Abstract Regulation of L-type Ca2+ channel current [ICa(L)] by cGMP-dependent protein kinase (PK-G) was investigated in ventricular myocytes from 2- to 21-day-old rats using whole-cell voltage clamp with internal perfusion. ICa(L) was elicited by a depolarizing pulse to +10 mV from a holding potential of −40 mV. Stimulated ICa(L) (by 2 μmol/L isoproterenol) was inhibited to the basal level by internal perfusion with 50 nmol/L PK-G (activated by 8Br-cGMP, 0.1 μmol/L). When ICa(L) was enhanced by Bay K 8644 (1 μmol/L), the enhanced basal ICa(L) was also reduced by PK-G. Basal ICa(L) (nonstimulated through the cAMP/cAMP-dependent protein kinase [PK-A] pathway) was also inhibited to various degrees (large, medium, or small) by internal application of PK-G (25 nmol/L). The average inhibition was 42.1% (n=36), and there were no differences in the inhibition during development. The inhibition by PK-G was blocked by the PK-G substrate peptide (cG-PKI, 300 μmol/L) and by heat inactivation of the PK-G. Relatively specific PK-G inhibitors (eg, cG-PKI and H-8) sometimes reversed the inhibition (5 of 25 cells), whereas isoproterenol stimulated ICa(L) (7 of 8 cells). When a holding potential of −80 mV was used, the inhibition produced by PK-G was much less. The inhibitory effects of PK-G were not mediated by activating phosphodiesterase or protein phosphatase but most likely by a direct phosphorylation of the Ca2+ channel or associated regulatory protein. The inhibitory effect of PK-G may be explained by a balance between activities of PK-A and PK-G in regulating the slow Ca2+ channels at two separate sites.
- Ca2+ slow channels
- whole-cell voltage clamp
- patch clamp
- cGMP-dependent protein kinase
- internal perfusion technique
The voltage- and time-dependent slow (L-type) Ca2+ channels in the cell membrane are the major pathway by which Ca2+ ions enter the cell during excitation for initiation and regulation of the force of contraction of the muscle. The channel can be controlled by extrinsic factors (such as autonomic nerve stimulation or circulating hormones) and by intrinsic factors (such as cellular pH or ATP level). One major mechanism for regulation of the L-type Ca2+ channels is elevation of cAMP level and stimulation of PK-A phosphorylation of the channel or a closely associated regulatory protein. Such phosphorylation results in an increase in the open probability of the channel1 2 and in the enhancement of peak amplitude of ICa(L).3
It has been proposed that cGMP plays a role antagonistic to that of cAMP.4 The antagonism produced by cGMP/PK-G was reported to involve the direct phosphorylation of the Ca2+ channel or a closely associated regulatory protein by PK-G in guinea pig and rat ventricular cells.5 6 7 8 9 It was reported that a single protein of ≈47 kD is specifically phosphorylated by PK-G (in the presence of 10−5 mol/L cGMP) in guinea pig sarcolemmal preparations,10 and this protein may be a possible mediator involved in regulation of Ca2+ channels of the heart by the cGMP/PK-G pathway. cGMP was also reported to increase cAMP degradation by a cGMP-stimulated PDE (PDE-II) in frog ventricular myocytes.11 12 It has been reported that under some conditions, cGMP may actually further stimulate the prestimulated ICa(L) rather than inhibit it, presumably by preventing cAMP degradation by cGMP inhibition of a phosphodiesterase (PDE-III).13
The effect of cGMP on basal ICa(L) (not stimulated through the cAMP/PK-A pathway) is still controversial. Although some investigators reported that there were no effects of cGMP on basal ICa(L),9 10 11 12 13 we have evidence that cGMP does have an inhibitory effect on basal ICa(L). Intracellular application of cGMP (by pressure injection,5 by the liposome method,14 or by using the membrane-permeable derivative 8Br-cGMP)6 8 inhibited the Ca2+-dependent slow action potentials in guinea pig papillary muscles and chick embryonic heart cells. We also reported that 8Br-cGMP inhibited basal ICa(L) in chick embryonic heart cells by using whole-cell voltage-clamp8 and cell-attached patch-clamp15 techniques. Recently, we found that in embryonic heart cells, intracellular application of PK-G via the patch pipette inhibited the basal and the prestimulated ICa(L)s.16
In the present study, whole-cell voltage clamp (with internal perfusion technique) was performed on young rat ventricular myocytes to examine the effect of PK-G on ICa(L). We found that PK-G inhibited both the basal ICa(L) as well as the prestimulated ICa(L).
Materials and Methods
Freshly isolated single cells were prepared from ventricles of young rats (Sprague-Dawley) as previously described.17 The rats were divided into four age groups: 3-day (2- to 4-day-old), 7-day (6- to 8-day-old), 11-day (10- to 13-day-old), and 20-day (19- to 21-day-old) rats. The rats were anesthetized by exposing them to an atmosphere of pure CO2 for a few minutes and then decapitated. The hearts were dissected and rinsed in oxygenated normal Tyrode’s solution and then immersed in nominally Ca2+-free Tyrode’s solution.17 The ventricles were dissected from the atria after the spontaneous contractions had ceased (≈15 minutes), and small pieces of ventricular tissue were enzymatically digested for 50 minutes (at 37°C) in the Ca2+-free Tyrode’s solution containing 1 mg/mL collagenase (Yakult). After incubation, the tissues were rinsed at least three times in modified KB solution17 at room temperature, and the cells were mechanically dispersed with a wide-bore pipette. The cell suspension was stored at 4°C in the modified KB solution. Cells were used between 2 and 6 hours after isolation.
Solutions and Drugs
The Na+-free and K+-free external solution contained (mmol/L) tetraethylammonium chloride 150, CaCl2 1.8, MgCl2 0.5, 4-AP 3, HEPES 5, and glucose 5.5 (pH 7.4 using HCl), so as to isolate the Ca2+ current from the Na+ and K+ currents. The internal (pipette) solution contained (mmol/L) CsOH 110, CsCl 20, l-glutamic acid 90, MgCl2 3, ATP-Na2 5, phosphocreatine disodium salt 5, EGTA 10, and HEPES 5 (pH 7.2 with CsOH).
ISO, H-8, and Rp-8-pCPT-cGMPS were dissolved in distilled water to provide a stock solution (1 mmol/L, 5 mmol/L, and 5 mmol/L, respectively). Bay K 8644 and MC were dissolved in ethanol to provide 10 mmol/L and 1 mmol/L stock solutions, respectively. IBMX was dissolved in DMSO to provide a 100 mmol/L stock solution. ISO, H-8, Rp-8-pCPT-cGMPS, Bay K 8644, and IBMX were diluted to appropriate concentrations and applied extracellularly. It was confirmed that the solvents used, ethanol (up to 0.05%) and DMSO (up to 0.1%), did not affect ICa(L). PK-G, cG-PKI, and MC were directly dissolved in the internal solution for intracellular application at a final concentration of 25 or 50 nmol/L for PK-G, 300 μmol/L for cG-PKI, and 0.5 μmol/L for MC. The PK-G solution was always accompanied by a potent and unhydrolyzable PK-G activator, 8Br-cGMP (0.1 μmol/L). PK-G and cG-PKI were stored at 4°C and used within 3 weeks. The sources of these drugs were as follows: ISO, H-8, 8Br-cGMP, IBMX, and MC from Sigma Chemical Co, PK-G and cG-PKI from Promega, Rp-8-pCPT-cGMPS from BioLog Life Science Institute, and Bay K 8644 from Dr Alexander Scriabine (Miles Laboratory).
Whole-Cell Voltage-Clamp Recording
Whole-cell voltage-clamp recordings, with the perfusion pipette technique, were made by using an EPC-7 patch-clamp amplifier (List-Electronic). The patch pipettes had resistances of 1.5 to 3 MΩ (when filled with the internal solution). Series resistance was partly compensated electrically (10% to 30%). The cells were placed into a perfusion chamber (which contained the external solution) located on an inverted microscope (Nikon) and allowed to settle for at least 5 minutes. Then they were constantly perfused with the external solution at 3 mL/min at room temperature (22°C to 25°C). All solutions containing drugs or proteins were applied extracellularly and/or intracellularly after the ICa(L) had stabilized (usually 3 minutes after breaking into the cell and obtaining the whole-cell configuration). The liquid junction potential between the internal solution and the external solution was −2.0±0.2 mV (mean±SEM, n=7). This value was so small and negligible that potentials given in the data were not corrected for the junction potential.
The ICa(L) currents were recorded at 0.1 Hz from an HP of −40 mV (pulses of 300-millisecond duration), unless otherwise mentioned, to exclude any ICa(T) or the possibility of Ca2+ traversing the fast Na+ channel. The currents were abolished completely by 2 mmol/L Co2+ or 0.5 mmol/L Cd2+, consistent with the current being carried through a Ca2+ channel. To see the effect of the PK-G, peak ICa(L) was generated with pulses to +10 mV. Current-voltage curves were obtained by applying voltage steps (300 milliseconds in duration) in 10-mV increments (−30 to +70 mV) from an HP of −40 mV. The inactivation curve was determined by using a standard double-pulse protocol: 3000-millisecond conditioning pulses of various amplitudes (from an HP of −40 mV) were followed by a test pulse to 10 mV (300 milliseconds in duration). A 5-millisecond gap was set between the conditioning pulse and test pulse (to −40 mV) to allow for resetting of the activation gate.
Leak and capacitance currents were subtracted by using currents elicited by small hyperpolarizing pulses (P/5 protocol), unless otherwise mentioned. Current signals were filtered with a cutoff frequency of 1 kHz (eight-pole Bessel) and sampled at 3 kHz. Current and voltage signals were stored on an IBM-AT–compatible personal computer; the pclamp program (Axon Instruments) was used for further analysis. Current density was calculated from the measured membrane capacitance. Membrane capacitance was determined by applying ramp voltage pulses (−0.5 V/s) from an HP of 0 mV; no ionic currents were activated during the voltage pulses.
Internal Perfusion Technique
PK-G and/or cG-PKI was applied intracellularly via the pipette by using a modified internal perfusion technique devised by Kameyama et al.3 A thin quartz tubing (Adams & List Associates) was connected with a long (≈80 cm) thin (internal diameter, 0.28 mm) polyethylene tubing (Clay Adams) and inserted into the glass pipette so that the tip of the quartz tubing was close to the pipette tip. A change of solution in the pipette (perfusion rate, 4 to 6 μL/min) was carried out by applying a negative pressure to the back end of the glass pipette. The perfusion quartz tubing was first filled with the control pipette solution to ≈10 cm from the tip to prevent premature diffusion into the pipette tip and the cell interior.
A tiny bubble (≈1 to 1.5 mm) was inserted into the thin tubing to separate the control internal solution and the subsequent experimental solution (containing the test substance). When two agents were to be internally perfused sequentially, another bubble was added for separation. The bubbles were useful to identify the change of solution, to control speed of perfusion by optically watching their movement, and to reduce the electrical noise pickup (60 Hz). In control experiments, the tiny bubble did not alter the peak amplitude of ICa(L) that was recorded when it reached the end of the quartz tubing (n=8, data not shown). Before perfusion was started, the polyethylene tubing was completely clamped in the middle by a string attached to a manipulator. The efficacy of the internal perfusion technique was checked by observing an increase in ICa(L) with cAMP (100 or 500 μmol/L) added to the pipette as a stimulatory agent (n=4, data not shown).
Current amplitudes were measured as the peak inward current. Experiments were discarded if the analyzed time course clearly showed substantial rundown during the control period before application of PK-G.
The dose-response curve was fitted by the Hill equation: where IC50 is the dose of half-inhibition, nH is the Hill coefficient, and Max is the maximum inhibition.
The inactivation curves were fitted using the conventional Boltzmann equation: where Vh is the potential at which half-inactivation occurs, k is the slope factor, and Max and Min are the maximum and minimum values, respectively, in the inactivation curve.
All data are presented as mean±SEM with the number of cells in parentheses. Statistical evaluation was made by the paired t test for PK-G effect and by ANOVA for comparing more than three groups. A value of P<.05 was considered to be statistically significant.
PK-G Inhibition of Prestimulated ICa(L)
ICa(L) of rat ventricular myocytes was prestimulated by 2 μmol/L ISO, and the effect of PK-G (50 nmol/L) was examined. The concentration of 2 μmol/L was reported to exert almost maximum stimulatory effect in newborn and adult rabbit and rat ventricular cells.18 19 We first checked the stimulatory effect of ISO (2 μmol/L) on ICa(L). The average stimulation by ISO was 83.4±14.9% for myocytes from the 3-day (n=4), 110±25.5% for 11-day (n=4), and 112±17.5% for 20-day (n=8) groups. ISO could maintain a sustained stimulatory effect for up to 15 to 20 minutes (n=3; Fig 1A⇓, inset). After ICa(L) stimulation reached maximum, PK-G (50 nmol/L) was perfused intracellularly via the perfusion patch pipette. A low concentration of 8Br-cGMP (0.1 μmol/L) always accompanied the PK-G in order to activate its enzymatic activity (since there is no catalytic subunit). Fig 1A⇓ shows a typical experiment. As can be seen, prestimulated ICa(L) was inhibited to about the basal level by PK-G application within 9 minutes after the start of perfusion. Similar results were obtained in a total of six cells in variable stages of development (n=2 for 3-day, n=1 for 11-day, and n=3 for 20-day groups). The steady state level of inhibition, which was almost the same as or lower than the control level (−9.2±13% from the control level, n=6), was reached within 4 to 10 minutes after PK-G application. After PK-G perfusion was stopped, no recovery was evident within 5 minutes (n=4, data not shown).
A second method of stimulating ICa(L), which does not involve the cAMP/PK-A pathway, was also used. ICa(L) was stimulated by the Ca2+ channel agonist Bay K 8644 (1 μmol/L). A representative experiment is shown in Fig 1B⇑. PK-G (50 nmol/L) reversed the Bay K 8644 stimulation of ICa(L) back to the control level within ≈3 minutes after PK-G application. Similar experiments were obtained in a total of three cells. The decrease in ICa(L) reached a steady state level, which was lower (−28.2±15.2%, n=3) than the control level, within 3 to 5 minutes after PK-G application. One control experiment showed that the enhancement produced by Bay K 8644 was maintained for 5 minutes (Fig 1B⇑, inset). Therefore, the inhibitory effect produced by PK-G on ICa(L) was not caused by desensitization to Bay K 8644. Since Bay K 8644 has been reported not to increase cAMP levels20 or to activate endogenous protein kinase activity in the heart,21 the inhibitory effect of PK-G on Bay K 8644–enhanced ICa(L) may reflect the endogenous activity of PK-A.
PK-G Inhibition of Basal ICa(L)
Similar experiments were performed on the basal (not stimulated through the cAMP/PK-A pathway) ICa(L). Since a low dose (0.1 μmol/L) of the potent and unhydrolyzable PK-G activator, namely 8Br-cGMP, was present to activate the enzymatic activity, we first tested its effect. 8Br-cGMP itself had only a slight inhibitory effect on basal ICa(L) (average inhibition, 10.5±1.9%; n=8; Fig 2⇓). However, when 25 nmol/L PK-G was perfused together with the 8Br-cGMP, ICa(L) was markedly inhibited: by 47.2±8.6% from the control level (n=5, Fig 2A⇓). As expected, higher doses of 8Br-cGMP (1 or 10 μmol/L) caused substantial inhibition of the basal ICa(L) (Fig 2B⇓). The steady state level of inhibition was reached within 6 minutes after PK-G application. As can be seen in Fig 2A⇓ (inset), after the application of PK-G, a small outward current sometimes appeared. The calculated values of IC50, the Hill coefficient, and percent inhibition (maximum) were 0.29 μmol/L, 1.32, and 52.5%, respectively (Fig 2B⇓).
The leak-subtracted current-voltage curves for basal ICa(L) before and after application of 25 nmol/L PK-G (n=11) are summarized in Fig 3A⇓. PK-G produced a decrease in ICa(L) at all test potentials without changing the apparent threshold potential (≈−30 mV) and potential for peak current (between +10 and +20 mV). The inhibitory effect was statistically significant at potentials of +10 mV and higher. At a test potential of +10 mV, PK-G inhibited the basal ICa(L) by 46.6% (from −8.35±1.94 to −4.46±0.70 pA/pF, n=11). The reversal potential was shifted to the left by ≈12 mV after PK-G application. Typical current traces of ICa(L) before and after application of PK-G (25 nmol/L) are illustrated in Fig 3B⇓. As shown, PK-G decreased basal ICa(L) at each test potential. At test potentials of +50 and +60 mV, a small outward current was produced. The outward currents were always seen after PK-G application, especially at test potentials from +50 mV and higher. Preliminary experiments suggest that this outward current is Cs+-insensitive, 4-AP–insensitive, and tetraethylammonium-insensitive K+ current, which is activated by PK-G, because application of a Cl− channel blocker (SITS, 3 mmol/L; Sigma) or change to low [Cl]o (15 mmol/L) failed to diminish the outward current (data not shown). The outward current persisted even when ICa(L) was blocked by 0.5 mmol/L Cd2+, thus indicating that the outward current was not dependent on Ca2+ influx. To prevent any change in the half-cell potential in the low [Cl]o experiment, an agar bridge was connected so that the Ag/AgCl electrode was immersed in a constant Cl− concentration.
Steady state inactivation curves were plotted as current densities (Fig 4A⇓) and as normalized currents (to the maximum ICa(L) obtained at a conditioning pulse of −80 mV) (Fig 4B⇓). PK-G (25 nmol/L) significantly decreased the current density between −40 and −20 mV (Fig 4A⇓), and the inhibition was not significant at −80 and −60 mV (P<.07). The normalized curves were fitted by the Boltzmann equation (Fig 4B⇓). Calculated values of the half-inactivation potential and slope factor for the control curve were −22 and 5.2 mV, respectively, and those for PK-G were −28 and 6.5 mV, respectively. There was a slight (6-mV) shift in the hyperpolarizing direction after PK-G application (Fig 4B⇓). The relative values at a conditioning pulse of −40 mV were 0.95±0.01 for the control condition and 0.84±0.05 after PK-G application (n=6). This slight shift in steady state inactivation cannot account for the overall inhibitory effect.
Some experiments were performed to examine the reversibility of the PK-G inhibition of basal ICa(L). Once the steady state level of inhibition was attained, PK-G (25 nmol/L) perfusion was stopped by clamping the perfusion tubing (n=12). In most experiments (10 of 12), the inhibitory effect of PK-G was long lasting, and ICa(L) remained inhibited for at least 5 minutes after PK-G perfusion was stopped; however, in two experiments, there was a partial (27% and 33%) recovery of ICa(L) (data not shown).
Mechanism of Inhibition by PK-G on Basal ICa(L)
To check whether the effect of PK-G on basal ICa(L) was due to its enzymatic activity, a stock solution containing only PK-G was incubated in a hot water bath (92°C) for 30 minutes, and this heated PK-G was used for the internal solution (25 nmol/L plus 0.1 μmol/L 8Br-cGMP). As can be seen in Fig 5A⇓, the heat-inactivated PK-G had only a slight inhibitory effect (9.0±2.2%, n=4), which is almost equal to the effect of the low dose of 8Br-cGMP alone (Fig 2⇑). However, subsequent addition of intact PK-G produced a much greater inhibition (35.3±4.9% of control, n=4) of the basal ICa(L) within 5 minutes. Similar results were produced in two other experiments.
To test whether the inhibitory effect of PK-G was due to its enzymatic activity, the effects of three different PK-G inhibitors were examined. Fig 5B⇑ shows a representative experiment in which simultaneous addition of the PK-G inhibitor, cG-PKI (300 μmol/L), prevented the effect of 25 nmol/L PK-G. As can be seen, cG-PKI prevented reduction of ICa(L) by PK-G, and once this inhibitory peptide was removed (its perfusion stopped), the effect of PK-G to inhibit ICa(L) was gradually exerted over the next 3 minutes. Similar results were obtained in a total of three experiments, and they confirmed that the inhibitory effect of PK-G was prevented by cG-PKI. These results suggest that the downregulation of ICa(L) by PK-G is due to its enzymatic activity.
The mechanism of inhibitory effect of PK-G may be explained by the following three possibilities: (1) direct phosphorylation of the channel or an associated protein, (2) activation of a PDE, or (3) activation of a PPase.
Some experiments were conducted to rule out the second possibility by using a nonspecific PDE inhibitor, IBMX. Akita et al18 reported that in rabbit cardiomyocytes at least 100 μmol/L (adult) and 300 μmol/L (newborn) of IBMX were required to exert maximum stimulatory effect on basal ICa(L). Therefore, these two doses of IBMX were tested on rat heart cells. External application of 100 μmol/L IBMX only slightly increased the basal ICa(L) (12.0±1.2%, n=4), and internal application of PK-G (25 nmol/L) markedly inhibited ICa(L) (Fig 6A⇓). Similar results were obtained in three cells. In two of the three experiments, there was a slight recovery after stopping PK-G perfusion (Fig 6A⇓). The higher dose of IBMX (300 μmol/L) had almost no effect on basal ICa(L) (−1.6±6.7%, n=7), but internal perfusion of PK-G (25 nmol/L) inhibited ICa(L) (n=2, Fig 6B⇓). In both experiments, ICa(L) slightly decreased after stopping the perfusion. As a control, the ability of IBMX to inhibit PDE was checked by using ISO. In the presence of IBMX (300 μmol/L), washout of ISO (2 μmol/L) did not reverse the stimulation within 20 minutes (n=3) (Fig 6B⇓, left inset), whereas it rapidly reversed in the absence of IBMX (Fig 6B⇓, right inset). These results indicate that the action of PK-G was not mediated through PDE activation.
A nonselective PPase (1 and 2A) inhibitor, MC (0.5 μmol/L) added in the pipette, was used to exclude the third possibility. MC (0.5 μmol/L) was reported to inhibit PPases in frog atrial cells.22 After the whole-cell configuration was formed, ICa(L) increased, reaching a steady level by 4 to 7 minutes (n=7, data not shown). As can be seen in Fig 7⇓, PK-G (25 nmol/L) was still capable of inhibiting ICa(L) in the presence of MC (n=5 cells). As a control, the ability of MC to inhibit PPases was checked by using ISO. In the presence of MC (0.5 μmol/L), washout of ISO (2 μmol/L) did not reverse the stimulation, and the effect persisted for almost 20 minutes (n=2; Fig 7⇓, inset). These results indicate that the action of PK-G was not mediated through PPase activation.
Variable Inhibitory Effect of PK-G on Basal ICa(L)
The average basal inhibition of ICa(L) by PK-G was 42.1±2.9% (n=36). However, PK-G (25 nmol/L) produced variable degrees of inhibition from cell to cell: large (54.5%, Fig 8A⇓), moderate (34.3%, Fig 8B⇓), or small (13.7%, Fig 8C⇓). The time required to reach a stable-state inhibition was 3 to 6 minutes after the start of PK-G application.
To test whether the variable response to PK-G may be due to possible developmental changes in the intrinsic PK-A and PK-G activities, the inhibition of basal ICa(L) by PK-G (25 nmol/L) was compared for rats in the four neonatal stages of development: 3-day, 7-day, 11-day, and 20-day groups. As shown in Fig 9⇓, there were no statistically significant differences among the four age groups with respect to the average inhibitory action of PK-G: 44.0±5.7% (n=7), 42.8±7.5% (n=7), 39.7±4.2% (n=15), and 44.7±8.7% (n=7), respectively. The variable degrees of inhibition from cell to cell were also seen in all four groups (Fig 9⇓).
Reversibility of the Inhibited ICa(L) by ISO and PK-G Inhibitors
As already described, simply stopping perfusion of PK-G was insufficient to reverse the downregulated ICa(L). Therefore, we checked the reversibility by using ISO and some PK-G inhibitors. After the PK-G (25 μmol/L) inhibition reached a steady state level, ISO (2 μmol/L) was applied from outside (in the 11-day group). The average inhibition of basal ICa(L) was 40.1±7.0% (n=7). As shown in Fig 10⇓, ISO could stimulate the inhibited ICa(L) in seven of the eight cells tested. These results suggest that upregulation through the PK-A pathway can still occur in the presence of PK-G inhibition.
Experiments were also performed in which the effects of PK-G were attempted to be reversed by application of PK-G inhibitors. After PK-G (25 nmol/L) reduced ICa(L) to a steady level, the PK-G inhibitors, cG-PKI (300 μmol/L), H-8 (1-5 μmol/L), and Rp-8-pCPT-cGMPS (2-10 μmol/L), were tested for recovery of ICa(L). H-8 and Rp-8-pCPT-cGMPS were applied externally because they are membrane permeable; cG-PKI was perfused internally. These respective concentrations were used because their reported Ki values (for respective values of PK-A versus PK-G) were 480 and 120 μmol/L for cG-PKI,23 1.2 and 0.48 μmol/L for H-8,24 and 8.3 and 0.5 μmol/L for Rp-8-pCPT-cGMPS (BioLog Co, personal communication, unpublished data, 1994). In most experiments (20 of 25), cG-PKI, H-8, and/or Rp-8-pCPT-cGMPS could not reverse the PK-G–inhibited ICa(L) (Fig 11A⇓ and 11B⇓), but sometimes (5 of 25 experiments) they produced some degree of reversibility. As shown in Fig 11C⇓, one experiment exhibited a full recovery by H-8.
Voltage Dependence of the PK-G Inhibition of Basal ICa(L)
Some experiments were performed to investigate whether the HP used might modulate the inhibitory effect of PK-G on basal ICa(L). To elicit ICa(L), test pulses to +10 mV (300 milliseconds in duration) were applied from a steady HP of −80 mV. Leak currents were subtracted by the P/5 protocol. The Na+-free and K+-free external solution contained 5 μmol/L tetrodotoxin (Calbiochem-Novabiochem Co) and 30 μmol/L Ni2+ to block INa and ICa(T). In these experiments, PK-G had a smaller inhibitory effect on basal ICa(L) when the HP was −80 mV than when it was −40 mV. The inhibition averaged only 8.8±1.4% (n=7, Fig 12⇓) compared with an average value of 42.1±2.9% (n=36) (at an HP of −40 mV).
In the present study, it was found that PK-G inhibited the basal ICa(L) as well as the stimulated ICa(L) in rat ventricular myocytes. The PK-G inhibition was variable in degree from cell to cell; large, moderate, or small degrees of inhibition were seen in all developmental stages. PK-G may directly regulate the L-type Ca2+ channels by phosphorylation of the channel protein or associated regulatory protein. The inhibition of basal ICa(L) was HP dependent: at an HP of −40 mV, the mean inhibition was 42.1% (n=36); at an HP of −80 mV, the mean inhibition was only 8.8% (n=7). Similar results have been obtained from chick embryonic heart cells, showing that intracellular application of PK-G via the patch pipette inhibited the basal and the stimulated ICa(L).16
To check whether the PK-G effect was due to its enzymatic activity, two types of experiments were performed. When a PK-G substrate inhibitory peptide (cG-PKI) was applied with PK-G, the inhibitory effect on ICa(L) was blocked (Fig 5B⇑). Heat-inactivated PK-G had little inhibitory effect on ICa(L) (mean, 9.0±2.2%) (Fig 5A⇑), thus indicating that the effect exerted by PK-G resulted from its enzymatic activity. This small inhibitory effect of heat-inactivated PK-G may have resulted from the action of 8Br-cGMP (0.1 μmol/L) in the test solution (mean, 10.5±1.9%; see Fig 2⇑).
We negated the possibility that the inhibitory action of PK-G was mediated through activation of either PDE or PPase. The possibility of involvement of the PDE was ruled out by demonstrating that the inhibitory action of PK-G was not different in the presence of a very high concentration of IBMX. That is, PK-G was still capable of depressing ICa(L) when the PDE was strongly inhibited by IBMX. The possible involvement of PPase was ruled out by demonstrating that the inhibitory action of PK-G was not different in the presence of the PPase inhibitor MC. That is, PK-G still had the ability to inhibit ICa(L) when the PPase was inhibited by MC. Therefore, the inhibitory mechanism may result from PK-G phosphorylation of the Ca2+ channel or an associated protein.
Other findings also suggest that PK-G acts via a phosphorylation-dependent mechanism. Méry et al9 reported that intracellular application of an active fragment of PK-G (prepared by action of a proteolytic enzyme) inhibited stimulated ICa(L). 8Br-cGMP inhibition of the Ca2+-dependent slow action potentials in guinea pig cardiac muscle was shown to occur without a decrease in cAMP level.6 Cuppoletti et al10 have reported that a single protein of ≈47 kD is specifically phosphorylated by PK-G in guinea pig sarcolemmal preparations.10 This protein may be a possible mediator involved in regulation of Ca2+ channels of the heart by the cGMP/PK-G pathway.
Recently, Haase et al25 reported that PK-A–mediated phosphorylation of the β-subunit of the cardiac L-type Ca2+ channel is the major mechanism for β-adrenergic regulation of the channel activity. In skeletal muscle, however, the α1-subunit is believed to be the relevant channel component for β-adrenergic regulation.26 27 Because the action of PK-G phosphorylation is antagonistic to that of PK-A phosphorylation, the target phosphorylation site for PK-G must be a different site from the site that is phosphorylated by PK-A. Such phosphorylations by PK-A and PK-G of the Ca2+ channel or an associated regulatory protein may regulate the channel activity by exerting an allosteric effect.
In the present study, 8Br-cGMP (0.1 μmol/L) alone failed to exert a significant inhibitory effect (by activating endogenous PK-G), whereas exogenous fully activated PK-G did (Fig 2A⇑). We used a low subthreshold dose (0.1 μmol/L) of 8Br-cGMP as a PK-G activator. 8Br-cGMP gives strong activation of PK-G28 but has only a weak action on PDE and is poorly hydrolyzed (ie, long lasting). Since the concentration used (0.1 μmol/L) was reported to give maximum activation of PK-G in vitro,28 the PK-G applied via the patch pipette should be maximally active. However, internal application of 8Br-cGMP (0.1 μmol/L) was almost subthreshold, whereas higher doses (1 or 10 μmol/L) significantly inhibited the basal ICa(L) (Fig 2B⇑). These results suggest that the dose (0.1 μmol/L) used was insufficient to activate endogenous PK-G. Méry et al9 showed that 10 μmol/L cGMP or 8Br-cGMP was needed (internal perfusion) to give a large inhibition (eg, 71%) of prestimulated ICa(L). Kameyama et al3 reported that the concentration of cAMP for half-maximal activation (internal perfusion) was 3 to 50 times higher than that reported for in vitro studies. Hence, there is some difference between the biochemical data and the physiological data with respect to the concentration of cGMP required to exert maximum effect. In summary, 0.1 μmol/L 8Br-cGMP was sufficient to maximally activate the exogenous PK-G in the pipette but activated the endogenous PK-G to only a small extent.
The variable effects of PK-G application on the basal ICa(L) may be explained by the working hypothesis that the level of endogenous PK-G activity may vary from cell to cell and that the basal ICa(L) is regulated by a balance of activity between PK-A and PK-G. If the endogenous PK-G activity is low, then application of exogenous activated PK-G (via the patch pipette) or addition of 8Br-cGMP into the bath should inhibit ICa(L). On the other hand, if the endogenous PK-G activity is already high, then introduction of exogenous activated PK-G should have little or no effect. This hypothesis can also explain why ISO-stimulated ICa(L) is inhibited by application of PK-G. That is, when the PK-A pathway is stimulated by ISO, the endogenous PK-G activity then becomes subordinate. Therefore, exogenous PK-G should always inhibit the stimulated ICa(L). These results are consistent with our hypothesis that PK-A and PK-G phosphorylate different target sites and that they exert antagonistic actions.
Once the applied PK-G had phosphorylated its target site, it was difficult to obtain full reversibility, even when relatively specific PK-G inhibitors were added. One possible reason for this may include the fact that the 8Br-cGMP, used as an unhydrolyzable and strong PK-G activator, causes prolonged activation of the PK-G, which results in long-lasting phosphorylation of the inhibitory regulatory subunit of the L-type Ca2+ channels. The variable effect of PK-G inhibitors on reversing the inhibition of ICa(L) can also be explained by differences in endogenous PPase activity from one cell to another. If the PPase activity is low, which may be the usual case, then the site phosphorylated by PK-G remains phosphorylated for a long time; hence, the addition of PK-G inhibitor should have little or no effect. If the PPase activity is high, then the site remains phosphorylated for only a short time and so must be rephosphorylated to remain active; therefore, the application of PK-G inhibitor should have a strong effect.
We showed that the inhibitory effect of PK-G is dependent on the HP used. So far, the reason for this is unknown. In accordance with the hypothesis described above, the endogenous PK-G activity may be generally low when the HP is −40 mV; if so, application of exogenous PK-G should exert a prominent inhibitory effect. On the contrary, when the HP is set at −80 mV, the endogenous PK-G activity may be generally high; if so, application of exogenous PK-G should give only a small inhibitory effect. HP-dependent regulation of the Ca2+ channel by PK-C activation is also reported in cultured neonatal rat ventricular cells.29 Using an HP of −80 mV, Méry et al9 reported that intracellular application of an active proteolytic fragment of PK-G gave only 5.0±3.2% inhibition of basal ICa(L) (n=4) in adult rat ventricular myocytes. The degree of inhibition is not statistically different from our finding of 8.8±1.4% inhibition (n=7) at the same HP of −80 mV. Thus, in this respect, the voltage-dependent inhibition by PK-G is similar to the inhibition produced by dihydropyridine30 and tetrodotoxin.31
Some hormones, such as ACh and ANF, are known to increase the cGMP level in mammalian cardiomyocytes.32 33 34 ACh depresses the basal ICa(L) in guinea pig atrial cells35 36 and rabbit sinoatrial nodal cells.37 38 ANF was reported to inhibit the basal ICa(L) in rat and guinea pig ventricular myocytes39 and human atrial cells.40 One possible mechanism for inhibition of the basal ICa(L) by ACh and ANF is cGMP production and the resultant PK-G phosphorylation of the Ca2+ channels or an associated protein.
In summary, in young rat ventricular myocytes, PK-G inhibited the basal ICa(L) to variable degrees (large, medium, or small), whereas PK-G always inhibited the prestimulated ICa(L) substantially back to approximately the basal level or beyond. The variable degrees of inhibition of basal ICa(L) produced by PK-G may be caused by different levels of activity of the endogenous PK-G. There were no differences in the average degree of inhibition of the basal ICa(L) during development. The inhibition of the basal ICa(L) was dependent on HP; ie, 42.1% inhibition occurred when the HP was −40 mV, whereas only 8.8% inhibition occurred when the HP was −80 mV. The inhibitory effects of PK-G were not mediated by activating PDE or PPase but most likely by a direct phosphorylation of the Ca2+ channel or an associated regulatory protein. The inhibitory effect of PK-G may be explained by a balance between activities of PK-A and PK-G in regulating the slow Ca2+ channels at two separate sites.
Selected Abbreviations and Acronyms
|ANF||=||atrial natriuretic factor|
|cG-PKI||=||PK-G peptide substrate|
|ICa(L)||=||whole-cell L-type Ca2+ current|
|ICa(T)||=||T-type Ca2+ current|
|PK-A||=||cAMP-dependent protein kinase|
|PK-G||=||cGMP-dependent protein kinase|
|Rp-8-pCPT-cGMPS||=||8-(4-chlorophenylthio)-guanosine-3′5′-cyclic monophosphorothioate, Rp isomer|
This study was supported by National Institutes of Health grant HL-31942. We thank Prof Issei Seyama, Hiroshima University School of Medicine, for providing Dr Sumii with the opportunity to conduct joint research at the University of Cincinnati. We thank Lisa Neumeier and Sheila Blanck for preparation of the cells.
- Received November 21, 1994.
- Accepted June 15, 1995.
- © 1995 American Heart Association, Inc.
Wahler GM, Rusch NJ, Sperelakis N. 8-bromo-cyclic GMP inhibits the calcium channel current in embryonic chick ventricular myocytes. Can J Physiol Pharmacol. 1989;68:531-534.
Méry PF, Lohmann SM, Walter U, Fischmeister R. Ca2+ current is regulated by cyclic GMP-dependent protein kinase in mammalian cardiac myocytes. Proc Natl Acad Sci U S A. 1991;88:1197-1201.
Cuppoletti J, Thakkar J, Sperelakis N, Wahler G. Cardiac sarcolemmal substrate of the cGMP-dependent protein kinase. Membr Biochem. 1989;7:135-142.
Bkaily G, Sperelakis N. Injection of guanosine 5′-cyclic monophosphate into heart cells blocks calcium slow channels. Am J Physiol. 1985;248:H745-H749.
Tohse N, Sperelakis N. cGMP inhibits the activity of single calcium channels in embryonic chick heart cells. Circ Res. 1991;69:325-331.
Akita T, Joyner RW, Lu C, Kumar R, Hartzell HC. Developmental changes in modulation of calcium currents of rabbit ventricular cells by phosphodiesterase inhibitors. Circulation. 1994;90:469-478.
Masuda H, Sumii K, Sperelakis N. Developmental changes in β-adrenergic and muscarinic modulation of Ca2+ currents in fetal and neonatal ventricular cardiomyocytes of rat. Reprod Fertil Dev. In press.
Kokubun S, Reuter H. Dihydropyridine derivatives prolong the open state of Ca channels in cultured cardiac cells. Proc Natl Acad Sci U S A. 1984;81:4824-4827.
Colbran JL, Francis SH, Leach AB, Thomas MK, Jiang H, McAllister LM, Corbin JD. A phenylalanine in peptide substrates provides for selectivity between cGMP- and cAMP-dependent protein kinases. J Biol Chem. 1992;267:9589-9594.
Mundina-Weilenmann C, Chang CF, Gutierrez LM, Hosey MM. Demonstration of the phosphorylation of dihydropyridine-sensitive calcium channels in chick skeletal muscle and the resultant activation of the channels after reconstitution. J Biol Chem. 1991;266:4067-4073.
Chang CF, Gutierrez LM, Mundina-Weilenmann C, Hosey MM. Dihydropyridine-sensitive calcium channels from skeletal muscle. J Biol Chem. 1991;266:16395-16400.
Lincoln TM, Corbin JD. Characterization and biological role of the cGMP-dependent protein kinase. Adv Cyclic Nucleotide Res. 1983;15:140-192.
Kass RS. Ionic basis of electrical activity in the heart. In: Sperelakis N, ed. Physiology and Pathophysiology of the Heart. 3rd ed. Boston, Mass: Kluwer Academic Publishers; 1995:77-90.
Cohen CJ, Bean BP, Colatsky TJ, Tsien RW. Tetrodotoxin block of sodium channels in rabbit Purkinje fibers. J Gen Physiol. 1981;78:383-411.
George WJ, Polson JB, O’Toole AG, Goldberg ND. Elevation of guanosine 3′,5′-cyclic phosphate in rat heart after perfusion with acetylcholine. Proc Natl Acad Sci U S A. 1970;66:398-403.
Cerbai E, Klöckner U, Isenberg G. Ca-antagonistic effects of adenosine in guinea pig atrial cells. Am J Physiol. 1988;255:H872-H878.
Sorbera LA, Morad M. Atrionatriuretic peptide transforms cardiac sodium channels into calcium-conducting channels. Science. 1990;247:969-973.