Cerebral Cortex, 2019;00: 1–14

doi: 10.1093/cercor/bhz174
Advance Access Publication Date:
Original Article

Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019

ORIGINAL ARTICLE

Synapsin I Synchronizes GABA Release in Distinct

Interneuron Subpopulations
N. Forte1,2,† , F. Binda1,2 , A. Contestabile3 , F. Benfenati1,2 and P. Baldelli2,4
1 Centerfor Synaptic Neuroscience and Technology, Istituto Italiano di Tecnologia, Largo Rosanna Benzi 10,
16132 Genova, Italy, 2 IRCSS, Ospedale Policlinico San Martino, Largo Rosanna Benzi 10, 16132 Genova, Italy,
3 Department of Neuroscience and Brain Technologies, Istituto Italiano di Tecnologia, Via Morego 30, 16163

Genova, Italy and 4 Department of Experimental Medicine, University of Genova, Viale Benedetto XV, 3, 16132
Genova, Italy
Address correspondence to Fabio Benfenati, Center for Synaptic Neuroscience and Technology, Istituto Italiano di Tecnologia, Largo Rosanna Benzi 10,
16132 Genova, Italy. Email: fabio.benfenati@iit.it
† Present address: Endocannabinoid Research Group, Institute of Biomolecular Chemistry, CNR, Viale Campi Flegrei, 34, 80078 Pozzuoli (NA), Italy

Benfenati F. and Baldelli P. contributed equally to this study

Abstract
Neurotransmitters can be released either synchronously or asynchronously with respect to action potential timing.
Synapsins (Syns) are a family of synaptic vesicle (SV) phosphoproteins that assist gamma-aminobutyric acid (GABA) release
and allow a physiological excitation/inhibition balance. Consistently, deletion of either or both Syn1 and Syn2 genes is
epileptogenic. In this work, we have characterized the effect of SynI knockout (KO) in the regulation of GABA release
dynamics. Using patch-clamp recordings in hippocampal slices, we demonstrate that the lack of SynI impairs synchronous
GABA release via a reduction of the readily releasable SVs and, in parallel, increases asynchronous GABA release. The
effects of SynI deletion on synchronous GABA release were occluded by ω-AgatoxinIVA, indicating the involvement of
P/Q-type Ca2+ channel-expressing neurons. Using in situ hybridization, we show that SynI is more expressed in
parvalbumin (PV) interneurons, characterized by synchronous release, than in cholecystokinin or SOM interneurons,
characterized by a more asynchronous release. Optogenetic activation of PV and SOM interneurons revealed a specific
reduction of synchronous release in PV/SynIKO interneurons associated with an increased asynchronous release in
SOM/SynIKO interneurons. The results demonstrate that SynI is differentially expressed in interneuron subpopulations,
where it boosts synchronous and limits asynchronous GABA release.

Key words: GABA interneurons, hippocampus, parvalbumin, somatostatin, Synapsin I

Introduction release represents a “delayed” release phase generated by

Neurotransmitter release is a multistep and highly regulated
process based on the participation of hundreds of tightly
regulated proteins. Alterations of this sophisticated machinery
lead to defects in synaptic transmission and consequently to
the development of neurological diseases such as epilepsy
or intellectual disability. Two distinct modalities of synaptic
communication exist, differently coupled in time to the action
potential (AP) invasion of the presynaptic terminal. The syn-
chronous release is tightly coupled to the Ca2+ rise in the close
vicinity of Ca2+ channels, at the active zone, while asynchronous
residual Ca2+ that slowly recruits groups of synaptic vesicles
(SVs) far from the Ca2+ sources (Kaeser and Regehr 2014). These
two release modalities characterize distinct subpopulations
of interneurons: parvalbumin (PV)-positive interneurons are
markedly synchronous in releasing γ -aminobutyric acid (GABA),
while cholecystokinin (CCK)- and somatostatin (SOM)-positive
interneurons predominantly release GABA asynchronously
(Hefft and Jonas 2005; Savanthrapadian et al. 2014; Liguz-Lecznar
et al. 2016; Tremblay et al. 2016). Both phasic synaptic inhibition,
derived from synchronous GABA release, and tonic shunting

© The Author(s) 2019. Published by Oxford University Press. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
 2 Cerebral Cortex, 2019, Vol. 00, No. 00
inhibition, largely contributed by asynchronous GABA release
(Medrihan et al. 2015), cooperate for the spatial and temporal
control of network activity. Alterations in the synchronous/asyn-
chronous release ratio (s/a-R) strongly impact on network
excitability and can trigger epileptogenesis (Jiang et al. 2012;
Medrihan et al. 2015).
The synapsins (Syns) compose a family of presynaptic phos-
phoproteins encoded by three distinct genes (Syn1, Syn2, and
Syn3). By interacting with SVs and the actin cytoskeleton, Syns
are involved in both predocking and postdocking stages of neu-
rotransmitter release, regulating SV trafficking and short-term
plasticity (Cesca et al. 2010). Mutations in the human SYN1 and
SYN2 genes were reported to associate with epilepsy, autism
spectrum disorder, or intellectual disability (Garcia 2004; Fassio
et al. 2011a; Giannandrea et al. 2013; Lignani et al. 2013; Corradi
et al. 2014; Nguyen et al. 2015; Guarnieri et al. 2017; Peron
et al. 2018). Similarly, mice lacking SynI and/or SynII display
an epileptic phenotype and cognitive disabilities (Corradi et al.
2008; Etholm et al. 2011, 2012, 2013; Michetti et al. 2017).
The deletion of SynI reduces the size of the reserve pool
(RP) of SVs alters short-term plasticity in excitatory synapses
(Rosahl et al. 1995; for review Cesca et al. 2010; Fassio et al.
2011b). In primary autaptic neurons, SynI deletion increases the
amplitude of evoked excitatory postsynaptic currents (eEPSCs),
an effect that is attributable, at least in part, to a decreased
activation of presynaptic GABAB receptors (Chiappalone et
al. 2009; Valente et al. 2017). On the contrary, SynI knockout
(KO) inhibitory synapses in primary neurons exhibit a decrease
in the amplitude of evoked inhibitory postsynaptic currents
(eIPSCs) and a reduction of readily releasable pool [RRP]) (Baldelli
et al. 2007; Chiappalone et al. 2009). Interestingly, SynIIKO
inhibitory synapses of the dentate gyrus of the hippocampus
exhibit an increased eIPSC amplitude and RRP size associated
with the disappearance of asynchronous release, resulting
in a dramatic increase of the s/a-R and in a sharp decrease
of tonic inhibition (Medrihan et al. 2013). The latter result
was recently confirmed in inhibitory synapses of the CA1
region of the hippocampus, where SynII deletion increased the
s/a-R without affecting the eIPSC amplitude (Feliciano et al.
2017). However, in this hippocampal region of SynIIKO mice,
the increased s/a-R observed in regular spiking (presumably
CCK/SOM) interneurons was associated with desynchronization
of GABA release in fast-spiking (presumably PV) interneurons,
suggesting distinct roles in various interneuron subpopulations
(Feliciano et al. 2017).
Based on these findings, we aimed at clarifying whether
SynI has a specific role in controlling GABA release dynamics
in the various interneuron subpopulations. To this aim, we used
patch-clamp recordings of granule cells in the dentate gyrus of
wild-type (Wt) and SynIKO hippocampi. Opposite to what we
found in SynIIKO mice in the very same area (Medrihan et al.
2013), we observed a strong reduction in synchronous release
and a concomitant increase in asynchronous release. We also
demonstrate that this effect was associated with a significantly
higher expression of SynI in PV interneurons than in CCK/SOM
interneurons. Optogenetic stimulation of specific interneuron
populations revealed that SynI deletion impairs synchronous
release in PV-interneurons and enhances asynchronous release
in SOM interneurons. The results demonstrate that SynI directly
contributes to the synchronization of GABA release and that the
expression of SynI/SynII can represent a push–pull mechanism
regulating the s/a-R of inhibitory transmission and thereby net-
work excitability.
Materials and Methods
Experimental Animals
All experiments were performed on 3–7 weeks old male
Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019
C57BL/6J, SynIKO, PV-cre, SOM-cre, PV-cre/SynIKO, or SOM-
cre/SynIKO. C57BL/6J were obtained from Charles River (Calco,
Italy), SynIKO mice were generated by homologous recombi-
nation and extensively backcrossed on a C57BL/6J background,
PV-cre and SOM-cre mice (Taniguchi et al. 2011) were obtained
from Jackson Laboratories (Bar Harbor, ME) and backcrossed
onto the C57BL/6J background. Since Syn1 is X-linked, we
crossed male PV-cre or SOM-cre homozygous mice with either
C57BL/6J or SynIKO females to obtain single-generation PV-
cre/SynIKO or SOM-cre/SynIKO hemizygous male mice. For
in situ hybridization experiments, male PV-Tomato Red and
SOM Tomato Red were obtained from Jackson Laboratories,
while CCK-GFP mice were kindly provided by Drs Steven A.
Siegelbaum and Peter Jonas. All experiments were performed
in accordance with the guidelines established by the European
Communities Council (Directive 2010/63/EU of 22 September
2010) and approved by the Italian Ministry of Health.
Preparation of Slices
Mice were anesthetized with an overdose of isoflurane, quickly
decapitated and the brain was cut in horizontal slices of 400 μm
thickness using a Microm HM 650V microtome equipped with a
Microm CU 65 cooling unit (Thermo Fisher Scientific, Waltham,
MA) at 2 ◦ C in a solution containing (in mM): 87 NaCl, 25 NaHCO3 ,
2.5 KCl, 0.5 CaCl2 , 7 MgCl2 , 25 glucose, 75 sucrose, and saturated
with 95% O2 and 5% CO2 . Slices were let to recover for 45 min at
35 ◦ C and 30 min at room temperature (RT) in recording solution.
Electrophysiology
Whole-cell recordings were performed with a Multiclamp
700B/Digidata1440A system (Molecular Devices, Sunnyvale, CA)
on visually identified mature dentate granule cells using a
BX51WI microscope (Olympus, Tokyo). All experiments were
performed at RT, except for a subset of data, for which the
bath temperature was maintained at 35 ± 0.5 ◦ C thanks to
an in-line solution heater in combination with a heated stage
chamber controlled by a feedback thermistor placed in the slice
chamber (Warner Instruments, Hamden, CT) The extracellular
solution used for recordings contained (in mM): 125 NaCl,
25 NaHCO3 , and 25 glucose, 2.5 KCl, 1.25 NaH2 PO4 , 2 CaCl2 ,
and 1 MgCl2 , pH 7.3 bubbled with 95% O2 –5% CO2 . A high-
chloride intracellular solution was used for all the experiments,
containing (in mM): 126 KCl, 4 NaCl, 1 MgSO4 , 0.02 CaCl2 , 0.1
BAPTA, 15 glucose, 5 HEPES, 3 ATP, and 0.1 GTP. pH 7.3 with
KOH, 290 mOsm/L. The patch pipette resistance was 3–5 MΩ
when filled with intracellular solution. All experiments were
performed in the presence of 50 μM D-APV, 10 μM CNQX, and
5 μM CGP55845 (all from Tocris) to block NMDA, AMPA, and
GABAB receptors, respectively. ω-Agatoxin IVA (AGA, 500 nM; The
Peptide Institute Inc.) was used to block P/Q-type Ca2+ channels.
Tetrodotoxin (TTX, 0.3 μM; Tocris) was added to block AP induced
by optogenetic activation of PV- and SOM-positive interneurons.
eIPSCs were recorded at a holding potential (Vh ) of −80 mV
in response to extracellular stimulation of the molecular layer
with a monopolar glass electrode placed in the proximity of the
recorded granule cell. The minimal response was found at a
stimulation intensity of 40 μA, lasting between 100 and 500 μs.

All data were acquired with Clampex and analyzed offline with
Clampfit 10.2 (Molecular Devices, Sunnyvale CA, USA), Excel,
and GraphPad.
Viral Injections and Optogenetic Stimulation
Virus injection was performed on PV-Cre, PV-Cre/SynIKO, SOM-
cre, and SOM-Cre/SynIKO mice. Mice from postnatal day 21–28
were anesthetized using isoflurane and placed in a stereotaxic
frame. An injection needle (Hamilton 7000, 1 μL syringe) was
lowered in the proximity of dentate gyrus and 1 μL of the Cre-
dependent AAV1-DIO-ChETA-eYFP viral vector (1012 vp/mL; Penn
Vector Core) was infused at a rate of 0.1 μL/min (injection coor-
dinates: anterior–posterior −2 mm from Bregma, lateral ±1 mm,
dorsal–ventral −2.40 mm). Mice were returned to their home
cage and allowed to recover at least 3 weeks after surgery. Optical
stimuli on hippocampal slices were generated using a Cairn
opto-LED system or a PRIZMATIX UHP-Mic-LED (wavelength: 473
nm; light power at the source: 50 mW) by passing the light beam
through the microscope optical system and conveying it to the
slice through an ×20 water-immersion objective. The light power
measured with a power meter at the level of the slice surface
was approximately 4 mW/mm2 . Light stimuli (1-ms long) were
TTL-triggered with Clampex digital output signals.
Data Analysis
Ten to twenty consecutive eIPSCs at 0.1 Hz were averaged
to calculate peak amplitude and kinetics of the eIPSC. We
considered the peak amplitude as the difference between the
baseline and the peak of the evoked postsynaptic current and
the charge as the area under the postsynaptic current. The
decay time was calculated fitting the deactivation response
with a mono-exponential equation. The size of the RRP
and the probability of synchronous release (RRPsyn and Pr ,
respectively) were calculated using the cumulative amplitude
analysis (Baldelli et al. 2007; Chiappalone et al. 2009; Medrihan
et al. 2013). The estimation of the RRPsyn is based on the
assumption that there is a rapid decrease in the number
of readily releasable quanta before reaching a steady state
phase limited by the constant recycling of SVs and that
equilibrium occurs between release and vesicle replenishment,
with Pr during the train approaching 1 (Schneggenburger
et al. 1999). RRPsyn was determined by summing up the peak
amplitudes of eIPSC responses during a 2-s stimulation train @
40 Hz. Lower stimulation frequencies under our experimental
conditions did not fulfill the above-mentioned requirements
of the cumulative amplitude analysis. The number of linearly
distributed data points starting from the last one (i.e., from
the 80th eEPSC), identified by adopting an objective criterion,
were fitted by linear regression and back extrapolated to time
0. The intercept with the Y-axis yielded the RRPsyn and the
ratio between the amplitude of the first eIPSC and the RRPsyn
yielded the Pr. To calculate paired-pulse depression (PPD), two
brief stimulation pulses were applied at interstimulus intervals
ranging between 10 ms and 4 s (10 sweeps for each trial) and the
ratio between the second and the first response (paired-pulse
ratio [PPR]) was calculated.
All the experiments for the evaluation of synchronous and
asynchronous release were performed using increasing stimula-
tion intensities. We started from the minimal stimulation (40 μA)
and increased the stimulation intensity of 3- (120 μA) and 4-fold
(160 μA) to progressively recruit more inhibitory fibers from
Synapsin I Synchronizes GABA Release Forte et al. 3
all the interneurons contacting the patch-clamped granule cell.
Each stimulation step was followed by a high-frequency train
(2 s @ 40 Hz) to evoke asynchronous release (Lu and Trussell
2000; Hagler and Goda 2001; Atluri 2006; Manseau et al. 2010).
Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019
The charge corresponding to the first 25 ms of the eIPSC was
used as an estimation of synchronous release to normalize asyn-
chronous release. The estimation of the synchronous charge
in this time frame (25 ms) avoids any possible contamination
from the asynchronous release component since a correlation
exists between the extent of asynchronous release and the decay
phase of the eIPSC (Isaacson and Walmsley 1995; Medrihan et
al. 2013; Iremonger and Bains 2016). The asynchronous release
was measured as the charge in the 10 s following the train
(2s@40Hz) for the extracellular stimulation experiments and was
normalized, for each individual neuron, to the first 25 ms charge
of the mean eIPSC average of 10 eIPSCs recorded at the same
stimulation intensity before applying the train. The temporal
profile of the delayed asynchronous release after the train (at
160 μA stimulation intensity) was calculated by dividing the
postsynaptic current recordings following the train in ten 1-
s area bins. Asynchronous release in optogenetic experiments
was analyzed as the charge in the 2 s following the light stimula-
tion and normalized with respect to the 2 s-area of spontaneous
activity before the train (Medrihan et al. 2013). For the calculation
of the GABA tonic current, we voltage clamped the granule
neurons at −80 mV and, after 30 s/1 min of baseline, we acutely
applied bicuculline (100 μM) and calculated the difference in the
holding current before and after the treatment. The amplitude
and frequency of spontaneous IPSCs (sIPSCs), recorded at −80
mV in the absence of TTX, were analyzed in a timeframe of 6 min
before the application of bicuculline and used for the analysis of
the GABAergic tonic current.
In situ Hybridization
PV-Tomato Red, SOM-Tomato Red, and CCK-GFP male mice
were deeply anesthetized with 20% urethane (0.1 mL/10 g
body weight) and perfused transcardially with 0.1 M phosphate
buffer containing 4% paraformaldehyde (pH 7.4). Brains were
postfixed overnight, cryoprotected in 30% sucrose, and frozen
in optimal cutting temperature. Horizontal sections (14 μm
thick) were cut using a CM 3050S Cryostat (Leica Microsystems,
Wetzlar, Germany) and collected on Superfrost slides. Before
starting hybridization, PV-tomato, SOM-tomato, and CCK-GFP
fluorescent signals in the hippocampal dentate gyrus area
were acquired using an Eclipse epifluorescence microscope
(Nikon, Tokyo, Japan). The antisense sequence designed to
target SynIa/SynIb messenger RNAs (mRNAs) (atgaactacct gcg-
gcgccgc ctgtcggaca gcaacttcat ggccaatctg ccgaatgggt acatgacaga
cctgcagcgc ccgcaaccgc ccccgccgcc tccctcggcc gccagccctg gggc-
cacgcc cggctccgcg acagcctctg ccgagagggc ctccacagct gctccagtgg
cttctccagc agcccctagt cctgggtcct cggggggcgg cggcttcttc tcgtcgctgt
ctaacgcggt caagcaaacc acagcagccg cagccgccac cttcagcgag
caggtgggcg gtggctctgg gggcgcaggc cgcgggggcg ccgccgccag ggt-
gctgctg gtcatcgacg aaccgcacac cgactgggca aaatacttca aagggaa-
gaa gatccatgga gaaattgaca ttaaagtaga gcaagctgaa ttctctgatc)
was labeled with digoxigenin (DIG) using the 3 -DIG label-
ing kit (Roche, Milano, Italy). Sections were permeabilized
with radioimmunoprecipitation assay buffer. After blocking
of positive charges with acetic anhydride (0.25% v/v) in
triethanolamine buffer (100 mM triethanolamine, 0.002% acetic
acid) and washing with phosphate buffered saline (PBS), slices
were prehybridized with hybridization solution (50% formamide,
 4 Cerebral Cortex, 2019, Vol. 00, No. 00
5× SSC, 5× Denhardts, 50 μg/mL Harring Sperm DNA, 250
μg/mL Yeast tRNA) at 70 ◦ C in hybridization solution. At day
2, sections were washed with posthybridization buffer (50%
formamide, 2× SSC, 0.1% Tween-20) and B1 buffer (10× Maleic
Acid, 0.1% Tween-20) at RT. Thereafter, slices were blocked with
B2 buffer (B1 buffer containing 10% normal goat serum) for 1
h and incubated with alkaline phosphatase-conjugated anti-
DIG antibody (1:2000; Roche) overnight at 4 ◦ C. At day 3, slices
were washed with buffer B1 and then with buffer B3 (100 mM
TrisCl pH 9.5, 50 mM MgCl2 , 100 mM NaCl, and 0.1%Tween-20)
once for 30 min and incubated overnight at 4 ◦ C with nitro blue
tetrazolium/5-bromo-4-chloro-3-indolyl phosphate (NBT/BCIP;
1 tablet in 10 mL ddH2 O; Roche). The reaction was terminated
by several washes with 0.1% Tween-20 in PBS. After Hoechst
staining (1:400 in PBS) for 10 min, slices were air-dried and
mounted with ProLong Gold (Invitrogen). All solutions were
prepared using diethyl pyrocarbonate water and the cryostat
cleaned with RNase ZAP (Sigma). The specificity of the SynI
probe was assessed using SynIKO hippocampal slices. The
mRNA signals were acquired using the Nikon Eclipse Microscope
and the signals of the interneuron markers and of SynI mRNA
were merged using Adobe Photoshop and analyzed with ImageJ.
Statistical Analysis
Data are reported as box plots, except for the graphs, in which
means ± SEM are shown for clarity. The box plots elements
are: center line, median (Q2); square symbol, mean; box lim-
its, 25th (Q1)-75th (Q3) percentiles; whisker length is deter-
mined by the outermost data points. Normal distribution was
assessed using D’Agostino–Pearson’s normality test. In the case
of two experimental groups, the statistical analysis was per-
formed using the two-tailed unpaired Student’s t-test for nor-
mally distributed data or the Mann–Whitney’s U-test for non-
normally distributed data. One-way ANOVA/Holm–Sidak’s tests
or Kruskal–Wallis/Dunn’s tests were used to analyze more than
two experimental groups with normally or non-normally dis-
tributed data, respectively.
Results
Deletion of SynI Reduces the Amplitude of eIPSCs
without Affecting Short-Term Plasticity
We previously reported that SynII deletion promotes upregu-
lation of synchronous GABA release and a concomitant loss
of delayed asynchronous release at GABAergic synapses of the
hippocampal dentate gyrus (Medrihan et al. 2013, 2015). Since
we previously showed that SynI deletion impairs GABAergic
transmission in primary hippocampal cultures (Baldelli et al.
2007; Chiappalone et al. 2009), we investigated whether it has
similar or distinct functions on synchronous and asynchronous
release with respect to SynII. To this aim, we performed whole-
cell voltage-clamp recordings of dentate gyrus granule neurons
in 3–5 weeks old SynIKO and Wt mice, a period in which SynIKO
mice are still nonepileptic (Rosahl et al. 1995; Corradi et al. 2008).
We placed an extracellular electrode in the molecular layer to
evoke the smallest eIPSC, putatively due to the activation of
a single presynaptic fiber onto granule cells (Medrihan et al.
2013). A significant reduction of eIPSC amplitude was observed
in SynIKO slices with respect to Wt slices (Fig. 1a,b). This effect
was paralleled by a reduction in the total eIPSC area (Fig. 1c),
in the absence of significant changes in the current kinetics
(Fig. 1d). In the same slices, we also ascertained whether these
changes were associated with alterations in GABA tonic inhi-
bition, as reported for SynII KO slices (Medrihan et al. 2015)
or spontaneous synaptic activity. Interestingly, in contrast with
Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019
SynII KO slices, no significant changes were observed in both
the tonic current (Fig. 1e) and the amplitude/frequency of sIPSCs
(Fig. 1f ) in SynIKO slices.
SynI has been implicated in short-term plasticity changes
(increase in paired-pulse facilitation) at excitatory synapses
from hippocampal SynIKO or triple Syn KO slices (Rosahl et
al. 1993; Farisello et al. 2013), while PPD was not affected in
inhibitory synapses of SynIKO primary hippocampal cultures
(Baldelli et al. 2007). Consistently, when inhibitory synapses
of the dentate gyrus were excited by two consecutive stimuli
administered at interpulse intervals ranging between 10 ms and
4 s, no differences in the paired-pulse ratios of monosynaptic
eIPSCs were found between Wt and SynIKO mice (Fig. 1g,h).
A Decrease in RRPsyn in SynIKO Slices Accounts for the
Reduced IPSC Amplitude
We next investigated the cellular basis of the observed decrease
in the eIPSC amplitude using the cumulative amplitude pro-
file analysis during a high-frequency stimulation train (Fig. 2a;
2 s @ 40 Hz (Schneggenburger et al. 1999). We confirmed the
smaller amplitude of the first eIPSC in the train of SynIKO
neurons (Fig. 2b) and observed no significant differences in the
eIPSC depression dynamics during the train (Fig. 2c). To per-
form quantal analysis, the size of RRPsyn was calculated using
a back extrapolation to time 0 of the linear portion of the
cumulative amplitude curve assuming that the slow linear rise
is attributable to the equilibrium between the release-induced
depletion and the constant replenishment of the RRPsyn with
Pr approaching 1 (Fig. 2d). As previously shown in primary neu-
rons (Baldelli et al. 2007; Chiappalone et al. 2009), the analy-
sis clearly indicated that the decreased eIPSC amplitude was
entirely attributable to a decline in RRPsyn , while Pr was not
significantly affected (Fig. 2f ).
The Balance between Synchronous and Asynchronous
Release of GABA Is Altered in SynIKO Mice
We previously showed that SynII deletion is associated with a
total loss of asynchronous GABA release at inhibitory synapses
of the hippocampal dentate gyrus (Medrihan et al. 2013). Thus,
we investigated whether also SynI deletion has an impact on the
s/a-R of GABA at the same synapses. We stimulated presynaptic
fibers by a single pulse of increasing intensity at RT. A weak
stimulus (40 μA) was set to a value capable to activate the
smallest postsynaptic current, putatively due to the activation
of a single presynaptic inhibitory fiber. Subsequently, we pro-
gressively increased the stimulation intensity (up to 160 μA)
to recruit a larger number of inhibitory fibers and yield an
average estimation of the total amount of synchronous and
asynchronous release impinging onto the patch-clamped gran-
ule cell. For each stimulation intensity, single pulses (each last-
ing 100–500 μs) were applied at 0.1 Hz over 1 min. To quantify
the amount of asynchronous release we used high-frequency
stimulation trains (2 s train @ 40 Hz) that evoke a delayed release
that persists for hundreds of milliseconds to seconds after the
end of the stimulation (Lu and Trussell 2000; Hagler and Goda
2001; Farisello et al. 2013).
 Synapsin I Synchronizes GABA Release Forte et al. 5

Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019

Figure 1. SynI deletion reduces the amplitude of the eIPSC without affecting PPD. (a) Representative traces of eIPSCs from Wt and SynIKO dentate gyrus granule
neurons. (b–d) Box plots of eIPSC amplitude (b; 49.88 ± 8.25 pA; n = 28 neurons from ten mice for Wt neurons; 34.93 ± 4.07 pA; n = 28 neurons from six mice for
SynIKO neurons; ∗ P < 0.05, unpaired Student’s t-test with Welch’s correction), charge (c; median 1142 and 643.7 (pA × ms) × 103 for Wt and SynIKO respectively;
∗∗ P < 0.01, Mann–Whitney U-test) and decay time (d; 27.97 ± 1.94 ms and 22.53 ± 2.02 ms for Wt and SynIKO respectively; P = 0.57, unpaired Student’s t-test with
Welch’s correction). (e) Representative traces of tonic inhibitory currents (left) and box plot of GABA tonic inhibition (right) in Wt and SynIKO granule neurons (P = 0.15,
unpaired Student’s t-test with Welch’s correction; n=11 neurons from four Wt and n = 10 from three SynIKO mice). (f ) Representative traces of spontaneous inhibitory
currents (sIPSC; left) recorded in the absence of TTX and box plots of sIPSC amplitude (middle) and frequency (right) in Wt and SynIKO neurons (sIPSC amplitude: P =
0.98, unpaired Student’s t-test with Welch’s correction; sIPSC frequency: P = 0.14, Mann–Whitney U-test; n = 11 neurons from four Wt and n = 10 from three SynIKO
mice). (g) Representative dual eIPSCs traces recorded from dentate gyrus Wt and KO neurons in response to paired stimulation of the inhibitory fibers at the indicated
interstimulus intervals. (h) Mean (± SEM) PPR observed in Wt and KO neurons plotted as a function of the interevent interval (n = 18 neurons from six Wt mice and n
= 19 neurons from seven SynIKO mice; P > 0.1 at all intervals, Student’s t-test with Welch’s correction).
 6 Cerebral Cortex, 2019, Vol. 00, No. 00

Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019

Figure 2. The effect of SynI deletion on eEPSC amplitude is due to a decrease of the RRPsyn . (a) Representative responses from Wt and SynIKO neurons induced by
stimulation of the inhibitory fibers in the molecular layer of dentate gyrus for 2 s @ 40 Hz. Stimulation artifacts were removed for clarity. (b) Box plot of the amplitude
of the first eIPSC in the train (median 114.90 pA; n = 13 neurons from seven mice for Wt; median 60.71 n = 15 neurons from five mice for SynIKO; ∗∗ P < 0.01, Mann–
Whitney U-test). (c) The plot of the normalized amplitude (means ± SEM) versus time showing the multiple-pulse depression during the 2 s @ 40 Hz train in Wt and KO
neurons. (d) Cumulative amplitude profile of a 2 s train @ 40 Hz. Data points of the last 30 responses were fitted by linear regression and the line was extrapolated to
time 0 to estimate the RRPsyn . (e) Box plot of RRPsyn in Wt and SynIKO mice (1017 ± 152 pA and 588.80 pA ± 77.58 pA for Wt and SynIKO mice respectively; ∗ P < 0.05,
unpaired Student’s t-test with Welch’s correction). (f ) Box plot of the Pr in Wt and SynIKO mice (0.162 ± 0.02 and 0.144 ± 0.020 for Wt and SynIKO mice respectively; P
= 0.52, unpaired Student’s t-test with Welch’s correction).

To yield an estimation of the synchronous charge not con-
taminated by any asynchronous component that may affect
the decay phase of the eIPSC, we referred to the charge cor-
responding to the first 25 ms of the eIPSC (Fig. 3a, left panel).
The reliability of this estimation was shown by the linear cor-
relation found between the RRPsyn and the respective 25 ms
area of the first eIPSC in the train (Fig. 3a, middle panel). The
synchronous charge transferred by GABA release was markedly
lower in SynIKO than in Wt slices at all stimulation intensi-
ties (Fig. 3a, right panel). On the contrary, the absolute charge
transferred by asynchronous release was 2-fold larger in SynIKO
than in Wt slices at 160 μA stimulation intensity (Fig. 3b,c). When
the asynchronous release was normalized to the 25-ms syn-
chronous charge, that is, to a parameter reflecting the num-
ber of activated presynaptic fibers and the quantal parameters
of release, its normalized values were strongly increased in
SynIKO slices at all stimulation intensities (Fig. 3d). The time-
course of asynchronous release in SynIKO synapses confirmed
the enhanced release from 2 to 5 s after the high-frequency
train and was characterized by a much slower return to baseline
than in Wt slices (Fig. 3e). The effect of SynI deletion was even
enhanced when the asynchronous release was normalized to
the area of sIPSCs recorded before the train (not shown).
Since it was shown that, in excitatory synapses, increasing
the temperature to physiological levels decreases asynchronous
release (Pyott and Rosenmund 2002), we repeated the above-
described experiments at 35 ◦ C (Fig. 4). Both, the decrease of
synchronous release (Fig. 4a) and the increase in asynchronous
GABA release (Fig. 4b–d), were preserved in SynIKO slices at 35 ◦ C.
Moreover, under these conditions, the increase in the asyn-
chronous release was significant even immediately after the end
of the train ∼5 s after the high-frequency stimulation (Fig. 4e).
These data indicate the existence of nonredundant functions of
Syns in the regulation of GABA release dynamics.
Blockade of P/Q Ca2+ Channels Selectively Occludes
the Impairment in Synchronous GABA Release Due
to SynI Deletion
It is widely accepted that synchronous synaptic transmission
mainly relies on the AP-induced activation of P/Q-type Ca2+
channels and, consistently, synchronous GABA release is largely
inhibited by 500 nM AGA, a selective P/Q-type channels antago-
nist (Hefft and Jonas 2005).
To evaluate the contribution of P/Q-type Ca2+ channels
to the imbalance of synchronous/asynchronous inhibitory
transmission of SynIKO mice, we repeated the experiments in
the presence of AGA (500 nM). In Wt slices, the synchronous
release elicited by 160 μA extracellular stimulation was
markedly reduced in the presence of AGA, while the AGA effect
was strongly attenuated in SynIKO slices (Fig. 5a,b). On the
contrary, AGA did not affect the extent of asynchronous release
in both genotypes (Fig. 5c). From the above data, we calculated
the contribution of P/Q-type Ca2+ channels to synchronous and
asynchronous release in the two genotypes by subtracting the
median current recorded in presence of AGA from that recorded
 Synapsin I Synchronizes GABA Release Forte et al. 7

Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019

Figure 3. The balance between synchronous and asynchronous GABA release is altered in SynIKO mice. (a) Left: Representative traces of eIPSCs at 160 μA of stimulation;
the gray 25-ms area of the eIPSC was taken as a purely synchronous fraction of the eIPSC charge used to normalize the asynchronous charge. Center: correlation between
the charge of the first 25 ms of eIPSCs in the train and the respective RRPsyn calculated using cumulative amplitude analysis. Pearson’s correlation coefficient, ∗∗ P =
0.04. Right: Box plots of the first 25-ms fraction of synchronous charge for eIPSCs evoked at increasing stimulation amplitudes in Wt (black symbols) and SynIKO (gray
symbols) mice. (b) Representative traces of asynchronous GABA release (gray areas) after a 2-s train stimulation (160 μA; 40 Hz) in Wt and SynIKO mice; the gray vertical
lines represent the ten 1-s bins used to evaluate the time course of the asynchronous charge in the two genotypes. (c) Box plots of asynchronous charge calculated in
the first 10 s after the train evoked at increasing stimulation amplitudes in Wt and SynIKO mice. (d) Box plots of asynchronous release as a function of the increasing
amplitude of stimulation and normalized to the 25-ms fraction of synchronous release. (e) Time course of asynchronous charge (mean absolute values ± SEM) after
the end of the stimulation train. For each experiment, at least 10 neurons from three mice per genotype were recorded. ∗ P < 0.05, ∗∗ P < 0.01, unpaired Student’s t-test
with Welch’s correction.

under control conditions (Fig. 5d,e). This analysis demonstrates
that the effect of AGA on synchronous release is occluded in
SynIKO slices and that the synchronous release impairment due
to SynI deletion largely corresponds to the AGA-sensitive, P/Q
Ca2+ channel-mediated fraction of GABA release (Fig. 5d). On
the contrary, AGA-sensitive asynchronous release is virtually
lacking in both genotypes, demonstrating its full independence
from P/Q-type Ca2+ channels (Fig. 5e).
Synchronous GABA Release in PV Interneurons Is
Strictly Dependent on Syn I Expression
It was previously shown that the dynamics of synchronous
GABA release from PV interneurons relies on the recruitment of
P/Q-type Ca2+ channels (Hefft and Jonas 2005). The observation
that P/Q Ca2+ channel blockade occludes the SynIKO-induced
reduction in synchronous GABA release led us to hypothesize
 8 Cerebral Cortex, 2019, Vol. 00, No. 00

Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019

Figure 4. The alteration of synchronous and asynchronous GABA release in SynIKO mice is preserved at physiological temperature. (a) Left: Representative traces of
eIPSC at minimal stimulation at 35 ◦ C; the gray 25-ms area of the eIPSC was taken as a purely synchronous fraction of the eIPSC charge used to normalize asynchronous
charge. Right: Box plots of the first 25-ms fraction of synchronous charge for eIPSCs evoked in Wt (black symbols) and SynIKO (gray symbols) mice. (b) Representative
traces of asynchronous GABA release (clear gray areas) after a 2-s train stimulation @ 40 Hz in Wt and SynIKO mice; the gray vertical lines define the 1-s bins used to
evaluate the time course of the asynchronous charge in the two genotypes. (c) Box plots of asynchronous charge calculated in the first 10 s after a 2-s train @ 40 Hz.
(d) Box plots of asynchronous release normalized to the 25-ms fraction of synchronous release. (e) Time course of asynchronous charge (mean absolute values ± SEM)
plotted as a function of time after the end of the stimulation train. For each experiment n = 13 neurons from three mice per genotype were recorded. ∗ P < 0.05, ∗∗ P <
0.01, unpaired Student’s t-test with Welch’s correction.

that it may specifically involve PV interneurons. Thus, we
crossed SynIKO mice with PV-cre mice and successively injected
a Cre-dependent AAV viral vector in the dentate gyrus to
selectively express the fast excitatory opsin ChETA in PV
neurons (Fig. 6a). We then recorded IPSCs induced by single
light pulses (at 470 nm) or trains of light stimuli (2 s @ 40 Hz)
in patched granule cells of the dentate gyrus (Fig. 6b,d). The
application of TTX inhibited the release of GABA from PV
interneurons, demonstrating that light-evoked GABA release
was strictly dependent on the AP generation, and not on local
depolarization of the presynaptic site (Fig. 6b). The selective
light-dependent activation of PV/SynIKO synapses onto granule
cells phenocopied the reduction in amplitude of synchronous
GABA release evoked by electrical stimulation (Fig. 6c,e), while no
changes were observed in the amount of asynchronous release
after high-frequency trains of light stimuli (Fig. 6f ).
SynI Is Predominantly Expressed in PV Interneurons of
the Dentate Gyrus
Distinct interneuron subpopulations display the propensity for
either synchronous or asynchronous release and are therefore
 Synapsin I Synchronizes GABA Release Forte et al. 9

Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019

Figure 5. ω-Agatoxin IVA occludes the effect of SynI deletion on synchronous GABA release. (a) Representative traces of eIPSCs from Wt (left) and SynIKO (right) slices
incubated in the absence or presence of the P/Q type Ca2+ channel inhibitor ω-agatoxin IVA (AGA; 500 nM). (b) Box plots of the fraction of synchronous release in the
four experimental groups, calculated in the first 25 ms of the synaptic response at 160 μA stimulation intensity (42.77 ± 4.89 pA × ms; n = 9 neurons from five mice
and 9.63 ± 1.63 pA × ms; n = 8 neurons from three mice for Wt and Wt/AGA groups, respectively; 22.86 ± 2.43 pA × ms; n = 10 neurons from three mice and 12.92 ±
2 pA × ms n = 9 neurons from three mice for SynIKO and SynIKO/AGA groups, respectively). ∗ P < 0.05, ∗∗∗ P < 0.001, one-way ANOVA/Holm–Sidak’s tests. (c) Box plots
of the asynchronous release calculated in the first 10 s after a 2s-train @ 40 Hz evoked at 160 μA stimulation intensity in the four experimental groups (6.90 ± 1.33
and 6.72 ± 1.34 pA × ms for Wt and Wt/AGA groups, respectively; 13.49 ± 3.09 and 15.42 ± 2.07 pA × ms; for SynIKO and SynIKO/AGA groups, respectively). ∗ P < 0.05,
one-way ANOVA/Holm–Sidak’s tests. (d,e) AGA-sensitive (black) and AGA-insensitive (gray) fractions of synchronous (d) and asynchronous (e) release. The percentages
of synchronous and asynchronous charge are shown.

characterized by different s/a-Rs. In PV interneurons, GABA
release is more temporally precise than in CCK/SOM interneu-
rons. Indeed, eIPSCs generated by CCK/SOM interneurons are
characterized by long delays, strong fluctuations in latency
and amplitude as well as high failure rates (∼16%), thereby
generating an unstable inhibitory output signal (Hefft and Jonas
2005; Daw et al. 2009; Ali and Todorova 2010; Savanthrapadian
et al. 2014; Yavorska and Wehr 2016; Matos et al. 2018).
Since it was demonstrated in cell cultures that hippocampal
interneurons collectively express all Syn isoforms (Song and
Augustine 2016), we evaluated the possibility that SynI and
SynII are differentially expressed in specific interneuron
subpopulations. Being Syns strictly localized to synaptic
terminals, we measured the codistribution of SynI mRNAs with
the specific markers (PV, CCK, and SOM) of the three main
GABAergic neuronal subpopulations in the dentate gyrus by
in situ hybridization (Fig. 7a). Interestingly, SynI mRNA was
expressed at significantly higher levels in "synchronous" PV
interneurons than in “asynchronous” CCK/SOM interneurons
(Fig. 7b).
Deletion of SynI Increases Asynchronous Release
in SOM Interneurons
The selective decrease of synchronous release in PV interneu-
rons, together with the differential expression of SynI,
 10 Cerebral Cortex, 2019, Vol. 00, No. 00

Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019

Figure 6. Synapsin I KO selectively reduces the synchronous release in PV interneurons. (a) Example of a hippocampal slice from PV-cre mice infected with AAV1-DIO-
ChETA-eYFP Cre-dependent viral vector; GC (Granule Cell Layer). (b) Representative traces of a single eIPSC (left) evoked by 470 nM blue light pulses in PV/Wt (black
trace) and PV/SynIKO slices (red trace). The application of TTX fully inhibited light-evoked GABA release (lower trace). (c) Box plot of eIPSCs amplitude (162.7 ± 21 pA, n
= 15 from four PV/Wt mice; 90.36 ± 14.16 pA, n = 16 from six PV/SynIKO mice). ∗∗ P < 0.01, unpaired Student’s t-test with Welch’s correction). (d) Representative traces
of inhibitory currents evoked by a 2 s-train of light pulses @ 40 Hz evoked in PV/Wt (black trace) and PV/SynIKO (red trace) slices. (e) Box plot of synchronous release
(18.14 ± 1.96 pA × ms × 102 , n = 15 from four PV/Wt mice; 8.6 ± 0.93 pA × ms × 102 , n = 16 from six PV/SynIKO mice; ∗∗∗ P < 0.001; Mann–Whitney U-test) from PV
interneuron of Wt and SynIKO mice calculated in the first 25 ms of the first eIPSC of the 2 s-train @ 40 Hz. (f ) Mean (±SEM) asynchronous release normalized to the 2
s-pretrain spontaneous release charge and plotted against time after the end of the stimulation train (P ≈ 0.5 for each point, Mann–Whitney U-test, n = 10 from five
PV/Wt mice; n = 6 from five PV/SynIKO).

suggests that SynI plays a fundamental role in the expres-
sion of synchronous GABA release, constraining the natural
tendency of neurons to release GABA asynchronously. If this
hypothesis holds true, the deletion of SynI from asynchronous
interneurons should increase the latter release modality.
Thus, we investigated the specific effects of SynI deletion in
these interneurons by generating SOM-cre/SynIKO mice and
expressing the fast channelrhodopsin ChETA specifically in
SOM interneurons by in vivo transduction of dentate gyrus
with the AAV1-DIO-ChETA-eYFP viral vector. Deletion of SynI
had no effects on the very low levels of synchronous release
evoked in these neurons by single light stimuli, when either
amplitude or transferred charge was considered (Fig. 7c,d).
However, when asynchronous GABA release was elicited by a 2
s-train of light stimuli @ 40 Hz, a significant increase in the early
poststimulation phase was observed at SOM/SynIKO synapses
(Fig. 7e,f ).
Discussion
Synchronous and asynchronous modalities of GABA release
play an important role in the regulation of network activity
and excitability. Modifications in the Ca2+ wave dynamics, Ca2+
sensitivity or spectrum of molecular actors participating in
the neuronal communication lead to substantial changes in
synaptic transmission that can result in pathological conditions,
such as epilepsy (Kaeser and Regehr 2014). The timing of
neurotransmitter release is an essential factor in converting the
digital signaling of the AP into the analog process of synaptic
transmission. The synchronous release is at the basis of the
cellular communication in the brain; it is fundamental for
all cerebral functions as demonstrated by the early death of
synaptotagmin I KO mice in which the synchronous release is
completely disrupted (Geppert et al. 1994).
The synchronous release is a hallmark of PV-interneurons
expressing P/Q Ca2+ channels, while asynchrony of release char-
acterizes the other interneuronal populations including CCK
interneurons predominantly expressing N-type Ca2+ channels
and SOM interneurons (Hefft and Jonas 2005; Savanthrapadian
et al. 2014). The time-locking of GABA release with the AP
(synchronous release) achieves a point-to-point space locking
of phasic inhibition signal, while asynchronous GABA release
is integrated in time and represents a more tonic signal that,
followed by spillover, generates a tonic inhibition that spreads
in the extracellular volume (volume control of excitability). A large
body of experimental evidence supports the view that syn-
chronous GABA release by PV interneurons in the hippocampus
plays important roles in network physiology such as feedback
and feedforward inhibition, oscillations, pattern identification,
temporal coding and information processing, network plasticity
and learning (Bartos et al. 2002; Hu et al. 2014; Gan et al. 2017;
Espinoza et al. 2018).
Although the Ca2+ transient following a single AP disappears
after few milliseconds, sustained firing activity gives rise to
 Synapsin I Synchronizes GABA Release Forte et al. 11

Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019

Figure 7. SynI is predominantly expressed in PV interneurons and increases the asynchronous release in SOM interneurons. (a) Representative images of the DG region
in brain slices from PV-tomato, CCK-GFP and SOM-tomato mice, respectively (red). SynI mRNA, analyzed by in situ hybridization, is visualized in green. Scale bar,
100 μm. (b) Box plots of the SynI mRNA expression in PV, CCK and SOM interneurons (61, 41 and 100 neurons from three PV-tomato, CCK-GFP and SOM-tomato mice
respectively). ∗∗ P < 0.01, ∗∗∗ P < 0.001, Kruskal–Wallis/Dunn’s tests. (c) Representative traces of a single eIPSCs evoked by 470 nM blue light pulses in SOM (black trace)
and SOM/SynIKO (red trace) slices. (d) Peak amplitude (SOM/Wt: 13.16 ± 3.78 pA, mean ± SEM, n = 7; SOM/SynIKO: 20.62 ± 5.85 pA, n = 8, from three animals per group;
P = 0.32) and area (SOM/Wt: 2.175 ± 0.6827 pA × ms × 102 , mean ± SEM, n = 7; SOM-SynIKO: 3.61 ± 1.191 pA × ms × 102 , mean ± SEM, n = 8; from three animals
per group; P = 0.33, Mann–Whitney U-test) of synchronous GABA release in SOM interneurons were unaffected by the deletion of SynI. (e) Representative traces of
inhibitory currents evoked by a 2 s-train of light pulses @ 40 Hz evoked in SOM/Wt (black trace) and SOM/SynIKO (red trace) slices. (f ) Mean (±SEM) asynchronous
release normalized to 2 s-pretrain spontaneous release charge and plotted against time after the end of the stimulation train. (SOM/Wt n = 6 neurons and SOM/SynIKO
n = 5 neurons from three mice per experimental group; ∗ P < 0.05, Mann–Whitney U-test).

an increase in the amount of the residual nonbuffered Ca2+ ,
stimulating release in a desynchronized fashion with respect
to the AP (Hagler and Goda 2001; Otsu 2004). Asynchronous
release is a relevant component of neuronal communication in
specific areas of the central nervous system, such as the spinal
cord dorsal horn, deep cerebellar nuclei, the inferior olive or
specific interneuron subtypes present in all brain areas (Hefft
and Jonas 2005; Best and Regehr 2009; Labrakakis et al. 2009; Ali
and Todorova 2010; Savanthrapadian et al. 2014). Asynchronous
GABA release could generate a fluctuating and long-lasting
inhibitory signal ideally suited for the purpose of gain control
in postsynaptic target cells (Mitchell and Silver 2003) and being
integrated over time, it is also causally related to the genera-
tion and extent of tonic shunting inhibition (Medrihan et al.
2013). Because of this, the asynchronous release was found to
be involved in neurological disorders, such as epilepsy, spinal
muscular atrophy, or Alzheimer’s disease (Yang et al. 2007; Ruiz
et al. 2010; Jiang et al. 2012)
Syns are important regulators of SV trafficking and thereby
participate in regulating the equilibrium between RRP and RP in
nerve terminals (Cesca et al. 2010). Deletion of either Syn gene
causes an epileptic propensity that phenotypically emerges at
2–3 months of age in KO mice, in parallel with the progressive
postnatal increase in the expression of Syns I/II at synapses
 12 Cerebral Cortex, 2019, Vol. 00, No. 00
(Bogen et al. 2009). This has led to the idea that Syn I and II may
have redundant functions, given also the sequence identity in
their large N-terminal region. We previously described that SynII
is essential for asynchronous GABA release to occur, and the
specific deficit of asynchronous GABA release in SynIIKO mice
promotes a marked decrease in tonic inhibition, resulting in
hyperexcitability and epileptogenesis. A recent paper confirmed
this result and showed, using specific Ca2+ channel blockers,
that SynII deletion suppressed the asynchronous release com-
ponent at synapses dependent on N-type, but not P/Q-type, Ca2+
channels (Feliciano et al. 2017).
SynI, in addition to its major role in regulating the entry/exit
of SVs into/from the RP, was shown to play a postdocking role in
facilitating the last obligatory steps of exocytosis in primary hip-
pocampal neurons, an effect that was specific for inhibitory neu-
rons (Baldelli et al. 2007; Chiappalone et al. 2009; Lignani et al.
2013). Based on these results, we wanted to clarify whether SynI
has similar or distinct roles in the dynamics of GABA release
with respect to SynII in the intact circuits of the dentate gyrus
of the hippocampus. We have found that, in SynIKO mice, syn-
chronous GABA release is impaired due to a reduction of RRPsyn
size, while the asynchronous release is significantly increased.
Notably, these changes in release dynamics are opposite to
what found in the very same synapses lacking SynII, in which
synchronous release is increased and asynchronous release is
virtually abolished (Medrihan et al. 2013). The potentiation of
synchronous release by SynI is occluded by blockade of P/Q-
type Ca2+ channels, confirming that it acts at SVs located in the
nanodomain associated with these active zone-targeted chan-
nels. Interestingly, SynI was highly expressed in PV interneu-
rons characterized by synchronous release and P/Q Ca2+ chan-
nels, while lower SynI levels characterize CCK/SOM interneu-
rons expressing asynchronous activity and non-P/Q Ca2+ chan-
nels. Selective optogenetic stimulation of SynI depleted PV-
interneurons was associated with a marked impairment of syn-
chronous GABA release that was paralleled by an increased
asynchronous GABA release from SynI depleted SOM interneu-
rons that phenocopied the dual effect observed with electrical
stimulation.
Several studies were performed to uncover the molecular
actors orchestrating the balance between synchronous and
asynchronous release in neurons. A task force of presynaptic
proteins favoring synchronous release exists, comprising P/Q-
type Ca2+ channels, synaptotagmins I/II (Nishiki and Augustine
2004; Bornschein and Schmidt 2018; Chang et al. 2018) VAMP2
(Maximov et al. 2009), Ca2+ buffering proteins (e.g., PV) (Hefft and
Jonas 2005) counterbalancing the pool of presynaptic proteins
favoring asynchronous release, whose members include N-
type Ca2+ channels (Hefft and Jonas, 2005), synaptotagmin
VII (Chen and Jonas 2017; Chen et al. 2017), Doc2 (Yao et al.
2011), VAMP4 (Raingo et al. 2012), and SynII (Medrihan et al.
2013, 2015; Bornschein and Schmidt 2018). These two pools
of presynaptic proteins likely act as a push–pull mechanism
controlling the timeliness of neurotransmitter release with
respect to AP pacing. This dynamic control is strongly affected
by the differential expression of these proteins in the particular
neuronal subpopulation, thereby generating groups of neurons,
in which the synchronous or asynchronous release modality
predominates. This is the case of the main hippocampal PV-
positive interneurons that are mostly synchronous and of
SOM/CCK interneurons that release GABA in a preferentially
asynchronous fashion.
Our results indicate that SynI can be added to the task
force of synaptic proteins favoring GABA synchronous release;
its expression, particularly prominent in PV-interneurons, is
essential to assure an efficient phasic release of GABA in these
Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019
neurons. In addition to boosting synchronous release, SynI could
also act as an inhibitory constraint for asynchronous release.
Indeed, SOM/CCK neurons constitutively expressing low SynI
levels display a prevalent asynchronous release that is further
increased by the genetic deletion of SynI. It is also possible that
low levels or absence of SynI, by simply removing a positive
effect on synchronous release, allow the asynchronous machin-
ery to take over.
The precise molecular mechanisms, by which SynI boosts
synchronous GABA release of PV interneurons and asyn-
chronous GABA release of SOM interneurons, remains to be
identified. We have demonstrated that SynI is critical for
maintaining the correct size of RRPsyn in inhibitory synapses
(Baldelli et al. 2007; Chiappalone et al. 2009). In PV-positive
terminals, RRPsyn is made of SVs docked and primed within
the “nanodomains” of P/Q-type Ca2+ channels (Hefft and Jonas
2005). The lack of SynI is likely to increase the mobility of the
docked/primed SVs that can escape the very small “nanodomain”
around the P/Q-type Ca2+ channels, thus decreasing RRPsyn .
This hypothesis suggests that the smaller RRPtot measured by
hypertonic stimulation in SynIKO autaptic inhibitory neurons
(Baldelli et al., 2007) could be due to the high frequency of PV
interneurons in cultured autaptic neurons from the hippocam-
pus. Indeed, in PV-positive terminals, the SVs leaving the RRPsyn
cannot populate the asynchronous release pool because of the
short lasting and spatially confined Ca2+ -transients resulting
from the strong Ca2+ buffering action of PV. The enhanced SV
mobility in the absence of SynI could also explain the increased
asynchronous release observed in SOM/CCK interneurons. In
these neurons, the higher SV mobility increases the probability
that SVs enter the large “microdomains” surrounding N-type
channels that characterize the asynchronous release dynamics
in CCK/SOM terminals devoid of presynaptic Ca2+ buffers.
The deletion of SynI did not apparently affect the small
fraction of synchronous release activity displayed by CCK/SOM
interneurons. If SynI is required by P/Q-type Ca2+ channels
to operate synchronous release in PV-positive terminals, the
lack of these channels, correlating with a release that occurs
within “microdomains,” in CCK/SOM terminals, together with
the lower SynI expression in these neurons, may explain the
lack of genotype effect on synchronous release in CCK/SOM
interneurons.
In conclusion, our data indicate that SynI and SynII represent
a system for fueling both phasic and tonic GABA release
and, although they play a distinct role for the two release
modalities, they both contribute to the inhibitory control of
network excitability and are causative genes for epileptic
synaptopathies.
Funding
Compagnia di San Paolo Torino (grant IDs ROL 20612 and 9344);
Ministero della Salute Ricerca Finalizzata (GR-2016-02363972);
EU Era-Net Neuron 2017 “Snareopathies” and ITN ECMED
(Grant agreement no. 642881); Italian Ministry of University
and Research (PRIN2017 Immune-synaptopathies: dissecting
the contribution of inflammation to synaptic dysfunctions).
 Synapsin I Synchronizes GABA Release
Notes
We thank Drs Steven A. Siegelbaum (Howard Hughes Medical
Institute, Columbia University Medical Center, New York State
Psychiatric Institute, New York, NY) and Peter Jonas (Institute
of Science and Technology, Vienna, Austria) for kindly providing
the CCK-GFP mice. We also thank Drs Caterina Michetti, Mon-
ica Morini (Italian Institute of Technology, Genova, Italy) and
Michele Cilli (IRCCS San Martino, Genova, Italy) for help in breed-
ing the mice, Silvia Casagrande (Department of Experimental
Medicine, University of Genova, Italy), Arta Mehilli and Diego
Moruzzo (Center for Synaptic Neuroscience, Istituto Italiano di
Tecnologia, Genova, Italy) for assistance in genotyping assays
and in the preparation of primary cultures.
Conflict of Interest: None declared.
References
Ali AB, Todorova M. 2010. Asynchronous release of GABA via
tonic cannabinoid receptor activation at identified interneu-
ron synapses in rat CA1. Eur J Neurosci. 31:1196–1207.
Atluri PP. 2006. The kinetics of synaptic vesicle Reacidification
at hippocampal nerve terminals. J Neurosci. 26:2313–2320.
Baldelli P, Fassio A, Valtorta F, Benfenati F. 2007. Lack of synapsin
I reduces the readily releasable pool of synaptic vesicles
at central inhibitory synapses. J Neurosci Off J Soc Neurosci.
27:13520–13531.
Bartos M, Vida I, Frotscher M, Meyer A, Monyer H, Geiger JRP,
Jonas P. 2002. Fast synaptic inhibition promotes synchronized
gamma oscillations in hippocampal interneuron networks.
Proc Natl Acad Sci U S A. 99:13222–13227.
Best AR, Regehr WG. 2009. Inhibitory regulation of electrically
coupled neurons in the inferior olive is mediated by asyn-
chronous release of GABA. Neuron. 62:555–565.
Bogen IL, Jensen V, Hvalby O, Walaas SI. 2009. Synapsin-
dependent development of glutamatergic synaptic vesicles
and presynaptic plasticity in postnatal mouse brain. Neuro-
science. 158:231–241.
Bornschein G, Schmidt H. 2018. Synaptotagmin Ca2+ sensors
and their spatial coupling to presynaptic Cav channels in
central cortical synapses. Front Mol Neurosci. 11:494.
Cesca F, Baldelli P, Valtorta F, Benfenati F. 2010. The synapsins:
key actors of synapse function and plasticity. Prog Neurobiol.
91:313–348.
Chang S, Trimbuch T, Rosenmund C. 2018. Synaptotagmin-1
drives synchronous Ca2+-triggered fusion by C2B-domain-
mediated synaptic-vesicle-membrane attachment. Nat Neu-
rosci. 21:33–40.
Chen C, Jonas P. 2017. Synaptotagmins: That’s why so many.
Neuron. 94:694–696.
Chen C, Satterfield R, Young SM, Jonas P. 2017. Triple function
of Synaptotagmin 7 ensures efficiency of high-frequency
transmission at central GABAergic synapses. Cell Rep.
21:2082–2089.
Chiappalone M, Casagrande S, Tedesco M, Valtorta F, Baldelli P,
Martinoia S, Benfenati F. 2009. Opposite changes in gluta-
matergic and GABAergic transmission underlie the diffuse
hyperexcitability of synapsin I-deficient cortical networks.
Cereb Cortex. 19:1422–1439.
Corradi A, Fadda M, Piton A, Patry L, Marte A, Rossi P, Cadieux-
Dion M, Gauthier J, Lapointe L, Mottron L et al. 2014. SYN2
is an autism predisposing gene: loss-of-function mutations
alter synaptic vesicle cycling and axon outgrowth. Hum Mol
Genet. 23:90–103.
Forte et al. 13
Corradi A, Zanardi A, Giacomini C, Onofri F, Valtorta F, Zoli
M, Benfenati F. 2008. Synapsin-I- and synapsin-II-null mice
display an increased age-dependent cognitive impairment. J
Cell Sci. 121:3042–3051.
Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019
Daw MI, Tricoire L, Erdelyi F, Szabo G, McBain CJ. 2009.
Asynchronous transmitter release from cholecystokinin-
containing inhibitory interneurons is widespread and target-
cell independent. J Neurosci Off J Soc Neurosci. 29:11112–11122.
Espinoza C, Guzman SJ, Zhang X, Jonas P. 2018. Parvalbumin+
interneurons obey unique connectivity rules and establish a
powerful lateral-inhibition microcircuit in dentate gyrus. Nat
Commun. 9:4605.
Etholm L, Bahonjic E, Heggelund P. 2013. Sensitive and critical
periods in the development of handling induced seizures in
mice lacking synapsins: differences between synapsin I and
synapsin II knockouts. Exp Neurol. 247:59–65.
Etholm L, Bahonjic E, Walaas SI, Kao H-T, Heggelund P. 2012. Neu-
roethologically delineated differences in the seizure behavior
of synapsin 1 and synapsin 2 knock-out mice. Epilepsy Res.
99:252–259.
Etholm L, Lindén H, Eken T, Heggelund P. 2011. Electroencephalo-
graphic characterization of seizure activity in the synapsin
I/II double knockout mouse. Brain Res. 1383:270–288.
Farisello P, Boido D, Nieus T, Medrihan L, Cesca F, Valtorta F,
Baldelli P, Benfenati F. 2013. Synaptic and extrasynaptic ori-
gin of the excitation/inhibition imbalance in the hippocam-
pus of synapsin I/II/III knockout mice. Cereb Cortex N Y N. 1991,
23:581–593.
Fassio A, Patry L, Congia S, Onofri F, Piton A, Gauthier J, Pozzi
D, Messa M, Defranchi E, Fadda M et al. 2011a. SYN1 loss-
of-function mutations in autism and partial epilepsy
cause impaired synaptic function. Hum Mol Genet.
20:2297–2307.
Fassio A, Raimondi A, Lignani G, Benfenati F, Baldelli P. 2011b.
Synapsins: from synapse to network hyperexcitability and
epilepsy. Semin Cell Dev Biol. 22:408–415.
Feliciano P, Matos H, Andrade R, Bykhovskaia M. 2017. Synapsin
II regulation of GABAergic synaptic transmission is depen-
dent on interneuron subtype. J Neurosci Off J Soc Neurosci.
37:1757–1771.
Gan J, Weng S-M, Pernía-Andrade AJ, Csicsvari J, Jonas P. 2017.
Phase-locked inhibition, but not excitation, underlies hip-
pocampal ripple oscillations in awake mice in vivo. Neuron.
93:308–314.
Garcia CC. 2004. Identification of a mutation in synapsin I, a
synaptic vesicle protein, in a family with epilepsy. J Med Genet.
41:183–186.
Geppert M, Goda Y, Hammer RE, Li C, Rosahl TW, Stevens CF,
Südhof TC. 1994. Synaptotagmin I: a major Ca2+ sensor for
transmitter release at a central synapse. Cell. 79:717–727.
Giannandrea M, Guarnieri FC, Gehring NH, Monzani E, Ben-
fenati F, Kulozik AE, Valtorta F. 2013. Nonsense-mediated
mRNA decay and loss-of-function of the protein underlie the
X-linked epilepsy associated with the W356× mutation in
synapsin I. PloS One. 8:e67724.
Guarnieri FC, Pozzi D, Raimondi A, Fesce R, Valente MM, Delvec-
chio VS, Van Esch H, Matteoli M, Benfenati F, D’Adamo P et al.
2017. A novel SYN1 missense mutation in non-syndromic X-
linked intellectual disability affects synaptic vesicle life cycle,
clustering and mobility. Hum Mol Genet. 26:4699–4714.
Hagler DJ, Goda Y. 2001. Properties of synchronous and asyn-
chronous release during pulse train depression in cultured
hippocampal neurons. J Neurophysiol. 85:2324–2334.
 14 Cerebral Cortex, 2019, Vol. 00, No. 00
Hefft S, Jonas P. 2005. Asynchronous GABA release gener-
ates long-lasting inhibition at a hippocampal interneuron-
principal neuron synapse. Nat Neurosci. 8:1319–1328.
Hu H, Gan J, Jonas P. 2014. Fast-spiking, parvalbumin+ GABAergic
interneurons: from cellular design to microcircuit function.
Science. 345:1255263–1255263.
Iremonger KJ, Bains JS. 2016. Asynchronous presynaptic gluta-
mate release enhances neuronal excitability during the post-
spike refractory period. J Physiol. 594:1005–1015.
Isaacson JS, Walmsley B. 1995. Counting quanta: direct measure-
ments of transmitter release at a central synapse. Neuron.
15:875–884.
Jiang M, Zhu J, Liu Y, Yang M, Tian C, Jiang S, Wang Y, Guo H,
Wang K, Shu Y. 2012. Enhancement of asynchronous release
from fast-spiking interneuron in human and rat epileptic
neocortex. PLoS Biol. 10:e1001324.
Kaeser PS, Regehr WG. 2014. Molecular mechanisms for syn-
chronous, asynchronous, and spontaneous neurotransmitter
release. Ann Rev Physiol. 76:333–363.
Labrakakis C, Lorenzo L-E, Bories C, Ribeiro-da-Silva A, De
Koninck Y. 2009. Inhibitory coupling between inhibitory
interneurons in the spinal cord dorsal horn. Mol Pain. 5:24.
Lignani G, Raimondi A, Ferrea E, Rocchi A, Paonessa F, Cesca
F, Orlando M, Tkatch T, Valtorta F, Cossette P et al. 2013.
Epileptogenic Q555X SYN1 mutant triggers imbalances in
release dynamics and short-term plasticity. Hum Mol Genet.
22:2186–2199.
Liguz-Lecznar M, Urban-Ciecko J, Kossut M. 2016. Somatostatin
and Somatostatin-containing neurons in shaping neuronal
activity and plasticity. Front Neural Circuits. 10:48.
Lu T, Trussell LO. 2000. Inhibitory transmission mediated by
asynchronous transmitter release. Neuron. 26:683–694.
Manseau F, Marinelli S, Méndez P, Schwaller B, Prince DA, Hugue-
nard JR, Bacci A. 2010. Desynchronization of neocortical net-
works by asynchronous release of GABA at autaptic and
synaptic contacts from fast-spiking interneurons. PLoS Biol.
8:e1000492.
Matos M, Bosson A, Riebe I, Reynell C, Vallée J, Laplante I, Panatier
A, Robitaille R, Lacaille J-C. 2018. Astrocytes detect and upreg-
ulate transmission at inhibitory synapses of somatostatin
interneurons onto pyramidal cells. Nat Commun. 9:4254.
Maximov A, Tang J, Yang X, Pang ZP, Südhof TC. 2009. Complexin
controls the force transfer from SNARE complexes to mem-
branes in fusion. Science. 323:516–521.
Medrihan L, Cesca F, Raimondi A, Lignani G, Baldelli P, Benfenati
F. 2013. Synapsin II desynchronizes neurotransmitter release
at inhibitory synapses by interacting with presynaptic cal-
cium channels. Nat Commun. 4:1512.
Medrihan L, Ferrea E, Greco B, Baldelli P, Benfenati F. 2015.
Asynchronous GABA release is a key determinant of tonic
inhibition and controls neuronal excitability: a study in the
Synapsin II-/- mouse. Cereb Cortex. 1991, 25:3356–3368.
Michetti C, Caruso A, Pagani M, Sabbioni M, Medrihan L, David
G, Galbusera A, Morini M, Gozzi A, Benfenati F et al. 2017. The
knockout of Synapsin II in mice impairs social behavior and
functional connectivity generating an ASD-like phenotype.
Cereb Cortex. 27:5014–5023.
Mitchell SJ, Silver RA. 2003. Shunting inhibition modulates neu-
ronal gain during synaptic excitation. Neuron. 38:433–445.
Nguyen DK, Rouleau I, Sénéchal G, Ansaldo AI, Gravel M, Ben-
fenati F, Cossette P. 2015. X-linked focal epilepsy with reflex
bathing seizures: characterization of a distinct epileptic syn-
drome. Epilepsia. 56:1098–1108.
Nishiki T, Augustine GJ. 2004. Synaptotagmin I synchronizes
transmitter release in mouse hippocampal neurons. J Neu-
Downloaded from https://academic.oup.com/cercor/advance-article-abstract/doi/10.1093/cercor/bhz174/5546054 by Airlangga University user on 07 November 2019
rosci Off J Soc Neurosci. 24:6127–6132.
Otsu Y. 2004. Competition between phasic and asynchronous
release for recovered synaptic vesicles at developing hip-
pocampal Autaptic synapses. J Neurosci. 24:420–433.
Peron A, Baratang NV, Canevini MP, Campeau PM, Vignoli
A. 2018. Hot water epilepsy and SYN1 variants. Epilepsia.
59:2162–2163.
Pyott SJ, Rosenmund C. 2002. The effects of temperature on
vesicular supply and release in autaptic cultures of rat and
mouse hippocampal neurons. J Physiol. 539:523–535.
Raingo J, Khvotchev M, Liu P, Darios F, Li YC, Ramirez DMO,
Adachi M, Lemieux P, Toth K, Davletov B et al. 2012. VAMP4
directs synaptic vesicles to a pool that selectively main-
tains asynchronous neurotransmission. Nat Neurosci. 15:
738–745.
Rosahl TW, Geppert M, Spillane D, Herz J, Hammer RE, Malenka
RC, Südhof TC. 1993. Short-term synaptic plasticity is altered
in mice lacking synapsin. Cell. 75:661–670.
Rosahl TW, Spillane D, Missler M, Herz J, Selig DK, Wolff JR, Ham-
mer RE, Malenka RC, Südhof TC. 1995. Essential functions
of synapsins I and II in synaptic vesicle regulation. Nature.
375:488–493.
Ruiz R, Casañas JJ, Torres-Benito L, Cano R, Tabares L. 2010.
Altered intracellular Ca2+ homeostasis in nerve terminals
of severe spinal muscular atrophy mice. J Neurosci Off J Soc
Neurosci. 30:849–857.
Savanthrapadian S, Meyer T, Elgueta C, Booker SA, Vida I, Bartos
M. 2014. Synaptic properties of SOM- and CCK-expressing
cells in dentate Gyrus interneuron networks. J Neurosci.
34:8197–8209.
Schneggenburger R, Meyer AC, Neher E. 1999. Released fraction
and Total size of a Pool of immediately available transmitter
quanta at a calyx synapse. Neuron. 23:399–409.
Song S-H, Augustine GJ. 2016. Synapsin isoforms regulating
GABA release from hippocampal interneurons. J Neurosci Off
J Soc Neurosci. 36:6742–6757.
Taniguchi H, He M, Wu P, Kim S, Paik R, Sugino K, Kvitsani D,
Fu Y, Lu J, Lin Y et al. 2011. A resource of Cre driver lines
for genetic targeting of GABAergic neurons in cerebral cortex.
Neuron. 71:995–1013.
Tremblay R, Lee S, Rudy B. 2016. GABAergic interneurons in
the neocortex: from cellular properties to circuits. Neuron.
91:260–292.
Valente P, Farisello P, Valtorta F, Baldelli P, Benfenati F. 2017.
Impaired GABAB-mediated presynaptic inhibition increases
excitatory strength and alters short-term plasticity in
synapsin knockout mice. Oncotarget. 8:90061–90076.
Yang L, Wang B, Long C, Wu G, Zheng H. 2007. Increased
asynchronous release and aberrant calcium channel activa-
tion in amyloid precursor protein deficient neuromuscular
synapses. Neuroscience. 149:768–778.
Yao J, Gaffaney JD, Kwon SE, Chapman ER. 2011. Doc2 is a Ca2+
sensor required for asynchronous neurotransmitter release.
Cell. 147:666–677.
Yavorska I, Wehr M. 2016. Somatostatin-expressing inhibitory
interneurons in cortical circuits. Front Neural Circuits.
10:76.