| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Universität Ulm, Abteilung für Angewandte Physiologie, D-89069 Ulm, Germany
Correspondence: Address reprint requests to Werner Melzer, Univ. of Ulm, Dept. of Applied Physiology, Albert-Einstein-Allee 11, D-89069 Ulm, Germany. Tel: ++49-731-500-23248; Fax: ++49-731-500-23260; E-mail: werner.melzer{at}medizin.uni-ulm.de.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
; Bers, 2001). A rapid depolarization-activated Ca2+ inward current triggers Ca2+ release in cardiac cells (Wang et al., 2001), whereas in skeletal muscle, Ca2+ release is activated first, followed by a Ca2+ current of much slower kinetics (Brum et al., 1987; Friedrich et al., 1999; Szentesi et al., 2001). In skeletal muscle cells, a direct protein-protein interaction between the 
1 subunit (1S) of the transverse tubular dihydropyridine receptor and the ryanodine receptor (RyR1) appears to be responsible for the activation of Ca2+ release and for a retrograde enhancement of the Ca2+ current (Grabner et al., 1999; Dirksen, 2002).
To identify structural determinants of excitation-contraction (EC) coupling, oligo-nucleated skeletal myocytes in culture (myotubes) have been extensively used for functional expression of chimeric and mutant proteins (e.g., García et al., 1994; Beam and Franzini-Armstrong, 1997). In particular, myotubes of the dysgenic and the dyspedic mouse, which are deficient in 1S and RyR1, respectively (Chaudhari, 1992; Takekura et al., 1995), have been employed. Corresponding permanent cell lines, GLT (Powell et al., 1996) and 1B5 (Moore et al., 1998) have also been generated and applied (e.g., Jurkat-Rott et al., 1998; Flucher et al., 2000; Fessenden et al., 2000).
The analysis of voltage-controlled Ca2+ release in myotubes lags behind studies of mature muscle fibers that provided most of the available information on this process (for reviews see e.g., Schneider, 1994; Baylor and Hollingworth, 2000; Ríos and Brum, 2002). Clearly, the kinetics of the release process in myotubes has to be studied in detail to fully evaluate alterations in EC coupling caused by the expression of mutant proteins. In the present investigation we carried out whole-cell patch-clamp experiments to control the membrane potential of myotubes that contained a millimolar concentration of EGTA as the major intracellular Ca2+ buffer, and we analyzed the kinetics of fura-2 fluorescence transients induced by step depolarization. Most of our experiments were conducted on myotubes derived from the C2C12 cell line. This diploid continuous cell line was established by subcloning the C2 cell line (Blau et al., 1983) originally derived from primary-cultured leg muscles of a C3H mouse (Yaffe and Saxel, 1977). It has become a frequently used model system to study skeletal muscle-specific events in differentiation and function.
The high concentration of intracellular EGTA used in our experiments ensures that most of the mobilized Ca2+ is bound to EGTA. Quantification of the time course of the Ca2+ mobilization (Ca2+ input flux), therefore, depends on the knowledge of the dynamic behavior of EGTA in the cell. Song et al. (1998) conducted a kinetic analysis on EGTA-loaded heart cells that allowed them to experimentally derive the apparent dissociation rate constant of EGTA (koff,EGTA) in the cell. Here we applied a modified version of this method that permits to estimate both kon,EGTA and koff,EGTA. In addition we determined the kinetic behavior of fura-2 for our experimental conditions. These kinetic parameters were used to derive the time course of total Ca2+ input flux to the myoplasm during pulse depolarization.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
80% confluency was reached) cells were detached by trypsin treatment (0.25%) in Ca2+- and Mg2+-free phosphate-buffered saline and reseeded after dilution (1:5 or 1:10) in fresh growth medium. To obtain myotubes, collagen-coated flasks were used and FCS in the medium was replaced by 2% horse serum 8–24 h after seeding. After one to two weeks in the differentiation medium, myotubes were detached by trypsin and seeded onto coverslips (22 x 22 mm, 0.17 mm thickness; Menzel-Gläser, Braunschweig, Germany) in 35-mm petri dishes. The coverslips were carbon-coated and subsequently ionized in a vacuum evaporation system (Model 306, BOC Edwards, Crawley, UK). To stabilize the carbon layer and for sterilization, the coverslips were heated at 185°C for 8 h. The carbon-coated sterile coverslips were covered with collagen to further improve cell adhesion.
The trypsin detachment followed by reattachment resulted in myotubes that were more compact than before the treatment and therefore better suited for patch clamping. Measurements were carried out one to two days after myotubes had been reseeded. Primary cultures of mouse myotubes were carried out as described by Ursu et al. (2001).
Dulbecco's modified Eagle's medium and trypsin were purchased from Gibco (Karlsruhe, Germany); FCS, horse serum, and phosphate-buffered saline from PAA Laboratories (Cölbe, Germany); and rat tail collagen (Typ 1, C 7661) from Sigma (Deisenhofen, Germany).
Experimental solutions
The solutions for patch-clamp measurements had the following composition (in mM): Bathing solution—140 tetraethylamonium hydroxide, 137 HCl, 10 CaCl2, 1 MgCl2, 10 HEPES, 2.5 4-aminopyridine, 0.00125 tetrodotoxin, pH adjusted to 7.4 with HCl. Pipette solution—145 CsOH, 110 HCl, 1.5 CaCl2, 10 HEPES, 15 EGTA, 4 MgATP (5.5 total Mg), 5 Na creatine phosphate, 0.2 K5-fura-2, pH adjusted to 7.2 with CsOH. 1 ml aliquots of the pipette solution (without fura-2) were stored frozen at -20°C. Solutions were filtered before use (pore size 0.25 µm).
Electrophysiology and data acquisition
Size (membrane capacitance 300 pF), compact shape, and a smooth surface were selection criteria to choose cells for the experiments. Myotubes were voltage-clamped in the whole-cell configuration with a patch-clamp amplifier (LMPC, List-Medical, Darmstadt-Eberstadt, Germany) with extended range of capacitance compensation (1000 pF). Pipettes were pulled from borosilicate glass (GC150TF-10, Clark Electromedical Instruments, Reading, UK) and the tips were fire polished. The pipettes had resistances between 1.5 and 2 M
when filled with the artificial intracellular solution. Simultaneously measured fluorescence and membrane current records were sampled at 4 kHz using a DA-AD interface (Digidata 1200, Axon Instruments, Union City, CA, USA) connected to a Pentium computer. The experiments were carried out at room temperature (20–23°C). For data acquisition, the pCLAMP 7.0 software package (Axon Instruments) was used.
Fluorimetry
Optical measurements were carried out using a photomultiplier tube (R268, Hamamatsu) attached to the trinocular head of an inverted microscope (Axiovert 100 with Objective "Fluar" 40 x 1.30 Oil, Zeiss, Germany). The cells were loaded with the indicator dye fura-2 by diffusion from the patch pipette. Fluorescence was excited at 380 nm (for Ca2+ signals) and 360 nm (isosbestic point) using interference filters of bandwidth 14 and 10 nm, respectively (Schott Glaswerke, Mainz, Germany, and Andover Corp., Salem, NH, USA) and was measured at 515 nm (bandwidth 30 nm, Zeiss, Oberkochen, Germany). Changes in free Ca2+ concentration were determined using the background-corrected ratio R (see Results) of the fluorescence signals (R = F380/F360) at the two excitation wavelengths and taking into account fura-2 kinetics according to Eq. 1:
 | (1) |
200 µm) of the calibration solution were applied from a patch pipette onto carbon-coated coverslips covered with a high viscosity silicone grease (Baysilone, Bayer, Leverkusen, Germany), and fluorescence was recorded with our microscope fluorimeter from areas of 50 µm x 50 µm. At each Ca2+ concentration, F380 was proportional to F360 when changing the droplet size, indicating that inner filter effects were insignificant. The ratio R was obtained from the slope of the F380 versus F360 plot determined by linear regression. R, plotted as a function of free Ca2+ concentration, was fitted with Eq. 2 and gave the following estimates for Rmin, Rmax, and KD,Fura: 2.69 ± 0.02, 0.77 ± 0.02, and 276 ± 22 nM, respectively.
 | (2) |
In vivo estimates of Rmax and Rmin were obtained as follows: Ratio values for full dye saturation in the cell (Rmax) were determined by perfusing myotubes with intracellular solutions of elevated Ca2+ concentration (2.7 mM free) and applying additional electrical stimuli to release stored Ca2+. Rmax obtained in this way was 0.68 ± 0.03 (n = 4). Rmin was calculated using estimates of the intracellular fluorescence ratio values R0 at the beginning of the experiment (see Results). The mean value of R0 in 62 cells was 2.69 ± 0.02. Assuming that the intracellular concentration during perfusion was identical to the buffered-free Ca2+ concentration in the pipette solution (20 nM) and using the KD,Fura value determined in vitro (276 nM), we calculated Rmin according to Eq. 2. The result was 2.84 ± 0.02. A comparison shows that these Rmax and Rmin values differ only slightly from the ones determined in vitro. The in vivo values were used in the analysis of measured fluorescence signals.
Volume-capacitance ratios (VC) were determined in voltage-clamped C2C12 myotubes that were loaded with the indicator fluo-3 and imaged with a confocal laser scanning microscope (Radiance 2000, Bio-Rad, Hemel Hempstead, UK) using the 488 nm Argon laser line.
Numerical analysis
General analysis and nonlinear curve fitting were carried out using Excel (Microsoft) and Origin (OriginLab, Northampton, MA, USA). Free concentrations of Mg2+ and Ca2+ were calculated using the program CalcV22 (Föhr et al., 1993). Simulations and numerical calculations of the total flux of Ca2+ into the myoplasm (Ca2+ input flux) from fluorescence measurements were carried out using a program written in Delphi (Borland, Scotts Valley, CA, USA) assuming the presence of fura-2 (0.2 mM), EGTA (15 mM), and optionally troponin C with 0.24 mM of fast Ca2+-specific T sites and 0.24 mM of slow Ca2+ Mg2+ sites with parvalbumin-type behavior (P sites) rate constants from Baylor and Hollingworth, 1998; differential equations, see Baylor et al., 1983, and Brum et al., 1988. For parameter values used in the calculations, see Results (Fig. 6 legend).
|
; Ratzlaff, 1987) of fixed bandwidth (5 ms) was normally used. Three successive filter runs were applied on each trace leading to a better elimination of high frequencies than just one run at higher bandwidth. In addition, a second filter algorithm was designed to automatically vary the bandwidth by estimating the local signal to noise relation. The filter uses a variable bandwidth kernel estimator as described in Müller and Stadtmüller (1987). However, instead of varying a globally optimized bandwidth (Rice, 1984; Timmer et al., 1998), an optimal local bandwidth was chosen by limiting the local bias of the filtered signal. To estimate the time course of the signal, its first and its second time derivative, we used kernels that were calculated by multiplying the cosine function plus offset with a constant, a linear, and a parabolic function, respectively. Commonly used polynomial kernels (boxcar and higher order as described in Ratzlaff, 1987; Gasser et al., 1985) showed similar results. This filter performed well on very noisy records due to its ability to track fast signal components and to strongly smooth regions with slow signal alterations (see Fig. 8 E and Fig. 9, F–H).
|
|
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
|
Fig. 2 A shows our experimental protocol to measure depolarization-induced Ca2+ signals. The cell was illuminated sequentially by the two excitation wavelengths 360 and 380 nm (horizontal bars, bottom). During the 380-nm interval, the cell was activated by a voltage clamp depolarization (a) that elicited Ca2+ inward current (b) and a fluorescence decrease resulting from the increase in intracellular Ca2+ concentration (c). A slow decline in fluorescence due to dye bleaching could be observed in most experiments. Even though this decline was generally small, it can influence the kinetic analysis described below. To correct for the loss of dye resulting from the irradiation, a single exponential function with time constant 
and end value 0 was fitted to the baseline preceding the voltage pulse of the background-corrected 380-nm record. For the ratio calculation during the 380-nm illumination interval, a virtual 360-nm record was calculated as a scaled-down version of the same exponential function. The scaling factor was F360/F380 determined using the linearly extrapolated background-corrected intensities (open and closed circle) at the time indicated by the arrow. Fig. 2 B demonstrates the procedure for a measurement containing a 100-ms pulse from -90 to +20 mV (a). Panel b shows the inward current and panel c the 380-nm signal before (left) and after (right) correction for bleaching.
|
). The advantage of introducing an extrinsic buffer of high concentration is that the determination of Ca2+ input flux becomes largely independent of intrinsic Ca2+ buffering as shown by Ríos and co-workers in their work on frog skeletal muscle fibers (González and Ríos, 1993, see also Pape et al., 1995; Struk et al., 1998). It also prevents contractile activation and movement artifacts. In the present study we used 15 mM EGTA in the pipette solution. When EGTA dominates intracellular binding, free Ca2+ exhibits a time course that shows kinetic characteristics of the Ca2+ input flux (Ríos and Pizarro, 1991; Song et al., 1998). A calculation by Song et al. (1998) showed that the change in free Ca2+ (f) in a cell loaded with a sufficiently high EGTA concentration is approximately the sum of a term fr proportional to the Ca2+ input flux and a term fs proportional to its time integral (f = fr + fs and 
). This group demonstrated that the proportionality factor of the integral term, which can be determined from the fluorescence measurements, allows the calculation of the EGTA dissociation rate constant koff,EGTA in the cell according to Eq. 3:
 | (3) |
Fig. 3 shows the procedure to derive from our experimental records. For an approximate first determination of , we converted the fluorescence ratio signal (Fig. 3 B) to linear Ca2+ (CaL2+) using the steady-state version of Eq. 1 (i.e., with dR/dt = 0). The CaL2+ record (Fig. 3 C) therefore is a low-pass-filtered version of free Ca2+. However, as will be shown below (Fig. 4), using CaL2+ instead of free Ca2+ makes only a small difference in the determination of .
|
|
, we consider F = [CaL2+] - [Ca2+]0 as the sum of a function Fr and a term Fs, which is the product of and the running time integral of Fr (i.e., F = Fr + Fs; 
). was determined by fitting a monotonously rising integral term Fs to the pedestal level at the end of the transient F. To achieve this, the integral equation 
was numerically solved for Fr with as a free parameter. was altered by iteration to ensure that Fr deviated minimally from zero in the interval 100–300 ms after the end of the pulse. After convergence of the fit, Fs resulted as the difference between F and the best-fit function Fr (Fig. 3 C).
Song et al. (1998) used to calculate an apparent koff,EGTA according to Eq. 3 by assuming intracellular values for [Ca2+ ]0 and KD,EGTA. In the present study, the use of a ratiometric dye permitted to study the [Ca2+ ]0 dependence of and to determine both rate constants koff,EGTA and kon,EGTA by linear least squares fitting using Eq. 4, which is identical to Eq. 3.
 | (4) |
To check the reliability of the method, we applied a model simulation and investigated whether the EGTA rate constants used in the simulation can be correctly determined. This is demonstrated in Fig. 4.
Artificial Ca2+ flux curves were constructed using a combination of rectangular pulses (Fig. 4 A) and were used as inputs to a system containing 200 µM fura-2 and 15 mM EGTA. In the simulations, short (10 ms) prepulses of different amplitudes generated different baseline levels of free Ca2+ concentration. Each prepulse was followed by the same test pulse paradigm consisting of a short initial pulse (10 ms) of relatively large amplitude (10 µM ms-1) and a four-times smaller adjacent step of 90-ms duration. The phasic shape of the input flux signal was necessary to generate artificial fluorescence signals similar to the ones observed in the experiments (see Fig. 5 A). The off-rate constant of fura-2 was assumed to be 30 s-1 (close to the value of 26 s-1 reported by García and Schneider (1993) for rat fibers. KD,Fura was assumed to be 276 nM as determined in the in vitro calibration. For EGTA we used kon,EGTA = 50 µM-1 s-1 and koff,EGTA = 2.5 s-1, i.e., values close to the ones determined in Fig. 3. Fig. 4 B shows the calculated linear Ca2+ signals [CaL2+] (no correction for dye kinetics) and Fig. 4 C the corresponding corrected free Ca2+ transients (thick lines, (1)). The thin lines (2) in Fig. 4, B and C, show the best-fit integral components Fs and fs, respectively (see Fig. 3 and Song et al., 1998).
|
for all simulated records when setting KD,EGTA in Eq. 3 to the true value used in the simulation. The and koff,EGTA values (in s-1) obtained in the fit were 2.87, 5.38, 13.6 and 2.66, 2.63, 2.69, respectively, for each of the three records shown in Fig. 4 B, and 2.73, 5.24, 13.4 and 2.53, 2.56, 2.67, respectively, for the free Ca2+ records shown in Fig. 4 C. The estimates of koff,EGTA are slightly above but very close to the true value (Fig. 4 D, horizontal line).
Fig. 4 E shows the result of determining both EGTA rate constants from the simulated data by using the [Ca2+]0 dependence of . The continuous line in the figure was calculated using the true parameter values. Deviations from the line became noticeable only for [Ca2+]0 values larger than 250 nM. In the range of [Ca2+]0 below 100 nM (inset) that was observed in the experiments of Fig. 3, the linear fit produced values very close to the ones used in the simulation: koff,EGTA = 2.65 s-1 (vs. 2.5 s-1) and kon,EGTA = 52.3 µM-1 s-1 (vs. 50 µM-1 s-1). The parameter estimates were essentially independent of the correction for dye kinetics (filled circles: CaL2+; open circles: free Ca2+). The values were koff,EGTA = 2.51 s-1 and kon,EGTA = 52.3 µM-1 s-1 when the free Ca2+ records were analyzed instead of [CaL2+](t).
In further simulations, troponin C as an additional intrinsic buffer was added to the reaction scheme, exhibiting fast Ca2+-specific (T type) and slow parvalbumin-like (P type) binding sites of 240 µM each (rate constants according to Baylor and Hollingworth, 1998, values; see legend of Fig. 6 B). The estimated values of the rate constants were unchanged, showing that physiological intrinsic Ca2+ binding sites are unlikely to contribute significantly to the results under our conditions.
Determining indicator dye kinetics in the cell
Fura-2 Ca2+ measurements include low-pass-filtering due to the noninstantaneous binding of Ca2+ to the dye. The deconvolution to determine the true time course of rapid changes in free intracellular Ca2+ requires knowledge of the dissociation rate constant of the indicator (Klein et al., 1988). We applied an approach previously described for frog muscle fibers (Struk et al., 1998) to approximately determine the fura-2 dissociation rate constant koff,Fura in the cytoplasm of myotubes.
Under the conditions used here (high EGTA concentration), free [Ca2+] falls abruptly from the level reached at the end of the depolarization to a lower almost steady level (González and Ríos, 1993; Struk et al., 1998) due to the rapid termination of the Ca2+ flux. The almost stepwise change is demonstrated in the numerical simulation of Fig. 4 C. Here input flux is terminated instantaneously, and consequently free Ca2+ shows a comparably rapid drop from the level reached at the end of the pulse to a new almost steady value (pedestal value). The fura-2 ratio signals (Fig. 5 A), on the other hand, decay with a single exponential time course that is determined by the fura-2 dissociation kinetics. The time constant of the exponential decay of dye-bound Ca2+, given a stepwise change of free Ca2+, depends on the free Ca2+ concentration (its level during the pedestal component) and the rate constants of the indicator (Eq. 5):
 | (5) |
Eq. 5 can be rewritten in the form of Eq. 6:
 | (6) |
and a final value R
. The fit is demonstrated in Fig. 5 A for the ratio signals corresponding to the three simulated free Ca2+ records of Fig. 4 C. The free Ca2+ value corresponding to R is a very good approximation of the almost constant free Ca2+ level after the end of repolarization (see Fig. 4 C). If Rmin is known, koff,Fura and Rmax can be determined by fitting a line to the plot of versus R as shown in Fig. 5 B for the complete set of ratio records generated in the simulation of Fig. 4. The examples shown in Fig. 5 A are indicated by the filled circles. The linear regression gave best-fit values of koff,Fura and Rmax very close to the ones used in the simulations: 29.0 s-1 (vs. 30 s-1) and 0.63 (vs. 0.68), respectively. When Rmin and Rmax were both set to their true values, koff,Fura was determined as 28.9 s-1.
Fig. 5 C shows that the decay of the ratio signal in our experiments could in fact be fitted by a single exponential (plus a very slowly sloping line). Fig. 5 D summarizes the result from a series of background- and bleaching-corrected records of 11 cells (indicated by different symbols) that covered a sufficient range of basal free Ca2+ (usually caused by appropriate prepulses). As in Fig. 5 B, the constant of the exponential component is plotted as a function of the end level R (after correction for the slow decay). The dependence could be fitted by a straight line according to Eq. 6. Values of the off-rate constant of fura-2 (koff,Fura) and the steady-state calibration parameter Rmax were returned by the least squares fit. Rmin was set to 2.84 (see Methods). The best-fit parameter values obtained in this analysis were 46.4 ± 1.2 s-1 for koff,Fura and 0.47 ± 0.13 for Rmax (n = 44). We consider the estimate for Rmax obtained in this fit less reliable than the in vivo determination of Rmax because of its large standard error and did not further use it for calculations. When setting both Rmin and Rmax to the values obtained in the in vivo calibration (2.84 and 0.68, respectively), a value for koff,Fura of 45.2 ± 0.9 s-1 was obtained that differed only slightly from the previous estimate. We used this value to calculate the time course of free myoplasmic Ca2+ from the fluorescence ratio signals according to Eq. 1 (see Methods).
Determination of free Ca2+ and Ca2+ input flux
Fig. 6 A demonstrates the kinetic deconvolution (using Eq. 1) for the record shown in Fig. 3 B. The calculation generates an initial peak in free myoplasmic Ca2+ that was not present in the indicator signal because the kinetics of Ca2+ binding to fura-2 is too slow. Its presence is only evident from the steeper slope at the beginning of the fluorescence trace. The free Ca2+ records obtained after the deconvolution were subjected to the same analysis as shown in Fig. 3. The integral component, here labeled fs according to Song et al. (1998), and the result of subtracting it from the free Ca2+ trace (fr, thin line) are shown in Fig. 6 A.
As in Fig. 3, a scaling factor resulted from fitting the integral component fs to the pedestal of f. Fig. 6 C plots as a function of the baseline Ca2+ concentration for the same set of records as shown in Fig. 3 D but now after correction for dye kinetics. The linear fit to the data resulted in values of 2.12 ± 0.28 s-1 and 55.8 ± 7.3 µM-1 s-1 for koff,EGTA and kon,EGTA, respectively. These values were finally used to determine the time course of EGTA-bound Ca2+ for the depolarization-induced Ca2+ recordings.
To determine the time course of the concentration of the total Ca2+ supplied to the myoplasmic space during the depolarization, bound and free Ca2+ concentration were summed. The input Ca2+ flux was calculated by taking the time derivative of the total Ca2+ change (Fig. 6 B). To obtain information on the possible contribution of intrinsic Ca2+ binding, we included in addition to fura-2 and EGTA also T and P sites originating from troponin C assuming concentration and properties as reported in the literature for mature muscle fibers (see Fig. 6 legend). The calculation showed that the troponin C components are negligibly small in comparison to the Ca2+ bound to EGTA. Correspondingly the resulting input Ca2+ fluxes with and without troponin C were essentially indistinguishable.
The phasic-tonic time course of the input flux seen in Fig. 6 B was found in most recordings, but Ca2+ transients and calculated input flux traces obtained in different myotubes showed considerable individual variability. We therefore averaged results of individual measurements that were obtained in different cells (Fig. 7). Panel b in Fig. 7 A (thick line) shows the mean fractional indicator occupancy derived from fluorescence ratio signals of 16 C2C12 cells when applying a voltage pulse of 100-ms duration to +20 mV (a). The thin lines in this and other panels indicate the standard errors of the mean. Panel c shows the mean of the individually determined input flux signals for the same group of measurements. On the other hand, panel d shows the input flux determined from the averaged ratio signal of panel b. Both approaches led to very similar results. Because calculating flux from the mean ratio signal is considerably less time-consuming than carrying out the individual calculations, it is convenient to use this calculation to obtain a rapid overview. Panel e of Fig. 7 A shows the corresponding averaged L-type current density and its standard error.
|
In the analysis of the fluorescence signals, we applied the same set of parameters as determined for C2C12 myotubes. Only Rmin and Rmax were different, because the two groups of measurements were done on different setups. The baseline Ca2+ levels were significantly higher and the peak amplitudes of the Ca2+ input flux were significantly smaller in the primary cultured myotubes (see Fig. 7 legend). We cannot rule out a functional connection between these two differences. For instance, the elevated Ca2+ level may have caused a partial inactivation of Ca2+ release. The differences could, however, also be related to the fact that the C2C12 group contained larger cells (mean capacitance was 317 ± 22 pF in C2C12 compared to 107 ± 22 pF in primary myotubes). On the other hand, the general time course of both Ca2+ inward currents and input flux signals were not different in the two preparations.
Contribution of the Ca2+ inward current
The signals shown in Fig. 7 (panels c and d) are estimates of the total rate of Ca2+ entry into the myoplasm from both extracellular and intracellular sources. The component of the total Ca2+ input flux that enters the cell from the outside (Ca2+ entry flux) is proportional to the Ca2+ inward current density iCa (referred to linear membrane capacitance) and can therefore be calculated using Eq. 7 (assuming that the current is only carried by Ca2+ ions).
 | (7) |
To assess VC, we measured the cell volume with a confocal laser scanning microscope (see Methods) and simultaneously determined the capacitance by whole-cell recording. Capacitance proved to be proportional to volume (Fig. 8 B) as previously shown by Satoh et al. (1996) for cardiac myocytes. With the value of VC = 0.23 l F-1 determined by linear regression and setting fV to 1, we calculated the flux shown in Fig. 8 D using the mean Ca2+ current record of Fig. 7 A (panel e). Fig. 8 C shows for comparison the mean Ca2+ input flux of Fig. 7 A (panel c) displayed at a 10 times smaller scale than the trace in panel D. Comparing panel C and D, the mean amplitude ratio (average of values between 25 and 75 ms after onset of the pulse) was 79.3 ± 11.9 (n = 16). The result indicates that Ca2+ inward current makes only a small contribution to the Ca2+ input flux. Therefore, it can be concluded that under these experimental conditions the input flux consists mainly of Ca2+ release from the sarcoplasmic reticulum (SR). In some experiments (six in a total of 26 that were analyzed in this investigation), the recorded Ca2+ transients and consequently the calculated Ca2+ input fluxes were considerably smaller than in the cases discussed so far. Fig. 8 E shows one example. Not only is the flux amplitude considerably smaller than the one of the record in Fig. 8 C, also its time course is strikingly different. It rises slowly and shows a sharp peak at the end of the pulse and resembles the flux derived from the simultaneously measured Ca2+ inward current (Fig. 8 F). Apparently, in these cases internal Ca2+ release is essentially absent and the dye senses only the Ca2+ entering the myoplasm from the extracellular space.
Usually we focused on measurements with large Ca2+ transients and discontinued experiments that showed very small fluorescence transients. However, in cells without internal release, the measured Ca2+ inward current can serve as a control to check the fidelity of the Ca2+ flux calculation. Fig. 9 investigates the question of how well the flux generated by the Ca2+ inward current can be determined from fluorescence measurements. The calcium inward current shown in Fig. 9 A was first converted to flux by Eq. 7 and then scaled to reach the same amplitude during the pulse as the flux (panel F) derived from the simultaneously measured fluorescence signal (panel E). It was then used as input to a model containing EGTA and fura-2 as described in conjunction with Fig. 4 (parameters of Fig. 6 B). This generated the artificial fluorescence ratio record of panel C. Subsequently noise taken from a fluorescence recording at a subthreshold depolarization of the same experiment was added and the analysis procedure was applied to the noisy record (panel D) to derive the underlying input flux. The result is shown in panel G. Fig. 9 H shows the average of five records like the one in panel G, which were generated in the same way but using different experimental noise traces. Traces G and H are very similar to trace F, showing that the differences in time course between the flux calculated from the current and the one determined from the fluorescence signal result exclusively from the digital filtering necessary to cope with the higher noise level of the optical recording. Fig. 9 demonstrates that the shape of the optically recorded Ca2+ signal is in fact fully compatible with an underlying input flux that has the time course of the measured inward current.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
). Both approaches provided valuable complementary information on the function of the Ca2+ mobilization mechanism in skeletal muscle. Because protein expression in cultured myotubes has become a central tool for structure-function analysis in EC coupling (Beam and Franzini-Armstrong, 1997), it is important to pursue similar investigations in these cells. In the present study we analyzed global Ca2+ signals in voltage-clamped mouse muscle-derived myotubes with the goal to quantify the underlying Ca2+ input flux.
First attempts to derive the Ca2+ input flux in myotubes have been made in our laboratory (Dietze et al., 1998; Dietze et al., 2000; Ursu et al., 2001) by using the approach of Baylor et al. (1983). This method requires assumptions for the rate constants of intracellular Ca2+ binding sites to quantify the bound fraction of the mobilized Ca2+. The present approach differs from the previous one by the direct experimental determination of kinetic parameters for the dominating intracellular Ca2+ buffer EGTA and the indicator dye fura-2 that were introduced into the cells. The procedure is derived from a method described by Song et al. (1998), who used millimolar concentrations of intracellular EGTA and a low affinity fast indicator dye to determine "Ca2+ spikes" in isolated heart cells. Because fura-2 does not bind calcium instantaneously, our experiments required a numerical deconvolution to determine the kinetics of free Ca2+. On the other hand, the higher affinity of the indicator facilitated the measurement of the small Ca2+ transients in the strongly EGTA-buffered myoplasm and served to better resolve the pedestal Ca2+ components (at the end of the pulses) that were investigated in the analysis. Because fura-2 permitted ratiometric measurements, changes in basal free Ca2+ concentration could be separated from alterations in background fluorescence or indicator concentration changes. The ratiometric Ca2+ determination allowed us to estimate both rate constants of EGTA.
Determination of EGTA rate constants in the cell
The Ca2+ input flux calculation requires the quantification of all relevant Ca2+ binding compartments in the cell. At the very high concentration of EGTA in the pipette solution, it is unlikely that components other than EGTA make a significant additional contribution to intracellular Ca2+ binding (see also Pape et al., 1995; Song et al., 1998). Therefore, the rate constants and the concentration of EGTA in the cell are the essential determinants for the estimation of the Ca2+ input flux. If other slowly Ca2+ binding mechanisms made a significant contribution, they would be lumped together with EGTA in the analysis.
As pointed out by Song et al. (1998), the rate constant determination in their procedure becomes independent of the EGTA concentration at sufficiently high concentrations. Our pipette concentration of EGTA was 15 mM (compared to 4 mM in the previous experiments on heart cells), thus the conditions for applying the approximate equations derived by Song and co-workers were even better fulfilled. Conducting the analysis on simulated Ca2+ transients in a system with predefined parameters showed that the rate constants could in fact be determined quite precisely as long as signals were small enough not to saturate the dye appreciably (Fig. 4). It should be noted that the analysis generates "apparent" EGTA rate constants, which depend on certain assumptions, in particular on the value assumed for the dissociation constant of the indicator. The KD,Fura in the myotubes may be higher than found in vitro due to the binding of the indicator to intracellular proteins (Konishi et al., 1988). To test the effect on the calculation results, we increased KD,Fura to 500 nM. This is the mean of values listed by Pape et al. (1993). Table 1 compares the results obtained with the two different assumptions. The estimated apparent KD,EGTA constant increased more than twofold (mainly due to a change in the on rate constant). The calculated input flux, on the other hand, was much less affected by the change. It increased by only 25% and did not alter its time course.
|
; Naraghi, 1997). We have no satisfying explanation for the difference, but noticed that other researchers, who likewise worked with high concentrations of EGTA in mammalian muscle cells and fitted EGTA rate constants to their measured Ca2+ transients, came to similar results: Shirokova et al. (1996) and Song et al. (1998) report koff,EGTA values of 3.4 s-1 (and higher) and 3 s-1, respectively.
An advantage of the use of EGTA is that it binds most of the total released Ca2+ without significantly disturbing the local Ca2+ feedback mechanisms (Pape et al., 1995; Jong et al., 1995). This is a consequence of its slow kinetics and contrasts to the effect of faster buffers like BAPTA, which was found to eliminate the initial Ca2+ release peak in frog muscle fibers (Csernoch et al., 1993). A progress would be the use of a high concentration of a Ca2+ indicator with the binding properties of EGTA. This would allow to determine the Ca2+ input flux directly. Using the indicator Quin2, this approach was tried but not further pursued (Dey et al., 1996).
Determination of the fura off-rate constant in the cell
The deconvolution of the fura-2 signal to derive free Ca2+ depends on the knowledge of the indicator off-rate constant under cytoplasmic conditions (Eq. 1), which differs from in vitro measurements. Previous determinations of fura-2 rate constants in muscle cells used a dual dye approach. Klein et al. (1988) and Baylor and Hollingworth (1988) fitted Ca2+ records that were measured with the fast metallochromic indicator antipyrylazo III to simultaneously measured fura-2 records. A drawback of this method is that the calibration parameters of the fast indicator in the cell are likewise different from those obtained in vitro. To determine koff,Fura in our experiments, we used an alternative procedure that does not require measurements with a second dye. We made use of the fact, supported by simulations (this study) and measurements in frog fibers (Struk et al., 1998), that free calcium in the presence of high EGTA concentrations shows a step-like change when the input flux is rapidly terminated by repolarization (Fig. 4 C). Under these conditions, the Ca2+ occupancy of the indicator decays exponentially to a new equilibrium and the off-rate constant can be determined from the Ca2+ dependence of the time constant. The resulting value of 45.2 s-1 in this study was intermediate between intracellular determinations in mature muscle fibers (García and Schneider, 1993: 26 s-1 at 14–17°C) and in vitro determinations using temperature-jump relaxation or stopped-flow methods: 97 s-1 (Kao and Tsien, 1988) and 84 s-1 (Jackson et al., 1987).
Components of Ca2+ input flux
Ca2+ input flux consists of two main components: intracellular Ca2+ release and entry from the extracellular space. Because the voltage-activated Ca2+ current is proportional to the Ca2+ entry flux and can be recorded with high time resolution, it provides an independent control for our quantification of the input flux of Ca2+. The input flux analysis in the cells with the smallest fluorescence changes led to results that showed characteristics of the electrically recorded inward current. Release from the SR is probably not functional in these myotubes and the recorded Ca2+ signal results from Ca2+ entering the cell with the L-type current. The reason for lack of internal release in part of the cells is unclear. If the expression of the ryanodine receptors was greatly reduced, L-type currents would be likewise reduced in amplitude due to compromised retrograde coupling. This was not observed. Probably gating of the ryanodine receptors is altered or loading of the SR not functional. In this context it should be mentioned that cells with apparently large Ca2+ release showed a variable degree of rundown (usually confined to the Ca2+ signals).
The electrically determined entry flux and the calculated input flux derived from the fluorescence transient shown in Fig. 8 differed in absolute scale. On average, when comparing the slowly changing sections of the signals (interval between 25 and 75 ms after pulse-on), the calculated input flux was larger by a factor of 10.5 ± 1.7 (n = 8). In these calculations, it was assumed that the effective cytoplasmic EGTA concentration is identical to the pipette concentration and that the whole cell volume is rapidly accessible to Ca2+ entering from the extracellular space (fv = 1 in Eq. 7). Both assumptions are certainly overestimates. The EGTA concentration is probably smaller in the cell than in the pipette, but, given the loading velocity of the myotubes shown in Fig. 1 and assuming that EGTA diffuses comparably rapidly as fura-2, we guess that it is normally at least half the pipette value. Assuming that only 50% of the total cell volume is accessible for Ca2+ from the extracellular space (fV = 0.5 in Eq. 7 as assumed by Sipido and Wier, 1991, for cardiac myocytes) the scale difference comes down to a factor of 2.5, which brings the differently obtained estimates quite close together.
The time course of the current-derived flux and the optically measured flux in cells without intracellular release was similar but not identical. Fig. 9 demonstrates that, due to noise in the optical signals and the necessary digital filtering, the sharp peak in the flux arising from the tail current was underestimated by a factor of 50% compared with the slowly changing signal components. However, these small signals with their low signal-to-noise ratio present particularly unfavorable conditions for the analysis. The signal-to-noise ratio is higher in the larger fluorescence signals. Thus, equally fast flux components underlying larger Ca2+ transients are likely to be less affected by the digital filtering and therefore better resolved.
The input flux derived from the larger fluorescence signals showed a completely different time course exhibiting a leading peak and a decline to a lower, slowly declining level. It originates mainly from intracellular Ca2+ release. C2C12 myotubes contain voltage- and ligand-activatable Ca2+ pools that form part of the same intracellular Ca2+ store (Lorenzon et al., 2000). This notion was based on experiments in which depolarizing solutions with elevated K+ concentration were applied and combined with caffeine stimulation. K+ responses could be suppressed by conditioning caffeine application and vice versa. The fact that depolarization elicited Ca2+ transients in the absence of extracellular Ca2+ shows that C2C12 myotubes develop skeletal muscle-type EC coupling. Whereas Györke and Györke (1996) reported that K+, like caffeine, released Ca2+ at discrete loci, Lorenzon et al. (2000) described a rapid, spatially synchronous elevation of intracellular Ca2+ when the cells were superfused with Ca2+-free depolarizing solutions. In voltage-clamped primary-cultured mouse myotubes, Shirokova et al. (1999) observed inhomogeneous Ca2+ release evidenced by narrow zones that were not responsive to depolarization. Even though the intracellular organization of the Ca2+ stores in myotubes is obviously more heterogeneous and less orderly than in mature muscle fibers, the previous and our present experimental evidence strongly suggests that C2C12 myotubes, like primary-cultured myotubes, share great similarities in EC coupling with mature fibers. Our experiments demonstrate that the characteristics of the depolarization-activated Ca2+ flux in C2C12 myotubes and mouse primary-cultured myotubes are very similar. The majority of measurements from both preparations showed phasic-tonic Ca2+ input flux signals that resemble the Ca2+ release rate records measured in mature mammalian skeletal muscle fibers under similar intracellular conditions (Shirokova et al., 1996). According to current models, the initial peak of the Ca2+ release flux has strong contributions from a Ca2+-dependent positive feedback mechanism (calcium-induced calcium release) whereas the release activity that follows the peak is predominantly voltage-gated (Csernoch et al., 1993; Stern et al., 1997). Further studies are required to verify if these concepts can be also applied in unmodified form to EC coupling in myotubes.
In conclusion, the results indicate that the present procedure leads to a reliable reconstruction of the time course of global Ca2+ input flux in myotubes. As in primary cultured myotubes, the flux estimated in C2C12 myotubes shows great similarities to the voltage-activated Ca2+ release flux in muscle fibers. Normally, Ca2+ inward current makes only a small contribution to the total flux signal even at the high concentration of 10 mM Ca2+ used in our external solution. The effect of the inward current will, however, become discernible when depletion of the SR takes place or when the Ca2+ tail current measured on repolarization is large or slowed down. The high-EGTA method is a straightforward way to gain insight into the dynamics of cellular Ca2+ fluxes and should therefore be of value for structure-function analyses in myotube preparations used for expression of EC coupling proteins.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
The work was supported by a grant of the Deutsche Forschungsgemeinschaft to W. Melzer (ME-713/10-2) and a scholarship of the Graduiertenkolleg 182 to R.P. Schuhmeier.
Submitted on August 26, 2002; accepted for publication October 16, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Baylor, S. M., and S. Hollingworth. 1988. Fura-2 calcium transients in frog skeletal muscle fibres. J. Physiol. 403:151–192.
Baylor, S. M., and S. Hollingworth. 1998. Model of sarcomeric Ca2+ movements, including ATP Ca2+ binding and diffusion, during activation of frog skeletal muscle. J. Gen. Physiol. 112:297–316.
