Presynaptic Release Probability and Readily Releasable Pool Size Are Regulated by Two Independent Mechanisms during Posttetanic Potentiation at the Calyx of Held Synapse

Abstract

At the immature calyx of Held, the fast decay phase of a Ca²⁺ transient induced by tetanic stimulation (TS) was followed by a period of elevated [Ca²⁺]_i for tens of seconds, referred to as posttetanic residual calcium (Ca_res). We investigated the source of Ca_res and its contribution to posttetanic potentiation (PTP). After TS (100 Hz for 4 s), posttetanic Ca_res at the calyx of Held was largely abolished by tetraphenylphosphonium (TPP⁺) or Ru360, which inhibit mitochondrial Na⁺-dependent Ca²⁺ efflux and Ca²⁺ uniporter, respectively. Whereas the control PTP lasted longer than Ca_res, inhibition of Ca_res by TPP⁺ resulted in preferential suppression of the early phase of PTP, the decay time course of which well matched with that of Ca_res. TS induced significant increases in release probability (P_r) and the size of the readily releasable pool (RRP), which were estimated from plots of cumulative EPSC amplitudes. TPP⁺ or Ru360 suppressed the posttetanic increase in P_r, whereas it had little effect on the increase in RRP size. Moreover, the posttetanic increase in P_r, but not in RRP size, showed a linear correlation with the amount of Ca_res. In contrast, myosin light chain kinase (MLCK) inhibitors and blebbistatin reduced the posttetanic increase in RRP size with no effect on the increase in P_r. Application of TPP⁺ in the presence of MLCK inhibitor peptide caused further suppression of PTP. These findings suggest that Ca_res released from mitochondria and activation of MLCK are primarily responsible for the increase in P_r and that in the RRP size, respectively.

Introduction

Posttetanic potentiation (PTP) is a transient enhancement of synaptic efficacy that lasts tens of seconds after a cessation of tetanic stimulation (TS). It is generally accepted that PTP is presynaptic in origin (for review, see Zucker and Regehr, 2002). The posttetanic residual calcium (Ca_res) has been a long-standing hypothesis for PTP mechanisms (Kamiya and Zucker, 1994). A linear relationship between the magnitudes of PTP and Ca_res concentration has been demonstrated at the calyx of Held (Habets and Borst, 2005; Korogod et al., 2005). Although a close temporal correlation between posttetanic recovery of presynaptic Ca²⁺ concentration ([Ca²⁺]_i) and synaptic enhancement during PTP has been shown at crayfish neuromuscular junctions (NMJs) (Delaney et al., 1989; Delaney and Tank, 1994), it has been recognized that the time course of [Ca²⁺]_i decay is faster than that of PTP at mammalian central synapses, suggesting that PTP is not solely dictated by Ca_res (Regehr et al., 1994; Korogod et al., 2005) (but see Habets and Borst, 2005). To account for the temporal mismatch between the Ca_res and PTP, a calcium binding site with slow kinetics has been assumed, but its identity remains elusive (Regehr et al., 1994).

Presynaptic release of synaptic vesicles is largely determined by both the readily releasable pool (RRP) size and release probability (P_r). Recently, it has been shown that increases in both P_r and the RRP size contribute to PTP at the calyx of Held synapse and that the increase in RRP size lasts longer than that in P_r (Habets and Borst, 2007). Interestingly, submicromolar elevation of basal [Ca²⁺]_i (<100 nm) induced by a small depolarization of the calyceal terminal caused a significant increase in P_r with no effect on the RRP size (Awatramani et al., 2005), implying that Ca_res may have differential effects on the posttetanic increase in P_r and in the RRP size. We investigated the source of Ca_res at the calyx of Held, and assessed contribution of Ca_res to the posttetanic increases in P_r and in the RRP size, which were estimated from the plot of cumulative EPSC amplitudes (Schneggenburger et al., 1999). We found that Ca_res is generated mainly by posttetanic release from mitochondria at the calyx of Held. When mitochondrial Ca²⁺ release was inhibited to reduce Ca_res, the increase in P_r, but not the RRP size, was affected. In contrast, inhibitors of myosin light chain kinase (MLCK) reduced specifically the posttetanic increase in the RRP size.

Materials and Methods

Preparation of brainstem slices.

Transverse 180- to 200-μm-thick brainstem slices containing the medial nucleus of trapezoid body (MNTB) were prepared from 6- to 8-d-old Sprague Dawley rats (17 ± 8 g). Rats were decapitated, and brainstems were chilled in ice-cold low-calcium artificial CSF (aCSF), which contained (in mm) 125 NaCl, 25 NaHCO₃, 2.5 KCl, 1.25 NaH₂PO₄, 2.5 MgCl₂, 0.5 CaCl₂, 25 glucose, 0.4 Na ascorbate, 3 myoinositol, and 2 Na pyruvate [pH 7.4 when saturated with carbogen (95% O₂, 5% CO₂); osmolarity, ∼320 mOsm]. Brainstem slices were made with a vibratome slicer (VT1000S; Leica), and slices were incubated at 37°C for 30 min in normal aCSF, the constituents of which are the same as low-calcium aCSF except 1 mm MgCl₂ and 2 mm CaCl₂, and thereafter maintained at room temperature (23–25°C) until required.

Electrophysiological recording and reagents.

Slices were transferred to a recording chamber in an upright microscope (BX50WI; Olympus). Whole-cell patch-clamp recordings from calyces of Held were made under visual control using differential interference illumination. In most experiments, to reduce saturation and desensitization of AMPA receptors, we added 1 mm kynurenic acid (Kyn; Tocris) or 4 mm γ-d-glutamylglycine (γ-DGG) to aCSF, which were denoted as “aCSF + Kyn” or “aCSF + γ-DGG,” respectively, in Results. To preload calyx terminals with Ca²⁺ indicator dye, we made a brief presynaptic whole-cell recording for ∼90 s with the K-gluconate-containing pipette solution composed of (in mm) 120 K-gluconate, 30 KCl, 20 HEPES, 4 Na₂-phosphocreatine, 4 Mg-ATP, 0.3 Na-GTP, and 0.2 fura-4F. The “preloading” technique was also used for delivering lipid-soluble drugs into the cytosol of calyx terminals, because it took too long a time (>50 min) for lipid-soluble drugs to take an effect, and drug effects were often too variable, when they were applied to the bath. Therefore, lipid-soluble drugs including blebbistatin, chelerythrine, Bis-I, and Bis-V were dissolved in dimethylsulfoxide (DMSO) and added to the internal solution such that final concentration of DMSO became 1:1000 (v/v). In these cases, the synapse whose calyx terminal was preloaded with vehicle only (DMSO) was regarded as a control. Postsynaptic recordings were made in voltage-clamp mode at a holding potential of −70 mV. Series resistance (R_s; range, 4–15 MΩ) was regularly checked and partially compensated by up to 75% for remaining uncompensated R_s value of 3–4 MΩ. Experiments were discarded when uncompensated R_s exceeded 15 MΩ. Patch pipettes with a resistance of 3.5–4.5 MΩ were used for recordings. Postsynaptic patch pipettes contained the following (in mm): 110 Cs-gluconate, 20 tetraethylammonium-Cl, 10 NaCl, 20 HEPES, 10 EGTA, 4 Mg-ATP, 5 Na₂-phosphocreatine, 5 lidocaine N-ethyl bromide (QX314; an intracellular blocker of Na⁺ currents), pH 7.3 (adjusted with CsOH). Recordings were made at room temperature (23–25°C) with an EPC10 amplifier (HEKA). EPSCs were evoked by stimulating presynaptic axons (0.1–0.2 ms, 10–20 V) with a bipolar stimulating electrode (TM53CCINSl; World Precision Instruments) placed at the midline of the brainstem. Fura-4F was purchased from Invitrogen. Ru360, ML-7, myosin light chain kinase inhibitor peptide 18 (MLCKip), Ro31-8220, Bis-1, and Bis-5 were obtained from Merck Biosciences. Kynurenate, γ-DGG, QX314, and cyclothiazide were purchased from Tocris. Blebbistatin was from Biomol. Other reagents, including tetraphenylphosphonium (TPP⁺), were purchased from Sigma.

Presynaptic [Ca²⁺] imaging.

Calyces of Held were preloaded with 200 μm fura-4F for typically 90 s via presynaptic whole-cell recording with the K-gluconate solution. After withdrawing the patch pipette, [Ca²⁺]_i at the calyx of Held was measured as previously described (Kim et al., 2005). Briefly, excitation light from a monochromator (Polychrome-II; TILL Photonics) was delivered to an upright microscope (BX50; Olympus). Imaging was performed using a 60× water-immersion objective (numerical aperture, 0.9; LUMPlanFl; Olympus) and an air-cooled slow-scan CCD camera (SensiCam; PCO) using on-chip binning (8 × 16 pixels). The fluorescence ratio (r = F_iso/F₃₈₀) at the isosbestic wavelength (360 nm; F_iso) to that at 380 nm (F₃₈₀) was converted to [Ca²⁺]_i according to the following equation:

Calibration parameters were determined by “in-cell” calibration. R_min values were measured using a Ca²⁺-free internal solution containing 10 mm BAPTA. R_max values were obtained from in vitro measurement, because calyces of Held did not endure internal dialysis with high CaCl₂ (10 mm). The values for the calibration ratio at intermediate [Ca²⁺]_i were measured in the calyx using a pipette solution containing 8 mm BAPTA and 6 mm CaCl₂ ([Ca²⁺]_i ≈ 660 nm) for fura-4F. The effective dissociation constant of fura-4F (K_eff) was calculated by measuring the fluorescence ratio at these intermediate [Ca²⁺]_i and by rearranging Equation 1 for K_eff. The K_d values of fura-4F were calculated as 0.75 μm from K_d = K_eff × (α + R_min)/(α + R_max), where α is the iso-coefficient. All the estimates of presynaptic [Ca²⁺] were presented as an excursion of [Ca²⁺]_i (Δ[Ca²⁺]_i) from its resting level ranging from 100 to 200 nm. The decay phase of a Ca²⁺ transient induced by TS (100 Hz for 4 s) is characterized by a period of slow [Ca²⁺] decay that lasted tens of seconds. Henceforth, we define posttetanic residual [Ca²⁺] as Δ[Ca²⁺]_i during the slow decay phase after a fast decay phase of a Ca²⁺ transient induced by TS. The amount of Ca_res was quantified as the time integral of Δ[Ca²⁺]_i from 2 s after the end of TS for a period two times longer than the decay time constant of residual [Ca²⁺]_i (denoted by ∫Ca_resdt).

Statistics and data presentation.

Data are presented as mean ± SEM, with n indicating the number of synapses analyzed. In figures, the number of synapses or calyceal terminals that we studied is indicated as a number in parentheses on each bar. In Results, the first and the second statistical values in parentheses intervened by “vs” represent statistical data under control conditions and under test conditions, respectively. Statistical analyses on data measured at the same synapse and on data at different synapses were performed using the Student's paired t test and the Student's t test, respectively.

Results

PTP results from increases in both the RRP size and P_r at the calyx of Held synapse

We investigated PTP of EPSCs evoked by afferent fiber stimulation at the immature calyx of Held synapse of the rat (postnatal day 6–8). Unless otherwise mentioned, we monitored EPSCs evoked by paired stimuli (interstimulus interval 20 ms) every 10 s in the presence of 1 mm Kyn to prevent desensitization and saturation of postsynaptic AMPA receptors. The mean basal EPSC amplitude was 1.53 ± 0.20 nA (n = 11), which corresponds to 40 ± 4.6% of that in the absence of 1 mm Kyn. At 20 s after TS (100 Hz train of 4 s duration), the EPSC amplitude increased by 100.33 ± 10.6% (n = 19) of control values (Fig. 1E), and the paired-pulse ratio was reduced from 0.91 ± 0.05 to 0.42 ± 0.03 (p < 0.01) (Fig. 1A), indicative of a presynaptic locus of the PTP expression. When PTP was induced repeatedly at the same synapse, no progressive decline of PTP was observed. We assessed TS-induced relative changes in the RRP size and P_r by comparing a plot of cumulative amplitudes of first 20 EPSCs evoked by TS with that of EPSCs evoked by short trains of APs (100 Hz, 20 pulses) applied 20 s after TS (Fig. 1C,D). TS increased both estimates for the RRP size and P_r by 16.5 ± 1.2% and by 70.6 ± 9.0% of control values, respectively (n = 19; p < 0.01, paired t test) (Fig. 1E).

Figure 1.

Changes in the RRP size and P_r during posttetanic potentiation at the calyx of Held synapse. A, EPSCs in response to a pair of fiber stimulation separated by 20 ms (aCSF + Kyn). Baseline EPSC (black) and EPSC recorded after TS (100 Hz for 4 s; gray) are superimposed. B, Time course for PTP of EPSC amplitudes. Time points of TS and stimulation with a short train of action potentials (short AP trains; 100 Hz, 20 pulses) are indicated by black and gray triangles, respectively. C, First 10 EPSCs induced by TS (black) or by short AP trains (gray) applied 20 s after the TS. D, Plot of cumulative (cum.) EPSC amplitudes evoked by TS (black circles) or by a short AP train shown in C. RRP size was estimated at the y-intercept of the back extrapolation line fitted to the last five points. P_r was estimated from the first EPSC amplitude divided by the estimate of RRP size. Both P_r and RRP were increased to 180% and 116% of control values, respectively. E, Mean values for PTP and associated changes in P_r and the RRP size measured 20 s after TS in the presence of 1 mm Kyn (n = 19), 2 mm Kyn (n = 5), or 4 mm γ-DGG (n = 11). Data are mean ± SEM.

To rigorously test whether the saturation and desensitization of AMPA receptors affected our estimates of relative changes in P_r and the RRP size, we repeated experiments in the presence of 2 mm Kyn or 4 mm γ-DGG instead of 1 mm Kyn (Wong et al., 2003). All estimates for the RRP size and P_r and their changes after PTP induction under the different pharmacological conditions are summarized in supplemental Table 1 (available at www.jneurosci.org as supplemental material). We found that relative changes in the RRP size and P_r in the presence of 1 mm Kyn were not different from those in the presence of 2 mm Kyn or 4 mm γ-DGG (Fig. 1E). Therefore, we henceforth presented the statistical data after pooling data obtained in the presence of 1 mm Kyn and those in the presence of 4 mm γ-DGG.

Effects of mitochondrial Na⁺-dependent Ca²⁺ efflux inhibitor, TPP⁺, on Ca_res and PTP

It has been demonstrated that the magnitude of PTP is linearly correlated with that of residual [Ca²⁺]_i at the calyx of Held synapse (Habets and Borst, 2005; Korogod et al., 2005), but the source of Ca_res is not known at this synapse. Mitochondria have been suggested as a source of Ca_res at the motor nerve terminals and at large terminals of hippocampal mossy fibers (Tang and Zucker, 1997; García-Chacón et al., 2006; Lee et al., 2007). TPP⁺ at low concentration (IC₅₀, 0.2 μm) (Wingrove and Gunter, 1986) is known as a potent and selective inhibitor of mitochondrial Na⁺-dependent Ca²⁺ efflux (NCE) with little adverse effects on mitochondrial respiration (Aiuchi et al., 1985). To test whether mitochondrial NCE contributes to presynaptic Ca_res, we studied the effects of 2 μm TPP⁺ on presynaptic Ca_res. For simultaneous recording of presynaptic [Ca²⁺] and EPSCs, we loaded the calyx with Ca²⁺-indicator dye via whole-cell patch pipette containing 200 μm fura-4F for a brief period (∼90 s), and pulled off the pipette. Because of the fluorescence of Kyn at UV excitation, when combined with Ca²⁺ imaging, EPSCs were recorded in the presence of 4 mm γ-DGG.

We compared the time courses of PTP and residual Δ[Ca²⁺] before and after bath application of 2 μm TPP⁺ at the same synapse (Fig. 2A). TPP⁺ had little effect on the basal transmission (0.52 ± 0.15 vs 0.49 ± 0.15 nA; n = 5; p = 0.11, paired t test), but significantly reduced PTP from 131.4 ± 33.6 to 53.4 ± 10.3% (n = 5; p = 0.04) (Figs. 2A, 3Ba). Concomitantly, TPP⁺ essentially abolished posttetanic Ca_res (Fig. 2A,D), whereas the Ca²⁺ buildup during TS was not affected (Fig. 2B,C). In control conditions, there was a significant temporal mismatch between the decay time courses of PTP and accompanying residual Δ[Ca²⁺]_i (Fig. 2A). The decay of TPP⁺-sensitive PTP (PTP in control conditions minus that in the presence of TPP⁺), however, occurred with a time course similar to the decay of residual Δ[Ca²⁺]_i (Fig. 2E,F), indicative of a linear relationship between Ca_res and TPP⁺-sensitive PTP.

Figure 2.

Effects of mitochondrial NCE inhibitor, TPP⁺, on PTP and presynaptic residual calcium (Ca_res) at the calyx of Held synapse. A, Time-dependent changes in EPSC amplitude (open circles, left ordinate) and in presynaptic Ca_res (gray dots; calibration: 200 nm) simultaneously recorded at the same synapse (aCSF + γ-DGG). TPP⁺ reduced the time integrals of Ca_res from 3.89 to 0.46 μm · s. B, Ca²⁺ transients during TS with expanded time scale (black, control; gray, TPP⁺). C, D, Mean values for peak [Ca²⁺] of Ca²⁺ transients (C) and time integral of Ca_res (D) induced by TS before and after bath application of TPP⁺. The estimate for ∫Ca_resdt was reduced from 2.03 ± 0.50 to 0.37 ± 0.07 μm · s (n = 5; p = 0.03), whereas the mean peak [Ca²⁺] was little affected (2.77 ± 0.63 μm and 2.75 ± 0.64 μm before and after application of TPP⁺, respectively; n = 5; p = 0.90). E, TPP⁺-sensitive PTP (solid line) and Ca_res (gray dots), calculated from A, decayed with a similar time course (dotted line, an exponential fit to TPP⁺-sensitive PTP with τ = 39.54 s; τ for Ca_res = 35.78 s). F, Mean values for decay time constants (τ) of TPP⁺-sensitive PTP (τ = 23.66 ± 4.34 s; n = 5) and residual Δ[Ca²⁺]_i (τ = 24.61 ± 4.56 s; n = 5). On the bar graphs (C, D, and F), data points measured from the same synapse are connected with a dotted line.

Figure 3.

Effects of TPP⁺ on posttetanic increases in P_r and in the RRP size. Aa, Ac, First 10 EPSCs induced by TS (black) or by a short AP train (gray) applied 20 s after the TS under control conditions (aCSF + γ-DGG; Aa) and in the presence of 2 μm TPP⁺ (Ac). Ab, Ad, Plots of cumulative (cum.) EPSC amplitudes recorded before (Ab) and after (Ad) the application of TPP⁺. EPSCs were recorded from the same synapse as in Figure 2A. The time points when these plots were obtained are indicated by triangles in Figure 2A (black, TS; gray, 20 s after TS). The posttetanic increase in P_r was reduced from 180.9 to 28.3% by TPP⁺, whereas the increase in RRP size was little affected (from 25.2 to 29.8%). Ba–Bc, Summary of the effects of TPP⁺ on PTP (Ba) and increases in P_r (Bb) and in the RRP size (Bc) under the “aCSF + γDGG” condition. Dots connected with a dotted line indicate data obtained from the same synapse. C, D, Plot of the posttetanic increase in P_r (C) and in RRP size (D) as a function of residual Δ[Ca²⁺]. The increase in P_r was linearly correlated with Ca_res (r = 0.91). Ea–Ec, Mean values for PTP (Ea) and increases in P_r (Eb) and in the RRP size (Ec) measured at the calyx of Held synapse, whose presynaptic terminal had not been preloaded with Ca²⁺ indicator dye (ctrl, under control conditions, aCSF + Kyn (n = 5) or γ-DGG (n = 6); TPP, in the presence of 2 μm TPP⁺).

Next, we investigated which component of PTP, when comparing increases in P_r and in the RRP size, is affected by a reduction of ∫Ca_resdt. TPP⁺ significantly suppressed the posttetanic increase in P_r (94.3 ± 25.7 vs 26.9 ± 6.4%; n = 5; p = 0.04) (Fig. 3Bb), whereas it had little effect on that in the RRP size (18.7 ± 2.4 vs 19.6 ± 3.2%; p = 0.63) (Fig. 3Bc). The posttetanic increase in P_r, but not in the RRP size, showed a linear relationship with residual Δ[Ca²⁺]_i (r = 0.91) (Fig. 3C,D).

Saturation of intracellular Ca²⁺ buffer can greatly influence the release probability and thus short-term plasticity (Felmy et al., 2003). To test whether the presence of exogenous Ca²⁺ buffer exaggerates PTP by the buffer saturation mechanism (Rozov et al., 2001), we assessed the effects of TPP⁺ on PTP under the conditions not disturbed by Ca²⁺-indicator dye and in the presence of 1 mm Kyn or 4 mm γ-DGG. TPP⁺ showed similar effects on the posttetanic increases in P_r and the RRP size, regardless of presynaptic loading with Ca²⁺-indicator dye (Fig. 3E). TPP⁺ suppressed PTP from 104.8 ± 10.1 to 52.3 ± 7.6% [n = 11; p < 0.01, paired t test; aCSF + Kyn (n = 5) or γ-DGG (n = 6)]. TPP⁺ exerts its effect on PTP by suppressing the posttetanic increase in P_r (78.6 ± 8.3 vs 37.1 ± 5.9%; n = 11; p < 0.01) (Fig. 3Eb), because TPP⁺ did not suppress the posttetanic increase in RRP size (16.1 ± 2.1 vs 16.0 ± 1.6%; n = 11; p = 0.96) (Fig. 3Ec).

Effects of Ru360 on PTP and residual Ca²⁺

Exposed to high [Ca²⁺]_i, mitochondria take up Ca²⁺ via the mitochondrial Ca²⁺ uniporter at the calyx of Held (Billups and Forsythe, 2002; Kim et al., 2005). To test whether mitochondrial Ca²⁺ uptake during TS is a prerequisite for the generation of Ca_res, we examined posttetanic Ca_res at the calyx of Held preloaded with 200 μm fura-4F plus 20 μm Ru360, a specific blocker of Ca²⁺ uniporter (Kirichok et al., 2004). The calyx terminal preloaded with Ru360 showed significantly higher [Ca²⁺] increments during TS (100 Hz, 4 s; 2.77 ± 0.63 μm vs 21.97 ± 5.21 μm; p < 0.01) (Fig. 4A), whereas posttetanic Ca_res was almost abolished [0.42 ± 0.11 μm · s; n = 6; p < 0.01] [Fig. 4Ba (gray dots), Cb]. These results indicate that mitochondrial Ca²⁺ uptake limits the Ca²⁺ build-up during TS, and is a prerequisite for Ca_res. In the presence of presynaptic Ru360, the same TS (100 Hz, 4 s) induced significantly lower PTP than under control conditions (55.0 ± 9.3%; n = 6; p = 0.04) (Fig. 4Ca), whereas the basal EPSC amplitude was not different from control values [0.52 ± 0.15 nA (n = 5) vs 0.43 ± 0.10 nA (n = 6); p = 0.62]. Lower PTP in the presence of Ru360 can be explained primarily by the lower increase in P_r (32.7 ± 7.3%; n = 6; p = 0.03) (Fig. 4Cc), because the increase in RRP size was not different from control values (16.7 ± 1.4%; n = 6; p = 0.48) (Fig. 4Cd). These results further confirm that posttetanic increase in P_r, but not in RRP size, depends on Ca_res at the calyx of Held synapse.

Figure 4.

Effects of Ru360 on presynaptic Ca²⁺ dynamics and PTP. The calyx of Held was preloaded with Ru360 (20 μm) together with fura-4F (200 μm). Aa, Ab, Representative Ca²⁺ transients (Aa) induced by the same TS (100 Hz, 4 s) and mean values for peak [Ca²⁺] of Ca²⁺ transients (Ab) under control conditions (black trace) and in the presence of Ru360 (gray trace). Ba, Time course of EPSC amplitude (open circles, left ordinate) and Ca_res (gray dots, right calibration bar) at the calyx of Held preloaded with Ru360 (aCSF + γ-DGG). Triangles indicate time points when TS (100 Hz, 4 s; black) or a short AP train (100 Hz, 20 pulses; gray) was applied. Bb, First 10 EPSCs during (black) and 20 s after TS (gray). Bc, Plots of cumulative (cum.) EPSC amplitudes during (black) and 20 s after (gray) TS. RRP and P_r were increased by 11.1 and 39.0% of control values, respectively. Ca–Cd, Mean values for PTP (Ca), amount of presynaptic Ca_res (Cb), and increases in the P_r (Cc) and in the RRP size (Cd) at control synapses (ctrl; aCSF + γ-DGG) and at those preloaded with 20 μm Ru360. Data are mean ± SEM; *p < 0.05; **p < 0.01; unpaired t test.

MLCK inhibitors abolish specifically the posttetanic increase in RRP size

MLCK has been suggested to mediate the mobilization of synaptic vesicles (SVs) in the reserve pool at the NMJ (Verstreken et al., 2005), but its role at hippocampal synapse is controversial (Ryan, 1999; Tokuoka and Goda, 2006). To test the involvement of MLCK in posttetanic change of the RRP size, we studied the effects of ML-7, a specific MLCK inhibitor, on PTP and TS-induced increase in the RRP size. ML-7 (20 μm) suppressed PTP from 131.0 ± 21.1 to 75.4 ± 10.2% [n = 10; p < 0.01, paired t test; aCSF + Kyn (n = 5) or γ-DGG (n = 5)] (Fig. 5Aa,Ba). ML-7 abolished the posttetanic increase in the RRP size (19.8 ± 2.0 vs 1.1 ± 2.0%; p < 0.01) (Fig. 5Bb), but not the increase in P_r (95.5 ± 18.7 vs 78.9 ± 8.9%; p = 0.22) (Fig. 5Bc). ML-7 increased the basal EPSC amplitude by 28.5 ± 12.1 and 42.3 ± 15.6% in the presence of Kyn or γ-DGG, respectively, probably because of recruitment of the fast-releasing SV pool (Srinivasan et al., 2008). This can result in occlusion of posttetanic increase in the RRP size. To further test whether MLCK is involved in the posttetanic increase of the RRP size, we investigated the effects of inhibitors for different steps of the signal transduction pathway involving MLCK: calmodulin inhibitor, MLCK inhibitory peptide, and myosin ATPase inhibitor.

Figure 5.

Inhibitors of MLCK or myosin ATPase abolished posttetanic increase in the RRP size with no significant effect on the increase in P_r. Aa, Time course of PTP before and 20 min after bath application of 20 μm ML-7. Triangles indicate time points when TS (100 Hz for 4 s; black) or short AP trains (20 pulses at 100 Hz; gray) was applied. Ab, Ad, First 10 EPSCs during (black) and 20 s after TS (gray) under control conditions (Ab, aCSF + Kyn) and in the presence of ML-7 (Ad). Ac, Ae, Plots of cumulative (cum.) EPSC amplitudes during (black) and 20 s after (gray) TS are superimposed under control conditions (Ac) and after application of ML-7 (Ae). ML-7 suppressed the increase in RRP size from 21.3 to 1.1%, but the increase in P_r was slightly enhanced (from 50.0 to 73.9%). Ba–Cc, Mean values for PTP (Ba, Ca) and increases in the RRP size (Bb, Cb) and P_r (Bc, Cc) under control conditions (ctrl) and 20 min after the bath application of 20 μm ML-7 (Ba–Bc) or 20 μm calmidazolium (CMZ; Ca–Cc). Ba–Bc, aCSF + Kyn (n = 5) or γ-DGG (n = 5); Ca–Cc, aCSF + Kyn (n = 5). Da–Dc, Mean values for PTP (Da) and increases in the RRP size (Db) and in P_r (Dc) at the synapse preloaded with vehicle only (1:1000 DMSO; DMSO, aCSF + γ-DGG; n = 5) or with 100 μm blebbistatin (Bleb; n = 6). Data are mean ± SEM; **p < 0.01.

Activation of MLCK requires binding of Ca²⁺/calmodulin (CaM) to its regulatory segment (Kamm and Stull, 2001). Bath application of calmidazolium (20 μm), a calmodulin antagonist, reduced PTP from 86.2 ± 17.7 to 59.9 ± 12.7% (n = 5; p = 0.02, paired t test; aCSF + Kyn) (Fig. 5Ca), whereas the basal EPSC amplitude was not significantly affected (1.58 ± 0.30 nA vs 1.78 ± 0.37 nA; p = 0.1). Similar to ML-7, calmidazolium suppressed posttetanic increase in the RRP size (16.5 ± 2.2 vs 5.7 ± 2.5%; p = 0.02) (Fig. 5Cb), with no effect on the posttetanic P_r increase (60.0 ± 15.1 vs 52.2 ± 11.0%; p = 0.4) (Fig. 5Cc).

MLCK phosphorylates the myosin regulatory light chain, which in turn enhances the myosin ATPase activity. Myosin II has been suggested to modulate neurotransmission at sympathetic neuron (Takagishi et al., 2005). To test whether myosin II is involved in the posttetanic increase in RRP size, 100 μm blebbistatin, a specific myosin II ATPase inhibitor, was introduced into the presynaptic terminal using the preloading technique. The basal EPSC amplitude was significantly lower at the synapse preloaded with blebbistatin than that at synapse preloaded with vehicle only [1:1000 DMSO; 0.85 ± 0.16 nA (n = 5) vs 0.35 ± 0.09 nA (n = 6); p = 0.02; aCSF + γ-DGG]. Blebbistatin abolished specifically the posttetanic increase in the RRP size (24.0 ± 2.8 vs 0.35 ± 1.8%; p < 0.01) with no significant effect on the increase in P_r (70.5 ± 10.0 vs 76.0 ± 12.4%; p = 0.75) (Fig. 5D).

Next, we tested the effects of MLCKip. The inhibitory mechanism of MLCKip is based on the autoinhibitory sequence of MLCK, independent of CaM, and it shows very high specificity for MLCK with no inhibition of Ca²⁺/CaM kinase II or protein kinase A (IC₅₀ = 50 nm) (Lukas et al., 1999). PTP was assessed at the calyx of Held synapse preloaded with 5 μm MLCKip. Similar to the effects of pseudosubstrate peptide of MLCK on EPSPs in sympathetic neurons (Mochida et al., 1994), MLCKip significantly reduced the basal EPSC amplitude [1.55 ± 0.15 nA (n = 19) vs 0.57 ± 0.09 nA (n = 10); p < 0.01; aCSF + Kyn]. At synapses preloaded with MLCKip, the mean value for PTP was significantly lower than that at control synapses (100.3 ± 10.6 vs 64.9 ± 10.5%; p = 0.04) (Fig. 6Ba). The lower PTP can be ascribed to the negligible posttetanic increase in the RRP size (16.5 ± 1.2 vs 1.2 ± 1.7%; p < 0.01) (Fig. 6Bb), because MLCKip had less effect on the increase in P_r (70.6 ± 9.0 vs 62.2 ± 9.4%; p = 0.56) (Fig. 6Bc). These results strongly suggest that activation of MLCK is responsible for the posttetanic increase in RRP size at the calyx of Held synapse. Finally, we tested whether the effects of TPP⁺ and MLCKip on PTP are additive. When 2 μm TPP⁺ was bath applied in the presence of MLCKip, PTP was further suppressed from 72.2 ± 13.8 to 20.5 ± 7.4% (n = 7; p < 0.01, paired t test; aCSF + Kyn) (Fig. 6A,Ba). In the presence of MLCKip, TPP⁺ reduced only the posttetanic increase in P_r from 70.2 ± 12.0 to 21.4 ± 8.0% (n = 7; p = 0.01) (Fig. 6Bc) with no effect on the RRP size (Fig. 6Bb).

Figure 6.

Additive inhibition of PTP by TPP⁺ and MLCKip. Aa, Time course of PTP at the calyceal synapse preloaded with 5 μm MLCKip (for 90 s) before and after bath application of TPP⁺. Triangles indicate time points when TS (100 Hz for 4 s, black) or short AP trains (20 pulses at 100 Hz; gray) were applied. Ab–Ae, First 10 EPSCs during (black) TS and 20 s after (gray) TS at the synapse preloaded with MLCKip before (Ab) and after (Ad) superfusion of 2 μm TPP⁺. Plots of cumulative (cum.) EPSC amplitudes during (black) and 20 s after (gray) TS at the same synapse are shown in Ac and Ae (black dots, during TS; gray dots, 20 s after TS). Ba–Bc, Mean values for PTP (Ba) and increases in the RRP size (Bb) and P_r (Bc) at the synapse preloaded with 5 μm MLCKip before and after bath application of TPP⁺. For comparison, statistical data under control conditions (ctrl, aCSF + Kyn; n = 19) were reproduced from Figure 1E. Statistical difference between ctrl and MLCKip: *p < 0.05, **p < 0.01, unpaired t test.

Summary for the effects of Ca_res and MLCK inhibitors

In Figure 7A, we summarize the effects of various inhibitors on PTP and increases in P_r and the RRP size. For comparison, we pooled data obtained with drugs targeting Ca_res into one group and those obtained with drugs targeting MLCK into another group, because the observed effects were statistically not different. In addition we summarize the results obtained for combined application of MLCKip plus TPP⁺. The bar graphs in Figure 7A demonstrate that P_r and the RRP size are independently regulated during PTP by Ca_res and MLCK, respectively. Finally, we examined the decay time constants of PTP under control conditions and PTP components sensitive to TPP⁺ or to ML-7 (Fig. 7B). We excluded from analysis synapses at which we lost the patch before PTP was fully decayed. At 11 of 21 synapses, the time course of control PTP was best fitted with biexponential function with fast and slow time constants of 24.2 ± 2.5 s and 50.0 ± 4.0 s, respectively (n = 11). At other synapses, PTP decayed monoexponentially with a time constant of 49.6 ± 2.5 s (n = 10). Regardless of whether the decay of PTP follows monoexponential or biexponential time course under control conditions, the PTP component sensitive to TPP⁺ (denoted by “TPP diff” in Fig. 7B) decayed with a time constant (25.7 ± 2.3 s; n = 13) similar to the fast component of biexponential decay of control PTP (denoted by “ctrl fast” in Fig. 7B). In contrast, the decay time constant of the PTP component sensitive to ML-7 (denoted by “ML-7 diff”; 47.1 ± 2.7 s; n = 8) was similar to the slow time constant of control PTP (p = 0.59), which is distinct from that of TPP⁺-sensitive PTP. These results suggest that PTP at the calyx of Held synapse is composed of fast and slow decaying components that are regulated by Ca_res and MLCK, respectively.

Figure 7.

Aa–Ac, Summary for the effects of mitochondrial Ca_res inhibitors and MLCK inhibitors on PTP (Aa) and increases in the RRP size (Ab) and in P_r (Ac). There was no statistical difference between the effect of TPP⁺ and that of Ru360, and the data were pooled. The data in the presence of ML-7, calmidazolium (CMZ), blebbistatin (Bleb), or MLCKip were pooled for the same reason. The statistical difference from a control value, **p < 0.01 (unpaired t test). ns, Not significant. B, Mean values for decay time constants of PTP under control conditions and the PTP components sensitive to TPP⁺ (denoted by “TPP⁺ diff”) or ML7 (denoted by “ML-7 diff”). The decay time constant of TPP⁺-sensitive PTP (reproduced from Fig. 2C) was significantly faster than that of ML7-sensitive PTP (**p < 0.01). The decay time constant for PTP remaining after the combined treatment with TPP⁺ and MLCKip was also similar to the slow time constant (49.7 ± 1.6 s; n = 6; p = 0.96). Numbers in parentheses are the number of synapses investigated.

Role of protein kinase C in PTP

The above results indicate that Ca_res and activation of MLCK can explain ∼80% of PTP. Recently, however, Korogod et al. (2007) have reported that PTP was abolished by protein kinase C (PKC) inhibitors, bisindolylmaleimide (Bis) drugs: Bis-I or Ro31-8220. To resolve this discrepancy, we compared PTP at the synapse preloaded with Bis drugs or an inactive analog, Bis-V. The results are summarized in Figure 8, and EPSC traces are presented in supplemental Fig. S1 (available at www.jneurosci.org as supplemental material). Not only Bis-1 (2 μm) and Ro31-8220 (2 μm) but also Bis-V (2 μm) significantly suppressed PTP (Fig. 8), indicating that PKC-independent mechanisms may be involved in the inhibitory effects of Bis drugs. Three drugs abolished the posttetanic increase in RRP size. Because posttetanic increase in P_r was significantly suppressed by Bis-I and Bis-V, but not by Ro31-8220, we tested the effect of these drugs on Ca_res. Consistently, Ca_res was abolished by Bis-I and Bis-V, but not by Ro31-8220 (supplemental Fig. S1A–C, available at www.jneurosci.org as supplemental material), suggesting that Bis-1 affects mitochondrial Ca²⁺ handling regardless of PKC. The Bis drugs inhibit PKC by competing with ATP (Toullec et al., 1991). We tested the effects of chelerythrine, which inhibits PKC by covalently reacting with catalytic domain (Gopalakrishna et al., 1995). PTP persisted at the synapse preloaded with 5 μm chelerythrine (Fig. 8; supplemental Fig. S1D, available at www.jneurosci.org as supplemental material). PTP and the increase in P_r were slightly reduced in the presence of chelerythrine with no statistical significance (p = 0.2 and p = 0.15, respectively). These results indicate that at least a part of inhibitory effects of Bis drugs may not be related to the inhibition of PKC.

Figure 8.

The effects of PKC inhibitors on PTP at the calyx of Held synapse. Mean values ± SEM for PTP and relative changes in the RRP size and P_r estimated at the synapse preloaded with DMSO (1:1000, v/v; n = 5), 2 μm Bis-I (n = 5), 2 μm Ro31-8220 (n = 4), 2 μm Bis-V (n = 4), or 5 μm chelerythrine (Che; n = 6) are shown. Note that the posttetanic increase in RRP size was abolished not only by Bis-I or Ro31-8220 but also by an inactive analog, Bis-V (−4.16 ± 2.77%; n = 4; p < 0.01). PTP at the synapse preloaded with 5 μm chelerythrine (84.0 ± 15.1%; n = 6; p = 0.23) was not different from that at the synapse with DMSO (1/1000, 112.0 ± 15.9%; n = 5). Statistical data in each condition were compared with data obtained from the synapse preloaded with DMSO unless specified otherwise. *p < 0.05; **p < 0.01.

Discussion

The principal findings of the present study are that (1) Ca_res is generated by posttetanic Ca²⁺ release via NCE subsequent to mitochondrial Ca²⁺ uptake during TS; (2) Ca_res and activation of MLCK independently mediate the posttetanic increases in P_r and the RRP size, respectively; and (3) the increase in RRP size lasted longer than Ca_res, and constitutes a part of the slow decay phase of PTP at the calyx of Held synapse. These findings may provide an explanation for the long-standing question of why presynaptic Ca_res decays faster than PTP at mammalian central synapses. The involvement of MLCK in PTP is of general importance for short-term synaptic plasticity.

Our results indicate that the increase in P_r contributes more to PTP than that in the RRP size. It should be noted, however, that the present study was done at immature synapses and at room temperature. Given that mitochondria in the calyx of Held compete for cytosolic Ca²⁺ with Na⁺/Ca²⁺ exchanger (Kim et al., 2005), developmental increase in Na⁺/Ca²⁺ exchanger activity (our unpublished observations) might limit mitochondrial uptake of Ca²⁺ during TS, a prerequisite for posttetanic Ca_res. Moreover, the decay of the posttetanic P_r increase is preferentially accelerated at physiological temperature with little effect on the change in RRP size (Habets and Borst, 2007). Therefore, residual calcium, which subserves the increase in P_r, might contribute less to PTP at mature synapses under physiological conditions.

Mechanisms underlying the posttetanic increase in P_r

Downstream mechanisms of Ca_res responsible for the increase in P_r are not clear. Saturation of endogenous Ca²⁺ buffers or increase in Ca²⁺ influx may underlie the increase in P_r (Felmy et al., 2003; Habets and Borst, 2006). Recently, Korogod et al. (2007) suggested that the latter is not sufficient to explain the P_r increase, and that activation of PKC contributes to the posttetanic P_r increase. The results in Figure 8, however, indicate that further experiments are necessary to clarify the role of PKC for PTP at the calyx of Held synapse.

Recently, we have reported that Ca_res derived from mitochondria and activation of PKC differentially contribute to PTP at large and small mossy fiber (MF) boutons of dentate granule cells (Lee et al., 2007). Ca_res-dependent PTP has been observed only at exceptionally large presynaptic terminals such as NMJs, large mossy fiber boutons, and calyces of Held, suggesting a higher contribution of mitochondrial Ca²⁺ uptake during TS to the Ca²⁺ clearance at large presynaptic terminals than at small boutons. In the same paper, we also reported that PTP at large MF synapses was not inhibited by PKC inhibitors (chelerythrine and Ro31-8220). Interestingly, we observed that both Bis-I and Bis-V suppressed PTP at large MF synapses to a similar extent (our unpublished observations). This result is consistent with our observation that Bis-I and Bis-V suppress Ca_res at the calyx of Held (supplemental Fig. S1, available at www.jneurosci.org as supplemental material), and thus one should be cautious in interpreting the effects of Bis-I on PTP.

Posttetanic increase in the RRP size

It has been firmly established that desensitization and saturation of AMPA receptors contribute to the synaptic depression during strong synaptic activity (Trussell et al., 1993; Otis et al., 1996; Wong et al., 2003). The absolute value for the RRP size should be underestimated in the presence of 1 mm Kyn without cyclothiazide (CTZ). Consistent with Korogod et al. (2005), we found that in the presence of 1 mm Kyn plus 100 μm CTZ, the relative increase in estimates for the RRP size was negligible. However, when Kyn was elevated from 1 to 2 mm, RRP size estimates were significantly higher after induction of PTP (supplemental Fig. S2B, available at www.jneurosci.org as supplemental material). This may indicate that the addition of 1 mm Kyn is insufficient to fully relieve AMPA receptor saturation during 100 Hz trains when CTZ is present in the bath. The decay time course of PTP was significantly slower in the presence of CTZ than that under control conditions, and the enhancement of EPSC did not subside at least until 5 min after TS; probably long-term potentiation of EPSC was induced (supplemental Fig. S2C, available at www.jneurosci.org as supplemental material). This result might be caused by nonspecific presynaptic effects of CTZ (Ishikawa and Takahashi, 2001). Therefore, we consider the use of CTZ problematic when studying the mechanisms of PTP at the calyx of Held synapse.

Because estimates of the RRP size from plots of cumulative EPSC amplitudes largely represent the fast-releasing pool (Schneggenburger et al., 2002; Sakaba, 2006), the TS-induced increase in the RRP size observed in the present study may represent an increase in the size of a fast-releasing pool rather than in a slow-releasing pool. In contrast, when the size of RRP was estimated by hypertonic sucrose, an increase in P_r, but not in the RRP size, was induced during augmentation in cultured hippocampal neurons (Stevens and Wesseling, 1999). The discrepancy may be ascribed to the fact that RRP estimates from cumulative EPSCs plots represent only a part of the RRP size reported by hypertonic sucrose (Moulder and Mennerick, 2005).

Role of MLCK in the posttetanic increase in the RRP size

It is not clear whether the mechanism responsible for replenishment of SV pools after depletion is similar to that for the posttetanic increase in RRP size. We showed that the posttetanic increase in RRP size was not affected by inhibition of Ca_res (Figs. 3B, 4C). Moreover, no linear correlation was found between the amount of posttetanic Ca_res and the change in RRP size (Fig. 3D), indicating that Ca_res plays little role in posttetanic regulation of the RRP size. Therefore, activation of MLCK induced by the Ca²⁺ build-up during TS seems to be primarily responsible for the posttetanic increase in RRP size. This result is contrasted with the recent report that replenishing rate of fast-releasing pool is linearly dependent on [Ca²⁺]_i (Hosoi et al., 2007). Activation of myosin II by MLCK is thought to be a mediator for mobilization and recruitment of SVs at NMJs, where inhibition of MLCK caused intermittent failure of neurotransmission during TS (Polo-Prada et al., 2005). However, we did not observed any difference in late steady-state neurotransmission during 100 Hz stimulation in the presence of MLCK inhibitor or blebbistatin, suggesting that inhibition of MLCK has little effect on the vesicle recruitment during TS. Consistent with this view, Srinivasan et al. (2008) have recently reported that ML-7 affects neither the steady-state EPSCs nor the rate of recovery from synaptic depression. Therefore, we imagine that the mechanisms underlying posttetanic increase in the RRP size might be different from that underlying replenishment of fast SV pool after depletion.

Possible physiological role of PTP in development of calyx of Held synapses

The calyx of Held synapse is a sign-inverting synapse, for which synaptic transmission with high fidelity of spike timing is essential. The task of the calyx of Held synapse might be interfered with by activity-dependent plastic changes of the synaptic connectivity, including PTP. Consistently, it has been reported that PTP is reduced as the synapse matures (Korogod et al., 2005). We consider a possibility that PTP contributes to the maturation of calyx of Held synapse. Recently, Joshi et al. (2007) reported that extrasynaptic NMDA receptors (NMDARs) are downregulated by pairing a depolarization of the postsynaptic neuron and burst stimulation of afferent axon fibers, probably by the mechanism of coincident activation of postsynaptic metabotropic glutamate receptors and NMDARs. It remains to be addressed whether such strong stimuli are possible under physiological conditions during a developmental period of hearing onset. We imagine that PTP at this synapse could provide an answer to this question, because an appreciable amount of PTP still remains at the age of hearing onset (Korogod et al., 2005). Sound stimulus elicits presynaptic bursts of high frequencies up to 600 Hz at the MNTB units in vivo (Kopp-Scheinpflug et al., 2003), suggesting that hearing onset will be accompanied with increased episodes of tetanic stimulation of MNTB neurons. It is conceivable that a prolonged tone burst might induce PTP of subsequent spontaneous synaptic activities. Enhanced glutamate release into synaptic cleft during PTP may facilitate the coincident activation and contribute to downregulation of extrasynaptic NMDARs.