INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

THE L-TYPE Ca²⁺ channel in the heart is a well-known target for -adrenergic regulation. For ₁- and ₂-adrenoceptors, the main signal transduction mechanism is believed to be cAMP-dependent phosphorylation of the channel. In several studies, however, differences were noted between the effects of selective activation of cardiac ₁- and ₂-adrenoceptors, respectively (2, 14, 33, 34). In particular, ₂-adrenergic stimulation of rat cardiomyocytes by the selective agonist zinterol increased whole cell Ca²⁺ current, but in contrast to the ₁-adrenergic effect, the time course of inactivation was slowed by zinterol (33). The authors concluded that this kinetic difference indicates a genuine difference between signal transduction pathways leading from the two types of -adrenoceptors to the Ca²⁺ channel. This statement, if true, would indeed be important for the understanding of ₁- and ₂-adrenoceptor function, because differential effects on the very same molecular target could not be explained by, for instance, differences in cAMP compartmentation (15) after the various agonists.

However, this conclusion needs to be corroborated by single-channel recordings for two reasons: 1) whole cell data cannot rule out the possibility of just a different spatial pattern of channel activation after ₁- and ₂-adrenoceptor activation, possibly because of a different distribution of receptors or cAMP compartmentation, and 2) inactivation kinetics under physiological conditions are strongly dominated by Ca²⁺-dependent inactivation (21, 28), and a major amount of this Ca²⁺ is released from the sarcoplasmic reticulum in the heart (24). Because zinterol affected intracellular Ca²⁺ handling to a lesser extent than ₁-adrenoceptor activation (33), the kinetic difference seen with the L-type Ca²⁺ current may have been secondary to the differences in the Ca²⁺ transient. Addition of 5 mM EGTA to the pipette solution (33) could not eliminate Ca²⁺-dependent inactivation and the confounding influence of intracellular Ca²⁺ handling.

Accordingly, we reassessed the effect of zinterol at the single-channel level compared with the effect of isoproterenol and a cAMP derivative. We chose our conditions to minimize any confounding influence of the intracellular Ca²⁺ homeostasis. The cell-attached configuration per se is suitable, because membrane depolarization is restricted to the patch area. Furthermore, the use of Ba²⁺ instead of Ca²⁺ minimizes inactivation by the charge carrier (13). Finally, we preincubated the cells with a cell-permeant form of the rapid (36) Ca²⁺ chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid (BAPTA). The involvement of ₂-adrenoceptors was checked using the ₂-subtype-selective antagonist ICI-118551 {(±)-1-[2,3-(dihydro-7-methyl-1H-inden-4-yl)oxy]-3-[(1-methylethyl)amino]-2-butanol hydrochloride}.

Under these conditions, it was possible to confirm, at the single-channel level, kinetic differences between zinterol and the other agonists in support of a true peculiarity of the signal transduction pathway of ₂-adrenergic receptors (2, 30, 32, 33).

	METHODS

Top Abstract Introduction Methods Results Discussion References

Isolation of ventricular myocytes. Adult male Wistar rats (180-350 g) were killed by cervical dislocation. Ventricular myocytes were enzymatically isolated as previously described by Mitra and Morad (26) by using a Langendorff perfusion. Briefly, hearts were perfused (48 mmHg, 37°C) in Ca²⁺-free, modified Tyrode solution (in mmol/l: 135 NaCl, 4 KCl, 0.3 NaH₂PO₄, 1 MgCl₂, 10 HEPES, 10 dextrose, pH 7.3). During the following 20 min, collagenase (1 mg/min; type CLS 1, Worthington Biochemical, Freehold, NJ) and protease (0.7 mg/min; type XIV, Sigma Chemical, St. Louis, MO) were added. After this treatment the hearts were perfused with enzyme-free modified Tyrode solution containing 0.2 mmol/l Ca²⁺ for 5 min. They were then minced, and cells were dispersed by gentle agitation. Before the electrophysiological measurements, aliquots of cell suspension were preincubated at room temperature for 30-180 min with 10 µM BAPTA-AM (Calbiochem).

Measurements of Ba²⁺ currents through single channels. Cells were placed in disposable perfusion chambers containing ~3 ml of bath solution. Single-channel recordings were performed in the cell-attached configuration. Pipettes (7-11 M) contained (in mmol/l) 70 BaCl₂, 110 sucrose, and 10 HEPES, with pH adjusted to 7.4 with tetraethylammonium hydroxide. Unlike previous studies in which guinea pig myocytes were used, we were unable to elicit in rat cells any considerable stimulation of currents by isoproterenol when using the conventional K⁺-rich depolarizing bath solution (cf. Ref. 37). In these preliminary experiments (n = 5), isoproterenol led to no increase in peak current [from 40.6 ± 8.8 to 38.2 ± 8.5 fA, not significant (NS)], availability (from 44.1 ± 4.3 to 37.7 ± 4.3%, P < 0.05), and open probability (P_o; from 6.61 ± 0.8 to 6.01 ± 0.8%, P = NS). Therefore, in all experiments reported here, a Tyrode solution was used, because successful stimulation of single rat channels by isoproterenol was reported previously by Scamps et al. (27). The composition of this solution was (in mmol/l) 135 NaCl, 4 KCl, 1 MgCl₂, 10 HEPES, 2 CaCl₂, and 10 dextrose, pH 7.3. A number of cells retained a negative resting potential under these conditions. Accordingly, Ba²⁺ currents were elicited by voltage steps (150 ms at 1.66 Hz) of a given amplitude (120 mV), but the absolute value of the applied potential (e.g., from 100 to +20 mV or from 40 to +80 mV) was chosen in each experiment to yield an approximate transpatch test potential of +20 mV guided by the single-channel amplitude (adjusted to about 0.7 pA, since current amplitude was 0.703 ± 0.016 pA at +20 mV under depolarizing conditions, n = 15). Experiments with a clear shift in single-channel amplitude during recording were excluded. Data were sampled at 10 kHz and filtered at 2 kHz (3 dB, 4-pole Bessel) by using an Axopatch 200 amplifier (Axon Instruments, Foster City, CA). pClamp software (version 6.0, Axon Instruments) was used for data acquisition and analysis of openings and closures.

Drug solutions. Zinterol (kindly provided by Bristol-Myers-Squibb) was prepared as a 10 mM stock solution in DMSO. 8-Bromo-cAMP (8-BrcAMP; Sigma Chemical) was dissolved as a 0.1 M stock solution in DMSO. Isoproterenol (Sigma Chemical) was prepared in distilled water containing the drug at 0.1 mM and 1 g/l of ascorbic acid to prevent oxidation. The stock solutions of ICI-118551 (Imperial Chemical Industries) and CGP-20712A {(±)-2-hydroxy-5-[2-[[2-hydroxy-3-[4-[1-methyl-4-(trifluoromethyl)-1H-imidazol-2-yl]phenoxy]propyl] amino] ethoxy]-benzamide methanesulfonate, Biotrend} contained the drug at 0.1 mM in 75% water and 25% DMSO (vol/vol). Drugs were added as bolus of 10 µl (ICI-118551 and CGP-20712A), 15 µl (zinterol), or 30 µl (isoproterenol and 8-BrcAMP) to the bath chamber. ICI-118551 and CGP-20712A were applied simultaneously with the respective agonists. The total volume of the bath was not always exactly 3 ml, but it was measured at the end of each experiment. Accordingly, the final concentration referred to as 50 µM zinterol reflects a certain range of concentrations (30-75 µM). The same applies to 0.3 µM ICI-118551 (0.23-0.36 µM), 0.3 µM CGP-20712A (0.25-0.34 µM), 1 µM isoproterenol (0.67-1.7 µM), and 1 mM 8-BrcAMP (0.63-1.67 mM).

Data analysis and statistics. Linear leak and capacity currents were digitally subtracted using averaged current traces of nonactive sweeps. The availability (percentage of sweeps containing 1 channel opening, i.e., fraction of active sweeps per total number of test pulses), P_o (i.e., fractional occupancy of the open state during active sweeps), and the peak ensemble-average current were corrected by the number of channels in the patch (n) in case of double-channel patches; n was derived from the maximum current amplitude observed divided by the unitary current amplitude. Experiments with n > 2 were entirely rejected from analysis. Peak current was normalized by division through n. The availability was corrected by the square root method: (1  availability_corr) is the nth root of (1  availability_uncorr), where availability_corr and availability_uncorr are corrected and uncorrected availability, respectively. The corrected P_o was calculated on the basis of the corrected number of active sweeps, i.e., total open time (in ms) within all sweeps of the ensemble, divided by (150 ms × n × availability_corr × number of test pulses). Openings and closures were identified by the half-height criterion. Open and closed time distributions were analyzed after logarithmic binning by using the maximum likelihood method (pStat version 6.0). Closed times, first latency, burst length, and the slow gating process (see below) were analyzed only when patches contained no more than one single channel. Burst length is defined as the interval between the first opening and the last closure of an active sweep. Values are means ± SE. Significance was checked by a paired t-test at P < 0.05. Increases in average current were checked in a two-tailed analysis, and a one-tailed test was applied for the individual gating parameters possibly contributing to the increase in ensemble-average current. The results are from 7 experiments for zinterol (5 single- and 2 double-channel patches), 6 for isoproterenol (5 single- and 1 double-channel patches), 6 for 8-BrcAMP (5 single- and 1 double-channel patches), 5 for zinterol plus ICI-118551 (2 single- and 3 double-channel patches), 8 for isoproterenol plus ICI-118551 (6 single- and 2 double-channel patches), and 7 for isoproterenol plus CGP-20712A (4 single- and 3 double-channel patches).

	RESULTS

Top Abstract Introduction Methods Results Discussion References

The effect of zinterol on cardiac L-type Ca²⁺ channels was tested in seven experiments. First, 240-360 control test pulses were recorded. Then high concentrations [presumably saturating (33)] of 50 µM zinterol were added to the bath solution, and another 180-840 sweeps were obtained. Figure 1 shows the effect of zinterol on the behavior of a single L-type Ca²⁺ channel. Channel activity was increased, as illustrated by the ensemble-average current (bottom) and the consecutive individual sweeps during control (left) and after exposure to the drug (right). The ensemble-average current was enhanced by two effects: 1) the availability, i.e., the probability of a channel to open at least once within a test pulse, was increased, and 2) within the active sweeps, P_o was enlarged. These findings could be reproduced in all experiments with zinterol: peak current increased from 38.5 ± 23.0 to 50.0 ± 27.0 fA, availability from 45.9 ± 8.1 to 65.5 ± 8.3%, and P_o from 5.4 ± 2.8 to 6.7 ± 3.2% (all P < 0.05, n = 7). The major contribution to an increase in ensemble current came from the availability (see below). An effect on the rate or extent of inactivation is not obvious from the ensemble current traces (Fig. 1). In the whole set of experiments, zinterol did not affect inactivation, as measured by the decay of the current from peak to the end of the test pulse (control, decrease by 33.0 ± 9.7%; after zinterol, 32.1 ± 7.3%).

View larger version (48K): [in this window] [in a new window]   Fig. 1.   Effect of zinterol on Ba2+ currents through a single cardiac L-type Ca2+ channel. Middle: sequential unitary current records before (control) and after zinterol. Bottom: ensemble-average currents calculated from sweeps before (240 sweeps) and after (840 sweeps) addition of zinterol to bath. Top: voltage steps (from 100 to +20 mV) delivered every 600 ms for 150 ms. Scale bars, 20 ms and 2 pA (middle) or 0.055 pA (bottom).

The lack of prolongation of inactivation by zinterol may be due to the small extent of inactivation after 150 ms under our conditions. However, because the other single-channel effects of zinterol are qualitatively similar to those reported for conventional -adrenergic or cAMP-dependent stimulation of Ca²⁺ channels (11, 37), we decided to compare our results with the action of equieffective concentrations of isoproterenol (Fig. 2) and 8-BrcAMP (Fig. 3). For each experiment, 240-420 control sweeps were recorded. Subsequently, 1 µM isoproterenol or 1 mM 8-BrcAMP was applied, and another 120-720 sweeps were acquired. Isoproterenol and 8-BrcAMP also increased ensemble-average current. Again, a significant enhancement of channel availability was found. P_o was significantly enhanced by 8-BrcAMP. Although isoproterenol tended to increase P_o by a similar amount, this effect did not reach significance, because in one experiment P_o decreased slightly. Figure 4 shows the effects of zinterol (n = 7), isoproterenol (n = 6), and 8-BrcAMP (n = 6) on the peak average channel current, availability, and P_o. It is evident that the changes caused by the three compounds are very similar. As with zinterol, neither isoproterenol (from 13.8 ± 5.6 to 16.3 ± 5.9%) nor 8-BrcAMP (from 17.5 ± 6.9 to 15.8 ± 5.6%) changed the extent of inactivation at the end of the 150-ms pulse.

View larger version (49K): [in this window] [in a new window]   Fig. 2.   Effect of isoproterenol on Ba2+ currents through a single cardiac L-type Ca2+ channel. Middle: sequential unitary current records before (control) and after isoproterenol. Bottom: ensemble-average currents calculated from sweeps before (300 sweeps) and after (420 sweeps) addition of isoproterenol to bath. Top: voltage steps (from 100 to +20 mV) delivered every 600 ms for 150 ms. Scale bars, 20 ms and 2 pA (middle) or 0.055 pA (bottom).

View larger version (43K): [in this window] [in a new window]   Fig. 3.   Effect of 8-bromo-cAMP (8-BrcAMP) on Ba2+ currents through a single cardiac L-type Ca2+ channel. Middle: sequential unitary current records before (control) and after 8-BrcAMP. Bottom: ensemble-average currents calculated from sweeps before (240 sweeps) and after (120 sweeps) addition of 8-BrcAMP to bath. Top: voltage steps (from 80 to +40 mV, in this case yielding a transpatch test potential of about +20 mV) delivered every 600 ms for 150 ms. Scale bars, 20 ms and 2 pA (middle) or 0.055 pA (bottom).

View larger version (22K): [in this window] [in a new window]   Fig. 4.   Effects of -adrenoceptor agonists and 8-BrcAMP on Ba2+ currents through single cardiac L-type Ca2+ channels. Drug effects on peak average current, channel availability (fraction of active sweeps containing 1 opening), and open probability (fraction of time spent in open state within active sweeps) are depicted with respective preceding control values. Open bars, controls; filled bars, zinterol; hatched bars, isoproterenol; striped bars, 8-BrcAMP. * P < 0.05 vs. control.

Because we might have overlooked a distinct effect on current ensemble kinetics because of the low signal-to-noise ratio of the ensemble-average traces, the single-channel gating kinetics were analyzed in detail for the three compounds. In particular, it was interesting to see how the increase in P_o was brought about. In preceding studies it was shown that P_o of L-type Ca²⁺ channels can be increased by a prolongation of the open times due to a shift toward a long opening mode (37) or by a shortening of the closed times (5, 11). We did not detect a significant change of mean open times after the addition of zinterol (from 0.47 ± 0.02 to 0.56 ± 0.08 ms), isoproterenol (from 0.54 ± 0.06 to 0.55 ± 0.05 ms), or 8-BrcAMP (from 0.53 ± 0.03 to 0.50 ± 0.02 ms). In this context it must be mentioned that sweeps containing several long openings (typical "mode 2") were seen here only rarely, in contrast to our previous work in guinea pig cells (29, 37). The results from the open time histogram are shown in Table 1. The time constants _open were obtained by monoexponential fits.

Surprisingly, zinterol was also unable to change the mean closed time (from 7.1 ± 1.6 to 7.0 ± 1.8 ms), whereas isoproterenol and 8-BrcAMP caused a significant reduction (from 10.2 ± 3.8 to 5.5 ± 1.8 and from 11.7 ± 2.0 to 7.5 ± 2.2 ms, respectively). To further examine this unexpected result, histograms of closed time distributions were analyzed. Double-exponential curves were fitted, yielding two time constants of closures (Fig. 5). Isoproterenol and 8-BrcAMP reduced the time constant and/or the proportion of long closures (Fig. 5), leading to an increased density of channel openings within active sweeps. The time constant of short closures remained unchanged. This fact explains the decline of the mean closed time and is in agreement with the results of preceding studies (5, 11). Interestingly, neither the time constant of long closures nor the proportion of long closures was influenced by zinterol (Fig. 5). The fast time constant was even slightly prolonged. Thus the mean closed time, i.e., the density of channel openings within active sweeps, was unaltered by zinterol. Therefore, the question arose of how zinterol may have increased P_o without affecting open or closed time distributions. The first latency, i.e., the waiting time from the beginning of the test pulse to the first opening, appeared to be reduced. However, this occurred to a small and insignificant extent, and this finding applied similarly to all three compounds: zinterol (from 43.1 ± 8.8 to 37.2 ± 9.1 ms), isoproterenol (from 30.6 ± 5.7 to 26.2 ± 7.9 ms), and 8-BrcAMP (from 43.9 ± 4.2 to 39.5 ± 6.9 ms). The one remaining explanation for the slight increase in P_o exerted by zinterol would be an enlargement of the burst length, defined here as the time from the beginning of the first opening to the end of the last opening within an active sweep (Fig. 6). Indeed, zinterol significantly increased mean burst length (from 53.3 ± 18.7 to 72.2 ± 17.4 ms). The size of this effect matches the observed extent of increase in the P_o. The burst length was also significantly increased by 8-BrcAMP (from 50.0 ± 8.4 to 62 ± 9.6 ms) but not by isoproterenol (from 78.0 ± 12.7 to 72.3 ± 17.7 ms, P = NS). To account for the problem that burst length duration may be underestimated because of the relatively short test pulse ("censor problem"), we analyzed the burst length distribution (Fig. 7). Indeed, burst length displayed a bimodal distribution. The main (long-lived) component of bursts was truncated by the 150-ms length of the voltage step: this truncation occurred over the region of 130-150 ms, corresponding to the pulse length minus the average first latency. Zinterol, without affecting this main component of bursts, reduced the short-lived component of bursts: the fraction of sweeps where burst length was 1 ms (percentage of all active sweeps) was reduced from 25.2 ± 5.3 to 11.8 ± 3.1% (P < 0.05), an effect more marked than with isoproterenol (from 19.9 ± 7.7 to 14.7 ± 7.9%, P = NS). These very short "bursts" are comprised of mostly one (rarely 2) very short opening.In summary, the ₂-adrenoceptor agonist zinterol increases P_o but cannot reduce long closed times. The only way zinterol enhanced P_o was to "prolong" burst length, which, however, turns out to be a redistribution phenomenon. The very characteristics of fast gating, including the P_o calculated within bursts, remained unaffected by zinterol (from 10.5 ± 1.1 to 12.2 ± 4.4%, n = 5, P = NS), whereas this parameter was raised by isoproterenol (from 9.35 ± 3.1 to 14.4 ± 3.8%, n = 5, P < 0.05), as expected from the changes in closed times.

View larger version (34K): [in this window] [in a new window]   Fig. 5.   Closed time histograms of 3 representative single-channel experiments before and after zinterol, isoproterenol, and 8-BrcAMP. Square root (SQR) of number of events per bin is plotted against their duration (logarithmic scale). Curves were generated using a maximum likelihood estimate for a double-exponential function before (dashed curves) and after drugs (solid curves). A scaled version of control curves is plotted with data after drugs for a better comparison.

View larger version (25K): [in this window] [in a new window]   Fig. 6.   Trace of unitary Ba2+ currents through a single Ca2+ channel to illustrate "burst length." Arrowheads, beginning of first and end of last channel opening. Time between these events is defined as burst length. Open bars, controls; filled bars, zinterol; hatched bars, isoproterenol; striped bars, 8-BrcAMP. * P < 0.05 vs. respective control.

View larger version (22K): [in this window] [in a new window]   Fig. 7.   Burst length distribution before and after isoproterenol (n = 5) and zinterol (n = 5). Only single-channel experiments were analyzed. Histograms from individual experiments were normalized (by setting largest bin value to 1.0) and then averaged. Note bimodal distribution throughout and marked reduction of short-lived component after zinterol.

With the concentrations of zinterol used here, a nearly complete activation of ₂-adrenoceptors, but also a considerable occupancy of ₁-adrenoceptors, should be expected (20). Therefore, we had to check whether part of the effect of zinterol was mediated by ₁-adrenoceptors. In five experiments, 50 µM zinterol was applied, together with 0.3 µM ICI-118551, a selective ₂-adrenoceptor antagonist (3, 38). ICI-118551 blocked the effect of zinterol almost completely. There was no change in the peak average current (from 11.8 ± 2.6 to 13.2 ± 3.1 fA, P = NS) and the mean P_o (from 2.1 ± 0.6 to 2.5 ± 0.6%, P = NS). Availability also remained constant (from 40.0 ± 5.6 to 40.2 ± 5.0%, P = NS).

On the other hand, isoproterenol is a nonselective -adrenoceptor agonist, and a part of its effect might have been mediated by ₂-adrenoceptors. To address this question, we tried to exert a pure ₁-adrenoceptor stimulation with a combination of 1 µM isoproterenol and 0.3 µM ICI-118551. Inclusion of ICI-118551 did not influence the previously described features of isoproterenol effects. The ensemble-average current was significantly increased (from 28.2 ± 7.3 to 65.3 ± 17.4 fA) because of a significant enhancement of availability (from 42.5 ± 3.7 to 63.6 ± 6.6%) and P_o (5.2 ± 1.56 to 9.5 ± 2.8%). P_o was mainly increased by a significant reduction of the mean closed time (from 8.9 ± 2.9 to 5.4 ± 1.6 ms) because of changes in the slow component of the closed time histogram (_C2 dropped from 14.6 ± 3.88 to 8.65 ± 1.8 ms, P = NS). Interestingly, in the experiments with isoproterenol combined with ICI-118551, the prolongation of mean open time reached statistical significance (from 0.53 ± 0.04 to 0.63 ± 0.06 ms), but the extent of this effect was again quite moderate. In the open time histogram analysis, _open increased slightly but significantly (from 0.42 ± 0.02 to 0.53 ± 0.05 ms), as with isoproterenol given alone (Table 1). In summary, after blockade of ₂-adrenoceptors, the extent and pattern of isoproterenol effects remained essentially the same.

On the contrary, a zinterol-like action should be elicited by isoproterenol if ₁-adrenoceptors are selectively blocked. The effect of 1 µM isoproterenol was therefore studied in the presence of the ₁-selective adrenoceptor antagonist CGP-20712A (0.3 µM, n = 7). Peak current was stimulated from 34.9 ± 5.1 to 64.6 ± 10.7 fA (P < 0.05), caused by an increase of availability (from 47.9 ± 6.4 to 76.9 ± 6.9%, P < 0.05) and P_o (from 8.5 ± 2.2 to 10.2 ± 2.7%, P < 0.05). Analysis of the single-channel patches (n = 4) revealed that neither the mean closed time (from 7.2 ± 3.9 to 7.5 ± 3.8 ms) nor its individual components were significantly changed. This pattern is entirely consistent with the action of zinterol, indicating that ₂-adrenoceptor stimulation exerts a unique form of single-channel activity, independent of the agonist used.

The most prominent effect of zinterol (and of isoproterenol plus CGP-0712A) was an increase in availability, comparable in size to that of the other drugs. Such an increase in availability can be further analyzed kinetically by sweep histogram analysis. It has been shown previously (10) that isoproterenol increases availability by shortening the periods of consecutive nonactive sweeps (blank runs) and by lengthening the periods of consecutively active sweeps (active runs). As shown in Fig. 8, zinterol exerts a very similar effect. Table 2 shows data from all appropriate experiments (single channel, sufficiently long recording). Two suitable experiments with zinterol plus ICI-118551 yielded no change in active or blank run length (not shown). In all other groups, although the number of experiments was often too small for statistical significance, both kinetic effects were present. In summary, there is no evidence for a special mechanism of ₂-adrenoceptor stimulation from the analysis of the slow (availability) gating process, unlike the peculiar effects on fast gating described above.

View larger version (20K): [in this window] [in a new window]   Fig. 8.   Slow-gating analysis of a single Ca2+ channel before (control) and after zinterol. Periods of consecutively active and nonactive sweeps were plotted cumulatively as active and blank runs, respectively. Histograms yield an estimate of lifetime of available state ( of monoexponential fits describing active runs) and of nonavailable state(s), calculated as mean duration of blank runs (10). Zinterol prolonged available state (top) from 0.62 to 2.19 s while shortening unavailable state duration from 2.17 to 1.42 s.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

This study confirms at the single-channel level that ₂-adrenergic stimulation of rat cardiomyocytes results in activation of L-type Ca²⁺ channels. First, to obtain a response of classical adrenergic agonists, we first had to adapt our standard experimental protocol (29, 37) from a depolarizing bath solution to a more physiological bath solution. This phenomenon, evidently representing a species difference between guinea pig and rat, cannot be explained. Membrane potential per se as a factor is unlikely (27), inasmuch as cells with a depolarized membrane potential (as estimated from the necessary clamp potentials) responded as well to agonists as those with a physiological resting potential in Tyrode solution.

The stimulatory effect of the ₂-agonist is in agreement with the measurement of whole cell Ca²⁺ current in rat cardiomyocytes (33) and in a fibroblast cell line coexpressing ₂-adrenoceptors and cardiac L-type Ca²⁺ channel subunits (23). However, ₂-adrenergic stimulation increased single L-type Ca²⁺ channel activity in a very specific manner. Analysis of single-channel gating revealed a typical increase of channel availability but an inability to reduce closed times completely different from the effect of isoproterenol (alone and in the presence of ₂-adrenoceptor blockade) or 8-BrcAMP. The latter compounds increased L-type Ca²⁺ current mainly because of a reduction of closed times, which is in agreement with previous results (5, 11). Importantly, the same pattern as with zinterol (i.e., increase of peak current, availability, and P_o, but without shortening of closed times) occurred with isoproterenol after ₁-adrenoceptor blockade with CGP-20712A.

Xiao and Lakatta (33) showed that stimulation of the ₂-adrenoceptor prolonged Ca²⁺ current inactivation time of the whole cell L-type Ca²⁺ current. In our experiments, however, the main mechanism whereby ₂-adrenoceptor stimulation increased ensemble currents was an increase in availability, which per se would not lead to a change in inactivation kinetics. Indeed, regarding the ensemble-average current, we found in control experiments only a slight inactivation of 15-30%, which was not visibly influenced by the test compounds. Even when we normalized and averaged the various experiments within data sets, no effect of zinterol on inactivation time course became obvious (not shown). One may argue that it would have required longer test pulses to reveal a more substantial inactivation of L-type Ba²⁺ current in cardiac myocytes (6) and to possibly observe a prolongation caused by ₂-adrenoceptor stimulation. The mechanism whereby zinterol affects fast gating, however, gives a clue to the apparent discrepancy between unitary events and "macroscopic" Ba²⁺ currents. Zinterol did not affect standard kinetic parameters such as open times or, in particular, closed times. Instead, the drug selectively reduced the relative abundance of very short-lived bursts (Fig. 7). The (mostly) singular opening events underlying this short-lived component of the burst length distribution resemble mode 0_a gating, originally described by Yue et al. (37). The finding that zinterol increases P_o simply by a redistribution of burst lengths 1) explains why zinterol increases P_o without altering closed times (there are almost no closed events associated with these short bursts), 2) resolves why there is no associated change in ensemble-average current (because mode 0_a contributes only negligibly to the average current), and 3) gives rise to a hypothesis to explain why whole cell currents with Ca²⁺ show a decrease in the rate of inactivation (33) [due to the proposed correspondence between mode 0_a and mode_Ca (13)]. We therefore consider that this effect might be a single-channel representation responsible for the prolonged inactivation time course under more physiological circumstances. This idea should be tested using single-channel recordings with Ca²⁺ as the charge carrier. However, the most important conclusion at this stage is that there is a kinetic difference (no effect on closed times) between selective ₂-adrenergic stimulation and the classical agonist. Importantly, the qualitatively different effects are exerted by zinterol at a comparable level of ensemble current stimulation (e.g., 1.58 ± 0.18- and 1.58 ± 0.07-fold for zinterol and isoproterenol, respectively), ruling out simple concentration-dependent phenomena (11). The main mechanism whereby ₂-adrenergic receptor stimulation leads to an increase in current is enhanced availability, brought about by the very same kinetic phenomena as with ₁-receptor-selective or -nonselective stimulation.

It is well established that ₁- as well as ₂-adrenoceptor subtypes increase the activity of adenylyl cyclase in cardiomyocytes via an interaction with G_s protein, raise the cellular cAMP concentration, and increase the level of cAMP-dependent protein phosphorylation (12, 16, 19). Interestingly, in several previous studies, distinct cellular responses after subtype-specific -adrenergic stimulation were reported (2, 14, 33, 34). Although it remained controversial for a long time whether rat cardiomyocytes express ₂-adrenoceptors (4) at all and whether the functional response of -adrenergic activation is mediated by ₂-receptors (14), it could recently be demonstrated that about one-third of -adrenoceptors of isolated rat ventricular cardiomyocytes are ₂-adrenoceptors (7). Furthermore, it could be shown that ₁- and ₂-adrenoceptor stimulation of cardiac myocytes from rat differ in effects on contraction, cytosolic Ca²⁺, and sarcolemmal Ca²⁺ current (33). It was also demonstrated that, in rat, ₁- and ₂-adrenoceptor activation increases intracellular cAMP content. However, after ₂-adrenoceptor stimulation, elevation of intracellular cAMP concentration was dissociated from the effect on intracellular Ca²⁺ transient and contraction amplitude (30).

These results indicate that the two receptor subtypes possess a different intracellular signal transduction. The mechanism for this is unclear, but an exclusive role for cAMP compartmentation is made unlikely by our single-channel approach, as outlined in the introduction. Furthermore, under the geometric conditions of the cell-attached configuration, a direct stimulation of the channel by G_s protein, as previously described for isoproterenol (35), is unlikely. This is supported by the phenomenological similarity between effects shown for isoproterenol and 8-BrcAMP. It is of great interest that ₂- but not ₁-adrenoceptor-mediated effects are potentiated by pertussis toxin (PTX) protein (32). This is evidence for the idea that ₂-adrenoceptors activate not only G_s protein, but also a PTX-sensitive G protein (31). PTX-sensitive G proteins inhibit the adenylyl cyclase and regulate the affinity of -adrenoceptors (1). However, the downstream elements leading to a special form of Ca²⁺ channel stimulation (33; present results) remain to be clarified. Activation of protein phosphatases has been implicated by Xiao et al. (32), and the idea is consistent with findings that ₂-adrenoceptor effects are still entirely cAMP dependent (39). Ca²⁺ channels can be dephosphorylated by protein phosphatase types 1 and 2A (PP1 and PP2A). Dephosphorylation of the channel by PP2A leads to a reduction of P_o (9, 29), whereas regulation of PP1 controls availability (10, 29). Our kinetic analysis of slow gating revealed no difference between ₂-adrenoceptor stimulation and classical agonists, which lead (presumably by inhibition of PP1) to a lengthening of active runs (10) (Table 2). Therefore, we speculate that our main result, the lack of effect of ₂-agonists on mean closed time, could be mediated by activation of PP2A via PTX-sensitive G proteins, which counteracts the cAMP-dependent phosphorylation of just one of several regulatory sites on the channel. Future experiments should examine the influence of phosphatase inhibitors on ₂-adrenergic effects, and the single-channel mechanism of ₂-adrenergic stimulation should be examined after pretreatment with PTX.

Characterizing -adrenoceptors in rat cardiac tissue, Kaumann and Molenaar (18) described a ₃-adrenoceptor. Moreover, a fourth "atypical" -receptor was found (17, 22). It is considered that the ₃-adrenoceptor couples to PTX-sensitive G proteins (8), whereas the atypical -receptor activates adenylyl cyclase (17). We cannot completely rule out that our findings are influenced by stimulation of the ₃-adrenoceptor or even the atypical -adrenoceptor. Zinterol effects due to stimulation of the ₃-adrenoceptor seem unlikely, because rodent ₃-adrenoceptors expressed in Chinese hamster ovary cells were not blocked by ICI-118551 at the concentration used here (3). However, the effect demonstrated here for zinterol was markedly attenuated by ICI-118551. Furthermore, 8-BrcAMP mimicked quite exactly the response of L-type Ca²⁺ channels to isoproterenol and isoproterenol plus ICI-118551, suggesting that isoproterenol effects are entirely due to an elevation of cAMP. Of course, we cannot rule out partial involvement of atypical receptors to the response toward isoproterenol, and a minor part of the stimulatory effect of zinterol may be due to occupancy of ₁- or atypical adrenoceptors. The important point, however, remains that zinterol effects, as well as isoproterenol effects in the presence of ₁-adrenoceptor blockade, differ qualitatively from effects of the other agents. This may turn out to be important under pathophysiological conditions. As stated above, there are several possibilities to further investigate the molecular basis of ₂-adrenoceptor action at the single-channel level. The development of transgenic animals with marked ₂-adrenoceptor overexpression (25) may be of great help for this purpose.