Introduction

Top Abstract Introduction Materials and Methods Results Discussion References

Extrinsic regulation of heart function usually involves opposing actions by the autonomic nervous system. The yin-yang hypothesis placed the postjunctional actions in the context of intracellular messengers, cAMP and cGMP, and their respective protein kinases having opposite effects on cellular processes (Watanabe and Besch, 1975). One target, the L-type calcium channel, has received considerable attention because the cAMP-activated protein kinase (PKA) increases current (I_Ca(L)) through the channel whereas the cGMP-activated kinase (PKG) decreases it (reviewed in McDonald et al., 1994).

In mammalian ventricle, cGMP-inhibited phosphodiesterase (PDE) and PKG appear to be the principal effectors of cGMP with the former causing I_Ca(L) to increase as it allows cAMP to accumulate (Ono and Trautwein, 1991; Shirayama and Pappano, 1996). Activation of PKG, in the presence of elevated cAMP, inhibits the L-type calcium current but the proposed mechanisms differ. In rat ventricular myocytes, inhibition of I_Ca(L) by cGMP was ascribed to phosphorylation by PKG of the L-type calcium channel or an associated regulatory protein (Méry et al., 1991; Sumii and Sperelakis, 1995). When the _1c subunit of the cardiac calcium channel is expressed in oocytes, 8-bromo-cGMP (8-Br-cGMP) inhibited basal current (Jiang et al., 2000). The PKG inhibitor KT5823 or replacement of serine by alanine at a consensus site (S533A) for PKG-dependent phosphorylation prevented inhibition by cGMP. In the _1c component of the L-type channel, PKA phosphorylates a serine residue 1928 (Gao et al., 1997). Thus, antagonistic modulation of I_Ca(L) by cyclic nucleotides could be referable to phosphorylation of different regulatory sites on the _1c subunit.

Activation of a phosphatase was excluded as a mechanism for inhibition by cGMP/PKG (Méry et al., 1991; Sumii and Sperelakis, 1995). This conclusion is at odds with the observation that acetylcholine (ACh), which increases cGMP, opposed PKA-dependent phosphorylation of proteins in guinea pig ventricular myocytes (Gupta et al., 1994). This was attributed to activation of protein phosphatase (PP). Okadaic acid, an inhibitor of PP, prevented dephosphorylation of proteins (Gupta et al., 1994) and inhibition of isoproterenol (ISO)-stimulated I_Ca(L) by ACh (Herzig et al., 1995). The phosphatase hypothesis was strengthened by evidence that inhibition of I_Ca(L) by carbachol (CCh) or 8-Br-cGMP was antagonized by dialysis with 300 nM okadaic acid (Sakai et al., 1999). As such, the results resemble those seen in GH₄C₁ cells where atrial natriuretic peptide increased a maxi-K channel current by activating a phosphatase through a cGMP/PKG pathway (White et al., 1993). Okadaic acid prevented maxi-K current stimulation by atrial natriuretic peptide. When Ba²⁺ current through _1c subunits expressed in human embryonic kidney cells is elevated by forskolin-induced activation of adenylyl cyclase, phosphatase inhibition by okadaic acid potentiates this current and the phosphorylation of serine 1928 (Gao et al., 1997).

We evaluated the phosphatase activation hypothesis for cGMP action on I_Ca(L). Our experiments were done with ATPS replacing ATP in the pipette solution [ATPS]_pip; 8-Br-cGMP was used to activate PKG. ATPS is a kinase substrate; the thiophosphorylated substrate resists phosphatase action (Yount, 1975). With 3 mM [ATPS]_pip, ISO increased I_Ca(L) more than with ATP and the stimulant effect persisted when ISO was removed (Kameyama et al., 1985; Sipido et al., 1995). PKA and PKG phosphorylate the L-type calcium channel at different sites. With ATPS present, we assume that if the effect of thiophosphorylation by PKA on I_Ca(L) cannot be reversed by activation of PKG, the antagonistic interaction in ATP occurs by dephosphorylation of the site phosphorylated by cAMP/PKA. On the other hand, if cGMP can oppose the effect of cAMP on I_Ca(L) in the presence of ATPS, the antagonism most likely occurs by thiophosphorylation of a different regulatory site on the _1c subunit.

	Materials and Methods

Top Abstract Introduction Materials and Methods Results Discussion References

Ventricular Myocyte Isolation. Single ventricular myocytes were isolated from the hearts of guinea pigs (250-450 g) anesthetized with sodium pentobarbital (30 mg/kg, intraperitoneally) and anticoagulated with heparin (1000 IU, i.p.). The heart was perfused according to the Langendorff technique with modified Tyrode's solution containing 135 mM NaCl , 5.4 mM KCl, 1.8 mM CaCl₂, 1.0 mM MgCl₂, 0.33 mM NaH₂PO₄, 10 mM HEPES, and 20 mM glucose (pH was 7.4 with NaOH). The procedures are the same as we reported previously (Imai et al., 2001). After collagenase and protease disrupted the extracellular matrix, the enzymes were washed out by perfusion with 50 ml of recovery solution containing 130 mM potassium aspartate, 5 mM K₂ATP, 5 mM HEPES, and 20 mM glucose; pH was adjusted to 7.4 with KOH. The ventricles were removed, and the cells were dispersed in recovery solution and kept at 4°C for at least 1 h. A cell suspension droplet was placed in a chamber (500-µl volume) on the stage of an inverted microscope. After 10 min, superfusion began with Tyrode's solution (2 ml/min) containing 10 mM glucose and 10 mM CsCl (35°C).

Electrophysiology. The patch rupture-whole cell voltage clamp technique used an EPC 7 patch-clamp amplifier (List Electronics, Darmstadt, Germany). Voltage commands and data acquisition were obtained with pClamp software (version 5.5; Axon Instruments, Union City, CA) and a Labmaster TL-1 interface (Axon Instruments). The pipette filling solution for glass electrodes (i.d.,1.1 mm; o.d.,1.3 mm) contained 135 to 140 mM cesium aspartate, 7 mM NaCl, 3.0 mM MgCl₂, 3.5 or 5 mM ATPS, 10 mM EGTA, and 10 mM HEPES; pH 7.3 (with CsOH). In some experiments, the pipette solution also contained 1 mM GTP. The electrode resistance was 1 to 3 M. Series resistance could be compensated up to 70% to values between 1 and 2.5 M.

Drugs and Application. Drugs were applied to myocytes by superfusion by gravity from a reservoir. The applied solutions were warmed to provide an experimental temperature of 35°C. CCh, ISO, 8-Br-cGMP, and 3-isobutyl-1-methylxanthine (IBMX) were prepared fresh daily from aqueous stock solutions.

Data Analysis. With Cs⁺-rich pipette solution to block membrane K⁺ currents, I_Ca(L) is taken as peak current minus end-of-pulse current at +10 mV. Measurements are reported as mean ± S.E.M. The statistical significance of mean differences was determined by Student's t test. p  0.05 was considered statistically significant.

	Results

Top Abstract Introduction Materials and Methods Results Discussion References

8-Br-cGMP inhibits I_Ca(L) augmented by ISO or IBMX. At 0.1 mM, 8-Br-cGMP inhibited I_Ca(L) in 0.1 mM IBMX by 34 ± 3.6% and in 10 nM ISO by 32 ± 7.1% (Imai et al., 2001). Accordingly, one would expect 0.1 mM 8-Br-cGMP to suppress cAMP-stimulated I_Ca(L) by about one-third. In the following experiments, a 10-fold greater concentration of 1 mM 8-Br-cGMP was used.

The results of a test with 8-Br-cGMP on IBMX-stimulated I_Ca(L) are shown in Fig. 1. The L-channel current was 0.7 nA upon patch rupture and gradually increased to 1.4 nA after 25 min, during which time the cell had been dialyzed with [ATPS]_pip of 5 mM. In this experiment, the pipette solution also contained 1 mM GTP. Addition of 0.1 mM IBMX to the superfusion fluid markedly increased I_Ca(L) to 4.6 nA at 30 min. When 1 mM 8-Br-cGMP was present for 5 min with IBMX, I_Ca(L) remained at 4.6 nA. Washout of 8-Br-cGMP ensued and I_Ca(L) remained elevated at 4.6 nA up to 40 min. This result is characteristic of that observed in a total of 21 cells in which 8-Br-cGMP was tested on cAMP-elevated I_Ca(L). In three of these cells, 10 nM ISO was used to stimulate I_Ca(L), and the results are included with those obtained with IBMX because the outcome was the same. Overall, the initial I_Ca(L) just before addition of cAMP-elevating agent was 1.0 ± 0.13 nA and it increased to 2.6 ± 0.27 nA (n = 21). On average, 1 mM 8-Br-cGMP changed cAMP-elevated I_Ca(L) by 3 ± 11.8% (n = 21), which is not significantly different from 0. 8-Br-cGMP reduced I_Ca(L) by 16 ± 5.6% (n = 8) in cells dialyzed with ATPS for <25 min, and by 5 ± 18.7% (n = 13) in cells dialyzed for >25 min (p = 0.2). Importantly, if 8-Br-cGMP reduced IBMX-stimulated I_Ca(L), the effect was transient.

View larger version (18K): [in this window] [in a new window]   Fig. 1.   8-Br-cGMP fails to inhibit IBMX-stimulated ICa(L) in a ventricular myocyte dialyzed with 5 mM ATPS. The pipette solution also had 1 mM GTP. Ordinate, ICa(L) in nanoamperes; abscissa, time in minutes. ICa(L) was 0.7 nA at patch rupture and slowly increased to ~1.4 nA at 25 min. IBMX (0.1 mM), added at 25 min, caused ICa(L) to increase rapidly to 4.6 nA at 30 min. Subsequent addition of 1 mM 8-Br-cGMP negligibly affected IBMX-stimulated ICa(L). Horizontal lines indicate the time and duration of drug additions. Individual current traces recorded at A to E are shown above with current calibration in the right margin.

There were two cells in which we did not observe this pattern; an example is shown in Fig. 2. The L-channel current increased from 1.4 to 2.8 nA during the first 10 min of dialysis; the pipette solution was the same as for the cell in Fig. 1. Addition of 0.1 mM IBMX at 10 min had a small stimulant effect on I_Ca(L) that eventually reached 3 nA at 16 min. When 1 mM 8-Br-cGMP was added, I_Ca(L) diminished rapidly over the next 4 min to 0.4 nA. The current did not change appreciably over the ensuing 15 min when 8-Br-cGMP and then IBMX were washed out. This result resembles that reported by Méry et al. (1991) in which cGMP irreversibly reduced cAMP-stimulated I_Ca(L). However, it is exceptional because it occurred in only 2 of the 23 cells tested and cannot be distinguished from rundown. In both instances, IBMX had been added at about 10 min after patch rupture and 8-Br-cGMP tested 5 min later. The time of IBMX addition ranged from 6 to 28 min with an average of 16 ± 2 min in the 21 cells where the effect of 8-Br-cGMP was suppressed.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   An example of an exceptional result when IBMX-stimulated ICa(L) is "irreversibly" reduced by 8-Br-cGMP. Ordinates and abscissa are as in Fig. 1. L-channel current increased from 1.4 nA at patch rupture to 2.7 nA at 10 min. IBMX (0.1 mM) increased ICa(L) to 3 nA by 16 min, and 8-Br-cGMP (1 mM) promptly reduced current to 0.4 nA, which is below the level seen at patch rupture. Washout of 8-Br-cGMP at 21 min with IBMX still present did not cause any further change of ICa(L). See text for details.

Inhibition of ISO- or IBMX-stimulated I_Ca(L) by CCh has been reported by many investigators. We found that 0.1 mM CCh inhibited ISO-stimulated I_Ca(L) by 77 ± 7.9% and IBMX-stimulated I_Ca(L) by 79 ± 6.3% (Imai et al., 2001). However, in cells dialyzed with ATPS, CCh had little effect when tested against IBMX-stimulated I_Ca(L). An example is shown in Fig. 3. In this cell, I_Ca(L) was 0.2 nA upon patch rupture and increased to 0.9 nA during the next 20 min as the cell was dialyzed with [ATPS]_pip of 5 mM and [GTP]_pip of 1 mM. Again, 0.1 mM IBMX, added at 20 min, promptly increased I_Ca(L) to 5 nA (Fig. 3). Current did not change upon addition of 0.1 mM CCh to the bath fluid for 5 min. When CCh was removed from the IBMX-containing solution, I_Ca(L) remained at 5 nA through 40 min.

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Dialysis with [ATPS]pip of 5 mM prevents inhibition by CCh of IBMX-stimulated ICa(L). The pipette solution also contained 1 mM GTP. Ordinates and abscissa as in Fig. 1. At patch rupture, ICa(L) was 0.2 nA, and it rose to 0.9 nA after 20 min of dialysis. Addition of 0.1 mM IBMX at 20 min caused ICa(L) to increase rapidly and to reach 4.8 nA at 25 min. Carbachol (0.1 mM) added on top of IBMX during min 26 to 32 did not change ICa(L), which remained elevated after CCh washout.

Overall, 0.1 mM CCh reduced cAMP-elevated I_Ca(L) by 32 ± 9.9% (n = 12). Although this is substantially less than that observed with ATP in the pipette solution (see above), it is greater than the inhibition observed with 8-Br-cGMP in separate experiments. However, when both were tested in the same six cells against IBMX-stimulated I_Ca(L), CCh reduced current by 29 ± 13.3% (32 ± 5.2 min) and 8-Br-cGMP reduced it by 21 ± 9.7% (26 ± 7.1 min). In nine cells, we obtained stable recordings during dialysis with [ATPS]_pip of 3.5 to 5 mM for 45 min. Under these conditions, tests with 1 mM 8-Br-cGMP (n = 5) or with 0.1 mM CCh (n = 7) revealed no inhibition of stimulated I_Ca(L) by either ligand (data not shown).

	Discussion

Top Abstract Introduction Materials and Methods Results Discussion References

With PDEs inactivated by IBMX, we assume that there are two pathways for inhibition of I_Ca(L) by cGMP/PKG, namely, phosphorylation of the _1c subunit that antagonizes the stimulant effect of cAMP/PKA and/or activation of PP that dephosphorylates the _1c subunit. Can 8-Br-cGMP or CCh suppress Ca²⁺ current through L-type channels in the presence of ATPS, which allows kinases to thiophosphorylate substrates that resist phosphatase action? Irreversible suppression by 8-Br-cGMP would indicate that thiophosphorylation of the channel had occurred at a site different from that acted upon by cAMP/PKA (Méry et al., 1991; Sumii and Sperelakis, 1995). Alternatively, failure of 8-Br-cGMP to suppress I_Ca(L) in ATPS would indicate that the thiophosphorylated channel was resistant to phosphatase. Our results are consistent with the PP hypothesis for inhibition by cGMP of current through L-type channels in guinea pig ventricular myocytes.

Neurotransmitter regulation of I_Ca(L) by reversible phosphorylation-dephosphorylation reactions is well known (McDonald et al., 1994; Herzig and Neumann, 2000). Serine 1928 of the _1c subunit has been reported to be an essential residue for phosphorylation by PKA (Gao et al., 1997). Our results point to the possibility that the site for phosphatase action on the _1c subunit is the same. This conclusion differs from that of others in several ways. Those who report that cGMP irreversibly inhibits I_Ca(L) by phosphorylation of the channel did experiments in rat ventricular myocytes (Méry et al., 1991; Sumii and Sperelakis, 1995). A similar conclusion emerged from experiments with mouse ventricular myocytes (Klein et al., 2000). We did not observe suppression of basal I_Ca(L) by 8-Br-cGMP unlike the results from rat ventricular myocytes (Sumii and Sperelakis, 1995). Our experiments and those reporting cGMP/PKG-induced activation of PP are from guinea pig ventricular myocytes (reviewed in Herzig and Neumann, 2000). Different species have distinct regulatory mechanisms for cGMP-dependent regulation of I_Ca(L). In amphibian ventricular myocytes, cGMP stimulated PDE, which reduces cAMP concentration (reviewed in Méry et al., 1997). This pathway is not prominent in mammalian ventricular myocytes and would be inoperative in our experiments because IBMX inhibits all PDEs. In human atrial trabeculae, 8-Br-cGMP had a negative inotropic effect (Nawrath et al., 1995; Flesch et al., 1997) but increased I_Ca(L) in single myocytes (Vandecasteele et al., 2001). Stimulation of I_Ca(L) by 8-Br-cGMP was attributed to activation of PKA. Although PKG was excluded as an effector of 8-Br-cGMP in single atrial myocytes (Vandecasteele et al., 2001), it may have been the target in experiments with human atrial trabecular contractions (Nawrath et al., 1995; Flesch et al., 1997). No information is available on effectors of 8-Br-cGMP in human ventricle, which conceivably differ from those in atrium (Vandecasteele et al., 2001).

8-Br-cGMP could sometimes inhibit I_Ca(L) in cells dialyzed for 25 min with ATPS. However, the inhibition was reversible, a finding inconsistent with cGMP/PKG causing irreversible inhibition via thiophosphorylation of the L-type calcium channel in rat ventricular myocytes (Méry et al., 1991). In only 2 of 23 cells, 8-Br-cGMP addition was accompanied by a sustained decrease of I_Ca(L) (Fig. 2). Although this could be an example of irreversible inhibition, it cannot be distinguished from current rundown. Sufficient time should be allowed for diffusion of ATPS from the pipette solution and for replacing ATP for participation in thiophosphorylation reactions including those at the L-type calcium channel.

That ATPS interfered with inhibition of I_Ca(L) by 8-Br-cGMP favors the participation of PKG-activated PP in the cGMP signaling pathway. Dialysis with either PP1 or PP2A opposed ISO-stimulated I_Ca(L), but not basal I_Ca(L), in guinea pig ventricular myocytes (Hescheler et al., 1987). Okadaic acid inhibits PP1 and PP2A with the latter being more sensitive (Herzig and Neumann, 2000). Okadaic acid (300 nM in pipette solution) prevented inhibition of I_Ca(L) by 8-Br-cGMP (Sakai et al., 1999). That okadaic acid inhibits membrane bound PP1 and PP2A activities in guinea pig ventricular myocytes (Neumann et al., 1993) indicates that neither enzyme can be excluded from participating in I_Ca(L) inhibition. Single channel recordings indicated that PP2A and PP1 dephosphorylate the L-type Ca²⁺ channel at different sites (Ono and Fozzard, 1993; Wiechen et al., 1995). Together with the present results, the data with okadaic acid favor activation of PP as a mechanism for cGMP/PKG signaling (see also White et al., 1993).

ATPS Prevents Inhibition by Carbachol of I_Ca(L). The ability of CCh to inhibit I_Ca(L) diminished in cells dialyzed with ATPS. This is in accordance with the evidence that muscarinic agonist inhibits I_Ca(L) by activating okadaic acid-sensitive PP (Herzig et al., 1995) and that cGMP is the signal molecule for this action (Imai et al., 2001), presumably by phosphorylating PP (Sakai et al., 1999). Similarly, ACh opposed the action of ISO or dibutyryl-cAMP on the inwardly rectifying K⁺ current by activation of okadaic acid-sensitive PP in guinea pig ventricular myocytes (Koumi et al., 1995).

Limitations. An alternative possibility is that cGMP/PKG not only activates PP but also phosphorylates the _1c subunit in guinea pig ventricular myocytes. That is, with ATP present, both reactions occur and contribute to suppression of cAMP/PKA-activated I_Ca(L). What happens in the presence of ATPS? With PP activity blocked, the most likely mechanism for inhibition is cGMP/PKG-induced thiophosphorylation of serine 533 on the _1c subunit. If this were so, one would have to conclude that these regulatory sites are not equal and that serine 1928 dominates because when thiophosphorylated by cAMP/PKA, thiophosphorylation of serine 533 by cGMP/PKG is unable to cause inhibition.