EphA4-Dependent Axon Guidance Is
Mediated by the RacGAP a2-Chimaerin
Heike Wegmeyer,1Joaquim Egea,2Nadine Rabe,4Henrik Gezelius,4Alessandro Filosa,2,3Anders Enjin,4
Frederique Varoqueaux,1Katrin Deininger,2Frank Schnu ¨tgen,5Nils Brose,1Ru ¨diger Klein,2Klas Kullander,4
and Andrea Betz1,*
1Department of Molecular Neurobiology and DFG Center for Molecular Physiology of the Brain, Max Planck Institute
of Experimental Medicine, D-37075 Go ¨ttingen, Germany
2Department of Molecular Neurobiology
3The International Max Planck Research School for Molecular and Cellular Life Sciences
Max Planck Institute of Neurobiology, D-82152 Martinsried, Germany
4Department of Neuroscience, Unit of Developmental Genetics, Uppsala University, S-75123 Uppsala, Sweden
5Department of Molecular Hematology, University of Frankfurt Medical School, D-60590 Frankfurt/Main, Germany
Neuronal network formation in the developing
nervous system is dependent on the accurate
navigation of nerve cell axons and dendrites,
which is controlled by attractive and repulsive
guidance cues. Ephrins and their cognate Eph
receptors mediate many repulsive axonal guid-
ance decisions by intercellular interactions
resulting in growth cone collapse and axon
retraction of the Eph-presenting neuron. We
show that the Rac-specific GTPase-activating
protein a2-chimaerin binds activated EphA4
and mediates EphA4-triggered axonal growth
display a phenotype similar to that of EphA4
mutant mice, including aberrant midline axon
guidance and defective spinal cord central pat-
tern generator activity. Our results reveal an a-
chimaerin-dependent signaling pathway down-
stream of EphA4, which is essential for axon
guidance decisions and neuronal circuit forma-
tion in vivo.
Many axonal guidance processes that take place during
neuronal network formation in mammals are regulated
by the ephrin/Eph system. Ephrinsare surface-associated
ligands that bind to Eph receptors, which constitute the
largest family of receptor tyrosine kinases. Thisinteraction
triggers intracellular signaling pathways in the contacting
nerve cells and typically causes the collapse of Eph-pre-
senting growth cones and axon repulsion. Eph receptor
signaling during axon guidance is initiated by a multistep
process that involves ephrin binding, activation of the
tyrosine kinaseactivity, autophosphorylation,and
higher-order clustering. Eph receptor complexes acti-
vated in this manner signal to Rho family GTPases to trig-
ger cell-cell detachment and reorganization of the actin
cytoskeleton, which ultimately results in growth cone col-
lapse and axon retraction (Egea and Klein, 2007). The
current knowledge about intracellular signaling proteins
that transduce Eph receptor signals to Rho family
GTPases is fragmentary. In fact, the only identified
RhoGTPase regulators that act downstream of activated
Eph receptors are guanine nucleotide exchange factors
(RhoGEFs), of which the best characterized are members
of the ephexin and Vav families (Sahin et al., 2005; Cowan
et al., 2005). Whether and how their activity is counterbal-
anced and complemented by Eph-dependent Rho/Rac
inhibitors such as RhoGTPase-activating proteins (Rho-
GAPs) is unknown.
Members of the chimaerin family of Rac-specific GAPs
are candidate regulators of the outgrowth and pathfinding
of neuronal axons and dendrites. The mouse genome
contains two chimaerin genes (a-chimaerin, chn1; b-chi-
maerin, chn2), each of which gives rise to the expression
of two isoforms, a1/a2 and b1/b2. All chimaerins contain
a RacGAP domain anda regulatory diacylglycerol-binding
C1 domain, while only a2- and b2-chimaerin contain an
additional SH2 domain at their N termini (Yang and Kaza-
nietz, 2003). Mammalian a2-chimaerin, which is mainly
expressed in the embryonic nervous system, is thought
to be involved in semaphorin3A-induced growth cone col-
lapse of rat dorsal root ganglion neurons (Brown et al.,
2004). In zebrafish, a2-chimaerin appears to be required
for morphogenic movements during early development
(Leskow et al., 2006). Mammalian a1-chimaerin, on the
other hand, is mainly expressed postnatally and is
involved in the pruning of dendritic spines and branches
of hippocampal neurons (Van de Ven et al., 2005; Buttery
et al., 2006). These functions are dependent on the
RacGAP domains of a1- and a2-chimaerin and were
shown to require C1 domain ligand binding (Brown et al.,
2004; Van de Ven et al., 2005; Buttery et al., 2006).
756 Neuron 55, 756–767, September 6, 2007 ª2007 Elsevier Inc.
In the present study, we used biochemical and mouse
genetic tools to examine the role of a2-chimaerin in vivo.
of activated EphA4 and is essential for EphA4-dependent
axon navigation during neuronal network formation in the
mutant mice causes deficient Eph-mediated growth cone
collapse in cortical and spinal cord neurons. The resulting
mutant phenotype thus mimics that of mouse mutants
with deficient EphA4-dependent forward signaling. It is
characterized by aberrant wiring of neuronal networks
that control coordinated limb movements, which in turn
causes an unusual hopping gait.
Chimaerins Interact with Eph Receptors
To identify interactors of the a2-chimaerin SH2 domain
that trigger or regulate a2-chimaerin signaling to Rac,
we performed a modified yeast two-hybrid (YTH) screen
using a rat brain cDNA prey library together with a fusion
protein of the a2-chimaerin SH2 domain and the Src
kinase domain as bait. Under these conditions, the Src
kinase domain can phosphorylate tyrosine residues of
prey proteins, a posttranslational modification that is typ-
ically required for SH2 domain interactions but would
otherwise not take place in yeast (Park and Yun, 2001).
We isolated multiple identical prey clones encoding the
kinase domain (residues 632–902) of EphB1 as interactors
of the a2-chimaerin SH2 domain. To test whether this
interaction is a common feature of chimaerins and
extends to other Eph receptors, we studied interactions
of a2-chimaerin-SH2-src-kinase and b2-chimaerin-SH2-
src-kinase fusion constructs with the kinase domains of
EphB1 and EphA4 (residues 634–902) in YTH assays.
The SH2-src-kinase fusion proteins of a2- and b2-chi-
maerin bound to the kinase domains of EphB1 and
EphA4 while the isolated SH2 or Src kinase domains did
not or only very weakly (Figure 1), indicating that the inter-
action is phosphorylation dependent and common to all
chimaerin SH2 domains and multiple Eph receptors.
Similar Phenotypic Changes after Genetic
Perturbation of a-Chimaerin or EphA4 Function
To analyze the function of a-chimaerins in vivo, we gener-
ated knockout mice (a-chimaerin?/?) lacking a-chimaerin
expression entirely and knockin mice (a-chimaerinmut/mut)
with a point mutation of the first histidine in the C1
domain (H81K and H206K in a1- and a2-chimaerin,
respectively; see Figure S1 in the Supplemental Data
available with this article online). The knockin was
intended to result in the expression of diacylglycerol
sion and expression of an aberrantly spliced a2-chimaerin
transcript. This led to a 26 residue deletion in the a2-chi-
maerin protein (D184-209) and a 95% reduction of total
a2-chimaerin protein expression (Figure S2).
Heterozygous animals of both mutant lines lacked any
obvious phenotypic changes.Similarly, newborn homozy-
gous mutants of both lines were viable and indistinguish-
able from wild-type (WT) and heterozygous littermates.
Shortly after birth, both a-chimaerin mutant lines dis-
played smaller body sizes but in most cases acquired
normal body size until early adulthood. With the onset of
locomotion, a rabbit-like hopping gait was apparent in a-
chimaerin?/?and a-chimaerinmut/mutmice. Instead of
a normal alternating gait, these animals moved first their
forelimbs in parallel and then their hindlimbs in parallel
(Figures 2A–2C). In addition, homozygous mutants of
both lines exhibited tonic seizures, and a-chimaerin?/?
mice showed a tendency for rightward circular locomo-
tion. Analyses of the brain morphology of a-chimaerin
mouse mutants revealed enlarged ventricles in some a-
chimaerinmut/mutand all a-chimaerin?/?mice but no other
obvious changes in the cytoarchitecture of the brain, in-
cluding the anterior commissure (AC) (Figure S3).
a-Chimaerin mutant mice display a hopping gait similar
to EphA4?/?mice (Dottori et al., 1998), ephrinB3?/?mice
(Yokoyama et al., 2001; Kullander et al., 2001a), and
knockin mice expressing EphA4 variants with defective
Figure 1. SH2 Domains of a2- and b2-Chimaerin Interact with
the Kinase Domains of EphB1 and EphA4 in YTH Assays
(A) The SH2 domain of a2-chimaerin (black bar) was fused to the
kinase domain of Src and used as bait in YTH screens. SH2, Src
homology domain 2; C1, C1 domain; RacGAP, Rac GTPase-activating
protein domain; aa, amino acids.
(B) Residues 632–902 of the EphB1 intracellular domain (black bar)
were identifiedto interact with
EphB1/A4-IC, EphB1 and EphA4 intracellular domains; Tyr kinase,
tyrosine kinase domain; SAM, sterile a motif domain; PDZ-bd, PDZ
(C) EphB1 and EphA4 prey constructs (black bar in [B]) or the empty
prey vector were tested for interaction with several bait constructs.
Relative growth of double transfected yeast colonies on selection
plates, indicative for the interaction of baits and preys, is shown.
Ratings were ?, +, ++, +++ for no, weak, medium, and strong growth.
Neuron 55, 756–767, September 6, 2007 ª2007 Elsevier Inc. 757
Chimaerins Mediate EphA4-Dependent Axon Guidance
forward signaling (Kullander et al., 2001b; Grunwald et al.,
tional involvement of a2-chimaerin in EphA4 forward
signaling. EphA4?/?mice and several EphA4 mutants
with defective forward signaling show perturbations of
spinal cord architecture due to aberrant axon guidance
decisions of mutant neurons (Dottori et al., 1998; Kul-
lander et al., 2001b). Such changes were also seen in
a-chimaerin mutant mice. Their dorsal funiculi, which con-
tain axons of the corticospinal tract (CST) originating from
neurons in the motor cortex, were broader (Figures 2D–
tant pathway in the regulation of voluntary distal limb
movements, aberrantly recrossed the spinal cord midline
(Figures 2H–2K). Most of the defects seen in the a-chi-
maerin mutants were more severe in a-chimaerin?/?
than in a-chimaerinmut/mutmice, indicating that the
remaining aberrantly spliced a2-chimaerin-D184-209 is
still partially active.
Elimination of EphA4 forward signaling in the mouse
also results in a dysfunction of the central pattern genera-
tor (CPG) for hindlimb movement (Kullander et al., 2003),
an oscillatory neuronal network that generates rhythmic
activities and thus patterns muscle contractions. EphA4-
positive neurons are a major component of ipsilaterally
projecting excitatory interneurons, which are likely part
of the locomotor CPG. In addition, the locomotor CPG
contains ipsilaterally projecting inhibitory interneurons,
which coordinate flexor-extensor activities, and inhibitory
and excitatory commissural interneurons (CINs), which
the activity of contralateral motor neurons and hence
coordinate left-right activities (Kiehn, 2006). We recorded
the activity of the locomotor CPG electrophysiologically
using isolated spinal cords from a-chimaerin?/?mice
and WT littermates. Recordings from WT mice showed
alternating left-right ventral root activities, which account
for alternating left-right limb movements (Figures 3A–
3C). In contrast, the left-right firing pattern of the ventral
roots in lumbar segment 2 (lL2-rL2) of a-chimaerin?/?an-
imals was rather uncoordinated and switched between
asynchronousand synchronousactivities, whereas
Figure 2. a-Chimaerinmut/mutand a-Chimaerin?/?Mice Dis-
play a Hopping Gait, Spinal Cord Malformations, and Abnor-
mal Midline Crossing of Corticospinal Tract Fibers
(A–C) Gait analyses of WT, a-chimaerinmut/mut, and a-chimaerin?/?
mice. Forepaws were stained with red, hindpaws with blue nontoxic
dyes (A). Determination of the distances covered by left and right
paws (b) and the distance covered by the same paw (a) for quantifica-
tion of the degree of parallel movement of hindlimbs (B). Comparison
of the b/a ratio of wild-type (n = 4), a-chimaerinmut/mut(n = 4), and
a-chimaerin?/?(n = 3) mice (C).
(D–F) Dark field photographs of spinal cord cross-sections at the
lumbar level of WT (D), a-chimaerinmut/mut(E), and a-chimaerin?/?
mice (F). The central canal is indicated by an arrowhead. Scale bar,
500 mm. DF, dorsal funiculus.
(G) Comparison of the percentage of the total height of the spinal cord
and a-chimaerin?/?mice (n = 3 for each genotype; 7 slices per animal).
(H–J) Cortical spinal tract axons of 10- to 11-week-old WT, a-chimaer-
inmut/mut, and a-chimaerin?/?mice were traced anterogradely with
biotin-dextran-amine by unilateral injection of the tracer into the right
hemisphere of the motor cortex. Many CST fibers (arrowheads)
recrossed the midline (dashed line) in the brachial part of the spinal
cord in a-chimaerinmut/mutand a-chimaerin?/?mice (analyzed 10–13
days after tracer injection). Scale bar, 100 mm.
(K) Quantification of recrossing fibers inWT, a-chimaerinmut/mut,and a-
chimaerin?/?mice (n = 3 for each genotype; 40–90 slices per animal).
Error bars indicate SEM; *p < 0.05; **p < 0.01 compared to wild-type
control (Student’s t test). mut/mut, a-chimaerinmut/mut; ?/?, a-
758 Neuron 55, 756–767, September 6, 2007 ª2007 Elsevier Inc.
Chimaerins Mediate EphA4-Dependent Axon Guidance
ipsilateral intersegmental flexor-extensor (lL2-lL5) activi-
ties were normal (Figures 3D–3F). This phenotype is
reminiscent of the consequences of defective EphA4
signaling, which results in the synchronization of left-right
hindlimb motor activities, most likely due to axons of
excitatory interneurons that aberrantly cross the midline
of the ventral spinal cord (Kullander et al., 2003; Butt
et al., 2005).
To examine whether the uncoordinated left-right spinal
ventral root activities observed in the a-chimaerin?/?mice
are due to a defect in the distribution of CINs, we applied
tracer substances to one side of lumbar segments of WT
and a-chimaerin?/?spinal cords to visualize CINs and
their processes. a-Chimaerin?/?mice showed fibers that
aberrantly crossed over the midline dorsally and ventrally
of the central canal, while no obvious defects in the
position and number of CINs were detectable (Figures
3G–3X and Figure S6). These defects are similar to the de-
fects observed in EphA4?/?mice (Kullander et al., 2003).
Taken together, the interaction between a2-chimaerin
and Eph receptors and the similar phenotypic conse-
quences of a-chimaerin mutations and perturbed EphA4
effectors of ephrin-triggered forward signaling through
a2-Chimaerin exhibits high expression levels in the rat
embryonic nervous system, while the expression of a1-
chimaerin essentially commences after birth (Hall et al.,
2001). Among other regions, a2-chimaerin mRNA is
Figure 3. Lost Left-Right Coordination of the Hindlimb CPG
and Abnormal Midline Crossing of Axons in Spinal Cords of
(A–F) Electrophysiological recordings of the CPG for hindlimb locomo-
tion.Recordedactivityinthelumbar ventralroots, atlevel 2and 5,after
application of NMDA, serotonin, and dopamine to the isolated spinal
cord of P0–P3 mice (A and D). Circular phase diagrams of intraseg-
mental left-right and intersegmental flexion/extension coordination
(B and E) of animals shown in (A) and (D), respectively. Each dot on
the unit circle represents the phase value of one locomotor cycle,
a phase value of 0.5 reflects alternating activity, and 0.0 reflects
synchronous activity. Combining 15 locomotor cycles covering 75
cycles from one animal gives a vector, the direction set by the mean
phase value, and the length given by the focus of phase values around
the mean. A vector endpoint outside the dotted circle marks a highly
significant coordination. Circular phase diagram of intrasegmental
left-right coordination of WT ([C]; n = 3) and a-chimaerin?/?mice ([F];
n = 6), each point represents the endpoint of a vector. l, left; r, right;
L, lumbar segment.
(G–X) Tracing of CINs in the spinal cord area of the hindlimb locomotor
CPG. Schematic drawing of tracer applications on the ventral spinal
cord (Stokke et al., 2002) (G and H). Intrasegmental tracing using fluo-
rescein-dextran-amine (FDA) to show local projecting CINs (G). Inter-
segmental tracing using FDA and rhodamine-dextran-amine (RDA) to
locate ascending (aCINs), descending (dCINs), and bifurcating CINs
(adCINs) (H). Photomicrographs of transverse sections of the L2 seg-
ment from WT littermate (I and J) and a-chimaerin?/?(K and L) spinal
cords traced according to (G), low-magnification (I and K) and close-
up (J and L), respectively. Note that aberrant fibers in a-chimaerin?/?
mice cross the midline dorsal from the central canal while contralateral
dorsal fibers in WT littermate spinal cords originate from dorsal inter-
neurons and cross the midline at the ventral commissure. Photomicro-
graphs from WT littermates (M–R) and a-chimaerin?/?(S–X) spinal
cords traced according to (H). RDA-labeled population in left column,
FDA-labeled population in middle column, and both populations in
a double exposure in right column. Aberrant fibers cross over the
midline in spinal cords of a-chimaerin?/?mice dorsal and ventral
from the central canal while no obvious defects are seen regarding po-
sition and number of aCINs, dCINs, and adCINs. Approximate posi-
tions of CINs are indicated with a dashed circle (M). White box in (I),
(M), and (S) indicates area of close-up (J, L, P–R, and V–X). The outline
ofthespinalcordisindicated byadashedline(I,K,M,and S),andcen-
Scale bar, 200 mm. ?/?, a-chimaerin?/?.
Neuron 55, 756–767, September 6, 2007 ª2007 Elsevier Inc. 759
Chimaerins Mediate EphA4-Dependent Axon Guidance
detected in the cortex and in the spinal cord of rat em-
bryos. In particular, at E12.5, a2-chimaerin mRNA is ex-
pressed at high levels in the spinal cord mantle layer, es-
pecially in differentiating motor neurons of the ventral
horn. By E16.5, an increased expression can be detected
in the cortical plate, from which the corticospinal tract
axons arise (Hall et al., 2001). This expression pattern is
coincident with the EphA4 expression in the spinal cord
and the cortical plate (Kullander et al., 2001b; Greferath
et al., 2002). We compared the expression of a2-chi-
maerin and EphA4 by in situ hybridizations in the mouse
spinal cord (E15.5 and P0) as well as motor cortex
(E18.5 and P0) and found both mRNAs to be expressed
in the same neuronal populations. We specifically studied
the motor cortex in these in situ hybridization experiments
because we used the neurons from this region for growth
cone collapse and tracing experiments. a2-Chimaerin
showed a widespread expression throughout the gray
matter in spinal cord segments analyzed, covering the
expression domain of EphA4 mRNA. In the motor cortex,
both EphA4 and a2-chimaerin were expressed in the cor-
tical plate and the ventricular and subventricular zones
(Figure S4).These datashow that a2-chimaerin isspatially
and temporally expressed in the same regions of the spi-
nal cord and motor cortex as EphA4 and may explain the
similar defects observed in CSTs and spinal cords of
a-Chimaerins Are Components of EphA Receptor
Forward Signaling Pathways
CST axons project from neurons located in the motor cor-
tex. Therefore, we used dissociated neuron cultures of
E16.5 motor cortices from a-chimaerin mutants and WT
littermates in order to directly address the involvement
of a-chimaerins in Eph forward signaling. We stimulated
the neurons with either ephrinB3 or ephrinA1 and counted
the number of collapsed growth cones. In these cultures,
ephrinB3 is a specific ligand for EphA4, whereas ephrinA1
activates additional EphA receptor subtypes. Cortical
neurons from knockin mice with disrupted EphA4 forward
signaling no longer respond to ephrinB3 but are still
responsive to ephrinA1 (Egea et al., 2005). In cultures of
a-chimaerinmut/mutand a-chimaerin?/?neurons, the num-
after stimulation with either ephrinB3 or ephrinA1 (Fig-
ure 4). This was not caused by loss of EphA4 expression
and/or activity, because the amount of EphA4 protein in
WT and a-chimaerin?/?neurons as well as the degree of
EphA4 phosphorylation after ephrinA1 stimulation were
unchanged (Figure 4D). These findings show that EphA4
and likely also other Eph receptors use a-chimaerins for
We next analyzed how EphA4 regulates a2-chimaerin.
We used full-length WT EphA4 (EphA4 WT) or mutant
EphA4 with an inactive kinase domain (EphA4 KD)
Figure 4. EphrinB3- and EphrinA1-In-
duced In Vitro Growth Cone Collapse Is
Impaired in a-Chimaerinmut/mutand a-
(A and B) Cortical neurons from WT, a-
chimaerinmut/mut, and a-chimaerin?/?E16.5
embryos were kept in culture for 36 hr, then
stimulated for 30 min with 2 mg/ml of preclus-
tered Fc (A and B), ephrinB3 (A), or ephrinA1
(B), fixed, stained with phalloidin, and scored
for collapsed or uncollapsed growth cones. In
each experiment, the average SEM of the
percentage of collapsed growth cones from
three to five coverslips (80–100 growth cones
per coverslip) was plotted. **p < 0,01; ***p <
0,001 (Student’s t test).
(C) Representative examplesof scored phalloi-
din-stained neurons. Neurons with an unaf-
fected growth cone (GC) with strong actin
staining (arrow) were considered as uncol-
lapsed, whereas neurons were scored as
collapsed when reduced actin staining was
observed and no filopodial extensions at the
Scale bar, 10 mm.
(D) Western blot analysis of immunoprecipi-
tated EphA4 and total cell lysate EphA4 from
cultured WT and a-chimaerin?/?cortical neu-
rons after stimulation with preclustered Fc
or ephrinA1 (cultures and stimulation like in
[A] and [B]). WT, wild-type; mut/mut, a-
WB, western blot.
760 Neuron 55, 756–767, September 6, 2007 ª2007 Elsevier Inc.
Chimaerins Mediate EphA4-Dependent Axon Guidance
(Kullander et al., 2001a) expressed in 293FT cells and
studied their binding to bacterially expressed GST-tagged
a-chimaerin constructs. We followed this experimental
strategy because immunoprecipitation of either EphA4
or a-chimaerins from brain extracts or coexpressing cul-
tured cells did not result in the coprecipitation of the cor-
responding binding partner. Full-length a2-chimaerin in-
teracted with EphA4 WT and to a lesser extent with
EphA4 KD. The same was true for an N-terminal a2-chi-
maerin fragment containing the SH2 domain (residues
1–183), while a corresponding mutant fragment with
1–183, R73L) bound only very weakly to EphA4 WT and
EphA4 KD (Figure 5A). Surprisingly, also a1-chimaerin,
which has a completely unrelated N terminus without an
SH2 domain, bound equally strongly to EphA4 WT and
EphA4 KD, indicating an interaction that is independent
of afunctional kinase domain(Figure 5A).An a2-chimaerin
variant lacking its SH2 domain bound more strongly to
EphA4 WT than to EphA4 KD (Figure S5A), demonstrating
that the SH2 domain is not the only EphA4 binding site in
a2-chimaerin. However, a conserved binding site in a-chi-
maerins for EphA4 would not explain the phosphorylation
independence of the a1-chimaerin-EphA4 interaction,
which rather reflects a specific binding site for a1-chi-
maerin in EphA4.
To determine whether EphA4 provides a second bind-
ing site for a-chimaerins outside its kinase domain, we
tested whether a-chimaerins can bind to the juxtamem-
brane (JM) region of EphA4 as well. We used two different
Figure 5. Protein Interactions and EphA4-Dependent Tyrosine Phosphorylation of a-Chimaerins
(A–C) GST-a-chimaerins interact with EphA4 (A and B), and GST-a2-chimaerin proteins but not GST-a1-chimaerin bind to Nck1 and Nck2 in cose-
dimentation experiments (C). Indicated EphA4 and Nck proteins were expressed in 293FT cells and used in cosedimentation experiments with bac-
terially expressed GST-tagged a-chimaerin fusion proteins. GST, glutathione S-transferase; GST-chim, GST-tagged a-chimaerin proteins; a1, a1-
chimaerin; a2, a2-chimaerin; a2-1-183 WT, a2-chimaerin 1-183; a2-1-183 R73L, a2-chimaerin 1-183R73L; EphA4 WT, wild-type EphA4; EphA4
KD, EphA4K653M; EphA4 FF, EphA4Y596F,Y602F; EphA4 EE, EphA4Y596E,Y602E; pTyr, phosphotyrosine; Lysate, 293FT cell lysate containing heterolo-
gously expressed proteins; WB, western blot.
(D–F) a-Chimaerin proteins undergo EphA4-dependent phosphorylation. Indicated myc-tagged a-chimaerin and untagged EphA4 proteins were
coexpressed in 293FT cells, and a-chimaerin proteins were immunoprecipitated (D and E). Cultured cortical neurons from P0 mice were infected
2 hr after plating with recombinant lentiviruses leading to the expression of EGFP-a2-chimaerin. At DIV5 these neurons were either control stimulated
with preclustered Fc or stimulated with preclusterd Fc-ephrinA1 (5 mg/ml each) for the indicated times (F). Myc-chim, myc-tagged a-chimaerin
proteins; a2, a2-chimaerin; a1, a1-chimaerin; a2-R73L, a2-chimaerinR73L; a2-Y202F, a2-chimaerinY202F; a2-Y303F, a2-chimaerinY303F; a2-Y333F,
a2-chimaerinY333F; a2-Y396F, a2-chimaerinY396F; a2-Y443F, a2-chimaerinY443F; EGFP-a2-chim, EGFP-a2-chimaerin; IP, immunoprecipitate; GFP,
green fluorescent protein; EGFP, enhanced green fluorescent protein.
Neuron 55, 756–767, September 6, 2007 ª2007 Elsevier Inc. 761
Chimaerins Mediate EphA4-Dependent Axon Guidance
EphA4 mutants, EphA4EEand EphA4FF, where the two ty-
rosine residues in the JM region (Y596, Y602), which nor-
mally are autophosphorylated upon EphA4 activation, are
replaced by glutamic acid and phenylalanine residues, re-
spectively. These mutants lack the ability to bind to SH2
domains (Zisch et al., 2000; Binns et al., 2000; Cowan
et al., 2005). We found that the efficient binding of full-
length a-chimaerins to EphA4 does not seem to require
its JM tyrosines, as both a1- and a2-chimaerin interacted
equally well with EphA4 WT, EphA4EE, and EphA4FF
(Figure 5B). However, we cannot exclude the possibility
ing in a way that evaded detection with our assay system.
In addition, the binding of both a-chimaerin variants to an
EphA4 mutant lacking its C-terminal SAM and PDZ bind-
ing domains was unaffected by the EphA4 mutation
(data not shown). These data confirm our initial finding
from YTH assays in which the kinase domain of EphA4
is sufficient to mediate the binding to a2-chimaerin and
indicate the presence of a binding site for a1-chimaerin
in the EphA4 kinase domain as well.
Some downstream effectors of activated EphA4 are
phosphorylated in an EphA4-dependent manner (Murai
and Pasquale, 2005). To test whether a1- and a2-chi-
maerin are tyrosine phosphorylated in an EphA4-depen-
dent manner, we coexpressed them with either EphA4
WT or EphA4 KD in 293FT cells and analyzed their phos-
phorylation status after immunoprecipitation by western
blotting. Full-length a1- and a2-chimaerin were tyrosine
phosphorylated in the presence of EphA4 WT but not in
the presence of EphA4 KD or in the absence of EphA4
(Figure 5D). a2-ChimaerinR73Lwas also tyrosine phos-
phorylated upon coexpression with EphA4 WT, albeit to
a lesser extent than WT a2-chimaerin (Figure 5E).
To test for tyrosine phosphorylation of a2-chimaerin by
neuronal EphA4, we expressed EGFP-a2-chimaerin in
cortical cultures and analyzed its phosphorylation after
ephrinA1 stimulation. We found that a2-chimaerin was
phosphorylated on tyrosine residues after stimulation of
endogenous EphA receptors (Figure 5F).
In an attempt to identify tyrosine phosphorylated resi-
dues in a-chimaerins, we first mutated several tyrosines
in the conserved C terminus and studied the phosphoryla-
tion of these mutants. Based on reduced phosphorylation
levels in the corresponding mutants, we identified Y202,
Y303, and Y333 of a2-chimaerin as putative phosphoryla-
tion sites (Figure 5E). In addition to the conserved C termi-
nus, the N terminus (residues 1–200) of a2-chimaerin was
also analyzed and found to be phosphorylated in an
EphA4-dependent manner (Figure S5D). These data
show that a-chimaerins contain multiple tyrosine phos-
The above data show that intact full-length a2-chi-
maerin binds to the kinase domain of EphA4 primarily
when EphA4 is activated, i.e., tyrosine phosphorylated,
while a1-chimaerin binds EphA4 in a constitutive manner.
The a2-chimaerin-EphA4 interaction is mediated by the
SH2 domain of a2-chimaerin and an additional binding
site. Both, a1- and a2-chimaerin are tyrosine phosphory-
lated in an EphA4-dependent manner. Apparently, this is
not required for EphA4 binding in experiments with bacte-
rially expressed proteins, where unphosphorylated a-chi-
maerin proteins interacted with EphA4.
The Eph Adaptors Nck1 and Nck2 Bind
to a2- and b2-Chimaerin
Nck1 and Nck2 (Grb4) are adaptor proteins with one SH2
domain and three SH3 domains that interact with Eph
receptors (Holland et al., 1997; Stein et al., 1998; Bisson
et al., 2007). a2-Chimaerin was recently shown to bind
Nck1 (Wells et al., 2006). WT full-length a2-chimaerin,
a1-chimaerin, bound to Nck1 and Nck2
(Figure 5C). The interaction requires the SH2 domain of
a2-chimaerin and is unaffected by the R73L mutation
(Figure 5C and Figure S5C). Additionally, a GST-Nck1
fusion protein lacking its SH2 domain bound to full-length
a2-chimaerin (Figure S5B). These data indicate that a2-
chimaerin and Nck form a stable complex involving the
chimaerin SH2 domain and one of the Nck SH3 domains.
Regulation of the RacGAP Activity of a-Chimaerins
We next examined the consequences of EphA4 and Nck
binding for the RacGAP activity of a-chimaerins. We
expressed a-chimaerins in the presence or absence of
EphA4 and Nck2 in 293FT cells and measured the GTP
load of endogenous Rac in pull-down assays. In the
absence of EphA4, WT a1- and a2-chimaerin exhibited
similar RacGAP activities, causing similar reductions in
RacGTP levels (Figure 6, bars 7 versus 1 and 11 versus 1).
To analyze the influence of the C1 domain on the
RacGAP activity of a2-chimaerin, we used its H206K
mutant variant in the same assay. Structural data on b2-
chimaerin showed that the molecule in its inactive state
is kept in a closed conformation by intramolecular interac-
tions of several regions with the C1 domain. In this confor-
mation, the N terminus protrudes into the RacGAP do-
main, thereby blocking Rac binding (Canagarajah et al.,
2004). Based on these data, it was predicted that DAG
and phospholipid binding to the C1 domain causes trans-
location of b2-chimaerin to the plasma membrane and un-
folding of the protein, which exposes its RacGAP domain
and thereby increases RacGAP activity (Canagarajah
et al., 2004). Indeed, b2-chimaerin exhibits an increased
RacGAP activity in RacGTP pull-down assays when stim-
ulated with the phorbol ester PMA, a functional analog of
DAG (Caloca et al., 2003). In COS cells, the induction of
EGF receptor signaling with subsequent DAG production
by phospholipase Cg (PLCg) causes b2-chimaerin to
translocate to the plasma membrane where it binds to
Rac1. In the same study, RNAi knock down of endoge-
nous b2-chimaerin in HEK293 cells resulted in a sustained
Rac activity after EGF stimulation (Wang et al., 2006).
Similar C1-domain-mediated effects were reported for
a2-chimaerin. When expressed in neuroblastoma cells,
a2-chimaerin translocates to cell membranes, reduces
endogenous RacGTP levels, and promotes neurite
762 Neuron 55, 756–767, September 6, 2007 ª2007 Elsevier Inc.
Chimaerins Mediate EphA4-Dependent Axon Guidance
retraction after PMA treatment (Brown et al., 2004). The
bited conformation in the cytosol (Brown et al., 2004). Be-
cause of the high similarity of a2- and b2-chimaerin in
terms of their primary amino acid sequence and spatial
distribution of structural domains within the molecules,
we hypothesized that inactive a2-chimaerin is folded in
a closed conformation similar to the structure described
for inactive b2-chimaerin (Canagarajah et al., 2004). To
mimic the unfolding of the protein that is naturally induced
by DAG and phospholipid binding, we mutated the invari-
ant first histidine of the a2-chimaerin C1 domain (H206K).
Although this mutation abolishes DAG binding by destruc-
tion of the C1 domain secondary structure (Quest et al.,
1994; Rhee et al., 2002), we expected that the autoinhibi-
toryintramolecular interactions within a2-chimaerinwould
be abrogated as well, rendering its tertiary structure an
open conformation. Indeed, the C1 domain mutant a2-
chimaerinH206Kshowed a robust increase in RacGAP
activity (Figure 6, bars 5 versus 7), causing further reduc-
tions in RacGTP levels as compared to WT a2-chimaerin.
The coexpression of EphA4 did not change the basal
RacGAP activities of a1- and WT a2-chimaerin (Figure 6,
bars 11 versus 12 and 7 versus 8). Similarly, influence of
alone (Figure 6, bars 15 versus 1). When Nck2 and EphA4
were coexpressed in the absence or presence of a2-
chimaerin, increases in RacGTP levels were observed
(Figure 6, bars 16 versus 15 and 10 versus 8). On the other
hand, the enhancement of Rac activity upon Nck2/EphA4
expression was completely reverted by the presence of
a1-chimaerin (Figure 6, bars 16 versus 14), which does
not interact with Nck2 (Figure 5C and Figure S5C).
With the limitations of this assay, which does not yield
equal expression levels of all three proteins tested, we
could not clarify the consequences of a concerted
EphA4/a2-chimaerin/Nck2 action on Rac. Three explana-
tions of the data set are plausible: (1) in the presence of
activated EphA4, Nck2 both inhibits the RacGAP activity
of a2-chimaerin and increases the GTP load of Rac; (2)
a2-chimaerin and EphA4-activated Nck2 regulate Rac ac-
tivity in an antagonistic manner; (3) a2-chimaerin and
Nck2 balance the activity of Rac in a cooperative and
EphA4-dependent manner. However, with this assay we
found that full RacGAP activity of a2-chimaerin seems to
depend on the unfolding of its tertiary structure. This
cannot be mediated by EphA4 and/or Nck binding or by
tyrosine phosphorylation, but it may occur in response
to additional DAG binding to the C1 domain of a2-chi-
maerin, which can be triggered by EphA4-dependent
activation of phospholipase Cg1 (Zhou et al., 2007).
The present study identifies the RacGAP a-chimaerin as
an essential component of the EphA4 receptor signaling
pathway that mediates axon guidance and growth cone
collapse. This conclusion is supported by five comple-
mentary lines of evidence: (1) both a1- and a2-chimaerin
interact with EphA4 and are phosphorylated in an EphA4-
dependent manner; (2) a-chimaerin-deficient cortical
neurons show a partially inhibited growth cone collapse
response to ephrins; (3) a-chimaerin mutant mice exhibit
developmental abnormalities that are observed in EphA4
ened dorsal funiculi in the spinal cord and abnormal
midline recrossing of CST axons; (4) like ephrinB3?/?,
EphA4?/?, and various EphA4 mutant lines with perturbed
EphA4 receptor forward signaling, deletion of a-chimaer-
ins in mice causes an abnormal rabbit-like hopping gait,
ing of axons in the lumbar spinal cord that participate in
Figure 6. RacGAP Activity of a-Chimaerins in the Presence or
Absence of EphA4 and Nck2
RacGTP levels were measured in 293FT cells after transfection with
indicated expression constructs. Transfected cells were lysed, and
endogenous GTP-loaded Rac was affinity precipitated with the p21
binding domain of PAK1 as a GST fusion protein.
system. RacGTP levels were normalized to total Rac levels in each
case. Values are normalized RacGTP levels expressed as percent of
control (EGFP alone or EGFP and EphA4) and represent the mean of
three independent experiments. Error bars indicate SEM. *p < 0.05,
**p <0.01, versus EGFP controls;and#p< 0.05,##p< 0.01 for compar-
followed by Tukey’s HSD test).
levels in experiments without (B) or with EphA4 (C). EGFP, enhanced
(RacGAP inactive mutant); a2-chimH206K, a2-chimaerinH206K(DAG-
binding-deficient mutant); a2-chim, a2-chimaerin; a1-chim, a1-
chimaerin; Nck2, Nck2-HA; PD, pull down.
Neuron 55, 756–767, September 6, 2007 ª2007 Elsevier Inc. 763
Chimaerins Mediate EphA4-Dependent Axon Guidance
the hindlimb motor CPG; (5) like in EphA4?/?and
ephrinB3?/?mice, the electrophysiological activity of the
CPG for hindlimb locomotion
Given that the CST and neuronal networks in the spinal
cord develop rather early in mice, the phenotypes we
observed here are most likely due to a2-chimaerin loss
because it is most strongly expressed during embryonic
development when a1-chimaerin expression is low (Hall
et al., 2001). The observation that ephrinB3-mediated
growthcone collapseisdeficientinthe absenceof a2-chi-
maerin (Figure 4) directly links a2-chimaerin function to
Eph receptor forward signaling and indicates that the
phenotypic changes described above are mostly due to
misled axon guidance decisions of EphA4-presenting
growth cones lacking a2-chimaerin during early network
formation. Because a1-chimaerin expression is low in
embryonic stages and increases postnatally, its interac-
tion with EphA4 is most likely relevant at later stages of
brain development and in the adult brain. Possible func-
tions may include the pruning of spines and dendrites or
Eph-mediated synaptic plasticity (Yamaguchi and Pas-
quale, 2004; Grunwald et al., 2004), but additional studies
are required to test this hypothesis in detail.
Several lines of evidence indicate that a2-chimaerin
may act downstream of multiple Eph receptors. Our initial
YTH screendemonstrated thata2-chimaerin bindsEphB1
(Figure 1), and our growth cone collapse assays showed
that a2-chimaerin loss interferes with ephrinB3- and eph-
rinA1-induced growth cone collapse (Figure 4). Further
support isprovided by
chimaerin?/?mice display a stronger phenotypic change
with severely desynchronized left-right ventral root activi-
ties as compared to EphA4?/?mice, which show a shift
from alternating to synchronized left-right ventral root
activities of the hindlimb locomotor CPG. This finding
may reflect additional mistargeting of CIN projections
within the CPG of a2-chimaerin?/?mice. EphA2 and
some EphB receptors are expressed by CINs (Imondi
et al., 2000; Brittis et al., 2002), and the EphB1/2/3 triple
knockout exhibits CIN projections that follow an incorrect
path on the contralateral side of the spinal cord (Kadison
et al., 2006). a2-Chimaerin may be involved in the path-
finding of growth cones presenting Eph receptors other
than EphA4, and subtle mistargeting of CIN projections,
which may have evaded detection in our tracing analysis,
may contribute to the severe phenotypic changes seen in
the CPG of a-chimaerin?/?mice. Similarly, a2-chimaerin
may also act in semaphorin-mediated guidance of CIN
projections, as it was shown to be involved in the
Sema3A-induced growth cone collapses of DRG neurons
(Brown et al., 2004).
Given that EphA4 functionally interacts with a2-chi-
maerin but does not alter its RacGAP activity, our data
are best compatible with a role of a2-chimaerin in control-
ling Rac activity at the site of activated EphA4. Here,
a2-chimaerin may either locally exert its basal RacGAP
activity or experience additional activation through DAG
binding to its C1 domain, resulting in a stronger but spa-
tially confined downregulation of Rac. However, the pro-
duction of DAG by PLCs in the ephrin-Eph-mediated
signaling steps that lead to growth cone collapse has
not been demonstrated yet. Furthermore, a2-chimaerin
a balanced activation of Rac. Indeed, Racundergoes bidi-
rectional activity changes during ephrin-induced growth
cone collapse. After cell-cell contact via ephrin-Eph bind-
ing, Rac activity initially decreases, but is quickly restored
(Wahl et al., 2000; Jurney et al., 2002).
ization, which terminates cell-cell contact and is essential
for growth cone retraction (Jurney et al., 2002; Zimmer
et al., 2003; Marston et al., 2003). In retinal ganglion cells,
RhoGEFs of the Vav family seem to beresponsible for Rac
ephrinA1-EphA endocytosis (Cowan et al., 2005). In con-
trast, the RacGAP a2-chimaerin appears to have no influ-
ence on receptor-ligand endocytosis, as cultured cortical
neurons from a-chimaerin?/?mice do not show changes
in absolute EphA4 levels, in the degree of EphA4 phos-
at the plasma membrane (data not shown).
In any case, the a-chimaerin mouse mutants described
here show that axon guidance decisions and subsequent
neuronal circuit formation require a delicate balance of
RhoGTPase activity. That the role of a2-chimaerin in this
process is not only intimately related to that of EphA4
but also to that of Ncks is strongly supported by the fact
that Nck mutant mice show defects in EphA4 forward
signaling that are very similar to those described here
(J.P. Fawcett and T. Pawson, personal communication).
YTH screens were performed essentially as described (Betz et al.,
1997;ParkandYun,2001).Abaitvector encoding theLexADNA-bind-
ing domain in frame with the SH2 domain of a2-chimaerin fused to the
kinase domain of Src (pLexN-a2-chimaerin-SH2-src-kinase) was used
to screen a P8 rat total-brain cDNA library in pVP16-3 (Vojtek et al.,
Antibodies, cDNA Constructs, and Reagents
Details are provided in the Supplemental Data.
Generation of a-Chimaerin?/?and a-Chimaerinmut/mutMice
Details are provided in the Supplemental Data.
Gait analyses on a-chimaerin?/?, a-chimaerinmut/mut, and WT litter-
mates were conducted according to previously published procedures
(Kullander et al., 2001a; Egea et al., 2005).
Corticospinal Tract Tracing
Tracing of the corticospinal tract with the anterograde tracer biotin
dextrane amine was performed according to previously published
procedures (Egea et al., 2005).
764 Neuron 55, 756–767, September 6, 2007 ª2007 Elsevier Inc.
Chimaerins Mediate EphA4-Dependent Axon Guidance
Spinal Cord Electrophysiology
The functional output of the locomotor CPG was studied using extra-
cellular electrophysiology on an in vitro preparation of the spinal
cord. Neonatal mice at postnatal day P0–P3 were anesthetized with
isofluran (Baxter, Sweden) and decapitated. During dissection, mice
(in mM: NaCl, 128; KCl, 4.69; NaHCO3, 25; KH2PO4, 1.18; MgSO4, 3.5;
CaCl2, 0.25; D-glucose, 22). Mice were crucified, ventral side up, in
a dissection chamber, eviscerated, and the spinal cord was exposed
by ventral laminectomy. The spinal cord was dissected out to preserve
the integrity of the ventral roots, pinned down in a perfusion chamber,
and perfused with oxygenated aCSF (in mM: NaCl, 128; KCl, 4.69;
NaHCO3, 25; KH2PO4, 1.18; MgSO4, 1.25; CaCl2, 2.5; D-glucose, 22)
at room temperature. Suction electrodes were attached to left and
right lumbar L2 and L5 ventral roots, and the spinal cord was equili-
brated at least 30 min before application of neurotransmitter chemi-
cals. A combination of N-methyl-D-asparate (NMDA, 5–10 mM), sero-
tonin (5-HT, 5–10 mM), and dopamine (50 mM) were added to the
perfusing aCSF to induce stable locomotor-like output. All chemicals
were obtained from Sigma. The signals were amplified 10,000 times
and band-pass filtered 10 Hz to 5 kHz (Model 1700, A-M systems).
The signals were A/D converted (Digidata 1322A, Axon Instruments)
before being recorded on a PC (Axoscope 9.0.2.05) for later off-line
analysis. Data analysis of the recorded signals was performed as
described (Kjaerulff and Kiehn, 1996); the onset of each burst was
determined directly from recordings without rectifying and filtering
the signals. The intrasegmental left-right alternations were analyzed
in L2 roots, and the intersegmental L2/L5 alternations were analyzed
oneither therightorleftsideineachanimal. Onlytracesshowing asta-
ble rhythmic activity were analyzed for coordination coupling. Vector
points were derived from 15 locomotor cycles with a five cycle interval
in between spanning 75 locomotor cycles. Each point represents the
endpoint of a vector, the direction given by the mean phase value
and the length given by the focus of phase values around the mean.
A phase value of 0.5 reflects alternating activity, and 0.0 reflects
CIN Tracing Experiment
Spinal cords were prepared from neonatal mice at P0–P3. Mice were
anesthetized with isofluran (Baxter, Sweden) and decapitated. During
dissection, neonatal mice were kept in ice-cold oxygenated (95% O2/
5% CO2) artificial cerebrospinal fluid (aCSF [in mM]: 128 NaCl, 4.69
KCl, 2.5 CaCl2, 1.25 MgSO4, 1.18 KH2PO4, 25 NaHCO3, 22 glucose).
Mice were pinned down in a dissection tray and eviscerated. The ver-
tebral column was cleaned of adherent tissue and opened ventrally.
After the spinal roots were cut, the spinal cord was removed and
pinned down. Fluorescent dextran-amines 3000 Da rhodamine-
dextran-amine (RDA) and 3000 Da fluorescein-dextran-amine (FDA)
(Invitrogen, Sweden) were used for retrograde tracing of commissural
were dissolved in a small drop of distilled water to allow collection of
minute quantities of tracer on the tip of a fine tungsten needle. Two
different application protocols were used (Figures 3G and 3H). First,
intrasegmental tracing with FDA was used to show local projecting
commissural interneurons. A sagittal cut was made parallel to the ven-
tral commissure along the entire L2 segment, and FDA tracer was im-
mediately applied. Second, intersegmental tracing with FDA and RDA
was used to locate ascending (aCINs), descending (dCINs), and bifur-
cating commissural interneurons (adCINs). Therefore, thoracic root 13
(T13) and lumbar root 3 (L3) were identified. With a 20 min interval,
a small cut was made across the ventral and ventrolateral funiculi im-
mediately caudal to each of these roots. A different tracer was applied
into each cut. For both application protocols, excess dye that diffused
away from the cut was removed by using a small pipette to avoid con-
tamination of adjacent areas. Preparations were then incubated in
a dark box in oxygenated aCSF (pH 7.4) at room temperature for 12–
16 hr. Samples were fixed in 4% paraformaldehyde (PFA) in 0.1 M
phosphate-buffered saline (PBS) (pH 7.4) and stored dark at 4?C for
1 week. Spinal cords were washed in PBS and embedded in 4% aga-
rose in 13TAE. 60 mm thick transverse sections were cut on a vibra-
tome (Leica), collected on glass slides, and mounted with 2.5%
DABCO (Sigma, Sweden) in glycerol containing 0.1 M Tris (pH 8.6).
Slides were stored in the dark at –20?C. Sections were viewed in an
Olympus BX61WI microscope (Olympus) with separate filters to view
RDA and FDA. The application sites were identified and photographs
with a 43 objective were taken of all sections of the L2 level (intraseg-
mental tracing) or between the two application sites (intersegmental
tracing). Higher-magnification photographs were taken with a 103
objective using the OptiGrid Grid Scan Confocal Unit (Qioptiq,
Rochester USA) and Volocity software (Improvision, Lexington USA).
Primary Cell Culture and Growth Cone Collapse Assay
The preparation of mouse cortical cultures as well as the stimulation of
growth cone collapses and their analysis were performed according to
previously published procedures (Kullander et al., 2001a; Egea et al.,
293FT Cell Culture and Cell Transfections
293FT cells (Invitrogen) were cultured in Dulbecco’s modified Eagle’s
medium (DMEM; Invitrogen) supplemented with 10% heat-inactivated
fetal calf serum (FCS; PAA Laboratories), MEM nonessential amino
acids, penicillin, streptomycin, and G418. For RacGAP assays, cells
were plated onto tissue culture dishes coated with poly-L-lysine
(Sigma). 293FT cells were transfected with different plasmids using
lipofectamine 2000 (Invitrogen).
Cosedimentation and Coimmunoprecipitation Experiments
293FT cells were transfected with the appropriate expression con-
structs and kept in culture after transfection for 24 hr before the cells
were lysed. The cells were harvested on ice with lysis buffer (20 mM
Tris/HCl [pH 7.5], 120 mM NaCl, 10% glycerol, 1% triton X-100, com-
plete proteinase inhibitors [Roche], 20 mM NaF, 10 mM Na-pyrophos-
phate, and 2 mM Na-orthovanadate), the proteins were solubilized by
agitating the cell lysates at 4?C for 30 min on a rotating wheel, and the
supernatants were cleared bycentrifugation at 100,000 3 g (4?C, 1 hr).
The cleared supernatants were used for either cosedimentaion exper-
iments with bacterially expressed GST fusion proteins or incubated
with appropriate antibodies for immunoprecipitation experiments. In
immunoprecipitation experiments, antibodies prebound to Protein A
Sepharose (Amersham Biosciences) or conjugated to agarose were
added to the lysates and incubated for 2 hr at 4?C on a rotating wheel.
The beads were then washed three times with washing buffer (lysis
buffer, but 0.5% triton X-100), and subjected to SDS-PAGE and west-
ern blotting. For cosedimentation assays, recombinant GST fusion
proteins were synthesized in E. coli (Guan and Dixon, 1991), purified
on glutathione sepharose (Amersham Bioscience), and used, immobi-
lized on the resin, for cosedimentation assays with proteins expressed
in 293FT cells. For this purpose, the cell lysates and 25 nmol of the se-
pharose bound GST fusion proteins were incubated for 2 hr at 4?C on
a rotator, then washed five times with washing buffer, and analyzed by
SDS-PAGE and western blotting.
RacGTP Pull-Down Assays
We used the p21 binding domain of Pak1 (aa 67–150) to isolate GTP-
loaded Rac from transfected fibroblasts (Benard et al., 1999). For this
purpose, 293FT cells were transfected with chimaerin expression con-
structs in pcDNA-IRES-EGFP, and with pcDNA3-EphA4 WT or pRK5-
Nck2-HA. For the expression of the EGFP control, the pcDNA-IRES-
EGFP vector was used. 24 hr after transfection, the cells were
harvested on ice with 1 ml ice-cold lysis buffer (20 mM Tris [pH 7.5],
150 mM NaCl, 5 mM MgCl2, 1 mM DTT, 10% glycerol, 0.5% nonidet
P40, 5 mM b-glycerophosphate, 1 mM PMSF, 0.5 mg/ml leupeptin,
1 mg/ml aprotinin). The proteins were then solubilized for 30 min by
incubation at 4?C on a rotating wheel, and the supernatants were
Neuron 55, 756–767, September 6, 2007 ª2007 Elsevier Inc. 765
Chimaerins Mediate EphA4-Dependent Axon Guidance
cleared by centrifugation at 16,000 3 g (4?C, 10 min). For the determi-
nation of the total amount of Rac and for monitoring the expression of
the transfected constructs, 100 ml of the cleared supernatents were
mixed with SDS-PAGE loading buffer, boiled, and analyzed by
SDS-PAGE and western blotting. To pull down GTP-loaded Rac, the
supernatants were supplemented with 12.5 mg GST-Pak1 fusion
protein immobilized on glutathione sepharose, incubated for 1 hr at
4?C on a rotating wheel, and then washed five times with ice-cold lysis
buffer using a cooled centrifuge. The beads were boiled in SDS-PAGE
loading buffer, separated on 15% SDS-polyacrylamide gels, and
transferred to a PVDF membrane for western blot analysis with an
anti-Rac antibody. The amount of total Rac in the supernatants and
of GTP-loaded Rac bound to GST-Pak1 was determined with an
Odyssey image reader (Li-Cor). The RacGTP/total Rac ratio of the
EGFP control was set as 100%, and the ratios obtained from the other
constructs were expressed as percentages relative to the control.
The Supplemental Data for this article can be found online at http://
We thank S. Kilinc, I. Thanha ¨user, D. Schwerdtfeger, and the staff of
the Animal Facility at the Max Planck Institute of Experimental Medi-
cine for technical assistance; N. Ghyselinck and J.M. Garnier for pro-
viding the pFlexHR gene targeting vector and giving advice on its use;
and T. Pawson and J.P. Fawcett for Nck expression constructs. This
work was supported by the Max Planck Society and the German Re-
search Foundation (Grant SFB406/A1 to A.B. and N.B.) and by grants
from the Swedish Medical Research Council (K2004-32P-15230,
gia, the foundations of Knut and Alice Wallenberg, A˚ke Wiberg, Mag-
nus Bergwall, A˚hle ´n, Hedlund, and Uppsala University (K.K.).
Received: June 25, 2007
Revised: July 20, 2007
Accepted: July 30, 2007
Published: September 5, 2007
Benard, V., Bohl, B.P., and Bokoch, G.M. (1999). Characterization of
rac and cdc42 activation in chemoattractant-stimulated human
neutrophils using a novel assay for active GTPases. J. Biol. Chem.
Betz, A., Okamoto, M., Benseler, F., and Brose, N. (1997). Direct inter-
action of the rat unc-13 homologue Munc13-1 with the N terminus of
syntaxin. J. Biol. Chem. 272, 2520–2526.
Binns, K.L., Taylor, P.P., Sicheri, F., Pawson, T., and Holland, S.J.
(2000). Phosphorylation of tyrosine residues in the kinase domain
and juxtamembrane region regulates the biological and catalytic activ-
ities of Eph receptors. Mol. Cell. Biol. 20, 4791–4805.
Bisson, N., Poitras, L., Mikryukov, A., Tremblay, M., and Moss, T.
(2007).EphA4 signaling regulates blastomere adhesion in theXenopus
embryo by recruiting Pak1 to suppress Cdc42 function. Mol. Biol. Cell
provides a mechanism for localized regulation at an intermediate tar-
get. Cell 110, 223–235.
Brown, M., Jacobs, T., Eickholt, B., Ferrari, G., Teo, M., Monfries, C.,
Qi, R.Z., Leung, T., Lim, L., and Hall, C. (2004). Alpha2-chimaerin,
cyclin-dependent Kinase 5/p35, and its target collapsin response me-
diator protein-2 are essential components in semaphorin 3A-induced
growth-cone collapse. J. Neurosci. 24, 8994–9004.
Butt, S.J., Lundfald, L., and Kiehn, O. (2005). EphA4 defines a class of
excitatory locomotor-related interneurons. Proc. Natl. Acad. Sci. USA
Buttery, P., Beg, A.A., Chih, B., Broder, A., Mason, C.A., and Scheif-
fele, P. (2006). The diacylglycerol-binding protein alpha1-chimaerin
regulates dendritic morphology. Proc. Natl. Acad. Sci. USA 103,
Caloca, M.J., Wang, H., and Kazanietz, M.G. (2003). Characterization
of the Rac-GAP (Rac-GTPase-activating protein) activity of b2-chi-
maerin, a ’non-protein kinase C’ phorbol ester receptor. Biochem.
J. 375, 313–321.
Canagarajah, B., Leskow, F.C., Ho, J.Y., Mischak, H., Saidi, L.F.,
Kazanietz, M.G., and Hurley, J.H. (2004). Structural mechanism for
lipid activation of the Rac-specific GAP, beta2-chimaerin. Cell 119,
Cowan, C.W., Shao, Y.R., Sahin, M., Shamah, S.M., Lin, M.Z., Greer,
P.L., Gao, S., Griffith, E.C., Brugge, J.S., and Greenberg, M.E.
(2005). Vav family GEFs link activated Ephs to endocytosis and axon
guidance. Neuron 46, 205–217.
Dottori, M., Hartley, L., Galea, M., Paxinos, G., Polizzotto, M., Kilpa-
trick, T., Bartlett, P.F., Murphy, M., Kontgen, F., and Boyd, A.W.
(1998). EphA4 (Sek1) receptor tyrosine kinase is required for the devel-
opment of the corticospinal tract. Proc. Natl. Acad. Sci. USA 95,
Egea, J.,and Klein, R.(2007). Bidirectional Eph-ephrin signaling during
axon guidance. Trends Cell Biol. 17, 230–238.
Egea, J., Nissen, U.V., Dufour, A., Sahin, M., Greer, P., Kullander, K.,
Mrsic-Flogel, T.D., Greenberg, M.E., Kiehn, O., Vanderhaeghen, P.,
and Klein, R. (2005). Regulation of EphA 4 kinase activity is required
for a subset of axon guidance decisions suggesting a key role for re-
ceptor clustering in Eph function. Neuron 47, 515–528.
Glover, J.C. (1995). Retrograde and anterograde axonal tracing with
fluorescent dextran-amines in the embryonic nervous system. Neuro-
sci. Prot. 30, 1–13.
Greferath, U., Canty, A.J., Messenger, J., and Murphy, M. (2002).
Developmental expression of EphA4-tyrosine kinase receptor in the
mouse brain and spinal cord. Mech. Dev. 119 (Suppl 1), S231–S238.
Grunwald, I.C., Korte, M., Adelmann, G., Plueck, A., Kullander, K.,
Adams, R.H., Frotscher, M., Bonhoeffer, T., and Klein, R. (2004). Hip-
pocampal plasticity requires postsynaptic ephrinBs. Nat. Neurosci. 7,
Guan, K.L., and Dixon, J.E. (1991). Eukaryotic proteins expressed in
Escherichia coli: an improved thrombin cleavage and purification
procedure of fusion proteins with glutathione S-transferase. Anal. Bio-
chem. 192, 262–267.
Hall, C., Michael, G.J., Cann, N., Ferrari, G., Teo, M., Jacobs, T., Mon-
fries, C., and Lim, L. (2001). alpha2-chimaerin, a Cdc42/Rac1 regula-
tor, is selectively expressed in the rat embryonic nervous system
and is involved in neuritogenesis in N1E-115 neuroblastoma cells. J.
Neurosci. 21, 5191–5202.
Holland, S.J., Gale, N.W., Gish, G.D., Roth, R.A., Songyang, Z., Cant-
ley, L.C., Henkemeyer, M., Yancopoulos, G.D., and Pawson, T. (1997).
Juxtamembrane tyrosine residues couple the Eph family receptor
EphB2/Nuk to specific SH2 domain proteins in neuronal cells. EMBO
J. 16, 3877–3888.
Imondi, R., Wideman, C., and Kaprielian, Z. (2000). Complementary
expression of transmembrane ephrins and their receptors in the
mouse spinal cord: a possible role in constraining the orientation of
longitudinally projecting axons. Development 127, 1397–1410.
Jurney, W.M., Gallo, G., Letourneau, P.C., and McLoon, S.C. (2002).
Rac1-mediated endocytosis during ephrin-A2- and semaphorin 3A-
induced growth cone collapse. J. Neurosci. 22, 6019–6028.
766 Neuron 55, 756–767, September 6, 2007 ª2007 Elsevier Inc.
Chimaerins Mediate EphA4-Dependent Axon Guidance
Kadison, S.R., Makinen, T., Klein, R., Henkemeyer, M., and Kaprielian, Download full-text
ventral midline of the embryonic mouse spinal cord. J. Neurosci. 26,
Kiehn, O. (2006). Locomotor circuits in the mammalian spinal cord.
Annu. Rev. Neurosci. 29, 279–306.
Kjaerulff, O., and Kiehn, O. (1996). Distribution of networks generating
and coordinating locomotor activity in the neonatal rat spinal cord
in vitro: a lesion study. J. Neurosci. 16, 5777–5794.
Kullander, K., Croll, S.D., Zimmer, M., Pan, L., McClain, J., Hughes, V.,
Zabski, S., DeChiara, T.M., Klein, R., Yancopoulos, G.D., and Gale,
N.W. (2001a). Ephrin-B3 is the midline barrier that prevents corticospi-
nal tract axons from recrossing, allowing for unilateral motor control.
Genes Dev. 15, 877–888.
Kullander, K., Mather, N.K., Diella, F., Dottori, M., Boyd, A.W., and
Klein, R. (2001b). Kinase-dependent and kinase-independent func-
tions of EphA4 receptors in major axon tract formation in vivo. Neuron
Kullander, K., Butt, S.J., Lebret, J.M., Lundfald, L., Restrepo, C.E.,
Rydstrom, A., Klein, R., and Kiehn, O. (2003). Role of EphA4 and Eph-
Leskow, F.C., Holloway, B.A., Wang, H., Mullins, M.C., and Kazanietz,
M.G. (2006). The zebrafish homologue of mammalian chimerin Rac-
GAPs is implicated in epiboly progression during development. Proc.
Natl. Acad. Sci. USA 103, 5373–5378.
Marston, D.J., Dickinson, S., and Nobes, C.D. (2003). Rac-dependent
Nat. Cell Biol. 5, 879–888.
Murai, K.K., and Pasquale, E.B. (2005). New exchanges in eph-depen-
dent growth cone dynamics. Neuron 46, 161–163.
Park, D., and Yun, Y. (2001). Tyrosine phosphorylation-dependent
yeast two-hybrid system for the identification of the SH2 domain-bind-
ing proteins. Mol. Cells 12, 244–249.
Quest, A.F., Bardes, E.S., and Bell, R.M. (1994). A phorbol ester bind-
ing domain of protein kinase Cg. Deletion analysis of the Cys2 domain
defines a minimal 43-amino acid peptide. J. Biol. Chem. 269, 2961–
Rhee, J.S., Betz, A., Pyott, S., Reim, K., Varoqueaux, F., Augustin, I.,
Hesse, D., Su ¨dhof, T.C., Takahashi, M., Rosenmund, C., and Brose,
N. (2002). Beta phorbol ester- and diacylglycerol-induced augmenta-
tion of transmitter release is mediated by Munc13s and not by PKCs.
Cell 108, 121–133.
Sahin,M.,Greer, P.L.,Lin,M.Z.,Poucher, H.,Eberhart, J.,Schmidt,S.,
Wright, T.M., Shamah, S.M., O’Connell, S., Cowan, C.W., et al. (2005).
Eph-dependent tyrosine phosphorylation of ephexin1 modulates
growth cone collapse. Neuron 46, 191–204.
Stein, E., Huynh-Do, U., Lane, A.A., Cerretti, D.P., and Daniel, T.O.
(1998). Nck recruitment to Eph receptor, EphB1/ELK, couples ligand
activation to c-Jun kinase. J. Biol. Chem. 273, 1303–1308.
Stokke, M.F., Nissen, U.V., Glover, J.C., and Kiehn, O. (2002). Projec-
tion patterns of commissural interneurons in the lumbar spinal cord of
the neonatal rat. J. Comp. Neurol. 446, 349–359.
Van de Ven, T.J., VanDongen, H.M., and VanDongen, A.M. (2005). The
nonkinase phorbol ester receptor alpha 1-chimerin binds the NMDA
sci. 25, 9488–9496.
Vojtek, A.B., Hollenberg, S.M., and Cooper, J.A. (1993). Mammalian
Ras interacts directly with the serine/threonine kinase Raf. Cell 74,
Wahl, S., Barth, H., Ciossek, T., Aktories, K., and Mueller, B.K. (2000).
Ephrin-A5 induces collapse of growth cones by activating Rho and
Rho kinase. J. Cell Biol. 149, 263–270.
Wang, H., Yang, C., Leskow, F.C., Sun, J., Canagarajah, B., Hurley,
J.H., and Kazanietz, M.G. (2006). Phospholipase Cg/diacylglycerol-
dependent activation of b2-chimaerin restricts EGF-induced Rac sig-
naling. EMBO J. 25, 2062–2074.
Wells, C.D., Fawcett, J.P., Traweger, A., Yamanaka, Y., Goudreault,
M., Elder, K., Kulkarni, S., Gish, G., Virag, C., Lim, C., et al. (2006). A
Rich1/Amot complex regulates the Cdc42 GTPase and apical-polarity
proteins in epithelial cells. Cell 125, 535–548.
Yamaguchi, Y., and Pasquale, E.B. (2004). Eph receptors in the adult
brain. Curr. Opin. Neurobiol. 14, 288–296.
Yang, C., and Kazanietz, M.G. (2003). Divergence and complexities in
DAG signaling: looking beyond PKC. Trends Pharmacol. Sci. 24, 602–
Yokoyama, N., Romero, M.I., Cowan, C.A., Galvan, P., Helmbacher,
naling mediated by ephrin-B3 prevents contralateral corticospinal
axons from recrossing the spinal cord midline. Neuron 29, 85–97.
Zhou, L., Martinez, S.J., Haber, M., Jones, E.V., Bouvier, D., Doucet,
G., Corera, A.T., Fon, E.A., Zisch, A.H., and Murai, K.K. (2007).
EphA4 signaling regulates phospholipase Cgamma1 activation, cofilin
membrane association, and dendritic spine morphology. J. Neurosci.
Zimmer, M., Palmer, A., Ko ¨hler, J., and Klein, R. (2003). EphB-ephrinB
bi-directional endocytosis terminatesadhesionallowing contactmedi-
ated repulsion. Nat. Cell Biol. 5, 869–878.
Zisch, A.H., Pazzagli, C., Freeman, A.L., Schneller, M., Hadman, M.,
Smith, J.W., Ruoslahti, E., and Pasquale, E.B. (2000). Replacing two
binding of SH2 domains without abrogating kinase activity and biolog-
ical responses. Oncogene 19, 177–187.
Neuron 55, 756–767, September 6, 2007 ª2007 Elsevier Inc. 767
Chimaerins Mediate EphA4-Dependent Axon Guidance