Fig 1.
DIP2A is localized to dendritic spines in excitatory neurons.
(A) Western blotting showing the level of DIP2A in cerebral cortex constantly increased during postnatal development. (B) Level of DIP2A in cultured neurons. (C) Western blot showing DIP2A in the cortical PSD composition (lane 4). Total, total homogenate; P2, crude synaptosomes; SV, crude synaptic vesicle fraction. (D) Representative image showing the distribution of MAP2 (blue), DIP2A (green), and PSD95 (red) in cultured neurons. Dashed white boxes represent the enlarged view. (E) Dip2a β-galactosidase (LacZ) reporter mice (Dip2alacZ/+) were used to display endogenous DIP2A-expressed cell types. CaMKII and GABA were stained in cortical sections of Dip2alacZ/+ mice to label excitatory and inhibitory neurons with antibodies, respectively. Arrowhead, GABA immunoreactive neurons; arrow, GABA and LacZ double staining neurons. (F) Relative quantification of LacZ-, CaMKII-, and GABA-positive cells. The underlying data for this figure can be found in S1 Data. CaMKII, Ca2+/calmodulin-dependent protein kinase II; DIP2A, disconnected-interacting protein homolog 2 A; DIV, day in vitro; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; LacZ, Dip2a β-galactosidase; MAP2, microtubule-associated protein 2; P, postnatal day; PSD, postsynaptic density; P2, crude synaptosomes; SV, crude synaptic vesicle fraction.
Fig 2.
Dip2a KO mice display impairment in synaptic morphology and function.
(A) Representative images of THY1-GFP–labeled dendrites from EPN and IPN. (B) Increased dendritic spines density of basal dendrite in Dip2a KO mice (P56, males; 4 mice per genotype; EPN, WT = 59 neurons, KO = 75 neurons, t132 = 3.831, ***P = 0.0002; IPN, WT = 71 neurons, KO = 64 neurons, t133 = 5.819, ***P < 0.0001). (C) Reduced radio of mushroom-like spines (EPN, stubby, t132 = 2.004, *P = 0.0471; mushroom-like, t132 = 2.134, *P = 0.0347; IPN, stubby, t133 = 3.1443, **P = 0.0020; mushroom-like, t133 = 2.815, **P = 0.0056). (D) Illustration of the measurement of dendritic spine size. (E-H) Quantitative assessment of spine width or length (EPN, WT = 203 spines, KO = 183 spines; E, t383 = 5.991, ***P < 0.0001; F, t381 = 3.632, ***P = 0.0003; IPN, WT = 158 spines, KO = 201 spines; G, t357 = 3.864, ***P = 0.0001; H, t353 = 2.549, *P = 0.0112). (I) Representative whole-cell voltage clamp recordings of spontaneous excitatory postsynaptic currents (sEPSCs) (n = 7 mice per genotype). (J) Reduced amplitude of sEPSCs in KO mice (t12 = 2.751, *P = 0.0176). (K) No significant difference was observed in the interventional interval (t12 = 0.2820, n.s. P = 0.7828). (L) Examples of 3D reconstruction depicting postsynaptic structure and electron micrographs. (M) Western blot showing the PSD fraction (4 mice per genotype in each experiment; data from 4 independent experiments). Protein levels were normalized to β-actin. The ratio in WT mice was set to 100% (NR1, t18 = 2.274, *P = 0.0355; NR2A, t10 = 4.156, **P = 0.0020; NR2B, t10 = 2.248, *P = 0.0484; GluR1, t18 = 1.082, *P = 0.0447; GluR2, t18 = 0.8829, n.s. P = 0.3889; PSD95, t10 = 0.0759, n.s. P = 0.9410). Data are represented as mean ± SEM and assessed with two-tailed unpaired t test. The underlying data for this figure can be found in S1 Data. Dip2a, disconnected-interacting protein homolog 2 A; EPN, external layer of pyramidal neurons; GFP, green fluorescence protein; GluR1, glutamate ionotropic receptor AMPA type subunit 1; IPN, internal layer of pyramidal neurons; KO, knockout; NMDAR, N-methyl-D-aspartate receptor; NR1, NMDAR1; NR2A, NMDAR2; NR2B, NMDAR2B; n.s., no significance; PSD, postsynaptic density; THY1, thymocyte antigen 1; WT, wild-type; 3D, three-dimensional.
Fig 3.
DIP2A interacts with cortactin and modulates cortactin acetylation (ac-cortactin).
(A) Interaction between endogenous DIP2A and cortactin from murine brain lysates. (B and C) Interaction between endogenous DIP2A, cortactin, and SHANK3 from murine brain lysates. (D) Representative image showing colocalization of DIP2A (green) and cortactin or SHANK3 (red) in cultured neurons. (E) Western blot showing DIP2A and cortactin, and SHANK3 in the cortical PSD composition (lane 4). Total, total homogenate; P2, crude synaptosomes; SV, crude synaptic vesicle fraction. (F) Western blot and bar graph showing Dip2a deletion reduced ac-cortactin without altering levels of total cortactin and phosphorylated cortactin (p-cortactin) (P56, males; n = 4 mice per genotype; data from 3 independent experiments). Levels of ac-cortactin and tubulin (ac-tub) were normalized to total cortactin and α-tubulin, respectively. The ratio in WT mice was set to 100% (ac-cortactin, t10 = 2.320, *P = 0.0428; p-cortactin, t10 = 0.8637, n.s. P = 0.4080; cortactin, t10 = 1.712, n.s. P = 0.1177; ac-tub, t10 = 0.1943, n.s. P = 0.8498). (G) Dip2a deletion reduced ac-cortactin without affecting total or p-cortactin levels in cultured cortical neurons (DIV15) (data from 3 independent experiments, approximately 5 × 106 cells per experiment; ac-cortactin, t10 = 3.258, **P = 0.0086; p-cortactin, t10 = 0.1551, n.s. P = 0.8798; cortactin, t10 = 0.4124, n.s. P = 0.6888; ac-tub, t10 = 0.1507, n.s. P = 0.8832). (H) Western blot showing overexpressed DIP2A in cultured cells with elevated ac-cortactin level. (I and J) Normalized grayscale value in (H) (data from 3 independent experiments and presented as mean ± SEM; one-way ANOVA; I, F3,20 = 19.13, post hoc LSD, *P = 0.0199, **P = 0.0020 and 0.0016, respectively; J, F3,20 = 17.76, post hoc LSD, n.s. P = 0.3227, ***P < 0.0001, *P = 0.0379). The underlying data for this figure can be found in S1 Data. ac-cortactin, cortactin acetylation; DIP2A, disconnected-interacting protein homolog 2 A; DIV, day in vitro; FLAG, FLAG tag with the sequence DYKDDDDK; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; HA, HA-tag (YPYDVPDYA); HDAC6, histone deacetylase 6; IgG, immunoglobulin G; IP, immunoprecipitation; KO, knockout; LSD, least significant difference; n.s., no significance; p-cortactin, phosphorylated cortactin; PSD, postsynaptic density; P2, crude synaptosomes; SHANK3, SH3 and multiple ankyrin repeat domains 3; SV, crude synaptic vesicle fraction; TSA, trichostatin A; tub, tubulin; WT, wild-type.
Fig 4.
The DIP2A-cortactin interaction is critical for mediating ac-cortactin.
(A) The SH3 domain of cortactin (aa 492–546) was required for binding with DIP2A in the GST pull-down assay. (B) The binding assay demonstrated that the DIP2A261–320 region with PXXP motifs was large enough to bind cortactin. (C) Diagram of GST-tagged DIP2A1–320 with site mutation strategies. Residues 271–310 with PXXP motifs are aligned below. (D) DIP2A-cortactin binding was minimally detectable when both PXXP motifs were simultaneously mutated (double). (E and F) Ac-cortactin levels were decreased in a concentration-dependent pattern (ac-cortactin, Pearson correlation = −0.656, *P = 0.0205; ac-tub, Pearson correlation = 0.279, n.s. P = 0.3798). (G and H) Ac-cortactin levels were decreased in time-dependent patterns (ac-cortactin, Pearson correlation = −0.814, **P = 0.0013; ac-tub, Pearson correlation = 0.181, n.s. P = 0.5729). The underlying data for this figure can be found in S1 Data. aa, amino acid; ac-cortactin, cortactin acetylation; DIP2A, disconnected-interacting protein homolog 2 A; FLAG, FLAG tag with the sequence DYKDDDDK; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GST, glutathione S-transferase tag; HA, human influenza hemagglutinin tag (YPYDVPDYA-tag); n.s., no significance; PXXP, proline-rich motif, where X is any residue; SH3, src homology 3; tub, tubulin.
Fig 5.
DIP2A-mediated ac-cortactin is dependent on Ac-CoA concentration.
(A) Ac-CoA concentration in the cerebral cortex decreased in Dip2a KO mice in comparison with WT littermates (P56, males; n = 9 mice per genotype; t16 = 2.291, *P = 0.0359). (B and C) Overexpressed DIP2A in HEK293 cells enhanced ac-cortactin in a dose-dependent manner (data from 4 independent experiments and represented as mean ± SEM). (D and E) Western blot and line graph showing that levels of ac-cortactin increased as the concentration of Ac-CoA increased in in vitro acetylation assays. The underlying data for this figure can be found in S1 Data. Ac-CoA, acetylated coenzyme A; ac-cortactin, cortactin acetylation; DIP2A, disconnected-interacting protein homolog 2 A; KO, knockout; WT, wild-type.
Fig 6.
Acetylation mimetic cortactin rescues synaptic defects.
(A) Representative image of dendritic segments of cultured neurons infected at DIV15 with Lv-eGFP (control) or Lv-eGFP-cortactin 8KQ, and imaged at DIV21. The right insets showing an example of the quantitative analysis of PSD95 distribution along the 2-μm vertical white line, which proceeds from the base of the dendritic shaft (set as 0) to the outline of PSD95. (B) Traces illustrating the fluorescence intensity of PSD95 (green) along the scanning path in (A). (C) Quantitative analysis of the distribution of PSD95. (D) Scatterplot graph showing the area ratio of PSD95-positive puncta to the total dendritic area (MAP2). Data from 4 independent experiments and represented as mean ± SEM; WT + control = 64 neurons, KO + control = 46 neurons, KO + 8KQ = 51 neurons; one-way ANOVA, F2, 158 = 16.56; post hoc LSD, ***P < 0.0001, n.s. P = 0.0513, related to WT + control. (E) Representative whole-cell voltage clamp traces of mEPSC from cultured neurons held at −70 mV. (F) Scatterplot graph showing mEPSC amplitude. Lv-eGFP-cortactin 8KQ infection of Dip2a KO neurons at DIV10 rescued the aberrant mEPSC amplitude (n = 16, 14, 18, and 21 neurons, respectively; data from 3 independent experiments; one-way ANOVA, F3, 65 = 34.16; post hoc LSD, ***P < 0.0001, n.s. P = 0.1521). (G) Scatterplot graph showing the equivalent mEPSC frequency (one-way ANOVA, F3, 61 = 1.01, P = 0.13946; post hoc LSD, n.s. P > 0.05). The underlying data for this figure can be found in S1 Data. DIV, day in vitro; KO, knockout; LSD, least significant difference; Lv-eGFP, lentivirus carrying enhanced green fluorescent protein tag; MAP2, microtubule-associated protein 2; mEPSC, miniature excitatory postsynaptic current; n.s., no significance; PSD, postsynaptic density; WT, wild-type; 8KQ, acetylation mimetic cortactin with eight lysine replaced with glutamine.
Fig 7.
Deletion of Dip2a in mice leads to an autism-like phenotype.
(A) Diagram of the evaluation of isolation-induced USVs (WT = 25 mice, KO = 24 mice). (B and C) Reduced call duration of USV (B, t47 = 2.711, **P = 0.0093) and calling time (C, t47 = 2.716, **P = 0.0092) in Dip2a KO pups. (D and E) Dip2a KO mice had increased self-grooming time in the open field (14 mice per genotype; D, t26 = 2.651, *P = 0.0135; E, t26 = 3.104, 2.425, and 2.456, **P = 0.0046, *P = 0.0226 and 0.0210, respectively). (F) Increased self-grooming time in home cage in Dip2a KO mice (18 mice per genotype, t34 = 2.508, *P = 0.0171). (G) In the marble burying test, Dip2a KO mice had increased digging time (WT = 14 mice, KO = 15 mice, t27 = 2.386, *P = 0.0243). (H) Exploration time of mice in chambers of the sociability trial (S1, the chamber for stranger 1; O, chamber with empty cage) (WT = 17 mice, t32 = 2.847, **P = 0.0076 for S1 versus O; KO = 16 mice, t30 = 2.982, **P = 0.0056). (I) Exploration time in the social novelty trial (S2, the chamber for stranger 2) (WT, t32 = 4.234, ***P = 0.0002 for S1 versus S2; KO, t30 = 0.8245, n.s. P = 0.4161). (J) One microliter AAV9 (titer > 1012 GC/mL) carrying CaMKII-GFP control or CaMKII-GFP-8KQ was injected into the cerebral cortex of 4-week-old mice and observed after 4 more weeks. Sites of injection: (1) anteroposterior (AP), 2.0 mm; mediolateral (ML), 0.5 mm; dorsoventral (DV), 1.3 mm; (2) AP, −1.0 mm; ML, 2.0 mm; DV, 0.8 mm. (K) The expression of cortactin-8KQ in cerebral cortex at 9 weeks old. (L) Cortactin 8KQ rescued the excessive self-grooming in Dip2a KO mice (WT + control = 11, KO + control = 9, KO + 8KQ = 8 mice; one-way ANOVA, F2, 25 = 5.283, P = 0.013; post hoc LSD, **P = 0.0038, n.s. P = 0.4001, related to WT + control). (M) A model illustrating the function of DIP2A in spine morphology. DIP2A interacted with cortactin and maintained ac-cortactin, which is essential for the mushroom-like morphology and synaptic transmission of dendritic spines. The underlying data for this figure can be found in S1 Data. AAV9, adeno-associated virus type 9; ac-cortactin, cortactin acetylation; AP, anteroposterior; CaMKII, Ca2+/calmodulin-dependent protein kinase II; Dip2a, disconnected-interacting protein homolog 2 A; DV, dorsoventral; GFP, green fluorescence protein; KO, knockout; LSD, least significant difference; ML, mediolateral; n.s., no significance; O, chamber with empty cage; S1, the chamber for stranger 1; USV, ultrasonic vocalization; WT, wild-type; 8KQ, acetylation mimetic cortactin with eight lysine replaced with glutamine.