Regulation of STEP61 and tyrosine-phosphorylation of NMDA and AMPA receptors during homeostatic synaptic plasticity

Background Sustained changes in network activity cause homeostatic synaptic plasticity in part by altering the postsynaptic accumulation of N-methyl-D-aspartate receptors (NMDAR) and α-amino-3-hydroxyle-5-methyl-4-isoxazolepropionic acid receptors (AMPAR), which are primary mediators of excitatory synaptic transmission. A key trafficking modulator of NMDAR and AMPAR is STriatal-Enriched protein tyrosine Phosphatase (STEP61) that opposes synaptic strengthening through dephosphorylation of NMDAR subunit GluN2B and AMPAR subunit GluA2. However, the role of STEP61 in homeostatic synaptic plasticity is unknown. Findings We demonstrate here that prolonged activity blockade leads to synaptic scaling, and a concurrent decrease in STEP61 level and activity in rat dissociated hippocampal cultured neurons. Consistent with STEP61 reduction, prolonged activity blockade enhances the tyrosine phosphorylation of GluN2B and GluA2 whereas increasing STEP61 activity blocks this regulation and synaptic scaling. Conversely, prolonged activity enhancement increases STEP61 level and activity, and reduces the tyrosine phosphorylation and level of GluN2B as well as GluA2 expression in a STEP61–dependent manner. Conclusions Given that STEP61-mediated dephosphorylation of GluN2B and GluA2 leads to their internalization, our results collectively suggest that activity-dependent regulation of STEP61 and its substrates GluN2B and GluA2 may contribute to homeostatic stabilization of excitatory synapses. Electronic supplementary material The online version of this article (doi:10.1186/s13041-015-0148-4) contains supplementary material, which is available to authorized users.


Background
In response to sustained changes in neuronal activity, homeostatic synaptic plasticity maintains synaptic strength and flexibility within physiological limit. This plasticity is expressed in part by dynamic changes in the postsynaptic levels of NMDARs and AMPARs that mediate excitatory synaptic transmission [1]. A key trafficking modulator of both NMDAR and AMPAR is STEP 61 , a protein tyrosine (Tyr) phosphatase in the central nervous system that has two main alternatively spliced forms, the cytosolic STEP 46 and the membrane-associated STEP 61 [2]. Tightly associated with the postsynaptic density, STEP 61 regulates the Tyr phosphorylation and surface density of NMDARs and AMPARs [3][4][5][6]. This regulation contributes to Hebbian long-term potentiation [4,7] and several neuropsychiatric disorders most notably Alzheimer's disease [4] and Fragile X-syndrome [2].
We have previously identified mRNA transcripts whose expressions are regulated by prolonged activity perturbation [8] due to a critical role of transcription in homeostatic synaptic plasticity [9,10]. Of these activity-regulated transcripts, we identified PTPN5 that encodes STEP [8]. The present study investigated whether STEP 61 contributes to homeostatic synaptic plasticity.

Results and discussion
Prolonged alterations of hippocampal network activity regulate STEP 61 level and activity Prolonged blockade of network activity for 48 h with the sodium channel blocker tetrodotoxin (TTX) induced synaptic scaling in dissociated hippocampal cultured neurons as demonstrated previously [11,12] (Fig. 1a-d), and reduced STEP 61 mRNA and protein expression compared to CTL treatment (Fig. 1e, f). Conversely, prolonged activity enhancement for 48 h using the GABA A receptor antagonist bicuculline (BC) increased STEP 61 protein level (Fig. 1g), but did not alter its mRNA level and the miniature excitatory postsynaptic current (mEPSC) (Fig. 1a-e).
To test whether prolonged TTX or BC treatment affects STEP 61 activity, we examined the phosphorylation of STEP 61 at Ser 221 within its kinase-interactive motif domain, which prevents STEP 61 interaction with all known substrates (Fig. 2a) [13]. TTX treatment for 36-48 h enhanced Ser 221 -phosphorylation of STEP 61 , indicating decreased STEP 61 activity (Fig. 2b, d). In contrast, 36-48 h BC treatment reduced Ser 221 -phosphorylation, indicating increased STEP 61 activity (Fig. 2c, d).
Prolonged alterations of hippocampal network activity regulate Tyr-phosphorylation of GluN2B and GluA2 in a STEP 61 -dependent manner STEP 61 dephosphorylates the NMDAR subunit GluN2B at Tyr 1472 [5,6] and reduces Tyr-phosphorylation of the AMPAR subunit GluA2 following group 1 metabotropic glutamate receptor (mGluR) stimulation [3]. Although the specific Tyr residues on GluA2 regulated by STEP 61 are b Normalized cumulative fraction of the mEPSC amplitudes. c-d Summary plots of average mEPSC amplitudes (c) and frequencies (d) for CTL (n = 19), TTX (n = 20), or BC (n = 16). TTX treatment for 48 h induced synaptic scaling whereas 48 h BC treatment had no effect. e Microarray (n = 4) and QPCR (n = 5) analyses revealed that 48 h application of TTX but not BC reduced the expression of PTPN5, which encodes STEP. f-g Immunoblot analysis of STEP 61 following 48 h administration of CTL, TTX (F), BC (g) (n = 9 per treatment). Data shown represent the mean ± SEM (*p < 0.05; **p < 0.01) unknown, the GluA2 phosphorylation state at Tyr 869 , Tyr 873 , and Tyr 876 (3Tyr) regulates AMPAR trafficking [14].
To determine if the TTX-or BC-induced changes in STEP 61 alter Tyr-phosphorylation of GluN2B and GluA2, we performed immunoblot analyses using specific antibodies to phosphorylated Tyr 1472 of GluN2B [6] and phosphorylated 3Tyr of GluA2 [14] (Fig. 3).
Consistent with the TTX-induced decrease in STEP 61 level and activity (Fig. 2b, d), prolonged TTX treatment increased the levels of Tyr 1472 -phosphorylated GluN2B (GluN2B-pY 1472 ) and 3Tyr-phopshorylated GluA2 (GluA2-p3Y) compared to CTL treatment without affecting their total protein expression ( Fig. 3a-c). In contrast, BC treatment for 24-48 h decreased the levels of GluN2B-pY 1472 and GluA2-p3Y ( Fig. 3d-f), concurrently with an increase in STEP 61 level and activity (Fig. 2c, d). Interestingly, total levels of GluN2B and GluA2 were reduced by 48 h BC application (Fig. 3d, e). We next examined if STEP 61 mediates the TTX-or BC-induced changes in Tyr-phosphorylation of GluN2B and GluA2. Transactivator of transcription (TAT) sequence was fused to STEP 46 and a myc tag (Fig. 4a), allowing the TAT fusion proteins to be membrane permeable (Additional file 1: Figure S1A, B) [15]. Preincubation for 30 min with active TAT-STEP wild-type (WT) but not control inactive TAT-myc reduced the levels of GluN2B-pY 1472 and GluA2-p3Y in CTL-treated neurons (Fig. 4b, c, Additional file 1: Figure S1C) and occluded the TTXinduced increase in GluN2B-pY 1472 and GluA2-p3Y levels compared to CTL application (Fig. 4b, c), suggesting that the increase in Tyr-phosphorylation of GluN2B and GluA2 is mediated by the TTX-induced reduction in STEP 61 .
In TAT-STEP C/S, a C300S point mutation inactivates STEP 46 , allowing it to bind constitutively to substrates but not to dephosphorylate them [13,15,16]. Consistently, introduction of TAT-STEP C/S in CTL-treated neurons Fig. 2 Prolonged alterations of hippocampal network activity regulate STEP 61 activity. a A schematic depicting the regulation of STEP 61 activity by its phosphorylation at Ser 221 within its kinase-interactive motif, a binding site for all STEP substrates. b-d Immunoblot analysis of STEP 61 and Ser 221 -phosphorylated STEP 61 (STEP 61 -pS 221 ) in hippocampal cultured neurons following CTL, TTX, or BC treatment for 24-48 h (n = 3 per treatment). b, d Prolonged TTX treatment reduced STEP 61 protein level and activity. c, d Prolonged BC treatment enhanced STEP 61 protein level and activity. d The relative phosphorylation state of STEP 61 as calculated by the ratio of STEP 61 -pS 221 level over total STEP 61 level. Data shown represent the mean ± SEM (*p < 0.05; **p < 0.01; ***p < 0.005) significantly increased the levels of GluN2B-pY 1472 and GluA2-p3Y compared to TAT-myc application (Fig. 4d, e, Additional file 1: Figure S1D). Preincubation with TAT-STEP C/S but not TAT-myc blocked the BC-induced reduction in the levels of GluN2B-pY 1472 , total GluN2B, and total GluA2 but not GluA2-p3Y (Fig. 4d, e). Since specific Tyr residues regulated by STEP 61 remain unknown, our analyses for GluA2-p3Y may not have revealed the effect of TAT-STEP C/S if STEP 61 causes dephosphorylation of only one Tyr. Nonetheless, these results suggest that STEP 61 mediates the BC-induced changes in Tyr 1472 -phosphorylation of GluN2B and abundance of GluN2B and GluA2.

Enhancement of STEP activity blocks synaptic scaling
Dephosphorylation of Tyr 1472 within a conserved endocytic motif of GluN2B [17] via STEP 61 reduces surface NMDAR level [4,6] by clathrin-mediated internalization [18]. Furthermore, AMPAR internalization can be induced by mGluR stimulation through STEP 61 [3] and by dephosphorylation of GluA2 at 3Tyr [14]. We hypothesized that prolonged activity blockade induces synaptic scaling (Fig. 1a-d) by inhibiting endocytosis of synaptic NMDARs and AMPARs upon STEP 61 reduction (Fig. 1f, Fig 2b). To test this, we enhanced STEP activity by administering TAT-STEP WT for 30 min prior to recording. In the presence of TAT-myc, 48 h TTX treatment increased the mEPSC amplitude but not frequency compared to CTL application ( Fig. 5a-d). However, this TTX-induced synaptic scaling was abolished by TAT-STEP WT preincubation (Fig. 5a-d), indicating that STEP 61 reduction contributes to synaptic scaling.
It remains unknown how prolonged activity perturbation regulates STEP 61 . Previous studies have reported that Ser 221 of STEP 61 is dephosphorylated by calciumdependent calcineurin upon NMDAR activation [13] and phosphorylated by protein kinase A (PKA) upon stimulation of dopamine D1 receptor [23]. Interestingly, synaptic scaling is shown to involve reduced calcium influx to the and 3Tyr-phosphorylation of GluA2 (GluA2-p3Y) (c). d-e TAT-STEP C/S blocked the BC-induced reduction in Tyr 1472 -phosphorylation and level of GluN2B (d) as well as GluA2 level but not 3Tyr-phosphorylation of GluA2 (e). Data shown represent the mean ± SEM following normalization to CTL values in the presence of TAT-myc (black bars), TAT-STEP WT (striped bars), and TAT-STEP C/S (gray bars) (*p < 0.05; **p < 0.01; ***p < 0.005) postsynaptic neuron [9], reduced calcineurin activity [24], and enhanced PKA activity at excitatory synapses [25]. Hence, prolonged activity blockade could increase Ser 221phosphorylation of STEP 61 (Fig. 2b, d) by reduced calcineurin activity and/or enhanced PKA activity, in addition to decreasing STEP 61 level by transcriptional down-regulation (Fig. 1e, f). Considering that a loss of PKA from synapses was found during synaptic downscaling [25], reduced PKA activity may contribute to the BC-induced decrease in Ser 221 -phosphorylation of STEP 61 (Fig. 2c, d).

Conclusions
In summary, we demonstrate a bidirectional modulation of STEP 61 level and activity by prolonged alterations of hippocampal network activity, resulting in correlative changes in Tyr-phosphorylation of STEP 61 substrates, GluN2B and GluA2. We also show that the reduction in STEP 61 contributes to synaptic scaling. Future studies should test if this regulation alters NMDAR and AMPAR surface density during homeostatic plasticity (Fig. 5e). Investigating how prolonged activity perturbation regulates STEP 61 should provide mechanistic insights into the dysregulation of STEP 61 expression, which are present in multiple neuropathologies [2].

Hippocampal neuronal culture
The Institutional Animal Care and Use Committee at the University of Illinois Urbana-Champaign approved all experimental procedures involving animals. Primary dissociated hippocampal cultures were prepared from Sprague-Dawley rat embryos at embryonic day 18 and plated at high density (330 cells/mm 2 ) as described [8]. At 10-13 days in vitro, neurons were treated for 24-48 h

Electrophysiology
Whole-cell patch-clamp recordings of mEPSCs (>150 events per neuron) were performed at 23-25°C from pyramidal neurons held at −60 mV in external solution containing 1 μM TTX and 20 μM BC as described [12,25] using a Multiclamp 700B amplifier, Digidata1440A, and the pClamp 10.2 (Molecular Devices). Signals were acquired 3 min after making the whole-cell configuration, filtered at 1 kHz, and sampled at 10 kHz on gap free mode (5 min). The mEPSCs were detected with a 10 pA thresholds and analyzed by Mini Analysis (Synaptosoft).

QPCR
The QPCR was performed with the StepOnePlus realtime PCR system (Applied Biosystems) using total RNA (1-2 μg) as described [8]. The forward and reverse primer sequences for PTPN5 were 5'-GGAGTCAGCCCATGAA TACC-3' and 5'-CAGACGTACCCTGCTGTGAG-3' respectively. The primer sequences for GAPDH has been previously described [8]. Following normalization to control GAPDH cDNA levels, the fold change of PTPN5 cDNA levels for each treatment compared to control was determined.

Immunocytochemistry
Permeabilized immunostaining were performed with antimyc antibodies (Thermo-Scientific) as described [12,25]. Fluorescence images of the neurons were acquired using the same exposure time and analyzed with ImageJ to compare their background-subtracted fluorescence intensities.

Statistical analyses
Using Origin 9.1 (Origin Lab), the Student's t test and one-way ANOVA with Tukey's and Fisher's multiple comparison tests were performed to identify the statistically significant difference with a priori value (p) < 0.05 between 2 groups and for >3 groups, respectively.

Additional file
Additional file 1: Figure S1. Membrane-permeable TAT-STEP WT or C/ S proteins alter STEP 61 -dependent Tyr-phosphorylation of GluN2B and GluA2. (A) Permeabilized immunostaining of cultured hippocampal neurons at 12 days in vitro that were incubated for 30 min with no fusion proteins (None), TAT-myc, TAT-STEP WT, or TAT-STEP C/S. Scale bars are 20 μm. (B) Background subtracted, mean intensity of myc fluorescence (n = 10-19 images per treatment). AU, arbitrary unit. (C-D) Quantification of the levels of Tyr 1472 -phosphorylated GluN2B (GluN2B-pY 1472 ) and the level of GluA2 that were phosphorylated at Tyr 869 , Tyr 873 , and Tyr 876 (GluA2-p3Y) in CTL-treated neurons (from Fig. 4b-e) that were incubated with TAT-fusion proteins for 30 min. (C) TAT-STEP WT decreases basal Tyr-phosphorylation of GluN2B and GluA2, confirming that TAT-STEP WT increases STEP activity. (D) TAT-STEP C/S increases basal Tyr-phosphorylation of GluN2B and GluA2, confirming its ability to block dephosphorylation of STEP substrates. Data shown represent the mean ± SEM (*p < 0.05; **p < 0.01). (PDF 969 kb)