SHP2 regulates GluA2 tyrosine phosphorylation required for AMPA receptor endocytosis and mGluR-LTD

Abstract

Posttranslational modifications regulate the properties and abundance of synaptic α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptors that mediate fast excitatory synaptic transmission and synaptic plasticity in the central nervous system. During long-term depression (LTD), protein tyrosine phosphatases (PTPs) dephosphorylate tyrosine residues in the C-terminal tail of AMPA receptor GluA2 subunit, which is essential for GluA2 endocytosis and group I metabotropic glutamate receptor (mGluR)-dependent LTD. However, as a selective downstream effector of mGluRs, the mGluR-dependent PTP responsible for GluA2 tyrosine dephosphorylation remains elusive at Schaffer collateral (SC)-CA1 synapses. In the present study, we find that mGluR5 stimulation activates Src homology 2 (SH2) domain-containing phosphatase 2 (SHP2) by increasing phospho-Y542 levels in SHP2. Under steady-state conditions, SHP2 plays a protective role in stabilizing phospho-Y869 of GluA2 by directly interacting with GluA2 phosphorylated at Y869, without affecting GluA2 phospho-Y876 levels. Upon mGluR5 stimulation, SHP2 dephosphorylates GluA2 at Y869 and Y876, resulting in GluA2 endocytosis and mGluR-LTD. Our results establish SHP2 as a downstream effector of mGluR5 and indicate a dual action of SHP2 in regulating GluA2 tyrosine phosphorylation and function. Given the implications of mGluR5 and SHP2 in synaptic pathophysiology, we propose SHP2 as a promising therapeutic target for neurodevelopmental and autism spectrum disorders.

Keywords: GluA2; SHP2; mGluR-LTD; mGluR5; trafficking.

Fig. 1.

mGluR5 stimulation activates SHP2 through phosphorylation at Y542. (A) Primary hippocampal neurons (DIV 14 to 16) were treated with 100 μM DHPG for 10 min. Representative western blots are shown. (B) Quantification data from panel (A) (n = 5, paired Student’s t test). (C) Primary hippocampal neurons were treated with 20 μM CPCCOEt (CPCC) or 10 μM MPEP for 20 min, and 100 μM DHPG was added for an additional 10 min. (D) Quantification data from panel (C) (n = 4, one-way ANOVA with Tukey’s multiple comparison test). (E) Primary hippocampal neurons were treated with 1 mM CHPG, a selective mGluR5 agonist, or 100 μM DHPG for 10 min. (F) Quantification data from panel (E) (n = 3; one-way ANOVA with Tukey’s multiple comparison test). (G) Primary hippocampal neurons were infected with lentivirus expressing shRNA against mGluR5 (mGluR5 KD) or a nonrelated target (Ctl KD) and then treated with 100 μM DHPG for 10 min. (H) Quantification data from panel (G) (n = 3, one-way ANOVA with Tukey’s multiple comparison test).

Fig. 2.

SHP2 protects phospho-Y869 of GluA2 through its SH2 domain. (A) Primary hippocampal neurons were infected with lentivirus expressing either control shRNA (Ctl KD) or SHP2 shRNA (SHP2 KD) for 10 d. At DIV 14 to 16, western blot analysis was performed. (B) Quantification data from panel (A) (n = 4, paired Student’s t test). (C) Primary hippocampal neurons were infected with lentivirus expressing either control shRNA (Ctl KD::GFP) or SHP2 shRNA (SHP2 KD::GFP), and simultaneously with SHP2 shRNA-resistant SHP2 full-length (SHP2 KD::SHP2-FL) or SHP2 SH2 domain fragment (SHP2 KD::SHP2-SH2) for 10 d. Representative western blots are shown. (D) Primary hippocampal neurons were treated with 10 μM SHP099 or 10 μM NSC87877 for 30 min and subjected to western blot assay. (E) Quantification data from panel (D) (n = 5, one-way ANOVA with Dunnett’s multiple comparison test). (F) Binding analysis of SHP2 to GluA2. Inhibitor-treated neuron lysates were immunoprecipitated with an anti-GluA2 antibody, followed by immunoblotting with an anti-SHP2 antibody. (G) Quantification of relative SHP2 binding to GluA2 from panel (F) (n = 6, paired Student’s t test).

Fig. 3.

SHP2 directly binds to GluA2 phosphorylated at Y869 and dephosphorylates GluA2 tyrosine residues. (A) GST pull-down assay to analyze direct binding between phosphorylated GluA2ct and recombinant SHP2. (B) Binding of recombinant SHP2 to pGluA2ct WT, Y869F, or Y876F. (C) The binding domain of SHP2 to pGluA2ct was analyzed using a GST pull-down assay. The arrowhead indicates the dimer of the N-SH2 fragment. (D) Recombinant SHP2 PTP/CT dephosphorylates GluA2 tyrosine residues. Src-phosphorylated GluA2ct was incubated with SHP2 PTP/CT. (E) Relative phosphorylation levels of each tyrosine residue from panel (D). The kinetics of SHP2-mediated dephosphorylation of GluA2 phospho-Y869 or phospho-Y876 was fitted using one-phase exponential decay curves (pY869, R² = 0.99, half-life = 14.4 min; pY876, R² = 0.98, half-life = 23.1 min; n = 4).

Fig. 4.

SHP2 regulates GluA2 tyrosine phosphorylation and surface expression. (A and B) Effects of SHP2 inhibitors on the tyrosine phosphorylation levels of GluA2. (A) Primary hippocampal neurons (DIV 14 to 16) were treated with 10 μM NSC87877 or 10 μM SHP099 for 30 min, and 100 μM DHPG was added for an additional 10 min. Representative western blots are shown. (B) Quantification data from panel (A) (n = 6, one-way ANOVA with Tukey’s multiple comparison test). (C and D) Effects of SHP2 inhibitors on the surface expression of GluA2. (C) Primary hippocampal neurons (DIV 14 to 16) were treated with 10 μM NSC87877 or 10 μM SHP099 for 30 min, and 100 μM DHPG was added for an additional 10 min. Neurons were subsequently incubated in conditioned medium for 15 min. For SHP099 experiments, SHP099 was added to the conditioned medium because of its reversible inhibitory properties. Surface-expressed GluA2 was immunostained under nonpermeabilized conditions. (Scale bar, 7 μm.) (D) Quantification of surface-expressed GluA2 from 30-μm-long secondary branches (n = 4, number of branches = 78−211, Kruskal–Wallis test with Dunn’s multiple comparison test).

Fig. 5.

GluA2 tyrosine phosphorylation and endocytosis are dysregulated in Ptpn11^Y542F/+ mice. (A) Crude synaptosomal P2 fractions were prepared from the hippocampi of 5- to 8-wk-old Ptpn11^Y542F/+ mice (Y542F/+) or their WT littermates (+/+). Representative western blots and quantification data are presented (n = 4, number of mice = 7 to 9, unpaired Student’s t test for SHP2, SHP2 pY580, GluA1 pS845, GluA2 pY876, mGluR5, PSD-95, ERK, Src, and Src pY527; Mann–Whitney U test for SHP2 pY542, GluA1, GluA2, GluA2 pY869, GluN2B, GluN2B pY1252, GluN1, pERK, and Src pY416). (B) Western blot analysis using primary cultured neurons. (C) Quantification data from panel (B) (n = 5 to 9, paired Student’s t test for GluA2 pY876 and SHP2 pY542; Wilcoxon signed-rank test for GluA2 pY869 and pERK). (D) After treating primary cultured neurons with 100 μM DHPG for 10 min, neurons were incubated in conditioned medium for 15 min. The surface-expressed GluA2 was visualized using a GluA2 N-terminal antibody under nonpermeabilized conditions. (Scale bar, 7 μm.) (E) Quantification of surface-expressed GluA2 from 30-μm-long secondary branches (n = 3, number of branches = 53 to 90, Kruskal–Wallis test with Dunn’s multiple comparison test).

Fig. 6.

mGluR-LTD is impaired by SHP2 dysfunction. (A−D) Effects of pharmacological inhibition of SHP2 on mGluR-LTD. Time course plot of fEPSP slopes of 100 μM DHPG-LTD (A) or PP-LFS-LTD (B) with or without 10 μM NSC87877 treatment. Time course plot of fEPSP slopes of DHPG-LTD (C) or PP-LFS-LTD (D) with or without 10 μM SHP099 treatment. For SHP099 experiments, SHP099 was continuously added to the artificial cerebrospinal fluid (ACSF) because of its reversible inhibitory properties. (E and F) mGluR-LTD in Ptpn11^Y542F/+ mice. Time course plot of fEPSP slopes of DHPG-LTD (E) or PP-LFS-LTD (F). Sample traces show representative fEPSPs 1 min before (black) and 44 min after (red) induction. (Scale bar, 0.5 mV, 5 ms.) Quantification of the average fEPSP slopes and the number of slices are presented in SI Appendix, Fig. S14.