Modulation of rat cardiac sodium channel by the stimulatory G protein α subunit

Abstract

1. Modulation of cardiac sodium currents (I_Na) by the G protein stimulatory α subunit (Gsα) was studied using patch-clamp techniques on freshly dissociated rat ventricular myocytes.

2. Whole-cell recordings showed that stimulation of β-adrenergic receptors with 10 μM isoprenaline (isoproterenol, ISO) enhanced I_Na by 68·4 ± 9·6 % (mean ±s.e.m.; n= 7, P < 0·05vs. baseline). With the addition of 22 μg ml⁻¹ protein kinase A inhibitor (PKI) to the pipette solution, 10 μM ISO enhanced I_Na by 30·5 ± 7·0 % (n= 7, P < 0·05vs. baseline). With the pipette solution containing both PKI and 20 μg ml⁻¹ anti-Gsα IgG or 20 μg ml⁻¹ anti-Gsα IgG alone, 10 μM ISO produced no change in I_Na.

3. The effect of Gsα on I_Na was not due to changes in the steady-state activation or inactivation curves, the time course of current decay, the development of inactivation, or the recovery from inactivation.

4. Whole-cell I_Na was increased by 45·2 ± 5·3 % (n= 13, P < 0·05vs. control) with pipette solution containing 1 μM Gsα27-42 peptide (amino acids 27-42 of rat brain Gsα) without altering the properties of Na⁺ channel kinetics. Furthermore, application of 1 nM Gsα27-42 to Na⁺ channels in inside-out macropatches increased the ensemble-averaged I_Na by 32·5 ± 6·8 % (n= 8, P < 0·05vs. baseline). The increase in I_Na was reversible upon Gsα27-42 peptide washout. Single channel experiments showed that the Gsα27-42 peptide did not alter the Na⁺ single channel current amplitude, the mean open time or the mean closed time, but increased the number of functional channels (N) in the patch.

5. Application of selected short amino acid segments (Gsα27-36, Gsα33-42 and Gsα30-39) of the 16 amino acid Gsα peptide (Gsα27-42 peptide) showed that only the C-terminal segment of this peptide (Gsα33-42) significantly increased I_Na in a dose-dependent fashion. These results show that cardiac I_Na is regulated by Gsα via a mechanism independent of PKA that results in an increase in the number of functional Na⁺ channels. In addition, a 10 residue domain (amino acids 33-42) near the N-terminus of Gsα is important in modulating cardiac Na⁺ channels.

The voltage-dependent Na⁺ channel is responsible for the upstroke of action potentials in most excitable cells. It is also a major target for modulation by pharmacological agents. The G protein-mediated signal transduction pathways are important mechanisms of cellular and physiological regulation (Hille, 1992). It is known that ion channels can be regulated by two different pathways involving G proteins (Breitwieser, 1991; Hille, 1992; Wickman & Clapham, 1995). One involves diffusible second messengers, including intracellular levels of Ca²⁺, cyclic nucleotides and diacylglycerol. G protein α subunits (Gα) and βγ subunits (Gβγ) can modulate specific effectors through regulation of intracellular second messenger pathways (Codina et al. 1987; Schwiebert et al. 1995, Neer & Smith, 1996; Herlitze et al. 1996; Ikeda, 1996; Ismailov et al. 1996). The other pathway is referred to as membrane-delimited G protein regulation, which is independent of second messengers. This type of regulation has been shown to modulate Na⁺ channels, K⁺ channels, Ca²⁺ channels and Cl⁻ channels (Kurachi et al. 1986; Scott & Dolphin, 1987; Tilly et al. 1991; Matsuda et al. 1992; Manavalan et al. 1995; Schreibmayer et al. 1996). The direct binding of Gβγ to ion channels has been reported for the G protein-gated inwardly rectifying K⁺ channel (GIRK1) (Reuveny et al. 1994; Huang et al. 1995), brain voltage-dependent Ca²⁺ channels (α_1A and α_1B) (Ikeda, 1996; Herlitze et al. 1996; De Waard et al. 1997) and brain Na⁺ channels (Ma et al. 1997). The direct binding of the stimulatory G protein α subunit (Gsα) to skeletal Ca²⁺ channels has also been reported by Hamilton et al. (1991).

Direct Gα modulation of effector proteins suggests an interaction between specific amino acid residues or sequences of the two protein molecules. A small linear stretch of native Gα sufficient to stimulate adenylyl cyclase was identified near the C-terminus (Osawa et al. 1990; Itoh & Gilman, 1991). Rarick et al. (1992) reported that a 22 amino acid peptide from the C-terminal region of G protein transducin was able to activate cGMP phosphodiesterase. In addition, it has been shown that a 16 amino acid peptide near the N-terminus of native Gsα was able to modulate cardiac Na⁺ channels (Matsuda et al. 1992). However, the mechanism of Gsα modulation of cardiac I_Na and the residues of Gsα that interact with and directly activate the Na⁺ channels still remain uncertain. In this study, we demonstrate that the activity of Na⁺ channels in rat ventricular myocytes is directly regulated by Gsα. This direct regulation enhances I_Na through an increase in the number of functional Na⁺ channels in the cell membrane.

METHODS

Preparation of single ventricular myocytes

Single ventricular myocytes were prepared as previously described (Isenberg & Klockner, 1982). These procedures were approved by the Animal Care and Use Committee at The University of Iowa. Briefly, Sprague-Dawley rats (200-250 g body weight) were anaesthetized with methoxyflurane. Hearts were rapidly excised and placed in ice-cold nominally Ca²⁺-free Tyrode solution containing (mM): NaCl, 138; KCl, 4·5; MgCl₂, 0·5; Na₂HPO₄, 0·33; Hepes, 10; and glucose, 5·5 (pH 7·38). The aorta was cannulated, and the heart was perfused using a modified Langendorff apparatus with nominally Ca²⁺-free Tyrode solution containing 0·1 % (w/v) bovine serum albumin (BSA) for 5 min at 37°C. The perfusate was switched over to nominally Ca²⁺-free Tyrode solution containing 0·6 mg ml⁻¹ collagenase (Worthington, CLS-2, 347 U mg⁻¹) and 0·1 % (w/v) BSA for 7 min at 37°C. The ventricles were removed and placed into 25 ml of fresh 0·6 mg ml⁻¹ collagenase solution for 5 min. The ventricles were cut into small pieces (approximately 1 mm³) and filtered through a medium mesh. After washing twice with nominally Ca²⁺-free Tyrode solution, single cells were maintained in KB solution containing (mM): KOH, 70; KCl, 40; L-glutamic acid, 50; taurine, 20; MgCl₂, 0·5; K₂HPO₄, 1·0; EGTA, 0·5; Hepes, 10; creatine, 5·0; pyruvic acid, 5·0; Na₂ATP, 5·0; and glucose, 10 (pH 7·38). All solutions were vigorously oxygenated for 30 min before starting the experiment.

Patch clamp measurements of I_Na

Voltage-dependent Na⁺ currents in isolated cardiac ventricular myocytes were measured using patch-clamp techniques (Hamill et al. 1981). Whole-cell I_Na and inside-out macropatch I_Na were recorded with an integrating amplifier (Axopatch 200B, Axon Instruments) and filtered with an eight-pole low-pass Bessel filter (902 LPF, Frequency Devices Inc., Haverhill, MA, USA) with a bandwidth (-3 dB) of 5 kHz and a sampling rate of 50 kHz (12-bit resolution). Borosilicate glass capillaries (Corning 7052, Warner Instruments, Inc.) were used to fabricate patch pipettes. Electrode resistance in the pipette solution (see below) ranged from 0·5 to 1 MΩ and seal resistances were 1-5 GΩ. Whole-cell series resistance was compensated to more than 80 % of the uncompensated value. The cell capacitance was 9·2 ± 0·21 pF (n= 60) and was compensated electronically. The whole-cell bath solution contained (mM): NaCl, 20; choline chloride, 130; CaCl₂, 1; CoCl₂, 2; MgCl₂, 2; KCl, 4·5; Hepes, 10; and glucose, 5·5 (pH 7·38). The pipette solution for whole-cell recording contained (mM): CsCl, 130; Na₂ATP, 5; GTP, 0·5; EGTA, 5·0; CaCl₂, 0·5; MgCl₂, 2·0; and Hepes, 10 (pH 7·25). The bath solution for the inside-out macropatches contained (mM): KCl, 140; MgCl₂, 2; GTP, 0·5; EGTA, 10; and Hepes, 10 (pH 7·25). The pipette solution for inside-out macropatch recording contained (mM): NaCl, 140; CaCl₂, 1; MgCl₂, 2; Hepes, 10; and nitrendipine, 0·002 (1 mM stock solution dissolved in DMSO) (pH 7·38). Because of the spontaneous shifts in the current inactivation-voltage relationship after establishing the whole-cell condition, I_Na data were recorded only after the currents became stable (at least 10 min).

Inside-out single channel recordings were performed with an improved patch clamp technique (Benndorf, 1994). The patch pipettes were pulled from thick-walled borosilicate glass (glass type 7740, Garner Glass Company) with an external diameter of 2·0 mm and an internal diameter of 0·5 mm. The pipette was coated with Sylgard 184 as close as possible to the tip before being fire polished. The resistance of the pipette filled with pipette solution was 5-10 MΩ and the typical seal resistance was 500-1000 GΩ. Single Na⁺ currents were recorded with an Axopatch 200B and an eight-pole low-pass Bessel filter with a bandwidth of 10 kHz and sampling rate of 100 kHz (12-bit resolution). Single Na⁺ channel current was identified by the amplitude and the fast time course of the ensemble-averaged current. The bath solution for single Na⁺ current recordings contained (mM): KCl, 250; CsCl, 20; MgCl₂, 2; EGTA, 5; Hepes, 10; and GTP, 0·5 (pH 7·25). The pipette solution contained (mM): NaCl, 250; CaCl₂, 1; MgCl₂, 2; Hepes, 10; and nitrendipine, 0·002 (pH 7·38). The bath solutions were superfused using a direct current-powered λ pump (Model 700, Instech Laboratories, Inc., Plymouth Meeting, PA, USA) at a rate of 1-2 ml min⁻¹ and solution exchanges were complete within 30 s. All experiments were performed at 21-23°C.

Experimental protocols and data analysis

Steady-state activation curves were constructed with the membrane potential held at -100 mV for 5 s, and a series of 20 ms test pulses were elicited which ranged from -80 to +50 mV in increments of 5 mV. The peak value of I_Na (I_Na,peak) at each membrane potential (V_m) was plotted. The relationship of peak I_Navs. V_m was fitted with the following equation:

where V_rev is the reversal potential of I_Na and G_Na is the voltage-dependent sodium current conductance. G_Na was fitted using a Boltzmann distribution equation:

where G_Na,max is the maximum conductance, V_½ is the membrane potential at half-maximal conductance and k is the slope factor.

The voltage-dependent steady-state inactivation relationship was investigated with a standard two-pulse protocol. For maximum recovery from slow inactivation of Na⁺ channels, a test potential to -20 mV was preceded by a 500 ms preconditioning pulse ranging from -160 to +20 mV with a 0·1 Hz pulse interval. Under these conditions, the slow inactivation-dependent shift in the inactivation curves was minimized. The normalized curves (I_Na/I_Na,max) were fitted using a Boltzmann distribution equation:

where I_Na,max is the peak Na⁺ current at -20 mV measured from the most negative preconditioning pulse potential (-160 mV), V_m is the preconditioning pulse potential, V_½ is the membrane potential of half-maximum I_Na, and k is the slope factor.

To characterize the recovery from steady-state inactivation, experiments were performed using a two-pulse protocol. A 500 ms preconditioning pulse from a holding potential of -100 mV to 0 mV was followed by a recovery period of variable duration from 1 to 2000 ms. The recovery period was followed by a 20 ms test pulse to 0 mV. The amplitude of the peak I_Na during the test pulse was normalized to the peak value after complete recovery from inactivation. The time course of recovery from inactivation was analysed by a two-exponential fit using the equation:

where I_Na,max is the peak Na⁺ current at 0 mV measured following a 2000 ms recovery period, A₁ and A₂ represent the relative area under the two-exponential fit (A₁+A₂= 1), and τ_f and τ_s represent fast and slow time constants, respectively.

To assess the development of inactivation, a conditioning pulse at -65 mV with a variable duration from 0·5 to 500 ms was evoked from a holding potential of -100 mV, followed by a test pulse to -10 mV. The peak current amplitude elicited by the test pulse was normalized to the peak current without conditioning prepulses. The normalized current amplitude was plotted as a function of the duration of the conditioning prepulses. This relationship was fitted with a mono-exponential equation.

For single channel data analysis, the capacitative transients and leak current were removed by subtraction of an averaged record from 50 empty traces. The subtracted traces were idealized at 10 kHz. The opening and the closing transitions of single channels were detected using the conventional 50 % threshold of the event amplitude. For analysis of the kinetic properties of single Na⁺ channels, patches containing only one channel were chosen. The mean open time was reliably measured between an upper limit (τ₀) and a lower limit which was given by the rise time (t_r) (Colquhoun & Sigworth, 1995), where τ₀= 0·05/kf_c exp(Φ²/2σ²), t_r= 0·33/f_c, k was a constant of 1·25f², and σ_noise (the background noise level) was 0·7. At the cutoff frequency f_c= 10 kHz, unitary current amplitude A₀= 2·2 pA and 50 % threshold Φ= 1·1 pA, the reliable measurement of the mean open time was 33-2000 μs. The distributions of dwell times of both channel events (t₀) and noise events (t_noise) were determined by the baseline method as described by Benndorf (1994). Computer-detected openings were confirmed by visual observation and used to generate idealized records from which histograms of amplitude, closed-time and open-time distributions were constructed. The last channel opening was determined in order to separate the closed state from an inactivation state. The closed time histogram was fitted with a two-exponential equation, ignoring the brief closed times less than t_r. The channel open probability (P_o) was determined using the equation:

where I_Na represents the macroscopic Na⁺ current averaged from single channel recordings, N is the number of functional channels in the patch, and i is single channel current amplitude. Single channel current amplitudes were determined by construction of amplitude histograms, which were fitted with Gaussian curves by using a χ² non-linear regression routine.

Data were collected and analysed using pCLAMP 6.04 software (Axon Instruments). Commercially available software modules from Origin 4.0 (Microcal Software, Inc., Northampton, MA, USA) or IGOR Pro 3.1 (WaveMetrics Inc., Lake Oswego, OR, USA) were used to fit curves.

Data are presented as means ±s.e.m. Student's paired t test was used to compare data obtained before (baseline) and after intervention (treatment group). Student's unpaired t test was used to compare two different groups. A one-way ANOVA followed by contrast testing was used to compare data from multiple groups. A statistically significant difference was defined as P < 0·05.

Development of anti-Gsα IgG and synthesis of Gsα peptide

A synthetic peptide that corresponds to amino acids 27-42 of the rat Gsα (C-K-Q-L-Q-K-D-K-Q-V-Y-R-A-T-H-R) coupled to haemocyanin was used for immunization of rabbits as described previously (Green et al. 1982). Details regarding the development of anti-Gsα antiserum and characterization of the immunoreactivity and specificity have previously been reported (Cai et al. 1993). Anti-Gsα IgG was isolated using immobilized protein A columns (Immuno Pure, Pierce). Using SDS-PAGE, the anti-Gsα IgG was shown to immunoreact with two bands (45 and 53 kDa) specific for Gsα and not with other membrane proteins in cardiac membranes (Cai et al. 1993). Synthesis of Gsα oligopeptides with sequences homologous to various segments of the native Gsα was performed by the Peptide Synthesis Core Facilities at the University of Iowa. Based on the 16 amino acid peptide of Gsα, short amino acid peptides corresponding to the amino terminus (Gsα27-36: C-K-Q-L-Q-K-D-K-Q-V), the carboxyl terminus (Gsα33-42: D-K-Q-V-Y-R-A-T-H-R), or the middle (Gsα30-39: L-Q-K-D-K-Q-V-Y-R-A) of the Gsα peptide were synthesized.

RESULTS

β-Adrenergic modulation of cardiac Na⁺ channels

Whole-cell recordings were used to investigate the effects of isoprenaline (isoproterenol, ISO) on I_Na in adult rat ventricular myocytes. Figure 1A and B shows that stimulation of the β-adrenergic receptor with 10 μM ISO increased peak I_Na by 68·4 ± 9·6 % (from 1·7 ± 0·3 nA at baseline to 2·5 ± 0·5 nA with ISO treatment; n= 7, P < 0·05). The rate of I_Na decay was well fitted with a two-exponential equation after treatment with ISO (τ_f= 0·54 ± 0·1 ms, τ_s= 2·16 ± 0·24 ms, n= 7), whereas in the control recordings, a mono-exponential fit was sufficient (τ= 1·67 ± 0·24 ms, n= 7). After treatment with ISO, the fast time constant (τ_f) of I_Na decay was faster than that of the control (P < 0·05). The I-V relationship showed the enhancement of whole-cell peak I_Na with the application of 10 μM ISO (Fig. 1C). The time course of enhancement of whole-cell peak I_Na by application of 10 μM ISO is plotted in Fig. 1D. The effect of ISO was reversible upon washout or by β-adrenergic receptor blockade with 10 μM propranolol in the presence of ISO.

Figure 1. β-Adrenergic modulation of cardiac I_Na.

Whole-cell I_Na was elicited from a holding potential of -100 mV to test potentials ranging from -80 to +50 mV in 5 mV step increments before (A) and after (B) application of 10 μM ISO to rat ventricular myocytes. C shows the I-V relationships, which represent the peak I_Na values before (○) and after application (•), and after washout (□) of 10 μM ISO (data from a typical experiment). The I-V relationship curves were fitted by the equation: I_Na=G_Na,max(V_m - V_rev)/{1 + exp[(V_½ - V_m)/k]} (see Methods). D, the time course of whole-cell peak I_Na enhancement (-40 mV) by 10 μM ISO, which was reversible upon ISO washout. The vertical lines indicate the start of solution changes.

Modulation of cardiac Na⁺ channels by Gsα and blockade by anti-Gsα antibody

To test the effects of β-adrenergic receptor-stimulated release of endogenous Gsα, PKA-dependent phosphorylation was inhibited with 22 μg ml⁻¹ protein kinase A inhibitor (PKI) in the pipette solution. I_Na was elicited from a holding potential of -100 mV to a test potential of -40 mV. With 22 μg ml⁻¹ PKI in the pipette solution, application of 10 μM ISO enhanced whole-cell I_Na by 30·5 ± 7·0 % (n= 8) (Fig. 2A) without altering the time constant of I_Na decay (Fig. 2B). There was a significant difference in the percentage of ISO-stimulated enhancement of I_Na in the presence and absence of PKI (30·5 ± 7·0 % with PKI, n= 8, vs. 68·4 ± 9·6 % without PKI, n= 7, P < 0·05).

Figure 2. Modulation of I_Na by Gsα.

A, with 22 μg ml⁻¹ PKI in the pipette solution, 10 μM ISO enhanced whole-cell I_Na, which was elicited from a holding potential of -100 mV to a test potential of -40 mV. B, superimposed normalized current tracings in control and 10 μM ISO from A. C, recordings of ISO effects on whole-cell I_Na with the pipette solution containing 22 μg ml⁻¹ PKI alone (○), 20 μg ml⁻¹ anti-Gsα IgG alone (▴), or 22 μg ml⁻¹ PKI plus 20 μg ml⁻¹ anti-Gsα IgG (•) plotted against time. ISO (10 μM) produced no change in I_Na when 20 μg ml⁻¹ anti-Gsα IgG or 20 μg ml⁻¹ anti-Gsα IgG plus 22 μg ml⁻¹ PKI were present in the pipette solution.

Complete inhibition of the PKA-dependent pathway was confirmed by application of 500 μM 8-(4-chlorophenylthio)-cAMP (8-CPT-cAMP), a membrane permeable cAMP analogue, in the presence of 22 μg ml⁻¹ PKI in the pipette solution (data not shown). These results showed that under these experimental conditions, 8-CPT-cAMP did not alter the size or the time course of I_Na. To confirm the direct effect of Gsα in our studies, 22 μg ml⁻¹ PKI plus 20 μg ml⁻¹ anti-Gsα IgG were included in the pipette solution. There was no change in the amplitude of I_Na or the time course of decay of I_Na on exposure to 10 μM ISO (Fig. 2C) (764 ± 65 pA baseline vs. 736 ± 52 pA ISO treatment; n= 8, P= n.s.). To determine whether the ISO effects on I_Na were specific to Gsα, we examined the effect of ISO in the presence of 20 μg ml⁻¹ anti-Gsα IgG alone in the pipette solution. ISO (10 μM) failed to stimulate I_Na with 20 μg ml⁻¹ anti-Gsα IgG in the pipette solution, suggesting that neither the adenylyl cyclase nor direct Gsα pathways could be activated (Fig. 2C). This suggests that β-adrenergic receptor modulation of I_Na involves two different G protein-dependent mechanisms, one via PKA and the other via a direct membrane-delimited G protein pathway. Accordingly, each pathway contributed to about 50 % enhancement of cardiac Na⁺ channel activity in the rat ventricular myocytes.

Gsα modulation of cardiac Na⁺ channel kinetics

To assess the mechanism of I_Na modulation by direct Gsα effects, the PKA-dependent effects were inhibited with 22 μg ml⁻¹ PKI in the pipette solution. Application of 10 μM ISO enhanced I_Na from a control value of 1·7 ± 0·4 nA to 2·2 ± 0·6 nA (Fig. 3C; n= 7, P < 0·05). The I-V relationship shows that there was no significant change in reversal potential (25 ± 1·5 mV at baseline vs. 26·4 ± 1·8 mV with ISO treatment; n= 7, P= n.s.). The effect of Gsα on I_Na was not due to a change in the half-activation potential (V_½= -45·2 ± 1·2 mV at baseline vs. -44·3 ± 1·2 mV with ISO treatment; n= 7, P= n.s.) or the slope factor (k= 2·3 ± 0·3 mV e-fold⁻¹ at baseline vs. 2·7 ± 0·2 mV e-fold⁻¹ with ISO treatment; n= 7, P= n.s.) of the steady-state activation curve (Fig. 3D). In addition, Gsα had no effect on the potential of half-inactivation (V_½= -77·2 ± 1·2 mV at baseline vs. -75·2 ± 1·6 mV with ISO treatment; n= 7, P= n.s.) or on the slope factor (k= 7·3 ± 0·4 mV e-fold⁻¹ at baseline vs. 7·5 ± 0·5 mV e-fold⁻¹ with ISO treatment; n= 7, P= n.s.) of the steady-state inactivation curve (Fig. 4A-C).

Figure 3. Effect of Gsα on G_max and activation of Na⁺ channels.

Whole-cell I_Na was elicited from a holding potential of -100 mV to test potentials ranging from -80 to +50 mV in 5 mV increments before (A) and after (B) application of 10 μM ISO, with pipette solution containing 22 μg ml⁻¹ PKI. C, the averaged I-V relationships (n= 7) (○, baseline; □, 10 μM ISO; ⋄, washout). ISO enhanced the G_Na,max of I_Na by 30·5 ± 7·0 %. D, the normalized channel conductance (G_Na) was fitted using a Boltzmann equation. Gsα did not change V_½ or the slope factor, k (n= 7).

Figure 4. Effect of Gsα on the voltage dependence of Na⁺ channel inactivation.

A, three superimposed inactivation curves obtained under control conditions, after application of 10 μM ISO and following washout of ISO. Experiments were performed with 22 μg ml⁻¹ PKI in the pipette solution. The normalized I_Na plotted against preconditioning pulse potential was fitted using a Boltzmann equation. I_Na traces elicited with preconditioning pulses of 500 ms duration from -160 to -80 mV to a test potential of +20 mV are shown in B (control) and C (10 μM ISO). D, recovery of I_Na from inactivation was studied using a standard two-pulse protocol. Normalized I_Na was plotted against the recovery duration and fitted with a two-exponential equation. There was no significant difference in τ_f and τ_s before (○) or after (□) application of 10 μM ISO. E, the normalized I_Na at -10 mV plotted against the variable prepulse intervals at -65 mV. The time constant of the development of inactivation at -65 mV was fitted with a mono-exponential equation and was not significantly different before (○) or after (▵) application of 10 μM ISO.

To further assess the effects of Gsα, we measured the time constant of recovery from inactivation at different holding potentials. The time constant of recovery from inactivation at -100 mV was best described by a two-exponential process. Application of 10 μM ISO with 22 μg ml⁻¹ PKI in the pipette solution did not affect the fast component (τ_f= 20·5 ± 1·9 ms at baseline vs. 22·3 ± 2·1 ms with ISO treatment; n= 5, P= n.s.) or the slow component (τ_s= 179·1 ± 15·5 ms at baseline vs. 198·08 ± 20·15 ms with ISO treatment; n= 5, P= n.s.) of channel recovery from inactivation (Fig. 4D). There were also no changes in τ_f and τ_s at a holding potential of -80 mV following application of ISO (data not shown).

The time course of the development of inactivation was measured with a standard double-pulse protocol with pipette solution containing 22 μg ml⁻¹ PKI. The membrane potential was depolarized from -100 to -65 mV, where no macroscopic current was observed, for durations ranging from 10 to 500 ms. Following this conditioning prepulse, a test pulse was elicited to -10 mV for 10 ms. Gsα did not alter the time constant of the onset of inactivation (τ= 38·3 ± 3·6 ms at baseline vs. 38·5 ± 2·7 ms with ISO treatment; n= 4, P= n.s.) (Fig. 4E). The time course from the open state to the inactivated state was determined by measuring the time constant of I_Na decay using a mono-exponential fit. Gsα had no effect on the time constant of I_Na decay at membrane potentials ranging from -50 to +20 mV (Table 1).

Table 1.

Effect of Gsα on the time constant of I_Na decay

Vm(mV)	Controlγ(ms)	Gsαγ(ms)	P
−50	1.44 ± 0.04	1.43 ± 0.03	n. s.
−40	1.14 ± 0.05	1.13 ± 0.05	n. s.
−30	0.82 ± 0.02	0.84 ± 0.03	n. s.
−20	0.66 ± 0.03	0.69 ± 0.03	n. s.
−10	0.53 ± 0.02	0.56 ± 0.03	n. s.
0	0.43 ± 0.02	0.45 ± 0.04	n. s.
+10	0.36 ± 0.02	0.37 ± 0.03	n. s.
+20	0.31 ± 0.02	0.31 ± 0.02	n. s.

Data are expressed as the mean ±s.e.m.; n= 7 for all potentials.

Modulation of cardiac Na⁺ channels by a 16 amino acid peptide of Gsα

A previous study has shown that a 16 amino acid peptide of Gsα containing residues 27-42 of the rat and human Gsα protein was able to increase the amplitude of I_Na in isolated rabbit ventricular myocytes (Matsuda et al. 1992). To further establish the effect of Gsα on rat ventricular myocytes, we tested the effects of a 16 amino acid peptide of native Gsα (amino acids 27-42 of rat brain Gsα; Gsα27-42) using inside-out macropatches. When the membrane potential was depolarized to -40 mV from a holding potential of -100 mV, application of 1 nM Gsα27-42 to the cytoplasmic side of the membrane produced a 32·5 ± 6·8 % (n= 4, P < 0·05vs. baseline) enhancement of I_Na (from 185 ± 34 pA at baseline to 243 ± 31 pA) (Fig. 5A). The time constant of I_Na decay did not change with Gsα27-42 peptide application (1·4 ± 0·03 ms at baseline vs. 1·40 ± 0·04 ms with Gsα27-42 peptide; n= 7, P= n.s.) (Fig. 5B). The Gsα27-42 effect on peak I_Na plotted against time is shown in Fig. 5C. The increase in I_Na amplitude was reversible upon washout of the Gsα27-42 peptide.

Figure 5. Effects of a 16 amino acid peptide of Gsα, Gsα27-42.

Three short amino acid oligopeptides corresponding to the N-terminal (Gsα27-36), middle (Gsα30-39), and C-terminal (Gsα33-42) segments of Gsα27-42 were synthesized (top of figure). A, inside-out macroscopic Na⁺ currents were elicited from a holding potential of -100 mV to a test potential of -40 mV before and after application of 1 nM Gsα27-42 peptide to the cytoplasmic side of the membrane. Each trace was averaged from 10 single episodes. B, normalized I_Na before and after application of Gsα27-42. C, the time course of the Gsα27-42 peptide effect on I_Na.

To examine the effect of the Gsα27-42 peptide on the properties of Na⁺ channel kinetics, we investigated the properties of the activation and inactivation curves, the development of inactivation and the recovery from inactivation using whole-cell recordings with the pipette solution containing 1 μM Gsα27-42 peptide. The peak I_Na increased by 45·2 ± 5·3 % (-232 ± 10 pA pF⁻¹ for control, n= 11, vs. -337 ± 31 pA pF⁻¹ with Gsα27-42 peptide, n= 13, P < 0·05) with the pipette solution containing 1 μM Gsα27-42 peptide (Fig. 6A). However, it did not alter the half-membrane potential (V_½) and slope factor (k) of activation (V_½= -46·2 ± 2·0 mV, k= 3·0 ± 0·3 mV e-fold⁻¹; n= 8, P= n.s. vs. control) and inactivation (V_½= -79·6 ± 0·5 mV, k= 8·1 ± 0·4 mV e-fold⁻¹; n= 8, P= n.s. vs. control) (Fig. 6B and C). Furthermore, 1 μM Gsα27-42 peptide did not change the time constant of development of inactivation (τ= 40·2 ± 0·3 ms; n= 6, P= n.s. vs. control) (Fig. 6D) and recovery from inactivation (τ_f= 24·7 ± 1·4 ms, τ_s= 186·3 ± 16·6 ms; n= 6, P= n.s. vs. control). Thus, in whole-cell experiments, the Gsα27-42 peptide could mimic the effect of native Gsα.

Figure 6. Effect of Gsα27-42 peptide on the kinetics of Na⁺ channels.

The kinetic properties of whole-cell Na⁺ currents were investigated with the pipette solution containing 1 μM Gsα27-42 peptide. Gsα27-42 (1 μM) increased the peak I_Na by 45 % (A) without significantly altering the activation (B) and the inactivation curves (C). D, the time constant of development of inactivation at a conditioning potential of -65 mV, fitted with a mono-exponential equation. Gsα27-42 peptide mimicked the effect of Gsα on the enhancement of I_Na, but did not alter the properties of Na⁺ channel kinetics.

Figure 7 (left panel) shows raw current traces of single Na⁺ channel recordings from an inside-out patch at a holding potential of -100 mV with different test potentials. Single Na⁺ channel current amplitudes plotted against membrane potential were fitted using a linear equation (right panel). The conductance (γ) of the control single Na⁺ channel was 28·1 ± 0·3 pS (n= 6). Application of 1 nM Gsα27-42 did not alter γ (27·9 ± 1·8 pS, n= 2).

Figure 7. Effect of Gsα27-42 peptide on single Na⁺ channel conductance.

Inside-out single Na⁺ currents were recorded from a holding potential of -100 mV to test potentials ranging from -60 to +10 mV (control, left). Single Na⁺ channel current amplitudes were plotted against membrane potential and fitted using a linear regression equation. Conductance (γ) of the control single Na⁺ channel (28·1 ± 0·3 pS, n= 6, •) and that in the presence of 1 nM Gsα27-42 peptide (○) were not different.

To examine the Gsα27-42 peptide effect on the kinetics of single Na⁺ channel current, single Na⁺ channel currents from inside-out patches were evoked from a holding potential of -100 mV to a test potential of -40 mV. In patches containing only one channel, application of 1 nM Gsα27-42 peptide to the cytoplasmic side of the membrane did not alter the single Na⁺ channel mean open time (τ₀= 224 ± 67 μs at baseline vs. 206 ± 83 μs with Gsα27-42 peptide; n= 5, P= n.s.), the mean closed time (τ_c= 272 ± 89 μs at baseline vs. 265 ± 74 μs with Gsα27-42 peptide; n= 5, P= n.s.), or the open probability (P_o) (0·30 ± 0·04 at baseline vs. 0·28 ± 0·05 with Gsα27-42 peptide; n= 4, P= n.s.). In these patches, Gsα27-42 did not increase the ensemble-averaged current (Fig. 8). However, in some patches, application of 1 nM Gsα27-42 peptide increased the ensemble-averaged current. The increases in current were due to an increase in the number of functional Na⁺ channels (Fig. 9). The time course of decay of the ensemble-averaged current (1·24 ± 0·4 ms with Gsα27-42 peptide, n= 3) was not significantly different from the baseline (1·26 ± 0·6 ms; n= 3, P= n.s.). These results suggest that direct modulation of I_Na by the Gsα peptide did not alter the Na⁺ channel kinetics but increased the number of functional cardiac Na⁺ channels.

Figure 8. Effects of Gsα27-42 peptide on the kinetics of single Na⁺ channels.

In inside-out patches containing only one channel, single Na⁺ currents were recorded before (A) and after (B) application of 1 nM Gsα27-42 peptide from a holding potential of -100 mV to a test potential of -40 mV for a duration of 10 ms. The ensemble-averaged current of a single Na⁺ channel patch is shown at the bottom of A and B. The open time and the closed time of Na⁺ channel recordings were fitted using a two-exponential equation in order to separate the basic noise events. The last channel opening was determined in order to separate the closed state from an inactivation state. Gsα27-42 did not change the mean open time (C, control; D, Gsα27-42) or the fast component of the mean closed time (E, control; F, Gsα27-42) and single Na⁺ current amplitude (2·1 pA at -40 mV) in the patch containing only one channel.

Figure 9. Increase in the number of functional Na⁺ channels by Gsα27-42.

In a patch showing only one functional channel at baseline, Gsα27-42 increased the number of functional Na⁺ channels in inside-out patches and the amplitude of the ensemble-averaged single Na⁺ channel current (A, control; B, 1 nM Gsα27-42). The time course of decay of the ensemble-averaged Na⁺ current was not altered. The time constant of the single exponential fit is shown superimposed on the ensemble-averaged current traces at the bottom of A and B. All-points histograms of single Na⁺ channel recordings are shown in C (control) and D (1 nM Gsα27-42). Gsα27-42 did not change the amplitude of the single Na⁺ channel current but activated a second Na⁺ channel in the patch.

Effects of Gsα oligopeptides on the modulation of cardiac Na⁺ channels

Using inside-out macropatches, we tested the effects of three short amino acid oligomers, Gsα27-36, Gsα33-42 and Gsα30-39, representing the N-terminal, C-terminal and middle segment of the active 16 amino acid peptide Gsα27-42, respectively (Fig. 5, top panel). Figure 10A shows enhancement of I_Na following application of 1 μM Gsα33-42 to the cytoplasmic side of the membrane. The time course of I_Na decay was unchanged by Gsα33-42 (Fig. 10B). Figure 10C shows that application of 1 nM Gsα27-36 (2·9 ± 1·4 % change vs. control; n= 4, P= n.s.) and 1 nM Gsα30-39 (-1·3 ± 4·9 % change vs. control; n= 6, P= n.s.) did not change I_Na. Application of the C-terminal segment oligopeptide (Gsα33-42) reproduced the whole-cell and single channel Gsα effect in a dose-dependent manner in all seven cells tested. With the application of 0·2 nM, 0·5 nM, 1 nM and 1 μM Gsα33-42, I_Na increased by 1·3 ± 0·6 % (n= 4, P= n.s.), 8·4 ± 1·4 % (n= 7, P < 0·05), 16·6 ± 3·5 % (n= 7, P < 0·05) and 29·7 ± 0·9 % (n= 4, P < 0·05), respectively. These results suggest that this segment of Gsα (amino acids 33-42) could represent the functional domain responsible for modulating the cardiac Na⁺ channels.

Figure 10. Effect of short amino acid oligopeptides from selected segments of Gsα on cardiac Na⁺ channels.

A, macroscopic Na⁺ currents from inside-out patches were elicited from a holding potential of -100 mV to a test potential of -40 mV before and after application of 1 μM Gsα33-42 to the cytoplasmic side of the membrane. Each trace was averaged from 10 episodes. B, normalized traces of I_Na before and after application of Gsα33-42, shown as superimposed current traces. C, bar graph showing the percentage change from the control peak I_Na after application of various concentrations of Gsα33-42 (0·2 nM, 0·5 nM, 1 nM and 1 μM), 1 nM Gsα27-42, 1 nM Gsα30-39 and 1 nM Gsα27-36. Data represent means ±s.e.m.*P < 0·05vs. control determined by a one-way ANOVA followed by contrast testing.

DISCUSSION

Modulation of I_Na by Gsα via a PKA-independent pathway

In the present study, we show that a significant portion of the enhancement of I_Na by Gsα is by a PKA-independent pathway. In rat ventricular myocytes, Gsα enhanced I_Na by about 30 % without changing the reversal potential or the time constant of I_Na decay. First, with 22 μg ml⁻¹ PKI in the pipette solution (Fig. 2), I_Na decay in the presence of ISO was best fitted by a mono-exponential equation. However, without PKI, ISO increased I_Na by about 68 %, shifted the apparent reversal potential to more positive values (Fig. 1), and significantly increased the rate of I_Na decay, which was best fitted as a two-exponential process. This finding suggests that there are two mechanisms of regulation of I_Na by β-adrenergic receptor stimulation as previously reported (Matsuda et al. 1992; Santana et al. 1998). Furthermore, the enhancement of cardiac I_Na by the Gsα pathway is almost half of the total effect of ISO. This suggests that the PKA-mediated pathway and the direct Gsα pathway make a similar contribution to the enhancement of rat cardiac I_Na. Second, with both PKI and 20 μg ml⁻¹ anti-Gsα IgG in the pipette solution, ISO failed to enhance I_Na. This finding suggests that the enhancement of I_Na by ISO is mediated by Gsα. In addition, with 20 μg ml⁻¹ anti-Gsα IgG alone in the pipette solution, ISO did not change I_Na, suggesting that the anti-Gsα IgG was bound to native Gsα or Gsα failed to dissociate from Gβγ. These results also suggest that Gβγ probably does not play an important role in modulating I_Na in rat ventricular myocytes. This is indicated by the observation that after application of anti-Gsα IgG, no residual G protein-mediated effects were observed. In addition, neither the rat cardiac Na⁺ channel isoform (rH1) nor the human Na⁺ channel isoform (hH1) contains the putative Gβγ-binding motif Q-X-X-E-R (Chen et al. 1995). In contrast, a Gβγ effect was observed in brain Na⁺ channels where the Gβγ-binding motif is found (Ma et al. 1997). Third, we believe that the 16 amino acid Gsα peptide, Gsα27-42, has a functional role in the enhancement of the rat Na⁺ channel current. This is suggested by the observation that we could fully mimic the effect of Gsα on the properties of Na⁺ channels by the application of Gsα27-42 peptide to the cytoplasmic side of the membrane (Fig. 6). Furthermore, with the application of Gsα27-42 peptide to inside-out patches, we could reproducibly and reversibly increase I_Na by about 30 % without any change in the time constant of current decay (Fig. 5).

Modulation of Na⁺ channel kinetic properties by Gsα

Our present study shows that approximately 50 % of the ISO enhancement of I_Na is not accountable by the phosphorylation mechanism. However, the PKI-insensitive portion of the ISO response of I_Na is completely inhibited after application of anti-Gsα IgG that is specific for Gsα (Cai et al. 1993). This suggests that the direct Gsα modulation of I_Na is an important regulatory mechanism. Although Gsα increased the steady-state maximum conductance of Na⁺ channels by about 30-45 %, it did not change the kinetic properties of the Na⁺ channels, including the determinants of steady-state activation and inactivation (Fig. 3). Moreover, we found that Gsα did not change the time course of the development of inactivation from the open state or recovery from the inactivated state (Fig. 4).

Our single channel studies demonstrated that Gsα did not alter the mean open time, the fast component of mean closed time, or the single channel conductance. The enhancement of the ensemble-averaged single Na⁺ channel current by Gsα was due to an increase in the number of functional Na⁺ channels without changing the time course of I_Na decay. These single channel results are consistent with our whole-cell experiments.

The identification of the amino acid residues of Gsα that interact with ion channels is still uncertain. However, there is evidence suggesting that the adenylyl cyclase modulation domain of Gsα is located near the C-terminus of Gsα (Osawa et al. 1990; Itoh & Gilman, 1991; Novotny et al. 1996). We tested a 16 amino acid peptide near the N-terminus of Gsα (Gsα27-42) and found that the 10 amino acid oligomer with the sequence D-K-Q-V-Y-R-A-T-H-R (Gsα33-42) was effective in increasing I_Na in a dose-dependent manner (Fig. 10). In this study, we confirm that both the PKA-mediated pathway and direct G protein activation contribute significantly to the enhancement of cardiac Na⁺ channel activity. We also show that in addition to the C-terminus of native Gsα, a specific domain near the N-terminus of Gsα can function to stimulate Na⁺ channels. Details of the interaction of Gsα with cardiac Na⁺ channels requires further investigation.

G protein regulation of the cardiac Na⁺ channels

There is evidence that Na⁺ channels in the heart are modulated by β-adrenergic receptor stimulation via dual G protein pathways: indirectly through G protein-regulated second messenger cascades and directly by a membrane-delimited pathway (Schubert et al. 1989; Matsuda et al. 1992; Ono et al. 1993; Cukierman, 1996; Frohnwieser et al. 1997). Although information on the exact structure and biochemistry of PKA-mediated phosphorylation of the cardiac Na⁺ channels is not available, the α subunits of both the rat and human cardiac Na⁺ channels (SkM2 and hH1, respectively) contain multiple PKA consensus sites (Schreibmayer et al. 1994; Fozzard & Hanck, 1996). In addition, incorporation of phosphate into rat heart Na⁺ channels via PKA has been reported (Cohen & Levitt, 1993). However, conflicting results exist regarding the direction of the I_Na response to PKA-mediated and direct Gsα-mediated effects. Schubert et al. (1989) reported that PKA and Gsα inhibited neonatal rat cardiac Na⁺ channels, whereas Matsuda et al. (1992) demonstrated that PKA and Gsα enhanced rabbit cardiac Na⁺ channels. The aetiology for such a discrepancy is still unresolved. Experiments using heterologous expression systems have demonstrated that PKA stimulation resulted in marked and significant increases in rat and human cardiac Na⁺ channels (Schreibmayer et al. 1994; Frohnwieser et al. 1997), suggesting that the behaviour of the human and rat heart Na⁺ channels in response to β-adrenergic receptor stimulation may resemble that of the rabbit heart rather than that of the neonatal rat heart. In adult rat ventricular myocytes, we found that the behaviour of Na⁺ channels in response to β-adrenergic receptor stimulation was opposite to that of the neonatal rat heart. It is very difficult to ascribe these opposite effects to species differences, but it may suggest that there exists a developmental difference in the regulation of β-adrenergic receptor stimulation. However, Ono et al. (1993) described voltage-dependent conductance and channel availability shifts in the hyperpolarizing direction by CPT-cAMP in cell-attached recordings. These shifts may account for some of the disparate findings previously reported. Recent reports showing that the β-adrenergic receptor could switch its coupling to G proteins (Daaka et al. 1997) and that the rat cardiac Na⁺ channel ionic selectivity can be transformed into a ‘slip mode conductance’ (Santana et al. 1998) suggest that β-adrenergic modulation of the cardiac Na⁺ channels could be much more complex than initially contemplated.