Different PPRs and Ca2+ sensitivities of glutamate/GABA co-transmission at SuM-GC synapses

(A) Diagram illustrating the injection of AAV-DIO-ChR2(H134R)-eYFP into the lateral part of SuM of VGluT2-Cre mouse. (B) (top) Fluorescence image showing the injection site of AAV in the SuM. (bottom) ChR2(H134R)-eYFP-expressing SuM axons are observed in the supragranular layer of the DG. ML, molecular layer; GCL, granule cell layer. Scale bars, Top, 200 μm; Bottom, 50 μm. (C) Z-stacked immunofluorescence images double stained for VGluT2 (red) and VIAAT (blue). The merged image demonstrates the colocalization of VGluT2 and VIAAT in the ChR2-eYFP-expressing SuM terminals (arrowheads). (D) Higher magnification xy, xz, and yz projection images outlined in boxed area in (C) show that a SuM bouton is co-stained with VGluT2 and VIAAT. (E) Proportion of VGluT2- and/or VIAAT-expressing boutons in the ChR2-eYFP labeled boutons (2205 boutons, 7 slices). Both VGluT2 and VIAAT (95.2 ± 0.3%), VGluT2 only (2.7 ± 0.2%), and VIAAT only (0%). (F) Representative traces of EPSC (red, Vh = -70 mV) and IPSC (blue, Vh = 0 mV) evoked with paired pulse illumination (100 ms interval). (G) Summary plot of PPR showing a significant difference between EPSC and IPSC. (H) Representative traces (left) and summary plots (right) of EPSCs and IPSCs in 2.5 and 1 mM extracellular Ca2+. (I) Summary plot of percent reduction in the amplitudes of EPSCs and IPSCs from 2.5 to 1 mM extracellular Ca2+. Data are presented as mean ± SEM. **p < 0.01, ***p < 0.001.

Different presynaptic modulation and Ca2+ chelator-sensitivities of glutamate/GABA co-transmission

(A) Time course summary plots showing the effects of ω-CgTx (500 nM) on co-transmission of glutamate (left) and GABA (right) at SuM-GC synapses. Insets indicate representative traces. (B) Time course summary plots showing that the co-transmission of glutamate (left) and GABA (right) at SuM-GC synapses were inhibited by ω-Aga-IVA (200 nM). Insets show representative traces. (C) (left) Representative traces before (EPSC, red; IPSC, blue) and after (gray) application of DCG-IV (0.1 μM). (right) Summary plot of concentration-response curves. Data are fitted to the Hill equation. Numbers in parentheses indicate the number of cells. (D) (left) Representative traces before (EPSC, red; IPSC, blue) and after (gray) application of baclofen (5 μM). (right) Summary plot of percent inhibition in the amplitudes of EPSCs and IPSCs by 5 μM baclofen. (E) Time course summary plots showing the sensitivity of EPSCs and IPSCs to 100 μM BAPTA-AM (left) and 100 μM EGTA-AM (right). Insets show representative traces. Calibration: 10 pA, 10 ms. Data are presented as mean ± SEM. *p < 0.05, **p < 0.01.

Uniquantal release-mediated synaptic responses evoked by minimal light stimulation or strontium-induced asynchronous release at SuM-GC synapses

(A) Representative traces of biphasic response elicited by light stimulation with maximum light power (left) and stochastically evoked PSCs with minimal light stimulation (100 trials for each cell) (right). Synaptic responses at SuM-GC synapses were recorded at holding potentials of -20 to -30 mV in the presence of TTX (1 μM), 4-AP (1 mM), D-AP5 (50 μM), and Ca2+ (1 mM). Average EPSC (red), IPSC (blue), and biphasic current (green) are superimposed on individual traces. (B) Scatter plot of the amplitude of IPSCs against the amplitude of EPSCs recorded from 17 cells. Success events of 172 PSCs are plotted. (Inset) A single experiment showing 100 trials with an interval of 10 sec elicited stochastically EPSCs (red) and IPSCs (blue). (C) Amplitude histograms of EPSCs (top) and IPSCs (bottom) in same data as in B. (D) Summary graph of PSCs probability. (E) Representative traces showing strontium-induced asynchronous release recorded from GC at holding potentials of -20 to -30 mV. The NMDA receptor antagonist D-AP5 (50 μM) was included in the extracellular solution to eliminate NMDA receptor-mediated EPSCs. Red, blue, and green points indicate detected EPSCs, IPSCs, and biphasic currents, respectively. Calibration: 10 pA, 50 ms. (F) (left) Summary bar graph showing the total number of asynchronous quantal responses (n = 11, 30 trials for each cell) (EPSC: 307 events; IPSC: 257 events; biphasic: 28 events). (right) Averaged number of asynchronous events showing that the majority of asynchronous events are EPSCs and IPSCs, and a few biphasic responses. (G) Cumulative distributions of the amplitudes of EPSCs (top) and IPSCs (bottom) for minimal light stimulation-evoked PSCs and strontium-induced PSCs (EPSC: minimal stim.: 86 events from 17 cells; strontium: 307 events from 11 cells, p = 0.72; IPSC: minimal stim.: 67 events from 17 cells; strontium: 256 events from 11 cells, p = 0.096, Kolmogorov-Smirnov test). Data are presented as mean ± SEM. ***p < 0.001.

Close association of GluN1 and GABAAα1 facing the same SuM terminals

(A, B) Double immunofluorescence for VIAAT (green) and VGluT2 (magenta) in the GC layer of the DG. (B) Higher magnification images of the boxed area in (A) show VGluT2-positive terminals are co-labeled with VIAAT (white arrowheads). (C, D) Post-embedding immunogold electron microscopy showing that immunogold particles for VIAAT (magenta, 5 nm) and VGluT2 (green, 10 nm) and are co-exist in the same SuM terminals of asymmetric (C) and symmetric (D) synapses. NT, nerve terminal; Som, soma. (E, F) Triple immunofluorescence for GluN1 (green), GABAAα1 (red), and VGluT2 (blue) in the GC layer of the DG. (F) Higher magnification images of the boxed region in E. Green and magenta arrowheads indicate GluN1 or GABAAα1 puncta that are apposed to VGluT2 puncta, respectively. (G) Proportion of GluN1 and GABAAα1 puncta in close proximity to 634 VGluT2-positive SuM terminals. (H) Cumulative distribution of the distance between GluN1 and GABAAα1 puncta derived from 241 VGluT2-positive SuM terminals that are double-labeled for GluN1 and GABAAα1. Individual and average values are plotted in inset. Data were obtained from two mice. Scale bars: A, 5 μm; B, E, 2 μm; C, D, 100 nm; F, 1 μm.

Frequency-dependent shift of glutamate/GABA co-transmission balance of SuM inputs in GCs

(A, D) Representative traces of EPSCs (Vh = -70 mV) (A) and IPSCs (Vh = 0 mV) (D) in response to 10 light stimuli at 5 Hz (left) or 20 Hz (right) in 2.5 mM (top) or 1 mM (bottom) extracellular Ca2+. (B) Summary graph of normalized EPSC amplitude plotted against the stimulus number in 2.5 mM extracellular Ca2+. Two-way repeated measures ANOVA, F(2,16) = 13.0, p < 0.001, n = 9 or 10; Tukey post hoc test: ***p < 0.001. (C) Same as (B), but recorded in 1 mM extracellular Ca2+. Two-way repeated measures ANOVA, F(2,16) = 11.4, p < 0.001, n = 9 or 10; Tukey post hoc test: **p < 0.01, ***p < 0.001. (E) Summary graph of normalized IPSC amplitude plotted against stimulus number in 2.5 mM extracellular Ca2+. Two-way repeated measures ANOVA, F(2,16) = 0.003, p = 0.997, n = 9. (F) Same as (E), but recorded in 1 mM extracellular Ca2+. Two-way repeated measures ANOVA, F(2,16) = 0.02, p = 0.981, n = 9. (G) Summary plots showing the normalized amplitudes of 10th EPSCs and IPSCs at 5 Hz, 10 Hz, and 20 Hz in 2.5 mM extracellular Ca2+ (G1: two-way repeated measures ANOVA, F(1,7) = 8.03, #p < 0.05, n = 9 or 10; Tukey’s post hoc test, EPSC versus IPSC, *p < 0.05, **p < 0. 01), or in 1 mM extracellular Ca2+ (G2: two-way repeated measures ANOVA, F(1,7) = 21.76, ##p < 0.01, n = 9 or 10; Tukey’s post hoc test, EPSC versus IPSC, ***p < 0. 001). n.s., not significant. Data are presented as mean ± SEM.

Frequency-dependent modulation of GC firing by SuM inputs

(A-C) (left) Representative traces showing GC spikes in response to sinusoidal current injections without (top) and with (middle) paired light stimulation of SuM inputs at 5 Hz (A), 10 Hz (B), and 20 Hz (C). Bottom trace showing the GC response to sinusoidal current injection paired with light stimulation in the presence of 100 μM picrotoxin (PTX). (right) Summary graph of spike probability against stimulus number. (D-F) Summary plots of spike probability at initial and last three stimulus numbers at 5 Hz (D), 10 Hz (E), and 20 Hz (F). The spike probabilities of stimulus numbers 1-3 and 8-10 were averaged. (D) one-way ANOVA, n = 14, F(2,123) = 28.8, p < 0.001 (#1-3), F(2,123) = 25.8, p < 0.001 (#8-10); Tukey’s post hoc test, ***p < 0. 001, **p < 0. 01. (E) one-way ANOVA, n = 13, F(2,114) = 26.3, p < 0.001 (#1-3), F(2,114) = 11.8, p < 0.001 (#8-10); Tukey’s post hoc test, ***p < 0. 001, **p < 0. 01, *p < 0.05, n.s., not significant. (F) one-way ANOVA, n = 11, F(2,96) = 31.4, p < 0.001 (#1-3), F(2,96) = 9.2, p < 0.001 (#8-10); Tukey’s post hoc test, ***p < 0. 001, **p < 0.01, n.s., not significant. Data are presented as mean ± SEM.

Working hypothesis of synaptic architecture of SuM-GC synapse

A single SuM terminal contains distinct glutamatergic and GABAergic vesicles, which are regulated by P/Q-type Ca2+ channels and also modulated by group II mGluRs and GABAB receptors. Glutamatergic and GABAergic vesicles are loosely and tightly coupled with Ca2+ channels, respectively. At the postsynaptic site of GCs, AMPA/NMDA and GABAA receptors are distributed separately. Based on their molecular composition, glutamatergic and GABAergic synapses are established independently, achieving the distinct glutamatergic and GABAergic transmission.

PPRs of glutamatergic and GABAergic co-transmission show no difference between over-axon and over-bouton illumination (Related to Figure 1)

(A) Diagram of the experimental setup. We recorded from GCs located in the most edge of upper blade of GC layer (∼150 μm). Once synaptic responses were observed by over-bouton illumination (over-the recorded cell soma, position b), then the objective lens was moved to ∼600 μm vertically away in the stratum radiatum of the CA1 region (position a). The light power was adjusted to evoke no response at position a, indicating that SuM-GC synapses on the recorded GCs were out of the illumination field. The objective lens was then returned to position b, and PSCs of SuM-GC synapses were recorded with paired stimulation. Then, the objective lens was moved ∼600 μm along the GC layer (position c) to achieve over-axon illumination. Scale bar: 200 μm. (B) Representative traces of SuM-GC EPSCs evoked with over-CA1 (a), over-bouton (b), and over-axon (c) illumination. (C, D) Representative traces (left) and summary plots of PPRs (right) of SuM-GC EPSCs (C) or IPSCs (D) indicating no significant difference between over-bouton and over-axon illumination of SuM inputs (EPSC: bouton: 0.45 ± 0.03; axon: 0.44 ± 0.04, n = 6, p = 0.72, paired t-test; IPSC: bouton: 0.58 ± 0.04; axon: 0.59 ± 0.04, n = 6, p = 0.32, paired t-test). EPSCs or IPSCs were recorded with over-bouton (top) or over-axon (middle) illumination, respectively. Bottom traces show that a trace by over-axon stimulation was scaled to the peak of a trace by over-bouton stimulation. Magnified traces of the dotted circle show a delayed onset latency of a trace by over-axon stimulation relative to a trace by over-bouton stimulation. (E) Summary plot of PPR indicating a significant difference between EPSC and IPSC evoked by over-axon illumination (EPSC: 0.44 ± 0.04; IPSC: 0.59 ± 0.04, n = 6, p < 0.05, unpaired t-test). (F) Summary plots of latency to the onset of the synaptic response. The delayed latency of the synaptic responses by over-axon illumination indicates that synaptic transmission was mediated by light-evoked action potential propagation toward SuM terminals (EPSC: bouton: 2.0 ± 0.1 ms; axon: 3.8 ± 0.2 ms, n = 7, p < 0.001, paired t-test; IPSC: bouton: 2.3 ± 0.1 ms; axon: 3.7 ± 0.2 ms, n = 6, p < 0.001, paired t-test). Data are presented as mean ± SEM. *p < 0.05, ***p < 0.001, n.s., not significant.

The effect of ω-CgTx on CA3-CA1 transmission (Related to Figure 2)

(A) (top) Schematic diagram illustrating the extracellular field excitatory postsynaptic potential recording (fEPSP) configuration. Both the recording and stimulating electrodes were placed in the stratum radiatum of CA1. (bottom) Representative fEPSP traces before (black) and after (red) application of ω-CgTx (500 nM). (B) Summary time course plot showing inhibition of CA3-CA1 transmission by ω-CgTx (44.5 ± 9.5% of baseline, n = 5, p < 0.01, paired t-test). Data are presented as mean ± SEM.

Minimal light stimulation-evoked synaptic responses at SuM-GC synapses (Related to Figure 3)

Sample traces recorded from GCs at holding potential of -20 to -30 mV in response to minimal light stimulation.

The amplitudes and kinetics of minimal light stimulation-evoked EPSCs or IPSCs at SuM-GC synapses were not altered by blockade of their counterpart currents (Related to Figure 3)

(A) Representative traces of EPSCs (left) and IPSCs (right) evoked by minimal light stimulation of SuM inputs at intermediate membrane potentials. EPSCs or IPSCs were recorded by repeating the light stimulation 100 times before, with the application of 100 μM picrotoxin (PTX) or 10 μM NBQX, respectively. (B) Summary plots show that the amplitudes of EPSCs and IPSCs were not changed after application of picrotoxin or NBQX, respectively (EPSC: before: 7.2 ± 0.6 pA; PTX: 7.4 ± 0.8 pA, n = 5, p = 0.71, paired t test; IPSC: before: 10.5 ± 0.7 pA; NBQX: 10.0 ± 0.9 pA, n = 5, p = 0.18, paired t test). (C) Summary plots of the decay time constant of EPSCs (tau, single-exponential fit) and rise time of IPSCs (10%-90% of IPSC peak amplitude) (tau: before: 3.5 ± 0.3 ms; PTX: 3.5 ± 0.4 ms, n = 5, p = 0.84, paired t test; rise time: before: 5.3 ± 0.9 ms; NBQX: 5.5 ± 1.1 ms, n = 5, p = 0.60, paired t test). Data are presented as mean ± SEM. n.s., not significant.

Short-term changes in EPSC/IPSC ratio during train stimulation at SuM-GC synapses (Related to Figure 5)

(A) Representative PSCs in response to 10 light pulses at 10 Hz recorded from GC at holding potentials of -30 mV before (black) and after application of 10 μM NBQX (blue). The EPSC (red) was isolated by subtracting the blue trace from the black trace. 1 μM TTX, 500 μM 4-AP, 50 μM D-AP5, 100 μM LY341495, and 3 μM CGP55845 were included in the extracellular solution throughout the experiments. (B) Summary graph of normalized EPSC and IPSC amplitudes plotted against stimulus number. Two-way repeated measures ANOVA, F(1,7) = 17.5, p < 0.01, n = 8. Data are presented as mean ± SEM. **p < 0.01.

Preventing postsynaptic saturation and desensitization does not alter short-term depression of EPSCs and IPSCs (Related to Figure 5)

(A and B) (left) Representative traces showing EPSCs in response to 10 Hz (A) or 20 Hz (B) light stimulation before (top) and after (bottom) application of 2 mM gDGG. (right) Summary graph of normalized EPSC amplitudes plotted against stimulus number. 10 Hz: Two-way repeated measures ANOVA, F(1,7) = 1.78, p = 0.22, n = 8. 20 Hz: Two-way repeated measures ANOVA, F(1,7) = 0.65, p = 0.45, n = 8. (C and D) (left) Representative traces showing IPSCs in response to 10 Hz (C) or 20 Hz (D) light stimulation before (top) and after (bottom) application of 300 mM TPMPA. (right) Summary graph of normalized IPSC amplitudes plotted against stimulus number. 10 Hz: Two-way repeated measures ANOVA, F(1,5) = 1.02, p = 0.44, n = 6. 20 Hz: Two-way repeated measures ANOVA, F(1,4) = 0.21, p = 0.67, n = 5. Data are presented as mean ± SEM.

List of primary antibodies used in the present study.