GLOSSARY

TOP ABSTRACT GLOSSARY INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

Abbreviations

RSP	Restricted subsarcolemmal space;
MYOF	Myofibrillar space;
SR	Sarcoplasmic reticulum;
RyR	Ryanodine receptor;
DHPR	Dihydropyridine receptor (L-type Ca2+ channel);
ATP	Adenosine triphosphate.

Geometric parameters and constants

R	Cell radius;
L	Cell length;
Cm	Cell capacitance;
Vcell	Cell volume;
dRSP	Restricted space thickness;
Vrd	Thin boundary volume between extracellular space and RSP;
rd	Boundary thickness;
Vacc	Accessible volume for Ca2+ in the cell;
S	Model cell surface through which Ca2+ enters.

Concentrations and reaction parameters

[Ca2+]i	Free intracellular Ca2+ concentration;
[Ca2+]o	Extracellular Ca2+ concentration;
[Ca2+]rest	Resting Ca2+ concentration;
[TN]	Total troponin concentration (low-affinity sites);
[CAL]	Total calmodulin concentration;
[PL]	Total phospholipid concentration (low-affinity sites);
[PH]	Total phospholipid concentration (high-affinity sites);
[FLUO]	Total Fluo-3 concentration;
[ATP]	Free ATP concentration;
[CaTN]	Ca2+-troponin concentration (low-affinity sites);
[CaCAL]	Ca2+-calmodulin concentration;
[CaPL]	Ca2+-phospholipid concentration (low-affinity sites);
[CaPH]	Ca2+-phospholipid concentration (high-affinity sites);
[CaFLUO]	Ca2+-Fluo-3 concentration;
[CaATP]	Ca2+-ATP concentration;
D	Diffusion coefficient for Ca2+ in RSP;
D	Diffusion coefficient for Ca2+ in MYOF;
D	Diffusion coefficient for CaCAL in RSP;
D	Diffusion coefficient for CaCAL in MYOF;
D	Diffusion coefficient for CaFLUO in RSP;
D	Diffusion coefficient for CaFLUO in MYOF;
D	Diffusion coefficient for CaATP in MYOF;
D	Diffusion coefficient for CaATP in RSP;
k	Ca2+ on-rate constant for troponin (low-affinity sites);
k	Ca2+ off-rate constant for troponin (low-affinity sites);
k	Ca2+ on-rate constant for calmodulin;
k	Ca2+ on-rate constant for calmodulin;
k	Ca2+ off-rate constant for phospholipid (low-affinity sites);
k	Ca2+ off-rate constant for phospholipid (low-affinity sites);
k	Ca2+ on-rate constant for phospholipid (high-affinity sites);
k	Ca2+ off-rate constant for phospholipid (high-affinity sites);
k	Ca2+ on-rate constant for Fluo-3;
k	Ca2+ off-rate constant for Fluo-3;
k	Ca2+ on-rate constant for ATP;
k	Ca2+ off-rate constant for ATP;
JICa	Ca2+ flux via L-type Ca2+ channels;
F	Faraday's constant;
ICa	L-type Ca2+ current;
I0	Constant;
ai, bj	Constants;
Jexch	Ca2+ flux through Na+/Ca2+ exchanger;
Vmax,x	Maximal velocity of Na+/Ca2+ exchanger;
Km	Ca2+ concentration at half Vmax,x;
n	Hill's coefficient;
Jexl	Inward Ca2+ leak flux via plasma membrane;
Lm	Ca2+ leak constant.

	INTRODUCTION

TOP ABSTRACT GLOSSARY INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

In cardiac and skeletal muscle cells mechanical activity is controlled by a transient elevation of the intracellular Ca²⁺ concentration ([Ca²⁺]_i) (Taylor et al., 1975; Cannell et al., 1987; Niggli, 1999). Compared to other cell types, striated muscle cells are quite large. Depending on the diameter of a given muscle cell, diffusion of Ca²⁺ from the cell membrane to the proteins regulating muscle force (i.e., troponin C) would introduce unacceptable delays in the activation of contraction. Reasons for this delay are the sheer distance and the presence of stationary Ca²⁺ buffers in the cell, which tend to slow the movement of Ca²⁺ (Crank, 1975; Neher and Augustine, 1992; Jafri and Keizer, 1995; Haddock et al., 1999). Therefore, several structural and functional systems have evolved to accelerate the spread of Ca²⁺ signals in muscle cells considerably. Skeletal and most cardiac muscle cells have developed deep invaginations of the extracellular space via infoldings of the cell membrane. These so-called T-tubules form a network of extracellular space, extending deep into the cell interior and allowing the fast electrical signal (i.e., the action potential) to be carried close to the subcellular location where Ca²⁺ is needed (Bers, 2001). In addition, intracellular Ca²⁺ stores are present in most species (i.e., the sarcoplasmic reticulum (SR)). In cardiac muscle Ca²⁺ release from these stores drastically reduces the amount of Ca²⁺ that has to enter from the extracellular space, while in skeletal muscle it represents the almost exclusive source of Ca²⁺. Besides acting as an amplifier of the trigger signals, Ca²⁺ release from the SR also acts as an accelerator for the spatial spread of Ca²⁺ signals (Dawson et al., 1999). In skeletal muscle, Ca²⁺ release from the SR initially occurs via Ca²⁺ release channels (ryanodine receptors; RyRs) that are under the control of voltage sensors located in the sarcolemma (Schneider and Chandler, 1973; Rios et al., 1991). In cardiac muscle, the RyRs are controlled by the Ca²⁺-induced Ca²⁺ release mechanism (CICR; Fabiato, 1983). The trigger Ca²⁺ is provided by influx via L-type Ca²⁺ channels (DHP receptors), which represent the structural and functional equivalent of the voltage sensors present in skeletal muscle. By working together, the T-tubules and the Ca²⁺ release from the SR ensure spatially homogeneous and synchronized Ca²⁺ release throughout each cell.

In many species atrial cardiac muscle cells have no T-tubules and are thus an interesting exception to the rule (Bers, 2001). Although generally exhibiting smaller diameters than ventricular myocytes, these cells might encounter Ca²⁺ diffusion delays if SR Ca²⁺ release from the SR fails (Lipp et al., 1990; Hüser et al., 1996; Kockskämper et al., 2001). They may have developed a dense network of Ca²⁺ stores close to the sarcolemma, but also deep inside the cell, to compensate partly for the lack of T-tubules. In the absence of functional T-tubules, Ca²⁺ signals are known to spread rapidly from one SR Ca²⁺ release site to the next, giving rise to saltatoric reaction-diffusion waves (Hüser et al., 1996; Cheng et al., 1996; Keizer and Smith, 1998; Kockskämper et al., 2001).

In the present study we examined Ca²⁺ diffusion in atrial cells that had been treated with ryanodine and thapsigargin to eliminate release and uptake of Ca²⁺ by the SR. Using a combination of experimental techniques and a mathematical model, we could analyze several important spatial and temporal features of Ca²⁺ diffusion and signaling in these cells. In this context, the goal was at least threefold. The first aim was to develop a mathematical model that would quantitatively predict our experimental results on Ca²⁺ influx, Ca²⁺ buffering, and Ca²⁺ diffusion in atrial cells, when the SR was inhibited. Second, we used the model to explore the parameter space beyond the experimentally accessible limits. The third task was to use the model to examine the importance of mobile and stationary Ca²⁺ buffers and the effect of altered restricted space geometry for the Ca²⁺ signals. The restricted space is also known as the "fuzzy space" (Lederer et al., 1990) and is below the optical resolution of confocal (optical) microscopes. Inclusion of the fuzzy space was not required to model the experimental results, but allowed explorations and predictions of Ca²⁺ signals in this space. Furthermore, our model calculations suggest an important role for mobile and stationary Ca²⁺ buffers, including the Ca²⁺ indicator dye used in our experiments. The model also predicts a significant acceleration of Ca²⁺ diffusion by physiological concentrations of the low-affinity Ca²⁺ buffer ATP. Preliminary results of this work have been presented to the Biophysical Society in abstract form (Michailova et al., 1999).

	MATERIALS AND METHODS

TOP ABSTRACT GLOSSARY INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

Cell isolation and solutions

Experiments were performed on single atrial myocytes isolated enzymatically from guinea pigs (Cavia porcellus). The isolation procedure used was a modification of the method reported by Kockskämper and Glitsch (1997). Adult animals were killed by cervical dislocation, the hearts rapidly removed, and retrogradely perfused for 3 min on a Langendorff perfusion system at 37°C. The perfusing solution consisted of basic Ca²⁺-free solution (in mM: sucrose 204, NaCl 35, KCl 5.4, MgCl₂ 1, HEPES 10, pH 7.4 adjusted with NaOH) with 2 mM ethylene glycol-bis-(-aminoethylether)-N,N,N',N'-tetraacetic acid (EGTA). Enzymatic digestion was started by switching to Ca²⁺-free solution containing collagenase B (0.2 mg/ml; Boehringer Mannheim, Rotkreuz, Switzerland), protease type XIV (0.05 mg/ml; Sigma, Buchs, Switzerland) and elastase (5 µl/ml; Serva, Heidelberg, Germany). To promote the digestion of the atria the large blood vessels were ligated and the heart was immersed in an organ bath.

After 15 min the atria were minced, placed in Ca²⁺-free solution containing 1 mg/ml bovine serum albumin (BSA) to stop the digestion, and left on a rocking table at room temperature (22°C) to allow for dispersion of the tissue. During this procedure (~1-h) the cells were adapted to calcium by dropwise addition of an equal volume of cell culture medium containing 1.26 mM Ca²⁺ (M199, Gibco, Basel, Switzerland) and supplemented with 100 IU/ml penicillin, 100 µg/ml streptomycin, and 10% fetal calf serum (all from Gibco). Finally, cells were taken from the supernatant and plated onto glass coverslips placed in culture dishes. The cells were incubated overnight at 37°C and 5% CO₂ and used the following day.

Current measurements

All experiments were carried out at room temperature (22°C). A coverslip with adherent cells was assembled into a recording chamber and mounted onto the stage of an inverted microscope (Diaphot TMD, Nikon, Küsnacht, Switzerland). The cells were constantly superfused (1-2 ml/min) with extracellular solution containing (in mM) NaCl 140, KCl 5, MgCl₂ 1, CaCl₂ 1, glucose 10, HEPES 10, pH 7.4 adjusted with NaOH. Patch-clamp recording electrodes were pulled from filamented borosilicate glass capillaries (GC150F, Clark Electromedical Instruments, Pangbourne, UK) on a horizontal puller (DMZ, Zeitz Instruments, Augsburg, Germany) and filled with intracellular solution containing (in mM): CsAsp 120, NaCl 10, TEA-Cl 20, HEPES 20, MgATP 5, MgCl₂ 1, Fluo-3 0.1, pH 7.2 adjusted with CsOH. The free [Ca²⁺] calculated for this solution was 99 nM (when assuming a typical Ca²⁺ contamination of 15 µM). Pipette resistances ranged from 2 to 4 MOhm. Cells were voltage-clamped in the whole-cell configuration and held at 70 mV without correction for the liquid junction potential (Axopatch 200, Axon Instruments, Foster City, CA). The voltage was stepped to 40 mV for 50 ms to inactivate the Na⁺ current and subsequently to 0 mV for 200 ms to elicit a Ca²⁺ current. The step to 40 mV was introduced to avoid any residual Na⁺ current that would contaminate the recording of the Ca²⁺ current despite the presence of 10 µM tetrodotoxin (TTX). In addition, the Ca²⁺ current was enhanced by application of 1 µM isoproterenol.

Series resistance and membrane capacitance were compensated with the built-in compensation circuit of the amplifier. The reading on the capacitance compensation dial of the amplifier was taken as the membrane capacitance of the cell. No leak subtraction was performed. The pure Ca²⁺ current was determined off-line by subtracting the current recorded in the presence of 5 mM Cd²⁺. Currents were low-pass filtered at 5 kHz and digitized at 10 kHz using the LabView data acquisition software (National Instruments, Ennetbaden, Switzerland). Data were stored on hard disk for later analysis with the IgorPro software (WaveMetrics, Lake Oswego, OR).

Thapsigargin and TTX were purchased from Alomone Labs (Jerusalem, Israel), ryanodine from Calbiochem (La Jolla, CA), isoproterenol from Fluka (Buchs, Switzerland), and Fluo-3 (penta-potassium) from TefLabs (Austin, TX). Cells were incubated with thapsigargin and ryanodine for 30 min before each experiment to block the SR Ca²⁺ pump and the ryanodine receptor. Thapsigargin was dissolved as 1 mM stock in ethanol and used at 0.1 µM. Ryanodine was dissolved at 10 mM in distilled water and used at 10 µM concentration. Isoproterenol stock solution (10 mM) was freshly prepared before each experiment in distilled water containing 1 mM L-ascorbic acid and added at 1 µM to the extracellular solution. TTX was dissolved in distilled water, kept in aliquots at 20°C as a stock solution (10 mM) and used at 10 µM. Fluo-3 was reconstituted in distilled water to 5 mM and diluted to 0.1 mM into the pipette filling solution. Drugs were delivered to the cells through a gravity-driven rapid superfusion system.

Confocal Ca²⁺ measurements

Cells were viewed with a 40× oil-immersion objective (Fluor, N.A. = 1.3, Nikon) and loaded with Fluo-3 through the recording pipette. Fluo-3 was excited with the 488 nm line of an argon laser (model 5000, Ion Laser Technology, Salt Lake City, UT) at 50 µW intensity on the cell. The fluorescence was detected at 540 ± 15 nm with a photomultiplier tube (PMT) of a laser-scanning confocal system (MRC 1000, Bio-Rad, Glattbrugg, Switzerland) operated in the line-scan mode. The recording chamber was rotated to position the cell's width in parallel to the scan direction. The scan speed was set to 2 ms per line. Synchronization of the Ca²⁺ signal with the voltage protocol was assured by simultaneously recording a red light-emitting diode, triggered by the acquisition software, with the second PMT of the confocal system (>600 nm).

To record the Ca²⁺ influx generated by the activation of L-type Ca²⁺ channels without contamination by CICR from the SR, the cells were treated with 10 µM ryanodine and 0.1 µM thapsigargin. Amplitude and time course of Ca²⁺ signals due to Ca²⁺ influx were computed off-line using a customized version of the NIH Image software (NIH, Bethesda, MD). Different regions of interest (width = 1-2 µm) were chosen to average the temporal Ca²⁺ concentration changes near the plasmalemma or in the center of the cell. Similarly, Ca²⁺ concentration profiles across the entire width of the cell were extracted for each time point (2 ms). The spatial profiles of [Ca²⁺]_i are limited by optical diffraction, while the mathematical simulation can exhibit a much better spatial resolution. The point-spread-function of our confocal microscope was examined with fluorescent beads (diameter 100 nm) and was determined to have a full-width at half-maximal amplitude (FWHM) of 350 nm · 350 nm · 900 nm (for the x, y, and z dimension, respectively). Ca²⁺ concentration was calculated from fluorescence images using an established self-ratio calibration procedure (Cheng et al., 1993). For the calibration we assumed a K_d value for Fluo-3 of 739 nM and a resting Ca²⁺ concentration of 100 nM at the beginning of each experiment. Surface plots were generated from line-scan images using the IDL software (Research Systems, Boulder, CO). Confocal x-y images were used to calculate the surface and the volume of the cells using the NIH Image software. The accuracy of the procedure was verified by comparing the results with the values obtained using the measured membrane capacitance (assuming 1 µF capacitance per cm² of membrane).

Mathematical model

We developed a mathematical model of Ca²⁺-signaling, Ca²⁺-diffusion, and Ca²⁺-buffering inside an atrial cardiac muscle cell. The goal was to simulate and analyze Ca²⁺ events, which were recorded on the confocal microscope and, in addition, to simulate Ca²⁺ signals that are not accessible experimentally. In view of the fact that the isolated atrial myocyte has an approximately cylindrical shape (see Fig. 1 A) and lacks T-tubules (Bers, 2001; Hüser et al., 1996; Kockskämper et al., 2001) a cylindrical geometry is assumed (see Fig. 2 A).

Model cell geometry

Restricted subsarcolemmal space

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

(9)

(10)

Myofibrillar space

(11)

(12)

(13)

(14)

Ca²⁺ current, Na⁺/Ca²⁺ exchanger, and Ca²⁺ leak

1

1

1

Initial Ca²⁺ and buffer concentrations and buffer rate and dissociation constants

1

1

1

1

1

1

1

1

1

1

1

Ca²⁺ and buffer diffusion coefficients

1

1

1

1

1

1

1

1

8

	RESULTS

TOP ABSTRACT GLOSSARY INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION CONCLUSIONS REFERENCES

Experimental recordings of Ca²⁺ influx and changes of [Ca²⁺]_i

The voltage-clamp protocol elicited inward currents with the typical signatures of the L-type Ca²⁺ current (Fig. 1 D). At the onset of the first voltage step from 70 to 40 mV a small current was observable, most likely attributable to incomplete blockade of Na⁺ channels by 10 µM TTX. In all cases a Ca²⁺ signal due to Ca²⁺ influx accompanied the inward current that coincided with the second voltage step from 40 to 0 mV, and thus corresponded to the activation of the L-type Ca²⁺ current. The Ca²⁺ current and the Ca²⁺ signal resulting from the Ca²⁺ influx were both blocked by 5 mM Cd²⁺ (data not shown). Please note that there was no Ca²⁺ release from the SR in our experiments, because the cells had been pretreated with ryanodine and thapsigargin.

Fig. 1 also shows additional data used to develop and validate the mathematical model. Panel A represents a false color x-y image of cytosolic Fluo-3 fluorescence under resting conditions, which was used to determine the geometrical parameters of the cell. The bright spot marks the tip of the patch-clamp pipette. This cell had a length of 125 µm, a diameter of 15.6 µm, and a membrane capacitance of 41 pF.

View larger version (31K): [in this window] [in a new window]   FIGURE 1   Ca2+ current and Ca2+ concentration recorded from an isolated guinea pig atrial myocyte (A), loaded with 100 µM Fluo-3 by dialysis from a patch-clamp pipette. A single line (yellow line in A) was then scanned to obtain a line-scan image of fluorescence versus space and time (E). (B) The time course of [Ca2+]i was obtained for the periphery (red) and center (blue) of the cell. Panel (C) shows the Ca2+ transient averaged along the complete line-scan image. The voltage-clamp protocol and the resulting L-type Ca2+ current are illustrated in (D). The spatial profile of [Ca2+]i is shown in (F), while (G) shows a surface plot computed from the line-scan image in (E). The red, green and blue bars in (E) indicate the region from which traces were averaged. Note that the gray portions of the traces in panels B-D belong to the pre-pulse protocol and where not simulated in the model.

A typical line-scan image acquired during a voltage-clamp protocol along the entire width of the cell is shown in Fig. 1 E. The colors correspond to fluorescence ratio values (F/F₀) reflecting the changes of Ca²⁺ concentration in time (vertical dimension of the image). A clear U-shaped profile extending across the cell is visible at the beginning of the Ca²⁺ signal resulting from Ca²⁺ influx (see also Fig. 1 F). A convenient way to visualize the relationship among space, time, and Ca²⁺ concentration is provided by the surface plot (panel 1 G). In this representation it is readily appreciable that the Ca²⁺ concentration increased faster and to a higher amplitude at the edges of the cell for a given time point, whereas in the center of the cell the signal reached a similar amplitude only after a considerable delay.

In Fig. 1 B the Ca²⁺ concentration changes extracted from the periphery (red) and the center of the line-scan image (blue) are superimposed to emphasize the delay between the two signals. The time course of the average Ca²⁺ concentration (calculated by averaging all points along the line-scan) is plotted in Fig. 1 C. The duration of the rising phase (200 ms) of the Ca²⁺ signal from the edge of the cell (red trace in Fig. 1 B) matched the duration of the L-type Ca²⁺ current. In contrast, the Ca²⁺ signal recorded from the center (blue) developed more slowly and continued to rise even when the current was already terminated.

Numerical simulation of the experimental data

The first set of modeling results (Fig. 2) describes our attempt to create a simulation that quantitatively approximates the experimental data. Fig. 2, A-G are arranged in analogy to the experimental Fig. 1; Fig. 2 H was obtained by convolving the model data with a simplified confocal point-spread function. Thus Fig. 2, E, G, and H illustrate the calculated temporal and spatial Ca²⁺ concentration changes as line-scan images and as a surface plot. The simulated local Ca²⁺ signals in the center (blue) and periphery (red) are shown in Fig. 2 B. The Ca²⁺ signal in the cell periphery was calculated by averaging the Ca²⁺ concentration across the first micrometer under the membrane. This average corresponds to the experimental measurement of peripheral [Ca²⁺], which is also a spatial average due to the limited optical resolution. As expected, the convolution of the simulated data with the point-spread function eliminated the signal in the fuzzy space and also introduced some edge effects at the boundary of the line-scan. These model results illustrate that the Ca²⁺ signal in the cell periphery increases faster and has a larger amplitude than the Ca²⁺ signal in the center, which peaks with a delay of ~100 ms. The simulations also suggest that the slower and smaller Ca²⁺ transient in the center can be explained by diffusion of Ca²⁺. The Ca²⁺ influx carried by the simulated Ca²⁺ current allowed us to predict the Ca²⁺ concentration levels developed in the narrow fuzzy space (RSP) that is not accessible experimentally (Fig. 2 B, green line). Thus, the model predicts steep Ca²⁺ concentration gradients within the signals recorded experimentally from the cell periphery. The simulated U-profile of Ca²⁺ at 100 ms extracted from the line-scan image (Fig. 2 E) is shown in Fig. 2 F. Note that, in contrast to the experimentally measured signal (Fig. 1 F), the calculated Ca²⁺ concentration near the plasmalemma peaks at 244 nM above resting concentration because the model is able to predict Ca²⁺ concentrations in the narrow RSP, which cannot be resolved optically. Fig. 2 C shows a Ca²⁺ transient averaged across the entire cell corresponding to the experimentally measured signal (Fig. 1 C).

View larger version (23K): [in this window] [in a new window]   FIGURE 2   Simulating the experimental data with the computer model. Panel (A) depicts the cylindrical model cell containing a myofilament space (MYOF) surrounded by a space with restricted diffusion (RSP). The radius of the cell was 7.8 µm, the thickness of RSP was 20 nm (unless noted otherwise). The length of the cylinder was adjusted to accommodate the accessible volume of the cytosol (~50% of total cytosolic volume in this case). The time courses of the Ca2+ concentration in the center (blue line) and in the RSP (green line) are superimposed in (B). In addition, the Ca2+ profile averaged over the first micrometer under the membrane is shown in red. This approximately corresponds to confocal recordings of Ca2+ in the cell periphery. Panel (C) shows the time course of the Ca2+ concentration averaged over the entire cell. The simulated Ca2+ influx, corresponding to the L-type Ca2+ current in Fig. 1, is illustrated in (D). The Ca2+ concentration changes were also visualized as line-scans (E) and as surface plots (G). From the line-scan image the time course traces and the Ca2+ concentration profile at 100 ms (F) were extracted and colored according to the respective symbols. Panel (H) was computed by convolving a simplified confocal point-spread function with the image shown in (E). This mimicks the limited optical resolution present in experimental data (compare with Fig. 1 E).

The good agreement between the theoretical and the experimental data suggested that the model, as implemented, correctly described the subcellular Ca²⁺ signaling in atrial myocytes. Furthermore, these quantitatively correct results provided an opportunity to examine and better understand how different model parameters beyond the experimentally accessible limits might influence the spatial and temporal characteristics of the Ca²⁺ transients.

Exploring the parameter space

Accessible volume fraction

View larger version (15K): [in this window] [in a new window]   FIGURE 3   Effects of changes in the accessible volume fraction and buffer mobility. Different estimates of the accessible volume fraction are compared in (A-C). (A) The surface plot obtained for an accessible volume of 70% of the total cell volume (compare also with Fig. 2 G, where accessible volume was 50%). In (B) Ca2+ profiles are shown for the cell center (blue), restricted space (green), and periphery (red). (C) The Ca2+ profile at 100 ms. These signals are compared with those from a volume of 50% (dashed lines in (B) and (C)). In panels (D-F) we illustrate the effect of buffer mobility. Panel (D) represents a surface plot when all buffers remain stationary (compare with Fig. 2 G, where Fluo-3 and calmodulin were mobile). (E) and (F) allow a quantitative comparison of the effects of buffer mobility. The color-coding is identical to (B) and (C). The Ca2+ signal in the center is dramatically slowed down while the Ca2+ concentration in the restricted space reaches much larger values when the buffers are immobilized.

Buffer mobility

1

1

Fluo-3 concentration

View larger version (23K): [in this window] [in a new window]   FIGURE 4   Effect of Fluo-3 concentration, ATP concentration, and presence of phospholipids on the sarcolemma. Surface plots were constructed from simulations with different concentrations of the mobile Ca2+ buffer Fluo-3 (A, 0 µM; B, 30 µM; C, 400 µM). The time course of the Ca2+ concentration in the center of the cell is illustrated in (D) for Fluo-3 concentrations from 0 µM to 1600 µM. Two effects of Fluo-3 become apparent: 1) at low concentrations, Fluo-3 accelerated the Ca2+ signal in the cell center because it carries bound Ca2+ while it diffuses; 2) at high concentration, the Ca2+ signals are suppressed because the buffering capacity of Fluo-3 dominates. This is also evident in (E) where the Ca2+ concentration at 100 ms is shownat various Fluo-3 concentrations for the restricted space (red), the periphery (green), the average (black), and the cell center (blue). In the center, the dual effect of Fluo-3 leads to low Ca2+ at both low and high Fluo-3 concentrations. Panels (F-G) show the effect of the mobile Ca2+ buffer ATP in the absence of Fluo-3. Panel (F) allows a quantitative comparison of the effect of ATP. Panel (