| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
and I
Charge Movement with Calcium Release in Frog Skeletal Muscle
Department of Cellular and Integrative Physiology, Indiana University Medical Center, Indianapolis, Indiana
Correspondence: Address reprint requests to Dr. Chiu Shuen Hui, Dept. of Cellular and Integrative Physiology, Indiana University Medical Center, 635 Barnhill Dr., Indianapolis, IN 46202. Tel.: 317-274-8238; Fax: 317-274-3318; E-mail: cshui{at}iupui.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
hump in the ON segments of charge movement traces remained invariant when the calcium release rate was greatly diminished. When the nominal free [Ca2+]i was 50 nM, which was close to the physiological level, the I humps were accelerated and a slow calcium-dependent I
component (or state) was generated. The peak of ON I synchronized perfectly with the peak of the calcium release rate whereas the slow decay of ON I followed the same time course as the decay of calcium release rate. Suppression of calcium release by TMB-8 reduced the amount of Q concomitantly but not completely, and the effects were partially reversible. The same simultaneous suppression effects were achieved by depleting the sarcoplasmic reticulum calcium store with repetitive stimulation. The results suggest that the mobility of Q needs to be primed by a physiological level of resting myoplasmic Ca2+. Once the priming is completed, more I is mobilized by the released Ca2+ during depolarization. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
humps appeared in the ON segments of charge movement traces (Horowicz and Schneider, 1981
; Hui and Chandler, 1990, 1991; Hui, 1990, 1991a,b; Chen and Hui, 1991a,b; Hui and Maylie, 1991; Csernoch et al., 1991; Garcia et al., 1991a,b; Pizarro et al., 1991; Hui and Chen, 1992a,b, 1994a,b, 1995; Jong et al., 1995; Pape et al., 1996; Francini et al., 2001; Pape and Carrier, 2002). Even with 3 mM EGTA, Vergara and Caputo (1983) were able to record charge movement traces with relatively prominent I humps. However, although a high concentration of EGTA facilitates the study of the I hump, it reduces the resting free [Ca2+] in the myoplasm to far below the physiological level. We therefore removed practically all the EGTA in the end-pool solution, keeping only 0.1 mM to chelate any contaminating Ca2+ in the solution, so that charge movement could be measured under conditions close to the physiological state of the fiber. We found that, under this condition, Q was very much reduced (Hui and Chen, 1997) and the associated I hump was much less resolvable. This finding is interesting because it explains a longstanding mystery of why the I hump was not present in charge movement traces published by some investigators (Kovacs et al., 1979; Melzer et al., 1986; Rios and Brum, 1987; Rios and Pizarro, 1988; Brum et al., 1988; Feldmeyer et al., 1990; Simon and Hill, 1992).
With the diminution of Q in a fiber exposed to an extremely low [EGTA]i, the total charge, Qtotal, in the fiber appeared to be reduced. It is now clear that the reduction in Qtotal was an artifact created by the pulse protocol and the baseline correction procedure that had been employed traditionally to record and analyze the signal. Specifically, the pulse duration was too short and insufficient length of the OFF segment of the current trace was digitized, leading to the truncation of the ON and OFF segments. When the lengths of both segments were increased, a new component of charge movement lasting hundreds of milliseconds became apparent (Hui, 1998). This new charge movement component depends on the presence of Ca2+ in the myoplasm. It was named I and its associated charge was named Q. With the presence of Q, Qtotal is actually increased rather than decreased in low [EGTA]i.
The appearance of the I component raised some important questions. First, is I related to I? One possibility is that Q and Q are entirely unrelated distinct charge entities, i.e., the decrease in Q and the appearance of Q are independent of each other. This was shown to be not the case by Pape et al. (1996), who reported a slowing of the kinetics of I caused by the feedback of calcium release. One can then extend their observation to postulate that Q could be converted to Q. In other words, Q and Q belong to the same charge entity but are manifestations of different kinetic states that depend on the level of free [Ca2+]i. However, I found that, under the conditions of my experiments, even if all the Q was converted to Q, there was insufficient Q to account for the large amount of Q (see Table 2 and associated text in Discussion). Thus, some additional Q is apparently mobilized by depolarization when the free [Ca2+]i is restored to the physiological level. Even if the first part of Q shares the same origin as Q, the additional part might not. If so, how are I and I associated with calcium release? The experiments reported in this article were aimed at providing some answers to the latter question. It was found that the results from the experiments are consistent with the idea that Q could be the trigger for calcium release and part of Q could be generated by the release. Thus, the longstanding controversy between the "trigger hypothesis" and the "feedback hypothesis" for Q might have been resolved. The information concerning whether Q and Q belong to the same charge entity or are separate distinct charge species will be presented in the Discussion. Because of the complexity of Q, a complete answer to this question is still unavailable.
|
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
, and Hui and Chandler, 1990). Special care was taken to minimize the amount of contaminating calcium introduced by the ion exchange procedure (Hui, 1998). With 20 mM EGTA in the internal solution, the amount of free Ca2+ was estimated to be <10–12 M. With either 0.4 or 1.8 mM total added calcium, the nominal free [Ca2+]i was estimated to be 
10 nM in solution B and 50 nM in solutions C and D. The latter value is assumed to be close to the physiological level. Antipyrylazo III (ApIII) was bought from ICN K & K Laboratories, Plainview, NY, and dissolved in the internal solution. 8-(N,N-diethylamino)octyl 3,4,5-trimethoxybenzoate hydrochloride (TMB-8) was purchased from Calbiochem and the appropriate amount was dissolved directly in the external solution.
|
4°C. In accordance with a procedure approved by the Institutional Animal Care and Use Committee, animals were killed by decapitation and destruction of the brain and spinal cord. The procedure for dissecting and mounting cut fibers from semitendinosus muscle was similar to that used by Kovacs et al. (1983) and Irving et al. (1987). Cut fiber segments were dissected in solution A. A stretched fiber segment was mounted in a double Vaseline-gap chamber. To facilitate the entry of the calcium indicator ApIII into the myoplasm, the outer membranes of the fiber in the end pools were permeabilized by a 2-min exposure to 0.01% saponin (Sigma). The beginning of the treatment marked time zero of an experiment. After the proper internal and external solutions were introduced, the voltage clamp was turned on at about the 20th–23rd minute, and the fiber was repolarized. The fiber was allowed to recover for 30 min during which various ions diffused into the myoplasm in the center-pool region. The temperature of the center-pool solution was kept at 13–14°C.
Data acquisition
The instrumentation for data acquisition was designed and fabricated by the Biomedical Instrumentation Laboratory of Yale Department of Cellular and Molecular Physiology. Ten analog signals were connected to the input channels of the module. They included six optical signals (see below), three electrical signals, and the temperature. The electrical signals were the potential in one end pool (V1), the potential in the other end pool (V2), and the total current injected into the latter end pool (I2). The cutoff frequency of the eight-pole Bessel filter in each channel was set at 0.6 kHz. Data was digitized at a rate of 10 µs per point and sent to a PDP 11/73 computer for processing. The points in each channel were compressed before storage. As a result, each point in a stored trace corresponds to 1 ms. Multiplexing of the channels was arranged in a way to synchronize the compressed points in all the channels in time.
Charge movement measurement
Holding potential was set at –90 mV. Control pulses were applied from –110 mV to the holding potential and test pulses from the holding potential to the potentials desired. The condition of the fiber was tracked by monitoring the holding current throughout an experiment. Subsequent data analysis included linear cable analysis of the control records. The analysis yielded information about myoplasmic resistance (ri), membrane resistance (rm), membrane capacitance (cm), and gap factor of the Vaseline seals defined by re/(re + ri) (Chandler and Hui, 1990). The terms re and ri represent the external and internal resistance per unit length of the fiber underneath the Vaseline seals. Each Icontrol trace was an average of four sweeps and all Itest traces were single-sweep. Each charge movement measurement is displayed as an Itest – Icontrol trace, which is obtained by subtracting a scaled Icontrol trace from a paired Itest trace.
To estimate the amount of Q in a charge movement trace, the procedure developed in Hui (1998) was applied to the OFF segment of the trace. The first step was to correct for the baseline. Each OFF segment was fitted by a sum of two exponentials and a sloping straight line, with the fastest phase of decay being excluded from the fit. For comparison, the fitting was repeated by replacing the sloping straight line with a constant. In principle, in the presence of a nonlinear time-dependent ionic current, such as a changing leakage current underneath the Vaseline seals, fitting with a sloping straight line should be more accurate, whereas in the absence of such a current, both fitting methods should yield similar results if a sufficiently long baseline was recorded. As will be shown in Figs. 6 and 8, the latter turned out to be true in the experiments presented. The straight line or constant resulting from the fit was used for baseline correction and the longer time constant was used to represent the decay time constant 
of the OFF I component. An integration of the complete OFF transient yielded the OFF Qtotal. The second step was to separate the OFF I component from the OFF Itotal. The OFF I component was found to have a rising phase with a time constant r (Hui, 1998). The waveform of the OFF I component can thus be expressed as
 | (1) |
.
|
|
and Maylie et al. (1987a,b) and had been used in our previous work (Maylie and Hui, 1991; Hui, 1999). The optical system was built on an upright microscope (model ACM, Carl Zeiss, New York, NY). Optical measurements were made with a 55.5-µm diameter spot of light focused on the axis of the fiber segment located in the center-pool region. Since three wavelengths are required to accurately describe the calcium indicator signal in muscle (Baylor et al., 1982), the transmitted light was separated into three beams with two beam-splitting cubes. The beams were made quasimonochromatic by passing through three interference filters with peak transmission wavelengths at 550 nm (10-nm bandwidth), 720 nm (30 nm), and 810 nm (30 nm). Each beam was further split into two beams of linear polarizations (0° and 90° with respect to the fiber axis) with a polarizing beam-splitting cube.
The intensities of the six resulting beams were monitored with photodiodes (model UV-100B, EG&G Electro-Optics Div., Salem, MA) and fed to the inputs of the optical channels of the data acquisition module. Absorbance measurements were made with unpolarized incident light (mode 1 in Irving et al., 1987). The absorbance at each wavelength A(
) was computed from the 1:2 average of A0:A90. The intrinsic absorbance signal 
Ai(810) was filtered by a 0.05-kHz digital Gaussian filter (Colquhoun and Sigworth, 1983). The procedures described in Maylie et al. (1987a) were used to subtract the contributions due to the intrinsic absorbance changes to yield the dye-related A(720).
Computation of calcium release rate
In this article, calcium release rate (Rel) will be used to refer to the d[CaT]/dt signal, in which CaT represents the total amount of calcium (free and bound) in the myoplasm. Rel was computed from the dye-related A(720), based on a model similar to the one used by Baylor et al. (1983), but modified to include the binding of Ca2+ to EGTA in addition to its binding to ApIII, troponin, and parvalbumin. The following values were used for the parameters in the model: 2.55 x 104 M–1 cm–1 for 
(550) of ApIII, 1.46 x 104 M–1 cm–1 for (720) of the Ca-ApIII complexes, 3.4 x 10–8 M2 for the apparent kD between Ca2+ and ApIII, 240 µM for the total concentration of troponin, 0.575 x 108 M–1 s–1 and 115 s–1 for k1 and k–1 of Ca2+ binding to troponin, 2 mM for the total concentration of parvalbumin, 1.25 x 108 M–1 s–1 and 0.5 s–1 for k1 and k–1 of Ca2+ binding to parvalbumin, 3.3 x 104 M–1 s–1 and 3.0 s–1 for k1 and k–1 of Mg2+ binding to parvalbumin, and 1.0 x 106 M–1 s–1 and 0.5 s–1 for k1 and k–1 of Ca2+ binding to EGTA. Throughout this article, only the computed Rel traces will be shown. For a typical computation from the dye-related A(720) traces to yield the free [Ca2+] traces, the [CaT] traces, and finally the Rel traces, refer to Figs. 1 and 2 A of Hui (1999). The peak amplitude of Rel, represented by Relp, and the time-to-peak in each trace were determined by fitting the peak with a parabolic function.
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
component as a prominent hump (see references in Introduction). Unfortunately, calcium transient cannot be monitored in these fibers with a metallochromic calcium indicator such as ApIII, presumably because most of the Ca2+ released from the sarcoplasmic reticulum (SR) is precluded from successful binding with ApIII by EGTA and the free [Ca2+]i is too low to replenish the SR calcium store after activities. To enable the observation of calcium transient, the [EGTA]i has to be reduced, but when the concentration is lowered to 0.1 mM, the I hump can hardly be resolved (Hui and Chen, 1997). The first experiment to be reported here is aimed at exploring whether there is any condition under which both prominent I hump and sizable calcium transient can be recorded simultaneously. The rationale is that perhaps the 20 mM EGTA should be retained and it might be possible to partially restore the free [Ca2+]i to a level just sufficient to enable the measurement of calcium transient without compromising the prominence of the I hump.
Fig. 1 shows the results from an experiment of this kind. The fiber was bathed in a TEA-Cl external solution. By trial and error, it was found that the optimal free [Ca2+]i was 10 nM (solution B). A sequence of depolarizing pulses with amplitudes ranging from 10 to 90 mV were applied at 2-min intervals to elicit pairs of charge movement (left column) and Rel (right column) traces. Only the traces at intermediate depolarizations are shown in Fig. 1 A to reveal the prominent I humps (as indicated by the arrowheads) in the electrical traces. Although the nominal free [Ca2+]i was 10 nM, the magnitudes of Rel were quite substantial in the early stage of the experiment. In this and the next four figures, only ON segments of the charge movement traces are shown for comparison with the Rel traces. The OFF segments are irrelevant as the inward charge movement had no correlation with the cessation of calcium release.
Another identical sequence of depolarizing pulses was applied almost half an hour later in the experiment and the pairs of traces in the same potential range are shown in Fig. 1 B. The Rel traces shown in the right column were very much reduced. From the values of Relp given in the figure legend, calcium release was reduced to 10% of those in Fig.1 A. This diminution can be attributed to the insufficient replenishment of the SR calcium store after multiple large depolarizations (between the two sequences, not shown) because the free [Ca2+]i was below the physiological level. In contrast to the reduction in Rel, a comparison of the charge movement traces in Fig. 1, A and B, showed that the I humps remained unchanged between the two sequences of runs. The same invariance was observed in two other fibers in which the same experimental protocol was used. This reinforces our previous conclusion that I cannot be generated by the feedback of calcium release from the SR.
The invariance in the shape of the I hump reported here is different from the finding that repetitive stimulation slowed the kinetics of I (Hui, 1991b) and that calcium release exerted feedback on the kinetics of I (Jong et al., 1995; Pape et al., 1996; Hui and Chen, 1997) but in agreement with the finding that the reduction of calcium transient by low concentrations of tetracaine had no influence on I (Csernoch et al., 1988). It should be noted, however, that the experimental conditions employed in the various studies were different.
Another interesting relationship between I and Rel is shown in Fig. 2. To facilitate the comparison of the time courses of I and Rel, the format for displaying the traces in Fig. 1 A is changed such that each pair of traces is lined up in time, with the electrical trace directly above the optical trace. After the rearrangement, it became obvious that the peak of I preceded the peak of Rel waveform at all potentials. The same observation was also reported by Csernoch et al. (1988).
This temporal relationship between I and Rel might provide additional support for the above conclusion that I cannot be generated by the feedback of calcium release. However, it should be noted that charge movement is a signal localized in the T-tubules whereas calcium transient is a global signal averaged over all regions of the sarcomeres illuminated by the light spot. Diffusion of the released Ca2+ from the triads to the A-bands should introduce some weighted delay in the global calcium transient. Thus, this temporal relationship between I and Rel is not as definitive as the invariance of I shown in Fig. 1 in supporting the conclusion.
Charge movement and calcium release in cut fibers with physiological level of myoplasmic Ca2+
The effect of calcium release on charge movement was quite different when the free [Ca2+]i was restored to the physiological level, as shown in Fig. 3. The traces shown in Fig. 3 were recorded from a fiber with the free [Ca2+] in the internal solution adjusted nominally to 50 nM (solution C). With the elevated free [Ca2+]i, it was anticipated that calcium-dependent Cl– current would be activated during depolarization (Hui and Chen, 1994b) and complicate the analysis of charge movement traces. To avoid this problem, the Cl– in the external solution was completely replaced with an impermeant anion. The traces shown in Fig. 3 were recorded from a fiber bathed in a TEA-CH3SO3 external solution. Between –55 and –35 mV, small and brief I humps (indicated by arrowheads) after the first Iß peak can be visualized in the charge movement traces. The higher free [Ca2+]i abbreviated the durations of the I humps as compared with those in Fig. 2. This is consistent with the finding reported in Hui and Chen (1997). Fig. 2 A of that article showed specifically how I evolved from a broad prominent hump that was well separated from Iß to a brief small hump that followed Iß closely when the free [Ca2+]i was elevated. For a closer visualization of Iß and I on an expanded timescale, refer to trace 6 or 7 of Fig. 2 B in that study. Based on the information provided by that figure, there should be no doubt that the small humps marked by the arrowheads in my Fig. 3 are I.
|
hump. The late hump can be visualized in all the traces and can be aligned perfectly with the peak of the Rel waveform, as indicated by the dashed lines. One might argue that this late hump could be I in the conventional sense, but it is impossible because I should be accelerated when the resting free [Ca2+]i is at a higher level and hence cannot peak later than that in the fiber of Fig. 2. However, it is possible that the late hump might reflect the flow of part of Q that is transformed to a slow kinetic state, as reported by Pape et al. (1996). The late hump was observed in many other fibers in which the same experimental protocol was used, 61 in the same solution, 6 in sulfate, and 26 in gluconate. It is worth mentioning that the late hump was also observed in 42 fibers bathed in Cl–. In those experiments, the late hump was not obscured by the calcium-dependent Cl– current because the ionic current was activated quite slowly even at large depolarizations. All these results suggest that the late hump could be a genuine signal rather than an artifact and could be closely associated with calcium release because of their tight temporal relationship. It is thus hypothesized that the newly discovered late hump is the peak of the slow calcium-dependent component of charge movement that has been named I (Hui, 1998). It should be emphasized that this peak of I was not apparent in the traces of Figs. 1 and 2 when the free [Ca2+]i was at an extremely low level.
Association of I with calcium release
It was explained in Hui (1998) that to study the slow ON and OFF I current, the durations of the ON and OFF segments of the charge movement traces should be sufficiently long to enable baseline fits. Consequently, a calcium-free external solution needs to be used to avoid the slow inward Ca2+ current. Fig. 4 shows three pairs of electrical and optical traces elicited by 2000-ms depolarizing pulses to –40, –30, and –20 mV. To increase the temporal resolution, only the first 1000 ms of the traces are shown in the figure. The slowly decaying I components in the traces resembled those shown in Hui (1998). The application of long pulse and abolition of inward Ca2+ current facilitate the checking of ON-OFF charge equality. The checking procedure was presented in Fig. 5 of Hui (1998). The main feature in this figure is that the slow decays of I and Rel shared similar time courses, although the waveforms of the two signals were not exactly identical. The same similarity was observed in 14 other fibers bathed in either a calcium-free TEA-CH3SO3 or calcium-free TEA-gluconate external solution. This suggests that the slow I is closely associated with calcium release.
|
|
was used. The obscurity is universally true in all experiments employing long depolarizing pulses, which presumably deplete the calcium content of the SR more effectively. The employment of a calcium-free external solution might also contribute to the differences (Hui, 1999). Concomitantly, the slow humps that were observed in Fig. 3 were much more inconspicuous in Fig. 4. In fact, this is the strongest piece of supporting evidence that can be used to associate the slow hump with calcium release, other than the close temporal relationship between the two signals. Since I is also closely associated with calcium release (Fig. 4), the slow hump could actually be part of I, as hypothesized in the preceding section. It thus appears that I contains a peak followed by a slow decay but the conditions for observing both parts are mutually exclusive. The long depolarizing pulses (in conjunction with a calcium-free external solution) that are required for the measurement of the slow decay phase of I in Fig. 4 are not optimal for the detection of the peak of I. On the other hand, Fig. 3 showed that prominent peaks of I can be recorded by applying short depolarizing pulses, but under this condition, the decay phase of I is truncated.
Concomitant suppression of I and calcium release by TMB-8
To gain further support of the hypothesis that I is associated with calcium release, two interventions were applied to reduce calcium release and investigate whether I is affected concomitantly. In each intervention, the amount of Q was compared with the magnitude of Relp as the experiment progressed. For the purpose of estimating the amount of Q, the OFF segments of charge movement traces are much more useful than the ON segments, because the procedure that was developed to separate Q from Qtotal is only applicable to the OFF segments (see Methods). Hence, unlike Figs. 1–4
which show ON segments of charge movement traces, Figs. 5 and 7 show OFF segments instead.
|
, TMB-8 inhibited contraction in skeletal muscle by suppressing calcium release. Since TMB-8 could serve as a useful agent for hindering the calcium release process (Janis et al., 1987), it was used in the first intervention. An experiment employing TMB-8 is shown in Fig. 5. Each pair of charge movement trace (in Fig. 5 A) and Rel trace (in Fig. 5 B) was elicited by a constant 150-ms pulse to –20 mV and recorded simultaneously. Although the short pulses truncated the ON charge movement, they were less damaging and so slowed the run-down of the fiber. Long (1800-ms) OFF segments of the charge movement traces were recorded to facilitate fitting of baselines. To increase temporal resolution, only 800 ms of the segments are shown in the figure. The pair of traces in the first row was taken before the application of TMB-8. A sizable OFF I can be visualized in the charge movement trace after a very fast and a somewhat slower phase of decay, which presumably corresponded to the OFF Iß and I components, respectively. The Rel trace showed an early peak.
The second pair of traces was taken 1 min after the application of 60 µM TMB-8 to the center pool. TMB-8 acted very fast. Within a minute, the OFF I was reduced and the peak of Rel was suppressed. After another 3 min, the third pair of traces was taken. Both the OFF I and the peak of Rel were further reduced. To examine the difference between the first and the third charge movement traces more closely, the difference trace is shown in Fig. 5 D. TMB-8 appeared to suppress all three components of OFF charge movement, but the dominating effect was on I, as revealed by the sizable slow decay in the difference trace. The ON segment of the difference trace is also shown in Fig. 5 C for comparison. It has an I hump that peaks at 18 ms, more or less matching the peak of Rel in trace a of Fig. 5 B.
Another pair of traces (not shown) was taken before TMB-8 was washed out. The fourth pair of traces in the figure was taken 12 min after washout. Both the slow I component and the Rel waveform recovered appreciably but not completely, implying that the effect of TMB-8 was reversible, at least partially. Subsequently, TMB-8 was reapplied. The fifth pair of traces showed that calcium release was completely suppressed by then, but the slow I component was only partially reduced.
To differentiate the effect of TMB-8 on the slow I component from that on the fast Iß and I components, the amount of OFF Q was estimated from each charge movement trace shown in Fig. 5 (and others not shown) by the procedure described in Methods. The baseline was first corrected by fitting with a sum of two exponentials and a sloping straight line. Since the rising time constant r was not measured in this fiber, it was adopted from the average value obtained from other fibers. As shown in Hui (1998), the value was 8.2 ms. An integration of Expression 1 with this value of r yielded the values of OFF Q, which are plotted as solid diamonds in Fig. 6 A. Finally, the difference between each OFF Qtotal and OFF Q gave the sum of the OFF Qß and Q, which is plotted as an open square in Fig. 6 A. No attempt was made to separate Qß and Q and their sum will be referred to as Qß from here on. For comparison, the value of Relp was estimated from each optical trace and is plotted in Fig. 6 B.
The values of Q, Qß, and Relp estimated from the first pair of traces in Fig. 5 A (marked by the letter a in Fig. 6, A and B), namely 51.3 nC µF–1, 17.0 nC µF–1, and 24.2 µM ms–1, were used as control. The left shaded area in each panel indicates the first application of TMB-8. At the instant the third pair of traces was recorded (marked by c), TMB-8 had suppressed Q and Relp to 47 and 32% of control, respectively. TMB-8 was not without effect on Qß but the reduction was much smaller, to 73% of control. This agrees with Fig. 5 D. Since the amount of charge is normalized by the membrane capacitance, there was concern that the reduction in the amount of charge could be caused by an increase in membrane capacitance. Fig. 6 C is shown to clarify this point. The membrane capacitance did increase slightly at the instant the third pair of traces was taken (it was 110% of control), but the increase was not large enough to account for the decrease in Qß and definitely far too small to account for the even larger decrease in Q. After washout (indicated by the white area in each panel between the two shaded areas), the data marked by d showed that Q and Relp recovered hand in hand to 77 and 71% of control, whereas Qß recovered almost fully to 98% of control. At that time, membrane capacitance returned halfway to 106% of control. The second application of TMB-8 (indicated by the right shaded area) completely suppressed Relp but reduced Q and Qß to 37 and 76% of control (data marked by e) whereas the membrane capacitance was increased to 114% of control.
The amounts of Q and Qß were also estimated by fitting the OFF segment of each charge movement trace with a sum of two exponentials and a constant. The values for Q are plotted in Fig. 6 A as solid circles, which show exactly the same pattern of change as the solid diamonds when TMB-8 was applied and washed out. The values for Qß are not shown because they are extremely close to the values represented by the open squares. The values represented by the solid circle and the solid diamond from the same charge movement trace differ by 9 nC µF–1, on average. These small differences suggest that the sloping baselines obtained with the first method of fitting are essentially horizontal, i.e., not much different from the constant baselines obtained with the second method of fitting. This similarity can only be achieved with sufficient duration of the OFF segments, namely 1800 ms in the experiment shown in Figs. 5 and 6 (and similar experiments).
The main points of Fig. 6 are: first, the effect of TMB-8 on Rel correlates better with its effect on Q than that on Qß; second, when calcium release is completely suppressed by TMB-8, part of Q still remains mobile. Similar results were obtained from five other fibers that were exposed to 60–100 µM TMB-8. When calcium release was suppressed to a negligible level, Q was reduced to 53 ± 5 (S.E.M.) % of control at a potential of –30 to –20 mV, averaged over all six fibers. The reversibility of the effects of TMB-8 was examined in three of the other five fibers. In two of them, TMB-8 was washed out after 40–45 min of exposure and neither fiber could release calcium any more. In the third fiber, TMB-8 was applied for 10 min and Relp was down to 27% of control. On washout, Relp recovered to 67%. In the experiment shown in Figs. 5 and 6, the first exposure lasted only 9 min, which was sufficient to suppress Relp to <10% of control. Even with exposures as short as 9–10 min, the recovery of Relp after washout was incomplete. It might be possible to achieve complete recovery if exposure was much shorter.
Concomitant suppression of I and calcium release by repetitive stimulation
Another intervention to reduce the amount of Ca2+ released from the SR is to reduce the supply of Ca2+ inside the store. This can be accomplished by stimulating the fiber repetitively to deplete the Ca2+ store, and, obviously, long pulses are more effective for this purpose. This approach was used in the experiment shown in Fig. 7. The fiber was bathed in a calcium-free TEA-CH3SO3 external solution. Each pair of charge movement trace (in Fig. 7 A) and Rel trace (in Fig. 7 B) was elicited by a 2000-ms TEST pulse to –30 mV. Only 5 of the 13 pairs of traces taken are shown in the figure. Some other pairs of traces taken at other potentials are also not shown. The first pair of traces (a) was taken when calcium release was fairly typical and served as control. After three other stimulations (traces not shown), the second pair of traces (b) was taken. By then the peak amplitude of the Rel trace was already down to less than half of the control, indicating that the SR store was emptied quite effectively. When the fifth pair of traces (e) was taken, calcium release was barely noticeable and the slow I component was reduced. To examine the difference between the first and fifth charge movement traces more closely, the difference trace is shown on an expanded scale in Fig. 7 C. It shows the reduction of I as well as of Iß.
Following the analysis shown in Fig. 6, the amounts of Qß and Q were estimated from the charge movement traces shown in Fig. 7 (and others not shown). The amounts of OFF Q estimated from the integral of Expression 1, after fitting the OFF segments with a sum of two exponentials and a sloping straight line, are plotted as solid diamonds in Fig. 8 A. The amounts of OFF Qß obtained from the differences between OFF Qtotal and OFF Q are plotted as open squares in the same panel. The corresponding values of Relp were calculated from the Rel traces and are plotted in Fig. 8 B.
The values of Q, Qß, and Relp estimated from the first pair of traces in Fig. 7 A (marked by the letter a in Fig. 8, A and B), namely 58.4 nC µF–1, 17.1 nC µF–1, and 21.2 µM ms–1, served as control. At the instant the second pair of traces was taken (marked by b), partial depletion of the SR calcium store reduced Q to 69% of control and Relp to 41% of control. The Qß was hardly affected: it was reduced to 95% of control. The SR calcium store was more depleted at the end of the experiment. Q, Qß, and Relp were reduced to 53, 89, and 10% of control (marked by e). Throughout the experiment, the membrane capacitance was decreasing slightly instead of increasing as in Fig. 6 C. Its value when the fifth pair of traces was taken was 94% of control.
As in Fig. 6 A, the estimation of the amounts of Q was repeated by fitting the OFF segment of each charge movement trace with a sum of two exponentials and a constant. The values are plotted in Fig. 8 A as solid circles. They differ from the values represented by the solid diamonds by 7 nC µF–1, on average. With this baseline fitting, the residual amount of Q at the end of the experiment (marked by e) was 54% of control, which is practically the same as the 53% obtained with the former fitting.
Similar experiments were performed on two other fibers. In one fiber, Q, Qß, and Relp were reduced to 62, 90, and 16% of control. In the other fiber, the quantities were reduced to 52, 86, and 8%, respectively. The results, together with those from Fig. 8, show that calcium release can be reduced effectively by depleting the SR calcium store. They strongly support the conclusions drawn in the preceding section that the reduction of calcium release correlates with a reduction of Q (with much less change in Qß) and that when calcium release is greatly reduced, a portion of Q still remains mobile. On the one hand, calcium depletion is a preferable intervention because it avoids the suspicion of any undesirable secondary pharmacological effect exerted by TMB-8, such as its effect on Qß. On the other hand, by the time the SR calcium store is successfully depleted, reversibility is very difficult to achieve. Also, complete depletion of the SR calcium store may never be achieved unless an ATPase inhibitor is used in conjunction with repetitive stimulation.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
with calcium release and Hui (1983). That led to the proposal that Q is generated by this feedback (Pizarro et al., 1991; Shirokova et al., 1994). This has been referred to as the "feedback hypothesis" for Q. However, results in this article showed that Q cannot be generated by the feedback of released Ca2+ (also supported by Jong et al., 1995; Chawla et al., 2002; Squecco et al., 2003). First, when calcium release is reduced to a negligible level, the waveform and size of I remain unaltered (Fig. 1; see also Csernoch et al., 1988). Second, the chronological order of the peaks of I and Rel precludes the possibility of Q being a consequence of calcium release (Fig. 2; see also Csernoch et al., 1988). However, the second evidence is not as definitive as the first (see text associated with Fig. 2).
If Q is not a consequence of calcium release, and since it is closely associated with calcium release (Huang, 1982; Hui, 1982; Vergara and Caputo, 1983), then it is quite likely that Q is a trigger for calcium release. This supports the "trigger hypothesis" for Q and resolves the controversy between the "trigger hypothesis" and "feedback hypothesis". If so, the positive feedback of calcium release on charge movement mentioned in the preceding paragraph is missing. The slow I can fill the gap, as will be explained below.
Waveform of I
When charge movement is measured in a cut fiber containing a physiological level of free [Ca2+]i, a slowly decaying current was observed in the ON and OFF segments of the traces. The capacitive nature of this slow current was established in the preceding article (Hui, 1998). Specifically, the possibility that the current arises as a result of permeation of cations or anions through the outer membranes was ruled out. The current was called the I component of charge movement. Experiments were also performed to determine the waveform of the OFF I component in an attempt to develop a method for separating Q from the Qtotal. It was found that OFF I has a slow rising phase with a time constant of the order of 10 ms and the amount of Q is much larger than the combined amount of Qß and Q. The ON I should also have a slow rising phase but its time constant varies as a function of the potential during depolarization, making it difficult to separate the ON Q from the ON Qtotal
In the experiments presented in this article, the ON I component was studied in greater detail to gain information about the relationship between I and calcium release. Although the exact shape of the rising phase of ON I is still not known with certainty, it appeared that ON I consists of a peak (Fig. 3) and a slow decay phase (Fig. 4), quite similar in shape to the Rel waveforms computed from absorbance signals recorded at the same time. Thus, in the ON segment of a charge movement trace, the peaks of I and I are manifested as the early and late humps, respectively, after the fast Iß peak (Fig. 3).
It should be noted that the conditions optimal for the observation of the peak and the slow decay of ON I are mutually exclusive. The observation of the slow phase of ON I necessitates the use of long depolarizing pulses (Fig. 4), which makes the peaks in Rel and ON I much less conspicuous. These peaks are most prominent when measured with short depolarizing pulses (Fig. 3), which truncate the slow decay of ON I. Because of this, it is difficult to establish the capacitive nature of the late hump, as baseline correction cannot be carried out with short pulses and ON-OFF charge equality cannot be verified. Hence, the only piece of evidence used to support the inclusion of the late hump as part of I (which is a capacitive current) is its presence when the Rel trace shows a sharp peak (Fig. 3) and its absence when the Rel trace does not show a sharp peak (Fig. 4). Combining the results from both groups of experiments, it can be postulated that the waveform of ON I has the same shape as that of Rel.
Existence of I requires physiological level of free [Ca2+]i
The presence of a sharp peak in the Rel trace does not guarantee the appearance of the peak of ON I. Fig. 2 shows that when the free [Ca2+]i was a small fraction of the normal physiological amount, the peak of ON I could not be detected. Another peculiar feature of the traces is that the I humps were still broad and prominent as if the acceleration of I kinetics observed by Jong et al. (1995), Pape et al. (1996), and Hui and Chen (1997) did not occur, or occurred to a much lesser extent.
One possible explanation for these observations is that perhaps the mobilization of Q and the conversion of Q to Q in response to calcium release require initial priming by some resting Ca2+ in the myoplasm. The resting concentration has to be close to the physiological level and the priming time course is quite slow. Thus, when the resting concentration is below the physiological level, such as 10 nM in the case of Figs. 1 and 2, even a sizable calcium release for a brief period of time could not exert much effect on the I hump nor mobilize any Q because there was not sufficient time for the priming to take place during the brief elevation of free [Ca2+]i. The peculiar features in Fig. 1 were captured at just the appropriate resting free [Ca2+]i. Had the resting free [Ca2+]i been lower, Rel might be too small to be detected with ApIII. Had the resting free [Ca2+]i been higher, the shape of the I hump most likely would not be invariant when Rel was diminished.
Association of Q with calcium release
One might argue that the similarity between the waveforms of I and Rel could be a coincidence. To establish the association of Q with calcium release more firmly, two interventions were applied to interfere with calcium release and to observe how Q is affected. The first intervention made use of an intracellular calcium antagonist TMB-8, which has been used widely as a pharmacological tool for inhibiting calcium release from stores inside a variety of cell types (for review, see Janis et al., 1987). According to Malagodi and Chiou (1974), TMB-8 inhibited contraction in skeletal muscle by suppressing calcium release. Fig. 6 shows that when calcium release was blocked by TMB-8, Q was reduced concomitantly, and if TMB-8 was washed out promptly, both calcium release and Q were restored in parallel. The reversibility of the blockades supports the association of Q with calcium release quite convincingly. Unfortunately, the utilization of TMB-8 was not without flaw. The most serious concern is the reduction in Qß when Q was reduced. Nonetheless, it is quite likely that the apparent decrease in Qß was caused by some other complications unrelated to calcium release, as discussed next.
First, the amounts of Qß and Q could decrease progressively over the course of the experiment as a result of fiber run-down which was inevitable. Fortunately, the restoration of Qß after washout of TMB-8 indicated that the decrease could not be entirely due to run-down. Second, the amounts of Qß and Q extracted from each charge movement trace relied heavily on the method used to separate the charge components in the OFF segment. Since the time constant of the rising phase of OFF I was not measured in the experiment of Fig. 6, the value 8.2 ms was adopted from the mean time constant from other fibers and used in Expression 1 to estimate the amount of Q. This could introduce some error in the amount of Q, and thus some error in the amount of Qß (see also Discussion in Hui, 1998). However, the difference trace in Fig. 5 D that was obtained from raw data shows that some reduction of Qß is real. Third, like most pharmacological agents, TMB-8 appeared to exert undesirable side effects on the fiber. TMB-8 at least increased the membrane capacitance slightly (Fig. 6 C), suggesting that TMB-8 has the ability to affect membrane electrical properties. This effect was unrelated to a change in calcium release, because a reduction in calcium release without TMB-8 in the experiment of Fig. 8 decreased the membrane capacitance instead. If so, there is no reason to expect that TMB-8 cannot affect Qß and Q directly, although to a relatively minor extent. In fact, the nature of multiple actions of TMB-8 has been noticed by some investigators (for example, Himmel and Ravens, 1990). Nonetheless, the parallel trend of diminution and reversible restoration of Q
