(A) Two-directional immunoprecipitations of neurofibromin and VCP. Adult rat brain extracts were used for immunoprecipitation using the indicated antibodies. Immunoblotting was then performed. (B) Co-immunoprecipitation of p47, VCP, and neurofibromin. HEK293T cells were transfected with Myc-tagged VCP and Myc-tagged p47 or vector control as indicated. One day later, cell extracts were harvested for immunoprecipitation using neurofibromin antibody. Immunoblotting using Myc tag antibody revealed the presence of both VCP and p47. (C–F) Interaction between neurofibromin and VCP in transfected HEK293T cells. Variant Myc-tagged VCP constructs and HA-tagged NF1 constructs were cotransfected into HEK293T cells as indicated, and cell lysates were harvested 24 hours later and immunoprecipitated using antibodies as indicated. (D) Myc-tagged D1D2 construct was cotransfected with variant neurofibromin fragments. GRD1^DN, dominant negative mutant of GRD1. (E) 2 mM DTT was added in lysates to reduce oxidation. (G) GST pull-down assay. His-tagged D1D2 of VCP was mixed with LRD or GRD1 GST fusion proteins. The protein complex was then pulled down using glutathione agarose and analyzed by immunoblotting.

(A) Immunostaining using Myc-tag antibody reveals the similar expression and distribution of Myc-tagged VCP, D1D2, and N-domain in cultured hippocampal neurons. (B) Myc-tagged VCP D1D2 construct and (E) Myc-tagged N-domain and vector control were cotransfected with GFP-actin, as indicated, into cultured hippocampal neurons at 12 DIV. Six days later, the neuronal morphology was monitored by detection of GFP immunoreactivity. In B, the lower panels show enlarged images (×5.3) of the boxed regions marked in the upper panels. (C and F) Cumulative probability distributions and (D and G) graph of protrusion densities obtained from B and E, respectively. More than 22 neurons and 86 dendrites for each group of experiments were analyzed. P < 0.05, DID2 versus control; **P < 0.01. Scale bars: 10 μm (A); 20 μm (B); 5 μm (E). Values are presented as mean plus SEM.

Mouse cultured cortical neurons were transfected at DIV7 (A) or DIV12 (D) and harvested for staining using neurofibromin and GFP antibodies at DIV10 (A) or DIV18 (D). For WT neurons, GFP-actin and vector control were transfected. For Nf1^+/– neurons, GFP-actin was cotransfected with full-length rNf1, full-length Y1587Δ mutant, or vector control, as indicated. B and C, and E and F are quantitative analyses of A and D, respectively. Values in B and E represent the mean plus SEM. The numbers of analyzed dendrites for each group were: (B and C) WT, 79; Nf1^+/– control, 106; Nf1^+/– rNf1, 37; Nf1^+/– Y1587Δ, 59; (E and F) WT, 94; Nf1^+/– control, 110; Nf1^+/– rNf1, 49; Nf1^+/– Y1587Δ, 44. Scale bars: 10 μm in the whole images; 2 μm in the enlarged images (A); 2 μm (D). In C, P < 0.01, Nf1^+/– rNf1 versus Nf1^+/– control; P < 0.001, Nf1^+/– control versus WT and Nf1^+/– rNf1 versus Nf1^+/– Y1587Δ. In F, P < 0.001, Nf1^+/– control versus WT and versus Nf1^+/– rNf1, and Nf1^+/– Y1587Δ versus Nf1^+/– rNf1. **P < 0.01, ***P < 0.001.

(A and B) Cultured cortical neurons isolated from Nf1^+/– mice and WT littermates were cotransfected with GFP-actin and Myc-tagged WT VCP or R95G mutant or vector control at 12 DIV. The density of dendritic spines was determined at 18 DIV by staining using GFP and Myc antibodies. (A) Mean values plus SEM of protrusion density. **P < 0.01, ***P < 0.001. (B) Cumulative probability of protrusion density. In total, 19–26 neurons and 31–82 dendrites were analyzed. P < 0.01, WT R95G versus WT WT and Nf1^+/– control versus WT control; P < 0.001, Nf1^+/– WT versus Nf1^+/– control and versus Nf1^+/– R95G. (C) Subcellular distribution of VCP in Nf1^+/– brain. Subcellular fractions of WT and Nf1^+/– brains were isolated by a series of centrifugations (95) and analyzed by immunoblotting. Synaptophysin (SVP38) was used as a loading control of LP2. H, total homogenate; P1, crude nuclei, unbroken cells, and debris; S1, supernatant of P1; P3, light membrane fraction (including ER and Golgi body); S3, soluble cytosolic fraction; LP1, lysed synaptosomal membrane; LP2, crude synaptic vesicles; LS2, soluble synaptic cytosol. The signals were quantified using Gel-Pro Analyzer (Media Cybernetics). The intensity ratios of WT to Nf1^+/– of each fraction are shown.

(A) First branches, (B) second branches, and (C) the remaining higher-order branches of apical dendrites of CA1 pyramidal cells. For each group, 2 representative images are shown. The right panels indicate the quantitative results of the spine density. Data represent mean plus SEM. The numbers of analyzed dendrites were (A) WT, 51; Nf1^+/–, 55; (B) WT, 21; Nf1^+/–, 12; (C) WT, 51; Nf1^+/–, 30. Scale bars: 2 μm. ***P < 0.001.

(A) Knockdown effect of the VCP miRNA construct in HEK293T cells. Myc-tagged WT VCP (VCP^WT) or a silent mutant of VCP (VCP^mt) and non-silencing control or VCP miRNA construct were cotransfected into cells. One day later, cells were harvested for immunoblotting analysis using Myc tag and GFP antibodies. (B) Knockdown of VCP in cultured hippocampal neurons. Transfection of VCP miRNA and non-silencing control was carried out at 12 DIV. Immunostaining using VCP antibody was performed at 18 DIV. The relative VCP protein levels revealed by immunostaining were quantified. Values are presented as means plus SEM. Non-silencing, n = 51; VCP miRNA, n = 35. Scale bar: 10 μm. (C) Cell morphology of VCP-knockdown neurons. Cultured hippocampal neurons were transfected with various plasmids (Control: non-silencing control plus GW1 vector control; miRNA: VCP miRNA plus GW1 vector control; Rescue: VCP miRNA plus miRNA-resistant VCP silent mutant) at 12 DIV and harvested for immunostaining using a GFP antibody at 18 DIV. The lower panels show enlarged images (×5.3) of the boxed regions marked in the upper panels. Scale bar: 20 μm. (D–F) Cumulative probability distributions and means plus SEM of protrusion density (D), protrusion width (E), and protrusion length (F). (D–F) More than 48 neurons, 193 dendrites, and 794 spines for each group from 2 independent experiments were analyzed. In D, P < 0.01, rescue versus control; P < 0.001, miRNA versus control and versus rescue. **P < 0.01, ***P < 0.001.

(A) Mutations in the LRD region of the NF1 gene identified from NF1 patients. Locations of mutation sites are indicated by asterisks. (B) Cotransfection of GFP-actin and vector control, HA-tagged WT LRD, or LRD mutants (Y1587Δ and C1909R) into cultured hippocampal neurons at 12 DIV. Six days later, immunostaining was performed using GFP and HA tag antibodies. Only GFP signals are shown. Scale bar: 5 μm. (C) Cumulative probability distributions and (D) graph of protrusion densities. More than 52 neurons and 224 dendrites collected from 2 independent experiments were analyzed. In C, P < 0.01, C1909R versus control and versus Y1587Δ; P < 0.001, C1909R, Y1587Δ, and control versus WT. ***P < 0.001. Values are presented as mean plus SEM. (E) Myc-tagged D1D2 and (F) Myc-tagged VCP were cotransfected with HA-tagged WT and mutant LRD into HEK293T cells as indicated. One day later, total cell extracts were harvested and immunoprecipitated with a Myc antibody. Immunoblotting analysis was then performed using sequentially HA and Myc tag antibodies.

(A) LRD and (B) neurofibromin interact weakly with IBMPFD mutants. Variant Myc-tagged VCP constructs and vector control were cotransfected with HA-tagged LRD or singly transfected into HEK293T cells. One day later, (A) Myc and (B) neurofibromin antibodies were used for immunoprecipitation. Immunoblotting analysis was performed using the antibodies as indicated. The relative intensities to WT lane are shown. Distribution of Myc-tagged VCP and IBMPFD mutants in (C) HEK293T cells and (H) cultured hippocampal neurons was examined by immunostaining using Myc and GFP antibodies. Cotransfection with GFP was performed to outline cell morphology. Expression of (D) IBMPFD mutants, (E and F) VCP K524A mutant, and (G) VCP miRNA construct did not obviously influence the total protein level of neurofibromin. HEK293T cells were transfected with variant VCP constructs, VCP miRNA, or vector control. Immunoblotting was then performed to examine the protein levels of neurofibromin. CASK and/or actin were used as internal controls. In E, F, and G, MG132 was added to cultures 4 hours before harvest or omitted. Ubiquitin antibody FK2 (αUb) was also included in immunoblotting analysis. For F and G, immunoprecipitation was performed using neurofibromin antibody to precipitate endogenous neurofibromin. Immunoblotting using FK2 antibody did not reveal obvious signals in the neurofibromin precipitates. (I) IBMPFD mutants decrease spine density. Cultured hippocampal neurons were cotransfected with GFP-actin and vector control, WT VCP, or IBMPFD mutants (R95G and R155H) at 12 DIV. Cells were fixed at 18 DIV and stained with Myc and GFP antibodies. The morphology of spines was then analyzed based on the GFP signals. (J) Protrusion density. Values are presented as mean plus SEM. (K) Cumulative probability distributions. More than 34 neurons and 146 dendrites for each group collected from 2 independent experiments were analyzed. Scale bars: (C) 10 μm; (H) upper panels, 10 μm; lower panels, 5 μm; (I) 2 μm. In K, P < 0.05, R155H versus control; P < 0.01, R95G versus WT; P < 0.001, R95G versus control. *P < 0.05, ***P < 0.001.

Cultured hippocampal neurons were cotransfected with GFP-actin and vector control or VCP miRNA at 12 DIV and treated with lovastatin (Lova.) at 15 DIV for 3 days or simvastatin (Simva.) at 14 DIV for 4 days. Quantitative measurements of spine density are shown as (A) cumulative probability distributions and (B and D) graph of protrusion densities. For A and B, more than 53 neurons and 213 dendrites for each group were analyzed. For D, more than 30 dendrites for each group were analyzed. In the presence of statins, knockdown efficiency of VCP miRNA assayed by immunostaining using VCP antibody was also evaluated in C and E. For VCP-knockdown neurons, somas were outlined based on the GFP signal. In C, sample sizes of each group were as follows: vehicle plus control, 51; vehicle plus miRNA, 35; lovastatin plus control, 59; lovastatin plus miRNA, 26. In A, P < 0.001, miRNA vehicle versus control vehicle and miRNA lovastatin 2 μM versus control lovastatin 2 μM. *P < 0.05, ***P < 0.001. Scale bars: 10 μm (C and E).