The Metabotropic Glutamate Receptor mGlu7 Activates Phospholipase C, Translocates Munc-13-1 Protein, and Potentiates Glutamate Release at Cerebrocortical Nerve Terminals*

At synaptic boutons, metabotropic glutamate receptor 7 (mGlu7 receptor) serves as an autoreceptor, inhibiting glutamate release. In this response, mGlu7 receptor triggers pertussis toxin-sensitive G protein activation, reducing presynaptic Ca2+ influx and the subsequent depolarization evoked release. Here we report that receptor coupling to signaling pathways that potentiate release can be seen following prolonged exposure of nerve terminals to the agonist l-(+)-phosphonobutyrate, l-AP4. This novel mGlu7 receptor response involves an increase in the release induced by the Ca2+ ionophore ionomycin, suggesting a mechanism that is independent of Ca2+ channel activity, but dependent on the downstream exocytotic release machinery. The mGlu7 receptor-mediated potentiation resists exposure to pertussis toxin, but is dependent on phospholipase C, and increased phosphatidylinositol (4,5)-bisphosphate hydrolysis. Furthermore, the potentiation of release does not depend on protein kinase C, although it is blocked by the diacylglycerol-binding site antagonist calphostin C. We also found that activation of mGlu7 receptors translocate the active zone protein essential for synaptic vesicle priming, munc13-1, from soluble to particulate fractions. We propose that the mGlu7 receptor can facilitate or inhibit glutamate release through multiple pathways, thereby exerting homeostatic control of presynaptic function.

At synaptic boutons, metabotropic glutamate receptor 7 (mGlu7 receptor) serves as an autoreceptor, inhibiting glutamate release. In this response, mGlu7 receptor triggers pertussis toxin-sensitive G protein activation, reducing presynaptic Ca 2؉ influx and the subsequent depolarization evoked release. Here we report that receptor coupling to signaling pathways that potentiate release can be seen following prolonged exposure of nerve terminals to the agonist L-(؉)-phosphonobutyrate, L-AP4. This novel mGlu7 receptor response involves an increase in the release induced by the Ca 2؉ ionophore ionomycin, suggesting a mechanism that is independent of Ca 2؉ channel activity, but dependent on the downstream exocytotic release machinery. The mGlu7 receptor-mediated potentiation resists exposure to pertussis toxin, but is dependent on phospholipase C, and increased phosphatidylinositol (4,5)-bisphosphate hydrolysis. Furthermore, the potentiation of release does not depend on protein kinase C, although it is blocked by the diacylglycerolbinding site antagonist calphostin C. We also found that activation of mGlu7 receptors translocate the active zone protein essential for synaptic vesicle priming, munc13-1, from soluble to particulate fractions. We propose that the mGlu7 receptor can facilitate or inhibit glutamate release through multiple pathways, thereby exerting homeostatic control of presynaptic function.
Metabotropic glutamate receptors belong to the G proteincoupled receptors (GPCRs) 2 superfamily and their eight recep-tor subtypes (mGlu1-8 receptors) are classified into three major groups. Most group III mGlu receptors (mGlu4, -6, -7, and -8 receptors) are located within the presynaptic active zone (1) where they act as autoreceptors mediating feedback inhibition of glutamate release (2)(3)(4). The signaling mechanism initiated by mGlu7 receptors to inhibit neurotransmitter release involves the activation of G i/o proteins that inhibit Ca 2ϩ channels and adenylyl cyclase (5), and probably the release process itself (6). However, mGlu7 receptor signaling is not restricted to these pathways. Thus, transfected mGlu7 receptors expressed in cerebellar granule cells inhibits somatic Ca 2ϩ currents by a mechanism that involves the activation of phospholipase C (PLC) and the hydrolysis of phosphatidylinositol (4,5)-bisphosphate thereby generating inositol trisphosphate that releases Ca 2ϩ from intracellular stores and diacylglycerol (DAG) that activates protein kinase C (PKC) (7). However, one important question that remains to be resolved is whether endogenous mGlu7 receptors at synaptic sites also signal via PLC and if so, what effect such signaling has on release modulation.
Phorbol esters, stable analogues of the endogenous product of PLC, DAG, potentiate synaptic transmission by increasing neurotransmitter release (8 -15). DAG signaling at synapses has long been thought to be mediated by PKC and both presynaptic K ϩ and Ca 2ϩ channels, as well as other proteins of the release machinery, have been identified as PKC substrates. In addition to PKC activation, phorbol esters can also activate munc13-1, a presynaptic protein with an essential role in synaptic vesicle priming (16,17). However, there is little information regarding the presynaptic receptors coupled to signaling pathways involved in synaptic potentiation. Presynaptic PLC activation in response to high frequency stimulation results in synaptic potentiation due to increased neurotransmitter release in hippocampal neurons (18). As high frequency stimulation increases synaptic glutamate enough to activate low affinity mGlu7 receptors (19), it remains unclear whether mGlu7 receptors may activate presynaptic PLC and whether this action enhances glutamate release.
Here we show, in a preparation of cerebrocortical nerve terminals, that mGlu7 receptors inhibit the release process, but can also potentiate it under specific conditions. Indeed, release potentiation is observed after a prolonged exposure of the receptor to agonist L-AP4. The new signaling elicited by the receptor involves PLC activation via a pertussis toxin (PTX)insensitive G protein and the subsequent hydrolysis of phosphatidylinositol (4,5)-bisphosphate. The mGlu7 receptor-dependent response potentiates release by a mechanism that could be dependent on the downstream release machinery because a parallel translocation of the active zone munc13-1 protein from the soluble to particulate fractions is observed upon receptor activation.

EXPERIMENTAL PROCEDURES
Synaptosomal Preparation-The handling and all procedures to sacrifice the animals used in this study were performed in accordance to the European Commission guidelines (86/ 609/CEE) and approved by the Animal Research Committee at the Complutense University. Synaptosomes were purified on discontinuous Percoll gradients (Amersham Biosciences) as described previously (5). Briefly, the cerebral cortex was isolated from adult male Wistar rats (2-3 months old) and homogenized in medium containing 0.32 M sucrose (pH 7.4). The homogenate was centrifuged for 2 min at 2,000 ϫ g and 4°C, and the supernatant spun again at 9,500 ϫ g for 12 min. From the pellets formed, the white loosely compacted layer containing the majority of the synaptosomes was gently resuspended in 8 ml of 0.32 M sucrose (pH 7.4). An aliquot of this synaptosomal suspension (2 ml) was placed onto a 3-ml Percoll discontinuous gradient containing: 0.32 M sucrose, 1 mM EDTA, 0.25 mM DLdithiothreitol, and 3, 10, or 23% Percoll (pH 7.4). After centrifugation at 25,000 ϫ g for 10 min at 4°C, the synaptosomes were recovered from the 10 and 23% Percoll bands, and they were diluted in a final volume of 30 ml of HEPES buffer medium (HBM): 140 mM NaCl, 5 mM KCl, 5 mM NaHCO 3 , 1.2 mM NaH 2 PO 4 , 1 mM MgCl 2, 10 mM glucose, and 10 mM HEPES (pH 7.4). Following further centrifugation at 22,000 ϫ g for 10 min, the synaptosome pellet was resuspended in 6 ml of HBM and the protein content was determined by the Biuret method. Finally, 1 mg of the synaptosomal suspension was diluted in 2 ml of HBM and spun at 3,000 ϫ g for 10 min. The supernatant was discarded and the pellets containing the synaptosomes were stored on ice. Under these conditions the synaptosomes remain fully viable for at least 4 -6 h, as judged by the extent of KCl-evoked glutamate release.
Glutamate Release-Glutamate release was assayed by online fluorimetry as described previously (5). Synaptosomal pellets were resuspended in HBM (0.67 mg/ml) and preincubated at 37°C for 1 h in the presence of 16 M bovine serum albumin (BSA) to bind any free fatty acids released from synaptosomes during the preincubation (20). A 1-ml aliquot was transferred to a stirred cuvette containing 1 mM NADP ϩ , 50 units of glutamate dehydrogenase (Sigma), and 1.33 mM CaCl 2 or 200 nM free Ca 2ϩ , and the fluorescence of NADPH was followed in a PerkinElmer LS-50 luminescence spectrometer at excitation and emission wavelengths of 340 and 460 nm, respectively. Traces were calibrated by the addition of 2 nmol of glutamate at the end of each assay. The data were obtained at 2-s intervals and corrected for Ca 2ϩ -independent release. Accordingly, the Ca 2ϩ -dependent release was calculated by subtracting the release obtained during a 5-min period of depolarization at 200 nM free [Ca 2ϩ ] from the release at 1.33 mM CaCl 2 .
The Cytosolic Free Ca 2ϩ Concentration ([Ca 2ϩ ] c ) in the Synaptosomal Population-The [Ca 2ϩ ] c concentration was measured with fura2. Synaptosomes were resuspended in HBM (2 mg/ml) with 16 M BSA in the presence of 1.3 mM CaCl 2 and 5 M fura2-acetoxymethyl ester (fura2-AM; Molecular Probes, Eugene, OR), and incubated at 37°C for 25 min. After fura2 loading, the synaptosomes were pelleted and resuspended in fresh HBM with BSA. A 1-ml aliquot was transferred to a stirred cuvette containing 1.3 mM CaCl 2 and the fluorescence was monitored at 340 and 510 nm. Data points were taken at 0.5-s intervals and the [Ca 2ϩ ] cyt , was calculated using the equations described previously (21). IP 1 Accumulation-IP 1 accumulation was determined using the IP-One kit (Cisbio, Bioassays, Bagnol sur-Cèze, France) (22). Synaptosomes (0.67 mg/ml) in HBM with 16 M BSA and adenosine deaminase (1.25 units/mg of protein) were incubated for 1 h at 37°C. After 25 min 50 mM LiCl was added to inhibit inositol monophosphatase and subsequently the agonist L-AP4 was added for 20 min prior to lysis. Other drugs were added as indicated in the figure legends. Synaptosomes were collected by centrifugation at 13,000 ϫ g for 1 min at 4°C and resuspended (1 mg/0.1 ml) in lysis buffer: 50 mM HEPES, 0.8 M potassium fluoride, 0.2% (w/v) BSA, and 1% (v/v) Triton X-100, pH 7.0). The lysed synaptosomes were transferred to a 96-well assay plate and the homogeneous time-resolved fluorescence (HTRF) components were added: the europium cryptate-labeled anti-IP 1 antibody, and the d2-labeled IP 1 analogue were both diluted in lysis buffer. After incubating the assays for 1 h at room temperature, the europium cryptate fluorescence and the time-resolved fluorescence resonance energy transfer signals were measured 50 s after excitation at 337 at both 620 and 665 nm, respectively, using a RubyStar fluorimeter (BMG Labtechnologies, Offenburg, Germany). The fluorescence intensities measured at 620 and 665 nm correspond to the total europium cryptate emission and the FRET signal, respectively. The specific FRET signal was calculated using the following equation: ⌬F% ϭ 100 ϫ (R pos Ϫ R neg )/(R neg ), where R pos is the fluorescence ratio (665/620 nm) calculated in the wells incubated with both donor-and acceptor-labeled antibodies, and R neg is the same ratio for the negative control incubated only with the donor fluorophore-labeled antibody. The FRET signal (⌬F%), which is inversely proportional to the concentration of IP 1 in the cells, was then transformed into the accumulated IP 1 value using a calibration curve prepared on the same plate.
cAMP Accumulation-cAMP accumulation was determined using a cAMP dynamic kit (Cisbio, Bioassays, Bagnol sur-Cèze, France). The assay was similar to that described for IP 1 except that during incubation the cAMP phosphodiesterase inhibitor Ro-20-1724 (0.1 mM) (Calbiochem, Damstard, Germany) was included for 35 min during incubation. The homogeneous time-resolved fluorescence assay was also similar to that described for IP 1 except that an anti-cAMP antibody and d2-labeled cAMP analogue were used.
Immunocytochemistry-The affinity purified guinea pig polyclonal antisera against mGlu4a, mGlu7a, and mGlu8a receptors used here have been described elsewhere (23). The polyclonal rabbit antiserum against synaptophysin 1 was obtained from Synaptic Systems, Gottingen, Germany. As a control of the immunochemical reactions, primary antibodies were omitted from the staining procedure whereupon no immunoreactivity could be detected that resembled that obtained with the specific antibodies.
Immunocytochemical Procedures-The synaptosomes were allowed to attach to polylysine-coated coverslips for 1 h and then fixed for 5 min in 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). Following several washes with 0.1 M phosphate buffer (pH 7.4), the synaptosomes were preincubated for 1 h in 10% normal goat serum diluted in 50 mM Tris buffer (pH 7.4) containing 0.9% NaCl (TBS) and 0.2% Triton X-100. Subsequently, they were then incubated for 24 h with the appropriate primary antiserum diluted in TBS with 1% normal goat serum and 0.2% Triton X-100:mGlu7a (1 g/ml), mGlu4a (1 g/ml), mGlu8a (1 g/ml) receptors, or synaptophysin (1:100). After washing in TBS, the synaptosomes were incubated with secondary antibodies diluted in TBS for 2 h: a goat anti-rabbit antibody coupled to the Cy2 cyanine-derived fluorochrome Cy2 (diluted 1:200; Jackson, West Grove, PA) or a Cy3-coupled donkey anti-guinea pig antibody (diluted 1:400, Chemicon International, Temecula, CA). After several washes in TBS, the coverslips were mounted with Prolong Antifade Kit (Molecular Probes) and the synaptosomes were viewed with a Nikon Diaphot microscope equipped with a ϫ100 objective, a mercury lamp light source, and fluorescein-rodamine Nikon filter sets.
Co-immunoprecipitation-The primary antibodies used for immunoprecipitations were a mouse monoclonal anti-munc13-1 (Synaptic Systems, Gottingem, Germany) or an affinity purified rabbit polyclonal antiserum against mGlu7a receptor (23). The secondary antibodies were horseradish peroxidase-conjugated anti-rabbit IgG TrueBlot TM (1:1000; eBioscience, San Diego, CA). The proteins recovered were resolved by SDS-PAGE on 6.5% polyacrylamide gels and transferred to polyvinylidene difluoride membranes using a semi-dry transfer system, which were then probed with the antibody indicated and a horseradish peroxidase-conjugated secondary antibody. The immunoreactive bands were visualized by chemiluminescence (Pierce) and detected in a LAS-3000 (FujiFilm Life Science, Woodbridge, CT). P 2 synaptosomes isolated from adult rats (Wistar) were homogenized using a VDI 12 homogenizer (VWR International, Barcelona, Spain) and solubilized for 30 min on ice in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl, pH 7.4, 100 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 0.2% SDS, and 1 mM EDTA) for 30 min on ice. The solubilized preparation was then centrifuged at 13,000 ϫ g for 30 min and the supernatant (2 mg/ml) was processed for immunoprecipitation, each step was conducted with constant rotation at 0 -4°C. The supernatant was incubated overnight with the antibody indicated and then 50 l of a suspension of either protein A cross-linked to agarose beads (Sigma) or TrueBlot anti-rabbit Ig IP beads (eBioscience, San Diego, CA) was added and the mixture and incubated overnight. Subsequently, the beads were washed with ice-cold RIPA buffer that was removed with a 28-gauge needle. Then, 100 l of SDS-PAGE sample buffer (0.125 M Tris-HCl, pH 6.8, 4% SDS, 20% glycerol, 0.004% bromphenol blue) was added to each sample, and the immune complexes were dissociated by adding fresh dithiothreitol (50 mM final concentration) and heating to 90°C for 10 min. Proteins were resolved and detected as described above.

RESULTS
Release Inhibition-It is known that mGlu7 receptors inhibit synaptic transmission by decreasing evoked glutamate release and Ca 2ϩ influx (5). The cerebrocortical nerve terminal prepa-ration from adult rats is enriched in this receptor, as witnessed by co-labeling with antisera against the vesicle marker synaptophysin and the L-AP4-sensitive mGlu receptors (4, 7, and 8) expressed in the brain. Indeed, among the nerve terminals that contained synaptophysin (1,148 particles from 7 fields), 30.0 Ϯ 1.2% also contained mGlu7a receptor (mean Ϯ S.E.; Fig. 1, A and D), whereas only 1.4 Ϯ 0.5% contained mGluR4a (1,616 particles from 8 fields) and 1.6 Ϯ 0.6% (989 particles from 8 fields) were immunoreactive for mGlu4a and mGlu8a receptors, respectively (Fig. 1, B-D). Thus the responses to L-AP4 are largely mediated by the mGlu7 receptor in this preparation.
Depolarization of nerve terminals with KCl increases Ca 2ϩ influx and glutamate release (5). In naive synaptosomes, the glutamate release evoked by 30 mM KCl (2.7 Ϯ 0.1 nmol of Glu/mg of protein Ϯ S.E., n ϭ 6) fell to 66.7 Ϯ 5% of control value in the presence of the group III mGlu receptor agonist, L-AP4 (n ϭ 6, p Ͻ 0.001, Fig. 1, E and F). This response paralleled the reduction in depolarization-induced Ca 2ϩ influx to 81.8 Ϯ 1.2% of that in the controls (n ϭ 8, p Ͻ 0.01, Fig. 1, G and H). However, there was no reduction in KCl-evoked release (97.0 Ϯ 2% of control release) in synaptosomes that were exposed to L-AP4 (1 mM) for 10 min (n ϭ 8, p Ͼ 0.05, Fig. 1, E and F). It is possible that the mGlu7 receptor undergoes activity-dependent internalization (19,24,25), leading to loss of the receptor response. However, this second agonist addition still decreased the Ca 2ϩ -influx in fura2-loaded synaptosomes (81.40 Ϯ 1%, n ϭ 8, p Ͻ 0.01, Fig. 1, G and H), indicating that the mGlu7 receptor remains at the surface and therefore that such receptor failure to reduce the evoked release is not due to receptor internalization. It is also possible that prolonged exposure to L-AP4 promotes a new response that augments release thereby counterbalancing the inhibition due to diminished Ca 2ϩ influx. Because 30 mM KCl evokes maximal glutamate release, this stimulation condition establishes a ceiling above that no facilitation can be observed. Thus, we used lower KCl concentrations to make release facilitation possible. Glutamate release in naive synaptosomes stimulated with 5 mM KCl (0.9 Ϯ 0.11 nmol, n ϭ 6, Fig 1, I and J) was reduced to 45.8 Ϯ 4% of the control value by the initial addition of L-AP4 (n ϭ 6, p Ͻ 0.05). Interestingly, in nerve terminals previously exposed to L-AP4 the glutamate release evoked by 5 mM KCl was potentiated (165.8 Ϯ 6.4% of control, n ϭ 6, p Ͻ 0.01). Furthermore, a second addition of L-AP4 failed to inhibit the release of glutamate (90.5 Ϯ 3.1% of control, n ϭ 6, p Ͼ 0.05) and thus, exposure of synaptosomes to L-AP4 promoted a new response that augmented the release and counteracted its inhibition due to diminished Ca 2ϩ influx.
As prolonged exposure to L-AP4 potentiated counterbalancing release inhibition despite impaired Ca 2ϩ influx, pharmacological blockage of the facilitating pathway should permit the recovery of release inhibition in response to a second addition of L-AP4. Inhibition of release by a second addition of L-AP4 was still evident in nerve terminals that were pre-exposed to L-AP4 but in the presence of the active PLC inhibitor U-73122 (61.5 Ϯ 4.9% of control, n ϭ 5, p Ͻ 0.001: Fig. 3, A and B). By contrast, exposure to L-AP4 in the presence of the inactive PLC inhibitor U-73343 results in the lack of release inhibition by a second addition of the agonist (99.2 Ϯ 1.9% of control, n ϭ 5, p Ͼ 0.05). Because release facilitation counterbalances release inhibition observed upon 30 mM KCl depolarization, altering the extent of release facilitation should also affect the extent of release inhibition. Exposing nerve terminals to different L-AP4 concentrations (0.5-1.0 mM) would be expected to produce graded release facilitation and hence, these conditions were used to test for the potentiation of ionomycin-induced release or for the inhibition of KCl-evoked release. At 1 mM L-AP4, maximal release potentiation was observed (172.0Ϯ.3.0%, n ϭ 5, p Ͻ 0.001), whereas this treatment completely abolished release inhibition in response to a second addition of the agonist (100.1 Ϯ 4.1%, n ϭ 5, p Ͼ 0.05: Fig. 3C). However, in synaptosomes that had been treated with lower L-AP4 concentrations, the facilitation of ionomycin-induced release decreased and a progressive recovery of release inhibition was observed. Thus, at 0.5 mM L-AP4, release potentiation was abolished (101.0 Ϯ 1.9%, n ϭ 5, p Ͼ 0.05), whereas release inhibition was FIGURE 3. Release modulation by mGlu7 receptors is a balance between inhibition and facilitation. Pharmacological blockage of PLC not only prevents L-AP4-induced release potentiation but also allows recovery of release inhibition by a second addition of L-AP4. Synaptosomes were exposed to 1 mM L-AP4 for 10 min in the presence of the PLC inhibitor, U-73122, or its inactive analogue, U-73343 (A). After washing, centrifuging, and resuspending, the synaptosomes were tested for release inhibition by a second addition of L-AP4 prior to depolarization with 30 mM KCl. B, diagrams show summarized data on release inhibition by a second addition of L-AP4. The release induced by 30 mM KCl in L-AP4-treated nerve terminals, but in the absence of a second addition of L-AP4, was taken as control. C, lowering the L-AP4 concentration of treated synaptosomes decreases the extent of release facilitation and allows the recovery of release inhibition by a second addition of L-AP4. Synaptosomes exposed to L-AP4 (0.5-1 mM) for 10 min were used to test release facilitation by estimating the ionomycin-induced release and to test release inhibition by depolarizing nerve terminals with 30 mM KCl after a second addition of L-AP4 (1 mM). In release facilitation experiments (green), the release induced by 2 M ionomycin in naive synaptosomes was taken as control. In release inhibition experiments (red), the release induced by 30 mM KCl in naive synaptosomes was taken as control. Data on the effect of different L-AP4 concentration pre-treatments both in release facilitation and inhibition are presented as compared with control. The data represent the mean Ϯ S.E. (n ϭ 5). NS, p Ͼ 0.05; ***, p Ͻ 0.001 (unpaired t test) when compared with corresponding control values. maximal (67.1 Ϯ 5.6% of control, n ϭ 5, p Ͻ 0.001). These results indicate that the extent of release facilitation by mGlu7 receptors dynamically controls the extent of release inhibition.
Given that the mGlu7 receptor and munc13-1 proteins both localize at the presynaptic active zone (1, 31), co-immunopre-cipitation experiments were performed using P 2 synaptosomes. The anti-munc13-1 antibody immunoprecipitated a band of around 200 kDa that appeared to correspond to the munc13-1 protein (29) (Fig. 5D), whereas the anti-mGlu7 receptor antibody was able to immunoprecipitate a band of around 200 kDa that corresponded to the mGlu7 receptor dimer (Fig. 5D), as demonstrated previously (23). Significantly, whereas the anti-munc13-1 antibody was able to co-immunoprecipitate the , and vGlut1 and Na ϩ ,K ϩ -ATPase (determined in Western blot) was estimated in each fraction, and the sum of the soluble and particulate fraction values taken as 100%. C, the ratio between soluble and particulate munc13-1 fractions were calculated in each experiment. Data represent the mean Ϯ S.E. (n ϭ 6). NS, p Ͼ 0.05; *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001 (unpaired t test), when compared with either the soluble or particulate fractions, or the S/P ratio from naive synaptosomes. D, co-immunoprecipitation of mGlu7 receptors and munc13-1 from the cerebrocortical synaptosomes. Solubilized extracts were subjected to immunoprecipitation (IP) with rabbit anti-FLAG antibody (5 g, lane 1), rabbit anti-munc13-1 polyclonal antiserum (5 g, lane 2), rabbit anti-mGlu7 receptor polyclonal antiserum (5 g, lane 3), and mouse anti-FLAG antibody (5 g, lane 4). Extracts (Crude) and immunoprecipitates were analyzed in Western blots probed with a rabbit anti-mGlu7 receptor polyclonal antiserum (2.5 g/ml) or a rabbit anti-munc13-1 polyclonal antiserum (1 g/ml). Horseradish peroxidase-conjugated anti-rabbit IgG TrueBlot (1:1000) was used as a secondary antibody to avoid IgG cross-reactivity and the immunoreactive bands were visualized by chemiluminescence. IB, immunoblot.
mGlu7 receptor, the anti-mGlu7 receptor antibody could also co-immunoprecipitate munc13-1 (Fig. 5D). These bands did not appear when an irrelevant rabbit or mouse IgG were used for immunoprecipitation (Fig. 5D, lanes 1 and 4, respectively), showing the specificity of the reaction. These results suggest that the mGlu7 receptor and munc13-1 might assemble into stable protein-protein complexes in the rat cortex that survive the solubilization and co-immunoprecipitation conditions employed. The stability of these oligomeric complexes indicates that they might be physiologically relevant in vivo.

DISCUSSION
It is well established that metabotropic glutamate receptor 7 inhibits glutamate release by decreasing presynaptic Ca 2ϩ influx through voltage-dependent calcium channels. Using a preparation of nerve terminals from the cerebral cortex, we found that in addition to mediating release inhibition, mGlu7 receptor also activates signaling pathways that potentiate release. Release potentiation requires at least a 10-min exposure to a receptor agonist and once the mechanism of release potentiation is established, it occludes the release inhibition produced by a second addition of the agonist. In contrast to the PTX-sensitive G protein that impedes Ca 2ϩ channel activity in mGlu7 receptor-dependent inhibition of release, the mGlu7 receptor response that potentiates release involves a PTX-resistant G protein that activates PLC, increasing phosphatidylinositol hydrolysis and promoting DAG-dependent translocation of the munc13-1 protein.
In addition to the well established inhibition of voltage-dependent Ca 2ϩ channels (5,(32)(33)(34) and adenylyl cyclase (5), mGlu7 receptor couples to other signaling pathways such as those that activate GIRK channels (35) or PLC (7). However, the physiological relevance of these signaling mechanisms in synaptic glutamate release modulation remain largely unknown. Thus, mGlu7 receptor-mediated GIRK channel activation is limited to heterologous expression systems and whether this signaling occurs at synaptic sites remains to be shown. When ectopic mGlu7 receptors are expressed in cerebellar granule cells, somatic Ca 2ϩ currents are inhibited via PLC activation (7), although the relevance of this signaling at synaptic sites remains unclear. Our finding that synaptic mGlu7 receptor activates a PLC-dependent pathway that leads to release poten-tiation is consistent with previous work showing that presynaptic PLC is involved in synaptic release potentiation in cultured hippocampal neurons (18). In this study, release potentiation induced by high frequency stimulation (HFS) was blocked by the active PLC inhibitor, U-73122 (18). As PLC is a Ca 2ϩ -dependent enzyme, it was proposed that PLC was activated by the increase in presynaptic Ca 2ϩ influx induced by HFS. Our data suggest that in addition to Ca 2ϩ , presynaptic PLC is also activated by PTX-insensitive G proteins. Nevertheless, high frequency stimulation increases synaptic release of glutamate to levels sufficient to activate mGlu7 receptors (19), despite the low affinity reported for this receptor for glutamate (36). The identity of the PLC isoform involved in L-AP4-mediated release potentiation at cerebrocortical synapses remains unclear. At least 13 PLC isoforms have been identified in mammals that belong to six families that are designated: -␤, -␥, -␦, -⑀, -, and PLC-, and some of which are abundantly expressed in the brain (PLC-␤, -␥, and -) (14,37,38). Although all PLCs catalyze the hydrolysis of phosphatidylinositol (4,5)-bisphosphate, thereby generating DAG and inositol trisphosphate , the presence of distinct regulatory domains in PLC isoforms renders them susceptible to different modes of activation. Thus, all PLCs require Ca 2ϩ with the ␦-, -, and -type being the most sensitive to this cation. In addition, G␣ q proteins activate PLC␤ but not the other isozymes, whereas G␤␥ can activate PLC␤, -⑀, and - (14,39). Thus, at least three PLCs can be activated by GPCRs.
Recent reports have shown that mGlu7 receptors bidirectionally control plasticity at hippocampal synapse (19). In naive slices, mGlu7 receptor activation during HFS generates long term depression, at synapses between mossy fibers and stratum lucidum interneurons due to a persistent decrease in glutamate release. However, the exposure of hippocampal slices to agonist L-AP4 internalizes mGlu7 receptors, unmasking the ability of these synapses to undergo presynaptic potentiation in response to the same HFS that induced long term depression in naive slices. L-AP4-mediated potentiation of release found at cerebrocortical boutons has some similarities with HFS-induced long term potentiation, and of mossy fiber synapses found in L-AP4-treated hippocampal slices (40). At both synapses release potentiation requires prolonged exposure to L-AP4 and it occurs despite the persistent reduction in Ca 2ϩ influx. However, whereas potentiation at MF-SLIN synapses is prevented by either the adenylyl cyclase inhibitor DDOA or the PKA inhibitor H-89, L-AP4-mediated potentiation at cerebrocortical nerve terminals is insensitive to these compounds, although it is blocked by U-73122.
The bidirectional control of glutamate release raises the question whether there is a single population of nerve terminals in which L-AP4 has a dual role in release modulation or whether there are two populations of regulated nerve terminals, one that FIGURE 6. Dual modulation of glutamate release by mGlu7 receptors. Scheme illustrating the putative signaling pathways activated by mGluR7. Exposure to agonist L-AP4 (1 mM, 30 s) initiates a signaling cascade that activates the pertussis toxin-sensitive G protein (G i/o ), resulting in the reduction of evoked Ca 2ϩ influx and glutamate release. Longer exposures to agonist (1 mM, 10 min) results in the potentiation of glutamate release. Release potentiation involves a PTX-resistant G protein that activates PLC, increasing phosphatidylinositol (4,5)-bisphosphate (PIP 2 ) hydrolysis and promoting DAG-dependent translocation of the munc13-1 protein.