Current address: School of Biological Sciences, Biology Department, University of Sussex, Falmer, Brighton BN1 9QG, UK.
Address correspondence and reprint requests to F. Anne Stephenson, Department of Pharmaceutical and Biological Chemistry, School of Pharmacy, University of London, 29/39 Brunswick Square, London WC1N 1AX, UK. E-mail: email@example.com
Coexpression of PSD-95c-Myc with NR1-1a/NR2A NMDA receptors in human embryonic kidney (HEK) 293 cells resulted in a decrease in efficacy for the glycine stimulation of [3H]MK801 binding similar to that previously described for l-glutamate. The inhibition constants (KIs) for the binding of l-glutamate and glycine to NR1-1a/NR2A determined by [3H]CGP 39653 and [3H]MDL 105 519 displacement assays, respectively, were not significantly different between NR1-1a/NR2A receptors coexpressed ± PSD-95c-Myc. The increased EC50 for l-glutamate enhancement of [3H]MK801 binding was also found for NR1-2a/NR2A and NR1-4b/NRA receptors thus the altered EC50 is not dependent on the N1, C1 or C2 exon of the NR1 subunit. The NR1-4b but not the NR1-1a subunit was expressed efficiently at the cell surface in the absence of NR2 subunits. Total NR1-4b and NR1-4b/NR2A expression was enhanced by PSD-95c-Myc but whole cell enzyme-linked immunoadsorbent assays (ELISAs) showed that this increase was not due to increased expression at the cell surface. It is suggested that PSD-95c-Myc has a dual effect on NMDA receptors expressed in mammalian cells, a reduction in channel gating and an enhanced expression of NMDA receptor subunits containing C-terminal E(T/S)XV PSD-95 binding motifs.
sodium dodecyl sulphate–polyacrylamide gel electrophoresis
N-Methyl-d-aspartate (NMDA) receptors are an important subclass of the excitatory, ionotropic glutamate neurotransmitter receptor family being pivotal for mechanisms of long-term potentiation and as therapeutic targets postischaemia. Seven genes encode the NMDA receptor subunits, NR1, NR2A-NR2D and NR3A-NR3B. The NR1 subunit undergoes extensive alternative splicing to yield eight variants termed, NR1-1a, NR1-1b to NR1-4a, NR1-4b. Native NMDA receptors are formed from the coassembly of NR1 and NR2 subunits, thus there is considerable potential for receptor diversity from both the NR1 splice variants and the different products of the NR2 subunit genes (reviewed in, e.g. Dingledine et al. 1999). NMDA receptors have been localized to synapses on dendritic spines where they are frequently found in the middle of the synaptic disc (e.g. Petralia et al. 1994; Takumi et al. 1999; Racca et al. 2000). At these defined sites, NMDA receptors are associated with a network of proteins including the postsynaptic density 95 protein (PSD-95; alternative nomenclature SAP90); chapsyn-110/PSD-93 and SAP102 (reviewed in Sheng 1996), Shank (Naisbitt et al. 1999), and Yotiao (Lin et al. 1998). NMDA receptors are also associated with the presynaptic protein, synaptic associated protein 97 (SAP97; alternative nomenclature, hdlg). The first of these anchoring proteins to be identified was PSD-95 (Kornau et al. 1995). PSD-95 interacts among others with the NR2 NMDA receptor subunits, Shaker-type K+ channel subunits and α1-adrenergic receptors via the amino acid sequence motif, E(T/S)XV, found at their respective C-termini (Kim et al. 1995; Kornau et al. 1995; Hu et al. 2000). PSD-95 colocalizes with and clusters NMDA receptor subunits following their coexpression in mammalian cells (Kim et al. 1996). It has been shown to be involved in long-term potentiation (LTP) and Ca2+-dependent signalling pathways via recruitment of associated proteins (Migaud et al. 1998). It has also recently been shown to play a role in the maturation of excitatory synapses (El-Husseini et al. 2000).
The functional and pharmacological properties of NMDA receptor subtypes have been characterized following the expression of different subunit combinations in mammalian cells or Xenopus oocytes. In a previous paper, we reasoned that these properties are not necessarily representative of native synaptic receptors since the host cells do not express the above-mentioned accessory proteins. Indeed, we showed that coexpression of NR1-1a/NR2A NMDA receptor clones with PSD-95 resulted in a decreased sensitivity to the enhancement by l-glutamate of [3H]MK801 binding activity (Rutter and Stephenson 2000). This agreed with the functional studies of Yamada et al. (1999) who reported that coexpression of PSD-95 with the NMDA receptor subunits, ε2ζ1 (i.e. NR2B/NR1) in Xenopus oocytes resulted in a decreased sensitivity of the channels to l-glutamate.
In this paper, we have characterized in more detail the pharmacological properties of defined NMDA receptor subtypes coexpressed with PSD-95. We report that glycine enhancement of [3H]MK801 binding to NR1-1a/NR2 receptors is reduced in the presence of PSD-95 in a manner analogous to that observed for l-glutamate. These parallel changes are not a result of a decrease in binding affinity for either glycine or l-glutamate. Further, the results of the association of PSD-95 with NMDA receptors containing two other NR1 variant subunits, i.e. NR1-2a/NR2A and NR1-4b/NR2A are described. PSD-95 has been reported to alter modulation of NMDA receptor activity by protein kinase C (Yamada et al. 1999; Liao et al. 2000) and the src tyrosine kinase (Liao et al. 2000). It can also change the phosphorylation state of NR2A subunits induced by fyn tyrosine kinase (Tezuka et al. 1999). The decreased sensitivity to the enhancement by l-glutamate of [3H]MK801 binding activity induced by PSD-95 may thus be due to changes in phosphorylation state. Since the splice cassettes giving rise to the NR1 splice variants contain distinct, known phosphorylation sites, characterization of the variant NR1 subunits may yield insight into the molecular basis for the reduced efficacy of l-glutamate. However, there was no difference in the l-glutamate modulation of [3H]MK801 binding between the NR1/NR2 splice forms. It is thus suggested that the association of NMDA receptors with PSD-95 may effect the channel gating mechanism.
Materials and methods
[3H]MDL 105 519 (74 Ci/mmol), horseradish-linked secondary antibodies and the ECL Plus Western Blotting System were from Amersham Pharmacia Biotech Ltd. (Little Chalfont, Buckinghamshire, UK). (+)-5-[3H]Methyl-10,11-dihydrobenzo[a,d]cyclohepten-5, 10-imine ([3H]MK801; 28.8 Ci/mmol) and d,l-(E)-2-amino-4-[3H]-propyl-5-phosphono-3-pentenoic acid ([3H]CGP 39653; 40 Ci/mmol) were from Du Pont (UK) Ltd. (Natick, MA, USA). Anti-NR1 911–920 (anti-NR1 C2), anti-NR2A 1381–1394, anti-NR2A/2B 1435–1445 and anti-NR2B 46–60 Cys antibodies were prepared and characterized as previously described (Chazot et al. 1994; Chazot and Stephenson 1997a,b). Anti-c-Myc 9E10 and anti-FLAG M2 mouse monoclonal antibodies were purchased from Sigma-Aldrich (Dorset, UK). K-Blue substrate was from Adgen Ltd. (Ayr, UK). NMDA receptor cDNAs were generous gifts from Professors S. Nakanishi (Kyoto, Japan) and M. Mishina (Nigata, Japan); pCISNR2BFLAG was as in Hawkins et al. (1999) with the FLAG epitope inserted between NR2B 53–54; pGW1PSD-95 and pGW1PSD95c-Myc were generous gifts from Dr Morgan Sheng (Boston, MA, USA). The c-Myc tag sequence, EQKLISEEDL, was inserted at amino acid 12 of PSD-95. All other materials were obtained from commercial sources.
Mammalian cell transfections and cell homogenate preparation
Human embryonic kidney (HEK) 293 cells were cultured and transfected with single subunit clones, pCISNR1-1a, pCISNR1-2a, pCISNR1-4b, pCISNR2A, pCISNR2B or pCIS NR2BFLAG, or pairwise combinations of pCISNR1-1a/pCISNR2A, pCISNR1-2a/pCISNR2A, pCISNR1-4b/pCISNR2A, pCISNR1-1a/pCISNR2B or pCISNR2BFLAG, pCISN1–4b/pCISNR2B or pCISNR2BFLAG cDNA clones in the presence and absence of either pGW1PSD-95c-Myc or pGW1PSD-95 using the calcium phosphate method using a total of 20 µg DNA (Cik et al. 1993). For transfections where cell surface NMDA receptor expression was measured, HEK 293 cells were subcultured overnight prior to transfection in poly lysine (50 µg/mL)-coated 24-well dishes and 0.8 µg total plasmid DNA was used per well. NMDA receptor: PSD-95 clones were transfected using a 1 : 1 ratio; NR1-1a: NR2 binary combinations used a 1 : 3 ratio thus a triple transfection used 2.5 µg pCISNR1-1a : 7.5 µg pCISNR2 : 10 µg pGW1PSD-95. Where appropriate, pSVα-galactosidase or pCIS were included in the transfection mix. Cells were harvested 24–36 h post-transfection, homogenates prepared, adjusted to 0.5 mg protein/mL and analysed by quantitative immunoblotting or radioligand binding. For [3H]MK801 and [3H]MDL 105 519 radioligand binding studies, cell homogenates were prepared as before (Chazot et al. 1992; Chazot et al. 1998). For the glutamate and glycine enhancement of [3H]MK801 radioligand binding studies, cell homogenates were washed extensively to remove endogenous glutamate and glycine. Flasks of transfected cells were pooled, collected by centrifugation (800 g for 5 min at 4°C) followed by homogenization in ice-cold 50 mm Tris-citrate, pH 7.1, containing 5 mm EDTA and 5 mm EGTA (buffer 1) using a Dounce (glass/glass) homogenizer (50 strokes). The homogenate was processed by five cycles of centrifugation (120 000 g for 30 min at 4°C) and homogenization (50 strokes) with final resuspension in buffer ∼ 1–1.5 mg protein/mL.
Preparation of membranes from adult rat forebrain
Well-washed, adult rat forebrain membranes were prepared using five freeze-thaw cycles to remove endogenous l-glutamate and glycine (Chazot et al. 1992).
Immunoblotting was performed as previously described using 25–50 µg protein/sample precipitated using the chloroform/methanol method and SDS-PAGE under reducing conditions in 7.5% polyacrylamide slab minigels (Duggan et al. 1991). Affinity-purified anti-NR1 (C2 exon), anti-NR2A (1381–1394), anti-NR2A/2B (1435–1445) and anti-NR2B (46–60) rabbit antibodies were used at final concentrations of 1–5 µg/mL. Anti-c-Myc 9E10 mouse monoclonal antibodies were used at a dilution of 1 : 2000. Rabbit and mouse immunoglobulin horseradish-linked whole antibodies were used at a final dilution of 1 : 2000 and immunoreactivities detected using the ECL Plus Western Blotting System. Immunoreactive bands were quantified by molecular densitometry using a Molecular Dynamics Personal Densitometer with Molecular Dynamics ImageQuant in the linear range of the film. For the quantitative comparative studies, equal amounts of protein were applied per gel lane. Both the protein concentration and the primary antibody concentrations were adjusted to establish the linear range of the detection system (Chazot and Stephenson 1997b).
Determination of cell surface expression of NMDA receptor subunits by enzyme linked immunoadsorbent assay (ELISA)
The measurement of cell surface expression of NMDA receptors was carried out using an ELISA method based on that of Bonnert et al. (1999) using antibodies directed against extracellular NMDA receptor subunit epitopes. Transfected HEK 293 cells were incubated for 24 h, the media aspirated and cells incubated with phosphate-buffered saline (PBS) containing 4% (w/v) milk powder (1 mL) for 30 min at room temperature (25°C). Cells were incubated for 1 h at room temperature with the primary antibody in PBS containing 4% (w/v) milk powder (triplicate samples) then washed for 5 min at room temperature with 1 mL PBS containing 4% (w/v) milk powder and incubated with a 1 : 1000 dilution of anti-rabbit or anti-mouse Ig horseradish peroxidase linked in PBS containing 4% (w/v) milk powder for 1 h at room temperature. Cells were washed 3 × 5 min in 1 mL PBS containing 4% (w/v) milk powder followed by 3 × 5 min in PBS. K-blue substrate was added (1 mL), samples incubated for 25 min at room temperature and the OD at λ = 650 nm measured.
Radioligand binding assays
[3H]MDL 105 519 binding assays were performed as previously described (Chazot et al. 1998). [3H]CGP 39653 radioligand binding was carried according to the method of Kendrick et al. (1996) using 0.1 mm l-glutamate for the determination of non-specific binding and a 0.6-nm[3H]CGP 39653 final concentration for displacement assays. [3H]MK801 saturation and displacement assays to transfected cell homogenates (50 µg protein) were performed in the presence of either 10 µm or 1 mm l-glutamate and 30 µm glycine using 100 µm thienylcyclohexylpiperidine (TCP) for the determination of non-specific binding, an incubation time of 2 h at 22°C and polyethyleneimine filtration for the separation of bound and free radioligand (Chazot et al. 1994). For the dose-dependent glutamate enhancement of [3H]MK801 binding studies, well-washed transfected cell homogenates (150 µg protein) were incubated with 2 nm[3H]MK801 and 30 µm glycine for 2 h at 22°C in the presence of a 0–10 mm range of glutamate concentration. Separation of bound and free radioligand was again by polyethyleneimine filtration. For the dose-dependent glycine enhancement of [3H]MK801 binding, well-washed transfected cell homogenates (150 µg protein) were incubated with [3H]MK801 (2 nm), 1 mm l-glutamate and a 0–10 mm concentration range of glycine. [3H]MDL105 519 displacement assays were carried out using 3 nm radioligand and a 0–1 mm concentration range of glycine. Binding curves were analysed by the Inplot program (Graph-Pad Software Inc., Prof M. A. Simmonds, School of Pharmacy, London, UK). Note that experiments comparing the radioligand binding properties to a defined receptor in the presence and absence of PSD-95 were always carried out in parallel.
Protein concentrations were determined using the method of Lowry et al. (1951) with bovine serum albumin as the standard protein.
The effect of coexpression of PSD-95c-Myc on the l-glutamate affinity of NR1-1a/NR2A receptors as determined by displacement of [3H]CGP 39653 radioligand binding
We previously reported that coexpression of PSD-95c-Myc resulted in an approximately four-fold increase in the EC50 value for the stimulation of [3H]MK801 radioligand binding by l-glutamate to NR1-1a/NR2A receptors (Rutter and Stephenson 2000). In order to determine if this increase was due to a change in agonist affinity, the l-glutamate recognition site was characterized by radioligand binding using the competitive antagonist, [3H]CGP 39653. There was no significant difference in the KD for the binding of [3H]CGP 39653 to NR1-1a/NR2A expressed in the presence and absence of PSD-95c-Myc but an ∼ 2.5-fold increase in binding sites was found. The binding parameters determined by [3H]CGP 39653 saturation studies were: NR1-1a/NR2A, KD = 4.7 ± 0.6 nm, Bmax = 760 ± 120 fmol binding sites/mg protein (n = 3); NR1-1a/NR2A + PSD-95c-MycKD = 6.6 ± 3.1 nm, Bmax = 1814 ± 224 fmol binding sites/mg protein (n = 3). l-Glutamate inhibition of [3H]CGP 39653 binding for both NR1-1a/NR2A ± PSD-95c-Myc was best fit by a single site with no significant difference in KI values between the two samples. The binding parameters for l-glutamate were: NR1-1a/NR2A, KI = 81 ± 13 nm, Hill coefficient = 1.03 ± 0.11 (n = 3); NR1-1a/NR2A + PSD-95c-MycKI = 71 ± 18 nm, Hill coefficient = 0.90 ± 0.05 (n = 3). Thus no change in the KI for l-glutamate for NR1-1a/NR2A receptors expressed in the presence of PSD-95c-Myc was found. Representative [3H]CGP 39653 saturation binding and l-glutamate inhibition curves are shown in Fig. 1.
The effects of coexpression of PSD-95c-Myc on the glycine sensitivity of [3H]MK801 binding to NR1-1a/NR2A receptors expressed in HEK 293 cells and in adult mammalian forebrain
Both l-glutamate and glycine are known to enhance the binding of [3H]MK801 in a dose-dependent manner. It was of interest therefore to investigate if PSD-95c-Myc also resulted in a decreased EC50 for the glycine stimulation of [3H]MK801. Thus, HEK 293 cells were transfected with NR1-1a/NR2A ± PSD-95c-Myc, cell homogenates, subjected to extensive washing to remove endogenous glycine and [3H]MK801 radioligand binding measured in the presence of fixed concentrations of l-glutamate (1 mm) and the glycine site antagonist, 7-chlorokynurenate (10 µm) and increasing concentrations of glycine. Typical results are shown in Fig. 2. Glycine caused a dose-dependent increase in specific [3H]MK801 binding for both NR1-1a/NR2A receptors expressed in the presence and absence of PSD-95c-Myc. It was found that as for l-glutamate, PSD-95c-Myc resulted in an increased EC50 for glycine stimulation of [3H]MK801 binding. The values were: NR1-1a/NR2A, EC50 = 3.1 ± 0.7 µm (n = 3); NR1-1a/NR2A + PSD-95c-Myc, EC50 = 7.3 ± 1.5 (n = 3). These values were compared with the EC50 for glycine enhancement of [3H]MK801 binding to membranes prepared from adult rat forebrain. Assayed under the same conditions for forebrain, the EC50 = 0.67 ± 0.06 µm (n = 3). Similarly, the EC50 for the enhancement of [3H]MK801 by l-glutamate for forebrain assayed in the presence of 30 µm glycine was also found to be an order of magnitude less than the value found for NR1-1a/NR2A ± PSD-95c-Myc, i.e. 0.46 ± 0.04 µm (n = 3; results not shown). This is compared to EC50 values of 1.8 µm and 7.8 µm for NR1-1a/NR2A in the absence and presence of PSD-95c-Myc, respectively (Rutter and Stephenson 2000).
In order to determine if the increase in EC50 for the glycine stimulation of [3H]MK801binding was due to a change in affinity of the receptor for glycine, the inhibitory constant, KI, for glycine was determined by displacement assay using the competitive glycine antagonist, [3H]MDL105 519. NR1-1a/NR2A NMDA receptors expressed in the presence and absence of PSD-95c-Myc have the same KD for [3H]MDL105 519 (Rutter and Stephenson 2000). Under the assay conditions used here, the glycine inhibition curve was best fit by a two-site binding model. No significant difference in the KI values for glycine binding to NR1-1a/NR2A ± PSD-95c-Myc were found. The binding parameters were: NR1-1a/NR2A, KI-high = 1.9 ± 1.0 µm, with 56 ± 16% site occupancy, KI-low = 34 ± 17 µm (n = 4); NR1-1a/NR2A + PSD-95c-Myc,KI-high = 1.9 ± 1.3 µm, with 49 ± 15% site occupancy, KI-low = 30 ± 19 µm. Thus there was no difference in the KI for glycine for NR1-1a/NR2A receptors expressed in the presence of PSD-95c-Myc. Representative [3H]MDL105 519 inhibition curves are shown in Fig. 2.
Are the effects of PSD-95 on the properties of expressed NR1/NR2 receptors dependent upon the splice variant of the NR1 subunit?
As described earlier, PSD-95c-Myc was initially shown to have two effects on NR1/NR2 NMDA receptors expressed in mammalian cells, the increased EC50 value for the stimulation of [3H]MK801 radioligand binding by l-glutamate and an increased expression of NR2 subunits (Rutter and Stephenson 2000). These experiments were carried out using only the NR1-1a NR1 splice variant. Thus similar experiments were carried out to investigate if these changes also occurred with other NR1 splice variants thus possibly yielding insights into their molecular mechanisms.
Expression levels in cell homogenates
NR1-1a/NR2A, NR1-2a/NR2A and NR1-4b/NR2A NMDA receptor subunit combinations were expressed transiently in parallel HEK 293 cell transfections. These three NR1 splice forms were selected because they encompass the complete range of NR1 exons, i.e. NR1-1a has exons C0, C1 and C2 at its C terminus; NR1-2a lacks the C1 exon; NR1-4b also lacks the C1 exon but has the N-terminal N1 exon cassette. The resultant cell homogenates were prepared and analysed by quantitative immunoblotting using an anti-NR1 pan antibody which recognizes all the NR1 variants. The results are shown in Fig. 3. Interestingly, it was found that the total expression level of the three NR1 splice variants coexpressed with NR2A was significantly different. Quantitative immunoblotting showed that the NR1-1a subunit was the most abundantly expressed with NR1-2a expressed at 0.6 ± 0.2 (n = 3; p < 0.05) and the NR1-4b at 0.3 ± 0.1 (n = 3; p < 0.005) relative to a value of 1.0 for NR1-1a. The expression level of NR2A was not significantly different between all combinations (Fig. 3). In the presence of PSD-95c-Myc, the expression level of NR2A was enhanced when expressed with each of the NR1 splice variants. There was no significant difference in the percentage enhancement of NR2A expression between NR1 splice variants (Fig. 3). As reported previously, NR1-1a was unchanged and the expression of NR1-2a ± PSD-95c-Myc was not significantly different, the ratio for NR1-2a/NR2A + PSD-95c-Myc: NR1-2a/NR2A – PSD-95c-Myc being 1.2 ± 0.2 (n = 3). However coexpression of PSD-95c-Myc with NR1-4b/NR2A resulted in an increased expression of NR1-4b with a ratio of 1.8 ± 0.1 (n = 3; p < 0.05; Fig. 3).
Cell surface expression
To determine if the enhanced expression of NR1-4b/NR2 + PSD-95c-Myc resulted in an increase of cell surface receptors, an ELISA was carried out using antibodies directed at extracellular NMDA receptor subunit epitopes. (Note, these experiments were performed with NR1/NR2B receptors because of a lack of appropriate anti-NR2A antibodies). The results are summarized in Fig. 4. Initially, single subunit transfections were carried out and in agreement with both Okabe et al. (1999), Standley et al. (2000) and Scott et al. (2001), it was found that NR1-4b but not NR1-1a or NR2B was expressed efficiently at the cell surface when expressed alone. In the presence of the NR2B subunit, NR1-1a/NR2B and NR1-4b/NR2B were both expressed at the cell surface. In the presence of PSD-95c-Myc there was no significant increase in cell surface NR1-1a or NR1-4b when expressed alone or in combination with NR2B. Cell surface NR2B was unchanged when expressed with NR1-4b however, in the presence of NR1-1a, an approximate six-fold decrease was found.
The effect of coexpression of PSD-95c-Myc on [3H]MK801 binding to NR1-1a/NR2A, NR1-2a/NR2A and NR1-4b/NR2A receptors
The results for the characterization of [3H]MK801 binding to NR1-1a/NR2A, NR1-2b/NR2A and NR1-4b/NR2A receptors are summarized in Fig. 4. Specific [3H]MK801 binding was carried out at a single radioactive concentration of 40 nm ± PSD-95c-Myc in the presence of 1 mm l-glutamate. Interestingly, it was found that the NR1-4b/NR2A combination resulted in a significant increase in the number of specific [3H]MK801 binding sites despite the fact that the NR1-4b subunit was expressed at the lowest level (Fig. 3). Each of the NR1/NR2A splice variant combinations have similar KDs for [3H]MK801 binding (results not shown). Coexpression of PSD-95c-Myc resulted in a significant increase in the number of [3H]MK801 binding sites. The fold-increase was the highest for NR1-1a/NR2A being ∼ three-fold and least for the NR1-4b/NR2A combination with an ∼1.4-fold increase in the presence of PSD-95c-Myc (Fig. 4b).
L-Glutamate sensitivity of NR1/NR2 receptors as determined by thel-glutamate enhancement of [3H]MK801 radioligand binding activity
To determine if NR1-2a/NR2A and NR1-4b/NR2A receptors had an increased EC50 for the l-glutamate stimulation of [3H]MK801 binding as discovered for NR1-1a/NR2A receptors, specific [3H]MK801binding at 2 nm concentration was carried out on the three different NR1/NR2 NMDA receptor combinations in the presence of either 0 mm, 10 µm or 1 mm l-glutamate. The results are shown in Fig. 5(c). It can be seen that in the presence of PSD-95c-Myc, there was a significant ∼1.4-fold increase in the ratio of [3H]MK801 binding in the presence of 1 mm: 10 µm l-glutamate for NR1-2a/NR2A and NR1-4b/NR2A. Thus the reduced efficacy of l-glutamate on the facilitation of [3H]MK801binding by PSD-95c-Myc is not dependent on the N1, C1 or C2 exon of the NR1 subunit.
In this paper, we have characterized in more detail the interaction between expressed NMDA receptors and PSD-95c-Myc. We have used the different NR1 splice variants to substantiate our previous findings to show that PSD-95c-Myc has two distinct effects on NMDA receptors namely, a decreased efficacy of both l-glutamate and glycine in the enhancement of specific [3H]MK801 binding and the stabilization of those NMDA receptor subunits containing the E(T/S)XV amino acid sequence motif.
Association of NR1/NR2 NMDA receptors with PSD-95c-Myc results in decreased efficacy of both l-glutamate and glycine in the enhancement of specific [3H]MK801 binding activity
We had previously shown that coexpression of PSD-95c-Myc with NR1-1a/NR2A NMDA receptors resulted in an approximate five-fold decreased efficacy of l-glutamate to enhance [3H]MK801 binding which was dependent upon an interaction between PSD-95c-Myc and the C-terminal NR2A ESDV sequence (Rutter and Stephenson 2000). This observation was in agreement with Yamada et al. (1999) who reported a similar decreased sensitivity of NR1-1a/NR2B receptors to l-glutamate. Here, we have now shown that the EC50 for glycine stimulation of [3H]MK801 binding is also increased although the magnitude of this increase was only ∼ two-fold compared to the five-fold change for l-glutamate. The KDs for the binding of l-glutamate and glycine to NR1-1a/NR2A receptors as determined by [3H]CGP 39653 and [3H]MDL105 519 displacement assays, respectively, were both unaffected by coexpression of PSD-95c-Myc. It was estimated that a two-fold change in the respective KD values should be readily detectable in the inhibition assays. It is concluded therefore that there is no significant difference in l-glutamate and glycine binding affinity to NR1-1a/NR2A receptors ± PSD-95c–Myc. The interaction with the scaffolding molecule is thus speculated to effect the gating of the channel rather than agonist binding.
The decreased efficacy of l-glutamate to enhance [3H]MK801 binding to cloned NMDA receptors in the presence of PSD-95c-Myc was also observed for the NR1 splice variants, NR1-2a and NR1-4b both coexpressed with NR2A. We had previously suggested that PSD-95c-Myc may facilitate phosphorylation of NMDA receptors thus resulting in the reduced responsiveness to l-glutamate. The NR1 C1 exon is notable for the presence of multiple phosphorylation consensus amino acid sequences (Raymond 1998); the C1 exon is absent in the NR1-4b splice form yet increased EC50 values are still apparent. It is therefore perhaps more likely that PSD-95 alters the gating properties of NMDA receptors purely by physical restriction due to the formation of receptor clusters rather than a change in phosphorylation state. Interestingly, a recent report showed that protein kinase C increases the opening rate of NR1-4b/NR2A NMDA channels thereby, in contrast to the results described here, probably increasing agonist sensitivity (Lan et al. 2001). Protein kinase C was also shown to induce a rapid delivery of functional NR1-4b/NR2A NMDA channels to the cell surface both in recombinant systems and in neurons (Lan et al. 2001), an action that may be facilitated by PSD-95 or mimicked by coexpression of NR1/NR2 subunits with PSD-95.
The EC50 value for l-glutamate enhancement of [3H]MK801 binding to NR1-1a/NR2A ± PSD-95c-Myc receptors is an order of magnitude larger than that found for rat cerebral cortical membranes (Ransom and Stec 1988; Rutter and Stephenson 2000; AR Rutter, unpublished observations). The same is also true for the glycine enhancement of [3H]MK801 binding (Fig. 2). Thus, either the native rat brain preparations comprise a mixture of synaptic and extra-synaptic receptors or alternatively, other components of the postsynaptic density contribute towards the regulation of NMDA receptor activation.
The enhanced expression of NR1-4b and NR2A/B in HEK 293 cells transfected with NR1/NR2 combinations + PSD-95c-Myc is due to accumulation/stabilization of assembled intracellular receptors
PSD-95c-Myc coexpression of NR1-1a/NR2A or NR1-1a/NR2B resulted in an approximate 2.5-fold increase in NR2 subunits, no detectable increase in the level of the NR1-1a subunit but an approximate three-fold increase in the number of [3H]MK801 binding sites (Rutter and Stephenson 2000). This apparent inconsistency was explained by the fact that a significant proportion of NR1-1a subunits exist as an intracellular pool of unassembled subunits and the lack of effect of PSD-95c-Myc on NR1-1a subunits but an overall increase in [3H]MK801 binding was due to a redistribution of NR1-1a subunits from the unassembled pool to assembled receptor complexes (Chazot and Stephenson 1997a; Rutter and Stephenson 2000). In this paper, it was found that although total NR1-4b subunits expressed in combination with NR2A subunits were expressed at the lowest level they resulted in the expression of the most [3H]MK801 binding sites compared to NR1-1a/NR2A and NR1-2a/NR2A combinations. It has recently been shown using Tac-NR1 chimeras that the C1 cassette of NR1 contains an endoplasmic reticulum (ER) retention signal (Standley et al. 2000). Only NR1-1a of the three NR1 splice forms tested has a C1 cassette. This suggests, in agreement with our earlier proposal, that indeed there is an intracellular pool of NR1-1a by virtue of the ER retention signal. Thus despite high levels of expression, the NR1-1a subunit does not assemble efficiently with NR2 subunits yielding low [3H]MK801 binding sites. Further support for the existence of an intracellular pool of NR1-1a subunits and the hypothesis that co-association with PSD-95c-Myc stabilizes interacting proteins in heterologous expression systems was evident from experiments in which the cell surface rather than total NR1 and NR2 subunits ± PSD-95c-Myc were measured post-transfection. In agreement with studies using wild-type NR1 single subunit expression (Okabe et al. 1999; Standley et al. 2000), Tac-NR1 chimeras (Scott et al. 2001) and NR1-1a/NR2A receptors (McIlhinney et al. 1998), it was shown that NR1-1a expressed alone showed minimal cell surface expression compared to the NR1-4b splice form which was expressed efficiently at the cell surface. For NR1/NR2 receptors expressed in the presence of PSD-95c-Myc, it was found that cell surface NR1-1a and NR2B surface expression was either not significantly different to expression levels in the absence of PSD-95c-Myc or actually decreased. This is despite the ∼2–3-fold increase in total subunits and the increase in both [3H]MK801 and [3H]CGP 39653 specific binding sites for NR1-1a/NR2A + PSD-95c-Myc. It is speculated that association with PSD-95 either prevents degradation of assembled receptors following internalization or alternatively, it may facilitate NR1 and NR2 subunit assembly without increasing export to the cell surface. Jugloff et al. (2000) reported that association of the Kv1.4 K+ channel with PSD-95 resulted in suppression of Kv1.4 internalization but without an increase in cell surface expression. There is thus a precedent for stabilization of channel proteins with no concomitant increase in cell surface expression. Roche et al. (2001) recently reported that NR1/NR2B receptors are rapidly internalized in heterologous cells. They went on to identify an internalization signal, YEKL, near the C-terminus of the NR2B subunit using TacNR2B constructs. Coexpression of PSD-95 with TacNR2B blocked NR2B internalization in HeLa cells and in hippocampal neurons in primary culture. One might therefore predict that this should lead to an increase in cell surface receptors for NR1-1a/NR2B + PSD-95c-Myc in HEK 293 cells. Differences may be due to the fact that the whole receptor complex was studied here compared to the use of TacNR2B chimeras in the absence of an NR1 subunit.
Association of NMDA receptors with the scaffolding protein, PSD-95, results in a decreased sensitivity of [3H]MK801 binding to enhancement by both l-glutamate and glycine. This, we speculate, is due to a decrease in channel gating and may thus be a fundamental distinguishing feature between synaptic and extra-synaptic NMDA receptors; extra-synaptic receptors requiring increased neurotransmitter sensitivity. Further, regulation of the alternative splicing of the NR1 subunit gene and of the association of NMDA receptors with PSD-95 influences the efficiency of cell surface NMDA receptor expression and thereby the molecular organization and efficacy of the excitatory postsynaptic membrane.
This work was funded by the Biotechnology and Biological Sciences Research Council (BBSRC) UK. ARR was funded by a postgraduate studentship from AstraZeneca, Loughborough, UK. We thank Dr Kieran Brickley and Dr Paul L Chazot for antibody production, Michalis Papadakis for help with Fig. 4 and Professors S. Nakanishi, M. Mishina and M. Sheng for the generous donation of cDNA clones.