INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

SINOATRIAL NODAL CELLS of the heart are characterized by slow and spontaneous diastolic depolarization, which is the basis of their pacemaking activity (11). The inwardly rectifying potassium channel (K_ACh) in the pacemaker cells of the sinoatrial node is an important ionic mechanism that underlies modulation of the chronotropic properties in the heart under vagal stimulation. ACh released on vagal stimulation slows the heart rate through activation of muscarinic receptors and subsequent opening of K_ACh. This process is in part mediated by hyperpolarization of pacemaker cells due to increased potassium conductance of the membrane. K_ACh is of clinical significance, because it plays an important part in determining the atrial resting membrane potential and the shape of the cardiac action potential during the final phase of repolarization. K_ACh opens primarily near the resting potential but closes at depolarized membrane potentials. Impaired activation can therefore decrease the excitation threshold and lead to premature generation of the action potential. In addition, it is the major effector of parasympathetic signal transduction in atrial myocytes.

I_KACh is present in the sinoatrial node, atria, atrioventricular node, and possibly Purkinje fibers of the mammalian heart (27). The cloning of GIRK1 and GIRK4, which constitute cardiac inward rectifier potassium channels by forming heterotetramers, has allowed a more detailed understanding of how heterotrimeric G proteins activate K_ACh (16). G-subunits that are released from pertussis toxin-sensitive G proteins after stimulation of M₂ muscarinic receptors (or adenosine-1 receptors) directly activate K_ACh via a membrane-delimited mechanism (1, 15, 18).

The aim of the present study was to investigate the physiological role of G protein -subunits for heart rate control and arrhythmia vulnerability in vivo using electrocardiographic telemetry and intracardiac electrophysiological stimulation. To that end, we generated a transgenic mouse model with a reduced amount of G-subunits in cardiocyte membranes due to cardiac-specific overexpression of nonprenylated G₂-subunits.

G protein -subunits are attached to the plasma membrane by a prenyl group on the carboxyl terminus of all G-subunits (5). In a multistep series of posttranslational modifications, the G-subunits are either farnesylated or geranylgeranylated at position 4 from the carboxyl terminus, followed by proteolytic removal of the last three amino acids and methyl esterification of the resultant terminal carboxyl group. Isoprenylation can be blocked by mutating the cysteine in the COOH-terminal CXXR motif to serine. Absence of the lipid modification does not prevent the formation of G complexes, but without this lipid modification, G is no longer membrane bound and unable to activate effectors (10).

We reasoned that overexpression of nonisoprenylated G-subunits in hearts of transgenic mice would diminish the amount of active, membrane-bound G because nonisoprenylated G should compete with endogenous G for assembly with G. We chose G₂[C68S] for these studies because it assembles well (35) with the G-subunits expressed in the heart: G₁, G₂, and G₄ (9, 30, 32). G₂, itself, is not found in the heart (9), but this is of no consequence, because the only function of its mutated, nonprenylated form is to trap as many G-isoforms as possible.

We hypothesized that a reduction in the amount of endogenous G-subunits in cardiac membranes would blunt the response of the intact animal to the negative chronotropic effects of the muscarinic receptor agonist carbachol and could potentially be protective against vagally mediated atrial arrhythmias.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Generation of transgenic mice. The cDNA for bovine G₂ was modified to carry an NH₂-terminal hemagglutinin (HA)-epitope and a C68S mutation by PCR using the primer pair 5'-GTCGACGCCGCCATGGTATCCTATCCATAC-3' and 5'-AAGCTTTTAAAGGATAGCAGAGAAAAAC-3' and a previously described HA₂ cDNA as template (21). The PCR product HA₂[C68S] was ligated into the T-Vector (Promega), and the DNA sequence was verified. The 0.28-kb SalI-HindIII HA₂[C68S] cDNA fragment was ligated into the corresponding sites of a pGEM-9Zf vector (kindly provided by Dr. W. J. Koch) containing the murine -myosin heavy chain promoter (5.5 kb) and a simian virus 40 intron/polyadenylation signal. A 6.63-kb linear cDNA fragment was released with SacI and NotI. Transgenic FVB mice were generated by the transgenic mouse facility of Harvard Medical School and were identified by Southern blot and PCR with the use of a cDNA probe or primers specific for the transgene.

Cardiocyte isolation. Hearts from wild-type and transgenic mice of different ages were excised and Langendorff perfused with Ca²⁺-free Tyrode's solution containing (in mM) 140 NaCl, 5.4 KCl, 1 MgCl₂, 1 Na₂HPO₄, 5 HEPES, and 10 glucose (pH 7.3). After 3-5 min of perfusion with noncirculating Tyrode's solution, the heart was perfused for 15 min with recirculating Ca²⁺-free Tyrode's solution containing collagenase II (0.6 mg/ml, Worthington), protease XXIV (0.1 mg/ml, Sigma), bovine serum albumin (1 mg/ml), and fetal calf serum (2%). The flaccid heart was removed from the cannula, the atria were trimmed away, and the ventricular tissue was minced in Tyrode's solution containing 0.2 mM Ca²⁺. The majority of cells were released by gentle pipetting. The supernatant containing the isolated cells was centrifuged at 50 g for 3 min, washed twice with Tyrode's solution (0.2 mM Ca²⁺) and once with phosphate-buffered saline, and then counted. The cells were kept in a low-Ca²⁺ solution because the two major cardiac adenylyl cyclase (AC) isoforms (AC V and AC VI) are sensitive to direct inhibition through Ca²⁺ (28). The final cell pellet was snap-frozen and stored at 80°C until further use.

Membrane and cytosol preparation. Atrial and ventricular cardiac tissue were homogenized as previously described (20). Isolated cardiocytes from one heart (1-2 × 10⁶ cells) were thawed and homogenized in 500 µl of ice-cold Tris buffer [50 mM Tris · HCl, pH 7.6, 1 mM EDTA, and 1 mM dithiothreitol plus proteinase inhibitors (3 mM benzamidine and 1 µg/ml of each soya and lima bean trypsin inhibitor and leupeptin)] with a hand-held tissue tearer (5 bursts of 10 s at 8-10,000 rpm). To separate the cytosolic and membrane fractions, both tissue and myocyte samples were centrifuged at 100,000 g for 30 min at 4°C. Protein was measured using the Bradford Microassay (Bio-Rad) with bovine serum albumin as standard (4).

Western blot analysis. Equal amounts of membrane and cytosolic protein were separated on 10% (G-subunits), 12% (G-subunits), or discontinuous 12%/17% gradient (G- and G-subunits) SDS-PAGE and transferred to nitrocellulose membranes. Immunoblots were performed as previously described (20) using antibodies against G_s (sc-383, 1:1,000, Santa Cruz Biotechnologies), G_i (AS/7, 1:1,000, NEN Life Science Products), G_q/11 (C-19, 1:1,000, Santa Cruz Biotechnologies), G (_common, 1:1,000, Upstate Biotechnologies), G₂ (₂EDPL, 1:500, gift from Dr. W. Simmonds), and the HA epitope (12CA5, 1:1,000, BAbCO). Band intensity was determined by scanning and computerized densitometry using the NIH Image 1.61 software.

Coimmunoprecipitation experiments. Ventricular cytosolic protein (500 µg) was incubated with a HA antibody (12CA5, 2 µl, BAbCo) in RIPA buffer containing 50 mM Tris · HCl, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% SDS plus proteinase inhibitors (3 mM benzamidine and 1 µg/ml of each soya and lima bean trypsin inhibitor and leupeptin) at 4°C overnight. Protein A Sepharose (35 µl of a 50% slurry of CNBr-activated Sepharose Cl-4B, Sigma) was added and tumbled for another 2 h. Samples were pelleted and washed three times with 1 ml of RIPA buffer followed by one wash in detergent-free buffer. The final pellets were resuspended in Laemmli sample buffer and heated to 95°C for 5 min. The supernatants, along with 100 µg of cytosolic starting material, were loaded onto a discontinous 12-17% gradient SDS-PAGE gel and transferred to nitrocellulose membrane. The upper and lower parts of each membrane were probed by Western blotting (see Western blot analysis) with antibodies against G and G₂, respectively.

Adenylyl cyclase assay. Ventricular cardiomyocyte membranes (15-20 µg protein/assay) from 9-wk-old HA₂-79h and wild-type mice were incubated for 20 min at 37°C, and the AC activity was determined as described previously (22). In brief, the 50-µl reaction cocktail contained 1 mM ATP (1 µCi [-³²P]ATP/tube), 5 mM Mg acetate, 0.1 mM cAMP, 1 mM dithiothreitol, 0.1 mg/ml bovine serum albumin, 1 mM isobutylmethylxanthine, 25 mM Tris acetate, pH 7.6, 5 mM creatine phosphate, 50 U/ml creatine kinase, and 10,000 cpm [³H]cAMP. For stimulation of AC activity, 100 µM GTPS, 10 µM isoproterenol, 100 µM forskolin, or a combination thereof, were added. cAMP was purified by precipitation of other nucleotides with ZnCO₃ and on alumina columns and quantitated by scintillation counting. The recovery was 70-80%.

Animals used for telemetry and electrophysiological studies. All experiments were performed on HA₂-79h mice (n = 21) and sex- and age-matched wild-type mice (n = 38) from the same inbred strain (FVB). The mean age was 14 wk (range 12-16 wk), and the average weight was 30.8 ± 1.8 g. The body weight was similar in HA₂-79h (30.2 ± 0.6 g) and wild-type mice (31.4 ± 2.5 g). Mice were housed two to four per cage at 24°C in a facility with 12:12-h light-dark cycles and allowed free access to water and food.

Heart rate and heart rate variability. The techniques used for recording ambulatory long-term electrocardiograms (ECGs) by telemetry and for the analysis of heart rate and heart rate variability (HRV) are described in detail elsewhere (8). Carbamylcholine chloride (carbachol, 0.5 mg/kg ip) was administered to evaluate the effect of muscarinic receptor stimulation on heart rate regulation. Atropine (0.5 mg/kg ip) and propranolol (1 mg/kg ip) were injected 5 min before carbachol to block the parasympathetic and sympathetic autonomic responses, respectively. Both antagonists were administered for combined autonomic blockade to assess "intrinsic heart rate." Baroreflex-mediated cardioinhibitory responses were tested in wild-type and transgenic mice that underwent pressure challenge with phenylephrine hydrochloride (3 mg/kg ip) after -adrenergic blockade with propranolol (1 mg/kg ip). ECG recordings and HRV analysis were performed 5-10 min after the drug administration to allow the heart rate to stabilize and preclude any direct effects of the intraperitoneal injection on ECG and HRV parameters (8).

Electrophysiology studies. Surface six-lead ECGs were obtained from anesthetized mice (ketamine hydrochloride and pentobarbital; 0.033 mg/g each) by placing subcutaneous 27-gauge electrodes in each limb. For the endocardial approach, an octapolar mouse 1.7-Fr electrophysiology catheter (CIBer mouse electrophysiology catheter, NuMED; Hopkinton, NY) with precise interelectrode distances was advanced from the right internal jugular vein, exposed by cutdown, through the right atrium to the right ventricle. The proximal electrodes paced and recorded from the atrium while the distal electrodes allowed stimulation and signal acquisition from the right ventricle.

Statistical analysis. Data are reported as means ± SD or SE for mice (n) as indicated. Statistical analysis was performed using paired and unpaired two-tailed Student's t-test, ANOVA with Scheffé's subgroup testing where appropriate, and analysis of interobserver variability. A P value of <0.05 was considered statistically significant.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Generation of transgenic mice. Two independent heterozygous mouse lines (HA₂-79 and HA₂-83) with cardiac-specific expression of nonisoprenylated, HA-tagged ₂ (HA₂[C68S]) were established. HA₂-79 mice were also bred to homozygousity (HA₂-79h). All mice were viable, had a normal lifespan, and showed no gross apparent phenotype compared with age-matched wild-type mice.

Expression of protein encoded by the transgene. G₂ is normally not expressed in the heart (9). Accordingly, we did not detect any endogenous G₂-subunits in the membrane (see Membrane/cytosol distribution and expression of G protein -subunits) or cytosol (Fig. 1A) of atrial and ventricular tissue from wild-type mice. The protein recognized by the G₂-specific antibody in the cytosol from wild-type (and transgenic) atria reflects cross-reactivity with a protein of unidentified origin. The protein encoded by the transgene, HA₂[C68S], was detected in the cytosol of both atria and ventricles from transgenic mice only (Fig. 1A). In the absence of endogenous G₂, the protein recognized by either the G₂- or the HA-specific antibody (Fig. 1A, top and bottom, respectively) represents HA₂[C68S]. We used the G₂-specific antibody in subsequent Western blots because it yielded a stronger signal than the antibody against the HA epitope. As expected for unprenylated G, no HA₂[C68S] was found in the membrane (see below, Fig. 2A).

View larger version (36K): [in this window] [in a new window]   Fig. 1.   HA2[C68S] expression in atria and ventricles of transgenic (T) mice. A: representative Western blots of cytosolic proteins (50 µg/lane) from atria (left) and ventricles (right) from wild-type mice (WT) and T mice (line HA2-79) of the indicated ages. Blots were probed with a hemagglutinin (HA)-specific antibody (12CA5, bottom), stripped, and reprobed with a G2-specific antibody (2EPDL, top). Position of HA2[C68S] is marked by arrows. Lower-molecular-weight protein recognized by the G2 antibody in the atria represents cross-reactivity with a protein of unidentified origin. B: time course of HA2[C68S] expression in the cytosol of ventricles from HA2-79 (n = 3 each) and HA2-79h mice (n = 3 each). Data are expressed in percentage of 3-wk-old HA2-79 and represent means ± SE from 3 independent experiments. #P < 0.05 vs. 3-wk time point for each genotype; *P < 0.05 for HA2-79h vs. age-matched HA2-79. C: representative Western blot showing coimmunoprecipitation of HA2[C68S] (bottom, probed with the 2EPDL antibody) and endogenous G (top, probed with the common antibody) in ventricular cytosol (500 µg/lane) from 9-wk-old WT and T mice after immunoprecipitation (IP) through the HA antibody. Shown to left is the starting material (SM, 100 µg/lane).

View larger version (43K): [in this window] [in a new window]   Fig. 2.   Effect of HA2[C68S] on G protein expression and distribution in cardiocytes of T mice. A: representative Western blots of cytosolic (35 µg/lane) and membrane (10 µg/lane) proteins from ventricular cardiocytes from 9-wk-old WT and T mice that were immunostained with common (top) and 2EPDL antibodies (bottom). Exposure times varied between lines. For a quantitative assessment of transgene expression and reduction in membrane-bound G, see Fig. 1B and 2B, respectively. B: expression of G (10 µg/lane), Gi (30 µg/lane), Gs (20 µg/lane), and Gq/11 (30 µg/lane) in ventricular myocyte membranes from HA2-79 and WT. Representative Western blots are shown as insets. In each group and at each time point (4, 15, and 22 wk) cardiocytes from 3 hearts were analyzed. All data are normalized to 4-wk-old WT and represent means ± SE from 3 independent experiments. #P < 0.05 vs. 4-wk time point for each genotype; *P < 0.05 for HA2-79 vs. age-matched WT controls. C: expression of G (10 µg/lane), Gi (30 µg/lane), Gs (20 µg/lane), and Gq/11(30 µg/lane) in ventricular myocyte membranes from 9-wk-old HA2-79h mice compared with age-matched WT controls.

Membrane-cytosol distribution and expression of G protein -subunits. The effect of unprenylated HA₂[C68S] on the membrane-cytosol distribution of G protein was tested in Western blots using a _common-antibody, which recognizes G₁, G₂, and G₄ (as confirmed with Sf9 cell extracts expressing different G isoforms; data not shown). No change in G distribution was detectable in atrial and ventricular tissue (data not shown), presumably due to the considerable number of nonmyocytes in cardiac tissue, which do not express HA₂[C68S] but contribute to the overall G pool. In contrast, the amount of membrane-bound G was greatly reduced in isolated cardiocytes from ventricles of mice from all lines, whereas, conversely, the amount of cytosolic G was increased (Fig. 2A). The limited number of atrial myocytes precluded us from carrying out a similar analysis in atria. Figure 2A also illustrates that HA₂[C68S] expression was restricted to the cytosol.

Expression of G protein -subunits. We examined whether the amount of G protein -subunits is altered in cardiocyte membranes from HA₂-79 mice (Fig. 2B). Similar results were obtained in HA₂-79h mice (see below) and HA₂-83 mice (data not shown). In wild-type mice, the amount of G_i (assessed with the AS/7 antibody, which recognizes all three G_i subtypes) increased twofold by 22 wk of age, whereas G_s and G_q/11 remained largely unchanged over time. In transgenic mice, the amount of both G_s isoforms was decreased compared with the wild-type mice over the entire period. G_i protein levels were slightly, but significantly, decreased in 22-wk-old transgenic mice. There was no change in the amount of G_q/11 at any time.

1

1

1

1

1

1

1

1

1

1

Heart rate and ECG intervals in conscious, unrestrained mice. The following experiments were performed on 12- to 16-wk-old HA₂-79h mice, which showed the most pronounced reduction of membrane-bound G among the different lines and featured comparable changes in the G protein -subunit expression (Fig. 2C). Age- and sex-matched wild-type mice were used as controls. ECG parameters obtained from telemetric recordings showed no difference between HA₂-79h and age-matched wild-type mice at baseline (Table 1). Within 2 min after carbachol injection, wild-type mice developed profound sinus bradycardia. The prolongation of the average sinus cycle length (and corresponding reduction in heart rate) was significantly blunted in the HA₂-79h mice (SCL increase by 99% in HA₂-79h vs. 274% in wild types, P < 0.05). There were no differences in the effects of carbachol on P-R interval, QRS duration, or corrected Q-T interval between groups.

View larger version (35K): [in this window] [in a new window]   Fig. 3.   Heart rate under basal conditions (A) and in response to muscarinic stimulation (B) in conscious, unrestrained WT and HA2-79h mice (T). Beat-to-beat heart rate was obtained from telemetered electrocardiogram (ECG) recordings. B: carbachol (0.5 mg/kg) was injected intraperitoneally at time 0. Data are means ± SE for each time point from WT mice (n = 19) and HA2-79h mice (n = 16).

Heart rate variability. The results of time-domain and frequency-domain measures of HRV are summarized in Table 2. Under baseline conditions, both time-domain and frequency-domain HRV parameters tended to be decreased in HA₂-79h mice compared with wild-type controls. Whereas some changes did not reach statistical significance, two time-domain indexes (standard deviation of the average R-R interval and coefficient of variance) and three frequency-domain indexes (low-frequency power, normalized frequency power, low frequency-to-high frequency power ratio) were significantly reduced. After muscarinic stimulation, wild-type mice exhibited a dramatic increase in all time- and frequency-domain parameters. In contrast, the effects of carbachol were profoundly blunted in HA₂-79h. Thus beat-to-beat fluctuations in the heart rate were reduced in transgenic mice compared with controls, particularly after carbachol challenge.

Effects of propranolol, atropine, and isoproterenol on heart rate. In a subgroup of mice (n = 5 each), we performed parasympathetic and sympathetic autonomic blockade (using atropine and propranolol, respectively) and autonomic stimulation (using isoproterenol) and compared the heart rate response with the baseline state. Propranolol caused a comparable decrease in heart rate in HA₂-79h mice (157 ± 24 beats/min) and wild-type controls (179 ± 21 beats/min). After atropine administration, the heart rate was indistinguishable between both groups (HA₂-79h: 14 ± 3 beats/min; wild types: 13 ± 5 beats/min). The intrinsic heart rate following dual autonomic blockade was also similar (HA₂-79h: 552 ± 39 beats/min; wild types: 517 ± 50 beats/min), as was the response to isoproterenol (HA₂-79h: 748 ± 25 beats/min, wild types: 722 ± 32 beats/min). Thus there was no difference between transgenic and wild-type mice in their response to autonomic blockade and sympathetic stimulation.

Cardiac electrophysiology. Table 3 summarizes the results of the endocardial electrophysiological study in the absence and presence of carbachol. Sinus node recovery times were slightly decreased in HA₂-79h mice compared with wild-type mice under baseline conditions, but the difference did not reach statistical significance. After muscarinic stimulation, sinus node recovery times were significantly shorter in the transgenic mice (see also Fig. 4, top). In wild-type mice, the SNRT at at 200 pacing lengths (SNRT₂₀₀) increased by 80% compared with 37% in the transgenic mice, SNRT₁₅₀ by 74% compared with 31%, and SNRT₁₀₀ by 61% compared with 39%. There were no discernible differences in the modulation of atrioventricular conduction properties, atrioventricular nodal, or ventricular effective refractory periods between HA₂-79h and wild-type mice either under baseline conditions or after carbachol administration.

View larger version (27K): [in this window] [in a new window]   Fig. 4.   Sinus node function and atrial arrhythmias in WT and HA2-79h mice. Top: illustration of sinus node recovery time (SNRT) following right atrial pacing (S1) at cycle length of 150 ms in WT (left) and transgenic (right) mice treated with carbachol (0.05 mg/kg). Surface ECG lead I is displayed, followed by right ventricular and right atrial intracardiac electrograms (IECG, in mV). After 15 s of atrial pacing, the WT mouse SNRT is 752 ms compared with 422 ms in the transgenic mouse. Bottom: programmed right atrial (RA) burst stimulation following carbachol administration (0.05 mg/kg) provokes a rapid sustained atrial tachycardia in a WT mouse, whereas the same protocol did not induce atrial arrhythmia in a transgenic mouse. Surface ECG lead I is displayed on top followed by the right ventricular and right atrial intracardiac electrograms (in mV). Overall, 73% of WT mice, compared with only 23% of transgenic mice had inducible atrial tachyarrhythmias.

Arrhythmia inducibility. Under baseline experimental conditions, no atrial or ventricular arrhythmias were inducible with programmed atrial and ventricular stimulation techniques or burst-pacing protocols in either wild-type or transgenic mice. As illustrated in Fig. 4, bottom, following pretreatment with carbachol, sustained atrial fibrillation with a mean duration of 79 ± 147 s was reproducibly inducible with programmed extrastimuli or burst-pacing protocols in 19 of 26 wild-type mice (73%). In contrast, in only 3 of 13 transgenic mice (23%) was atrial fibrillation with a mean duration of 23 ± 21 s inducible during the electrophysiological study (P < 0.05). In a subgroup of mice that were reproducibly inducible for arrhythmias, reinduction of atrial fibrillation was attempted after administration of atropine. None of these mice was any longer inducible. We did not observe supraventricular tachycardias, ventricular arrhythmias, or sudden cardiac death in any mouse under any condition tested.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Transgenic mouse model. Genetically modified mouse models are often utilized to examine the physiological role of signaling components under physiological conditions in the intact animal. We choose a transgenic approach to create a mouse model in which we could investigate the functional role of G-subunits for the regulation of heart rate control, cardiac conduction, and arrhythmogenesis. In the present study, we demonstrate that it is possible to significantly reduce the amount of endogenous G-subunits in cardiocyte membranes by trapping them in the cytosol with overexpressed nonprenylated HA₂[C68S]. As a result, less G is available to form dimers with endogenous G-subunits, causing a reduction in the amount of functional G-dimers available for signal transduction. In contrast to targeted deletion, this approach causes a reduction in the amount of membrane-bound G protein that is not limited to a particular isoform, because the G-subunits used as a trap, G₂, form dimers with all G-isoforms expressed in the heart equally well (35). This was critical for the present study because it has been shown that G-dimers containing G₁-G₄ bind to and activate K_ACh current equally well (17, 34).

Expression of G protein -subunits. The reduction in the amount of membrane-bound G protein in transgenic mice was accompanied by a reduction in the amount of G_s (and to a lesser degree G_i). This observation may be indicative of coordination between G and G levels. It has been shown in brains of mice with a targeted deletion of G_o that the amount of G protein is reduced (presumably due to enhanced degradation of "excess" G), so that it matches the amount of remaining G subunits (22). It is conceivable that the amount of G_s (and G_i) in mice expressing HA₂[C68S] is reduced because less prenylated G is available for heterotrimer formation. The affinity of nonprenylated G for G-subunits is, in contrast to prenylated G, very low (10).

Adenylyl cyclase activity. The response of AC to stimulation with isoproterenol (which activates G_s indirectly through -adrenoceptor stimulation), GTPS (which activates G proteins directly), and forskolin [which acts directly on the catalytic unit of AC but in synergy with G_s (2)] was blunted in the transgenic mice. This could at least partly be due to the decreased amount of G_s protein. However, we cannot exclude the possibility that the protein expression of the AC itself is also reduced in the transgenic mice, potentially contributing to the observed results. In contrast, the reduction in functional G in the transgenic hearts is unlikely to play a role because the predominant cardiac AC isoforms (AC V and AC VI) are not modulated by G (28).

Modulation of electrophysiological characteristics. The main electrophysiological finding of our investigation was that G is an important component in the regulation of pacemaker activity and in the parasympathetic branch of signal transduction. Although evidence for a role of K_ACh in vagal heart rate regulation has been shown in mice with targeted disruption of the GIRK 4 gene (33), this report addresses the contribution of its upstream modulator G to in vivo cardiac electrophysiological function and chronotropic heart rate regulation.

Heart rate variability. HRV analysis was performed to interrogate abnormalities in autonomic regulation of cardiac rhythm that are not necessarily reflected in changes in mean heart rate. As the frequency components of the heart rate spectra are affected by both sympathetic and parasympathetic nervous systems inputs, HRV analysis allows quantification of their respective contributions. We found that the direct, membrane-delimited activation of K_ACh is important for beat-to-beat modulation of heart rate by the autonomic nervous system, which functions on a second time scale, in contrast to the involvement of a second messenger system (cAMP) operating on a tens-of-seconds time scale (25).

Atrium and sinus node modulation. Transgenic mice exhibited shorter sinus node recovery times than wild-type mice after carbachol administration. This suggests that G-subunits play a role in sinus node automaticity and pacemaking function. Two potentially overlapping mechanisms could contribute to this effect. First, deficiency in functional G is likely to cause malfunction of the repolarizing current K_ACh, leading to impairment of K⁺ efflux, hyperpolarization, and inadequate decrease in phase 4 of the action potential. Thus one important vagal bradycardia mechanism could be impaired. In the sinoatrial node, most of the basal K⁺ conductance is due to K_ACh, and in pacemaker cells any changes in its activity can alter excitability and heart rate (11). Second, in addition to K_ACh, ACh modulates the pacemaker current (I_f) in sinoatrial nodal cells (25, 36). There are some important differences between I_f and K_ACh modulation. Heart rate control by inhibition of I_f is achieved with moderate vagal stimulation or low doses of ACh, whereas stronger vagal activity or a higher concentration of ACh is required for greater bradycardia due to activation of K_ACh (7). Furthermore, I_f regulation by muscarinic receptors is mediated by a second messenger (cAMP) and not by direct activation through G (25). Our data implicate alterations in K_ACh regulation as the major underlying mechanism, because diminution in I_f should have caused abnormalities in heart rate in transgenic mice under baseline conditions, (i.e., before muscarinic stimulation) and, in addition, prevention of vagal-provoked atrial fibrillation would have been unlikely. However, we cannot rule out secondary effects due to I_f blockade and/or compensatory I_f changes.

Arrhythmia vulnerability. Although one could have anticipated an increased propensity to atrial arrhythmias in the transgenic mice due to malfunction of a key regulatory conductance in pacemaker cells, there were no ambient arrhythmias detected during ambulatory ECG recordings without or with pharmacological manipulation. In contrast, disruption in the parasympathetic branch of signal transduction in HA₂-79h mice turned out to be protective for vagally mediated pacing-inducible atrial fibrillation in the presence of carbachol, because the incidence of this arrhythmia in both frequency and duration was markedly reduced in transgenic mice compared with their wild-type counterparts. Together with our recent finding that GIRK4 knockout mice are resistant to carbachol-mediated, pacing-induced atrial fibrillation (14), the present study demonstrates the crucial role G plays in vivo for K_ACh-mediated, parasympathetic regulation of atrial arrhythmic vulnerability.

Conclusions and future implications. The present study demonstrates the feasibility of creating an in vivo mouse model with a deficiency in membrane-bound G-dimers by transgenic overexpression of nonprenylated G. We utilized this model to examine the physiological role of G for heart rate regulation, sinus node activity, and arrhythmia vulnerability, particularly under vagal stimulation. Our data indicate that G is a physiologically important component in the parasympathetic branch of cardiac signal transduction and critical in heart rate regulation. Reduction of functional G produced abnormalities of beat-to-beat-control of cardiac activity and of spontaneous depolarization within the sinus node, which presumably reflects a malfunction of the atrial K_ACh channel. We thereby provide direct in vivo electrophysiological evidence that the integrity of G-mediated signaling is vital to cardiac pacemaker activity in the sinoatrial node, effective heart rate regulation, cardiac automaticity, and arrhythmia susceptibility.