© 2016. Published by The Company of Biologists Ltd | Journal of Cell Science (2016) 129, 1802-1814 doi:10.1242/jcs.184580

RESEARCH ARTICLE TrkA mediates retrograde 3A signaling through A4 to regulate dendritic branching Naoya Yamashita1,2,*, Masayuki Yamane1, Fumikazu Suto3 and Yoshio Goshima1,*

ABSTRACT subcellular compartments of – , and cell – Semaphorin 3A (Sema3A), a secretory semaphorin, exerts various body need to communicate with each other. biological actions through a complex between -1 and plexin- Semaphorin 3A (Sema3A), one of the first repulsive As (PlexAs). Sema3A induces retrograde signaling, which is involved guidance molecules to be recognized (Raper, 2000; Tran et al., in regulating dendritic localization of GluA2 (also known as GRIA2), 2007), plays an important role in proper neuronal network formation an AMPA subunit. Here, we investigated a possible by regulating not only motility but also dendritic – interaction between retrograde signaling pathways for Sema3A and development and maturation through neuropilin-1 (NRP1) Plexin- nerve growth factor (NGF). Sema3A induces colocalization of PlexA4 A (PlexA) receptor complexes (Campbell et al., 2001; Morita et al., (also known as PLXNA4) signals with those of tropomyosin-related 2006; Polleux et al., 2000; Sasaki et al., 2002; Shelly et al., 2011; kinase A (TrkA, also known as NTRK1) in growth cones, and these Suto et al., 2003; Yamashita et al., 2014). Sema3A accelerates colocalized signals were then observed along the axons. The time- endocytosis during growth cone collapse (Fournier et al., 2000; lapse imaging of PlexA4 and several TrkA mutants showed that the Hida et al., 2012), and also induces bidirectional axonal transport kinase and dynein-binding activity of TrkA were required for Sema3A- (Goshima et al., 1999, 1997; Li et al., 2004; Yamane et al., 2012). induced retrograde transport of the PlexA4–TrkA complex along the We have previously shown that Sema3A facilitates dendritic as well axons. The inhibition of the phosphoinositide 3-kinase (PI3K)–Akt as axonal transport in cultured hippocampal neurons. Sema3A signal, a downstream signaling pathway of TrkA, in the distal axon initiates signals at axonal growth cones, and these signals are suppressed Sema3A-induced dendritic localization of GluA2. The propagated to the somatodendritic compartments by retrograde knockdown of TrkA suppressed Sema3A-induced dendritic axonal transport of Sema3A and PlexA4 (also known as PLXNA4), localization of GluA2 and that suppressed Sema3A-regulated one of the PlexA isoforms. PlexA4 induces dendritic localization of dendritic branching both in vitro and in vivo. These findings suggest the AMPA receptor GluA2 (also known as GRIA2) through that by interacting with PlexA4, TrkA plays a crucial role in redirecting cis-interaction with GluA2 at the immunoglobulin-like plexin- local Sema3A signaling to retrograde axonal transport, thereby transcription-factor domain (PlexA-IPT) in the somatodendritic regulating dendritic GluA2 localization and patterning. compartment (Yamashita et al., 2014, 2016). This finding indicates that, at least among PlexAs, PlexA4 plays an important KEY WORDS: Hippocampus, Semaphorin 3A, NGF, TrkA, PlexA4, role in mediating retrograde Sema3A signaling through a form of Axonal transport signaling endosome. Thus, it is possible that the long-range signaling between the growth cone and the cell body of NGF and INTRODUCTION Sema3A might share some common components during neuronal Various neurotrophic factors, such as nerve growth factor (NGF), development. the prototypical growth factor, are required for the survival and In the present study, we provide evidence that, by acting at the maintenance of neurons. After NGF is released from the target cells, distal axons, Sema3A induces a dynein-dependent retrograde it binds to and activates its high-affinity receptor, tropomyosin- axonal transport of a TrkA–PlexA4 complex. Specifically, our related kinase (TrkA, also known as NTRK1), and is internalized findings suggest that Sema3A regulates dendritic patterning through into the responsive . The signaling endosome containing retrograde transport of the TrkA–PlexA4 complex along the axons, TrkA activated by NGF is subsequently trafficked back to the cell which in turn regulates the dendritic localization of GluA2. This body where the downstream signaling is activated. This movement represents a new mode of crosstalk between the signaling pathways of NGF from the axon tip to the soma is thought to be involved in the involving molecules and neurotrophins. long-range signaling between the growth cone and the cell body (Harrington and Ginty, 2013; Zweifel et al., 2005). In addition, RESULTS coordinated regulation of local and global neuronal signaling Sema3A induces PlexA4 and TrkA colocalization in cultured elicited by the extracellular environment is thought to be essential DRG neurons for the proper formation of neural networks. In this process, the To investigate the relationship between Sema3A and neurotrophin signaling, we first determined whether PlexA4 interacts with Trk 1Department of Molecular Pharmacology and Neurobiology, Yokohama City receptors. In HEK293T cells, we found that PlexA4 interacted University School of Medicine, Yokohama 236-0004, Japan. 2Department of 3 with TrkA and TrkC (also known as NTRK3) but not with TrkB Biology, Johns Hopkins University, Baltimore, MD 21218, USA. National Center of – Neurology and Psychiatry, National Institute of Neuroscience, Department of (also known as NTRK2) (Fig. 1A), and PlexA4 TrkA Ultrastructural Research, 4-1-1, Ogawahigashi, Kodaira, Tokyo 187-8502, Japan. immunoprecipitation efficiency was higher than that of PlexA4– TrkC. The relative ratio of immunoprecipitated TrkC to PlexA4 *Authors for correspondence ([email protected]; [email protected]) (ratio of TrkA to PlexA4 as a percentage) was 68.7±25.1% (mean± s.e.m.; n=3), suggesting that PlexA4 has a higher affinity for TrkA

Received 9 December 2015; Accepted 26 February 2016 than for TrkC. Journal of Cell Science

1802 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1802-1814 doi:10.1242/jcs.184580

Fig. 1. Sema3A induces interaction between PlexA4 and TrkA. (A) Flag-tagged PlexA4 expressed in HEK293T cells interacted with EGFP-tagged TrkA and TrkC, but not with TrkB. Immunoprecipitation (IP) was performed with anti-Flag-tag . A, B and C represent TrkA, TrkB and TrkC, respectively. (B) Schematic representation of TrkA and constructed TrkA mutants. TrkA contains signal (SP) and C1, LRR, C2, Ig1, Ig2, transmembrane (TM) domains and cytosolic domains. Amino acids 464–483 of the cytosolic domain make up the dynein-binding domain (DB). Full (F), full-length; ΔC1-C2 (C2), lacking the region from the C1 domain to the C2 domain; ΔC1-Ig1 (Ig1), lacking the region from the C1 domain to the Ig1 domain; K537A, catalytic inactive; Δ464–483, lacking amino acids 464–483. (C) Flag-tagged PlexA4 interacted with full-length TrkA and the C2 mutant but not with the Ig1 mutant of TrkA–EGFP. (D) After Sema3A treatment, the fluorescence clusters of PlexA4 and TrkA moved along the axons towards the cell bodies in a time-dependent manner (arrowheads). Scale bar: 10 µm. (E) Time dependence of the mean relative ratio of immunostaining intensity in the axon to that in the growth cone (n=12 neurons from three independent cultures). (F) Sema3A treatment increased the number of double-positive clusters (arrowheads) of anti-PlexA4 and anti-TrkA in the growth cones (main panels), and subsequently in the axons (rectangular panels below main image), in a time-dependent manner. The colocalization of PlexA4 and TrkAis presented using colocalization highlighter (white color). Scale bars: 5 µm (main panels), 2 µm (rectangular panels). (G) Time-dependence of mean colocalization coefficients in growth cones (GC) and axons (n=12 neurons from three independent cultures). Data are presented as mean±s.e.m. *P<0.05; **P<0.01 (one-way ANOVA compared to 0 min); #P<0.05; ##P<0.01 (one-way ANOVA compared to 3 min). Journal of Cell Science

1803 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1802-1814 doi:10.1242/jcs.184580

We used several TrkA mutants to determine the region of the et al., 2001; Zweifel et al., 2005). However, NGF neither altered the that interacted with PlexA4. Both full-length TrkA and TrkA localization of PlexA4 (Fig. 3A,B) nor induced the PlexA4 and mutant lacking the region from C1 to C2 (TrkA ΔC1-C2, Fig. 1B) TrkA colocalization in the growth cones and the axons (Fig. 3C,D). region interacted with PlexA4 (Fig. 1C). In contrast, a TrkA mutant These data clearly show that the interaction between PlexA4 and lacking the region from C1 to Ig1 region (TrkA ΔC1-Ig1, Fig. 1B) TrkA, and retrograde axonal transport of the PlexA4–TrkA complex did not interact with PlexA4, thereby indicating that the TrkA Ig1 occur in response to Sema3A, but not to NGF, thereby suggesting domain is required for the interaction between TrkA and PlexA4 distinct signaling pathways for Sema3A and NGF. (Fig. 1C). To further investigate the interaction between PlexA4 and TrkA, The interaction between PlexA4 and TrkA mediates we next performed immunostaining of PlexA4 and TrkA in cultured Sema3A-induced dendritic localization of GluA2 mouse dorsal root ganglion (DRG) neurons. Sema3A potently We have previously shown that the Sema3A retrograde signal induces axon repulsion or growth cone collapse of the NGF- regulates the dendritic localization of GluA2 in hippocampal sensitive neurons in the DRG (Messersmith et al., 1995). Anti- neurons by means of a PlexA4-dependent mechanism (Yamashita PlexA4 immunostaining signals were mainly observed in the et al., 2014). To examine the roles of TrkA in Sema3A signaling, we growth cones of neurons when not treated with Sema3A, which is analyzed dendritic localization of GluA2 in cultured hippocampal consistent with the results observed in cultured hippocampal neurons where TrkA was knocked down by means of small neurons (Yamashita et al., 2014). The specificity of the anti- interfering RNA (siRNA). The TrkA expression level in PlexA4 antibody used was confirmed by the absence of hippocampal neurons was undetectable by western blot analysis immunofluorescence signals in plexA4-deficient mouse brain (data not shown). Semi-quantitative RT-PCR analysis revealed that tissue (Suto et al., 2007). After Sema3A treatment, the fluorescent TrkA was expressed in the cultured hippocampal neurons. The clusters of PlexA4 moved along the axons towards the cell bodies in expression levels of Trka mRNA were reduced in the TrkA-siRNA- a time-dependent manner. A similar change in the localization of the treated cultured hippocampal neurons (Fig. 4A). The relative level TrkA signals was observed (Fig. 1D,E). These findings suggest that of Trka mRNA in TrkA-siRNA-treated neurons normalized to the Sema3A induces retrograde transport of the PlexA4 and TrkA along amount of glyceraldehyde 3-phosphate dehydrogenase (Gapdh) the axon. At 3 min after stimulation with Sema3A, PlexA4 and mRNA in control neurons (100.0±17.1) was 17.3±2.7% (mean± TrkA colocalization increased in the growth cone but not in the s.e.m.; n=3; P<0.01 by t-test). TrkA protein production was reduced axon. At 5 min after stimulation, the increased colocalization in the by introduction of TrkA siRNA into PC12 cells (Fig. 4B). As shown growth cone returned to the basal level, and was followed by an previously (Yamashita et al., 2014), Sema3A enhanced the increase in colocalization in distal axons (Fig. 1F,G). Taken immunostaining of GluA2 in dendrites (Fig. 4D,E). This effect together, these data suggest that, in cultured DRG neurons, was markedly attenuated in neurons transfected with the TrkA Sema3A induces formation of the PlexA4–TrkA complex in the siRNA plasmid and the attenuation was rescued by expressing a growth cone and retrograde transport of the complex along the axon. siRNA-resistant form of TrkA (Fig. 4C–E). Time-lapse imaging analysis further supports the notion that the To examine the possible role of TrkB, a major neurotrophin PlexA4–TrkA complex is retrogradely transported along the axons. receptor in the central (Bartkowska et al., 2010), we In DRG neurons coexpressing PlexA4–EGFP and TrkA–mCherry, performed siRNA-mediated knockdown of TrkB. We confirmed Sema3A induced colocalization of PlexA4–EGFP and TrkA– that the siRNA knockdown of TrkB markedly reduced the mCherry in the growth cones, and then retrograde transport of expression of TrkB (Fig. 5A) and suppressed the effect of brain- PlexA4–EGFP colocalized with TrkA–mCherry along the axons derived neurotrophic factor (BDNF) on primary formation (Fig. 2A,B). In contrast, in neurons coexpressing PlexA4–EGFP (Fig. 5B,C). In contrast to TrkA, TrkB knockdown did not affect and TrkAΔC1-Ig1–mCherry (Fig. 1B), Sema3A induced neither Sema3A-enhanced immunostaining of GluA2 (Fig. 5D,E). This colocalization of PlexA4–EGFP and TrkAΔC1-Ig1–mCherry in the result indicates that TrkA, but not TrkB, mediates the effect of growth cones, nor retrograde transport of PlexA4–EGFP and Sema3A on dendritic localization of GluA2, which is consistent TrkAΔC1-Ig1–mCherry (Fig. 2C,D). Time-lapse imaging was with the finding that PlexA4 interacts selectively with TrkA and not performed using TrkA K537A, a kinase-dead mutant (Fig. 1B) TrkB (Fig. 1A). These findings suggest that TrkA mediates the (Yano et al., 2001). Sema3A increased colocalization of PlexA4– Sema3A retrograde signal that regulates the dendritic GluA2 EGFP and TrkA-K537–mCherry in the growth cones but did not localization. induce retrograde transport of the colocalized signals (Fig. 2C,D). A similar result was obtained with a TrkAΔ464-483 deletion mutant PI3K–Akt at the distal axon mediates the retrograde (Fig. 1B), which lacks the interaction domain for the light chain of signaling of Sema3A cytoplasmic dynein (Fig. 2C,D) (Yano et al., 2001). These data The kinase activity of TrkA mediates Sema3A retrograde signal suggest that Sema3A induces an interaction between PlexA4 and along the axon in cultured DRG neurons (Fig. 2). Upon NGF TrkA in the growth cones and then induces retrograde transport of binding to TrkA, TrkA phosphorylates PLCγ1, Akt and the PlexA4–TrkA complex along the axons through the kinase and ERK1 and ERK2 (ERK1/2, also known as MAPK3 and MAPK1, dynein-binding activity of TrkA. respectively, and P44/42 MAPK) (Delcroix et al., 2003; Heerssen We next examined the effect of NGF on the localization of et al., 2004; Wu et al., 2007; Zweifel et al., 2005). To determine PlexA4 and TrkA in neurons. For this purpose, we performed which of these TrkA signaling pathways are involved in Sema3A double immunostaining with PlexA4 and TrkA after NGF retrograde signaling, we examined the phosphorylation levels of stimulation in cultured mouse DRG neurons. NGF treatment these proteins in cultured cortical neurons in the presence or absence induced a time-dependent change in localization of TrkA from the of Sema3A. The phosphorylation levels of Akt were increased growth cones to the axons after NGF treatment (Fig. 3A,B). This 5 min after Sema3A treatment and returned to the basal level 10 min localization change might be caused by NGF-induced retrograde after stimulation; after 30 min the levels decreased to below the transport of the signaling endosome (Delcroix et al., 2003; Yano levels observed in the absence of Sema3A (Fig. 6A,B). The Journal of Cell Science

1804 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1802-1814 doi:10.1242/jcs.184580

Fig. 2. Sema3A induces colocalization of PlexA4–EGFP and TrkA–mCherry, and retrograde transport of PlexA4–EGFP and TrkA–mCherry colocalized particles. (A) Time-lapse imaging of PlexA4–EGFP and TrkA–mCherry in growth cones and axons of living DRG neurons. Sema3A induced colocalization of full-length of PlexA4 and full-length TrkA in the growth cone and subsequently enhanced the retrograde transport of colocalized PlexA4 and TrkA signals along the axon (white arrowheads). Scale bar: 5 µm. (B) Mean number of colocalized PlexA4 and TrkA signals in the growth cones (upper panel), and the mean number of signals for retrograde transport of PlexA4–TrkA signals along the axon (lower panel) (n=12 neurons from three independent cultures). (C) Time-lapse imaging of PlexA4–EGFP and mutant forms of TrkA–mCherry in growth cones and axons of living DRG neurons. Scale bar: 5 µm. (D) Mean number of colocalization events in the growth cone from 60 to 120 s after Sema3A stimulation (left), and the mean number of signals for retrograde transport of colocalized PlexA4 and TrkA signals along the axon from 120 to 180 s after Sema3A stimulation (right) (n=12 neurons from three independent cultures). Each arrowhead in A and C is coded in color (white or light blue) and represents individual PlexA4–TrkA colocalized particles. *P<0.05, **P<0.01 (t-test compared to vehicle control in B, or by one-way ANOVA compared to full-length of TrkA in D). Data are presented as mean±s.e.m. Other details are as in the legend of Fig. 1. phosphorylation level of ERK2 was not altered by Sema3A initial Akt response might be different from those for the late-phase (Fig. 6A,B). A phosphorylation signal of PLCγ1 and ERK1 was response to Sema3A. not detectable under our experimental conditions (Fig. 6A). To elucidate the role of the Akt signaling pathway in the Sema3A has been reported to induce apoptosis and decrease Akt retrograde Sema3A signaling, we used a microfluidic chamber to phosphorylation in cultured differentiated phodocytes (Guan et al., examine the effect of local application of LY294002, an inhibitor of 2006). In our experiment, the level of phosphorylated Akt was phosphoinositide 3-kinase (PI3K)-dependent Akt phosphorylation. increased at 5 min, but decreased at 30 min after exposure to When applied locally to the distal axonal region, LY294002

Sema3A, respectively (Fig. 6B). Thus signaling mechanisms for the (50 µM) suppressed the Sema3A-induced increase in GluA2 Journal of Cell Science

1805 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1802-1814 doi:10.1242/jcs.184580

Fig. 3. NGF does not induce colocalization of the PlexA4 and TrkA. (A) After NGF treatment, the localization of fluorescence clusters (arrowhead) of TrkA moved along the axons towards the cell bodies in a time-dependent manner. Scale bar: 10 µm. (B) Time dependence of relative mean ratio of immunostaining intensity in the axon to that in the growth cone (n=12 neurons from three independent cultures). (C) NGF did not increase the number of double-positive clusters of anti-PlexA4 and anti-TrkA in the growth cones or in the axons. Scale bars: 5 µm (main panels), 2 µm (rectangular panels below main image). The colocalization of PlexA4 and TrkA is presented using colocalization highlighter (white). (D) The time dependence of mean colocalization coefficients in growth cones (GC) and axons (n=12 neurons from three independent cultures). Data are presented as mean±s.e.m. *P<0.05 (one-way ANOVA compared to 0 min). immunostaining levels in dendrites, whereas application of the knockdown attenuated Sema3A-induced dendritic branching, and inhibitor to the somatodendrite region did not suppress the effects of this effect was rescued by expression of a siRNA-resistant form of Sema3A (Fig. 6C,D). Sema3A did not enhance the phosphorylation TrkA (Fig. 7A,B). of ERK2 (Fig. 6A,B). Consistent with this, local application of We next performed in vivo knockdown of TrkA in the CA1 PD98059 (50 µM), an inhibitor of MEK-dependent MAPK hippocampus. CA1 hippocampal neurons were transfected with phosphorylation, to either the axonal or the somatodendrite region plasmids expressing TrkA siRNA and EGFP by means of in utero did not suppress the effects of Sema3A action (Fig. 6C,D). These electroporation at E17, and dendritic morphology was investigated at data suggest that Akt pathway activation in the distal axon mediates postnatal day (P)15. In CA1 hippocampal neurons transfected with the the retrograde Sema3A signaling. control siRNA plasmid, the first branch was observed mainly near the stratum lacunosum-moleculare region. The apical dendrites of TrkA mediates Sema3A-regulated dendritic patterning both neurons transfected with the TrkA siRNA plasmid branched shortly in vitro and in vivo after they entered the stratum radiatum region, showing proximal To further investigate the role of TrkA in biological activities of bifurcation (Fig. 7C,D). This phenotype is similar to that in neurons in Sema3A, we performed knockdown of TrkA and examined the which Sema3A-induced dendritic GluA2 localization is impaired, effect of Sema3A on dendritic branching in cultured hippocampal such as CA1 pyramidal neurons in sema3A-deficient mice (Nakamura neurons. Neurons at 6 days in vitro (DIV) were stimulated with et al., 2009; Yamashita et al., 2014). GluA2 plays an important role

Sema3A for 24 h (7 DIV), and then fixed for immunostaining. TrkA in dendritic development (Passafaro et al., 2003) by recruiting Journal of Cell Science

1806 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1802-1814 doi:10.1242/jcs.184580

Fig. 4. TrkA is required for Sema3A-induced dendritic localization of GluA2. (A,B) Levels of TrkA after siRNA knockdown of TrkA by introduction of pSilencer- based siRNA constructs targeting TrkA (TrkA KD) or control vector (Cont). (A) At 3 days after transfection, total RNA was extracted from cultured hippocampal neurons and then subjected to semi-quantitative reverse-transcription PCR analysis. Equal amounts of cDNA were used, as indicated by the internal standard (Gapdh). (B) At 3 days after transfection, total protein was extracted from PC12 cells and then subjected to western blot analysis. Equal amounts of protein were analyzed, as indicated by the loading control (anti-β-). (C) Characterization of the siRNA-resistant form of TrkA–EGFP in HEK293T cells. Expression of wild-type (TrkAwt–EGFP), but not the siRNA-resistant form of TrkA (TrkAres–EGFP), was decreased by siRNA against TrkA. (D) Knockdown of TrkA attenuated Sema3A-enhanced immunostaining of GluA2 in dendrites. This attenuation was rescued by expression of TrkAres–EGFP. Magnified images of the boxed areas in the main panels are shown below. Arrowheads indicate the axon. Scale bars: 50 µm (main panel), 10 µm (magnified panels). (E) Normalized mean immunostaining intensity of dendrite per neuron (n=47–59 neurons from three independent cultures). Data are presented as mean±s.e.m. *P<0.01 (two-way ANOVA compared to vehicle control).

N-cadherin through direct interaction (Saglietti et al., 2007). Hence, it of N-cadherin-regulated dendritic development causing a bifurcation is likely that a decrease in the density of GluA2 in distal dendrites due phenotype. These results suggest that TrkA is required for Sema3A to to an impairment of the Sema3A signaling might result in weakening regulate dendritic patterning both in vitro and in vivo. Journal of Cell Science

1807 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1802-1814 doi:10.1242/jcs.184580

Fig. 5. TrkB does not mediate Sema3A-induced dendritic localization of GluA2. (A) Efficacy of siRNA knockdown of TrkB. Cultured hippocampal neurons were transfected with pSilencer-based siRNA constructs targeting TrkB (TrkB KD) or control vector (Cont). After 3 days, total protein was extracted and then subjected to western blot analysis. Equal amounts of protein were analyzed as indicated by the loading control (anti-β-actin). (B) Knockdown of TrkB attenuated BDNF-induced formation of primary dendrites (arrowheads). Scale bar: 50 µm. (C) Normalized mean number of primary dendrites per neuron (n=38–41 neurons from two independent cultures). Arrowheads indicate primary dendrites. (D) Knockdown of TrkB did not attenuate Sema3A-enhanced immunostaining of GluA2 in dendrites. Arrowheads indicate axons. (E) Normalized mean immunostaining intensity of dendrites per neuron (n=57–70 neurons from three independent cultures). D shows magnified images of boxed areas in the main panels below. Scale bars: 50 µm (main panels), 10 µm (magnified panels). Data are presented as mean±s.e.m. *P<0.01 (two-way ANOVA compared to vehicle control).

DISCUSSION required for Sema3A-induced retrograde transport of PlexA4– We demonstrate that TrkA plays a crucial role in mediating TrkA. The retrograde transport of PlexA4–TrkA is a prerequisite for Sema3A-induced retrograde signaling by interacting with PlexA4. Sema3A to regulate dendritic patterning. This finding provides a

The kinase activity and dynein-binding activity of TrkA were new mode of crosstalk between the two signaling pathways of NGF Journal of Cell Science

1808 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1802-1814 doi:10.1242/jcs.184580

Fig. 6. PI3K–Akt signaling at the distal axon mediates Sema3A-induced dendritic localization of GluA2. (A) Sema3A increased the phosphorylation level of Akt at 5 min but decreased the level at 30 min. (B) Normalized mean relative intensity of phospho-Akt or phospho-ERK2 to total Akt (P-Akt/Akt) or total ERK2 (P-ERK2/ERK2), respectively (n=8 dishes from four independent cultures). (C) Pretreatment with LY294002 (LY), using the microfluidic culture system, in the axonal region (axon), but not in the somatodendritic region (S+D), attenuated the enhanced immunostaining of GluA2 caused by application of Sema3A applied to the axonal region. Pretreatment with PD98059 (PD) had no effect. Magnified images of the boxed areas in the main panels are shown below. Arrowheads indicate axons. (D) Normalized mean immunostaining intensity of dendrite per neuron (n=23–43 neurons from six independent cultures). Scale bars: 20 µm (main panels), 10 µm (magnified panels). Data are presented as mean±s.e.m. #P<0.05; ##P<0.01 (t-test compared to 0 min); *P<0.01 (two-way ANOVA compared to vehicle control). and Sema3A. Whether similar types of crosstalk between Reichardt, 2003; Tojima et al., 2011), it is likely that the Trk neurotrophin receptors and axon guidance molecules in general family of neurotrophin receptors might be involved in Sema3A- play some roles in the neuron wiring system is an intriguing issue to induced axonal transport. Indeed, we found that PlexA4 interacted be addressed. with TrkA and TrkC (Fig. 1). Sema3A induced colocalization of Sema3A induces anterograde and retrograde axonal transport PlexA4 and TrkA in cultured DRG neurons (Fig. 1). These through growth cone NRP1 (Goshima et al., 1999, 1997), but its colocalized signals were then retrogradely transported along the mechanism and physiological relevance have long been ill-defined axons in a manner requiring the dynein-binding and kinase activity (Hida et al., 2015, 2012; Li et al., 2004). We recently found that of TrkA (Figs 1 and 2). TrkA knockdown suppressed Sema3A- local application of Sema3A to distal axons induces internalization induced localization of GluA2 in cultured hippocampal neurons of Sema3A at the distal axons, and then retrograde transport of (Fig. 4). These findings together suggest that interaction between Sema3A along the axon. Sema3A also enhances retrograde axonal TrkA and PlexA4 in the axons mediates Sema3A retrograde transport of PlexA4, which then escorts the AMPA receptor GluA2 signaling from distal axon to somatodendrite. The signaling from the cell body to distal dendrites (Yamashita et al., 2014). This pathway of Sema3A, however, might be distinct from that of finding suggests that the Sema3A signal is transmitted from distal NGF. Unlike Sema3A, NGF did not induce colocalization of NGF axons to somatodendritic compartments by the signaling endosome. and PlexA4 (Fig. 3). Given that share some downstream effectors with An immunoprecipitation assay revealed that PlexA4 interacted neurotrophins that induce signaling endosome from distal axons with TrkA and TrkC but not with TrkB. Consistently, siRNA-

(Castellani et al., 2004; Henley and Poo, 2004; Huang and mediated knockdown of TrkA but not TrkB suppressed Sema3A- Journal of Cell Science

1809 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1802-1814 doi:10.1242/jcs.184580

Fig. 7. TrkA mediates Sema3A-induced dendritic branch formation in vitro and in vivo. (A) Knockdown of TrkA suppressed Sema3A-induced dendritic branching of cultured hippocampal neurons. This knockdown effect was rescued by expressing TrkAres–EGFP. Arrowheads indicate MAP2- positive dendritic branching points. Scale bar: 25 µm. (B) Normalized mean number of dendritic branching points per neuron (n=24–38 neurons from three independent cultures). (C) Compared to control-vector-electroporated neurons, apical dendrites branched shortly after they entered the stratum radiatum (sr) region in the TrkA knocked down neurons (arrowheads). sp, stratum pyramidale; slm, stratum lacunosum-moleculare. Scale bar: 50 µm. (D) The relative position of the first branch point of apical dendrites (n=111–123 neurons from three individual brains). The relative length of the stratum radiatum was set to 1 (position 0 is the border between the stratum pyramidale and stratum radiatum; position 1 is the border between the stratum radiatum and stratum lacunosum-moleculare). *P<0.01 (two-way ANOVA in B, or by t-test in D compared to vehicle control or control-vector-electroporated

neurons). Journal of Cell Science

1810 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1802-1814 doi:10.1242/jcs.184580 induced localization of GluA2 in dendrites (Figs 4 and 5). Although accumulates in the cytoplasm of the CA1 hippocampal pyramidal we could not rule out the possibility that TrkC is also involved, our neurons (Good et al., 2004). Here, it might induce neuronal data suggest that TrkA at least partially mediates the effect of degeneration because Sema3A can induce neuronal cell death (Ben- Sema3A to regulate dendritic localization of GluA2. Our deletion Zvi et al., 2006). By contrast, early and predominant loss of NGF- mutant study revealed that TrkA Ig1 domain was required for the sensitive basal forebrain cholinergic neurons is the major functional interaction between TrkA and PlexA4 (Fig. 1C). In addition to the basis for the cognitive impairment in Alzheimer’s diseases (Bartus Ig2 domain, a neurotrophin-binding domain, the Ig1 domain is also et al., 1980; Hefti and Weiner, 1986), and compensatory important for Trk receptor dimerization (Arevalo et al., 2000). hyperactivity of the NGF–TrkA complex has been implicated in Given that Trk receptor dimerization is vital for its kinase activity the pathogenesis of Alzheimer’s disease (Zhang et al., 2014). Thus, (Lemmon and Schlessinger, 2010), the interaction of PlexA4 with it is possible that disruption in the balance between Sema3A and TrkA Ig1 domain might activate TrkA. NGF occurs, thereby causing bias towards cell death signaling NGF-bound TrkA is internalized into signaling endosomes that pathways, further accelerating neurodegenerative process in the include components of the Ras-MAPK, PLC-γ and PI3K pathways, brains affected with Alzheimer’s disease. which are activated by the kinase activity of TrkA (Delcroix et al., In conclusion, we demonstrate a new mechanism by which 2003). Our study showed that Sema3A-induced colocalization of Sema3A induces a retrograde signaling that regulates dendritic PlexA4 and TrkA K537A mutant. However, these colocalized GluA2 localization and branching. The interaction between PlexA4 particles were stationary, and were not transported along the axons and TrkA provides evidence for a crosstalk between Sema3A and (Fig. 2). This result suggests that upon Sema3A binding to its NGF that might play a crucial role in neuronal network formation, receptor complex, TrkA interacts with the Sema3A signaling neuronal survival and other neuronal functions. complex, thereafter inducing internalization of the complex through its kinase activity. Among signaling molecules downstream of MATERIALS AND METHODS TrkA, Sema3A enhanced the phosphorylation of Akt in primary Cell culture and recombinant protein culture of cortical neurons (Fig. 6). In addition, the application of Dorsal root ganglion (DRG) neurons from chicks of either sex at embryonic LY294002, a PI3K–Akt inhibitor, to the axonal regions but not day (E7) or mouse of either sex at E12 were cultured for 1 day on a 0.01% ’ somatodendritic regions suppressed Sema3A-induced localization poly-L-lysine-coated dish in Ham s F-12 medium (Wako Pure Chemicals) of GluA2 in dendrites of cultured hippocampal neurons (Fig. 6). Akt supplemented with 10% fetal bovine serum (FBS) and 10 ng/ml NGF (Wako Pure Chemicals). To starve the cells of NGF, we replaced the culture can influence various aspects of neuronal development, such as medium after 1 day in vitro (DIV) with Leibovitz’s L-15 Medium (Life neurite elongation, branching and dendritic maturation (Manning Technologies) and cultured the cells for 4 h. The cultured DRG neurons and Cantley, 2007; Read and Gorman, 2009). Although its were treated with Sema3A (5 nM) or NGF (40 ng/ml) and then fixed for downstream substrates involved are unknown, the PI3K–Akt staining. Hippocampal or cortical neurons from Wistar rats of either sex at pathway, which is activated through the kinase activity of TrkA, E18 (Charles River Laboratories) were cultured for 3 days (unless otherwise might mediate local signaling at the growth cone elicited by stated) on a 0.01% poly-L-lysine-coated dish in neurobasal medium (Life Sema3A. Indeed, inhibition of PI3K prevents accumulation of Technologies) supplemented with 2% B27 supplement (Life Technologies) NGF–TrkA-containing endosomes in the cell body (Kuruvilla et al., and 0.5 mM L-glutamine (Sigma). 2000). Recombinant Sema3A was prepared using TALON Metal Affinity Resin We have previously shown that interaction between PlexAs and (Takara Bio) from its histidine-tagged form. The activity of Sema3A was estimated by the growth cone collapse assay as previously described GluA2 is required for proper dendritic patterning in CA1 (Goshima et al., 1999). One unit of collapsing activity is defined as the hippocampal pyramidal neurons. PlexA interacts with GluA2 at amount required to collapse 50% of growth cones of chick DRG neurons in the PlexA-IPT domain in somatodendritic regions. Overexpression a 1 ml culture (Goshima et al., 1995). All procedures on animals were of PlexA-IPT causes a similar phenotypic defect to that seen in performed according to Policies on the Use of Animals and Humans in sema3A-deficient mice in vivo (Yamashita et al., 2014). In the Neuroscience Research and the guidelines outlined in the Institutional present study, knockdown of TrkA suppressed Sema3A-induced Animal Care and Use Committee of Yokohama City University School of dendritic branching in vitro (Fig. 7). We further found that TrkA Medicine. knockdown caused a phenotypic defect in dendrite branching sema3A Immunoprecipitation and immunoblot analysis similar to that observed in -deficient and PlexA-IPT- 5 overexpressing neurons (Fig. 7). These findings are consistent with HEK293T cells were seeded at 4.0×10 cells/dish in a 6-cm dish. After 1 day, the cells were transfected with a Flag–PlexA4 (mouse) expression the idea that TrkA is required for retrograde Sema3A signaling for Δ in vivo vector together with a expression vector for full-length, C1-C2 (lacking proper dendritic branching, and also support relevance for amino acids 90–195 of mouse TrkA) or ΔC1-Ig1 (lacking amino acids 90– the interaction in the developing nervous system. Expression of 296 of mouse TrkA) variants of TrkA–EGFP; full-length of TrkB–EGFP TrkA in the hippocampus is a controversial topic (Fortress et al., (mouse); or full-length of TrkC–EGFP (mouse) using FuGENE6 2011). Our present findings also showed that the TrkA expression Transfection Reagent (Promega). After 48 h, cells were lysed with level was undetectable by western blot analysis (data not shown), immunoprecipitation buffer [20 mM Tris-HCl pH 8.0, 150 mM NaCl, ρ ρ and was detectable by semi-quantitative RT-PCR analysis in the 10 mM NaF, 1 mM Na3VO4, 1 mM EDTA, 50 µM -APMSF ( - hippocampal neurons, suggesting a low expression level of TrkA in amidinophenylmethanesulfonyl fluoride) and 1% Nonidet P-40] and then the hippocampal neurons (Fig. 4A). However, our current study immunoprecipitated with 1 µg of anti-Flag mouse monoclonal antibody demonstrates a crucial role of TrkA in mediating Sema3A actions (F1804, Sigma). The samples were used for immunoblot analysis with anti- for dendritic branching of the CA1 hippocampal neurons (Fig. 7). GFP mouse monoclonal (sc-9996, 1:2500 dilution, Santa Cruz Biotechnology) and anti-Flag mouse monoclonal antibodies (1:5000 Several lines of evidence suggest that Sema3A signaling ’ dilution). is involved in the pathogenesis of Alzheimer s disease (Good To determine the phosphorylation levels of PLCγ1, Akt and ERK1/2, et al., 2004; Isono et al., 2013; Uchida et al., 2005; Yoshida we stimulated cultured cortical neurons with Sema3A (5 nM) and then lysed et al., 1998). During the progression of Alzheimer’s diseases, the neurons with immunoprecipitation buffer. The resulting samples were immunohistochemical studies have revealed that Sema3A used for immunoblot analysis with anti-PLCγ1 (#2822, 1:1000 dilution, Journal of Cell Science

1811 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1802-1814 doi:10.1242/jcs.184580

Cell Signaling), anti-phospho-PLCγ1 (Tyr783, #2821, 1:1000 dilution, Cell transcription polymerase chain reaction (RT-PCR) of Trka mRNA (forward Signaling), anti-Akt (#9272, 1:1000 dilution, Cell Signaling), anti- primer 5′-CCTACTGAGGGCAAAGGCTC-3′; reverse primer 5′-TGGG- phospho-Akt (Ser473 #9271, 1:1000 dilution, Cell Signaling), anti- ACCGGAGGAAACGGTT-3′). Expression of Gapdh mRNA was used as ERK1/2 (#06-182, 1:1000 dilution, Millipore) and anti-phospho-ERK1/2 an internal standard (forward primer; 5′-GGGTGTGAACCACGAGAAA- (Thr183 and Tyr185, V8031, 1:1000 dilution, Promega) rabbit polyclonal T-3′, reverse primer 5′-ACTGTGGTCATGAGCCCTTC-3′). The efficacy antibodies. of the siRNA expression after plasmid transfection by Lipofectamine 2000 transfection reagent (Invitrogen) into PC12 cells was evaluated by Immunocytochemistry immunoblotting analysis with anti-TrkA (1:1000 dilution) and anti-β- Immunocytochemistry was performed by a standard protocol under actin mouse monoclonal antibodies (A5316, 1:10,000 dilution; Sigma). To permeabilization with 0.1% Triton X-100. The antibodies used for rescue siRNA-mediated knockdown of TrkA, we generated a mutant that staining the cultured neurons were as follows: anti-TrkA rabbit polyclonal did not change the amino acid sequence but was no longer targeted by the antibodies (#06-574 1:1000 dilution, Millipore), anti-PlexA4 hamster siRNA; specifically, the section of mouse TrkA mRNA (NM_001033124) monoclonal antibodies (1:500 dilution) (Suto et al., 2007), anti-GluA2 corresponding to nucleotides 1506–1530 was replaced with CCCACAAT- mouse monoclonal antibodies (MAB397, 1:500 dilution, Millipore), anti- ATTTTTCCGATACCTGT. This mutant was tagged with EGFP and MAP2 mouse monoclonal antibodies (M1406, 1:1000 dilution, Sigma), co-transfected with the siRNA vector. anti-GFP rabbit polyclonal antibodies (598, 1:1000 dilution, MBL). For knockdown of TrkB, we used RNAi duplex oligonucleotides (MBL). Immunostained neurons were analyzed with a laser scanning microscope The target sequence was 5′-ATTCACTGTACATAGACAAAGGCCTTA- (LSM510) equipped with a water-immersed 40× objective lens (C- 3′. Stealth RNA™ siRNA Negative Control (Invitrogen) was used as the Apochromat/1.2W corr) and with an Axioplan 2 imaging microscope negative control. The efficacy of TrkB siRNA transfection with Amaxa™ (Carl Zeiss) Rat Neuron Nucleofector Kit into cultured hippocampal neurons was evaluated by immunoblot analysis with anti-TrkB mouse monoclonal Quantitative immunocytochemistry (#610101, 1:2500 dilution, BD Transduction Laboratories) and anti-β-actin ImageJ software was used to measure the mean intensities of mouse monoclonal antibodies. immunostaining by anti-PlexA4, anti-TrkA or anti-GluA2 antibodies. To To assess Sema3A-enhanced immunostaining of GluA2, we used the quantify the change in the distributions of PlexA4 and TrkA in cultured DRG Amaxa™ Rat Neuron Nucleofector Kit to co-transfect neurons with siRNA neurons, we measured the mean intensity of PlexA4 or TrkA in the growth together with EGFP or EGFP-tagged siRNA-resistant mutant expression cone (within 10 µm of the axon tip) and axon (within 25–35 µm of the vectors. On 3 DIV, immunocytochemistry was performed after stimulation initiation point) according to the morphology observed by differential with Sema3A (5 nM) for 30 min. To examine the Sema3A-induced interference contrast microscopy. The ratio of axon immunostaining dendritic branching, we used Lipofectamine 2000 to co-transfect neurons intensity to the growth cone immunostaining intensity was then calculated. with TrkA siRNA expression vector together with EGFP or siRNA-resistant To quantify the colocalization of PlexA4 and TrkA immunoreactive signals TrkA-expressing vector on 3DIV. On 6 DIV, bath application of Sema3A in growth cones or axons of cultured DRG neurons, we calculated Pearson’s (5 nM) was performed. After incubation with Sema3A for 24 h, neurons correlation coefficients (colocalization coefficients) using LSM 5 Image were fixed for immunostaining with anti-GFP rabbit polyclonal and anti- software. This coefficient is +1 for perfect correlation, 0 for no correlation MAP2 mouse monoclonal antibodies. The number of MAP2-positive and −1 for perfect anti-correlation. The colocalization of PlexA4 and TrkA dendritic branch points of EGFP-positive neurons was then counted. To in Figs 1F and 3C was presented using colocalization highlighter (ImageJ). examine the BDNF-induced primary dendrite formation, we co-transfected The colocalization signal is visualized as white color. neurons with TrkB siRNA together with EGFP-expressing vector with the Cultured hippocampal neurons were treated with Sema3A (5 nM) for Amaxa™ Rat Neuron Nucleofector Kit. On 3 DIV, BDNF (50 ng/ml) was 30 min and then fixed for staining with an anti-GluA2 antibody. To obtain applied, and the neurons were cultured for an additional 3 days before the mean intensity of dendrite per neuron, we normalized the fixation. To visualize primary dendrites, we immunostained cultured immunofluorescence level of each EGFP-positive dendrite by the volume hippocampal neurons with anti-GFP rabbit polyclonal and anti-MAP2 of dendrites. The longest neurite was defined as the axon (indicated by mouse monoclonal antibodies. The MAP2-positive primary dendrites of arrowheads in figures) (Yamashita et al., 2014). To obtain the mean EGFP-positive neurons were then counted. intensity of neurite per neuron, immunostaining intensity of each individual neurite was summed and then divided by the number of neurites. Intensity Cell culture in microfluidic chamber measurement was performed by a researcher who was blind to the treatment. Dissociated hippocampal neurons were first transfected with an EGFP- expression vector using Amaxa™ Rat Neuron Nucleofector Kit. Neurons Time-lapse observation were then plated onto one of the reservoirs of AXIS™ Axon Isolation Cultured chick DRG neurons were transfected with expression plasmids Device (AX150, Millipore) and cultured for 3 days. For local stimulation, encoding PlexA4–EGFP and TrkA–mCherry using an Amaxa Basic Neuron the reservoirs for the axonal and somatodendritic compartments were SCN Nucleofector Kit (Lonza), according to the manufacturer’s maintained in fluidic isolation by using a low hydrostatic pressure, which instructions. After 1 day, a growth cone was imaged at 1.5 s intervals enabled us to locally stimulate the axonal or somatodendritic region by using a LSM5 PASCAL with an oil-immersed 63× objective lens (C- isolating these two parts (Park et al., 2006). The EGFP-positive neurons that Apochromat/1.4F corr) and an Axioplan 2 imaging microscope (Carl Zeiss). extended long neurites into the axonal regions were used for the analysis The mean number of PlexA4 and TrkA colocalization events in each growth (Yamashita et al., 2014). cone and the mean number of colocalized particle retrogradely transported from growth cone to axon were counted with LSM 5 Image software. The In utero electroporation colocalization of PlexA4 and TrkA was presented using a colocalization Timed-pregnant Wistar rats were anesthetized at E17 with isoflurane and the highlighter (white color). uterine horns were exposed by means of a laparotomy. A total of 5 μgof plasmid DNA mixture (EGFP, 0.25 mg/ml; TrkA siRNA or control siRNA siRNA vector: 1 mg/ml) containing Fast Green was injected through the uterine For knockdown of TrkA, we used siRNA based on pSilencer 3.0 (Life wall into one of the lateral ventricles of the embryo. Then the head of the Technologies). The target sequence was 5′-GCCCACAGTACTTCAGTG- embryo was electroporated with forceps-type electrodes (BEX) [50-ms ATACCTGT-3′. A sequence that did not share with electric pulses of 100 V were delivered five times at intervals of 950 mc with any rat mRNAs (5′-CAGTCGCGTTTGCGACTGG-3′) was used as the a square electroporator (BEX)]. To examine the dendritic morphology of negative control (control vector). The efficacy of siRNA expression plasmid CA1 hippocampal neurons, we intracardially perfused male or female pups transfection by means of Amaxa™ Rat Neuron Nucleofector Kit into at P15 with 4% paraformaldehyde. The morphology of apical dendrites was cultured hippocampal neurons was evaluated by semi-quantitative reverse determined by visualization of EGFP florescence with an LSM510 with a Journal of Cell Science

1812 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1802-1814 doi:10.1242/jcs.184580

20× (Plan-NEO FLUAR 20×/50), and the z-axis series of images were Goshima, Y., Kawakami, T., Hori, H., Sugiyama, Y., Takasawa, S., Hashimoto, superimposed onto one image by means of LSM 5 Image software. The Y., Kagoshima-Maezono, M., Takenaka, T., Misu, Y. and Strittmatter, S. M. relative position of the first branch point was quantified. The relative length (1997). A novel action of collapsin: collapsin-1 increases antero- and retrograde of the stratum radiatum was set to 1. Position 0 is the border between the axoplasmic transport independently of growth cone collapse. J. Neurobiol. 33, 316-328. stratum pyramidale and stratum radiatum, and position 1 is the border Goshima, Y., Hori, H., Sasaki, Y., Yang, T., Kagoshima-Maezono, M., Li, C., between the stratum radiatum and the stratum lacunosum-moleculare Takenaka, T., Nakamura, F., Takahashi, T., Strittmatter, S. M. et al. (1999). (Yamashita et al., 2014). Growth cone neuropilin-1 mediates collapsin-1/Sema III facilitation of antero- and retrograde axoplasmic transport. J. Neurobiol. 39, 579-589. Statistical significance Guan, F., Villegas, G., Teichman, J., Mundel, P. and Tufro, A. (2006). Autocrine class 3 semaphorin system regulates slit diaphragm proteins and podocyte Data are shown as mean±s.e.m. The statistical significance of the results was ’ t survival. Kidney Int. 69, 1564-1569. analyzed using one-way or two-way ANOVA or Student s -test. All Harrington, A. W. and Ginty, D. D. (2013). Long-distance retrograde neurotrophic statistics were performed assuming Gaussian distribution. factor signalling in neurons. Nat. Rev. Neurosci. 14, 177-187. Heerssen, H. M., Pazyra, M. F. and Segal, R. A. (2004). Dynein motors transport Acknowledgements activated Trks to promote survival of target-dependent neurons. Nat. Neurosci. 7, We thank Drs K. Takei and F. Nakamura for useful discussions and constructive 596-604. input and T. Gotoh, A. Yamamoto, S. Chen, and M. Ogawara for technical Hefti, F. and Weiner, W. J. (1986). Nerve growth factor and Alzheimer’s disease. assistance. Ann. Neurol. 20, 275-281. Henley, J. and Poo, M.-M. (2004). Guiding neuronal growth cones using Ca2+ Competing interests signals. Trends Cell Biol. 14, 320-330. Hida, T., Yamashita, N., Usui, H., Nakamura, F., Sasaki, Y., Kikuchi, A. and The authors declare no competing or financial interests. Goshima, Y. (2012). GSK3beta/axin-1/beta-catenin complex is involved in semaphorin3A signaling. J. Neurosci. 32, 11905-11918. Author contributions Hida, T., Nakamura, F., Usui, H., Takeuchi, K., Yamashita, N. and Goshima, Y. N.Y., M.Y. and Y.G. designed research; N.Y. and M.Y. performed research; F.S. (2015). Semaphorin3A-induced axonal transport mediated through provided anti-PlexA4 antibody; N.Y. and Y.G. wrote the paper. phosphorylation of Axin-1 by GSK3beta. Brain Res. 1598, 46-56. Huang, E. J. and Reichardt, L. F. (2003). Trk receptors: roles in neuronal signal Funding transduction. Annu. Rev. Biochem. 72, 609-642. This work was supported by Grants-in-Aid for Scientific Research in a Priority Area to Isono, T., Yamashita, N., Obara, M., Araki, T., Nakamura, F., Kamiya, Y., Alkam, Y.G. [grant number 17082006], Targeted Proteins Research Program to Y.G. [grant T., Nitta, A., Nabeshima, T., Mikoshiba, K. et al. (2013). Amyloid-beta25-35 number 0761890004], and by Global COE Program, Innovative Integration between induces impairment of cognitive function and long-term potentiation through Medicine and Engineering Based on Information Communication Technology to phosphorylation of collapsin response mediator protein 2. Neurosci. Res. 77, M.Y. and Y.G. [grant number 1542140002], and Creation of Innovation Centers for 180-185. Advanced Interdisciplinary Research Areas Program in the Project for Developing Kuruvilla, R., Ye, H. and Ginty, D. D. (2000). Spatially and functionally distinct roles Innovation Systems to Y.G. [grant number 42890001] from the Ministry of Education, of the PI3-K effector pathway during NGF signaling in sympathetic neurons. Science, Sports and Culture; and a Grant-in-Aid for Young Scientists (B) [grant Neuron 27, 499-512. number 21700411] and a Grant-in-Aid for Scientific Research (C) [grant number Lemmon, M. A. and Schlessinger, J. (2010). Cell signaling by receptor tyrosine 24500444] from the Japan Society for the Promotion of Science (JSPS) to N.Y. N.Y. kinases. Cell 141, 1117-1134. received a JSPS postdoctoral fellowship for research abroad. Li, C., Sasaki, Y., Takei, K., Yamamoto, H., Shouji, M., Sugiyama, Y., Kawakami, T., Nakamura, F., Yagi, T., Ohshima, T. et al. (2004). Correlation between semaphorin3A-induced facilitation of axonal transport and local activation of a References translation initiation factor eukaryotic translation initiation factor 4E. J. Neurosci. Arevalo, J. C., Conde, B., Hempstead, B. L., Chao, M. V., Martin-Zanca, D. and 24, 6161-6170. Perez, P. (2000). TrkA immunoglobulin-like ligand binding domains inhibit Manning, B. D. and Cantley, L. C. (2007). AKT/PKB signaling: navigating spontaneous activation of the receptor. Mol. Cell. Biol. 20, 5908-5916. downstream. Cell 129, 1261-1274. Bartkowska, K., Turlejski, K. and Djavadian, R. L. (2010). Neurotrophins and their Messersmith, E. K., Leonardo, E. D., Shatz, C. J., Tessier-Lavigne, M., receptors in early development of the mammalian nervous system. Acta Goodman, C. S. and Kolodkin, A. L. (1995). Semaphorin III can function as a Neurobiol. Exp. 70, 454-467. selective chemorepellent to pattern sensory projections in the spinal cord. Neuron Bartus, R. T., Dean, R. L., Goas, J. A. and Lippa, A. S. (1980). Age-related 14, 949-959. changes in passive avoidance retention: modulation with dietary choline. Science Morita, A., Yamashita, N., Sasaki, Y., Uchida, Y., Nakajima, O., Nakamura, F., 209, 301-303. Yagi, T., Taniguchi, M., Usui, H., Katoh-Semba, R. et al. (2006). Regulation of Ben-Zvi, A., Yagil, Z., Hagalili, Y., Klein, H., Lerman, O. and Behar, O. (2006). dendritic branching and spine maturation by semaphorin3A-Fyn signaling. Semaphorin 3A and neurotrophins: a balance between apoptosis and survival J. Neurosci. 26, 2971-2980. signaling in embryonic DRG neurons. J. Neurochem. 96, 585-597. Nakamura, F., Ugajin, K., Yamashita, N., Okada, T., Uchida, Y., Taniguchi, M., Campbell, D. S., Regan, A. G., Lopez, J. S., Tannahill, D., Harris, W. A. and Holt, Ohshima, T. and Goshima, Y. (2009). Increased proximal bifurcation of CA1 C. E. (2001). Semaphorin 3A elicits stage-dependent collapse, turning, and pyramidal apical dendrites in sema3A mutant mice. J. Comp. Neurol. 516, branching in Xenopus retinal growth cones. J. Neurosci. 21, 8538-8547. 360-375. Castellani, V., Falk, J. and Rougon, G. (2004). Semaphorin3A-induced receptor Park, J. W., Vahidi, B., Taylor, A. M., Rhee, S. W. and Jeon, N. L. (2006). endocytosis during axon guidance responses is mediated by L1 CAM. Mol. Cell. Microfluidic culture platform for neuroscience research. Nat. Protoc. 1, Neurosci. 26, 89-100. 2128-2136. Delcroix, J.-D., Valletta, J. S., Wu, C., Hunt, S. J., Kowal, A. S. and Mobley, W. C. Passafaro, M., Nakagawa, T., Sala, C. and Sheng, M. (2003). Induction of dendritic (2003). NGF signaling in sensory neurons: evidence that early endosomes carry spines by an extracellular domain of AMPA receptor subunit GluR2. Nature 424, NGF retrograde signals. Neuron 39, 69-84. 677-681. Fortress, A. M., Buhusi, M., Helke, K. L. and Granholm, A.-C. (2011). Cholinergic Polleux, F., Morrow, T. and Ghosh, A. (2000). Semaphorin 3A is a chemoattractant degeneration and alterations in the TrkA and p75NTR balance as a result of pro- for cortical apical dendrites. Nature 404, 567-573. NGF injection into aged rats. J. Aging Res. 2011, 460543. Raper, J. A. (2000). Semaphorins and their receptors in vertebrates and Fournier, A. E., Nakamura, F., Kawamoto, S., Goshima, Y., Kalb, R. G. and invertebrates. Curr. Opin Neurobiol 10, 88-94. Strittmatter, S. M. (2000). Semaphorin3A enhances endocytosis at sites of Read, D. E. and Gorman, A. M. (2009). Involvement of Akt in neurite outgrowth. receptor-F-actin colocalization during growth cone collapse. J. Cell Biol. 149, Cell. Mol. Life Sci. 66, 2975-2984. 411-422. Saglietti, L., Dequidt, C., Kamieniarz, K., Rousset, M.-C., Valnegri, P., Good, P. F., Alapat, D., Hsu, A., Chu, C., Perl, D., Wen, X., Burstein, D. E. Thoumine, O., Beretta, F., Fagni, L., Choquet, D., Sala, C. et al. (2007). andKohtz,D.S.(2004). A role for semaphorin 3A signaling in the Extracellular interactions between GluR2 and N-cadherin in spine regulation. degeneration of hippocampal neurons during Alzheimer’s disease. Neuron 54, 461-477. J. Neurochem. 91, 716-736. Sasaki, Y., Cheng, C., Uchida, Y., Nakajima, O., Ohshima, T., Yagi, T., Goshima, Y., Nakamura, F., Strittmatter, P. and Strittmatter, S. M. (1995). Taniguchi, M., Nakayama, T., Kishida, R., Kudo, Y. et al. (2002). Fyn and Collapsin-induced growth cone collapse mediated by an intracellular protein Cdk5 mediate semaphorin-3A signaling, which is involved in regulation of dendrite

related to UNC-33. Nature 376, 509-514. orientation in cerebral cortex. Neuron 35, 907-920. Journal of Cell Science

1813 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1802-1814 doi:10.1242/jcs.184580

Shelly, M., Cancedda, L., Lim, B. K., Popescu, A. T., Cheng, P.-L., Gao, H. and Yamane, M., Yamashita, N., Yamamoto, H., Iizuka, A., Shouji, M., Usui, H. and Poo, M.-M. (2011). Semaphorin3A regulates neuronal polarization by Goshima, Y. (2012). Semaphorin3A facilitates axonal transport through a local suppressing axon formation and promoting dendrite growth. Neuron 71, 433-446. calcium signaling and tetrodotoxin-sensitive voltage-gated sodium channels. Suto, F., Murakami, Y., Nakamura, F., Goshima, Y. and Fujisawa, H. (2003). Biochem. Biophys. Res. Commun. 422, 333-338. Identification and characterization of a novel mouse plexin, plexin-A4. Mech. Dev. Yamashita, N., Usui, H., Nakamura, F., Chen, S., Sasaki, Y., Hida, T., Suto, F., 120, 385-396. Taniguchi, M., Takei, K. and Goshima, Y. (2014). Plexin-A4-dependent Suto, F., Tsuboi, M., Kamiya, H., Mizuno, H., Kiyama, Y., Komai, S., Shimizu, M., retrograde semaphorin 3A signalling regulates the dendritic localization of Sanbo, M., Yagi, T., Hiromi, Y. et al. (2007). Interactions between plexin-A2, GluA2-containing AMPA receptors. Nat. Commun. 5, 3424. plexin-A4, and semaphorin 6A control lamina-restricted projection of hippocampal Yamashita, N., Aoki, R., Chen, S., Jitsuki-Takahashi, A., Ohura, S., Kamiya, H., mossy fibers. Neuron 53, 535-547. Goshima, Y. (2016). Voltage-gated calcium and sodium channels mediate Tojima, T., Hines, J. H., Henley, J. R. and Kamiguchi, H. (2011). Second Sema3A retrograde signaling that regulates dendritic development. Brain Res. messengers and membrane trafficking direct and organize growth cone steering. 1631, 127-136. Nat. Rev. Neurosci. 12, 191-203. Yano, H., Lee, F. S., Kong, H., Chuang, J., Arevalo, J., Perez, P., Sung, C. and Tran, T. S., Kolodkin, A. L. and Bharadwaj, R. (2007). Semaphorin regulation of cellular morphology. Annu. Rev. Cell Dev. Biol. 23, 263-292. Chao, M. V. (2001). Association of Trk neurotrophin receptors with components of Uchida, Y., Ohshima, T., Sasaki, Y., Suzuki, H., Yanai, S., Yamashita, N., the cytoplasmic dynein motor. J. Neurosci. 21, RC125. Nakamura, F., Takei, K., Ihara, Y., Mikoshiba, K. et al. (2005). Semaphorin3A Yoshida, H., Watanabe, A. and Ihara, Y. (1998). Collapsin response mediator ’ signalling is mediated via sequential Cdk5 and GSK3beta phosphorylation of protein-2 is associated with neurofibrillary tangles in Alzheimer s disease. J. Biol. CRMP2: implication of common phosphorylating mechanism underlying axon Chem. 273, 9761-9768. guidance and Alzheimer’s disease. Cells 10, 165-179. Zhang, Q., Descamps, O., Hart, M. J., Poksay, K. S., Spilman, P., Kane, D. J., Wu, C., Ramirez, A., Cui, B., Ding, J., Delcroix, J.-D. M., Valletta, J. S., Liu, J.-J., Gorostiza, O., John, V. and Bredesen, D. E. (2014). Paradoxical effect of TrkA Yang, Y., Chu, S. and Mobley, W. C. (2007). A functional dynein- inhibition in Alzheimer’s disease models. J. Alzheimers Dis. 40, 605-617. network is required for NGF signaling through the Rap1/MAPK pathway. Traffic 8, Zweifel, L. S., Kuruvilla, R. and Ginty, D. D. (2005). Functions and mechanisms of 1503-1520. retrograde neurotrophin signalling. Nat. Rev. Neurosci. 6, 615-625. Journal of Cell Science

1814