Full-length Stx1A and Stx1A-GFP were created as described previously (Cotrufo et al., 2011). UNC5A-myc, UNC5B-myc and UNC5C-myc were a gift from Prof. Lindsay Hinck (University of California, Santa Cruz, Santa Cruz, CA, USA), pEGFP-C1 (Clontech), pcDNA3.1 (Invitrogen, Thermo Fisher Scientific), GFP-EpsΔ95/295 vector was a gift from Dr. Francesc Tebar (University of Barcelona, Spain). For shRNA experiments, the pLVTHM plasmid was kindly provided by Prof. Didier Trono (École polytechnique fédérale de Lausanne, Lausanne), including specific oligonucleotides for Stx1A and Stx1B sequence: gatcccc CCAGAGGCAGCTGGAGATCACttcaagagaGGTCTCCGTCGAC CTCTAGTGttttt (Forward), and agctaaaaaCCAGAGGCA GCTGGAGATCACtctcttgaaGGTCTCCGTCGACCTCTAGTGggg (Reverse).

HEK-293T cells were maintained in DMEM (11995065; GIBCO-Thermo Fisher Scientific) medium supplemented with 10% fetal bovine serum (FBS; 26140079; GIBCO-Thermo Fisher Scientific), 1% GlutaMAX (35050061; GIBCO-Thermo Fisher Scientific) and 1% penicillin/streptomycin (15140122; GIBCO-Thermo Fisher Scientific). PC12 cells were maintained in DMEM containing 1% GlutaMAX, 5% FBS, 5% horse serum (HS; 26050-088; GIBCO-Thermo Fisher Scientific), and 1% penicillin/streptomycin.

Primary cultures and explants of cerebellar EGL neurons were prepared from P3 to P5 postnatal CD1 strain mice (Charles River). Animals were sacrificed by decapitation in accordance with institutional and governmental ethical guidelines and regulations. All the experiments using animals were performed in accordance with the European Community Council directive and the National Institutes of Health guidelines for the care and use of laboratory animals. Experiments were also approved by the ethical committee from the Generalitat of Catalonia. Postnatal cerebellums were isolated, mechanically disaggregated and trypsinized as previously described (Rosello-Busquets et al., 2019; Hernaiz-Llorens et al., 2020). Briefly, after centrifugation, neurons were resuspended in 2 mL of DMEM medium and EGL neurons were isolated by centrifugation (3000 rpm, 10 min at 4°C) in a discontinuous percoll gradient (35 and 60% of percoll). After washing with PBS, EGL neurons were plated on poly-D-lysine pre-coated dishes in DMEM, 1% penicillin/streptomycin, 1% glutamine, 4.5% D-(+)-glucose (G-8769; Sigma), 5% HS and 10% FBS. This medium was maintained for 24 h and then changed to a medium where HS and FBS were replaced by 2% B27 (17504001; GIBCO-Thermo Fisher Scientific) and 1% N2 (11520536; GIBCO-Thermo Fisher Scientific). Neurons were cultured for 72 h (3DIV).

To isolate and culture EGL explants, cerebellums from P3 to P5 mice were isolated and chopped in to 300 μm slices. Selected slices were further dissected using fine tungsten needles to extract small tissue pieces from the EGL. Explants were carefully placed inside a 3D collagen matrix on poly-D-lysine pre-coated dishes, prepared as previously described (Gil and Del Rio, 2019). Explants were cultured for 48 h (2DIV) in DMEM, 1% penicillin/streptomycin, 1% glutamine, 4.5% D-(+)-glucose, 2% B27.

To downregulate the expression of both Stx1A and Stx1B, a shRNA directed both Stx1A and Stx1B was used to downregulate the expression levels of the two paralogs, and a scramble shRNA was used as negative control [21]. The coding mRNA sequence for both Stx1A and Stx1B in mouse (mus musculus) and rat (rattus norvegicus) share 97 and 96% identity, respectively, and are identical in the shRNA-recognized region. One day before transfection, cells were counted and plated in 100 mm dishes, to have a 70% of confluence on the day of transfection. HEK-293T or PC12 cells were transfected with Lipofectamine 2000 (Thermo Fisher Scientific) following the manufacturer’s instructions.

After 2DIV, neuronal cultures were transfected using Lipofectamine 2000. Before transfection, 3/5 of the medium was removed and kept aside to be returned later. For each 35 mm plate, 4 μg of DNA and 8 μL of Lipofectamine 2000 were used. The final mixture of DNA-Lipofectamine 2000 was carefully added to the cultures and further incubated at 37°C in 5% CO₂ for 60 min. The transfection medium containing DNA and Lipofectamine 2000 was finally replaced by the medium removed at the beginning. An equal amount of freshly prepared medium was added. Cultures were incubated until the following day.

Explants were electroporated using the Invitrogen Neon system. Briefly, dissected explants were washed three times in PBS and resuspended in buffer R containing 2 to 4 μg of the DNA to electroporate. The conditions for the electroporation were voltage: 500 V; width: 50 ms; 5 pulses. Explants were immediately washed once in Neurobasal medium and then mounted in a 3D collagen matrix.

PC12 cells and primary neurons were fixed with a solution of 4% paraformaldehyde (PFA) in PBS for 10 min at room temperature. Neuronal explants were fixed by incubation with the same solution for 30 min at room temperature. They were rinsed with PBS, and permeabilized with a solution of 0.1% Triton-X-100 in PBS for 10, or 30 min for the explants. Cells were then washed with PBS and incubated in blocking solution (10% HS in PBS) for 1 h at room temperature, or 3 h for the explants. After blocking, the cells were incubated with the respective primary antibodies diluted in blocking solution for 2 h at room temperature or overnight at 4°C for the explants. Cells and explants were washed with PBS and incubated in the secondary antibody solution (PBS 1x with 10% HS) for 1 h at room temperature, or 3 h for the explants. Finally, cells and explants were washed and mounted in Mowiol (Sigma-Aldrich) for imaging.

Cholera toxin B-subunit (CTxB) Alexa Fluor 488-conjugated (C22841; Invitrogen-Thermo Fisher Scientific). Colocalization of UNC5C receptor with Stx1 and the lipid raft membrane marker CTxB was quantified in growth cones manually selected using Fiji wand tool, and performing thresholding of the corresponding channels (UNC5C, Stx1 or CTxB) to obtain binary masks that were overlapped. Colocalization was considered as the fraction of UNC5C area that overlaps with Stx1 or the raft marker.

Cells were lysed in lysis buffer (50 mM Tris pH 7.2, 150 mM NaCl, 5 mM EDTA, 1% Triton, 10% glycerol, 10 μg/ml Leupeptin, 10 μg/ml Aprotinin, 1 mM PMSF). For the immunoprecipitation assays, 50–500 μg of total protein per brain tissue sample or transfected HEK-293T cell homogenates were used. Briefly, samples were first pre-cleared with 30 μL of protein G-sepharose beads (P3391, Sigma). Additional protein G-sepharose beads was blocked in 5% bovine serum albumin (BSA) and stored at 4°C. After 1 h of pre-clearing and blocking, samples were incubated overnight at 4°C with mouse anti-myc (Sigma-Aldrich; 1/5000), mouse anti-GFP (11814460001, Sigma-Aldrich; 1/500) or mouse anti-Sytx1 (HPC-1 clone; Sigma-Aldrich; 1:500) antibodies. Negative controls were performed by immunoprecipitation with Sepharose beads alone or with the non-interacting antibody rabbit anti-Egr1 (Invitrogen Thermo Fisher-Scientific; 1:50) (Cotrufo et al., 2011). Blocked protein G-sepharose beads was added and incubated for 2 h at 4°C. After five washes with wash buffer (10 mM Tris pH 7, 500 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 0.5% NP-40), SDS-sample buffer was added to the beads and the proteins were analyzed by SDS-PAGE and Western Blot. Proteins were transferred onto nitrocellulose membranes, which were blocked with 5% non-fat dry milk in Tris-HCl buffered saline (TBS) containing 0.1% Tween 20, and incubated overnight at 4°C with goat anti-UNC5C (ab106949; Abcam; 1/500), mouse anti-Sytx1 (1:3000), rabbit anti-myc (1:1000), or rabbit anti-GFP (A11122; Invitrogen-Thermo Fisher Scientific; 1:1000) antibodies. After incubation with the appropriate HRP-conjugated secondary antibodies, blots were developed following the ECL method (Amersham Pharmacia Biotech). To separate the Sytx1A and Stx1B isoforms, brain protein samples were analyzed by Tris-urea/SDS-PAGE (18% acrylamide, 6 M urea, 750 mM Tris-HC1, pH 8.85, 50 mM NaCI, 0.1% SDS). The resulting bands were manually selected and quantified using ImageJ.

External granule layer primary neurones were cultured for 3DIV and then treated with specific reagents or toxins, or the respective controls (DMSO or PBS), which were included 10 to 30 min before the addition of Netrin-1, and maintained during the whole incubation. 5-(N-ethyl-N-isopropyl) amiloride (EIPA) was purchased from SIGMA-MERCK, and botulinum neurotoxin type A (BoNT/A) and type C1 (BoNT/C1) were purchased from Metabiologics, INC. In growth cone collapse experiments, neurons were then incubated with Netrin-1 (300 ng/mL) or control (BSA 0.1%) for 45 min at 37°C in DMEM containing 1% glutamine and 4.5% D-(+)-glucose. After this incubation, neurons were fixed with 4% PFA and permeabilized with PBS-Triton-X-100 (0.1%). Finally, actin filaments were stained by incubation with phalloidin-TRITC (1 μM) for 30 min. Cells were mounted on Mowiol and used for imaging. Actin staining was used to identify growth cones, which were outlined based on differential staining for actin in this compartment with respect to the adjacent axon. When actin was not stained, growth cones were outlined based on their WGA-TRITC staining that highlighted their surface. An intensity threshold mask was created using Fiji (Schneider et al., 2012) and the growth cone perimeter was selected using the wand tool. Collapsing growth cones were manually identified based on their morphology. In contrast to normal extended growth cones with evident filopodia and lamellipodia, collapsing growth cones lose their spread morphology, rapidly (within minutes) acquiring a shriveled, round-tipped pencil-like shape largely devoid of lamellae or filopodia (Luo et al., 1993; Goshima et al., 1995). The percentage of growth cones containing macropinocytic (dextran-positive) endosomes, and the percentage of collapsing growth cones was calculated and plotted to evaluate the degree of growth cone collapse and macropinocytosis.

During the plating procedure, explants embedded into the 3D collagen matrix were confronted at a distance of 150–300 μm with aggregates of HEK-293T cells stably expressing Netrin-1 (Cotrufo et al., 2011), or transiently transfected with semaphorin 3A or semaphorin 3F. Netrin-1 expression in stable cells was regularly checked by western blot (data not shown). If required, explants were treated with EIPA, DMSO (control), or with specific BoNTs (25 nM BoNT/A or 15 nM BoNT/C1) by adding them to the medium 3 to 4 h after being cultured. After 48 h (2DIV) of culture, explants were fixed, and immunocytochemistry against βIII-tubulin and GFP was performed. Repulsion was analyzed by measuring the proximal/distal (P/D) ratio as previously described (Gil and Del Rio, 2019; Hernaiz-Llorens et al., 2020). Briefly, the P/D ratio is the fraction of axons growing in the proximal quadrant of the explant (closer to the aggregate of HEK-293T cells) vs. those growing in the distal quadrant. A ratio close to 1 indicates a radial growth pattern, below 1 indicates chemorepulsion and above 1 indicates chemoattraction. Only the GFP-positive axons were used for quantification in experiments with explants containing neurons expressing GFP.

Results were analyzed statistically using GraphPad Prism software (GraphPad Software, Inc). The D’Agostino and Pearson test was used to test for normality. The unpaired two-tailed Student’s t-test was used to compare two groups. For datasets comparing more than two groups, ANOVA followed by the Dunn test, Sidak’s or the Dunnet test corrections for multiple comparisons was used. Statistical comparisons were performed on a per-explant or per-experiment basis. If the number of elements analyzed for each condition was above 15, it is indicated within each bar. If the number was below 15, each value is plotted. When the comparison is performed on a per-experiment basis, individual data points are shown as the number is <15. Each of the experiments analyzed a large number of individual cones/axons, as stated in the corresponding figure legends. The neurons analyzed were randomly selected within the dishes and blindly analyzed, and were collected from at least three independent experiments. Values are represented as the mean ± SEM. The tests used are indicated in the respective figure legends. A p-value below 0.05 was accepted as significant. “ns” means non-significant.

We have previously demonstrated that the t-SNARE Stx1 interacts with the guidance receptor DCC (Cotrufo et al., 2011, 2012) and the tropomyosin-related kinase (TrkB) receptor (Fuschini et al., 2018), with these associations being necessary for the Netrin-1-dependent chemoattraction and neurotrophin-dependent outgrowth of axons, respectively. These results indicate that Stx1 regulates exocytosis and membrane retrieval in growth cones and is required for Netrin-1-induced chemotropic guidance in vitro and in vivo (Ros et al., 2015). On the other hand, it has also been suggested that SNARE proteins can be involved in certain types of endocytosis within synapses (Xu et al., 2013; Zhang et al., 2013), and that some SNAREs interact with proteins involved in endocytosis (Galas et al., 2000; Koo et al., 2011; Miller et al., 2011). In addition, we have found that Stx1 is required for the repulsion of commissural axons (Ros et al., 2018). Here, we first investigated whether Stx1 interacts with the chemorepulsive Netrin-1 receptors UNC5A, UNC5B, and UNC5C, which are essential in controlling the repulsive response of axonal tracts during cerebellar development (Alcantara et al., 2000; Kim and Ackerman, 2011). First, we tested this interaction using an in vitro system in which HEK-293T cells were co-transfected with Stx1A or Stx1A-EGFP, together with UNC5B-myc, and tested for co-immunoprecipitation (Figure 1). Our results revealed that immunoprecipitation with anti-myc antibodies led to Stx1 association (Figure 1A). The reverse immunoprecipitation with anti-Stx1 antibodies also yielded myc-tagged UNC5B (Figure 1A). This interaction was maintained using anti-GFP antibodies to pull down Stx1A-EGFP (Figure 1B). In similar co-immunoprecipitation experiments, co-transfecting UNC5A-myc or UNC5C-myc with Stx1A-EGFP revealed that Stx1 also co-associates with UNC5A and UNC5C receptors (Figure 1C). Furthermore, control immunoprecipitations with beads alone did not show co-association (Figures 1B, C). Lack of immunoprecipitation with irrelevant antibodies (anti-Egr1) further suggested that the observed interaction was specific (Supplementary Figure 1). Next, we asked whether Stx1 is associated with UNC5 receptors in vivo. Using cultured postnatal cerebellar EGL neurons, we found that both proteins partially colocalize in control (Netrin-1 untreated) growth cones (Figure 1D) to a similar extent to the colocalization with lipid rafts microdomains labeled with cholera toxin subunit B (CTxB) (Figure 1E, inset in Figure 1D) (Hernaiz-Llorens et al., 2020). Moreover, immunoprecipitation of postnatal day 4 (P4) cerebellar cortical tissue with Stx1 revealed interaction with UNC5C (Figure 1F). Finally, to investigate whether Stx1 interacts with other receptors that mediate chemorepulsion, we co-transfected HEK-293T cells with Stx1A-EGFP and Neuropilin-1-HA or Plexin-A1-VSV, the receptors for class III Semaphorins (Gil and Del Rio, 2019). Our results showed that Stx1 does not interact with these Semaphorin receptors (Supplementary Figure 2).

Netrin-1 exerts long-range chemorepulsion that is important for axonal guidance (Round and Stein, 2007), and plays a major role in the organization of EGL neurons during early postnatal development of the cerebellum, controlling the growth of parallel fibers (Przyborski et al., 1998; Alcantara et al., 2000; Peng et al., 2010; Consalez et al., 2020). EGL neurons have been shown to co-express DCC, neogenin, DSCAM, UNC5B, and UNC5C that sense Netrin-1 expressed in the EGL and interneurons of the molecular layer of the cerebellum (Alcantara et al., 2000; Kim and Ackerman, 2011; Hernaiz-Llorens et al., 2020). To investigate whether Stx1 is necessary for Netrin-1-dependent repulsion, we co-cultured postnatal EGL explants in type I collagen 3D hydrogels and exposed them to aggregates of either control HEK-293T cells or Netrin-1-expressing HEK-293T cells for 2 days (Alcantara et al., 2000; Hernaiz-Llorens et al., 2020). EGL explants cultured with control HEK-293T cells exhibited a radial pattern of axonal growth (Figure 2A). In contrast, those confronted with Netrin-1-expressing cells showed strong chemorepulsion, with most axons growing in the distal quadrant (Figure 2A). Botulinum neurotoxins (BoNTs) are metalloproteases that reduce synaptic vesicle exocytosis and neurotransmitter release by cleaving specific SNARE proteins (Schiavo et al., 2000). Botulinum neurotoxin A (BoNT/A) exclusively cleaves SNAP25, whereas botulinum neurotoxin C1 (BoNT/C1) cleaves both Stx1 and SNAP25 (Blasi et al., 1993; Schiavo et al., 2000). Cleavage of Stx1 by BoNT/C1 in EGL cultures was confirmed by western blot (Supplementary Figure 3A). It has been shown that BoNTs cleave SNARE proteins in neuronal explants without affecting the secretion of Netrin-1 from stably transfected HEK-293T cells (Cotrufo et al., 2011). EGL-derived explants co-cultured with Netrin-1 in the presence of BoNT/C1 or BoNT/A displayed different phenotypes. While Netrin-1 elicited chemorepulsion in EGL explants incubated with BoNT/A, treatment with BoNT/C1 resulted in radial axonal sprouting (Figures 2A–C). Quantification of the length of axons sprouting from the explants confirmed a specific decrease in the length of explants treated with BoNT/A or BoNT/C1, but showed that only the use of BoNT/C1 toxin affected the chemorepulsive response toward Netrin-1 (Figure 2C).

As Stx1 interacts with UNC5 receptors but not with Neuropilin or Plexin receptors, we next investigated whether BoNTs altered Semaphorin 3A/3F-induced chemorepulsion (Supplementary Figure 3B). Hippocampal explants exhibited strong repulsion when confronted with Semaphorin 3A- or Semaphorin 3F-expressing cells. This repulsion was not altered when the explants were cultured in the presence of BoNT/A or BoNT/C1 (Supplementary Figure 3B). Taken together, these data indicate that the cleavage of Stx1 specifically abolishes Netrin-1-induced repulsion, but not class III Semaphorin-induced chemorepulsion.

Stx1 consists of two similar paralogs in mammals, Stx1A and Stx1B (Gerber et al., 2008; Ros et al., 2018). We investigated the expression of Stx1A and Stx1B in P3-5 postnatal cerebellum by western blot. We found that, although both paralogs are expressed, the expression pattern of the paralogs differs in the postnatal cerebellum and the forebrain (Figures 3A, B). It has previously been shown that mice deficient for Stx1A do not display anatomical abnormalities and only exhibit minor physiological neurotransmitter release deficits, most likely due to functional redundancy with Stx1B (Fujiwara et al., 2006). To evaluate the importance of Stx1B in Netrin-1-dependent chemorepulsion, we used neurons isolated from Stx1B knock-out mice (Ros et al., 2018). The level of Stx1B paralog was depleted in the knock-out mice (−/−), compared with the wild-type (+/+) or the heterozygous (±) animals (Figure 3C). Similar to what has already been described in other studies using independently generated Stx1B knock-out mice, newborn homozygous Stx1B (−/−) mice were slightly smaller than their control wild-type littermates (Supplementary Figure 4A; Kofuji et al., 2014; Wu et al., 2015). Whereas homozygous Stx1B knock-out mice usually survived until P7–P15, heterozygous Stx1B (±) targeted mice were viable until adulthood. We next performed Netrin-1 repulsion assays in EGL explants derived from P5 mice. EGL explants from control Stx1B (+/+) littermates displayed strong axonal chemorepulsion when confronted with Netrin-1-expressing cells. In contrast, chemorepulsion to Netrin-1 was significantly reduced in explants from homozygous Stx1B-deficient mice (Figures 3D, E) but was not completely abolished. To ascertain whether a functional redundancy with Stx1A may compensate for the lack of Stx1B, we generated double Stx1B/Stx1A-deficient mice. However, as double Stx1B/Stx1A knock-out animals die at birth (Ros et al., 2018), we used an alternative approach to target both Stx1 paralogs. We created a Stx1-shRNA sequence complementary to a conserved region in both Stx1A and 1B, enabling us to knock down both Stx1 paralogs. The efficiency of Stx1 downregulation was confirmed in PC12 cells, in which expression of the shRNA Stx1A-1B construct decreased Stx1 protein levels detected by western blot and made Stx1 undetectable by immunocytochemistry (Supplementary Figures 4B, C). Similar results were found after expressing the shRNA Stx1A-1B construct in EGL neurons and observing the decrease in Stx1 level (Supplementary Figure 4D). After electroporating EGL explants with the shRNA Stx1A-1B, the downregulation of both Stx1 paralogs resulted in a dramatic decrease in chemorepulsion, when compared to control cultures (Figures 3F, G). Together with the above results, these observations suggest that both Stx1 paralogs participate in Netrin-1-mediated chemorepulsion of EGL axons.

To better understand the mechanisms involved in Netrin-1 chemorepulsion, we performed collapse assays in postnatal cerebellar EGL neurons. Primary neuronal cultures were exposed to Netrin-1 for 15–45 min and stained with phalloidin to label F-actin and analyze growth cone morphologies. Control cultures exhibited typical growth cones with triangular shapes and numerous lamellipodia and filopodia (Figure 4A). Incubation with Netrin-1 resulted in the initiation of the collapse of up to 75% of growth cones (Figures 4B, C), which exhibited the typical shriveled, round-type, pencil-like shape, largely devoid of filopodia or lamellipodia, with a redistribution of actin filaments toward the central domain of growth cones (Figure 4A).

Previous studies have shown that Semaphorin 3A (Fournier et al., 2000; Tojima et al., 2010) and Slit2 (Piper et al., 2006) induce endocytic events in growth cones. We therefore investigated whether Netrin-1 leads to endocytosis in EGL growth cones, using two typical markers of endocytosis, the styryl dye FM1-43 (Gaffield and Betz, 2007) and a fluorescently labeled low-molecular-weight (LMW, <10 kDa) dextran (Alexa Fluor 488-LMW-dextran) (Fournier et al., 2000). After incubation with Netrin-1 and the above dyes, EGL cultures were incubated with wheat germ agglutinin conjugated with TRITC (WGA-TRITC) to label the cell surface of growth cones (Figures 4D, H). After 15 min of Netrin-1 treatment, the growth cone areas were slightly reduced, ranging from 7 to 24% reduction (Figures 4C, E, I). The total area of internalized membranes (endosome area) within growth cones was also used to quantify the resulting endocytosis during growth cone collapse. Both markers used to monitor endocytosis showed that Netrin-1 triggers a marked increase in endosome area per growth cone (Figures 4F, J), as well as an increase in the percentage of the growth cone area occupied by internalized structures (endosome occupancy) (Figures 4G, K), showing that membrane internalization is an indicative parameter to appreciate changes during the initial phase of growth cone collapse, suggesting that the growth cone starts collapsing through the internalization of its cell surface components.

To test whether Netrin-1-induced growth cone collapse requires clathrin-dependent endocytosis, we used a combination of pharmacological and dominant negative strategies. Monodansylcadaverine (MDC) is a potent inhibitor of clathrin-dependent endocytosis (Schutze et al., 1999), and Tyrphostin A23 (AG18) is an inhibitor of the binding of internalized cargo to AP-2 (Banbury et al., 2003). We found that neither AG18 nor MDC was able to prevent the collapse of growth cones (Figures 5A, B) or the formation of Netrin-1-induced internalization of vesicles in growth cones (Figure 5C). Eps15-Δ95/295 is a dominant-negative form of Eps15 (EGFR pathway substrate clone 15) lacking the Eps15 homology domains. This truncated molecule inhibits the plasma membrane targeting of AP-2 and clathrin and the formation of coated pits and clathrin-dependent endocytosis (Benmerah et al., 1999; Poupon et al., 2002). Mutant Eps15-Δ95/295 overexpression in EGL neurons did not prevent growth cone collapse in neurons treated with Netrin-1 (Figures 5D, E). These results suggested that Netrin-1-induced large vesicle endocytosis and growth cone collapse are independent of clathrin-dependent internalization mechanisms.

Macropinocytosis is a clathrin-or-caveolin-independent form of endocytosis that forms large endocytic vacuoles (Ferreira and Boucrot, 2018; Lin et al., 2020). The incorporation of high-molecular-weight (HMW, >10 kDa) dextran has been associated with membrane retrieval by macroendocytic vesicles (Kabayama et al., 2009; Kolpak et al., 2009). We investigated Netrin-1-induced membrane retrieval through macropinocytosis by analyzing the internalization of Fluorescein-HMW-dextran in the absence or presence of 5-(N-ethyl-N-isopropyl) amiloride (EIPA), a potent analog of amiloride channels that has been used as a macropinocytosis-specific inhibitor (Meier et al., 2002; Koivusalo et al., 2010; Sundaramoorthy et al., 2013; Tian et al., 2014; Zeineddine and Yerbury, 2015; Costa Verdera et al., 2017). Our results revealed that incubation with Netrin-1 induced the internalization of HMW-dextran (Figure 6A, arrowhead). Pre-incubation with EIPA resulted in a minor increase in basal macropinocytosis, but the Netrin-1-dependent macropinocytosis was entirely blocked by pre-treatment with EIPA (Figures 6A, C). We next asked whether the blockade of macropinocytosis also influenced Netrin-1-induced growth cone collapse. Incubation with EIPA was also able to dramatically abolish Netrin-1-induced growth cone collapse in EGL growth cones (Figures 6B, D). To evaluate whether axonal chemorepulsion to Netrin-1 was also associated with macropinocytosis, postnatal EGL explants were grown confronted to Netrin-1 in the presence of EIPA. Our results showed that the blockade of macropinocytosis with EIPA could inhibit the axonal chemorepulsion toward Netrin-1 (Figures 6E, F). Taken together, our data indicate that macropinocytosis, but not clathrin-dependent endocytosis, mediates Netrin-1-induced growth cone collapse and axon repulsion in EGL axons.

It has recently been reported that Stx1 is involved in rapid and slow endocytosis at synapses, which can be either clathrin-independent or -dependent (Xu et al., 2013). As Stx1 interacts with UNC5 receptors, which mediate chemorepulsion and growth cone collapse, we next determined whether Netrin-1-induced growth cone macropinocytosis and collapse were Stx1-dependent using BoNTs. BoNT/A or BoNT/C1 was added to the medium 10 min before incubation with Netrin-1. The percentage of macropinocytic vesicle-containing growth cones increased about twofold when Netrin-1 was added alone or in the presence of BoNT/A (Figures 7A, C). In contrast, Netrin-1-induced formation of macropinocytic vesicles was reduced specifically by cleaving Stx1 using BoNT/C1 (Figures 7A, C). We then tested whether BoNT/A or BoNT/C1 altered growth cone collapse. Neurons treated with control medium or with control medium supplemented with BoNT/A or BoNT/C1 exhibited normal growth cone morphologies with lamellipodia and filopodia (Figure 7B). After incubation with Netrin-1, about 70% of EGL growth cones exhibited collapsing and round-tipped shapes (Figures 7B, D). This effect was maintained when EGL neurons were co-incubated with BoNT/A, but was markedly reduced when the cultures were co-incubated with BoNT/C1 (Figure 7D). These results suggested that Stx1 is necessary for both Netrin-1-induced growth cone collapse and macropinocytosis.

To confirm these findings, we downregulated both Stx1 paralogs using the Stx1A-1B shRNA construct. Electroporation of EGL cells with Stx1A-1B shRNA, but not with control scrambled shRNA sequence, blocked the formation of macropinocytic vesicles in EGL growth cones incubated with Netrin-1 (Figures 7E, G). Similarly, the knockdown of Stx1 blocked Netrin-1-dependent collapsing of growth cones (Figures 7F, H). Our results, therefore, demonstrate that Stx1 is required for both Netrin-1-dependent growth cone collapse and macropinocytosis.