(Received for publication, August 16, 1996, and in revised form, November 22, 1996)
From the Vollum Institute, Oregon Health Sciences University, Portland, Oregon 97201
The trafficking and phosphorylation of the NR1 and NR2 subunits of the N-methyl-D-aspartate-type glutamate receptor complex were studied in cultured rat hippocampal neurons. Surface expression was examined by modifying surface receptors via treatment of intact neurons with either the protease chymotrypsin or the cross-linking reagent bis(sulfosuccinimidyl)suberate, followed by quantification of anti-NR1 and anti-NR2B Western blot immunostaining. These studies revealed that only 40-50% of total NR1 immunoreactivity is found at the cell surface, as compared to more than 90% of total NR2B immunoreactivity. Metabolic labeling of the cultures with 32P revealed that NR2 subunits are highly phosphorylated under basal conditions, whereas basal phosphorylation of NR1 subunits is barely detectable. Following stimulation of the cultures with glutamate/glycine or phorbol esters, NR1 phosphorylation was found to be enhanced by 3-5-fold, whereas phosphorylation of NR2 subunits was enhanced by less than 2-fold. To determine whether the difference in the basal NR1 versus NR2 phosphorylation could be due to tyrosine phosphorylation of NR2, phosphoamino acid analyses of NR2 were performed. These analyses revealed phosphorylation on serine but not on threonine or tyrosine; immunoprecipitation and deglycosylation experiments using anti-phosphotyrosine antibodies confirmed that NR2 subunits in the primary hippocampal cultures are not detectably phosphorylated on tyrosine residues. These results demonstrate that the NR1 and NR2 subunits, which assemble into heteromeric complexes to form functional N-methyl-D-aspartate receptors, are trafficked in neurons with differential efficiency to the plasma membrane and exhibit different levels of basal versus stimulated serine phosphorylation.
Fast excitatory synaptic transmission in the mammalian brain is
mediated by two classes of receptor; both classes contain intrinsic ion
channels and are responsive to the neurotransmitter glutamate. These
receptor classes are defined by the specific agonists that activate
them: N-methyl-D-aspartate
(NMDA)1 and
-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA). A third
class of mammalian glutamate-gated ion channels, the kainate receptors,
has been characterized and cloned, but the functional significance of
these receptors in brain circuits is unclear (1).
The molecular structure of glutamate receptors has been the subject of intensive study. NMDA receptors are comprised of NR1 subunits, which exist as a number of splice variants derived from a single gene product, and NR2 subunits, which are four different gene products referred to as NR2A-D. AMPA receptors are comprised of the GluR1-4 subunits, four different gene products that may undergo alternative splicing and RNA editing (2). Both NMDA receptors (3) and AMPA receptors (4, 5) are believed to exist as heteropentameric assemblies of the aforementioned subunits. The nature of these complexes, however, is quite different for the two classes of glutamate receptor. Whereas the four AMPA receptor subunits are very similar in both size and primary sequence (6), the NR1 subunits of the NMDA receptor share less than 20% primary sequence identity with the NR2 subunits and are 120 kDa in size compared to 180 kDa for the NR2 subunits (7). This divergence in the structure of the subunits of the NMDA receptor suggests the possibility that the different subunits may be posttranslationally regulated in very different manners, with potentially interesting implications for NMDA receptor function.
A number of important questions remain unanswered about the posttranslational regulation of NMDA receptor subunits. Although it is known that both the NR1 (8, 9) and NR2 subunits (10-12) can be phosphorylated, it is not known whether the level of NR subunit phosphorylation can be altered by physiological stimulation. Moreover, although immunohistochemical studies of NMDA receptor subunits have revealed staining in both synaptic and intracellular regions (13, 14), quantitative analyses of the surface expression of the various subunits have not been performed. In the present study, we examined the cell surface trafficking of the NR1 and NR2 subunits in cultured hippocampal neurons and also quantified NMDA receptor subunit phosphorylation under basal conditions and following stimulation. We report that the NR1 and NR2 subunits of the NMDA receptor differ with regard to both their level of surface expression and their basal versus stimulated levels of phosphorylation.
Timed-pregnant Sprague-Dawley rats
were obtained from Bantin and Kingman. Culture plates were from Falcon.
Minimal essential medium (without glutamine) and heat-inactivated fetal
calf serum were from Life Technologies, Inc. Mito+ serum extender
was from Collaborative Research. Chymotrypsin and papain were from
Worthington. BS3 was from Pierce. Monoclonal anti-NR1 was
from Pharmingen. Monoclonal anti-NR2B was from Transduction Labs.
Monoclonal anti-phosphotyrosine 4G10 was obtained from Upstate
Biotechnology, Inc.. Polyclonal anti-NR1 and anti-NR2A/B antisera were
gifts from Robert Wenthold (National Institutes of Health). Anti-actin
monoclonal antibodies and microcystin-LR were from Boehringer Mannheim.
Anti--tubulin monoclonal antibodies were from
Sigma. Goat anti-rat brain
calcium/calmodulin-dependent protein kinase II (
subunit) antibodies were developed by Bethyl Laboratories. Donkey
anti-rabbit and donkey anti-mouse horseradish peroxidase-linked
secondary antibodies were from Amersham Corp.; swine anti-goat was from
Boehringer Mannheim. Cellulose plates were from Merck, whereas
nitrocellulose sheets were from Schleicher & Schuell. ReadySafe liquid
scintillation fluid was obtained from Beckman Instruments. TPA was
obtained from Calbiochem. The Renaissance chemilluminescence kit was
obtained from DuPont NEN. Radiolabeled orthophosphate was from ICN.
Brain-derived neurotrophic factor and neurotrophin-3 were gifts from
Amgen. All other reagents were from Sigma.
Primary cultures of rat hippocampal neurons
were prepared from rat pups at 24-48 h of postnatal life. Hippocampi
were dissected out into room temperature osmotically balanced saline
solution (SS, 137 mM NaCl, 5.3 mM KCl, 0.17 mM Na2HPO4, 0.22 mM
KH2PO4, 10 mM Hepes, 33 mM glucose, 44 mM sucrose, and 0.024 g/liter
phenol red, pH 7.3, 325 mosm), cut into five to six pieces, and
incubated for 1 h with 10 ml of 20-unit/ml papain solution in SS.
The papain solution was removed, and the hippocampal fragments were
washed once in complete medium (minimal essential medium, 5%
heat-inactivated fetal calf serum, 1 µl/ml serum extender, 21 mM glucose, 10 µg/ml 5-fluoro-2-deoxyuridine, and 25 µg/ml uridine). Cells were then dissociated into fresh complete
medium via 15-20 passes through a Pasteur pipette and plated onto
sterile, poly-L-lysine-coated 35-mm culture plates at an
approximate density of two hippocampi per dish. Cultures were fed with
almost complete changes of medium on postculture days 1, 4, and 7 and
typically used for experiments on culture day 9 or 10. Cultures
prepared via this method are of high density (roughly 1-2 million
cells/plate) and have a very high neuron:glia ratio because they are
cultured in the presence of a mitotic inhibitor
(5-fluoro-2
-deoxyuridine) from the very first day. Prior to use in
experiments, cultures were washed with room temperature SS three times
and then exposed to the various treatments described below.
Following a wash incubation of 20 min at 37 °C, cultures were incubated with 1 mg/ml chymotrypsin in SS for 10 min with agitation at 37 °C. The SS was then aspirated, and the plates were washed three times in ice-cold harvest buffer (SS containing 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 10 µM leupeptin, 50 mM ethanolamine, 50 mM sodium orthovanadate, 50 mM NaF, 1 µM microcystin-LR, and 1 mM EDTA). Preliminary experiments demonstrated that phenylmethylsulfonyl fluoride has a rapid and irreversible inhibitory effect on chymotrypsin activity. After the third wash, 1.0 ml of fresh, ice-cold harvest buffer was added, and the cells were scraped, transferred into a 1.5-ml snap-cap vial, briefly sonicated, and then frozen.
Cross-linkingFollowing a wash incubation of 20 min at 37 °C, cultures were incubated with 1 mg/ml BS3 in SS for 10 min with agitation at 37 °C (these conditions were optimized in preliminary experiments). The SS was then aspirated, and the plates were washed three times in ice-cold harvest buffer (ethanolamine is present in the harvest buffer to quench any unreacted BS3). After the third wash, 1.0 ml of fresh, ice-cold harvest buffer was added, and the cells were scraped, transferred into a 1.5 ml snap-cap vial, briefly sonicated, and then frozen.
SDS-PAGE and Western BlottingLysate samples were thawed on ice and then mixed with SDS-PAGE sample buffer to final concentrations of 2% SDS, 5% 2-mercaptoethanol, and 5% glycerol. The samples were then loaded into 9% acrylamide gels and run at 150 V for 1 h. The dye front was not allowed to exit the bottom of the gel. Proteins were then blotted onto nitrocellulose for 4 h at 60 V. Blots were blocked for 30 min with 2% milk in Tris-buffered saline (TBS, 50 mM NaCl, 10 mM Tris, pH 7.4) with 0.1% Tween 20 and then incubated with primary antibodies in 2% milk-TBS with 0.1% Tween 20 ("wash buffer") for 1 h at room temperature. Blots were then washed three times (10 min each) in wash buffer and incubated for 1 h at room temperature with an appropriate horseradish peroxidase-linked secondary antibody at a dilution of 1:2000. Following three more washes with wash buffer, blots were developed via a 1-min incubation with the Renaissance chemilluminescence reagent, followed by exposure to sheets of Kodak X-OMAT film for varied lengths of time. Films were developed such that all bands resulting from a given blot exposure were in the most linear range of intensity, as determined from preliminary experiments in which standard curves were constructed by plotting the relative absorbance of the immunoreactivities for increasing concentrations of lysed membranes versus the amount of membranes loaded per lane. Reductions in immunoreactivity in experimental samples were determined from these standard curves.
Metabolic Labeling and ImmunoprecipitationHippocampal cultures were preincubated for 2 h in phosphate-free minimal essential medium and then labeled with [32P]orthophosphate (1 mCi) in 1.0 ml phosphate-free minimal essential medium for 2 h at 37 °C. The cultures were washed and incubated for 15 min more at 37 °C with either SS (for control cultures) or SS containing glutamate/glycine or TPA (for stimulated cultures). Following two quick washes with cold SS, the cells were harvested into 0.5 ml of ice-cold solubilization buffer containing (unless otherwise indicated) 50 mM Hepes, pH 7.4, 0.5 M NaCl, 0.5% Triton X-100, 0.1% deoxycholate, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 10 mM NaF, and 1 µM microcystin-LR. The harvested cells were kept on ice and then briefly sonicated to achieve maximal solubilization. Each solubilized sample was then incubated for 2 h at 4 °C with 10 mg of protein A-Sepharose beads and 2 µg of either anti-NR1 or anti-NR2A/B polyclonal antibody. Following this incubation, the supernatants were removed, and the beads were washed five times with solubilization buffer. After the final wash, the beads were resuspended in 100 µl of SDS sample buffer, incubated with agitation for 5 min, and briefly centrifuged; the supernatants were loaded on 9% SDS-PAGE gels as described. Incorporation of 32P was quantified by exposing dried gels to film for varying lengths of time and then performing densitometric scanning as described. Western blots were performed on all immunoprecipitated samples from the metabolically labeled cultures to verify that equal amounts of NR1 and NR2 protein were being precipitated under all conditions.
Enzymatic DeglycosylationTo remove N-linked
glycosyl residues from the NMDA receptor subunits, lysate or
immunoprecipitated samples were denatured via boiling for 5 min in the
presence of 0.5% SDS and 250 mM -mercaptoethanol. The
samples were then diluted five times in a buffer containing 50 mM Hepes, pH 7.4, 50 mM sodium orthovanadate,
50 mM sodium fluoride, 1 mM
phenylmethylsulfonyl fluoride, 1 mM benzamidine, 100 µM leupeptin, 1 µM microcystin-LR, and 1 mM EDTA. The diluted samples were incubated for 12 h
at 35 °C in the absence or presence of 10 units/ml
peptide-N-glycosidase F. The samples were then prepared for
SDS-PAGE as described.
Two-dimensional phosphoamino acid analyses were performed according to a method described previously (15). Briefly, excised bands from gels containing 32P-labeled samples were rehydrated, digested overnight at 37 °C with trypsin, lyophilized, and then boiled for 1 h in 200 µl of 6 M HCl. Following another lyophilization step and resuspension in a small volume of buffer, samples were loaded onto cellulose plates along with phosphoamino acid standards and separated in two dimensions as described in the aforementioned reference. The cellulose plates were dried and sprayed with ninhydrin to visualize standards and then exposed to Kodak X-OMAT film for varying lengths of time to visualize the location of 32P-containing species.
Western blots of lysate from hippocampal cultures were probed with commercially available antibodies raised against the NMDA receptor subunits NR1 and NR2B. NR2A and NR2B are the most prevalent NR2 subunits in hippocampal neurons, but NR2A does not appear in high levels until late in development (16, 17). Thus, NR2B should be the major NR2 subunit found in these hippocampal cultures, which are prepared from neonatal rats. Blots probed with a monoclonal anti-NR1 antibody revealed a tight cluster of bands at 120 kDa. These species of slightly different sizes probably represent alternatively spliced forms of NR1; because these bands were too compacted to allow individual quantitation, the 120-kDa anti-NR1 staining was considered in the present study as a single broad band. Blots of the same samples probed with a monoclonal anti-NR2B antibody revealed a sharp band of 180 kDa.
When intact hippocampal cultures were treated before harvest with 1 mg/ml chymotrypsin for 10 min to proteolyze externally accessible
proteins, the major bands for both anti-NR1 and anti-NR2B Western blot
immunostaining of the resulting lysates were reduced (Fig.
1). Longer treatments or higher concentrations of
chymotrypsin did not result in significantly larger losses of
immunoreactivity. Along with the reductions in immunoreactivity of the
major bands came the appearance of breakdown products of approximately
65 kDa for the NR1 subunit and approximately 130 and 85 kDa for the NR2B subunit. It is likely that this pattern of immunostaining represents cleavage of surface-exposed NMDA receptor subunits by
chymotrypsin, because preliminary experiments revealed that intracellular proteins were not affected by the chymotrypsin
incubation. Western blot immunoreactivities for the cytoskeletal
proteins actin and tubulin and the abundant intracellular enzyme
calcium/calmodulin-dependent protein kinase II were
unaltered by exposure of intact cells to chymotrypsin, as reported
previously (18). All of these proteins were, however, readily degraded
by chymotrypsin when the protease was simply added to a lysate
preparation, supporting the conclusion that they were not affected in
the whole-cell experiments because chymotrypsin did not have access to
intracellular compartments.
Also shown in Fig. 1 is the result of experiments where intact
hippocampal cultures were exposed to 1 mg/ml of the membrane-impermeant cross-linking reagent BS3. This treatment, like the
chymotrypsin treatment, resulted in reductions in intensity of the
major immunoreactive bands for both anti-NR1 and anti-NR2B staining;
instead of lower molecular weight breakdown products, however, the
lanes of samples derived from the cross-linked cultures exhibited
higher molecular weight aggregated species, which presumably reflect
NMDA receptor subunits cross-linked together or with associated
proteins. Longer incubations or higher concentrations of
BS3 did not result in significantly larger reductions in
the immunoreactivities of the major bands. As with the
chymotrypsin-treated samples, immunoreactivities for the intracellular
proteins actin, tubulin, and calcium/calmodulin-dependent
protein kinase II were unaffected by the BS3 treatment, as
reported previously (18), suggesting that only externally accessible
proteins were cross-linked by this protocol. The reductions in
immunostaining induced by the chymotrypsin and BS3
treatments were quantified as described under "Materials and Methods." The results of this quantification are shown in Fig. 2. For anti-NR1 immunoreactivity, the two surface
treatments removed between 40 and 50% of the total 120-kDa
immunoreactivity. For anti-NR2B immunoreactivity, the 180-kDa bands of
the treated samples were reduced by more than 90% relative to
control.
Stimulation of the cultures for 15 min with either 10 µM glutamate/1 µM glycine or with 2 µM of the phorbol ester TPA had no significant effect on the fraction of NR1 and NR2 subunits in the plasma membrane. The percentages of anti-NR1 immunoreactivity remaining in Western blots of stimulated cultures following cleavage of surface receptors with chymotrypsin were 52 ± 6% and 55 ± 8% for glutamate/glycine and TPA stimulation, respectively (n = 4); for anti-NR2B staining, the values were 6 ± 3% and 5 ± 3%, respectively (n = 4). Although the glutamate/glycine and TPA treatments had no acute effects on surface expression, both of these treatments were observed to result in substantial increases in NMDA receptor subunit phosphorylation (see below).
Phosphorylation of NR1 and NR2 SubunitsHippocampal cultures
were metabolically labeled with [32P]orthophosphate,
harvested into a solubilization buffer containing 0.5% Triton
X-100/0.1% deoxycholate, and then subjected to immunoprecipitation using a polyclonal anti-NR1 antibody (13). Under these conditions, Western blot analyses revealed that a majority of total NR1 is precipitated, whereas a small but detectable fraction of total NR2 is
coprecipitated (data not shown); these data are consistent with
previous observations (3) that NR2 subunits coimmunoprecipitate with
NR1 subunits but that some NMDA receptor complexes are disrupted when
receptors are solubilized with Triton X-100. When the anti-NR1 precipitates were subjected to SDS-PAGE gels and autoradiography to
detect 32P incorporation, a single major band of
approximately 180 kDa was evident (Fig. 3). Because this
band exactly comigrated with the major anti-NR2B immunoreactive band on
Western blots, it presumably represents phosphorylated NR2 subunits
coprecipitating with the NR1 subunits. This correlation was
strengthened by experiments in which precipitates were enzymatically
deglycosylated overnight with peptide-N-glycosidase F; these
experiments demonstrated that removal of N-linked glycosyl
residues decreased the apparent size of the major NR1-precipitable
phosphorylated band by exactly the same amount as the band seen in
Western blots of NR2B (data not shown).
If the labeled cultures were stimulated for 15 min before harvesting
with either 10 µM glutamate/1 µM glycine or
with 2 µM TPA, 32P incorporation into the
180-kDa band was somewhat increased. More strikingly, a species of 120 kDa, which was barely detectable under control conditions, was evident
as a prominent phosphoprotein following stimulation. This species
exactly comigrated with the major anti-NR1 immunoreactive band on
Western blots and shifted in apparent size by the same amount as the
NR1 subunit following enzymatic deglycosylation, and therefore,
probably represents phosphorylated NR1 subunits. The intensities of all
of the bands on the autoradiograms were quantified as described under
"Materials and Methods," and the increases in the 32P
incorporation into the NR1 and NR2 subunits with stimulation are shown
in Fig. 4. These results demonstrate that
phosphorylation of the NR2 subunits is enhanced less than 2-fold by
stimulation with either glutamate/glycine or TPA, whereas NR1
phosphorylation is enhanced approximately 3-fold by stimulation with
glutamate/glycine and more than 5-fold by stimulation with TPA. The
increases in phosphorylation induced by glutamate/glycine were
completely blocked by coapplication of the specific NMDA receptor
antagonist AP-5; radiolabeled phosphate incorporation into the NR1 and
NR2 subunits was 107 ± 9% and 101 ± 8% of control,
respectively, when the labeled cultures were treated with 10 µM glutamate/1 µM glycine in the presence
of 100 µM AP-5 (n = 4).
Evidence for Phosphorylation of NR2 Subunits on Serine but Not Threonine or Tyrosine
Phosphorylation sites for NR1 have been
well characterized (8, 11), but much less is known about the sites of
NR2 phosphorylation. Because the NR2 subunits have been reported to be
the major tyrosine-phosphorylated species in rat brain postsynaptic
density (10) and synaptic plasma membrane preparations (11, 12), a
potential explanation for the much higher basal phosphorylation
observed for NR2 relative to NR1 in the present study is that NR2
subunits are phosphorylated by basally active tyrosine kinases, whereas
NR1 subunits are not. As shown in Fig. 5A,
extracts of hippocampal culture lysate probed on Western blots with an
anti-phosphotyrosine monoclonal antibody did, indeed, reveal a major
band at 180 kDa, which approximately comigrated with anti-NR2B
staining. When the extracts were boiled and deglycosylated, however,
the major anti-NR2B band shifted to an apparent size of 165 kDa,
whereas no change was evident in the anti-phosphotyrosine main band.
This suggests that the two species are not the same.
This point was further examined via the immunoprecipitation studies shown in Fig. 5B. Hippocampal extract was solubilized with 1% SDS to achieve full solubilization and then diluted five times with buffer containing 1% Triton X-100, according to the method of Lau and Huganir (11). An immunoprecipitation was then performed using an anti-NR2A/B antibody (14) and protein A-Sepharose beads. Following this protocol, roughly one-half of total anti-NR2B immunostaining was precipitated. Immunostaining of the 180-kDa major band of anti-phosphotyrosine immunostaining, however, was not altered. Moreover, no anti-phosphotyrosine staining was detected in the precipitated samples, even upon overexposure of the blots.
Phosphoamino acid analyses of the NR2 phosphorylation under basal conditions revealed phosphorylation on serine residues but no detectable phosphorylation on threonine or tyrosine residues (Fig. 5C). This finding correlates with the lack of NR2 tyrosine phosphorylation observed using the anti-phosphotyrosine antibody and suggests that the NR2 subunits in these primary hippocampal cultures are not tyrosine-phosphorylated. We tried a number of manipulations to attempt to induce tyrosine phosphorylation of the NR2 subunits, including exposure of the cultures for varying lengths of time to glutamate/glycine, TPA, epidermal growth factor, brain-derived neurotrophic factor, and neurotrophin-3. We also attempted to induce integrin-mediated tyrosine kinase activation (19) by plating cultures on different substrata, such as laminin or fibronectin. All of these treatments were followed by harvest of the cultures, solubilization, immunoprecipitation with anti-NR2A/B antibody, and probing of Western blots with anti-phosphotyrosine antibody as described. None of these treatments was found to lead to detectable tyrosine phosphorylation of NR2 (data not shown).
The data presented here demonstrate that the NR1 and NR2 subunits of the NMDA receptor are regulated in quite different ways. In terms of cell surface expression, almost all NR2B subunits are found in the plasma membrane, as compared to less than one-half of total NR1 subunits. This observed difference between the subunits is unlikely to be an artifact due simply to differences in the availability or conformation of NR1 versus NR2B extracellular domains, because two very different techniques (proteolytic cleavage with chymotrypsin and cross-linking with the small, membrane-impermeant molecule BS3) yielded strikingly similar estimates of NR1 versus NR2B surface expression. Given that recent studies in nonneuronal cells have demonstrated that NR1 subunits require coexpression with NR2 subunits to be targeted efficiently to the cell surface (20), our data suggest that neurons possess a large intracellular pool of NR1 subunits that await heteromeric assembly with NR2 subunits before they can be expressed in the plasma membrane.
The large discrepancy between NR1 and NR2B surface expression reported here represents a more quantitative description of observations made in electron microscopic immunohistochemical studies of brain slices that staining for NR2A/B subunits (14) seems to be concentrated more specifically with postsynaptic densities than staining for NR1 subunits (13), which is observed frequently in what seem to be intracellular regions. As for a more detailed mechanistic explanation as to why NR2 subunits are trafficked so much more efficiently than NR1 subunits to the plasma membrane, it has been found recently that NR2 subunits contain specialized motifs, termed tSXV domains, which facilitate interactions with various membrane proteins including the postsynaptic density protein PSD-95 (21). Such interactions may explain, at least in part, the efficiency with which NR2 subunits are trafficked to the plasma membrane.
Protein kinase C phosphorylation of the NR1 subunit expressed in quail fibroblasts has been shown to modulate receptor clustering (9). In the present study, we did not find any evidence for an acute change in NR1 surface expression following large increases in NR1 phosphorylation induced by stimulation with either glutamate/glycine or TPA. These data suggest that although protein kinase C phosphorylation can regulate the distribution of NR1 subunits within the plasma membrane, it has no acute effect on the trafficking of NR1 subunits to and from the plasma membrane. The present data also reveal that the surface expression of NR2B subunits is not rapidly malleable in response to stimulation; the lack of an acute effect of stimulation on NMDA receptor surface expression mirrors previous observations made on the surface expression of the AMPA receptor subunits GluR1-3 (18).
It has been shown previously that phosphorylation of the NR1 subunit in neurons can be enhanced by stimulation with phorbol esters (8). The present results confirm and extend this finding by showing that NR1 phosphorylation in neurons is increased by both phorbol esters and by stimulation of NMDA receptors with their endogenous agonist, glutamate. It was also shown in the present study that stimulation with either phorbol esters or glutamate leads to an increase in NR2 subunit phosphorylation. Direct phosphorylation of NMDA receptors following NMDA receptor activation provides a potential feedback mechanism for NMDA receptors to regulate their own activity. In almost every case where a physiological effect of kinase activation or phosphatase inhibition on NMDA receptor function has been reported, the effect has been to increase the size of currents mediated by the receptor (22). Thus, phosphorylation of NMDA receptors in response to NMDA receptor activation, as observed in the present study, provides a potential positive feedback loop that might be expected to produce transient increases in synaptic strength of the type commonly reported at hippocampal synapses in response to brief bursts of synaptic activity (23).
Basal phosphorylation of NR1 under the culture conditions examined in the present study is quite low, and phosphorylation is markedly enhanced by stimulation. The situation is different for the NR2 subunits, which exhibit a high degree of basal phosphorylation and more modest increases in response to stimulation. Direct evidence for serine/threonine phosphorylation of NR2 subunits has not been reported previously, but the NR2B subunit has been shown to be a major phosphotyrosine-containing protein in brain tissue (10-12). Thus, it might seem likely that tyrosine phosphorylation would account for the high basal NR2 phosphorylation observed in the present study. Such a scenario would also account for the absence of this basal phosphorylation in the NR1 subunits, which have been shown to be phosphorylated on serine but not on tyrosine (8, 11).
It was found in the present study that the major phosphotyrosine-containing protein in hippocampal culture extract is approximately 180 kDa in size, as reported previously for rat brain postsynaptic density (10) and synaptic plasma membrane preparations (11, 12). However, this species in the hippocampal culture extract does not represent tyrosine-phosphorylated NR2 subunits, based on the following three lines of evidence: (i) the 180-kDa major anti-phosphotyrosine band does not shift in size with enzymatic deglycosylation, as the NR2 subunits do; (ii) NR2 subunits immunoprecipitated from hippocampal culture extract do not exhibit any anti-phosphotyrosine immunoreactivity; and (iii) phosphoamino acid analysis of NR2 reveals phosphorylation only on serine residues. Thus, in our primary hippocampal cultures, NR2 subunits are not detectably phosphorylated on tyrosine.
We attempted to induce NR2 tyrosine phosphorylation in the hippocampal cultures via a variety of treatments, as described under "Results." None of these treatments resulted in detectable NR2 tyrosine phosphorylation. Thus, at the present time, it is not certain what factor or factors present in the brain but absent from primary hippocampal cultures are responsible for inducing tyrosine phosphorylation of NR2 subunits. The lack of NR2 tyrosine phosphorylation in the cultures studied here may simply reflect the fact that these cultures are prepared from neonatal rats; it is known that the tyrosine phosphorylation levels of many proteins in rat brain are very low at birth and reach high levels only after the third or fourth week of postnatal life (24). At any rate, the phosphorylation of NR2 subunits exclusively on serine residues observed in the present study is of interest because it may help direct future studies aimed at identifying NR2 phosphorylation sites. There are a multitude of consensus phosphorylation sites for various kinases located on both serine and threonine residues within the C-terminal tail region of the NR2 subunits (7), but in the present study, we found evidence only for phosphorylation on serine.
In conclusion, the present data demonstrate that NMDA receptor NR1 and NR2 subunits are regulated differentially in hippocampal neurons in terms of both their cell surface trafficking and their basal versus stimulated phosphorylation state. These regulatory mechanisms allow neurons tight posttranslational control over the formation and function of NMDA receptors. Multi-subunit receptors like NMDA receptors in which each subunit can be regulated distinctly are susceptible to more levels of regulation than are single-subunit receptors or multi-subunit receptors composed of subunits that are closely homologous to one another. Given that both underactivation (25) and overactivation (26) of NMDA receptors can lead to neuronal dysfunction and death, it is perhaps not surprising that neuronal NMDA receptors are subject to multiple layers of regulation to constrain their activity within narrowly defined parameters.
We thank Robert Wenthold (National Institutes of Health) for providing us with samples of anti-NR1 and anti-NR2A/B antibodies and also for technical advice and comments on the manuscript. We also thank Amgen for providing us with samples of brain-derived neurotrophic factor and neurotrophin-3, as well as Andrés Barria, Hervé Enslen, and Peter Vanderklish for helpful discussions concerning the work described here.