Intracellular signalling mechanism responsible for modulation of sarcolemmal ATP-sensitive potassium channels by nitric oxide in ventricular cardiomyocytes

Abstract

The ATP-sensitive potassium (K_ATP) channels are crucial for stress adaptation in the heart. It has previously been suggested that the function of K_ATP channels is modulated by nitric oxide (NO), a gaseous messenger known to be cytoprotective; however, the underlying mechanism remains poorly understood. Here we sought to delineate the intracellular signalling mechanism responsible for NO modulation of sarcolemmal K_ATP (sarcK_ATP) channels in ventricular cardiomyocytes. Cell-attached patch recordings were performed in transfected human embryonic kidney (HEK) 293 cells and ventricular cardiomyocytes freshly isolated from adult rabbits or genetically modified mice, in combination with pharmacological and biochemical approaches. Bath application of the NO donor NOC-18 increased the single-channel activity of Kir6.2/SUR2A (i.e. the principal ventricular-type K_ATP) channels in HEK293 cells, whereas the increase was abated by KT5823 [a selective cGMP-dependent protein kinase (PKG) inhibitor], mercaptopropionyl glycine [MPG; a reactive oxygen species (ROS) scavenger], catalase (an H₂O₂-degrading enzyme), myristoylated autocamtide-2 related inhibitory peptide (mAIP) selective for Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) and U0126 [an extracellular signal-regulated protein kinase 1/2 (ERK1/2) inhibitor], respectively. The NO donors NOC-18 and N-(2-deoxy-α,β-d-glucopyranose-2-)-N²-acetyl-S-nitroso-d,l-penicillaminamide (glycol-SNAP-2) were also capable of stimulating native sarcK_ATP channels preactivated by the channel opener pinacidil in rabbit ventricular myocytes, through reducing the occurrence and the dwelling time of the long closed states whilst increasing the frequency of channel opening; in contrast, all these changes were reversed in the presence of inhibitors selective for soluble guanylyl cyclase (sGC), PKG, calmodulin, CaMKII or ERK1/2. Mimicking the action of NO donors, exogenous H₂O₂ potentiated pinacidil-preactivated sarcK_ATP channel activity in intact cardiomyocytes, but the H₂O₂-induced K_ATP channel stimulation was obliterated when ERK1/2 or CaMKII activity was suppressed, implying that H₂O₂ is positioned upstream of ERK1/2 and CaMKII for K_ATP channel modulation. Furthermore, genetic ablation (i.e. knockout) of CaMKIIδ, the predominant cardiac CaMKII isoform, diminished ventricular sarcK_ATP channel stimulation elicited by activation of PKG, unveiling CaMKIIδ as a crucial player. Additionally, evidence from kinase activity and Western blot analyses revealed that activation of NO–PKG signalling augmented CaMKII activity in rabbit ventricular myocytes and, importantly, CaMKII activation by PKG occurred in an ERK1/2-dependent manner, placing ERK1/2 upstream of CaMKII. Taken together, these findings suggest that NO modulates ventricular sarcK_ATP channels via a novel sGC–cGMP–PKG–ROS(H₂O₂)–ERK1/2–calmodulin–CaMKII (δ isoform in particular) signalling cascade, which heightens K_ATP channel activity by destabilizing the long closed states while facilitating closed-to-open state transitions. This pathway may contribute to regulation of cardiac excitability and cytoprotection against ischaemia–reperfusion injury, in part, by opening myocardial sarcK_ATP channels.

Introduction

Vital in the adaptive response to (patho)physiological stress, the ATP-sensitive potassium (K_ATP) channel functions as a high-fidelity metabolic sensor, which couples intracellular metabolic state to membrane excitability (Ashcroft, 1988; Miki & Seino, 2005; Nichols, 2006) and serves a homeostatic role ranging from blood glucose regulation to cardioprotection (Olson & Terzic, 2010). The K_ATP channel is a hetero-octameric protein composed of four inwardly rectifying potassium channel subunits (Kir6.x) and four sulphonylurea receptors (SURx; Shyng & Nichols, 1997; Babenko et al. 1998), whose molecular (subunit) composition exhibits tissue specificity. For example, in cardiac (ventricular) and skeletal muscles the K_ATP channels are composed of Kir6.2 and SUR2A subunits (Inagaki et al. 1996; Okuyama et al. 1998), whereas in central neurons and pancreatic β-cells they consist of Kir6.2 and SUR1 subunits (Aguilar-Bryan et al. 1998). Whilst it is appreciated that K_ATP channels are directly regulated by intracellular ATP, MgADP (Nichols, 2006) and phosphatidylinositol-4,5-bisphophate (PIP₂; Fan & Makielski, 1997; Baukrowitz et al. 1998; Shyng & Nichols, 1998), how these important channels are modulated by more complicated intracellular signalling processes is far less understood.

The gaseous messenger nitric oxide has a fundamental biological role in protecting the heart against ischaemia–reperfusion injury (Bolli, 2001). It has been suggested that NO shortens action potential duration (APD)₉₀ and increases maximal diastolic potential in the heart, by activating sarcolemmal K_ATP (sarcK_ATP) channels via a cGMP-dependent mechanism (Baker et al. 2001). Nitric oxide also potentiates the action of potassium channel openers (KCOs) on the K_ATP channel in single ventricular cells, yet with conflicting findings on whether cGMP is involved (Shinbo & Iijima, 1997; Han et al. 2002). The intracellular mechanism by which NO modulates cardiac K_ATP channels has remained largely unknown.

In the present study, we combined single-channel patch-clamp recordings with pharmacological and biochemical approaches to delineate the intracellular signalling mechanism responsible for NO modulation of cardiac sarcK_ATP channels. Human embryonic kidney (HEK) 293 cells expressing recombinant cardiac-type K_ATP (i.e. Kir6.2/SUR2A) channels and ventricular cardiomyocytes freshly isolated from adult rabbits as well as from CaMKIIδ gene-null and wild-type mouse models expressing endogenous K_ATP channels were used. Specifically, we investigated the involvement in NO signal transduction of soluble guanylyl cyclase (sGC), cGMP-dependent protein kinase (PKG), reactive oxygen species (ROS), hydrogen peroxide (H₂O₂), calmodulin, calcium/calmodulin-dependent protein kinase II (CaMKII) and extracellular signal-regulated protein kinase (ERK)1/2 of the mitogen-activated protein kinase (MAPK) family. Here we show that functional modulation of ventricular sarcK_ATP channels by NO induction is mediated by intracellular signalling via a novel sGC–cGMP–PKG–ROS(H₂O₂)–ERK1/2–calmodulin–CaMKII (CaMKIIδ isoform in particular) signalling pathway that alters the open and closed properties of the channel, enhancing channel activity.

Methods

Ethical approval

All protocols involving animals were approved by the institutional Animal Care and Use Committee at the University of California, Davis, and experiments were performed in strict accordance with the Guide for the Care and Use of Laboratory Animals 8th edition (2011) of the National Research Council, USA and conformed to the principles of UK regulations as described by Drummond (2009).

Construction of cDNAs

To reconstitute cardiac ventricular-type K_ATP channels, cDNAs encoding the pore-forming subunit Kir6.2 (mouse; gift from Dr. Susumu Seino at Kobe University, Chuo-ku, Japan) and the regulatory subunit SUR2A (rat; gift from Dr. Joseph Bryan at Baylor College of Medicine, Houston, TX, USA) were subcloned into mammalian expression vectors pIRES-EGFP (Clontech, Mountain View, CA, USA) and pcDNA3 (Invitrogen, Carlsbad, CA, USA), respectively. The plasmids to be used for transient transfection were prepared with Qiagen maxipreps and verified by DNA sequencing (Qiagen, Valencia, CA, USA).

Mammalian cell culture and transient transfection

The HEK293 cells (ATCC, Manassas, VA, USA) were maintained in Dulbecco's modified Eagle's medium DMEM/F12 (Mediatech, Herndon, VA, USA; supplemented with 2 mm l-glutamine, 10% fetal bovine serum, 100 IU ml⁻¹ penicillin and 100 μg ml⁻¹ streptomycin) at 37°C in humidified air supplemented with 5% CO₂. Cells were transiently transfected with expression plasmids containing cDNAs of interest using a modified calcium phosphate–DNA coprecipitation method (Chen & Okayama, 1987; Jordan et al. 1996). Positive transfection was marked by cistronic EGFP expression provided by the vector pIRES-EGFP. The cells were replated the following day at a density of 5000–20,000 cells per dish onto 12 mm glass coverslips precoated with fibronectin (∼0.5 μg per coverslip, or 0.5 μg cm⁻²; Sigma-Aldrich, St Louis, MO, USA) to be recorded 48–72 h after transfection as previously described (Lin et al. 2000).

Isolation of ventricular cardiomyocytes

Rabbits

Left ventricular myocytes were enzymatically isolated from adult New Zealand White rabbits as described before (Chai et al. 2011). Rabbits were deeply anaesthetized by intravenous injection of pentobarbital sodium (80–100 mg kg⁻¹). Hearts were excised and quickly placed on a Langendorff apparatus and perfused retrogradely for 5–7 min with nominally Ca²⁺-free Dulbecco's minimal essential medium solution. Perfusion was then switched to the same solution containing 1 mg ml⁻¹ collagenase with up to 0.1 mg ml⁻¹ neutral protease. Once the heart became flaccid (∼15–30 min), the ventricles were dispersed and filtered. The cell suspension was washed several times with medium containing ∼150 μm Ca²⁺.

Mice

CaMKIIδ-null mice (generated as reported previously; Ling et al. 2009) and their littermate/wild-type controls were anaesthetized with isoflurane at 3–5% in 100% oxygen via a Bickford veterinary vapourizer with a flow rate of 1–2 l min⁻¹, followed by decapitation. Hearts were excised, and myocytes were dissociated from ventricles by enzymatic treatment. Isolated ventricular myocytes were subsequently plated on 12 mm glass coverslips freshly coated with laminin (∼1 μg per coverslip, or 1 μg cm⁻²; Invitrogen, Carlsbad, CA, USA) to enhance cell adhesion. Rod-shaped cells with clear margin and striation were used for immediate recordings.

Electrodes, recording solutions and single-channel recordings

The recording electrodes were pulled from thin-walled borosilicate glass with an internal filament (MTW150F-3; World Precision Instruments, Sarasota, FL, USA) using a P-97 Flaming Brown puller (Sutter Instrument Co., Novato, CA, USA) and were firepolished to a resistance of 5–10 MΩ. Cell-attached single-channel recordings (Hamill et al. 1981) were performed using a recording chamber (RC26; Warner Instruments, Hamden, CT, USA) filled with the intracellular (bath) solution, and the recording pipette was filled with the extracellular solution. For HEK293 cells, the intracellular (bath) solution consisted of (mm): KCl, 110; MgCl₂, 1.44; KOH, 30; EGTA, 10; HEPES, 10; and sucrose, 30; pH adjusted to 7.2 with KOH. The extracellular (intrapipette) solution consisted of (mm): KCl, 140; MgCl₂, 1.2; CaCl₂, 2.6; and HEPES, 10; pH adjusted to 7.4 (with KOH). For cardiomyocytes, the intracellular (bath) solution consisted of (mm): KCl, 127; MgCl₂, 1; KOH, 13; EGTA, 5; HEPES, 10; and glucose, 10; pH adjusted to 7.2 (with KOH). The extracellular (intrapipette) solution consisted of (mm): KCl, 140; MgCl₂, 1; CaCl₂, 2; HEPES, 10; and glucose, 10; pH adjusted to 7.4 (with KOH). The use of symmetrical recording solutions (140 mm K⁺) resulted in an equilibrium potential for potassium (E_K) and a resting membrane potential (V_m) around 0 mV, as determined from the I–V relationship of the K_ATP channel. All recordings were carried out at room temperature, and all patches were voltage clamped at −60 mV (i.e. with +60 mV intrapipette potentials) unless specified otherwise. Single-channel currents were recorded with an Axopatch 200B patch-clamp amplifier (Molecular Devices: Axon Instruments, Sunnyvale, CA, USA), low-pass filtered (3 dB, 2 kHz) and digitized at 20 kHz online using Clampex 9 software (Axon Instruments) via a 16 bit A/D converter (Digidata acquisition board 1322A; Axon Instruments).

Preparations of drugs

Working solutions of N-(2-deoxy-α,β-d-glucopyranose-2-)-N²-acetyl-S-nitroso-d,l-penicillaminamide (glycol-SNAP-2), DETA NONOate (NOC-18), 1,4-dihydro-5-(2-propoxyphenyl)-7H-1,2,3-triazolo[4,5-d]pyrimidine-7-one (zaprinast), pinacidil, 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ), KT5823, N-(2-mercaptopropionyl)glycine (MPG), 5-hydroxydecanoate (5-HD), fluphenazine-N-2-chloroethane (SKF-7171A), myristoy-lated autocamtide-2 related inhibitory peptide for CaMKII (mAIP), 1,4-diamino-2,3-dicyano-1,4-bis(2-aminophenylthio)butadiene (U0126) and 2′-amino-3′-methoxyflavone (PD98059) were diluted from aliquots with bath recording solutions prior to use. Stock solutions were prepared as follows: zaprinast, pinacidil, KT5823, ODQ, SKF-7171A, U0126 and PD98059 in DMSO; and glycol-SNAP-2, NOC-18, MPG, 5-HD and mAIP in H₂O; all were stored at −80°C in aliquots. Working solutions of catalase (human erythrocyte) and H₂O₂ were prepared directly from original stocks immediately before use. All working drug solutions were put on ice and kept away from light. Drugs were applied through a pressure-driven perfusion system (BPS-8; ALA Scientific Instruments, Westbury, NY, USA) to the recording chamber via a micromanifold positioned closely to the patches. Reagents and chemicals were purchased from EMD Millipore (Calbiochem, Billerica, MA, USA) or Sigma-Aldrich (St Louis, MO, USA). For pharmacological blockade, individual groups of cells were pretreated with respective inhibitors (except catalase) at room temperature for at least 15 min before being subjected to functional assays.

Electrophysiological data analysis

Data were analysed as described before (Lin et al. 2000, 2004; Mao et al. 2007; Chai & Lin, 2008, 2010; Lin & Chai, 2008; Chai et al. 2011), using individual data files of 120 s durations.

Single-channel currents

Individual, digitized single-channel records of 120 s duration (gap-free) were detected with Fetchan 6.05 (events list) of pCLAMP (Axon Instruments) using the 50% threshold crossing criterion and analysed with Intrv5 (gift from Dr. Barry S. Pallotta, formerly at University of North Carolina, Chapel Hill, NC, USA, and Dr. Janet Fisher at University of South Carolina, Columbia, SC, USA). Analysis was performed at the main conductance level (approximately 70–80 pS) for K_ATP channels. Only patches with infrequent multiple-channel activity were used for single-channel analysis. Duration histograms were constructed as described by Sigworth & Sine (1987), and estimates of exponential areas and time constants were obtained using the method of maximal likelihood estimation. The number of exponential functions required to fit the duration distribution was determined by fitting increasing numbers of functions until additional components could not significantly improve the fit (Horn, 1987; McManus & Magleby, 1988). Events with duration less than 1.5 times the system dead time were not included in the fit. Mean durations were corrected for missed events by taking the sum of the relative area (a) of each exponential component in the duration frequency histogram multiplied by the time constant (τ) of the corresponding component. Each of the single-channel properties was then normalized to the corresponding controls obtained in individual patches (taken as one).

Multiple-channel currents

In patches where multiple-channel activities of K_ATP channels were observed for more than 10% of the recording time, the digitized current records were analysed using Fetchan 6.05 (browse) of pCLAMP to integrate currents in 120 s segments. The current amplitude (I) values (current amplitude = integrated current/acquisition time) were then normalized to the corresponding controls obtained from the same patches to yield normalized open probability (NP_o; control value taken as one), because the normalized current amplitude is equivalent to the normalized NP_o obtained from single-channel analysis when the single-channel conductance remains the same (Mao et al. 2007). The normalized NP_o values obtained from both single-channel and multiple-channel patches were then pooled. In Fig. 1 and all other figures illustrating raw single-channel current records, representative traces (taken from individual 120 s files used for data analysis) with segments marked with a horizontal bar on top are displayed at increasing temporal resolution in successive traces (arranged from top to bottom).

Figure 1.

A–F, representative single-channel current traces of Kir6.2/SUR2A obtained from cell-attached patches before (upper panel of traces) and during (lower panel of traces) application of DETA NONOate (NOC-18, 300 μm; A) or NOC-18 plus one of the following: the KT5823 (1 μm; B); N-(2-mercaptopropionyl)glycine (MPG, 500 μm; C); catalase (500 U ml⁻¹; D); U0126 (10 μm; E); or myristoylated autocamtide-2 related inhibitory peptide selective for CaMKII (mAIP, 1 μm; F), showing that the NO donor NOC-18 increases recombinant Kir6.2/SUR2A channel activity in intact HEK293 cells, whereas the increase induced by NOC-18 is abated when PKG, ROS, H₂O₂, ERK1/2 or CaMKII is selectively inhibited. Patches were voltage clamped at −60 mV. Downward deflections represent openings from closed states. Segments of current traces (taken from individual 120 s data files) marked with a horizontal line atop are displayed in successive traces at increasing temporal resolution. Horizontal scale bars represent 1 s, 300 ms and 100 ms (top to bottom in each three-trace group), and vertical scale bars represent 4 pA. G, averaged normalized open probability (NP_o) of Kir6.2/SUR2A channels obtained from individual groups (control taken as one, indicated by dashed line; mean ± SEM of 7–15 patches), demonstrating that the stimulatory effect of NOC-18 on the normalized NP_o (i.e. relative channel activity) of Kir6.2/SUR2A channels is dependent on PKG, ROS, H₂O₂, ERK1/2 and CaMKII. *P < 0.05, ^**P < 0.01 and ^****P < 0.0001 (Student's two-tailed, one-sample t test within groups, and one-way ANOVA followed by Dunnett's multiple comparison tests among groups).

CaMKII activity assay

Isolated rabbit ventricular myocytes were treated with NOC-18 (300 μm; chemical NO donor) or zaprinast (50 μm; selective inhibitor of cGMP-specific phosphodiesterases V and IX, capable of activating PKG) in the absence and presence of KT5823 (1 μm; selective PKG inhibitor) or U0126 [10 μm; selective mitogen-activated protein kinase kinase or MAPK kinase (MEK) inhibitor] for up to 30 min at room temperature. Immediately after the treatment, myocytes were homogenized using sonication in an ice-cold lysis buffer containing 50 mm HEPES, pH 7.5, 2 mg ml⁻¹ bovine serum albumin, 5 mm EDTA and phosphatase inhibitor cocktail. CaMKII activity assays were then performed on fresh lysates as previously described (Wu et al. 2002; Erickson et al. 2008). Briefly, CaMKII activity was measured as a function of ³²P-ATP incorporation into a synthetic substrate, syntide-2, by scintillation counter. Assays were performed at 30°C. Background measurements lacking syntide-2 were subtracted from experimental values. Kinase activity is expressed relative to baseline radiation from samples containing no cellular lysate. Each experiment was done in triplicate and repeated three times, unless otherwise noted.

Western blot analysis

Rabbit ventricular myocytes were treated and lysed as described under CaMKII activity assay (above). Immunoblotting for total and T287 phosphorylated CaMKII was performed via standard protocols. Equal amounts of protein were loaded and electrophoresed on 10% SDS–polyacrylamide gel before being transferred to a polyvinylidene difluoride membrane. Total CaMKII antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) was used at 1:1000 dilution. Phospho-specific CaMKII antibody (Affinity Bioreagents, Golden, CO, USA) was used at 1:1500 dilution. After incubation with the horseradish peroxidase-labelled secondary antibody, blots were developed using enhanced chemiluminescence (Pierce SuperSignal; Thermo Scientific, Rockford, IL, USA).

Statistics

Data are presented as means ± SEM. Statistical comparisons were made using Student's two-tailed one-sample, paired or unpaired t tests, or one-way ANOVA followed by Dunnett's multiple comparison tests to test the significance of difference in the following: normalized data (in response to treatment) in individual groups (Student's one-sample t tests); raw data pairs obtained before and during treatment in the same group (Student's paired t tests); normalized data between two separate groups (Student's unpaired t tests); or normalized data among multiple groups (one-way ANOVA followed by Dunnett's multiple comparison tests). Significance was assumed when P < 0.05. Statistical comparisons were performed using Prism (GraphPad Software, San Diego, CA, USA).

Results

Stimulation of Kir6.2/SUR2A channels by NO induction in intact HEK293 cells depends on PKG activation

To elucidate the underlying mechanism responsible for functional modulation of cardiac K_ATP channels by NO, we first examined how Kir6.2/SUR2A (i.e. ventricular-type K_ATP) channels transiently expressed in HEK293 cells respond to NO induction. Single-channel recordings were performed in the cell-attached patch configuration to preserve integrity of the intracellular milieu for potential signalling. Bath perfusion of NOC-18 (300 μm), an NO donor which spontaneously releases NO in aqueous solution, markedly enhanced the single-channel activity of Kir6.2/SUR2A channels (Fig. 1A shows a representative patch); the apparent opening frequency and the open duration were both increased, whereas the single-channel conductance remained the same. The averaged normalized NP_o (i.e. relative channel activity) was increased to 4.84 ± 0.68 (control taken as one; Fig. 1G, filled bar; P < 0.0001, Student's two-tailed, one-sample t test; n = 15). In contrast, although pretreatment with the selective PKG inhibitor KT5823 did not alter the basal activity of these channels (Fig. 1A and B), K_ATP channel stimulation evoked by NOC-18 was reduced by more than 50% in the presence of 1 μm KT5823 (following 15 min pretreatment; Fig. 1B and G, open bar; P < 0.01; n = 10), revealing significant attenuation of the NOC-18 effect by KT5823 (Fig. 1G, filled vs. open bars; P < 0.05, Dunnett's multiple comparison test following one-way ANOVA). The specificity of KT5823 at 1 μm to selectively inhibit activation of PKG but not that of cAMP-dependent protein kinase (PKA) has been verified in our recent study (Chai & Lin, 2010). These data thus indicate that NOC-18 stimulated Kir6.2/SUR2A channels in intact HEK293 cells primarily via activation of PKG.

Effects of ROS scavengers and catalase on Kir6.2/SUR2A channel stimulation by NO induction

ROS are identified as important mediators in intracellular signalling (Dröge, 2002; Finkel, 2011). The NO donor S-nitroso-N-acetyl penicillamine (SNAP) has been shown to induce ROS generation in isolated rat cardiomyocytes (Xu et al. 2004). Are ROS involved in cardiac K_ATP channel stimulation by NO? We evaluated this possibility by examining whether ROS removal affects the action of NO donors on Kir6.2/SUR2A channels. Following pretreatment for at least 15 min, MPG (500 μm; an ROS scavenger) was applied together with NOC-18 (300 μm) to cell-attached patches obtained from transfected HEK293 cells. Coapplication of NOC-18 and MPG did not alter the single-channel currents of Kir6.2/SUR2A channels (Fig. 1C and G, third bar from left), in sharp contrast to the increase rendered by NOC-18 when applied alone (Fig. 1G, filled vs. third bars; P < 0.01). We also examined the effect of the H₂O₂-decomposing enzyme catalase on NO donor-induced channel stimulation. H₂O₂ is a relatively stable form of ROS, an attractive candidate for cell signalling (Scherz-Shouval & Elazar, 2007). In the presence of catalase (500 U ml⁻¹), which provides a sink for endogenously generated H₂O₂, NOC-18 (300 μm) failed to elevate Kir6.2/SUR2A channel activity (Fig. 1D and G, fourth bar from left), showing nearly complete blockade of the NOC-18 effect (Fig. 1G, filled vs. fourth bars; P < 0.01). These data indicate that ROS, and especially H₂O₂, were indispensible signals for NO stimulation of cardiac-type K_ATP channels in intact HEK293 cells.

Inhibition of ERK1/2 abrogates Kir6.2/SUR2A channel stimulation by NO induction

ERK1/2, a member of the MAPK family, is ubiquitously expressed and has many diverse cellular and physiological functions (Rose et al. 2010). ERK1/2 may be activated by H₂O₂ (Nishida et al. 2000). We showed above that NO stimulation of Kir6.2/SUR2A channels required ROS/H₂O₂; however, little is known about whether ERK plays a signalling role in acute NO modulation of ion channel function. To address this question, following pretreatment with U0126, which blocks activation of ERK1/2 through selectively inhibiting MEK1 and MEK2, cell-attached recordings were conducted in the continuous presence of U0126. Intriguingly, we found that NOC-18 (300 μm) was incapable of facilitating Kir6.2/SUR2A channel opening when U0126 (10 μm) was coadministered (Fig. 1E and G, fifth bar from left); that is, the increase in the normalized NP_o by NOC-18 was abrogated by blocking ERK1/2 activation (Fig. 1G, filled vs. fifth bars; P < 0.01). These data indicate that ERK1/2, presumably activated downstream of ROS, was required for NO stimulation of cardiac-type K_ATP channels.

Effect of CaMKII inhibitory peptides on NO stimulation of Kir6.2/SUR2A channels

Calcium/calmodulin-dependent kinases (CaMKs) influence processes as diverse as gene transcription, cell survival, apoptosis, cytoskeletal reorganization and learning and memory. CaMKII is the CaMK isoform predominantly found in the heart (Maier, 2009). Nevertheless, the potential involvement of CaMKII in NO signalling for cardiac K_ATP channel modulation has never been investigated. In this set of experiments, we tested whether blocking CaMKII activation with mAIP (1 μm), a myristoylated autocamtide-2 related inhibitory peptide for CaMKII, interferes with Kir6.2/SUR2A channel stimulation induced by NOC-18 (300 μm). Subsequent to 15 min pretreatment with mAIP, coapplication of NOC-18 and mATP resulted in no significant change in the activity of Kir6.2/SUR2A channels acquired in cell-attached patches (Fig. 1F and G, sixth bar from left), uncovering that mAIP nullified the stimulatory action of NOC-18 (Fig. 1G, filled vs. sixth bars; P < 0.01). These results thus indicate that NO modulation of Kir6.2/SUR2A channels in intact HEK293 cells relied on activation of CaMKII.

Effect of NO induction on sarcK_ATP channels in intact rabbit ventricular myocytes: the dependence on sGC and PKG

To evaluate the physiological relevance of NO signalling in cardiac K_ATP channel modulation, cell-attached recordings as performed on HEK293 cells were conducted on ventricular cardiomyocytes freshly isolated from adult rabbits. In these native cells, pinacidil (100–200 μm), a KCO, was applied first to induce baseline sarcK_ATP channel activity comparable to that seen in transfected HEK293 cells. The NO donors glyco-SNAP-2 (300 μm; Fig. 2A) and NOC-18 (300 μm; Fig. 2B) were then added, and both evoked marked increases in the opening and bursting frequencies and the bursting duration of ventricular sarcK_ATP channels; the normalized NP_o was raised to 8.29 ± 2.71 (control value in pinacidil taken as one; Fig. 2E, grey bar; P < 0.05) and 5.79 ± 1.51 (Fig. 2E, filled bar in black; P < 0.01), respectively, whereas the single-channel conductance remained unchanged. Moreover, to ensure that the stimulatory effect of NO induction on the normalized single-channel activity of rabbit ventricular sarcK_ATP channels is not biased toward increases due to the low basal activity in the cell-attached patch configuration, the absolute NP_o (i.e. NP_o without normalization) values obtained in control and NOC-18-treated conditions were directly compared (Supplemental Fig. S1; a scatter plot). The averaged absolute NP_o values were significantly increased, manifesting a positive effect of NOC-18 (nine data pairs; P < 0.05); the shift in the median points (Supplemental Fig. S1, golden bars) was also consistent with an upward change caused by NOC-18. These results thus indicate that NO induction stimulated pinacidil-preactivated sarcK_ATP channels in native ventricular cardiomyocytes, reinforcing our findings made on recombinant cardiac K_ATP channels. By contrast, NOC-18 did not increase sarcK_ATP channel activity in excised, inside-out patches (data not shown), excluding the possibility that the stimulation results from direct chemical modification of the channel by NO.

Figure 2.

A–D, representative single-channel current traces of ventricular sarcK_ATP channels induced by pinacidil (200 μm) in cell-attached patches obtained from rabbit cardiomyocytes before and during addition of glycol-SNAP-2 (300 μm; A), NOC-18 (300 μm; B), or NOC-18 plus 1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one (ODQ, 50 μm; C) or KT5823 (1 μm; D), illustrating that NO donors enhance ventricular sarcK_ATP channel activity but the enhancement is reversed in the presence of inhibitors selective for sGC or PKG. Recording settings and scale bars are the same as described in the legend to Fig. 1. E, averaged, normalized NP_o in individual groups of cell-attached patches (n = 4–12), showing that the significant increase of sarcK_ATP single-channel activity in intact ventricular cardiomyocytes induced by NO donors is abolished by inhibition of sGC or PKG. *P < 0.05; ^**P < 0.01 (Student's one-sample t test within groups, and one-way ANOVA followed by Dunnett's multiple comparison tests among groups).

To identify signalling partners involved in NO modulation of the channel in native cardiomyocytes, we next examined whether NO modulation of ventricular sarcK_ATP channels requires activation of sGC and PKG, by applying NOC-18 (300 μm) together with the selective sGC inhibitor ODQ (50 μm) or the PKG inhibitor KT5823 (1 μm), following pretreatment with respective inhibitors. The NOC-18 did not potentiate the single-channel activity of sarcK_ATP channels preactivated by pinacidil in the presence of ODQ (Fig. 2C and E, open bar) or KT5823 (Fig. 2D and E, hatched bar), revealing annihilation of the stimulatory effect of NO donors (Fig. 2E, P < 0.05 vs. filled bar in black). These results indicate that NO induction was capable of enhancing the function of sarcK_ATP channels in native ventricular cardiomyocytes and that the enhancement was sGC-and PKG-dependent.

Suppression of ERK1/2 activity obliterates sarcK_ATP channel stimulation elicited by NO donors in intact ventricular cardiomyocytes

Our findings obtained from the cloned K_ATP channel Kir6.2/SUR2A expressed in HEK293 cells (see Fig. 1) revealed, for the first time, that ERK1/2 was required for NO modulation of cardiac K_ATP channels. To substantiate these findings in a native cell setting, cell-attached patch-clamp recordings were conducted on rabbit ventricular myocytes pretreated with the ERK1/2 inhibitor U0126. Application of NOC-18 (300 μm) in the continuous presence of U0126 (10 μm) failed to elevate pinacidil-preactivated sarcK_ATP single-channel activity (Fig. 3A and E, open bar); the increase in the normalized NP_o induced by NOC-18 was completely abolished (Fig. 3E, filled vs. open bars; P < 0.05). Likewise, in ventricular myocytes pretreated with PD98059, another ERK1/2 inhibitor, NOC-18 was unable to stimulate sarcK_ATP channels when PD98059 (20 μm) was coapplied (Fig. 3B and E, third bar from left; P < 0.05 vs. filled bar). These data consistently supported our hypothesis that activation of ERK1/2 mediates NO stimulation of sarcK_ATP channels in ventricular myocytes.

Figure 3.

A–D, representative single-channel current traces of pinacidil-preactivated sarcK_ATP channels in cell-attached patches before and during addition of NOC-18 (300 μm) together with one of the following inhibitors: U0126 (10 μm; A); PD98059 (20 μm; B); SKF-7171A (10 μm; C); or mAIP (1 μm; D), illustrating that the stimulatory effect of NOC-18 on native ventricular sarcK_ATP channels is nullified when ERK1/2, calmodulin or CaMKII activity is suppressed. See Fig. 1 for definition of scale bars. E, summary data of the averaged normalized NP_o obtained in individual groups of cell-attached patches (n = 4–12), demonstrating that stimulation of sarcK_ATP channels by NO induction in intact ventricular cardiomyocytes requires activities of ERK1/2, calmodulin and CaMKII. The NOC-18 group data, the same as those shown in Fig. 2, are included here for comparison purposes. *P < 0.05; ^**P < 0.01 (Student's one-sample t test within groups, and one-way ANOVA followed by Dunnett's multiple comparison tests among groups).

Effects of antagonizing calmodulin and CaMKII on ventricular sarcK_ATP channel stimulation caused by NO donors

To define the roles played by calmodulin (a ubiquitous calcium-binding protein) and CaMKII (activation of which depends on Ca²⁺/calmodulin binding) for sarcK_ATP channel stimulation elicited by NO in ventricular cardiomyocytes, SKF-7171A, a selective calmodulin antagonist, and mAIP, the membrane-permeable inhibitory peptide for CaMKII, were respectively coapplied with NOC-18 during cell-attached patch-clamp recordings (following pretreatment). When coapplied with SKF-7171A (10 μm; Fig. 3C) or mAIP (1 μm; Fig. 3D), NOC-18 (300 μm) did not enhance ventricular sarcK_ATP channel currents preactivated by pinacidil (Fig. 3E, fourth and fifth bars from left), yielding significant abrogation of the stimulatory effect of NOC-18 (Fig. 3E; P < 0.05 vs. filled bar for both groups). In agreement with the findings made in HEK293 cells (see Fig. 1), these results indicate that the stimulatory action of NO induction on ventricular sarcK_ATP channels required activation of calmodulin and CaMKII.

Inhibition of ERK and CaMKII abolishes potentiation of sarcK_ATP channel activity rendered by exogenous H₂O₂ in ventricular cardiomyocytes

We showed in the preceding subsections that inhibition of ROS/H₂O₂, ERK and CaMKII could blunt functional stimulation of ventricular K_ATP channels induced by NO donors in intact cells, revealing the involvement of these molecules as intracellular signalling partners mediating K_ATP channel stimulation downstream of NO (induction). It is important to determine how ERK1/2 and CaMKII are positioned relative to ROS in the NO signalling pathway that enhances K_ATP channel function. To address this, we examined whether the ability of exogenous H₂O₂ to stimulate ventricular K_ATP channels in intact cells is affected by inhibition of ERK1/2 and CaMKII (Supplemental Fig. S2). The rationale is as follows. If H₂O₂ is generated endogenously after, and hence positioned downstream of, activation of ERK1/2 and CaMKII, the effectiveness of exogenous H₂O₂ to stimulate sarcK_ATP channels should not be compromised by suppression of either kinase. The same outcome is expected in the event that H₂O₂ modulates sarcK_ATP channels independently of these kinases. Conversely, if H₂O₂ stimulates sarcK_ATP channels via activation of ERK and/or CaMKII, the K_ATP channel-potentiating capability of exogenous H₂O₂ ought to be hampered by functional suppression of respective kinases. Interestingly, while application of H₂O₂ (1 mm) reliably enhanced sarcK_ATP single-channel activity preactivated by pinacidil in cell-attached patches obtained from rabbit ventricular cardiomyocytes, H₂O₂ failed to elicit changes in K_ATP channel activity when the MEK1/2 inhibitor U0126 (10 μm) or the CaMKII inhibitory peptide mAIP (1 μm) was coapplied (Supplemental Fig. S2), revealing total abolition of the stimulatory action of H₂O₂ by inhibition of ERK1/2 and CaMKII (P < 0.05 vs. H₂O₂ applied without kinase inhibitors). These results indicate that both ERK1/2 and CaMKII were crucial for exogenous H₂O₂ to potentiate ventricular K_ATP channel activity successfully, hence placing ERK1/2 and CaMKII downstream of H₂O₂ for stimulation of K_ATP channels in intact ventricular cardiomyocyes.

Effects exerted by NO signalling on ventricular sarcK_ATP single-channel open and closed properties

Our foregoing data indicate that NO donors enhanced the activity of ventricular K_ATP channels via intracellular signalling. To delineate whether NO signalling affects the gating (i.e. opening and closing) of ventricular sarcK_ATP channels, we analysed K_ATP single-channel activity to determine whether the NO donor NOC-18 causes more frequent entry into the open state (i.e. increases the opening frequency), prolongs stay in the open state (i.e. increases the open time constant of certain open state), decreases dwelling time in the closed states (i.e. decreases the closed time constant of certain closed state), stabilizes or destabilizes the occurrence of a particular state (i.e. shifts the relative distribution among states) or induces any combination of the above. The fitting results revealed that in the control condition, the open-and closed-duration distributions of rabbit ventricular sarcK_ATP channels in the cell-attached patch configuration could be described best by a sum of two open components and a sum of four closed components, respectively (Fig. 4A, control; a representative patch), implying that there are at least two open states and four closed states. Moreover, NOC-18 treatment altered the closed duration distribution (Fig. 4A, closed; top vs. bottom panels); the relative areas and/or the time constants under the longer and longest closed states were reduced [Fig. 4A, inset; magenta colour (depicting NOC-18-treated condition) vs. black (depicting control)], while the shorter closed states were stabilized, resulting in shortening of the mean closed duration to 231.1 from 734.3 ms in this representative patch. Indeed, pooled data showed that NOC-18 decreased the normalized mean closed duration (control taken as one; 0.31 ± 0.07; P < 0.0001; n = 7) and increased the normalized opening frequency (5.10 ± 1.60; P < 0.05), thereby elevating the normalized NP_o (i.e. relative channel activity; see Figs. 1G and 2E). Meanwhile, the longer open state became increasingly prominent in the presence of NOC-18 (Fig. 4A, open; top vs. bottom panels), although neither the corrected mean open duration (1.65 ms in control condition vs. 1.75 ms in NOC-18) obtained from this patch nor the normalized mean open duration averaged from the whole group (control taken as one; 1.16 ± 0.15; n = 7) was significantly increased. By contrast, NOC-18 failed to evoke similar changes in the opening frequency (data not shown) and the open and closed duration distributions of ventricular sarcK_ATP channels when the PKG inhibitor KT5823 (Fig. 4B), the ERK1/2 inhibitor U0126 (Fig. 4C) or the CaMKII inhibitory peptide mAIP (data not shown) was coadministered, explicating the absence of NOC-18-induced increases in NP_o in these conditions (see Figs. 1G, 2E and 3E). Our findings thus indicate that NO induction potentiated ventricular K_ATP channel activity by shortening and destabilizing long closures, whilst increasing the opening transitions of the channel, in a PKG-, ERK1/2-and CaMKII-dependent manner.

Figure 4.

A–C, frequency histograms of open-duration and closed-duration distributions fitted from events detected before (upper panels) and during (lower panels) application of NOC-18 (300 μm; A), NOC-18 plus KT5823 (1 μm; B) or NOC-18 plus U0126 (10 μm; C) in representative cell-attached patches obtained from rabbit ventricular myocytes. Insets show superimposed curve fittings of duration distributions of the two longer closed components in control (black) versus treatment conditions (colours) to highlight NOC-18 effects. The NOC-18 left-shifts the longest closed component and reduces the relative areas under longer/longest closed components, effectuating destabilization of the longer/longest closed components. By contrast, inhibition of PKG (with KT5823) or ERK1/2 (with U0126) prevents these changes induced by NOC-18 from occurring, which demonstrates that the NO donor effects on duration distributions are mediated by intracellular signalling through activation of PKG and ERK1/2.

Genetic ablation of CaMKIIδ prevents PKG stimulation of ventricular sarcK_ATP channels

The predominant CaMKII isoform in the heart is CaMKIIδ (Tobimatsu & Fujisawa, 1989). To evaluate the role of CaMKIIδ in mediating cGMP/PKG stimulation of cardiac K_ATP channels, a CaMKIIδ-null mouse model (plus littermate controls) was employed. Application of the PKG activator zaprinast (50 μm) to cell-attached patches obtained from wild-type mouse ventricular myocytes significantly enhanced the activity of sarcK_ATP channels preactivated by pinacidil (Fig. 5A and C, filled bar; normalized NP_o = 4.57 ± 1.29; P < 0.05); however, this stimulatory effect was absent in CaMKIIδ-null cardiomyocytes (Fig. 5B and C, open bar); that is, genetic ablation of CaMKIIδ prevented ventricular sarcK_ATP channel stimulation caused by activation of PKG (Fig. 5C, filled vs. open bars; P < 0.01). These results indicate that CaMKIIδ was required for functional enhancement of ventricular sarcK_ATP channels elicited by PKG activation in intact cells, unveiling a previously unrecognized role played by CaMKIIδ. As activation of PKG represented a key step linking NO induction to functional enhancement of cardiac K_ATP channels (see Figs. 1 and 2), the findings obtained from CaMKIIδ-null ventricular cardiomyocytes thus lend additional support to our hypothesis that CaMKII, especially CaMKIIδ, is indispensable in the NO–PKG signalling cascade for functional modulation of myocardial K_ATP channels.

Figure 5.

A–C, electrophysiological analysis of sarcK_ATP channel activity in response to PKG activation in intact ventricular myocytes isolated from CaMKIIδ-null versus littermate/wild-type (WT) mice, showing that genetic ablation of CaMKIIδ obliterates PKG stimulation of ventricular sarcK_ATP channels. Representative single-channel current traces of pinacidil-preactivated sarcK_ATP channels in response to addition of zaprinast (50 μm; PKG activator) in cell-attached patches obtained from the wild-type (A) and CaMKIIδ-null mouse ventricular myocytes (B) illustrate that potentiation of pinacidil-preactivated ventricular sarcK_ATP single-channel activity by zaprinast is obliterated in CaMKIIδ-null mouse cardiomyocytes. Recording settings and scale bars are the same as described in Fig. 1. Summary data (C) obtained from individual groups demonstrate that, compared with wild-type counterparts, the increase in the averaged normalized NP_o (control taken as one; dashed line) by PKG activation is diminished in CaMKIIδ-null ventricular myocytes (n = 7–9). *P < 0.05; ^**P < 0.01 (Student's one-sample t test within groups, and unpaired t test between groups). D and E, biochemical analysis of CaMKII activity, showing that the activity of CaMKII in intact rabbit ventricular myocytes is increased by NO–PKG activation in an ERK1/2-dependent manner. Cardiomyocytes were treated with NOC-18 (300 μm) or zaprinast (50 μm) in the absence and presence of KT5823 (1 μm) or U0126 (10 μm) for 30 min, followed by preparation of cell lysates. The CaMKII activity was then assayed by Western blotting of phospho-CaMKII (p-CaMKII) relative to total CaMKII and by estimating ³²P incorporation of a synthetic CaMKII substrate. Representative Western blots (D) and the mean densitometric values of relative CaMKII activity (E) estimated by ³²P incorporation (filled bars) and by Western blots (p-CaMKII relative to total CaMKII values; open bars; n = 3) reveal that CaMKII activity in cardiomyocytes is elevated by NO induction and PKG activation, but the increase is attenuated when PKG or ERK1/2 activity is inhibited. Values are means ± SEM of three experiments of independent cell preparations. The kinase activity assay was conducted in triplicate each time. *P < 0.05; ^**P < 0.01 (Student's one-sample t test within groups, and one-way ANOVA followed by Dunnett's multiple comparison tests among groups).

Effects of NO induction and PKG activation on CaMKII activity in ventricular myocytes: involvement of ERK1/2

To seek direct evidence for CaMKII activation by NO and PKG in intact cells, two independent biochemical assays, Western blotting that measures autophosphorylation of CaMKII at T287 (p-CaMKII) and a kinase activity assay that detects ³²P-ATP incorporation into syntide-2, a synthetic substrate for CaMKII, were conducted. Isolated adult rabbit ventricular myocytes were treated with the NO donor NOC-18 (300 μm) and the PKG activator zaprinast (50 μm), respectively, for 30 min in the absence and presence of KT5823 (1 μm; PKG inhibitor) or U0126 (10 μm; ERK1/2 inhibitor), followed by preparation of cell lysates for subsequent assays to estimate CaMKII activity. Western blotting assays revealed that both zaprinast and NOC-18 elevated the p-CaMKII level (relative to total CaMKII; Fig. 5D, upper panel, lanes 2 and 4 from left; Fig. 5E, open bars; P < 0.01, Student's two-tailed, one sample t test; control taken as one); however, these increases were attenuated by KT5823 (Fig. 5D, upper panel, lanes 3 and 5 from left; Fig. 5E, open bars; P < 0.01 for NOC-18 vs. NOC-18 + KT5823 and P < 0.05 for zaprinast vs. zaprinast + KT5823, Dunnett's multiple comparison test following one-way ANOVA) and by U0126 (Fig. 5D, lower panel; Fig. 5E, P < 0.01 for zaprinast vs. zaprinast + U0126). In accordance with Western blot data, CaMKII activity measured by ³²P-ATP incorporation was also increased by NOC-18 and by zaprinast (Fig. 5E, filled bars; three independent runs of triplicates each time; P < 0.01 for both treatment groups), and the changes were significantly abated when KT5823 or U0126 was coadministered (Fig. 5E, filled bars; P < 0.01 vs. NOC-18 or zaprinast administered alone). These results indicate that CaMKII was activated by NO–PKG signal transduction in ventricular cardiomyocytes; additionally, the ERK1/2 dependence of CaMKII activation implies that ERK1/2 is likely to be positioned upstream of CaMKII in the signalling cascade triggered by NO–PKG.

Discussion

sGC and PKG are required for NO stimulation of cardiac K_ATP channels

NO represents one of the most important defenses against myocardial ischaemia–reperfusion injury (Jones & Bolli, 2006); meanwhile, the K_ATP channel has been regarded as mandatory in acute and chronic cardiac adaptation to imposed haemodynamic load, protecting against congestive heart failure and death (Yamada et al. 2006). NO may potentiate the action of KCOs on K_ATP channels in ventricular cardiomyocytes (Shinbo & Iijima, 1997; Han et al. 2002) and activate sarcK_ATP channels in normoxic and chronically hypoxic hearts (Baker et al. 2001). However, little is known about the intracellular mechanism responsible for NO modulation of cardiac K_ATP channels. In the present study, we showed that induction of NO by chemical donors resulted in increases in Kir6.2/SUR2A (i.e. recombinant cardiac-type K_ATP) and KCO-induced native sarcK_ATP single-channel activities in cell-attached patches obtained from intact HEK293 cells and ventricular cardiomyocytes, respectively. Moreover, the stimulatory action of NO donors was attenuated or abolished by selective inhibition of sGC and PKG, suggesting that NO induction enhances the function of cardiac K_ATP channels in intact cells via activation of sGC and PKG. In contrast to a K_ATP-potentiating effect observed in intact cells, NO donors did not increase ventricular sarcK_ATP channel activity in excised, inside-out patches (data not shown), which is consistent with a working model that NO modulates K_ATP channel function via intracellular signalling rather than direct chemical modification of the channel.

ROS, in particular H₂O₂, act as intermediate signals in NO-induced stimulation of cardiac K_ATP channels

ROS are generated by all aerobic cells, and most endogenously produced ROS are derived from mitochondrial respiration (Liu et al. 2002). They have been shown to contribute to cardioprotection afforded by ischaemic preconditioning (Baines et al. 1997). Among all ROS, H₂O₂ is an attractive candidate for cell signalling, because it is relatively stable and long lived and its neutral ionic state allows it to exit the mitochondria easily (Scherz-Shouval & Elazar, 2007). In the present study, increases in Kir6.2/SUR2A channel activity rendered by NO donors in intact HEK293 cells were aborted not only by the ROS scavenger MPG but also by the H₂O₂-decomposing enzyme catalase. These results suggest that ROS, and in particular H₂O₂, presumably produced downstream of PKG activation, mediate NO-induced stimulation of cardiac K_ATP channels in intact cells. In line with our findings that support an NO–PKG–ROS signalling model, the NO donor SNAP has been demonstrated to increase ROS generation in isolated cardiomyocytes, which, importantly, requires activation of PKG (Xu et al. 2004). It has also been shown that late and early preconditioning induced by NO donors is blocked by the ROS scavenger MPG, implying that ROS are involved in cardioprotection induced by (exogenous) NO (Takano et al. 1998; Nakano et al. 2000); in light of the present findings, protection by NO in the heart may involve ROS-dependent activation of myocardial sarcK_ATP channels.

In addition to ROS, an involvement of the putative mitochondrial K_ATP (mitoK_ATP) channel in mediating NO stimulation of cell-surface cardiac K_ATP channels was also investigated. Opening of mitoK_ATP channels has been suggested as a downstream event of PKG activation (Xu et al. 2004). Our findings indicate that 5-hydroxydecanoate (5-HD), the specific antagonist for the putative mitoK_ATP channel, significantly attenuated the increase in Kir6.2/SUR2A channel activity rendered by NOC-18 in intact HEK293 cells (Supplemental Fig. S3). The results thus suggest that the mitoK_ATP channel (or ‘the 5-HD-sensitive factor’; see Chai & Lin, 2010), like ROS, is an intermediate signal crucial for mediating functional enhancement of cardiac K_ATP channels caused by NO. Activation of the mitoK_ATP channel and ROS generation may be sequential or parallel events induced by NO. However, because ROS scavengers in intact cells completely abolish the stimulatory effect on cardiac K_ATP channels rendered by NO induction (Fig. 1) and by activation of PKG (Chai et al. 2011), whereas the stimulatory effect of exogenous H₂O₂ on cell-surface K_ATP channels is unaffected by 5-HD treatment (Chai & Lin, 2010), it is conceivable that the mitoK_ATP channel or the 5-HD-sensitive factor is positioned upstream of, not in parallel to, ROS/H₂O₂ (generation) for K_ATP channel modulation in the NO–PKG signalling pathway. Collectively, these results support our working model (Fig. 6), where the putative mitoK_ATP channel mediates ROS generation induced by NO induction to stimulate cell-surface K_ATP channel activity. MitoK_ATP channels and ROS are implicated in the cardioprotective effect of ischaemic preconditioning (Vanden Hoek et al. 1998; Pain et al. 2000) and the anti-infarct effect of NO in intact, isolated heart (Xu et al. 2004). It is possible that NO exerts its cardiac protection by activating sarcK_ATP channels via a PKG–mitoK_ATP–ROS signalling mechanism.

Figure 6.

Based on evidence obtained from the present study, we suggest that induction of NO leads to sGC activation and cGMP generation, which in turn activates PKG and triggers downstream signalling that consists of (in sequence) ROS, ERK1/2, calmodulin and CaMKII, resulting in sarcK_ATP channel stimulation. Signalling components involved are shown in rectangular or oval shapes (shaded); pharmacological reagents or genetic ablation employed in the present study targeting individual signalling components are also depicted, with inhibitory approaches positioned on the left and activators on the right.

ERK1/2 mediates NO-and H₂O₂-induced stimulation of cardiac K_ATP channels

ERKs play pivotal roles in many aspects of cell functions and are activated by oxidative stress in some types of cells (Aikawa et al. 1997; Nishida et al. 2000). Our present investigation revealed that increases in cardiac K_ATP single-channel activity induced by NO donors in both ventricular cardiomyocytes and transfected HEK293 cells were abolished by inhibition of MEK1 and MEK2 (both upstream kinases of ERK1/2) with U0126 or PD98059. These results thus suggest that, like ROS, ERK1/2 is a key relay signal evoked by NO to mediate cardiac K_ATP channel stimulation. But what is the relationship between ROS and ERK in the NO–K_ATP channel signalling pathway?

Most aspects of oxidant signalling have been linked to the more stable derivative, H₂O₂ (Finkel, 2003). It has been reported that in cardiac myocytes, ERKs are activated by H₂O₂ transiently and in a concentration-dependent manner (Aikawa et al. 1997). H₂O₂ may regulate K_ATP channel activity in ventricular cardiomyocytes (Goldhaber et al. 1989; Ichinari et al. 1996; Tokube et al. 1996). Befittingly, exogenous H₂O₂ enhances the single-channel activity of pinacidil-preactivated sarcK_ATP channels in a concentration-dependent manner in intact rabbit ventricular myocytes (Chai et al. 2011). In the present study, we found that the stimulatory action of exogenous H₂O₂ on sarcK_ATP channels in intact cardiomyocytes was abrogated when the ERK1/2 inhibitor U0126 was coapplied (Supplemental Fig. S2). These results suggest that ERK1/2 is positioned downstream of H₂O₂ to mediate H₂O₂-induced sarcK_ATP channel stimulation in ventricular cardiomyocytes. Complementing evidence presented in the foregoing subsections that ROS/H₂O₂ and ERK1/2 were required for NO stimulation of cardiac K_ATP channels, it is therefore conceivable that activation of ERK1/2 takes place following ROS generation in the NO–K_ATP channel signalling cascade. Indeed, this hypothesis is compatible with biochemical evidence demonstrated by Xu et al. (2004) using isolated cardiomyocytes that the NO donor SNAP enhances phosphorylation of ERK in a ROS scavenger-sensitive manner, which suggests phosphorylation/activation of ERK as the downstream signalling event of NO-induced ROS generation. Collectively, our data suggest that ROS/H₂O₂ activates ERK1/2 in the intracellular signalling cascade initiated by NO induction, leading to ventricular sarcK_ATP channel stimulation.

Calmodulin and CaMKII are indispensible for stimulation of cardiac K_ATP channels induced by NO and H₂O₂

CaMKII is one of the major regulators of Ca²⁺ homeostasis in the heart, phosphorylating cardiac contractile regulatory proteins and modulating the function of cardiac ion channels (Zhang et al. 2004; Wagner et al. 2009). Binding of Ca²⁺/calmodulin activates CaMKII, by disinhibiting the autoregulatory domain of the kinase (Hudmon & Schulman, 2002). We showed in the present study that potentiation of pinacidil-preactivated sarcK_ATP channels by NO donors in ventricular cardiomyocytes was diminished by both mAIP, a cell-permeable, inhibitory peptide selective for CaMKII, and SKF-7171A, a potent and irreversible calmodulin antagonist; likewise, mAIP treatment abolished NO donor-induced stimulation of recombinant Kir6.2/SUR2A channels expressed in HEK293 cells. These results coherently suggest that NO induction enhances cardiac K_ATP channel function via activation of calmodulin and CaMKII. By contrast, application of CaMKII to excised, inside-out patches did not reproduce the positive action of NO donors on ventricular sarcK_ATP channel activity (data not shown); it thus seemed unlikely that direct CaMKII phosphorylation of the channel protein is responsible for NO potentiation of K_ATP channel function in intact cells. Additionally, we demonstrated that the increase in ventricular sarcK_ATP channel activity rendered by exogenous H₂O₂ was reversed by mAIP in intact cardiomyocytes (Supplemental Fig. S2), implying that activation of CaMKII mediates the stimulatory effect of exogenous H₂O₂. Taken together, these results suggest that CaMKII is positioned downstream of ROS/H₂O₂ in the NO signalling pathway to mediate functional enhancement of cardiac K_ATP channels.

On the other hand, activation of CaMKII has recently been reported to promote internalization (endocytosis) of cardiac K_ATP channels, reducing surface expression (Sierra et al. 2013). It is possible that, through different downstream mechanisms, activity and surface expression of cardiac K_ATP channels are differentially regulated by activation of CaMKII, as previously reported for cardiac inwardly rectifying potassium channels, IRK (i.e. cardiac Kir2.x channels that give rise to I_K1 currents; Wagner et al. 2009). Notably, for IRK channels the increase in function predominates over the reduction in expression when CaMKII is activated (Wagner et al. 2009), resulting in an overall effect of channel stimulation. Our findings evidently support a working model where calmodulin and CaMKII serve as indispensible elements in the NO signalling pathway mediating functional enhancement, not suppression, of cardiac K_ATP channels.

Involvement of CaMKIIδ

The CaMKII family consists of four closely related yet distinct isoforms (α, β, γ and δ). The major isoform of CaMKII in the heart is CaMKIIδ (Tobimatsu & Fujisawa, 1989). Importantly, the present study revealed that genetic ablation of CaMKIIδ (i.e. CaMKIIδ knockout) diminished PKG stimulation of ventricular sarcK_ATP channels, suggesting a crucial role of CaMKIIδ in mediating enhancement of ventricular sarcK_ATP channel activity elicited by PKG activation. As PKG activation was required for NO stimulation of cardiac K_ATP channels, these results thus suggest that CaMKIIδ is primarily responsible for functional effects rendered by NO elevation on sarcK_ATP channels in intact ventricular myocytes. Increased short-term CaMKII activity may serve as beneficial negative feedback for calcium on repolarization of cardiomyocyte membranes (Wagner et al. 2009). Further study is required to identify the direct target(s) of CaMKII(δ) for K_ATP channel stimulation.

Activation of NO signalling modifies the open and closed properties of ventricular sarcK_ATP channels to potentiate channel activity

Based on the open-and closed-duration distributions of sarcK_ATP channels in intact rabbit ventricular cardiomyocytes, we suggest that the cardiac K_ATP channel exhibits at least two open states and four closed states. The enhanced K_ATP channel activity (as evidenced by higher NP_o values) observed in the presence of NO donors could be accounted for by an increase in the opening frequency and by shifts in the closed-duration distributions, the latter of which included reductions in the occurrence (i.e. the relative area of individual exponential components shown in the frequency histogram) of the two longer closed states relative to that of the two shorter ones, and a shortened dwelling duration (i.e. the time constant) of the longest closed state. These results suggest that NO potentiates ventricular sarcK_ATP channel activity by destabilizing the long closed conformations and by facilitating the closed-to-open transitions. Importantly, the aforementioned changes caused by NO donors in the channel open and closed properties were prevented by the PKG inhibitor KT5823, by the MEK1/2 inhibitor U0126 and by the CaMKII inhibitory peptide mAIP, suggesting the involvement of PKG, ERK1/2 and CaMKII as molecular transducers in mediating the effect of NO on cardiac K_ATP channel gating.

NO–PKG signalling augments cardiac CaMKII activity in an ERK1/2-dependent manner

Calcium/calmodulin binding activates CaMKII by disinhibiting the autoregulatory domain, which initiates intraholoenzyme autophosphorylation. Autophosphorylation of CaMKII at T287 produces Ca²⁺-autonomous activity by preventing reassociation of the kinase domain by the autoinhibitory region (Hudmon & Schulman, 2002). Our biochemical evidence revealed that both the PKG activator zaprinast and the NO donor NOC-18 activated CaMKII in intact rabbit ventricular cardiomyocytes, as manifested by increases in autophosphorylation of CaMKII and incorporation of ³²P into CaMKII substrates. Importantly, activation of CaMKII induced by NOC-18 and by zaprinast was significantly attenuated by the PKG inhibitor KT5823, suggesting that CaMKII is activated by NO–PKG signal transduction in ventricular cardiomyocytes. Moreover, enhancement of CaMKII activity by zaprinast was reduced in the presence of the MEK1/2 inhibitor U0126, which further suggests that ERK1/2 mediates PKG-elicited activation of CaMKII, hereupon placing CaMKII downstream of ERK1/2 in the signalling cascade initiated by NO–PKG. In addition, we also examined the effect of coapplication of NOC-18 and zaprinast on CaMKII phosphorylation. Data obtained from this group revealed that coapplication of NOC-18 and zaprinast increased CaMKII phosphorylation (Supplemental Fig. S4; n = 3), but the magnitude of increase did not exceed that rendered by zaprinast administered alone (see Fig. 5D and E). These results thus suggest that PKG and NO act through the same signalling mechanism to enhance CaMKII activity in cardiomyocytes, providing additional evidence supportive of our hypothesis that PKG mediates stimulation of CaMKII activity caused by NO. While H₂O₂ can directly drive autonomous CaMKII activation in a Ca²⁺/calmodulin-dependent manner (Erickson et al. 2008), our electrophysiological data showing that cardiac K_ATP channel stimulation by exogenous H₂O₂ and by NO donors was both abrogated by inhibition of ERK1/2, complemented by biochemical evidence discussed above, suggest that ERK is likely to be positioned downstream of ROS/H₂O₂ but upstream of CaMKII in the NO signalling pathway, at least for cardiac K_ATP channel modulation. In other words, these results collectively support a working model (see Fig. 6), in which Ca²⁺/calmodulin-dependent activation of CaMKII takes place after sequential activation of NO (induction), sGC, PKG, ROS/H₂O₂ (generation) and ERK1/2 to mediate cardiac K_ATP channel stimulation. In this NO–K_ATP channel signalling pathway, the ability of ROS to activate CaMKII directly (Erickson et al. 2008) appears to be non-essential.

The residual effect caused by NO donors on K_ATP channel potentiation in the presence of KT5823 observed in HEK293 cells (Fig. 1B and G) seemed to imply that in HEK293 cells, but not in ventricular cardiomyocytes, some yet-to-be-identified signal(s) besides PKG is also activated by NO induction to mediate K_ATP channel stimulation. Even if NO induces PKG-independent signalling in addition to activation of PKG, the ‘divergent’ signals probably converge to one common pathway at or above the level of ROS in HEK293 cells, as evidenced by total abrogation of the NO donor effect by scavenging of ROS, or respective suppression of the more downstream signalling partners ERK1/2 and CaMKII (Fig. 1C–F). It is worth mentioning that many of the intermediate signals required for mediating K_ATP channel potentiation in the signalling mechanism proposed in this study (Fig. 6) may intersect with other signalling pathways in different intracellular conditions, and therefore, our findings do not exclude a possibility that the signalling molecules involved in K_ATP channel modulation downstream of NO may also affect K_ATP channel activity through some parallel signalling pathways. Further studies will be required to elucidate this possibility.

In conclusion, here we report, for the first time, that the function of ventricular sarcK_ATP channels is modulated by NO induction via an intracellular signalling pathway consisting of sGC, PKG, ROS/H₂O₂, ERK1/2, calmodulin and CaMKII (CaMKIIδ in particular) that facilitates opening transitions while destabilizing long closures of the channel. Specifically, our study suggests that ERK1/2 mediates NO/PKG activation of CaMKII, thereby relaying the signal from elevation of NO (and ROS) to the sarcK_ATP channel in cardiomyocytes, rendering heightened channel activity. The present study highlights the relevance of intracellular signalling mechanisms as effective functional regulators for K_ATP channels. The signalling mechanism described herein may offer the framework to permit fine-tuning of K_ATP channel activity in different intracellular conditions. Mechanistic understanding of K_ATP channel regulation may provide insights into the development of strategies for the management of cardiovascular injury. It is noteworthy that K_ATP channels, NO, PKG, ROS and ERK1/2 have also been implicated in cardiac protection/tolerance against ischaemic injury. Cardiac protection by NO from exogenous sources or endogenously released during the short episode of sublethal ischaemia may be mediated partly by K_ATP channel stimulation. Hence, this NO–sGC–PKG–ROS–ERK1/2–calmodulin–CaMKII (CaMKIIδ in particular)–sarcK_ATP signalling pathway may regulate cardiomyocyte excitability and contribute to endogenous cytoprotection in the heart.

Key points

• Both the ATP-sensitive potassium (K_ATP) channel and the gaseous messenger nitric oxide (NO) play fundamental roles in protecting the heart from injuries related to ischaemia.

• NO has previously been suggested to modulate cardiac K_ATP channels; however, the underlying mechanism remains largely unknown.

• In this study, by performing electrophysiological and biochemical assays, we demonstrate that NO potentiation of K_ATP channel activity in ventricular cardiomyocytes is prevented by pharmacological inhibition of soluble guanylyl cyclase (sGC), cGMP-dependent protein kinase (PKG), Ca²⁺/calmodulin-dependent protein kinase II (CaMKII) and extracellular signal-regulated protein kinase 1/2 (ERK1/2), by removal of reactive oxygen species and by genetic disruption of CaMKIIδ.

• These results suggest that NO modulates cardiac K_ATP channels via a novel cGMP–sGC–cGMP–PKG–ROS–ERK1/2–calmodulin–CaMKII (δ isoform in particular) signalling cascade.

• This novel intracellular signalling pathway may regulate the excitability of heart cells and provide protection against ischaemic or hypoxic injury, by opening the cardioprotective K_ATP channels.

Glossary

APD₉₀

action potential duration at 90% repolarization

CaMKII

calcium/calmodulin-dependent protein kinase II

E_K

equilibrium potential for potassium

ERK

extracellular signal-regulated kinase

5-HD

5-hydroxydecanoate

HEK293

human embryonic kidney 293 (cell line)

H₂O₂

hydrogen peroxide

IRK

inwardly rectifying Kir2.x (channel)

K_ATP

ATP-sensitive potassium (channel)

KCO

potassium channel opener

Kir

inwardly rectifying potassium (channel)

mAIP

myristoylated autocamtide-2 related inhibitory peptide selective for CaMKII

MAPK

mitogen-activated protein kinase or MAP kinase

MEK

mitogen-activated protein kinase kinase or MAPK kinase

mitoK_ATP

mitochondrial K_ATP (channel)

MPG

N-(2-mercaptopropionyl)glycine

NO

nitric oxide

NOC-18

DETA NONOate

NP_o

open probability

ODQ

1H-[1,2,4]oxadiazolo[4,3-a]quinoxalin-1-one

p-CaMKII

autophosphorylated CaMKII

PIP₂

phosphatidylinositol-4,5-bisphophate

PKA

cAMP-dependent protein kinase

PKG

cGMP-dependent protein kinase

ROS

reactive oxygen species

sarcK_ATP

sarcolemmal K_ATP

sGC

soluble guanylyl cyclase

glycol-SNAP-2

N-(2-deoxy-α,β-d-glucopyranose-2-)-N²-acetyl-S-nitroso-d,l-penicillaminamide

SNAP

S-nitroso-N-acetyl penicillamine

SUR

sulfonylurea receptor

V_m

membrane potential

Additional Information

Competing interests

None declared.

Author contributions

The experiments in this study were conducted in the University of California Davis at the Department of Physiology and Membrane Biology as well as at the Department of Pharmacology. Y.-F.L. directed the study, contributing to the conception and design of the experiments, analysis and interpretation of the data and drafting of the manuscript. Y.C., D.-M.Z. and J.R.E. contributed to the collection and analysis of the data. D.-M.Z., D.M.B. and J.H.B. critically reviewed the manuscript and contributed to important intellectual content. All authors approved the final version of the manuscript.

Funding

This study was supported by funds from the American Heart Association and the UC Davis to Y.-F.L., the National Institutes of Health to D.M.B. and J.H.B. and Fondation Leducq to D.M.B.

Authors’ present addresses

Y. Chai: Institute of Biophysics, Chinese Academy of Science, Beijing, PR China.

J. R. Erickson: Department of Physiology, Otago School of Medical Sciences, University of Otago, Dunedin, New Zealand.

D.-M. Zhang: Department of Cardiology, Nanjing First Hospital Affiliated to Nanjing Medical University, Nanjing, Jiangsu, PR China.