DBZ regulates cortical cell positioning and neurite development by sustaining the anterograde transport of Lis1 and DISC1 through control of Ndel1 dual-phosphorylation

Abstract

Cell positioning and neuronal network formation are crucial for proper brain function. Disrupted-in-Schizophrenia 1 (DISC1) is anterogradely transported to the neurite tips, together with Lis1, and functions in neurite extension via suppression of GSK3β activity. Then, transported Lis1 is retrogradely transported and functions in cell migration. Here, we show that DISC1-binding zinc finger protein (DBZ), together with DISC1, regulates mouse cortical cell positioning and neurite development in vivo. DBZ hindered Ndel1 phosphorylation at threonine 219 and serine 251. DBZ depletion or expression of a double-phosphorylated mimetic form of Ndel1 impaired the transport of Lis1 and DISC1 to the neurite tips and hampered microtubule elongation. Moreover, application of DISC1 or a GSK3β inhibitor rescued the impairments caused by DBZ insufficiency or double-phosphorylated Ndel1 expression. We concluded that DBZ controls cell positioning and neurite development by interfering with Ndel1 from disproportionate phosphorylation, which is critical for appropriate anterograde transport of the DISC1-complex.

Keywords: anterograde transport; cortical development; microtubule; migration; neurite extension.

Figure 1.

Neurite development is impaired by DBZ deletion. A, DBZ mRNA was localized primarily in the subventricular/intermediate zone (SVZ). A representative coronal section of an E14.5 cortex is shown (left). The section was counterstained with Nissl staining (right). No obvious expression was observed in the ventricular zone (VZ), whereas strong signals were observed in the SVZ. Scale bar, 200 μm. B, Histological examination of the cortices of the DBZ^−/− (DBZ KO) and the littermate DBZ^+/+ mice (WT). High-magnification views of the red boxes are shown on the right. Scale bars, 200 μm. C, Representative photomicrographs show coronal sections of the cortices at E17.5 and at P3 stained using Bodian's method. The DBZ^−/− mice showed poor neurite development. High-magnification images in the squares on the left are shown in the next right panels. Scale bars, 100 μm. D, Examples of Golgi-impregnated P3 cerebral cortices of a WT and a DBZ KO mouse are shown. Magnified images are shown in the right. Pyramidal neurons with thin and less-branched dendrites were frequently observed in the DBZ KO. Scale bars, 40 μm.

Figure 2.

Cell arrangement is impaired in the cortex due to DBZ deletion. A, Neuronal cell arrangement was altered by DBZ deletion. The thymidine homologs CldU and IdU were consecutively injected at E13.5 and E16.5, respectively, and the localization of labeled cells was observed at P2 in layer II/III (II/III). B, Distributions of CldU- and IdU-labeled cells in the cortex. The cortex was divided into 10 bins, and the labeled cells in each bin were counted. C, Immunohistochemistry against the proliferation marker Ki67 in an E15.5 mouse cortex. Scale bar, 200 μm. D, The labeled cells in each bin were counted. The values represent the mean ± SEM; *p < 0.05.

Figure 3.

Acute DBZ insufficiency leads to impaired radial migration, poor neurite elongation and a loss of centrosome-nucleus coupling. A, The efficiencies of RNA interference with different shRNAs against DBZ (Nos. 1–3) were evaluated. No. 1 markedly suppressed EGFP-DBZ expression and was subsequently used for all DBZ-knockdown experiments. B, shRNAs against DBZ No. 1 suppressed endogenous DBZ expression in N2a cells. GAPDH was used as a loading control. C, The mU6pro scramble vector (mU6 sc) was used as a control (left). DBZ-knockdown radially migrating cells did not migrate far beyond the intermediate zone (right). CP, Cortical plate; IZ, intermediate zone. The neocortex was subdivided evenly into five bins (1–5) from the ventricular zone (VZ) to the CP. The occurrences of GFP-labeled cells (cell bodies) in each bin were counted. D, DBZ-knockdown neurons that were cotransfected with the hDBZ migrated into the cortical plate (left), whereas sole DBZ-knockdown neurons did not migrate (right, the same image of C). Scale bar, 100 μm. E, Neurons in the cortical plate were examined in the brain slices. Neurons, their centrosomes and their nuclei, were labeled with DS-Red2 (shown in red), EGFP-CETN2 (green), and Hoechst (blue), respectively. The nucleus–centrosome distances were greater in neurons with RNAi against DBZ compared with controls. CETN2, Centrin 2. Arrows indicate centrosomes, and arrowheads represent the edges of nuclei. Scale bar, 5 μm. F, At E17.5, DBZ-expressing neurons were collected from the cortices and dissociated for primary culture. The lengths of the primary neurites were measured after culture for 96 h. Acute DBZ-insufficient neurons had shorter primary neurites than controls. Scale bar, 100 μm. The values represent the mean ± SEM; *p < 0.05.

Figure 4.

Acute DBZ insufficiency does not significantly alter the migration of radially migrating neurons in DBZ KO mouse. A, B, The mU6pro scramble vector (mU6 sc) was used as a control (left). DBZ-knockdown radially migrating neurons of the DBZ KO mouse did not exhibit any significant difference in terms of localization. Vectors were injected at E14.5 and brains were observed at E18.5. The neocortex was subdivided evenly into five bins. Occurrences of GFP-labeled cells (cell bodies) in each bin were counted. Scale bar, 100 μm. Values represent the mean ± SEM.

Figure 5.

DBZ regulates Ndel1 phosphorylation, and the phosphorylation-status of Ndel1 is critical for radial migration. A, Cell lysates from the cerebral cortices of E18.5 DBZ KO and WT were subjected to Western blot analyses. The phosphorylation of threonine 219 and serine 251 of Ndel1 increased in the DBZ KO. B, Radial migration was impaired by the ectopic expression of a dual-phospho-mimetic form of Ndel1 (T219E:S251E Ndel1). Scale bar, 100 μm. C, The ratio of cells with morphologically abnormal nuclei increased in T219E:S251E Ndel1-overexpressing cells. T219E:S251E Ndel1- or T219A:S251A Ndel1-expressing vectors were transfected into HeLa cells. Forty-eight hours later, the cells were immunostained with anti-tyrosinated tubulin antibody and observed with confocal microscopy. Scale bar, 5 μm. D, The ratio of cells with morphologically abnormal nucleus is shown. Values represent the mean ± SEM; *p < 0.05.

Figure 6.

Cdk5 and Aurora A are involved in Ndel1 phosphorylation. A, B, Examination of the effect of GST-DBZ on phosphorylation of GST-Ndel1 by GST-Cdk5/p35 (A) and GST-Aurora A (B). GST-Ndel1 was incubated with GST-DBZ and GST-Cdk5/p35 (A) or GST-Aurora A (B). Phosphorylation of GST-Ndel1 decreased in the presence of DBZ (top) in both cases. CBB staining (bottom). C, Coimmunoprecipitation assay of DBZ and Ndel1. HEK293T cells were transfected with GFP-DBZ in combination with full-length Ndel1; myc-Ndel1-FL-FLAG (Ndel1-FL), the N-terminus of Ndel1; myc-Ndel1-Nt (Ndel1-Nt, residues 1–205), or the C-terminus of Ndel1; Ndel1-Ct-FLAG (Ndel1-Ct, residues 206–345). Immunoprecipitation was performed with the anti-GFP antibody, and Western blotting was performed with anti-myc and anti-FLAG antibodies.

Figure 7.

DISC1 rescues impaired migration and impaired nucleus-centrosome coupling due to DBZ insufficiency. A, Mouse brain lysates at E14.5, E17.5, P0, or P5 were immunoprecipitated with the anti-DBZ antibody. Immunoprecipitates were subjected to Western blotting analyses with the anti-DISC1 antibody (left). The anti-DBZ antibody was examined with mouse brain lysate at P5 using normal IgG (right). B, C, Using in utero electroporation gene transfer, the indicated vectors, including an EGFP-expressing vector, were transfected into the lateral ventricle of mouse embryos at E14.5, and the migration of transfected cells was observed at E18.5. The neocortex was evenly subdivided into five bins. The incidences (%) of GFP-labeled cells (cell bodies) in each bin were counted. The arrest of migration in DBZ-knockdown neurons (C, left); the same image as Fig. 3C,D) was rescued by the coexpression of human DISC1 (hDISC1). Scale bar, 100 μm. D, hDISC1 rescue of impaired migration due to the expression of a dual phosphorylation mimetic form of Ndel1 (T219E:S251E). Expression vectors for T219E:S251E Ndel1 and hDISC1 were transfected together with an EGFP-expressing vector into the ventricle at E14.5, and the transfected cortices were observed at E18.5. Comparisons of the distribution of EGFP-labeled cells between the T219E:S251E Ndel1-transfected and hDISC1/T219E:S251E Ndel1-transfected cortices. hDISC1 coexpression rescued the impaired migration caused by T219E:S251E overexpression. E, The impaired nucleus-centrosome coupling due to DBZ insufficiency was rescued by hDISC1 coexpression. Additionally, we measured the nucleus-centrosome distance of neurons with exogenous hDISC1 alone. There were no significant changes in the nucleus-centrosome distance in this case. Scale bar, 5 μm. The values represent the mean ± SEM; *p < 0.05.

Figure 8.

Mouse DBZ (mDBZ) overexpression did not rescue the migration arrest due to DISC1 knockdown. A, B, The neocortex was subdivided evenly into five bins. The occurrences of GFP-labeled cells (cell bodies) in each bin were counted. Scale bar, 100 μm. The values represent the mean ± SEM.

Figure 9.

Depleting DBZ or expressing a phosphorylation mimetic form of Ndel1 results in a slower extension of EB3-EGFP, and coexpressing DISC1 rescues this phenotype. A, Using in utero electroporation gene transfer, the indicated vectors were transfected into the lateral ventricles of mouse embryos at E14.5. At E16.5, EB3-EGFP-expressing neurons were collected from the cortices and dissociated for primary culture. Forty-eight hours later, time-lapse images were obtained. Scale bar, 20 μm. EB3 binds to microtubule tips, and EB3 movements between the arrow and the arrowhead are shown in the kymographs. The results demonstrated that the EB3-EGFP velocity significantly decreased in DBZ-knockdown neurons. B, T219E:S251E Ndel1-expressing vectors or T219A:S251A Ndel1-expressing vectors were transfected together with the EB3-EGFP-expressing vector. The EB3 velocity in neurites significantly decreased in the T219E:S251E Ndel1-expressing neurons. C, D, The EB3 velocity in the neurites decreased in DBZ^−/− neurons. E, F, Human DISC1 (hDISC1)-HA-expressing vectors were cotransfected into the lateral ventricle of mouse embryos at E14.5. Unlike DBZ^−/− neurons without DISC1, no significant alteration in terms of EB3 velocity was observed among DISC1-overexpressing DBZ^+/+ neurons, DISC1-overexpressing DBZ^−/− neurons and neurons from WT. The values represent the mean ± SEM; *p < 0.05.

Figure 10.

Deleting DBZ or overexpressing a phosphorylation mimetic form of Ndel1 impairs the anterograde transport of DISC1 and Lis1 to the proximal ends. A, FRAP analysis of DBZ^+/+ and DBZ^−/− cortical neurons transfected with the mDISC1-EGFP expression vector. The indicated areas (white circles) of a growth cone were bleached, and fluorescence recovery was monitored every 1 s for 1 min. B, Quantitative fluorescence recovery data were calculated from the fluorescence intensities of the bleached areas of the DBZ^+/+ and DBZ^−/− cortical neurons over a period of 1 min following photobleaching. The data were normalized such that the fluorescence intensity of the prebleach sample was set to one, and the initial postbleach sample was set to zero. C, The half-time of recovery was calculated from each recovery curve. Scale bar, 5 μm. D, DISC1-HA-expressing vectors were transfected into DBZ^+/+ neurons and DBZ^−/− neurons. Neurons were stained with the anti-HA antibody. Magnified images of the white boxes are shown in the bottom. The occurrences of HA signals within 100 μm of the tip of axon were counted. The number of HA signals significantly decreased in the DBZ^−/− neurons. Scale bar, 100 μm. E, Cortical neurons cultured from DBZ^+/+ or DBZ^−/− mouse embryos (E16.5) were stained with antibodies against Lis1 (green) and βIII-tubulin (Tuj1; red) at DIV3. Magnified images of the white boxes are shown on the right. F, Relative immunofluorescence intensities of Lis1 over Tuj1 within 50 μm of the tip of the growth cone were plotted. The fluorescence intensity of Lis1 decreased in the proximal side of the growth cones of the DBZ^−/− neurons. G, The control empty vector or T219E:S251E Ndel1 expression vector, together with the tdTomato expression vector, were transfected at E14.5. At E16.5, tdTomato-positive neurons were collected from the cortices and dissociated for primary culture. Cortical neurons were stained with antibodies against Lis1 (green) and βIII-tubulin (Tuj1; blue) at DIV3. Magnified images of the white boxes are shown in the right. H, Relative immunofluorescence intensities of Lis1 over Tuj1 within 50 μm of the tip of the growth cone were plotted. The fluorescence intensity of Lis1 decreased in the proximal side of the growth cones in T219E:S251E Ndel1-expressing neurons. Scale bars, 50 μm (left) and 10 μm (right). I, The DBZ RNAi vector and the EGFP expression vector were transfected into the lateral ventricle at E14.5, and cortical neurons were collected and dissociated for primary culture at E16.5. Ninety-six hours after the beginning of primary culturing, the lengths of the primary neurites were measured. A GSK3β inhibitor rescued the poor neurite extension caused by acute DBZ insufficiency. The values represent the mean ± SEM; *p < 0.05.

Figure 11.

Cell biological aspects of DBZ activities. DBZ^+/+ (top) and DBZ^−/− (bottom). In the absence of DBZ, Ndel1 was disproportionately phosphorylated in a neuron, leading to poor transportation of DISC1 and Lis1 to the neurite ends. DISC1 suppresses GSK3β activity, which inhibits neurite extension. If insufficient amounts of DISC1 are transported, then neurites do not fully elongated. Anterogradely transported Lis1 is then retrogradely transported and plays a critical role in neuronal migration in an amount-dependent manner.