A functional coupling between CRMP1 and Na_v1.7 for retrograde propagation of Semaphorin3A signaling

Neural circuit formation requires the orchestration of multiple developmental events such as cell fate specification, cell migration, axon guidance, synaptic target selection, synaptogenesis and the regulation of neuronal activities. Semaphorin3A (Sema3A), a secreted type of repulsive axon guidance molecule, is involved in some of these processes, including growth cone collapse or axon repulsion in dorsal root ganglion (DRG) neurons through microtubule and actin cytoskeleton reorganization (Fan and Raper, 1995; Goshima et al., 2002; Tran et al., 2007; Zhou et al., 2008). A complex of neuropilin-1 (NRP1) and plexinA proteins mediates Sema3A signals. NRP proteins and plexinA proteins are ligand-binding and signal-transducing subunits of class-3 semaphorin receptor complexes, respectively (Takahashi et al., 1999; Tamagnone et al., 1999). Sema3A stimulation enhances the interaction between PlexA proteins and collapsin response mediator proteins (CRMPs) (Schmidt et al., 2008), which were identified as an intracellular mediator of Sema3A signaling (Goshima et al., 1995). The sequential phosphorylation of CRMP1 and/or CRMP2 by cyclin-dependent kinase 5 (Cdk5) and GSK3β is necessary for the growth cone collapse through microtubule reorganization (Nakamura et al., 2014; Uchida et al., 2005; Yamashita et al., 2007).

Other biological actions of Sema3A include regulation of dendritic patterning, synapse maturation, neural cell polarity and axonal transport (Goshima et al., 1999, 1997; Li et al., 2004; Nakamura et al., 2009; Sasaki et al., 2002; Tran et al., 2009; Yamashita et al., 2007, 2014). Sema3A-induced axonal transport could play an essential role in mediating some of the biological activities of Sema3A (Goshima et al., 2016; Yamashita et al., 2014, 2016c). Sema3A signaling at the axonal growth cone is propagated towards the cell body by retrograde axonal transport and drives the AMPA receptor GluA2 to the distal dendrites, thereby regulating dendritic development (Yamashita et al., 2014). Sema3A induces dynein-dependent retrograde axonal transport of a tropomyosin kinase A (TrkA)–PlexA4 complex along the axons (Yamashita et al., 2016c). To achieve its biological activity, Sema3A and its signaling complex PlexA4–TrkA must be retrogradely transported along the axon from the growth cone to the cell body. For example, upon stimulation with Sema3A, even a kinase-dead mutant TrkA (K537A), as well as wild-type (WT) TrkA, is capable of colocalizing with PlexA4 in the growth cones. However, the kinase and dynein-binding activities of TrkA are required for Sema3A-induced retrograde transport of the PlexA4–TrkA complex. The kinase activity of TrkA is also required for Sema3A to induce dendritic localization of GluA2 and branching since expression of the K537A mutant does not rescue the knockdown effect of TrkA (Yamashita et al., 2016c). In addition, ion signaling plays an essential role in this transport process (Yamane et al., 2012; Yamashita et al., 2016a). Notably, the facilitation of axonal transport by Sema3A entirely depends on tetrodotoxin (TTX)-sensitive Na⁺ channels (Yamane et al., 2012; Yamashita et al., 2016a). However, the mechanism by which the Sema3A-induced local signaling event is propagated through the TTX-sensitive ion channels to the cell body is unknown. In this study, we identified Na_v1.7 as a voltage-gated Na⁺ channel that mediates the retrograde Sema3A signaling. CRMP1 expression lowered the threshold of Na_v1.7 in HEK293 cells, whereas the threshold value in crmp1^−/− neurons was higher than that in WT neurons. We here propose the functional coupling between Na_v1.7 and CRMP1 as the prerequisite process for retrograde Sema3A signaling, which plays an important role in axon guidance and neuronal network formation.

The initial signaling event of Sema3A involves endocytosis of Sema3A and the formation of the GSK3β–Axin-1–β-catenin complex (Hida et al., 2012). Furthermore, we have recently provided evidence that an interaction between PlexA4 and TrkA occurs in response to Sema3A, which is essential for retrograde Sema3A signaling (Yamashita et al., 2016b). To further characterize the endocytotic pathway for Sema3A signaling, we tested the effect of nimodipine, a voltage-dependent L-type Ca²⁺ channel blocker, and of monodancylcadaverine (MDC), an inhibitor of clathrin-dependent endocytosis. As reported previously (Yamashita et al., 2016b), in the growth cones, a transient increase in the colocalization of PlexA4- and TrkA-positive clusters was observed 3 min after Sema3A stimulation but returned to the basal levels 5 min after the stimulation (Fig. 1B). In the axons, an increase in the colocalized clusters of PlexA4 and TrkA was observed 5 min after the stimulation (Fig. 1B). This time-dependent imaging of double-positive signals suggests that the PlexA4–TrkA complex may form first in the growth cones and then be transported to the axons after Sema3A stimulation. Indeed, by using the fluorescence recovery after photobleaching (FRAP) method, we have previously demonstrated that axonal transport of PlexA4 tagged with enhanced green fluorescence protein (EGFP) is accelerated in response to Sema3A (Yamashita et al., 2016a). We found that nimodipine suppressed Sema3A-induced colocalization of PlexA4 and TrkA in both the growth cones and distal axons (Fig. 1). This result indicates that nimodipine-sensitive Ca²⁺ channels, including R-type Ca²⁺ channel, mediate Sema3A-induced colocalization of PlexA4 and TrkA in both growth cones and axons (Nishiyama et al., 2011; Treinys et al., 2014). MDC also suppressed Sema3A-induced colocalization of PlexA4 and TrkA in the growth cones and axons (Fig. 2). These findings suggest that Ca²⁺ channels- and clathrin-dependent endocytosis form the upstream signaling pathway for Sema3A-induced colocalization of PlexA4 and TrkA in growth cones and axons.

The Sema3A-induced local Ca²⁺ signaling in growth cones is propagated to the cell body in a TTX-sensitive manner (Yamane et al., 2012; Yamashita et al., 2016a). To investigate a possible involvement of the TTX-sensitive process, we examined the localization of PlexA4 and TrkA in growth cones and axons after Sema3A stimulation with or without TTX (Fig. 3). Colocalization was analyzed as the percentage of PlexA4 and TrkA staining that overlapped (see Materials and Methods). As described in Fig. 1B, the peak effect of Sema3A on the percentage of PlexA4 and TrkA double-positive signals was seen 3 min after Sema3A application in the growth cones and after 5 min in the axons, respectively. Treatment with TTX did not affect the Sema3A-induced colocalization of PlexA4 and TrkA in growth cones (Fig. 3B). This growth cone localization of the double-positive signals was sustained upto 5 min after the stimulation in the presence of TTX, whereas colocalization returned back to basal levels in the absence of TTX (Fig. 3B). By contrast, TTX suppressed the increase in the colocalized clusters of PlexA4 and TrkA in the distal axons observed 5 min after Sema3A stimulation (Fig. 3B). Thus, this result suggests that TTX, unlike nimodipine and MDC, does not prevent PlexA4 and TrkA from being endocytosed into the growth cones but suppresses transport of these molecules from growth cones to axons and their colocalization in axons after Sema3A stimulation.

To specify the TTX-sensitive voltage-dependent Na⁺ channels involved, we examined the effects of siRNA knockdown of Na_v1.6 and Na_v1.7 (encoded by SCN8A and SCN9A, respectively), the two major Na⁺ channels expressed in the DRG neurons (Dib-Hajj et al., 2013). We confirmed that transfection of siRNAs against Na_v1.6 and Na_v1.7 into dissociated DRG neurons effectively reduced Na_v1.6 and Na_v1.7 expression levels, respectively (Fig. 4A). The siRNA-mediated Na_v1.6 knockdown did not affect the Sema3A-induced colocalization of PlexA4 and TrkA in the growth cones and axons of mouse DRG neurons (Fig. 4B,C). By contrast, the siRNA knockdown of Na_v1.7 induced accumulation of the colocalized signals of PlexA4 and TrkA in the growth cones, and suppressed the colocalization in the distal axons, a remarkably similar phenotype to that observed with TTX treatment. In order to rule out any off-target effects, we performed siRNA knockdown of Na_v1.7 using the two different target sequences individually. Again, treatment with the two different kinds of Na_v1.7 siRNA resulted in prolonged localization of the colocalized signals of PlexA4 and TrkA in the growth cone, thereby showing the specificity of the siRNA (Fig. S1). This result demonstrates that Na_v1.7 is the voltage-dependent Na⁺ channel that is responsible for the Sema3A-induced localization of PlexA4 and TrkA clusters in distal axons.

We have previously shown that knockdown of either CRMP1 or CRMP2 suppresses Sema3A-induced growth cone collapse (Uchida et al., 2005). To elucidate the possible interaction between Na_v1.7 and CRMP1 or CRMP2, we first performed immunocytochemistry to examine whether or not Sema3A promoted colocalization between Na_v1.7 and CRMP1 and/or CRMP2 in DRG growth cones. We found that Sema3A transiently increased the levels of colocalization of CRMP1 and Na_v1.7 in the growth cones, which returned to the basal levels 5 min after stimulation. No change was observed in the colocalization levels of CRMP2 and Na_v1.7 (Fig. 5A,B), suggesting that CRMP1 but not CRMP2 colocalizes with Na_v1.7 in response to Sema3A and that CRMP1 plays a role in mediating Sema3A-induced axonal transport. This finding suggests that CRMP1 selectively colocalizes with Na_v1.7 in the growth cones to alter the ion-channel properties of Na_v1.7 in response to Sema3A.

To further confirm functional coupling between Na_v1.7 and CRMP1, we examined the localization of PlexA4 and TrkA in crmp1^−/− or crmp2^−/− neurons. In crmp1^−/− DRG neurons, the colocalization of PlexA4 and TrkA in the growth cones was sustained, while that in the axons was suppressed after the Sema3A stimulation, compared to WT neurons. No such phenotype was observed in crmp2^−/− neurons (Fig. 5C,D). This phenotype of crmp1^−/− neurons was similar to that observed in TTX-treated or Na_v1.7-knockdown DRG neurons (Figs 3 and 4). These findings suggest a functional coupling between Na_v1.7 and CRMP1, but not CRMP2, in inducing retrograde transport of PlexA4 and TrkA from growth cones to axons.

To investigate whether CRMP1 was functionally coupled with Na_v1.7, we characterized the biophysical properties of Na⁺ channels in WT and crmp1^−/− DRG neurons or HEK293 cells (Figs 6 and 7; Table S1). We performed current–voltage (I-V) curve analysis of cultured DRG neurons from WT, crmp1^+/− and crmp1^−/− mice. The value of the holding potential causing half-maximal current (V_1/2) in crmp1^−/− was significantly higher than that in WT DRG neurons (Fig. 6; Table S1).

CRMP1 is phosphorylated at residues Thr509 and Ser522 by Cdk5, and this phosphorylation process is involved in Sema3A signaling (Cole et al., 2006; Uchida et al., 2005; Yamashita et al., 2007). We next examined the V_1/2 in HEK293 cells expressing Na_v1.7 and nonphosphorylatable or phosphomimetic mutant CRMP1. Introduction of CRMP1 WT lowered the V_1/2 in HEK293 cells expressing Na_v1.7 (Fig. 7B; Table S1). The above findings suggest that CRMP1 lowers the threshold of Na_v1.7 and regulates the function of the voltage-dependent Na⁺ channels. To determine if the phosphorylation of CRMP1 is involved in the regulation of the Na⁺ channels, we next investigated the effect of the introduction of the nonphosphorylatable or phosphomimetic mutants of the CRMP1 gene in HEK293 cells. There was no enhancement of the value of V_1/2 in HEK293 cells that co-expressed Na_v1.7 and the nonphosphorylatable CRMP1 mutant Thr509Ala,Ser522Ala (T509A/S522A) when compared to HEK293 cells expressing Na_v1.7 alone (Fig. 7C; Table S1). Introduction of WT and phosphomimetic CRMP1 mutant Thr509Asp,Ser522Asp (T509D/S522D) into Na_v1.7-expressing HEK293 cells lowered the V_1/2 value (Fig. 7C; Table S1). The effect of the phosphomimetic form of CRMP1(T509D/S522D) in enhancing the V_1/2 value was superior to that of WT CRMP1 (Fig. 7C; Table S1).

There are two possibilities for the functional coupling between CRMP1 and Na_v1.7. One is that the colocalization between CRMP1 and Na_v1.7 per se could trigger Sema3A-induced localization of PlexA4 and TrkA in the growth cones and axons. The other is that CRMP1 alters the channel properties, thereby augmenting Na⁺ influx to propagate local Sema3A signaling. To test these possibilities, we examined the holding potential causing half-maximal current in crmp1^−/− mouse DRGs in the presence of a high and low concentration of extracellular Na⁺ ([Na⁺]_o) (Table S1). The holding potential causing half-maximal current (V_1/2) in crmp1^−/− DRG neurons was again higher than that in crmp1^+/− neurons, confirming that CRMP1 lowered the V_1/2 of the Na⁺ channels. This ion channel phenotype of the crmp1^−/− neuron was rescued by increasing [Na⁺]_o from 140 to 160 mM (Fig. 8A; Table S1). This finding suggests that CRMP1 modulates the channel properties, thereby accelerating Na⁺ influx to induce the retrograde Sema3A signaling. Consistently, accumulation of colocalized PlexA4 and TrkA in the growth cones in crmp1^−/− DRG neurons 3 min after Sema3A stimulation was rescued by increasing the concentration of [Na⁺]_o. Again, lack of localization of PlexA4 and TrkA in the axons 5 min after Sema3A in crmp1^−/− neurons was rescued by increasing [Na⁺]_o (Fig. 8B,C).

We have previously demonstrated that dynein-dependent and TTX-sensitive retrograde signaling of Sema3A plays a crucial role in the regulation of dendritic GluA2 localization and branching (Morita et al., 2006; Yamashita et al., 2016a,b). Our present study provides the first evidence for the molecular mechanism of TTX-sensitive propagation of the retrograde Sema3A signaling (Yamane et al., 2012; Yamashita et al., 2015). Furthermore, CRMP1, a mediator of Sema3A signaling (Nakamura et al., 2014; Uchida et al., 2005; Yamashita et al., 2007), lowered the threshold of voltage-dependent Na_v1.7 channel. The I-V curve analysis showed that the value of the holding potential causing half-maximal current in crmp1^−/− mouse DRG neurons was higher than that in WT or crmp1^+/− neurons. Electrophysiological examination in the HEK293 system further supports the functional coupling between Na_v1.7 and CRMP1. Moreover, TTX treatment and knockdown of Na_v1.7 both induced the same phenotypic defect as seen in crmp1^−/− neurons – i.e. the suppression of Sema3A-induced colocalization and dynein-dependent retrograde transport of PlexA4 and TrkA in growth cones and axons, which are essential for the retrograde Sema3A signaling (Yamashita et al., 2016b). Thus, we propose that voltage-dependent Na_v1.7 coupled with the dynein-dependent retrograde transport play a synergistic role in transmitting the retrograde Sema3A signal from the growth cone to the cell body (Fig. S2).

The differential effects of nimodipine, MDC and TTX suggest that colocalization of PlexA4 and TrkA in the growth cone and axon occurs through related but distinct mechanisms. After Sema3A treatment, the fluorescent clusters of PlexA4 and TrkA moved along the axons toward the cell bodies in a time-dependent manner – the colocalized signals of PlexA and TrkA emerged as punctate signals from the growth cones leading up to axons. The colocalization of PlexA4 and TrkA in both the growth cones and distal axons was suppressed by nimodipine or MDC (Figs 1 and 2). This finding suggests that the colocalization of these two molecules occurred downstream of Sema3A-induced Ca²⁺-dependent endocytosis (Castellani et al., 2004; Fournier et al., 2000; Hida et al., 2012; Tojima et al., 2010). Notably, at 5 min after stimulation, the increased colocalization in the growth cones returned to the basal level and was followed by an increase in colocalization in distal axons (Figs 1 and 2). TTX suppressed the return to the basal levels, thereby leading to sustained level of the colocalized signals in growth cones, and it also suppressed the colocalization in distal axons (Fig. 3), providing a contrast to the effect of nimodipine. These effects of TTX were mimicked upon RNAi knockdown of Na_v1.7 but not of Na_v.1.6 (Fig. 4). This finding indicates that Na_v1.7 is the target of TTX that could be responsible for the accumulation and suppression of colocalized signals of PlexA4 and TrkA. Consistently, Na_v1.7 is a major TTX-sensitive Na⁺ channel that is expressed in DRG neurons (Catterall et al., 2005; Wada, 2006).

Nimodipine and TTX suppress Sema3A-induced dendritic localization of GluA2 and dendritic branch formation in cultured hippocampal neurons (Yamashita et al., 2016a). Therefore, both endocytosis and axonal transport are essential cellular responses to Sema3A in order to execute its biological activity. We have also provided evidence that Sema3A-induced retrograde transport of endosomes containing TrkA–PlexA is required for its biological activity to induce dendritic branching (Yamashita et al., 2016c). This property of Sema3A signaling is similar to that of nerve growth factor (NGF)–TrkA signaling (Harrington et al., 2011). The maturation from endosomes to retrograde-transport-component TrkA signaling endosomes requires activation of Rac1–cofilin and actin depolymerization (Harrington et al., 2011). In contrast, the neurotrophin NT3 is incapable of eliciting retrograde survival signaling, presumably due to the absence of the actin-regulatory endosomal components and its inability to form retrogradely transported TrkA signaling endosomes. Therefore, Sema3A signaling may also involve Rac1–cofilin and an actin depolymerization process to induce retrograde axonal transport. Indeed, phosphorylation of cofilin by LIM kinases is a critical signaling event in Sema3A-induced growth cone collapse that is associated with depolymerization of actin filaments (Aizawa et al., 2001).

Sema3A induced colocalization of Na_v1.7 and CRMP1 in the growth cones, without changing the colocalization levels of Na_v1.7 and CRMP2 (Fig. 5A,B). This finding suggests that functional coupling between Na_v1.7 and CRMP1 plays a role in mediating Sema3A-induced axonal transport. It is unknown whether or not Sema3A induced colocalization of Na_v1.7 and CRMP1 at the plasma membrane or inside the growth cone. However, Na⁺ channels must be inserted into the plasma membrane to function in neurons, and there is a large pool of intracellular Na⁺ channels, thereby suggesting that alterations in the mode of Na⁺ channel trafficking could lead to quick changes in channel and neuron function (Bao, 2015). These findings support the idea that the colocalization of Na_v1.7 and CRMP1 may reflect or result from colocalization of Na_v1.7 and CRMP1 at the plasma membrane. As in TTX-treated neurons, the Sema3A-induced regulation of PlexA4 and TrkA localization in growth cones and distal axons was disrupted in crmp1^−/− DRG neurons (Fig. 5). In contrast, the Sema3A-induced localization of PlexA4 and TrkA was unaffected in crmp2^−/− neurons (Fig. 5C,D). This finding suggests a selective functional coupling between Na_v1.7 and CRMP1. The involvement of CRMP1 was also seen in Sema3A-induced colocalization of TrkA in the growth cones (Fig. 5C,D). These findings suggest that CRMP1 mediates Sema3A-induced retrograde axonal transport. It has been reported that CRMP2 is involved in kinesin-1-dependent transport of the Sra1–WAVE1 complex (Arimura et al., 2009) and in regulation of trafficking by linking endocytic regulatory proteins to dynein motors (Rahajeng et al., 2010). CRMP2 has been reported to be able to directly bind to voltage-dependent Na⁺ channels, modulating the channel's slow inactivation (Dustrude et al., 2013; Wang et al., 2010). CRMP2 also modulates Na_v1.7 trafficking via SUMOylation (Dustrude et al., 2013). Thus our finding that CRMP2 is not required for the TTX-sensitive axonal transport elicited by Sema3A, which is responsible for the localization of PlexA4 and TrkA in axons, is unexpected. Although the exact reasons are unknown, this is probably related to the different subcellular locations of CRMP1 and CRMP2 in neuronal cells (Higurashi et al., 2012; Makihara et al., 2016). We have previously conducted immunocytochemistry analyses by staining for the CRMPs with actin or tubulin (Higurashi et al., 2012). CRMP1 is uniformly localized throughout the growth cone with a subtle predominance in lamellipodia, and CRMP1 in the peripheral domain roughly colocalized with actin. CRMP2 is distributed in the central domain, showing a distribution pattern similar to that of tubulin. Furthermore, inactivation of CRMP1 or CRMP2 in the whole growth cone disturbs neurite outgrowth in different ways. To examine the local function of CRMP1 and CRMP2, we have performed previously microscale-chromophore-assisted light inactivation (micro-CALI), which enables investigation of localized molecular function with highly spatial and temporal resolutions (Higurashi et al., 2012). The lamellipodial retraction rate in micro-CALI analysis of CRMP1 is higher than that of CRMP2 (Higurashi et al., 2012). The predominant distribution and function of CRMP1 in the actin cytoskeleton may be related to the functional coupling between CRMP1 and Na_v1.7 in naive DRG neurons (Figs 5 and 6). Interestingly, interaction between CRMP1 and Filamin-A regulates actin reorganization in growth cones upon Sema3A stimulation (Nakamura et al., 2014), and filamin regulates Na⁺ channels in a Xenopus laevis renal tubular cell line (Cantiello et al., 1991). These findings suggest that, in comparison with CRMP2, the function of CRMP1 might be more closely involved in actin cytoskeleton dynamics in neurons.

Our electrophysiological analyses indicate that CRMP1 and its phosphorylation regulate the threshold of Na_v1.7. The I-V curve measurement in DRG neurons revealed that the V_1/2 value in crmp1^−/− neurons was lower than that in wild-type neurons. This finding is consistent with that obtained in HEK293 cells expressing Na_v1.7 and wild-type or phosphomimetic forms of CRMP1 (Table S1; Fig. 7). The further decrease in the V_1/2 value in the phosphomimicking CRMP1 mutant suggests that phosphorylation of CRMP1 at residues Thr509 and/or Ser522 regulates the functional interaction between Na_v1.7 and CRMP1. Indeed, Cdk5 can directly phosphorylate CRMP1 at Thr509 and Ser522 (Cole et al., 2006), and Sema3A stimulation induces activation of Cdk5 (Sasaki et al., 2002), thereby leading to the phosphorylation of CRMP1 in the growth cones (Uchida et al., 2005). This finding is consistent with the immunocytochemical observation that, upon Sema3A stimulation, CRMP1 became colocalized with Na_v1.7 at the growth cones, although it remains to be clarified whether or not this functional coupling between CRMP1 and Na_v1.7 is based on their direct interaction.

If CRMP1 mediates the retrograde Sema3A signaling by regulating Na⁺ influx by lowering the threshold of Na_v1.7, the electrophysiological properties of crmp1^−/− neurons may be modified by altering the concentration of [Na⁺]_o. Expectedly, raising the concentration of [Na⁺]_o rescued the phenotype of crmp1^−/− DRG neurons. When the concentration of Na⁺ was raised from a normal (140 mM) to a high (160 mM) level, the increased V_1/2 value in crmp1^−/− DRG neurons was decreased to the V_1/2 level in WT neurons (Fig. 8A; Table S1). Likewise, impaired localization of PlexA4 and TrkA in growth cones and distal axons were rescued by increasing [Na⁺]_o (Fig. 8).

In Xenopus commissural interneurons, it has been suggested that Sema3A increases the number of Ca_v2.3 channels in the growth cone plasma membrane rather than modulating their gating properties (Nishiyama et al., 2011). Further studies will be required to determine whether CRMP1 regulates membrane trafficking of Na_v1.7 and/or it modulates the channel properties of Na_v1.7.

CRMP1 mutants – T509A/S522A and T509D/S522D – were constructed by using site-directed mutagenesis. Transfection into HEK293 cells of plasmid DNAs encoding Na_v1.7 (Klugbauer et al., 1995), CRMP1 or its mutants was performed using Lipofectamine 2000 (Life Technologies).

DRGs were removed from mouse embryos at embryonic day (E)12.5 (male or female) and were treated with 0.25% trypsin (Life Technologies) in Ca²⁺- and Mg²⁺-free PBS at 37°C for 5 min, and the reaction was stopped by adding F-12 Ham's medium (Wako) containing 10% FBS and 10 ng/ml of the 2.5S subunit of NGF. The DRGs were washed with the same medium used to stop the reaction. The cells were cultured onto culture dishes that had been previously coated with 100 µg ml⁻¹ poly-L-lysine (Wako Pure Chemical Industries Ltd., Osaka, Japan) in F-12 Ham's medium containing 10% FBS and 10 ng/ml 2.5S NGF. To starve the cells of NGF, we replaced the culture medium after 1 day in vitro (DIV) with Leibovitz's L-15 Medium (Life Technologies) and cultured the cells for 4 h. Cultured mouse DRG neurons were stimulated with 0.1 or 0.5 nM Sema3A for 0, 3 and 5 min. In some experiments, 1 µM nimodipine, 100 µM MDC, 100 nM TTX or vehicle control dimethyl sulfoxide (volume-matched) was added to the explants 7 min before Sema3A stimulation.

Non-targeting scrambled siRNA (Stealth RNAi™ siRNA Negative Control) was used as a control siRNA (Invitrogen, Med GC, #12935300). The sense strands for the two 19-oligonucleotide Na_v1.6 siRNAs were synthesized with the following sequences: siRNAa: 5′-CGAAUGGCUUGUACUAUUA-3′; siRNAb: 5′-CUAGUAAUGUUCAGCGAUU-3′; siRNAc: 5′-CAACCCACGUAAACUGUAA-3′ (#SC149786, Santa Cruz, CA). The sense strands for the two 19-oligonucleotide Na_v1.7 siRNAs were synthesized with the following sequences: siRNAa: 5′-GUUCUUAUCUGCCUCAAUA-3′; siRNAb: 5′-CUCACUCGUGUUAAUAAGA-3′; siRNAc: 5′-CACCAGGUGUUUAGUAUGA-3′ (#SC149784A, Santa Cruz, CA). The siRNAs against Na_v1.6 and Na_v1.7, mixed to use the same amount of the three siRNAs for each of the corresponding target genes, were introduced into cultured dissociated mouse DRG neurons using DharmaFECT-3 (Thermo Scientific, Hudson, NH, USA) according to the manufacturer's protocol (Hida et al., 2012). Cultured dissociated mouse DRG neurons were transfected with each siRNA (final concentration, 50 nM) 10-12 h after the start of cultures. These cells were further incubated for 24 h before use in a series of experiments.

The dissociated DRG neurons were fixed with fresh 2% or 4% PFA including 4% sucrose at room temperature for 20 min, washed twice with PBS, permeabilized for 10 min using 0.1% Triton X-100 in PBS and then blocked for 1 h in blocking buffer (5% goat serum in PBS). The DRG neurons were incubated with anti-PlexA4 antibody (1:500-1:2500 dilution) (Suto et al., 2007); anti-TrkA rabbit polyclonal antibodies (1:1000 dilution; 06-574, Millipore); anti-CRMP1 (1:400) antibody (2E7G) and anti-CRMP2 antibody (9F) (1:400) (Wako Pure Chemical) (Makihara et al., 2016); or anti-Na_v1.7 antibody (1:1000 dilution; ASC-008, Alomone Labs) diluted in Can Get Signal (Toyobo, Osaka, Japan) at 4°C overnight, and then incubated at room temperature for 1 h with Cy3- (Jackson) or Alexa-Fluor-594 (Life Technologies)-conjugated secondary antibodies diluted with TBS containing 0.05% Tween 20. The fluorescence microscopy images were generated using a laser confocal microscope (LSM510) with a water-immersed objective set at 40× (C-Apochromat) equipped with an Axioplan 2 imaging microscope (Zeiss, Oberkochen, Germany).

HEK293 cells and DRG neurons from E11 to E14 C57BL6N mice of either sex were cultured on cover glasses. Those cells on the cover slip were perfused in the extracellular solution containing (in mM): 140 NaCl, 3 KCl, 1 MgCl₂, 1 CaCl₂, 10 D-glucose, 20 HEPES, pH 7.3. EGFP-positive HEK293 cells, or cultured DRG neurons for which the diameter of the somata was less than 30 µm were voltage clamped at −80 mV using Axopatch 200B patch clamp amplifier (Molecular Devices, Sunnyvale, CA) at room temperature. A borosilicate glass pipette (GD-1.5, Narishige, Tokyo, Japan) had a resistance of 3-5 MΩ when filled with an intracellular solution that contained (in mM): 100 CsCl, 40 CsF, 10 NaCl, 1 EDTA, 300 sucrose, 10 HEPES, pH 7.3. After whole-cell configuration, cells were allowed to settle for 5 min to stabilize. Thereafter, membrane currents were given a depolarizing test pulse of 0.2 s at 10-mV steps from −60 mV to 30 mV and were recorded using a standard P/4 leak subtraction protocol (Bendahhou et al., 1999). Responses were filtered at 1 kHz with an eight-pole Bessel filter, digitized at 3 kHz. Data acquisition and analysis were performed using pCLAMP 10.2 software (Molecular Devices). Membrane potential was corrected for the liquid junction potential.

To create an I-V curve, the ratios to the peak current of each depolarization test pulse were plotted. Using this current-voltage relationship, conductance [G_Na=I_Na/(V−V_rev)] was calculated and was approximated using the Boltzmann distribution {G_Na/G_MAX=(1+exp [ze (V−V_1/2)/kT])⁻¹} to calculate the half-activation potential V_1/2. In these formulae, I_Na represents each current value at depolarizing test pulse (V), V_rev is the equilibrium potential calculated using the Nernst equation. G_MAX is the peak conductance, k is the Boltzmann constant, z is the gate charge. Under the conditions of 22°C, kT/e=25 mV.

To quantify the colocalization of immunoreactive signals of PlexA4, TrkA, Na_v1.7, CRMP1 or CRMP2 in growth cones (within 10 μm of the axon tip) or axons (within 25-35 μm of the axon tip) of cultured DRG neurons, we calculated Pearson's correlation coefficients (colocalization coefficients) using LSM 5 Image software (Zeiss). The result is +1 for perfect correlation, 0 for no correlation and −1 for perfect anti-correlation. The colocalization signal of PlexA4 and TrkA, and of Na_v1.7 and CRMP1 or CRMP2 was analyzed using colocalization highlighter (ImageJ) and visualized as white in figures.

Dissociated mouse DRG neurons transfected with each siRNA were washed twice with ice-cold PBS, and total RNAs were obtained from these cultured cells. The RNA was extracted using the RNeasy kit (Qiagen) according to the standard protocol. To obtain first-strand cDNA, 500 ng of the obtained total RNA was incubated at 65°C for 5 min with 1 µl 50 µM oligo(dt)₂₀, 1 µl 10 mM dNTPmix and water that had been treated with diethylpyrocarbonate, which was quickly cooled on ice. It was treated with 4 µl 5× first-strand buffer, 1 µl 0.1 M dithiothreitol, 0.5 µl 200 U µl⁻¹ SuperScript™ III reverse transcriptase (Life Technologies, Co.) and incubated at 50°C for 50 min and at 70°C for 15 min. In semi-quantitative real-time (RT)-PCR analysis, Go Taq^® Green Mastermix (Promega) was used. RT-PCR was conducted at 95°C for 60 s, followed by 30 cycles of 95°C for 30 s and 58°C for 30 s and 72°C for 90 s (Na_v1.6), or 30 cycles of 95°C for 30 s and 60°C for 30 s and 72°C for 90 s (Na_v1.7), or 30 cycles of 95°C for 30 s and 56°C for 30 s and 72°C for 90 s (β-actin). PCR samples were loaded onto 1% agarose gel for electrophoresis. Sequences of primers were 5′-CTTTCACCCCCGAGTCGCTGGCAAA-3′ and 5′-GCTAAATCTGAAGAGAGTTTTCCCT-3′ for Na_v1.6; 5′-CCCTTGGATCAGAATCCGCAGGTGCACTCA-3′ and 5′-CTCCGTAGATGAAGGGTAGCTGTTTACCTG-3′ for Na_v1.7; and 5′-GTGGACATCCGCAAAGACCTGTA-3′ and 5′-CGCCGATCCACACGGAGTACT-3′ for β-actin.

Data are shown as mean±s.e.m. The statistical significance of the results was analyzed using one-way ANOVA or Student's t-test.

We thank Drs Takuya Takahashi, Kiwamu Takemoto, Kohtaro Takei, Yukio Sasaki, Akihiko Wada and Shigeo Ohno for useful discussions and constructive inputs. We thank Dr Stephen G. Waxman (Yale School of Medicine, New Haven) for his kind supply of the plasmid containing Na_v1.7 cDNA. We also thank Dr Fumikazu Suto (National Center of Neurology and Psychiatry, Tokyo) for kindly supplying the anti-PlexA4 antibody.