ATP-sensitive potassium channels in the sinoatrial node contribute to heart rate control and adaptation to hypoxia

ATP-sensitive potassium channels (KATP) contribute to membrane currents in many tissues, are responsive to intracellular metabolism, and open as ATP falls and ADP rises. KATP channels are widely distributed in tissues and are prominently expressed in the heart. They have generally been observed in ventricular tissue, but they are also expressed in the atria and conduction tissues. In this study, we focused on the contribution and role of the inwardly rectifying KATP channel subunit, Kir6.1, in the sinoatrial node (SAN). To develop a murine, conduction-specific Kir6.1 KO model, we selectively deleted Kir6.1 in the conduction system in adult mice (cKO). Electrophysiological data in single SAN cells indicated that Kir6.1 underlies a KATP current in a significant proportion of cells and influences early repolarization during pacemaking, resulting in prolonged cycle length. Implanted telemetry probes to measure heart rate and electrocardiographic characteristics revealed that the cKO mice have a slow heart rate, with episodes of sinus arrest in some mice. The PR interval (time between the onset of the P wave to the beginning of QRS complex) was increased, suggesting effects on the atrioventricular node. Ex vivo studies of whole heart or dissected heart regions disclosed impaired adaptive responses of the SAN to hypoxia, and this may have had long-term pathological consequences in the cKO mice. In conclusion, Kir6.1-containing KATP channels in the SAN have a role in excitability, heart rate control, and the electrophysiological adaptation of the SAN to hypoxia.

ATP-sensitive potassium channels (K ATP ) 2 contribute to membrane currents in a number of tissues. They are opened by falling cellular ATP and/or rising ADP levels and are important metabolic sensors (1). Substantial K ATP currents were first described in cardiomyocytes and then subsequently in pancreatic ␤ cells, skeletal muscle, smooth muscle cells, and neurons (2,3). K ATP channels are critical in mediating a number of physiological processes. For example, they are central in stimulus secretion coupling underlying insulin release from the endocrine pancreas (4). At the molecular level the channel is a heterooctamer of four sulfonylurea receptors (SUR1, SUR2A) and four pore-forming inwardly rectifying Kir6.0 subunits (Kir6.1, Kir6.2) (5-7). Recent ground breaking structural studies using cryo-EM have suggested models for how the channels are able to sense nucleotides at the molecular level (8,9).
Ironically, despite the initial description of the K ATP current in cardiomyocytes and the substantial current density in ventricular cells, it is less clear as to the physiological role this channel plays in cardiac function (1). Mice with global genetic deletion of Kir6.2, which is thought to underlie ventricular cardiac K ATP currents, have impaired exercise tolerance and cardiac performance under sympathetic stimulation, suggesting a role in adaptation of the myocyte to the metabolic demands of exercise (10). Furthermore, K ATP currents are present in other chambers in the heart, including the atria and conduction system (11,12). Indeed these currents may have unique subunit composition and properties.
The sinoatrial node (SAN) is the dominant pacemaker in the mammalian heart and has unique electrophysiological characteristics underlying this role. There have been few studies on K ATP currents in these cells, but it is clear they exist and may at least partially be based upon Kir6.2 subunit expression (13). However, Kir6.1-containing channels may also have a role. Global Kir6.1 knockout mice die prematurely from SAN failure and heart block. Initially this was attributed to loss of K ATP channels in coronary vascular smooth muscle leading to vasospasm and cardiac ischemia (14). However, we have generated mice in which it is possible to selectively delete Kir6.1 in various tissues and at various times using cre‫گ‬loxP technology (15,16). Mice with deletion in vascular smooth muscle or endothelium do not recapitulate this phenotype, suggesting there may be a direct role for Kir6.1-containing K ATP channels in the conduction system (15,16). In this study we explore this idea by using cre‫گ‬loxP technology using our previously described mice and crossing them with a Cre recombinase murine line that allows  cro ARTICLE selective deletion in the cardiac conduction system after the administration of tamoxifen (17).

Generation of the conduction-specific Kir6.1 KO mouse
HCN4creϩ Kir6.1(flx/flx) (cKO) mice were generated by crossing Kir6.1(flx/flx) mice with a conduction system-specific Cre into which the tamoxifen-inducible Cre-ER T2 had been "knocked in" into the HCN4 locus (see "Experimental Procedures" and Fig. 1A) (15)(16)(17). PCR confirms the presence of cre recombinase (900 bp) at the genomic level in ear tissue from HCN4creϩ Kir6.1(flx/flx) mice but not in that of control animals (WT) (Fig. 1B). The "floxed" allele is recognized by a PCRspanning exon 1, the 5Ј loxP site, and the recombined flippase recognition target (FRT) site and associated sequence. In the "floxed" mouse this leads to a 600-bp product, whereas it generates a 474-bp product in the WT allele (which does not include these additional sequence elements) (Fig. 1B). The occurrence of the recombination event in HCN4creϩ Kir6.1(flx/flx) mice selectively in the SAN following tamoxifen induction was confirmed using PCR on genomic DNA isolated from SAN, atria, and ventricles (Fig. 1B).
To confirm the reduced expression of Kir6.1 (Kcnj8) in the SAN of cKO mice, quantitative real-time PCR was used on RNA isolated from SAN, atria, and ventricles from cKO and littermate controls after tamoxifen administration (Fig. 1C). Kir6.1 expression was reduced by ϳ50% in SAN of cKO mice compared with WT SAN (n ϭ 5, p Ͻ 0.01). Expression of Kir6.1 was unaffected in both atria and ventricles, confirming conduction system-specific deletion (p Ͼ 0.05). Interestingly, expression of Kir6.2 (Kcnj11) in SAN was increased by ϳ50%, suggesting possible compensation for Kir6.1 deletion (n ϭ 5, p Ͻ 0.01). SUR1 (Abcc8) and SUR2 (Abcc9) expression was unchanged in any of the cKO tissues assayed compared with WT (n ϭ 5, p Ͼ 0.05).

Kir6.1 deletion reduces the K ATP current in acutely isolated SAN cells from cKO mice
To investigate whether the deletion of Kir6.1 in the SAN has functional consequences, we subjected acutely isolated SAN cells from cKO and WT mice to whole-cell patch clamping (Fig.  2). The K ATP current in the SAN has not been extensively characterized, so some basic pharmacological experiments were first performed. Not all SAN cells were found to express a pinacidil/diazoxide-sensitive current (15/29 cells, n ϭ 29 cells from 6 mice). Of the WT cells that did express a K ATP current, ϳ7% were responsive to pinacidil, whereas ϳ41% were activated by diazoxide (Fig. 2). In cells from cKO mice, ϳ4 and ϳ11% responded to pinacidil and diazoxide, respectively, (n ϭ 28 cells from 4 mice, p ϭ 0.015 for diazoxide sensitivity when compared with WT). Both pinacidil-and diazoxide-sensitive currents were inhibited by tolbutamide. These data suggest that a K ATP channel comprising Kir6.1 is present in a proportion of SAN cells and that it is predominantly activated by diazoxide and inhibited by tolbutamide.

Kir6.1 cKO mice have a bradycardic phenotype
The potential physiological consequences of Kir6.1 deletion in SAN on heart rate and rhythm were then studied. ECG radiotelemetry was used in conscious freely moving WT and cKO mice (Fig. 3). Both WT and cKO mice showed a diurnal variation in heart rate before and after tamoxifen induction (p Ͻ 0.01 compared night to day). HR, RR (the time between 2 consecutive R waves on an ECG), and PR intervals (times between the onset of the P wave to the beginning of QRS complex) were not significantly different before and after tamoxifen administra- K ATP in the sinoatrial node tion in ECG recordings from littermate control mice (n ϭ 5, p Ͼ 0.05). Conversely, cKO mice developed a bradycardic phenotype following administration of tamoxifen with significant changes in HR, PR, and RR interval both during the day and night (n ϭ 8, p Ͼ 0.05). HR decreased by ϳ50 bpm during the day and at night, and this was reflected in an increase in both RR and PR intervals. A closer investigation of the ECG recordings revealed a much more severe bradycardic phenotype in 3 of 8 cKO mice (Fig. 3E). In these mice, there were pronounced bradycardic episodes with evidence of sinus pauses reflected in RR intervals of greater than 1 s in some instances. We quantified these episodes by counting the number of RR intervals greater than 200 ms in all mice following tamoxifen induction (Fig. 3F). cKO mice had a greater frequency of prolonged beats compared with WT mice (p Ͻ 0.01).
The change in heart rate in cKO mice was further investigated by carrying out heart rate variability (HRV) analysis (Fig.  4). We looked at ECG recordings during a period of low activity (12-2 p.m.) to analyze possible changes in autonomic influence in HRV. Frequency domain analysis showed no change in HRV in WT and cKO mice following tamoxifen induction (n ϭ 5, p Ͼ 0.05). There was no change in the power of both low (0.4 -1.5 Hz) and high (1.5-4 Hz) frequency spectra or the total power, suggesting that the deletion of Kir6.1 from the conduction system does not grossly affect autonomic control of heart rate.

Action potential duration is prolonged in isolated SAN cells from cKO mice
The effects of conduction-specific Kir6.1 deletion on spontaneous SAN action potentials were investigated using currentclamp (Fig. 5). Deletion of Kir6.1 reduced the firing rate of spontaneous action potentials in cKO SAN cells but did not change SAN cell resting membrane and maximum diastolic potential ( Fig. 5 and Table 1). Initial upstroke velocity was unchanged, but the time to peak was increased in cKO mice (n ϭ 11-15 cells from 7-9 mice, p Ͻ 0.05). Repolarization was delayed throughout the action potential but particularly during early repolarization (Fig. 5). This was reflected in a ϳ50% decrease in the action potential duration (APD) ratio, APD90/50 (p Ͻ 0.005) and an increase in the APD30 -40/APD70 -80 ratio (p Ͻ 0.05) in SAN cells from cKO mice (Fig. 5, F and G). These data suggest that deletion of Kir6.1 changes the duration and shape of the SAN action potential but does not contribute significantly to resting membrane potential in these cells.
To investigate the phenotype further in vitro, multielectrode arrays (MEAs) were used to measure spontaneous field potentials in SAN/atrial preparations from WT and cKO mice (Fig. 6,  A and B). MEA recordings showed a significantly reduced beating frequency in preparations from cKO mice compared with WT mice (n ϭ 12, p Ͻ 0.05). In support of these observations, SAN cells acutely isolated from cKO mice after tamoxifen

Sinus node recovery time (SNRT) is prolonged, and its adaptation to stress is compromised in cKO mice
To analyze SAN function in the intact heart, we used a flexMEA to measure the SNRT in the Langendorff-perfused heart ( Fig. 6, C-E). SNRT, following pacing with S1-intervals of 100 ms, was significantly increased in cKO mouse hearts compared with WT hearts (n ϭ 10 -12, p Ͻ 0.05), suggesting potential SAN dysfunction in cKO mice. K ATP channels are known to be protective against metabolic challenge (1). To investigate whether deletion of Kir6.1 in the SAN changes adaptation to stress, we challenged WT and cKO hearts with hypoxia (Krebs bubbled with 95% N 2 , 5% CO 2 ) and measured the SNRT. In WT hearts SNRT was increased following 10 min of hypoxia (Fig. 6E, p Ͻ 0.05 compared with control). Interestingly, hypoxia had no effect on SNRT in cKO hearts (p Ͼ 0.05). SNRT of WT hearts returned close but not completely to basal levels following 10 min of reperfusion.

Increased expression of pro-fibrotic markers in hypomorphic cKO SANs
We examined whether deletion of Kir6.1 in the SAN had long-term consequences for SA node morphology and pathology. Masson's trichrome staining was used to investigate whether Kir6.1 deletion increases fibrosis and changes the morphology of the SAN in cKO mice (Fig. 7). The SA node appeared smaller and thinner in SAN sections from cKO mice, with fewer viable myocytes. There was more fibrous tissue compared with WT SANs. To quantify these observations, RT-qPCR was performed for known fibrosis-related genes. An increase in expression of a number of pro-fibrotic markers within cKO SANs was observed (Fig. 7C). Specifically, the expression of genes for collagen (Col1a3), Tgf-b1, and Tgf-b2 were significantly increased by ϳ50% (n ϭ 5, p Ͻ 0.05). These data suggest that deletion of Kir6.1 in the SAN can lead to an increase in pro-fibrotic markers and fibrosis. We also examined whether there were any changes in expression of ion channels other than Kir6.1 in the cKO mice indicative of a remodeling process. We used RT-qPCR (n ϭ 4 mice, means from triplicate measurements of ⌬C T ) to measure the expression of Hcn4 (WT, 7.73 Ϯ 0.46; cKO, 7.52 Ϯ 0. 16

Discussion
Using a novel murine model, we have deleted the K ATP channel subunit Kir6.1 in the SAN. Uniquely, our data indicate that

K ATP in the sinoatrial node
Kir6.1 underlies a K ATP current in a proportion of SAN cells and influences repolarization during the spontaneous action potential. This leads to mice having a bradycardic phenotype with episodes of sinus arrest in some mice, changes intrinsic to the SAN. Furthermore, the impaired adaptive responses of the SAN to hypoxia suggest that the K ATP channel promotes protection in these conditions. Our studies define a novel and little appreciated contributor to mammalian pacemaking.
Comparatively little literature exists on the nature of K ATP channels in the SAN. In rabbit SAN cells, the potassium channel openers pinacidil and cromakalim abolished pacemaker activity through membrane hyperpolarization (18). Electrophysiological studies identified an ATP-sensitive current with a unique single channel conductance of ϳ50 pS rather than one closer to the 80 pS found for channels in ventricular cells (18). In mice with global genetic deletion of Kir6.2, the SAN in the intact perfused heart no longer responded to hypoxia with an increase in coupling length (13). Our data show that a prominent K ATP current could be demonstrated in ϳ50% of murine SAN myocytes. This is perhaps not surprising because the SAN is a heterogeneous structure, and a number of different cell types and regions can be distinguished (19,20). In these studies, the number of cells containing K ATP currents was significantly reduced (yet not wholly abolished) in cKO mice, indicating that Kir6.1 is a major contributor to these membrane currents. The single-channel conductance observed by Han et al. (18) is closer to that reported for Kir6.1 containing complexes or that for various heteromultimers between Kir6.1 and Kir6.2 (21). The pharmacology is also interesting with diazoxide more frequently activating these currents, whereas tolbutamide as well as glibenclamide can inhibit the responses. This is more repre-

K ATP in the sinoatrial node
sentative of SUR1-containing K ATP channel complexes than SUR2, although diazoxide can act on SUR2B-containing heterooctamers (22). SUR1-containing K ATP channels are already known to be present in murine atrial myocytes (23). In cKO mice there was an increase in Kir6.2 expression as assessed by RT-qPCR but not in SUR1 or SUR2 expression. Furthermore, K ATP currents were essentially abolished with no response to diazoxide and only a very few cells, comparable with controls, responding to pinacidil. Thus the increased expression of Kir6.2 mRNA does not have prominent functional sequelae. The cKO mice are clearly bradycardic when assayed by three independent assays: single cell, ex vivo and in vivo studies. In the latter studies, no evidence was found supporting changes in autonomic or hormonal influences on heart rate in the cKO mice. In other words, the process observed was solely intrinsic to the SA node, as would be expected from the specificity of expression of the cre recombinase.
Our action potential recordings demonstrate an increase in action potential duration particularly prominent during the early stages of repolarization. Thus, even under resting conditions and without metabolic challenge, Kir6.1-containing K ATP channels are active and contribute to potassium currents involved in shaping the SAN action potential. Furthermore, the maximum diastolic depolarization and resting membrane potential were not prominently changed in knockout animals.

Figure 6. Reduced beat frequency in intact right atrial/SAN preparations and increased SNRT in intact hearts from cKO mice.
A, representative field potentials from WT and cKO right atrial tissue with the SAN intact recorded using two-dimensional MEA. B, mean beat frequency from RA/SAN tissue from WT (n ϭ 11) and cKO (n ϭ 12) mice. C, SNRT was measured using a flexMEA on isolated heart preparations. Hearts were paced for 20 s at S1S1 cycle lengths of 100 ms and defined as the time interval from the last pacing spike to the first recorded atrial activity. Hearts were challenged with hypoxia (95% N2) and reperfusion. D, mean basal SNRT measured from WT (n ϭ 12) and cKO (n ϭ 10) hearts. E, mean relative SNRT during hypoxia and reperfusion in WT (n ϭ 12) and cKO (n ϭ 10) hearts. The data are shown as means Ϯ S.E. *, p Ͻ 0.05 compared WT.

K ATP in the sinoatrial node
The time-to-peak depolarization was also delayed in the cKO mice. The net upshot of these cumulative effects is a prolongation of the action potential. Given an equivalent or unchanged time period for diastolic depolarization, a consequence would be to increase the SAN cycle length and thus decrease heart rate. K ATP currents generated by Kir6.1-containing subunits are actually outwardly rectifying (see Fig. 2). However, currents do not deactivate with hyperpolarization, as would a classic voltage-gated potassium channel, and it is surprising that they do not have a significant influence on the pacemaker depolarization. It is known that cAMP signaling is highly developed in the SAN (24), and it is possible that phasic oscillations in PKA activity might amplify the K ATP current during repolarization. PKA is known to prominently regulate Kir6.1-containing channel complexes (25). However, there is precedent for analogous observations in the literature: deletion of the gene encoding TREK-1, a background potassium current, led to sinus bradycardia and pauses (26).
The Hcn4 promoter-driven cre is also active in other regions of the conduction system. Although a formal investigation of this fell outside of the scope of this study, an increase in PR interval on the ECG reflective of delayed AV nodal conduction was observed in these animals. This implies that Kir6.1 also plays a role in the physiological function of the AV node. Indeed, gain-of-function mutations in the Kir6.1 subunit found in patients with Cantu syndrome and mouse models with Kir6.1 gain of function show AV conduction abnormalities (27,28).
In the global Kir6.1 knockout mouse, one of the major contributors to premature death is sudden cardiac death caused by bradyarrhythmia resulting from sinus arrest and heart block (14,15). Selected cKO mice displayed features of this phenomenon, but not as prominently as in the global knockouts. Opening of K ATP channels can contribute to cellular protection in several pathological conditions (1,29). We therefore examined the response of the SAN to hypoxia in ex vivo intact hearts, using sinus node recovery from burst pacing (SNRT) as a comparison metric. In WT littermate controls, the SNRT was prolonged by hypoxic challenge, whereas in the cKO mice this did not occur. It is plausible that this response by the SAN is protective because it allows some degree of electrical silencing in conditions of metabolic challenge. The electrical silencing likely occurs through action potential failure because of overwhelming repolarizing currents. This will prevent cellular calcium overload and cell death. The observed hypoplasia of the SA node and increased expression of profibrotic markers support the hypothesis of cell death and fibrosis in cKO mice. It is difficult to unambiguously separate the functional effects of a direct effect of Kir6.1 deletion and the pathological changes seen in SAN on SAN electrophysiology. We think both contribute; however, it is worth noting that we have performed singlecell studies on SAN myocytes where a reduced spontaneous firing rate is present in cKO mice, and the relative change in heart rate in the knockout is similar to that in the ex vivo and in vivo studies. It is also possible that the fibrotic process leads to remodeling of ion channel expression in the SAN. Thus we examined the expression of a selected number of ion channels in the SAN using RT-qPCR. We did not find prominent changes in expression between WT control SAN cells and those with deletion of Kir6.1.
The more severe electrophysiological phenotype seen in global Kir6.1 knockout mice can be explained by likely episodes of coronary vasospasm leading to hypoxia in the SAN combining with failures of appropriate protective responses. The phenotype in the cKO mice is milder, and sinus arrest occurs in only a limited number of mice. However, the primacy of the cardiac electrophysiology effect is supported by the observation that a pure vascular phenotype, as in a murine line with specific smooth muscle deletion, did not display a bradycardiac phenotype under resting conditions or with ergonovine (15). It is interesting that sudden infant death syndrome has been associated with loss-of-function mutations in KCNJ8 (30).
This study is the first to formally define the role of Kir6.1 in the sinus node and significantly adds to the understanding of the complexities of sinus node physiology. However, there are several limitations to this study. First, this study investigated the murine SAN given that cre/loxP technology is as yet not easily replicated in other species. Mutations in Kir6.1 have been implicated in causing atrial fibrillation and early repolarization syndromes in man (31,32). However, there are data indicating that Kir6.1 may be expressed more widely in murine cardiac tissue than in man for example (33), and this may confound translation to human disease. Second, although we found evidence for the cellular protective effects of Kir6.1, we have not performed long-term aging studies that may be required to fully define the precise molecular effects on these pathways.
In conclusion, we show that K ATP channels constituted of Kir6.1 are present in SAN and influence repolarization. Cycle length is prolonged, resulting in bradycardia in cKO mice. In addition, these channels may also protect against sinus node dysfunction for example during hypoxic conditions.

Animal husbandry
All experiments were conducted in accordance with the Guide for the Care and Use of Laboratory Animals published by the British Home Office and the U.S. National Institutes of Health (Publication 85-23, revised 1996). The animal work was approved by the Queen Mary University of London (QMUL) Animal Welfare and Ethical Review Body and covered by British Home Office Project Licenses PPL/6732 and PPL/7665.

K ATP in the sinoatrial node
Genotyping DNA was extracted from mouse ear biopsies by proteinase K digestion. The presence of the Kir6.1 floxed allele was confirmed using PCR with the following primer set: sense, 5Ј-GAGATCTTAACTCAGTTCTGGAGGACCAACA-3Ј; and antisense, 5Ј-AGCGAAGAAAACTGCTTCCTGTTCATT-AAG-3Ј yielding a WT band of 474 bp and a floxed allele band of 600 bp. PCR cycle conditions were denaturation at 94°C for 2 min, 35 cycles of 94°C for 30 s, 63°C for 30 s, 68°C for 1 min, and extension at 68°C for 8 min. The presence of the cre recombinase gene in the Kir6.1 conditional knockout line was determined using the following primer set: sense, 5Ј-CCA-ATTTACTGACCGTACACC-3Ј; and antisense, 5Ј-GTTTCA-CTATCCAGGTTACGG-3Ј yielding a band of 900 bp in crepositive mice and no band in cre-negative mice. PCR cycle conditions were denaturation at 94°C for 2 min, 35 cycles of 94°C for 1 min, 60°C for 1 min, 72°C for 1 min, and extension at 72°C for 5 min. Both the floxed and cre PCRs used Taq polymerase (New England Biolabs).

Isolation of SAN myocytes
SAN myocytes were isolated using the method of Mangoni and Nargeot (34). Briefly, the mice were injected with heparin sodium (250 IU) and anesthetized with a combination of ketamine (0.1 mg/g) and xylazine (0.01 mg/g). The hearts were rapidly excised, and the SAN was dissected and placed in Tyrode's solution consisting of 140 mM NaCl, 5.4 mM KCl, 1.8 mM CaCl 2 , 1 mM MgCl 2 , 5 mM HEPES-NaOH, and 5.5 mM D-glucose (adjusted to pH 7.4 with NaOH). The SAN was then placed in low Mg 2ϩ /Ca 2ϩ solution containing 140 mM NaCl, 5.4 mM KCl, 0.5 mM MgCl 2 , 0.2 mM CaCl 2 , 1.2 mM KH 2 PO 4 , 50 mM taurine, 5.5 mM D-glucose, 1.0 mg/ml BSA, and 5.0 mM HEPES-NaOH (adjusted to pH 6.9 with NaOH). The tissue was digested with 1 mg/ml collagenase type II (Worthington), 1 mg/ml protease (Sigma-Aldrich), 1 mg/ml BSA (Sigma-Aldrich), and 200 M CaCl 2 for ϳ20 min at 37°C with consistent agitation with a fire-polished Pasteur pipette. The tissue was then washed three or four times in modified Kraft-Bruhe solution containing 70 mM L-glutamic acid, 20 mM KCl, 80 mM KOH, 10 mM (Ϯ) D-␤-OH-butyric acid, 10 mM KH 2 PO 4 , 10 mM taurine, 1 mg/ml BSA, and 10 mM HEPES-KOH, pH 7.4, with KOH. SAN cells were dissociated manually using a fire-polished Pasteur pipette in Kraft-Bruhe solution and allowed to rest for 5 min prior to readaptation. The cells were allowed to readapt to physiological levels of Na ϩ and Ca 2ϩ by incremental addition of a solution containing 10 mM NaCl and 1.8 mM CaCl 2 and finally normal Tyrode's solution with 1 mg/ml BSA. Readapted cells were stored at room temperature in storage solution containing 100 mM NaCl, 35 mM KCl, 1.3 mM CaCl 2 , 0.7 mM MgCl 2 , 14 mM L-glutamic acid, 2 mM (Ϯ)D-␤-OH-butyric acid, 2 mM KH 2 PO 4 , 2 mM taurine, and 1 mg/ml BSA, pH 7.4. SAN cells were plated onto 13-mm coverslips coated with laminin 30 min prior to patch-clamp recordings.

Patch-clamp recordings from isolated SAN myocytes
Whole-cell patch-clamp recordings were performed as described previously (15). Capacitance transients and series resistance in whole-cell recordings was compensated electronically by using amplifier circuitry (Axopatch 700B). The data were filtered at 1 kHz using the filter provided with the Axopatch 700B (4-pole Bessel) and sampled at 5 kHz using a Digidata 1550 (Axon Instruments). The currents were acquired and analyzed using pClamp10 (Axon Instruments). Whole-cell currents were recorded using a voltage-ramp protocol (Ϫ150 to ϩ50 mV for 1 s). Pipette solution contained 107 mM KCl, 1 mM MgCl 2 , 10 mM EGTA, and 10 mM HEPES with 3 mM MgATP, and 1 mM Na 2 ADP, pH 7.2, using KOH. The bath solution contained 137 mM NaCl, 5.4 mM KCI, 1 mM MgCl 2 , 1.8 mM CaCl 2 , 5 mM glucose, 10 mM HEPES, pH 7.4, using NaOH. Drugs were applied to the bath using a gravity-driven perfusion system. Whole-cell action potentials were recorded using the currentclamp configuration. The pipette solution for current-clamp recordings contained 110 mM potassium gluconate, 20 mM KCl, 10 mM HEPES, 0.05 mM EGTA, 0.5 mM MgCl 2 , 5 mM MgATP, 0.3 mM Na 2 -GTP, 5 mM Na 2 -phosphocreatine, pH 7.4, using KOH. The data were initially analyzed using Clampfit 10 (Axon Instruments).

Masson's trichrome staining
Trichrome staining was used to distinguish between muscle, collagen, and connective tissue in serial 10 M SAN sections embedded in paraffin. We used the Sigma-HT15 trichrome stain (Masson) kit as per the manufacturer's instructions (Sigma-Aldrich).

ECG telemetry
Briefly, the recording leads were tunneled subcutaneously in a conventional lead II ECG configuration connected to a telemetry device (TA-F10, Data Sciences International) implanted either subcutaneously or intra-abdominally. Surface ECG was continuously recorded via radiotelemetry after 2 weeks of post-K ATP in the sinoatrial node operative recovery. To record standard surface ECG parameters, consecutive individual ECG complexes recorded over 2 min during sinus rhythm at high sampling frequency (2 kHz) were analyzed using Ponemah P3 plus analysis software (Data Sciences International). Diurnal variation of heart rate was measured by averaging heart rate over a 12 h period (7 a.m. to 7 p.m. (day) and 7 p.m. to 7 a.m. (night)) corresponding to the light/ dark cycle in the animal facility over at least a 48-h period of recording.

Analysis of heart rate variability
Changes in the RR interval variability were analyzed using the HRV module in LabChart v7.0 (AD Instruments). ECG recordings taken at a low activity period (12 p.m. to 2 p.m.) were analyzed using power spectral analysis as described previously (35). R waves were detected using a threshold-sensing algorithm, and ectopic beats were excluded from the analysis. Fast Fourier transformation was performed using 1024 spectral points and a half overlap within a Welch window, and power spectral density plots were determined. The total power (TP) (s 2 /Hz) was calculated as the integral sum of total variability after Fourier transformation over the frequency range 0 -4.0 Hz. Cut-off frequencies determined to be accurate for mice were used to divide the signal into three components: very low frequency (VLF: 0.0-0.4 Hz), low frequency (LF: 0.4 -1.5 Hz), and high frequency (HF: 1.5-4.0 Hz). Data from segments was also normalized to exclude VLF. Normalized low frequency (nLF) ϭ LF/(TP Ϫ VLF) ϫ 100 and normalized high frequency (nHF) ϭ HF/(TP Ϫ VLF) ϫ 100.

MEA recordings
MEA recordings were taken from atrial/SAN preparations and from the intact heart. We used two types of MEAs to record spontaneous electrical activity (field potentials) from in vitro atrial preparations (60-electrode MEA, 60 pMEA200/30iR-Ti and MEA2100 system Multi Channel Systems, Reutlingen, Germany) and the intact heart (32 electrodes flexMEA, Multi Channel Systems). Details of electrophysiological recordings with the MEA2100 system have been described previously (36). Briefly, the right atrial appendage and SA node were carefully and quickly dissected from the heart following excision and placed on the MEA dish and secured for good contact with electrodes with a holder. Tissue was continuously perfused with warm (37°C) and oxygenated (95% O 2 /5% CO 2 ) Krebs solution (consisting of 118 mM NaCl, 4.75 mM KCl, 1.19 mM MgSO 4 ⅐7H 2 O, 25 mM NaHCO 3 , 1.19 mM KH 2 PO 4 , 5 mM D-glucose, 1.4 mM CaCl 2 , and 2 mM sodium pyruvate). The tissue was allowed to equilibrate for 30 min prior to recording of spontaneous extracellular field potentials using MC Rack software (version 4.6.2, Multi Channel Systems).
For experiments on intact isolated heart preparations, hearts were excised as above and quickly cannulated via the aorta and mounted onto a modified Langendorff setup. Hearts were continuously retrogradely perfused with Krebs (as above). The flexMEA was placed on the boundary of the right ventricle and right atria allowing for measurement of spontaneous electrical signals from both atria and ventricles using MC Rack software. SNRT was measured using the flexMEA and a unipolar elec-trode to pace at S1S1 cycle lengths of 100 ms. SNRT was measured from the last pacing spike to the first atrial signal. The data were analyzed using LabChart v7 and Clampfit v10.

Statistical analysis
The data are presented as means Ϯ S.E. The data were analyzed using Microsoft Excel, Microcal Origin, and GraphPad Prism. Student's t test and analysis of variance were used to compare means where appropriate. p Յ 0.05 was taken to be significant.