PSD (postsynaptic density)-95, a scaffold protein that tethers NMDA (N-methyl-D-aspartate) receptors to signal molecules, is implicated in pathological events resulting from excitotoxicity. The present study demonstrates that brain ischaemia and reperfusion increase the tyrosine phosphorylation of PSD-95 in the rat hippocampus. PP2, a specific inhibitor of SrcPTKs (Src family protein tyrosine kinases), prevents the ischaemia-induced increases not only in the tyrosine phosphorylation of PSD-95, but also in the interaction between PSD-95 and Src kinases. PSD-95 is phosphorylated either by purified Src/Fyn kinases in vitro or by co-expression of constitutively active Src/Fyn in COS7 cells. The results suggest that SrcPTKs are involved in PSD-95 phosphorylation. The single Tyr523 mutation to phenylalanine (Y523F) reduces the Src/Fyn-mediated phosphorylation of PSD-95 in COS7 cells and in vitro. As shown with a rabbit polyclonal antibody against phospho-PSD-95 (Tyr523), Tyr523 phosphorylation is responsible for the increased tyrosine phosphorylation of PSD-95 induced by ischaemia in the rat hippocampus. In cultured hippocampal neurons, overexpression of PSD-95 Y523F, but not PSD-95 Y533F, abolishes the facilitating effect of PSD-95 on the glutamate- or NMDA-mediated currents, implying that PSD-95 Tyr523 phosphorylation contributes to the post-ischaemic over-activation of NMDA receptors. Thus the present study reveals an additional mechanism for the regulation of PSD-95 by tyrosine phosphorylation. This mechanism may be of pathological significance since it is associated with excitotoxicity in the ischaemic brain.
PSD (postsynaptic density)-95, also referred to as SAP (synapse-associated protein)-90, belongs to the PSD-95 subfamily of MAGUKs (membrane-associated guanylate kinases) and serves as a scaffold protein in excitatory synapses of the brain [1,2]. All PSD-95 subfamily members, including PSD-95, PSD-93, SAP-97 and SAP-102, share an organized domain structure, consisting of three N-terminal PDZ domains, an SH (Src homology) 3 domain, and a C-terminal GK (guanylate kinase) domain. With these distinct structural domains, PSD-95 subfamily proteins bind to the membrane-associated excitatory glutamate receptors, as well as to various downstream signal molecules, including CaMKII (Ca2+/calmodulin-dependent protein kinase II) [3,4], SrcPTKs (Src family protein tyrosine kinases) [5,6] and Pyk2 (proline-rich tyrosine kinase 2) [7,8].
The NMDA (N-methyl-D-aspartate) subtype of glutamate receptors co-localizes with PSD-95 in the excitatory PSD . Functional NMDA receptor channels are hetero-oligomeric protein complexes composed of necessary NMDA receptor subunit 1, regionally localized NMDA receptor subunit 2 (A–D) and, more rarely, NMDA receptor subunit 3 (A or B) [10,11]. In the hippocampus, NMDA receptor subunits 2A and 2B are the major subtypes of NMDA receptor subunit 2 and determine the functional properties of NMDA receptors [12,13]. PSD-95 associates with the C-terminal of NMDA receptor subunits 2A and 2B by the PDZ1/2 domains [9,14]. As an organizer of the excitatory synapse, PSD-95 is involved in the NMDA-receptor-dependent excitatory synaptic transmission and synaptic plasticity [1,15,16]. At the same time, PSD-95 has been implicated in the pathological events resulting from brain ischaemia and excitotoxicity mediated by over-activation of NMDA receptors [17–19].
SrcPTKs, the largest family of membrane-associated non-receptor tyrosine kinases, interact with the PDZ3 domain of PSD-95 via its SH2 domain . There are five members of SrcPTKs (Src, Fyn, Yes, Lck and Lyn) in the mammalian central nervous system. Both Src and Fyn possess high enzyme activity and are highly expressed in the brain . PSD-95 is a substrate for the intracellular serine/threonine kinases CaMKII [3,4], cyclin-dependent kinases , p38γ and ERK (extracellular-signal-regulated kinase) 1/2 . Although PSD-95 binds to SrcPTKs, there is little information related to its tyrosine phosphorylation.
Previous work by our laboratory has indicated that brain ischaemia and reperfusion enhance the interactions between PSD-95 and Src/Fyn kinases [23,24]. In the present study we found that, in the rat hippocampus, PSD-95 is tyrosine-phosphorylated after brain ischaemia, suggesting a pathological role of such phosphorylation. This process is mediated by SrcPTKs in vivo and in vitro. The mutation of Tyr523 in PSD-95 not only decreases Src/Fyn-dependent phosphorylation of PSD-95 in COS7 cells and in vitro, but also eliminates the facilitating effect of PSD-95 on glutamate/NMDA-mediated currents in rat hippocampal neurons, implicating the tyrosine phosphorylation of PSD-95 in the up-regulation of NMDA receptors. These results contribute to the elucidation of the molecular mechanisms underlying ischaemic excitotoxicity.
Mouse monoclonal anti-PSD-95 antibody (clone 7E3-1B8), mouse monoclonal anti-pY (phosphotyrosine) antibody (clone PT-66), rabbit polyclonal anti-c-Myc antibody, alkaline-phosphatase-conjugated secondary antibodies and non-specific mouse IgG and rabbit IgG were from Sigma. Rabbit polyclonal anti-Src antibody, mouse monoclonal anti-Fyn (Fyn15) antibody and rabbit polyclonal anti-actin antibody were from Santa Cruz Biotechnology. Rabbit polyclonal anti-[phospho-PSD-95 (Tyr523)] antibody was raised in New Zealand White rabbits with a KLH (keyhole-limpet haemocyanin)-conjugated pY-containing synthetic peptide (REDSVLSpYETVT) as the immunogen, and the antibody was purified by two-step immunoaffinity chromatography using a column conjugated to the non-phosphopeptide to eliminate antibodies against non-phosphorylated PSD-95, followed by purification with another column conjugated with the immunogen (AbMART).
The vector pcDNA3.1 (+) (Invitrogen) was used for the expression of Src, Fyn, PSD-95 and various mutant constructs. SrcY529F- and SrcK297R-pcDNA3.1, encoding constitutively active and inactive forms of Src, were subcloned from SrcY529F- and SrcK297R-pUSEamp (Upstate Biotechnology) respectively. The cDNAs of Fyn Y531F and Fyn K299M, encoding constitutively active and inactive forms of Fyn respectively, were amplified by RT-PCR (reverse transcription-PCR) and overlap extension. The cDNAs of Myc–ΔSG (PSD-95 with SH3 and GK domains deleted) and Myc–ΔG (PSD-95 with GK domain deleted), encoding SH3/GK domain-deletion mutants of PSD-95, were PCR amplified from Myc–PSD-95-GW1 (a gift from Dr M. Sheng, Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, MA, U.S.A.). The cDNAs of full-length PSD-95 and PSD-95 single tyrosine mutants Y432F, Y439F, Y523F, Y533F, Y573F, Y580F, Y604F, Y609F, Y701F and Y715F were amplified by RT-PCR and overlap extension respectively. All constructs were verified by DNA sequencing.
Brain ischaemia and drug treatments
Brain ischaemia was induced by the four-vessel occlusion method as described previously . Briefly, adult male Sprague–Dawley rats (220–270 g) were anaesthetized with chloral hydrate (300 mg/kg, intraperitoneal) and vertebral arteries were electrically coagulated. On the following day, ischaemia was induced for various time periods by the occlusion of both common carotid arteries with aneurysm clips. Animals not meeting the following criteria were excluded: completely flat bitemporal EEG (electroence-phalogram), maintenance of dilated pupils or absence of a corneal reflex. Rat rectal temperature was maintained at 37–37.5 °C during ischaemia and 2 h of reperfusion. Sham-operated rats received the same treatments, except for occlusion of the carotid artery. PP2 or PP3 (3 μg/μl in DMSO; Calbiochem) was infused into the rat cerebral ventricle (from the bregma: anteroposterior, −0.8 mm; lateral, 1.5 mm; depth, 3.5 mm) through a step-motorized microsyringe (Stoelting) 30 min before occlusion. All experiments were performed in accordance with the guidelines of the local Animal Care Committee.
Rats were decapitated at indicated times after ischaemia, and the hippocampi were removed and quickly frozen in liquid nitrogen. Samples were homogenized in ice-cold homogenization buffer [50 mM Mops (pH 7.4), 320 mM sucrose, 100 mM KCl, 0.5 mM MgCl2, 0.2 mM dithiothreitol and phosphatase and protease inhibitors (20 mM β-glycerophosphate, 20 mM sodium pyrophosphate, 50 mM NaF, 1 mM each of EDTA, EGTA, sodium orthovanadate, p-nitrophenyl phosphate, PMSF and benzamidine, and 5 μg/ml each of aprotinin, leupeptin and pepstatin A)]. The homogenates were centrifuged at 800 g for 10 min at 4 °C, and the supernatants were collected. Protein concentrations were determined by the method of Lowry [24a]. Samples were stored at −80 °C until use.
The protein samples were diluted 4-fold with immunoprecipitation buffer containing 50 mM Hepes (pH 7.1), 10% (v/v) glycerol, 150 mM NaCl, 1 mM ZnCl2, 1.5 mM MgCl2, 1% Triton X-100, 0.5% NP-40 (Nonidet P40) and protease and phosphatase inhibitors as described above. The samples were incubated with 1–2 μg of the indicated antibodies for 4 h or overnight at 4 °C. Protein A–Sepharose (25 μl) was added and the incubation was allowed to continue for 2 h at 4 °C. Immunoprecipitates were collected by centrifugation at 10000 g for 2 min at 4 °C, and the bound proteins were eluted from Protein A–Sepharose by boiling for 5 min in Laemmli sample buffer [62.5 mM Tris/HCl (pH 6.8), 2% (w/v) SDS, 1% 2-mercaptoethanol, 10% (v/v) glycerol and 0.003% Bromophenol Blue]. After centrifugation at 10000 g for 1 min, the supernatants were used for immunoblotting or for kinase assays in vitro.
As described previously , the immunoprecipitates or equal amounts of protein samples were separated by SDS/PAGE and then transferred on to nitrocellulose membranes (Amersham Pharmacia). After blocking with 3% (w/v) BSA for 3 h, the membranes were incubated with primary antibodies or alkaline-phosphatase-conjugated secondary antibodies and developed with the colour substrate, BCIP (5-bromo-4-chloroindol-3-yl phosphate) and NBT (Nitro Blue Tetrazolium) (Promega) following the manufacturer's protocol.
In vitro kinase assay
Immunoprecipitates of PSD-95 proteins were incubated with 3 units of purified Src or 5 m-units of Fyn kinase (Upstate Biotechnology) and ATP (100 μM) in kinase reaction buffer [40 mM Hepes (pH 7.4), 25 mM MgCl2, 5 mM MnCl2, 0.4 mM EGTA, 0.5 mM sodium orthovanadate and 0.4 mM dithiothreitol] for 30 min at 30 °C. The kinase reaction was initiated by the addition of ATP and was stopped by the addition of Laemmli sample buffer. After boiling for 5 min, the reaction mixtures were centrifuged at 10000 g for 1 min and the supernatants were subjected to immunoblot analysis with an anti-pY antibody to visualize the tyrosine phosphorylation of PSD-95.
Hippocampal synaptosome preparation
Rat hippocampal samples were homogenized in ice-cold homogenization buffer (see above) and were centrifuged at 800 g for 15 min at 4 °C. The collected supernatants were centrifuged at 14000 g for 15 min at 4 °C. The pellets were lysed in ice-cold homogenization buffer and then subjected to immunoprecipitation and kinase assays in vitro.
COS7 cell culture and plasmid transfection
COS7 cells were cultured in DMEM (Dulbecco's modified Eagle's medium) (Gibco/Invitrogen) supplemented with 10% (v/v) fetal calf serum. Recombinant plasmids based on the pcDNA3.1 vector were transfected into cells at 90–95% confluency with Lipofectamine™ 2000 (Invitrogen) following the manufacturer's protocol. After 48 h of transfection, the cultures were lysed in ice-cold homogenization buffer (see above) followed by ultrasonic treatment. The samples were stored in liquid nitrogen and were used later for immunoprecipitation, immunoblotting or kinase analysis in vitro.
Primary hippocampal neuron culture
Hippocampal neurons were isolated from the hippocampi of embryonic-day-18 rats as described previously . Dissociated neurons were plated at a density of 0.3×105 cells/cm2 on cover-slips coated with poly-L-lysine in DMEM supplemented with 10% (v/v) fetal calf serum and 10% (v/v) horse serum. After 24 h, the medium was replaced with Neurobasal serum-free medium supplemented with 2% (v/v) B27 (Gibco/Invitrogen), 100 units/ml penicillin, 100 units/ml streptomycin and 0.5 mM L-glutamine (Sigma). Subsequently, the culture medium was replaced every 3 days. At 7–8 days in vitro, the neurons were transfected, by use of Lipofectamine™ 2000, with EGFP (enhanced green fluorescent protein)-tagged full-length PSD-95, PSD-95 Y523F, PSD-95 Y533F or EGFP alone. After 24 h of transfection, the neurons were subjected to electrophysiological studies.
Whole-cell patch-clamp recording was performed in voltage-clamp mode using an Axon-700B amplifier (Axon Instruments) at a membrane voltage of −60 mV. The pipette had a resistance of 3–4 MΩ when filled with the intracellular solution [10 mM Hepes (pH 7.2), 120mM CsCl, 20 mM triethylamine, 1 mM MgCl2, 10 mM EGTA and 2 mM ATP. Osmolarity was adjusted to 290–310 mOsmol/l with sucrose]. The recording chamber was perfused with a standard extracellular solution [10 mM Hepes (pH 7.4), 150 mM NaCl, 5 mM KCl, 2 mM CaCl2 and 10 mM glucose, 310–330 mOsmol/l]. The whole-cell currents were examined in the presence of glutamate (50 μM, with 5 μM glycine) or NMDA (100 μM, with 5 μM glycine). The results were presented as current densities (pA/pF), defined as currents divided by capacitances. All results are means±S.E.M. (n=9–16), with the statistical significance determined by ANOVA.
Tyrosine phosphorylation of PSD-95 after brain ischaemia
Since brain ischaemia promotes the PSD-95–Src/Fyn interactions [23,24], the tyrosine phosphorylation of PSD-95 was measured after transient (15 min) brain ischaemia in the rat hippocampus, one of the most vulnerable brain regions. As demonstrated by immunoprecipitation with a mouse anti-PSD-95 antibody followed by immunoblotting with a mouse anti-pY antibody, or by immunoprecipitation with an anti-pY antibody and then immunoblotting with an anti-PSD-95 antibody, a 95 kDa protein band was observed, the amount of which was greater in the ischaemia groups when compared with the sham-operated groups (Figure 1a). In contrast, immunoprecipitation with a non-specific mouse IgG resulted in no bands observed at 95 kDa. Immunoblot analysis showed no difference in the expression level of PSD-95 between the ischaemia and sham-operated groups (Figure 1a).
Brain ischaemia and reperfusion increase the tyrosine phosphorylation of PSD-95 in the rat hippocampus
The tyrosine phosphorylation of PSD-95 increased consistently during 5, 15 and 30 min of brain ischaemia (Figure 1b1). There was a decrease after reperfusion following 15 min of ischaemia, then tyrosine phosphorylation of PSD-95 increased again to a maximum at 6 h, and remained evident at 24 h reperfusion (Figure 1b2).
To clarify the role of SrcPTKs in the tyrosine phosphorylation of PSD-95 after brain ischaemia, a selective SrcPTK inhibitor, PP2, was administered to rats 30 min before ischaemia. PP3, an inactive derivative of PP2, was used as a negative control. Pre-treatment with PP2, but not with the same dose of PP3 or with vehicle (DMSO), attenuated the increased tyrosine phosphorylation of PSD-95 and the enhanced interaction between PSD-95 and Src. These results suggest that SrcPTKs are involved in ischaemia-induced tyrosine phosphorylation of PSD-95 (Figure 1c).
Src/Fyn-mediated phosphorylation of PSD-95 in vitro and in COS7 cells
The anti-PSD-95 antibody (clone 7E3-1B8) slightly cross-reacts with PSD-93, a member of the PSD-95 subfamily . PSD-93 has been reported to be a substrate for Fyn kinases . Therefore active Src or Fyn kinases were used in vitro and in COS7 cells to verify the tyrosine phosphorylation of PSD-95. For the in vitro kinase assay, PSD-95 proteins immunoprecipitated from the hippocampal synaptosome fractions or COS7 cells transfected with PSD-95-pcDNA3.1 were used as the substrates. The results showed that PSD-95 proteins from synaptosomes or COS7 cells were phosphorylated after incubation with purified active Src or Fyn kinases for 30 min (Figure 2a). In COS7 cells, the construct for full-length PSD-95 was co-transfected with plasmids encoding constitutively active (Src Y529F/Fyn Y531F) or inactive (Src K297R/Fyn K299M) forms of Src or Fyn. The results showed that PSD-95 was tyrosine-phosphorylated only when it was co-expressed with active Src or Fyn, but not with the inactive kinases (Figure 2b).
Src/Fyn induces tyrosine phosphorylation of PSD-95 in vitro and in COS7 cells
Identification of tyrosine phosphorylation sites in PSD-95 by Src/Fyn
To locate the phosphorylation sites in PSD-95 by Src/Fyn, recombinant proteins expressing Myc-tagged full-length PSD-95, ΔSG (amino acids 1–424 of PSD-95) and ΔG (amino acids 1–533 of PSD-95) were generated (Figure 3a). In all constructs, the PDZ3 domain was preserved for binding to Src and Fyn. In COS7 cells, phosphorylation was observed when Myc–PSD-95 and Myc–ΔG were co-expressed with constitutively active Src or Fyn kinases, but Myc–ΔSG was not phosphorylated; and the interactions between Myc–PSD-95, Myc–ΔG or Myc–ΔSG and Src/Fyn were not affected (Figure 3b). In vitro, Myc–ΔG, but not Myc–ΔSG, was phosphorylated in a Src- or Fyn-dependent manner (Figure 3c). These results suggest that the major phosphorylation sites of PSD-95 for Src and Fyn are located in amino acids 425–533. However, amino acids 534–724 of PSD-95 cannot be excluded as a location for tyrosine-phosphorylation sites because of possible multi-site phosphorylation.
Various degrees of tyrosine phosphorylation of the ΔG and ΔSG PSD-95 mutants by Src/Fyn in COS7 cells and in vitro
There are four tyrosine residues in amino acids 425–533 of PSD-95 (Figure 4a). In COS7 cells, the single tyrosine mutants of full-length PSD-95, including Y432F, Y439F, Y523F or Y533F, were co-expressed with constitutively active Src or Fyn kinases. A reduction in tyrosine phosphorylation was evident with PSD-95 Y523F compared with full-length PSD-95 (Figures 4b1 and 4b2). The single Tyr523 mutation did not alter the interactions between PSD-95 and Src or Fyn kinases (Figure 4b3). Furthermore, PSD-95 Y523F could not be efficiently phosphorylated by Src or Fyn kinases in vitro relative to PSD-95 (Figure 4c). Thus Tyr523 is the major phosphorylation site in PSD-95.
PSD-95 Tyr523 is a principal phosphorylation site for Src/Fyn in COS7 cells and in vitro
There are six tyrosine residues located in amino acids 534–724 of PSD-95 (Figure 4a). Each of the single PSD-95 tyrosine mutants, Y573F, Y580F, Y604F, Y609F, Y701F and Y715F, displayed a similar phosphorylation level to wild-type full-length PSD-95 when co-expressed with constitutively active Src or Fyn kinases (Figure 4d), suggesting that there are no phosphorylation sites in the C-terminal 200 amino acids of PSD-95.
PSD-95 Tyr523 phosphorylation in COS7 cells and rat hippocampus
To evaluate the phosphorylation of Tyr523in vivo, a rabbit polyclonal antibody which selectively recognizes Tyr523-phosphorylated PSD-95 was raised using the pY-containing synthetic peptide KLH-CREDSVLSpYETVT as the immunogen (Figure 5a).
The phosphorylation of Tyr523 is responsible for the increased tyrosine phosphorylation of PSD-95 after brain ischaemia
Immunoblot analysis with an anti-[phospho-PSD-95 (Tyr523)] antibody showed that Tyr523 was phosphorylated in COS7 cells co-transfected with PSD-95 and constitutively active Fyn, but not with inactive Fyn kinases; the mutation of Tyr523 to Phe523 eliminated the phosphorylation of PSD-95 Tyr523 (Figure 5b). The results establish the specificity of the anti-[phospho-PSD-95 (Tyr523)] antibody. In the rat hippocampus, PSD-95 Tyr523 was phosphorylated, especially in the ischaemia groups; but there was only a weak signal in the sham-operated groups (Figure 5c). These results indicate that, in the rat hippocampus, the phosphorylation of Tyr523 is responsible for the increased tyrosine phosphorylation of PSD-95 after brain ischaemia.
Effects of PSD-95 Tyr523 on glutamate/NMDA-mediated currents
PSD-95 has been implicated in the up-regulation of NMDA-receptor function, but the role of PSD-95 Tyr523 phosphorylation in the function of NMDA receptors is not clear. Whole-cell patch-clamp experiments were performed to record the glutamate- or NMDA-stimulated whole-cell currents in the cultured primary hippocampal neurons with over-expressed EGFP-tagged PSD-95, PSD-95 Y523F, PSD-95 Y533F or with EGFP alone (Figure 6a). EGFP was used for identifying the candidate cells. The capacitance is directly proportional to the surface membrane area of a cell; the amplitude of glutamate- and NMDA-induced currents, on the other hand, relates to the total number of functional NMDA receptors. Therefore the current density (defined as the current divided by the capacitance) provides an estimate of the activity of functional NMDA receptors in different neurons. The results showed that the over-expression of full-length wild-type PSD-95 led to an increase in current density stimulated by glutamate or NMDA; over-expression of the PSD-95 Y533F mutant had an effect similar to that observed for wild-type PSD-95; but PSD-95 Y523F did not affect glutamate- or NMDA-induced currents (Figures 6b and 6c). These results reveal that Tyr523 is involved in the facilitating effect of PSD-95 on the activity of glutamate receptors, including NMDA receptors.
Tyr523 is associated with PSD-95-mediated up-regulation of glutamate receptors in cultured hippocampal neurons
Excitotoxicity is implicated in the pathogenesis of neurodegenerative disorders resulting from brain ischaemia. After an ischaemic episode, excessively released glutamate causes the over-activation of glutamate receptors, mainly of the NMDA subtype, and activates intracellular calcium-dependent signal cascades that eventually lead to neuronal cell death, a process known as excitotoxicity . However, the clinical application of NMDA-receptor antagonists is limited by severe side effects. The postsynaptic molecular mechanisms associated with the over-activation of NMDA receptors may provide a potential target for the therapy of ischaemic stroke. PSD-95 is involved in the up-regulation of NMDA receptors, but the regulation of PSD-95 after brain ischaemia remains to be elucidated.
Reversible protein phosphorylation is a ubiquitous mechanism for the regulation of biological function of proteins and the control of signal transduction. Although an interaction between PSD-95–Src/Fyn is established, the regulation of PSD-95 by tyrosine phosphorylation has not been reported previously. The reason may be the result of a relatively weak signal under physiological conditions . In the post-ischaemic brain, PSD-95 is highly tyrosine-phosphorylated by SrcPTKs, suggesting a pathological association of such phosphorylation. PSD-93, another member of the PSD-95 subfamily in the hippocampal postsynapse, is a major tyrosine-phosphorylated protein in the NMDA receptor complex of the normal mouse brain . Immunoprecipitation with an anti-pY antibody, followed by blotting with a specific antibody against PSD-93 (catalogue number 34-4700; Invitrogen) showed that PSD-93 had a high basal level of tyrosine phosphorylation in the rat hippocampus; and in comparison with sham-operated animals, the tyrosine phosphorylation of PSD-93 was not changed following 15 or 30 min of brain ischaemia (results not shown). Thus the tyrosine phosphorylation of PSD-93 may correlate with its physiological function.
The present study provides evidence that Tyr523 is a major phosphorylation site in vitro and in vivo. Since mutation of Tyr523 of PSD-95 did not eliminate Src/Fyn-mediated phosphorylation, screening for other tyrosine phosphorylation sites in PSD-95 was performed. Every tyrosine mutant construct in amino acids 425–724 except for Y523F displayed a tyrosine phosphorylation level similar to that for wild-type PSD-95, suggesting that the additional tyrosine phosphorylation sites exist among amino acids 1–424. However, the ΔSG mutant, containing amino acids 1–424, could not be efficiently phosphorylated by active Src/Fyn kinases, revealing the possibility that amino acids 425–533 of PSD-95 are required for the phosphorylation of PSD-95 by SrcPTKs.
Pyk2, an upstream activator of SrcPTKs, binds to the SH3 domain of PSD-95, and SrcPTKs bind to the PDZ3 domain of PSD-95. This structural adjoin allows SrcPTKs to be activated by Pyk2 . Tyr523 is located between the SH3 and GK domains of PSD-95. There is a direct intramolecular interaction between the SH3 domain and the GK domain in the PSD-95 subfamily . Protein phosphorylation usually results in a conformational change in the modified protein. It is possible that phosphorylation of PSD-95 at Tyr523 opens up this stable intra-molecular structure, facilitates the association of Pyk2 with the SH3 domain of PSD-95 and increases the activation of SrcPTKs.
Brain ischaemia induces the increase in the phosphorylation of PSD-95 Tyr523. The mutation of Tyr523 not only reduces the phosphorylation of PSD-95, but also abolishes the effect of PSD-95 on the up-regulation of NMDA receptor function. The results imply that PSD-95 Tyr523 phosphorylation contributes to the excessive openness of NMDA receptors after brain ischaemia, but the mechanism needs to be clarified. The tyrosine phosphorylation of NMDA receptor subunit 2A by SrcPTKs enhances the activity of NMDA receptors , and PSD-95 allows Src to potentiate the NMDA receptor subunit 1/NMDA receptor subunit 2A currents . Previous work has shown that the antisense oligonucleotide-mediated down-regulation of PSD-95 expression inhibits NMDA receptor subunit 2A tyrosine phosphorylation and diminishes hippocampal CA1 pyramidal cell death after brain ischaemia [19,33]. The tyrosine phosphorylation of PSD-95 appears to precede that of NMDA receptor subunit 2A after transient brain ischaemia. Thus phosphorylation of PSD-95 Tyr523 could promote the tyrosine phosphorylation of NMDA receptor subunit 2A. On the basis of this concept, the rapid phosphorylation of PSD-95 Tyr523 induced by brain ischaemia could open up the intramolecular SH3–GK interaction and facilitate the activation of SrcPTKs by Pyk2, which would result in increased tyrosine phosphorylation of NMDA receptor subunit 2A and subsequent over-activation of NMDA receptors. This hypothesis remains to be evaluated.
In conclusion, the present study reveals an additional mechanism for the regulation of PSD-95 involving tyrosine phosphorylation. Such phosphorylation, accomplished by SrcPTKs, occurs after brain ischaemia, indicating that it has a pathological significance. Tyr523 in PSD-95 is a main phosphorylation site for Src/Fyn kinases, and it contributes to the up-regulation of NMDA receptors. Elucidation of the molecular mechanisms responsible for the over-activation of NMDA receptors through tyrosine phosphorylation of PSD-95 may provide a novel targets for the therapy of ischaemic stroke.
This work was supported by grants from the National Natural Science Foundation of China (30300069), the Program for New Century Excellent Talents in University (NCET-04-0520) and the Natural Science Foundation of Jiangsu Province (BK2007503). We thank Dr M. Sheng (Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, MA, U.S.A.) for providing the GW1/Myc-PSD-95 clone, and Dr R. Zhang and Dr D. Hill (Department of Pharmacology and Toxicology, University of Alabama at Birmingham, Birmingham, AL, U.S.A) for in-depth pre-acceptance editing of the paper. We also thank J. Z. Yan, S. Q. Hu, Y. Y. Zong and Y. Xu for excellent technical assistance.
Ca2+/calmodulin-dependent protein kinase II
Dulbecco's modified Eagle's medium
enhanced green fluorescent protein
PSD-95 with GK domain deleted
proline-rich tyrosine kinase 2
PSD-95 with SH3 and GK domains deleted
Src family protein tyrosine kinase
These authors contributed equally to this work.