Native presynaptic metabotropic glutamate receptor 4 (mGluR4) interacts with exocytosis proteins in rat cerebellum.

Background: mGluR4 negatively modulates neurotransmission at glutamatergic synapses. Results: Using proteomic approaches, we show that native mGluR4 interacts with exocytosis proteins. Conclusion: Native mGluR4 could inhibit glutamate release via interactions with exocytosis proteins. Significance: Because mGluR4 plays key roles in regulation of neurotransmission and is considered as a promising target for treatment of various brain diseases, the identification of its mechanisms of action is fundamental. The eight pre- or/and post-synaptic metabotropic glutamatergic receptors (mGluRs) modulate rapid excitatory transmission sustained by ionotropic receptors. They are classified in three families according to their percentage of sequence identity andtheirpharmacologicalproperties.mGluR4belongstogroupIII and is mainly localized presynaptically. Activation of group III mGluRs to depression of excitatory transmission, a process that is exclusively provided by mGluR4 at parallel fiber-Purkinje cell synapse in rodent cerebellum.

The eight pre-or/and post-synaptic metabotropic glutamatergic receptors (mGluRs) modulate rapid excitatory transmission sustained by ionotropic receptors. They are classified in three families according to their percentage of sequence identity and their pharmacological properties. mGluR4 belongs to group III and is mainly localized presynaptically. Activation of group III mGluRs leads to depression of excitatory transmission, a process that is exclusively provided by mGluR4 at parallel fiber-Purkinje cell synapse in rodent cerebellum. This function relies at least partly on an inhibition of presynaptic calcium influx, which controls glutamate release. To improve the understanding of molecular mechanisms of the mGluR4 depressant effect, we decided to identify the proteins interacting with this receptor. Immunoprecipitations using anti-mGluR4 antibodies were performed with cerebellar extracts. 183 putative partners that co-immunoprecipitated with anti-mGluR4 antibodies were identified and classified according to their cellular functions. It appears that native mGluR4 interacts with several exocytosis proteins such as Munc18-1, synapsins, and syntaxin. In addition, native mGluR4 was retained on a Sepharose column covalently grafted with recombinant Munc18-1, and immunohistochemistry experiments showed that Munc18-1 and mGluR4 colocalized at plasma membrane in HEK293 cells, observations in favor of an interaction between the two proteins. Finally, affinity chromatography experiments using peptides corresponding to the cytoplasmic domains of mGluR4 confirmed the interaction observed between mGluR4 and a selection of exocytosis proteins, including Munc18-1. These results could give indications to explain how mGluR4 can modulate glutamate release at parallel fiber-Purkinje cell synapses in the cerebellum in addition to the inhibition of presynaptic calcium influx.
Glutamate mediates fast excitatory synaptic transmission by activating ionotropic receptors in the central nervous system. This neurotransmitter can also activate G protein-coupled receptors such as metabotropic glutamate receptors so called mGluRs. 3 These receptors display differential expression on pre-and postsynaptic membranes and modulate the fast excitatory glutamatergic neurotransmission. This family is divided into three groups (I, II, III) according to their sequence identities, pharmacological profile, and signal transduction mechanisms. Group III mGluRs have been described to be negatively coupled to adenylate cyclase (1) and have been shown to inhibit glutamate release (2,3). These receptors are activated by released glutamate in the synaptic cleft and act as autoreceptors to inhibit excitatory synaptic transmission. They are also involved in the induction of two forms of short-term synaptic plasticities: the paired-pulse facilitation and the post-tetanic potentiation (4,5). The mGluR4 subtype, which belongs to the presynaptic group III mGluRs, is mainly expressed in the cerebellar cortex (6). The most compelling demonstration for a physiological function of presynaptic mGluR4 is found at the glutamatergic synapses between parallel fibers and Purkinje cells in the rodent cerebellar cortex. At these synapses, pharmacological activation of these receptors acutely depresses excitatory transmission and inhibits presynaptic calcium influx (7,8). In the rodent cerebellum, depression of parallel fiber-Purkinje cell excitatory transmission is exclusively due to mGluR4 (8).
Little is known about the mechanisms of action of mGluR4 and the presynaptic targeting of these receptors. Most of the studies conducted to identify mGluR4-interacting proteins were performed in GST pulldown assays or in heterologous systems. mGluR4 has been shown to interact with microtubuleassociated protein or with proteins linked to cytoskeleton (␤-spectrin, myosin) or involved in endocytosis (clathrin and adaptor complex proteins) (9). The Ca 2ϩ sensor calmodulin has also been reported to interact with the C-terminal tail of mGluR4 (10) as well as syntenin, glutamate receptor interacting protein (11), and calcium-binding protein 1 (12). Finally, Munc18-1 was also recently identified as a binding partner for mGluR4 by co-immunoprecipitation and GST pulldown experiments (13). This protein has been suggested to assist in the formation of soluble N-ethylmaleimide-sensitive factor attachment protein receptor (SNARE) complex syntaxin/synaptosomal-associated protein 25 (SNAP-25)/synaptobrevin during neurotransmitter vesicles exocytosis (14) via a Ca 2ϩ / calmodulin-dependent pathway.
In an attempt to identify new molecular partners involved in the negative regulation of neurotransmission induced by native mGluR4 in the rat cerebellum, we decided to carry out a proteomic study. Several papers reported proteomic approaches for the identification of proteins interacting with G-proteincoupled receptor such as 5-hydroxytryptamine type 2 receptor (15) and mGluR5 (16). These works led to the discovery of direct but also indirect interactions, all partners included in the term "receptosome." To study the receptosome of mGluR4, we used a proteomic approach consisting of co-immunoprecipitation of the native receptor and identification of the protein partners by mass spectrometry analysis. In this study we demonstrated that mGluR4 was able to interact with numerous exocytosis proteins including Munc18-1 and SNAREs proteins such as syntaxin and SNAP-25. The interaction of Munc18-1 with native mGluR4 from rat cerebellum was confirmed using chromatography on a Sepharose column grafted with recombinant Munc18-1. Interaction domains of mGluR4 with several exocytosis proteins were further studied by affinity chromatography experiments using peptides corresponding to the intracellular parts of mGluR4. These interactions could be involved in the molecular mechanisms underlying the depressant presynaptic action of this activated receptor in the cerebellar cortex through the control of glutamate release.
Glutamate Assay-Glutamate assay was performed on total protein lysate of rat cerebellum according to the instructions of the manufacturer (EnzyChrom™ Glutamate Assay kit, BioAssay Systems, Hayward, CA).
Immunoprecipitation-Animal care and all experiments procedures adhered to governmental guidelines. Animals were stunned and then decapitated. Proteins were manually extracted using a potter from cerebellum of 30 -37-day-old male Sprague-Dawley rats with lysis buffer in a ratio of 0.25 ml/10 mg of tissue. The buffer contained 150 mM NaCl, 25 mM HEPES-NaOH, pH 8.0, 2% digitonin, 0.5 mM EDTA, 0.25 mg/ml 4-(2-aminoethyl)-benzenesulfonyl fluoride hydrochloride, 0.01 mg/ml E-64, 0.05 mg/ml antipaïn, and 1 mM Na 3 VO 4 . The homogenate was incubated on a rotary shaker during 1 h at 8°C then centrifuged at 14,000 ϫ g for 20 min at 4°C. Supernatant (1 ml) was used for immunoprecipitation. Protein concentration was determined by the Bradford method. Anti-mGluR4 and anti-Cascade Blue antibodies were cross-linked to protein A-Sepharose magnetic beads (Ademtech, Pessac, France) according to instructions of the manufacturer. The supernatant (3 mg of proteins) was incubated with 9 g of anti-mGluR4 or anti-CB overnight at 4°C. After several washing steps with 150 mM NaCl, 25 mM HEPES-NaOH, pH 8.0, buffer, proteins were eluted with 30 l of 50 mM glycine, pH 2.5, loaded on 10% acrylamide SDS-PAGE without heating, and then revealed with silver nitrate staining or Western blot.
Protein Identification by Mass Spectrometry-After one-dimensional electrophoresis, bands of interest were excised, then treated and digested using the automated system Digest Pro 96 (Intavis AG, Bremen, Germany). Bands were first destained by two washing steps with a freshly prepared solution containing 15 mM K 3 [Fe(CN) 6 ] and 50 mM Na 2 S 2 O 3 . Then, proteins were reduced and alkylated by successive incubations with 10 mM DTT in 50 mM NH 4 HCO 3 for 30 min at 57°C and 55 mM iodoacetamide in 50 mM NH 4 HCO 3 for 20 min at room temperature. In-gel digestion was then performed with modified trypsin (Promega), and peptides were extracted with formic acid-CH 3 CN (1:60 ratio). Peptide mixture was submitted to nano-LC-ion trap (IT)-MS/MS using an Agilent 1200 nanoflow LC system coupled to a 6330 Ion Trap equipped with the Chip Cube orthogonal ionization system (Agilent Technologies, Santa Clara, CA) as described in Marchand et al. (17). For protein identification, a MASCOT MS/MS Ions search was used, and searches were performed against the NCBI data base (release20100116; 10.343.571 sequences) with taxonomic specification to Rattus norvegicus. Mass accuracy tolerance was set to 100 ppm on the parent ion mass and 0.5 Da in MS/MS mode. One missed cleavage per peptide was allowed, and some modifications were taken into account: carbamido-methylation for cysteines as fixed modification and methionine oxidation as variable modifications. Proteins were considered with confidence if at least two unique peptides were identified with a score Ն34 and if the p value was Ͻ0.03. The Mowse score for the protein is calculated as Ϫ10 ϫ log (p).
Protein Munc18-1 Overexpression and Purification-Coding sequence of rat Munc18-1 was cloned under the T7 promotor control into PIVEX2.4d plasmid, allowing the addition of a sixhistidine tag at the N terminus of the protein. We used a syn-thetic gene optimized for heterologous expression and supplied by MWG. BL21 (DE3/pDIA17) was used as expression strain. Cells were grown under aerobic conditions at 37°C in Luria-Bertani medium in the presence of 0.1 mg/ml ampicillin, and protein expression was induced for 2 h with 0.5 mM isopropyl ␤-D-1-thiogalactopyranoside. Cells were harvested by centrifugation and broken in French press at 4°C in lysis phosphate buffer, pH 7.5, containing 2 mM MgSO 4 , 5% sucrose, 100 mM NaCl, 0.01 mg/ml DNase, and a mixture of anti-proteases (Complete Protease Inhibitor Mixture Tablet, Roche Applied Science). After cell debris and inclusion bodies had been pelleted, cytoplasmic and membrane proteins were separated by ultracentrifugation (90,000 ϫ g, 1 h, 4°C). The Munc18-1 recombinant protein found in the supernatant was affinitypurified on a Ni 2ϩ -IDA 1000 column (Macherey Nagel, Düren, Germany) according to the manufacturer's instructions. The protein was eluted from the affinity column with 250 mM imidazole, then submitted to dialysis against buffer, pH 8.3, containing 100 mM NaHCO 3 and 500 mM NaCl. The concentration of purified Munc18-1 was determined spectrophotometrically (⑀ 280 ϭ 61,770 M Ϫ1 cm Ϫ1 ).
Colocalization Quantification-Colocalization was determined using ImageJ software (National Institute of Health) and the Coloc_2 plugin. Basically, the plugin performed a pixel intensity spatial correlation analysis. We used the Mander's coefficient (r) (0 is no colocalization, 1 means 100% colocaliza-tion) to evaluate the signal proportion in one channel that colocalizes with the other channel. In this study we chose to calculate the mean value of the Mander's coefficient for the Munc18-1 channel over mGluR4 channel and, thus, to evaluate the proportion of Munc18-1 signal which colocalizes with the mGluR4 signal (r ϭ x Ϯ S.E.) (18).

Native mGluR4 Interacts with Exocytosis Proteins
20178 JOURNAL OF BIOLOGICAL CHEMISTRY min at 4°C. The supernatant was centrifuged at 100,000 ϫ g for 1 h at 4°C, whereas the pellet was solubilized in the same buffer supplemented with 1% Triton X-100 and centrifuged again at 14,000 ϫ g for 20 min at 4°C. Both extracts were used for affinity chromatography experiments. Samples (3-3.5 mg of proteins at a concentration of 0.5 mg/ml) were loaded on the resin, and the column was washed with 10 ml of buffer. Elution was performed with 7 ml of buffer containing 25 mM HEPES-NaOH, pH 8.0, 500 mM NaCl, 0.5 mM EDTA. Eluates were concentrated with Centricon devices (UFC201024PL, Millipore) and loaded on 10% SDS-PAGE gel for Western blot. Controls were performed with the same resin without peptide.
Immunoblot Analysis-After electrophoresis, proteins were transferred onto polyvinylidene difluoride membranes (Immobilon-P; Millipore) or to nitrocellulose membranes (Hybond-C Extra, Amersham Biosciences) for 1 h at 100 V with buffer containing 25 mM Tris, 192 mM glycine, 0.1% SDS. The membrane was incubated for 1 h in Odyssey buffer (LICOR Biosciences, NE) and TBS (50 mM Tris, 150 mM NaCl, pH 7.5) at the ratio 1:1 as blocking buffer. Analysis of immunoblot was carried out by the Odyssey Infrared Imaging system (LICOR). Secondary antibodies labeled with near-infrared dyes were used for protein detection; IRDye 800 has a max of 780 nm and a emission of 815 nm, and IRDye 680 has a max of 683 nm and a emission of 710 nm. Parameters such as dilution of the antibody, temperature, and incubation time were optimized for each antibody.

RESULTS
We verified the specificity of anti-mGluR4 antibodies (Invitrogen) used in co-immunoprecipitation (co-IP) by Western blot analysis. These experiments were performed on cerebellum extracts from rat, wild type mice, and mGluR4 knock-out mice in which the gene coding for the receptor had been deleted (5) (Fig. 1). A band at 250 kDa (likely corresponding to dimeric mGluR4) was observed both in rat and wild type mice extracts but not in those from mGluR4 KO mice. Similar results were obtained with several commercial anti-mGluR4 antibodies (data not shown). In co-IP experiments, an additional band at 100 kDa was observed, as shown in Fig. 2a, corresponding to the monomeric form of mGluR4. These results demonstrate that antibodies used for co-IP and Western blot are specific for mGluR4. It should be noted that when samples were heated before loading on SDS-PAGE, the receptor aggregated and remained in the stacking gel. For mGluR4 to enter the resolving gel and, as such, migrate properly, the samples should not be preheated.
An assay was performed to determine whether concentration of glutamate, the endogenous agonist, was sufficient to activate mGluR4 in our experimental conditions. Glutamate was measured at a concentration of 1 mM (Ϯ0.25 mM, an average of three separate experiments) in the supernatant used for co-IP (data not shown). Under these conditions, mGluR4 should be activated and thus able to interact with its putative transduction pathway partners as the EC 50 of glutamate for this receptor is 3-20 M (20).
To define the optimal working buffer for co-IP experiments, two detergents, Triton X-100 (TX-100) and digitonin, were used at different concentrations, 0.5-1 and 1-2%, respectively, and at two different pHs (7.4 and 8.0). The relative amount of native mGluR4 extracted from rat cerebellum in each buffer condition was assessed by Western blot. Co-immunoprecipitation was then performed using anti-CB antibodies as a negative control as described in Farr et al. (16). The proteins that coeluted either with anti-mGluR4 or anti-CB antibodies were separated on SDS-PAGE and stained with silver nitrate.
TX-100 at a concentration of 1% and at pH 8.0 provided the most efficient extraction of mGluR4 but failed to give a significant differential pattern (data not shown). On the other hand, a buffer containing 2% digitonin at pH 8.0 extracted a lower amount of receptor than 1% TX-100 but allowed successful immunoprecipitation of mGluR4 with its putative partners, indicating that this buffer is mild enough to maintain protein interactions (Fig. 2a). Indeed, numerous proteins were co-immunoprecipitated with anti-mGluR4 antibodies as opposed to the control antibodies (Fig. 2b). In addition, the efficiency of co-IP was assessed by Western blot, showing that all of the   JUNE 8, 2012 • VOLUME 287 • NUMBER 24 JOURNAL OF BIOLOGICAL CHEMISTRY 20179 mGluR4 receptor present in the supernatant had been immunoprecipitated with the anti-mGluR4 antibodies under these conditions (Fig. 3).

Native mGluR4 Interacts with Exocytosis Proteins
Three separate experiments were performed, and each resulted in similar differential patterns after silver nitrate staining. For protein identification, the whole SDS-PAGE lanes were cut into small bands, submitted to tryptic digestion, and analyzed independently by nanoLC-MS/MS. The eluates obtained with anti-mGluR4 and control antibodies were treated in the same way. Proteins identified in these control bands were considered as nonspecific interacting proteins and, therefore, removed from the list of mGluR4 partners. As expected, the two bands that gave a specific signal in Western blot experiments at 250 and 100 kDa were identified as mGluR4 by mass spectrometry. We identified a total of 184 proteins (mGluR4 and 183 potential partners) in the data pooled from these three experiments (supplemental Table 1). Fig. 4 shows the classification of these proteins according to their principal cellular functions such as proteins involved in metabolism (54), cytoskeleton (33), trafficking and exocytosis (21), receptors and channels (15, including mGluR4), gene regulation (13), oxidative stress (11), signaling (10), Ca 2ϩ regulation (9), molecular chaperones (8), axon growth (5), and non-classified proteins (5). Specifically, our experimental approach identified several proteins that have been previously described as partners of mGluR4, such as ␤-spectrin, myosin, clathrin, microtubule-associated protein, and calcium-binding protein 1 (see supplemental Table 1). It should be noted that no protein of the post-synaptic element, such as abundant post-synaptic density proteins, was found in any of the three experiments, suggesting that co-IP was rather specific.
We then focused on proteins belonging to the membrane presynaptic vesicular trafficking proteins family because mGluR4 is localized at the presynaptic zone (21), close to the exocytosis machinery and is known to negatively regulate synaptic neurotransmission (8). These proteins are shown in Table  1. Among them, Munc18-1 had already been described as a potential mGluR4 partner in co-IP experiments with synaptosomal preparations and GST pulldown experiments (13). In addition, we identified presynaptic vesicle proteins such as synapsins I and II, synaptic vesicle glycoprotein 2A (SV2A), which could regulate the initial steps of exocytosis, target-SNAREs such as SNAP-25 and Syntaxin-1B, ATPase N-ethylmaleimidesensitive factor, which dissociates SNARE complex and permits recycling of these proteins after membrane fusion, and ␣synuclein, which is involved in exocytosis regulation (22,23). Finally, proteins from the Ras-protein superfamily such as Rab, Rac, and Rho, which are involved in vesicular trafficking (24), were also found in our study. The presence of Munc18-1, SNAP-25, synapsins I and II, and syntaxin in the eluates of co-IP experiments was validated by Western blot analysis as shown in Fig. 5.
To confirm native interactions between presynaptic vesicular trafficking proteins and mGluR4, reverse co-IPs were performed on rat cerebellum extracts with antibodies directed against Munc18-1, Syntaxin, and SNAP-25. However, this receptor could not be detected in co-IP eluates. Although mGluR4 is relatively abundant in the cerebellum compared with other brain structures, it remains very weakly expressed in comparison with exocytosis and trafficking proteins. Therefore, the probability of detecting mGluR4 as a partner of Munc18-1, syntaxin, or SNAP-25 in reverse co-IPs was quite low.
To circumvent this limitation, we developed another reverse strategy to demonstrate that an exocytosis protein could target mGluR4. We performed affinity chromatography with recombinant Munc18-1. This protein has already been identified as a putative partner of mGluR4 but only when the C terminus of mGluR4 was the bait (13). Therefore, Munc18-1 was overexpressed in Escherichia coli as a polyhistidine-tagged protein, purified, and covalently bound on activated Sepharose matrix. For these experiments, rat cerebellar proteins were extracted with TX-100 to obtain the maximal amount of mGluR4. After loading the cerebellar extract onto the Munc18-1 column, retained proteins were eluted with a concentrated saline solution, and the presence of mGluR4 was investigated by Western blot analysis. A blank was performed under the same conditions using the resin alone. Native mGluR4 was specifically detected in the eluates from the Munc18-1 column, which confirms the ability of the native receptor to interact with this protein (Fig.  6). In addition, proteins such as syntaxin and SNAP-25, which have already been described as Munc18-1 partners (25)(26)(27), and synapsin I, were also present in the eluates. The experiment was performed twice and gave similar results. Together with FIGURE 3. Control of the mGluR4 co-IP efficiency. Immunoprecipitation with anti-mGluR4 or anti-CB antibodies was performed as described under "Experimental Procedures." Proteins that were not co-immunoprecipitated and remained in the flow-through after co-IP were loaded on 10% SDS-PAGE and analyzed by Western blot (50 g of proteins/lane) using anti-mGluR4 antibodies to validate the complete immunoprecipitation of the receptor. co-IP results, these data suggest that Munc18-1, SNAP-25, synapsins, syntaxin, and mGluR4 form stable complexes.
To further characterize Munc18-1 and mGluR4 interaction, we investigated their co-localization in HEK293 cells. Munc18-1 expressed alone was highly concentrated in the cytoplasm (Fig. 7a), which is consistent with previous observations (28), whereas mGluR4 expressed alone was mainly observed at the cell membrane (Fig. 7b). Capping experiments with mGluR4 allowed us to specifically visualize membrane expression of the receptor (Fig. 7c1). We used the technique of co-capping to detect an association between Munc18-1 and mGluR4. We showed that among the population of Munc18-1 clusters (Fig. 7c2), a significant proportion colocalized with caps of mGluR4 (Fig. 7c3), which was supported by calculation of Mander's coefficient (r) (r ϭ 0.27 Ϯ 0.07, n ϭ 10, see "Experimental Procedures"). Thus, the localization of Munc18-1 toward mGluR4 caps within the plasma membrane was quite striking and strongly supported our proteomic and biochemical data concerning Munc18-1/mGluR4 interaction.
The interaction domain of mGluR4 with exocytosis proteins was further studied by affinity chromatography experiments. Peptides corresponding to the second intracellular loop and to the entire or partial C-terminal tail of the receptor were synthesized and covalently fixed on a Sepharose column (Table 2 and

partners involved in exocytosis and cellular trafficking
Shown is a list of trafficking and exocytosis proteins identified by mass spectrometry after co-IP. 9 of 21 identified proteins (in bold) are required for synaptic exocytosis (M r , molecular weight; pI, isoelectric point). NSF, N-ethylmaleimide sensitive fusion protein.    Affinity chromatography experiments performed with the entire C-terminal peptide, called AI-59, validated the interaction of Munc18-1, syntaxin, and synapsin I with mGluR4 (Fig.  9a). The same proteins were found to also interact with the second intracellular loop of the receptor called NQ-25 (Fig. 9b). By contrast, SNAP-25 was not detected using this approach. All the results from affinity chromatography experiments were analyzed by Western blot are summarized in Table 3. To obtain supplemental information on the interaction domains of mGluR4 with its binding partners, we used truncated fragments of the AI-59 peptide, namely AN-23, VN-30, and GI-30, from the N terminus to the C terminus, respectively. Munc18-1 was found to interact with the AN-23 peptide (Fig. 9c), whereas VN-30 and GI-30 peptides showed no interaction ( Table 3). The positive signal with AN-23 was observed on proteins extracted both with and without detergent, although it was more intense with the use of detergent (1% TX-100). The experiments were performed at least twice and gave similar results. This affinity chromatography approach thus allowed us to validate the interaction of several exocytosis proteins with mGluR4 with potential binding sites on the second intracellular loop and/or on the C-terminal tail of the receptor.

DISCUSSION
A total of 183 proteins was found to interact with native mGluR4 in our co-IP experiments. Proteins were classified in different families according to their cellular function. Among them, metabolism and cytoskeleton proteins account for nearly 50% of the proteins identified. Due to their high abundance in the synapse, these protein groups are usually found on pre-or post-synaptic receptors in proteomic experiments. This is the c, co-capping experiments of mGluR4 and Munc18-1 is shown. c1, mGluR4 capping was induced by cross-linking with HA antibody on the cell surface and detected with Alexa488-labeled secondary antibody (green). Alexa596 secondary antibody was used as a control to ensure no cross-reactivity with anti-HA antibody, used to detect HA-tagged mGluR4 (c1, inset). c2, Munc18-1 labeled with anti-Munc18-1 antibody and detected with Alexa546-labeled secondary antibodies (red) is shown. Note that no cross-reactivity was observed (c2, inset), although mGluR4 was well capped (c1, inset). c3, Munc18-1 was recruited to mGluR4 caps and thus partially co-localized with mGluR4 at the plasma membrane (c1-3, white arrows). Note that co-capping of mGluR4 and Munc18-1 was observed on 79% of mGluR4-expressing cells (n ϭ 210). Bar, 10 m.

TABLE 2
Characteristics of mGluR4 intracellular peptides used in affinity chromatography AI-59 corresponds to the C-terminal tail of mGluR4, except that the first Pro was changed into Ala. AN-23, VN-30, and GI-30 correspond to overlaying fragments of the full length C-terminal tail of mGluR4 (from proximal part to distal part respectively). NQ-25 corresponds to the second intracellular loop.  (29), or the GluR4 subunit of post-synaptic AMPA receptor (30). In all cases co-immunoprecipitated proteins are found specifically in the synaptic compartment containing the receptors. That was also the case in our study, as no post-synaptic protein was identified. Several partners already proposed in other studies (9, 12) have been found in our work, such as ␤-spectrin, myosin,  clathrin, adaptor protein complex, calcium-binding protein 1 (see supplemental Table 1), and should be further studied. In contrast, some other putative interacting proteins (syntenin, glutamate receptor interacting protein) highlighted in heterologous systems (11) were not identified in our study. Although protein interacting with c-kinase 1 (PICK1) was found to interact with mGluR4 and mGluR7 in heterologous systems (31), we did not identified this protein as a native mGluR4 partner. In cultured cerebellar granule cells, the PICK1/mGluR7 interaction leads to protein kinase C (PKC) activation (32), indicating a functional role of PICK1 in cellular mechanisms of mGluR7. This activation induces inhibition of presynaptic P/Q-type voltage-gated calcium channels, leading to reduction of glutamatergic synaptic transmission (32).

Name
Recently, it appears that this inhibition would involve the phosphorylation of SNAP-25 on Ser-187 by PKC (33,34). These results show the tight link between mGluR7 and scaffold/exocytosis proteins in the control of neurotransmission regulation via voltage-gated calcium channels. Although we did not identify these channels as interacting proteins of native mGluR4 in our study, we cannot exclude that the negative modulation of neurotransmission induced by mGluR4 activation could involve transient inhibition of presynaptic voltage-gated calcium channels. Indeed, the nature of the intermediate molecular players involved in the signaling cascades responsible for the transient inhibition of these channels remains an open question. Otherwise, we did not find any Homer protein in mGluR4 receptosome, which can be explained by their specific interactions with group I mGluRs. Due to the crucial roles of Homer proteins in membrane expression of mGluR5 (35) and in the functional coupling efficiency of group I receptors (16,36), further studies should be performed to determine whether similar proteins could play such a role for mGluR4.
In this study we focused on exocytosis/trafficking proteins (21 proteins identified here) because of their proximity to mGluR4. Among trafficking proteins, 9 of the 21 identified proteins belong to exocytosis family. Proteins involved in the regulation of glutamate release were identified, such as synaptotagmin-2, which is a Ca 2ϩ sensor involved in neurotransmitter release at central nervous system synapses (37). Recently, synaptotagmins were found to interact with SV2A (also identified in our experiments), and SV2A proteins would regulate their expression and internalization (38). Synapsins I and II, which may play a role in the stabilization of synaptic vesicles to maintain a pool of neurotransmitters available (39), were also identified in our experiments. Therefore, mGluR4 could be involved in the mobilization of synaptic vesicles through direct or indirect interactions with synaptotagmin-2, synapsins I and II, and SV2A.
Moreover, our proteomic approach revealed an interaction between mGluR4 and several native proteins of the SNARE complex. Indeed, syntaxin-1 and SNAP-25 (target SNAREs), which were identified in this study, associate with synaptobrevin-2 (VAMP-2, a vesicle SNARE) to create a ternary trans-SNARE complex. This complex is involved in the physical linkage of synaptic vesicles on the plasma membrane necessary for neurotransmitter release. We also identified ␣-synuclein, which is thought to facilitate the assembly of the trans-SNARE complex (23). Thus, mGluR4 may be able to regulate the SNARE complex assembly via its interactions, direct or indirect, with syntaxin, SNAP-25, and ␣-synuclein.
To gain insight into the interaction domains between mGluR4 and these exocytosis proteins, affinity chromatography experiments were performed on different peptides corresponding to the C-terminal tail (full-length or truncated in three fragments) and the second intracellular loop of mGluR4. Indeed, we confirmed the interaction of mGluR4 with Munc18-1, synapsin I, and syntaxin, both on the C-terminal peptide and on the second intracellular loop. Each of these proteins may bind to mGluR4 either directly via one or several binding sites or indirectly. Moreover, the AN-23 peptide, which is in close proximity to the membrane, was also able to retain Munc18-1. This result is consistent with the work of Nakanishi and co-workers (13) that demonstrated a direct interaction between Munc18-1 and mGluR4 via the membrane proximal region of C-terminal receptor using a GST pulldown approach. It is interesting to note that AN-23 is predicted to exhibit an ␣-helix secondary structure, whereas VN-30 and GI-30 peptides are less structured. This feature probably favors the implication of AN-23 sequence in native interactions. Finally, our affinity chromatography experiments highlighted the role of the second intracellular loop as an additional interaction domain of Munc18-1 with mGluR4. Notably, SNAP-25 was not found with this approach, although this protein belongs to the SNARE complex. Whatever, these results confirm that mGluR4 clearly interacts with a pool of proteins involved in exocytosis.
In particular, Munc18-1 was identified as a partner of native mGluR4 in all three co-IP experiments. In addition, recombinant Munc18-1 covalently grafted on Sepharose matrix was able to retain native mGluR4, which consolidated the co-IP results and suggests a direct physical interaction between Munc18-1 and mGluR4, as already reported by Nakajima et al. (13), for recombinant Munc18-1 and C-terminal mGluR4 constructs. Moreover, we demonstrated that these proteins colocalized at the plasma membrane in HEK293 cells. Studies on SNAREs complexes reconstituted in liposomes have defined a facilitator effect of Munc18-1 on synaptic fusion (40). In vitro experiments have shown that Munc18-1 interacts with the syn- taxin-1 C-terminal part (41) but also with the SNARE complex itself (40,42). Recently Munc18-1 has been shown to interact with mGluR7 and mGluR4 and play a role in short term plasticity (13). In our affinity chromatography experiments, Munc18-1 was found to interact with two different regions of mGluR4; that is, the proximal C-terminal and the second intracellular loop. Based on these results, it would be informative to further study the mechanisms of interaction between Munc18-1 and mGluR4 in the negative regulation of neurotransmitter release.
Our results suggest that mGluR4 could interact directly with exocytosis proteins with the purpose of limiting neurotransmitter release by sequestering one or more of these proteins. Whether these interactions correspond to direct physical interactions should be further investigated to better understand the molecular mechanisms involved in this process. Taken together our results suggest that modulation of synaptic transmission by group III mGluRs may be at least partly independent of the classical G-protein transduction pathway.