The extent to which Ca2+-induced Ca2+ release (CICR) affects transmitter release is unknown. Continuous nerve stimulation (20–50 Hz) caused slow transient increases in miniature end-plate potential (MEPP) frequency (MEPP-hump) and intracellular free Ca2+ ([Ca2+]i) in presynaptic terminals (Ca2+-hump) in frog skeletal muscles over a period of minutes in a low Ca2+, high Mg2+ solution. Mn2+ quenched Indo-1 and Fura-2 fluorescence, thus indicating that stimulation was accompanied by opening of voltage-dependent Ca2+ channels. MEPP-hump depended on extracellular Ca2+ (0.05–0.2 mM) and stimulation frequency. Both the Ca2+- and MEPP-humps were blocked by 8-(N,N-diethylamino)octyl3,4,5-trimethoxybenzoate hydrochloride (TMB-8), ryanodine, and thapsigargin, but enhanced by CN. Thus, Ca2+-hump is generated by the activation of CICR via ryanodine receptors by Ca2+ entry, producing MEPP-hump. A short interruption of tetanus (<1 min) during MEPP-hump quickly reduced MEPP frequency to a level attained under the effect of TMB-8 or thapsigargin, while resuming tetanus swiftly raised MEPP frequency to the previous or higher level. Thus, the steady/equilibrium condition balancing CICR and Ca2+ clearance occurs in nerve terminals with slow changes toward a greater activation of CICR (priming) during the rising phase of MEPP-hump and toward a smaller activation during the decay phase. A short pause applied after the end of MEPP- or Ca2+-hump affected little MEPP frequency or [Ca2+]i, but caused a quick increase (faster than MEPP- or Ca2+-hump) after the pause, whose magnitude increased with an increase in pause duration (<1 min), suggesting that Ca2+ entry-dependent inactivation, but not depriming process, explains the decay of the humps. The depriming process was seen by giving a much longer pause (>1 min). Thus, ryanodine receptors in frog motor nerve terminals are endowed with Ca2+ entry-dependent slow priming and fast inactivation mechanisms, as well as Ca2+ entry-dependent activation, and involved in asynchronous exocytosis. Physiological significance of CICR in presynaptic terminals was discussed.

Introduction

Ca2+ in presynaptic terminals plays key roles not only in the impulse-induced exocytosis of neurotransmitters, but also in other slow regulatory actions on transmitter release, such as the turnover of synaptic vesicles and short- and long-term plasticities (see Katz, 1969; reviewed by Schweizer et al., 1995; Südhof, 1995; Zucker, 1996). While impulse-induced release of transmitter appears to be caused by a high [Ca2+]i existing in a short period at the orifice of voltage-activated Ca2+ channels (Llinás et al., 1992; Heidelberger et al., 1994; Landò and Zucker, 1994), slow regulatory actions of Ca2+ could result from a longer, more global, but moderate rise in cytosolic Ca2+ concentration ([Ca2+]i), as has been well known for the action of residual [Ca2+]i remaining after preceding impulses on short-term plasticities of transmitter release (Katz and Miledi, 1968; Tanabe and Kijima, 1992; see Zucker, 1996).

Such a moderate, prolonged increase in [Ca2+]i, however, could more effectively be achieved by the release of Ca2+ from intracellular Ca2+ stores in response to impulses (see review by Kuba, 1994). Erulkar and Rahamimoff (1978) suggested that tetanic and post-tetanic increases in miniature end-plate potential (MEPP)1 frequency in a Ca2+-free (EGTA) solution are the result of intracellular Ca2+ release activated by Na+ entering during nerve impulses. On the other hand, much evidence has been accumulated that the activation of Ca2+-induced Ca2+ release (CICR) via ryanodine receptors occurs in response to Ca2+ entry in the cell soma and dendrites of neurones (Kuba and Nishi, 1976; Kuba, 1980; Thayer et al., 1988; Lipscombe et al., 1988; Friel and Tsien, 1992; Llano et al., 1994; Garaschuk et al., 1997; see review by Kuba, 1994). The evidence, however, for the involvement of CICR in exocytosis at presynaptic terminals is rather indirect. Caffeine and theophylline, activators of ryanodine receptor, have been known to enhance both the evoked and spontaneous release of transmitters in several types of terminals, including frog motor nerve terminals (Elmqvist and Feldman, 1965; Onodera, 1973; Ohta and Kuba, 1980; Lockerbie and Gordon-Weeks, 1986; Tóth et al., 1990). Recent studies suggested ryanodine-sensitive components of tetanus-induced rise in [Ca2+]i and transmitter release in autonomic nerve terminals (Peng, 1996; Smith and Cunnane, 1996).

In this study, we have tested the hypothesis that Ca2+ entry produced by impulses activates CICR in presynaptic terminals, amplifies the rise in [Ca2+]i, and enhances transmitter release and other Ca2+-dependent functions. We have measured changes in the frequency of MEPPs and [Ca2+]i in motor nerve terminals in response to tetanic stimulation in combination with pauses in tetanus of various duration and timing at frog neuromuscular junctions and observed effects of blockers of Ca2+ release and uptake. The results suggest that CICR occurs via the activation of ryanodine receptors in response to Ca2+ entries accompanying impulses and enhances asynchronous release of transmitter. They further revealed unique Ca2+ entry-dependent priming, depriming, inactivation, and deinactivation processes of ryanodine receptors as well as Ca2+ entry-dependent activation process. The CICR mechanism, which is normally deprimed under the resting condition, is slowly primed, subsequently activated, and then inactivated by Ca2+ entry accompanying nerve activity, and eventually deprimed slowly in the absence of Ca2+ entry.

Materials And Methods

Electrophysiology

Frogs (Rana nigromaculata) were decapitated and sartorius muscles with the innervating nerve were isolated. Most experiments were carried out after the incubation of preparations at 4°C for 3–12 h to place them in a condition similar to that for measurement of [Ca2+]i. Such a treatment did not essentially affect the amplitude and time course of changes in MEPP frequency in response to continuous tetanus. The composition of normal Ringer's solution (mM) was 100 NaCl, 2 KCl, 2.5 CaCl2, 3.0 MgCl2, 8.0 Tris, pH 7.4. Low Ca2+, high Mg2+ solutions were made by lowering CaCl2 (0.02–0.2 mM) and adding MgCl2 (10 mM) with the adjustment of osmolarity by changing NaCl concentration. A Ca2+-free (EGTA), Mg2+ solution was made by removing CaCl2 and adding 1 mM MgCl2 and 1 mM EGTA. Neostigmine bromide (10−6 g/ml) was added to all solutions to enhance the amplitude of MEPP for the accurate measurement of its frequency. MEPPs were recorded by a conventional intracellular recording technique using electrodes filled with 3 M KCl at 20–24°C. They were stored on a PCM recorder (40 kHz, RD101T or RD125T; TEAC, Tokyo, Japan) and fed into a computer (PC-9801-DA, NEC; Tokyo, Japan) through a GPIB interface (GP-300 or GP-350; TEAC) that also controlled the transfer of the data into the computer memory. All the digitized MEPPs were low-pass filtered at 5 kHz (six poles, Butterworth; installed in RD101T and RD125T). In most analyses, MEPPs displayed on the screen were identified by eye and their number was counted. When the summation of MEPPs was so large that the number of MEPPs buried in the record could not be ignored (usually >200/s), MEPP frequency was estimated from the integration of voltage records (Mulkey and Zucker, 1993). The voltage record was divided into 1-s segments. The areas enclosed by the voltage trace in each segment and the base line measured before the beginning of tetanus were integrated. To eliminate the contribution of stimulation artifacts, the record of one segment that contained artifacts but not MEPP was integrated and subtracted from the integration of the record that contained MEPPs. The result was divided by the area of the averaged single MEPP (or the product of the mean amplitude and the sum of the decay time constant and the rise time) recorded discretely at low frequency, yielding MEPP frequency. A blocker of ryanodine receptor, 8-(N,N-diethylamino)octyl3,4,5-trimethoxybenzoate hydrochloride (TMB-8), reduced the amplitude of MEPP to 81% of the control (see also Hunt et al., 1990). This action, however, did not affect the estimation of MEPP frequency for the following reasons. First, the extent of the decrease in MEPP amplitude was not large enough to affect identification by eye. Second, both the voltage trajectory of summated MEPPs and a single MEPP should be similarly affected by TMB-8 so that their division could be unchanged.

[Ca2+]i Measurements

Changes in [Ca2+]i in frog motor nerve terminals were measured from those of cutaneous pectoris muscles. Frogs were decapitated and cutaneous pectoris muscles with the innervating nerve were isolated. The composition of normal Ringer's solution (mM) was 112 NaCl, 2 KCl, 1.8 CaCl2, 2.4 NaHCO3, pH 7.4 (when equilibrated with air), with or without 5.0 glucose or 111 NaCl, 2 KCl, 1.8 CaCl2, 5.0 Hepes-Na, pH 7.4, 5.0 glucose. Low Ca2+, high Mg2+ and Ca2+-free (EGTA), Mg2+ solutions were prepared similarly as in the experiments for recording MEPPs. K-salt of Indo-1, Oregon Green BAPTA-1, or Fura-2 was loaded into the terminals from the cut end of the nerve bundle (Peng and Zucker, 1993). A drop (0.2–0.5 μl) of a Ca2+-free solution (buffered with 10 mM HEPES-Na, pH 7.4) containing K-salt of Indo-1, Oregon Green BAPTA-1, or Fura-2 (30 mM) was placed at the cut end of the nerve in a moist chamber and incubated for 8–12 h at 20–22°C (Osanai et al., 1996). During this period, Ca2+ probes were transported or diffused through the axon cytoplasm down to the terminals. In most experiments, cutaneous pectoris muscles, whose nerve terminals were loaded with Ca2+ probes, were set on a plastic plate, which was held upside down in a recording chamber placed on the stage of an inverted microscope (TMD or TMD-300; objective, 40× water/N.A. 1.15; Nikon, Tokyo, Japan). The innervating nerves were stimulated with a pair of platinum wires set in a plastic tube attached to the plate, on which the muscles were fixed.

There are two reasons for the use of different preparations for recording MEPPs and [Ca2+]i in nerve terminals. First, we have initially started the experiments by recording MEPPs from frog sartorius muscles and most data that suggest the existence of CICR in nerve terminals were obtained from measurement of MEPPs. Second, small, thin muscle preparations, such as cutaneous pectoris muscles, were relevant for recording fluorescence from the motor nerve terminals of neuromuscular preparations with an inverted microscope. Limitations for the use of different preparations and the rationale that the data obtained from independent experiments on different preparations could lead to the conclusion of the present study are described in the discussion.

In most experiments, confocal laser scanning microscopes (MRC-500 or -600; Nippon Biorad, Tokyo, Japan) equipped with an ultraviolet Argon laser (351 nm; Kuba et al., 1994) or Krypton-argon laser (488 nm) were used to measure fluorescence changes of Indo-1 or Oregon Green BAPTA-1. Indo-1 fluorescence intensity was measured at two wavelength ranges separated at 445 nm and the ratio of fluorescence (F412/F475) was converted to a [Ca2+]i value using Eq. 5, described by Grynkiewicz et al. (1985). The apparent dissociation constant, the maximum and minimum ratios for conversion were obtained from the relationship in vitro between the ratio of Indo-1 fluorescence and the known concentration of Ca2+ in a pseudo-intracellular solution (140 mM KCl, 10 mM EGTA, 10–25 μM Indo-1, 10 mM HEPES-K, pH.7.2, and various concentrations of CaCl2) containing aldolase (2–4 mg/ml, Type IV; Sigma Chemical Co., St. Louis, MO), which was assumed to mimic endogenous cytosolic proteins affecting a Ca2+ dependence of Indo-1 fluorescence (Konishi et al., 1988). The dissociation constant of Indo-1 was found to be ∼450 nM in the presence of 2–4 mg/ml aldolase, about twice that in a protein-free solution. Since both maximum and minimum ratios were reported to depend on the concentration and type of proteins (Baker et al., 1994), absolute values of [Ca2+]i estimated would be subject to ambiguity by a factor of 2. Nevertheless, this possible error would have little effect on the interpretation of the results throughout this paper. In the analyses of fluorescence images of Oregon Green BAPTA-1, the ratios of the images during and after nerve stimulation to that before stimulation (F/FC) were taken. This corrected differences in path length and dye concentration variation (Neher and Augustine, 1992). Ratios were converted to [Ca2+]i values by the equation expressing the relationship between Ca2+ concentration and F/FC.

 
\begin{equation*}[Ca^{2+}]_{i}=K_{d}(AF/F_{C}-1)/(F_{max}/F_{min}-AF/F_{C}),\hspace{.167em}\hspace{.5em}\end{equation*}

where

\begin{equation*}A=(F_{max}/F_{min}+K_{d}/[Ca^{2+}]_{C})/(1+K_{d}/[Ca^{2+}]_{C}),\end{equation*}
Fmax, and Fmin are the maximum and minimum fluorescence intensities, respectively, of Oregon Green BAPTA-1 at a high (enough to saturate the probe) and zero concentration of Ca2+, respectively, and [Ca2+]C was the [Ca2+]i before stimulation. The values for [Ca2+]C and the Kd of Oregon Green BAPTA-1 used in calculation were 72 ± 9 nM (n = 18; previously obtained from Indo-1 fluorescence measurement) and 240 nM (obtained similarly as in the case of Indo-1; our unpublished observations), respectively. The values for Fmax and Fmin were obtained in vitro by measuring the fluorescence intensities of pseudo-intracellular solutions containing 10 μM Ca2+ and no Ca2+, respectively, and their ratio was found to be 15.0.

Scanning of the nerve terminals was made by an x–y scan mode with or without the summation of two to four images. The averages of the ratios of Indo-1 fluorescence or the intensities of Oregon Green fluorescence were measured from small areas (2.25–4 μm2) of the images of individual terminals.

In some experiments, examining quenching effects of Mn2+ on fluorescence of Fura-2 loaded in nerve terminals, the fluorescence was measured with a cooled CCD-camera (Argus/HiSca, C6790-81; Hamamatsu Photonics, Hamamatsu, Japan) through an image intensifier attached to an upright microscope (Zeiss Axioscope, objective 100× water/N.A. 0.95; Carl Zeiss Japan, Tokyo, Japan) and analyzed by image analysis software (Argus; Hamamatsu Photonics).

Statistical Tests

Student's t tests were made for the data in which they were apparently required.

Drugs

Indo-1/K, Oregon Green BAPTA-1/K, and Fura-2/K were obtained from Molecular Probes, Inc. (Eugene, OR). TMB-8 was from Sigma Chemical Co. (St. Louis, MO) or Tokyo Kasei Kogyo Co. (Tokyo, Japan). Ryanodine, HEPES-Na or -K, aldolase, and neostigmine bromide were from Sigma Chemical Co. Thapsigargin was from Research Biochemicals, Inc. (Natick, MA) or Sigma Chemical Co.

Experimental Parameters Reflecting the Rate of Ca2+ Release

The approximate estimation of changes in the rate of Ca2+ release from changes in [Ca2+]i and MEPP frequency was made based on a simple model for intracellular Ca2+ dynamics (Fig. 1). Ca2+ diffusion can be neglected in the formation of rate equations for Ca2+ dynamics in response to Ca2+ entry and Ca2+ release, since most changes recorded in [Ca2+]i were much slower than the time for spatial equilibrium of Ca2+, and the [Ca2+]i values measured are the average over those within a small region (a few square micrometers) of the terminal. It is further assumed that the rates of Ca2+ binding to, and unbinding from, proteins are so fast that the reactions can be considered to be at equilibrium (Neher and Augustine, 1992). Likewise, the rates of Ca2+ uptake into, and release from, mitochondria are assumed to be faster than those of other Ca2+ stores. This assumption can be supported by similar time courses of Ca2+-humps in the absence and presence of CN (see Fig. 6 C). Then, the rate of changes in [Ca2+]i (dx/dt) is expressed by the sum of the rates of Ca2+ entry during tetanus (JCa*), Ca2+ release from Ca2+ stores (γz), Ca2+ uptake into Ca2+ stores (αx), and Ca2+ extrusion at the cell membrane (px). Thus, an equation similar to that previously proposed for CICR (Kuba and Takeshita, 1981) was derived with the inclusion of Ca2+ buffering by Ca2+ binding proteins and mitochondria (1 + κs):

 
\begin{equation*}(1+{\kappa}_{s})dx/dt=J_{Ca}*+{\gamma}z-({\alpha}+p)x\end{equation*}
1

or

 
\begin{equation*}{\gamma}z=(1+{\kappa}_{s})dx/dt\;+({\alpha}+p)x-J_{Ca}*,\end{equation*}
2

where x and z are [Ca2+]i and the Ca2+ concentration of Ca2+ stores ([Ca2+]s), respectively, γ, p, and α are rate constants, and κs is the sum of the Ca2+ binding capacity of endogenous Ca2+ binding proteins (Neher and Augustine, 1992) and the Ca2+ uptake capacity of mitochondria. In the equation, Ca2+ influx into Ca2+ stores through Ca2+ release channels was implicitly ignored. The rate constants, γ, p, and α, include the total number of Ca2+ release channels (for γ) or Ca2+ pumps (for p and α), open probability of Ca2+ release channels (for γ) or rate constants of the single Ca2+ pump (for p and α), and the volume of the free cytoplasm or Ca2+ stores (for γ and p, respectively). JCa* is a function of the total number, open probability, and permeability of Ca2+ channels, the extracellular Ca2+ concentration, cell membrane potential, and the total volume of cytoplasm. It may safely be assumed that JCa* is constant, since Ca2+-dependent inactivation of voltage-dependent Ca2+ channels appears to be not so large, if it occurs, as suggested from the quick quenching effect of Mn2+ (see Fig. 4). Furthermore, JCa* in low Ca2+ (0.05–0.2 mM), high Mg2+ (10 mM) solutions used in the present study is negligibly small when compared with the net rate of Ca2+ release (γz − αx), seen as a transient rise in [Ca2+]i (see Figs. 3 and 10 A). Then, Eq. 2 implies that the sum of (1 + κs)dx/dt and (α + p)x would reflect the rates of Ca2+ release.

When (α + p)x is much greater than (1 + κs)dx/dt (or dx/dt is negligibly small: steady/equilibrium approximation; Kijima and Kijima, 1982), it becomes,

 
\begin{equation*}{\gamma}z\;\approx ({\alpha}+p)x.\end{equation*}
3

On the other hand, when (α + p)x is smaller than (1 + κs)dx/dt,

 
\begin{equation*}{\gamma}z\;\approx (1+{\kappa}_{s})dx/dx.\end{equation*}
4

When the values of dx/dt are measured at similar values of x, changes in dx/dt would reflect changes in the rate of Ca2+ release (see Eq. 2).

In addition to [Ca2+]i measurement, we have used changes in MEPP frequency as a reflection of changes in [Ca2+]i with the understanding that MEPP frequency depends on the power of [Ca2+]i (Ravin et al., 1997), the efficiency of exocytosis, and the amount of transmitter available for release. The amount of transmitter pool may be decreased by an increase in MEPP frequency during tetanus. Such a reduction, however, would have been ∼300–500 quanta/s, assuming that each MEPP consists of a single quantum. This is roughly equivalent to that caused by the stimulation at 2 Hz in normal Ringer solution, since 200 quanta are usually released by a nerve impulse at frog motor nerve terminals (Gage, 1976). Thus, the reduction of available pools of transmitter may be negligible. On the other hand, the efficiency of exocytosis may change during continuous tetanic stimulation. This indeed took place in the present experiments. Changes in MEPP frequency due to alterations in exocytotic efficiency, however, were relatively small when compared with those produced by CICR (Fig. 2 D). This conventional interpretation of MEPP frequency may therefore be used in the analyses of CICR. In fact, the time course of changes in MEPP frequency and [Ca2+]i in response to tetanus, its interruption, and actions of blockers of Ca2+ release and uptake were similar. However, the extent of changes in MEPP frequency exceeded those for [Ca2+]i due to the power law dependence of MEPP frequency on [Ca2+]i.

Results

All the experiments were done in low Ca2+ (0.05–0.2 mM), high Mg2+ (10 mM) or Ca2+-free (EGTA), and Mg2+ (1 mM) solutions, where no end-plate potentials were evoked by nerve stimulation (Fig. 2 A). Under this condition, the contribution of CICR to impulse- induced rise in [Ca2+]i would be greater, if it occurs. MEPPs were recorded from the end plates of frog sartorius muscles, while changes in [Ca2+]i were measured from the motor nerve terminals of frog cutaneous pectoris muscles.

A Transient Increase in MEPP Frequency during Continuous Tetanic Stimulation

Continuous nerve stimulation at 50 Hz produced a large transient rise in MEPP frequency (MEPP-hump) in frog sartorius muscles. The MEPP-hump grew over a few minutes, reached a peak (288.0 ± 43.6/s, SEM, n = 17 from the resting value of 0.58 ± 0.13/s, n = 17, P < 0.001) and decayed to a level (77.6 ± 8.4/s, n = 17, P < 0.001 when compared with the resting value) higher than the control at a rate slightly faster than that of growth (Fig. 2, A and B). This was followed by a slow progressive rise (122.0 ± 26.5/s, n = 14, at 2 min after the end of a transient rise, P < 0.2 when compared with that immediately after the MEPP-hump; Fig. 2, A and B). Decreasing tetanus frequency to 20 Hz reduced the magnitude of MEPP-hump and delayed its peak (Fig. 2,C). Lowering [Ca2+]o from 0.2 to 0.1 mM reduced MEPP-hump to 76.7 ± 13.2/s (n = 8, P < 0.001), and the later slow rise in MEPP frequency to 66.6 ± 9.5/s (n = 8, P < 0.2; Fig. 2 B). MEPP-hump was almost abolished at 0.05 mM [Ca2+]o. It may be noted that the strong potentiation of exocytotic mechanism may also occur in the late slow rise in MEPP frequency. The mechanism for this slow rise was not studied here, although it is of interest.

A Transient Rise in [Ca2+]i in Motor Nerve Terminals during a Tetanus

Continuous tetanic stimulation (50 Hz) also caused a transient increase in [Ca2+]i in motor nerve terminals (Ca2+-hump) in frog cutaneous pectoris muscles. Changes in [Ca2+]i were measured by recording alterations in the fluorescence ratio of Indo-1 or Oregon Green BAPTA-1 in a region (a few square micrometers) of the images of the terminals. The Ca2+-hump recorded with Indo-1 fluorescence rose in a few minutes to peak (from 69 ± 8 nM to 1.58 ± 0.2 μM, n = 12; P < 0.001) and decayed to a level (320 ± 14 nM; P < 0.001 when compared with the resting value) higher than that before stimulation (Figs. 3 and 4,A). Ratio images of Oregon Green BAPTA-1 fluorescence showed a better signal to noise ratio (Fig. 5). Ca2+-humps measured by Oregon Green BAPTA-1 fluorescence were in general similar in time course and space, but varied to some extent in amplitude (see also Fig. 10 A).

Ca2+-humps were blocked by lowering [Ca2+]o (not shown). The time course of Ca2+-hump resembled that of MEPP-hump, although the magnitude of relative changes to that before stimulation was much greater in the latter (obviously for the power law dependence of MEPP frequency on [Ca2+]i). The results suggest that similar mechanisms underlie in both MEPP- and Ca2+-humps, although the activation of the exocytotic process is also involved in the former.

Ca2+-hump was followed by the late slow rise in [Ca2+]i, which was maintained further (354 ± 17 nM at 2 min; P < 0.001 in respect to the [Ca2+]i before stimulation, P < 0.2 when compared with that immediately after the Ca2+-hump). The magnitudes of the late slow rises in [Ca2+]i were variable among terminals and smaller than those of MEPP frequency when compared in amplitude relative to respective humps.

Reduction of Ca2+ Entry Does Not Explain the Decline of Ca2+- and MEPP-Humps

Ca2+- and MEPP-humps can be accounted for by several possible mechanisms. First, continuous tetanic stimulation at a high frequency (50 Hz) for several minutes in a low Ca2+, high Mg2+ solution may have blocked the conduction of impulses to motor nerve terminals. Second, even if no conduction failure occurred, it might be possible that the increase in [Ca2+]i induced by continuous activation of Ca2+ entry during a tetanus would cause Ca2+-dependent inactivation of voltage-dependent Ca2+ channels (Eckert and Tillotson, 1981; Galli et al., 1994). These mechanisms would have caused the rapid decay of the Ca2+- and MEPP-humps. An experiment to test whether voltage-dependent Ca2+ channels are opened by each impulse during tetanus is to apply Mn2+ externally and to observe its quenching effects on fluorescent Ca2+ probes loaded in the terminals. Mn2+ (1.8 mM), applied externally after the end of Ca2+-hump, quickly quenched the fluorescence of Indo-1 at two wavelength ranges (Fig. 4,A; n = 4) or the fluorescence of Fura-2 measured at 380 nm (Fig. 4,C; n = 1) (Merritt et al., 1989), indicating the opening of voltage-dependent Ca2+ channels by each impulse during this period. On the other hand, Mn2+ had little effect on Indo-1 or Fura-2 fluorescence when no stimulation was given to the motor nerve (Fig. 4 B, not shown for Indo-1 fluorescence). The results thus indicate that there is neither conduction failure of impulses along motor nerve axons nor the considerable inactivation of Ca2+ channels during a continuous tetanus.

Effects of Blockers of Ryanodine Receptors on Ca2+- and MEPP-Humps

The most likely mechanism for Ca2+- and MEPP-humps could be that CICR occurs in response to Ca2+ entry and ceases due to depletion of Ca2+ stores and/or inactivation of the mechanism. If CICR occurs, it would involve ryanodine receptors rather than inositoltrisphosphate (IP3) receptors (Bezprozvanny et al., 1991; see also Berridge, 1993), since no agonist of receptors to raise IP3 level was given under the present experimental condition. Then, if CICR occurs via the activation of ryanodine receptor, its antagonists, TMB-8 (Chiou and Malagodi, 1975) and ryanodine (Sutko et al., 1979) are expected to block Ca2+- and MEPP-humps. This was the case. TMB-8 (4–8 μM) blocked reversibly MEPP-humps, but not later slow progressive rises in MEPP frequency (Fig. 2,D; n = 11). So did ryanodine (10 μM), but almost irreversibly (n = 4; not shown). Likewise, TMB-8 (10 μM; Fig. 3; n = 3) and ryanodine (n = 1, not shown) completely blocked Ca2+-humps. The results strongly suggest that ryanodine receptors in motor nerve terminals are activated by repetitive Ca2+ entries during a continuous tetanus, causing CICR and transiently increasing [Ca2+]i and MEPP frequency. It might be possible, however, that an increase in [Ca2+]i by the CICR via ryanodine receptors activates phospholipase C to produce IP3, which results in further CICR via IP3 receptors.

It may be noted that TMB-8 (8 μM) decreased the amplitude of MEPP to 81.2 ± 4.8% of the control (n = 5; P < 0.05; see Hunt et al., 1990). This action, however, did not affect much the above suggestion (see materials and methods). The possibility that the decline of MEPP-hump results from the depletion of quanta available for release can be ruled out for two reasons. First, the time course of MEPP-hump was similar to that of Ca2+-hump. Second, the MEPP frequency of several hundreds per second is equivalent to the number of quanta released by nerve impulses at 1 or 2 Hz in normal Ringer solution (Gage, 1976), which would cause little depletion of quanta in nerve terminals.

Effects of Inhibitors of Ca2+ Uptake at Ca2+ Stores

Thapsigargin (1 μM), a blocker of Ca2+ pumps at the endoplasmic reticulum membrane, suppressed Ca2+-humps (to 27.4 ± 4.4%, n = 5; P < 0.001; Fig. 6,A) and MEPP-humps (to 45.5 ± 9.8%, n = 4; P < 0.02; Fig. 6 B). These results indicate that thapsigargin-sensitive Ca2+ stores are involved in CICR in frog motor nerve terminals. The basal levels of [Ca2+]i before tetanic stimulation was not significantly changed in the presence of thapsigargin (89.2 ± 18.9 nM, n = 5; controls, 69 nM). The [Ca2+]i values at the end of Ca2+-hump or at the equivalent time in the presence of thapsigargin (131.3 ± 21.9 nM, n = 5) were also similar to those of the controls (137.2 ± 33.6 nM, n = 5).

Mitochondria are known to retain Ca2+ under resting conditions and also to act as Ca2+ buffers at a relatively high [Ca2+]i level in some types of neurones (Duchen et al., 1990; Thayer and Miller, 1990; Friel and Tsien, 1994; Nohmi et al., 1994). It has also been known that mitochondria of hepatocytes and heart muscles have a permeability-transition pore, large nonselective ion channel (Bernardi et al., 1993; Szabó and Zoratti, 1993), which is activated by an increase in [Ca2+]i (Ichas et al., 1997). It might therefore be possible that such large nonselective ion channels may exist in the mitochondria of frog motor nerve terminals and be activated by Ca2+ entry. The resultant Ca2+ release may have caused Ca2+-humps. Application of a mitochondrial poison, CN (2 mM), to frog motor nerve terminals increased the basal level of [Ca2+]i from 69 to 93.2 ± 4.1 nM (n = 4; P < 0.2). Although the difference was not statistically significant, this may suggest that the mitochondria in frog terminals indeed retain Ca2+ in a certain amount. The poison, however, did not block the Ca2+-hump, but instead enhanced it (to 557.0 ± 119.3% of the control in CN, n = 3; P < 0.05; Fig. 6 C). Furthermore, the basal [Ca2+]i at the end of Ca2+-hump was significantly increased from 99.2 ± 25.7 to 152.1 ± 15.7 nM (n = 4; P < 0.1) in the presence of CN. Thus, it is obvious that mitochondria are not involved in CICR in motor nerve terminals, but in the buffering of the resultant rise in [Ca2+]i.

The Time Course of the Activation of CICR

The component of CICR in the time course of a Ca2+- or MEPP-hump may be demonstrated in three ways. First, it can obviously be shown by a difference between the time courses of [Ca2+]i during a tetanus in the absence and presence of TMB-8 and also by a difference between those of MEPP frequency (see Figs. 2,D and 3). There were similarities between the time courses of the TMB-8-sensitive components of Ca2+- and MEPP-humps. The second method to estimate the time course of CICR activation would be to stop Ca2+ entry, and so the activation of CICR, by a short pause (5 s) in tetanus and to measure the magnitude of resultant reductions in [Ca2+]i or MEPP frequency. Pauses during a MEPP-hump quickly reduced MEPP frequency to a low level (Fig. 7,A), equivalent to that solely caused by Ca2+ entry under the blockade of CICR by TMB-8 (Fig. 2,D). Thus, the magnitude of net reduction in MEPP frequency produced by a pause slowly increased up to a time at the peak of MEPP-hump, and then declined to zero after the end of MEPP-hump so that MEPP frequency was little changed during a short pause after the disappearance of the TMB-8-sensitive component. The time course of the magnitude of pause-induced decrease in MEPP frequency resembled that of the TMB-8-sensitive component of MEPP-hump. The third method to estimate the time course of changes in the rate of Ca2+ release during the course of tetanus could be to measure alterations in the initial rate of rise in MEPP frequency after a pause (see materials and methods). They increased during the rising phase of a MEPP-hump and decreased during the falling phase to a moderate level, which remained constant as long as tetanus continued (Fig. 7 B). (The reason for the incomplete reduction of the initial rate of rise to that at the beginning of tetanus was found to be due to recovery from inactivation of the Ca2+ release mechanism during a pause; see below.) Thus, CICR begins to be slowly activated over minutes and subsides at the rate faster than the growth of its activation. Furthermore, the appearance of the component of CICR at the initial phase of tetanic stimulation indicates that Ca2+ stores in frog motor nerve terminals are normally filled with Ca2+ in an amount at least enough for release in response to Ca2+ entry (see discussion for details).

Ca2+ Entry-dependent Inactivation of the CICR Mechanism

Another important effect of applying pause in tetanus during the falling phase of MEPP-hump is a rebound increase in MEPP frequency after the pause to a value greater than the level of the hump (Fig. 7,A). Furthermore, resuming tetanus after a pause given in a period after MEPP-hump induced a quick increase in MEPP frequency, although there was little change in MEPP frequency during the pause (Figs. 7,A and 8,A). These post-pause increases in MEPP frequency were blocked by TMB-8 (8 μM; n = 3; Fig. 8,B) and thapsigargin (1 μM; to 21 and 19%, n = 2; Fig. 6,B). Likewise, a sharp rise in [Ca2+]i occurred after a pause in tetanus given after the end of a Ca2+-hump, although there was a negligible change during the pause (Figs. 9 and 10,A). This post-pause rise in [Ca2+]i was blocked by ryanodine (not shown; 10 μM; n = 1). Thus, the post-pause increases in MEPP frequency and [Ca2+]i must be caused by the reactivation of CICR. In addition, the fact that stopping tetanus at a time after the end of MEPP- or Ca2+-hump had little effect on MEPP frequency and [Ca2+]i level strongly indicates that CICR is inactivated under this condition. Then, the quick increases in [Ca2+]i and MEPP frequency after the resumption of tetanus suggest that the inactivation of the CICR mechanism is removed during the pause. If so, the extent of recovery from inactivation should depend on the duration of pause. This was indeed the case. The magnitude of the post-pause rise in MEPP frequency was increased, as pause duration was made longer (Figs. 8,A and 10,B). Likewise, the magnitude of post-pause rise in [Ca2+]i was augmented, as pause duration was increased (Figs. 9 and 10, A and B). The results suggest that the recovery of Ca2+ entry-dependent inactivation of CICR occurs rather quickly over several tens of seconds. The possibility that the dependence of post-pause rises in [Ca2+]i and MEPP frequency on pause duration reflects a process of Ca2+ uptake into Ca2+ stores can be ruled out by the lack of a reduction in [Ca2+]i or MEPP frequency during a pause (see discussion).

Ca2+ Entry-dependent Priming and Depriming of the CICR Mechanism

A gradual increase in the rate of Ca2+ release during tetanus, as shown in Figs. 2,B, 3, and 7, A and B, suggests that the mechanism of CICR slowly changes into an activable state in response to tetanus. This process may be called “priming.” The dependence of the slow onsets of MEPP (Fig. 2,B) and Ca2+-humps on [Ca2+]o suggests the role of Ca2+ entry in the mechanism of priming. Fig. 11 shows an experiment to test this possibility. A 50-Hz tetanus was given for 7 min in a Ca2+-free (1 mM EGTA), Mg2+ (1 mM) solution, and then a low Ca2+ (0.2 mM), high Mg2+ (10 mM) solution was superfused for 5 min without applying tetanus. Resumption of tetanus after this “conditioning treatment” caused only a small and slow increase in MEPP frequency (Fig. 11, A a and B a). This small increase by 50 Hz tetanus may presumably be due to depletion of Ca2+ in Ca2+ stores. On the other hand, after a 50-Hz tetanus for 7 min with the subsequent pause for 5 min in the low Ca2+, high Mg2+ solution, MEPP frequency was sharply increased in response to the resumption of tetanus (Fig. 11 Ab, Ad, Bb, and Bd). Thus, Ca2+ entry during “the conditioning tetanus” is necessary for priming the mechanism of CICR in frog motor nerve terminals.

When tetanus was stopped for a sufficiently long period, 79 min, after the end of a MEPP-hump, the subsequent resumption of tetanus produced an increase in MEPP frequency (Fig. 11,Ac) whose rate was similar to that of MEPP-hump (Fig. 2,D). Accordingly, the absence of nerve activity would decrease the readiness for activation of the CICR mechanism. This process may be defined as the depriming of the CICR mechanism. The time course of depriming can be estimated by giving longer pauses of variable duration (1–60 min) after a MEPP-hump. As pause duration was increased, the time course of post-pause rise in MEPP frequency became slower, resembling that of the initial MEPP-hump (Fig. 12,A). The initial rate of post-pause rise to 10% of the maximum amplitude, reflecting the rate of Ca2+ release, slowly decreased over tens of minutes (Fig. 12 B).

Discussion

The present study demonstrates the slow, transient activation of CICR in response to repetitive Ca2+ entries in frog motor nerve terminals. Continuous tetanic nerve stimulation (20–50 Hz) produced slow transient increases in [Ca2+]i in the motor nerve terminals and MEPP frequency in a low Ca2+, high Mg2+ solution, which were reduced in amplitude by lowering [Ca2+]o and decreasing stimulation frequency and blocked by antagonists of ryanodine receptors, TMB-8, and ryanodine, and an antagonist of intracellular Ca2+ pump, thapsigargin. Further analyses of the mode of CICR activation revealed the unique Ca2+ entry-dependent priming, depriming, inactivation, and deinactivation of the CICR mechanism.

The evidence for this conclusion was obtained by independent measurements of MEPP frequency and [Ca2+]i in the motor nerve terminals from different muscle preparations. Accordingly, relevancy and limitations for the use of different muscles must be stated before discussing the evidence and rationale leading to the proposition of the unique mechanisms of CICR in frog motor nerve terminals. First, both the sartorius and cutaneous pectoris muscles of frogs are of fast type. Each muscle fiber is innervated by a single or a few nerve(s) and the nerve terminals are those of an elongated shape (Birks et al., 1960; Peper and Dreyer, 1974). Second, changes in [Ca2+]i and MEPP frequency in response to tetanus and its interruption are quite similar. Third, the modes of actions of drugs that affect Ca2+ release and uptake are essentially identical. An apparent limitation in the interpretation of the data obtained from different preparations would be that changes in [Ca2+]i cannot be directly correlated to those in MEPP frequency.

There are at least three factors that decide the magnitude of CICR from Ca2+ stores; the amount of trigger Ca2+, the rate constant of Ca2+ release (γ in Eq. 1 or 2), and the amount of Ca2+ in Ca2+ stores.

Experimental Parameters Reflecting the Rate of Ca2+ Release

It may be safe to assume that the amount of Ca2+ influx accompanying each impulse during tetanus does not change to an extent affecting the activation of CICR for two reasons. First, there was only a small change in the level of [Ca2+]i increased only by Ca2+ entry during the course of continuous tetanus under the blockade of CICR (Figs. 3 and 6,A). Second, Mn2+ applied externally quenched Indo-1 or Fura-2 fluorescence after the generation of Ca2+-hump (Fig. 4, A and C), indicating the opening of voltage-dependent Ca2+ channels throughout the course of tetanic stimulation. Accordingly, Eq. 2, expressing the rate of Ca2+ release, can be reduced to

 
\begin{equation*}{\gamma}z\;\approx (1+{\kappa}_{s})dx/dt+({\alpha}+p)x.\end{equation*}
5

Thus, changes in the rate of Ca2+ release can be predominantly reflected in either or both of two experimental variables: the levels of the [Ca2+]i (x) or MEPP frequency (the power function of x) via constants (α + p) and the initial rate of rise in [Ca2+]i (dx/dt) or MEPP frequency via constants, (1 + κs). When dx/dt is slow, enough for the steady/equilibrium approximation, the rate of Ca2+ release could be represented predominantly by the level of [Ca2+]i or MEPP frequency, as shown in Eq. 3. On the other hand, when the initial rate of post-pause rise is measured from a relatively low and constant level, it would also reflect the rate of Ca2+ release, as shown in Eq. 4. Relevancy of applying these ideas (see also materials and methods) to the present experimental data was tested.

The rate of changes in [Ca2+]i at the half maximum point of the rising phase of a Ca2+-hump was 10 nM/s (Fig. 3), yielding 4.5 μM/s for (1 + κs)dx/dt in Eq. 5, assuming 450 (90 × 5) of κs (90 for Ca2+ binding proteins taken from measurements by Neher and Augustine, 1992; mitochondrial factor was assumed to be 5 based on the effect of CN on Ca2+-hump). The rate constant for Ca2+ removal (α + p) from frog motor nerve terminals can be estimated from the rate of a rise in [Ca2+]i to a plateau in response to a short tetanus in normal Ringer. A rise in [Ca2+]i produced by 50 Hz in normal Ringer solution reached a plateau in 0.2 s (our unpublished observations). Solving Eq. 1 for γ = 0 (since Ca2+ release channels would not be primed for a short period of 0.2 s), the time constant (0.2 s) to reach the plateau is (1 + κs)/(α + p), yielding a value of 2,255 s−1 [= (1 + 450)/0.2] for (α + p). Then, taking a value of 1.4 μM at the midpoint of the rising phase of Ca2+-hump, (α + p)x during the course of a Ca2+-hump would be 3,157 μM/s (= 2,255 s−1 × 1.4 μM). Thus, (α + p)x is greater than (1 + κs)dx/dt (= 4.5 μM/s) during the rising phase of Ca2+-hump. Accordingly, (α + p)x is the major determinant factor for γz, the rate of Ca2+ release, in Eq. 5, indicating that the level of [Ca2+]i (and also MEPP frequency) itself reflects the rate of Ca2+ release (Eq. 3).

When the initial rates of post-pause rise in [Ca2+]i (dx/dt in Eq. 5) were compared at similar low values of [Ca2+]i during the course of Ca2+-hump, it would also represent the rate of Ca2+ release. This idea may be supported by a relatively small value of (α + p)x at the end of each pause, from which the initial rate of post-pause rise was measured. When the level of [Ca2+]i is as low as a few hundred nanomolars, the time constant of Ca2+ removal is slow and in the order of seconds (our unpublished observations), say 5 s, yielding a value of 90.2 s−1 [= (1 + 450)/5] for (α + p). Then, (α + p)x during a pause in tetanus, during which [Ca2+]i was 0.15 μM, would be 13.5 μM/s (= 90.2 × 0.15 μM/s). This is still larger than (1 + κs)dx/dt measured from a post-pause rise in [Ca2+]i after a Ca2+ hump [say 6.8 μM/s (15 nM/s × 451) from the data shown in Fig. 10,A] and presumably also that during Ca2+-hump, although it was not measured in the present study. Nevertheless, when the rates of Ca2+ release are compared at similar values of [Ca2+]i, their changes may be safely represented by changes in the initial rate of post-pause rise. In support of this, the time course of changes in the initial rate of post-pause rise in MEPP frequency reflecting changes in [Ca2+]i during the course of a MEPP-hump paralleled that of the hump itself (Fig. 7, A and B).

Priming, Depriming, Inactivation, Deinactivation of the Ca2+ Release Mechanism

The rate of Ca2+ release (γz in Eq. 5) is the product of rate constant (γ) for Ca2+ release (the number and open probability of Ca2+ release channels) and [Ca2+]s (z). The rate of Ca2+ release, reflected at the level of MEPP frequency or [Ca2+]i or the initial rate of post-pause rise in MEPP frequency during MEPP- or Ca2+-hump, increased during the rising phase of the humps. In contrast, [Ca2+]s should have decreased during this phase (see the next section). Consequently, the rate constant for Ca2+ release (γ) must have increased during the rising phase of Ca2+- or MEPP-hump. This slow rise in rate constant for Ca2+ release, called the priming of the CICR mechanism, could be due to an increase in the readiness for activation or in open probability or number of Ca2+ release channels. The priming process was found to be dependent on Ca2+ entry, as seen in the dependence of MEPP-hump on [Ca2+]o (Fig. 2,B) and the lack of priming effect of tetanus in a Ca2+-free solution to produce a quick rise in MEPP frequency like a post-pause rise (Fig. 11 A a).

Based on the rationale discussed above, the rate of Ca2+ release decreased during the falling phase of Ca2+- and MEPP-humps and became zero after the end of humps. This is also supported in part by the decrease in the initial rate of post-pause rise in MEPP frequency during the decay phase of MEPP-hump (Fig. 7 B), although it did not decrease to zero because of the effects of pause on the readiness for the activation of Ca2+ release (see below). The slow decrease in the rate of Ca2+ release to zero may be caused by depletion of Ca2+ in stores, the reversal of priming process, or the inactivation of Ca2+ release channels. The generation of post-pause rises in [Ca2+]i or MEPP frequency after the end of Ca2+- or MEPP-hump indicates that Ca2+ stores involved in CICR are filled with Ca2+ in a considerable amount for release under this condition (see below). Furthermore, the fact that the rate of the post-pause rises in [Ca2+]i and MEPP frequency immediately after the end of Ca2+- and MEPP-hump, respectively, was much faster than the rate of rise of each hump suggests that the decay phase of Ca2+- and MEPP-humps cannot be explained by the depriming of the CICR mechanism. (If it were so, the rate of post-pause rise must have been as slow as that of humps.) Accordingly, the CICR mechanism would lie in another state after the end of Ca2+- and MEPP-humps, which is inactivation produced by Ca2+ entry. Then, the increase in the amplitude of post-pause rise in MEPP frequency or [Ca2+]i with an increase in pause duration would therefore represent the rate of recovery from Ca2+-dependent inactivation. Ca2+-dependent inactivation of Ca2+ release channels in neurones was recently suggested by Hernandez-Cruz et al. (1997). The fact that a short pause changed little the level of [Ca2+]i during it, but produced a sharp rise in [Ca2+]i after the pause, suggests that Ca2+ entry itself or the high [Ca2+]i close to the intracellular orifice of voltage-dependent Ca2+ channels at the cell membrane may inactivate Ca2+ release channels, while the absence of Ca2+ entry results in recovery from inactivation in frog motor nerve terminals.

On the other hand, a longer pause (>1 min) decreased the rate of post-pause rise in MEPP frequency to that of MEPP-hump produced by the initial application of tetanus. Thus, this slow dependence of the rate of post-pause rise in MEPP frequency on pause duration would indicate the depriming process of CICR.

The tentative mechanisms of priming, activation, inactivation, and depriming of ryanodine receptor/Ca2+ release channels in frog motor nerve terminals may be summarized as follows. Ryanodine receptor/Ca2+ release channels are slowly primed in response to repetitive Ca2+ entries. Once Ca2+ release channels are primed, they are swiftly activated by subsequent Ca2+ entry. Further continuation of Ca2+ entry inactivates the Ca2+ release channels so that inactivation occurs after priming and/or activation. The Ca2+ entry-dependent inactivation is quickly removed by stopping Ca2+ entry for a short period (<1 min). A longer cessation (>1 min) of Ca2+ entry resulted in the depriming of Ca2+ release channels over tens of minutes.

Changes in the State of Ca2+ Stores Involved in CICR

Changes in the state of Ca2+ stores before and during tetanic stimulation may be estimated from the time course of the TMB-8-sensitive component of MEPP- or Ca2+-hump based on the model in Fig. 1. Since the amount of Ca2+ influx by each nerve impulse is negligible compared with that of Ca2+ release throughout most of Ca2+-hump, the rate of changes in [Ca2+]i (dx/ dt) in the TMB-8-sensitive component of Ca2+- and MEPP-hump must be reflected in changes in the rates of Ca2+ release, uptake, and extrusion as shown in the equation:

 
\begin{equation*}(1+{\kappa}_{s})dx/dt\;\approx {\gamma}z-({\alpha}+p)x.\end{equation*}
6

During the rising phase of TMB-8-sensitive component of Ca2+- or MEPP-hump (dx/dt > 0), the rate of Ca2+ release (γz) must be greater than the sum of the rates of Ca2+ extrusion and uptake [(α + p)x], indicating a progressive decrease in the amount of Ca2+ stores during the rising phase of Ca2+- or MEPP-hump. Furthermore, the fact that the TMB-8-sensitive component of a rise in [Ca2+]i or MEPP frequency occurred without a delay after the beginning of tetanus (see Figs. 2,D and 3) suggests that CICR takes place from the beginning of tetanus. This implies that Ca2+ stores must be normally filled with Ca2+ before the beginning of tetanus.

During the falling phase of MEPP- or Ca2+-hump (dx/dt < 0), the sum of the rates of Ca2+ uptake and extrusion [(α + p)x] must be greater than the rate of Ca2+ release (γz). Although it is not known how much Ca2+ is taken up into Ca2+ stores after the end of Ca2+- or MEPP-hump, there is evidence to suggest the extent of filling of Ca2+ stores after the end of Ca2+- or MEPP-hump. A short pause applied immediately after the humps, which affected little [Ca2+]i or MEPP frequency during it, produced a sharp and significant increase in [Ca2+]i or MEPP frequency after the pause. These post-pause rises in [Ca2+]i and MEPP frequency were blocked by TMB-8 or thapsigargin, indicating that they are induced by CICR from the Ca2+ stores that are involved in the generation of Ca2+- and MEPP-humps. If Ca2+ stores were to be depleted of Ca2+ after the end of Ca2+- or MEPP-hump, Ca2+ must have been taken up into Ca2+ stores during the pause to produce such a CICR. This should have caused a marked reduction in [Ca2+]i during the pause, which is equivalent in time integration to the post-pause rises in [Ca2+]i. Thus, it is likely that a considerable amount of Ca2+ remains in Ca2+ stores after the end of Ca2+- or MEPP-hump. This could be due to a large amount of Ca2+ stored in Ca2+ stores under the resting condition or to an effective Ca2+ uptake during the decay phase of Ca2+- or MEPP-hump. Such a Ca2+ uptake may be achieved by the Ca2+ pump, which is activated by the Ca2+ in high [Ca2+]i domains produced around the site of Ca2+ release, as suggested for Ca2+ stores in bullfrog sympathetic ganglion cells (Cseresnyé et al., 1997).

Comparison with CICRs in Other Neurones

The mode of activation of CICR in frog motor nerve terminals may be compared with those of other neurones in several respects. First, the Ca2+ stores in frog motor nerve terminals are normally filled with Ca2+ under the resting condition. This characteristic resembles that of Ca2+ stores in bullfrog sympathetic ganglion cells (Kuba and Nishi, 1976; Hua et al., 1993), but differs from those in rat sensory neurones (Usachev et al., 1993) and bullfrog sympathetic preganglionic COOH terminals (Peng, 1996). Secondly, the graded activation of CICR in frog motor nerve terminals is similar to those in bullfrog sympathetic ganglion cells (Hua et al., 1993) and rat cerebellar neurones (Llano et al., 1994), but different from the full activation of CICR by Ca2+ entry accompanying a single action potential in rabbit otic ganglion cells (Yoshizaki et al., 1995). Thirdly, Ca2+-dependent inactivation of the CICR mechanism in frog motor nerve terminals is similar to that suggested for mammalian sympathetic neurones (Hernandez-Cruz et al., 1997), but differs in the mode of Ca2+ dependence. In frog motor nerve terminals, Ca2+ release channels appeared to be inactivated by Ca2+ entry, not the [Ca2+]i in the bulk phase, while in mammalian sympathetic ganglion cells CICR seems to be inactivated by the rise in [Ca2+]i produced by Ca2+ release. Finally, the most unique characteristics of CICR in frog motor nerve terminals, which are slow priming and depriming mechanisms, are not seen, to our knowledge, in other types of neurones or in muscles. Thus, these properties in frog motor nerve terminals are quite different from those in other cells.

The Physiological Significance of CICR in the Presynaptic Nerve Terminals

It is not known how the activation of CICR, found in the present study, affects transmitter release from motor nerve terminals during synaptic transmission under the physiological condition. This problem is now under study. Preliminary experiments indicate that the enhancement of end-plate potentials produced by a combination of tetani of different frequencies is blocked by ryanodine (Narita et al., 1998). This is indeed consistent with the previous study that caffeine inhibited the post-tetanic potentiation of evoked release of transmitters, one form of short-term plasticities, at frog motor nerve terminals (Onodera, 1973) and presynaptic terminals in rat hippocampal slices (Lee et al., 1987). The induction of long-term depression in hippocampal neurones was blocked by ryanodine applied externally, but not internally, to postsynaptic neurones (Reyes and Stanton, 1996). Furthermore, Smith and Cunane (1996) recently reported that after the blockade of the fast release of transmitter, ATP, by blocking N-type Ca2+ channels with ω-conotoxin GVIA, high frequency stimulation produced ryanodine-sensitive component of transmitter release. Thus, it is likely that CICR plays some roles in the modulation or plasticity of transmitter release. The novel priming and depriming mechanisms of CICR found in frog motor nerve terminals would be more relevant in general for the mechanism of plasticity in central synapses. The priming and depriming of CICR could be one of the mechanisms for conditioning and unconditioning of plasticity. In this context, search for such a mechanism in central synapses may well be necessary.

Another possible role of CICR at the presynaptic terminals would be the activation of oxidative phosphorylation in mitochondria. An electron transfer inhibitor, CN, markedly enhanced Ca2+-humps, indicating that mitochondrial Ca2+ uptake normally occurs during the tetanus-induced activation of CICR in frog motor nerve terminals. This would imply that the activation of CICR enhances ATP synthesis, since Ca2+-sensitive dehydrogenases play key roles in it (Hansford, 1985; Denton and McCormack, 1990). Furthermore, the oscillations of [Ca2+]i, which were reported to be needed to maintain a high activity of Ca2+-sensitive dehydrogenases (Hajnóczky et al., 1995), were indeed seen occasionally in the present study (data not shown), in the lizard motor nerve terminals (Melamed et al., 1993), and the growth cones of cultured neurones (Gomez et al., 1995; Gu and Spitzer, 1995). Thus, the presynaptic terminals and growth cones, which appear to have high-energy demand, may require CICR to raise mitochondrial Ca2+ concentration and its oscillation in addition to impulse-induced Ca2+ entry.

Acknowledgments

This study was supported in part by a Research Project Grant (K. Narita) from Kawasaki Medical School and Grants in Aid for Scientific Research (K. Kuba) from the Japanese Ministry of Education, Science and Culture.

Abbreviations used in this paper

     
  • CICR

    Ca2+-induced Ca2+ release

  •  
  • IP3

    inositoltrisphosphate

  •  
  • MEPP

    miniature end-plate potential

  •  
  • TMB-8

    8-(N,N-diethylamino)octyl3,4,5-trimethoxybenzoate hydrochloride

references

references
Baker
AJ
,
Brandes
R
,
Schreur
HM
,
Camacho
SR
,
Weiner
MW
Protein and acidosis alter calcium-binding and fluorescence spectra of the calcium indicator Indo-1
Biophys J
1994
67
1646
1654
[PubMed]
Bernardi
P
,
Veronese
P
,
Petronilli
V
Modulation of the mitochondrial cyclosporin A-sensitive permeability transition pore
J Biol Chem
1993
268
1005
1010
[PubMed]
Berridge
MJ
Inositoltrisphosphate and calcium signalling
Nature
1993
361
315
325
[PubMed]
Bezprozvanny
I
,
Watras
J
,
Ehrlich
BE
Bell-shaped calcium response curves of Ins (1,4,5-) P3 - and calcium-gated channels from endoplasmic reticulum of cerebellum
Nature
1991
351
751
754
[PubMed]
Birks
R
,
Huxley
HE
,
Katz
B
The fine structure of the neuromuscular junction of the frog
J Physiol (Camb)
1960
150
134
144
[PubMed]
Cseresnyés
Z
,
Bustamante
AI
,
Klein
MG
,
Schneider
MF
Release-activated Ca2+transport in neurons of frog sympathetic ganglia
Neuron
1997
19
403
419
[PubMed]
Chiou
CY
,
Malagodi
MH
Studies on the mechanism of action of a new Ca2+ antagonist, 8-(N,N-diethylamino)octyl 3,4,5-trimethoxybenzoate hydrochloride in smooth and skeletal muscles
Br J Pharmacol
1975
53
279
285
[PubMed]
Denton
RM
,
McCormack
J
Ca2+as a second messenger within mitochondria of the heart and other tissues
Annu Rev Physiol
1990
52
451
466
[PubMed]
Duchen
MR
,
Valdeoolmillos
M
,
O'Neill
SC
,
Eisner
DA
Effects of metabolic blockade on the regulation of intracellular calcium in dissociated mouse sensory neurones
J Physiol (Camb)
1990
424
411
426
[PubMed]
Eckert
R
,
Tillotson
D
Calcium-mediated inactivation of the calcium conductance in caesium-loaded giant neurones of Aplysia californica.
J Physiol (Camb)
1981
314
265
280
[PubMed]
Elmqvist
D
,
Feldman
DS
Calcium dependence of spontaneous acetylcholine release at mammalian motor nerve terminals
J Physiol (Camb)
1965
181
487
497
[PubMed]
Erulkar
SD
,
Rahamimoff
R
The role of calcium ions in tetanic and post-tetanic increase of miniature end-plate potential frequency
J Physiol (Camb)
1978
278
501
511
[PubMed]
Friel
DD
,
Tsien
RW
A caffeine- and ryanodine-sensitive Ca2+ store in bullfrog sympathetic neurons modulates the effects of Ca2+ entry on [Ca2+]i
J Physiol (Camb)
1992
450
217
246
[PubMed]
Friel
DD
,
Tsien
RW
An FCCP-sensitive Ca2+ store in Bullfrog sympathetic neurons and its participation in stimulus-evoked changes in [Ca2+]i
J Neurosci
1994
14
4007
4024
[PubMed]
Gage
PW
Generation of end-plate potentials
Physiol Rev
1976
56
177
247
[PubMed]
Galli
A
,
Ferroni
A
,
Bertollini
L
,
Mazzanti
M
Inactivation of single Ca2+ channels in rat sensory neurons by extracellular Ca2+
J Physiol (Camb)
1994
477
15
26
[PubMed]
Garaschuk
O
,
Yaari
Y
,
Konnerth
A
Release and sequestration of calcium by ryanodine sensitive stores in rat hippocampal neurones
J Physiol (Camb)
1997
502
13
30
[PubMed]
Gomez
TM
,
Snow
DM
,
Letourneau
PC
Characterization of spontaneous calcium transients in nerve growth cones and their effect on growth cone migration
Neuron
1995
14
1233
1246
[PubMed]
Grynkiewicz
G
,
Poenie
M
,
Tsien
RY
A new generation of Ca2+indicators with greatly improved fluorescence properties
J Biol Chem
1985
260
3440
3450
[PubMed]
Gu
X
,
Spitzer
NC
Distinct aspects of neuronal differentiation encoded by frequency of spontaneous Ca2+transients
Nature
1995
375
784
787
[PubMed]
Hajnóczky
G
,
Robb-Gaspers
LD
,
Seitz
MB
,
Thomas
AP
Decoding of cytosolic calcium oscillations in the mitochondria
Cell
1995
82
415
424
[PubMed]
Hansford
RG
Relation between mitochondrial calcium transport and control of energy metabolism
Rev Physiol Biochem Pharmacol
1985
102
1
72
[PubMed]
Heidelberger
R
,
Heinemann
C
,
Neher
E
,
Matthews
G
Calcium dependence of the rate of exocytosis in a synaptic terminal
Nature
1994
371
513
515
[PubMed]
Hernández-Cruz
A
,
Escobar
AL
,
Jiménez
N
Ca2+-induced Ca2+ release phenomena in mammalian sympathetic neurons are critically dependent on the rate of rise of trigger Ca2+
J Gen Physiol
1997
109
147
167
[PubMed]
Hua
S-Y
,
Nohmi
M
,
Kuba
K
Characteristics of Ca2+ release induced by Ca2+influx in cultured bullfrog sympathetic neurones
J Physiol (Camb)
1993
464
245
272
[PubMed]
Hunt
JM
,
Silinsky
EM
,
Hirsh
JK
,
Ahn
D
,
Solsona
C
The effects of TMB-8 on acetylcholine release from frog motor nerve: interactions with adenosine
Eur J Pharmacol
1990
178
259
266
[PubMed]
Ichas
F
,
Jouaville
LS
,
Mazat
J-P
Mitochondria are excitable organelles capable of generating and conveying electrical and calcium signals
Cell
1997
89
1145
1153
[PubMed]
Katz, B. 1969. The Release of Neural Transmitter Substances. Charles C. Thomas. Springfield, IL.
Katz
B
,
Miledi
R
The role of calcium in neuromuscular facilitation
J Physiol (Camb)
1968
195
481
492
[PubMed]
Kijima
H
,
Kijima
S
‘Steady/equilibrium approximation' in relaxation and fluctuation. I. Procedure to simplify first-order reaction
Biophys Chem
1982
16
181
192
[PubMed]
Konishi
M
,
Olson
A
,
Hollingworth
S
,
Baylor
SM
Myoplasmic binding of fura-2 investigated by steady-state fluorescence and absorbance measurements
Biophys J
1988
54
1089
1104
[PubMed]
Kuba
K
Release of calcium ions linked to the activation of potassium conductance in a caffeine-treated sympathetic neurone
J Physiol (Camb)
1980
298
251
269
[PubMed]
Kuba
K
Ca2+-induced Ca2+release in neurones
Jpn J Physiol
1994
44
613
650
[PubMed]
Kuba
K
,
Hua
S-Y
,
Hayashi
T
A UV laser-scanning confocal microscope for the measurement of intracellular Ca2+
Cell Calcium
1994
16
205
218
[PubMed]
Kuba
K
,
Nishi
S
Rhythmic hyperpolarizations and depolarization of sympathetic ganglion cells induced by caffeine
J Neurophysiol
1976
39
547
563
[PubMed]
Kuba
K
,
Takeshita
S
Simulation of intracellular Ca2+oscillation in a sympathetic neurone
J Theor Biol
1981
93
1009
1031
[PubMed]
Landò
L
,
Zucker
RS
Ca2+ cooperativity in neurosecretion measured using photolabile Ca2+chelators
J Neurophysiol
1994
72
825
830
[PubMed]
Lee
W-L
,
Anwyl
R
,
Rowan
M
Caffeine inhibits post-tetanic potentiation but does not alter long-term potentiation in the rat hippocampal slice
Brain Res
1987
426
250
256
[PubMed]
Llano
I
,
Dipolo
R
,
Marty
A
Calcium-induced calcium release in cerebellar Purkinje cells
Neuron
1994
12
663
673
[PubMed]
Llinás
R
,
Sugimori
M
,
Silver
RB
Microdomains of high calcium concentration at a presynaptic terminal
Science
1992
256
677
679
[PubMed]
Lipscombe
D
,
Madison
DV
,
Poenie
M
,
Reuter
H
,
Tsien
RW
,
Tsien
RY
Imaging of cytosolic Ca2+ transients arising from Ca2+ stores and Ca2+channels in sympathetic neurons
Neuron
1988
1
355
365
[PubMed]
Lockerbie
RO
,
Gordon-Weeks
PR
Further characterization of [3H] γ-aminobutyric acid release from isolated neuronal growth cones: role of intracellular Ca2+stores
Neuroscience
1986
17
1257
1266
[PubMed]
Melamed
N
,
Helm
PJ
,
Rahamimoff
R
Confocal microscopy reveals coordinated calcium fluctuations and oscillations in synaptic boutons
J Neurosci
1993
13
632
649
[PubMed]
Merritt
JE
,
Jacob
R
,
Hallam
TJ
Use of manganese to discriminate between calcium influx and mobilization from internal stores in stimulated human neutrophils
J Biol Chem
1989
264
1522
1527
[PubMed]
Mulkey
RM
,
Zucker
RS
Calcium released by photolysis of DM-nitrophen triggers transmitter release at the crayfish neuromuscular junction
J Physiol (Camb)
1993
462
243
260
[PubMed]
Narita
K
,
Ochi
K
,
Akita
T
,
Kuba
K
Short term plasticity caused by Ca2+ induced Ca2+release in frog motor nerve terminals
Neurosci Res Suppl
1998
22
S97
Neher
E
,
Augustine
AG
Calcium gradients and buffers in bovine chromaffin cells
J Physiol (Camb)
1992
450
273
301
[PubMed]
Nohmi
M
,
Liu
C
,
Kuba
K
Mitochondrial Ca2+-induced Ca2+release in response to cell membrane depolarization in rabbit autonomic neurones
Neurosci Res Suppl
1994
19
S27
Ohta
Y
,
Kuba
K
Inhibitory action of Ca2+ on spontaneous acetylcholine release at motor nerve terminals in a high K+solution
Pflügers Arch
1980
386
29
34
[PubMed]
Onodera
K
Effect of caffeine on the neuromuscular junction of the frog, and its relation to external calcium concentration
Jpn J Physiol
1973
23
587
597
[PubMed]
Osanai
M
,
Suzuki
S
,
Suzuki
N
,
Kuba
K
,
Kijima
H
Presynaptic Ca2+-imaging at the frog neuromuscular junction
Neurosci Res
1996
20
S36
Peng
Y-Y
Ryanodine-sensitive component of calcium transients evoked by nerve firing at presynaptic nerve terminals
J Neurosci
1996
16
6703
6712
[PubMed]
Peng
Y-Y
,
Zucker
RS
Release of LHRH is linearly related to the time integral of presynaptic Ca2+elevation above a threshold level in bullfrog sympathetic ganglia
Neuron
1993
10
465
473
[PubMed]
Peper
K
,
Dreyer
F
,
Sandri
C
,
Akert
K
,
Moor
H
Structure and ultrastructure of the frog motor end-plate. A freeze-etching study
Cell Tissue Res
1974
149
437
455
[PubMed]
Ravin
R
,
Spira
ME
,
Parnas
H
,
Parnas
I
Simultaneous measurement of intracellular Ca2+and asynchronous transmitter release from the same crayfish bouton
J Physiol (Camb)
1997
501
252
262
Reyes
M
,
Stanton
PK
Induction of hippocampal long-term depression requires release of Ca2+from separate presynaptic and postsynaptic intracellular stores
J Neurosci
1996
16
5951
5960
[PubMed]
Szabó
I
,
Zoratti
M
The mitochondrial permeability transition pore may comprise VDAC molecules
FEBS Lett
1993
330
201
205
[PubMed]
Schweizer
FE
,
Betz
H
,
Augustine
AG
From vesicle docking to endocytosis: intermediate reactions of exocytosis
Neuron
1995
14
689
696
[PubMed]
Südhof
TC
The synaptic vesicle cycle: a cascade of protein– protein interactions
Nature
1995
375
645
653
[PubMed]
Smith
AB
,
Cunnane
TC
Ryanodine-sensitive calcium stores involved in neurotransmitter release from sympathetic nerve terminals of the guinea-pig
J Physiol (Camb)
1996
497
657
664
[PubMed]
Sutko
JL
,
Willerton
JT
,
Templeton
GH
,
Jones
LR
,
Besch
HR
Ryanodine: its alteration of cat papillary muscle contractile state and responsiveness to inotropic interventions and a suggested mechanism of action
J Pharmacol Exp Ther
1979
209
961
964
Tanabe
N
,
Kijima
H
Ca2+-dependent and -independent components of transmitter release at the frog neuromuscular junction
J Physiol (Camb)
1992
455
271
289
[PubMed]
Thayer
SA
,
Hirning
LD
,
Miller
RJ
The role of caffeine-sensitive calcium stores in the regulation of the intracellular free calcium concentration in rat sympathetic neurons in vitro.
Mol Pharmacol
1988
34
664
673
[PubMed]
Thayer
SA
,
Miller
RJ
Regulation of the intracellular free calcium concentration in single rat dorsal root ganglion neurones in vitro.
J Physiol (Camb)
1990
425
85
115
[PubMed]
Tóth
PY
,
Török
TL
,
Magyar
K
Depolarization promotes caffeine induced [3H]-noradrenaline release in calcium-free solution from peripheral sympathetic nerves
Cell Calcium
1990
11
557
563
[PubMed]
Usachev
Y
,
Shmigol
A
,
Pronchuk
N
,
Kostyuk
P
,
Verkhratsky
A
Caffeine-induced calcium release from internal stores in cultured rat sensory neurons
Neuroscience
1993
57
845
859
[PubMed]
Yoshizaki
K
,
Hoshino
T
,
Sato
M
,
Koyano
H
,
Nohmi
M
,
Hua
S-Y
,
Kuba
K
Ca2+-induced Ca2+release and its activation in response to a single action potential in rabbit otic ganglion cells
J Physiol (Camb)
1995
486
177
187
[PubMed]
Zucker
RS
Exocytosis: a molecular and physiological perspective
Neuron
1996
17
1049
1055
[PubMed]

Portions of this work were previously published in abstract form (Narita, K., K. Kuba, T. Mitsumoto, T. Shirasaki, M. Osanai, and H. Kijima. 1996. Neurosci. Res. 20:S49).

Author notes

Address correspondence to Dr. Kenji Kuba, Department of Physiology, School of Medicine, Nagoya University, 65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan. Fax: +81-52-744-2049; E-mail: kubak@med.nagoya-u.ac.jp