Huntingtin-Interacting Protein 1-Related Protein Plays a Critical Role in Dendritic Development and Excitatory Synapse Formation in Hippocampal Neurons

Abstract

Huntingtin-interacting protein 1-related (HIP1R) protein is considered to be an endocytic adaptor protein like the other two members of the Sla2 family, Sla2p and HIP1. They all contain homology domains responsible for the binding of clathrin, inositol lipids and F-actin. Previous studies have revealed that HIP1R is highly expressed in different regions of the mouse brain and localizes at synaptic structures. However, the function of HIP1R in the nervous system remains unknown. In this study, we investigated HIP1R function in cultured rat hippocampal neurons using an shRNA knockdown approach. We found that, after HIP1R knockdown, the dynamics and density of dendritic filopodia, and dendritic branching and complexity were significantly reduced in developing neurons, as well as the densities of dendritic spines and PSD95 clusters in mature neurons. Moreover, HIP1R deficiency led to significantly reduced expression of the ionotropic glutamate receptor GluA1, GluN2A and GluN2B subunits, but not the GABA_A receptor α1 subunit. Similarly, HIP1R knockdown reduced the amplitude and frequency of the miniature excitatory postsynaptic current, but not of the miniature inhibitory postsynaptic current. In addition, the C-terminal proline-rich region of HIP1R responsible for cortactin binding was found to confer a dominant-negative effect on dendritic branching in cultured developing neurons, implying a critical role of cortactin binding in HIP1R function. Taken together, the results of our study suggest that HIP1R plays important roles in dendritic development and excitatory synapse formation and function.

Keywords: HIP1R; cultured neurons; dendritic development; proline-rich domain; synapse formation.

Figure 1

Huntingtin-interacting protein 1-related (HIP1R) expression in rat brain tissues and cultured hippocampal neurons. (A–C) Western blotting showed that HIP1R was highly expressed in different brain regions of rats (A) expressed with a gradual increase in the hippocampus during postnatal development (B) and located partially in the postsynaptic density (PSD) fraction (C). (D) Immunohistochemical staining of the hippocampus revealed a strong immunoreactivity of HIP1R in the pyramidal layer in the CA1 and CA3 regions, and the dentate granule cell layer. The insets showed that HIP1R was also distributed on the CA3 and CA1 stratum radiatum. Scale bar = 200 μm. (E) Co-immunofluorescence staining of primary cultured hippocampal neurons at days in vitro 21 (DIV21) showed that HIP1R exhibited small punctate structures on the soma and along dendrites marked by anti-MAP2 antibody. Scale bar = 10 μm; and (F,G) HIP1R puncta were partially colocalized with PSD95, but barely with synaptophysin (arrows). Scale bar = 5 μm. Uncropped images of blots are shown in Supplementary Figure S1.1A–C.

Figure 2

Knockdown of HIP1R expression suppresses dendrite growth and spine formation. (A) Cultured hippocampal neurons were infected by HIP1R-shRNA or scrambled control lentivirus and harvested for western blotting at DIV13. A representative western blot shows that the expression level of HIP1R was dramatically reduced in the shRNA group (shRNA) compared to the scrambled control (Scrambled), and viral infection itself had no effect on HIP1R expression (Blank). (B) Densitometric analysis of western blotting from 4 independent experiments; Blank, 1.00 ± 0.07; Scrambled, 1.04 ± 0.12; shRNA, 0.17 ± 0.03. (C) Cultured hippocampal neurons were transfected at DIV6 with HIP1R-shRNA or scrambled plasmid vectors. Immunofluorescence staining at DIV13 showed a significant decrease in HIPIR immunoreactivity in HIP1R-KD neurons (shRNA) but not in the control (Scrambled). Arrows indicate transfected neurons marked by GFP expression. Scale bar = 10 μm. (D–G) Images of HIP1R-shRNA transfected neurons at DIV13 immunostained with anti-MAP2 and used for counting dendrites. Scale bar = 10 μm (D). Unpaired two-tailed t-test analysis showed a significant decrease in the total dendritic branch number and total dendritic length in the HIP1R-KD group compared to the scrambled control; 11.69 ± 0.67, n = 45 vs. 41.64 ± 2.15, n = 45 (E) and 388.98 ± 16.63 μm, n = 45 vs. 989.91 ± 28.38 μm, n = 45 (F), respectively. Sholl analysis revealed dendritic complexity dramatically decreased in the HIP1R-KD group compared to the scrambled control (n > 40, ^#P < 0.0001) (G). (H–K) Images of HIP1R-shRNA transfected neurons at DIV21 immunostained with anti-GFP and used for counting dendritic spines. Scale bar = 10 μm (H) The spine density, spine width and length were also significantly reduced in the HIP1R-KD group compared to the control with spine number per μm: 0.12 ± 0.01, n = 48 vs. 0.48 ± 0.01, n = 48 (I); spine width: 0.56 ± 0.02 μm, n = 62 vs. 0.73 ± 0.03 μm, n = 62 (J); and spine length: 0.79 ± 0.03 μm, n = 62 vs. 1.09 ± 0.04 μm, n = 62 (K) respectively. All data are presented as mean ± SEM, ^#P < 0.0001. Uncropped images of blots are shown in Supplementary Figure S1.2A.

Figure 3

Over-expression of HIP1R enhances dendritic growth and spine formation. (A–D) Cultured hippocampal neurons were transfected at DIV6 with GFP-HIP1R or GFP-expressing vectors. Images of the transfected neurons at DIV12 immunostained with anti-MAP2 were used for counting dendrites. Scale bar = 10 μm (A) Sholl analysis indicated an increased complexity of dendrites in HIP1R over-expressed neurons (GFP-HIP1R) compared to the control (GFP-Vector) (B). Quantitative data showed that both total dendrite number and total dendritic length were significantly increased in HIP1R over-expressed neurons compared to the control; 64.70 ± 2.326, n = 43 vs. 52.17 ± 3.66, n = 42 (C) and 1518.42 ± 45.80 μm, n = 43 vs. 1249.72 ± 65.02 μm, n = 42 (D), respectively. (E–H) Images of the transfected neurons at DIV21 visualized by anti-GFP immunostaining were used for analyzing dendritic spines by MetaMorph software. Scale bar = 10 μm (E). Quantitative data revealed that spine density, width and length were all increased in HIP1R over-expressed neurons (GFP-HIP1R) compared to the control (GFP-Vector), with spine number per μm: 0.63 ± 0.02 n = 48 vs. 0.49 ± 0.01 n = 48 (F), spine width: 1.17 ± 0.03 μm, n = 61 vs. 0.74 ± 0.03 μm, n = 61 (G), and spine length: 1.73 ± 0.09 μm, n = 61 vs. 1.10 ± 0.04 μm, n = 61 (H), respectively. All data are presented as mean ± SEM, **P < 0.01, ***P < 0.001, ^#P < 0.0001.

Figure 4

HIP1R knockdown decreases dynamics and density of dendritic filopodia. (A–C) Cultured hippocampal neurons were transfected at DIV5 with HIP1R-shRNA and Scrambled vectors. Confocal fluorescence microscopy images were taken at DIV8 and used for analyzing spine morphology by MetaMorph software. Scale bar = 20 μm (A). Quantitative data showed both the density and length of dendritic filopodia were remarkably reduced in HIP1R-KD neurons compared to the scrambled control, with filopodia number per μm: 0.36 ± 0.03, n = 30 vs. 0.59 ± 0.03, n = 65 (B) and filopodia fiber length: 1.42 ± 0.15 μm, n = 30 vs. 1.94 ± 0.13 μm, n = 65 (C), respectively. (D–F) Time-lapse images of living transfected neurons were also taken for analyzing the dynamic activity of filopodia. The linear trajectory shows the movement pathway of each filopodium tip during 20 min and the size of the sphere represents the volume of the filopodium tip each minute. Scale bar = 10 μm (D). The average speed and track length of filopodia tips were obviously reduced in HIP1R-KD neurons compared to those of the control, with average speed: 1.15 ± 0.03 μm/s, n = 177 vs. 1.66 ± 0.06 μm/s, n = 166 (E) and track length: 30.25 ± 0.90 μm, n = 177 vs. 54.23 ± 1.76 μm, n = 166 (F), respectively. Imaris software was used to track the tips of fibers in each min and analyze speed and track length. All data are represented as mean ± SEM, *P < 0.05, ^#P < 0.0001.

Figure 5

HIP1R knockdown reduces expression of NMDARs, AMPARs and PSD95, but not GABA_A. Cultured hippocampal neurons were infected with HIP1R-shRNA or Scrambled lentivirus at DIV6 and analyzed by western blotting at DIV13. (A) A representative western blot for total GluN2A, GluN2B, GluA1 and GABA_A-α1 expression. (B) Quantification of blots from repeated independent experiments showed that the total expression levels of GluN2A, GluN2B and GluA1 were significantly reduced in the HIP1R-KD group compared to the control with 0.52 ± 0.11, n = 7 vs. 1.02 ± 0.18, n = 7 for GluN2A; 0.46 ± 0.07, n = 6 vs. CTL, 1.00 ± 0.20, n = 6 for GluN2B; and 0.52 ± 0.06, n = 6 vs. 1.00 ± 0.08, n = 6 for GluA1, respectively. (C) A representative western blot for surface GluN2A, GluN2B, GluA1 and GABA_A-α1 expression. (D) The ratio of surface to total expression levels of GluN2A, GluN2B, GluA1 and GABA_A-α1 did not differ between the HIP1R-KD group and the control. (E) Surface expression levels of GluN2A, GluN2B and GluA1 were significantly reduced in the HIP1R-KD group compared to the control, when normalized for GABA_A-α1. There was no significant change in the total and surface expression levels of GABA_A-α1 in the HIP1R-KD group. The ratio of HIP1R-KD vs. the control: 0.47 ± 0.05, n = 4 for GluN2A, 0.66 ± 0.06, n = 3 for GluN2B; 0.36 ± 0.01, n = 5 for GluA1 vs. 1.00 ± 0.05, n = 5 for GABA_A-α1, respectively. (F) Representative images of dendrites immunostained using anti-PSD95 antibody at DIV16 in cultured hippocampal neurons transfected at DIV6 with HIP1R-shRNA and Scrambled vectors, respectively. (G) Quantification showed a significant decrease in the cluster density of PSD95 in the HIP1R-KD group compared to the control with 0.12 ± 0.01, n = 55 vs. 0.48 ± 0.02, n = 55, respectively. (H) Representative western blots for PSD95 of lysates from cultured hippocampal neurons at DIV13 6 days after infection with HIP1R-shRNA or Scrambled lentivirus. (I) Quantification of western blots from three independent experiments showed decreased expression of PSD95 in the HIP1R-KD group compared to the control with 0.68 ± 0.06, n = 3 vs. 1.00 ± 0.07, n = 3, respectively. All data are presented as mean ± SEM. *P < 0.05, **P < 0.01, ^#P < 0.0001. Uncropped images of blots are shown in Supplementary Figure S1.5A,C,H.

Figure 6

Amplitude and frequency of miniature excitatory post-synaptic current (mEPSC), but not miniature inhibitory post-synaptic current (mIPSC), are reduced in HIP1R-knockeddown neurons. Cultured hippocampal neurons were infected by HIP1R-shRNA or Scrambled lentivirus at DIV6, and whole-cell patch-clamp recording was carried out at DIV13. (A) Example traces of mEPSC. (B) Graph of cumulative probability and bar graph of amplitude mEPSC (HIP1R-KD, 17.93 ± 1.13 pA, n = 19 vs. the control, 22.38 ± 1.03 pA, n = 22) and (C) Graph of cumulative probability and bar graph of inter-event interval of mEPSC (HIP1R-KD, 1.95 ± 0.28 s, n = 19 vs. the control, 0.74 ± 0.17 s, n = 22) showed a significant reduction in mEPSC amplitude and frequency in HIP1R-KD neurons compared to the control. (D) Example traces of mIPSC. (E) Graph of cumulative probability and bar graph of amplitude mIPSC (HIP1R-KD, 53.33 ± 2.32 pA, n = 25 vs. the control, 50.98 ± 3.36 pA, n = 21), and (F) graph of cumulative probability and bar graph of inter-event interval of mIPSC (HIP1R-KD, 1.33 ± 0.27 s, n = 25 vs. the control, 1.37 ± 0.31 s, n = 21) did not show marked differences in mIPSC amplitude and frequency between the two groups. All data are presented as mean ± SEM, unpaired two-tailed t-test. **P < 0.01, ^#P < 0.0001.

Figure 7

Transfection of mouse HIP1R rescues the dendrite growth defects in HIP1R-knockeddown neurons. (A) A schematic illustration showing mouse full-length and several truncated forms of HIP1R cDNA in the myc-tagged expressing vector. (B) Cultured hippocampal neurons at DIV6 were co-transfected with HIP1R-shRNA for knocking down endogenous HIP1R expression, and full-length or truncated HIP1R vectors for re-expressing exogenous cognates. Transfected neurons were immunostained with myc and MAP2 antibodies at DIV12. GFP-expressing and positively myc stained neurons were used for morphological analysis. (C) ANOVA analysis with repeated measures of Sholl analysis showed that full-length HIP1R re-expression fully rescued the complexity impairment in HIP1R KD neurons. HIP1R_{1–350/766–1068} had a partial rescue effect. (D,E) Unpaired two-tailed t-test analysis of total dendrite length and number. Total dendrite length: 1076.65 ± 36.67 μm, n = 31 for the control group (CTL); 1098.99 ± 56.12 μm, n = 32 for KD+myc-HIP1R; 425.94 ± 20.50 μm, n = 33 for KD+myc-HIP1R_350–1068; 817.70 ± 42.27 μm, n = 36 for KD+myc -HIP1R_{1–350/766–1068}; 437.96 ± 28.07 μm, n = 37 for KD+myc-HIP1R_1–655; and 374.80 ± 26.39 μm, n = 31 for KD + Vector. Total dendrite number: 44.45 ± 2.99, n = 31 for the control group (CTL); 43.80 ± 1.81, n = 32 for KD+myc-HIP1R; 13.85 ± 0.93, n = 33 for KD+myc-HIP1R_350–1068; 32.53 ± 2.56, n = 36 for KD+myc-HIP1R_{1–350/766–1068}; 15.19 ± 1.22, n = 37 for KD+myc-HIP1R_1–655; and 11.83 ± 0.83, n = 31 for KD + Vector. All data are presented as mean ± SEM. **P < 0.01, ^#P < 0.0001.

Figure 8

Expression of the HIP1R C-terminus and its proline-rich region confer a dominant negative effect on dendrite development. (A) A schematic illustration showing full-length and several truncated forms of HIP1R cDNA in the myc-tagged expressing vector. (B) Cultured hippocampal neurons were transfected at DIV6 with each of the myc-tagged HIP1R mutants. An empty vector was taken as a negative control and GFP-shRNA as a positive control. Representative images of transfected neurons are shown immunostained at DIV12 with myc and MAP2 antibodies. Scale bar=20 μm. (C) ANOVA with repeated measures for Sholl analysis showed a significant decrease in dendrite complexity in HIP1R_766–1068 and HIP1R_1018–1068 transfected neurons. (D,E) Quantitative analysis with unpaired two-tailed t-tests showed a partial decrease in dendrite number and total length in HIP1R_766–1068 and HIP1R_1018–1068 transfected neurons compared to the negative and positive controls. Dendrite number: 44.69 ± 2.18, n = 42 for the negative control (CTL); 43.70 ± 3.30, n = 46 for myc-HIP1R_1–350; 42.30 ± 2.95, n = 44 for myc-HIP1R_350–655; 30.21 ± 2.03, n = 48 for myc-HIP1R_766–1068; 42.98 ± 2.23, n = 42 for myc-HIP1R_766–1017; 29.70 ± 1.59, n = 40 for myc-HIP1R_1018–1068; 12.17 ± 0.68, n = 42 for the positive control (KD). Dendrite total length: 1041.91 ± 29.58 μm, n = 42 for the negative control (CTL); 957.21 ± 42.99 μm, n = 46 for myc-HIP1R_1–350; 965.05 ± 47.24 μm, n = 44 for myc-HIP1R_350–655; 642.64 ± 34.09 μm, n = 48 for myc-HIP1R_766–1068; 942.51 ± 37.93 μm, n = 42 for myc-HIP1R_766–1017; 711.39 ± 30.27 μm, n = 40 for myc-HIP1R_1018–1068; 401.25 ± 21.66 μm, n = 42 for the positive control (KD). (F,G) Myc-vector or myc-HIP1R1018–1068 plasmid was transfected into DIV6–7 hippocampal neurons. Transfected neurons were stained with anti-myc antibody to analyze the spine density at DIV18–21. Spine density per μm: 0.18 ± 0.01, n = 55 for myc-HIP1R1018–1068 vs. 0.49 ± 0.01, n = 55 for the vector control. All data are presented as mean ± SEM. ^#P < 0.0001.