Two young adult male Wistar rats (P30, 42), 3 adult (P50) male C57BL/6J mice, 47 adult (P50-60) Tg(Chrna2-Cre)OE25Gsat/Mmucd (RRID:MMRRC_036502-UCD, on C57BL/6J background (Leao et al., 2012) crossed with reporter line Ai9 [Gt(ROSA)26Sor_CAG/LSL_tdTomato], 15 adult (P50–70) male C57BL/6N-Elfn1^{tm1.1(KOMP)Vlcg}/MbpMmucd (RRID:MMRRC_047527-UCD, on C57BL/6N background (Tomioka et al., 2014) and 5 heterozygous littermate control mice, and 35 adult P50–70 C57BL/6N-Unc13b^{TM 1a(KOMP)Wtsi}/MbpMmucd (RRID:MMRRC_050292-UCD, on C57BL/6N background) were used. Mice of both sexes were used. The animals were housed in the vivarium of the Institute of Experimental Medicine in a normal 12 h/12 h light/dark cycle and had access to water and food ad libitum. All the experiments were carried out in accordance with the Hungarian Act of Animal Care and Experimentation 40/2013 (II.14) and with the ethical guidelines of the Animal Committee of the Institute of Experimental Medicine, Budapest.

Mice were anesthetized with a mixture of ketamine, xylasine, pypolphene (0.625, 6.25, 1.25 mg/ml respectively, 10 μl/g body weight). The pAAV-Ef1a-mCherry-IRES-Cre was a gift from Karl Deisseroth (1.8 × 10¹³ Addgene viral prep # 55632-AAV8¹ ; RRID:Addgene_55632) (Fenno et al., 2014) or pENN.AAV.CamKII 0.4.Cre.SV40 was a gift from James M. Wilson (Addgene viral prep # 105558-AAv9² ; RRID:Addgene_105558) (1:10 dilution 2.8 × 10¹³, Penn Vector Core) were injected into the dorsal hippocampus (200 nl, coordinates from the Bregma in mm: antero posterior/dorso ventral/lateral: 2.1/1.1/1.3 and/or 2.2/1.5/1.2). After 2 weeks, the mice were either perfused or in vitro acute slices were prepared from the dorsal hippocampus as below.

The following animals were used: 44 adult (P50-60) Tg(Chrna2-Cre)OE25Gsat/Mmucd (RRID:MMRRC_036502-UCD, on C57Bl/6J background) crossed with reporter line Ai9 [Gt(ROSA)26Sor_CAG/LSL_tdTomato]; 10 adult P50-70 C57BL/6N-Elfn1^{tm1.1(KOMP)Vlcg}/MbpMmucd. Nineteen adult P50-70 C57BL/6N-Unc13b^{TM 1a(KOMP)Wtsi}/MbpMmucd mice were injected with pAAV-Ef1a-mCherry-IRES-Cre more than 2 weeks prior to the slice preparation. Mice were stably anesthetized with a ketamine, xylasine, and pypolphene cocktail (0.625, 6.25, 1.25 mg/ml, respectively, 10 μl/g body weight), and then decapitated; the brain was quickly removed and placed into an ice-cold cutting solution containing the following (in mM): sucrose, 205.2; KCl, 2.5; NaHCO₃, 26; CaCl₂, 0.5; MgCl₂, 5; NaH₂PO₄, 1.25; and glucose, 10, saturated with 95% O₂ and 5% CO₂. Coronal slices (300 μm thick) were then cut from the dorsal hippocampus using a Leica Vibratome (VT1200S) and were incubated in a submerged-type holding chamber in an artificial cerebral spinal fluid (ACSF) containing the following (in mM): NaCl, 126; KCl, 2.5; NaHCO₃, 26; CaCl₂, 2; MgCl₂, 2; NaH₂PO₄, 1.25; and glucose, 10, saturated with 95% O₂ and 5% CO₂, pH 7.2–7.4, at 36°C for 30 min, then kept at 22–24°C until use. Recordings were performed in the same ACSF supplemented with 2 μM AM251 to block presynaptic CB1 receptors at 32°C up to 6 h after slicing.

Cells were visualized using infrared differential interference contrast (DIC) imaging on a Nikon Eclipse FN1 microscope with a 40X water immersion objective (NA = 0.8). CA1 PCs were identified from their position and morphology, and the virally expressed mCherry was identified using fluorescent illumination. Whole-cell current-clamp recordings were performed from CA1 PCs using MultiClamp 700B amplifiers (Molecular Devices). Recorded traces were filtered at 3–4 kHz and digitized online at 50 kHz. Patch pipettes (resistance 3–6 MΩ) were pulled from thick-walled borosilicate glass capillaries with an inner filament. Intracellular solution contained the following (in mM): K-gluconate, 130; KCl, 5; MgCl₂, 2; EGTA, 0.05; creatine phosphate, 10; HEPES, 10; ATP, 2; GTP, 1; and biocytin, 7; glutamate, 10 (for presynaptic PCs only) pH = 7.3; 290–300 mOsm. Pyramidal cells were held at –65 mV (with a maximum of ± 100 pA DC current) and trains of 3 APs at 40 Hz were evoked with 1.2–1.5 ms-long depolarizing current pulses (1.5–2 nA). Peak amplitude and full width at half-maximal amplitude of the APs were monitored and the cells were rejected if any of these parameters changed greater than 10%. Postsynaptic O-LM cells were identified in the stratum oriens of the CA1 region by tdTomato fluorescence in the Tg(Chrna2-Cre)OE25Gsat/Mmucd animals crossed with reporter line Ai9 or by their morphology and firing pattern obtained with de- or hyperpolarizing square current injections (600 ms, from –250 to 300 pA with 50 pA steps). Paired whole-cell recordings were performed with a dual-channel amplifier (MultiClamp 700B; Axon Instruments, CA, United States). Data were filtered at 3–4 kHz (Bessel filter), digitized online at 50 kHz, recorded and analyzed using pClamp 10.3 (Molecular Devices, CA, United States). Postsynaptic INs were held at –65 mV (with a maximum of ± 200 pA DC current) in the voltage-clamp mode with the access resistance below 20 MOhm.

After recordings, the slices were fixed in a solution containing 4% formaldehyde (FA, Molar Chemicals, Budapest, Hungary), 0.2% picric acid in 0.1 M phosphate buffer (PB), pH = 7.4, at 4°C for 12 h. Immunolabeling was carried out without re-sectioning. Slices were washed in 0.1 M PB and blocked in normal goat serum (NGS, 10%; Vector Laboratories, CA, United States) for 1 h made up in Tris-buffered saline (TBS; pH 7.4), incubated in the solutions of the primary antibodies: guinea pig anti-mGluR1α (1:1,000, Frontier Institute Co., Ltd.; Cat# mGluR1a-GP-Af660, RRID:AB_2531897), or a cocktail of this Ab and a mouse anti-Cre Ab (IgG1, 1:1,000, Millipore, Cat# MAB3120, RRID:AB_2085748); diluted in TBS containing 2% NGS and 0.2% TritonX-100. Biocytin was visualized with Cy3-conjugated streptavidin (1:1,000; Jackson Immunoresearch Laboratories, PA, United States, RRID:AB_2337244). After several washes, the following secondary antibodies were applied: Alexa488-conjugated goat anti-mouse IgG1 (Jackson Immunoresearch Laboratories, PA, United States, Code: 115-547-185, RRID:AB_2632534), and Cy5-conjugated donkey anti-guinea pig IgGs (Jackson Immunoresearch, Code 706-175-148, RRID: AB_2340462). Sections were mounted in Vectashield. Image stacks were acquired with an Olympus FV1000 confocal microscope with 20x and 60x (oil immersion) objectives. A PC was considered virally infected if its nucleus had a detectable Cre signal. A cell was considered an O-LM cell, if the axon arborized in the stratum lacunosum-moleculare.

Two young-adult Wistar rats (P30, 42), 3 adult (P50) male C57BL/6J mice, 3 adult (P50–60) Tg(Chrna2-Cre)OE25Gsat/Mmucd, (RRID:MMRRC_036502-UCD, on C57BL/6J background) crossed with reporter line Ai9 [Gt(ROSA)26Sor_CAG/LSL_tdTomato], 5 adult (P50–70) male C57BL/6N-Elfn1^{tm1.1(KOMP)Vlcg}/MbpMmucd (RRID:MMRRC_047527-UCD, on C57BL/6N background) and 5 heterozygous littermate control mice, and 16 adult P50-70 C57BL/6N-Unc13b^{TM 1a(KOMP)Wtsi}/MbpMmucd (RRID:MMRRC_050292-UCD, on C57BL/6N background) were deeply anesthetized and transcardially perfused with a fixative containing 1% FA (Molar Chemicals, Budapest, Hungary) and 0.2% picric acid in 0.1 M PB or in 0.1 M sodium acetate buffer for 15–20 min. The brains were then quickly removed from the skull and placed in 0.1 M PB. Next, 60–100 μm thick sections were cut from the dorsal hippocampus. After blocking in NGS or normal donkey serum (NDS, 10%) for 1 h made up in TBS, the sections were incubated in the following primary antibodies: guinea pig anti-mGluR1α (1:1,000; Frontier Institute Co., Ltd, Nittobo Medical Co. Ltd, Tokyo, Japan; Cat# mGluR1a-GP-Af660, RRID:AB_2531897), goat anti-mGluR1α (1:1,000, Frontier Institute Co., Ltd; Cat# mGluR1a-Go-Af1220, RRID:AB_2571800), rabbit anti-Munc13-1 (1:500, Synaptic Systems, Göttingen, Germany; Cat# 126113, RRID:AB_887734); rabbit anti-Munc13-2 (1:250; Synaptic Systems, Göttingen, Germany; Cat# 126203, RRID:AB_2619807, recognizing the brain-specific isoform, bMunc13-2), rabbit anti-mGluR7 (1:1,000; Millipore, MA, United States; Cat# 07-239, RRID: AB_310459), rabbit anti-Elfn2 (1:500, Sigma-Aldrich, MO, United States, Cat# HPA000781, RRID: AB_1079280) that recognizes both Elfn1 and Elfn 2 (Sylwestrak and Ghosh, 2012), rabbit anti-Elfn1 (1:500; Synaptic Systems, Göttingen, Germany; Cat#448-003, RRID:AB_2884915), guinea pig anti-mGluR7b (1:1,000 gift from Prof. Shigemoto, Shigemoto et al., 1997) diluted in TBS containing 2% NGS or NDS and 0.2% TritonX-100. After several washes, the following secondary antibodies were applied: Cy3-conjugated donkey anti-guinea pig (Cat#: 706-165-148, RRID: AB_2340460) or donkey anti-goat IgG (Cat#: 705-165-147, RRID: AB_2307351) and Cy5-conjugated donkey anti-rabbit IgG (Cat#711-175-152, RRID: AB_2340607; all from Jackson Immunoresearch Laboratories, PA, United States). Sections were either mounted in Vectashield or processed for multiplexed postembedding immunolabeling (see the section below). Image stacks were acquired with an Olympus FV1000 confocal microscope (Olympus Europa SE & Co., Hamburg, Germany) with 20x and 60x (oil immersion) objectives.

For details of this method (see Holderith et al., 2020). Briefly, sections following a preembedding immunoreaction (see the section above) for mGluR1α (1:1,000 visualized with Cy3-coupled donkey anti-goat Ab) and Munc13-2 (1:250; visualized with Cy5-coupled donkey anti-rabbit Ab) were washed several times in 0.1 M PB, dehydrated (without treatment with OsO₄), and embedded into Durcupan. Preembedding immunolabeling of Munc13-2 was essential, because after epoxy embedding, our antibody did not recognize its epitope. 500 nm thick serial sections of the stratum oriens containing double immunolabeled IN dendrites were cut from the top 5 μm of the sections (to avoid heterogeneity from potential unequal penetration of the antibody into the tissue) and placed onto Superfrost Ultra plus slides. First, selected regions of interest (ROIs) were analyzed and imaged with Olympus FV1000 confocal microscope at high magnifications (60X objective; NA = 1.35). The resin was etched with Na-ethanolate (saturated solution of NaOH in absolute ethanol) for 5 min, then rinsed with 96% ethanol, and finally rinsed with distilled water. Retrieval of the proteins were carried out in 0.02 M Tris Base (pH = 9) containing 0.5% sodium dodecyl sulfate (SDS) at 80^°C for 60 min. After several washes in TBS containing 0.05–0.1% Triton X-100 (TBST, pH = 7.6), the sections were blocked in TBST containing 6% Blotto A (Santa Cruz Biotechnology), 10% NGS, and 1% BSA (Sigma) for 30 min, then incubated in the guinea pig primary Ab against Munc13-1 (1:500, Synaptic Systems, Göttingen, Germany; Cat# 126104, RRID: AB_2619806) diluted in a blocking solution at room temperature, overnight. After several washes in TBST, the secondary Ab (Alexa 488 donkey anti-guinea pig, 1:200, Jackson ImmunoResearch Laboratories, PA, United States, Cat# 706-545-148, RRID: AB_2340472) was applied in TBST containing 10% of blocking solution for 2 h. After several washes in TBST, the slides were rinsed with distilled water, and the sections were mounted in SlowFade Diamond (Invitrogen, MA, United States). Images of ROIs were taken with an Olympus FV1000 confocal microscope and a 60x (oil immersion) objective or with an Abberior ExpertLine confocal microscope and a 100x (oil immersion) objective. Immunoglobulins were removed with a 5-min incubation in TBS containing 1% SDS (pH = 7.7) at 80^°C. After 5 min of washing in TBST, new rounds of immunolabeling were performed using a guinea pig anti-PSD95 Ab (1:500, Synaptic Systems, Göttingen, Germany; Cat# 124014, RRID: AB_2619800), a guinea pig anti-panAMPA receptor Ab (1:100 Frontiers Cat#Af580; RRID: AB_257161), a chicken anti-Bassoon Ab (1:200 Synaptic Systems, Göttingen, Germany; Cat#141-016; RRID: AB_2661779), a guinea pig anti-Cav2.1 Ab (1:100 Synaptic Systems, Göttingen, Germany; Cat#152 205; RRID: AB_2619842), and finally with a guinea pig anti-Rim1/2 Ab (1:100 Synaptic Systems, Göttingen, Germany; Cat#140-205; RRID:AB_2631216) with the appropriate Alexa488-coupled secondary Abs (Alexa488-coupled donkey anti-guinea pig, 1:200; Jackson ImmunoResearch Laboratories, PA, United States; as above or Alexa488-coupled goat anti-chicken, 1:200; Jackson ImmunoResearch Laboratories, PA, United States, Cat# 103-545-155, RRID: AB_2337390).

Confocal images were imported in ImageJ where circular ROIs (for postembedding immunolabeling) were placed over synaptic fluorescent clusters and over the unlabeled neuropil to determine the specific and non-specific labeling, respectively. The integral of the fluorescence was measured, and the background was subtracted for each ROI. For postembedding multiplexed labeling, images of the serial 500 nm thick sections were aligned using a custom-made module of ImageJ (“HyperStackStitcher;” available on the web site).³ To pool data from different experiments obtained with different confocal microscopes and laser settings, PSD95- normalized intensities of each immunosignal were further normalized to the population mean of the random synapses.

To quantify Munc13-2, Elfn, mGluR7, and mGluR1α immunolabeling intensity in Elfn KO and virally injected animals, freehand contour ROIs were used to outline mGluR1α+ dendrites in both the KO and the control or the injected and the contralateral hemispheres with same microscope settings and in the same depth of the tissue. The integral of the fluorescence was measured, and the background was subtracted for each ROI. About 40–150 ROIs were measured per animal.

Data are presented as mean ± SD and median throughout this study. All data were tested for normality using Shapiro–Wilk test. Non-normally distributed datasets were compared with Mann–Whitney U-test. Correlations were determined with Spearman’s rank correlation and the regression coefficient (r_s); related p-values were calculated from two-tailed Student’s t-distribution. For multiple comparison, Kruskal–Wallis ANOVA was used with post hoc Dunn’s test. Statistically significant difference was assumed at p < 0.05.

All the chemicals are obtained from Sigma, unless otherwise stated.

Immunostaining of bMunc13-2 (brain specific isoform of the Munc13-2, referred as Munc13-2 in the study) in the dorsal hippocampus of adult mice [Figure 1, n = 3 C57BL6/J, n = 3 Tg(Chrna2-Cre)OE25Gsat/Mmucd, Leao et al., 2012] crossed with the reporter line Ai9 [Gt(ROSA)26Sor_CAG/LSL_tdTomato] and rats (n = 2, Supplementary Figure 1) revealed punctate labeling of neuronal processes in the stratum oriens and in the alveus of the CA1 area. Double immunofluorescent labeling of Munc13-2 and mGluR1α demonstrated that most of the Munc13-2 immunopositive puncta decorate mGluR1α+ dendrites and the majority of the mGluR1α+ dendrites are decorated by Munc13-2 puncta (Figures 1A,B). Since mGluR7 is present in presynaptic glutamatergic AZs that innervate mGluR1α+ INs (Shigemoto et al., 1996) and has a rather similar labeling pattern to that of Munc13-2, we performed colocalization of these two proteins. The majority of the mGluR7b puncta along the small diameter dendrites were Munc13-2 immunopositive and vice versa, and most Munc13-2 puncta were labeled for mGluR7b in rat (Supplementary Figures 1C,D). This suggested that Munc13-2 is present in excitatory synapses.

To further test the molecular composition of these Munc13-2 immunopositive synapses in mice, we applied postembedding multiplexed immunolabeling of a number of synaptic proteins on ultrathin sections obtained from epoxy resin-embedded tissue (Holderith et al., 2020) that had been immunolabeled for Munc13-2 and mGluR1α. We discovered that dehydration and resin embedding without OsO₄ treatment retains the preembedding immunosignal when the reactions are visualized with Cy3- or Cy5-coupled secondary antibodies (Figure 1C). In 500 nm thick epoxy resin-embedded sections, Munc13-2 immunopositive puncta were clearly visible as they surrounded mGluR1α+ small diameter dendrites but spared the perisomatic and proximal dendritic membranes (Figure 1C). Following the removal of the resin with Na-ethanolate and antigen retrieval with an SDS solution, we immunolabeled the sections, first for Munc13-1, imaged the ROI, eluted the immunoglobulins, and relabeled the sections for PSD95, then for AMPA receptors, Bassoon, Cav2.1 voltage-gated Ca²⁺ channel (VGCC) subunit, and finally for Rim1/2 (Figure 1D). Qualitative assessment of the images revealed that the Munc13-2 positive puncta were immunopositive for all of these synaptic proteins, but to a different degree. We then quantified the fluorescent intensities for each protein in circular ROIs placed over the Munc13-2 positive puncta. We found that all Munc13-2 positive puncta contained PSD95 immunosignal (Figure 1E, n = 40 and 80 puncta in two mice) indicating that they are excitatory glutamatergic synapses. As the amount of PSD95 correlates almost perfectly with the size of the synapse (Cane et al., 2014; Karlocai et al., 2021), we normalized the immunosignal for each synaptic protein to that of PSD95 of the same synapse, resulting in density values that should be independent of the synapse size. To compare the data from different experiments, Munc13-2 density values were further normalized to the population mean of randomly selected synapses. Following this normalization, Munc13-2 density values displayed large variability (coefficient of variation: CV = 0.87 and 0.86 for mouse #1 and #2, respectively) among individual synapses and they did not correlate with normalized Munc13-1 density values (Figure 1F), suggesting that their amounts in the AZ are independently regulated. Interestingly, the PSD95 normalized densities of Munc13-1, AMPA receptors, Bassoon, Cav2.1, and Rim1/2 were significantly higher in synapses on mGluR1α+ dendrites compared to the randomly selected surrounding synapses (Munc13-1: 1.27 ± 0.58; AMPAR: 1.46 ± 0.50; Bassoon: 1.47 ± 0.85; Cav2.1: 1.38 ± 0.57; Rim1/2: 1.22 ± 0.6, n = 194 mGluR1α targeting and n = 160 random synapses from 2 mice; Figure 1G). To assess the selectivity of the Munc13-2 expression in mGluR1α+ dendrites targeting synapses, we measured their Munc13-2 content and compared them with that of randomly selected synapses in the surrounding neuropil in 2 mice (n = 101 and n = 60 mGluR1α+ dendrite targeting synapses, and n = 1,000 and n = 500 random synapses). Our quantification revealed that only 4 ± 6% and 4 ± 10% of the immunoreactivity found in mGluR1α+ dendrite targeting synapses are present in the surrounding randomly sampled synapses (Figure 1H).

Munc13-2 did not colocalize with vesicular inhibitory amino acid transporter (VIAAT; Figures 1I,J, n = 152 Munc13-2, and n = 222 VIAAT positive puncta in 2 mice, and Supplementary Figure 1F in 1 rat) or with vesicular glutamate transporter-2 (vGluT2; Figures 1K,L, n = 43 Munc13-2 and n = 33 vGluT2 positive puncta in 1 mouse, Supplementary Figure 1G in 1 rat). This indicates that most of the Munc13-2 labeled puncta are present on the axon terminals of local (CA1 and/or CA3) PCs.

Elfn1 has been shown to be selectively expressed in Som/mGluR1α+ hippocampal INs (Sylwestrak and Ghosh, 2012), where it recruits mGluR7 to the presynaptic AZ (Tomioka et al., 2014; Stachniak et al., 2019). The selective knock-down of Elfn1 from these INs results in a decreased short-term facilitation. Next, we tested the potential role of Elfn1 in the above-described selective recruitment of Munc13-2. In five Elfn1 knock-out (KO) mice, immunolabeling of Elfn1/2 were clearly absent in the stratum oriens/alveus (Figures 2A,E). Quantitative measurement of Elfn1 immunolabeling showed a 97.7 ± 0.53% decrease in KO mice (n = 5) compared to the control littermates (n = 4) demonstrating that our antibody against Elfn1/2 provides specific labeling here for Elfn1 and also that the protein is missing in the KO mice. In line with previous results, we also found that immunolabeling of mGluR7 is dramatically reduced compared to the littermate heterozygous controls (Figures 2C,G) without any apparent change in the postsynaptic mGluR1α expression (Figures 2B,F, 107 ± 16% of control littermates, n = 5 KO mice, n = 4 control mice). Finally, we found that immunolabeling of Munc13-2 is apparently absent in the KO mice (3.7 ± 1.7% of control littermates; n = 5 KO mice and n = 4 control mice) compared to the littermate controls where punctate Munc13-2 immunolabeling is clearly visible along the mGluR1α+ dendrites (Figures 2D,H). The lack of both mGluR7 and Munc13-2 in Elfn1 KO mice raises the question whether the increased P_v at the PC-Som+ IN synapses found following Elfn1 knock-down is attributable to the lack of either mGluR7, or that of Munc13-2, or both.

To investigate the contribution of Munc13-2 to the properties of CA1 PC to mGluR1α+ IN synapses, Munc13-2 was conditionally knocked out from CA1 PCs in sixteen C57BL/6N-Unc13b^{TM 1a(KOMP)Wtsi}/MbpMmucd mice with Cre-recombinase expressing AAVs injected unilaterally into the dorsal hippocampus (Figure 3). After 14 days, the mice were transcardially perfused and Cre expression was visualized with an anti-Cre antibody (Figure 3 and Supplementary Figure 2). In the non-injected hemisphere, no detectable anti-Cre immunoreactivity was observed in the nuclei of CA1 PCs and the immunolabeling patterns of Munc13-2, mGluR1α, Elfn1, mGluR7, and Munc13-1 (Figures 3A–E) were indistinguishable from those seen in the control mice and rats (Figure 1 and Supplementary Figure 1). In the central part of the injected area, the nuclei of apparently every PC were intensely labeled for Cre-recombinase and the mGluR1α+ dendrite associated specific immunosignal for Munc13-2 decreased by 92 ± 10% in the stratum oriens/alveus (n = 3 mice, Supplementary Figure 2G and Figure 3F). At the edges of the injection zone, the frequency of Cre-immunopositive nuclei decreased and Munc13-2 labeled structures emerged (Supplementary Figures 2A–C). In contrast, despite the expression of Cre and the lack of Munc13-2, the labeling patterns and the intensity of labeling for mGluR1α, Elfn1, mGluR7, and Munc13-1 were unchanged (Figures 3G–J and Supplementary Figure 2G, 101 ± 18%, 104 ± 5%, 104 ± 4%, 99 ± 1% of controls, respectively. n = 3 mice for mGluR1α, Elfn1, mGluR7, n = 2 mice for Munc13-1). In our injections, neither CA3 nor subicular PCs were transfected with Cre-expressing AAVs. Our results demonstrate that the majority of Munc13-2 immunosignal in the stratum oriens of the dorsal CA1 area originates from the axons of CA1 PCs and 2 weeks is sufficient for the significant removal of the protein from these synapses.

Postsynaptic responses between CA1 PCs and mGluR1α+ INs have small amplitudes and display marked short-term facilitation. We recorded from a large population of connected CA1 PC to mGluR1α+ IN pairs in a transgenic mouse line that expresses the red fluorescent protein, tdTomato in O-LM INs in the stratum oriens of the CA1 region (Figures 4A,B). This allowed us to efficiently select mGluR1α+ /O-LM INs (50 out of 52 post hoc identified INs had O-LM morphology, and only 2 were bistratified cells, Figure 4A). We recorded unitary excitatory postsynaptic currents (uEPSCs) evoked by a short train of presynaptic action potentials (APs) at 40 Hz (Figure 4C, black). The first uEPSC had a small amplitude (9.6 ± 9.4 pA, median = 7.2 pA, n = 80 pairs in 44 mice; Figure 4D, black) and the paired pulse ratio (PPR, 2nd uEPSC/1st uEPSC) was 2.19 ± 0.78 (median = 2.09, n = 66 in 38 animals, in 14 cell pairs, the 1st uEPSC peak amplitude was 0 pA, precluding the calculation of the PPR, Figure 4E, black). The amplitude of the 3rd uEPSC increased further resulting in a third uEPSC/first uEPSC ratio of 2.99 ± 1.25 (median = 2.74, n = 66 in 38 mice). In Elfn1 KO mice, the amplitude of the first uEPSC was significantly larger (29.0 ± 28.9 pA, median = 13.9 pA, n = 14 pairs in 10 mice; Figures 4C,D magenta) than that in the control mice. In these KO animals, presumed mGluR1α+ /O-LM INs were preselected based on the location, size, shape of their somata, and on their firing patterns upon DC current injections. Following the recordings, we characterized the cells post hoc and found that 10 out of 14 cells had O-LM morphology and the remaining 4 cells had truncated axon, but they were immunopositive for mGluR1α. The 3-times increase in the first uEPSC is accompanied by a significant decrease of the PPR (1.46 ± 0.41, n = 14 pairs in 10 mice; Figure 4E). As discussed above, the altered uEPSCs in Elfn1 KO mice could be the consequence of the loss of mGluR7 and/or Munc13-2. To address the contribution of Munc13-2, we performed paired recordings between CA1 PCs and presumed mGluR1α+ /O-LM INs in slices from conditional Munc13-2 KO mice in which the dorsal CA1 region was injected with Cre- and mCherry-expressing AAVs. In these animals, PCs were selected based on the expression of mCherry, and the presence of Cre-recombinase in the recorded PCs was verified post hoc by immunolabeling for Cre. Only those pairs were included where the Cre immunopositivity was unequivocally identified in the presynaptic PC (Figure 4F). Similar to the Elfn1 KO mice, in this transgenic line, mGluR1α+ /O-LM INs were selected in the slice based on the same criteria as described above and were molecularly and morphologically characterized post hoc (10 out of 20 cells were O-LM cells, 10 out of 20 cells have truncated axons, but they were immunopositive for mGluR1α; Figure 4G). The first uEPSC of the short train had a small amplitude (6.7 ± 7.9 pA, median = 4.3 pA, n = 20 pairs in 13 animals) which was not significantly different from that recorded in the control mice (Figure 4D). The amplitudes of the second (14.0 ± 13.7 pA, median = 7.1 pA) and the third uEPSCs (18.7 ± 19.9 pA, median = 7.6 pA) were also similar to those in the control, resulting in a PPR of 2.30 ± 1.64 (median = 1.66, n = 13 pairs in 10 mice, for 7 cell pairs the amplitude of the 1st uEPSC peak was 0 pA, precluding the calculation of the PPR; Figure 2E), which was not significantly different from that in the control. Since the Munc13-2 KO mice had a different genetic background (C57BL/6N) compared to the control mice (C57BL/6J), we recorded from the contralateral, non-injected hemisphere of Munc13-2 conditional KO mice as an independent control population and found that the amplitude (first: 7.7 ± 5.5 pA, median = 6.7 pA; second: 17.0 ± 11.8 pA, median = 17.3 pA; third: 27.1 ± 18.6 pA, median = 27.1 pA, n = 12 pairs in 6 mice, Figures 4C,D) and PPR (2.4 ± 1.5; Figure 4E) of uEPSCs were not significantly different from those recorded from the control or the Munc13-2 KO mice.