Collapsin Response Mediator Protein 2 and Endophilin2 Coordinate Regulation of AMPA Receptor GluA1 Subunit Recycling

Abstract

The dynamic trafficking of AMPA receptors (AMPARs), which enables the endocytosis, recycling, and exocytosis of AMPARs, is crucial for synaptic plasticity. Endophilin2, which directly interacts with the GluA1 subunit of AMPARs, plays an important role in AMPAR endocytosis. Collapsin response mediator protein 2 (CRMP2) promotes the maturation of the dendritic spine and can transfer AMPARs to the membrane. Although the mechanisms of AMPAR endocytosis and exocytosis are well known, the exact molecular mechanisms underlying AMPAR recycling remain unclear. Here, we report a unique interaction between CRMP2 and endophilin2. Our results showed that overexpression of CRMP2 and endophilin2 increased the amplitude and frequency of miniature excitatory synaptic currents (mEPSCs) and modestly enhanced AMPAR levels in hippocampal neurons. Furthermore, the CRMP2 and endophilin2 overexpression phenotype failed to occur when the interaction between these two proteins was inhibited. Further analysis revealed that this interaction was regulated by CRMP2 phosphorylation. The phosphorylation of CRMP2 inhibited its interaction with endophilin2; this was mainly affected by the phosphorylation of Thr514 and Ser518 by glycogen synthase kinase (GSK) 3β. CRMP2 phosphorylation increased degradation and inhibited the surface expression of AMPAR GluA1 subunits in cultured hippocampal neurons. However, the dephosphorylation of CRMP2 inhibited degradation and promoted the surface expression of AMPAR GluA1 subunits in cultured hippocampal neurons. Taken together, our data demonstrated that the interaction between CRMP2 and endophilin2 was conductive to the recycling of AMPA receptor GluA1 subunits in hippocampal neurons.

Keywords: AMPA receptor; GluA1 subunit; collapsing response mediator protein 2; endophilin2; recycling.

Figure 1

Collapsin response mediator protein 2 (CRMP2) interacts with endophilin2. (A) Electrostatic surface potential maps of CRMP2 (left) and endophilin2 SH3 domain (right). The blue color represents the positive potential, and red color represents the negative potential. The electrostatic potential energy of the protein surface is calculated by APBS (figures are produced by using PyMOL). (B) The structures of the SH3 domain of endophilin2 (PDB code: 3C0C) and CRMP2 (5X1A) were downloaded from the Protein Data Bank www.rcsb.org. (C) Purified glutathione-S-transferase (GST) and GST-Endo2 proteins were incubated with 1-month-old Sprague–Dawley rat brain lysates. Input and bound proteins were analyzed by immunoblotting, using antibodies against CRMP2. (D) Purified GST and GST-CRMP2 proteins were incubated with 1-month-old Sprague–Dawley rat brain lysates. Input and bound proteins were analyzed by immunoblotting, using antibodies against Endo2. (E,F) HEK293 cells were cotransfected with green fluorescence protein (GFP)-Endo2 and Flag-CRMP2, followed by the co-IP assay. Protein complexes coeluted with the anti-Flag antibody were detected using the anti-GFP antibody (E) and protein complexes coeluted with the anti-GFP antibody were detected using the anti-Flag antibody (F). All experiments were repeated independently at least three times. (G) Representative image of DIV12 cultured hippocampal neurons stained for endophilin2 and CRMP2. White arrowheads indicate endophilin2 puncta that colocalized with CRMP2. Scale bars of 20 and 5 μm in the magnified dendrite. (H) Intensity trace of the endophilin2 and CRMP2 along the dendrite.

Figure 2

C-terminal residues 303–368 of endophilin2 interact with C-terminal residues 381–562 of CRMP2. (A,B) Schematic diagrams of endophilin2 and CRMP2 truncation mutants are shown according to their different domains. (C) Purification of glutathione-S-transferase (GST) and truncated segments of GST-endophilin2 (GST-Endo2) fusion proteins. Bovine serum albumin (BSA) standard protein: 4, 8, and 12 μg. (D) Purification of GST and truncated segments of GST-CRMP2 fusion proteins. BSA standard protein: 4, 8, and 12 μg. (E) Pull-down assays were conducted with GST-Endo2 and its truncation mutants using 1-month-old Sprague–Dawley rat brain lysates. Western blotting was performed with an anti-CRMP2 antibody. (F) Pull-down assays were conducted with GST-CRMP2 and its truncation mutants using 1-month-old Sprague–Dawley rat brain lysates. Western blotting was performed with an anti-endophilin2 antibody. All experiments were repeated independently at least three times.

Figure 3

CRMP2 and endophilin2 coordinate with each other to promote the current level of AMPA receptors (AMPARs). (A) Representative tracings of miniature excitatory synaptic current (mEPSC) recorded from cultured hippocampal neurons transfected with mCherry-Endo2 and GFP, mCherry-Endo2 and GFP-CRMP2, and mCherry-Endo2 and GFP-CRMP2^ΔC381. (B,C) Cumulative probabilities of the mEPSC amplitude and interevent interval of neurons in these groups. Insets: histogram plots of mEPSC amplitude and frequency of neurons in these groups, n = 15 cells. *p < 0.05, as compared to the mCherry-Endo2 and GFP group; ^#p < 0.05, as compared to the mCherry-Endo2 and GFP-CRMP2 group. (D) Representative tracings of mEPSC recorded from cultured hippocampal neurons transfected GFP-CRMP2 and mCherry, GFP-CRMP2 and mCherry-Endo2, and GFP-CRMP2 and mCherry-Endo2^ΔC303. (E,F) Cumulative probabilities of the mEPSC amplitude and interevent interval of neurons in these groups. Insets: Histogram plots of mEPSC amplitude and frequency of neurons in these groups, n = 15 cells. *p < 0.05, as compared to the GFP-CRMP2 and mCherry group, ^#p < 0.05, as compared to the GFP-CRMP2 and mCherry-Endo2 group.

Figure 4

CRMP2 and endophilin2 coordinate to promote the surface expression of AMPA receptor (AMPAR) GluA1 subunit. (A) Cultured hippocampal neurons transfected with mCherry-Endo2 and GFP, mCherry-Endo2 and GFP-CRMP2, and mCherry-Endo2 and GFP-CRMP2^ΔC381 were immunostained with surface GluA1. Scale bar, 50 and 10 μm in the magnified dendrite. (B) Quantitative analysis of fluorescence intensity of surface GluA1 of neurons in these groups, n = 15 cells per group. *p < 0.05, as compared to the mCherry-Endo2 and GFP group; ^#p < 0.05, as compared to the mCherry-Endo2 and GFP-CRMP2 group. (C) Cultured hippocampal neurons transfected with GFP-CRMP2 and mCherry, GFP-CRMP2 and mCherry-Endo2, and GFP-CRMP2 and mCherry-Endo2^ΔC303 were immunostained with surface GluA1. Scale bar, 50 and 10 μm in the magnified dendrite. (D) Quantitative analysis of fluorescence intensity of surface GluA1 of neurons in these groups, n = 15 cells per group. *p < 0.05, as compared to the GFP-CRMP2 and mCherry group; ^#p < 0.05, as compared to the mCherry-Endo2 and GFP-CRMP2 group.

Figure 5

The phosphorylation of CRMP2 decreases its interaction with endophilin2. (A) Schematic of GSK3β and Cdk5 and RhoK phosphorylation sites within the rat CRMP2 sequence. Numbers represent amino acid residues within the CRMP2 sequence. (B) The expression and purification of glutathione-S-transferase (GST), GST-CRMP2, and its phosphomimetic and dephosphomimetic mutants. BSA standard protein: 2, 4, and 8 μg. (C) Pull-down assays were conducted using GST-CRMP2 and its phosphomimetic and dephosphomimetic mutants in brain lysates of 1-month-old Sprague–Dawley rats. Western blotting was performed using an anti-endophilin2 antibody. (D) Pull-down assays were conducted with GST-CRMP2 and its phosphomimetic and dephosphomimetic mutants in brain lysates of 1-month-old Sprague–Dawley rats. Western blotting was performed with an anti-tubulin antibody. Each result is representative of three separate experiments with similar results.

Figure 6

CRMP2 phosphorylation reduces the surface expression of GluA1 in hippocampal neurons. (A) Representative tracings of mEPSC recorded from cultured hippocampal neurons transfected with mCherry-Endo2 and GFP, mCherry-Endo2 and GFP-CRMP2, mCherry-Endo2 and GFP-CRMP2^QmA, and mCherry-Endo2 and GFP-CRMP2^QmD. (B,C) Cumulative probabilities of the mEPSC amplitude and interevent interval of neurons in these groups. Insets: histogram plots of mEPSC amplitude and frequency of neurons in these groups, n = 15 cells per group. *p < 0.05, as compared to the mCherry-Endo2 and GFP group, ^#p < 0.05, as compared to the mCherry-Endo2 and GFP-CRMP2^QmD group. (D) Cultured hippocampal neurons transfected with mCherry-Endo2 and GFP, mCherry-Endo2 and GFP-CRMP2, mCherry-Endo2 and GFP-CRMP2^QmA, and mCherry-Endo2 and GFP-CRMP2^QmD were immunostained with surface GluA1. Scale bar, 50 and 10 μm in the magnified dendrite. (E) Quantitative analysis of fluorescence intensity of surface GluA1 of neurons in these groups, n = 15 cells per group. *p < 0.05, as compared to the mCherry-Endo2 and GFP group, ^#p < 0.05, as compared to the mCherry-Endo2 and GFP-CRMP2^QmD group.

Figure 7

The phosphorylation of CRMP2 at Thr514/Ser518 reduces its interaction with endophilin2. (A) The expression and purification of glutathione-S-transferase (GST)-CRMP2 and its phosphorylated and dephosphorylated mutants. (B) Pull-down assays were conducted with GST-CRMP2 and its phosphorylated and dephosphorylated mutants in brain lysates of 1-month-old rats. Western blotting was performed with an anti-endophilin2 antibody (up); quantitative analysis of the relative binding of Endo2 to CRMP2 was performed (down), n = 3, **p < 0.01, ***p < 0.001 compared to GST-CRMP2 group. (C) HEK293 cells were cotransfected with Flag-Endo2 and GFP, GFP-CRMP2, and its dephosphomimetic and dephosphomimetic mutants, after which the co-IP assay was performed. Protein complexes that were coeluted with the anti-GFP antibody were detected using the anti-Flag antibody (up); quantitative analysis of the relative binding of Endo2 to CRMP2 was performed (down), n = 3, **p < 0.01 compared to the CRMP2 group. (D) Cultured hippocampal neurons were infected with Ad-HA vector and Ad-HA-GSK3β S9A (constitutively active GSK3β mutant containing a serine-to-alanine substitution at residue 9) for 24 h. The cell lysates were immunoprecipitated with anti-CRMP2 or the control immunoglobulin G (IgG) antibody, followed by the immunoblotting of immunoprecipitated samples with anti-endophilin2 antibody. (E) Quantification of Endo2-immureactive bands after normalization with coprecipitated CRMP2 levels, n = 3, **p < 0.01.

Figure 8

The phosphorylation of CRMP2 at Thr514/Ser518 reduces the surface expression and increases degradation of GluA1. (A) Cultured hippocampal neurons transfected with mCherry-Endo2 and GFP, and mCherry-Endo2 and GFP (incubated with CHIR99021), mCherry-Endo2 and GFP-CRMP2, mCherry-Endo2 and GFP-CRMP2^T514A/S518A, and mCherry-Endo2 and GFP-CRMP2^T514D/S518D; then, they were immunostained with surface GluA1. Scale bar, 50 and 10 μm in the magnified dendrite. (B) Quantitative analysis of fluorescence intensity of surface GluA1 of neurons in these groups, n = 15 cells per group, *p < 0.05, as compared to the mCherry-Endo2 and GFP group; ^#p < 0.05, as compared to the mCherry-Endo2 and GFP-CRMP2 group. (C,D) Representative images (C) and quantitative analysis (D) show the colocalization of EEA1 and internalized GluA1 in hippocampal neurons. Neurons were coimmunostained at DIV12, using antibodies against GluA1 and EEA1. Scale bar, 10 μm. (E,F) Representative images (E) and quantitative analysis (F) show the colocalization of LAMP1 and internalized GluA1 in hippocampal neurons. Neurons were coimmunostained at DIV12, using antibodies against GluA1 and LAMP1. Scale bar, 10 μm. n = 15 cells per group. *p < 0.05, as compared to the mCherry-Endo2 and GFP group; ^#p < 0.05, as compared to the mCherry-Endo2 and GFP-CRMP2 group.