Synapsin II desynchronizes neurotransmitter release at inhibitory synapses by interacting with presynaptic calcium channels

Abstract

In the central nervous system, most synapses show a fast mode of neurotransmitter release known as synchronous release followed by a phase of asynchronous release, which extends over tens of milliseconds to seconds. Synapsin II (SYN2) is a member of the multigene synapsin family (SYN1/2/3) of synaptic vesicle phosphoproteins that modulate synaptic transmission and plasticity, and are mutated in epileptic patients. Here we report that inhibitory synapses of the dentate gyrus of Syn II knockout mice display an upregulation of synchronous neurotransmitter release and a concomitant loss of delayed asynchronous release. Syn II promotes γ-aminobutyric acid asynchronous release in a Ca(2+)-dependent manner by a functional interaction with presynaptic Ca(2+) channels, revealing a new role in synaptic transmission for synapsins.

Figure 1. Syn II deletion modifies the dynamics of the evoked GABA response.

(a) Representative traces of eIPSCs from WT and KO dentate gyrus granule neurons (the average trace is shown in black). (b) Mean (±s.e.m.) amplitude and charge of eIPSCs (closed and open circles represent single experiments for WT and KO neurons, respectively). (c) The decay of the WT eIPSC (left panel; black trace) was normalized to the peak amplitude of the KO eIPSC (grey trace). A two-exponential model was used to fit the fast and slow components of the decay (τ_fast and τ_slow) and their average values (±s.e.m.) are shown in the middle and right panels, respectively. *P<0.05, two-tailed unpaired Student’s t-test. n=23 and n=30 neurons for WT (10 mice) and KO (15 mice), respectively. Analysis of eIPSC kinetics showed that the latency (2.06±0.19 and 1.78±0.28 ms, for WT and KO neurons, respectively, P=0.45) and rise time (2.74±0.34 and 2.85±0.32 ms, for WT and KO neurons, respectively, P=0.81) were similar between genotypes (data not shown). (d) Representative traces of mIPSCs recorded in dentate gyrus granule neurons from WT (black) and KO (grey) mice. (e) Aligned dot plots of frequency, amplitude and kinetic parameters (rise time; τ_fast; τ_slow ) of mIPSCs in neurons from WT (closed symbols) and KO (open symbols) mice. Each dot represents one experiment. *P<0.05, two-tailed unpaired Student’s t-test. n=9 and n=13 neurons from WT (4 mice) and KO (6 mice), respectively.

Figure 2. Syn II deletion does not induce major changes in short-term plasticity at inhibitory synapses.

(a) Representative dual eIPSCs traces recorded from dentate gyrus WT (black) and KO (grey) neurons in response to paired stimulation of the medial perforant path at the indicated ISIs. (b) The mean per cent (±s.e.m.) PPD observed in WT (black symbols; n=8–12) and KO (grey symbols; n=6–13) neurons is plotted as a function of the ISI. (c) Plot of the amplitude of the response to the second pulse (A₂) versus the amplitude of the response to the first pulse (A₁) (10 responses/neuron from n=13 KO and 12 WT neurons,) at 2-s ISI in 2 mM Ca²⁺. Both amplitudes are normalized to the mean A₁ in the recorded ensemble. Linear regression analysis of the data points (black and grey lines for WT and KO, respectively) shows no significant correlation between A₁ and A₂. (d) The inverse of the square coefficient of variation (CV⁻²) of the second IPSC (A₂) normalized by the CV₁ of the first IPSC (A₁) was plotted against the PPR (A₂/A₁) (n=12 and n=16 pairs of WT and KO neurons stimulated at 2 s ISI, respectively). Note that most of the data points of both WT and KO neurons fall below the unitary line, indicating a presynaptic mechanism of short-term depression. (e) Representative traces from WT (black) and KO (grey) in response to stimulation of the medial perforant path for 2 s at 10 Hz. (f) Plot of normalized peak amplitude (±s.e.m.) versus time showing the multiple-pulse depression at 10 Hz in WT (black) and KO (grey) neurons.

Figure 3. Syn II deletion increases the RRP size of synchronously released GABA vesicles.

(a) Representative responses from WT (black) and KO (grey) neurons induced by stimulation of the medial perforant path for 2 s at 40 Hz. Stimulation artifacts were removed for clarity. (b) Plot of the normalized amplitude (means±s.e.m.) versus time showing the multiple-pulse depression during the 40-Hz train in WT (black) and KO (grey) neurons. (c) Cumulative amplitude profile of a 2-s train at 40 Hz. Data points from the last second of the response were fitted by linear regression and the line was extrapolated to time 0 to estimate the RRP_syn. (d) Mean values (±s.e.m.) of the amplitude of the first eIPSC in the train, RRP_syn and P_ves for WT (black bars) and KO (open bars) neurons. (e) Cumulative area profile of a 2-s train at 40 Hz. Data points from the last second of the response were fitted by linear regression and the line was extrapolated to time 0 to estimate the RRP_tot. (f) Mean values (±s.e.m.) of the first eIPSC area in the train and RRP_tot for WT (black bars) and KO (open bars) neurons. For electrophysiology experiments: **P<0.01, two-tailed unpaired Student’s t-test and *P<0.05, Welch’s t-test. n=16 and n=14 neurons from WT (5 mice) and KO (8 mice), respectively. (g) Micrographs of GABA-immunopositive synaptic terminals (stained by 10 nm gold particles) making contact with granule cell dendrites in the molecular layer of the dentate gyrus of WT and KO slices. Scale bar, 200 nm. (h) Mean values (±s.e.m.) of the density of total SVs and the number of docked SVs in presynaptic terminals of WT (black bars) and KO (open bars) neurons. For EM experiments: *P<0.05, unpaired Student’s t-test; n=100/39 synapses from 4 WT mice and 69/72 synapses from 4 KO mice.

Figure 4. Syn II deletion dramatically reduces the delayed asynchronous GABA release.

(a,b) Representative traces of the inhibitory delayed asynchronous response in WT (a; black) and KO (b; grey) dentate gyrus granule cells after stimulation of the medial perforant path. Inset traces represent the first second after the end of the stimulation. (c) Representative traces of sIPSCs recorded in dentate gyrus granule neurons from WT (black) and KO (grey) mice. (d) Frequency (means±s.e.m.) of sIPSC measured at −80 mV in the presence of 50 μM D-APV, 10 μM CNQX, 5 μM CGP 55845 (n=11 and n=10 neurons from 5 WT mice and 9 KO mice, respectively). Each dot represents a single experiment. (e) Plot of the mean (±s.e.m.) charge of the delayed GABA release as a function of time after the end of the stimulation train. Data (black and grey symbols/lines for WT and KO, respectively) were normalized to the pre-train spontaneous release. *P<0.05, **P<0.01 and ***P<0.001, Welch’s t-test. n=26 and n=24 neurons from WT (10 mice) and KO (12 mice), respectively.

Figure 5. Ca²⁺-dependency of delayed asynchronous GABA release in Syn II KO neurons.

(a) Representative traces of eIPSCs from the same WT (black) and KO (grey) dentate gyrus granule neurons in 2 mM Ca²⁺ and after replacement of Ca²⁺ with 2 mM Sr²⁺. (b) Plot of the peak amplitude of eIPSCs before and after the switch to 2 mM Sr²⁺ for WT (black) and KO (grey) neurons. (c) Mean values (±s.e.m.) of the per cent depression in the peak eIPSC amplitude for WT and KO neurons (closed and open bars, respectively) after replacement of Ca²⁺ with Sr²⁺. (d) Representative traces of the eIPSC decay. The trace recorded in WT neurons (black) was normalized to the peak amplitude of eIPSC in KO neurons (grey trace) in 2 mM Sr²⁺. A two-exponential model was used to fit the fast and the slow components of the decay (τ_fast and τ_slow). *P<0.05 and **P<0.01, two-tailed paired (b) and unpaired (c) Student’s t-test. n=5 and n=4 neurons from WT (3 mice) and KO (3 mice), respectively. (e) Representative traces of the delayed asynchronous inhibitory response in 2 mM Sr²⁺ recorded in WT (black) and KO (grey) dentate gyrus granule neurons after stimulation of the medial perforant path. Inset traces represent the first second after the end of the stimulation. (f) Plots of the mean (±s.e.m.) charge of the delayed asynchronous release as a function of time after the end of the stimulation train in WT and KO neurons. Data recorded in the presence of either 2 mM Ca²⁺ (closed symbols) or 2 mM Sr²⁺ (open symbols) were normalized to the respective pre-train spontaneous release. *P<0.05, two-tailed paired Student’s t-test n=11 and n=4 neurons from WT (3 mice) and KO (3 mice), respectively.

Figure 6. Syn II deletion modulates the synchronous/asynchronous release in a Ca²⁺-dependent manner.

(a) Representative traces of eIPSCs from WT (black) and KO (grey) dentate gyrus neurons at low (0.5 mM), normal (2 mM) and high (6 mM) extracellular Ca²⁺ concentrations. (b) Representative traces of the first second after a 2-s train at 40 Hz from WT (black) and KO (grey) neurons at different Ca²⁺ concentrations. (c,d) eIPSCs peak amplitude (c) and normalized area of the first second after the train (d) in WT and KO slices as a function of the increasing extracellular Ca²⁺ concentration. *P<0.05, **P<0.01, Welch’s t-test. n=10 neurons/6 mice for WT and n=9 neurons/5 mice for KO.

Figure 7. Syn II modulates GABA asynchronous release by interacting with presynaptic Ca²⁺ channels.

(a) Representative traces of eIPSCs from WT and KO dentate gyrus granule neurons before (black and grey for WT and KO, respectively) and after blockade of P/Q-type Ca²⁺ current with 0.5 μM ω-Agatoxin-IVA (blue) or of N-type Ca²⁺ current with 1 μM ω-Conotoxin-GVIA (green). (b) Aligned dot plot of the per cent reduction in peak amplitude of eIPSCs in WT (closed circles) and KO (open circles) neurons after the addition of either ω-Agatoxin-IVA (AGA; 0.5 μM) or ω-Conotoxin-GVIA (CONO; 1 μM). (c) Representative traces of the delayed asynchronous inhibitory response in WT and KO neurons (black and grey, respectively) and its modulation by Ca²⁺-channel blockers. Inset traces represent the first second after the end of the stimulation under control conditions and after the addition of ω-Agatoxin-IVA (0.5 μM; blue) or ω-Conotoxin-GVIA (1 μM; green). (d) Plots of the mean (±s.e.m.) charge of the delayed asynchronous release as a function of time after the end of the stimulation train in WT and KO neurons. Data recorded under control conditions (black/grey circles) or in the presence of ω-Agatoxin-IVA (0.5 μM; closed/open blue circles) were normalized to the respective pre-train spontaneous release. *P<0.05, two-tailed paired Student’s t-test. n=5 neurons from both WT (4 mice) and KO (4 mice), respectively. (e) Same as for d, but in the presence of ω-Conotoxin-GVIA (1 μM; closed/open green circles) **P<0.01, two-tailed paired Student’s t-test. n=5 and n=6 neurons from WT (4 mice) and KO (3 mice), respectively. (f,g) Cortical neuron lysates were immunoprecipitated with anti-SynII antibodies or anti-HA antibodies, as indicated (immunoprecipitation (IP)). After electrophoretic separation of the immunocomplexes, membranes were probed with antibodies specific for anti-Cav2.1 (P/Q-type), anti-Cav2.2 (N-type) and anti-Cav2.3 (R-type) Ca²⁺ channel subunits, as indicated (western blotting (WB)). A representative immunoblot is shown (f), together with the quantification of the immunoreactive signal in the immunoprecipitated samples (g), normalized to the binding to the HA control (means±s.e.m.; n=3 independent experiments). T, 100 μg of total lysate; FT, 100 μg of flow-through after immunoprecipitation.