Dysbindin regulates hippocampal LTP by controlling NMDA receptor surface expression

Abstract

Abnormalities in NMDA receptor (NMDAR) function have been implicated in schizophrenia. Here, we show that dysbindin, a schizophrenia-susceptibility gene widely expressed in the forebrain, controls the surface expression of NMDARs in a subunit-specific manner. Imaging analyses revealed a marked increase in surface NR2A, but not NR2B, in hippocampal neurons derived from dysbindin-null mutant mice (Dys−/−). Exogenous expression of dysbindin reduced NR2A surface expression in both wild-type and Dys−/− neurons. Biotinylation experiments also revealed an increase in surface expression of endogenous NR2A in Dys−/− neurons. Disruption of the dysbindin gene dramatically increased NR2A-mediated synaptic currents, without affecting AMPA receptor currents, in hippocampal CA1 neurons. The Dys−/− hippocampal slices exhibited an enhanced LTP, whereas basal synaptic transmission, presynaptic properties, and LTD were normal. Thus, dysbindin controls hippocampal LTP by selective regulation of the surface expression of NR2A. These results reveal subunit-specific regulation of NMDARs by dysbindin, providing an unexpected link between these two proteins implicated in schizophrenia.

The DTNBP1 gene, which encodes dybindin at chromosome 6p.22.3, is a susceptibility gene for schizophrenia (1). In schizophrenic patients, dysbindin mRNA and protein are reduced in the hippocampus (2) and the prefrontal cortex (3), areas known to contribute to cognitive deficits in schizophrenia. A recent study also reported a reduction in dysbindin mRNA in the hippocampal formation of schizophrenic patients (4). Dysbindin is a key component of biogenesis of lysosome-related organelles complex-1 (BLOC-1), which regulates the trafficking of proteins in the lysosomal pathway (5, 6). Dysbindin has been shown to regulate cell surface expression of the dopamine receptor D2, but not D1 (7). It is interesting to note that D1 is recycled to the plasma membrane after endocytosis, whereas D2 is trafficked to the lysosomal pathway and degraded (8–11). However, how aberrant dysbindin expression contributes to the pathophysiology of schizophrenia has remained elusive.

Abnormal NMDAR function in hippocampus has also been implicated in schizophrenia (12). NMDARs are tetrameric receptors consisting of two obligatory NR1 subunits and two NR2 subunits (13). NR2A and NR2B are the major NR2 subunits expressed in the hippocampus (14), with expression patterns that change during development. In cortex and hippocampus, NR2B is predominantly expressed early in development and gradually decreases. In contrast, NR2A expression is low at birth but gradually increases during development (14–17). In adult cortical and hippocampal neurons, the NR2A subunit is preferentially localized in synaptic sites, whereas the NR2B subunit is expressed in both synaptic and extrasynaptic membranes (18–20). Furthermore, trafficking of NMDARs is subunit-specific (21–25). For example, NR2A and NR2B sort to different intracellular pathways following endocytosis. Specifically, NR2A is sorted to the lysosomal pathway, whereas NR2B preferentially traffics to recycling endosomes (23).

Given that dysbindin is thought to be involved in protein trafficking in the lysosomal pathway, we asked whether it may differentially regulate the trafficking of NMDAR subunits in hippocampal neurons. We took advantage of a natural dysbindin mutant line (Dys−/−), the Sandy mouse, and backcrossed on the C57BL/6J background for >10 generations. Imaging and biochemical analyses showed that the surface expression of NR2A, but not NR2B, was significantly increased in Dys−/− neurons. The wild-type (WT) levels of NR2A surface expression was rescued by expression of exogenous dysbindin in Dys−/− cells. Electrophysiological recordings revealed a dramatic increase in evoked NMDAR-mediated EPSCs, as well as NMDAR-dependent LTP, in Dys−/− hippocampal slices. Together, these data identify dysbindin as an important regulator of NMDAR in a subunit-specific manner. Our findings also support the view that a dysbindin deficit may contribute to glutamatergic imbalances characteristic of schizophrenia.

Results

Increased Surface NR2A in Hippocampal Neurons Derived from Dys−/− Mice.

To determine whether disruption of the dysbindin gene affects the trafficking of NMDAR subunits, we compared the surface expression of NR2A and NR2B expressed in cultured hippocampal neurons derived from Dys−/− or WT mice. Primary hippocampal cultures derived from embryonic (E16–18) WT and Dys−/− mice were transfected with exogenous NR2A or NR2B containing a GFP tag in the extracellular N-terminal domain. We measured the steady-state levels of surface NR2A and NR2B using an antibody against GFP. As shown in Fig. 1A, the levels of NR2A on the surface (Left), but not intracellular (Middle), were markedly increased in hippocampal neurons derived from Dys−/− mice, compared with that from WT mice. In comparison, surface levels of NR2B were not increased (Fig. 1B). Quantitative analyses, which normalized the fluorescence signal on the surface to total fluorescence (surface + intracellular), revealed an approximate 50% increase in surface NR2A in Dys−/− neurons (Fig. 1C).

Fig. 1.

Increased surface expression of NR2A, but not NR2B, in Dys−/− neurons. Cultured hippocampal neurons prepared from WT and Dys−/− E18 embryos were transfected with GFP-NR2A or GFP-NR2B at DIV9–10. After 5 days, live neurons were incubated with an anti-GFP antibody (red) to label surface-expressed receptors. The neurons were fixed, permeabilized and labeled again with anti-GFP antibody to detect intracellular NR2A or NR2B (green). (A and B) Representative images showing increased surface expression of (A) GFP-NR2A but not (B) GFP-NR2B in Dys−/− neurons. Images were collected using 40× objective at 0.34 μm. (Scale bars, 25 μm.) (C) Quantification of the imaging experiments in A and B. Fluorescence intensities were measured using Metamorph software. (n = 27–36 cells; n = 4 independent experiments) and the ratios of surface (red) to total (red + green) fluorescence for WT and Dys−/− neurons are presented. Error bars, mean ± SEM. *: P < 0.02, Student's t test. (D and E) Increased surface expression of endogenous NR2A, but decreased NR2B in cultured Dys−/− cortical neurons. (D) Biotinylated surface proteins from WT and Dys−/− cortical neurons were isolated, resolved by SDS/PAGE, and probed with NR2A, NR2B, or α-tubulin antibodies. (E) Quantification of the biotinylation experiments. The immuno-reactive signals for surface NR2A and NR2B were normalized to total input and presented as a bar graph. n = 3 independent experiments. Error bars, mean ± SEM.; *, P < 0.02, Student's t test.

To evaluate if surface levels of endogenous NMDARs are also increased by loss of dysbindin, we performed surface biotinylation assays to measure the amount of endogenous NR2A and NR2B on the neuronal surface. Due to the difficulty in obtaining a large number of hippocampal neurons from mouse embryonic brains, we used cultured cortical neurons for biotinylation experiments. Biotinylated proteins from the crude membrane lysates were extracted, pulled-down with streptavidin-Sepharose beads, resolved by SDS/PAGE, and blotted with various antibodies (Fig. 1D). In cortical neurons prepared from Dys−/− mice, we observed a substantial increase in NR2A surface expression compared to the WT control, which is consistent with our imaging studies. However, we also found that surface NR2B was reduced in Dys−/− neurons (Fig. 1 D and E).

It is important to note that the dysbindin mutation does not affect the expression level of endogenous NR2A and NR2B in the hippocampus. Western blot analysis demonstrated that the total levels of NR2A or NR2B were not changed in the adult hippocampus (4-week-old) of Dys−/− mice compared to age-matched WT control animals (Fig. 2 A and B). There was also no difference in total expression levels of NR1 or GluR1 between WT and Dys−/− mice (Fig. 2 A and B).

Fig. 2.

NR2A and NR2B expression is not changed in Dys−/− neurons. Hippocampi were dissected from 4-week-old WT or Dys−/− mice. P2 crude synaptosomes were extracted, and 25 μg of protein were loaded in each lane. Expression of NR2A, NR2B, NR1, GluR1, and α-tubulin was detected by Western blot analysis using antibodies as indicated. (A) Representative Western blots of hippocampal lysates from WT or Dys−/− mice. (B) Quantification of Western blots. Relative intensities of ECL bands were quantified by National Institutes of Health ImageJ software. Immunoreactivities of NR2A, NR2B, NR1, and GluR1 were normalized to that of α-tubulin in the same sample, and the signals from Dys−/− hippocampal neurons were compared to that from WT control. n = 4 independent experiments. Error bars, mean ± SEM; in all cases, P > 0.4, Student's t test.

To verify whether the increase in surface NR2A in Dys−/− neurons is due to the disruption of dysbindin, we performed rescue experiments by introducing dysbindin into WT or Dys−/− neurons. We measured the steady-state level of surface-expressed NR2A in hippocampal neurons expressing GFP-NR2A and HA-dysbindin or HA, and we found that expression of exogenous HA-dysbindin significantly reduced the surface NR2A expression in Dys−/− neurons, compared with HA alone (Fig. 3A). Quantitative analyses of fluorescence images revealed a 25% reduction in surface NR2A in Dys−/− neurons (Fig. 3B) upon exogenous dysbindin expression, confirming that the reduction in dysbindin directly leads to increased surface expression of NR2A. Furthermore, expression of exogenous dysbindin also reduced surface NR2A in WT neurons, compared to the HA control (Fig. 3 A and B). To confirm the specificity of our assay, we quantified surface expression of GFP-NR2B in Dys−/− hippocampal neurons expressing exogenous HA-dysbindin. As shown in Fig. S1, Dys−/− neurons expressing HA-dysbindin or HA control exhibited similar levels of surface NR2B. These results demonstrate that NMDAR surface expression is regulated by dysbindin in a subunit-specific manner, and specifically the level of NR2A surface expression is modulated by changes in the amount of dysbindin expressed in neurons.

Fig. 3.

Expression of exogenous dysbindin reduces NR2A surface expression in both WT and Dys−/− hippocampal neurons. Cultured hippocampal neurons from Dys−/− and WT mice were transfected with GFP-NR2A and HA-dysbindin or HA. Neurons were processed for staining of surface and intracellular GFP-NR2A with GFP antibody. Transfected cells were detected by HA antibody. (A) Representative images showing surface and intracellular NR2A in transfected Dys−/− and WT neurons, respectively. Note that transfection of HA-dysbindin, but not HA alone, rescued WT levels of NR2A surface expression in Dys−/− neurons expressing exogenous NR2A, and reduced surface expression of NR2A in WT neurons. Images were collected using 40× objective at 0.34 μm. (Scale bars, 25 μm.) (B) Quantification of surface NR2A in HA-Dysbindin transfected Dys−/− and WT hippocampal neurons. Image analysis was performed as described in Fig. 1C (n = 27–37 cells; n = 3 independent experiments).

Increase in NR2A-Mediated EPSCs in Dys−/− Neurons.

Our imaging and biochemical analyses demonstrated a consistent increase in surface expression of NR2A in Dys−/− neurons. However, the surface biotinylation assays also revealed a decrease in surface NR2B in Dys−/− cortical neurons (Fig. 1 D and E). The differing results using these two assays prompted us to directly examine the function of synaptic NMDARs in hippocampal slices. Whole-cell, voltage-clamp recordings were performed on CA1 pyramidal neurons of hippocampal slices derived from adult (26- to 31-day-old) mice. Excitatory postsynaptic currents (EPSCs) were evoked by stimulating Schaffer collaterals in the presence of the GABA_A receptor antagonist bicuculline (10 μM) (Fig. 4A). When membrane potentials were held at −80 mV (Vh = −80 mV), the EPSCs were predominantly mediated by AMPA receptors. Input-output curves, derived by plotting peak AMPA current amplitudes against the intensities of presynaptic stimulation, were indistinguishable between WT and Dys−/− mice (Fig. 4C). Evoked NMDAR-mediated currents were then recorded at a holding potential of +50 mV (Vh = 50 mV) in the presence of the AMPA receptor antagonist CNQX (50 μM) (Fig. 4B). The NMDAR-mediated synaptic currents recorded from the Dys−/− neurons were greater with faster decay time, compared with those from the WT control (Fig. 4B). Input-output curves revealed consistent differences in NMDAR currents between genotypes in the entire range of stimulation intensities (Fig. 4D). Moreover, the NMDA to AMPA ratio in Dys−/− neurons was also significantly higher than the WT control (Fig. S2 A and B).

Fig. 4.

Increase in evoked NMDAR-mediated synaptic currents in hippocampal CA1. Pyramidal neurons from Dys−/− mice. (A–D) Representative traces (A and B) and quantification (C and D) of amplitudes of evoked EPSCs, showing normal AMPAR-EPSCs but increased NMDA-EPSCs in Dys−/− neurons. Whole-cell, voltage-clamp recordings were performed to record total EPSCs (Vh = −80 mV) and NMDA-EPSCs (Vh = +50 mV, in the presence of CNQX) in hippocampal slices derived from WT and Dys−/− mice (p26–p35). Traces from WT neurons (black) and those from Dys−/− neurons (red) are superimposed. Input-output curves were generated by plotting peak EPSC amplitudes against the intensities of presynaptic stimulation. (E–I) Selective enhancement of NR2A- but not NR2B-mediated NMDA currents. (E) Superimposed traces of NMDA currents recorded from WT and Dys−/− neurons in the presence or absence of the NR2B antagonist ifenprodil (Ifen, 3 μM, lower trace). Note that Dys−/− neurons exhibited much higher amplitude and faster decay time compared with WT neurons. (F) Changes in NMDA currents (Vh = +50 mV, in the presence of CNQX) induced by ifenprodil in neurons from young adult WT and Dys−/− mice (p27–p31). After 2.5 min of recording, ifenprodil (3 μM, indicated by the blue bar) was applied to the hippocampal slices. (G) Total NMDA currents, normalized to the first 2.5 min of recording, were reduced similarly in WT and Dys−/− neurons. (H) Marked increases in total and ifenprodil-insensitive NMDA currents in Dys−/− neurons. Data for total, and non-NR2B NMDA currents was obtained by averaging the first and last 2.5 min of recordings, respectively. Data for NR2B currents was derived by subtracting the non-NR2B current from the total currents for each recording. (I) Decreased decay time constants of total (before ifenprodil) and NR2A (after ifenprodil) NMDA currents in Dys−/− neurons. The decay phase of NMDA currents was fitted by single-exponential functions to calculate decay time constants.

We next investigated the relative contributions of NR2A or NR2B to the changes in NMDAR synaptic currents. Application of the selective NR2B antagonist ifenprodil (3 μM) significantly reduced NMDAR-mediated synaptic currents in both WT and Dys−/− neurons (Fig. 4 E and F). Likewise, a more specific NR2B antagonist Ro25–6981 (1 μM) elicited a similar decrease in NMDAR-EPSCs in CA1 pyramidal neurons from both genotypes (Fig. S2C). Given that both antagonists provide a maximal and selective inhibition of NR1/NR2B diheteromeric receptors and that the majority of NMDARs in CA1 neurons contain NR2A and/or NR2B, the remaining (non-NR2B) currents reflected those mediated by NR2A-containing receptors. We found that hippocampal CA1 synapses in Dys−/− mice exhibited a much greater proportion of NR2A-mediated currents, compared with WT mice (Fig. 4F and Fig. S2C). The normalized NMDAR currents decreased within a few minutes following perfusion of ifenprodil or Ro25–6981 onto the slices (Fig. 4G and Fig. S2D). Quantitative analyses revealed a significant increase in the total NMDAR currents (averaged from the first 2.5 min of recording before ifenprodil or Ro25–6981 application) in Dys−/− neurons (Fig. 4H and Fig. S2E, Left columns). Further, NR2A-mediated EPSCs, obtained by averaging the last 2.5 min of the currents recorded in the presence of NR2B antagonists, also exhibited a big difference between WT and Dys−/− neurons (Fig. 4H and Fig. S2E, Middle columns). In contrast, when the NR2A-mediated currents were subtracted from the total NMDAR-EPSCs for each neuron, no significant difference was found in the ifenprodil- or Ro25–6981- sensitive, NR2B-mediated currents between WT and Dys−/− neurons (Fig. 4H and Fig. S2E, Right columns). Given that NR2A-mediated currents usually decay faster than NR2B-mediated currents (26), we further analyzed the decay time constants of NMDAR currents. Indeed, there was a significant decrease in the decay time of NMDAR-EPSCs in Dys−/− neurons, both before and after ifenprodil application (Fig. 4I). Taken together, these results suggest that the increase in NMDAR-EPSCs in Dys−/− neurons is due largely to elevated NR2A at the hippocampal CA1 synapses.

Elevated LTP at Dys−/− CA1 Synapses.

To investigate whether the changes in synaptic NR2A could alter synaptic plasticity in the hippocampus, we performed field recordings of excitatory postsynaptic potentials (fEPSPs) in hippocampal slices. Because long-term depression (LTD) is prominent earlier in development, 2-week-old animals were used. Conventional LTD, induced by low frequency stimulation (LFS, 1 Hz, 900 pulses), was not affected in the Dys−/− slices (Fig. 5 A and B). Control experiments showed no change in input-output of fEPSPs or paired-pulse facilitation (PPF) in Dys−/− mice at this age (Fig. 5 C and D). Long-term potentiation (LTP), induced by a single tetanus (4 TBS, 100 Hz) was investigated in slices derived from 8-week-old mice. We found that LTP was significantly increased in Dys−/− mice compared to WT control (Fig. 5 E and F). The fEPSP was 125 ± 5% in wild-type, but 151 ± 5% in Dys−/−, respectively (P < 0.003). However, input-output curves were normal in Dys−/− mice (Fig. 5H). PPFs were virtually identical over a wide range of interpulse intervals for Dys−/− and WT slices (Fig. 5G). These data suggest that the observed change in LTP in Dys−/− mice is due to a postsynaptic rather than presynaptic mechanism (27, 28).

Fig. 5.

Significant increase in LTP, but not LTD, in Dys−/− hippocampus. (A and B) Dys−/− mice display normal LTD. LTD was induced by 900 pulses at 1 Hz (black bar in B in the CA1 area of acutely prepared hippocampal slices from 2-week-old Dys−/− mice and WT littermates. Example fEPSP recordings before (lower curves) and 60 min after LTD induction (upper curves) are shown in A and the complete time courses are shown in B. LTD is comparable between Dys−/− and WT mice (69 ± 2% vs. 68 ± 3% at 70–75 min, respectively). (E and F) Slices from Dys−/− mice display elevated LTP. LTP was induced by 4 TBS (four bursts, each of four pulses at 100 Hz) (arrows in F) in 8-week-old animals. The magnitude of LTP was nearly double in Dys−/− that of WT mice (Dys−/−: 151 ± 5%, WT: 125 ± 5%, P < 0.003, t test). (C and G) Normal paired pulse facilitation (PPF). The ratios of the second and first EPSP slopes were calculated, and mean values are plotted against different interpulse intervals (IPI, 10 to 1,000 ms). (D and H) Normal basal synaptic transmission. Input-output curves were generated by plotting the postsynaptic response (initial slope of fEPSP) as a function of the presynaptic fiber volley amplitudes.

Discussion

Although dysbindin has been implicated in schizophrenia (1), little is known about its cellular function in neurons. By using a dysbindin null mutant mouse line (Dys−/−), we have uncovered an unexpected role of dysbindin in NMDAR regulation. A combination of imaging and biochemical approaches revealed that disruption of the dysbindin gene increases NR2A surface expression in neurons. This dramatic increase in NR2A surface expression was mitigated by introducing exogenous dysbindin in Dys−/− neurons, suggesting that dysbindin directly regulates NR2A surface expression, rather than indirectly by regulating neuronal development. In hippocampal slices from Dys−/− mice, NMDA-, but not AMPA-mediated synaptic currents were dramatically increased, and this effect is due largely to an elevated expression of NR2A at the excitatory synapses of CA1 pyramidal neurons. We have also observed a marked increase in LTP, but not LTD, at CA1 synapses in Dys−/− slices. Together, our data provide a link between dysbindin and NMDARs. These results have implications in synaptic plasticity and perhaps schizophrenia.

By using a variety of techniques, we demonstrated a specific enhancement of NR2A expression on the surface of neurons and at synapses. We did not examine surface expression of endogenous NR2A in our imaging experiments because specific antibodies against the N-terminal domain of NR2A or NR2B are not available. However, our surface biotinylation assays demonstrated that disruption of dysbindin resulted in more endogenous NR2A on the surface of cortical neurons. Electrophysiological recordings revealed an increase in NR2A-mediated synaptic currents in Dys−/− mice pyramidal neurons, suggesting that more endogenous NR2A has been incorporated into excitatory synapses. Unexpectedly, while no change in NR2B-mediated synaptic currents was observed in Dys−/− neurons, our biotinylation experiments also showed a reduction in NR2B on the surface of Dys−/− neurons. Given that NR2B is largely expressed in extrasynaptic sites (20, 22, 29, 30), it is possible that surface biotinylation assays revealed a decrease in extrasynaptic but not synaptic NR2B in Dys−/− neurons. Alternatively, NR2B surface expression in cortex may be different from that in hippocampus. Moreover, ifenprodil and Ro25–6981 are known to be most effective in blocking NR2B in NR1/NR2B diheteromeric NMDARs but less effective in blocking triheteromeric (NR1/NR2A/NR2B) NMDARs (31), thus the possibility of an increase in triheteromeric NMDA receptor in Dys−/− neurons cannot be excluded.

NMDARs undergo a subunit-specific postendocytic sorting, with NR2A being sorted to the lysosomal pathway, whereas NR2B preferentially traffics to recycling endosomes (23). It should also be noted that we do not know whether the increased surface expression of NR2A in Dys−/− hippocampal neurons is due to an increase in insertion or a decrease in endocytosis of NR2A. However, it is intriguing that dysbindin is thought to regulate protein sorting from early endosomes to lysosome-related organelles (LROs), which have recently been implicated in the delivery of membrane proteins to the cell surface (5, 6, 32, 33). Moreover, knockdown of dysbindin in rat cortical neurons has been shown to increase the surface expression of D2R, a subtype of dopamine receptor known to traffic to the lysosomal pathway (8–11). Thus, knockout of dysbindin may divert endocytosed membrane proteins normally destined to lysosomes back to the cell surface. Consistent with this idea, we found an enhancement in NR2A surface expression in neurons lacking dysbindin, whereas NR2B was not affected (Fig. 1). Furthermore, expression of exogenous dysbindin also reduced NR2A surface expression in both WT and Dys−/− neurons (Fig. 3). We also observed an increase in NR2A-mediated NMDAR currents in Dys−/− neurons (Fig. 4H and Fig. S2E), suggesting that dysbindin regulates the number of NR2A-containing NMDARs at synaptic sites. Together these results support a model in which dysbindin controls the amount of NR2A at excitatory synapses, potentially by regulating postendocytic sorting of NR2A from early endosomes into lysosomes or LROs.

Activation of NMDARs is required for both LTP and LTD in hippocampal CA1 synapses. Several studies suggest that NR2A-containing NMDARs are preferentially involved in LTP whereas NR2B-containing NMDARs are critical for LTD (34–36). Although this model remains controversial (37, 38), our findings revealed an increase in LTP in adult Dys−/− CA1 pyramidal neurons, whereas there is no change in LTD (Fig. 5). It is conceivable that an increase in the number of NR2A-containing NMDARs present in Dys−/− synapses leads to an enhancement in LTP without interfering with LTD.

Our finding that knockout of dysbindin leads to increased NR2A expression at synapses brings a twist to our understanding of the pathogenesis of schizophrenia. Several lines of evidence have pointed to a hypofunction of NMDAR in schizophrenia brains (12). In situ hybridization of postmortem brains revealed a reduced level of NMDARs, particularly NR1 in the hippocampus of schizophrenic patients (39, 40). Moreover, administration of NMDAR antagonists such as PCP or ketamine to healthy volunteers produces hallucination and psychosis (41, 42). These results at first glance appear at odds with our findings. However, methods used to detect NMDARs in the postmortem brains cannot distinguish whether changes in NMDARs occur at synaptic or extrasynaptic sites. It is also unclear whether systemic administration of NMDAR antagonists act on hippocampus or other brain regions to achieve their effects. Furthermore, whether and how impairments in hippocampal synaptic plasticity contribute to the pathogenesis of schizophrenia have not been established. We here show a specific increase in NR2A surface expression in Dys−/− hippocampus, most likely at excitatory synapses in pyramidal neurons, leading to a selective increase in LTP but not LTD. The specific increase in surface NR2A, but not NR2B, unavoidably changes the NR2A/NR2B ratio, which is thought to be important for regulating the critical period in the visual cortex as well as the threshold for bidirectional synaptic plasticity (43). An imbalance between NR2A and NR2B in the brain should therefore be considered as a potential factor involved in schizophrenia. Thus, our findings that dysbindin controls NR2A surface expression and hippocampal LTP open up an avenue for schizophrenia research.

Methods

Animals, Primary Cultures, and Transfection.

The use and care of animals in this study followed the guidelines of the National Institutes of Health Animal Research Advisory Committee. The Sandy mouse occurred spontaneously in the inbred DBA/2J strain at The Jackson Laboratory and carries an autosomal recessive mutation, which results in lighter coat color. The Sandy mouse expresses no dysbindin protein owing to a deletion in the gene dystrobrevin-binding protein 1 (Dtnbp1; encoding dysbindin). The mutation in Sandy mice (Dys−/−), was transferred to the C57BL/6J genetic background by >10 generations of backcrossing. Dys−/− mice and WT (+/+) littermates were bred by heterozygous (dys +/−) mating. Mice were identified by tail DNA genotyping. Primary hippocampal or cortical neurons were prepared from E18 WT or Dys−/− mouse embryos for biochemical and immunostaining experiments. Cultures were maintained for 14–16 days in vitro (DIV). Hippocampal neurons were transfected at DIV 9–10 using Lipofectamin 2000 (Invitrogen), and analyzed 5 days later.

Constructs and Antibodies.

GFP-NR2A and GFP-NR2B were obtained as gifts from Stefano Vicini (Georgetown University, Washington, DC), with a GFP tag inserted in the N-termini of NR2A and NR2B following the signal sequences. HA-dysbindin was generated in the lab, in the pcDNA3.1 vector, with HA tagged to the N terminus of dysbindin. Monoclonal rabbit anti-GFP antibody was obtained from Molecular Probes. Monoclonal mouse IgG₁ anti-HA antibody was obtained from Covance. Unconjugated rabbit IgG, anti-α-tubulin and anti-NR2B antibodies were obtained from Sigma. Anti-NR2A antibody was obtained from Upstate. Anti-NR1 antibody was obtained from BD PharMingen. Anti-GluR1 antibody was obtained from Millipore.

Additional methods can be found in SI Text.