Phosphorylation and Assembly of Glutamate Receptors After Brain Ischemia
Background and Purpose—Overassembly of synaptic glutamate receptors leads to excitotoxicity. The goal of this study is to investigate phosphorylation and assembly of α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid and N-methyl-D-aspartate receptors after brain ischemia with reperfusion (I/R).
Methods—Rats were subjected to 15 minutes of global ischemia followed by 0.5, 4, and 24 hours of reperfusion. Phosphotyrosine peptides of glutamate receptors in synaptosomal fraction after I/R were identified and quantified by state-of-the-art immuno-affinity purification of phosphotyrosine peptides followed by liquid chromatography/mass spectrometry/mass spectrometry analysis (immunoaffinity purification-coupled liquid chromatography/mass spectrometry/mass spectrometry). Glutamate receptor phosphorylation and synaptic assembly after I/R were studied by biochemical methods.
Results—Numerous phosphotyrosine-sites of α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid and N-methyl-D-aspartate were upregulated by approximately 2- to 37-fold after I/R. A core glutamate receptor kinase, Src kinase, was significantly activated. GluR2/3 and NR2A/B were rapidly clustered from extrasynaptic to synaptic membrane fractions after I/R. GluR2/3 was then translocated into the intracellular pool, whereas NR2A/B remained in the synaptic fraction for as long as 24 hours. Consistently, trafficking-related phosphorylation of GluR2/3-S880 was significantly but transiently upregulated, whereas NR2A/B-Y1246 and NR2A/B-Y1472 were significantly and persistently upregulated after I/R.
Conclusions—Phosphorylation of glutamate receptors at synapses may lead to overassembly of glutamate receptors, probably via activation of Src family kinases, after I/R. This study provides global proteomic information about glutamate receptor tyrosine phosphorylation after brain ischemia.
- brain ischemia
- glutamate receptor
- mass spectrometry
- Src family kinases
- tyrosine phosphorylation
An episode of complete ischemia with reperfusion (I/R) or incomplete ischemia in the penumbral area after focal ischemia produces minimal apparent initial damage and does not result in acute neuronal death. Rather, these milder insults are associated with profound alterations in synaptic ultrastructure, molecular organization, and neurological dysfunction.1,2 Ultimately, functional recovery after brain ischemia requires precise synaptic communication and is probably the most important issue for stroke patients. For that reason, it may be important to investigate the effects of ischemia on synaptic glutamate receptor tyrosine phosphorylation, even in the brain regions that do not show any obvious lesion under the light microscopy. There are several possible rationales of such a study. (1) Brain ischemia often leads to neuronal death in selective brain regions. (2) Our previous studies have shown that there are profound and long-lasting synaptic morphological and molecular composition changes in these brain areas where overt neuronal damage does not occur after ischemia.1,2 (3) Functional changes such as language representations, mood, and memory after ischemia may be correlated with synaptic connections in the brain area where neuronal death may not necessarily occur after ischemia.3 (4) The most vulnerable area where neuronal death occurs, such as CA1, after I/R may lose synaptic connections to the projecting cortical areas. (5) Postischemic neurons may modify existing synapses or form new synapses via changes in tyrosine phosphorylation-mediated signaling pathways for adapting ischemic brain injury.4 Therefore, studying of growth-related tyrosine phosphorylation of synaptic proteins after ischemia may provide more information about functional changes and neuroplasticity after brain ischemia.
Ionotropic glutamate receptors include N-methyl-D-aspartate (NMDA) receptors (NR1, NR2A-2D) and α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid (AMPA) receptors (GluR1-4). Both NMDA and AMPA receptors are not only located in the postsynaptic density (PSD)-associated membrane; a significant proportion of them are distributed to the perisynaptic membranes and intracellular light membranes, known as extrasynaptic receptor and intracellular receptor pools, respectively. In response to changes in synaptic activities, NMDA and AMPA receptors are translocated rapidly among different pools under physiological conditions.5–7 Src family kinases have been suggested to act as a core kinases to integrate synaptic signalings to synaptic activities.8–11
In the present study, we identified numerous phosphotyrosine (Ptyr) sites of ionotropic glutamate receptors by a state-of-the-art immunoaffinity purification–coupled liquid chromatography/mass spectrometry/mass spectrometry (IAP-LC-MS/MS) analysis. Most of them were significantly upregulated as high as 37-fold after I/R. We also found that AMPA and NMDA receptors were significantly clustered at synapse during the reperfusion phase. The results provide new insights into phosphorylation and synaptic assembly of glutamate receptors after I/R.
All procedures were approved by the Animal Care and Use Committee at the University of Maryland. Brain ischemia followed by reperfusion was produced using the 2-vessel occlusion model in Wistar rats as described before.1 Blood gases were measured and adjusted to Pao2>90 mm Hg, Paco2 35 to 45 mm Hg, and pH 7.35 to 7.45 by changing the tidal volume of the respirator, and brain temperature was maintained at 37°C before, during, and after ischemia. A sham-operated control group and groups of 15 minutes of ischemia followed by 0.5, 4, and 24 hours of reperfusion were used as described before.1 For histopathology, acid fuchsin and celestine blue–stained 8-μm brain sections were examined by light microscopy. For biochemical analysis, brains were frozen in situ with liquid nitrogen according to the method of Poten et al.11 Frozen brains were sliced into 1- to 2- mm coronal sections. The dorsal cortical areas above the rhinal fissure and between bregma +1.5 mm and bregma −2.5 mm were dissected in a glove freezer box at −12°C. The cortical tissues was cut into very small pieces and mixed thoroughly to avoid uneven sampling among different experimental groups.
Preparation of Synaptosomal Fraction
The synaptosomal fraction was isolated according to the procedure described previously.1 Briefly, cortical tissue homogenate was centrifuged at 10 000g to obtain a pellet fraction. The pellet fraction was loaded onto a sucrose density gradient of 0.85 mol/L and 1.0 mol/L and 1.2 mol/L and centrifuged at 82 500g for 2 hours at 4°C. The synaptosomal fraction was collected from the 1.0 mol/L to 1.2 mol/L sucrose interface. After washing with 1% Triton X-100, the synaptosomal pellet and synaptosomal supernatant fractions were collected by centrifugation. The synaptosomal pellet was dissolved and denatured in 8 mol/L urea lysis buffer (20 mmol/L Hepes, 8 mol/L urea, 1 mmol/L sodium vanadate, 2.5 mmol/L sodium pyrophosphate, 1 mmol/L b-glycerol-phosphate), reduced by adding dithiothreitol (4.5 mmol/L), alkylated by adding iodoacetamide (8.3 mmol/L), and digested by trypsin. Protein digested was purified over C18 Sep-pak cartridge (Waters). Eluted peptides were lyophilized and stored at −80°C.
IAP-LC-MS/MS Analysis of Ptyr Peptides
IAP of phosphotyrosine peptides was performed according to the method of Rush et al12 using Phospho-tyrosine PhosphoScan Kit (Cell Signaling Technology, Danvers, MA). LC-MS/MS analysis was performed with an LTQ Orbitrap Mass Spectrometer (Thermo Fisher Scientific), and a peptide mass accuracy of ±3 ppm was 1 of the filters used for peptide identification. Details were described previously.13 Sequest (Thermo Fisher Scientific) searches were performed against the NCBI rat database, allowing for tyrosine phosphorylation (Y+80) and oxidized methionine (M+16) as differential modifications. The PeptideProphet probability threshold was chosen to give a false-positive rate of 5% for the peptide identifications.14
Quantitative MS/MS Analysis of Ptyr Peptides
To quantify Ptyr peptides, the intensity value of the peak apex for each peptide was determined for each individual LC-MS/MS run and compared against other LC-MS/MS runs.13 For LC-MS/MS runs in which the peptide was not identified by MS/MS, peptide mass-to-charge ratio (m/z) and retention time (RT) values were used to search for peaks in the MS1 channel, which would correspond to the peptide ion. Similar to several previous studies,13 approximately 30% to 40% peptide intensities were derived from MS1 channel. To reduce the selection of the incorrect peak when an MS/MS was not triggered, we used several criteria. For example, in cases in which an MS/MS was not triggered for a peptide m/z, we used peptide m/z values in common between runs to define RT differences by simple linear regression. Once RT drift was defined, we required the m/z value to be present in a narrow RT window of 2 minutes. Within that RT window, we required a mass tolerance of 10 ppm. Peaks identified within the RT window and mass tolerance were further scrutinized for the correct charge state by requiring at least 2 C13-containing isotopes along with the monoisotopic m/z. Finally, only 1 peak was allowed during passing each of these criteria. If an identified peak passed each of these criteria, then its intensity at the peak’s apex was extracted and used for quantification.
Preparation of Synaptic, Extrasynaptic, and Intracellular Light Membrane Fractions
Neocortical tissues were homogenized with a Dounce homogenizer (50 strokes) in a homo-buffer by the method described previously.1,15 The homogenate was centrifuged at 10 000g for 10 minutes at 4°C to obtain a pellet and a supernatant S2. The pellet was washed with the homo-buffer containing 2% Triton X-100 and 500 mmol/L KCl, then centrifuged at 18 000g for 20 minutes. The pellet, referred to as P(1+2)p or synaptic fraction (SYN), containing detergent-insoluble and salt-insoluble proteins tightly associated to PSDs.16 The supernatant was referred to as P(1+2)s or extrasynaptic fraction (EXTRA), containing detergent-soluble and salt-soluble proteins not associated with PSDs.16 The S2 fraction was further centrifuged at 165 000g for 60 minutes at 4°C to obtain a pellet or intracellular light membrane fraction (P3 or light membrane [LM]) and cytosol (S3 or CYTO). Protein concentration was determined by the microbicinchonic acid method (Pierce Biochemicals, Rockford, IL).
Western Blot Analysis
Western blot analysis was performed with 8% SDS-PAGE. Equal amounts of samples containing 30 µg protein from homogenate, P(1+2)s, S3 fractions, and 50 µg from P(1+2)p and P3 fractions were subjected to SDS-PAGE and then transferred to Immobilon-P membranes (Millipore). After incubation with primary antibodies and then horseradish peroxidase-conjugated secondary antibodies, blots were developed with an ECL detection system (Amersham, Piscataway, NJ). The films were scanned and the optical densities of the protein bands were quantified using NIH ImageJ software. One-way ANOVA followed by Dunnett post hoc test were used to assess statistical significance (P<0.05 and P<0.01).
No morphological changes in brain sections were seen under the light microscopy between sham-operated control rats and rats subjected to 15 minutes of ischemia followed by 4 hours of reperfusion. Delayed neuronal death occurred after 48 hours of reperfusion selectively in dorsal CA1 neurons. Only a few, if any, neocortical neurons were damaged at 7 days of reperfusion (Supplementary Figure IB, arrows). All these results are consistent with previous reports.17,18 The objective of this study was to investigate synaptic NMDA and AMPA receptor assembly in brain tissues without apparent pathological damage. Therefore, neocortical tissues from sham control and 4 hours of reperfusion (I/R-4 hours) were used in the IAP-LC-MS/MS analysis.
IAP-LC/MS/MS Analysis of Synaptic NMDA Receptor and AMPA Receptor Ptyr Motifs
All Ptyr peptide sequences presented in this study were reproducibly captured at least 3 times. As demonstrated in recent studies, the IAP-LC-MS/MS analysis is highly specific for detecting Ptyr peptides.12,13 Figure 1 shows representative MS spectra of GluR2 and NR2, respectively. Western blot validations revealed that phosphorylations of GluR2/3-Y867 and NR2B-Y1246 were significantly reduced in brain tissue because of ATP depletion during ischemia and were markedly increased at 4 hours of reperfusion. Several Ptyr sites at C-termini of GluR2/3 were identified (Table and Supplementary Table II). In addition to the previously known GluR2-Y876 sites, GluR3-Y871/881/869/Y873/Y874 sites were newly identified. Similarly, NR2A Ptyr-1184/1187/1246/1267 and NR2B Ptyr-1070/1155/1225/1304 sites were newly identified and they were mostly increased at 4 hours of reperfusion (Table and Supplementary Table II). Figure 2 shows the quantities of those Ptyr sites with >5-fold increase after ischemia.
Activation of Synaptic Src Kinase After I/R
Activation of Src kinase requires phosphorylation of the active loop Y416 and dephosphorylation of the inhibitory site Y527.10 The level of phospho-Src Y416 was significantly upregulated at 30 minutes and 4 hours, and then returned to the control level at 24 hours of reperfusion (Figure 3A). In contrast, phospho-Src Y527 was significantly downregulated at 30 minutes and 4 hours, and then returned to the control level at 24 hours of reperfusion (Figure 3B). Upregulation of the activation site Y416 and downregulation of the inhibitory site Y527 were observed only in Triton X-100–insoluble SYN, whereas no changes were detected in Triton X-100–soluble EXTRA (Figure 3). The data showed that Src was significantly activated only in the synaptic fraction after I/R.
Synaptic Glutamate Receptor Clustering
To understand glutamate receptor clustering after I/R, we prepared SYN, EXTRA, and intracellular LM fractions (Supplementary Figure IIIA). β-actin, as a protein-loading control, was unchanged in all subcellular factions after I/R (Supplementary Figure IV). PSD93 and PSD95 are PSD proteins and thus are located mainly in the synaptic fraction (SYN) and, to a much lesser degree, in EXTRA and LM fractions. Clathrin and Rab4 are early endosome markers. They are mainly located in the LM fraction and, to a much lesser degree, also in the cytosol (Supplementary Figure IIIB).
GluR2/3 was significantly increased in SYN and concomitantly decreased in EXTRA and LM fractions at 30 minutes of reperfusion. GluR2/3 was then reduced to below the control levels in SYN and EXTRA fractions but recovered in LM (intracellular pool) fraction (Figure 4A). The data suggested that GluR2/3 was translocated from EXTRA and intracellular LM pools to synapses initially, and then returned to the intracellular pools, probably via the endocytic pathway.19–21
Phosphorylation of GluR2-S880, NR2A-Y1246, and NR2B-Y1472
To verify further the proteomic MS data and synaptic clustering of glutamate receptors shown, we studied all subcellular fractions with antibodies to GluR2-S880, NR2A-Y1246, and NR2B-Y1472. These sites are related to glutamate receptor trafficking.19–25 The Western blot data showed that the levels of GluR2-S880, NR2A-Y1246, and NR2B-Y1472 were significantly increased at 0.5 and 4 hours of reperfusion in the SYN fractions after I/R (Figure 5). In the perisynaptic (EXTRA) fraction, GluR2-S880, NR2A-Y1246, and NR2B-Y1472 were initially decreased at 0.5 hours, transiently returned to or above the control level at 4 hours of reperfusion, and then decreased again at 24 hours of reperfusion (Figure 5). These sites were not detectable in the intracellular LM fraction after I/R (data not shown). These results suggested that the glutamate receptor trafficking–related phosphorylation sites were highly upregulated in the synaptic fraction after I/R.
The present study showed that I/R led to upregulation of phosphorylated glutamate receptors at multiple Ptyr sites (Table, Supplemental II). Many novel Ptyr sites were identified in this study. Consistently, a core glutamate receptor kinase, Src kinase, was activated after I/R. GluR2/3, as well as NR2A/B, was rapidly clustered from extrasynaptic to synaptic membranes after I/R. Synaptic GluR2/3 was then translocated into the intracellular LM pool, whereas NR2A/B remained in the synaptic fraction for as long as 24 hours of reperfusion after I/R. Consistently, GluR2/3 trafficking related phosphorylation site S880 was significantly but transiently upregulated after I/R, whereas NR2A and NR2B trafficking–related phosphorylation sites Y1246 and Y1472 were significantly and persistently upregulated after I/R. The results suggest that tyrosine phosphorylation of glutamate receptors at synapses may lead to overassembly of glutamate receptors after I/R. This study provides global proteomic information about ionotropic glutamate receptor tyrosine phosphorylation after brain ischemia.
Identification of Glutamate Receptor Ptyr Sites by IAP-LC/MS/MS
Although most neocortical neurons will not die after a brief episode of brain ischemia, there are profound changes in ultrastructure and molecular composition at the synapses.1,2,15 In collaboration with Cell Signaling Technology (Danvers, MA), this study used state-of-the-art IAP-LC/MS/MS technology to study the overall glutamate receptor tyrosine phosphorylation status in the synaptic fractions after I/R. The selection of postischemic synaptic fractions is based on our previous observation that tyrosine phosphorylation in synaptic fractions is the most dramatic as well as the most selective among different subcellular fractions after brain ischemia.1,15 A key reason may be that a brief episode of global ischemia induces overall depolarization of synapses, followed by repolarization on reperfusion in the entire forebrain.18 Therefore, unlike physiological stimuli that activate a few or a group of neuronal networks, an episode of ischemia followed by reperfusion leads to overall activation of all forebrain synapses on a global scale.18 This may explain why changes in the Ptyr number and degree are so dramatic because all currently known Ptyr proteins are detected in postischemic synaptic fraction by the IAP-LC/MS/MS analysis. This is further supported by the fact that most, if not all, glutamate receptor–related synaptic events can be detected after I/R, including neurotransmitter release, induction of Long-term potentiation, translocation of protein kinases, activation of the neurotrophin/receptor pathways, and activation of extracellular signal–regulated kinases, and others.1,15 This IAP-LC-MS/MS analysis of Ptyr sites has tremendous advantages over more traditional methods in which a single protein or motif is usually characterized.26 Existing evidence proves that the IAP-LC-MS/MS analysis is highly efficient and specific for detecting Ptyr sites.12,13
It should be pointed out that the synaptosomal fraction we used in this study contains mainly synaptic structures from neurons, as demonstrated by the morphological assessment. However, this fraction also might contain small amount of membranes from non-neuronal cells. Therefore, the changes in glutamate receptor tyrosine phosphorylation is highly likely to be mainly neuronal, but also may occur in neuronal-to-astrocyte synapses after brain ischemia.27
Redistribution of Glutamate Receptors After I/R
AMPA and NMDA receptors are present in SYN, EXTRA, and intracellular pools of neurons.5–7 AMPA receptors are highly mobile, rapidly shuttling between these different pools. This process is regulated by phosphorylation of S880.19,23,25 Similarly, the synaptic translocation of NMDA receptors also depends on phosphorylation of NR2A-Y1246 and NR2B-Y1472. By using detergent extraction and different receptor pool separation, this study analyzed glutamate receptor redistribution and showed that both GluR2/3 and NR2A/B were translocated to the synaptic fraction after I/R (Figures 4 and 5). These data are consistent with previous studies showing that synaptic localization of AMPA and NMDA receptors are tightly regulated by phosphorylation and synaptic activities.5,19–25
Previous studies have shown that tyrosine phosphorylation of glutamate receptors by Src family kinases enhances their synaptic localization.28 Inhibition of Src kinase protects brain from ischemic injury.29–31 This is consistent with the fact that Src family kinases are responsible for GluR2/3 and NR2A/B synaptic clustering. In fact, many ischemia-upregulated phosphorylation sites of glutamate receptors listed in the Table (also in Supplementary Table II) are Src kinase substrates. The evidence is also in agreement with the significant upregulation of Src active loop Y416 and downregulation of Src inhibitory domain Y527 in a time course similar to that of NR2A/B tyrosine phosphorylation after I/R (Figure 3).
In summary, we have identified numerous Ptyr sites of ionotropic glutamate receptors by a state-of-the-art IAP-LC-MS/MS method. Most of them are Src kinase substrates and are dramatically upregulated after I/R. Such upregulation is in accordance with GluR2/3 and NR2A/B synaptic clustering after I/R. These alterations provide an explanation of the synaptic dysfunction after brain ischemia.
Source of Funding
This work was supported by National Institutes of Health grants NS36810 and NS030291, and by American Heart Association grant EIA 090042N.
The online-only Data Supplement is available with this article at http://stroke.ahajournals.org/lookup/suppl/doi:10.1161/STROKEAHA.112.667253/-/DC1.
- Received June 29, 2012.
- Revision received September 26, 2012.
- Accepted October 17, 2012.
- © 2012 American Heart Association, Inc.
- Hu BR,
- Park M,
- Martone ME,
- Fischer WH,
- Ellisman MH,
- Zivin JA
- Martone ME,
- Jones YZ,
- Young SJ,
- Ellisman MH,
- Zivin JA,
- Hu BR
- Carmichael ST
- Overman JJ,
- Clarkson AN,
- Wanner IB,
- Overman WT,
- Eckstein I,
- Maguire JL,
- et al
- Sinai L,
- Duffy S,
- Roder JC
- Hu BR,
- Martone ME,
- Jones YZ,
- Liu CL
- Fu J,
- deSouza S,
- Ziff EB
- Seidenman KJ,
- Steinberg JP,
- Huganir R,
- Malinow R
- Nada S,
- Shima T,
- Yanai H,
- Husi H,
- Grant SG,
- Okada M,
- et al