INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

PHOSPHORYLATION of cardiac regulatory proteins plays an important role in the regulation of cardiac contractility and mediating inotropic and relaxant effects of -adrenergic catecholamines. In the intact isolated heart, -adrenergic stimulation leads to phosphorylation of phospholamban (PLB) on serine-16 by cAMP-dependent protein kinase and on threonine-17 by Ca²⁺/calmodulin-dependent protein kinase II (Ca²⁺/CaM II) (27). PLB is an intrinsic protein of the sarcoplasmic reticulum (SR) and inhibits the function of SR Ca²⁺-ATPase. Stimulation of -adrenoceptors increases the phosphorylation state of PLB and thereby enhances Ca²⁺-ATPase activity. This augments the rate of SR Ca²⁺ uptake, and intracellular Ca²⁺ declines faster. Moreover, the SR can release more Ca²⁺ for the subsequent contraction. These steps contribute to the inotropic and relaxant effects of -adrenergic stimulation in the mammalian heart (25). In addition, stimulation of -adrenoceptors increases the phosphorylation state of the inhibitory subunit of troponin (TnI), which is localized in the thin filaments of cardiac contractile apparatus. Both, PLB phosphorylation (PLB-P) and TnI phosphorylation (TnI-P) can hasten cardiac relaxation (28).

The two phosphorylation sites of PLB are substrates for type 1 and 2A protein phosphatases (PP) in vitro. PP1 and PP2A could dephosphorylate PLB in membranes from the rabbit heart (phosphorylated in vitro on serine or threonine), whereas PP2B and PP2C were nearly inactive (13). Similarly, PP1 and PP2A, separated from human ventricles, dephosphorylate recombinant PLB (20).

Cell membrane-permeant PP inhibitors like okadaic acid, cantharidin, and calyculin A increase the phosphorylation state of cardiac regulatory proteins independently of receptor activation and without increasing intracellular cAMP content (15, 16, 19). Cantharidin increased the force of contraction and hastened relaxation in isolated electrically driven guinea pig papillary muscles and enhanced the phosphorylation state of PLB and TnI in ³²P-labeled guinea pig ventricular cardiomyocytes (19). Cantharidin inhibited PP2A purified from guinea pig ventricles ~10-fold more potently than PP1 (19). In our previous work, however, we did not study the site-specific phosphorylation of PLB. To study the correlation between inotropic function (under isometric conditions) and protein phosphorylation in more detail, it is necessary to measure both parameters in the very same preparations. Here, a couple of limitations need to be considered. Regrettably, it is not feasible to label papillary muscles with ortho[³²P]phosphoric acid to perform phosphoamino acid analysis on contracting papillary muscles. On the other hand, in isolated ³²P-labeled myocytes, defined isometric contractions cannot be measured. One can measure force of contraction and phosphorylation in ³²P-labeled isolated Langendorff-perfused hearts (17, 18). However, those preparations are not contracting under isometric but auxotonic conditions. They do not perform work and do not act at defined preload and afterload. Finally, at concentrations where cantharidin (or okadaic acid) exerts a positive inotropic effect on cardiomyocytes, it also constricts coronary arteries (8). Hence, in perfused hearts, a PP inhibitor would affect the coronary perfusion at positive inotropic concentrations. This would make the interpretation rather impossible. Because of these limitations, the effects of PP inhibitors and the functional role of phosphatases cannot be easily studied in Langendorff preparations. In the past, we and others have used perfused hearts because this preparation provides much more ventricular tissue for the subsequent biochemical assays than papillary muscles (500 vs. 5 mg). Fortunately, the necessity to use whole ventricles to measure phosphorylation in contracting preparations has been overcome when phosphorylation-specific antibodies for PLB (5) and TnI (1, 2) became available. Here, we adapted the method to samples as small as papillary muscles. This offered us the unique possibility to measure isometric contractility (developed force, time to peak tension, and time of relaxation) in preparations in which we studied PLB-P and TnI-P. This enabled us to elucidate the role of PP (by means of an inhibitor) in PLB-P and TnI-P and their mechanical correlation in greater detail.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Contraction experiments in isolated perfused hearts. Rats and guinea pigs of either sex (250-300 g; Harlan Winkelmann, Borchen, Germany) were heparinized (500 IU heparin sodium ip; Braun Melsungen, Melsungen, Germany). The chest was opened, and hearts were rapidly excised and submerged in ice-cold Krebs-Henseleit buffer. The ascending aorta was cannulated, connected to a modified Langendorff apparatus, and perfused at a constant flow of 10 ml/min (21) with Krebs-Henseleit buffer consisting of (in mM) 118 NaCl, 25 NaHCO₃, 2.5 CaCl₂, 4.7 KCl, 1.2 KH₂PO₄, 1.2 MgSO₄, and 11 glucose. The buffer was gassed with 95% O₂-5% CO₂ at 37°C (pH = 7.4). The temperature of the perfusion buffer, measured at the aortic cannula, was maintained at 37°C, and hearts were contained in a heating jacket at the same temperature and electrically driven (frequency 4.5 Hz; duration 5 ms). After a 60-min equilibration period, hearts were perfused with isoproterenol (1 µM for 5 min) or cantharidin (100 µM for 30 min). At the end of the perfusion, ventricular tissue was immediately freeze-clamped using aluminium tongs precooled in liquid nitrogen and then pulverized in liquid nitrogen (4). Samples were stored at 80°C for further analysis.

Contraction experiments in papillary muscles. Experiments were performed in electrically driven (frequency 1 Hz; duration 5 ms; intensity 20% greater than threshold) papillary muscles from guinea pig right ventricles (3). Preparations were isolated, mounted, and individually suspended in glass tissue chambers for recording isometric contractions. The bathing solution (10 ml) was a modified Tyrode solution containing (in mM) 119.9 NaCl, 5.4 KCl, 1.8 CaCl₂, 1.05 MgCl₂, 0.42 NaH₂PO₄, 22.6 NaHCO₃, 0.05 Na₂EDTA, 0.28 ascorbic acid, and 5.0 glucose. It was continuously gassed with 95% O₂-5% CO₂ and maintained at 35°C (pH = 7.4). Force of contraction was measured by an inductive force transducer, and the time course of individual isometric contractions was calculated from three subsequent twitches at high chart speed. The onset of time to peak tension was measured at 10%, and the endpoint of relaxation time was set at 10% of peak force. The basal force of contraction of all papillary muscles included in the study amounted to 1.38 ± 0.05 mN (n = 95). There were no differences in the predrug values in the experiments in the concentration-response and time course studies (1.37 ± 0.06 mN, n = 56, vs. 1.40 ± 0.08 mN, n = 39). This is also in good agreement with our earlier studies dealing with the effects of PP inhibitors (15, 19). The concentration-response curves were performed in a noncumulative manner, and papillary muscles were incubated with only a single concentration of cantharidin for 30 min. At the end of the incubation, papillary muscles were immediately frozen using aluminium tongs precooled in liquid nitrogen (4). The length and diameter of papillary muscles were not measured in order to freeze-clamp as close as possible to the given time point. Any delay due to the microscopical examination might affect the protein phosphorylation. Samples were stored at 80°C for further analysis.

Phosphorylation of guinea pig ventricular cardiomyocytes. Guinea pig ventricular cardiomyocytes were isolated as described (15). For phosphorylation experiments, adenosine deaminase (10 U/ml) was added to avoid interference from endogenous adenosine on treatment (6). The drug solution (150 µl) was preincubated at 37°C for 2 min before mixing with 150 µl of the diluted cardiomyocytes and kept at 37°C. The reaction was stopped by adding 150 µl of SDS stop solution (11). Samples were frozen at 80°C for further analysis.

Measurement of cytosolic Ca²+. Isolated guinea pig ventricular cardiomyocytes were placed in a small bath (300 µl) on the stage of a modified inverted microscope (Diaphot 200, Nikon, Tokyo, Japan) in Tyrode solution containing the cell permeant indo 1-acetoxymethyl ester (25 µM) and 2.5% of the nonionic surfactant Pluronic F-127 (both from Molecular Probes, Eugene, OR). After a incubation period of 3 min at room temperature, the cells were perfused with Tyrode solution (0.8 ml/min) for at least 20 min to allow the washout of the extracellular dye and intracellular indo 1 deesterfication. Cardiomyocytes were electrically stimulated via platinum wire electrodes at 1 Hz. The emission field was restricted by a square window containing only the cell under study. Background fluorescence was obtained as bath fluorescence from the same emission field as used to record the cell fluorescence. Both loading of indo 1 into the myocytes and the experiments were done at room temperature to minimize loss of the Ca²⁺ indicator from the cells. Intracellular Ca²⁺ was recorded from a single cardiomyocyte using a dual-emission microfluorescence system (PTI, Princeton, NJ). The 75-W xenon lamp output passed through a monochromator (365 ± 5 nm) and was reflected by a dichroic mirror (DM 385). The light beam was focussed onto the cells through a ×40 oil-immersion objective (CF Fluor NA 1,3, Nikon). The emitted light was transmitted by the dichroic mirror and split by a second mirror (DM 455) in the two emission wavelengths (405 and 495 nm) before it reached two photomultiplier tubes. The 405-to-495-nm ratio was used as an index of intracellular free Ca²⁺ concentration. The fluorescence data were acquired at 20 Hz and processed via an analog-to-digital converter. Data acquisition and processing were performed by a software (FeliX, version 1.1, PTI) for intracellular Ca²⁺ measurement.

Gel electrophoresis and Western blotting. Protein extracts were prepared as described (14), and SDS buffer was added according to Laemmli (11). The samples were heat treated for 10 min at 30°C. Homogenate sample proteins (30 µg) were loaded per lane. Electrophoresis was run using 8% polyacrylamide separating gels. Gels were initially run at 40 mA per gel for 30 min, and the current was then increased to 60 mA. Proteins were electrophoretically transferred to nitrocellulose membranes (BA 85, pore size 0.45 µM, Schleicher & Schuell, Dassel, Germany) in 50 mM of sodium phosphate buffer (pH 7.4) 180 min at 1.5 Å at 4°C. Membranes were then treated with Tris-buffered saline containing 2.0% bovine serum albumin for 30 min and incubated overnight at 4°C with polyclonal antibodies directed against the PLB peptide (residues 9-19) phosphorylated at serine-16 or threonine-17 (PLB-Ser-16 and PLB-Thr-17, respectively) (PhosphoProtein Research, Bardsey, UK; see Ref. 5) and subsequently with ¹²⁵I-labeled protein A (ICN Biomedicals, Eschwege, Germany). The specificity of phosphorylation-specific antibodies for PLB was tested in membrane vesicles from guinea pig ventricles. The membrane vesicles were completely dephosphorylated by potato acid phosphatase and thereafter backphosphorylated by exogenous PKA or Ca²⁺/CaM-dependent protein kinase, respectively. Fittingly, the antibody directed against PLB-Ser-16 provided signals only in the samples pretreated with PKA. On the other hand, the PLB-Thr-17 antibody recognized PLB only in the membrane vesicles phosphorylated by Ca²⁺/CaM-dependent protein kinase (data not shown). Moreover, the phosphorylation-specific antibodies for PLB have been extensively characterized previously by Drago and Colyer (5). Recently, Kuschel et al. (10) also confirmed the specificity of PLB-Ser-16 and PLB-Thr-17 antibodies by means of competitive immunoblotting using monophosphorylated and diphosphorylated PLB peptides. To exclude possible differences due to unequal protein loading, the blots were probed with a polyclonal antibody against calsequestrin (7) and thereafter with ¹²⁵I-labeled protein A. Radioactive bands were visualized and quantified using PhosphorImager (Molecular Dynamics, Sunnyvale, CA). The signals obtained from the phosphorylation site-specific PLB antibodies were normalized to the corresponding calsequestrin values.

Data analysis. Data shown are means ± SE. Statistical significance was estimated with Student's t-test for paired or unpaired observations. The correlations were calculated using linear correlation coefficients. A P value of <0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

To validate the measurement of phosphorylation using specific antibodies, we performed experiments in isolated electrically driven hearts perfused according to Langendorff. In initial experiments, we used rats because the antibodies were first characterized in rat cardiac preparations (5). However, to facilitate comparison with our previous work (15, 16, 19), we then switched to guinea pig cardiac preparations (isolated hearts, isolated cardiomyocytes, and papillary muscles).

In isolated perfused rat hearts, the -adrenoceptor agonist isoproterenol (1 µM for 5 min) increased PLB phosphorylation at serine-16 to 215 ± 30% of control and at threonine-17 to 365 ± 100% of control, respectively (n = 5, P < 0.05). The PP inhibitor cantharidin (100 µM for 30 min) induced comparable increases of PLB phosphorylation. The effects amounted to 192 ± 20% of control for serine-16 and 366 ± 66% of control for threonine-17, respectively (n = 8, P < 0.05). Similarly, isoproterenol and cantharidin led to an increase of PLB phosphorylation at both phosphorylation sites in isolated perfused guinea pig hearts, as shown in Fig. 1. Moreover, isoproterenol, as well as cantharidin, enhanced phosphorylation of the inhibitory subunit of troponin (Fig. 1). Thus we could measure isoproterenol-stimulated increases in the phosphorylation state of regulatory proteins by means of antibodies in both rat and guinea pig cardiac preparations. Moreover, cantharidin increased the phosphorylation state of PLB-Ser-16 and PLB-Thr-17 as well as of TnI in the guinea pig heart (Fig. 1).

View larger version (29K): [in this window] [in a new window]   Fig. 1.   Autoradiogram of nitrocellulose strips. Effects of isoproterenol (Iso) (1 µM for 5 min) and cantharidin (Cant) (100 µM for 30 min) on phosphorylation of phospholamban (PLB) at serine-16 (PLB-Ser-16) and threonine-17 (PLB-Thr-17) and the inhibitory subunit of troponin (TnI-P) in isolated electrically driven guinea pig hearts perfused according to Langendorff. At the end of the perfusion, ventricular tissue was immediately freeze-clamped, and phosphorylation was measured using phosphorylation-specific antibodies as described in METHODS. Left: molecular weight markers (in kDa) are indicated. Ctr, control.

We then performed similar experiments in nonstimulated guinea pig ventricular cardiomyocytes to compare the data with our previous studies in ³²P-labeled guinea pig ventricular myocytes (19). There, isoproterenol and cantharidin enhanced PLB phosphorylation, but, in that report, we did not determine which amino acid on PLB was phosphorylated (19). In the present study, cantharidin augmented the phosphorylation state of PLB-Ser-16 and PLB-Thr-17 in a concentration-dependent manner. Notably, cantharidin started to increase the phosphorylation state of PLB-Ser-16 and PLB-Thr-17 at the same concentration (10 µM). In contrast to cantharidin and to the effect in electrically driven perfused hearts, isoproterenol enhanced phosphorylation of PLB-Ser-16, whereas we did not detect increased phosphorylation of PLB-Thr-17 (Table 1). Suitably, we have obtained the same results in ³²P-labeled myocytes after phosphoamino acid analysis (data not shown).

It may be asked why we detected isoproterenol-stimulated phosphorylation of PLB-Ser-16 and PLB-Thr-17 in perfused hearts but only on PLB-Ser-16 in nonstimulated cardiomyocytes. This difference might be due to different Ca²⁺ levels in nonstimulated versus stimulated myocytes. Hypothetically, only rhythmical changes in Ca²⁺ levels are able to activate the Ca²⁺/CaM-dependent protein kinase [this was suggested for perfused hearts, e.g., by Wegener et al. (27)]. Cantharidin increased the free intracellular Ca²⁺ concentration in suspensions of nonstimulated guinea pig ventricular cardiomyocytes but only to a very small extent (19). Here, extending our previous work, the effects of cantharidin on Ca²⁺ transients were studied in electrically driven (and contracting) single cells. Cantharidin concentration dependently increased the free intracellular Ca²⁺ concentration starting at 10 µM. Cantharidin was most effective at 100 µM, and the increase amounted to 153 ± 14% of control (n = 6, P < 0.05). Isoproterenol (3 nM) led to a comparable increase of intracellular Ca²⁺ (to 162 ± 9% of control, n = 20). In contrast, neither cantharidin nor isoproterenol increased free intracellular Ca²⁺ concentration in nonstimulated single cardiomyocytes (data not shown). Hence, it is important to measure phosphorylation in contracting preparations to mimic the physiological situation as closely as possible.

On the basis of this conclusion, we studied the effects of cantharidin on contractility and phosphorylation in guinea pig papillary muscles. It was not feasible to study protein phosphorylation in noncontracting papillary muscles because cantharidin induced spontaneous contractions, probably by increasing intracellular Ca²⁺. Therefore, we present here only data on electrically driven papillary muscles. Others have shown that these preparations exhibit changes in intracellular Ca²⁺ (21) that are increased by PP inhibitors (12). Figure 2 depicts original tracings of single contractions in electrically driven guinea pig papillary muscles at high time resolution. Cantharidin (30 min) started to increase the force of contraction at a concentration of 3 µM without affecting the time course of isometric contraction (Fig. 2A). The shortening of relaxation time was apparent at only 100 µM cantharidin, the highest concentration studied. As depicted in Fig. 2B, this concentration of cantharidin markedly enhanced the force of contraction after 6 min of incubation. However, only at a longer time of incubation (30 min), 100 µM cantharidin shortened relaxation time. Under all the conditions studied, cantharidin did not affect the time to peak tension. The time to peak tension under control conditions amounted to 101.4 ± 5.7 and 94.3 ± 5.0 ms after 30 min of incubation with 100 µM cantharidin (n = 11 each, P > 0.05).

View larger version (14K): [in this window] [in a new window]   Fig. 2.   A: concentration-dependent effects of cantharidin (after 30 min incubation) on time course of isometric contraction in isolated electrically driven guinea pig papillary muscles. B: time-dependent effects of cantharidin (100 µM) on time course of isometric contraction in isolated electrically driven guinea pig papillary muscles. Contraction experiments were performed as described in METHODS, and the time course of individual isometric contraction was measured at high chart speed. Bars indicate time scale (in ms) and calibration of force of contraction (in mN), respectively.

Representative Western blots in Fig. 3 show the concentration-dependent effects of cantharidin on TnI-P (A) and phosphorylation of PLB-Ser-16 (B) and PLB-Thr-17 (C). Cantharidin (30 min) started to increase TnI-P and phosphorylation of PLB-Thr-17 at 1 µM, whereas a 10-fold higher concentration (10 µM) was the threshold for PLB-Ser-16 phosphorylation.

View larger version (78K): [in this window] [in a new window]   Fig. 3.   Autoradiogram of nitrocellulose strips. Concentration-dependent effects of cantharidin (for 30 min) on TnI-P (A), PLB-Ser-16 (B), and PLB-Thr-17 (C) in isolated electrically driven guinea pig papillary muscles. After incubation with the indicated concentrations of cantharidin, papillary muscles were freeze-clamped and probed with phosphorylation-specific antibodies as described in METHODS. Left: molecular weight markers (in kDa) are indicated.

Figure 4 summarizes the results from several experiments and contains concentration-response curves for the effects of cantharidin on force of contraction and relaxation time in isolated electrically driven guinea pig papillary muscles. The phosphorylation state of TnI and PLB-Thr-17 started to go up at 1 µM cantharidin, and this was accompanied by an increase in force of contraction. Interestingly, phosphorylation of PLB-Thr-17 and force of contraction induced by increasing concentrations of cantharidin correlated positively (r = 0.504, P < 0.05, n = 28). In contrast, the phosphorylation state of PLB-Ser-16 started to increase at 10 µM cantharidin and preceded the shortening of relaxation time observed at 30 µM.

View larger version (20K): [in this window] [in a new window]   Fig. 4.   Concentration-dependent effects of cantharidin (30 min) on force of contraction (FOC), relaxation time (RT), TnI-P, and phosphorylation of PLB-Ser-16 and PLB-Thr-17 in isolated electrically driven papillary muscles (n = 6-12). After incubation with indicated concentrations of cantharidin, papillary muscles were freeze-clamped and probed with phosphorylation-specific antibodies as described in METHODS. All data are given as the percentage of the corresponding control value (appropriate amount of DMSO). *Significant differences from control (P < 0.05).

We then investigated the time course of cantharidin effects on force of contraction, relaxation time, and protein phosphorylation. As depicted in Fig. 5B, cantharidin (100 µM) rapidly increased phosphorylation of PLB-Thr-17. In contrast, an increase of PLB-Ser-16 (Fig. 5A) phosphorylation was slower and apparent from 10 min onward.

View larger version (85K): [in this window] [in a new window]   Fig. 5.   Autoradiogram of nitrocellulose strips. Time-dependent effects of cantharidin (100 µM) on phosphorylation of PLB-Ser-16 (A) and PLB-Thr-17 (B) in isolated electrically driven guinea pig papillary muscles. At the indicated incubation times, papillary muscles were freeze-clamped and probed with phosphorylation-specific antibodies as described in METHODS. Left: molecular weight markers (in kDa) are indicated.

Figure 6 summarizes the effects of cantharidin on the time course of contractile parameters and the site-specific phosphorylation of PLB. The increase of TnI-P and PLB-Thr-17 phosphorylation at 3 min was paralleled by a positive inotropic effect of cantharidin. A shortening of relaxation time accompanied the increase in the phosphorylation state of PLB-Ser-16 at 20 min.

View larger version (20K): [in this window] [in a new window]   Fig. 6.   Time-dependent effects of cantharidin (100 µM) on FOC, RT, TnI-P, and phosphorylation of PLB-Ser-16 and PLB-Thr-17 in isolated electrically driven papillary muscles (n = 4-8). After the indicated incubation times, papillary muscles were freeze-clamped and probed with phosphorylation-specific antibodies as described in METHODS. All data are given as the percentage of the corresponding control value. *Significant differences from control (P < 0.05).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Phosphorylation of TnI and PLB is an important mechanism by which -adrenergic catecholamines increase the force of contraction and rate of diastolic relaxation in the heart. PKA and Ca²⁺/CaM-dependent protein kinase II phosphorylate PLB in vitro at two different sites, serine-16 and threonine-17 (24). The stimulatory effects of the two protein kinases on SR Ca²⁺ uptake are additive (9, 23, 26). -Adrenergic stimulation phosphorylates PLB at both sites in the intact isolated heart. The initial phosphorylation at serine-16 within seconds was followed by phosphorylation of the threonine-17 residue within minutes (30 s vs. 3 min) (27). Wegener et al. (27) explained this time course as follows. First, isoproterenol increases intracellular cAMP-levels and thereby activates PKA. PKA phosphorylates PLB-Ser-16 and, in addition, also the L-type Ca²⁺ channel. Thus more Ca²⁺ passes through the L-type Ca²⁺ channel, and intracellular Ca²⁺ increases to an extent that it can activate Ca²⁺/CaM-dependent protein kinase, leading to phosphorylation of PLB-Thr-17.

In the present study, the cantharidin-induced increase in the phosphorylation state of PLB-Thr-17 (and TnI-P) preceded the elevation of the phosphorylation state of PLB-Ser-16. This is in opposition to the time course noted after isoproterenol stimulation (10, 27) and warrants consideration. The increases in the phosphorylation states of PLB-Thr-17 and TnI were not accompanied by a shortening of relaxation time. Although the phosphorylation states of PLB-Thr-17 and TnI as well as force of contraction were already high, only a low phosphorylation state of PLB-Ser-16 was detectable, and no effect of cantharidin on relaxation time was seen. We observed this dissociation between the phosphorylation state of PLB-Thr-17 and shortening of relaxation time in studying the concentration dependency of cantharidin effects, and this finding is also supported by the results from the time course study. In agreement with our study, the recent data from Kuschel et al. (10) also indicate that the phosphorylation of PLB-Ser-16 is the key regulator of cardiac relaxation. The authors noted a very close correlation between the phosphorylation of PLB-Ser-16 induced by cAMP-increasing interventions (isoproterenol, forskolin, and 3-isobutyl-1-methylxanthine) and shortening of cardiac relaxation. Moreover, even an eightfold increase of PLB phosphorylation at threonine-17 at submaximal concentrations of isoproterenol had no impact on the rate of cardiac relaxation. However, the time sequence of PLB phosphorylation apparently differs between PP inhibition (present study) and -adrenergic stimulation (10). In contrast to our data, Kuschel et al. reported a time-delayed phosphorylation of PLB-Thr-17, which was preceded by the phosphorylation of PLB-Ser-16 in the presence of isoproterenol. Nevertheless, these findings and our present data concur that the duration of cardiac relaxation is most probably regulated by the phosphorylation of PLB at serine-16 independently of phosphorylation at threonine-17.

It could be asked why cantharidin first elevated the phosphorylation state of PLB at threonine-17 and thereafter at serine-16. In contrast to isoproterenol, cantharidin does not elevate cAMP levels in the heart (19). Hence, PKA is not activated and cannot be expected to phosphorylate PLB-Ser-16. However, cantharidin increased the current through L-type Ca²⁺ channels, probably by PP inhibition in the sarcolemma (19). This should lead to an elevation of the free intracellular Ca²⁺ concentration due to Ca²⁺-induced Ca²⁺ release. Actually, we measured this increase in electrically stimulated single cells (present study). Hence, the phosphorylation state of PLB-Thr-17 should be increased by Ca²⁺/CaM-dependent protein kinase. Additionally, the phosphorylation state of PLB can be increased through inhibition of dephosphorylation by cantharidin. Thus an enhanced phosphorylation state of PLB-Thr-17 should be observed in electrically driven preparations, as it was. We were not able to detect any effect of cantharidin on Ca²⁺ transients in nonstimulated single cells. In an earlier paper (19), we found a slight increase of the intracellular Ca²⁺ concentration by cantharidin, which was also in nonstimulated cells. However, in that study, Ca²⁺ transients were measured in a very simple experimental setting: using fura 2 with a sampling rate of 1 Hz. In addition, the measurements were performed only in a cell suspension without monitoring the cell viability by a microscope. In the present study, the intracellular Ca²⁺ concentration was measured much more precisely: we used single cells and a high sampling frequency (20 Hz), and the cells were continuously monitored during the measurement. We suspect that in our earlier experiments, a part of the cells in the suspension might have been damaged, and this could explain the observed increase of Ca²⁺ concentration. Thus given the much more reliable and modern method used, we consider the present data under improved experimental conditions as definite, and cantharidin does not increase Ca²⁺ transients in nonstimulated cardiomyocytes.

In contrast to phosphorylation at PLB-Thr-17, an increase of PLB-Ser-16 phosphorylation can only result from the inhibition of dephosphorylation. Fittingly, only higher concentrations of cantharidin or longer periods of PP inhibition were sufficient to increase the PLB phosphorylation at serine-16 and to shorten the relaxation time. Hence, this phosphorylation site appears to be important for the relaxation and may mediate the relaxant effect of cantharidin. A caveat is in order. Cantharidin increased the phosphorylation state of numerous proteins in addition to PLB and TnI in isolated ³²P-labeled guinea pig cardiomyocytes (19). Possibly, the increased phosphorylation state of these proteins (most of which are unidentified) may also contribute to the contractile effects of cantharidin.

Another explanation for the higher potency of cantharidin in increasing the phosphorylation state of TnI and PLB-Thr-17 is the greater sensitivity of PP(s) dephosphorylating TnI and PLB-Thr-17 to cantharidin. Given the close proximity of PLB phosphorylation sites the existence of several pools of PPs, differentially accessible for cantharidin seems to be a remote possibility. Cantharidin more potently inhibited the catalytic subunit of PP2A than of PP1 (IC₅₀ = 0.13 µM vs. 2.70 µM; see Ref. 19). Thus the phosphatase(s) dephosphorylating PLB-Thr-17 and TnI might be of type 2A, whereas PP1 might cause the dephosphorylation of PLB-Ser-16. On the other hand, TnI is probably dephosphorylated by PP localized in the contractile apparatus. Thus differences between TnI dephosphorylation and the dephosphorylation of PLB-Ser-16 might be explained by different subcellular pools of PP.

Other hypotheses may also be put forward. Protein phosphorylation reflects the balance between the activities of phosphorylating protein kinases and dephosphorylating PP. Assuming that cantharidin activates Ca²⁺/CaM-dependent protein kinase via increasing intracellular Ca²⁺, both threonine-17 and serine-16 could be dephosphorylated by the same PP; a partial inhibition of this PP would enhance the threonine-17 phosphorylation, whereas complete inhibition by high concentrations of cantharidin is necessary for an increase of serine-16 phosphorylation.

The data from nonstimulated cells argue for the latter hypothesis. In nonstimulated cells, isoproterenol increases only cAMP levels but does not elevate intracellular Ca²⁺ (19). This leads to the phosphorylation of PLB-Ser-16 but not of PLB-Thr-17 (Table 1). Similarly, cantharidin does not increase intracellular Ca²⁺ in nonstimulated single cells. Hence, in nonstimulated cardiomyocytes, inhibition of PP(s) will be the only mechanism increasing PLB phosphorylation. Because cantharidin per se neither stimulates PKA nor Ca²⁺/CaM-dependent protein kinase, their basal activity will slowly lead to PLB phosphorylation on both sites (threonine-17 and serine-16). If the same PP or the mixture of two PP (consisting of equal amounts of PP1 and PP2A) dephosphorylates both phosphorylation sites, a similar concentration dependency for cantharidin-induced phosphorylation on serine-16 and threonine-17 would be expected. This is what we measured in nonstimulated ventricular cardiomyocytes (Table 1). Moreover, cantharidin is ~10× less potent in increasing PLB-Thr-17 phosphorylation in nonstimulated cardiomyocytes than in electrically stimulated preparations. This "rightward shift" of the concentration-response curve implies an additional mechanism operating in electrically driven preparations. Thus in contracting papillary muscles, differences in protein kinase activation (only Ca²⁺/CaM-dependent protein kinase being stimulated) can explain the differences in cantharidin-induced phosphorylation of PLB-Ser-16 and PLB-Thr-17.

In summary, we have shown that increases in the phosphorylation state of TnI and PLB-Thr-17 accompany the positive inotropic effect of cantharidin without affecting the time course of contraction under isometric conditions. The cantharidin-induced increase in the phosphorylation state of PLB-Ser-16 precedes (and may cause) the relaxant effect of cantharidin.