Published online 2015 Oct 12. doi: 10.1113/JP270704
PMID: 26337248
What You Need in a Paint Protection Film: XPEL, LLumar/SunTek, and 3M all have it. Now before we get into the differences of these films, we should stop and say that those three items I mentioned above are really the most important factors in a Paint Protection Film.
This article has been cited by other articles in PMC.
Abstract
Key points
- Short‐term facilitation takes place at GABAergic synapses between cerebellar Purkinje cells (PCs).
- By directly patch clamp recording from a PC axon terminal, we studied the mechanism of short‐term facilitation.
- We show that the Ca2+ currents elicited by high‐frequency action potentials were augmented in a [Ca2+]i‐dependent manner.
- The facilitation of synaptic transmission showed 4–5th power dependence on the Ca2+ current facilitation, and was abolished when the Ca2+ current amplitude was adjusted to be identical.
- Short‐term facilitation of Ca2+ currents predominantly mediates short‐term facilitation at synapses between PCs.
Abstract
Short‐term synaptic facilitation is critical for information processing of neuronal circuits. Several Ca2+‐dependent positive regulations of transmitter release have been suggested as candidate mechanisms underlying facilitation. However, the small sizes of presynaptic terminals have hindered the biophysical study of short‐term facilitation. In the present study, by directly recording from the axon terminal of a rat cerebellar Purkinje cell (PC) in culture, we demonstrate a crucial role of [Ca2+]i‐dependent facilitation of Ca2+ currents in short‐term facilitation at inhibitory PC–PC synapses. Voltage clamp recording was performed from a PC axon terminal visualized by enhanced green fluorescent protein, and the Ca2+ currents elicited by the voltage command consisting of action potential waveforms were recorded. The amplitude of presynaptic Ca2+ current was augmented upon high‐frequency paired‐pulse stimulation in a [Ca2+]i‐dependent manner, leading to paired‐pulse facilitation of Ca2+ currents. Paired recordings from a presynaptic PC axon terminal and a postsynaptic PC soma demonstrated that the paired‐pulse facilitation of inhibitory synaptic transmission between PCs showed 4–5th power dependence on that of Ca2+ currents, and was completely abolished when the Ca2+ current amplitude was adjusted to be identical. Thus, short‐term facilitation of Ca2+ currents predominantly mediates short‐term synaptic facilitation at synapses between PCs.
Abbreviations
- AAV
- adeno‐associated virus
- AHP
- afterhyperpolarization
- AP
- action potential
- [Ca2+]i
- intracellular Ca2+ concentration
- CaM
- calmodulin
- Cm
- membrane capacitance
- DCN
- deep cerebellar nuclei
- EGFP
- enhanced green fluorescent protein
- GC
- granule cell
- IN
- inhibitory interneuron
- PC
- Purkinje cell
- PPD
- paired‐pulse depression
- PPF
- paired‐pulse facilitation
- PPR
- paired‐pulse ratio
- PSC
- postsynaptic current
Introduction
At many synapses in the central and peripheral nervous system, repetitive activation induces short‐term plasticity of the transmission efficacy for tens of milliseconds to minutes, which has been considered to play an important role in neural information processing (Zucker & Regehr, 2002; Abbott & Regehr, 2004; Fioravante & Regehr, 2011; Regehr, 2012). Different forms of short‐term plasticity have been reported at different synapses. For example, excitatory synapses on a cerebellar Purkinje cell (PC) from a granule cell exhibit paired‐pulse facilitation (PPF) of transmission efficacy (Perkel et al. 1990; Atluri & Regehr 1996; Dittman et al. 2000), whereas those from a climbing fibre show paired‐pulse depression (PPD) (Perkel et al. 1990; Atluri & Regehr 1996; Dittman et al. 2000). Dynamic changes of presynaptic intracellular Ca2+ concentration ([Ca2+]i) and the downstream process of transmitter release have been considered to mediate short‐term plasticity. Subsequent to the pioneering studies conducted at the neuromuscular junction by Katz and Miledi (1968), PPF has been extensively studied and several mechanisms have been suggested, such as temporal summation of residual Ca2+ (Katz & Miledi, 1968; Kamiya & Zucker, 1994), modification of action potential (AP) waveform (Jackson et al. 1991; Geiger & Jonas, 2000), saturation of Ca2+ buffer proteins (Neher, 1998; Rozov et al. 2001; Matveev et al. 2004), facilitated Ca2+ currents through voltage‐gated Ca2+ channels (Forsythe et al. 1998; Borst & Sakmann, 1998; Cuttle et al. 1998), recruitment of an extra pool or reluctant vesicles (Valera et al. 2012; Brachtendorf et al. 2015) and Ca2+‐dependent positive modulation of transmitter release machinery (Bertram et al. 1996; Atluri & Regehr 1998). However, the relative contribution of each mechanism to PPF still remains unclear, especially at inhibitory synapses, mainly as a result of difficulty in precisely measuring the Ca2+ dynamics and vesicular exocytosis in a presynaptic terminal (but see excitatory mossy fibre works by Vyleta & Jonas, 2014). In the present study, using direct patch clamp recording from an axon terminal of the cultured cerebellar PC, we have clarified the mechanism of PPF at GABAergic synapses between PCs.
PCs are the sole output neurons of the cerebellar cortical circuit, and they form synapses not only on deep cerebellar nuclei (DCN) neurons, but also on neighbouring PCs through axon collaterals. PC synapses on another PC show PPF upon high frequency activation (Orduz & Llano, 2007; Bornschein et al. 2013), whereas those on DCN neurons exhibit potent short‐term depression (Telgkamp & Raman, 2002; Pedroarena & Schwarz, 2003). It remains to be clarified how the PC output synapses on a PC and those on a DCN neuron show opposite forms of short‐term plasticity. Regarding the candidate mechanism for PPF, high expression of two Ca2+ buffering proteins, calbindin and parvalbumin, in a PC might possibly cause a supralinear increase of free cytosolic Ca2+ by saturation of Ca2+ buffers during repetitive stimulation, leading to facilitated transmitter release. However, this possibility was excluded by a recent study showing resistance of PPF to the genetic ablation of calbindin and parvalbumin in PCs (Bornschein et al. 2013). Thus, the mechanism of PPF at PC terminals still remains unclear. Direct patch clamp recording from PC axon terminals recently showed that frequency‐dependent attenuation of APs underlies the depression of PC output synapses on DCN neurons (Kawaguchi & Sakaba, 2015). In the present study, using the direct recording technique from a small presynaptic terminal (1–3 μm), we attempted to clarify the mechanism of PPF at PC–PC synapses in culture. We demonstrate that the main factor driving synaptic facilitation at the PC terminals is the [Ca2+]i‐dependent facilitation of Ca2+ current (Forsythe et al. 1998; Borst & Sakmann, 1998; Cuttle et al. 1998).
Methods
Ethical approval
All animal experiments were performed in accordance with the principles of UK regulations, as well as the guidelines regarding care and use of animals for experimental procedures of the National Institutes of Health, USA and Doshisha University, and were approved by the local committee for handling experimental animals in Doshisha University.
Culture
The method for preparing cerebellar neuronal culture was similar to that employed in previous studies (Kawaguchi & Hirano, 2007). Briefly, newborn rats of both sexes were decapitated and their cerebella were obtained, followed by incubation in Ca2+ and Mg2+‐free Hanks’ balanced salt solution containing 0.1% trypsin and 0.05% DNase for 15 min at 37ºC. Cells were then dissociated by trituration and seeded on poly‐d‐lysine‐coated cover slips in Dulbecco's modified Eagle's medium/Ham's F12‐based medium, with 1% fetal bovine serum. One day after seeding, 75% of the medium was replaced with serum‐free Eagle's basal medium. One week after seeding, PCs were infected with adeno‐associated virus (AAV) vector carrying enhanced green fluorescent protein (EGFP) under the control of CAG promoter (AAV‐CAG‐EGFP) (Kaneko et al. 2011). PCs could be visually identified by their large cell bodies and thick dendrites, as well as EGFP fluorescence that preferentially labelled PCs by relatively specific AAV (serotype 2) infection. Immunocytochemistry confirmed that almost all of EGFP‐positive neurons were positive for calbindin, a molecular marker of PCs in the cerebellum (data not shown). An axon of PC was clearly different from dendrites that have a high density of spines. We selected PC axon terminals for patch clamp recording based on the EGFP fluorescence. Each week after seeding, half of the medium was replaced with fresh one containing 4 μm cytosine β‐d‐arabino‐furamoside to inhibit glial proliferation. Electrophysiological experiments were performed 3–6 weeks after seeding.
Electrophysiology
Electrophysiological experiments were performed similarly to previous studies (Kawaguchi & Sakaba, 2015) at room temperature (20–24°C). Whole‐cell patch clamp recording from a cultured PC was performed with an amplifier (EPC10; HEKA, Lambrecht, Germany) in an extracellular solution containing (in mm) 145 NaCl, 5 KCl, 2 CaCl2, 1 MgCl2, 10 Hepes and 10 glucose (pH 7.3). In some experiments, 2,3‐dioxo‐6‐nitro‐1,2,3,4‐tetrahydrobenzo[f]quinoxaline‐7‐sulfonamide (NBQX, 10 μm; Tocris Cookson, Bristol, UK), tetraethylammonium (2 mm) and TTX (1 μm; Wako Pure Chemical Industries, Tokyo, Japan) were used to inhibit glutamatergic EPSCs, K+ channels and APs, respectively.
We searched for a synaptically connected PC pair based on axonal paths of a PC visualized by EGFP, showing clear several contacts on the soma and/or dendrites of another PC. PC pairs with a relatively short soma‐to‐soma distance (∼200 μm) were selected for the experiments, as in slice preparation (Orduz & Llano, 2007). Presynaptic PC soma was whole‐cell voltage clamped with a patch pipette (3–5 MΩ) filled with 140 (in mm) potassium‐d‐gluconate, 7 KCl, 5 EGTA, 10 Hepes, 2 ATP and 0.2 GTP. The postsynaptic PC was voltage clamped at –70 mV using a patch pipette (3–5 MΩ) filled with 147 (in mm) CsCl, 5 EGTA, 10 Hepes, 15 CsOH, 2 ATP, 0.2 GTP and 2 QX‐314. The postsynaptic PC was visualized by applying CF568 or CF633 fluorescent dye (100 μm; Biotium, Hayward, CA, USA). Synaptic transmission was triggered by depolarizing the presynaptic PC soma (−70 mV to 0 mV, 1–2 ms) and the postsynaptic currents (PSCs) were recorded from the postsynaptic PC. We analysed recordings in which the series resistance for postsynaptic PC was <25 MΩ and the leak current lower than –200 pA. The paired pulse ratio of PSCs was calculated by dividing the amplitude of the second PSC by that of the first one applied at different frequencies (200, 100, 50 and 20 Hz). The second PSC amplitude was obtained by measuring the difference between the peak and the residual current of the first PSC, which was estimated by extrapolation based on the decay time constant. The paired pulse stimulation was applied every 2 s. The time difference between the onset of presynaptic PC soma stimulation and the visually identified onset of PSC was measured as the onset delay.
For terminal recordings, EGFP‐labelled varicosities impinging on the soma or the proximal dendrites of other PC were selected. We recorded from a terminal using a patch pipette (16–18 MΩ) filled with 147 or 152 (in mm) CsCl, 10 Hepes, 2 ATP, 0.2 GTP and either 5 or 0.5 EGTA. The Ca2+ current into the presynaptic terminal was recorded in the presence of 1 μm TTX and 2 mm tetraethylammonium to avoid activation of neighbouring terminals. The effective space clamp at the PC terminal on a PC in culture was estimated by recording capacitive transients in response to a small depolarizing or hyperpolarizing pulse (10 mV) to an axon terminal. Capacitive transients at an axon terminal followed a dual exponential with time constants of 0.15 ± 0.01 ms and 3.4 ± 0.7 ms (n = 11 cells). Assuming that the terminal was connected to a cylinder shape of axon, the terminal size was estimated to be 1.5 ± 0.3 pF, and a clamped axonal region had 4.6 ± 0.8 pF of capacitance. These membrane capacitances corresponded to a terminal with a diameter of 3 μm and an axon of length 20 μm interconnected through a resistance of ∼700 MΩ. Considering the relatively sparse synaptic formation between PCs, the area controlled by the voltage command was probably confined to the patched terminal. This idea was supported by simulation of the electrical circuit consisting of a terminal and an axon. In some experiments, EGTA was replaced with 20 mm BAPTA accompanied with appropriate reduction of CsCl to adjust osmolality. Basal membrane capacitance (Cm) was 2–3 pF, and series resistance of the terminal recordings (typically 70 MΩ) was compensated for by ∼30–80%. Measurements of Cm were carried out using the sine + DC technique (Neher & Marty, 1982) implemented via Patchmaster software (HEKA). Presynaptic terminals were held at –80 mV and the sine wave (1000 Hz and peak amplitude of 30 mV) was applied on the holding potential. Because of large conductance changes during the depolarizing pulse, capacitance was usually measured between 200 and 300 ms after the pulse.
Ca2+ imaging
To record the Ca2+ increase at the presynaptic terminals, we applied CF633 (100 μm; Biotium) and either Oregon Green BAPTA‐1 (OGB‐1, 200 μm, Invitrogen, Carlsbad, CA, USA; Kd = 0.2 μm) or Oregon Green BAPTA‐6 (OGB‐6, 200 μm, Invitrogen; Kd = 3 μm) to the soma of a granule cell (GC), an inhibitory interneuron(IN) or a PC through a patch pipette. The dyes were allowed to diffuse throughout the cell for 25 min before recording started. Axonal varicosities were identified by CF633 fluorescence. The Ca2+ increase upon an AP, which was triggered at the soma by a depolarizing pulse (0 mV, 2 ms), was detected as the fluorescence increase of OGB‐1/OGB‐6. Five to twenty trails were performed per cell, with a 1 min interval between trails. Images of PC axon terminals were obtained with a Zyla sCMOS camera (Andor Technology Ltd, Belfast, UK) and analysed with SOLIS (Andor) or Image J software (NIH, Bethesda, MD, USA).
Statistical analysis
Data are presented as the mean ± SEM. Statistical significance was assessed using a paired t test, an unpaired Students’ t test, the Mann–Whitney U test or ANOVA.
Results
PPF of synaptic transmission at PC–PC synapses in culture
We first tested whether short‐term facilitation takes place at inhibitory synapses between cultured PCs as in slice preparation (Orduz & Llano, 2007). Some of PCs were fluorescence labelled with EGFP using AAV vector (Kaneko et al. 2011), which allowed us to identify a synaptically connected pair of PCs (Fig. (Fig.11A). PCs tend to form synapses on a DCN neuron (Kawaguchi & Sakaba, 2015) and synaptically connected PC–PC pairs were rarely found in culture. We performed whole‐cell patch clamp recordings from a pair of PCs that were located relatively near, as in slice preparation (Orduz & Llano, 2007), and the postsynaptic PC was visualized by fluorescence using CF568 or CF633 applied through a patch pipette. The distance from the axon hillock to the first branching point of the axon collateral was 130 ± 20 μm, and the distance between the presynaptic PC soma and the closest synapse was 530 ± 99 μm (n = 8 pairs).
PPF of synaptic transmission at the cultured PC–PC synapseA, fluorescence image of a synaptically‐connected PC–PC pair. Green: a presynaptic PC expressing EGFP. Red: a postsynaptic PC visualized by CF633 applied through a patch pipette (removed when the image acquisition). To show several synapses around the postsynaptic PC soma and proximal dendrites, the area surrounded by a white rectangle is presented as an enlarged image. B, representative traces of presynaptic Na+ and K+ currents and PSCs upon paired pulse stimulation at a 10 ms interval (depolarization pulses to 0 mV for 2 ms). Presynaptic currents through voltage‐dependent channels were isolated by subtracting the 7‐fold capacitive currents upon the 10 mV depolarization pulse. C, normalized PSC traces upon paired‐pulse stimulation at 5, 10, 20 or 50 ms intervals. D, averaged ratio of PSC amplitudes (PSC2/PSC1) is plotted against the interval of paired pulse stimulation (n = 6–7 pairs).
Both pre‐ and postsynaptic PCs were voltage clamped at −70 mV. Paired pulse stimulation consisting of two depolarization pulses (to 0 mV, 2 ms) was applied to the presynaptic PC soma (to elicit APs) at different intervals (5, 10, 20 and 50 ms) and synaptic transmission was recorded from the postsynaptic PC. The PSC was recorded as inward currents because of a high concentration of internal Cl−, and had an average amplitude of 409 ± 153 pA (n = 8 pairs with six synaptic contacts on average), onset delay of 2.6 ± 0.2 ms, 10–90% rise time of 2.0 ± 0.2 ms and half‐height width of 15 ± 2 ms. As shown in Fig. Fig.11B, the amplitude of the second PSC was larger than that of the first one, and the extent of facilitation depended on the stimulation interval (50 ms, 107 ± 2%; 20 ms, 122 ± 4%; 10 ms, 134 ± 6%; 5 ms, 150 ± 9%; n = 6–7 cells) (Fig. (Fig.11C and D). The time course of PPF was comparable to that reported in slice preparation (Orduz & Llano, 2007). Thus, the facilitation property of GABAergic synapses between PCs is preserved in the dissociated culture.
Short‐term facilitation of Ca2+ currents in a PC terminal
To study the mechanism underlying short‐term facilitation at PC–PC synapses, we directly recorded from the EGFP‐positive presynaptic terminal (1–3 μm in diameter) impinging on the soma or proximal dendrite of a PC (Fig. (Fig.22A). By taking advantage of direct voltage clamp recording from a presynaptic terminal, we first recorded the presynaptic Ca2+ currents through voltage‐gated Ca2+ channels upon an AP and the resultant postsynaptic response. Accordingly, the terminal was stimulated by a voltage command with an AP waveform that was recorded from a PC terminal on a DCN neuron in a previous study (Kawaguchi & Sakaba, 2015), which was similar to the AP waveform recorded from a PC terminal on a PC (Fig. (Fig.22B). To restrict the stimulation to the patched terminal, we patch clamp recorded from a relatively isolated terminal in the presence of 1 μm TTX. Paired recordings from a PC axon terminal and a postsynaptic PC soma showed that the presynaptic AP command evoked the PSCs with mean amplitude of 116 ± 32 pA (n = 10 pairs). When the amplitude of a single AP command to the presynaptic terminal was changed to elicit different amplitudes of Ca2+ current, the amplitude of PSC also changed. As the AP command became larger in amplitude, the Ca2+ current almost linearly increased (Fig. (Fig.22C). On the other hand, the PSC amplitude changed supralinearly, and showed 4–5th power dependenceon the Ca2+ current amplitude (Fig. (Fig.22C). Because PCs receive inputs from many neurons, miniature PSCs were isolated by analyzing asynchronous events following a strong pulse to the terminal. Miniature PSCs, observed as asynchronous releases after a square pulse to the terminal, had a mean amplitude of 52 ± 3 pA (n = 6 pairs) (Fig. (Fig.22D). Thus, a few synaptic vesicles are exocytosed upon the AP command, which is consistent with a previous study (Kawaguchi & Sakaba, 2015). In addition, the amplitude of PSCs was in a reasonable range considering that a PC pair showing the mean PSC amplitude (409 pA) under somatic paired recordings had approximately six synaptic connections located at various postsynaptic compartments (Fig. (Fig.1).1). Taken all of these observations together, the AP voltage command to the PC axon terminal causes synaptic transmission compatible with the physiologically evoked one.
PC–PC synaptic transmission upon an AP voltage command at a terminalA, representative image for paired recordings from the presynaptic PC terminal (highlighted by a white arrowhead) and the postsynaptic PC soma. Both pre‐ and postsynaptic PCs are EGFP‐positive in this case. B, AP waveform recorded from a PC terminal impinging on a PC (left) and the AP waveform used for voltage command in the present study that was recorded from a PC terminal on a DCN (right). C, left: representative traces of different amplitudes of AP commands (Vcom), the Ca2+ currents (ICa2+) in a presynaptic PC terminal and the PSCs simultaneously recorded from the postsynaptic PC. Right: PSC amplitudes upon various amplitudes of AP commands were plotted against the ICa2+ amplitudes. The grey line represents the 4.5th power relationship between x‐ and y‐axis. D, left: a representative trace of PSC evoked by the 5 ms depolarization pulse to 0 mV in the presynaptic terminal. For hundreds of milliseconds after the large PSC, asynchronous events could be observed. These events were used for sampling mPSCs because this is the most reliable way of collecting miniature events from a given presynaptic PC terminal. Right: amplitude histogram of asynchronous PSCs. Pooled data from six cells are shown.
To examine whether the Ca2+ currents are modulated by high‐frequency AP arrivals, we stimulated the presynaptic terminal by paired pulses of the AP command (Fig. (Fig.33A). As shown in Fig. Fig.33B, the waveform of an AP pair recorded from a PC terminal upon somatic high frequency stimulation showed afterhyperpolarization (AHP) between two APs. Therefore, we used a similar voltage command with AHP for the paired pulse stimulation of the PC terminal, unless otherwise stated. The amplitude of Ca2+ current upon the second AP waveform was larger than that upon the first one, showing PPF dependent on the stimulation interval (3.33 ms, 112 ± 2%; 5 ms, 111 ± 2%; 10 ms, 111 ± 1%; 20 ms, 107 ± 1%; 50 ms, 102 ± 2%) (Fig. (Fig.33B, C and E). There are two possible candidate mechanisms for the Ca2+ current facilitation upon the paired‐pulse stimulation of the AP command. One is that AHP between the first and the second AP commands may recover some fractions of voltage‐gated Ca2+ channels from inactivation at resting membrane potential (Brachaw et al. 1997), leading to increased availability of Ca2+ channels at the second stimulation. To test this possibility, we omitted AHP between two AP commands and measured the PPF of Ca2+ currents (Fig. (Fig.33D). As shown in Fig. Fig.33D and E, the PPF of Ca2+ currents without AHP was similar to that observed in the presence of AHP (3.33 ms, 113 ± 3%; 5 ms, 113 ± 2%; 10 ms, 109 ± 1%; 20 ms, 107 ± 1%; 50 ms, 103 ± 1%). The other candidate mechanism is the Ca2+‐dependent facilitation of Ca2+ current, which is caused by direct modulation of voltage‐gated Ca2+ channels by Ca2+‐binding proteins such as calmodulin (CaM) and neuronal Ca2+ sensor (Catterall & Few, 2008; Ben‐Johny & Yue, 2014). To test this possibility, the PPF of Ca2+ current was measured in the presence of higher concentration of Ca2+ chelator in the intracellular solution. Increase of intracellular concentration of EGTA (from 0.5 mm to 5 mm) did not suppress the PPF of Ca2+ current at short intervals but tended to accelerate the recovery of the PPF of Ca2+ current (3.33 ms, 112 ± 4%; 5 ms, 109 ± 2%; 10 ms, 107 ± 2%; 20 ms, 105 ± 1%; 50 ms, 102 ± 0%) (Fig. (Fig.33E). Application of BAPTA (20 mm) was necessary to completely abolish the Ca2+ current facilitation (3.33 ms, 103 ± 1%; 5 ms, 100 ± 2%; 10 ms, 101 ± 0%; 20 ms, 101 ± 1%; 50 ms, 100 ± 1%) (Fig. (Fig.33E). Taking all of these results together, it was suggested that the Ca2+ current through voltage‐gated Ca2+ channels is facilitated upon high‐frequency stimulation depending on the increase in [Ca2+]i.
Ca2+‐dependent facilitation of presynaptic Ca2+ currents at PC–PC synapsesA, image of direct patch clamp recording from an EGFP‐positive axon terminal on a proximal dendrite of another PC. B, pair of APs recorded from a PC terminal on a PC (top), representative traces of paired‐pulse voltage commands with an AP waveform (middle, Vcom) and the resultant Ca2+ currents (bottom, ICa2+). C, normalized Ca2+ currents upon paired‐pulse AP commands at 3.33, 5, 10, 20 or 50 ms interval. The dotted line represents the average peak amplitude of the first Ca2+ current. D, representative traces of the paired‐pulse voltage command with an identical AP waveform without AHP and the resultant Ca2+ currents. E, averaged Ca2+ current facilitation upon various intervals of paired‐pulse AP waveforms in the presence of 0.5 or 5 mm EGTA or of 20 mm BAPTA.The data obtained without AHP of the AP waveform in the presence of 0.5 mm EGTA are also shown (no AHP) [n = 4 (0.5 mm EGTA, 5 mm EGTA and 20 mm BAPTA); n = 5 (no AHP)].
PPF of synaptic transmission is determined by that of Ca2+ currents
Taking into consideration that transmitter release is highly sensitive to the [Ca2+]i (Fig. (Fig.22C), the PPF of Ca2+ current might be a powerful mechanism leading to the short‐term facilitation of synaptic transmission. To study the quantitative relationship between the PPF of Ca2+ current and the PPF of synaptic transmission, we performed paired recordings from a presynaptic PC axon terminal and a postsynaptic PC soma. Paired‐pulse AP waveforms with different intervals caused facilitation in both the presynaptic Ca2+ current amplitude (3.33 ms, 116 ± 2%; 5 ms, 114 ± 2%; 10 ms, 112 ± 2%; 20 ms, 110 ± 2%; 50 ms, 103 ± 2%) and synaptic transmission (3.33 ms, 202 ± 14%; 5 ms, 209 ± 8%; 10 ms, 174 ± 3%; 20 ms, 170 ± 20%; 50 ms, 108 ± 14%) (Fig. (Fig.44A–C). The facilitation of PSC could be fitted to the 4–5th power of the facilitation of Ca2+ currents (Fig. (Fig.44B and C). When the recovery from Ca2+ current facilitation was accelerated by the presence of 5 mm intracellular EGTA (3.33 ms, 114 ± 1%; 5 ms, 112 ± 1%; 10 ms, 109 ± 1%; 20 ms, 105 ± 1%; 50 ms, 102 ± 1%), the PPF of PSC also recovered faster (3.33 ms, 198 ± 14%; 5 ms, 170 ± 9%; 10 ms, 151 ± 12%; 20 ms, 118 ± 16%; 50 ms, 99 ± 11%), showing 4–5th power dependence of the PSC facilitation on the Ca2+ current facilitation (Fig. (Fig.44C). Thus, the PPF of PSC was tightly correlated with that of the Ca2+ currents. The time course of the PSC facilitation observed by paired terminal‐soma recordings in the presence of 5 mm rather than 0.5 mm EGTA was similar to that observed under paired soma–soma recordings from synaptically‐connected PCs (Fig. (Fig.1).1). Thus, the strong Ca2+ buffering capacity of PC terminals containing calbindin and parvalbumin (Bornschein et al. 2013) might be somewhat similar to the situation containing a millimolar‐order of EGTA.
Tight coupling of Ca2+ current facilitation and synaptic facilitation at PC–PC synapsesA, representative traces of ICa2+ (average of 20 traces) and PSCs upon paired pulse voltage commands consisting of an identical AP waveform. Black PSC trace is the average of 20 PSC traces shown in grey. B and C, averaged PPR of ICa2+ (B) and that of PSC (C) in the presence of intracellular 0.5 or 5 mm of EGTA are plotted against the interstimulus interval. Grey or black continuous lines in (C) are the 4.5th power of ICa2+ (shown as dotted lines, which are same as continuous lines in B) [n = 6 pairs (5 mm EGTA); n = 4 pairs) (0.5 mm EGTA)].
To study the causal relationship between the PPF of Ca2+ current and that of PSC, we next changed the amplitude of Ca2+ currents during the second AP by systematically altering the AP amplitude (to 1.05, 1, 0.95, 0.9 or 0.85 times) (Fig. (Fig.55A). As the amplitude of the second AP command decreased, that of the Ca2+ current was also decreased (1.05AP, 123 ± 3%; 1AP, 113 ± 1%; 0.95AP, 104 ± 1%; 0.9AP, 93 ± 2%; 0.85AP, 80 ± 3%) (Fig. (Fig.55A and C). As a result of the decreased Ca2+ current, PPF of synaptic transmission changed into PPD (1.05AP, 236 ± 31%; 1AP, 175 ± 27%; 0.95AP, 136 ± 15%; 0.9AP, 83 ± 15%; 0.85AP, 64 ± 13%)(Fig. 5B and C). When the paired‐pulse ratio (PPR) of PSC was plotted against that of Ca2+ current, their relation matched to the 4–5th power dependence (Fig. (Fig.55C). This result, together with that shown in Fig. Fig.4,4, indicates that the paired pulse ratio of PSCs critically depends on the 4–5th power of Ca2+ current, and the PSC amplitude does not change when the presynaptic Ca2+ current is identical between the first and second pulses. Thus, it appears that the PPF of synaptic transmission is predominantly brought about by the PPF of Ca2+ currents at the PC–PC synapses.
ICa2+ facilitation determines the synaptic facilitationA and B, representative traces of voltage commands consisting of paired‐pulse AP waveforms (Vcom) and the Ca2+ currents (ICa2+) in a presynaptic PC terminal (A) and of PSCs (B). The second AP amplitude was changed (1.05, 1, 0.95, 0.9 or 0.85 times that of the first) to alter the second Ca2+ current amplitude. The PSC amplitudes were normalized to the first one. C, the averaged PPR of PSCs was plotted against that of ICa2+. The data were obtained from (A) and (B) (n = 6 pairs).
Marginal residual Ca2+ in a PC terminal
The above results precluded the contribution of other mechanisms of facilitation, such as the residual Ca2+ hypothesis, which suggested that temporal summation of [Ca2+]i during paired AP arrivals increases the transmitter release probability (Katz & Miledi, 1968). To test this issue further, we performed Ca2+ imaging using OGB‐1 (200 μm, Kd = 0.2 μm) or OGB‐6 (200 μm, Kd = 3 μm) applied into the soma through a patch pipette (Fig. (Fig.66A). After 25 min of dye diffusion, we evoked an AP by applying a single depolarization pulse in the PC soma (to 0 mV, 2 ms) and the fluorescence changes in the axonal varicosities were recorded. The Ca2+ influx into the varicosity resulted in the increase in fluorescence signal of OGB‐1 or OGB‐6, and the relative fluorescence increase (ΔF/F) reflected the amplitude of Ca2+ increases. The ΔF/F upon a single AP was very small in a PC terminal (OGB‐1, 0.11 ± 0.01) compared to the varicosities of an IN (0.38 ± 0.12) or a GC (0.45 ± 0.16). An IN is known to express parvalbumin but not calbindin, whereas a GC lacks both but expresses calretinin (Bastianelli, 2003). The ΔF/F was extremely small in a PC terminal when using OGB‐6 (PC: 0.03 ± 0.01; IN: 0.19 ± 0.05; GC: 0.23 ± 0.03) (Fig. (Fig.66B and C). Thus, the residual [Ca2+]iincrease is tiny in a PC terminal compared to axon terminals of an IN or a GC. The small change of OGB‐1 and OGB‐6 fluorescence upon a single AP in a PC terminal corresponded to an∼10–20 nm of [Ca2+]i increase, which is similar to the value estimated in previous slice studies (Schmidt et al. 2003; Orduz & Llano, 2007). A high expression level of Ca2+ buffer proteins, calbindin and parvalbumin, is probably responsible for the low residual [Ca2+]i increase (Bornschein et al. 2013). Considering our previous estimate of local [Ca2+]i during a single AP as 5–10 μm in a PC terminal (Kawaguchi & Sakaba, 2015), the residual Ca2+ increase by 20 nm is estimated to increase the transmitter release by ∼1% assuming 4–5th power Ca2+ dependence. Thus, in line with the idea that the PPF of PSC is predominantly mediated by that of Ca2+ current, the temporal summation of residual Ca2+probably plays no role in the PPF at PC–PC synapses.
Residual Ca2+ increase at PC terminalsA, representative fluorescence image of a PC loaded with 200 μm OGB‐1. At an axon varicosity (highlighted by an white arrow), fluorescence change was measured before and after the action potential evoked at the soma. B, time courses of normalized fluorescence intensity change of OGB‐1 or OGB‐6 recorded from an axon varicosity from a PC, an IN or a GC. Representative traces for each condition are also shown in grey. C, ΔF/F upon single AP in a GC (n = 6 cells for OGB‐1 and 8 for OGB‐6), IN (n = 5 cells for OGB‐1 and 6 for OGB‐6) or PC (n = 7 cells for OGB‐1 and 5 for OGB‐6). *P < 0.05.
Release properties of PC axon terminals on a PC similar to those on a DCN neuron
The above results indicated that the facilitative property of PC synapses on another PC is predominantly mediated by Ca2+‐dependent facilitation of Ca2+ currents through voltage‐gated Ca2+ channels. On the other hand, PC synapses on a DCN neuron show depression upon high frequency activation (Telgkamp & Raman, 2002) as a result of AP attenuation around terminals (Kawaguchi & Sakaba, 2015). To study the mechanism underlying the target‐dependent opposite forms of short‐term plasticity, we next examined whether the facilitation of Ca2+ current and the resultant enhancement of transmitter release is a unique mechanism of PC synapses on another PC but not on DCN neurons. As shown in Fig. Fig.7,7, cultured PC axons preferentially formed high density of synapses around a type of neuron, which is estimated to be a potential DCN neuron based on morphology, electrophysiological properties and molecular marker expression (Kawaguchi & Sakaba, 2015). We performed paired whole‐cell recordings from the presynaptic PC axon terminal and the postsynaptic DCN cell. Similar to PC–PC synapses, paired‐pulse voltage commands of an AP waveform to the presynaptic terminal caused facilitation of presynaptic Ca2+ currents (3.33 ms interval, 114 ± 1%; 5 ms, 113 ± 2%; 10 ms, 114 ± 1%; 20 ms, 107 ± 2%; 50 ms, 103 ± 2%) (Fig. (Fig.77B and C). Consequently, the synaptic transmission also exhibited PPF rather than depression (3.33 ms, 177 ± 9%; 5 ms, 181 ± 10%; 10 ms, 182 ± 26%; 20 ms, 144 ± 9%; 50 ms, 123 ± 9%) (Fig. (Fig.77B and D). Interestingly, the extents of facilitation of Ca2+ current and that of PSC were similar to those observed at PC–PC synapses, and the relationship between facilitation of Ca2+ currents and that of PSCs also showed 4–5th power dependence. In addition, the Ca2+ current facilitation at PC synapses on DCN cells was shortened by increasing the EGTA concentration (3.33 ms interval, 111 ± 3%; 5 ms, 110 ± 2%; 10 ms, 108 ± 2%; 20 ms, 104 ± 2%; 50 ms, 100 ± 1%), as for PC–PC synapses (Fig. (Fig.77C). Consequently, the PPF of synaptic transmission was also shortened (3.33 ms, 164 ± 13%; 5 ms, 149 ± 21%; 10 ms, 133 ± 15%; 20 ms, 117 ± 11%; 50 ms, 91 ± 5%) (Fig. (Fig.77D). Thus, the Ca2+‐dependent facilitation of Ca2+ current causing short‐term facilitation is a common feature in PC axon terminals, irrespective of the target neuron type, as long as identical APs are elicited.
Ca2+‐dependent PPF of Ca2+ currents and PSCs at PC–DCN synapsesA, representative image for paired recordings from a presynaptic PC terminal and a postsynaptic DCN neuron. B, representative traces of the paired‐pulse AP command (Vcom) and the Ca2+ currents (ICa2+) in a presynaptic PC terminal (top) and the PSCs simultaneously recorded from a postsynaptic DCN neuron (bottom). Red trace is the averaged PSC from 20 PSC traces (grey). C and D, averaged PPR of ICa2+ (C) and that of PSC (D) in the presence of intracellular 0.5 or 5 mm of EGTA are plotted against the interstimulus interval. Black or grey lines in (D) are the 4.5th power of ICa2+ (shown as dotted lines, which is same as the continuous lines in C) (n = 5 and 6 for 5 and 0.5 mm EGTA, respectively).
In addition, the size of total readily releasable pool of synaptic vesicles was also similar when measured by a change of Cm. Ca2+ influx into the PC terminal caused by different durations (1–50 ms) of square depolarization pulses increased Cm reflecting exocytosis of synaptic vesicles (Fig. (Fig.8).8). The amplitude of Cm increase could be fitted by a single exponential curve. The PC–PC presynaptic terminal showed 83 ± 24 fF increase upon the 50 ms depolarization pulse, whereas the PC–DCN terminal showed 93 ± 27 fF (P > 0.39). Thus, assuming the Cm of single synaptic vesicle to be ∼70 aF (Kawaguchi & Sakaba, 2015), ∼1000 releasable synaptic vesicles are contained in each terminal. Consistent with the previous data (Kawaguchi & Sakaba, 2015), the Cm increase was not suppressed by increasing the intracellular EGTA concentration from 0.5 to 5 mm in PC terminals (74 ± 23 fF at terminals on PC and 95 ± 38 fF on DCN neuron, respectively) (Fig. (Fig.8),8), suggesting that transmitter release is tightly coupled to the Ca2+ influx.
Cm increases at a PC terminal on a PC or DCN neuronTop: representative traces of presynaptic ICa2+ and Cm recorded from a PC terminal on a PC (A) or that on a DCN neuron (B). Bottom: Cm increases recorded with intracellular 0.5 or 5 mm EGTA were plotted against the depolarization pulse duration [n = 7 (0.5 mm EGTA) and n = 13 (5 mm EGTA) terminals on a PC; n = 7 (0.5 mm EGTA) and n = 5 (5 mm EGTA) terminals on a DCN neuron].
Taken all of these results together, we conclude that the Ca2+‐dependent transmitter release mechanisms and the facilitative property relying on the Ca2+‐dependent facilitation of Ca2+ current are common in all PC axon terminals. However, because of AP attenuation, PC–DCN synapses exhibit depression rather than facilitation (Kawaguchi & Sakaba, 2015).
Discussion
By direct patch clamp recording from inhibitory presynaptic terminals of cultured cerebellar PCs, we have studied a mechanism of PPF at GABAergic synapses between PCs. This technique allowed us to reveal that (1) voltage‐gated Ca2+ current was facilitated in a Ca2+‐dependent manner in PC axon terminals and (2) short‐term facilitation of Ca2+ currents almost exclusively determines the short‐term facilitation of synaptic transmission between PCs. We conclude that PC–PC synapses show short‐term facilitation of transmitter release predominantly depending on the Ca2+‐dependent Ca2+ current facilitation. Our data further indicated that the synaptic facilitating property relying on the Ca2+ current facilitation was common at PC output synapses, irrespective of the target cells, when identical paired AP commands were applied to the terminal.
Ca2+ current facilitation
In the present study, we demonstrated that Ca2+ current in PC axon terminals is facilitated by ∼10% in a Ca2+‐dependent manner upon high‐frequency paired APs. Previous studies on PPF at PC–PC synapses in slice preparation could not detect facilitated increase in [Ca2+]i by fluorescence imaging (Orduz & Llano, 2007; Bornschein et al. 2013). Considering the tiny fluorescence change by single AP‐triggered Ca2+ influx as a result of the strong Ca2+ buffering (Fig. (Fig.6),6), it appears to be hard to detect a small facilitation of Ca2+ influx of ∼10% by Ca2+ imaging. Similar Ca2+‐dependent facilitation of Ca2+ currents has been reported at glutamatergic presynaptic terminal in the calyx of Held (Forsythe et al. 1998; Borst & Sakmann, 1998; Cuttle et al. 1998), which partly contributes to short‐term facilitation of synaptic transmission when the basal release probability is lowered (Felmy et al. 2003; Müller et al. 2008; Hori & Takahashi, 2009).
CNS neurons have several types of voltage‐gated Ca2+ channels, such as N, L, P/Q and R‐types. Among these, mature PCs express P/Q type of Ca2+ channels abundantly, and its genetic ablation results in severe ataxia (Jun et al. 1999) and reduced synaptic facilitation (Inchauspe et al. 2004). The detailed mechanism of Ca2+‐dependent Ca2+ current facilitation has been studied using a heterologous expression system (Lee et al. 2003; DeMaria et al. 2001; Catterall & Few, 2008; Ben‐Johny & Yue, 2014). Similar to other types of Ca2+ channels, the P/Q type has a CaM binding site. The Ca2+ binding to C‐lobe of CaM is reported to rapidly (<1 ms) facilitate the Ca2+ currents by increasing the channel open probability. On the other hand, CaM bound with Ca2+ at N‐lobe has been shown to inactivate the P/Q type Ca2+ channels relatively slowly (tens of milliseconds). In addition, neuronal Ca2+ sensor protein was also suggested to mediate the Ca2+‐dependent facilitation of Ca2+ currents in a similar manner to CaM but with stronger Ca2+ affinity (Tsujimoto et al. 2002). If the Ca2+ current is facilitated by a mechanism coupled loosely with the Ca2+ influx, an increase of exogenous buffering by EGTA is expected to weaken the Ca2+ current facilitation (Neher, 1998). However, as shown in Figs Figs33 and and7,7, a high concentration of EGTA accelerated the recovery but did not change the peak amplitude of Ca2+ current facilitation in PC terminals (Alturi & Regehr, 1998). Thus, the Ca2+ current facilitation appears to be tightly coupled with local Ca2+ influx, which is consistent with the idea that Ca2+‐binding to CaM associated with the P/Q‐type Ca2+ channels at rest contributes to Ca2+ current facilitation (Erickson et al. 2001; Ben‐Johny & Yue, 2014).
Interestingly, the Ca2+ current facilitation by Ca2+/CaM is specific to P/Q‐type Ca2+ channels, and other types of Ca2+ channels are rather inactivated by Ca2+/CaM association (Catterall & Few, 2008; Ben‐Johny & Yue, 2014). The types of Ca2+ channels contributing to transmitter release in presynaptic terminals developmentally change at some synapses, including PC output synapses, excitatory synapses in the calyx of Held and thalamic inhibitory synapses (Iwasaki et al. 2000; Miki et al. 2013). Thus, short‐term plasticity by Ca2+ channel modulation might also developmentally change. Indeed, in contrast to PPF at synapses between PCs in slices prepared from postnatal day 7–19 mice (Orduz & Llano, 2007; Bornschein et al. 2013), PPD was observed in younger mice (postnatal day 4–6 mice) (Watt et al. 2009). A developmental switch of Ca2+ channels in the PC terminal from mixture of N‐, R‐ and P/Q‐type to P/Q‐type‐specific (Iwasaki et al. 2000) might be responsible for the age‐dependent differences in short‐term plasticity.
Mechanisms of PPF
The averaged PSC amplitude (409 pA) at a PC pair with approximately six synaptic contacts and the miniature PSC amplitude (52 pA) imply that one or two synaptic vesicles are exocytosed at each presynaptic terminal. This estimation is in accordance with the number of exocytosed vesicles (∼2) based on the PSC amplitude (116 pA) upon an AP command at single terminal. Thus, the release probability per synapse is almost 1 in PC terminals. On the other hand, the Cm measurement suggests that ∼1000 vesicles are readily releasable in a single terminal within 10 ms if a strong Ca2+ increase takes place (Fig. (Fig.8).8). Thus, the release probability per each RRP vesicle is estimated to be very low (∼0.2%). It should be noted that the number of synaptic vesicles exocytosed upon a square depolarization pulse might be larger than the releasable vesicle pool upon high frequency APs. In addition, if the vesicle replenishment is very fast (i.e. several milliseconds) (Valera et al. 2012; Brachtendorf et al. 2015), the capacitance measurement might overestimate the RRP size by 2‐ to 3‐fold. In any cases, a PC axon terminal has a large amount of RRP vesicles and the release probability of each release‐ready vesicle upon an AP is low. The large RRP and low release probability are typical properties of facilitating synapses (Dittman et al. 2000) and the increase of either RRP size or release probability leads to PPF.
Paired recordings from a PC axon terminal and a postsynaptic PC soma demonstrated that PPF is almost exclusively mediated by the Ca2+‐dependent facilitation of Ca2+ current into a presynaptic terminal (Figs (Figs44 and and5).5). Using the expression of mutant Ca2+ channels in cultured superior cervical ganglion neurons, a tight correlation between Ca2+ current modulation and short‐term plasticity was demonstrated (Mochida et al. 2008). When the release probability is decreased to suppress depletion of the releasable synaptic vesicle pool, the calyx of Held synapses also show short‐term facilitation, ∼50% of which relies on the Ca2+‐dependent Ca2+ current facilitation (Felmy et al. 2003; Müller et al. 2008; Hori & Takahashi, 2009). Thus, Ca2+‐dependent Ca2+ current facilitation may contribute to synaptic facilitation at many synapses. However, our data do not necessarily exclude other mechanisms suggested in PPF at other synapses. The residual Ca2+ hypothesis is the simplest one arising from studies on short‐term facilitation at the neuromuscular junction (Katz & Miledi, 1968). It was postulated that the summation of residual Ca2+ remaining in the cytoplasm after the first AP and the local Ca2+ during the following AP contributes to the facilitated transmitter release as a result of 4th power dependence of transmitter release on [Ca2+]i. However, in most cases, the residual [Ca2+]i increase is estimated to be too small compared to the local [Ca2+]i to account for facilitation. This is particularly the case for PC axon terminals (Fig. (Fig.6)6) (Orduz & Llano, 2007), presumably as a result of very high expression of high affinity Ca2+ buffer calbindin. We estimated that the residual [Ca2+]i increase accounts for only 1–2% facilitation at PC–PC synapses.
Saturation of mobile Ca2+ buffer proteins in presynaptic terminals has also been considered as a candidate mechanism for short‐term facilitation. If Ca2+ entering into the terminal upon the first AP remains bound to the large part of Ca2+ buffer molecules, free [Ca2+]i is expected to become higher upon the following APs, leading to larger transmitter release. Indeed, calbindin was shown to contribute to the PPF at inhibitory synapses in the cerebral cortex and excitatory mossy fibre–CA3 synapses in the hippocampus (Blatow et al. 2003). On the other hand, despite extensive expression of calbindin in cerebellar PCs throughout the cell, it was recently demonstrated that PPF at PC–PC synapses was not affected by the genetic ablation of either calbindin or parvalbumin, another Ca2+ buffer protein (Bornschein et al. 2013). The lack of effect by calbindin ablation on PPF might be a result of the limited amount of Ca2+ influx into the presynaptic terminal upon a single AP (half‐width of ∼0.4 ms) (Fig. (Fig.3),3), which might be far from saturating the abundant calbindin. Also, the buffer saturation model requires loose coupling between Ca2+ channels and vesicles (Neher 1998; Rozov et al. 2001; Vyleta and Jonas, 2014), which is not the case at this (Fig. (Fig.8)8) and other inhibitory synapses (Bucurenciu et al. 2008).
Another hypothesis is that Ca2+ causes facilitation acting at presynaptic release sensors (Atluri & Regehr, 1998; Bertram et al. 1996; Regehr, 2012). If there is a Ca2+‐binding site for release with high Ca2+ affinity and slow kinetics, the site is expected to be occupied with Ca2+ during the small increase of residual Ca2+, resulting in facilitated release upon the following stimulation. However, the identity of the molecule playing this role remains obscure. Using a kinetic simulation based on the five‐site Ca2+ sensor model (Schneggenburger & Neher, 2000), the slow and high‐affinity Ca2+ sensor for vesicular release was suggested to underlie the PPF observed at PC–PC synapses (Bornschein et al. 2013). By contrast, we now show that the PPF of PSCs is suppressed when the amplitude of Ca2+ current upon the second AP is identical to that upon the first (Fig. (Fig.5),5), indicating that the first Ca2+ influx exerts little positive effect on the release machinery remaining at the time of next AP arrival. Alternatively, we demonstrate that the Ca2+‐dependent Ca2+ current facilitation plays a role as a ‘Ca2+ sensor’ for the PPF at PC–PC synapses.
Activity‐dependent recruitment of an extra pool or reluctant vesicles to the RRP is also a possible mechanism for PPF (Valera et al. 2012; Brachtendorf et al. 2015). However, our Cm measurement data showing the uniform releasable vesicle pool in PC terminals imply that RRP size change may not be implicated in the PPF at PC–PC synapses.
Target‐dependent plasticity and physiological implication
We demonstrated that the Ca2+‐dependent facilitation of Ca2+ currents and the resultant synaptic short‐term facilitation are a common property of PC terminals irrespective of their target neurons (i.e. another PC or DCN cells) when the amplitudes of APs are identical (using AP commands). In accordance with a previous study on PC terminals on a DCN neuron (Kawaguchi & Sakaba, 2015), PC terminals on another PC also exhibited the large readily releasable synaptic vesicle pool (∼1000 vesicles) and the low release probability (<1%) upon an AP arrival, which are typical properties of facilitating synapses (Dittman et al. 2000). However, PC–PC synapses exhibit short‐term facilitation (Orduz & Llano, 2007; Bornschein et al. 2013), whereas PC–DCN synapses undergo short‐term depression in situ (Telgkamp & Raman, 2002). These apparently opposite directions of plasticity dependent on the target neurons can simply be explained by the difference in AP conduction fidelity. Direct recordings from an axon and/or a terminal previously demonstrated that AP amplitudes attenuated around axon terminals on DCN neurons located far from the PC soma in both culture and slice preparation (Kawaguchi & Sakaba, 2015). On the other hand, previous studies have shown that axonal AP conduction from the soma to the proximal region of axon is reliable up to 200 Hz (Khaliq & Raman, 2005; Monsivais et al. 2005). Consistently, our preliminary data suggest that an AP train at 100 Hz faithfully propagates from the soma to the proximal terminal on a PC with little attenuation of amplitude (Takeshi Sakaba and Shin‐ya Kawaguchi, unpublished observation). Consequently, the PC–PC synapses exhibit short‐term facilitation through the facilitated Ca2+ current. On the other hand, taking the 4th power dependence of synaptic transmission on the Ca2+ current (Fig. (Fig.2)2) into consideration, a small decrease of Ca2+ influx caused by a slight AP amplitude change has a large impact on synaptic efficacy, converting the intrinsically facilitating synapses to depressing synapses at PC–DCN synapses.
Target‐dependent short‐term synaptic plasticity has been reported at various synapses, such as those in neocortex, cerebellum and hippocampus (Markram et al. 1998; Rozov et al. 2001; Koester & Johnston, 2005; Pelkey et al. 2006; Beierlein et al. 2007; Bao et al. 2010). For example, excitatory synapses from a cortical pyramidal neuron on basket cells exhibit short‐term depression, whereas those on Martinotti cells show facilitation (Markram et al. 1998). It is generally assumed that synapses usually contain both the mechanisms for facilitation and those for depression, and transmitter release probability is considered to determine which of the two mechanisms prevails. Fluorescence imaging techniques have shown that the facilitating presynaptic varicosities tend to show a small Ca2+ increase and low synaptic release probability, whereas the depressing ones show larger Ca2+ transients and higher release probability (Koester & Johnston, 2005). Considering the large impact of AP amplitude on the presynaptic Ca2+ influx and the resultant synaptic transmission, we suggest that slight differences of membrane excitability in presynaptic terminals may also play a role in the target‐dependent plasticity. To test this idea at various synapses in the future, direct recording from the terminals and/or development of voltage‐sensitive dyes are essential (Hoppa et al. 2014; Vyleta & Jonas, 2014).
Additional Information
Competing interests
The authors declare that they have no competing interests.
Author contributions
SK and TS designed the study. FD and SK performed the experiments. FD, TS and SK interpreted the data and wrote the manuscript. All authors approved the final version of the manuscript submitted for publication.
Funding
This work was supported by grants from the JSPS/MEXT, Japan (KAKENHI grant numbers 15K06722 and 15KT0082 to SK; 15H04261, 15K14321, 26110720 and Core‐to‐Core Program A Advanced Research Networks to TS) and the Naito Science Foundation to SK.
Acknowledgements
We are grateful to Dr Tomoyuki Takahashi and Dr Mitsuharu Midorikawa for critically reading the manuscript and for making helpful comments.
References
- Abbott LF & Regehr WG (2004). Synaptic computation. Nature431, 796–803. [PubMed] [Google Scholar]
- Atluri PP & Regehr WG (1996). Determinants of the time course of facilitation at the granule cell to Purkinje cell synapse. J Neurosci16, 5661–5671. [PMC free article] [PubMed] [Google Scholar]
- Atluri PP & Regehr WG (1998). Delayed release of neurotransmitter from cerebellar granule cells. J Neurosci18, 8214–8227. [PubMed] [Google Scholar]
- Bao J, Reim K & Sakaba T (2010). Target‐dependent feedforward inhibition mediated by short‐term synaptic plasticity in the cerebellum. J Neurosci30, 8171–8179. [PubMed] [Google Scholar]
- Bastianelli E (2003). Distribution of calcium‐binding proteins in the cerebellum. Cerebellum2, 242–262. [PubMed] [Google Scholar]
- Ben‐Johny M & Yue DT (2014). Calmodulin regulation (calmodulation) of voltage‐gated calcium channels. J Gen Physiol143, 679–692. [PMC free article] [PubMed] [Google Scholar]
- Beierlein M, Fioravante D & Regehr WG (2007). Differential expression of posttetanic potentiation and retrograde signaling mediate target‐dependent short‐term synaptic plasticity. Neuron54, 949–959. [PMC free article] [PubMed] [Google Scholar]
- Bertram R, Sherman A & Stanley EF (1996). Single‐domain/bound calcium hypothesis of transmitter release and facilitation. J Neurophysiol75, 1919–1931. [PubMed] [Google Scholar]
- Bornschein G, Arendt O, Hallermann S, Brachtendorf S, Eilers J & Schmidt H (2013). Paired‐pulse facilitation at recurrent Purkinje neuron synapses is independent of calbindin and parvalbumin during high‐frequency activation. J Physiol591, 3355–3370. [PMC free article] [PubMed] [Google Scholar]
- Borst JG & Sakmann B (1998). Facilitation of presynaptic calcium currents in the rat brainstem. J Physiol513, 149–155. [PMC free article] [PubMed] [Google Scholar]
- Blatow M, Caputi A, Burnashev N, Monyer H & Rozov A (2003). Ca2+ buffer saturation underlies paired pulse facilitation in calbindin‐D28k‐containing terminals. Neuron38, 79–88. [PubMed] [Google Scholar]
- Branchaw JL, Banks MI & Jackson MB (1997). Ca2+‐ and voltage‐dependent inactivation of Ca2+ channels in nerve terminals of the neurohypophysis. J Neurosci17, 5772–5781. [PMC free article] [PubMed] [Google Scholar]
- Brachtendorf S, Eilers J & Schmidt H (2015). A use‐dependent increase in release sites drives facilitation at calretinin‐deficient cerebellar parallel‐fiber synapses. Front Cell Neurosci. 9, 27. [PMC free article] [PubMed] [Google Scholar]
- Bucurenciu I, Kulik A, Schwaller B, Frotscher M & Jonas P (2008). Nanodomain coupling between Ca2+ channels and Ca2+ sensors promotes fast and efficient transmitter release at a cortical GABAergic synapse. Neuron57, 536–545. [PubMed] [Google Scholar]
- Catterall WA & Few AP (2008). Calcium channel regulation and presynaptic plasticity. Neuron59, 882–901. [PubMed] [Google Scholar]
- Cuttle MF, Tsujimoto T, Forsythe ID & Takahashi T (1998). Facilitation of the presynaptic calcium current at an auditory synapse in rat brainstem. J Physiol512, 723–729. [PMC free article] [PubMed] [Google Scholar]
- DeMaria CD, Soong TW, Alseikhan BA, Alvania RS & Yue DT (2001). Calmodulin bifurcates the local Ca2+ signal that modulates P/Q‐type Ca2+ channels. Nature411, 484–489. [PubMed] [Google Scholar]
- Dittman JS, Kreitzer AC & Regehr WG (2000). Interplay between facilitation, depression, and residual calcium at three presynaptic terminals. J Neurosci20, 1374–1385. [PubMed] [Google Scholar]
- Erickson MG, Alseikhan BA, Peterson BZ & Yue DT (2001). Preassociation of calmodulin with voltage‐gated Ca(2+) channels revealed by FRET in single living cells. Neuron31, 973–985. [PubMed] [Google Scholar]
- Felmy F, Neher E & Schneggenburger R (2003). Probing the intracellular calcium sensitivity of transmitter release during synaptic facilitation. Neuron37, 801–811. [PubMed] [Google Scholar]
- Fioravante D & Regehr WG (2011). Short‐term forms of presynaptic plasticity. CurrOpinNeurobiol21, 269–274. [PMC free article] [PubMed] [Google Scholar]
- Forsythe ID, Tsujimoto T, Barnes‐Davies M, Cuttle MF & Takahashi T (1998). Inactivation of presynaptic calcium current contributes to synaptic depression at a fast central synapse. Neuron20, 797–807. [PubMed] [Google Scholar]
- Geiger JR & Jonas P (2000). Dynamic control of presynaptic Ca2+ inflow by fast‐inactivating K(+) channels in hippocampal mossy fiber boutons. Neuron28, 927–939. [PubMed] [Google Scholar]
- Hoppa MB, Gouzer G, Armbruster M & Ryan TA (2014). Control and plasticity of the presynaptic action potential waveform at small CNS nerve terminals. Neuron84, 778–789. [PMC free article] [PubMed] [Google Scholar]
- Hori T & Takahashi T (2009). Mechanisms underlying short‐term modulation of transmitter release by presynaptic depolarization. J Physiol587, 2987–3000. [PMC free article] [PubMed] [Google Scholar]
- Inchauspe CG, Martini FJ, Forsythe ID & Uchitel OD (2004). Functional compensation of P/Q by N‐type channels blocks short‐term plasticity at the calyx of held presynaptic terminal. J Neurosci24, 10379–10383. [PubMed] [Google Scholar]
- Iwasaki S, Momiyama A, Uchitel OD & Takahashi T (2000). Developmental changes in calcium channel types mediating central synaptic transmission. J Neurosci20, 59–65. [PubMed] [Google Scholar]
- Jackson MB, Konnerth A & Augustine GJ (1991). Action potential broadening and frequency‐dependent facilitation of calcium signals in pituitary nerve terminals. Proc Natl Acad Sci USA. 88, 380–384. [PMC free article] [PubMed] [Google Scholar]
- Jun K, Piedras‐Rentería ES, Smith SM, Wheeler DB, Lee SB, Lee TG, Chin H, Adams ME, Scheller RH, Tsien RW & Shin HS (1999). Ablation of P/Q‐type Ca(2+) channel currents, altered synaptic transmission, and progressive ataxia in mice lacking the alpha(1A)‐subunit. Proc Natl Acad Sci USA96, 15245–15250. [PMC free article] [PubMed] [Google Scholar]
- Kamiya H & Zucker RS (1994). Residual Ca2+ and short‐term synaptic plasticity. Nature371, 603–606. [PubMed] [Google Scholar]
- Kaneko M, Yamaguchi K, Eiraku M, Sato M, Takata N, Kiyohara Y, Mishina M, Hirase H, Hashikawa T & Kengaku M (2011). Remodelling of monoplaner Purkinje cell dendrites during cerebellar circuit formation. Plos ONEe20108. [PMC free article] [PubMed] [Google Scholar]
- Kawaguchi S & Hirano T (2007). Sustained structural change of GABAA receptor‐associated protein underlies long‐term potentiation at inhibitory synapses on a cerebellar Purkinje neuron. J Neurosci27, 6788–6799. [PubMed] [Google Scholar]
- Kawaguchi S & Sakaba T (2015). Control of inhibitory synaptic outputs by low excitability of axon terminals revealed by direct recording. Neuron85,1273‐1288. [PubMed] [Google Scholar]
- Katz B & Miledi R (1968). The role of calcium in neuromuscular facilitation. J Physiol195, 481–492. [PMC free article] [PubMed] [Google Scholar]
- Khaliq ZM6Raman IM (2005). Axonal propagation of simple and complex spikes in cerebellar Purkinje neurons. J Neurosci25, 454–463. [PubMed] [Google Scholar]
- Koester HJ & Johnston D (2005). Target cell‐dependent normalization of transmitter release at neocortical synapses. Science308, 863–866. [PubMed] [Google Scholar]
- Lee A, Zhou H, Scheuer T & Catterall WA (2003). Molecular determinants of Ca(2+)/calmodulin‐dependent regulation of Ca(v)2.1 channels. Proc Natl Acad Sci USA100, 16059–16064. [PMC free article] [PubMed] [Google Scholar]
- Markram H, Wang Y & Tsodyks M. (1998). Differential signaling via the same axon of neocortical pyramidal neurons. Proc Natl Acad Sci USA95, 5323–5328. [PMC free article] [PubMed] [Google Scholar]
- Matveev V, Zucker RS & Sherman A (2004). Facilitation through buffer saturation: constraints on endogenous buffering properties. Biophys J86, 2691–2709. [PMC free article] [PubMed] [Google Scholar]
- Miki T, Hirai H & Takahashi T (2013). Activity‐dependent neurotrophin signaling underlies developmental switch of Ca2+ channel subtypes mediating neurotransmitter release. J Neurosci33, 18755–18763. [PubMed] [Google Scholar]
- Mochida S, Few AP, Scheuer T & Catterall WA (2008). Regulation of presynaptic Ca(V)2.1 channels by Ca2+ sensor proteins mediates short‐term synaptic plasticity. Neuron57, 210–216. [PubMed] [Google Scholar]
- Monsivais P, Clark BA, Roth A & Häusser M (2005). Determinants of action potential propagation in cerebellar Purkinje cell axons. J Neurosci25, 464–472. [PubMed] [Google Scholar]
- Müller M, Felmy F & Schneggenburger R (2008). A limited contribution of Ca2+ current facilitation to paired‐pulse facilitation of transmitter release at the rat calyx of Held. J Physiol586, 5503–5520. [PMC free article] [PubMed] [Google Scholar]
- Neher E (1998). Vesicle pools and Ca2+ microdomains: new tools for understanding their roles in neurotransmitter release. Neuron20, 389–399. [PubMed] [Google Scholar]
- Neher E & Marty A (1982). Discrete changes of cell membrane capacitance observed under conditions of enhanced secretion in bovine adrenal chromaffin cells. Proc Natl Acad Sci USA79, 6712–6716. [PMC free article] [PubMed] [Google Scholar]
- Orduz D & Llano I (2007). Recurrent axon collaterals underlie facilitating synapses between cerebellar Purkinje cells. Proc Natl Acad Sci USA104, 17831–17876. [PMC free article] [PubMed] [Google Scholar]
- Pedroarena CM & Schwarz C (2003). Efficacy and short‐term plasticity at GABAergic synapses between Purkinje and deep cerebellar nuclei neurons. J Neurophys89, 704–716. [PubMed] [Google Scholar]
- Pelkey KA, Topolnik L, Lacaille JC & McBain CJ (2006). Compartmentalized Ca2+ channel regulation at divergent mossy‐fiber release sites underlies target cell‐dependent plasticity. Neuron52, 497–510. [PubMed] [Google Scholar]
- Perkel DJ, Hestrin S, Sah P & Nicoll RA (1990). Excitatory synaptic currents in Purkinje cells. Proc Biol Sci. 241, 116–121. [PubMed] [Google Scholar]
- Regehr WG (2012). Short‐term synaptic plasticity. Cold Spring Harb Perspect Biol4, a005702. [PMC free article] [PubMed] [Google Scholar]
- Rozov A, Burnashev N, Sakmann B & Neher E (2001). Transmitter release modulation by intracellular Ca2+ buffers in facilitating and depressing nerve terminals of pyramidal cells in layer 2/3 of the rat neocortex indicates a target cell‐specific difference in presynaptic calcium dynamics. J Physiol531, 807–826. [PMC free article] [PubMed] [Google Scholar]
- Schmidt H, Stiefel KM, Racay P, Schwaller B & Eilers J (2003). Mutational analysis of dendritic Ca2+ kinetics in rodent Purkinje cells: role of parvalbumin and calbindin D28k. J Physiol551, 13–32. [PMC free article] [PubMed] [Google Scholar]
- Schneggenburger R & Neher E (2000). Intracellular calcium dependence of transmitter release rates at a fast central synapse. Nature406, 889–893. [PubMed] [Google Scholar]
- Telgkamp P & Raman IM (2002). Depression of inhibitory synaptic transmission between Purkinje cells and neurons of the cerebellar nuclei. J Neurosci22, 8447–8457. [PubMed] [Google Scholar]
- Tsujimoto T, Jeromin A, Saitoh N, Roder JC & Takahashi T (2002). Neuronal calcium sensor 1 and activity‐dependent facilitation of P/Q‐type calcium currents at presynaptic nerve terminals. Science295, 2276–2279. [PubMed] [Google Scholar]
- Valera AM, Doussau F, Poulain B, Barbour B & Isope P (2012). Adaptation of granule cell to Purkinje cell synapses to high‐frequency transmission. J Neurosci32, 3267–3280. [PubMed] [Google Scholar]
- Vyleta NP & Jonas P (2014). Loose coupling between Ca2+ channels and release sensors at a plastic hippocampal synapse. Science343, 665–670. [PubMed] [Google Scholar]
- Watt AJ, Cuntz H, Mori M, Nusser Z, Sjöström PJ & Häusser M (2009). Traveling waves in developing cerebellar cortex mediated by asymmetrical Purkinje cell connectivity. Nat Neurosci12, 463–473. [PMC free article] [PubMed] [Google Scholar]
- Zucker RS & Regehr WG (2002). Short‐term synaptic plasticity. Ann Rev Physiol64, 355–405. [PubMed] [Google Scholar]
Articles from The Journal of Physiology are provided here courtesy of The Physiological Society
Published online 2002 Aug 16. doi: 10.1113/jphysiol.2002.026609
PMID: 12381819
This article has been cited by other articles in PMC.
Abstract
Paired recordings between CA3 interconnected pyramidal neurons were used to study the properties of short-term depression occurring in these synapses under different frequencies of presynaptic firing (n = 22). In stationary conditions (0.05–0.067 Hz) pairs of presynaptic action potentials (50 ms apart) evoked EPSCs whose amplitude fluctuated from trial to trial with occasional response failures. In 15/20 cells, paired-pulse ratio (PPR) was characterized by facilitation (PPF) while in the remaining five by depression (PPD). Increasing stimulation frequency from 0.05–0.067 Hz to 0.1–1 Hz induced low frequency depression (LFD) of EPSC amplitude with a gradual increase in the failure rate. Overall, 9/12 cells at 1 Hz became almost ‘silent’. In six cells in which the firing rate was sequentially shifted from 0.05 to 0.1 and 1 Hz, changes in synaptic efficacy were so strong that PPR shifted from PPF to PPD. The time course of depression of EPSC1 could be fitted with single exponentials with time constants of 98 and 36 s at 0.1 and 1 Hz, respectively. In line with the inversion of PPR at 1 Hz, the time course of depression of EPSC2 was faster than EPSC1 (7 s). Recovery from depression could be obtained by lowering the frequency of stimulation to 0.025 Hz. These results could be explained by a model that takes into account two distinct release processes, one dependent on the residual calcium and the other on the size of the readily releasable pool of vesicles.
Short-term forms of synaptic plasticity are crucial for regulating the temporal code and information processing between neurons in a network (). These vary from synapse to synapse and in the same synapse according to its previous history (; ; ; ). One common form of short-term plasticity lasting from seconds to minutes is depression upon repeated use (; ). This may provide a dynamic gain control over a variety of presynaptic afferent firing action potentials at different rates (; ; ). This form of plasticity may reflect mainly presynaptic depletion of a readily releasable pool of vesicles (; ; ; ). However, other mechanisms, such as presynaptic calcium channel inactivation (Gingrich & Byrne, 1985), negative feedback through inhibitory autoreceptors () or receptor desensitization (; ) cannot be ruled out.
In most cases, synaptic depression should be dependent on previous release () in such a way that, in a paired-pulse protocol, the ratio between the mean amplitudes of the second EPSC over the first EPSC (paried-pulse ratio, PPR) is inversely related to the initial release probability (). In general, the smaller is the probability of release to the first pulse, the more facilitated is the response to the second pulse. This phenomenon, known as paired-pulse facilitation (PPF), is accounted for by the residual calcium hypothesis, according to which the small fraction of calcium entering the terminal during the first spike increases the probability of transmitter release to a second action potential (). It follows that if repetitive stimulation of presynaptic neurons causes a reduction in the release probability and, therefore, in the amplitude of the first EPSC (), it should induce a further facilitation of the second EPSC, leading to an increase in PPR. The present experiments were undertaken to see whether the relation between PPR and probability of release was still maintained during use-dependent depression.
In the hippocampus, synaptic depression has been studied mainly with minimal stimulation of afferent inputs (; ; ). With this method it is impossible to ascertain that the same presynaptic axon is activated trial after trial as different axons are likely to be stimulated. Moreover failures of presynaptic release cannot be unambiguously distinguished from failures of the stimulus to trigger an action potential in the presynaptic fibre. To overcome these problems we used whole-cell recordings from pairs of interconnected CA3–CA3 pyramidal neurons in hippocampal slice cultures (; ). We found that in the majority of neurons low frequency depression (LFD) was associated with a reduction in EPSC amplitude. The magnitude of LFD varied between pairs and depended on the rate of presynaptic firing. Unexpectedly PPR was reduced at higher presynaptic firing frequency (1 Hz) and, on average, PPF was converted into PPD suggesting that during LFD release probability results from the balance of at least two distinct processes, one depending on the residual calcium and the other on the number of available vesicles.
METHODS
Organotypic cultures
Hippocampi were removed from 4- to 7-day-old rats killed by decapitation and organotypic cultures were prepared following the method already described (). The procedure was performed in accordance with the regulations of the Italian Animal Welfare Act and was approved by the local authority veterinary service. Transverse 400 μm thick slices were cut with a tissue chopper and attached to coverslips in a film of reconstituted chicken plasma (Cocalico, Reamstown, PA, USA) clotted with thrombin (Sigma, Milan, Italy). The coverslips were transferred to plastic tubes containing 0.75 ml of medium. The tubes were placed in a roller drum (6 revolutions h−1) inside an incubator at 36 °C. The medium contained: basal medium (BME, Eagle, with Hanks' salts, without l-glutamine; Gibco, 100 ml), Hanks' balanced salt solution (HBSS; Gibco, 50 ml), horse serum (Gibco, 50 ml), l-glutamine (Gibco, 200 mm, 1 ml), 50 % d-glucose in sterile water for tissue culture (Gibco, 2 ml).
Whole-cell patch-clamp recordings
After 10-14 days incubation, slices, which had flattened near-monolayer thickness, were transferred to a recording chamber fixed to the stage of an upright microscope. Cultured slices in the recording chamber were superfused at room temperature (22-24 °C) with a bath solution containing (mm): 150 NaCl, 3 KCl, 2 CaCl2, 1 MgCl2, 10 Hepes, 10 glucose (adjusted to pH 7.3 with NaOH). Electrophysiological experiments were performed on CA3 pyramidal neurons using the whole-cell configuration of the patch-clamp technique in current (presynaptic cell) and voltage clamp mode (postsynaptic cell). Patch electrodes were pulled from borosilicate glass capillaries (Hingenberg, Malsfeld, D). They had a resistance of 3-6 MΩ when filled with an intracellular solution containing (mm): 135 KMeSO4, 10 KCl, 10 Hepes, 1 MgCl2, 2 Na2ATP, 0.4 Na2GTP for the presynaptic cell and 135 CsMeSO4, 10 CsCl, 10 Hepes, 5 QX 314, 0.5 EGTA, 2 MgATP, 0.3 NaGTP for the postsynaptic cell; pH was adjusted to 7.3 with KOH and CsOH, respectively. Pairs of presynaptic action potentials (50 ms interval) were evoked at resting membrane potential (−58.1 ± 0.73 mV) by short (5 ms) depolarizing current pulses at different frequency (from 0.025 to 1 Hz). AMPA-mediated EPSCs were recorded at a holding potential of −60 mV. A liquid junction potential of 9 mV was not corrected. In some experiments (n = 4) the NMDA receptor antagonist 2-amino-5-phosphopentanoic acid (AP5, obtained from Tocris Cookson Ltd, Bristol, UK) was added to the perfusion solution to block NMDA receptors.
Data were sampled at a rate of 20 kHz and filtered with a cut off frequency of 1 kHz. Series resistance compensation was used only for the presynaptic cell.
Data acquisition and analysis
EPSCs were stored on a magnetic tape and transferred to a computer after digitization with an A/D converter (Digidata 1200). Data acquisition was done using pCLAMP (Axon Instruments, Union City, CA, USA). EPSCs were analysed with AxoGraph 4.6 (Axon Instruments), which uses a detection algorithm based on a sliding template. The onset of the EPSC was given by the intersection of a line through the 20 and 80 % of EPSC rise time with the baseline. The same program was used to fit the decay phase of the EPSCs with a monoexponential function. Failures (N0) were identified visually. In a set of experiments (n = 12), in order to control the adequacy of the visual selection, the fraction of responses with amplitude > 0 pA, corresponding to failures, was calculated. Since failures should be symmetrically distributed around zero, N0 was calculated by doubling this fraction (). A similarity and a high correlation between N0 estimated by the two methods were found (mean failure rates were 84.9 ± 4.5 and 81.5 ± 6.0 %, respectively, r = 0.95; P < 0.001).
For any given frequency, mean EPSC amplitude was obtained by averaging successes and failures.
The time course of synaptic depression and recovery could be fitted with a monoexponential function (using a Levenberg-Marquardt algorithm implemented in SigmaPlot 2001, Jandel, Germany). Data points were obtained by averaging for any given frequency EPSC amplitudes over four to eight consecutive trials and normalizing them to the first value in the train (I/I1st, in the case of depression) or to the mean EPSC amplitude obtained in control conditions (I/Icontrol, in the case of recovery from depression).
Values are given as means ± s.e.m. Significance of differences was assessed using Student's t test or the Wilcoxon signed rank test. The differences were considered significant when P was < 0.05.
RESULTS
Stable long-lasting (>30 min) whole-cell recordings were obtained from 22 CA3 pyramidal cell pairs. Interconnected neurons were identified as pyramidal cells both visually and on the basis of their firing properties, i.e. their ability to accommodate in response to long (800 ms) depolarizing current pulses. The identity of the connected cells as pyramidal neurons was confirmed in three experiments in which cell pairs were labelled with neurobiotin. In control conditions, pairs of presynaptic action potentials (50 ms apart), delivered at the frequency of 0.05–0.067 Hz, evoked two sequential EPSCs that fluctuated in amplitude from trials to trials with occasional response failures (Fig. 1). Paired-pulse modulation was quantified by calculating the paired-pulse ratio (PPR) as the ratio between the mean amplitude of the second and the first EPSC. In agreement with previous reports (; ) a strong heterogeneity in PPR across different connections was observed. The majority of cells (15 of 20) exhibited PPF (1.4 ± 0.1, see Fig. 1), while the remaining five exhibited PPD (0.7 ± 0.1).
Low frequency depression induced in a synaptically connected pair of CA3 neurons by increasing the presynaptic firing frequency from 0.05 to 1 HzA and B, ten postsynaptic responses (B) induced by pairs of depolarization-evoked action potentials (50 ms apart) in the presynaptic cell (A) are superimposed for each frequency. The resting potential of both neurons was −60 mV. C, all EPSCs evoked at a given frequency (including failures) are averaged. D, time course of EPSC amplitude to the first and second pulse for the same neuron shown in A-C.
Properties of low frequency depression
Increasing frequency of presynaptic firing from 0.05–0.067 to 0.1–1 Hz induced depression of postsynaptic responses (n = 14). The degree of synaptic depression was heterogeneous across different pairs but was clearly dependent on the testing frequency, being larger at higher frequencies. As shown in the example of Fig. 1, consecutive changes in presynaptic firing frequency from 0.05, to 0.1 and 1 Hz induced a gradual reduction in the number of successes, leading only to transmission failures. Overall, in 9 of 12 cells at the end of a 1 Hz low frequency train, synaptic transmission was almost completely abolished (Fig. 1D and Fig. 2C). In the remaining three cells 1 Hz stimulation induced a reduction of the mean amplitude of synaptic responses of 60 ± 10 and 70 ± 10 %, for EPSC1 and EPSC2, respectively, although only a few failures were observed. Successful transmission reappeared in all cells examined (n = 8) after switching to a lower frequency (0.025 Hz).
Short-term depression in a synapse with high release probabilityA, superimpositions of ten EPSC pairs evoked at different frequencies by presynaptic action potentials. B, all EPSCs evoked at a given frequency (including failures) are averaged (scale bar is the same for A and B). C, time course of EPSC amplitude to the first (filled circles) and second pulse (open circles) for the same neuron shown in A and B.
Interaction between paired-pulse modulation and low frequency depression
In the example of Fig. 1, increasing the stimulation frequency from 0.05 to 0.1 Hz produced a reduction in the amplitude of both EPSC1 and EPSC2. This was associated with an enhanced number of failures and a slight decrease in PPR. The figure shows also that a further increase of stimulation frequency to 1 Hz strongly depressed both responses. Unexpectedly EPSC2 became silent before EPSC1, leading to the conversion of PPF into PPD. Responses reappeared after setting the frequency of stimulation to 0.025 Hz (Fig. 1D and E). In this particular case EPSC1 recovered less than EPSC2 leading to a very strong PPF.
LFD was observed also in cases with initial high probability of release as in the example of Fig. 2, where only successes were recorded in control conditions (at 0.05 Hz). Increasing the stimulation frequency led to a decrease in the mean EPSCs amplitude and to the appearance of some transmission failures. Again, at 1 Hz, the responses were almost abolished with sporadic responses to EPSC2. A complete recovery of both EPSC1 and EPSC2 was obtained at 0.025 Hz (Fig. 2A-C). Similar depression and recovery were observed in an additional four neurons that were sequentially tested at 0.05, 0.1, 1 and 0.025 Hz. Mean values of failure rate, changes in EPSC amplitudes and PPR are shown in Fig. 3. While at 0.05 Hz the failure rate in response to the first spike was significantly higher than that to the second (filled and open columns of Fig. 3A, respectively, P < 0.05), at 0.1 Hz this difference was less pronounced. At 1 Hz, the failure rate of synaptic response to the second spike was significantly (P < 0.05) higher than that to the first one. A partial recovery was observed at 0.025 Hz (n = 4). The higher failure rate found at 0.025 Hz in comparison with that at 0.05 Hz reflects the slowness of the recovery process since, for any given frequency, we analysed responses occurring during the entire period of stimulation. This also accounts for the smaller mean EPSC amplitude measured at 0.025 Hz with respect to controls (Fig. 3B, see also Fig. 2B and C). In fact, the percentage of recovery obtained after the first 5 min was 70 ± 20 and 80 ± 20 %, for EPSC1 and EPSC2, respectively (n = 3). This rules out the possibility of long-term depression () or response run down. At 0.1 and 1 Hz, changes in failure rate were associated with a reduction in the mean EPSC amplitude (Fig. 3B). The greater reduction of EPSC2 in comparison with EPSC1 is in line with the frequency-dependent PPR changes shown in Fig. 3C. Similar results were found in an additional eight neurons in which use-dependent depression was investigated using different stimulation frequencies.
Summary data of changes in failures rate, amplitude and paired-pulse ratio of synaptic responses evoked at different frequenciesA, failure rates to the first (filled columns) and second (open columns) pulse under different frequencies of stimulation (n = 5). B, mean EPSC amplitude plotted for three different frequencies as a percentage of control (six connected pairs). C, PPR calculated at different frequencies for the same six pairs of cells (0.025 Hz refers to 3 neurons). Reference marks (* and +) indicate significant differences from control or from 1 Hz values, respectively. Single reference marks indicate P < 0.05 and double reference marks P < 0.001.
Time course of low frequency depression
The amount of depression varied from cell to cell according to their previous history (i.e. to the duration and frequency of stimulation). In the graphs of Fig. 4, the time courses of depression and recovery are shown. While at 0.05 Hz, the mean EPSC amplitudes remained stable (Fig. 4A), LFD appeared at the presynaptic firing frequency of 0.1 Hz and strongly increased at 1 Hz (Fig. 4B and C). The time course of depression of EPSC1 could be fitted with a mono-exponential function having a time constant of 98 and 36 s at 0.1 and 1 Hz, respectively. Interestingly, the time course of depression of EPSC2 did not follow that of EPSC1. While at 0.1 Hz, depression of EPSC2 was slower than EPSC1 (time constants of 145 versus 98 s), at 1 Hz it was faster (7 versus 36 s). All synapses slowly recovered from depression during 0.025 Hz stimulation. The time course of recovery was well described by a single exponential function having a time constant of 594 and 603 s for EPSC1 and EPSC2, respectively (Fig. 4D). The slowness of this process may account for the incomplete (70 ± 20 %) recovery of EPSC1 after 5 min. In line with a presynaptic site of action, LFD did not modify EPSC kinetics (Fig. 4E). As shown in Fig. 4F, the decay time constants of single EPSCs evoked at different frequencies were not significantly different (P = 0.89). Moreover, taking advantage of the double recordings technique, possible modifications of the presynaptic action potential were investigated. In spite of a small reduction in spike amplitude (from 92 ± 9 to 86 ± 9 mV), no change in spike half-width value (from 1.7 ± 0.1 to 1.66 ± 0.07 ms) was observed between 0.05 and 1 Hz.
Time course of frequency depression and recoveryA-D, time course of normalized mean EPSC amplitude evoked at different frequencies (n = 5-6 neurons for A-C and 3 for D, see methods for normalization). Filled and open circles refer to the first and second EPSC, respectively. Data points were fitted with one exponential function. Continuous and dotted lines represent the fit to the first and the second EPSC, respectively. E, an example of normalized and superimposed mean EPSC evoked at 0.05 (continuous line) and 1 Hz (dashed line) in one representative neuron. F, mean decay time constant of EPSC evoked at 0.05 and 1 Hz (n = 4).
NMDA receptors are not involved in short-term low frequency depression
Recently it has been reported that in the same preparation at CA3–CA3 connections, presynaptic action potentials (600 pulses) at 1 Hz associated with a slight depolarization (−55 mV) of the postsynaptic cell were able to induce a long-term depression (LTD) of EPSCs of almost 80 % (). This effect required the activation of NMDA receptors and was prevented by the NMDA receptor antagonist AP5 (). To see whether a similar type of mechanism could account for the present results, in four experiments LFD was induced in the presence of the NMDA receptor antagonist AP5 (50 μM). A similar degree of depression of EPSC1 was attained when the presynaptic firing was switched from 0.05 to 1 Hz (34 ± 6 and 40 ± 10 % in control and AP5, respectively). Also in the presence of AP5, this effect was associated with a reduction of the paired-pulse ratio (from 1.3 ± 0.3 to 0.82 ± 0.03, data not shown). In two cases, LFD was studied in the same pairs before and after application of AP5. Thus, after EPSC depression at 1 Hz and recovery, presynaptic firing was set again at 1 Hz in the presence of AP5. Also in these cases a similar degree of depression was obtained in control (31 ± 9 %) and in the presence of AP5 (28 ± 4 %). These data exclude the involvement of NMDA receptors in LFD.
DISCUSSION
The present results indicate that LFD is a common form of short-term synaptic plasticity at CA3–CA3 connections in hippocampal slice cultures. The reduction in synaptic strength varied across different connections according to presynaptic firing frequency and was associated with a modulation of the paired-pulse ratio, with PPD prevailing over PPF at higher stimulation frequencies.
Mechanisms of low-frequency synaptic depression
Frequency depression constitutes the predominant form of short-term dynamics in many CNS structures including sensorimotor cortex (; ; ), auditory pathways (), cerebellum () and hippocampus (). Although in the present experiments comparatively low frequencies were used, they were able to produce a gradual reduction in EPSCs amplitude. Moreover, especially at 1 Hz, synaptic depression could be so strong as to make synapses almost ‘silent’. The LFD observed here differed from the LTD recently described in the same preparation by since responses recovered after switching to a low frequency stimulation. Unlike LTD, the induction of LFD was NMDA independent because it was insensitive to AP5. Moreover, LFD could be induced from a holding potential of −69 mV (including the correction for liquid junction potential, see Methods), while LTD from −55 mV.
As in other CNS structures (; ; ; ), use-dependent depression was found to be presynaptic in origin as suggested by the increase in failure rate and changes in PPR. Presynaptic mechanisms interfering with synaptic vesicle release include changes in action potential waveform, inactivation of calcium currents, modulation of presynaptic calcium channels by G-protein coupled receptors, etc. Changes in action potential waveform may depend on sodium channel inactivation. However, this mechanism does not seem to play a crucial role in paired-pulse depression under basal conditions (). Moreover, the observation that small changes in spike amplitude were not associated with modifications in action potential half-width rules out sodium channel inactivation playing an important role in LFD (; von Gersdorff & Borst, 2001). Presynaptic group II and III metabotropic glutamate receptors could potentially exert a strong inhibitory effect on transmitter release (; ). These receptors, however, are not present on hippocampal associative commissural fibres ().
Although the locus of depression is likely to be presynaptic, a postsynaptic effect, such as receptor desensitization, cannot be completely ruled out. In our case, EPSC kinetics was unchanged under different stimulation frequencies, suggesting that modifications in AMPA receptor gating are not involved in this form of plasticity.
Frequency-dependent modulation of paired-pulse ratio
Several independent approaches suggest that both frequency depression and paired-pulse modulation depend largely on changes in release probability (Pr). A reduction in Pr during LFD predicts an increase in the PPR. Indeed, PPR has been shown to be enhanced when the extracellular calcium/ magnesium ratio was lowered (; ). Unexpectedly, we found that PPF was converted into PPD. The reduction in PPR can be attributed to the activation of two temporally distinct processes: (i) modulation of the release probability by residual calcium and (ii) changes in size of the readily releasable pool of vesicles (; ). The idea of expressing release probability as the balance between two processes, one related to the residual calcium and the other to vesicle availability, is similar in many aspects to that put forward by , and . Unlike previously proposed models, we did not impose any constraint on the number of docked vesicles per release site, even though the model proposed by could lead to equations similar to those presented in the Appendix. Moreover, we did not consider desensitization (), since our data exclude this possibility. Finally, in our model the interaction between paired-pulse ratio and stimulation frequency was described by a differential equation for the mean population of docked vesicles. It should be stressed that, although the number of docked vesicles may be related to the release probability (), this cannot be generalised to all synapses (see ).
According to our model, the probability of release at a single active zone (Pr) can be written as:
(1)
where Pr(Ca2+) and Pr(Ves) are the probability of release dependent on residual calcium and on the size of the available pool, respectively. The interplay between Pr(Ca2+) and Pr(Ves) at the moment of arrival of the second spike would determine the direction of the paired-pulse modulation (PPF or PPD). Thus, PPF observed in the majority of neurons in stationary conditions would be mainly dictated by the residual calcium (), being Pr(Ves) constant. In fact, at lower stimulation frequencies the size of the pool should be larger and any ‘loss’ of vesicles during the first pulse will be negligible. At higher stimulation frequencies, when the available pool becomes too small, an additional depletion due to the release from the first presynaptic spike in the paired-pulse protocol could diminish the size of the pool to such a critical level that at the moment of the arrival of the second spike Pr(Ves) would be close to zero. In this case, according to eqn (1), Pr would be close to zero independently of the value of Pr(Ca2+). Therefore, PPR is expected to increase at relatively low frequencies, and to decrease at higher frequencies when fewer vesicles are available for release. To validate these assumptions, a simple model has been developed (see Appendix). As predicted by the model in stationary conditions, when the time-dependent term of the process is neglected (f(t) of eqn (A8), Appendix), PPR is dependent on the frequency of stimulation (Fig. 5A). It is clear from the figure that the switch from PPF to PPD occurs between 0.1 and 1 Hz.
Expected frequency-dependent modulation of paired-pulse ratioA, stationary PPR obtained for different stimulation frequencies according to the proposed model (see eqn (A6), Appendix). B, predicted time course of the probability of release to the first (filled circles) and second pulse (open circles) at different stimulation frequencies. C, PPR calculated from the data of Fig. 4B for respective frequencies. PPR was obtained as the ratio between the mean values of p2 and p1. Model parameters: α= 0.48, γ= 0.28, nd = 18, λr = 0.0026 Hz, τ= 100 ms, Nc = 5.
The time dependency of the process is introduced by describing the dynamics of the population of docked vesicles via a first-order differential equation that considers both depletion and refilling of the pool. These two processes are characterized by time constants λd and λr, respectively (see also ). While λd depends on the stimulation frequency ω, the question as to whether λr also depends on the same factor is still open. Recently it has been suggested that the endocytosis rate, which influences the refilling of the pool (λr), depends on the frequency of stimulation (). However, assuming that λr is independent of ω, we underestimated the frequency-dependent decrease in the mean pool size. The time course of the probability of release obtained at different frequencies (Fig. 5B) is similar to that of EPSC amplitudes obtained in experimental conditions, including the switch from PPF to PPD at 1 Hz. This suggests that a decrease in λr can account for our observations, without requiring changes in quantal size or in the number of release sites. By averaging the probability of release over all trials (at each frequency) an estimate of PPR in non-stationary conditions was obtained (Fig. 5C). As in experimental conditions (compare with Fig. 3C) PPR decreased with increasing frequency leading to PPD at 1 Hz. This effect fully recovered after switching to 0.025 Hz.
In conclusion, from our experiments it appears that presynaptic changes in Pr accounts for frequency-dependent synaptic depression. Release probability is directly correlated with a morphologically defined pool of docked vesicles (; but see ). The size of this pool would vary enormously between different synapses. This may account for the different time course of depletion and replenishment and for distinct frequency-dependent modulation of paired-pulse ratio found in various brain structures. Thus, a large pool of vesicles at the calyx of Held would ensure reliable synaptic transmission even at high frequencies when only a small fraction of synaptic vesicles are available for release ().
Acknowledgments
We wish to express our gratitude to B. Gähwiler and L. Ballerini for helping us setting the hippocampal organotypic cultures. This work was supported by grants from Ministero dell'Università e Ricerca Scientifica (MURST) to E.C., from INTAS to E.C. and L.L.V. and from RFBR and the Welcome Trust to L.L.V.; L.P.S. was supported by the Program for Training and Research in Italian laboratories, International Center for Theoretical Physics, Trieste, Italy.
APPENDIX
The dependency of Pr on the size of the available pool can be written as:
(A1)
where k is a constant, p1 is the probability of release to the first pulse, λ the fusion rate for a vesicle integrated over the duration of the presynaptic pulse () and N is the number of ready releasable vesicles.
The probability of release to the second pulse (p2) is given by:
(A2)
where P2(1) is the probability of release to the second pulse, after release to the first and P2(0) the same probability after no release to the first pulse.
Taking into account also the effect of the residual calcium, we can write:
(A3)
and
(A4)
where α is a constant representing the sensitivity to residual calcium and τ is the time constant of decay of the higher probability of release, which is related to the rate of removal of residual calcium ().
Hence the paired-pulse ratio is:
(A5)
Thus, by substituting P2(1), P2(0) and p1 from eqns (A3) and (A4) into eqn (A5):
(A6)
The number of available vesicles N can be modelled by writing:
(A7)
where λd and λr are the depletion and refilling constants, respectively, while Nc is the maximum size of the ready releasable pool of vesicles.
This equation states that there is a use-dependent depletion, whose rate depends on: (i) the frequency of stimulation (ω) with λd = ω/nd, and (ii) the refilling process controlled by λr. The solution of this first-order differential equation gives us the mean occupancy of each release site in our model synapse at time t:
(A8)
where the time-dependant term f(t) is equal to a exp(−(λd + λr)t) and the equilibrium term Neq is given by:
REFERENCES
- Abbott LF, Varela JA, Sen K, Nelson SB. Synaptic depression and cortical gain control. Science. 1997;275:220–224. [PubMed] [Google Scholar]
- Atluri PP, Regehr WG. Determinants of the time course of facilitation at the granule cell to Purkinje cell synapse. Journal of Neuroscience. 1996;16:5661–5671.[PMC free article] [PubMed] [Google Scholar]
- Berretta N, Rossokhin AV, Kasyanov AM, Sokolov V, Cherubini E, Voronin LL. Postsynaptic hyperpolarisation increases the strength of AMPA mediated synaptic transmission at large synapses between the mossy fibres and CA3 pyramidal cells. Neuropharmacology. 2000;39:2288–2301. [PubMed] [Google Scholar]
- Brody DL, Yue DT. Release-independent short-term synaptic depression in cultured hippocampal neurons. Journal of Neuroscience. 2000;20:2480–2494. [PubMed] [Google Scholar]
- Canepari M, Cherubini E. Dynamics of excitatory transmitter release: analysis of synaptic responses in CA3 hippocampal neurons after repetitive stimulation of afferent fibers. Journal of Neurophysiology. 1998;79:1977–1988. [PubMed] [Google Scholar]
- Debanne D, Guerineau NC, Gähwiler BH, Thompson SM. Paired-pulse facilitation and depression at unitary synapses in rat hippocampus: quantal fluctuation affects subsequent release. Journal of Physiology. 1996;491:163–176.[PMC free article] [PubMed] [Google Scholar]
- Dittman JS, Kreitzer AC, Regehr WG. Interplay between facilitation, depression and residual calcium at three presynaptic terminals. Journal of Neuroscience. 2000;20:1374–1385. [PubMed] [Google Scholar]
- Dobrunz LE, Stevens CF. Heterogeneity of release probability, facilitation and depletion at central synapses. Neuron. 1997;18:995–1008. [PubMed] [Google Scholar]
- Forsythe ID, Clements JD. Presynaptic glutamate receptors depress excitatory monosynaptic transmission between mouse hippocampal neurones. Journal of Physiology. 1990;429:1–16.[PMC free article] [PubMed] [Google Scholar]
- Gähwiler BH. Organotypic monolayer cultures of nervous tissue. Journal of Neuroscience Methods. 1981;4:329–342. [PubMed] [Google Scholar]
- Gasparini S, Saviane C, Voronin LL, Cherubini E. Silent synapses in the developing hippocampus: lack of functional AMPA receptors or low probability of glutamate release. Proceedings of the National Academy of Sciences of the USA. 2000;97:9741–9746.[PMC free article] [PubMed] [Google Scholar]
- He Y, Zorumski CF, Mennerick S. Contribution of presynaptic Na+ channel inactivation to paired-pulse synaptic depression in cultured hippocampal neurons. Journal of Neurophysiology. 2002;87:925–936. [PubMed] [Google Scholar]
- Larkman AU, Jack JJB, Stratford KJ. Quantal analysis of excitatory synapses in rat hippocampal CA1 in vitro during low-frequency depression. Journal of Physiology. 1997;505:457–471.[PMC free article] [PubMed] [Google Scholar]
- Markram H, Tsodyks M. Redistribution of synaptic efficacy between neocortical pyramidal neurons. Nature. 1996;382:807–810. [PubMed] [Google Scholar]
- Markram H, Wang Y, Tsodyks M. Differential signaling via the same axon of neocortical pyramidal neurons. Proceedings of the National Academy of Sciences of the USA. 1998;95:5323–5328.[PMC free article] [PubMed] [Google Scholar]
- Meyer AC, Neher E, Schneggenburger R. Estimation of quantal size and number of functional active zones at the calix of Held synapse by nonstationary EPSC variance analysis. Journal of Neuroscience. 2001;21:7889–7900. [PubMed] [Google Scholar]
- Montgomery JM, Madison DM. State-dependent heterogeneity in synaptic depression between pyramidal cell pairs. Neuron. 2002;33:765–777. [PubMed] [Google Scholar]
- Neher E, Sakaba T. Combining deconvolution and noise analysis for the estimation of transmitter release rates at the calix of Held. Journal of Neuroscience. 2001;21:444–461. [PubMed] [Google Scholar]
- Nelson SB, Turrigiano GG. Synaptic depression: a key player in the cortical balancing act. Nature Neuroscience. 1998;1:539–541. [PubMed] [Google Scholar]
- Nicholls J, Wallace BG. Quantal analysis of transmitter release at an inhibitory synapse in the central nervous system of the leech. Journal of Physiology. 1978;281:171–185.[PMC free article] [PubMed] [Google Scholar]
- O'Donovan MJ, Rinzel J. Synaptic depression: a dynamic regulator of synaptic communication with varied functional roles. Trends in Neurosciences. 1997;20:431–433. [PubMed] [Google Scholar]
- Oleskevich S, Clemens J, Walmsley B. Release probability modulates short-term plasticity at a rat giant terminal. Journal of Physiology. 2000;524:513–523.[PMC free article] [PubMed] [Google Scholar]
- Otis T, Zhang S, Trussel LO. Direct measurement of AMPA receptor desensitization induced by glutamatergic synaptic transmission. Journal of Neuroscience. 1996;16:7496–7504.[PMC free article] [PubMed] [Google Scholar]
- Pavlidis P, Madison DV. Synaptic transmission in pair recordings from CA3 pyramidal cells in organotypic culture. Journal of Neurophysiology. 1999;81:2787–2797. [PubMed] [Google Scholar]
- Quastel DM. The binomial model in fluctuation analysis of quantal neurotransmitter release. Biophysical Journal. 1997;72:728–753.[PMC free article] [PubMed] [Google Scholar]
- Rosenmund C, Stevens CF. Definition of the readily releasable pool of vesicles at hippocampal synapses. Neuron. 1996;16:1197–1207. [PubMed] [Google Scholar]
- Scheuss V, Neher E. Estimating parameters from mean, variance, and covariance in trains of synaptic responses. Biophysical Journal. 2001;81:1970–1989.[PMC free article] [PubMed] [Google Scholar]
- Silver RA, Momiyama A, Cull-Candy SG. Locus of frequency dependent depression identified with multiple-probability fluctuation analysis at rat climbing fibre-Purkinje cell synapses. Journal of Physiology. 1998;510:881–902.[PMC free article] [PubMed] [Google Scholar]
- Sun J, Wu X, Wu L. Single and multiple vesicles fusion induce different rates of endocytosis at a central synapse. Nature. 2002;417:555–559. [PubMed] [Google Scholar]
- Takahashi T, Forsythe ID, Tsujimoto T, Barnes-Davies M, Onodera K. Presynaptic calcium current modulation by a metabotropic glutamate receptor. Science. 1996;274:594–597. [PubMed] [Google Scholar]
- Thomson AM, Deuchars J. Temporal and spatial properties of local circuits in neocortex. Trends in Neurosciences. 1994;17:119–126. [PubMed] [Google Scholar]
- Thomson AM, Deuchars J, West DC. Large, deep layer pyramid-pyramid single axon EPSPs in slices of rat motor cortex display paired pulse and frequency-dependent depression, mediated presynaptically and self-facilitation mediated postsynaptically. Journal of Neurophysiology. 1993;70:2354–2369. [PubMed] [Google Scholar]
- Tsodkys M, Markram H. The neural code between neocortical pyramidal neurons depends on neurotransmitter release probability. Proceedings of the National Academy of Sciences of the USA. 1997;94:719–723.[PMC free article] [PubMed] [Google Scholar]
- Von Gersdorff H, Borst GJ. Short term plasticity at the calyx of Held. Nature Reviews Neuroscience. 2002;3:53–64. [PubMed] [Google Scholar]
- Von Gersdorff H, Schneggenburger R, Weis S, Neher E. Presynaptic depression at a Calyx synapse: the small contribution of metabotropic glutamate receptors. Journal of Neuroscience. 1997;17:8137–8148.[PMC free article] [PubMed] [Google Scholar]
- Voronin LL. On the quantal analysis of hippocampal long-term potentiation and related phenomena of synaptic plasticity. Neuroscience. 1993;56:275–304. [PubMed] [Google Scholar]
- Wang L-Y, Kaczmarek LK. High-frequency firing helps replenish the readily releasable pool of synaptic vesicles. Nature. 1998;394:384–388. [PubMed] [Google Scholar]
- Wu L-J, Saggau P. Presynaptic inhibition of elicited neurotransmitter release. Trends in Neurosciences. 1997;20:204–212. [PubMed] [Google Scholar]
- Xu-Friedman MA, Harris KM, Regehr WG. Three-dimensional comparison of ultrastructural characteristics at depressing and facilitating synapses onto cerebellar Purkinje cells. Journal of Neuroscience. 2001;21:6666–6672. [PubMed] [Google Scholar]
- Zucker RS. Changes in the statistics of transmitter release during facilitation. Journal of Physiology. 1973;229:787–810.[PMC free article] [PubMed] [Google Scholar]
- Zucker RS. Short-term synaptic plasticity. Annual Review of Neuroscience. 1989;12:13–31. [PubMed] [Google Scholar]
Articles from The Journal of Physiology are provided here courtesy of The Physiological Society