| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Department of Physiology and Biophysics, University of Miami School of Medicine, Miami, Florida 33136
Submitted 8 January 2003; accepted in final form 24 March 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
3.4-fold, and this increase was even greater (transiently) after
mitochondrial depolarization. In both 2 and 0.5 mM
[Ca2+], mitochondrial depolarization increased
asynchronous release during the 50-Hz train and increased the total vesicular
release (phasic and asynchronous) measured by destaining of the styryl dye
FM2-10. These results suggest that by limiting the stimulation-induced
increase in cytosolic [Ca2+], mitochondrial
Ca2+ uptake maintains a high ratio of phasic to
asynchronous release, thus helping to sustain neuromuscular transmission
during repetitive stimulation. Interestingly, the quantal content of the epp
reached during 50-Hz stimulation stabilized at a similar level (20
quanta) in both 2 and 0.5 mM Ca2+. A similar convergence
was measured in oligomycin, which inhibits mitochondrial ATP synthesis without
depolarizing mitochondria, but quantal contents fell to <20 when
mitochondria were depolarized in 2 mM Ca2+. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
;
Colegrove et al.
2000a,b;
Friel and Tsien 1994;
Thayer and Miller 1990;
Werth and Thayer 1994).
Mitochondria take up Ca2+ passively via a uniporter,
down the electrochemical gradient across the inner mitochondrial membrane.
Most of this gradient is attributable to the membrane potential gradient
m (–150 to –200 mV), created by H+
extrusion via the complexes of the electron transport chain (reviewed in
Gunter and Gunter 1994;
Gunter and Pfeiffer 1990). In
some cells, this mitochondrial Ca2+ uptake begins at
≤300 nM cytosolic [Ca2+]
(Babcock et al. 1997;
Colegrove et al. 2000a;
David et al. 1998).
Mitochondrial Ca2+ uptake appears to be especially
important in nerve terminals, which contain abundant mitochondria and
experience a large Ca2+ influx during activity. During
repetitive action potential discharge, cytosolic [Ca2+]
in frog, crayfish, lizard, and mouse motor terminals increases at first
rapidly, then more slowly (David et al.
1998; David and Barrett
2000; Suzuki et al.
2000,
2002;
Tang and Zucker 1997;
Wu and Betz 1996). These
studies showed that when mitochondrial Ca2+ uptake is
blocked, cytosolic [Ca2+] continues to increase rapidly
throughout the stimulus train, reaching much higher final values. Steunkel
(1994) demonstrated similar results in rat neurohypophysis depolarized using
voltage-clamp pulses.
The present study investigated how inhibition of mitochondrial
Ca2+ uptake affects transmitter release from lizard
motor terminals. Previous studies showed that m depolarizing
agents increase the asynchronous release recorded in physiological saline,
Ca2+-free saline, and media containing high
[K+] (Alnaes and Rahamimoff
1975; Calupca et al.
2001; Molgo and
Pecot-Dechavassine 1988;
Washio 1982). In low bath
[Ca2+], m -depolarizing agents increased
the rate at which end-plate potential (epp) amplitudes increased during 20-Hz
stimulation of frog motor terminals
(Zengel et al. 1994). In
contrast, in physiological [Ca2+] m
depolarization accelerated the depression of epp amplitudes in mouse motor
terminals (David and Barrett
2003). We measured total vesicular release using destaining of
FM2-10, and phasic and asynchronous quantal release using electrophysiological
techniques, in lizard motor terminals in both low (0.5 mM) and physiological
(2 mM) bath [Ca2+]. We report that m
depolarization increased total vesicular release and asynchronous quantal
release during 50-Hz stimulation in both bath [Ca2+].
However, the effects of m depolarization on phasic release
during repetitive stimulation were markedly different, with a transient
increase in 0.5 mM Ca2+ but an accelerated depression in
physiological [Ca2+]. An unexpected finding was that epp
quantal contents in 0.5 and 2 mM Ca2+ converged to a
common value during repetitive stimulation.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Neuromuscular preparations from external intercostal muscles were dissected from the lizard Anolis sagrei. Animals were decapitated and pithed after euthanasia in 100% CO2. The preparation was mounted in a silicon chamber constructed on a No. 1 glass coverslip and bathed in normal lizard saline (NLS, pH = 7.4) composed of (mM) 157 NaCl, 4 KCl, 2 CaCl2, 2 MgCl2, 5.5 glucose, and 1 HEPES buffer. In some experiments, bath [Ca2+] was reduced to 0.5 mM. Multiple muscles (4–6), each innervated by its own motor nerve, were accessible in each preparation. All experiments were performed at room temperature.
Action potentials were evoked in the motor nerve via a suction electrode by
applying brief (0.3 ms) suprathreshold depolarizing pulses. Nerve stimulation
was applied at 1 Hz throughout the experiment, with trains of 50 Hz for 10 s
(or sometimes 50 s) superimposed. Intervals between stimulus trains were
≥10 min to ensure complete recovery of mitochondrial
[Ca2+] (David
1999). At the beginning of each experiment, the viability of the
preparation was confirmed by observing muscle contractions. During imaging
experiments, contractions were blocked by 15 µM tubocurarine. Techniques
used to block or minimize contractions during electrophysiological recordings
are described in a later section.
m was depolarized using carbonyl cyanide
m-chlorophenyl hydrazone (CCCP, 1 µM), a protonophore that
dissipates the proton electrochemical gradient produced by H+
extrusion via complexes I, III, and IV of the mitochondrial electron transport
chain, or by antimycin A1 (2 µM), which inhibits complex III (reviewed in
Gunter and Pfeiffer 1990).
Responses to these two m depolarizing agents were similar
except that the effects of brief exposure to CCCP were partially reversible,
whereas the effects of antimycin were not. Depolarizing m
reduces not only the electrical gradient that permits mitochondrial
Ca2+ uptake, but also the proton motive force that
permits mitochondrial ATP synthesis via complex V
(F1,F0-ATP synthetase) of the electron transport chain.
In addition, when m is depolarized,
F1,F0 ATP synthetase can operate in reverse mode,
hydrolyzing ATP and maintaining partial m polarization. To
prevent this extra ATP hydrolysis and to ensure complete m
depolarization, we inhibited complex V by adding oligomycin (5–10
µg/ml) to all solutions containing CCCP or antimycin A1. In addition,
solutions containing oligomycin alone were tested to measure the effect of
inhibiting mitochondrial ATP production without depolarizing
m. In most experiments, each preparation was monitored in
control solution (2 or 0.5 mM Ca2+) followed by
application of oligomycin alone or of oligomycin plus a m
depolarizing agent. Experimental measurements were made between 20 and 120 min
after drug application. We did not test the effects of Ru360, an inhibitor of
the mitochondrial Ca2+ uniporter, because in lizard
motor terminals Ru360 also appears to inhibit Ca2+
influx across the plasma membrane (David
1999).
Imaging changes in cytosolic and mitochondrial [Ca2+]
Measurements of stimulation-induced changes in cytosolic
[Ca2+] were made using the Ca2+
indicator Oregon Green BAPTA 5N (OG-5N), injected ionophoretically via a
microelectrode inserted into the motor axon
(David et al. 1997). OG-5N was
excited at 488 nm and emitted light collected at >515 nm. The low affinity
of OG-5N (Kd 40–60 µM) ensured minimal
disruption of cellular Ca2+-dependent processes and
prevented dye saturation at the higher cytosolic [Ca2+]
attained during repetitive stimulation with m
depolarization.
Rhod dyes were used to monitor changes in mitochondrial
[Ca2+]. Rhod-2 (Kd 0.5 µM)
or rhod-5F (Kd 1.9 µM) were targeted into the
mitochondria by bath-loading 50 µg/ml of the acetoxymethylester (AM) form
for ≤2 min, followed by washout via 5–10 bath exchanges of NLS for 1
h. Rhod dyes were excited at 528 nm, and emission was monitored at >570 nm.
Mitochondrial localization was confirmed using the following morphological,
pharmacological, and kinetic criteria. First, the fluorescence increase
following trains of stimuli was punctate and localized within the nerve
terminal (not the axon), consistent with the known clustering of mitochondria
within motor terminals. Second, the stimulation-induced increase in
fluorescence was abolished by m depolarizing agents
(Fig. 1A). Third, the
rate of rise of mitochondrial [Ca2+] was slower than
that of cytosolic [Ca2+], and when stimulation ceased,
mitochondrial [Ca2+] decayed slowly, without the rapid
initial component characteristic of cytosolic [Ca2+]
(David 1999;
David et al. 1998;
Jou et al. 1996).
|
Terminals were imaged with a confocal laser-scanning microscope (Noran
Odyssey, Noran Instruments, Middleton WI) using a Nikon x40
water-immersion lens or a Nikon x20 air lens. Laser power was kept low
to minimize photodamage. Data were collected using an Indy workstation
(Silicon Graphics) with Noran InterVision software. Images were sampled at a
rate of 0.533 or 1.066 s/frame and analyzed using V++ software
(version 4.0, Digital Micro Optics, Auckland NZ). Regions of interest (ROIs)
were drawn around the fluorescent terminals and the same ROIs were used across
all images collected for a given terminal. Background staining intensity was
averaged from ROIs drawn outside of and adjacent to the boutons. Fluorescence
was reported as 
F/F, calculated as
 | (1) |
).
Briefly, the classic equation of Grynkiewicz et al.
(1985)
 |
F/Frest ratios
 | (2) |
= Frest/Fmin =
(Kd +
([Ca2+]rest)(Fmax/Fmin))/(Kd
+ [Ca2+]rest). For a low affinity indicator
like OG-5N (Kd approximately 50 µM), the fluorescence
at resting [Ca2+] (assumed to be 100 nM) should be close
to Fmin, so the value of should be only slightly
>1. The Fmax/Fmin ratio measured
for OG-5N is 50 (W.G.L. Kerrick, personal communication). Imaging vesicular release
Vesicular release was monitored using a styryl dye FM2-10 (50 µM)
(Betz et al. 1996;
Richards et al. 2000), which
was loaded into presynaptic vesicles by incubating in 60 mM KCl for 15 min and
then washing with NLS for 30–60 min. Richards et al.
(2000) showed that in frog
motor terminals FM2-10 is internalized with a half-time of 8 min. Thus
the 15-min incubation used to load FM2-10 would be expected to achieve
internalization of ≥86% of the FM2-10 taken up, consistent with our finding
that incubation times >15 min did not result in greater dye retention. Our
protocol thus corresponds to their "delayed wash" conditions, in
which FM2-10 would be expected to label all vesicles. Terminal destaining was
measured during trains of 2,500 stimuli at 50 Hz. Trains were repeated at
15-min intervals until no further decrease in fluorescence was detected, in
most cases after 12,500 stimuli (5 trains). Images were recorded every 1.066
s. Values reported were for the first train. Calculation of percent dye loss
began by subtracting background fluorescence to yield net fluorescence
intensity. Pixel brightness, measured in arbitrary pixel intensity units
(PIU), was then converted into percent dye loss, using the following equation
 | (3) |
Electrophysiological measurements
For these recordings, we needed to block muscle contractions in a manner
that permitted recording of both phasic and asynchronous release. The long
duration of the recordings (multiple 50-Hz trains separated by ≥10-min
intervals) precluded use of voltage clamping or focal extracellular recording.
A toxin (µ-conotoxin GIIIB) that blocks muscle but not nerve
Na+ channels in frogs and mice does not have this selective effect
in lizard preparations. Thus we used two other approaches to reduce the
detrimental effect of muscle contractions on intracellular recordings. First
we selected fibers in which electrode damage had partially depolarized the
resting potential to –40 to –50 mV for recordings in 2 mM
Ca2+ or to –60 to –70 mV for recordings in
0.5 mM Ca2+ (see
Table 1). This depolarization
increased the threshold for action potential generation and thus minimized
contractions. Second, we used a floating microelectrode
(Woodbury and Brady 1956),
manufactured by wrapping the middle cm of a 3-cm silver wire electrode around
a 0.2-mm-diam rod to create a small spring. One end of the silver wire
was pushed into the barrel of a microelectrode and secured with a small piece
of modeling clay. The other end was mounted on the micromanipulator.
Microelectrodes (8–20 M
) were pulled from borosilicate glass and
filled with 3 M KCl. Data were collected using an Axoclamp 2A preamplifier
(Axon Instruments, Union City CA), using Axoclamp 8.1 software.
|
Epp amplitudes were measured using Clampfit software (Axon Instruments).
These amplitudes were then corrected using an equation that combines Martin's
(1976, Eq. 6)
correction for nonlinear summation, and Kelly's
(1978, Eq. 1)
correction to a common resting potential, to facilitate comparison of epp
amplitudes recorded from fibers with different resting potentials.Martin's
(1976) formula can be written
as
 |
) correction factor is
(Vc –
Vr)/(Vo –
Vr), where Vc is the selected common
resting potential (see following text). Combining Martin's formula with
Kelly's correction yields
 | (4) |
Mepps were detected and analyzed using Mini Analysis Program (version 5.3.5, Synaptosoft, Jaejin Software, Leonia NJ) with visual inspection of recordings to exclude noise and stimulus artifacts. To estimate mepp frequency during the interstimulus interval, we counted mepps during the last 10 ms of the 20-ms interstimulus interval and multiplied by 2 to estimate the total number of mepps in the interstimulus interval. However, the decay of large epps sometimes extended over most of the interstimulus interval, resulting in underestimation of the frequencies of superimposed mepps. Mepp amplitude was averaged from 20 to 60 mepps measured in the resting fiber and corrected using Eq. 4.
Quantal content was calculated by dividing the corrected epp amplitude by the corrected mepp amplitude. Quantal contents (Figs. 5 and 6) were plotted using both geometric and arithmetic averaging but since both analyses yielded similar patterns, only arithmetic averages are shown.
|
|
Statistical analysis
Average values are reported as mean ± SE. Statistical significance was assessed using Student's t-test, ANOVA, and Student-Newman-Keuls (SNK) or Tukey's multiple comparison test using InStat/Prism software. Data were plotted using GraphPad Prism version 4.10 for Windows software (GraphPad Software, San Diego CA).
Reagents
Fluorescent indicator dyes were purchased from Molecular Probes (Eugene, OR). Other reagents were from Sigma.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Figure 1A shows
that the increase in rhod-5F fluorescence produced by 50-Hz stimulation in
physiological saline was blocked by m depolarization with
CCCP, verifying the mitochondrial localization of AM-loaded rhod dyes.
Figure 1B compares
rhod-2 F/F responses in 2 and 0.5 mM
Ca2+. In physiological [Ca2+],
mitochondrial [Ca2+] rose above resting levels within
50–100 stimuli, and by 250–300 stimuli reached a plateau level
that was maintained for the duration of the train, followed by a slow
posttrain decay. Decreasing [Ca2+] to 0.5 mM resulted in
a slower rate of rise in mitochondrial [Ca2+], though
the final plateau value was unchanged. This plateau is not due to dye
saturation because similar plateaus are measured in mitochondria loaded with
lower-affinity dyes (David et al.
2003). David
(1999) estimated that at this
plateau level the free [Ca2+] in the mitochondrial
matrix is 1 µM.
Figure 2 demonstrates the
effects of bath [Ca2+] and m
depolarization on stimulation-induced increases in cytosolic
[Ca2+], calculated from fluorescence changes of
ionophoretically injected OG-5N, as described in METHODS.
Figure 2A shows that
cytosolic [Ca2+] increased rapidly within the first
1–2 s after the onset of stimulation, followed by a slower rate of rise
for the duration of the train as reported by David et al.
(1998). When stimulation
ceased, cytosolic [Ca2+] showed a rapid initial decay,
followed by a slower decay to baseline. By the end of the 50-Hz, 10-s train,
the increase in cytosolic [Ca2+] above the assumed
resting level of 0.1 µM was 0.65 ± 0.02 µM in 2 mM
Ca2+, significantly greater than that in 0.5 mM
Ca2+ (0.48 ± 0.02 µM, P < 0.01).
Stimulation after m depolarization with antimycin A1 resulted
in larger increases in cytosolic [Ca2+] in both 2 and
0.5 mM [Ca2+] with the magnitude of the response
increasing over time (Fig. 2, B
and C). Stimulation-induced cytosolic
[Ca2+] responses increased more rapidly and reached
higher peak values in 2 than in 0.5 mM Ca2+. The
measurements of vesicular and quantal release during m
depolarization reported below were made between 20 and 120 min after
m depolarization and thus would have been accompanied by
increased cytosolic [Ca2+] responses like those in
Fig. 2, B and
C.
|
The mitochondrial and cytosolic F/F responses in
Figs. 1 and
2 demonstrate that mitochondria
contribute to handling the Ca2+ loads associated with
repetitive stimulation in low, as well as in physiological, bath
[Ca2+]. In these terminals, mitochondria contribute more
to Ca2+ sequestration than endoplasmic reticulum (ER),
because cyclopiazonic acid, which blocks the ER Ca-ATPase, has no significant
effect on cytosolic or mitochondrial F/F responses to
50-Hz stimulation (David
1999).
Total vesicular release during 50-Hz stimulation increases after
m depolarization
Figure 3 shows the effect of
m depolarization on total vesicular release (phasic plus
asynchronous) during 50-Hz stimulation. The styryl dye FM2-10 was loaded into
synaptic vesicles in 2 mM Ca2+, and preparations were
then stimulated for 50 s (2,500 stimuli) in the indicated solutions.
Destaining was measured and converted into % dye loss, as described in
METHODS. In 2 mM Ca2+
(Fig. 3A) stimulation
produced a 42 ± 4% dye loss in both control solution and oligomycin
alone. After m depolarization, total dye loss rose to 55
± 4% by the end of the train, reflecting a significant increase in
total vesicular release (P < 0.01, Tukey's multiple comparison
test). This increase in release was also significant after 10 s (500 stimuli,
as used in Figs. 2 and
4,
5,
6,
7,
8,
9,
10; 19 ± 3% in CCCP vs.
11 ± 2% in control, P < 0.05).
|
|
|
|
|
|
This effect of m depolarization on vesicular release was
also seen in 0.5 mM Ca2+.
Figure 3B shows that
stimulation under control conditions produced a 16 ± 3% dye loss,
40% of that measured in 2 mM Ca2+. Dye loss was
similar in oligomycin alone but after m depolarization, the
percent dye loss increased to 31 ± 4%.
m depolarization accelerates epp
depression in 2 mM [Ca2+], but transiently increases epp
enhancement in 0.5 mM Ca2+
Electrophysiological recordings were made to determine whether the
m depolarization-induced increase in vesicular release during
50-Hz stimulation was due mainly to phasic or asynchronous quantal release.
Figure 4 shows representative
traces of the first and last five epps evoked by a train of 500 stimuli. In 2
mM Ca2+, epp amplitude depressed, and this depression
was increased during m depolarization
(Fig. 4A). In
contrast, in 0.5 mM Ca2+, epp amplitudes increased both
before and after m depolarization
(Fig. 4B). In both
cases, there was evidence for increased asynchronous release at the end of the
train during m depolarization (see following text).
Figure 5 plots the averaged
time course of epp depression during 50-Hz stimulation in 2 mM
Ca2+. Epp amplitudes were corrected for nonlinear
summation, and quantal contents calculated as described in METHODS.
Figure 5A plots
averaged quantal contents measured in control, oligomycin alone, and
oligomycin plus m depolarization. Final quantal contents in
control and oligomycin alone were similar, but the final quantal content was
significantly smaller during m depolarization.
Treatments that increase depression of phasic release during tetanic
stimulation often increase the pretrain (baseline) quantal content.
Table 1 indicates how these
drug treatments affected phasic and asynchronous release from terminals at
rest and during low-frequency stimulation (1 Hz). The low muscle resting
membrane potentials were the result of having to depolarize the muscle fiber
to minimize action potentials and subsequent contractions (see
METHODS). Reducing bath [Ca2+] from 2 to 0.5
mM reduced the average epp quantal content by 86%, from 50 to
7. However, for a given bath [Ca2+], the applied
drugs did not significantly alter resting membrane potentials or quantal
content. Thus the increased depression during m depolarization
in 2 mM Ca2+ was not associated with a higher initial
quantal content.
To examine the time course with which depression of phasic release
developed, Fig. 5B
plots quantal contents normalized to the pretrain quantal content, comparing
control with oligomycin alone (left) and oligomycin alone with oligomycin plus
m depolarization (right). In all conditions there was an
initial increase in phasic release during the first 4 stimuli, followed by
depression. At early times (<1 s) depression was greater in oligomycin
alone than in control or m -depolarized solutions, and in this
normalized plot the final depression in oligomycin (67 ± 7%) exceeded
that measured in control (53 ± 6%), though this difference failed to
reach significance. By the end of the train the depression during
m depolarization (80 ± 7%) was greater than that
measured in control (P < 0.05). The semi-logarithmic plot in
Fig. 5C shows that the time
course of the late component of depression during m
depolarization was faster than that in control solution or oligomycin alone.
Thus the increased depression of phasic release during m
depolarization was not due solely to inhibition of mitochondrial ATP
synthesis. Figure 5D shows the
early components of depression, after subtraction of the late component. This
plot suggests that during m depolarization a transient
facilitatory process temporarily delayed the extra depression of phasic
release that later ensued.
In contrast to the depression seen in 2 mM Ca2+,
50-Hz stimulation in 0.5 mM Ca2+ increased epp amplitude
and quantal content under all conditions
(Fig. 6A). Average
normalized data in Fig.
6B show that in control solution phasic release increased
during the first 150 stimuli and then stabilized for the remainder of the
train at 3.4 times the pretrain value. In oligomycin alone, phasic
release also increased and then plateaued at 2.4 times the pretrain
value. Stimulation after m depolarization enhanced phasic
release by a maximum of 4.5-fold, significantly greater than the
enhancement in control and oligomycin-only solutions. After this peak at
3–5 s, phasic release declined slightly.
Figures 5B and 6B plot quantal contents normalized to their pretrain values, the traditional way of plotting depression and facilitation. However, in both physiological and low [Ca2+] conditions, the pretrain quantal contents in oligomycin alone tended to be higher than those measured in control solution, although this difference did not reach statistical significance (Table 1). This tendency toward higher initial quantal contents might have contributed to the finding that in oligomycin alone the normalized quantal contents showed greater depression than control values as stimulation progressed.
During 50-Hz stimulation average quantal contents in 0.5 and 2 mM Ca2+ converge
Comparison of Figs.
5A and
6A demonstrates the
surprising finding that, under five of the six experimental conditions
examined, the final average quantal contents achieved during 50-Hz stimulation
were similar. Figure 7 plots
the time course with which these quantal contents converged. In both control
solutions (Fig. 7A)
and oligomycin alone (Fig.
7B) the initial quantal contents were large (50–57)
in 2 mM Ca2+ and small (7–8.5) in 0.5 mM
Ca2+, but during 50-Hz stimulation the quantal contents
converged to a similar value (20) within 6 s. Toward the end of the
50-Hz train, epps in 0.5 mM Ca2+ occasionally produced
an action potential (not shown). Because for this study we selected partially
depolarized muscle fibers in which the threshold for action potential
generation was higher than normal, we suspect that in normally polarized
muscle, the depolarization produced by 20 quanta would have been sufficient to
elicit an action potential most of the time. After m
depolarization (Fig.
7C) in 0.5 mM Ca2+, quantal
contents reached a similar value, but in 2 mM Ca2+
quantal contents continued to depress throughout the course of
stimulation.
Repetitive stimulation produces more asynchronous release after
m depolarization
Figure 8 plots the average
increase in asynchronous release measured during stimulus trains applied under
control conditions, in oligomycin alone, and after m
depolarization in 2 mM (A) and 0.5 mM Ca2+
(B). At both Ca2+ levels, the rate of
asynchronous release in control and oligomycin alone increased slightly during
the train, but remained <25 s–1, consistent
with the modest increase in cytosolic [Ca2+] measured
under these conditions (Fig.
2). After m depolarization, there was a much
greater increase in asynchronous release rates during the train, to ≥450
s–1 in 2 mM Ca2+ and 250
s–1 in 0.5 mM Ca2+
(P < 0.001, SNK). These values are probably underestimates, for
several reasons. First, the extensive mepp overlap at higher frequencies
(>100 s–1) made it difficult to detect all
mepps. Second, the falling phase of epps sometimes extended throughout most of
the 20-ms interstimulus interval, making it difficult to detect all
superimposed mepps. Posttrain measurements often revealed higher mepp
frequencies not discernable during the train. An example is shown in
Fig. 9, where the measured
asynchronous release rate increased from 300
s–1 during tetanic stimulation to almost 600
s–1 immediately after stimulation. Third, the need
to use partially depolarized muscle fibers (especially in 2 mM
Ca2+, see METHODS) reduced mepp amplitudes,
making them harder to detect. A decrease in vesicular acetylcholine content
during repetitive stimulation could also contribute to an underestimation of
asynchronous release (Naves and van der
Kloot 2001). For these reasons, asynchronous release rates during
trains must have been greater than the values measured here, with larger
errors likely in 2 mM Ca2+ and at the higher mepp
frequencies.
Histograms in Fig. 10 show
average cumulated phasic and asynchronous quantal release measured during the
stimulus trains. In 2 mM Ca2+ m
depolarization significantly decreased summed phasic release but increased
summed asynchronous release by almost 20-fold. In 0.5 mM
Ca2+, summed asynchronous release increased ≥10-fold
during m depolarization, but summed phasic release was not
significantly different from control. In both 2 and 0.5 mM
Ca2+, total release (phasic plus asynchronous) was not
significantly different across treatments, although m
depolarization increased the percentage of release that was asynchronous. In
contrast, the dye-destaining study of Fig.
3 indicated greater total vesicular release during
m depolarization. This discrepancy joins with the reasons
noted in the preceding text to suggest that asynchronous release during
m depolarization was undercounted. Thus the data of
Fig. 10 likely underestimate
the degree to which release was desynchronized during 50-Hz stimulation when
m was depolarized.
More detailed comparisons between vesicular release estimated by dye-destaining and quantal release measured by electrophysiology were not possible because, in addition to the undercounting of asynchronous quanta, the duration of stimulation used for the electrophysiological experiments (10 s at 50 Hz) was short compared with the train durations usually used for dye-destaining experiments. Dye measurements over short intervals are complicated by the kinetics of destaining, which usually begins with a lag after the onset of stimulation. (In our experience, this lag was less for FM2-10 than for FM1-43, perhaps because FM2-10 is less hydrophobic than FM1-43.) We were constrained to use shorter durations of stimulation in the electrophysiological experiments because of the need to administer multiple stimulus trains in most experiments. Use of longer durations would have required intertrial intervals even longer than the 10 min used here.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
m (Fig.
2). Our finding that stimulation-induced increases in
mitochondrial [Ca2+] reached similar plateau values in
both low and physiological [Ca2+] agrees with previous
work showing that the plateau amplitude does not vary when stimulation
frequency is varied from 20 to 100 Hz or when Ca2+
influx per action potential is increased or decreased
(David 1999;
David et al. 2003). David
(1999) hypothesized that this
constant plateau amplitude results from reversible formation of an osmotically
inactive calcium complex within the matrix.
Effects of inhibiting mitochondrial ATP synthesis without
m depolarization
Inhibition of mitochondrial ATP synthesis with oligomycin (which does not
depolarize m) had relatively little effect on
stimulation-induced changes in total vesicular release
(Fig. 3) or on phasic and
asynchronous quantal release (Fig.
10). This finding is consistent with the previous demonstration
that oligomycin had no significant effect on stimulation-induced increases in
cytosolic [Ca2+] in this preparation
(David 1999). Oligomycin did
appear to reduce phasic release during 50-Hz stimulation in normalized records
in both physiological and low-bath [Ca2+] (Figs.
5B and
6B), but this effect
may have been related to the slightly higher initial quantal contents measured
in oligomycin. In nonnormalized measurements, it was clear that the brief
oligomycin exposures used here had little or no effect on the terminals'
ability to sustain phasic release during maintained stimulation. Because some
steps in vesicular trafficking require ATP
(Heidelberger et al. 2002;
Klenchin and Martin 2000;
Ohyama et al. 2002), our
findings suggest that oligomycin alone did not produce severe ATP depletion in
these motor terminals. Perhaps ATP sufficient to sustain release diffused into
the terminal from the axon, which is partially protected from the effects of
bath-applied drugs by myelin and the perineural sheath. Motor terminals and
axons can also make ATP by glycolysis. Peripheral motor axons have been shown
to be more resistant to ischemia than central axons, perhaps due to a greater
capacity for anaerobic metabolism (reviewed by
Stys et al. 1995). The
glycolytic capacity of motor axons/terminals probably helps to sustain release
as the partial pressure of O2 falls to low levels within
tetanically stimulated muscles (Koch
2002; Richardson et al.
1999; Wagner
2001).
Desynchronization of release during m
depolarization
For the ≤2-h exposure times used here, m depolarization
had no significant effect on the baseline mepp frequency or on epp quantal
content during 1-Hz stimulation. However, m depolarization had
marked effects on the asynchronous and phasic release measured during 50-Hz
stimulation.
m depolarization increased the rate of asynchronous release
during stimulus trains in both low and physiological bath
[Ca2+]. One likely reason is the greater elevation of
cytosolic [Ca2+] recorded in response to stimulation
during m depolarization. Under control conditions,
mitochondrial Ca2+ sequestration restricted the
elevation of spatially averaged cytosolic [Ca2+] to
≤1 µM, but much higher levels were attained when this sequestration was
inhibited by m depolarization
(Fig. 2, B and
C) (David
1999; David et al.
1998). The marked increase in asynchronous release measured under
these conditions is consistent with reports that motor and some other
presynaptic terminals and secretory cells contain release mechanisms with
Kds < 10 µM
(Angleson and Betz 2001;
Augustine and Neher 1992;
Bittner and Holz 1992;
Bollmann et al. 2000;
Ravin et al. 1997;
Rieke and Schwartz 1996;
Schneggenburger and Neher
2000). As expected, the stimulation-induced elevations of both
cytosolic [Ca2+] and asynchronous release during
m depolarization were greater in 2 mM than in 0.5 mM
Ca2+, but our data were not sufficient to permit
quantification of the relationship between cytosolic
[Ca2+] and asynchronous release. One reason was that
during m depolarization the stimulation-induced elevations in
cytosolic [Ca2+] changed over time
(Fig. 2), and we have not yet
succeeded in measuring cytosolic [Ca2+] and release
simultaneously from the same terminal. Another reason was that asynchronous
release rates evoked by stimulation during m depolarization
exceeded the ability for accurate measurement by the counting method as also
found by David and Barrett
(2003) in mouse motor
terminals.
m depolarization increased phasic release during 50-Hz
stimulation in 0.5 mM Ca2+, but this increase was
transient, because as stimulation continued, release returned to the levels
measured in non-m depolarized preparations
(Fig. 6B). Effects of
m depolarization during stimulation in 2 mM
Ca2+ were qualitatively similar, but in this case, the
early facilitation was smaller (Fig.
5D) and the ensuing depression of phasic release was more
profound (Fig. 5C).
The marked difference between the effects of oligomycin alone and
m depolarization indicates that the effects of
m depolarization on phasic release were not due solely to
inhibition of mitochondrial ATP synthesis. Rather, it appears that these
effects were due, directly or indirectly, to the greater stimulation-induced
elevations of cytosolic [Ca2+] recorded during
m depolarization or to the combination of elevated cytosolic
[Ca2+] and reduced ATP synthesis. The modest elevations
of cytosolic [Ca2+] at the onset of stimulation during
m depolarization would be expected to facilitate phasic
release (in accord with "residual Ca2+"
models, reviewed in Zucker and Regehr
2002), but greater or longer-lasting [Ca2+]
elevations have been reported to depress phasic release
(Adams et al. 1985;
Augustine and Neher 1992;
Hsu et al. 1996). This
depression was likely not due to a decrease in the sensitivity of postsynaptic
receptors, because mepps were readily detected throughout and after the 50-Hz
tetanus (Fig. 9). Nor was the
depression likely due to failure of action potential conduction because epps
continued to be recorded throughout the stimulus train. Reduced phasic release
when cytosolic [Ca2+] becomes excessively elevated might
be caused by direct effects, e.g., "adaptation" of the release
mechanism (Hsu et al. 1996) or
inhibition of N-type Ca2+ channel activity
(Shirokov 1999) or by indirect
effects associated with the increased asynchronous release. During 50-Hz
stimulation in both 2 and 0.5 mM Ca2+, the ratio of
asynchronous release to phasic release increased, i.e., release became
desynchronized (see also David and Barrett
2003; Kirischuk and Grantyn
2003). Excess asynchronous release might inhibit phasic release
via activation of inhibitory autoreceptors on the motor terminal (reviewed by
Wu and Saggau 1997) or via
competition if both types of release share a vesicle pool (as suggested by
Hagler and Goda 2001) and/or
if release of one vesicle transiently depresses the likelihood of further
release from the same active zone (Stevens
and Tsujimoto 1995).
Convergence of phasic release in physiological and low bath [Ca2+] during repetitive stimulation
A striking feature of our data was that the phasic release measured in 2
and 0.5 mM Ca2+ converged to a common quantal content
(20) during repetitive stimulation
(Fig. 7). This quantal content
would have been adequate to sustain neuromuscular transmission in muscle
fibers with a normal resting potential and thus may be just sufficient to
sustain release during prolonged tetanic stimulation. This ability to attain
suprathreshold release rates even in low-bath [Ca2+]
would preserve neuromuscular transmission if [Ca2+] in
the synaptic cleft falls during repetitive stimulation (a possibility
discussed in Borst and Sakmann
1999; Ginsburg and Rahamimoff
1983; Stanley
2000). The convergence of epp quantal content measured here is
reminiscent of the convergence of total release (phasic plus asynchronous)
noted by Hagler and Goda
(2001) in hippocampal autaptic
synapses in vitro during 20-Hz stimulation in varying bath
[Ca2+].
This convergence of quantal content might not seem surprising given the
abundant evidence that during sustained repetitive stimulation, phasic release
in physiological [Ca2+] depresses whereas phasic release
in lower [Ca2+] potentiates. But the mechanisms by which
the terminal achieves this convergence may shed light on synaptic control
mechanisms. A full examination of such mechanisms is beyond the scope of this
study, but the data of Fig.
2A suggest that one mechanism contributing to the
convergence involves control of the stimulation-induced elevation of cytosolic
[Ca2+]. If this elevation were proportional to bath
[Ca2+], one would have expected a fourfold greater
elevation in 2 than in 0.5 mM [Ca2+], but the measured
elevations indicated a difference of only 1.3-fold near the end of the
stimulus train. This similarity of the elevations of cytosolic
[Ca2+] in 0.5 and 2 mM Ca2+
suggests that during repetitive stimulation Ca2+ influx
becomes a less-than-linear function of bath [Ca2+]
and/or that intraterminal Ca2+
sequestration/buffering/extrusion mechanisms become greater-than-linear
functions of cytosolic [Ca2+]. Mechanisms that might
make Ca2+ influx a less-than-linear function of bath
[Ca2+] include inhibition of N-type
Ca2+ channels by elevated cytosolic
[Ca2+] or by inhibitory autoreceptors (references cited
in the preceding text) or the Ca2+-activated
K+ conductance in these terminals
(Morita and Barrett 1990),
which shortens the duration of the action potential as bath
[Ca2+] increases. An intraterminal
Ca2+ handling mechanism with a greater-than-linear
dependence on cytosolic [Ca2+] is the opening of the
mitochondrial uniporter (reviewed by
Gunter and Pfeiffer 1990).
Interestingly, the quantal content at the end of a 500-stimulus, 50-Hz
train was similar not only over a range of bath [Ca2+],
but also when mitochondrial ATP synthesis was inhibited by oligomycin. Phasic
release was even well sustained during brief (<2 h) m
depolarization in low bath [Ca2+]. In all these
conditions, the stimulation-induced increase in cytosolic
[Ca2+] remained modest (<5 µM). But phasic release
was not well sustained during m depolarization in
physiological [Ca2+], where the stimulation-induced
elevation of cytosolic [Ca2+] rapidly reached higher
levels. These findings suggest that mitochondrial Ca2+
uptake is most critical at higher rates of Ca2+ influx
and that limiting the increase in spatially averaged cytosolic
[Ca2+] during repetitive stimulation is important for
sustaining phasic release and thereby neuromuscular transmission.
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
This work was supported by National Institute of Neurological Disorders and Stroke Grants NS-12404 and 2T32 NS-07044.
|
FOOTNOTES |
|---|
Address for reprint requests: Corresponding Author: Ellen F. Barrett, Department of Physiology and Biophysics R-430, University of Miami School of Medicine, P.O. Box 016430, Miami, FL 33101 (E-mail: ebarrett2{at}med.miami.edu).
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGMENTSREFERENCES |
|---|
Alnaes E and
Rahamimoff R. On the role of mitochondria in transmitter release from
motor nerve terminals. J Physiol
248: 285–306,
1975.
Angleson JK and
Betz WJ. Intraterminal Ca2+ and spontaneous
transmitter release at the frog neuromuscular junction. J
Neurophysiol 85:
287–294, 2001.
Augustine GJ and Neher E. Calcium requirements for secretion in bovine chromaffin
cells. J Physiol 450:
247–271, 1992.
Babcock DF,
Herrington J, Goodwin PC, Park YB, and Hille B. Mitochondrial
participation in the intracellular Ca2+ network.
J Cell Biol 136:
833–844, 1997.
Babcock DF and Hille B. Mitochondrial oversight of cellular Ca2+ signaling. Curr Opin Neurobiol 8: 398–404, 1998.[ISI][Medline]
Betz WJ, Mao F, and Smith CB. Imaging exocytosis and endocytosis. Curr Opin Neurobiol 6: 365–371, 1996.[ISI][Medline]
Bittner MA and
Holz RW. Kinetic analysis of secretion from permeabilized adrenal
chromaffin cells reveals distinct components. J Biol
Chem 267:
16219–16225, 1992.
Bollmann JH,
Sakmann B, and Borst JG. Calcium sensitivity of glutamate release in a
calyx-type terminal. Science
289: 953–957,
2000.
Borst JG and
Sakmann B. Depletion of calcium in the synaptic cleft of a calyx-type
synapse in the rat brain stem. J Physiol
521: 123–133,
1999.
Calupca MA, Prior C, Merriam LA, Hendricks GM, and Parsons RL. Presynaptic function is altered in snake K+-depolarized motor nerve terminals containing compromised mitochondria. J Physiol 532: 217–227, 2001.
) or 0.5 mM Ca2+
(
). Decreasing bath [Ca2+] decreased the rate of
rise of mitochondrial [Ca2+], but did not alter the
plateau amplitude. Data were averaged from 2 terminals in a preparation
different from that in A. The rate of rise of mitochondrial
[Ca2+] at a given bath [Ca2+]
showed considerable variability between terminals and preparations, but
repeated trains administered to the same terminals yielded highly reproducible
responses.
), and oligomycin plus 
). C: semi-logarithmic plot of the late component (5–10
s) of depression of normalized quantal content. Linear regressions yielded
time constants and relative amplitudes (at t = 0) of (respectively)
control: 36.2 s, 0.60; oligomycin: 31.1 s, 0.45; 
), in oligomycin alone (
) and 31 min (
; 53
min, 
= oligomycin alone,
