Bidirectional modulation of glutamatergic synaptic transmission by metabotropic glutamate type 7 receptors at Schaffer collateral–CA1 hippocampal synapses

Abstract

Key points

• Neurotransmitter release is inhibited by metabotropic glutamate type 7 (mGlu₇) receptors that reduce Ca²⁺ influx, yet synapses lacking this receptor also produce weaker release, suggesting that mGlu₇ receptors may also prime synaptic vesicles for release. Prolonged activation of mGlu₇ receptors with the agonist l‐AP4 first reduces and then enhances the amplitude of EPSCs through a presynaptic effect.

• The inhibitory response is blocked by pertussis toxin, while the potentiating response is prevented by a phospholipase C inhibitor (U73122) and an inhibitor of diacylglycerol (DAG) binding (calphostin C), suggesting that this receptor also couples to pathways that generate DAG.

• Release potentiation is associated with an increase in the number of synaptic vesicles close to the plasma membrane, which was dependent on the Munc13‐2 and RIM1α proteins. The Glu7 receptors activated by the glutamate released following high frequency stimulation provoke a bidirectional modulation of synaptic transmission.

Abstract

Neurotransmitter release is driven by Ca²⁺ influx at synaptic boutons that acts on synaptic vesicles ready to undergo exocytosis. Neurotransmitter release is inhibited when metabotropic glutamate type 7 (mGlu₇) receptors provoke a reduction in Ca²⁺ influx, although the reduced release from synapses lacking this receptor suggests that they may also prime synaptic vesicles for release. These mGlu₇ receptors activate phospholipase C (PLC) and generate inositol trisphosphate, which in turn releases Ca²⁺ from intracellular stores and produces diacylglycerol (DAG), an activator of proteins containing DAG‐binding domains such as Munc13 and protein kinase C (PKC). However, the full effects of mGlu₇ receptor signalling on synaptic transmission are unclear. We found that prolonged activation of mGlu₇ receptors with the agonist l‐AP4 first reduces and then enhances the amplitude of EPSCs, a presynaptic effect that changes the frequency but not the amplitude of the mEPSCs and the paired pulse ratio. Pertussis toxin blocks the inhibitory response, while the PLC inhibitor U73122, and the inhibitor of DAG binding calphostin C, prevent receptor mediated potentiation. Moreover, this DAG‐dependent potentiation of the release machinery brings more synaptic vesicles closer to the active zone plasma membrane in a Munc13‐2‐ and RIM1α‐dependent manner. Electrically evoked release of glutamate that activates mGlu₇ receptors also bidirectionally modulates synaptic transmission. In these conditions, potentiation now occurs rapidly and it overcomes any inhibition, such that potentiation prevails unless it is suppressed with the PLC inhibitor U73122.

Key points

• Neurotransmitter release is inhibited by metabotropic glutamate type 7 (mGlu₇) receptors that reduce Ca²⁺ influx, yet synapses lacking this receptor also produce weaker release, suggesting that mGlu₇ receptors may also prime synaptic vesicles for release. Prolonged activation of mGlu₇ receptors with the agonist l‐AP4 first reduces and then enhances the amplitude of EPSCs through a presynaptic effect.

• The inhibitory response is blocked by pertussis toxin, while the potentiating response is prevented by a phospholipase C inhibitor (U73122) and an inhibitor of diacylglycerol (DAG) binding (calphostin C), suggesting that this receptor also couples to pathways that generate DAG.

• Release potentiation is associated with an increase in the number of synaptic vesicles close to the plasma membrane, which was dependent on the Munc13‐2 and RIM1α proteins. The Glu7 receptors activated by the glutamate released following high frequency stimulation provoke a bidirectional modulation of synaptic transmission.

Introduction

Metabotropic glutamate type 7 (mGlu₇) receptors belong to group III mGlu receptors (mGluRs), along with mGluR types 4, 6 and 8. These mGlu₇ receptors are G protein coupled receptors (GPCRs) that are widely expressed in the brain (Kinzie et al. 1995; Ohishi et al. 1995; Corti et al. 1998) and they localize to the presynaptic active zone of glutamatergic synapses (Shigemoto et al. 1996, 1997; Kinoshita et al. 1998) where they dampen neurotransmitter release (Millán et al. 2002, 2003; Martín et al. 2007). The inhibition of glutamate release mediated by mGlu₇ receptors requires activation of a pertussis toxin (PTx) sensitive Gi/o protein and it reduces Ca²⁺ channel activity (Perroy et al. 2000, 2002; Millán et al. 2002, 2003; Pelkey et al. 2006; Martín et al. 2007). This effect is consistent with the weaker synaptic transmission induced by the group III mGluR agonist l‐2‐amino‐4‐phosphonobutyrate (l‐AP4) (Baskys & Malenka, 1991; Gereau & Conn, 1995; Capogna, 2004; Pelkey et al. 2005; Ayala et al. 2008; Klar et al. 2015). Moreover, prolonged activation of mGlu₇ receptors de‐represses excitatory synaptic transmission due to receptor internalization (Pelkey et al. 2005). However, synaptic transmission is depressed in mGlu₇ receptor knockout mice when high frequency stimulation (HFS) activates this receptor at Schaffer collateral (SC)–CA1 hippocampal synapses, suggesting that mGlu₇ receptors also potentiate neurotransmitter release at these synapses (Bushell et al. 2002).

The mGlu₇ receptor can also couple to phosphatidylinositol bisphosphate (PIP₂) hydrolysis through the activation of phospholipase C (PLC: Perroy et al. 2000; Martín et al. 2010; Ferrero et al. 2011), although the effects of this signalling on synaptic transmission have not yet been addressed. PLC activation hydrolyses phospholipid PIP₂ to generate inositol trisphosphate (IP₃) and diacylglycerol (DAG), the former releasing Ca²⁺ from intracellular stores while the latter activates proteins containing DAG‐binding domains, such as Munc13 and protein kinase C (PKC). Phorbol esters, the stable analogues of DAG, enhance synaptic transmission through PKC (Malenka et al. 1986; Herrero et al. 1992; Parfitt & Madison, 1993) and through the non‐kinase active zone protein Munc13, which is essential for synaptic vesicle (SV) priming (Betz et al. 1998; Wu & Wu, 2001; Rhee et al. 2002; Wierda et al. 2007; Lou et al. 2008). Therefore, it might be anticipated that the coupling of mGlu₇ receptors to PLC and the ensuing generation of DAG would potentiate synaptic transmission. However, there is currently no evidence of such PLC‐coupled mGlu₇ receptor mediated potentiation of synaptic transmission. Nevertheless, group III mGluRs at the calyx of Held do augment the readily releasable pool (RRP) of SVs, although with little impact on synaptic amplitude since the receptor also decreases the release probability (Billups et al. 2005).

Here we studied how mGlu₇ receptors influence synaptic transmission at SC–CA1 hippocampal pyramidal cell synapses where these receptors are expressed strongly (Shigemoto et al. 1997). Indeed, this receptor is the only group III mGluR present at adult SC–CA1 synapses (Baskys & Malenka, 1991; Ayala et al. 2008). We found that mGlu₇ receptors bidirectionally modulate synaptic transmission at SC–CA1 hippocampal synapses. Indeed, receptor activation with the agonist l‐AP4 first reduces and then enhances the excitatory postsynaptic current (EPSC) amplitude through a presynaptic effect. The inhibition of release reflects the rapid coupling of the receptor to a PTx sensitive G protein, while mGlu₇ mediated potentiation is slower and results from receptor coupling to a PTx resistant pathway. Synaptic potentiation, which is prevented by inhibitors of PLC‐ and DAG‐binding, drives SVs closer to the active zone (AZ) plasma membrane in a Munc13‐2‐ and RIM1α‐dependent manner. Both these responses (release inhibition and potentiation) provoked by mGlu₇ receptors occur simultaneously following electrically evoked glutamate release, and while potentiation prevails, inhibition can be revealed by blocking this potentiation.

Methods

Animals

All procedures for handling and killing animals were performed in accordance with European Commission guidelines (2010/63/UE) and they were approved by the Animal Research Committee at the Complutense University. The animal experiments also comply with the policies and regulations of The Journal of Physiology (Grundy, 2015). Hippocampal slices were obtained from C57BL/6 (30‐ to 45‐day‐old) wild‐type (WT) mice and in some experiments, slices from Munc13‐2 and RIM1α knockout (KO) mice were used (generously provided by Dr N. Brose (Varoqueaux et al. 2002) and Dr S. Schoch (Schoch et al. 2002), respectively). In all instances adequate measures were taken to minimize the animal's pain and discomfort.

Hippocampal slice preparation

Animals were anaesthetized with isoflurane (1.5–2% in a mixture of 80% synthetic air–20% oxygen) and decapitated. The brain was quickly removed and placed in ice‐cold artificial CSF (aCSF, in mm: NaCl 124, KCl 2.69, KH₂PO₄ 1.25, MgSO₄ 2, NaHCO₃ 26, CaCl₂ 2, ascorbic acid 0.4, and glucose 10 (pH 7.3)). Coronal vibratome hippocampal slices (325 μm thick: Leica VT 1200S, Leica Microsystems, Wetzlar, Germany) were obtained and maintained at room temperature (21–24°C) in a holding chamber (≥1 h) in aCSF continuously bubbled with carbogen (95% O₂ and 5% CO₂). The slices were then transferred to an immersion recording chamber and superfused at 1 mL min⁻¹ with gassed aCSF containing 50 μm picrotoxin to block GABA_A receptors. To measure miniature EPSCs (mESPCs: Fig. 3 A‐G), tetrodotoxin (TTx, 1 μm) was added to the aCSF. Pyramidal CA1 cells were visualized on an Eclipse FN1 Nikon microscope using a 40× water immersion objective with a Nomarski condenser combined with infrared microscopy (Nikon Instruments, Tokyo, Japan).

Figure 3. The mGlu₇ receptors increase the frequency but not the amplitude of mEPSCs.

A, l‐AP4 (500 μm, 10 min) increases the mEPSC frequency. B, quantification of the changes in the mEPSC frequency before and after l‐AP4 (500 μm, 10 min, P = 0.0101, n = 21 cells/21 slices/10 mice). C, cumulative frequency plots showing the increase in the mEPSC frequency induced by l‐AP4 (500 μm, 10 min). D, sample traces recorded before and after exposure to l‐AP4 (500 μm, 10 min). E, l‐AP4 (500 μm, 10 min) fails to alter the mEPSC amplitude. F, quantification of the changes in mEPSC amplitude before and after exposure to l‐AP4 (500 μm, 10 min, P = 0.7016). G, cumulative frequency plots show that l‐AP4 has no effect on the mEPSC amplitude. The data were compared to the basal state prior to l‐AP4 addition using an unpaired Student's t test and the values represent the mean ± SD. Scale in D: 10 pA and 10 ms.

Electrophysiology experiments

Pyramidal neurons from the hippocampal CA1 region were identified by their morphology and characteristic electrophysiological properties, and recorded in the whole‐cell configuration. Series and input resistances were monitored throughout the experiment using a −5 mV pulse, and the recordings were considered stable when the series and input resistances, resting membrane potential and stimulus artifact duration did not change >20%. Cells that did not meet these criteria were discarded.

Patch pipettes were pulled from thick‐walled borosilicate glass (1.5 mm outer diameter) on a P‐97 puller (Sutter Instrument, Novato, CA, USA). The pipettes (3–4 MΩ) were filled with the internal solution containing (in mm): potassium gluconate 135, KCl 10, Hepes 10, MgCl₂ 1, ATP‐Na 2 (pH adjusted to 7.3 with KOH; osmolarity 280–290 mOsm L⁻¹). Recordings were obtained with a PC‐ONE amplifier under voltage‐clamp conditions and the membrane potential was held at −70 mV to record glutamatergic evoked EPSCs (eEPSCs) from SC afferents. Signals were fed to a Pentium‐based PC through a DigiData1322A interface board (Molecular Devises, Sunnyvale, CA, USA). The pCLAMP 10.2 software was used to generate stimuli, as well as for data display, acquisition, storage and analysis. Experiments were performed at 25°C using a temperature controller (Warner Instruments, Hamden, CT, USA).

Theta capillaries (2–5 μm tip) filled with aCSF were used for bipolar stimulation. The electrodes were connected to a stimulator (S38, Grass, Artisan Technology Group, Champaign, IL, USA) through an isolation unit and placed in the stratum radiatum layer at the border of the CA3–CA1 regions to stimulate glutamatergic afferents. Stimuli were delivered at 0.33 Hz and the paired pulse ratio (PPR) was obtained as (2nd EPSC/1st EPSC) after delivering paired pulses (50 ms interval) at 0.33 Hz. The stimulus intensity was adjusted to yield EPSC amplitudes between 20 and 200 pA and it remained unchanged during the experiment. To activate mGlu₇ receptors by electrically induced release of glutamate, a HFS protocol was applied that consisted of two (1 s at 50 Hz, every 20 s) or three trains (1 s at 100 Hz, every 20 s).

For statistical analysis, an average of 30 consecutive EPSCs were compared (normalized to the baseline), delivered at 0.33 Hz (90 s) before, during (90 s after the onset of the treatment) and 10 min after l‐AP4 perfusion. In the mEPSC studies (Fig. 3), the frequency and mean amplitude of the events that appeared in the 5 min prior to and after l‐AP4 application were compared. In all the electrophysiology experiments only one neuron was analysed per slice.

Analysis of synaptic vesicle distribution at the active zone by electron microscopy

Hippocampal slices (325 μm thick) obtained as described above for the electrophysiology experiments were transferred to an immersion recording chamber and superfused at 1 mL min⁻¹ with gassed aCSF including 50 μm picrotoxin. In some cases, the slices were also treated with 500 μm l‐AP4 for 10 min. The slices were fixed immediately afterwards by immersion for 45 min at 37°C in 3.5% glutaraldehyde in 0.1 m phosphate buffer (PB, pH 7.4) and they were then left in glutaraldehyde solution for 30 min at room temperature before storing them for 20 h at 4°C. The slices were then rinsed six times with large volumes of 0.1 m PB and the hippocampus was dissected out, post‐fixing it in 1% OsO₄–1.5% K₃Fe(CN)₆ for 1 h at room temperature. After dehydrating through a graded series of ethanol (30, 50, 70, 80, 90, 95 and 100%), the samples were then embedded using the SPURR embedding kit (TAAB, Aldermaston, UK). Ultrathin ultramicrotome sections (70–80 nm thick: Leica EM UC6 Leica Microsystems, Wetzlar, Germany) were stained routinely with uranyl acetate and lead citrate, and images were obtained on a Jeol 1010 transmission electron microscope (Jeol, Tokyo, Japan). Randomly chosen areas of the CA1 striatum radiatum were then photographed at 80,000× magnification and only asymmetric synapses with clearly identifiable electron‐dense postsynaptic densities were analysed by measuring them with ImageJ software. The relative percentage of SVs per AZ was calculated in 10 nm bins at the AZ of the inner layer membrane and the total number of SVs per synaptic terminal was also determined. The data were analysed blind to the genotype and treatment.

Immunohistochemistry for electron microscopy

Electron microscopy examination of immunoreactivity for mGlu₇, Munc13‐1 and Munc13‐2 in the hippocampus was performed as described previously, using the pre‐embedding immunoperoxidase and immunogold double‐labelling method (Luján et al. 1996). Briefly, 30‐day‐old C57BL/6J mice were perfused with 4% paraformaldehyde, 0.05% glutaraldehyde and 15% (v/v) saturated picric acid in 0.1 m PB. Vibratome brain sections (60 μm) were processed for immunohistochemical detection of Munc13‐1 or Munc13‐2 using the horseradish peroxide (HRP) reaction, and of mGlu₇ using a silver‐enhanced immunogold reaction. Briefly, sections were incubated for 1 h in 10% normal goat serum (NGS) diluted in Tris buffer (TB, 50 mm, pH 7.4) containing 0.9% NaCl (TBS) and 0.2% Triton X‐100. The sections were incubated in a mixture of anti‐mGlu₇ (1–2 μg mL⁻¹ diluted in TBS containing 1% NGS) and anti‐Munc13‐1 or anti‐Munc13‐2 (1–2 μg mL⁻¹ diluted in TBS containing 1% NGS). Subsequently, they were incubated overnight at 4°C with a mixture of goat anti‐mouse IgG coupled to 1.4 nm gold (Nanoprobes Inc., Stony Brook, NY, USA) and biotinylated goat anti‐rabbit IgG (Vector Laboratories, Burlingame, CA, USA) in TBS containing 1% NGS. After several washes in TBS, the sections were post‐fixed in 1% (v⁄v) glutaraldehyde and washed in double‐distilled water, prior to performing silver enhancement of the gold particles with an HQ Silver kit (Nanoprobes Inc.). Subsequently, the sections were incubated with an avidin–biotin–peroxidase complex diluted 1:100 (ABC kit, Vector Laboratories). Bound peroxidase enzyme activity was revealed using 3,3′‐diaminobenzidine tetrahydrochloride (DAB: 0.05% in TB, pH 7.4) as the chromogen and 0.01% H₂O₂ as the substrate. After washing with TB and PB, the sections were then treated with osmium tetroxide (1% in 0.1 m PB), block‐stained with uranyl acetate, dehydrated in a graded series of ethanol and flat‐embedded in Durcupan resin (Fluka). Regions of interest were cut at 70–90 nm on an ultramicrotome (Reichert Ultracut E, Leica, Austria) and collected on single slot pioloform‐coated copper grids. Staining was performed on drops of 1% aqueous uranyl acetate followed by Reynolds's lead citrate, and the ultrastructure was analysed in a Jeol‐1010 electron microscope.

To test the specificity of the electron microscopy staining, the primary antibody was omitted or replaced with 5% (v/v) NGS from the species of the primary antibody. Under these conditions, no selective labelling was observed. In addition, some sections were incubated with only gold‐labelled and biotinylated secondary antibodies, followed by the ABC complex and peroxidase reaction without silver intensification. This resulted in an amorphous HRP end‐product, and no metal particles were detected. When the same protocol was performed but with silver intensification and without the HRP reaction, silver grains were evident but there was no amorphous HRP end‐product. Under these conditions, only infrequent small patches of HRP end‐product were detected and the patches were not selectively associated with any particular cell profile. In addition, the selective location of the signals in structures labelled with only one or the other of the signalling products in the same section, as well as when there were adjacent double‐labelled structures, indicated that the method used does not produce false‐positive double‐labelling results.

Quantitative analysis at EM level

To establish the relative abundance of mGlu₇ in hippocampal axon terminals expressing Munc13‐1 and Munc13‐2, immunolabelling of the CA1 stratum radiatum was quantified in 60 μm coronal slices, as described previously (Luján et al. 1996). Briefly, three tissue samples from three animals were embedded in blocks, totalling nine blocks. To minimize false negatives, serial ultrathin electron microscopy sections were cut close to the surface of each block as the immunoreactivity decreased with depth. We estimated the quality of the immunolabelling by always selecting areas with optimal gold labelling at approximately the same distance from the cutting surface, within 5–10 μm of the surface. Randomly selected areas were then photographed from the selected ultrathin sections at a final magnification of 50,000×. To determine the frequency of mGlu₇ in axon terminals, immunogold labelling was quantified relative to the total area of ∼5000 μm². Munc13‐1‐ and Munc13‐2‐positive axon terminals were assessed for the presence of mGlu₇ immunoparticles and the percentage of these terminals containing mGlu₇ immunoparticles was calculated in the hippocampal stratum radiatum. In total, 75 dendrites were selected and the area of the individual dendritic profiles was measured. The percentage of immunoparticles for HCN1 at postsynaptic and presynaptic sites was calculated in layer III.

Statistical analysis

The data are expressed as the mean ± standard deviation (SD) and the results were compared using an unpaired Student's t test. Statistical differences were established at P < 0.05 (^*), P < 0.01 (^**) and P < 0.001 (^***).

Results

While group III mGluRs with low and high affinity for the l‐AP4 agonist inhibit synaptic transmission at hippocampal SC–CA1 synapses in young animals (Baskys & Malenka, 1991), only mGlu₇ receptors inhibit such synaptic transmission in adult mice (Ayala et al. 2008). Thus, adult (>30‐day‐old) SC–CA1 synapses are suitable to study mGlu₇ receptor mediated modulation of synaptic transmission without interference from the high l‐AP4‐affinity mGlu₄ and mGlu₈ receptors. l‐AP4 (20 μm) did not alter the EPSC amplitude at adult (30‐ to 45‐day‐old) SC–CA1 hippocampal synapses (98.0 ± 15.3% of the baseline, n = 9 cells/9 slices/9 mice, P = 0.9013: Fig. 1 A and B). However, short applications (3 min) of l‐AP4 (500 μm) did reversibly reduce the EPSC amplitude (59.3 ± 20.7% of the baseline, n = 9 cells/9 slices/9 mice, P = 0.0004: Fig. 1 A and B). We also found that prolonged application (10 min) of l‐AP4 (500 μm) caused a rapid reduction in the EPSC amplitude (60.5 ± 13.8% of the baseline, n = 12 cells/12 slices/10 mice; P ˂ 0.0001: Fig. 1 C and D) that subsequently faded, resulting in a net increase (139.2 ± 25.2% of the baseline, n = 12 cells/12 slices/10 mice, P = 0.0002: Fig. 1 C and D). The specific group III mGluR antagonist 2‐Methyl‐O‐phosphonoserine (MSOP) completely abolished the responses to l‐AP4 (basal 100.4 ± 12.9%; inhibition 99.1 ± 17.8%, P = 0.8886; potentiation 99.0 ± 13.4%, P = 0.8586, n = 6 cells/6 slices/6 mice: Fig. 1 C and D). This bidirectional modulation of EPSC amplitude by l‐AP4 is associated with contrasting changes in the EPSC2/EPSC1 PPR (basal, 1.42 ± 0.31; inhibition, 2.12 ± 0.27 (P ˂ 0.0001); potentiation, 0.99 ± 0.31, P = 0.0027: Fig. 1 E and F). Hence, this bidirectional modulation of EPSC amplitude by mGlu₇ receptors appeared to be due to changes in neurotransmitter release.

Figure 1. The potentiation but not the inhibition of synaptic transmission by mGlu₇ receptors requires prolonged activation with l‐AP4.

EPSCs were recorded from CA1 pyramidal cells after stimulation of Schaffer collaterals with an electrode placed in the stratum radiatum. A, the application of a low concentration of l‐AP4 (20 μm, 3 min) does not produce any change, whereas a higher concentration of l‐AP4 (500 μm, 3 min) reversibly reduces the EPSC amplitude. EPSC sample traces represent the mean of 30 consecutive EPSCs at 0.33 Hz 90 s before (1) and 90 s after the addition of l‐AP4 at 20 μm (2) or 500 μm (3). B, quantification of the changes in EPSC amplitude with l‐AP4 (20 μm, 3 min, P = 0.9013 or 500 μm, 3 min, P = 0.0004: n = 9 cells/9 slices/9 mice). C, a longer application of l‐AP4 (500 μm, 10 min) first reduces and then increases the amplitude of the EPSC. The group III mGluR antagonist MSOP (200 μm, 30–60 min) blocked the l‐AP4 induced changes in EPSC amplitude. The EPSC sample traces represent the mean of 30 consecutive EPSCs at 0.33 Hz from experiments on control (black) and MSOP treated slices (grey) taken 90 s before (1, basal) or 90 s after the addition of l‐AP4 (2, inhibition), and 10 min after l‐AP4 perfusion (3, potentiation). D, quantification of the changes in EPSC amplitude induced by l‐AP4 (500 μm, 10 min) during inhibition (2) and potentiation (3) in both control (P ˂ 0.0001 and P = 0.0002) and MSOP treated slices (P = 0.8886 and P = 0.8586, respectively): control, n = 12 cells/12 slices/10 mice; MSOP treated slices, n = 6 cells/6 slices/6 mice. E, the data shown in C were used to determine the l‐AP4 induced changes in the paired pulse ratio (PPR: EPSC₁/EPSC₂) from control and MSOP treated slices. The sample traces are the mean of 30 consecutive EPSCs. F, quantification of the changes in the PPR induced by l‐AP4 (500 μm, 10 min) during inhibition (2) and potentiation (3) in both control (P ˂ 0.0001 and P = 0.0027) and MSOP treated slices (P = 0.8886 and P = 0.8586, respectively). The data were compared to the basal conditions prior to l‐AP4 addition using an unpaired Student's t test and the values represent the mean ± SD. Scale bars in A, C, E: 50 pA and 15 ms.

l‐AP4 is a weak agonist of NMDA receptors (NMDARs: Lafon‐Cazal et al. 1999). However, the l‐AP4 induced bidirectional modulation of synaptic transmission does not involve NMDARs, as l‐AP4‐induced inhibition (63.0 ± 9.8% of the baseline, P = 0.0013) and potentiation (141.3 ± 29.0% of the baseline, P = 0.0381, n = 5 cells/5 slices/4 mice) was unchanged in the presence of the NMDAR antagonist d‐2‐amino‐5‐phosphonovalerate (d‐AP5) (50 μm: Fig. 2 A and B). l‐AP4 can also activate mGlu₇ receptors at GABA terminals and reduce GABA release (Klar et al. 2015), thereby attenuating the inhibition of glutamate release via GABA_B receptors. However, again l‐AP4‐induced inhibition (61.7 ± 8.6% of the baseline, P = 0.0003) and potentiation of synaptic transmission (135.6 ± 11.3% of the baseline, P = 0.0157, n = 5 cells/5 slices/3 mice) was not altered in the presence of the GABA_B antagonist CGP 55845 (5 μm: Fig. 2 C and D). Hence, it appears that neither NMDARs nor GABA_B receptors are involved in the bidirectional modulation of synaptic transmission induced by l‐AP4.

Figure 2. There is no change in the inhibition and potentiation of synaptic transmission by l‐AP4 in the presence of the NMDAR and GABA_B receptor antagonists d‐AP5 and CGP 55845.

A, l‐AP4 (500 μm, 10 min) first reduces and then increases the amplitude of the EPSC in the continued presence of the NMDAR antagonist d‐AP5 (50 μm). B, quantification of the changes in EPSC amplitude induced by l‐AP4 (500 μm, 10 min) during inhibition (2) and potentiation (3: P = 0.0013 and P = 0.0381, n = 5 cells/5 slices/4 mice). C, l‐AP4 (500 μm, 10 min) first reduces and then increases the amplitude of the EPSC in the continued presence of the GABA_B receptor antagonist CGP 55845 (5 μm). D, quantification of the changes in EPSC amplitude induced by l‐AP4 (500 μm, 10 min) during inhibition (2) and potentiation (3: P = 0.0003 and P = 0.0157, n = 5 cells/5 slices/3 mice). In A and C, the EPSC sample traces represent the mean of 30 consecutive EPSCs at 0.33 Hz taken 90 s before (1, basal) or 90 s after the addition of l‐AP4 (2, inhibition), and 10 min after the end of l‐AP4 addition (3, potentiation). The data were compared to the basal state prior to l‐AP4 addition using an unpaired Student's t test and the values represent the mean ± SD.

l‐AP4 increased the frequency of the mEPSCs measured with the Na⁺ channel blocker tetrodotoxin (TTx, 1 μm; 1.08 ± 0.25 Hz to 2.25 ± 0.82 Hz, n = 21 cells/21 slices/10 mice, P  =  0.0101: Fig. 3 A–C) without changing the mEPSC amplitude (before 11.84 ± 2.1 pA, after 11.64 ± 1.8 pA, P = 0.7420: Fig. 3 D–G). Moreover, l‐AP4 did not decrease the mEPSC frequency because voltage dependent Ca²⁺ channels (VDCCs) that mediate release inhibition by mGlu₇ (Millán et al. 2002) are not active in the presence of TTx. Indeed, the VDCC blocker Cd²⁺ (100 μm) failed to decrease the mEPSC frequency (1.18 ± 0.40 Hz in the presence and 1.16 ± 0.48 Hz in the absence of Cd²⁺, n = 4 cells/4 slices/2 mice: data not shown: see also Parfitt & Madison, 1993; Gereau & Conn, 1995). Thus, the modulation of synaptic transmission by the mGlu₇ receptor has a presynaptic origin.

There was an increase in the amplitude of the second EPSC during potentiation compared to basal conditions in PPR experiments (Fig. 1 E), suggesting that an increase in the RRP size may contribute to synaptic strength. The RRP size was calculated from cumulative amplitude plots as the y‐intercept from a linear fit of the steady state levels attained during the high frequency train (Schneggenburger et al. 2002; Gioia et al. 2016). The RRP size of l‐AP4 (500 μm, 10 min) treated slices stimulated at 40 Hz for 2.5 s increased from 1150.2 (±268.3) to 1589.7 (±309.9) pA, P = 0.0035: Fig. 4 A and B). Hence, an increase in the number of release competent SVs may contribute to the mGlu₇ receptor mediated potentiation of synaptic transmission.

Figure 4. The mGlu₇ receptor mediated potentiation of synaptic transmission involves an increase in the RRP size.

Slices were treated with l‐AP4 (500 μm, 10 min) and after a 5 min wash, they were stimulated at 40 Hz for 2.5 s. The RRP size was calculated from the cumulative amplitude plots as the y‐intercept from a linear fit of the steady state level attained during the high frequency train. A, cumulative amplitude plot of a 2.5 s 40 Hz train in control and l‐AP4 treated cells. B, quantification of the changes in RRP size induced by treating cells with l‐AP4 (500 μm, 10 min, P = 0.0035): control, n = 10 cells/10 slices/4 mice; l‐AP4, n = 10 cells/10 slices/4 mice. The data were compared to the controls using an unpaired Student's t test and the values represent the mean ± SD.

PLC inhibition impairs mGlu₇ receptor mediated potentiation of synaptic transmission

While mGlu₇ receptors activate PLC (Perroy et al. 2000; Martín et al. 2010), it is not clear how this signalling affects neurotransmitter release as both inhibition (Perroy et al. 2000) and potentiation (Martín et al. 2010) have been reported. l‐AP4 caused a sustained reduction in the EPSC amplitude (62.3 ± 17.0% of the baseline, n = 8 cells/8 slices/8 mice, P ˂ 0.0001: Fig. 5 A and B) and it failed to increase this parameter in slices treated with the PLC inhibitor U73122 (2 μm, 30 min; 97.9 ± 26.8%, n = 8 cells/8 slices/8 mice, P = 0.7689: Fig. 5 A and B). By contrast, l‐AP4 did provoke transient inhibition in the presence of the inactive PLC inhibitor U73343 (60.8 ± 9.5% of baseline, n = 6 cells/6 slices/6 mice, P = 0.011: Fig. 5 A and B) followed by potentiation of the EPSC amplitude (142.5 ± 16.3% of the baseline, n = 6 cells/6 slices/6 mice, P = 0.0434: Fig. 5 A and B). Thus, PLC appeared to be required for mGlu₇ receptor mediated potentiation of synaptic transmission at SC–CA1 synapses. Moreover, these results also indicate that the fading of mGlu₇ receptor mediated inhibition could be the result of the simultaneous coupling of the receptor to an additional signalling pathway that potentiates synaptic transmission, first counteracting the release inhibition and then resulting in a net potentiation.

Figure 5. The l‐AP4‐induced potentiation is sensitive to the PLC inhibitor U73122 and to the DAG binding inhibitor calphostin C, but not to pertussis toxin.

A, in the presence of the PLC inhibitor U73122 (2 μm, 30 min), l‐AP4 (500 μm, 10 min) fails to enhance the EPSC amplitude but, rather, it causes sustained inhibition of this parameter. By contrast, the inactive PLC inhibitor U73343 (2 μm, 30 min) caused a transient inhibition that subsequently faded, resulting in the net potentiation of the EPSC amplitude. B, quantification of the changes in EPSC amplitude induced by l‐AP4 (500 μm, 10 min) during inhibition (2) and potentiation (3) in the presence of U73122 (P ˂ 0.0001 and P = 0.7689, n = 6 cells/6 slices/6 mice) or U73343 (P = 0.011 and P = 0.0434, n = 8 cells/8 slices/8 mice). C, in the presence of PTx (2 μg mL⁻¹ for 2–4 h at 30°C), l‐AP4 (500 μm, 10 min) failed to reduce, but rather it enhanced, the EPSC amplitude, whereas in control slices incubated for 2–4 h at 30°C, l‐AP4 (500 μm, 10 min) caused a transient inhibition that subsequently faded into a net potentiation of the EPSC amplitude. D, quantification of the changes in EPSC amplitude induced by l‐AP4 (500 μm, 10 min) during inhibition (2) and potentiation (3) in both PTx treated (P = 0.1709 and P = 0.0048, n = 8 cells/8 slices/8 mice) and in control slices (P ˂ 0.0001 and P = 0.0002, n = 8 cells/8 slices/8 mice). E, in the presence of the DAG binding inhibitor calphostin C (0.1 μm, 30 min), l‐AP4 (500 μm, 10 min) induced sustained inhibition of the EPSC amplitude and it failed to enhance the EPSC amplitude. By contrast, the ATP binding antagonist and PKC inhibitor RO 31‐8220 (1 μm, 30 min) did not alter the l‐AP4 responses, as the transient inhibition was followed by a net potentiation of the EPSC amplitude. F, quantification of the changes in EPSC amplitude induced by l‐AP4 (500 μm, 10 min) during inhibition (2) and potentiation (3) in the presence of calphostin C (P ˂ 0.0001 and P = 0.4741, n = 10 cells/10 slices/10 mice) or RO 31‐8220 (P = 0.0001 and P ˂ 0.0001, n = 9 cells/9 slices/6 mice). In A, C and E, the EPSC sample traces represent the mean of 30 consecutive EPSCs at 0.33 Hz taken 90 s before (1, basal) or 90 s after the addition of l‐AP4 (2, inhibition), and 10 min after l‐AP4 perfusion (3, potentiation). The data were compared to the basal state (1) using an unpaired Student's t test and the values represent the mean ± SD. Scale in A, C and E: 50 pA and 15 ms.

Pertussis toxin impairs inhibition but not mGlu₇ receptor mediated potentiation of synaptic transmission

As mGlu₇ receptors coupled to Gi/o proteins inhibit Ca²⁺ channels and adenylyl cyclase (Perroy et al. 2000; Millán et al. 2002, 2003; Capogna, 2004; Pelkey et al. 2005; Martín et al. 2007, 2010; Kammermeier, 2015), we tested whether blocking Gi/o proteins with PTx alters mGlu₇ receptor mediated potentiation of synaptic transmission. In hippocampal slices incubated with PTx (2 μg mL⁻¹ for 2–4 h at 30°C), l‐AP4 failed to reduce (103.8 ± 9.9% of the baseline, n = 8 cells/8 slices/8 mice, P = 0.1709: Fig. 5 C and D), while the increase of EPSC amplitude was not altered (137.0 ± 26.4% of the baseline, n = 8 cells/8 slices/8 mice, P = 0.0048: Fig. 5 C and D). By contrast, l‐AP4 first inhibited (57.1 ± 8.9% of the baseline, n = 8 cells/8 slices/8 mice, P ˂ 0.0001: Fig. 5 C and D) and then potentiated the amplitude of the EPSCs in control slices that were not exposed to PTx (135.4 ± 14.9% of the baseline, n = 8 cells/8 slices/8 mice, P = 0.0002: Fig. 5 C and D). Therefore, the modulation of EPSCs by mGlu₇ receptors is the result of the temporal summation of PTx‐sensitive and PTx‐resistant components of synaptic transmission. The PTx‐sensitive component develops rapidly upon exposure to l‐AP4 as a consequence of the well‐established inhibition of Ca²⁺ channels by mGlu₇ receptors that activate Gi/o proteins (Okamoto et al. 1994; Millán et al. 2002, 2003; Perroy et al. 2002; Martín et al. 2007, 2010; Kammermeier, 2015). By contrast, the PTx‐resistant component of synaptic transmission develops more slowly, probably reflecting the mGlu₇ receptor signalling pathway described in synaptosomes, a pathway targeting the release machinery through PIP₂ hydrolysis coupled to the generation of IP₃ and DAG (Martín et al. 2010).

The inhibitor of DAG binding, calphostin C, prevents mGlu₇ receptor mediated potentiation of synaptic transmission

To further understand the mechanism underlying mGlu₇ receptor mediated potentiation of synaptic transmission, we distinguished between PKC and Munc13 (Lou et al. 2008) as targets of the DAG generated by mGlu₇ receptors. To study these targets, we used the PKC inhibitor RO 31‐8220, which prevents ATP binding, and calphostin C, which prevents DAG binding and therefore also inhibits non‐kinase DAG‐binding proteins. Slices exposed to RO 31‐8220 (1 μm, 30 min) responded to l‐AP4 with a transient inhibition of the EPSC amplitude (62.9 ± 7.6% of the baseline, n = 9 cells/9 slices/6 mice, P = 0.0001: Fig. 5 E and F), which then faded into net potentiation (139.8 ± 13.3% of the baseline, n = 9 cells/9 slices/6 mice, P ˂ 0.0001: Fig. 5 E and F). By contrast, the inhibition of the EPSC amplitude in hippocampal slices exposed to calphostin C (0.1 μm, 30 min) was sustained until the agonist was removed (59.7 ± 10.1% of the baseline, n = 10 cells/10 slices/10 mice, P ˂ 0.0001: Fig. 5 E and F), with no potentiation of the EPSC amplitude (94.7 ± 23.1% of the baseline, n = 10 cells/10 slices/10 mice, P = 0.4741: Fig. 5 E and F). These data suggest that Munc13 proteins and not PKC participate in mGlu₇ receptor‐mediated potentiation of synaptic transmission at SC–CA1 hippocampal synapses.

Lack of mGlu₇ receptor mediated potentiation in Munc13‐2 KO mice

The AZ protein Munc13‐1 is a phorbol ester receptor essential for SV priming that plays an important role in neurotransmitter release (Betz et al. 1998; Rhee et al. 2002). There are three Munc13 isoforms expressed in the brain (Munc13‐1–3), although only Munc13‐1 and Munc13‐2 are found in the hippocampus (Augustin et al. 1999b). However, Munc13‐1 KO mice (Augustin et al. 1999a) are not viable and thus they cannot be analysed by electrophysiology. Munc13‐2 KO mice are viable and basal synaptic transmission at SC–CA1 hippocampal synapses is unaffected by the absence of Munc13‐2 (Varoqueaux et al. 2002; Breustedt et al. 2010). Thus, we tested whether mGlu₇ receptor modulation of synaptic transmission is maintained in slices lacking Munc13‐2. In hippocampal slices from Munc13‐2 KO mice there was a sustained dampening of the EPSC amplitude in the presence of l‐AP4 (65.5 ± 6.5% of the baseline, n = 11 cells/11 slices/4 mice, P < 0.0001: Fig. 6 A and B) and no subsequent potentiation of the EPSC amplitude (97.7 ± 13.3% of the baseline, n = 11 cells/11 slices/4 mice, P = 0.5488: Fig. 6 A and B). By contrast, WT littermates showed bidirectional modulation of the EPSC amplitude (Fig. 6 A and B). We also assessed whether mGlu₇ receptor mediated potentiation of synaptic transmission was related to changes in SV distribution. In control slices from WT littermates, l‐AP4 increased the number of SVs close to the plasma membrane (from 6.8 ± 1.5%, n = 86 synapses/2 control mice, to 10.9 ± 1.2%, 90 synapses/2 l‐AP4 mice, P < 0.0001: Fig. 6 C–E). Conversely, in slices from Munc13‐2 KO mice there was a similar number of SVs close to the AZ plasma membrane to those in the WT (6.6 ± 1.2%, n = 145 synapses/3 mice, P = 0.7551: Fig. 6 G and I) but l‐AP4 did not increase the number of SVs close to the plasma membrane (6.9 ± 1.2%, n = 147 synapses/3 mice, P = 0.5815: Fig. 6 H and I). Interestingly, l‐AP4 did not alter the number of SVs at the AZ in WT (38.2 ± 4.8 and 34.5 ± 3.9, P = 0.0745 before and after exposure to l‐AP4: Fig. 6 F) or in the Munc13‐2 KO synapses (36.8 ± 3.9 and 34.6 ± 3.6, P = 0.2147 before and after l‐AP4: Fig. 6 J). Thus, mGlu₇ receptor mediated potentiation of synaptic transmission requires an increase in the size of the RRP of SVs that does not occur in cells lacking Munc13‐2.

Figure 6. The lack of mGlu₇ receptor mediated potentiation of synaptic transmission in Munc13‐2 knockout mice is associated with the failure of l‐AP4 to increase the number of synaptic vesicles close to the active zone plasma membrane.

A, l‐AP4 fails to enhance the EPSC amplitude in Munc13‐2 KO mice. The EPSC sample traces represent the mean of 30 consecutive EPSCs at 0.33 Hz in slices from wild‐type (WT) and Munc13‐2 KO mice taken 90 s prior to (1, basal), 90 s after (2, inhibition) and 10 min after l‐AP4 perfusion (3, potentiation). B, quantification of the changes in the EPSC amplitude induced by l‐AP4 (500 μm, 10 min) in WT (n = 5 cells/5 slices/3 mice) and Munc13‐2 KO mice (n = 11 cells/11 slices/4 mice), during inhibition (2, P = 0.0002 and P = 0.0397) and potentiation (3, P < 0.0001 and P = 0.5488). The data were compared to the basal state (1) using an unpaired Student's t test. C and D, l‐AP4 treatment brings SVs close to the plasma membrane in WT mice. E, quantification of the l‐AP4 induced increase (P < 0.0001) in SV distribution less than 10 nm from the AZ membrane. F, l‐AP4 does not change (P = 0.0745, n = 86 synapses/2 mice) the total number of SVs per AZ in WT mice compared to the controls (n = 90 synapses/2 mice). The data from l‐AP4 treated slices were compared with the control slices (untreated) using an unpaired Student's t test. G and H, l‐AP4 treatment fails to bring SVs less than 10 nm from AZ plasma membrane in Munc13‐2 KO mice. I, quantification of the l‐AP4 induced effect on SV distribution in Munc13‐2 KO mice. l‐AP4 fails (P = 0.5815) to bring SVs to less than 10 nm from the AZ membrane. J, l‐AP4 does not change (P = 0.2147, n = 145 synapses/3 mice) the total number of SVs per AZ in Munc13‐2 KO mice (n = 147 synapses/3 mice). The values represent the mean ± SD and the data from l‐AP4 treated slices were compared with control (untreated) slices using an unpaired Student's t test. Scale in A: 50 pA and 15 ms. Scale bar in C, D, G and H: 100 nm.

The mGlu₇ receptors co‐localize with Munc13‐2 and Munc13‐1 at glutamatergic nerve terminals

The precise subcellular localization of mGlu₇ receptors in axon terminals expressing Munc13‐1 or Munc13‐2 was studied by double labelling pre‐embedding immunogold/HRP electron microscopy. Munc13‐1, Munc13‐2 and mGlu₇ immunoreactivity was found at presynaptic sites (Fig. 7 A and B) and when dual mGlu₇/Munc13‐1 labelling was quantified in the neuropil of the stratum radiatum, 128 (38.55%) of the 332 axon terminals labelled for mGlu₇ were also labelled for Munc13‐1. In terms of Munc13‐2, this proportion increased to 201 (63.21%) double‐labelled terminals from a total of 318 axon terminals labelled for mGlu₇.

Figure 7. Nerve terminals containing the mGlu₇ receptor also express the Munc13‐1 or Munc13‐2 proteins.

A and B, double‐labelling pre‐embedding immunogold/HRP electron microscopy demonstrates mGlu₇ immunolabelling (gold/silver‐intensified immunogold reaction) in the presynaptic AZ (white arrows) of axon terminals (at) immunopositive (HRP reaction) for Munc13‐1 (A) and Munc13‐2 (B) that make synapses with dendritic spines (S) of pyramidal cells. Scale bar: 0.5 μm. Den, dendrite.

Lack of mGlu₇ receptor mediated potentiation in RIM1α KO mice

RIMs are multidomain proteins that play a central role in the organization of the AZ. RIM proteins tether Ca²⁺ channels to the presynaptic AZ (Kaeser et al. 2011) and they activate SV priming by reversing the auto‐inhibitory homodimerization of Munc13 (Deng et al. 2011). In addition, RIM1α also recruits and stabilizes Munc13 proteins at the AZ (Betz et al. 2001; Schoch et al. 2002), and it drives the distribution of SVs (Fernández‐Busnadiego et al. 2013). Given the physical and functional association between the Munc13 and RIM1α proteins, we studied whether mGlu₇ receptor modulation of synaptic transmission is altered in slices from RIM1α KO mice. Inhibition of EPSC amplitude was sustained in hippocampal slices lacking RIM1α (66.4 ± 14.3% of the baseline, n = 10 cells/10 slices/10 mice, P = 0.0024: Fig. 8 A and B), which reversed upon l‐AP4 washout, although there was no subsequent potentiation of the EPSC amplitude (97.4 ± 16.2% of the baseline, n = 10 cells/10 slices/10 mice, P = 0.8095: Fig. 8 A and B). l‐AP4 also failed to promote the accumulation of SVs close to the plasma membrane of RIM1α KO synapses (4.3 ± 1.2%, n = 152 synapses/3 mice in the presence of l‐AP4 and 5.0 ± 1.2%, n = 147 synapses/3 mice in the absence of l‐AP4, P = 0.2456: Fig. 8 G–I). Thus, mGlu₇ mediated potentiation of synaptic transmission is supported by the recruitment of SVs to the plasma membrane, which requires the Munc13‐2 and RIM1α proteins. By contrast, the number of SVs close to the AZ plasma membrane (Fig. 8 G and I) was similar in control slices from RIM1α KO and their WT littermates (Fig. 8 C and E), supporting basic synaptic transmission and its inhibition by l‐AP4 (Fig. 8 A and B).

Figure 8. The lack of mGlu₇ receptor mediated potentiation of synaptic transmission in RIM1α knockout mice is associated with the failure of l‐AP4 to increase the number of synaptic vesicles close to the active zone plasma membrane.

A, l‐AP4 fails to potentiate the EPSC amplitude in RIM1α KO mice. EPSC sample traces represent the mean of 30 consecutive EPSCs at 0.33 Hz from experiments in wild‐type (WT) and RIM1α KO mice taken 90 s prior to (1, basal), 90 s after (2, inhibition) and 10 min after l‐AP4 perfusion (3, potentiation). B, quantification of the changes in the EPSC amplitude induced by l‐AP4 (500 μm, 10 min) in WT (n = 5 cells/5 slices/5 mice) and RIM1α KO mice (n = 10 cells/10 slices/10 mice) during both inhibition (2, P = 0.0004 and P = 0.0061) and potentiation (3, P = 0.0024 and P = 0.8095). The data were compared to the basal state (1) using an unpaired Student's t test. C and D, l‐AP4 treatment brings SVs closer to the plasma membrane in WT mice. E, quantification of the l‐AP4 induced increase (P = 0.0013) in the SVs at least 10 nm from the AZ membrane. F, l‐AP4 does not change (P = 0.944, n = 112 synapses/3 mice) the total number of SVs per AZ in WT mice. The data from l‐AP4 treated slices were compared with the control slices (untreated) using an unpaired Student's t test. G and H, l‐AP4 treatment fails to bring SVs close to the plasma membrane in RIM1α KO mice. I, quantification of the l‐AP4 induced change in SV distribution in RIM1α KO mice. l‐AP4 fails (P = 0.2456) to bring SVs at least 10 nm from the AZ membrane. J, l‐AP4 does not change (P = 0.3616, n = 152 synapses/3 mice) the total number of SVs per AZ in RIM1α KO mice (n = 147 synapses/3 mice). The data from l‐AP4 treated slices were compared with control (untreated) slices using an unpaired Student's t test and the values represent the mean ± SD. Scale in A: 50 pA and 15 ms. Scale bar in C, D, G and H: 100 nm.

Electrically evoked release of glutamate also activates mGlu₇ receptors

We tested whether mGlu₇ receptors activated by electrically evoked release of glutamate also modulate excitatory synaptic transmission bidirectionally. As mGlu₇ receptors have a low affinity for both l‐AP4 and glutamate agonists (Okamoto et al. 1994), we used three‐train HFS (1 s at 100 Hz, every 20 s) to increase the synaptic glutamate. Accordingly, a MSOP‐sensitive component of synaptic potentiation (50.7 ± 24.2%, P = 0.0052) that lasted several minutes after HFS was unmasked (10 cells/10 slices/5 mice: Fig. 9 A and B). Nevertheless, as only a rapidly developing potentiation was observed, activation of mGlu₇ receptors in these conditions did not result in the bidirectional modulation of synaptic transmission seen with the exogenous agonist l‐AP4. While the PLC inhibitor U73122 prevented mGlu₇ induced potentiation, it did not unravel any inhibitory response as it completely eliminated the MSOP‐sensitive component of the EPSC (6.6 ± 24.4%, P = 0.7464, n = 8 cells/8 slices/4 mice: Fig. 9 A and B).

Figure 9. HFS‐induced activation of mGlu₇ receptors bidirectionally modulates synaptic transmission, although a potentiating response prevails.

A, changes in EPSC amplitude induced by three trains (1 s at 100 Hz, every 20 s) in control slices (untreated: n = 10 cells/10 slices/5 mice) or in slices exposed to: MSOP (200 μm, 30–60 min, n = 10 cells/10 slices/5 mice), U73122 (2 μm, 30 min, n = 8 cells/8 slices/4 mice), or MSOP plus U73122 (n = 8 cells/8 slices/4 mice). B, quantification of the MSOP sensitive component of the EPSC normalized to the basal values prior to HFS. C, changes in EPSC amplitude induced by two trains (1 s at 50 Hz, every 20 s) in control slices (untreated, n = 9 cells/9 slices/3 mice) or slices exposed to: MSOP (200 μm, 30–60 min, n = 9 cells/9 slices/ 3 mice), U73122 (2 μm, 30 min, n = 9 cells/9 slices/3 mice) or MSOP plus U73122 (n = 9 cells/9 slices/3 mice). D, quantification of the MSOP sensitive component of the EPSC normalized to the basal values prior to HFS. E, quantification of the MSOP sensitive component of the EPSC normalized to the basal values prior to HFS in d‐AP5 (50 μm) plus CGP 55845 (5 μm) treated slices (n = 10 cells/10 slices/3 mice) and in calphostin C (0.1 μm, 30 min) treated slices (n = 10 cells/10 slices/3 mice). Unpaired Student's t test compared to basal values. Scale bars in A and C: 50 pA and 15 ms. EPSC sample traces in A and C represent the mean of 10 consecutive EPSCs at 0.33 Hz 90 s before (basal) and 20 s after HFS. Thin and thick traces indicate in the absence and presence of MSOP, respectively.

The effects of the inhibitory response might be counteracted by the strong Ca²⁺ influx during HFS, therefore we used a weaker stimulation with only two trains (1 s at 50 Hz, every 20 s). With this protocol, a MSOP‐sensitive component of synaptic potentiation was evident (48.8 ± 23.5%, P = 0.0003) (n = 9 cells/9 slices/3 mice: Fig. 9 C and D), indicating that potentiation of synaptic transmission by mGlu₇ receptors also dominates under these conditions. However, U73122 now revealed an inhibitory effect of mGlu₇ (32.0 ± 24.9%, P = 0.0055, n = 9 cells/9 slices/3 mice: Fig. 9 D) that lasted less than the potentiating one.

We assessed the presynaptic origin of this MSOP‐sensitive component of synaptic transmission by measuring the corresponding changes in the PPR. The PPR decreased to −0.29 ± 0.21 (P = 0.0098) after HFS compared with basal (−0.03 ± 0.16, n = 9 cells/9 slices/3 mice) in the control slices, whereas the PPR increased to 0.25 ± 0.17 (P = 0.0011) after HFS compared to basal (−0.03 ± 0.11%, n = 9 cells/9 slices/3 mice) in U73122 treated slices (data not shown). We also confirmed that the modulation of synaptic transmission by HFS‐induced activation of mGlu₇ receptors was independent of NMDARs and GABA_B receptors, as the NMDAR antagonist d‐AP5 (50 μm) and the GABA_B antagonist CGP 55845 (5 μm), added together, did not alter the MSOP component of synaptic transmission (52.5 ± 17.2%, P = 0.0001, n = 10 cells/10 slices/3 mice, Fig. 9 E). We also tested that the mGlu₇ receptor dependent potentiation was abolished in the presence of calphostin C, revealing an inhibitory response (43.2 ± 20.5%, P = 0.0001, n = 10 cells/10 slices/3 mice: Fig. 9 E).

The bidirectional modulation of synaptic transmission by the HFS‐induced activation of mGlu₇ receptors was tested for its sensitivity to the lack of Munc13‐2 and RIM1α proteins. In Munc13‐2 KO the potentiation was abolished, although the inhibitory response was preserved (27.0 ± 24.7%, P = 0.0075, n = 11 cells/11 slices/4 mice). In contrast, WT littermates showed potentiation (54.5 ± 24.9%, P = 0.0001, n = 11 cells/11 slices/3 mice, Fig. 10 A and B). Similarly, in RIM1α KO mice the potentiation was abolished and unchanged inhibition was observed (30.3 ± 26.0%, P = 0.0043, n = 11 cells/11 slices/4 mice) while WT littermates showed potentiation (50.3 ± 29.1%, P = 0.0012, Fig. 10 C and D). Thus, HFS‐induced activation of mGlu₇ receptors provokes the bidirectional modulation of synaptic transmission. However, as both responses occur simultaneously, the potentiating response alone prevails while the inhibitory one is only evident when potentiation is blocked with calphostin C, or when there is no Munc13‐2 or RIM1α protein.

Figure 10. The mGlu₇ dependent HFS‐induced potentiation is abolished by the lack of Munc13‐2 or RIM1α proteins.

A, changes in the EPSC amplitude induced by two trains (1 s at 50 Hz, every 20 s) in control (untreated) and MSOP (200 μm, 30–60 min) treated slices from Munc13 KO (n = 11 cells/11 slices/4 mice) and WT littermates (n = 11 cells/11 mice/3 mice). B, quantification of the MSOP sensitive component of the EPSC normalized to the basal values prior to HFS. C, changes in EPSC amplitude induced by two trains (1 s at 50 Hz, every 20 s) in control (untreated) and MSOP (200 μm, 30–60 min) treated slices from RIM1α KO (n = 11 cells/11 slices/4 mice) and WT littermates (n = 9 cells/9 mice/3 mice). D, quantification of the MSOP sensitive component of the EPSC normalized to the basal values prior to HFS. Unpaired Student's t test compared to basal values. Scale bars in A and C: 50 pA and 15 ms. EPSC sample traces in A and C represent the mean of 10 consecutive EPSCs at 0.33 Hz 90 s before (basal) and 20 s after HFS (thin and thick traces in the absence and presence of MSOP, respectively). E, scheme of the proposed signalling for the bidirectional modulation of synaptic transmission by mGlu₇ receptors. The mGlu₇ receptor stimulates a PTx‐sensitive G protein that inhibits voltage dependent Ca²⁺ channels and decreases neurotransmitter release. The mGlu₇ receptors also activate a PTx‐resistant G protein to enhance PLC activity, thereby increasing PIP₂ hydrolysis and the generation of DAG. This DAG binds to and activates/translocates the non‐kinase DAG‐binding protein Munc13‐2 to promote glutamate release with the help of RIM1α. DAG, diacylglycerol; HFS, high frequency stimulation; PIP₂, phosphatidylinositol bisphosphate; PLC, phospholipase C; PTx, pertussis toxin; VDCC, voltage dependent Ca²⁺ channel. The blockade/inhibition (dashed line) of Gi/o proteins by PTx, mGlu₇ receptors by MSOP, PLC by U73122 and Munc13 by calphostin C is also indicated, as well as the activation (arrows) of mGlu₇ receptors with the agonist l‐AP4 or by HFS induced glutamate release.

Discussion

Activation of mGlu₇ receptors with the agonist l‐AP4 or following glutamate release induced by HFS provokes a bidirectional modulation of excitatory synaptic transmission at SC–CA1 synapses. This inhibition is sensitive to the blockage of Gi/o proteins with PTx, while potentiation is prevented by the PLC inhibitor U73122 and by the inhibitor of DAG binding, calphostin C. This DAG‐dependent modulation of the release machinery requires the Munc13‐2 or RIM1α proteins. Both inhibition and potentiation occur simultaneously when mGlu₇ receptors are activated by the glutamate released in response to HFS, although potentiation prevails and the inhibitory response can only be revealed after the blockade of potentiation.

We found that mGlu₇ receptors bidirectionally modulate synaptic transmission through PTx‐sensitive and PTx‐resistant signalling pathways, which inhibit and enhance neurotransmitter release, respectively. The inhibition of synaptic transmission mediated by mGlu₇ (Baskys & Malenka, 1991; Gereau & Conn, 1995; Ayala et al. 2008; Klar et al. 2015) is the result of Ca²⁺ channel inhibition (Millán et al. 2002, 2003; Perroy et al. 2000, 2002; Pelkey et al. 2006), yet it is independent of changes in cAMP unless adenylyl cyclase is activated (Millán et al. 2002; Martín et al. 2007; Ferrero et al. 2016). Direct evidence that mGlu₇ receptors potentiate neurotransmitter release was obtained from synaptosomes (Martín et al. 2010), and our electrophysiological data are consistent with that biochemical study. First, the potentiating response is resistant to PTx but dependent on PLC, and second, the potentiating response is resistant to PKC inhibitors but blocked by the DAG‐binding antagonist calphostin C. Finally, the potentiation of neurotransmitter release does not involve a switch in receptor signalling from inhibition to potentiation but, rather, the coupling of the receptor to an additional PLC‐dependent signalling cascade. This is consistent with the capacity of the mGlu₇ receptor to simultaneously activate several signalling pathways (Martín et al. 2010; Ferrero et al. 2016). Thus, mGlu₇ receptor coupling to PLC‐dependent pathways in synaptosomes results in the net potentiation of release, despite the ongoing inhibition of Ca²⁺ channels (Martín et al. 2010).

The mGlu₇ receptor mediated potentiation of synaptic transmission requires the AZ protein Munc13‐2, consistent with the central role of Munc13 proteins in neurotransmitter release. The hippocampal CA regions express Munc13‐1 and Munc13‐2 (Augustin et al. 1999a), which are essential for evoked and spontaneous release, and to maintain the RRP (Varoqueaux et al. 2002; Imig et al. 2014). We could not test the role of Munc13‐1 in mGlu₇ modulation of synaptic transmission because Munc13‐1 KO mice are not viable (Augustin et al. 1999a). By contrast, mice lacking Munc13‐2 (Varoqueaux et al. 2002) maintain basal synaptic transmission and the number of SVs close to the AZ plasma membrane, and mGlu₇ receptor mediated inhibition persists in these mice, although they failed to potentiate synaptic transmission because no further SVs are recruited to the AZ plasma membrane. This is consistent with the fact that these receptors are distributed in both Munc13‐1‐ and Munc13‐2‐containing synapses. Munc13 proteins have a DAG binding site (Betz et al. 1998), through which PLC‐coupled GPCRs activate/translocate these proteins (Bauer et al. 2007; Martín et al. 2010; Ferrero et al. 2013) and initiate the PLC dependent potentiation of synaptic transmission. Munc13‐2 is more abundant in cytosolic fractions than Munc13‐1 (Augustin et al. 1999a) and thus, Munc13‐2 could provide extra priming of SVs in response to DAG generation by mGlu₇ receptors. In addition, Munc13‐2 proteins have been associated with the potentiation of synaptic transmission at several synapses (Gioia et al. 2016; Kawabe et al. 2017) in a PLC dependent manner (Rosenmund et al. 2002). The mGlu₇ receptor mediated potentiation of both glutamate release (Martín et al. 2010) and synaptic transmission is also PLC dependent, suggesting that mGlu₇ receptors are involved in this form of synaptic plasticity. In fact, short‐term potentiation is attenuated in mGlu₇ KO mice (Bushell et al. 2002).

The mGlu₇ receptor mediated potentiation of synaptic transmission requires RIM1α proteins, consistent with its essential role in transmitter release and synaptic plasticity (Castillo et al. 2002; Schoch et al. 2002; Kaeser et al. 2011). RIM1α supports the function of Munc13 proteins by disrupting inactive Munc13 homodimers and stabilizing the protein (Schoch et al. 2002; Deng et al. 2011). In addition, RIM1α organizes AZ proteins and recruits Ca²⁺ channels to the AZ (Kaeser et al. 2011). Finally, the dependence of mGlu₇ receptor mediated potentiation on RIM1α is also consistent with the co‐immunoprecipitation of these two proteins in the hippocampus (Pelkey et al. 2008).

The mGlu₇ receptor mediated potentiation of synaptic transmission is faster when glutamate is released by HFS than by the agonist l‐AP4. The activation of mGlu₇ receptors also depends on Ca²⁺/calmodulin binding to the intracellular C domain (O'Connor et al. 1999; Caldwell et al. 2008; Scheschonka et al. 2008). Thus, the Ca²⁺ requirements for optimal receptor activation can be better fulfilled during HFS than with agonist activation (0.33 Hz). Strikingly, mGlu₇ receptors respond to much lower l‐AP4 concentrations after light‐induced channel rhodopsin‐2 mediated Ca²⁺ influx (Caldwell et al. 2008). Therefore, AZ mGlu₇ receptors (Shigemoto et al. 1996, 1997) exposed to high concentrations of extracellular glutamate and intracellular Ca²⁺ during HFS may be more active than when exposed to l‐AP4. High Ca²⁺ levels may also favour the activation of other Ca²⁺ dependent proteins in the mGlu₇ receptor signalling cascade that are involved in the potentiation of neurotransmitter release (Martín et al. 2011), such as Munc13 (Junge et al. 2004).

The HFS‐induced release of glutamate activates mGlu₇ receptors that signal simultaneously through pathways leading to the potentiation and inhibition of neurotransmitter release. As the potentiating effect is stronger and lasts longer than the inhibitory one, the potentiation of synaptic transmission prevails, consistent with the role of these receptors in short term potentiation (Bushell et al. 2002). The duration of the inhibitory response is probably related to the time during which synaptic glutamate persists above the threshold for mGlu₇ receptor activation, which exhibits a low affinity for glutamate. By contrast, mGlu₇ receptor potentiation lasts longer because a more complex signalling cascade is associated with this response. Thus, mGlu₇ receptors activate a PTx resistant G protein that enhances PLC activity and generates DAG, which in turn translocates Munc13‐2 proteins to plasma membrane and recruits SVs to the AZ, with the help of RIM1α (see Fig. 10 E and Martín et al. 2010). Accordingly, the duration of the potentiating response could be dependent on the DAG levels and/or on the time that Munc13 priming lasts. The rapid potentiation of synaptic transmission by mGlu₇ may have important implications for the short‐term plasticity of excitatory synapses. Moreover, mGlu₇ receptors are also present at GABAergic terminals, although it is not known if these receptors exert bidirectional control of inhibitory synaptic transmission. Additional studies will be required to address these issues.

Additional information

Competing interests

None declared.

Author contributions

R.M., M.T. and J.S.‐P. conceived and designed the experiments with input from all the co‐authors; R.M. performed and analysed the electrophysiological recordings; R.M., J.J.F. and A.C.‐A. performed and analysed the electron microscopy experiments at Universidad Complutense; C.A. and R.L. performed and analysed the immunohistochemistry for electron microscopy at Universidad Castilla‐La Mancha. J.S.‐P. wrote the paper with input from all the co‐authors. All authors have approved the final version of the manuscript and agree to be accountable for all aspects of the work. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

This work was supported by grants from the “Spanish Ministry for Economy and Competitivity” MINECO (grant numbers BFU 2013‐43163R to J.S.‐P. and BFU 2012‐32105 to M.T., and BFU 2015‐63769‐R to R.L.) and the “Junta de Comunidades de Castilla‐La Mancha” (PPII‐2014‐005‐P to R.L.).