The F-BAR domain of srGAP2 induces membrane protrusions required for neuronal migration and morphogenesis

Abstract

During brain development, proper neuronal migration and morphogenesis is critical for the establishment of functional neural circuits. Here we report that srGAP2 negatively regulates neuronal migration and induces neurite outgrowth and branching through the ability of its F-BAR domain to induce filopodia-like membrane protrusions resembling those induced by I-BAR domains in vivo and in vitro. Previous work has suggested that in nonneuronal cells filopodia dynamics decrease the rate of cell migration and the persistence of leading edge protrusions. srGAP2 knockdown reduces leading process branching and increases the rate of neuronal migration in vivo. Overexpression of srGAP2 or its F-BAR domain has the opposite effects, increasing leading process branching and decreasing migration. These results suggest that F-BAR domains are functionally diverse and highlight the functional importance of proteins directly regulating membrane deformation for proper neuronal migration and morphogenesis.

Figure 1. srGAP2 is expressed in neuronal progenitors and post-mitotic neurons and localizes to sites of membrane protrusion

(A) In situ hybridization for srGAP2 in developing cortex at embryonic day (E)13 and E15 and postnatal day (P)1. (B) Domain organization of srGAP2 which contains an F-BAR domain, a RhoGAP and a SH3 domain from N-to C-terminal ends (1–1045 aa, predicted MW 118kDa). The black bar indicates the localization of the antigen (A2, aa 873–890) used to affinity purify the srGAP2-specific polyclonal antibody to the C-terminus of srGAP2 (Yao et al., 2008). (C) Western blot for srGAP2 protein levels during cortical development at the indicated time points (E15, P1, P15 and Adult) obtained by SDS-PAGE and immunoblotting with A2-rabbit polyclonal antibody (Yao et al, 2008). (D–J) Immunofluorescence staining of srGAP2 protein expression on fixed coronal sections of E15 mouse cortex. srGAP2 protein colocalizes (arrowheads) with MAP2 (postmitotic neuron marker) in the CP (D–F) and also colocalizes with Nestin (arrowheads) (neuronal precursor marker) in the VZ (G–I). (K–P) Immunofluorescence staining of srGAP2 protein in early dissociated cortical neuron cultures (E15+ 24 hours in vitro –hiv). srGAP2 protein is found close to the plasma membrane of immature cortical neurons (arrow in K-M) and to F-actin-rich filopodia (stained with Alexa546-phalloidin; arrowheads in K–M and N–P).

Figure 2. F-BAR induced filopodia required F-actin for their dynamic formation but not for their structural maintenance

(A–C) COS7 cell expressing the F-BAR-EGFP fusion protein not treated with cytochalasin D (control). Note the cortical localization of the F-BAR domain and the numerous F-actin-rich filopodia (phalloidin in B and C). (D–F) COS7 cell expressing the F-BAR-EGFP fusion protein incubated with 400μM cytochalasin D 30 minutes. Note that the complete loss of F-actin (phalloidin; E) had no effect on the localization of the F-BAR domain or on the structure of the F-BAR mediated protrusions. (H–K) Time series showing the dynamics of F-BAR-EGFP-induced filopodia in COS7 cells. Time 0, 5, and 10 minutes are pseudo-colored in red, green, and blue respectively. Note there is little colocalization of filopodia at the cell periphery (K). This is in stark contrast to COS7 cells expressing F-BAR-EGFP treated with Cytochalasin D (30 minutes) (L–O) where the protrusions remain static and do not grow or retract for the same period of time shown in control cells. (P) Schema depicting tubulation assay in Q. (Q) F-BAR domain of srGAP2 added to preformed liposomes. Note the inward dimpling or “scalloping” of the liposome surface. (R) Schema depicting tubulation assay in (S) where F-BAR domain of srGAP2 was added to liposomes after extrusion. This results in a fraction of the F-BAR domain resident inside the liposome. Note the formation of tubule protrusion from the liposome. (S) High magnification of liposome/F-BAR mixture after sonication. Note the absence of striations or an obvious protein coat on the lipid tubule, a hallmark of canonical F-BAR tubulation. These tubules are 83 nm +/−15 nm (average +/− SD, N=38) after being partially flattened by the negative staining procedure.

Figure 3. Knockdown of srGAP2 in cortical neurons reduces axonal and dendritic branching

(A) Western blot probed with ant-GFP and anti-actin antibodies from COS7 cells co-transfected with either control shRNA plus srGAP2-EGFP (lane 1), srGAP2 shRNA plus srGAP2-EGFP (Dha2, lane 2) or (Dha5, lane3) (B) Western blot probed with anti-GFP and anti-actin antibodies from COS7 cells co-transfected with either control shRNA plus srGAP2-EGFP (lane 1), srGAP2 shRNA plus srGAP2-EGFP (lane 2), a mutated form of srGAP2*-EGFP (resistant to srGAP2 shRNA) plus control shRNA (lane 3), or srGAP2*-EGFP plus srGAP2 shRNA (lane 4). srGAP2 shRNA significantly knocks down srGAP2 expression compared to control shRNA which can be rescued by expression of srGAP2*-EGFP (compare lanes 3 and 4). (C–E, G–I) E15 dissociated cortical neurons were cultured for 4 days after ex vivo electroporation with control shRNA, srGAP2 shRNA, or srGAP2 shRNA + srGAP2*-EGFP. Control shRNA transfected neurons display frequent primary branches from the axon (arrowheads in B) and the primary dendrite (arrowheads in F). Both effects were markedly reduced in srGAP2 shRNA transfected neurons (D and H) and rescued by co-transfection of srGAP2 shRNA with srGAP2*-EGFP (E and I). (E) Quantification of the number of branches from the longest neurite (axon) as shown in C–E. (I) Quantification of the number of primary dendritic branches as shown in G–I. (Control shRNA, n=42 cells; srGAP2 shRNA, n=95; srGAP2*-EGFP + srGAP2 shRNA, n=39. Cells were taken from 3 independent experiments and analyzed blind to the treatment. Mann-Whitney Test * p<0.05; ** p<0.01; *** p<0.001.

Figure 4. srGAP2 promotes filopodia formation and neurite outgrowth in an F-BAR dependent manner

(A–E) Stage 1 cortical neurons expressing various srGAP2 constructs. All cells are stained with neuron-specific β-III tubulin (blue) to reveal presence of microtubules (see also Fig. S7) and phalloidin (red) to visualize F-actin. (F) Quantification of filopodia normalized per cell perimeter in all conditions. EGFP n= 20 cells; srGAP2-EGFP n= 21; srGAP2^ΔF-BAR-EGFP n= 20; F-BAR-EGFP n= 20; F-BAR^Δ49-EGFP n= 20. Cells were taken from 3 independent experiments and analyzed blind to the treatment. (G–K) Stage 2 cortical neurons expressing various srGAP2 constructs. All cells are stained with β-III tubulin (blue) and phalloidin (red) as in panels A–F. Arrows point to primary neurites and arrowheads point to neurite branches. (L) Quantification of neurite number normalized per cell perimeter in all conditions and primary branch number per neurite. Note srGAP2 and F-BAR are potent inducers of neurite outgrowth while srGAP2 ^ΔF-BAR and F-BAR^Δ49 are not. Mann-Whitney Test * p<0.05; ** p<.001; *** p<0.001. Green stars indicates comparison to EGFP and blue stars indicates comparison to srGAP2-EGFP.

Figure 5. Knockdown of srGAP2 promotes neuronal migration and reduces leading process branching

(A) E15 cortical slices cultured for 3 days after electroporation with EGFP + control shRNA. Slices were stained with anti-Nestin antibody revealing radial glial scaffold and Draq5 to illustrate cytoarchitecture. (B) E15 cortical slices cultured for 3 days after electroporation with EGFP + Dha2 (B, top panel) or Dha5 (B lower panel). Slices were stained with anti-Nestin antibody revealing radial glial scaffold and Draq5 to illustrate cytoarchitecture. (E–L) E15 cortical slices cultured for 2 days ex vivo after electroporation with Nuclear EGFP (3NLS) along with control shRNA (E–H) or srGAP2 shRNA (I–L) were imaged using time-lapse confocal microscopy. Neurons transfected with srGAP2 shRNA undergo faster translocation within 4 hrs (I–L and no colocalization in L) than control shRNA-transfected neurons. (M) Quantification of effects of srGAP2 knockdown on cell speed. Neurons with reduced level of srGAP2 (shRNA) migrated approximately 23% faster (6.91 μicrons/h compared to 5.59 μicrons/h) compared to control shRNA-transfected neurons. Control shRNA, n=95 cells; srGAP2 shRNA n=84. Cells were taken from 3 independent experiments. Mann-Whitney Test * p<0.05; ** p<.001; *** p<0.001. (N–O) High magnification images (N) and reconstructions (O) of control shRNA (left panel) or srGAP2 shRNA (right panel) expressing neurons in layers 5/6. Note the branched morphology of the leading process of control shRNA expressing neurons (red arrowheads pointing to leading process tips in N) whereas srGAP2 shRNA expressing neurons displayed a simpler, less branched morphology (green arrowhead in N pointing to single leading process tip). (P) Quantification of the leading process branch number in control shRNA or srGAP2 shRNA expressing neurons. Control shRNA, n=19 cells; srGAP2 shRNA n=17 cells. Cells were taken from 3 independent experiments. Mann-Whitney Test * p<0.05; ** p<.001; *** p<0.001.

Figure 6. srGAP2 mediated inhibition of migration requires F-BAR-mediated membrane deformation

(A–T) E15 cortical slices cultured for 5 days after co-electroporation of monomeric Red Fluorescence Protein (mRFP for cytoplasmic filing) together with EGFP (A–D), full-length srGAP2-EGFP (E–H), srGAP2^ΔF-BAR-EGFP (I–L), F-BAR-EGFP (M–P) and F-BAR^Δ49-EGFP fusion proteins (Q–T). Slices were stained with nuclear marker Draq5 in order to reveal the cytoarchitecture. (U) Quantification of CP/IZ ratio. EGFP n= 13 slices; srGAP2-EGFP n= 14 slices; srGAP2^ΔF-BAR-EGFP n= 8 slices; F-BAR-EGFP n= 10 slices; F-BAR^Δ49-EGFP n= 6 slice. Slices were taken from 4 different experiments and CP/IZ ratio analyzed using Mann-Whitney Test * p<0.05; ** p<.001; *** p<0.001. Green stars indicates comparison to EGFP and blue stars indicates comparison to srGAP2-EGFP. (V) Quantification of percentage of cells with multipolar morphology in EGFP, srGAP2-EGFP, or F-BAR-EGFP transfected slices. Multipolar cells were defined as cells possessing ≥ 3 processes. EGFP n= 66 cells; srGAP2-EGFP n= 42; F-BAR-EGFP n= 57. Cells were taken from 3 different experiments and analyzed using Fisher’s exact test * p<0.05; ** p<.001; *** p<0.001. (W–Y) Individual frames using time-lapse confocal microscopy of E15 cortical slices cultured for 3 days after electroporation with EGFP, srGAP2-EGFP, or F-BAR-EGFP (co-transfected with Venus plasmid). Arrows indicate leading processes and arrowheads the cell body. (Z) Quantification of leading process branch number from cells expressing EGFP, srGAP2-EGFP, or F-BAR-EGFP in layer 5/6. EGFP n= 17 cells; srGAP2-EGFP n= 21 cells; F-BAR-EGFP n= 9 cells. Cells were taken from 3 independent slices. Mann-Whitney Test * p<0.05; ** p<.001; *** p<0.001. Green stars indicates comparison to EGFP and blue stars indicates comparison to srGAP2-EGFP.

Figure 7. Model for srGAP2 regulated membrane protrusion in neuronal migration

(A–F) Representative images of optically isolated neurons translocating radially through layer 5/6 following electroporation at E15 (+5div) with indicated srGAP2 constructs containing a F-BAR domain. (G–H) Hypothetical model of the molecular mechanisms underlying srGAP2 function in membrane protrusion during neuronal migration and morphogenesis (G). Summary of srGAP2 effects on neuronal migration and morphogenesis during cortical development (H). See text for details.