| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1The Otto Loewi Minerva Center for Cellular and Molecular Neurobiology, Department of Neurobiology, Hebrew University of Jerusalem, Jerusalem; 2Department of Neurobiology, Weizmann Institute of Science, Rehovot, Israel; and 3Laboratory of Bioorganic Chemistry, National Institute of Diabetes and Digestive Kidney Diseases, Bethesda, Maryland
Submitted 3 November 2004; accepted in final form 3 February 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
), one of which involves direct modulation of proteins of the release machinery (for example, Blackmer et al. 2001; Capogna et al. 1996; Scholz and Miller 1992; Silinsky 1984; Trudeau et al. 1996; for review see Miller 1998).
We recently proposed that control of the time course of neurotransmitter release, both of initiation and termination, is achieved by a direct effect of presynaptic inhibitory autoreceptors on key proteins of the release machinery (Parnas et al. 2000). In particular, we proposed that at rest (i.e., at resting potential and resting low level of transmitter in the cleft) the release machinery is under tonic block. The block is achieved as a result of an interaction of the transmitter-occupied inhibitory autoreceptor with the core proteins of the release machinery. At resting potential the receptor is in a high-affinity state and thus can be occupied even at the low tonic concentration of transmitter. Initiation of release is achieved by depolarization shifting the autoreceptor to a low-affinity state, resulting in fast dissociation of the transmitter from the receptor. The free autoreceptor rapidly detaches from the proteins of the release machinery; the free proteins of the release machinery, together with the Ca2+ that had entered on depolarization, then act together to promote release. Termination of release occurs on membrane repolarization, which causes the autoreceptors to revert rapidly to their high-affinity state, rebind transmitter, and reassociate with the release machinery. This results in release being blocked. (In the DISCUSSION we mention and discuss experiments that seem to contradict our hypothesis.)
The above hypothesis is based on experiments implicating the presynaptic M2-muscarinic receptor (M2R) in the release of acetylcholine (ACh) at neuromuscular junctions. Detailed justification for the various assumptions involved in our hypothesis are presented in Parnas et al. (2000). Here we list only the key experimental findings supporting it.
). Furthermore, interaction was seen only when the M2R was occupied by a nonhydrolyzable muscarinic receptor agonist, muscarine. Addition of methoctramine, a specific M2/M4 antagonist, abolished the interaction (Ilouz et al. 1999). ; Morita et al. 1982; Peteris and Ogren 1988; Slutsky et al. 1999). ).
A formal mathematical model of our hypothesis was provided by Yusim et al. (1999). This model was further tested, and some of its key assumptions were supported by experimental results. In particular, it was shown that if binding of ACh to the M2R is retarded, termination of evoked ACh release on repolarization is slowed, as predicted by the model (Slutsky et al. 2001). Furthermore, in wild-type (WT) mice the time course of ACh release is insensitive to experimental manipulations known to affect Ca2+ entry and removal. In contrast, in M2R knockout mice, these same experimental manipulations significantly altered the time course of ACh release; release was briefer when Ca2+ removal was accelerated and the duration of release was prolonged when more Ca2+ had entered (Slutsky et al. 2003). Furthermore, it was shown that these changes in the time course of release were not the result of changes in Ca2+-current kinetics (Slutsky et al. 2003). These results were interpreted by Slutsky et al. (2003) as meaning that in the M2R knockout mice, in which the tonic block by the receptor cannot be imposed, the release machinery is permanently free. Thus the time course of ACh release is determined by the second limiting factor, that is, influx and removal of Ca2+, whereas in WT mice it is determined by the M2R. This interpretation is further supported by data showing that release in M2R knockout mice starts earlier and lasts longer than that in WT mice (Slutsky et al. 2003).
Here we subject our hypothesis concerning initiation of release to a further test. Specifically, we examine whether depolarization indeed plays 2 roles in the initiation of phasic release: opening of voltage-gated Ca2+ channels and relief of a tonic block imposed by the transmitter-occupied inhibitory autoreceptor M2R in the preparations studied. To reveal this additional role of depolarization, the terminal was depolarized to different levels; the relationship between quantal content and the depolarizing pulse amplitude (PA) was then measured under control conditions and under conditions in which the M2R-imposed tonic block had been drastically reduced or completely abolished.
The rationale underlying these experiments is as follows: One of the means of establishing the role of Ca2+ in promoting release has been to measure, at constant depolarization, the relationship between the quantal content m and Ca2+ ([Ca2+]o, Dodge and Rahamimoff 1967; ICa, Augustine et al. 1985; [Ca2+]i, Bollmann et al. 2000; Ravin et al. 1999; Schneggenburger and Neher 2000). The log/log plot of quantal content versus [Ca2+]o had a slope of about 4, indicating that 4 Ca2+ ions are required for the release of one quantum (Dodge and Rahamimoff 1967). To establish the role of depolarization in promoting release, in the experiments below we measure the dependency of m on depolarization at fixed [Ca2+]o. If depolarization indeed plays the above-mentioned 2 roles in promoting neurotransmitter release then this relationship, expressed as the maximal slope of log m versus log PA, should reflect the quantitative contributions of both roles of depolarization, that is, in Ca2+ entry and in relief of the tonic block. It is thus predicted that the slope of log m/log PA should be reduced when the additional role of depolarization, relieving the tonic block, is abolished. It should then reflect only the role of depolarization in permitting Ca2+ influx.
Such experiments were performed both on the frog neuromuscular junction (nmj) and on those of wild-type and M2R knockout mice, and the predictions made above were confirmed.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Frogs. Frogs (Rana ridibunda) were killed by stunning and double pitting in accordance with institutional guidelines and the Israel animal protection law. The cutaneous pectoris neuromuscular preparation was isolated and pinned in a chamber with a Sylgard bottom and shallow walls (0.4 x 1.5 x 4 cm3). The chamber was secured on the stage of an upright microscope (Axioscope, Zeiss), which was modified to hold the micromanipulators also. The chamber was continuously perfused (Gilson Minipulse 3 pump) with a bathing solution that passed through a cooling device. The temperature was controlled at 8–10 ± 1°C. The standard Ringer solution contained (in mM): NaCl, 116; KCl, 2; MgCl2, 1; CaCl2, 1; Tris, 5. The pH was adjusted to 7.4 with NaOH. (Small changes in CaCl2 or MgCl2 were not compensated for.)
Mice.
M2R knockout mice (M2-KO) (Gomeza et al. 1999), 1.5–3 mo of age, were used. They had a mixed genetic background (129J1 X CF-1; 50%/50%). Age-matched wild-type (WT) mice of the same genetic background served as controls. Mouse colonies were amplified at Taconic Farms (Germantown, NY).
Mice were anesthetized with CO2 and decapitated, in accordance with institutional guidelines and the Israel animal protection law. Hemidiaphragm neuromuscular preparations were isolated and pinned in the chamber described above. The standard bathing solution contained (in mM): NaCl, 160; KCl, 2,5; MgCl2, 1; CaCl2, 3; Hepes 10, glucose, 8, bubbled with 95% O2-5% CO2. The temperature (30 ± 1°C) was maintained by circulating (Gilson Minipulse 3 pump) the fluid through a heat exchanger. The pH was adjusted to 7.4 with NaOH.
Stimulation and recording
The macropatch technique (Dudel 1981) was used for local depolarization of a small region of the terminal and for concomitant recording of single quanta events. With this technique depolarization is produced by shifting the extracellular potential to more negative values. This is done by passing constant negative current pulses, which can vary in amplitude and duration, through the macropatch electrode. The seal resistance (180–200 k
) determines the maximal possible depolarization of the membrane of the terminal. The time constant of the depolarization thus produced does not depend on the resistance–capacitance (RC) of the membrane, but rather on the time constant of the amplifier (10 kHz, 0.1 ms); thus for brief current pulses of 0.1 ms, the shift in external potential is truncated (Dudel 1981).
The macropatch technique suffers from a major drawback, inasmuch as the depolarization level is not known. The current that passes through the electrode is shunted and not all of it reaches the extracellular space around the terminal. One way to roughly estimate the level of depolarization is to measure what level of current pulse produces a similar quantal content to that produced by an action potential. However, the amplitude and duration of the action potential at the nerve terminal are not known. Nevertheless, because the frog nerve terminal is excitable (Katz and Miledi 1965a), we estimate the duration of the action potential and its amplitude to be 1–1.2 ms and 100 mV, respectively. At the same recording site [after addition of 0.2 µM tetrodotoxin (TTX)] we applied current pulses of 1- to 1.2-ms duration and varying amplitudes. In the example given in Fig. 1, the quantal content (m) for nerve stimulation (NS) was 0.66. A pulse of 1.2 ms and –0.5 µA [direct stimulation (DS)] produced a similar quantal content (m = 0.68). In 2 additional experiments, for a pulse of 1 ms, –0.7 µA was required to produce release similar to that of an action potential. Because the seal resistance was 200 k the maximal possible depolarization for the 3 experiments was 100–140 mV. Thus there must be some shunting (
40%) of the current pulse. Qualitatively, we can assume that larger current pulses produce stronger depolarization and greater release (Fig. 2). (For further technical details see Dudel 1981; Ravin et al. 1997; see also Katz and Miledi 1965b.)
|
|
). The fire-polished tip (about 8 µ) was slightly bent to allow positioning of the macropatch electrode over a branch of the endplate in the small space between the objective and the preparation. Pulse duration was 0.3 ms for mice and 0.7 ms for frogs, pulse amplitude usually varied between –0.3 and –0.7 µA, and frequency of stimulation was 1–3 Hz. TTX (0.2 µM) was present in the solution to prevent sodium excitability, thus permitting graded depolarization of the axon terminal (Dudel 1981). Determination of quantal content
At 8–10°C (frog experiments), and even at 30°C (mouse experiments), the quanta appeared after the stimulus artifact, and could be easily discerned and counted even when 2 or 3 quanta were released concomitantly (Figs. 1 and 2). Figure 2 shows, for the mouse nmj, that for any pulse, irrespective of its amplitude, the number of quanta released after each depolarizing pulse varies. Furthermore, for a pulse of a constant duration (0.4 ms), the quantal content increased with current pulse amplitude. To determine the quantal content, the quanta were counted for a period of 10 ms after the beginning of each depolarization pulse (2,000 pulses for m <0.25 and 512 for m >0.3). The total number of quanta divided by the number of applied pulses yields the quantal content (the average number of quanta released per pulse). Such a procedure was used for several pulse amplitudes (given randomly) at the same recording site (Fig. 2). The number of spontaneously released quanta was counted for the period starting 10 ms after the pulse until the following pulse. The number of spontaneous releases varied in the various preparations and was higher in M2-KO mice than that in WT mice. It also varied according to the experimental conditions. However, in all cases the probability of spontaneous release occurring during the period of 10 ms after each pulse was <0.0001 (2–3 quanta for 200 s), precluding the need to subtract spontaneous from evoked release. In the experiment depicted in Fig. 2, 4 pulse amplitudes, given in a random manner, were used. The quantal content for each pulse amplitude is given below each column of traces (m was determined for 2,000 pulses). Figure 2B shows that the slope of log m/log PA was 6.8. Current traces were digitized using a Neurodata (Neuro-Corder DR-484) A/D converter at 50 kHz and transferred to a Pentium III-500 computer using a Labview (AT-MIO-16F-5,NIDAQ 4.9.0 driver software) interface.
|
The G2 dimeric form of AChE was purified from Torpedo californica electric organ by affinity chromatography, after solubilization with phosphatidylinositol-specific phospholipase C (Futerman et al. 1985). Its specific activity was about 3,000 units per mg protein, one unit corresponding to hydrolysis of 1 µmol min–1 of acetylthiocholine, assayed according to Ellman et al. (1961) (see also Slutsky et al. 2001).
Statistical evaluation
Significance was checked by the Student’s paired (the same experiment) and unpaired (different experiments) 2-tail t-test. Results are given as means ± SD throughout.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
We begin by testing whether depolarization does indeed play 2 roles in initiation of phasic release, that is, opening of voltage-gated Ca2+ channels and relief of a tonic block of release imposed by ACh-occupied M2R. To do so, we compared the dependency of the quantal content (m) on depolarizing pulse amplitude (PA) at fixed [Ca2+]o (expressed as the slope of log m/log PA), both under control conditions and in the presence of the selective M2R/M4R antagonist methoctramine. As an M2R antagonist, methoctramine is expected to reduce, or even abolish, the M2R-imposed tonic block of ACh release, resulting in reduction of the slope of log m/log PA.
Figure 3A shows the average results for 10 such experiments. Because both M1R and M2R are present at frog nerve terminals, and M1R enhances release (Slutsky et al. 1999), all experiments were conducted in the presence of the M1R antagonist pirenzepine (10 µM). First, the control quantal content was measured at several low-to-medium PAs administered in a random manner. After establishing the control curve (Fig. 3A, filled squares), methoctramine (1 µM) was applied and the experimental protocol was repeated. As shown earlier (Slutsky et al. 1999), methoctramine increased the quantal content in a voltage-dependent manner (Fig. 3A, open squares). At a low-pulse amplitude (–0.3 µA) methoctramine increased the quantal content by 251%, but at a larger amplitude (–0.5 µA), by only 3%. Thus the slope of log m/log PA declined in the presence of methoctramine (see slope values below). An M2R agonist should compete with methoctramine for binding to M2R, thus diminishing or abolishing its effect. Indeed, when 100 µM muscarine was added to the methoctramine, the curve relating m to PA was similar to that obtained in the control (Fig. 3A, open circles).
|
AChE also reduces the maximal slope of log m/log PA
Another way to diminish occupancy of the M2R by ACh, and thereby to reduce the tonic inhibition, is to reduce the concentration of ACh in the synaptic cleft by adding the potent ACh-hydrolyzing enzyme AChE (Slutsky et al. 2001). Thus the experiments described in Fig. 3A were repeated, but in the presence of 80 µg/ml AChE (Fig. 3B). After establishing the control curve (Fig. 3B, average of 3 experiments, filled squares) AChE was added and 5 min later the dependency of m on PA was measured (Fig. 3B, open squares). AChE, similarly to methoctramine, increased the quantal content in a depolarization-dependent manner: an increase of 277% at –0.3 µA versus an increase of 1% at –0.45 µA. Thus AChE, like methoctramine, reduced the slope of log m/log PA (see following text). Again, addition of muscarine restored the slope to control values.
Specifically, the slope of log m/log PA was 7.0 ± 0.2 in the control (Fig. 3B, filled squares); it was reduced to 4.2 ± 0.3 after addition of AChE (Fig. 3B, open squares; extremely significant, P < 0.0001) and increased to the control slope (7.0 ± 0.44, P > 0.142; Fig. 3B, open circles) after subsequent application of muscarine.
Two initial conclusions may be drawn: 1) methoctramine and AChE indeed reduce a tonic block of release imposed by the M2R; 2) high levels of depolarization completely relieve the tonic block. This second conclusion derives from the finding that neither methoctramine nor AChE produce elevation of release at high pulse amplitudes.
The slope of log m/log PA is higher in WT than in M2-KO mice
Experiments similar to those shown in Fig. 3 were performed on diaphragm muscles taken from 2 mouse strains: wild-type (WT) mice possessing functional M2R and knockout mice lacking functional M2R (M2-KO) (Gomeza et al. 1999; Slutsky et al. 2003). We predicted that in the M2-KO mice, in which the M2R is not functional, the release machinery should constantly be in a "free" state. Thus the additional role of depolarization in relieving the tonic block should not be observed and, consequently, the behavior of the M2-KO mice should be similar to that obtained in both the frog and in WT mice in the presence of methoctramine or AChE. In particular, release should be higher in M2-KO than in WT mice, but only at the low-pulse amplitudes and, consequently, the slope of log m/log PA should be lower in M2-KO mice than in WT mice.
The data presented in Fig. 4 confirm this prediction. Figure 4 presents average quantal contents measured in WT (n = 13, Fig. 4A, filled squares) and in M2-KO mice (n = 10, Fig. 4A, open squares) at 4 (low-to-medium) depolarizing pulse amplitudes administered in a random manner. At low-pulse amplitudes the quantal content, on average, was higher in the M2-KO than in the WT mice (Fig. 4A). The average m values in WT mice were: 0.06 ± 0.01, 0.137 ± 0.03, 0.28 ± 0.10, and 0.44 ± 0.10 at –0.5, –0.55, –0.6, and –0.65 µA, respectively. The average m values in M2-KO mice were: 0.14 ± 0.01, 0.19 ± 0.03, 0.27 ± 0.05, and 0.44 ± 0.07 for the same pulse amplitudes, respectively. Thus release was about 2.3-fold higher at –0.5 µA, whereas at higher PA values the quantal contents were very similar in M2-KO and WT mice (P > 0.6). The average slope of log m/log PA was 7.47 ± 0.56 (n = 13) in WT mice (Fig. 4A, filled squares), but only 4.34 ± 0.31 (n = 10, P < 0.0001) in M2-KO mice (Fig. 4A, open squares), a slope similar to that obtained in the frog preparation after application of either methoctramine or AChE (Fig. 3).
If the lower slope indeed reflects a lack of tonic inhibition of the release machinery in M2-KO mice, addition of methoctramine should not further reduce the slope. In contrast, addition of methoctramine to WT mice should reduce the slope, as was the case in the frog. Figure 4B shows that in WT mice the control slope of 7.47 ± 0.56 was reduced to 4.35 ± 0.36 after addition of methoctramine. In contrast, in M2-KO mice the control slope of 4.34 ± 0.31 remained unaltered after addition of methoctramine (4.3 ± 0.21) (n = 8).
In frog, a strong and brief depolarizing prepulse increases test pulse (low-to-medium amplitude) release and reduces the slope of log m/log PA
The results presented thus far are compatible with the hypothesis according to which the agonist-occupied M2R imposes a tonic block, at resting potential, and that for initiation of ACh release to take place, this block must be alleviated. Furthermore, the reduction in the slope of log m/log PA on addition of methoctramine, or in the presence of AChE, is in accordance with the notion that it is strong depolarization that alleviates the M2R-imposed tonic block, thus permitting release of ACh to occur.
Because this last notion is our principal contention in the present study, we sought to substantiate the role of depolarization in relieving the block by an independent set of experiments. Slutsky et al. (1999, 2002) previously showed that the M2R-imposed inhibition of ACh release is Ca2+-independent but voltage-dependent, where inhibition is strong at low depolarizations, declining as depolarization increased, and being completely abolished at high depolarizations. It follows that, if a depolarizing test pulse (of low-to-medium PA) is preceded by a strong depolarizing prepulse, it will produce release higher than a test pulse alone. This is because a depolarizing test pulse (of low or medium amplitude) administered shortly after a strong depolarization should encounter release machinery that had been "freed" by the strong depolarization that preceded the test pulse. We thus conducted experiments similar to those used to study the mechanism of membrane-delimited inhibition of Ca2+ channels mediated by G-protein–coupled receptors (GPCRs) (see, e.g., Kuo and Bean 1993; for reviews, see Arnot et al. 2000; Bertram and Behan 1999; Dolphin 1998; Hille 1994; Ikeda 1996; Jarvis et al. 2000; Zamponi and Snutch 1998). Accordingly, we compared release produced by a low test pulse alone with release when it was preceded by a strong depolarizing prepulse.
Considerations underlying choice of the experimental protocol for prepulse administration
A strong depolarization both relieves inhibited Ca2+ channels (see references above) and activates the proteins of the release machinery, either by dissociating them from M2R (Ilouz et al. 1999; Linial et al. 1997) or by an unknown Ca2+-independent mechanism (Slutsky et al. 1999, 2002). Thus if we observe prepulse-mediated enhancement of release it may be attributable either to disinhibition of Ca2+ channels or to detachment of the release machinery from the M2R, or to both.
In an attempt to at least partially distinguish between the 2 possibilities, we took into account that a relatively long prepulse, in the range of a few milliseconds, is required for disinhibition of Ca2+ channels. Furthermore, the time constant of reinhibition on repolarization is also long, in the range of 50–100 ms (Arnot et al. 2000; Kasai 1992). We therefore tested whether a strong, but very brief depolarizing prepulse, that would not be likely to cause disinhibition of Ca2+ channels, would nevertheless enhance release by a subsequent test pulse.
Accordingly, the experiments shown in Figs. 3 and 4 were repeated, but with a strong and brief depolarizing prepulse (–1.2 µA, 0.1-ms duration, denoted hereafter as the "standard prepulse") preceding the test pulse.
The alternating stimulus regime (Slutsky et al. 2003) was used. As a control, we randomly administered test pulses of various amplitudes, such that the control (no prepulse, Fig. 5 A, middle column) and the experiments (with prepulse, Fig. 5A, right column) were given successively. This ensured that the quantal contents of all test pulses, with and without a prepulse, would be determined over the same period of time.
|
To obtain a maximal effect of the prepulse, it should ideally be administered with zero interval before the test pulse. Under such conditions, however, the prepulse should be considered as being composed of 2 components, each having a different effect on the test pulse. The first component, having the same amplitude as that of the test pulse (Fig. 5, B and C, inset, gray column), should not be expected to relieve the tonic block more than the test pulse itself (Slutsky et al. 1999; Yusim et al. 1999; Zamponi and Snutch 1998), but should increase Ca2+ influx during the test pulse simply by increasing its duration. We denote this effect of the prepulse as the "prolongation" effect. The prepulse is indeed very brief (0.1 ms), but so is the test pulse (0.7 ms for frog and 0.3 ms for mouse). Consequently, by combination with the prepulse, the test pulse is prolonged by 15 and 25% in frog and mouse, respectively.
The second component, the higher amplitude of the prepulse (Fig. 5, B and C, inset, empty column), is responsible for relief of the tonic block, and it is in this component that we are interested. Figure 5, B and C presents an experimental protocol designed to distinguish between the 2 components.
Figure 5, B and C shows the average results for 5 experiments. Figure 5B shows the control curve for log m/log PA (filled squares). Application of the standard prepulse (–1.2 µA, 0.1 ms) with zero interval before the test pulses (open squares) increased m in a voltage-dependent manner; thus m increased 11-fold at a PA of –0.3 µA and 2.4-fold at –0.7 µA. When the prepulse had an amplitude identical to that of each of the test pulses (shaded squares), its effect was significantly different from that of the strong (–1.2 µA) standard prepulse. It increased test-pulse release in a voltage-independent manner. Furthermore, the increase was much smaller (average increase about 2-fold) than the maximal increase (at low-pulse amplitudes) produced by the 1.2-µA prepulse (11-fold).
The same experiment, at the same recording site, was repeated, but with a 1-ms interval between the pre- and test pulses. Figure 5C shows that with 1-ms interval the "prolongation" effect (shaded squares) was completely abolished; however, the strong (–1.2 µA) standard prepulse (empty squares) still significantly increased release, with the increase being largest at the low-pulse amplitudes, as in Fig. 5B. The data presented in Fig. 5, B and C thus indicate that the prepulse does indeed enhance release by 2 mechanisms. One, arising from prolongation of the test pulse, is short-lived, dissipating within about 1 ms. The other mechanism, relief of the tonic block, is primarily responsible for enhancement of test-pulse release and persists for several milliseconds (see following text).
The actual prepulse experiments
To avoid possible distortion of the results arising from the "prolongation" effect of the prepulse, in the following experiments we used an interval of 1 ms or more between the pre- and test pulses. Figure 6 depicts the results obtained, with the experimental protocol being shown in the inset of Fig. 6A. Figure 6A shows that the standard prepulse (–1.2 µA, 0.1 ms), administered with intervals of –1, –2, and –4 ms before the test pulses, increased the quantal content of test-pulse release in a depolarization-dependent manner, similarly to methoctramine and AChE. For example, with a 1-ms interval, m increased 16-fold for a –0.4-µA test pulse, but only 2.5-fold for a –0.7-µA test pulse. Furthermore, the increase in release produced by the prepulse declined as the interval between the prepulse and the test pulse increased. The slope of log m/log PA decreased from 7.3 ± 0.82 in the control (Fig. 6A, filled squares) to 4.44 ± 0.89 (Fig. 6A, open squares) when the standard prepulse was administered 1 ms before the test pulse. With a 4-ms interval (open triangles), the slope was 6.8 ± 0.7. Figure 6B depicts the dependency of the slope (taken from the data shown in Fig. 6A) on the time interval between the pre- and test pulses. It can be seen that the effect of the prepulse was essentially abolished within about 4–5 ms.
|
Although the results displayed in Fig. 6 fulfill our theoretical predictions that the prepulse enhances test-pulse release resulting from a relief of a tonic block, it is still necessary to consider the possibility that, rather than relieving a tonic block, the prepulse (even though it did not produce release itself) admitted a small amount of Ca2+, a residual amount of which produced facilitation (Katz and Miledi 1968) of the test-pulse release. To check for this possibility, we applied a train of 20,000 pulses at 100, 200, and 300 Hz, with the parameters of the standard prepulse. It would be expected that if some Ca2+ were to enter during the prepulse, it would accumulate. Thus it would further be expected that increases in evoked and spontaneous release should be seen. Furthermore, the rates of both modes of release should increase progressively along the time course of the train.
The results obtained do not confirm these expectations. In one such experiment only 3 spontaneously released quanta (no stimulation) were counted for a period of 200 s. On stimulation at the same recording site, the number of quanta recorded for 20,000 pulses at 100, 200, and 300 Hz were 3, 2, and 3, respectively, and these quanta were not "locked" to the stimulus. Thus the "quantal content" was the same as the rate of spontaneous release, about 0.0001. Similar results were obtained in an additional 3 experiments in which 10,000 pulses were administered. These results, together with those displayed in Fig. 6, indicate that no significant entry of Ca2+ occurred during the standard prepulse. Thus the prepulse-mediated enhancement of release seen in Fig. 6 cannot be ascribed to entry of Ca2+ during the prepulse. The results presented in Fig. 9, which show that the standard prepulse did not affect twin-pulse facilitation, further corroborate this conclusion.
|
Irrespective of the exact mechanism, the data shown in Fig. 6 are compatible with the notion that the strong and brief depolarizing prepulse, similarly to methoctramine and AChE, relieves a tonic block of release imposed by the transmitter-occupied M2R. To further test this conclusion we examined whether application of methoctramine + AChE would occlude the effect of the prepulse. Figure 7 shows that this is indeed the case. Methoctramine + AChE increased release, as before, in a depolarization-dependent manner and, consequently, reduced the slope of log m/log PA from 7.7 ± 0.81 (filled squares) to 5.34 ± 0.35 (filled diamonds, n = 7). Subsequent application of the prepulse did not further increase release and did not further reduce the slope (open diamonds). We may thus conclude that both the standard prepulse and either methoctramine or AChE affect release by the same mechanism, which is likely to be relief of a tonic block imposed by the M2R.
|
Another way to confirm whether the depolarizing prepulse indeed relieves a tonic block imposed by the M2R is to study its effect in M2-KO mice lacking functional M2R. Accordingly, we repeated the experiments done in frog for both WT (6 experiments) and M2-KO mice (3 experiments). The results obtained in WT mice closely resemble those obtained in frog (Fig. 8, A–C). In particular, the standard prepulse, when given at zero delay, exhibits both the "prolongation effect" (Fig. 8A, shaded squares) and the relief from tonic block (Fig. 8A, open squares). As in frog, the control slope of 7.6 ± 0.2 decreased to 4.59 ± 0.15 when the standard prepulse was applied. Furthermore, as in the frog, the prolongation effect is both relatively small and independent of pulse amplitude (slope = 7.59 ± 0.12). Also, as in the frog, the prolongation effect was abolished if there was a 1-ms interval between the pre- and test pulses (Fig. 8B, shaded squares).
|
Finally, Fig. 8C shows that, as in frog, the prepulse had no additive effect when administered after perfusion with methoctramine + AChE. The control slope (filled squares) of 7.44 ± 0.16 decreased to 4.54 ± 0.11 in the presence of methoctramine + AChE (shaded diamonds), and remained at 4.55 ± 0.14 after subsequent application of the standard prepulse with a 1-ms interval (open diamonds).
The observations for M2-KO mice (n = 3) were radically different. In the M2-KO mice only the prolongation effect of the prepulse persisted (Fig. 8D, shaded squares) and this effect, as in both frog and WT mice, disappeared when the prepulse preceded the test pulse by 1 ms (Fig. 8E, shaded squares; Slutsky et al. 2003). This result further supports our earlier report that properties of Ca2+ channels are similar in M2-KO and WT mice.
The second component of the prepulse effect, relief from a tonic block, is completely absent in the M2-KO mice (Fig. 8, D and E, open squares). Moreover, as already shown, methoctramine has no effect on release in M2-KO mice. The average control slope was 4.5 ± 0.27 and it was 4.4 ± 0.14 and 4.3 ± 0.1 after application of the "prolongation" and standard prepulses, respectively. Furthermore, neither methoctramine + AChE nor the standard prepulse had any effect on test-pulse release. Thus the average control slope was 4.5 ± 0.18, remaining the same in the presence of methoctramine +AChE (4.4 ± 0.17, P = 0.38) and after subsequent application of the standard prepulse (4.5 ± 0.1, P = 0.52).
In summary, the data displayed in Fig. 8 further support the notion that the strong and brief depolarizing prepulse indeed relieves a tonic block imposed by the M2R.
The strong and brief depolarizing prepulse does not affect twin-pulse facilitation in all three preparations
The following experiments attempt to elucidate the mechanism that underlies tonic inhibition of release. As mentioned before, one possibility that must be considered is that the tonic block is attributed to M2R-mediated voltage-dependent membrane-delimited inhibition of Ca2+ channels (Arnot et al. 2000; Jarvis and Zamponi 2001a, b; Jarvis et al. 2000; Zamponi and Snutch 1998). Another possibility is that the tonic block is achieved by a Ca2+-independent, but voltage-dependent mechanism (Slutsky et al. 1999, 2002). Such a mechanism might involve physical interaction of the M2R with SNARE proteins (Ilouz et al. 1999; Linial et al. 1997; Parnas et al. 2000).
An optimal way to distinguish between these 2 putative mechanisms would be by direct measurement of Ca2+ currents in the nerve terminal, at various pulse amplitudes, both before and after application of a depolarizing prepulse as done, such as in a transfected cell line (Arnot et al. 2000). This is currently impossible. The available techniques for focal measurement of Ca2+ currents (Brigant and Mallart 1982; Dudel 1990; Slutsky et al. 2001, 2002, 2003) permit measurement only of Ca2+ currents induced by action potentials, not by graded depolarization. We were thus forced to adopt a less direct but nevertheless sufficiently sensitive approach.
Because it is well accepted that twin-pulse facilitation (Katz and Miledi 1968) depends on residual calcium (Kamiya and Zucker 1994; Zucker 1999), we compared the effect of the standard prepulse applied before the first of twin pulses, on release evoked by the first and second of the twin pulses. Such experiments were performed in frog and in WT and M2-KO mice. The rationale underlying these experiments is as follows: If the increase in release during the first test pulse is attributed to increased entry of Ca2+, the residual Ca2+ encountered by the second test pulse should also be larger, and twin-pulse facilitation should be enhanced. If, in contrast, the depolarizing prepulse does not relieve a block of Ca2+ channels, but rather a block of the release machinery, and in addition, this effect is brief (4–5 ms), then the twin-pulse facilitation should not be affected by the prepulse. Such experiments can distinguish between these 2 possibilities only if the time course of the decline of prepulse-enhanced release differs significantly from the time course of facilitation.
Figure 9, A and B shows that this is indeed the case for both frog and WT mice. For clarity we denote the prepulse as pp; the quantal content of the first and second test pulses as m1 and m2, respectively; the quantal content of the first test pulse, when preceded by a prepulse, as m1p; and the quantal content of the second test pulse, when the first is preceded by a prepulse, as m2p. With the standard prepulse, the decline in prepulse-enhanced release of the first test pulse (m1p/m1), both in frog (Fig. 9A, open circles) and in WT mice (Fig. 9B, open circles), is rapid. The enhancing effect disappears, similarly to the effect on the slope (Fig. 6), after about 5 ms. The ratio m1p/m1 was 1.3 after 4 ms and 1.16 after 5 ms. In the M2-KO mice the standard prepulse, as expected, did not enhance release of the first pulse (open circles in C). Twin-pulse facilitation (m2/m1), in contrast, lasted about 20 ms in all 3 preparations (Fig. 9, A–C, filled squares). The results in Fig. 9 clearly show that even for these low test pulses (small entry of Ca2+), the time course of twin-pulse facilitation is much slower (by about an order of magnitude) than the time course of decline in prepulse-enhanced release.
We thus measured the effect of the standard prepulse, applied 1 ms before the first test pulse, on release produced by the second test pulse (the ratio m2p/m1), that is, on twin-pulse facilitation. It is seen that even though m1p/m1 was 16 in the frog (Fig. 9A, open circles), twin-pulse facilitation (m2p/m1) was not affected (Fig. 9A, open squares on the facilitation curve). A 16-fold increase of release of the first pulse, all other parameters being equal, requires that entry of Ca2+ during the pulse is at least doubled (m exhibiting a power of 4 relationship to Ca2+) (Dodge and Rahamimoff 1967). We expected that such an increase in Ca2+ influx, had it existed, would be detected in the facilitation experiments (see confirmation for this expectation below).
The same behavior was seen in WT mice (Fig. 9B). In this case, m1p/m1 was 8.2 (Fig. 9B, open circles) and, as in the frog, twin-pulse facilitation was not affected (Fig. 9B, open squares on the facilitation curve). In M2-KO mice, prepulse-mediated enhancement of release was, as before, completely lacking (open circles) and, naturally, the standard prepulse did not affect facilitation (open squares on the facilitation curve).
Together, these results indicate that, even though release in the first of the twin pulses increased in frog and WT mice when preceded by a prepulse, the increase was not produced by prepulse-induced higher Ca2+ influx during the first of the twin pulses. It may be argued that twin-pulse facilitation is not sufficiently sensitive to detect small changes in Ca2+ entry. To test for this possibility, we changed the parameters of the prepulse such that it admitted Ca2+ by itself. To this end, we lowered the amplitude of the prepulse and prolonged its duration. In selecting the final stimulation parameters of the new prepulse, we aimed at achieving comparable enhancement of release (m1p/m1) to that produced by the standard prepulse. With this new prepulse we repeated the experiment described before. For the frog, the release produced by the new prepulse (0.4 ms, –0.5 µA) plus the first test pulse was increased, where m1p/m1 was 13 (Fig. 9A, asterisk), slightly less than the increase produced by the standard prepulse (m1p/m1 = 16). However, in contrast to the case of the standard prepulse, facilitation was significantly increased (Fig. 9A, open squares with asterisks).
In WT mice, the longer and lower-amplitude new prepulse (0.3 ms, –0.4 µA) enhanced release produced by the first of the twin-test pulses 10-fold (Fig. 9B, asterisk), a little more than the standard prepulse. Here also, as in the frog, this prepulse increased twin-pulse facilitation (Fig. 9B, squares with asterisk). As mentioned earlier, both in frog and in WT mice we adjusted the parameters of the longer and lower-amplitude prepulse such that it produced enhancement of release similar to that produced by the standard prepulse. In the case of M2-KO mice, such a criterion could not be applied because the standard prepulse did not enhance release. Consequently, for the longer and lower-amplitude new prepulse we used the same parameters that were used for WT mice (0.3 ms, –0.4 µA). Figure 9C shows that, as for frog and WT mice, also in the M2-KO mice the test pulse preceded by the new prepulse produced release 17-fold higher than the test pulse alone (Fig. 9C, asterisk), and twin-pulse facilitation increased concomitantly (Fig. 9C, open squares with asterisks).
The data presented in Fig. 9 show that when either the prepulse admits Ca2+, or the prepulse increases Ca2+ influx during the first of the twin pulses, the increase in Ca2+ influx can be detected as residual Ca2+ by twin-pulse facilitation. Therefore both the lack of effect of the standard prepulse on twin-pulse facilitation, and the increase in facilitation produced by the lower-amplitude but longer new prepulse, support the conclusion that the standard prepulse alone does not admit Ca2+, nor does it increase Ca2+ influx during the first of the twin pulses. This conclusion is further strengthened by the following observation. In the frog, with the longer new prepulse, release was increased 13-fold (Fig. 9A), whereas the standard prepulse increased test-pulse release 16-fold. If the level of release is proportional only to the level of [Ca2+]i, these results mean that the broader prepulse, together with the test pulse, admitted less Ca2+ than the standard prepulse together with the test pulse. Yet, only the broader prepulse affected facilitation. Finally, the results of Fig. 9 further corroborate an earlier conclusion (see above and Slutsky et al. 2003) that the mechanisms of Ca2+ influx and removal, and thus of twin pulse-facilitation, are not altered in M2-KO mice. What is lacking is the M2R-imposed tonic block of release. Thus depolarization plays only one role in M2-KO mice, that of opening voltage-gated Ca2+ channels.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
The evidence in support of the above conclusions derives from experiments in which the quantal content (m) was measured at various presynaptic depolarizations [achieved by varying the depolarizing pulse amplitude (PA)], and establishing the slope of log m/log PA under various experimental conditions. The main evidence is as follows: 1) The slope of log m/log PA was 7–8 in frog and in WT mice, and only about 4 in M2-KO mice. 2) Addition of methoctramine, or AChE, treatments previously shown to abolish the M2R-imposed block (see references in INTRODUCTION), increased release in a voltage-dependent manner in both frog and WT mice. Enhancement became stronger as the depolarizing pulse decreased. In M2-KO mice, neither methoctramine nor AChE had any effect on release. As a corollary to the above, both methoctramine and AChE reduced the slope of log m/log PA in both frog and WT mice, but had no effect on the slope in M2-KO mice. 3) Administration of a brief (0.1-ms) but strong (–1.2 µA) depolarizing pulse before a test pulse enhanced release of the test pulse similarly to methoctramine. Thus it enhanced release in a voltage-dependent manner in frog and in WT mice but did not affect release in M2-KO mice. 4) A cocktail of methoctramine + AChE occluded the effect of the depolarizing prepulse. 5) The prepulse-enhanced release decayed in about 4–5 ms in frog and WT mice, at 8 and 30°C, respectively, whereas the duration of twin-pulse facilitation was >20 ms in all 3 preparations. 6) The strong and brief prepulse did not affect the amplitude of twin-pulse facilitation in any of the 3 preparations. In contrast, a prepulse lower in amplitude and of longer duration increased twin-pulse facilitation in all 3 preparations.
Relationship between the dependency of m on depolarization and on Ca2+ entry
The slopes of log m/log [Ca2+]o, or ICa, or [Ca2+]i have values of 3–5 (Augustine et al. 1985; Dodge and Rahamimoff 1967; Ravin et al. 1999). We show here that the slope of log m/log PA is about 7–8 in both frog and WT mice. We have interpreted the higher slope of log m/log PA, relative to log m/log [Ca2+], as meaning that for release to be initiated, depolarization must fulfill 2 roles: relief from a tonic block and opening of voltage-gated Ca2+ channels (Parnas et al. 2000). However, other explanations should be kept in mind. For example, the higher slope of log m/log PA might result from a nonlinear relationship between Ca2+ currents and depolarization. In particular, Simon and Llinas (1985) suggested that for identical ICa at 2 levels of depolarization, the Ca2+ concentration at the release site is expected to be higher at the higher depolarization. This is because the number of open channels increases with depolarization, and thus overlapping of Ca2+ diffusing from adjacent channels is expected. Also, other mechanisms for a supralinear relationship between Ca2+ influx and depolarization could exist. It should be noted that a nonlinear relationship between Ca2+ currents and depolarization cannot explain the lower slope in the M2-KO mice. We argue, however, that even if some nonlinearity of the types mentioned above does exist, it is the lower slope in the M2-KO mice and the reduction in the slope in frog and WT mice on addition of methoctramine or applying a brief depolarizing prepulse that constitute the key points. These findings provide evidence that, irrespective of the exact mechanism, depolarization does have an additional role (in initiation of release) over and above opening of Ca2+ channels.
Possible mechanism of the tonic block
The main thrust of this study had been to demonstrate that for initiation of ACh to take place, depolarization must relieve a tonic block imposed by the M2R. Our experiments were not aimed at directly studying the mechanism that underlies the tonic block. Nevertheless, we discuss 2 possible mechanisms. One possibility is that under resting conditions the voltage-gated Ca2+ channels are tonically inhibited by a mechanism that involves M2R-mediated voltage-dependent membrane-delimited inhibition (see, e.g., Jarvis and Zamponi 2001a, b; Jarvis et al. 2000; Zamponi and Snutch 1998). The second possibility is that the tonic inhibition of release is achieved by the M2R interacting with SNARE proteins at resting potential (Ilouz et al. 1999; Linial et al. 1997), thereby blocking the release machinery (Parnas et al. 2000). In both cases the block would be relieved by depolarization (see references above).
Although we cannot completely rule out the first suggestion, we consider this possibility less likely. This assertion is based on the following considerations. 1) Slutsky et al. (2002) showed that muscarine, at concentrations as high as 70 µM, does not reduce Ca2+ currents produced by the physiological action potential. It is thus hard to imagine that the low concentration of ACh (about 10 nM; Katz and Miledi 1977) present in the synaptic cleft at rest would be capable of maintaining the Ca2+ channels in a blocked configuration. 2) The brief, but strong, depolarizing prepulse-enhanced release evoked by the first of the twin-test pulses, but did not affect facilitation. In contrast, a longer but lower-amplitude prepulse enhanced release of the first test pulse, but concomitantly increased twin-pulse facilitation (Fig. 9). 3) The time constant observed here for inhibition (decline of prepulse enhanced test-pulse release) is of the order of a few milliseconds (4–5 ms, Fig. 9), whereas the time constant for inhibition of voltage-gated Ca2+ channels is significantly slower [e.g., about 50 ms for a depolarizing prepulse >20 mV and even slower for lower prepulse depolarizations (Arnot et al. 2000; Kasai 1992)]. 4) In the preceding studies, for the strongest depolarizing prepulse (>20 mV) to be able to relieve inhibition of the Ca2+ channels, its duration had to be about 4 ms. We find that even an extremely brief (0.1-ms) depolarizing prepulse substantially relieves the block.
We should note, however, that under physiological conditions the action potential is wider than 0.1 ms, and it is quite possible that the action potential also relieves, in addition to the M2R imposed but Ca2+-independent block, an M2R-imposed membrane-delimited inhibition of voltage-gated Ca2+ channels [but recall point 1) above].
It is interesting to note that the short time course of reinstatement of the M2R-imposed block (Fig. 9) is similar (but somewhat longer) to that of termination of action potential–evoked release. This is compatible with the suggestion that termination of ACh release occurs because on membrane repolarization the M2R rapidly shifts to its high-affinity state, rebinds ACh, and reimposes the tonic block (Slutsky et al. 2001, 2003). If so, then why is termination of release faster than the relaxation of disinhibition after a 0.1-ms prepulse observed here? One reason could be that because the prepulse did not induce release, the ACh concentration in the cleft encountered by the M2R remained very low (about 10 nM; Katz and Miledi 1977). Thus the first step in reinstatement of the block, after the prepulse (i.e., binding of ACh) is slower. In contrast, under physiological conditions, ACh concentration in the cleft is high because of release induced by the action potential and binding of ACh to the M2R is fast.
Termination of action potential induced release may be even faster than anticipated based on the considerations mentioned above. This is because elevation of transmitter concentration in the cleft, attributed to release induced by the action potential, may enable even the low-affinity M2R (prevailing during depolarization and shortly thereafter) to bind ACh and reinstate the block even sooner. Thus termination of release could begin on membrane repolarization, even before the M2R shifts eventually back to its high-affinity state. This interpretation is consistent with findings of Slutsky et al. (2002)
, n = 13) and M2R knockout (M2-KO) (
, n = 10) mice. B: log m/log PA slopes for the data presented in A (controls) and after addition of methoctramine. Slope of log m/log PA in WT mice was 7.47 ± 0.56 (black column), declining to 4.35 ± 0.36 (extremely significant, P < 0.0001) in the presence of 1 µM methoctramine (n = 12) (striped column). In M2-KO mice the control slope of 4.34 ± 0.31 (n = 10) (black column) remained unaltered after addition of methoctramine (4.3 ± 0.21, P = 0.23, n = 8) (striped column).
, on subsequent addition of 100 µM muscarine. Slope of log m/log PA was 7.2 ± 0.2 in the control, was reduced to 4.7 ± 0.3 (extremely significant, P < 0.0001) after addition of methoctramine, and increased to 6.9 ± 0.4 on subsequent addition of muscarine (an insignificant difference from the control, P = 0.54). B: average results of 3 experiments: 
, 4-ms interval; 
, 10 µM methoctramine + 80 µg/ml AChE reduced the log m/log PA slope to 5.34 ± 0.35 (extremely significant, P = 00007); 
, standard prepulse applied 1 ms before the test pulses did not further reduce the log m/log PA slope, which remained at 5.31 ± 0.31 (insignificant difference, P = 0.417).
, with 4-ms delay the slope already increased to 5.32 ± 0.33 (P = 0.24); 
, in the presence of methoctramine + AChE, slope = 4.54 ± 0.11 (extremely significant, P < 0.0001);