Abstract
Regulation of neuronal NMDA receptors (NMDARs) by group I metabotropic glutamate receptors (mGluRs) is known to play a critical role in synaptic transmission. The molecular mechanisms underlying mGluR1-mediated potentiation of NMDARs are as yet unclear. The present study shows that in Xenopus oocytes expressing recombinant receptors, activation of mGluR1 potentiates NMDA channel activity by recruitment of new channels to the plasma membrane via regulated exocytosis. Activation of mGluR1α induced (1) an increase in channel number times channel open probability, with no change in mean open time, unitary conductance, or reversal potential; (2) an increase in charge transfer in the presence of NMDA and the open channel blocker MK-801, indicating an increased number of functional NMDARs in the cell membrane; and (3) increased NR1 surface expression, as indicated by cell surface Western blots and immunofluorescence. Botulinum neurotoxin A or expression of a dominant negative mutant of synaptosomal associated protein of 25 kDa molelcular mass (SNAP-25) greatly reduced mGluR1α-mediated potentiation, indicating that receptor trafficking occurs via a SNAP-25-mediated form of soluble N-ethylmaleimide sensitive fusion protein attachment protein receptor-dependent exocytosis. Because group I mGluRs are localized to the perisynaptic region in juxtaposition to synaptic NMDARs at glutamatergic synapses in the hippocampus, mGluR-mediated insertion of NMDARs may play a role in synaptic transmission and plasticity, including long-term potentiation.
- metabotropic glutamate receptors
- NMDA receptors
- receptor trafficking
- channel gating
- protein kinase C
- (1S,3R)-1-amino-cyclopentane-1,3-dicarboxylic acid
- ACPD
Activation of metabotropic glutamate receptors (mGluRs) modulates NMDA receptor (NMDAR) activity and is implicated in synaptic transmission and activity-dependent synaptic plasticity (Pin and Duvoisin, 1995; Conn and Pin, 1997; Nicoletti et al., 1999). mGluRs are encoded by a gene family of eight members; further diversity is generated by alternative RNA splicing of the receptor C-terminal domains. By sequence homology, agonist selectivity, intracellular signaling mechanisms, and differential targeting in neurons, mGluRs can be divided into three classes. Group I mGluRs (mGluR1 and mGluR5) couple to diverse intracellular signaling transduction pathways. Group I mGluRs can couple positively to phospholipase C, activation of which leads to stimulation of protein kinase C (PKC) and release of intracellular Ca2+ (Conn and Pin, 1997), or to adenylyl cyclase, activation of which stimulates cAMP formation (Aramori and Nakanishi, 1992; Joly et al., 1995). Group I mGluRs couple negatively to NMDARs via a G-protein-dependent, membrane-delimited signaling pathway in cortical neurons (Yu et al., 1997) and depolarize CA3 neurons via a G-protein-independent, Src-dependent signaling cascade (Heuss et al., 1999).
Immunocytochemical studies reveal strikingly different distributions of group I mGluRs within the hippocampus. Although mGluR1 is localized primarily to somatostatin-positive GABAergic interneurons, mGluR5 is highly expressed in pyramidal neurons (Shigemoto et al., 1997). At excitatory glutamatergic synapses, mGluR1 and mGluR5 are concentrated in the perisynaptic zone at the periphery of postsynaptic densities (PSDs; Baude et al., 1993; Lujan et al., 1996). Localization of group I mGluRs in close proximity to NMDARs and AMPA receptors (AMPARs), which are concentrated at PSDs, and to critical signaling proteins is consistent with a modulatory role in glutamatergic transmission.
Activation of group I mGluRs induces potentiation of NMDA EPSCs in hippocampal (Aniksztejn et al., 1992; Challiss et al., 1994), striatal (Pisani et al., 1997), and subthalamic nucleus (Awad et al., 2000) slices and potentiation of NMDA currents in lamprey motoneurons (Krieger et al., 2000) and frog spinal cord neurons (Holohean et al., 1999). However, the molecular mechanisms underlying potentiation of NMDARs by group I mGluRs are as yet unclear. Mice with null mutations of either mGluR1 (Aiba et al., 1994a,b; Conquet et al., 1994; Bordi, 1996) or mGluR5 (Lu et al., 1997) display reductions of hippocampal long-term potentiation (LTP) and abnormalities of motor coordination and associative learning. Metabotropic glutamate receptor signaling is also implicated in neurodegenerative diseases (Conn and Pin, 1997;Nicoletti et al., 1999), cortical development (Kaczmarek et al., 1997), and addiction (Wolf, 1998).
Activation of PKC induces rapid delivery of NMDARs to the cell surface of Xenopus oocytes and hippocampal neurons (Lan et al., 2001). The present study was undertaken to examine the hypothesis that potentiation of NMDARs by mGluR1 occurs at least in part by receptor trafficking. We used Xenopus oocytes to express a homogeneous population of receptors of known subtype in a geometrically simple system. Moreover, the molecular machinery for protein trafficking is highly conserved from yeast to mammals (Bennett and Scheller, 1993). Experiments involving patch-clamp recording, charge transfer measurements, cell surface Western blots, and immunofluorescence reveal that activation of mGluR1α promotes delivery of new NMDA channels to the plasma membrane by regulated exocytosis, a mechanism that accounts entirely for potentiation of NMDAR activity. These studies provide molecular insight into synaptic function of group I mGluRs and define a novel mechanism for potentiation of NMDAR function.
MATERIALS AND METHODS
Expression constructs. Rat NMDAR subunit NR1-4b (NR1100; Zukin and Bennett, 1995) cDNA was cloned in this laboratory (Durand et al., 1992); rat NR1-1a (NR1011) and mGluR1α cDNAs were gifts from Dr. S. Nakanishi (Kyoto University, Kyoto, Japan); mouse ε1, ε2, and ε3 (corresponding to rat NR2A, NR2B, and NR2C, respectively) cDNAs were gifts from Dr. M. Mishina (Niigata University, Tokyo, Japan). cDNAs were subcloned into the pBluescript SK(−) vector (Stratagene, La Jolla, CA) for oocyte expression. Synaptosomal associated protein of 25 kDa molecular mass (SNAP-25) and SNAP-25(Δ20) cDNAs were gifts from Dr. R. Y. Tsien (Howard Hughes Medical Institute, University of California, San Diego, CA). Capped mRNAs were synthesized as runoff transcripts from linearized plasmid cDNAs with T3 or T7 polymerase (mMessage mMachine transcription kit; Ambion, Austin, TX; 2 hr at 37°C). The concentration and integrity of mRNAs were assessed after staining with ethidium bromide by direct comparison of sample mRNAs with an RNA standard ladder (Life Technologies, Gaithersburg, MD).
Electrophysiology of recombinant NMDA receptors expressed inXenopus oocytes. Selected stage V and VI oocytes from adult female Xenopus laevis (Xenopus I, Ann Arbor, MI) were injected with a mixture of in vitrotranscribed mRNAs (20 ng of mRNA/cell; NR1:NR2:mGluR1α ratio, 1:2:1) using a Nanoject injector (Drummond Scientific, Broomall, PA) as described previously (Zheng et al., 1997). Oocytes were maintained at 18°C in culture buffer (in mm: 103 NaCl, 2.5 KCl, 2 MgCl2, 2 CaCl2, 5 HEPES, pH 7.5).
Whole-cell recording. Whole-cell currents were recorded from oocytes (2–6 d after injection) at ambient temperature using a GeneClamp 500 amplifier (Axon Instruments, Foster City, CA) by two-microelectrode voltage clamp, filtered at 20 Hz, and digitized on-line at 100 Hz (Zheng et al., 1997). Currents were elicited by bath application of NMDA (300 μm with 10 μmglycine) at a holding potential of −60 mV. Oocytes were perfused in Mg2+-free, normal frog Ringer's solution consisting of (in mm): 116 NaCl, 2.0 KCl, 1.0 CaCl2, and 10 HEPES, pH 7.2 (Zheng et al., 1997). Because Ca2+ influx through the NMDAR can cause Ca2+ amplification (Zheng et al., 1997), Ca2+ inactivation (Legendre et al., 1993), and activation of chloride channels endogenous to the oocyte (Leonard and Kelso, 1990) and thereby affect measurements of PKC potentiation, all experiments except those illustrated in Figures 1-3were performed in Ringer's solution in which CaCl2 was replaced by BaCl2. The pipette resistance ranged from 0.5 to 2 MΩ with an internal solution consisting of (in mm): 3000 KCl and 10 HEPES, pH 7.2. For mGluR1α activation, oocytes were incubated in (1S,3R)-1-amino-cyclopentane-1,3-dicarboxylic acid (ACPD; 100 μm, 2 min; Sigma, St. Louis, MO) and washed. ACPD potentiation is defined as the ratio of the NMDA-elicited current measured after ACPD application to that measured before ACPD application. For direct PKC activation, oocytes were incubated in 100 nm12-O-tetradecanoylphorbol 13-acetate (TPA; Sigma) for 10 min. TPA potentiation is defined as the ratio of the NMDA-elicited current measured after TPA application to that measured before TPA application.
Single-channel recording. Single-channel currents were recorded from outside-out patches excised from devitellinized oocytes (2–7 d after injection) as described previously (Araneda et al., 1999). Pipette resistance ranged from 7 to 12 MΩ with an internal solution consisting of (in mm): 115 CsCl, 5.0 EGTA, and 10 HEPES, pH 7.2. External solutions were as described for whole-cell recording to permit direct comparisons between NMDA-elicited whole-cell and single-channel currents. NMDA (100 μm with 10 μm glycine) solution was delivered to the patch from a multibarrel array fed by gravity.
Single-channel current amplitudes were determined from means of Gaussian fits to all-point amplitude histograms. The number of active channels times channel open probability (npo) was calculated as the total channel open time divided by recording time. Open time durations were calculated from single-channel openings above baseline but may include events from more than one channel in the same patch. All data are presented as mean ± SEM for 4–12 experiments performed with different oocytes. Statistical significance was assessed by the Student's t test (SigmaPlot 3.0) and ANOVA (CLR ANOVA 1.3).
Surface Western blot analysis. Control and ACPD-treatedXenopus oocytes expressing mGluR1α and NR1-4b/NR2A receptors were screened for NMDA currents in the range of 100–300 nA and ACPD potentiation of ∼2–3. Oocytes were incubated in external recording solution in the absence or presence of ACPD (100 μm, 2 min) and washed twice, and surface proteins were biotinylated with the membrane-impermeant reagent sulfosuccinimidyl 2-(biotinamido) ethyl-1,3′-dithiopropionate (sulfo-NHS-SS-biotin, 1.5 mg/ml, 30 min at 4°C; Pierce, Rockford, IL) according to the method of Chen et al. (1999). Cell extracts were prepared as described by Hollmann et al. (1994). To isolate biotinylated surface proteins from nonsurface proteins, cell extracts were incubated with Neutravidin-linked beads (Pierce) for 2 hr at 4°C, centrifuged, and washed. Bound proteins were eluted from beads by incubation with SDS-PAGE gel loading buffer containing DTT (which releases the proteins from the biotin moiety) and subjected to gel electrophoresis together with aliquots of total protein.
Cell surface immunolabeling of oocytes. Xenopusoocytes expressing NR1-4b/NR2A receptors with mGluR1α were labeled with monoclonal antibody 54.1 directed to the extracellular loop of the NR1 subunit (Gazzaley et al., 1996). Oocytes were screened for NMDA currents in the range of 100–300 nA and ACPD potentiation of approximately twofold to threefold. Oocytes were incubated in 20 ml of external recording solution in the absence or presence of ACPD (100 μm, 2 min), devitellinized, and fixed in 4% paraformaldehyde and 2% sucrose (1 hr at room temperature). Oocytes were then incubated with NR1 antibody (10 μg/ml, 4°C overnight), followed by biotinylated horse anti-mouse IgG (Vector Laboratories, Burlingame, CA; 10 μg/ml, 1.5 hr at room temperature) and FITC-conjugated avidin (Vector Laboratories; 10 μg/ml, 1.5 hr at room temperature). Oocytes were rinsed, placed in a chamber 1.5 mm deep with a coverslip forming the bottom (MatTek, Ashford, MA), and covered with ProLong mounting medium (Molecular Probes, Eugene, OR) to reduce fluorescence quenching. Cross-sectional and tangential images of oocytes were viewed by a Bio-Rad MRC 600 Kr/Ar laser scanning confocal microscope and acquired using COMOS software (Bio-Rad, Hercules, CA). FITC fluorescence was measured (excitation wavelength, 488 nm; emission, 515 nm) with a 40× objective lens (aperture, 1.3); laser intensity, photomultiplier gain, and pinhole aperture were kept constant.
RESULTS
Activation of mGluR1α potentiates NMDA responses
To examine the interaction between group I mGluRs and NMDARs, we coexpressed mGluR1α with NR1-4b/NR2A receptors in Xenopusoocytes and examined the effect of mGluR activation on NMDA responses by whole-cell recording under voltage clamp. NR1-4b/NR2A and -B have the highest cell surface expression (Okabe et al., 1999) and highest degree of PKC potentiation (Zheng et al., 1999). Application of the mGluR agonist ACPD (100 μm) in Ca2+-free Ringer's solution to the oocyte elicited a large inward current (∼750 nA) that decayed to near baseline in <30 sec (Fig.1A). The ACPD-elicited current is ascribable to Cl− efflux through Ca2+-activated Cl− channels endogenous to the oocyte (Saugstad et al., 1996). It is well documented that activation of mGluR1 can lead to an increase of [Ca2+]i by activation of phospholipase C and release of Ca2+ from inositol triphosphate (IP3)-sensitive intracellular stores (Pin and Duvoisin, 1995). Hence, the ACPD-evoked current indicates effective expression of mGluR1α and targeting of functional receptors to the cell surface of the oocyte.
To examine the effect of mGluR1α activation on NMDAR responses, we recorded whole-cell currents elicited by NMDA (300 μmwith 10 μm glycine) before and after bath application of ACPD (100 μm, 2 min) to the oocyte. Treatment with ACPD significantly increased NMDA-elicited responses (Fig.1A; compare first and second NMDA responses). In Ca2+ Ringer's solution, ACPD potentiation (defined as the ratio of NMDA current amplitude measured after ACPD to that measured before ACPD) was 2.97 ± 0.30 (n = 7) (Fig. 1C).
To investigate the effect of alternative splicing of the NR1 subunit on mGluR1-mediated potentiation of NMDAR responses, we coexpressed mGluR1α with either NR1-1a/NR2A (NR1011/NR2A) or NR1-4b/NR2A (NR1100/NR2A) receptors inXenopus oocytes (Fig. 1B). NR1-1a/NR2A and -B have the lowest cell surface expression (Okabe et al., 1999) and lowest degree of PKC potentiation (Zheng et al., 1999). Application of ACPD potentiated NMDA responses of NR1-4b/NR2A receptors to 2.97 ± 0.30 times the control response in Ca2+Ringer's solution (n = 7) (Fig.1A,C). ACPD potentiated responses of NR1-1a/NR2A receptors to 2.03 ± 0.18 times the control response in Ca2+ Ringer's solution (n= 7) (Fig. 1C). Differences in ACPD-induced potentiation between splice variants did not reach statistical significance in either Ca2+- containing or Ca2+-free (Ba2+) Ringer's solution (one-way ANOVA followed by Bonferroni test, p > 0.05 for comparisons between each receptor pair). The smaller magnitude of the mGluR1α potentiation may have obscured an effect of NR1 splicing similar to that observed for PKC potentiation (Zheng et al., 1999).
To investigate the effects of the NR2 subunit on ACPD-induced potentiation, we coexpressed mGluR1α with NR1-4b/NR2A, NR1-4b/NR2B, or NR1-4b/NR2C receptors in Xenopus oocytes. The mGluR1 agonist ACPD (100 μM) significantly potentiated NMDA-induced currents of NR1-4b/NR2A and NR1-4b/NR2B receptors (n = 3) (Fig.2A,B). Potentiation was to 3.33 ± 0.88 (n = 3) (Fig.1A) and 2.57 ± 0.14 (n = 3) times the control responses in Ca2+Ringer's solution, respectively (Fig. 2D). The degree of ACPD potentiation did not differ significantly for NR1-4b/NR2A versus NR1-4b/NR2B receptors. As observed for TPA potentiation (Mori et al., 1993), ACPD did not potentiate responses of NR1-4b/NR2C receptors (n = 3) (Fig. 2C); potentiation was to 1.02 ± 0.04 times the control response in Ca2+ Ringer's solution (n= 3) (Fig. 2D). Similar results were observed for the three receptor subtypes in Ba2+ Ringer's solution. Thus, the NR2 subunit composition of NMDAR affected the degree of ACPD potentiation.
To rule out the possibility that ACPD directly activates NMDA receptors (the mGluR agonist 2-(2′,3′)-dicarboxycyclopropylglycine at high concentrations may act on the NMDA receptor as an agonist) (Opitz et al., 1994; Yu et al., 1997), we applied a high concentration of ACPD (100 μm) to oocytes expressing NR1-4b/NR2A receptors in the absence of mGluR1α. Under these conditions, ACPD elicited no response and did not potentiate NMDA currents (data not illustrated), indicating that the Cl− response was mGluR-dependent and that ACPD did not activate or potentiate NMDA channels directly.
ACPD increases npo but not conductance or open time
To examine effects of mGluR1 activation on NMDA channel conductance and gating, we recorded channel activity in outside-out patches excised from oocytes expressing mGluR1α and NR1-4b/NR2A receptors before and after ACPD treatment. We compared patches excised before and after ACPD treatment, because patch formation inhibits exocytosis, presumably by deformation of the membrane and disruption of association with cytoplasmic elements (Lan et al., 2001)
In control patches, NMDA (10 μm) activated single channels with a conductance of γ = 44 ± 1 pS at −60 mV, which did not vary with voltage (n = 5) (Fig.3A,D). Occasional transitions from the main conductance state to sublevels (total dwell time of <1% of that of the main state; data not shown) were excluded from the analysis. NMDA channel activity in patches excised after application of ACPD (100 μm, 2 min) was markedly potentiated (Fig. 3A,B). ACPD increasednpo by approximately fourfold from 0.050 ± 0.006 before ACPD (n = 5) to 0.209 ± 0.027 after ACPD (n = 5; p < 0.001) (Fig. 3C).
mGluR1 activation did not alter NMDA single-channel conductance (n = 5) (Fig. 3A,B) or reversal potential (Erev ∼ 0), as evidenced by comparison of current–voltage relationships (n = 5) (Fig. 3D). Moreover, ACPD did not significantly change the mean duration of openings (Fig. 3E,F). The distribution of open times was fit by a single exponential, consistent with the presence of a single open state (Fig. 3E,F) (τ = 6.35 ± 0.26 msec before ACPD; n = 5; vs τ = 6.08 ± 0.29 msec after ACPD; n = 5). Potentiation of npo without change in γ was also induced by the mGluR agonist (S)-3,5-dihydroxyphenylglycine (100 μm; data not shown).
mGluR1 activation induces an increase in the number of active NMDA channels
The results reported thus far indicate that ACPD increases the number of active channels in the cell membrane times the open probability. To distinguish between effects of mGluR1 activation on the number of functional channels in the membrane, n, and channel open probability, po, we recorded NMDA whole-cell currents in the presence of MK-801 (5 μm) from control (Fig.4B, left trace) and ACPD-treated (Fig. 4B, right trace) oocytes by a modification of the method of Jahr (1992) as adapted by Rosenmund et al. (1995). This method takes advantage of the essentially irreversible block of NMDA-elicited currents by the open-channel blocker MK-801. To determine the number of channels in the whole oocyte, N, we calculated the cumulative charge transfer, Q, which is the integral of NMDA-elicited current during the time required for complete block by MK-801. The number of channels, N, can be calculated from Q as follows: where tbl is the time constant for MK-801 block (tbl = 1/kbl [MK-801], andkbl = 2.5 × 107m/sec; Jahr, 1992). From the single-channel recordings (Fig. 3), ACPD does not change single-channel conductance, and we would not expected it to change tbl. The channel number,N, for control and ACPD-treated oocytes was normalized to the NMDA-elicited whole-cell current, which corrects for any differences in levels of expression within and between the two groups. For control oocytes, which exhibited control currents in the range of 100–300 nA, the mean number of channels per 100 nA was 6.8 ± 0.8 × 105, and the mean channel density for 100 nA was 0.02 μm−2, assuming a surface area for the oocyte of 3 × 107 μm2(Zampighi et al., 1999). ACPD increased N ∼1.6-fold (NACPD/Ncontrol= 1.6; p < 0.01). This analysis indicates that mGluR1 activation increases the number of functional NMDA channels per cell.
The data in Figure 4B also enabled us to analyze the effect of mGluR1 activation on NMDA channel gating. We first calculated channel open probability in control and ACPD-treated oocytes from the NMDA-elicited whole cell current, I, the single-channel current, i (which is not changed by mGluR1 activation) (Fig.3D), and the number of channels per cell, N, as follows: from which po, ACPD/po, control = 1.2. The ACPD-induced increase in channel open probability was not significant.
As an independent measure of the effect of mGluR1 activation on channel gating, we analyzed the decay of NMDA-elicited current in the presence of MK-801 from the data in Figure 4B normalized to the same peak amplitude (Fig. 4C). Provided thatkbl [MK-801] ≫kα andkβ (the closing and opening rates, respectively), the decay can be described by a single exponential with a rate constant, kβ, for channel opening. Activation of mGluR1 did not significantly changekβ (1.9 ± 0.2/sec for control oocytes vs 2.1 ± 0.2/sec for ACPD-treated oocytes) (Fig.4G). From measurements ofpo andkβ and the relationpo =kβ/(kα+ kβ), we calculated values ofkα = 12.7 ± 1.5/sec in control oocytes versus 11.1 ± 0.8/sec in ACPD-treated oocytes (Fig.4H). These values ofkα are substantially less than the value of ∼167/sec predicted from the ∼6 msec mean open time of single channels. The apparent discrepancy arises because thepo equation as applied to macroscopic current assumes an equilibrium between unliganded and bursting states and thus will yield a kα value corresponding to termination of bursting rather than of individual openings. Moreover, kα calculated in this manner will be overestimated in that the probability that a receptor is liganded is greater than the probability that it is open,po, calculated from Nand I.
mGluR1 activation delivers NMDA channels to the cell membrane via exocytosis
The results reported thus far indicate that mGluR1 activation induces an increase in the number of active channels at the cell surface but does not distinguish between insertion of new NMDA channels and unmasking of silent channels. To distinguish between these possibilities, we performed two additional experiments. First, we examined the effects of loading oocytes with the light chains of type A botulinum neurotoxin (BoNT A), which is known to inactivate SNAP-25 and to prevent SNAP-25-dependent exocytosis (Montecucco and Schiavo, 1995). Treatment of oocytes with BoNT A reduced the degree of mGluR1-mediated potentiation of NMDA-elicited currents by ∼50% (Fig.5A–C). A mixture of type A, B, and E BoNTs reduced ACPD potentiation by ∼60% (Fig.5C). No significant change in the resting potential, input resistance, and basal NMDA response was caused by BoNTs (data not shown). BoNT A also reduces TPA potentiation by ∼50% (Lan et al., 2001). No effects on mGluR1-mediated potentiation were observed when only the vehicle (10 mm DTT) was injected. Presumably, toxin treatment attenuated the number of new channels inserted by mGluR1 activation.
Second, we examined the ability of a dominant negative mutant of SNAP-25 to block mGluR1-mediated potentiation. SNAP-25(Δ20), a truncation mutant lacking the C-terminal 20 amino acids, corresponds to yeast sec9-Δ17, a dominant negative mutant of the yeast SNAP-25 homolog (Yao et al., 1999). Mutant and wild-type SNAP-25 were expressed in oocytes by injection of their cRNAs 48 hr after injection of NMDAR cRNAs, and ACPD potentiation was measured ∼20 hr later. ACPD-induced potentiation of NMDA currents in control oocytes (2.5 ± 0.3;n = 3) (Fig.6A,D) and in oocytes expressing full-length SNAP-25 (2.9 ± 0.2; n = 4) (Fig. 6B,D) was greater than in oocytes expressing SNAP-25(Δ20) (1.5 ± 0.2; n = 5;p < 0.01) (Fig. 6C,D). Expression of SNAP-25(Δ20) did not affect the NMDA-elicited current or endogenous Ca2+-activated Cl− currents. These findings indicate that NMDA channels inserted at the cell surface by mGluR1α activation are delivered, at least in part, by SNAP-25-mediated solubleN-ethylmaleimide sensitive fusion protein attachment protein receptor (SNARE)-dependent exocytosis.
Although the data indicate that ACPD increases the number of functional NMDA channels at the cell surface via a SNARE-dependent mechanism, they do not exclude a contribution from reduction in internalization (or another mechanism of silencing). As an estimate of the rate of constitutive exocytosis in the oocyte, we measured reappearance of functional channels at the cell surface after quasi-irreversible block by MK-801. MK-801 (1 μm) blocked completely the NMDA-elicited current (Fig.7A, first response). A test application of NMDA at 3 min elicited a small response, indicative of a small number of newly inserted channels, which were opened by the agonist (Fig. 7A, second response). The small size of the test response is indicative of a relatively slow rate of constitutive exocytosis and is consistent with the half-time of ∼40 min reported for recovery of the acetylcholine response after irreversible block of nicotinic acetylcholine receptors expressed in Xenopusoocytes (Akabas et al., 1992). Because under steady-state conditions, the rate of internalization equals the rate of exocytosis, the rate of internalization is slow (Fig. 7A), and a reduction in the rate of internalization (or of silencing) could account for, at most, a very small fraction of the potentiation by ACPD.
As a further test of ACPD-induced recruitment of new channels to the cell surface, we blocked NMDA currents with MK-801 as in Figure7A and then applied ACPD in the presence of MK-801 (Fig.7B). In these oocytes, a test application of NMDA elicited a much larger response, ∼4.5 times that observed in control oocytes at 3 min after MK-801 block (Fig. 7A) (p< 0.01) and approximately the same size as the control response (p < 0.001) (Fig. 7C). If ACPD potentiation were attributable entirely to insertion of new channel molecules (Fig. 4), and constitutive insertion were negligible, the amplitude of the test NMDA current after ACPD application in the presence of MK-801 (IACPD, MK-801) would equal the amplitude of the test NMDA current after ACPD alone (IACPD) minus the amplitude of the initial control current (Icontrol).IACPD, MK-801 was slightly smaller than the predicted value, possibly an effect of residual MK-801 (Fig.5D, bars 1, 3). These findings demonstrate that ACPD causes a rapid increase in the number of functional NMDARs in the cell surface.
ACPD increases surface expression of NMDARs
The electrophysiological experiments presented thus far argue for mGluR1-mediated increase in the number of active channels via exocytosis but do not provide a measure of the total number of channel molecules expressed at the surface. To examine the effect of mGluR1α on the surface expression of NMDARs, we performed cell surface Western blots by a modification of the method of Chen et al. (1999). Control and ACPD-treated oocytes expressing mGluR1α and NR1/NR2A receptors were surface-labeled with sulfo-NHS-SS-biotin, and surface proteins were separated from intracellular proteins by reaction with Neutravidin beads and centrifugation. Samples of surface and total cell protein were treated with DTT (which releases the biotin moiety from surface proteins) and subjected to electrophoresis. Membranes were probed with antibody 54.1, a monoclonal antibody directed to the extracellular loop of the NR1 subunit (Gazzaley et al., 1996) (Fig.8A). Analysis of band densities indicated an increase in surface NR1 expression to 1.7 ± 0.2 times control (n = 4; p < 0.01) (Fig. 8B,D), with no change in total cell NR1 protein (Fig. 8C).
In addition, we assessed the effect of mGluR1α activation on receptor surface expression by immunofluorescence. Control and ACPD-treated nonpermeabilized oocytes expressing NR1-4b/NR2A and mGluR1α receptors were reacted with antibody 54.1, followed by a biotinylated secondary antibody and fluorescein-conjugated avidin. Optical sections through the oocyte were viewed by confocal microscopy and captured by COMOS software. Control oocytes expressing NMDARs exhibited clear immunofluorescence, which appeared concentrated at the external surface (Fig. 9A,C). ACPD dramatically increased NR1 surface immunolabeling (Fig. 9B,D). Little or no immunofluorescence was observed for oocytes reacted with the FITC-tagged secondary antibody in the absence of the primary antibody (Fig. 9E) or for water-injected oocytes labeled with the N-terminal antibody (Fig. 9F). These findings indicate specificity of the immunofluorescence labeling. ACPD did not increase the resting conductance of the oocytes after the Ca2+-activated Cl− current subsided. Moreover, control experiments indicate that TPA treatment does not increase the permeability of the oocyte membrane to the antibody (Lan et al., 2001). We therefore infer that ACPD is unlikely to affect the permeability of the oocyte membrane.
DISCUSSION
Modulation of ligand-gated channels by G-protein-linked receptors is likely to be an important mode of receptor regulation under physiological conditions. The present study demonstrates that mGluR1α-mediated potentiation of NMDARs in Xenopus oocytes is mediated by recruitment of new channel molecules to the cell surface via regulated exocytosis. We show that loading cells with BoNT A, which interferes with exocytosis by cleaving SNAP-25, markedly reduces ACPD potentiation of NMDAR responses. Moreover, expression of a dominant negative mutant of SNAP-25, SNAP-25(Δ20), nearly abolishes ACPD potentiation. Specificity of mGluR1α-stimulated insertion of NMDARs at the cell surface is indicated by the observation that NR1/NR2A and NR1/NR2B but not NR1/NR2C receptors exhibit ACPD-induced potentiation. The present study is the first to examine biophysical mechanisms at the single-channel level in a system in which activation of group I mGluRs induces potentiation of NMDAR activity. We show that activation of mGluR1α increases the number of functional NMDA channels at the cell surface, with no change in single-channel conductance, mean open duration, or open probability. Our study is also the first to demonstrate mGluR-mediated rapid trafficking of NMDARs in oocytes via a SNARE-dependent mechanism. A simple scenario is that activation of mGluR1α regulates membrane fusion events of NMDAR-containing vesicles by activation of endogenous protein kinases, which in turn phosphorylate a protein involved in receptor trafficking. One potential target is SNAP-25, which can be directly phosphorylated in an activity-dependent manner (Shimazaki et al., 1996; Genoud et al., 1999). The intracellular signaling pathway linking mGluR1 activation with an increase in regulated exocytosis is, however, unknown, and association of SNAP-25 or its binding partners with NMDARs remains to be established.
Recent studies by our laboratory show that activation of PKC induces rapid delivery of NMDARs to the cell surface of oocytes and hippocampal neurons and an increase in the NMDA channel opening rate (Lan et al., 2001). Several findings indicate that the molecular mechanisms underlying mGluR1-mediated potentiation of NMDARs differ from those of PKC potentiation. First, ACPD regulates receptor trafficking in the absence of a significant effect on channel gating (although the present data cannot exclude a proportionately smaller change in the channel opening rate). Second, whereas the degree of TPA potentiation is markedly affected by NR1 splicing (Zheng et al., 1999), the degree of ACPD potentiation is affected only modestly by NR1 splicing, an effect that was not statistically significant in the present study. Third, whereas PKC-induced potentiation of NMDAR activity is observed in oocytes and dissociated hippocampal neurons, ACPD potentiation is observed in oocytes and in hippocampal slices but not in dissociated neurons, suggesting that the intracellular signaling pathway linking mGluR activation to NMDAR potentiation requires preservation of neuronal circuitry (P. J. Conn, personal communication).
Finally, group I mGluRs couple to diverse intracellular signaling pathways and potentiate NMDA EPSCs and NMDA-elicited currents by PKC-independent signal transduction in some neuronal cell types and preparations and in expression systems. Electrophysiological studies of CA3 pyramidal neurons from rat hippocampal slice cultures indicate that synaptic activation of group I mGluRs (mGluR1) by mossy fiber stimulation evokes an EPSC by a G-protein-independent signaling pathway (Heuss et al., 1999). These studies show that an Src-family tyrosine kinase is an integral component in a signaling pathway that functionally links mGluR1 activation with a transient cationic conductance increase. Because NR1/NR2A NMDARs are functionally modified by tyrosine kinases such as Src (Wang and Salter, 1994), it is reasonable to propose that mGluR1 potentiates NMDA currents via activation of Src. Consistent with a role for tyrosine kinase-dependent modulation of NMDAR activity, we recently reported that insulin promotes rapid exocytosis of NMDARs in Xenopus oocytes (Skeberdis et al., 2001a). Alternatively, group I mGluRs might signal via a membrane-delimited signaling pathway, as has been observed for mGluR-mediated inhibition of NMDAR activity in embryonic mouse cortical neurons (Yu et al., 1997).
The role of PKC-mediated signaling cascades in mGluR potentiation of NMDARs is, at best, controversial. mGluR-mediated potentiation of NMDA EPSCs is reported to occur independently of PKC signaling in hippocampal (Harvey and Collingridge, 1993) and cerebellar slice preparations (Kinney and Slater, 1993), as is group I mGluR-mediated potentiation of NMDA currents in lamprey motoneurons (Krieger et al., 2000) and frog spinal cord neurons (Holohean et al., 1999). On the other hand, the intracellular pathway linking mGluRs to potentiation of NMDA responses appears to involve PKC in at least two preparations, although neither the mGluR nor NMDAR subtype or receptor class was identified in these preparations. PKC inhibitors block at least in part mGluR-mediated potentiation of NMDA EPSCs in hippocampal slice (Aniksztejn et al., 1992) and potentiation of NMDA currents in oocytes expressing rat brain mRNA (Kelso et al., 1992). Thus, multiple pathways might contribute to the group I mGluR-dependent regulation of NMDARs observed in the present study. This concept is supported by pharmacological studies; whereas mGluR-dependent potentiation of NMDARs is only partly blocked by selective PKC inhibitors, it is completely blocked by the broad-spectrum kinase inhibitor staurosporine (Skeberdis et al., 2001b).
mGluR potentiation is via regulated exocytosis
In the present study, ACPD potentiation of NMDA-elicited currents was inhibited by ∼50% injection of BoNT A and nearly completely abolished by expression of SNAP-25(Δ20) (Figs. 5, 6). These findings provide important new evidence for a mechanism involving regulated exocytosis of NMDARs to the cell surface of the oocyte. Inhibition of exocytosis by BoNT A is incomplete in a number of other systems, including exocytosis of Isoc channels (Yao et al., 1999) and NMDARs in Xenopus oocytes (Lan et al., 2001). The action of BoNT A on regulated exocytosis of NMDARs was relatively specific in that (1) ACPD-induced insertion of new channel molecules was not inhibited by BoNT B or E; and (2) constitutive expression of NMDARs and endogenous Ca2+-activated Cl− channels (Yao et al., 1999; our unpublished observations) and transfected epithelial Na+ channels (Yao et al., 1999) in oocytes is not reduced by BoNT A.
Structural basis of localization of group I mGluRs and NMDARs
Immunogold studies reveal a highly ordered arrangement of group I mGluRs and NMDARs at the postsynaptic membrane of CA1 synapses and indicate that the two classes of receptors are concentrated in microdomains in close proximity to one another (Baude et al., 1993;Lujan et al., 1996; Lujan et al., 1997; Shigemoto et al., 1997). Studies involving the yeast two-hybrid system and immunoprecipitation suggest that mGluR1α and mGluR5 may be cross-linked to NMDARs at synapses. The C-terminal tails of mGluR1 and mGluR5 contain a recognition sequence for binding by members of the Homer family of anchoring proteins (Tu et al., 1998). Homer proteins, in turn, link via Shank and PSD to each other and to IP3 receptors, NMDARs, and AMPARs (Tu et al., 1999). Thus, a scaffolding structure consisting of Homer and Shank may directly connect group I mGluRs to NMDARs, thereby providing strategic positioning for regulation of NMDARs by mGluRs at hippocampal synapses.
Physiological significance of mGluR1 potentiation of NMDARs
The physiological relevance of mGluR1 actions is underscored by observations that activation of group I mGluRs induces potentiation of NMDA EPSCs in rat hippocampal (Aniksztejn et al., 1992; Challiss et al., 1994), striatal (Pisani et al., 1997), and subthalamic nucleus (Awad et al., 2000) slices and potentiation of NMDA currents in lamprey motoneurons (Krieger et al., 2000) and frog spinal cord neurons (Holohean et al., 1999). Turnover and biotinylation studies of cerebellar granule neurons indicate a large pool of NR1 subunits (∼60% total NR1) in the cytoplasm of dendritic shafts and spines, where they are poised for rapid assembly with NR2 subunits and insertion at synaptic sites (Huh and Wenthold, 1999). Rapid trafficking of AMPARs between the plasma membrane and an intracellular compartment in response to synaptic activity (Nishimune et al., 1998; Osten et al., 1998; Song et al., 1998) and synaptic plasticity (Carroll et al., 1999) is thought to represent recycling.
Findings from the present study suggest a mechanism whereby group I mGluRs can augment the number of synaptic NMDARs, thereby modulating neuronal excitability and lowering the threshold for LTP and long-term depression. Moreover, activation of NMDARs potentiates mGluR5 activity (Challiss et al., 1994; Luthi et al., 1994; Alagarsamy et al., 1999); mechanisms implicated in this potentiation include a rise in intracellular Ca2+, activation of calcineurin, and inhibition of PKC (De Blasi et al., 2001). Thus, positive interactions between mGluRs and NMDARs may be reciprocal in nature and important to NMDAR-dependent LTP. The present study extends earlier studies by providing insight into the biophysical and molecular mechanisms underlying mGluR1-mediated regulation of NMDARs in a pure population of receptors. Insertion and retrieval of NMDARs to and from the plasma membrane are likely to be common and powerful mechanisms for regulating excitatory synaptic transmission.
Footnotes
This work was supported by National Institutes of Health Grants NS 20752 and NS 31282 (R.S.Z.) and NS 07512 (M.V.L.B). M.V.L.B is the Sylvia and Robert S. Olnick Professor of Neuroscience. We thank Alice P. Wang for technical support. We acknowledge the Analytical Imaging Facility of the Albert Einstein College of Medicine (Michael Cammer, director).
Correspondence should be addressed to Dr. R. Suzanne Zukin, Department of Neuroscience, Albert Einstein College of Medicine, 1300 Morris Park Avenue, Bronx, NY 10461. E-mail: zukin{at}aecom.yu.edu.