Wild-type (WT) C57BL/6 J mice were purchased from Japan SLC. PlexinD1-flox mice (Zhang et al., 2009) were described previously. Sema3E-KO mice (Gu et al., 2005) were provided by Dr. Fanny Mann (Institut de Biologie du Developpement de Marseille) and Dr. Christopher E. Henderson (Columbia University). RhoJ-KO mice were described previously (Kim et al., 2014). Both male and female animals were used in this study. All animals were maintained with their mother mouse during weaning and within 7 mice per cage after weaning on a 12 h light/dark cycle with ad libitum access to food and water. All of the animal experiments were performed in accordance with the guidelines and regulations of Nagoya City University (Approval No. 21-028).

CSII-CMV-RfA-IRES2-Venus and CSII-EF-Venus lentiviral vectors were provided by Dr. Hiroyuki Miyoshi (RIKEN Tsukuba BioResource Center). CSII-CMV-PlexinD1-IRES2-Venus and CSII-CMV-PlexinD1ΔRBD-IRES2-Venus were described previously (Sawada et al., 2018). The pLV-CMV-tdTomato-IRES-Cre was provided by Dr. Magdalena Götz (Munich University). These viral vectors and the packaging vectors (pCAG-HIVgp and pCMV-VSV-G-RSV-Rev) were co-transfected into HEK293T cells using polyethylenimine to generate lentiviral particles, and then the culture supernatants were concentrated by centrifuging at 8,000 rpm for 16 h at 4°C using a refrigerated microcentrifuge (MX-307, Tomy) and resuspended with sterile phosphate buffered saline (PBS).

pEGFPC2 and pEGFPC2-RhoJ were described previously (Kusuhara et al., 2012; Fukushima et al., 2020). pCAGGS-DsRed was described previously (Sawada et al., 2018). These plasmids were amplified using E. coli and purified using a PureLink HiPure Plasmid Midiprep kit (Invitrogen).

The V-SVZ cell culture was performed as described previously (Sawada et al., 2018). Briefly, the V-SVZ tissues were dissected from P0-1 WT, PlexinD1^+/fl, and PlexinD1^fl/fl pups and dissociated with trypsin–EDTA (Invitrogen). The cells were washed two times in L-15 medium (Invitrogen) containing 40 μg/mL DNase I (Roche), seeded on coverglass (Matsunami) in 24-well cell culture plates, and cultured in Neurobasal medium (GIBCO) containing 10% fetal bovine serum, 2% NeuroBrew-21 (MACS Miltenyi Biotec), 2 mM L-glutamine (GIBCO), and 50 U/mL penicillin–streptomycin (GIBCO). At 4 days in vitro (div), 1 μL of lentiviral particles was added into the culture medium. At 10 div, cells were fixed with 4% paraformaldehyde (PFA) in 0.1 M phosphate buffer (PB) for 15 min at room temperature (RT) and subjected to immunocytochemistry.

Injection of lentiviral suspension and plasmid solution was described previously (Ota et al., 2014; Sawada et al., 2018). For lentiviral infection, a 2 μL volume of lentiviral suspension was stereotaxically injected into the V-SVZ (1.8 mm anterior, 1.4 mm lateral to lambda and 1.5–2.0 mm deep) of male and female pups at postnatal day 1 (P1). For in vivo electroporation of P1 male and female pups, a 2 μL volume of plasmids (3.5 μg/μL) containing 0.05% Fast Green was stereotaxically injected into the lateral ventricle (1.8 mm anterior, 1.2 mm lateral to lambda and 2.0 mm deep), and electroporation was performed using an electroporator (CUY-21SC, Nepagene) with a forceps-type electrode (CUY650P7, Nepagene). Lentivirus-injected or electroporated pups were allowed to recover on a heating pad and returned to their home cage.

Immunohistochemistry was performed as described previously (Sawada et al., 2018). Briefly, the brain was fixed by transcardiac perfusion with 4% PFA in 0.1 M PB, and postfixed overnight in the same fixative at 4°C. Sixty-micrometer-thick coronal sections were prepared using a vibratome (VT-1200S, Leica) and incubated for 30 min at RT in blocking solution (10% normal donkey serum and 0.2% Triton X-100 in PBS). The sections were then incubated overnight at 4°C with the primary antibodies, and then for 2 h at RT with biotin- or Alexa Fluor-conjugated secondary antibodies (1:1,000, Invitrogen) in the blocking solution. For immunocytochemistry, fixed cells were incubated for 30 min at RT in blocking solution, overnight at 4°C with the primary antibodies, and then for 2 h at RT with Alexa Fluor-conjugated secondary antibodies (1:1,000, Invitrogen) in the blocking solution. The signal amplification and visualization were performed using the Vectastain Elite ABC kit (Vector Laboratories) and Tyramide Signal Amplification (Thermo Fisher Scientific), respectively. The following primary antibodies were used: mouse anti-βIII-tubulin (Tuj1) (1:1,000, Sigma); rat anti-CD31 (1:100, BD Pharmingen); rabbit anti-doublecortin (Dcx) (1:200, Cell Signaling Technology); guinea pig anti-Dcx (1:500, Chemicon); rabbit anti-DsRed (1:1,000, Clontech); rabbit anti-GFP (1:500, MBL); rat anti-GFP (1:500, Nakalai Tesque); rabbit anti-NeuN (1:1,000, abcam); goat anti-PlexinD1 (1:100, abcam); goat anti-PlexinD1 (1:100, R&D systems) antibodies. Nuclei were stained with Hoechst 33342 (1:5,000, Sigma).

Image acquisition and dendritic tracing of granule cells were performed as previously described (Sawada et al., 2018). Images of labeled granule cells of the postnatal OB were acquired by scanning at 2 μm intervals using an LSM 700 confocal laser-scanning microscope (Carl Zeiss) with a 20× objective lens (NA 0.8). For characterization of RhoJ^+/GFP-positive cells, three regions-of-interest (ROIs) were randomly selected from three consecutive sections (one ROI per section) from every sixth 60 μm-thick OB section and image acquisition was performed, and all of the positive cells in the ROIs were counted. The obtained proportions of RhoJ^+/GFP-positive cells were reported as mean ± SEM (n = 3 mice). Acquired z-stack images of virally labeled neurons were traced, reconstructed, and quantified using Neurolucida and Neurolucida Explorer software (MBF Bioscience) (Sawada et al., 2018). For analysis of apical and basal dendrites, all of the labeled granule cells showing complete neuronal morphologies observed in every sixth section were traced and analyzed. For analysis of dendritic branching in the proximal domain of the apical dendrite, all of the labeled granule cells showing >100 μm-length apical dendrites from soma were analyzed in this study. Dendritic branching in all domains of granule cells was expressed as branching in the dendritic domain per cell. For validation of PlexinD1 expression in neuronal culture, the signal intensity in the cell surface of labeled cells was measured using ZEN software (Carl Zeiss), and the average signal intensity (per μm), and normalized value (average value of control groups is 1.0) were calculated as reported previously (Sawada et al., 2018).

All of the data were two-tailed and analyzed using EZR (Kanda, 2013). The data distribution was analyzed by the Kolmogorov–Smirnov test. The equality of variance between groups was analyzed by the F test. Comparisons between two groups were analyzed by unpaired t-test or Mann–Whitney U-test. Comparisons among multiple groups were analyzed by one-way ANOVA test followed by a post-hoc Tukey–Kramer test, or Kruskal-Wallis test followed by a post-hoc Steel-Dwass test. All the numerical data are presented as the mean ± SEM and p-values less than 0.05 were considered statistically significant.

In the GCL of the postnatal OB, while Sema3E is expressed in mature granule cells, PlexinD1 is expressed in the leading process of migrating newborn neurons and is involved in suppressing FLP formation to maintain their immature morphology (Sawada et al., 2018). In this study, we investigated whether Sema3E-PlexinD1 signaling is also involved in the suppression of dendritic branching in the proximal domain of the apical dendrite of newborn granule cells that have terminated their migration.

First, we injected lentivirus encoding Venus into the V-SVZ of Sema3E^+/− (control) and Sema3E^−/− (Sema3E-KO) mice (Gu et al., 2005) at postnatal day 1 (P1) and analyzed the dendritic morphology of Venus+ granule cells in the GCL at 10 days post infection (dpi) (Figure 1A). The branch number in the proximal domain of the apical dendrite was significantly increased in Sema3E-KO mice compared with control mice (Figures 1B–D), suggesting that Sema3E suppresses dendritic branching in the proximal domain of the apical dendrite of newborn granule cells.

Next, to investigate the expression and role of PlexinD1 in this process, brain sections were stained with a PlexinD1 antibody. The PlexinD1 signal was observed in the proximal domain of the apical dendrite of differentiating granule cells (Supplementary Figure S1A). For functional analyses of PlexinD1, we used lentivirus encoding tdTomato and Cre recombinase. The decrease of PlexinD1 proteins by these vectors was confirmed in primary neuronal cultures (Figures 1E,F). To examine the function of PlexinD1 in vivo, we injected these lentiviral vectors into the V-SVZ of PlexinD1^+/flox (control) and PlexinD1^flox/flox (PlexinD1-cKO) mice (Zhang et al., 2009) at P1 and analyzed the dendritic morphology of tdTomato+ granule cells (Figure 1G). At 10 dpi, similar to the phenotype observed in Sema3E-KO mice, the branch number in the proximal domain of the apical dendrite was significantly increased by PlexinD1 deficiency (Figures 1H–J). Moreover, at 28 dpi, when granule cells morphologically mature (Petreanu and Alvarez-Buylla, 2002), the whole dendritic morphology of tdTomato+ granule cells was traced (Figure 1K). We found that the dendritic branch number in the proximal but not the distal or basal domain in tdTomato+ granule cells was significantly increased in PlexinD1-cKO mice compared to that in control mice (Figures 1L–P). The dendritic length of the apical and basal dendrites was not significantly different between control and PlexinD1-cKO mice (Supplementary Figures S1B–E). Taken together, these results suggest that Sema3E-PlexinD1 signaling is specifically involved in the suppression of dendritic branching in the proximal domain of the apical dendrite of newborn granule cells in the postnatal OB.

The intracellular domain of Plexins has a Rho binding domain (RBD), which regulates cytoskeletal dynamics through binding to various Rho family small GTPases (Gay et al., 2011) (Figure 2A). To study the role of PlexinD1 and its RBD in dendritic development of newborn granule cells, we generated lentiviral vectors encoding PlexinD1 or its deletion mutant of RBD (PlexinD1ΔRBD), and confirmed their overexpression in primary neuronal cultures (Figures 2A–D). To examine the effects of PlexinD1 overexpression on the dendritic branching in vivo, we injected these lentiviruses into the V-SVZ and analyzed the dendritic morphology of infected granule cells in the OB at 10 dpi (Figure 2E). Overexpression of PlexinD1 decreased the number of dendritic branches in the proximal domain of the apical dendrite (Figures 2F,G,I). This effect was partially diminished when lentiviruses encoding PlexinD1ΔRBD were injected (Figures 2H,I). These results suggest that RBD is involved in the suppression of dendritic branching in the proximal domain of the apical dendrites of newborn granule cells by PlexinD1 overexpression.

RhoJ, a member of Rho family small GTPases, is highly expressed in vascular endothelial cells and involved in Sema3E’s repulsive effect by directly binding to the PlexinD1RBD to promote F-actin depolymerization (Fukushima et al., 2011, 2020). In the developing retina, RhoJ is expressed in not only vascular endothelial cells but also neurons (Fukushima et al., 2020). However, the function of RhoJ in the CNS remains unknown. Since RBD is involved in the PlexinD1-induced suppression of dendritic branching in the proximal domain of apical dendrites (Figure 2), we hypothesized that RhoJ contributes to this process.

First, to identify RhoJ-expressing cells in the V-SVZ-OB pathway, we analyzed GFP-expressing cell types in RhoJ^+/GFP mice (Kim et al., 2014), in which EGFP replaces exon 1 of the RhoJ gene and is expressed under the control of the endogenous RhoJ promoter. EGFP was strongly expressed in the CD31+ vascular endothelial cells not only in the retina (Fukushima et al., 2020) but also in the brain (Figure 3A). We found that EGFP was not expressed in Dcx + cells in the V-SVZ, but was expressed in a subset of Dcx + and NeuN+ cells in the core and GCL of the OB (Figures 3B–E). These results suggest that newborn neurons express RhoJ during migration and maturation in the postnatal OB.

Next, to investigate the effect of forced expression of RhoJ on the dendritogenesis of newborn granule cells, plasmids encoding an EGFP-fused RhoJ were introduced into the V-SVZ by in vivo electroporation, and the dendritic morphology of labeled granule cells in the GCL was analyzed at 10 dpi (Figure 3F). The forced expression of RhoJ in granule cells significantly suppressed both their dendritic branching in the distal and proximal domains (Figures 3G–J) and dendritic length in the distal domain (Figure 3G; control, 340.0 ± 28.5 μm; RhoJ, 224.5 ± 29.0 μm; p = 0.0063, unpaired t-test). These results suggest that RhoJ overexpression has an inhibitory effect on dendritic branching and outgrowth in vivo.

Finally, to study the function of RhoJ on dendritogenesis of newborn granule cells, we introduced plasmids encoding DsRed into the V-SVZ of RhoJ^+/GFP (control) and RhoJ^GFP/GFP (RhoJ-KO) mice by in vivo electroporation and analyzed their dendritic morphology (Figure 3K). The dendritic branch number in the proximal but not distal or basal domain in DsRed+ granule cells was significantly increased in RhoJ-KO mice (Figures 3L–O). The dendritic length of the apical and basal dendrites was not significantly different between control and RhoJ-KO mice (Supplementary Figures S1B,F–H). Furthermore, the increase of dendritic branching in the proximal domain in RhoJ-KO neurons was partially diminished by PlexinD1 overexpression (Figures 3L–O). Together, these results suggest that RhoJ is specifically involved in the suppression of dendritic branching in the proximal domain of the apical dendrite of newborn granule cells in the postnatal OB.