Figure 1.
Dissociation of the phosphorylated proteins from CCVs.
Rat brain CCVs (20 µg) were incubated with 0.1 mM [γ-32P]-ATP with or without GST-Dyrk1A497 (3.7 µg) as described in MATERIALS AND METHODS. After mixing with EDTA and phosphatase inhibitors, the samples were transferred on ice and ultracentrifuged at 70,000 rpm for 15 min using a Beckman TLA-100 rotor. The supernatants (S) were collected, and the precipitates (P) were suspended in the original volume of the kinase buffer containing protease- and phosphatase- inhibitors. Each sample was subjected to SDS-PAGE followed by Coomassie Blue staining (CB) and autoradiography (32P). Half of each SDS sample was applied per lane. The numbers in the right panel refer to the phosphorylated protein bands 1–5. Dyrk1A497 was used in most of our experiments, because this truncated form 1) is highly purified in contrast to the full-length protein, which always contains kinase bands degraded to various extents [20], and 2) exhibits the similar kinase activity as the full-length protein [21]. *, denotes clathrin heavy chain (CHC). (n = 3; the assay was repeated three times by using different CCV preparations giving similar results.)
Figure 2.
Identification of the phosphorylated protein bands 1 and 2 as MAP1A and MAP2.
(A) MAP1A and MAP2 in the phosphorylated CCVs and MTs. CCVs (20 µg) and purified MTs (5 µg) were incubated with [γ-32P]-ATP with (+) or without (−) GST-Dyrk1A497 as described in Fig. 1, followed by SDS-PAGE without ultracentrifugation. Approximately 9 µg and 2 µg of CCVs and MTs, respectively, were applied per lanes. After transferring proteins, each lane of the PVDF membranes was cut into two strips for immunostaining either with anti-MAP1A (a) or anti-MAP2 (b) antibody, or for Coomassie Blue staining (c). The strips were reassembled (WB/CB) and subjected to autoradiography (32P). (n = 2). (B) Coomassie Blue-staining of the CCV and MT preparations. Ten and five µg of CCVs and MTs, respectively, were applied per lane. (C) Immunoprecipitation of MAP1A and MAP2 from the phosphorylated MTs. MTs (200 µg) were phosphorylated for 1 hr with GST-Dyrk1A497 (18 µg) and 0.2 mM [γ-32P]-ATP in a final volume of 250 µl. After the reaction, the soluble fraction was subjected to immunoprecipitation (IP) by using anti-MAP2 (M2) or anti-MAP1A (M1A) antibody as described in MATERIALS AND METHODS. A negative control for the immunoprecipitation (−) was obtained without primary antibody. The immunoprecipitates were applied to SDS-PAGE followed by Coomassie Blue staining (CB) and autoradiography (32P). (n = 1; various preliminary performances carried out to lead the final assay conditions are not included). Scanning of the MAP1A and MAP2 bands from the original material used for immunoprecipitation (Ori) gave the arbitrary units for these proteins as 3306 and 6323, respectively, whereas those for the radioactivity were 6056 and 12582, respectively. (D) Immunoprecipitation of MAP1A from the extract of the phosphorylated CCVs. CCVs (60 µg) were incubated with GST-Dyrk1A497 (7 µg) and 0.2 mM [γ-32P]-ATP for 1 hr in a final volume of 120 µl. The phosphorylated CCVs were extracted with 0.5 M Tris-HCl, diluted the Tris-HCl concentration, and used for immunoprecipitation with anti-MAP1A antibody (IP). The immunoprecipitates were subjected to blotting (WB) with anti-MAP1A antibody followed by autoradiography (32P). (n = 2). St, pre-stained standard proteins; M1A, MAP1A; M2, MAP2.
Table 1.
Summary of the MS analysis for identification of the rat proteins in the Coomassie Blue-stained bands.
Figure 3.
Identification of the phosphorylated protein bands 3 and 4.
(A) Migration patterns of AP180 and band 3 in SDS-PAGE. After [32P]-phosphorylation, CCVs were subjected to SDS-PAGE followed by blotting with anti-AP180 antibody (WB), Coomassie-Blue staining (CB), and autoradiography (32P) as in Fig. 2A. (n = 2). Approximately 1.3 µg proteins were applied per lane, and a single lane of the PVDF membrane was cut into two strips. Arrowheads, AP180; asterisk, CHC. Two lower bands in the autoradiogram are band 4 and the autophosphorylated GST-Dyrk1A497. (B) Immunoprecipitation of AP180 from the extract of the phosphorylated CCVs. CCVs were incubated with GST-Dyrk1A497 and [γ-32P]-ATP, extracted with Tris-HCl, and used for immunoprecipitation (IP) using anti-AP180 antibody (anti-SNAP91), as described in Fig. 2D. The resultant immunoprecipitates were subjected to blotting using anti-AP180 antibody (WB:AP180) followed by autoradiography (32P). (n = 2). (−) and (+), immunoprecipitation without and with anti-AP180 antibody, respectively; Ori, the CCV extract used for immunoprecipitation. (C) Immunoprecipitation of β-adaptin. The CCV extract prepared as in (B) was subjected to immunoprecipitation without (−) or with (+) anti-β-adaptin antibody (IP:β). (n = 2). One lane of the starting extract (Ori) was cut into two strips; one was probed with anti-β-adaptin (strip 1) together with the immunoprecipitates (WB:β), and the other (strip 2) and lane 3 were incubated with anti-α-adaptin antibody (α). All strips were reassembled for autoradiography (32P). The lower band in the WB panel is IgG heavy chain. (D) Immunoprecipitation of β-adaptin but not α-adaptin by anti-β-adaptin antibody. The membranes containing lane (+) and strip 1 from (C) were re-blotted with anti-α-adaptin (WB:β+α) and re-assembled with strip 2 from (C). The ratios in relative intensities of α- (APα) to β (APβ)-adaptins in the IP:β lane and strip 1 shown in here were 1∶40.5 and 1∶2.7, respectively. (E) Immunoprecipitation of α-adaptin. After the first immunoprecipitation using anti-β-adaptin antibody (C), the unbound fraction was incubated with (+) or without (−) anti-α-adaptin antibody (IP:α) for the second immunoprecipitation. The precipitates were subjected to immunoblotting using anti-α-adaptin antibody (WB:α) followed by autoradiography (32P). (n = 1). (F) Co-precipitation of α- and β-adaptins from the extracts of unphosphorylated CCVs. CCVs in two tubes were diluted in kinase buffer, mixed with either H2O (Mg2+) or 10 mM EDTA, and extracted with 0.5 M Tris-HCl for immunoprecipitation with anti-β-adaptin antibody as in (C). (n = 3). The immunoprecipitates (IP) and the original extracts (Ori) were blotted using antibodies against α- or β-adaptin (WB: α, β).
Figure 4.
Release of MAP1A and MAP2 into the soluble fraction after phosphorylation.
(A) Release of MAP1A. CCVs (20 µg) were phosphorylated and ultracentrifuged as described in Fig. 1 except for using 2 mM cold ATP and with or without GST-Dyrk1A497. The resultant supernatants were blotted with anti-MAP1A antibody (M1A). (n = 3). (B) Phosphorylation-dependent dissociation of MAP1A and MAP2. CCVs were incubated as in (A) in duplicate tubes containing ATP and either wild type (wt) or double mutant (DF) GST-Dyrk1A497. After adding EDTA, one set of tubes was kept on ice, whereas the other set was ultracentrifuged to collect the soluble fractions. Aliquots of whole mixtures (W) and the supernatants (S) were subjected to immunoblotting. Both MAP1A (M1A) and MAP2 (M2) were identified in the same lane by blotting the PVDF membrane first with anti-MAP2 then with anti-MAP1A antibodies. (n = 1, various preliminary performances carried out to lead the final assay conditions are not included).
Figure 5.
Effect of Dyrk1A on dissociation of adaptor proteins from CCVs.
(A) Effect of phosphorylation. CCVs (3.5 µg) were incubated with ±1 mM cold ATP and ± Dyrk1A (0.5 µg), as indicated in the panel, and ultracentrifuged as described in Fig. 1. The resultant supernatants were immunoblotted (WB) using the antibodies indicated in the figure. (n = 2). (B) Release of α- and β-adaptins by Dyrk1A without phosphorylation. CCVs (3.75 µg) were incubated with various amounts of GST-Dyrk1A497, but without ATP, either in kinase buffer (pH 7.4) or MES buffer (pH 6.5) containing 0.1 M NaCl, 5 mM MgCl2. Appropriate amounts of GST were added into each tube to compensate for the different kinase amounts. After chilling on ice, the reaction mixtures were ultracentrifuged as described in Fig. 1. The supernatants were blotted by using anti-α- or β-adaptin antibody. (n = 2). (C) Release of multiple CCV proteins by Dyrk1A without phosphorylation. The same samples from (B) at pH 7.4 were analyzed by using various antibodies as indicated. Lanes (−) and 1–5 represent the kinase concentrations at 0, 0.1, 0.25, 0.5, 1.0, and 2 µg/assay. One µg Dyrk1A/assay is equal to 4.2×10−7 M by assuming the enzyme purity as 100%.
Figure 6.
Effect of phosphorylation on dissociation of adaptor proteins in the presence of Mg2+.
(A) Effect of Mg2+ on AP180 dissociation. CCVs (3 µg) were incubated with Dyrk1A and cold ATP as described in Fig. 5A, but with modified conditions including 75 mM HEPES (pH 7.0), 1 mg/ml BSA, and a mixture of phosphatase-inhibitors. After the reaction, the tubes were mixed with either 10 mM EDTA (EDTA) or with H2O (Mg2+) and ultracentrifuged. The resultant supernatant fractions were blotted with anti-AP180 antibody. (n = 3). Duplicated samples were shown. (B) Effect of phosphorylation on adaptor protein dissociation. CCVs were incubated as in (A) with ± Dyrk1A and ± ATP as indicated in the panel. The reaction mixtures were ultracentrifuged without adding EDTA. The same volumes of the supernatants and whole reaction mixtures (W) were immunoblotted using the indicated antibodies. Duplicated samples were shown. (n = 1, various preliminary performances carried out to lead the final assay conditions are not included).
Figure 7.
Phosphorylation of the membrane-unbound adaptor proteins.
(A) Immunoprecipitation of AP180 from the phosphorylated PNP extract. The Tris-HCl extract from PNP (diluted, 300 µl) prepared as in MATERIALS AND METHODS was phosphorylated for 1 hr in a mixture containing 0.2 mM [γ-32P]-ATP, 9 µg Dyrk1A, 5 mM MgCl2, 0.12 M NaCl, 0.1 mM EGTA, and 0.2 mM DTT. After the reaction, the mixture was subjected to immunoprecipitation by using anti-AP180 antibody as in Fig. 3B. (n = 2). The immunoprecipitate (IP) and an aliquot of the original phosphorylation mixture (Ori) were subjected to immunoblotting (WB) and autoradiography (32P). (B) Phosphorylation of the immunoprecipitated AP180. The PNP extract was first subjected to immunoprecipitation with (+) and without (−) anti-AP180 antibody. The pellets of Protein A/G resins were washed three times with PBS-T and once with kinase buffer, and suspended in a small volume of kinase buffer (20 µl). Phosphorylation reaction was carried out in the presence of 0.2 mM [γ-32P]-ATP and Dyrk1A (2 µg) in a final liquid volume of 25 µl. Aliquots of the reaction mixtures were subjected to SDS-PAGE followed by Coomassie-Blue staining (CB) and autoradiography (32P). Arrow, AP180. (n = 1). (C, D) Immunoprecipitation of α- and β-adaptins after phosphorylation reaction. The PNP extract (C) and cytosol (D) were incubated with Dyrk1A and [32P]-ATP and subjected to immunoprecipitation with (IP) and without (−) corresponding antibodies. The immunoprecipitates and the aliquots of original reaction mixtures (Ori) were subjected to immunoblotting and autoradiography. (n = 3). (E) Incubation of the immunoprecipitated adaptins with Dyrk1A. Cytosol and the PNP extract were first subjected to immunoprecipitation without (−) and with (+) anti-α-adaptin antibody. The resultant precipitates were incubated with Dyrk1A and [γ-32P]-ATP followed by immunoblotting (WB) and autoradiography (32P) as described in (B). (n = 1, various preliminary performances carried out to lead the final assay conditions are not included). α, α-adaptin; β, β-adaptin; PNP-Ext, PNP extract; arrowheads, α-adaptin; arrow, β-adaptin; *, autophosphorylated Dyrk1A.
Figure 8.
Ratios of adaptin subunits recovered in cytosol to the precipitated fractions from cultured cells.
Cytosol (Sup) and precipitated (Ppt) fractions from CHO and PC12 cells were prepared as described in MATERIALS AND METHODS. SDS-PAGE was carried out by applying equal volumes of the soluble and insoluble fractions in each lane side-by-side. Immunoblotting was performed by using corresponding antibodies derived from mouse and rabbit. The first blots with mouse and rabbit antibodies were stripped (stripping buffer, PIERCE) and re-blotted with rabbit and mouse antibodies, respectively, for detecting other adaptin subunits in the same membranes. The antibodies used were mouse monoclonal antibodies against anti-α and γ-adaptins, rabbit monoclonal anti-β-adaptin antibody, and rabbit polyclonal anti-μ adaptin antibody. Each adaptin band in the Sup and Ppt was scanned, and the Sup to Ppt ratio was calculated. Four independent samples per each cell type were shown.
Figure 9.
Relative kinase concentrations in the CCV fraction.
Rat brain fractions S1, P2, P3 were prepared as described in MATERIALS AND METHODS. These fractions together with CCVs were subjected to SDS-PAGE followed by immunoblotting using 8D9 (Dyrk1A). Fifteen µg of protein were applied per each lane. After stripping, the same membrane was re-blotted with anti-CHC antibody. (n = 1, various preliminary performances carried out to lead the final assay conditions are not included).
Figure 10.
Co-localization of endogenous Dyrk1A and clathrin in the adult mouse brains.
Forty micrometer sections of the mouse brain tissues were incubated with polyclonal anti-Dyrk1A (red, a, d, and g) and monoclonal anti-CHC (green, b, e, and h) antibodies as described in MATERIALS AND METHODS. Presented are Z-stack images collected from the subiculum (a–c), the medio-posterior thalamic nucleus (d–f), and individual confocal images from the cerebellar Purkinje cell layer (g–i). Structures showing co-localization of Dyrk1A and CHC are visible on merged images (yellow in c, f, i). The dotted squares shown in a–i are enlarged in the upper right corners. Scale bar = 10 µm. Panels (j–m) show three consecutive scanning images from subiculum. Pearson's correlation coefficients for three consecutive scanning images from subiculum and the medio-posterior thalamic nucleus were calculated as 0.439 and 0.499, respectively, by using NIH Image J with the JACoP plug-in [65].
Figure 11.
Localization of Dyrk1A and clathrin in the primary cultured neurons.
E-18 rat hippocampal neurons were differentiated, fixed with 2% formaldehyde, and immunostained with monoclonal anti-CHC followed by Alexa Fluor 488 conjugated goat anti mouse IgG. After completion of CHC staining, the cells were incubated with monoclonal anti-Dyrk1A antibody (7F3), directly conjugated with Alexa Fluor 568 as described in MATERIALS AND METHODS. (a), single confocal images; (b and c), Z-stack images. (n = 2). Pearson's correlation coefficient for single scanning images was 0.645 as calculated as in Fig. 10.
Figure 12.
Proposed Dyrk1A functions in regulating synaptic vesicle endocytosis.
Dyrk1A phosphorylated AP180 in cytosol. The phosphorylated AP180 may decrease its binding affinity to the AP2 complex; we hypothesize here that the decrease in such binding affinity would reduce recruitment of clathrin at the nucleation and invagination sites. Phosphorylation of dynamin 1, amphiphysin 1, and synaptojanin 1 at their PRD reduces the interaction between the PRDs and the SH3 domains of amphiphysin and endophilin. This likely slows down the invagination and fission steps of synaptic vesicle formation. Once endocytosed, the vesicle-associated proteins are quickly removed from the membranes (uncoating). The Dyrk1A-mediated phosphorylation releases first AP180 and β-adaptin from the vesicle membranes, while both α- and μ-adaptin subunits remain bound with the membranes. Additional factor(s) are required to release the membrane-bound subunits. Clathrin release from the vesicles is independent from Dyrk1A. We speculate that the adaptin subunits released in cytoplasm may reassemble into the AP2 complex after dephosphorylation.