Combined patch-clamp and Fura-2 measurements were performed on chinese hamster ovary (CHO) cells co-expressing two channel proteins involved in skeletal muscle excitation-contraction (E-C) coupling, the ryanodine receptor (RyR)-Ca2+ release channel (in the membrane of internal Ca2+ stores) and the dihydropyridine receptor (DHPR)-Ca2+ channel (in the plasma membrane). To ensure expression of functional L-type Ca2+ channels, we expressed α2, β, and γ DHPR subunits and a chimeric DHPR α1 subunit in which the putative cytoplasmic loop between repeats II and III is of skeletal origin and the remainder is cardiac. There was no clear indication of skeletal-type coupling between the DHPR and the RyR; depolarization failed to induce a Ca2+ transient (CaT) in the absence of extracellular Ca2+ ([Ca2+]o). However, in the presence of [Ca2+]o, depolarization evoked CaTs with a bell-shaped voltage dependence. About 30% of the cells tested exhibited two kinetic components: a fast transient increase in intracellular Ca2+ concentration ([Ca2+]i) (the first component; reaching 95% of its peak <0.6 s after depolarization) followed by a second increase in [Ca2+]i which lasted for 5–10 s (the second component). Our results suggest that the first component primarily reflected Ca2+ influx through Ca2+ channels, whereas the second component resulted from Ca2+ release through the RyR expressed in the membrane of internal Ca2+ stores. However, the onset and the rate of Ca2+ release appeared to be much slower than in native cardiac myocytes, despite a similar activation rate of Ca2+ current. These results suggest that the skeletal muscle RyR isoform supports Ca2+-induced Ca2+ release but that the distance between the DHPRs and the RyRs is, on average, much larger in the cotransfected CHO cells than in cardiac myocytes. We conclude that morphological properties of T-tubules and/or proteins other than the DHPR and the RyR are required for functional “close coupling” like that observed in skeletal or cardiac muscle. Nevertheless, some of our results imply that these two channels are potentially able to directly interact with each other.
Introduction
The dihydropyridine receptor (DHPR)1 and the ryanodine receptor (RyR) are the major proteins thought to be involved in the initial step of excitation-contraction (E-C) coupling, the process that links depolarization of the transverse (T)-tubular membranes to release of Ca2+ from the sarcoplasmic reticulum (SR). The DHPRs are located in the T-tubular membranes and are structurally identified as tetrads; the RyRs are located in the SR membranes and are structurally identified as the so-called “feet” which span the gap between the T-tubule and SR membranes (for current reviews see Rios and Pizarro, 1991; Franzini-Armstrong and Jorgensen, 1994; Meissner, 1994).
In skeletal muscle, the DHPRs are considered to primarily function as voltage sensors (Rios and Brum, 1987; Tanabe et al., 1988) for inducing SR Ca2+ release possibly via a direct interaction with the RyR-Ca2+ release channels (Takeshima et al., 1994; Nakai et al., 1996). Consistent with this is the presence of intramembrane charge movement (Chandler et al., 1976), which indicates molecular rearrangement of the DHPRs (Rios and Brum, 1987; Adams et al., 1990). The voltage sensor hypothesis would account for the following functional features of skeletal muscle (Rios and Pizarro, 1991): (a) voltage-gated Ca2+ release starts to occur without requiring a second messenger, such as Ca2+ or InsP3; (b) Ca2+ transients (CaTs) exhibit a sigmoidal voltage dependence; and (c) Ca2+ release flux that occurs following depolarization starts to decay within a millisecond of fiber repolarization. Thus, it is generally believed that the voltage sensors (DHPRs) are mechanically linked to the release channels (RyRs) (Chandler et al., 1976; Rios et al., 1993), a notion strongly supported by ultrastructural evidence (Block et al., 1988) as well as biochemical evidence (Brandt et al., 1990; Marty et al., 1994).
In cardiac myocytes, the DHPR represents a major pathway for entry of extracellular Ca2+, which in turn causes Ca2+ release from the SR via the Ca2+-induced Ca2+ release (CICR) mechanism (Fabiato, 1985; Beuckelman and Wier, 1988; Näbauer et al., 1989, Cleemann and Morad, 1991). In contrast to skeletal E-C coupling, Ca2+ release through the RyRs in cardiac muscle appears to be tightly controlled by both the amplitude and duration of Ca2+ current (Cannell et al., 1987; Barcenas-Ruiz and Wier, 1987; Callewaert et al., 1988; Cleemann and Morad, 1991; Wier et al., 1994). Although this behavior is not easily reconciled with the regenerative nature of CICR, it does support the notion that Ca2+ release is triggered by Ca2+ entry through the Ca2+ channels. A recent study using confocal microscopy has demonstrated that Ca2+ release occurs within a few milliseconds of depolarization both at the edge and at the deep myoplasm (Cheng et al., 1994). This together with biochemical data (Brandt and Basset, 1986) suggests that a large fraction of the DHPRs reside in the T-tubule membranes, and thereby allow efficacious communication with the RyRs. Consistent with this, a local increase in cytoplasmic free calcium concentration ([Ca2+]i) due to Ca2+ entry through the DHPR-Ca2+ channels is much more effective in inducing SR Ca2+ release than the spatially averaged (global) increase in [Ca2+]i induced by Ca2+ entry through the Na-Ca exchanger (Sham et al., 1995a; Lopez-Lopez et al., 1995). Furthermore, Lopez-Lopez et al. (1995) and Cannell et al. (1995) have suggested, again based on confocal microscopy data, that even a single Ca2+ channel can induce local Ca2+ release from one or a small group of RyRs (Ca2+ release element) in a stochastic manner, despite the rather small conductance of the Ca2+ channel (Rose et al., 1992). Thus, even in cardiac muscle, the DHPRs and the RyRs appear to be located in close proximity with each other, presumably within 20 nm (Sham et al., 1995a; Cannell et al., 1995). This unique spatial arrangement of both receptors possibly in combination with the recently described phenomenon of ryanodine receptor adaptation (Györke and Fill, 1993; Yasui et al., 1994; Valdivia et al., 1995) may be the primary reason for the “paradoxical” stable large amplification of CICR observed in cardiac myocytes, provided that there is no significant interaction among elementary Ca2+ release units (Stern, 1992; Cannell et al., 1995; Lopez-Lopez et al., 1995).
In the present study, we have co-expressed in chinese hamster ovary (CHO) cells the skeletal RyRs (Takeshima et al., 1989), α2, β, and γ DHPR subunits and a chimeric DHPR α1 subunit which possesses the regions critical for skeletal-type E-C coupling while conducting cardiac-like Ca2+ current (Tanabe et al., 1990). The purpose of the present study was to investigate possible interactions between these two channels in nonmuscle cells. The cotransfected CHO cells released Ca2+ through the RyRs following depolarization, but the Ca2+ release required entry of extracellular Ca2+. Furthermore, the rate of increase in [Ca2+]i due to Ca2+ release was much slower than that of cardiac myocytes, although the activation rate of Ca2+ current was similar. These results are compatible with the spatial arrangement of both receptors described for this preparation (Takekura et al., 1995b). Some aspects of this work appeared in an abstract form (Suda et al., 1996).
Methods
Cell Preparation and Transfection
For details about construction of expression plasmids, transfection, RNA-, and immunoblotting analysis and other molecular biological methods, see Takekura et al. (1995b). Briefly, CHO cells were transfected with plasmids pRRS11 (for skeletal RyR) and pCAS8 (for chimeric DHPR α1 subunit) in addition to pCAS7, pCAS14, and pCAS15 (for α2, β, and γ subunits of the DHPR, respectively). G418-resistant clones were screened by RNA-blotting analysis (Takeshima et al., 1989) which gave positive results for all expected components. Expression of the RyRs and the chimeric DHPRs were confirmed by immunoblotting analysis (Takeshima et al., 1989) using monoclonal antibodies against the skeletal RyR and the cardiac DHPR, respectively. [3H]Ryanodine binding assay and PN200-110 binding assay also confirmed expression of the RyRs and the chimeric DHPRs, respectively.
Cells were plated on coverslips and cultured at 37°C and 10% CO2 in MEM-α (without nucleoside) medium (GIBCO BRL, Gaithersburg, MD) supplemented with FBS (10%), glutamine (2 mM), penicillin (0.06 mg/ml), streptomycin (0.11 mg/ml), and Na-pyruvate (1 mM).
Electrophysiological and Fluorescence Recordings
Combined patch-clamp and Fura-2 measurements were performed with a computer-controlled data acquisition system (EPC-9, HEKA, Lambrecht, Germany). The tight-seal whole-cell configuration was used to control membrane voltage and measure ionic current. Liquid junction potential correction (+8 mV) was made before seal formation when patch-pipette solution contained glutamate as the major anion. Capacitive currents were determined and compensated before each voltage pulse using the automatic capacitance compensation feature of the EPC-9. The holding potential was usually −70 mV. Currents were low-pass filtered (8-pole Bessel) at a corner frequency of 3.0 kHz and digitized at 100-μs intervals. For gating current experiments, the currents were filtered at 5 kHz and digitized at 40-μs intervals. To remove residual linear capacitive and leakage currents in the test currents, 4 or 6 scaled, hyperpolarizing control voltage steps (each one-fourth or one-sixth the magnitude of the corresponding test step) were given before applying test pulses (P/−4 or P/ −6). For presentation of figures, currents were digitally filtered at 3 kHz. Spatially averaged [Ca2+]i was monitored with a photomultiplier based system where two fluorescence intensities were sampled at 2–10 Hz by a computer-driven charting program. Emission was measured at 500 nm, while excitation centered at 360 and 390 nm (in some experiments excitation was at 350 and 380 nm). An in situ calibration was made to estimate the absolute value of [Ca2+]i as described previously (Suda and Penner, 1994). We used BAPTA [bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetate] (Molecular Probes, Inc., Eugene, OR) for adjusting [Ca2+]i in the calibration solutions, assuming a dissociation constant of 225 nM (Harrison and Bers, 1987). Usually, depolarizing pulses were started 200–250 s after obtaining the whole-cell configuration when fluorescence intensities approached steady levels. Experiments were performed at room temperature (22–25°C).
Solutions and Chemicals
Standard extracellular saline contained (in mM): TEA-Cl 150, CaCl2 3-5, MgCl2 1, glucose 11, Hepes 10, pH 7.2. Nominally Ca2+-free saline was made by replacing Ca2+ with an equimolar concentration of Mg2+. Gating currents were measured while applying nominally Ca2+-free saline (plus appropriate amounts of MgCl2 as suggested by Pizarro et al., 1989) or a standard saline to which 1–2 mM Cd2+ and 0.1 mM La3+ were added. For Ca2+ current measurements as illustrated in Fig. 1, external saline contained (in mM): NaCl 145, KCl 2.8, CaCl2 10, MgCl2 1, TEA-Cl 10, Hepes 10, glucose 11. Sylgard-coated patch-pipettes (Kimax) had resistances between 2–3 MΩ after filling with internal solution containing (in mM): Cs-glutamate 145, Fura-2 (potassium salt; Molecular Probes) 0.07–0.1, NaCl 8, Mg-ATP 6, Na2-ATP 2, Hepes-CsOH 10, pH 7.2. The free magnesium concentration is calculated to be 0.35 mM, with an assumed affinity constant of 6.9 × 103 M−1 for Mg2+ binding to ATP. In some experiments, the following internal solution was used (in mM): Cs-glutamate 145, Fura-2 0.07–0.1, NaCl 8, Mg-ATP 7, Na2-ATP 1, Hepes-CsOH 10, pH 7.2. The free magnesium concentration of this solution is calculated to be 0.6 mM. For Ca2+ current measurements as illustrated in Fig. 1, the pipette solution contained (in mM): N-methyl- d-glucamine145, NaCl 8, MgCl2 2, Cs-EGTA 20, Mg-ATP 4, GTP 0.3, Hepes 10, pH 7.2. Drug application was made by local ejection from wide-tipped pipettes. When tetracaine (100 μM, Sigma Chemical Co., St. Louis, MO) was used, the saline contained 7 mM Ca2+. Usually, 10 mM caffeine (Sigma Chemical Co.) was applied in nominally Ca2+-free saline. But in the experiments as illustrated in Fig. 5, caffeine (2 mM) was applied in 3 mM Ca2+ solution. All chemicals were of analytical grade.
Ca2+ Current Density in Cardiac Myocytes and Transformed CHO Cells
The apparent Ca2+ current density of the cotransfected CHO cells, that exhibited CICR, varied from 34 to 160 pA/pF and was on average 74 ± 44 pA/pF (mean ± SEM) at +20 mV (n = 10). These values were measured 250–400 s after obtaining the whole-cell configuration with 0.1 mM Fura-2 as the only exogenous Ca2+ chelator. In cardiac myocytes, peak inward Ca2+ current is at most 2 nA for both rat and guinea-pig ventricular cells and the cell surface area (including T-tubules) is estimated to be larger than 6–7 × 10−5 cm2 (e.g., Callewaert et al., 1988; Cannell et al., 1995) Thus, assuming a specific membrane capacitance of 1 μF/ cm2, the current density would be at most 30 pA/pF. Actual reported values in the literatures are much smaller (<10 pA/pF) than this (e.g., Sham et al., 1995b). Therefore, it is clear that the apparent current densities of the cotransfected CHO cells that exhibited CICR were much larger than the upper limit of those of cardiac myocytes. With “low” Ca2+ current densities (<20 pA/ pF), CICR was not detected in this preparation.
Results
Ca2+ Currents of Transformed CHO Cells Are Similar to Ca2+ Currents in Cardiac Cells
Because the putative transmembrane regions of the chimeric DHPR expressed in the transfected CHO cells are of cardiac origin, the characteristics of the Ca2+ current are expected to be similar to those of cardiac myocytes (Tanabe et al., 1990). In Fig. 1,A, the cell was bathed in a saline containing 5 mM extracellular Ca2+ concentration ([Ca2+]o). Depolarizing pulses of 150-ms duration were applied from a holding potential of −70 mV to various test potentials, in 20-mV increments. Inward currents were evoked by depolarizations between −30 mV and +80 mV, while more positive voltages evoked small outward currents. Peak inward currents were measured around +10 mV (Fig. 1,B). A fit to the mean current-voltage curve yielded values for V1/2 of +4 mV and a slope factor of +11 mV. The activation time constant of the current was on the order of 1–3 ms (Fig. 1,C) and similar to L-type Ca2+ current present in native cardiac myocytes (data not shown). Inward currents were strongly reduced in nominally Ca2+-free saline and were completely blocked by low concentrations (0.5–1 μM) of the dihydropyridines, PN200-110 (n = 4; see Fig. 1 D), nifedipine (n = 3), and nimodipine (n = 2), or 1–2 mM Cd2+ (n = 10). Peak inward currents were significantly increased and the decay of the tail current was slowed by 1 μM Bay-K 8644 (n = 2). These results, taken together, confirm that the current reflected Ca2+ flux through the dihydropyridine-sensitive Ca2+ channels. No such currents were detected in nontransfected CHO cells (n = 10; data not shown).
Cotransfected CHO Cells Exhibit Different Patterns of Ca2+ Transients
The transformed cells used in the present study also expressed type-1 RyRs, which are intracellular Ca2+ release channels not normally found in CHO cells (Takekura et al., 1995b). We used relatively low concentrations of Mg2+ (0.35–0.6 mM) and a high concentration of total ATP (8 mM) in the internal solutions to provide favorable conditions for activation of RyRs (Meissner, 1994), with 0.1 mM Fura-2 being the only exogenous Ca2+ chelator. Under these conditions, basal [Ca2+]i was usually between 50 and 300 nM, except when spontaneous [Ca2+]i oscillation occurred. 5 mM [Ca2+]o was the upper limit to maintain the basal [Ca2+]i below 300 nM.
Application of 5–10 mM caffeine confirmed that large CaTs could be induced in the absence of [Ca2+]o (e.g., Fig. 4,C; n > 24). Furthermore, ∼10% of the tested cells exhibited spontaneous [Ca2+]i oscillations in 3–5 mM [Ca2+]o, even though the membrane potential was clamped at −70 mV (n = 11; Fig. 2,A). In these cells, basal [Ca2+]i was usually higher than 0.4 μM and depolarizing voltage pulses, which normally would produce a CaT (see below), occasionally failed to induce a CaT, possibly due to inactivation of Ca2+ currents at high concentrations of global [Ca2+]i. When removing [Ca2+]o, basal [Ca2+]i decreased and the oscillation frequency was reduced (Fig. 2 A; n = 3). Addition of 100 μM tetracaine not only decreased the basal [Ca2+]i but terminated the oscillations completely (n = 5; data not shown), suggesting that these oscillatory events occurred through pulsatile release of Ca2+ through RyRs. This, together with the Ca2+ current data presented above, establishes that the DHPRs and the RyRs form functional channels in the surface membrane and the membranes of internal stores, respectively. This stands in good agreement with immunoblotting analysis (Takekura et al., 1995b).
The next question we asked was whether these two channels located in distinct membranes are able to communicate with each other. Fig. 2, B–D show examples of three representative CaTs evoked by depolarizing pulses to +20 mV (250 ms) at 60-s intervals in different cells. In all cases, a transient increase in [Ca2+]i occurred immediately after depolarization. Such depolarization-induced CaTs were only observed when the cells were bathed in Ca2+-containing saline (usually 5 mM [Ca2+]o). In some cells, the CaT returned to baseline with a monoexponential time course, whereas in others the rapid [Ca2+]i spike was followed by a secondary, slower rise in [Ca2+]i (Fig. 2, C and D).
To assess the CaTs in a more quantitative manner, we analyzed them as illustrated in Fig. 3. The criterion used to distinguish the two phases is based on the break point where the time derivative of the Ca2+ signal decreases (following the rapid rise in the Ca2+ signal). Thus, the fast initial component appears in <0.75 s after the onset of depolarization and reaches 95% of its peak value within 0.6 s. Subsequently, [Ca2+]i continues to rise in cells that exhibit a second component, but at significantly decreased rates. Our results suggest that the first component primarily reflects an increase in [Ca2+]i due to Ca2+ current, whereas the second component reflects an increase in [Ca2+]i due to Ca2+ release through the RyRs (Figs. 5 and 6). Throughout this article, we will refer to the ratio of the magnitude (i.e., the difference between peak and basal [Ca2+]i) of the secondary increase in [Ca2+]i to that of the initial peak of the Ca2+ transient elicited by depolarization as the CaT2/CaT1 Ratio (“CaTR”).
Based on the analysis of 98 cells (excluding those with [Ca2+]i oscillations), we estimated that ∼60% of the tested cells (n = 60) exhibited a rapid increase in [Ca2+]i (the first component in Fig. 3) followed by single-exponential decay to baseline (Fig. 2,B), whereas the others showed a sustained secondary increase in [Ca2+]i (the second component in Fig. 3) lasting for 5–10 s (n = 38). In the latter case, the ratio of the amplitude of the peak of the second component to that of the first component (“CaTR”) varied from cell to cell. While some cells exhibited a rather small ratio (e.g., Fig. 2,C: CaTR = 1.64), others had large ratios (e.g., CaTR = 2.69; Fig. 2 D). The size of the second component also varied from cell to cell, ranging between 0.7 and 3.4 μM. In general, the second component was gradually reduced and ultimately disappeared at later times during a lengthy experiment. However, as long as a relatively large ratio appeared at the first pulse (e.g., CaTR > 2.0), the second component was invariably present at least for the first 6–7 pulses (n = 12).
Recently, Takeshima et al. (1994) and Nakai et al. (1996) directly demonstrated that the skeletal RyR isoform functions as the physiological voltage-gated Ca2+ release channel. Because the chimeric DHPR expressed in the transformed CHO cells possesses the regions critical for interacting with the release channels (Tanabe et al., 1990), the following possibilities may explain the two components of the CaT, provided that the second component does not reflect a saturation of Ca2+-buffering capacity: (1) voltage-gated Ca2+ release from internal stores, (2) Ca2+ influx through Ca2+ channels, (3) CICR from internal stores, and (4) a combination of two or all of these factors.
Cotransfected CHO Cells Lack Voltage-gated Ca2+ Release
First, we examined whether voltage-gated Ca2+ release, an essential property of skeletal-type E-C coupling, is reconstituted in cotransfected CHO cells. As illustrated in Fig. 4,A, depolarizing pulses to +20 mV (250 ms) in the presence of 5 mM [Ca2+]o evoked a large CaT with two kinetic components. In nominally Ca2+-free saline, however, the CaT evoked by the same pulse was markedly reduced (n = 13; Fig. 4,B). The small CaT might reflect a residual Ca2+ current since we had not added EGTA to the external solution. Furthermore, stronger depolarizations (e.g., +90 mV), where the net Ca2+ influx probably was zero, did not evoke any detectable rise in [Ca2+]i (Fig. 4,B). The reduction or abolition of a CaT in nominally Ca2+-free saline was not attributable to depletion of stored Ca2+, as 10 mM caffeine still induced a large rapid increase in [Ca2+]i under the same conditions (Fig. 4 C). Depolarization also failed to induce a CaT in nominally Ca2+-free saline at quite negative holding potentials (−100 to −120 mV). Thus, absence of voltage-gated Ca2+ release in nominally Ca2+-free saline is unlikely to be due to an effect of Ca2+ deprivation on the voltage sensors (Brum et al., 1988a, b; Pizarro et al., 1989).
As illustrated in Fig. 4 E, the cells exhibited large gating currents which appeared at −30 mV and gradually increased, reaching 66.8 ± 25.6 nC/μF (n = 6) at +60 or +70 mV. These values are larger than Qmax values obtained from native mammalian skeletal muscle (15.8 − 46.2 nC/μF; Rios and Pizarro, 1991). Because charge movement does not significantly differ among skeletal, cardiac, and chimeric DHPRs when expressed in dysgenic myotubes (Adams et al., 1990), the expression level of the DHPRs in the transfected CHO cells is at least as high as functional voltage sensors in native skeletal muscles, assuming the same gating charge to be required to activate either species of channel. No such gating currents were detected in nontransfected CHO cells (data not shown).
Clearly, in contrast to skeletal muscle, CaTs did not exhibit a sigmoidal voltage dependence (Fig. 4 D ; n = 7) and CaTs were smaller at stronger depolarization (n = 17), presumably due to a smaller amount of Ca2+ influx. This suggests that, unlike in skeletal muscle, DHPRs and RyRs do not interact directly in cotransfected CHO cells (but see Discussion). Thus, CaTs elicited by depolarization in the presence of [Ca2+]o probably reflected Ca2+ influx alone or a combination of both Ca2+ influx and CICR.
The Second Component of Ca2+ Transient Is Due to Release from Internal Stores
To see whether CICR is involved in producing CaTs, we tested the effect of tetracaine, a CICR inhibitor, on the second component of the CaT. Since tetracaine caused a slight inhibition of Ca2+ currents, we increased [Ca2+]o in the tetracaine-containing solution to 7 mM to compensate for the effect of tetracaine on DHPRs. This produced almost the same amount of total Ca2+ influx as compared to control (e.g., compare the 2nd and 3rd pulses in Fig. 5 B).
Fig. 5,A illustrates a control experiment, in which the ratio of the amplitude of the second component to that of the first component was relatively large (CaTR > 2.5). It is seen that repetitive stimulation evoked similar-sized Ca2+ currents and CaTs under control conditions (the first four CaTs of the cell are shown). However, application of 100 μM tetracaine strongly reduced the second component (Fig. 5 B). Similar results were obtained in seven other cells, two of which exhibited complete abolition of the second component.
Conversely, cells that exhibited only the first component of CaT revealed a large second component in the presence of a subthreshold concentration of caffeine (2 mM), despite a smaller amount of Ca2+ influx (n = 6; Fig. 5 C).
No Indication of CICR in CHO Cells Expressing Only DHPR Subunits
We conducted control experiments using CHO cells expressing only DHPRs (−RyR) to confirm that solely CHO cells co-expressing DHPRs and RyRs (+RyR) are able to show Ca2+ release. In Fig. 6 A, CaTs from single-transfected (−RyR) and cotransfected (+RyR) CHO cells are superimposed in order to compare the time course of CaTs (induced by 250 ms depolarization to +20 mV in 5 mM [Ca2+]o). Depolarization-evoked CaTs in single-transfected CHO cells decayed monoexponentially after the first rapid peak (0.6 s after depolarization in this case) and invariably failed to exhibit any secondary increase in [Ca2+]i. The shape of the CaTs in the single-transfected cells (−RyR) was similar to those CaTs in (+RyR) cells that did not show a secondary phase.
Fig. 6,B plots the maximum depolarization-induced increase of [Ca2+]i against the normalized Ca2+ current integral of both (−RyR) and (+RyR) cells without a plateau phase. It is seen that the two regression lines fitted to the respective data sets do not differ significantly, supporting the idea that monoexponentially decaying CaTs reflect Ca2+ influx through Ca2+ channels rather than Ca2+ release. The analysis of cotransfected cells producing a plateau phase is shown in Fig. 6 C. Here, the peaks of the first component as well as the second component are plotted against the normalized charge (current integral) of the respective Ca2+ currents. It is seen that while the maximal amplitudes of the first component (▪) are well described by the regression line fitted to (+RyR) cells, the peaks of the second component (□) deviate significantly. This further supports the notion that the first component appears to primarily reflect Ca2+ influx through Ca2+ channels and that CICR is associated with expression of RyRs.
Separation of Two Kinetic Components in All-or-None Type Ca2+ Transient
To further establish that the second component is indeed caused by Ca2+ release from internal stores and not a consequence of saturation of the Ca2+-buffering capacity of the cell, we reduced CICR activity by increasing [Mg2+]i to 0.6 mM and reducing [Ca2+]o to 2 mM. These conditions were deemed sufficient to inhibit CICR, because the skeletal muscle RyR isoform is less sensitive to Ca2+ than the cardiac isoform and because the distance between the DHPRs and the RyRs appears to be much larger than in cardiac myocytes, as will be discussed later. CaTs exhibiting a secondary component were rare under these conditions. However, four cells did show CaTs with a plateau phase that seemed to occur almost in an all-or-none manner, since the same pulse (+20 mV, 200 ms) produced completely different patterns of CaTs within one cell.
The data in Fig. 7 A provide an example in which the first and the second components were clearly separated. It should be pointed out that the secondary component developed after the first component already started to decay, which indicates that the second component does not reflect a saturation of Ca2+-buffering capacity. Similarly dissociated CaTs were observed in five other cells, regardless of the individual CaTR.
The possibility remains that not only the amount of Ca2+ entry but also the filling state of internal Ca2+ stores may be an important determinant of the amount of Ca2+ release in this preparation. We therefore checked whether the pulse intervals of 40–60 s used in this study were appropriate to allow refilling of caffeine-sensitive pools. To this end we applied 10 mM caffeine during an interpulse interval to see whether such a caffeine-induced CaT would alter the shape of a subsequent depolarization-induced CaT (Fig. 7 B). However, the size of the CICR component evoked by a depolarization to +20 mV was not significantly affected by a preceding caffeine-induced Ca2+ release (n = 5). This suggests that a large fraction of the Ca2+ ions released by caffeine could be taken up again by the stores within 30 s and contribute to the next release phase.
Global Increase in [Ca2+]i Rather Than Local Increase in [Ca2+]i Determines the Amount of Ca2+ Release
Finally, we posed the question whether the amount of Ca2+ release is determined by either a global increase in [Ca2+]i or rather a local one. If the global increase in [Ca2+]i is the important determinant of triggering Ca2+ release, then the integral of the Ca2+ current rather than the peak amplitude will determine the amount of Ca2+ release. To test this, we conducted experiments where the cell was depolarized at 40-s intervals to the same potential (+20 mV) but with increasing pulse duration (Fig. 8). It is seen that the CaTs elicited by depolarizing pulses of longer than 100 ms clearly exhibited two kinetic components in which the second component grew as the depolarizing pulse duration was increased (Fig. 8,A, 175-ms and 250-ms pulse durations). Similar results were obtained when reversing the stimulus protocol by decreasing the pulse duration (Fig. 8 B). According to Stern's theory (1992) a low-gain system would require only very little additional Ca2+ influx in order to produce such a graded response in CICR. It is noteworthy that in this preparation DHPRs inactivated and thus only a small change in total Ca2+ influx was able to produce a relatively large change in [Ca2+]i. Therefore, one might argue that the total amount of Ca2+ influx rather than the peak current amplitude appeared to determine the amount of Ca2+ release in these cells (n = 9).
Discussion
CICR Is Reconstituted in the Cotransfected CHO Cells
The present study demonstrates that CHO cells, when co-expressed with the skeletal muscle RyR isoform and a chimeric DHPR, express both proteins as functional Ca2+ release channels and voltage-gated Ca2+ channels, respectively. Previous studies have established that type-1 RyR is responsible for voltage-gated Ca2+ release (Takeshima et al., 1994; Nakai et al., 1996) and that the chimeric DHPR expressed in CHO cells can in principle function as the voltage sensor of E-C coupling in dysgenic mouse myotubes (Adams et al., 1990). Although both channel proteins are expressed in high levels in CHO cells, there is no apparent reconstitution of voltage-gated Ca2+ release. However, we have obtained clear evidence for cross-talk between DHPRs and RyRs in the form of CICR.
The cotransfected CHO cells apparently exhibit two different patterns of CaTs. One pattern is similar to the CaT of control cells (expressing only with the DHPR), where a transient increase in [Ca2+]i is followed by an exponential decay. The other pattern exhibits two kinetic components: a transient increase in [Ca2+]i that is followed by a sustained secondary increase in [Ca2+]i lasting 5–10 s. Because the peak of the second component is reduced by addition of tetracaine and is enhanced by subthreshold concentrations of caffeine, this component appears to reflect Ca2+ release from internal stores. Consistent with this, the second component is not observed in single-transfected CHO cells. Further confirmation comes from the fact that, under certain conditions, the same pulse produces completely different patterns of CaTs: either the first component or a combination of both the first and second components.
When the calcium-transient ratio (CaTR) is relatively large (e.g., R > 3.0), the Ca2+ release occurrs in an all-or-none rather than graded manner, and occasionally the second (CICR) component develops after a partial decay of the first component (Fig. 7). It is likely that once a fraction of the RyRs starts to release Ca2+ following a global increase in [Ca2+]i, the released Ca2+ ions in turn trigger neighboring RyRs, resulting in a further increase in [Ca2+]i. It is also possible that development of the second component after a partial decay of the first component reflects a transient inhibition of Ca2+ release by excessive Ca2+ influx, which resumes after dissipation of [Ca2+]i. On the other hand, when the CaTR is small (e.g., R < 1.5), the Ca2+ release occurs in a more graded manner (e.g., Fig. 8), which might be due to the much smaller overall gain of CICR in the transfected cells as compared to cardiac myocytes (Stern, 1992). However, the present results do not rule out other explanations for graded CICR such as adaptation (Györke and Fill, 1993; Yasui et al., 1994; Valdivia et al., 1995) or even as yet unknown mechanisms. In any case, it seems safe to assume that the second component reflects a regenerative Ca2+ release process and not a saturation of Ca2+-buffering capacity.
It seems unlikely that the second component is due to Ca2+ release through the InsP3 receptors, which are also present in these cells, because the amplitude of the InsP3-induced CaT is much smaller than that of the caffeine-induced CaT, as previously reported (Penner et al., 1989). That the second component appears in the presence of a subthreshold concentration of caffeine (Fig. 5 C), an inhibitor of InsP3-induced Ca2+ release, also argues against a major role for InsP3-induced Ca2+ release. To test this point more directly, we dialyzed cells with 0.5 mg/ml heparin, a blocker of InsP3 receptors. Under these conditions, the second component could be observed in 4 out of 4 cells and the largest CaTR observed was 2.5 (data not shown).
CICR of the Cotransfected CHO Cells Differs from That of Cardiac Myocytes
While the present study suggests that Ca2+ entering through DHPRs can trigger CICR, the following features of CICR observed in the cotransfected CHO cells differ from those in cardiac myocytes and suggest that the spatial distance between the DHPRs and the RyRs is, on average, much larger in cotransfected CHO cells than in cardiac myocytes: (1) the activation rate of CICR is very slow (it takes 5–10 s to reach its peak), although the activation rate of Ca2+ current is similar to that of native cardiac myocytes. (2) Likewise, the onset of CICR appears to be slow and may develop after a substantial delay, since the secondary rise in [Ca2+]i occasionally develops after a partial decay of the Ca2+ current component (cf. Fig. 7 A). Even though we cannot completely exclude the possibility that CICR may contribute to the rapid first component, the time to peak of the first component is usually several hundreds of milliseconds and is considerably slower than that of CaTs in cardiac myocytes (<40 ms with Fura-2). Also, the onset of CICR in native cardiac myocytes is extremely fast, occurring within 2–4 ms of depolarization (Cheng et al., 1994; Cannell et al., 1995). (3) The size of the Ca2+ release component is larger when the pulse duration is prolonged. It proved difficult to induce the CICR by pulses briefer than 100 ms, even though the Ca2+ current density is higher than that of cardiac myocytes. When the Ca2+ current density is lower than 20 pA/pF, the second component is rarely detected. By contrast, depolarizations as brief as 10 ms are sufficient to elicit maximal Ca2+ release in cardiac myocytes (Cleemann and Morad, 1991; Cannell et al., 1995). (4) When the CaTR is relatively large, CICR is all-or-none, whereas cardiac myocytes exhibit a graded CICR, despite a large gain (Cannell et al., 1987; Beuckelman and Wier, 1988; Wier et al., 1994). (5) When the concentration of Fura-2 was higher than 200 μM (data not shown), the CICR component was rarely observed, indicating that Ca2+ ions released from the RyRs when captured by moderate increases in buffering capacity with a fast and high affinity Ca2+ buffer, reduced the efficacy of CICR. In cardiac myocytes, even 10 mM EGTA fails to alter the effect of SR Ca2+ release on the Ca2+-dependent inactivation of Ca2+ current (Sham et al., 1995a). (6) Tail current-induced Ca2+ transients, as occasionally observed in cardiac myocytes (Cannell et al., 1987; Cleemann and Morad, 1991; Cannell et al., 1995) were never observed in the cotransfected CHO cells.
Thus, the spatially averaged (global) increase in [Ca2+]i rather than the local increase in [Ca2+]i is likely to be involved in the initial triggering of Ca2+ release in the cotransfected CHO cells. It should be pointed out that the rate of increase in [Ca2+]i due to CICR remained slow compared with that of cardiac myocytes even when optimizing conditions that are known to increase the sensitivity of the RyRs to Ca2+ (i.e., in the presence of 2 mM caffeine, low Mg2+ [0.35 mM] and high total ATP [8 mM]). This would suggest that the slowness of the Ca2+ release process in this preparation is not due to low Ca2+ sensitivity of the skeletal RyR isoform compared with that of the cardiac RyR isoform, but rather to the relative spatial distribution of the expressed receptors.
Skeletal-Type E-C Coupling Is Not Reconstituted in the Cotransfected CHO Cells
The spatial distance between the DHPRs (in the surface membrane) and the RyRs (in the membrane of internal Ca2+ stores) appears to be, on average, much larger in the transfected CHO cells than in cardiac myocytes, and hence, it is not surprising that there is no skeletal-type interaction between the DHPRs and the RyRs. The following evidence is in support of this view: (1) The CaT is strongly reduced and the second (CICR) component disappears in nominally Ca2+-free saline. (2) The CaT exhibits a bell-shaped voltage dependence. (3) Repolarization fails to even partially curtail (or change the time derivative of) the depolarization- induced CaT. In other words, [Ca2+]i continues to increase for several seconds following repolarization. (4) Repolarization fails to induce any dip in the caffeine-activated CaT, which suggests that activated RyRs are not influenced by the DHPR. In mammalian skeletal muscle, repolarization terminates caffeine-induced Ca2+ release (RISC phenomenon; Suda and Penner, 1994). Together, these results lead us to the conclusion that there is no direct interaction between the DHPRs and the RyRs in the transfected CHO cells (but see below).
Possible Indications of Direct Coupling between the DHPR and the RyR in Cotransfected CHO Cells
We cannot completely exclude the possibility that a small fraction of the RyRs might be directly coupled to the DHPRs and thereby contributes to voltage-gated Ca2+ release, resulting in a larger amount of Ca2+ release with a longer pulse (Brum et al., 1988a). Although our results suggest that Ca2+ release does not occur to any significant extent in the absence of [Ca2+]o, despite the caffeine-sensitive Ca2+ stores being sufficiently full (cf. Fig. 4,C), this may not unambiguously rule out some voltage-gated Ca2+ release component. One finding that might point towards a “facilitatory” role of depolarization on [Ca2+]i, that is not easily reconcilable with a simple CICR scheme, is that Ca2+ influx elicited by stronger depolarizations appears to release Ca2+ more efficiently than that by weak depolarizations (cf. Fig. 4 D). Also, one should recall that when the chimeric DHPR (cytoplasmic loop between repeats II and III is of skeletal origin and the remainder is cardiac) was expressed in dysgenic myotubes, voltage-gated Ca2+ release occurred more efficiently in the presence of [Ca2+]o than in the absence of [Ca2+]o (Tanabe et al., 1990). Furthermore, it is possible that the caffeine-sensitive pools that contribute to direct coupling of both receptors are located very close to the surface membrane and thus are easily emptied upon removal of extracellular Ca2+. Because the total capacity of these pools may be small compared to other caffeine-sensitive pools, caffeine might still be able to induce Ca2+ release from the pools which are located far away from the surface membrane.
The results presented in this study are basically in accordance with structural features of the same transformant (Takekura et al., 1995b), in which no junctions are formed. Recently, Takekura et al. (1995a) have shown in “dyspedic” myotubes that even in the absence of feet (corresponding to the RyRs), junctions are formed between the T-tubule membranes and the SR membranes, although the length of the junctions are shorter than those observed in normal muscle, probably because of the lack of the feet structures. This would suggest that proteins other than the DHPR and the RyR are necessary for the initial forming of junctions between the T-tubule and the SR membranes. CHO cells may lack the necessary proteins for linking the surface membrane to ER membranes (Takekura et al., 1995a,b). Such proteins may also be involved in diadic junctions in cardiac muscle to accomplish tight control of SR Ca2+ release by relatively small Ca2+ entry through the DHPR Ca2+ channels.
Acknowledgments
We are grateful to the late Dr. S. Numa for support and constant encouragement during the early stage of this project, Dr. T. Tanabe for the cardiac-skeletal DHPR chimera, Drs. V. Flockerzi and F. Hofmann for β and γ subunits of the DHPR, Dr. K.P. Campbell for antibody to the skeletal DHPR. We also thank Dr. E. Neher for constant encouragement, support, and comments on the manuscript, Dr. S.H. Heinemann for help with gating current measurement, Drs. K.G. Beam, A. Korotzer, and N.M. Losenzon at Colorado State University for critical reading and comments on the manuscript, Drs. K. Imoto and J. Nakai at National Institute for Physiological Sciences (Japan) for helpful discussions and comments on the manuscript, Dr. Y. Hakamata at Kyoto University for stimulating discussions during the early stage of this project. Finally, we thank Frauke Friedlein and Michael Pilot for excellent technical assistance.
This work was supported by the following institutions: Deutsche Forschungsgemeinschaft, Sonderforschungsbereich 236, Hermann-und Lilly-Schilling-Stiftung to R. Penner; The Ministry of Education, Science and Culture (Japan) and The Japan Heart Foundation (I.B.M.) to H. Takeshima. N. Suda was supported by Ciba-Geigy Foundation for the Promotion of Science (Japan) and a grant from the Max-Planck-Gesellschaft. The manuscript was written while N. Suda received a grant from the Human Frontier Science Program.
references
Address reprint requests to R. Penner, Department of Membrane Biophysics, Max-Planck-Institut fur biophysikalische, Am Fassberg, 37077 Gottingen, Germany. Fax: 49-551-201-1688.
Dr. N. Suda's present address is Department of Anatomy and Neurobiology, Colorado State University, Fort Collins, CO 80523.
Dr. M. Hoth's present address is Department of Molecular and Cellular Physiology, Stanford University, Stanford, CA 94305-5426.
Dr. S. Nishimura's present address is Department of Molecular and Cellular Biology, Nippon Boehringer Ingelheim Co., LTD., Kawanishi Pharma Research Institute, Yato, Takada 103, Kawanishi, Hyogo 666-01, Japan.
Abbreviations used in this paper: CaT, Ca2+ transient; CHO, Chinese hamster ovary; CICR, Ca2+-induced Ca2+ release; CaTR, calcium-transient ratio; DHPR, dihydropyridine receptor; E-C, excitation-contraction; RyR, ryanodine receptor; SR, sarcoplasmic reticulum, T, transverse.