Intracellular cyclic AMP inhibits native and recombinant volume-regulated chloride channels from mammalian heart

Abstract

1. ClC-3 encodes a volume-regulated Cl⁻ channel (I_Cl,vol) in heart. We studied the regulation of native and recombinant cardiac I_Cl,vol by intracellular cyclic AMP (cAMP_i).

2. Symmetrical high Cl⁻ concentrations were used to effectively separate outwardly rectifying I_Cl,vol from other non-rectifying Cl⁻ currents, such as the cystic fibrosis transmembrane conductance regulator (CFTR) and Ca²⁺-activated Cl⁻ currents (I_Cl,CFTR and I_Cl,Ca, respectively), which are concomitantly expressed in cardiac myocytes.

3. 8-Bromo-cyclic AMP (8-Br-cAMP) significantly inhibited I_Cl,vol in most guinea-pig atrial myocytes. In ≈30 % of the atrial myocytes examined, 8-Br-cAMP increased macroscopic Cl⁻ currents. However, the 8-Br-cAMP-stimulated difference currents exhibited a linear current-voltage (I–V) relation, consistent with activation of I_Cl,CFTR, not I_Cl,vol.

4. In canine atrial myocytes, isoprenaline (1 μM) consistently reduced I_Cl,vol in Ca²⁺-free hypotonic bath solutions with strong intracellular Ca²⁺ (${\text{Ca}}_{\text{i}}^{2+}$) buffering. In Ca²⁺-containing hypotonic bath solutions with weak ${\text{Ca}}_{\text{i}}^{2+}$ buffering, however, isoprenaline increased net macroscopic Cl⁻ currents. Isoprenaline-stimulated difference currents were not outwardly rectifying, consistent with activation of I_Cl,Ca, not I_Cl,vol.

5. In NIH/3T3 cells transfected with gpClC-3 (the gene encoding I_Cl,vol), 8-Br-cAMP consistently inhibited I_Cl,ClC-3. These effects were prevented by a protein kinase A (PKA) inhibitor, KT5720, or by mutation of a single consensus protein kinase C (PKC) phosphorylation site (S51A) on the N-terminus of ClC-3, which also mediates PKC inhibition of I_Cl,ClC-3.

6. We conclude that cAMP_i causes inhibition of I_Cl,vol in mammalian heart due to cross-phosphorylation of the same PKC consensus site on ClC-3 by PKA. Our results suggest that contamination of macroscopic I_Cl,vol by I_Cl,CFTR and/or I_Cl,Ca may account for some of the inconsistent and controversial effects of cAMP_i on I_Cl,vol previously reported in native cardiac myocytes.

A volume-regulated chloride current (I_Cl,vol) is widely expressed in cardiac and many other tissues of a variety of species (Sorota, 1992; Hagiwara et al. 1992; Tseng, 1992; Vandenberg et al. 1994; Oz & Sorota, 1995; Hall et al. 1995; Duan et al. 1995; Strange et al. 1996; Vandenberg et al. 1996; Okada, 1997; Duan et al. 1997a). Several unique properties of I_Cl,vol distinguish these channels from other types of Cl⁻ channels: outward rectification with symmetrical high Cl⁻ concentrations, time-dependent inactivation at positive potentials, higher permeability to I⁻ than Cl⁻, and an intermediate single-channel conductance of 28–60 pS (Strange et al. 1996; Vandenberg et al. 1996; Okada, 1997). Recent evidence suggests that ClC-3, a member of the ClC Cl⁻ channel family, may be responsible for native I_Cl,vol in many mammalian cells (Kawasaki et al. 1994; Duan et al. 1997b, 1999; Yamazaki et al. 1998; von Weikersthal et al. 1999). In heart, I_Cl,vol plays an important role in the regulation of cell volume homeostasis and electrical activity (Vandenberg et al. 1996; Hiraoka et al. 1998; Hume et al. 2000). I_Cl,vol function is known to be regulated by phosphorylation and dephosphorylation (Hall et al. 1995; Okada, 1997) and we recently described the role of protein kinase C (PKC) phosphorylation in the regulation of native I_Cl,vol in heart and expressed ClC-3 channels by changes in cell volume (Duan et al. 1999). The present study was designed to examine the regulation of native I_Cl,vol and expressed ClC-3 channels by intracellular cyclic AMP (cAMP_i).

In many cardiac cells, elevation of cAMP_i is known to activate a distinct class of Cl⁻ channels through protein kinase A (PKA) phosphorylation (I_Cl,PKA) (Harvey & Hume, 1989; Bahinski et al. 1989; Gadsby et al. 1995; Hume et al. 2000). In contrast to I_Cl,vol, I_Cl,PKA exhibits a linear current-voltage relation in symmetrical high Cl⁻, and is generated by small (8-12 pS) unitary conductance channels encoded by the CFTR gene (Gadsby et al. 1995; Hart et al. 1996). Whether or not cardiac I_Cl,vol can also be regulated by cAMP_i and cAMP-dependent PKA phosphorylation is controversial (Sorota, 1992; Hagiwara et al. 1992; Tseng, 1992; Oz & Sorota, 1995; Hall et al. 1995; Du & Sorota, 1997). In canine ventricular myocytes, I_Cl,vol has been reported to be insensitive to the PKA inhibitor H-89 (Tseng, 1992), and the stretch-induced Cl⁻ current in rabbit heart has also been reported to be insensitive to a specific peptide inhibitor of PKA (Hagiwara et al. 1992). In contrast, in cultured chick cardiac myocytes, I_Cl,vol was inhibited by forskolin, and activation of I_Cl,vol during cell swelling was prevented when the pipette filling solution contained cAMP and isobutylmethylxanthine in the presence of okadaic acid (Hall et al. 1995). An increase in cAMP_i levels by isoprenaline or forskolin was reported to stimulate I_Cl,vol in canine atrial myocytes (Sorota, 1992) and in human atrial and ventricular myocytes (Oz & Sorota, 1995), although these effects may be complex, with some cells responding with stimulation, inhibition, no response, or a biphasic stimulation and inhibition (Du & Sorota, 1997). Finally, in guinea-pig atrial and ventricular myocytes, no consistent effects attributable to cAMP_i on I_Cl,vol were observed (Vandenberg et al. 1994).

At this time, no clear explanation is available to account for these reported disparate effects of cAMP_i on I_Cl,vol. It is possible, however, that some of these inconsistencies might arise due to the combined complex effects of cAMP_i and PKA on different types of Cl⁻ channels (in addition to I_Cl,vol) which may be concomitantly expressed in the same myocyte. For example, many cardiac myocytes that express I_Cl,vol also express CFTR Cl⁻ currents (I_Cl,CFTR) and/or Ca²⁺-activated Cl⁻ currents (I_Cl,Ca) (Harvey, 1996; Hiraoka et al. 1998; Hume et al. 2000), making it difficult to distinguish and interpret the effects of cAMP stimulation on channel subtypes. This problem is accentuated when membrane currents are only studied using asymmetric Cl⁻ gradients, since I_Cl,vol, I_Cl,CFTR and I_Cl,Ca all exhibit indistinguishable outwardly rectifying macroscopic current-voltage (I–V) relations under these conditions (Harvey, 1996; Hume et al. 2000). In the present study, we used symmetrical high Cl⁻ concentrations to effectively separate the outwardly rectifying I_Cl,vol from other possible contaminating Cl⁻ currents like I_Cl,CFTR and I_Cl,Ca (which have linear I–V relations in symmetrical high Cl⁻) and re-examined the cAMP_i regulation of native I_Cl,vol in guinea-pig and canine cardiac myocytes. We also examined cAMP_i regulation of the cloned I_Cl,vol in gpClC-3-transfected NIH/3T3 cells. Under these conditions, we find that elevation of cAMP_i consistently causes inhibition, not stimulation, of outwardly rectifying I_Cl,vol exhibited by both native and recombinant channels. Finally, experiments employing site-directed mutagenesis of the ClC-3 gene product, identify one primary phosphorylation site involved in the inhibitory modulation of I_Cl,vol by both PKC and PKA phosphorylation.

METHODS

Preparation of single cardiac myocytes

Single cardiac myocytes were obtained from guinea-pig and dog hearts using techniques as described previously (Collier et al. 1996; Yamazaki & Hume, 1997; Duan et al. 1997b, 1999). Briefly, guinea-pigs (350-450 g) were killed by a blow to the neck, and the hearts were quickly removed. The aorta was cannulated, and the heart was retrogradely perfused first with a modified Hepes-buffered Tyrode solution at 37°C, then with a nominally Ca²⁺-free Tyrode solution (until the heart ceased to beat), and finally with the same solution containing 0.1-0.3 mg ml⁻¹ collagenase (Type II, Sigma) and 1.0 % bovine serum albumin for 10 min. The left atrium was removed and further dissected into small pieces, and cell dissociation was achieved by gentle mechanical agitation in the high-K⁺ cell storage solution.

Adult mongrel dogs of either sex were anaesthetized with pentobarbital sodium (45 mg kg⁻¹i.v.) and their hearts were quickly removed and also perfused in the Langendorff mode. Atrial and ventricular myocytes were isolated in the same way as for guinea-pigs except for the enzyme concentration and duration of treatment. The heart was perfused with 0.5-1.5 mg ml⁻¹ collagenase for 15–30 min. All cells studied were rod shaped, exhibited clear cross striations, and lacked any visible blebs under isotonic conditions. Data from blebbed cells after hypotonic cell swelling were not included in the final analysis.

Functional expression of guinea-pig cardiac ClC-3 in NIH/3T3 cells

Full-length gpClC-3 was ligated to the mammalian expression vector pZeoSV (Invitrogen) and transfected into NIH/3T3 cells by electroporation as previously described (Duan et al. 1997b, 1999). For transient transfections, dishes were transfected with an appropriate combination of CD8 (a lymphocyte cell surface antigen) in the πH3-CD8 plasmid construct as a marker for transfection (4 μg) and gpClC-3 in the pZeoSV vector (20 μg). Transfected cells were identified by their binding to CD8-coated beads (M-450 CD8; Dyna-beads, Dynal Inc., Oslo, Norway). Cells were subcultured on glass coverslips for electrophysiological recordings.

Site-directed mutagenesis

The serine at position 51 was altered by an S51 to alanine site-specific mutation introduced into gpClC-3 cDNA as previously described (Duan et al. 1999). The mutation was confirmed by nucleotide sequencing of both strands of the mutated cDNA.

Electrophysiological techniques

The whole-cell patch-clamp technique (Hamill et al. 1981) was used to record currents in isolated guinea-pig and canine cardiac myocytes and gpClC-3-transfected NIH/3T3 cells. Patch pipettes were made from borosilicate glass capillaries and had a tip resistance of 2–5 MΩ. Ag-AgCl wires were immersed in the bath and pipette solutions and connected to a patch-clamp amplifier (Axopatch-1D, Axon Instruments). A 3 M KCl-agar bridge between the bath and the Ag-AgCl reference electrode was used to minimize changes in liquid junctional potential during some experiments. To obtain whole-cell Cl⁻I–V relations, cells were held at the holding potential (0, -40 or -60 mV) and test potentials (400 or 150 ms) were applied from -100 to +120 mV in 20 mV increments at an interval of 5 s (voltage-clamp protocol is shown in each figure). In some cases, I–V relations were measured using depolarizing voltage ramps applied at a rate of 0.2 V s⁻¹ following a voltage step from 0 mV to -100 mV. To examine time-dependent changes in current amplitude before and after different interventions, cells were clamped from a holding potential of 0 mV to hyperpolarizing potential of -100 mV for 100 ms, back to 0 mV for 10 ms, and then to a depolarizing potential of +100 mV for 100 ms. The same hyperpolarizing and depolarizing pulses were imposed every 30 s. Currents were filtered at a frequency of 1 kHz and digitized on-line at 5 kHz using an IBM PC/AT compatible computer and pCLAMP 6 software (Axon Instruments). Experiments were performed at room temperature (22-24°C).

Solutions and drugs

The modified Tyrode solution for cell isolation contained (mM): NaCl 126, KCl 5.4, CaCl₂ 2.0, MgCl₂ 1.0, NaH₂PO₄ 0.33, glucose 10 and Hepes 10; pH was adjusted to 7.4 with NaOH. The high-K⁺ cell storage solution contained (mM): KCl 20, KH₂PO₄ 10, glucose 10, L-glutamic acid 70, β-hydroxybutyric acid 10, taurine 10 and EGTA 10; along with 1 % alubumin, pH 7.4 (KOH). Osmolality was adjusted to 300 mosmol (kg H₂O)⁻¹ by adding mannitol.

Bath and pipette solutions were chosen to facilitate Cl⁻ current recording. Hypotonic (220 mosmol (kg H₂O)⁻¹, measured by freezing point depression, Osmomette; Precision Systems Inc.) bath solution for recordings in cardiac myocytes contained (mM): NaCl 90, MgCl₂ 1.0, BaCl₂ 3.0, NaH₂PO₄ 0.33, CsCl 10, Hepes 10 and glucose 5.5; pH = 7.4, [Cl⁻]_o = 108 mM. The isotonic bath solutions were the same as the hypotonic bath solutions except that the osmolality was adjusted to 300 mosmol (kg H₂O)⁻¹ by adding mannitol. Pipette solution for recordings in cardiac myocytes contained (mM): NMDG-Cl 108, BAPTA 10, Mg-ATP 5 and Hepes 10; pH = 7.4, 300 mosmol (kg H₂O)⁻¹, [Cl⁻]_i = 108 mM. In experiments using Ca²⁺-containing bath solution, 1 mM BaCl₂ in the bath solutions was replaced by equimolar CaCl₂, and BAPTA in pipette solutions was replaced by 2 mM EGTA.

Hypotonic (250 mosmol (kg H₂O)⁻¹) bath solution for whole-cell current recordings in NIH/3T3 cells contained (mM): NaCl 125, MgCl₂ 2.5, CaCl₂ 2.5 and Hepes 10; pH = 7.2, [Cl⁻]_o = 135 mM. Osmolality was adjusted to 300 mosmol (kg H₂O)⁻¹ by adding mannitol for isotonic bath solution. Pipette solution for recordings in NIH/3T3 cells contained (mM): NMDG-Cl 135, EGTA 2, Mg-ATP 5 and Hepes 10; pH = 7.2, 300 mosmol (kg H₂O)⁻¹), [Cl⁻]_i = 135 mM.

Isoprenaline and 8-Br-cAMP (both from Sigma, St Louis, MO, USA) were freshly dissolved in the bath solution. KT5720 (Calbiochem, San Diego, CA, USA) was prepared as a 1 mM stock solution in DMSO. The stock solution was diluted to the desired final concentration immediately before use. The final concentration of DMSO was < 0.1 %, which, by itself, did not affect Cl⁻ currents. All other chemicals were purchased from Sigma.

Data analysis

All data are expressed as means ±s.e.m. Statistical analysis was performed using Student's t test; a two-tailed probability (P) of < 0.05 was taken to indicate statistical significance.

RESULTS

Effects of 8-Br-cAMP on I_Cl,vol in guinea-pig atrial cells

We first studied the effect of an increase in intracellular cAMP (cAMP_i) on I_Cl,vol in guinea-pig atrial myocytes, since: (i) I_Cl,Ca is lacking in these myocytes (Sipido et al. 1995); (ii) the majority (90 %) of these myocytes express I_Cl,vol while only ∼10-21 % myocytes may express I_Cl,CFTR (Vandenberg et al. 1994; James et al. 1996); and (iii) the gene encoding I_Cl,vol, gpClC-3, was originally cloned from guinea-pig heart (Duan et al. 1997b). As shown in Fig. 1, hypotonic cell swelling activated an outwardly rectifying Cl⁻ current in guinea-pig atrial myocytes under symmetrical high Cl⁻ conditions (Fig. 1D and G). The properties of this swelling-induced current in guinea-pig myocytes, including its Cl⁻ dependence and pharmacological properties, have been described in more detail in a previous study from this laboratory (Duan et al. 1999). In the continued presence of the hypotonic bath solution, subsequent application of the membrane-permeable cAMP analogue 8-Br-cAMP (500 μM) to the superfusate significantly decreased Cl⁻ current amplitudes (Fig. 1C and F). This inhibition was voltage independent (37 ± 6 % inhibition at +80 mV, 44 ± 9 % inhibition at -80 mV, n = 6, P = not significant (n.s.)). The 8-Br-cAMP-sensitive current was indistinguishable from I_Cl,vol and had an outwardly rectifying I–V relation with symmetrical high Cl⁻ concentrations.

Figure 1. Effect of 8-Br-cAMP on I_Cl,vol in guinea-pig atrial cells.

Representative current traces from the guinea-pig atrial cells under isotonic (A), hypotonic (B) and hypotonic + 8-Br-cAMP (500 μM) (C) conditions are shown. D, difference currents induced by cell swelling. E, 8-Br-cAMP-sensitive currents. Voltage-clamp protocol is shown in the inset. The cell was first equilibrated with the isotonic solution, then exposed to the hypotonic solution. Thereafter, 8-Br-cAMP was applied. Mean I–V relations (n = 6) of I_Cl,vol under isotonic (^), hypotonic (□) and hypotonic + 8-Br-cAMP (▵) conditions are shown in F. I-V relations of hypotonic-induced currents (•) and 8-Br-cAMP-sensitive currents (▪) are shown in G and H, respectively. The rectification ratios at ±80 mV of the hypotonic-sensitive and the 8-Br-cAMP-sensitive difference currents were 2.50 ± 0.47 and 2.22 ± 0.54, respectively (P = n.s.).

To confirm that 8-Br-cAMP-induced inhibition of I_Cl,vol in guinea-pig atrial myocytes is indeed mediated by an increase in cAMP_i activity, we simultaneously monitored the well-known stimulatory effect of cAMP_i on L-type Ca²⁺ channels (Bean et al. 1984) by examining the effects of 8-Br-cAMP on Ba²⁺ currents (I_Ba) in the same cell. As shown in Fig. 2, when the cell was held at more negative potentials (-60 mV) to allow the activation of Ba²⁺ currents, hypotonic cell swelling resulted in the activation of I_Cl,vol, which overlapped with I_Ba. Addition of 8-Br-cAMP markedly increased the transient inward I_Ba while it significantly inhibited the cell swelling-activated outwardly rectifying I_Cl,vol.

Figure 2. Effect of 8-Br-cAMP on Ba²⁺ currents (I_Ba) and macroscopic Cl⁻ currents in guinea-pig atrial cells.

Representative current traces from guinea-pig atrial cells under isotonic (A), hypotonic (B) and hypotonic + 8-Br-cAMP (500 μM) (C) conditions are shown. The cell was held at -60 mV to record I_Ba through the voltage-dependent Ca²⁺ channel and I_Cl,vol simultaneously (voltage-clamp protocol is shown in the inset).

In the majority (7/10, 70 %) of guinea-pig atrial myocytes examined, 8-Br-cAMP consistently caused inhibition of I_Cl,vol. However, in a small percentage (3/10, 30 %) of atrial myocytes, 8-Br-cAMP elicited an apparent increase in total macroscopic Cl⁻ current amplitude. In contrast to the well-known outwardly rectifying properties of I_Cl,vol in symmetrical high Cl⁻ (see Fig. 3D and G), the 8-Br-cAMP-induced difference currents in these cells, although reversing near the estimated value of E_Cl (0 mV; the equilibrium potential for Cl⁻), exhibited a linear I–V relation (Fig. 3E and H). Since I_Cl,CFTR is known to exhibit a linear I–V relation in symmetrical high Cl⁻ solutions (Harvey et al. 1990) and a small percentage of guinea-pig atrial myocytes express I_Cl,CFTR (Vandenberg et al. 1994; James et al. 1996), it seems likely that the 8-Br-cAMP-induced Cl⁻ currents recorded in these cells may be contaminated by activation of I_Cl,CFTR.

Figure 3. Increase in macroscopic Cl⁻ currents by 8-Br-cAMP in guinea-pig atrial cells.

Representative current traces from the guinea-pig atrial cells under isotonic (A), hypotonic (B) and hypertonic + 8-Br-cAMP (500 μM) (C) conditions are shown. D, hypotonically induced currents. E, 8-Br-cAMP-induced currents. Mean I–V relations (n = 3) of macroscopic I_Cl under isotonic (^), hypotonic (□) and hypotonic + 8-Br-cAMP (▵) conditions are shown in F. I-V relations of hypotonically induced currents (•) and 8-Br-cAMP-induced currents (▪) are shown in G and H, respectively.

Effects of isoprenaline on I_Cl,vol in canine atrial myocytes

Highly variable effects of cAMP_i and PKA stimulation on I_Cl,vol have been reported in canine cardiac myocytes (Sorota, 1992; Tseng, 1992; Oz & Sorota, 1995; Du & Sorota, 1997). It is possible that the presence of mixed types of overlapping Cl⁻ currents may be partially responsible for the reported variable effects of cAMP_i and PKA stimulation on I_Cl,vol in canine atrial and/or ventricular myocytes. Contamination of macroscopic I_Cl,vol by I_Cl,CFTR as observed in some guinea-pig atrial myocytes is not expected, since most evidence to date has failed to show functional or molecular expression of CFTR Cl⁻ channels in canine cardiac myocytes (Sorota et al. 1991; Horowitz et al. 1993; Harvey, 1996; Hume et al. 2000). However, these cells express I_Cl,Ca in significant abundance, which would become activated as cytoplasmic Ca²⁺ rises in response to elevation of cAMP_i (Zygmunt, 1994; Collier & Hume, 1995; Yue et al. 1996). Therefore, it is possible that contamination from I_Cl,Ca could potentially confound any interpretation of the effects of cAMP_i on I_Cl,vol. We, therefore, first performed experiments in canine atrial and ventricular myocytes under conditions where possible contamination of I_Cl,vol by I_Cl,Ca was minimized by strongly buffering [Ca²⁺]_i and performing experiments in Ca²⁺-free bath (external) solutions.

Figure 4 shows the effects of isoprenaline on I_Cl,vol in a canine atrial myocyte under [Ca²⁺]_o-free and high (10 mM) intracellular BAPTA conditions. While only small currents could be detected in isotonic solutions (Fig. 4A), subsequent exposure of the cell to hypotonic solutions activated substantial macroscopic Cl⁻ currents (Fig. 4B). The hypotonically induced currents showed outward rectification and time-dependent inactivation at more positive potentials and had a reversal potential near the predicted value of E_Cl (0 mV) with symmetrical high Cl⁻ concentrations (Fig. 4D and G). These properties are identical to those of cardiac I_Cl,vol previously described in canine atrial (Sorota, 1992), and ventricular myocytes (Tseng, 1992) and in the heart of guinea-pig and other species (Vandenberg et al. 1994; Shuba et al. 1996; Duan et al. 1999; Hume et al. 2000). Isoprenaline (1 μM), a β-adrenergic receptor agonist which elevates cAMP_i, consistently decreased I_Cl,vol in the five different canine atrial myocytes examined (Fig. 4C and F). The inhibition by isoprenaline was voltage independent (48 ± 7 % inhibition at +80 mV vs. 39 ± 7 % inhibition at -80 mV, n = 5, P = n.s.). The isoprenaline-sensitive difference currents also showed time-dependent inactivation at positive potentials and had an outwardly rectifying I–V relation (Fig. 4E and H), confirming that the effects observed are due to inhibition of I_Cl,vol.

Figure 4. Effect of isoprenaline on I_Cl,vol in canine atrial cells in nominally Ca²⁺-free solutions.

Representative current traces from dog atrial cells under isotonic (A), hypotonic (B) and hypotonic + isoprenaline (1 μM) (C) conditions are shown. D, hypotonically induced currents. E, isoprenaline-sensitive currents. Mean I–V relations (n = 5) of I_Cl,vol under isotonic (^), hypotonic (□) and hypotonic + isoprenaline (▵) conditions are shown in F. I-V relations of hypotonically induced currents (•) and isoprenaline-sensitive currents (▪) are shown in G and H, respectively.

We also simultaneously verified the well-known stimulatory effect of isoprenaline on L-type Ca²⁺ channels in canine atrial cells using a ramp voltage-clamp protocol to simultaneously monitor the effects of isoprenaline on I_Cl,vol and I_Ba under [Ca²⁺]_o-free and high (10 mM) intracellular BAPTA conditions. While isoprenaline (1 μM) inhibited the outwardly rectifying I_Cl,vol, it also caused a significant increase in the amplitude of I_Ba (data not shown).

We next tested the effects of isoprenaline on I_Cl in canine atrial myocytes using conditions (Ca²⁺ (1.0 mM)-containing bath solutions and low [Ca²⁺]_i buffering) similar to those previously used in this preparation in which variable stimulatory and inhibitory effects of isoprenaline and forskolin on I_Cl,vol were reported (Sorota, 1992; Du & Sorota, 1997). However, we used symmetrical high Cl⁻ concentrations in an attempt to distinguish effects of isoprenaline on I_Cl,vol from possible effects on other types of macroscopic Cl⁻ currents, which may also be modulated by cAMP_i. Figure 5 shows an example of the effect of isoprenaline on hypotonic cell swelling-activated Cl⁻ currents in canine atrial myocytes recorded with a bath (external) solution containing 1.0 mM Ca²⁺ and the pipette (internal) solution containing 2 mM EGTA. Hypotonic cell swelling initially activated an outwardly rectifying Cl⁻ current, but subsequent application of isoprenaline (1 μM) caused a larger increase in the amplitude of both inward and outward currents, and the increase was greater for inward currents compared with outward currents (Fig. 5C). Thus, the isoprenaline-induced difference currents were no longer outwardly rectifying, but actually exhibited some inward rectification, and failed to exhibit time-dependent inactivation at positive membrane potentials (Fig. 5E and H). Thus, these isoprenaline-induced difference currents cannot be attributed solely to an effect on I_Cl,vol, since in symmetrical high Cl⁻ solutions, I_Cl,vol is outwardly rectifying and exhibits time-dependent inactivation at positive membrane potentials (cf. Figs 1 and 4). Similar effects of isoprenaline in canine atrial myocytes were also observed when the intracellular pipette solution contained 4 mM EGTA, suggesting that even this concentration of EGTA may not be enough to buffer [Ca²⁺]_i when cells are stimulated with isoprenaline.

Figure 5. Effect of isoprenaline on macroscopic Cl⁻ currents in canine atrial cells in Ca²⁺-containing solutions.

Representative current traces from dog atrial cells under isotonic (A), hypotonic (B) and hypotonic + isoprenaline (1 μM; C) conditions are shown. D, hypotonically induced currents. E, isoprenaline-induced currents. Mean I–V relations (n = 3) of macroscopic I_Cl under isotonic (^), hypotonic (□) and hypotonic + isoprenaline (▵) conditions are shown in F. I-V relations of hypotonically induced currents (•) and isoprenaline-induced currents (▪) are shown in G and H, respectively.

In summary, in canine atrial cells in Ca²⁺-free bath (external) solution and in the presence of strong intracellular Ca²⁺ buffering, isoprenaline consistently caused a monotonic decrease in I_Cl,vol in every cell examined, and the inhibited currents could be attributed to I_Cl,vol since the isoprenaline-induced difference currents were outwardly rectifying with symmetrical high Cl⁻ concentrations. However, when 1.0 mM Ca²⁺ was included in the external bath solution and the intracellular Ca²⁺ buffer was changed from 10 mM BAPTA to 2–4 mM EGTA, we also observed a stimulatory effect of isoprenaline on I_Cl,vol, suggesting that the stimulatory effects caused by isoprenaline are Ca²⁺ dependent.

Effects of 8-Br-cAMP on recombinant ClC-3 channels expressed in NIH/3T3 cells

We next studied the regulation of I_Cl,vol by cAMP in a heterologous gene expression system in which the gene encoding I_Cl,vol cloned from guinea-pig heart, gpClC-3, was stably or transiently expressed in NIH/3T3 cells (Duan et al. 1997b, 1999). Quantitative RT-PCR has recently revealed significant expression of ClC-3 mRNA in both atrium and ventricles of canine heart suggesting that ClC-3 may be responsible for the native I_Cl,vol in this tissue as well (F. Britton, C. Rossow, D. Duan, J. R. Hume & B. Horowitz, unpublished observations).

Stable or transient transfection of gpClC-3 into NIH/3T3 cells yields a basally active chloride conductance that is strongly modulated by cell volume (Fig. 6A and B). Properties of this expressed I_Cl,ClC-3 which resemble those reported for native I_Cl,vol in heart include: an outwardly rectifying unitary slope conductance of 40 pS, an anion selectivity of I⁻ > Cl⁻ > Asp⁻, inactivation at positive potentials, activation by extracellular hypotonicity, and inhibition by hypertonicity, by extracellular nucleotides, by phorbol esters, by stilbene derivatives, and by tamoxifen (Duan et al. 1997b). As shown in Fig. 6A, B and E, the expressed currents were outwardly rectifying under both isotonic and hypotonic conditions. 8-Br-cAMP (500 μM) consistently decreased the amplitude of I_Cl,ClC-3 in gpClC-3-transfected NIH/3T3 cells under hypotonic conditions (493 ± 20 vs. 334 ± 34 pA pF⁻¹ at +80 mV (n = 7, P < 0.02) and -301 ± 15 vs. -195 ± 23 pA pF⁻¹ at -80 mV (n = 7, P < 0.03) in the absence and presence of 8-Br-cAMP, respectively). Although the degree of rectification of I_Cl,ClC-3 in NIH/3T3 cells appeared to be somewhat less than that observed for I_Cl,vol in native cardiac myocytes, the difference in rectification ratios of I_Cl,vol in native guinea-pig atrial cells (2.50 ± 0.47, n = 6) and I_Cl,ClC-3 in NIH/3T3 cells (1.66 ± 0.21, n = 7) was not statistically significant (P = 0.113). As in native guinea-pig and canine myocytes (cf. Figs 1 and 4), the inhibition by 8-Br-cAMP was voltage independent (32 ± 7 % inhibition at +80 mV vs. 32 ± 7 % inhibition at -80 mV, n = 7, P = n.s.). The 8-Br-cAMP-sensitive currents (Fig. 6D and F) showed time-dependent inactivation at depolarizing potentials and had an outwardly rectifying I–V relation (rectification ratio = 1.62 ± 0.14, n = 7, P = n.s. vs. control), indicating that the cAMP-inhibited current can be attributed to I_Cl,ClC-3. No evidence for stimulation of I_Cl,ClC-3 was observed in any of the seven gpClC-3-transfected NIH/3T3 cells tested. To test whether these inhibitory effects of 8-Br-cAMP are due to activation of protein kinase A (PKA) in these experiments, we examined further the effects of a specific PKA inhibitor, KT5720 (Kase et al. 1987). As shown in Fig. 7A, pretreatment of 3T3 cells with KT5720 (200 nM) prevented the inhibition of I_Cl,ClC-3 by 500 μM 8-Br-cAMP (421 ± 58 vs. 452 ± 76 pA pF⁻¹ at +80 mV (n = 4, P = n.s.) and -255 ± 45 vs. -256 ± 44 pA pF⁻¹ at -80 mV (n = 4, P = n.s.) in the absence and presence of 8-Br-cAMP, respectively).

Figure 6. Effect of 8-Br-cAMP on recombinant ClC-3 channels expressed in NIH/3T3 cells.

Representative current traces from the ClC-3 expressed in NIH/3T3 cells under isotonic (A), hypotonic (B) and hypotonic + 8-Br-cAMP (500 μM) (C) are shown. D, 8-Br-cAMP-sensitive currents. Mean I–V relations (n = 7) of I_Cl,vol under hypotonic (^) and hypotonic + 8-Br-cAMP (□) conditions are shown in E. I-V relations of 8-Br-cAMP-sensitive currents (▵) are shown in F. The rectification ratios at ±80 mV of the hypotonic-sensitive and the 8-Br-cAMP-sensitive difference currents were 1.66 ± 0.21 and 1.62 ± 0.14, respectively (P = n.s.).

Figure 7. Both PKA inhibitor and site-directed mutation of the putative PKC phosphorylation site on the N-terminus of ClC-3 prevented the inhibitory effect of 8-Br-cAMP on ClC-3 channels.

A, whole-cell currents recorded from NIH/3T3 cells expressing wild-type ClC-3 under hypotonic conditions. Panels a and b show the representative currents in the presence of 200 nM KT5720 alone and KT5720 + 500 μM 8-Br-cAMP, respectively. Panel c shows the mean I–V curves from 4 different cells under conditions identical to those in panels a (^) and b (□). B, whole-cell currents recorded from NIH/3T3 cells expressing S51A ClC-3 mutant under hypotonic conditions. Panels a and b show the representative currents in the absence and presence of 500 μM 8-Br-cAMP, respectively. Panel c shows the mean I–V curves from 5 different cells under conditions identical to those in panels a (^) and b (□).

In a final series of experiments, we tested whether or not the inhibitory effects of 8-Br-cAMP on I_Cl,ClC-3 might be due to direct phosphorylation of the ClC-3 protein by PKA. One cytoplasmic consensus PKC phosphorylation site (serine 51 on the N-terminus) has recently been identified and functionally implicated in the regulation of ClC-3 channel activity by changes in cell volume (Duan et al. 1999). Since stimulation of PKA seems to mimic the inhibitory effects of PKC on I_Cl,ClC-3, and there is precedence showing that some putative phosphorylation sites often exhibit cross-recognition by different kinases (Kennelly & Krebs, 1991), we examined whether or not mutation of serine 51 to an alanine (S51A), which effectively eliminates PKC inhibition of I_Cl,ClC-3 (Duan et al. 1999), might also alter the ability of 8-Br-cAMP to inhibit I_Cl,ClC-3.

As previously reported, S51A ClC-3 channels expressed in NIH/3T3 cells were constitutively open under isotonic conditions and hypotonic cell swelling did not significantly change current amplitudes further (Duan et al. 1999). As shown in Fig. 7B, exposure of these S51A ClC-3-transfected NIH/3T3 cells to 8-Br-cAMP (500 μM) failed to produce any significant inhibition of I_Cl,ClC-3 in hypotonic solutions (530 ± 46 vs. 455 ± 43 pA pF⁻¹ at +80 mV (n = 5, P = n.s.) and -373 ± 25 vs. -314 ± 37 pA pF⁻¹ at -80 mV (n = 5, P = n.s.) in the absence and presence of 8-Br-cAMP, respectively). These results suggest that the observed weak inhibitory effects of 8-Br-cAMP on I_Cl,ClC-3 (Fig. 6) may be due to PKA phosphorylation of the same amino terminus serine residue (S51) that is predominantly phosphorylated by PKC, and also leads to inhibition of channel activity.

DISCUSSION

The regulation of cardiac I_Cl,vol by cAMP_i has been a controversial issue (Sorota, 1992; Hagiwara et al. 1992; Tseng, 1992; Oz & Sorota, 1995; Hall et al. 1995; Du & Sorota, 1997). In the present study, we examined cAMP_i regulation of native I_Cl,vol in atrial myocytes from guinea-pig and canine hearts and on recombinant I_Cl,vol encoded by gpClC-3, expressed in NIH/3T3 cells. In our experiments, we attempted to examine the effects of cAMP_i on I_Cl,vol in isolation and avoid possible confounding influences due to contamination from multiple Cl⁻ channel subtypes concomitantly expressed in the same cell. To this end, symmetrical high Cl⁻ concentrations were used to separate outwardly rectifying I_Cl,vol from other possible contaminating Cl⁻ currents (e.g. I_Cl,CFTR and I_Cl,Ca) which have linear I–V relations under these conditions. The results indicate that 8-Br-cAMP and agonists such as isoprenaline, which elevate cAMP_i, consistently inhibit the outwardly rectifying I_Cl,vol in native cardiac myocytes from both guinea-pig and canine hearts. Furthermore, identical effects of cAMP_i on recombinant I_Cl,vol were confirmed in gpClC-3-transfected NIH/3T3 cells, where the potential problem of multiple overlapping Cl⁻ currents is mimimized.

Our results are most consistent with those of Hall et al. (1995) in chick cardiac myocytes, where increases in cAMP_i caused inhibition and even prevented the activation of I_Cl,vol. Similar to our observations in the ClC-3-transfected NIH/3T3 cells (Fig. 7A), they also found that the inhibitory effect of cAMP_i on I_Cl,vol in chick cardiac myocytes is due to a cAMP-PKA-dependent phosphorylation process. They proposed that the activation of I_Cl,vol in chick heart cells requires dephosphorylation of a cAMP-dependent phosphorylation site. We now provide molecular evidence that mutation of a single serine residue S51 at the N-terminus of ClC-3, which may be a phosphorylation consensus site for both PKC and PKA, abolished the inhibitory effect of 8-Br-cAMP (Fig. 7B).

Our results are in marked contrast to those reported by Sorota and colleagues (Sorota, 1992; Oz & Sorota, 1995; Du & Sorota, 1997) in canine and human atrial myocytes. They originally found that elevation of cAMP_i stimulated I_Cl,vol (Sorota, 1992; Oz & Sorota, 1995) but later suggested that the effects of cAMP_i could actually be more complicated with both inhibitory and stimulatory effects seen in different cells or even in the same cell (Du & Sorota, 1997). They further suggested that the inhibitory effects were due to PKA phosphorylation, but the stimulatory effects were mediated by a PKA-independent pathway. These experiments relied solely upon the use of ramp voltage-clamp protocols to record whole-cell currents under asymmetric Cl⁻ gradients (low [Cl⁻]_i). Unfortunately, ramp voltage-clamp protocols do not provide kinetic information to distinguish I_Cl,vol from other types of Cl⁻ currents expressed in heart (Shuba et al. 1996) and with asymmetrical Cl⁻ conditions, most cardiac Cl⁻ currents exhibit nearly identical outwardly rectifying I–V relations (Harvey et al. 1990; Hume et al. 2000). Under these experimental conditions, therefore, contamination from multiple overlapping Cl⁻ current subtypes concomitantly expressed in the same cell is possible. With symmetrical high Cl⁻ concentrations, only I_Cl,vol remains outwardly rectifying, which effectively separates it from other possible contaminating Cl⁻ currents (Strange et al. 1996; Shuba et al. 1996; Okada, 1997; Hume et al. 2000). Thus, in the present study, we used voltage step protocols to investigate the effects of isoprenaline on I_Cl,vol under symmetrical high Cl⁻ conditions. We found that cell swelling-induced I_Cl,vol and the currents inhibited by elevation of cAMP_i in both guinea-pig and canine atrial myocytes and in the gpClC-3-transfected NIH/3T3 cells were all outwardly rectifying. These currents also showed time-dependent inactivation at positive potentials. Since the cAMP_i-inhibited currents have properties identical to I_Cl,vol and I_Cl,ClC-3, it is clear that they can be attributed to inhibition of I_Cl,vol. It is possible that an earlier report (Vandenberg et al. 1994) which failed to see inhibition of I_Cl,vol by isoprenaline might also be complicated by the presence of contaminating overlapping Cl⁻ channel subtypes.

In a small percentage of guinea-pig cardiac myocytes, we observed some stimulation of whole-cell Cl⁻ currents by 8-Br-cAMP. However, these currents showed no time-dependent inactivation and had linear I–V relations with symmetrical high Cl⁻ concentrations. It is well known that cAMP activates I_Cl,CFTR in guinea-pig atrial and ventricular myocytes (Bahinski et al. 1989; Harvey et al. 1990; Vandenberg et al. 1994; Shuba et al. 1996). Thus, the currents induced by 8-Br-cAMP in a small percentage of guinea-pig atrial myocytes in the present study have properties most consistent with activation of I_Cl,CFTR. However, the effects actually observed in these cells are complicated and probably reflect the combined inhibitory effects of cAMP_i on I_Cl,vol as well as the stimulatory effects of cAMP_i on I_Cl,CFTR.

In canine atrial myocytes, stimulatory effects of isoprenaline on whole-cell Cl⁻ currents were only observed in Ca²⁺-containing bath solutions and with ${\text{Ca}}_{\text{i}}^{2+}$ weakly buffered. This is consistent with the results reported previously by Sorota and colleagues (Sorota, 1992; Du & Sorota, 1997) with a similar (4 mM) concentration of EGTA in the intracellular pipette solution but a higher [Ca²⁺]_o (1.8 vs. 1.0 mM) in the bath solutions. In contrast, with Ca²⁺-free bath solutions and with ${\text{Ca}}_{\text{i}}^{2+}$ strongly buffered, we found that isoprenaline consistently caused inhibition of the hypotonic cell swelling-activated Cl⁻ currents. It is well known that isoprenaline causes a marked elevation of intracellular Ca²⁺ due to stimulation of L-type Ca²⁺ channels (Bean et al. 1984) and I_Cl,Ca can be activated as [Ca²⁺]_i rises in response to an elevation of cAMP_i (Zygmunt, 1994; Kawano et al. 1995; Zygmunt et al. 1998). It is also possible that exposure to hypotonic solutions alone, which has been reported to cause inhibition of Na⁺-Ca²⁺ exchange (Wright et al. 1995), could also contribute to a rise in [Ca²⁺]_i in isolated cardiac myocytes. Although I_Cl,Ca was initially described as a transient outward current (I_to2) with a bell-shaped I–V relation (Zygmunt & Gibbons, 1992), later studies in canine ventricular myocytes found that I_Cl,Ca showed little or no voltage- or Ca²⁺-induced inactivation and is essentially time and voltage independent when intracellular Ca²⁺ is pseudo-clamped to constant levels (Zygmunt, 1994; Yue et al. 1996). In addition, I_Cl,Ca exhibits a linear I–V relation with symmetrical high Cl⁻ (Kawano et al. 1995; Yamazaki et al. 1999). Thus, the isoprenaline-activated Cl⁻ currents observed in canine atrial myocytes have properties most consistent with activation of I_Cl,Ca, although the current changes measured under these conditions probably reflect the combined inhibitory effects of cAMP_i on I_Cl,vol as well as the stimulatory effects of cAMP_i on I_Cl,Ca.

It is possible that the cAMP_i-stimulated linear current observed in both guinea-pig and dog atrial myocytes might be due to modulation of I_Cl,vol if isoprenaline alters the voltage dependence of I_Cl,vol and linearizes its I–V relation. However, this possibility seems highly unlikely since such an effect was not observed in canine atrial myocytes in Ca²⁺-free bath (external) solutions with ${\text{Ca}}_{\text{i}}^{2+}$ strongly buffered (Fig. 4), or in gpClC-3-transfected NIH/3T3 cells, a system with minimal contamination from I_Cl,CFTR and I_Cl,Ca. In these cells, 8-Br-cAMP caused only consistent inhibition and no alteration in the rectification and voltage dependence of I_Cl,ClC-3 (Fig. 6).

In addition, we demonstrated that the inhibitory effects of 8-Br-cAMP on I_Cl,ClC-3 are due to phosphorylation by PKA, since these effects were prevented by the specific PKA inhibitor KT5720. We also showed that this inhibition probably involves direct PKA phosphorylation of the ClC-3 protein itself as the S51A mutation abolished the inhibition by 8-Br-cAMP. While the serine at position 51 in gpClC-3 is within a perfect consensus PKC phosphorylation sequence (RINSKK), this sequence also represents a less well-utilised PKA phosphorylation sequence (R-X2-S, where X represents any amino acid) (Kennelly & Krebs, 1991). Thus, phosphorylation of S51 by either PKC or PKA appears to inhibit I_ClC-3, although the former effect appears to be more potent. The transcriptional expression of ClC-3 in both canine atrial and ventricular myocytes has recently been demonstrated (F. Britton, C. Rossow, D. Duan, J. R. Hume & B. Horowitz, unpublished observations), but further studies are required to determine if S51 plays a similar role to that in gpClC-3.

In summary, our results suggest that contamination of macroscopic I_Cl,vol by I_Cl,CFTR and/or I_Cl,Ca may account for some of the inconsistent and controversial effects of cAMP_i on I_Cl,vol reported previously in native cardiac myocytes. When contributions from contaminating macroscopic Cl⁻ currents are prevented, elevation of cAMP_i consistently inhibits I_Cl,vol in both guinea-pig and canine native cardiac myocytes and inhibits recombinant ClC-3 channels expressed in NIH/3T3 cells. These experiments emphasize the potential pitfalls of assessing I_Cl,vol regulation using experimental conditions (asymmetric Cl⁻ gradients with low [Cl⁻]_i and voltage ramps) which do not allow reliable separation of overlapping macroscopic Cl⁻ currents. In the case of I_Cl,vol, such a separation can easily be achieved using symmetrical high Cl⁻ concentrations, since these channels are unique in retaining their outwardly rectifying properties under these conditions. It is tempting to speculate that similar complications due to the presence of multiple overlapping macroscopic Cl⁻ currents, as illustrated in the present study, may also be at least partially responsible for recent controversies regarding PKC modulation of I_Cl,vol in heart (Duan et al. 1995, 1997b, 1999; Clemo & Baumgarten, 1998; Clemo et al. 1999; Du & Sorota, 1999) and other mammalian tissues (Okada, 1997), or even the reported sensitivity of I_Cl,vol to niflumic acid and other putative Cl⁻ channel antagonists (Sorota, 1994) as well.