A cytoplasmic factor, calpastatin and ATP together reverse run-down of Ca²⁺ channel activity in guinea-pig heart

Abstract

1. The cytoplasmic extract of bovine heart was separated into four fractions by gel filtration: H (molecular mass > 300 kDa), P (250-300 kDa), L₁ (180-250 kDa) and L₂ (< 180 kDa). The effects of these fractions on the run-down of L-type Ca²⁺ channel activity were investigated in guinea-pig ventricular myocytes.

2. After run-down induced by inside-out patch formation, Ca²⁺ channel activity was restored by P or H (+ 3 mM ATP) to 7·5 and 5·8 % of that in the cell-attached mode, respectively, but to as high as 86 % by P + H + ATP.

3. The reversal of run-down brought about by the P fraction was mimicked by calpastatin.

4. The restorative effect of calpastatin + ATP showed a biphasic time course: 38 % in the early transient phase and 11 % in the late phase. However, calpastatin + H + ATP showed a sustained effect: 66 % in the early transient phase, and 87 % in the late phase.

5. The effective component of the H fraction showed a protein-like nature: heat and trypsin sensitivity.

6. The activities of cAMP-dependent protein kinase, casein kinase I, casein kinase II, protein tyrosine kinase, protein serine/threonine or tyrosine phosphatases were measured. However, these kinases and phosphatases were not confirmed as the effective component of cytoplasm or the H fraction.

7. Run-down was not prevented by 2 μM phalloidin or 2 μM paclitexal, suggesting that neither actin filaments nor microtubules are directly involved in the run-down.

8. Our results support the view that the basal activity of the Ca²⁺ channel is maintained by at least three factors: a protein-like factor in the H fraction, calpastatin, and ATP.

Voltage-operated L-type Ca²⁺ channels play important roles in cellular excitation related to rhythmic firing, synaptic transmission, muscle contraction and secretion. Electrophysiological studies of the channels are usually carried out by using patch-clamp techniques. It has been noticed, however, that activity of the channels decreases when the cytoplasmic side of the channels is perfused with an artificial physiological solution. This phenomenon is referred to as run-down, and is most pronounced in the cells dialysed internally in whole-cell recording or in inside-out patches (for reviews see Hagiwara & Byerly, 1983; Kostyuk, 1984; McDonald et al. 1994). Thus run-down has been an obstacle to studies of the Ca²⁺ channel. Furthermore, only the L- but not the T-type Ca²⁺ channel and certain other channels exhibit run-down (Fenwick et al. 1982; Nilius et al. 1985), suggesting that the basal activity of these channels is maintained by mechanisms specific to the species of channel. Such a mechanism, however, has not yet been established. Therefore, studies on run-down are important, not only for the promotion of research on Ca²⁺ channels, but also for the understanding of the regulation of channel activity.

So far, some mechanisms have been suggested for run-down, including proteolysis (Chad & Eckert, 1985; Belles et al. 1988; Romanin et al. 1991) and dephosphorylation of the channels (Kostyuk, 1984; Armstrong & Eckert, 1987; Ono & Fozzard, 1992). Alternatively, we have suggested that, in the inside-out patch mode, cardiac cytoplasm + ATP restores channel activity to levels observed in the cell-attached mode, suggesting that a cytoplasmic factor(s) is involved in maintenance of channel activity, and that washout of the factor(s) is the cause of run-down (Kameyama, M. et al. 1988; Kameyama, A. et al. 1997). Furthermore, our recent study indicates that the properties of one possible factor in cytoplasm resemble those of calpastatin (Kameyama, A. et al. 1998), an endogenous inhibitor of Ca²⁺-activated neutral protease. However, fractionated cytoplasm or purified calpastatin shows a limited ability to recover channel activity (Kameyama, A. et al. 1998), even when replenished with ATP which is known to enhance the action of the cytoplasm (Yazawa et al. 1997). To account for this, three possibilities have been suggested: (1) the factor might be calpastatin, but it is partially degraded during the purification procedure; (2) the factor might not be calpastatin but another(s) active component that is co-purified with calpastatin; (3) there might be other cofactor(s) discarded during the fractionation procedure (Kameyama, A. et al. 1998).

This study was carried out in order to evaluate the possibilities described above and elucidate the nature of the cytoplasmic factors. We report here that another protein-like factor, together with calpastatin and ATP, restores L-type Ca²⁺ channel activity.

METHODS

Preparation of single myocytes

Single ventricular myocytes from guinea-pig hearts were dispersed by collagenase and protease as described previously (Yazawa et al. 1990). Briefly, a female guinea-pig (weight 400-600 g) was anaesthetized with sodium pentobarbitone (30 mg kg⁻¹i.p.), and the aorta cannulated in situ under artificial respiration. The heart was then dissected out, mounted on a Langendorff apparatus and perfused with Tyrode solution (3 min at 37°C), followed by nominally Ca²⁺-free Tyrode solution for 5 min and a Ca²⁺-free Tyrode solution containing collagenase (0.08 mg ml⁻¹, Yakult, Japan) for 10 to 15 min. The heart was then washed out with a high-K⁺ and low-Ca²⁺ solution (storage solution) and the ventricular myocytes dispersed and filtered through a stainless-steel mesh (105 μm). The myocytes were washed twice by centrifugation (800 r.p.m. for 3 min) and kept at 4°C. In some cases, myocytes were treated with protease (Nagase NK-103, 0.05 mg ml⁻¹) and DNase I (Sigma Type IV, 0.02 mg ml⁻¹) to improve the success rate in attaining a gigaohm seal. This study was carried out with permission from the committee of Animal Experimentation, Faculty of Medicine, Kagoshima University, Japan.

Solutions

Tyrode solution contained (mM): NaCl, 135; KCl, 5.4; NaH₂PO₄, 0.33; MgCl₂, 1.0; glucose, 5.5; CaCl₂, 1.8; and Hepes-NaOH buffer, 10 (pH 7.4). The storage solution was composed of (mM): KOH, 70; glutamic acid, 50; KCl, 40; KH₂PO₄, 20; taurine, 20; MgCl₂, 3; glucose, 10; Hepes, 10; and CaEGTA buffer, 0.5; pH was adjusted to 7.4 with KOH. The pipette solution contained (mM): BaCl₂, 50; TEACl, 70; EGTA, 0.5; and Hepes-CsOH buffer, 10 (pH 7.4). The basic internal solution consisted of (mM): potassium aspartate, 90; KCl, 30; KH₂PO₄, 10; EGTA, 1; MgCl₂, 0.5; CaCl₂, 0.5; and Hepes-KOH buffer, 10 (pH 7.4).

Materials

The agents used in this study were as follows: aprotinin and pepstatin A from Boehringer Mannheim, Germany; iodoacetamide from Katayama, Japan; trichloroacetic acid (TCA) and trypsin from Wako, Japan; N-(2-methylaminoethyl)-5-isoquinolinesulphonamide hydrochloride (H8) from Seikagaku Kogyo, Japan; adenosine 5′-[γ³²P]triphosphate ([³²P]ATP) from Amersham Life Science, Tokyo, Japan; ATP (Mg⁺ salt), anthracene-9-carboxylic acid (9AC), benzamidine, histone type II-A (from calf thymus), leupeptin, phenylmethylsulphonyl fluoride (PMSF), paclitaxel (taxol) and phalloidin from Sigma, USA; cAMP-dependent protein kinase inhibitor (PKI 5-24) from Peninsula Laboratories, Inc., Belmont, CA, USA; assay systems of casein kinase I (CK-1) and casein kinase II (CK-2), protein tyrosine kinase (PTK) from Promega, Madison, WI, USA; assay systems of protein serine/threonine phosphatase (PSP) and protein tyrosine phosphatase (PTP) from New England BioLabs, Beverly, MA, USA. Bay K 8644 and okadaic acid (OA) were generous gifts from Bayer, Leverkusen, Germany, and Dr A. Takai, Nagoya University, respectively. F-actin was prepared by the incubation of G-actin (Sigma) with 3 mM MgATP for 2 h in the basic internal solution or a solution containing 142 mM KCl and 5.5 mM Hepes (pH 7.4).

Preparation of cytoplasmic fractions

Bovine heart ventricle was obtained from a slaughter house, and was homogenized at 4°C in three volumes (v w⁻¹) of a solution containing 20 mM Tris-Hepes (pH 7.2) and various protease inhibitors (starting buffer). The stock solution of protease inhibitors (× 1000 concentration, protease inhibitor cocktail) contained 10 mM benzamidine, 10 mg ml⁻¹ iodoacetamide, 10 mM PMSF, 2 mg ml⁻¹ aprotinin, 5 mg ml⁻¹ leupeptin and 0.5 mg ml⁻¹ pepstatin A (Kameyama, M. et al. 1988, Kameyama, A. et al. 1997). The homogenate was centrifuged at 10 000 g for 20 min and then again at 100 000 g for 30 min. The supernatant fraction was then filtered through a 0.22 μm millipore filter. The filtrate was applied to ion exchange chromatography, using a DEAE-5PW column (8 × 75 mm) pre-equilibrated with the starting buffer (at pH 6.8). The eluted protein at 50-360 mM KCl was concentrated and applied to gel filtration, using a column of Toyopearl HW 60 (26 × 680 mm) pre-equilibrated with the basic internal solution containing protease inhibitors, at a rate of 4.7-6 ml min⁻¹. The fractions were then divided into four fractions (H, P, L₁ and L₂) for the systematic analysis of cytoplasmic factors (Fig. 1).

Figure 1. Protein profile of bovine heart cytoplasm on gel filtration.

Cytoplasm was passed through a column of Toyopearl HW 60 (26 × 680 mm), pre-equilibrated with basic internal solution, at a rate of 4.7 ml min⁻¹. Fractions (10 ml) were collected by a fraction collector. According to their apparent molecular mass, the fractions were grouped to H (apparent molecular mass > 300 kDa), P (250-300 kDa), L₁ (180-250 kDa) and L₂ (< 180 kDa). Continuous line, protein profile; dotted line, molecular mass.

Trypsin treatment of the H fraction

The H fraction (see below) was incubated with trypsin (200 μg ml⁻¹) at room temperature (23-25°C) for 3 h. The trypsin reaction was then stopped by adding the protease inhibitor cocktail.

Purification of calpastatin

Calpastatin from bovine cardiac muscle or the P fraction was purified according to Takano & Murachi (1982). Bovine cardiac muscle (obtained from a slaughter house) was minced and homogenized with a Waring Blendor in four volumes of 20 mM Hepes-Tris buffer (pH 7.0) containing 1 mM EGTA, 1 mM EDTA and 5 mM 2-mercaptoethanol (CS buffer). The homogenate was centrifuged at 10 000 g for 30 min and then at 100 000 g for 60 min. The supernatant or the P fraction (dialysed against CS buffer) was mixed with DEAE-cellulose gel which had been equilibrated with the starting buffer containing 50 mM KCl. Calpastatin was eluted at 150 mM KCl and concentrated by ultrafiltration with a membrane filter (Amicon PM30, Beverly, MA, USA). The concentrated protein was heated to 80°C for 10 min, and the supernatant was chromatographed on a Toyopearl HW60 column (26 × 680 mm). The calpastatin fractions (250-300 kDa) were collected and concentrated with the Amicon ultrafiltration membrane. During the final concentration, the starting buffer was replaced by the basic internal solution for the patch-clamp experiments. Calpastatin activity was measured as the potency to inhibit calpain activity (Takano & Murachi, 1982; Kameyama, M. et al. 1998). The activity of the calpastatin fractions was 5.2-8.3 U (mg protein)⁻¹.

Enzyme assay

Protein kinase A (PKA), CK-1, CK-2 and PTK in crude cytoplasm and cytoplasmic fractions were determined by measuring the incorporation of radioactive phosphate from [γ-³²P]ATP into proteins or specific peptide substrates. The reaction, 20-50 μl in volume, was started by adding 2-5 μl of enzyme preparation, and lasted for 5-10 min at 30°C. For the determination of PKA (Beavo et al. 1974), CK-1 (Agostinis et al. 1989) and CK-2 (Kuenzel & Krebs, 1985), the reaction was terminated by adding 200 μl of 20 % (w/v) TCA. The precipitate of the reaction was rinsed three times with 20 % TCA and solubilized in 100 μl of 50 mM Tris aminomethane (Tris)-HCl solution (pH 7.0), the radioactivity of which was then measured in 2 ml scintillation fluid. The reaction for PTK (Toomik et al. 1992) was terminated by adding 7.5 M guanidine hydrochloride. A 12.5 μl aliquot of the reaction mixture was spotted onto a streptavidin matrix membrane. The membrane was washed with a solution containing 1 % H₃PO₄ (v/v) and 2 M NaCl for 30 s once, for 2 min three times, and with deionized water for 30 s twice, then the radioactivity of the membrane was measured. The assay condition of each kinase reaction was as follows. PKA (mM): Mes, 25; MgCl₂, 5; ATP, 0.1 or 3; dithiothreitol (DTT), 1; cAMP, 0.01; OA, 0.01; 9AC, 1; histone (type II-A mixture), 25 or 50 mg ml⁻¹; and with or without PKA inhibitors, PKI, 1 or 10 μm; or H8, 50 or 100 μM. PKA activity is taken as the component enhanced in the presence of 10 μM cAMP. CK-1 (mM): Tris-HCl, 40 (pH 7.5); MgCl₂, 20; ATP, 0.1; CK-1 peptide substrate, 0.8; and bovine serum albumin (BSA), 0.1 mg ml⁻¹. CK-2 (mM): Tris-HCl, 40 (pH 7.5); MgCl₂, 20; NaCl, 150; ATP, 0.1; CK-2 peptide substrate, 0.05; and BSA, 0.1 mg ml⁻¹. PTK (mM): imidazole hydrochloride, 8 (pH 7.3); EGTA, 0.2; MnCl₂, 1; β-glycerophosphate, 8; MgCl₂, 20; sodium vanadate, 0.125; ATP, 0.025; biotinylated peptide substrate, 2.5; and BSA, 0.1 mg ml⁻¹.

PSP and PTP activity was determined by measuring the release of inorganic phosphate (TCA-soluble radioactivity) from ³²P-labelled myelin basic protein (MyBP) prepared with PKA for PSP (Cohen, 1991) and with Abl PTK (product of the abl gene) for PTP (Tonks et al. 1991) in the presence of [γ-³²P]ATP. The reaction was started by adding 10 μl of ³²P-labelled substrate and incubated for 5-10 min at 30°C, and then terminated by 200 μl of 20 % TCA. After centrifugation at 12 000 g for 10 min, the radioactivity of the supernatant (200 μl) was measured.

Patch-clamp and data analysis

The activity of the L-type Ca²⁺ channel was monitored by the patch-clamp technique in the myocytes superfused with the basic internal solution at 31-35°C using a patch pipette (2-4 MΩ) containing 50 mM Ba²⁺ and 3 μM Bay K 8644, a Ca²⁺ channel modulator. After Ca²⁺ channel activity was recorded in the cell-attached mode, the membrane patch was excised from the cell to establish the inside-out patch mode. For the application of the cytoplasmic fractions, the patch was moved to a small inlet of the perfusion chamber, which was connected to a microinjection system.

Barium currents through the Ca²⁺ channels were elicited by depolarizing pulses from -70 to 0 mV with 200 ms duration at a rate of 0.5 Hz, recorded with a patch-clamp amplifier (EPC-7, List, Darmstadt, Germany), and fed to a computer at a sampling rate of 3.3 kHz. The capacity and leakage currents in the current traces were digitally subtracted. The mean current during the period 5-105 ms after the onset of the test pulses (I) was measured and divided by the unitary current amplitude (i) to yield NP_o (since I = N×P_o×i), where N is the number of channels in the patch and P_o is the time-averaged open-state probability of the channels. A mean NP_o value was measured for 2 min just before excision of the patch (for the cell-attached mode), and for 5 min from 5 min after the application of cytoplasmic fractions (for the inside-out mode). When a patch was broken during this period, the mean NP_o was measured for at least 2 min, otherwise data were discarded. Such cases were 13 % (7 out of 54 cases).

Data are presented as means ±s.e.m. Student's t test was used to estimate statistical significance and a probability value (P) of less than 0.05 was considered to be significant.

RESULTS

P fraction contains a factor

The apparent molecular mass of calpastatin estimated by gel filtration is about 250-300 kDa, therefore the fraction strictly corresponding to this molecular mass (P fraction, see Fig. 1) was first examined in this study. The fraction was supplemented with 3 mM ATP in every experiment, based on our previous finding that the effect of cytoplasm requires ATP (Yazawa et al. 1997). As shown in Fig. 2A, a Ba²⁺ current through the Ca²⁺ channel was first recorded in the cell-attached mode with an NP_o of 0.31 ± 0.30. After excision of the patch in the basic internal solution, channel activity decreased to less than 0.01 in 1 min. After application of P fraction + ATP, Ca²⁺ channel activity was evoked once again. The NP_o value from 5 to 10 min after the application was 0.009 ± 0.049, 2.9 % of that in the cell-attached mode. In six similar experiments, the mean NP_oin the presence of P fraction + ATP in the inside-out mode relative to that in the cell-attached mode was 7.5 ± 1.7 %. Considering that the NP_o in the inside-out mode in basic internal solution was almost zero, we concluded that the P fraction contained a factor capable of maintaining Ca²⁺ channel activity. Nevertheless, the effect of the P fraction was very small compared with the channel activity induced by crude cytoplasm. One possibility to account for this is that the factor in the P fraction alone was not sufficient: a cofactor in addition to the factor in the P fraction might be required for full channel activity. The following experiments were designed to examine this possibility.

Figure 2. Effects of various fractions of cytoplasm on Ca²⁺ channel activity.

The NP_o for each stimulation was calculated and plotted against time. Inside-out patch mode (I.O) was initiated at the time indicated by the arrow. Fractions of the cytoplasm, supplemented with 3 mM MgATP, were applied as indicated by the boxes. A, effects of the P fraction on Ca²⁺ channel activity. Ba, effects of the P fraction + H fraction on Ca²⁺ channel activity. Bb, consecutive current traces in the cell-attached mode (a), in the inside-out patch mode in basic internal solution (b) and in the presence of P fraction + H fraction (c). Sample traces were recorded at the times a, b and c, as shown in Ba. C, effects of the H fraction on Ca²⁺ channel activity. B₁. C. Effects of H fraction on Ca²⁺ channel activity.

H fraction contains a cofactor

We considered that the presumed cofactor would be within the cytoplasm since our previous results suggested that crude cytoplasm prepared from heart tissue could restore ∼90 % of Ca²⁺ channel activity (Kameyama, M. et al. 1988, Kameyama, A. et al. 1997). Therefore, we designed the following experiments. First, by using gel filtration, cytoplasm was separated into four fractions (Fig. 1): H (molecular mass > 300 kDa), P (250-300 kDa), L₁ (180-250 kDa), and L₂ (< 180 kDa). Second, P fraction was combined with other fractions and 3 mM ATP: (1) P + H + ATP, (2) P + L₁+ ATP, (3) P + L₂+ ATP. Then, the effect of each combination was checked in the patch-clamp experiments. We expected that the combination which could recover Ca²⁺ channel activity with an efficacy similar to that of crude cytoplasm would contain the predicted cofactor.

The combination of P + H + ATP was found to be effective. One of the experiments with P + H + ATP is illustrated in Fig. 2Ba and Bb. In the cell-attached mode, NP_o was 0.50 ± 0.06 and declined in the inside-out mode to 0.086 ± 0.035 (1 min, before addition of cytoplasmic fractions), only 17 % of that in the control cell-attached mode. After the application of P + H + ATP, channel activity recovered gradually and reached 0.41 ± 0.46 5 min after the application, 81 % of control. In eight similar experiments, the relative NP_o value in the inside-out mode in the presence of P + H + ATP was 86 ± 23 %. This efficacy was similar to that for crude cytoplasm (∼90 %), suggesting that a cofactor was present in the H fraction. The stabilizing effect was also seen in the absence of Bay K 8644 in the pipette solution (87 ± 43 %, n = 3), suggesting that Bay K 8644 was not required to reverse run-down.

The effects of the other combinations were also studied. The relative NP_o was 7.9 ± 0.4 % (n = 3) and 7.7 ± 1.5 % (n = 3) in the inside-out mode in the presence of P + L₁+ ATP and P + L₂+ ATP, respectively. These values were not significantly different from that for P + ATP (P > 0.05, for both), implying that neither L₁ nor the L₂ fraction had any effect on Ca²⁺ channel activity.

The effect of the H fraction alone was then examined as a control. As demonstrated in Fig. 2C, in the presence of H + ATP in the inside-out mode, the NP_o was 0.002 ± 0.014, 1.3 % of that in the cell-attached mode (0.15 ± 0.03). In a total of ten experiments, the relative NP_o was 5.8 ± 2.0 % of control, suggesting that, although the H fraction enhanced the effect of the P fraction, this fraction by itself had negligible ability to recover channel activity.

Figure 3 is a summary of the above results, in which the effect of P + H + ATP was much higher than the sum of P + ATP and H + ATP, implying that there is synergism between the factors in the P and H fractions. Therefore we concluded that, in addition to ATP, at least two cytoplasmic factors are required for the Ca²⁺ channel activity, one existing in the P fraction, and the other in the H fraction. The following experiments were designed to investigate the nature of these factors.

Figure 3. Recovery of Ca²⁺ channel activity in the inside-out patch mode by various fractions of cytoplasm.

Data are presented as means and s.e.m. The number of experiments was six for P, ten for H, eight for P + H, three for P + L₁ and three for P + L₂. All the fractions were supplemented with 3 mM MgATP.

Calpastatin may be the factor in the P fraction

Our previous report suggested that the physicochemical characteristics of the P fraction were similar to those of calpastatin, the endogenous inhibitor of calpain (Kameyama, M. et al. 1998). In order to test the hypothesis that calpastatin may be the effective component of the P fraction, we examined the effect of calpastatin + H + ATP. Figure 4Aa and Ab is an example of such an experiment, in which the NP_o of the Ca²⁺ channel was 0.75 ± 0.08 in the cell-attached mode and decreased to 0.23 ± 0.09 within 0.5 min following the inside-out patch formation. The application of calpastatin + H + ATP evoked the channel activity once again. The channel activity 5 min after the application was 0.30 ± 0.40, 40 % of that in the cell-attached mode. Four experiments produced a relative channel activity of 87 ± 35 %. This value was not significantly different from that for the effect of P + H + ATP (86 %), indicating that calpastatin may be the effective component of the P fraction.

Figure 4. Effect of calpastatin with or without the H fraction of the cytoplasm on Ca²⁺ channel activity.

The NP_o value for each stimulation was calculated and plotted against time. Inside-out patch mode (I.O) was initiated at the time indicated by the arrow. Calpastatin (purified from the P fraction) with (Aa and Ab) or without (B) the H fraction, supplemented with 3 mM MgATP, was applied as indicated by the boxes. Ab, consecutive current traces in the cell-attached mode (a), in the inside-out patch mode in basic internal solution (b), and in the presence of calpastatin + H fraction (c and d). Sample traces were recorded at the times a, b, c and d as indicated in Aa.

Biphasic action of calpastatin

As a control, the effect of calpastatin + ATP was investigated. We found that calpastatin seemed to act in a biphasic manner comprising an early transient phase and a late phase. As demonstrated in Fig. 4B, when calpastatin + ATP was applied following excision of the patch into the basic internal solution, the channel activity was promptly evoked. In the first 1.5 min, the NP_o was 0.42 ± 0.05, 46 % of that in the cell-attached mode (0.92 ± 0.07). Nevertheless, in a manner different to that observed in the recovery by calpastatin + H + ATP or P + H + ATP, channel activity did not increase further but decreased progressively. From 5 to 10 min after application, the NP_o was reduced to 0.029 ± 0.061, 3.6 % of that in cell-attached mode. A similar biphasic action was also seen in four other experiments and in some experiments with P + ATP. In five experiments of calpastatin + ATP, the transient phase started within 0.2 min after application and lasted for 1.5 to 7 min, with a mean of 3.5 ± 0.9 min. Therefore, the NP_o measured in the first 1.5 min was taken to represent the channel activity of the transient phase, and that 5 min after the application of the cytoplasmic fractions was considered to represent the channel activity of the late phase. Figure 5 summarizes those measurements. In the presence of calpastatin + ATP, a relatively higher recovery was seen in the transient phase, 38 ± 5 % of channel activity in the cell-attached mode. However, the activity was 11 ± 6 % in the late phase (P < 0.05). For calpastatin + H + ATP, although the recovery (66 ± 43 %, n = 4) was not significantly different from that for calpastatin + ATP (P > 0.05) in the transient phase, it was greater (P < 0.05) in the late phase (87 %). A more pronounced tendency was observed for P + H + ATP: the relative channel activity was small in the transient phase (3.8 ± 2.6 %, n = 4), while it was as high as that for calpastatin + H + ATP in the late phase. The results in Fig. 5 indicate that the recovery induced by calpastatin + ATP was transient and less potent, whereas the recovery induced by calpastatin + H + ATP or P + H + ATP was long-lasting and more potent, implying that calpastatin alone was not sufficient to maintain Ca²⁺ channel activity and that the addition of the H fraction is required for the full activity. The recovery induced by calpastatin + H + ATP or P + H + ATP could last until the patch rupture, with the longest recorded recovery of 37 min.

Figure 5. Biphasic action of calpastatin to reverse Ca²⁺ channel run-down.

The effects of calpastatin, calpastatin + H fraction, and P fraction + H fraction to reverse Ca²⁺ channel run-down during the transient phase () and during the late phase (▪) are presented. Data are expressed as means and s.e.m. Student's t test showed whether the results were significantly different (*P < 0.05, **P < 0.01) or not significantly different (ns) for the combinations joined by brackets. The number of experiments was five for calpastatin, four for calpastatin + H and four for P + H. All the test solutions were supplemented with 3 mM MgATP.

Some characteristics of the H factor

In order to find the effective component of the H fraction (H factor), we carried out some experiments to characterize this factor. When this fraction was heated to 100°C for 5 min, the effect of the H fraction was abolished. As shown in Fig. 6A, Ca²⁺ channel activity was 0.69 ± 0.13 in the cell-attached mode, and 0.56 ± 0.09 in the following inside-out mode. When heat-treated H fraction + P fraction + ATP was applied, a marked recovery was not observed. However, the following application of control H + P + ATP evoked an activity which reached 0.28 ± 0.03 (time 14-16 min in Fig. 6A), which was 42 % of that in the cell-attached mode. Four similar experiments showed a mean channel activity in the presence of heat-treated H + P + ATP of 3.5 ± 1.6 %, which was significantly lower than that of control H + P + ATP (P < 0.05), suggesting that the H factor was heat sensitive. Similar results were also obtained when the H fraction was heated to 65°C for 15 min (n = 2, data not shown).

Figure 6. Characteristics of the H fraction of cytoplasm.

The NP_o value for each stimulation was calculated and plotted against time. Inside-out patch mode (I.O) was initiated at the time indicated by the arrow. The H fraction (heat-treated or trypsin-treated) + P fraction, supplemented with 3 mM MgATP, was applied as indicated by the boxes. Aa, effect of P fraction + heat-treated H fraction of the cytoplasm on Ca²⁺ channel activity. Ab, consecutive current traces in the cell-attached mode (a), in P fraction + heat-treated H fraction (b) and in the presence of P + H fractions (c). Sample traces were recorded at the times a, b and c, as indicated in Aa. B, effect of P fraction + trypsin-treated H fraction of cytoplasm on Ca²⁺ channel activity.

The heat sensitivity of the H fraction implied that the H factor was likely to be a protein, therefore we examined whether this factor was sensitive to protease. One of the experiments is shown in Fig. 6B, in which Ca²⁺ channel activity, after the formation of an inside-out patch, fell to 0.13 ± 0.04 within 1 min from 1.29 ± 0.12, 10 % of that in the cell-attached mode. After run-down, trypsin-treated H fraction + P fraction + ATP was applied. Although a transient recovery was evoked in the first 1.5 min (29 % of control), it decreased and stayed at a relatively low level. From 5 to 10 min after application, the NP_o was 0.069 ± 0.114, 5.3 % of that in the cell-attached mode. The mean relative channel activity in seven experiments was 6.4 ± 1.5 %, which was significantly lower (P < 0.01) than that of H + P + ATP (86 %). As in this experiment, the proteolysis by trypsin was stopped by trypsin inhibitors, the effect of H + trypsin inhibitors + P + ATP was also examined as a control. The control experiment showed that the relative channel activity was 139 ± 88 % (n = 4), suggesting that the trypsin inhibitors had a negligible effect on channel activity. These results suggested that the factor in the H fraction, which restored channel activity, was heat and trypsin sensitive.

Activity of protein kinases and phosphatases in the H fraction

To characterize the factor in the H fraction further, we determined the activity of several protein kinases and phosphatases.

PKA activity was detected in both the crude cytoplasm and H fraction, but not much was detected in the P fraction (Fig. 7A) or in the L₁ and L₂ fractions (not shown). The PKA activity was blocked completely by 1 or 10 μM PKI (a specific PKA inhibitor) in crude cytoplasm, and to 82 % by 1 μM PKI and completely by 10 μM PKI in the H fraction (Fig. 7A). The PKA activity in crude cytoplasm and the H fraction was also blocked completely by H8 (50 or 100 μM), a non-specific protein kinase inhibitor (not shown). PKA activity was also determined in the presence of 3 mM ATP. The results were essentially similar to those with 100 μM ATP, and PKA activity was blocked completely by 10 μM PKI and 50 μM H9 (not shown). Basal PKA activity, measured as PKI-inhibitable kinase activity in the absence of cAMP, was not detectable in crude cytoplasm and its fractions.

Figure 7. Activity of protein kinases and protein phosphatases.

A, cAMP-dependent protein kinase (PKA) activity and the block by PKI in crude cytoplasm, H and P fractions. B, casein kinase I (CK-1), casein kinase II (CK-2) and protein tyrosine kinase (PTK) activities in crude cytoplasm and cytoplasmic fractions as indicated. C, protein serine/threonine phosphatase (PSP) activity and the block by 10 μM okadaic acid (OA) or 1 mM anthracene-9-carboxylic acid (9AC) in crude cytoplasm and cytoplasmic fractions as indicated. D, protein tyrosine phosphatase (PTP) activity in crude cytoplasm and cytoplasmic fractions as indicated. Data are presented as means and s.e.m. The number of experiments was 4-5.

Figure 8. Effects of cytoskeleton-modifying agents on the run-down of Ca²⁺ channel activity.

A, effects of phalloidin on the run-down of Ca²⁺ channel activity. The NP_o value for each stimulation was calculated and plotted against time. Inside-out patch mode (initiated at the time indicated by I.O and the arrow) was made in basic internal solution containing 2 μM phalloidin. B, effects of paclitaxel on the run-down of Ca²⁺ channel activity. The NP_o value was plotted as in A. Inside-out patch mode (I.O, arrow) was made in basic internal solution containing 2 μM paclitaxel. C, effect of F-actin on the run-down of Ca²⁺ channel activity. The NP_o was plotted as in A. Inside-out patch mode (I.O) was initiated at the time indicated by the arrow. F-actin (1 mg ml⁻¹), supplemented with P fraction + 3 mM MgATP, was applied as indicated by the box.

The activities of CK-1, CK-2 and PTK were also measured. As shown in Fig. 7B, the activities of CK-1 and CK-2 in the H fraction were low, whilst in the L₁ fraction CK-1 and CK-2 activities were high. The PTK activity was highest in the P fraction and intermediate in the H fraction (Fig. 7B).

Protein phosphatases were also measured. As shown in Fig. 7C, the activity of PSP in the H and P fractions was relatively high compared with that in the L₁ and L₂ fractions. In the H fraction, 65 % of the total PSP activity was sensitive to 10 μM OA, a blocker of type 1 and type 2A phosphatase, and 33 % was sensitive to 1 mM 9AC, a blocker of a non-identified phosphatase (Zhou et al. 1997), whereas, in the P fraction, 95 % of the total PSP was sensitive to OA but only 4.3 % was sensitive to 9AC. Fig. 7D shows the activity of PTP. The H, P and L₁ fractions contained similar activities of PTP, whilst the L₂ fraction contained a lower activity.

Actin is unlikely to be the effective component of the H fraction

In the following experiment, we also examined the possibility that a component of the cytoskeleton might be the factor in the H fraction.

We first investigated whether the time course of run-down was slowed by phalloidin, a stabilizer of actin and microfilaments. As shown in Fig. 7A, channel activity (0.44 ± 0.05) was first recorded in the cell-attached mode in the basic internal solution containing 2 μM phalloidin, and then the patch was excised. The run-down, however, was neither prevented nor slowed. The NP_o decreased to 0.025 ± 0.007 (1-2 min after the formation of an inside-out patch), 5.7 % of that in the cell-attached mode. Although short openings of the channel were seen occasionally, this phenomenon was also observed in the absence of phalloidin (data not shown). In four experiments, the mean channel activity in the second 1 min after the onset of the inside-out patch was 12.2 ± 8.7 %. This value was not significantly different from that without phalloidin (5.3 ± 1.5 %), suggesting that breakdown of actin filaments was not responsible for the progress of run-down of the Ca²⁺ channel in these conditions. Similar results were also obtained even when the concentration of phalloidin was increased to 20 μM (n = 2).

We then studied the effects of paclitaxel (taxol), a stabilizer of microtubules. As shown in Fig. 7B, run-down was even accelerated in the internal solution containing 2 μM paclitaxel. Channel activity was diminished within 0.3 min after the formation of an inside-out patch in the paclitaxel-containing solution. In three other experiments, channel activity was also diminished within 0.5 to 1 min after patch excision, indicating that the breakdown of microtubules was also unlikely to be involved in the run-down in these conditions.

Considering that the P fraction (or calpastatin) and ATP are necessary to maintain Ca²⁺ channel activity in the inside-out mode, we examined the effect of F-actin + P fraction + ATP. Figure 7C is an example of experiments in which F-actin (1 mg ml⁻¹) + P fraction + 3 mM ATP was applied after run-down of the channels in the basic internal solution. Channel activity was evoked 2.5 min after application, but at a low level. From 5 to 7 min after the application, the channel activity was 0.001 ± 0.007, 2.8 % of that in the cell-attached mode (0.36 ± 0.04). Four of these experiments gave a mean relative NP_o of 5.9 ± 5.7 %. This low-level recovery of channel activity was not different from that seen with the P fraction + ATP. Therefore, the possibility of direct involvement of actin breakdown in the run-down of the Ca²⁺ channel was not supported.

DISCUSSION

In the present study, we have confirmed the previous finding that the run-down of cardiac L-type Ca²⁺ channel activity is reversed by fractions of bovine cardiac tissue extract, supporting the view that basal activity of the Ca²⁺ channel is maintained by cytoplasmic factors.

In this study, we compared the effects of calpastatin and the P fraction of cytoplasm (molecular mass 250-300 kDa on gel filtration) and found that: (1) both calpastatin and the P fraction restored Ca²⁺ channel activity after run-down with a similar potency, i.e. about 10 % of that seen before run-down; and (2) the restorative effects were enhanced to about 90 % in the presence of the H fraction. These results indicate that the action of the P fraction on calcium channel activity is mimicked by calpastatin, suggesting that calpastatin may be the effective component of the P fraction. The previous finding that the effect of calpastatin is lower than that of crude cytoplasm (Kameyama, M. et al. 1998) may be explained by the absence of the H fraction. Therefore, we think that calpastatin has the ability to maintain Ca²⁺ channel activity.

Calpastatin is the endogenous inhibitor of calpain, a Ca²⁺-activated neutral protease, found in a wide variety of tissues (Suzuki et al. 1988; Murachi, 1989; Melloni et al. 1992). The mechanism by which calpastatin prevents the run-down of Ca²⁺ channels may be considered to be: (1) inhibition of proteolysis by calpain (Romanin et al. 1991), or (2) direct interaction with the Ca²⁺ channel, based on the findings of the reversibility and repeatability of channel activity induced by calpastatin or crude cytoplasm (Kameyama, M. et al. 1988, 1990; Kameyama, A. et al. 1997, 1998). Recently, it has been reported that the effect of calpastatin is not mimicked by synthetic calpain inhibitors (Seydl et al. 1995). This finding suggests that a simple inhibition of calpain is not responsible for the mechanism of prevention of run-down. In this study, we have confirmed that calpastatin together with ATP can induce a relative channel activity of about 40 % in the early transient phase and about 10 % in the late phase in the inside-out patch mode. We have also found that, even without calpastatin, channel activity could be observed, although at a low level, in the presence of the H fraction and ATP and even occasionally in the basic internal solution. These results suggest that, at least in our experimental conditions, the channels are not in an irreversibly inactivated state but in a silent (low-active) state during run-down. Considering the above findings together with others, we conclude that the action of calpastatin on the Ca²⁺ channel is not related to inhibition of proteolysis. A simple scheme for the mechanism may be that calpastatin interacts directly with the Ca²⁺ channel.

The most important result of this study is the effect of the H fraction. The H fraction itself has little ability to restore Ca²⁺ channel activity even in the presence of ATP (about 6 % relative NP_o), whereas it shows a great ability when applied together with calpastatin (about 90 % relative NP_o). This suggests that there is some factor in the H fraction which affects cardiac Ca²⁺ channel activity. This finding explains why this fraction was discarded during the former systematic analysis of the cytoplasmic fractions, and why the calpastatin containing fraction showed a significantly increased ability to restore channel activity when the molecular mass covered by the fraction was increased in the high molecular mass direction (Kameyama, A. et al. 1998). We have also examined the effect of P + L₁ or L₂, and noted little effect on channel activity, suggesting that neither fraction L₁ nor L₂ is likely to contain any Ca²⁺ channel activity regulator.

The heat and trypsin sensitivity suggest that the H factor may be a protein. To characterize the effective component of the H fraction further, we measured the activity of protein kinases (PKA, CK-1, CK-2 and PTK) and protein phosphatases (PSP and PTP) in the H and other fractions. It was found that the activity of PKA in the H faction was high whereas the activities of CK-1, CK-2 and PTK were low. The activities of PSP and PTP detected in the H and P fractions were nearly equal. These results suggest that CK-1, CK-2, PTK, PSP and PTP are not likely to be the effective components of the H fraction. Amongst those investigated, PKA seems to be the only possible candidate for the effective component of the H fraction. However, PKA in the cytoplasm used in this study is totally inactive unless cAMP is elevated. Furthermore, our previous study has revealed that: (1) crude cytoplasm can reverse run-down of Ca²⁺ channel activity even in the presence of PKI (1 μM) and H8 (50 or 100 μM); (2) MgATP can be replaced by non-hydrolysable ATP analogues (AMP-PNP and AMP-PCP) or K₂ATP; and (3) cytoplasm from kidney, which contains PKA activity similar to that from heart, has a minimal effect on the channel (Kameyama, A. et al. 1997; Yazawa et al. 1997). Thus, it is difficult to think that PKA was the effective component of crude cytoplasm. However, at present we cannot exclude the possibility that the effective component of the H fraction is some other protein kinases or phosphatases. It is noted that an OA-insensitive and/or 9AC-sensitive component of PSP activity is present in the H fraction. Whether or not this component contributes to regulation of the Ca²⁺ channel has still to be determined.

The H factor has a large molecular mass (> 300 kDa). Therefore we examined the possibility that cytoskeleton might be the effective component of the H fraction in restoring Ca²⁺ channel activity. It has been reported that the cytoskeleton is involved in the regulation of Na⁺ channels in epithelial cells (Cantiello et al. 1991) and central neurons (Srinivasan et al. 1988), Ca²⁺ channels in neurons (Fukuda et al. 1981; Johnson & Byerly, 1994), a K⁺ channel in renal cortical collecting ducts (Wang et al. 1994) and a Cl⁻ channel in renal cortical collecting ducts (Schwiebert et al. 1994). It has also been reported that run-down of the ATP-sensitive K⁺ channel in cardiac myocytes is related to depolymerization of F-actin, the main constituent of microfilaments (Furukawa et al. 1996). However, our experiments show that the run-down of cardiac Ca²⁺ channel activity is neither prevented by phalloidin, a microfilament stabilizer, nor restored by F-actin even in the presence of calpastatin-containing P fraction and ATP. Therefore, it is suggested that actin is not likely to be the factor in the H fraction. The effect of taxol, a stabilizer of microtubules, on run-down is also negligible, indicating that microtubules are also probably not the effective component of the H fraction.

Another possible candidate for the H factor may be the β-subunit of the Ca²⁺ channel. This subunit has been found to enhance the current amplitude in whole-cell recording (e.g. Waard & Campbell, 1995) and the NP_o in single channel recording (Wakamori et al. 1993). However, it has been reported that the binding between the α₁- and β-subunit of skeletal muscle is so tight that they cannot be separated by a mild detergent such as digitonin (Takahashi et al. 1987). Thus it is difficult to imagine that the Ca²⁺ channels release β-subunits during run-down of the channels. Recently we have found that Ca²⁺ channels expressing α₁-subunits alone show run-down in the inside-out patch mode, and that this run-down is reversed by the application of cytoplasm. Thus, it is unlikely that the β-subunit is the factor in the H fraction (Hao et al. 1998).

In our preliminary experiments, treatment of the H fraction with phospholipase A₂ reduced the effect of the H fraction on the channel activity (L.-Y. Hao & A. Kameyama, unpublished observations). The possibility that a lipoprotein might be involved in the action of the H factor remains to be examined.

Based on the previous and present findings, it is suggested that run-down of cardiac Ca²⁺ channel activity can be reversed by at least three factors: a protein-like factor in the H fraction, calpastatin (the factor in the P fraction) and ATP. It should be noted that the action of calpastatin is not the inhibition of proteolysis of the channel, and that the action of ATP is not phosphorylation of the channel protein. Thus the exact mechanism by which these factors regulate Ca²⁺ channel activity is not clear. Some further findings may be relevant for considering the molecular mechanism of the factors. (1) The factor in the H fraction potentiates the effect of calpastatin in a synergistic manner. (2) The action of calpastatin is composed of two phases: a transient phase and a late phase. We propose a possible mechanism in which these three factors bind to the channel itself or to a protein close to the channel and interact with the channel. However, the binding may not be as tight as that between the subunits of the channel. Thus when a patch is excised from the host cell in a basic solution, these factors would be washed out, resulting in run-down of the channel. It is further speculated that the washout of the H factor might be slower than that of calpastatin. If so, this may explain the relatively higher restoration of channel activity by calpastatin in the early transient phase. In the late phase, the H factor is washed out extensively, and thereby the restorative effect of calpastatin alone on the channel activity is decreased. However, when calpastatin is applied together with the H fraction, the channel activity is restored to a great extent, even in the late phase. We think that this hypothesis can also explain why the channel activity is prevented from run-down by calpastatin + ATP (without the H factor) when an inside-out patch is made in the calpastatin-containing internal solution (Romanin et al. 1991), but restored only partially when applied after complete run-down (Kameyama, M. et al. 1998).

The results in the present study together with our previous studies suggest that the basal activity of cardiac L-type Ca²⁺ channels is maintained by factors in the cytoplasm and ATP. The concentration of ATP is dependent on the metabolic state of the cell, and calpastatin, the putative factor in the P fraction, is known to translocate to the membrane upon elevation of cellular Ca²⁺ and phosphorylation (for review see Murachi, 1989; Melloni et al. 1992). Our preliminary study suggests that the presumed H factor is possibly associated with the membrane. Taken together, it is possible that the novel mechanism for maintaining the Ca²⁺ channel activity plays some role in the modulation of the channel activity in physiological and/or pathophysiological conditions. On the other hand, the Ca²⁺ channels are regulated by phosphorylation with protein kinases and by a direct interaction with G-proteins (for review see McDonald et al. 1994). Although the relationship between the action of the cytoplasmic factors and the regulation by phosphorylation or G-proteins is poorly understood, it would constitute the complicated linkage between the Ca²⁺ channel activity and the cellular metabolism and signalling.