Facilitation of presynaptic calcium currents in the rat brainstem

Abstract

1. To study use-dependent changes in the presynaptic Ca²⁺ influx and their contribution to transmitter release, we made simultaneous voltage clamp recordings from presynaptic terminals (the calyces of Held) and postsynaptic cells (the principal cells of the medial nucleus of the trapezoid body) in slices of the rat auditory brainstem.

2. Following a short (2 ms) prepulse to 0 mV, calcium channels opened faster during steps to negative test potentials. During trains of action potential waveforms the Ca²⁺ influx per action potential increased. At the same time, however, the amplitude of the EPSCs decreased.

3. The facilitation of the calcium currents appeared to depend on a build-up of intracellular Ca²⁺, since its magnitude was proportional to the Ca²⁺ influx and it was reduced in the presence of 10 mm BAPTA.

4. Facilitation of the presynaptic calcium currents may contribute to short-term facilitation of transmitter release, observed when quantal output is low. Alternatively, it may counteract processes that contribute to synaptic depression.

The main function of presynaptic action potentials is to open voltage-dependent calcium channels in the vicinity of synaptic vesicles. The subsequent influx of Ca²⁺ triggers their fusion with the presynaptic membrane. Successive action potentials are not equally effective in evoking release. During a train of action potentials, release may be facilitated or depressed, depending largely on the size of the ‘local release fraction’, that is the fraction of the releasable vesicle pool that is released by a single action potential (reviewed in Zucker, 1996; Fisher et al. 1997).

Residual intracellular Ca²⁺ is involved in synaptic facilitation (Katz & Miledi, 1968). An increase in the Ca²⁺ influx during an action potential is not necessary for paired-pulse facilitation of transmitter release (e.g. Charlton et al. 1982; Wright et al. 1996). However, if it were to occur, an increase in the calcium current could increase synaptic facilitation. An increase in the calcium current during successive voltage steps has been observed for somatic calcium currents or currents from cloned calcium channels. Different mechanisms may underlie this facilitation (Dolphin, 1996).

Because of the lack of accessibility of most presynaptic terminals in the central nervous system (CNS), the contribution of changes in presynaptic Ca²⁺ influx to synaptic facilitation has been difficult to assess. Here we addressed this issue in an axosomatic synapse in the medial nucleus of the trapezoid body (MNTB). Each principal cell in the MNTB is innervated by a giant terminal, called the calyx of Held. The MNTB is part of an auditory pathway that is involved in the localization of sound (Oertel, 1997). It is possible to voltage clamp both the pre- and the postsynaptic part of this synapse in whole-cell recordings (Borst et al. 1995; Borst & Sakmann, 1998), making it possible to correlate directly the size and time course of the Ca²⁺ influx with release. Although at physiological extracellular Ca²⁺ concentrations ([Ca²⁺]_o) synaptic depression is observed, at reduced Ca²⁺ concentrations synaptic facilitation is displayed (Barnes-Davies & Forsythe, 1995; Borst et al. 1995). Here we study whether facilitation of the presynaptic Ca²⁺ influx occurs during double-pulse protocols or during trains of action potential waveforms.

METHODS

In a few experiments, terminals were voltage clamped using two-electrode voltage clamp recordings, as described previously (Borst & Sakmann, 1998).

Simultaneous pre- and postsynaptic voltage clamp recording

Eight- to ten-day-old Wistar rats were decapitated without prior anaesthesia using a small animal guillotine, in accordance with national guidelines. Parasagittal slices (200 μm) were cut with a vibratome in ice-cold saline, containing (mm): 125 NaCl, 2.5 KCl, 4 MgCl₂, 0.1 CaCl₂, 25 dextrose, 1.25 NaH₂PO₄, 0.4 ascorbic acid, 3 myo-inositol, 2 sodium pyruvate, 25 NaHCO₃. Slices were equilibrated for 30 min at 37°C in the same solution, except that the Ca²⁺ and Mg²⁺ concentrations were 2 and 1 mm, respectively. Afterwards, slices were transferred to a recording chamber and during recordings of the calcium current, they were perfused with a solution containing (mm): 105 NaCl, 20 TEA-Cl (Fluka, Buchs, Switzerland), 0.1 3,4-diaminopyridine (Sigma), 2.5 KCl, 1 MgCl₂, 2 CaCl₂, 25 dextrose, 1.25 NaH₂PO₄, 0.4 ascorbic acid, 3 myo-inositol, 2 sodium pyruvate, 0.001 tetrodotoxin (TTX, RBI), 0.05 D-2-amino-5-phosphonovalerate (D-APV, Tocris Neuramin), 25 NaHCO₃, pH 7.4 when bubbled with 95 % O₂, 5 % CO₂. Bath temperature was 22–24°C.

Presynaptic electrodes (4–6 MΩ) were filled with a solution containing (mm): 125 caesium gluconate, 20 CsCl, 10 di-sodium phosphocreatine, 4 MgATP, 0.3 GTP, 0.05 fura-2 (Molecular Probes), 10 Hepes (pH 7.2 with CsOH). Postsynaptic pipettes (2.5–3.5 MΩ) were filled with (mm): 125 potassium gluconate, 20 KCl, 10 di-sodium phosphocreatine, 4 MgATP, 0.3 GTP, 0.5 EGTA, 10 Hepes (pH 7.2 with KOH).

Simultaneous pre- and postsynaptic whole-cell recordings were made with two Axopatch 200A amplifiers. Passive properties of terminals and postsynaptic cells were routinely measured (Borst & Sakmann, 1998). In most terminals the passive response could be well fitted with a single exponential function, whereas more than one exponential function was needed for the passive response of the postsynaptic cell. Presynaptic and postsynaptic series resistance (typically 15 and 8 MΩ, respectively) were compensated to 90 % (lag < 8 μs, prediction when applicable 60 %). Holding potential was −80 mV. Potentials were corrected for a liquid junction potential of −11 mV. Voltage clamping of terminals with action potential waveforms was done as described previously (Borst & Sakmann, 1996). Subtraction of passive responses was with the P/5 or the P/-5 method (Borst & Sakmann, 1998). Currents were filtered at 5 kHz and digitized at 50 kHz with a 16 bit analog-to-digital converter (ITC-16, Instrutech, Greatneck, NY, USA). Interpulse interval was 15–60 s.

Fluorescence measurements

The presynaptic volume-averaged Ca²⁺ concentration was measured by forming ratios between continuous recordings of fluorescence at a calcium-insensitive (357 nm) wavelength followed by a calcium-sensitive wavelength (380 nm), after background subtraction (Helmchen et al. 1997; Wu et al. 1998). Fura-2 was excited using a polychromatic illumination system (T.I.L.L. Photonics, Munich, Germany). Excitation light was coupled into an upright microscope (Axioskop, Zeiss) via a light guide. The microscope was equipped with a × 40 water-immersion objective (NA 0.75, Zeiss), a dichroic mirror (500 nm) and a long-pass (520 nm) emission filter. Two photodiodes (area, 1.1 mm × 1.1 mm; Hamamatsu) situated in the image plane were used for signal and background subtraction (Borst & Helmchen, 1998). Photodiode signals were filtered at 20 Hz (8-pole Bessel filter). Excitation light was generally attenuated by 60 %, reducing bleaching of fura-2 (or the background fluorescence) to less than 1 % per second. When necessary, fluorescence signals were corrected for bleaching by linear interpolation.

Minimal and maximal fluorescence ratios (R_min and R_max) were measured in terminals using intracellular solutions containing 20 mm EGTA and 20 mm CaCl₂, respectively (Helmchen et al. 1997). R_min was 0.56 ± 0.04 (n = 4) and R_max was 5.9 ± 0.5 (n = 3). The K_d for Ca²⁺ of fura-2 was assumed to be 273 nm (Helmchen et al. 1997).

Acquisition and analysis

Acquisition and analysis were done with Pulse Control 4.7 (Herrington & Bookman, 1994), combined with IGOR (WaveMetrics, Lake Oswego, OR, USA) macros. To measure the shift in the isochronal current-voltage (I–V) relationships following a brief prepulse, the I-V relationships were interpolated with a cubic spline function. The shift was measured at around −20 mV.

Simulation of the calcium currents evoked by an action potential waveform or step depolarizations were done using a Hodgkin- Huxley model with two gating particles (m). Parameters were taken from Borst & Sakmann (1998):

where V (in mV) is the voltage command potential and α_m and β_m are the rate constants (in ms⁻¹) for the opening and closing of the gates, respectively. To assess the effect of the facilitation of the calcium current as a result of brief prepulses on the currents during action potentials waveforms, α_m was shifted to more negative values by trial and error, until the shift in simulated I-V relationships matched the experimentally observed shift. This shift, S, was −2.2 and −4.0 mV for the prepulses to +60 and 0 mV, respectively. Calyces from 8- to 10-day-old rats contain a mixture of N-, P/Q- and R-type calcium channels (Wu et al. 1998; L.-G. Wu, J. G. G. Borst & B. Sakmann, unpublished results). This simple Hodgkin-Huxley model neglects the heterogeneity in both the biophysical and the pharmacological properties of the calcium currents (Borst & Sakmann, 1998).

Data are presented as average ± standard error of the mean (s.e.m.).

RESULTS

Facilitation of presynaptic calcium currents during voltage steps

A 2 ms prepulse to 0 or +60 mV resulted in faster activation of presynaptic calcium channels during subsequent test pulses at negative command voltages (n = 4; Fig. 1A). At potentials above +10 mV no difference was observed (Fig. 1B and C). The 2 ms prepulses shifted the isochronal I-V relationship to more negative values at potentials lower than 0 mV (n = 4; Fig. 1C). This increase in the inward current during the test pulse at negative potentials was not due to a contaminating current or to imperfect leak subtraction: Firstly, the tail currents increased in parallel (horizontal arrows in Fig. 1A). Secondly, the volume-averaged Ca²⁺ concentration evoked by the voltage steps in the calyx increased proportionally to the calcium current (not shown). This also suggests that calcium-induced calcium release did not contribute to the measured increases in the Ca²⁺ concentration (Helmchen et al. 1997). Thirdly, in two of four experiments the increase in the inward current following the prepulse to +60 mV resulted in larger EPSCs at these potentials than in the absence of a prepulse (Fig. 1A, bottom).

Figure 1. Facilitation of presynaptic calcium currents.

A, from top to bottom: presynaptic command voltage (V_pre), presynaptic calcium current (I_pre) and postsynaptic currents (I_post). A 2 ms prepulse to 0 mV (thin line) resulted in a facilitation of the calcium current during a 5 ms step to −20 mV compared with the current in the absence of a prepulse (thick line). A somewhat smaller effect on the presynaptic current was observed after a prepulse to +60 mV (dotted line). The vertical arrow indicates where the isochronal current-voltage (I–V) relationship was evaluated. The horizontal arrows point to the peak amplitude of the tail currents with (lower arrow) and without (upper arrow) prepulse. B, no effect of the prepulse on the current during the test pulse was observed when the test pulse was to +20 mV. C, isochronal I-V relationship. Amplitudes of the currents during the test pulse were measured 1.5 ms after the start of the depolarizing step. The I-V relationship for the prepulse to 0 mV (□, thin line) was shifted to the left for negative voltages compared with the I-V relationship in the absence of a prepulse (•, thick line). An intermediate effect was observed after the prepulse to +60 mV (▴, dotted line). D, changes in volume-averaged Ca²⁺ concentration ([Ca²⁺]_i) resulting from the prepulse to 0 mV (thin line) or to +60 mV (dotted line). The thick line shows [Ca²⁺]_i in the absence of a voltage step. Different experiment from the one shown in A–C.

During the prepulse to 0 mV, 3.8 ± 0.3 pC of charge carried by Ca²⁺ entered the terminal, raising the volume-averaged Ca²⁺ concentration by about 330 nm (Fig. 1D). The interpolated isochronal I-V relationship was shifted by −2.8 ± 0.3 mV (n = 4). The maximal calcium current, elicited at around 0 mV, was about 5 % larger (Fig. 1C). In the same experiments, during a prepulse to +60 mV, about 1.2 ± 0.1 pC of calcium charge entered, mostly as a tail current when the terminal was repolarized to the holding potential of −80 mV. This raised the volume-averaged Ca²⁺ concentration by around 120 nm (Fig. 1D). In all experiments, the shift in the isochronal I-V relationship was somewhat smaller following the prepulse to +60 mV, being on average −1.6 ± 0.2 mV (Fig. 1C). The significantly (P < 0.01; paired t test) smaller shift that followed this larger voltage step argues against a dominant contribution of voltage per se. Instead, this facilitation of the calcium current could be due to an effect of intracellular calcium ions on calcium channels.

Effect of voltage steps on transmitter release

In the bottom panels of Fig. 1A and B, the synaptic currents evoked by the different presynaptic calcium currents are shown. The delay and amplitude of the EPSC depended on the speed of activation and the peak amplitude of the calcium current. The 2 ms step to +60 mV did not evoke release until after the step, whereas the 2 ms step to 0 mV evoked a much larger EPSC, which had already begun during the step. The calcium tail current evoked by the step to +60 mV evoked an EPSC whose properties were similar to action potential-evoked EPSCs (Barnes-Davies & Forsythe, 1995; Borst et al. 1995). For example, 20–80 % rise times of EPSCs were around 0.25 ms in both cases.

The amplitude of the EPSCs showed run-down (cf. Fig. 1A and B). Because the different prepulses were intermixed, the effects of run-down on transmitter release were minimized. Facilitation of transmitter release following the prepulse to −20 mV was observed in only two of four experiments (e.g. Fig. 1A, bottom panel). Most likely, synaptic depression induced by the transmitter release during the brief prepulse could overcome the effects of the increase in the calcium current on transmitter release. At a test pulse of +20 mV, the calcium currents did not increase following the prepulses and EPSCs were reduced by the prepulse in all four experiments (Fig. 1B).

Facilitation of presynaptic calcium currents during action potential waveform commands

Terminals were voltage clamped with trains of identical action potential waveforms at 100 Hz. They were either perfused with a low concentration of Ca²⁺ buffer (50 μm fura-2) during simultaneous pre- and postsynaptic whole-cell voltage clamp experiments, or with a high concentration of exogenous Ca²⁺ buffer (10 mm BAPTA) during two-electrode voltage clamp experiments. The two-electrode voltage clamp allowed us to check the voltage control within the terminal, but the postsynaptic cell could not be recorded from in these experiments. The measured voltage change induced by the action potential waveform did not change during the train. Under both conditions, the peak amplitude of the Ca²⁺ influx increased during the first few action potential waveforms. In the presence of 50 μm fura-2 the increase was significantly larger than with 10 mm BAPTA (Fig. 2), supporting the view that the calcium current facilitation depended upon a build-up of intracellular Ca²⁺. In the presence of low Ca²⁺ buffer concentrations, the volume-averaged Ca²⁺ concentration increased by several hundred nanomolar during the trains (Fig. 2A, bottom panel).

Figure 2. Facilitation of presynaptic calcium currents during a train of action potential waveforms.

A, increase of the Ca²⁺ influx per action potential during a 100 Hz train of identical action potential waveforms. Simultaneous pre- and postsynaptic voltage clamp recording. From top to bottom, the panels show: presynaptic voltage command, presynaptic calcium currents, EPSCs and volume-averaged calcium concentration. Presynaptic pipette solution contained 50 μm fura-2. B, lack of a clear increase in the Ca²⁺ influx per action potential during a 100 Hz train of identical action potential waveforms in the presence of 10 mm BAPTA in the presynaptic pipette solution. Two-electrode voltage clamp recording of the calyx. Top panel, presynaptic voltage command; bottom panel, presynaptic calcium currents. Same scale applies as to the two top panels of A. C, average increase in the peak amplitudes of the calcium currents evoked by an action potential waveform in the presence of 10 mm BAPTA (○, n = 5) or 50 μm fura-2 (•, n = 6). Amplitudes were normalized to the first one. Following the second stimulus, normalized amplitudes in the presence of low and high Ca²⁺ buffer concentrations were significantly different (P < 0.05, unpaired t test for unequal variances).

Effect of action potentials on transmitter release

Despite the facilitation of the calcium currents, synaptic depression was observed in the presence of 50 μm fura-2 (Fig. 2A). In four of the five experiments, the second EPSC was smaller than the first one, in agreement with earlier results obtained in current clamp recordings or with intact terminals (Borst et al. 1995). The strong synaptic depression that was observed while the calcium currents were facilitated suggested that under these conditions a mechanism other than calcium current inactivation mediated synaptic depression.

We tested whether a large increase of the calcium current at the end of a 0.5 s, 100 Hz action potential waveform train could overcome synaptic depression. The first 45 action potentials had the same shape as the one in Fig. 2 or Fig. 3B, whereas for the last 5 action potentials of the train, a 1 ms plateau to +60 mV was added to the action potential to increase the Ca²⁺ influx (Borst & Sakmann, 1998). The large increase in the calcium current that was evoked by the plateau phase in the last 5 action potential waveforms resulted in only a small increase in the EPSC (n = 4, not shown).

Figure 3. Hodgkin-Huxley simulation of calcium currents.

A, to mimic the effects of a 2 ms prepulse to +60 mV or to 0 mV on the calcium currents, the voltage dependence of the rate constant (α_m) for channel opening was shifted by −2.2 and −4.0 mV, respectively (see Methods). As a result, the simulated isochronal I-V relationship was shifted by −1.6 mV (dotted line) and −2.8 mV (thin line) at around −20 mV compared with control (thick trace) and the current evoked by a step to 0 mV increased by 4.4 % and 7.7 %, respectively, similar to that observed experimentally. B, simulation of calcium currents during an action potential using the same three conditions as in A. Top panel, voltage template. Middle panel, calculated open probability (P_o) of the calcium channels, displayed as a fraction of the maximum open probability. Lower panel, simulated calcium currents.

Simulations

We performed a Hodgkin-Huxley simulation based on the measured activation and deactivation of the calyceal calcium channels (Borst & Sakmann, 1998), to investigate whether the small shift in the isochronal I-V relationships induced by a prepulse was sufficiently large enough to account for the facilitation during the action potential waveforms. To mimic the effects of a 2 ms prepulse to +60 or 0 mV on the calcium currents, the voltage-dependent rate constant (α_m) for opening was changed to produce a shift in the isochronal I-V relationships similar to that observed experimentally (Fig. 3A). As a result, the current evoked by a step to 0 mV also increased. No change in the isochronal I-V relationships was seen above +20 mV. A change in the voltage-dependent activation can therefore be responsible for the observed increase in the peak amplitude of the isochronal I-V relationship. Using these parameters, the peak amplitude of the simulated action potential-evoked calcium current increased by 5.8 and 10.3 % for α_m shifts of −2.2 and −4.0 mV, respectively (Fig. 3B). During the prepulses we used, the amount of Ca²⁺ that entered was comparable to what normally enters with 1–2 (prepulse to 60 mV) or 4 action potentials (prepulse to 0 mV). The measured increase in the calcium currents during an action potential waveform train in simultaneous pre- and postsynaptic voltage clamp recordings was 7.4 ± 1.4 % for the third and 9.3 ± 1.7 % for the fifth action potential in the train (Fig. 2C), similar to the increases that were observed in the simulations. This indicates that the observed small change in the I-V relationship is sufficiently large to account for the observed increase in the calcium current evoked by action potential waveforms.

DISCUSSION

Calcium currents in the calyx of Held were facilitated during action potentials. The underlying mechanism appeared to be an acceleration of the activation kinetics of the calcium channels at negative membrane potentials. The lack of an effect at voltages more positive than +10 mV suggested that there was no change in calcium channel conductance or maximal open probability. The effect of the hyperpolarizing shift in the activation of calcium channels was that the current during the repolarization phase of the action potentials increased. A simulation confirmed that the observed small shift in the isochronal I-V relationship was sufficient to explain the observed increases in the calcium current during action potential waveforms.

The reduction in the facilitation of the calcium currents during action potential waveforms in the presence of high concentrations of BAPTA and the positive correlation of the magnitude of facilitation during a test pulse with the amount of Ca²⁺ entry during a prepulse suggested that Ca²⁺ entry was important for calcium current facilitation. More experiments are needed to further elucidate the biochemical mechanism that is responsible for the facilitation, for example to investigate whether G-proteins are involved (Dolphin, 1996).

Facilitation of calcium currents during action-potential-like voltage commands has previously been observed for somatic calcium currents (Song & Surmeier, 1996) and cloned P/Q-type calcium channels (Brody et al. 1997). Facilitation of calcium currents has typically not been observed in nerve terminals (Charlton et al. 1982; Regehr & Atluri, 1995; Wright et al. 1996; David et al. 1997), although it was recently observed in the calyx of Held during brief step depolarizations (Forsythe et al. 1998). Activation of GABA_B receptors slows the activation time course of calyceal calcium channels (Takahashi et al. 1998; Wu et al. 1998). Facilitation of calcium currents may also be important to overcome the blocking effects of presynaptic G-protein-coupled receptors (Heidelberger & Matthews, 1991; Dolphin, 1996).

Despite its small size, the calcium current facilitation may have physiological relevance. Because of the strong dependence of transmitter release on Ca²⁺ influx, even small changes in Ca²⁺ influx may lead to appreciable changes in transmitter release. The facilitation of the calcium current may therefore contribute to synaptic facilitation, or it may counteract the processes that lead to synaptic depression.