| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
MS 8415 Received 29 June 1998; accepted after revision 12 October 1998.
 |
ABSTRACT |
|---|
 TopAbstract IntroductionMethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Classical electrophysiological studies at the neuromuscular junction have demonstrated that Sr2+ can replace Ca2+ in activating neurotransmitter release (Miledi, 1966; Dodge et al. 1969; Meiri & Rahamimoff, 1971; Bain & Quastel, 1992). More recently, Sr2+-mediated neurotransmitter release has been demonstrated at excitatory (Goda & Stevens, 1994; Abdhul-Ghani et al. 1996; Oliet et al. 1996) as well as at inhibitory (Ohno-Shozaku et al. 1994; Morishita & Alger, 1997; Behrends & ten Bruggencate, 1998) synapses of the central nervous system (CNS). In all preparations studied, a characteristic distinction of Sr2+- versus Ca2+-mediated synaptic transmission is a reduced early (or phasic) component and an augmented late (or asynchronous) tail of individual miniature-like events called 'late release'. Thus, transmission in Sr2+ is desynchronized with respect to the normal situation.
Understanding this non-physiological phenomenon may entail important insights into the mechanisms which under physiological conditions control the kinetics of evoked transmitter release. Thus, Goda & Stevens (1994) have proposed that preferential activation by Sr2+ of late, asynchronous release is due to binding at a secondary, high-affinity divalent cation binding site, where the action of Sr2+ is stronger than that of Ca2+. A high-affinity site in addition to the putative low-affinity receptor involved in synchronous release (cf. Heidelberger et al. 1994) has also been postulated in order to account for use-dependent synaptic facilitation (Yamada & Zucker, 1992). The two-binding-site hypothesis has gained indirect support from experiments on excitatory synapses lacking the presynaptic Ca2+-binding protein synaptotagmin I, where phasic release was suppressed and asynchronous release was intact or increased (Geppert et al. 1994).
Alternatively, it has been suggested that a single release-inducing binding site may suffice to explain Sr2+-induced late release if buffering, extrusion or sequestration are assumed to be less efficient for Sr2+ than for Ca2+ so that free Sr2+ concentration remains elevated for longer times after a presynaptic impulse (Bain & Quastel, 1992; Goda & Stevens, 1994; Abdhul-Ghani et al. 1996; Behrends & ten Bruggencate, 1998).
Since the work of Goda & Stevens, reports have focused on the use of Sr2+ substitution as a tool to determine quantal size and its variation at CNS synapses (Abdhul-Ghani et al. 1996; Oliet et al. 1996; Morishita & Alger, 1997; Behrends & ten Bruggencate, 1998). This approach might become questionable, if, indeed, Sr2+-dependent asynchronous release was mediated through a particular reaction that may differ in many properties from normal synchronous transmission. Here, we specifically address the question of why Sr2+ desynchronizes synaptic transmission at inhibitory striatal synapses in cell culture. Using paired recordings we show that exogenous buffers selectively curtail late release, suggesting that it requires the prolonged presence of free Sr2+ and diffusion out of the immediate vicinity of the Ca2+ channel. Experiments where Sr2+-mediated asynchronous release and Ca2+-evoked phasic transmission occur simultaneously suggest that both processes are highly independent of each other. Our findings suggest that asynchrony of release is due to inefficient buffering and, consequently, occurs by diffusional spread and binding of Sr2+ to remote release activation sites that are not reached by the Ca2+ signal following a single presynaptic action potential.
Some of the data reported here have previously been published in abstract form (Rumpel & Behrends, 1997, 1998).
|
METHODS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Cells and solutions
Cell culture methods were those described by Gottmann et al. (1994). Briefly, following deep ether anaesthesia, pregnant Wistar rats were decapitated and the uterus dissected out and placed on an ice-cold glass dish. Embryos (embryonic day 17) were removed and their brains placed into ice-cold culture medium (Eagle's minimal essential medium). The lateral ganglionic eminences were dissected and subjected to mechanical dissociation. Cells were plated at low density in order to enhance the probability of synaptic connections between neighbouring neurones. Cultures were used for recordings from 12 to 22 days in vitro. For paired recordings of synaptically coupled neurones pipettes were filled with a solution containing (mM): KCl, 110; MgCl2, 5; EGTA, 0·6; Hepes, 10; Na-ATP, 2 (pH set to 7·3 with KOH, 230 mosmol l-1). When extracellular stimulation was applied, 5 mM 2(triethylamino)-N-(2,6-dimethylphenyl)acetamide (QX-314) was added to block Na+-dependent action currents and KCl was replaced by CsCl. The control bath solution was composed of (mM): NaCl, 125; KCl, 1; MgCl2, 1; Hepes, 20; glucose, 10; CaCl2, 2 (pH set to 7·35 with NaOH, 270 mosmol l-1). All bath solutions were prepared by adding glucose and CaCl2 (or SrCl2, as indicated) to a stock solution containing all other constituents. All constituents were purchased from Sigma Germany (Munich). In most experiments cells were continuously superfused with standard or test solution, employing a gravity-driven local application system (Y-tube). For rapid solution changes, a double-barrelled theta glass application pipette was used with each channel filled to the tip with a different saline. Electromagnetic valves controlled the gravity-driven flow of one solution at a time (cf. Taschenberger et al. 1995).
The acetoxymethyl ester form of BAPTA (BAPTA AM; Research Biochemicals International) and EGTA AM (Molecular Probes, Eugene, OR, USA) were dissolved in dimethyl sulfoxide (DMSO) and diluted in external solution to 50 µM. Final DMSO concentration was 0·5 %. DMSO at 0·5 % or 1 % had no effect on the inhibitory postsynaptic currents (IPSCs) in control solution (n = 2). Bovine serum albumin (0·1 % w/v) was added to solutions with BAPTA AM to improve solubility.
Electrophysiological recording
Patch-clamp experiments were performed at room temperature under direct visual control on a Zeiss IM 35 inverted microscope. Borosilicate pipettes with an outer tip diameter of 1·5-2 µm and an open tip resistance of 3-5 M
were used. In paired recordings IPSCs were evoked by short (3-5 ms) depolarizing step commands to the presynaptic neurone in voltage-clamp mode (holding potential near -70 mV) that elicited action currents. Alternatively, extracellular stimulation (duration 0·15-0·4 ms, amplitudes 2·5-5 V) was applied via a bipolar metal electrode. Stimuli were delivered every 5-10 s. Two EPC-7 patch-clamp amplifiers (HEKA-Electronics, Darmstadt, Germany) were used for recording without series resistance compensation (< 20 M). Output signals were filtered at 3 kHz and digitized on-line at 24 kHz with a National Instruments (Austin, TX, USA) NB-MIO 16L 14-bit A/D converter in a Macintosh 8100/100 computer.
Data analysis
The mean and standard deviation of the waveform of 5-20 IPSCs was calculated off-line using software written in LabView (National Instruments). The kinetics of averaged responses were first analysed using the built-in curve-fitting routine of IGORPro (Wavemetrics, Lake Oswego, OR, USA) to approximate a double-exponential function of the form:
y(t) = Afexp(-t/
f) + Asexp(-t/s)
to their decay phase, where f and s are the time constants of the fast and slow components (Af and As, respectively) of the IPSC waveform, respectively. A robust criterion for response synchrony was then obtained by approximating the integral of the responses from the relationship:
0 ydt = Aff + Ass
and dividing by the peak amplitude of the waveform. This peak-normalized integral has the dimension of time and may be thought of as a weighted global time constant. We introduce it here to measure the degree to which a waveform is dominated by a single peak. In the presence of such a dominant peak (i.e. with synchronous responses) the normalized integral is small; it gradually increases when the response becomes desynchronized. The advantage of this parameter is that it is sensitive not only to changes in the rate constants of processes underlying the time course studied but is strongly affected by changes in their relative amplitudes. It will faithfully reflect kinetic changes independently of whether the function used to fit the data is physically meaningful, as long as the fit is adequate and is more sensitive to small degrees of asynchrony than the half-width of the response.
Statistical analysis was done by Student's paired or unpaired t test, as applicable, using StatView software (Abacus Concepts, Berkeley, CA, USA). Mean values are given ± the standard error of the mean.
|
RESULTS |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Characteristics of Sr2+-mediated, desynchronized transmission
When the presynaptic neurone of a pair was stimulated in control extracellular solution (2 mM Ca2+), the resulting inhibitory postsynaptic current (IPSC) was dominated by a rapid, well-synchronized early component. Late release occurred at very low frequency, typically not above one or two events per response (Fig. 1A). Equimolar (2 mM) substitution of Sr2+ for Ca2+ resulted in a decrease in peak amplitude of the IPSC which was accompanied by a characteristic enhancement of late release, i.e. a barrage of asynchronous inhibitory postsynaptic currents (Fig. 1B) that continued for more than 1 s following the presynaptic impulse. These asynchronous IPSCs have been shown to be due to transmitter release at single synaptic contacts (Behrends & ten Bruggencate 1998). The increase in late release can also be assessed from the slowing of the rate of decay of averaged synaptic currents (Fig. 1C). Under control conditions (2 mM Ca2+), the decays of IPSCs were in general best fitted with two exponentials, a fast component (f) averaging 22·9 ± 2·01 ms and a slower one (s) of 85·3 ± 11·37 ms (n = 26 cell pairs). The decay phases of responses recorded in 2 mM Sr2+ were not always as clearly biexponential as in the case illustrated here (Fig. 1C). When biexponential fits to the data were done, however, both the early and fast components of decay were significantly slowed down with respect to Ca2+-mediated IPSCs: in 2 mM Sr2+, f and s averaged 53·9 ± 18·14 and 486·7 ± 167·70 ms, respectively (n = 7, P < 0·005).
The insets in Fig. 1A and B show the early phases of the IPSCs on a faster time scale. As previously observed by Morishita & Alger (1997), Sr2+ substitution appears to desynchronize the IPSC from its beginning. Indeed, an increase in time to peak of the averaged IPSC (see inset in Fig. 1C) was systematically observed with equimolar substitution of Sr2+ for Ca2+ (from 4·3 ± 0·3 ms to 9·8 ± 2·5 ms, P < 0·005).
Figure 1Da illustrates the effect of equimolar Sr2+ substitution on the total synaptic charge transferred during responses. As can be seen from the average time courses of current integrals, late release in the presence of Sr2+ more than compensates for the reduced peak amplitude of the responses (see figure legend for details). On average, total synaptic charge was 110 ± 41 % of control (P > 0·5, n = 7) after equimolar Sr2+ substitution. Thus, Sr2+-induced transmitter release is, on average, roughly equivalent to that provoked by Ca2+ influx, but it is less concentrated around the peak of the response. Figure 1Db illustrates a quantitative measure for this fact: dividing the integral waveforms by the peak amplitude of the response yields the normalized integral. Note the strong and significant increase in this parameter for the asynchronous response obtained in 2 mM Sr2+. On average, this parameter increased from 69·3 ± 2·54 ms in 2 mM Ca2+ to 208·1 ± 55·58 ms in 2 mM Sr2+.
 |
View larger version [in this window] [in a new window] |
|
|
A and B, superimposed traces show IPSCs (n = 5) recorded in response to a presynaptic action potential in the presence of 2 mM Ca2+(A) or 2 mM Sr2+(B). Insets: the same responses displayed at an expanded time scale. Note the massive desynchronization of the response in Sr2+. C, superimposition of the averages (n = 15) of IPSCs recorded in Ca2+ and Sr2+. Inset: averages shown scaled to peak and at an expanded time scale. Note the longer time to peak of the Sr2+ response. Da, time course of the integrals of the averaged IPSCs recorded in Ca2+ ( | ||
Lack of effect of rapid removal of extracellular Sr2+
The particular relevance of Sr2+-induced late release to the molecular physiology of synaptic transmission hinges on the assumption that late release cannot be simply explained by a prolonged time course of presynaptic influx of Sr2+ with respect to that of Ca2+. For instance, failure of Sr2+ to activate Ca2+-dependent K+ channels (Gorman & Hermann, 1979) might unmask a slow after-depolarization of the presynaptic action potential leading to additional divalent cation influx. Altered Ca2+ channel gating and inactivation (Hagiwara & Ohmori, 1982; Eckert & Chad, 1984) might produce long-lasting tail currents or Sr2+ might itself activate a persistent Sr2+-permeable conductance mechanism (Niggli, 1989). In addition, Sr2+ substitution may alter axonal excitability (Lüscher et al. 1994) to promote repetitive presynaptic discharges. None of these possibilities has, to our knowledge, been excluded previously. Here we argue that, if any mechanism resulting in prolonged influx of Sr2+ is responsible for late release, its time course should be affected by rapid removal of extracellular Sr2+. For this experiment, a single neurone was recorded in the whole-cell configuration, and synapses were stimulated using an extracellular stimulating electrode (see Methods). This simpler recording situation allowed us to position a double-barrelled theta pipette into the immediate vicinity of the cell in order to produce rapid changes of extracellular solution. Tests using the GABAA receptor antagonist bicuculline (10 µM) indicated that this set-up permits changes in the extracellular solution to be completed within 100 ms. This was shown by the block and recovery of individual Sr2+-evoked peak responses when the stimulus was moved in and out of a 250 ms time window during which the cell was superfused with solution containing the antagonist (Fig. 2Aa). As expected, when bicuculline-containing solution was rapidly applied for 250 ms during the late phase of Sr2+-dependent release, a reversible reduction in average synaptic current resulted. The difference trace computed between the average IPSCs with and without switch into the test solution clearly indicates the time course of bicuculline action (Fig. 2Ab).
 |
View larger version [in this window] [in a new window] |
|
|
A, validation of the rapid solution exchange using bicuculline. IPSCs were elicited by extracellular stimulation in 10 mM Sr2+. Aa, stimuli were repeatedly applied at different intervals with respect to the onset of a 250 ms rapid perfusion of the cell with bicuculline (10 µM)-containing solution (filled bar). Note the decrease of the peak amplitude of responses elicited in the time window of bicuculline perfusion. Ab, Sr2+-mediated IPSCs were elicited and averaged (n = 15) with or without the perfusion of bicuculline-containing solution for 250 ms starting 50 ms after the stimulus. Intervals of 5 s were allowed between stimuli. Averaged responses are shown superimposed and scaled to peak to enable comparison of kinetics. A clear deviation in time course due to the action of the antagonist becomes visible almost immediately after activation of the switch. The difference trace shown above the dashed line clearly illustrates the onset and offset of bicuculline action. Data in Aa and b are from the same neurone. B, rapid replacement of 10 mM Sr2+ by 10 mM Mg2+. Ba, protocol similar to Aa: the stimulus was repeatedly given at varying intervals with respect to the onset of rapid perfusion of Sr2+-free high-Mg2+ solution for 400 ms. Moving the stimulus into the Mg2+ perfusion time window (filled bar) leads to a progressive loss of synaptic current, indicating that Sr2+ is being washed out rapidly. Bb, protocol similar to Ab: superimposition of averaged, scaled (n = 15) Sr2+-mediated IPSCs elicited with or without a rapid switch from Sr2+- into Mg2+-containing solution 50 ms after the stimulus. Note the absence of any difference in the time course. Data in Ba and Bb are from the same neurone. | ||
When the rapidly perfused test solution contained no Sr2+ but 10 mM Mg2+ to minimize influx of Sr2+ through voltage-dependent Ca2+ channels, stimuli applied after the onset of the solution switch elicited responses that gradually became smaller when the delay between the onset of the solution change and stimulus became longer (Fig. 2B a). In this experiment, a complete block of the response was observed when the stimulus was applied 150 ms after perfusion of Sr2+-free, high-Mg2+ solution began, indicating that the concentration of Sr2+ present at this time was too small to sustain a cation influx sufficient to trigger release (Fig. 2Ba). In contrast, application of Sr2+-free test solution 50 ms after the stimulus, i.e. during the course of late release, did not produce any detectable change in the shape of the averaged IPSCs in four experiments (Fig. 2Bb). The same result was obtained when Ca2+ was applied at 10 mM in Sr2+-free solution (n = 5, data not shown).
Effects of exogenous chelators
The result of the experiment reported above corroborates the idea that the continued extracellular presence of Sr2+ is not required for late release to occur. The cause of the prominent desynchronization of release in the presence of Sr2+ must, therefore, lie in the particular manner of interaction of this cation either directly with sensors involved in triggering exocytosis or with sites involved in shaping the time course and spatial extent of the rise in cation concentration, such as mobile Ca2+ buffers (for review see Neher, 1998). In the latter scenario, the introduction into the intracellular space of an efficient exogenous chelator for Sr2+ should strongly affect late release. On the other hand, if the asynchrony of release was determined by the kinetics of unbinding from a putative high-affinity trigger site, no such effect would be expected.
BAPTA is known as a rapid, high-affinity chelator for Ca2+, and has been used to buffer Sr2+ (Bain & Quastel, 1992). However, information about its Sr2+-binding properties is not available. In order to obtain a rough estimate, we made use of the fact that BAPTA changes its UV absorption spectrum when bound by divalent cations (Tsien, 1980). Figure 3 shows a comparison of the dependence on total divalent cation concentration of the decrease in absorption at 255 nm with Ca2+, Sr2+ and Mg2+. The addition of Sr2+ reduced the UV absorption of BAPTA samples with an EC50 of about 50 µM (Ca2+: 42 µM). In contrast, Mg2+ was essentially ineffective. While this experiment does not permit any quantitative statement, the result clearly shows that Sr2+ binding by BAPTA is more similar to Ca2+ binding than to Mg2+ binding. Thus, BAPTA is a potential tool to elucidate whether Sr2+-mediated late release can be modified by altering intracellular Sr2+ buffering.
 |
View larger version [in this window] [in a new window] |
|
|
Increasing amounts of Ca2+ and Sr2+, respectively, were added to two aliquots of 100 µM BAPTA (filled symbols). In a control experiment, the same comparison was made for Sr2+ and Mg2+ (open symbols). Each data point is the mean of two or three repetitive absorption measurements at a wavelength of 255 nm. | ||
In the following experiments, BAPTA was used in its membrane-permeant acetoxymethyl ester form at 50 µM. In order to obtain larger postsynaptic currents with Sr2+-mediated transmission, Sr2+ was present at 5 mM. None of the kinetic parameters of the response was significantly affected by changes in Sr2+ concentration between 2 and 10 mM.
In 2 mM Ca2+, BAPTA AM application reduced the peak amplitude of the IPSC by 67·4 ± 7·8 % (n = 7) and in 5 mM Sr2+ peak responses were decreased by 64·1 ± 15·3 % (n = 7). The reduction in peak IPSC amplitude by BAPTA AM was thus not significantly different for both divalent cations (P > 0·1). However, BAPTA AM application strongly reduced the magnitude of late release during Sr2+-mediated transmission, while the time course of Ca2+-mediated synaptic responses was not significantly affected. Figure 4A shows a kinetic comparison of averaged IPSCs recorded in 2 mM Ca2+, 5 mM Sr2+ and in Ca2+ or Sr2+ solutions following a 4 min application of BAPTA AM. In this experiment, BAPTA AM had a negligible effect on the kinetics of IPSCs recorded in Ca2+ solution (Fig. 4Aa). In contrast, BAPTA AM strongly affected the shape of the Sr2+-dependent IPSC: in essence, the added intracellular buffer abolished the kinetic differences between Sr2+- and Ca2+-dependent transmission (Fig. 4Ab and c). As shown in Fig. 4B, this was associated with a return of the fast time constant (f) to control values. The slow time constant s, in contrast, was only weakly reduced by BAPTA AM application, despite the fact that synchrony of the response was essentially restored. Figure 4C displays a summary of numeric results from seven such experiments. Using the peak-normalized integral as a quantitative measure for IPSC synchrony (see Methods), we found that after BAPTA AM application there is no longer a significant difference between IPSCs evoked in 5 mM Sr2+ and those recorded in 2 mM Ca2+ (P > 0·05, Fig. 4Ca). The same applies to the fast time constant of decay (f) while s remains significantly elevated despite the return to normal of the overall shape of the responses (Fig. 4Cb and c). The time to peak of synaptic responses in 5 mM Sr2+ was also decreased by the action of BAPTA AM from 6·1 ± 0·61 ms to 5·0 ± 0·48 ms (P < 0·05), and was then no longer significantly different from the control value in Ca2+ (4·5 ± 0·57 ms, P > 0·1, data not shown), corroborating the notion that asynchronous release is active from the very beginning of the response and affects the time course of responses early on (Morishita & Alger, 1997).
 |
View larger version [in this window] [in a new window] |
|
|
A, superimposed peak-scaled traces show the differential effect of BAPTA AM on the kinetics of IPSCs evoked in 2 mM Ca2+ and in 5 mM Sr2+. The Ca2+-mediated response is kinetically unaffected (a). Sr2+ substitution induces the characteristic asynchrony of the response (b). Note, however, that hardly any difference is detectable between the response recorded in 2 mM Ca2+ and that obtained in 5 mM Sr2+ after application of BAPTA AM (c). B, semilogarithmic plot of the responses shown in A, illustrating the changes in time constants. Note the return to control of the fast time constant of the Sr2+-mediated response after application of BAPTA AM. C, summary of the results of 7 experiments (see text for details). All of 7 cell pairs were tested in all 4 conditions. | ||
The addition of exogenous buffer thus effectively suppresses Sr2+-dependent late asynchronous release. In fact, in the presence of intracellular BAPTA Sr2+ with respect to Ca2+ becomes a similarly short-acting, but less efficient, secretagogue. This finding is a strong indication that late release depends on the continued presence of unbound Sr2+ ions in the intracellular space of the synaptic terminals and that, therefore, diffusional spread of Sr2+ ions may be important in generating the late asynchronous phase of release.
Because of its rapid binding kinetics, BAPTA is thought to affect intracellular Ca2+ changes within a few hundred nanometres of the Ca2+ channel (Neher, 1998). It is, therefore, of interest to determine whether a molecule with slower binding kinetics is also effective in suppressing the late release phase induced by Sr2+. EGTA has been used as such a slow Ca2+ buffer (Adler et al. 1991; Borst & Sakmann, 1996). Because this compound does not undergo a shift in its UV absorption spectrum upon binding of divalent cations, we were unable to estimate its Sr2+-binding properties in the manner done for BAPTA. However, our experiments showed that EGTA AM at 50 µM, while somewhat less potent than BAPTA AM, was clearly effective in reducing late release. Peak IPSC amplitudes were reduced by 21 ± 3 % in 2 mM Ca2+ (n = 5) and by 33 ± 10 % in 5 mM Sr2+ (n = 5). As shown in Fig. 5, the compound was without a significant effect on the kinetics of Ca2+-dependent unitary IPSCs, while the late, asynchronous component of those evoked in Sr2+-containing solution was clearly reduced. However, in contrast to the result with BAPTA AM the peak-scaled integral of the responses on average remained significantly elevated above control (Fig. 5C a). This partial resynchronization of transmission was, as in the case of BAPTA AM, associated with a significant shortening of the first time constant, while the slower time constant was not significantly affected (Fig. 5B, and Cb and c).
 |
View larger version [in this window] [in a new window] |
|
|
A, the effect of EGTA AM is shown as in Fig. 4A for BAPTA AM. Note the reduced but persistent asynchrony after application of EGTA AM (c). B, as shown for BAPTA AM in Fig. 4B. Note: the fit for 5 mM Sr2+ ( | ||
Interaction of presynaptic Ca2+ influx with Sr2+-mediated release
The experiments reported above strongly suggest that Sr2+-evoked late release requires the prolonged presence of free Sr2+ inside the presynaptic terminal.
It is well known that asynchronous release can also be potentiated by repetitive presynaptic stimulation in Ca2+-containing solution (Cummings et al. 1996). The postulated high-affinity site for asynchronous release would, therefore, have to bind both Ca2+ and Sr2+ ions. If Sr2+ preferentially activates this rather than the 'classical' site, one would expect Sr2+ influx to specifically potentiate the late release component of a response to Ca2+ influx. Alternatively, other types of interaction between the two cations, such as partial agonism or competitive antagonism by Ca2+, might lead to a decrease in late release when both cations are present in the terminal. We therefore attempted to obtain evidence for such interactions.
In order to avoid spurious effects due to interactions at the Ca2+ channel, we attempted to establish a situation where Ca2+ and Sr2+ could enter the terminals as independently as possible. To this end, we again used the rapid perfusion apparatus and extracellular stimulation. Figure 6A shows two superimposed averaged IPSCs with and without a rapid change of the extracellular solution from one containing 10 mM Sr2+ to one with 10 mM Ca2+. These high concentrations were used to reduce the relative importance of possible residual ions of the other species following a solution change. The IPSC waveform obtained following the switch from Sr2+ into Ca2+ solution was compared with one recorded in the continuous presence of 10 mM Ca2+ in order to judge the quality of solution exchange. Of nine such experiments, six were found to have been conducted with adequate control of the extracellular medium.
Figure 6B shows the actual experiment performed to detect a possible interaction of Ca2+ and Sr2+ ions on a late release-inducing binding site. Paired stimuli were delivered at intervals between 100 and 300 ms and the evoked responses recorded under three different protocols: (1) both stimuli in Ca2+ solution, (2) first stimulus in Sr2+ solution, second stimulus in Ca2+ solution and (3) both stimuli in Sr2+ solution. For each case, responses were averaged and the waveform of the second IPSC was isolated by subtraction of the extrapolated double-exponential time course of the first IPSC. As shown in Fig. 6C, the isolated waveform of the second IPSC was identical if the second stimulus had been applied in Ca2+ solution, irrespective of whether the first stimulus had been given in Ca2+ or in Sr2+-containing medium. This result was obtained in all six experiments: the normalized integral of the second response for the Ca2+, Ca2+ sequence was, on average, 106 ± 9 % (P > 0·5) of that for the Sr2+, Ca2+ sequence. Thus, prior Sr2+ influx did not lead to enhanced facilitation of late release by a subsequent Ca2+ influx. Conversely, Ca2+ influx did not have a suppressive action on on-going Sr2+-dependent late release: the fact that the waveform of the second IPSC is identical in the two cases shows that the Sr2+-induced signal continues unchanged. In contrast, when both stimuli were delivered with Sr2+ present, a strong potentiation of late release resulted (Fig. 6Bc and C). Thus, these data indicate that late release by Sr2+ is not detectably influenced by a simultaneous activation of synchronous release by Ca2+, and vice versa.
 |
View larger version [in this window] [in a new window] |
|
|
A, set-up of the rapid perfusion device: IPSCs recorded either in Sr2+-containing solution or after a rapid switch into Ca2+ saline. Note the clear difference in kinetics. Ba-c, averaged IPSCs (n > 5) with conditioning and test pulse delivered in the solutions indicated. Dashed lines: extrapolated time courses of conditioning IPSCs. Insets: superimposed time course of first response (I) and second response (II) in each sequence. Note that facilitation of late release occurred in all cases except the Sr2+, Ca2+ sequence. C, superimposed, isolated test IPSCs (see text) resulting from the three sequences. Note the absence of any difference between the kinetics of test IPSCs of the Ca2+, Ca2+ and Sr2+, Ca2+ sequence. | ||
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
The most conspicuous changes in the postsynaptic response upon replacement of Ca2+ by Sr2+ are the decrease in peak amplitude and a prominent tail of small, asynchronous synaptic currents. It is highly probable that asynchronous IPSCs are due to discrete releases of transmitter at individual contact sites, as they are kinetically indistinguishable from miniature events recorded in normal extracellular medium from the same postsynaptic neurone (Behrends & ten Bruggencate, 1998).
Potentiation of late release by Sr2+ substitution increased both the fast and slow component of IPSC decay (cf. Goda & Stevens, 1994). This finding shows that a simple attribution of fast and slow time constants to early and late release is not feasible, the waveform of the IPSC being a convolution of release probability as a function of time with the time course of miniature events.
By showing that continued Sr2+ influx is not required for the asynchronous phase of release, our experiments using rapid solution changes clearly corroborate the assumption that its main cause must be sought in particular interactions of Sr2+ - as opposed to Ca2+ - with those constituents of the intraterminal space that govern the time course of transmitter release. It is nevertheless important to note that changes in the waveform of the presynaptic action potential by Sr2+ substitution are likely (Lüscher et al. 1994) and may affect the total amount of presynaptic charge flowing through Ca2+ channels. A prolonged influx of Sr2+ by a broadening of the presynaptic action potential to up to 100 ms would not be prevented by our solution exchange mechanism. The important result is, however, that the time course of late asynchronous release (from 100 to > 1000 ms post stimulus) was unaffected by the removal of Sr2+; therefore, the time course of late release is determined by the intracellular action or fate of Sr2+ and not by the time course of its flux across the membrane.
A secondary binding site for asynchronous release?
In the secondary-site model, the late, asynchronous component of transmitter release is due to the slow unbinding of the divalent cation from a high-affinity site. A suitable candidate is the Ca2+-binding protein synaptotagmin III which, in contrast to synaptotagmin I, has high Ca2+ affinity and functions in the presence of Sr2+ as well as Ca2+ (for review see Geppert & Südhof, 1998).
A more traditional explanation is the 'residual ion' model (Dodge et al. 1969; Bain & Quastel, 1992; van der Kloot & Molgó, 1993) in which the asynchronous phase is due to Sr2+ accumulating temporarily in the presynaptic terminal (for review see van der Kloot & Molgó, 1994). A mainstay of this hypothesis regarding late release mediated by Sr2+ influx was that, as shown here for central inhibitory synapses, BAPTA AM curtailed Sr2+-induced late release. We have now provided the missing independent evidence that at least BAPTA binds Sr2+ and have shown that it suppresses Sr2+-dependent late release at central inhibitory synapses. Regarding EGTA, its effectiveness may appear surprising in view of the fact that Sr2+ is known to bind to the unprotonated form of EGTA with about 270-fold lower affinity than Ca2+. On the other hand it has ca 35-fold higher affinity than Ca2+ for the singly protonated form of EGTA (Martell & Smith, 1974). It is likely, that a doubly protonated form will predominate at physiological pH, for which, however, there is no information on Sr2+ binding. The fact that EGTA has been used to buffer Sr2+ at physiological pH in the calibration of a fluorescent indicator, indo (Spencer & Berlin, 1997), indicates that this chelation is efficient.
The suppressant effects of both BAPTA AM and EGTA AM on Sr2+-induced late release do not favour any model where the time course of release is governed by the kinetics of interaction of Sr2+ with a release-inducing binding site. In contrast, apparently, late release depends on the prolonged presence of intracellular free divalent cation as in the 'residual ion' model. Without exogenous buffers, Sr2+ remains unbound for longer times after an influx compared with Ca2+; with exogenous buffers added, the difference is much reduced. Additionally, the selective sensitivity to exogenous intracellular buffers of the late phase of Sr2+-dependent postsynaptic responses suggests that late release requires diffusion of the divalent cation towards effector sites that are relatively distant from the source, i.e. outside the immediate vicinity of the Ca2+ channel. Because of this longer diffusion path, the ions triggering late release can more easily be intercepted by exogenous buffers than those that induce early synchronous release (cf. Neher, 1998). This scenario is also in line with our finding that prior activation of Sr2+ influx is not more efficient than Ca2+ influx at facilitating Ca2+-activated late release: assuming that paired-pulse facilitation of late release occurs by the superposition of the residual cation concentrations remaining from each pulse, then the spatial extent of that superposition in the intraterminal space is limited by the diffusional spread of the most efficiently buffered cation species, i.e. in this case, Ca2+. Because Ca2+ concentration is likely to be buffered to a very steep spatial profile, the Ca2+ ions entering during the second stimulus superimpose with residual Sr2+ only to the extent that they would superimpose with residual Ca2+; therefore, it does not make a difference to the facilitation of late release whether the conditioning stimulus evoked an influx of Sr2+ or of Ca2+.
In addition to being compatible with this buffered-diffusion variant of the residual cation hypothesis, as explained above, this result also constitutes evidence against the possibility that Sr2+ preferentially activates a secondary site which, rather than directly triggering release, mediates a facilitatory effect when activated (van der Kloot & Molgó, 1994). If Sr2+ were a more efficient agonist than Ca2+ at such a site, the asynchronous component of a subsequent Ca2+-dependent response would, contrary to our observations, be potentiated more by a prior influx of Sr2+ than of Ca2+.
The selective persistence of asynchronous release in synaptotagmin I knockouts can be explained without recourse to a secondary binding site: an alternative explanation has been proposed where the lack of synaptotagmin-mediated recruitment of Ca2+ channels to the fusion complex (cf. Mochida et al. 1996) would result in the exposure of the cation-binding trigger site to smaller, slower and more generalized increases in intracellular Ca2+, resulting in a desynchronization much like during the late phase of Sr2+-induced release (Neher & Penner, 1994).
Finally, this experiment also rules out the possibility that Sr2+, by some unknown mechanism, disrupts synchronous release, because synchronous responses could be elicited by Ca2+ influx even when Sr2+-mediated release was still on-going.
Differential buffering of Sr2+ and Ca2+
A requirement of our hypothesis is that endogenous Ca2+-binding proteins have a much lower affinity for Sr2+ than for Ca2+. Indeed, many of these, such as the calbindins, calretinin and parvalbumin share an EF-hand Ca2+-binding motif, which is known to have strong selectivity regarding ionic size, and to have an approximately 1000-fold lower affinity for Sr2+ than for Ca2+ (see Falke et al. 1994, for review). In contrast other Ca2+-handling systems, such as the Na+-Ca2+-exchanger and the reticular Ca2+ stores seem to accept Sr2+ quite well (Kimura et al. 1987; Spencer & Berlin, 1997).
The view emerging from the present results underlines the importance of diffusion and buffering mechanisms for the control of transmitter release kinetics. Because the intracellular concentrations and Sr2+-binding properties of BAPTA and EGTA are unknown, we are unable to draw quantitative conclusions such as those recently reviewed by Neher (1998). However, our findings are compatible with the idea that due to a low affinity of endogenous buffers for Sr2+, instead of the precisely shaped spatiotemporal domain structure of the Ca2+ signal (Simon & Llinàs, 1985; Naraghi & Neher, 1997), Sr2+ influx gives rise to a more generalized increase in divalent cation concentration. The difference in reaction rate between BAPTA and EGTA is due to the fact that protons have to dissociate from EGTA before Ca2+ can bind (Naraghi, 1997). Therefore, EGTA is likely to bind Sr2+ with a slow time constant as well. Thus, the fact that EGTA was efficient at all in reducing asynchronous release in our experiments is an indication that vesicles released during the late component must be relatively far away from the Ca2+ channel mouth (Neher, 1998).
In summary, our findings are not supportive of the hypothesis that a secondary binding site is responsible for asynchronous transmitter release. In contrast, we propose on the basis of our results that Sr2+-induced late release is due to reduced efficiencies of intracellular buffers for Sr2+ ions, resulting in a more spread-out spatiotemporal concentration increase.
|
REFERENCES |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
| Abdul-Ghani, M. A., Valiante, T. A. & Pennefather, P. S. (1996). Sr2+ and quantal events at excitatory synapses between mouse hippocampal neurons in culture. The Journal of Physiology 495, 113-125. | [Abstract] |
| Adler, E. M., Augustine, G. J., Duffy, S. N. & Charlton, M. P. (1991). Alien intracellular calcium chelators attenuate neurotransmitter release at the squid giant synapse. Journal of Neuroscience 11, 1496-1507. | [Abstract] |
| Bain, A. I. & Quastel, D. M. J. (1992). Quantal transmitter release mediated by strontium at the mouse motor nerve terminal. The Journal of Physiology 450, 63-87. | [Abstract] |
| Behrends, J. C. & ten Bruggencate, G. (1998). Changes in quantal size distributions upon experimental variations in the probability of release at striatal inhibitory synapses. Journal of Neurophysiology 79, 2999-3011. | [Abstract/Full Text] |
| Borst, J. G. G. & Sakmann, B. (1996). Calcium influx and transmitter release in a fast CNS synapse. Nature 383, 431-434. | [Medline] |
| Cummings, D. D., Wilcox, K. S. & Dichter, M. A. (1996). Calcium-dependent paired pulse facilitation of miniature EPSC frequency accompanies depression of EPSCs at hippocampal synapses in culture. Journal of Neuroscience 16, 5312-5323. | [Abstract/Full Text] |
| Dodge, F. A., Miledi, R. & Rahamimoff, R. (1969). Strontium and quantal release of transmitter at the neuromuscular junction. The Journal of Physiology 200, 267-283. | [Medline] |
| Eckert, R. & Chad, J. E. (1984). Inactivation of Ca channels. Progress in Biophysics and Molecular Biology 44, 215-267. | [Medline] |
| Falke, J. J., Drake, S. K., Hazard, A. L. & Peersen, O. B. (1994). Molecular tuning of ion binding to calcium signaling proteins. Quarterly Reviews of Biophysics 27, 219-290. | [Medline] |
| Geppert, M., Goda, Y., Hammer, R. E., Li, C., Rosahl, T. W., Stevens, C. F. & Südhof, T. C. (1994). Synaptotagmin I: a major Ca2+ sensor for transmitter release at a central synapse. Cell 79, 717-727. | [Medline] |
| Geppert, M. & Südhof, T. C. (1998). RAB3 and synaptotagmin: the Yin and Yang of membrane fusion. Annual Review of Neuroscience 21, 75-95. | [Abstract/Full Text] |
| Goda, Y. & Stevens, C. F. (1994). Two components of transmitter release at a central synapse. Proceedings of the National Academy of Sciences of the USA 91, 12942-12946. | [Medline] |
| Gorman, A. L. F. & Hermann, A. (1979). Internal effects of divalent cations on potassium permeability in molluscan neurones. The Journal of Physiology 296, 393-410. | [Abstract] |
| Gottmann, K., Pfrieger, F. W. & Lux, H. D. (1994). The formation of glutamatergic synapses in cultured central neurons: selective increase in miniature synaptic currents. Developmental Brain Research 81, 77-88. | [Medline] |
| Hagiwara, S. & Ohmori, H. (1982). Studies of calcium channels in rat clonal pituitary cells with patch electrode voltage clamp. The Journal of Physiology 331, 231-252. | [Medline] |
| Heidelberger, R., Heinemann, C., Neher, E. & Matthews, G. (1994). Calcium dependence of the rate of exocytosis in a synaptic terminal. Nature 371, 513-515. | [Medline] |
| Kimura, J., Miyame, S. & Noma, A. (1987). Identification of sodium-calcium exchange current in single ventricular cells of guinea-pig. The Journal of Physiology 384, 199-222. | [Abstract] |
| Lüscher, C., Streit, J., Lipp, P. & Lüscher, H.-R. (1994). Action potential propagation through embryonic dorsal root ganglion cells in culture. II. Decrease of conduction reliability during repetitive stimulation. Journal of Neurophysiology 72, 634-643. | [Medline] |
| Martell, A. E. & Smith, R. M. (1974). Critical Stability Constants, p. 269. Plenum Press, New York. | |
| Meiri, U. & Rahamimoff, R. (1971). Activation of transmitter release by strontium and calcium ions at the neuromuscular junction. The Journal of Physiology 215, 709-726. | [Medline] |
| Miledi, R. (1966). Strontium as a substitute for calcium in the process of transmitter release at the neuromuscular junction. Nature 212, 1233-1234. | |
| Mochida, S., Sheng, Z. H., Baker, C., Kobayashi, H. & Catterall, W. A. (1996). Inhibition of neurotransmission by peptides containing the synaptic protein interaction site of N-type Ca2+-channels. Neuron 17, 781-788. | [Medline] |
| Morishita, W. & Alger, B. E. (1997). Sr2+ supports depolarization-induced suppression of inhibition and provides new evidence for a presynaptic expression mechanism in rat hippocampal slices. The Journal of Physiology 505, 307-317. | [Abstract] |
| Naraghi, M. (1997). T-jump study of calcium binding kinetics of calcium chelators. Cell Calcium 22, 255-268. | [Medline] |
| Naraghi, M. & Neher, E. (1997). Linearized buffered Ca2+ diffusion in microdomains and its implications for calculation of [Ca2+] at the mouth of a calcium channel. Journal of Neuroscience 17, 6961-6973. | [Abstract/Full Text] |
| Neher, E. (1998). Vesicle pools and Ca2+ microdomains: new tools for understanding their roles in neurotransmitter release. Neuron 20, 389-399. | [Medline] |
| Neher, E. & Penner, R. (1994). Mice sans synaptotagmin. Nature 372, 316-317. | [Medline] |
| Niggli, E. (1989). Strontium-induced creep currents associated with tonic contractions in cardiac myocytes isolated from guinea-pigs. The Journal of Physiology 414, 549-568. | [Abstract] |
| Ohno-Shozaku, T., Sawada, S., Hirata, K. & Yamamoto, C. (1994). A comparison between potencies of external calcium, strontium and barium to support GABAergic transmission in rat cultured hippocampal neurons. Neuroscience Research 20, 223-229. | [Medline] |
| Oliet, S. H. R., Malenka, R. C. & Nicoll, R. A. (1996). Bidirectional control of quantal size by synaptic activity in the hippocampus. Science 271, 1294-1297. | [Abstract] |
| Rumpel, E. & Behrends, J. C. (1997). Is late/asynchronous transmitter release mediated through a distinct pathway? Investigation on Ca2+ and Sr2+-mediated synaptic responses. Society for Neuroscience Abstracts 23, 1172. | |
| Rumpel, E. & Behrends, J. C. (1998). Ca2+-chelators accelerate the decay of Sr2+-mediated late release at striatal synapses. Pflügers Archiv 435 (suppl.), R138. | |
| Simon, S. M. & Llinàs, R. R. (1985). Compartmentalization of the submembrane calcium activity during calcium influx and its significance in transmitter release. Biophysical Journal 48, 485-498. | [Abstract] |
| Spencer, C. I. & Berlin J. R. (1997). Calcium-induced release of strontium ions from the sarcoplasmic reticulum of rat cardiac ventricular myocytes. The Journal of Physiology 504, 565-578. | [Abstract] |
| Taschenberger, H., Engert, F. & Grantyn, R. (1995). Synaptic current kinetics in a solely AMPA-receptor-operated glutamatergic synapse formed by rat retinal ganglion neurons. Journal of Neurophysiology 74, 1123-1136. | [Medline] |
| Tsien, R. Y. (1980). New calcium indicators and buffers with high selectivity against magnesium and protons: design, synthesis, and properties of prototype structures. Biochemistry 19, 2396-2404. | [Medline] |
| van der Kloot, W. & Molgó, J. (1993). Facilitation and delayed release at about 0°C at the frog neuromuscular junction: effects of calcium chelators, calcium transport inhibitors, and okadaic acid. Journal of Neurophysiology 69, 717-729. | [Medline] |
| van der Kloot, W. & Molgó, J. (1994). Quantal acetylcholine release at the vertebrate neuromuscular junction. Physiological Reviews 74, 899-990. | [Medline] |
| Yamada, W. M. & Zucker, R. S. (1992). Time course of transmitter release calculated from simulations of a calcium diffusion model. Biophysical Journal 61, 671-682 | [Abstract] |
This research was performed in the laboratory of Professor Gerrit ten Bruggencate, whom we would like to thank for his support and encouragement. We are grateful to Luise Kargl and Anke Grünewald for expert technical assistance. This work was supported by the Deutsche Forschungsgemeinschaft (Grant Be-1739/1 and /2).
Corresponding author
J. C. Behrends: Physiologisches Institut, Universität München, Pettenkoferstrasse 12, 80336 München, Germany.
Email: j.behrends{at}lrz.uni-muenchen.de
This article has been cited by other articles:
|
|
 |
|
  N. Zhang, W. Wei, I. Mody, and C. R. Houser Altered Localization of GABAA Receptor Subunits on Dentate Granule Cell Dendrites Influences Tonic and Phasic Inhibition in a Mouse Model of Epilepsy J. Neurosci., July 11, 2007; 27(28): 7520 - 7531. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Saitow, H. Suzuki, and S. Konishi {beta}-Adrenoceptor-mediated long-term up-regulation of the release machinery at rat cerebellar GABAergic synapses J. Physiol., June 1, 2005; 565(2): 487 - 502. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 O. Wiser, R. Cohen, and D. Atlas Ionic dependence of Ca2+ channel modulation by syntaxin 1A PNAS, March 19, 2002; 99(6): 3968 - 3973. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. Rumpel and J. C. Behrends Postsynaptic Receptor Occupancy During Evoked Transmission at Striatal GABAergic Synapses In Vitro J Neurophysiol, August 1, 2000; 84(2): 771 - 779. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Xu-Friedman and W. G. Regehr Probing Fundamental Aspects of Synaptic Transmission with Strontium J. Neurosci., June 15, 2000; 20(12): 4414 - 4422. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
O. Caillard, Y. Ben-Ari, and J.-L. Gaiarsa Mechanisms of Induction and Expression of Long-Term Depression at GABAergic Synapses in the Neonatal Rat Hippocampus J. Neurosci., September 1, 1999; 19(17): 7568 - 7577. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
) and Sr2+ (
). Error bars: standard error of the mean. In this example, the mean total synaptic charge flowing in Sr2+ exceeds that transferred during the Ca2+-dependent response. Note however, that due to the large variance of integrals the difference between the two final values is not statistically significant (11·62 ± 1·76 pC vs. 17·30 ± 3·79 pC, P > 0·05). Db, time courses of the peak-normalized integrals, obtained by dividing the integrals in a by the peak amplitude of the response (see Methods). 
, Ca2+; 
, Sr2+. The difference between the final values is highly significant (51·16 ± 0·77 ms vs. 182·14 ± 39·95 ms, P < 0·005).