows 5-7 bold) with the sequence
"PYATTQLIQANLSNN." Interestingly, the sequence comprising the region of highest
conservation (Fig. lb, box) is insufficient for EVHI binding (Fig. 4, row 4). Although the
sequence conservation is less strict, the sequence immediately C-terminal CCI in vertebrate
Robol/2 as well as D. melanogaster Robo contains many polar uncharged residues, which may
additionally be important for binding.
131
Figure 4
IA|
ROBO CC1I
DLSNKINEMKTFNSPNLKDGRFVNP
I NEMKTFNSPNLKDGRFVNPSGQPT
TFNSPNLKDGRFVNPSGQPTPYATT
NLKDGRFVNPSGQPTPYATTQL IQA
RFVNPSGQPTPYATTQLIQANLSNN
SGQPTPYATTQL IQANLSNNMNNGA
PYATTQL IQANLSNNMNNGAGDSSE
QLIQANLSNNMNNGAGDSSEKHWKP
Figure 4 - Specific binding of EVH1 to CC1.
(A) Peptide sequences walking through CCl of mouse Robol shown on sequential lines.
Interaction motif mapped to a 15 amino acid peptide shown in bold.
132
4.3.3 EVH1 binding of CC3 may be phospho-dependent
A recent high throughput study of cell division in HeLa cells sought to find new proteins
regulated by serine and threonine phosphorylation(Dephoure et al., 2008). Mass spec analysis in
this screen revealed new sites within the cytoplasmic tail of human Robo 1 that can be
phosphorylated. Two of these exist in close proximity to the consensus EVH 1 binding site found
in CC3. Phosphorylation of the serine, threonine, or both residues (Figure 6, residues colored
red) found in this screen ablates the EVH1:CC3 interaction (Figure 6). Estimation of the relative
change in binding in these data is impossible as interaction with the phospho-peptides in not
detectable. While this represents the possibility of a novel means of regulation of EVH 1 binding
to it's canonical ligands, phosphorylation at these residues in neuronal tissue has yet to be
demonstrated and therefore remains preliminary.
133
Figure 6
E-I
CA E-I
4 0
WA
O
O
ROBO CC3
4
W
Q4 0
YADDLPPPPVPPPAIKSPTVQSKAQ
YADDLPPPPVPPPAIKSPTVQSKAQ
YADDLPPPPVPPPAIKSPTVQSKAQ
YADDLPPPPVPPPAIKSPTVQSKAQ
YADDLPPPPVPPPAIKSPTVQSKAQ
YADDLPPPPVPPPAIKSPTVQSKAQ
Figure 6 - Phospho-regulation of CC3 binding.
(A) Binding of EVH 1 to CC3 motif is ablated by phosphorylation at either of two sites outside of
the core consensus binding sequence. Position of phospho-residues demarcated in red. Each
sequence is duplicated to show repeatability.
134
4.4 Discussion
Studies have shown direct binding in vitro of D. melanogasterEna to the cytoplasmic tail
of Robo. This binding was attenuated, but not lost by either deletion of CCl or CC2 while
deletion of CC3 displayed no effect. The earlier work did not map these interactions or examine
the overall conformation of Robo in the presence of these deletions leaving the possibility that
folding rather than specific binding is affected(Bashaw, 2000). Here, SPOTs arrays performed
with amino acid sequence taken from vertebrate CC1 and CC2 imply conservation and
specificity of these reported interactions. Interestingly, the D. melanogasterCC3 motif shows
weak conservation near it's N-terminus, where the acidic and hydrophobic residues present in
vertebrate CC3 and shown to be important for EVH 1 binding are not present(Ball et al., 2000).
This may underlie the observation that purified EVH1 can bind to vertebrate CC3 in a specific
fashion, while this domain was dispensable for interaction in D. melanogaster(Bashaw,2000).
As noted previously, while there is a demonstrated genetic interaction between Ena and Robo in
flies, loss of function mutations show weaker phenocopy than what I observe between mmvvee
and Slit1/2 or Robol/2 knockout mice. This additional EVH1 binding site supports the
hypothesis that Ena/VASP proteins play an expanded role in vertebrate growth cone sensitivity
to the Slit/Robo pathway.
The well described ideal consensus sequence for EVH 1 binding is
(D/E)(F/L)PXCDP(D/E)(D/E)(Niebuhr et al., 1997). SPOTs arrays demonstrate that substitution
of the phenylalanine residue in position 2 with leucine is tolerated but confers a reduction in
binding affinity(Ball et al., 2000). An LPPPP motif found in Lamellipodin also binds Ena/VASP
with reduced affinity, confirming this finding in a physiological context(Krause et al., 2004).
Interestingly, instead of phenylalanine, leucine is highly conserved in Robo in vertebrates and
135
invertebrates at this position. Furthermore, while not essential, the acidic residues C-terminal to
the last proline have been shown to potentiate EVH1 binding (Krause et al., 2004)and are absent
in both CC2 and CC3 of all of the M musculus Robos. Conservation of residues which are
permissive yet sub-optimal for Ena/VASP binding may imply a number of additional regulatory
mechanisms for EVHl:Robo interaction. The SH3 domain containing adaptor protein Nck has
been shown to be an important link between Robo and Rac activation, and to bind to CC2 of
Robo(Yang and Bashaw, 2006). While the exact residues in CC2 binding Nck remain to be
mapped, if it binds as other SH3 domains to a PXXP motif, the only binding sites within CC2
would be either nested within or overlapping the mapped EVH1 binding site. Similarly the
RhoGAP, SrGAP 1, has been shown to bind to CC3 directly, though this interaction has been
mapped more precisely. Formation of an SrGAP:CC3 complex has been shown to require both
the proline rich region of CC3 and the N-terminal acidic residue also crucial for EVHI binding,
demonstrating overlap between the binding sites of these two proteins. Together these data
suggest the exciting but untested possibility of competition between Nck or SrGAP 1 and
Ena/VASP for binding within the cytoplasmic tail of Robo, a balance that could potentially be
upset by evolutionary mutations enhancing Ena/VASP binding. Possible regulation of the
interaction between Ena/VASP in Robo is further supported by my observation that
phosphorylation of a either a serine or threonine residue outside of CC3 leads to a complete
ablation of Ena/VASP binding. Phosphorylation of this residue has been only demonstrated in
non-neuronal cells however, so verification of its physiological relevance in axon guidance will
be a necessary next step. While kinase predictions are notoriously inaccurate it is tantalizing to
note that the NetPhosK algorithm predicts that both of these sites are predicted to be targets of
136
PKC, a kinase with demonstrated roles in repulsive axon guidance mediated in the
RGMa/Neogenin pathway (Conrad et al., 2007).
SPOTs arrays also confirm the interaction reported to exist in drosophila between CCl
and EVH1. The EVH 1 of Mena, VASP, and EVL bind to a sequence that overlaps the most
highly conserved region of CCl. This makes CC 1 the second member on a list of non-canonical
EVHI motifs described, so far comprised of only Tes. BLAST searches find no significant
similarity between this sequence and proteins outside of the Robo family. Further mapping may
reveal that only a subset of the residues in this sequence are important for EVHI binding, which
may allow discovery of other proteins with the core consensus sequence. While residues outside
of the area of highest conservation are also important for EVH1 binding, it is worth noting that in
both vertebrate Robol and drosophila Robo, five of the seven amino acids in this region are
polar, uncharged residues implying a level of conservation that may be sufficient for binding in
both sequences.
The structure of the cytoplasmic tail of Robo is one that contains three Ena/VASP
binding sites, each of which has either a phosphorylation site that affects binding, an overlapping
binding site for other proteins known to bind Robo, or both. Tightly coordinated context
dependent changes in growth cone sensitivity and response to guidance cues are among the most
appreciated hallmarks of axon guidance, yet given their complexity, our understanding of how
these changes occurs remains incomplete. Understanding the functional significance of the
Ena/VASP:Robo interaction represents the next major question raised by this study.
It is particularly interesting that there are multiple EVH 1 binding sites in the cytoplasmic
tail of Robo given that Ena/VASP proteins are obligate tetramers. Together these observations
137
lead me to propose a model in which the relatively reduced affinity of EVH 1 binding sites in
Robo are compensated for by an increase in avidity.
4.5 Methods
Peptides in SPOTs arrays are covalently linked at the carboxy terminus to the nitrocellulose
membrane where they are synthesized. SPOTs arrays were overlaid with tris-buffered saline
containing 5ug/ml of either GST:EVH1(Mena), GST:EVH1(VASP), GST:EVH1(EVL), or GST
alone. After washing, membranes were incubated in Rabbit anti-GST antibody (Sigma #G778 1)
at a dilution of 1:2000. After washing, membranes were then incubated in Horseradish
peroxidase conjugated Donkey anti Rabbit IgG (Jackson Immunoresearch) at a dilution of
1:5000. Ternary complex of protein:primary:secondary were visualized with standard
ECL(Amersham).
4.5 References
Ball, L. J., Kifhne, R., Hoffmann, B., Hafner, A., Schmieder, P., Volkmer-Engert, R., Hof, M.,
Wahl, M., Schneider-Mergener, J., Walter, U., et al. (2000). Dual epitope recognition by the
VASP EVH1 domain modulates polyproline ligand specificity and binding affinity. EMBO
J. 19, 4903-4914.
Bashaw, G. J. (2000). Repulsive Axon GuidanceAbelson and Enabled Play Opposing Roles
Downstream of the Roundabout Receptor. Cell 101, 703-715.
Blaikie, P., Immanuel, D., Wu, J., Li, N., Yajnik, V., and Margolis, B. (1994). A region in Shc
distinct from the SH2 domain can bind tyrosine-phosphorylated growth factor receptors. J.
138
Biol. Chem. 269, 32031-32034.
Bodda, B., Briggs, D. C., Higgins, T., Garvalov, B. K., Fadden, A. J., McDonald, N.
Q., and
Way, M. (2007). Tes, a specific Mena interacting partner, breaks the rules for EVH1
binding. Mol. Cell 28, 1071-1082.
Brindle, N. P., Holt, M. R., Davies, J. E., Price, C. J., and Critchley, D. R. (1996). The focaladhesion vasodilator-stimulated phosphoprotein (VASP) binds to the proline-rich domain in
vinculin. Biochem. J. 318 (Pt 3), 753-757.
Conrad, S., Genth, H., Hofmann, F., Just, I., and Skutella, T. (2007). Neogenin-RGMa Signaling
at the Growth Cone Is Bone Morphogenetic Protein-independent and Involves RhoA,
ROCK, and PKC. Journal of Biological Chemistry 282, 16423-16433.
Dephoure, N., Zhou, C., Villen, J., Beausoleil, S. A., Bakalarski, C. E., Elledge, S. J., and Gygi,
S. P. (2008). A quantitative atlas of mitotic phosphorylation. Proc Natl Acad Sci USA 105,
10762-10767.
Drees, B., Friederich, E., Fradelizi, J., Louvard, D., Beckerle, M. C., and Golsteyn, R. M. (2000).
Characterization of the interaction between zyxin and members of the Ena/vasodilatorstimulated phosphoprotein family of proteins. J. Biol. Chem. 275, 22503-22511.
Fedorov, A. A., Fedorov, E., Gertler, F., and Almo, S. C. (1999). Structure of EVH1, a novel
proline-rich ligand-binding module involved in cytoskeletal dynamics and neural function Nature Structural & Molecular Biology. Nat. Struct Biol. 6, 661-666.
Hu, H., Li, M., Labrador, J. P., McEwen, J., Lai, E. C., Goodman, C. S., and Bashaw, G. J.
139
(2005). Cross GTPase-activating protein (CrossGAP)/Vilse links the Roundabout receptor to
Rac to regulate midline repulsion. Proc Natl Acad Sci USA 102, 4613-4618.
Kavanaugh, W., and Williams, L. (1994). An alternative to SH2 domains for binding tyrosinephosphorylated proteins. Science 266, 1862-1865.
Krause, M., Leslie, J. D., Stewart, M., Lafuente, E. M., Valderrama, F., Jagannathan, R.,
Strasser, G. A., Rubinson, D. A., Liu, H., Way, M., et al. (2004). Lamellipodin, an
Ena/VASP ligand, is implicated in the regulation of lamellipodial dynamics. Dev. Cell 7,
571-583.
Lafuente, E. M., van Puijenbroek, A. A. F. L., Krause, M., Carman, C. V., Freeman, G. J.,
Berezovskaya, A., Constantine, E., Springer, T. A., Gertler, F. B., and Boussiotis, V. A.
(2004). RIAM, an Ena/VASP and Profilin ligand, interacts with Rap 1-GTP and mediates
Rapl-induced adhesion. Dev. Cell 7, 585-595.
Lundstr6m, A., Gallio, M., Englund, C., Steneberg, P., Hemphali, J., Aspenstr6m, P., Keleman,
K., Falileeva, L., Dickson, B. J., and Samakovlis, C. (2004). Vilse, a conserved Rac/Cdc42
GAP mediating Robo repulsion in tracheal cells and axons. Genes Dev. 18, 2161-2171.
Niebuhr, K., Ebel, F., Frank, R., Reinhard, M., Domann, E., Carl, U. D., Walter, U., Gertler, F.
B., Wehland, J., and Chakraborty, T. (1997). A novel proline-rich motif present in ActA of
Listeria monocytogenes and cytoskeletal proteins is the ligand for the EVH1 domain, a
protein module present in the Ena/VASP family. EMBO J. 16, 5433-5444.
Prehoda, K. E., Lee, D. J., and Lim, W. A. (1999). Structure of the enabled/VASP homology 1
domain-peptide complex: a key component in the spatial control of actin assembly. Cell 97,
140
471-480.
Stein, E., and Tessier-Lavigne, M. (2001). Hierarchical organization of guidance receptors:
silencing of netrin attraction by slit through a Robo/DCC receptor complex. Science 291,
1928-1938.
Yang, L., and Bashaw, G. J. (2006). Son of sevenless directly links the Robo receptor to rac
activation to control axon repulsion at the midline. Neuron 52, 595-607.
141
Chapter 5
Conclusions and future directions
142
5.1 Conclusions and future directions
The founding member of the Ena/VASP family of actin regulatory proteins, Enabled, was
discovered in D. melanogasterin a screen for dominant suppressors of the Ableson non-receptor
tyrosine kinase(Gertler et al., 1990). It was soon thereafter described that homozygosity for loss
of function mutations in Ena cause axon guidance defects in both the peripheral and central
nervous system(Gertler et al., 1995). Ena/VASP proteins have since been implicated in
mediating guidance responses to both repulsive(Bashaw, 2000)and attractive(Gitai et al., 2003)
guidance molecules in invertebrates.
Addressing the question of how conserved a role Ena/VASP proteins play in vertebrate
axonal pathfinding is compounded by a number of problems. While invertebrates have only one
Ena/VASP homolog, in mammals, there are three: Mena, VASP, and EVL. These three proteins
show not only largely overlapping expression patterns during development, but overlapping
functions as evidenced by relatively weak phenotypes observed in mice deficient for any one
family member. While mice deficient in only Mena, or both Mena and VASP (Menzies et al.,
2004)(Lanier et al., 1999)display defects in formation of the hippocampal commissure and the
corpus callosum, these phenotypes are only partially penetrant. Defects observed in formation of
the optic chiasm in Mena, VASP double mutants argue for a conserved role of Ena/VASP in
axon pathfinding, but the guidance molecules responsible have not yet been described.
Achieving an accurate picture of axon guidance in the absence of Ena/VASP proteins in
vertebrates requires the production of animals deficient for all three family members. In our
previous studies we describe the production and initial characterization of these mmvvee
animals. Mice genetically deficient for the three Ena/VASP paralogs display pleiotropic
phenotypes including edema and hemorrhage, (Furman et al., 2007), excencephaly, and a failure
143
in cortical neuritogenesis(Kwiatkowski et al., 2007)(Dent et al., 2007). While this defect in
cortical neuritogenesis represents an novel role for Ena/VASP proteins, it is obviously
confounding to the study of subsequent axon guidance.
In this thesis, I demonstrated that in spinal cords of mmvvee animals sufficient axonal
outgrowth occurs to allow subsequent examination of the role of Ena/VASP in axon guidance.
These animals show profound defects in pathfinding in both the central and peripheral nervous
systems, just as was described originally in D. melanogaster.The defects observed in the dorsal
root ganglia of the peripheral nervous system share substantial phenotypic overlap with those
observed in mice lacking Slitl/2 or Robol/2. These data suggest that the role of Ena/VASP in
regulating Slit/Robo mediated repulsive axon guidance is remarkably conserved in vertebrates.
Defects observed in the spinal motor axons of the central nervous system are quite striking, but
have not been investigated in detail in this study and therefore remain a promising line of
investigation for future studies.
I have observed that Ena/VASP deficient growth cones of the dorsal root ganglia display
a club-like morphology, devoid of filopodia ex vivo and in vitro. These growth cones display a
defect in sensitivity but not in collapse per se upon application of the chemorepellent Slit. While
these data argue for a role of Ena/VASP proteins upstream of Robo in sensitization of the growth
cone to the presence of Slit, they do not argue against further roles of Ena/VASP downstream of
Robo in proper growth cone turning response. Club-like growth cone morphologies are seen in
wild-type animals in vivo during points in an axons pathway in which growth cone turning is not
a necessity(Moon and Gomez, 2005), suggesting the exciting but untested possibility that
temporal or spatial regulation of Ena/VASP induced filopodia formation could underlie changes
in responsiveness to axon guidance cues.
144
Filopodia have a somewhat controversial role in axon guidance. While many studies have
shown a correlation between filopodia formation and growth cone turning, others have shown
examples where filopodia are dispensable for proper guidance (Dwivedy et al., 2007)(Kim et al.,
2002)(Chang et al., 2006). The notion that filopodia are involved in sensitizing a growth cone to
the presence of guidance cues may present a model that reconciles these observations. While
filopodia may be crucial during a subset of axon guidance decisions, there may be other points in
an axons path in which loss of some sensitivity to guidance cues is insufficient to manifest as a
guidance phenotype in vivo. Further evidence for the involvement of filopodia in guidance factor
sensitivity comes from the observation that in wild-type growth cones, Slit application causes a
dramatic filopodia elongating activity, and that this activity appears to happen asymmetrically
during a turning response. Expanded analysis and quantitation of this response will be required
to determine how frequently filopodial elongation occurs during repulsive turning, and whether
this effect is specific to Slit/Robo mediated repulsion. While these data point toward a sensory
role for filopodia, they leave unresolved the mechanism by which this sensitization may be
occurring. The presence of filopodia on a growth cone increases the area that can be sampled in
search of guidance cues. Furthermore, filopodia may represent structures that contain high
concentrations of signaling scaffolds and effector molecules, priming them for intracellular
responses to guidance factor encounter.
Finally, I show that Ena/VASP proteins can bind to three specific sites in the cytoplasmic
tail of Robo, and that two of these interactions can be regulated by phosphorylation in vitro.
SPOTs arrays are powerful tools for high throughput examination of Protein:Peptide interactions,
but also have drawbacks. While relative binding affinities can be estimated, dissociation
constants cannot be determined. Furthermore while the in vitro SPOTs arrays detected
145
interactions between single EVH1 domains and a single consensus binding motifs, this is almost
certainly not how these interactions are occurring in vivo. Ena/VASP proteins are obligate
tetramers and thus may interact with the multiple EVH 1 binding sites in Robo simultaneously,
thereby increasing avidity of this interaction. Multiple binding sites of lower affinity may
represent a way for this interaction to be affected by competition at one or more sites. To address
these questions I have used site directed mutagenesis to engineer expression constructs encoding
Robo with all combinations of Ena/VASP binding motif mutated in ways predicted to prevent
binding. Using these constructs, I will be able to determine the importance of multiple binding
sites for Ena/VASP:Robo interaction in vivo. Furthermore, while this interaction is extremely
intriguing I have yet to determine it's functional relevance. The next major step in this project
will be to determine if this interaction has any bearing on either collapse or turning.
Recent evidence regarding a filopodia promoting activity of the Robo4 receptor provides
an obvious first line of investigation. Overexpression of the sole exclusively non-neuronal
member of the Robo family, Robo4 has recently been shown to potently induce filopodia
formation in human vascular endothelial cells. Interestingly, this filopodial induction requires
Robo 1. Given my observation that DRG growth cones extend very long filopodia in response to
Slit during turning and collape, it may be that Slit application may either promote or strengthen a
Robol:Ena/VASP interaction leading to this elaboration. Observation of the collapse and turning
responses of Robo 1/2 mutants transfected with constructs expressing Robo receptors lacking
combinations of Ena/VASP binding sites will allow me to answer unequivocally the role of the
interaction between Ena/VASP and Robo in axon guidance.
146
Bashaw, G. J. (2000). Repulsive Axon GuidanceAbelson and Enabled Play Opposing Roles
Downstream of the Roundabout Receptor. Cell 101, 703-715.
Chang, C., Adler, C. E., Krause, M., Clark, S. G., Gertler, F. B., Tessier-Lavigne, M., and
Bargmann, C. I. (2006). MIG-10/lamellipodin and AGE-1/PI3K promote axon guidance and
outgrowth in response to slit and netrin. Curr. Biol. 16, 854-862.
Dent, E. W., Kwiatkowski, A. V., Mebane, L. M., Philippar, U., Barzik, M., Rubinson, D. A.,
Gupton, S., van Veen, J. E., Furman, C., Zhang, J., et al. (2007). Filopodia are required for
cortical neurite initiation. Nat. Cell Biol. 9, 1347-1359.
Dwivedy, A., Gertler, F. B., Miller, J., Holt, C. E., and Lebrand, C. (2007). Ena/VASP function
in retinal axons is required for terminal arborization but not pathway navigation.
Development 134, 2137-2146.
Furman, C., Sieminski, A. L., Kwiatkowski, A. V., Rubinson, D. A., Vasile, E., Bronson, R. T.,
Fassler, R., and Gertler, F. B. (2007). Ena/VASP is required for endothelial barrier function
in vivo. J. Cell Biol. 179, 761-775.
Gertler, F. B., Comer, A. R., Juang, J. L., Ahern, S. M., Clark, M. J., Liebl, E. C., and Hoffmann,
F. M. (1995). enabled, a dosage-sensitive suppressor of mutations in the Drosophila Abl
tyrosine kinase, encodes an Abl substrate with SH3 domain-binding properties. Genes Dev.
9, 521-533.
Gertler, F. B., Doctor, J. S., and Hoffmann, F. M. (1990). Genetic suppression of mutations in
the Drosophila abl proto-oncogene homolog. Science 248, 857-860.
147
Gitai, Z., Yu, T., Lundquist, E., and Tessier-Lavigne, M. (2003). ScienceDirect - Neuron: The
Netrin Receptor UNC-40/DCC Stimulates Axon Attraction and Outgrowth through Enabled
and, in Parallel, Rac and UNC- 115/AbLIM. Neuron.
Kim, M. D., Kolodziej, P., and Chiba, A. (2002). Growth cone pathfinding and filopodial
dynamics are mediated separately by Cdc42 activation. J Neurosci 22, 1794-1806.
Kwiatkowski, A. V., Rubinson, D. A., Dent, E. W., Edward van Veen, J., Leslie, J. D., Zhang, J.,
Mebane, L. M., Philippar, U., Pinheiro, E. M., Burds, A. A., et al. (2007). Ena/VASP Is
Required for neuritogenesis in the developing cortex. Neuron 56, 441-455.
Lanier, L. M., Gates, M. A., Witke, W., Menzies, A. S., Wehman, A. M., Macklis, J. D.,
Kwiatkowski, D., Soriano, P., and Gertler, F. B. (1999). Mena Is Required for Neurulation
and Commissure Formation. Neuron 22, 313-325.
Menzies, A. S., Aszodi, A., Williams, S. E., Pfeifer, A., Wehman, A. M., Goh, K. L., Mason, C.
A., Fassler, R., and Gertler, F. B. (2004). Mena and vasodilator-stimulated phosphoprotein
are required for multiple actin-dependent processes that shape the vertebrate nervous system.
J Neurosci 24, 8029-8038.
Moon, M.-S., and Gomez, T. M. (2005). Adjacent pioneer commissural interneuron growth
cones switch from contact avoidance to axon fasciculation after midline crossing. Dev. Biol.
288, 474-486.
148