Current Biology 24, 2355–2365, October 20, 2014 ª2014 Elsevier Ltd All rights reserved

http://dx.doi.org/10.1016/j.cub.2014.08.028

Article EphA4 Receptor Shedding Regulates Spinal Motor Axon Guidance Graziana Gatto,1 Daniel Morales,2,3 Artur Kania,2,3,4,5 and Ru¨diger Klein1,6,* 1Department of Molecules-Signaling-Development, Max Planck Institute of Neurobiology, Am Klopferspitz 18, 82152 Martinsried, Germany 2Institut de Recherches Cliniques de Montre ´ al, Montre´al, QC H2W 1R7, Canada 3Integrated Program in Neuroscience, McGill University, Montre´al, QC H3A 2B4, Canada 4Division of Experimental Medicine, Department of Anatomy and Cell Biology and Department of Biology, McGill University, Montre´al, QC H3A 2B2, Canada 5Faculte ´ de Me´decine, Universite´ de Montre´al, Montre´al, QC H3C 3J7, Canada 6Munich Cluster for Systems Neurology (SyNergy), 80336 Munich, Germany

Summary Background: Proteolytic processing of axon guidance receptors modulates their expression and functions. Contact repulsion by membrane-associated ephrins and Eph receptors was proposed to be facilitated by ectodomain cleavage, but whether this phenomenon is required for axon guidance in vivo is unknown. Results: In support of established models, we find that cleavage of EphA4 promotes cell-cell and growth cone-cell detachment in vitro. Unexpectedly, however, a cleavage resistant isoform of EphA4 is as effective as a wild-type EphA4 in redirecting motor axons in limbs. Mice in which EphA4 cleavage is genetically abolished have motor axon guidance defects, suggesting an important role of EphA4 cleavage in nonneuronal tissues such as the limb mesenchyme target of spinal motor neurons. Indeed, we find that blocking EphA4 cleavage increases expression of full-length EphA4 in limb mesenchyme, which—via cis-attenuation—apparently reduces the effective concentration of ephrinAs capable of triggering EphA4 forward signaling in the motor axons. Conclusions: We propose that EphA4 cleavage is required to establish the concentration differential of active ephrins in the target tissue that is required for proper axon guidance. Our study reveals a novel mechanism to regulate guidance decision at an intermediate target based on the modulation of ligand availability by the proteolytic processing of the receptor. Introduction The correct assembly of neural circuits relies on guidance decisions made by growing axons navigating the extracellular environment. A great variety of axon guidance decisions are controlled by a rather small number of ligand-receptor systems, suggesting the presence of mechanisms that amplify and diversify their outputs. Proteases have been shown to be important in regulating guidance receptor-ligand expression and signaling [1], by regulating the number of functional

*Correspondence: [email protected]

guidance receptors on the plasma membrane [2–4], releasing a functional ectodomain (ECD), which can act as a soluble axon guidance cue [5], and/or releasing a functional intracellular domain (ICD) with signaling characteristics distinct from the full-length receptor [6, 7]. The indiscriminate substrate preference of most metalloproteases and the absence of a standard ADAM-type protease consensus target sequence have in part hindered in vivo studies of how cleavage regulates the function of guidance receptors. Proteolytic cleavage is also an important component of cellcell repulsion mediated by Eph tyrosine kinase receptors and their membrane-tethered ephrin ligands [8–12]. Cleavage of Eph and/or ephrin ectodomains is a mechanism to break the molecular tether between two opposing cells and thereby allow cell detachment [9–11]. Proteolytic cleavage may therefore have an analogous or parallel function to bidirectional endocytosis of the Eph-ephrin complex to promote cell detachment [13, 14]. Moreover, cleavage of Eph receptors was shown to promote repulsive signaling in vitro [12] and to be regulated by ephrinB interaction [15]. However, the genetic evidence that Eph or ephrin cleavage directly causes a change in cellular responses in vivo has been absent. The navigation of spinal motor axons toward their specific target muscles in the developing limb is an excellent model system to study the molecular mechanisms of ephrin-Eph signaling in axon guidance. Axons of the lateral population of lateral motor column (LMC) neurons express EphA receptors and are repelled into the dorsal limb away from the ephrinA-enriched ventral limb via ephrinA:EphA4 forward signaling (Figure 1A) [16–18]. The fidelity of this event is enhanced by a parallel mechanism: EphA receptors, including EphA4, are expressed in the dorsal limb mesenchyme, where they act as ligands and attract lateral LMC axons through EphA:ephrinA reverse signaling (Figure 1B) [19–21]. Medial LMC neurons innervate the ventral limb and express EphA receptors, but they also coexpress high levels of ephrinA5 that cis- inhibits EphA receptors, thus making their axons ‘‘blind’’ to the ephrinA-rich ventral limb (Figure 1C) [20]. The contribution of EphA4 proteolytic cleavage to these guidance mechanisms has not been investigated. Here we show that EphA4 cleavage is widespread and temporally regulated during mouse embryonic development. By generating a cleavage resistant (CR) isoform of EphA4 we find that EphA4 cleavage promotes cell detachment in vitro. However, EphA4CR is as effective as EphA4WT in redirecting LMC axons in limb mesenchyme, demonstrating that cleavage of motor axon EphA4 is dispensable for forward signaling in vivo. Nevertheless, EphA4CR/CR mutant mice display defects in LMC axon projections. This phenotype correlates with higher levels of EphA4 and reduced abundance of free ephrinAs distribution in their limbs. We propose that EphA4 cleavage is a constitutive mechanism that regulates the availability of ephrin ligands in the target tissue. Results EphA4 Cleavage Is Temporally and Spatially Regulated during Development To begin investigating the role of EphA4 cleavage, we analyzed the extent of cleavage of endogenous EphA4 in cultured

Current Biology Vol 24 No 20 2356

A

B

C

D

F

H

E

G

I

Figure 1. EphA4 Cleavage Is Temporally and Spatially Regulated during Development (A–C) Schematics showing repulsive ephrinA:EphA forward signaling (A), attractive EphA:ephrinA reverse signaling (B), and inhibition of trans-interactions by cis-attenuation (C). (D) Representative western blots of total cell lysates (TCL) or conditioned media (CM) prepared from E16.5 cortical neurons (2 days in vitro, 2 DIV) from EphA4+/2 and EphA42/2 embryos as indicated. Samples were probed with antibodies against the extracellular domain (ECD) of EphA4 (EphA4ECD) or FLRT3, as control. EphA4 full-length runs at 120 kDa, EphA4-ECD (asterisk), and FLRT3 ECD at approximately 75 kDa [5]. (E) Representative western blots of TCLs and CM obtained from wild-type E16.5 cortical neurons kept in culture for 1 (1DIV) or 7 days (7DIV) as indicated. Blots with TCL samples were probed with antibodies against the intracellular domain of EphA4 (EphA4-ICD) and tubulin, whereas the blot with CM samples was probed with EphA4-ECD antibody. EphA4-ICD runs at around 50 kDa (asterisk). (F and G) Representative western blots of wild-type (+/+) and EphA42/2 spinal cord (F) and hindlimb (G) lysates prepared from several developmental stages probed with EphA4-ICD, EphA4-ECD, and tubulin antibodies. EphA4 full-length runs at 120 kDa, EphA4-ECD at 75 kDa (orange asterisk),

mouse cortical neurons. Conditioned media from EphA4+/2, but not EphA42/2 cultures, contained a peptide of 75 kDa that was detected with antibodies raised against the EphA4 ectodomain (referred to as EphA4-ECD) (Figure 1D). Total cell lysates contained EphA4 peptides of around 120 kDa, corresponding to full-length EphA4, and of about 50 kDa that were detected with antibodies raised against the EphA4 intracellular domain (referred to as EphA4-ICD) (Figure 1E and Figure S1A available online). The relative abundance of these peptide species changed with prolonged culture times: the levels of full-length EphA4 in cell lysates decreased after 7 days in culture, the levels of EphA4-ICD remained constant, and EphA4-ECD accumulated in the conditioned media, suggesting constitutive cleavage of EphA4 under these culture conditions (Figure 1E). We also estimated the rate of EphA4 cleavage in tissues in which EphA4 is known to have important functions [16, 17, 21]. In embryonic spinal cords, the abundance of EphA4ECD and EphA4-ICD were modest throughout development compared to full-length EphA4 (Figures 1F and 1H). In contrast, the amount of EphA4-ECD in the hindlimbs changed markedly throughout development and at times was more abundant than full-length EphA4 (Figures 1G and 1I). Conditional ablation of EphA4 from the spinal cord or the limb indicated that the bulk of EphA4 and EphA4-ECD were produced by hindlimb mesenchymal cells rather than by ingrowing axons (Figures S1B and S1C). These data suggest that EphA4 cleavage in hindlimb is under stronger temporal control than in spinal cord. EphA4 Cleavage Regulates Cell-Cell Detachment Since proteolytic cleavage of EphB2 and ephrins were previously shown to regulate repulsive guidance and cell-cell detachment, respectively [9, 12], we hypothesized that EphA4 cleavage would also regulate these processes. To test this, we generated a cleavage resistant isoform of EphA4 by narrowing down the region of cleavage to 15 amino acids (Figures S2A and S2B) and then by gradually converting them to the corresponding 13 amino acids in the related EphA3 receptor, which is not cleaved in the same manner (Figures 2A and 2B and Figure S2C). The resulting isoform was no longer cleaved into EphA4-ECD and EphA4-ICD and was therefore termed EphA4CR (CR, cleavage resistant). We confirmed that EphA4CR was expressed normally at the cell surface (Figures 2C–2F) and was phosphorylated to a similar degree as EphA4WT in response to preclustered ephrinA5 stimulation (Figures 2G and 2H). We first asked if EphA4 cleavage were required for the cellular collapse response typically seen when Eph forward signaling is induced in transfected cells [22]. To perform liveimaging experiments, we used mCherry-tagged EphA4, a

and EphA4-ICD below 50 kDa (purple asterisk). In spinal cord and to a lesser extent in hindlimb lysates probed with the EphA4-ICD antibody, there is a weak 120 kDa band in knockout controls that partially overlaps with the specific EphA4 band. This nonspecific band is absent from knockout samples probed with the EphA4-ECD antibodies. (H and I) Line graphs showing EphA4-ECD levels normalized to full-length EphA4 in spinal cord (H) and hindlimb (I) lysates from several developmental stages (mean 6 SEM). Lysates from one embryo at E11.5 and three to four embryos for all the other developmental stages were analyzed. For the different developmental stages, except E11.5 (not possible since n = 1), the ratio of EphA4-ECD and full-length EphA4 were compared using oneway ANOVA, followed by Bonferroni’s post hoc comparison test (*p < 0.05, ***p < 0.001). See also Figure S1.

EphA4 Cleavage Regulates Motor Axon Guidance 2357

A

Figure 2. Generation of a Cleavage Resistant Version of EphA4

B UT wt ∆4 ∆6 ∆8 CR Flag CM kDa - 120

EphA4 ICD

TCL

- 50

*

Tubulin EphA4 EphA3 EphA4CR surface D

C

total E

surface

EphA4WT

Fc

pEph kDa 120 -

EphA4 ICD 50 -

*

EphA4CR

ephrinA5 ephrinA5 + ++ +++ Fc + ++ +++

total

EphA4CR

H pospho/total EphA4

EphA4WT

G

F

0.4

*

0.3

* 0.2

* *

***

*

EphA4WT EphA4CR

0.1

0

Fc

ephrinA5 ephrinA5 ephrinA5 2 μg/ml 5 μg/ml 1 μg/ml

fusion protein previously shown not to impair forward signaling [22, 23]. EphA4CR-mCherry behaved similarly to EphA4CR-FLAG in being cleavage resistant (data not shown). During the first 2 hr, transfected cells were stimulated with control Fc, followed by an additional 2 hr with preclustered ephrinA5-Fc. In response to ephrinA5-Fc, upon increase in phosphotyrosine signaling, cells reduced their surface area (Figures 3A–3D; Figures S3A–S3D). EphA4WT-mCherry and EphA4CR-mCherry did not differ in the extent of collapse in response to ephrinA5-Fc (Figure 3E), suggesting that EphA4 cleavage does not regulate cell collapse induced by soluble ephrins. To test the role of EphA4 cleavage in cell-cell detachment, we challenged EphA4 expressing HeLa cells with SKN neuroblastoma cells, which express endogenous ephrinBs (T. Gaitanos and R.K., unpublished data). HeLa cells expressing EphA4WT-mCherry after the initial contact with SKN cells detached and moved away. This effect was mildly, yet significantly, reduced in HeLa cells expressing EphA4CR-mCherry (p < 0.05, Fisher’s exact test; Figures 3F–3H). Next we asked if under these conditions cells would compensate by enhancing the rate of trans-endocytosis, e.g., of EphA4 into ephrinB-expressing cells. SKN cells in contact with EphA4CR-expressing cells had significantly more EphA4containing vesicles than when in contact with EphA4WT-expressing cells (p = 0.0002, two-tailed Student’s t test; Figures 3I–3K). Both ‘‘donor’’ cells showed comparable EphA4 surface

(A) Schematic drawing showing the domain swapping (red arrow) between EphA4 and EphA3 to generate the EphA4CR (cleavage resistant) mutant. (B) Representative western blots of TCLs and CM obtained from HeLa cells transfected with EphA4WT and several EphA4 mutants obtained by site-direct mutagenesis. All the constructs carried a FLAG-tag in their N-terminal domain, so the shed ectodomains were detected by probing the conditioned media samples with a FLAG antibody. TCL samples were probed with EphA4-ICD and tubulin antibodies. (C–F) Representative images of HeLa cells transfected with EphA4WT (C and D) and EphA4CR (E and F) stained with anti-FLAG antibody before and after permeabilization to detect surface (C and E) and total (D and F) expression of EphA4, respectively. Quantification of surface versus total ratio did not reveal differences between the EphA4 isoforms (data not shown). Scale bar, 25 mm. (G) Representative western blots of TCLs prepared from HeLa cells transfected with EphA4WT and EphA4CR and stimulated as indicated. +, ++, and +++ represent stimulation with 1, 2, and 5 mg/ml recombinant protein, respectively. EphA4 phosphorylation was assessed using an anti-phosphoEphA antibody. Full-length protein and EphA4-ICD were detected using the EphA4ICD antibody. (H) Bar graph showing the levels of phosphoEphA4 normalized to the total EphA4WT or EphA4CR. Data represent the mean 6 SEM of four experiments. Statistical analysis was done using two-tailed Student’s t test (*p < 0.05, ***p < 0.001). No significant difference is observed between EphA4wt and EphA4CR. See also Figure S2.

expression (Figure S3E). As expected, EphA4 trans-endocytosis required the presence of the ephrin ligand binding domain in EphA4 (Figure S3F). These data suggest that abolishing EphA4 cleavage slows the process of cell-cell detachment, an effect that is accompanied by enhanced trans-endocytosis. EphA4 Cleavage Regulates Growth Cone-Cell Detachment To assess the role of EphA4 cleavage in axon guidance, we focused on spinal motor neurons, which are excellent models of ephrin-mediated axon guidance. We generated an EphA4CR knockin mouse by inserting the appropriate segment of the mouse EphA4CR cDNA in frame into exon 3 of the EphA4 gene locus (Figures S4A and S4B), a previously used strategy [24–26]. Western blot analysis indicated that the mutation abolished EphA4 cleavage in vivo: in EphA4CR/CR embryos, the EphA4-ICD and the EphA4-ECD were no longer detected (Figure 4A; data not shown). Moreover, the CR mutation did not affect EphA4 expression on lateral and medial LMC motor axons (Figures 4B–4D; Figures S4C–S4E). Next, we tested the response to ephrinAs of embryonic EphA4CR/CR spinal motor axons labeled with the transgenic marker Hb9-GFP [27]. The collapse responses of GFP+ spinal motor neurons to a range of soluble, preclustered ephrinA2/ A5-Fc concentrations were unchanged in EphA4CR/CR compared with controls (Figures 4E and 4F). Likewise, live imaging of cocultured motor neurons with HeLa cells expressing

Current Biology Vol 24 No 20 2358

A

E

Fc B

0’

60’

120’

ephrinA5

0’

60’

120’

EphA4CR

C

100

Area [%]

EphA4WT

120

80

**** **

60

*** ***

40 Fc

0’

60’

WT/Fc CR/Fc WT/ephrinA5 CR/ephrinA5

120’

D

20 0

ephrinA5

0’

60’

0 10 20 30 40 50 60 70 80 90 100 110 120

Time (minute)

120’

F

EphA4WT SKN cells F’

0’

24’

48’

72’

96’

0’

24’

48’

72’

96’

G

EphA4CR SKN cells

G’

% detachment

H

J 80 60 40

vesicles per cell

surface

CMB total surface CMB

*

20 0

I

total

EphA4WT

WT CR 6 5

***

K

4 3 2 1 0

WT CR

EphA4CR

Figure 3. EphA4 Cleavage Regulates Cell-Cell Detachment (A–D) Representative time-lapse images of HeLa cells, expressing equal levels of EphA4WT (A and B) or EphA4CR (C and D) stimulated with 1 mg/ml preclustered Fc (A and C) or ephrinA5-Fc (B and D) for 120 min. Scale bar, 25 mm. (legend continued on next page)

EphA4 Cleavage Regulates Motor Axon Guidance 2359

membrane-tethered ephrinA5-mCherry revealed normal rates and timing of growth cone collapse (Figure S4G–S4I). In contrast, in a growth cone-cell detachment assay, control axons retracted from the ephrinA5 source in 60% of the cases, whereas only 18% of EphA4CR/CR axons detached from the HeLa cells during the 2 hr imaging period (p < 0.01, Fisher’s exact test) (Figure 4G–4I0 ; Figure S4F). These results suggest that cleavage of EphA4 may be a mechanism for efficient detachment of motor axons from membrane-tethered ephrinAs. Cleavage of Neuronal EphA4 Is Dispensable for Axon Guidance In Vivo Next we tested whether impeding EphA4 cleavage might have a consequence on axonal forward ephrinA:EphA4 signaling in vivo. We coelectroporated either EphA4CR or EphA4WT and a myristoylated GFP plasmid or the GFP expression plasmid alone in HH st. 18 chicken embryos (Figures 4J–4L [20, 28]), reasoning that if cleavage were required for forward EphA4 signaling, then the redirection of LMC trajectory induced by EphA4WT overexpression [17, 18] should be attenuated by EphA4 cleavage site mutation. Coelectroporation of EphA4WT and GFP or EphA4CR and GFP expression plasmids led to similar levels of EphA4 overexpression (Figures S5C– S5F) and did not result in any significant changes in neuronal fate (Figures S5G–S5L). Following incubation until HH st. 25/ 26, as expected, in EphA4WT-GFP-coexpressing embryos, a significantly greater proportion of GFP+ axons entered the dorsal nerve branches when compared with GFP controls (Figures 4M and 4N; p < 0.005, Fisher’s exact test). Intriguingly, EphA4CR-expressing axons were also present in the dorsal limb in significantly greater numbers compared with GFP-expressing axons (Figures 4M and 4O; p < 0.05, Fisher’s exact test), showing that EphA4CR-expressing axons were as efficiently rerouted to the dorsal limb as EphA4WT-expressing axons (Figures 4N and 4O; p = 0.33, Fisher’s exact test). Together, these results demonstrate that cleavage of motor axon EphA4 is dispensable for forward signaling-mediated repulsion from ephrinAs in vivo. In further support of this model, the dorsal funiculus in the spinal cord whose formation depends on ephrin:EphA4 forward signaling [26, 29, 30] develops normally in EphA4CR/CR mice (Figures S4J and S4K). Since ephrinA:EphA interaction elicits bidirectional signaling, and since EphA4 cleavage may be a required process to initiate ephrin reverse signaling, we examined the formation of the anterior commissure (AC) of the forebrain, which depends on EphA4:ephrinB reverse signaling [26, 30, 31]. The

AC formation proceeds normally in EphA4CR/CR mice (Figure S4L), arguing that EphA4 cleavage is not required to initiate EphA4:ephrin reverse signaling. EphA4CR/CR Mice Have Abnormal Spinal Motor Axon Guidance In the chick overexpression experiment, EphA4CR-expressing motor axons navigate through wild-type limbs. In contrast, in EphA4CR/CR mice EphA4 cleavage is abolished in all tissues, allowing us to determine whether EphA4 cleavage in the limb might contribute to spinal motor axon guidance. We injected rhodamine dextran (RD) into the ventral or dorsal limbs of EphA4CR/CR and control mice, and we counted the numbers of neurons colabeled with RD and markers of medial (Islet1) or lateral LMC subpopulations (Lim1) [18]. In control embryos, ventral limb RD injections labeled medial LMC motor neurons expressing Islet1 and a very small fraction of lateral LMC motor neurons that express Lim1 (4% 6 2% of all RD+ LMC neurons; Figure 5B). In contrast, in EphA4CR/CR embryos the fraction of lateral LMC neurons labeled by ventral limb RD injections was increased more than 3-fold (14% 6 5%, p < 0.05, MannWhitney test; Figure 5C), indicating that lateral LMC axons were misguided ventrally (Figures 5A–5D). For comparison, ventral RD injections label approximately 30% of lateral LMC axons in EphA4 full knockout embryos [21]. In contrast, injections of RD into the dorsal limb labeled similar proportions of Islet1-expressing medial LMC neurons in wild-type, EphA4wt/CR, and EphA4CR/CR mice (Figures 5E–5H). These results indicate that EphA4 cleavage is required for normal spinal motor axon guidance and together with the chick overexpression of EphA4CR argue that EphA4 cleavage in the limb mesenchyme has a significant role in this process. EphA4 Cleavage Inhibition Increases EphA4 Protein Abundance To explore the role of EphA4 cleavage in hindlimb, we first quantified the overall EphA4 protein levels in embryos with the CR mutation by western blot. Full-length EphA4 was increased in EphA4CR/CR hindlimb, spinal cord, and forebrain compared with wild-type controls (Figures 6A–6D; Figure S6A). These effects were not due to the knockin strategy, since they were not observed in embryos with two copies of knocked-in wild-type EphA4 cDNA (EphA4wtKI/wtKI) [26] (Figures S6B– S6D). They were also not caused by changes in EphA4 mRNA levels (Figures S6E and S6F). These results suggest that EphA4 cleavage constitutes a posttranslational mechanism to regulate the levels of full-length EphA4.

(E) Line graph showing the area changes over time of HeLa cells transfected with EphA4WT-mCherry (blue and green lines, labeled as WT) and EphA4CRmCherry (red and purple lines, labeled as CR) in response to 1 mg/ml preclustered Fc (red and blue lines) or ephrinA5-Fc (green and purple lines). Cell collapse was scored by measuring the cell area at each time point. Data represent mean 6 SEM of 43 cells for EphA4WT-mCherry Fc, 30 cells for EphA4CR-mCherry Fc, 39 cells for EphA4WT-mCherry ephrinA5, and 29 cells for EphA4CR-mCherry ephrinA5. Statistical significance was determined using two-way ANOVA, followed by Bonferroni’s post hoc test, and is shown with green stars (EphA4WT-mCherry Fc versus ephrinA5) and purple stars (EphA4CR-mCherry Fc versus ephrinA5); *p < 0.05, **p < 0.01, ***p < 0.001. (F–G0 ) Representative fluorescent (F, G) and brightfield (F0 , G0 ) images of HeLa cells transfected with EphA4WT-mCherry(EphA4WT) (F) or EphA4CR-mCherry (EphA4CR) (G) and cocultured with SKN cells endogenously expressing ephrins, labeled with cell tracker green. SKN and transfected HeLa cells are defined by the dashed green and red lines, respectively. Scale bar, 25 mm. (H) Bar graph showing the percentage of cell-cell detachment (EphA4WT n = 24 and EphA4CR n = 22 events observed). Statistical analysis was performed using Fisher’s exact test. (I) Bar graph showing the mean 6 SEM number of vesicles per SKN cells (n = 344 SKN cells in contact with EphA4WT-expressing cells and n = 365 SKN cells in contact with EphA4CR-expressing cells). The experiment was repeated four times. Statistical analysis was performed as in Figure 2H. (J–K) Representative examples of HeLa cells transfected with EphA4WT (J) and EphA4CR (K) and cocultured with SKN-H2B-RFP cells (nuclei visible by stable expression of histone2B-RFP). Cells were stained with anti-FLAG antibody before and after permeabilization to detect surface and total expression of EphA4, respectively. Cell outlines were labeled by Cell Mask Blue (CMB) staining. Insets show higher magnification of the vesicles. Arrowheads in the color-combined pictures point to endocytosed vesicles. Scale bar, 10 mm. See also Figure S3.

Current Biology Vol 24 No 20 2360

+/+

+/CR

CR/CR

EphA4

NF

EphA4 ECD

** wild-type

E12.5

EphA4CR/CR

E12.5

C

- 75 Tubulin

Fc

0

Peroneal nerve

Collapsed GC [%]

wild-type

Tibial nerve

G 80

60

wild-type, EphA4CR/wt EphA4CR/CR

40 20 0

**

EphA4CR/CR

80

ephrinA2/5

H

1

F 100

Phalloidin Hb9-GFP

E

Figure 4. EphA4 Cleavage Is Dispensable for Axon Guidance In Vivo

2

kDa

- 120

*

D

Fc

% detachment

B

forebrain

EphA4/NF

A

0.2 0.5 1 ephrinA2/A5 (μg/ml)

60 40

*

20 0

ctrl

CR

wild-type

ephrinA5 Hb9-GFP

00:00

00:20

00:40

01:00

01:20

00:40

01:00

01:40

02:00

H’

I

EphA4CR/CR

00:00

I’

J

Tuj1

GFP

Tuj1 M GFP D V 0

GFP

K

20 40 60 80 100 % GFP Fluorescence

*

N D

*

V 0

EphA4WT/GFP

L

20 40 60 80 100 % GFP Fluorescence

O D

V 0

EphA4CR/GFP

20 40 60 80 100 % GFP Fluorescence

In embryonic hindlimbs the increases of EphA4CR expression were most prominent during the period of motor axon ingrowth (E11.5–E13.5) (Figures 6B and 6D), raising the possibility that EphA4 cleavage could be triggered by interaction with ephrinAs [15]. However, stimulation of transfected cells expressing EphA4 and of cortical neurons endogenously expressing EphA4 with different concentrations of ephrinA5-Fc (see Figure 2G) and with

ns

(A) Western blots of E12.5 wild-type (+/+), EphA4wt/CR (CR/+), and EphA4CR/CR (CR/CR) forebrain lysates probed with EphA4-ECD and tubulin antibodies. EphA4 full-length runs at 120 kDa and EphA4-ECD at 75 kDa (asterisk). (B and C) Immunostaining of E12.5 wild-type (B) and EphA4CR/CR (C) hindlimb nerves with EphA4 and neurofilament (NF) antibodies. Arrowheads point to the peroneal nerve, and arrows point to the tibial nerve. EphA4 expression is higher in the peroneal nerve. Scale bar, 25 mm. (D) Graph representing mean 6 SEM of EphA4 staining intensity on E12.5 peroneal and tibial nerves of wild-type (N = 3) and EphA4CR/CR (N = 3) embryos. Statistical analysis was performed as in Figure 2H (**p < 0.01). (E) Representative examples of noncollapsed and collapsed growth cones from motor neuron explant cultures stimulated with preclustered Fc (as a control) or ephrinA2/A5-Fc (1:1 mix). In green is Hb9-GFP and in red Phalloidin-568 staining. Scale bar, 20 mm. (F) Graph represents the percentage of collapsed growth cones (GC) in controls (wild-type or EphA4wt/CR) versus EphA4CR/CR embryos. Four to six explants per condition were analyzed from four embryos per genotype. For each genotype, the percentage of collapsed growth cones at different conditions was compared using one-way ANOVA followed by Bonferroni’s post hoc test. For the control group: Fc versus 0.5 mg/ml ephrins, p < 0.05; Fc versus 1 mg/ml ephrins, p < 0.001. For the EphA4CR/CR group: Fc versus 0.5 mg/ml ephrins, p < 0.01; Fc versus 1 mg/ml ephrins, p < 0.01. To compare the two genotypes, statistical analysis performed done using two-way ANOVA followed by Bonferroni’s post hoc test, and there was no statistical significant difference. Error bars indicate SEM. (G) Bar graph showing the percentage of axoncell detachment (wild-type and EphA4wt/CR n = 26 and EphA4CR/CR n = 33 encounters). Statistical analysis was performed as in Figure 3H. (H–I0 ) Representative fluorescence (H and I) and brightfield (H0 and I0 ) images of time-lapse movies of wild-type (H and H0 ) or EphA4CR/CR (I and I0 ) neurons (green, Hb9-GFP+) cocultured with ephrinA5-mCherry transfected HeLa cells (red). Arrows point to the tip of the axons. HeLa cell borders are indicated with red dashed lines in brightfield images. Scale bar, 10 mm. (J–O) Tuj1 and GFP detection in the limb nerve branches in the crural plexus of HH st. 25/26 electroporated chick embryos. Quantification of GFP signals is expressed as percentage in dorsal and ventral limb nerves (% GFP Fluorescence). Data represent mean 6 SEM (N = 5 for GFP, N = 5 for EphA4WT-GFP, N = 9 for EphA4CR-GFP). Statistical analysis was performed as in Figure 3H. ns, not significant. Scale bar, 100 mm. See also Figures S4 and S5.

membrane-anchored ephrinA5, respectively, did not significantly increase EphA4-ICD levels (Figures S6G–S6I). To the contrary, prolonged stimulation of transfected cells with ephrinA4-Fc significantly decreased EphA4-ICD levels (Figures 6E and 6F). Together these results suggest that EphA4 cleavage is ephrin independent and constitutes a mechanism to persistently regulate the levels of full-length EphA4.

EphA4 Cleavage Regulates Motor Axon Guidance 2361

B

C

wild-type

Islet1

Lim1

RD

D % Lim1+, RD+ neurons

A

Lim1 Islet1 RD EphA4CR/CR

*

40 30 20 10 0

lateral LMC medial LMC RD

wil

F

G

wild-type

Islet1

Lim1

RD

Lim1 Islet1 RD EphA4CR/CR

d-t

yp

e

E

H % Islet1+, RD+ neurons

E

50

A ph

CR

4

/C

R

50 40 30 20 10 0

wil

d-t

yp

e Ep

R

/C

wt

4 hA

Ep

CR

h

A4

/C

R

Figure 5. EphA4 Cleavage Is Required for Lateral LMC Axon Guidance In Vivo (A and E) Schematic drawings showing the dorsal (A) and ventral (E) injection sites. (B and C) Confocal pictures of ventral retrograde tracings in wild-type (B) and EphA4CR/CR (C) E12.5 embryos. Rhodamine dextran (RD) was injected in the ventral shank of the hindlimb, and sections were stained with Islet1 and Lim1 to label medial LMC and lateral LMC (populations are delineated by dashed lines), respectively. Scale bar, 50 mm. (D) Graph represents the percentage of misprojections (neurons positive for RD and Lim1 staining) in relation to all RD-labeled cells in ventral retrograde tracings of five wild-type embryos and eight EphA4CR/CR embryos. Each dot in the graphs represents one embryo. The black line represents the mean. Statistical analysis was performed using the nonparametric Mann-Whitney test (*p < 0.05). (F and G) Confocal pictures of dorsal retrograde tracings in wild-type (F) and EphA4CR/CR (G) E12.5 embryos. Rhodamine dextran (RD) was injected in the dorsal shank of the hindlimb and sections were stained as above. (H) Graph represents the percentage of misprojections (neurons positive for RD and Islet1) in relation to all RD-labeled cells in dorsal retrograde tracings. The black line represents the mean. Seven wild-type embryos, four EphA4wt/CR, and seven EphA4CR/CR were analyzed. Each dot in the graphs represents one embryo.

EphA4 Cleavage Regulates the Balance of cis or trans Modes of Receptor-Ligand Interactions How might more EphA4 protein in limbs of EphA4CR/CR mutants influence LMC axon guidance? Higher EphA4 levels in the ventral limb could attract more lateral LMC axons there, through EphA4:ephrinA5 reverse signaling [19–21]. Additionally, higher EphA4 levels in the ventral limb could cis-mask ephrinAs expressed there, thereby reducing the effective concentration of free ephrinAs that normally repulses lateral LMC axons from the ventral limb. To test the latter scenario, we estimated EphA4 abundance by determining EphA4 immunofluorescence intensity in E11.5 and E12.5 wild-type and EphA4CR/CR limb sections (Figures 7A–7C; Figures S7A– S7C). EphA4 levels were significantly increased in dorsal and ventral hindlimbs from EphA4CR/CR embryos compared with wild-type controls (Figure 7D; Figure S7D). Receptor-unbound ephrinAs (ƩephrinAFREE) in tissue sections are typically detected with recombinant EphA7-Fc preclustered with a fluorescent antibody [18–20, 32]. Before doing so, we analyzed which parameters determine changes in free ephrins. When ephrinA5 and EphA4 were coexpressed in HeLa cells, whereas the amount of surface expression of ephrinA5 depended only on the total ephrin expression, the amount of receptor-unbound ephrinAs (ƩephrinAFREE) inversely correlated with the ratio of EphA4-ephrinA5 in the cells (Figures S7E–S7M). We did not observe any differences

in the behavior of EphA4CR and EphA4WT, raising the possibility that in EphA4CR/CR mice the increased levels of full-length receptor might reduce the levels of ƩephrinAFREE. Indeed, the amounts of ƩephrinAFREE in the ventral and dorsal mesenchyme were significantly reduced in EphA4CR/CR embryos at both developmental stages examined (Figures 7E–7J; Figures S7N and S7O). The levels of ƩephrinAFREE in the ventral mesenchyme of EphA4CR/CR embryos were comparable to the levels of ƩephrinAFREE in the dorsal mesenchyme of wild-type embryos, a concentration compatible with ingrowth of lateral LMC axons (Figures 7H and 7J; two-tailed unpaired Student’s t test, p > 0.2). Since we did not detect alterations in ephrinA2 and ephrinA5 mRNA levels (Figures S6E and S6F), our findings suggest that spinal motor axon guidance errors in EphA4CR/CR mice occur in part through increased EphA4 expression levels squelching the ƩephrinAFREE concentration differential between ventral and dorsal hindlimb, arguing that EphA4 cleavage is required for the accurate restriction of the distribution of ephrinA that is capable to initiate forward EphA signaling. Discussion EphA4 Cleavage in Cell Detachment and Growth Cone Collapse Paradoxically, ephrin:Eph forward signaling that results in repulsion between ligand and receptor-bearing cells, initially

Current Biology Vol 24 No 20 2362

A kDa 120-

EphA4 ECD

B

spinal cord +/+

hindlimb +/+

+/CR CR/CR

+/CR

CR/CR

*

75-

*

Tubulin

120-

EphA4 ICD

*

50-

**

Tubulin

full length/tubulin

C

1.5

1.0

wild-type EphA4wt/CR EphA4CR/CR

spinal cord

*

*

0.5

* 0

E11.5

full length/tubulin

D

2.0

E12.5

E13.5

hindlimb

1.5

*

*

E17.5

E15.5

wild-type

*

EphA4wt/CR

EphA4CR/CR

**

1.0

*

0.5

0

E11.5

E

E12.5

E13.5

E15.5

E17.5

ephrinA4 UT Fc 30’ 60’ 120’ 240’

EphA4 ICD

*

-50

Tubulin

ICD/full length

pEphA F

kDa -120

0.3 0.2

* 0.1 0 Fc

30

60

120 240

Figure 6. EphA4 Cleavage Inhibition Increases EphA4 Protein Abundance (A and B) Representative western blots of E12.5 wild-type (+/+), EphA4wt/CR (CR/+), and EphA4CR/CR (CR/CR) spinal cord (A) and hindlimb (B) lysates probed with EphA4-ECD, EphA4-ICD, and tubulin antibodies. EphA4 fulllength runs at 120 kDa, EphA4-ECD at 75kDa (orange asterisk), and EphA4-ICD below 50 kDa (purple asterisk). The purple arrowhead points to a faint band that is unlikely to be a residual shed ectodomain since no EphA4-ICD is detected, but it might suggest a minimal residual proteolytic processing of EphA4 in the hindlimb of these mutant embryos. (C and D) Graphs representing mean 6 SEM full-length EphA4 expression normalized to tubulin at several developmental stages in wild-type, EphA4wt/CR, and EphA4CR/CR spinal cord (C) and hindlimb (D). Lysates from three to five embryos per developmental stage were analyzed. Statistical analysis was performed as in Figure 2H (*p < 0.05, **p < 0.01).

leads to a ligand-receptor complex that bridges the two cells, de facto increasing adhesion [33]. Repulsion requires that the initial adhesive ligand-receptor complex is disrupted, and one way to achieve this is through proteolytic cleavage of ephrin and Eph ectodomains [9, 10, 12]. Our in vitro experiments on cells expressing the cleavage resistant isoform EphA4CR provide support for this model. Although the molecular mechanism underlying cell contact-triggered cleavage of EphA4 remains to be elucidated, our data argue against soluble ephrins triggering EphA4 cleavage. They suggest that cellcell contact is necessary, perhaps to allow EphA4 cleavage in trans by proteases expressed by the ephrin-expressing cell. Alternatively, activation of ephrinA reverse signaling in the opposing cell may trigger signals that ultimately activate proteases in the EphA4 cell. Our experiments with EphA4CR/CR motor neurons cocultured with cells expressing ephrinA5 showed normal rates of growth cone collapse, arguing that cleavage is specifically required for cell detachment and not for growth cone collapse, similar to our previous observations on Eph-ephrin trans-endocytosis [14]. In chick, EphA4CR overexpression in spinal motor neurons redirects them as efficiently as overexpression of the wildtype isoform. This result was unexpected, since this rerouting is dependent on ephrinA-EphA4-mediated repulsion that requires the detachment of motor axons from ephrinA+ limb mesenchyme. More importantly, together, these experiments contrast the report that cleavage of the related EphB2 receptor is required for growth cone collapse [12]. It is unlikely that in chick alternative EphA4 cleavage sites are used, because our biochemical analysis indicates cleavage of wild-type EphA4 and lack of cleavage of EphA4CR protein in electroporated limbs (Figures S5A and S5B). It is also unlikely that overexpressed EphA4CR protein binds to ephrinAs in cis on medial LMC motor axons, thereby freeing endogenous wild-type EphA4 that is normally cis masked by ephrinAs. If this was true, overexpression of signalingdefective EphA4 should have the same effect and lead to rerouting of medial LMC motor axons. However, this is not the case [34]. A more likely explanation could be that the lack of EphA4 cleavage is compensated by alternative mechanisms such as ephrinA cleavage or Eph-ephrin endocytosis. In fact, we did observe enhanced trans-endocytosis of EphA4CR into ephrin-expressing opposing cells in vitro (Figures 3I–3K). To corroborate the limited role of EphA4 cleavage for forward repulsion, we analyzed another structure whose development depends on EphA4-mediated repulsion, the formation of the dorsal funiculus (DF) in the spinal cord. DF formation depends on proper positioning of EphA4+ spinal neurons at either side of the spinal cord midline [29], a process that depends on EphA4 forward signaling [26, 29, 30]. In EphA4CR/CR mice, DF formation is normal (Figures S4J and S4K), arguing that cleavage of neuronal EphA4 is not the dominant mechanism to regulate EphA4 forward repulsion in vivo.

(E) Representative western blot of TCLs prepared from HEK293 cells transfected with EphA4WT and stimulated with 1 mg/ml soluble preclustered Fc or ephrinA4-Fc for the indicated times. Samples were probed using EphA4ICD, phosphoEphA (pEphA), and tubulin antibodies. (F) Bar graph showing the levels of EphA4-ICD normalized to EphA4 fulllength expression in cells stimulated as described in V. Data represent mean 6 SEM of six experiments. Statistical analysis was performed as in Figure 2H. See also Figure S6.

EphA4 Cleavage Regulates Motor Axon Guidance 2363

NF

A

NF EphA4

EphA4

E11.5 wild-type

B

E11.5 EphA4CR/CR

C 800

25

wild-type EphA4CR/CR

400

10

179

268

357

446

∑ephrinAFREE

60 30 1

70 139 208 277 346 415 ns

∑ephrinA

FREE

50

30

E12.5

40 20

322

429

536

643

E11.5

*

wild-type EphA4CR/CR

20 10 dorsal

ventral

ns 25

*

20

** *

15 10 5 0

215

**

*

J

∑ephrinAFREE

wild-type EphA4CR/CR

*

40

0

80

108

90

H

EphA4CR/CR

1

wild-type EphA4CR/CR

E11.5

0

wild-type E11.5

60

ventral

dorsal G 120 Hb9-GFP 535

∑ephrinAFREE

90

E11.5

E

∑ephrinAFREE

*

0

1

0

wild-type EphA4CR/CR

5

0

I

***

15

200

F

**

*

20

EphA4

600

EphA4

ns

D

dorsal

ventral

E12.5 wild-type EphA4CR/CR

Figure 7. EphA4 Cleavage Regulates the Balance of cis or trans Modes of Receptor-Ligand Interactions (A and B) Representative images of E11.5 wildtype and EphA4CR/CR hindlimbs stained with EphA4 and neurofilament (NF) antibodies. Arrowheads point to the peroneal nerve, and arrows point to the tibial nerve. Yellow stippled lines indicate the segment that was analyzed to determine EphA4 levels in the hindlimb, and blue stippled lines indicate the dorso-ventral border of the hindlimb. Scale bar, 100 mm. (C) Representative line graph showing EphA4 staining intensity along the yellow lines in (A) and (B). The x axis represents the distance (pixel) from the most dorsal point of the stippled yellow lines. The black line indicates the dorso-ventral border. (D) Bar graph showing the mean 6 SEM EphA4 staining intensity in wild-type (N = 5) and EphA4CR/CR (N = 8) embryos plotted as the average intensity in the dorsal and ventral shanks. Statistical analysis was performed as in Figure 2H. (E and F) Representative images of EphA7-Fc overlay (ƩephrinA) on cryosections from E11.5 Hb9-GFP+ wild-type (E) and EphA4CR/CR (F) embryos. Yellow stippled lines indicate the segment that was analyzed to determine EphA4 levels in the hindlimb, and the blue stippled line indicates the dorso-ventral border. Arrowheads point to the peroneal nerve, and arrows point to the tibial nerve. Scale bar, 50 mm. (G and I) Representative line graphs showing EphA7-Fc overlay staining intensity measured along the yellow lines depicted in (E) and (F) [E11.5] (G) or in Figures S5K and S5L [E12.5] (I) in wild-type and EphA4CR/CR cross-sections. The x axis represents the distance (pixel) from the most dorsal point of the stippled yellow lines. The black line indicates the dorso-ventral border. (H and J) Graph showing the mean 6 SEM EphA7-Fc overlay staining intensity in E11.5 (H) and E12.5 (J) wild-type (E11.5 N = 5, E12.5 N = 8) and EphA4CR/CR (E11.5 N = 8, E12.5 N = 9) embryos plotted as average intensity in the dorsal and ventral shank. ns, not significant. Statistical analysis was performed as in Figure 4D. See also Figure S7.

Current Biology Vol 24 No 20 2364

EphA4 Cleavage and Spatial Restriction of ephrinA Activity The EphA4CR mouse mutation abolishes EphA4 cleavage in all tissues and results in more lateral LMC axons entering the ventral limb. What is the underlying mechanism? One possibility could be that EphA4CR is unable to initiate ephrinA reverse signaling, which normally attracts lateral LMC axons in the dorsal limb [19–21]. We consider this a rather unlikely possibility, since the anterior commissure of the forebrain, an axon tract whose development is mediated by EphA4:ephrin reverse signaling [26, 30, 31], is normal in EphA4CR/CR mice (Figure S4L). We also exclude a role for elevated EphA4 expression in the dorsal limb of EphA4CR/CR embryos, since this would increase EphA4:ephrinA reverse signaling and result in more axons entering the dorsal limb [19–21], which is opposite to the phenotype observed in EphA4CR/CR mice. Another important mechanism that controls Eph signaling is cis-inhibition of EphA4 by coexpressed ephrinAs [35, 36]. Previous work showed that raising EphA4 levels in motor neurons leads to ephrinA cis-attenuation and abolishes ephrinEph trans-interactions [32]. Our analysis of EphA4CR/CR mice suggests that cis-attenuation in the target tissue is a likely explanation for the observed lateral LMC misrouting. We propose that the increased levels of full-length EphA4 in the hindlimbs of EphA4CR/CR mice mask ephrinAs in cis and thereby reduce ligand availability for trans-interactions. Thus, our findings suggest that cleavage of EphA4 in the target tissue is a mechanism for the accurate distribution of its signalingcompetent ligand. Contrasting this observation, previous reports have described cell-autonomous protease functions to regulate DCC or PlexinA1 receptor availability on axons [2, 4, 6]. One recent report exploring the molecular mechanism of axon guidance cue distribution has implicated dystroglycan modification via glycosylation [37], but our findings are the first example of receptor cleavage controlling ligand distribution variegating the repertoire of protease functions in axon guidance. More generally, our data add protease-mediated receptor cleavage to the expanding portfolio of strategies that diversify the outputs of a relatively small number of axon guidance ligand-receptor systems, which have yet to match the complexity of neuronal connections in our nervous system. EphA4 Cleavage Is a Constitutive Mechanism to Regulate Full-Length EphA4 Expression Contrary to what was described so far for other Eph receptors [15, 38], we find that ephrin stimulation has little effect on EphA4 cleavage in cultured cells. We propose that proteolytic cleavage of EphA4 is a constitutive posttranslational mechanism to regulate full-length EphA4 expression, independently of ephrinA-engagement. Whether proteolytic cleavage operates in parallel to other mechanisms such as local translation [39] or involves negative feedback loops, such that the cleaved EphA4-ICD lowers EphA4 production, remains to be established. It would also be interesting to explore if the released EphA4-ECD is bioactive, e.g., as an antagonist of Eph-ephrin signaling. In summary, our findings establish EphA4 cleavage as an important regulator of axon guidance in vivo by regulating the availability of free ephrinAs in the axon intermediate target. It remains to be shown whether adult functions of EphA4 such as the regulation of neuronal plasticity and spine growth are modulated by cleavage and whether EphA4’s role as disease modifier of amyotrophic lateral sclerosis depends on proteolytic cleavage [40–42].

Experimental Procedures Animals Hb9-GFP, Prx1-cre, Nestin-cre, PGK-cre, EphA4lx, EphA4wtKI, and EphA42/2 mice have been described previously [26, 27, 43–46]. EphA4CR knockin mice were generated as described in Kullander et al., 2001 [26]. All the mutants were maintained in a comparable mixed 129/Svev 3 C57Bl/6 background. Fertilized chick eggs (Couvoir Simetin) were stored for a maximum of 1 week at 18 C and then incubated at 38 C and staged according to standard protocols [28]. Animal procedures were approved by the institutional animal safety representative. Retrograde Tracings E12.5 embryos were eviscerated and kept in DMEM/F-12 medium (Invitrogen) aerated with 5% CO2/95% O2. A 6% lysine-fixable tetramethylrhodamine-dextran (molecular weight 3000, Invitrogen) solution in PBS with 0.4% Triton X-100 was injected into the ventral and dorsal shanks of the hindlimb and allowed to diffuse for 6 hr at room temperature. Images were acquired with an Olympus FV1000 confocal microscope using a 203 objective. Statistical analysis was done using the nonparametric MannWhitney test. Primary neuronal cultures, immunostaining, live imaging, chick overexpression, and western blot were performed as described in [20, 47, 48]. See the Supplemental Experimental Procedures for a more detailed description. Supplemental Information Supplemental Information includes Supplemental Experimental Procedures and seven figures and can be found with this article online at http://dx.doi. org/10.1016/j.cub.2014.08.028. Acknowledgments We thank P. Alcala´, J. Lindner, L. Gaitanos, and the transgenic service unit for technical assistance in generating the EphA4CR knockin mouse; D. HabaSchneider and R. Goetz for mouse genotyping; D. Haba-Schneider for helping with cryosections and immunofluorescence; A. Huber-Bro¨samle for kindly providing reagents; T. Gaitanos for sharing the SKN-H2B-RFP cell line; S. Paixa˜o for sharing some of the EphA4 mutant constructs; I. Dudanova for sharing embryonic lysates from EphA4 conditional knockouts; and M. Liang for technical assistance. This study was supported by the Max Planck Society, by a grant from the German Israeli Foundation (to R.K.), and by grants from the Canadian Institutes of Health Research (MOP-77556 and MOP-97758), the Natural Sciences and Engineering Research Council of Canada, Brain Canada, and the W. Garfield Weston Foundation (to A. K.). D.M. holds a Mexican Council for Science and Technology (CONACYT) graduate scholarship. Received: April 30, 2014 Revised: July 22, 2014 Accepted: August 13, 2014 Published: September 25, 2014 References 1. O’Donnell, M., Chance, R.K., and Bashaw, G.J. (2009). Axon growth and guidance: receptor regulation and signal transduction. Annu. Rev. Neurosci. 32, 383–412. 2. Charoy, C., Nawabi, H., Reynaud, F., Derrington, E., Bozon, M., Wright, K., Falk, J., Helmbacher, F., Kindbeiter, K., and Castellani, V. (2012). gdnf activates midline repulsion by Semaphorin3B via NCAM during commissural axon guidance. Neuron 75, 1051–1066. 3. Galko, M.J., and Tessier-Lavigne, M. (2000). Function of an axonal chemoattractant modulated by metalloprotease activity. Science 289, 1365–1367. 4. Nawabi, H., Brianc¸on-Marjollet, A., Clark, C., Sanyas, I., Takamatsu, H., Okuno, T., Kumanogoh, A., Bozon, M., Takeshima, K., Yoshida, Y., et al. (2010). A midline switch of receptor processing regulates commissural axon guidance in vertebrates. Genes Dev. 24, 396–410. 5. Yamagishi, S., Hampel, F., Hata, K., Del Toro, D., Schwark, M., Kvachnina, E., Bastmeyer, M., Yamashita, T., Tarabykin, V., Klein, R., and Egea, J. (2011). FLRT2 and FLRT3 act as repulsive guidance cues for Unc5-positive neurons. EMBO J. 30, 2920–2933.

EphA4 Cleavage Regulates Motor Axon Guidance 2365

6. Bai, G., Chivatakarn, O., Bonanomi, D., Lettieri, K., Franco, L., Xia, C., Stein, E., Ma, L., Lewcock, J.W., and Pfaff, S.L. (2011). Presenilin-dependent receptor processing is required for axon guidance. Cell 144, 106–118. 7. Coleman, H.A., Labrador, J.P., Chance, R.K., and Bashaw, G.J. (2010). The Adam family metalloprotease Kuzbanian regulates the cleavage of the roundabout receptor to control axon repulsion at the midline. Development 137, 2417–2426. 8. Georgakopoulos, A., Litterst, C., Ghersi, E., Baki, L., Xu, C., Serban, G., and Robakis, N.K. (2006). Metalloproteinase/Presenilin1 processing of ephrinB regulates EphB-induced Src phosphorylation and signaling. EMBO J. 25, 1242–1252. 9. Hattori, M., Osterfield, M., and Flanagan, J.G. (2000). Regulated cleavage of a contact-mediated axon repellent. Science 289, 1360–1365. 10. Janes, P.W., Saha, N., Barton, W.A., Kolev, M.V., Wimmer-Kleikamp, S.H., Nievergall, E., Blobel, C.P., Himanen, J.P., Lackmann, M., and Nikolov, D.B. (2005). Adam meets Eph: an ADAM substrate recognition module acts as a molecular switch for ephrin cleavage in trans. Cell 123, 291–304. 11. Janes, P.W., Wimmer-Kleikamp, S.H., Frangakis, A.S., Treble, K., Griesshaber, B., Sabet, O., Grabenbauer, M., Ting, A.Y., Saftig, P., Bastiaens, P.I., and Lackmann, M. (2009). Cytoplasmic relaxation of active Eph controls ephrin shedding by ADAM10. PLoS Biol. 7, e1000215. 12. Lin, K.T., Sloniowski, S., Ethell, D.W., and Ethell, I.M. (2008). Ephrin-B2induced cleavage of EphB2 receptor is mediated by matrix metalloproteinases to trigger cell repulsion. J. Biol. Chem. 283, 28969–28979. 13. Marston, D.J., Dickinson, S., and Nobes, C.D. (2003). Rac-dependent trans-endocytosis of ephrinBs regulates Eph-ephrin contact repulsion. Nat. Cell Biol. 5, 879–888. 14. Zimmer, M., Palmer, A., Ko¨hler, J., and Klein, R. (2003). EphB-ephrinB bi-directional endocytosis terminates adhesion allowing contact mediated repulsion. Nat. Cell Biol. 5, 869–878. 15. Litterst, C., Georgakopoulos, A., Shioi, J., Ghersi, E., Wisniewski, T., Wang, R., Ludwig, A., and Robakis, N.K. (2007). Ligand binding and calcium influx induce distinct ectodomain/gamma-secretase-processing pathways of EphB2 receptor. J. Biol. Chem. 282, 16155–16163. 16. Helmbacher, F., Schneider-Maunoury, S., Topilko, P., Tiret, L., and Charnay, P. (2000). Targeting of the EphA4 tyrosine kinase receptor affects dorsal/ventral pathfinding of limb motor axons. Development 127, 3313–3324. 17. Eberhart, J., Swartz, M.E., Koblar, S.A., Pasquale, E.B., and Krull, C.E. (2002). EphA4 constitutes a population-specific guidance cue for motor neurons. Dev. Biol. 247, 89–101. 18. Kania, A., and Jessell, T.M. (2003). Topographic motor projections in the limb imposed by LIM homeodomain protein regulation of ephrin-A:EphA interactions. Neuron 38, 581–596. 19. 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. 20. Kao, T.J., and Kania, A. (2011). Ephrin-mediatedcis-attenuation of Eph receptor signaling is essential for spinal motor axon guidance. Neuron 71, 76–91. 21. Dudanova, I., Kao, T.J., Herrmann, J.E., Zheng, B., Kania, A., and Klein, R. (2012). Genetic evidence for a contribution of EphA:ephrinA reverse signaling to motor axon guidance. J. Neurosci. 32, 5209–5215. 22. Schaupp, A., Sabet, O., Dudanova, I., Ponserre, M., Bastiaens, P., and Klein, R. (2014). The composition of EphB2 clusters determines the strength in the cellular repulsion response. J. Cell Biol. 204, 409–422. 23. Lauterbach, J., and Klein, R. (2006). Release of full-length EphB2 receptors from hippocampal neurons to cocultured glial cells. J. Neurosci. 26, 11575–11581. 24. 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 receptor clustering in Eph function. Neuron 47, 515–528. 25. Grunwald, I.C., Korte, M., Adelmann, G., Plueck, A., Kullander, K., Adams, R.H., Frotscher, M., Bonhoeffer, T., and Klein, R. (2004). Hippocampal plasticity requires postsynaptic ephrinBs. Nat. Neurosci. 7, 33–40. 26. Kullander, K., Mather, N.K., Diella, F., Dottori, M., Boyd, A.W., and Klein, R. (2001). Kinase-dependent and kinase-independent functions of EphA4 receptors in major axon tract formation in vivo. Neuron 29, 73–84. 27. Wichterle, H., Lieberam, I., Porter, J.A., and Jessell, T.M. (2002). Directed differentiation of embryonic stem cells into motor neurons. Cell 110, 385–397.

28. Hamburger, V., and Hamilton, H.L. (1951). A series of normal stages in the development of the chick embryo. J. Morphol. 88, 49–92. 29. Paixa˜o, S., Balijepalli, A., Serradj, N., Niu, J., Luo, W., Martin, J.H., and Klein, R. (2013). EphrinB3/EphA4-mediated guidance of ascending and descending spinal tracts. Neuron 80, 1407–1420. 30. Yokoyama, N., Romero, M.I., Cowan, C.A., Galvan, P., Helmbacher, F., Charnay, P., Parada, L.F., and Henkemeyer, M. (2001). Forward signaling mediated by ephrin-B3 prevents contralateral corticospinal axons from recrossing the spinal cord midline. Neuron 29, 85–97. 31. Cowan, C.A., Yokoyama, N., Saxena, A., Chumley, M.J., Silvany, R.E., Baker, L.A., Srivastava, D., and Henkemeyer, M. (2004). Ephrin-B2 reverse signaling is required for axon pathfinding and cardiac valve formation but not early vascular development. Dev. Biol. 271, 263–271. 32. Bonanomi, D., Chivatakarn, O., Bai, G., Abdesselem, H., Lettieri, K., Marquardt, T., Pierchala, B.A., and Pfaff, S.L. (2012). Ret is a multifunctional coreceptor that integrates diffusible- and contact-axon guidance signals. Cell 148, 568–582. 33. Egea, J., and Klein, R. (2007). Bidirectional Eph-ephrin signaling during axon guidance. Trends Cell Biol. 17, 230–238. 34. Kao, T.J., Palmesino, E., and Kania, A. (2009). SRC family kinases are required for limb trajectory selection by spinal motor axons. J. Neurosci. 29, 5690–5700. 35. Carvalho, R.F., Beutler, M., Marler, K.J., Kno¨ll, B., Becker-Barroso, E., Heintzmann, R., Ng, T., and Drescher, U. (2006). Silencing of EphA3 through acis interaction with ephrinA5. Nat. Neurosci. 9, 322–330. 36. Falivelli, G., Lisabeth, E.M., Rubio de la Torre, E., Perez-Tenorio, G., Tosato, G., Salvucci, O., and Pasquale, E.B. (2013). Attenuation of eph receptor kinase activation in cancer cells by coexpressed ephrin ligands. PLoS ONE 8, e81445. 37. Wright, K.M., Lyon, K.A., Leung, H., Leahy, D.J., Ma, L., and Ginty, D.D. (2012). Dystroglycan organizes axon guidance cue localization and axonal pathfinding. Neuron 76, 931–944. 38. Sugiyama, N., Gucciardo, E., Tatti, O., Varjosalo, M., Hyytia¨inen, M., Gstaiger, M., and Lehti, K. (2013). EphA2 cleavage by MT1-MMP triggers single cancer cell invasion via homotypic cell repulsion. J. Cell Biol. 201, 467–484. 39. Brittis, P.A., Lu, Q., and Flanagan, J.G. (2002). Axonal protein synthesis provides a mechanism for localized regulation at an intermediate target. Cell 110, 223–235. 40. Filosa, A., Paixa˜o, S., Honsek, S.D., Carmona, M.A., Becker, L., Feddersen, B., Gaitanos, L., Rudhard, Y., Schoepfer, R., Klopstock, T., et al. (2009). Neuron-glia communication via EphA4/ephrin-A3 modulates LTP through glial glutamate transport. Nat. Neurosci. 12, 1285–1292. 41. Inoue, E., Deguchi-Tawarada, M., Togawa, A., Matsui, C., Arita, K., Katahira-Tayama, S., Sato, T., Yamauchi, E., Oda, Y., and Takai, Y. (2009). Synaptic activity prompts gamma-secretase-mediated cleavage of EphA4 and dendritic spine formation. J. Cell Biol. 185, 551–564. 42. Van Hoecke, A., Schoonaert, L., Lemmens, R., Timmers, M., Staats, K.A., Laird, A.S., Peeters, E., Philips, T., Goris, A., Dubois, B., et al. (2012). EPHA4 is a disease modifier of amyotrophic lateral sclerosis in animal models and in humans. Nat. Med. 18, 1418–1422. 43. Herrmann, J.E., Pence, M.A., Shapera, E.A., Shah, R.R., Geoffroy, C.G., and Zheng, B. (2010). Generation of an EphA4 conditional allele in mice. Genesis 48, 101–105. 44. Lallemand, Y., Luria, V., Haffner-Krausz, R., and Lonai, P. (1998). Maternally expressed PGK-Cre transgene as a tool for early and uniform activation of the Cre site-specific recombinase. Transgenic Res. 7, 105–112. 45. Logan, M., Martin, J.F., Nagy, A., Lobe, C., Olson, E.N., and Tabin, C.J. (2002). Expression of Cre Recombinase in the developing mouse limb bud driven by a Prxl enhancer. Genesis 33, 77–80. 46. Tronche, F., Kellendonk, C., Kretz, O., Gass, P., Anlag, K., Orban, P.C., Bock, R., Klein, R., and Schu¨tz, G. (1999). Disruption of the glucocorticoid receptor gene in the nervous system results in reduced anxiety. Nat. Genet. 23, 99–103. 47. Dudanova, I., Gatto, G., and Klein, R. (2010). GDNF acts as a chemoattractant to support ephrinA-induced repulsion of limb motor axons. Curr. Biol. 20, 2150–2156. 48. Gatto, G., Dudanova, I., Suetterlin, P., Davies, A.M., Drescher, U., Bixby, J.L., and Klein, R. (2013). Protein tyrosine phosphatase receptor type O inhibits trigeminal axon growth and branching by repressing TrkB and Ret signaling. J. Neurosci. 33, 5399–5410.