Mechanisms underlying presynaptic facilitatory effect of cyclothiazide at the calyx of Held of juvenile rats

Abstract

1. Excitatory postsynaptic currents (EPSCs) were recorded using the whole-cell patch-clamp technique at the calyx of Held synapse in the medial nucleus of the trapezoid body (MNTB) in auditory brainstem slices from juvenile rats.

2. Bath application of cyclothiazide (CTZ, 100 μm) significantly increased the amplitude of EPSCs mediated by α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) and N-methyl-d-aspartate (NMDA) receptors. Cyclothiazide increased the magnitude of paired-pulse depression of both AMPA-EPSCs (intervals, 50 and 500 ms) and NMDA-EPSCs (interval, 20 ms). In low Ca²⁺, high Mg²⁺ solution, CTZ decreased the number of failures and increased the mean amplitude of AMPA-EPSCs more than three-fold.

3. Presynaptic Ca²⁺ currents and K⁺ currents were directly recorded from the calyceal nerve terminals. These currents were attenuated by CTZ in a reversible manner. The magnitude of inhibition of presynaptic K⁺ currents by CTZ (100 μm) was comparable to that by 5 μm 4-aminopyridine (4-AP). Both CTZ and 4-AP slowed the repolarizing phase of presynaptic action potentials.

4. The inhibitory effects of CTZ on presynaptic ion channels were mimicked by a solution having reduced Ca²⁺ concentration and 5 μm 4-AP. This solution facilitated EPSCs, but the magnitude of facilitation was significantly less than that caused by CTZ.

5. In the presence of tetrodotoxin (TTX), CTZ increased the mean frequency of miniature EPSCs three-fold. CTZ prolonged their decay time but had no effect on their amplitude. The facilitatory effect of CTZ on the miniature frequency was neither blocked by a protein kinase C inhibitor nor occluded by phorbol ester, suggesting that a distinct mechanism underlies the effect of CTZ.

6. We conclude that CTZ facilitates transmitter release through suppression of presynaptic potassium conductance and stimulation of exocytotic machinery downstream of Ca²⁺ influx.

The benzothiadiazide drug cyclothiazide (CTZ) is a potent blocker of desensitization of AMPA-type glutamate receptors (Yamada & Rothman, 1992; Patneau et al. 1993; Trussell et al. 1993; Yamada & Tang, 1993). However, CTZ also has a presynaptic facilitatory effect on transmitter release (Barnes-Davies & Forsythe, 1995; Diamond & Jahr, 1995; Dittman & Regehr, 1998; Bellingham & Walmsley, 1999). Considering the wide use of this compound for studies of synaptic transmission, it is important to re-examine its presynaptic effect and to clarify the underlying mechanism.

The calyx of Held is a giant nerve terminal of globular bushy cells in the anteroventral cochlear nucleus forming an axo-somatic synapse onto the principal cell in the MNTB. At this synapse, it is possible to record presynaptic ionic currents as well as the postsynaptic responses (Forsythe, 1994; Borst et al. 1995; Takahashi et al. 1996). Using this preparation, we have investigated the presynaptic mechanism underlying the effect of CTZ. Our results suggest that the suppression of presynaptic K⁺ conductance and stimulation of the exocytotic machinery underlie the presynaptic facilitatory effect of CTZ.

METHODS

Experiments were performed in accordance with the guidelines of the Physiological Society of Japan. Transverse brainstem slices (150-200 μm thick) containing the MNTB were prepared from 13- to 16-day-old Wistar rats, killed by decapitation under halothane anaesthesia (Forsythe & Barnes-Davies, 1993). The MNTB principal cells and calyces were visually identified using a × 60 or × 63 water immersion lens (Olympus or Zeiss) attached to an upright microscope (Axioskop, Zeiss). Each slice was superfused with an artificial cerebrospinal fluid (aCSF) containing (mm): NaCl 125, KCl 2.5, NaHCO₃ 26, glucose 10, NaH₂PO₄ 1.25, CaCl₂ 2, MgCl₂ 1, myo-inositol 3, sodium pyruvate 2 and ascorbic acid 0.5 (pH 7.4; 95 % O₂-5 % CO₂). For recording EPSCs, the superfusate routinely contained bicuculline methiodide (10 μm, Sigma) and strychnine hydrochloride (0.5 μm, Sigma) to block inhibitory synaptic responses. The patch pipette solution for postsynaptic recordings contained (mm): CsF 110, CsCl 30, Hepes 10, EGTA 5, and MgCl₂ 1. When TTX was not present in the aCSF, N-(2,6-diethylphenylcarbamoylmethyl)-triethyl-ammonium chloride (QX314, 5 mm) was included in the postsynaptic pipette solution to block action potential generation. For recording presynaptic Ca²⁺ currents, the aCSF contained TTX (1 μm) and tetraethylammonium (TEA) chloride (10 mm), and the pipette solution contained (mm): CsCl 110, Hepes 40, TEA-Cl 10 and EGTA 0.5. For recording presynaptic K⁺ currents, the aCSF contained TTX (1 μm) and the pipette solution contained (mm): potassium gluconate 97.5, KCl 32.5, Hepes 10 and EGTA 5. For recording presynaptic action potentials, the pipette solution contained (mm): potassium gluconate 87.5, KCl 32.5, Hepes 10, EGTA 0.5 and l-glutamic acid 10. The presynaptic pipette solutions also routinely contained (mm): MgCl₂ 1, sodium phosphocreatinine 12, ATP-Mg 2 and GTP 0.5. Cyclothiazide (Tocris, Bristol, UK) was dissolved in dimethyl sulphoxide (DMSO) to a final concentration of less than 0.3 % (usually 0.1 %), and the same concentration of DMSO was included in control solutions.

Excitatory postsynaptic currents were evoked at a holding potential of -70 mV (unless otherwise noted) using a platinum bipolar electrode positioned half-way between the midline and the MNTB. The EPSCs from the calyx of Held displayed a clear threshold and had amplitudes larger than 1 nA (Forsythe & Barnes-Davies, 1993). The EPSCs could be blocked by GYKI 52466 (100 μm), indicating that they are mediated by AMPA receptors. The access resistance of postsynaptic recordings was 6-15 MΩ (electrode resistance, 2-4 MΩ) and data were discarded if this value changed by > 30 %. The presynaptic access resistance was 8-20 MΩ (electrode resistance, 5-8 MΩ) and compensated by 70-80 % for current recordings under voltage clamp. Leak currents were subtracted for presynaptic Ca²⁺ and K⁺ currents by the scaled pulse (P/n) protocol (Cuttle et al. 1998). Currents were recorded using a patch-clamp amplifier (Axopatch-1D or Axopatch 200B, Axon Instruments). Presynaptic action potentials were recorded using a high impedance amplifier (IE-251A, Warner Instrument Corp., Hamden, CT, USA). The records were low-pass-filtered at 5 kHz and digitized at 20-50 kHz by Digidata 1200A analog-digital converter with pCLAMP7 software (Axon Instruments) or by a CED 1401 interface (Cambridge Electronic Design, Cambridge, UK). No correction was made for the liquid junction potentials between the pipette solution and the aCSF. Drugs were applied to the bath through superfusates flowing at 1-1.5 ml min⁻¹. All experiments were conducted at room temperature (23-29 °C). Values are given as means ±s.e.m., and the significance of difference was evaluated by the Student's paired two-tailed t test (unless otherwise noted). In the paired-pulse protocol with 10-50 ms intervals, the mean amplitude of the second EPSC was measured after subtracting the tail of the averaged single-pulse EPSC.

RESULTS

Effects of CTZ on EPSCs in normal solution

When CTZ (100 μm) was applied in normal aCSF, EPSCs gradually increased both in amplitude and decay time (Fig. 1a and B) as reported (Trussell et al. 1993; Barnes-Davies & Forsythe, 1995; Diamond & Jahr, 1995). The magnitude of potentiation was small (25.5 ± 5.5 %, n = 7 cells) but significant (P < 0.05). CTZ also increased the synaptic latency by 16.8 ± 3.4 % (n = 7, P < 0.005, Fig. 1B). At this synapse in rats of this age, a paired-pulse stimulation causes a depression of the second EPSC by 10-20 % for inter-pulse intervals ranging from 5 ms to 5 s (S. Iwasaki & T. Takahashi, unpublished observation). At a 500 ms interval, the amplitude of the second EPSC relative to the first (paired-pulse ratio) was 0.81 ± 0.02 (n = 7, Fig. 1B). At this interval, postsynaptic AMPA receptors completely recover from desensitization (Otis et al. 1996), therefore this paired-pulse depression (PPD) must arise presynaptically, presumably from partial depletion of readily releasable synaptic vesicles. When CTZ was applied, the paired-pulse ratio decreased in all cells examined to 0.78 ± 0.02 (n = 7, P < 0.05), suggesting that the drug has a presynaptic effect. Similarly, at a 50 ms interval, CTZ decreased the paired-pulse ratio from 0.86 ± 0.02 to 0.83 ± 0.02 (n = 5, P < 0.01). At a 10 ms interval, however, CTZ increased the paired-pulse ratio from 0.80 ± 0.12 to 0.85 ± 0.12 (n = 3). This facilitation might be caused by the postsynaptic effect of CTZ reducing the AMPA-receptor desensitization (Trussell et al. 1993; Otis et al. 1996), or through the presynaptic exocytotic mechanism postulated for PPD at the endbulb of Held synapse (Bellingham & Walmsley, 1999).

Figure 1. Facilitatory effect of CTZ on AMPA- and NMDA-EPSCs.

A, the effect of CTZ (100 μm) on AMPA-EPSCs in normal aCSF in the presence of (±)-2-amino-5-phosphono-pentanoic acid (d-AP5, 25 μm). The holding potential was set at -5 mV to reduce the EPSC amplitude, for better voltage-clamp performance. Data from 7 cells were normalized to the mean amplitude of the EPSCs before CTZ application (1.10 ± 0.22 nA). B, the effect of CTZ on the paired-pulse depression at a 500 ms interval (repeated at 0.05 Hz). Top traces, EPSCs (average of 18 events, superimposed) before and during CTZ application. Dotted lines indicate the amplitude of the first EPSC. Bottom traces, EPSCs normalized in amplitude to the 1st EPSCs. Note the delayed onset during CTZ application. C, the effect of CTZ on NMDA-EPSCs recorded at +20 mV in the presence of CNQX (20 μm). Data from 5 cells normalized to the mean amplitude of EPSCs before CTZ application (352 ± 91 pA). D, the effect of CTZ on the paired-pulse depression of NMDA-EPSCs at a 20 ms interval. Top traces in D, EPSCs (average of 12 events) before and during CTZ application (superimposed). Bottom traces in D, EPSCs normalized in amplitude to the 1st EPSC. Data points and bars indicate mean ±s.e.m.

To separate the presynaptic effect of CTZ from the postsynaptic effect on AMPA receptors, we recorded EPSCs mediated by NMDA receptors (NMDA-EPSCs) at a holding potential of +20 mV in the presence of 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX, 20 μm). CTZ significantly increased the amplitude of the NMDA-EPSC (32.8 ± 9.7 %, n = 5, P < 0.05; Fig. 1C), to a similar extent as with AMPA-EPSCs (see above, P > 0.5, two sample t test), further suggesting that the facilitatory effect of CTZ is presynaptic. In contrast to AMPA-EPSCs, the kinetics of NMDA-EPSCs were not affected by CTZ (Fig. 1D). CTZ decreased the paired-pulse ratio of NMDA-EPSCs at a 20 ms interval from 0.61 ± 0.06 to 0.48 ± 0.05 (n = 5, P < 0.001; Fig. 1D) as observed for AMPA-EPSCs. These results are consistent with those reported at the hippocampal (Diamond & Jahr, 1995) and cerebellar (Dittman & Regehr, 1998) synapses but are different from those reported at the endbulb of Held, where the paired-pulse ratio of NMDA-EPSCs at a 20 ms interval was increased by CTZ (Bellingham & Walmsley, 1999).

Effects of CTZ on EPSCs in low Ca²⁺ solution

The facilitatory effect of CTZ on EPSCs might be underestimated if the transmitter release mechanism or postsynaptic receptors were close to saturation in normal aCSF. Therefore we re-evaluated the effect of CTZ after reducing the release probability in low Ca²⁺ (0.5 mm), high Mg²⁺ (5.8 mm) solution (Fig. 2). In this solution, 61.6 ± 5.2 % of stimulation trials failed to elicit EPSCs and the mean amplitude of EPSCs was reduced to 20.2 ± 5.7 pA (including failures, n = 6). Application of CTZ markedly decreased the proportion of failures to 17.6 ± 6.2 % (P < 0.001, Fig. 2A, B and D) and increased the mean amplitude of EPSCs by 233 ± 13 % (P < 0.001; Fig. 2A–C), indicating that the drug increases the mean quantal content (del Castillo & Katz, 1954). The coefficient of variation (c.v.) was significantly decreased by the application of CTZ from 1.62 ± 0.19 to 0.738 ± 0.078 (n = 6, P < 0.001) as previously reported (Barnes-Davies & Forsythe, 1995). Thus, the potent presynaptic facilitatory effect of CTZ was revealed by lowering the release probability. These results suggest that relatively small facilitatory effects reported for CTZ in normal solutions (Diamond & Jahr, 1995; Isaacson & Walmsley, 1996; Dittman & Regehr, 1998; Dzubay & Jahr, 1999) may arise from a partial saturation of synaptic currents.

Figure 2. Facilitatory effect of CTZ on EPSCs at low release probability.

A and B, amplitude histograms of AMPA-EPSCs elicited at 1 Hz in the low Ca²⁺ (0.5 mm), high Mg²⁺ (5.8 mm) aCSF. Events < 10 pA were counted as failures and were fitted by a Gaussian distribution as background noise. A, control (975 events). B, 6-23 min after the application of CTZ (975 events). Sample traces are 10 consecutive EPSCs (superimposed). C and D, the effect of CTZ on the mean amplitude of EPSCs (C) and the proportion of failures (D) in 6 different cells.

Effects of CTZ on presynaptic Ca²⁺ currents

To understand the mechanism underlying the presynaptic effect of CTZ, we tested the effect of CTZ on presynaptic Ca²⁺ currents (I_pCa) recorded directly from the calyx of Held nerve terminals. The presynaptic Ca²⁺ currents in rats at this age have been identified as P/Q-type (Takahashi et al. 1996; Forsythe et al. 1998; Iwasaki & Takahashi, 1998). CTZ significantly suppressed I_pCa (Fig. 3) without appreciably affecting its voltage dependence. The mean magnitude of suppression was 23.2 ± 1.6 % at voltage steps to -5 mV (n = 6, ranging from 20.3 to 26.5 % between -30 and +25 mV, measured at 1-1.5 ms from the current onset). This effect was partially reversible after washout of the drug. Similarly, CTZ (100 μm) has been reported to attenuate L-type Ca²⁺ currents at the goldfish retinal bipolar cell terminals, albeit to a lesser extent (8 %, von Gersdorff et al. 1998).

Figure 3. CTZ suppressed presynaptic Ca²⁺ currents.

The effect of CTZ (100 μm) on presynaptic Ca²⁺ currents. Ca²⁺ currents in sample records were evoked by a depolarizing pulse from -80 mV holding potential to -10 mV (control, during CTZ application and after washout, superimposed). The current-voltage relationship of Ca²⁺ currents before (○), during CTZ application (•) and after washout (▵) recorded from 6 calyceal terminals. The amplitude of the Ca²⁺ current was measured 1-1.5 ms after the onset of the command pulse and normalized to control at 0 mV. Data points and bars indicate means ±s.e.m. Fitting lines are drawn by eye.

Effects of CTZ on presynaptic K⁺ currents

Next we examined the effect of CTZ on the presynaptic voltage-gated K⁺ currents, which are highly sensitive to 4-AP (Forsythe, 1994). CTZ significantly and reversibly reduced K⁺ currents (Fig. 4a). This effect of CTZ was concentration dependent, with a 50 % inhibitory concentration (IC₅₀) of about 14 μm, which is comparable to the EC₅₀ of CTZ in augmenting the AMPA receptor responses in hippocampal cells (9 μm, Patneau et al. 1993). The maximal effect of CTZ on K⁺ currents was found at around 100 μm (Fig. 4B). CTZ had no effect on the holding current. Given that CTZ attenuated presynaptic K⁺ currents, it might change the waveform of presynaptic action potentials. As illustrated in Fig. 4C, CTZ (100 μm) slightly but significantly slowed the repolarizing phase of the action potential, with the 50 % decay time being prolonged by 18.9 ± 6.4 % (n = 5, P < 0.05). However, CTZ had no effect on the peak amplitude of the presynaptic action potential (0.0 ± 1.9 %) or the resting potential (0.0 ± 0.6 %). These results suggest that CTZ prolongs presynaptic action potential duration through attenuation of presynaptic K⁺ currents, thereby facilitating transmitter release.

Figure 4. CTZ suppressed presynaptic K⁺ currents and slowed the repolarization of presynaptic action potentials.

A, the effect of CTZ on presynaptic K⁺ currents. K⁺ currents in sample records were evoked by depolarizing a calyx from -80 to 0 mV. For the current-voltage relationship, the mean K⁺ current amplitude was measured at 2-3 ms from the onset of the command pulse and normalized to control (○) at 0 mV (5 calyces). •, suppression of K⁺ currents during CTZ application; ▵, recovery after washout. Fitting lines drawn by eye. B, concentration dependence of CTZ-induced inhibition of K⁺ currents evoked by depolarization to -20 mV. Data from 4-7 experiments at each dose (10-300 μm) were normalized to the amplitude before CTZ application. A curve fitted to data points is drawn according to the equation: magnitude of K⁺ current inhibition (%) = maximal K⁺ current inhibition (%)/[1 + (IC₅₀/CTZ concentration)ⁿ], where the maximal inhibition was 36.7 %, IC₅₀ was 13.9 μm and the Hill coefficient n was 1.7. C, action potentials recorded from a calyceal terminal (average of 10 traces), before and during application of CTZ. In control, the mean resting potential was -65.8 ± 1.4 mV, and the mean peak amplitude and 50 % decay time of the action potential were 68.3 ± 8.3 mV and 0.27 ± 0.02 ms, respectively (mean ±s.e.m.; n = 5). Data points and bars indicate mean ±s.e.m.

Experimental simulation of effects of CTZ

Suppression of presynaptic Ca²⁺ and K⁺ currents by CTZ should affect transmitter release in opposite directions. Suppression of presynaptic Ca²⁺ currents will reduce EPSCs and this effect can be mimicked by reducing [Ca²⁺]_o. Suppression of presynaptic K⁺ currents will increase EPSCs and this can be mimicked by 4-AP. A low concentration of 4-AP (5 μm) clearly attenuated the presynaptic K⁺ current (Fig. 5a), as reported for higher concentrations (200 μm, Forsythe, 1994). The magnitude of suppression was concentration dependent, blocking 50 % of presynaptic K⁺ currents at around 10 μm (Fig. 5B). The voltage dependence of the effect of 4-AP on the presynaptic K⁺ current was slightly different from that of CTZ, with stronger inhibition at a more positive membrane potential (Fig. 5C). At -20 mV, the inhibitory effect of 100 μm CTZ was comparable to that of 5 μm 4-AP. The 50 % decay time of the presynaptic action potential was prolonged by 4-AP by 40.4 ± 18.2 % at 5 μm (n = 3) and 137 ± 53 % at 20 μm (n = 3) without appreciably affecting the peak amplitude or resting potential (data not shown).

Figure 5. Effect of 4-AP on presynaptic K⁺ currents and EPSCs in experimental simulation.

A, the effect of 4-AP (5 μm) on presynaptic K⁺ currents (n = 3). B, concentration-dependent effect of 4-AP on presynaptic K⁺ currents evoked by depolarization to -20 mV. 4-AP was applied sequentially from low to high concentrations (5-20 μm) at 4 calyces. The effect of 4-AP was reproducible by repeated applications (not shown). A dotted line indicates the mean magnitude of suppression by 100 μm CTZ (Fig. 4B). C, voltage dependence of the inhibitory effects of 100 μm CTZ (○) and 5 μm 4-AP (•) on K⁺ currents. The mean K⁺ current amplitude during drug applications relative to the control is plotted against membrane potential. Data derived from Fig. 4A (n = 5) and Fig. 5A (n = 3). D and E, the relative changes in the EPSC amplitude (D) and the rate of failures (E) during application of the test solution (5 μm 4-AP and reduced [Ca²⁺], ▪, n = 6) and those during CTZ application (□, data taken from Fig. 2C and D and normalized). Sample records in D are averaged EPSCs (of 50 events) before and during application of the test solution (superimposed). Data points and bars indicate mean ±s.e.m. Fitting lines drawn by eye.

Subsequently we compared the facilitatory effects of CTZ (100 μm) and 4-AP (5 μm) on EPSCs. To mimic the inhibitory effect of CTZ on I_pCa (24 % reduction), we simply reduced [Ca²⁺]_o from 0.5 to 0.38 mm because the amplitude of I_pCa is in linear proportion to the external Ca²⁺ concentration below 1 mm (Takahashi et al. 1996; Borst & Helmchen, 1998). When the superfusate was switched from control (0.5 mm Ca²⁺) to the test solution (0.38 mm Ca²⁺ and 5 μm 4-AP), EPSCs increased in amplitude by 133 ± 34 % (n = 6; Fig. 5D) and the rate of failures decreased to 58.0 ± 7.6 % (Fig. 5E). Also, the synaptic latency appeared slightly increased (Fig. 5D) as observed during CTZ application (Fig. 1B). Thus the facilitatory effect of 4-AP (5 μm) surpassed the inhibitory effect of [Ca²⁺]_o reduction and mimicked the effect of CTZ. Compared with CTZ, however, the magnitude of potentiation by the test solution was significantly smaller (P < 0.05, two sample t test). These results suggest that the presynaptic facilitatory effect of CTZ is mediated by its attenuating effect on the presynaptic K⁺ currents, and additionally by a mechanism involving the exocytotic machinery downstream of Ca²⁺ influx.

Effects of CTZ on miniature EPSCs

As reported for cultured hippocampal neurons (Yamada & Tang, 1993; Diamond & Jahr, 1995; Mennerick & Zorumski, 1995), CTZ increased the frequency of spontaneous miniature (m) EPSCs (0.55 ± 0.13 Hz, n = 6) recorded in the presence of TTX by 234 ± 74 % (P < 0.05; Fig. 6a). But a test solution containing 4-AP (5 μm) with reduced [Ca²⁺]_o (by 24 % to 1.52 mm) had no effect on the mEPSC frequency (98.5 ± 20.2 %, n = 6). At this auditory synapse, CTZ had no effect on the mEPSC amplitude (Fig. 6B) as at the endbulb of Held synapse (Isaacson & Walmsley, 1996), whereas it is reported to increase the mEPSC amplitude at the hippocampal synapses (Yamada & Tang, 1993; Diamond & Jahr, 1995; Mennerick & Zorumski, 1995). The decay time of mEPSCs (time constant, 0.87 ± 0.13 ms) was prolonged in the presence of CTZ (1.67 ± 0.16 ms, P < 0.05, inset of Fig. 6B) as previously reported (Yamada & Tang, 1993; Diamond & Jahr, 1995).

Figure 6. The effect of CTZ on miniature EPSCs.

A, cumulative interval histograms of mEPSCs recorded in the presence of TTX. The frequency of mEPSCs was increased by CTZ. Sample traces are 20 consecutive traces superimposed. B, cumulative amplitude histograms of mEPSC from the same cell. Sample traces are averaged mEPSCs (of 175 and 375 events) before and during CTZ application superimposed. Data in histograms were sampled 6-12 min after CTZ application. Similar results were obtained from 5 other cells. The mean amplitudes of mEPSCs before and during CTZ application were 32.0 ± 3.5 pA and 31.2 ± 3.7 pA, respectively (n = 6).

At this synapse, phorbol ester facilitates transmitter release, through activation of protein kinase C (PKC) and the Doc2α-Munc13-1 interaction, which directly activates the exocytotic machinery downstream of Ca²⁺ influx (Hori et al. 1999). This effect of phorbol ester can be largely inhibited by the PKC inhibitor bisindolylmaleimide (BIS). We examined whether CTZ shares a common mechanism with phorbol ester in synaptic facilitation. In the presence of BIS (1 μm), CTZ increased the mEPSC frequency by 164 ± 60 % (n = 6, P < 0.05). This magnitude of increase is not significantly different from control (P > 0.4, two sample t test), suggesting that PKC is probably not involved in the effect of CTZ. We next examined whether phorbol ester might occlude the CTZ effect. In the presence of phorbol 1,2-dibutrate at its saturating concentration (0.5 μm, Hori et al. 1999), CTZ (100 μm) still significantly facilitated the mEPSC frequency (by 101 ± 17 %, n = 5). These results suggest that the mechanism underlying the facilitatory effect of CTZ may be distinct from that of phorbol ester.

DISCUSSION

A presynaptic facilitatory effect of CTZ has been suggested from various kinds of observations including the decrease in the paired-pulse ratio (Diamond & Jahr, 1995; Dittman & Regehr, 1998), the increase in the mEPSC frequency (Diamond & Jahr, 1995; Mennerick & Zorumski, 1995), the decreased proportion of failures (Diamond & Jahr, 1995) and the increase in the c.v. (Barnes-Davies & Forsythe, 1995). Our present study confirmed these reports, and further demonstrated that the presynaptic effect of CTZ is substantial and mediated by multiple mechanisms. One important mechanism is its inhibitory effect on presynaptic K⁺ currents. CTZ also attenuates presynaptic Ca²⁺ currents, but this effect is surpassed by that on K⁺ currents. Suppression of K⁺ currents by CTZ slowed the repolarizing phase of presynaptic action potentials. This will increase Ca²⁺ influx through voltage-dependent Ca²⁺ channels (Augustine, 1991), thereby increasing transmitter release. Repolarization of nerve terminals following depolarization provides a large driving force for Ca²⁺ influx (Katz & Miledi, 1971). Therefore, slowing of the repolarization phase of the action potential by CTZ or 4-AP would delay the presynaptic Ca²⁺ influx, thereby increasing the synaptic latency, as reported for the effect of 3,4-diaminopyridine on the synaptic latency at the squid giant synapse (Augustine, 1991).

Our results show that 4-AP at 10 μm inhibited 50 % of presynaptic K⁺ currents at the calyx of Held. This sensitivity is comparable to that reported for the K⁺ currents at mice cerebellar basket cell terminals (Southan & Robertson, 2000), suggesting that a common type of potassium channel might be expressed at these nerve terminals. The inhibitory effects of CTZ on presynaptic K⁺ and Ca²⁺ currents were not strongly dependent upon voltage, suggesting that the drug may not affect the voltage sensors of the channels. Details of the mechanisms by which CTZ inhibits the voltage-dependent channel currents remain to be studied. CTZ had no effect on either the peak amplitude of the action potential or the resting potential, suggesting that it does not affect voltage-dependent sodium channels or inwardly rectifying potassium channels, the latter of which contribute to the resting membrane conductance at this nerve terminal (Takahashi et al. 1998).

Although inhibition of presynaptic K⁺ currents by CTZ was substantial, this did not fully explain the facilitatory effects of CTZ on evoked and spontaneous transmitter release, suggesting that the exocytotic machinery downstream of Ca²⁺ influx is additionally involved in the presynaptic effect of CTZ. Similar effects are seen with phorbol ester, which facilitates both evoked and spontaneous release of transmitter by directly affecting the downstream mechanism through activation of PKC and the Doc2α-Munc13-1 interaction (Hori et al. 1999). However, the lack of effect of the PKC antagonist BIS suggests that PKC is not involved in the effect of CTZ. Also, the failure of a saturating concentration of phorbol ester to occlude the CTZ effect argues against the possibility that CTZ shares a common target with phorbol ester. Although the target of CTZ in the exocytotic machinery is unknown, our present results suggest that synaptic facilitation can be caused by multiple exocytotic mechanisms at this synapse.

At the endbulb of Held, CTZ increased the paired-pulse ratio of both AMPA- and NMDA-EPSCs at short inter-pulse intervals (< 50 ms, Bellingham & Walmsley, 1999). In contrast, at the calyx of Held CTZ decreased the paired-pulse ratios of AMPA- and NMDA-EPSCs at 50 and 20 ms inter-pulse intervals, respectively. Thus, our results at the calyx of Held are different from those reported at the endbulb of Held. The PPD can be caused by multiple mechanisms including depletion of synaptic vesicles (Kusano & Landau, 1975), Ca²⁺-dependent adaptation of exocytosis (Hsu et al. 1996), inactivation of presynaptic Ca²⁺ currents (Forsythe et al. 1998), presynaptic conduction failures (Brody & Yue, 2000) and postsynaptic receptor desensitization (Trussell et al. 1993; Otis et al. 1996). Which of these factors is responsible for the PPD may depend upon inter-pulse intervals within a synapse and may also differ among synapses. Bellingham & Walmsley (1999) suggested that their results might arise from the adaptation of exocytosis. This mechanism, however, is unlikely to be involved in the PPD at the calyx of Held synapse.

In conclusion, beside its inhibitory effect on desensitization of AMPA receptors, CTZ attenuates presynaptic Ca²⁺ and K⁺ channels and also directly stimulates the exocytotic machinery. These effects, in combination, facilitate the synaptic transmission at the calyx of Held synapse.