Activation of the cAMP–protein kinase A pathway facilitates Na⁺ translocation by the Na⁺–K⁺ pump in guinea-pig ventricular myocytes

METHODS

Experimental procedure and solutions

A culture dish containing isolated ventricular myocytes was placed on the stage of an inverted microscope (Diaphot, Nikon). The bath solution was pre-warmed to ∼35°C by means of a Peltier device. The medium contained (mM): NaCl, 144; NiCl₂, 5.0; BaCl₂, 2.0 or 4.0; CaCl₂, 1.8; MgCl₂, 0.5; Hepes, 10; pH 7.4 (adjusted with NaOH). For measurements of steady-state Na⁺-K⁺ pump currents 10 mM KCl was added.

Na⁺-free solution contained (mM): choline chloride, 144; KCl, 1.5; NiCl₂, 5.0; BaCl₂, 4.0; CaCl₂, 1.8; MgCl₂, 0.5; Hepes, 10; atropine, 0.002-0.005; pH 7.4 (adjusted with tetraethylammonium hydroxide (TEA-OH)). All solutions, including the pipette solutions (see below), were selected so as to isolate Na⁺-K⁺ pump-mediated currents and the cAMP-dependent chloride current (I_Cl(cAMP)).

The cell under study was additionally superfused by means of a multibarrelled, solenoid-operated pipette with a tip diameter of 500 μm. Gravity-fed solution flow was ∼0.4 ml min⁻¹ and solution exchange near the cell was complete within 1 s. Under these conditions, the temperature in the vicinity of the myocyte was ∼30°C. Ventricular cells were whole-cell voltage clamped with low resistance patch pipettes made from borosilicate glass capillaries (GC150TF-10, Clark Electromedical Instruments, Reading, UK). Pipettes were pulled on a horizontal puller (DMZ, Zeitz, Munich, Germany) and had initial resistances between 1 and 2 MΩ when backfilled with the respective pipette solution. Two pipette solutions were used, a high- and a low-chloride solution. The high-chloride pipette solution contained (mM): TEA-Cl, 130; TEA-OH, 20; NaCl, 5; MgCl₂, 3; EGTA, 6; Hepes, 16; MgATP, 10; pH 7.2-7.4 (adjusted with TEA-OH). Assuming a contamination of 1 μM Ca²⁺, free Ca²⁺ and Mg²⁺ concentrations were calculated to be 0.013 nM and 2.4 mM, respectively (according to Fabiato & Fabiato, 1979). The low-chloride pipette solution contained only 36 mM Cl⁻ instead of 141 mM Cl⁻, with chloride being replaced by aspartate. Thus, when the high- or low-chloride pipette solution was used, the chloride equilibrium potential (E_Cl) equalled -4 or -40 mV, respectively. When the pipette Na⁺ or K⁺ concentration ([Na⁺]_pip and [K⁺]_pip) was increased to 50 or 90 mM, respectively, [TEA⁺] was reduced accordingly. For some experiments, the pipette free Ca²⁺ concentration ([Ca²⁺]_pip) was increased to 15 nM by addition of 0.7 mM CaCl₂ to the pipette solution.

Potential differences between the pipette and superfusion solution were nulled immediately before patch formation. After establishing a gigaohm seal between the pipette tip and the cell membrane, gentle suction was applied to the pipette to achieve the whole-cell configuration. Cell membrane capacitance (C_m) was estimated using at least one of the following procedures: (i) integration of the capacitative current transient during a hyperpolarizing voltage step from 0 to -10 mV, (ii) compensation of the capacitative current transient by means of the built-in devices of the amplifier, or (iii) a program routine applying ± 10 mV voltage ramps from a holding potential (V_h) of 0 mV. All estimates gave nearly identical results. C_m of the myocytes used in this study was 137 ± 6 pF (n = 78).

Steady-state pump currents in Na⁺-containing external solution were estimated as external K⁺ (${\text{K}}_{\text{o}}^{+}$)-activated currents. Control experiments revealed that these currents were almost identical to cardiac glycoside-sensitive currents. When external Na⁺ (${\text{Na}}_{\text{o}}^{+}$)-free solution was used, however, I_p was measured as cardiac glycoside-sensitive current (1 mM dihydro-ouabain; DHO) due to the very high apparent ${\text{K}}_{\text{o}}^{+}$ affinity of the pump under these conditions (K_0.5, ∼0.08 mM; Bielen et al. 1993).

Transient pump currents were recorded as cardiac glycoside-sensitive currents (200 μM DHO) in response to a voltage step (100 ms) under conditions of electroneutral Na⁺-Na⁺ exchange (see Läuger, 1991). The superfusion medium contained 144 mM Na⁺ and zero K⁺. The pipette solution included 5 mM Na⁺ and 10 mM ATP. V_h was either 0 or -40 mV. We refer to these transient pump currents as Na⁺-dependent transient charge movements or Na-TCM. To speed up the voltage clamp, Na-TCM were recorded after cancellation of C_m and compensation of series resistance (R_s; 30–80 %). The remaining capacitative transient decayed with a time constant, τ, of 0.02-0.10 ms, much faster than the fastest Na-TCM (τ = 1.5-2.0 ms). The amount of moved charge (Q) and the rate constant of current decay (k) of Na-TCM were analysed as a function of membrane potential (V). k values were obtained as the reciprocal of the time constant of a monoexponential function fitted to the declining part of the transient pump current beginning 1–3 ms after the voltage step. Q values were then calculated by integration of the monoexponential function between t = 0 ms and t = 100 ms. The sigmoid Q-V relationship was fitted according to the following Boltzmann equation:

where ΔQ_max is the maximum amount of movable charge, Q_min is the bottom of the curve, V_0.5 is the midpoint voltage, and α is a steepness factor corresponding to the equivalent charge. F is Faraday's constant, R is the universal gas constant and T is absolute temperature.

RESULTS

Steady-state Na⁺-K⁺ pump currents of isolated guinea-pig ventricular myocytes were studied at a V_h of -20 mV using low-chloride pipette solutions. E_Cl was -40 mV under these conditions. Thus, an increase in intracellular [cAMP] would result in an outward current shift due to activation of I_Cl(cAMP). I_p was measured at near physiological [Na⁺] of 5 mM pipette Na⁺ (${\text{ Na}}_{\text{pip}}^{+}$) and 144 mM ${\text{Na}}_{\text{o}}^{+}$. To saturate external K⁺-binding sites of the pump, 10 mM ${\text{K}}_{\text{o}}^{+}$ was used for maximal I_p activation. The amplitude of I_p was estimated as the inward current shift upon application of ${\text{K}}_{\text{o}}^{+}$-free solution (see Methods). Current recording was begun after the holding current had reached a steady state, usually 2–3 min after access to the cell interior was achieved. Figure 1A shows an original current trace. At the beginning, I_p under control conditions was estimated by application of ${\text{K}}_{\text{o}}^{+}$-free solution, as indicated by the short horizontal bars above the trace. Its amplitude was 34 pA. The respective trace designated a is shown on an enlarged scale in Fig. 1B. Superfusion of the myocyte with 10 μM 1,9-dideoxyforskolin, a forskolin analogue that does not stimulate adenylyl cyclases (Seamon & Daly, 1985), neither activated I_Cl(cAMP) nor increased I_p. In the presence of 1,9-dideoxyforskolin I_p was slightly reduced to 30 pA (Fig. 1Ab and Bb). By contrast, application of 4 μM forskolin, a concentration previously shown to maximally stimulate I_Cl(cAMP) in these cells (Traebert et al. 1998), activated, after a lag of approximately 20 s, an outward current which was identified as I_Cl(cAMP) (see Fig. 8). Furthermore, as is evident from the repeated application of ${\text{K}}_{\text{o}}^{+}$-free solution, the drug induced a delayed stimulation of I_p that reached maximal amplitude roughly 3 min after the start of the superfusion with forskolin-containing solution. At this point, I_p amplitude was 59 pA(Fig. 1Ac and Bc) or 174 % of the control. Upon wash-out of the adenylyl cyclase activator both the activation of I_Cl(cAMP) and the stimulation of the Na⁺-K⁺ pump current were reversed. Thus, the holding current shifted in the inward direction and the I_p amplitude was reduced to 28 pA (Fig.1Ad and Bd). In a total of 15 ventricular cells, 4 μM forskolin augmented I_p to 133 ± 4 % of the control (see also Fig. 3). Stimulation of I_p was almost always delayed compared with activation of I_Cl(cAMP). Maximal I_p in the presence of forskolin occurred about 2–5 min after superfusion of the myocytes with the drug was begun. Application of H-89, a specific inhibitor of PKA, in the continued presence of forskolin, reduced Na⁺-K⁺ pump current to values below the control. In three cells, in which forskolin had increased I_p to 123 ± 9 % of the control, H-89 (10 μM) reduced I_p to 64 ± 8 % of the control (not shown). The results presented so far suggest that forskolin increased I_p at near physiological concentrations of intra- and extracellular Na⁺ by activation of the cAMP- PKA pathway. Next, we sought to elucidate the mechanism of the forskolin-induced stimulation of I_p.

Corresponding experiments using 50 mM ${\text{ Na}}_{\text{pip}}^{+}$ to nearly saturate internal Na⁺-binding sites of the pump (Nakao & Gadsby, 1989) revealed that forskolin stimulated I_p to 142 ± 3 % of the control (n = 5; see Fig. 3), i.e to nearly the same extent as with a low [Na⁺]_pip of 5 mM. Once again, I_p stimulation by the adenylyl cyclase activator occurred with a considerable time lag compared with I_Cl(cAMP) activation. Whereas I_Cl(cAMP) was completely activated after about 60 s, I_p became maximal only after 2–5 min. Almost identical results were obtained when free [Ca²⁺]_pip was increased from 0.013 to 15 nM. Under these conditions, I_p amplitude was increased by the drug to 138 ± 4 % of the control (n = 9; see Fig. 3). Similarly, when 90 mM pipette K⁺ (${\text{K}}_{\text{pip}}^{+}$) was used to alter intracellular K⁺ deocclusion/release, I_p stimulation by forskolin was unaffected (129 ± 6 % of the control, n = 7; see Fig. 3). In addition, the activation of I_Cl(cAMP) was nearly identical under all conditions (0.39 ± 0.05 pA pF⁻¹ at 5 mM ${\text{ Na}}_{\text{pip}}^{+}$ (n = 19), 0.34 ± 0.08 pA pF⁻¹ at 50 mM ${\text{ Na}}_{\text{pip}}^{+}$ (n = 4), 0.29 ± 0.06 pA pF⁻¹ at 50 mM ${\text{ Na}}_{\text{pip}}^{+}$/15 nM ${\text{Cl}}_{\text{pip}}^{-}$ (n = 9), and 0.29 ± 0.04 pA pF⁻¹ at 50 mM ${\text{ Na}}_{\text{pip}}^{+}$/90 mM ${\text{K}}_{\text{pip}}^{+}$ (n = 5); all P values > 0.05).

Next, we investigated whether forskolin-dependent stimulation of I_p persisted in the absence of extracellular Na⁺ using choline as a Na⁺ substitute. ${\text{Na}}_{\text{o}}^{+}$ is known to cause a voltage-dependent inhibition of steady-state Na⁺-K⁺ pump current (Nakao & Gadsby, 1989) due to voltage-dependent rebinding to the pump molecule in a putative external access channel in which part of the electrical field across the membrane drops (Gadsby et al. 1993). This inhibitory effect of ${\text{Na}}_{\text{o}}^{+}$ is abolished in ${\text{Na}}_{\text{o}}^{+}$-free solution. Figure 2 illustrates an experiment conducted under these conditions. V_h was -20 mV, as before. Switching to a superfusate containing 1 mM DHO shifted the holding current in the inward direction by 80 pA under control conditions (I_p(control)), as shown in the left-hand part of the figure. During wash-out of the cardiac glycoside (not shown) membrane current slowly returned to its initial value. The cell was then challenged with 4 μM forskolin. Forskolin caused an outward shift of the holding current by 35 pA. Application of DHO in the presence of the adenylyl cyclase activator revealed that I_p was increased to 96 pA (I_p(FSK)) or to 120 % of the control. Afterwards, first DHO and then forskolin were washed out (not shown). Four minutes after switching to drug-free solution, the holding current had shifted in the inward direction and I_p was only 44 pA (I_p(wash)) demonstrating that the forskolin effects on I_Cl(cAMP) and I_p were completely reversible. In a total of 10 ventricular myocytes studied using this protocol, forskolin elevated I_p to 116 ± 3 % of the control. It should be noted that application of DHO in the presence of forskolin was always started about 2–3 min after superfusion of the cell with forskolin-containing solution had begun, since the measurements in ${\text{Na}}_{\text{o}}^{+}$-containing superfusate had revealed that I_p stimulation was maximal only after this time lag (see above). The fact that the I_p amplitude after wash-out of forskolin was lower than the control value before forskolin was a consistent finding. This was observed under all experimental conditions (see also Fig. 1) and may be attributed to a rundown of pump current, as described previously in cardiac myocytes (Gadsby & Nakao, 1989; Glitsch & Tappe, 1993).

Figure 3 summarizes the effects of forskolin on I_p in ${\text{Na}}_{\text{o}}^{+}$-free or ${\text{Na}}_{\text{o}}^{+}$-containing superfusate and at various concentrations of pipette Na⁺, K⁺, and Ca²⁺. In ${\text{Na}}_{\text{o}}^{+}$-containing solution, I_p was increased by forskolin to ∼130-140 % of the control, regardless of the composition of the pipette solution. Since the magnitude of I_p stimulation by the drug was almost identical at low and high [Na⁺]_pip, binding of intracellular Na⁺ to the pump seems to be unaltered. Hence, an increase in the apparent affinity of the Na⁺-K⁺ pump to intracellular Na⁺ can be ruled out as the mechanism underlying the stimulation of I_p by the cAMP-PKA pathway. In addition, intracellular K⁺ in the range 0–90 mM and intracellular free Ca²⁺ in the nanomolar to subnanomolar range both do not affect the cAMP-PKA-dependent regulation of I_p.

In ${\text{Na}}_{\text{o}}^{+}$-free superfusate, I_p was still increased by forskolin. The increase, however, was less pronounced than in ${\text{Na}}_{\text{o}}^{+}$-containing solution. Whereas in the presence of extracellular Na⁺ the I_p amplitude was augmented by forskolin to ∼130-140 % of the control, this effect was only about one-half as large (116 %) in the absence of extracellular Na⁺.

To examine whether partial reactions in the Na⁺ limb of the pump cycle are involved in the stimulation of the Na⁺-K⁺ pump current by forskolin, we studied the effects of the drug on transient pump currents under conditions of electroneutral Na⁺-Na⁺ exchange (Na-TCM). Na-TCM indicate voltage-dependent alterations of the equilibrium between the two main conformations of the pump, E1 and E2. In the E1 conformation Na⁺-binding sites face the cytosol, whereas in the E2 conformation they are orientated towards the extracellular space. Figure 4A illustrates original current traces of Na-TCM of a ventricular myocyte. V_h was 0 mV. Na-TCM were obtained in response to a voltage step (indicated by the vertical arrows) to +60 mV (upper traces) or -80 mV (lower traces) 5 min (left), 10 min (middle) and 15 min (right) after access to the cell interior was achieved. At -80 mV (generally at negative potentials) transient inward currents were observed indicating that the E1-E2 equilibrium was shifted in favour of E1 (see Nakao & Gadsby, 1986). The opposite was true for the voltage jump to +60 mV (generally to positive voltages). The currents declined monoexponentially with voltage-dependent time constants as shown above or below the traces. Current decay occurred much faster at -80 than at +60 mV. In addition, prolonged cell dialysis resulted in an increase in the time constants at both potentials. The respective k-V relationships are shown in Fig. 4B (^, 5 min; □, 10 min; ▵, 15 min). Rate constants increased at negative potentials and were almost voltage independent at positive voltages. Furthermore, after 10 or 15 min k values were decreased at all potentials investigated. The relative decrease in the rate constants, however, was more pronounced at positive potentials. The Q-V relationships are displayed in Fig. 4C (symbols as in Fig. 4B). Five minutes after establishment of the whole-cell configuration, the midpoint voltage, which is a measure of the voltage-dependent equilibrium between the E1 and the E2 conformation, was -43.1 mV. It was shifted to the more positive values of -9.2 and -7.9 mV after 10 and 15 min, respectively, indicating that the E1-E2 transition was decelerated. Simultaneously, ΔQ_max was reduced by ∼25 % from 39.0 fC pF⁻¹ after 5 min to 30.5 fC pF⁻¹ after 15 min. On average, V_0.5 and ΔQ_max, respectively, after 5 min were -42.6 ± 2.7 mV and 44.4 ± 2.4 fC pF⁻¹ and after 10 min -24.8 ± 4.4 mV and 37.2 ± 2.8 fC pF⁻¹ (n = 8). Whereas the drop in V_0.5 was statistically significant (P < 0.01) the difference in ΔQ_max was not (P = 0.07). Assuming a specific membrane capacitance of 1 μF cm⁻² and translocation of one elementary charge per pump molecule during Na-TCM, a ΔQ_max of 35–45 fC pF⁻¹ translates into 2200–2800 Na⁺-K⁺ pumps per square micrometre. This number is close to previous estimates for guinea-pig (1200-3200 μm⁻², Nakao & Gadsby, 1986; Collins et al. 1992) or rat ventricular myocytes (2600 μm⁻², Dobretsov et al. 1998).

Figure 5 shows the effect of 2 μM forskolin on Na-TCM. In this experiment, V_h was -40 mV. Original current traces of Na-TCM during voltage steps to +20 mV (upper traces) and -100 mV (lower traces) are illustrated in Fig. 5A. First, Na-TCM were recorded under control conditions (left). Then, the myocyte was challenged with forskolin. Na-TCM in the presence of the drug were obtained 10 min after access to the cell interior was achieved (middle). Finally, Na-TCM were recorded after wash-out of forskolin (right). In the presence of the adenylyl cyclase activator the time constants were hardly changed compared with the control before application of forskolin. This can also be seen in the respective k-V relationships in Fig. 5B. Rate constants in the presence of forskolin (▪) were similar to the values obtained before application of the drug (^). Hence, compared with control conditions (Fig. 4B) the k values 10 min after establishment of the whole-cell configuration were increased by the drug. At the same time, the Q-V relationship (Fig. 5C) was shifted to the left by forskolin (▪), i.e. in favour of E2, from a V_0.5 of -52.1 mV to a V_0.5 of -59.4 mV. The drug effect was reversible as is evident both from the slowing of the rate constants (▵, Fig. 5B) and the positive shift of the Q-V curve (V_0.5: -21.1 mV, ▵, Fig. 5C) after wash-out.

The forskolin effect on Na-TCM was due to activation of the cAMP-PKA pathway, as illustrated in Fig. 6. Midpoint voltages are displayed as a function of time under various experimental conditions. Under control conditions (Fig. 6A), V_0.5 became progressively more positive with the duration of cell dialysis: -42.2 mV after 5 min and -10.0 mV after 15 min. By contrast, application of 2 μM forskolin (Fig. 6B) shifted V_0.5 from -52.1 to -59.4 mV. After wash-out of the drug the midpoint voltage was much more positive amounting to -21.1 mV (values taken from Fig. 5). When 10 μM PKI was present in the pipette solution to inhibit PKA, forskolin had no effect (Fig. 6C). Under these conditions, a marked positive shift of V_0.5 from -63.7 to -30.5 mV was observed despite superfusion of the cell with a medium containing 2 μM forskolin. Finally, application of 500 μM 8-bromo-cAMP mimicked the forskolin effect (Fig. 6D). The cAMP analogue caused a negative shift of V_0.5 from -41.6 mV under control conditions to -52.0 and -46.6 mV after 10 and 15 min, respectively. Mean values for V_0.5 obtained after 10 min of cell dialysis under various conditions are shown in Fig. 6E. In the absence of any drugs V_0.5 was -24.8 ± 4.4 mV (n = 8). In the presence of forskolin (2-10 μM), the midpoint voltage was -53.7 ± 2.3 mV (n = 11, P < 0.01 compared with the control). The significant negative shift of V_0.5 induced by the adenylyl cyclase activator was counteracted by the inclusion of PKI in the pipette solution. Under these conditions, V_0.5 reached -37.7 ± 4.4 mV (n = 5, P < 0.01 compared with forskolin). Mean values of ΔQ_max for the same batches of cells were 37.2 ± 2.8 fC pF⁻¹ under control conditions, 35.8 ± 2.4 fC pF⁻¹ in the presence of forskolin and 34.8 ± 3.9 fC pF⁻¹ in the additional presence of PKI (not shown, P values > 0.05). Thus, forskolin did not change the number of Na⁺-K⁺ pumps involved in Na⁺-Na⁺ exchange. Figures 5C and 6B demonstrate that forskolin shifted the V_0.5 of the Q-V curve to potentials more negative than the initial control values. Thus, the drug facilitates the E1-E2 transition in the Na⁺ limb of the pump cycle by a mechanism which must be different from a simple prevention of rundown of V_0.5 with time. However, the results presented in Fig. 5B may be interpreted to mean that forskolin inhibits the rundown of pump activity. To address this point, the time-dependent changes in V_0.5 were compared in drug-free and drug-containing media. In this series of experiments, forskolin was applied 2 min after access to the cell interior was achieved to facilitate comparison of the V_0.5 values obtained during the whole course of the measurements. Furthermore, to increase PKA-dependent phosphorylation some experiments were conducted in the additional presence of 10 μM microcystin in the pipette solution. Microcystin is a specific inhibitor of protein phosphatases 1 and 2A (MacKintosh et al. 1990). Thus, if forskolin affected the V_0.5 rundown via activation of PKA, this effect would be expected to increase in the additional presence of microcystin. Figure 7A–C shows plots of V_0.5versus time. Each circle represents a V_0.5 measurement. In order to quantify the rundown of V_0.5, regression lines were arbitrarily fitted to the data. Irrespective of the conditions chosen, the midpoint voltage monotonously increased towards more positive potentials. The slopes of the fitted lines were almost identical. Extrapolation to zero time yields midpoint voltages of -45 mV under control conditions (Fig. 7A) and of approximately -65 mV in media containing forskolin (Fig. 7B and C). After 5 min of whole-cell recording the mean value of V_0.5 was -42.6 ± 2.7 mV (n = 8) under control conditions. The corresponding values from forskolin-treated myocytes were calculated to be -60.4 ± 4.5 mV (forskolin, n = 9) and -60.5 ± 3.5 mV (forskolin + microcystin, n = 6), both being significantly more negative than the control value (P < 0.05). Therefore, forskolin facilitates the E1-E2 transition by a mechanism different from that responsible for the rundown of V_0.5.

Apart from accelerating the E1-E2 transition of the pump cycle forskolin activated a membrane conductance that was identified as the cAMP-dependent chloride conductance, as shown in Fig. 8 (see also Figs 1 and 2). Figure 8A displays original membrane currents in response to voltage jumps to potentials between +80 and -120 mV. V_h was -40 mV (same recording as in Fig. 5). Currents obtained in the absence and presence of 200 μM DHO are shown superimposed. They differed immediately after the voltage step (most easily seen in the middle panel) but were identical in the steady state, as expected for Na-TCM. It is clear that application of forskolin reversibly increased the membrane conductance. Steady-state membrane currents measured at the end of each voltage jump before (^), during (▪) and after (▵) application of forskolin are plotted versus membrane potential in Fig. 8B. The I–V relationship of the forskolin-activated membrane conductance is illustrated in Fig. 8C. The values shown represent the difference between the currents measured in the presence and in the absence of forskolin (from Fig. 8B). The drug-induced current showed slight outward rectification and reversed its sign at approximately -7 mV. Similarly, the mean I–V curve obtained from seven cells under these conditions (• in Fig. 8D) exhibited slight outward rectification and a reversal potential of -5 mV, which was close to the calculated E_Cl of -4 mV. When 10 μM PKI was included in the pipette solution (^ in Fig. 8D, n = 5) almost no current was activated by forskolin implying that it was dependent on PKA activity. Finally, when E_Cl was set to -40 mV (▪ in Fig. 8D, 36 mM ${\text{Cl}}_{\text{pip}}^{-}$, n = 4) the forskolin-induced current reversed at -37 mV and displayed even more pronounced outward rectification than at near symmetrical chloride concentrations. These properties are characteristic for the cAMP-dependent chloride current known to be present in these cells.

DISCUSSION

The main findings of the present study are that (i) the Na⁺-K⁺ pump current of guinea-pig ventricular myocytes is increased by forskolin via activation of the cAMP-PKA pathway and (ii) this increase is mediated, at least in part, by a modulation of one or more partial reactions in the Na⁺ limb of the pump cycle.

The first conclusion is based on the observation that forskolin, which is known to raise intracellular cAMP levels by activation of adenylyl cyclases, reversibly increased I_p under various experimental conditions (Fig. 3). In contrast, 1,9-dideoxyforskolin, which does not activate adenylyl cyclases, was unable to mimic the forskolin effect. Furthermore, a cAMP analogue produced forskolin-like effects on Na-TCM (Fig. 6D) and two inhibitors of PKA, H-89 and PKI, abolished the forskolin action on I_p or Na-TCM, respectively. The results, therefore, demonstrate that a cAMP-dependent activation of PKA was responsible for the observed effects of forskolin on both steady-state and transient pump currents in these cells.

The forskolin-induced increase in I_p was ∼130-140 % of the control in Na⁺-containing extracellular solution, irrespective of the composition of the pipette solutions used. Similar increases in I_p upon activation of β-adrenergic receptors by catecholamines were described in ventricular myocytes (140-145 %, Dobretsov et al. 1998) and skeletal myoballs (150 %, Li & Sperelakis, 1994) of the rat. Also, Gao et al. (1994) found that at micromolar concentrations of ${\text{Cl}}_{\text{pip}}^{-}$ Na⁺-K⁺ pump current in guinea-pig ventricular myocytes was augmented by cAMP-increasing drugs to about 125 % of the control. Thus, a general picture seems to emerge that activation of the cAMP-PKA cascade in striated muscle, at least in rat and guinea-pig, results in an increased I_p. However, the molecular mechanism underlying this increased pump current may be different between the two species, as discussed below.

A major difference between the results presented above and those of Gao and co-workers is the [Ca²⁺]_pip dependence of the I_p regulation by the cAMP-PKA pathway. Whereas our measurements clearly showed that I_p was stimulated in a cAMP-dependent manner at nanomolar and subnanomolar concentrations of ${\text{Cl}}_{\text{pip}}^{-}$ (Fig. 3), Gao et al. (1992, 1996, 1998) reported that at concentrations < 150 nM ${\text{Cl}}_{\text{pip}}^{-}$I_p was decreased in the very same cell type under otherwise comparable experimental conditions. Only at concentrations > 150 nM ${\text{Cl}}_{\text{pip}}^{-}$ did they observe an increased I_p upon isoprenaline application (Gao et al. 1992, 1994, 1996). The reason for this discrepancy remains unknown at present and clearly needs to be investigated further. It should be noted, however, that the catecholamine-induced stimulation of I_p in rat ventricular myocytes also occurred at nanomolar [Ca²⁺]_pip (Dobretsov et al. 1998) and that an elevation of [Ca²⁺]_pip up to 1 μM likewise resulted in an increased I_p amplitude in response to isoprenaline (Stimers & Dobretsov, 1998).

Besides increasing I_p, forskolin also activated the cAMP-dependent chloride current (I_Cl(cAMP)), which was identified by its dependence on activation of the cAMP-PKA pathway and its I–V characteristics (Fig. 8). I_Cl(cAMP) served as a positive control for the effects on I_p of agents intervening with the cAMP signalling cascade. Thus, when the forskolin action on I_p was blocked by PKA inhibitors, activation of I_Cl(cAMP) was also abolished. In our measurements of steady-state pump current (e.g. Fig. 1), activation of I_Cl(cAMP) due to forskolin application was visible as an outward shift of the holding current. This outward current developed about 15–25 s after starting superfusion of the myocyte with forskolin and was complete within another minute, very similar to the time course previously reported by Pelzer and co-workers (Pelzer et al. 1997). Interestingly, stimulation of I_p only occurred once this activation of I_Cl(cAMP) had commenced. In most cases, I_p amplitude began to rise after a time lag of about 50–120 s and was maximal roughly 2–5 min after application of forskolin was begun. The reason for these different time courses remains unknown, but it can be speculated that one or more components of the cAMP-PKA pathway may have preferential access to the chloride channels. Further studies to clarify this issue are underway.

In principle, several (sufficiently slow) partial reactions in the Na⁺-K⁺ pump cycle may be affected by forskolin to cause the observed increase in I_p. According to Dobretsov et al. (1998), the relative increase in I_p in rat ventricular myocytes depends on [K⁺]_pip. At zero ${\text{K}}_{\text{pip}}^{+}$, isoprenaline augmented I_p amplitude to ∼120 % of the control while at 100 mM ${\text{K}}_{\text{pip}}^{+}$I_p was elevated to ∼160 % of the control. The data were fitted with a cyclic model of the Na⁺-K⁺ pump and it was concluded that enhanced intracellular deocclusion/ release of K⁺ was responsible for the observed stimulation of the pump current. A similar mechanism, however, does not seem to be operative in guinea-pig ventricular myocytes, since our measurements showed that the forskolin-induced increase in I_p amplitude was almost identical at 0 and 90 mM ${\text{K}}_{\text{pip}}^{+}$ (Fig. 3). In addition, although not studied in detail, an increased apparent affinity of the pump towards extracellular K⁺ can probably be excluded as the reason for the augmented I_p, since saturating concentrations of this external activator cation of the pump were used in all our experiments. This conclusion is further strengthened by the observation of Gao et al. (1992) that an increase in [K⁺]_o from 4.6 to 20 mM did not affect the β-adrenergic stimulation of I_p in the same cell type. Hence, neither intra- nor extracellular K⁺ binding/release seem to be involved in the forskolin-induced increase in I_p. We cannot, however, exclude the possibility that other partial reactions in the K⁺ limb of the pump cycle are modified by the cAMP-PKA cascade.

To isolate partial reactions in the Na⁺ limb of the pump cycle transient pump currents under conditions of electroneutral Na⁺-Na⁺ exchange were recorded. Under these conditions, and the E2 conformation of the Na⁺-K⁺ pump exists. Stepping a voltage-dependent equilibrium between the E1 membrane voltage to positive potentials shifts the equilibrium in favour of E2 while clamping the membrane voltage to negative potentials shifts it towards E1. Five minutes after establishment of the whole-cell configuration a midpoint voltage, V_0.5, of approximately -45 mV was obtained, which is in good agreement with previous measurements on rat (-25 to -30 mV, Dobretsov et al. 1998) or guinea-pig ventricular myocytes (-20 mV, Nakao & Gadsby, 1986) or giant membrane patches thereof (-45 mV, Hilgemann, 1994; -37 to -52 mV, Friedrich & Nagel, 1997). During sustained cell dialysis this value became more positive and was -25 mV after 10 min. This rightward shift of the Q-V curve indicates that the E1-E2 transition was impaired. The phenomenon is reminiscent of the rundown of I_p as described in this and other studies on cardiac myocytes (Gadsby & Nakao, 1989; Glitsch & Tappe, 1993). By contrast, forskolin caused a significant leftward shift of the Q-V curve. In its presence, V_0.5 was significantly more negative than the initial control value (Figs 6B and 7) indicating that the E1-E2 transition was facilitated by a mechanism different from prevention of V_0.5 rundown, which was not affected by forskolin (Fig. 7). In line with this finding rate constants under forskolin were larger than the respective controls at the same time (compare Figs 4B and 5B). The fact that the rate constants of transient pump current decline in the presence of forskolin did not exceed the initial values in its absence (Fig. 5B) is not inconsistent with this conclusion. One has to take into account that the measured rate constants k are the sum of the rate constants of the forward and backward reaction of the E1-E2 transition, k_1-2 and k_2-1. It is easily possible that forskolin increases k_1-2 and simultaneously decreases k_2-1 without large changes in k. The effects of forskolin on Na-TCM thus demonstrate that activation of the cAMP-PKA pathway facilitates the E1-E2 transition in the Na⁺ limb of the pump cycle, a result that may well explain the observed increase in I_p.

A further important conclusion can be drawn from these measurements. Since forskolin did not change the number of pump molecules involved in Na⁺-Na⁺ exchange, as judged from the identical ΔQ_max values in its presence or absence, respectively, the recruitment of additional Na⁺-K⁺ pumps to the plasma membrane can be excluded as the underlying cause for the augmented I_p. Such a mechanism has recently been shown to be responsible for the cAMP-dependent increased cation transport and hydrolytic activity of the Na⁺-K⁺-ATPase in rat kidney (Carranza et al. 1998).

From the above reasoning it is clear that a modulation of partial reactions in the Na⁺ limb of the pump cycle is probably at least one cause for the increase in I_p observed upon activation of the cAMP-PKA cascade in guinea-pig ventricular myocytes. The question, then, concerning which partial reaction(s) is (are) involved remains. Although from our results no definite answer can be given to that question, some limitations may apply. Firstly, an increase in the apparent affinity of the Na⁺-K⁺ pump for intracellular Na⁺ cannot account for the augmented I_p, since the quantitative rise in pump current at both low (5 mM) and high (50 mM) ${\text{ Na}}_{\text{pip}}^{+}$ was almost identical. The same conclusion was drawn previously by Gao et al. (1992). Secondly, extracellular Na⁺ deocclusion/release may be altered and could be at least one mechanism for the increased I_p. This hypothesis is based on the finding that forskolin stimulated I_p significantly less in the absence of extracellular Na⁺. One possible mechanism involved may be, for example, that the voltage-dependent rebinding of Na⁺ in the putative external access channel, which is absent in ${\text{Na}}_{\text{o}}^{+}$-free solution, is reduced by forskolin. Consequently, I_p would be increased by the drug in Na⁺-containing solution. This interpretation must, however, be considered with caution for the following two reasons. (i) I_p ${\text{Na}}_{\text{o}}^{+}$-free solution was measured as DHO-sensitive current. Wash-out of DHO occurred rather slowly (∼1.5-3 min) so that there was a considerable delay between the estimation of I_p before and following forskolin application. During this time rundown of I_p may have proceeded and I_p amplitude in the presence of the adenylyl cyclase activator may have been underestimated. (ii) Because of this time-consuming wash-out of DHO only one determination of I_p magnitude during forskolin treatment of a cell was possible. This estimation occurred 2–3 min after forskolin application had started. However, maximal I_p in the presence of the drug was reached at quite variable times, usually 2–5 min from the beginning of application. Thus, maximal stimulation may not have been reached at that time resulting in an (additional) underestimate of I_p. Taking this into account, the only firm conclusion from our experiments is that (a) partial reaction(s) in the Na⁺ limb is (are) modified by the cAMP-PKA pathway leading to a facilitated E1-E2 transition.

Sympathetic stimulation of the heart leads, via activation of β-adrenoceptors, to increased action potential frequencies and passive Na⁺ and K⁺ fluxes. In order to maintain the Na⁺ and K⁺ gradients across the sarcolemma, which are essential for cardiac excitation, active Na⁺-K⁺ transport would also have to be increased under these conditions. The stimulation of the Na⁺-K⁺ pump by the cAMP-PKA pathway, as documented in this study, may be one means by which large changes in intracellular Na⁺ and K⁺ are prevented.